Mutational Analysis of Genes Involved in Pilus Structure, Motilityand Transformation Competency in the Unicellular Motile CyanobacteriumSynechocystis sp. PCC6803

Abstract

The relevance of pilus-related genes to motility, pilus structure on the cell surface and competency of natural transformation was studied by gene disruption analysis in the unicellular motile cyanobacterium Synechocystis sp. PCC 6803. The genes disrupted in this study were chosen as related to the pil genes for biogenesis of the type IV pili in a Gram-negative bacterium Pseudomonasaeruginosa. It was found that motility of Synechocystis cells was lost in the mutants of slr0063, slr1274, slr1275, slr1276, slr1277 and sll1694 together with a simultaneous loss of the thick pili on the cell surface. Competency of the natural transformation was lost in the mutants listed above and slr0197-disruptant. The gene slr0197 was previously predicted as a competence gene by a search with sequence-independent DNA-binding structure [Yura et al. (1999)DNA Res. 6: 75]. It was suggested that both DNA uptake for natural transformation and motility are mediated by a specific type IV-like pilus structure, while a putative DNA-binding protein encoded by slr0197 is additionally required for the DNA uptake. Based on the homology with the pil genes in P. aeruginosa, slr0063, slr1274, slr1275, slr1276, slr1277 and sll1694 were designated pilB1, pilM, pilN, pilO, pilQ and pilA1, respectively. The gene slr0197 was designated comA.

(Received August 23, 2000; Accepted October 25, 2000).

Introduction

Cyanobacteria are the most ancient organisms to perform oxygenic photosynthesis and are believed to be ancestors of chloroplasts in plants. Not all but many cyanobacteria show motility in a form of gliding, twitching or swimming, although none of them employ flagella which is in contrast to many other bacteria (Diehn et al. 1979, Häder 1987, Waterbury et al. 1985). The molecular mechanism of such motility has been poorly understood even though physiology of motility has been extensively studied by using filamentous cyanobacteria for many decades. However, to date only a few genes have been shown to be responsible for motility in unicellular cyanobacteria. In marine Synechococcus which shows swimming motility (Waterbury et al. 1985), swmA encoding a cell-surface-associated protein of 130 kDa was shown to be essential for translocation of cells in liquid (Brahamsha 1996). On the other hand, the transformable cyanobacterium Synechocystis sp. PCC 6803, of which the complete genome was determined (Kaneko et al. 1996), shows sporadic motility of twitching (Stanier et al. 1971). Among over 3000 ORFs in the genome, Bhaya et al. (1999) reported that an alternative sigma factor, sigF, and putative pilus subunit genes, sll1694 and sll1695, are essential for motility. Okamoto and Ohmori (1999) also reported that slr0161 (pilT) gene product having ATPase activity is essential for motility in the same organism. These suggest that a kind of pili, which resemble the type IV pili in many Gram-negative bacteria (Mattick et al. 1996), may be the motility machinery at least in some cyanobacteria, although the pili themselves or genes for pilus biogenesis have not yet been characterized.

The type IV pili are widely distributed in many Gram-negative bacteria (Mattick et al. 1996). They are typical extracellular structures located at the pole of the cell and feature unique pilin subunits (Sastry et al. 1983). Each subunit is synthesized as a precursor and a unique leader sequence is cleaved off on the cytoplasmic side by a specific prepilin peptidase, which recognizes a region around the cleavage site including the mature N-terminal part (Nunn and Lory 1991). Accordingly, the N-terminal region of the mature pilin protein is highly conserved, while the leader sequence or C-terminal region of the protein is not conserved (Sastry et al. 1985). It is well known that the type IV pili are involved in various cellular functions such as motility, transformation, infection and adhesion (Strom and Lory 1993). In an opportunistic Gram-negative pathogen Pseudomonasaeruginosa, the type IV pilus has been shown to be responsible for the twitching motility (Darzins and Russell 1997, Mattick et al. 1996). In a soil bacterium Myxococccusxanthus, the social motility of the gliding cells requires a number of pil genes for the type IV pilus (Spormann 1999). In a pathogenic bacterium Neisseriagonorrhoeae, the type IV pilus is involved in both twitching motility, competence for natural transformation and infection to human cells (Fussenegger et al. 1997, Wolfgang et al. 1998). Thus, it is suggested that the type IV pili are the common apparatus for the twitching or gliding motility in many Gram-negative bacteria. However, the molecular mechanism of motility or its regulatory processes is still totally unknown in contrast to the well-characterized swimming motility by flagella.

Based on the complete genome sequence of the unicellular cyanobacterium Synechocystis sp. PCC 6803, we have been studying physiological roles of a number of genes by directed gene disruption. It was found that a putative protein phosphatase gene, slr2031, is required for motility, although its target phosphoprotein is yet to be determined (Kamei et al. 1998). We also found that many regulatory components such as transcription factor, sensor histidine kinase, eukaryotic-type serine/threonine protein kinase and photoreceptor are also required for normal motility (Kamei et al. 2001, Yoshihara et al. 2000). For thorough understanding, it is a prerequisite to characterize the apparatus for motility as a target of those regulatory components. To end this, we created mutants of the ORFs which are homologous to the genes involved in biogenesis of the pilus structure in P. aeruginosa and examined the phenotype. It was found that a specific type IV-like pilus structure supported by a certain set of pil genes is indispensable for motility and transformation competency.

Materials and Methods

Culture and growth condition

The motile strain of the unicellular cyanobacterium Synechocystis sp. PCC 6803 was obtained from Pasteur Culture Collection and a clone showing vigorous motility with positive phototaxis (substrain PCC-P) was selected as a parent strain for gene disruption. Cells were grown in liquid BG11 medium (Stanier et al. 1971) bubbled with air containing 1% (v/v) CO₂ at 31°C at light intensity of 50 µE m^–2 s^–1. Kanamycin was included at 20 µg ml^–1 when mutants were screened and maintained.

Insertional mutagenesis

The protein database derived from the genome of Synechocystis sp. PCC 6803 was searched for homology using the BLAST program (Altschul et al. 1997) with deduced amino acid sequences of the known pil genes from the γ-proteobacterium, P.aeruginosa (Mattick et al. 1996). To inactivate each pil-like gene thus highlighted, a DNA harboring the gene was amplified with primers (Table 1) with AmpliTaq DNA polymerase (PE Applied Biosystems, Foster City, U.S.A.) under the conditions recommended. DNA was cloned into pT7Blue-T vector (Novagen, Madison, U.S.A.) and then a Tn5-derived kanamycin resistant cassette, in which transcription was not terminated (a read-through type), was inserted at a unique restriction site into the target gene (Table 1). The DNA, thus constructed, was introduced into the motile substrain PCC-P of Synechocystis sp. PCC 6803 as described previously (Hihara and Ikeuchi 1997). Complete segregation of the mutation was confirmed by polymerase chain reaction using the same primers as used for cloning. Mutant strain names were abbreviated as in Table 1; e.g. Msll1694 for the disruptant of sll1694.

Motility assay

Motility was estimated by colony morphology on 0.8% (w/v) agar (Bacto-Agar, Difco, Detroit, U.S.A.) (Bradley 1980). Cells grown in liquid at late log-phase (A₇₃₀=0.8–1.0) were spread on BG11 solid medium containing 0.8% (w/v) agar and 0.3% (w/v) sodium thiosulfate and cultivated at 31°C under white fluorescent lamp of light intensity at 50 µE m^–2 s^–1 for 4 d.

Detection of pili by electron microscopy

Cells of each strain growing on agar plates were gently suspended in BG11 medium and examined after staining with 0.8% (w/v) phosphotungstic acid (pH 7.0) by transmission electron microscope (model 1200EX, JEOL, Tokyo, Japan) basically according to Vaara and Vaara (1988).

Transformation competency assay

Competency of each mutant was estimated by transformation efficiency with a test DNA, in which Synechocystis genomic DNA of 3.1 kbp carrying pmgA was interrupted with a spectinomycin-resistant cassette in pUC18 plasmid (Hihara and Ikeuchi 1997). 10⁸ cells of each mutant, as well as wild type, growing at late log-phase were mixed with 0.1 µg of the test DNA in 1 ml, immediately spread on a nitrocellulose membrane (Millipore, Bedford, U.S.A.) on the BG11 agar plate and incubated for 24 h under illumination at 50 µE m^–2 s^–1 at 31°C. Then, cells on the membranes were transferred onto another BG11 agar plate containing 20 µg ml^–1 spectinomycin. After incubation for about a week under the same conditions as above, the number of green transformants were counted. Usually, about 600–850 transformants were obtained from the motile substrain PCC-P.

Results

Using the known pil genes of P.aeruginosa, we performed homology search of the genome sequence of Synechocystis sp. PCC 6803 (Altschul et al. 1997). More than 30 ORFs were highlighted, although most of them were not annotated as pil genes but just hypothetical or gsp/hof genes for general secretion pathway in the original report (Kaneko et al. 1996). Of these, we chose 15 ORFs for disruption analysis in this work, since they are homologous to the genes involved in biogenesis of the pilus structure in P. aeruginosa (Table 2). These ORFs were classified into four groups: (a) a group of genes homologous to the pilus subunit (pilA), which has a unique sequence feature of the type IV prepilin near the mature N-terminus (see Discussion); (b) two genes homologous to gspE or pilB encoding ATPase; (c) a cluster of genes weakly homologous to genes of the pilM operon in P. aeruginosa and (d) a gene homologous to pilD for prepilin peptidase. Besides these genes, the Synechocystis genome carried many pil-like ORFs but we did not study them in this report because many of them were just homologous to a universal part of regulatory components such as the two component phosphotransfer system, which is known to regulate expression of pil genes in P. aeruginosa (Ishimoto and Lory 1992). We disrupted separately the 15 ORFs in the list and examined their phenotype with regard to motility, pilus structure and transformation competency. Complete segregation of the mutants except slr1120-disruptant was achieved and confirmed by polymerase chain reaction, while the slr1120 mutant did not segregate the wild-type genome within this work.

Motility of the mutants

Motility ofSynechocystis cells was examined by colony morphology. Cells of the motile substrain PCC-P formed flat sheet-like colonies showing irregular shapes due to motility of individual cells (Fig. 1A). On the other hand, Mslr0063 formed domed colonies of round shape, indicative of non-motile phenotype (Fig. 1E). Disruption of the other gspE-like slr0079 did not affect the colony morphology (Fig. 1C).

All the disruptants of the pilM-like gene cluster (Mslr1274, Mslr1275, Mslr1276 and Mslr1277) showed domed colonies (Fig. 1F-I), indicative of the non-motile phenotype. Since each of the clustered genes was disrupted by insertion of the read-through type cassette with the same direction as the gene cluster, we may exclude the possibility of pleiotropic effects on the downstream genes, although they were expressed from the promoter of the kanamycin cassette but not from the intrinsic promoter. Thus, we concluded that all of them were required for motility in Synechocystis.

Bhaya et al. (1999) reported that the double mutant of sll1694 and sll1695 lost motility. We disrupted them separately and also created the similar double mutant. It was found that Msll1694 and the double mutant but not Msll1695 lost motility and thus formed domed colonies (Fig. 1). Cells in colonies of Msll1694 and the double mutant appeared to be more tightly aggregated than the parent PCC-P and it was difficult to collect those cells from agar plates with a loop. When grown in liquid BG11 medium, cells of Msll1694 and the double mutant tended to form a large clump. These properties were not observed in PCC-P or the other non-motile mutants. We also disrupted other pilin-like genes (slr1046, slr1456, slr1928, slr1929, slr1930 and slr1931). All of them did not show any defects in motility as judged by colony morphology (not shown).

Although the slr1120 mutant retained genome DNA of wild type under the selective conditions and did not segregate, it showed minute domed colonies (not shown). This indicates that the slr1120 product is required not only for motility but also for normal growth. This suggests that a putative peptidase of Slr1120 processes not only prepilins but also other proteins. In P. aeruginosa, a single PilD peptidase is shared with the two distinct systems, the type IV pilus and the general secretion pathway (Nunn and Lory 1992).

Pilus structure on the cell surface

The pilus structure on the cell surface of the mutants as well as the parent PCC-P was examined by electron microscopy after negatively staining with phosphotungstic acid. In Fig. 2, typical images of pili on the cell surface of PCC-P are presented. Clearly, we could see three types of pili on the surface of cells. One type is a thick pilus, with external diameter of 5 nm and length of more than 2 µm (Fig. 2, long arrows). They are often observed apart from the cell (Fig. 2, arrowheads), probably because they are too long compared with the cell size (about 2 µm) and detached from the cells during the transferring process from the agar plate to the grid. The second type is a thin pilus, which was also present in a great number on the PCC-P cells (Fig. 2, short arrows). Although these thin pili were not clearly visualized by negative staining, their diameter was about 3–4 nm. The thin pili were observed only on the cell surface. The third type is a bundle of pili (Fig. 2, white arrows). The diameter of the bundles was less than 45 nm and gradually decreased from the proximal end on the cells surface to the tip. In some case, we could see the thin pili projecting from the tip of the bundle, indicating that these bundles consist of the thin pili but not of the thick pili. In short, PCC-P cells bear on the cell surface both the thick and thin pili, the latter often aligned to form bundles.

In the non-motile mutants of sll1694, slr0063, slr1274, slr1275, slr1276, and slr1277, the thick pili were clearly absent (Fig. 3D-I). Moreover, Msll1694 was largely depleted of the thin pili. The other non-motile mutants retained more or less the bundles of the thin filaments. On the other hand, the thick pili were present in the motile mutants such as Msll1695 and Mslr0079 (Fig. 3B, C). The number of the thick pili in Msll1695 was much more than in PCC-P (Fig. 3A), although we could not detect a marked difference in motility between PCC-P and Msll1695. These results strongly suggest that the thick pili are responsible for motility and their assembly or formation strictly depends on the genes, sll1694, slr0063, slr1274, slr1275, slr1276, and slr1277.

Transformation competency of the mutants

It is well known that Synechocystis cells are naturally transformable to uptake extracellular DNA (Grigorieva and Shestakov 1982). The incorporated DNA is usually integrated by homologous recombination with double crossing over in this cyanobacterium. Thus, we estimated transformation competency of the mutants as well as the parent PCC-P by using a gene disruption construct with the different antibiotic resistant gene (aadA) (Table 3). Under the experimental conditions we employed, PCC-P gave about 600–850 transformants out of 10⁸ recipient cells, while the non-motile mutants of sll1694, slr0063, slr1274, slr1275, slr1276 and slr1277 gave no transformants reproducibly. On the other hand, the motile mutants of sll1695 and slr0079 retained competency, although the efficiency was lower than PCC-P. No spontaneous spectinomycin-resistant colonies emerged without addition of DNA. This strongly suggests that the thick pili are essential for transformation competency as well as motility.

Competence-specific gene

Yura et al. (1999) detected an ORF in the Synechocystis genome by a search with novel algorithms for sequence-independent DNA-binding proteins. Slr0197, thus obtained, has hybrid features of two known proteins: the N-terminal part was a tandem repeat of a domain homologous to endonuclease/cardiolipin synthetase, while the C-terminal part was homologous to DNA-binding competence protein of bacteria. To test whether slr0197 is involved in motility or competence, we created a disruption mutant. Complete segregation of the mutant genome was confirmed by polymerase chain reaction (not shown). Mslr0197 showed normal motility on the agar plate (Fig. 4A), whereas it lost transformation competency completely (Table 3). Electron microscopic examination of the negatively stained cells revealed the presence of both thick and thin pili, although the number of the thick pili seemed to be slightly fewer than PCC-P (Fig. 4B). This suggests that Slr0197 is specifically involved in the transformation process, but not in the motility process.

Discussion

By examination of the mutant phenotype of pil-like genes, we have demonstrated that slr0063, slr1274, slr1275, slr1276, slr1277 and sll1694 are required for assembly of the thick pili together with motility and transformation competency. Based on homology with the gene designation in P. aeruginosa, we designated slr0063, slr1274, slr1275, slr1276, slr1277 and sll1694 as pilB1, pilM, pilN, pilO, pilQ and pilA1, respectively. We also propose to name slr1120pilD, although the phenotype of the slr1120 mutant was only partially studied due to incomplete segregation. Combined with the observation that a gene cluster of pilT/pilC is essential for motility and transformation competency (Okamoto and Ohmori 1999), almost whole set of the pil genes, which are essential for the biogenesis of the type IV-like thick pilus structure, were now established in the unicellular cyanobacterium Synechocystis sp. PCC 6803. On the other hand, the thin pili may be composed of PilA1 (Sll1694), although they are not supported by the other pil genes. If so, the thin pili should be categorized into the type IV, too. Thus, “type IV pilus structure” may be too broad to classify the thick and thin pilus structures in Synechocystis sp. PCC 6803. We would like to reserve the conclusive classification of pili for future work.

It is well known that the general secretion system (also called the type II protein secretion system) in Gram-negative bacteria consists of a set of gsp genes, which resemble, in essence, the pil genes for the type IV pili (Russel 1998). In those, gspE and pilB encode a putative ATPase, while gspD and pilQ encode proteins belonging to a large secretin family, which form pores on the outer membrane in Gram-negative bacteria (Bitter et al. 1998). They are related to each other but they play distinct roles in the same organism such as P.aeruginosa (Martin et al. 1993, Akrim et al. 1993). In this report, we unambiguously demonstrated that slr0063 and slr1277 are required for formation of the thick pili, cell motility and transformation competency. This indicates that they are categorized as the pil genes but not gsp genes. On the other hand, slr0079 may be a homolog of gspE as already annotated in the genome (Kaneko et al. 1996), although we tentatively designated it pilB2. We could not find any homolog of gspD other than slr1277 (pilQ) in the Synechocystis genome. It should be noted that the thin pili, which often form bundles, were able to be assembled in the absence of Slr1277 (PilQ). This may indicate that Mslr1277 still assemble a kind of pore, from which the thin pili could be projected. There are many unidentified genes in the Synechocystis genome, which are seemingly targeted into the outer membrane as judged from the sequence homology. Some of them may be involved in the assembly of the thin pili.

There is a characteristic difference in gene arrangement of pilB, pilC and pilT between cyanobacteria and P. aeruginosa. pilT and its homologs are unique to the type IV pili in many bacteria but not known in the general secretion system. PilB and PilT are homologous to each other, having the ATP-binding “Walker” motif of GXXGXGKT (Walker et al. 1982). In P. aeruginosa, pilT mutant was featured by no twitching motility and hyper-piliation (Wall and Kaiser 1999, Whitchurch et al. 1991), while pilB mutant did not assemble the pili (Nunn et al. 1990). The same results were obtained in slr0161 (pilT) mutant (Okamoto and Ohmori 1999) and slr0063 (pilB1) mutant (this work). Furthermore, two ORFs, slr0162 and slr0163, at downstream of slr0161 was found to be derived from a frameshift mutation of a single gene, which is homologous to pilC in P. aeruginosa (Okamoto, S. and Ohmori, M. unpublished results). Thus, pilC is clustered with pilT (slr0161) but not with pilB1 (slr0063) in Synechocystis, whereas pilC is located just downstream of pilB but far from pilT in P. aeruginosa (Mattick et al. 1996). Homologs of pilB1 (slr0063), pilC, pilT (slr0161) were also detected in the genome of a filamentous cyanobacterium Anabaena sp. PCC 7120 (http://www.kazusa.or.jp/cyano/anabaena/). Interestingly, ORF670 (pilB), ORF370 (pilT) and ORF407 (pilC) are clustered together in this order in the Anabaena genome. Taking into consideration that many gene clusters are more preserved in Anabaena genome than in Synechocystis, the original cluster of pilB/pilT/pilC in Anabaena was probably split into the two, pilB1 and pilT/pilC in Synechocystis. In any case, these gene arrangement in cyanobacteria is distinctly different from pilB/pilC and pilT in P. aeruginosa.

Although the homology of Slr1274, Slr1275 and Slr1276 to PilM, PilN and PilO of P. aeruginosa is weak, our observation that those mutants of Synechocystis completely lost the thick pili strongly supported the idea that they are distant homologs of the Pseudomonasgenes. Strangely, we could not find in the Synechocystis genome an ORF homologous to pilP, despite that pilP is located in the pilM cluster and essential for the biogenesis of the type IV pili in P. aeruginosa (γ-proteobacteria) or Neisseriagonorrhoeae (β-proteobacteria) (Martin et al. 1995, Drake et al. 1997). PilP protein is assumed to be a lipoprotein located in the periplasm and play an important role in functioning of PilQ, the outer membrane pore protein in N. gonorrhoeae (Drake et al. 1997). The apparent absence of pilP in Synechocystis genome suggests that the cell surface structure of Synechocystisis somewhat different from that of the proteobacteria.

Bhaya et al. (1999) reported that a double deletion mutant of the tandemly arranged sll1694 and sll1695 lost most of pili and was non-motile. We disrupted them separately and found that only sll1694 (pilA1) was responsible for the thick pilus structure, motility and transformation competency. In Msll1694, the thin pili still remained but in a greatly reduced quantity. This may suggest that not only the thick pili but also the thin pili consist of pilin subunit of Sll1694 in common. On the other hand, we could not know the role of sll1695, although it is highly expressed as shown by DNA microarray analysis (Kamei, A., Hihara, Y. and Ikeuchi, M. unpublished results). In the Synechocystis genome, we could detect more than 8 genes including sll1694 and sll1695 for possible prepilin-like genes based on the conserved features of the prepilin genes (Fig. 5). Namely, the N-terminal hydrophobic regions of the mature proteins are highly conserved, while the leader sequences or C-terminal parts were not conserved. In addition, Gly residue at position –1 and Glu residue at position +5 of the mature proteins are totally conserved and Phe residue at +1 is also highly conserved. These fit well with the substrate specificity of PilD peptidase (Pasloske and Paranchych 1988, Strom and Lory 1991). We disrupted them separately by insertional mutagenesis but all of them except sll1694 retained motility. Such genes may be a component of the thin pili or the general secretion system, although the latter has not yet been proved experimentally. We are currently trying to create double mutants by introduction of slr0063-disruption in each mutant of the prepilin-like gene. Examination of the thin pili in these mutants would clarify the complex structure on the cell surface in cyanobacteria.

We demonstrated that slr0197 was essential for transformation competency but not for motility. This agrees with the prediction that slr0197 codes for a hybrid protein of endonuclease/cardiolipin synthetase and DNA-binding competence protein (Yura et al. 1999). The natural transformation is widely distributed in Gram-positive as well as Gram-negative bacteria. In Bacillussubtilis, it is known that exogenous DNA binds to ComEA on the cell surface, undergoes limited fragmentation by a certain nuclease, and then one strand is incorporated into the cell (Provvedi and Dubnau 1999). ComEA is a small DNA-binding protein, which shows homology to the C-terminal part of Slr0197. On the other hand, the hypothetical nuclease has not yet been identified in B. subtilis or in other bacteria. The predicted hybrid structure of Slr0197 suggests that the nuclease-like domain may digest the double strand DNA on the cell surface of Synechocystis (Yura et al. 1999). Taking these into consideration, we propose to designate slr0197comA in Synechocystis sp. PCC 6803. Homology search of the Synechocystis genome with other com genes in B. subtilis revealed the presence of comEC-like ORF (sll1929). To extend our hypothesis, we tried to disrupt this ORF but could not achieve the complete segregation of the mutant genome till now (not shown).

It is of note that the apparatus for DNA uptake in Gram-positive B. subtilis shows limited homology to the type IV pilus, which is widely distributed in Gram-negative bacteria. For example, a number of pilin-like genes and their specific peptidase are encoded by comGC/comGD/comGE/comGG and comC in B. subtilis, respectively (Dubnau 1997). However, the apparatus for DNA uptake or pilus structure cannot be detected on the Bacillus cells by electron microscopy. On the other hand, the Gram-negative bacterium N. gonorrhoeae shows the close relationship between the natural transformation competency and the type IV pilus structure (Fussenegger et al. 1997). However, slr0197(comA)-like gene has not yet been identified in this organism. Thus, Synechocystis may provide a unique experimental model to study the complex pili system in relation to motility and natural competence.

Acknowledgements

This work was supported by Grants-in-Aid for Scientific Research on Priority Areas C “Genome Biology” (12206002) (to M.I.) and for Scientific Research (08836002, 11554035, 09NP1501) (to M.I.) from the Ministry of Education, Science and Culture, Japan and by a grant for Scientific Research from the Human Frontier Science Program (to M.I.).

3

Corresponding author: E-mail, [e-mail]; Fax, [phone].

Fig. 1 Colony morphology of wild type and mutants on the agar plate. Cells were grown on 0.8% agar plate for 4 d under illumination at 50 µE m^–2 s^–1. A, wild type (PCC-P); B, Msll1695; C, Mslr0079; D, Msll1694; E, Mslr0063; F, Mslr1274; G, Mslr1275; H, Mslr1276; I, Mslr1277. Bars show 0.2 mm in A–C and 0.1 mm in D–I, respectively.

Fig. 2 Electron micrographs of negatively stained cells of wild type. Panel A, typical image with many thick pili, thin pili and a few bundles of the thin pili on the surface of the cell; B, a portion of panel A was enlarged three times; C and D, enlarged images of wild-type cells to show the thin pili, the bundle of the thin pili and detached thick pili. A bar in panel A shows 1 µm, while bars in panels B, C and D show 0.2 µm. Long arrows, arrowheads, short arrows and white arrows indicate the thick pili projected from the cell, the thick pili detached from the cell, the thin pili and the bundles of the thin pili, respectively.

Fig. 3 Electron micrographs of negatively stained images of wild type and mutants. Panel A, wild type (PCC-P); B, Msll1695; C, Mslr0079; D, Msll1694; E, Mslr0063; F, Mslr1274; G, Mslr1275; H, Mslr1276; I, Mslr1277. Bars show 1 µm.

Fig. 4 Colony morphology (A) and electron micrograph of negatively stained image (B) ofMslr0197 mutant. Bars show 0.1 mm in panel A and 1 µm in panel B, respectively.

Fig. 5 Sequence alignment of pilA-like gene products of Synechocystis with PilA of P. aeruginosa. Residues identical to Pseudomonas PilA are shown in reverse type letters. N-terminal and C-terminal sequences, which are not shown in this figures are indicated by dots. Asterisks represent the special residues recognized by PilD peptidase in P. aeruginosa. An arrow shows the possible cleavage sites by the PilD peptidase.

Table 1

Primers and positions of the insertional disruption for the pil-related genes

Disrupted gene (Mutant strain)	Forward primer	Reverse primer	Position of insertional disruption/ amplified DNA (bp)
sll1694 (Msll1694)	5-AGCGGGAAGAATCCGCAT-3	5-AAATGAGTTGAACTTGGG-3	518 / 2277
sll1695 (Msll1695)	5-AGCGGGAAGAATCCGCAT-3	5-AAATGAGTTGAACTTGGG-3	882 / 2277
slr0063 (Mslr0063)	5-CATATGACATCTTCCTCCTCT-3	5-CTCCGCAATGTCCATGTC-3	404 / 903
slr0079 (Mslr0079)	5-CATATGGTCTTTTCCTCCGATT-3	5-GGATTCCTGGGCTAATCT-3	711 / 1143
slr1274 (Mslr1274)	5-GGATGTTAGCCTCTCGGCAA-3	5-CTCTTCGCTGTAGCTGGTCA-3	1517 / 3140
slr1275 (Mslr1275)	5-GGATGTTAGCCTCTCGGCAA-3	5-CTCTTCGCTGTAGCTGGTCA-3	2527 / 3140
slr1276 (Mslr1276)	5-TTGAGGCAAAAACGGCGG-3	5-AATTACCCTTTAGGGGGC-3	772 / 1463
slr1277 (Mslr1277)	5-GCCAAGTGGCAGTAAACG-3	5-TAACTATGGCTGAGCCTG-3	475 / 1091
slr0197 (Mslr0197)	5-CATATGGGAAAACTAAACAGAGG-3	5-GATTGCTCGGCGATCGGG-3	362 / 903

Note that the same sets of primers were used for disruption of sll1694/sll1695 and slr1274/slr1275. This was because the neighboring ORFs in the same DNA fragment were insertionally interrupted at different positions.

Table 2

List of possible pil-related genes in the genome of Synechocystis sp. PCC 6803

	Synechocystis sp. PCC 6803					Pseudomonas aeruginosa
Gene ID	Original annotation	Assignment in this work a	References b		pil genes	Proposed function	References b
sll1694	hofG	pilA1	1, this work		pilA	prepilin (pilus subunit)	3, 4
sll1695	hofG	(pilA2)	1, this work		pilA	prepilin (pilus subunit)	3, 4
slr1046	hypothetical	(pilA3)	this work		pilA	prepilin (pilus subunit)	3, 4
slr1456	gspD	(pilA4)	1, this work		pilA	prepilin (pilus subunit)	3, 4
slr1928	hypothetical	(pilA5)	this work		pilA	prepilin (pilus subunit)	3, 4
slr1929	hypothetical	(pilA6)	1, this work		pilA	prepilin (pilus subunit)	3, 4
slr1930	hypothetical	(pilA7)	1, this work		pilA	prepilin (pilus subunit)	3, 4
slr1931	hypothetical	(pilA8)	this work		pilA	prepilin (pilus subunit)	3, 4
slr0063	gspE	pilB1	this work		pilB	pilus assembly	5
slr0079	gspE	(pilB2)	this work		pilB	pilus assembly	5
slr1274	pilM	pilM	this work		pilM	unknown	6
slr1275	hypothetical	pilN	this work		pilN	unknown	6
slr1276	hypothetical	pilO	this work		pilO	unknown	6
slr1277	gspG	pilQ	this work		pilQ	pore at the outer membrane	7
slr1120	tapD	pilD	this work		pilD	prepilin peptidase	8, 9
slr0197	comE ORF1	comA	2, this work		—

^a Gene names in parenthese are tentative.

^b 1, Bhaya et al. 1999; 2, Yura et al. 1999; 3, Pasloske and Paranchych 1988; 4, Sastry et al. 1985; 5, Nunn et al. 1990; 6, Martin et al. 1995; 7, Martin et al. 1993; 8, Nunn and Lory 1991; 9, Strom and Lory 1993.

Table 3

Transformation efficiencies of wild type and mutants

Strain	Transformation efficiency (%) a
wild type (PCC-P)	100
Msll1694 (pilA1)	0 b
Msll1695 (pilA2)	52
Mslr0063 (pilB1)	0
Mslr0079 (pilB2)	37
Mslr1274 (pilM)	0
Mslr1275 (pilN)	0
Mslr1276 (pilO)	0
Mslr1277 (pilQ)	0
Mslr0197 (comA)	0

^a Value (%) represents the frequency of transformation relative to wild type.

^b Zero means below the detection limit (0.1%).

Abbreviations

 
• !ORF

  open reading frame.