The Two-Component Signal System in Rice (Oryza Sativa

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Genomics 89 (2007) 697–707 www.elsevier.com/locate/ygeno

The two-component signal system in rice (Oryza sativa L.): A genome-wide study of cytokinin signal perception and transduction ⁎ ⁎ Liming Du a,1, Fangchan Jiao a,1, Jun Chu a, Gulei Jin b, Ming Chen a,c, , Ping Wu a,

a The Key State Laboratory of Plant Physiology and Biochemistry, College of Life Sciences, Zijingang Campus, Hangzhou 310058, People’s Republic of China b Institute of Bioinformatics, Zhejiang University, Huajiachi Campus, Hangzhou 310029, People’s Republic of China c Department of Bioinformatics, College of Life Sciences, Zhejiang University, Zijingang Campus, Hangzhou 310058, People’s Republic of China Received 23 June 2006; accepted 22 February 2007 Available online 6 April 2007

Abstract

In this report we define the genes of two-component regulatory systems in rice through a comprehensive computational analysis of rice (Oryza sativa L.) genome sequence databases. Thirty-seven genes were identified, including 5 HKs (cytokinin-response histidine protein kinase) (OsHK1–4, OsHKL1), 5 HPs (histidine phosphotransfer proteins) (OsHP1–5), 15 type-A RRs (response regulators) (OsRR1–15), 7 type B RR genes (OsRR16–22), and 5 predicted pseudo-response regulators (OsPRR1–5). Protein motif organization, gene structure, phylogenetic analysis, chromosomal location, and comparative analysis between rice, maize, and Arabidopsis are described. Full-length cDNA clones of each gene were isolated from rice. Heterologous expression of each of the OsHKs in yeast mutants conferred histidine kinase function in a cytokinin-dependent manner. Nonconserved regions of individual cDNAs were used as probes in expression profiling experiments. This work provides a foundation for future functional dissection of the rice cytokinin two-component signaling pathway. © 2007 Elsevier Inc. All rights reserved.

Keywords: Cytokinin two-component system; Rice; Histidine kinase; Histidine phosphotransfer proteins; Response regulator; Genomic analysis; Phylogenetic analysis; Yeast experiment; RT-PCR

Cytokinins have been implicated in many different develop- Three Arabidopsis HK genes (AHK4/CRE1/WOL, AHK2, mental and physiological processes of plants [1–5]. The histidyl– and AHK3) and three maize HK genes (ZmHK1, ZmHK2, and aspartyl (His–Asp) phosphorelay systems, also known as ZmHK3a) have been identified [10,14–17]. Plant HKs typically two-component regulatory systems and which are involved in contain a conserved transmitter domain and a C-terminal cytokinin signal transduction, have been studied extensively in receiver domain. The transmitter domain includes characteristic bacteria, Arabidopsis, and maize, which use these systems to sequence motifs and a conserved His residue that is the site of sense and respond to diverse environmental stimuli [6–10].The autophosphorylation [18]. All HKs share a domain in the signaling cascade typically consists of three functional modules: predicted extracytoplasmic region, designated CHASE (cyclase/ a sensory histidine kinase (HK), a histidine phosphotransfer histidine kinase-associated sensory extracellular), which is the protein (HP), and a response regulator (RR) [11].Inthese putative recognition site for cytokinins [17,19,20]. Cytokinin systems, signals are transmitted by a phosphoryl group transfer receptors have functions in shoot growth, leaf senescence, seed between His and Asp residues. Cytokinins are sensed by size, germination, root development, and cytokinin metabolism membrane-located HK receptors that transmit signals via HPs [21]. In rice, five genes encoding putative sensory histidine to nuclear RRs that activate or repress transcription [12,13]. kinases (CRL-1a, CRL-1b, CRL2, CRL3, CRL4), now renamed OsHK1–4 and OsHKL1, have been identified [22]. ⁎ Corresponding authors. M. Chen is to be contacted at Department of In Arabidopsis, the downstream signaling components of the Bioinformatics, College of Life Sciences, Zhejiang University, Zijingang cytokinin signal-transduction pathway consist of five HPs Campus, Hangzhou 310058, People’s Republic of China. – E-mail addresses: [email protected] (M. Chen), [email protected] (AHPs) [23] and 22 RRs (ARRs) [13,24 27]. Three ZmHP (P. Wu). genes and 10 ZmRR genes have been found in maize 1 These authors contributed equally to this work. [5,9,10,28]. The HP genes contain the highly conserved

XHQXKGSSXS motif that mediates the transfer of a phosphate ssp. indica rice and provide a framework for future functional from the receiver domain of an activated hybrid histidine kinase dissection of the cytokinin two-component systems in plant to the receiver domain of a response regulator in a multistep signal transduction. phosphorelay signal transduction pathway [25]. The RRs are divided into two major classes, type A and type Results B, based on a phylogenetic analysis of their amino acid sequences and their domain structures. The type-A ARRs are Genes comprising the rice cytokinin two-component system primary cytokinin response genes [29] and consist of a receiver domain with a short extension at both the N- and the C-terminal Conserved domains derived from an alignment of sequences ends. The transcription levels of some type-A ARRs are also from well-known plants (Arabidopsis and maize) were used to responsive to various environmental stresses and nitrogen levels identify two-component system genes in rice. After cross- [12,30]. The type-B ARRs contain a receiver domain and a searching the japonica rice genome (cv. Nipponbare) and large C-terminal output domain containing features of tran- the indica genome (cv. 9311) through the NCBI, TIGR, and scription factors [13,31,32] as well as a type-B signature GARP Gramene databases, we identified the unique hits and removed motif of ∼60 amino acids that is distantly related to the Myb duplications from our data set. Based on the conserved do- repeat of transcription factors [33,34]. The GARP domain main analysis, a total of 5 hybrid histidine kinases (OsHK1–4, binding sites are present in the promoters of type-A ARRs, OsHKL1), 5 histidine phospho-transmitters (OsHP1–5), 15 suggesting that type-B ARRs may bind to these promoters and type-A response regulator genes (OsRR1–15), 7 type-B response activate transcription [34]. Consistent with their function as regulator genes (OsRR16–22), and 5 predicted pseudo-response transcription factors, the C-terminal region of type-B ARRs regulators (OsPRR1–5) were identified as components of the contains nuclear localization sequences and P/Q-rich, acidic cytokinin two-component signaling system. domains [31]. Structural analysis suggests that monomeric A summary including FL-cDNA numbers, BAC accession type-B ARRs might bind DNA [35]; and they were found in numbers, protein length, chromosome location, BAC/clone the nucleus [25,34,36,37]. Unlike the expression pattern of coordinates, and corresponding ESTs for each of the predicted type-A ARR genes, the steady-state levels of type-B ARR genes of the cytokinin two-component signaling system in rice is transcripts are not affected by the application of either presented in Table 1. Among all the genes, one hybrid histidine cytokinins or nitrate or other plant hormones [12,13,31]. kinase containing only a full CHASE domain without other Various Arabidopsis type-B ARR mutants showed decreased conserved domains was identified in japonica and designated as sensitivity to cytokinin, including effects on root elongation, OsHK-like 1 (OsHKL1). Five genes that were structurally similar lateral root formation, callus induction and greening, and to Arabidopsis APRRs were designated OsPRRs [43,44]. induction of cytokinin primary response genes [38]. AHK3 mediated leaf longevity through a specific phosphor- Multiple sequence alignments ylation of a response regulator, ARR2. However, the actual molecular mechanisms and the specificity of cytokinin signal- Prior to examining in detail the domain organization of HK ing in controlling the various cytokinin-mediated responses are proteins, multiple sequence alignments of the cytokinin receptor still largely unknown [39]. Elucidation of expression patterns of protein sequences from rice, Arabidopsis, and maize were con- His–Asp phosphorelay components and their distribution at the ducted (Supplemental Fig. 1A). Alignments of the proteins of subcellular level will be important for understanding the His–Asp phosphorelay signaling factors revealed high homol- cytokinin-mediated nitrogen signal transduction in photosyn- ogy in rice, Arabidopsis, and maize (Supplemental Fig. 1). thetic cells at the molecular level [1,9,40]. The four OsHK proteins possess three conserved domains In rice, 10 type-A OsRRs have been reported recently [41]. (CHASE domain, transmitter domain, and receiver domain) and Transcript levels of most of the OsRR genes increased rapidly two or more transmembrane domains (Fig. 1). However, on exogenous cytokinin application even in the presence of OsHKL1 (OsCRL4) [22] possesses only the CHASE motif. All cycloheximide. The induction of OsRR genes by cytokinin even OsHKs have the conserved Thr that characterizes the CHASE in the absence of de novo protein synthesis qualifies them to be motif and mutation of Thr to Ile in Arabidopsis AHK4 will lead primary cytokinin response genes. The induction of OsRR6 in to a loss of function [15]. Each of the four cytokinin receptors response to various environmental stimuli including salinity, possesses the five conserved blocks H, N, G1, F, and G2 dehydration, and low temperature stress indicates its role in characteristic of canonical histidine kinases [45,46]. cross talk between abiotic stress and cytokinin signaling. The HP proteins are important components of the His-to-Asp identification and functional analysis of Ehd1, a type-B OsRR, phosphorelay that serve as intermediates by acquiring and trans- suggest that a novel two-component signaling cascade is ferring phosphate to the receiver domain of a response regulator. integrated into the conserved pathway in the photoperiodic Five rice genes, OsHP1–5, possess HP domains (Fig. 1B). control of flowering in rice [42]. OsHP1 and OsHP2 appear to be functional intermediates However, an integral cytokinin two-component system in because they possess the highly conserved XHQXKGSSXS rice has not been reported. In this paper we present a genome- motif, but OsHP3–5 have a Gln-for-His substitution and are, wide analysis of the rice cytokinin two-component signaling therefore, probably not functional intermediates (Supplemental system genes from Oryza sativa ssp. japonica and Oryza sativa Fig. 1B) [25]. L. Du et al. / Genomics 89 (2007) 697–707 699

Table 1 Features of the cytokinin two-component signaling system genes in rice Gene FL-cDNA TIGR locus ID Length Chromosome Genomic locus EST No. (aa) BAC name cM position Rice cytokinin receptors OsHK1 AK060291 LOC_Os02g50480 1024 2 AP005112 131–134.5 ESTI OsHK2 AK121000 LOC_Os10g21810 1117 10 AC092548 19.0–21.8 ESTJ/I OsHK3 AK067355 LOC_Os01g69920 1023 1 AP004672 160.4 ESTJ/I OsHK4 AK060709 LOC_Os03g50860 938 3 AC091532 128.3 ESTI OsHKL1/OsCRL4 AK122032 LOC_Os12g26940 948 12 AL954854 55.9 ESTI

Rice histidine phosphotransfer protein gene family OsHP1 AK072521 LOC_Os09g39400 149 9 AP005558 93.1 ESTJ/I OsHP2 AK061111 LOC_Os08g44350 148 8 AP003928 119.9 ESTJ/I OsHP3 NF LOC_Os01g54050 187 1 AP003376 127.3 ESTJ OsHP4 AK120693 LOC_Os05g44570 153 5 AC130611 107.4 ESTJ/I OsHP5 NF LOC_Os05g09410 150 5 AC097174 36.4 ESTJ/I

Rice type-A response regulator genes OsRR1 AJ938070 LOC_Os04g36070 231 4 BX548156 65.3 ESTJ OsRR2 AJ938071 LOC_Os02g35180 252 2 AP005808 81.7–83.6 ESTJ/I OsRR3 AJ938072 LOC_Os02g58350 131 2 AP004043 157.9 ESTI OsRR4 AJ938073 LOC_Os01g72330 233 1 AP003683 169.5 ESTJ/I OsRR5 AJ938074 LOC_Os04g44280 134 4 AL663010 78.2–81.7 OsRR6 AJ938075 LOC_Os04g57720 170 4 AL606652 123.8 ESTJ/I OsRR7 AJ938076 LOC_Os07g26720 207 7 AP003802 53.4 ESTJ OsRR8 AJ938077 LOC_Os08g28950 201 8 AP003962 60.1–60.4 ESTJ/I OsRR9 AJ938078 LOC_Os11g04720 188 11 AC123526 8.1–8.6 ESTJ/I OsRR10 NF LOC_Os12g04500 185 12 AL928752 12.2 ESTJ/I OsRR11 AK099560 LOC_Os02g42060 167 2 AP004768 102.6–102.9 ESTJ/I OsRR12 NF LOC_Os08g26990 135 8 AP005523 58.1–58.4 ESTJ OsRR13 NF LOC_Os04g13480 132 4 AL606634 18.3–19.6 ESTJ/I OsRR14 AK108126 LOC_Os03g53100 127 3 AC096855 136.5–137.6 OsRR15 AJ938077 LOC_Os08g28900 201 8 AP003962 60.1 ESTJ

Rice type-B response regulator genes OsRR16 AK062250 LOC_Os01g67770 555 1 AP005523 58.4 ESTJ/I OsRR17 AK100530 LOC_Os02g08500 626 2 AP006838 157.6 ESTJ/I OsRR18 AK065276 LOC_Os02g55320 688 2 AP004087 22 ESTJ/I OsRR19 AK111864 LOC_Os03g12350 691 3 AP004094 114.7–146.6 ESTJ/I OsRR20 AK073624 LOC_Os06g08440 696 6 AC107226 31 ESTI OsRR21 NF LOC_Os06g43910 694 6 AP007226 15.5 ESTJ OsRR22/Ehd1 NF LOC_Os10g32600 341 10 AC027038 44.3 ESTI/J

Rice pseudo-response regulators OsPRR1 NF LOC_Os05g32890 368 5 AP003962 60.1–60.4 ESTJ OsPRR2 NF LOC_Os04g28120 252 4 AL662963 30.8 ESTJ OsPRR3 NF LOC_Os04g28160 380 4 AL731588 30.8 OsPRR4 NF LOC_Os04g28130 620 4 AL662963 30.8–41.5 OsPRR5 NF LOC_Os05g32880 390 5 AC134932 75 ESTJ/I Systematic designation given to rice cytokinin two-component signaling system genes locus ID for each gene on rice chromosome pseudo-molecules is available at TIGR (Release 4.0). Accession numbers of the full-length cDNA sequence are available at KOME. ESTJ, japonica EST; ESTI, indica EST; NF, not found.

Analysis of the rice genome revealed the presence of 22 RRs possess a receiver domain and an extended C-terminal response regulator (OsRR) genes characterized by the presence output domain, which contains a GARP, a C-terminal transacti- of the hallmark residues D-D-K (Fig. 1C), as well as the presence vation domain and a variable extension. A Myb motif is highly of five pseudo-response regulators (OsPRR1–5, Table 1). conserved in all the type-B OsRRs (Supplemental Fig. 1C). The Analysis of the amino acid sequences and domain structures of alignments of deduced amino acid sequences of RRs and the OsRR proteins suggested that OsRR1–15 represent type-A pseudo-RR proteins in rice, Arabidopsis, and maize show that OsRRs, whereas OsRR16–22 are type-B OsRRs (Supplemental most type-A and all type-B OsRRs contain a highly conserved Fig. 1C). The type-A RRs contain a receiver domain along with a Lys. The sequence similarities between the OsPRR genes are short variable extension (less than 90 amino acids). The type-B low, but OsPRR1 and OsPRR4 have D-D-K residues. 700 L. Du et al. / Genomics 89 (2007) 697–707

Phylogenetic analysis of the two-component system genes in rice, Arabidopsis, and maize

An unrooted phylogenetic tree based on the alignments of the reading frames of the two-component system genes was constructed (Fig. 2). For statistical reliability, the bootstrap analysis was conducted with 1000 replicates. The histidine kinase family was found to fall into two groups. OsHK1 and OsHK4 form one clade, and OsHK2 and OsHK3 form another clade. OsHK1 and OsHK4 contain similar numbers of introns with conserved positions. Within each group, the rice and maize genes appear more closely related to one another than to the HKs from Arabidopsis. This suggests that the expansion of HK genes occurred after the divergence of monocot and dicot plants (Fig. 2A). The phylogenetic analysis also showed that most Arabi- Fig. 1. Primary domain structure of representative cytokinin two-component elements in rice. (A) Domain structure of cytokinin receptor in rice. TM, dopsis AHPs (with the exception of AHP1 and AHP4) form a transmembrane domain; ED, extracellular putative input domain (CHASE); KD, clade distinct from the OsHP and ZmHP clusters (Fig. 2B). The kinase domain; RD, receiver domain; H, His; D, Asp. (B) Histidine phos- ZmHPs and OsHP1 belong to the same clade, indicating that photransfer proteins. (C) Response regulators. RD, receiver domain. HPs expanded in a species-specific manner from the common ancestral genes that existed prior to the divergence of monocots and dicots. The gene structure of the two-component system genes We evaluated the relationships of response regulators between rice, maize, and Arabidopsis (Fig. 2C). The type-A Both OsHK1 and OsHK4 have 11 exons and 10 introns, RRs comprised five groups and the type-B RRs comprised three with the size of the exons being fairly well conserved groups. The OsRRs and ZmRRs are closely related, supporting (Supplemental Fig. 2A). OsHK2 has 13 exons, and OsHK3 the idea that the common ancestor of the RR genes predates the has 10 exons, whereas OsHKL1 has a single exon and no separation of monocots and dicots. The PRR clades lie outside intron. The lengths of exons 2–10 in OsHK1 and OsHK4 and the branch of the remaining RRs. The unrooted tree indicates of exons 1–8inOsHK2 and OsHK3 are invariant. Each of that the PRRs evolved from type-B RR ancestral genes. the OsHKs is disrupted by highly conserved phase introns at perfectly conserved positions in the domain (Supplemental Distribution and duplications of the two-component genes in Fig. 1A). the rice genome Both OsHP1 and OsHP2 contain five introns and the length of each exon is also well conserved. But the numbers and The positions of each gene and segmental genome duplica- lengths of exons among OsHP3, OsHP4, and OsHP5 are tions of rice were determined. There are regions on chromo- different(Supplemental Fig. 2B). All OsHP genes have the somes 1, 2, 4, 5, and 8 with a high density of OsRR genes. conserved intron phase except OsHP5, which does not have the Conversely, of chromosomes 7, 9, 11, and 12, each carries only a first intron phase (Supplemental Fig. 1B). single OsRR gene. All the OsHKs and OsHPs, and most OsRRs, Four types of intron–exon structure patterns were revealed are represented by a single gene at any given chromosomal among type-A OsRR and OsPRR genes (Supplemental Fig. location. However, the OsPRRs are found in clusters, e.g., 2C). The lengths of exons within the same type of OsRRs and OsPRR2, OsPRR3,andOsPRR5 on chromosome 4 and OsPRRs are invariant, with the exceptions of OsPRR1 and OsPRR1 and OsPRR4 on chromosome 5. OsPRR4. The genomic region encoding the receiver domain of OsHP4 was located on chromosome 5 in japonica. However, most type-A RRs contains two conserved intron phasings, in indica OsHP4 was split, with the amino-terminal 46 amino except for OsRR11, OsRR13, and OsRR14, which have different acids on chromosome 5 and the remainder on chromosome 6. intron phasings from the other type-A OsRR genes (Supple- Locations of adjacent genes were not changed. This split mental Fig. 1C). There are two types of intron–exon pattern presumably resulted from a chromosomal exchange event among the type-B OsRR genes (Supplemental Fig. 2C). All between chromosomes 5 and 6. genes encoding type-B OsRRs are disrupted by highly To determine the contribution of whole-genome duplica- conserved phase introns at perfectly conserved positions in tion and reshuffling, we compared chromosomal duplication the receiver domain. Five OsPRR genes have the same intron patterns with the organization and phylogeny of cytokinin phasing as type-B RRs. All but OsPRR5 lack the Myb motif at two-component signaling genes. As shown in Fig. 3, among the C terminus. Their predicted protein sequences are longer the 37 genes, 17 were found in nonduplicated regions of the than those of the type-A RRs, and the D-D-K amino acid genome. Of the genes found in duplicated regions, 11 OsRR residues are not conserved (Supplemental Fig. 1C). No EST and 4 OsHP genes were singular (genes not in tandem sequences were found to match the five OsPRRs. repeats) with a close relative found in the corresponding L. Du et al. / Genomics 89 (2007) 697–707 701 duplicated regions. No homolog of OsHK1, OsHK3, OsHP5, OsRR4, OsRR6,orOsRR22 was found within the duplicated region, suggesting the involvement of gene loss or more localized duplications. Two genes (OsRR5 and OsRR6) are found as repeats on the top arm of chromosome 4. This area of chromosome 4 is included in a segmental duplication with a region on the upper arm of chromosome 2 that contains the OsRR11 gene (Fig. 3). OsRR1 and OsRR2, OsRR9 and OsRR10, OsRR17 and OsRR21, OsRR18 and OsRR20, OsHP1 and OsHP2, and OsHP3 and OsHP4 are found within a segmental duplications. The clustering of the duplication pairs in the phylogenetic tree suggests an evolutionary relationship that is reinforced by the distribution of these genes in the genome. It is thus likely that the duplication of the segment harboring OsRR11 gave rise to the tandem repeats on chromosome 4 through unequal crossing over. Also, OsPRR2, OsPRR3, and OsPRR5; OsPRR1 and OsPRR4; and OsRR11, OsRR12, and OsRR15 are located in three clusters, on chromosomes 4, 5, and 8, respectively, supporting the interpretation that large-scale segmental duplications, in conjunction with tandem duplications, are partly responsible for the expansion of the gene family in rice.

Analysis of the type-A RR promoters

Cytokinin-dependent induction of the type-A RRs is partially dependent on transcriptional regulation by type-B RRs [36,47]. The rice type-A RR promoters (1.5 kb upstream of the translation start sites) were searched for the presence of various putative cis-elements known to confer regulation by cytokinin (http://www.dna.affrc.go.jp/htdocs/PLACE/). The results showed that the cytokinin-responsive type-B ARR1- binding element AGATT [34] was found in OsRR8 and OsRR14. An additional ARR1-binding element, GATCTT [34,48], was found in OsRR8 and OsRR14. The cytokinin- enhanced protein binding element TATTAG [49] was found in OsRR1-4, OsRR8, OsRR10, and OsRR12-15. The presence of these elements suggests that rice type-A RR genes may be regulated by type-B OsRRs and cytokinin.

Function of the OsHK genes

We examined whether the putative OsHKs are able to function as histidine kinases based on a yeast genetic system. In budding yeast SLN1 encodes an osmosensing histidine kinase, the deletion of which is lethal. The downstream components of SSK1 are always dephosphorylated in the yeast sln1Δ mutant, which results in overactivation of the downstream MAPK

Fig. 2. Phylogenetic tree of histidine kinases, histidine phosphotransfer proteins, and response regulators and pseudo-response regulator proteins in rice, Arabi- dopsis, and maize. The phylogenetic tree was constructed using the neighbor- joining method with bootstrap test by MEGA 3.1. The numbers at the branches are confidence values. (A) Histidine kinases. (B) Histidine phosphotransfer proteins. (C) Response regulators and pseudo-response regulator proteins. Bar, 0.1 sequence divergence. 702 L. Du et al. / Genomics 89 (2007) 697–707

Fig. 3. Locations and duplications of putative cytokinin two-component regulators in the rice chromosomes (cv. Nipponbare). Ovals on the chromosomes represent centromeres. pathway [50,51]. The yeast strain TM182 [50], which carries We fused individual putative OsHKs into a yeast expression the sln1Δ mutation, can grow on galactose- but not on glucose- vector, pGADT7, carrying the LEU2 marker and transformed it containing medium. into the yeast strain TM182. As shown in Fig. 4, TM182

Fig. 4. Cytokinin-dependent growth phenotype of yeast in which SLN1 has been replaced with OsHKs. (A) sln1Δ strains were transformed with the vector pGADT7 or pGADT7− OsHKs. Transformants grew on SD/Leu medium containing the plant hormone 6-BA and galactose (+Gal) or glucose (+Glc). (B) Transformants grown on SD/Leu medium containing DMSO and galactose (+Gal) or glucose (+Glc). 1, pGADT7; 2, pGADT7-OsHKL1; 3, pGADT7-OsHK1; 4, pGADT7-OsHK2; 5, pGADT7-OsHK3; 6, pGADT7-OsHK4. L. Du et al. / Genomics 89 (2007) 697–707 703 transformants that express one of the OsHK1, OsHK2, OsHK3, OsRR21, which showed higher transcript levels in stems and or OsHK4 genes can grow on both galactose- and glucose- spikelets. Although OsRR genes exhibited different expression containing medium if complemented with the exogenous levels in various tissues, they were not strictly tissue-specific cytokinin 6-benzylaminopurine (6-BA). However, TM182 but were expressed in most tissues. transformants that express OsHKL1 fail to grow on glucose- containing medium, indicating that OsHKL1 cannot rescue the Response of the two-component genes to cytokinin sln1Δ strain. Importantly, none of the transformants were capable of growing on glucose-containing medium without As shown in Fig. 6, exogenous application of cytokinin 6- added cytokinin 6-BA. These results indicate that OsHK1–4 BA induced the apparent accumulation of transcripts of can function as histidine kinases in yeast in a cytokinin- OsHK1–4 in roots, but not in leaves. Transcript abundance of dependent manner. The inability of OsHKL1 to display histidine most type-A ARR genes increased rapidly (within 10 min) in kinase activity in the yeast complementation assay suggests that response to cytokinin [29]. Similar to Arabidopsis,the conserved domains other than the CHASE domain are essential transcripts of type-A OsRR genes in rice accumulated rapidly for histidine kinase activity. in roots and leaves after 6-BA treatment for 4 h. In contrast, transcriptional levels of OsHKL1, OsHPs, and type-B OsRR Expression profile of the two-component genes genes in roots and leaves were not affected by 6-BA treatment.

The nonconserved cDNA regions of individual two- Discussion component genes were used as gene-specific probes. Tran- scripts of OsHK1–4 were detected in all tissues tested but Recent genetic and molecular biological studies have revealed expression levels varied. OsHK1 and OsHK4 were higher in the major components of the cytokinin signal pathway in Arabi- leaves than in other tissues, while OsHK2 and OsHK3 were dopsis. Although cytokinin receptors have been described higher in roots and leaves than in stems and spikelets. OsHKL1 previously in both Arabidopsis and maize, our analysis of rice was expressed only in roots and spikelets. OsHP1 and OsHP2 orthologs provides new insights into cytokinin signaling. were expressed in every tissue, and the other three OsHP genes The four cytokinin receptor histidine kinases (OsHK1–4) (OsHP3–5) were predominantly expressed in leaves. The contain a conserved CHASE domain at the N terminus, expression of OsHP3 and OsHP5 in other tissues could be followed by a transmitter (histidine kinase) and two receiver readily detected, while the transcript level of OsHP4 was very domains, whereas OsHKL1 contains only the CHASE domain low (Fig. 5). and a putative Ser/Thr protein kinase domain (Fig. 1), The majority of the type-A OsRR genes (OsRR2–9 and suggesting that OsHK1–4 have a histidine kinase function and OsRR11) were expressed at various levels in roots, stems, that OsHKL1 may have lost the cytokinin receptor function in leaves, and spikelets. However, OsRR3 was hardly detected in rice. Our yeast mutant experiments validated this hypothesis. stems, OsRR6 was expressed mostly in roots and leaves, and Phylogenetic analysis of cytokinin receptors in Arabidopsis, OsRR8 was expressed only in roots. Transcripts of the type-B maize, and rice showed one-to-one correspondence between OsRR genes (OsRRs16–21) were detected in all tissues, but OsHKs, ZmHKs, and AHKs (Fig. 2). Similar correspondence were expressed mainly in roots and leaves, with the exception of was not found in the downstream elements, including OsHPs,

Fig. 5. Expression analysis RT-PCR. RNA was isolated from roots, stems, leaves, and spikelets from Nipponbare rice growing in soil. RNA was also harvested from roots of plants growing in liquid nutrient medium. OsACTIN primers were used as a template control. 704 L. Du et al. / Genomics 89 (2007) 697–707

Fig. 6. The analysis of the response of the system genes family to cytokinin by RT-PCR. RNA was isolated from roots and leaves from Nipponbare rice growing in liquid nutrient medium either with 1 μM 6-benzylaminopurine (6-BA) for 4 h or with no 6-BA. OsACTIN primers were used as a template control.

ZmHPs, and AHPs and OsRRs, ZmRRs, and ARRs [9,10]. The divergence of monocot and dicot plants. Among type-A RRs, conservation of HKs implies that these cytokinin receptors the OsRR13, OsRR14, and ARR22 cluster did not show close diverged from the ancestral genes into the three isogenes prior relations with the others. The sequence alignment analyses to the divergence of monocots and dicots and that they played show that their intron phasings differ from the others. We physiologically similar roles in cytokinin signaling in all speculate that they may have lost or gained introns. The five angiosperms. OsPRR genes display the same intron phasing pattern as the Five OsHPs and 22 OsRRs were identified in the rice type-B OsRR genes but have lost the Myb-B motif. The results genome, compared to 5 AHPs and 22 ARRs in Arabidopsis. suggested that these OsPRRs were introduced via tandem The results of phylogenetic analysis indicate that most genes duplication and exon shuffling (leading to loss of Myb-B) of the from monocots (rice and maize) and dicots (Arabidopsis) are type-B OsRR genes. clustered in species-specific clades, suggesting a species- The identification of putative cis-elements in the promoters, specific expansion in plants. Especially, phylogenetic analysis in conjunction with the expression patterns, suggests that the RR of the HP subfamily revealed that the number of HP proteins genes in rice, as in Arabidopsis, might be hormonally regulated increased after the divergence of the monocot and dicot plants. and might have analogous functions. However, the fact that there The type-A response regulators in rice have been recently is species-specific expansion in rice and Arabidopsis would also reported by Jian et al. [41]. We identified five additional type-A suggest the existence of species-specific functions. OsRRs and seven type-B OsRRs, including Ehd1 [42]. Type-A RRs are partially dependent upon transcriptional Duplication analyses, mapping genes on chromosomes, and regulation by type-B RRs [36,47]. The ARR1-binding element localizations of all OsRRs show that the RR gene family arose AGATT [34] was found in the promoter of some type-A OsRR through duplication during evolution. Phylogenetic analysis genes. Thus, we speculate that the two-component system is revealed the presence of many clusters (Fig. 2). In some cases, conserved between rice and Arabidopsis at the level of system the phylogenetic relationship can be correlated with the regulation as well as gene structure. distribution of the genes on the genome. For example, the The expression patterns of the five OsHKs, five OsHPs, clusters OsRR1 and OsRR2, OsRR5 and OsRR6 and OsRR11, eight type-A OsRRs, and six type-B OsRRs revealed that the OsRR9 and OsRR10, OsRR18 and OsRR20 are linked to several rice cytokinin two-component genes were expressed in a broad duplication events, suggesting an evolutionary relationship. range of tissues. Different components had different expression However, considering differences in the genome size, a levels in roots, stems, leaves, and spikelets, and transcriptional number of HKs and RRs are detected between rice and Arabi- levels of the same gene in different tissues were different dopsis, which implies that gene loss events should be although expression patterns were overlapping. Molecular and considered. The gene intron phasing does not change much genetic analysis of Arabidopsis loss-of-function mutants and represents the conserved genes. Differences in intron indicated that the two-component elements involved in phasing for the type-A and type-B OsRR genes were observed, cytokinin signaling had redundant and overlapping functions and the same subfamily RRs contain the same intron numbers [52]. Our results showed that the rice two-component system and phasing patterns between rice and Arabidopsis. These factors might also have redundant and overlapping functions. results suggest that the type-A and type-B OsRRs were The steady-state transcripts of most OsRR genes increased conserved fairly well and evolved differently before the rapidly (within 15 min) on exogenous cytokinin application L. Du et al. / Genomics 89 (2007) 697–707 705

[41] and were also increased by exogenous cytokinin 6-BA bioedit.html). Phylogenetic analysis of the sequences was conducted using (within 4 h) in our experiment. Expression analysis indicated MEGA software (version 3.1; http://www.megasoftware.net/). The phylogenetic trees were constructed using the neighbor-joining method with bootstrap tests. that the rice type-A OsRR genes, like their counterparts in Ar- abidopsis, were primary cytokinin response factors. In contrast results reported by Jain et al. [41], in our experiments Promoter motif search transcriptional levels of OsRR3 and OsRR8 were significantly The 1.5-kb upstream genome sequences of the annotated translation start elevated by 6-BA treatment. In Arabidopsis, ARR2, ARR10, sites were retrieved from the TIGR databases for rice and Arabidopsis, and ARR11 have been detected in all adult tissues examined; respectively. Known consensus cis-regulatory sequence motifs that were however, some reports showed that the transcripts of ARR2 and detected were compared manually and using the PLACE database (http:// ARR10 were present only in the roots [13,31,32,37]. These www.dna.affrc.go.jp/htdocs/PLACE/). In addition, manual searches were performed for three regulatory motifs: (1) the sequence of the cytokinin- contradictory results may reflect a strong effect of environ- responsive ARR1-binding element AGATT, which is found in the promoter of mental factors on the expression of these genes. ARR6 [47]; (2) the ARR1-binding element GATCTT, previously determined by expression profiling [34,48]; and (3) the cytokinin-enhanced protein binding Materials and methods element TATTAG [49].

Sequence retrieval and database search Plant material and culture experiment

Multiple database searches were performed to identify members of the rice The japonica variety of rice, O. s. L. cv. Nipponbare, was used in this work. two-component system. The O. s. ssp. indica genome sequence was downloaded Uniform seeds germinated at 37°C for 2 days. The germinated seeds were grown from BGI-RIS (http://rise.genomics.org.cn/). The genome sequence of O. s. in a nutrient solution (320 μMPi) for an additional 14 days. The pH of the japonica cv. Nipponbare and the predicted protein sequences were downloaded culture solution was adjusted to 5.0 using 1 M NaOH each day. The plants were from the TIGR Rice Genome Annotation Database (version 4.0, http://www.tigr. grown in a growth chamber under a photosynthetic photon flux density of − − org/tdb.e2k1/osa1/pseudomolecules). To detect putative genes involved in the ∼200 μmol photons m 2 s 1 under a 16 h light (28°C)/8 h dark (22°C) rice cytokinin two-component signaling, sequences of 3 AHKs, 6 AHPs and 22 photoperiod. Humidity was controlled at ∼60%. For cytokinin stress, the 16- ARRs from Arabidopsis and 3 HKs, 3 HPs and 10 RRs from maize (from the day-old seedlings were transferred to a nutrient solution supplemented with NCBI databases) were used as queries. The two rice genome databases searched 1 μM 6-BA for 4 h. The tissues of the plants were harvested and frozen in liquid using TBLASTN with an E-value cutoff of 10−10 and at least 30% identity. nitrogen for RNA isolation. HMMER software (version 3.3.2; http://hmmer.wustl.edu/) with default parameters was then employed to screen all predicted proteins against the Semiquantitative RT-PCR analysis Pfam database (version 20.0), with the Pfam domain profiles including HK A (phosphoacceptor) domain (GenBank Accession No. PF00512), Hpt (GenBank Total RNA was extracted from roots, stems, and leaves of 16-day-old Accession No. PF01627), RR RD (GenBank Accession No. PF0072), CHASE seedlings and spikelets of mature seeds with Trizol reagent (Gibco BRL) domain (GenBank Accession No. PF03924), and HATPase (GenBank according to the manufacturer’s instructions. To examine the expression Accession No. PF02518). All significant hits with positive scores were classed patterns of OsHKs, OsHPs, and OsRRs genes, 37 sets of specific primers for and examined individually. The BLAST search capabilities (TBLASTN and the corresponding genes were designed for RT-PCR analysis. All primers were BLASTP) available from the NCBI (http://www.ncbi.nlm.nih.gov) and TAIR designed to span introns to eliminate potential genomic DNA contamination. (http://www.arabidopsis.org) databases and Gramene (http://www.gramene.org) The gene-specific primers are listed in Supplemental Table 1. cDNA was were also used to search for additional members of the entire rice genome. To synthesized using a first-strand cDNA synthesis kit (Invitrogen, Carlsbad, CA, increase the accuracy and extent of the database search results and to minimize USA). The rice actin gene was used as an internal control. Reactions were the exclusion of real hits caused by incorrect annotation (missing exons or performed with Taq polymerase (Takara Biotechnology, Japan) on a Peltier introns annotated as exons), we retrieved the nucleotide sequence for each of the thermal cycler (PCT-220; MJ Research, USA) with the following profile: 95 °C unique hits obtained. The database KOME (Knowledge-Based Oryza Molecular for 5 min; 28 cycles of 95°C for 30 s, 58°C for 30 s, 72°C for 30 s; and a final Biological Encyclopedia; http://cdna01.dna.affrc.go.jp/cDNA/) was used to 7-min extension at 72°C. obtain the full-length cDNAs of all predicted genes. The numbers and positions of exons and introns for individual OsHK, OsHP, and OsRR genes were Yeast experiments determined on FL-cDNA and compared with the annotations in TIGR. Genomic sequences that appeared to be misannotated by comparison with available Primers (Supplemental Table 2) were designed for full-length cDNA cDNA sequences (full-length cDNAs, ESTs) were corrected for subsequent amplification. The entire coding regions of both OsHK1-4 and OsHKL1 analysis. The EST analysis was performed using gene expression evidence cDNAs were cloned and recombined with the yeast expression vector search tools against the rice data available at TIGR. We analyzed the expression pGADT7 (Clontech K1612-1) under control of the ADH promoter, generating differences between japonica (cv. Nipponbare) and indica (cv. 9311) using the fusion vector pGADT7±OsHKs. After sequence confirmation, plasmids were ESTs in the database (http://www.Plantgdb.org). introduced into sln1Δ (strain TM182) [50]. Suspensions of the resultant transformants were lined onto drop-out medium with 10 μM exogenous plant Sequence and phylogenetic analysis hormone 6-BA as indicated in Fig. 4, with either 2% glucose or 2% galactose. To confirm the presence of conserved domains in the predicted genes, motifs in candidate genes were identified using the Pfam database (http://www.sanger. ac.uk). In addition, BLASTP was used with the InterPro (http://www.ebi.ac.uk/ Acknowledgments interpro/) databases to detect conserved domains. Each putative cytokinin two- component signaling gene was positioned on the rice chromosome pseudomo- Useful comments by Dr. Chuanzao Mao and the anonymous lecules available at TIGR by BLASTN search. Segmental genome duplications referees are gratefully acknowledged. This work was supported of rice were determined (http://www.tigr.org/tdb/e2k1/osa1/segmental_dup/). Multiple sequence alignments were performed using the ClustalW program by the National Key Basic Research and Development Plan (version 1.83; http://www.ebi.ac.uk/clustalw/). Alignments were visualized (project 973) (no. 2005CB120900), the National Natural Science using the BioEdit program (version 5.0.9; http://www.mbio.ncsu.edu/BioEdit/ Foundation of China (no. 30500106) and by the Department of 706 L. Du et al. / Genomics 89 (2007) 697–707

Science and Technology (no. 2007C22G2010039), the Depart- size, germination, root development, and cytokinin metabolism, Plant Cell – ment of Education of Zhejiang Province, China (no. 20051038). 18 (2006) 40 54. [22] Q.M. Han, H.W. Jiang, X.P. Qi, J. Yu, P. Wu, A CHASE domain containing protein kinase OsCRL4, represents a new AtCRE1-like gene family in rice, Appendix A. Supplementary data J. Zhejiang Univ. Sci. 5 (2004) 629–633. [23] T. Suzuki, K. Sakurai, C. Ueguchi, T. Mizuno, Two types of putative Supplementary data associated with this article can be found, nuclear factors that physically interact with histidine-containing phospho- in the online version, at doi:10.1016/j.ygeno.2007.02.001. transfer (Hpt) domains, signaling mediators in His-to-Asp phosphorelay, in Arabidopsis thaliana, Plant Cell Physiol. 42 (2001) 37–45. [24] G. Haberer, J.J. Kieber, Cytokinins: new insights into a classic References phytohormone, Plant Physiol. 28 (2002) 354–362. [25] I. Hwang, H.C. Chen, J. Sheen, Two-component signal transduction [1] M.C. Mok, Cytokinins and plant development: an overview, in: D.W.S. pathways in Arabidopsis, Plant Physiol. 129 (2002) 500–515. Mok, M.C. Mok (Eds.), Cytokinins: Chemistry, Activity, and Function, [26] J. Lohrmann, K. Harter, Plant two-component signaling systems and the CRC Press, Boca Raton, FL, 1994, pp. 155–166. role of response regulators, Plant Physiol. 128 (2002) 363–369. [2] D.W. Mok, M.C. Mok, Cytokinin metabolism and action, Annu. Rev. Plant [27] G.E. Schaller, D.E. Mathews, M. Gribskov, J.C. Walker, Two-component Physiol. Plant Mol. Biol. 52 (2001) 89–118. signaling elements and histidyl-aspartyl phosphorelays, in: C. Sommer- [3] B.G. Forde, Local and long-range signaling pathways regulating plant ville, E. Meyerowitz, M.D. Rockville (Eds.), The Arabidopsis Book, Am. responses to nitrate, Annu. Rev. Plant Biol. 53 (2002) 203–224. Soc. Plant Biol., American Society of Plant Biologists, Rockville, MD, [4] A.C. Martín, et al., Influence of cytokinins on the expression of phosphate 2002. starvation responsive genes in Arabidopsis, Plant J. 24 (2000) 559–567. [28] H. Sakakibara, A. Hayakawa, A. Deji, S.W. Gawronski, T. Sugiyama, His– [5] H. Sakakibara, et al., A response-regulator homolog possibly involved in Asp phosphotransfer possibly involved in the nitrogen signal transduction nitrogen signal transduction mediated by cytokinin in maize, Plant J. 14 mediated by cytokinin in maize: molecular cloning of cDNAs for two- (1998) 337–344. component regulatory factors and demonstration of phosphotransfer [6] A. Heyl, T. Schmülling, Cytokinin signal perception and transduction, activity in vitro, Plant Mol. Biol. 41 (1999) 563–573. Curr. Opin. Plant Biol. 6 (2003) 480–488. [29] I.B. D’Agostino, J. Deruère, J.J. Kieber, Characterization of the response [7] J.B. Stock, A.J. Ninfa, A.M. Stock, Protein phosphorylation and regulation of the Arabidopsis response regulator gene family to cytokinin, Plant of adaptive responses in bacteria, Microbiol. Mol. Biol. Rev. 53 (1989) Physiol. 124 (2000) 1706–1717. 450–490. [30] T. Urao, B. Yakubov, S.K. Yamaguchi, K. Shinozaki, Stress-responsive [8] C.E. Hutchison, J.J. Kieber, Cytokinin signaling in Arabidopsis, Plant Cell expression of genes for two-component response regulator-like proteins in 14 (2002) S47–S59. Arabidopsis thaliana, FEBS Lett. 427 (1998) 175–178. [9] Y. Asakura, et al., Molecular characterization of His–Asp [31] J. Lohrmann, et al., Differential expression and nuclear localization of phosphorelay signalling factors in maize leaves: implications of the signal response regulator-like proteins from Arabidopsis thaliana, Plant Biol. 1 divergence by cytokinin-inducible response regulators in the cytosol and (1999) 495–505. the nuclei, Plant Mol. Biol. 52 (2003) 331–341. [32] H. Sakai, T. Aoyama, H. Bono, A. Oka, Two-component response [10] K. Yonekura-Sakakibara, M. Kojima, T. Yamaya, H. Sakakibara, regulators from Arabidopsis thaliana contain a putative DNA-binding Molecular characterization of cytokinin-responsive histidine kinases in motif, Plant Cell Physiol. 39 (1998) 1232–1239. maize: differential ligand preferences and response to cis-zeatin, Plant [33] J.L. Riechmann, et al., Arabidopsis transcription factors: genome-wide Physiol. 134 (2004) 1654–1661. comparative analysis among eukaryotes, Science 290 (2000) 2105–2110. [11] T. Mizuno, His–Asp phosphotransfer signal transduction, J. Biochem. 123 [34] H. Sakai, T. Aoyama, A. Oka, Arabidopsis ARR1 and ARR2 response (1998) 555–563. regulators operate as transcriptional activators, Plant J. 24 (2000) 703–711. [12] T. Kiba, et al., Differential expression of genes for response regulators in [35] K. Hosoda, et al., Molecular structure of the GARP family of plant Myb- response to cytokinins and nitrate in Arabidopsis thaliana, Plant Cell related DNA binding motifs of the Arabidopsis response regulators, Plant Physiol. 40 (1999) 767–771. Cell 14 (2002) 2015–2029. [13] A. Imamura, et al., Compilation and characterization of Arabidopsis [36] I. Hwang, J. Sheen, Two-component circuitry in Arabidopsis cytokinin thaliana response regulators implicated in His–Asp phosphorelay signal signal transduction, Nature 413 (2001) 383–389. transduction, Plant Cell Physiol. 40 (1999) 733–742. [37] J. Lohrmann, et al., The response regulator ARR2: a pollen-specific [14] T. Inoue, et al., Identification of CRE1 as a cytokinin receptor from transcription factor involved in the expression of nuclear genes for Arabidopsis, Nature 409 (2001) 1060–1063. components of mitochondrial complex in Arabidopsis, Mol. Gen. [15] A.P. Mahonen, et al., A novel two-component hybrid molecule regulates Genomics 265 (2001) 2–13. vascular morphogenesis of the Arabidopsis root, Genes Dev. 14 (2000) [38] M.G. Mason, et al., Multiple type-B response regulators mediate cytokinin 2938–2943. signal transduction in Arabidopsis, Plant Cell 17 (2005) 3007–3018. [16] T. Suzuki, et al., The Arabidopsis sensor His-kinase, AHK4, can respond [39] H.J. Kim, et al., Cytokinin-mediated control of leaf longevity by AHK3 to cytokinins, Plant Cell Physiol. 42 (2001) 107–113. through phosphorylation of ARR2 in Arabidopsis, Proc. Natl. Acad. Sci. [17] H. Yamada, et al., The Arabidopsis AHK4 histidine kinase is a cytokinin USA 103 (2006) 814–819. binding receptor that transduces cytokinin signals across the membrane, [40] T. Suzuki, C. Cretin, T. Omata, T. Sugiyama, Transcription and Plant Cell Physiol. 42 (2001) 1017–1023. posttranscriptional regulation of nitrogen-responding expression of phos- [18] A.H. West, A.M. Stock, Histidine kinases and response regulator proteins phoenolpyruvate carboxylase gene in maize, Plant Physiol. 105 (1994) in two-component signaling systems, Trends Biochem. Sci. 26 (2001) 1223–1229. 369–376. [41] M. Jain, A.K. Tyagi, J.P. Khurana, Molecular characterization and [19] V. Anantharaman, L. Aravind, The CHASE domain: a predicted ligand- differential expression of cytokinin-responsive type-A response regulators binding module in plant cytokinin receptors and other eukaryotic and in rice (Oryza sativa), BMC Plant Biol. 6 (2006) 1 (available at: http:// bacterial receptors, Trends Biochem. Sci. 26 (2001) 579–582. www.biomedcentral.com/content/pdf/1471-2229-6-1.pdf). [20] C. Ueguchi, H. Koizumi, T. Suzuki, T. Mizuno, Novel family of sensor [42] K. Doi, et al., Ehd1, a B-type response regulator in rice, confers short-day histidine kinase genes in Arabidopsis thaliana, Plant Cell Physiol. 42 promotion of flowering and controls FT-like gene expression indepen- (2001) 231–235. dently of Hd1, Genes Dev. 18 (2004) 926–936. [21] M. Riefler, O. Novak, M. Strnad, T. Schmulling, Arabidopsis cytokinin [43] A. Matsushika, S. Makino, M. Kojima, T. Mizuno, Circadian waves of receptor mutants reveal functions in shoot growth, leaf senescence, seed expression of the APRR1/TOC1 family of pseudo-response regulators in L. Du et al. / Genomics 89 (2007) 697–707 707

Arabidopsis thaliana: insight into the plant circadian clock, Plant Cell [49] N. Fusada, et al., Identification of a novel cis-element exhibiting cytokinin- Physiol. 9 (2000) 1002–1012. dependent protein binding in vitro in the 5′-region of NADPH- [44] T. Mizuno, N. Nakamichi, Pseudo-response regulators (PRRs) or true protochlorophyllide oxidoreductase gene in cucumber, Plant Mol. Biol. oscillator components (TOCs), Plant Cell Physiol. 46 (2005) 677–685. 59 (2005) 631–645. [45] A.M. Stock, V.L. Robinson, P.N. Goudreau, Two-component signal [50] T. Maeda, M.S.M. Wurgler, H. Saito, A two-component system that transduction, Annu. Rev. Biochem. 69 (2000) 183–215. regulates an osmosensing MAP kinase cascade in yeast, Nature 369 (1994) [46] A.H. West, A.M. Stock, Histidine kinases and response regulator proteins in 242–245. two-component signaling systems, Trends Biochem. Sci. 26 (2001) 369–376. [51] F. Posas, et al., Yeast HOG1 MAP kinase cascade is regulated by a [47] H. Sakai, et al., ARR1, a transcription factor for genes immediately multistep phosphorelay mechanism in the SLN1-YPD1-SSK1 “two- responsive to cytokinins, Science 294 (2001) 1519–1521. component” osmosensor, Cell 86 (1996) 865–875. [48] A.M. Rashotte, S.D. Carson, J.P. To, J.J. Kieber, Expression profiling of [52] F.J. Ferreira, J.J. Kieber, Cytokinin signaling, Curr. Opin. Plant Biol. 8 cytokinin action in Arabidopsis, Plant Physiol. 132 (2003) 1998–2011. (2005) 518–525.