H-NS suppresses pilE intragenic transcription and antigenic variation in Neisseria gonorrhoeae

Abstract

Initially, pilE transcription in Neisseria gonorrhoeae appeared to be complicated, yet it was eventually simplified into a model where integration host factor activates a single -35/ -10 promoter. However, with the advent of high-throughput RNA sequencing, numerous small pil-specific RNAs (sense as well as antisense) have been identified at the pilE locus as well as at various pilS loci. Using a combination of in vitro transcription, site-directed mutagenesis, Northern analysis and quantitative reverse transcriptase PCR (qRT-PCR) analysis, we have identified three additional non-canonical promoter elements within the pilE gene; two are located within the midgene region (one sense and one antisense), with the third, an antisense promoter, located immediately downstream of the pilE ORF. Using strand-specific qRT-PCR analysis, an inverse correlation exists between the level of antisense expression and the amount of sense message. By their nature, promoter sequences tend to be AT-rich. In Escherichia coli, the small DNA-binding protein H-NS binds to AT-rich sequences and inhibits intragenic transcription. In N. gonorrhoeae hns mutants, pilE antisense transcription was increased twofold, with a concomitant decrease in sense transcript levels. However, most noticeably in these mutants, the absence of H-NS protein caused pilE/pilS recombination to increase dramatically when compared with WT values. Consequently, H-NS protein suppresses pilE intragenic transcription as well as antigenic variation through the pilE/pilS recombination system.

Fig. 1.

Northern analysis of pilE antisense RNA. Total RNA and size-fractionated sRNA were isolated from N. gonorrhoeae strain MS11. The Northern blot was then probed with an oligonucleotide designed to hybridize with pilE-specific antisense RNA.

Fig. 2.

Identification of the pilE midgene sense and antisense promoters. (a) Putative promoter elements were initially identified by the bprom and Neural Network online prediction tools. (b) Mutated promoter sequences. (c) Expression levels of WT and mutated constructs through strand-specific qRT-PCR analysis using in vitro-transcribed RNA. Strand-specific cDNA was made utilizing primers 09305 and 09264 for antisense and sense, respectively. WT expression for both antisense and sense expression was normalized to 100 %. Data represent mean ± sd; n = 4.

Fig. 3.

clustal analysis of Neisseria pil gene sequences. The accession numbers for each strain are as follows: Neisseria gonorrhoeae MS11 GenBank accession CP003909.1 (pilE = NGFG_01821; pilS1 copy 1 = NGFG_02253 (2122534–2123035), copy 2 = NGFG_02253 (2123072–2123479), copy 3 = NGFG_02487, copy 4 = NGFG_00014 (2124309–2124659), copy 5 = NGFG_00014 (2124696–2125043); pilS2 copy 1 = NGFG_02484, copy 2 = NGFG_02485; pilS5 = NGFG_02405; pilS6 copy 1 = NGFG_02481, copy 3 = NGFG_02482; pilS7 = NGFG_02431); Neisseria gonorrhoeae FA1090 GenBank accession NC_002946.2 (pilE = 2038021–2037383*; pilS1 copy 1 = 2043862–2044354*, copy 2 = 2044392–2044736*, copy 3 = 2044773–2045127*, copy 4 = 2045506–2045842*, copy 5 = 2045843–2046256*; pilS2 copy 1 = 2040723–2041203*, copy 2 = 2041256–2041603*, copy 3 = 2041646–2042026*, copy 4 = 2042470–2042800*, copy 5 = 2042801–2043147*, copy 6 = 2043148–2043522*; pilS3 copy 1 = 2013715–2014201*, copy2 = 2014238–2014567*, copy 3 = 2014626–2015018*; pilS6 NGO2041a copy 1 = 2015343–2015868, copy 2 = 2016302–2016669, copy 3 = 2016707–2017060; pilS7 = 1480693–1481248*); Neisseria gonorrhoeae NCCP 11945 GenBank accession NC_011035.1 [pilE = NGK_RS08820, pilS1 = NGK_RS09875 (1818155–1818627), pilS2 = NGK_RS09875 (1818628–1819032), pilS3 = NGK_RS09875 (1819033–1819327)]; Neisseria meningitidis FAM18 GenBank accession NC_008767.1 (pilE = NMC_RS01135, pilS1 = NMC_RS00015, pilS2 = NMC_RS00020); Neisseria meningitidis 053442 GenBank accession NC_010120.1 (pilE = NMCC_RS00970; pilS1 = NMCC_RS00130; pilS2 = NMCC_RS00125 copy 1 = 19974–20401, copy 2 = 19532–19973 ; pilS3 = NMCC_RS03380; pilS4 = NMCC_RS07650; pilS5 = NMCC_RS07655; pilS6 = NMCC_RS08075; pilS7 = NMCC_RS08080); Neisseria meningitidis MC58 GenBank accession NC_003112.2 (pilE = NMB0018; pilS1 = NMB0019; pilS2 = NMB0020; pilS3 = NMB0021; pilS4 = NMB0022; pilS5 = NMB0023; pilS6 = NMB0024; pilS7 = NMB0025 copy 1 = 23169–23502, copy 2 = 22825–23168; pilS8 = NMB0026); *Not annotated in final assembly. The midgene antisense promoter is indicated by the left arrow; the midgene sense promoter is indicated by the right arrow. The colour-shaded regions highlight the locations of the antisense and sense promoters within the sequences.

Fig. 4.

Identification of the antisense promoter at the 3′ end of pilE. (a) Location of the predicted antisense promoter at the 3′ end of pilE. (b) 3′ Mutated promoter. (c) Expression of pilE antisense RNAs from recombinant WT and the 3′ mutated constructs in E. coli using strand-specific qRT-PCRs. WT expression is normalized to 100 %. Error bars represent mean ± sd; n = 10.

Fig. 5.

Strand-specific qRT-PCR analyses of in vivo-transcribed gonococcal RNAs. Total RNA was isolated from gonococci carrying either a non-mutated pilE gene, a pilE midgene antisense promoter mutation, a pilE 3′ antisense mutation, or a pilE midgene antisense promoter mutation as well as a pilE 3′ antisense mutation. Strand-specific cDNAs were synthesized using primer 09172 for antisense expression and 09085 for sense expression. (a) AT-richness of the various promoter sequences and surrounding DNA. The promoter is underlined. (b, c) Expression of pilE antisense (b) and sense (c) RNAs in all four GC constructs. The left side of each panel reflects RNAs extracted from WT bacteria; the right side of each panel reflects RNAs extracted from an isogenic hns mutant. Expression is normalized relative to the WT value, which is set to 1. Error bars represent mean ± sd; n = 10.

Fig. 6.

pilE RNA expression over time. (a) Northern blot of total RNA samples isolated from liquid-grown cells at the indicated time points. The total RNAs from each time point were quantified by spectrophotometric analysis and equal amounts of RNA were applied to the gel. The blot was probed with oligonucleotide probe 245, which recognizes pilE sense RNA at the 5′ end (Wachter et al., 2015). The arrow indicates the full-length pilE message. (b) Antisense : sense ratio of pilE RNA. The strand-specific C_t values were determined through qRT-PCR. The initial strand-specific cDNA was made using primer 09172 for antisense expression and primer 09085 for sense expression. As a C_t value is inversely related to the extent of expression, the higher the ratio that is observed, the more sense expression is apparent when compared with antisense expression. Error bars represent mean ± sd; n = 4.

Fig. 7.

pilE/pilS recombination assay. (a) pilE/pilS recombination assessed at the DNA level. Identical quantities of chromosomal DNA were used to assess pilE/pilS5 recombinants using qRT-PCR. (b) pilE/pilS recombination assessed at the RNA level. Identical amounts of total RNA were used to assess pilE/pilS5 recombinant cDNAs levels using qRT-PCR. The left grouping reflects a WT genetic background (hns⁺); the right grouping reflects an hns⁻ genetic background. The primers that were used for the recombination assay were the pilE-specific primer (RC3) and the pilS5-specific primer (pilS5c1). The relative level of recombinant RNAs/DNAs was initially standardized against a recA control. Each group was then standardized against WT pilE (hns⁺), which was set to 1. The data reflect mean ± sd; n = 3.

Fig. 8.

Preferential inhibition of transcription of recombinant genes by H-NS protein. The overall extent of pilE transcription was determined by qRT-PCR analysis using primers RC3/173, which specifically amplify all pilE transcripts in a total RNA sample. The pilE recombinant scores were determined using the primers that were used for the recombination assay, which were the pilE-specific primer (RC3) and the pilS5-specific primer (pilS5c1). The data represent the ratios of the recombinant pilE C_t scores to the overall pilE C_t score. The left grouping reflects a WT genetic background (hns⁺); the right grouping reflects an hns⁻ genetic background. Each group was also standardized against WT pilE (hns⁺), which was set to 1. The data reflect mean ± sd; n = 3.