CcpA mediates proline auxotrophy and is required for Staphylococcus aureus pathogenesis

Abstract

Human clinical isolates of Staphylococcus aureus, for example, strains Newman and N315, cannot grow in the absence of proline, albeit their sequenced genomes harbor genes for two redundant proline synthesis pathways. We show here that under selective pressure, S. aureus Newman generates proline-prototrophic variants at a frequency of 3 x 10(-6), introducing frameshift and missense mutations in ccpA or IS1811 insertions in ptsH, two regulatory genes that carry out carbon catabolite repression (CCR) in staphylococci and other Gram-positive bacteria. S. aureus Newman variants with mutations in rocF (arginase), rocD (ornithine aminotransferase), and proC (Delta(1)-pyrroline 5-carboxylate [P5C] reductase) are unable to generate proline-prototrophic variants, whereas a variant with a mutation in ocd (ornithine cyclodeaminase) is unaffected. Transposon insertion in ccpA also restored proline prototrophy. CcpA was shown to repress transcription of rocF and rocD, encoding the first two enzymes, but not of proC, encoding the third and final enzyme in the P5C reductase pathway. CcpA bound to the upstream regions of rocF and rocD but not to that of proC. CcpA's binding to the upstream regions was greatly enhanced by phosphorylated HPr. The CCR-mediated proline auxotrophy was lifted when nonpreferred carbohydrates were used as the sole carbon source. The ccpA mutant displayed reduced staphylococcal load and replication in a murine model of staphylococcal abscess formation, indicating that carbon catabolite repression presents an important pathogenesis strategy of S. aureus infections.

FIG. 1.

Two pathways for proline synthesis in S. aureus. The two pathways share the first step (from arginine to ornithine), but they diverge at the step(s) of ornithine's conversion to proline. Gene names and locus tags in the Newman genome are indicated in parentheses.

FIG. 2.

Growth patterns of the ccpA mutant and the complemented strain in a proline-deficient medium. The complementation plasmid, pOS1-ccpA, was constructed by cloning ccpA in the E. coli-S. aureus shuttle plasmid pOS1. Cells were grown in CDM-P at 37°C for 30 h, and their optical density was measured at 600 nm. 7791, ccpA transposon insertion mutant ΦΝΞ-7791; WT, wild type.

FIG. 3.

Identification of mutations in ccpA (A) and ptsH (B). ccpA and ptsH were amplified from nine proline-prototrophic variants by PCR, and the products were sequenced. Insertion mutations are shown by inverted triangles, while other types of mutations are indicated by vertical arrows; a delta indicates a deletion. The variant identification numbers are also shown, to the right of the mutations. Amino acid changes are indicated in parentheses; an asterisk indicates a stop codon.

FIG. 4.

A proline prototroph emergence test. Proline synthesis mutants were grown in CDM-P for 48 h, and then the emergence of proline prototrophs was measured by the optical density at 600 nm. Mutated genes are shown under the graph. A mutant of argD encoding acetylornithine aminotransferase (NWMN_0129) involved in arginine synthesis was used as a negative control. Standard deviations are also indicated on the graph. wt, wild-type.

FIG. 5.

Transcript analysis for proline synthesis genes. (A) The wild type (wt) and proline-prototrophic variants (NMpro+5, NMpro+18, NMpro+24, NMpro+33, and NMpro+36) were grown in TSB, and total RNA was purified. Transcript levels of the proline synthesis genes in the purified RNA were measured by RT-PCR with 17 cycles of amplification. 16S RNA was used as a control. (B) The wild-type strain (wt), the ccpA transposon mutant ΦΝΞ-7791 (7791), and the ΦΝΞ-7791 mutant carrying the vector pCL55 (pCL55) or the complementation plasmid pccpA (pccpA) were all grown in TSB. After total RNA was purified from the cells, transcript levels were analyzed by RT-PCR with 18 cycles of amplification. The numbers below the images represent the results of densitometry analysis normalized to the transcript level of the wild-type strain. N, no reverse transcriptase.

FIG. 6.

CcpA binding to CRE sequences. (A) CRE sequences in rocF and rocD. The CRE sequence of pckA was used to identify CRE in rocF and rocD. Identical nucleotides are indicated with a black dot. The numbers flanking the CRE sequence represent the distance from the start codon of the corresponding gene. (B) HPr phosphorylation by HPrK/P. His₆-HPr (6 μM) was mixed with His₆-HPrK/P (2 μM) in the presence of [γ-³²P]ATP (5 μCi), and the reaction products were analyzed by SDS-PAGE and autoradiography. (C) EMSA analysis of CcpA binding to CRE. The 50-bp double-stranded DNAs containing CREs were mixed with various concentrations of His₆-CcpA in either the presence or absence of HPr∼P. As a negative control, the 50-bp DNA sequence upstream of proC was used. The DNA-protein complex was analyzed by PAGE and autoradiography. Lanes 1 and 6, 0.0625 μM; lanes 2 and 7, 0.125 μM; lanes 3 and 8, 0.25 μM; lanes 4 and 9, 0.5 μM; lanes 5 and 10, 1 μM. (D) Gel-filtration analysis of CcpA binding to rocF DNA. The mixture of His₆-CcpA (5.8 μM), HPr (8.3 μM), and His₆-HPrK/P (1.1 μM) was incubated at room temperature in either the absence (upper panel) or presence (lower panel) of ATP (1 mM). After addition of the rocF DNA and incubation at room temperature, the protein-DNA complexes were applied to an FPLC gel-filtration column and eluted at 1 ml/min. Gel-filtration chromatograms are shown, along with the SDS-PAGE analysis of the four peaks (A, B, C, and D). The control proteins (i.e., CcpA and HPrK/P for the upper panel and CcpA and HPrK/P and HPr for the lower panel) are also shown. The gels were stained with Coomassie dye.

FIG. 7.

Carbon source effects on proline synthesis. The wild-type strain Newman was grown in CDM or CDM-P with each of the indicated carbohydrates as the sole carbon source at 37°C for 16 h. Then, the proline synthesis was measured by the cell growth in CDM-P. Error bars show standard deviations.

FIG. 8.

CcpA's contribution to staphylococcal virulence in the murine abscess formation model. The wild-type strain Newman, the ccpA transposon mutant ΦΝΞ-7791, and the ΦΝΞ-7791 strain carrying either pCL55 or pccpA (10⁷ CFU) were administered to 10 BALB/c mice via retro-orbital injection. Four days later, the mice were sacrificed and bacterial loads in liver and kidneys were measured by counting CFU on TSA plates. Horizontal bars indicate the means of all observations. Bacterial loads were compared with those in mice infected with the wild type (NM). Statistical significance was measured by one-way ANOVA with Bonferroni's posttest. NM, Newman strain; 7791, ΦΝΞ-7791 strain; **, P < 0.05; n.s., not significant.