Background

Clostridium innocuum, identified as an Erysipelotrichia species, is a Gram-positive, anaerobic, spore-forming bacterium [1]. C. innocuum was avirulent in intramuscular and intraperitoneal infection models in guinea pigs and did not have known toxin expression; therefore, this anaerobe was defined as the intestinal commensal bacterium [2, 3]. Nonetheless, studies indicate that it is the second most common species that causes extra-intestinal clostridial infections and a potential contributor to antibiotic-associated diarrhea (AAD) [4, 5]. A retrospective case-control study in Taiwan revealed that patients infected with C. innocuum experienced more complications related to the gastrointestinal tract than those infected with Clostridioides difficile [6]. Furthermore, Ha et al. [7] showed that C. innocuum is related to the extra-intestinal manifestation of Crohn’s disease, a form of inflammatory bowel disease. These studies suggest that C. innocuum may have undefined pathogenic roles within the human intestinal tract. More importantly, David et al. [8] suggested that C. innocuum could synthesize the peptidoglycan precursor that terminating in serine to result in low affinity for vancomycin; as a result, C. innocuum is intrinsically resistant to vancomycin, with a minimal inhibitory concentration in 8 µg/mL [2].

A phylogenetic analysis using multilocus sequencing indicated that isolates of C. innocuum from blood, ascites, and feces could not be differentiated based on their phylogenetic position [5]. Moreover, a cross-sectional study with comparative genomic analysis found no phylogenetic associations between C. innocuum isolates from AAD patients and asymptomatic controls [9]. Nonetheless, a recent genomic analysis identified four genetically distinguishable clades among C. innocuum isolates based on single nucleotide polymorphisms of the core genome and average nucleotide identity [10]. These results suggest that, although C. innocuum is a genetically diverse species, elucidating the genetic variance of C. innocuum could help clarify its role in human disease.

The natural habitat of C. innocuum is in the human intestine [2, 3]. It has been suggested that the dominant survival niche of colonized bacteria in the gut is in the biofilm [11]. Biofilm can retain water, shield bacteria from antibiotics and immune clearance, and facilitate quorum sensing detection and horizontal gene transfer [12], offering advantages for survival in the intestinal niche. As the intestinal bacterium, the biofilm production activity of C. innocuum may contribute to its colonization, persistence, and potential pathogenesis in the intestine. Nonetheless, it remains unclear whether C. innocuum can produce biofilms and whether different isolates exhibit varying levels of biofilm formation activities.

Given that C. innocuum is a genetically diverse species [10], additional genetic and phenotypical analyses are necessary to establish the molecular epidemiological information for C. innocuum. To achieve this objective, there is an urgent need for rapid and cost-effective molecular typing methods for this anaerobic bacterium. This study employed amplicon sequencing, also known as Illumina targeted sequencing, to establish a sequence type for C. innocuum isolates. Furthermore, the biofilm production activity of C. innocuum isolates was analyzed, and the correlation between the biofilm production activity and the phylogenetic position of these isolates was established.

Methods

Bacterial isolates and culture conditions

Non-repetitive isolates of Clostridium innocuum from stool were provided by the Bacteria Bank, Chang Gung Memorial Hospital at Linkou in Taiwan (Supplementary Table S1). All C. innocuum isolates included in this study were verified by polymerase chain reaction (PCR) with species-specific primers for C. innocuum [13]. It should be noted that clinical history of the patients is currently not available. The C. innocuum isolates were cultured on anaerobic blood agar or grew in brain heart infusion (BHI) broth (Becton Dickinson and Company; Sparks, MD, USA) supplemented with 0.1% L-cysteine at 37ºC under anaerobic conditions.

DNA manipulations

The total DNA of C. innocuum was extracted using the phenol/chloroform method [14]. Briefly, C. innocuum pellets (after 12–16 h incubation) were collected (2850 ×g at 4ºC) and washed with ddH2O twice. The bacterial pellets were treated with lysozyme (0.8 mg at 37ºC for 30 min), protease K (0.1 mg at 56ºC for 30 min), and RNase A (0.1 mg at 37ºC for 30 min). Bacterial DNAs were extracted by phenol/chloroform (1:1 mix ratio by volume) and precipitated with isopropanol (1:2 mix ratio by volume) at room temperature. The extracted DNAs were air dried, dissolved in ddH2O and adjusted to a concentration of 100 ng/µL for PCR amplification. Primers were designed using primer3 (https://primer3.ut.ee) according to the C. innocuum genome deposited in NCBI [15] and are described in Table 1. The target genes were amplified by PCR in conditions 95ºC 30”, 57ºC/60ºC 30”, and 72ºC 30” for 35 cycles (Table 1). The PCR products used for amplicon sequencing were generated using primers with Nextra transposase adaptors (Illumina, Inc.; San Diego, CA, USA).

Table 1 Primers used for multilocus sequence typing analysis
Full size table

Amplicon sequencing

PCR products were checked by 2% agarose gel electrophoresis, mixed, and purified using the Agencourt AMPure XP PCR Purification Kit (Fisher Scientific International Inc.; Pittsburgh, PA, USA). Quantification was performed using the Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific) on a Qubit 3.0 Fluorometer (Thermo Fisher Scientific) according to the manufacturer’s instructions. The final purified libraries were applied for cluster generation and sequencing on MiSeq using the V3 600 cycles kit.

Alignment and identify variations of amplicons

The FASTX-Toolkit (http://hannonlab.cshl.edu/fastx_toolkit) was utilized to process the raw read data files. BWA alignment tools were utilized to align the filtered reads with the reference sequences. The amplicon included 10 targets, atpA, ddl, dxr, groL, gyrA, gyrB, mdh, pgk, recA and rpoB. The nucleotide sequence of these 10 genes in CIN141 was amplified by polymerase chain reaction and determined by Sanger sequencing as the reference sequence. The alignment results of paired-end reads were categorized under two conditions: first, read 1 and read 2 were aligned on the same amplicon target, and second, the reads were exclusively aligned to a single amplicon target. Variant calling was done by Clair3 with default options. The qualified variants were defined by fitting with a sequencing depth of over 100× and an allele fraction of over 80%. For the genome sequences fetched from NCBI, the sequences of the selected housekeeping genes were obtained by aligning the fetched genome to the reference sequence by Blast (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Phylogenetic analysis

The concentrated sequence of 7 selected targets from each isolate was aligned by Clustal Omega [16], and the phylogenetic distance of the input sequences was calculated by maximum likelihood estimation with the Generalized Time Reversible (GTR) model and bootstrap 1000 (Unipro UGENE v45.0) [17, 18]. The resulting phylogenetic tree was visualized using Figtree (v1.4.4) distributed with Bioconda [19].

Crystal violet biofilm assay

The overnight bacterial cultures in BHI broth supplemented with 0.1% L-cysteine were diluted 100-fold with fresh BHI broth (supplemented with 0.1% L-cysteine and 0.1 M glucose) in a 24-well polystyrene microplate (Costar 3513, Corning Life Sciences) and the biofilm matrix was determined by crystal violet staining as previously described with modifications [20]. Following a 24 h incubation at 37ºC, the wells were washed with sterile water to remove non-adherent bacterial cells. The washed wells were air dried, and the biofilm matrix on the well surface was fixed with 95% ethanol at room temperature for 30 min. The biofilm matrix was stained with 0.5% crystal violet at room temperature for 30 min. After three washes with tap water, crystal violet was dissolved with 30% acetic acid and quantified on an ELISA reader at a wavelength of 562 nm (OD562). The well without C. innocuum inoculation served as the negative control. Two biological replicates were performed for biofilm production activity.

Biofilm inhibitory concentration assay

The biofilm susceptibility assay was performed as described in previously established procedures with modifications [21]. C. innocuum isolates were grown in BHI broth supplemented with 0.1% L-cysteine for 24 h at 37ºC. After diluting these cultures to OD600 of 0.1 with BHI broth (supplemented with 0.1 M glucose), 150 µL of bacterial suspension was transferred to a 96-well flat-bottom plate (Nunc™ 269787 MicroWell, Thermo Fisher Scientific Inc.; Waltham, MA, USA) and covered with a modified polystyrene microtiter lid (Nunc™ 445497 Immuno TSP Lids). After incubation, the peg lids were rinsed in sterile water three times and placed on a flat bottom 96-well plate containing 4–256 µg/mL vancomycin and incubated at 37ºC for an additional 24 h. After vancomycin treatments, the peg lids were rinsed with sterile water three times and transferred to the BHI broth in the flat bottom 96-well plate. The biofilm on the peg was detached by sonication at room temperature for 5 min (DC150H, Delta Ultrasonic Co., LTD.; New Taipei City, Taiwan), and the peg lids were discarded and replaced by a standard lid. The OD600 was measured before and after incubation at 37ºC for 24 h. Bacterial growth was defined by the difference in.

OD600 (OD600 at 24 h – OD600 at 0 h) ≥ 0.05. The biofilm minimal inhibitory concentration was defined as the lowest concentration of drugs that inhibited bacterial growth. Three biological replicates were performed to determin the vancomycin BIC.

Statistical analysis

Statistical analyses were performed using Prism software, version 6 (GraphPad Software, Inc.; San Diego, CA, USA). Significant differences between multiple groups were determined using ANOVA and Tukey’s multiple comparisons test. Statistical significance was set at P < 0.05.

Results

Multilocus sequence typing of Clostridium innocuum isolates

We selected housekeeping genes used for typing Clostridium and Clostridioides species from the PubMLST database (https://pubmlst.org/organisms). The homologous of ten house-keeping genes (atpA, ddl, dxr, groL, gyrA, gyrB, mdh, pgk, recA, and rpoB) were identified in C. innocuum, amplified by polymerase chain reaction, and subjected to amplicon sequencing. Simpson’s index of diversity [22, 23], used to evaluate the discriminatory power of the 10 selected genes, was calculated for 52 isolates of C. innocuum from Chang Gung Memorial Hospital at Linkou (CGMH, Taiwan) and 60 sequence-available C. innocuum isolates from the National Center for Biotechnolgy Information (NCBI) Genome Database (https://www.ncbi.nlm.nih.gov/datasets/taxonomy/1522/), as shown in Table 2. Multilocus sequence typing (MLST) analyses typically use seven loci to type bacterial isolates (https://pubmlst.org). Simpson’s index of diversity of groL, pgk, and recA from 112 isolates was less than 0.7 (CGMH + NCBI, Table 2); therefore, these genes were excluded from MLST analysis. Finally, seven genes, including atpA, ddl, dxr, gyrA, gyrB, mdh, and rpoB, were utilized to type C. innocuum isolates.

Table 2 Simpson’s index of the selected genes
Full size table

Among 52 isolates from CGMH (Taiwan) and 60 isolates from the NCBI Genome Database, 36 and 25 different sequence types (STs) were identified, respectively (Fig. 1A). Among 58 STs, ST1 (5.4%, 6/112), ST6 (11.6%, 13/112) and ST42 (5.4%, 6/112) exhibited a prevalence greater than 5%. Fifty-three STs were exclusive to isolates from specific geographic regions (Fig. 1B). ST1, ST6, and ST51 were found in C. innocuum isolates from both CGMH and the NCBI Genome Database, representing isolates from Taiwan and the USA (Fig. 1A and B). Furthermore, ST37 was identified in C. innocuum isolates from Australia and China, and ST42 was identified in isolates from the USA and Canada (Fig. 1B). Noticeable, there is only one isolate from Australia, Switzerland and the UK; therefore, these results cannot represent the prevalence of STs in these geographic regions.

Fig. 1
figure 1

Multilocus sequence typing for 52 C. innocuum isolates in Taiwan (CGMH) and 60 sequence-available C. innocuum isolates from the NCBI Genome Database. The housekeeping genes atpA, ddl, dxr, gyrA, gyrB, mdh, and rpoB were utilized for typing C. innocuum isolates. (A) The sequence types (STs) of C. innocuum from CGMH and NCBI Genome Database. (B) The STs of C. innocuum with corresponding geographic distribution information

Full size image

Phylogenetic analysis and biofilm production activity of C. innocuum isolates

The intestinal microbiota can form a biofilm adhering to the intestinal mucus surface under healthy conditions [11]. To evaluate the biofilm production activity of C. innocuum, we cultured 52 isolates from CGMH for 24 h and analyzed their biofilm production using the crystal violet staining assay. All analyzed C. innocuum isolates were found to be biofilm producers (OD562 > background; acetic acid washes from empty wells). The OD562 values ranged from 0.069 to 3.746, with a median value of 0.184 (Fig. 2A). We categorized C. innocuum isolates according to biofilm formation levels: isolates with OD562 < 0.184 (median) exhibited low-level biofilm formation (25/52, 48%), isolates with 0.184 < OD562 < 0.276 (1.5× median) exhibited medium-level biofilm formation (9/52, 17%), isolates with 0.276 < OD562 < 0.552 (3× median) exhibited high-level biofilm formation (7/52, 14%), and isolates with OD562 > 0.552 exhibited very high-level biofilm formation (11/52, 21%) (Fig. 2B). 52% of C. innocuum isolates (27/52) showed medium to very high levels of biofilm production activity. C. innocuum isolates with ST2, 4, 15, 18, 21, 26, 28, and 58 exhibited a very high-level biofilm mass (Fig. 2A).

Fig. 2
figure 2

The phylogenetic distance and biofilm production activity of C. innocuum isolates. (A) The sequence type (ST) and biofilm production activity of 52 C. innocuum isolates from the Chang Gung Memorial Hospital at Linkou (CGMH, Taiwan). (B) The percentage distribution of C. innocuum isolates exhibited low (OD562 < 0.184), medium (0.184 < OD562 < 0.276), high (0.276 < OD562 < 0.552), and very high (OD562 > 0.552) levels of biofilm. 0.184 is the median value of the biofilm mass of 52 analyzed C. innocuum isolates. The maximum likelihood phylogenetic tree constructed from the sequence of concentrated housekeeping genes of C. innocuum isolates (C) from the CGMH (n = 51; excluding the outgroup isolate CIN117) or (D) in both CGMH (n = 52) and NCBI Genome Database (n = 60). The scale bar represents nucleotide substitutions per site. The phylogenetical clade and biofilm production activity of C. innocuum isolates from CGMH are labeled in (C). Two biological replicates were performed for biofilm production activity

Full size image

To assess the relationship between the biofilm production phenotype and the genetic phylogenetics of C. innocuum, the concentrated sequences of the 7 selected loci were used for the phylogenetic analysis based on maximum likelihood estimation. The results showed that the 52 C. innocuum isolates from CGMH could be separated into phylogenetical clade I (13/52, 25%) and clade II (39/52, 75%) (Fig. 2C), with CIN117 as an outgroup isolate (Supplementary Fig. S1). Noticeably, among 11 isolates exhibiting very high-level biofilm formation, 8 isolates were classified as clade I (Fig. 2C), suggesting that the strong biofilm production phenotype was associated with a specific clade of C. innocuum. The phylogenetic distance of 52 isolates from CGMH and 60 isolates from the NCBI database was shown in Fig. 2D. Most analyzed isolates showed close phylogenetic distance except 4 outgroup isolates (CIN117, OF1-2LB, DFI.7.33, and DFI.1.206, Fig. 2D).

Vancomycin treatments did not activate the production of biofilm in C. innocuum

C. innocuum is intrinsically resistant to vancomycin [4, 5]. As vancomycin is one of the primary drugs used to treat Clostridioides difficile infection, it may serve as an external signal or environmental stress for C. innocuum. To evaluate whether vancomycin influences the biofilm production activity of C. innocuum, we selected isolates from clade I and clade IIa (Fig. 2C) with varying levels of biofilm production activities (Fig. 3A) for analysis. The growth of the selected isolates was inhibited by 8 µg/mL vancomycin (except for CIN152) while remaining unaffected by 2 µg/mL and 4 µg/mL vancomycin treatments (Fig. 3B). The crystal violet staining assay showed that the biofilm production of the selected isolates remained similar under 0 µg/mL and 4 µg/mL vancomycin culture conditions (Fig. 3C). This suggests that vancomycin does not appear to be a factor that affect the biofilm production activity of C. innocuum isolates.

Fig. 3
figure 3

The sublethal concentration of vancomycin treatments did not induce biofilm production in C. innocuum isolates. (A) Biofilm production activity of the selected clade I and IIa C. innocuum isolates. (B) The growth activity of the selected C. innocuum isolates under vancomycin concentrations ranged from 0 to 8 µg/mL. (C) Biofilm production activity of the selected C. innocuum isolates under 0 and 4 µg/mL vancomycin treatments. Two biological replicates were performed for growth activity and biofilm production activity

Full size image

Biofilm-embedded C. innocuum survived the high concentration of vancomycin treatments

The biofilm provides protection for bacteria against antibiotic elimination. Therefore, this study evaluated whether vancomycin treatments could effectively eliminate C. innocuum inside the biofilm. The selected isolates of clade I and IIa were cultured in the 6-well polypropylene plate to form biofilms under vancomycin-free conditions. After 24 h of incubation, planktonic cells were removed and the biofilm mass in the plate was exposed to vancomycin at concentrations ranging from 4 to 16 µg/mL for an additional 24 h. Crystal violet staining showed that only CIN152 and CIN164 exhibited a statistically significant decrease (ANOVA, p < 0.01) in biofilm mass under 16 µg/mL vancomycin treatments (Fig. 4A).

Fig. 4
figure 4

Biofilm protection by C. innocuum isolates against high concentrations of vancomycin treatments. (A) Biofilm mass of the C. innocuum isolates after exposure to vancomycin treatments at 0, 4, 8, and 16 µg/mL. (B) The vancomycin biofilm inhibitory concentration (BIC) of C. innocuum isolates, categorized based on their levels of biofilm formation, ranging from low to very high. Three biological replicates were performed for biofilm production and vancomycin BIC. *, P < 0.05

Full size image

To further verify whether the biofilm could protect C. innocuum isolates from high concentrations of vancomycin treatments, and whether strong biofilm producers would be more resistant to vancomycin than weak biofilm producers, the vancomycin biofilm inhibitory concentration (BIC) [21] was compared for the selected C. innocuum isolates with different levels of biofilm production. Among the 9 isolates that exhibited low to medium levels of biofilm formation, 8 isolates had vancomycin BIC 8 µg/mL, and one isolate (CIN146) showed vancomycin BIC > 256 µg/mL (Fig. 4B). Among 8 isolates exhibiting high to very-high levels of biofilm formation, 7 isolates had vancomycin BIC > 256 µg/mL (Fig. 4B). These results suggest that strong biofilm producers of C. innocuum can survive very high concentrations of vancomycin treatments.

Discussion

MLST and sequence-based phylogenetic analyses revealed significant genetic diversity among C. innocuum isolates [10]. The extensive genetic variability within the C. innocuum isolates suggests a possible natural existence as a commensal organism in the human intestinal tract. The capacity to form spores provides C. innocuum a competitive advantage in colonizing the intestine [1]. Furthermore, this study showed that the majority of analyzed C. innocuum isolates exhibit a remarkable ability to produce substantial amounts of biofilm. This attribute would enhance the ability of C. innocuum to interact with or compete with other commensal bacteria within the intestinal environment and would be related to its pathogenicity.

Biofilm is the structure produced by bacteria that plays a pivotal role in the survival of bacteria under stressful conditions and their ability to establish prolonged colonization [24]. This study showed that all analyzed C. innocuum isolates exhibited biofilm-producing capabilities, with 52% (27/52) of the isolates producing medium to very high-levels of biofilm. These results suggest that the biofilm production capacity is a significant and conserved trait within the evolutionary process of C. innocuum. In addition to serving as a physical barrier that protects bacteria from the effects of antibiotics, biofilm mass plays crucial roles in interactions among multiple bacterial species [25, 26]. Slater et al. [27] showed that, in mixed-species biofilms, Bacteroides fragilis could inhibit C. difficile growth; therefore, biofilm and bacteria inside the biofilm may have intricate functions within the intestinal niche. With its intrinsic vancomycin-resistant activity, C. innocuum could still produce substantial biofilm masses under 4 µg/mL vancomycin culture conditions. Additionally, the strong biofilm production activity was related to the high vancomycin biofilm inhibitory concentration in C. innocuum isolates. Consequently, the biofilm produced by C. innocuum may serve as a protective barrier, not only for C. innocuum itself but also for other intestinal commensals and pathogens, such as C. difficile, shielding them from vancomycin-induced elimination. Although the potential for C. innocuum to form multi-species biofilm communities within the intestine requires further investigation, this possibility expands our understanding of the diverse roles of C. innocuum in intestinal and extra-intestinal infections.

Stevens et al. [28] showed that subinhibitory concentrations of the ß-lactam antibiotic nafcillin induce the transcription of toxin genes and increase toxin production in Staphylococcus aureus. Furthermore, Chen et al. [29] showed that vancomycin and ampicillin treatments induce α-hemolysin expression and enhance the cytotoxicity of vancomycin-resistant S. aureus. In C. difficile, Gerber et al. [30] observed that sub-MIC metronidazole and vancomycin treatments were associated with earlier toxin production and increased toxin gene transcription. These results suggest that antibiotics at subinhibitory concentrations may serve as signals to activate bacterial toxin expression. As an intrinsic vancomycin-resistant bacterium, vancomycin would not be able to eliminate C. innocuum but act as an external signal to alter the phenotype of C. innocuum. This study revealed that biofilm-embedded C. innocuum exhibited survival even when exposed to high concentrations of vancomycin. These further suggest that C. innocuum within the biofilm can adapt its phenotype in response to vancomycin stimuli.

Conclusions

The pathogenetic mechanism of C. innocuum remains largely unknown. The present study showed the correlation between phylogenetical position and biofilm production activity and suggests that its intrinsic resistant to and robust biofilm production activity likely play a significant role, not only in its pathogenesis, but also in its interactions with intestinal commensals and pathogens.