Nitrogen Metabolism in Pseudomonas putida: Functional Analysis Using Random Barcode Transposon Sequencing

Abstract

Pseudomonas putida KT2440 has long been studied for its diverse and robust metabolisms, yet many genes and proteins imparting these growth capacities remain uncharacterized. Using pooled mutant fitness assays, we identified genes and proteins involved in the assimilation of 52 different nitrogen containing compounds. To assay amino acid biosynthesis, 19 amino acid drop-out conditions were also tested. From these 71 conditions, significant fitness phenotypes were elicited in 672 different genes including 100 transcriptional regulators and 112 transport-related proteins. We divide these conditions into 6 classes, and propose assimilatory pathways for the compounds based on this wealth of genetic data. To complement these data, we characterize the substrate range of three promiscuous aminotransferases relevant to metabolic engineering efforts in vitro. Furthermore, we examine the specificity of five transcriptional regulators, explaining some fitness data results and exploring their potential to be developed into useful synthetic biology tools. In addition, we use manifold learning to create an interactive visualization tool for interpreting our BarSeq data, which will improve the accessibility and utility of this work to other researchers. IMPORTANCE Understanding the genetic basis of P. putida's diverse metabolism is imperative for us to reach its full potential as a host for metabolic engineering. Many target molecules of the bioeconomy and their precursors contain nitrogen. This study provides functional evidence linking hundreds of genes to their roles in the metabolism of nitrogenous compounds, and provides an interactive tool for visualizing these data. We further characterize several aminotransferases, lactamases, and regulators, which are of particular interest for metabolic engineering.

Keywords: BarSeq; Pseudomonas putida; RB-TnSeq; amino acid; aminotransferase; aminotransferases; biosensor; biosensors; lactam; metabolism; nitrate; nitrite; nitrogen; nitrogen metabolism; nucleotide; polyamine; t-SNE; transposon.

FIG 1

Global analysis of the P. putida KT2440 BarSeq data. (A) Significant genes (|fitness| > 1 and |t | > 5) from all 71 tested conditions sorted by their cluster of orthologous groups (COGs) based on the eggNOG database (149, 153, 154). “Multiple COGs” indicates that there was more than one COG assigned. (B) Image of the interactive t-SNE visualization (Fig. I1 available at https://ppnitrogentsne.lbl.gov) showing the legend, t-SNE clustering (left) and cluster centroids (right). By clicking on a substrate in the legend, the corresponding cluster (left) and centroid (right) is highlighted, opening a list of cluster members and additional information. By clicking the highlighted centroid (right), the user is redirected to the Fitness Browser (https://fit.genomics.lbl.gov) (20), where the fitness data for all significant genes in the condition cluster is shown. An additional t-SNE visualization including COG identifiers is presented in Fig. I3 (available at https://ppnitrogentsne.lbl.gov). More information about the interactive figures can be found in the description of Fig. S3.

FIG 2

(A) Scatterplot of the average fitness values (n = 2) for ntrC and gltBD in all the tested nitrogen conditions. For ntrC, nitrogen conditions are grouped based on whether the fitness phenotype of ntrC is < −1.5 (blue) or > −1.5 (green). Conditions where ntrC fitness is >−1.5 (green) may be less dependent on ntrC activation and are shown in Table 2. gltBD phenotypes are sorted based on putative glutamate (brown), glutamate+ammonium (pink), or ammonium (gray) release during nitrogen source utilization. Fitness values for conditions resulting in gltBD fitness >−1.5 are shown in Table 3. (B) The role of the GS/GOGAT cycle in the beta-alanine (red), 3-aminoisobutyrate (green), 4-aminobutyrate (blue), and 5-aminovalerate (orange) nitrogen source conditions. Average fitness values (n = 2) are shown for gltBD, davT, and the pyruvate dependent transaminase PP_0596. *GlnA lacks fitness data because the library has no insertions. **Pyruvate and alanine are the specific nitrogen acceptor and product of PP_0596.

FIG 3

The assimilatory nitrate reduction system in P. putida KT2440. Average fitness values (n = 2) exhibited in the nitrate (green), nitrite (red), and ammonium (yellow) sole nitrogen source experiments. Shown are putative transporters, action of the NasST regulatory system, the assimilatory pathway, and the role of the bis-molybdopterin guanine dinucleotide (bis-MGD) cofactor. bis-MGD is the required cofactor for NarB (155). The fitness phenotypes for bis-MGD biosynthesis cluster together with the nitrate phenotypes and can be found in the interactive t-SNE visualization (Fig. I1 available at https://ppnitrogentsne.lbl.gov).

FIG 4

Quaternary amine and ethanolamine degradation in P. putida. (A) Putative routes for the quaternary amine catabolism in P. putida KT2440. The figure shows the degradation of choline (red), carnitine (green), and betaine (blue). The corresponding average fitness scores (n = 2) are shown next to each gene. (B) Ethanolamine degradation pathway, shown with fitness values (n = 2) and regeneration of the AdoCbl cofactor. (C) Heatmap with average fitness scores (n = 2) of genes that are putatively involved in P. putida’s adenosylcobalamin biosynthesis. No fitness scores (gray) could be obtained for the genes PP_1680 (cobV) and PP_3410 (cobM).

FIG 5

Putative routes for purine (A) and pyrimidine (B) catabolism in P. putida KT2440. Shown are the average fitness scores (n = 2) for genes involved in adenine (teal), guanine (yellow), cytosine (blue), uracil (red), and hydantoin (green) degradation.

FIG 6

(A) Heatmap with fitness scores (n = 2) of genes putatively involved in the hydrolysis of caprolactam, valerolactam, butyrolactam, and 5-oxoproline in P. putida KT2440. (B) LC-MS analysis of caprolactam degradation in P. putida KT2440. Wild-type cells were grown in MOPS minimal media with ammonium (blue) and caprolactam (orange) as the sole source of nitrogen. Shown is the OD (squares) and concentration of adipate in the supernatant (circles) over a time course of 72 h. The dashed line marks the time point at which caprolactam was no longer detected in the media.

FIG 7

Heatmaps of polyamine and ω-amino acid catabolism in P. putida. For each enzyme class shown, HMMER was used to identify all putative genes present in P. putida corresponding to that pFam in P. putida. Data was filtered to find genes in each class with fitness values >–1.0 and t > |5| in the nitrogen sources shown. (A) Heatmap with fitness values (n = 2) for genes putatively involved in polyamine transamination in P. putida KT2440. (B) Heatmap with fitness values (n = 2) for genes putatively involved in polyamine γ-glutamylation in P. putida KT2440. (C) Heatmap with fitness values (n = 2) for genes putatively involved in omega-amino acid transamination in P. putida KT2440. Spermidine (SPRM) = teal; 1,3 diaminopropane (1,3 DAP) = yellow; β-alanine = yellow; putrescine/1,4-diaminobutane (1,4 DAB) = black; γ-aminobutyric acid (GABA) = black; cadaverine/1,5-diaminopentane (1,5 DAP) = blue; 5-aminovalerate (5AVA) = blue; 1,6-diaminohexane (1,6 DAH) = green; 6-aminocaproic acid (6ACA) = green.