A glance at the following article's methods section reveals the extent of EFI's multidisciplinary approach: docking and modeling, protein over-expression, crystallization, metabolomics, gene disruption, proton NMR, polarimetry, transcriptomics... Spearheaded by researchers in the Microbiology Core (with support provided by the EN Bridging Project, Protein Core, Structure Core, and Computation Core), this work demonstrates that the powerful coupling of genome neighborhood analysis and computational modeling can result in the expeditious discovery of not only catabolic pathways, but also mechanisms for the regulation of gene expression.
Through the use of genetic, enzymatic, metabolomic, and structural analyses, we have discovered the catabolic pathway for proline betaine, an osmoprotectant, in Paracoccus denitrificans and Rhodobacter sphaeroides. Genetic and enzymatic analyses showed that several of the key enzymes of the hydroxyproline betaine degradation pathway also function in proline betaine degradation. Metabolomic analyses detected each of the metabolic intermediates of the pathway. The proline betaine catabolic pathway was repressed by osmotic stress and cold stress, and a regulatory transcription factor was identified. We also report crystal structure complexes of the P. denitrificans HpbD hydroxyproline betaine epimerase/proline betaine racemase with l-proline betaine and cis-hydroxyproline betaine. IMPORTANCE At least half of the extant protein annotations are incorrect, and the errors propagate as the number of genome sequences increases exponentially. A large-scale, multidisciplinary sequence- and structure-based strategy for functional assignment of bacterial enzymes of unknown function has demonstrated the pathway for catabolism of the osmoprotectant proline betaine.
Figure 1: “Genome contexts of HpbD in P. bermudensis and the orthologous genes in P. denitrificans and R. sphaeroides. The genes encoding orthologues are given the same color, and the sequence identities relating orthologous genes are indicated. More detail is given in Table S1 in the supplemental material. Genes separated by a string of dots are unlinked in the genome. Note that in vivo study of P. bermudensis was precluded by its slow growth and very low maximal cell density and by a lack of genetic tools.
Figure 2: (A) Comparison of two homology models, pdHpbD (magenta) and rsHpbD (orange), with the X-ray structure of pdHpdD (cyan; PDB 4J1O). (B) The Glide XP docking positions of proline betaine in the two models are highly similar to that of the cocrystallized proline betaine of the X-ray structure. Asp194, Glu219, and Asp242 are metal (shown in spheres) binding residues. Lys164 and Lys266 are two catalytic bases, whereas Trp312 and Asp293 are important substrate binding residues.
Figure 3: Effects of L- or D-Pro-B on growth of wild-type P. denitrificans as the carbon source and as an osmoprotectant. (A) Growth without added NaCl. Growth on glucose is also shown. (B) Growth in the presence or absence of 0.5 M NaCl with or without L- or D-Pro-B supplementation. The L- or D-isomers of ProB were used at a concentration of 20 mM. (C) Effects of addition of the L- or D-isomers of ProB on growth of wild-type R. sphaeroides in minimal media containing 0.5 M NaCl. Glucose (20 mM) was present in all cultures.
Figure 4: Proposed pathway for L-Pro-B and D-Pro-B catabolism based on the genome neighborhood contexts in P. bermudensis, P. denitrificans, and R. sphaeroides. The interconversion of L-Pro-B and D-Pro-B is catalyzed by HpbD; L-Pro-B undergoes two N demethylations to L-Pro; and finally, L-Pro oxidation is catalyzed by a bifunctional enzyme PutA to give L-glutamate, which is readily deaminated to ammonia plus 2-ketoglutarate, the citric acid cycle intermediate.
Figure 5: Growth phenotypes. (A) Growth of the P. denitrificans wild-type strain (Pd1222) or mutant strains (RPd1 and RPd10) in minimal medium with L- or D-Pro-B. The hpbD strain was RPd1, and the complemented strain was RPd10. (B) Growth of the P. denitrificans wild-type strain (Pd1222) or mutant strains (RPd4 and RPd11) in minimal medium with D-Pro-B or N-Me-Pro. The strain missing both Rieske demethylases (genes hpbB1 and hpbC1 and genes hpbB2 and hpbC2) was RPd4, and the complemented strain was RPd11. (C) Growth on N-Me-Pro as the carbon source. The P. denitrificans hpbA strain was RPd6, and the complemented strain was RPd12.
Figure 6: Accumulation of D-proline betaine catabolic metabolites measured by LC-FTMS. (A to D) Extracted ion chromatograms of representative LC-FTMS runs of extracts of the P. denitrificans hpbD, hpbB1, and hpbC1 and hpbB2 and hpbC2 strain, RPd14, fed L-Pro-B for 1 h (black curves), wild-type P. denitrificans PD1222 grown on D-glucose (red curves), and the wild-type strain grown on L-Pro-B (blue curves) and standards (purple curves). The cells were extracted with 10 mM ammonium bicarbonate (pH 9.2)-buffered acetonitrile (Materials and Methods). Panels A to D correspond to the intensities (5-ppm mass accuracy) for proline betaine (M+; 144.10191 Da), N-Me-Pro (M+; 130.08626 Da), proline (M+; 116.07060 Da), and glutamate (M−; 148.06043 Da), respectively. (E) Averaged integrated intensities of the four peaks for the two strains, with conservation of the color coding of panels A to D. Standard errors are shown for four biological replicates (three for strain RPd14); each biological replicate value was the average of the results of three technical replicates. More details are given in Fig. S5 in the supplemental material.
Reprinted with permission from mBio. Copyright © 2014 American Society of Microbiology.