Plasmids
For biosensor screening and PAL variant characterization, the gene encoding StlA from Photorhabdus luminescens11 and StlA variants were expressed from anhydrotetracycline (aTc)-inducible plasmids containing low-copy origins of replication (pSC101). These plasmids encoded an autoregulated TetR and a constitutive spectinomycin resistance cassette; 100 µg/mL spectinomycin was provided in all growth steps for StlA plasmid maintenance. Variant library design and build approaches are described below. The sensor system (engineered TCA biosensor and biosensor-regulated expression of gfp) was contained on a single high-copy plasmid with a constitutive ampicillin resistance marker; 100 µg/mL carbenicillin was provided in all growth steps for sensor plasmid maintenance.
Chromosomally engineered strain construction
Escherichia coli Nissle 1917 (EcN) was purchased from the German Collection of Microorganisms and Cell Cultures (DSMZ Braunschweig, E. coli DSM 6601). Phe-degrading clinical candidate strains were constructed through the insertion of genes into the EcN chromosome. SYNB1618 construction has been described previously10. The intergenic loci, malEK, araBC, yicS/nepI, agaI/rsmI, exo/cea, and rhtBC were all identified as suitable insertion sites. These intergenic regions consist of divergent promoters or convergent open reading frames separated by a significant length of DNA such that any inserted sequences would not be expected to lead to polar effects on neighboring genes or promoters. Chromosomal insertions into the EcN genome were performed using the well-characterized lambda Red recombineering approach51. For each insertion, (1) a pKD3 or pKD4-based plasmid containing 1000 bp of 5′ and 3′ EcN genome homology for recombination was built, followed by (2) insertion of the gene/promoter of interest into the plasmid by the isothermal assembly (HiFI DNA Assembly Master Mix, NEB), (3) amplification of the insertion fragment from the plasmid by PCR (including EcN homology regions and a flippase recognition target (frt) site flanked chloramphenicol or kanamycin resistance cassette for subsequent antibiotic cassette removal, Q5 High-Fidelity Master Mix, NEB), (4) recombineering of the insertion fragment by electroporation via pKD46 and subsequent pKD46 removal, and (5) the removal of antibiotic resistance cassettes via pCP20 and subsequent pCP20 removal. All DNA sequences for genomic insertions used in the construction of SYNB1618 and SYNB1934 are available upon request.
For deletion of the dapA gene, two rounds of PCR were performed using nested primers. For the first round of PCR, pKD3 was used as the template DNA. The primers were designed to generate a dsDNA fragment that contained homology adjacent to the dapA gene locus in the EcN chromosome and a chloramphenicol resistance gene flanked by frt sites. The primers used in the second round of PCR used the PCR product of the first round as template DNA. EcN containing pKD46 was transformed with the dapA knockout fragment by electroporation. Colonies were selected on LB agar containing chloramphenicol (30 μg/ml) and diaminopimelate (Sigma, D1377; 100 μg/ml).
All electroporation was performed in an Eppendorf Eporator (1.8-kV pulse, 1-mm gap length electro-cuvettes). Transformed cells were selected as colonies on LB agar (Sigma, L2897) containing carbenicillin at 100 μg/ml of kanamycin at 50 μg/ml where appropriate.
Assay media
Phenylalanine was obtained from VWR (catalog number 97062-556) or Sigma (catalog number P2126), and M9 salts were purchased from Fisher (catalog number DF0485-17). Activated biomass was assayed for whole-cell activity in M9 (6.8 g/L Na2HPO4 + 3 g/L KH2PO4 + 0.5 g/L NaCl + 1 g/L NH4Cl) + 0.5% glucose + 40 mM phenylalanine. An adapted minimal M9 glucose recipe with reduced phenylalanine concentration was used for sensor-based screens: 1% glucose + 13.6 g/L Na2HPO4 + 6 g/L KH2PO4 + 1 g/L NaCl + 2 g/L NH4Cl + 1% LB Lennox + trace metals (0.0002% C6H8FeNO7, 1.8 mg/L ZnSO4·7 H2O, 1.8 mg/L CuSO4·5 H2O, 1.2 mg/L MnSO4·H2O, 1.8 mg/L CoCl2·6 H2O) + antibiotic for plasmid maintenance + 200 ng/mL aTc for stlA variant expression + 10 mM L-phenylalanine as substrate.
Biosensor engineering
A specific TCA-responsive aTF biosensor was engineered using Zymergen’s proprietary sensor development platform. It was enriched from a library of designed sensors by selecting for high per cell GFP in the presence of TCA and low background per cell GFP in the absence of TCA on a FACS. Top biosensor candidates were further characterized for their specificity for TCA over related ligands in the production host EcN, as well as their correlations to produced TCA using StlA variants of known activity levels (Fig. 1b).
Homology model generation
A homology model of StlA was constructed using RosettaCM32. Templates were identified by using HHblits and HHsearch52 searching the Pfam and PDB70 databases. The 10 highest scoring PDB templates were then used by RosettaScripts53 with the beta energy function54 to construct an ensemble of homology structures; 200 decoys were generated, and the lowest energy was selected. The code used in this analysis, along with instructions, can be found in the following GitHub repository: https://github.com/Zymergen/AdolfsenIsabella_NatComm.
Library design
To design libraries for the target enzymes, we used three approaches: phylogenetic analysis, co-evolutionary analysis, and structural analysis.
To perform the phylogenetic analysis, a PSSM was generated using PSI-BLAST33 (typically using the default settings of e value = 0.01 and iterations = 3) and used to evaluate single substitutions of the wild-type StlA sequence. The resulting matrix contains a score for every possible amino acid at every position in the protein, corresponding to the amino acid’s frequency at that position in the set of homologous sequences returned by PSI-BLAST. These potential substitutions were compared to the score for the wild-type residue at that position and, if significantly higher-scoring substitutions (typical cutoff: z score >2) were available from the PSSM, the amino acid at that given position was selected for experimental verification.
To expand the analysis to include residue couplings, a coevolution analysis was performed for StlA using GREMLIN34, and all couplings were analyzed and optimized. For positions where the coupling was not optimal according to the analysis, the most optimal amino acid substitution was selected to be tested experimentally.
At last, to target the active site, positions near the active site were identified using the homology model (model generation described in the previous section). If those positions had been observed in the phylogenetic analysis (PSSM score >0), they were included in the library for experimental validation.
Library build
The library was built into four templates: wild-type StlA and three variants that had demonstrated improved performance in early engineering efforts (H133M_I167K_V207I, S92G_H133F_Y437N, and A93C_H133M_I167K, all of which had 20–40% improvement in activity over wild-type StlA). The StlA variant templates, each harbored within a low-copy aTc-inducible plasmid backbone, were mixed in an equimolar ratio to provide the plasmid template for an oligo-based library build approach. A separate primer pair (Supplementary Data) was designed to introduce each of the target mutations through a one-piece Golden Gate reaction, using inverse PCR55 to amplify the entire plasmid at the target mutation site. For the 130 target mutations, 130 separate PCRs (Q5 High-Fidelity Master Mix, NEB) were performed. A subsample of each reaction was run on an agarose gel to allow for quantification of relative band intensity, which was used to normalize and pool together the 130 PCR products. The pooled products were digested by DpnI and gel purified, then circularized using NEB Golden Gate mix (BsaI-v2) at 37 °C for 1 h followed by 60 °C heat inactivation for 5 min. The pooled circularized library of 130 mutations ×4 templates was transformed into NEB10β electrocompetent cells using NEB’s protocol. Recovered cultures were transferred into LB with antibiotics for overnight selection of transformants, and a subsample was plated on LB agar plates with antibiotics to confirm adequate library coverage. This process was repeated two more times, generating combinatorial libraries of one and two combinations of designed mutations after a second cycle, and one, two, and three combinations of designed mutations after a third. For the second and third cloning cycles, library plasmid was purified from the overnight selection culture and used as the template in the subsequent round of PCR mutagenesis. Following three cycles of cloning, the low-copy aTc-inducible >1-million member library was transformed into a wild-type E. coli Nissle (EcN) host strain containing the sensor system on a high-copy plasmid.
Droplet-based incubation
To reduce crosstalk during TCA production and sensor response, cells were encapsulated in microfluidically generated water-in-oil droplets. Libraries were inoculated into 10 mL LB+ antibiotic (refer to “Plasmids” above) in 25 mm test tubes from a sufficient volume of −80 °C glycerol storage stock to maintain library complexity. The cultures were incubated overnight ~16 h to saturation, at which point 1 mL of the culture was centrifuged at 17,800 × g for 1 min and the pellet resuspended in an equal volume of filtered sensor response medium (see “Assay media” above). The OD600 of the washed cells was measured on a Genesys 20 spectrophotometer with 20× dilution, and the cells were diluted to an OD600 of 0.03 in the same medium. An OD600 of 0.03 results in about one cell being encapsulated per 40 µm droplet on average at time 0.
Droplets were generated in Sphere Fluidics oil (Pico-SurfTM 1, 2% in NovecTM 7500, Lot # 031117-1 – diluted to 1% surfactant in additional NovecTM 7500) at a rate of ~3.4 kHz in a flow-focusing microfluidic device, fabricated in-house as PDMS chips on glass. Droplets were collected in 1.5 mL Eppendorf tubes and incubated at 33 °C in a standing incubator without tumbling for ~16 h. Following incubation, emulsions were broken by adding a sufficient volume of 1H, 1H, 2H, 2H-Perfluoro-1-octanol (Sigma), vortexing for ~15 s, and pulse spinning ~1 s in a VWR Galaxy minister centrifuge to separate the organic and aqueous layers. The aqueous layer contains the cells which have been released from the droplets. The aqueous phase from the broken emulsions was diluted 100× into filtered phosphate-buffered saline (PBS) prior to cell sorting.
Fluorescence-activated cell sorting
Cells were sorted on a Bio-Rad S3E Cell Sorter at a target event rate of 1500 events/s with the sample and collection chambers held at room temperature using ProSortTM Software Version 1.6.0.12. Events were collected in 5 mL polypropylene snap cap tubes (Falcon) containing an initial volume of 0.5 mL LB. Events were gated using three metrics: an elliptical SSC-Area vs FSC-Area gate to isolate EcN cells of a similar size, an FSC-Height vs FSC-Area gate to exclude any doublets, and a GFP-Area (FITC-A) gate to select for cells based on strong sensor response. This gating strategy is exemplified in Supplementary Fig. 6. The sorted volume was diluted into sufficient LB Lennox + antibiotic for overnight recovery at 33 °C 220 rpm. To avoid concerns about genetic drift from multiple outgrowths, the StlA plasmids from recovered populations were purified and transformed into fresh EcN containing the sensor plasmid prior to performing subsequent sorts.
Plate-based activated biomass assay for TCA production from strains
150 µL LB Lennox + antibiotic in standard 96-well plates (VWR catalog number 82050-772) was inoculated from colonies of E. coli Nissle containing StlA variant expression plasmids and sensor plasmid. These pre-culture plates were incubated overnight at 33 °C 750 rpm in an Incu-Mixer MP incubator (Benchmark Scientific), covered in VWR porous film (VWR catalog number 60941-086).
The overnight pre-cultures were inoculated 1:100 into 1 mL LB Lennox + antibiotic in deep-well 96-well plates (VWR catalog number 89047-264). The plates were covered in VWR porous film and incubated in an Incu-Mixer at 33 °C 1500 rpm for 2 h into exponential phase. After 2 h, all wells were induced with 200 ng/mL aTc (final concentration, VWR catalog number 200002-828) and placed back in the Incu-Mixer for 4 h at 33 °C 1500 rpm, again covered in VWR porous film. All remaining steps were performed at 4 °C/on ice. After 4 h, the plates were centrifuged at 3214 × g for 10 min. The supernatant was decanted, and the pellets were washed in 250 µL cold PBS. The plate was again centrifuged at 3214 × g for 10 min and decanted. Pellets were resuspended in 30 µL cold PBS (~40 µL final with cell pellet). To this, 40 µL cold 50% glycerol was added, leading to an OD600~25 preparation. The cell preparations were transferred to chilled 96-well PCR plates, covered in foil, and stored at −80 °C until the assay.
For the assay, activated biomass preparation 96-well PCR plate aliquots were thawed at 4 °C, and 4 µL of each OD600 = 25 activated biomass was aliquoted across a 96-well PCR plate and stored on ice until assay initiation. To initiate the assay, 96 µL of pre-warmed (37 °C) assay buffer (M9 0.5% glucose 40 mM phenylalanine) was added to the 4 µL activated biomass, mixed, covered in foil, and placed in the 37 °C static incubator. After 4 h, the cells were pelleted by centrifugation at 3214 × g 4 °C for 10 min. After pelleting, TCA was quantified from supernatants in a Synergy H1 microplate reader at absorbance 290 nm in UV-Star microplates (Greiner catalog number 655801) following dilution in water to within the linear range of the instrument. The final volume in the plate was 150 µL. A standard curve was used to translate A290 measurements to TCA (mM) using trans-cinnamic acid purchased from Sigma (catalog number C80857). TCA concentrations produced from select variants were confirmed by high-performance liquid chromatography. For fold over wild-type (FOWT) normalizations, variant TCA levels were divided by the averaged wild-type TCA values from the same assay.
Activated biomass assay at various pH or after exposure to low pH
During further characterization of top PAL variants, activated biomass was more meticulously normalized to OD600 = 1 during activity assays. Culture tubes with 10 mL LB Lennox + antibiotic were inoculated 1:100 from 3 mL LB Lennox + antibiotic 33 °C 220 rpm overnight pre-cultures. These were incubated for 2 h at 33 °C 220 rpm, at which point they were induced with 200 ng/mL aTc and placed back in the 33 °C 220 rpm incubator. All remaining preparation steps were done on ice/at 4 °C. After 4 h of induction, the cultures were transferred to 15 mL conical tubes (Falcon) and centrifuged at 3214 × g for 10 min, resuspended in 1 mL cold PBS, and transferred to 1.5 mL Eppendorf tubes. They were then washed one more time in 1 mL cold PBS (17,800 × g 1 min in a microcentrifuge), and resuspended a final time in 300 µL PBS. OD600 measurements were performed in cuvettes, and the cells were normalized to OD600 = 50 in 300 µL cold PBS. To the 300 µL OD600 = 50 in PBS samples, 300 µL cold 50% glycerol was added for a final OD600 = 25. Aliquots were distributed across PCR tubes for each strain and stored at −80 °C until the activated biomass assay could be performed.
For the assays performed at various pH (Fig. 4b), activated biomass was thawed at 4 °C and diluted to OD600 = 1 in M9 0.5% glucose 40 mM Phe titrated to pH 5, 6, 7, or 8 in 96-well PCR plates. The plate was covered in foil and incubated at 37 °C for 4 h. The PCR plates were then centrifuged at 3214 × g 4 °C for 10 min and the supernatants were quantified by absorbance at 290 nm as described in “Plate-based activated biomass assay for TCA production from strains” above.
For assays performed after recovery from low pH (Fig. 4c), the activated biomass aliquots were thawed and diluted to OD600 = 1 in M9 0.5% glucose at pH 5 (no Phe) in 96-well PCR plates. The plates were incubated at 37 °C for 1 h without shaking, centrifuged at 3214 × g 4 °C for 10 min and washed in PBS, and then were assayed for activated biomass activity at neutral pH as described in “Plate-based activated biomass assay for TCA production from strains” above. To control for cell loss during wash steps to remove pH 5 medium, fresh activated biomass was added to 96-well PCR plates, washed alongside the samples incubated at pH 5, and then assayed for activated biomass activity; these samples are labeled “control” in Fig. 4c. For FOWT normalizations, variant TCA levels were divided by the averaged wild-type TCA values from the same assay.
Comparison of whole cells vs lysates for examination of PAL expression, abundance, and activity
A series of PAL-expressing strains were constructed in EcN, which included strains (2) with a single chromosomal insertion of stlA at separate locations, a strain containing both chromosomal stlA gene copies in the same background, a strain with a low-copy stlA-expressing plasmid (pSC101 origin), and a strain with a high-copy stlA-expressing plasmid (pUC origin). In each of these strains, the same stlA coding sequence, ribosome binding site, and aTc-inducible promoter were used so that the only difference between strains was the locations of chromosomal insertion and/or stlA gene copy number. For strain growth and PAL induction, 50 mL baffled flasks with 10 mL of LB Lennox broth containing appropriate antibiotics were inoculated 1:100 from overnight cultures. Flasks were grown shaking at 37 °C 250 rpm in an Eppendorf orbital shaker for 2 h to bring cultures into exponential phase, at which point aTc was added at 200 ng/mL aTc (final concentration, VWR catalog number 200002-828). Cells were allowed to grow induced for an additional 4 h before centrifugation for 10 min at 5000 × g and resuspension of pellets in an equal volume of ice-cold PBS.
To measure PAL activity of whole cells, strains were resuspended to an OD600 of 0.1 in 1 mL Phe assay buffer (M9 0.5% glucose-containing 40 mM Phe) in microfuge tubes and placed in a 37 °C heat block. Samples were removed every 30 min for 2 h and TCA concentration was determined by OD290 measurement as described above. Rates of TCA production were extrapolated from the concentration of TCA measured in the supernatant over time.
To measure PAL activity from lysates, a 6 mL volume of resuspended cells were passed through a Microfluidics™ LV1 Low volume Microfludizer® homogenizer at a pressure of 18,000 psi. Each strain was passed through the homogenizer three times to ensure complete lysis. The resulting lysates were spun down at 12,000 × g for 20 min to pellet insoluble material. Total soluble protein concentration for each cleared lysate was determined via BCA assay (Pierce, catalog number 23227). Rates of TCA production were determined similar to the method used for whole cells described above, with the exception that 30 mg of soluble protein was used to provide PAL for the assay rather than resuspension of cells.
Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) gels were run using whole cells of WT EcN and the engineered strain series described above. Cells were grown and induced as described. Whole cells of each induced strain were resuspended to an OD600 of 0.5 and boiled in 1× lithium dodecyl sulfate sample loading buffer for 15 min (Invitrogen, catalog number NP0007). A 20 μL volume of each sample was loaded onto 4–12% bis-tris stacking SDS–PAGE gel (Invitrogen, catalog number NP0322). For standards, iBright Prestained Protein Ladder (Invitrogen, catalog number LC5615) was included to estimate protein size. Gels were run in MES SDS Buffer (Invitrogen, catalog number NP0002) at 200 V for 30 min and stained with SimplyBlue SafeStain (Invitrogen, catalog number LC6065) according to the manufacturer’s instructions.
Michaelis–Menten kinetic parameters determined from cell lysate
Activated biomass was prepared from 10 mL culture tubes as described in the section above, and then lysed for kinetic parameter determination. To prepare lysate, thawed biomass samples were diluted and sonicated using a Branson Digital Sonifier with microtip, then the soluble fraction of the lysate samples were used for the kinetic assay. Total protein in the lysate samples was measured via Bradford Assay, and all samples were normalized to 10 µg total protein loading per well for the kinetic assay. The lysate samples were incubated in M9 0.5% glucose with Phe concentrations ranging from 40 mM Phe down to 39 μM with twofold dilutions (assay buffer without Phe was also included as a control). The kinetic assay was performed in UV-star 96-well microplates (Greiner) with TCA quantified by A290 measurements every minute using a BioTek Synergy H1 microplate reader set to 37 °C static incubation. Michaelis–Menten model fitting was performed using a nonlinear regression (nls function in R with formula V = (Vmax * [S])/(KM + [S])) for the rate data from three batch replicates. Example model fits can be found in Supplementary Fig. 7. The rate V used in the nonlinear regression was calculated from the first hour of activity for each Phe concentration tested, where activity remained linear.
Bioreactor growth of integrated strains, lyophilization, and determination of viability
All strains were grown in fermentation media, which was prepared as followed: yeast extract (40 g/L), K2HPO4 (5 g/L), KH2PO4 (3.5 g/L), (NH4)2HPO4 (3.5 g/L), MgSO4*7H2O (0.5 g/L), FeCl3 (1.6 mg/L), CoCl2*6H2O (0.2 mg/mL), CuCl2 (0.1 mg/L), ZnCl2 (0.2 mg/L), NaMoO4 (0.2 mg/L), H3BO3 (0.05 mg/L), Glycerol (25 g/L), Antifoam 204 (125 μL/L). Production of strains began by thawing a vial from a cell bank and culturing 1 mL of the thawed vial in 50 mL fermentation media supplemented with diaminopimelate (300 µg/mL) in a 500 mL Ultra-Yield™ flask (Thomson Instrument Company). Cells were grown with shaking at 375 rpm until an OD600 of 10–15 was reached, at which point the cultures were used to inoculate 24.5 L of media in a HyPerforma™ Thermo Scientific 30 L Single-Use Fermenter at a starting OD600 of 0.000016. The fermenter was controlled at 37 °C, 60% dissolved oxygen (DO) concentration, and pH 7 using ammonium hydroxide. For SYNB1618, at an OD600 of 1.5, cells were activated by the creation of a low oxygen environment (10% DO), and the addition of Isopropyl β-d-1-thiogalactopyranoside (IPTG, 1 mM). A nutrient feed was also started at this time (15 mL/L-h of 271.75 g/L yeast extract, 86.27 g/L glycerol) and continued until the end of fermentation. After 3.5 h from the addition of IPTG, the nutrient feed rate was doubled to 30 mL/L-h. For SYNB1934, at an OD600 of ~1.5, cells were activated by the addition of IPTG (1 mM), and DO was maintained at 30%. A nutrient feed was also started at this time (11.25 mL/L-h of 271.75 g/L yeast extract, 86.27 g/L glucose) and continued for 3.5 h. After 3.5 h, the nutrient feed rate was doubled to 22.5 mL/L-h for a period of 2.75 h. For both strains, l-arabinose was added to the fermentation (10 mM final concentration) for the final hour of fermentation. Control strain SYN094, which contains no Phe degradation components, was grown in a 5-L bioreactor in fermentation media supplying a steady DO content of 30% until the stationary phase was reached. Strains were harvested by tangential flow filtration (TFF).
At the end of TFF, the supernatant was discarded and cells were resuspended to a final concentration of 150 OD600 in lyoprotectant buffer (10% wt/vol Trehalose, 50 mM Tris, pH 7.5). The formulated cell suspension was used to fill 2 mL glass amber vials and lyophilized to a final water content of <5%. Lyophilized material was stored at 4 °C. Dried lyophilized powder was reconstituted with PBS (Quality Biological, 114-056-101) to match the pre-lyophilization volume. This reconstituted cell suspension was used for measuring activity, viability.
To determine bacterial cell viability, resuspended cells were diluted and stained with SYTOX Green nucleic acid stain (Life Technologies). Live and dead stained cells were counted directly on a Nexcelom Bioscience Cellometer X2 image cytometer per manufacturer’s protocol.
IVS of the gut environment assays
The in vitro gastric simulation model was designed to simulate key aspects of oral administration in humans, including gastric oxygen concentration, pepsin secretion, and gastric pH. The IVS assay is comprised of incubations in 96-well microtiter plate format designed to simulate human stomach conditions56. In brief, lyophilized cells were resuspended in PBS at room temperature. Bacterial cell concentrations were determined by counts of viable and/or total cells. Aliquots of cells were resuspended in 0.077 M sodium bicarbonate buffer at 5.0 × 109 cells per mL. This solution was then mixed with equal parts of simulated gastric fluid56 containing 20 mM Phe, and incubated for 2 h at 37 °C with shaking in a polycarbonate in vitro hypoxic chamber (Coy Lab Products) calibrated to 2% oxygen. The resulting SYNB1618 cell density in SGF was 2.5 × 109 cells/mL. To determine PAL activity, SGF aliquots were collected periodically and centrifuged at 5000 × g for 5 mins using a tabletop centrifuge, followed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) quantification of metabolites, including Phe and trans-cinnamate (PBS).
Studies of strain activity in NHPs
NHP studies were performed at Charles River Labs (Shrewsbury, MA) in compliance with all applicable sections of the Final Rules of the Animal Welfare Act regulations (Code of Federal Regulations, Title 9), the Public Health Service Policy on Humane Care and Use of Laboratory Animals from the Office of Laboratory Animal Welfare, and the Guide for the Care and Use of Laboratory Animals from the National Research Council. Twelve male cynomolgus monkeys aged 2–5 years were used (2.5–4 kg), and were maintained on International Certified Primate Chow (PMI nutrition, 5048). Standard operating procedures related to NHP studies have been reviewed and approved by Charles River Laboratories’ Institutional Animal Care and Use Committee. All animals in the cohort were in a good health at the beginning of the study and washed out for at least 7 days between studies. Three single-dose studies were performed to compare SYNB1934 with SYN1618. On each of the experimental days, six NHP subjects were dosed with SYNB1618 or SYNB1934 and the data presented above are the combined results of three experiments.
The animals were fasted overnight the day before dosing (Day 1) and throughout the procedures on Day 1 without exceeding a maximum of 24 h. On the morning of Day 1, a baseline blood sample was drawn from each monkey by venipuncture. The animals were temporarily restrained for dose administration, but not sedated. Lyophilized bacteria were resuspended and administered orally at 1011 Live Cell doses to each animal, together with 5 mL of 0.36 M sodium bicarbonate, 7.7 mL of 20 mg/mL l-phenyl-D5-alanine (C/D/N Isotopes Inc.), and 6.1 mL of 500 g/L Peptone peptic digest (Sigma). Plasma and cumulative urine were collected for further analysis. Following dosing, animals were then returned to their cages, and a clean urine collection pan was placed at the bottom of each cage. In all, 6 h after initial dosing, the cumulative volume of urine was measured and recorded, and urine samples were stored at −80˚C. Blood samples were collected at 0.5, 1, 2, 4, and 6 h post dose, and plasma was prepared and frozen at −80 °C. Concentrations of metabolites were quantified using a derivatization assay with LC-MS/MS detection.
LC-MS/MS performance
Quantification of TCA, d5-TCA, HA, and d5-HA were performed using targeted multiple reaction monitoring (MRM) mode in a Thermo TSQ Quantum Max triple quadrupole Liquid Chromatography-tandem Mass Spectrometry (LC-MS/MS) system. The standards used were trans-Cinnamic acid (Acros, 158570050), trans-Cinnamic acid-d5 (CDN Isotopes, D-5284), Hippuric acid (Sigma, 112003), and Hippuric acid-d5 (CDN Isotopes, D-5588).
Standards were prepared in water with the following concentrations: 0.032, 0.16, 0.8, 4, 20, 100, and 250 μg/mL. Samples were stored at –80 °C prior to analysis. Urine samples were diluted 40-fold in water prior to sample processing. Creatinine (Sigma, 60275) was added to the standard mixture when analyzing urinary HA and d5-HA (0.32, 1.6, 8, 40, 200, 1000, and 2500 μg/mL). In a 96-well plate, 10 µL of the standards and samples were transferred, followed by the addition of 90 µL derivatization solution (50 mM of 2-hydrazinoquinoline, dipyridyl disulfide, and triphenylphosphine in acetonitrile with 1 μg/mL of isotopically labeled internal standard 13C9-15N-Phe (Cambridge Isotopes, CNLM-575-H-PK) and d5-creatinine (CDN Isotopes, D-7707). The plate was heat-sealed with a ThermASeal foil, mixed, and incubated at 60 °C for 1 h to derivatize the samples. The derivatized samples were then centrifuged at 3200 × g for 5 m. To another plate, 20 µL of the derivatized samples was transferred and further diluted with 180 μL of 0.1% formic acid in water/acetonitrile (140:40). The injection volume used was 10 µL, and the run time was 4.25 m at a flow rate of 0.5 mL/minute. Mobile phase A was 0.1% formic acid in the water, and mobile phase B was 0.1% formic acid in acetonitrile/isopropanol (90:10). Chromatographic separation was carried out using a Phenomenex C18 column (3 µm, 100 × 2 mm) with the following gradient: 10% B from 0 to 0.5 m, 10 to 97% B from 0.5 to 2 m, 97% B from 2 to 4 m, and 10% B from 4 to 4.25 m. Multiple reaction monitoring in positive mode was used for tandem mass spectrometry analysis. The following mass transitions were monitored for quantitation: TCA (290/131), d5-TCA (295/136), HA (321/160), d5-HA (326/160), and creatinine (114/44).
Statistical analysis
Group means, standard errors/deviations, and linear regressions were calculated in Microsoft Excel. To calculate p values, unpaired student t tests, one-way ANOVA followed by Tukey’s multiple comparison tests, or Welch’s ANOVA with Dunnett’s T3 multiple comparison tests were performed in Graphpad Prism or in Excel. Areas under the curve were calculated with the linear-trapezoidal method using Graphpad Prism and baselines were set by the average values at time 0 for each applicable experiment.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
