- Top of page
- Materials and methods
- Supporting Information
Cytokinins are plant hormones that influence diverse processes of growth and development, such as cell proliferation and differentiation, vascular morphogenesis, shoot development, chloroplast morphogenesis, leaf senescence, and axillary bud dormancy (Cline, 1991; Gan & Amasino, 1995; Inoue et al., 2001; Mok & Mok, 2001; Fukuda, 2004; Kim et al., 2006; Mahonen et al., 2006; Tanaka et al., 2006). Cytokinins have been implicated in the regulation of cell proliferation in the vascular cambium. In woody perennials like Populus, growth of this meristem produces secondary phloem (bark) and secondary xylem (wood), and results in stem girth increase. In addition to the vascular cambium, stems also contain a secondary shoot apical meristem (axillary meristem) in each node (axil of each leaf) that can generate sylleptic branches under appropriate developmental and environmental conditions. Direct cytokinin treatments to Populus buds promote the elongation of sylleptic buds to generate new branches during the same season in which they are formed without an intervening rest period (Cline et al., 1997). Cytokinins have also been implicated in below-ground meristematic activities (Mahonen et al., 2000; Inoue et al., 2001; Werner et al., 2001). Exogenous cytokinin application can inhibit both primary root elongation and lateral root formation, resulting in reduced root growth and biomass (Werner et al., 2001; Higuchi et al., 2004). Because of their importance in controlling crown architecture through the induction of sylleptic branching and their unquestionable effect on root and shoot growth, cytokinin signaling genes are likely to be key elements coordinating the production and distribution of biomass in trees.
The cytokinin signaling pathway resembles bacterial and yeast two-component signal transduction pathways in which an external signal is perceived by a sensor protein and transmitted to a response regulator by transfer of a phosphate group (Mizuno, 1998; West & Stock, 2001). Recent genetic and molecular studies in Arabidopsis have identified three key components of this signaling pathway in plants: sensor histidine kinases (HKs), histidine-containing phosphotransfer proteins (HPs) and response regulators (RRs) (Mok & Mok, 2001; Kakimoto, 2003; Ferreira & Kieber, 2005). Cytokinin responses are initiated when cytokinin binds to the HK in a conserved extracellular domain and induces autophosphorylation on a histidine residue within the cytoplasmic transmitter domain. The phosphate group is then transferred to a HP which has the ability to phosphorylate RR proteins.
Cytokinin RRs are key elements in this phosphorelay cascade because they modulate downstream signaling through transcriptional activation and regulation of protein activity. Based on their domain structure and amino acid sequence, RRs are classified as type As, type Bs and pseudos. The relative abundance of type As and Bs, 23 in Arabidopsis (Ferreira & Kieber, 2005) and 26 in rice (Ito & Kurata, 2006; Schaller et al., 2007), indicates that they have the potential to coordinate many physiological processes regulated by cytokinin. The type As have a receiver domain with conserved aspartate-aspartate-lysine (D-D-K) residues and a short C-terminus of unknown function (Sakai et al., 2000). Type As are cytokinin primary response genes whose transcripts accumulate rapidly after cytokinin treatment without the requirement for previous protein synthesis (Brandstatter & Kieber, 1998; D’Agostino et al., 2000; Taniguchi et al., 1998). Analyses of gain and loss-of-function Arabidopsis RRs have shown that type As decrease cytokinin sensitivity and negatively regulate their own transcription (Hwang & Sheen, 2001; To et al., 2004). The type Bs are characterized by the presence of a receiver domain with the conserved D-D-K residues and a large C-terminal extension. The C-terminal extension contains a Myb-like DNA-binding region, referred to as a GARP domain (Imamura et al., 1998; Sakai et al., 1998), that is common to a class of plant-specific transcription factors that includes maize GOLDEN2, Arabidopsis ARRs, and Chlamydomonas Psr1 (Riechmann et al., 2000). This region is highly variable and rich in glutamine and proline residues, a feature usually observed in transcriptional activators (Triezenberg, 1995), and contains putative nuclear localization signals (Sakai et al., 1998; Lohrmann et al., 1999) that localize the RR to the nucleus when fused to reporter genes (Lohrmann et al., 1999; Sakai et al., 2000; Imamura et al., 2001). In contrast to type As, exogenous cytokinin has not been found to alter steady-state transcript abundances of type B RRs (Imamura et al., 1998; Kiba et al., 1999). Pseudo-RRs are genes that encode proteins that resemble authentic RRs (i.e. they contain a receiver-like domain); however, they have a glutamate in place of the central aspartate in the conserved D-D-K domain that prevents phosphorylation (Stock et al., 1989; Imamura et al., 1998; Makino et al., 2000).
It is thought that the evolution of many gene families in Arabidopsis and Populus was influenced by three genome duplications. The most recent genome duplication (‘salicoid’ event) occurred in Populus between 8 and 13 Myr ago in an ancestor of the Salicaceae and affected roughly 92% of the genome (Sterck et al., 2005) and generated nearly 8000 pairs of paralogous genes (Tuskan et al., 2006). The gene content of Populus is predicted to be 45 000, almost twice the number in Arabidopsis (Tuskan et al., 2006). The difference in gene number between Populus and Arabidopsis is largely the result of gene family expansion, as the relative frequency of protein domains present in the two species is similar (Tuskan et al., 2006). Families that have undergone expansion in Populus include genes involved in wood formation, such as cellulose biosynthesis, flavonoid biosynthesis, and membrane transport (Tuskan et al., 2006). However, certain gene families involved in hormone homeostasis and signal transduction have not expanded. This is the case for families encoding cytokinin homeostasis-related enzymes such as isopentenyl transferases and cytokinin oxidases, where the number of members is similar between Populus and Arabidopsis (Tuskan et al., 2006). The recent completion of the Populus genome sequence (Tuskan et al., 2006) significantly improves our ability to understand the structure and function of gene families involved in cytokinin signal transduction. In the present study, we identified 11 type As, 11 type Bs and 11 pseudo-RRs in Populus. Using microarray analysis and semiquantitative RT-PCR, we show expression data for these genes in different organs and tissues as well as cytokinin transcript inducibility using a detached-leaf system.
- Top of page
- Materials and methods
- Supporting Information
The present study identified 33 type A, type B and pseudo-RRs in Populus. This number is approximately equivalent to the RR family size in Arabidopsis (32 genes) and rice (36 genes). The Populus genome is predicted to contain 45 000 genes (Tuskan et al., 2006), almost twice the number in Arabidopsis (c. 27 000) and approximately the same number as rice (c. 41 000). Based on the JGI gene models, Populus has more protein-coding genes than Arabidopsis, averaging 1.4–1.6 putative Populus homologs for each Arabidopsis gene (Tuskan et al., 2006), yet the similarity in protein domain frequencies in the two genomes suggests that the greater gene number in Populus is largely the result of gene family expansion and not the presence of unique Populus genes (Tuskan et al., 2006). Families that have undergone expansion in Populus include genes involved in lignocellulosic wall formation (cellulose synthases, 93 in Populus vs 78 in Arabidopsis), lignin and phenylpropanoid biosynthesis (34 in Populus vs 18 in Arabidopsis), flavonoid biosynthesis (chalcone synthases, six in Populus vs one in Arabidopsis), and disease resistance (NBS coding R genes, 399 in Populus vs c. 200 in Arabidopsis) (Tuskan et al., 2006). The Populus type As, type Bs and pseudo-RRs, however, show no expansion compared with Arabidopsis and rice. Because of the importance of cytokinin in plant growth and development, the conservation in family size observed among Populus, Arabidopsis and rice type As, type Bs and pseudo-RRs, may reflect selection against substantial changes in the stoichiometry of the components of this signaling cascade. Genes encoding regulatory molecules involved in signal transduction pathways, like transcription factors, tend to be dosage-dependent; changing the concentration of a regulator could change the concentration of its targets (Birchler et al., 2001).
Populus type As and type Bs exhibit the invariant D-D-K residues in the receiver domain originally identified in bacterial signal transducers (Parkinson & Kofoid, 1992) and are predicted to be phosphorylated at the central aspartate in a cytokinin-dependent manner. Such phosphorylation is hypothesized to modulate the DNA binding activity of the C-terminal domain in the type Bs, but the effects on type A protein function remain elusive (Miyata et al., 1998; Sakai et al., 2000). However, site-directed mutagenesis of the central aspartate to a glutamine in the Arabidopsis type B ARR2 had no effect on DNA-binding (Hwang & Sheen, 2001), suggesting that phosphorylation is not required for target DNA binding.
Searching the Populus genome revealed the presence of atypical RRs referred to here as pseudo-RRs. The proteins encoded by these genes have the central aspartate of the invariant D-D-K substituted by glutamate (D E). We speculate that Populus pseudo-RRs may be components of the plant biological clock, as pseudo-RRs in Arabidopsis and rice have been implicated in circadian controlled events such as flowering time and photomorphogenic responses (Makino et al., 2000; Murakami et al., 2004; Mizuno & Nakamichi, 2005). APRR1, an Arabidopsis pseudo-RR also referred to as TOC1 (TIMING OF CAB EXPRESSION 1), is a component of the central oscillator of the circadian clock (Somers et al., 1998; Strayer et al., 2000). Transcripts of several Arabidopsis and rice pseudo-RRs, including TOC1 and its putative ortholog in rice (OsPRR1), have been detected in leaves and exhibit circadian regulation (Makino et al., 2000; Murakami et al., 2005). High expression of several Populus pseudo-RRs in mature leaves may indicate a role for these genes in clock-regulated events such as stomatal opening and phenylpropanoid accumulation (Harmer et al., 2000).
The classification scheme for response regulators was recently expanded to include an additional subclass (type Cs; Schaller et al., 2007), the members of which have a receiver domain more similar to that of histidine kinases than to the type As, Bs or pseudo-RRs. The rice and Arabidopsis genomes each encode two type Cs; however, the Populus genome appears to have 15 (data not shown). It will be of interest to test the functional significance of this expansion in Populus.
Our analysis of the relationship of the Populus RR gene family revealed that 78% of the genes grouped in pairs. Comparing the chromosomal distribution of the Populus RR gene family revealed that 71% of the sister pairs were located in paralogous genome regions as defined by Tuskan et al. (2006). The high similarity in amino acid sequence and gene structure among these sister pairs suggest their origination from the duplication of a common ancestor. We also found that Arabidopsis and rice RRs grouped in sister pairs. All of these results are consistent with the proposed genome duplications that Populus (Sterck et al., 2005; Tuskan et al., 2006), Arabidopsis and rice have undergone (Bowers et al., 2003). Typically, an individual member of a sister pair can be lost without affecting fitness, as the pair has redundant functions immediately after the duplication event (Blanc & Wolfe, 2004). However, gene loss or retention is not random and regulatory genes involved in signal transduction, such as RRs, tend to be dosage-dependent and preferentially retained (Blanc & Wolfe, 2004). Intriguingly, the Populus RR sister pairs that arose during the salicoid duplication, the type As (four), type Bs (five) and pseudo-RRs (four), reflect a large increase in the gene family but evenly distributed across the subfamilies. In Arabidopsis, functional redundancy among RR sister pairs is positively correlated with overlapping expression profiles (Mason et al., 2004, 2005; To et al., 2004) and may explain the weak phenotypes exhibited by single loss-of-function mutants (To et al., 2004). In contrast, the RR gene family in rice appears less functionally redundant, as single RR mutants often exhibit stronger phenotypes (Hirose et al., 2007). Our finding that Populus RR sister pairs exhibit overlapping expression profiles suggests potential functional redundancy in this gene family.
High transcript abundance of Populus RRs in nodes and phloem is consistent with a role for members of this gene family in regulating meristematic activity in the vascular cambium and axillary apical meristems. Roots also appear to be sites of active cytokinin signaling in Populus, as transcripts for most of the Populus type As and Bs (17 out of 22) were detected in this organ. Similarly, expression of 13 RRs and HKs have been detected in Arabidopsis roots (Imamura et al., 1999; D’Agostino et al., 2000; Higuchi et al., 2004; Nishimura et al., 2004; To et al., 2004). Arabidopsis AHK4/CRE1/WOL, the first cytokinin receptor identified, was found to be preferentially expressed in roots and required for root xylem differentiation (Mahonen et al., 2000; Inoue et al., 2001; Higuchi et al., 2004). The cre1 mutant exhibits a reduced number of cell files within the vascular bundle because of the lack of periclinal procambial cell divisions (Mahonen et al., 2000). Because of the abundant evidence that cytokinin has negative effects on root growth in Arabidopsis and tobacco, we speculate that Populus RRs negatively regulate root growth by blocking lateral root formation, restricting the size of the meristem in root tips, and preventing protoxylem specification.
Cytokinins have also been identified as one of the signaling molecules required for the formation of floral meristems (Bernier & Perilleux, 2005). In Sinapis alba, flowering is inhibited when long-distance signaling is interrupted by phloem removal and restored upon application of cytokinin to the apex (Havelange et al., 2000). Additional documented roles of cytokinin in reproductive tissue development include maintaining cell division within the embryo, greening of sepals, and enhancing sink strength of seeds (Herbers & Sonnewald, 1998). We detected expression of 77% of the Populus type A and B RRs in prereceptive and postreceptive catkins, with type As being more abundant in prereceptive catkins. Arabidopsis cytokinin RRs and HKs have been detected in different floral tissues, including petals, stigmas, styles, immature siliques, the abscission zone of flowers, the junction of sepals and pedicels (Urao et al., 1998; Imamura et al., 1999; Mason et al., 2004, 2005; Tajima et al., 2004). Two Populus type As appear to be specifically expressed during the reproductive phase of development, as they were not detected in vegetative tissues or in any publicly available EST resources. Interestingly, these two genes grouped with the two Arabidopsis type As (ARR16 and ARR17) whose expression is also high in reproductive tissues (https://www.genevestigator.ethz.ch/). Such similarities in both protein sequence and tissue preference may indicate identical gene function in the two species.
Consistent with findings in Arabidopsis, rice and maize (Sakakibara et al., 1998; Brandstatter & Kieber, 1998; D’Agostino et al., 2000; Asakura et al., 2003; Jain et al., 2006), transcripts of seven Populus type As were induced after 1 h of cytokinin treatment. Type As are thought to be negative regulators that mediate a feedback mechanism by which the plant decreases its sensitivity to the hormone (To et al., 2004) and are predicted to inhibit type B activation by competing for phosphotransfer from upstream HP proteins (To et al., 2004). Although Arabidopsis type Bs do not appear induced after cytokinin treatment (Imamura et al., 1999; Rashotte et al., 2003; Kiba et al., 2004; Brenner et al., 2005), three Populus type Bs (PtRR13, PtRR18 and PtRR22) showed increased transcript abundances after exogenous cytokinin treatment. We speculate these results could reflect a novel regulation of particular Populus type Bs.
In the present study, we have identified the members of the Populus RR gene family. They exhibit typical features of other plant RRs, such as transcript induction in response to exogenous cytokinin, the presence of a receiver domain and a GARP domain (characteristic of the type Bs). A significant proportion of the genes in this family seem to be the product of the recent salicoid whole-genome duplication event. Most of the type As and Bs are preferentially expressed in stem tissues, while pseudo-RRs are preferentially expressed in mature leaves. Unraveling the contributions of individual cytokinin RRs in Populus will contribute to our understanding of the roles that cytokinin can play in perennial plant growth and development. There is a growing realization that combustion of fossil fuels and other human activities, including deforestation and other changes in land use, are driving an imbalance in the global carbon cycle (DeLucia et al., 2005). Forest trees such as Populus are seen as ‘clean’ alternatives to reducing anthropogenic carbon dioxide in the atmosphere because of their capacity to store large quantities of biomass in below-ground organs (http://www.science.doe.gov/grants/Fr02-23.html) and as a feedstock for renewable bioenergy. A better understanding of the hormone response pathways that govern productivity should improve opportunities for genetic enhancement of woody biomass.