A microbial endocrinology-based simulated small intestinal medium for the evaluation of neurochemical production by gut microbiota

A microbial endocrinology-based simulated small intestinal medium for the evaluation of... ABSTRACT Microbial endocrinology represents the union of microbiology and neurobiology and is concerned with the ability of neurochemicals to serve as an evolutionary-based language between host and microbiota in health and disease. The recognition that microorganisms produce, modify and respond to the same neurochemicals utilized in the various signaling pathways of their mammalian hosts is increasingly being recognized as a mechanism by which the host and microbiota may interact to influence the progression of infectious disease as well as influence behavior through the microbiota-gut-brain axis. While the capacity for bacteria to produce neurochemicals has been recognized for decades, the degree to which this occurs in the environment of the gastrointestinal tract is still poorly understood. By combining techniques used in analytic chemistry, food science and environmental microbiology, a novel culture-based method was developed which generates a medium utilizing animal feed which resembles the contents of the small intestine. The usage of this medium allows for the in vitro growth of bacteria native to the gastrointestinal tract in an environment that is reflective of the small-intestinal host-based milieu. We describe a detailed protocol for the preparation of this medium and the quantification of neurochemicals by microorganisms grown therein. Catecholamines including dopamine and its precursor L-3,4-dihydroxyphenalanine (L-DOPA) as well as biogenic amines including tyramine and its precursor tyrosine, serve as prototypical examples of neurochemicals that are quantifiable with the methods described herein. microbial endocrinology, neurochemicals, gut microbiota, small intestine INTRODUCTION Microbial endocrinology is the study of the ability of microorganisms to produce and recognize neurochemicals that originate either within the microorganisms themselves or within the host they inhabit. Discoveries in the field of microbial endocrinology (Lyte 1993; Lyte 2014; Neuman et al.2015; Lyte 2016) pose questions not easily addressed solely by an examination of an organism's genome. In the natural setting of the mammalian gastrointestinal (GI) tract, questions arise concerning the degree to which microorganisms may produce physiologically relevant concentrations of neurochemicals that can influence the host, especially given that the in situ environment is nutritionally distinct from that encountered using standard microbiological media. Likewise, under what conditions are these chemicals made and who are the key microbial players responsible for this production? To address these questions, an approach which combined augmented culture-based techniques utilizing the same food that an animal would consume coupled with modern tools in biochemical analysis was employed. Simulated small intestinal media (sSIM) is a medium developed to evaluate the potential for bacteria to produce neurochemicals in an in vitro environment that approximates the in situ environment. The development of sSIM is based on the Infogest Consensus Method which is a static digestion technique used to subject food and other bioactives to the same chemical conditions present during the first three phases of digestion: oral passage, gastric digestion and proximal intestinal digestion (Mackie and Rigby 2015). The Infogest Consensus Method was originally developed to closely approximate the physiological electrolyte composition, pH, digestive time and the digestive enzyme content of each phase of digestion. What makes the Infogest Consensus Method unique in terms of its application to the study of microbial endocrinology is that it utilizes the same food that is used for animal feeding. In this regard, the resulting medium, sSIM, differs dramatically from conventional microbiological media which are designed for rapid microbial growth. Since it is composed of actual animal feed, sSIM provides an in vitro environment for the gut microbiota that very closely approximates the in vivo GI environment. By creating a medium which approximates the in vivo environment, microbiota samples from animals can thus be further examined in vitro in a system that utilizes the same food environment upon which their growth and survival in vivo is dependent. This approach enables the dissection of relevant microbial-host based mechanisms, such as those involving microbial endocrinology, to be elucidated. This is relevant, for instance, in experiments evaluating the capacity of the microbiota-gut-brain axis to influence cognition (Bercik 2011; Lyte and Cryan 2014). Included in the feed-based matrix are neurochemical precursors, nutrients and cofactors necessary to support the growth of microbes and neurochemical production. Components include choline, B vitamins, bile, hemin, insoluble fiber, digestive enzymes and mucin which are inherent to the GI contents yet seldom included in conventional media recipes. This complexity directly influences the behavior of microorganisms and their production of neurochemicals. Thus, using the Infogest Consensus Method, an in vitro digestion of food material produces a medium which is functionally similar to the material found in the proximal small intestine. It can be argued that in order to assess the production of neurochemicals from microorganisms in a physiologically relevant way, it is first necessary to grow those organisms in an appropriate environment possessing the resources likely to be available in vivo. Past attempts to obtain media that simulate the GI environment have included autoclaving GI contents such as rumen contents or feces (Lagier et al.2012). A drawback to approaches that rely on in vivo acquired material is that this material has the potential to demonstrate variability dependent on host factors. The protocol described for the production of sSIM is an effort to simulate the natural complexity of food which undergoes the digestive process while minimizing variability. In this study, we outline an approach for both the preparation of an in vivo-like medium and detail processing it for the analysis of neurochemicals including most catecholamines and the aromatic amines tyramine and tyrosine. As a demonstration of this approach, we followed the conversion of tyrosine into tyramine by several strains of Enterococcus faecium. Tyramine is a biogenic amine with cardiovascular, immunological and neurochemical properties. Foods rich in tyramine have the potential to trigger a reaction which includes increased cardiac output, peripheral vasoconstriction, elevated blood glucose and the release of norepinephrine (Bargossi et al.2015). It has also been recognized that tyramine also has the capacity to serve as a neurotransmitter in some species (Alkema et al.2005). Tyramine is produced from tyrosine by the enzyme tyrosine decarboxylase (TDC), an enzyme encoded on the TDC operon along with the genes tyrosine/tyramine permease (tyrP) and Na+/H+ antiporter (nhaC-2). Though the TDC operon is common and highly conserved among Enterococci, a large amount of intraspecies heterogeneity in the production of tyramine has been reported (Bargossi et al.2015). Much of this heterogeneity appears to arise from phenotypic variation in the capacity to respond to a given environment. Thus, examining the behavior of E. faecium in an environment reflective of in vivo conditions provided for a proof-of-concept demonstration of the utility of sSIM. MATERIALS AND METHODS Adaptation of the Infogest method for the preparation of sSIM With modifications as discussed below, the preparation of sSIM follows the Infogest Consensus Method (Mackie and Rigby 2015) in which a food substrate is sequentially digested with an oral, gastric and intestinal phase. We have found that 60 grams of feed is a convenient mass to scale to and our approach (Fig. 1) is based on that mass. This yields a final sSIM volume of 420mL plus the volume of the digested substrate. Figure 1. View largeDownload slide The step-wise process in which sSIM is made. (1) The oral phase is made by mixing feed into simulated salivary fluid (SSF) in a 1:1 ratio. (2) The reaction proceeds for two minutes. (3) The entirety of the contents from the oral phase (≈120mL) is mixed 1:1 with a simulated gastric fluid (SGF) to initialize the gastric phase. (4) The gastric phase progresses over two hours in a stomacher. (5) The entirety of the contents from the gastric phase (≈240mL) is mixed 1:1 with a SIF to initiate the intestinal phase. (6) The reaction proceeds for two hours in a stomacher. Post reaction, the mixture is supplemented with hemin and is then flash frozen in liquid nitrogen as well degassed. Figure 1. View largeDownload slide The step-wise process in which sSIM is made. (1) The oral phase is made by mixing feed into simulated salivary fluid (SSF) in a 1:1 ratio. (2) The reaction proceeds for two minutes. (3) The entirety of the contents from the oral phase (≈120mL) is mixed 1:1 with a simulated gastric fluid (SGF) to initialize the gastric phase. (4) The gastric phase progresses over two hours in a stomacher. (5) The entirety of the contents from the gastric phase (≈240mL) is mixed 1:1 with a SIF to initiate the intestinal phase. (6) The reaction proceeds for two hours in a stomacher. Post reaction, the mixture is supplemented with hemin and is then flash frozen in liquid nitrogen as well degassed. As our laboratory utilizes mice to examine the ability of stress to influence behavior through the microbiota-gut-brain axis, we have employed an autoclavable formulation of the same feed as fed to the mice (Teklad Global Diet #2019S, Envigo, Madison, WI, USA). This is an autoclavable diet which we autoclave in advance of use to ensure that no organisms associated with the feed can contaminate the process. Three stock electrolyte solutions corresponding to the three sequential phases of digestion are made in advance of the digestion: the simulated salivary fluid stock electrolyte (SSFSE), simulated gastric fluid stock electrolyte (SGFSE), and simulated intestinal fluid (SIF) stock electrolyte (SIFSE) (Table 1). In advance of any mixing, all enzyme solutions and stock electrolytes were warmed to 37°C. Table 1. Composition of stock electrolyte solutions. Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 The concentration of electrolytes used to make the stock electrolyte solutions. Stock electrolyte solutions are used to simulate physiological conditions and provide the environment necessary for digestive enzyme function. View Large Table 1. Composition of stock electrolyte solutions. Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 The concentration of electrolytes used to make the stock electrolyte solutions. Stock electrolyte solutions are used to simulate physiological conditions and provide the environment necessary for digestive enzyme function. View Large The oral phase is initiated when ground and autoclaved feed is mixed with a simulated salivary fluid at a ratio of 1 gram of feed per 1mL of salivary fluid (Table 2). The mixing of feed with the salivary fluid occurred in a sterile paddle blender bag (Boekel Scientific, Cat. A1905994, Feasterville, PA, USA). The oral phase lasts two minutes, a time sufficient for wetting all of the feed by kneading the bag by hand. The salivary fluid contains 11.7mL sterile water, 300μL 0.3 Maqueous calcium chloride, 42mL SSFSE and 6 mLalpha amylase solution. Alpha-amylase solution was made by dissolving 30mg alpha-amylase (Millipore-Sigma, St. Louis, MO, USA) into 10mL of SSFSE. Table 2. Preparation of digestive phase solutions. Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg SSFSE: Simulated salivary fluid stock electrolyte, SGFSE: Simulated gastric fluid stock electrolyte, SIFSE: Simulated intestinal fluid stock electrolyte. The combination of stock electrolyte solutions (see Table 1) with the constituents above yield the respective digestive fluids used. Together these fluids contribute 420 mLto the final volume of approximately 450 mL. The exact final volume is dependent on the substrate digested and the solubility of liberated products. View Large Table 2. Preparation of digestive phase solutions. Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg SSFSE: Simulated salivary fluid stock electrolyte, SGFSE: Simulated gastric fluid stock electrolyte, SIFSE: Simulated intestinal fluid stock electrolyte. The combination of stock electrolyte solutions (see Table 1) with the constituents above yield the respective digestive fluids used. Together these fluids contribute 420 mLto the final volume of approximately 450 mL. The exact final volume is dependent on the substrate digested and the solubility of liberated products. View Large The gastric phase began immediately following the salivary phase. To the products of the salivary phase, a gastric solution comprised of 72mL SGFSE, 60μL of aqueous 0.3M calcium chloride, 2.4mL 1M hydrochloric acid, 21.5mL sterile water, 120mg type III mucin (Millipore-Sigma) and 24mL pepsin enzyme solution was added (Table 2).The pepsin enzyme solution was prepared by mixing 2.0g pepsin (Millipore-Sigma, St. Louis, MO, USA) with 25mL SGFSE. Once the salivary products and complete gastric solution were mixed, the entire mixture was paddle blended at 180rpm using a triple mix paddle blender (Boekel Scientific) for two hours at 37°C. The intestinal phase began immediately following the gastric phase. To the gastric phase, an intestinal solution composed of 30mL bile salt solution, 60mL pancreatin enzyme solution, 132mL SIFSE, 480μL of aqueous 0.3M calcium chloride, 1.8mL 1N sodium hydroxide and 15.7mL sterile water was added (Table 2). Bile salt solution was made by dissolving 2.0g bile salts (Millipore-Sigma, St. Louis, MO, USA) into 30mL sterile water. Pancreatin enzyme solution consisted of 6.9g porcine pancreatin (Millipore-Sigma) dissolved into 60mL SIFSE. Once the gastric phase products and complete intestinal solution were mixed, the mixture was paddle blended at 120rpm for two hours at 37°C. Post-phase, 480µL of filter sterilized hemin solution was added. The hemin solution consisted of 0.5% hemin and 128mM potassium phosphate dibasic K2HPO4 dissolved into sterile water. Degassing was carried out by flash freezing in liquid nitrogen and followed by thawing under a vacuum for a total of three cycles. sSIM was stored at −80°C. Sterility was determined by streaking onto TSA with 5% bovine blood (Remel Inc., Lenexa, KS, USA) as well as Miller LB agar (BD, Sparks, MD, USA) and Lactobacilli MRS agar (BD, Sparks) and incubated aerobically and anaerobically. Brain heart infusion broth (BD, Sparks) and MRS broth were also inoculated with 100µL of sSIM. Cultures were examined daily for five days with no growth in any condition. Analysis by ultra high pressure liquid chromatography with electro-chemical detection Cultures grown in sSIM for 24 hours were acidified with the addition of 10μL of 10N hydrochloric acid (HCl) for every 1mL of medium to yield a pH of between 3.0–3.3. Culture medium was centrifuged (3000 × g, 4°C for 15 minutes) to remove insoluble fiber, denatured proteins and other precipitates. The sample supernatant was further purified by passage through a 3kDa molecular weight cut off filter. Samples were stored at −80°C. The separation and quantification of neurochemicals were performed by Ultra High Performance Liquid Chromatography with electrochemical detection (UHPLC-ECD) on a Dionex UHPLC system which consisted of the following components: a Dionex Ultimate 3000 autosampler, a Dionex Ultimate 3000 pump and a Dionex Ultimate 3000 RS electrochemical detector (Thermo Scientific, Sunnyvale, CA, USA). Optimized UHPLC conditions for the separation and quantification of catecholamines and aromatic amines were developed by a review of similar procedures used by other groups (Asano et al.2012; Anantharam et al.2017). Briefly, separation was achieved using buffered 10% acetonitrile mobile phase (MD-TM mobile phase, Thermo Scientific), a 150 mm, 3 μm Hypersil BDS C18 column (Thermo Scientific) and flow rate of 0.6mLmin−1. Prior to injection, samples were held at 4°C by the autosampler. For catecholamines, electrochemical detection was achieved with a 6041RS glassy carbon electrode set to 400mV. For the aromatic amines, tyrosine and tyramine, which lack the catechol moiety, the glassy carbon electrode was set to 700mV. Data capture and analysis was facilitated by the Chromeleon 7.2 software package (Thermo Scientific). Chemical identification was achieved by matching retention times to corresponding analytic standards purchased from Millipore–Sigma. Separation and quantification of acetylcholine was achieved using the buffered phosphate mobile phase described in Thermo Fisher's Guide to acetylcholine analysis, a 150 mm, 3 μm Hypersil BDS C18 column (Thermo Scientific) and a flow rate of 0.7mLmin−1. A post-column derivatization reaction through an acetylcholinesterase-based solid phase reactor (Thermo Scientific) was necessary to create electrochemically active and detectable metabolites. Prior to injection, samples were held at 4°C by the autosampler. Detection was achieved with a platinum electrode set to 275mV. Data capture and analysis was facilitated by Chromeleon 7.2. Chemical identification was achieved by matching retention times to analytic standards purchased from Millpore–Sigma. Optimization of the clean-up and validation of the recovery method Online tools which allow the reliable prediction of the physical properties of organic molecules are available. The company ChemAxon supplies Log D prediction software modeled after the work of Viswanadhan et al. (Viswanadhan et al.1989). This software allows for the in silco preparation of Log D vs. pH plots for many organic compounds including the common biogenic amines. With the data generated using these tools (Fig. 2), we refined our approach to optimize clean-up conditions of sSIM for analysis by UHPLC-ECD. Log D vs. pH plots for several common catecholamines were created and utilized to optimize clean-up conditions of sSIM for analysis by UHPLC-ECD (Fig. 2). From this analysis, it was determined that adjusting sSIM to a pH of 3.0 was ideal. Over the range of pH 2–4, lipophilicity (represented by logD) remains stable for the compounds L-DOPA, dopamine, noradrenaline, adrenaline and serotonin and these compounds predominantly associate with the aqueous layer. Figure 2. View largeDownload slide The tendency for various catecholamines to associate with the aqueous layer over a lipid one. Lower LogD values correspond to a greater tendency for a solute to associate with the aqueous layer. 97% of dissolved solute associates with the aqueous layer if a logD of −1.5 is obtained. Figure 2. View largeDownload slide The tendency for various catecholamines to associate with the aqueous layer over a lipid one. Lower LogD values correspond to a greater tendency for a solute to associate with the aqueous layer. 97% of dissolved solute associates with the aqueous layer if a logD of −1.5 is obtained. As validation of our recovery method for biogenic amines, the recovery of isoproterenol, an isopropyl aminomethyl analog of adrenaline widely used as an internal standard for catecholamines (Nohta et al.1987), was tested by spiking the compound into sSIM, processing the medium by our recovery method and then performing an analysis by UHPLC-ECD. (−)-Isoproterenol hydrochloride (Millipore–Sigma) was spiked into sSIM to yield final concentrations of 100μgmL−1, 10μgmL−1 and 1μgmL−1. As shown in Table 3, there was good correlation between the spiked concentration and measured concentration, with recovery being linear throughout the microgram range (Table 3). The practical limits of quantification were between 0.1–1ngmL−1 (1–10pg on column) for most catecholamines in sSIM when a 10μL injection was used. Practical limits of quantification were established from signals exceeding 10 times the baseline variation. Table 3. High agreement between spiked internal standard and recovered chemical. Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 View Large Table 3. High agreement between spiked internal standard and recovered chemical. Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 View Large Isolation of microbes Clinical bacterial isolates were generously provided by the Iowa State Veterinary Diagnostic Laboratory as well as samples isolated from the environment. Environmental isolation measures included mixing polymicrobial samples (feces, cecal contents, rumen contents, probiotic supplements) with sSIM. Environmental samples were standardized by weight to 100mg and then mixed with 5mL of sSIM media. Mixtures were allowed to incubate either aerobically or anaerobically, at 37°C for 24 hours with continuous stirring. Post growth, media samples were serially diluted into peptone water and then plated onto varying nutrient plates including TSA with 5% ovine blood (Remel, Lenexa) and sSIM plates made by addition of 1.5% agar. Plates were incubated under the same conditions as their respective medium tubes. Distinct colonies were streaked for isolation and identified by a Bruker MALDI Biotyper (Bruker Daltonik Gmb H, Bremen, Germany). MALDI-TOF scores of >2.0 were needed to achieved in order to definitively assign genus and species level identification.’ Bacterial production of tyramine in sSIM Three strains of E. faecium designated ML1085, ML1087 and ML1089 (isolated from canine urine, feline urine and canine bile; respectively) were grown overnight at 37°C anaerobically on TSA with 5% ovine blood. Colonies from each strain were suspended in peptone water and standardized to an OD600 of 0.20. Cultures with an initial population density on the order of 6 to 7 logs per milliliter were prepared by inoculating 25mL of sSIM with 500μL of suspension. These samples were then grown at 37°C anaerobically with agitation. Initially, and for every four hours over a 24 hour period, 1mL of material was removed from each culture for neurochemical analysis by UHPLC-ECD and plate counts. Sampling was done in duplicate. Bacterial production of acetylcholine in sSIM Multiple Lactobacilli isolates, including Lactobacillus plantarum were recovered from the environment and screened for the production of acetylcholine as described above. Isolated organisms were grown overnight, at 37°C anaerobically on TSA blood with 5% ovine blood. Colonies from each strain were suspended in peptone water and standardized to an OD600 of 0.20. Cultures were prepared by inoculating 5mL of sSIM with 100uL of peptone suspension and then incubated at 37°C anaerobically with agitation. After 24 hours, the material was processed for analysis of acetylcholine by UHPLC-ECD, specifically the approach used for acetylcholine as discussed above. Each organism was cultured and analyzed in triplicate. Of the organisms tested, only L. plantarum demonstrated significant acetylcholine production and was used for subsequent examination. RESULTS Reproducibility of sSIM media preparation The preparation of sSIM required multiple sequential digestive processes dependent on the particular added enzymes to mimic the natural process of digestion (Mackie and Rigby 2015). Because of this complexity, one concern we investigated was whether or not such a media could be consistently produced, particularly with respect to constitutive neurochemicals and the associated precursor molecules. Accordingly, five separate batches of media were processed for UHPLC-ECD using the clean-up method described. From each batch, five 5mL samples were taken and analyzed in duplicate. We sought to quantify chemicals amendable to detection by ECD that were also present in the media at sufficient quantities to be relevant. These included L-3,4-dihydroxyphenylalanine (L-DOPA), dopamine and tyrosine (Fig.3) as prototypical neurochemicals for proof-in-concept validation. Figure 3. View largeDownload slide The native abundance of three metabolites found in sSIM across five distinct batch preparations. Figure 3. View largeDownload slide The native abundance of three metabolites found in sSIM across five distinct batch preparations. As shown in Fig. 3, with respect to L-DOPA, batch 1 measured 57.33µM with a standard error of the mean (SEM) of 0.41µM; batch 2: 57.15µM (SEM: 0.42µM); batch 3: 78.07µM (SEM: 0.50µM); batch 4: 64.10µM (SEM: 1.20µM); batch 5: 60.5µM (SEM: 1.03µM). The greatest deviation occurred in batch 3 which possessed a mean deviation of 30.7% relative to the mean of all other samples. All other samples fell within a mean deviation of 8% from the average mean. As shown in Fig. 3, dopamine in batch 1 measured 5.28µM (SEM: 0.05µM); batch 2: 5.38µM (SEM: 0.04µM); batch 3: 6.40µM (SEM: 0.05 µM); batch 4: 10.86µM (SEM 0.24µM); batch 5: 9.66µM (SEM: 0.15µM). Batches four and five appeared distinct from batches 1 to 3, with batch four deviating 62.6% from the mean of all other batches. As shown in Fig. 3, the concentration of tyrosine in batch 1 measured 5.08mM (SEM: 0.09mM); batch 2: 5.11mM (SEM: 0.07mM); batch 3: 4.72mM (SEM: 0.03mM); batch 4: 3.29mM (SEM: 0.03mM); batch 5: 3.50mM (SEM: 0.02mM). Here again, batches 4 and 5 varied from the other three batches with batch 4 deviating 28.8% from the average of all other batches. Batches 1–3 once again remained within 6.0% of the mean of batches 1–3. An evaluation of the microbiota's endocrine capacity is better facilitated when the medium in which organisms are grown is well defined and consistent. Overall, the results shown in Fig.3 demonstrate that consistent production of sSIM and its integral chemical constituents can be achieved. Tyramine production in sSIM by E. faecium As shown in Fig. 4, ML1085 demonstrated the most rapid utilization of tyrosine, reaching complete tyrosine utilization and the highest level of tyramine reported by 4 hours. ML1087 exhibited a classic sigmoidal growth pattern with population density reaching a maximal point by 16 hours, far later than either other strain (Fig. 4). Tyrosine utilization and tyramine accumulation changed marginally over the first twelve hours and then rapidly changed during the period of logarithmic growth. ML1089 reached a peak population of 109 CFUmL−1 by 8 hours (Fig. 4). Unlike ML1085, the strain ML1089 only underwent a marginal loss of population, subsisting at levels around 108 CFUmL−1 for the remainder of the experiment (Fig. 4). Although ML1089 subsisted and reached population levels matching or exceeding the highest observed in ML1085 or ML1087, the utilization of tyrosine to make tyramine was never exhaustive. Maximal conversion occurred by 12 hours and remained constant for the remainder of the experiment. Figure 4. View largeDownload slide The tyrosine utilization (A), tyramine production (B) and growth profiles (C) of three strains of Enterococcus faecium. Figure 4. View largeDownload slide The tyrosine utilization (A), tyramine production (B) and growth profiles (C) of three strains of Enterococcus faecium. Production of acetylcholine in sSIM by L. plantarum L. plantarum has been documented to produce acetylcholine when cultured in a medium which supplies key chemicals including pantothenic acid (Rowatt 1948). As shown in Fig. 5, in sSIM, L. plantarum produced an average of 4.02μgmL−1 of acetylcholine with a standard error of the mean of 0.59μgmL−1. The amount of acetylcholine produced in sSIM is consistent with that reported by Rowatt in which the organism was reported to produce 4.8μgmL−1 acetylcholine in a peptone-based media supplemented with pantothenic acid (Rowatt 1948). As such, the use of sSIM and the biochemical environment it represents supports the production of acetylcholine by L. plantarum. Figure 5. View largeDownload slide Production of acetylcholine in sSIM by L. plantarum. Tracing A is representative of an uninoculated sample which possesses choline but no appreciable acetylcholine. Tracing B is a sample inoculated with L. plantarum. A distinct 150nA signal for acetylcholine is present, which is easily distinguished from baseline noise (typically <1nA). Though not quantified, the apparent decreased signal strength of the choline peak in the inoculated sample suggests utilization of choline by L. plantarum. Figure 5. View largeDownload slide Production of acetylcholine in sSIM by L. plantarum. Tracing A is representative of an uninoculated sample which possesses choline but no appreciable acetylcholine. Tracing B is a sample inoculated with L. plantarum. A distinct 150nA signal for acetylcholine is present, which is easily distinguished from baseline noise (typically <1nA). Though not quantified, the apparent decreased signal strength of the choline peak in the inoculated sample suggests utilization of choline by L. plantarum. DISCUSSION Justification for usage of and alterations to the Infogest Method The process by which sSIM is made is not a radical departure from the Infogest Consensus method (Mackie and Rigby 2015). However, the Infogest Method was not developed specifically to make a medium suitable for microbial growth. Rather, the process was developed to closely simulate digestion to facilitate investigations into how pharmaceuticals and foodstuff constituents behaved when subjected to the digestive process. In order to adapt this approach for the preparation of a microbial medium, several modifications were made (Table 4). Table 4. Comparison between Infogest and sSIM. Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing View Large Table 4. Comparison between Infogest and sSIM. Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing View Large Unlike the original Infogest Consensus Method, sSIM is supplemented with type III gastric mucin and hemin. Mucin is a complex glycosylated protein that can serve as a rich carbon and energy source for intestinal microbiota. It is readily colonized by bacterial biofilms and can influence the composition of the microbiota (Gibson, Cummings and Macfarlane 1988; Derrien et al.2010). Fully processed, the media contains 250μgmL−1 of mucin and in vivo, mucin levels well in excess of this have been reported in animals (Miner-Williams, Moughan and Fuller 2013). Porphyrins are naturally found in the GI tract as various breakdown products of heme and are excreted into GI tract via bile (Fevery 2008). Some examples include bilirubin or urobilinoid pigments which can be found in the stools of adults in these respective levels 5–20mgday−1 and 50–250mgday−1 (Vítek et al.2006). Since porphyrin bound iron is a required microbial factor for the growth of some enteric species such as Prevotella intermedia, 5μgmL−1 (7.7μM) hemin was supplemented into sSIM (Leung and Folk 2002). Another aspect of sSIM preparation is that the gastric and intestinal phases make use of a triple mix paddle blender. This is unlike Infogest Consensus Method which relies on magnetic agitation. The more rigorous blending provided by the paddle blender approximates the mechanical forces of the stomach and allows for a more homogeneous mixture. Autoclaving sSIM can produce a usable media, however much of the inherent digestive enzyme activity will be lost. To avoid some of the enzymatic inactivation while still achieving a sterile medium, a cryogenic approach was used. This involved several cycles of flash freezing in liquid nitrogen followed by thawing under a vacuum of 750 mbar (11psi). As the medium thaws under vacuum, dissolved gasses such as oxygen evacuate the solution. The pressure chosen proved sufficient for the formation of copious quantities of air bubbles within the media, indicative of successful degassing. This has the advantage of making the medium suitable for anaerobic work. The medium is stored at −80°C to limit further digestion by the enzymes present. Medium prepared in this manner shows no growth by direct microscopy or by broth and plate culture after five days of incubation at 37°C under anaerobic and aerobic conditions. Considerations for the clean-up, recovery and quantification strategy Approaches to the processing of complex matrices for the quantification of neurochemicals, such as catecholamines, have existed for decades (Bertler, Carlsson and Rosengren 1958). Our adaptation is similar to many other commonly used approaches (Asano et al.2012). Approaches typically used to process complex matrices, such as tissue, for the quantitation of the catecholamines require that samples are first acidified. Acidification offers multiple advantages including the precipitation of many of the proteins which interfere with ultrafiltration and electrochemical analysis. A similar acidification-based approach was adopted for sSIM, since steps which ensure liberation of catecholamines from proteins and insoluble material are particularly relevant. In a complex mixture such as sSIM, the recovery of any specific metabolite can dramatically vary due to pH. Depending on the starting feed stock used, the media may contain particles which can sequester and remove target molecules during centrifugation or ultra filtration. By optimizing the logD, we have found a pH range that ensures high solubility and recovery of most of the common biogenic amines from this matrix. During the clean-up, molecules of interest will remain in solution and not bound to hydrophobic surfaces which may layer off or exist within the protein precipitates. This translates into greater reproducibility as within the pH range of 3.0–3.3, overall recovery is likely to be consistent between samples even though pH values may differ slightly between samples. The acidic environment also disfavors oxidative processes which convert the catechol group of catecholamines to their respective quinones (Fig.6). Quinones, being already oxidized, are not detectable by ECD under the conditions described herein. Figure 6. View largeDownload slide The equilibrium between the quinone and diol forms of catecholamines favors the diol in a low pH environment with reducing agents present. Figure 6. View largeDownload slide The equilibrium between the quinone and diol forms of catecholamines favors the diol in a low pH environment with reducing agents present. In our evaluation of the consistency of sSIM media, we found some deviation among specific batches in relation to specific metabolites. We believe that some of this deviation is attributable to the age of the starting material, as batches 4 and 5 used slightly older starting material than batches 1–3. The importance of these deviations will vary by application and the amounts present. For example, a batch with a 62.6% deviation in dopamine suggests a large deviation mathematically. However, this deviation results from only a 0.77µgmL−1 mass difference between batches and may not in fact be biologically important. Overall, our data demonstrates that acceptable reproducibility across batches can be achieved using the described methodology as was evidenced by the very low deviation between batches 1–3. Regardless, because of minute variation which can occur with different operators and material heterogeneity, experiments should be designed with controls from the same batch of media. Variable production of tyramine by E. faecium grown in sSIM Preliminary work suggested that E. faecium isolates ML1085, ML1087 and ML1089 had distinct and reproducible chromatographic differences including variation in the capacity to utilize tyrosine and produce tyramine. A review of the literature revealed that among Enterococci, there is often a large amount of phenotypic heterogeneity with regards to tyramine production (Bargossi et al.2015). Our results from this experiment were consistent with these findings and highlight what are likely several strain differences. The isolate ML1085 demonstrated the most rapid utilization of tyrosine, reaching complete tyrosine utilization and the highest overall level of tyramine reported. Interestingly however, this population also precipitously dropped over four logs early after its growth phase. Presumably, the consumption of easily available resources or the production of toxic byproducts of metabolism exceeded a survivable threshold before most members of the population could enter a sustainable dormancy. Generally, tyramine production overlapped with the logarithmic phase of cell growth but leveled off or stopped soon after. This was evident in the behavior of ML1089 which grew to a population of 109 CFUmL−1 by 8 hours but then receded and persisted at a level of 108 CFUmL−1 from 12 hours onward. After 12 hours no additional tyramine production was noted even though tyrosine remained. Quantification of acetylcholine With the exception of acetylcholine, the neurochemicals discussed within this paper are primary amines containing aromatic hydroxyl groups amendable to oxidation and detection on an electrochemical detector. Acetylcholine however is a fully methylated tertiary amine which is not quantifiable by ECD under standardly used conditions. Efforts to analyze acetylcholine by ECD require that the molecule first be converted to electrochemically active molecules, often using enzymes bound in a solid phase reactor (Van Zoonen et al.1987). A reactor based approach has been applied to urine and serum but to our knowledge this approach has never been used on material consistent with that derived from the GI tract. The clean-up process used on sSIM for the other amine neurochemicals also proved sufficient for analysis using the reactor. However, the technology still retains a number of drawbacks including reactor degradation, limited shelf life and expense. Data presented in the present study demonstrates the utility of sSIM as a medium to assess the ability of bacteria to produce acetylcholine. Building towards greater complexity with polymicrobial communities It is acknowledged that the behavior of microbes in a polymicrobial community is radically different from that in a pure culture. Prior to working up to the complexity of polymicrobial communities, we first wanted to demonstrate the utility of sSIM for single organism inoculations. Future work will include experiments built upon polymicrobial communities. By comparing group metabolic changes with the inclusion or absence of specific inoculants, information regarding the individual impact of a given species can be obtained. For example, having identified a metabolic behavior in a species of interest such as E. faecium, it is possible to survey other available isolates of the same species to see if similar behavior is observed. Polymicrobial co-inoculations can be built upon mixing isolates with the potential to influence each other. An example might include mixing an organism that produces a precursor molecule with an organism that uses that precursor. Competition assays can also be envisioned; perhaps if a TDC containing E. faecium is mixed with a tyrosine hydrolase containing organism something interesting may occur. Concluding remarks In contrast to conventional microbiological medium, sSIM more closely approximates the physiological conditions expected in the mammalian proximal small intestine. By deriving this media from the same feed stocks as fed to experimental animals, experiments can be designed in which gut microorganisms are grown using the resources and conditions they would normally be exposed to within the GI environment. With careful design, this approach allows the quantification of neurochemicals produced by the microbiome which under in vivo conditions are often difficult if not impossible to distinguish from host generated neurochemicals. Thus, by employing sSIM, it is possible to dissect and elucidate relevant microbial-host based mechanisms. Questions posed by microbial endocrinology, which seeks to understand the production and utilization of neurochemicals by microorganisms, are not easily answered. The complexity of the contents of the mammalian small intestine has required the development of new culture-based approaches to study the ability of the microbiome to produce neurochemicals that can impact host physiology. The combination of a simulated digestion medium, sSIM, with UHPLC-ECD allows for a broad range of neurochemical experiments to be designed and conducted for a wide range of animal species. ACKNOWLEDGEMENTS The assistance of Dr. Alan Mackie (University of Leeds, UK) in aspects of the Infogest adaptation is gratefully acknowledged. The assistance of Meicen Liu (Iowa State University, USA) in providing experimental support is acknowledged. Additionally, technical consultation with Mr. Reza Hussain (Tucson, AZ) in reviewing pertinent organic chemistry considerations is gratefully acknowledged. This study was supported by the United States Department of Defense, Office of Naval Research award #N00014–15-1–2706 to ML. All authors declare that there are no conflicts of interests. Conflict of interest. None declared. REFERENCES Alkema MJ , Hunter-Ensor M , Ringstad N et al. Tyramine Functions independently of octopamine in the Caenorhabditis elegans nervous system . Neuron . 2005 ; 46 : 247 – 60 . Google Scholar CrossRef Search ADS PubMed Anantharam P , Whitley EM , Mahama B et al. Characterizing a mouse model for evaluation of countermeasures against hydrogen sulfide-induced neurotoxicity and neurological sequelae . Ann N Y Acad Sci . 2017 ; 1400 : 46 – 64 . Google Scholar CrossRef Search ADS PubMed Asano Y , Hiramoto T , Nishino R et al. Critical role of gut microbiota in the production of biologically active, free catecholamines in the gut lumen of mice . Am J Physiol Gastrointest Liver Physiol . 2012 ; 303 : G1288 – 95 . Google Scholar CrossRef Search ADS PubMed Bargossi E , Gardini F , Gatto V et al. The capability of tyramine production and correlation between phenotypic and genetic characteristics of enterococcus faecium and enterococcus faecalis Strains . Front Microbiol . 2015 ; 6 : 1371 . Google Scholar PubMed Bercik P . The microbiota-gut-brain axis: learning from intestinal bacteria ? Gut . 2011 ; 60 : 288 – 9 . Google Scholar CrossRef Search ADS PubMed Bertler Å , Carlsson A , Rosengren E . A method for the fluorimetric determination of adrenaline and noradrenaline in tissues.1 . Acta Physiol Scand . 1958 ; 44 : 273 – 92 . Google Scholar CrossRef Search ADS PubMed Derrien M , van Passel MWJ , van de Bovenkamp JHB et al. Mucin-bacterial interactions in the human oral cavity and digestive tract . Gut Microbes . 2010 ; 1 : 254 – 68 . Google Scholar CrossRef Search ADS PubMed Fevery J . Bilirubin in clinical practice: a review . Liver Int . 2008 ; 28 : 592 – 605 . Google Scholar CrossRef Search ADS PubMed Gibson GR , Cummings JH , Macfarlane GT . Use of a three-stage continuous culture system to study the effect of mucin on dissimilatory sulfate reduction and methanogenesis by mixed populations of human gut bacteria . Appl Environ Microbiol . 1988 ; 54 : 2750 – 5 . Google Scholar PubMed Lagier JC , Armougom F , Million M et al. Microbial culturomics: paradigm shift in the human gut microbiome study . Clin Microbiol Infect . 2012 ; 18 : 1185 – 93 . Google Scholar CrossRef Search ADS PubMed Leung KP , Folk SP . Effects of porphyrins and inorganic iron on the growth of Prevotella intermedia . FEMS Microbiol Lett . 2002 ; 209 : 15 – 21 . Google Scholar CrossRef Search ADS PubMed Lyte M . The role of microbial endocrinology in infectious disease . J Endocrinol . 1993 ; 137 : 343 – 5 . Google Scholar CrossRef Search ADS PubMed Lyte M . Microbial endocrinology and the microbiota-gut-brain axis . Adv Exp Med Biol . 2014 ; 817 : 3 – 24 . Google Scholar CrossRef Search ADS PubMed Lyte M . Microbial endocrinology in the pathogenesis of infectious disease . Microbiol Spectr . 2016 ; 4 : 1 – 2 . Lyte M , Cryan JF . Microbial Endocrinology: The Microbiota-Gut-Brain Axis in Health and Disease. Advances in Experimental Medicine and Biology . New York, NY : Springer , 2014 . Mackie A , Rigby N . Infogest Consensus Method . In: Verhoeckx K , Cotter P , López-Expósito I et al. (eds). The Impact of Food Bioactives on Health: in vitro and ex vivo models . Cham : Springer International Publishing , 2015 ; 13 – 22 . Google Scholar CrossRef Search ADS Miner-Williams WM , Moughan PJ , Fuller MF . Analysis of an ethanol precipitate from ileal digesta: evaluation of a method to determine mucin . Sci Rep . 2013 ; 3 : 3145 . Google Scholar CrossRef Search ADS PubMed Neuman H , Debelius JW , Knight R et al. Microbial endocrinology: the interplay between the microbiota and the endocrine system . FEMS Microbiol Rev . 2015 ; 39 : 509 – 21 . Google Scholar CrossRef Search ADS PubMed Nohta H , Mitsui A , Umegae Y et al. Determination of free and total catecholamines in human urine by HPLC with fluorescence detection . Biomed Chromatogr . 1987 ; 2 : 9 – 12 . Google Scholar CrossRef Search ADS PubMed Rowatt E . The Relation of Pantothenic Acid to Acetylcholine Formation by a Strain of Lactobacillus plantarum . J Gen Microbiol . 1948 ; 2 : 25 – 30 . Google Scholar CrossRef Search ADS Van Zoonen P , Gooijer C , Velthorst NH et al. HPLC detection of choline and acetylcholine in serum and urine by an immobilized enzyme reactor followed by chemiluminescence detection . J Pharm Biomed Anal . 1987 ; 5 : 485 – 92 . Google Scholar CrossRef Search ADS PubMed Viswanadhan VN , Ghose AK , Revankar GR et al. Atomic physicochemical parameters for three dimensional structure directed quantitative structure-activity relationships. 4. Additional parameters for hydrophobic and dispersive interactions and their application for an automated superposition of certain naturally occurring nucleoside antibiotics . J Chem Inf Model . 1989 ; 29 : 163 – 72 . Google Scholar CrossRef Search ADS Vítek L , Majer F , Muchová L et al. Identification of bilirubin reduction products formed by Clostridium perfringens isolated from human neonatal fecal flora . J Chromatogr . 2006 ; 833 : 149 – 57 . © FEMS 2018. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png FEMS Microbiology Ecology Oxford University Press

A microbial endocrinology-based simulated small intestinal medium for the evaluation of neurochemical production by gut microbiota

Loading next page...
 
/lp/ou_press/a-microbial-endocrinology-based-simulated-small-intestinal-medium-for-sTCplk6lCp
Publisher
Blackwell
Copyright
© FEMS 2018.
ISSN
0168-6496
eISSN
1574-6941
D.O.I.
10.1093/femsec/fiy096
Publisher site
See Article on Publisher Site

Abstract

ABSTRACT Microbial endocrinology represents the union of microbiology and neurobiology and is concerned with the ability of neurochemicals to serve as an evolutionary-based language between host and microbiota in health and disease. The recognition that microorganisms produce, modify and respond to the same neurochemicals utilized in the various signaling pathways of their mammalian hosts is increasingly being recognized as a mechanism by which the host and microbiota may interact to influence the progression of infectious disease as well as influence behavior through the microbiota-gut-brain axis. While the capacity for bacteria to produce neurochemicals has been recognized for decades, the degree to which this occurs in the environment of the gastrointestinal tract is still poorly understood. By combining techniques used in analytic chemistry, food science and environmental microbiology, a novel culture-based method was developed which generates a medium utilizing animal feed which resembles the contents of the small intestine. The usage of this medium allows for the in vitro growth of bacteria native to the gastrointestinal tract in an environment that is reflective of the small-intestinal host-based milieu. We describe a detailed protocol for the preparation of this medium and the quantification of neurochemicals by microorganisms grown therein. Catecholamines including dopamine and its precursor L-3,4-dihydroxyphenalanine (L-DOPA) as well as biogenic amines including tyramine and its precursor tyrosine, serve as prototypical examples of neurochemicals that are quantifiable with the methods described herein. microbial endocrinology, neurochemicals, gut microbiota, small intestine INTRODUCTION Microbial endocrinology is the study of the ability of microorganisms to produce and recognize neurochemicals that originate either within the microorganisms themselves or within the host they inhabit. Discoveries in the field of microbial endocrinology (Lyte 1993; Lyte 2014; Neuman et al.2015; Lyte 2016) pose questions not easily addressed solely by an examination of an organism's genome. In the natural setting of the mammalian gastrointestinal (GI) tract, questions arise concerning the degree to which microorganisms may produce physiologically relevant concentrations of neurochemicals that can influence the host, especially given that the in situ environment is nutritionally distinct from that encountered using standard microbiological media. Likewise, under what conditions are these chemicals made and who are the key microbial players responsible for this production? To address these questions, an approach which combined augmented culture-based techniques utilizing the same food that an animal would consume coupled with modern tools in biochemical analysis was employed. Simulated small intestinal media (sSIM) is a medium developed to evaluate the potential for bacteria to produce neurochemicals in an in vitro environment that approximates the in situ environment. The development of sSIM is based on the Infogest Consensus Method which is a static digestion technique used to subject food and other bioactives to the same chemical conditions present during the first three phases of digestion: oral passage, gastric digestion and proximal intestinal digestion (Mackie and Rigby 2015). The Infogest Consensus Method was originally developed to closely approximate the physiological electrolyte composition, pH, digestive time and the digestive enzyme content of each phase of digestion. What makes the Infogest Consensus Method unique in terms of its application to the study of microbial endocrinology is that it utilizes the same food that is used for animal feeding. In this regard, the resulting medium, sSIM, differs dramatically from conventional microbiological media which are designed for rapid microbial growth. Since it is composed of actual animal feed, sSIM provides an in vitro environment for the gut microbiota that very closely approximates the in vivo GI environment. By creating a medium which approximates the in vivo environment, microbiota samples from animals can thus be further examined in vitro in a system that utilizes the same food environment upon which their growth and survival in vivo is dependent. This approach enables the dissection of relevant microbial-host based mechanisms, such as those involving microbial endocrinology, to be elucidated. This is relevant, for instance, in experiments evaluating the capacity of the microbiota-gut-brain axis to influence cognition (Bercik 2011; Lyte and Cryan 2014). Included in the feed-based matrix are neurochemical precursors, nutrients and cofactors necessary to support the growth of microbes and neurochemical production. Components include choline, B vitamins, bile, hemin, insoluble fiber, digestive enzymes and mucin which are inherent to the GI contents yet seldom included in conventional media recipes. This complexity directly influences the behavior of microorganisms and their production of neurochemicals. Thus, using the Infogest Consensus Method, an in vitro digestion of food material produces a medium which is functionally similar to the material found in the proximal small intestine. It can be argued that in order to assess the production of neurochemicals from microorganisms in a physiologically relevant way, it is first necessary to grow those organisms in an appropriate environment possessing the resources likely to be available in vivo. Past attempts to obtain media that simulate the GI environment have included autoclaving GI contents such as rumen contents or feces (Lagier et al.2012). A drawback to approaches that rely on in vivo acquired material is that this material has the potential to demonstrate variability dependent on host factors. The protocol described for the production of sSIM is an effort to simulate the natural complexity of food which undergoes the digestive process while minimizing variability. In this study, we outline an approach for both the preparation of an in vivo-like medium and detail processing it for the analysis of neurochemicals including most catecholamines and the aromatic amines tyramine and tyrosine. As a demonstration of this approach, we followed the conversion of tyrosine into tyramine by several strains of Enterococcus faecium. Tyramine is a biogenic amine with cardiovascular, immunological and neurochemical properties. Foods rich in tyramine have the potential to trigger a reaction which includes increased cardiac output, peripheral vasoconstriction, elevated blood glucose and the release of norepinephrine (Bargossi et al.2015). It has also been recognized that tyramine also has the capacity to serve as a neurotransmitter in some species (Alkema et al.2005). Tyramine is produced from tyrosine by the enzyme tyrosine decarboxylase (TDC), an enzyme encoded on the TDC operon along with the genes tyrosine/tyramine permease (tyrP) and Na+/H+ antiporter (nhaC-2). Though the TDC operon is common and highly conserved among Enterococci, a large amount of intraspecies heterogeneity in the production of tyramine has been reported (Bargossi et al.2015). Much of this heterogeneity appears to arise from phenotypic variation in the capacity to respond to a given environment. Thus, examining the behavior of E. faecium in an environment reflective of in vivo conditions provided for a proof-of-concept demonstration of the utility of sSIM. MATERIALS AND METHODS Adaptation of the Infogest method for the preparation of sSIM With modifications as discussed below, the preparation of sSIM follows the Infogest Consensus Method (Mackie and Rigby 2015) in which a food substrate is sequentially digested with an oral, gastric and intestinal phase. We have found that 60 grams of feed is a convenient mass to scale to and our approach (Fig. 1) is based on that mass. This yields a final sSIM volume of 420mL plus the volume of the digested substrate. Figure 1. View largeDownload slide The step-wise process in which sSIM is made. (1) The oral phase is made by mixing feed into simulated salivary fluid (SSF) in a 1:1 ratio. (2) The reaction proceeds for two minutes. (3) The entirety of the contents from the oral phase (≈120mL) is mixed 1:1 with a simulated gastric fluid (SGF) to initialize the gastric phase. (4) The gastric phase progresses over two hours in a stomacher. (5) The entirety of the contents from the gastric phase (≈240mL) is mixed 1:1 with a SIF to initiate the intestinal phase. (6) The reaction proceeds for two hours in a stomacher. Post reaction, the mixture is supplemented with hemin and is then flash frozen in liquid nitrogen as well degassed. Figure 1. View largeDownload slide The step-wise process in which sSIM is made. (1) The oral phase is made by mixing feed into simulated salivary fluid (SSF) in a 1:1 ratio. (2) The reaction proceeds for two minutes. (3) The entirety of the contents from the oral phase (≈120mL) is mixed 1:1 with a simulated gastric fluid (SGF) to initialize the gastric phase. (4) The gastric phase progresses over two hours in a stomacher. (5) The entirety of the contents from the gastric phase (≈240mL) is mixed 1:1 with a SIF to initiate the intestinal phase. (6) The reaction proceeds for two hours in a stomacher. Post reaction, the mixture is supplemented with hemin and is then flash frozen in liquid nitrogen as well degassed. As our laboratory utilizes mice to examine the ability of stress to influence behavior through the microbiota-gut-brain axis, we have employed an autoclavable formulation of the same feed as fed to the mice (Teklad Global Diet #2019S, Envigo, Madison, WI, USA). This is an autoclavable diet which we autoclave in advance of use to ensure that no organisms associated with the feed can contaminate the process. Three stock electrolyte solutions corresponding to the three sequential phases of digestion are made in advance of the digestion: the simulated salivary fluid stock electrolyte (SSFSE), simulated gastric fluid stock electrolyte (SGFSE), and simulated intestinal fluid (SIF) stock electrolyte (SIFSE) (Table 1). In advance of any mixing, all enzyme solutions and stock electrolytes were warmed to 37°C. Table 1. Composition of stock electrolyte solutions. Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 The concentration of electrolytes used to make the stock electrolyte solutions. Stock electrolyte solutions are used to simulate physiological conditions and provide the environment necessary for digestive enzyme function. View Large Table 1. Composition of stock electrolyte solutions. Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 Simulated salivary fluid stock electrolyte (SSFSE) mM Simulated gastric fluid stock electrolyte (SGFSE) mM Simulated intestinal fluid stock electrolyte (SIFSE) mM Potassium chloride 18.91 8.61 8.50 Potassium phosphate monobasic 4.59 1.12 1.00 Magnesium chloride 0.19 0.32 0.87 Sodium bicarbonate 17.02 31.24 106.30 Ammonium carbonate 0.07 0.60 – Sodium chloride – 5.90 48.77 pH 7.00 3.00 7.00 The concentration of electrolytes used to make the stock electrolyte solutions. Stock electrolyte solutions are used to simulate physiological conditions and provide the environment necessary for digestive enzyme function. View Large The oral phase is initiated when ground and autoclaved feed is mixed with a simulated salivary fluid at a ratio of 1 gram of feed per 1mL of salivary fluid (Table 2). The mixing of feed with the salivary fluid occurred in a sterile paddle blender bag (Boekel Scientific, Cat. A1905994, Feasterville, PA, USA). The oral phase lasts two minutes, a time sufficient for wetting all of the feed by kneading the bag by hand. The salivary fluid contains 11.7mL sterile water, 300μL 0.3 Maqueous calcium chloride, 42mL SSFSE and 6 mLalpha amylase solution. Alpha-amylase solution was made by dissolving 30mg alpha-amylase (Millipore-Sigma, St. Louis, MO, USA) into 10mL of SSFSE. Table 2. Preparation of digestive phase solutions. Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg SSFSE: Simulated salivary fluid stock electrolyte, SGFSE: Simulated gastric fluid stock electrolyte, SIFSE: Simulated intestinal fluid stock electrolyte. The combination of stock electrolyte solutions (see Table 1) with the constituents above yield the respective digestive fluids used. Together these fluids contribute 420 mLto the final volume of approximately 450 mL. The exact final volume is dependent on the substrate digested and the solubility of liberated products. View Large Table 2. Preparation of digestive phase solutions. Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg Simulated salivary fluid (SSF) Simulated gastric fluid (SGF) Simulated intestinal fluid (SIF) SSFSE 42.0 mL SGFSE 72 mL SIFSE 132.0 mL 0.3 M Calcium chloride 300 μL 0.3 M Calcium chloride 60 μL 0.3 M Calcium chloride 480 μL Water 11.7 mL Water 21.5 mL Water 15.7 mL α-amylase in SSFSE (3 mgmL−1) 6.0 mL Type III mucin 120.0 mg Sodium hydroxide (1M) 1.8 mL Pepsin in SGFSE (80 mgmL−1) 24.0 mL Bile salts (66 mgmL−1) 30.0 mL Hydrochloric acid (1M) 2.4 mL Pancreatin in SIFSE (115mgmL−1) 60.0 mL Hemin 2.4 mg SSFSE: Simulated salivary fluid stock electrolyte, SGFSE: Simulated gastric fluid stock electrolyte, SIFSE: Simulated intestinal fluid stock electrolyte. The combination of stock electrolyte solutions (see Table 1) with the constituents above yield the respective digestive fluids used. Together these fluids contribute 420 mLto the final volume of approximately 450 mL. The exact final volume is dependent on the substrate digested and the solubility of liberated products. View Large The gastric phase began immediately following the salivary phase. To the products of the salivary phase, a gastric solution comprised of 72mL SGFSE, 60μL of aqueous 0.3M calcium chloride, 2.4mL 1M hydrochloric acid, 21.5mL sterile water, 120mg type III mucin (Millipore-Sigma) and 24mL pepsin enzyme solution was added (Table 2).The pepsin enzyme solution was prepared by mixing 2.0g pepsin (Millipore-Sigma, St. Louis, MO, USA) with 25mL SGFSE. Once the salivary products and complete gastric solution were mixed, the entire mixture was paddle blended at 180rpm using a triple mix paddle blender (Boekel Scientific) for two hours at 37°C. The intestinal phase began immediately following the gastric phase. To the gastric phase, an intestinal solution composed of 30mL bile salt solution, 60mL pancreatin enzyme solution, 132mL SIFSE, 480μL of aqueous 0.3M calcium chloride, 1.8mL 1N sodium hydroxide and 15.7mL sterile water was added (Table 2). Bile salt solution was made by dissolving 2.0g bile salts (Millipore-Sigma, St. Louis, MO, USA) into 30mL sterile water. Pancreatin enzyme solution consisted of 6.9g porcine pancreatin (Millipore-Sigma) dissolved into 60mL SIFSE. Once the gastric phase products and complete intestinal solution were mixed, the mixture was paddle blended at 120rpm for two hours at 37°C. Post-phase, 480µL of filter sterilized hemin solution was added. The hemin solution consisted of 0.5% hemin and 128mM potassium phosphate dibasic K2HPO4 dissolved into sterile water. Degassing was carried out by flash freezing in liquid nitrogen and followed by thawing under a vacuum for a total of three cycles. sSIM was stored at −80°C. Sterility was determined by streaking onto TSA with 5% bovine blood (Remel Inc., Lenexa, KS, USA) as well as Miller LB agar (BD, Sparks, MD, USA) and Lactobacilli MRS agar (BD, Sparks) and incubated aerobically and anaerobically. Brain heart infusion broth (BD, Sparks) and MRS broth were also inoculated with 100µL of sSIM. Cultures were examined daily for five days with no growth in any condition. Analysis by ultra high pressure liquid chromatography with electro-chemical detection Cultures grown in sSIM for 24 hours were acidified with the addition of 10μL of 10N hydrochloric acid (HCl) for every 1mL of medium to yield a pH of between 3.0–3.3. Culture medium was centrifuged (3000 × g, 4°C for 15 minutes) to remove insoluble fiber, denatured proteins and other precipitates. The sample supernatant was further purified by passage through a 3kDa molecular weight cut off filter. Samples were stored at −80°C. The separation and quantification of neurochemicals were performed by Ultra High Performance Liquid Chromatography with electrochemical detection (UHPLC-ECD) on a Dionex UHPLC system which consisted of the following components: a Dionex Ultimate 3000 autosampler, a Dionex Ultimate 3000 pump and a Dionex Ultimate 3000 RS electrochemical detector (Thermo Scientific, Sunnyvale, CA, USA). Optimized UHPLC conditions for the separation and quantification of catecholamines and aromatic amines were developed by a review of similar procedures used by other groups (Asano et al.2012; Anantharam et al.2017). Briefly, separation was achieved using buffered 10% acetonitrile mobile phase (MD-TM mobile phase, Thermo Scientific), a 150 mm, 3 μm Hypersil BDS C18 column (Thermo Scientific) and flow rate of 0.6mLmin−1. Prior to injection, samples were held at 4°C by the autosampler. For catecholamines, electrochemical detection was achieved with a 6041RS glassy carbon electrode set to 400mV. For the aromatic amines, tyrosine and tyramine, which lack the catechol moiety, the glassy carbon electrode was set to 700mV. Data capture and analysis was facilitated by the Chromeleon 7.2 software package (Thermo Scientific). Chemical identification was achieved by matching retention times to corresponding analytic standards purchased from Millipore–Sigma. Separation and quantification of acetylcholine was achieved using the buffered phosphate mobile phase described in Thermo Fisher's Guide to acetylcholine analysis, a 150 mm, 3 μm Hypersil BDS C18 column (Thermo Scientific) and a flow rate of 0.7mLmin−1. A post-column derivatization reaction through an acetylcholinesterase-based solid phase reactor (Thermo Scientific) was necessary to create electrochemically active and detectable metabolites. Prior to injection, samples were held at 4°C by the autosampler. Detection was achieved with a platinum electrode set to 275mV. Data capture and analysis was facilitated by Chromeleon 7.2. Chemical identification was achieved by matching retention times to analytic standards purchased from Millpore–Sigma. Optimization of the clean-up and validation of the recovery method Online tools which allow the reliable prediction of the physical properties of organic molecules are available. The company ChemAxon supplies Log D prediction software modeled after the work of Viswanadhan et al. (Viswanadhan et al.1989). This software allows for the in silco preparation of Log D vs. pH plots for many organic compounds including the common biogenic amines. With the data generated using these tools (Fig. 2), we refined our approach to optimize clean-up conditions of sSIM for analysis by UHPLC-ECD. Log D vs. pH plots for several common catecholamines were created and utilized to optimize clean-up conditions of sSIM for analysis by UHPLC-ECD (Fig. 2). From this analysis, it was determined that adjusting sSIM to a pH of 3.0 was ideal. Over the range of pH 2–4, lipophilicity (represented by logD) remains stable for the compounds L-DOPA, dopamine, noradrenaline, adrenaline and serotonin and these compounds predominantly associate with the aqueous layer. Figure 2. View largeDownload slide The tendency for various catecholamines to associate with the aqueous layer over a lipid one. Lower LogD values correspond to a greater tendency for a solute to associate with the aqueous layer. 97% of dissolved solute associates with the aqueous layer if a logD of −1.5 is obtained. Figure 2. View largeDownload slide The tendency for various catecholamines to associate with the aqueous layer over a lipid one. Lower LogD values correspond to a greater tendency for a solute to associate with the aqueous layer. 97% of dissolved solute associates with the aqueous layer if a logD of −1.5 is obtained. As validation of our recovery method for biogenic amines, the recovery of isoproterenol, an isopropyl aminomethyl analog of adrenaline widely used as an internal standard for catecholamines (Nohta et al.1987), was tested by spiking the compound into sSIM, processing the medium by our recovery method and then performing an analysis by UHPLC-ECD. (−)-Isoproterenol hydrochloride (Millipore–Sigma) was spiked into sSIM to yield final concentrations of 100μgmL−1, 10μgmL−1 and 1μgmL−1. As shown in Table 3, there was good correlation between the spiked concentration and measured concentration, with recovery being linear throughout the microgram range (Table 3). The practical limits of quantification were between 0.1–1ngmL−1 (1–10pg on column) for most catecholamines in sSIM when a 10μL injection was used. Practical limits of quantification were established from signals exceeding 10 times the baseline variation. Table 3. High agreement between spiked internal standard and recovered chemical. Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 View Large Table 3. High agreement between spiked internal standard and recovered chemical. Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 Mass isoproterenol spiked (μgmL−1) Amount recovered (μgmL−1) % Agreement 100.0 90.6 90.6 10.0 9.2 92.2 1.0 1.2 81.0 View Large Isolation of microbes Clinical bacterial isolates were generously provided by the Iowa State Veterinary Diagnostic Laboratory as well as samples isolated from the environment. Environmental isolation measures included mixing polymicrobial samples (feces, cecal contents, rumen contents, probiotic supplements) with sSIM. Environmental samples were standardized by weight to 100mg and then mixed with 5mL of sSIM media. Mixtures were allowed to incubate either aerobically or anaerobically, at 37°C for 24 hours with continuous stirring. Post growth, media samples were serially diluted into peptone water and then plated onto varying nutrient plates including TSA with 5% ovine blood (Remel, Lenexa) and sSIM plates made by addition of 1.5% agar. Plates were incubated under the same conditions as their respective medium tubes. Distinct colonies were streaked for isolation and identified by a Bruker MALDI Biotyper (Bruker Daltonik Gmb H, Bremen, Germany). MALDI-TOF scores of >2.0 were needed to achieved in order to definitively assign genus and species level identification.’ Bacterial production of tyramine in sSIM Three strains of E. faecium designated ML1085, ML1087 and ML1089 (isolated from canine urine, feline urine and canine bile; respectively) were grown overnight at 37°C anaerobically on TSA with 5% ovine blood. Colonies from each strain were suspended in peptone water and standardized to an OD600 of 0.20. Cultures with an initial population density on the order of 6 to 7 logs per milliliter were prepared by inoculating 25mL of sSIM with 500μL of suspension. These samples were then grown at 37°C anaerobically with agitation. Initially, and for every four hours over a 24 hour period, 1mL of material was removed from each culture for neurochemical analysis by UHPLC-ECD and plate counts. Sampling was done in duplicate. Bacterial production of acetylcholine in sSIM Multiple Lactobacilli isolates, including Lactobacillus plantarum were recovered from the environment and screened for the production of acetylcholine as described above. Isolated organisms were grown overnight, at 37°C anaerobically on TSA blood with 5% ovine blood. Colonies from each strain were suspended in peptone water and standardized to an OD600 of 0.20. Cultures were prepared by inoculating 5mL of sSIM with 100uL of peptone suspension and then incubated at 37°C anaerobically with agitation. After 24 hours, the material was processed for analysis of acetylcholine by UHPLC-ECD, specifically the approach used for acetylcholine as discussed above. Each organism was cultured and analyzed in triplicate. Of the organisms tested, only L. plantarum demonstrated significant acetylcholine production and was used for subsequent examination. RESULTS Reproducibility of sSIM media preparation The preparation of sSIM required multiple sequential digestive processes dependent on the particular added enzymes to mimic the natural process of digestion (Mackie and Rigby 2015). Because of this complexity, one concern we investigated was whether or not such a media could be consistently produced, particularly with respect to constitutive neurochemicals and the associated precursor molecules. Accordingly, five separate batches of media were processed for UHPLC-ECD using the clean-up method described. From each batch, five 5mL samples were taken and analyzed in duplicate. We sought to quantify chemicals amendable to detection by ECD that were also present in the media at sufficient quantities to be relevant. These included L-3,4-dihydroxyphenylalanine (L-DOPA), dopamine and tyrosine (Fig.3) as prototypical neurochemicals for proof-in-concept validation. Figure 3. View largeDownload slide The native abundance of three metabolites found in sSIM across five distinct batch preparations. Figure 3. View largeDownload slide The native abundance of three metabolites found in sSIM across five distinct batch preparations. As shown in Fig. 3, with respect to L-DOPA, batch 1 measured 57.33µM with a standard error of the mean (SEM) of 0.41µM; batch 2: 57.15µM (SEM: 0.42µM); batch 3: 78.07µM (SEM: 0.50µM); batch 4: 64.10µM (SEM: 1.20µM); batch 5: 60.5µM (SEM: 1.03µM). The greatest deviation occurred in batch 3 which possessed a mean deviation of 30.7% relative to the mean of all other samples. All other samples fell within a mean deviation of 8% from the average mean. As shown in Fig. 3, dopamine in batch 1 measured 5.28µM (SEM: 0.05µM); batch 2: 5.38µM (SEM: 0.04µM); batch 3: 6.40µM (SEM: 0.05 µM); batch 4: 10.86µM (SEM 0.24µM); batch 5: 9.66µM (SEM: 0.15µM). Batches four and five appeared distinct from batches 1 to 3, with batch four deviating 62.6% from the mean of all other batches. As shown in Fig. 3, the concentration of tyrosine in batch 1 measured 5.08mM (SEM: 0.09mM); batch 2: 5.11mM (SEM: 0.07mM); batch 3: 4.72mM (SEM: 0.03mM); batch 4: 3.29mM (SEM: 0.03mM); batch 5: 3.50mM (SEM: 0.02mM). Here again, batches 4 and 5 varied from the other three batches with batch 4 deviating 28.8% from the average of all other batches. Batches 1–3 once again remained within 6.0% of the mean of batches 1–3. An evaluation of the microbiota's endocrine capacity is better facilitated when the medium in which organisms are grown is well defined and consistent. Overall, the results shown in Fig.3 demonstrate that consistent production of sSIM and its integral chemical constituents can be achieved. Tyramine production in sSIM by E. faecium As shown in Fig. 4, ML1085 demonstrated the most rapid utilization of tyrosine, reaching complete tyrosine utilization and the highest level of tyramine reported by 4 hours. ML1087 exhibited a classic sigmoidal growth pattern with population density reaching a maximal point by 16 hours, far later than either other strain (Fig. 4). Tyrosine utilization and tyramine accumulation changed marginally over the first twelve hours and then rapidly changed during the period of logarithmic growth. ML1089 reached a peak population of 109 CFUmL−1 by 8 hours (Fig. 4). Unlike ML1085, the strain ML1089 only underwent a marginal loss of population, subsisting at levels around 108 CFUmL−1 for the remainder of the experiment (Fig. 4). Although ML1089 subsisted and reached population levels matching or exceeding the highest observed in ML1085 or ML1087, the utilization of tyrosine to make tyramine was never exhaustive. Maximal conversion occurred by 12 hours and remained constant for the remainder of the experiment. Figure 4. View largeDownload slide The tyrosine utilization (A), tyramine production (B) and growth profiles (C) of three strains of Enterococcus faecium. Figure 4. View largeDownload slide The tyrosine utilization (A), tyramine production (B) and growth profiles (C) of three strains of Enterococcus faecium. Production of acetylcholine in sSIM by L. plantarum L. plantarum has been documented to produce acetylcholine when cultured in a medium which supplies key chemicals including pantothenic acid (Rowatt 1948). As shown in Fig. 5, in sSIM, L. plantarum produced an average of 4.02μgmL−1 of acetylcholine with a standard error of the mean of 0.59μgmL−1. The amount of acetylcholine produced in sSIM is consistent with that reported by Rowatt in which the organism was reported to produce 4.8μgmL−1 acetylcholine in a peptone-based media supplemented with pantothenic acid (Rowatt 1948). As such, the use of sSIM and the biochemical environment it represents supports the production of acetylcholine by L. plantarum. Figure 5. View largeDownload slide Production of acetylcholine in sSIM by L. plantarum. Tracing A is representative of an uninoculated sample which possesses choline but no appreciable acetylcholine. Tracing B is a sample inoculated with L. plantarum. A distinct 150nA signal for acetylcholine is present, which is easily distinguished from baseline noise (typically <1nA). Though not quantified, the apparent decreased signal strength of the choline peak in the inoculated sample suggests utilization of choline by L. plantarum. Figure 5. View largeDownload slide Production of acetylcholine in sSIM by L. plantarum. Tracing A is representative of an uninoculated sample which possesses choline but no appreciable acetylcholine. Tracing B is a sample inoculated with L. plantarum. A distinct 150nA signal for acetylcholine is present, which is easily distinguished from baseline noise (typically <1nA). Though not quantified, the apparent decreased signal strength of the choline peak in the inoculated sample suggests utilization of choline by L. plantarum. DISCUSSION Justification for usage of and alterations to the Infogest Method The process by which sSIM is made is not a radical departure from the Infogest Consensus method (Mackie and Rigby 2015). However, the Infogest Method was not developed specifically to make a medium suitable for microbial growth. Rather, the process was developed to closely simulate digestion to facilitate investigations into how pharmaceuticals and foodstuff constituents behaved when subjected to the digestive process. In order to adapt this approach for the preparation of a microbial medium, several modifications were made (Table 4). Table 4. Comparison between Infogest and sSIM. Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing View Large Table 4. Comparison between Infogest and sSIM. Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing Infogest sSIM Digests wide variety of foodstuffs Digests wide variety of foodstuffs (including animal feeds) Flash frozen once, not validated for sterility Cycles of freeze thaw to achieve sterility Oxygen remains in matrix Cryogenically degassed for anaerobes, little remaining oxygen Mucin omitted Type III mucin included, reflective of actual gut environment Hemin omitted Hemin included, reflective of actual gut environment Agitation in digestion provided by stir bar Material churned during digestion with a stomacher to simulate gastric mixing View Large Unlike the original Infogest Consensus Method, sSIM is supplemented with type III gastric mucin and hemin. Mucin is a complex glycosylated protein that can serve as a rich carbon and energy source for intestinal microbiota. It is readily colonized by bacterial biofilms and can influence the composition of the microbiota (Gibson, Cummings and Macfarlane 1988; Derrien et al.2010). Fully processed, the media contains 250μgmL−1 of mucin and in vivo, mucin levels well in excess of this have been reported in animals (Miner-Williams, Moughan and Fuller 2013). Porphyrins are naturally found in the GI tract as various breakdown products of heme and are excreted into GI tract via bile (Fevery 2008). Some examples include bilirubin or urobilinoid pigments which can be found in the stools of adults in these respective levels 5–20mgday−1 and 50–250mgday−1 (Vítek et al.2006). Since porphyrin bound iron is a required microbial factor for the growth of some enteric species such as Prevotella intermedia, 5μgmL−1 (7.7μM) hemin was supplemented into sSIM (Leung and Folk 2002). Another aspect of sSIM preparation is that the gastric and intestinal phases make use of a triple mix paddle blender. This is unlike Infogest Consensus Method which relies on magnetic agitation. The more rigorous blending provided by the paddle blender approximates the mechanical forces of the stomach and allows for a more homogeneous mixture. Autoclaving sSIM can produce a usable media, however much of the inherent digestive enzyme activity will be lost. To avoid some of the enzymatic inactivation while still achieving a sterile medium, a cryogenic approach was used. This involved several cycles of flash freezing in liquid nitrogen followed by thawing under a vacuum of 750 mbar (11psi). As the medium thaws under vacuum, dissolved gasses such as oxygen evacuate the solution. The pressure chosen proved sufficient for the formation of copious quantities of air bubbles within the media, indicative of successful degassing. This has the advantage of making the medium suitable for anaerobic work. The medium is stored at −80°C to limit further digestion by the enzymes present. Medium prepared in this manner shows no growth by direct microscopy or by broth and plate culture after five days of incubation at 37°C under anaerobic and aerobic conditions. Considerations for the clean-up, recovery and quantification strategy Approaches to the processing of complex matrices for the quantification of neurochemicals, such as catecholamines, have existed for decades (Bertler, Carlsson and Rosengren 1958). Our adaptation is similar to many other commonly used approaches (Asano et al.2012). Approaches typically used to process complex matrices, such as tissue, for the quantitation of the catecholamines require that samples are first acidified. Acidification offers multiple advantages including the precipitation of many of the proteins which interfere with ultrafiltration and electrochemical analysis. A similar acidification-based approach was adopted for sSIM, since steps which ensure liberation of catecholamines from proteins and insoluble material are particularly relevant. In a complex mixture such as sSIM, the recovery of any specific metabolite can dramatically vary due to pH. Depending on the starting feed stock used, the media may contain particles which can sequester and remove target molecules during centrifugation or ultra filtration. By optimizing the logD, we have found a pH range that ensures high solubility and recovery of most of the common biogenic amines from this matrix. During the clean-up, molecules of interest will remain in solution and not bound to hydrophobic surfaces which may layer off or exist within the protein precipitates. This translates into greater reproducibility as within the pH range of 3.0–3.3, overall recovery is likely to be consistent between samples even though pH values may differ slightly between samples. The acidic environment also disfavors oxidative processes which convert the catechol group of catecholamines to their respective quinones (Fig.6). Quinones, being already oxidized, are not detectable by ECD under the conditions described herein. Figure 6. View largeDownload slide The equilibrium between the quinone and diol forms of catecholamines favors the diol in a low pH environment with reducing agents present. Figure 6. View largeDownload slide The equilibrium between the quinone and diol forms of catecholamines favors the diol in a low pH environment with reducing agents present. In our evaluation of the consistency of sSIM media, we found some deviation among specific batches in relation to specific metabolites. We believe that some of this deviation is attributable to the age of the starting material, as batches 4 and 5 used slightly older starting material than batches 1–3. The importance of these deviations will vary by application and the amounts present. For example, a batch with a 62.6% deviation in dopamine suggests a large deviation mathematically. However, this deviation results from only a 0.77µgmL−1 mass difference between batches and may not in fact be biologically important. Overall, our data demonstrates that acceptable reproducibility across batches can be achieved using the described methodology as was evidenced by the very low deviation between batches 1–3. Regardless, because of minute variation which can occur with different operators and material heterogeneity, experiments should be designed with controls from the same batch of media. Variable production of tyramine by E. faecium grown in sSIM Preliminary work suggested that E. faecium isolates ML1085, ML1087 and ML1089 had distinct and reproducible chromatographic differences including variation in the capacity to utilize tyrosine and produce tyramine. A review of the literature revealed that among Enterococci, there is often a large amount of phenotypic heterogeneity with regards to tyramine production (Bargossi et al.2015). Our results from this experiment were consistent with these findings and highlight what are likely several strain differences. The isolate ML1085 demonstrated the most rapid utilization of tyrosine, reaching complete tyrosine utilization and the highest overall level of tyramine reported. Interestingly however, this population also precipitously dropped over four logs early after its growth phase. Presumably, the consumption of easily available resources or the production of toxic byproducts of metabolism exceeded a survivable threshold before most members of the population could enter a sustainable dormancy. Generally, tyramine production overlapped with the logarithmic phase of cell growth but leveled off or stopped soon after. This was evident in the behavior of ML1089 which grew to a population of 109 CFUmL−1 by 8 hours but then receded and persisted at a level of 108 CFUmL−1 from 12 hours onward. After 12 hours no additional tyramine production was noted even though tyrosine remained. Quantification of acetylcholine With the exception of acetylcholine, the neurochemicals discussed within this paper are primary amines containing aromatic hydroxyl groups amendable to oxidation and detection on an electrochemical detector. Acetylcholine however is a fully methylated tertiary amine which is not quantifiable by ECD under standardly used conditions. Efforts to analyze acetylcholine by ECD require that the molecule first be converted to electrochemically active molecules, often using enzymes bound in a solid phase reactor (Van Zoonen et al.1987). A reactor based approach has been applied to urine and serum but to our knowledge this approach has never been used on material consistent with that derived from the GI tract. The clean-up process used on sSIM for the other amine neurochemicals also proved sufficient for analysis using the reactor. However, the technology still retains a number of drawbacks including reactor degradation, limited shelf life and expense. Data presented in the present study demonstrates the utility of sSIM as a medium to assess the ability of bacteria to produce acetylcholine. Building towards greater complexity with polymicrobial communities It is acknowledged that the behavior of microbes in a polymicrobial community is radically different from that in a pure culture. Prior to working up to the complexity of polymicrobial communities, we first wanted to demonstrate the utility of sSIM for single organism inoculations. Future work will include experiments built upon polymicrobial communities. By comparing group metabolic changes with the inclusion or absence of specific inoculants, information regarding the individual impact of a given species can be obtained. For example, having identified a metabolic behavior in a species of interest such as E. faecium, it is possible to survey other available isolates of the same species to see if similar behavior is observed. Polymicrobial co-inoculations can be built upon mixing isolates with the potential to influence each other. An example might include mixing an organism that produces a precursor molecule with an organism that uses that precursor. Competition assays can also be envisioned; perhaps if a TDC containing E. faecium is mixed with a tyrosine hydrolase containing organism something interesting may occur. Concluding remarks In contrast to conventional microbiological medium, sSIM more closely approximates the physiological conditions expected in the mammalian proximal small intestine. By deriving this media from the same feed stocks as fed to experimental animals, experiments can be designed in which gut microorganisms are grown using the resources and conditions they would normally be exposed to within the GI environment. With careful design, this approach allows the quantification of neurochemicals produced by the microbiome which under in vivo conditions are often difficult if not impossible to distinguish from host generated neurochemicals. Thus, by employing sSIM, it is possible to dissect and elucidate relevant microbial-host based mechanisms. Questions posed by microbial endocrinology, which seeks to understand the production and utilization of neurochemicals by microorganisms, are not easily answered. The complexity of the contents of the mammalian small intestine has required the development of new culture-based approaches to study the ability of the microbiome to produce neurochemicals that can impact host physiology. The combination of a simulated digestion medium, sSIM, with UHPLC-ECD allows for a broad range of neurochemical experiments to be designed and conducted for a wide range of animal species. ACKNOWLEDGEMENTS The assistance of Dr. Alan Mackie (University of Leeds, UK) in aspects of the Infogest adaptation is gratefully acknowledged. The assistance of Meicen Liu (Iowa State University, USA) in providing experimental support is acknowledged. Additionally, technical consultation with Mr. Reza Hussain (Tucson, AZ) in reviewing pertinent organic chemistry considerations is gratefully acknowledged. This study was supported by the United States Department of Defense, Office of Naval Research award #N00014–15-1–2706 to ML. All authors declare that there are no conflicts of interests. Conflict of interest. None declared. REFERENCES Alkema MJ , Hunter-Ensor M , Ringstad N et al. Tyramine Functions independently of octopamine in the Caenorhabditis elegans nervous system . Neuron . 2005 ; 46 : 247 – 60 . Google Scholar CrossRef Search ADS PubMed Anantharam P , Whitley EM , Mahama B et al. Characterizing a mouse model for evaluation of countermeasures against hydrogen sulfide-induced neurotoxicity and neurological sequelae . Ann N Y Acad Sci . 2017 ; 1400 : 46 – 64 . Google Scholar CrossRef Search ADS PubMed Asano Y , Hiramoto T , Nishino R et al. Critical role of gut microbiota in the production of biologically active, free catecholamines in the gut lumen of mice . Am J Physiol Gastrointest Liver Physiol . 2012 ; 303 : G1288 – 95 . Google Scholar CrossRef Search ADS PubMed Bargossi E , Gardini F , Gatto V et al. The capability of tyramine production and correlation between phenotypic and genetic characteristics of enterococcus faecium and enterococcus faecalis Strains . Front Microbiol . 2015 ; 6 : 1371 . Google Scholar PubMed Bercik P . The microbiota-gut-brain axis: learning from intestinal bacteria ? Gut . 2011 ; 60 : 288 – 9 . Google Scholar CrossRef Search ADS PubMed Bertler Å , Carlsson A , Rosengren E . A method for the fluorimetric determination of adrenaline and noradrenaline in tissues.1 . Acta Physiol Scand . 1958 ; 44 : 273 – 92 . Google Scholar CrossRef Search ADS PubMed Derrien M , van Passel MWJ , van de Bovenkamp JHB et al. Mucin-bacterial interactions in the human oral cavity and digestive tract . Gut Microbes . 2010 ; 1 : 254 – 68 . Google Scholar CrossRef Search ADS PubMed Fevery J . Bilirubin in clinical practice: a review . Liver Int . 2008 ; 28 : 592 – 605 . Google Scholar CrossRef Search ADS PubMed Gibson GR , Cummings JH , Macfarlane GT . Use of a three-stage continuous culture system to study the effect of mucin on dissimilatory sulfate reduction and methanogenesis by mixed populations of human gut bacteria . Appl Environ Microbiol . 1988 ; 54 : 2750 – 5 . Google Scholar PubMed Lagier JC , Armougom F , Million M et al. Microbial culturomics: paradigm shift in the human gut microbiome study . Clin Microbiol Infect . 2012 ; 18 : 1185 – 93 . Google Scholar CrossRef Search ADS PubMed Leung KP , Folk SP . Effects of porphyrins and inorganic iron on the growth of Prevotella intermedia . FEMS Microbiol Lett . 2002 ; 209 : 15 – 21 . Google Scholar CrossRef Search ADS PubMed Lyte M . The role of microbial endocrinology in infectious disease . J Endocrinol . 1993 ; 137 : 343 – 5 . Google Scholar CrossRef Search ADS PubMed Lyte M . Microbial endocrinology and the microbiota-gut-brain axis . Adv Exp Med Biol . 2014 ; 817 : 3 – 24 . Google Scholar CrossRef Search ADS PubMed Lyte M . Microbial endocrinology in the pathogenesis of infectious disease . Microbiol Spectr . 2016 ; 4 : 1 – 2 . Lyte M , Cryan JF . Microbial Endocrinology: The Microbiota-Gut-Brain Axis in Health and Disease. Advances in Experimental Medicine and Biology . New York, NY : Springer , 2014 . Mackie A , Rigby N . Infogest Consensus Method . In: Verhoeckx K , Cotter P , López-Expósito I et al. (eds). The Impact of Food Bioactives on Health: in vitro and ex vivo models . Cham : Springer International Publishing , 2015 ; 13 – 22 . Google Scholar CrossRef Search ADS Miner-Williams WM , Moughan PJ , Fuller MF . Analysis of an ethanol precipitate from ileal digesta: evaluation of a method to determine mucin . Sci Rep . 2013 ; 3 : 3145 . Google Scholar CrossRef Search ADS PubMed Neuman H , Debelius JW , Knight R et al. Microbial endocrinology: the interplay between the microbiota and the endocrine system . FEMS Microbiol Rev . 2015 ; 39 : 509 – 21 . Google Scholar CrossRef Search ADS PubMed Nohta H , Mitsui A , Umegae Y et al. Determination of free and total catecholamines in human urine by HPLC with fluorescence detection . Biomed Chromatogr . 1987 ; 2 : 9 – 12 . Google Scholar CrossRef Search ADS PubMed Rowatt E . The Relation of Pantothenic Acid to Acetylcholine Formation by a Strain of Lactobacillus plantarum . J Gen Microbiol . 1948 ; 2 : 25 – 30 . Google Scholar CrossRef Search ADS Van Zoonen P , Gooijer C , Velthorst NH et al. HPLC detection of choline and acetylcholine in serum and urine by an immobilized enzyme reactor followed by chemiluminescence detection . J Pharm Biomed Anal . 1987 ; 5 : 485 – 92 . Google Scholar CrossRef Search ADS PubMed Viswanadhan VN , Ghose AK , Revankar GR et al. Atomic physicochemical parameters for three dimensional structure directed quantitative structure-activity relationships. 4. Additional parameters for hydrophobic and dispersive interactions and their application for an automated superposition of certain naturally occurring nucleoside antibiotics . J Chem Inf Model . 1989 ; 29 : 163 – 72 . Google Scholar CrossRef Search ADS Vítek L , Majer F , Muchová L et al. Identification of bilirubin reduction products formed by Clostridium perfringens isolated from human neonatal fecal flora . J Chromatogr . 2006 ; 833 : 149 – 57 . © FEMS 2018. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

FEMS Microbiology EcologyOxford University Press

Published: May 21, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off