A New Class of Signal Transducer in His-Asp Phosphorelay Systems*

Nitrate transport activity of the LtnT permease of the cyanobacterium Synechococcus elongatus is activated when LtnA, a response regulator without an effector domain, is phosphorylated by LtnB, a hybrid histidine kinase. We identified a protein (LtnC) that is required for activation of LtnT. LtnC consists of an N-terminal histidine-containing phosphoacceptor (HisKA) domain, a receiver domain, and a unique C-terminal domain found in some cyanobacterial proteins. Because LtnC lacks an ATP-binding kinase domain of a histidine kinase, it is incapable of autophosphorylation, but LtnC is phosphorylated by LtnA. The histidine residue in the HisKA domain but not the aspartate residue in the receiver domain is essential for phosphorylation of LtnC and activation of LtnT. LtnC phosphorylation leads to oligomerization of the protein. Fusion of the C-terminal domain of LtnC to glutathione S-transferase, which forms oligomers, also activates LtnT, suggesting that oligomerization of the LtnC C-terminal domain causes LtnT activation. These results indicate that the C-terminal domain of LtnC acts as an effector domain that directs the output of the signal from the phosphorelay system. The two-step (His-Asp-His) phosphorelay system, composed of the LtnB, LtnA, and LtnC proteins, is distinct from the known phosphorelay systems, namely, the typical two-component system (His-Asp) and the multistep phosphorelay system (His-Asp-His-Asp), because the HisKA domain of LtnC is the terminal phosphoacceptor that determines the signal output. LtnC is a new class of signal transducer in His-Asp phosphorelay systems that contains a HisKA domain and an effector domain.

ization of a pseudorevertant of NA3 (NA3R) capable of uptake of low concentrations of nitrate identified a sulfate permeaselike protein (LtnT) having latent nitrate transport activity and whose C-terminal domain inhibits the transport activity (10). Although the physiological role of this permease, whose affinity for nitrate is much lower than that of ABC-type NRT, remains to be elucidated, the presence of a novel mechanism for regulation of the permease was revealed by the characterization of NA3R. A two-component system involving a hybrid sensor histidine kinase with two REC domains (LtnB) and a response regulator with no effector domain (LtnA) is required to activate the LtnT permease (10). In vitro experiments using recombinant LtnA and LtnB proteins showed that autophosphorylated LtnB transfers the phosphoryl group from its HisKA domain to the Asp 52 residue in the REC domain of LtnA. Overexpression of LtnA in NA3, but not an LtnA derivative carrying an Asp to Glu amino acid substitution at the phosphorylation site, conferred on the mutant the ability to assimilate nitrate, showing that phosphorylation of LtnA by LtnB alleviates the inhibitory effect of the LtnT C-terminal domain on transport activity. Because the NA3R pseudorevertant has a nonsense mutation, which results in expression of a truncated LtnB protein with no REC domain, it was deduced that the role of the LtnB REC domains is to interfere with the signal transmission to LtnA by quenching the phosphorylation signal in the HisKA domain (10).
In this study, we have identified and characterized a previously unidentified component (LtnC) of this signal transduction system. LtnC consists of the N-terminal HisKA domain, the central REC domain, and the C-terminal domain of a unique amino acid sequence. Our data show that LtnC receives the phosphoryl group of LtnA at the histidine residue in its HisKA domain. Phosphorylation of the protein promotes oligomerization of LtnC, which in turn activates the LtnT activity. The aspartate residue in the REC domain of LtnC is not required for activation of LtnT, indicating that the histidine residue in the HisKA domain acts as the terminal phosphoacceptor that directs transmission of the signal from LtnC to downstream components of the pathway.

EXPERIMENTAL PROCEDURES
Strains and Growth Conditions-A derivative of S. elongatus that was cured of the resident small plasmid pUH24 (R2-SPc (11), hereafter designated as the wild-type strain) and the mutant strains derived from it were grown photoautotrophically at 30°C under continuous illumination provided by fluorescent lamps (70 microeinsteins m Ϫ2 s Ϫ1 ). The basal medium used was a nitrogen-free medium obtained by modification of the BG11 medium (12) as described previously (13). Ammonium-containing medium was prepared by adding 3.75 mM (NH 4 ) 2 SO 4 to the basal medium. Nitrate-containing medium was prepared by adding KNO 3 at indicated concentrations to the basal medium. Solid medium was prepared by adding 1.5% Bacto Agar (Difco) to the liquid medium. All media were buffered with 20 mM HEPES-KOH (pH 8.2). When appropriate, kanamycin (25 g/ml) or chloramphenicol (10 g/ml) or both was added to the media. The Synechococcus strains and plasmids used in this study are listed in Table 1.
Construction of the Insertional Mutants-Site-directed insertional mutants were constructed as described previously by Williams and Szalay (14). DNA fragments that contained complete or partial sequences of the target open reading frames (ORFs) of Synechococcus were amplified by PCR and cloned into pGEM-T Vector (Promega Corp.). A chloramphenicol-resistant (Cm R ) marker (15) was subsequently inserted at a suitable restriction site in each of the ORFs in the same orientation as the ORFs. The resulting plasmids were used to transform of Synechococcus to chloramphenicol resistance through homologous recombination. The transformants were allowed to grow on solid medium supplemented with chloramphenicol. After serial streak purifications to promote segregation of alleles and to isolate homozygous mutants, genomic DNA from selected clones was analyzed by PCR to confirm the presence and position of the Cm R marker. Expression of Plasmid-encoded Proteins in Synechococcus-A shuttle expression vector (pSE1) (9) was used to express cloned genes in Synechococcus. The coding regions of ltnC and ltnD were amplified from the Synechococcus chromosomal DNA by PCR. The 5Ј primers carried mismatches with the genomic sequence, which created a BspHI recognition site at the translation start site without changing the encoded amino acid sequence, and the 3Ј primers carried a BglII or a BamHI recognition sequence immediately downstream of the termination codon. Two ltnC derivatives were generated by overlap extension PCR (16): one had CAT-to-GAG base substitutions at nucleotides 46 -48 that change the histidine residue at position 16 to a Glu residue (H16E), and the other had a G-to-A base substitution at nucleotide 667 that changes the Asp residue at position 223 to an Asn residue (D223N). To express the C-terminal portion of the LtnC and LtnD proteins, the relevant regions of ltnC and ltnD were amplified by PCR using 5Ј primers carrying mismatches with the genomic sequence to create an NcoI recognition site, which provides the initiation codon, at nucleotide position 898 of ltnC and 532 of ltnD. The PCR-amplified ltnC, ltnD, and their derivatives were digested with a combination of BspHI and BglII, BspHI and BamHI, NcoI and BglII, or NcoI and BamHI and cloned between the NcoI and BamHI sites of pSE1.
To express the translational fusions of glutathione S-transferase (GST) and the C-terminal region of LtnC and LtnD, a shuttle expression vector (pSE1GST) was constructed from the pSE1 vector as follows. The GST coding sequence corresponding to amino acids 1-225 of GST of Schistosoma japonicum was amplified by PCR with the plasmid DNA of pGEX-4T3 (Amersham Biosciences). For cloning, the 5Ј primer carried mismatches with the GST sequence, which created a BspHI recognition site at the translation start site without changing the encoded amino acid sequence, and the 3Ј primer carried an NcoI recognition sequence. The PCR-amplified GST gene was cloned into the NcoI site of pSE1 after digestion with BspHI and NcoI. The 3Ј region of the ltnC and ltnD genes amplified by PCR as described above was digested with a combination of NcoI and BglII or NcoI and BamHI and cloned between the NcoI and BamHI sites of pSE1GST. The resulting plasmids were introduced into the Synechococcus cells after verification of the nucleotide sequence.
Preparation of Recombinant Proteins-The expression vector pQE30 (Qiagen) was used to overexpress LtnC and its derivatives as His-tagged proteins. The coding sequences of the ltnC gene and its derivatives encoding LtnC(H16E) and LtnC(D223N) were prepared essentially as described above, except that the 5Ј primers used for PCR carried mismatches with the genomic sequence to create a BglII recognition sequence immediately upstream of the translation start codon. The PCR products were digested with BglII and cloned into the BamHI site of pQE30. After verification of the nucleotide sequence, the resulting plasmids were transformed into Escherichia coli M15 (pREP4) (Qiagen), and His-tagged proteins were purified on nickel-nitrilotriacetic acid resin. To overexpress GST and its fusion with the C-terminal region of LtnC, the pSE1GST plasmid and its derivative carrying the 3Ј region of ltnC were transformed into E. coli JM105, expression of the encoded proteins was induced by 50 M isopropyl-thio-␤-Dgalactopyranoside, and the proteins were purified on glutathione-Sepharose resin. To coexpress LtnA, the histidine kinase domain of LtnB (amino acids 408 -735) and LtnC in E. coli, transcriptional fusion of the ltnA gene and the truncated ltnB gene was constructed by overlap extension PCR (17). The PCR product was digested with a combination of BspHI and XbaI and cloned between the NcoI and XbaI sites of pSE1. The resulting plasmid and the pQE30 derivative carrying ltnC (see above) were cotransformed into E. coli JM105. Expression of the protein was induced with 50 M isopropyl-thio-␤-D-galactopyranoside, and the His-tagged LtnC protein coexpressed with the untagged LtnA and LtnB proteins was purified on nickel-nitrilotriacetic acid resin.
Gel-filtration Chromatography-Purified proteins were concentrated by an Amicon Ultra-4 centrifugal filter device (Millipore) and fractionated on a prepacked HiLoad 16/60 Superdex 200-pg column (Amersham Biosciences) equilibrated with a buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM KCl, and 5 mM dithiothreitol. Proteins were eluted at a flow rate of 0.5 ml/min, and eluted fractions were collected as 2 ml/fraction. The column was size-calibrated with ferritin, catalase, aldolase, bovine serum albumin, chymotrypsinogen A, and cytochrome c.
Other Methods-Chromosomal DNA was extracted from Synechococcus cells and purified as described by Williams (19). Manipulations and analyses of DNA were performed according to standard protocols. Protein concentration was determined with the Bio-Rad Protein Assay (Bio-Rad).

Identification of the ltnC Gene Required for Activation of
LtnT-In a previous study (10), we identified three genes (ltnA, ltnB, and ltnT) required for nitrate transport by the NA3R mutant, a pseudorevertant of the NA3 mutant of S. elongatus lacking the ABC-type NRT. Downstream of the ltnA and -B genes involved in activation of the permease encoded by ltnT are two ORFs syc2279_d and syc2280_d, which are oriented in the same direction as ltnAB and which overlap with the upstream genes by 8 and 77 bases, respectively (Fig. 1). syc2279_d encodes a protein of 411 amino acids, which is similar to hybrid histidine kinase and carries a HisKA domain in the N-terminal region and a REC domain in the central region, although it lacks an HATPase domain. syc2280_d encodes a protein of 271 amino acids and has an effector domain of the cAMP receptor protein family (CAP_ED domain, NCBI accession number cd00038) in the N-terminal region. The C-terminal regions of Syc2279_d and Syc2280_d are 47% identical, but they do not belong to known domains proposed in the conserved domain data base of NCBI, and their function is unknown. To characterize the syc2279_d and syc2280_d genes, we inactivated them by inserting a Cm R cassette in the NA3R pseudorevertant ( Fig. 1) and compared the growth of the resulting mutants with that of the wild-type strain, the NRT-deficient mutant (NA3), the NA3R pseudorevertant, and the previously constructed NA3R derivatives (Fig. 2, strains a-g). The wildtype strain grew equally well on ammonium-or nitrate-containing medium (a). The NA3 mutant grew as well as the wildtype strain on ammonium-containing medium but hardly grew on a medium containing nitrate and showed yellow color because of reduced pigmentation (b), as described previously (9). NA3R cells grew as well as the wild-type strain on plates containing nitrate (c), but interruption of ltnA or ltnB in NA3R abolished the ability to grow on nitrate (d and e), as described previously (10). Interruption of syc2279_d but not syc2280_d in NA3R abolished the ability to grow on nitrate (f and g), indicating that syc2279_d but not syc2280_d was required for the LtnTdependent nitrate transport activity of NA3R. syc2279_d was named ltnC.
Structure-Function Relationship of LtnC-To study the structure-function relationship of LtnC in activating LtnT, we used the pSE1 shuttle expression vector (9) to introduce the ltnC gene and its derivatives into the NA3R derivative carrying insertionally inactivated ltnC and examined the growth of the resulting transformants (Fig. 2). Expression of the LtnC protein and its derivative, whose Asp 223 residue in the REC domain had been replaced with Asn, supported cell growth on plates containing 5 mM nitrate (i and k). In contrast, expression of an LtnC derivative whose His 16 residue in the HisKA domain had been replaced with Glu did not support cell growth on 5 mM nitrate (j). These results indicate that the His 16 residue in the HisKA domain of LtnC is required for activation of the LtnT permease but that the Asp 223 residue in the REC domain of LtnC is not.
When radiolabeled, the 32 P signal was stronger in LtnC than in LtnA (Fig. 3B, lanes 4 -5 and 10 -11), indicating that the phosphoryl group of the HisKA domain of LtnC is more stable than that of the REC domain of LtnA. Adding LtnC(H16E) to the reaction mixture, including LtnA, LtnB, and [␥-32 P]ATP increased the 32 P signal of LtnA (Fig. 3B, lanes 6 and 12), indicating that LtnC(H16E) inhibits the dephosphorylation of LtnA. This suggests that LtnC(H16E) interacts with LtnA even though it cannot receive the phosphoryl group and that His 16 of the HisKA domain is not required for the interaction of LtnC with LtnA.
Promotion of LtnC Oligomerization by Phosphorylation-To investigate the roles of the C-terminal region of the LtnC protein, ltnC and its derivatives were introduced into the S. elongatus NA3 mutant using the shuttle expression vectors pSE1 and pSE1GST. The LtnT activity is latent in NA3, because LtnA is not phosphorylated by LtnB. The pSE1GST vector was used to express a protein as a fusion to GST, which mediates dimer formation. The ability of the resulting transformants to grow on nitrate was examined (Fig. 2). Whereas expression of the genes encoding full-length LtnC or the C-terminal region of LtnC lacking the HisKA and REC domains did not confer the ability to grow on agar plates containing 5 mM nitrate on NA3 (Fig. 2,  n and o), expression of the translational fusion of the C-terminal region of LtnC to GST supported cell growth on nitrate (Fig.  2, p). Expression of GST alone did not support nitrate utilization (Fig. 2, m). These results suggest that the transport activity of the LtnT permease is activated by dimerization of the C-terminal region of LtnC. However, gel-filtration chromatography analysis (Fig.  4) showed that the His-tagged LtnC protein, which has a calculated molecular mass of 47,844 Da, was eluted in a peak corresponding to a calibrated molecular mass of 100 kDa (Fig. 4A, open square), indicating that unphosphorylated LtnC is in a dimeric form. The GST protein, which has a calculated molecular mass of 27,795 Da, was eluted in a peak corresponding to a molecular mass of 50 kDa (Fig. 4A, open circle), confirming that GST is in a dimeric form, as reported previously (20). The fusion of the C-terminal region of LtnC and GST, having a calculated molecular mass of 39,279 Da, was eluted in a broad peak corresponding to a molecular mass range of Ͼ230 kDa (Fig. 4A, closed circle), indicating that the fused protein forms hexamer and higher order oligomers. These results indicate that dimerization of LtnC is not sufficient for activation of LtnT and that higher order oligomerization of the protein is required. To examine whether phosphorylation of LtnC causes high order oligomerization of the protein, His-tagged LtnC protein was coexpressed with untagged LtnA and LtnB-(408 -735) proteins in E. coli cells, purified to homogeneity, and subjected to gelfiltration chromatography. The purified His-tagged LtnC was eluted in two peaks (Fig. 4A, closed square), one of which corresponded to a molecular mass range of Ͼ400 kDa but was not observed when LtnC alone was expressed in E. coli (Fig. 4A,  open square). These results show that phosphorylation of LtnC results in the formation of octamer or higher order oligomers. The results also suggest that LtnT permease is activated by formation of high order oligomers of the C-terminal region of LtnC in the cell.
Identification of the ltnD Gene-Although Syc2280_d is dispensable for manifestation of the LtnT activity in NA3R (discussed above), its C-terminal region is 47% identical to the C-terminal region of LtnC, suggesting that the protein may activate LtnT under certain circumstances. Expression of the full-length Syc2280_d protein or the C-terminal region of Syc2280_d did not confer on NA3 the ability to grow on nitrate (Fig. 2, q and r). However, expression of the fusion of the C-terminal region of Syc2280_d to GST supported cell growth on plates containing 5 mM nitrate (Fig. 2, s). These results suggest that the C-terminal region of Syc2280_d, which is similar to the corresponding region of LtnC, can activate the LtnT permease upon formation of high order oligomers. Thus, we named syc2280_d ltnD, although it is currently unknown when and how the protein forms oligomers in vivo.

DISCUSSION
By characterizing the NA3R mutant of S. elongatus, in which a nitrate transport system distinct from the ABC-type NRT is activated, we previously identified two regulatory genes (ltnA and ltnB) required for activation of the permease encoded by the ltnT gene (10). ltnB encodes a hybrid histidine kinase and ltnA encodes a response regulator with no effector domain. Phosphotransfer from LtnB to LtnA is essential for activation of the LtnT permease (10). However, the next step of this signal transduction pathway was unknown. In the present study, we identified the third component of this phosphorelay system, which is encoded by the gene (ltnC) located downstream of the ltnA and ltnB genes (Fig. 1). LtnC contains a HisKA domain and a REC domain, but because it lacks an HATPase domain, the protein is incapable of autophosphorylation (Fig. 3, lane 1). LtnC is nevertheless capable of accepting the phosphoryl group from LtnA (Fig. 3, lanes 5 and 11), showing that it comprises a phosphorelay system with LtnB and LtnA. Analyses including site-specific amino acid substitution in LtnC showed that the His 16 residue in the HisKA domain is essential for phosphorylation of LtnC in vitro (Fig. 3, lanes 6 and 12) and activation of LtnT in vivo (Fig. 2, i and j). HPt domains are known to mediate signal transmission in multistep (His-Asp-His-Asp) phosphorelay systems. By contrast, the HisKA domain of LtnC seems to be the first example of this type of His-containing domain that accepts a phosphoryl group from a REC domain.
Another unique feature of LtnC is that its C-terminal region acts as a molecular switch that activates the LtnT permease upon formation of high order oligomers (discussed below). Although LtnC has a REC domain, Asp 223 in the REC domain is not required for acceptance of a phosphoryl group from LtnA in vitro or activation of LtnT in vivo. These findings indicate that the phosphorylation of the HisKA domain triggers signal output from the phosphorelay system. The role of the LtnC REC domain is unknown. It may dephosphorylate the HisKA domain to modulate the activity of LtnC in signaling. In the conventional His-Asp phosphorelay systems, namely, a typical two-component system (Fig. 5A) and a multistep phosphorelay  A, the E. coli osmoregulatory system uses a single phosphoryl transfer event between the sensor histidine kinase (EnvZ) and its cognate response regulator (OmpR) (4). B, the Bacillus subtilis sporulation control system uses three phosphoryl transfer events. Spo0F, which contains an intermediate REC domain, receives a phosphoryl group from a sensor histidine kinase (KinA) and subsequently transfers the phosphoryl group to Spo0B, which contains an HPt domain. The phosphoryl group is then transferred to the terminal response regulator (Spo0A) (5). C, the S. elongatus regulatory system of transport activity of the LtnT permease uses two phosphoryl transfer events. LtnA, which contains an intermediate REC domain, receives a phosphoryl group from a sensor histidine kinase (LtnB) and transfers it to the HisKA domain of LtnC. Further details are given in the text. DECEMBER 8, 2006 • VOLUME 281 • NUMBER 49 system (Fig. 5B), response regulators, whose aspartate residues of the REC domains are terminal phosphoacceptors, act as molecular switches to certain cellular events. Thus, the LtnB, LtnA, and LtnC proteins comprise a newly described two-step (His-Asp-His) phosphorelay system that terminates with a "HisKA-containing regulator" carrying an effector domain (Fig. 5C).

Two-step (His-Asp-His) Phosphorelay System
The NCBI conserved domain search program (www.ncbi. nlm.nih.gov/Structure/cdd/wrpsb.cgi) shows 203 proteins that have a HisKA domain but lack an HATPase domain in addition to LtnC. Among these, 37 lack the conserved histidine residue of the HisKA domains and hence would not receive a phosphoryl group. Eight of the others consist only of the HisKA domain and may function as intermediate phosphotransfer molecules like HPt proteins. The remaining 158 proteins, found in the genomes of 4 archaea, 101 bacteria, and 2 eukaryotes, may include HisKA-containing regulators (supplementary Fig. S1). However, only two of these proteins have known effector domains. One, from Neisseria gonorrhoeae FA 1090, has a predicted permease domain (COG0730), a sigma-54 interaction domain (pfam00158), and a Fis (factor for inversion stimulation) domain (HTH_8; pfam02954) (Fig. 6). Downstream of the gene encoding this protein is the pivNG gene encoding a sitespecific recombinase (21), suggesting possible involvement of the HisKA-containing protein in DNA rearrangements. The other putative HisKA-containing regulator, from Bradyrhizobium japonicum USDA 110, has a REC domain and an adenylate cyclase catalytic domain (CyaA; COG2114) ( Fig. 6), but its function remains to be identified. As in the case of LtnC, other HisKA-containing proteins might have uncharacterized effector domains. Because these proteins have no known sequence signatures for DNA binding, they would function as post-translational regulators rather than as transcriptional regulators. It should be noted that 61 of these HisKA-containing proteins have putative signal-accepting domain(s), such as GAF (pfam01590) (22), HAMP (smart00304) (23), PAS (pfam00989) (24), PAC (C-terminal to a subsets of PAS domains; pfam00785), REC, and the predicted periplasmic ligand-binding sensor (COG3452) domains ( Fig. 6 and supplemental Fig.  S1), suggesting that they also accept signal(s) other than HisKA phosphorylation that comprise complex signal-transduction mechanisms.
The C-terminal portion of LtnC activates LtnT when expressed as a fusion protein with GST in the S. elongatus NA3 mutant (Fig. 2, p). Because expression of GST does not activate LtnT in NA3 (Fig. 2, m), we conclude that the C-terminal portion of LtnC acts as the effector domain that activates LtnT permease. However, expressing the domain alone in NA3 failed to activate LtnT (Fig. 2, o). Because the effector domain formed insoluble material when expressed in E. coli, 3 we infer that the fusion to GST contributes to stabilizing the LtnC effector domain in NA3 cells. Whereas GST by itself forms a dimer in solution (Ref. 20; see also Fig. 4A), the fusion of the LtnC effector domain to GST forms hexamer or higher order oligomers (Fig. 4A), indicating that the LtnC effector domain can promote the formation of the high order oligomers. However, LtnC is in a dimeric structure when expressed in E. coli, and the oligomers are formed when coexpressed with LtnB and LtnA in E. coli cells (Fig. 4A). These observations and the known ability of the HisKA domain to mediate dimer formation (25) suggest that LtnC dimer is formed by the interaction between the HisKA domains of two unphosphorylated LtnC molecules and, in these conditions, that the ability of the effector domain to promote oligomerization is inhibited. Phosphorylation of the HisKA domain presumably causes conformational change in LtnC to allow its C-terminal domain to promote oligomer formation. Because phosphorylation of the HisKA domain of LtnC is essential for activation of LtnT in vivo (Fig. 2, i and j), we conclude that phosphorylation-induced oligomerization of the effector domain somehow triggers activation of LtnT. We therefore define the effector domain of LtnC as a "high order oligomerization-dependent regulatory (HODR)" domain.
The gene located downstream of ltnC, which we named ltnD, encodes a protein whose C-terminal portion can activate LtnT when expressed as a fusion to GST (Fig. 2, s). Because the C-terminal portion of this protein and the LtnC effector domain are 47% identical in amino acid sequence (Fig. 7A), it seems reasonable to suppose that the C-terminal domain of LtnD is the second member of HODR domains, although oligomer formation remains to be verified experimentally. Because the N-terminal portion of the LtnD protein comprises a CAP_ED domain, this domain seems to accept an as-yet unidentified signal to promote oligomerization of the protein to activate the LtnT permease. It is currently unknown whether the oligomerized HODR domains interact with LtnT or whether one or more additional signaling proteins are involved in activation of LtnT.
Further study is required to clarify the molecular mechanism of LtnT activation.
Proteins containing HODR-like sequences are found only in proteins from cyanobacteria Thermosynechococcus elongatus BP-1 (Tll2374), Crocosphaera watsonii WH 8501 (Cwat_3664), Nostoc sp. PCC 7120 (All2034), Anabaena variabilis ATCC 29413 (Ava_1137), and Nostoc punctiforme PCC 73102 (Npun02006412, Npun02004888, and Npun02006607) with the exception of Methylococcus capsulatus str. Bath (MCA0452) (Fig. 7A). In all cases, the putative HODR domains are located in the C-terminal region of the respective protein (Fig. 7B). All of these proteins except for Cwat_3664 contain putative signal-accepting domain(s), such as HAMP, PAS, Cache (pfam02743) (26), and COG4192 (Fig. 7B). Thus, these proteins are likely to be involved in signal-transduction mechanisms in which their reception of a specific signal initiates the formation of high order oligomers to regulate certain cellular processes.