Introduction

Colonoscopy is a physical and diagnostic method for colonic disease or routine physical examinations [1, 2]. There are 100 million colonoscopies performed in China per year. Bowel preparation is performed half a day before colonoscopy and consists of large doses of laxatives to evacuate most of the stool from the colon and obtain a clear picture during screening [3, 4]. Patients experience the excretion of watery feces 10–20 times [3]. During colonoscopy, a large amount of physiological saline is used to thoroughly flush the intestinal tract [2]. These processes cause drastic changes in gut microbiota and nearly one-third of adults and 10% of children experience complications during and after colonoscopy including abdominal pain and diarrhoea [5,6,7]. Therefore, it is necessary to explore the changes in the gut during and after colonoscopy and devise a strategy to reduce adverse reactions.

Anaerobic microorganisms in the mammalian large intestine ferment and produce a large number of short-chain fatty acids (SCFAs) including acetic acid, propionic acid, and butyric acid [8]. These SCFAs provide energy to the colon with butyric acid being the main energy source for intestinal mucosal cells [9, 10]. Butyric acid also promotes intestinal mucosal repair and functional recovery in human intestinal cells [11,12,13]. Clostridium, Eubacterium, Fusobacterium, and Bacteroidaceae are anaerobic bacteria which efficiently produce butyric acid [14,15,16]. The final steps in butyrate synthesis occurs through two main pathways: the butyryl-CoA:acetate CoA-transferase (but) pathway and the butyrate kinase (buk) pathway [17, 18]. Briefly, butyryl-CoA:acetate CoA-transferase transfers the CoA moiety to external acetate forming acetyl-CoA and butyrate, while butyrate kinase converts butyryl-CoA into butyrate [19]. Bacteria contain two buk homologs: buk I and buk II [20]. Previous detection of but and buk in gut microbiota has mainly focused on one type of bacteria or a genus of bacteria [21,22,23]. A systematic review of cultivated gut bacteria with the buk gene is needed given the important role of the gut microbiota in butyric acid metabolism.

In this study, we aimed to: (1) characterise gut microbiota and SCFA changes during and after colonoscopy; (2) identify bacteria encoding buk in human gut microbiota and investigate buk sequence variability among different bacteria; and (3) explore whether supplementation with butyric acid-producing Clostridium butyricum accelerates gut microbiome and SCFA stabilisation.

Materials and methods

Fecal samples and intestinal content collection

The research protocol was approved by the Ethics Committee of Nanfang Hospital, China (NFEC-2020-036). Eleven participants were enrolled in this study and their clinical characteristics are shown in Additional file 2: Table S1. The participants did not consume any food or medication containing probiotics within the 3 weeks prior to colonoscopy. In the group treated with C. butyricum (n = 7), C. butyricum administration began immediately after colonoscopy and lasted for 20 days. Three pills of C. butyricum oral capsules, live (Eastsea Pharmacology, Qingdao, China) were administered to each participant twice a day. Biosamples were provided by the Department of Gastroenterology at Nanfang Hospital of Southern Medical University. Fecal samples were collected before bowel lavage (Pre), during colonoscopy (d 0 h) and after colonoscopy (day 1, 3, 7, 14, 30, and 60) with participants asked to bring a ‘fresh’ stool sample to the research location within 6 h after collection (Fig. 1, Additional file 1: Fig. S1 and Additional file 2: Table S1). Intestinal contents collected during colonoscopy were extracted using a sterile injection syringe. Twenty-five mL was centrifuged at 12,000 × g for 5 min at 4 °C. The supernatant was aspirated and discarded and the remaining pellet was stored at − 80 °C. Samples were temporarily stored at − 80 °C at the research location before transportation to LC-Bio Technology Co., Ltd. (Hangzhou, China).

Fig. 1
figure 1

Overview of the study design and samples collection. Pre denotes the time point before bowel lavage, d 0 h denotes the time point during colonoscopy, while 1d, 3d, 7d, 14d, 30d, and 60d denote time points 1, 3, 7, 14, 30, and 60 days after colonoscopy

Full size image

16S ribosomal RNA (rRNA) gene sequencing

16S rRNA gene sequencing was performed as previously described [24]. Briefly, total genomic DNA from stool samples was extracted using a TIANamp Stool DNA Kit (TIANGEN Biotech, Beijing, China). The V4 region of the 16S rRNA gene was amplified using the forward primer 515F (5ʹ-GTGCCAGCMGCCGCGGTAA-3ʹ) and reverse primer 806R (5ʹ- GGACTACHVGGGTWTCTAAT-3ʹ). Eleven samples were sequenced on an Illumina NovoSeq6000 platform (Illumina, CA, USA) by LC-Bio Technology. The Quantitative Insights Into Microbial Ecology (QIIME2, v 2021.2) pipeline was employed to process the sequencing data, as previously described [25]. The lowest number of readings was normalized to the total number of reads obtained in the sequencing run. An operational taxonomic unit (OTU) table was further generated to record the abundance of each OTU in each sample and the taxonomy of these OTUs. OTUs containing less than 0.001% of total sequences across all samples were discarded. To minimize the difference of sequencing depth across samples, an averaged, rounded rarefied OTU table was generated by averaging 100 evenly resampled OTU subsets under 90% of the minimum sequencing depth for further analysis. Alpha diversity and principal coordinate analyses (PCoA) were conducted using a data visualisation web server ImageGP (www.ehbio.com/ImageGP/) [26].

SCFA extraction and GC/MS analysis

Fecal samples and intestinal contents from eleven participants stored at − 80 °C were thawed at room temperature. One gram of sample was mixed with 1 mL OmniSolv (EMD Millipore Corporation, Billerica, MA, USA) pure water, homogenized for 10 min at 4 °C using a water bath sonicator (Sonics & Materials, Connecticut, USA), then centrifuged at 15,000 × g for 10 min at 4 °C. One mL of supernatant (fecal homogenate) was transferred into a new 4 mL centrifuge tube containing 10 µL of 50% (v/v) sulfuric acid to bring the pH of the fecal solution to 2. The sample was spiked with 10 µL 2-ethyl butyric acid (100 μg/mL in methanol) as an internal control. The acidified fecal homogenate was extracted by adding 2 mL methanol, vortexed for 30 s, incubated on ice for 5 min, then centrifuged at 10,000 × g for 10 min at 4 °C. The upper layer containing SCFAs was transferred to a new microtube containing anhydrous Na2SO4 to remove any residual water.

GC/MS analysis was performed using a 7890 B gas chromatograph/5977 single quadrupole (Agilent Technologies, Santa Clara, CA, USA) with a DB-FFAP125-3237 capillary column (30 m × 0.53 mm × 0.5 µm film thickness) (Agilent Technologies). The injector, ion source, quadrupole, and GC/MS interface temperatures (°C) were 200, 230, 200, and 280, respectively. The flow rate of helium carrier gas was maintained at 1 mL/min. One microliter of the derivatized sample was injected with a 3.5 min solvent delay and a split ratio of 10:1. The initial column temperature (90 °C) was held for 1 min, ramped to 200 °C at 12 °C/min and held for 2.33 min. Ionization was carried out in the electron impact mode at 70 eV. The MS data were acquired in full scan mode from m/z 40–400 with an acquisition frequency of 12.8 scans per second. The analytes were quantified in the selected ion-monitoring mode. The target ion (m/z) of the SCFAs was previously described [27]. Data were analyzed using Agilent MassHunter software and the SCFA content was calculated using external standard methods. The spiked recoveries were calculated using the following equation: recovery % = (final concentration − initial concentration)/spiked concentration × 100%.

Taxonomic characterization of buk gene

The Culturable Genome Reference (CGR) database contains the reference genomes of 1,520 cultivated human gut bacteria, and it can be obtained from the NCBI database under accession code PRJNA482748 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA482748) [28]. De novo assembly was conducted according to a previous protocol. The genes and related proteins from these bacterial genomes were predicted using Prokka (v1.3) [29], and taxonomic information regarding these genes and proteins was directly extracted from the strain names. All buk sequences are shown in Additional file 3: Table S2. Pairwise amino acid sequence alignments and multiple alignments of buk sequences were performed using ClustalW in MEGA software (v11.0) [30]. Multiple sequence alignment was presented in Additional file 5. A phylogenetic tree was constructed using the maximum likelihood method and Kimura 2-parameter model in MEGA with 1000 bootstrap replications. An interactive tree of life (iTOL) (v5.0, https://itol.embl.de) was used to embellish the phylogenetic tree by adjusting the labels and colourisation [31]. The nine classified buk phylotypes represented were buk-T1-9. The n-buk represented bacteria with no buk gene in the genome.

RNA isolation and RT-qPCR

Total RNA was extracted from bacteria (0.1 g) using TRIzol reagent. RNA was converted to cDNA using a reverse transcription kit. Gene expression was determined using qPCR SYBR Green Master Mix and the 7500 real-time quantitative PCR system (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA). 16S was used for normalisation. Relative quantification of the target genes was performed using the 2−ΔΔCT method. The forward and reverse primers of buk were 5ʹ-TGCTGTWGTTGGWAGAGGYGGA-3ʹ and 5ʹ- GCAACIGCYTTTTTTGATTTAATGCATGG-3ʹ. The forward and reverse primers of 16S were 5ʹ-CCTACGGGNGGCWGCAG-3ʹ and 5ʹ-GACTACHVGGGTATCTAATCC-3ʹ.

Statistical analysis

Statistical significance between changes in bacterial abundance at the phylum level was determined using Student’s t-test. Statistically significant differences between groups were determined using the Wilcoxon rank-sum test with false discovery rate correction. Cluster analysis was performed using Bray–Curtis similarity. Statistically significant differences in SCFAs and buk at different time points were calculated using the paired-samples t-test. Relationship analysis was performed using Pearson correlations. All analyses were performed using R version 3.6.1 and statistical significance was set at p < 0.05.

Results

Fecal samples and intestinal contents were collected from four healthy untreated subjects at nine time points. The microbial diversity in the fecal samples fluctuated significantly over time both during colonoscopy and after colonoscopy in all human subjects, and especially in Ctrl 2 (Fig. 2a). The microbiota of each subject clustered together at the different time points indicating conservation of the gut microbiota amongst individuals (Fig. 2b). The microbial community prior to colonoscopy served to determine species turnover (i.e., the change in microbial structure) after colonoscopy. Firmicutes appeared on day 3 after colonoscopy and Bacteroidetes increased during colonoscopy (Fig. 2c). Firmicutes and Bacteroidetes changed significantly during colonoscopy and on days 1 and 3 after colonoscopy (P < 0.05, Fig. 2d). Almost all bacterial phyla were at their highest or lowest abundances on day 3 and Firmicutes decreased by 30% (p < 0.05, Fig. 2d). The Firmicutes to Bacteroidetes ratio decreased during colonoscopy and on the third day after colonoscopy then returned to baseline levels on day 7 (Additional file 1: Fig. S2). There were three clusters based on the longitudinal gradient of gut microbiota: the early, middle, and mature phases with changes in genera change patterns within each phase (Fig. 2e). The bacteria that underwent changes in the early phase (0 h and 1 d) were bacteria with quick recovery; the bacteria that increased in the middle phase (3 d and 7 d) were primary bacteria and formed a foundation for the successive bacteria; and the bacteria that reproduced in the mature phases (14 d, 30 d and 60 d) but had not completely recovered at the end of this period were members of long-term changes (Fig. 2e). Bacteria requiring long-term recovery included Bifidobacterium, Clostridium, Lachnospira, and Rothia (p < 0.05, Fig. 2f).

Fig. 2
figure 2

Longitudinal dynamics of human gut microbiota and SFCA in the control group. a Alpha diversity of control group gut microbiota over time. b PCoA analysis of control group gut microbiota at each time point. c Tracing the OTUs at each time point. Fecal samples of the former time point were taken as potential sources of the latter time point. The bands of each color indicate the different phylum. d The temporal changes of bacterial abundance at the phylum level and different time points. The significant points represent the changes from Pre. e Clustering of the gut microbiota at the genus level and different time points, a log10 transformation was performed on the OTU. f Significantly different genera at each time point, compared to the Pre time point with a p-value < 0.05 (using Wilcoxon rank-sum test), are shown. Colour in the heatmap represent z-values obtained from the Wilcoxon rank sum test. g The changes of SCFAs in fecal samples, normalized to Pre. *p < 0.05, **p < 0.01

Full size image

Butyric acid, propanoic acid, hexanoic acid, and heptanoic acid levels decreased in stool samples during day one and recovered close to baseline levels 1 month after colonoscopy (Fig. 2g). Most SCFAs increased from 3 to 14 d after colonoscopy, however, acetic acid levels remained steady over the entire time period (Fig. 2g).

Due to the long recovery time after colonoscopy, we hypothesised that Clostridium could be a good probiotic to help accelerate restoration of the gut microbiome and SCFAs. Bioinformatic analysis of buk gene was performed. Bacteria were assigned to 22 genera including Bacteroides, Clostridium, Coprococcus, Eubacterium, and Odoribacter from four phyla (Fig. 3a, Additional file 4: Table S3). More than half of the buk containing bacteria belong to the Bacteroidetes phyla (349 strains, 78.07%) and Firmicutes phyla (132 strains, 16.58%) (Additional file 1: Fig. S3). A total of 95.87 and 45.16% of the bacteria strains encoding buk belong to the Bacteroides and Clostridium genus, respectively (Fig. 3b). Meanwhile, 69.03% (321 strains) of Bacteroides and 8.60% (40 strains) of Clostridium strains encoded only one buk gene, while 45.45% (10 strains) of Clostridium and 18.18% (4 strains) of Bacteroides strains had two or three paralogous buk genes (Fig. 3c, d). This suggests that Clostridium strains have higher butyric acid production proficiency. The 515 buk genes in the CGR were classified into nine phylotypes by phylogenetic analysis with the number of sequences stated in parenthesis: buk—T1 (18), buk—T2 (37), buk -T3 (27), buk -T4 (79), buk -T5 (17), buk -T6 (5), buk -T7 (107), buk -T8 (104), and buk -T9 (120) (Fig. 3e). The results of qPCR indicated that two C. butyricum strains, not only the strain of C. butyricum used in this study, but also the strain of C. butyricum in the American Type Culture Collection, had significantly higher mRNA levels of buk compared with those in other probiotics (Fig. 3f).

Fig. 3
figure 3

Taxonomic characterization of buks obtained from the CGR database. a Taxonomic identification of buks in different main genera. b Quantity of buks containing bacterial strains at the genus level. c The proportion of bacterial strains at genus level with one buk gene. d The proportion of bacterial strains at genus level with two or three buk gene. e Taxonomic characterization of reclassified buks. The width of the clade indicates the value of bootstrap, normalized to 0–1. f The mRNA level of buk in different bacterias. *p < 0.05, **p < 0.01, ***p < 0.001

Full size image

C. butyricum is the only Clostridium species that has been commercialized and approved by the National Medical Products Administration in China. The Fecal microbial alpha diversity slightly fluctuated during and after colonoscopy in human subjects compared with the former Control group, except for CB 2 (Fig. 4a and b). The microbiota of each subject clustered together at different time points (Fig. 4c). Firmicutes decreased by 16.1% in C. butyricum-treated subjects on three days after colonoscopy compared with 27.9% in the control, while Bacteroidetes increased by 13.4% on three days after colonoscopy compared with 22.4% in control subjects (Fig. 4d and e). Firmicutes and Bacteroidetes changed significantly from during and 1d after colonoscopy (P < 0.05, Fig. 4e). Almost all of the bacterial phyla were at their highest or lowest abundance after 1 d in the C. butyricum-treated subjects, which is 2 d ahead of the Control group (Fig. 4e). The ratio of Firmicutes to Bacteroidetes decreased during the first day after colonoscopy and returned to baseline levels on the third day (Additional file 1: Fig. S4). Three clusters were grouped based on the longitudinal gradient of fecal samples: early phase (0 h), middle (1 d, 3 d and 7 d) phase and mature (14 d, 30 d and 60 d) phase (Fig. 4f). The middle gut microbiota phase indicated a faster recovery of gut microbiota changes beginning from 1 d. Bacteroides, Ruminococcus, Parabacteroides, Oscillospira, Veillonella, and Desulfovibrio increased in the middle phase as the first colonizers, while Coprococcus and Collinsella increased in the mature phase as successive colonizers (p < 0.05, Fig. 4g). Butyric acid and other SCFAs decreased during colonoscopy and recovered to baseline levels after 1–3 d, which is four days faster than the Control group (Fig. 4h). Collectively, these results show that C. butyricum accelerates the restoration of gut microbiota balance and provides SCFAs in the gut.

Fig. 4
figure 4

Longitudinal dynamics of human gut microbiota and SFCA in the Clostridium Butyricum (CB) treated group. a The alpha diversity of the CB group gut microbiota over time. b Chaos diversity index of gut microbiota in the control group and the CB group. c The PCoA analysis of CB group gut microbiota at each time point. d Tracing the OTUs in each time point. The bands of each color indicate the different phyla. e The temporal changes of bacterial abundance at the phylum level. The significant points represent changes from Pre. f Clustering of the gut microbiota at the genus level and different time points, a log10 transformation was performed on the OTU. g Significantly different genera at each time point, compared to the Pre time point with a p-value < 0.05 (using Wilcoxon rank-sum test), are shown. Colours in the heatmap represent z-values obtained from the Wilcoxon rank sum test. h The changes of SCFAs in fecal samples, normalized to Pre. *p < 0.05, **p < 0.01

Full size image

Furthermore, a direct statistical analysis was conducted between the control group of four participants and the experimental group of seven individuals supplemented with C. butyricum. The microbial diversity between the control and C. butyricum groups was not significantly different, as shown in Fig. 5a. However, the abundance of the phylum Actinobacteria was significantly higher, while the abundance of Bacteroidetes was significantly lower in the C. butyricum group at time point Pre (Fig. 5b). Notably, the abundance of Actinobacteria remained significantly higher in the C. butyricum group at 3 d, 7 d, and 30 d after colonoscopy when compared to the control group (Fig. 5b). Using LEfSe analysis, the Coriobacteriaceae family and Bifidobacterium genus were significantly higher in the C. butyricum group at time point Pre, while the Granulicatella and Neisseria genera were significantly higher during colonoscopy. The Roseburia genus was significantly higher in the C. butyricum group at 1 d, and the Bifidobacterium and Coprococcus genera were significantly higher at 3 d, with the Coprococcus genus continuing to maintain higher levels at 7 d and 14 d. At 30 d, Bifidobacterium and Rothia genera were significantly higher in the C. butyricum group (Fig. 5c). No statistically significant differences were observed in bacteria obtained from the LEfSe analysis at day 60d. Furthermore, the SCFA concentrations of the two groups were compared. The acetic acid levels in the C. butyricum group were significantly higher than those in the control group at d 0 h (Fig. 5d). Although there was no statistically significant, acetic acid levels in the C. butyricum group were higher than those in the control group at 30 d and 60 d. Hexanoic acid levels in the C. butyricum group were higher than those in the control group at 30 d, and valeric acid levels in the C. butyricum group were higher than those in the control group at 60 d (Fig. 5d).

Fig. 5
figure 5

Comparison of human gut microbiota and SCFAs in the control (Ctrl) and C. Butyricum (CB) treated groups. a The Shannon diversity index of the two groups over time. b Changes in bacterial abundance at the phylum level. Significant points indicate a statistically significant difference in phylum levels between the two groups at each time point. c Genera that differed significantly between the two groups at each time point were identified using LefSe analysis, with only the genera with LDA > 2 and p < 0.05 shown. d Changes in SCFA levels in fecal samples at different time points. *p < 0.05

Full size image

The fecal mRNA levels of the buk gene were examined in all study participants. The qPCR results revealed a noticeable reduction in the buk mRNA levels immediately following colonoscopy, followed by a gradual recovery over time (Fig. 6a, b). In subjects who did not receive C. butyricum supplementation, the buk gene levels returned to their baseline levels at fourteen days after the procedure (Fig. 6a). Conversely, in the C. butyricum-treated group, the buk gene levels reached the baseline as early as three days after the colonoscopy and steadily increased with continued consumption of the supplement (Fig. 6b). Additionally, a comparison of buk abundance between the two groups at Pre, 0 h,1 d, 7 d, 14 d revealed a higher level of buk in the C. butyricum group compared to that in the control group, although there is no statistical difference (Fig. 6c).

Fig. 6
figure 6

Transcription levels of buk in eleven subjects across all time points. a Transcription levels of buk in the control group at all time points, with all data compared to time point Pre and performed log10 transformation. Paired sample t-test was performed with Pre and other time points. b Transcription levels of buk in the C. Butyricum group at all time points, with all data compared to the Pre performed log10 transformation. Paired sample t-test was performed with Pre and other time points. c Transcription levels of buk in two groups at various time points. All data performed log10 transformation

Full size image

Discussion

There are few studies that have analyzed buk paralogs in bacterial genomes. We systematically analyzed 1520 cultivated bacteria in the human gut and identified that Bacteroides and Clostridium species have the buk gene indicating that they have the potential to generate butyric acid which is consistent with previous studies using these cultures [14,15,16].

SCFAs play an essential role in intestinal mucosal repair [11, 12]. However, little is known about SCFA changes during or after colonoscopy. Butyric acid is the most common SCFA generated by gut microbiota. In this study, dynamic changes in SCFAs including butyric acid were observed one day after colonoscopy, and SCFAs may require 1 week to recover to baseline levels. Therefore, complications arising one week after colonoscopy may be alleviated by supplying SCFAs, fatty acid salts, or other biological agents to patients.

Despite several studies showing changes in intestinal microbiota two-eight weeks after colonoscopy in adults [32], little attention is paid during the early stages (1–7 d) after colonoscopy. Bowel preparation before colonoscopy resulted in substantial immediate changes in the intestinal microbiota which persisted 2 weeks and 4 weeks after colonoscopy in a randomized controlled trial involving 23 healthy subjects [33]. The total microbial load significantly decreased and nearly one-fifth of participants lost subject-specificity of their gut microbiota. There was a significant decrease and increase in Firmicutes and Proteobacteria abundance, respectively immediately after colon cleansing and 1 month after colonoscopy using 16S rRNA Ion Torrent profiling of fecal samples from 10 patients [34]. In this study, dynamic changes in gut microbiota were observed between the first and seventh days after colonoscopy using 16S rRNA sequencing. There was a significant decrease in Firmicutes and an increase in Bacteroidetes after 3 days. The middle phase (3 d) was distinct from the early (1 d) and mature (7 d) phase. Bifidobacterium, Clostridium, Lachnospira, and Rothia were identified as bacterial members involved in long-term changes. The use of these bacteria as a probiotic was hypothesized to accelerate the restoration of the gut microbiota balance.

Bowel preparation before colonoscopy causes microbiota-related symptoms for days, resulting in the loss of working days. Therefore, there is a need to rapidly counterbalance microbiota alterations. Probiotics may offer a potential therapeutic option to restore the altered gut microbiota. For example, tablets containing Bifidobacterium tetragenous restore intestinal dysbacteriosis on the seventh day after bowel preparation [35]. A combination of Lactobacillus acidophilus and Bifidobacterium species producing lactic acid results in significant pain reduction following colonoscopy, shortened from 1.99 to 2.78 days [36]. However, a probiotic mixture (Lactobacillus plantarum, Lactobacillus lactis subspecies cremoris, and Lactobacillus delbrueckii) only has a benefit in patients with pre-existing symptoms and does not help ease post-procedural symptoms in healthy subjects [37]. In this study, patients supplemented with C. butyricum had significantly improved microbiotal balance with butyric acid levels restored significantly faster than those in untreated patients and B. tetragenous treated patients as previously reported (see Additional file 5).

This study has three main limitations. First, only 11 subjects provided continuous fecal samples which may have resulted in sample bias, and additional subjects are required to draw more accurate conclusions. Conducting these types of studies requiring consistent fresh fecal samples was challenging given that individuals with normal colonoscopy results do not visit hospitals frequently. We proposed recruiting college students or clinical staff members as potential study participants to increase the sample size. Second, the symptoms of the participants after colonoscopy were not recorded. Third, the molecular mechanism by which C. butyricum resulted in quicker recovery and stabilisation of gut microbiota and fecal SCFAs remains unknown. This unresolved issue will be the focus of future research.

This study revealed the changes in microbiota and SCFA metabolites from stool samples in the early stages (one week) after colonoscopy. C. butyricum produces more butyric acid and other SCFAs than other probiotics and has a potential therapeutic role in restoring the balance of the gut microbiome. Routine application of C. butyricum after colonoscopy might be a potentially effective method to limit the complications associated with this procedure. Future investigations should focus on improving restoration of the gut microbiome by supplying probiotics and prebiotics.