Characterization of Zoospore Type IV Pili in Actinoplanes missouriensis

Bacterial zoospores are interesting cells in that their physiological state changes dynamically: they are dormant in sporangia, show temporary mobility after awakening, and finally stop swimming to germinate in niches for vegetative growth. However, the cellular biology of a zoospore remains largely unknown. This study describes unprecedented zoospore type IV pili in the rare actinomycete Actinoplanes missouriensis. Similar to the case for the usual bacterial type IV pili, zoospore pili appeared to be retractable. Our findings that the zoospore pili have a functional role in the adhesion of zoospores to hydrophobic solid surfaces and that the zoospores use both pili and flagella properly according to their different purposes provide an important insight into the cellular biology of the zoospore.

ABSTRACT

The rare actinomycete Actinoplanes missouriensis produces terminal sporangia containing a few hundred flagellated spores. After release from the sporangia, the spores swim rapidly in aquatic environments as zoospores. The zoospores stop swimming and begin to germinate in niches for vegetative growth. Here, we report the characterization and functional analysis of zoospore type IV pili in A. missouriensis. The pilus gene (pil) cluster, consisting of three apparently σ^FliA-dependent transcriptional units, is activated during sporangium formation similarly to the flagellar gene cluster, indicating that the zoospore has not only flagella but also pili. With a new method in which zoospores were fixed with glutaraldehyde to prevent pilus retraction, zoospore pili were observed relatively easily using transmission electron microscopy, showing 6 ± 3 pili per zoospore (n = 37 piliated zoospores) and a length of 0.62 ± 0.35 μm (n = 206), via observation of fliC-deleted, nonflagellated zoospores. No pili were observed in the zoospores of a prepilin-encoding pilA deletion (ΔpilA) mutant. In addition, the deletion of pilT, which encodes an ATPase predicted to be involved in pilus retraction, substantially reduced the frequency of pilus retraction. Several adhesion experiments using wild-type and ΔpilA zoospores indicated that the zoospore pili are required for the sufficient adhesion of zoospores to hydrophobic solid surfaces. Many zoospore-forming rare actinomycetes conserve the pil cluster, which indicates that the zoospore pili yield an evolutionary benefit in the adhesion of zoospores to hydrophobic materials as footholds for germination in their mycelial growth.

IMPORTANCE Bacterial zoospores are interesting cells in that their physiological state changes dynamically: they are dormant in sporangia, show temporary mobility after awakening, and finally stop swimming to germinate in niches for vegetative growth. However, the cellular biology of a zoospore remains largely unknown. This study describes unprecedented zoospore type IV pili in the rare actinomycete Actinoplanes missouriensis. Similar to the case for the usual bacterial type IV pili, zoospore pili appeared to be retractable. Our findings that the zoospore pili have a functional role in the adhesion of zoospores to hydrophobic solid surfaces and that the zoospores use both pili and flagella properly according to their different purposes provide an important insight into the cellular biology of the zoospore.

INTRODUCTION

Pili, also referred to as fimbriae, are hairlike nonflagellar appendages that are located over cell surfaces in a wide range of bacteria. These dynamic extracellular organelles range between 5 and 9 nm in diameter and can reach several micrometers in length. There are five known types of pili in bacteria: chaperone-usher pili, type IV pili, conjugative type IV secretion pili, curli fibers, and type V pili (1). Type IV pili serve diverse functions, including motility along solid surfaces, adhesion to host cells, microcolony or biofilm formation, electron transfer, and DNA uptake. For example, the Gram-negative deltaproteobacterium Myxococcus xanthus relies on type IV pili for social motility and for fruiting body and biofilm formation (2). The human pathogen Neisseria gonorrhoeae moves on surfaces by attaching and retracting type IV pili (3). Among Gram-positive bacteria, on the other hand, mutants lacking type IV pili have been shown to be deficient in twitching and gliding motility in Clostridium difficile and Clostridium perfringens, respectively (4, 5). These functions are dependent on the three basic activities of type IV pili: (i) extension, i.e., lengthening the pilus through the polymerization of filament-constituting pilin subunits, (ii) adhesion, i.e., the ability of pilus subunits to bind to target surfaces or specific biomolecules, and (iii) retraction, i.e., shortening the pilus through pilin depolymerization (3, 6).

The type IV pilus system is similar to the type II secretion system, which translocates folded proteins from the periplasm into the extracellular environment in Gram-negative bacteria (1, 7). In the type IV pilus system, the substrates for translocation are the pilin subunits. The prepilin protein encoded by pilA has the following three characteristic structures: (i) a signal peptide; (ii) a recognition site for a prepilin peptidase, GFXXXE (where X is any amino acid); and (iii) an N-terminal transmembrane-like α-helix (8). In the first step of type IV pilus biogenesis, prepilin subunits are inserted into the plasma membrane by the Sec machinery. The signal peptide is then cleaved off in the membrane by the prepilin peptidase PilD (9, 10). During the elongation of pilus filaments, the mature pilin subunits are extracted from the membrane and incorporated into the base of growing filaments by the type IV pilus biogenesis machinery, which is composed of several subcomplexes (11). The motor subcomplex is composed of the membrane protein PilC and the cytoplasmic ATPases PilB and PilT, which are responsible for pilus elongation and retraction, respectively (3, 12). In Gram-negative bacteria, the alignment subcomplex, which is composed of PilM, PilN, PilO, and PilP, bridges the motor subcomplex and the outer membrane secretin subcomplex, which is a gated pore for pilus assembly and disassembly (13, 14). The final structural component of the type IV pilus system is the helical pilus filament, which is composed of a major pilin subunit. In some bacteria, minor pilin subunits and adhesins are also components of the filament (15).

Actinomycetes are high-GC Gram-positive, mainly soil-inhabiting bacteria. Many of them show filamentous growth and are often characterized by complex morphological development (16–19). Streptomyces is the most representative genus, isolated with very high frequency, and actinomycetes (especially filamentous actinomycetes) other than those in the genus Streptomyces are often called rare actinomycetes. Members of the genus Actinoplanes are rare actinomycetes with remarkable morphological development. They form a substrate mycelium from a germinating spore and subsequently produce terminal sporangia growing from the substrate mycelium through short sporangiophores (20). Terminal sporangia contain flagellated spores and open up to release the spores in response to water. This process is referred to as sporangium dehiscence (21, 22). Spores are termed zoospores after release from the sporangia, because they can swim in aquatic environments and show chemotactic properties toward various substances. When reaching a niche suitable for vegetative growth, zoospores stop swimming and begin to germinate (23, 24). Although they have such an interesting life cycle, molecular biological studies on their morphological development have been very limited.

The extensively characterized species Actinoplanes missouriensis produces terminal sporangia in a round shape when cultivated on humic acid-trace element (HAT) agar. Each sporangium contains a few hundred spherical flagellated spores (25). Recently, we revealed that the transcriptional regulator TcrA globally controls sporangium formation, spore dormancy, and sporangium dehiscence in A. missouriensis (26). TcrA is predominantly produced during sporangium formation and activates the transcription of genes responsible for the developmental processes by binding to the 21-bp direct repeat sequence 5′-NNGCA(A/C)CCG-N₄-GCA(A/C)CCGN-3′ (TcrA box) (N, any nucleotide). Based on comparative RNA sequencing (RNA-Seq) analysis comparing wild-type and isogenic tcrA null (ΔtcrA) mutant strains, we listed a total of 263 genes whose transcription is downregulated in the ΔtcrA mutant (26). We found a gene cluster consisting of nine genes that encode putative type IV pilus system components in this TcrA-dependent gene list, which suggests that the zoospores of A. missouriensis are piliated (26). To the best of our knowledge, pili are unprecedented in the cellular biological studies of a zoospore. In this study, we genetically analyzed the pilus gene cluster and functionally characterized A. missouriensis zoospore pili.

RESULTS

In silico analysis of the A. missouriensis type IV pilus gene cluster.

Among the 263 genes whose transcript levels were downregulated (over 4.0-fold) in the ΔtcrA mutant, we found nine consecutive genes, AMIS_8980 to AMIS_9060, that constitute a putative pilus gene (pil) cluster. AMIS_9030 is annotated as pilA, which encodes a structural component of pilus fiber, and the deduced amino acid sequences of five other gene products are homologous to those of the type IV pilus system components in C. difficile (Table 1). Thus, this gene cluster encodes sufficient proteins for assembling a functional type IV pilus of Gram-positive bacteria: PilA (prepilin), PilB (ATPase for pilus elongation), PilC (integral membrane protein required for pilus elongation), PilD (prepilin peptidase), PilM and PilO (components of the alignment subcomplex), and PilT (ATPase for pilus retraction). One of the remaining two genes, AMIS_9000, is predicted to encode a minor pilin subunit (see Discussion). The function of the other gene product, AMIS_8980, is unknown because this protein shows no significant homology with any characterized proteins (Table 1).

TABLE 1.

Genes in the pil cluster in A. missouriensis

Gene ID (AMIS no.)	Length (amino acids)	Gene product	Putative function	Homolog in C. difficile	Identity/similarity (%)
8980	104	—a	Unknown
8990	204	PilOb	Component of alignment subcomplex
9000	232	—	Unknown
9010	341	PilM	Component of alignment subcomplex	CD630_32930	13/36
9020	259	PilD	Prepilin peptidase	CD630_35040	28/55
9030	158	PilA	Prepilin	CD630_32940	33/55
9040	410	PilC	Component of motor subcomplex	CD630_35110	26/56
9050	372	PilT	ATPase for pilus retraction	CD630_35050	47/70
9060	563	PilB	ATPase for pilus elongation	CD630_32960	37/62

^a—, the protein exhibits no significant sequence homology to any characterized protein.

^bThe gene product is homologous to PilO in Neisseria meningitidis (17% and 45% identity and similarity, respectively). There is no PilO homolog in C. difficile R20291.

The genome sequences of 16 Actinoplanes species, including A. missouriensis, have been registered in the NCBI genome database with gene annotation (https://www.ncbi.nlm.nih.gov/genome/). We searched for homologous genes of the pil cluster using the A. missouriensis pil genes as queries and found that all 16 Actinoplanes species have a very similar pilus gene cluster. The gene organization of the pil cluster in Actinoplanes lutulentus is the same as that in A. missouriensis (Fig. 1). The pil clusters of the other 14 Actinoplanes species have the same gene organization as the A. missouriensis pil cluster except that all of them lack an AMIS_8980 ortholog. Furthermore, gene clusters showing high homology with the A. missouriensis pil cluster were found in other zoospore-forming rare actinomycetes. Couchioplanes caeruleus and Catenuloplanes japonicus, both of which produce segmental motile spores, have a gene cluster that is quite similar to the A. missouriensis pil cluster; they also lack an AMIS_8980 ortholog (Fig. 1). Dactylosporangium aurantiacum, which produces motile sporangiospores, also has a gene cluster showing high homology to the A. missouriensis pil cluster; it has an additional gene between the pilM and AMIS_9000 orthologs. Partially homologous gene clusters are also carried on the actinobacteria genomes of the members of the genera Planomonospora, Kineosporia, and Spirillospora, all of which produce motile sporangiospores, and in the genera Kineococcus (motile cocci) and Angustibacter (nonmotile cocci to rods). In contrast, no pil cluster was found in Streptomyces species. These results indicate that the type IV pilus system is evolutionarily conserved among zoospore-producing rare actinomycetes.

FIG 1.

Gene organization of the pil clusters in A. missouriensis (A), A. lutulentus (B), and C. japonicus (C). Arrows indicate the locations of the open reading frames, including their lengths and directions. Gene identifiers (IDs) and names are shown above and below the arrows, respectively. Gene names are not shown when the functions of the gene products are unknown.

Transcriptional analysis of the pil cluster.

Previously, we performed an RNA-Seq analysis to obtain transcriptional profiles during sporangium formation using the total RNA extracted from wild-type cells cultivated for 1, 3, 6, and 40 days on HAT agar (27). On this agar plate, small sporangium-like structures were observed after 2 or 3 days of cultivation, and then mature sporangia that can release spores under dehiscence-inducing conditions were formed after 5 to 7 days of incubation. Based on these results, the transcriptional profile of the pil cluster is presented in Fig. 2. All genes of the pil cluster were scarcely transcribed on day 1. However, transcription was significantly activated on day 3. The transcript levels of the pilA- and pilT-containing operons increased substantially with the analyzed time course just like those of group iv genes in the flagellar gene cluster (27), while the transcripts of pilB were detected at nearly equal levels on days 3, 6, and 40, as for group iii genes in the flagellar gene cluster (27). In this way, both the flagellar and pil gene clusters are actively transcribed during sporangium formation and seem to be controlled by common regulatory mechanisms. This transcriptional profile of the pil cluster indicates that components for pilus biogenesis are produced during sporangium formation. Pilus structures are presumably assembled in the process of sporangium formation.

FIG 2.

Transcriptional profile of the pil cluster during sporangium formation. Distributions of the mapped RNA-Seq read counts in the 1-, 3-, 6-, and 40-day cultures are denoted by lines with different colors. All genes of the cluster were hardly transcribed on day 1. Bold arrows indicate open reading frames. The arrowhead indicates a tRNA gene. The read counts mapped to the tRNA gene were eliminated from the profile. The transcriptional start points are shown by bent arrows. The three major transcriptional units are shown by blue arrows.

RNA-Seq analysis also clarified the transcriptional units of the pil genes; the pil cluster consists of three major transcriptional units, as shown in Fig. 2. While pilB is transcribed as a monocistronic transcript, the other genes are transcribed as polycistronic operons. We determined transcriptional start points in the pil cluster (Fig. 2). The transcriptional start site of pilB was determined to be the first nucleotide of the translational start codon by using the 5′ rapid amplification of cDNA ends (5′-RACE) procedure, whereas the transcriptional start points upstream from the pilT- and pilA-containing operons were determined using high-resolution S1 nuclease mapping. The transcriptional start points of the pilT-containing operon were 28 and 29 nucleotides upstream from the start codon of pilT, and those of the pilA-containing operon were 229 and 231 nucleotides upstream from the start codon of pilA (see Fig. S1 in the supplemental material). Alignment of the promoter regions upstream from the transcriptional start sites revealed that all of the pil genes shared similar promoter elements that were in accord with the putative σ^FliA-dependent promoter, 5′-CTCA-(n_15–17)-GCCGA(A/T)-3′ (26) (Fig. 3). This result suggests that a σ^FliA-family sigma factor(s) initiates the transcription of all genes in the pil cluster (see Discussion).

FIG 3.

Sequence alignment of promoter regions upstream from the transcriptional start points in the pil cluster. The first genes downstream of each transcriptional start site are shown on the left side. The putative σ^FliA-dependent promoter element is shown below the alignment. Conserved promoter elements are shaded. The transcriptional start points are shown with bent arrows. n, any nucleotide; W, A or T.

Observation of zoospore pili.

To examine whether zoospores are piliated, a transmission electron microscopic (TEM) analysis was performed to observe zoospores released from sporangia of the wild-type strain by using a negative-staining method for flagellar observation (21, 27). At first, we were unable to observe any pili, while many flagella were clearly observed. After repeated experiments, we obtained only one picture that shows the presence of pilus filaments extending from the surface of a zoospore (Fig. 4A to C). The pilus filaments are thinner (diameter, approximately 5 nm) than flagellar filaments (approximately 13 nm) and appear to be somewhat linear compared with the gently curved flagellar filaments (Fig. 4A to C). The extremely low probability of pilus observation prompted us to examine a new method for the observation of zoospore pili. Based on the assumption that zoospores should withdraw their pili in response to physical contact with the grids for TEM observation, we fixed zoospores with glutaraldehyde before putting them on the grids. This new method drastically increased the frequency of pilus observation (up to 20% of the observed zoospores), and although the images were rather less clear (Fig. 4D and E), we used this method thereafter in this study unless otherwise mentioned.

FIG 4.

Observation of the wild-type zoospores by TEM. Panels B, C, and E are enlarged views of portions of panels A, B, and D, respectively. Arrowheads indicate the pilus filaments. The thick filaments are flagella. A conventional method for flagellar observation (21, 27) was used for panels A, B, and C, while a new method was used for panels D and E. Scale bars, 500 nm.

To clearly show that the observed thin filaments around the cell surfaces are distinct from flagellar filaments, we observed zoospores of the flagellin-encoding fliC deletion (ΔfliC) mutant constructed in our previous study (27). TEM observation of the nonmotile ΔfliC mutant zoospores revealed that they were not flagellated but were piliated (Fig. 5A and B). Furthermore, we constructed a pilA deletion (ΔpilA) mutant to show that the observed thin filaments are produced by the type IV pilus system encoded by the pil cluster. The ΔpilA mutant grew and formed sporangia normally, and the sporangia released zoospores by normal dehiscence. We did not observe any pilus structures on the ΔpilA mutant zoospores with TEM analysis (Fig. 5C), confirming that pilA encodes the prepilin subunit. The piliation of the zoospores was restored by the introduction of the pilA gene with its own promoter region into the ΔpilA mutant (Fig. 5D and E [this picture was taken by using the conventional method without glutaraldehyde fixation]). Next, we generated a pilT deletion (ΔpilT) mutant to reduce the frequency of the pilus retraction of the zoospores. The ΔpilT mutant also formed sporangia normally, and the sporangia released zoospores under dehiscence-inducing conditions. As expected, the pilT deletion improved the frequency of pilus observation; without glutaraldehyde treatment, we were able to observe piliated zoospores with a much higher frequency (more than 50%) in our TEM analysis (Fig. 5F and G), indicating that the pilT gene product is responsible for the pilus retraction.

FIG 5.

(A to G) Observation of the ΔfliC (A and B), ΔpilA (C), pilA-complemented ΔpilA/pilA⁺ (D and E), and ΔpilT (F and G) zoospores by TEM and the numbers and lengths of pili in the ΔfliC (H and I) and ΔpilT (J) mutants. The ΔpilA/pilA⁺ strain harbors the pilA complementation plasmid on the chromosome. Panels B, E, and G are enlarged views of portions of panels A, D, and F, respectively. Arrowheads indicate pilus filaments. The thick filaments in panels C to G are flagella. Scale bars, 500 nm. (H) Distribution of pilus number per piliated ΔfliC mutant zoospore (n = 37). (I) Distribution of length of ΔfliC zoospore pili (n = 206). (J) Distribution of pilus number per piliated or apparently nonpiliated ΔpilT mutant zoospore (n = 21) and distribution of length of ΔpilT zoospore pili (n = 69).

We counted the pili extending from each zoospore of the ΔfliC mutant in the electron microscopic images. Zoospores with five pili were most abundant; the average (±standard deviation) was 6 ± 3 pili per zoospore (n = 37 piliated zoospores) (Fig. 5H). It should be noted that some zoospores were observed to be nonpiliated, presumably owing to the retraction of pili and limitation of the method, and these apparently nonpiliated zoospores were excluded from the calculation. The largest number of pili observed on a zoospore was 13. In addition, we measured the length of each pilus in the images using ImageJ (http://rsb.info.nih.gov/ij/) (28). The average (±standard deviation) pilus length was 0.62 ± 0.35 μm (n = 206) (Fig. 5I). The length of the longest pilus observed in the microscopic images was 2.49 μm. Furthermore, we also counted the pili of the ΔpilT mutant zoospores in the TEM images without glutaraldehyde fixation. Because the frequency of the pilus observation was improved in the ΔpilT mutant zoospores, nonpiliated zoospores were included in the calculation. The average number of pili (±standard deviation) was 3 ± 3 pili per zoospore (n = 21), and the average (±standard deviation) pilus length was 0.45 ± 0.22 μm (n = 69) (Fig. 5J). It should be noted that the ΔpilT zoospores have both flagella and pili (see Discussion). Interestingly, the pilus filaments appeared to extend from only a restricted area of the zoospore surface (Fig. 5A) (see Discussion).

Zoospore pili are required for adhesion to hydrophobic solid surfaces.

As mentioned in the introduction, bacterial type IV pili have been reported to serve fundamental functions in diverse cellular processes. In A. missouriensis, we hypothesized that the zoospore pili have an important function in adhesion to solid surfaces upon the cessation of swimming behavior before the onset of germination. Thus, we analyzed the ability of zoospores to adhere to the surface of a plastic dish. The average proportion (±standard deviation) of the zoospores that adhered to the dish surface was 40.6% ± 2.5% in the wild-type strain (Fig. 6A). In the ΔpilA mutant, however, it was reduced to only 3.9% ± 1.0%, indicating that zoospore pili are required for sufficient adhesion to the surface of the plastic dish under the test conditions (Fig. 6A). The proportion of the ΔpilA mutant zoospores that adhered to the dish surface was partially restored to 20.2% ± 1.7% by the introduction of the pilA gene with its own promoter region (Fig. 6A).

FIG 6.

Zoospore adhesion to solid surfaces. (A) Proportions of zoospores adhering to the hydrophobic plastic (polystyrene) surface. Data are the mean values from three biological replicates ± standard deviations. Microscopic images of the adhesive zoospores are shown below the graph. Scale bars, 10 μm. (B) Proportions of zoospores of the wild-type strain adhering to the hydrophilic glass and BSA-coated plastic surfaces.

We also examined the ability of zoospores to adhere to a hydrophilic glass surface. The proportion of the wild-type zoospores that adhered to the glass surface was only 7.2% ± 1.3% under the same conditions as for the plastic dish (Fig. 6B), suggesting that zoospores can adhere predominantly to hydrophobic solid surfaces. We speculate that a small proportion of zoospores adhered to the glass surface through flagella. To analyze the hypothesis that zoospores adhere predominantly to hydrophobic solid surfaces, we further examined whether zoospores adhere to the surface of a plastic dish treated with 1% (wt/vol) bovine serum albumin (BSA) solution. As expected, only 0.2% ± 0.1% zoospores adhered to the BSA-treated dish surfaces under the test conditions, indicating that the zoospore pili possess a much higher affinity to hydrophobic solid surfaces than to hydrophilic ones (Fig. 6B).

DISCUSSION

In this study, we successfully identified and characterized unprecedented zoospore type IV pili of A. missouriensis. We also demonstrated the property of adhesion of the A. missouriensis zoospore to hydrophobic surfaces. The zoospore type IV pili were indicated to play a pivotal role in the adhesion of the zoospore. This function is expected for bacterial type IV pili. However, when we consider the unique features of the zoospore, our findings become more important. Zoospores of filamentous actinomycetes are highly differentiated cells that aim for the rapid expansion of their habitat; they can swim far away to seek niches and settle themselves in the niches to grow as mycelia. Flagella and pili are required for swimming and solid surface adhesion (i.e., the initiation of colonization), respectively, both of which are very important in the biology of the zoospore. Therefore, to understand the physiology of the zoospore, we should pay attention not only to flagella but also to pili.

While type IV pilus filaments are composed of repeating units of the major pilin, other proteins such as minor pilins and adhesins can be incorporated into the filament in some cases. For example, the pilus filament of Pseudomonas aeruginosa is composed of major (PilA) and several minor (PilE, PilV, PilW, PilX, and FimU) pilin subunits and adhesion molecules such as PilY1 (29). One of the minor pilins is an initiator pilin that forms the template upon which the polymerization of major pilins begins. The initiator pilins lack a conserved Glu residue at position 6 in the recognition site for the prepilin peptidase because this Glu residue forms a salt bridge with the N terminus of the previously incorporated pilin subunit and hence is not required for the first pilin subunit. Furthermore, the initiator pilins are larger than their cognate major pilins (15). In the pil cluster of A. missouriensis, AMIS_9000 encodes a protein consistent with these two features, suggesting that the gene product is the initiator pilin located at the tip of the pilus filament. For AMIS_8980, the precise function of the gene product remains elusive. Considering that homologs of AMIS_8980 are not found in the type IV pilus gene clusters of most Actinoplanes species and many other rare actinomycetes, we postulate that the gene product is not required for the pilus biogenesis.

Putative σ^FliA-dependent promoters were identified in the upstream regions from all three transcriptional start points of the pil genes (Fig. 3). In a previous study, we revealed that TcrA activates transcription of the genes involved in sporangium formation, spore dormancy, sporangium dehiscence, flagellar biogenesis, and chemotaxis (26). Considering that three σ^FliA-family sigma factors, σ^FliA1, σ^FliA2, and σ^FliA3, are under transcriptional control of TcrA, we predict that one or more of these sigma factors are responsible for transcription in the pil cluster, leading to lower transcript levels of the pil genes in the ΔtcrA mutant than in the wild-type strain (26). Detailed functional analysis of the σ^FliA-family sigma factors is in progress and will be published elsewhere.

In the adhesion test, the proportion of adhesive zoospores of the pilA complemented strain was not fully restored to that of the wild-type strain (Fig. 6A). We speculate that the transcription level of pilA in the pilA complemented strain is lower than that in the wild-type strain and that this explains the partial restoration of adhesion activity. The transcriptional profile of the pil cluster indicates that pilA is transcribed not only from its own promoter but also as a read-through from the pilT-containing operon (Fig. 2). However, the pilA gene is transcribed only from its own promoter in the pilA complemented strain, which probably results in an insufficient transcript level of pilA.

In the ΔpilT zoospore, the frequency of the pilus retraction was greatly reduced; we could observe piliated zoospores of the ΔpilT mutant without glutaraldehyde treatment with high frequency (more than 50%). This enabled us to count the pili and determine their length in the flagellated zoospores, because it was very difficult to observe pili on the flagellated wild-type zoospore with glutaraldehyde fixation, which reduced the clearness of the TEM images. The numbers of pili per piliated zoospore and the average length of observed pili were not so varied between nonflagellated (ΔfliC) and flagellated (ΔpilT) zoospores. Apparently, this result suggests no link between flagellation and piliation. However, it may be somewhat strange that only the frequency of pilus observation was greatly improved and hyperpiliation (in number and/or length) was not induced by the deletion of pilT. Further investigation of the possible link between flagellation and piliation, including their specific locations (see below), is our future research subject.

We previously reported that flagellar formation was observed in a restricted area of the zoospore surface in A. missouriensis (27). In this study, pilus filaments were also often observed in a restricted area of the zoospore surfaces. Interestingly, some TEM images indicate that flagella and pili are extended from areas that are opposite each other (Fig. 4D and E). The molecular mechanism and physiological role of the specific localization of flagella and pili remain elusive. However, we think that possible different localization of flagella and pili is very reasonable, because pili would be an obstacle for flagellar rotation if the pili and flagella extended from the same region. Our ongoing studies on the zoospore’s swimming behavior are also important for understanding the zoospore biology. We believe that our multilateral analyses of the A. missouriensis zoospore are revealing the molecular mechanisms of the species’ characteristic survival strategy.

MATERIALS AND METHODS

Bacterial strains, plasmids, media, and primers.

A. missouriensis 431^T (NBRC 102363^T) was obtained from the National Institute of Technology and Evaluation (NITE), Chiba, Japan. A. missouriensis was grown on YBNM (yeast extract-meat extract-NZ amine-maltose) or HAT agar at 30°C for a solid culture and in PYM (peptone-yeast extract-MgSO₄) broth at 30°C for a liquid culture, as previously described (30). MS (2% mannitol and 2% soy flour) or modified ISP4 (ISP medium 4 [Difco] supplemented with 0.05% yeast extract [Difco] and 0.1% tryptone [Difco]) agar medium was used for transformation by conjugation with Escherichia coli ET12567(pUZ8002). MS and modified ISP4 agar media were supplemented with MgCl₂·6H₂O at a final concentration of 40 mM. MS agar was used for the construction of the recombinant strain for the complementation test, and modified ISP4 agar was used for the construction of the ΔpilA mutant. E. coli ET12567(pUZ8002) was obtained from the John Innes Centre (Norwich, UK) and used as the donor in intergeneric conjugation. E. coli JM109 and pUC19 were purchased from TaKaRa Biochemicals (Shiga, Japan). The media and growth conditions for E. coli were as described by Maniatis et al. (31). Apramycin (50 μg/ml), spectinomycin (50 μg/ml), and ampicillin (50 μg/ml) were added when necessary. The primers used in this study are listed in Table S1 in the supplemental material.

RNA extraction.

A. missouriensis cells for RNA extraction were prepared as previously described (26). Cells were disrupted by rubbing with a mortar and pestle, and the cell lysate was mixed with the lysis/binding solution of the RNAqueous total RNA isolation kit (Thermo Fisher Scientific, Waltham, MA). After the debris was removed by centrifugation at 21,000 × g for 5 min, total RNAs were extracted according to the manufacturer’s instructions. The total RNAs were treated with DNase I to eliminate contaminating genomic DNA and purified by phenol-chloroform extraction and ethanol precipitation.

S1 nuclease mapping.

S1 nuclease mapping was performed using a method described by Bibb et al. (32) and Kelemen et al. (33). Hybridization probes were prepared using PCR and labeled at both 5′ ends with [γ-³²P]ATP (220 TBq/mmol) using T4 polynucleotide kinase. Labeling at one side of the 5′ ends was eliminated by restriction enzyme digestion. For hybridization, 40 μg of total RNA was used. Protected fragments were analyzed on 6% polyacrylamide DNA sequencing gels according to the method of Maxam and Gilbert (34).

5′-RACE.

Mapping of the 5′ end was carried out using a Full RACE Core Set (TaKaRa Biochemicals) according to the manufacturer’s instructions. The PCR products were cloned into pUC19 and sequenced by Fasmac Corp. (Kanagawa, Japan).

TEM observation.

Zoospores were released from the sporangia by pouring 10 ml of 25 mM NH₄HCO₃ onto one HAT plate and incubating the plate at 30°C for 1 h. After being collected from the plate, zoospores were observed by TEM according to the method reported previously (21, 27). In this method, the zoospore samples were negatively stained with 1% (wt/vol) phosphotungstic acid (pH 7.0) and observed with a JEM-1010 electron microscope (Jeol, Tokyo, Japan) using Formvar-coated copper grids. As depicted in Fig. 4D and E and Fig. 5A to C, we used a new method as follows. The zoospore samples were incubated in 1% (wt/vol) glutaraldehyde solution for a few minutes at room temperature. These samples were centrifuged, and the supernatants were removed. The zoospore pellets were resuspended in 25 mM NH₄HCO₃. The samples were negatively stained with 2% (wt/vol) ammonium molybdate and observed with a JEM-1400 electron microscope using carbon-coated copper grids, as described previously (35).

Construction of the ΔpilA and ΔpilT mutant strains.

For construction of ΔpilA and ΔpilT mutant strains, 2.5-kbp upstream and downstream regions of pilA and pilT were amplified by PCR. The amplified DNA fragments for pilA were cloned into pUC19 digested with XbaI and HindIII using the In-Fusion HD cloning kit (TaKaRa Biochemicals) according to the manufacturer’s instructions, generating pUC19-ΔpilA. The amplified upstream and downstream fragments for pilT were digested with EcoRI plus XbaI and with XbaI plus HindIII, respectively, and cloned together into pUC19 digested with EcoRI and HindIII, generating pUC19-ΔpilT. Plasmids pUC19-ΔpilA and pUC19-ΔpilT were sequenced to confirm that no PCR-derived error was introduced. pUC19-ΔpilA and pUC19-ΔpilT were then digested with XbaI plus HindIII and with EcoRI plus HindIII, respectively, and the insert fragments were cloned into pK19mobsacB digested with the same restriction enzymes, whose kanamycin resistance gene had been replaced with the apramycin resistance gene aac(3)IV (30), generating pK19mobsacB-ΔpilA and pK19mobsacB-ΔpilT. Plasmids pK19mobsacB-ΔpilA and pK19mobsacB-ΔpilT were introduced into A. missouriensis by conjugation as described previously (27). Apramycin-resistant colonies resulting from a single-crossover recombination were isolated. One of them was cultivated in PYM liquid medium for 2 days, and the mycelia, suspended in a 0.75% NaCl solution, were spread onto Czapek-Dox broth agar medium (BD, Franklin Lakes, NJ) containing extra sucrose (final concentration, 5%). After incubation at 30°C for 5 to 7 days, the sucrose-resistant colonies were inoculated onto YBNM agar with or without apramycin to confirm that they were sensitive to apramycin. The apramycin-sensitive and sucrose-resistant colonies resulting from the second crossover recombination were isolated as candidates for the ΔpilA and ΔpilT mutant strains. The disruption of pilA and pilT was confirmed by PCR (data not shown).

Construction of the recombinant strain for complementation testing.

A 0.9-kbp DNA fragment containing the promoter and coding sequences of pilA was amplified by PCR. The amplified fragment was cloned into pTYM19-Apra (26) digested with EcoRI and HindIII using the In-Fusion HD cloning kit, resulting in pTYM19-Apra-pilA. Plasmid pTYM19-Apra-pilA was sequenced to confirm that there was no PCR-derived error and was then introduced into the ΔpilA mutant by conjugation as described previously (27). Apramycin-resistant colonies were obtained.

Zoospore adhesion test.

A cover glass was put on a plastic (polystyrene) dish (Iwaki number 1000-035; AGC, Shizuoka, Japan) using two narrow double-sided tapes that were arranged at both side edges as parallel lines to seal the edges, and a zoospore-containing solution (approximately 10⁴ cells/μl) was poured into the space between the dish and cover glass. After 10 min, the whole zoospores were photographed with a high-speed camera by scanning the microscopic fields along the vertical direction. Then, to remove zoospores that did not adhere to the solid surfaces, 25 mM NH₄HCO₃ was poured into the chamber from one side, and the overflowed solution was absorbed with a paper filter on the other side. The images of the zoospores that adhered to the dish surface were then recorded. From the microscopic images, the adhesion ratios (proportions of adhesive zoospores to whole zoospores) were calculated. The whole zoospores in the solution were counted using the Color_Footprint plug-in for ImageJ (36). A glass dish (Iwaki number 3970-035; AGC) was used to test the adhesion to the hydrophilic glass surface. For the BSA coating, the plastic dish was treated with 1% (wt/vol) BSA for a few minutes at room temperature and washed with 25 mM NH₄HCO₃ twice.

Optical microscopy.

Cells were visualized under a phase-contrast microscope (IX73; Olympus, Tokyo, Japan) equipped with an objective lens (LUCPLFLN 20×PH; Olympus), a complementary metal-oxide semiconductor (CMOS) camera (DMK33UX174; Imaging Source, Bremen, Germany), and an optical table (HAX-0605; JVI, Shizuoka, Japan). For high-speed imaging, a lab recorder system (LRH1540; Digimo, Tokyo, Japan) was used at a speed of 200 frames per second. The cell images were captured as 8-bit images and converted into TIF files without compression. All data were analyzed using ImageJ and its plug-ins.