The nucleotide messenger (p)ppGpp is an anti-inducer of the purine synthesis transcription regulator PurR in Bacillus

Abstract

The nucleotide messenger (p)ppGpp allows bacteria to adapt to fluctuating environments by reprogramming the transcriptome. Despite its well-recognized role in gene regulation, (p)ppGpp is only known to directly affect transcription in Proteobacteria by binding to the RNA polymerase. Here, we reveal a different mechanism of gene regulation by (p)ppGpp in Firmicutes: (p)ppGpp directly binds to the transcription factor PurR to downregulate purine biosynthesis gene expression upon amino acid starvation. We first identified PurR as a receptor of (p)ppGpp in Bacillus anthracis. A co-structure with Bacillus subtilis PurR reveals that (p)ppGpp binds to a PurR pocket reminiscent of the active site of phosphoribosyltransferase enzymes that has been repurposed to serve a purely regulatory role, where the effectors (p)ppGpp and PRPP compete to allosterically control transcription. PRPP inhibits PurR DNA binding to induce transcription of purine synthesis genes, whereas (p)ppGpp antagonizes PRPP to enhance PurR DNA binding and repress transcription. A (p)ppGpp-refractory purR mutant in B. subtilis fails to downregulate purine synthesis genes upon amino acid starvation. Our work establishes the precedent of (p)ppGpp as an effector of a classical transcription repressor and reveals the key function of (p)ppGpp in regulating nucleotide synthesis through gene regulation, from soil bacteria to pathogens.

Figure 1.

DRaCALA screen identifies the transcription factor PurR as a (p)ppGpp target in Firmicutes. (A, B) Z-scores of ³²P-pppGpp binding to an open reading frame (ORF) library from B. anthracis. His indicates hexahistidine-tagged ORFs, and HisMBP indicates hexahistidine maltose binding protein tagged ORFs. ORFs with Z-scores greater than 2.5σ are filled. The dashed horizontal line is at 2.5σ. The red circle is the PurR ORF. Yellow symbols indicate ORFs that were hits without GTP competition but were not hits with GTP (see C, D). These data were previously reported in Yang et al. (2020). (C, D) DRaCALA of ORF libraries from (A) and (B) with 100 μM non-radioactive GTP as a competitor to reveal (p)ppGpp-specific targets. ³²P-pppGpp is at ∼0.2 nM. (E) Schematic of PurR regulation of de novo ATP and GTP synthesis. (F) Cladogram constructed from 106 bacterial 16S rRNA and three eukaryotic 18S rRNA sequences. Tree branches colored according to the phylum. A red dot on the branch tip indicates that the species contains a B. subtilis PurR homolog. (G) Binding curve between purified untagged B. subtilis PurR and ³²P-ppGpp (gray) and ³²P-pppGpp (purple) obtained with DRaCALA. Error bars represent SEM of technical triplicate. Error bars are not visible when they are smaller than the height of the symbols. (H) Binding parameters from (G). mean ± standard error. h = Hill coefficient. (I) DRaCALA of ³²P-ppGpp binding to 30 μM purified HisMBP-tagged PurR proteins from Clostridium sporogenes (Csp), Enterococcus faecalis (Efa), and Streptococcus mutans (Smu). Control is protein storage buffer.

Figure 2.

ppGpp binds to the effector binding domain of PurR. (A) B. subtilis PurR dimer crystallized with ppGpp. The N-terminal DNA binding domain is blue and the C-terminal effector binding phosphoribosyltransferase (PRT) domain is salmon. (Right) Structure from left rotated 90° to show the ppGpp binding pockets on the effector domains. (B) ppGpp binding pocket with select interacting residues indicated in silver. Loops I–IV of the PRT binding pocket are labeled. See Supplementary Figure S1B for an omit electron density map. (C) Frequency logo of the (p)ppGpp binding residues from 938 PurR proteins homologous to B. subtilis PurR. The logo was made with WebLogo (UC Berkeley; https://weblogo.berkeley.edu/logo.cgi). (D) Overlay of apo PurR (green; PDB ID 1O57) and ppGpp-bound PurR (salmon). Bridging loop that is unresolved in apo but not with ppGpp is indicated. (E, F) Poisson-Boltzmann continuum electrostatics of the effector binding domain of apo (E) and ppGpp-bound (F) PurR. The scale of electrostatic potential is –5 (red) to + 5 (blue).

Figure 3.

ppGpp competes with PRPP to allosterically regulate PurR-DNA interaction. (A) Overlay of PurR-ppGpp (salmon) with PurR-cPRPP (yellow; PDB ID 1P4A). The side chains of Y102 and D203 are shown. Loop II is hidden for clarity. (B) Competition between ³²P-ppGpp (∼0.2 nM) and PRPP for binding to PurR and PurR D203A. Binding of ³²P-ppGpp was measured with DRaCALA. Reactions were performed in technical triplicate, and error bars represent SEM. Error bars are not visible when their height is smaller than the height of the symbol. (C) EMSA showing PurR interaction with FAM-labeled 221 bp DNA with increasing PRPP concentrations and without ppGpp. (D) EMSA showing PurR interaction with DNA with increasing ppGpp concentration and with PRPP. The PurR concentration is 100 nM in (C) and (D). Similar EMSA results were observed with a non-labeled 202 bp probe from the same control region with a lower KCl concentration and no nonspecific DNA in the EMSA reaction (Supplementary Figure S3). (E) DNase I footprinting of PurR-PurBox interaction. ³²P-labeled PurBox fragment was incubated with or without PurR (50 nM), ppGpp (1 mM), and PRPP (16 μM), and then briefly digested by DNase I (3.2 μg/ml). The digestion product was analyzed by electrophoresis. The uncropped gel is in Supplementary Figure S4. (F) Relative intensity of PurBox 2 in PurR-PurBox DNA footprint. Relative intensity is the intensity of the PurBox 2 area normalized to the intensity of a reference area (raw data in Supplementary Figures S5 and S6). Error bars represent standard error of the mean for 4 replicates. A two-tailed two-sample equal-variance Student's t test was performed between samples for statistical significance (*P ≤ 0.05; n.s. P > 0.05).

Figure 4.

(p)ppGpp enhances PurR-DNA interaction during stress response. (A) Genome-wide view of mean PurR ChIP reads per million reads from B. subtilis grown with or without the nucleobases adenine, cytosine, guanine, and uracil. (B) Schematic of the GTP and ATP synthesis pathway in B. subtilis. Red arrows indicate PurR-controlled steps, as deduced from the respective genes being downstream of a PurR binding site. 10-fTHF = 10-formyl tetrahydrofolate; ser = serine; gly = glycine; thr = threonine; pyr = pyruvate; OAA = oxaloacetate; asp = aspartate; hpx = hypoxanthine; xan = xanthine; gua = guanine. Placements of pycA and steT in this pathway are inferred based on enzyme function and have not been verified. (C) PurR represses the 12-gene pur operon in B. subtilis, and PRPP is an inducing ligand. (D) Sequence logos of PurBox1 (top) and PurBox2 (bottom) sequences from 12 of the 15 PurR ChIP peaks. Logos were created using WebLogo (https://weblogo.berkeley.edu/logo.cgi). See Supplementary Figure S10 for complete sequences. (E) A plot showing the distance of PurBoxes from the ChIP enrichment peaks. The peak is represented by the dotted horizontal line at Y = 0. Peak location was determined at 10 bp resolution from the PurR ChIP sample obtained from (p)ppGpp-induced WT B. subtilis. The distance was calculated from the center of each PurBox (7th nt out of 15 nt PurBox) to the peak location. (F) Genome-wide view of mean PurR ChIP reads per million reads in WT and (p)ppGpp⁰ before and after arginine hydroxamate (RHX) treatment. Before RHX are the same data as without nucleobases in A. (G) PurR ChIP enrichment at the pur operon in WT and (p)ppGpp⁰. The solid trace is the mean enrichment of biological triplicate, and SD is shown as the shaded region. The inset shows the PurBox sites (shaded box), −10 and −35 promoter sequences (blue box), and the transcription start site (arrow) upstream of the pur operon. The sequence for this site is shown in Supplementary Figure S12.

Figure 5.

(p)ppGpp regulation of PurR is important for nutrient stress adaptation. (A) Change in transcript level of PurR-regulated genes after RHX treatment. In wild-type B. subtilis, genes are highly downregulated, with some genes in the pur operon being downregulated over 30-fold. In comparison, there is only a minor effect of RHX on expression of PurR-regulated genes in (p)ppGpp⁰. These data are from Kriel et al. (2014). Mean of triplicate ± SD is shown. (B) DRaCALA binding curves showing ³²P-ppGpp interaction with PurR Y102A/F205A and Y102A/K207A. Experiments were performed in technical triplicate. Error bars representing SEM may be shorter than the height of the symbols. (C) Expression of a P_purE-GFP reporter in wild-type, (p)ppGpp⁰, purR^Y102A/F205A, and purR^Y102A/K207AB. subtilis. (D) Doubling times of B. subtilis wild type, ΔpurR, purR^D203A, purR^Y102A/F205A, and purR^Y102A/K207A in minimal medium (Min) and Min supplemented with 20 amino acids (Min + 20aa). An unpaired parametric two-tailed t-test with Welch's correction was used to compare wild type and mutants’ doubling times (** P ≤ 0.01; all other pairings are not significant, P > 0.05). (E) Growth of wild type, purR^Y102A/F205A and purR^Y102A/K207A in minimal medium following a nutrient downshift from Min + 20aa to Min media. All growth experiments were performed in biological triplicate, and error bars are SEM of triplicate. For OD and fluorescence curves, error bars are represented by dotted lines.

Figure 6.

Models of (p)ppGpp regulation of PurR and global (p)ppGpp regulation of purine nucleotide synthesis in B. subtilis. (A) Schematic of PurR evolution from PRT enzymes like XPRT. PurR repurposes the active site of PRT enzymes to serve as an effector binding pocket, which PRPP and ppGpp bind to allosterically regulate DNA binding. (B) GTP and ATP synthesis in B. subtilis. (p)ppGpp regulates GTP synthesis at multiple points, including inhibition of GTP synthesis enzymes (IMPDH, GMK, HPRT, XPRT) and binding the transcription factor PurR. Red arrows represent steps in the PurR regulon.