Plasmid-encoded toxin of Escherichia coli cleaves complement system proteins and inhibits complement-mediated lysis in vitro

Plasmid-encoded toxin (Pet) is an autotransporter protein of the serine protease autotransporters of Enterobacteriaceae (SPATE) family, important in the pathogenicity of Escherichia coli. The pet gene was initially found in the enteroaggregative E. coli (EAEC) virulence plasmid, pAA2. Although this virulence factor was initially described in EAEC, an intestinal E. coli pathotype, pet may also be present in other pathotypes, including extraintestinal pathogenic strains (ExPEC). The complement system is an important defense mechanism of the immune system that can be activated by invading pathogens. Proteases produced by pathogenic bacteria, such as SPATEs, have proteolytic activity and can cleave components of the complement system, promoting bacterial resistance to human serum. Considering these factors, the proteolytic activity of Pet and its role in evading the complement system were investigated. Proteolytic assays were performed by incubating purified components of the complement system with Pet and Pet S260I (a catalytic site mutant) proteins. Pet, but not Pet S260I, could cleave C3, C5 and C9 components, and also inhibited the natural formation of C9 polymers. Furthermore, a dose-dependent inhibition of ZnCl2-induced C9 polymerization in vitro was observed. E. coli DH5α survived incubation with human serum pre-treated with Pet. Therefore, Pet can potentially interfere with the alternative and the terminal pathways of the complement system. In addition, by cleaving C9, Pet may inhibit membrane attack complex (MAC) formation on the bacterial outer membrane. Thus, our data are suggestive of a role of Pet in resistance of E. coli to human serum.

In fact, the SPATEs EspP, Pic and Sat can cleave diverse complement system proteins of the classical (CP), the alternative (AP) and the lectin (LP) pathways (Orth et al., 2010;Abreu et al., 2015;Abreu et al., 2016;Freire et al., 2022).The complement system is an important arm of the innate immunity composed of a set of proteins that can be activated in a sequential enzymatic cascade, playing an important role in the defense against Gram-negative and Gram-positive bacteria (Joiner et al., 1984;Bhakdi et al., 1987;Nesargikar et al., 2012;Berends et al., 2014;Bjanes and Nizet, 2021).
Considering the presence of the pet gene in E. coli strains causing extraintestinal infections, and the proteolytic activity of some SPATEs of E. coli on proteins of the complement system, this study investigated whether Pet can also contribute to serum resistance in vitro.According to our data, Pet can potentially interfere with the alternative pathway of complement system activation and the formation of important by-products by cleaving key components of the cascade.In addition, Pet can also inactivate the terminal pathway by targeting C9, thus preventing C9 polymerization and lytic pore formation.As the ability to circumvent lysis by the complement system facilitates bacterial survival in the bloodstream, Pet may also play an important role in the pathogenesis of sepsis caused by E. coli.

Proteins and bacterial strains
Pet and Pet S260I were obtained from culture concentrated supernatants of E. coli HB101(pCEFN1) and HB101(pCEFN2), respectively (Eslava et al., 1998;Navarro-Garcıá et al., 1999).pCEFN1 corresponds to the pet gene from EAEC 042 cloned into pSPORT1, while pCEFN2 resulted from a site directed mutagenesis that replaced the serine residue present in the catalytic triad and in the serine protease motif with an isoleucine, inactivating Pet proteolytic action.Concentrated supernatant of HB101(pSPORT1) was employed as negative control, as previously described (Abreu et al., 2015).

Proteolytic activity of Pet on complement components
To evaluate the proteolytic activity of Pet on complement components, concentrated supernatants (1 µg) were incubated with purified complement molecules for 5 and 24 h at 37°C, and cleavage products were analyzed by immunoblotting using specific antibodies as described in previous works (Orth et al., 2010;Abreu et al., 2015;Freire et al., 2022).

C9 polymerization assay with ZnCl 2 catalyst
Since C9 is cleaved by Pet, we performed new assays to evaluate if this protease would impair C9 polymerization.For that, ZnCl 2induced C9 polymerization was performed as previously described (Tschopp, 1984;da Silva et al., 2015;Conde et al., 2016).Firstly, concentrated supernatants of HB101(pCEFN1) (1, 2.5 and 5 µg) or HB101(pCEFN2) (5 µg) were pre-incubated with 3 µg of C9 for 40 min at 37°C in 20 mM Tris-HCl buffer (pH 7.2).After preincubation, 50 µM of ZnCl 2 , diluted in the same buffer, were added to the reactions, and incubated for 2 h at 37°C.A control reaction containing 3 µg of C9 and 50 µM of ZnCl 2 in 20 mM Tris-HCl buffer (pH 7.2) was also included in the assay.A second reaction was carried out to initially induce the formation of C9 polymers for 2 h followed by an incubation of 5 µg of the HB101(pCEFN1) concentrated supernatant for 40 min.This second reaction was carried out to verify if the serine protease could degrade the C9 polymers previously formed in the presence of ZnCl 2 .

E. coli DH5a resistance assay in human serum pre-treated with Pet
To evaluate the capacity of E. coli DH5a to survive in Pet-treated human serum, assays were performed as previously described using 50% of commercial human serum and a final reaction volume of 200 µL (Freire et al., 2022).Considering that C3 usual concentration in human serum is 1500 µg/mL (Barnum, 2018), approximately 300 µg of protein were needed to perform the assay.
After preparing the human serum reactions, 20 mL of E. coli DH5a inoculum (with an OD 600 nm of 0.6) were added to each reaction tube and incubated for 1 h at 37°C.Samples were collected immediately after adding the inoculum to the reaction (t0), as well as after 30 min (t30), and 60 min (t60) of incubation.Each sample was then serially diluted and plated onto MacConkey agar plates for CFU/mL counting, following incubation for 18 h at 37°C.The assays were performed in triplicate, and the values obtained from CFU/mL counting were statistically analyzed using two-way ANOVA and Tukey's multiple comparison test, with a 95% confidence interval.

Pet cleaves key components of the complement system
Pet and Pet S260I production was previously detected by immunobloting showing a 104 kDa band corresponding to mature form of protein (Supplementary Figure S1).Specific degradation products of C3, C5 and C9 were only observed after the incubation with concentrated supernatant of HB101(pCEFN1) containing Pet. Pre-treatment with PMSF inhibited degradation of all complement components, confirming that complement cleavage by Pet relies on its serine protease activity.Furthermore, no degradation products were observed in the presence of Pet S260I (Figure 1).To further assess the minimum time required for degradation of C3, C5 and C9, shorter incubation periods were tested.For all three substrates, specific degradation products were only observed in the presence of the concentrated supernatant containing Pet after 1 h, but not after in 30 min of incubation (Figure 2).Therefore, cleavage of C3, C5 and C9 molecules was time-dependent, as more degradation products were observed in 24 h of incubation.Besides, Pet S260I was not able to cleave C3, C5 and C9 even when total protein amount was enhanced to 2 µg (data not shown).

Pet inhibits ZnCL 2 -induced C9 polymerization and cleaves C9 polymers previously formed
Pet was first incubated with C9 for 40 min and polymerization was subsequently induced by ZnCl 2 for 2 h.C9 monomers were dose-dependently degraded by Pet, preventing polymer formation, which are usually detected in the range of 100-250 kDa (Figure 3A).Almost complete degradation of C9 monomers and abrogation of C9 polymerization were observed after incubation with 5 µg of Pet and no degradation products were detected with Pet S260I (Figure 3A).Interestingly, Pet could also degrade C9 from preformed polymers induced by ZnCl 2 for 2 h.After 40 min of incubation with Pet, C9 was degraded, and lower amounts of polymers were detected (Figure 3B).

E. coli DH5a survives in human serum pre-treated with supernatant containing Pet
Figure 4 shows that E. coli DH5a survived in Pet-HS as well as in HI-HS but were lysed when incubated with HB101-HS or NHS.
Curiously, Pet S260I was also able to inhibit complement action, although less efficiently.Taken together, these data suggest that Pet plays a role in the inactivation of complement molecules, thus contributing to bacterial serum resistance.
Members of the SPATE family are commonly classified into class 1 (cytotoxic activities) and class 2 (immunomodulatory functions) (Ruiz-Perez and Nataro, 2014;Pokharel et al., 2019;Navarro-Garcia, 2023).However, studies published in recent years have shown that these proteases may have overlapping functions by acting depending on the bacterial environment.Class 1 SPATEs, like EspP, Sat and Pet, have immunomodulatory functions by targeting the complement system and by stimulating the inflammatory response (Orth et al., 2010;Rocha-Ramıŕez et al., 2016;Freire et al., 2022).Also, SepA, a class 2 SPATE, induces both cytopathic and pro-inflammatory effects in epithelial cell lines of human origin (Maldonado-Contreras et al., 2017;Meza-Segura et al., 2021).Therefore, it would be appropriate to revise the SPATEs classification, since these proteases can display both immunomodulatory and cytotoxic activities simultaneously.
In fact, we showed in our study that Pet, a class 1 cytotoxic SPATE, presented immunomodulatory activities as time-dependent cleavage of C3, C5 and C9 key components of the complement cascade.Cleavage products were detected after 1 h of incubation and no specific degradation products were observed either in the presence of Pet S260I or PMSF.Thus, the cleavage of these components can be attributed to the previously described serine protease activity (Ruiz-Perez and Nataro, 2014).
Proteins of the complement system are common targets of members of the SPATE family.(Table 1).EspP, Pic, Sat and Pet cleave C3, while C5 is degraded by EspP, Sat and Pet (Orth et al., 2010;Abreu et al., 2015;Freire et al., 2022).The amino acid sequence similarity observed between Pet and Sat (53%) can be an explanation why these two SPATEs share common complement substrates.In fact, Pet and Sat also degrade factor V and spectrin and display cytopathic effect on HEp-2 cells (Dutta et al., 2002).
The complement component C9 is essential for human serum bactericidal activity.C9 monomers are recruited by the C5b-8 complex, forming the MAC, a lytic pore, which promotes disarrangement of both bacterial outer and inner membranes, leading to cell lysis (Bjanes and Nizet, 2021;Doorduijn et al., 2021).Interestingly, Pet degrades C9 both in its monomeric and polymeric forms, and thus can potentially hamper pore formation and disarrange pre-formed pores.Differently from LcpA, a Leptospira spp.membrane protein (da Silva et al., 2015), and NS1, a membrane-associated glycoprotein of dengue virus (Conde et al., 2016), which bind to C9 to inhibit the formation of polymers, Pet interferes with MAC formation by degrading C9 molecules.Understanding how Pet prevents C9 polymerization is important in the context of an infection and may represent one of the strategies employed by E. coli to evade the complement system in the bloodstream.
In addition to MAC assembled, complement activation also leads to the formation of important by-products of the immune and inflammatory response, such as anaphylatoxins and opsonins.Anaphylatoxins C3a, C4a and C5a are important in processes such as chemotaxis, activation of immune system cells, antiinflammatory processes, chemokine synthesis and modulation of the adaptative immunity (Merle et al., 2015;Laumonnier et al., 2017).C3a may also have an important antimicrobial role, since high amounts are detected during bacterial infection and sepsis (Nordahl et al., 2004).C5a also plays an important role in the modulation of inflammation induced by bacteria (Czermak et al., 1999;Jain et al., 2015).Opsonins C3b and C4b, on the other hand, are important as they deposit on the surface of the target pathogen to facilitate recognition and destruction by cells of the immune system, such as neutrophils and macrophages (Merle et al., 2015).Since Pet cleaved C3 and C5, both alternative and terminal pathways of complement system activation as well as the biological processes involving C3a, C3b, C5a and C5b could be potentially inhibited by its proteolytic action.Previous studies have shown that proteases secreted by Staphylococcus aureus and Pseudomonas aeruginosa degrade C3 and C5, reducing the formation of anaphylatoxins C3a and C5a (Jusko et al., 2014;Mateu-Borraś et al., 2021) or generating active fragments like C5a from C5 degradation (Jusko et al., 2014).
E. coli DH5a, highly susceptible to complement-mediated killing, survived incubation with Pet pre-treated human serum, similarly to what has been shown for Pic and Sat (Henderson et al., 1999;Freire et al., 2022).Therefore, these SPATEs may collectively contribute to complement inactivation.It is important to emphasize that E. coli resistance to host defense mechanisms is multifactorial.Pathogenic E. coli, both intestinal and extraintestinal, have an extensive genetic framework of virulence factors that promotes evasion to the immune system and/or dissemination in the host (Santos et al., 2013;Freire et al., 2020;Santos et al., 2020;Santos et al., 2023).Besides, Pet S260I was also able to inhibit complement action, although less efficiently.Our hypothesis is that Pet S260I can partially inhibit complement action by a direct binding mechanism, but more experiments would be necessary to elucidate this mechanism, including mapping Pet S260I sites involved in C9 binding.Thus, Pet can be considered one of the several virulence factors harbored by pathogenic E. coli that may contribute to serum resistance and thereby to the host dissemination.
Due to the genetic plasticity of E. coli, the Pet-encoding gene can be found in extraintestinal isolates, as observed in cases of urinary infections and sepsis (Abe et al., 2008;Park et al., 2009;Nazemi et al., 2011;Herzog et al., 2014;Nunes et al., 2017).Moreover, a case of hemolytic uremic syndrome (HUS) resulting from a STEC infection that harbored EAEC virulence factors (Stx-EAEC O59: NM [H19]), among them the pet gene, was also reported (Carbonari et al., 2020).
Despite having been firstly identified and characterized in EAEC 042 (Eslava et al., 1998), the prevalence of the pet gene in DEC and E. coli isolated from bloodstream infections seems to be lower compared to other SPATE members (Boisen et al., 2009;Boisen et al., 2012;Abreu et al., 2013;Lima et al., 2013;Imuta et al., 2016;Andrade et al., 2017;Havt et al., 2017;Freire et al., 2020;Petro et al., 2020).In contrast, Mandomando and colleagues described that the pet gene is more prevalent in E. coli strains that cause bacteremia than in fecal EAEC strains (Mandomando et al., 2020).
Considering all the data presented in this study and in other previously published works (Navarro-Garcıá et al., 1999;Navarro-Garcıá et al., 2001;Abe et al., 2008;Park et al., 2009;Nazemi et al., 2011;Herzog et al., 2014;Nunes et al., 2017;Espinosa-Antuńez et al., 2019;Carbonari et al., 2020;Freire et al., 2020;Mandomando et al., 2020;Schüroff et al., 2021;Nascimento et al., 2022), we suggest that Pet producer-EAEC could cause damage to the intestinal epithelium, translocate through the intestinal barrier and reach the bloodstream, degrading complement components and promoting bacterial immune evasion.Likewise, Pet-producing ExPEC could evade the complement system by direct cleavage of its components and successfully spreading throughout the host.For this, our group intends to perform in vivo bacterial translocation assays to support this hypothesis, verifying whether the bacteria or the protease have the potential to translocate the intestinal barrier and whether the bloodstream would be a suitable environment for the secretion and activity of this serine protease.
Our results show that Pet is an important E. coli virulence factor that degrades components of the complement system in vitro, Pet and Pet S260I confer serum resistance to E. coli DH5a.E. coli DH5a (20 µL of inoculum -OD 600nm = 0.6) was incubated for 30 and 60 min with 50% NHS pre-treated with supernatant containing Pet (Pet-HS), Pet S260I (Pet S260I-HS) or containing the empty vector pSPORT (HB101-HS) (300 ug), with 50% NHS or HIS.Data are represented as log 10 CFU/mL.Results were analyzed using Two-way ANOVA, with Tukey's multiple comparison test and a 95% confidence interval (P < 0.05), using the GraphPad Prisma software (version 8.4.3).Even though the studies with EspP and Pic (Orth et al., 2010;Abreu et al., 2015) didn't evaluate all complement components, members of the SPATE family have some targets in common, such as C3 and C5.In this study, we used shorter incubation times and cleavage of C3, C5 and C9 by Pet were already observed in 1 h of incubation.ND, degradation not determined; +, positive molecule degradation.
mediates resistance to the bactericidal activity of human serum and, consequently, contributes the immune system evasion by E. coli.Therefore, the presence of pet in different ExPEC and other DEC pathotypes should be more investigated in order to elucidate the role of Pet in extraintestinal infections, mainly in E. coli collections that cause bacteremia and sepsis.

TABLE 1
Cleavage of complement components by EspP, Pic, Sat and Pet in in vitro assays.