A Bacterial Homolog of a Eukaryotic Inositol Phosphate Signaling Enzyme Mediates Cross-kingdom Dialog in the Mammalian Gut

Summary Dietary InsP6 can modulate eukaryotic cell proliferation and has complex nutritive consequences, but its metabolism in the mammalian gastrointestinal tract is poorly understood. Therefore, we performed phylogenetic analyses of the gastrointestinal microbiome in order to search for candidate InsP6 phosphatases. We determined that prominent gut bacteria express homologs of the mammalian InsP6 phosphatase (MINPP) and characterized the enzyme from Bacteroides thetaiotaomicron (BtMinpp). We show that BtMinpp has exceptionally high catalytic activity, which we rationalize on the basis of mutagenesis studies and by determining its crystal structure at 1.9 Å resolution. We demonstrate that BtMinpp is packaged inside outer membrane vesicles (OMVs) protecting the enzyme from degradation by gastrointestinal proteases. Moreover, we uncover an example of cross-kingdom cell-to-cell signaling, showing that the BtMinpp-OMVs interact with intestinal epithelial cells to promote intracellular Ca2+ signaling. Our characterization of BtMinpp offers several directions for understanding how the microbiome serves human gastrointestinal physiology.


Purification of recombinant and selenomethionine-labeled BtMinpp
The 405 amino acids product of the BtMinpp C-terminal region excluding the 20 amino acid Nterminal predicted signal peptide was purified using the His-Tag technique. A PCR fragment was generated using the primer pair MINPPam (CATGCATATGCAAACTAAGATACAGAAGTA) and MINPPav (CATGGGATCCCTATTCATTGAAAACTGA) and the resulting fragment was cloned into the NdeI/BamHI restriction sites of the pET-15b expression vector (Novagen), which carries an pH optimum and pH stability. The assay was performed using a variety of buffers: pH 1.0-3.5, 0.2 M glycine-HCl; pH 4.0-6.0, 0.05 M sodium acetate; pH 6.5-9.5, 0.2 M Tris-HCl. To determine the optimal pH, 2 ng of BtMinpp was incubated for 5 minutes at 37°C at different pH in the presence of 100µM InsP 6 and Pi release was measured under the standard InsP 6 dephosphorylation assay conditions. pH stability was determined after dilution and incubation of 10 ng of enzyme in different buffers and storage at 4°C for 24 hours.
Effect of temperature on enzyme activity. The temperature profile of the purified enzyme was determined in the temperature range from 10 C° to 80 C° using the standard assay. To check thermal stability, the purified enzyme was incubated at different temperatures for 10 minutes, cooled to 4 C°, and assayed using the standard InsP 6 dephosphorylation assay.
Effect of pepsin on enzyme activity. Enzyme inactivation by pepsin was determined by incubating 20 mU of BtMinpp in 0.2 M glycine-HCl pH 2.5, containing 3000 U pepsin for 30 min at 37 C°. After incubation, InsP 6 phosphatase activity was determined using the standard InsP 6 dephosphorylation assay.

HPLC
Inositol phosphate products of assays using 32 P-labelled substrate were resolved by HPLC on a 23.

Crystal structure determination
Recombinant BtMinpp was purified prior to crystallization by gel filtration using a Superdex 75 16/60 column and concentrated to 1 mg/ml. Crystallization proceeded by vapour diffusion using dataset on beamline I-24 of the DLS at the Se K edge (λ = 0.9799 Å) to a resolution of 2.50 Å.
Datasets were processed using MOSFLM (Leslie, 1999) and SCALA (Evans, 2006) as part of the CCP4 package (Collaborative Computational Project, 1994). The structure of SeMet Minpp was determined by SAD phasing using SHELX (Sheldrick, 2008) and chain-traced with BUCCANEER (Cowtan, 2006). This interim structural model was transferred to the cell of the native protein and refined against the native protein structure factors using PHENIX (Adams et al, 2010), alternating with manual rebuilding in COOT (Emsley & Cowtan, 2004). Refinement converged to give a final model for native Minpp displaying R-work and R-free values of 16.6 % and 21.3 %, respectively ( Figure S4). When analyzed for stereochemical quality using MOLPROBITY ) the final structure has only 2 residues (Ala 220 in both molecular copies in the AU) in the disallowed region of the Ramachadran plot. Following refinement this structure was found to contain a single inorganic phosphate ion bound at the active site. To produce crystals of substrate analogue-bound enzyme, phosphate-bound crystals were soaked in mother liquor solution containing 1 mM myoinositol hexakis (hydrogen sulfate) hexapotassium salt (InsS 6 ) and 25 % (v/v) ethylene glycol.
Collection of X-ray diffraction data followed on beamline I-24 of the DLS giving a dataset to a resolution of 2.42 Å. The structure of the InsS 6 -bound complex was solved and refined to give Rwork and R-free values of 15.6 % and 21.7 %, respectively. When analyzed for stereochemical quality using MOLPROBITY , the InsS 6 -bound structure again has only residue Ala 220 in both molecular copies in the disallowed region of the Ramachandran plot. Full data collection, phasing and refinement statistics for all structures reported herein are presented in Table S3III.

Molecular docking calculations
Molecular docking experiments using torsionally-flexible phytate as ligand and the crystal structures of BtMinpp and A.niger PhyA (Kostrewa et al, 1997) as receptors were carried out using AutoDock Vina (Trott & Olson, 2010). A D-2 axial model for InsP 6 (D-2 axial and five equatorial phosphates) was used as representative of the predominant conformation at the acidic pH used in our hydrolysis assays (Isbrandt & Oertel, 1980). Atomic coordinates for the ligand obtained from the Hic-Up database (Kleywegt et al, 2003). The structures of ligand and receptor were formatted with His59 imidazole Nε2 atom to phosphate phosphorous distance of 4 Å or less. These were classified according to the phosphate groups bound in the S3 and S2 specificity subsites (Table SIV).

Site-directed mutagenesis
The A31Y and R183D BtMinpp protein variants were generated using the PCR-based Phusion Site-Directed Mutagenesis kit (Thermo Scientific) according to the manufacturer's instructions. The primer pairs A31Y (GAAGTATGCAGGGACGTACATGCCCTATCCTAATAG),

A31Y_antisense
(TGTATCTTAGTTTGCATATGGCTGCCG) and R183D (CAATATAATCATATCCTTGATTTTTTTGATCTGAATAAATC), R183D_antisense (TTTTCCTTCACTTCGCTGTAC) were used to obtain the A31Y and R183D amino acid substitutions, respectively, using pGH07 as a template. The resulting plasmids were used to transform Rosetta2(DE3)pLysS E. coli cells and protein expression and purification were carried out as described above.

Comparison of the structure of BtMinpp with other Branch 2 Histidine Phosphatases
The refined crystal structure of BtMinpp was compared to a representative subset of the available branch 2 histidine phosphates of known structure (Rigden, 2008). Dali-Lite (Holm & Park, 2000) was  (CATGGGATCCGCCAGCCGTTATGCGGCAGC) was used to amplify a 1252 bp region encoding ermF from the plasmid pFD516 (Smith et al, 1995). The ermF fragment was digested with NdeI and cloned into NdeI digested (blunted) and NsiI digested pGH035, to replace the existing 1803 bp tetQ portion of pGH035, creating pGH038. For overexpression purposes the BspHI and BamHI digested minpp fragment described above was cloned into similar sites of the medium-level expression vector pGH020 (Wegmann et al, 2013) creating pGH036. This was followed by the conversion from tetracycline resistance to erythromycin resistance as described above, creating pGH037. The transformation of the minpp deletion mutant GH59 with pGH037 resulted in the creation GH115.

Periplasmic protein extraction
Bacteroides species were grown in 20 mL of BHI supplemented with 0.001% hemin for 16 hours. The cells were centrifuged at 3500 g for 10 minutes and the periplasmic fraction was prepared according to the method described by Osborn et al. (Osborn et al, 1972). Briefly, the cell pellet was resuspended in 4 ml of fractionation buffer (Tris 30 mM, sucrose 20%, EDTA 1mM, pH8) and incubated for 10 minutes at 20 o C. The cell suspension was centrifuged for 10 minutes at 3000 g and the cell pellet was resuspended in 0.8 ml of ice-cold 5 mM MgSO 4 and the suspension was left on ice for 10 minutes.
The osmotic shock fluid was harvested by centrifugation for 10 minutes at 3000 g, 4°C. To verify that no cross-contamination of cytoplasmic proteins occurred, glucose-6-phosphate dehydrogenase activity was assessed (described in S1 Materials and Methods) in the different periplasmic fractions and no activity was detected.
Glucose-6-phosphate dehydrogenase assay 100 µl of periplasmic or cytoplasmic extract was added to the following solution, pre-incubated at 30°C: 2.5 mL of 50 mM glycylglycine buffer pH 7.4, 0.2 mL of 150 mM MgCl 2, ; 0.1 mL of 20 mM NADP and 0.1 mL of 60 mM glucose-6-phosphate (Noltmann et al, 1961). The increase in A 340 /min for 4 to 5 minutes was recorded (UVIKON XS spectrophotometer, NorthStar Scientific) and the A 340 /minute from the initial linear portion of the curve was calculated.

Outer membrane vesicles protein extraction
Cultures of bacteria were centrifuged at 5000 g for 15 minutes at 4ºC and the supernatant filtered through a 0.22 µm PES membrane (Sartorius) to remove debris and cells. Supernatants were concentrated by molecular weight (100 kDa MWCO, Sartorius) and the retentate ultracentrifuged (150,000 g for 2 h at 4 °C in a Ti70 rotor (Beckman Instruments). The supernatant was carefully aspirated from the tubes and the vesicle pellet was resuspended with 25 mM Tris buffer (pH 7.4). The pellet was washed with the same Tris buffer and centrifuged at 16000 g for 30 min. The pellet was resuspended in Tris buffer and OMV sterility examined by checking for growth of any contaminating bacterial cells on BHIH agar. The OMVs were disrupted by sonication. The OMV protein content was determined using the Total Protein Micro protein assay reagent kit (Sigma).

InsP 6 phosphatase activity measurement
Contaminating inorganic phosphate was removed from Bacteroides cytoplasmic, periplasmic and OMV fractions using PiBind™ resin (Innova Biosciences) prior to InsP 6 phosphatase activity measurements. The phosphatase activity was measured after diluting 10 µl of phosphate-depleted extracts with Tris buffer (25mM, pH 7.5) in the presence of 10 µM of InsP 6 (Merck) and the mixture was incubated at 37ºC for 1 hour. InsP 6 dephosphorylation was measured using a molybdate/malachite green-based activity assay for the quantification of phosphate release (PiColorLock Gold Phosphate Detection System, Innova Biociences) according to the manufacturer's instructions.

Measurement of Minpp activity in the extracellular medium
20 mL of Bacteroides 16-hour cultures in BHI-hemin were centrifuged at 3500 g for 10 minutes and the supernatants were filtered through a 0.22 µm PES membrane (Sartorius) to remove debris and cells. The supernatants were then filtered through a 1000 kDa MWCO (Sartorius) filter to remove OMVs and the filtrate was further filtrated through a 10 kDa MWCO filter (Sartorius) to concentrate the 20 mL solution containing Minpp (49 kDa) down to 200 µl (100x). InsP 6 phosphatase activity was measured following depletion of inorganic phosphate with PiBind™ resin (Innova Biosciences) in 10 µl of the concentrated solution using the malachite green activity assay.

Measurement of the catalytic activity of BtMinpp in resuspended OMVs
Bacterial cells from 500 mL cultures were centrifuged at 5000 g for 20 minutes at 4ºC and the supernatant filtered through a 0.22 µm PES membrane (Sartorius) to remove debris and cells.
Supernatants were concentrated by molecular weight (100 kDa MWCO) using a Vivaflow 50 ultrafiltration device (Sartorius) at 4ºC. To rinse the vesicles, the 5 mL of residual retentate was diluted 100 times in Tris buffer (25mM, pH 7.5) and the resulting 500 mL concentrated to 5 mL. The vesicles were further concentrated to 2 mL with a Vivaspin 20 centrifugal concentrator (Sartorius) and degradation experiments, caecal contents from 4 C57BL/6 mice were homogenised in distilled water, centrifuged at 10,000 g for 5 min and the pellet was washed once with distilled water. Supernatants at each washing step were collected and the volume was adjusted to a final dilution factor of 1:10 (wt:v).
To remove microorganisms, the supernatants were filtered through a 0.22 µm PES membrane (Sartorius). The pH of the extracts was measured at 6.5. 10 µl of vesicle suspension was added to 90 µl of extract to which 1mM of InsP 6 was added.