Bacterial microcompartment utilization in the human commensal Escherichia coli Nissle 1917

ABSTRACT

Bacterial microcompartments (BMCs) are self-assembled protein structures often utilized by bacteria as a modular metabolic unit, enabling the catalysis and utilization of less common carbon and nitrogen sources within a self-contained compartment. The ethanolamine (EA) utilization (eut) BMC has been widely demonstrated in enteropathogens, such as Salmonella enterica, and current research is exploring its activity in the commensal species that populate the human gut. Escherichia coli Nissle 1917 (EcN) is a strong colonizer and probiotic in gut microbial communities and has been used extensively for microbiome engineering. In this study, the utilization of ethanolamine as a sole carbon source and the formation of the eut BMC in EcN were demonstrated through growth assays and visualization with transmission electron microscopy. Subsequently, flux balance analysis was used to further investigate the metabolic activity of this pathway. It was found that not only is the utilization of the eut BMC for the degradation of EA as a carbon source in EcN comparable with that of Salmonella enterica but also that ammonium is released into solution as a byproduct in EcN but not in S. enterica. Control of EA-dependent growth was demonstrated using different concentrations of the operon inducer, vitamin B₁₂. We show that vitamin B₁₂-dependent EA utilization as the sole carbon source enables growth in EcN, and demonstrate the concurrent formation of the BMC shell and inducible control of the eut operon.

IMPORTANCE

The human gut is a complex environment of different bacterial species, nutrient sources, and changing conditions that are essential for human health. An imbalance can allow for the emergence of opportunistic pathogens. Bacterial microcompartments (BMCs) are utilized by bacteria to metabolize less common nutrients, conferring a growth advantage. Although widely studied in enteropathogens, there is limited research on BMC activity in commensal species. We demonstrate the formation of the eut BMC and utilization of ethanolamine as a carbon source in the human gut commensal Escherichia coli Nissle 1917 (EcN). Additionally, we found increased ammonium production when EcN utilized ethanolamine but did not see the same in Salmonella enterica, highlighting potential differences in how these species affect the wider microbial community.

INTRODUCTION

The human gut comprises a complex interplay of microorganisms, metabolites, and enzymes, where dysbiosis (disruption in the healthy state of the microbiota) is linked to a multitude of health problems (1–4). The ecology of this community is complex and can be difficult to restore when perturbed. For example, under specific circumstances, a commensal strain of bacteria may become an opportunistic pathogen, which requires the administration of antibiotics to remove (5). As many antibiotics are non-specific, they can have widespread detrimental effects throughout the gut microbial community, as well as contributing to the increasing prevalence of antimicrobial resistance (6). Therefore, to understand how to control and treat dysbiosis in a more nuanced predictable manner, it is important to understand not only the colonisation mechanisms of these bacteria but also their interactions within the gut environment.

Escherichia coli Nissle 1917 (EcN) plays an important role in preventing pathogen-driven dysbiosis, where EcN is commonly used as a probiotic and target of bioengineering to promote and maintain a diverse and stable gut microbial population (7–9) and has been utilized as a chassis for therapeutics currently in phase 1 trials (10). EcN both directly and indirectly antagonizes multiple enteropathogens through competition and in preventing systemic circulation (11, 12), as well as reducing the invasion capacity of Shiga-like toxin E. coli, Listeria monocytogenes, Salmonella enteritidis, and Legionella pneumophila (13–15). Additionally, EcN elicits anti-inflammatory and immune responses (16, 17), as well as contributing to the healthy function of the intestinal epithelial barrier (18).

Bacterial microcompartments (BMCs) encapsulate enzymatic pathways in self-assembling protein shells and can be found across 23 bacterial phyla, expanding their metabolic potential (19–21). The compartment can be beneficial in preventing the exposure of the cell to toxic intermediates, such as aldehydes, by coupling its sink reactions to the recycling of an exclusive cofactor pool (22–25), and in increasing the rate of reaction through the high concentration of enzymes (26). Many examples of BMC-associated metabolic pathways are present in the human gut (27, 28), where they enable the catabolism of alternative substrates. They commonly confer a competitive advantage to enteric pathogens when more typical substrates like glucose are restricted, enabling them to overcome colonization resistance and rapidly populate the microbiota (29).

For example, in Salmonella enterica the capacity for EA catabolism enables the bacteria to sense different gut environments, and therefore enact optimal proliferation strategies, and interact with the gut environment in order to outcompete the other resident bacteria (30). In this way, not only does S. enterica utilize EA to increase the range of nutrients it can survive on and thus gain a competitive advantage but it also uses this pathway to change the gut environment for its advantage and adapt to it itself. In anaerobic conditions, EA does not support significant growth through fermentation; however, S. enterica induces the gut to produce tetrathionate as an alternative electron acceptor, enabling it to utilize EA even in anaerobic conditions (31, 32). It is held that S. enterica uses this niche to rapidly proliferate and cause infection and dysbiosis in the human gut (33, 34). However, this report looks to investigate whether the human gut commensal E. coli Nissle 1917 can also utilize this niche and perhaps alter the gut environment. This could reveal a point of competition that can be exploited for methods of bacterial community control.

This study focuses on the eut BMC, which contains the enzymes necessary for the ethanolamine utilization pathway (21) (Fig. 1A). Ethanolamine is a product of phosphatidylethanolamine degradation, a common lipid component of plant and animal cell membranes (35). Due to the high epithelial turnover in the gut caused by mechanical and chemical stress, EA is a naturally abundant carbon source (36). The eut BMC takes up ethanolamine, where, via the toxic and volatile intermediate acetaldehyde, it is catabolized into ethanol, acetyl-phosphate, and acetyl-CoA, which is then sequestered into the TCA cycle and contributes to cell growth (37). In this way, some bacteria can utilize ethanolamine as an alternative carbon and/or nitrogen source when nutrient-limited (34), for example, enteropathogens including strains of E. coli, S. enterica, and Enterococcus faecalis (33, 34, 37, 38).

Fig 1.

Schematic of the catabolism of ethanolamine by the ethanolamine utilization (eut) pathway within the bacterial microcompartment (BMC) of E. coli Nissle 1917. (A) The production of ethanolamine (EA) from lipid membrane degradation in the intestinal lumen. (B) EA enters the microcompartment and is metabolized to ethanol, acetyl-phosphate, and acetyl-CoA. The enclosed environment of the eut BMC prevents dissipation of the volatile and toxic intermediate acetaldehyde into the cytosol. Metabolic enzymes encapsulated within the BMC include EA ammonia lyase (EAL, EutBC), EAL reactivase (EutA), alcohol dehydrogenase (EutG), aldehyde dehydrogenase (EutE), and phosphotransacetylase (EutD). (C) The eut operon in E. coli Nissle 1917 and S. enterica. Transcription is induced from promoter PI in the presence of both EA and vitamin B₁₂, and promoter PII regulates weak constitutive expression of the transcription factor EutR (30). (D) The eut operon in E. coli BW25113 is interrupted by the CPZ-55 prophage encoding nine genes translated in the opposing direction between EutA and EutB, preventing the functional expression of the eut operon (31, 32).

Although the eut operon is present in 10.7% of bacteria common to the human gut (28), not all of these have the capacity for the expression and formation of the BMC, due to genetic mutations and prophage interruptions which prevent the expression of the eut operon (39). Much of the previous research into BMC-dependent growth has focused on its role in pathogen colonization (19, 33, 40–42), but recent studies have found its presence in gut commensal species too and call for a broader research focus on the field (43, 44), perhaps demonstrating further complexity in gut bacterial community dynamics than previously thought. Many commensal strains of E. coli were thought not to have the capacity for EA degradation, perhaps because of the prophage insertion in the common lab strain K12 from which the E. coli reference genome is based (45). A recent study from de Gouveia et al. provides an exciting new overview of the widespread capacity for the degradation of EA as a nitrogen source among 35 E. coli human gut commensals and as a carbon source among three of these (44). Although EcN has been demonstrated to utilize EA as a nitrogen source (46), it has not yet been demonstrated to utilize EA as a sole carbon or energy source. Additionally, although eut BMC activity has been demonstrated in other E. coli strains, the utility of EcN in microbiome engineering studies makes this research an exciting extension to its capacity as a bioengineering tool (10).

This study demonstrates the utilization of EA as a sole energy source in the human gut commensal EcN through the expression and utilization of the eut BMC. The capacity of EcN for EA degradation as a carbon source was shown through growth assays, and the formation of the BMC structure through transmission electron microscopy (TEM). The expression of key genes encoding the signature enzyme, shell protein, and regulatory protein under conditions of EA utilization was confirmed with qPCR. Additionally, flux balance analysis was used to make predictions about changes to metabolic output when catabolizing EA, which was tested experimentally through ammonia and acetate assays. Finally, the control of eut BMC-dependent growth by varying the induction of the native promoter was shown in EcN and S. enterica. These results contribute to the understanding of the complex microbial interactions within the human gut.

RESULTS

EA utilization in EcN

The capacity of EcN to catabolize EA as the sole carbon source to support cell growth was explored through growth assays under either glucose or EA and vitamin B₁₂ supplementation. E. coli BW25113 (EcBW), which contains the eut operon but with a prophage insertion preventing its functional expression (47) (Fig. 1D), was used as a negative control to demonstrate growth under these conditions with the absence of the eut BMC and its signature enzymes. S. enterica, which has previously been demonstrated to express the genes of the eut operon and utilize the associated BMC under these conditions, was used as a positive control (37).

Increased growth in EcN when supplemented with both EA and vitamin B₁₂ was observed (Fig. 2A through C), indicating catabolism of EA under conditions expected for eut BMC utilization. The absence of increased growth in supplementation with either EA or vitamin B₁₂ confirms the vitamin B₁₂ dependence of the eut BMC in EcN. Low levels of growth are observed even in the absence of a carbon source (Fig. 2), as the pre-culture media (M9 containing glucose) was unable to be washed while maintaining the required biosafety conditions for this species.

Fig 2.

EA-dependent growth in E. coli Nissle 1917. Growth curves of EcBW (A), EcN (B), and S. enterica (C) when grown in M9 supplemented with either glucose, EA, and/or vitamin B₁₂, in triplicate. The mean and standard error are shown with a solid line and a pale ribbon, respectively, and the data were fitted with the Gompertz function (D) shown by the respective dashed lines. The estimated growth rates (E) and carry capacities (F) are shown with standard error bars. Where appropriate, P-values are given, with exact values displayed in Table S1.

Growth tests supplementing the minimal media with EA, vitamin B₁₂, and glucose together were performed to assess the difference in growth cultures compared with glucose alone (Fig. S1). There was minimal observed difference between the conditions in EcBW, suggesting that the presence of EA and vitamin B₁₂ do not have a detrimental or toxic effect on a glucose-metabolizing cell. In EcN, the growth with glucose, EA, and vitamin B₁₂ was higher than when catabolizing EA as the sole carbon source (Fig. 2B), but lower than glucose alone, suggesting that glucose is both a preferable carbon source and less burdensome than microcompartment formation, as supported by the lower growth rates when utilizing EA alone. This also indicates that the expression and utilization of the eut BMC is dependent not only on the presence of both EA and vitamin B₁₂ but also on the absence of glucose.

The Gompertz model (Fig. 2D) was used to estimate the growth rates and carry capacities (maximum cell density) of the cultures. Although the estimated growth rates were comparable or lower (Fig. 2E), the estimated carry capacity of EcN and S. enterica is higher in minimal media supplemented with EA and vitamin B₁₂ compared with either alone or no addition (Fig. 2F). This suggests that although growth is observed under eut BMC utilizing conditions, the catabolism of EA is burdensome to the cell. This would support the expression of the eut BMC only in the absence of glucose in times of nutrient depletion when EA is an adequate alternative, where EA is not a preferred carbon source, but is an adequate alternative in times of nutrient depletion. Although S. enterica has the capacity for de novo vitamin B₁₂ synthesis (48), this is only possible under anaerobic conditions, and hence, exogenous vitamin B₁₂ is supplied under aerobic conditions in this study.

eut BMC expression in EcN

Next, we explored whether BMC structures could be observed directly using TEM. BMC structures were observed in EcN when supplemented with EA and vitamin B₁₂ (Fig. 3A), but not with unsupplemented media, or either EA or vitamin B₁₂ alone. This suggests that both EA and vitamin B₁₂ are required to induce the production of the BMC structure. Additionally, EcBW did not produce BMC structures when supplemented with both EA and vitamin B₁₂, indicating that the prophage insertion is sufficient in disrupting the expression of the eut operon. Structures appear under the same conditions where growth is demonstrated in media supplemented with EA and vitamin B₁₂. The BMC structures are the same approximate size (50–200 nm) and shape as previously reported BMCs in other species and E. coli strains, as they had not been previously visualized in EcN (49–51).

Fig 3.

Microcompartment formation in E. coli Nissle 1917. (A) Transmission electron micrographs of EcBW and EcN transformed with DVA plasmid for antibiotic resistance grown in media supplemented with and without glucose, ethanolamine, and vitamin B₁₂, where suspected BMC structures are indicated by white arrows. (B and C) Change in expression of three key eut operon genes in EcN and EcBW, respectively. qPCR was performed in M9 supplemented with either EA and vitamin B₁₂, or glucose. The qPCR was completed in both biological and technical triplicate, where the compounded standard error bars are shown. Where appropriate, P-values are given, with exact values displayed in Table S2.

To further verify the structures observed in the TEM micrographs, qPCR with three key eut genes was performed under conditions predicted to utilize the eut pathway compared with standard glucose metabolism (Fig. 3B and C). The gene eutS encodes one of the shell proteins, eutB encodes part of the signature enzyme for the pathway, and eutR is a transcription factor for the operon. As the presence of glucose has previously been reported to suppress eut expression (38), and this was observed in the growth assays in EcN (Fig. 2B), this was a suitable control condition. The mRNA corresponding to the three key eut genes had a 12.7-fold to 30.0-fold increase in expression in EcN when supplemented with EA as the sole carbon source and vitamin B₁₂ as the required cofactor for the eut pathway. This result, in combination with the growth assays (Fig. 2) and the TEM micrographs (Fig. 3A), supports the expression and formation of the eut BMC shell, and the expression and utilization of the enzymes of the eut metabolic pathway to utilize EA as a sole carbon source.

Interestingly, the effect of the prophage insertion in the eut operon of EcBW in between the genes encoding eutA and eutB (Fig. 1D) is evident from the qPCR results. It was unknown whether the prophage disrupts the expression of the whole operon or just renders the metabolic pathway non-functional, as it interrupts the genes encoding the signature enzyme, preventing EA catabolism. The qPCR results (Fig. 3C) indicate that there is increased expression of eutR and eutS in EA conditions, but not of eutB. Despite the shell protein eutS being expressed, no BMC structures are observed in identical conditions when imaged with TEM (Fig. 3A).

Ammonium production in EcN

Flux balance analysis (FBA) can be a powerful tool for the exploration and understanding of flux dynamics occurring within a cell, given a set of constraints (Fig. 4A) (52, 53). One use for FBA is the generation of experimental hypotheses that can be readily tested in vitro, where given all the metabolites and reactions in the cell model, the media composition (metabolite input), and active reaction pathways can be altered to explore possible metabolic outputs. In this case, after modifying the models to reflect the experimental findings in this study, we supplied the model with either EA or glucose as a carbon source so we could observe if any metabolites of interest could be produced during the cell’s growth cycle.

Fig 4.

Metabolite flux prediction and experimental verification in E. coli Nissle 1917. (A) Overview of genome-scale metabolic reconstruction and subsequent flux balance analysis. (B) Schematic representing the predicted direction of flux through the eut pathway using FBA with iHM1533_CC. The dotted gray line represents pathways without flux. (C and D) Predicted flux from the FBA when the model is supplied with either EA or glucose as the sole carbon source, respectively. (E) NH${\displaystyle {}_4^{+}}$ and acetate were detected in cell culture after 16/24 h growth in media supplemented with either EA and vitamin B₁₂, or glucose, in triplicate. The results are normalized to the OD₇₀₀ of the solution, and standard error bars are shown. Where appropriate, P-values are given, with exact values displayed in Table S3.

For FBA in EcN, we used a modified version of the iHM1533 model (54) (here called iHM1533_CC), which originally predicted that EcN could not catabolize EA as a carbon source, leading to the removal of the EA transport reaction to render the eut pathway non-functional by van t’ Hof et al. However, since our experimental data demonstrated the ability of EcN to utilize the eut pathway (Fig. 2 and 3), this transport reaction was restored for this study. Similarly, the iJO1366 model (55) was used for EcBW and predicts EA catabolism, as from the genome-scale model, it would appear to express the requisite proteins. However, due to the prophage insertion (Fig. 1D), and confirmed by our growth assays (Fig. 2A), this is not the case, and hence, the model was modified to reflect this, creating iJO1366_CC Tables S3 and S4). The modified models were verified, and the flux pathways were visualized (Fig. 4B).

Within this study, we used FBA models to search for metabolites that would be secreted only if the eut pathway was metabolically active, indicating the impact of eut utilization on the environment and wider microbial community. The ammonium generated by the EutBC catabolism of ethanolamine to acetaldehyde can be utilized by the cell as a source of reduced nitrogen for processes such as amino acid synthesis (38, 56). Many studies report the utilization of the EA as a primary nitrogen source, where this ammonia generated would produce an additional nitrogen supply to the cell (56, 57). As expected, the models all performed similarly to each other in FBA in terms of their metabolite uptake and secretion, with the exception of iJO1366_CC, which cannot utilize EA (Fig. 4C and D; Fig. S2). Additionally, the use of EA as a nitrogen source as well as a carbon source can be demonstrated here. When glucose is supplied as the carbon source, there is a requirement for NH${\displaystyle {\displaystyle {}_4^{+}}}$ as a nitrogen source, but when ethanolamine is supplied as the carbon source, there is a surplus of NH${\displaystyle {\displaystyle {}_4^{+}}}$ which is excreted (Fig. 4C and D). This was expected to be due to the NH${\displaystyle {\displaystyle {}_4^{+}}}$ produced when EutBC converts EA into acetaldehyde, and this was confirmed by following the direction of flux in the FBA (Fig. 4B). The FBA predicts that the products of eut degradation are sequestered directly into the TCA cycle under aerobic conditions and therefore aid cell growth, and the secretion of NH${\displaystyle {\displaystyle {}_4^{+}}}$ (Fig. 4B).

When tested experimentally, it was found that EcN does indeed secrete more net ammonium when utilizing the eut BMC compared with glucose metabolism but that S. enterica does not (Fig. 4E). A baseline signal is observed in all conditions, as EA or glucose is supplied as a carbon source, while the media for the growth cultures contains NH${\displaystyle {\displaystyle {}_4}}$Cl as a nitrogen source. Therefore, the net ammonium production between the conditions is compared. These opposing metabolic outputs have not previously been reported, and although the eut pathway is very similar in both species, this highlights a key difference in the surrounding metabolism and adaptations between the two.

Control of EA-dependent growth

In order to explore the potential applications of BMCs within microbial engineering, the expression of the operon needs to be controllable to optimize the formation of the microcompartment and the utilization of the alternative metabolic pathway it contains. In E. coli, 16 of the 17 genes encoded within the eut operon are under the inducible control of vitamin B₁₂ (Fig. 1D) as previously demonstrated in S. enterica (58). To explore the degree of this control in the eut-dependent growth of both EcN and S. enterica, growth time courses in minimal media supplemented with EA and varying concentrations of vitamin B₁₂ were carried out (Fig. 5). In EcN, concentrations of 0–32.5 nmol results in a proportional positive effect on growth, with no further increase in growth seen above 40 nmol (Fig. 5A and C). In S. enterica, the plataeu is reached at a lower concentration (25–40 nmol) (Fig. 5B and D), but until this point, the growth of the bacteria is proportional to the concentration of exogenous vitamin B₁₂ supplied. This indicates that the expression of the operon and utilization of the resulting BMCs are able to be reasonably controlled by the exogenous application of vitamin B₁₂.

Fig 5.

Control of EA-dependent growth in E. coli nissle 1917 and S. enterica. (A and B) Growth curves of EcN and S. enterica when grown in M9 supplemented with EA and different concentrations of vitamin B₁₂. The mean and standard error are shown in black and grey, respectively. Gompertz model gits are given by the solid lines. (C and D) The estimated carry capacity from the growth curves, with the calculated R${\displaystyle {\displaystyle {}^2}}$ shown above.

DISCUSSION

Here, we demonstrate that EcN shows increased growth when grown in minimal media supplemented with both EA and vitamin B₁₂ in the absence of glucose (Fig. 2B), when compared with EcBW, which cannot utilize the eut BMC, and S. enterica that can. These results, alongside the TEM micrographs that show BMC-like structures comparable with that of the literature, and the qPCR which showed increased expression of mRNA of three key genes of the eut operon, support the hypothesis that EcN can utilize the eut BMC to degrade EA not only as a nitrogen source (46) but also as the sole carbon source. Additionally, the expression of the shell protein EutS in EcBW (Fig. 3C) but lack of BMC shell formation (Fig. 3A) could indicate that the prophage insertion has more widespread effects, for example, on the other shell proteins EutL and EutK further downstream. Additionally, the increased secretion of NH${\displaystyle {\displaystyle {}_4^{+}}}$ during EA metabolism compared with glucose metabolism in EcN, but not in S. enterica, indicates different adaptations and subsequent metabolism of the products of the eut pathway despite the similarities in the eut operon and BMC structure between the two. This complements the recent work of Jallet et al. in the E. coli K-12 derivative W3110, which does not colonize the intestine, where they found that NH${\displaystyle {\displaystyle {}_4^{+}}}$ is secreted when EA is catabolized as a nitrogen and partial carbon source (59). The effects of this difference on supporting the growth of the extended microbial community in the gut is not yet known but is an interesting avenue for investigation.

In the context of eut BMC use in enteropathogens for colonization advantage, this places EcN within a niche that was previously thought to be predominantly occupied by pathogenic species (19, 33). Additionally, the secretion of NH${\displaystyle {\displaystyle {}_4^{+}}}$ from EA degradation could provide insights into how EA degradation interacts with the environment and perhaps indirectly affects the growth of other populations in the gut microbiome. By understanding how these dynamics work in pathogens and their commensal nutrient competitors, we can further investigate the methods of bacterial community control. This work also shows the control of EA-dependent growth using the native promoter and lays the groundwork for the potential engineering of the promoter to be under the control of an inducer that does not naturally occur in the desired environment. In this way, the growth of EcN could be tunable and could potentially be controlled to compete with other eut BMC utilizing bacteria, for example, S. enterica. Here, we demonstrated the ability to control the EA-dependent growth of EcN and S. enterica by inducing the eut operon.

The FBA was used to compare genome-scale models to predict metabolite uptake and secretion to identify any potential cross-feeding or interaction dynamics that could be exploited for synthetic community design. When the models were updated based on the experimental findings of this study, the FBA was utilized to generate hypotheses regarding metabolite secretion into culture media, which were then tested experimentally. Through this method, it was found that EcN secretes increased levels of NH${\displaystyle {\displaystyle {}_4^{+}}}$ when catabolizing EA as opposed to glucose (Fig. 4), which could contribute to changes in the ecological landscape of the bacterial community. However, S. enterica was not found to increase NH${\displaystyle {\displaystyle {}_4^{+}}}$ secretion under EA catabolizing conditions, perhaps indicating their different effects on the wider gut microbial community.

As FBA does not capture biological dynamics, such as nutrient-sensing mechanisms or promoter induction, the results need to be contextualized, given literature and experimental findings to interpret them correctly. Additionally, the directive of FBA is to optimize the flux, given a set objective function, where this may not accurately capture the true biological mechanisms. For example, the direction of flux through the eut pathway predicted by the FBA in this study (Fig. 4B) predicts that there will be no flux through the ethanol production branch of the pathway. However, Huseby et al. found that the cofactors involved in this reaction, namely the NAD${\displaystyle {\displaystyle {}^{+}}}$, NADH, and CoA, are actually recycled within the compartment and act to pull the acetaldehyde through the reactions quicker to convert it into less toxic and volatile forms (25). This is important because the pores of the microcompartment shell are not selective enough to prevent acetaldehyde escape, given sufficient time and concentration. However, although these limitations exist, they are significantly mitigated through critical application and verification in the context of existing literature and new experimental data. In this way, FBA proved a powerful resource in generating hypotheses of metabolite activity to test in the lab and allowed the rapid assessment and selection of metabolic pathways that were most relevant to our investigation of the effects of eut BMC utilization on the gut microbial community.

This study extends the previous work of this field in characterizing the role of BMC-mediated EA utilization as a carbon source in the absence of glucose in a common gut commensal, E. coli strain Nissle 1917. The prevalence of this mechanism in a gut commensal, supporting the work of recent studies (44, 46), highlights that the community dynamics between pathogen and commensal in the gut could be more complex than previously thought and that there could be competition for a niche that was established as a mechanism for pathogen colonization. This work contributes to the potential to leverage BMC-dependent growth of a commensal strain to deplete the nutrient pool of a pathogenic strain and thus outcompete it (27, 60). The difference in ammonium production between EcN and S. enterica perhaps suggests different commensal communities’ support to maintain suppression of a pathogen after it has been outcompeted, indicating more complex dynamics in this niche than previously thought. As outlined in previous reports and reviews (19, 44), understanding not only the mechanisms of the eut BMC but also this interaction between the gut microbiota and potential pathogens provides many opportunities for gut community control for health and microbial engineering applications. This also lays the ground work for the use of the eut BMC as a modular metabolic unit within engineering biology, where it could be used to control community dynamics. Additionally, we have shown that FBA can be a useful tool for hypothesis generation when characterizing pathways experimentally.

MATERIALS AND METHODS

Strains

All strains used within this study are listed in Table 1. Ampicillin-resistant strains were created from electrotransformation of electrocompetent cells of the relevant strain with the DVA plasmid encoding ampicillin resistance obtained from the CIDAR MoClo kit (61).

TABLE 1.

Bacterial strains used in this study

Strain	Description	Source
Escherichia coli BW25113	Assumed BMC-negative lab strain	Keio collection (62)
Escherichia coli Nissle 1917	Assumed BMC-positive commensal	Prof Ian Henderson
Escherichia coli BW25113 - DVA	Ampicillin-resistant lab strain	This study
Escherichia coli Nissle 1917 - DVA	Ampicillin-resistant commensal	This study
Salmonella enterica subsp. enterica serotype Typhimurium	BMC-positive enteropathogen	Leibniz Institute DSMZ

Culture media and growth conditions

Overnight cultures of BW25113 and EcN were grown at 37°C with shaking (200 rpm) from single colonies picked from LB agar plates, in 5 mL of supplemented M9 minimal media (1X M9 Salts (Difco, UK), 0.2% casamino acids (vol/vol), 2 mM MgSO${\displaystyle {\displaystyle {}_4}}$, 0.1 mM CaCl${\displaystyle {\displaystyle {}_2}}$, and 0.4% glucose (vol/vol) where stated), and 100 µg/mL Ampicillin where appropriate.

The ability of the strains to grow with EA as a sole carbon source was tested on minimal media supplemented with 150 nmol 100% cyanocobalamin as vitamin B₁₂ (Merck Life Science Limited, UK) and 30 mmol ${\displaystyle {\displaystyle \ge }}$99.0% EA hydrochloride (Sigma-Aldrich, UK) instead of 0.4% glucose (vol/vol), unless specified otherwise.

Growth assays

The overnight cultures were diluted to OD₇₀₀ 0.05 in M9 media containing Ampicillin, EA, and vitamin B₁₂ as appropriate. Then, 125 ${\displaystyle {\displaystyle \mu }}$L of the diluted culture was added to a 96-well microtitre plate (Greiner, DE) covered with a Breathe-Easy sealing membrane (Diversified Biotech, USA). The plate was inserted into a Tecan Spark plate reader (Tecan, UK) and grown for 16 h with continuous double orbital shaking (2 mm, 150 rpm) at 37°C. Absorbance at 700 nm was measured every 20 min. Calibration and normalization of plate reader data were performed in R using the freely available FlopR package (63). Plots were made in R (64) using ggplot2 (65) and dplyr (66).

Additional growth assays were completed using spectrophotometer (Thermo Scientific Multiskan GO) measurements of the absorbance at 700 nm every 2 h for 6 h. The overnight cultures were diluted to OD₇₀₀ 0.05 in M9 media containing Ampicillin, EA, and vitamin B₁₂ as appropriate. Then, 5 mL of the diluted culture was added to 14 mL culture tubes and incubated at 37°C with shaking (200 rpm) for the duration of the assay.

Fitting data Gompertz model (67, 68) was used in R to fit the cell growth curves (Fig. 2 and 5) for each sample with measured cell density (OD₇₀₀). The growth curves were defined as the logarithm of the cell density (y) plotted against time (t):

$${\displaystyle \begin{array}{r}y=A\cdot \text{exp}\left\{-\text{exp}\left[\frac{\mu \cdot \text{exp}}{A}\right]\lambda -\mathit{\text{t}}+1\right\}\end{array}}$$

where ${\displaystyle {\displaystyle \mu }}$ stands for the growth rate of the bacteria at exponential growth phase; A is the maximum cell density that could be achieved by the cell culture; and ${\displaystyle {\displaystyle \lambda }}$ is the lag time before the bacteria enter exponential growth phase.

Transmission electron microscopy

Culture preparation and cell embedding

After growth (as described above), 10 mL of each culture was pelleted by centrifugation. The pellet was fixed in 2 mL of 2.5% glutaraldehyde diluted in 100 mM sodium cacodylate, pH 7.2 for 2 h with gentle rotation. The bacteria were washed twice with 100 mM sodium cacodylate for 10 min each. The bacteria were then stained for 2 h in 1 mL 1% (wt/vol) osmium tetroxide diluted in 100 mM sodium cacodylate. The pellets were washed four times with 100 mM sodium cacodylate for 5 min each. The pellets were dehydrated through an ethanol (EtOH) gradient as follows: 50% (vol/vol) EtOH for 15 min, 70% EtOH overnight, 90% EtOH for 15 min, and 100% EtOH for 15 min four times. The pellets were then washed twice in propylene oxide for 15 min. The pellets were embedded in 1 mL propylene oxide-low-viscosity (LV) resin (1:1) for 30 min, followed by incubation in fresh LV resin twice for 90 min each. The pellets were resuspended in 1 mL LV resin and transferred to a conical-bottom tube, and then, following centrifugation, they were left to polymerize at 60°C overnight.

Sectioning and visualization of samples

Samples were ultra-thin sectioned on an RMC MT-XL ultra-microtome with a diamond knife (diatome 45°) sections (60–70 nm) were collected on un-coated 300 mesh copper grids. Grids were stained by incubation in 4.5% uranyl acetate in 1% acetic acid solution for 45 min followed by washing in a stream of dH₂0. Grids were then stained with Reynolds lead citrate for 7 min followed by washing in a stream of dH₂0.

Electron microscopy was performed using a JEOL-1230 transmission electron microscope equipped with a Gatan multiscan digital camera operated at an accelerating voltage of 80 kV.

Ammonia production assay

The overnight cultures including either EA (+ B₁₂) or glucose as a carbon source, and NH${\displaystyle {\displaystyle {}_4}}$Cl as a nitrogen source, were diluted to OD${\displaystyle {\displaystyle {}_{7}00}}$ = 0.05 or 1:1000 in accordance to appropriate BioSafety Level conditions. After 16 and 24 h growth for E. coli strains and S. enterica, respectively, to ensure completion of the growth phase, the ammonia detection assay was completed as per the manufacturer’s instructions (Sigma-Aldrich).

qPCR

Total RNA from 16 h growth cultures (as described above) was extracted using the Qiagen RNeasy Plus kit (Qiagen, Chatsworth, CA). RNA was treated with DNaseI (amplification grade; Invitrogen) to remove any remaining trace amounts of DNA. cDNA was generated with 450 ng of RNA by use of the Qiagen Whole Transcriptome Amplification Kit as described by the manufacturer. Primers for qPCR can be found in Table 2. The qPCR was performed with SYBR Green reagent on a Chromo4 PTC-200 Real-Time PCR Detector system (Bio-Rad). PCR conditions were 94°C for 2 min, followed by 40 three-step cycles of 94°C for 15 s, 60°C for 30 s, and 72°C for 30 s. rrsA was used for endogenous reference gene control. Each primer set amplified a single PCR product of predicted size as determined by melt-curve analysis following PCR and by agarose gel electrophoresis and had approximately equal amplification efficiencies when validated with a serial dilution of representative cDNA. Each qPCR was performed in triplicate, and relative quantification was determined according to the ${\displaystyle {\displaystyle \delta \delta }}$C(t) method (69).

TABLE 2.

Primers used in qPCR

Name	Direction	Sequence
EcN_eutS_F	Forward	TTATGACGCTAACTCCCGGAGAA
EcN_eutS_R	Reverse	CCGAGCCATAAATCACCAGCG
EcN_eutB_F	Forward	GCGTGGCGGCAAAGC
EcN_eutB_R	Reverse	CACCGGATTATTGCGGATGT
EcN_eutR_F	Forward	ATCCGGGCAAGTTTCATGGT
EcN_eutR_R	Reverse	CGGAATGCCAAACCAGAACG
EcBW_eutS_F	Forward	GAAAATCGGCGTTCCGGATG
EcBW_eutS_R	Reverse	CCGCTGAATCGGTCGAGAAA
EcBW_eutB_F	Forward	GACGACGTGCAAAGTATCGC
EcBW_eutB_R	Reverse	AGATGGTATCCAACACGCGG
EcBW_eutR_F	Forward	TCCGTGGTCGCAAAGTATGA
EcBW_eutR_R	Reverse	CAGTGACGGCTTCTCGGAAA
rrsA_qPCR1_F	Forward	CTCTTGCCATCGGATGTGCCCA
rrsA_qPCR1_R	Reverse	CCAGTGTGGCTGGTCATCCTCTCA

Flux balance analysis

Flux balance analysis was completed on a MacBook Pro A2338 in both MATLAB using the COBRA Toolbox (70) with Gurobi (71) as the solver and in Python using the COBRApy Toolbox (72) using Jupyter Notebook. The iHM1433 (54), iJO1366 (55) STM_v1.0 (73) models were used for E. coli Nissle 1917, E. coli BW25113 (where the model is based on its ancestor E. coli K-12 MG1655), and S. enterica LT2, respectively.

FBA code repository can be found at https://github.com/ChaniaClare/FBA-eut-BMC/.

Statistical analysis

Where appropriate, statistical analysis was completed in R to determine the significance of observed differences using two-tailed unpaired t-tests with a significance level of 0.05.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00269-24.

Supplemental tables and figures. jb.00269-24-s0001.pdf.

Tables S1 to S7; Fig. S1 and S2.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.