The complex logic of stringent response regulation in Caulobacter crescentus: starvation signalling in an oligotrophic environment

Abstract

Bacteria rapidly adapt to nutritional changes via the stringent response, which entails starvation-induced synthesis of the small molecule, ppGpp, by RelA/SpoT homologue (Rsh) enzymes. Binding of ppGpp to RNA polymerase modulates the transcription of hundreds of genes and remodels the physiology of the cell. Studies of the stringent response have primarily focused on copiotrophic bacteria such as Escherichia coli; little is known about how stringent signalling is regulated in species that live in consistently nutrient-limited (i.e. oligotrophic) environments. Here we define the input logic and transcriptional output of the stringent response in the oligotroph, Caulobacter crescentus. The sole Rsh protein, SpoT(CC), binds to and is regulated by the ribosome, and exhibits AND-type control logic in which amino acid starvation is a necessary but insufficient signal for activation of ppGpp synthesis. While both glucose and ammonium starvation upregulate the synthesis of ppGpp, SpoT(CC) detects these starvation signals by two independent mechanisms. Although the logic of stringent response control in C. crescentus differs from E. coli, the global transcriptional effects of elevated ppGpp are similar, with the exception of 16S rRNA transcription, which is controlled independently of spoT(CC). This study highlights how the regulatory logic controlling the stringent response may be adapted to the nutritional niche of a bacterial species.

Figure 1. SpoT_CC responds to glucose and ammonium, but not amino acid or phosphate starvation

A) ppGpp accumulation in wild-type cells in glucose, ammonium and phosphate starvation. For glucose and ammonium starvation, n=4; for phosphate starvation, n=2. B) ppGpp production in a series of auxotrophic strains starved for their respective amino acids at 5 minute and 120 minute time points. Data for separate auxotrophic strains starved individually for 17 different amino acids are similar and have been averaged (black line). ppGpp starvation curves for individual amino acid auxotrophs are presented in Figure S1. Also, presented, ppGpp accumulation in wild-type cells cultured in the presence of all 11 non-inhibitory amino acids and then simultaneously deprived of these amino acids (gray line; N=2). Error bars refer to standard deviation for all experiments.

Figure 2. SpoT_CC associates with the ribosome

A) sucrose-gradient polysome profile (top) and western blot (bottom) of pooled fractions from the soluble, 70S and polysome peaks. The top blot strip is from the HA-SpoT strain, the bottom strip is from the wild-type strain, which serves as a negative control. B) Western blots of supernatant (S) and pellet (P) fractions from cell lysates spun through a 1M sucrose cushion to separate the ribosomes from other cell constituents (See Figure S2). HA-SpoT co-purifies with the ribosome when the C-terminal domains are removed (ΔCTD) and in the L11 P22L ribosomal mutant. C) HA-SpoT co-purifies with the ribosome in M2G and after 15 minutes of glucose starvation in M2. FixJ is presented as a non-ribosome-associated control.

Figure 3. SpoT_CC response to starvation is abrogated by certain ribosome poisons

A) ppGpp decay in wild-type cells starved for glucose for two hours and then treated with chloramphenicol (50 µg/ml), tetracycline (50 µg/ml), kanamycin (250 µg/ml), spectinomycin (1250 µg/ml) or streptomycin (50 µg/ml). B) ppGpp accumulation in wild-type cells grown in M2G, and then washed and resuspended in M2 without glucose and with an antibiotic at the same concentration as in A. C) Western blot of HA-SpoT culture aliquots taken before (0) and after 50 µg/ml tetracycline was added to the culture. D) Western blots of supernatant (S) and pellet (P) fraction from an HA-SpoT culture that was treated with 50 µg/ml tetracycline for 15 minutes before being lysed and separated through a sucrose cushion. FixJ is a non-ribosome-associated control. E) ppGpp decay in wild-type cells starved for glucose for two hours and then treated with 50 µg/ml tetracycline or 0.2% glucose for 20 minutes: the tetracycline-treated cells were then treated with 0.2% glucose and analyzed for ppGpp content for an additional 10 minutes. F) ppGpp decay in wild-type cells starved for ammonium for two hours, treated with 50 µg/ml tetracycline for 20 minutes, then treated with 9.3 mM NH₄Cl and analyzed for ppGpp content for an additional 10 minutes. N=2 for each experiment, error bars refer to the standard deviation.

Figure 4. Glucose starvation in the presence of amino acids does not induce significant ppGpp accumulation

A) ppGpp accumulation in Δfbps (glucose auxotroph) and wild-type cells starved for glucose with all 20 amino acids supplemented. Cells were grown in M2G with all 20 amino acids supplemented at 100 µg/ml each. Cells were then washed and resuspended in media with all 20 amino acids but no glucose, and ppGpp measured at times after glucose removal. The data for both strains starved for glucose without amino acid supplementation is included for comparison. B) ppGpp decay in Δfbps cells starved for glucose for two hours and then treated with addition of all 20 amino acids, after 30 minutes 0.2% glucose was added and further time points were taken. N=2 for each experiment, error bars refer to the standard deviation.

Figure 5. Glucose and ammonium starvation are sensed by different mechanisms

A) Schematic showing domain structure of the full length SpoT protein and domain-deletion alleles. Locations of IS insertion elements found in hydrolase suppressor mutants are indicated by black triangles. (See last section of Results). B) In vivo ppGpp accumulation by the wild-type and mutant SpoT proteins in M2G, or after 2 hours of glucose or ammonium starvation. The stars indicate statistically significant difference, by student’s t-test, from the ΔCTD strain for each condition where *= p<0.01 and **=p<0.001. N=3. C) Western blot showing the stability of each domain-deletion protein in cells. N-terminal HA-tagged versions of each mutant were used. FixJ was used as a loading control. D) ppGpp accumulation in an L11 P22L mutant starved for glucose or ammonium. N=2. Error bars refer to standard deviation for all experiments

Figure 6. Microarray analysis of genes regulated by C. crescentus SpoT_CC

The genes included in this figure are protein-coding open reading frames that were regulated at least 3-fold in the wild-type versus ΔspoT_CC Affymetrix data set. The size of each pie is proportional to the number of genes in that category. In total, 379 genes were upregulated at least 3-fold (17-anabolic, 120-catabolic, 126-other processes, 8-cell cycle/nutrient granule, 108-unknown function), and 382 genes were downregulated at least 3-fold by spoT (166-anabolic, 44-catabolic, 94-other processes, 6-cell cycle/nutrient granule, 71-unknown function) (see Table S1). Genes of unknown function and cell cycle/nutrient granule genes in Table 2 are not included in the pie charts

Figure 7. rRNA transcriptional control

A) Schematic of 16S transcriptional reporter plasmid, which produces an unstable transcript from the 16S A rRNA promoter. Numbers below genes represent the number of nucleotides. Location of qRT-PCR amplicon primers are indicated by small arrows. B) 16S rRNA promoter activity in wild-type, ΔspoT and ΔdksA strains before (0 minutes) and during an hour of glucose starvation. N=3. C) 16S rRNA promoter activity in the prototrophic strain (NA1000 pxyl∷pP16Slacz) and the auxotrophic strains (NA1000 ΔgltB pxyl∷pP16Slacz and NA1000 ΔpheA pxyl∷pP16Slacz) in M2G with and without added glutamate or phenylalanine. The cells were grown up in M2G + glu or M2G + phe + ala, and then washed and resuspended in M2G or M2G + ala and grown in those conditions for two hours before the starvation samples were taken. All data were normalized to the signal from the ruvA amplicon on the same biological sample (N=3). Error bars refer to standard deviation for all experiments.

Figure 8. Summary of the regulatory logic of the stringent response in a copiotrophic and an oligotrophic species, and a ribosome model

A) Comparison of the signaling proteins, signaling logic, and signaling output of the stringent response in the oligotroph, C. crescentus, and the copiotroph, E. coli. The dashed lines from the OR gate of C. crescentus represent the different signals that activate the stringent response during carbon and nitrogen starvation. B) Cartoon of the bacterial 70S ribosome. SpoT_CC is in green. SpoT_CC is presumed to be near the 70S A site and ribosomal protein L11 based on regulatory data presented herein.