What this is
- This research investigates how the human gut microbiota metabolizes dietary precursors like choline and L-carnitine into trimethylamine ().
- Using an in-vitro batch fermentation model, the study measures production over 24-48 hours.
- Findings reveal distinct metabolic pathways for production from various substrates, with choline being the most efficiently converted.
Essence
- Choline is the primary substrate for production in the human gut, followed by L-carnitine and Ī³-butyrobetaine. The model accurately reflects human metabolic pathways for production.
Key takeaways
- production from dietary precursors occurs primarily through the choline -lyase pathway, with minimal involvement from other pathways.
- L-carnitine conversion to is slower than that of choline, requiring Ī³-butyrobetaine as an intermediate.
- Betaine does not serve as a significant intermediate for production from choline or L-carnitine in humans.
Caveats
- The study's model only simulates conditions in the colon, potentially overlooking transformations in the upper gut.
- Variability in production across different stool samples suggests that results may not be universally applicable.
Definitions
- TMA: Trimethylamine, a compound produced by gut microbiota from dietary precursors.
- TMAO: Trimethylamine N-oxide, a metabolite formed from TMA in the liver, associated with metabolic disease risk.
AI simplified
Introduction
It is well established that the human gut microbiota produce substrates that are both beneficial and deleterious to health [1ā3]. In particular, the metabolism of dietary substrates such as choline and l-carnitine to trimethylamine (TMA) and subsequently to trimethylamine oxide (TMAO) by a hepatic enzyme flavin-containing monooxygenase 3 (FMO3) is of interest due to the strong associations of high circulating TMAO concentrations with the risk of numerous metabolic diseases and the risk of death from heart failure [4ā6]. Whether or not TMAO is causally linked to disease risk remains contentious, but there is a growing consensus that changes in microbial production of TMA are indicative of alterations in gut microbiota composition and diversity [7ā10], i.e. that high levels of circulatory TMAO reflect gut dysbiosis in some way. In addition, because they are essential for methylation processes, lipid membrane generation, lipid metabolism and neurotransmitter production, excessive metabolism of choline and l-carnitine by the gut microbiota may deprive the host of these essential substrates and increase the risk of metabolic diseases such as diabetes, cardiovascular diseases and neurodegenerative diseases [11]. Accordingly, there is considerable interest in developing effective and sustainable interventions that effectively reduce TMA production, and, therefore, host TMAO burden.
The requirement for the gut microbiota in the metabolism of choline to TMA was elegantly demonstrated by Tang et al. who showed that the metabolism of phosphatidylcholine (PC)/lecithin to TMA, and subsequently TMAO, could be suppressed by the administration of broad-spectrum antibiotics, without suppressing the production of choline and betaine [10]. Metagenomic and transcriptomics approaches have played a crucial role in identifying genes, enzymes and pathways that are involved in TMA production [12, 13]. However, the presence of genes in microbial genomes that are predicted to be capable of degrading choline and l-carnitine is not always indicative of functional capacity of the microbe to produce TMA [14, 15]. Consequently, animal, human and in-vitro single-strain studies have been employed to functionally characterise the pathways and species involved in the metabolism of choline, l-carnitine, betaine and Ī³-BB. Of note are the requirement of a glycyl radical enzyme choline TMA-lyase (CutC) along with its activating enzyme (CutD) for the metabolism of choline to TMA, [15ā17]. The importance of bacterial microcompartments in choline metabolism has also been demonstrated using bacterial pure cultures [18]. In mice, using PC as a substrate, it was reported that TMA can be formed as a result of multiple steps (PC ā choline ā betaine ā TMA), although betaine was a poor substrate for the production of TMAO [19, 20]. Another study incorporating experiments with single bacterial strains, live human faecal microbes and mouse caecal cell lysates provided evidence that the metabolism of choline to TMA could be inhibited by a choline analogue 3,3-dimethyl-1-butanol (DMB) [21]. Nevertheless, in humans it is still unclear which pathways are predominant for the metabolism of choline to TMA. The absolute requirement of the gut microbiota in the metabolism of l-carnitine to TMA has also been demonstrated, in single-strain, animal and human studies [22, 23]. Using a combination of bioinformatics and strain-specific approaches, it was shown that the Rieske-type two component l-carnitine oxygenase/reductases (CntA/Bs) are required for microbial dependent degradation of l-carnitine to TMA [14]. In another report, it was shown that the concentration of TMAO in plasma peaked 24 h (h) after oral l-carnitine administration [22]. Using both human and animal studies, Koeth et al.provided evidence that Ī³-BB is an obligate intermediate formed during the metabolism of l-carnitine to TMA with the major site for Ī³-BB production being the small bowel, proximal to that of TMA production [22].
Here we describe the use of an in-vitro batch fermentation model to study the metabolic pathways for TMA production from dietary substrates such as choline, l-carnitine, betaine and Ī³-BB in humans. This model has been widely used to investigate the microbial metabolism of various other nutrients [24, 25]. We hypothesised that the conversion of these substrates into TMA in a batch model faithfully replicate the findings from previously reported studies of TMA production in human, animal and in-vitro studies. In testing the human colon model for this purpose, we anticipate being able to study the human gut microbiota-dependent conversion of these substrates to TMA in more detail, and demonstrate that this model is suitable for studies that seek to find dietary interventions that reduce the production of TMA. Our specific objective was to measure the changes in the concentration of TMA and associated metabolites over an extended period (24 h) in the colon model when incubated with TMA precursors. We also explored whether DMB, which had previously been shown to inhibit TMA production from choline in-vivo and in single-strain models [21], would inhibit the metabolism of choline to TMA in the human in-vitro colon model. In addition, we investigate whether the conversion of these substrates to TMA could be replicated using a simple anaerobic environment without pH control.
Materials and methods
Materials
Trichloroacetic acid (TCA), heptafluorobutyric acid (HFBA), Ī³-BB (3-(carboxypropyl)trimethylammonium chloride) and choline chloride-(trimethyl-d9) were purchased from SigmaāAldrich. All unlabelled methylamines except Ī³-BB, all solvents unless otherwise stated, acetic acid, and ammonium acetate were purchased from Fisher Scientific Limited. Labelled internal standards [(d9-TMA), d9-TMAO and d9-carnitine] were purchased from Cambridge Isotope Laboratories. Fermac 260 pH control units were from Electrolab, and Stomacher 400 from Seward. 3,3-Dimethyl-1-butanol (DMB) was purchased from VWRā¢.
Study design
An in-vitro batch fermentation model was used to study the microbial metabolism of choline, betaine, l-carnitine and Ī³-BB. Faecal samples were obtained from participants recruited onto the QIB Colon Model study. Men and women aged 18 years or older who live/work within 10 miles of the Norwich Research Park were recruited if they satisfied the following criteria: participants who were assessed to have a normal bowel habit, regular defecation between three times a day and three times a week, with an average stool type of 3ā5 on the Bristol Stool Chart, and no diagnosed chronic gastrointestinal health problems, such as irritable bowel syndrome, inflammatory bowel disease, or coeliac disease. Participants were asked further questions immediately prior to donating a stool sample to confirm that they had not taken antibiotics or probiotics within the last 4 weeks, had not experienced a gastrointestinal complaint, such as vomiting or diarrhoea, within the last 72 h, were not currently pregnant or breast-feeding, and had not recently had an operation requiring general anaesthetic. The study was approved by the Quadram Institute Bioscience (formally Institute of Food Research) Human Research Governance committee (IFR01/2015), and LondonāWestminster Research Ethics Committee (15/LO/2169). The informed consent of all participating subjects was obtained, and the trial is registered at http://www.clinicaltrials.govā (NCT02653001ā). Twenty-one fresh human faecal slurry from five different donors were used for choline inoculations. For l-carnitine, betaine and Ī³-BB inoculations, samples from three donors were used in eight, four and six fermentation experiments, respectively, with each donor being used at least once for each substrate.
Each substrate was added to vessels to give a target concentration of 2000 Āµmol/L. A blank vessel was inoculated with PBS and human faecal slurry. Concentrations of TMA and the related precursors were measured at baseline, then at 2, 4, 6, 8, 10, 12 and 24 h. In three subsequent experiments, 2000 or 10,000 Āµmol/L DMB was added to vessels treated with choline. Finally, similar substrate concentrations for choline, l-carnitine, betaine, Ī³-BB were also used for the anaerobic fermentations without pH control.
In-vitro batch fermentations
| Substrate | Theoretical conc(Āµmol/L)n | Measured in 1% faecal matrix(Āµmol/L) | Measured in PBS(Āµmol/L) | Measured TMA conc(Āµmol/L)n | Sum of substrate+āTMA (Āµmol/L)(PBS) |
|---|---|---|---|---|---|
| Blank | 0 | n.d | n.d | n.d | 0 |
| Choline | 2000 | 1742 | 1685 | 234 | 1919 |
| Betaine | 2000 | 1883 | 1936 | n.d | 1936 |
| -Carnitinel | 2000 | 1658 | 2087 | n.d | 2087 |
| Ī³-BB | 2000 | 1536 | 1611 | n.d | 1611 |
Metabolite quantification using LCāMS/MS
| Compound | Retention time (min) | Precursor ion (m/z) | Product ion (m/z) |
|---|---|---|---|
| Betaine | 0.84 | 118.2 | 58.1 |
| -Carnitinel | 0.95 | 162.2 | 43.1 |
| D9--Carnitinel | 0.93 | 171.2 | 43.1 |
| Choline | 1.59 | 104.2 | 45.1 |
| D9-Choline | 1.59 | 113.8 | 49.2 |
| TMA | 1.7 | 60.6 | 44.1 |
| D9-TMA | 1.68 | 69.4 | 49.2 |
| TMAO | 1.43 | 76.1 | 58.1 |
| D9-TMAO | 1.45 | 85.2 | 66.1 |
| Ī³-BB | 1.12 | 146.2 | 86.9 |
Statistical analysis
To assess whether choline, betaine, l-carnitine and Ī³-BB are at all metabolised to TMA in the multi-species batch fermentation human in-vitro colon model, we compared the increase in the concentrations of TMA over 24 h to the increase in the untreated āBlankā fermentations. To account for within-sample correlation in outcomes, and the observation that standard deviation of TMA concentration was proportional to the mean values in each condition, a linear weighted mixed model was used to estimate the differences between production from different substrates, with regression weights set to a power of the residual variance. A random effect of substrate for groups defined by sample was included in the model. The average marginal effect of adding each substrate on TMA produced was estimated, as well as all pairwise differences between substrates. The packages nlme (version 3.1ā139) and emmeans (version 1.3.4) in R statistical software (version 3.6.1) were used to perform the statistical analysis [26ā28]. All trajectories of metabolite levels over time periods under each treatment condition, and the difference in choline metabolism without DMB or with 2 or 10 mM/L DMB are described graphically, both as means within treatment groups and for each fermentation individually.
Results
Significant production of TMA from choline,-carnitine and Ī³-BB but limited production from betaine l
Change in TMA between 0 and 24Ā h in each batch fermentation seeded with faecal samples of five different donors. Fermentation stratified by substrate added. Colours correspond to donors and marker shapes to individual experiments. There is a high intra-class correlation between replicate fermentations within experiments, but little intra-class correlation within donors beyond this. Variation in the levels of TMA produced across experiments is high for each substrate, despite similar levels of substrate being utilised
Estimated mean TMA produced from each substrate over 24Ā h. Estimates were obtained by mixed effects regression models of the difference between TMA concentration at 0 and 24Ā h (as described in the statistical methods). The effect of each substrate is calculated with correction for the production of TMA observed in blank vessel without added substrates. Error bars represent 95% confidence intervals
Metabolic pathways for individual substrates
The average trajectory of all metabolites. With no added substrate () and following supplementation with each substrate (ā). Error bars represent standard errors a b e
The direct choline TMA-lyase pathway is the predominant route of colonic TMA production from choline with little or no contribution from the betaine aldehyde-choline dehydrogenase-betaine reductase pathway
In the choline inoculated fermentations, the concentration of choline started to decline between 6 and 8 h and continued to do so until 24 h by which time it was almost completely utilised in all experiments (Fig. 3b). TMA levels started to increase from the baseline concentration between 6 and 8 h which corresponded with the decrease in baseline levels of choline. It is possible that betaine could be an intermediate in the metabolism of choline to TMA which in theory can occur via the aldehyde/choline dehydrogenase /betaine reductase pathway. However, we did not observe a significant decrease in betaine values when it was used as the substrate (Fig. 3c). However, we did observe modest increases in the concentrations of TMA in betaine-supplemented incubations, and although some of this could be accounted for by reductions in the concentrations of choline and other substrates present in the baseline samples, the final concentration of TMA was significantly higher that observed in the blank incubations, suggesting that some of this TMA may have come from the metabolism of betaine. Additional evidence that there was at least a slow production of TMA from betaine was the observation that background betaine concentrations declined slowly across 24 h in all the non-betaine-supplemented incubations (Fig. 3a, b, d, e). It is also interesting to note that the absolute concentration of betaine in baseline samples, which originates from the faecal inoculum, was substantial (almost 500 Āµmol/L), which is consistent with a very slow rate of conversion of betaine to TMA and, therefore, the accumulation of unmetabolized betaine in the human colon. Overall, and taken together with the observation here that appearance of TMA closely mirrored the disappearance of choline, the data presented do not support the notion that the indirect pathway via betaine is a significant route of TMA production from choline by human gut microbiota.
The metabolism of-carnitine to TMA requires the Ī³-butyrobetainyl-CoA:carnitine CoA transferase-carnitine TMA-lyase pathway with formation of Ī³-BB as an intermediate and does not occur via the Rieske-type monooxygenase or the-carnitine dehydrogenase pathway l l
In the l-carnitine supplemented fermentations, the concentration of l-carnitine declined steadily from 6 h until 12 h, by which time it was no longer detectable (Fig. 3d), whilst only a minimal increase in the concentration of TMA was observed during this time. However, there were substantial increases in Ī³-BB concentrations up to and peaking at 12 h, and these closely corresponded with the disappearance of l-carnitine. Taken together, the observations that (i) the choline observed at baseline also declined from 6 h, (ii) at no timepoint did the concentration of betaine increase and (iii) the concentration of betaine was lower at 24 h than at 0 h, all support the notion that little if any betaine is formed from l-carnitine by the human gut microbiota (Supplementary Fig. 4). Therefore, the production of TMA from carnitine likely occurs via the Ī³-butyrobetainyl-CoA:carnitine CoA transferase-carnitine TMA-lyase pathway with the formation of Ī³-BB as an intermediate.
y-BB is metabolised to TMA but not to-carnitine, and at a slower rate than the conversion of choline to TMA l
In the Ī³-BB-supplemented ermentations, the concentration of Ī³-BB declined slowly and by only approximately one third (equivalent to ~ 450 Āµmol/L) after 24 h, following a 6ā8 h lag phase (Fig. 3e). A minimal increase in TMA concentrations were observed across 0ā8 h and then increased moderately up to 24 h, reaching a concentration similar to the loss of Ī³-BB (Fig. 3e). l-carnitine concentrations did not change over the 24 h period. Similar to that observed with choline, the background choline in Ī³-BB-supplemented fermentations declined after a lag of around 8ā10 h (Supplementary Fig. 4). These observations are consistent with TMA being produced directly from Ī³-BB (albeit rather slowly) and not via l-carnitine as an intermediate.
The concentrations of TMA from l-carnitine and Ī³-BB inoculated fermentations were also measured at 48 h in a single experiment with two replicates per treatment (Supplementary Fig. 5). In the l-carnitine supplemented fermentations, production of TMA increased further between 24 and 48 h to a final concentration of 1541 Āµmol/L, which corresponded to 93 mol% of the starting concentration of available l-carnitine (1658 Āµmol/L, Table 1). In the Ī³-BB-supplemented fermentations, the concentration of TMA reached 1311 Āµmol/L at 28 h, accounting for 85 mol% of available Ī³-BB (1536 Āµmol/L, Table 1). These observations demonstrate that the rate of production of TMA from choline > carnitine > Ī³-BB, which is reflected in the concentrations of TMA at 24 h, but that given longer incubation (e.g., 48 h) there is almost complete transformation of all three substrates.
No evidence of inhibition of choline metabolism to TMA by dimethyl-1-butanol (DMB)
The effects of DMB on TMA production from choline in three independent experiments. The graph is stratified by experiment to enable a direct comparison between paired fermentations with and without DMB added. There was no effect of adding DMB on TMA production or choline concentration at any stage over the time period tested
pH control is required TMA production in batch fermentation of human colon models
The fermentation of TMA substrates under anaerobic conditions without pH control. There was no fermentation of any of the substrate and no TMA production in anaerobic conditions without pH control, although only a few experiments were carried out (=ā3 for choline and 2 for betaine,-carnitine and Ī³-BB) n l
Discussion
The overall aim of the research reported here was to investigate the metabolism of choline, l-carnitine, betaine and Ī³-BB to TMA in an in-vitro batch fermentation model, and in doing so determine how closely the model replicates the production of TMA that has been previously reported in humans. The main findings were that (i) TMA was produced from all three dietary precursors, as well as from the metabolic intermediate Ī³-BB, (ii) that the relative rates of production of TMA from the substrates was choline > carnitine > Ī³-BB > betaine, (iii) that all the previously described metabolic routes to TMA production reported in humans were replicated in this in-vitro pH-controlled batch colon model and (iv) that no metabolism to TMA occurred from any substrate when fermentations were performed in the absence of pH control.
Unlike in mice, we show that betaine in humans is not formed as an intermediate of choline to TMA metabolism. Instead, almost all choline is metabolised via the direct choline to TMA route, potentially involving the choline TMA-lyases, suggesting that it is the predominant route. There was minimal TMA production in the betaine inoculations and very little betaine disappearance, indicating that in humans the contribution of betaine to TMA production is minimal if at all. Interestingly, background betaine also declined in the inoculations where choline, and l-carnitine were used as substrates. This may suggest that betaine metabolism requires bacteria or enzymes involved in the metabolism of choline, l-carnitine, and Ī³-BB, i.e., that are only induced in response to available substrates.
In line with others, we demonstrate that the metabolism of l-carnitine to TMA involves the formation of Ī³-BB as an intermediate and that the Rieske-type carnitine oxygenase/reductase pathways are not involved in this metabolism by human colonic microbiota [22].
We further demonstrate that while the metabolism of choline to TMA is fast, the production of TMA from l-carnitine is much slower. We provide evidence that this is due to the requirement for Ī³-BB to be formed as an intermediate. More specifically, we have shown that the rate of conversion of Ī³-BB to TMA is very slow (Fig. 3e), and this is consistent with our observation of significant accumulation of Ī³-BB in carnitine supplemented fermentations (Fig. 3d), i.e., the very slow rate of conversion of Ī³-BB to TMA is rate-limiting in carnitine conversion. Nevertheless, if given sufficient time for complete conversion (24 h for choline and 48 h for l-carnitine), similar levels of TMA were achieved. In a review by Fennema et al. it was proposed that betaine may be formed as an intermediate of l-carnitine by l-carnitine dehydrogenase and metabolised further by betaine reductase to form TMA [29], but this was not observed in the model used here.
After inoculation, the concentration of choline started to decline at 6 h, and in most cases it had completely disappeared within 20 h. Increases in TMA concentrations started after 6 h and coincided with the timepoint at which choline concentrations started to decline, strongly suggesting that TMA was directly synthesised from choline. Betaine has been proposed as a possible intermediate in the metabolism of choline to TMA in mice [19ā21, 29, 30]. However, at no time-point did we detect betaine in choline supplemented fermentations. Since the rate of production of TMA from betaine was slow relative to the rate of production of TMA from choline (Fig. 3b, c, respectively), these data are in keeping with the notion that all the TMA produced in choline supplemented fermentations is a result of direct conversion via a single enzyme-catalysed step, i.e., via the choline TMA-lyase (CutC/CutD).
As in all experiments where choline was not supplemented as a substrate, background choline declined significantly, we can postulate that some of the TMA may have originated from choline and some from the metabolism of betaine via the glycine betaine transmethylase and then to TMA by decarboxylation. However, we did not measure the concentration of dimethylglycine in this study. Further studies using isotopically labelled betaine would be required to ascertain this. As the expression of CutC/D enzymes has been demonstrated in 100% of all people examined, while between 6 and 21% of people are estimated to lack the betaine metabolising grdH enzymes, it is possible faecal samples from participants involved in our study lacked the grdH enzymes [31, 32].
Interestingly, although there were high intra-class correlation in TMA production between replicates within the same experiment there were no such correlations with respect to different stool samples from the same donors. This indicates that the capacity to produce TMA in faecal samples collected from the same donor on different days was variable. This is consistent with previous reports that have shown there is high intra-individual variation in plasma TMAO levels over a 1ā2 year period, although the source of these variations is not clear and warrants further investigation [33ā35]. Nevertheless, future human studies investigating the impact of interventions aimed at reducing TMA production will need to take these variabilities into account by multiple sampling over a period of time and using stable isotopes where necessary.
We sought to replicate previous studies to determine whether DMB could also inhibit the metabolism of choline to TMA in the in-vitro batch fermentation model. We were unable to show inhibition of choline metabolism to TMA by DMB at 2 and 10 mM. This shows that investigating TMA precursors in the context of complex bacterial mixtures is crucial as only a fraction of bacterial strains may be inhibited by DMB [21]. Wang et al. further demonstrated that DMB administered orally, but not subcutaneously, was effective at reducing TMAO production from choline indicating that this is a gut microbial dependent inhibition [21]. These discrepancies could be due to differences in the microbiota composition between humans and animals.
In the fermentations with l-carnitine we demonstrate the disappearance of l-carnitine from 6 h after the start of the fermentations, which did not correspond with the observed increase in TMA concentration. However, the disappearance of l-carnitine corresponded with the appearance of Ī³-BB at 8 h which peaked and plateaued at 12 h and slowly disappeared at 20ā24 h, corresponding with TMA production. These observations confirm previous findings in animal and human studies showing that the formation of Ī³-BB as an intermediate of l-carnitine was obligatory for the formation of TMA. We further confirmed the direct metabolism of Ī³-BB to TMA by carrying out fermentations in the presence of Ī³-BB substrate, and we show metabolic dynamics similar to choline where the disappearance of Ī³-BB corresponds with the appearance of TMA, albeit at a far lower rate in comparison to choline at 24 h. In contrast to the proposed pathway of TMA formation from Ī³-BB via l-carnitine, where Ī³-BB may be metabolised to l-carnitine by Ī³-BB hydroxylase, which is subsequently metabolised to TMA by the Rieske-type carnitine reductase/oxidase, there was no significant formation of TMA production at 8 and 12 h when l-carnitine was completely used up [14, 29]. In addition, at no point did we observe an increase in l-carnitine in the Ī³-BB fermentations. This once again highlights the need to study microbial functions in the context of the complex microbiota. These observations complement studies by Koeth et al. and Rath et al. who demonstrated that only a small proportion of individualās stool samples expressed l-carnitine oxygenase CntA/B and that their function is dependent on the availability of molecular oxygen [12, 22].
![Pathways for the metabolism of choline, betaine,-carnitine and Ī³-BB by human gut microbiota. This is based on data reported here and previously by others [,,,ā,]. In humans, choline is metabolised to TMA via the choline TMA-lyase pathway, betaine is not formed as an intermediate of choline to produce TMA via the choline dehydrogenase (CHDH)/betaine aldehyde dehydrogenase (BADH)ā>ābetaine reductase pathway, nor is it formed as an intermediate of-carnitine via the-carnitine dehydrogenase pathwayThere is no direct conversion of-carnitine to TMA via the Rieske-type C-carnitine oxygenase/reductase pathway, instead-carnitine is first converted to Ī³-BB by Ī³-butyrobetainyl-CoA:carnitine CoA transferases which is then converted to TMA by the-carnitine TMA lyases. It is possible that betaine may be converted to dimethylglycine by glycine betaine transmethylase and then to TMA by decarboxylation although the evidence for this is weak. Dashed black lines are pathways shown not to be functional in this model; solid green lines indicate pathways we have demonstrated to be important for TMA production in thehuman colonic fermentation model l l l l l l l [5] [10] [12] [19] [22] [32] . in-vitro](https://europepmc.org/articles/PMC8437865/bin/394_2021_2572_Fig6_HTML.jpg.jpg "Click to view full size")
Pathways for the metabolism of choline, betaine,-carnitine and Ī³-BB by human gut microbiota. This is based on data reported here and previously by others [,,,ā,]. In humans, choline is metabolised to TMA via the choline TMA-lyase pathway, betaine is not formed as an intermediate of choline to produce TMA via the choline dehydrogenase (CHDH)/betaine aldehyde dehydrogenase (BADH)ā>ābetaine reductase pathway, nor is it formed as an intermediate of-carnitine via the-carnitine dehydrogenase pathwayThere is no direct conversion of-carnitine to TMA via the Rieske-type C-carnitine oxygenase/reductase pathway, instead-carnitine is first converted to Ī³-BB by Ī³-butyrobetainyl-CoA:carnitine CoA transferases which is then converted to TMA by the-carnitine TMA lyases. It is possible that betaine may be converted to dimethylglycine by glycine betaine transmethylase and then to TMA by decarboxylation although the evidence for this is weak. Dashed black lines are pathways shown not to be functional in this model; solid green lines indicate pathways we have demonstrated to be important for TMA production in thehuman colonic fermentation model l l l l l l l [5] [10] [12] [19] [22] [32] . in-vitro
Conclusion
Using the pH controlled in-vitro batch fermentation human colon model, we show that in humans, the choline TMA-lyase pathway is the major pathway for the production of TMA from choline. Unlike in mice, betaine is not formed as an intermediate of choline for the production of TMA, nor is betaine produced as an intermediate of l-carnitine to produce TMA, although very small quantities of TMA may be produced from betaine. Complementing previously reported data from human and animal studies we show that Ī³-BB is an intermediate for the metabolism of l-carnitine to TMA. We could not show the direct metabolism of l-carnitine to TMA via the Rieske-type carnitine oxygenase/reductase pathway. While strain-specific models are important for identifying bacterial species and compartments for the production of TMA from its dietary precursors, they must always be accompanied by models which include complex microbiota mixtures similar to those found in the human gut and under pH regulation. This model offers an invaluable tool for increasing our understanding of the metabolic pathways of TMA production including identifying further microorganisms that are involved in TMA production and metabolism, such as Archaea. We envisage that this model will also be useful for studies aiming to identify potential dietary or pharmacological inhibitors of TMA production in the human gut.
Strengths and limitations
One strength of this study is that we were able to use a validated human colon model to investigate the metabolism of TMA precursors separately and investigate their microbial dependent degradation in the presence of complex microbiota species and independent of host enzymes that may also degrade some of these TMA precursors. Another is that the results presented here are consistent with those reported previously from in-vivo studies using human participants and animals, with the exception of betaine, which we show is not an intermediate in the metabolism of choline and l-carnitine. A weakness of this research is that the colon model only reflects conditions in the colon, and our observations may not reflect transformations that potentially occur in the upper gut, particularly the ileum, where the structure of the bacterial communities differ. However, the consistency of our observations with those from animal and human intervention studies suggests that it is an excellent model for studying TMA production in the human colon.
Supplementary Information
Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 633 kb)