Characterization of the Nuclear- and Plastid-Encoded secA-Homologous Genesin the Unicellular Red Alga Cyanidioschyzon merolae

Yosuke KOYAMA,1 Koji TAKIMOTO,1 Asuka KOJIMA,1 Kei ASAI,1 Satoshi MATSUOKA,1

Toshiaki MITSUI,2 Kouji MATSUMOTO,1;y Hiroshi HARA,1 and Niji OHTA1

1Department of Biochemistry and Molecular Biology, Graduate School of Science and Engineering,Saitama University, 255 Shimo-Ohkubo, Sakura-ku, Saitama 338-8570, Japan2Graduate School of Science and Technology, Niigata University, Niigata 950-2181, Japan

Received April 28, 2011; Accepted August 19, 2011; Online Publication, October 7, 2011

[doi:10.1271/bbb.110338]

SecA is an ATP-driven motor for protein trans-location in bacteria and plants. Mycobacteria andlisteria were recently found to possess two functionallydistinct secA genes. In this study, we found thatCyanidioschyzon merolae, a unicellular red alga, pos-sessed two distinct secA-homologous genes; one encodedin the cell nucleus and the other in the plastid genome.We found that the plastid-encoded SecA homologshowed significant ATPase activity at low temperature,and that the ATPase activity of the nuclear-encodedSecA homolog showed significant activity at hightemperature. We propose that the two SecA homologsplay different roles in protein translocation.

Key words: secA; Cyanidioschyzon merolae; ATPaseactivity; temperature dependence

Protein translocation via Sec translocase is a well-studied protein transport mechanism. In bacteria, thebulk of protein export across the cytoplasmic membraneis carried out by Sec translocase. Its components havebeen well-characterized in studies with Escherichia coliand Bacillus subtilis.1) Bacterial Sec translocase consistsmainly of a highly conserved protein-conducting chan-nel, SecYEG, and a peripheral ATPase motor SecA.ATPase SecA couples ATP hydrolytic energy to proteintranslocation.2) Through repeated cycles of ATP bindingand hydrolysis, SecA undergoes dynamic conforma-tional changes that drive the stepwise translocation ofpre-protein through the SecYEG channel.3)

Sec translocase is conserved also in protein trans-location pathways into the endoplasmic reticulum (ER)and the thylakoid lumen or membrane of the chloroplastin eukaryotic cells.4–6) The post-translational transloca-tion pathway into the ER consists mainly of Sec61���(a homolog of SecYEG), the tetrameric Sec62/63complex, and luminal chaperone BiP. It is believed thatthis pathway does not include a peripheral ‘‘pushing’’motor protein, like SecA.7)

In plants and algae, SecYE homologs exist onthylakoid membranes and probably form the protein-conducting channel,8) and a SecA homolog exists in thechloroplast stroma.9) Chloroplast SecA is responsiblefor the translocation of thylakoid proteins, such as the33-kDa polypeptide of the oxygen-evolving complex of

photosystem II, plastocyanin, the F subunit of photo-system I, and cytochrome f.10,11) Chloroplast SecA isbelieved to drive the translocation of the precursorproteins into thylakoid lumen or membrane from stromain a manner similar to that employed by bacterialSecA,12) and is thought to play an essential role inchloroplast biogenesis.13)

Most bacteria possess only a single essential secAgene, but mycobacteria and listeria possess two non-redundant secA-homologous genes.14,15) One is essentialfor viability (secA1) like the secA genes of manybacteria, but the other (secA2) is not. It has beensuggested that SecA2 is dedicated to exporting specificsubsets of proteins related to bacterial virulence.16,17)

It shows intrinsic ATPase activity. Hence it has beensuggested that it functions in a manner similar to E. coliSecA in undergoing cycles of ATP hydrolysis.18)

Cyanidioschyzon merolae is an ultra-small unicellularred alga living in acidic hot springs. Most algaeinvestigated to date encode only a single secA-homol-ogous gene, which is found in the plastid genome. Incontrast, complete genome sequencing identified twosecA-homologous genes in C. merolae, one in the cellnucleus and the other in the plastid genome (http://merolae.biol.s.u-tokyo.ac.jp).19–21) The differences be-tween and the significance of the two secA genes inmycobacteria and listeria have been well investigated,but they have not received attention in plants untilrecently. Here we report the first characterization of apair of secA-homologous genes in a single plant species,and we include an examination of the ATPase activityof the two corresponding SecA homologs. In this paper,we refer to the nuclear-encoded secA-homologousgene (C. merolae genome project ID: CMQ393C) assecA(nuc), and to the plastid-encoded secA-homologousgene (C. merolae genome project ID: CMV071C) assecA(pt) to distinguish them. We conducted ATPaseassay at various temperatures and found that SecA(nuc)and SecA(pt) can be distinguished by the way its activitydepends on temperature. The data provided here giveclues to understanding the functions of the two secA-homologous genes in C. merolae.The protein encoded by secA(nuc) is a 113-kDa

protein of 1,011 amino acids, 46% identical to thechloroplast pre-protein translocase of Arabidopsis thali-

y To whom correspondence should be addressed. Tel: +81-48-858-3406; Fax: +81-48-858-3384; E-mail: koumatsu@mail.saitama-u.ac.jp

Biosci. Biotechnol. Biochem., 75 (10), 2073–2078, 2011

Communication

ana (AtcpSecA) and 44% identical to the SecA ofSynechocystis sp. PCC 6803 (SySecA). The otherhomologous gene, secA(pt), encodes a 90-kDa proteinof 774 amino acids, 59% and 58% identical to AtcpSecAand the SySecA respectively. SecA(nuc) was 46%identical to SecA(pt). Figure 1A shows the amino acidsequence alignments of the N-terminal regions, includingthe SecA DEAD-like domains of AtcpSecA, SecA(nuc),SecA(pt), and SySecA. The conserved Walker A and Bmotifs, which are required for optimal ATPase activity,were found in the amino acid sequences of both SecAhomologs of C. merolae. The conserved lysine residue inthe Walker A motif was also found in both SecAhomologs (Fig. 1A, boldface). Many genes that functionin chloroplasts are encoded in cell nuclei and include asequence coding for N-terminal transit peptides. The

secA gene of A. thaliana is also encoded in the cellnucleus. A possible cleavage site of the transit peptidewas found between positions 22 and 23 in AtcpSecA andbetween 41 and 42 in SecA(nuc) (Fig. 1A, arrow) bySignalP. The subcellular localizations of AtcpSecA andSecA(nuc) led to an inference of ‘‘Chloroplast’’ by theWoLF PSORT program (http://wolfpsort.org/).22)

GFP fused N-terminal 130 amino acids of SecA(nuc)(SecAnuc-N130-GFP) were targeted to the plastid, whenthe plasmid DNA was introduced by particle bombard-ment into onion epidermal cells. As a control, we usedWxTP-DsRed, containing the first 1 to 111 amino acidresidues, including the transit peptide, of the rice waxygene, for plastid targeting (Fig. 2A).23) WxTP-DsRedand SecAnuc-N130-GFP were co-localized in the plastid(Fig. 2A, Merged), indicating that the N-terminal region

B

A

Fig. 1. Multiple Alignment and Putative Domains of C. merolae SecA Homologs.A, Multiple alignment of amino acid sequences of N-terminal regions including the SecA DEAD-like domains of SecA proteins. The amino

acid sequences of the SecA homologs, deduced from the nucleotide sequences of A. thaliana, C. merolae nuclear-encoded, C. merolae plastid-encoded, and the Synechocystis sp. PCC 6803 secA genes, were aligned with the Clustal X program. Amino acid residues shared by all the SecAhomologs are shown by asterisks. Putative Walker A and B motifs are enclosed by squares. The conserved lysine residues in the Walker A motifare indicated in boldface. The arrow indicates a possible cleavage site of the C. merolae nuclear-encoded SecA homolog predicted by SignalP.B, Putative protein domains of C. merolae SecA homologs. Three domains (SecA DEAD-like, pre-protein cross-linking, and Wing and Scaffold)characteristic of SecA proteins are shown. These domains were predicted by the Pfam database (http://pfam.sanger.ac.uk/).

2074 Y. KOYAMA et al.

of SecA(nuc) includes the transit peptide for plastidtargeting. GFP from vector plasmid pUC18-CaMV35S-sGFP(S65T)-NOS was observed in the cytoplasm of theepidermal cells (Fig. 2B).24)

The SecA protein includes several domains requiredfor protein translocation (Fig. 1B). Putative domainscharacteristic of SecA proteins, the SecA DEAD-likedomain (Pfam ID: PF07517) representing the N-terminalATP-dependent helicase domain, the SecA-preproteincross-linking domain (Pfam ID: PF01043), the SecAWing and Scaffold domain (Pfam ID: PF07516) forassociation with SecY, all predicted in the Pfam data-base (http://pfam.sanger.ac.uk/),25) were found in theC. merolae SecA homologs.

To determine whether the SecA(nuc) and SecA(pt)function as SecA, we performed complementation testsusing E. coli strain MM52, a temperature-sensitive secAmutant.26) Full-length SecA(nuc) and SecA(pt) did notcomplement the temperature-sensitive secA at 42 �C.Hence we tested constructs encoding chimeric proteinswhose C-terminal regions SecA(nuc) or SecA(pt) werereplaced with the C-terminal regions of E. coli SecA.Only the chimeric protein composed of the N-terminalregion of SecA(pt) and the C-terminal region of E. coliSecA partially complemented temperature-sensitivesecA (data not shown). That of SecA(nuc) did not.Chloroplast SecA and E. coli SecA are reported to havepreferences for the distinct lipid and signal peptide.27)

Thus there might be distinct preferences for interactionwith SecY proteins between them. These distinct

preferences between chloroplast and E. coli SecA mightbe one of the reasons that the full-length C. merolaeSecA homologs did not complement the temperature-sensitive SecA of E. coli.Expression of the secA(nuc) and secA(pt) genes in

C. merolae was examined by RT-PCR. Total RNAisolated from C. merolae cells was used as the templatefor cDNA synthesis. PCR products amplified withspecific primers for the various secA-homologous geneswere detected with ethidium bromide. No bands weredetected in the absence of reverse transcriptase, indicat-ing that the reaction mixtures were uncontaminated bygenomic DNA (Supplemental Fig. S1A, �RTase; seeBiosci. Biotechnol. Biochem. Web site). In contrast, thebands derived from transcripts of the secA-homologousgenes appeared clearly when reverse transcriptase wasadded (Supplemental Fig. S1A, þRTase), indicatingthat the two homologous genes are not pseudogenes andhave certain functions. The full open reading frames ofsecA(nuc) and secA(pt) and their expression wereconfirmed by these results and previous DNA micro-array analyses.28,29)

In order to determine the expression patterns ofsecA(nuc) and secA(pt) during light-dark cycles, relativetranscript levels were assessed by real-time RT-PCR.C. merolae was cultivated under a 12-h light/12-h darkcycle, and total RNA was isolated every 4 h under lightand every 2 h under dark (Supplemental Fig. S1B).In the first half of the dark period (from L12 to D6),the transcript levels of both secA(nuc) and secA(pt)decreased gradually and consequently showed minimallevels in the middle of the dark period (D6). In the latterhalf of the dark period (from D6 to D12), the transcriptlevels appeared remain at the minimal level or toincrease gradually toward the light period. There did notappear to be obvious differences in the expressionpatterns of secA(nuc) and secA(pt) under this light-darkcycle. The transcript level of secY in the dark perioddecreased slightly. This decrease in transcript level didnot appear to be remarkable under the light-dark cycle,compared with that of the secA-homologous genes. Thisdecrease and increase of the transcripts might have beendue to the periodic switching of the lighting conditions.The gradual decrease may serve to keep a minimalrequired transcript level, and the gradual increase maybe preparation for the adaptation to the light period.Indeed, chloroplast SecA is responsible for the trans-location of proteins related to photosynthesis,9–11) andthe secA gene of A. thaliana is light-inducible and isexpressed in green tissues.13)

In order to gain an understanding of the evolutionaryrelationship between secA(nuc) and secA(pt), we con-structed a phylogenetic tree inferred from the putativeSecA proteins by the neighbor-joining method(Fig. 3).30) Homologous sequences were obtained byBLAST against the NCBI database with E-values of<1� e�10. Alignments of homologous sequences weremade with the Clustal W program in MEGA 4.0software using the default parameter sets (http://www.megasoftware.net/). In the phylogenetic tree, the bacte-rial SecA proteins were used as the out-group. The plantand cyanobacterial SecA contributes to the proteinimport into the thylakoid lumen, whereas the bacterialSecA contributes to protein secretion.

ASecAnuc-N130-GFP Control (GFP alone)

Bright field

GFP

RFP(WxTP-DsRed)

Merged

B

Fig. 2. Transient Expression of the SecAnuc-N130-GFP FusionProtein in Onion Epidermal Cells.To determine whether the N-terminal region of SecA(nuc)

includes a transit peptide for localization to plastids, we constructedpSecAnuc-N130-GFP. To control GFP alone, we used vector plasmidCaMV35S-sGFP(S65T)-NOS. The plasmid DNAs were introducedinto onion epidermal cells by particle bombardment. WxTP-DsRed(red fluorescence) was used as control for plastid targeting. A, Onionepidermal cells were transformed with pSecAnuc-N130-GFP andpWxTP-DsRed. B, Epidermal cells were transformed withCaMV35S-sGFP(S65T)-NOS plasmid and pWxTP-DsRed. Scalebars, 100mm.

Characterization of Two SecA Homologs in C. merolae 2075

C. merolae SecA(nuc) constituted a cluster distinctfrom that of SecA(pt) (the node is indicated by anarrow). In Phaeodactylum tricornutum, the nuclear-encoded SecA homolog also constituted a clusterdistinct from that of the plastid-encoded SecA homolog.The phylogenetic tree suggests that the nuclear-encodedsecA-homologous genes belong to the plant/cyanobac-terial cluster, and may have branched off in an earlydivergent evolutional step or may have been acquiredthrough endosymbiotic/horizontal gene transfer events.On the other hand, two SecA homologs of Chlamydo-monas reinhardtii (here denoted SecA1 and SecA2 forexpedience), which adjoin in the cell nucleus, wereclosely related. The tree led us to assume that thetwo nuclear-encoded secA-homologous genes of C.reinhardtii are closely related phylogenetically andarose in a gene duplication event. Additionally, weinvestigated the evolutionary relationship between bac-terial SecA1 and SecA2. The SecA2 proteins ofM. tuberculosis and L. monocytogenes, which have beenreported to function distinctly from the canonicalSecA1,17–20) formed clusters distinct from those ofcanonical SecA1 proteins (data not shown). Consideringour phylogenetic analysis, the distant relationshipbetween secA(nuc) and secA(pt) suggests distinct func-tions of the two SecA homologs in C. merolae.

To determine whether SecA(nuc) and SecA(pt)possess intrinsic ATPase activities, the recombinantproteins were overexpressed as N-terminal His-tag fusedproteins and purified with a Niþ-charged column closeto homogeneity. Overexpression of the proteins wasconfirmed by Western blotting using anti-His6 mono-clonal antibody (Supplemental Fig. S2A). The purity of

the eluted fraction was checked by Coomassie BrilliantBlue (CBB) staining (Supplemental Fig. S2B).We performed ATPase assays of these purified SecA

homologs using malachite green. The purified SecAhomologs showed ATPase activities. When the con-served lysine residues in the Walker A motif of theproteins were replaced with arginine residues by site-directed mutagenesis, the mutated SecA homologs,SecA(nuc)K88R and SecA(pt)K87R, showed littleATPase activity (Fig. 4A). These results indicate thatSecA(nuc) and SecA(pt) indeed possess intrinsicATPase activities and that the conserved lysine residueis required for the ATPase activities of both C. merolaeSecA homologs. Substitutions at the invariable lysineresidue in the Walker A motif block the translocationactivity of E. coli and B. subtilis SecA.31) The ATPaseactivity is thus almost certainly essential to the proteintranslocation activity of SecA. Based on these observa-tions and our data, we surmise that both SecA(nuc) andSecA(pt) function in protein translocation.The intrinsic ATPase activity of the purified SecA

homologs was measured at various temperatures(Fig. 4B). SecA(pt) showed high ATPase activity at30, 35, and 40 �C (approximately 100 pmol Pi/mg/min),but showed very low activity at 45, 50, 55, and 60 �C(less than 5 pmol Pi/mg/min). The ATPase activity ofSecA(nuc) increased gradually as the temperature rose,and it was highest at 55 �C (approximately 70 pmol Pi/mg/min). Even at 60 �C, SecA(nuc) showed significantATPase activity (approximately 20 pmol Pi/mg/min).C. merolae lives in acidic hot water, as in hot springs,

at about 45 �C. Temperatures of less than 40 �C might below for C. merolae. Since SecA(pt) showed high

Plant /Cyanobacteria

Bacteria

Cyanidium caldarium (plastid)Cyanidioschyzon merolae (plastid)

Odontella sinensis (plastid)Phaeodactylum tricornutum (plastid)

Porphyra yezoensis (plastid)

Guillardia theta (plastid)Rhodomonas salina (plastid)

Prochlorococcus marinus

Gloeobacter violaceus PCC 7421Synechococcus elongatus PCC 6301

Synechocystis sp. PCC 6803Nostoc sp. PCC 7120

Anabaena variabilis ATCC 29413

Chlamydomonas reinhardtii SecA1 (nuclear)Chlamydomonas reinhardtii SecA2 (nuclear)

Spinacia oleracea (nuclear)

Arabidopsis thaliana (nuclear)Pisum sativum (nuclear)

Phaeodactylum tricornutum (nuclear)Cyanidioschyzon merolae (nuclear)

Escherichia coli

Pseudomonas aeruginosaListeria monocytogenes SecA2

Bacillus subtilis

Listeria monocytogenes SecA1Clostridium thermocellum

Clostridium difficile

99

99

99

99

9399

97

99

97

99

91

99

6597

99

99

99

90

57

66

88

0.05

Fig. 3. Phylogenetic Tree Inferred from the Amino Acid Sequences of SecA-Homologous Proteins.The amino acid sequences were deduced from the nucleotide sequences of genes annotated as secA. The accession numbers of the secA genes

of the organisms that appear in the phylogenetic tree are shown in Supplemental Table S1. The tree was constructed by the neighbor-joiningmethod with MEGA 4.0 software. The neighbor-joining consensus tree used 1,000 bootstrap replicates. The numbers represent percentages ofbootstrap values. The bacterial SecA proteins were used as out-group. The scale bar represents 0.05 mutations/site, and branch lengths are drawnto scale. The arrow indicates the node that separates a cluster including SecA(pt) from that of SecA(nuc). (plastid), plastid-encoded; (nuclear),nuclear-encoded.

2076 Y. KOYAMA et al.

ATPase activity at low temperatures (at 30, 35, and40 �C), it appears seem to possess protein translocationactivity at low temperatures, but it had low ATPaseactivity at high temperatures (at 45, 50, 55, and 60 �C).In contrast, SecA(nuc) showed significant ATPaseactivity at the higher temperatures. Based on thisdifference in the temperature dependence of the ATPaseactivity, we assume that SecA(nuc) is required forprotein translocation at high temperatures. ChloroplastSecA is essential for photosynthetic development inA. thaliana.13) C. merolae is an obligete photoautotro-phic organism that can grow at 50 �C. Although there isa possibility that SecA(nuc) transports a subset ofproteins distinct from the substrate proteins of SecA(pt),such as mycobacterial and listerial SecA2, we supposethat C. merolae preserved (or acquired) and evolved thesecA(nuc) gene in adapting to high temperature environ-ments. A recent study indicates that A. thaliana pos-sesses two secA genes with distinct functions.32) Thestudies described here are intended to lay the ground-work for further investigation of the two SecA homologsin plants.

Acknowledgments

We thank Dr. Yasuo Niwa and Dr. Toshihisa Kotakefor providing pUC18-CaMV35S-sGFP(S65T)-NOS vec-tor, and Dr. Daisuke Takezawa and Dr. Shin Kore-edafor suggestions and technical support with PDS-1000/He. We also thank Professor Emeritus Yoshito Sadaie,

Professor Yasuhiro Takahashi, Professor Ikuo Nishida,and Dr. Yoshitaka Nishiyama for valuable advice andhelpful discussion.

References

1) Driessen AJM and Nouwen N, Annu. Rev. Biochem., 77, 643–

667 (2008).

2) Lill R, Cunningham D, Brundage LA, Ito K, Oliver D, and

Wickner W, EMBO J., 8, 961–966 (1989).

3) Hunt JF, Weinkauf S, Henry L, Fak JJ, McNicholas P, Oliver

DB, and Deisenhofer J, Science, 297, 2018–2026 (2002).

4) Rapoport TA, Nature, 450, 663–669 (2007).

5) Mori H and Cline K, Biochim. Biophys. Acta, 1541, 80–90

(2001).

6) Aldridge C, Cain C, and Robinson C, FEBS J., 276, 1177–1186

(2009).

7) Osborne AR, Rapoport TA, and van den Berg B, Annu. Rev.

Cell Dev. Biol., 21, 529–550 (2005).

8) Schuenemann D, Amin P, Hartmann E, and Hoffman NE,

J. Biol. Chem., 274, 12177–12182 (1999).

9) Nakai M, Goto A, Nohara T, Sugita D, and Endo T, J. Biol.

Chem., 269, 31338–31341 (1994).

10) Cline K and Henry R, Annu. Rev. Cell Dev. Biol., 12, 1–26

(1996).

11) Nohara T, Asai T, Nakai M, Sugiura M, and Endo T, Biochem.

Biophys. Res. Commun., 224, 474–478 (1996).

12) Hulford A, Hazell L, Mould RM, and Robinson C, J. Biol.

Chem., 269, 3251–3256 (1994).

13) Liu D, Gong Q, Ma Y, Li P, Li J, Yang S, Yuan L, Yu Y,

Pan D, Xu F, and Wang NN, J. Exp. Bot., 61, 1655–1669

(2010).

14) Braunstein M, Brown AM, Kurtz S, and Jacobs Jr WR,

J. Bacteriol., 183, 6979–6990 (2001).

A

B

AT

Pas

e ac

tivi

ty (

pm

ol P

i/µµg

/min

)A

TP

ase

acti

vity

( p

mo

l Pi/

µg/m

in )

Fig. 4. ATPase Activities of C. merolae SecA Homologs.A, Effects of the conserved lysine residue substitution in the Walker A motifs on the intrinsic ATPase activities of the C. merolae SecA

homologs. The lysine residues in the Walker A motifs were replaced with arginine residues by site-directed mutagenesis. The mutated SecAhomologs are designated SecA(nuc)K88R and SecA(pt)K87R. The ATPase activity levels of SecA(nuc) wild-type(WT) and SecA(nuc)K88Rand those of SecA(pt)WT and SecA(pt)K87R were determined at 50 �C and 30 �C respectively by malachite green assay to determine freeinorganic phosphate. B, Intrinsic ATPase activities of C. merolae SecA homologs at 30–60 �C. Each assay was performed in triplicate. Averagesand standard deviations are shown.

Characterization of Two SecA Homologs in C. merolae 2077

15) Lenz LL and Portnoy DA, J. Mol. Biol., 45, 1043–1056

(2002).

16) Lenz LL, Mohammadi S, Geissier A, and Portnoy DA, Proc.

Natl. Acad. Sci., 100, 12432–12437 (2003).

17) Kurtz S, McKinnon KP, Runge MS, Ting JPY, and Braunstein

M, Infect. Immun., 74, 6855–6864 (2006).

18) Hou JM, D’Lima NG, Rigel NW, Gibbons HS, McCann JR,

Braunstein M, and Teschke CM, J. Bacteriol., 190, 4880–4887

(2008).

19) Matsuzaki M, Misumi O, Shin-I T, Maruyama S, Takahara SM,

Miyagishima SY, Mori T, Nishida K, Yagisawa F, Nishida K,

Yoshida Y, Nishimura Y, Nakao S, Kobayashi T, Momoyama

Y, Higashiyama T, Minoda A, Sano M, Nomoto H, Oishi K,

Hayashi H, Ohta F, Nishizaka S, Haga S, Miura S, Morishita T,

Kabeya Y, Terasawa K, Suzuki Y, Ishii Y, Asakawa S, Takano

H, Ohta N, Kuroiwa H, Tanaka K, Shimizu N, Sugano S, Sato

N, Nozaki H, Ogasawara N, Kohara Y, and Kuroiwa T, Nature,

428, 653–657 (2004).

20) Ohta N, Matsuzaki M, Misumi O, Miyagishima SY, Nozaki H,

Tanaka K, Shin-I T, Kohara Y, and Kuroiwa T, DNA Res., 10,

67–77 (2003).

21) Ohta N, Sato N, and Kuroiwa T, Nucleic Acids Res., 26, 5190–

5198 (1998).

22) Horton P, Park KJ, Obayashi T, Fujita N, Harada H, Collier

CJA, and Nakai K, Nucleic Acids Res., 35, W585–W587 (2007).

23) Kitajima A, Asatsuma S, Okada H, Hamada Y, Kaneko K,

Nanjo Y, Kawagoe Y, Toyooka K, Katsuoka K, Takeuchi M,

Nakano A, and Mitsui T, Plant Cell, 21, 2844–2858 (2009).

24) Chiu W-I, Niwa Y, Zeng W, Hirano T, Kobayashi H, and Sheen

J, Curr. Biol., 6, 325–330 (1996).

25) Kumar S, Dudley J, Nei M, and Tamura K, Brief. Bioinform., 9,

299–306 (2008).

26) Oliver DB and Beckwith J, J. Bacteriol., 150, 686–691 (1982).

27) Sun C, Rusch SL, Kim J, and Kendall DA, J. Bacteriol., 189,

1171–1175 (2007).

28) Minoda A, Nagasawa K, Hanaoka M, Horiuchi M, Takahashi H,

and Tanaka K, Plant Mol. Biol., 59, 375–385 (2005).

29) Imamura S, Kanesaki Y, Ohnuma M, Inouye T, Sekine Y,

Fujiwara T, Kuroiwa T, and Tanaka K, Proc. Natl. Acad. Sci.

USA, 106, 12548–12553 (2009).

30) Saitou N and Nei M, Mol. Biol. Evol., 4, 406–425 (1987).

31) Klose M, Schimz KL, van der Wolk J, Driessen AJM, and

Freudl R, J. Biol. Chem., 268, 4504–4510 (1993).

32) Skalitzky CA, Martin JR, Harwood JH, Beirne JJ, Adamczyk

BJ, Heck GR, Cline K, and Fernandez DE, Plant Physiol., 155,

354–369 (2011).

2078 Y. KOYAMA et al.