The role of site-2-proteases in bacteria: a review on physiology, virulence, and therapeutic potential

Abstract Site-2-proteases are a class of intramembrane proteases involved in regulated intramembrane proteolysis. Regulated intramembrane proteolysis is a highly conserved signaling mechanism that commonly involves sequential digestion of an anti-sigma factor by a site-1- and site-2-protease in response to external stimuli, resulting in an adaptive transcriptional response. Variation of this signaling cascade continues to emerge as the role of site-2-proteases in bacteria continues to be explored. Site-2-proteases are highly conserved among bacteria and play a key role in multiple processes, including iron uptake, stress response, and pheromone production. Additionally, an increasing number of site-2-proteases have been found to play a pivotal role in the virulence properties of multiple human pathogens, such as alginate production in Pseudomonas aeruginosa, toxin production in Vibrio cholerae, resistance to lysozyme in enterococci and antimicrobials in several Bacillus spp, and cell-envelope lipid composition in Mycobacterium tuberculosis. The prominent role of site-2-proteases in bacterial pathogenicity highlights the potential of site-2-proteases as novel targets for therapeutic intervention. In this review, we summarize the role of site-2-proteases in bacterial physiology and virulence, as well as evaluate the therapeutic potential of site-2-proteases.


Introduction
Bacteria depend on tr ansmembr ane signaling to sense and quic kl y r espond to a c hanging envir onment. Fast r esponse is particularly crucial for pathogenic bacteria, which encounter rapidly c hanging envir onment within the host thr oughout the infection pr ocess. Regulated intr amembr ane pr oteol ysis (RIP) is a signaling mechanism known to mediate transmembrane signaling in Gr am-positiv e and Gr am-negativ e bacteria (Urban 2009, Schneider and Glickman 2013, Wettstadt and Llamas 2020. RIP is mediated by a unique class of intr amembr ane pr oteases (IMP). IMP are divided into four families based on their catalytic mechanisms: The rhomboid proteases (serine proteases), the s ite-2metallo p roteases (S2P; zinc metalloproteases), Rec1 (glutamyl protease) and the presenilin/signaling pe ptide pe ptidases (aspartyl proteases). While the catalytic mechanism varies between the differ ent IMP families, structur al determination of pr ototypic members has confirmed the membrane-embedded location of the catal ytic activ e site for eac h famil y (Wu et al. 2006, Feng et al. 2007, Li et al. 2013, Manolaridis et al. 2013, Bai et al. 2015. All members ar e pr oposed to contain a hydrophilic cavity or c hannel, deliv ering water from either side of the membrane to the active site, thereby allowing hydr ol ysis of peptide bonds within the cell membr ane (Wu et al. 2006, Feng et al. 2007, Li et al. 2013, Manolaridis et al. 2013, Bai et al. 2015. Such intramembrane hydrolysis of the peptide substrate is often a critical step in cell signaling. In the present r e vie w, the r ole of S2P-mediated RIP signaling in bacteria will be discussed. The remaining IMP families have been extensively reviewed in other articles (Wolfe and Kopan 2004, Urban 2009, Sun et al. 2016, Verhelst 2017, Beard et al. 2019 ).
In bacteria, gene expression is commonly controlled at the tr anscriptional le v el. S2P-mediated RIP signaling allow r egulated gene expression in bacteria by activating an alternative sigma factor of the extracytoplasmic function (ECF) family in response to en vironmental cues . As r e vie wed in detail her e and by others, S2Pmediated RIP signaling contributes to a v ast arr ay of cellular processes, including cell polarity in Caulobacter crescentus , iron acquisition in various Gram-negative bacteria and sporulation in Bacillus subtilis (Chen et al. 2005, King-Lyons et al. 2007, Yoko y ama et al. 2021. Most notably, RIP-mediated activation of sigma factors has been linked to virulence in se v er al human pathogens (Schöbel et al. 2004, Makinoshima and Glickman 2006, Urban 2009, Sc hneider and Glic kman 2013. While r egulating a v ast arr ay of cellular processes, S2P-mediated RIP signaling generally adheres to the same general principle ( Fig. 1 ) (Chen and Zhang 2010 ). A membrane-bound anti-sigma factor, commonly a single-spanning membr ane pr otein sequesters the ECF sigma factor to the membr ane, ther eby negativ el y r egulating ECF sigma factor activity. Following the detection of an environmental signal, a primary site-1pr otease (S1P) cleav es the anti-sigma factor (site-1-cleav a ge). This primary cleav a ge triggers an intr amembr ane pr otease (IMP) to cleave the anti-sigma factor within the lipid bilayer (site-2-cleave), ther eby r eleasing the activ e σ -factor fr om the membr ane . T he RIP cascade is a widely distributed signaling pathway in the bacterial world, and variations of this blueprint ar e emer ging as the role of S2P in different bacteria are explored (Fig. 2 , Table 1 ).
All S2P contain a catalytic core region consisting of three tr ansmembr ane segments, with the conserved zinc-binding motif HExxH and the zinc-coordinating region of the LDG motif lo-Figur e 1. T he general S2P-mediated RIP paradigm. In the absence of environmental cues (left panel), the cleavage of the anti-sigma factor is inhibited, leaving the bound σ -factor in an inactive state. When extracytoplasmic stimuli are detected (right panel), a site-1-protease (S1P) cleaves a membrane-bound anti-sigma factor on the periplasmic side . T he primary cleavage triggers a secondary cleavage by a site-2-protease (S2P), resulting in release of the σ -factor (orange) from the membrane, and activation of genes involved in adaptive response.

Figur e 2.
T he role of S2P in RIP signaling mechanism in selected bacteria. (A) The Escherichia coli RseP activation of the extracytoplasmic stress r esponse serv es as a blueprint for S2P-mediated signaling. Accum ulation of unfolded outer membr ane pr oteins (OMPs) is a known str ess cue sensed by the PDZ domain of the S1P DegS. The accumulation of OMPs results in sequential cleavage of the anti-sigma factor RseA by DegS (site-1-cleavage) and RseP (site-2-cleav a ge). The RseP cut r e v eals a degron in the anti-sigma domain which is subsequently cleaved by ClpXP in the cytosol. The r esulting r elease of σ E activ ates genes involv ed in the ada ptiv e str ess r esponse. RseB, a periplasmic pr otein, negativ el y r egulates σ E activity by blocking DegS/RseP cleav a ge under non-str ess conditions . (B-F) Variations of RIP in v olving S2P are sho wn for various species . In all panels: T he suggested stimuli signal is shown in yellow, the substrate in blue, presumed S1P in red, S2P in green, sigma factors (A-D) or lipoprotein (E) in active form in orange, and additional components in grey. See the text for more details.
cated on separate transmembrane segments . T he conserved glutamate residue of the HExxH motif is suggested to activate a water molecule bound to the zinc, thereby initiating the nucleophilic attack on the peptide bond. Mutations in the conserved catalytic motifs (HExxH, LDG) has been associated with reduced protease activity of S2P in multiple species (Feng et al. 2007, Koide et al. 2007, Saito et al. 2011. Based on phylogenetic analysis, the S2P family can be divided into four subgroups (Kinch et al. 2006 ). Group I, the largest subgroup, is characterized by a variable number of PDZ domains (IPR001478, structural domain commonly found in signaling pr oteins). Gr oup III contains a cystathionine βsynthase (CBS) domain (IPR000644, structural domain commonly found in a variety of proteins), but lacks the PDZ domain. With the exception of the S2P of B. subtilis and Methanocaldococcus jannaschii , all S2P discussed in this r e vie w belongs to subgroup I. The two r emaining subgr oups (Gr oup II and IV), lac ks an y additional domains; ho w e v er, none of these putativ e pr oteases hav e been extensiv el y c har acterized (Kinc h et al. 2006 ). For mor e details on the bioc hemical and structur al featur es of S2P, the r eader is dir ected to other r e vie ws (Kr oos and Akiyama 2013, Sun et al. 2016. In this r e vie w, the r oles and mec hanisms of RIP and S2P in bacteria will be explored, starting with the implications of S2P in bacterial adaptation to a rapidly changing environment. Next, RIPmediated virulence will be r e vie wed, focusing on the role of S2P in important human pathogens. Finally, we will discuss the therapeutic potential of S2P and highlight the challenges and opportunities of targeting this unique class of proteins.

T he r ole of S2P in response to c hanges in the extracytoplasmic environment
Bacteria depend on signal transduction across the membrane to r espond a ppr opriatel y to envir onmental cues. Pr oteol ysis is a r a pid pr ocess , thus , multiple signaling pathways depend on proteases to generate an immediate response (Wettstadt and Llamas 2020 ). Following the discovery of the S2P RseP in E. coli and the S2P SpoIVFB in B. subtilis at the beginning of this millennium, se v er al S2P have been implied in adaptive response (Table 1 , Rudner et al. 1999, Yu and Kroos 2000, Kanehara et al. 2001 (Fig. 2 ). The following section will discuss the role of S2P in response to changes in the extracytoplasmic environment.

T he r ole of RseP in regula tion of the σ E -dependent stress response in E. coli
Esc heric hia coli RseP ( R egulator of s igma E , p r otease; Ec RseP), pr e viously known as Yael, is one of the most extensiv el y studied members of the S2P family and was first identified as a k e y regulator of the σ E -dependent stress response ( Fig. 2 A) (Kanehara et al. 2001, Alba et al. 2002, Kanehara et al. 2002. In E. coli , the extracytoplasmic stress response is initiated following the sequential digestion of the anti-sigma factor RseA by the S1P DegS and the S2P Ec RseP (Ades et al. 1999, Alba et al. 2002, Kanehara et al. 2002. Ec RsePmediated cutting of RseA is one of the best understood transmembrane signaling pathways involving a S2P and is viewed as the blueprint for the bacterial RIP-cascades. In the absence of stress, σ E is sequestered in the membrane by the anti-sigma factor RseA (De Las Peñas et al. 1997, Missiakas et al. 1997 (Fig. 2 A). During str ess conditions, suc h as heat, unfolded or misfolded OMPs will accumulate in the periplasm. Unfolding of OMPs results in exposure of hydrophobic amino acids that is dir ectl y r ecognized by the periplasmic PDZ domain of the S1P DegS (Walsh et al. 2003 ). Recognition of unfolded OMPs triggers DegSmediated cleav a ge of the periplasmic r egion of RseA, whic h subsequently initiates a secondary cleavage by the S2P Ec RseP (Ades et al. 1999, Alba et al. 2002, Kanehara et al. 2002, Walsh et al. 2003, Akiyama et al. 2004. Ec RseP cleaves RseA within the transmembr ane segment, ther eby r eleasing the cytoplasmic r egion of RseA in complex with σ E (Akiyama et al. 2004, Li et al. 2009 ). The protease ClpXP performs the final cleav a ge of RseA to release the σ E into the cytosol, which activates the transcription of target genes involved in the stress response (Flynn et al. 2004 ).
The RIP-mediated activation of σ E is tightly regulated at multiple le v els to ensur e that σ E activ ation is strictl y str ess-dependent ( Fig. 2 A). Most importantly, Ec RseP can only efficiently process RseA following site-1-cleav a ge (Li et al. 2009 ). Man y pr oteins involved in RIP, including both DegS and Ec RseP, harbors extracellular PDZ domains, which are thought to be important for structuring of signal complexes. It has been suggested that the Ec RseP PDZ domain acts as a size-exclusion filter, pr e v enting inter action with lar ger substr ates suc h as full-length RseA (Hizukuri et al. 2014 (Hizukuri et al. 2014, Imaizumi et al. 2022, ther eby bloc king the access of bulky substrates to the active site. Notably, the newly exposed Cterminal of DegS-cleav a ge RseA is suggested to inter act dir ectl y with the PDZ domain (Li et al. 2009 ), further highlighting the importance of this domain in substrate discrimination. In addition, the periplasmic stress sensor protein RseB function as a negative regulator of σ E activity. RseB interacts with the periplasmic region of RseA, pr esumabl y inhibiting DegS cleav a ge by masking the site-1-cleav a ge site (De Las Peñas et al. 1997, Missiakas et al. 1997, Cezairliyan and Sauer 2007, Kim et al. 2010. During stress, misfolded OMPs and lipoprotein intermediates are suggested to promote RseB r elease fr om RseA (Wollmann and Zeth 2007, Chaba et al. 2011, Lima et al. 2013. These regulation mechanisms collectiv el y ensur e that Ec RseP-mediated activ ation of σ E is strictl y stress-dependent.

SpoIVFB and endospore formation in B. subtilis
SpoIVFB (Stage IV sporulation protein FB), the S2P of B. subtilis contr olling the tr anscription factor σ K , is another well-c har acterized member of the S2P family ( Fig. 2 B) (Rudner et al. 1999, Yu and Kroos 2000, Zhou et al. 2009 ). Under harsh environmental conditions, activ el y gr owing B. subtilis differ entiates into a dormant state, known as endospores, one of the most advanced and longlasting forms of stress response found among bacteria (Tan and Ramamurthi 2014 ). The regulatory mechanisms of endospore formation are complex, involving a set of sigma factors which are active at specific time points during sporulation. The transcription factor σ K controls the final stage of the endospore formation. σ K is synthesized as an inactive precursor, pro-σ K , which is sequestered in the outer forespore membrane (Zhang et al. 1998 ). SpoIVFB cleaves the N-terminal pro-region of pro-σ K , resulting in mature σ K being released into the mother cell (Rudner et al. 1999, Zhou et al. 2009 ). Once r eleased, σ K activ ates genes needed in the final stage of endospore maturation (Fig. 2 B).
Analogous to Ec RseP cleav a ge of RseA ( Fig. 2 A), the SpoIVFB cleav a ge of pr o-σ K is tightl y r egulated ( Fig. 2 B), to make sur e that σ K is activated only at the correct developmental stage of sporulation. As discussed abo ve , regulation of Ec RseP activity largely relies on substrate priming by a S1P and the size-excluding role of the Ec RseP PDZ domain. In contr ast, the pr ocessing of pr o-σ K by SpoIVFB, which does not contain a PDZ domain, occurs independently of a S1P. BofA and SpoIVFA, two regulatory membrane proteins, form a complex with SpoIVFB in the outer for espor e membrane Losick 2002 , Zhou andKroos 2004 ). The formation of the complex is suggested to allow BofA to interact directly with the SpoIVFB active site, thereby inhibiting Pro-σ K processing (Olenic et al. 2022a , Zhou andKroos 2004 ). The serine proteases SpoIVB and CtpB ar e suggested to sequentiall y cleav e the BofA: SpoIVFA complex, thereby liberating the active SpoIVFB to cleave pro-σ K (Dong and Cutting 2003, Zhou and Kroos 2005, Campo and Rudner 2006. Importantl y, tr anscription of serine pr otease is onl y activated in the forespore after it has been completely engulfed (Campo andRudner 2006 , Tan andRamamurthi 2014 ), thereby ensuring that σ K will be released into the mother cell to activate genes at the onset of the final sporulation stage. In this sense, SpoIVB acts as the primary mediator of signaling from the forespore to σ K activation in the mother cell.
T he SpoIVFB-mediated clea v a ge of pr o-σ K differs fr om the RIPcascade paradigm in E. coli in multiple aspects. Most notably, SpoIVFB dir ectl y cleav es the transcription factor, not an antisigma factor, to produce the active form of σ K . Moreover, SpoIVFB mediated processing of pro-σ K appears to be independent of canonical S1P cleav a ge, although pr oteol ytic cleav a ge of the BofA: SpoIVFA complex is needed for SpoIVFB activation (Fig. 2 B).

S2P-mediated lysozyme resistance in Gram-positive bacteria
Lysozyme is a k e y component of the innate immune response. Lysozyme acts as a hydr olase, degr ading the glycosidic bonds between N-acetylm ur amic acid (NAM) and N-acetylglucosamide (NAG) in peptidoglycan in the bacterial cell wall, resulting in increased cell wall permeability and ultimately cell death (Ragland and Criss 2017 ). Se v er al bacteria hav e de v eloped sophisticated mechanisms to counteract the effect of lysozyme (Bera et al. 2005, Ho et al. 2011, Varahan et al. 2013, Ragland and Criss 2017, Ho and Ellermeier 2019, Pannullo and Ellermeier 2022. In B. subtilis, l ysozyme r esistance is mediated thr ough S2Pde pendent acti vation of the ECF sigma factor σ V (Ho et al. 2011, Hastie et al. 2013 ). In the absence of lysozyme, σ V activity is hindered by the anti-sigma factor RisV (Yoshimura et al. 2004, Hastie et al. 2013 ). In the presence of lysozyme, an extracellular domain of RisV dir ectl y binds l ysozyme r esulting in conformational changes within the anti-sigma factor, thereby initiating the RIPcascade (Hastie et al. 2014, Hastie et al. 2016 ). The signal peptidases SipS and SipT ar e suggested to cleav e the extr acellular region of RisV (Hastie et al. 2014, Castro et al. 2018. Following this primary cleav a ge, the S2P RasP ( r egulator of s igma p rotease, Bs RasP) cleav es RisV, ther eby r eleasing the cytoplasmic r egion of RisV containing σ V (Hastie et al. 2013 ). It is presumed that cytosolic proteases further process RisV, ho w ever, such proteases have not yet been identified. Once released into the cytosol, σ V activates genes involved in lysozyme resistance, including oatA and dltA . OatA and DltA modify cell wall properties and c har ge thr ough O-acetylation of the peptidoglycan and D-alanylation of teichoic acids, r espectiv el y (Guariglia-Or opeza and Helmann 2011 , Ho et al. 2011 ).
A similar cascade is observed in the opportunistic pathogens C. difficile and E. faecalis . Deletion of RasP ( Cd RasP) and RseP ( Efs RseP, also known as Eep) the S2Ps of C. difficile and E. f aecalis, r espectiv el y, r esults in decreased σ V activity and lysozyme resistance (Varahan et al. 2013, Pannullo and Ellermeier 2022, Rouchon et al. 2022 ). In C. difficile , Cd RasP mediates σ V activation by cleaving RsiV in response to lysozyme in a dose-dependent manner (Pannullo and Ellermeier 2022 ). C. difficile σ V regulates the expression of se v er al genes in r esponse to l ysozyme, including the peptidogl ycan deacetylase PdaV, whic h r emov es acetyl gr oups fr om peptidoglycan in the cell wall (Ho et al. 2014, Woods et al. 2016, Kaus et al. 2020, P arthasar athy et al. 2021. In this bacterium, deacetylation of the peptidoglycan results in reduced affinity for lysozyme, ther eby incr easing l ysozyme r esistance. Inter estingl y, deletion of pd aV results in only a modest increase in lysozyme susceptibility. Ho w e v er, the combined loss of pdaV and another deacetylase, pgdA , resulted in a 1 000-fold reduction in l ysozyme r esistance, suggesting that these two deacetylases are redundant (Kaus et al. 2020  . It is possible that a similar redun-dancy as observed in C. difficile , is found in E. faecalis , or that other σ V -dependent genes also contribute to the decrease in lysozyme r esistance observ ed for the E. f aecalis str ains depleted of RseP and σ V . T hus , the S2P-mediated activation of σ V dependent lysozyme r esistance a ppears highl y conserv ed in B. subtilis , C. difficile, and E. faecalis . To date, no S1P has been identified for RsiV in E. faecalis or C. difficile . Ho w e v er, a signal peptidase cleav a ge site has been identified in RsiV for both species, suggesting that the S1P is a so far unknown signal peptidase (Hastie et al. 2014 , Pannullo and. In E. faecalis , σ V is also suggested to contribute to a general stress response, as rseP and sigV significantly atten uated survi val in response to multiple stressors, including heat, acid, and ethanol (Benachour et al. 2005, Varahan et al. 2013 ). Similar observations have been made in Enterococcus f aecium , wher e m utations within rseP r esults in a 6-8-fold incr ease in l ysozyme susceptibl y as well as a r eduction in desiccation tolerance (Reinseth et al. 2021 ). A more general role in stress response has so far not been observed for RasP and σ V in B. subtilis or C. difficile.

RseP and iron acquisition in Gram-negative bacteria
Iron acquisition is a k e y factor for colonization and infection of se v er al pathogens (Sheldon et al. 2016 , Llamas andSánc hez-Jiménez 2022 ). For pathogenic bacteria, iron resources in the envir onment ar e limited, as most of the host iron is present in high affinity iron binding complexes, such as hemoglobin and transferrin. To circumvent these limitations, bacteria have developed se v er al sophisticated iron acquisition systems, including the production of sider ophor es and heme utilization (Sheldon et al. 2016 , Llamas andSánchez-Jiménez 2022 ). To ensure rapid adaption to c hanging envir onment, ir on acquisition is r egulated by ECF sigma factors in se v er al Gr am-negativ e bacteria.
In P. aeruginosa , an opportunistic human pathogen causing a range of different infections, particularly in immunocompromised and cystic fibrosis patients, the involvement of ECF sigma factors in iron acquisition has been extensively reviewed (Llamas et al. 2014, Che v alier et al. 2019, Llamas and Sánc hez-Jiménez 2022. Inter estingl y , the S2P RseP , also known as MucP, is r equir ed for complete pr oteol ysis of the anti-sigma factor and thus ECF sigma factor activity, in at least five pathways related to iron acquisition in P. aeruginosa : The p y over dine, ferric hr ome and ferrioxamine sider ophor e pathwa ys , and the Has and Hxu heme pathways (Dr a per et al. 2011, Bastiaansen et al. 2014, Bastiaansen et al. 2015, Otero-Asman et al. 2019. S2P mediated iron acquisition appears to be conserved in Pseudomonas species, as the deletion of r seP ad ditionall y er adicates the activity of m ultiple ir on acquisition pathways in P. putida (Bastiaansen et al. 2014 ). Inter estingl y, se v er al of the RseP-r egulated ir on acquisition pathways have also been implied in the pathogenicity of P. aeruginosa (Damron et al. 2016, Minandri et al. 2016, Cai et al. 2022. Most notably, the deletion of the sigma factor PvdS, which regulates p y over dine production in a RseP-dependent manner, resulted in attenuated virulence in mouse lung infection (Wilderman et al. 2001, Dr a per et al. 2011 ) . In addition to p y over dine production, σ PvdS r egulated the expr ession of the major virulence factors exotoxin A and Pr pL, ther eby having a dual role in P. aeruginosa virulence (Wilderman et al. 2001, Hunt et al. 2002, Ochsner et al. 2002, Minandri et al. 2016 ). The role of S2P and the RIP-cascade in P. aeruginosa virulence will be further discussed below. S2P-mediated iron acquisition has also been implied in other gr am-negativ e species . In B . bronchiseptica , a pathogen colonizing the r espir atory tr act of animals and humans, the S2P HurP ( h eme u tilization r eceptor p rotease) is essential for heme utilization through the expression of the outer membrane receptor BhuR (King-Lyons et al. 2007 ). The complete regulation mechanism has not been fully elucidated; ho w ever, it is suggested that HurP cleaves the anti-sigma factor HurR in a heme-dependent manner. The modification of HurR is proposed to release the sigma factor HurI, which subsequently regulates the cellular response to the iron depleted environment and heme uptake . T he S1P of this proposed RIP cascade has not yet been identified (King-Lyons et al. 2007 ). Curiousl y, hurP complements m utations in the S2P rseP in V. cholerae while rseP from E. coli complements hurP deletions in B. bronchiseptica (King-Lyons et al. 2007 ), indicating that substr ate r ecognition and pr ocessing m ust be highl y conserv ed in these species. To date, RseP has not been implied in iron acquisition in V. cholerae , ho w ever, Ec RseP w as recently found to activate the ECF sigma factor FecI in response to ferric-citrate (Yoko y ama et al. 2021 ).

RasP and resistance to antimicrobials in Bacillus spp .
Se v er al bacteria utilize alternative sigma factors to regulate genes in response to stressors such as antibiotics and antimicrobial peptides (Woods and McBride 2017 ). In B. thuringiensis , a Gr am-positiv e bacterium commonly found in the soil, resistances to selected β-lactams are controlled by the ECF sigma factor σ P (Fig. 2 C). Upon activ ation, σ P r egulates expr ession of m ultiple β-lactamases and penicillin binding proteins (PBP) , providing B. thuringiensis with an arsenal of pr otection str ategies . T he activity of σ P is hindered by the anti-sigma factor RsiP, which, in response to selected β-lactams, is successiv el y cleav ed by the S1P SipP and the S2P RasP . It is suggested that penicillin-binding protein P (PbpP) acts as a sensor for β-lactams, somehow activating the S1P cleavage of the RIP cascade (Nauta et al. 2021 ). pbpP , sipP , and rasP increase susceptibility to ampicillin and cefoxitin , Nauta et al. 2021, underpinning the RIP cascade's role in σ P mediated βlactam resistance. Similar cascades is observed in B. subtilis , where σ W participates in response to a variety of cell wall inhibiting antibiotics and antimicrobial peptides (Cao et al. 2002, Pietiäinen et al. 2005, Butcher and Helmann 2006. Activation of σ W depends on the RIP-mediated cleav a ge of the anti-sigma factor RsiW by the S1P PrsW and the S2P RasP (Schöbel et al. 2004, Ellermeier and Losick 2006, Heinrich and Wiegert 2006. Although primarily considered as an antibiosis regulon, the B. subtilis σ W responds to v arious str essors compr omising cell w all integrity (P etersohn et al. 2001, Schöbel et al. 2004, Pietiäinen et al. 2005, Butcher and Helmann 2006, Ellermeier and Losick 2006. In contrast, B. thuringiensis σ P is activated only by a specific subset of β-lactams, not other cell wall-targeting antibiotics, suggesting that σ P is not activated in response to general cell wall stress . It should be noted that σ P also has been linked to r esistance a gainst βlactams in Bacillus cereus and Bacillus anthracis (Ross et al. 2009. RasP is conserved in these closel y r elated human pathogens, suggesting a role of RIP-mediated antibiotic resistance in these species. Ho w e v er, the r ole of RIP-mediated r esistance to antibiotics in B. cereus and B. anthracis has not been extensiv el y studied. In this context, it is also interesting to note that a variation to the canonical RIP cascade involving the S2P RasP in B. subtilis was r ecentl y r eported (Brunet et al. 2022 ). In this system, whic h is involved in sensing cell wall homeostasis, the anti-sigma factor RsgI is constitutiv el y cleav ed b y a S1P. Ho w e v er, the two cleava ge pr oducts r emain associated as long as the cell wall is intact, thereby hindering site-2 cleavage by RasP. When a cell wall defect occurs, this is sensed by the extracellular domain on the cleaved RsgI, resulting in release of the cleavage product, follo w ed b y site-2 cleav a ge by RasP to r elease activ e σ I . This sigma factor activ ates genes needed to repair cell wall defects and has also been linked to changes in sensitivity to β-lactams (Patel et al. 2020, Brunet et al. 2022 ).

T he r ole of S2P in bacterial pa thogenesis
As discussed abo ve , S2P-mediated activation of ECF sigma factors plays a k e y r ole in the quic k ada ptation to extr acytoplasmic stimuli in many bacteria. Rapid adaptation to changing environments is crucial for establishing infections, ho w e v er S2P-mediated RIP have also been implied in the direct activation of virulence related genes (Fig. 2 , Table 1 ). In the following section the role of S2P in prominent human pathogens will be discussed.

MucP and mucoid production in P. aeruginosa
Pseudomonas aeruginosa is an opportunistic pathogen causing a wide range of nosocomial infections. Most notably , P . aeruginosa is a major cause of morbidity among patients with cystic fibrosis. Clinical P. aeruginosa isolated from patients with cystic fibrosis gener all y exhibit mucoid phenotype due to overproduction of alginate (Pritt et al. 2007 , Hogardt andHeesemann 2010 ). It is suggested that alginate production contributes to protecting the bacteria a gainst v arious str essors, including antibiotics and o xidati ve str ess, ther eby pr omoting persistence in the lungs and thereby virulence . In P. aeruginosa , genes in volv ed in alginate pr oduction is controlled by the ECF sigma factor σ 22 , also known as σ AlgT or σ AlgU (Wood et al. 2006 , Wood andOhman 2009 ). Activation of σ 22 is regulated in a RIP-depended manner in response to envelope str ess, highl y r esembling the RseP-mediated activ ation of the extr acytoplasmic str ess r esponse in E. coli (Fig. 2 D) (Wood et al. 2006, Wood and Ohman 2009, Damron and Goldberg 2012. It has indeed been suggested that alginate production plays a part in a general str ess r esponse, as σ 22 also r egulates genes involv ed in pr otection a gainst extr acytoplasmic str ess (Wood et al. 2006 , Wood andOhman 2009 ).
Under non-stress conditions, the anti-sigma factor MucA sequesters σ 22 in the membr ane, ther eby inactiv ating the sigma factor (Li et al. 2019 ) (Fig. 2 D). The PDZ domain of P aeruginosa DegS ( Pa DegS), also known as AlgW, is suggested to recognize Cterminal residues of the envelope protein MucE and possibly other OMP, whic h ar e accum ulated under cell wall str ess (Qiu et al. 2007 , Cezairliyan andSauer 2009 ). This interaction initiates Pa DegS mediated cleav a ge of MucA, whic h is pr oposed to subsequentl y trigger a secondary cleav a ge by Pa RseP (Qiu et al. 2007, Cezairliyan and Sauer 2009, Wood and Ohman 2009, Damron and Hongwei 2011, Damr on and Goldber g 2012. MucA is suggested to be further processed by proteases such as ClpXP in the cytosol, finally resulting in the release of activated σ 22 (Qiu et al. 2008 ). Under non-stress conditions, MucB, the equivalent to RseB in E. coli , binds to MucA, thereby inhibiting RIP-mediated cleavage of the antisigma factor (Fig. 2 A and D)  It is interesting to note that clinical Pseudomonas isolates from cystic fibrosis patients often contain mutation within MucA, commonl y r esulting in a C-terminal truncation of the anti-sigma factor and subsequent incr eased σ 22 activity (Br a gonzi et al. 2006, Pulcrano et al. 2012. In E. coli , truncation of RseA results in Ec RseP cutting of RseA independent of site-1 cleav a ge (Li et al. 2009, Hizukuri et al. 2014. A similar mechanism is proposed for the cleav a ge of truncated MucA in P. aeruginosa , explaining the constitutiv e m ucoid phenotype commonl y observ ed for MucA m utants (Damr on and Goldber g 2012 ). As discussed abov e, RseP of P. aeruginosa has also been implicated in iron-uptake and regulation of virulence determinators exotoxin A and PrpL. The combined role of RseP-mediated regulation of iron uptake, alginate production and expression of virulence enhancing genes, makes it tempting to speculate that a rseP deletion would have a striking effect on P. aeruginosa pathogenicity. Ho w e v er, the r ole of pseudomonal RseP has not yet been investigated in infection models in vivo.

Rip1 and virulence in M. tuberculosis
Mycobacterium tuberculosis is an obligate human pathogen and a leading cause of infection worldwide (Antimicrobial-Resistance-Collaborators 2022 ). The complex composition of the M. tuberculosis cell envelope makes these infections difficult to treat and is known to play a k e y role in the persistence and virulence of this pathogen (Garcia-Vilanova et al. 2019, Maitra et al. 2019. Transcriptional analysis in the presence or absence of detergents revealed that the M. tuberculosis S2P Rip1 regulates genes involved in lipid metabolism in response to a c hanging envir onment (Makinoshima and Glickman 2005 ). In line with this evidence, rip1 displays altered cell morphology and loss of ability to cord (Makinoshima and Glickman 2005 ), which is an established virulence tr ait r el ying on a gl ycolipid kno wn as the cor d-factor in the mycobacterial cell-en velope . When c hallenged in an aer osol m urine model, rip1 sho w ed impair ed gr owth during both the acute and c hr onic phase of infection (Makinoshima and Glickman 2005, Schneider et al. 2014, Buglino et al. 2021. Specifically, by 22 weeks, the bacterial titers from the lungs were 10 000-fold lo w er for rip1 compared to wild-type, while rip1 was completely abolished from the liver (Makinoshima and Glickman 2005 ), and by 43 w eeks, rip1 w as completely cleared from the lungs, showing no detectable colony forming units (CFU) in the majority of the mice (Schneider et al. 2014 ). The deletion phenotype was rescued when complemented with the wild-type Rip1, showing that attenuated virulence is Rip1 dependent (Makinoshima and Glickman 2005 ).
Although the se v er e atten uation of M. tuber culosis virulence caused by the single gene deletion of rip1 is striking, the mechanism of Rip1 in M. tuberculosis virulence remains somewhat elusive. In fact, Rip1 cuts four independent anti-sigma factors controlling the activity of σ K , σ L , σ M and σ D (Sklar et al. 2010, Schneider et al. 2014. Inter estingl y, sigKLM does not attenuate bacterial growth in aerosol murine model, suggesting that the Rip1dependent virulence observed for rip1 is independent of these pathways (Schneider et al. 2014 ). This points to a k e y role of σ D in the virulence mec hanism, whic h notabl y contr ols expr ession of multiple virulence factors, including fbpA and the resuscitation promoting factor RpfC (Raman et al. 2004, Calamita et al. 2005. T hus , it was surprising to note that single deletion sigD only caused modest decrease in virulence compared to the Rip1 (Raman et al. 2004, Calamita et al. 2005. So far, the effect of the quadruple mutant ( sigKLMD ) has not been investigated in vivo.
Adding to the complexity of Rip1s role in virulence, a recent study r e v ealed that Rip1 contributes to defense against host-imposed stressors (Buglino et al. 2021 ). When challenging rip1 with v arious str essors, Rip1 was found to be essential for protection against metal and nitrosative stress. Most importantly, the observed attenuated growth of rip1 in acute infection in mice was r e v ersed in the absence of nitric oxide production, suggesting that Rip1 dir ectl y contributes to pathways defending M. tuberculosis a gainst host-pr oduced nitric oxide (Buglino et al. 2021 ). Interestingly, no Rip1-dependent defects were observed when challenging the rip1 mutant with starvation, lysozyme, iron deficiency or an acidic environment (Buglino et al. 2021 ), as pr e viousl y observ ed for other S2P mutants (as discussed above).

RseP and pheromone production in E. faecalis
S2P was gener all y belie v ed onl y to r egulate tr anscription factor activity. Ho w e v er non-anti-sigma factor substr ates hav e been identified for S2P in multiple species, with E. faecalis being the earliest example (Chen et al. 2005, Matson and DiRita 2005, Mukherjee et al. 2009, Saito et al. 2011, Sc hilc her et al. 2020. Although considered commensal enterococci have emerged as important healthcare-associated pathogens in the last decades, causing a wide range of infections, including urinary tract infections (UTIs), endocarditis, and bacteremia (Cattoir andLeclercq 2013 , Reinseth et al. 2019 ). The S2P of E. faecalis , Efs RseP, also known as Eep ( e nhanced e xpression of p her omone), incr eases the pr oduction of multiple sex pheromones, including cCF10, cAD1 and cPD1 , An and Clewell 2002, Chandler et al. 2005 , Chandler and Dunn y 2008 ), whic h ar e critical for conjugation of plasmids between bacterial cells. A model has been proposed in which lipopr otein pr ecursor is sequentiall y digested by a type II signaling peptidase and Efs RseP, before a ABC transporter actively transport the resulting mature pheromones outside the cell (Varahan et al. 2014 ) (Fig. 2 E). S2P involv ement in pher omone matur ation has also been suggested in se v er al species of Streptococcus , L. monoc ytogenes , and S. aureus , indicating that S2P-mediated pheromone processing may be a conserved mechanism among some gr am-positiv e bacteria (Denham et al. 2008, Chang et al. 2011, Pér ez-P ascual et al. 2015, Xayarath et al. 2015, Cheng et al. 2020, Sc hilc her et al. 2020. In addition to its role in pheromone processing, Efs RseP has been suggested as a k e y factor in E. faecalis virulence (Frank et al. 2012, Frank et al. 2013. When challenged in a rabbit endocarditis model, rseP was se v er el y attenuated, showing a 4-log 10 decr ease in CFU r ecov er ed fr om the heart v alv es compar ed to the wild-type . T he attenuation defects observed for rseP could be rescued by in trans expression of rseP (Frank et al. 2012 ). Moreov er, rseP r esulted in a reduced bacterial load r ecov er ed fr om the kidneys in a murine catheter-associated UTI model (Frank et al. 2013 ). It should be noted that the defects in virulence observed for the rseP is likely not attributed to loss of plasmid conjugation, as both in vivo studies were performed using E. faecalis OG1RFa well c har acterized labor atory str ain whic h almost completel y lacks mobile genetic elements (Bourgogne et al. 2008 ). Efs RseP dependent virulence is ther efor e most likel y not r elated to lac k of sex pheromones maturation. On the other hand, σ V may be involv ed. As discussed abov e, Efs RseP induces l ysozyme r esistance through RIP-mediated activation of the ECF σ V (Varahan et al. 2013 ). Indeed, sigV attenuated E. faecalis virulence in a murine UTI and systemic infection (Le Jeune et al. 2010 ). This indicates that, at least to some extent, reduction in σ V activation may contribute to the reduced virulence observed for strains deficient in Efs RseP. Ne v ertheless, the exact mec hanisms underl ying Efs RsePmediated virulence remains to be confirmed.

Eep and virulence in S. aureus
S. aureus is a major human pathogen causing a wide range of both nosocomial and community acquired infections. Small linear peptides are suggesting a k e y role in multiple cellular processes in Gr am-positiv e bacteria, including signaling, competence de v elopment, and virulence (Håvarstein et al. 1995, Thoendel and Horswill 2010, Chang et al. 2011, Xayarath et al. 2015. Similar to what was discussed above for E. faecalis , the S. aureus S2P ( Sa Eep) is involved in maturation of multiple small linear peptides, including the sex pheromone cAM373 (Cheng et al. 2020, Schilcher et al. 2020. A pr oteomic anal ysis sho w ed that eep in S. aureus affected the levels of more than 55 proteins, including a significant decrease of se v er al pr oteins involv ed in bacterial adhesion (i.e. SpA, SasG and FnbA) (Cheng et al. 2020 ). Indeed, eep sho w ed reduced adhesion to human epithelial cells two hours post-infection. Ov er expr ession of Spa, SasG, and FnbA increased adhesion in the eep strain, suggesting that the reduced ability to adhere is dir ectl y linked to r educed pr oduction of important adhesion pr oteins in the m utant (Cheng et al. 2020 ). Furthermor e, when c hallenging a m urine blood infection model, eep sho w ed significantly decreased virulence, and survival (Cheng et al. 2020 ). In this study, CFU counts r etrie v ed fr om the liv er, but not the kidneys, could be r estor ed to wild-type le v els when SpA was ov er expr essed in the eep m utant, suggesting that Sa Eep may affect other aspects of infection besides adhesion (Cheng et al. 2020 ).
The ECF factor σ S has been suggested to play a role in S. aureus virulence (Shaw et al. 2008, Miller et al. 2012. Inter estingl y, a S. aureus homolog of the S1P PrsW in B. subtilis is suggested to be involv ed in σ S activ ation (Krute et al. 2015 ). These observ ations, taken together with the fact that Sa Eep controls the expression of a v ast r ange of pr oteins, make it tempting to speculate that Sa Eep and RIP-mediated activation of ECF factors play roles in S. aureus virulence. Ho w e v er, the potential r ole of Eep in S. aureus ECF factor activity remains to be elucidated.

RseP and virulence regulation in V. cholerae
As r e vie wed abov e, RIP is a common mechanism for activating transcriptional factors by releasing them from the membrane. In contrast, TcpP, a transcription factor needed for activation of the master virulence regulator ToxT in V. cholerae , is active when bound to the membrane but inactivated following proteol ysis (Fig. 2 F). ToxT dir ectl y activ ates m ultiple virulence associated genes in V. cholerae , including genes encoding c holer ae toxin and the toxin-co-regulated pilus (Matson et al. 2007 ). Under non-virulence inducing conditions, TcpP is degraded in a RIPdependent manner, thereby negativ el y r egulating the expr ession of toxT and downstream virulence associated genes DiRita 2005 , Teoh et al. 2015 ). In this process, the periplasmic protein Tsp (tail-specific protease) performs the primary cleav a ge of TcpP, while the S2P cleav a ge is performed by V. cholerae RseP ( Vc RseP, also known as Yael) DiRita 2005 , Teoh et al. 2015 ). During virulence promoting conditions, another protein known as TcpHis suggested to interact with TcpP, thereby protecting it from RIP mediated clea vage , thus allowing toxT expression (Beck et al. 2004 , Matson andDiRita 2005 ).
Another le v el of r egulatory complexity in this system is added by another membrane-bound transcription factor known as ToxR and its effector protein ToxS (Fig. 2 E). ToxR is thought to addi-tionally enhance the TcpP mediated activation of toxT (Fig. 2 E

Targeting S2P in the context of antimicr obial ther apy
Despite the evident role of S2P in bacterial physiology and virulence, only a limited number of inhibitors have been shown to tar get S2P. Se v er al c hallenges hamper the de v elopment of specific S2P inhibitors. Most notably, S2P are conserved among all kingdoms of life (Kinch et al. 2006 ), making adverse interaction between an inhibitor and the host S2P a potential concern. Increased biochemical and structural understanding of S2P may contribute to de v eloping specific inhibitors with minimal off-target effects. Nov el pr otease inhibitors ar e gener all y r ationall y designed based on the known structure of the protease itself, or by mimicking the interaction with known protease substrates and/or inhibitors. To date, only three S2P has been fully structurally characterized: Ec RseP, RseP of Kangiella koreensis and the S2P of M. jannaschii (Feng et al. 2007, Imaizumi et al. 2022. Despite these challenges, S2P remains an attractive drug target, with a limited number of inhibitors been identified.
Most notably, a subgroup of bacteriocins known as the LsbB famil y, tar get the S2P of selected gr am-positiv e bacteria (Uzelac et al. 2013, Ovc hinnik ov et al. 2014, Ovc hinnik ov et al. 2017, Kr anjec et al. 2021. Bacteriocins are small, ribosomally synthesized peptides produced by bacteria that inhibit the growth of other bacteria in competition for nutrients and ecological niches . T hese antimicr obial peptides hav e pr edominantl y been exploited as food pr eserv ativ es; ho w e v er, bacteriocins ar e r eceiving incr eased scientific interest as promising alternatives and/or complements to antibiotics due to their potent activity against multiple human pathogens (Cotter et al. 2013, Soltani et al. 2021 ). Among the members of the LsbB famil y, enter ocin EJ97 (EntEJ97) and enterocin K1 (EntK1) show the greatest therapeutic potential. EntEJ97 and EntK1 show potent activity a gainst v ancomycin-r esistant enterococci (VRE) at nanomolar concentrations in vitro (Ovchinnikov et al. 2017, Reinseth et al. 2021. Notably, EntEJ97 and EntK1 are autoclav able and r etain antimicr obial activity in blood, highlighting their ther a peutic potential a gainst systemic infections (Reinseth et al. 2021 ). The potent activity of EntK1 and EntEJ97 have been attributed to a specific interaction with the enterococcal S2P (Ovc hinnik ov et al. 2017 , Kristensen et al. 2022 ). As discussed abo ve , enterococcal RseP is suggested to be important for stress tolerance and virulence in enterococci (Benachour et al. 2005, Frank et al. 2012, Frank et al. 2013, Varahan et al. 2013, Reinseth et al. 2021. Indeed, while spontaneous EntK1-and EntEJ97 resistant mutants can emerge due to mutations in rseP , these mutants show a 6-8-fold increase in lysozyme susceptibility, reduced tolerance to heat and a significant reduction in desiccation tolerance (Ovc hinnik ov et al. 2017, Reinseth et al. 2021. T hus , resistance development seems to come at a high cost for the bacteria cell.
EntK1 and EntEJ97 are particularly attractive bacteriocins, not only because they act on VRE, but also because they are synthesized without a N-terminal leader sequence and contain no post-translational modifications (Ovchinnikov et al. 2017 , Kris-tensen et al. 2022 ). This makes EntK1 and EntEJ97 an ideal starting point for bioengineering of novel S2P targeting antimicrobials. Indeed, the construction of a EntK1/EntEJ97 hybrid bacteriocin, designated Hybrid 1, exhibited a novel inhibition spectrum and impr ov ed activity compar ed to the par ental bacteriocins (Kr anjec et al. 2021 ). It was r ecentl y shown that the antimicrobial activity of EntK1 is dependent on its interactions with the PDZ-domain and Asn359 in the extended LDG motif of RseP in E. faecium (Kristensen et al. 2022 ). Inter estingl y, these r egions hav e been implicated in substrate interaction for other bacterial S2P (e.g Ec RseP, B. subtilis SpoIVFB) (Olenic et al. 2022b, Koide et al. 2008, Zhang et al. 2013, Akiyama et al. 2015, Akiyama et al. 2017, and are conserved among se v er al of the human pathogens discussed in this r e vie w. While these bacteriocins hold significant potential as novel antimicr obials tar geting S2P, the in vivo efficacy and potential cytotoxicity of these peptides remain to be investigated in detail. Besides antimicrobial peptides, two commerciall y av ailable pr otease inhibitors hav e been shown to have a surprising effect on S2P acti vity. Nelfinavir (Viace pt), an FDA-a ppr ov ed pr otease inhibitor used in HIV ther a py, is suggested to inhibit human S2P. The drug results in increased accumulation of the human S2P substrates AFT6 and SREBP1 (Guan et al. 2011, Guan et al. 2015. In humans, S1P and S2P ar e involv ed in lipid metabolism and unfolded pr otein r esponse (endoplasmatic r eticulum str ess) (Dan yuk ov a et al. 2021 ). Inter estingl y, nelfinavir shows pr omising anti-cancer pr operties, and the ther a peutic potential of nelfinavir is curr entl y being investigated for several types of cancers (Gills et al. 2007, Gantt et al. 2013, Guan et al. 2015, Kawabata et al. 2021. Nelfinavir and nelfinavir-analogs inhibit the pr oteol ytic activity of Mj S2P, the S2P in M . jannaschii, in vitro, indicating that this drug may be a potent inhibitor of bacterial S2P (Guan et al. 2015 ). Ho w e v er, the antimicrobial potential of nelfinavir is largely unexplored.
Additionally, K ono valo va and colleagues recently identified batimastat, a broad-spectrum protease inhibitor known to target eukaryotic metalloproteases, as a potent inhibitor of Ec RseP (Konov alov a et al. 2018 ). Structur al anal ysis of the Ec RseP in complex with batimastat r e v ealed that the binding mode of batimastat to Ec RseP depends on residues and regions known to be crucial for interaction with the native substrate RseA (Fig. 1 A) (Imaizumi et al. 2022 ). Notably, as also seen for RseP: EntK1 interaction, this includes the extended LDG domain (Imaizumi et al. 2022, Kristensen et al. 2022 ). Ho w e v er, batimastat was identified as Ec RseP inhibitor in a screen using an efflux-deficient ( tolC ) strain, and batimastat have no inhibitory effect on the wild-type efflux-intact strains (K ono valo va et al. 2018 ). This likely hampers the therapeutic potential of batimastat in antimicrobial infections.

Future perspective and concluding remarks
S2P-mediated RIP is a well conserved signaling mechanism common to all forms of life. Notably, S2P and RIP play a k e y role in physiology and virulence in multiple human pathogens, including enter ococci, sta phylococci, and M. tuberculosis . The r ecent advances in the field have increased our understanding of the various molecular mec hanisms underl ying S2P-mediated signaling and has raised the exciting possibility that S2P may serve as novel antimicr obial tar gets. So far, the antimicr obial peptides of the LsbB bacteriocin family are the only known antimicrobials specificall y tar geting S2P; ho w e v er, as our structur al and bioc hemical insights into S2P incr ease, mor e inhibitors are likely to be developed. Tools like Alphafold and artificial intelligence (AI) are revolutionizing the field of drug discovery by pr oviding ne w ways to pr edict pr otein structur es and identify potential inhibitors . T he use of these technologies is expected to grow in the future as they will help r esearc hers to understand the structure of S2P better and screen large libraries of potential inhibitors for those that effectiv el y bind and inhibit S2P.

Ac kno wledgments
The funder had no role in the conceptualization, data collection, or inter pr etation, or the decision to submit the work for publication.

Funding
This study was financed by the Research Council of Norway thr ough pr oject number 275190.