Effect of hcp1 of the Type VI Secretion System on Biological Characteristics of F18 Shiga Toxin-Producing Escherichia coli

Yang Yang1,2, Hong Zhou3, Ji Shao1,2, Xinyi Zhang1,2, Pengpeng Xia1,2, Mingxu Zhou4,5, Qiangde Duan1,2 and Guoqiang Zhu1,2* 1College of Veterinary Medicine, Yangzhou University, Yangzhou 225009, China 2Jiangsu Co-Innovation Center for Important Animal Infectious Diseases and Zoonoses, Yangzhou 225009, China 3School of Basic Medical Science, Southwest Medical University, Luzhou 646000, China 4Institute of Veterinary Immunology and Engineering, National Research Center of Engineering and Technology for Veterinary Biologicals, Jiangsu Academy of Agricultural Sciences, Nanjing 210014, China 5Jiangsu Key Laboratory for Food Quality and Safety-State Key Laboratory Cultivation Base, Ministry of Science and Technology, Nanjing 210014, China


INTRODUCTION
S higa toxin-producing Escherichia coli (STEC) expressing F18 fimbriae is one of important pathogens causing diarrhea and edema in weaned piglets, which has brought huge losses to pig industry in China . F18 E. coli has many remarkable biological characteristics. It can express F18 fimbriae, type I fimbriae, AIDA and flagella. After adherence and invasion towards pig intestinal epithelial cells, F18 STEC secrete Shiga toxin (Stx2e) and damage pig intestines (Da et al., 2001). These virulence factors are closely related to its pathogenicity, and may be regulated by type VI secretory system (T6SS).
Recently, bacterial secretory systems have become one of the frontier research fields of pathogenic microorganisms (Yi et al., 2019). Many studies have found that T6SS plays an important role in bacterial pathogenicity, and mediates interaction between bacteria and environment (Zheng et al., 2011). It has also been found to influence bacterial interspecific competition. T6SS is encoded by a gene cluster containing 15-20 conservative open reading frames (ORFs) with an average length of more than 20 kb. About 25% Gram-negative bacteria contain T6SS gene cluster (Williams et al., 1996;Pukatzki et al., 2009). T6SS has been reported to be closely involved with bacterial pathogenicity (Peng et al., 2016). Shrivastava et al. (2008) found that the heme carrier protein (HCP), a heme-regulated protein that was encoded by hcp gene, can form a hexamer ring structure that can participate in the formation of transport channels connecting the inner and outer membrane of bacteria (Filloux et al., 2008). HCP is involved in the pathogenic process of a variety O n l i n e F i r s t A r t i c l e of bacteria. It have been reported to participate in the resistance of Vibrio cholerae and Burkholderia to the predation of Amoeba (Pukatzki et al., 2007;Aubert et al., 2008), in the inhibition of the phagocytosis of Aeromonas by macrophages (Suarez et al., 2008), in the increase of Salmonella survival ability in host cells (Parsons and Heffron, 2005), in the biofilm formation of Vibrio cholera (Yahr, 2006), and in promotion of acute and chronic infection of some pathogens (Yahr, 2006;Filloux et al., 2008;Hood et al., 2010), etc. E. coli contains multiple hcp clusters. Zhou et al. (2012) found that Hcp1 of E. coli could induce cytoskeleton rearrangement, apoptosis, IL-6 and IL-8 release of human brain microvascular endothelial cells. Hcp2 was related to the adherence and invasion of bacteria to host cells. Ding et al. (2018) proved that in avian pathogenic E. coli (APEC), Hcp1 and Hcp2 were jointly involved in biological characteristics such as adherence and viability. Hcp1 affected Quorum sensing, while as Hcp2 participated in biofilm formation. Hu et al. (2021) discovered three hcp gene clusters (hcp1, hcp2 and hcp3) in porcine extraintestinal pathogenic E. coli. All 3 gene clusters were involved into bacterial pathogenicity and participate in the colonization and persistence of bacteria in the host. Song et al. (2020) discovered that the deletion of hcp2a in APEC could improve biofilm formation, reduce resistance to chicken serum and suppress the expression of IL8 and IL1β in host cells. It was found that hcp1 and hcp2 gene clusters also exist in STEC, which may participate in the secretion and assembly of virulence factors and play an important role in bacterial pathogenesis. However, in F18 STEC, exact function of T6SS hcp1 and its relationship with bacterial virulence have not been explored clearly.
To further analyze the effect of T6SS hcp1 in bacterial virulence, F18ab STEC 107/86 reference strain was selected, and hcp1 deletion strain was constructed by λ Red-based recombination system. Function of hcp1 was further discussed by comparing the differences in growth, biofilm, motility, adherence, Stx2e production between wild type and hcp1 deletion strain.

Materials and the construction of F18ab Δhcp1 and F18ab
Δhcp1/phcp1 The strains and plasmids are listed in Table I. F18ab STEC 107/86 was stored in our laboratory . All strains were grown at 37°C in LB medium with 100μg/mL Ampicillin (Amp) or 34μg/mL chloramphenicol (Cm+) added if requested. The intestinal porcine enterocyte cell line (IPEC-J2) was grown in RPMI 1640-F12 medium supplemented with 10% neonatal bovine serum (37°C, 5% CO 2 ). Vero cells were cultured in DMEM medium supplemented with 10% fetal bovine serum (37°C, 5% CO 2 ). Plasmids pKD3, pKD46 and pCP20 used for λ Red-based recombination system were preserved in our laboratory. 4-week-old ICR mice were purchased from the Institute of Comparative Medicine of Yangzhou University.
The deletion strain F18abΔhcp1 was constructed by λ Red-based recombination system using △hcp1-1/△hcp1-2 primers (Table II) (Datsenko and Wanner, 2000). Construction of the hcp1 deletion strain was further verified by DNA sequencing. The ORF of hcp1 was amplified by PCR by using Hcp1-1/Hcp1-2 primers (Table  II). After insertion of hcp1 into pBR322, the complement plasmid pBR-hcp1 was constructed. The plasmid was transformed into the deletion strain F18abΔhcp1 for the construction of F18abΔhcp1/phcp1.

Growth curves, biofilm formation and mobility assays
The wild type (F18ab), deletion mutant (F18abΔhcp1) and complement strain (F18abΔhcp1/phcp1) were inoculated in LB medium and cultured overnight at 37 °C. OD 600 of each strain was measured every hour. The experiment was repeated 3 times to obtain the growth curves (Yang et al., 2018(Yang et al., , 2021. Amp r , λRed recombinase expression Kept and provided by our own lab pCP20 Amp r ,Cm r ; Flp recombinase expression Kept and provided by our own lab O n l i n e

F i r s t A r t i c l e
The wild type (F18ab), deletion mutant (F18abΔhcp1) and complement strain (F18abΔhcp1/phcp1) were inoculated in biofilm induction medium as previously described Zhou et al., 2014). The OD 600 values of each well were recorded to measure the amount of biofilm production using a crystal violet staining method. Each strain was tested using six replicates, and the experiment was conducted in triplicate.
The strains were inoculated in LB medium and cultured overnight at 37 °C. 1 μL of each strain was transferred to the middle of semi-solid medium (1% tryptone, 0.25% NaCl, 0.3% agar) respectively, and incubated at 37 °C for 16h. Motility halos were measured as previously described (Duan et al., 2013).

Adherence and invasion assay towards IPEC-J2
RPMI 1640 and Ham's F12 medium (1:1) containing 10% newborn bovine serum were used for IPEC-J2. Cells were cultured and transferred into 96 well plate. Briefly, 10 7 CFU of bacteria was added to each well of a 96-well tissue culture plate with a monolayer of approximately 10 5 . After incubation for 1h, cells were washed with PBS (pH 7.2) for 3 times, and then lysed with 0.5% Triton X-100 for 20 min. Solutions were diluted to 1:10 in PBS, and were spread on the LB plate for bacterial counting (Yang et al., 2021).
Invasion assays were performed as previously described . The monolayer cells were washed with PBS buffer (pH 7.2) for 3 times, and the strains were added with MOI at 1:100 respectively. After incubation for 2h, cells were washed with PBS (pH 7.2) for 3 times, and gentamicin (50 μg/mL) was added. After additional 2h incubation, 0.5% Triton X-100 was used for cell lysis. The number of bacteria that invaded IPEC-J2 cells was enumerated. DH5α was used as negative control.

Bacterial colonization assay in mice intestine
12 ICR mice (4-week-old) were divided into 3 groups: Wild type group, deletion strain group and complement strain group. 48 h before challenge, 5 g/L streptomycin was added to drinking water to eradicate the normal resident flora in the intestine. Fructose (6.7%) was added to increase the amount of drinking water. Cimetidine (50 mg/kg) was administered intraperitoneally 3 h before challenge to reduce gastric acid secretion (van der Velden et al., 1998). 2×10 8 CFU bacteria were injected intragastrically into mice. Feces pellets were collected every day after challenge. After dissolving with PBS, 10fold dilution series were plated on MacConkey agar plates to determine the number of CFU of E. coli. Experiment lasted for 5 days.

Vero cell cytotoxicity assay
Wild type (F18ab), deletion strain (F18abΔhcp1), complementation (F18abΔhcp1/phcp1) and DH5α were inoculated in LB medium at 37 °C. Strains were cultured to an OD 600 of 0.3, then mitomycin C (0.25 µg/mL) was added into each tube. After 12 h induction, culture supernatants were filtered through 0.22 µm filters. 100µL supernatants was added to the Vero cells and incubated for 20 h, with DMEM as negative control . According to the instructions of Cell Proliferation Reagent WST-1 Kit, 100 μL supernatants were added into Vero cells followed by 25 μL Cell Proliferation WST-1 Reagent for 4h incubation. OD 450 was measured.

RNA extraction and real-time fluorescence quantitative PCR
Each strain was growing to OD 600 at 2.0. Total RNA of each strain were extracted by TRIzol method . The primers used for fliC, fimH, fedf, AIDA and stx2e genes were listed in Table II. gapA gene was used as housekeeper gene. SYBR®Premix Ex Taq II (Takara, Shiga, Japan) and ABI 7500 Real Time System (Applied Biosystems) were employed for subsequent experiments.

O n l i n e F i r s t A r t i c l e
The obtained data were processed based on the 2 -△△ Ct method.

Statistical analyses
All statistical analyses were performed using SPSS 15.0 software (SPSS Inc., Chicago, IL, USA). Differences in data were analyzed with t-test.

F18abΔhcp1 and F18abΔhcp1/phcp1
The hcp1 deletion strain F18abΔhcp1 was constructed from its wild type F18ab 107/86 by using λ Red-based recombination system. The complete reading frame of hcp1 was amplified by PCR, inserted into pBR322 plasmid and transformed into F18abΔhcp1. The complement strain F18abΔhcp1/phcp1 was also successfully constructed. F18abΔhcp1 was verified verified by combined PCR and sequencing data. The results of growth curve and biofilm formation assay showed that there was no significant difference between F18ab, F18abΔhcp1 and F18abΔhcp1/ phcp1 (Fig. 1). Fig. 1. Results of growth curve (A) and motility assay (B). Notes: A, there was no significant difference in growth rate among the three strains. B, there was no significant difference in biofilm formation ability among the three strains.

Results of bacterial adhesion and colonization assays
The results showed that the diameter of the motion circle of F18abΔhcp1 decreased by about 18.8%, compared with the wild type. The motility of F18abΔhcp1/phcp1 was slightly improved (Fig. 2A), compared with F18abΔhcp1. In the adhesion assay, compared with wild type, the adherence ability of F18abΔhcp1 to IPEC-J2 decreased by about 40% (Fig. 2B).
The invasion ability of each strain was analyzed. Extracellular bacteria which failed to invade IPEC-J2 were killed by antibiotics, and cells were lysed by 0.5% Triton X-100. The intracellular bacteria were released for plate counting. Compared with wild type, the invasion ability of F18abΔhcp1 decreased by 28.3%, and invasion of F18abΔhcp1/phcp1 was recovered (Fig. 2C).
MacConkey plates counting assay (Fig. 2D) showed that, after the challenge, the bacterial shedding of wild type strain and F18abΔhcp1/phcp1 displayed a gradual upward trend. The excretion of F18abΔhcp1 gradually decreased. The knockout of hcp1 gene may block the secretion pathway of T6SS and hindered the secretion of protein molecules related to the process of adhesion and colonization.

Cytotoxicity assay
The results showed that, in hcp1 deletion strain group, the number of surviving cells is seven times that of wild type group (Fig. 3), which proved that the deletion of hcp1 inhibited the synthesis or secretion of Stx2e. The cytotoxicity of the deletion strain is similar to that of the negative control group. The cytotoxicity of F18abΔhcp1/ phcp1 was significantly higher than that of deletion strain.

O n l i n e F i r s t A r t i c l e
Effect of T6SS hcp1 on F18 STEC 5

The transcription levels of virulence genes
Total RNA of F18ab, F18abΔhcp1 and F18abΔhcp1/ phcp1 were extracted, respectively, and were reverse transcribed into cDNA. Housekeeping gene gapA was used in this experiment. The expression levels of five important virulence factors fimH, fedF, fliC, AIDA and stx2e were detected. As shown in Figure 4, the transcription level of fliC decreased by about 28%. The expression of fimH (type I fimbriae) decreased by 31%, and there was no significant difference in expression of fedF, AIDA and stx2e (Fig. 4). Previous studies have shown that 2-3 hcp gene clusters (hcp1, hcp2 and hcp3) have been found in meningitis-causing E. coli, avian pathogenic E. coli and porcine extraintestinal pathogenic E. coli (Ding et al., 2018;Song et al., 2020;Hu et al., 2021). Three hcp gene clusters exhibit different functions. However, the role of T6SS and HCP in the pathogenesis of F18ab STEC has not been well elucidated.
In this study, the growth trend of wild type, deletion mutant, and complement strains is the same. There is no significant difference in the lag phase, logarithmic growth phase, stable phase and decay phase, indicating that Hcp1 did not affect the growth of F18 E. coli (Fig. 1A).
STEC mediates the adherence of bacteria to host cells through F18 fimbriae, flagella and other adhesins (Da et al., 2001). After adherence, bacteria would invade into host cells or forms biofilm to colonize in host, and then releases Stx2e and other toxins, causing edema lesions. In this study, the related characteristics of bacterial adhesins influenced by hcp1 were investigated, and the transcription levels of major adhesins were detected by fluorescence quantitative PCR. F18ab E. coli mainly relies on its flagella to achieve its motor function. Most flagella subunits are assembled outside the plasma membrane, thus the protein molecules used for assembly need to be synthesized in the cytoplasm and then transported to the final assembly position through the inner membrane and outer membrane. Flagella not only mediate bacterial motility, but also acts as an important adhesin to mediate the adherence process between bacteria and host cells. The results showed that the motility of hcp1 deletion strain decreased by 18.8% ( Fig. 2A). T6SS may be involved in the transport process of flagella subunit protein, or the secretion of other cofactors involved in the assembly process of flagella subunit protein. The deletion of hcp1 leads to the decrease of assembly efficiency of flagellum, resulting in the weakening of bacterial motility. In this study, it was found that the adherence ability of Δhcp1 deletion strain was reduced by 40% (Fig. 2B), and its invasion ability decreased by 28.3% (Fig. 2C). The transcription levels of several main adhesion factor genes showed that the expression of flagella fliC, type I fimbriae fimH decreased by 28 and 31%, respectively (Fig. 4). The results indicated that hcp1 is related to the expression of type I fimbriae and flagella in F18 STEC strain, and deletion of hcp1 can inhibit the motility, adherence and invasion of F18 STEC.
Biofilm is a bacterial morphological change for adaptability, which can enhance the bacterial resistance to the external environment. Biofilm can greatly improve bacterial colonization in the host (Duan et al., 2012). After hcp1 deletion, there was no significant difference in biofilm formation ability between STEC deletion strain and wild strain (Fig. 1B), which proved that Hcp may not be involved in biofilm formation.
F18 STEC attached to porcine intestinal epithelial cells through adhesins, and then colonizes and secreted Stx2e. Through subunit B, Stx2e binds to Gb5 receptor of intestinal epithelium and can be ingested by the intestinal epithelial cells through endocytosis. Stx2e can induce apoptosis through subunit A, influence the vascular permeability, causes osmotic pressure disorder, and finally leads to edema (Niewerth et al., 2001). The expression level of Stx2e toxin gene in deletion strain F18abΔhcp1 did not change (Fig. 4), but the toxicity to Vero cells decreased significantly (Fig. 3). The numbers of surviving cells in wild and complement strain group were less than hcp1 deletion strain, indicating that Stx2e proteins in the supernatant of deletion strain was reduced. T6SS may be an important secretory pathway for Stx2e. The difference in cytotoxicity between the hcp1deletion strain and the O n l i n e

F i r s t A r t i c l e
Y. Yang et al. wild type did not come from the difference in the synthesis and expression of toxin, but from the release and secretion process.

CONCLUSION
In conclusion, T6SS hcp1 affects flagella and type I fimbriae in F18 STEC. Deletion of hcp1 suppress bacterial motility, adherence, invasion ability, and colonization in mice. Moreover, it is closely related to the release and secretion process of Stx2e, and is deeply involved in bacterial cytotoxicity. The exploration of the effect of hcp1 laid a foundation for further clarifying the complex function of T6SS system in F18 STEC.

O n l i n e F i r s t A r t i c l e
Effect of T6SS hcp1 on F18 STEC