2.1 DNA replication initiation

Mini-chromosome maintenance (MCM) proteins are composed of six subtypes, MCM2, MCM3, MCM4, MCM5, MCM6, and MCM7. All subunits integrate into a hetero–hexameric complex and act as a replicative DNA helicase to unwind the parental DNA.¹² Beyond that, other MCM proteins, MCM8, MCM9, and MCM10, are reportedly essential for the DNA replication and remaining genome maintenance.^13-15

In eukaryotic cells, activation of replication origins is a prerequisite of DNA replication, manifesting the bidirectional movement.^{16, 17} The prereplication complex (pre-RC) forms at the origin recognition complex (ORC1-6), which serves as the actuated operator of DNA replication. In order to maintain genome integrity, ORC proteins are essential for establishing pre-RC at origins since the distribution and density of origins have to be adequate to replicate the entire genome without leaving any regions un-replicated. In the early G1 phase, cell division cycle 6 (CDC6) and DNA replication factor 1 (Cdt1) are recruited to the replication origins, subsequently attracting MCM2-7 complex to load onto chromatin.^{18, 19} MCM2-7 hexamer itself has restricted helicase activity, while executing integrated helicase activity in combination with Cdc45 and GINS (CMG) during G1/S transition.^{20, 21} Those proteins could compose of preinitiation complex (pre-IC), which then preparing to form bi-directional replication forks once MCM double hexamers separating into two single units.²² Additional factor, MCM10, collaborates with polymerase ε (polε) and polymerase δ (polδ) in replication origins for replication initiation, meanwhile interacting with CMG helicase to stabilize the replisome.^{23, 24} Moreover, one recent study found that MCM10 is necessary for CMG to transit between double-strand DNA (dsDNA) and single-strand DNA (ssDNA). Additionally, MCM10 migrates along with the replication fork and energizes replication elongation (Figure 1A).

In light of the bulky size of genomes and abundant amounts of chromosomes, eukaryotic cells contain numerous replication origins to duplicate their genomes. Nonetheless, massive replication origins assist high-efficiency genetic information transfer with more hazard since their distribution and proceeding have to be tightly controlled. Eukaryotic cells have an extensive system to guarantee precise DNA replication. During the S phase, disparate genome regions or domains are duplicated at a staggered time, and the origin licensing system is carried out from firing at distinct cell cycle phases. In addition, in the G1 phase, license origins are much more than they used in the subsequent S phase, while the inactive origins are named as “dormant origins.” The plain fact is that dormant origins constitute the tremendous majority of licensed origins, which serve as backup to sustain the replication fork regular progression under conditions of RS.

Since overabundance and distribution patterns on chromatin, the definite and accessional functions of MCM proteins are always contradictory. The issue is defined as “MCM paradox,” which is chiefly embodied in two aspects: (1) MCM2-7 complexes massively exist in nonreplicated DNA; (2) excess MCM hetero–hexamers attach to chromatin instead of replication origins and ORCs.^{25, 26} Apparently, the excess MCMs are involved in other biological processes. Numerous studies proved that MCMs serve as biomarkers in multiple tumors, which are closely related to tumorigenesis, development, and even in tumor therapeutic.

2.2 DNA replication elongation

Afterward, dozens of distinct proteins consistently coordinate to promote DNA replication elongation. Owing to the DNA antiparallel structure and DNA polymerases’ 3′–5′ direction of forward motion, the running replication forks separate into two single strands, which are continuously synthesized leading strand and inconsecutive synthesized lagging strand, respectively.²⁷ In the lagging strand, discontinuous and short fragments are considered as Okazaki fragments, which require DNA ligase to assemble into the complete lagging strand rapidly and ultimately.²⁸

DNA elongation and polymerization is catalyzed by multifarious enzymes, which are responsible for DNA synthesis and progression of the DNA replication. Polymerase α (polα)/primase mainly partakes in the initial stage of DNA synthesis.²⁹ Four subunit enzymes of polα/primase catalyze RNA oligonucleotide synthesis, which subsequently can be applied to extend by a short stretch of DNA. After this initiation step, polα is immediately switched into replicative polymerase via an ATP-dependent manner.³⁰ Two replicative polymerases, polε and polδ, principally execute DNA synthesis to elongate the nascent leading strand and lagging strand, respectively. Both polε and polδ are four subunit enzymes with intrinsic 3′–5′exonuclease proofreading activities, which increase replication fidelity with a lower mutation rate.^{31, 32} Moreover, multiple evidence suggested that polymerase activities of polδ are stimulated by protein proliferating cell nuclear antigen (PCNA), serving as a platform to coordinate numerous proteins interaction at the replication fork. Polδ cooperates with PCNA to promote long stretches of DNA synthesis.³³ Nevertheless, PNCA could not load onto DNA without replication factor C (RFC) assistance, which could wrap PCNA homo-trimeric ring to promote its DNA loading via ATP-dependent manner.³⁴ Additionally, flap endonuclease1 (FEN1) and Dna2, two endonucleases, are mainly needed for DNA and RNA flap structure cleaving, which are mediated by replication protein A (RPA).³⁵ Ultimately, flap cleavage generated DNA nick is sealed by DNA ligase I (Figure 1B).³⁶

2.3 DNA replication termination

In contrast to the initiation and elongation steps, DNA replication termination still remains several queries, even though it occurs on neighboring replication origins encounter. Due to the torsional strain caused by DNA helicase, positive supercoils structure must be removed by DNA topoisomerases to maintain the replication fork progression and genomic integrity. Type I and type II topoisomerases unwrap supercoils primarily to rotate the direction of fork evolution into clockwise, which could transfer the topological stress. Additionally, type II topoisomerase specifically removes precatenanes to assure converging replisomes unwind and DNA complete synthesis.^{37, 38}

Like the ultraprecise instrument, every module of DNA replication all links with one another. During the S phase, reloading of MCMs is inhibited to ensure that no genome segment is re-replicated to preserve genome integrity.³⁹ Except for MCM proteins, several proteins are involved in this node. Cdt1, as the component of pre-RC, is the prime modulator to prevent re-replication.⁴⁰ CDK-mediated phosphorylation of Cdt1 is inhibited from interacting with Orc6 once the DNA replication initiation in Saccharomyces cerevisiae.⁴¹ In eukaryotes, however, Cdt1 is degraded upon S phase entry through two independent ubiquitin-mediated pathways.^{42, 43} In addition, other components of pre-RC, ORC1-6, are also the critical point for preventing re-replication.⁴⁴ During S phase entry, ORC1 is released from replicating sequence via CDK-mediated manner, which prevents ORC from entering second round of licensing.⁴⁵ CDC6 also manifests in preventing re-replication via a distinct mechanism. Some studies support that the SCF^{cyclin F} ubiquitin ligase complex impedes DNA re-replication by proteasomal degradation of CDC6 in the cell cycle.⁴⁶

In eukaryotes, neighboring CMG complexes meet each other on different strands, which is propitious to stable and orderly replication progression. Leading and lagging strand separation promotes CMG of one replisome to straightway transfer into the lagging strand template. Rapid encounter of adjacent CMG complex without pausing could preserve genome stability, whereas suspending for a while when CMG complex confronts a covalent DNA–protein.⁴⁷ This observation points out that cells set the defense mechanism to prevent conflict between adjacent CMG complexes during DNA replication termination.

CMG removal from the strand is a remarkable event in eukaryotic replication termination when CMG cooperated with dsDNA. Previous reports suggested that converging CMG complexes proceed migration along the leading strand template until the downstream Okazaki fragment, which no longer performs dsDNA unwinding at all. Ultimately, CRL2^Lrr1-mediated MCM7 polyubiquitination leads to CMG unloading, subsequently removed by CDC48/p97 segregase (Figure 1C).⁴⁸

DNA replication is an intricate process with a coordinated interplay of multiple proteins. As we summarized, each step of DNA replication must be strictly regulated to preserve genome integrity, while internal or external DNA-damage agent always threatens DNA replication to activate DDR system. Meanwhile, dysfunction of DNA replication and DDR causes severe diseases, which highlights the role of DNA replication in tumorigenesis and development.

3.1 CMG complex

In eukaryotes, DNA replicative helicase CMG complex binds to dsDNA at replication origins, subsequently transfers to ssDNA for DNA unwinding. As we described above, Cdc45 and GINS cooperate with MCM2-7 during S phase entry, forming CMG helicase for bidirectional replication forks (Figure 2A).⁴⁹

MCM proteins were firstly identified in S. cerevisiae, which were deemed to MCM. Based on electron microscopy investigations, each MCM monomer involves two conserved main domains exercising respective functions.^{50, 51} MCM2-7 complexes could motion via the nuclear localization signals on N-terminal region of MCM2 and the C-terminal of MCM3, whereas nuclear export signals on the central part of MCM3.⁵² In contrast, recent data suggest that MCM complexes are shuttled in interphase cells basically relying on the nuclear export signals in MCM3.⁵³

The N-terminal domain (MCM_N) possesses three consistent crystal structure subdomains: an OB (oligonucleotide/oligosaccharide binding)-fold, a peripheral helical bundle, and usually a zinc-binding motif (Figure 2B).^{54, 55}

The OB-fold subdomain links two single hexamers as head-to-head form and is also for DNA binding.⁵¹ Mcm4^Chao3 (chromosome aberrations occurring spontaneously 3) mutation occurring in mouse Mcm4 OB fraction disrupts routine DNA binding process, resulting in genomic instability.^{56, 57}

The peripheral helical bundle interacts with the OB fraction via a slight linker promoting the interaction with DNA.⁵⁸ The observation indicated that a helical bundle might be essential for protein–protein interplay and protein–DNA interactions during the initiation step. However, molecular mechanics studies presented that deleting a helical bundle exerts a limited influence on MCM function.^{59, 60}

The X-ray crystal structures of MCM_N suggested a zinc-binding domain, which presented two conserved arginine residues in Pyrococcus furiosus MCM (pfMCM). Studies with the pfMCM verified that the zinc-binding domain is probably needed for ssDNA binding.⁵⁴ Mutation of these two conserved arginine residues in MCM4/6//7 interfered with the loading of MCM2-7 complex onto DNA, further resulting in growth defect in S. cerevisiae. These findings suggested that zinc-binding domain of MCM4/6/7 is the vital region in ssDNA binding and origin melting.^{61, 62} In eukaryotes, the zinc-binding motif of MCM3 lacking a prominent motif impacts the MCM2-7 complex original function, suggesting that zinc-binding motifs play a vital role in MCM2-7 activities.⁶³

MCM proteins contain an AAA⁺ ATPase domain in C-terminal with two subunits terming as Walker A and Walker B, which are integrant for ATP hydrolysis and ATP binding.^{64, 65} Mutation of nearly any residues of the MCM AAA⁺ ATPase domain eradicates ATPase activity.⁶⁶ Despite all the MCMs harboring ATP-binding motifs at the intersubunit interfaces, the ATP-binding mode is quite different.⁶⁷ MCM4/6/7 proteins exhibit distinct functions when the ATP binding sites undergo mutations.⁶⁸ It should also be mentioned that MCM7 is required for ATP hydrolysis and DNA helicase via its ATP binding motif.⁶⁹

Nuclear magnetic resonance (NMR) structure studies revealed that the C-terminus of MCMs comprises a winged helix (WH) domain. Furthermore, the WH domain connecting to the AAA⁺ ATPase domain exhibits ATPase activity, promoting the domain shift via a flexible linker with the protein core.^{70, 71} In contrast, archaeal MCM exhibits increasing ATPase activity and dsDNA unwinding activity when partial deleting of WH domain.⁷² Thus, WH domain may reserve the latent function during dsDNA unwinding and may take effect in initiating helicase activity.⁷³

Except for conserved MCM2 and MCM3, MCM8 and MCM9 also possess a nuclear localization signal to shuttle between cytoplasm and nuclear.⁷⁴ Some studies indicated that MCM2 and MCM3 distribute in the cytoplasm but temporally and spatially shift to nucleus in a cell cycle-dependent manner.⁵² However, the distribution of MCM2 may also be associated with DNA damage. Envelope protein gp70 directly recognized MCM2 nuclear localization signal in the cytoplasm, thus enhancing DNA damage-induced apoptosis.^{75, 76} However, limited researches discuss the purpose of the MCM proteins motion.

The eukaryotic GINS complex consists of four subunits, Sld5, Psf1, Psf2, and Psf3, pronounced as Sld-go, Psf-ichi, Psf-ni, and Psf-san in Japanese. Despite its central role in CMG complex, GINS also modulates massive protein interaction during DNA replication and DNA repair.⁷⁷ Each subunit of GINS interacts with each other extensively, meanwhile, each of them possesses the related two-domain (A-domain: α-helical region; B-domain: β-rich region) structure, whose structural similarity causes pseudo-twofold symmetry in whole GINS architecture (Figure 2C).⁷⁸

In eukaryotic GINS, Psf1 only has an intact A-domain, yet B-domain is invisible in the crystal lattice, even though the similar B-domain of Psf1 to the three other subunits via sequence alignment. Some reports indicated that the complementarity of the B-domain into Psf1 disturbs GINS packing, which implies an essential role in CMG formation and Cdc45 binding.⁷⁹ However, the Psf3 B-domain is widely considered to interact with the MCM complex, strengthening the MCM3–MCM5 interface.⁷⁸

In the CMG complex, Cdc45 cooperates with the MCM2-7 complex to shut down the MCM2–MCM5 gate, which is crucial for ATPase site forming and CMG translocation on ssDNA. Cdc45 possesses a distinct helical motif, which is proximal to the catalytically active domain of polε. N-terminus of polε crosslinks with Cdc45 on the tip of the protrusive helix of Cdc45, indicating Cdc45 impacts on CMG helicase and polε polymerase activity.⁸⁰

3.2 PCNA and its binding proteins/enzymes

Eukaryotic sliding clamp protein, PCNA, is a ring-shaped homo-trimer with each subunit containing two domains, which presents a pseudo-six-fold symmetry pattern. Each subunit of eukaryotic PCNA is formed from two independent and semblable folded domains, which is ultimately confirmed by X-ray crystal structure analysis. PCNA could be roughly separated into two domains, domain A and domain B, connected by an extended β sheet across the interdomain frontier. Moreover, a flexible linker concatenates two domains are named the interdomain connector loop. The assembled pattern among three subunits performs end to end structure, precisely as one domain A connects with the adjacent subunit's domain B (Figure 2D).^{81, 82}

Due to its essential role in DNA replication, PCNA embraces the DNA and travels along it, conducting for DNA polymerases and DNA replication proteins. DNA could cooperate with three equivalent sites of PCNA since its symmetry patter. PCNA sliding along DNA counts on its basic residue interactions with the phosphates of DNA, which promoting the rotation of PCNA around the DNA. One convincing model supports that PTMs of PCNA alter its positive charges on the inner side, leading to unconscionable movement. Thermal and chemical denaturation researches demonstrated that human PCNA is much more unstable than S. cerevisiae homolog though they share the homogeneous three-dimensional structure. Furthermore, human PCNA performs tough backbone dynamics, especially at helix of ring inner surface. Due to the highly dynamic and plastic property, PCNA evolves as platform to facilitate interacting with multiple proteins.^{83, 84}

A huge collaborative network of proteins engages for high fidelity DNA repair and accurate DNA damage repair. PCNA is regarded as entire hub in DNA replication that interacts with abundant proteins involved in multiple DNA-related processes. By occasion of homo-trimer shape of PCNA, three identical pockets could cooperate distinct partners simultaneously and coordinate various functions spatiotemporally. Numerous PCNA-interacting proteins (PIP) interact with PCNA via their PIP box. A typical consensus amino acid sequence of PIP motif is (Q-x-x-(I/L/M)-x-x-(F/Y)-(F/Y)).⁸³

The PCNA ring has three independent PIP-box binding sites with three distinct ligands for binding proteins. To secure normal replication, three promoters of the PCNA trimer convene DNA ligase I, polδ, and FEN1 simultaneously to ensure stable Okazaki fragment synthesis. Constitutive complex has been demonstrated in yeast called the “PCNA tool belt,” which could be modulated by diverse PTMs. FEN1 interacts with PCNA via its canonical PIP box exhibiting lower affinity, while increasing the affinity by replenishing 20-residue long PIP fragments.⁸¹ These observations indicate that PIP box of proteins mediates their interaction affinity to PCNA, which is also modulated by PTMs. Thus, targeting such a binding site may interfere with DNA replication and DNA damage repair, thereby serving as attractive targets for cancer therapy.

Indispensable DNA replicative polymerases are required for DNA synthesis with high efficiency and accuracy. The general architectures of DNA polymerases present right-hand aspect with three main functional domains, which also contain exonuclease activity site for proofreading. Eukaryotic DNA replication primarily depends on three B-family DNA polymerases: polα, polδ, and polε. Polε and polδ are chiefly for high accurate DNA synthesis on the leading and lagging stands via interaction with PCNA, respectively. All eukaryotic replicative polymerases contain two conserved motifs with cysteines (CysA and CysB), which was regard as Zn-finger motifs originally. Except for detectable PIP box sequence in polδ, CysA motif could also directly interact with PCNA to promote efficient loading and synthesis of DNA.³¹ However, little is known about how pol ε with PCNA in mammals.

Insight into the general architecture of the replication proteins assists us in clarifying more accurate molecular regulatory mechanisms. Due to the complex interaction network, spontaneous or revulsive mutations of MCM2-7 complex disturb the normal biological processes such as DNA replication, cell proliferation, and DDR.^{57, 85} Moreover, MCMs load onto DNA via particular binding domains, thus it is possible to interrupt the chromosome remodeling through interfering these specific domains.^{26, 86} PTMs of replication proteins in diverse residues distributed in different domains may present a special effect due to their topological alteration in positive or negative patterns. Nevertheless, the precise regulatory for how PTMs in different domain affecting downstream processes is still unclear.

5.1 MCMs phosphorylation

Individual MCMs are subject to phosphorylation in a cell cycle-specific manner, which may be consistent with their cell cycle-specific functions. Due to various kinase types, phosphorylation of MCM proteins undergoes distinct regulatory mechanisms. Moreover, exceptional phosphorylation of MCMs might disrupt DNA replication progression, further causing DNA damage and leading to diseases or cancers (Table 1).¹³⁴

5.2 MCMs Ubiquitination

Protein Ub is a well-known pathway for target protein degradation. Otherwise, protein Ub also modulates multiple cellular biological processes such as DNA replication, cell cycle checkpoint activation, and DNA repair. Mass spectrometry (MS) results showed that all the MCM proteins in eukaryotes are ubiquitinated in human cells. Of those, MCMs are ubiquitinated by diverse E3 ligases when cells are threatened by DNA damage or RS (Table 1).¹⁵⁵

The Kelch-like ECH-associated protein 1 (KEAP1) is one crucial candidate of the Cullin3 (CUL3)–RBX1 E3 ligase complex, which ubiquitinates MCM3 in actively proliferating cells.¹⁵⁶ KEAP1-mediated MCM3 Ub regulates cell cycle progression and genome stability by controlling the MCM2-7 complex helicase activation. Actually, KEAP1 itself serves as the crucial component in response to oxidative stress, which may be achieved through MCM2-7 complex chromatin loading.

Recent Ub proteomic analysis revealed that ubiquitin protein ligase E3A (UBE3A) could interact with HERC2 and MCM6 with unknown functions.¹⁵⁷ Apparently, HERC2 is a crucial DNA damage repair factor participating in HR repair at DSB sites. In addition, HERC2, with RNF8, has been shown to promote translesion synthesis (TLS) at stalled replication forks.^{158, 159} Thus, not far to seek, UBE3A-mediated MCM6 Ub may interact with HERC2 to keep the chromosome stable and further play a role in DNA repair.

During DNA replication termination, CDC49/p97 complex targets polyubiquitinated MCM7 to disengage CMG complex, thus terminating DNA replication.^{160, 161} George et al.¹⁶¹ supported that polyubiquitylation of MCM7 has a modest effect to interstrand cross-links (ICLs) repair, which suggests that MCM7 proteasomal degradation may play a more active role in response to DNA damage. Moreover, illustrious HR repair-associated factor BRCA1 serves as upstream of MCM7 Ub.¹⁶² BRCA1 recruits additional E3 ligases to promote MCM proteins and CMG complex Ub.¹⁶³ It is necessary that helicases remove from the damaged DNA after accomplishing recovery. During ICL repair, BRCA1-mediated CMG Ub assists their disassembly, positioning a distinct regulatory signal to ensure unloading initiation. Thus, Ub-mediated MCMs unloading provides an appropriate occasion to resolve RS.

The best-characterized E3 ligase comes from S. cerevisiae, cullin 1 ligase SCF^Dia2, drives CMG ubiquitylation to perform CMG helicase disassembly. However, CMG depolymerizing has various pathways during eukaryotic evolution.¹⁶⁴ Subsequent works indicated that cullin 2 ligase CUL2^LRR1-mediated MCM7 Ub is essential to preserve genome stability during DNA replication termination both in yeast and in human cells.^{165, 166} Further work indicated that two crucial E2 enzymes, UBE2R1/R2 and UBE2G1/G2, connect with CUL2^LRR1 to extend a polyubiquitin chain on MCM7.¹⁶⁵ Ub of MCMs departs from chromatin due to their topological alternation, which may form a functional MCM2-7 hexamer in their de-Ub pattern.

Deng et al.¹⁶⁷ suggested that tumor necrosis factor receptor-associated factor-interacting protein (TRAIP) acts as ubiquitin ligase associating with the CMG replisome, thus triggering replication fork collapse. Previous work indicated that TRAIP is crucial in regulating normal cell cycle to keep genome stability in eukaryotes.¹⁶⁸ However, Favrizio et al.¹⁶⁶ found that TRAIP-mediated Ub of MCM7 triggers CMG disassembling in mouse embryonic stem cells. ICL-induced DNA replication stalling is repaired by endonuclease 8-like protein 3 (NEIL3) glycosylase and FA pathway.¹⁶⁹ Wu et al.¹⁷⁰ identified stalled replication fork triggers TRAIP-mediated MCM7 Ub to participate in ICL repair. TRAIP-mediated MCM7 Ub causes distinct ICL-repair pathways, NEIL3 recognized short ubiquitin chains to cleave directly, while long ubiquitin chains recognized by p97 complex to trigger FA pathway.

In S. cerevisiae, MCM10 is mono-ubiquitinated at two distinct lysine sites, subsequently, interacts with PCNA.¹⁷¹ Expression level of MCM10 is precisely mediated by the CUL4^DDB1 complex.¹⁷² De-Ub of MCM10 results in hypersensitive to HU owing to dysregulation of the interaction between MCM10 and PCNA.

In summary, K48-linked MCM7 degradation leads to disassembly of the MCM2-7 complex, which is critical in DNA replication termination. However, BRCA1-mediated Ub of MCM7 takes part in HR and ICL repair but not in replication termination. It provides us a novel sight that Ub in different MCM subunits or various sites performs distinct functions. It also possible that specific E3 triggers distinct Ub-chains to ubiquitinate MCMs for degradation or activation pattern.

5.3 MCMs SUMOylation

SUMO is a protein modifier that plays crucial roles in a wide range of cellular processes, making it essential for the viability of most eukaryotes. SUMOylation is a multistep process modulated by specific E1, E2, and E3 enzymes, like Ub. Compared with Ub, SUMOylation of proteins do not mediate their degradation, modulating their subcellular compartmentalization and reinforcing their stability (Table 1).^{173, 174}

Previous studies showed that DNA alkylating agents stimulated SUMOylation of MCMs except for MCM3 and MCM7. However, MCM2, 3, 4, and 7 were SUMOylated in response to heat shock in human cell, indicating that MCM SUMOylation may modulate cells against cytotoxic stress.¹⁷⁵

The SUMO-target ubiquitin ligase Slx5/Slx8 in S. cerevisiae are crucial in modulating DNA repair via SUMOylation repair factors.¹⁷⁶ Coincidentally, Slx5-based proteomic research revealed that MCM2-7 complex may be as potential substrates of Slx5/Slx8. These data suggest the Slx5/Slx8-mediated SUMOylation of MCM2-7 may take effect during DNA replication and DDR.¹⁷⁷

SUMO modification of MCM3 at K767 and K768 may work together to directly or indirectly promote MCMs loading onto chromatin. Site-specific mutagenesis of MCM3 K767/768 leads to MCM2-7 complex disassembly and CMG complex collapse, delaying the chromosomal DNA replication and leading to genome instability. Uniformly, factitious de-SUMOylation of MCM3 may generate spontaneous DSBs due to incomplete DNA replication, which is quite lethal to cells. ¹⁷⁵

S. cerevisiae harbors three SUMO E3 ligases: Siz1, Siz2, and Mms21, which are necessary for controlling intracellular activities.¹⁷⁸ Except for SUMO E3 ligases, SUMOylation is also modulated by SUMO isopeptidase Ulp1 and Ulp2, which performing de-SUMOylation effect. Ulp2 is inutile for cell viability but necessary for the accumulation of poly-SUMO chains.¹⁷⁹ de Albuquerque et al.¹⁸⁰ found that loss of Ulp2 aggravates SUMOylation of MCM4 and MCM7, while partially downregulating MCM6 SUMOylation. Mms21, but not Siz1 and Siz2, mediates SUMOylation of MCM3 under HU stimulation, suggesting that Mms21-dependent SUMOylation of MCM3 might contribute to regulating DNA replication and respond to DNA RS.¹⁸⁰ Siz1/Siz2-mediated SUMOylation of MCMs has been detected in unperturbed cells, whereas Mms21 preferentially interacting with MCM2 and MCM3.¹⁸¹ Wei and Zhao found that SUMOylation of MCMs exhibits preference for chromatin-bound MCM subunits including MCM4, MCM6 and MCM7. SUMOylation of MCM proteins leads to decreased CMG protein levels and inhibits DNA replication initiation.¹⁸²

Tian et al.¹⁸³ identified a germline variant rs2274110 in MCM10 that confers an inferior survival of esophageal squamous cell carcinoma (ESCC) patients. This functional variant can increase MCM10 SUMOylation resulting in aberrant overexpression, substantially facilitating ESCC progression via fueling DNA over-replication and genomic instability. These findings underline that PTMs of MCM proteins may serve as potential therapeutic targets in tumor treatment.¹⁸³

5.4 MCMs acetylation

Lysine acetylation is a widespread and versatile protein PTM. Indeed, nonhistone protein acetylation is deemed as a key regulatory component in multiple biological processes such as DNA replication, DNA damage repair, autophagy, and metabolism. Several studies revealed that MCM proteins are substrates for acetylation (Table 1).¹⁸⁴

MCM3AP acts as an acetyltransferase to acetylate MCM3, which promotes the translocation of MCM3 from the cytoplasm into the nuclei. Moreover, MCM3AP-mediated acetylation of MCM3 can inhibit DNA replication.¹⁸⁵

HBO1 complexes belong to the MYST family and are major acetyltransferases aiming for histone H4 acetylation in vivo. More recently, HBO1 was deemed to modulate the replication origin of Kaposi's sarcoma-associated herpes virus. These functional interactions implied a putative function of HBO1 in pre-RC formation and replication licensing.¹⁸⁶ Lizuka et al.¹⁸⁷ demonstrated that HBO1 significantly acetylates DNA replication-associated proteins, such as ORC2, MCM2, CDC6, and Geminin. HBO1-mediated MCM2 might regulate the initiation of DNA replication.¹⁸⁷ During DNA replication initiation step, recruitment of HBO1 to origin by Cdt1 is required for MCM2-7 complex loading in human cells, which may stabilize the interaction of MCM complex with chromatin.¹⁸⁸ It provides a novel insight that HBO1 may acetylate MCMs to perform conformation alternation, further modulating the DNA replication or DNA damage repair.

SIRT1 is a histone deacetylase that has been implicated in containing chromatin structure and DNA repair, serving as a crucial guard to maintain genomic stability. Samuel et al.¹⁸⁹ demonstrated that deacetylation of MCM10 by SIRT1 is one of the vital regulatory events in preserving genome stability. Moreover, MS and biochemical analysis indicated that twelve lysine residues of MCM10 acetylated by p300 are involved in DNA binding.¹⁸⁹ These results indicated that the dynamic balance of MCM10 acetylation has to be tightly regulated for proper fork initiation and stable genome integrity.

5.5 Other PTMs of the MCM proteins

Protein O-GlcNacylation is involved in multiple biological processes, especially in stress response. O-GlcNacylation of proteins is catalyzed by O-GlcNAc transferase (OGT) to transfer the GlcNAc group onto serine or threonine residues of proteins. Reciprocally, O-GlcNAcase (OGA) reverses these PTMs by removing the GlcNAc residue.¹⁹⁰ In mammalian cells, O-GlcNAcylation levels are dynamic alternation during the cell cycle, thus abnormal O-GlyNAc dynamic cycling disrupts the cell cycle and causes RS. Using a mass-tagging strategy, Leturcq et al.¹⁹¹ identified that MCM2-7 all subunits are O-GlcNAcylated by OGT in human cells, especially in MCM3, MCM6, and MCM7. Each subunit of MCM2-7 complex gradually disperses in knockout OGT cells, subsequently departing from chromatin. Thus, it is tempting to speculate that O-GlcNAcylation of MCMs assists MCM assembly and regulates dynamic balance during DNA damage and DNA replication.¹⁹¹

Lysine methylation usually occurs in histones, whereas in nonhistones in recent decades. Methylation can alter the conformation of proteins thus changing their function. Xia et al.¹⁹² identified that recombinant Sulfolobus MCM (sisMCM), an archaeal homolog of MCM2-7 eukaryotic replicative helicase, is mono-methylation by aKMT4 in vitro, which is characterized as the first archaeal lysine methyltransferase. Interestingly, MCM methylation (me-MCM) upregulates MCM complex DNA unwinding ability, modulating their helicase activity. More intriguingly, me-MCM also enhances heat resistance, which supports that methylation of MCM proteins also impacts protein thermal properties (Table 1).

5.6 PCNA Ubiquitination and SUMOylation

During DNA replication, PCNA serves as the pivot to recruit replicative polymerases polε and polδ to perform high-fidelity DNA synthesis.¹⁹³ When replicative DNA polymerase encounters damaged DNA, the progression of the polymerases is blocked, and the replication fork is stalled. If this problem were not be resolved, the replication fork would be collapsed, resulting in cell death. As we summarized above, the blooey replication fork activates a unique DNA repair pathway called postreplication repair (PRR), such as error-prone translesion DNA synthesis (TLS) and error-free template switching (TS).¹⁹⁴ Specialized DNA TLS polymerases have been identified in yeast and mammalian, including polymerase η (polη), polymerase ι (polι), and polymerase κ (polκ). For instance, polη mediates efficient and precise TLS past UV-induced thymine-thymine CPD (T-T CPD), whereas resulting in the high frequency of mutations in routine replication progression.¹⁹⁵ Low-fidelity TLS polymerase causes incorrect nucleotide insertion in normal replication, subsequently stimulating NER, BER, or HR pathways to fix the errors.¹⁹⁶

PCNA is mono-ubiquitinated at K164 by RAD6–RAD18 E2–E3 complex in response to replication fork stalling.¹⁹⁷ Mono-ubiquitinated PCNA increases affinity with polη, thus further promoting efficient TLS.¹⁹⁸ Hence, dysregulation of replication and TLS progression cause severe genomic instability and tumorigenesis. Except for the Ub of PCNA, multiple modifications are involved in regulating PCNA functions, thereby impacting DNA replication, DNA repair, and even carcinogenesis.

Ub of PCNA undergoes regulation and control from the chromatin microenvironment by various factors. The chromatin structure could be regulated by PTMs of histone and nucleosome remodeling, which is vital for DNA replication, cell cycle, and DNA damage repair. Mutation of histone H3 and H4 disrupts DNA packaging, resulting in reduced expression of PCNA Ub under methyl methanesulfonate (MMS) or UV irradiation.¹⁹⁹

RAD6–RAD18-mediated PCNA–K164 mono-Ub is widely known in a variety of organisms. Meanwhile, polyubiquitination of PCNA has also been observed at K164 residue in yeast and mammal cells, which is essential for inducing error-free pathways upon damage response.^{200, 201} K63-linked polyubiquitination of PCNA requires a ternary complex composed of RING-finger E3 ligase RAD5 and Mms2–Ubc13 complex, protecting cellular DNA against genomic mutations via a TS pathway.^{200, 202, 203}

RNF8 is a crucial E3 ligase in regulating histone Ub during DNA damage repair, while it is also identified as a novel E3 ligase for PCNA. Nevertheless, RNF8 cooperates with distinct E2 such as Ubc15 and Mms2 to activate mono-Ub and polyubiquitination, respectively.²⁰⁴ Depletion of RNF8 in medulloblastoma cells significantly suppresses PCNA mono-Ub under UV damage, which is associated with UV-induced p53 targeting and checkpoint activation.

In response to stalling DNA replication fork, PCNA could also be modified by SUMOylation at K164 residue, which is commonly observed in yeast and mammals. Furthermore, some reports also supported that PCNA might be SUMOylated at K127 residue and K164 in response to DNA damage.²⁰⁵ SUMOylation is closely related to Ub that modulates protein coordinated interaction, whereas possibly antagonizes Ub via competing the same residue in substates.²⁰⁶ In contrast, forceful evidence supported that SUMOylation of PCNA prohibits its Ub and DNA damage repair progression, implying SUMO of PCNA functions on regulating normal DNA replication.

In S. cerevisiae, SUMOylation of PCNA recruit helicase Srs2 via its C-terminus SUMO-interaction motif (SIM), which disrupt RAD51 single-stranded presynaptic filaments, ultimately interrupt HR progression.²⁰⁷ Furthermore, RFC is required for PCNA SUMOylation, which facilitating PCNA loading onto chromatin. SUMOylation of PCNA also cooperates with PIP, suppressing inappropriate HR.²⁰⁸ Moreover, Gali et al.²⁰⁹ revealed that SUMOylation of PCNA could decrease DSBs formation using neutral comets assay, even lower spontaneous recombination frequencies, and enhance damage resistance. Thus, impaired PCNA SUMOylation facilitates DSB formation at stalled replication fork, which highlight its role in preserving genome integrity.

One intriguing perspective supports that SUMOylation could conduct for a signal for Ub. SUMO-targeting ubiquitin ligases (STUbLs) intervenes protein SUMOylation and Ub of SUMO segments via its SIMs. Mutation of PCNA SUMOylation site K164/K127 and RAD18 SIMs residues directly decrease PCNA Ub level, implying SUMOylation of PCNA facilitates RAD18 to target PCNA.²¹⁰ Therefore, it points out that the SUMO and Ub crosstalk may be essential for DNA replication and DNA damage repair (Table 2).

6.1 MCMs in tumorigenesis and development

Genome instability is a hallmark of cancer. Conventional expression of the MCM2-7 complex ensures routine DNA replication progression, which is a prerequisite for genome stability.²¹³ Nevertheless, numerous studies have indicated that both deficiency and overexpression of MCMs are associated with cancer development. Mcm3-deficient mouse model was used to determine the impact on gene function in hematopoietic stem cells. The results indicated that downregulation of MCM3 results in RS, further leading to fetal anemia during embryonic development.²¹⁴ In contrast, MCM6 was overexpressed in clear-cell renal cell carcinoma.²¹⁵ Upregulation of MCM6 also exists in non-small cell lung cancer and breast cancer with worse survival and higher histological grade.^216-218 The reasons for abnormal MCMs expression remain unclear. There are two possible speculations: (a) CDK-mediated MCM complex dissociating prevents DNA re-replication. However, dysregulation of the cell cycle-dependent kinase CDK permits MCMs constantly binding to each other, resulting in continuous cell division and high expression.^{219, 220} (b) Aberrant DNA replication license system induces abnormal DNA replication, increasing genomic instability, and carcinogenesis.^{221, 222} In addition, spontaneous mutation of MCMs increases chromosome elimination and DNA damage,²²³ whereas a more than twofold reduction of MCM protein expression could lead to genomic instability in S. cerevisiae. These findings demonstrated that MCMs are indeed involved in tumorigenesis, but the detailed mechanism is still unclear.²¹²

In lung cancer, omics data of MCM2 overexpression is analyzed using the Gene Expression Omnibus database, which is associated with large tumor size, different malign degrees, and clinical stages.²²⁴ Using RT-PCR analysis, except for MCM3 and MCM5, MCMs are upregulated in cervical cancer in vitro and in vivo, which are critical in tumor progression.²²⁵ Spontaneous dominant leukemia (Sdl) mice model is a murine model for heritable T cell lymphoblastic leukemia/lymphoma, which harbors a spontaneous mutation in Mcm4 (Mcm4^D573H). Mcm4^D573H could not alter the total expression of MCM2-7 complex, whereas it significantly promotes tumor formation.²²⁶ In laryngeal squamous cell carcinoma cells, knockout of MCM4 using siRNA also suppresses cell proliferation and inhibits of tumor progression.²²⁷

MCM7 is regarded as an extensive mark for cancer development.²²⁸ Some studies indicated that MCM7 genome sequence embodies a cluster of miRNAs (miR-106b, miR-93, miR-25), which can downregulate the expression of oncogenes, including p21, E2F1, BIM, and PTEN.²²⁹ PRMT5 acts as one methyltransferase to methylate multiple proteins in histones,²³⁰ which is also deemed a potential target in colorectal cancer development and progression.²³¹ Recent studies revealed that PRMT5 physically interacts with MCM7 in HCT8 cells, while MCM7 depletion impairs cancer cell migration and invasion.²³² Using TCGA analysis, the expression of MCM7 enhanced by approximately 12% in ESCC. Silencing of MCM7 via siRNA significantly impaired KYSE510 cell proliferation and migration in vitro.²³³ Furthermore, miR-214 targets overexpression of MCM5 and MCM7 in hepatocellular carcinoma (HCC) cells to inhibit cell replication and colony formation.²³⁴

Massive rearrangements are one of the characteristics of aggressive cancer genomes. MCM8, unclassical MCM proteins, is deemed to interrelate with chromosome rearrangement.²³⁵ Knockdown of MCM8 in mice diminished xenografted tumor volume, which implied the critical role of MCM8 in tumor metastasis in vivo.²³⁶

Despite novel and striking findings, additional investigations are still needed to be addressed. Since closely associated with tumorigenesis, MCMs can be used as precise cancer therapy via molecular targets.

6.2 MCMs as diagnostic and prognostic biomarkers

As mentioned above, aberrant expression of MCMs is closely related to tumorigenesis and development. Therefore, MCMs are also used as tumor biomarkers and indicators of prognosis.²³⁷ For example, MCM2 is regarded as a novel proliferation biomarker for oligodendroglioma,²³⁸ ESCC,²³⁹ and breast cancer.^{240, 241} However, the detailed interventional mechanism is different due to MCM2's diversiform role in the distinct process. Knockdown of MCM2 abolishes DNA damage in ESCC cells, interfering with DNA replication in breast cancer cells. Meanwhile, MCM2 is also suggested to be a prognostic marker for some tumors such as renal cell carcinoma,²⁴² laryngeal carcinoma,^{243, 244} and gastric cancer.^{245, 246} In oral squamous cell carcinoma^{247, 248} and large B-cell lymphoma,^{249, 250} MCM2 also nominates as a prognostic marker significantly related to malignant progression and the 2-year survival rate of patients, respectively. In addition, abnormal expression of MCM3 reflects advanced tumor stage and metastatic status in cervical cancer,^{251, 252} breast cancer,^{253, 254} oral squamous cell carcinoma,²⁵⁵ malignant salivary gland tumors,^{256, 257} and HCC.^{258, 259} Simultaneously, based on the TCGA and GEO analysis, some reports persist that MCM4 mainly serves as a prognostic indicator for HCC,^{260, 261} which is relevant to poor prognosis with MCM4 overexpression pattern. Furthermore, overexpression of MCM4 may also be a diagnostic signal in esophageal cancer,^{262, 263} colorectal cancer,^{264, 265} cervical cancer,^{225, 266, 267} ovarian cancer,^268-270, breast cancer,⁵⁶ and gastric cancer.^{271, 272} In addition, upregulated expression of MCM5 is mainly related to poor prognosis and malignant status.^{115, 225, 273-275} Consistent with this notion, overexpression of MCM5 also associates with tumor stages, which appears to be potential diagnostic and prognostic markers in thyroid cancer,^{276, 277} ovarian cancer,^{269, 278-280} bladder cancer,^281-283 and renal cell carcinoma.²⁸⁴

On the contrary, WGCNA combining with TCGA and GEO analysis identified that exceptional expression of MCM6 reflects pathologic stage of multiple tumors, serving as putative biomarkers for breast cancer,^285-287 gastric cancer,^{288, 289} renal cell carcinoma,^{215, 290, 291} HCC,^292-295 and non-small cell lung carcinoma.^{217, 296, 297} Meanwhile, MCM7 is widely regarded as an extensive biomarker in multiple tumor types since its overexpression pattern.^{233, 234, 298-310} Except for conventional MCM proteins, MCM8, MCM9, and MCM10 are also novel prognostic markers in multiple tumors. In tissues, high expression of MCM8 may act as a valuable prognostic indicator for different cancer therapy, consistent with gastric and cervical cancer.^311-313 MCM9 and MCM10 are associated with additional tumor types such as colorectal cancer, breast cancer, ovarian cancer, and HCC (Table 3).^314-322

Compared with the common-used proliferation biomarker, MCM proteins are more sensitive and particular than PCNA and Ki-67,³²³ accurately reflecting cell proliferation status and predicting prognostic tumor patient's survival rate.

6.3 MCMs as therapeutic target

Since increased DDR preventing cancer cells from effective therapy, MCM2-7 proteins as intermediates are involved in intricate progression, which may serve as a pivotal role in influencing therapy response.

Conventional chemotherapy is one of the staple cancer treatment strategies, with emerging resistance causing the limited anticancer effects.³²⁴ Previous reviews highlighted that chemotherapy uniting with MCMs knockout approach inhibits tumor cell proliferation.³²⁵ Recently, combination chemotherapies have been widely used in cancer treatment, especially with the knockdown of MCM proteins. Carboplatin-based chemotherapy is for the initial treatment of ovarian cancer, while carboplatin resistance results in treatment failure.³²⁶ Deng et al.³²⁷ found that knockdown of MCM2 could enhance the carboplatin sensitivity of A2780 cells and increase cells’ UV sensitivity, which may owe to accumulation of damaged DNA and activation of the p53-dependent apoptotic response. Oxaliplatin or etoposide-mediated chemotherapy combined with knockdown of MCM7 could reduce the proliferation of colorectal carcinoma cells and induce tumor apoptosis in vitro.³²⁸ In pancreatic ductal adenocarcinoma cells, reduction of MCM4 or MCM7 clearly exhibits more sensitive to gemcitabine and 5-FU exposure, which may be caused by MCM suppression-induced RS.³²⁸ Classic hypercholesterolemia curative simvastatin was demonstrated that reduces the expression of MCM7.³²⁹ Thus, it implied that simvastatin combining with chemotherapeutic drugs may be the putative cancer therapy for some chemo-resistance cancer treatment. Liang et al.³³⁰ demonstrated that simvastatin combines with tamoxifen impaired breast cancer cell proliferation and resulted in apoptosis in vivo and in vitro. Knockout of MCM8 or MCM9 selectively increases cisplatin sensitivity in specific cancer cells such as HCT116 and HeLa cells.^{331, 332} However, siRNA-mediated MCM8 silencing could not alter the cisplatin sensitivity of normal HFF2/T fibroblasts, indicating MCM8 may act as a molecular target just in cancer cells. Consistent with cisplatin, silencing MCM8 and MCM9 selectively hypersensitizes cancer cells to Olaparib, which may rely on MCM8-9's role in resolving RS.^{333, 334}

Radiotherapy is an alternative cancer therapy known to induce cancer cells’ DNA damage and autophagy to reach the treatment goals, while acquired radio-resistance disturbs the effectiveness.³³⁵ Next-generation sequencing of mRNA (RNA-seq) results revealed that MCM7 is significantly upregulated in radio-resistant PC-3 cells after 2Gy IR treatment. Diminishing expression of MCM7 might increase radiotherapy response in prostate cancer.³³⁶ Consistent with this notion that upregulated expression of MCM2 is involved in the radio-resistant cervical cancer cell, indicating that MCM2 is one potential regulatory factor in increasing radio-sensitivity in cancer treatment.³³⁷ As mentioned above, MCM3 is one prognosis marker for HCC with high expression in vitro and in vivo. Using MTT and TUNEL methods, low MCM3 expression HCC cell line performs low growth, whereas high MCM3 expression induces lower apoptosis under radiotherapy. In addition, overexpression of MCM3 further promotes HCC cell radio-resistance, revealing MCM3 prevents HCC radiotherapy efficiency via activating NF-κB pathway.²⁵⁹

Specific small molecule inhibition for MCM proteins is an additional invaluable approach for various cancer treatments.³³⁸ There are three typical MCMs-based small molecule inhibitors with potential chemotherapeutic effects: (a) Enzyme inhibitors such as DNA helicase-targeting small molecule inhibitors.³³⁹ Ciprofloxacin targets MCM2-7 complex to block the helicase activity, further inhibiting cell proliferation.^{340, 341} (b) The inhibitors prevent interaction among MCM subunits.^{112, 342, 343} (c) The inhibitors regulate the expression of MCM proteins.³⁴⁴ Widdrol could downregulate the expression of MCM proteins to inhibit cancer cell proliferation in G1 phase.^{345, 346} In addition, trichostatin A targets MCM2 to inhibit its expression.^{347, 348} Recent studies revealed that Breviscapine downregulates the expression of MCM7 and impairs tumor progression in prostate cancer via activating DNA damage-induced apoptosis (Table 4).³⁴⁹

6.4 GINS and Cdc45 as prognostic markers and the target for therapy

Except for MCMs, concurrent overexpression of GINS and Cdc45 are also observed in various cancers. Since Psf1 promoter activity is related with 17β-estradiol (E2)-based estrogen receptor pathway, aberrant expression of Psf1 might be a signal in breast cancer. Nakahara et al.³⁵⁰ revealed that expression of Psf1 is remarkably increased in breast cancer both in vivo and in vitro, whereas siRNA-mediated depletion of Psf1 inhibits DNA replication and cell growth. This evidence provides that Psf1 could be deemed as a novel breast cancer biomarker and therapeutic benefit for breast cancers with overexpression of Psf1.³⁵⁰ Moreover, uprelated expression of Psf1 is also detected in non-small cell lung cancer, which implies Psf1 as a prognostic biomarker and potential target for lung cancer therapy.^{351, 352} Using the tumor cell xenograft model, Nakahama et al. found higher expression of Psf1 is correlative with higher proliferative ability and metastatic capability, implicating Psf1 in tumorigenesis and its conceivable role as a therapeutic target. Furthermore, anomalous expression of Psf1 also exists in prostate cancer and HCC,^{353, 354} significantly correlated with tumor grade and clinical stage.

Using cDNA microarray analysis, Psf2 is frequently upregulated in cholangiocarcinoma, while knockdown of Psf2 drastically reduces cell proliferation and inhibits cell growth.³⁵⁵ Furthermore, obvious upregulation of Psf2 is detectible in cervical cancer cell, whereas downregulation of Psf2 inhibits cell multiplication and tumorigenic ability.³⁵⁶ Combined with the observations, Psf2 might be a novel and valuable prognostic biomarker for ovarian cancer and leukemia,^{357, 358} subsequently an underlying molecular target for cancer diagnosis and treatment. In triple negative breast cancer (TNBC) cell lines, similar with Psf1, the expression of Psf2 is also enriched correlated with the advanced stages of tumor. Intriguingly, silencing of Psf2 decreases the expression of matrix metallopeptidase 9, which is necessary for tumor invasion, hence suppress tumor cell migration and invasion.³⁵⁹

Aberrant expression of Psf3 in colon carcinoma cell line associates with tumor cell proliferation and tumor progression, simultaneously, the similar results have also been found in non-small cell lung cancer, lung adenocarcinoma, and colorectal cancer patients.^360-363 Finally, abundant Sld5 expression is closely related to bladder cancer and gastric cancer, since defect of Sld5 causes severe tumor cell proliferation disorder and inhibits cell growth.^{364, 365}

Upregulated expression of Cdc45 is also a predictor for multiple cancers, including breast cancer, leukemia, lung cancer and osteosarcoma.^{366, 367} Using the Cancer Genome Atlas, prominent overexpression of Cdc45 is discovered in papillary thyroid cancer tissue, which impacts tumor sizes and cancer stages.³⁶⁸ In tongue squamous cell carcinomas (SCC), higher expression of Cdc45 with severe lymph node status is observed in malignant tumor than mild precancerous epithelial dysplasia, which implies its role in distinguish precancerous dysplasia from SCC.³⁶⁹ Myc is one crucial factor in regulating cell growth and tumorigenesis, overexpression of Myc enhances the Cdc45 DNA binding ability and replication fork stalling, which indicates Myc-induced RS collaborates with Cdc45 in tumor development.^{370, 371}

In summary, excessive expression of CMG is generally related with tumor size and malignancy but to varying extents, depending on the type of tumor. Thus, except for MCMs, GINS and Cdc45 could also serve as diagnostic and prognostic biomarkers for multiple tumors, which also as a druggable target to treat cancers (Table 5).

Cancer immunotherapy is a novel biological treatment for malignant tumors, which could be collaborated with traditional radiotherapy and chemotherapy to improve the curative rate. Cancer immunotherapy elicits the immune system by identifying specifically expressed molecules on the tumor surface, thereby eliminating the cancer cells.³⁷² Cytotoxic T lymphocytes (CTLs) are specific T cells with potent nocuity to cancer cells, which recognizing cancer-specific antigenic peptides human leukocyte antigen (HLA). Though mass spectrometric analyses and bioinformatic analysis, Yoshida et al.³⁷³ isolates a Psf1-derived peptide presented by HLA. Moreover, they found no other cancer vaccine target proteins except for Psf1. Detailed peptide is verified using the mouse model, suggesting Psf1_79–87 peptide induces CTL response in vitro and in vivo, which provides a novel cancer immunotherapy for targeting cancer stem cells. Previously, the similar approach was utilized in Cdc45, demonstrating that strongly immunogenic Cdc45-derived peptides stimulated CTLs to be reactive to lung cancer cells.³⁷³

6.5 PCNA as prognostic markers and the target for therapy

Due to its role in cell proliferation, PCNA is deemed as the tumor marker for diagnosis and patient prognosis. Overexpression of PCNA is observed in lung cancer in vitro and in vivo, while silencing of PCNA reduces cell invasion ability and 95D cells proliferations.³⁷⁴ Identical results also occur in prostate carcinoma and breast cancer that PCNA connects with pathological stage and cellular grade, suggesting PCNA might be a crucial prognostic indicator of malignant tumors.^{375, 376} Furthermore, upregulated PCNA is associated with various digestive system tumors including colorectal carcinoma, ESCC, and HCC (Table 5).^377-379

PCNA is a critical factor in DNA replication and DNA damage repair. Hence, PCNA is a target for designing antiproliferation and anticancer drugs. Since multiple PTMs of PCNA interrupt its chromatin binding ability, developing therapeutics of modified PCNA will be conducive to target cancer cells. Previous research found that Y211 phosphorylation of PCNA is a crucial event to preserve the PCNA stability, promoting DNA damage repair and DNA synthesis. Therefore, mutant of Y211 phosphorylation inhibits tumor cells proliferation including prostate cancer and breast cancer.^{380, 381} In addition, specific small molecular inhibitors targeting PTMs of PCNA are also identified to disrupt cell proliferation and enhance chemosensitivity or radiosensitivity. RAD6 selective small molecular inhibitor SMI#9 leads to PCNA mono-Ub defect and mitochondrial function reduction, suggesting RAD6 inhibitor serves as a promising strategy for TNBC treatment.³⁸²

Although PTMs of PCNA have several implications in carcinogenesis; however, the detailed antitumor mechanism needs further investigation. Moreover, novel inhibitors or strategies inhibiting tumor development or proliferation via targeting PTMs of PCNA require identification.