Gonococci undergo frequent and efficient natural transformation. Transformation occurs so often that the population structure is panmictic, with only one long-lived clone having been identified. This high degree of genetic exchange is likely necessary to generate antigenic diversity and allow the persistence of gonococcal infection within the human population. In addition to spreading different alleles of genes for surface markers and allowing avoidance of the immune response, transformation facilitates the spread of antibiotic resistance markers, a continuing problem for treatment of gonococcal infections. Transforming DNA is donated by neighbouring gonococci by two different mechanisms: autolysis or type IV secretion. All types of DNA are bound non-specifically to the cell surface. However, for DNA uptake, Neisseria gonorrhoeae recognizes only DNA containing a 10-base sequence (GCCGTCTGAA) present frequently in the chromosome of neisserial species. Type IV pilus components and several pilus-associated proteins are necessary for gonococcal DNA uptake. Incoming DNA is subject to restriction, making establishment of replicating plasmids difficult but not greatly affecting chromosomal transformation. Processing and integration of transforming DNA into the chromosome involves enzymes required for homologous recombination. Recent research on DNA donation mechanisms and extensive work on type IV pilus biogenesis and recombination proteins have greatly improved our understanding of natural transformation in N. gonorrhoeae. The completion of the gonococcal genome sequence has facilitated the identification of additional transformation genes and provides insight into previous investigations of gonococcal transformation. Here we review these recent developments and address the implications of natural transformation in the evolution and pathogenesis N. gonorrhoeae.
Bacteria use three major mechanisms for genetic exchange: conjugation, transduction and transformation. Neisseria gonorrhoeae is one of at least 44 species of bacteria that are naturally competent for genetic transformation (Lorenz and Wackernagel, 1994; Sparling, 1966). Phages able to infect N. gonorrhoeae are not known (Goldberg et al., 1978) and, although conjugative plasmids are present in some gonococcal isolates, transformation is the only mechanism for mobilization of gonococcal chromosomal loci (Sox et al., 1978; Koomey, 1998).
Natural transformation is the ability of bacteria to take up and incorporate macromolecular DNA efficiently. It is hypothesized that natural transformation evolved as a mechanism for using DNA as a food source (Finkel and Kolter, 2001) or to aid in the repair of damaged chromosomes (Solomon and Grossman, 1996). For gonococci, it is clear that transformation has been harnessed as a powerful mechanism for generating genetic diversity, spreading advantageous alleles and mediating some forms of antigenic variation (Hobbs et al., 1994; Fudyk et al., 1999; Snyder et al., 2004).
Neisseria spp. are unusual in that they do not regulate competence like many other naturally competent bacteria, including Streptococcus pneumoniae, Bacillus subtilis and Haemophilus influenzae; rather, the Neisseriae are competent for natural transformation during all phases of growth (Biswas et al., 1977; Goodgal, 1982). The competence of N. gonorrhoeae is not dependent on any soluble competence factor (Biswas et al., 1977). Like H. influenzae, N. gonorrhoeae only takes up DNA that contains a genus-specific uptake sequence. In N. gonorrhoeae, the DNA uptake sequence (DUS) is a 10-base sequence (GCCGTCTGAA) that appears very frequently in its own chromosome and that of other Neisseria spp. The 2.15 Mb genome sequence of strain FA1090 contains 1965 copies of the DUS (Davidsen et al., 2004), thus averaging one DUS for every 1096 bp. The DUS sometimes occurs in inverted repeats between genes forming possible transcriptional terminators (Goodman and Scocca, 1988). Transformation of N. gonorrhoeae with DUS-bearing supercoiled or linear DNA is equally efficient. It was reported that the transformation efficiency of single-stranded DNA (ssDNA) is comparable to that of double-stranded DNA (dsDNA) (Stein, 1991). However, this idea is still controversial (see Remaining questions below).
This review will focus on the studies that have elucidated the four major steps in gonococcal transformation: donation of DNA, binding and uptake of transforming DNA, processing, and integration into the gonococcal chromosome. It should be noted that many of the mechanisms involved in N. gonorrhoeae transformation are the same in Neisseria meningitidis transformation. Therefore, where appropriate, insights derived from studies of N. meningitidis will also be described and significant differences will be mentioned.
Step 1. DNA donation
Most descriptions of natural transformation begin with DNA uptake. The DNA is simply presumed to be in the environment, and the donor is thought to have no role other than having made the supreme sacrifice. However, the donor may be an important player. What set of events has resulted in the presence of naked DNA in proximity to the competent recipient? In the case of N. gonorrhoeae, the bacteria only live inside human beings and only take up DNA containing the genus-specific DUS; therefore, the donated DNA must be from siblings of the infecting strain or from a co-infecting gonococcal strain. In rare instances the DNA might come from another species, including other Neisseria or H. influenzae (Kroll et al., 1998). Transfer from other Neisseria might be expected, as all Neisseria species use the same DUS. In fact, transfer from commensal Neisseria spp. appears to be responsible for the creation of penA alleles that confer penicillin resistance in N. gonorrhoeae (Spratt et al., 1992). Similarly, argF of N. meningitidis has regions that appear to have been acquired from commensal Neisseria (Zhou and Spratt, 1992). Transfer from H. influenzae would not be as likely as Haemophilus spp. use a different DUS. However, the H. influenzae chromosome was found to contain four copies of the Neisseria DUS, and one of these is thought to have facilitated the transfer of the bio locus from Haemophilus to N. meningitidis and N. gonorrhoeae (Kroll et al., 1998). Most species are thought to donate DNA for subsequent transformation by autolysis. However, N. gonorrhoeae can also donate DNA for transformation by type IV secretion.
Type IV secretion
Type IV secretion systems (T4SS) are one of more than five multicomponent secretion systems utilized by Gram-negative bacteria to secrete macromolecular substrates across the cell envelope. T4SS are noted for their ability to secrete both protein and DNA substrates, although not all systems secrete both. The presence of a T4SS encoded on the gonococcal genetic island (GGI) has recently been described. The GGI is present in approximately 80% of N. gonorrhoeae strains and appears to have been horizontally acquired (Dillard and Seifert, 2001). Many of the T4SS homologues encoded in the GGI have been shown to be necessary for the secretion of gonococcal chromosomal DNA into the extracellular milieu during growth (Hamilton et al., 2001; 2005). The DNA secreted by donors is taken up by recipient bacteria when donors and recipients are cultured together (Dillard and Seifert, 2001; Hamilton et al., 2001).
The DNA has been detected in culture medium by use of a fluorescent DNA-binding dye, as well as by co-culture transformation. The addition of DNase I to the culture supernatants eliminates the increased fluorescence seen in the wild-type culture medium as compared with that from the T4SS mutants (Hamilton et al., 2001; 2005), confirming that DNA is the material being detected. Also DNase I reduces transfer of chromosomal markers 250-fold, indicating that DNA is present in the medium and is not transferred by conjugation (Dillard and Seifert, 2001). Theoretically, the DNA could be released by autolysis of a portion of the population. However, all measures of autolysis have shown no differences between the wild-type strain and the T4SS mutants. Neither the release of a cytoplasmic protein (Dillard and Seifert, 2001) nor the release of radiolabelled RNA (our unpublished observation) was different between wild type and mutants during growth. Furthermore, cell death was not reduced in the T4SS mutants as compared with wild type when analysed by live/dead staining and fluorescent microscopy or fluorometry (Hamilton et al., 2005). Thus, the T4SS releases DNA into the medium by a method different from autolysis.
Little is known about the secreted DNA; how it is processed, if secretion is linked to chromosome replication, or if secreted DNA is single- or double-stranded. However, it is easy to imagine type IV secretion as a mechanism of DNA donation for lateral gene transfer within a population of gonococci without sacrificing vertical transmission or risking lysis and death of the entire population.
Autolysis is a curious phenomenon. As many microbiologists study survival strategies, the idea that a bacterium would act to promote its own death and lysis seems inherently wrong. In spite of our feelings on the subject, N. gonorrhoeae autolyses readily; in vitro, autolysis normally occurs during stationary phase or during non-growth conditions, such as exhaustion of nutrients, non-optimal temperature, pH or osmolarity (Hebeler and Young, 1975). The major autolysin in vitro is thought to be an N-acetylmuramyl-l-alanine amidase (Hebeler and Young, 1976). An enzyme with this specificity is also the major autolysin in S. pneumoniae, and recent evidence suggests that the pneumococcal enzyme acts in the lysis of a portion of the population to provide donor DNA for transformation (Steinmoen et al., 2002). Two amidase homologues are encoded in the gonococcal chromosome sequence (GenBank Accession No. AE004969). One of these is similar to Escherichia coli AmpD and is predicted to act in the breakdown of peptidoglycan monomers in the cytoplasm, but not in breakdown of the cell wall. The second is similar to E. coli AmiC and would be predicted to act on macromolecular peptidoglycan. The AmiC homologue is likely responsible for the peptidoglycan hydrolysis activity identified by Hebeler and Young (1976) and might play an important part in gonococcal autolysis. Membrane instability also contributes to autolysis (Cacciapuoti et al., 1978) and a recent report describes mutations in a phospholipase that result in decreased autolysis in N. meningitidis and N. gonorrhoeae (Bos et al., 2005).
DNA released through gonococcal autolysis can contribute to DNA donated for natural transformation (Norlander et al., 1979). For the 20% of gonococcal strains that lack the T4SS and for all N. meningitidis strains, autolysis is likely the only mechanism of DNA donation. A number of questions arise when considering donation of DNA by cell death. Is cell lysis mediated by host defences, or do the bacteria commit suicide by autolysis? How can autolysis be controlled such that one sibling dies while another does not? Is autolysis part of a programmed cell death event that processes DNA to make it better for transport to, or incorporation into, another cell? Which is better for transformation, DNA released by autolysis or by secretion? Both from an immunological standpoint and for understanding DNA donation, the process of autolysis warrants further investigation.
One possible mechanism for initiating a programmed cell death event was described for N. gonorrhoeae. Gonococci were found to carry a phase-variable type III restriction-modification enzyme, designated NgoX. The coding sequence of the methylase subunit carried 8–12 pentanucleotide repeats near the 5′ end of the gene, allowing the sequence to slip in and out of frame by slipped-strand mispairing, similar to the well-characterized mechanism of opacity protein variation in gonococci (Stern et al., 1986; Belland et al., 1996; Snyder et al., 2001a). The type III restriction-modification system methylase is required for sequence recognition. Therefore, the switching on of the methylase with the restriction subunit already produced could result in destruction of the chromosome (Bourniquel and Bickle, 2002). A second type III methylase (NG0641) carries 37 copies of a 4 bp repeat and would similarly be expected to phase-vary (Snyder et al., 2001a). Phase-variable expression of these lethal activities could allow for random death of a portion of the population.
Most gonococci also contain a toxin-antitoxin gene pair on the chromosome (J.P. Dillard and H.S. Seifert, unpublished). Although the gonococcal toxin–antitoxin (TA) system has not been characterized, related protein-based killer systems cause cell death or cell cycle arrest. Interestingly, the TA systems appear to respond to the growth state of the cell (Gerdes et al., 2005). As gonococci are known to autolyse under non-growth conditions, the TA system would be one explanation for the linkage of autolysis to growth conditions and could explain the massive death and lysis occurring in cultures following the exhaustion of nutrients (Morse and Bartenstein, 1974).
Step 2. DNA binding and uptake
The process of DNA binding and uptake by the ever-competent gonococcus has been intensely studied and recently reviewed (Chen and Dubnau, 2004). A major remaining question is which, if any, protein acts as a receptor for transforming DNA. Because of its strong preference for DUS-containing transforming DNA, N. gonorrhoeae must distinguish between self DNA (carrying the DUS) and non-self DNA (lacking the DUS). Thus, DNA binding events are classified into two types: non-specific and specific. Once specific binding of DUS-containing DNA has occurred, uptake across the outer membrane, the murein layer and the inner membrane follow.
Nearly all naturally transformable bacteria, both Gram-positive and Gram-negative, utilize type IV pili (Tfp) or a type IV pilus-like apparatus for the process of DNA uptake. Although the Gram-positive species do not have type IV pili, the DNA uptake proteins show significant similarity to proteins required for assembly of Tfp or the related type II secretion systems (Chen and Dubnau, 2004). Very early in the studies of gonococcal competence, a correlation was established between the expression of type IV pili and competence. Piliated gonococci were found to be more transformable than non-piliated bacteria (Sparling, 1966). At high DNA concentrations, the transformation frequency can exceed 20% (Gunn and Stein, 1996), whereas non-piliated variants yield transformation frequencies near 1 × 10−7 (Sparling, 1966). Both non-specific and specific DNA binding has been measured (Dougherty et al., 1979). Only strains expressing retractable Tfp bound DNA non-specifically, suggesting the hypothesis that electrostatic interactions between DNA and positively charged patches on Tfp might lead to non-specific binding (Aas et al., 2002a). It is unclear whether the pilus or a different pilus-like apparatus mediates uptake of DUS-carrying DNA. Mutations in genes involved in pilus biogenesis [pilE (pilin subunit), pilF (putative NTP-binding protein), pilD (prepilin peptidase), pilG (putative inner membrane protein) and pilQ (outer membrane secretin protein)] reduce natural transformation to undetectable or nearly undetectable levels, and reduce binding of DUS-containing DNA to the level of non-specific DNA (Drake and Koomey, 1995; Freitag et al., 1995; Boyle-Vavra and Seifert, 1996; Aas et al., 2002a). PilC, which is localized to the pilus tip or to the outer membrane, has also been implicated in competence (Jonsson et al., 1991; Rudel et al., 1995).
Even small amounts or altered forms of pilin are sufficient for gonococcal transformation. S- and L-pilin forms result from recombination events in the pilE expression locus; S-pilin variants display intermediate piliation and secrete a soluble form of the pilin monomer, while L-pilin variants are the result of a tandem duplication in pilE that generates an overlong pilin that is not assembled into pili. Gibbs et al. (1989) demonstrated that S-pilin variants display wild-type transformation frequencies, while the competence of gonococcal L-pilin variants is reduced approximately 35-fold. S-pilin variants make some normally processed pilin (Long et al., 1998), and it is likely that L-pilin variants make a small amount as well, thereby fulfilling the need for pilin, if not pili, for transformation. The fact that PilE and the pilus assembly proteins are necessary for natural transformation suggested that Tfp mediate DNA transformation directly; however, this hypothesis has not been proven, leaving open the possibility that a pilus-like apparatus might be enough to mediate natural transformation. In fact, Long et al. (2003) demonstrated that gonococcal pilin variants that elaborate virtually no observable pili but that still express pilin are competent for natural transformation and that this transformation is dependent on the pilus-related proteins PilQ and PilT. This finding generates support for the hypothesis that full, extended pili are not necessary for natural transformation and that a pilus-like apparatus is sufficient.
Two proteins known to be involved in gonococcal competence do not have a demonstrated role in Tfp biogenesis: PilT and ComE (Biswas et al., 1989; Chen and Gotschlich, 2001; Aas et al., 2002a,b). The dud-1 mutant was one of the first reported transformation-deficient mutants (Biswas et al., 1989) and the mutation was later mapped to pilT (Wolfgang et al., 1998). PilT is necessary for twitching motility, which is believed to occur via retraction of gonococcal Tfp. PilT, whose homologue has been shown to be an NTPase (Herdendorf et al., 2002; Aukema et al., 2005), is not necessary for specific binding of DNA but is rather involved in uptake of DNA (Aas et al., 2002a; Herdendorf et al., 2002). Uptake of DNA across the outer membrane, i.e. the transfer to a DNase-resistant state, has been demonstrated to be distinct from DNA binding (Aas et al., 2002a), and components involved in competence have been linked to each step. ComE was identified as the gonococcal orthologue of the B. subtilis competence protein, ComEA. There are four copies of comE in the gonococcal chromosome and sequential mutation of comE1–4 additively diminished transformation efficiency (Chen and Gotschlich, 2001). ComE binds DNA non-specifically and its presence is necessary for efficient DNA uptake (Chen and Gotschlich, 2001). As ComE is likely not localized to the gonococcal surface, it has been proposed that ComE binds DNA within the periplasm after specific DNA uptake has occurred (Chen and Gotschlich, 2001; Aas et al., 2002a).
Two pilin-like proteins, ComP and PilV, have been shown to affect the level of specific DNA binding (Aas et al., 2002b). Mutants lacking one or the other of these proteins still produce pili. However, mutants defective in comP and pilV are greatly reduced in piliation. Increased levels of ComP result in high transformation frequencies, whereas increased PilV decreases transformation. Although no regulation of competence has been observed in wild-type Neisseria, control of the relative levels of these two proteins could be used to modulate competence (Aas et al., 2002b). Mutants defective in pilC1 and pilC2, or mutants lacking one of five newly discovered pilin-like proteins PilH–L were not transformable or were significantly reduced in transformation competence and were found not to localize ComP to the pili (Winther-Larsen et al., 2005).
Mutations in the additional genes comA, comL, dca and tpc also diminish natural transformation. ComL, competence lipoprotein, is associated with the murein layer and is thought to be involved either directly or indirectly in puncturing the peptidoglycan layer for incoming DNA entry into the gonococcus (Facius et al., 1996; Fussenegger et al., 1996a). In addition to competence defects, mutants in tpc display a reduction in total murein hydrolase activity and have cell separation defects; they grow in tetrapacs rather than as diplococci (Fussenegger et al., 1996b). The involvement of ComL and Tpc might indicate the need for localized openings in the gonococcal peptidoglycan for DNA uptake to occur, or may be due to pleiotropic effects on cell wall metabolism. Dca is encoded in the division and cell wall gene cluster, but was not found to have any effect on peptidoglycan metabolism. It is predicted to be an inner membrane protein, and it is required for competence in N. gonorrhoeae but not N. meningitidis (Snyder et al., 2001b). ComA is predicted to be an inner membrane protein involved in transport of DNA into the cytosol (Facius and Meyer, 1993). This idea is supported by studies of the related protein ComEC in B. subtilis. ComEC is a polytopic membrane protein required for DNA uptake and is thought to form a channel for DNA transport (Draskovic and Dubnau, 2005).
Taken together, these data suggest a model for gonococcal competence (Fig. 1). Many components involved in Tfp biogenesis are necessary for competence, and DNA is bound by an as yet unidentified component that is likely associated with the pilus or a pilus-like apparatus. ComP and PilV influence specific binding of DUS-containing DNA. Once across the outer membrane, ComE binds DNA and aids in transformation, while Tpc and ComL may allow the DNA to cross the peptidoglycan layer. PilT is necessary for DNA uptake, theoretically by retracting the Tfp or the subunits of a pilus-like apparatus. ComA might help the DNA cross the inner membrane into the cytosol. Once within the cytoplasm, this DNA might be processed by gonococcal enzymes.
Step 3. DNA processing
The fate of transforming DNA in N. gonorrhoeae is dependent on a number of factors, including its size and nature. Transformation of N. gonorrhoeae with plasmid DNA is 1000-fold less efficient than with chromosomally derived loci (Eisenstein et al., 1977). Sox et al. (1979) demonstrated that transformation of N. gonorrhoeae with a native 7.1 kb, penicillinase-producing plasmid often generated larger or smaller versions of the plasmid in gonococcal transformants. E. coli-propagated versions of the same gonococcal plasmids did not transform recipient gonococci. These experiments provided evidence that plasmid DNA is subject to restriction by the gonococcus during transformation (Sox et al., 1979). N. gonorrhoeae encodes on the order of 16 methyltransferases, many of which have corresponding endonucleases that, together, generate an impressive restriction barrier for transforming plasmids (Stein et al., 1995). The capacity for plasmid transformation can be improved by inactivating several of these restriction-modification systems (Gunn and Stein, 1996). Large plasmids are processed into linear dsDNA pieces during transformation, before any restriction of DNA that occurs inside the gonococcus (Biswas et al., 1986).
The restriction barrier to plasmid DNA transformation in N. gonorrhoeae is not evident during transformation with chromosomal loci. Although initial studies of gonococcal transformation suggested that transforming DNA enters the cell as double-stranded molecules (Biswas and Sparling, 1981), more recent studies suggest that at least some ssDNA is formed during transformation and that this conversion occurs primarily in the periplasm (Chaussee and Hill, 1998). While the enzyme(s) responsible for this activity have not been identified, a similar mechanism of the degradation of one DNA strand during transformation exists in Gram-positive bacteria (reviewed in Dubnau, 1999). The lack of an observable restriction barrier during transformation of chromosomal loci argues that this DNA, at least once it is in the cytoplasm, is single-stranded and therefore resistant to endonucleases. Incoming ssDNA would not be restricted, and the integrated heteroduplex would be methylated on the resident strand and therefore not restricted. As plasmid transformation results from the re-assembly of an intact plasmid by the annealing of overlapping imported strands, both strands would be unmodified and therefore susceptible to restriction. A similar restriction barrier is observed for plasmids introduced into N. gonorrhoeae by conjugation (Butler and Gotschlich, 1991).
The size of the imported DNA has not been extensively studied. Experiments with plasmids (mentioned above) show that it is easy to recreate a 7 kb plasmid following transformation, but impossible to recreate a 42 kb plasmid. This result might indicate that the 42 kb plasmid was cut at multiple sites before import, and suggests that imported fragments are less than 42 kb. Linkage studies performed using chromosomal markers conferring antibiotic resistance can now be re-analysed for size using the chromosome sequence data. Sarubbi et al. (1974) showed linkage of antibiotic resistance markers rif-1 and str-7, mutations known to occur in rpoB and rpsL respectively. These genes are separated by 11 kb in strain FA1090. Similarly, linkage was found with str-7 (rpsL) and spc-3 (rpsE), which are 13 kb apart. So, 11 kb and 13 kb fragments can be imported. However, no linkage was found between rif-1 and spc-3, indicating that the ∼25 kb fragment carrying both of these markers is rarely or never imported.
Step 4. Homologous recombination into the gonococcal chromosome
Homologues of enzymes involved in homologous recombination in E. coli have been identified in N. gonorrhoeae, and a number of these proteins are required for gonococcal transformation. Gonococcal genes encoding RecA, B, C, D, O, Q and X are present in the gonococcal chromosome, suggesting the existence of both RecBCD and RecF pathways for homologous recombination, both of which are dependent on RecA (Koomey and Falkow, 1987; Mehr and Seifert, 1998; Stohl and Seifert, 2001). RecA is necessary to maintain newly acquired chromosomal markers via homologous recombination into the genome, and RecA-deficient gonococci are not transformable (Koomey and Falkow, 1987).
Mehr and Seifert demonstrated that the RecBCD pathway of homologous recombination, but not the RecF pathway, is necessary for efficient gonococcal transformation. However, gonococci are only 40-fold reduced in transformation efficiency in a RecBCD-pathway mutant as well as a RecBCD-/RecF-pathway mutant, suggesting the possibility of a third, as yet undescribed pathway for homologous recombination during natural transformation (Mehr and Seifert, 1998). Additionally, the homologous recombination activity that remains in these mutants might reflect a difference in the nature of the transforming DNA; single-stranded transforming DNA might bypass RecBCD and bind directly to RecA to mediate recombination (Mehr and Seifert, 1998). Finally, gonococci lacking RecX, which is presumed to regulate RecA activity, are fivefold reduced in DNA transformation (Stohl and Seifert, 2001; Stohl et al., 2003).
Advantages and implications
It is notable that many naturally transformable species are human pathogens. This might be an artefact of our human-centric view of the world or the current state of funding. However, it is clear that transformation impacts N. gonorrhoeae pathogenesis and consequently, its evolutionary survival. The genome sequence of N. gonorrhoeae strain FA1090 reveals the presence of at least six genetic islands, presumably acquired via transformation from different bacterial species, which may encode any number of fitness advantages to the organism (GenBank Accession No. AE004969). The GGI, which is present in 80% of gonococcal strains (but not FA1090), has the characteristics of a horizontally acquired mobile genetic element that might be transferred between gonococcal strains by natural transformation (Hamilton et al., 2005). In addition to the acquisition of large genetic islands, gonococcal transformation is known to generate hybrid porin alleles and is likely to also generate other advantageous alleles which affect gonococcal pathogenesis (Hobbs et al., 1994; Cooke et al., 1998).
Although the gonococcus is an excellent model for the study of natural transformation, many questions still remain. How is DNA donation controlled? Is cell lysis a stochastic or a regulated process? What are the identities of the autolysins? Both a phospholipase and a peptidoglycan hydrolase appear to be involved in this process, and mechanisms of autolysis were characterized decades ago (Hebeler and Young, 1976; Cacciapuoti et al., 1978). However, no one has yet been able to identify a nonautolytic mutant. Does type IV secretion of DNA occur in all cells or only a portion of the population, and how is it regulated?
What binds the DUS on DNA during transformation? As mentioned previously, the receptor for species-specific DNA for natural transformation has not yet been identified. The prime candidates for this function would appear to be PilE, ComP, or another protein that is controlled by the presence of ComP. However, no DNA binding to PilE or ComP could be demonstrated (Mathis and Scocca, 1984; Aas et al., 2002a), and a ComP-dependent DNA-binding protein has not yet been found.
Are gonococci as efficiently transformed with ssDNA as with dsDNA? Stein has demonstrated that ssDNA generated by M13 phage transformed gonococci at similar levels as double-stranded donor DNA (Stein, 1991); however, some researchers adhere to the belief that only dsDNA transforms. ssDNA was also reported to transform H. influenzae with an efficiency on the order of ∼50% that of dsDNA (Postel and Goodgal, 1966), although this result has also been challenged (Mulder and Doty, 1968). Recently a third Gram-negative bacterial species, Pseudomonas stutzeri, was shown to be transformed by ssDNA (Meier et al., 2002). Additionally, N. meningitidis PilQ, which forms the outer membrane ‘pore’ of the pilus apparatus and is required for natural transformation, has been recently found to bind ssDNA better than dsDNA (Frye et al., 2004). These facts argue that ssDNA should be a good substrate for transformation of N. gonorrhoeae. Additionally, DNA secreted by the gonococcal T4SS, which is known to transform recipient bacteria, is presumably single-stranded (Hamilton et al., 2001).
Are there effects on recipient gonococci apart from acquisition of genes? N. meningitidis shows increased phase variation in response to transforming DNA from heterologous Neisseria species or N. meningitidis strains, due to titrating DNA repair proteins (Alexander et al., 2004). A similar phenomenon may also occur in N. gonorrhoeae. In addition, it might be advantageous for gonococci to use the recognition of incoming DNA as the indication of the presence of significant numbers of other gonococci, i.e. DNA could act as a quorum-sensing molecule. Presumably genes might be turned on in response to transforming DNA, much like genes are regulated in the response to detection of other well-known quorum-sensing molecules.
Could the methylation state of incoming DNA aid in distinguishing self gonococcal DNA from non-self gonococcal DNA? The many restriction-modification systems of N. gonorrhoeae might differentially methylate genomic DNA. Taking up DNA of non-self rather than self DNA (i.e. DNA from a different gonococcal strain) could be more advantageous for the purposes of allelic diversification and fitness.
Does N. gonorrhoeae possess a competence nuclease to degrade one incoming strand during transformation with dsDNA? It is interesting that most models of transformation in both Gram-positive and Gram-negative bacteria depict a competence nuclease that degrades one strand of the double-stranded transforming DNA, but only one competence nuclease has yet been identified, EndA of S. pneumoniae (reviewed in Dubnau, 1999). DNA degradation studies like those performed by Provvedi et al. (2001) in B. subtilis might aid in elucidating the fate of incoming DNA in N. gonorrhoeae. Along a similar line what is the identity of the specific enzymes that linearize circular DNA molecules during transformation with plasmid DNA?
Natural transformation appears to be an ever-present and essential mechanism for the acquisition and exchange of genetic material in the gonococcus. This process has undoubtedly contributed to the success of N. gonorrhoeae as a human pathogen. N. gonorrhoeae remains an ideal organism for the study of natural transformation and will be a crucial tool for understanding bacterial competence for many years to come.
We acknowledge the Gonococcal Genome Sequencing Project supported by USPHS/NIH Grant ♯AI38399, and B.A. Roe, L. Song, S.P. Lin, X. Yuan, S. Clifton, T. Ducey, L. Lewis and D.W. Dyer at the University of Oklahoma. We would like to acknowledge the support from our laboratory through NIH Grant AI47958 to J.P.D. and traineeship on the NIH T32 Grant AI055397 to H.L.H. We apologize to those whose work contributed to the fields discussed in this review but were not specifically mentioned.