Regulation of Sporangium Formation, Spore Dormancy, and Sporangium Dehiscence by a Hybrid Sensor Histidine Kinase in Actinoplanes missouriensis: Relationship with the Global Transcriptional Regulator TcrA

Actinoplanes missouriensis goes through complex morphological differentiation, including formation of flagellated spore-containing sporangia, sporangium dehiscence, swimming of zoospores, and germination of zoospores to filamentous growth. Although the orphan response regulator TcrA globally activates many genes required for sporangium formation, spore dormancy, and sporangium dehiscence, its partner histidine kinase remained unknown. Here, we analyzed the function of an orphan hybrid histidine kinase, HhkA, and proposed that HhkA constitutes a cognate two-component regulatory system with TcrA. That HhkA and TcrA homologues are highly conserved among the genus Actinoplanes and several closely related rare actinomycetes indicates that this possible two-component regulatory system is employed for complex morphological development in sporangium- and/or zoospore-forming rare actinomycetes.

ABSTRACT

The rare actinomycete Actinoplanes missouriensis forms terminal sporangia containing a few hundred flagellated spores. In response to water, the sporangia open and release the spores into external environments. The orphan response regulator TcrA functions as a global transcriptional activator during sporangium formation and dehiscence. Here, we report the characterization of an orphan hybrid histidine kinase, HhkA. Sporangia of an hhkA deletion mutant contained many distorted or ectopically germinated spores and scarcely opened to release the spores under sporangium dehiscence-inducing conditions. These phenotypic changes are quite similar to those observed in a tcrA deletion mutant. Comparative RNA sequencing analysis showed that genes controlled by HhkA mostly overlap TcrA-regulated genes. The direct interaction between HhkA and TcrA was suggested by a bacterial two-hybrid assay, but this was not conclusive. The phosphorylation of TcrA using acetyl phosphate as a phosphate donor markedly enhanced its affinity for the TcrA box sequences in the electrophoretic mobility shift assay. Taking these observations together with other results, we proposed that HhkA and TcrA compose a cognate two-component regulatory system, which controls the transcription of the genes involved in many aspects of morphological development, including sporangium formation, spore dormancy, and sporangium dehiscence in A. missouriensis.

IMPORTANCE Actinoplanes missouriensis goes through complex morphological differentiation, including formation of flagellated spore-containing sporangia, sporangium dehiscence, swimming of zoospores, and germination of zoospores to filamentous growth. Although the orphan response regulator TcrA globally activates many genes required for sporangium formation, spore dormancy, and sporangium dehiscence, its partner histidine kinase remained unknown. Here, we analyzed the function of an orphan hybrid histidine kinase, HhkA, and proposed that HhkA constitutes a cognate two-component regulatory system with TcrA. That HhkA and TcrA homologues are highly conserved among the genus Actinoplanes and several closely related rare actinomycetes indicates that this possible two-component regulatory system is employed for complex morphological development in sporangium- and/or zoospore-forming rare actinomycetes.

INTRODUCTION

Two-component regulatory systems are widespread in microorganisms. A prototypical two-component regulatory system consists of a sensor histidine kinase and its cognate response regulator (1, 2). The sensor histidine kinase comprises a set of sensor and kinase domains and displays considerably wide architectural varieties. The kinase domain is further dissected into the dimerization and histidine phosphotransfer (DHp) domain and the catalytic ATP-binding (CA) domain. Meanwhile, the response regulator harbors a receiver domain, and the majority of response regulators have additional domains, such as an output effector domain, which often displays DNA-binding activity. In several cases, the effector domain shows RNA-binding activity or functions as an enzyme (3, 4). In addition, atypical single-domain response regulators such as CheY and Spo0F utilize the receiver domain for both input and output reactions and exert their function through protein-protein interactions (5, 6). The sensor histidine kinase perceives a wide variety of chemical and physical stimuli via its sensor domain, such as small molecules, ions, dissolved gases, pH, temperature, osmotic pressure, light, and the redox state of the cell; however, actual signals remain unknown for most sensor kinases (7, 8). Then, the sensor histidine kinase conveys the perceived environmental or physiological information to its partner response regulator, which triggers appropriate cellular responses, including cell division, metabolic modulation, morphological development, chemotaxis, antibiotic resistance, and pathogenicity (9). In response to the external stimuli, the sensor histidine kinase undergoes autophosphorylation, where the conserved His residue in the DHp domain is phosphorylated by the CA domain. Subsequently, the phosphoryl group on the His residue is transferred onto the conserved Asp residue in the receiver domain of its cognate response regulator, which allosterically modifies the activity of the response regulator. The phosphorylated response regulator induces the cellular responses through the transcriptional regulation of specific genes or protein-protein interaction. Among well-studied response regulators serving as transcriptional regulators are OmpR, NarL, and NtrC (10–12).

The hybrid sensor histidine kinase contains an extra region serving as the receiver domain in its amino acid sequence. In frequent cases, phosphotransfer reactions involving hybrid sensor histidine kinases are more complex than the phosphotransfer reaction in the typical two-component regulatory system (13). Following the autophosphorylation of the His residue in its kinase domain, the hybrid sensor histidine kinase transfers the phosphoryl group onto the Asp residue in its receiver domain. Then, a histidine phosphotransfer (Hpt) protein shuttles the phosphoryl group via its conserved His residue between the two Asp residues in the receiver domains of the hybrid sensor kinase and its partner response regulator. For example, four hybrid sensor histidine kinases, PA1611, PA1976, PA2824, and RetS, transfer the phosphoryl groups onto the response regulator PA3346 via the Hpt protein HptB in Pseudomonas aeruginosa PAO1 (14).

Actinomycetes are mainly soil-dwelling Gram-positive bacteria, many of which show filamentous growth. Filamentous actinomycetes are characterized by complex morphological development. In particular, morphological development of the most representative genus, Streptomyces, has been extensively studied (15, 16). Filamentous actinomycetes other than those of the genus Streptomyces are often called rare actinomycetes because they are isolated at low frequencies by conventional isolation methods. Members of the genus Actinoplanes are rare actinomycetes with complex morphological development, and Actinoplanes missouriensis is a well-characterized species in this genus. It forms a branched substrate mycelium during vegetative growth and subsequently produces globose or subglobose terminal sporangia that grow from the substrate mycelium through short sporangiophores on sporangium-forming humic acid-trace element (HAT) agar (17–19). Each sporangium contains a few hundred spherical flagellated spores, and spaces among spores inside a sporangium are filled with an intrasporangial matrix. Sporangia open and release the spores in response to external water through the process referred to as sporangium dehiscence (20, 21). Spores are termed zoospores after being released from sporangia because they can swim in aquatic environments using flagella. When they reach a niche suitable for vegetative growth, zoospores stop swimming and begin to germinate (22, 23). On HAT agar, small sporangium-like structures are observed after 2 or 3 days of cultivation. Then, mature sporangia that can release spores under sporangium dehiscence-inducing conditions are formed after incubation for 5 to 7 days (24, 25).

In A. missouriensis, an orphan response regulator, TcrA, globally activates many genes involved in sporangium formation, spore dormancy, and sporangium dehiscence by binding to a 21-bp direct repeat sequence (TcrA box), 5′-NNGCA(A/C)CCGN₄GCA(A/C)CCGN-3′ (26). We proved that Asp-52 in the receiver domain, which is the predicted phosphorylation site of TcrA, is indispensable for the in vivo function of TcrA by a gene complementation experiment with a tcrA-null (ΔtcrA) mutant using the wild-type and mutated (D52N) tcrA genes (26). This suggested that TcrA is phosphorylated at Asp-52 by an unknown sensor histidine kinase(s). Very recently, we reported that three FliA family sigma factor genes (fliA1, fliA2, and fliA3) are the direct targets of TcrA and that these three FliA proteins constitute a complex regulatory network for sporangium formation, spore dormancy, and sporangium dehiscence (27). In addition, we previously conducted a comparative proteome analysis using cellular proteins extracted from zoospores and synchronously germinating cells, and successfully identified 27 proteins whose amounts are larger in zoospores than in germinating cells (25). Here, we focused on AMIS_64930, one of those 27 proteins that seem to be unnecessary for filamentous growth. Because AMIS_64930 is predicted to be a hybrid histidine kinase, we named it HhkA. We assumed that HhkA is involved in the biology of either zoospores (swimming or chemotaxis behavior) or sporangia (sporangium formation, spore dormancy, or sporangium dehiscence). We conducted detailed characterization of HhkA and finally proposed that HhkA and TcrA make up a cognate two-component regulatory system that controls the reproduction process involving morphological differentiation in A. missouriensis.

RESULTS

HhkA is widely conserved among Actinoplanes bacteria and some other rare actinomycetes.

In our previous work, we identified 27 proteins that were at least three times as abundant in zoospores as in germinating cells (25). HhkA is one of those 27 proteins and is composed of 682 amino acids; although the annotated coding sequence of hhkA encodes a protein of 685 amino acids, we concluded in this study that the genuine HhkA is composed of 682 amino acids as described below. A protein database search using InterPro version 73.0 (http://www.ebi.ac.uk/interpro/) showed that HhkA harbors a signal sensor PAS domain (accession number IPR000014; residues 174 to 294), a histidine kinase domain (IPR005467; residues 315 to 535), and a signal transduction response regulator receiver domain (IPR001789; residues 552 to 668). Therefore, HhkA is predicted to function as a hybrid sensor histidine kinase in a two-component regulatory system. Because no signal peptide or transmembrane region is predicted by the SignalP5.0 server (http://www.cbs.dtu.dk/services/SignalP/) or the SOSUI engine version 1.11 (http://harrier.nagahama-i-bio.ac.jp/sosui/sosui_submit.html), respectively, HhkA is predicted to be a cytoplasmic protein. Although two-component system histidine kinases are typically encoded in the vicinity of their cognate response regulator genes, they are also found encoded by orphan genes (28–30). While a total of 87 two-component system histidine kinase and response regulator gene pairs are annotated on the A. missouriensis genome sequence, 50 and 18 orphan genes also encode putative two-component system histidine kinases and response regulators, respectively. In addition, two and four genes encoding putative histidine kinase CheA and response regulator CheY proteins, respectively, and a gene encoding a putative CheA-CheY fusion protein were found in three putative chemotaxis-related gene clusters. We consider HhkA an orphan histidine kinase, because neither a response regulator nor an Hpt protein is encoded in the vicinity of hhkA; genes encoding a putative ribose-phosphate pyrophosphokinase and a putative serine/threonine protein kinase are located upstream and downstream, respectively, of hhkA. His-318 in the histidine kinase domain and Asp-603 in the response regulator receiver domain are probable phosphorylation sites of HhkA; His-318 and Asp-603 are predicted to be phosphorylated by the autophosphorylation and phosphotransfer reactions, respectively (see Fig. S1 in the supplemental material).

We performed a BLAST search in the NCBI genome database (https://www.ncbi.nlm.nih.gov/genome/) and found that HhkA homologues are highly conserved (E values < e⁻¹⁵²) in all 22 Actinoplanes species, including A. missouriensis, whose genome sequences and gene annotations have been registered in the database (Fig. S2). Furthermore, proteins showing high similarities with HhkA are also conserved in four species of other rare actinomycetes: Couchioplanes caeruleus, which forms flagellated arthrospores by fragmentation of aerial hyphae; Catenuloplanes japonicus, which forms segmental motile spores; Pseudosporangium ferrugineum, which produces irregular pseudosporangia directly from aggregated spore chains; and Krasilnikovia cinnamomea, which forms pseudosporangia on short sporangiophores above the surface of substrate mycelium (Fig. S2) (31–34). These in silico analyses demonstrate that HhkA homologues are evolutionarily conserved among the members of the genus Actinoplanes and some related rare actinomycetes in the family Micromonosporaceae.

hhkA is actively transcribed during sporangium formation and dehiscence.

We analyzed the transcript levels of hhkA in the wild-type strain by reverse transcription-quantitative PCR (qRT-PCR) analysis. Few hhkA transcripts were detected in substrate hyphae cultivated on HAT agar at 30°C for 24 h and in the mixture of substrate hyphae and immature sporangium-like structures cultivated for 48 and 72 h (Fig. 1, bars S₁, S₂, and S₃). In contrast, the transcripts were abundantly detected in the mixture of substrate hyphae and sporangia formed on HAT agar at days 4 and 5 (Fig. 1, bars S₄ and S₅). The transcript level decreased at day 6 on HAT agar (Fig. 1, bar S₆). The transcripts were kept at high levels during sporangium dehiscence (Fig. 1, bars D₀ and D₁₅) and rapidly decreased after completion of sporangium dehiscence (Fig. 1, bar D₁₂₀). Thus, we revealed that hhkA is actively transcribed in the process of sporangium formation and dehiscence, similar to tcrA, fliA1, fliA2, and fliA3 (26, 27).

FIG 1.

Transcript levels of hhkA under various culture conditions. The transcripts were examined by qRT-PCR analysis. RNA samples were prepared from substrate hyphae or mixtures of substrate hyphae and sporangia grown on HAT agar for 1 to 6 days (S₁ to S₆) and sporangia (including some substrate hyphae) incubated in 25 mM histidine solution to induce sporangium dehiscence for 0, 15, and 120 min (D₀, D₁₅, and D₁₂₀, respectively). The rpoB transcript encoding the RNA polymerase β subunit was used as an internal standard. Data are means from three biological replicates ± standard errors.

We determined the transcriptional start point of hhkA to be nine nucleotides downstream from the first nucleotide of the annotated translational start codon by high-resolution S1 nuclease mapping (Fig. S3A). Taking the transcriptional start position into consideration, we concluded that the hhkA mRNA should be the leaderless mRNA and that the genuine translational start site of hhkA should be nine nucleotides (three codons) downstream from the annotated translational start site (Fig. S3B), which is in accordance with the translational start sites of most hhkA homologues in other rare actinomycetes (Fig. S2). In the upstream region from hhkA, we were not able to find any sequences that match the regulatory motifs identified in A. missouriensis, such as the AmBldD and TcrA boxes and the FliA (FliA1, FliA2, and/or FliA3)-recognizing promoters (Fig. S3B) (25–27). Furthermore, binding of AmBldD to the upstream region from hhkA was not detected in the chromatin immunoprecipitation sequencing (ChIP-Seq) experiment in our previous study (25), and significant changes in hhkA transcript levels were not observed in the ΔtcrA mutant or any of the ΔfliA1 to -3 mutants compared with the wild-type strain in our previous transcriptome sequencing (RNA-Seq) analyses (26, 27). Therefore, the regulatory mechanism for the growth phase-dependent transcriptional activation of hhkA is elusive.

HhkA regulates a signal transduction pathway involved in sporangium formation, spore dormancy, and sporangium dehiscence.

To examine the in vivo function of HhkA, we generated an hhkA null (ΔhhkA) mutant strain. No difference was observed between the wild-type and ΔhhkA mutant strains by macroscopic observation of mycelia or sporangia formed on yeast extract-beef extract-NZ amine-maltose monohydrate (YBNM) and HAT agars (data not shown). To examine sporangium formation in detail, we observed mycelia and sporangia of the wild-type and ΔhhkA mutant strains grown on HAT agar at 30°C for 7 days by scanning electron microscopy (SEM), but both strains produced normal globose or subglobose sporangia with short sporangiophores (Fig. S4). Then, we observed ultrathin sections of the wild-type and ΔhhkA mutant sporangia grown under the same conditions as for SEM analysis by transmission electron microscopy (TEM) to examine spore maturation inside sporangia. The wild-type strain produced normal round sporangiospores of similar sizes (Fig. 2A). In contrast, the ΔhhkA mutant produced many distorted sporangiospores of various sizes (Fig. 2B). Furthermore, several sporangiospores ectopically germinated inside the sporangium, indicating that spore dormancy was not maintained properly in the ΔhhkA mutant (Fig. 2B). This phenotypic change is quite similar to that observed in the ΔtcrA mutant (Fig. 2C), which suggests that HhkA and TcrA could constitute a two-component signal transduction pathway (26) (see Discussion).

FIG 2.

TEM observation of the ultrathin sections of the wild-type (A), ΔhhkA (B), and ΔtcrA (C) sporangia produced on HAT agar through 7 days of cultivation. Ectopically germinating spores in the ΔhhkA and ΔtcrA mutants are indicated by arrows. A part of the sporangium wall appears to be ruptured in the ΔhhkA and the ΔtcrA mutants. However, these mutant sporangia were deficient in sporangium dehiscence. This apparently inconsistent observation can be explained by our unpublished observation that degradation of the sporangium wall was not sufficient for the release of sporangiospores. Bars, 1 μm.

Next, we tried to analyze the motility of zoospores by optical microscopy, but the ΔhhkA mutant spores were rarely observed under the dehiscence-inducing conditions used in this assay, suggesting that sporangium dehiscence is severely inhibited in the ΔhhkA mutant. Thus, we quantified the spores released from the wild-type and ΔhhkA mutant sporangia, both of which contained the chromosome-integrating pTYM19-Apra vector, by counting colonies formed on YBNM agar after incubation at 30°C for 2 days. While the wild-type sporangia formed on one HAT agar plate released over 10⁶ spores, the ΔhhkA mutant sporangia released only over 10⁴ spores per one HAT agar plate under the same conditions (Fig. 3), which is consistent with the notion that sporangium dehiscence is severely inhibited in the ΔhhkA mutant. In a gene complementation test, the introduction of hhkA on pTYM19-Apra into the ΔhhkA mutant resulted in the complete restoration of the number of spores released from sporangia (Fig. 3). As described above, His-318 and Asp-603 of HhkA are probable residues that are phosphorylated in the signal transduction process (Fig. S1). Therefore, we generated and introduced two mutated hhkA genes encoding HhkA (H318F) and HhkA (D603N) into the ΔhhkA mutant in parallel. As a result, the number of spores released from the ΔhhkA mutant sporangia was not restored at all by the introduction of the mutated hhkA genes, suggesting that phosphorylation at His-318 and Asp-603 is essential for the in vivo function of HhkA (Fig. 3). Alternatively, it is possible that the mutated gene products were unstable and the decrease of protein levels caused the failure in the complementation test. In any case, these results clearly demonstrate that HhkA is required for the formation of mature sporangia that can normally dehisce and release sporangiospores under dehiscence-inducing conditions. The severe decrease in colony counts was also observed in the ΔtcrA mutant (26), suggesting the possibility that HhkA and TcrA constitute a two-component signal transduction pathway required for proper sporangium formation (see Discussion).

FIG 3.

Numbers of spores released from sporangia. The strains for the complementation test harbor the wild-type or mutated hhkA gene on the chromosome. Each strain was cultivated on HAT agar at 30°C for 7 days. Zoospores released from the sporangia by pouring 25 mM NH₄HCO₃ solution were counted as CFU on YBNM agar as described in Materials and Methods. Data are means from at least three biological replicates ± standard errors.

Each of six Hpt domain-containing proteins is not essential for sporangium formation.

As described in the introduction, phosphotransfer reactions between a hybrid sensor histidine kinase and a response regulator can be mediated by an Hpt protein. Thus, we examined whether HhkA transfers its phosphoryl group to TcrA via an Hpt protein. Because AMIS_37460 is a sole gene that encodes a protein composed of only the Hpt domain on the A. missouriensis genome, we first investigated whether AMIS_37460 is required for the signal transduction pathway involving HhkA and TcrA. We generated an AMIS_37460-null (ΔAMIS_37460) mutant and compared its phenotypes with those of the wild-type strain. The ΔAMIS_37460 mutant grew normally and formed globose or subglobose sporangia as in the wild-type strain (Fig. S5A to D). Furthermore, the ΔAMIS_37460 mutant sporangia dehisced and released zoospores similarly to the wild-type sporangia under sporangium dehiscence-inducing conditions (Fig. S5K). These results demonstrated that AMIS_37460 is not required for sporangium formation and dehiscence and that the Hpt protein is not involved in the phosphotransfer reaction between HhkA and TcrA. In addition to AMIS_37460, five genes (AMIS_18620, AMIS_25430, AMIS_36070, AMIS_68710, and AMIS_76570) were predicted to encode multidomain proteins with an Hpt domain on the A. missouriensis genome. Next, we examined the possible involvement of these five proteins in the phosphotransfer reaction between HhkA and TcrA by the observation of phenotypic changes in each gene deletion mutant. We generated null mutants of AMIS_18620, AMIS_25430, and AMIS_36070. These 3 mutants showed no phenotypic change compared with the wild-type strain in sporangium formation (Fig. S5E to J) and dehiscence (Fig. S5K). The remaining two genes, AMIS_68710 (cheA-2) and AMIS_76570 (cheA-1), belong to the che gene clusters (che2, from AMIS_68640 to AMIS_68770, and che1, from AMIS_76480 to AMIS_76570, respectively) predicted to be involved in the chemotactic properties of zoospores. We had generated and analyzed mutant strains that lacked either the whole che1 or che2 gene cluster. Both of the mutants grew normally and formed globose or subglobose sporangia that dehisced and released zoospores similarly to the wild-type sporangia (data to be published elsewhere). Taken together, these results demonstrated that each of the Hpt domain-containing proteins is dispensable for the signal transduction involving HhkA and TcrA, although we cannot exclude the possibility that two or more Hpt domain-containing proteins (including the Hpt protein AMIS_37460) mediate the phosphotransfer reaction between HhkA and TcrA. We note that transcript levels of all of these Hpt domain-containing protein genes in the wild-type strain increased greatly (over 10-fold) during sporangium formation (at day 3 and/or day 6 compared to day 1 on the HAT agar) according to our previous RNA-Seq analysis (24).

Genes regulated by HhkA overlap mostly with TcrA-regulated genes.

Considering the significant phenotypic changes observed in the ΔhhkA mutant (Fig. 2 and 3), we postulated that the signal transduction pathway involving HhkA regulates many genes. Therefore, we compared the transcriptomes between the wild-type and ΔhhkA mutant strains using RNA sequencing (RNA-Seq) analysis. In this analysis, we also analyzed the transcriptome of the ΔtcrA mutant to verify the functional relevance between HhkA and TcrA. We prepared total RNAs extracted from the mixtures of vegetative hyphae and sporangia grown on HAT agar for 6 days in triplicate for each strain. From the obtained RNA-Seq data, we extracted the genes that meet the following criteria as up- or downregulated genes in the ΔhhkA and ΔtcrA mutant strains compared with the wild-type strain: (i) average numbers of reads per kilobase of coding sequence per million mapped reads (RPKM) in the mutant strains over 2.0-fold more (upregulation) or below 0.5-fold less (downregulation) than those in the wild-type strain and (ii) statistical q values under 0.05. As a result, the transcript levels of 194 genes were significantly changed in the ΔhhkA mutant compared with the wild-type strain, with 140 and 54 genes being down- and upregulated in the mutant, respectively (Fig. 4A; also, see Tables S1 and S2 in the supplemental material). This result is consistent with the assumption that HhkA primarily functions as a positive transcriptional regulator via the function of its partner response regulator. In the ΔtcrA mutant, the transcript levels of 446 genes were significantly changed compared with those in the wild-type strain, with 349 and 97 genes being down- and upregulated in the mutant, respectively (Fig. 4B; Tables S3 and S4).

FIG 4.

RNA-Seq analysis and HhkA- or TcrA-regulated genes. (A and B) Volcano plots of differential expression. Each gene was plotted based on the fold change versus the wild-type strain and the q value in the ΔhhkA (A) and ΔtcrA (B) mutant strains. Genes differentially expressed compared with the wild-type strain are highlighted by color; blue and red dots indicate the up- and downregulated genes, respectively, in the ΔhhkA or ΔtcrA mutant. Dotted lines indicate the threshold of the q value (0.05). (C) Scatterplot of correlated expression in the ΔtcrA mutant. Expression levels of each gene were plotted based on the fold changes in the ΔtcrA mutant compared with the wild-type strain analyzed in this and our previous studies (26). Differentially expressed genes are highlighted by color; red dots indicate the downregulated genes in the ΔtcrA mutant in both studies. Blue and green dots indicate the downregulated genes in this study and in our previous analyses, respectively. The regression line is shown in the graph (R² = 0.424). The dotted lines indicate the thresholds of the fold change (2.0- and 4.0-fold in this and our previous studies, respectively). (D) Venn diagram of downregulated genes in the ΔtcrA mutant. The numbers of the downregulated genes compared with the wild-type strain are shown. (E and F) Venn diagrams of differentially expressed genes. The numbers of the genes significantly downregulated (E) or upregulated (F) (over 2.0-fold) in the ΔhhkA and ΔtcrA mutant strains compared with the wild-type strain are shown. (G) Scatterplot of correlated expression in ΔhhkA and ΔtcrA mutants. Expression levels of each gene were plotted based on the fold changes in the ΔhhkA and ΔtcrA mutant strains compared with the wild-type strain. The regression line is shown in the graph (R² = 0.365). (H) Sequence logos of the TcrA box. The panels are based on the 24 and 11 TcrA-binding sites among the upstream regions of the 125 TcrA- and HhkA-dependent genes and 224 TcrA-dependent and HhkA-independent genes, respectively. Arrows indicate the direct repeat.

In our previous work, we reported the comparative transcriptome analysis by the RNA-Seq between the wild-type and ΔtcrA mutant strains (26). Because we used biologically single RNA samples extracted from the mixture of vegetative hyphae and sporangia of the wild-type and ΔtcrA mutant strains, we defined 263 genes whose RPKM values in the wild-type strain were over 4.0-fold higher than those in the ΔtcrA mutant as TcrA-dependent genes (26). In the present study, we analyzed biologically triplicate RNA samples in each strain and adopted the different criteria to extract TcrA-dependent genes mentioned above. Therefore, we compared the transcriptomic profiles obtained in our previous and present studies. The variations of the transcript levels of all annotated protein-coding genes were correlated between the two analyses (decision coefficient [R²] = 0.424), confirming that reproducible results were obtained (Fig. 4C). Of the 263 TcrA-dependent genes identified in our previous work (26), 227 (86%) were also found to be TcrA dependent in this study (Fig. 4C and D; Table S3). Furthermore, we extracted 122 additional genes as TcrA-dependent genes in the present study (Fig. 4C and D; Table S3). We assume that the statistically and biologically precise TcrA-dependent genes were successfully identified in the current RNA-Seq analysis.

Next, we compared the genes whose transcript levels were significantly changed in the ΔhhkA mutant with those in the ΔtcrA mutant. As expected, the genes regulated by HhkA mostly overlapped those regulated by TcrA; 125 of the 140 genes (89%) downregulated in the ΔhhkA mutant were also downregulated in the ΔtcrA mutant (Fig. 4E; Table S3), while 36 of the 54 genes (67%) upregulated in the ΔhhkA mutant were also upregulated in the ΔtcrA mutant (Fig. 4F; Table S4). Within the 125 genes that were downregulated in both ΔhhkA and ΔtcrA mutants, four flagellar genes (flbD, motA, motB, and fliL), 12 chemotaxis-related genes, including those for methyl-accepting chemotaxis protein (MCP) and MCP-like protein, two FliA family sigma factor genes (fliA3 and fliA4), and a cell wall hydrolase gene (gsmA) were found (Tables S1 and S3), supporting the notion that HhkA and TcrA are involved in the same signal transduction pathway regulating sporangium formation, dehiscence, and motility of zoospores (24, 27, 35). Furthermore, the variations in the transcript levels of all protein-coding genes between the ΔhhkA and ΔtcrA mutants compared with the wild-type strain correlated with each other (R² = 0.365) (Fig. 4G).

In our previous study, we demonstrated that a recombinant TcrA protein binds to TcrA box-containing regions upstream from 34 TcrA-dependent genes by electrophoretic mobility shift assay (EMSA) (26). In our recent study, we also showed that TcrA binds to the upstream region from fliA2, which encodes a FliA family sigma factor involved in sporangium formation, spore dormancy, and sporangium dehiscence (27). In the present study, we identified 125 genes as both HhkA- and TcrA-dependent genes, and we defined 224 genes as TcrA-dependent, HhkA-independent genes, including direct and indirect regulation (Fig. 4E; Tables S1 and S3). Then, we searched the regions upstream from these genes for the TcrA box sequences to which the recombinant TcrA protein bound in our previous study (26, 27). As a result, we found 24 and 11 genuine TcrA boxes in the regions upstream from the set of both HhkA- and TcrA-dependent genes and the set of TcrA-dependent and HhkA-independent genes, respectively (Table S3). The direct repeat sequences of the TcrA box upstream from the former 24 genes showed lower conservation than those from the latter 11 genes, indicating that HhkA may be required for TcrA binding to the low-affinity TcrA box sequences that contain a relatively large number of mismatches to the consensus TcrA box sequence (Fig. 4H). Because phosphorylation of TcrA enhanced its affinity for a low-affinity TcrA box sequence, as described below, this result suggests that HhkA is involved in the phosphorylation of TcrA.

Possible interaction between HhkA and TcrA.

Because HhkA and TcrA are predicted to constitute a two-component system, we assume that HhkA and TcrA interact with each other. To test this possibility, we performed a bacterial two-hybrid assay in Escherichia coli and observed weak interaction between HhkA and TcrA, which was indicated by slow growth on the assay plate containing maltose as the sole carbon source (Fig. 5). In a parallel experiment, similar weak interaction was detected between a truncated HhkA protein without the response regulator receiver domain in its C-terminal region (HhkAΔRec) and TcrA (Fig. 5), suggesting the possibility that the phosphotransfer reaction occurs directly from the His residue in the kinase domain of HhkA to the Asp residue in the receiver domain of TcrA (see Discussion). In control experiments, all of the HhkA, HhkAΔRec, and TcrA fusion proteins exhibited no interaction with the T18 and T25 domains of adenylate cyclase, showing that HhkA and HhkAΔRec proteins specifically interacted with TcrA in E. coli cells (Fig. 5). Next, we attempted to quantify the protein-protein interactions via a β-galactosidase assay using the above-mentioned E. coli transformants in liquid culture. However, we failed to detect significant changes in enzymatic activities between the cells cotransformed with the hhkA- and tcrA-expressing plasmids and those with the empty vectors, further suggesting that the interaction between HhkA and TcrA is very weak and probably transient (Fig. S6).

FIG 5.

Bacterial two-hybrid analysis of HhkA and TcrA. E. coli BTH101 cells carrying the plasmids encoding protein fusions to the T18 or T25 domains were spotted onto M63 agar plates supplemented with X-Gal, IPTG, and maltose. E. coli BTH101 cells carrying the plasmids encoding the T18 or T25 domains were used as a negative control. The plates were incubated at 30°C for 4 days and imaged. The cells can grow and turn blue only when the T18 domain binds to the T25 domain through the interaction between two proteins fused individually to the T18 and T25 domains.

In vitro phosphorylation of TcrA enhances its affinity for the TcrA box.

We repeatedly tried but failed to detect the phosphotransfer reaction between the recombinant HhkA and TcrA proteins produced in E. coli due to the lack of autophosphorylation of HhkA (data not shown). Thus, the polyhistidine-tagged TcrA (His-TcrA) protein was phosphorylated using acetyl phosphate as a phosphate donor, and the effect of the phosphorylation on its affinity for the TcrA box sequences was assessed by EMSAs. The His-TcrA protein was incubated in the absence or presence of acetyl phosphate at 30°C for 15 min and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using a Phos-tag acrylamide gel, which has been used to separate the phosphorylated protein as a retarded signal from the nonphosphorylated form (36, 37). As a result, a portion of His-TcrA was detected in a phosphorylated form after incubation with acetyl phosphate (Fig. 6A). We calculated the proportion of the phosphorylated protein at approximately 18% of total His-TcrA used in this assay by quantification of signal intensities in the gel (Fig. 6A). In a parallel experiment, the mutated His-TcrA(D52N) protein, in which a probable phosphorylation site in the receiver domain (Asp-52) was changed to Asn, was also incubated in the absence or presence of acetyl phosphate. Under the same conditions, the phosphorylated form of His-TcrA(D52N) was not detected, clearly indicating that the Asp-52 residue in His-TcrA was specifically phosphorylated with acetyl phosphate (Fig. 6A).

FIG 6.

In vitro phosphorylation and DNA-binding activity of the His-TcrA protein. (A) In vitro phosphorylation of the His-TcrA protein. His-TcrA and His-TcrA(D52N) proteins were incubated in the absence or presence of 100 mM acetyl phosphate at 30°C for 15 min and analyzed by SDS-PAGE using a Phos-tag acrylamide gel. Phosphorylated and nonphosphorylated proteins are indicated with open and closed triangles, respectively. (B) Nucleotide sequences of upstream regions from AMIS_76300 and AMIS_35720. Direct repeat sequences of the TcrA box are shaded in yellow. Arrows indicate the direction of the direct repeat sequences. A FliA-dependent promoter is shown in red. Bent arrows indicate the transcriptional start points. (C to E) DNA-binding activity of the His-TcrA protein. DNA probes were ³²P labeled and used for EMSAs. The His-TcrA or His-TcrA(D52N) protein was incubated in the absence (left panels) or presence (middle panels) of acetyl phosphate prior to and during the binding reactions. Protein concentrations used in the reactions are shown above the lanes. Protein-DNA complexes and wells are indicated with closed and open triangles, respectively. Using the results of EMSAs, fractions of DNA bound by His-TcrA or His-TcrA(D52N) (retarded DNA/total DNA) were quantified using Quantity One software, and the K_D values were calculated using nonlinear regression lines (right panels) (38). The DNA probes and proteins used were the AMIS_76300 promoter-containing region (probe 1) and His-TcrA (C), the AMIS_35720 promoter-containing region (probe 2) and His-TcrA (D), and the probe 2 and His-TcrA(D52N) (E), respectively.

We performed EMSAs using two DNA probes, 1 and 2, both of which contain the TcrA box (Fig. 6B). Probe 1 covers an upstream region from AMIS_76300, which is one of the TcrA-dependent and HhkA-independent genes (Table S3). The direct repeat in this probe exactly matches the consensus TcrA box sequence (Fig. 6B) (26). Probe 2 contains an upstream region from AMIS_35720, which is a member of the set of both HhkA- and TcrA-dependent genes (Table S3). The direct repeat in probe 2 differs slightly from the consensus sequence (Fig. 6B). In our previous study, we reported that His-TcrA bound to these probes with different affinities, as determined by a competitive EMSA, and that His-TcrA did not bind to the mutated probes lacking the upstream or downstream halves of the direct repeat of the TcrA box in probe 1 by EMSAs (26). In EMSAs, the affinity of His-TcrA for probe 1 was scarcely affected by phosphorylation (the dissociation constants [K_D values] [38] are 0.04 and 0.03 μM in the absence and presence of acetyl phosphate, respectively) (Fig. 6C), while that for probe 2 was markedly enhanced through the phosphorylation (the K_D values are 0.3 and 0.1 μM in the absence and presence of acetyl phosphate, respectively) (Fig. 6D). Considering that only 18% of His-TcrA was phosphorylated, the phosphorylated form seemed to bind to probe 2 with more than 10-times-greater affinity than the unphosphorylated form. In addition, the affinity of His-TcrA(D52N) for probe 2 was hardly affected by incubation with acetyl phosphate (the K_D values are 0.8 and 0.7 μM in the absence and presence of acetyl phosphate, respectively) (Fig. 6E), which corresponded to the observation that no phosphorylated form was detected in the SDS-PAGE using the Phos-tag acrylamide gel (Fig. 6A). Taken together, these results support the notion that the phosphorylation in the receiver domain of TcrA enhances its affinity for the low-affinity TcrA box sequences upstream from the HhkA- and TcrA-dependent genes (see Discussion).

DISCUSSION

In the current study, we carried out a functional characterization of the hybrid histidine kinase HhkA in A. missouriensis. We proposed that HhkA and TcrA compose a cognate two-component system essential for sporangium formation, spore dormancy, and sporangium dehiscence according to the following findings and experimental results: (i) HhkA and TcrA homologues are highly conserved in Actinoplanes bacteria (see Fig. S2 in the supplemental material) (26); (ii) very similar phenotypic changes were observed in the ΔhhkA and ΔtcrA mutants, in which spores ectopically germinated inside sporangia and sporangium dehiscence was severely inhibited (Fig. 2 and 3) (26); (iii) the genes regulated by HhkA mostly overlapped those regulated by TcrA (Fig. 4); (iv) each of the six Hpt domain-containing proteins was not essential for sporangium formation and dehiscence (Fig. S5); and (v) interaction between HhkA and TcrA was suggested by a bacterial two-hybrid assay (Fig. 5).

As described in the introduction, the phosphotransfer reactions between hybrid histidine kinases and cognate response regulators are mediated by Hpt proteins. However, some hybrid sensor histidine kinases transfer the phosphoryl group on the His residue in the kinase domain directly onto the Asp residue in the receiver domain of their respective partner response regulators. The AtsR-AtsT two-component system that regulates the quorum sensing in Burkholderia cenocepacia is an example. In this regulatory system, the hybrid histidine kinase AtsR autophosphorylates a His residue, and then the phosphoryl group is transferred to the response regulator AtsT and partly to the Asp residue of the receiver domain of AtsR (39). In Myxococcus xanthus, the hybrid histidine kinase SgmT also autophosphorylates His-336 in the kinase domain and directly transfers the phosphoryl group to Asp-53 in the receiver domain of the partner response regulator DigR without the phosphotransfer reaction to Asp-615 in the receiver domain of SgmT (40). Because of the above-mentioned experimental results (no. 4 and 5), we assumed the direct phosphotransfer reaction from His-318 in the kinase domain of HhkA to Asp-52 in the receiver domain of TcrA. To show this, we repeatedly tried in vitro assays using the recombinant HhkA and TcrA proteins produced in E. coli. Unexpectedly, however, the autophosphorylation activity was not detected at all in the assays using the HhkA proteins in their full-length and truncated forms that lacked the sensor or receiver domains (data not shown). Considering that both His-318 and Asp-603 are required for the in vivo function of HhkA in the gene complementation test (Fig. 3), we suppose that an unknown histidine kinase(s) (or possibly an intracellular phosphate donor such as acetyl phosphate) transfers a phosphoryl group onto Asp-603 of HhkA, thereby allosterically modifying the activity of HhkA to allow the autophosphorylation of His-318 (Fig. 7A). Then, the activated HhkA mediates the phosphotransfer reaction from His-318 of HhkA to Asp-52 in the receiver domain of TcrA (Fig. 7A). In this working model, the HhkA-TcrA two-component system in A. missouriensis is different from the SgmT-DigR system of M. xanthus (40), in that the phosphorylation at Asp-603 in the receiver domain of HhkA is essential for the signal transduction pathway in A. missouriensis. However, it should be noted that we cannot completely exclude the possibility that the in vivo phosphotransfer reaction from Asp-603 of HhkA to Asp-52 of TcrA is mediated by any of the two (or more) of six Hpt domain-containing proteins. In such a case, it would be understandable that the single gene mutants of the six candidate proteins showed no phenotypic changes (Fig. S5).

FIG 7.

Proposed regulatory model of gene expression for morphological development by HhkA and TcrA. (A) Regulatory model of phosphotransfer in the HhkA-TcrA regulatory system. An unknown histidine kinase mediates the phosphotransfer to Asp-603 in the HhkA receiver domain, followed by the autophosphorylation at His-318 in the kinase domain. The active HhkA transfers the phosphoryl group to Asp-52 in the TcrA receiver domain to enhance its affinity for TcrA boxes. (B) Regulatory model of gene expression by the HhkA-TcrA regulatory system. Arrows indicate positive control, including direct and indirect regulation. Open arrows indicate the involvement of gene products in the biological phenomena described in boxes. TcrA activates the target genes partly through the alternative sigma factors FliA1 and FliA2. The phosphorylated TcrA activates the wider range of genes required for the developmental processes. The regulatory pathways and genes (or gene products) activated by phosphorylated TcrA are indicated in red.

In the current study, we analyzed the transcriptomes of the wild-type, ΔhhkA, and ΔtcrA strains by RNA-Seq techniques (Fig. 4). To the best of our knowledge, no comparative transcriptomic analyses using RNA-Seq between a histidine kinase gene mutant and its partner response regulator gene mutant have been conducted so far, although the RhpS-RhpR two-component system in Pseudomonas savastanoi has been characterized by RNA-Seq, in which the transcriptomes of the histidine kinase-encoding rhpS deletion mutant and rhpS rhpR deletion mutant were compared with the wild-type transcriptome (41). In our analysis, a considerable number of TcrA-dependent, HhkA-independent genes were found, while most of the genes under the control of HhkA were also regulated by TcrA (Fig. 4E; Table S3). This result indicates that unphosphorylated TcrA can bind to some TcrA box sequences to activate the transcription of target genes (Fig. 7B). In fact, our EMSA results showed that unphosphorylated TcrA bound to a high-affinity TcrA box sequence, which completely matches the consensus TcrA box sequence, similar to phosphorylated TcrA, while phosphorylation at Asp-52 greatly enhanced the binding affinity of TcrA to a low-affinity TcrA box sequence (Fig. 6). Similar to other histidine kinases of the two-component regulatory system, HhkA seems to enhance the affinity of TcrA to the TcrA boxes through the phosphorylation of TcrA, which induces the conformational change (probably dimerization) of TcrA. In this respect, phosphorylation of the response regulator OmpR has been reported to increase its affinity for the OmpR-binding site by facilitating cooperative interactions between the phosphorylated OmpR proteins located at adjacent OmpR-binding sites (42). We repeatedly tried but failed to detect the phosphorylated TcrA in the cellular proteins, which were extracted from vegetative hyphae, a mixture of hyphae and sporangia, and sporangia under dehiscence-inducing conditions, by Western blot analysis using the Phos-tag acrylamide gel and a polyclonal anti-TcrA antibody (data not shown). This is probably due to the low level of phosphorylated TcrA. Determination of the accurate timing and frequency of in vivo phosphorylation of TcrA is the subject of our future studies.

In Bacillus subtilis, a response regulator, DegU, of the DegS-DegU two-component regulatory system has different functions by means of its unphosphorylated and phosphorylated forms; the former and latter are required for degradative enzyme production and competence gene expression, respectively (43). Thus, DegU has two active (unphosphorylated and phosphorylated) conformations for the control of different regulatory pathways. In addition, unphosphorylated and phosphorylated forms of the response regulators OmpR, AlgR, SsrB, and RhpR were reported to regulate distinct biological processes (41, 44). Different from the response regulators in these two-component regulatory systems, both unphosphorylated and phosphorylated forms of TcrA regulate the same biological process, i.e., morphological differentiation in A. missouriensis. In the HhkA-TcrA system, the genes activated by the unphosphorylated form of TcrA are also activated by the phosphorylated form of TcrA; phosphorylation of TcrA results in the extension of the activated genes in number (Fig. 7B), and some of the TcrA-dependent genes are expected to be activated more strongly by phosphorylated TcrA than unphosphorylated TcrA. Importantly, our results indicate that approximately one-third of the TcrA-dependent genes seem to be activated only by phosphorylated TcrA directly or indirectly (Fig. 4E; Table S3).

Consistent with our previous gene complementation test using the mutated tcrA(D52N) gene (26), the in vitro phosphorylation experiment using acetyl phosphate confirmed that Asp-52 is the major phosphorylated residue in TcrA (Fig. 6A). Taking the in vivo interaction between HhkA and TcrA into consideration, we proposed that the phosphorylation of TcrA at Asp-52 is mainly attributed to the phosphotransfer from HhkA, although we cannot completely exclude the possibility that a low level of phosphorylation of TcrA could be carried out by another histidine kinase(s) due to a phenomenon described as cross talk (45). The presence of a considerable number of TcrA-dependent, HhkA-independent genes can explain the slight phenotypic difference between the ΔhhkA and ΔtcrA mutants. For example, apparently normal sporangia were observed by SEM in the ΔhhkA mutant (Fig. S4), while squashed sporangia were observed in the ΔtcrA mutant (26).

We recently reported that TcrA regulates several hundreds of genes through the transcriptional activation of three regulatory genes, fliA1, fliA2, and fliA3, which encode the FliA family sigma factors (27). Among them, two paralogous FliA family sigma factors, FliA1 and FliA2, form an important hub in the regulatory network for sporangium formation, spore dormancy, and sporangium dehiscence, while FliA3 does not have an apparent impact on morphological differentiation. In this study, we revealed that the transcription of fliA1 and fliA2 is under the control of TcrA but not HhkA, while the transcription of fliA3 is regulated by both TcrA and HhkA (Tables S1 and S3). This means that TcrA can likely activate fliA1 and fliA2 without phosphorylation. The observation that the ΔhhkA mutant was deficient in the formation of mature sporangia that can normally dehisce and release the spores despite the transcriptional activation of fliA1 and fliA2 implies that the whole complex regulatory network involving TcrA, HhkA, and three FliA family sigma factors is required for normal morphological differentiation (Fig. 7B). Inevitably, there remains a possibility that the production or function of FliA1 and FliA2 is repressed at posttranscriptional levels in the ΔhhkA mutant. Considering that the genes highly similar to hhkA, tcrA, and multiple fliA genes are conserved among the genus Actinoplanes and several other rare actinomycetes (Fig. S2) (26, 27), we expect that the regulatory network involving the possible HhkA-TcrA two-component regulatory system and the FliA family sigma factors is evolutionarily conserved for the reproduction processes in these developmentally complex bacterial members.

MATERIALS AND METHODS

General methods.

Bacterial strains, plasmid vectors, and media used in this study were described previously (25, 46). Primers used in this study are listed in Table S5 in the supplemental material. Preparation of A. missouriensis cells and RNA extraction were described previously (26). qRT-PCR and high-resolution S1 nuclease mapping were performed as described previously (27). The rpoB gene was used as an internal standard in qRT-PCR. All reactions were performed in triplicate, and the data were normalized using the average for the internal standard. TEM was performed with an H-7600 electron microscope (Hitachi, Tokyo, Japan) as described previously (26). SEM was performed with an S-4800 scanning electron microscope (Hitachi) as described previously (47).

Construction of the gene deletion mutant strains.

For construction of the ΔhhkA, ΔAMIS_18620, ΔAMIS_25430, ΔAMIS_36070, and ΔAMIS_37460 mutant strains, the upstream and downstream regions of the genes were amplified by PCR. The amplified DNA fragments were cloned into pUC19 and sequenced to confirm that no PCR-derived error had been introduced. The cloned fragments were digested and cloned together into pK19mobsacB (48), whose kanamycin resistance gene had been replaced with the apramycin resistance gene aac(3)IV (25). Using these plasmids, ΔhhkA, ΔAMIS_18620, ΔAMIS_25430, ΔAMIS_36070, and ΔAMIS_37460 mutant strains were generated by the method described previously (27). The disruption of each gene was confirmed by PCR (data not shown). The ΔtcrA mutant was obtained in our previous study (26).

Construction of the recombinant strains for the complementation test.

A DNA fragment containing the promoter and coding sequences of hhkA was amplified by PCR. The amplified fragment was digested with EcoRI and HindIII and cloned between the EcoRI and HindIII sites of pTYM19-Apra (26). To construct the mutated hhkA genes, H318F and D603N mutations were introduced by PCR and the resultant DNA fragments were also cloned into pTYM19-Apra. The generated plasmids were sequenced and introduced into the ΔhhkA mutant by conjugation as described previously (24). Plasmid pTYM19-Apra was also introduced into the wild-type and ΔhhkA mutant strains for the vector control strains. Then, apramycin-resistant colonies were obtained.

Counting of spores released from sporangia.

Strains were precultured in nutrient-rich PYM broth with shaking at 30˚C for 2 days. A fixed amount of the mycelium (1 ml of preculture) was inoculated onto a HAT plate, and the plate was incubated at 30°C for 7 days to produce sporangia. Zoospores were released from the sporangia by pouring 10 ml of 25 mM NH₄HCO₃ solution into one HAT plate and incubating the plate at 30°C for 1 h. After being collected from the plate, the zoospore-containing solution was filtrated through a 5-μm Acrodisc membrane filter (Pall Corporation, NY) to eliminate hyphae and sporangia. A portion of the filtrate was inoculated onto YBNM agar, and the plate was incubated at 30°C for 2 days. From the number of colonies formed on YBNM agar, the number of zoospores released from sporangia on one HAT plate was estimated.

RNA-Seq and in silico analyses.

RNAs were extracted from the wild-type, ΔhhkA, and ΔtcrA mutant strains as described previously (26). The qualities and quantities of total RNAs were assessed with a Bioanalyzer DNA1000 (Agilent Technologies, CA). Sequencing libraries were prepared with 3 μg of RNA as the starting material, and the sequencing was performed with a HiSeq 2500 sequencer (Illumina, CA) to generate nondirectional paired-end 150-nucleotide reads. At least 1.2 Gb of sequencing data was obtained from each cDNA library. Library construction and sequencing were performed by Novogene (Beijing, China). The reads were filtered by sequence quality using a CLC Genomics Workbench 6.05 (CLC Bio, Aarhus, Denmark) and mapped to the A. missouriensis genome sequence.

Bacterial two-hybrid assay.

The bacterial two-hybrid assay was conducted using a BACTH system kit (Euromedex, Strasbourg, France). For construction of the T18 or T25 domain fusion plasmids, the coding sequences of hhkA and tcrA were amplified by PCR. The truncated hhkA gene without the receiver domain-coding sequence was also amplified. The DNA fragments were cloned into pUC19 and sequenced. Then, the cloned fragments were digested with restriction enzymes and cloned into the vectors pUT18C (hhkA) and pKT25 (tcrA). Because the annotated coding sequence of hhkA was amplified by PCR, the T18-HhkA fusion protein contained an additional three residues in the linker region (Fig. S2). E. coli BTH101 competent cells were cotransformed with the T18 and T25 domain fusion plasmids, and transformants were selected on LB agar containing ampicillin and kanamycin. To test for protein-protein interaction, three individual colonies per assay were grown overnight in LB broth with ampicillin and kanamycin. The resulting cultures were spotted onto M63 agar containing maltose, isopropyl-β-d(−)-thiogalactopyranoside (IPTG), 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal), ampicillin, and kanamycin according to the kit manufacturer’s instructions. The plates were incubated at 30°C for 4 days and photographed. To quantify protein-protein interaction, the transformants were grown overnight at 30°C in LB broth with ampicillin and kanamycin. The cultures were inoculated into LB broth with ampicillin, kanamycin, and IPTG and cultivated at 30°C for 40 h. The β-galactosidase activities were quantified as described previously (49).

In vitro phosphorylation of the His-TcrA protein.

The recombinant His-TcrA protein was produced using pColdI-TcrA-NHis in E. coli BLR(DE3) and purified as described previously (26). A DNA fragment containing the TcrA(D52N)-coding sequence was amplified by overlap extension PCR using pColdI-TcrA-NHis as the template. The fragment was digested with NdeI and XhoI and cloned between the NdeI and XhoI sites of pColdI, generating pColdI-TcrA(D52N)-NHis. The plasmid was introduced into E. coli BLR(DE3), and the recombinant His-TcrA(D52N) protein was produced and purified. The proteins were incubated with or without 100 mM lithium potassium acetyl phosphate in a kination buffer (100 mM Trizma base [Sigma-Aldrich, St. Louis, MO], 10 mM MgCl₂, 125 mM KCl [pH 7.0]) at 30°C for 15 min. Then, a third volume of an SDS-PAGE loading buffer (200 mM Trizma base, 400 mM dithiothreitol, 8% SDS, 0.4% bromo phenol blue, 40% glycerol [pH 6.8]) was added. The protein samples were analyzed by SDS-PAGE using a 12.5% Phos-tag acrylamide gel (Fujifilm Wako Pure Chemicals, Osaka, Japan) and stained with Coomassie brilliant blue R-250 (Nacalai Tesque, Kyoto, Japan).

EMSAs.

DNA fragments, including the AMIS_76300 promoter- and AMIS_35720 promoter-containing regions, were prepared by PCR as probe 1 (255 bp) and probe 2 (320 bp), respectively. The PCR products were ³²P labeled at the 5′ ends using [γ-³²P]ATP (PerkinElmer, MA) and T4 polynucleotide kinase (TaKaRa Biochemicals, Shiga, Japan). Recombinant His-TcrA and His-TcrA(D52N) proteins were incubated with or without 100 mM lithium potassium acetyl phosphate in the kination buffer at 30°C for 15 min. Then, the protein mixtures were incubated with the labeled DNA probes at 30°C for 15 min in a binding buffer [25 mM NaH₂PO₄, 7.5% glycerol, 100 ng/μl bovine serum albumin, 25 ng/μl poly(dI-dC)], and the mixtures were subjected to 6% PAGE at 34 mA for 2 h.

Data availability.

Nucleotide sequence data from the RNA-Seq analysis have been deposited in the DDBJ Sequence Read Archive under accession number DRA009652.