Metabolic Reprogramming of Clostridioides difficile During the Stationary Phase With the Induction of Toxin Production

Abstract

The obligate anaerobe, spore forming bacterium Clostridioides difficile (formerly Clostridium difficile) causes nosocomial and community acquired diarrhea often associated with antibiotic therapy. Major virulence factors of the bacterium are the two large clostridial toxins TcdA and TcdB. The production of both toxins was found strongly connected to the metabolism and the nutritional status of the growth environment. Here, we systematically investigated the changes of the gene regulatory, proteomic and metabolic networks of C. difficile 630Δerm underlying the adaptation to the non-growing state in the stationary phase. Integrated data from time-resolved transcriptome, proteome and metabolome investigations performed under defined growth conditions uncovered multiple adaptation strategies. Overall changes in the cellular processes included the downregulation of ribosome production, lipid metabolism, cold shock proteins, spermine biosynthesis, and glycolysis and in the later stages of riboflavin and coenzyme A (CoA) biosynthesis. In contrast, different chaperones, several fermentation pathways, and cysteine, serine, and pantothenate biosynthesis were found upregulated. Focusing on the Stickland amino acid fermentation and the central carbon metabolism, we discovered the ability of C. difficile to replenish its favored amino acid cysteine by a pathway starting from the glycolytic 3-phosphoglycerate via L-serine as intermediate. Following the growth course, the reductive equivalent pathways used were sequentially shifted from proline via leucine/phenylalanine to the central carbon metabolism first to butanoate fermentation and then further to lactate fermentation. The toxin production was found correlated mainly to fluxes of the central carbon metabolism. Toxin formation in the supernatant was detected when the flux changed from butanoate to lactate synthesis in the late stationary phase. The holistic view derived from the combination of transcriptome, proteome and metabolome data allowed us to uncover the major metabolic strategies that are used by the clostridial cells to maintain its cellular homeostasis and ensure survival under starvation conditions.

Keywords: Clostridioides difficile; Clostridium difficile; Stickland reactions; metabolism; starvation; toxin formation.

FIGURE 1

Growth curve of C. difficile 630Δerm. Shown is the growth curve of C. difficile 630Δerm in casamino acids medium of four biological replicates, lag phase corrected, and the sampling time points (^∗) in the exponential phase (exp), the transient phase (trans), and three times in the stationary phase (stat1-3). The curve was fitted according to the biphasic Hill Equation using OriginPro, 2016 software (OriginLab Corporation, MA, USA); reduction of the optical density after time point stat1 was induced by aggregation of the cells in the culture medium.

FIGURE 2

Overview of changed pathways on transcriptomic (T), proteomic (P), and metabolic (M) level. Shown are the log2 fold changes of the transcriptomic and the proteomic data from the sampling time points trans, stat1, stat2, and stat3 with the samples of the exponential phase as reference. Yellow squares represent an increase of gene expression/protein production with log2 FC 0.5–1 for the light color and >1 for the dark color. Blue squares represent a decrease of gene expression/protein production with log2 FC –0.5 to –1 for the light color and <–1 for the dark color. Gray squares represent no changes with log2 FC between –0.5 and 0.5, white squares: not detected. Metabolite values represent a trend compared to the samples of the exponential phase. Toxin A and B in the proteomic data were detected with an ELISA (see section “Materials and Methods”), TCA, tricarboxylic acid; BCAA, branched chain amino acids; PPP, pentose phosphate pathway; ox, oxidative; red, reductive.

FIGURE 3

Time resolved systematic overview of reductive and oxidative Stickland reactions in C. difficile. On the left side the principal metabolic pathways are depicted. On the right, the individual metabolic pathways of the various amino acids of the reductive (Upper) and oxidative (Lower) Stickland reactions with the corresponding transcriptomic and proteomic data of representative genes/proteins are shown. The bars and the squares point out the sample time points exp, trans, stat1, stat2, and stat3 from left to right. The squares shows the log2 FC of each time point in comparison with the exponential phase of the cytosolic proteome (Upper row) and the transcriptome (Lower row). The bars represent the relative abundance of the metabolites based on the highest concentration; blue: intracellular compounds, orange: extracellular compounds, gray bar in exometabolome: initial concentration in the medium. Missing intermediates were not detected in the GC-MS or LC-MS analysis. ^∗ not detected in proteomic approach.

FIGURE 4

Proposed biosynthesis-pathway of C. difficile from the glycolysis substrate 3-phosphoglycerate to serine and cysteine. (A) Shown is the proposed pathway from 3-phosphoglycerate via serine to cysteine with the detected metabolite levels of 3-phosphoglycerate (3-P-glycerate) and the intracellular and extracellular L-serine values and the corresponding transcriptomic and proteomic data. The bars and the squares point out the sample time points exp, trans, stat1, stat2, and stat3 from left to right. The vertical axis shows the relative abundance based on the highest concentration; blue: intracellular compounds, orange: extracellular compounds, gray bar in exometabolome: initial concentration in the medium. Missing intermediate metabolites were not detected in the GC-MS or LC-MS analysis. The proteomic data of the cytosolic fraction (top squares) and the transcriptomic data (lower squares) show the log2 FC of each time point in comparison with the exponential phase. ^∗We proposed CDIF630erm_01132, a protein of unknown function located in the same operon with weak homologies to hydrolases, as a candidate for the generation of serine by the removal of the phosphate. (B) NADPH formation in the enzyme assay of crude extracts from exponential (dark gray) and stationary phase (light gray) performed at 30°C in a photometer. (C) Serine content in the enzyme assay detected in samples of crude extracts from exponential (dark gray) and stationary phase (light gray) analyzed by HPLC with (+) and without (–) the substrate 3-P-glycerate performed at 22°C.

FIGURE 5

Time resolved analysis of central carbon metabolism of glucose via fermentation by C. difficile. The different fermentation pathways from glucose via pyruvate with the detected intra- and extracellular metabolites and the corresponding transcriptomic and proteomic data of representative genes/proteins are shown. The bars and the squares point out the sample time points exp, trans, stat1, stat2, and stat3 from left to right. The squares shows the log2 FC of each time point in comparison with the exponential phase of the cytosolic proteome (Upper row) and the transcriptome (Lower row). The bars represent the relative abundance based of the metabolites based on the highest concentration; blue: intracellular compounds, orange: extracellular compounds, gray bar in exometabolome: initial concentration in the medium. Missing intermediate metabolites were not detected in the GC-MS or LC-MS analysis.

FIGURE 6

Toxin formation of C. difficile 630Δerm. Toxins were quantified at all sample time points discussed above and after 48 h of growth. Toxin A (yellow) and toxin B (orange) were quantified in the culture supernatant using an immunoassay and were calculated per mg of C. difficile dry weight; ^∗ no toxin detectable.