|Home | About | Journals | Submit | Contact Us | Français|
Neisseria gonorrhoeae produces a type IV secretion system that secretes chromosomal DNA. The secreted DNA is active in the transformation of other gonococci in the population and may act to transfer antibiotic resistance genes and variant alleles for surface antigens, as well as other genes. We observed that gonococcal variants that produced type IV pili secreted more DNA than variants that were nonpiliated, suggesting that the process may be regulated. Using microarray analysis, we found that a piliated strain showed increased expression of the gene for the putative type IV secretion coupling protein TraD, whereas a nonpiliated variant showed increased expression of genes for transcriptional and translational machinery, consistent with its higher growth rate compared to that of the piliated strain. These results suggested that type IV secretion might be controlled by either traD expression or growth rate. A mutant with a deletion in traD was found to be deficient in DNA secretion. Further mutation and complementation analysis indicated that traD is transcriptionally and translationally coupled to traI, which encodes the type IV secretion relaxase. We were able to increase DNA secretion in a nonpiliated strain by inserting a gene cassette with a strong promoter to drive the expression of the putative operon containing traI and traD. Together, these data suggest a model in which the type IV secretion system apparatus is made constitutively, while its activity is controlled through regulation of traD and traI.
Neisseria gonorrhoeae (the gonococcus) is a strictly human pathogen and the causative agent of the sexually transmitted disease gonorrhea. N. gonorrhoeae has an extraordinary capacity to undergo antigenic variation of surface-exposed structures, such as lipooligosaccharide, outer membrane proteins, pilin, and porin (14, 17, 24, 25, 41). Furthermore, it undergoes natural transformation at high frequency and is unusual in that it is continuously competent for transformation (45). Competence contributes to the acquisition of new alleles for antigenic variation and the acquisition of antibiotic resistance genes (reviewed in reference 20). The rapid spread of antibiotic resistance and the appearance of multidrug resistance among clinical isolates makes the effective treatment of gonorrhea more difficult (52). Our studies have demonstrated that ~80% of gonococcal strains carry the genes for a type IV secretion system (T4SS) (11, 21). The T4SS transports chromosomal DNA into the extracellular milieu, where the secreted DNA is effective in transforming other gonococci in the population (11, 22, 39). Thus, type IV secretion functions as a mechanism of DNA donation for natural transformation.
T4SSs are ubiquitous within the microbial world (32). T4SSs are present in a number of bacterial pathogens (e.g., Bordetella pertussis, Legionella pneumophila, and Helicobacter pylori) and transport proteins or DNA and proteins (e.g., Agrobacterium tumefaciens) that are required for pathogenesis (1). Conjugation systems are T4SSs that transfer plasmid or mobilizable DNA (encoding drug resistance, virulence factors, and/or metabolic processes) directly from one bacterium into another (7). Therefore, type IV secretion is an important mechanism of pathogenesis and horizontal gene exchange.
The expression and regulation of type IV secretion genes have been shown to involve two-component systems, such as LuxR/LuxI regulators in A. tumefaciens (15); host factors, including IHF (integration host factor) (9, 16), H-NS (nucleoid-associated protein) (47, 54), and Hfq (host RNA chaperone) (55); and type IV secretion components, including F-plasmid TraJ (primary activator), TraY (secondary regulator), TraM (relaxosomal protein), and TraI (relaxase) (13, 19). The gonococcal T4SS is an F-like system but does not encode homologues of TraJ, TraY, and TraM, which are the primary activators of F-plasmid conjugation (13, 21). Type IV secretion of DNA in N. gonorrhoeae has been shown to occur during growth in log phase in liquid culture (11, 21), but nothing is known regarding what controls DNA secretion in N. gonorrhoeae, whether it also occurs in other stages of growth, and whether it occurs during a short window of time or continuously. Since producing the T4SS and secreting chromosomal DNA likely incur a significant energy expense for the gonococcal cell, it would seem advantageous for the bacterium to regulate this process.
In this study, we show that DNA secretion is increased in gonococcal variants that produce type IV pili. Despite the similarity in their names, the gonococcal type IV pilus is unrelated to the T4SS. Gonococcal type IV pili mediate attachment to host cells (49), autoagglutination (51), and twitching motility (50), and the pilus machinery acts in the uptake of DNA for natural transformation (3). Production of type IV pili is a variable phenotype that changes at high frequency. Although there are several mechanisms for revertable pilus phase variation (43), gonococci can also become irreversibly nonpiliated by deletion of the 5′ end of the pilin gene (2). In this study, we examine differences in type IV secretion in piliated (P+) strains and their nonpiliated (P−), nonreverting derivatives.
Time course analysis of DNA secretion by P+ and P− gonococci revealed that P+ gonococci secrete significantly more DNA than P− gonococci over time. Increased DNA secretion in P+ gonococci suggests that the piliation status of the cell either stimulates DNA secretion in P+ gonococci or downregulates DNA secretion in P− gonococci. A gene expression profile of P+ versus P− gonococci revealed an increase in traD transcript levels in P+ gonococci relative to the levels in P− gonococci. Mutational analysis suggested that traD is part of a larger transcript that also includes traI, yaa, and yaf. All of these genes are transcribed divergently from the rest of the T4SS genes. Therefore, the microarray data suggest that the expression levels of these genes might determine DNA secretion levels in N. gonorrhoeae. We found that while increasing the expression of yaf, traI, traD, and yaa had no effect on DNA secretion in P+ gonococci, it resulted in a 2-fold increase in DNA secretion in P− gonococci. However, this increase was not enough to reach P+ levels of DNA secretion, suggesting that other components may also act in the regulation of type IV secretion in N. gonorrhoeae.
The N. gonorrhoeae strain MS11 was used as the wild-type (WT) strain and for the construction of the gonococcal strains described in Table Table1.1. Growth of gonococci on GC medium base (GCB) agar (Difco) was performed as previously described (10). Escherichia coli strains were grown on Luria-Bertani (LB) agar plates or in LB broth (40). Graver-Wade (GW) defined medium (pH 6.8) was used for the growth of gonococci in liquid culture for DNA secretion assays and growth rate determinations (53). For microarray experiments, gonococci were grown in GCB liquid medium with supplements as described previously (10). For plasmid selection in E. coli, chloramphenicol was used at a concentration of 25 μg ml−1 and erythromycin was used at a concentration of 500 μg ml−1. For selection in gonococci, erythromycin was used at a concentration of 10 μg ml−1 and chloramphenicol was used at a concentration of 10 μg ml−1. Nonpiliated N. gonorrhoeae variants that arose spontaneously were screened for the presence of a deletion in pilE by PCR as described by Sechman et al. (42).
The DNA secretion assay was performed as described previously (39). Briefly, piliated (P+) gonococcal colonies were harvested from a GCB plate and grown for 2 h at 37°C with aeration in GW medium. The cultures were vortexed, and a 500-μl volume transferred to 2.4 ml of fresh GW medium. For nonpiliated (P−) variants, gonococci were harvested from a GCB plate and grown for 2 h at 37°C with aeration in GW medium. The cultures were vortexed and diluted to an optical density at 540 nm of 0.1 in a 3-ml culture. This method of dilution of the P+ and P− cultures resulted in equal protein amounts in the cell pellet at time zero. P+ and P− cultures were then grown for an additional 2.5 h. Supernatants were collected at the beginning (t = 0 h) and end (t = 2.5 h) of the second round of growth. DNA in culture supernatants was detected using the fluorescent DNA-binding dye PicoGreen (Invitrogen), and the amount normalized to the amount of total cell protein (Bradford assay). The results given below are the averages of results from at least four independent experiments. The average background fluorescence, determined by performing the secretion assay with the T4SS mutant [traI(Y93F)] N. gonorrhoeae strain (WSP5) (39), was subtracted from the average result for all the strains. For the time course analysis of secreted DNA, P+ and P− individual cultures were prepared for each time point (1.0, 1.5, 2.0, and 2.5 h) as described above. Supernatants were collected at the beginning (t = 0 h) and at each time point, and the amount of total cell protein was determined for all time points. The data are plotted as the change in the amount of DNA over time (DNA amount at a particular time point − DNA amount at 0 h) over the change in the amount of protein (protein amount at a particular time point − protein amount at 0 h). Statistical significance for the DNA secretion assay or protein assay was determined using the t test on the Simple Interactive Statistical Analysis (SISA) Website (http://www.quantitativeskills.com/sisa/).
DNA microarrays were made by amplification of 2,043 open reading frames (ORFs) of the N. gonorrhoeae genome, 1,985 from strain FA1090 (GenBank accession number AE004969) (8) and 58 from the gonococcal genetic island (GGI) of N. gonorrhoeae strain MS11 (11, 21, 22). The sizes of the amplicons spotted ranged from 143 bp to 3,485 bp in length and corresponded to the predicted ORF of each gene. An additional 362 amplicons corresponding to internal ORF sequences were generated for genes with insufficient DNA observed upon agarose gel electrophoresis of the initial PCR. PCR amplicons were spotted in duplicate onto SuperAmine glass slides (TeleChem International, Inc.) at the Michigan State University Research Technology Support Facility (MSU RTSF).
N. gonorrhoeae strains MS11 and MS11-307 (ΔpilE1::erm ΔpilE2) (35) were grown in GC base liquid medium with supplements (27) and 0.042% sodium bicarbonate in a humidified 5% CO2 environment. RNA was isolated from mid-log-phase cultures using Trizol reagent (Invitrogen). RNA was quantified spectrophotometrically, and quality assessed by agarose gel electrophoresis. RNA was labeled with either Cy3 or Cy5 dye using a CyScribe first-strand labeling kit (Amersham Biosciences), and the cDNA probes were purified using the Qiagen PCR cleanup kit (Qiagen). Probes were combined and concentrated to 15 μl using Microcon YM-30 centrifugal filtration units (Millipore).
DNA array slides were UV-cross-linked at 60 mJ of energy and then washed twice in 0.1% SDS for 3 min at room temperature to remove the unbound DNA. The slide was then washed in H2O for 2 min at room temperature and boiled in H2O for 3 min to denature the double-stranded DNA on the slide. The slide was plunged briefly into room temperature H2O to cool and dried by centrifugation for 1 min at 800 × g in a 50 ml-conical tube. Prehybridization was performed by incubating the slide with 60 μl of prewarmed prehybridization buffer (5× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.1% SDS, 1% bovine serum albumin) in a Corning hybridization chamber (Corning Life Sciences) at 42°C for 45 min. The labeled probes were denatured by boiling for 5 min along with 1 μl of 1 μg ml−1 sheared, sonicated herring sperm DNA and chilled on ice. DNA was then mixed with 15 μl of 4× hybridization buffer (Amersham) and 30 μl of formamide, pipetted onto the prehybridized array slide, assembled in the hybridization chamber, and hybridized at 42°C overnight. Following hybridization, the coverslip was removed and the slide was washed as follows: once in 1× SSC, 0.2% SDS at 42°C; twice in 0.1× SSC, 0.2% SDS at room temperature; and once in 0.1× SSC. After the last wash, the slide was dried by centrifugation and scanned using a GenePix 4000B scanner (Axon Instruments). Images were processed and analyzed using GenePix 4.0 software. Data normalization and analysis were done as described previously (12).
Oligonucleotide primers for real-time PCR were designed using the Primer Express 1.5 program from Applied Biosystems. The primers used were as follows: for traD, traD-F (GCGCGAAAACATGAGATTGA) and traD-R (CCATGCCGATTTCCGAGTTA); for traI, traI-F (GGCTTCCCGCCAGTGAG) and traI-R (CATCGAGTGTATGGCGGACTAA); and for rpoB, rpoB-F (TGCCGTACATGGCGGAC) and rpoB-R (ATACGGGAAGGTACGCCCA). Piliated and nonpiliated N. gonorrhoeae strains were grown in GC base liquid medium with Kellogg's supplements and 0.042% sodium bicarbonate in log-phase culture (27). Trizol reagent was used to isolate RNA from 2 ml of culture at approximately 5 × 108 CFU/ml. The RNA from each culture was suspended in 50 μl of H2O with 2 μl of Promega recombinant RNasin RNase inhibitor. Possible contaminating DNA was removed by digestion using 6 units of Promega RQ1 RNase-free DNase and 4 units of Invitrogen DNase I. Digestion was allowed to proceed for 30 min at room temperature. RNA was further purified using a Qiagen RNeasy midi kit according to the RNA cleanup protocol. Real-time reverse transcription (RT)-PCRs were performed using an Applied Biosystems power SYBR green RNA-to-Ct 1-step kit. Reaction mixtures were prepared in a 25-μl volume and run in triplicate for each gene. N. gonorrhoeae strain MS11 chromosomal DNA was used to establish standard curves. All reactions were performed using an Applied Biosystems ABI Prism 7700 sequence detector. Three biological replicates were performed.
Plasmids used for mutagenesis and complementation are described in Table Table1.1. The traD deletion removes bases 4 to 1,978, 97% of the coding sequence, but it maintains the predicted ribosome-binding site and start codon for the downstream gene yaa and is thus predicted not to have polar effects on yaa. The traD deletion or the traI frameshift mutation was introduced into gonococci by allelic exchange without selection (10). Complementation was achieved by introduction of the wild-type copy of traD or traI into the gonococcal chromosome between aspC and lctP along with the cat marker via the pKH37 vector (29). An ermC cassette with a strong promoter (22) was inserted upstream of the yaf, traI, traD, and yaa promoter region to increase transcription of these genes.
Our previous studies using piliated gonococcal strains (39) showed more DNA secretion than our studies that used nonpiliated variants (21, 22). To determine if the piliation status of the bacteria has an effect on the amount of DNA secreted into the culture medium, we directly compared a piliated strain and a nonpiliated variant using the fluorometric DNA secretion assay. We used a time course so that we could determine if there is a secretion window, i.e., a short period of time during which more DNA secretion occurs, analogous to the competence window seen for bacteria that regulate DNA uptake (36, 46). For piliated-strain measurements, DNA secretion was assessed in WT P+ gonococci and in a DNA secretion-deficient P+ strain [traI(Y93F) mutant] (39) over a 2.5-h period of growth in log phase (Fig. (Fig.1A).1A). The amount of DNA or protein measured at each time point was subtracted from the DNA or protein measured at the beginning of the time course (t = 0 h), respectively. Therefore, the values represent the change in DNA in the medium over the change in protein in the cell pellet (ΔDNA/Δprotein). The amount of DNA detected in the medium from the WT P+ strain was significantly different from that in the medium from the traI(Y93F) mutant P+ strain at 1 h and increased steadily over time up to 2 h, where there was a sharp increase in the amount of DNA measured in the culture medium (Fig. (Fig.1A).1A). Conversely, the DNA measured in the culture medium of the traI(Y93F) mutant P+ strain remained at the same level over time, indicating that there was no active secretion of DNA in this strain (Fig. (Fig.1A1A).
To evaluate DNA secretion by nonpiliated gonococci during growth in log phase, we isolated nonpiliated variants from WT P+ and traI(Y93F) mutant P+ strains. These variants carried a deletion of the pilE promoter region and 5′ coding sequence and thus were unable to make pilin. As opposed to WT P+ gonococci, where the amount of DNA measured in the culture medium was significantly different from that measured for traI(Y93F) mutant P+ gonococci at 1 h, DNA secretion by WT P− gonococci was not significant until 2 h of growth and increased thereafter (Fig. (Fig.1B).1B). To directly compare DNA secretion by P+ and P− gonococci over time, background fluorescence values from traI(Y93F) mutant P+ or P− gonococci were subtracted from those of the wild-type P+ or P− strains. WT P+ cultures exhibited significantly more secreted DNA than WT P− cultures at each time point, with the greatest fold differences observed at 1 h (19-fold higher in P+ cultures), 1.5 h (4.7-fold higher), and 2.5 h (3.7-fold higher) (Fig. (Fig.1C).1C). The smallest difference in DNA secretion between WT P+ and P− cultures was observed at 2 h of growth, where WT P+ cultures contained 1.7 times more secreted DNA than P− cultures. Therefore, these results demonstrate that P+ gonococci secrete more DNA than P− gonococci during growth in log phase.
We sought to evaluate differences in gene expression in P+ versus P− gonococci that could explain the marked differences in DNA secretion. A gonococcal DNA microarray was used to compare gene expression between a piliated gonococcal strain (MS11 P+) and a nonpiliated strain (MS11-307 P−). MS11-307 cannot produce type IV pili due to deletions in the pilin expression locus, pilE (35). Strains were grown in GC base liquid medium, and total RNA was isolated and labeled with Cy3 or Cy5 by reverse transcription. The labeled cDNAs were hybridized to the gonococcal DNA microarray slide as described in Materials and Methods. A total of three independent experiments (including one dye swap) were done, and data normalization and analysis were performed as described previously (12). Outliers were identified as those with expression ratios (log2 transformed) greater than 2.5 standard deviations from the mean and are listed in Tables Tables22 to to44.
The greatest difference in hybridization signal was observed with pilE and some pilS sequences, which was expected as pilE is deleted in the P− strain and pilS loci share some sequence homology with pilE (data not shown). bfrA and bfrB showed the greatest increases in expression in P+ gonococci relative to their levels in P− gonococci, with ratios of 3.8 and 4.0, respectively (Table (Table2).2). These two gene products comprise bacterioferritin, an important iron storage protein for bacteria. traD, encoding a predicted type IV secretion coupling protein, showed the third-largest increase in expression in P+ gonococci, with a ratio of 3.44, and it was the only gene of the GGI transfer region included in the array that was increased in expression. ydhB, encoding a hypothetical protein, was the other GGI gene with increased expression in P+ gonococci (2.36-fold increase). The increased expression of several other genes in P+ gonococci is also worth mentioning. One such gene is ppiB, which encodes a putative rotamase, an enzyme involved in protein folding (26). The expression of this gene was increased 2.82-fold in P+ gonococci relative to its level in P− gonococci. Additionally, the expression levels of himA and recN were slightly increased in P+ gonococci. himA encodes a subunit of IHF (integration host factor), a site-specific DNA-binding protein with roles in DNA processing for type IV secretion and DNA recombination, transcription, and replication (16, 23, 37, 38, 56). recN encodes a DNA repair protein that is required for both DNA repair and DNA transformation in N. gonorrhoeae (44).
Compared to the number of genes whose expression was increased in P+ gonococci, fewer genes were identified as significantly increased in expression in P− gonococci (Tables (Tables33 and and4).4). The expression levels of several genes were consistently increased ~1.5-fold in P− gonococci, and most of these genes are involved in transcription and translation. These changes in gene expression may explain the increased growth rate that is apparent in P− versus P+ cultures in liquid medium, as was previously noted by Swanson (51). It is difficult to assess whether P− gonococci truly grow faster than P+ gonococci, since P+ gonococci show a high degree of clumping in liquid culture, making measurements of optical density or colony counts inaccurate. Therefore, we compared the growth of P+ and P− gonococci in liquid culture by measuring the protein content of cell pellets over time. We found that a P− strain does indeed grow faster than a P+ strain (Fig. (Fig.2).2). While in P+ gonococci, total protein in the cell pellet increased from ~14 mg ml−1 at 0 h to ~38 mg ml−1 at 2.5 h (a 2.7-fold increase), in P− gonococci, total cell protein increased from ~10 mg ml−1 to 45 mg ml−1 in the same time period (a 4.5-fold increase) (Fig. (Fig.2).2). This increased growth rate is consistent with the higher expression levels observed for many of the genes involved in transcription and translation.
yaf, traI, traD, and yaa (located in the GGI transfer region) are transcribed divergently from the rest of the T4SS genes (Fig. (Fig.3A).3A). Sequence analysis suggests that these genes may be part of the same transcript. traD encodes a putative coupling protein, traI encodes a relaxase homologue, and yaf and yaa are small genes that encode hypothetical proteins. traI is located upstream of traD and is required for type IV secretion (39). Unfortunately, the microarray data for traI was not of sufficient quality to determine differential expression. Since traD is increased in expression in piliated gonococci, the levels of this putative transcript may control DNA secretion. In characterized type IV secretion systems, the relaxase and the coupling protein are key players in the processing and recruitment of the DNA substrate (7). The relaxase cleaves the DNA and serves as a pilot protein for secretion (30), and the coupling protein recognizes the DNA substrate and brings it to the apparatus for secretion (4, 5, 18).
To evaluate the validity of the microarray results for traD, we used real-time RT-PCR to examine traD expression. Similarly, we examined traI expression by real-time RT-PCR to determine if it might show the same expression pattern as traD. Total RNA was prepared from log-phase cultures of WT piliated or nonpiliated strains. The levels of traI or traD transcripts were normalized to those of the gene for RNA polymerase, rpoB. The piliated strain consistently showed increased traD levels, with an average ratio of 1.97 (±0.11)-fold (mean ± standard deviation) as much traD message as in the nonpiliated variant. Although this ratio is not as great as was seen in the microarray, these data confirm that traD transcript levels are increased in piliated cells. Similarly, the transcript levels for traI, the gene adjacent to traD, were increased 2.08 (±0.25)-fold in the piliated strain relative to their levels in the nonpiliated variant. The fold increases for the traD and traI transcripts were not different from each other as determined by t test (P = 0.5). These data suggest that the expression of traI and traD is regulated similarly and that increased expression of these genes may lead to increased DNA secretion in piliated gonococci.
The results of mutation and complementation analyses suggest that traI and traD are on the same transcript, as described below. A point mutation was introduced into traI that results in a frameshift and a premature stop codon 4 amino acid residues from the start. We tested this strain in a DNA secretion assay, and as expected from previous results with traI mutants (39), it is deficient in DNA secretion (Fig. (Fig.3B).3B). However, complementation of this strain with a wild-type copy of traI at a distant site on the chromosome did not restore DNA secretion, while complementation restored DNA secretion in a strain carrying the nonpolar traI(Y93F) mutation (Fig. (Fig.3B)3B) (39).
Lack of wild-type traI complementation in the strain carrying the traI frameshift mutation suggests that the mutation has a polar effect on downstream genes that are required for secretion. The gene just downstream of traI is traD, which encodes the putative coupling protein. Since the start codon of traD overlaps the stop codon of traI, translation of traD may depend on translation of the complete traI sequence. However, other mutations in traD had been found to cause a slight reduction or no reduction in DNA secretion, suggesting that traD might not be required for DNA secretion (21). We constructed a traD deletion strain in N. gonorrhoeae and found that this strain is deficient in DNA secretion (Fig. (Fig.3B).3B). Complementation with wild-type traD restored DNA secretion to wild-type levels. Altogether, the requirement for traD for DNA secretion and the inability of traI to restore DNA secretion in the strain carrying a traI frameshift suggest that traD and traI are transcriptionally and translationally coupled.
To determine if traD and traI levels control DNA secretion in N. gonorrhoeae, we constructed a strain carrying an ermC cassette with a strong promoter (22) upstream of the yaf, traI, traD, and yaa promoter region to increase transcription from their native site (Fig. (Fig.4A)4A) (39). We also created strains that overexpressed traI or traD from a distant site on the chromosome (Fig. (Fig.4B).4B). We then isolated nonpiliated, nonreverting variants of these strains and tested all of these strains for DNA secretion. Overexpression of traI or traD alone had no effect on DNA secretion in either P+ or P− gonococci (Fig. 4C and D). Interestingly, P− gonococci with traI and traD driven by the ermC promoter exhibited a 2-fold increase in DNA secretion, while the same construct had no effect on DNA secretion in P+ gonococci (Fig. 4C and D). However, the increase in DNA secretion observed in P− gonococci falls short of reaching the levels of DNA secretion of P+ gonococci. To determine if the failure to reach levels of DNA secretion equivalent to those of P+ gonococci was due to insufficient transcription, we used real-time RT-PCR to measure traD transcript levels in the strains overexpressing traI and traD from the native site driven by the ermC promoter or traD from the distant site driven from the lac promoter. Transcript levels for traD in the ermC construct strain WSP38 were 4.0 (±1.5)-fold the levels measured in the WT nonpiliated strain, whereas transcript levels for traD in the lac construct strain KL503 were 123 (±55)-fold the levels of the WT nonpiliated strain. These values were significantly different from the levels in the WT nonpiliated strain (P < 0.05). Thus, insufficient transcript does not explain either the inability of the ermC construct to restore the P− strain to WT levels of DNA secretion or the inability of the traD overexpression strain to show increased DNA secretion. These data suggest that, in addition to traD and traI levels, other factors are also involved in the regulation of type IV secretion in N. gonorrhoeae.
We have shown that gonococcal variants that produce type IV pili secrete more DNA than variants that lack type IV pili. A time course analysis showed more DNA in the culture medium of piliated strains than in that of nonpiliated strains at all time points. However, piliated strains showed the largest increase in DNA secretion in the later stage of log-phase growth (Fig. (Fig.1).1). Previous studies have demonstrated that this DNA release is not through autolysis, in that the kinetics of DNA release are different from those of cytoplasmic protein release (11) or RNA release (29), which occur by autolysis. Furthermore, the DNA in the medium at these time points was shown to be single stranded (39). However, after the secretion window between the 2-h and 2.5-h time points, the bacteria begin to lyse. This secretion window just prior to the onset of the stationary/death phase is reminiscent of the competence window seen in other naturally transformable bacteria. Species like Haemophilus influenzae and Streptococcus pneumoniae express competence proteins for a short period just prior to the stationary phase or during growth limitation (36, 46). This phenomenon has been proposed to be part of the catastrophe kit, as a mechanism that would increase genetic reassortment and diversity in the population and potentially produce a genetic variant that would survive the conditions that caused the growth reduction (28). Since competence is not regulated in Neisseria gonorrhoeae (45), it might be useful for gonococci to regulate DNA donation and increase transformation during periods of growth reduction.
The results of microarray analysis suggest that increased DNA secretion in piliated variants may be due to transcriptional regulation. Piliated strains showed increased transcription of traD, the gene encoding the T4SS coupling protein. traD appears to be part of an operon that includes traI (encoding the relaxase) and yaf and yaa (genes of unknown function) (21). These results suggest a model in which the T4SS apparatus may be constitutively expressed and secretion could be activated by the production of the coupling protein and the relaxase. While increased expression of traD in P+ gonococci was of interest to us, there was no evidence to suggest that the product of this gene was required for DNA secretion. Strains containing an insertion interrupting the C-terminal coding region of traD showed intermediate levels of DNA secretion, and a deletion removing one of the two Walker A box sequences had no effect on DNA secretion (21). To determine if traD was required for DNA secretion, we constructed a strain where traD was almost completely deleted. The traD deletion strain was found to be deficient in DNA secretion, and complementation with wild-type traD restored secretion. These results provide evidence that the putative coupling protein is involved in DNA secretion in N. gonorrhoeae.
T4SSs are encoded by a single polycistronic transfer region in most bacteria, while in others they are encoded as split loci (6). The genes in the transfer region usually have overlapping stop-and-start codons, which allows the expression of a large set of genes from a few promoters and regulation of both transcription and translation. yaf, traI, traD, and yaa also have overlapping stop-and-start codons, and a single putative promoter upstream of yaf was found by sequence analysis of the region. The sequence arrangement of these genes together with the presence of one promoter suggests that they are part of the same transcript. To evaluate this hypothesis, we constructed a strain carrying a traI frameshift mutation very close to the 5′ end of the coding sequence. As expected, this strain showed a deficiency in DNA secretion. However, what is more relevant is that complementation with wild-type traI did not restore DNA secretion. This is in contrast to the restored levels of DNA secretion obtained by complementation of a strain carrying a traI point mutation with wild-type traI (previously described in reference 39). These results suggest that traD and traI are both transcriptionally and translationally coupled.
If the expression levels of traI and traD determine the amount of DNA secretion, P− gonococci could be induced to secrete higher levels of DNA by increasing the expression of these genes. To test this hypothesis, we introduced an ermC cassette with a strong promoter upstream of the putative promoter for traI and traD. This construct had been previously used (and shown) to overexpress traI (39) and for driving the expression of essential genes downstream of insertion mutations (34). Increased expression of traI and traD had no effect on DNA secretion in P+ gonococci. However, DNA secretion was increased by ~2-fold in P− gonococci containing this construct. This increased DNA secretion in P− gonococci did not reach the level of DNA secretion seen in wild-type P+ gonococci, even though transcript levels were found to be 4 times as great as those of the nonpiliated strain and nearly twice those of the piliated strain. These data indicate that additional mechanisms act in the regulation of DNA secretion.
It is not clear why piliation should affect type IV secretion. One possibility that could explain these results is if T4SS gene expression and pilin expression were regulated by the same factors. However, it is unclear if, or how, pilin transcription is regulated, and the sigma factor that regulates type IV pilin expression in other bacteria (sigma 54) is only represented with a pseudogene in N. gonorrhoeae (31). Another possibility is that gonococci respond to signals released by other piliated gonococci that indicate that they are competent for genetic transformation. Most piliated gonococcal variants produce significant amounts of secreted pilin monomers, known as S-pilin (33). It is possible that gonococci could sense S-pilin as a method for identifying piliated gonococci nearby. N. gonorrhoeae may also be able to sense physiological conditions, such as growth rate. The F-plasmid tra operon was recently found to be repressed by the host nucleoid-associated protein, H-NS (54). H-NS protein levels remain constant during growth, and changes in gene expression are driven by nutritional and environmental signals that cause changes in chromosomal supercoiling, thereby affecting H-NS binding (54). Similarly, in N. gonorrhoeae, a higher growth rate (typical of nonpiliated gonococci) could be a physiological signal that causes nucleoid-associated protein repression of type IV secretion gene expression.
This work was supported by NIH grant R01 AI047958 to J.P.D. and NIH grant R21 AI064292 to C.G.A. W.S.-P. received support through traineeship on NIH grant T32 AI055397.
We thank Chris van der Does for pointing out the existence of the yaa open reading frame.
Published ahead of print on 5 February 2010.