Volume 4, Issue 3 e261
REVIEW
Open Access

Protein posttranslational modifications in health and diseases: Functions, regulatory mechanisms, and therapeutic implications

Qian Zhong

Qian Zhong

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Qian Zhong, Xina Xiao, Yijie Qiu, Zhiqiang Xu, and Chunyu Chen contributed equally to this work.

Search for more papers by this author
Xina Xiao

Xina Xiao

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Qian Zhong, Xina Xiao, Yijie Qiu, Zhiqiang Xu, and Chunyu Chen contributed equally to this work.

Search for more papers by this author
Yijie Qiu

Yijie Qiu

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Qian Zhong, Xina Xiao, Yijie Qiu, Zhiqiang Xu, and Chunyu Chen contributed equally to this work.

Search for more papers by this author
Zhiqiang Xu

Zhiqiang Xu

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Qian Zhong, Xina Xiao, Yijie Qiu, Zhiqiang Xu, and Chunyu Chen contributed equally to this work.

Search for more papers by this author
Chunyu Chen

Chunyu Chen

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Qian Zhong, Xina Xiao, Yijie Qiu, Zhiqiang Xu, and Chunyu Chen contributed equally to this work.

Search for more papers by this author
Baochen Chong

Baochen Chong

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Search for more papers by this author
Xinjun Zhao

Xinjun Zhao

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Search for more papers by this author
Shan Hai

Shan Hai

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Search for more papers by this author
Shuangqing Li

Shuangqing Li

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Search for more papers by this author
Zhenmei An

Zhenmei An

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Search for more papers by this author
Lunzhi Dai

Corresponding Author

Lunzhi Dai

Department of Endocrinology and Metabolism, General Practice Ward/International Medical Center Ward, General Practice Medical Center and National Clinical Research Center for Geriatrics, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, China

Correspondence

Lunzhi Dai, National Clinical Research Center for Geriatrics and General Practice Ward/International Medical Center Ward, General Practice Medical Center, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu 610041, China.

Email: [email protected]

Search for more papers by this author
First published: 02 May 2023
Citations: 17

Abstract

Protein posttranslational modifications (PTMs) refer to the breaking or generation of covalent bonds on the backbones or amino acid side chains of proteins and expand the diversity of proteins, which provides the basis for the emergence of organismal complexity. To date, more than 650 types of protein modifications, such as the most well-known phosphorylation, ubiquitination, glycosylation, methylation, SUMOylation, short-chain and long-chain acylation modifications, redox modifications, and irreversible modifications, have been described, and the inventory is still increasing. By changing the protein conformation, localization, activity, stability, charges, and interactions with other biomolecules, PTMs ultimately alter the phenotypes and biological processes of cells. The homeostasis of protein modifications is important to human health. Abnormal PTMs may cause changes in protein properties and loss of protein functions, which are closely related to the occurrence and development of various diseases. In this review, we systematically introduce the characteristics, regulatory mechanisms, and functions of various PTMs in health and diseases. In addition, the therapeutic prospects in various diseases by targeting PTMs and associated regulatory enzymes are also summarized. This work will deepen the understanding of protein modifications in health and diseases and promote the discovery of diagnostic and prognostic markers and drug targets for diseases.

1 INTRODUCTION

Protein posttranslational modifications (PTMs) refer to the breaking or generation of covalent bonds on the backbones or amino acid side chains of proteins and are also called covalent modifications.1-4 By covalently modifying proteins, cells, tissues, and biological individuals expand the chemical composition and information of twenty amino acids. PTMs escape from genetic confinement in nature.5 Rapid changes in gene sequences on evolutionary timescales are not suitable for organisms to develop and survive.6 PTMs can dynamically change the properties of amino acids according to the requirements on developmental 00and physiological timescales.7 Consequently, numerous PTMs lead to an explosion in the number of proteins with potential molecular states, which provides the basis for the emergence of organismal complexity.8 More than 650 types of protein modifications, such as the most well-known phosphorylation, acetylation, methylation, ubiquitination, glycosylation, acylation, cysteine oxidation, SUMOylation, ADP-ribosylation, neddylation, citrullination, and carbamylation, have been described to date (http://www.uniprot.org/docs/ptmlist.txt), and the inventory is still increasing.9, 10

The PTM process is divided into the following classes (Figure 1). First, modifiers such as small chemicals and complex biomolecules are added to the amino acid side chains. Small chemicals such as phosphate, sugar, methyl group, and acetyl group are usually electrophilic. In contrast, the amino acid side chains that receive modifiers are usually rich in electrons and act as nucleophiles during the modification process, such as lysine and cysteine side chains.11 Second, there are changes in the chemical properties of amino acids, such as deamination, deamidation, citrullination, and oxidation.12 Notably, some types of redox modifications can also be recognized as the addition of small chemicals onto the side chain of cysteine such as S-nitrosylation (SNO) and S-glutathionylation. Third, the cleavage of protein backbone. This process can be conducted by enzyme catalysis or by protein autocatalysis. The cleavage process controls protein localization in or around the cell, protein activity, and protein turnover.11 Most PTMs are dynamically reversible, and the addition and removal of these PTMs are enzymatically regulated.13 These protein modifications occur faster than the synthesis of new proteins, which allows cells or organisms to respond rapidly to changes in the surrounding environment,14 making the PTM process essential for signal transduction and life processes.15 PTMs can occur at various stages of a protein's “life cycle.” New proteins can be modified immediately after synthesis to mediate their folding into the correct structures,16 while stable proteins are modified in response to stimuli to trigger or block downstream signaling pathways.17

Details are in the caption following the image
Common types of protein posttranslational modifications. Small molecules, lipids, carbohydrates, and polypeptides can be added to amino acid side chains to form modifications. In addition, changes in the chemical properties of amino acids are also common modifications, such as citrullination.

By changing protein conformation, activity, charges and stability and interactions with DNA, RNA, and other proteins within and between cells, PTMs ultimately alter the phenotypes and biological functions of cells18 and participate in the regulation of numerous cellular processes and pathways, such as cell cycle,19 cell differentiation,20 transcriptional regulation,21 cell metabolism,17 immunity,22 signal transduction,23 and autophagy.24 For example, phosphorylation is involved in cell signal transduction and the cell cycle25; acetylation and methylation are associated with transcriptional regulation and cell metabolism26, 27; glycosylation plays an important role in protein folding and cell adhesion28; and ubiquitination regulates protein degradation and localization.29

Abnormal PTMs may cause changes in protein properties and loss of protein biological functions, directly participating in the occurrence and development of diseases.30 For example, Tau hyperphosphorylation usually leads to neurodegenerative diseases such as Alzheimer's disease (AD).31 Low palmitoylation of the mutant huntingtin (HTT) protein in the nervous system results in increased neurotoxicity and greater susceptibility to aggregate formation, which may induce Huntington's disease (HD).32 Protein acetylation is a critical regulator of insulin sensitivity and metabolism, global SIRT1 overexpression can improve insulin sensitivity, glucose tolerance, and hepatic steatosis.33 The disorder of glucose and lipid metabolism in type 2 diabetes mellitus (T2DM) may be related to the malfunction of key enzymes caused by malonylation.34 In cancers, many signaling pathways are in a state of continuous activation and are mainly conducted through a cascade of reversible phosphorylation of different proteins, such as the MAPK, JAK/STAT, and PI3K/AKT signaling pathways.35 Moreover, the continuous ubiquitination of tumor suppressors causes protein degradation and functional loss, also contributing to the development of various tumors.36 In addition to nonhistone modifications, the roles of histone modifications in health and diseases are also very important. Ataxin-3 protein causes spinocerebellar ataxia by altering histone acetylation profiles and inducing transcriptional defects.37 Loss of H4K16ac and H4K20me3 are key hallmark of human cancer.38 In hematological malignancies, hypermethylation frequently occurs at H3K7939 and H3K4.40 Thus, deciphering PTMs is of great significance for the prevention, diagnosis, and treatment of diseases.19

In this review, we systematically examine the various PTMs, including phosphorylation, acetylation, acylation with short- or long-chain fatty acids, methylation, ubiquitination, SUMOylation, glycosylation, citrullination, carbamylation, cysteine oxidation, and other modifications. We discuss their characteristics, regulatory mechanisms, and functions in both health and diseases, including development and aging, immune diseases, metabolic disorders, cancers, neurodegenerative diseases, and cardiovascular diseases (CVDs). Moreover, the therapeutic prospects in various diseases by targeting PTMs and associated regulatory enzymes are also summarized.

2 PHOSPHORYLATION

Protein phosphorylation, formed by adding a phosphate group from ATP to the side chains of amino acids by kinases, usually turns hydrophobic nonpolar proteins into hydrophilic polar proteins. Phosphorylation is a reversible PTM, and the reverse process of phosphorylation is called dephosphorylation catalyzed by phosphatases (Figure 2A).41 Phosphorylation modifications occur most commonly on serine, followed by threonine and tyrosine residues, accounting for 86.4, 11.8, and 1.8%, respectively.42 However, it is important to note that kinases can also act on the side chains of other amino acids, such as cysteine, lysine, histidine, arginine, aspartic acid, and glutamic acid, although with reduced frequency.43 Histidine and aspartate phosphorylation are much less stable than other modifications.42 The phosphosites can be recognized and bound by specific phosphorylation-binding proteins.44 Therefore, the protein phosphorylation system consists of kinases, phosphatases, phosphorylation substrates, and phosphorylation-binding proteins.45

Details are in the caption following the image
The phosphorylation process and the potential functions of phosphorylation. (A) The phosphorylation and dephosphorylation process. Phosphorylation is catalyzed by kinases, and dephosphorylation is mediated by phosphatases. Most phosphorylation events occur on serine, threonine, and tyrosine residues. (B) Representative functions of protein phosphorylation are shown. Protein phosphorylation extensively affects cellular signal transduction, protein stability, activity, localization, conformation, protein–protein interactions, gene transcription, and so on. Representative phosphosites with related functions are shown.

Protein kinases are widely distributed in cells throughout the nucleus, cytosol, mitochondria, and microsomes. To date, 518 protein kinases have been identified and verified.42 The 518 protein kinases are mainly divided into the following three categories according to the type of amino acids on which protein phosphorylation occurs, including serine/threonine protein kinases (STKs),46 protein tyrosine kinases (PTKs),47 dual-specificity kinases (DSKs),42 and histidine protein kinases (HPKs) (Figure 3).48 The STKs are enzymes that phosphorylate serine or threonine and are activated by different events such as DNA damage and chemical signals. STKs include protein kinase A (PKA), protein kinase C (PKC), PKG, calcium/calmodulin-regulated kinase (CaMK), CMGC, CK1, and so on.49 According to whether PTK is a cell membrane receptor, PTKs can be divided into nonreceptor type and membrane receptor type.50 Receptor-type tyrosine kinases include EGFR, VEGFR, and FGFR. Abnormal activation of these kinases is related to angiogenesis, tumor invasion, and metastasis.51, 52 Nonreceptor tyrosine protein kinases mainly contain BTK, JAK, and FAK, which are related to cell proliferation and migration.53 DSKs can phosphorylate STKs and PTKs.42 HPKs are a large class of enzymes involved in signal transduction by auto-phosphorylating conserved histidine residues.48 In addition, based on sequence similarity in the kinase domain, protein kinases can be divided into the following categories: tyrosine kinase (TK) family, tyrosine kinase-like (TKL) family, sterile 20 serine/threonine (STE) kinase family, casein kinase 1 (CK1) family, protein A, G, and C (AGC) kinase family, CAMK family, CMGC family (including cyclin-dependent kinases (CDKs), mitogen-activated protein kinases (MAP kinases), glycogen synthase kinases (GSK) and CDK-like kinases, receptor guanylate cyclase family (RGC), and others (Figure 3).54

Details are in the caption following the image
The kinases and phosphatases can be classified into different groups based on their phosphorylation substrates or based on their sequence similarity.

In contrast, many phosphatases are thought to be passive housekeeping enzymes and seem less important than protein kinases.55 According to the pH required for their proper functions, phosphatases can be divided into alkaline phosphatases and acid phosphatases.56 Protein phosphatases can also be classified into three main families based on their substrate specificity, including the phosphoprotein phosphatase (PPP) family, metallo-dependent protein phosphatase (PPM) family and protein-tyrosine phosphatase (PTP) family57, with three additional families: HADs, LPs, and NUDT (Figure 3).58 The PPP and PPM families are serine/threonine-specific phosphatases that appear to have evolved independently of each other.41 The phosphatases PP1, PP2A and PP2B and the newly discovered subfamilies PP4, PP5, PP6 and PP7 belong to the PPP family.42, 59 PP2C belongs to the magnesium ion-dependent PPM family.60 Notably, most PTPs belong to the same class but can be assigned to different subfamilies based on their selectivity for tyrosine or tyrosine/serine/threonine phosphorylation substrates.61 The first type is the classical PTPs, which are specific to tyrosine phosphorylation. The second type is dual-specificity phosphatases, which can dephosphorylate both serine and threonine residues in addition to tyrosine residues.62 Of all the phosphatases, at least 100 belong to those that dephosphorylate tyrosine residues, such as the tyrosine-specific phosphatase subfamily, Cdc25 family, myotubularin-related phosphatase and low molecular weight tyrosine phosphatase.42 PTP can also dephosphorylate aspartate-based phosphatases such as FCP/SCP (small CTD phosphatase) and TAD (haloacid dehalogenase) family enzymes and nonprotein targets such as carbohydrates, mRNA, and phosphoinositides.42 According to the structural folds, protein phosphatases can be classified into 10 types, including CC1, CC2, CC3, PPM, PPPL, HAD, AP, HP, PHP, and RTR1 (Figure 3).63

Protein phosphorylation is one of the most abundant PTMs in humans and is involved in the regulation of numerous physiological processes, such as protein activity,64 protein stability, protein conformation, protein–protein interaction (PPI), growth signal response, cell cycle, cellular stress response, neuronal function, and immune response.65 It also plays important roles in cellular activities such as cell proliferation,66 transcriptional regulation,67 DNA repair,68, 69 subcellular localization,70 and tumor development (Figure 2B).71 Mutation and abnormal expression of kinases lead to abnormal activation or dysregulation of downstream signaling pathways72 and have been found to be the causes of many human diseases,73 such as immune diseases,74 hyperuricemia,75 neurodegenerative diseases,76 and cancers.77, 78

2.1 Phosphorylation in development

Protein phosphorylation is critical in growth and development.79 It is essential for the precise regulation of cell proliferation, cell cycle arrest, and differentiation into various cell types during embryonic development.80 During early embryonic development, the metabolism of mammalian totipotent stem cells is tightly regulated by the kinases HK and PFK1. In addition, phosphorylation regulates the process of embryonic development by mediating chromosome condensation and spindle assembly.81 EGF promotes AKT1 phosphorylation through PI3K, which further stimulates the proliferation of stem cells and precursor mesenchymal cells while blocking their differentiation.82 Phosphorylation of the RNA-binding protein MSY2 during oocyte-to-embryo transition drives maternal mRNA degradation and converts a highly differentiated oocyte to totipotent blastomeres.83 PKD stimulates the phosphorylation of MAPK for spindle organization and cofilin for actin assembly and plays an important role in meiotic maturation of porcine oocytes.84

Growth inhibitory signaling is regulated by the Raf/MEK/ERK pathway, which plays an important role in early development and neuronal differentiation.85 The persistent activation of ERK1/2 is a common feature of growth inhibitory signaling in the Raf/MEK/ERK pathway.86 The target of rapamycin (TOR) is a kinase that regulates cell growth and metabolism by stimulating cell growth through anabolism and inhibiting catabolism.87 In mammalian cells, cyclin E plays a role in the G1 and S phases of cell cycle. Cyclin E1 and cyclin E2 affect cell growth and development by activating the cyclin-dependent kinase CDK2 and then phosphorylating a series of proteins involved in cell cycle progression, male meiosis, and stem cell maintenance.88 Deficiency of both cyclin E1 and cyclin E2 in mice is embryonic lethal.89

2.2 Phosphorylation in aging

Aging is a process characterized by declines in both organism and organ functions, which can result in various diseases.90 This process is characterized by cell cycle arrest,91 abnormal accumulation of senescent cells in tissues,92 and altered neurotransmission and response ability to external stimuli.93 In quiescent cells, most protein phosphorylation does not change significantly with age.94 However, some protein phosphorylation significantly changes with age and has crucial physiological functions. For example, αB-crystallin is a lenticular protein, and its phosphorylation can be boosted by aging, stress, and diseases.95 αB-crystallin phosphorylation is also increased in aged muscle tissues and eye lenses.96, 97 Modulation of αB-crystallin phosphorylation is a potential strategy to address aging-related complications.98

p53 is an important tumor suppressor,99 and its ability to suppress tumors is related to the function of p53 in regulating the transcription of genes associated with cell cycle arrest and senescence.100 Phosphorylation of the p53 DNA-binding domain can reduce its activity and prevent senescence.98 p53-triggered senescence is also mediated by phosphorylation of other proteins, such as MDM2 at Ser183, which can activate p53-mediated senescence and delay tumor progression.100

The brain is one of the most functionally affected organs during aging, and dysregulation of protein phosphorylation is common during brain aging.101 Protein phosphorylation signals in the brain are rich and diverse and mediated by kinases such as PKA, PKC, and CAM during aging.102, 103 The phosphorylation levels of B50/GAP-43 protein, which plays a role in long-term memory, are significantly reduced in the hippocampus of aging rats. An imbalance in protein phosphorylation, including Tau phosphorylation, acts as a key factor causing brain aging.101 Specifically, the accumulation of Tau phosphorylation at Ser396/404 in mitochondria is associated with cognitive dysfunction.104

Sarcopenia is characterized by the loss of skeletal muscle mass and strength with age.105, 106 The elderly population may experience basal hyperphosphorylation of mTORC1, which could potentially contribute to insulin resistance and the age-related anabolic resistance of skeletal muscle protein metabolism in response to nutrition and exercise.107 The decreased phosphorylation of myosin regulatory light chain (RLC), a critical protein involved in the modulation of muscle contractility, at Ser14/15 with age is the cause of sarcopenia-associated muscle dysfunction (Table 1).108

TABLE 1. Representative phosphorylation events in health and diseases.
Diseases and biological processes Protein substrates Effects
Aging αB-crystallin αB-crystallin phosphorylation increases in muscle tissues and eye lens with age.96, 97
p53 Phosphorylation of p53 DNA-binding domain reduces p53 activity and prevents senescence.98
MDM2 phosphorylation at Ser183 activates p53-mediated senescence and delays tumor progression.100
B50/GAP-43 protein Phosphorylation B50/GAP-43 is critical for long-term memory and reduced in the hippocampus of aging rats.193
Tau Accumulation of Tau phosphorylation at Ser396/404 in mitochondria contributes to cognitive dysfunction during aging.104
mTORC1 Basal mTORC1 hyperphosphorylation in the elderly may contribute to insulin resistance and the age-related anabolic resistance of skeletal muscle protein metabolism to nutrition and exercise.107
RLC Decreased phosphorylation of RLC at Ser14/15 with age causes sarcopenia-associated muscle dysfunction.108
Development MeCP2 S421 phosphorylation controls the ability of MeCP2 to regulate dendritic patterning, spine morphogenesis, and the activity-dependent induction.194
AKT1 EGF promotes AKT1 phosphorylation, which further stimulates the proliferation of stem cells and precursor mesenchymal cells while blocks their differentiation.195
MSY2 Phosphorylation of the MSY2 drives maternal mRNA degradation and converts a highly differentiated oocyte to totipotent blastomeres.196
Immune regulation Infection STAT1 Serine phosphorylation of STAT1 is required for the body's resistance to viral infection.115
STAT6 STAT6 regulates the innate immunity by transducing signals from extracellular cytokines through phosphorylation by TBK1.119
STAT2 STAT2 phosphorylation at S734 inhibits IFN-α-induced antiviral responses.120
TRAF4 TRAF4 phosphorylation downregulates innate immune signaling.197
STING Cyclic dinucleotides trigger ULK1 (ATG1) phosphorylation of STING to prevent sustained innate immune signaling.198
NF-κB NF-κB activation regulated by phosphorylation controls the expression of a series of inflammatory cytokine genes and triggers the antiviral innate immune response.199
MAVS Phosphorylation of MAVS and STING by IKK/TBK1 induces type-I IFNs and other antiviral molecules.117
MITA MITA phosphorylation by TBK1 during antiviral immunity activates IRF3.200
YAP Viruses activate the kinase IKKε to further phosphorylates YAP at Ser403 and trigger YAP degradation to antagonize innate antiviral immunity.121
Tumor immunology RAB7 RAB7 phosphorylation by TBK1/IKKε regulates innate immune signaling in triple-negative breast cancer.201
PDHE1α Phosphorylated PDHE1α at S327 by ERK2 in cytoplasm can induce its transfer to mitochondria and improve NF-κB signal in the cytoplasm, which increases resistance to cytotoxic lymphocytes and promotes tumor immune escape.125
IκBα In GBM, HK2 binds to IκBα and phosphorylates it at Thr291, which increases PD-L1 expression and promotes tumor immune escape.126
p73 CDK4/6 of tumor cells phosphorylate p53 family member p73 to prevent DR5 activation and promote antitumor immunity.202
NEDD4 In urothelial carcinoma, activated FGFR3 phosphorylates NEDD4 and further regulates Lys48-linked ubiquitination of PD-L1 to activate CD8+ T cell infiltration and antitumor activity.203
METTL3 TBK1 phosphorylates m6A methyltransferase METTL3 to enhance the interaction between METTL3 and the translation complex, which promotes antitumor immune response.204
Metabolic disorders DM GLP1 Phosphorylation at Arg91 may inhibit processing of glucagon precursor to GLP1 to affect the blood glucose levels.132
PPARγ Phosphorylation of PPARγ at S273 induces insulin resistance by upregulating Gdf3 expression and inhibiting BMP signaling pathway.133
Afadin Phosphorylation of Afadin at S1795 promotes insulin resistance in the early stages of diet-induced obesity.134
Obesity PPARγ Blocking PPARγ phosphorylation at Thr166 prevents obesity-related metabolic dysfunction.138
Cancers Multiple cancers HK1 c-Src phosphorylates HK1 at Tyr732 to promote the glycolysis rate of tumor cells and their proliferation, invasion, and metastasis abilities.71
Multiple cancers IκBα Aerobic glycolysis promotes tumor immune escape through phosphorylation of IκBα at T291 mediated by HK2.126
Breast cancer HK2 Phosphorylation of HK2 at Thr473 by PIM2 enhances HK2 stability and activity and promotes glycolysis, tumor growth, and drug resistance to paclitaxel.165
Glioma PFKP Phosphorylation of PFKP by AKT at Ser386 inhibits PFKP degradation and promotes aerobic glycolysis of glioma cells and tumor growth.167
Melanoma PFKFB2 RSK phosphorylates PFKFB2 to increase PFKFB2 activity and the glycolysis pathway, which accelerates the growth of BRAF-mutated melanoma.168
Multiple cancers PKM2 Phosphorylation of PKM2 at Tyr105 mediates the transformation of tumor cell metabolic mode to aerobic glycolysis.170
Multiple cancers PDHA Hyperphosphorylation of PDHA at Ser295 and Ser314 redirects tumor metabolism to TCA cycle. This protects spread cancer cells from metabolic and oxidative stress-induced cell death and promotes tumor metastasis.171
Gastric cancer ULK1 DAPK3 directly phosphorylates Ser556 of ULK1 to increase ULK1 activity and promote the formation of ULK1 complex, leading to inhibition of the proliferation of gastric cancer cells.173
GBM ACSS2 Phosphorylation at Ser267 of ACSS2 by CDK5 inhibits the degradation of ACSS2 and promote the growth of GBM tumor cells.174
Colon cancer Drp ERK phosphorylates Drp1 at Ser616 to activate it. Activated Drp1 facilitates the oxidation of fatty acids to promote the proliferation of colon cancer cells.176
Breast cancer RNF12 AKT promotes TGF-β-driven breast cancer metastasis by mediating RNF12 phosphorylation and enhancing RNF12 stability.177
Bladder cancer AKT KNSTRN phosphorylates AKT at Thr308 and Ser473 to activate AKT and promotes bladder cancer metastasis.178
Breast cancer PKM2 Phosphorylation of PKM2 at Ser37 is a prominent feature of invasive breast cancer.180
PC PD-L1 NEK2 phosphorylates PD-L1 at Thr194/Thr210 to maintain its stability, leading to less effectiveness of PD-L1-targeted therapy in PC.205
Neurodegenerative diseases PD Parkin Dyrk1A phosphorylates Parkin at Ser131 to inhibit its E3 Ub ligase activity, which may be involved in the pathogenesis of PD.206
XBP1s PINK1 can control XBP1s transcriptional activity by phosphorylating XBP1s at Ser61 and The48, which consequently enhances PINK1's own transcription.186
AD Tau Transient Tau hyperphosphorylation has a protective effect on neurons. While persistent accumulation of phosphorylated Tau causes neurodegeneration.189 Hyperphosphorylated Tau depolymerizes normal microtubule-associated proteins after forming neuronfibrillary tangle, disrupts cellular dynamic structures, blocks intracellular material exchange and cell signaling, inhibits Ub–proteasome activity.190, 191
  • Abbreviations: AD, Alzheimer's disease; DM, diabetes mellitus; GBM, glioblastoma; PC, pancreatic cancer; PD, Parkinson's disease; RLC, regulatory light chain.

2.3 Phosphorylation in immune regulation

Phosphorylation, a common PTM, plays a crucial role in regulating innate and acquired immunity, a process that is coregulated by kinases and phosphatases.109, 110 Phosphorylation and other PTMs work together to regulate the signaling networks of the immune system. For example, the MAP4K family of kinases play an important role in immune cell signaling, immune response, and inflammation; PKC is involved in regulating the important signaling pathways of innate and adaptive immunity, and plays an intermediary role in the signaling process of immune cells through immune synapses; PKA is involved in multiple processes that regulate immune activation and immune control, not only regulating lymphocyte activation, but also modulating antigen receptor-induced signaling by altering protein interactions and altering enzyme activity of substrate proteins.111-113

Normally, phosphorylation and dephosphorylation maintain a dynamic balance in maintaining the immune homeostasis of organisms. On the one hand, protein phosphorylation is widely involved in immune regulation, for example, the receptors on immune cells trigger phosphorylation signals through the recruitment of TKs, resulting in the activation of immune cells.114 Shuai et al.115 showed that serine phosphorylation of STAT1, an important signal converter in IFN signaling, is required for the body's resistance to viral infection. On the other hand, dephosphorylation of proteins is also widely involved in immune responses and this process is mediated by phosphatases.110

Innate immunity is the first line of defense against pathogen invasion. Phosphorylation plays an important role in innate immunity. It has been shown that the transcription factor (TF) interferon regulatory factor 3 (IRF-3) regulates gene expression in innate immune responses, and IRF-3 activation is mediated by phosphorylation of kinase IKK/TBK1.116 The toll-like receptors (TLRs) are principal sensors capable of sensing multiple microbial stimuli and inducing innate immune responses through a cascade of phosphorylation signals. TLR signaling reaches its peak during the activation of nuclear factor-kappaB (NF-κB), which is mediated by phosphorylation and controls the expression of a series of inflammatory cytokine genes and further triggers the innate immune response against viruses. During viral infection, the adaptor proteins MAVS and STING are phosphorylated by the kinase IKK/TBK1 in response to stimulation, inducing type I interferons (IFNs) and other antiviral molecules.117 Chen et al. found that ionizing radiation leads to phosphorylation of phosphoribosyl pyrophosphate synthetase 1/2 at T228, triggering innate immune response in the body.118

The antiviral immune response also requires phosphorylation to mediate. STAT6 is essential for antiviral innate immunity. After viral infection, STAT6 is aggregated in the endoplasmic reticulum and phosphorylated by TBK1, which then dimerizes into the nucleus and regulates the expression of antiviral immunity genes.119 And phosphorylation of STAT2 at S-734 inhibits IFN-α-induced antiviral response.120 The virus also activates the kinase IKKε, which phosphorylates YAP at Ser403, triggering degradation of YAP in lysosomes and antagonizing innate antiviral immunity.121 In addition, kinase complex mTORC2, which is involved in phosphorylation of AKT and GSK3β kinase, can maintain reactive oxygen species balance in mitochondria and maintain the lifespan of virus-specific memory CD4+ T cells in vivo, playing an important role in antiviral immunity.122

Phosphorylation plays a key role in signal transduction during tumor immunity, mediating immune escape in a variety of tumors. PD-1 is crucial for inhibiting the activation of T cells in vitro and in vivo, and its immunosuppression process also requires phosphorylation mediated by the specific mechanism as follows: PD-1 binds to its ligand PD-L1, then aggregates with T cell receptors (TCRs) and binds briefly to phosphatase SHP2 to initiate dephosphorylation of TCR, resulting in inhibition of T cell activation.123 Inhibition of CDK4/6 in vivo has been shown to inhibit cyclin D-CdK4-mediated Spoz protein phosphorylation, thereby increasing PD-L1 protein levels, and this can increase the number of tumor infiltrating lymphocytes and enhance tumor immunity.124 Yang et al. found that phosphorylated PDHE1α (pyruvate dehydrogenase complex E1 subunit α) at S327 by ERK2 in cytoplasm can induce its transfer to mitochondria and improve NF-κB signal in the cytoplasm, which increases resistance to cytotoxic lymphocytes and promotes tumor immune escape.125 In human glioblastoma cells, a high glucose environment promotes mitochondrial separation of hexokinase 2 (HK2), which binds to the T291 site of IκBα and phosphorylates it, subsequently mediating upregulation of PD-L1 and promoting tumor immune escape.126

2.4 Phosphorylation in metabolic disorders

Abnormal phosphorylation may lead to the blockage of cell signaling and in turn result in metabolic disorders in the human body.42 Diabetes mellitus is a metabolic syndrome characterized by long-term hyperglycemia, 90% of which is T2DM. Insulin resistance is a fundamental mechanism leading to T2DM.127 Glucose homeostasis is maintained by insulin in insulin-responsive tissues, while phosphorylation is a critical mechanism for regulating insulin secretion and insulin signaling processes.33, 128, 129 Over 1000 phosphorylation events are dysregulated in T2DM.130 The effect of phosphorylation on diabetes occurs mainly through the cascade of kinases and phosphatases that regulate insulin signaling.129 Several key molecules in the insulin pathway, such as IR, IRS1, IRS2, PDK, and mTORC1, are phosphorylated upon insulin stimulation.131 In mammals, GLP1 acts as an incretin to promote the release of insulin from pancreatic B cells. It is speculated that phosphorylation at Arg91 may inhibit processing of the glucagon precursor to GLP1 to affect blood glucose levels.132 In addition, phosphorylation of obesity-associated PPARγ at S273 induces insulin resistance by upregulating Gdf3 expression and inhibiting the BMP signaling pathway1.133 Phosphorylation of Afadin at S1795 also promotes insulin resistance in the early stages of diet-induced obesity.134

Obesity, a common metabolic disorder, results from the accumulation of adipose tissue caused by energy imbalances.135 Phosphorylation plays a role in the pathogenesis of obesity by regulating adipogenesis and metabolism. For example, S6K1 participates in many key metabolic pathways, including lipid synthesis in the body, by mediating the phosphorylation of H2BS36 in obese patients. S6K1 is a potential therapeutic target for obesity.136 Mammalian white adipose tissue (WAT) is critical for whole-body homeostasis. Smyd2 is abundant in WAT and regulates STAT2 phosphorylation to regulate adipocyte differentiation.137 PPARγ is indispensable in the process of adipocyte differentiation, and the phosphorylation level of PPARγ at Thr166 is positively correlated with obesity status. Specifically, blocking PPARγ phosphorylation at Thr166 prevents obesity-related metabolic dysfunction (Table 1).138

2.5 Phosphorylation in cancers

Abnormal kinase activity and expression are implicated in various types of cancers. In recent years, with the increasing development of mass spectrometry (MS) technology, the Clinical Proteomic Tumor Analysis Consortium and many other teams have conducted phosphoproteomics investigations in various cancers, such as lung cancer,139-142 colorectal cancer (CRC),143-146 liver cancer,147 breast cancer,148 prostate cancer,149 gastric cancer,150, 151 head and neck cancer,152 esophageal cancer,153 pancreatic cancer (PC),154, 155 kidney cancer,156, 157 melanoma,158 skin cancer,159 leukemia,160 pancreatic ductal adenocarcinoma (PDAC),161 pituitary neuroendocrine tumors,162 cholangiocarcinoma,163 and urothelial carcinoma of the bladder.164

Protein phosphorylation mediates metabolic reprogramming of tumors. Studies have found that c-Src phosphorylates HK1 at Tyr732, which promotes the glycolysis rate of tumor cells and their proliferation, invasion, and metastasis abilities.71 Aerobic glycolysis can promote tumor immune escape through IκBαT291 phosphorylation mediated by HK2.126 Moreover, HK2 can be phosphorylated at Thr473 by the kinase PIM2, which increases its stability and enzymatic activity and promotes glycolysis and breast tumor growth, enhancing its drug resistance to paclitaxel.165 AKT2 may also be an upstream kinase leading to HK2 Thr473 phosphorylation in CRC.166 PFK also plays an important role in the regulation of tumor metabolism. The homologous isoform PFKP of PFK1 can be phosphorylated by AKT at Ser386, which inhibits the degradation of PFKP and promotes aerobic glycolysis in glioma cells and tumor growth.167 RSK directly phosphorylates PFKFB2 to increase PFKFB2 activity and glycolysis, which accelerates the growth of BRAF-mutated melanoma.168 PFKFB3 in the cytoplasm is phosphorylated and activated by AMPK. Targeted inhibition of PFKFB3 improves the sensitivity of chemotherapy drugs such as cisplatin.169 Phosphorylation at Tyr105 of PKM2 is significantly increased in various tumors to mediate the transformation of the tumor cell metabolic mode to aerobic glycolysis.170 Hyperphosphorylation of Ser295 and Ser314 of PDHA redirects tumor metabolism to the tricarboxylic acid (TCA) cycle by increasing PDH activity. This protects cancer cells from metabolic and oxidative stress-induced cell death and promotes tumor metastasis.171

Protein phosphorylation extensively regulates cancer cell proliferation, metastasis, and invasion.172 For example, DAPK3 directly phosphorylates Ser556 of ULK1, which increases the activity of ULK1 and promotes the formation of the ULK1 complex, leading to inhibition of the proliferation of gastric cancer cells. The downregulation of DAPK3 in gastric cancer patients is related to poor prognosis.173 Phosphorylation at Ser267 of ACSS2 by CDK5 kinase inhibits the degradation of ACSS2 and promotes the growth of GBM tumor cells.174 BZW1 enhances the phosphorylation of eIF2α to promote tumor progression. This process can be prevented by the PERK/eIF2α phosphorylation inhibitors GSK2606414 and ISRIB.175 The kinase ERK catalyzes Drp1 phosphorylation at Ser616 to activate Drp1. Activated Drp1 changes the metabolic pathway, facilitates the oxidation of fatty acids, and promotes the proliferation of colon cancer cells.176

Dysregulated phosphorylation can promote tumor metastasis. For example, AKT promotes TGF-β-driven breast cancer metastasis by mediating RNF12 phosphorylation and enhancing RNF12 stability.177 KNSTRN, a component of the mitotic spindle, phosphorylates AKT at Thr308 and Ser473 to activate AKT and promote bladder cancer metastasis.178 TKT, a key metabolic enzyme in the pentose phosphate pathway (PPP), interacts with GRP78 to promote glycolysis by increasing AKT phosphorylation, which promotes CRC metastasis.179 PKM2 phosphorylation at Ser37 is a prominent feature of invasive breast cancer. The use of the pyruvate kinase activator TEPP-46 or the potent CDK inhibitor dinaciclib to bind to phosphorylation sites can reduce its nuclear localization and inhibit cancer cell migration and invasion (Table 1).180

2.6 Phosphorylation in neurodegenerative diseases

Parkin is a tumor suppressor gene, and its overexpression can inhibit the growth of tumor cells. Parkin mutations exist in a variety of malignant tumors, such as colon cancer,181 PC,182 and cervical cancer.183 However, Parkin is also a causative gene related to Parkinson's disease (PD). It has a neuroprotective effect, and mutations in Parkin lead to the loss of dopaminergic neurons in the substantia nigra.184 Parkin is almost inactive in vitro, and its activation is regulated by PINK1-mediated phosphorylation.185 After phosphorylation, the protein conformation, solubility, and affinity with the substrate of Parkin are changed. Parkin amplifies the PINK1-induced signaling pathway through positive feedback, which enhances mitophagy and selectively degrades defective mitochondria to maintain the stability of the intracellular environment. Abnormalities in this pathway may cause PD.184 In addition, PINK1 controls XBP1s transcriptional activity by phosphorylating XBP1s at Ser61 and Thr48, which consequently enhances PINK1 transcription, and triggers a promitophagic phenotype.186 Notably, functional deficiency of Parkin leads to ineffective ubiquitination and a large accumulation of cyclins. These cyclins are responsible for initiating the cell cycle in both neurons and mitotically active cells. However, due to the lack of mitogenic capacity in neurons, their inability to undergo cell division ultimately leads to apoptosis.187

Tau hyperphosphorylation has an intrinsic link with neurodevelopment and degeneration, and the phosphorylation level of Tau in the AD brain is three to four times higher than that of normal peers.188 Transient Tau hyperphosphorylation is protective on neurons. However, persistent accumulation of hyperphosphorylated Tau may cause neurodegeneration.189 Hyperphosphorylated Tau depolymerizes normal microtubule-associated proteins after forming neuronfibrillary tangles, disrupts cellular dynamic structures, blocks intracellular material exchange and cell signaling, inhibits ubiquitin (Ub)–proteasome activity, and finally leads to neurodegenerative diseases.190, 191 In a cohort study of 593 elderly people with an average age of 64 years, it was found that compared with cognitively normal controls, the plasma concentrations of P-tau217 and P-tau181 are increased in clinical AD patients, suggesting that P-tau217 and P-tau181 may be useful biomarkers for AD diagnosis (Table 1).192

2.7 Phosphorylation-associated targeted therapies

Compared with traditional cytotoxic anticancer drugs, targeted anticancer drugs have the advantages of high efficiency, low toxicity, and strong specificity.207 Given the important roles of protein kinases in tumor growth and metastasis, if the kinase signaling pathway is effectively blocked, the malignant progression of tumors may be prevented.208 To date, the United States Food and Drug Administration (US FDA) has approved 68 small molecule kinase inhibitors.209 These kinase inhibitors can be roughly divided into four classes according to the way they bind to protein kinases. Type I kinase inhibitors are by far the most US FDA-approved drugs, such as bosutinib, dasatinib, and crizotinib.210 Dasatinib acts on multiple targets, such as BCR-Abl and the SRC kinase family, and is mainly used for the treatment of leukemia.211 Crizotinib has been confirmed in tumor patients with abnormal ALK, ROS kinase, and HGFR/c-MET activities.212 Type II kinase inhibitors, including the BCR-Abl inhibitors imatinib and nilotinib, are mainly used for the treatment of chronic myeloid leukemia (CML).213, 214 Another representative drug, sorafenib,215 is a typical multitarget drug targeting TKs such as VEGFR2 and PDGFR-β, as well as the serine/threonine kinase Raf-1,216 and can be used for the treatment of hepatocellular carcinoma (HCC) and renal cell carcinoma (RCC).217, 218 Allosteric kinase inhibitors are another type of kinase inhibitor.215 Trametinib and cobimetinib are allosteric kinase inhibitors targeting MEK1/2, both of which can be used for the treatment of non-small cell lung cancer (NSCLC).219 Allosteric kinase inhibitors do not bind to the ATP binding site, so they act together with ATP-competitive inhibitors, which makes allosteric inhibitors useful for overcoming the low selectivity, off-target effects and resistance of small molecule inhibitors.220 The fourth type of kinase inhibitors are covalent inhibitors, such as afatinib, neratinib, ibrutinib and acalabrutinib.215, 221 Afatinib acts on EGFR and is mainly used to NSCLC.222 Neratinib inhibits HER2 and is used for the treatment of HER2-positive breast cancer.223 The BTK inhibitors ibrutinib and acalabrutinib are mainly used for the treatment of chronic lymphocytic leukemia (CLL) and mantle cell lymphoma (MCL).224, 225 Notably, acalabrutinib significantly prolonged the progression-free survival of patients with CLL (Table 2).221

TABLE 2. Representative approved kinase inhibitors and their clinical uses.
Classifications Targets Drugs Clinical uses
TKIs ALK Alectinib, brigatinib, ceritinib, crizotinib, lorlatinib ALK-positive NSCLC238
BCR-Abl Bosutinib, dasatinib, nilotinib, ponatinib CML239
Imatinib Ph+ CML/ALL, GIST, aggressive systemic mastocytosis, chronic eosinophilic leukemias, dermatofibrosarcoma protuberans, hypereosinophilic syndrome, myelodysplastic, and myeloproliferative disease207, 209, 210
BTK Acalabrutinib MCL, CLL, SLL240, 241
Ibrutinib CLL, MCL, marginal zone lymphomas, graft-versus-host disease209
c-MET Cabozantinib Metastatic medullary thyroid cancer242
Crizotinib Metastatic ALK-, c-MET-, or ROS-1-positive NSCLC243
c-KIT Axitinib RCC244
Cabozantinib Metastatic medullary thyroid cancer245
Erlotinib NSCLC246
Nilotinib Ph+ CML247
Pazopanib Advanced RCC, advanced soft tissue sarcoma248, 249
CSF1R Pexidartinib Tenosynovial giant cell tumors250
EGFR Erlotinib NSCLC, PC251, 252
Afatinib, dacomitinib, gefitinib, osimertinib NSCLC253
Lapatinib, neratinib HER2-positive breast cancer254
FGFR Erdafitinib Urothelial bladder cancers255
Nintedanib IPF256
FLT3 Gelteritinib AML257
Midostaurin AML, mastocytosis, mast cell leukemias258, 259
JAKs Fedratinib, ruxolitinib Myelofibrosis260
Baricitinib, tofacitinib RA261
PDGFR Axitinib RCC244
Erlotinib NSCLC207
Nilotinib Ph+ CML210
Pazopanib Advanced RCC, advanced soft tissue sarcoma210, 262
Sorafenib RCC, HCC210, 263, 264
Sunitinib RCC, GIST265, 266
RET Alectinib NSCLC267
Cabozantinib Medullary thyroid cancers, RCC, HCC268
Lenvatinib Differentiated thyroid cancers269
ROS1 Crizotinib, entrectinib ROS1-postive NSCLC270
SRC Bosutinib Ph+ CML271
Dasatinib Ph+ CML/ALL210
Syk Fostamatinib, R406 Chronic immune thrombocytopenia272
TRKA/B/C Entrectinib, larotrectinib Solid tumors with NTRK fusion proteins273, 274
Tyk Ruxolitinib Myelofibrosis, polycythemia vera209
VEGFR Axitinib RCC275
Cabozantinib Medullary thyroid cancers, RCC, HCC276-278
Lenvatinib Differentiated thyroid cancer279
Pazopanib RCC, soft tissue sarcomas248, 280
Regorafenib CRC281
Sorafenib RCC, HCC, differentiated thyroid cancer282-284
Sunitinib GIST, RCC, pancreatic neuroendocrine tumors285-287
Vandetanib Medullary thyroid cancers288
STK inhibitors BRAF Dabrafenib BRAFV600E/K melanomas, BRAFV600E NSCLC, BRAFV600E anaplastic thyroid cancers289-291
Encorafenib, vemurafenib BRAFV600E/K melanoma292
CDKs Abemaciclib, palbociclib, ribociclib Breast cancer293
FKBP12/mTOR Everolimus HER2-negative breast cancers, pancreatic neuroendocrine tumors, RCC, angiomyolipomas, subependymal giant cell astrocytomas294-298
Sirolimus Kidney transplants, lymphangioleiomyomatosis (LAM)299, 300
Temsirolimus RCC301
ROCK1/2 Netarsudil Glaucoma302
Double specific protein kinase inhibitors MEK1/2 Binimetinib, cobimetinib BRAFV600E/K melanoma303
Trametinib BRAFV600E/K melanomas/BRAFV600E NSCLC304
  • Abbreviations: ALL, acute lymphocytic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CML, chronic myeloid leukemia; CRC, colorectal cancer; GIST, gastrointestinal stromal tumors; HCC, hepatocellular carcinoma; IPF, idiopathic pulmonary fibrosis; MCL, mantle cell lymphoma; NSCLC, non-small cell lung cancer; PC, pancreatic cancer; RA, rheumatoid arthritis; RCC, renal cell carcinoma; SLL, small lymphocytic lymphoma.

Tyrosine kinase inhibitors (TKIs) are currently the most widely studied. The TK EGFR is mutated or overexpressed in a variety of tumors. Abnormal expression of EGFR is closely related to the occurrence of cancer. Thus, the development of drugs targeting EGFR is a research hotspot.226 The current small-molecule inhibitors designed to target EGFR have been developed into the fourth generation.227 The first three generations of inhibitors are widely used in the clinic and have gradually become the first choice for NSCLC treatment, mainly by inhibiting the phosphorylation of the intracellular TK domain.228 First-generation EGFR-TKIs, including gefitinib and erlotinib, are reversible inhibitors.229, 230 Second-generation EGFR-TKIs, including dacomitinib and afatinib, are irreversible.230 Third-generation EGFR-TKIs mainly target T790M mutant EGFR and are irreversible as well. The representative drug is osimertinib.231 Although TKIs, represented by third-generation EGFR-TKIs, have achieved remarkable success in the field of cancer treatment, clinical results show that there are still inevitable toxic side effects in the gastrointestinal tract, skin and other organs.232 In addition, TKIs have also been used to treat T1DM and T2DM.233 For example, c-Abl234 and VEGFR2235 inhibitors have been shown to enhance β cell survival and insulin secretion, while PDGFR236 and EGFR237 inhibitors have been demonstrated to improve insulin sensitivity.

3 ACETYLATION

Acetylation is a process in which acetyl group donors, such as acetyl-CoA and acetyl phosphate, covalently bind to the protein N-terminus and lysine side chains in an enzymatic or nonenzymatic manner, forming N-terminal (Nα) and internal (Nε) acetylation (Figure 4).305, 306 The regulation of Nα-acetylation remains unclear,307 while the process of Nε-acetylation is dynamic and reversible.308 Nε-acetylation changes with the physiological state of cells and the external environment. It serves as a regulatory switch for protein conformation and activity changes. However, when Nε-acetylation becomes abnormal, it can lead to the development of diseases.306

Details are in the caption following the image
Representative scheme of reversible acetylation regulated by HATs and HDACs is shown. The classification of well-known HATs and HDACs are organized. KATs are classified into three major families: GCN5, p300 and MYST. The remaining KATs belong to basal TF family and NCoA family. HDACs are divided into two categories: the classical Zn2+-dependent HDACs and NAD+-dependent sirtuin deacetylases. HDACs can be further grouped into class I, class Iia, class Iib, class III, and class IV.

The homeostasis of lysine acetylation is regulated by lysine acetyltransferase (HATs/KATs) and lysine deacetylase (HDACs/KDACs).309 Acetylase KATs are mainly divided into three families (Figure 4), including the GNAT superfamily (GCN5, pCAF, HAT1, and MEC-17), the MYST family (MOZ, TIP60, MORF, HBO1, and MOF), and the CBP/p300 family. Apart from the above three main categories, there are also two other KAT families, including the basal TF family and the nuclear receptor coactivator (NCoA) family.310 The deacetylases are divided into two large families (Figure 4). The classical large family includes 11 members, HDAC1–11, which are similar in secondary structure to the yeast Hda1/Rpd3 protein, and all rely on Zn2+ to promote deacetylation. The deacetylases in the second major family are all NAD+-dependent yeast Sir2 homologous proteins, including seven members SIRT1–7.311

Acetylation is a widespread PTM involved in gene transcription, metabolism, DNA damage repair, signal transduction, PPIs, stress response, proteolysis, autophagy, and many other biological processes (Figure 5).20, 312, 313 In particular, histone acetylation is closely related to transcriptional activity, and hyperacetylated histones are specifically aggregated in active chromatin.314 Mechanically, negatively charged acetyl groups covalently added to specific lysine residues in histones can diminish the electrostatic affinity between histone proteins and DNA, thus disrupting the interaction of these histones with DNA and leading to chromatin relaxation that enables the activation of gene transcription.314 For example, SIRT2 catalyzes H4K16 deacetylation to maintain a condensed heterochromatin state and shut down gene transcription, whereas the histone acetyltransferase Sas2 counteracts this effect.315 In contrast, acetylation of H4K16 (rather than H4K5, H4K8, and H4K12) contributes to the folding of nucleosome arrays, which is essential for transcriptional regulation in vivo,316 suggesting that the position of acetylation in histone protein is much more important than the number of acetylation modifications. p300/CBP is a coactivator of various TFs, such as p53,317 HIF-1α,318 and c-Myc,319 which can remodel chromatin and transcription processes through the activity of acetyltransferase. In addition, acetylation also acts on almost all metabolic enzymes. By changing the PPI, localization, stability, and activity of metabolic enzymes, acetylation is extensively involved in metabolism regulation.320 In mouse hepatocyte mitochondria, more than 20% of mitochondrial proteins have been acetylated, including many growth factors and metabolic enzymes. In human liver tissue, 1300 lysine acetylation sites from 1047 proteins have been identified. Interestingly, almost all intermediate metabolic enzymes are acetylated.321 Acetylation is involved in cellular antioxidant processes. SOD2, IDH2, and G6PD are all regulated by acetylation. Deacetylation of SOD2322, 323 and IDH2324, 325 by SIRT3 and deacetylation of G6PD326 by SIRT2 can increase the catalytic capacity of SOD2, IDH2, and G6PD, as well as the level of NADPH, consequently reducing cellular oxidative damage.

Details are in the caption following the image
Representative functions of histone and nonhistone acetylation are shown. Protein acetylation is mainly involved in the regulation of gene transcription, metabolism, DNA damage repair, signal transduction, stress response, signal transduction, protein stability, protein activity, protein–protein interaction, and autophagy.

The regulation of protein stability by acetylation is usually achieved by competing with ubiquitination. For example, FASN is a key enzyme of nascent adipogenesis, and HDAC3 can reduce the interaction of FASN with the E3 Ub ligase Trim21 by deacetylating FASN.327 K163, K174, K180, and other lysine sites of Tau can be acetylated by p300, which inhibits normal ubiquitination-dependent protein degradation and microtubule assembly and promotes abnormal aggregation of Tau protein, leading to neurodegenerative diseases such as AD (Figure 5).328

Autophagy can be regulated by acetylation. Acetylation of histones and TFs regulates autophagy-related gene expression and their activities. The rapid and precise regulation of autophagy contributes to the maintenance of cellular homeostasis.329 p300 appears to acetylate many ATG proteins that regulate autophagy at multiple steps. p300 depletion or p300 inhibitors induce autophagy, while p300 overexpression inhibits autophagy.330 The acetylation of the mTORC1 component raptor is increased through a p300-dependent pathway, which activates mTORC1 and inhibits autophagy (Figure 5).331

Acetylation and deacetylation coordinated by acetyltransferase and deacetylase are in a dynamic balance to maintain normal physiological and biochemical processes of cells. However, once this balance is broken, it will lead to disordered regulation of gene expression and the occurrence of diseases.332 During normal aging, gene expression controlled by multiple epigenetic factors, including histone acetylation, is weakened. Interestingly, histone deacetylation can prolong lifespan by promoting autophagy and inhibiting oxidative stress and necrosis.333 However, many age-related diseases are often characterized by lower levels of histone acetylation.334

The occurrence of various diseases, such as metabolic diseases, tumors, CVDs, neurodegenerative diseases, and immune diseases, is related to the imbalance of protein acetylation and deacetylation.335, 336 Studies have identified a large number of acetylated proteins in the cytoplasm and mitochondria, most of which are related to metabolism. Lysine acetylation affects the functions of metabolic enzymes by regulating their activity and stability.321 Loss of their acetylation regulation may lead to metabolic disorders and the accumulation or insufficient synthesis of some metabolic intermediates, resulting in metabolic-related diseases.

3.1 Acetylation in development

Previous studies have shown that HDACs regulate histone acetylation to affect the proliferation, differentiation, apoptosis, migration, and synapse regeneration of nerve cells.337 HDAC1 and HDAC2 are essential for cortical lamination and play a crucial role in maintaining the progenitor pool during cortical development. Deletion of both HDAC1 and HDAC2 results in a deficiency in neocortical development.338 HDACs also play roles in the development and differentiation of various immune cells.312 Inactivating HDAC3 during the double-negative stages of thymocyte development will cause significant damage at the CD8 immature single-positive (ISP) stage and the CD4/CD8 double-positive stage, resulting in the production of few mature CD4(+) or CD8(+) single-positive cells.339 Deletion of HDAC3 in early B-progenitor cells caused a defect in VDJ recombination and failure in B cell development.340 In addition, HDAC3 can also indirectly regulate the development and function of these immune cells through stromal cells or target cells interacting with immune cells. For example, HDAC3 is an important component of the Notch signaling pathway that regulates the development of medullary thymic epithelial cells (mTECs).341 Loss of HDAC3 expression can lead to developmental arrest of mTECs with impaired T-cell negative selection. However, whether the regulation of HDAC3 on the development of neurons and immune cells depends on its deacetylation function remains to be further studied.

3.2 Acetylation in aging

As a key metabolite, acetyl-CoA is an important donor of acetylation modifications.342, 343 Previous studies have shown that fasting and caloric restriction (CR) reduce glucose-derived metabolic flux and cytoplasmic acetyl-CoA levels through ATP-citrate lyase,344 which decreases p300 activity to stimulate long-lived autophagy. However, increased nuclear acetyl-CoA can promote lifespan by increasing the levels of histone acetylation.343, 345, 346 Sirtuins are epigenetic enzymes that are key regulators of aging and CR.347 In yeast, CR prolongs lifespan by increasing the activity of Sir2.347, 348 In mammals, the role of SIRT1–7 in extending lifespan is also largely based on their deacetylase functions.349

Nicotinamide mononucleotide (NMN) supplementation not only inhibits the aging-associated increase in protein acetylation but also modulates fatty acid β-oxidation, TCA cycle, and valine degradation. Aged livers show increased acetylation compared with young livers, but NMN supplementation decreases acetylation. These results reveal the potential of NMN in combating aging and aging-related functional declines.350 Inflammatory aging of the brain is a hallmark of age-related neurodegenerative diseases. Integrated analysis of H3K27ac and gene expression data in human and mouse brains shows that genes upregulated and downregulated with aging are correlated with different H3K27ac modification patterns.351 By using aging mouse models under inflammatory conditions, it has been found that the pattern recognition receptor NLRP3 is acetylated in macrophages and deacetylated by NAD+-dependent sirtuins. Dysregulation of the NLRP3 inflammasome acetylation switch may be the cause of aging-associated chronic inflammation.352

During the aging process of mesenchymal stem cells, histone acetylation on the promoters and enhancers of osteogenic genes, as well as the chromatin accessibility, decreases, which leads to the downregulation of osteogenic gene expression and a decrease in osteogenesis.353 Comparing the changes in sirtuins in experimental animals of three different age groups, young, middle-aged and old, it has been found that the expression of sirtuin family proteins in skeletal muscle increases during the aging process, but acetylation is not effectively reduced, which is associated with a severe reduction in NAD+ content.354

In addition, the acetyltransferase KAT7 can promote H3K14ac-related gene expression and induce cell senescence. Inactivation of KAT7 reduces H3K14ac and represses the transcription of p15INK4b, which attenuates the senescence of human peritoneal mesothelial cells.355 DNA damage can activate ATM and inhibit LARP7-regulated SIRT1 activity, leading to increased p53 and p65 acetylation and transcriptional activity to promote cellular senescence. Activation of this pathway exacerbates aging and atherosclerosis in ApoE-knockout mice, while inactivation of this pathway can reverse these phenotypes (Table 3).356

TABLE 3. Representative acetylation substrates and their functions in health and diseases.
Diseases and the biological processes Substrates Effects
Aging NLRP3 SIRT2 and NLRP3 deacetylation prevent and can be targeted to reverse, aging-associated inflammation, and insulin resistance.352
H3K14 KAT7 promotes H3K14ac-related gene expression and induces cell senescence.355
p53, p65 DNA damage can increase p53 and p65 acetylation and transcriptional activity to promote cellular senescence.356
Metabolic disorders Diabetes IRS1 HDAC2 reduces IRS1 acetylation in hepatocytes, to reduce pancreatic insulin formation and secretion.367
Obesity H3K27 p300/CBP-mediated H3K27ac in the PPARγ complex promotes adipogenesis.385
CVDs Cardiometabolic diseases CypD Decrease of SIRT3 in failing hearts from patients with obesity and metabolic syndrome leads to CypD hyperacetylation, mitochondrial permeability transition pore opening, and cardiac dysfunction.431
p53 Activation of SIRT1 protects against advanced glycation end products (AGEs)-induced apoptosis in endothelial cells in diabetes through decreasing p53 acetylation.432
MPC2 Increased MPC2 acetylation at K19/26 impairs mitochondrial pyruvate transport activity and metabolic inflexibility in Akita diabetic hearts.433
Myocardial infarction p53 p53 acetylation at K118 increases infarct size, and its inhibition promotes NOS3-mediated cell survival and cardioprotection.434
Prdx1 Tubastatin A (TubA) selectively inhibits HDAC6 and promotes Prdx1 acetylation at K197, which offers cardioprotection against injury in rats and H/R-induced cell death in H9c2 cells.435
Cardiac hypertrophy, remodeling and heart failure NF-κB HDAC inhibitor TSA attenuates transverse aortic constriction (TAC)-induced hypertrophy by regulating histone acetylation on promoters of NF-κB target genes.436
MHC HDAC3 aggravates cardiac hypertrophy by deacetylating cardiac myosin heavy chain (MHC) isoforms.437
H3 Prenatal exposure of PM2.5 leads to lower birth weight and cardiac hypertrophy in adulthood by increasing CBP/p300 and H3K9ac.438
Hypertension H3 In spontaneously hypertensive rats, HDAC inhibition suppresses cardiac hypertrophy and fibrosis through increasing H3 acetylation on promoters of mineralocorticoid receptor (MR) target genes.439
SOD2 SIRT3 depletion causes hyperacetylation of mitochondrial SOD2 and overproduction of oxidative stress, which results in endothelial dysfunction, vascular inflammation, and hypertension in mice.440
Cardiac arrhythmias Connexin 43 Chronic tachypacing leads to abnormal ventricular activation and increases acetylation of connexin 43 in canines.441
Cancers Cancer cell proliferation H4K16 HDAC3 promotes histone H4K16ac, which acts on PI3K and enhances the interaction between LC3 and ATG4 to trigger autophagy that affects cancer cell proliferation.420
GBM PAK1 Phosphorylation of ATG5 at Thr101 in GBM is positively regulated by PAK1 acetylation, which promotes tumor growth.423
PC LDHA K5 acetylation of LDH-A is reduced in human PC, and K5 acetylation of LDH-A inhibits LDH-A activity.425
HCC H3K27 HDAC-mediated suppression of FBP1 is correlated with decreased H3K27ac in the FBP1 enhancer. Treatment of HCC cells with HDAC inhibitors restores FBP1 expression and inhibits HCC cell growth.426
Breast cancer MORC2 MORC2 acetylation is associated with elevated NAT10 expression in breast cancer. Acetylated MORC2 binds to phosphorylation at H3T11 and contributes to DNA damage-induced G2 checkpoint activation.442
Pulmonary cancer H3K27 H3K27 acetylation activated-COL6A1 promotes osteosarcoma lung metastasis by repressing STAT1 and activating pulmonary cancer-associated fibroblasts.443
Prostate cancer TPD52 Acetylation-dependent regulation of TPD52 modulates CMA oncogenic function in prostate cancer.444
PDAC BCAT2 BCAT2 is acetylated at K44. K44R mutant promotes BCAA catabolism, cell proliferation, and pancreatic tumor growth.430
Neurodegenerative diseases AD H2B The p300/CBP activator CSP-TTK21 can rescue Aβ-impaired synaptic plasticity induced by various pathways, presumably through reversing Aβ-induced dysregulation of H2B acetylation and gene expression.405
Axon dysfunction Miro1 Deacetylation of Miro1 by HDAC6 blocks mitochondrial transport and mediates axon growth inhibition.445
PD SOD2, ATP synthase β PGC-1α/ERRα-Sirt3 pathway protects against DAergic neuronal death by directly deacetylating SOD2 (K130) and ATP synthase β (K485) in PD.416
  • Abbreviations: AD, Alzheimer's disease; CVDs, cardiovascular diseases; GBM, glioblastoma; PC, pancreatic cancer; HCC, hepatocellular carcinoma; PDAC, pancreatic ductal adenocarcinoma; PD, Parkinson's disease.

3.3 Acetylation in metabolic disorders

Diabetes and obesity are related to mutations in the acetylation sites of metabolic enzymes.357, 358 Persistent hyperglycemia in diabetes can increase acetyl-CoA and protein acetylation levels, which may impair protein functions.359 In diabetic rat models, organs with high protein acetylation are susceptible to diabetic complications.360 The acetylation level of NF-κB in the hearts of diabetic rats is elevated, and the expression of Nrf2-related genes and mitochondrial activity are impaired. Consequently, this results in the persistence of inflammation, impairs the functions of the heart to resist oxidative stress, and increases the risk of cardiovascular complications in diabetes.361 HDACs and sirtuins play key roles in diabetes by affecting insulin signaling and secretion.362, 363 The GLUT4 gene promoter is composed of an MEF2-binding domain and domain I. Transcriptional activity is highest when MEF2 is bound to the MEF2-binding domain and the GLUT4 enhancer GEF is bound to domain I.364 HDACs downregulate the transcription of MEF2-related genes.365 HDAC2 reduces acetylation by binding to IRS1 in hepatocytes, thereby reducing insulin receptor-mediated IRS1 tyrosine phosphorylation366 and downregulating pancreatic insulin formation and secretion.367 Notably, SIRT1 can enhance glucose-induced pancreatic insulin secretion.368 HDACs and p300/CBP mediate STAT3 acetylation to regulate gluconeogenesis.369

The p300/CBP family and SIRT3/SIRT6 are involved in the process of obesity,370-372 and the histone acetylation level is positively correlated with adipogenic differentiation.373-375 p300/CBP in the PPARγ complex is the main enzyme that activates gene transcription, which can increase the expression of CEBPα and PPARγ and promote adipogenesis.376-378 This process has been associated with p300/CBP-mediated H3K27ac.379, 380 p300/CBP double knockout mice develop severe lipodystrophy with hepatic steatosis, hyperglycemia, and hyperlipidemia.381 In a cardiac-specific SIRT6 knockout mouse model fed a high-fat diet (HFD), loss of SIRT6 function exacerbates cardiac injury, including left ventricular hypertrophy and lipid accumulation.382 Enzymes of the HDAC family are also involved in the regulation of obesity. For example, by regulating fat metabolism, HDAC3 can promote fat absorption and diet-induced obesity.372 HDACs also inhibit adipogenesis by downregulating histone acetylation.383, 384 HDAC1 but not HDAC2 can inhibit adipogenesis by reducing CEBPα and PPARγ expression (Table 3).385

3.4 Acetylation in CVDs

The modulation of HDAC functions can improve CVDs such as cardiac hypertrophy, heart failure, arrhythmia, myocardial infarction, hypertension, atherosclerosis, and fibrosis.386, 387 Although both class I and class II HDACs have conserved HDAC domains, they have completely different functions in CVDs. Class I HDACs have procardiac hypertrophic effects, whereas class II HDACs are expressed in a relatively tissue-specific manner and have anticardiac hypertrophic effects.388 HDAC7 is localized in the cardiac cytoplasm, and its overexpression induces the expression of cardiac hypertrophy and heart failure-related genes such as Nppa and Nppb.389

Vascular endothelial dysfunction is the main cause of CVDs, and one of the characteristics of endothelial dysfunction is insufficient synthesis of nitric oxide (NO). The main enzyme responsible for the synthesis of NO in endothelial cells is endothelial nitric oxide synthase (eNOS). The interaction between SIRT1 and eNOS can activate eNOS by reducing its acetylation level, which promotes NO production and vasodilation.390 CKIP-1 regulates physiological cardiac hypertrophy by inhibiting HDAC4 phosphorylation.391 In an AngII-induced mouse model of pathological cardiac hypertrophy, the pan-HDAC inhibitor (HDACI) emodin ameliorates hypertrophy by inhibiting the activity of class I, IIa, and IIb HDACs.392 HDACs are also involved in myocardial fibrosis. Overexpression of class I HDACs significantly enhances the proliferation of cardiac fibroblasts and the expression of proteins associated with fibrosis. Silencing of HDAC3 upregulates miR-18a and reduces ADRB3 expression, thereby inhibiting cardiomyocyte fibrosis and hypertrophy.393 The HDAC8 inhibitor PCI34051 mediates the p38 MAPK pathway to alleviate isoproterenol-induced cardiac hypertrophy and fibrosis394 and attenuates myocardial fibrosis induced by transverse aortic constriction in mice through downregulation of Ace1.395 The sirtuin family also plays key roles in preventing cardiomyocyte fibrosis, regulating cardiomyocyte apoptosis, improving cellular energy metabolism remodeling and inflammation, and maintaining cardiac homeostasis.396 Stimulation of SIRT3 reduces ROS and protein kinase levels and prevents cardiac hypertrophy, which may be a mechanism to inhibit cardiac remodeling.397 Last, p300 is a potential therapeutic target for heart failure. Mice with p300 knockout exhibit remarkable cardiac defects and embryonic lethality (Table 3).398, 399

3.5 Acetylation in neurodegenerative diseases

The imbalance between acetylation and deacetylation processes is related to neurodegenerative diseases such as AD and HD.400, 401 Abnormal histone acetylation in AD affects the expression of memory-related genes and dysregulates several signaling pathways, including cell differentiation, apoptosis, inflammation, and neuronal and vascular remodeling.402, 403 In a transgenic AD fly model, loss of Tip60 activity significantly increases the transcriptional expression of amyloid precursor protein (APP), leading to neuronal apoptosis, while overexpression of Tip60 HAT activity can potentially serve as a neuroprotective agent.404 p300/CBP is widely expressed in the nervous system. It has been proposed that inhibiting the activity of CBP/p300 acetyltransferase may affect the death of brain neurons and the long-term memory of animals.405 Among various HDACs, HDAC2 modulates chromatin plasticity to regulate the expression of learning and memory-related genes, and its dysregulation leads to the dysfunction of cholinergic nbM neurons, neurofibrillary tangle (NFT) pathology, and cognitive decline in AD.406 HDAC3 controls gene expression during the development and maintenance of neural stem cells,407 while HDAC4 may also play a role in the area of learning and memory. Selective deletion of HDAC4 in the brain leads to impaired long-term synaptic plasticity.408 HDAC6 plays a leading role in neuronal health or dysfunction. Selective inhibition of HDAC6 can promote growth cone function, synaptic plasticity, transport, and autophagosomal degradation, which can help protect neurons.409 Notably, HDAC6 is significantly elevated in the brains of AD patients. Sirtuins restore protein microenvironmental homeostasis mainly by reducing toxic protein aggregates. They also improve neural plasticity by increasing gene transcription activity, which can reduce oxidative stress, enhance mitochondrial function, and improve learning and memory abilities.410 SIRT3 expression is significantly increased in the temporal cortex in AD patients.411 High SIRT3 expression can promote antioxidant effects in mutant HTT cells, enhance mitochondrial function, and exert neuroprotective effects in HD.412 In PD mice, a neuroprotective effect of SIRT3 has also been found.413 SIRT3 may play a protective role in neurons by scavenging free radicals in mitochondria.414 Decreased SIRT3 function increases mitochondrial oxidative stress and cell death in substantia nigra dopaminergic neurons in PD models.415 The expression of SIRT3 is significantly reduced in MPTP-induced PD cell models, and overexpression of SIRT3 inhibits cell apoptosis. PGC-1α can promote the transcription of SIRT3 and inhibit the loss of dopaminergic neurons (Table 3).416

3.6 Acetylation in cancers

Abnormal acetylation exists in various cancers.310 Most histones are in a hypoacetylated state in tumor cells, and mutations in the acetyltransferases CBP and p300 are often found in tumors. An imbalance in acetylation leads to dysregulated gene expression related to cancer cell proliferation, differentiation, migration, invasion, and apoptosis.417, 418 For example, H4K16ac alters the chromatin state and promotes gene transcription to regulate tumorigenesis and development.419 miR24-2 inhibits histone deacetylase HDAC3 through miR675 to promote histone H4K16ac, which acts on PI3K and enhances the interaction between LC3 and ATG4, consequently triggering autophagy that affects cancer cell proliferation.420 The acetylation of the cytoskeleton is related to tumorigenesis. The acetylation of α-tubulin, a component of the cytoskeleton, is an important indicator of microtubule stability. Tubulin is the target of many anticancer drugs.421 Tumors are resistant to apoptosis. PDCD5, a protein related to apoptosis, can bind to Tip60 and increase p53 acetylation at K120, which affects the expression of apoptosis-related genes such as Bax.422 The acetylation of the hypoxia-induced autophagy regulator PAK1 regulates the phosphorylation of ATG5 at Thr101 in GBM and is important for hypoxia-induced autophagy and tumor growth.423

Increasing evidence shows that carcinogenesis is affected by metabolism in the body. Most metabolic proteins are substrates of lysine acetylation,421 such as ATM, ABL1, CDK9, BTK, and CDK1. PKM2 is the last rate-limiting enzyme in the glycolytic pathway responsible for the conversion of phosphoenolpyruvate to pyruvate. In a high glucose environment, PCAF acetylates PKM2 at K305, which reduces its binding to the substrate PEP, inhibits its enzymatic activity and promotes its chaperone-mediated autophagy and lysosome-dependent degradation.424 The acetylation of LDHA at K5 inhibits its enzymatic activity and is recognized and mediated by the heat shock protein HSC70, which downregulates the level of LDHA. The acetylation level of LDHA at K5 in early PC tissues is significantly lower than that in adjacent tissues, suggesting that acetylation of LDHA at K5 may be related to the occurrence of PC.425 FBP1 is the rate-limiting enzyme in gluconeogenesis and is lost in many types of cancer. Reduced FBP1 is associated with poor prognosis in HCC. HDAC-mediated repression of FBP1 expression is associated with a reduction in H3K27ac in the FBP1 enhancer.426 PDC is located within the mitochondria and is responsible for the irreversible conversion of pyruvate to acetyl-CoA. Phosphorylation of PDP1 at Tyr381 triggers SIRT3 to detach from the PDC center but recruits the acetyltransferase ACAT1 to the PDC center to acetylate PDP1 at K202 and PDHA at K321. This reconstructs the structure of the PDC center and inhibits PDC activity, further promoting tumor cell proliferation and growth.427 6PGD is an important enzyme in the PPP. Acetylation of 6PGD at K294 promotes the formation of highly active 6PGD dimers, thereby further activating the 6PGD and PPP pathways to produce more ribulose-5-phosphate and NADPH for nucleic acid synthesis and resisting oxidative free radical damage.428 Fatty acid metabolism is important for tumor growth and metastasis. Acetylation of FASN in the fatty acid synthesis pathway promotes its degradation. The deacetylation process is regulated by HDAC3, which functions in the initiation and development of liver cancer.429 Furthermore, SIRT4 can regulate branched-chain amino acid catabolism by deacetylating BCAT2 and promote PDAC growth (Table 3).430

3.7 Acetylation-associated targeted therapies

Due to the important functions of acetylation in diseases, HDACIs have now shown good application prospects in the treatment of various diseases, such as heart disease, diabetes, and cancers.446 Currently, HDACIs can be divided into four classes, including short-chain fatty acids (SCFAs) predominantly inhibiting class I HDACs (e.g., butyrate, phenylbutyrate, and valproate), hydroxamic acids inhibiting class I and II HDACs (e.g., trichostatin A (TSA) and suberoylanilide hydroxamic acid (SAHA)), cyclic tetrapeptides displaying class I HDAC selectivity in vitro, and benzamides inhibiting class I HDACs (e.g., RGFP136 and MS-275).447 Restoring normal protein acetylation may be a new approach for the treatment of malignant tumors.448

Upregulation of HDAC expression is a characteristic of various malignant cancers,449 such as prostate cancer,450 gastric cancer,451 breast cancer,452 renal cancer,453 and Hodgkin's lymphoma.454 HDACIs can inhibit tumor cell proliferation by inducing cell differentiation, growth arrest, and apoptosis.455, 456 A variety of HDACIs have been approved or entered clinical trials. HDACIs not only show direct inhibitory effects on tumor cells but also overcome the resistance of tumors to other drugs, which makes the combination of HDACIs and other antitumor drugs possible.457

HDACIs have bidirectional effects on inflammation and anti-inflammatory effects. HDAC inhibition may not only increase inflammation but also attenuate the expression of specific genes to reduce infiltrating inflammatory cells, leading to a beneficial result.458 Additionally, HDACIs can reverse neuronal degeneration and aging and enhance synaptic plasticity in mouse models. On the one hand, HDACIs have a direct influence on gene transcription by remolding histone acetylation. On the other hand, HDACIs can increase the acetylation of transcriptional regulators such as HNF4a to indirectly regulate gene expression.459 HDACIs have been found to exhibit neuroprotective effects on neurological diseases such as PD, AD, amyotrophic lateral sclerosis (ALS), and HD.460 For example, the class I HDACI valproic acid may exert neuroprotective effects by regulating the BDNF/TrkB signaling axis.461 In a cellular model of PD patients, selective inhibition of SIRT2 increases tubulin acetylation and improves microtubule-mediated transport.462 Moreover, SIRT2 inhibitors such as AGK2, AK-7, and AK1 have been demonstrated to decrease neuroinflammation and cytotoxicity induced by toxins or mutant protein aggregation.463 Additionally, the HDAC6 inhibitor tubastatin A, which increases autophagic flux and protects neurons in HD patients, is a potential drug for the treatment of HD.464

Deacetylase inhibitors are also used to treat metabolic disorders. SAHA has been shown to target eNOS uncoupling and oxidative stress in diabetes.465 TSA can prevent ischemia-induced left ventricular remodeling by inhibiting TNF-α transcription. In addition, it also enhances AKT phosphorylation to promote angiogenesis and cardiomyocyte survival.466 However, a SIRT6 inhibitor aggravates diabetes-induced cardiomyocyte apoptosis and fibrosis in mice by increasing the levels of inflammatory factors and ROS.467

4 OTHER SCFA MODIFICATIONS

SCFAs are products of food digestion and dietary fiber fermentation in the gut containing fewer than six carbons.468 Microbial-derived metabolites have deleterious and beneficial effects on human health.468 They have a wide range of functions in signaling, cellular metabolism, and immunity.469 SCFAs can be transformed into acyl-CoAs,470 which act as donors of protein lysine acylation.471 In addition to acetylation, the most extensively studied SCFA modification, introduced above, many other types of SCFA-derived modifications have been identified, including propionylation (Kpr),472 butyrylation (Kbu),472 2-hydroxyisobutyrylation (Khib),473 succinylation (Ksucc),474 isobutyrylation (Kisobu),475 malonylation (Kmal),476 glutarylation (Kglu),477 crotonylation (Kcr),478 β-hydroxybutyrylation (Kbhb),479 and lactylation (Kla) (Figure 6).480 To date, hundreds of histone acylation sites have been identified (Figure 7), and numerous studies have demonstrated the important roles that SCFA modifications play in both health and disease.481

Details are in the caption following the image
Chemical structures and representative biological functions of SCFA-derived lysine acylation modifications, including propionylation (Kpr), butyrylation (Kbu), succinylation (Ksucc), 2-hydroxyisobutyrylation (Khib), isobutyrylation (Kisobu), malonylation (Kmal), glutarylation (Kglu), crotonylation (Kcr), β-hydroxybutyrylation (Kbhb), and lactylation (Kla).
Details are in the caption following the image
Distribution of reported lysine acylation modifications on histones.

The acylation modifications derived from SCFAs are regulated by “writers” and “erasers.” Writers are enzymes that promote lysine acylation modifications, and erasers are enzymes that remove acylation modifications (Figure 8).482 In the past few years, an increasing number of studies have shown that classic acetyltransferases and deacetylases, such as p300, TIP60, HDACs, and SIRTs, also regulate other types of acylation modifications (Table 4).34, 483, 484

Details are in the caption following the image
Origin and regulation of SCFA. SCFA-derived lysine acylation modifications. The SCFA is converted into the corresponding acyl-CoA in the presence of enzymes such as ACCS2. The acyl groups can be transferred onto proteins to modify the side chain of lysine residues by HATs. In addition, the acylated proteins can be deacylated by HDACs and sirtuins.
TABLE 4. The enzymatic specificities of different writers and erasers.
Acylations from SCFAs Writers Erasers
Propionylation GCN5,485 PCAF,486 P300/CBP,472 MYST (MOF, MOZ, HBO1),487 KAT6A488 SIRT1, SIRT2, SIRT3, SIRT5489
Butyrylation p300,472, 490 P300/CBP,472 GNAT491 SIRT3,492 SIRT5,493
Malonylation NA SIRT3,493 SIRT2, SIRT5494
Succinylation KAT2A495 SIRT5,483 SIRT7484
2-Hydroxyisobutyrylation P300,487 Tip60, EP300496 SIRT5,497 HDAC2,496 HDAC3498
β-Hydroxybutyrylation P300/CBP, MYST, GNAT499 HDAC1, HDAC2, SIRT1–3500
Crotonylation P300/CBP,501 GNAT, MYST,502 HBO1, KAT6A, MOF, PCAF, TIP60503 SIRT1–3, HDAC1–3, HDAC8502
Glutarylation P300,479 KAT2A504 SIRT5,477, 483 SIRT7505
Lactylation p300480 HDAC1–3506

Kpr is a widely distributed PTM. The propionyl-CoA donor for Kpr is derived from odd-chain fatty acid oxidation (FAO) and branched-chain amino acid catabolism.507 Kpr is mainly found in proteins involved in energy production and conversion, participates in various metabolic processes, and plays an important role in protein breakdown.508-510 Histone propionylation, acetylation, and butyrylation levels also change in response to cellular metabolic changes, and these modifications regulate chromatin structure and function as important markers of such changes.491

Butyryl-CoA is a donor of Kbu derived from even-chain fatty acids.511 Kbu not only regulates transcription but also regulates RNA metabolism, chromatin organization, and DNA repair.512 The binding of a testis-specific member, Brdt, can be inhibited by histone butyrylation, which affects the differentiation of male germ cells.513

Kisobu is a recently discovered isomeric modification of Kbu. Isobutyrylation and butyrylation are derived from different donors, isobutyryl-CoA and butyryl-CoA, whose biosyntheses are different in mammalian cells. Butyryl-CoA is derived from the metabolism of fatty acids, while valine metabolism contributes to isobutyryl-CoA production. Kisobu is also involved in the regulation of gene transcription.475

The crotonylation donor crotonyl-CoA can be produced by ACCS2-catalyzed crotonic acid metabolism514 or converted by the butyrate β-oxidation pathway.515 During FAO, ACADS and ACOX3 are key enzymes that catalyze the conversion of butyryl-CoA to crotonyl-CoA,516 and Kcr is involved in several physiological processes in humans, such as DNA damage repair,517 chromatin reorganization, RNA processing, and regulation of protein activity and localization.518

Lactic acid serves as a carbon source in organisms and is the precursor of Kla,480, 519 which is derived from lactyl-CoA generated by glycolytic conversion of glucose.480 Lactyl-CoA is then transferred to the lysine side chain of proteins through transferases.480 Kla can be inhibited by glycolysis inhibitors and boosted by mitochondrial inhibitors or hypoxia, all of which affect lactate production.492Lactate stimulates histone Kla and influences gene transcription480 and is involved in important life activities such as anti-inflammation,520 immune regulation,521 and gene reprogramming.522 In addition, histone Kla also inhibits the activation of inflammatory macrophages by promoting M2-like polarization.492

Kmal refers to the addition of malonyl groups to lysine side chains. 483 Malonyl-CoA, as a reactive donor of Kmal, can inhibit glycolysis-related enzyme activities by modifying them.523 In addition, Kmal can modify many proteins and affect the related signaling pathways, including fatty acid synthesis and oxidation,34, 524 mitochondrial respiration,524 glycolysis,524, 525 and histones.526 Kmal also acts as a signal to regulate macrophage mRNA binding to promote inflammation. 483 In cells lacking FASN, malonyl-CoA accumulation can lead to mTOR malonylation and affect mTORC1 signaling.527

Ksucc is a process of covalently attaching a succinyl group to the lysine side chain in an enzymatic or nonenzymatic manner.479 Ksucc occurs mainly in the mitochondria,528 where succinyl-CoA is produced by amino acid metabolism or the TCA cycle.529 Moreover, succinyl-CoA and Ksucc are highly abundant in tissues, such as the heart, brown adipose tissue and liver, with greater numbers of mitochondria.530 Ksucc participates in energy metabolism in vivo,481 causes protein conformation changes,17 and regulates nuclear function.531

Kglu, a reversible, dynamic and conserved modification, is produced by covalently binding glutaryl groups to lysine residues504 and occurs mainly in mitochondria.532 Kglu plays an important role in regulating protein structural changes,532 oxidative damage,533 mitochondrial functions,534 and sperm motility.535

2-Hydroxyisobutyryl-CoA is a potential donor for Khib. Tip60 and p300 are identified as 2-hydroxyisobutyryltransferases, while HDAC2 and HDAC3 are de-2-hydroxyisobutyrylases.473, 496 Khib plays critical roles in the regulation of gene transcription, cell growth and cellular metabolism. It not only affects the binding interaction between histones and DNA536, 537 but also participates in the regulation of metabolic pathways such as glycolysis/gluconeogenesis and the TCA cycle. In addition, Khib also affects the motility of human sperm.538

3-Hydroxybutyrate is a metabolic component of ketone bodies that provides energy for the heart and brain during periods of starvation.499, 539 Hypoglycemia leads to ketogenesis, producing β-hydroxybutyrate.479 β-Hydroxybutyrate forms covalent bonds with lysine side chains in proteins during ketogenesis, leading to Kbhb.536, 540 Kbhb is a sensitive indicator of changes in energy metabolism. Under starvation conditions, histone Kbhb levels in the mouse liver are significantly elevated, which can impact metabolic pathways such as amino acid catabolism.536 Histone Kbhb can promote the transcription of the BDNF gene.536 In addition, Kbhb also contributes to the regulation of chromatin structure.512

Protein acylation not only regulates cellular processes such as gene transcription and cellular metabolism, but also plays a role in the regulation of health and disease.541 The microbiota-dependent synthesis of many metabolites, particularly SCFAs, affects human health.542 SCFA modification, as an epigenetic mechanism, regulates key functions of various proteins related to growth, metabolism, cell differentiation and apoptosis, inflammation, aging, and angiogenesis. It plays a role in many diseases, including cancer, neurological and psychiatric disorders, CVD, diabetes, hepatitis, and kidney disease (Figure 9).479, 503, 543, 544

Details are in the caption following the image
The functions of SCFA-derived acylation modifications in various human diseases. Representative examples are shown.

4.1 CVDs

SCFA modifications play a role in regulating the progression of CVDs through enzymatic switches and oxidative stress. Additionally, they can affect the cellular localization and PPIs of many cardioprotective proteins.545 Maintaining metabolic stability of SCFAs is essential for cardiac and vascular function.543 p53 Kbhb mediates the protective effect of β-hydroxybutyrate on vascular cell senescence, possibly by reducing the expression of p21 and PUMA, which leads to cell growth arrest and reduced apoptosis.499 Kpr can cause FAO disorders and impair mitochondrial functions.524 H3K14pr is highly expressed in promoter regions of transcriptionally activating genes, including FAO-related genes associated with CVD progression.491 SIRT5 defects lead to an increase in succinyl-coA in the heart.504 Elevated levels of Ksucc can lead to hypertrophic obstructive cardiomyopathy,528 while oxidative stress caused by Kglu may be an important mechanism for inducing CVDs.546 A high-fat diet causes adverse effects on cardiovascular health, especially under stressful conditions. The levels of H3k9bu affected by ACADS can moderate the expression of stress-regulated genes.547

4.2 Metabolism-associated diseases

Metabolic disorders cause dysregulated SCFA modifications, which can lead to various metabolic diseases.17, 548 Obesity can lead to sperm DNA damage and decreased sperm quality,549 and SCFA modifications cause male reproductive dysfunction in obese men. In high-fat diet mice, acetylation and crotonylation decrease in the testes, while other metabolism-related lysine acylations, including propionylation, malonylation, succinylation, glutarylation, 2-hydroxyisobutyrylation, and benzoylation, increase.550 Obesity and metabolic syndrome accelerate the occurrence of osteoarthritis during the aging process, and SIRT5-regulated malonylation may impair chondrocyte metabolism.551 The mechanism by which energy restriction improves fat metabolism may be involved in Ksucc. Acute fasting regulates Ksucc through SIRT5 to modulate lipid metabolism in adipose tissues and improve obesity.552 In addition, specific inhibition of p300-mediated butyrylation at H4K5 by LTK-14A in adipocytes and liver improves obesity.490

Decreased SCFA modifications, such as malonylation, butyrylation, and propionylation, are found in liver histones of obese mice induced by a high-fat diet.553 Dapagliflozin treatment leads to elevated 3-hydroxybutyrate in the plasma and adipose tissues of obese diabetic mice, which further induces H3K9 3-hydroxybutyrylation to promote the expression of apolipoproteins in adipocytes. Apolipoproteins are anti-inflammatory and antiatherosclerotic, indicating that 3-hydroxybutyrate is protective against obesity-associated diabetes by modulating H3K9 3-hydroxybutyrylation.554 In addition, the dysregulation of Kmal in fatty acid oxidative metabolism can also lead to mitochondrial fatty acid metabolism diseases such as malonyl-CoA synthetase ACSF3 deficiency.532 These studies provide additional evidence for the link between metabolic disorders and epigenetic regulation by SCFA modifications.

4.3 Kidney diseases

Kidney disease includes acute kidney injury (AKI) and chronic kidney disease (CKD), some of which progress to end-stage renal disease (ESRD). SCFA modification, as a type of epigenetic mechanism, is involved in the progression of kidney disease.481 Histone lysine crotonylation was observed in mouse and human renal tubular cells, and histone crotonylation was observed to be increased in renal tissues during AKI. Crotonate supplementation can increase overall histone crotonylation and have a protective effect on the kidneys.555 Crotonyl-CoA hydratase CDYL regulates the crotonylation of histone H3K18 and affects the disease process of ADPKD.556 Diabetic kidney disease (DKD) is the main cause of ESRD, and inflammation and fibrosis are key processes in its development. Butyrate inhibits the expression of renal inflammation and fibrosis genes through p300-mediated histone Kbu and improves DKD.557 Other studies have found that crotonylation and 2-hydroxyisobutyrylation also play a significant role in ESRD. By analyzing crotonylation and 2-hydroxyisobutyrylation in PBMCs of patients with ESRD, it is speculated that crotonylation and 2-hydroxyisobutyrylation may affect immune cell numbers and induce immune senescence, which may be due to regulation of the glycolytic/gluconeogenesis pathway and protein processing.558 In STZ-induced diabetic SD rats, 3-hydroxybutyrate treatment increases H3K9 3-hydroxybutyrylation in the gene promoter to upregulate MMP-2, which reduces collagen IV content and glomerular fibrosis.559

4.4 Cancers

Metabolic reprogramming is a common feature of cancer.560 The “Warburg effect” is an important feature of tumor cell metabolism.548 Tumor cells predominantly rely on aerobic glycolysis for energy production and metabolite synthesis, resulting in extracellular acidification due to lactate accumulation, which is a hallmark of cancer.561 Lactate is involved in the regulation of the tumor microenvironment, which promotes macrophage polarization into an M2-like phenotype, thereby inhibiting the immune response in the tumor microenvironment.480 Histone Kla has been proven to promote the development of tumors. Histone Kla regulates the transcription of the m6A reader YTHDF2, which can recognize the m6A modification on the 3'UTR of the tumor suppressor genes PER1 and TP53 mRNA, resulting in their degradation and further impacting the development of melanoma.562 Kla is more abundant in gastric cancer tumor tissues than in adjacent normal tissues, indicating its potential as a prognostic indicator for gastric cancer.563 In PDAC, elevated tumor-mediated Kla levels can enhance the expression of cancer-associated fibroblasts (CAFs) to promote cancer cell invasiveness.522

Besides lactylation, other SCFA modifications also contribute to cancer development. In PDAC, SUCLA2-coupled regulation of GLS succinylation promotes tumor growth.564 Lower levels of the mitochondrial protein GCDH result in Kglu of the TF NRF2, leading to cell death, indicating that GCDH pathway inhibition is a potential therapeutic strategy for melanoma treatment.565 Khib is widely distributed in PC, and treatment with the TIP60 inhibitor MG149 can significantly reduce the total Khib level in PC, which leads to the inhibition of PC cell proliferation, migration, and invasion.498 Kcr is also involved in cancer regulation, with decreased levels observed in liver, gastric, and kidney cancers and increased levels in thyroid, esophageal, and PC.566 In addition, p300-mediated crotonylation enhances the expression of HNRNPA1, promoting HeLa cell malignancies.567 By knocking out HDACs or adding the HDACI TSA, the level of Kcr is increased, which inhibits the motility and proliferation of HCC cells.566 The level of Kcr in prostate cancer tissue is higher than that in adjacent tissues, and its level increases as the malignancy increases.568 Hyperpropionylation of H3K23 in the leukemia cell line U937 may serve as a stage-specific biomarker in hematopoiesis and leukemogenesis.569 Moreover, propionate can induce global Kpr, which inhibits colon cancer development by upregulating the expression of MICA/B.570 Ksucc can affect the synthesis of thyroid hormones. Radiation-induced thyroid cancer and cancer cell metastasis in apoptotic cell lines can be inhibited by Ksucc.571 Downregulation of TFAM can induce Kmal of mDia2 to promote its nuclear translocation, which further induces lung metastasis of mouse liver cancer.572

4.5 Neurological diseases

SCFA modifications are closely related to neurological function. In cases of syndromic intellectual disability, there is an alteration in histone H3K23 propionylation.488 The level of Kbu is significantly changed in the brains of rats with vascular dementia compared to the control group.573 Crotonylation and succinylation levels are increased in the cerebral cortex of mice with developmental disorders of the central nervous system (CNS).574 Neural excitation can modulate Kla levels in brain cells, suggesting that protein Kla in the brain may be associated with neuropsychiatric disorders.575 CDYL mediates histone crotonylation, which regulates gene transcription and promotes the development of stress-induced depression in rodents, providing a potential therapeutic target for major depression.576 Lysine succinylation and malonylation are related to protein regulation, glycolysis, and energy metabolism. Mitochondrial dysfunction causes an imbalance in succinylation, which in turn leads to schizophrenia and other psychiatric disorders.577 Histone Kbhb is enriched in the promoter region of active genes and affects the organism by reprogramming the epigenetic map.536 Kbhb plays an important role in the development of neurological diseases, and 3-hydroxybutyrate can be used to treat certain neurological diseases, such as epilepsy and AD.578, 579 Experiments in mice have shown that 3-hydroxybutyrate may alleviate depressive behavior by increasing histone H3k9 3-hydroxybutyrylation.580

In patients with AD, the succinylation levels of various mitochondrial proteins are decreased, while the succinylation of APP is increased. This disrupts the normal proteolytic process of APP and leads to abnormal protein deposition in the brain.581 In the AD mouse model, elevated levels of H4K12 lactylation in microglia activate the transcription of glycolytic genes, resulting in proinflammatory activation of microglia.582 During mammalian development, histone crotonylation and lactylation are widely distributed in the brain and play important roles in neurodevelopmental processes by contributing to transcriptome remodeling.583

5 LONG-CHAIN FATTY ACID MODIFICATIONS

Straight-chain fatty acids with 12 or more carbon atoms are referred to as long-chain fatty acids (LCFAs), such as myristic acid, oleic acid, linoleic acid and palmitic acid.584 Similar to SCFAs, LCFAs can also be attached to the N-terminus and amino acid side chains of proteins through the action of enzymes. The most commonly modified proteins by LCFAs are palmitoylated and myristoylated.585 Palmitoylation, the covalent attachment of the palmitoyl group to protein amino acid side chains, is a widespread modification in organisms586 and plays a crucial role in regulating protein translocation, localization and stability.587 Myristoylation, also known as N-myristoylation, is an important PTM resulting from the covalent attachment of myristic acids to the N-terminus of proteins catalyzed by N-myristoyltransferase (NMT).588 It plays significant roles in innate immunity,589, 590 signal transduction,591 and cancer progression.592

5.1 Palmitoylation

The regulation of protein palmitoylation balance is carried out by palmitoyl acyltransferases (PATs) and depalmitoylating enzymes (Figure 10A).587 PATs belong to the PAT family and contain a conserved DHHC (aspartic acid-histidine-histidine-cysteine) motif, hence, they are referred to as DHHC-PAT .593 These enzymes are also known as zinc finger-containing DHHCs (ZDHHCs) as the DHHC motifs form zinc finger domains. Palmitoylation can be categorized into three types based on the way palmitoyl groups are attached to proteins: S-type, N-type, and O-type.594 S-type palmitoylation refers to the attachment of palmitoyl groups to the cysteine residues of proteins through an unstable thioester bond. N-type palmitoylation refers to the attachment of palmitoyl groups to the amino groups of various amino acids (e.g., glycine, cysteine, lysine), while O-type palmitoylation is the attachment of a few palmitoyl groups to the hydroxyl groups of serine or threonine. S-palmitoylation, which dominates the majority of palmitoylated proteins and is a reversible process, represents typical palmitoylation. To date, 23 PATs have been discovered (Table 5).595 The thioester bond is hydrolyzed by depalmitoylases, leading to the dissociation of palmitoyl groups from cysteine residues. Five depalmitoylases have been found to date,596-600 including PPT1, PPT2, APT1, APT2, and ABHD17. PPT1, which is located mainly in lysosomes, is a thioesterase that mediates the depalmitoylation of various palmitoylated proteins in neurodegenerative diseases.596 PPT2,which has a different crystal structure from PPT1, is essential for depalmitoylation in protein degradation.601 APT1, which is mainly localized in the cytoplasm of yeast and mammalian cells, is a highly conserved α/β hydrolase containing the S-H-D catalytic triad and the G-X-S-X-G motif, with palmitoylated Ras proteins being its main substrates.602 APT2 is highly homologous to APT1.603 ABHD17 is essential for N-Ras depalmitoylation and the relocalization of N-Ras to internal cellular membranes.604

Details are in the caption following the image
The process of S-palmitoylation and N-myristoylation. (A) The S-palmitoylation and depalmitoylation process. Step1, palmitoyl acyltransferases (PATs) undergo auto-palmitoylation, and the palmitoyl group is transferred to PAT; Step2, the palmitoyl group is transferred from PAT to protein substrates. (B) The process and functions of N-myristoylation. Myristic acid and coenzyme A are converted to myristic acid coenzyme A by acetyl-CoA synthetase. If the starting amino acid of the protein is methionine, it needs to be removed by methionyl aminopeptidase (MetAP2) before N-myristoylation. N-myristoyltransferase (NMT) is responsible for the addition of the myristoyl group to the glycine residue at the N-terminal of the protein.
TABLE 5. Twenty-three palmitoyl acyltransferases and their localization and functions.
Gene Localization Biological functions
ZDHHC1 ER Ablation of ZDHHC1-mediated p53 palmitoylation help cancer cells escape from the suppression of p53.605
ZDHHC2 ER, Golgi, dendritic vesicle in neuron ZDHHC2 plays a critical role in inflammatory response of psoriasis606; ZDHHC2 is critical for the proliferation and the survival of B cells607; ZDHHC2 shows association with neurological diseases608-610; The C-terminal domain of ZDHHC2 can regulate intracellular localization in neurons611; palmitoylation of SARS-CoV-2 spike protein is critical for virus entry.612
ZDHHC3 Golgi Palmitoylation of PD-L1 by ZDHHC3 inhibits antitumor immunity in vitro613; palmitoylation of ACE2 by ZDHHC3 is critical for the membrane-targeting of extracellular vesicles secretion614; elevated expression of ZDHHC3 is correlated with poor survival in breast cancer615; high ZDHHC3 levels inhibits synaptic plasticity and memory in high-fat diet (HFD) mice616; ZDHHC3 regulates the infection of primary and latent herpes simplex Virus 1.617
ZDHHC4 ER, Golgi ZDHHC4 palmitoylates KAI1 and affects its localization, inhibiting angiogenesis618; GSK3β palmitoylation mediated by ZDHHC4 promotes tumorigenicity of GBM stem cells619; palmitoylation of D2R by ZDDHC4 is important for cell surface expression of the receptor.620
ZDHHC5 Plasma membrane, endosomes in dendritic shafts ZDHHC5 plays important role in synaptic plasticity, cardiac function, cell adhesion, and fatty acid uptake; ZDHHC5 interacts with SARS-CoV-2 spike protein and affects their subcellular localization and pseudovirus entry621; Circ-ZDHHC5 accelerates esophageal squamous cell carcinoma progression in vitro.621
ZDHHC6 ER ZDHHC6-mediated palmitoylation restrains FLT3-ITD surface expression, signaling, and colonogenic growth in AML622; intracellular MYD88 palmitoylation by ZDHHC6 is a therapeutic target of sepsis623; ZDHHC6 palmitoylates NRas, contributing to its subcellular localization, and improves the downstream proproliferative signaling cascades in cancers.624
ZDHHC7 Golgi ZDHHC7 regulates neuronal development and plasticity and modulates structural connectivity between hippocampus and medial prefrontal cortex in mice625; ZDHHC7 palmitoylates sex steroid hormone receptors and correlates with mental disorders626; palmitoylation of CD36 by ZDHHC7 are critical in NASH development.627
ZDHHC8 Golgi, dendritic vesicles, spines in neuronal cells ZDHHC8 regulates seizure susceptibility in epilepsy628; ZDHHC8 palmitoylates scribble and Ras64B and controls growth and viability in Drosophila629; ZDHHC8 and ZDHHC5 are present in dorsal root ganglion (DRG) axons and control retrograde signaling via the Gp130/JAK/STAT3 pathway.630
ZDHHC9 ER, Golgi ZDHHC9-mediated GLUT1 S-palmitoylation promotes GBM glycolysis and tumorigenesis595; ZDHHC9 is essential for dendrite outgrowth and inhibitory synapse formation631; ZDHHC9 plays a critical role in intellectual disability.632
ZDHHC11 ER ZDHHC11 is a positive modulator in NF-κB signaling633; ZDHHC11 mediates MITA-dependent innate immune responses against DNA viruses634; ZDHHC11 is a critical novel component of the oncogenic Myc-miR-150-MYB network in Burkitt lymphoma.635
ZDHHC12 ER, Golgi CLDN3 palmitoylated by ZDHHC12 contributes to plasma membrane localization and protein stability of CLDN3, thus promoting the progression of ovarian cancer636; ZDHHC12 can promote the proliferation and migration of glioma cells637; palmitoylation of gephyrin by ZDHHC-12 contributes to coordinated neurotransmission.638
ZDHHC13 ER, Golgi MC1R palmitoylation mediated by ZDHHC13 activates MC1R signaling, and affects senescence and melanomagenesis639; Drp1 palmitoylation by ZDHHC13 impacts brain bioenergetics and anxiety640; ZDHHC13 regulates skin barrier development by controlling protein stability641; ZDHHC13 regulates mitochondrial functions and metabolism in liver.642
ZDHHC14 ER Palmitoylation induced by ZDHHC14 is of vital importance in control of neuronal excitability643; the expression of ZDHHC14 is inhibited, leading to increased proliferation and decreased apoptosis in coronary artery disease644; ZDHHC14 is involved in the palmitoylation of SARS-CoV-2 spike protein and contributes to virus entry.612
ZDHHC15 Golgi ZDHHC15 mutations lead to psychiatric diseases645; ZDHHC15-mediated palmitoylation may be a novel regulatory mechanism of dopamine in the striatum of mice646; ZDHHC15 regulates the formation of dendrite morphology and excitatory synapse647; ZDHHC15-mediated GP130 palmitoylation is critical in the growth and self-renewal of GBM stem cells.648
ZDHHC16 ER Reduced ZDHHC16 contributes to p53 activation in GBM649; ZDHHC16 plays a crucial role in regulating neural stem/progenitor cells proliferation during zebrafish telencephalic development650; ZDHHC16 is involved in early stages of DNA damage response.651
ZDHHC17 Golgi, intracellular vesicles, presynaptic terminals ZDHHC17 activates JNK and p38 MAPK and drives multiforme development and malignant progression in GBM652; ZDHHC17 interacts with CALCOCO1 and mediates selective Golgi autophagy653; ZDHHC17 is involved in the control of somal and distal axon integrity.654
ZDHHC18 Golgi ZDHHC18 negatively regulates cGAS-mediated innate immunity through palmitoylation655; MDH2 palmitoylation by ZDHHC18 sustains mitochondrial respiration and promotes the progress of ovarian cancer656; ZDHHC18 can regulate the cellular plasticity of glioma stem cells and contributes to their survival.657
ZDHHC19 ER Zdhhc19 is dispensable for spermatogenesis and sperm functions in mice658, 659; ZDHHC19 accelerates tumor progression through wnt/β-catenin pathway in osteosarcoma660; Flotillin-1 palmitoylation turnover by APT-1 and ZDHHC-19 promotes cervical cancer progression.661
ZDHHC20 Plasma membrane ZDHHC20-mediated palmitoylation controls SARS-CoV-2 membrane lipid organization and enhances its fusion capacity662; palmitoylation by ZDHHC20 targets ORAI1 channels to lipid rafts for efficient Ca2+ signaling in immune responses663; palmitoylation of IFITM3 by ZDHHC20 enhances its antiviral activity.664
ZDHHC21 Golgi, plasma membrane 5-HT1AR is palmitoylated by ZDHHC21 and reduced 5-HT1AR palmitoylation is involved in depression665; ZDHHC21 mediates signaling events required for gut hyperpermeability induced by inflammation666; DHHC21 can palmitoylate α1 adrenergic receptor and regulate vascular functions667; DHHC21 mediates endothelial dysfunction in systemic inflammatory response syndrome.668
ZDHHC22 ER, Golgi Palmitoylation of mTOR by ZDHHC22 can restrain breast cancer growth669; ZDHHC22 interacts with CCN3 and affects neuronal axon growth.670
ZDHHC23 ER, Plasma membrane ZDHHC23 dynamically regulates the functional coupling with β1-subunits and may be involved in cell-specific control of ion-channel physiology671; ZDHHC23 acts as potential regulators of tumor-infiltrating immune cells and glioma progression.672
ZDHHC24 ER High mRNA expression is an unfavorable prognostic marker in GBM.
  • Abbreviation: ER: endoplasmic reticulum

Palmitoylation is essential for protein localization. The palmitoylation of Cdc42 at Cys188 plays a crucial role in its localization to the plasma membrane, which regulates gene transcription and neuronal morphology in hippocampal neurons.673 The precise localization of calcineurin CNAβ1 and CD36 is also regulated by palmitoylation.674 Palmitoylation enhances the hydrophobicity of CD36,675 increasing its ability to bind to the membrane and absorb fatty acids.676 In contrast, inhibiting CD36 palmitoylation reduces its hydrophobicity and localization to the cytoplasmic membrane.677 The palmitoylation of CD36 is precisely regulated by DHHC4 and DHHC5.678 In addition, the palmitoyltransferase ZDHHC5 mediates the palmitoylation of NOD1/2 to promote its membrane recruitment and immune signaling, which are extremely important for microorganisms to establish an effective immune response.679

Cellular palmitoylation maintains a dynamic balance to ensure normal life activities, and any disruptions of this balance can lead to various diseases, such as autoimmune diseases,613 neurodegenerative diseases,593 T2DM,33 and nonalcoholic fatty liver disease,677 tumors,680-683 Friedreich ataxia, and peripheral artery disease.684, 685 Elevated palmitoylation of NOD2 mutants in autoinflammatory diseases leads to inflammation and inhibits autophagic degradation (Figure 11).686 In autoimmune diseases, palmitoylation activates the STING signal associated with the type I IFN response and induces the expression of inflammatory genes through recruitment of TBK1 and IRF3. Treatment with the palmitoylation inhibitor 2-bromopalmitate (2-BP) can abolish the type I IFN response by inhibiting the palmitoylation of STING.687 In major depressive disorder, deletion of ZDHHC21 reduces palmitoylation of 5-HT1AR and affects its signaling function.665 Reduced palmitoylation of Cdc42 in ZDHHC8-deficient neurons interferes with Akt/Gsk3β signaling, leading to schizophrenia.688 Increased palmitoylation in the brains of HD mice can alleviate the anxiety and depression behaviors of mice689 and reduce cytotoxicity in YAC128 neurons.32 CD36 palmitoylation on the plasma membrane in nonalcoholic steatohepatitis (NASH) is significantly increased, but inhibition of CD36 palmitoylation protects against NASH in mice.677 Aberrant protein palmitoylation mediates cell barrier disruption and leads to spermatogenesis dysfunction in spermatodysplastic patients.690 ZDHHC17 mediates palmitoylation of Oct4A in human GBM and protects it from lysosomal degradation, which maintains tumorigenicity.691 Palmitoylated PCSK9 can activate the PI3K/AKT pathway, confer drug resistance in HCC cells, and promote cancer cell proliferation.692 In addition, palmitoylated EGFR in TKI-resistant EGFR-mutant NSCLC cells positively regulates FASN and further promotes cancer cell growth through the Akt pathway.693

Details are in the caption following the image
The functions of LCFA-derived acylation modifications in various diseases, including palmitoylation and myristoylation. Representative examples are shown.

At present, various therapeutic strategies have been developed to address diseases caused by imbalanced palmitoylation. FLT3 is palmitoylated in primary human AML cells, which hinders the activation of AKT signaling and AML progression. A novel therapeutic strategy has been developed for FLT3-ITD+ leukemia by promoting FLT3 depalmitoylation.622 The palmitoyltransferase ZDHHC3 mediates the palmitoylation of PD-L1, which further inhibits PD-L1 ubiquitination and degradation. The palmitoylation inhibitor 2-BP can lower PD-L1 palmitoylation and boost PD-L1 degradation through the lysosomal pathway, thereby increasing the immune response of T cells against tumors.613 Furthermore, 2-BP can block STING palmitoylation and impair the type I IFN response.687 ABD957 acts as a potent and selective inhibitor of ABHD17 depalmitoylase. Specifically, it inhibits N-Ras depalmitoylation in AML cells and disrupts the balance of N-Ras palmitoylation, suggesting ABHD17 as a promising target for the treatment of N-Ras mutant tumors.604

5.2 Myristoylation

Protein myristoylation is a process mediated by NMT, with the majority of this transfer occurring at the amino group of glycine.585, 694 On rare occasions, it takes place at the side chain of lysine.695 Until now, no demyristoylase has been discovered, but a study has shown that a cysteine protease IpaJ expressed by Shigella bacteria can cleave myristoylglycine from the N-terminus of host proteins, serving a similar function as a hypothetical demyristoylase.696 The N-terminus of the myristoylated protein contains a conserved sequence Met-Gly-XXX-Ser/Thr (XXX could be any natual amino acid). The N-terminal amino acid methionine must be removed by methionine aminopeptidase to expose glycine before myristoylation (Figure 10B).697 There are two NMTs (NMT1 and NMT2) in humans, NMT1 and NMT2, which have 77% sequence similarity and partially overlapping substrates and biological functions.695, 698

Myristoylation affects the subcellular localization of proteins and regulates PPIs and protein–membrane interactions. Myristoylated proteins located on the membrane can trigger subsequent cellular responses by sensing and transmitting signals.699, 700 For example, the myristoylation of the signal peptide of the virus envelope glycoproteins is essential for the fusion of the virus with the cell membrane and promotes the virus infection of host cells.701 EV71 has a myristoylation modification site on the glycine residue of VP4, which improves membrane permeability. When this modification is missing, the viral genome replication is impaired.702 ZYG11B and ZER1 are E3 ligase complexes that control the quality of N-myristoylated proteins.585

N-myristoylation endows proteins with stronger hydrophobicity, and the disruption of cellular N-myristoylation balance may lead to the occurrence of malignant tumors, CVDs, and immune diseases.699, 703, 704 N-myristoylation of EZH2 promotes phase separation of EZH2 with its substrate STAT3, leading to the activation of STAT3 signaling and growth of lung cancer cells, making N-myristoylation of EZH2 a potential target for lung cancer therapy.705 In rheumatoid arthritis (RA), T cell deficits in NMT1 can lead to inflammation in synovial tissue due to impaired lysosomal transfer and AMPK activation.704 During vascular lesions, N-myristoylation of LMCD1 specifically suppresses E2F1 and NFATc1, resulting in increased CDC6 and IL-33, which further affects VSMC proliferation and migration.706

A variety of treatments have also been developed for diseases caused by N-myristoylation imbalance. NMT1 is significantly upregulated in bladder cancer, and high NMT1 expression is linked to poor patient prognosis. NMT1 mediates the myristoylation of LAMTOR1 at Gly2 to increase LAMTOR1 stability and lysosomal localization, which is critical for amino acid sensing and mTORC1 activation. The inhibitor B13 can abrogate the functions of NMT1 and suppress tumor growth, suggesting that targeting NMT1 is a potential treatment for bladder cancer.707 B13 inhibits NMT1 activity by blocking Src myristoylation and reducing its cytoplasmic membrane localization, which inhibits prostate cancer cell proliferation.708

6 METHYLATION

Protein methylation, formed by transferring a methyl group from S-adenosylmethionine (SAM) to a specific methyl acceptor, usually occurs at the side chains of lysine, arginine, histidine, asparagine, and glutamine, among which the methylation of lysine and arginine is the most common.709, 710

Lysine methylation, mediated by protein lysine methyltransferases (PKMTs), has three different methylation forms, including monomethylation (Kme1), dimethylation (Kme2), and trimethylation (Kme3) (Figure 12A), linked to heterochromatin formation, X chromosome inactivation, and transcriptional silencing or activation.711, 712 PKMTs can be divided into two broad categories. One category is methyltransferases with a conserved SET domain, and most of the known PKMTs belong to this broad category.713 Another class is methyltransferases without the SET conserved domain, and most of this class of methyltransferases belongs to the seven-β-strand methyltransferase family. This PKMT family is characterized by a twisted β-fold structure and can affect chromatin structure and gene regulation expression by modifying lysine sites at specific positions of proteins.714

Details are in the caption following the image
The methylation and demethylation process. (A) Lysine residues undergo mono-, di- or trimethylation through the addition of a methyl group to its side chain. (B) There are three types of methylation occurring at the side chain of arginine, including monomethylated arginine (MMA), asymmetric dimethylated arginine (ADMA), symmetric dimethylated arginine (SDMA). (C) The common lysine methyltransferases and demethylases are listed.

Lysine methylation is a reversible modification. According to the different types of catalytic reactions, lysine demethylases can be mainly divided into two families: LSDs and Jumonji C (JmjC) domain-containing demethylases (Figure 12C). The discovery of the first lysine demethylase LSD1 in 2004, also known as KDM1A or AOF2, was a milestone.715 In 2005, the first Jumonji C domain-containing lysine demethylase KDM2A was reported and was mainly responsible for the demethylation of H3K36.716, 717 There are two LSD enzymes (LSD1 and LSD2) encoded in the human genome. LSD enzymes have the amine oxidase catalytic domain commonly found in metabolic enzymes and the SWIRM domain associated with chromatin binding that stabilizes the overall structure of the protein.718, 719 LSD1/2 can only demethylate monomethylated and dimethylated lysines. The demethylation of trimethylated lysines requires demethylases containing the Jumonji C domain.720 To date, approximately 30 Jumonji C domain-containing proteins have been identified in the human genome.721 The family of demethylases containing the Jumonji C domain can hydrolyze methyl groups on monomethylated, demethylated, and trimethylated lysines.720

Protein arginine methyltransferases (PRMTs) mediate arginine methylation722 and can be divided into three types: monomethylarginine (MMA), asymmetric dimethylarginine (ADMA), and symmetric dimethylarginine (SDMA).723 To date, nine PRMTs have been identified, among which type I arginine methylases can catalyze the formation of MMA and ADMA, including PRMT1, PRMT2, PRMT3, PRMT4, PRMT6, and PRMT8; type II arginine methyltransferases include PRMT5 and PRMT9, which can catalyze the formation of MMA and SDMA; and the type III arginine methyltransferase is PRMT7, which only catalyzes the formation of MMA (Figure 12B).23

Histone arginine demethylation is mainly accomplished by two enzymes. One is PAD4, which plays a critical role in regulating the methylation of arginine residues on histones by catalyzing the conversion of methyl-arginine to citrulline, resulting in the release of methylamine. PAD4 targets various sites in histones H3 and H4, including those that are methylated by coactivators CARM1 (such as H3 Arg17) and PRMT1 (such as H4 Arg3).724 Another enzyme is a JmjC domain-containing JMJD6. JMJD6 is a specific histone arginine demethylase dependent on Fe2+ and ketoglutarate725. JMJD6 can catalyze the demethylation of histones H3R2 and H4R3 and convert them into formaldehyde through hydroxylation726. JMJD6 affects the demethylation of monomethylated, symmetric dimethylated and asymmetric dimethylated arginine residues.711, 727

Histone methylation refers to the methylation of the lysine or arginine side chains of histone H3 or H4 mediated by histone methyltransferase (HMT), and some of them also occur on histidine residues.728 The most widely studied histone lysine methylation sites include H3K4, H3K9, H3K27, H3K36, H3K79, and H4K20, and the most widely studied histone arginine methylation sites include H3R2, H3R8, H3R17, H3R26, and H4R3 (Figure 13).728 Methylation of H3K4, H3K36, and H3K79 is generally associated with transcriptional activation of genes, whereas methylation of H3K9, H3K23, H3K27, H3K56, and H4K20 is generally related to transcriptional repression of genes.729, 730 Notably, different methylation states of the same lysine residue, such as mono-, di-, and trimethylation, may play different roles in chromatin state and gene transcriptional regulation.731

Details are in the caption following the image
Protein methylation on histone H3 and H4 and their regulatory enzymes. The same residues can be regulated by multiple methylases and demethylases.

PRMT and PKMT can not only catalyze the methylation of histones, but also catalyze the methylation of nonhistone proteins.732 For example, the DNA damage response proteins MRE11 and 53BP1 can be methylated by PRMT1 to regulate its DNA exonuclease activity and localization at DNA damage sites.733 Some immunomodulatory proteins, such as Vav1 and NIP45, can also be modified by arginine methylation.734 In addition, KMT1C, KMT1D, and KMT1E in the KMT1 family; KMT2F in the KMT2 family; KMT3B, KMT3C, and KMT3E in the KMT3 family; KMT5A in the KMT5 family; and KMT7 have also been reported to catalyze nonhistone lysine methylation.735-738 For example, KMT3C/SMYD2 methylates Rb at Lys860.739

Methylation is involved in the regulation of protein stability, protein activity, PPIs, nuclear-cytoplasmic shuttling, DNA damage repair, transcriptional regulation, ribosome assembly, RNA processing and trafficking, heterogeneous RNA ribosomal protein maturation, protein translation and processing, and intracellular signal transduction (Figure 14).27, 740 Normal methylation modification is of great significance to the growth and development of cells and organisms. For example, EHMT2, also known as G9A, has a SET domain and acts as a transcriptional cooperator or a corepressor.741, 742 EHMT2 can catalyze the mono- or di-methylation of H3K9, which is involved in the regulation of embryonic development and DNA replication.743 The Ehmt2−/− SET domain deletion mutation is embryonic lethal.744 G9a-mediated nonhistone MyoD methylation plays a key regulatory role during muscle development. G9a methylates MyoD at Lys104 to limit its transcriptional activity. Mutation of Lys104 makes MyoD activity refractory to G9a transferase inhibition, resulting in enhanced myogenic activity.745

Details are in the caption following the image
Functions of protein methylation. Representative methylation substrates are presented. Protein methylation is extensively involved in signal transduction, protein stability, protein activity, protein–protein interaction, mRNA splicing, transcriptional regulation, DNA damage repair, nuclear, and cytoplasmic shuttling.

6.1 Methylation in aging

Histone methylation and methylated proteins have recently been shown to play a role in regulating lifespan and tissue aging in organisms. EZH2 is involved in the regulation of aging. Compared with young mice, more EZH2 was recruited to the SDF1 promoter region in aged mice to inhibit SDF1 expression and promote skin tissue regeneration and repair in aged mice.746 KDM2B is a regulator of mouse embryonic fibroblast (MEF) lifespan.747 KDM2B inhibits MEF senescence through demethylation of H3K36me2, resulting in cell immortalization (Table 6).748

TABLE 6. Representative methylation events in health and diseases.
Diseases and biological processes Substrates Effects
Aging H3K27 Pharmacologic inhibition of EZH2 restores SDF1 induction and prevents tissue regeneration.746
H3K36 KDM2B inhibits MEF senescence by demethylating H3K36me2.748
Metabolic disorders Metabolic syndrome SHP PRMT5 catalyzes Arg57 methylation of SHP to augments the SHP repression function and mitigates the risk of metabolic syndrome.790
Vascular dysfunction in T2DM H3K4 In endothelial cells, high glucose induced sustained expression of NF-κB p65 subunit and inflammatory genes by increasing H3K4me1 via SETD7 activation.764
Diabetes H3K4, H3K9 Hyperglycemia induces aberrant changes in H3K4me2 and H3K9me2 in human monocytes.791
Obesity H3K36 Nsd2-mediated H3K36 methylation affects adipose tissue development and function.792
DN H4K20 H4K20 methylation is a direct target of KMT5A. KMT5A and RFX1 modulate ENO1, and are involved in hyperglycemia-mediated EndMT in glomeruli of DN.793
Immune diseases Autoimmunity FOXP3 FOXP3 is dimethylated by PRMT5 (or PRMT1, PRMT6) at R48 and R51, which attenuates the expression of immunosuppressive genes and may lead to autoimmunity.27
Secondary bacterial infection H3K9 Setdb2 is upregulated and induces the repressive H3K9me3 of at Cxcl1 promoter, leading to reduced neutrophil infiltration and attenuated host defense against secondary bacterial infection.794
Innate antiviral immunity TBK1 PRMT1 interacts with TBK1 and catalyzes asymmetric methylation of R54, R134, and R228 on TBK1. Myeloid-specific Prmt1-knockout mice are more susceptible to infection with DNA and RNA viruses than Prmt1fl/fl mice.795
Neurodegenerative diseases HD H2A, H4 Mutant HTT inhibits the activity of PRMT5 and reduces symmetrical dimethylation of H2A and H4 in HD brain.775
H3K27 Reduced PRC2 in adult neurons inhibits the expression of PRC2 target genes. Loss of neuronal function and survival enhances the ongoing dysregulation of PRC2, as well as other H3K27me3-regulated enzymes, likely leading to systemic neurodegeneration in HD.796
AD H3K4 KMT2A can monomethylate and trimethylate H3K4 to promote neuronal gene expression. Mice heterozygous for loss-of-function mutations in the KMT2A gene exhibit learning and memory deficits.771
H3K9 EHMT1/2 inhibitors can reverse histone hyper-methylation and lead to the recovery of glutamate receptor expression and excitatory synaptic function in prefrontal cortex and hippocampus in FAD mice.797
Neurodevelopmental disorders H4K20 Demethylation of H4K20me1 by Phf8 results in transcriptional suppression of RSK1 and homeostasis of mTOR signaling and causes cognitive impairments.774
PD H3K9 α-Synuclein overexpression enhances H3K9me2 level in SNAP25 promoter region by ΕΗΜΤ2 to affect α-synuclein-regulated synaptic vesicle fusion events and leads to synaptic dysfunction in PD.776
H3K4 Increase of H3K4me3 at the SNCA promoter reverts the deregulated expression of α-synuclein in neurons in the context of PD.798
H3K4 Upregulating H3K4me3 by GSK-J4 confers neuroprotection from oxidative stress and alleviate motor deficits in PD.799
ALS H3K9, H3K27, H3K79, H4K20 Reduced mRNA levels from pathogenic C9orf72 are associated with enhanced binding of trimethylated lysine residues in histones H3 and H4.800
CVDs Coronary and ventricular defects H3K36 SETD2 deletion reduces H3K36me3 and affects the expression of cardiac development-related genes Rspo3 and Flrt2, resulting in coronary and ventricular defects.749
Cardiac malformations H3K4 H3K4 methyltransferases, SETD7 and SMYD3, are highly expressed during the development of zebrafish heart. Knockout or overexpression of both Setd7 and Smyd3 can induce cardiac malformations.752
Diabetic vascular complications H3K4, H3K9 H3K4me1, H3K9me2, and H3K9me3 promote endothelial dysfunction in diabetic vasculature by inducing ROS.801
CAD H3K4 ANRIL can promote the combination of WDR5 and HDAC3 complexes and active histone marks such as H3K4me3 to upregulate the ROS level and promote the transformation of HASMC phenotype802
Cardiac hypertrophy H3K9 G9a mediates cardiomyocyte homeostasis by repressing antihypertrophic genes through H3K9 methylation and interaction with EZH2 and MEF2C.803
H4R3 PRMT5 ameliorates cardiomyocyte hypertrophy and induces the methylation of H4R3me2 via the transcriptional activation of Filip1L and subsequent enhancement of β-catenin degradation.804
Cancers Breast cancer EZH2 EZH2 R342 methylation by PRMT1 increases EMT of breast cancer cells.780
H3K9 G9a exerts its oncogenic function in breast cancer by repressing hephaestin and destruction cellular iron homeostasis.805
PT H3K36 Alteration of SETD2 and downstream H3K36me3 may be involved in the development of PT.806
Neuroblastoma H4R3 Downregulation of PRMT1 in neuroblastoma leads to decreased expression of H4R3me2a enrichment at ATF5 promoter and inhibits tumor cell growth.781
ccRCC H4R3 DCPT1061 inhibits ccRCC cell proliferation and induces G1 phase arrest by decreasing the expression of ADMA and PRMT1-mediated H4R3me2a.782
HCC H3K9 Knockdown of G9a reduces H3K9me2 and impairs HCC cell growth and sphere formation.807
Leukemogenesis H3K36 SETD2 mutations affect the expression of leukemigenic genes, hinder the repair of H3K36me3-mediated DNA damage.808
CRC H3K36, H3K9 Overexpression of KDM4C reduces H3K36me3 and H3K9me3 at the promoter of MALAT1, thereby up-regulating MALAT1 expression and enhancing β-catenin signaling pathway.809
PDAC H3K36 SETD2 loss reduces H3K36me3 occupancy at Fbxw7, leading to decreased Fbxw7 expression and increased Myc protein.810
LUAD H3K36 SETD2 inhibits CXCL1 expression by promoting H3K36me3 within the promoter of CXCL1 to reduce the proliferation of LUAD cells and the growth of tumors.811
  • Abbreviations: AD, Alzheimer's disease; ALS, amyotrophic lateral sclerosis; CAD, coronary atherosclerotic heart disease; ccRCC, clear cell renal cell carcinoma; CRC, colorectal cancer; CVDs, cardiovascular diseases; DN, diabetic nephropathy; HCC, hepatocellular carcinoma; HD, Huntington's disease; LUAD, lung adenocarcinoma; PD, Parkinson's disease; PDAC, pancreatic ductal adenocarcinoma; PT, phyllodes tumor of the breast.

6.2 Methylation in heart development and CVDS

Abnormal protein methylation or mutations in methyltransferases often lead to many diseases, such as CVDs, metabolic diseases, immune diseases, neurodegenerative diseases, and tumors. Mutations in multiple SETD family members have been associated with abnormal development of the cardiovascular system, mainly involving SETD2,749 SETD5,750 and SETD7.751 For example, the loss of Setd2 in cardiac progenitor cells leads to obvious coronary vascular defects and ventricular noncompaction, which causes the fetus to die in mid-gestation. The mechanism may be that Setd2 deletion significantly reduces the level of H3K36me3 and affects the expression of the heart development-related genes Rspo3 and Flrt2.749 Both SETD7 and SMYD3 are H3K4 methyltransferases that are highly expressed during heart development in zebrafish. Knockout and overexpression of both Setd7 and Smyd3 induce severe defects in cardiac morphogenesis, suggesting that SETD7 and SMYD3 have a synergistic effect on heart development.752 Moreover, abnormal expression of SETD family members is potentially related to pulmonary hypertension. SETD3 may be protective factors against hypoxic pulmonary hypertension,753 while SETD2, SETD8, and SETD9 may be pathogenic factors in hypoxic pulmonary hypertension or pulmonary fibrosis.754-756 In human and mouse hypertrophic hearts, the expression of JMJD1C is increased, and the methylation level of H3K9 is decreased. Knockdown of Jmjd1c can inhibit Ang II-induced expression of hypertrophy-related genes and cardiomyocyte hypertrophy, while overexpression of JMJD1C can promote cardiomyocyte hypertrophy. Elevated JMJD1C expression induced by pathological conditions reduces the levels of H3K9me1/2/3 at the CaMKK2 promoter, which was associated with CaMKK2 gene silence. CAMKK2 could facilitate the development of metabolic dysfunction and cardiac hypertrophy.757

6.3 Methylation in metabolic disorders

Histone methylation may be responsible for diabetic complications, including diabetic neuropathy, and the phenomenon of “metabolic memory” of long-term changes, and plays a crucial role in pathways of fibrosis, inflammation, and oxidative stress.758 PRMT1 is associated with abnormal glucose tolerance by affecting hepatic glucose metabolism and insulin secretion. PRMT1 knockdown reduced the activation of insulin signaling and inhibited the expression of gluconeogenic genes in hepatocytes.759 High expression of SHP can inhibit the activity of some metabolic enzymes, which increases glucose tolerance and reduces the levels of bile acid and triglycerides. PRMT5 catalyzes Arg57 methylation of SHP to augments the SHP repression function and reduce the occurrence of metabolic syndrome.760 Increased expression of SETD7 is one of the mechanisms of vascular dysfunction in T2DM.761, 762 In endothelial cells, high glucose induced sustained expression of NF-κB p65 subunit and inflammatory genes by increasing H3K4me1 via SETD7 activation.763, 764 Moreover, high-glucose stimulation can reduce the level of H3K9me3 and increase the expression of inflammatory genes in normal human vascular smooth muscle cells (VSMCs).758 On the contrary, SETD8 is a protective factor against endothelial damage in hyperglycemic patients (Table 6).765

6.4 Methylation in immune diseases

PRMTs play a critical role in the establishment and maintenance of lymphoid and myeloid cell lines. PRMT1 is essential for lymphocyte development, proliferation, and differentiation in vivo, as well as for cytokine production by Th cells. CARM1 regulates the differentiation of early thymocyte progenitors by methylating the T cell-specific factor TARPP at R650, while PRMT5-mediated arginine methylation is crucial for the recruitment of TFs during cytokine gene expression in activated T cells. PRMTs have a role in regulating inflammation, with PRMT1 acting as a negative regulator. It interacts with and methylates the NF-κB subunit, RelA/p65, at R30 to suppress its activation by TNF-α. Asymmetric dimethylation of RelA/p65 at R30 inhibits its function as a TF.766 On the other hand, PRMT5 is a positive regulator of inflammation, as it contributes to the activation of IKK and NF-κB, and the induction of several NF-κB target genes.767 Additionally, PRMT6 and CARM1 also positively regulate inflammation.768

PRMT1 is involved in both acute and chronic asthma in epithelial cells and fibroblasts.768 The PRMT5 inhibitor C220 can reduce T cell proliferation and cytokine production, thereby alleviating acute graft-versus-host disease.769 Moreover, selectively inhibiting PRMT5 may prove to be an effective therapeutic strategy for RA and ulcerative colitis.768 PRMT7 is an essential contributor to B cell lymphomagenesis.768 FOXP3, a TF critical for Treg cell identity and immunosuppressive function, is dimethylated by PRMT5 (or PRMT1, PRMT6) at R48 and R51. This methylation reduces the expression of immunosuppressive genes, consequently leading to autoimmunity, tumor shrinkage, and associated CD8+ T cell infiltration.27 Anti-hnRNP reactivity in RA, systemic lupus erythematosus (SLE), and mixed connective tissue diseases (MCTD) is mainly derived from arginine-methylated proteins such as hnRNP A1, A2, and K (Table 6).770

6.5 Methylation in neurogenerative diseases

Protein methylation is closely related to neurological diseases.771 PRMT5 is highly expressed in mammalian neurons. In neurons, β-amyloid peptide (Aβ) deposition reduces PRMT5 expression, increasing E2F-1 expression and activating GSK-3β and NF-κB, leading to caspase-3-dependent neuronal apoptosis.772 KMT2A can monomethylate and trimethylate H3K4 to promote neuronal gene expression. Mice heterozygous for loss-of-function mutations in the KMT2A gene exhibit learning and memory deficits.771 The histone demethylase KDM2B is a candidate gene associated with intellectual disability, autism, epilepsy, and craniofacial abnormalities.773 KDM7B is a key factor in learning and memory. KDM7B-knockout mice show impaired learning and memory, accompanied by abnormal long-term potentiation in the hippocampus.774

The abnormal protein interactions of mutant HTT have been implicated in the pathogenesis of HD. Normal HTT stimulates PRMT5 activity in vitro. However, the presence of mutant HTT reduced the symmetrical dimethylation of arginine (sDMA) of histones H2A and H4 in primary cultured neurons and in HD brain, consistent with impaired gene transcription and RNA splicing in HD.775 In terms of PD, increased levels of α-synuclein can boost the H3K9 methylation activity of EHMT2, which may affect SNARE complex assembly and effectively vesicle fusion events.776 Methyltransferase KMT2A (MLL1) and G9a are associated with AD. KMT2A plays a protective role in AD, while G9a plays a harmful role. Inhibition of the G9a/GLP complex promotes long term potentiation and synaptic tagging/capture in the hippocampus (Table 6).771

6.6 Methylation in cancers

Aberrant methylation is closely associated with the occurrence and development of cancer.734 Compared with normal tissues, PRMT1, PRMT4, and PRMT6 showed higher expression in lung cancer tissues.777, 778 Similarly, PRMT1, PRMT2, PRMT3, PRMT4, and PRMT7 are highly expressed in breast cancer tissues.779 EZH2 R342 can be methylated by PRMT1, which increases epithelial–mesenchymal transition (EMT) in breast cancer cells and predicts poor prognosis in breast cancer patients.780 Downregulation of PRMT1 expression in neuroblastoma results in reduced activity of the prosurvival factor ATF5 and inhibits tumor cell growth.781 In clear cell RCC (ccRCC), a novel potent inhibitor, DCPT1061, was found to induce G1 cell cycle arrest by targeting PRMT1 activity.782 G9a is highly expressed in diverse tumors and indicates poor prognosis. G9a can induce H3K9me2 to affect cancer cell growth and apoptosis.783 Set7-mediated methylation of Gli3 at K436 and K595 in the Sonic Hedgehog pathway promotes NCSLC.784 Aberrant SMYD3 expression may contribute to carcinogenesis.785-787 In PDAC, SMYD3-catalyzed MAP3K2 methylation at lysine 260 is involved in the regulation of oncogenic Ras signaling.788 SMYD3 also exhibits a proto-oncogenic role in prostate cancer due to its methyltransferase enzymatic activity.789 Moreover, key methylation sites, such as H3K27me3, have also been found to be upregulated in many cancers including prostate cancer, breast cancer, and lymphoma, indicating their involvement in tumor progression (Table 6).728

6.7 Methylation-associated targeted therapies

EZH2, a key histone methyltransferase and EMT inducer, is overexpressed in diverse carcinomas. Given its role in tumorigenesis and progression, EZH2 has emerged as a potential antitumor therapeutic target.812, 813 It has been reported that EZH2 can enhance adhesion turnover and accelerate tumorigenesis by increasing cytoskeletal regulatory protein, Talin1 methylation and cleavage. However, this capacity is abolished by targeted disruption of the EZH2 interaction with cytoskeleton remodeling effector, VAV. The interaction of EZH2 with VAV family proteins in the cytoplasm contributes to initial tumor transformation and may maintain cancer stem cells by regulating adhesion dynamics and STAT3 signaling pathways.814 Some anticancer drugs targeting mutant or wild-type EZH2, such as GSK126, were used to inhibit EZH2-mutant lymphoma cells,815 and EPZ-6438 in a phase I/II clinical trial was designed for treating patients with relapsed or refractory B-cell non-Hodgkin lymphoma or advanced solid tumors.812 Many G9a inhibitors, such as diazepinquinazolin-amines and benzimidazoles, have also been developed. The current G9a inhibitors are roughly classified into three types according to their binding modes, including substrate competitive inhibitors, SAM cofactor competitive inhibitors, and inhibitors whose mechanism of inhibition remains elusive. In general, substrate-competitive inhibitors show better selectivity for G9a than SAM inhibitors.783 In addition, the PRMT1 inhibitor GSK3368715 has also entered phase I clinical trials,816 and other inhibitors, including AMI-1, allantodapsone and furamidine, have also started preclinical studies.817

7 UBIQUITINATION

Ubiquitination (also termed ubiquitylation) is the covalent attachment of Ub monomers or Ub chains to lysine residues of proteins. In addition to lysines, the side chains of serine, threonine and cysteine can also undergo ubiquitination.818 Ub, a protein of 76 amino acids that is highly conserved in eukaryotes, contains seven lysine residues K6, K11, K27, K29, K33, K48, and K63 through which the ubiquitination chain extends.819 Protein modification can occur through monoubiquitination (a single Ub moiety) or polyubiquitination (Ub chains) via the isopeptide linkage between two Ub moieties.820 Monoubiquitination refers to the modification of a target protein by a single Ub molecule, while multimonoubiquitination involves the simultaneous modification of multiple lysine residues of a target protein by a single Ub molecule. If a single lysine residue of the target protein is labeled by Ub chains, polyubiquitination occurs.821 The N-terminal methionine residue (Met1) can also be modified by Ub molecules, further increasing the diversity and complexity of ubiquitination.822 Ubiquitination is regulated by three enzymes: Ub-activating enzyme (E1), Ub-conjugating enzyme (E2), and Ub-ligase enzyme (E3).823 First, E1 (UBA1, UBA6) forms a high-energy thioester bond with the Ub molecule to activate it. Then, the activated Ub is covalently attached to the cysteine residue of E2 (UBE2B, UBE2D2) through a thioesterification reaction. Finally, activated Ub is either directly attached to substrates through E2 or transferred to substrates in the presence of E3 (RNF6, TRAF6, SKP2, and Nedd4).824 Ubiquitination is a reversible process. The removal of ubiquitination is mainly carried out by deubiquitinases (DUBs). Both Ub-modifying enzymes and deubiquitinating enzymes work together to regulate the transmission of intracellular Ub signaling to maintain normal cellular activities (Figure 15).825

Details are in the caption following the image
The protein ubiquitination pathway. The ubiquitin (Ub) moiety is activated by E1 through the cysteine (Cys) residue of E1. Ub at E1 is transferred to the Cys residue of E2. Ub conjugated with E2 is transferred to the lysine residue of a substrate protein by E3. Ubiquitinated proteins with ubiquitination are subjected to 26S proteasome-dependent degradation or execute other functions and activities. Deubiquitination is the opposite mechanism of ubiquitination mediated by DUBs. The process includes reversing ubiquitin conjugation and recycling ubiquitin molecules through the UPS. Based on the enzymatic cleavage mechanism, the DUB family is divided into two subfamilies. The cysteine protease family consists of USP, UCH, OTH, MJD, MINDY, MCPIP, and ZUFSP. Metalloprotease family includes JAMM.

Thus far, only two human E1 Ub-activating enzymes, UBA1 and UBA6, have been identified.826 However, in the human proteome, eight E1 enzymes are known to activate Ub-like proteins (UBLs), including UBA1 (UAE), NAE, SAE, UBA6, UBA7, UBA4, UBA5, and ATG 7.827 In contrast, more than 40 human E2 Ub-conjugating enzymes have been reported.828 There are more than 600 E3 Ub ligases, which can be roughly divided into three categories: RING (truly interesting new gene) E3 ligases, HECT (homologous to E6AP C-terminus) E3 ligases, and RBR (RING-between-RING) E3 ligases.829 Ring E3 ligases are the most predominant Ub ligases in the human body.830 There are few identified DUBs, and more than 100 DUBs have been identified to date.818 Based on sequence homology, deubiquitinating enzymes can be divided into eight classes (Figure 15), including Ub C-terminal hydrolases (UCHs), Ub-specific proteases (USPs), Machado-Joseph domain-containing proteases (MJDs), ovarian tumor proteases (OTUs), motif interacting with Ub-containing novel DUB family (MINDYs), JAMMs (JAB1/MPN/MOV34), monocyte chemotactic protein-induced protein (MCPIP) families, and ZUFSP DUB family.831 According to the mechanisms of action, deubiquitinating enzymes can also be divided into cysteine proteases (including UCHs, USPs, MJDs, OTUs, and MINDYs and MCPIPs) and metalloprotease JAMMs.832

A variety of Ub combinations form a variety of structures and are involved in different physiological functions.833-835 Among the seven types of polyubiquitin chains, the most common K48 and K63 polyubiquitin chains mainly regulate proteasomal degradation of substrates and intracellular signaling, respectively.836, 837 K48-linked ubiquitination is the most prevalent signal for proteasomal degradation, although other ubiquitination types such as K11- or K29-linked ubiquitination and multiple monoubiquitination are also signals for proteasomal degradation.838, 839 In contrast, K63-linked ubiquitination regulates “proteasome-independent” processes such as inflammatory signal transduction, neurodegeneration, DNA repair, endocytosis, and selective autophagy.840, 841 K27 is critical for cellular immunity842 and the DNA damage response.843 The M1 chain (linear chain) is a positive regulator of NF-κB signaling844 and a negative regulator of type I IFN signaling.845 The K6 chain is involved in the regulation of the UV-induced DNA damage response and mitochondrial homeostasis.846, 847 The K11 chain is another proteasomal degradation signal involved in cell cycle regulation.848 K11-linkages are also implicated in regulating membrane trafficking and the innate immune response.845 The K29 chain is involved in proteotoxic stress responses, cell cycle and AMPK regulation.845, 849 The K33 chain acts on protein exchange at the Golgi membrane and is related to the regulation of the innate immune response.850 However, polyubiquitin chains can be heterogeneous, consisting of more than one type of connection, and are divided into branched/forked chains and mixed (hybrid) chains.837, 851 A large proportion (10–20%) of branched chains are present in the aggregated form of ubiquitination.852 Branched Ub chains have two degradative linkages, including K11/K48 or K29/K48. The K48/K63 branched Ub chain can enhance NF-κB signaling.837, 853 Similarly, mixed chains consisting of two NF-κB-associated junctions (M1 and K63) are formed during NF-κB activation.851 The discovery that the K63 chain is modified by the M1 chain in a hybrid or branched structure addresses the question of whether the K63 chain acts on inflammatory signaling or NF-κB activation.851 In contrast to polyubiquitination, monoubiquitination plays critical roles in DNA repair, receptor endocytosis, vesicle sorting, and gene silencing,854 and multimonoubiquitination is involved in the regulation of receptor endocytosis, protein interactions and localization (Figure 16).855

Details are in the caption following the image
Ubiquitin linkage types and their roles. Ubiquitination can occur as single or multiple monoubiquitin or as homotypic/heterotypic/branched chains linked through K6, K11, K27, K29, K33, K48, or K63, as well as M1. The functional consequences of ubiquitin signals are determined by cellular and substrate-context, ubiquitin chain position, linkage type and conformation, ranging from proteasomal degradation to nonproteolytic functions.

7.1 Ubiquitination in development

An increasing number of studies have shown that members of various E2 and E3 families are involved in sperm capacitation, oocyte maturation, and embryonic development in mammals.856 For example, the absence of UCH-L1, an neuronal deubiquitinating enzyme, impacts the maintenance of spermatogonial stem cells homeostasis and metabolism and impacts the differentiation competence.857 Testicular macrophage USP2 promotes sperm motility, activation, and capacitation.858 Polycomb repressive complex 1 (PRC1) is known to play a crucial role in stem cell and tissue development. It has been demonstrated that PRC1, in conjunction with H2AK119ub, influences early embryonic development. High levels of expression of Polycomb repressive DUB complex (PR–DUB) in zygotes can rapidly reduce H2AK119ub levels, resulting in developmental arrest at the 4-cell stage.859 Furthermore, Cbls play a role in promoting the ubiquitination and degradation of FLT3, which results in the inhibition of FLT3 signaling and limits the development of CD8α+/CD103+ DC1 (cDC1). When Cbls are absent, activated FLT3 cannot be efficiently removed, leading to constant FLT3 signaling that favors cDC1 development and expansion.860

7.2 Ubiquitination in aging

During aging, there is a buildup of damaged and aggregated proteins, which can lead to a decline in cellular function. Ub-dependent proteolytic pathways are critical for the efficient turnover of defective proteins.861 However, age-related impairment of these pathways can lead to a greater accumulation of damaged proteins, thereby exacerbating the aging process.861, 862 For example, in aged C. elegans, 192 proteins with low levels of ubiquitination accumulate, further contributing to the decline in cellular function.863 Notably, ageing causes a global loss of ubiquitination that is triggered by increased DUB activity. Parkin-mediated mitophagy is essential to ensure mitochondrial quality control in myocardium. The main mechanism of action is that the interaction between Parkin and TBKI promotes the K63 polyubiquitination of TBK1, which in turn promotes TBK1 phosphorylation to enhance mitophagy and alleviate cardiac aging.864 However, excessive or inappropriate protein ubiquitination may also shorten longevity. For example, the Ub ligase RLE-1 selectively polyubiquitinates daf-16, a key component in the insulin/IGF signaling pathway, leading to its degradation by the proteasome. As a result, inhibition of RLE-1 extends lifespan in C. elegans.865 In human fibroblasts, the degradation of BMAL1 is mediated by the E3 Ub ligase STUB1. Reduced BMAL1 can attenuate cellular senescence induced by hydrogen peroxide.866 Nrf2 is an important regulator in healthy aging, and its activity is also affected by ubiquitination. For example, p62 prevents Nrf2 from being ubiquitinated by combining with Keap1. Some studies have found that the expression of p62 decreases with age. Hrd1 is a negative regulator of Nrf2. It interacts with Nrf2 through its Neh4–5 domain to enhance its ubiquitination.867, 868 Stem cell dysfunction and reduced regenerative capacity are hallmarks of aging. NANOG, one of homeobox proteins, plays a crucial role in regulating self-renewal and pluripotency for embryonic stem cells (ESCs).The deubiquitinating enzyme USP21 increases NANOG levels to maintain pluripotency in ESCs by deubiquitinating NANOG to reduce NANOG proteasomal degradation (Table 7).869

TABLE 7. Representative ubiquitination events in health and diseases.
Diseases and biological processes Protein substrates Effects
Aging TBK1 K63 polyubiquitination of TBK1 promotes TBK1 phosphorylation to enhance mitophagy and attenuate cardiac aging.864
daf-16 The Ub ligase RLE-1 selectively poly-ubiquitinates daf-16 and promotes proteasomal degradation. Inhibition of RLE-1 prolongs lifespan in C. elegans.865
BMAL1 STUB1 ubiquitinates and degrades the substrate BMAL1, attenuating hydrogen peroxide-induced cellular senescence.938
Immune regulation Viral infection IRF7 NEURL3 promotes innate antiviral responses by catalyzing K63-linked poly-ubiquitination of IRF7 at K375.872
Viral infection RIG-I The CoV nucleocapsid (N) protein of SADS-CoV interacts with RIG-I and promotes its K27-, K48-, and K63-linked ubiquitination to induce RIG-I degradation, which further inhibits the host IFN-β response.879
Viral infection RIG-I lncRNAs can promote influenza A virus (IAV) replication and immune evasion by restricting RIG-I K63 ubiquitination mediated by TRIM25.880
Viral infection PPE Mycobacterial PPE protein ubiquitination mediated by MKRN1 suppresses the innate immune response.881
Viral infection MAVS Viral infection enhances the interaction between USP18 and MAVS and promotes the K63-linked ubiquitination of MAVS to upregulate the production of IFN-I.882 RNF115 interacts with MAVS to promote K48 ubiquitination of MAVS, and loss of RNF115 enhances antiviral signaling triggered by RNA viruses.883
Autoimmune disease c-Rel Peli1 negatively regulates T cell activation and inhibit the development of autoimmunity through K48 ubiquitination dependent degradation of c-Rel.875
Metabolic disorders Insulin resistance MG53 MG53 acts as an E3 ligase targeting insulin receptor and IRS1 for Ub-dependent degradation. Overexpression of MG53 is sufficient to induce muscle insulin resistance and metabolic syndrome.887
Diabetes EZH2 Suppressive role of E3 Ub ligase FBW7 in type I diabetes in nonobese diabetic mice through mediation of ubiquitination of EZH2.886
Cancers Multiple cancers p53 The tumor suppressor p53 is degraded by ubiquitination mediated by MDM2, resulting in immortal cancer cell proliferation.895
Liver cancer Smad4 USP10 stabilizes Smad4 by ubiquitinating it, activates TGF-β signaling, and promotes liver cancer metastasis.900
PC TRAF6 USP4 is highly expressed in PC. It stabilizes TRAF6 and activates the NF-κB signaling pathway to enhance the proliferation, migration and invasion of PC cells.902
Ovarian cancer BCL6 USP14 expression is increased in cisplatin-resistant ovarian cancer cells. It inhibits ovarian cancer cell apoptosis by stabilizing BCL6, which increases ovarian cancer cisplatin resistance.904
PDAC DRYK1A USP22 is highly expressed in PDAC. It enhances the growth and colony formation ability of cancer cells by regulating DRYK1A.905
Gastric cancer SMAD2 USP32 is highly expressed in gastric cancer and is closely related to the stage and prognosis of gastric cancer patients. Downregulation of USP32 significantly reduces SMAD2 expression, thereby inhibiting the proliferation, migration, and resistance to cisplatin of gastric cancer cells.906
Liver cancer TBLR1 USP1 promotes the survival of liver circulating tumor cells in the bloodstream by deubiquitinating and stabilizing TBLR1.907
Liver cancer RAB1A USP2a is highly expressed in HCC tissues and is positively correlated with poor prognosis. USP2a can deubiquitinate and stabilize RAB1A to promote HCC progression.908
Multiple cancers K-Ras Ubiquitination of K-Ras can enhance its interaction with PI3K, leading to abnormal activation of the PI3K/AKT signaling pathway. This is one of the mechanisms by which the G12V mutation of K-Ras causes malignant cell proliferation.909
CVDs Cardiac fibrosis RIP1 Peli1 silencing abrogates mechanical stretch-induced polyubiquitination of TRAF6 and RIP1 and consequently decreases the DNA binding activity of NF-κB in neonatal rat cardiac fibroblasts.913
Vascular lesions HIF The loss-of-function mutation of VHL can inhibit the normal degradation of its downstream substrate HIF. Accumulated HIF activates downstream target genes such as VEGF, leading to the formation of vascular lesions.939
Neurodegenerative diseases AD C/EBPβ Peli1 can directly ubiquitinate and degrade C/EBPβ, inhibits the phagocytosis of microglial cells, thus slowing down Aβ clearance in the brain of AD mice.929
AD C/EBPβ Loss of COP1 results in rapid accumulation of the transcription factor C/EBPβ, which drives the expression of proinflammatory and neurodegeneration-related genes and accelerates the neurodegeneration of AD.928
AD TAp73 Aβ42 can lead to hyperphosphorylation of Itch by abnormally activating the JNK signaling pathway. Hyperphosphorylated Itch ubiquitinates and degrades TAp73, leading to abnormal expression of important neuronal cyclins and causing neuronal apoptosis, which accelerates AD progression.926
AD ESR2 The HECT family protein E6AP can activate the transcription of ESR2, which reduces Aβ deposition in the hippocampus and improves learning and memory in AD rats.927
PD Synphilin 1 The UbcH7-parkin complex promotes the ubiquitination and degradation of several proteins via the 26S proteasome. Cellular accumulation of the UbcH7-parkin targets, α-synuclein and synphilin-1, has been associated with PD.940
ALS TDP-43 The occurrence of ALS is related to neuronal cell death caused by the abnormal aggregation of highly phosphorylated and ubiquitinated pathological TDP-43.941 Insufficient degradation of abnormally aggregated TDP-43 protein leads to cell death and inflammation, which is one of the important mechanisms in the pathogenesis of ALS.935
  • Abbreviations: AD, Alzheimer's disease; ALS, amyotrophic lateral sclerosis; PC, pancreatic cancer; PDAC, pancreatic ductal adenocarcinoma; PD, Parkinson's disease.

7.3 Ubiquitination in immune regulation

Ubiquitination modulates the function of immune cells by regulating biological processes such as protein degradation and signal pathway transduction.870 K63- or K48-linked ubiquitination is common in these processes.871 For example, the E3 Ub ligase NEURL3 promotes host antiviral immune response by catalyzing K63-linked polyubiquitination of IRF7 at K375.872 The removal of K63 ubiquitination of TBK1 by USP15 negatively regulates TBK1 activity to suppress macrophage antiviral innate immune responses.873 Pellino 1 (Peli1) can mediate TLR-stimulated K63 ubiquitination of c-IAP2 in microglia to trigger the Ub ligase activity of c-IAP2, which catalyzes the K48 ubiquitination and degradation of TRAF3 and consequently induces proinflammatory cytokine production and the recruitment of autoimmune T cells in peripheral lymphoid organs to the CNS.874 Peli1 also negatively regulates T-cell activation and inhibits the development of autoimmunity through K48 ubiquitination-dependent degradation of c-Rel.875 The unanchored K48-polyubiquitin chain synthesized by the E3 ligase TRIM6 can promote the binding of DHX16 to RIG-I and mediate the production of IFN-I and the expression of IFN-stimulated genes (ISGs).876

During thymocyte development, multiple E3 ligases have been shown to play a role in T cell development. Protein ubiquitination regulates T cell development and differentiation. For example, Ub ligases, including Itch, LNX, DTX, Mib1, Mib2, Neur1, and Neur2, catalyze Notch ubiquitination, which is critical to the early stage of T cell development.877 Ubiquitination also affects the proliferation and development of B cells by regulating NF-κB signaling and participating in BCR and BAFFR signal transduction. E3 Ub ligase Hrd1 mediates the downregulation of pre-BCR through ubiquitination and promotes the maturation of pre-B cells.878

Viruses can escape from immunity by degrading target proteins. For example, the CoV nucleocapsid (N) protein of SADS-CoV interacts with RIG-I and promotes its K27-, K48-, and K63-linked ubiquitination to induce RIG-I degradation, which further inhibits the host IFN-β response.879 lncRNAs can promote influenza A virus (IAV) replication and immune evasion by restricting RIG-I K63 ubiquitination mediated by TRIM25.880 Mycobacterial PPE proteins are ubiquitinated by MKRN1, which suppresses the innate immune response.881 Additionally, ubiquitination plays a role in regulating early innate immune responses triggered by human respiratory syncytial virus. RIG-I, MAVS, TRAF3/6, and NEMO are the main proteins involved in these processes.871 Activation of MAVS is indispensable for antiviral immunity. Viral infection enhances the interaction between USP18 and MAVS and promotes the K63-linked ubiquitination and subsequent aggregation of MAVS to upregulate the production of IFN-I.882 RNF115 interacts with MAVS to promote K48 ubiquitination of MAVS, and loss of RNF115 enhances antiviral signaling triggered by RNA viruses (Table 7).883

7.4 Ubiquitination in metabolic disorders

Dysfunction of the Ub-proteasomal system can lead to obesity-related metabolic disorders such as diabetes and fatty liver. Chronic insulin stimulation inhibits hepatocyte ubiquitination by activating USP14, which also increases the nuclear translocation of the lipogenic TF SREBP-1c to inhibit mature SREBP-1c.884 USP7 is increased in diabetic foot ulcers and human umbilical vein endothelial cells (HUVECs). USP7 inhibition can suppress AGEs-induced cell cycle arrest and cellular senescence in HUVECs by promoting p53 ubiquitination.885 E3 Ub ligase FBW7 prevents type I diabetes in nonobese diabetic mice by mediating EZH2 ubiquitination.886 MG53 acts as an E3 ligase targeting insulin receptor and IRS1 for Ub-dependent degradation. Overexpression of MG53 is sufficient to induce muscle insulin resistance and metabolic syndrome.887 Diabetic cataract is also a common complication of diabetes. The E3 Ub ligase MDM2 may promote high glucose-induced EMT and oxidative stress damage by downregulating LKB1. The EMT of lens epithelial cells is an important step in the development of diabetic cataracts.888 Pregnant women with obesity or gestational diabetes have reduced blood levels of adiponectin, which is thought to be associated with an increased risk of obesity or obesity-related insulin resistance and fetal overgrowth. Adiponectin ubiquitination is increased in the visceral fat of obese pregnant women compared to their lean counterparts, and it is a key mechanism through which obesity curtails adiponectin secretion during pregnancy (Table 7).889

7.5 Ubiquitination in cancers

Dysregulation of ubiquitination may cause a range of adverse consequences, such as abnormal activation or inactivation of signaling pathways, abnormal protein complex formation, accumulation of misfolded proteins and mislocalization of proteins,890 and even cancers.834 Due to the specificity of E3 in recognizing protein substrates, E3 has received increasing attention.891 Some E3 ligases are carcinogenic factors, some are tumor suppressors, and some have both functions dependent on the context.892 E3 usually participates in tumorigenesis and development by regulating the stability of oncoproteins and tumor suppressors. Members of the Cbl family of E3 ligases are involved in tumorigenesis and development by mediating lysosomal sorting and degradation of activated RTKs.893 Mutations and aberrant expression of C-CBL are most common in myelodysplastic syndromes.894 The tumor suppressor p53 is degraded by ubiquitination mediated by the E3 ligase MDM2, resulting in immortal cancer cell proliferation.895

DUBs also affect cancer signaling pathways by deubiquitination. The changes in DUBs may cause continuous activation or abnormal blockade of downstream signal transduction molecules such as PI3K/AKT896 and NF-κB897 to affect the progression of malignant tumors. USP7 can deubiquitinate and stabilize MDM2 (Murine double minute 2) oncoproteins, which is the major negative regulator of the p53 tumor suppressor, thereby inducing the initiation, progression, and metastasis of human cancers.898 On the other hand, USP10 can counteract the effects of MDM2-induced p53 nuclear export and degradation by deubiquitinating p53.899 Additionally, USP10 can also stabilize Smad4 by ubiquitinating it, which contributes to liver cancer metastasis.900 USP4 can interact directly with and deubiquitinates ARF-BP1, leading to the stabilization of ARF-BP1 and subsequent reduction of p53 levels.901 In addition, USP4 is highly expressed in PC tumors. It stabilizes TRAF6 and activates the NF-κB signaling pathway to enhance the proliferation, migration and invasion of PC cells.902 In cervical cancer cell lines SiHa and Caski, silencing USP18 resulted in the inhibition of cell proliferation, induction of apoptosis, and promotion of cleaved caspase-3 expression.903 USP14 has increased expression in cisplatin-resistant ovarian cancer cells. It inhibits ovarian cancer cell apoptosis by stabilizing the level of BCL6, which increases ovarian cancer cisplatin resistance.904 USP22 is highly expressed in human PDAC tissues. It enhances the growth and colony formation ability of cancer cells by regulating the expression of DRYK1A.905 USP32 is highly expressed in gastric cancer and is closely related to the high T-staging and poor prognosis of gastric cancer patients. Downregulation of USP32 can significantly inhibit the expression of SMAD2, thereby inhibiting the proliferation, migration, and chemoresistance to cisplatin of gastric cancer cells.906 TBLR1 plays an important role in regulating the Wnt signaling pathway. USP1 promotes the survival of liver circulating tumor cells in the bloodstream by deubiquitinating and stabilizing TBLR1.907 USP2a is highly expressed in HCC tissues and is positively correlated with poor prognosis. USP2a can deubiquitinate and stabilize RAB1A to promote HCC progression.908

In addition to impacting protein activity and degradation, ubiquitination is also implicated in cancer signaling regulation by modulating PPIs. Monoubiquitination of K-Ras at K147 can lead to enhanced GTP loading and increases its affinity for specific downstream effectors PI3K and Raf, which results in anomalous activation of the PI3K–AKT signaling pathway. This is one of the mechanisms by which the G12V-K-Ras mutant spurs malignant cell proliferation (Table 7).909

7.6 Ubiquitination in CVDS

Ubiquitination is also critical in the development of CVDs. Elevated levels of myocardial ubiquitinated proteins have been observed in most primary causes of heart failure, such as cardiac muscle loss910 and cardiomyopathy.911 Changes in Ub protein ligases targeting myofibrillar and other cardiac proteins, such as atrogin-1 (MAFbx) and MURF-1, are associated with pathological cardiac remodeling.912 Moreover, ubiquitination is linked to atherosclerosis, with Ub–proteasome system (UPS) regulating eNOS activity and oxidative stress in the initiation and development of atherosclerosis. It also activates the NF-κB pathway, affecting adhesion molecule expression, cytokine release, and proliferation. The UPS also influences foam cell formation and maintenance, which can impact atherosclerosis progression.912 Ubiquitination also plays an important role in the development of myocardial fibrosis. By inhibiting the expression of the E3 Ub ligase Pellino1, it is possible to prevent the production of α-SMA, collagen I and collagen III, and thereby attenuate myocardial interstitial fibrosis. Pellino1 has been shown to facilitate the binding of NF-κB and AP-1 to the TGF-β promoter, which regulates the fibrogenic capability of cardiac fibroblast cells and contributes to the development of fibrosis in the heart (Table 7).913

7.7 Ubiquitination in neurodegenerative diseases

Abnormal UPS function is closely related to the formation of protein aggregates in neurodegenerative diseases.914 The accumulation of insoluble Aβ in extracellular plaques and hyperphosphorylated tau protein (P-tau) in NFTs within neuronal cytoplasm is a significant pathological factor observed in the brains of AD patients.915 Reduced UPS efficiency and the inhibited autophagy-lysosomal pathway are significantly positively correlated with the abnormal accumulation of Tau at synaptic terminals.916, 917 Molecular misreading allows the formation of mutant proteins in the absence of gene mutations. Ubb+1, a frameshift mutation product of Ub protein in the brains of AD patients, can inhibit the function of the 26S proteasome and lead to an accumulation of a large number of pathogenic proteins, such as Aβ.918, 919 In addition, many ubiquitination-related enzymes are abnormally expressed in AD, such as increased E2 ligase E2-25K/Hip-2920 and E3 ligase CHIP921 and RNF182,922 and decreased E3 ligases Parkin,923 HRD1,924 and the DUB UCHL1.925

Aβ42 can lead to hyperphosphorylation of the E3 Ub ligase Itch by abnormally activating the JNK signaling pathway. Hyperphosphorylated Itch ubiquitinates and degrades TAp73, leading to abnormal expression of important neuronal cyclins and causing neuronal apoptosis, which accelerates AD progression.926 The HECT family protein E6AP can activate the transcription of the ESR2 gene encoding ER-β, which reduces Aβ deposition in the hippocampus and improves learning and memory in AD rats.927 Loss of the E3 Ub ligase COP1 results in rapid accumulation of the TF C/EBPβ, which drives the expression of proinflammatory and neurodegeneration-related genes and accelerates the neurodegeneration of AD.928 However, C/EBPβ is also the main TF responsible for the transcription of the scavenger receptor CD36. The E3 Ub ligase Peli1 can directly ubiquitinate and degrade C/EBPβ, which further reduces CD36 and inhibits the phagocytosis of microglial cells, slowing the clearance of Aβ in the brains of AD mice.929

While the precise mechanisms of PD are not yet entirely clear, it is widely acknowledged that α-synuclein plays key pathophysiological roles as the main constituent of the cytoplasmic inclusions known as Lewy bodies. SIAH is an E3 Ub ligase that plays a key role in stress-induced cell death and α-synuclein degradation. This suggests that ubiquitination could protect against PD.930 However, some studies have found that SIAH can actually promote α-synuclein aggregation and enhance its toxicity,931, 932 leading to more inclusions in dopaminergic neurons. Inhibiting SIAH could prevent Lewy bodies formation and be a potential therapy for PD.933

The accumulation of RNA-binding protein TDP-43 in neuronal cytoplasmic and intranuclear aggregates is a defining feature of neurodegenerative disorders, including ALS and frontotemporal lobar degeneration. TDP-43 is typically modified with polyubiquitin chains that are mainly K48- or K63-linked.934 Insufficient degradation of abnormally aggregated TDP-43 protein leads to cell death and inflammation, which is an critical mechanism in the pathogenesis of ALS.935 Several studies have examined the potential therapeutic value of targeting TDP-43 ubiquitination by preventing the removal of Ub chains, with conflicting results. Inhibition of the DUB USP14 promotes TDP-43 clearance by maintaining Ub chains.936 However, in Drosophila, knockdown of the DUB UBPY increased TDP-43 toxicity, despite retaining Ub chains (Table 7).937

7.8 Ubiquitination-associated targeted therapies

An increasing number of studies have shown that ubiquitination-related enzymes are a class of important drug targets. For example, the E3 ligase inhibitors thalidomide, lenalidomide, and permadomide have been used to treat multiple myeloma.942 These inhibitors bind to CRBN, activate the activity of CRL4CRBN E3 Ub ligase, and induce the degradation of two important TFs, Ikaros/Aiolos, to kill cancer cells.943 With the continuous advancement of drug screening technology and chemical synthesis technology, an increasing number of novel small molecules targeting ubiquitination have been discovered, such as the E2 ligase UBE2 inhibitor NSC697923,944 E3 ligase MDM2 inhibitor BI-0252,945 CRL inhibitors 33-11 and KH-4-43,946 XIAP and cIAP1 inhibitor AT-IAP,947 VHL inhibitor VH298,948 DUB USP2 inhibitor 6TG,949 USP7 inhibitor FT671,950 USP9X inhibitor G9,951 USP14 inhibitor IU1–47,952 and PSMD14 inhibitor THL.953 For proteins that pose challenges for targeting, regulating their upstream ubiquitination-related enzymes has emerged as a novel drug development strategy.954

In 2001, the proteolysis-targeting chimera (PROTAC) was first proposed as a chemical biology tool for targeted therapies. After 20 years of development, PROTAC technology has matured.955 The structures of PROTACs are similar to a dumbbell, connecting the “ligand of target protein” and “recruitment ligand of E3 ubiquitin ligase” through a linker.691, 692, 956 The tagged target proteins are recognized and degraded by the intracellular 26S proteasome.955 The most significant advantage of PROTAC technology is its ability to convert potentially undruggable targets into druggable ones.957 Moreover, PROTACs overcome drug resistance.958-960 Some PROTACs developed against cancer-associated proteins outperform traditional small-molecule inhibitors for cancer therapy. For example, oral PROTACs targeting ER and AR (ARV-110 and ARV-471) are used for the treatment of breast and prostate cancer, respectively.961-963

8 SUMOYLATION

SUMOylation is a PTM that covalently attaches small Ub-like modifiers (SUMOs) to specific lysine residues in proteins.964 SUMO is a family of highly conserved small-molecule proteins widely found in eukaryotes. There are currently five SUMO proteins (SUMO1–5) found in eukaryotes.964 The size of SUMOs is approximately 11 kDa. SUMO1, SUMO2, and SUMO3 contain 101, 103, and 95 amino acid residues, respectively.965 SUMO2 and SUMO3 cannot be distinguished by antibodies because their sequence similarity is as high as 97%, so they are usually collectively referred to as SUMO2/3. However, the sequence similarity between SUMO2/3 and SUMO1 is only 46%, and they usually show different biological functions in the body.966, 967 For example, SUMO1 mainly modifies some proteins in the physiological state, and SUMO2/3 mainly modifies stress proteins.968 SUMO1, SUMO2, and SUMO3 are widely expressed in all cells and organs.969 In contrast, SUMO4 is specifically expressed only in certain organs, such as the kidney, lymph nodes, and spleen,968 and SUMO5 is mainly expressed in the lung and spleen.970

SUMOylation is similar to the Ub modification process, requiring SUMO-activation enzyme (E1), SUMO-conjugating enzyme (E2), and SUMO-ligating enzyme (E3).966 The process of protein SUMOylation includes four steps. (1) Maturation of SUMO proteins. In this process, several amino acids of the C-terminal sequence of SUMO precursor proteins are excised by SENP to expose the diglycine GG motif, which matures the SUMO proteins. (2) Activation. Under the action of ATP, mature SUMO is linked to the cysteine of the E1 activating enzyme (SAE1/SAE2 heterodimer in humans, called AOS1/Uba2 in yeast) through a thioester bond to activate the SUMO molecule. (3) Conjugation. The SUMO–E1 complex transfers SUMO to the E2 ligase Ubc9. (4) Ligation. SUMO is transferred from E2 to the lysine residues of protein substrates under the action of ligase E3 (Figure 17A).968 To date, the reported E3 ligases include protein PIAS family members (PIAS1, PIAS3, PIASxα, PIASxβ, PIASy), hPC2 (also known as PC2 and CBX4) and RanBP2.967, 971 SUMOylation is a highly dynamic and reversible process. The process of deSUMOylation is mainly regulated by SENPs.972 There are seven SENPs, namely, SENP1, SENP2, SENP3, SENP5, SENP6, SENP7, and SENP8.973, 974 The seven SENPs are divided into three families. The first family includes SENP1 and SENP2, which have broad substrate specificity and can bind to SUMO1/2/3. The second family includes SENP3 and SENP5, located in the nucleolus, and is mainly responsible for the removal of SUMO2/3. The third family includes SENP6 and SENP7, which remove SUMO2/3 from poly-SUMO chains and are mainly localized in the nucleoplasm (Figure 17A).966, 975

Details are in the caption following the image
SUMOylation process and its functions. (A) The catalytic cycle of SUMOylation. SENP have endopeptidase activity to cleave SUMO precursors by exposing the carboxy-terminal diglycine motifs essential for the conjugation to lysine residues in target proteins. SUMOylation is catalyzed by SAE1-SAE2 (E1) and UBC9 (E2). The E3 ligases can facilitate the last step of SUMO conjugation. SENPs isopeptidase activity allows for the release of SUMO from target proteins. (B) Functions of SUMOylation. SUMOylation is extensively involved in signal transduction, and the regulation of protein activity, protein stability, protein–protein interaction, cell apoptosis, translocation, and DNA damage repair. Representative SUMOylation proteins and substrates are shown.

SUMOylation is associated with the regulation of protein expression, localization, stability, and activity and is involved in various cellular processes, such as PPIs, intracellular localization, DNA repair, nucleocytoplasmic transport, TF activation, apoptosis, cell cycle, and gene transcription (Figure 17B).964, 976 The way SUMOylation works can be divided into two categories. First, SUMOylation directly affects protein functions by covalently modifying the protein. Second, the target protein indirectly regulates the biological function of the protein through noncovalent binding of the SUMO interact motif (SIM).977 Although SUMO and Ub are very similar in structure and occurrence process, their effects are often opposite. Ubiquitinated proteins are usually sent to the proteasome for degradation, while SUMOylated proteins are often more stable and less susceptible to degradation.978 For example, IκBα K21 can be modified by both SUMO and Ub, and SUMOylation can antagonize ubiquitination.979

8.1 Sumoylation in immune regulation

SUMOylation plays a significant role in the host immune response, as numerous SUMOylated proteins are involved in the development and activation of various immune cells.980 For example, SUMOylation of PKC-θ is required for T cell activation and formation of a mature immunological synapse.981 Viruses can manipulate the process of SUMOylation through the SUMO pathway, while SUMOylation can also eliminate viral infections by regulating host antiviral immune components.982 Additionally, SUMOylation has also been linked to autoimmune diseases, especially RA. SUMOylation bidirectionally regulates immune pathways, which can prevent hyperresponsiveness of the immune system and inhibit the development of inflammatory and autoimmune diseases.983 In RA patients, the expression of SUMO1 and SUMO2 is elevated in fibroblast-like synoviocytes (FLSs), especially SUMO1, and FLSs are important for promoting RA pathogenesis. SUMO-1 suppression could be protective against joint destruction in RA by inhibiting aggressive behavior of RA FLSs.984 Moreover, SUMOylation is crucial in maintaining the stability of the intestinal epithelial barrier by changing the intestinal flora, regulating immune cells, and regulating cytokines, such as IL-6, TNF-α, and IFN-γ. SUMOylation of intestinal epithelial cells (IECs) can reduce the severity of inflammatory bowel disease (IBD) by inhibiting the activity of master regulators, including the serine-threonine kinase AKT1 (Table 8).985

TABLE 8. Representative SUMOylation events in health and diseases.
Diseases Protein substrates Effects
Immune diseases Inflammation NLRP3 TRIM28 binds to NLRP3, catalyzes SUMO1, SUMO2, and SUMO3 modification of NLRP3, promotes NLRP3 expression, and enhances NLRP3 inflammasome activation.1057
RA IκB-α SUMOylation of IκB-α prevents NF-κB from ubiquitination, which further inhibit nuclear migration and production of inflammatory mediators.983
IBD IECs SUMOylation of IECs reduces the severity of IBD by inhibiting the activity of master regulators.985
Neurodegenerative diseases AD Tau SUMOylation promote Tau phosphorylation and inhibit Tau degradation to promote NFT formation.1058
APP SUMO1 modification of APP promotes the generation of Aβ plaques in AD mouse models.1042
PD α-Synuclein SUMO1 modification promotes its aggregation to form Lewis bodies.1045
ALS TDP-43 SUMOylation promotes the formation of TDP-43 aggregates and affects the nuclear localization of TDP-43, involving in the pathological process of ALS.1049
HD HTT SUMOylation of HTT increases its insolubility and toxicity, leading to the accumulation of HTT.1059
CVDs Ischemia Drp1 DUSP6 SUMOylation at K234 is antiapoptotic during reperfusion.1060
Myocardial IRI PPAR-γ Overexpression of PIAS1 alleviates injury of myocardial I/R by increasing SUMOylation of PPAR-γ at K365 and downregulating NF-κB pathway.1061
IRI SERCA2a SERCA2a SUMOylation at K585 enhances intracellular mitochondrial membrane potential and reduces cell apoptosis, which promotes the recovery of cardiac function and reduces the infarct area in vivo.1062
Cancers Various cancers β-catenin SUMOylation of β-catenin prevents its ubiquitination and degradation.1063
Prostate cancer AR Modification by SUMO1 attenuates AR's transcriptional activity.1064
Prostate cancer HK2 SUMOylation-deficient HK2 promotes the growth of prostate cancer cells that resist chemotherapeutic drug-induced apoptosis.1028
APL PML/RARA Arsenic-enhanced PML/RARA SUMOylation promotes degradation.1065
Breast cancer hTERT CBX4 regulates SUMOylation of hTERT to promote the migration and invasion of breast cancer cells.1036
HCC METTL3 SUMOylation of METTL3 regulates HCC progression by controlling Snail mRNA homeostasis in a m6A methyltransferase activity dependent manner.1037
Aging Aging UBC9 The SUMOylation of UBC9 at K49 is conducive to its relocation to PML-NBs and promotes the translocation of target proteins into nucleus, which can transmit the antiaging phenotype.1012
Sp1 A gradual decrease in Prdx6 expression is associated with increased Sp1 SUMOylation and decreased Sp1 expression during aging.1013, 1014
Metabolic disorders Diabetes ICA512 PIASy reduces the interaction between ICA512 and STAT5 through SUMOylation of ICA512 and inhibits insulin secretion.1019
Diabetes mellitus and myocardial infarction ERK5 ERK5 SUMOylation enhances the inhibition of ROS-mediated ERK5 transcription, which leads to the deterioration of left ventricular function after myocardial infarction in diabetic patients.1021
Obesity ERp44 SUMOylation of ERp44 enhances Ero1α ER retention, thereby resulting in ER stress associated with aberrant lipid metabolism and obesity.1066
  • RA, rheumatoid arthritis; IBD, inflammatory bowel disease; IECs, intestinal epithelial cells; AD, Alzheimer's disease; NFT, neurofibrillary tangle; APP, amyloid precursor protein; PD, Parkinson's disease; ALS, amyotrophic lateral sclerosis; HD, Huntington's disease; HTT, huntingtin; IRI, ischemia–reperfusion injury; AR, androgen receptor; APL, acute promyelocytic leukemia; HCC, hepatocellular carcinoma.

8.2 Sumoylation in development

The role of SUMOylation in embryonic development has been confirmed in Drosophila,986 nematodes,987 zebrafish,988 Xenopus laevis,989 silkworms,990 and various plants.991 For example, an imbalance of SUMO leads to defects in embryonic patterning in Drosophila,992 while the absence of SUMO activity disrupts multiple signaling pathways and causes neural tube and heart defects in Xenopus embryos.989 In mammals, dysregulation of SUMO leads to defects in embryonic development, craniofacial defects,993 and even embryonic lethality. SUMO-deficient mice suffer from severe developmental disabilities, and mice die during embryonic development. Further studies have found that only SUMO2-deficient mice die earlier in embryonic development, while SUMO1- or SUMO3-deficient mice survive and reproduce well.969, 994, 995

The balance of SUMOylation in organisms is critical to the development of tissues and organs such as the heart,996 blood vessels,997 reproductive system,998 lung,999 and nervous system.1000 Too high or too low SUMO levels may lead to organ dysfunction. SUMO1 has important and specific functions in normal heart development. Both hetero- and homozygous SUMO-1 knockout mice exhibited atrial septal defects and ventricular septal defects with high mortality rates, which were rescued by cardiac reexpression of the SUMO-1 transgene.1001 Similarly, high expression of SENP2 can enhance deSUMOylation in the mouse heart, leading to congenital heart defects and cardiac dysfunction in mice.1002 Angiogenesis is essential for embryonic development and tissue growth, and the NOTCH pathway is a significant negative regulator of endothelial sprouting and vascular growth. SUMOylation negatively regulates angiogenesis by targeting endothelial NOTCH signaling. Endothelial SENP1 deletion in newly generated mice significantly delayed retinal vascularization by maintaining prolonged NOTCH1 signaling.997 The dynamic SUMOylation of endothelial FGFR1 regulates the balance of the angiogenesis core pathways VEGF/VEGFR and FGF/FGFR and enables the body to complete angiogenesis in different microenvironments.1003 SUMOylation also plays an important role in the generation of germ cells. SUMO1 and SUMO2/3 function at different stages of male meiosis and precisely regulate the formation of sex chromosomes.1004 C/EBPα is a core TF that regulates cell growth and differentiation. During lung development, C/EBPα is SUMOylated and participates in C/EBPα-mediated lung growth and differentiation.1005 Utf1 is a key SUMOylation target during neurogenesis and determines normal neurogenesis.1006 SENP2 can regulate the calcium homeostasis of mouse neurons through the SENP2–PLCβ4 signaling axis and then regulate neurogenesis in the hippocampus.1007 Furthermore, SUMOylation is important to maintain the development of human induced pluripotent stem cells.1008

8.3 Sumoylation in aging

During normal aging, SUMOylation is very important.1009, 1010 In C. elegans, enhanced insulin/IGF signaling activity promotes SUMOylation of the germ cell protein CAR-1, resulting in shortened lifespan and impaired proteostasis.1011 The degree of SUMOylation at Lys49 of UBC9 increases during aging. It has been found that the SUMOylation of UBC9 at Lys49 is conducive to its relocation to PML-NBs and promotes the translocation of target proteins into the nuclear bodies, which can transmit the antiaging phenotype. Whereas SUMOylation of proteins by the non-SUMOylated UBC9 promotes senescence.1012 Persistent DNA damage triggers cells to undergo apoptosis or senescence to prevent replicating a damaged genome. Sp1, a protein involved in double-strand break (DSB) repair, has been linked to aging, with Sp1 levels decreasing with age. Proteasomal degradation of Sp1 in senescent cells is mediated via SUMOylation, where SUMOylation of Sp1 on lysine 16 is increased in senescent cells. Prdx6 is important in maintaining redox homeostasis and localizes to ROS-producing organelles. A gradual decrease in Prdx6 expression is associated with increased Sp1 SUMOylation and decreased Sp1 expression during aging.1013, 1014 SUMOylation of KLF1 at K74 regulates its transcriptional activity during erythroid differentiation. K74 SUMOylation deficiency contributes to health and longevity in mice.1015, 1016 SUMO can also affect lifespan by affecting the mitochondrial unfolded protein response (UPR). During mitochondrial stress, ULP4 prolongs lifespan in C. elegans by removing SUMOylated DVE1 and ATFS1 (Table 8).1017

8.4 Sumoylation in metabolic disorders

SUMOylation also plays an important role in the regulation of cellular metabolism.966 SUMOylation acts on factors related to cholesterol homeostasis, including SREBPs and members of the nuclear receptor superfamily, such as LXR, FXR, LRH1, and PPAR. These receptors are potential therapeutic targets for lipid metabolism disorders.1018 Studies have found that insufficient insulin secretion or output disorders will eventually lead to diabetes. The E3 SUMO ligase PIASy can inhibit insulin secretion by reducing the interaction between ICA512 and STAT5 through SUMOylation of ICA512.1019 Furthermore, SUMOylation can prevent stress-induced β-cell apoptosis by upregulating the levels of antioxidant genes, including Ho-1, Cat, and Nqo-1.1020 ERK5 is one of the major targets of SUMOylation in diabetic hearts. ERK5 SUMOylation enhances the inhibition of ROS-mediated ERK5 transcription, which leads to the deterioration of left ventricular function after myocardial infarction in diabetic. The phenotype can be significantly reversed by inhibiting ERK5 SUMOylation.1021 Overexpression of SUMO4 promotes SUMOylation of IκBα, which inhibits the activation of NF-κB by external stimuli. This modification is considered to be related to type 1 diabetes (Table 8).1022

8.5 Sumoylation in cancers

The important role of SUMOylation in human tumorigenesis has gradually emerged. SUMOylation enzymes E1, E2 and E3 are highly expressed in many types of tumors. For example, Ubc9 is associated with the occurrence and development of ovarian carcinoma, advanced melanomas, colon cancer and primary prostate cancer.968 SUMO E3 ligase PIAS1 has been implicated in the regulation of several oncogenes and tumor suppressors. In B-cell lymphoma, PIAS1-mediated SUMOylation of Myc leads to a longer half-life of the protein and increased oncogenic activity, contributing to the development of B-cell lymphoma.1023 PIAS1 is highly expressed in prostate cancer and leads to accelerated tumor cell proliferation by inhibiting p21 expression.1024 High expression of PIAS3 is very common in CRC.1025

SUMOylation has an impact on cancer cell signaling and gene networks that regulate DNA damage, metabolism, inflammation and immunity, which provides link with carcinogenesis, proliferation, metastasis and apoptosis. SUMOylation is important for mammalian DNA damage response. BRCA1 participates in the DNA damage response and mutations in BRCA1 are associated with a high risk of breast and ovarian cancer. Hybrid SUMO-Ub chains are synthesized by RNF4, a SUMO-targeted Ub E3 ligase, and recognized by RAP80 to promote BRCA1 recruitment and repair of DNA DSBs. PIAS1 and PIAS4 are necessary for efficient Ub-adduct formation by RNF8, RNF168, and BRCA1 at DNA damage sites.1026, 968

SUMOylation is also involved in the regulation of cancer metabolism. For example, K270 of PKM2 can be modified by SUMO1, which promotes its transformation from a tetramer to a dimer. After entering the nucleus, it binds to the SIM on RUNX1, recruits RUNX1, and regulates the differentiation process of leukemia cells.1027 K315 and K492 are the SUMOylation sites of HK2. SUMOylation-deficient HK2 enhances its binding to mitochondria, which reduces mitochondrial oxidative phosphorylation and increases glycolysis and lactic acid production. This process promotes the growth of prostate cancer cells that resist chemotherapeutic drug-induced apoptosis.1028

SENP3 is involved in the regulation of immune cell function, which in turn affects the progression of tumor development.1029 Knockdown of SENP3 in DCs inactivates the STING-dependent type-I IFN signaling pathway and weakens the antitumor immune response. In the tumor microenvironment, SENP3 senses oxidative stimulation from DCs by deSUMOylating IFI204 and activating the STING signaling pathway.1030 SENP7 can regulate the metabolic homeostasis and antitumor activity of CD8+ T cells in response to oxidative stress. The key deSUMOylation substrate of SENP7 in this process is PTEN.1031

SUMOylation plays a role in cell differentiation and carcinogenesis. Myc mutation results in the constitutive expression of Myc protein, which leads to the uncontrolled expression of various genes, including those that drive cell proliferation, ultimately leading to cancer. Loss of SAE1/2 enzymatic activity is synthetically lethal with Myc. SUMOylation-dependent Myc switchers (SMS genes) are necessary for Myc-driven tumorigenesis. Patients with breast cancer who have Myc overexpression and low expression of SAE1/2 show significantly reduced cancer cell metastasis and improved survival compared with those with high SAE1/2 expression. Similarly, in a Myc-overexpressing PDAC model, the highly selective SAE small-molecule inhibitor ML-93 significantly inhibited protein SUMOylation and tumor growth.1032 In addition, BIRC5, EG5, and TPX2 are synthetic lethal partners of Myc. The functions of these three proteins are also dependent on SUMOylation.1033, 1034 Moreover, SUMOylation is associated with tumor metastasis.1035 hTERT is the catalytic component of human telomerase, and SUMOylation of hTERT promotes the migration and invasion of breast cancer cells.1036 The SUMOylated E2-conjugating enzyme Ubc9 modifies METTL3 through SUMO1, and the SUMOylated METTL3/Snail axis is correlated with high metastatic potential of liver cancer (Table 8).1037

8.6 Sumoylation in neurodegenerative diseases

Tight control of the CNS by SUMOylation is critical for maintaining neuronal cell viability, function, and connectivity.1038 SUMOylation plays important roles in the repair of DNA damage in neurons, axonal mRNA transport, and the regulation of synaptic plasticity.1039 Dysregulation of SUMOylation has been observed in the AD brain, with increased levels of hippocampal SUMO1 transcription possibly contributing to Aβ aggregations and impaired learning and memory abilities.1040 In fact, several proteins involved in the physiopathological process of AD, such as BACE1, GSK3-β tau, Aβ precursor protein (AβPP), and JNK, are in fact subject to protein SUMOylation or interactions.1041 Mature Aβ is produced by hydrolysis of APP. SUMO1 modification of APP promotes the generation of Aβ plaques in AD mouse models.1042 In addition, melatonin can induce the SUMOylation of the APP intracellular domain (AICD). SUMOylation of AICD activates the transcription of two Aβ-degrading enzymes, which promotes Aβ degradation and delays the occurrence of AD.1043 SUMO can also modify Tau. SUMOylation and phosphorylation of Tau promote each other and inhibit ubiquitination-dependent Tau degradation, thereby promoting NFT formation.1044 α-synuclein has two SUMOylation sites, Lys96/102, which are modified by SUMO1.1044 SUMOylation may regulate the normal and pathological functions of α-synuclein, including degradation, intracellular distribution, and PPIs and aggregation.1045 SUMOylation promotes PD onset by preventing proteasomal degradation of α-synuclein.1046 SUMOylation is also involved in regulating the pathogenesis of ALS.1047 Aggregation of SOD1 is characteristic of patients with SOD1 variant-induced ALS. The modification of SOD1 by SUMO3 enhances the aggregation of familial ALS (fALS)-linked SOD1 mutants, while SENP1 decreases the number of cells exhibiting SOD1-mutant aggregation.1048 TDP-43 can also be SUMOylated. SUMOylation promotes the formation of TDP-43 aggregates and affects the nuclear localization of TDP-43, which is involved in the pathological process of ALS (Table 8).1049

8.7 Sumoylation in CVDS

SUMOylation is also closely related to the development, metabolism, and pathology of the heart. For example, SUMO1 is essential for normal cardiac development, and cardiac-specific overexpression of SUMO1 improves cardiac functions. SUMO1 mutant or knockout mice are more prone to congenital heart defects.1050 Specifically, UBC9 is the sole E2 enzyme essential for GATA4's role in cardiac development and function. It boosts GATA4's transcriptional activity and affects its nuclear localization. UBC9 adds a SUMO group to GATA4 at K366, which activates particular gene expression in pluripotent cardiac cells.1051

Protein SUMOylation and its dysregulation are implicated in various CVDs, including atherosclerosis, heart failure, and ischemic cardiomyopathy. Several factors, including ERK5, NF-κB, p53, and PKC, undergo SUMOylation, which contributes to atherosclerosis progression.1051 For example, SUMO1-mediated SUMOylation of NF-κB inhibits IκBα degradation and reduces NF-κB activation.1052 In contrast, SUMO2/3 modification promotes IκBα detachment from NF-κB and enhances NF-κB activation, ultimately inducing atherosclerosis.1052, 1053 Additionally, SUMOylation of HSF2, myocardin, PARIS, PPARγ1, and SERCA2a plays crucial roles in heart failure progression.1051 The protein levels of SUMO1 and SUMOylated SERCA2a are significantly decreased in failing hearts, whereas increased SUMOylation of SERCA2a improves myocardial contractility and ventricular function in heart failure mice.1054 Conversely, hypoxia induces SUMOylation of HIF-1α, which promotes HIF-1α degradation, whereas SENP1 deSUMOylates HIF-1α and enhances its stability.1055 Reduced SENP1 levels exacerbate ischemia/reperfusion (I/R) injury in cardiomyocytes through the HIF-1α pathway, while HIF-1α overexpression counteracts the detrimental impact of SENP1 downregulation on cell death (Table 8).1056

8.8 Sumoylation-associated targeted therapies

Targeted cancer therapy may be achieved by inhibiting the SUMO pathway.1067 In 2021, Bellail's team screened the hit compound CPD1 that can specifically target and degrade SUMO1 from the drug-like compound library of the National Institutes of Health (NCI). CPD1 specifically reduces SUMO1 protein levels without affecting SUMO1 mRNA levels. Further druggability optimization identifies the first highly selective SUMO1 degrader, HB007. HB007 inhibits the proliferation of various tumor cells by selectively degrading SUMO1.1068 The highly selective SAE inhibitor TAK-981 can significantly upregulate the expression of IFN1 and activate IFN1-dependent innate immune cells, including macrophages, NK cells, DCs, and T cells, promoting antitumor immune responses.1069 Moreover, TAK-981 is currently in phase 1 clinical trials in patients with solid tumors and lymphomas.1069 Although there are still many blind spots on the mechanism of SUMOylation involved in the occurrence and development of tumors, AD/PD, and CVDs at this stage, a large number of experiments have confirmed the role of SUMOylation in these diseases. Therefore, drugs targeting the SUMOylation mechanism may represent a promising treatment strategy.1070

9 GLYCOSYLATION

Protein glycosylation is a process in which sugar groups are transferred to proteins catalyzed by glycosyltransferases, which predominantly occurs in the ER and Golgi apparatus. Most glycans exist on the surface of cells and secreted proteins, with intricate and varied structures.1071-1073 In contrast, the types of glycosylation existing in the nucleus and cytoplasm have simple structures and are highly dynamic.1074, 1075 The structural diversity and extensive distribution of protein glycosylation make it one of the most prevalent forms of PTMs in humans.1076

Protein glycosylation is a complicated process that involves multiple steps. The human genome contains approximately 700 genes related to glycosylation and deglycosylation, including enzymes, transporters, and chaperones.1077-1079 Of these genes, approximately 200 encode glycosyltransferases involved in the construction of complex glycans on proteins.1078 These glycans are assembled from ten monosaccharides, including N-acetylgalactosamine (GalNAc), N-acetylglucosamine (GlcNAc), xylose (Xyl), fucose (Fuc), galactose (Gal), mannose (Man), glucose (Glc), glucuronic acid (GlcA), iduronic acid (IdoA), and sialic acid (SA).1076, 1077 The monosaccharides are linked to nucleotides or lipids to form an activated donor substrate, which is then extended through the action of glycosyltransferases to form approximately 1012 different glycan structures.1077, 1080, 1081 There are four ways of linking glycans to proteins, including N-glycosylation to asparagine (Asn) residues, O-glycosylation to serine (Ser), threonine (Thr) or tyrosine (Tyr) residues (O-linked monosaccharides include GalNAc, GlcNAc, Gal, Glc, Man, Fuc, and Xyl), C-mannosylation to tryptophan (Trp) residues, and glypiation. Based on these linking methods, protein glycosylation is divided into 14 different types, including N-glycosylation, 11 types of O-glycosylation, C-mannosylation, and glypiation.1076, 1082 These glycosylation modifications play important roles in regulating various intracellular and extracellular protein functions and are involved in a variety of biological processes in humans (Table 9).

TABLE 9. Protein glycosylation through different linkages and their function.
Types Linkages Enzymes Modified sequence in glycoproteins Functions of glycosylation
Sequence motifs Specific domain
N-glycosylation GlcNAc-β-Asn OST complex (STT3A/STT3B)

N-X-T/S, X≠P

N-G; N-X-C/V, X≠P

None Protein stability,1118 protein folding and quality control,1119 self/nonself recognition,1120 cell adhesion1121, immunotherapy,1122 receptor activation and endocytosis,1123 glycoediting and drug delivery.1124
O-glycosylation GalNAc-α-Ser/Thr GALNT1–20 Weak isoform specific motifs1093 None O-glycan shielding is essential for secretion of an active protein,1125 participates in immunological recognition of the immune system,1082 protects membrane proteins from ectodomain shedding,1090 increases the half-life of peptide hormones in circulation,1126 modulates the interaction between viral proteins and host surface receptors.1127 O-glycosylated mucins expressed at mucosal surfaces such as the respiratory and GI tract can form an effective barrier against pathogens.1128 The particular glycosylation of LDLR class A repeat linker regions by GalNAc-T11 alters the rate of uptake by cargo receptors.1129 O-glycan is also important for leukocyte extravasation.1130
GALNT11 C6-X3-5-T-C1 1131 LA
GlcNAc-β-Ser/Thr EOGT C5-X2-(G/P/S)-(Y/F/W)-(T/S)-G-X2-C6 1132 EGF
GlcNAc-β-Ser/Thr OGT None None
Gal-β-Hyl COLGALT1–2 X-Hyl-Gly Collagen repeats
Glc-α-Tyr GYG Tyr194 of GYG None
Glc-β-Ser POGLUT1 C1-X-S-X-(A/P)-C2 EGF
POGLUT2–3 C3-X-N-T-X-G-S-(F/Y)-X-C4
Fuc-α-Ser/Thr POFUT1 C2-X-X-X-X-(S/T)-C3 EGF
POFUT2 C-X-X-(S/T)-C-X-X-G TSR
Man-α-Ser/Thr POMT1–2 None None
TMTC1–4 None EC
Unknown None IPT
Xyl-β-Ser XYLT1–2 a-a-a-a-G-S-G-a-(a/G)-a (“a” represents Asp or Glu) None
C-mannosylation Man-α-Trp DPY19L1–4 W-X-X-W TSR Support folding, enhance stability of thrombospondin repeats,1106 regulate protein folding and maturation,1133 play a key role in the folding, sorting and secretion of substrate proteins.1107, 1108
Glypiation Pr-C(O)EthN-6-P-Man Transamidase Carboxy-terminal hydrophic segment None Regulate protein location on cell membrane,1116 cell signal transduction, cell adhesion, and immune recognition.1117

N-glycosylation occurs on Asn residues of proteins. The sugar complex with GlcNAc2Man3 as the core is linked to the N atom on the Asn side chain via GlcNAc, and various enzymes are recruited to remove or add monosaccharides. Based on the polysaccharide structure, glycans can be classified into high-mannose N-glycans, hybrid N-glycans, and complex N-glycans.1083 The oligosaccharyltransferase (OST) complex catalyzes the initiation of N-glycosylation in the ER, transferring a 14-saccharide precursor structure (GlcNAc2Man9Glc3) to the Asn-X-Ser/Thr (X represents other amino acids except Pro) motif. This glycosylation process can be cotranslational or posttranslational, and regulated by the STT3A and STT3B catalytic subunits of OST, respectively.1084, 1085 The OST–STT3A complex is mainly responsible for the cotranslational glycosylation of nascent peptides when they enter the ER cavity, while the OST–STT3B complex is mainly responsible for the release of the glycans on misfolded N-glycoproteins into oligosaccharides, which are the main source of oligosaccharides.1084-1086 After processing in the ER, the glycans in the precursor structures are moved to the cis-Golgi and modified by a series of specific mannosidases and then transferred to the inside of the Golgi for further processing and maturation (Figure 18).

Details are in the caption following the image
Overview of human N- and O-glycosylation in the ER and Golgi apparatus. On the right side, the biosynthesis of complex-type N-glycans is shown. On the left side, the biosynthesis O-glycosylation is shown.

Protein O-glycosylation mainly occurs on amino acids with functional hydroxyl groups on the side chains, such as Ser and Thr. The monosaccharides linked to Ser and Thr residues in humans are mainly GalNAc and GlcNAc.1087-1089 GalNAc-type O-glycosylation is present on extracellular and secreted glycoproteins, such as mucins,1090, 1091 and this type of O-glycosylation is initiated in the Golgi apparatus and regulated by up to 20 GalNActransferases (GALNTs), of which 15 isozymes have been demonstrated to be active enzymes.1091, 1092 GALNTs exhibit some specificity, but there is no specific amino acid motif for recognition in substrate proteins. The site and type of protein O-glycosylation are coregulated by different transferases in a cooperative manner (Figure 18).1091, 1093 O-glycosylation usually has the same glycan core structure (e.g., cores 1–4, sialyl-Tn antigens and terminal GalNAc), and the glycan is further extended on the basis of the core structure, which protects glycoproteins and cell surfaces from external stress and microbial infection, which affects the self-recognition process of the immune system.1094, 1095 GlcNAc-type O-glycosylation mainly exists on glycoproteins in the cytoplasm, mitochondria, and nucleus.1096-1098 Unlike O-GalNAcylation, O-GlcNAcylation starts in the ER. It does not typically occur in the Golgi apparatus, and the glycan does not extend further. The formation of O-GlcNAcylation is mainly regulated by O-GlcNAc transferases (OGTs) and O-GlcNAcases (OGAs).1089, 1099, 1100 O-GlcNAcylation also plays an important role in cells and is closely associated with protein stability and localization, intracellular signal transduction, chromatin remodeling, and mitochondrial function.1101 Although OGTs and OGAs exist in different subcellular compartments in different forms, they share the same function of adding GlcNAc to protein substrates or removing GlcNAc from protein substrates. This maintains the homeostasis of O-linked GlcNAc and plays an important role in regulating cellular functions.1102, 1103

C-Mannosylation is a relatively rare form of protein glycosylation. The monomeric D-mannopyranose bound with the C-2 position of the pyrrole ring of the tryptophan residue to form a carboncarbon bond.1104 C-Mannosylation occurs on proteins in the ER and is regulated by four dpy-19-like C-Man transferases (DPY19),1105 which transfer monomeric α-mannose to the substrate on the first Trp residue in the Trp-X-X-Trp/Cys motif.1106 C-Mannosylation plays an important role in protein folding, sorting, and secretion,1107, 1108 and it has been determined that 18% of human proteins undergo C-mannosylation during secretion and transmembrane transport.1109 The known protein substrates of C-mannosyltransferases include the TSR superfamily and type I cytokine receptor family.

Glypiation is a special type of glycosylation that localizes proteins to the cell membrane via glycosylphosphatidylinositol (GPI). The GPI anchor contains a phosphoethanolamine linker, which binds to the C-terminus of the target protein. The GPI sugar chain core structure contains a phospholipid tail, which anchors the structure to the membrane.1110 Similar to the synthesis of glycan precursors required for N-glycosylation, the biosynthesis of GPI anchors begins on the cytoplasmic side of the ER. Glypiation-modified proteins usually have two signal sequences. The N-terminal and C-terminal signal sequences determine transport into the ER and recognition by the GPI transamidase complex GPIT, respectively.1111 C-terminal sequence recognized by GPIT for covalent binding to the GPI anchor. During the synthesis of GPI anchors, sugars on membrane-embedded phosphatidylinositol (PI) molecules originate from sugar nucleotides and dolichol-P-mannose around the ER. The residues at the phosphoethanolamine (EtN-P) linker are provided by phosphatidylethanolamine in the ER cavity.1112, 1113 GPI anchors recruit specific proteins to the cell membrane for their crucial roles, and enzymes such as phospholipase C are responsible for the cleavage of GPI anchors, which regulate cell membrane protein localization.1114-1116 The diverse glypiation makes it critical in cell signaling, cell adhesion, and immune recognition.1117

9.1 Glycosylation in development

During the developmental process, protein glycosylation occurs at various times and locations, and these characteristic glycans cover the surface of nearly all cells. The numerous and complex structures of glycans provide strong support for cell-to-cell communication during growth and development.1134, 1135 Genetic defects in glycosylation are often embryonic lethal.1136 Congenital disorders of glycosylation (CDGs) are diseases caused by disorders of glycoprotein synthesis, with various clinical manifestations, such as the appearance of special facial features and lesions in the organs of the body, which can be divided into Type I and Type II CDGs.1137, 1138 The causative factors of CDGs include abnormal activation, presentation, or transport of glycolipid precursors, abnormal expression or activity of glycosidases or glycosyltransferases, and abnormal functions of proteins that control glycosylation or maintain the Golgi apparatus.1082 Furthermore, during embryonic development, N-glycosylation plays a key role in the generation of hematopoietic stem cells (HSCs) from arterial endothelial cells through the process of endothelial to hematopoietic transition (EHT) and is a determinant of hematopoietic fate.1139 CDGs often have serious consequences, suggesting the important roles of protein glycosylation in maintaining the normal growth and development of individuals (Figure 19).

Details are in the caption following the image
Functions of glycosylation in health and diseases, such as congenital diseases, immune regulation, neurodegenerative diseases, IBD, cancers, aging, and infectious diseases.

9.2 Glycosylation in aging

Protein glycosylation is a dynamic process highly sensitive to aging,1140 and acts as a potential important molecular effector in aging and age-related diseases.1141 Numerous studies have shown that the glycan at Asn279 on IgG heavy chains undergoes changes during aging that are often similar to those found in inflammatory states,1142 but the mechanism by which the glycans at Asn279 on IgG heavy chains are altered is unclear (Figure 19).1143, 1144 In addition, reduced expression of the glycosyltransferase B4GALT1 can prevent senescence-associated IgG glycan changes and improve the senescence phenotype.1145 The glycosyltransferase ST6GAL1 can alter the glycans of fibroblasts, leading to the transition of fibroblasts to proinflammatory cells.1146 Sialic acid-binding immunoglobulin-like lectins (Siglecs) act as inhibitory receptors on the immune cell surface.1147 The activity of Siglecs appears to correlate with longevity, possibly due to their ability to suppress aging-related inflammation.1148

9.3 Glycosylation in immunity

A large number of glycoproteins on immune cell surfaces are extensively involved in many immune processes by receiving signals from the extracellular environment.1149 Neutrophils are the most abundant innate immune cells and act as the host's first line of defense against pathogen invasion, utilizing bioactive glycoproteins assembled in the cytoplasm to fight pathogenic infections.1150, 1151 N-glycosylated and O-glycosylated proteins exhibit structural and functional diversity in different life stages of neutrophils during bone marrow maturation, blood circulation, and sterilization of inflammatory peripheral tissues.1152 Numerous modified granule glycoproteins are present in neutrophils, including neutrophil elastase, myeloperoxidase, and cathepsin G.1153 In addition, neutrophils possess unique glycans that are not typically observed, such as hypertruncated chitobiose core- and paucimannosidic-type N-glycans and monoantennary complex-type N-glycans.1154, 1155 The glycoproteins major histocompatibility complex (MHC) classes I and II play key roles in adaptive immunity. They are closely associated with the presentation of cell surface antigen peptides and circulating T lymphocyte recognition and activation. Glycosylated protein antigens are critical for antigen uptake by cells, proteolysis, antigen presentation by MHC, and activation and initiation of T cells.1156, 1157 In the adaptive immune system, protein glycosylation also has multifaceted roles in the differentiation of B cells and T cells, cell–cell interactions and the recognition of glycosylated antigens (Figure 19).1157-1159 Protein glycosylation has broad impacts on the function of the immune system.

9.4 Glycosylation in the gut

There are numerous glycosylation modifications on the surface proteins of IECs that form a physiological barrier to protect the intestinal tract from bacterial infection and invasion. For example, Golgi glycosyltransferases modify mucin in the gut to separate IECs from commensal microorganisms.1160 Changes in the glycosylation level of proteins on the surface of IECs can result in the destruction of the mucus layer and the occurrence of intestinal IBD (Figure 19).1161 Glycosylation of IEC surface proteins can also alter the structure and function of the microbiota.1162 IL-22-mediated glycosylation of intestinal cells facilitates the growth of succinate-consuming Bacillus in the gut microbiome, which reduces the availability of succinate, a key metabolite for Clostridium difficile growth, and prevents Clostridium difficile infection.1163

9.5 Glycosylation in neurodegenerative diseases

A growing number of studies have shown significant differences in the levels of N-glycosylation and O-glycosylation in brains between AD patients and healthy individuals.1164-1168 APP is concentrated at the synapses of neurons. After being cleaved by proteases, it produces toxic Aβ protein that accumulates in AD patients.1169 N-glycosylation- and O-glycosylation-modified APP is found in the cerebrospinal fluid of AD patients.1170 The structure of N-glycans can affect APP transport and Aβ production.1171 The modification of APP-linked N-glycans by sialylation may affect APP processing, resulting in increased APP secretion and Aβ production.1172 In addition, various O-glycosylation sites have also been identified on APP in human cerebrospinal fluid,1173 and O-glycosylation modification can also increase nonamyloidogenic α-secretase processing, thus affecting APP processing and reducing Aβ secretion.1174 In AD patients, Tau proteins modified by N-glycosylation and O-GlcNAcylation can also be detected,1175, 1176 and the levels of N-glycosylation in AD patients are higher than those in healthy people,1177 which may affect the aggregation of Tau proteins.1175 Interestingly, the level of O-GlcNAcylation is reduced in AD patients.1178

TREM2 expressed on myeloid cells is a PD-related protein that has multiple ligands, including APOE and lipids. By interacting with DNAX activating protein to transduce signals, TREM2 plays an anti-inflammatory role in various diseases, including PD,1179 and can be modified by sialylated and fucosylated complex glycans. N-glycan changes alter TREM2 conformation and affect the stability and antioxidant capacity of the protein.1180 α-Synuclein can also be modified by O-GlcNAcylation.1181 O-GlcNAcylation of α-synuclein affects its phosphorylation and blocks the toxicity of α-synuclein, suggesting that an increase in O-GlcNAcylation may prevent α-synuclein aggregation (Figure 19).1182

9.6 Glycosylation in viruses

Protein glycosylation is involved in the regulation of host–pathogen interactions and mediates the adhesion, recognition, invasion and immune evasion of pathogens in host cells, which affects pathogen virulence or host cell resistance.1183-1188 COVID-19 caused by SARS-CoV-2 is currently a major global health problem.1189 Protein glycosylation affects the toxicity and viability of SARS-CoV-2, which utilizes its highly glycosylated modified spike (S) protein to interact with the glycosylated host receptor ACE2 and to facilitate SARS-CoV-2 invasion of host cells.1127

Acquired immune deficiency syndrome caused by human immunodeficiency virus (HIV) seriously endangers human health. On the HIV-1 envelope (Env), there are approximately 25 and 4 glycosites on each gp120 monomer and gp41 subunit, respectively, and 18–33 glycans on each gp120 monomer. These glycans are dominated by the high mannose type and sialylated complex type.1190, 1191 which can bind to the chemokine receptors CD4 and DC-SIGN of host cells or mannose receptors of macrophages and regulate HIV-1 invasion of host cells (Figure 19).1192, 1193 Highly glycosylated HIV binds to C-type lectin receptors (CLRs) of different antigen-presenting cell subsets, possibly regulating T cell priming and B cell activation.1194, 1195 N-glycosylation of gp120 in HIV-1 is critical for CD4+ T cell recognition.1196 HIV infection also leads to altered glycosylation of host IgG. The levels of galactosylation in HIV+ patients are lower than those in healthy people, and it is more pronounced in the IgG1 subtype.1197 Low sialylation of IgG is also found in HIV-infected patients.1197 During the evolution of HIV, the glycans on the HIV-1 envelope become more complicated under the guidance of natural selection, which also makes HIV more diverse.1198

9.7 Glycosylation in cancer

O-GlcNAc of the key cell cycle regulators participates in the processes of cell division, DNA repair, and cell death and is dynamically changed in a cell cycle stage-dependent manner.1199 MUC1, a transmembrane glycoprotein associated with the cell cycle,1200 is overexpressed and aberrantly glycosylated in a variety of epithelial cancers and plays an important role in disease progression.1201 Dysregulation of glycosyltransferases such as ST6GalNAcI, C1GalT1, and ST3GalI alters the level of O-glycosylation of MUC1.1202-1204 In addition, changes in glycosylation motifs on MUC1 also affect cancer immune surveillance. For example, the binding of sialylated MUC1 to the cell surface lectin CD169 can enhance macrophage activation and promote tumor growth.1205

The growth of tumor cells is accompanied by immune escape. There are ligands of programmed death receptors on the surface of tumor cells, which can bind to programmed death receptors (such as PD1) on the surface of T cells, making T cells exhausted and unable to kill tumor cells. Aberrant glycosylation on the surface of tumor cells can alter how the immune system senses tumors and induce immunosuppressive signals. Thus, specific glycans on tumor cells represent a novel immune checkpoint. Aberrant glycosylation on the surface of tumor cells may affect antitumor responses. Changing the level of glycosylated proteins on the surface of tumor cells can enhance the killing effect of CAR-T cells on solid malignant tumors.1206 Moreover, the glycosylation of tumor cell surface proteins also provides new antigen targets for tumor-specific T cells (Figure 19).1207 The most common tumor-associated glycans include sialylated glycans, Tn antigen and Lewis antigen. The level of sialylation on melanoma cells correlates with the level of tumor growth in vivo, which is associated with the accumulation of Treg cells, reduction of effector T cells, and decreased activity of NK cells.1208 Poor survival in patients with stage III colon cancer is associated with BRAF mutation and increased Tn antigen.1209 In addition, increased Lewis antigen in the tumor microenvironment can drive innate immune suppression.1207

9.8 Therapeutic glycosylated proteins

Almost all therapeutic proteins are glycosylated, such as EPO, ENPP1, and IgG antibodies. Carbohydrate components play an important role in the safety and pharmacokinetic properties of these protein-based drugs.1210 Rapid advances in the field of glycobiology have provided more opportunities for the development of glycoprotein therapeutics.1211 Technologies are being developed for glycan processing of therapeutic proteins using chemical, chemoenzymatic and genetic approaches in different cell types.1210, 1212, 1213 In addition, engineered cells provide a more powerful tool for developing more complex protein glycan structures and improving the pharmacodynamic properties of therapeutic proteins.1211

EPO is an endogenous glycosylated hormone that can stimulate erythropoiesis and has a variety of important physiological functions.1214 It can be used to treat anemia caused by CKD and cancer. The recombinant EPO used for clinical treatment can be divided into four types according to their different glycosylation levels. Sialylation and branching N-glycans on EPO can prolong its half-life in serum,1215, 1216 while EPO lacking sialylation exhibits neuroprotective effects in vivo.1217 This shows the importance of the type of glycosylation for therapeutic EPO.

Mutations in ENPP1 cause generalized arterial calcification in infancy (GACI), an extremely rare neonatal disease associated with extensive arterial calcification and narrowing. ENPP1 cleaves ATP into PPi and AMP extracellularly. Recombinant ENPP1-Fc protein has preventive and therapeutic effects on GACI animal models.1218 Enhancing the sialylation level of recombinant ENPP1-Fc protein has a significant effect on prolonging the protein half-life and improving drug efficacy.1219

Glycosylation-dependent IgGs are therapeutic antibodies for cancers.1220, 1221 Glycosylation of the IgG Fc region affects the safety and clinical efficacy of therapeutic antibodies. Biantennary complex oligosaccharide modification of Fc at Asn297 is essential for the effector function of antibodies. Fucose and outer arm sugars linked to the core heptasaccharide create structural heterogeneity that also exhibits unique biological activities.1222 Clarifying the glycosylation profile of Fc is the key to the development and quality control of therapeutic antibodies.

Glycosylated vaccines also have broad application prospects. For example, changes in the glycans of the glycoprotein gp120 expressed on the HIV-1 envelope may lead to immune escape of the virus.1223 In the process of vaccine design, adding a new glycosylated epitope on recombinant gp120 is beneficial to improve the ability of neutralizing monoclonal antibodies to recognize HIV-1, thereby optimizing the design of viral vaccines.

10 CITRULLINATION

Citrullination refers to the irreversible process of converting arginine residues in proteins into citrulline residues under the action of protein arginine deiminases (PADs) (Figure 20A).1224 The conversion of positively charged arginine to charged citrulline affects hydrogen bond formation, protein folding, hydrophobicity, and protein interactions, ultimately leading to protein denaturation.1225, 1226 For example, the change of 5% of arginine residues into citrullination in hyalin or fibroin can affect its tertiary structure, and the citrullination of more than 10% of arginine residues may completely destroy the protein structure and denature it.1227 PAD includes five isozymes (PAD1–4 and PAD6) with tissue specificities (Figure 20B). PADs 1–4 are catalytically active, while PAD6 is not because of active site mutations.1228 The most common citrullination substrates of PADs are keratin, fibroin, vimentin, actin, histones, collagen, and myelin basic protein (Figure 20B).1229, 1230 Notably, free arginine cannot be citrullinated by PADs.1228 PADs have a high degree of sequence homology but differ in their tissue distribution. Both PAD1 and PAD3 are mainly in the hair follicle and epidermis, and PAD1 is also present in the uterus.1231, 1232 PAD2 and PAD4 are widely distributed throughout tissues. For example, PAD2 is expressed in skeletal muscle, brain, spleen, and secretory glands and is the most widely expressed PAD in the human body.1233, 1234 There are two main substrates of PAD2: myelin basic protein in the CNS and glial fibrillary acidic protein (GFAP). Citrullination of the former is involved in the pathogenesis of multiple sclerosis (MS), while the latter may be associated with senile dementia.1235 In contrast, PAD4 is present in neutrophils, macrophages, mammary cells, and tumor cells.1236-1238 PAD4 mainly catalyzes histones,1239 whose citrullination not only disrupts their structures but also causes them to lose many positive charges, thereby depolymerizing nucleosomes, breaking DNA, and ultimately leading to apoptosis. Notably, PAD4 is involved in histone citrullination in neutrophils. After translocation of activated PAD4 to the nucleus of neutrophils, neutrophil extracellular traps (NETs) are produced to trap bacteria and other pathogens.1240, 1241 Histone modifications control NETosis, which is linked to the development of autoimmune diseases such as RA, ulcerative colitis, and SLE.1242 PAD6 was originally discovered through sequence alignment and is located in early embryos, eggs, and ovaries.1238, 1243 PAD6 expression is correlated with the degree of citrullination in female germ cells despite lacking catalytic activity,1244 and its function may be involved in early embryonic development.1245

Details are in the caption following the image
The schematic reaction for citrullination regulated by PADs. (A) Cartoon depicting PAD-mediated conversion of arginine to citrulline. (B) Table listing the tissue distribution of PADs and their representative citrullination substrates.

PAD-mediated citrullination is regulated by various factors, such as Ca2+ concentration and amino acid sequence1246. The intracellular Ca2+ concentration is maintained at low levels to keep PAD inactive under physiological conditions.1247 However, certain PAD-mediated processes are operated under physiological Ca2+ concentrations, implying other unknown mechanisms of PAD action.1248, 1249 Apoptosis is dependent on high intracellular Ca2+ concentrations,1250, 1251 and protein citrullination is increased in apoptotic cells.1247

Protein citrullination has diverse biological functions. For example, citrullination is involved in cell apoptosis, which may be related to cell morphological changes and DNA fragmentation during apoptosis. The elevation of Ca2+ is involved in the early signal transduction and execution stage of apoptosis.1265, 1266 PAD2-mediated vimentin citrullination may result in changes in cell morphology.1247 Activated nuclear PAD4 induces nonspecific citrullination of histones, which disrupts protein structure, depolymerizes nucleosomes, and makes DNA more susceptible to nuclease cleavage, ultimately leading to apoptosis.1252 Citrullination can promote terminal differentiation of cells, and the expression of keratin changes as the epidermis progresses to terminal differentiation. Keratin 1, keratin 10, keratin 5, and keratin 14 are four major keratins expressed in the human epidermis. Keratin citrullination in the epidermis promotes terminal differentiation of keratinocytes.1244 Citrullination can regulate gene expression.1253 PAD2-mediated citrullination of histone H3 is associated with the regulation of the expression of more than 200 ER-related genes, such as HER2.1254 Moreover, PAD2 expression is significantly higher in blood and tissues from breast cancer patients than in those from normal controls.1255 When PAD2 is inhibited, the expression of ACSL4 and baculovirus-containing IAP repeats is reduced in the breast cancer cell line MCF-7.1256 Abnormal expression of these genes is related to dysregulated lipid metabolism and tumor cell invasion.1257

10.1 Citrullination in immune diseases

Inflammation, chronic pain, and polyarthritis are the main features of RA.1258 A growing amount of evidence shows that the autoimmune reaction to RA is driven by dysregulation of protein citrullination mediated by PAD4, as 75% of patients have anticitrullinated protein antibodies (ACPAs).1239 ACPAs are the most specific autoantibodies in RA serum.1259 Most of these autoantibodies can be detected early in the plasma, making them useful diagnostic markers for RA.1260, 1261 In inflammatory synovial tissues, macrophages and neutrophils express more PAD2 and PAD4, respectively. Both PAD2 and PAD4 are released into joint citrullinated proteins, such as fibrin, fibrinogen and vimentin,175 which further initiate immune responses and induce autoantibodies (Figure 21).1262, 1263 Blood coagulation factor citrullination is also important in RA.1264 Thrombin can generate inflammatory mediators that promote inflammation, resulting in excessive capillary formation and fibrin deposition in synovial tissues (Figure 21). PAD4-mediated citrullination of antithrombin inhibits thrombin activity, resulting in an increase in the coagulation rate.1265 The levels of citrullinated antithrombin are significantly elevated in RA patients compared to healthy controls.1229 Inhibiting hyperactivated PAD enzymes to reduce citrullination is a promising therapeutic strategy for RA patients.1266

Details are in the caption following the image
Representative citrullination substrates and their functions in various diseases, such as RA, cancer, MS, and SLE.

MS is characterized by inflammatory demyelinating lesions of the CNS. The pathogenesis of MS is complex and is currently thought to be caused by a combination of genetic and environmental factors.1264 MS is primarily associated with PAD2-mediated over-citrullination of myelin basic protein and GFAP, leading to demyelination and affecting nerve signal transduction (Figure 21).1267, 1268 Overexpression of PAD2 in transgenic mice increases the amount of citrullinated myelin basic protein and accelerates demyelinating changes.1269 Citrullinated myelin basic protein alters its processing and presentation by T cells. Despite the presence of T cells specific for citrullinated myelin peptides, no autoantibodies against citrullinated proteins have been found in the serum of MS patients.1270 Furthermore, the upregulation of PAD2 and inflammatory signaling may locally increase PAD4 to further aggravate inflammatory disease. Collectively, myelin basic protein citrullination mediated by PAD2 and PAD4 promotes proteolysis, demyelination, and signaling blockade, ultimately leading to MS.

SLE is manifested by the immune system attacking healthy cells and tissues throughout the body. SLE immune system activation is characterized by B cell and T cell hyperreactivity and loss of immune tolerance to self-antigens.1271 Autoantigens in SLE patients mainly come from apoptosis and the formation of NETs, and citrullination is involved in these processes and affects the occurrence and development of SLE. For example, LL37 binds to self-DNA/RNA and stimulates plasmacytoid dendritic cells to produce IFN-I, leading to an autoimmune response. In the skin and kidney of SLE, citrullinated LL37 (cit-LL37) is significantly increased, and LL37-specific T cells show a significant response to cit-LL37 (Figure 21).1272 Many NETs-related proteins are posttranslationally modified, especially histones found to be methylated, acetylated, and citrullinated, indicating that NETs may be a source of self-antigens in autoimmune diseases (Figure 21).1273 Aberrant apoptotic pathways prevent immune cell clearance, which prolongs the exposure of self-antigens and induces the production of autoantibodies.1264, 1274 Among the autoantibodies associated with SLE, many antigens, including nuclear DNA and proteins, can be detected in NETs.1283, 1293

10.2 Citrullination in cancers

p53 is a well-known tumor suppressor and TF.1275 Based on pathological studies of patient samples, PAD4 is highly expressed in a variety of tumors, including colon cancer, esophageal cancer, ovarian cancer, PC, and gastric cancer,1237, 1276 suggesting possible involvement of PAD4 in tumorigenesis. PAD4 expression is regulated by p53. PAD4 citrullinates the growth inhibitor ING4, which subsequently prevents the binding of p53 to ING4 to inhibit p53 expression, further inhibiting downstream p21 expression and promoting tumor growth (Figure 21). Notably, PAD4 forms negative feedback to regulate p53 through histone citrullination.1277 PAD inhibitors may also prevent the expression of genes related to cancer cell invasion and metastasis.1256

10.3 Citrullination in inflammatory diseases

Ulcerative colitis is a chronic inflammatory disease occurring in the colonic mucosa. Ulcerative colitis patients often suffer from serious complications, the most common of which is peripheral arthropathy.1278 Immunohistochemical analysis shows increased PAD2 and PAD4 in damaged tissues in ulcerative colitis patients.1279 The upregulation of PAD4 is associated with neutrophils and NETs in the colonic mucosa of ulcerative colitis patients, even in remission.1280 Proteins that promote NETs formation are potential therapeutic targets to reduce inflammation in ulcerative colitis.1281 The severity of acute colitis in PAD4-deficient mice with considerable rectal bleeding is evidence that PAD4 is essential for maintaining colitis mucosal homeostasis and regulating rectal bleeding.1282

11 CARBAMYLATION

Protein carbamylation is a nonenzymatic modification mediated by cyanate. Cyanate reacts with the amino groups of proteins to form carbamylation.1283 Spontaneous carbamylation can alter the molecular weight, isoelectric point, and other physical properties of the protein and can also lead to an irreversible decrease in protein activity.1284 In theory, all proteins in the human body are prone to carbamylation. However, the probability of carbamylation of each protein depends on the number and activity of amino groups and the lifetime of the protein. Protein carbamylation preferentially occurs at the side chains of lysine residues, also known as homocitrullination.1285

There are two main pathways for the formation of cyanate (Figure 22A). First, urea is a product of protein catabolism that decomposes slowly and spontaneously in aqueous solution to form cyanic acid and cyanate.1286 Under normal physiological conditions, the cyanate concentrations in body fluids are also extremely low and do not cause extensive carbamylation in the body.1287 Second, under certain inflammatory conditions, MPO alters the balance between cyanate and thiocyanate, causing an increase in cyanate synthesis and leading to carbamylation in the body.1288 In addition, smoking also increases the concentration of cyanate and carbamylation in the body, which leads to the occurrence of diseases.1289, 1290

Details are in the caption following the image
The different pathways of carbamylation and its roles in several diseases. (A) There are two main pathways for the production of isocyanic acid for carbamylation. (B) Representative protein substrates and roles of carbamylation in various diseases.

Protein carbamylation can alter protein structures, PPIs, and protein–cell interactions. By removing the positive charges on proteins, carbamylation can alter the protein's interaction with water and disrupt ionic interactions on the protein's surface. As these interactions are able to stabilize the secondary and tertiary structures of proteins, their absence may result in dramatic changes in protein conformation.1291 Changes in protein structure will affect its functions and biological activities and cause diseases in the body. Protein carbamylation alters the native structure of plasma proteins, which participates in the pathogenesis of diabetes.1292 In addition, there are also proteins that acquire new functions after carbamylation. Carbamylation of LL-37 reduces its bactericidal properties, and converts anti-inflammatory LL-37 to proinflammatory LL-37.1293 Moreover, carbamylation affects the assembly of homologous or heterologous protein monomers into fibers or filaments.1294 For example, actin or collagen cannot form intact filaments or fibers after carbamylation.1295

11.1 Carbamylation in aging

Studies have found that various physiological and pathological processes are related to protein carbamylation, such as aging, cataracts, CKD, atherosclerosis, RA, and neurological diseases (Figure 22B).1283 By measuring the changes in homocitrulline, a typical carbamylation derivative product (CDP), with age, carbamylation occurs throughout the life cycle and contributes to carbamylated protein accumulation in organs. The accumulation rate of CDPs is negatively correlated with lifespan, suggesting that longer-lived species may have efficient turnover, repair, or degradation systems to limit carbamylated protein accumulation in organs.1296 Modifications of salivary proteins increase with age, as evidenced by decreased total thiol levels and increased carbamylated proteins in the saliva of older adults.1297 Therefore, protein carbamylation may serve as a marker of mammalian aging.1285 In addition, protein carbamylation in peripheral blood is associated with age-related oxidative damage.1298 In elderly cataract patients, crystallins, including α-, β-, and γ-crystallin, can be modified by carbamylation, with seven lysine residues modified in α-crystallin.1299 An important function of α-crystallin in the lens is to ensure the activity of its chaperones, which limit protein aggregation and thus keep the lens transparent. However, α-crystallin carbamylation significantly affects its chaperone activity.1300

11.2 Carbamylation in kidney diseases

Carbamylated protein levels have been found to be significantly elevated in CKD (Figure 22B), and carbamylated albumin is considered an important biomarker for risk of death.1301 Carbamylated proteins may cause CKD complications such as atherosclerosis. Higher levels of carbamylated proteins are detected in the plasma of 75% of kidney-removed CKD mice.1302 C-Alb carbamylation levels are associated with higher mortality in diabetic patients with ESRD.1303 Carbamylated HDL but not carbamylated LDL in plasma is independently related to CKD progression in T2DM patients.1304 By investigating the levels of fibrinogen carbamylation in patients with renal disease and the effect of carbamylation on thrombin fibrinogen cleavage, fibrin polymerization and in vitro crosslinking, it has been found that although carbamylation itself does not affect thrombin cleavage, it alters fibrin polymerization kinetics and impairs cross-linking and clot degradation.1295

11.3 Carbamylation in atherosclerosis

Protein carbamylation is closely associated with atherosclerosis (Figure 22B).1305 “Uremic dyslipidemia” in CKD patients is characterized by normal low-density lipoprotein cholesterol, low high-density lipoprotein cholesterol, and high triglyceride plasma levels.1306 cLDL induces the prothrombotic effects of vascular cells and platelets by activating LOX-1 receptors and promotes thrombus formation.1307 An increased incidence of acute thrombosis has been observed in patients with CKD. In addition, the slight carbamylation of LDL leads to a decrease in the uptake of LDL by fibroblasts, which results in a lower clearance rate and prolongs the residence time of these particles in the blood circulation. This consequently increases the chance of further carbamylation of LDL. High cLDL leads to the accumulation of cholesteryl esters in macrophages.1308, 1309 Meanwhile, cLDL causes damage to vascular endothelial cells and induces abnormal proliferation of VSMCs, eventually leading to atherosclerosis.1310 Compared with cLDL, HDL has antiatherosclerotic properties and protective effects on the heart. However, carbamylation of HDL is involved in the occurrence and development of atherosclerotic CVD.1311 Carbamylated HDL engages in the formation of foam cells and becomes proatherosclerotic lipoproteins.1312 In addition, after carbamylation, vascular elastin fiber morphology and susceptibility to elastase degradation remain unchanged, but elastic fiber stiffness increases. These changes in the mechanical properties of the vascular wall may lead to aortic stiffness.1313

11.4 Carbamylation in immune diseases

In addition to ACPAs,1314 autoantibodies against carbamylated proteins (anti-CarP) have been detected in the serum of RA patients (Figure 22B).1315 Anti-CarP has important implications in the pathophysiology of RA and can be used to assess the risk level of RA patients. RA patients also develop autoantibodies against carbamylated NET (cNET) antigens, and the levels of these antibodies correlate with anti-CarP levels, making them a new biomarker for RA.1316, 1317 In addition, carbamylated histone-IgG immune complexes can promote osteoclast differentiation and enhance the matrix resorption of osteoclasts, suggesting that carbamylated proteins in NETs can increase pathogenic immune responses and bone destruction. This explains the link between anti-CarP levels and erosive arthritis in RA patients.1317

11.5 Carbamylation in neuropathy

Carbamylation has also been related to neuropathy (Figure 22B). The higher the level of carbamylation in the rat brain, the more severe the memory loss.1283 Carbamylation has also been implicated in AD,1318 and prevention of carbamylation may protect against cyanate-induced neuropathy.1319 Collectively, protein carbamylation is a potential biomarker for various human diseases and has important clinical significance.

12 REDOX MODIFICATIONS

As a main element of life, sulfur has multiple oxidation states (from −2 to +6) and participates in various redox reactions. Cysteine is a sulfur-containing amino acid in proteins, and its sulfur atom provides a wide range of chemical reactivity and structural flexibility for proteins.1320 Although the theoretical abundance of cysteine distribution in the human proteome is only 3.3%, as much as 22% of protein active sites are formed.1321 Redox is the main pathway for the regulation of cysteine functions, with two main classes of drivers.1322 The first is ROS/RNS/RSS generated by endogenous metabolism or exogenous stimuli, which mediate various redox modifications on cysteine, such as the oxidation of sulfhydryl (-SH) on the side chain of cysteine to sulfenylation (-SOH), which not only participates in the formation of intramolecular or intermolecular disulfide bonds (-SS-) or undergoes S-glutathionylation (-SSG) but can also be further oxidized to generate sulfinylation (-SO2H) and sulfonylation (-SO3H) modifications. The second driver is the complex and diverse reductase system in cells, which can reduce modifications other than sulfonylation. For instance, glutathionylation can be reduced by GRX/GST,1323 and sulfenylation can be specifically reduced by two isomerases, DsbG/C.1324 Currently, a number of oxidative PTMs (oxPTMs) on the thiols of cysteines, including SNO, sulfenylation, sulfinylation, sulfonylation, S-glutathionylation, and disulfide bonds, have been described (Figure 23A).

Details are in the caption following the image
Redox modifications of cysteine. (A) Different-types-of-redox modifications of cysteine. (B–E) The mechanisms of different types of redox modifications and associated physiological functions.

12.1 S-Nitrosylation

The formation of SNO is a reversible nonenzymatic catalyzed reaction in which NO is covalently bound to the thiols of cysteines to form SNO.1325 In general, there are three pathways for the synthesis of SNO (Figure 23B). NO reacts directly with the thiol group of the cysteine through NO+,1326 the SNO-modified small molecule or the proximal SNO-modified protein provides NO+,1327 or the thiol of cysteine can be activated to form a sulfur radical (S•), which then reacts with a NO radical (NO•) to generate SNO.1326 S-nitrosothiols are biologically stable reservoirs of NO,1328 have selective and transient modification properties and are excellent signal sensors.1329 For each protein, SNO can occur at a single cysteine or multiple cysteines.1330 There are more than 3000 proteins that are regulated by SNO and are involved in the regulation of protein stability, DNA damage repair, transcriptional regulation, cell growth, differentiation, and apoptosis (Figure 23B).1330 In addition, SNO also regulates the protein conformation of overlapping or nonoverlapping residues, PPIs, and other PTMs (e.g., phosphorylation, acetylation, ubiquitination, and disulfide linkage).1331 Dysregulation of SNO contributes to many diseases.1332-1334

12.2 S-Glutathionylation

Glutathione is covalently bound to reactive cysteines in proteins through disulfide bonds, termed S-glutathionylation.1335 S-Glutathionylation can occur via nucleophilic sulfur, where the thioanion (S-) reacts with oxidized glutathione (GSSG), or via the reaction between GSH and electrophilic sulfur intermediates such as sulfenic acid, S-nitrosothiol, and thiol radical (Figure 23C).1336 S-glutathionylation is involved in redox signaling and protects the thiols of cysteines from irreversible oxidation during oxidative stress. Many enzymes are known to catalyze glutathionylation/deglutathionylation reactions, including GSTP and GRX.1335 GSTP possesses both molecular chaperone and catalytic properties and controls the redox balance in the oxidative endoplasmic reticulum. It catalyzes the S-glutathionylation of target proteins and impacts the function of unfolded proteins.1337 GRX plays a crucial role in removing S-glutathionylation, preserving cellular redox homeostasis by regulating the S-glutathionylation of essential proteins such as phosphatases, kinases, and TFs.1338 Therefore, S-glutathionylation balance serves as a feature of normal cellular redox homeostasis.1335 Moreover, S-glutathionylation also regulates the activities of mitochondrial enzymes, heat shock proteins, TFs, and cytoskeletal proteins.1339 Hsp90 is a widely distributed molecular chaperone that interacts with a variety of proteins and regulates a variety of cellular processes. S-glutathionylation of Hsp90 results in inactivation of ATPase.1340 S-glutathionylation of C/EBPβ stabilizes the protein and increases its levels, promoting 3T3l1 cell differentiation.1341 Additionally, S-glutathionylation plays a role in regulating apoptosis. For example, S-glutathionylation of GADPH may transmit signals to the nucleus where GADPH trans-glutathionylates nuclear proteins such as Sirt1 to trigger apoptosis. GRX removes S-glutathionylation of GAPDH and prevents its nuclear transport.1342 FASL-induced activation of airway epithelial cell apoptosis is accompanied by an increase in protein S-glutathionylation.1343

12.3 Sulfenylation

Cysteine thiols are oxidized by hydrogen peroxide to produce sulfenic acids. Sulfenylation is a nonenzymatic modification that can also be converted from other oxidized forms of cysteine, such as SNO (Figure 23D).1326 Sulfenic acids have long been considered transient reaction intermediates formed by cysteine thiols under oxidative stress.1344 However, they have been discovered to play much more significant roles in cellular biology. These sulfenic acids serve not only as indicators of oxidation-sensitive cysteines and intermediate oxidation states, but also as key regulators of protein function, signal transduction, and initiators of disulfide bond formation.1345, 1346 For example, the formation of sulfenic acid is associated with hydrogen peroxide-mediated inactivation of PTPs, and the oxidation of cysteine thiol affects cellular PTP activity.1344 Sulfenylation increases the kinase activity of EGFR by oxidizing the active site Cys797.1347 The activity of Src is regulated by a redox-dependent mechanism, in which sulfenylation at Cys185 and Cys277 can enhance its activity.1348 Platelet CD36 signaling can promote hydrogen peroxide-mediated sulfenylation of Src family kinases, which is critical for oxLDL/CD36 proaggregation and procoagulant functions.1349 UCP1 is a protein required for thermogenesis in adipose tissue. Cys253 of UCP1 is sulfenylated during thermogenesis and affects sensitivity to UCP1-dependent thermogenesis.1350 The redox state of a single cysteine can alter biological processes in response to changes in cellular redox homeostasis.1351 For example, ROS-induced sulfonylation of C663 inhibits the UPR and stimulates the antioxidant response mediated by p38 MAPK signaling.1352 Reversible sulfenylation can also regulate the enzymatic activity of transcription and transduction factors, such as Mfn2, which undergoes sulfenylation after inflammatory stimulation and negatively regulates the transcriptional activity of β-catenin.1353 In addition, protein sulfenylation can also directly or indirectly affect many PTMs in cells, such as the oxidation of specific cysteines in PTPs, PTK, and cysteine proteases.1354

12.4 Sulfinylation and sulfonylation

Cysteine sulfinic and sulfonic acids are the peroxidation products of cysteine.1345 In the presence of excess oxidants, sulfenic acid can be oxidized to sulfinic acids or even sulfonic acids (Figure 23D). Sulfonic acids are the most oxidized form of cysteine. Although sulfinic acid was once thought to be an irreversible oxidative modification, the oxidation of the active site cysteine of Prx I or Prx II to sulfinic acids is reversible.1355 With the discovery of sulfiredoxin,1356, 1357 the regulation of sulfinylation in biology has also received attention. By using electrophilic diazene probes (DiaAlk), hundreds of previously unreported protein sulfinylation sites have been identified in mammalian cells.1358 Prx is an important class of human antioxidant enzymes that are present in high concentrations in human cells such as erythrocytes1359 and can rapidly oxidize sulfenic to sulfinic acids.1360 Under normal conditions, Prx utilizes the thiol/sulfenic acid oxidative cycle to detoxify hydrogen peroxide. However, under conditions of oxidative stress, the hydrogen peroxide concentration exceeds the reducing power of Prx, resulting in the formation of sulfinic acids. Cysteine residues in Prx sense the intracellular hydrogen peroxide concentration.1360 DJ-1 is a protein associated with hereditary Parkinson's syndrome. Sulfinylation at Cys106 regulates the protective function of DJ-1.1361 Cysteine residues usually exist in metal-binding motifs and form coordinate bonds with metal ions, such as zinc, copper, and iron. However, sulfinylation of these proteins can lead to the release of zinc ions and changes in protein conformation, which in turn alter protein function.1362

12.5 Disulfides

Disulfide bonds in proteins are a widespread cysteine modification that plays an important role in protein folding and stability. Disulfide bonds form intermolecularly or intramolecularly, depending on the accessibility and proximity to other cysteine groups. Disulfide bonds can be produced intracellularly through thiol-disulfide exchange, coupling between thiol radicals, and thiol reaction with nitrosylated cysteine or sulfenic acids (Figure 23E).1345 Thiol-disulfide exchange can be achieved by nonenzymatic or enzymatic reactions, such as Trx, Grx, and PDI, which can accelerate disulfide bond formation.1345 Disulfide bonds can introduce conformational constraints in peptides and proteins. Peptides containing disulfide bonds are expected to be used as drug leads or scaffold materials.1363 Hinge disulfide bonds in the human IgG2 CD40 antibody regulate receptor signaling by modulating conformation and flexibility.1364 Kinases are now also thought to be regulated by redox. The formation of intermolecular disulfide bonds between homodimers activates PKG1α and ATM, while intermolecular disulfide bonds between Src monomers inhibit kinase activity.1365 MLL1 is a redox-regulated HMT, and peroxide-induced intramolecular disulfide bond formation results in inactivation of the HMT SET-1/MLL1.1366 The protease activity of human ATG4B is also affected by reversible redox modification. A previous study found that C292 and C361 of ATG4B form a disulfide bond, which affects autophagy by regulating the activity of ATG4B.1367 Disulfide bonds also affect the subcellular localization of proteins, PPIs,1368 and the activity of cytokines. HMGB1 is a nuclear protein with extracellular inflammatory cytokine activity. HMGB1 requires mild oxidation to form a C23–C45 disulfide bond and unoxidized C106 to induce phosphorylation of the NF-κB p65 subunit and TNF-α production.1369 Oxidative stress induces the intermolecular disulfide bond formation of TFEB/TFE3 in mammals, which enhances the activity of the TF.1370 STING is a key regulator in the innate immune type I IFN pathway, and its C206 oxidation to form intermolecular disulfide bonds leads to a conformational change in the protein that prevents excessive activation of STING.1371 In addition, cellular structure is also dependent on cysteine disulfides, as microtubule assembly is partly mediated by disulfide bonds.1345

12.6 Redox modifications in health and diseases

Redox-associated PTMs are a part of normal cell signaling and can regulate the activity of a wide variety of proteins involved in energy metabolism, protein folding and degradation, and gene transcription. Imbalances in redox homeostasis are associated with aging1372 and various diseases, such as neurodegenerative diseases,1373 CVDs,1374 and cancers (Figure 24).1334

Details are in the caption following the image
Representative redox modification events in aging, metabolic disorders, neurological disorders, cancers, and CVDs are shown.

12.7 Redox modifications in aging

Aging-caused imbalance between RONS production and cellular antioxidant capacity may lead to oxidative stress. Proteins aberrantly modified by redox modifications become dysfunctional.1372 Protein cysteine redoxomics in various tissues of young and old mice has shown that many redox modifications in young tissues disappear in old tissues, but some new redox modifications also appear in old tissues.1375 Supplementation with antioxidants can neutralize ROS, and glutathione and its precursor N-acetylcysteine (NAC) are common dietary supplements. However, long-term use of GSH and NAC inhibits skn-1-mediated gene transcription and accelerates aging.1376 Endoplasmic reticulum sulfhydryl oxidase Ero1α produces SNO to reduce its activity, leading to reduced stress in the ER and compromised ER proteostasis and UPR, which in turn promotes cell senescence.1377 The KEAP1–NRF2 system is a key defense mechanism to prevent oxidative stress and aging.1378

Sarcopenia is a hallmark of human aging.1379 Muscle wasting is accompanied by decreased oxygen consumption and increased ROS production in sarcopenia.1380 The restoration of redox balance by elamipretide (SS-31) in aged mice can enhance mitochondrial function and improve skeletal muscle function. S-glutathionylation has been significantly reversed in aged mice, and the gastrocnemius muscle of SS-31-treated mice has greater fatigue resistance and mass.1381 S-glutathionylation is significantly increased in mitochondrial complexes I, II and V, ACO2, GAPDH, and MDH1 after rest and fatigue contraction, indicating that redox plays important roles in the control of muscle physiology, metabolism, and exercise adaptation response.1382

12.8 Redox modifications in metabolic disorders

The production and metabolism of active substances and the recovery and removal of the antioxidant system are highly synchronized processes in normal physiological conditions. When these systems are disturbed, a series of metabolic diseases, such as obesity, metabolic syndrome, and T2DM, can occur (Figure 24).1383, 1384 Excessive production of ROS in vivo can lead to oxidation of proteins, leading to decreased insulin secretion and increased insulin resistance, and contributing to the development of diabetes.1385 For example, ROS-induced oxidative stress can activate the IKKβ/NF-kβ pathway, leading to pancreatic β-cell dysfunction. Cys179 of IKKβ may be a vulnerable site to redox modification.1386 Oxidative modification of cysteine residues in Keap1 also affects the development of diabetes by regulating the Nrf2/Keap1/ARE pathway.1385

Increased intracellular NO production and reduced bioavailability are key factors leading to imbalances in redox homeostasis in metabolic diseases.1387 Protein SNO is important to the entire process of insulin action, including processing and secretion by pancreatic β cells, transport by endothelial cells, signal transduction, and degradation of insulin.1388, 1389 For example, glucokinase SNO at Cys371 dissociates glucokinase from ISG and facilitates its transition to the active conformation to increase insulin secretion.1390

Obesity is a metabolic disorder affected by oxidative stress.1391, 1392 SNO of IRE1α in obese mice leads to a decrease in glucose homeostasis.1393 ROS can activate thermogenesis through direct redox modification of UCP1. UCP1 Cys253 in brown and beige adipose tissues undergoes sulfenylation during thermogenesis, and the heat generated by these adipocytes can fight obesity and diabetes.1394, 1395

In addition to the influence of ROS/RNS, antioxidant enzymes also play a role in metabolic diseases. Numerous studies have demonstrated that the activity of antioxidant enzymes can be used as potential biomarkers of metabolic diseases. For example, elevated levels of glutathione S-transferase have been observed in the blood plasma of individuals with T2DM.1396 The activities of enzymes such as superoxide dismutases, catalases, and glutathione peroxidases in PBMCs of obese patients are found to be significantly lower.1397 In addition, the regulation of the Trx/trxR system has been identified as a potential target for treating metabolic syndrome, T2DM and hypertension.1398, 1399, 14

12.9 Redox modifications in the cardiovascular system

The heart is one of the organs most severely affected by SNO, which plays a key role in regulating redox homeostasis in the stressed heart.1400 In mouse cardiac proteins, a total of 1974 SNO sites from 761 proteins have been identified. In the cardiovascular system, NO signaling regulates vasodilation and myocardial contraction, and SNO of proteins may represent the third functional dimension of NO signaling in the cardiovascular system to ensure optimal cardiac function.1401 SNO at Cys589 of Hsp90 promotes cardiac hypertrophy.1402 Moreover, the SNO levels of MLP are significantly elevated in patients with hypertrophic myocardium and in spontaneously hypertensive rats and mice with transverse aortic constriction.1403 Cardiac fibrosis is an irreversible pathological process, and inhibition of SNO of Hsp90 can alleviate myocardial fibrosis through the TGFβ/SMAD3 signaling pathway (Figure 24).1404 SNO at C116/C163 of JNK accelerates cardiac fibrosis.1405 SNO of Hsp90 affects cardiac hypertrophy and myocardial fibrosis. A recent study found that SNO of C521 of Hsp90 can inhibit the interaction between Hsp90 and AHA1, promote the recruitment of CDC37, and aggravate atherosclerosis.1406 In addition, SNO mediates the coupling of GNAI2 to CXCR5, activating YAP-dependent endothelial inflammation to drive diabetes-accelerated atherosclerosis.1407 SNO is also involved in calcific aortic valve diseases. USP9X SNO is reduced in calcified human aortic valves. SNO of USP9X can stabilize MIB1, which activates the NOTCH1 signaling pathway in adjacent cells to prevent calcification.1408 In addition to SNO, protein S-glutathionylation also plays a role in CVDs. During the development of calcific aortic valve stenosis, abnormal S-glutathionylation promotes tissue phenotypic switching in the aortic valve, eventually leading to calcium deposition.1409

12.10 Redox modifications in neurological disorders

Redox homeostasis is linked to neurological disorders. ALS is a motor neuron disease. The G93A mutation in the antioxidant enzyme SOD1 is a gain-of-function mutation that causes SOD1 aggregation and motor neuron degeneration. Aggregation of SOD1 is associated with the oxidation of cysteine residues on SOD1 and increases when Cys6 and Cys111 are oxidized (Figure 24).1410, 1411 The pathogenesis of AD is related to Aβ, neuronal hyperexcitability, and aging-related neuroinflammation. Excessive NO production leads to aberrant SNO of various proteins.1412 Moreover, noncanonical transnitrosylation can lead to synaptic loss, which may be one of the pathological causes of cognitive decline in AD patients.1412 SNO also affects the occurrence and progression of PD. HMGB1 is a DNA-binding protein that regulates gene transcription and genome stability in the nucleus. SNO in C106 of HMGB1 can promote the secretion of HMGB1 and enhance its binding force with microglial Mac1, which is one of the key mechanisms for the development of PD.1413 In addition, SNO of foreign α-synuclein also promotes Parkinopathy.1414

12.11 Redox modifications in cancers

ROS-induced oxidative stress is a fundamental feature of cancer. Tumor cells produce more ROS than normal cells, resulting in abnormal redox homeostasis.1415 Oxidative stress can regulate intracellular signaling pathways and promote the growth of tumor cells, while excessive oxidative stress may lead to oxidative damage and even tumor cell death.1416 Tumor cells can rely on their own powerful antioxidant systems to withstand oxidative stress with high levels of ROS at different stages. For example, in the early stages of tumors, they can adapt to oxidative stress by activating antioxidant TFs or increasing NADPH through the PPP. During tumor growth and metastasis, tumor cells can activate AMPK, PPP, and reductive glutamine to increase NADPH, allowing cells to survive under conditions of high ROS.1417 During the whole process, the redox modifications of proteins in tumors play a crucial role (Figure 24). For example, SNO at C277 of VEGFD inhibits its expression and promotes the occurrence and development of LUAD.1418 In early CRC, PTPS is highly expressed and phosphorylated at Thr58 under hypoxic conditions, which promotes binding to LTBP1 and drives LTBP1 SNO, thereby maintaining tumor cell growth under hypoxic conditions.1419 Tumor cells can resist high levels of ROS by activating the activity of antioxidant TFs such as Nrf2. Nrf2 deletion affects the redox proteome of PC and NCSLC, which has impacts on mRNA translation machinery and the glycolytic pathway.1420, 1421 ROS can also affect tumor cells by regulating the functions of metabolic enzymes through redox modification. For example, TPβ is a redox-sensitive protein and a key rate-limiting enzyme in the FAO process. Under the stimulation of ROS, C458 of TPβ will be oxidized and inactivated. In oxidative stress caused by glucose starvation, Nur77 can enter the mitochondria and be oxidized by ROS, protecting C458 of TPβ from oxidation, resulting in the production of FAO-mediated NADPH to relieve intracellular oxidative stress and promote melanoma cell survival and metastasis. This suggests that Nur77 may be a potential target for the treatment of melanoma.1422 In the glycolysis pathway, PFKM is one of the most important regulatory enzymes, and SNO at C351 of PFKM can stabilize the tetramer of PFKM, which contributes to the metabolic reprogramming of ovarian cancer cells.1423 Additionally, elevated expression of Nm23-H1 is linked to a favorable prognosis in patients with breast cancer, and activation of Nm23-H1 through redox regulation can inhibit breast cancer metastasis.1424 PTEN is a tumor suppressor. The oxidative modification of the cysteine of PTEN makes it inactive. The recovery of its activity mainly depends on the availability of Trx and Prx.1425 Methionine is an amino acid residue prone to redox modifications that are regulated by MSRA. MSRA deletion promotes the oxidation of methionine on PKM2, which in turn promotes mitochondrial respiration and cell metastasis.1426

13 OTHER MODIFICATIONS

13.1 ADP-ribosylation

ADP-ribosylation is a process in which ADP-ribosyl transferases transfer ADP-ribosyl to the target protein's ADP-ribose binding domain using NAD+ as the substrate.1427 Currently, over 800 proteins have been found to contain the ADP-ribose binding domain.1428 The structure and function of many proteins, including nuclear proteins topoisomerase I, DNA ligase II, endonuclease, histones H1, H2B and H4, DNA polymerases and cytoplasmic proteins adenyl cyclase and elongation factor eEF-2, are regulated by ADP-ribosylation.1429 ADP-ribosylation is involved in numerous physiological and pathological processes, such as signal transduction, protein transport, transcription, DNA damage repair, cell cycle regulation, apoptosis, and necrosis.1427 Poly-ADP-ribosylation (PAR) of proteins may decease during aging because the activity of poly-ADP-ribose polymerase (PARP) in senescent human fibroblasts is reduced with donor age and continuous passage in vitro.1429 PAR modification plays an important role in DNA damage repair,1430 and many DNA damage repair factors can recognize PAR signals, leading to the rapid recruitment of these factors for efficient repair.1431 Loss of PAR modification inhibits SSB and DSB repair.1432 Therefore, PARP inhibitors (PARPi), such as olaparib, rucaparib, niraparib, and talazoparib, have been developed as a class of targeted drugs for cancer treatment.1433 These drugs compete with PARP1/2 for intracellular NAD+ and inhibit its catalytic activity to block DNA damage repair signals. Additionally, PARP1/2 can become trapped in damaged DNA, forming a PARP–DNA complex and blocking its release, causing replication fork stagnation and leading to cancer cell death.1434, 1435

13.2 Benzoylation

Lysine benzoylation (Kbz) was the first discovered aromatic fatty acid modification and primarily occurs in the N-terminal tails of histones.1436 Sodium benzoate (SB) is a widely used food additive and a clinical treatment for hyperammonemia.1437 It can be converted into benzoyl-CoA in mammalian cells and as a precursor of Kbz.1438 Moreover, benzoyl-CoA is also a central intermediate for the degradation of numerous aromatic growth substrates in bacteria and gut microbes.1439 Previous studies have showed that HBO1 and the Spt-Ada-Gcn5 acetyltransferase complex are writers of histone Kbz.1440, 1441 In vivo studies have found that SIRT2, unlike other sirtuins or histone deacetylases (HDACs), acts as an eraser of histone Kbz.1436 NAD+-dependent histone deacetylase Hst2 has debenzoylase activity in yeast.1440 Human DPF and YEATS but not BRD domains are the readers for histone benzoylation.1442

Histone Kbz is a mark enriched in gene promoters and is associated with gene expression.1436 Furthermore, it has a different physiological relevance than histone acetylation. Kbz primarily targets gene promoters and regulates glycerophospholipid metabolism, ovarian steroid synthesis, and hydrolytic phospholipase signaling pathways.1438 Excessive intake of SB increases Kbz levels, leading to motor coordination disorders and increased risk of diseases such as ADHD.1443

13.3 Neddylation

Neddylation is a PTM that can conjugate the Ub-like protein NEDD8 to target proteins.1444 Neddylation has a broad range of functions and can alter various aspects of protein function, including protein conformation, stability, subcellular localization, affinity for DNA, and binding of protein substrates.1445, 1446 NEDD8 is a highly conserved protein composed of 81 amino acids. Of all the Ub-like protein families, NEDD8 is the molecule with the highest sequence and structural similarity to Ub.1444 After passing through the E1–E2–E3 cascade, NEDD8 covalently binds to the substrate and undergoes Neddylation on the lysine of the substrate protein clock, thereby regulating the biological function of the substrate.1444

NEDD8 activating enzyme E1 is a heterodimer composed of APPBP1/UBA3.1447 Through its ATP-dependent catalytic subunit, it catalyzes the formation of a high-energy sulfolipid bond at the C-terminal glycine of NEDD8 molecule and the cystine active site of UBA3 to activate the NEDD8 molecule.1448 The NEDD8-loaded NEDD8-activating enzyme (NAE) is transferred to E2-conjugating enzymes UBC12/UBE2M or UBE2F via a trans-thiolation reaction. Ultimately, substrate specific E3 ligases transfer NEDD8 from E2 to lysine residues in their target proteins by covalent attachment. NEDD8 ligase E3 can be divided into RING finger and HECT types.1449 Many neddylation E3s have been discovered, including RBX1/2, ROC1/2, SAG, c-CBL, MDM2, FBXO11, c-CBL, DCNL1–5, IAPs, RNF111, TFB3, and TRIM40.1448, 1449 Neddylation is a reversible process, and under the action of the de-neddylation enzyme, NEDD8 can be dissociated from the substrate protein.1450

Similar to ubiquitination, the neddylation process also affects cell growth in multiple aspects such as proliferation, apoptosis, and migration in tumor cells. Many recent studies have shown that NEDD8 or enzymes related to the neddylation pathway are overexpressed in various cancers, including lung cancer,1451 liver cancer,1452 CRC,1449 and esophageal squamous cell carcinoma.1453 At present, the primary approach to studying the neddylation pathway involves interfering with the expression of particular genes through siRNA or inhibiting the activation enzyme NAE using small molecule inhibitors such as MLN4924 which deactivate the entire neddylation pathway.1454 The formation of a stable covalent bond with NEDD8 and competitive inhibition of NAE's activation of NEDD8 leads to inhibition of the neddylation pathway and partially blocks the ubiquitination pathway dependent on CRL. This inhibition results in the accumulation of substrates, triggering various cellular responses, such as cell cycle arrest, apoptosis, aging, and autophagy. By utilizing this mechanism, it is possible to achieve a therapeutic effect for treating tumors.1455 In addition, the dysregulation of neddylation is involved in the development of various diseases such as neurodegenerative diseases,1456 inflammation,1457 and CVDs,1445 suggesting that it can be a potential target for disease treatment. In fact, the NAE inhibitor MLN4924 not only has significant antitumor activity, but also has activities such as antiviral1458 and anti-inflammatory effects.1459 Collectively, neddylation is a versatile PTM that can modulate a wide range of protein functions, with implications for many physiological and pathological processes. The neddylation pathway and its inhibitors hold significant potential for research into various disease treatments.

14 PERSPECTIVE

Protein modification signaling pathways play important roles in physiological and pathological processes.10 Compared with genomics, proteomics can instantly reflect the profiles of proteins and protein modifications of individuals in the disease state and normal state and has more prospects for diagnosis and treatment in the clinic.1460, 1461 Deciphering the complex biological functions of proteins requires a deep understanding of protein modification, referred to as the “PTM code.”1462-1464 In recent years, the rapid development of MS-based proteomics technology has greatly advanced the research progress of protein PTMs. Despite these advances, high-throughput characterization of protein modifications remains a challenging task.1465-1467 First, there is a lack of effective antibodies and reagents to enrich the modified peptides of some modification types, such as methylation.1468 Second, large-scale PTMomics analysis of clinical samples can be difficult due to the limited availability of samples from clinical cohorts.1469 Third, the dynamic range of the human proteome exceeds several orders of magnitude, exacerbating the identification of low-abundance modified proteins.1470 Fourth, complete structural characterization of proteins requires more substance and analysis time than “simple” identification based on some peptide fragments.1470 Finally, the modification groups on some proteins such as antibodies, often have poor chemical stability and low abundance, which also puts forward higher requirements for more sensitive MS detection.1471

When a protein PTM performs its function, it is often not the PTM at a single site that does it. In fact, crosstalk between protein modifications is common.1472-1474 It includes not only the crosstalk of different modifications of the same protein at the same site but also the crosstalk of different modification sites of the same protein.1475, 1476 In addition, there is also extensive crosstalk between modifications on different proteins.1477 However, research on this PTM crosstalk is still in its infancy. The development of top-down proteomics technology, high-throughput PTM omics technology and novel PTM crosstalk analysis algorithm will provide the most powerful help to reveal the crosstalk of PTMs.1478-1480

An increasing number of studies have demonstrated the close relationship between protein modifications and diseases.1465, 1481, 1482 Thanks to the rapid development of simple and efficient phosphorylation enrichment technology, the phosphoproteomes of some diseases, such as tumors and neurodegenerative diseases, have been revealed.1483-1485 Some phosphosites are closely related to the occurrence and development of these diseases. In addition, preliminary progress has also been made in the acetylome, ubiquitinome and glycosylome in clinical samples.140, 141, 148, 1486-1489, 161 However, studies of the relationship between many other types of protein modifications and diseases in large clinical cohorts remain lacking. The establishment of a multiomics molecular network based on the genome, proteome, PTMome, and metabolome will provide new breakthroughs for the discovery of new biomarkers and drug targets.1490, 1491 In addition, some irreversible protein modifications, such as citrullination and carbamylation, may generate neoantigens.1492-1494 Uncovering the roles of these neoantigens in diseases will also be the focus of future research.

Although more than 650 types of protein modifications have been identified (http://www.uniprot.org/docs/ptmlist.txt), we believe that with the rapid development of new technologies and algorithms, the number of novel PTMs may still increase, which will help to elucidate the code of life.

AUTHOR CONTRIBUTIONS

L.D. organized the team and revised the manuscript. Q.Z., X.X., Y.Q. Z.X., C.C., B.C. X.Z., S.H., S.L., and Z.A. drafted the manuscript and participated in the discussion. All authors have read and approved the final manuscript.

ACKNOWLEDGMENTS

The authors thank the National Key R&D Program of China, National Natural Science Foundation of China and West China Hospital for financial supports. All the images and elements used in the manuscript are original and there are no copyright issues. This work was supported by the National Key R&D Program of China (No. 2018YFC2000305 and 2022YFA1303202), National Natural Science Foundation of China (No. 82073221, 31870826, and 82073246), Science and Technology Project of Sichuan Province (No. 2021YFS0134 and No. 2023YFS0247), National Clinical Research Center for Geriatrics of West China Hospital (No. Z20201002 and Z2021JC005), and West China Hospital 135 project (No. ZYYC20007).

    CONFLICT OF INTEREST STATEMENT

    The authors declare no competing interests.

    ETHICS STATEMENT

    No ethical approval was needed.

    DATA AVAILABILITY STATEMENT

    No additional data are included.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.