(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Glucosyltransferase-dependent and independent effects of Clostridioides difficile toxins during infection

['F. Christopher Peritore-Galve', 'Department Of Pathology', 'Microbiology', 'Immunology', 'Vanderbilt University Medical Center', 'Nashville', 'Tennessee', 'United States Of America', 'Vanderbilt Institute For Infection', 'Inflammation']

Date: 2022-04

Clostridioides difficile infection (CDI) is the leading cause of nosocomial diarrhea and pseudomembranous colitis in the USA. In addition to these symptoms, patients with CDI can develop severe inflammation and tissue damage, resulting in life-threatening toxic megacolon. CDI is mediated by two large homologous protein toxins, TcdA and TcdB, that bind and hijack receptors to enter host cells where they use glucosyltransferase (GT) enzymes to inactivate Rho family GTPases. GT-dependent intoxication elicits cytopathic changes, cytokine production, and apoptosis. At higher concentrations TcdB induces GT-independent necrosis in cells and tissue by stimulating production of reactive oxygen species via recruitment of the NADPH oxidase complex. Although GT-independent necrosis has been observed in vitro, the relevance of this mechanism during CDI has remained an outstanding question in the field. In this study we generated novel C. difficile toxin mutants in the hypervirulent BI/NAP1/PCR-ribotype 027 R20291 strain to test the hypothesis that GT-independent epithelial damage occurs during CDI. Using the mouse model of CDI, we observed that epithelial damage occurs through a GT-independent process that does not involve immune cell influx. The GT-activity of either toxin was sufficient to cause severe edema and inflammation, yet GT activity of both toxins was necessary to produce severe watery diarrhea. These results demonstrate that both TcdA and TcdB contribute to disease pathogenesis when present. Further, while inactivating GT activity of C. difficile toxins may suppress diarrhea and deleterious GT-dependent immune responses, the potential of severe GT-independent epithelial damage merits consideration when developing toxin-based therapeutics against CDI.

Clostridioides difficile is an anaerobic spore-forming bacterium that is the leading cause of antibiotic-associated diarrhea and pseudomembranous colitis in the USA. This pathogen produces two protein toxins, TcdA and TcdB, that enter host colon cells to cause inflammation, fluid secretion, and cell death. The enzymatic activity of TcdB is a target for novel C. difficile infection (CDI) therapeutics since it is considered the major factor in causing severe CDI. However, necrotic cell death due to non-enzymatic TcdB-host interactions has been reported in cell culture and colonic explant experiments. Here, we generated C. difficile mutant strains with enzyme-inactive toxins to evaluate each toxin’s role in an animal model of CDI. We observe an additive role for TcdA and TcdB in disease, and both glucosyltransferase-dependent and independent phenotypes. These findings are expected to inform the development of toxin-based CDI therapeutics.

Funding: This project was supported by the funding from the National Institutes of Health (NIH) and the United States Department of Veterans Affairs (VA). FCPG was supported by NIH grant T32 DK007673, and DBL received funding through NIH grant AI957555 and VA grant VA BX002943. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

CDI remains a pervasive disease across the world and poses a significant burden on patients and the healthcare system. The main goal of this study was to define GT-dependent and independent effects of each toxin during CDI using the mouse model of infection and a BI/NAP1/PCR-ribotype 027 epidemic strain. The results from this study provide a framework to further understand the complex molecular interplay between C. difficile toxins and the host colon during infection.

Historically, TcdA was considered the toxin responsible for CDI pathogenesis [ 20 ]. The identification of patient-derived clinical isolates that are TcdA-negative and TcdB-positive, along with functional analyses of toxin mutants in animal models indicated however that TcdB is necessary and sufficient to cause severe CDI [ 21 , 22 ]. Other data demonstrate that TcdA alone can cause symptoms in the hamster model of infection, suggesting a role for both toxins during disease [ 23 ]. Overall, the respective roles of TcdA and TcdB during infection are not fully understood.

In addition to GT-dependent effects of C. difficile toxins, TcdB can induce GT-independent necrotic cell death at high concentrations in vitro (> 0.1 nM) [ 12 , 13 ]. This effect occurs through the stimulation of reactive oxygen species via the NADPH oxidase complex [ 14 ]. In contrast, TcdA induces GT-dependent cell death through the apoptosis pathway at both high and low concentrations [ 15 ]. TcdA and TcdB can also stimulate cytokine release in vitro through both GT-dependent and independent pathways [ 9 ]. The relevance of GT-independent cell death during infection has been unclear since it is unknown if cells encounter the concentration of TcdB required to stimulate necrotic cell death. A recent study concluded that GT activity was necessary to cause CDI in mouse and hamster models of infection, but this does not preclude an additional role for GT-independent events as the infection progresses [ 16 ]. The study also involved GT-defective C. difficile mutants in the 630 strain, which is genetically tractable, but causes only mild disease symptoms in animal models [ 17 ]. In contrast, the BI/NAP1/PCR-ribotype 027 epidemic strains have been associated with multiple human epidemics and cause significant pathology in animal models of infection [ 16 – 19 ]. Among several differences, the 630 strain has been characterized as one with low expression levels of TcdA and TcdB [ 17 , 19 ].

CDI is mediated by two large homologous protein toxins, TcdA (308 kDa) and TcdB (270 kDa) that are secreted during infection to cause disease symptoms. The toxins bind host cell receptors and become internalized into vesicles via endocytosis [ 5 ]. Endosome acidification elicits structural changes in the toxins, stimulating pore formation and translocation of the N-terminal glucosyltransferase (GT) and autoprocessing (AP) domains into the cytosol [ 5 ]. Inositol-6-phosphate-induced autoprocessing releases the GT domain in the host cell and permits access to Rho family GTPases. GT activity transfers glucose from UDP-glucose onto Rho GTPases, irreversibly inactivating these regulatory proteins [ 5 – 8 ]. This inactivation causes cytoskeletal rearrangement, leading to the disruption of cell adhesion junctions, cytopathic changes, and apoptosis [ 5 ]. These effects stimulate proinflammatory cytokines and neutrophil chemoattractants that generate acute inflammatory responses. Prolonged host inflammation during CDI increases the severity of tissue damage and the probability of lethal disease outcomes [ 9 – 11 ].

Clostridioides difficile infection (CDI; formerly Clostridium difficile) is the leading cause of hospital-acquired diarrhea and pseudomembranous colitis in the USA [ 1 , 2 ]. C. difficile is a Gram-positive, spore-forming anaerobe that infects the colon, causing mild to severe symptoms including diarrhea, pseudomembranous colitis, toxic megacolon, and in severe cases, death [ 3 ]. CDI is prevalent among elderly and immunocompromised individuals in healthcare settings, typically following treatment with broad spectrum antibiotics. However, the rate of community-acquired infections among healthy individuals has increased over the past two decades due to the emergence of novel epidemic C. difficile strains [ 3 , 4 ]. Despite the clinical importance of CDI, we do not have a complete understanding of molecular host-microbe interactions during infection, which hinders our progress towards developing effective prevention and treatment strategies.

(A) Number of MPO + cells per single mucosal layer per 20x field of view (FOV). Each point is the average of three 20x FOVs from one animal. Bars depict the group mean and error bars are the standard error of the mean. Statistical differences between R20291 and A GTX B+ and the other treatments as determined by Tukey’s HSD are shown in a bracket (*** p < 0.001). (B) Representative MPO immunohistochemistry images showing high amounts of immune cell influx into submucosal and mucosal layers in R20291 and A GTX B+, and lower amounts in A+ B GTX and A GTX B GTX . Scale bar, 80 μm.

Epithelial damage can be induced or exacerbated by immune cell infiltrates, and pseudomembranes are often formed by sloughed epithelial cells and intraluminal neutrophils bound in a fibrinous matrix [ 18 , 26 ]. Therefore, immunohistochemical analyses of infected cecal tissues was employed to test GT-dependent effects on MPO + immune cell recruitment, and to determine if epithelial damage observed in A GTX B GTX -infected mice occurred through toxin-host interactions or aggressive immune cell infiltration. The number of MPO + cells per 20x field of view were quantified in single mucosal layers (n = 3 animals per treatment). This approach revealed that R20291 and A GTX B+ caused the most significant immune cell influx during disease (~230 cells per FOV; p < 0.001; Fig 5 ). Although A+ B GTX elicited mild MPO + cell recruitment, it was not significantly higher than that observed in mock-inoculated mice (p = 0.2587; Fig 5 ). Finally, A GTX B GTX caused considerable epithelial damage and pseudomembrane formation, but there were few MPO + immune cells recruited to sites of epithelial damage ( Fig 5 ).

Phenotypes of cecal inflammation were more nuanced than edema. Once again, R20291, A GTX B+, and A+ B GTX , caused the highest average inflammation scores (2–3), hallmarked by moderate to severe neutrophilic inflammation and submucosal to mural involvement (p < 0.001; Fig 3C and 3D ) [ 18 ]. However, A GTX B GTX caused minimal to moderate inflammation and did not significantly differ from inflammation induced by A+ B GTX (p = 0.1736; Fig 3C and 3D ).

To examine histopathological phenotypes in infected tissues, excised ceca were fixed and frozen in OCT media, then cryosectioned and H&E stained. Edema, inflammation, and epithelial damage were scored on a 1–4 scale by a board-certified gastrointestinal pathologist based on previously published criteria [ 18 ]. Cecal edema was observed in mice inoculated with R20291, A GTX B+, and A+ B GTX , and was scored on average between 2 and 3 based on the presence of moderate to severe edema with widespread multifocal submucosal expansion (p < 0.001; Fig 3C and 3D ) [ 18 ].

Whole organ inflammation was determined by measuring cecum area and colon length ( Fig 3B ). The area of ceca from mice inoculated with R20291, A GTX B+, and A+ B GTX were half the size of those from mice inoculated with A GTX B GTX , ΔtcdA B GTX , ΔtcdA ΔtcdB, and the mock control (p < 0.001; Fig 3B ). Colon length was altered in response to weight loss-inducing strains, albeit less drastically than differences observed in ceca ( Fig 3B ). Colons from mice inoculated with R20291, A GTX B+, and A+ B GTX were significantly shorter than mock-inoculated mice (p < 0.05; Fig 3B ). However, none were significantly shorter than mice inoculated with ΔtcdA ΔtcdB (p > 0.05; Fig 3B ).

(A) Representative images of ceca and colons from mice inoculated with each treatment at 2 dpi. White bars denote 1 cm. (B) Cecum area and colon length as metrics of organ inflammation. Bars are the mean of each group and points are individuals within the group. Error bars represent standard error of the mean. Differences in cecum area depict that R20291, A GTX B+, and A+ B GTX are significantly smaller than the four other groups. Differences in colon length are comparing R20291, A GTX B+, and A+ B GTX to mock-inoculated colons. (C) Epithelial damage, inflammation, and edema mean scores as determined by a gastrointestinal pathologist. Points represent each individual animal and error bars are the standard error of the mean (n = 6–9 per treatment). Statistical differences as determined by Tukey’s HSD are shown in brackets (* p < 0.05; *** p < 0.001; **** p < 0.0001) or by letters (p < 0.05). (D) Representative H&E images of mouse ceca inoculated with each strain at 2 dpi. Scale bar, 80 μm.

To visualize diarrhea and inflammation on a macroscale, mice were euthanized at 2 dpi, then ceca and colons were excised and imaged (n = 6 per treatment). The ceca and colons of mice inoculated with the wildtype R20291 strain were severely inflamed and contained very little wet stool ( Fig 3A ). Soft, discolored stool was typically observed in ceca and colons from mice inoculated with A GTX B+ and A+ B GTX ( Fig 3A ). Stool discoloration from A GTX B GTX , ΔtcdA B GTX , and ΔtcdA ΔtcdB-inoculated mice was noted in ceca and colons, relative to well-formed, normal colored stool from mock-inoculated mice ( Fig 3A ).

To quantify C. difficile burden during infection, daily fecal samples were collected, weighed, homogenized, then dilution plated onto semi-selective media. At 1 dpi, strains colonized to a high density (~10 7 CFU g -1 stool), however, R20291 was significantly more abundant in stool compared to ΔtcdA B GTX (p = 0.031; Fig 2D ). There were no significant differences in C. difficile stool burden between strains at 2 dpi, when symptoms were most severe. Colonization density of ΔtcdA ΔtcdB began to significantly reduce at 3 and 4 dpi, while titers of other strains remained more consistent. As expected, no mock-inoculated mice contained C. difficile ( Fig 2D ). Since there were no notable differences in C. difficile burden in stool, it is expected that symptom severity phenotypes caused by each strain are due to molecular toxin-host interactions and not colonization burden.

Weight loss is typically caused by increased fluid loss during CDI-associated diarrhea. To visually test the effects of GT activity of each toxin on diarrhea severity, mouse stool was collected daily for visual assessment of moisture, color, and consistency. Stool scores from 1-to-4 were assigned based on pre-defined criteria. All mice inoculated with wildtype R20291 had a stool score of 4 at 2 dpi based on the presence of watery diarrhea and wet tail ( Fig 2C ). Although some mice inoculated with A GTX B+ developed severe watery diarrhea, most had a stool score of 3, characterized by soft and discolored stool ( Fig 2C ). No mice inoculated with A+ B GTX developed watery diarrhea, and typically displayed soft, discolored stool ( Fig 2C ). Stool from mutant strains that induced no weight loss (A GTX B GTX , ΔtcdA B GTX , and ΔtcdA ΔtcdB) was well formed, yet discolored ( Fig 2C ).

(A) Visual abstract of the cefoperazone mouse model of CDI used for this study. The figure was created with BioRender (B) Percent weight loss from day 0 for R20291 (n = 22), A GTX B+ (n = 15), A+ B GTX (n = 15), A GTX B GTX (n = 15), ΔtcdA B GTX (n = 9), ΔtcdA ΔtcdB (n = 22), and mock (n = 13). Points represent group averages at each day and error bars denote standard error of the mean. (C) Stool scores at 2 dpi. (D) Daily C. difficile burden in stool. Each point is an individual mouse, and the crossbars represent group daily means. Significantly different groups as determined by Dunn’s test are shown in brackets (* p < 0.05; *** p < 0.001).

The non-lethal mouse model of infection was used to define GT-dependent and independent effects of each toxin during CDI. Mice were pre-treated with cefoperazone antibiotics in drinking water for five days, then were returned to regular drinking water for two days prior to inoculation with C. difficile ( Fig 2A ). C. difficile spores of each strain (10 5 CFU/mL) were administered via oral gavage, and metrics of infection and symptom severity were assessed daily until 2- or 4-days post-inoculation (dpi). The most severe decline in animal weight occurred at 2 dpi, then mice began to recover by 4 dpi ( Fig 2B ). Thus, the starkest differences between treatments were observed at 2 dpi ( S2 Fig ). At 2 dpi, the wildtype R20291 strain induced 12% weight loss on average, which was significantly more than all other groups, even when compared to A GTX B+ (p = 0.0075) and A+ B GTX (p < 0.0001) (Figs 2B and S2 ). A GTX B+ caused the second highest average weight loss at 2 dpi (8.9%), but was not significantly different from A+ B GTX , which caused 6.4% average weight loss (p = 0.2273; Figs 2B and S2 ). Mice inoculated with A GTX B GTX , ΔtcdA B GTX , ΔtcdA ΔtcdB, or mock (sterile PBS) lost little to no weight (Figs 2B and S2 ). Similar trends between treatments were observed at 3 and 4 dpi (Figs 2B and S2 ).

Finally, we confirmed mutant and wildtype toxin function in vitro using cell rounding assays. Strain supernatants were obtained from bacteria cultured in toxin-inducing TY medium, followed by centrifugation and filter sterilization to remove cell contaminants. Dilutions of strain supernatants, TY, and purified recombinant TcdA (100 pM) and TcdB (1 pM) were placed into individual wells each containing 20,000 Vero cells and cytopathic rounding was imaged every hour. This experiment confirmed that cell rounding required functional TcdA, TcdB or both toxins ( S1 Fig ). Consistent with previously published data, TcdB demonstrated faster kinetics of rounding compared to TcdA ( S1 Fig ). Validation of mutant genotypes, toxin secretion, bacterial growth, and toxin function allowed us to confidently move into animal infection experiments.

The novel strains were tested for toxin secretion and bacterial growth in vitro to ensure that the introduced mutations had no unintended effects on these processes. Toxin secretion was determined in mutant and wildtype strains cultured for 24 hours in toxin-inducing TY medium. After 24 hours of growth, cell pellets and supernatants were separated by centrifugation, then supernatants were filter-sterilized to remove any cell contaminants. Immunoblot analyses of TcdA and TcdB determined that there were no GT-dependent effects on toxin secretion and confirmed the knockout phenotypes of ΔtcdA B GTX and ΔtcdA ΔtcdB ( Fig 1B ). In vitro growth of each strain was assessed in BHIS medium for 24 hours under anaerobic conditions. Bacterial growth was measured every hour, revealing that there were no significant differences in growth between mutant and wildtype strains ( Fig 1C ).

(A) tcdA and tcdB gene alignments of mutant and wildtype strains to the reference genome of R20291. Red boxes highlight SNMs that deactivate GT catalytic activity. (B) Representative Western blot images of mutant and wildtype strains to confirm secretion (or lack thereof) of TcdA and TcdB. Images directly below each Western blot are stain-free gel images shown as loading controls. The graphs show densitometry analyses from Western blot experiments (n = 3). Dots signify each replicate, bars denote the mean, and error bars represent the standard error of the mean. Significant differences were determined by Tukey’s HSD (* p < 0.05). (C) In vitro growth curves of each strain and mock-inoculated controls (n = 5 per treatment) cultured in BHIS for 24 hours. Dots are the average at each given timepoint, and the error bars depict the 95% confidence interval.

To define GT-dependent and independent effects of TcdA and TcdB during infection, GT-deficient mutants were generated in the epidemic C. difficile BI/NAP1/PCR-ribotype 027 R20291 background (hereafter referred to as R20291) using homologous allelic exchange [ 24 ]. This approach allowed for the introduction of single nucleotide mutations (SNMs) in the GTD of either or both toxins, causing single point mutations at positions TcdA::D285N/D287N and/or TcdB::D286N/D288N ( Fig 1A ). Four novel mutant strains were created, including GT-deficient mutants for each toxin (A GTX B+ and A+ B GTX ), a GT-deficient mutant of both toxins (A GTX B GTX ), and a total tcdA knockout with GT-deficient TcdB (ΔtcdA B GTX ) ( Table 1 ). Whole genome sequencing was performed on all strains used in this study to ensure the proper mutations were present and that there were no off-target mutations that might affect their phenotypes ( Table 1 and Fig 1A ) [ 25 ]. There were no unexpected variations in nucleotides for any of the mutant strains compared to the wildtype background and the reference R20291 genome (NC_013316). All sequencing reads are available in the NCBI Sequencing Read Archive under accession number PRJNA762329.

Discussion

CDI continues to be a significant nosocomial disease in the USA and worldwide, and a more thorough understanding of the function of TcdA and TcdB during infection may open new avenues for prevention and treatment therapies. This includes the development of toxin-inhibiting drugs such as small molecule inhibitors or neutralizing antibodies and nanobodies [19]. These compounds are designed to inhibit specific toxin domains such as the GTD, APD, or the combined oligopeptide repeat (CROPS) domain [5,19]. In recent years, efforts have focused on inhibiting the GT activity of TcdB since TcdB is considered the major driver of severe CDI symptoms. While conceptually an ideal target, multiple investigators have noted that glucosyltransferase-deficient toxin mutants are still cytotoxic or capable of eliciting cytokine responses [9,12–14,27,28].

The role of glucosyltransferase-independent activities during infection are not fully understood. Previous attempts to recapitulate GT-independent effects in murine models have included instillation of recombinant GT-defective TcdB by intrarectal instillation or cecal injection [29,30]. These studies found no GT-independent effects. However, these systems may have lacked the sensitivity of the C. difficile infection model, in which toxins are constantly produced and interact with the host epithelium over a course of several days. Recently, GT activity of TcdB was shown to be necessary to cause CDI, and there were no observed GT-independent effects in the mouse model of infection using defined GTX mutants in the C. difficile 630 strain [16]. This study had two limitations: the first being that the C. difficile 630 strain produces few CDI symptoms in the mouse model of infection when compared to epidemic C. difficile strains, and it also does not cause epithelial damage [16–18]. The second limitation was that histopathology was assessed at 5 dpi, a timepoint in which mice begin to recover from CDI [16,18]. The aim of our study was to define the GT-dependent and independent effects of C. difficile toxins during CDI using the epidemic BI/NAP1/PCR-ribotype 027 R20291 strain in the mouse model of infection and to look for evidence of epithelial damage at the time when weight loss is most severe.

In recent years, TcdB has been accepted as the main virulence factor necessary for fulminant CDI [21,22]. However, TcdA alone can cause mild to moderate CDI symptoms in animal models of infection [21–23]. In this study, mice were inoculated with novel isogenic mutants with deactivated GTDs of either or both toxins, as well as wildtype, double toxin KO, and mock-inoculated controls. The result of these experiments demonstrated that GT activity of TcdA or TcdB alone can cause moderate weight loss during infection, and that GT activity is required for CDI-induced weight loss. Mice inoculated with the wildtype R20291 strain lost the most weight, indicating that the toxins had an additive effect in causing the most severe weight loss outcomes. Additionally, the size of colons and ceca were measured, revealing that GT activity of either toxin was sufficient to cause significant inflammation and subsequent size reduction of both organs. Unexpectedly, we observed that R20291 was the only strain capable of causing severe watery diarrhea, whereas A+ B GTX and A GTX B+ only caused stool to become softer and discolored. Collectively, these results highlight the nuanced and additive effects of each toxin to CDI symptoms including weight loss, diarrhea, and organ inflammation.

Histological damage was assessed in mouse ceca, which is the site of the most severe histopathology in the mouse model of CDI [18]. Strains that had functional GT activity of one or both toxins caused severe multifocal edema, whereas GT-inactive or toxin knockout strains caused little to no edema. Inflammation was the most severe in mice inoculated with R20291, A GTX B+, and A+ B GTX . However, the severity of inflammation between A+ B GTX and A GTX B GTX was not significantly different, suggesting a mild GT-independent effect on inflammation during infection. Wildtype R20291, A GTX B+, A+ B GTX , and A GTX B GTX all caused severe epithelial damage during infection. This phenotype was observed as dead and dying cells sloughing off into the lumen, and by the presence of pseudomembranes. Unexpectedly, ΔtcdA B GTX did not elicit epithelial damage different from the double KO and mock controls, which may indicate that GT-independent effects of TcdA are necessary to potentiate epithelial damage during CDI. Indeed, both TcdA and TcdB cause GT-independent effects on immune responses during intoxication in vitro and in vivo which may synergize to increase the severity of epithelial damage [9,27,31]. Together, these results demonstrate that edema and inflammation are GT-dependent and are elicited by TcdA or TcdB during infection, although mild inflammation can be caused in a GT-independent manner. They also indicate that epithelial injury occurs through GT-independent mechanisms of TcdA and TcdB during infection.

Since inflammation can contribute to epithelial damage in some models [26,32–35], immunohistochemical analyses were used to quantify GT-dependent and independent effects on MPO+ immune cell influx. Strikingly, high amounts of MPO+ cell infiltration were detected in mice inoculated with only R20291 or A GTX B+. In contrast, relatively low levels of MPO+ cells were observed in ceca from mice inoculated with A+ B GTX and A GTX B GTX despite their induction of epithelial injury and inflammation. These results indicate that MPO+ immune cell infiltration is dependent on GT activity of TcdB, and that A GTX B GTX causes severe epithelial damage and pseudomembrane formation in the absence of MPO+ immune cell recruitment.

In conclusion, our data demonstrate that GT activity of either TcdA and TcdB is required for weight loss and organ inflammation, but GT activity of both toxins together causes the most severe weight loss and diarrhea phenotypes. Histological assessment of infected tissues revealed that GT activity of TcdA or TcdB can cause significant edema and inflammation, however GT-activity was not necessary to cause epithelial damage and pseudomembrane formation. Finally, analysis of MPO+ immune cell recruitment demonstrated that GT activity of TcdB was required to elicit MPO+ immune cells to sites of infection and that GT-independent epithelial damage did not require or elicit MPO+ immune cell influx.

While we have observed GT-independent effects during infection, it is possible that therapeutic approaches targeting the GTD will be successful to halt downstream GT-dependent diarrhea and acute inflammatory responses that exacerbate disease severity. However, we must consider that there may be consequences from GT-independent epithelial damage and pseudomembrane formation that might occur when GT activity is neutralized, such as effects on CDI recurrence and/or colon-related diseases. Future studies will aim to elucidate these effects.

[END]

[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1010323

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.