Prevalence and Characteristics of Plasmid-Encoded Serine Protease EspP in Clinical Shiga Toxin-Producing Escherichia coli Strains from Patients in Sweden

Shiga toxin-producing Escherichia coli (STEC) infection can cause a broad spectrum of symptoms spanning from asymptomatic shedding to mild and bloody diarrhea (BD) and even life-threatening hemolytic-uremic syndrome (HUS). As a member of the serine protease autotransporters of Enterobacteriaceae (SPATE) family, EspP has the ability to degrade human coagulation factor V, leading to mucosal bleeding, and also plays a role in bacteria adhesion to the surface of host cells. Here, we investigated the prevalence and genetic diversity of espP among clinical STEC isolates from patients with mild diarrhea, BD, and HUS, as well as from asymptomatic individuals, and assessed the presence of espP and its subtypes in correlation to disease severity. We found that 130 out of 239 (54.4%) clinical STEC strains were espP positive, and the presence of espP was significantly associated with BD, HUS, and O157:H7 serotype. Eighteen unique espP genotypes (GTs) were identified and categorized into four espP subtypes, i.e., espPα (119, 91.5%), espPγ (5, 3.8%), espPδ (4, 3.1%), and espPε (2, 1.5%). espPα was widely distributed, especially in strains from patients with BD and HUS, and correlated with serotype O157:H7. Serogroup O26, O145, O121, and O103 strains carried espPα only. Ten GTs were identified in espPα, and espPα/GT2 was significantly associated with severe disease, i.e., BD and HUS. Additionally, espP was strongly linked to the presence of eae gene, and the coexistence of espPα and stx2/stx2a + stx2c was closely related to HUS status. To sum up, our data demonstrated a high prevalence and genetic diversity of the espP gene in clinical STEC strains in Sweden and revealed an association between the presence of espP, espP subtypes, and disease severity. espP, particularly the espPα subtype, was prone to be present in more virulent STEC strains, e.g., “top-six” serotypes strains.


Introduction
Shiga toxin-producing Escherichia coli (STEC) is a foodborne, gram-negative bacterium belonging to the Enterobacteriaceae family and can cause a variety of human diseases ranging from asymptomatic shedding to mild/bloody diarrhea (BD) or even life-threatening diseases such as hemolytic uremic syndrome (HUS) [1].STEC infection is one of the leading causes of acute kidney injury in children, and STEC-infected individuals aged over 60 are more prone to mortality, irrespective of clinical conditions [2,3].Although O157:H7 has been considered the top causative serotype of STEC-linked disease and outbreaks, non-O157 strains with various genetic backgrounds are increasingly recognized by their association with HUS and linkage to large outbreaks, particularly strains of the "top-six" serogroups (i.e., O26, O45, O103, O111, O121 and O145) [4][5][6][7].Shiga toxin (Stx) is the most important virulence factor in STEC.It contains two main types, assigned Stx1 and Stx2, with four Stx1 subtypes (a, c, d, and e) and twelve Stx2 subtypes (a-l) [8].Stx2 is more critical than Stx1 in the development of HUS [9,10], and strains carrying stx2a with/without stx2c genes are significantly associated with severe clinical diseases [11].Intimin, encoded by eae gene located within the locus of enterocyte effacement (LEE) pathogenicity island, is an important aggravating factor involved in gut colonization of STEC.Intimin can induce attaching and effacing (A/E) lesions on intestinal epithelial cells and contribute to human diseases, including the development of hemorrhagic colitis (HC) and HUS [12,13].
STEC induces intestinal impairment through the release of virulence factors without invading tissues [14,15].The release of secreted proteins, such as proteases, is crucial for the generation of A/E lesions and is involved in a variety of processes associated with infection [16].Extracellular serine protease P (EspP) is one of the most abundant proteins in culture supernatants of STEC strains and has been described as a member of the serine protease autotransporter of Enterobacteriaceae (SPATEs) protein family encoded on large virulence plasmids, such as pO157, pO113, and pO26-Vir in STEC strains [17][18][19][20].The espP gene comprises a 3900 bp open reading frame encoding the 1300 amino acid (aa) EspP protein with a molecular weight of 142 kDa, and the mature secreted passenger domain with a molecular weight of 104 kDa is generated through cleavage of the N-terminal signal peptide and the C-terminal β-domain and secreted into the extracellular milieu, showing serine protease activity [21].By cleaving coagulation factor V and complement C3, C3b, and C5, EspP could impact host proteins, which are important for coagulation and complement activation, thus enhancing the severity of infections [22,23].EspP might also be involved in the regulation of virulence, as shown by the cleavage of hemolysin [21].A recent study reported that pooled immunoglobulins (IgG) on the course of disease in a mouse model could bind to EspP, block its enzymatic activity, and protect the host from O157:H7 STEC infection [24].Additionally, EspP could stimulate electrogenic ion transport in human colonic monolayers, leading to watery diarrhea that is often followed by HC and extra-intestinal complications, including HUS, while neither Stx nor numerous components of the type-III secretion system have been found to independently elicit fluid secretion [25].Collectively, EspP could promote colonic cell injury, bacterial adherence to intestinal cells, and the uptake of Stx by intestinal cells [25,26], and its role in blood coagulation, pathophysiology, and immune-modulation can contribute to STEC pathogenesis [27,28].Five EspP subtypes have currently been identified (EspPα-EspPε) [29].EspPα participates in biofilm formation and also plays a role in adhesive and cytopathic effects [23,30,31].EspPγ is able to cleave pepsin and human coagulation factor V, while EspPβ and EspPδ either remained un-secreted or exhibited proteolytic activity [22].Intriguingly, EspPα has been shown to be more prevalent in human isolates, while other espP subtypes are more prevalent in reservoir animals and the environment [22,28].
Although EspP is frequently found in STEC strains [27], the role of EspP in STEC pathogenesis is not well-studied, and the molecular characteristics of espP-positive STEC strains, especially clinical strains, have rarely been described.Therefore, in this study, we investigate the prevalence of espP and its subtypes and polymorphisms among clinical STEC strains isolated from patients with varying disease outcomes in Sweden.Furthermore, Microorganisms 2024, 12, 589 3 of 10 we assess its correlations with serotypes, other virulence factors such as eae and stx, and clinical outcomes.

Ethic Statement
The study was approved by both the regional ethics committees in Gothenburg (2015/335-15) and Stockholm (2020-02338), Sweden.Patient consent was waived due to a retrospective review of the patients' medical records.Patient data were anonymous, and no consent was required to work with the bacterial strains.

Bacterial Strains
A total of 239 STEC strains were included in this study.These strains were isolated from STEC-infected individuals in Sweden in the period of 1994-2018.The isolation of STEC strains was performed as described previously [32].Clinical data of STEC-infected patients, such as age, sex, and clinical symptoms, were collected by reviewing medical records and utilizing the standard practices employed for STEC surveillance in Sweden, with clinical symptoms categorized into non-bloody stool (NBS), bloody diarrhea (BD) and HUS.The duration of bacterial shedding was defined as described previously [33].
Bacterial DNA of all STEC strains were extracted and then subjected to whole-genome sequencing using Illumina HiSeq X platform at SciLifeLab (Stockholm, Sweden) as described elsewhere [34], and Ion Torrent S5 XL platform (Thermo Fisher Scientific, Waltham, MA, USA) at The Public Health Agency of Sweden as described elsewhere [35].The Illumina sequencing reads underwent de novo assembly using SKESA (version 2.3.0),where the reads were assembled into longer contiguous sequences to rebuild an approximate sequence of the original genome [33].The Ion Torrent sequencing reads were de novo assembled utilizing SPAdes (version 3.12.0) in its "careful mode", a specialized setting designed to enhance the accuracy, comprehensive coverage, and fidelity of the assembly process, resulting in a more reliable reconstruction of the genomic sequence, and then the sequences were annotated with Prokka (version 1.14.6)[33].The genomic assemblies in this study were deposited in GenBank with accession numbers, as shown in Table S1.
Serotype determination was achieved by comparing assemblies to the SerotypeFinder database (DTU, Denmark) (http://www.genomicepidemiology.org/ (accessed on 6 August 2020)) with the use of BLAST+ (version 2.2.30) [33].An in-house stx subtyping database was constructed with ABRicae (version 0.8.10) (https://github.com/tseemann/abricate(accessed on 6 August 2020)), incorporating representative nucleotide sequences of all identified stx1 and stx2 subtypes, and then stx subtypes were identified using the assemblies to search against this stx subtyping database.The presence of intimin-encoding gene eae was determined according to the genome annotation as previously described [33].Multilocus sequence typing (MLST) analysis was performed by comparing sequences of seven housekeeping genes (adk, fumC, gyrB, icdF, mdh, purA, and recA) against the E. coli MLST database with the use of an online tool provided by the Warwick E. coli MLST scheme website (https://enterobase.warwick.ac.uk/species/ecoli/allele_st_search (accessed on 8 August 2020)) as mentioned before [33].The allelic profile of these seven housekeeping genes was used to generate a specific sequence type (ST) for each STEC strain.The metadata of all isolates is shown in Table S1.

espP Subtyping
The sequences of the espP gene were retrieved from the genomic assemblies in accordance with the genome annotation.The unique espP sequences in this study were then aligned with reference nucleotide sequences of different espP subtypes that have been previously reported and downloaded from GenBank [22,36,37].After alignment using MEGA 11 software (version 11.0.13)(Center for Evolutionary Medicine and Informatics, Tempe, AZ, USA), the genetic distances of the espP sequences were calculated with the maximum composite likelihood method, and a neighbor-joining phylogenetic tree was constructed using 1000 bootstrap replicates with maximum composite likelihood model.The espP subtypes were determined by the phylogenetic structure and genetic distance.Based on espP sequence polymorphism, espP genotypes (GTs) were used to determine the diversity within each espP subtype as described previously [38].

Data Analysis
Statistical correlations between the presence of espP/espP subtypes and characteristics of the strain (serogroups, stx subtypes, the presence of eae) or clinical outcomes (HUS, BD, and NBS) were examined using Fisher's exact test in R software (version 4.3.1)(https://www.r-project.org)(accessed on 20 November 2023).A p-value less than 0.05 was considered statistically significant.
The presence of espP was significantly associated with BD, HUS, O157:H7, and the presence of eae (Table 1).However, no association was found between espP and age groups, nor was the duration of bacterial shedding (Table S2).HUS-hemolytic uremic syndrome; BD-bloody diarrhea; NBS-non-bloody stool.# The association was analyzed between espP gene and clinical symptoms (non-HUS and HUS; NBS and BD; NBS and BD + HUS), bacterial features (serotype O157 and non-O157; eae-positive or eae-negative), age groups (child: <10 years; adult: ≥10 years) or duration of bacterial shedding (long: >24 days; short: ≤24 days); only differences with statistical significance were shown.The figures represent the number of espP-positive or -negative STEC strains, and the percentage is shown in the following brackets.* Statistically significant difference.
Among 18 espP GTs, 2 major GTs (espPα/GT2 and espPα/GT1) contained 52 and 49 strains, respectively, and 9 GTs contained only 1 strain, while the rest contained 2 to 5 strains (Figure 1).espPα/GT2 was more common in strains from patients with BD, HUS, and BD + HUS, whereas espPα/GT1 was more prevalent in strains from individuals with NBS and non-HUS, and espPα/GT6 was more prevalent in strains with NBS (Table 2).No association was found between other espP subtypes/GTs and clinical symptoms (Table S4).NBS-non-bloody stool; BD-bloody diarrhea; HUS-hemolytic uremic syndrome.# The association was analyzed between espPα/GTs and clinical symptoms (HUS and non-HUS; BD and NBS; HUS + BD; and NBS).
The number represents the number of strains, and the percentage is shown in the following brackets.* Statistically significant difference.
Table 3. Association between stx subtypes + espP subtypes and clinical symptoms.NBS-non-bloody stool; BD-bloody diarrhea; HUS-hemolytic uremic syndrome.Combinations of stx subtypes and espP subtypes showed associations with clinical symptoms.The presence of stx2a + stx2c + espPα was significantly higher in strains from patients with HUS and BD + HUS (p < 0.0001), while strains with stx1a + espPα were more prevalent in patients without HUS or BD + HUS (p = 0.0003 and 0.0036).Additionally, stx1a + stx2c + espPα was more commonly found in strains associated with non-HUS (p = 0.0500), and stx2c + espPα showed a higher prevalence in strains with NBS (p = 0.0383 and 0.0172) (Table 3).

Discussion
This study reported a high prevalence of espP (54.4%) in clinical STEC strains from patients with various disease outcomes, especially in strains of O157 (98.5%),O26 (81.6%),O121 (69.2%) and O145 (66.7%) serogroups, and espP was detected in 75.0% of strains from patients with HUS, 74.5% of strains with BD and 74.8% of strains with BD + HUS.The prevalence and distribution of the espP gene in human-derived STEC strains have also been investigated in previous studies.For instance, espP was detected in 55.0% of STEC strains implicated in human disease in Africa [39], whereas 65.0% of clinical STEC strains harbored espP in Austria, with the majority being serogroup O157, O26, and O145 strains espP was observed in the majority of O145 STEC strains (88.0%) from patients with watery diarrhea, BD HUS, and from Germany [40].In a report from Canada, espP was present in 86 (76.8%) out of 112 STEC strains of highly pathogenic serogroups O157, O26, O103, O111, and O145 from humans, including 42 (77.8%)strains from patients with severe diseases (BD + HUS) [41].Meanwhile, espP was not detected in stx1cand stx2eharbouring eae-negative STEC isolates from patients in this study, the same as previously described [42,43].These findings showed that espP tended to be prevalent in clinical STEC strains, especially in highly pathogenic serogroups, suggesting its role in the pathogenic process and clinical outcomes.
There are very limited data on the association of espP and disease severity, although an antibody response against EspP was discovered during the development of STEC infection [18].We found that espP was strongly associated with severe outcomes, e.g., BD and/or HUS, in contrast with a previous study in Canada reporting no significant association between EspP protease and disease in humans [41].We were interested to understand if different espP subtypes contribute to the disease severity.A previous study revealed significant functional differences among various EspP subtypes, where subtype α and γ isolates showed proteolytic activity, whereas subtype β and δ either lacked proteolytically activity or were not secreted, and these differences correlated with point mutations around the active serine protease site [22].Subtype ε was first found in O91:H14 strains, with no functional study till now [37].Four espP subtypes, i.e., espPα, espPγ, espPδ, and espPε were identified in our strain collection, in which espPα was the most predominant subtype, accounting for 97.5% of strains belonging to serotypes O157:H7, O26:H11, O121:H19, O103:H2, and O145:H28, while other serotypes harbored espPγ, espPδ or espPε.espPα was the predominant subtype, especially in strains from patients with BD and HUS, and was statistically associated with O157:H7.Although no association was found between four espP subtypes and clinical outcomes, within ten espPα genotypes, espPα/GT2 was significantly associated with BD, HUS, and BD + HUS, as compared to other espPα genotypes, indicating that certain espPα genotypes could be considered as a predictor for severe disease outcome.Further studies are necessary to understand the functional differences and mechanisms of different espPα genotypes underlying STEC pathogenesis.
The coexistence of stx and other virulence genes, i.e., eae, is more prone to enhance the virulence of STEC and exacerbate the STEC-associated disease severity [44].However, there is limited literature describing the relationship between the coexistence of stx subtypes and espP in relation to disease severity.Our study showed that stx2a + stx2c was the most prevalent stx subtype among espP-positive clinical strains.Interestingly, the presence of stx2a + stx2c + espPα (mostly espPα/GT2) was strongly associated with BD and HUS, indicating that espPα might play a more important role in the pathogenesis of STEC strains with stx2a + stx2c.In accordance with a previous study showing that 97 out of 106 espP-positive strains (91.5%) from humans were positive for eae [28], 124 out of 130 espP-positive strains (95.4%) contained eae in this study, and the presence of espP was significantly associated with eae-positive strains.Intimin encoded by eae is an outer membrane protein and responsible for intimate adherence to target eukaryotic cells as an important virulence factor, whereas EspP is an autotransporter that can translocate through the periplasm and the outer membrane of bacteria.The role of EspP in bacterial adhesion was supported by a transposon mutagenesis investigation performed in the O157:H7 STEC strain EDL933, in which EspP was identified as one of the virulence factors directly involved in biofilm formation and adherence to T84 intestinal epithelial cells, probably through the polymerization of EspP and generation of "rope-like structures" [17,30].It has also been demonstrated that human STEC isolates that carry eae along with espP adhere more strongly to HEp-2 cell cultures [28,41].Combined with our findings, there might be some functional associations between the two proteins, which need further verification.
In conclusion, espP was highly prevalent in clinical STEC strains in Sweden, which was also strongly linked to the presence of the eae gene and significantly associated with severe disease outcomes, i.e., BD and HUS.Four espP subtypes were identified, among which espPα was the most predominant, carried by strains of virulent serogroups O157, O26, O145, O121, and O103, and correlated with serotype O157:H7.espPα, along with stx2a + stx2c, was closely related to HUS, while genotype espPα/GT2 was distinctively correlated with BD and HUS, compared to other espPα genotypes.Our results revealed that espPα, particularly espPα/GT2, is prone to be present in highly virulent STEC clinical strains, highlighting its significant clinical relevance.The pathogenicity of espP-positive strains associated with human diseases requires further exploration.

Figure 1 .
Figure 1.Phylogenetic relationships of 18 different espP sequences identified in this study and 15 espP subtypes reference sequences based on the Neighbor-Joining method.At each node, the black circles represent values of bootstraps that were more than 60.The corresponding espP subtype (number of strains), strain name, serotype (number of strains), stx subtype (number of strains), and ST types (number of strains) are shown.The espP subtypes/GTs in this study are indicated in bold and different colors.Scale bar indicates genetic distance.

Figure 1 .
Figure 1.Phylogenetic relationships of 18 different espP sequences identified in this study and 15 espP subtypes reference sequences based on the Neighbor-Joining method.At each node, the black circles represent values of bootstraps that were more than 60.The corresponding espP subtype (number of strains), strain name, serotype (number of strains), stx subtype (number of strains), and ST types (number of strains) are shown.The espP subtypes/GTs in this study are indicated in bold and different colors.Scale bar indicates genetic distance.

Table 1 .
Prevalence of espP gene in 239 STEC clinical strains in correlation to clinical symptoms and bacterial features # .