Fig 1.
Gaps in amino acid biosynthesis in 10 bacteria.
For each amino acid, we identified the missing reactions or gaps in the IMG predictions, and we show a single-letter code with the classification of each gap. A cell may have multiple codes, one for each gap. “Clear candidate” means that at least two annotation resources identified a gene for the gap. If there was more than one gap with a clear candidate, this is shown with a number (i.e., “c2” for two gaps with clear candidates). “Known” means that the step is described in the literature but not in the databases. Some reactions are involved in the biosynthesis of multiple amino acids, so some gaps are shown multiple times.
Table 1.
Classification of the 173 gaps in amino acid biosynthesis pathways in 10 bacteria.
Fig 2.
Mutant fitness data for clear candidates that have auxotrophic phenotypes.
We selected one gene at random from each organism that has such a gene. In each panel, the x axis shows gene fitness (values below -7 are shown at -7), and the y axis separates experiments by whether most amino acids were available (green points) or not (red points). In between, we show experiments in which only the relevant amino acid was provided (blue points, if any). Within each category, the y axis is random.
Fig 3.
Synthesis of methionine and threonine in Desulfovibrio vulgaris Miyazaki F.
We grew a pool of transposon mutants of D. vulgaris in a defined lactate/sulfate medium with or without added amino acids. L-methionine or L-threonine were supplemented at either 1 mM or 10 mM. D,L-cysteine, D,L-histidine, or L-serine were supplemented at 1 mM.
Fig 4.
Methionine synthesis in Phaeobacter inhibens by a three-part methionine synthase and two vitamin B12 reactivation proteins.
(A) Domain content of MetH from E. coli and of the three-part methionine synthase of P. inhibens. (B) Fitness data of the methionine synthesis genes. We grew pools of mutants of P. inhibens aerobically in defined media with a variety of carbon and nitrogen sources. (C) Phylogenetic profile of the presence or absence of the vitamin B12 reactivation proteins across 158 α-Proteobacterial genomes from MicrobesOnline [29]. The bacteria are ordered by evolutionary relationships and some of the genera that contain these proteins are labeled.
Fig 5.
Serine biosynthesis in Burkholderia phytofirmans.
(A) A heatmap of gene fitness of the putative phosphoglycerate dehydrogenase (“FAD-linked oxidase”) and another serine biosynthesis gene (the aminotransferase serC). Our standard minimal media for this organism contains glucose and ammonia, and CAS is short for casamino acids, which contains both L-serine and L-histidine. (B) A comparison of gene fitness during growth in minimal media (x axis) or in media that was supplemented with 1 mM L-serine (y axis). The lines show x = 0, y = 0, or x = y. We highlight the genes from part (A) as well as the catabolic serine dehydratase. The other point near serC and the dehydrogenase is a putative cell wall synthesis gene (BPHYT_RS14855, ADP-L-glycero-D-manno-heptose-6-epimerase). Each point shows the average of two replicate experiments.
Fig 6.
Identification of novel histidinol-phosphate phosphatases.
We compared gene fitness in minimal media (x axis) or in minimal media supplemented with 1 mM L-histidine (y axis) for (A) Burkholderia phytofirmans, (B) Pseudomonas stutzeri, (C) Marinobacter adhaerens, and (D) Herbaspirillum seropedicae. In each panel, we highlight the putative phosphatase as well as genes for other steps in histidine biosynthesis. For H. seropedicae, the phosphatase is not shown because of a lack of data. The lines show x = 0, y = 0, and x = y. Each point shows the average gene fitness from two replicate experiments.
Fig 7.
DVU1186 is required for histidine synthesis.
We grew transposon mutants of DVU1186 (strain GZ8414) and of DVU2938 (strain GZ9865), as well as wild type D. vulgaris Hildenborough, in defined lactate-sulfate medium (panel A) or in the same medium with 0.1 mM L-histidine added (panel B). For each strain and condition, we show three replicates.
Table 2.
Sources of wild-type strains and their minimal media.