1. Introduction

Proteases hydrolyze and cleave peptide bonds to degrade proteins and peptides. They are classified into several categories based on their catalytic mechanism: cysteine, glutamic, metalloproteases, serine, and threonine proteases. These proteases are central to the regulation of biological processes and can play an important role as effector proteins in various host–microbe and microbe–microbe interactions. Many of these activities have ties to human health and disease, making proteases potential targets for therapeutics [1, 2]. In addition, proteases have broad applications in food, agriculture, and waste degradation, which have been examined elsewhere [3, 4].

In this minireview, we focus on the role of bacterial proteases from the microbiome in human disease. Proteases from pathogenic organisms can be key virulence factors that facilitate infection and colonization that could lead to severe morbidity and mortality. Both the human microbiome and proteases have each been exciting topics of study in recent years, but there has been a lack of review focusing specifically on the role of microbiome-derived proteases in human health. Here, we provide a succinct overview of research on specific microbiome-derived bacterial proteases that play a role in disease pathogenesis, with a focus on the gut, skin, vaginal, and urinary microbiomes. The proteases involved in disease pathogenesis at these locations could be the target for therapeutics research.

2. Gut Microbiome

Bacterial proteases have been previously reviewed in the context of intestinal infections and diseases [1, 5]. Both pathogenic and commensal bacteria secrete various proteases that could increase susceptibility or exacerbate intestinal conditions that lead to disease. These luminal proteases directly affect macromolecules themselves to drive changes in cell signaling networks, the microbes in the microbiome, immune components, and physical barriers.

In a study of proteolytic activity from enteric microbes in healthy and irritable bowel syndrome (IBS) cohorts, fecal samples with high protease activity were compositionally distinct from low protease activity samples and also had lower species diversity [6]. Bacteria from Lactobacillales, Lachnospiraceae, and Streptococcaceae were associated with fecal protease activity. Studies with IBD (inflammatory bowel disease) patients also revealed upregulated total protease activity compared to healthy controls [7]. Using inhibitors targeting each protease family, fecal samples of IBD patients showed activity from serine proteases as well as trypsin, elastase, cathepsin G, and proteinase 3-like activity. Further analysis examining serine proteases from Clostridium was able to identify the serine protease SP-1 [8]. SP-1 shares sequence similarity to subtilisin and other serine proteases from Coprobacillus, Anaerotruncus, and Methanobacterium. Numerous subtilisin-like proteases are found in Clostridium genomes, and SP-1 itself does not match to proteases in bacteria associated with IBD. Lactococcus lactis expressing SP-1 was shown to exacerbate oral DSS-induced colitis and inflammation. Omics data also supports the idea that gut bacterial proteases are linked to diseases such as ulcerative colitis (UC) and Crohn’s disease (CD). Analysis of multiomics datasets from multiple cohorts of patients, including those with UC and CD, showed that proteases from Bacteroides vulgatus could induce colitis and that these proteases were enriched in a subset of clinically active UC patients [9]. Dipeptides and oligopeptides, which included fragments from human proteins such as collagen and mucin, were closely correlated with UC and patents with more Bacteroides protease production. Using Caco-2 and mouse models, proteases from B. vulgatus were also shown to cause colonic epithelial damage and inflammatory cell infiltration. These proteases also contributed to colitis in germ-free mice in a feces transplant as assessed by gross indicators for colitis and histopathology. Thus, proteolytic activity from the gut microbiome could lead to, or create environments favorable for, the development of intestinal diseases.

Other proteases also explicitly target cells of the immune system and even post-translational protein modifications. The Pic protease targets human leukocyte adhesion proteins, impairs chemotaxis, and causes programmed cell death in T cells [10]. Another strategy for defeating host defenses includes affecting protein processing and modifications, such as ubiquitination, within cells. A previously uncharacterized protease that affects ubiquitination was found in intestinal pathogenic Escherichia coli strains [11]. ElaD is a cysteine, deubiquitinating, and specific protease and is found in all intestinal pathogenic strains. While the study did not further characterize ElaD, other work reviewing deubiquitinating proteins has shown that many pathogenic bacteria have similar activity as a form of disrupting antibacterial response [12]. Numerous bacterial effector proteins that act as deubiquitinating enzymes have been identified and vary in function from suppressing the NF-κB pathway to inhibiting autophagy. Such enzymes from the host can be pro- or anti-inflammatory in IBD [13]. Further work will be needed to fully characterize the contribution of deubiquitinating enzymes to intestinal pathologies. Together, these activities show that bacterial modulation of host cells and activities could enhance bacterial survival, colonization, and the development of intestinal disease.

In addition to degrading immune and host cell protein components, bacterial proteases can also target the intestinal lining to facilitate infection and result in a disease phenotype. Histology of tissue samples collected from rats given purified E. coli Pet protease showed evidence of hemorrhage, necrosis, and ulceration, thereby altering the intestinal epithelium [14]. The commensal but opportunistic pathogen Enterococcus faecalis secretes a metalloprotease, GelE, that affects intestinal barrier function and contributes to intestinal inflammation [15]. This was observed with both mouse models as well as cell culture using E. faecalis from patients with CD and UC. Mucus lines the intestinal tract and has both a physically protective and regulatory role. Enteropathogenic E. coli (EPEC) was shown to degrade mucin while nonpathogenic strains could not [16]. Serine proteases from enterohemorrhagic E. coli (EHEC) have been shown to contribute to diarrhea [17]. Both EspP and serine protease A stimulated ion transport in human colonoid monolayers in a voltage clamp experiment. Commensal bacterial strains can also contribute to survival and colonization by pathogens. Proteases from Bacteroides thetaiotaomicron can cleave Type III secretion system (T3SS) proteins, increase T3SS function, and facilitate effector movement into host cells [18]. Cleavage activity facilitates pore formation that can affect EHEC disease pathogenesis. Bacterial translocation across the intestinal layers is also important for the development of CD and can rely on host protein expression in certain sections of the intestines [19]. Adherent-invasive E. coli (AIEC) is associated with CD and produces Vat-AIEC, a protease that helps in AIEC mucosal colonization in CD patients [20]. Vat-AIEC is secreted, degrades mucin, and aids in AIEC colonization of ileal and colonic tissues in a mouse model of AIEC infection before colitis develops. By degrading mucin, the protease helps AIEC access epithelial cells instead of being trapped in the mucin layer. The chemical composition of the ileum was also found to activate Vat-AIEC, suggesting its site-specific activity to enhance bacterial colonization. The gastroenteritis-causing Campylobacter is also implicated in affecting the intestinal barrier and could facilitate the development of postinfection IBS [21, 22]. This disruption of barrier function has been shown to be facilitated by the secreted serine protease HtrA [23]. In vitro experiments show that HtrA alters the localization of occludin in the tight junction during infection by Campylobacter jejuni and can also cleave occludin. Colon biopsies from infected and noninfected subjects revealed that C. jejuni redistributes occludin as opposed to samples from noninfected controls. Similarly, HtrA has also been found to cleave another member of tight junctions, claudin-8 [24]. The activities of the protease could collectively result in C. jejuni moving between cells and allowing access to deeper tissues, thus contributing to its colonization. Due to its importance in compromising the integrity of the intestinal barrier and other activities as a virulence factor in bacteria, HtrA has been proposed as an antibiotic target; however, challenges exist, and more research is needed [25].

While their full contributions to disease pathogenesis and effects on both the host and other bacteria are yet to be fully characterized, bacterial proteases play an important role in intestinal infections and have been implicated in intestinal diseases. Recent articles provide further insight into proteolytic activity and disease in the gut microbiome [26, 27].

3. Skin Microbiome

Numerous bacterial secreted proteases have been implicated in the pathogenesis of skin diseases or contributing to virulence [28, 29]. Here, we detail several major bacterial species and the proteases that correspond to certain skin diseases as well as their role in microbial competition, which could further contribute to skin microbiome dysbiosis.

Secreted proteases are widely prevalent among Staphylococcus aureus strains and serve as regulated virulence factors that degrade host defenses and barriers [30]. These proteases have also been shown to impact immune signaling and allergic responses [31]. Colonization by S. aureus is often found in atopic dermatitis (AD) patients and not only exacerbates AD but also contributes to microbial dysbiosis and allergies in AD patients [32]. V8 protease, aureolysin, staphopain A, and staphopain B are encoded in a vast majority of S. aureus clinical isolates [33]. Some of these proteases were also detected at statistically significant levels in sera of S. aureus-infected patients compared to healthy controls. Expression of scpA also increases while the bacterium is inside an infected cell and was able to induce host cell death when produced from a noncytotoxic S. aureus strain [34]. Proteases are also important for bacterial growth in various environmental and host conditions. S. aureus growth is protease dependent in nutrient-limited conditions, and its proteases contribute to survival in the presence of antimicrobial peptides as well as in evading the innate immune system [35, 36]. Other S. aureus proteases such as the Spl group of proteases, EpiP, and exofoliative toxins A and B promote bacterial invasion, cleave structural components, and affect the skin barrier function [37]. An N-terminomics approach identified 280 potential cleave sites in 85 host–protein targets of the secreted V8 protease that included immune proteins and clotting cascade proteins as well as host-derived protease inhibitors [38].

In addition to S. aureus, the commensal bacterium Staphylococcus epidermidis is often found in high abundance on AD patient skin, and its extracellular protease activity could break down both chemical defenses and physical barriers of the skin, leading to damage and inflammation [39]. Bacteria collected from healthy and AD patient skins were assayed for proteolytic activity. Testing of strains and further characterization revealed that S. epidermidis produces an extracellular agr quorum sensing–regulated cysteine protease EcpA that degrades elastin, collagen I, and collagen IV. EcpA expression in S. epidermidis was high and caused disruption in the epidermal barrier in a mouse model of AD as well as the degradation of the barrier protein DSG-1. The EcpA protease was also shown to cleave the host defense antimicrobial peptide LL-37. Interestingly, S. epidermidis isolates have also been shown to produce extracellular proteases, especially serine proteases, that can inhibit S. aureus biofilms [40]. S. epidermidis also produces other extracellular proteases such as Esp and SepA that participate in biofilm formation [41]. Thus, it is likely that both S. epidermidis and S. aureus proteases can contribute to AD pathogenesis.

In acne vulgaris, genome sequencing of the ubiquitous Propionibacterium acnes identified an extracellular subtilisin-like protease (PPA598) as well as a tripeptidyl aminopeptidase (PPA247) [42]. Mass spectrometry of P. acnes culture supernatants identified the same PPA0598 protease [43]. However, the impact that these specific proteases have on host cell structure and acne development remains to be seen.

The various skin microbiome–derived proteases noted here contribute to their survival against host defenses, colonization, and potential link to disease pathogenesis. Evaluating the proteolytic capacity of known pathogens and commensals in patients with skin conditions will help provide a more thorough understanding of the disease and offer insight into treatment approaches.

4. Vaginal Microbiome

Microbial protease activity could provide additional context to the development and progression of vaginal diseases like bacterial vaginosis (BV). Various bacteria such as Gardnerella, Prevotella, Mobiluncus, and Porphyromonas contribute to the polymicrobial state seen in BV [44]. One recent study characterized how the gram-negative obligate anaerobe Porphyromonas in the vagina can secrete proteases that degrade extracellular matrix components such as collagen and fibrinogen and act as virulence factors [45]. Collagenase assays show that Porphyromonas asaccharolytica and Porphyromonas uenonis secrete proteases that degrade Type I and Type IV collagen, while commensal Lactobacillus cells were not able to. The Porphyromonas bacterium also completely degrades fibrinogen alpha and beta chains, which then delay fibrin clot formation. A combination of bioinformatic and functional assays reveals that a secreted metalloprotease related to the PepO endopeptidase functions as a virulence factor in the cervicovaginal niche. Proteases from the vaginal microbiome of individuals with BV have also been shown to degrade the protective glycoprotein mucin layer [46]. Anaerobic bacteria recovered from women with BV, normal, and intermediate flora were able to degrade mucin via glycine and arginine aminopeptidases. Bacterial concentrations from vaginal swabs obtained from the three groups also correlated with patient-reported abnormal and thin or watery discharge. Some bacteria can also form symbiotic relationships with other microbes found in BV microbiomes. Prevotella bivia forms a commensal relationship with Peptostreptococcus anaerobius through shared nutrient pathways [47]. These pathways include amino acid utilization that results from the metabolism of peptides by proteases and peptidases by P. bivia. One study noted that metabolic signatures in BV versus non-BV patients show statistically significant decreases in amounts of amino acids among BV cases [48]. The authors note that the lower amounts of amino acids and higher amounts of amino acid catabolites in BV cases versus non-BV patients could mean that amino acids are used more in BV patients.

Taken together, vaginal microbes linked to BV produce proteases that degrade structural and protective components present in normal and healthy vaginal environments. This alteration then likely creates a permissive environment for further microbial colonization that contributes to and maintains a disease phenotype. In addition, bacterial symbioses could exist that rely on proteolytic activities that provide amino acids that create optimal nutrient settings for BV-associated pathogens to grow.

5. Urinary Microbiome

A deeper understanding of how various biotic and abiotic factors contribute to urobiome dysbiosis and urinary tract infections (UTIs) could help create new avenues for treatment. To that end, virulence factors such as proteases in pathogenic bacteria that promote colonization and resistance are important to study.

A primary transposon screen in mice to evaluate genes associated with uropathogenic E. coli (UPEC) fitness during systemic infection revealed a number of genes encoding proteases [49]. Important genes include the zinc metallopeptidase pqqL and autotransporter serine proteases pic and vat. Enterobacteriaceae are known to harbor a family of serine protease autotransporters called SPATE. A coinfection model in mice using intravenously delivered wild-type and mutant bacteria lacking fitness genes identified from the primary screen confirmed that pic and vat deletions lead to a fitness defect during systemic infection. The pic gene has also previously been shown to be an active protease in UPEC pathogenesis. An in silico search of the UPEC CFT073 strain found uncharacterized autotransporters, including PicU, a serine protease that is homologous to the Pic protein in Shigella and enteroaggregative E. coli [50]. The picU gene was found in 22.5% of all UTI E. coli isolates tested and 25% among E. coli samples from recurring UTIs. This is in comparison to picU presence in 12% of rectal E. coli.

Other members of the SPATE family affect epithelial cells that line the urinary tract. The secreted SPATE proteins TagB, TagC, and Sha from extraintestinal pathogenic E. coli (ExPEC) are internalized by 5637 bladder epithelial cells and disrupt the actin cytoskeleton [51]. TagB and TagC reduced actin stress fibers. Sha treatment also caused a loss of actin stress fibers and resulted in actin puncta in the cytoplasm. These effects, as well as entry into the bladder cells, were found to be dependent on the serine protease activity. Thus, SPATE family proteases can play a role in bacterial colonization and contribute to systemic infection. Additional ExPEC proteases also contribute to UTIs. Using a mouse model for UTIs and global proteomic analyses, the Prc protease was shown to be needed for ExPEC virulence [52]. A UTI model through transurethral infection by a mutant strain containing a prc deletion and a control strain resulted in a colonization defect of the prc mutant. The deletion also altered outer membrane protein expression and decreased ExPEC mobility by decreasing expression of the flagellar regulon flhDC and thus reducing the ability of ExPEC to cause UTIs. The lack of Prc protease also led to an increase in the peptidoglycan hydrolase Spr that resulted in destabilization of the bacterial envelope.

Outer membrane proteins in E. coli have also been shown to be important in bacterial survival and could impact UTI pathogenesis. The OmpT protease degrades low molecular weight cationic peptides [53]. These cationic peptides from human urine, possibly containing disulfide bridges, exhibited antimicrobial activity against E. coli, S. aureus, and C. albicans. Enzymatic activity of OmpT confers resistance to these urinary peptides and could provide an advantage to E. coli growth in the urinary tract, thus increasing the risk of UTIs. Previous work with OmpT from clinical UPEC isolates shows that the protease could lend a fitness advantage to UPEC through degradation of antimicrobial peptides [54]. In 58 clinical E. coli isolates, OmpT activity was higher in strains from patients with symptomatic infections, although this activity was heterogeneous. The ompT gene was found in 89% of UPEC isolates from symptomatic infections as well. Strains with the most proteolytic activity contained the arlC gene (also called ompTp) that encodes an OmpT-like protease. While ompT is encoded on the chromosome, arlC was encoded on large plasmids that ranged in size from 150 to 195 kbp. The UPEC isolate ArlCs were an identical match to ArlC from AIEC strain NRG857c from the gut.

The establishment, growth, and colonization of UTI-causing bacteria appear to be reliant on different proteases exposed to the host environment. These proteolytic virulence factors modify host cell structural components, affect mobility, provide resistance against antimicrobial peptides, and result in a fitness advantage that could progress to UTIs.

6. Conclusion

Proteases produced by bacteria in the microbiome can be important virulence factors that facilitate disease pathogenesis in humans. These effectors provide a competitive advantage in skewing host–microbe and microbe–microbe interactions, promoting growth and colonization by pathogens. In the various microbiome sites, proteases can compromise the mucosal and cellular barriers, affecting immune signaling, targeting host defense compounds, and creating more favorable environmental factors for growth. Due to their role in many diseases, bacterial proteases and their inhibitors have potential for further study for antibacterial compounds and other therapeutics that mitigate or prevent specific diseases of the microbiome. Advances in structural modeling programs and also drug discovery platforms could help identify potential treatments. However, more research is needed to provide safe and effective inhibitors that can be used in the clinic.

Taken together, proteases from the microbiome can affect host physiology. A better understanding of their contributions to diseases of the microbiome can set the foundation for innovative treatments and better human health.

Conflicts of Interest

The author declares no conflicts of interest.

Funding

The research did not receive specific funding.