The nucleotide sequences reported in this paper have been submitted to the DDBJ database under accession numbers AB429410 (PnLHY1 cDNA), AB433347 (PnLHY1 genomic DNA), AB429411 (PnLHY2 cDNA) and AB433348 (PnLHY2 genomic DNA).
• LHY/CCA1 genes play a key role in the plant circadian clock system and are highly conserved among plant species. However, the evolutionary process of the LHY/CCA1 gene family remains unclear in angiosperms. To obtain details of the phylogeny of these genes, this study characterized LHY/CCA1 genes in a model woody plant, Populus tree.
• The evolutionary process of angiosperm LHY/CCA1 genes was elucidated using three approaches: comparison of exon–intron structures, reconstruction of phylogenetic trees and examination of syntenic relationships. In addition, the molecular evolutionary rates and the expression patterns of Populus LHYs were analyzed.
• Gene duplication events of Populus LHYs and Arabidopsis LHY/CCA1 had occurred independently by different chromosomal duplication events arising in each evolutionary lineage. Populus LHYs were under purifying selection by estimating substitution rates of these genes. Further, Populus LHYs conserved diurnal expressions in leaves and stems but the transcripts of LHY2 were more abundant than those of LHY1 in Populus plants.
• This study uncovered phylogenetic relationships of the LHY/CCA1 gene family in angiosperms. In addition, the transcript abundance and the evolutionary differences between Populus LHY1 and LHY2 imply that Populus LHY2, rather than LHY1, may have a major role in the Populus clock system.
Circadian rhythms, one of the most widespread phenomena in living organisms, are generated by endogenous circadian clock systems. The circadian clock systems in plants control the timing of endogenous responses under complex circumstances such as 24 h light : dark cycles and daily temperature fluctuations. Those responses include various diurnal biological processes, such as leaf movements, stomatal conductance, photosynthetic capacity and volatile emission (reviewed in Yakir et al., 2007). Furthermore, plant circadian clock systems regulate long-term developmental processes, such as transition from vegetative to reproductive development and from growing to dormant stages, in response to long-period circannual changes in environmental factors (Böhlenius et al., 2006; Hecht et al., 2007).
In the past decade, aided by Arabidopsis genetics and systems biology, a wealth of information about plant clock systems has been accumulated (reviewed in Más, 2005; McClung, 2006). The plant clock system is proposed to be a three transcriptional-feedback loop model (loop I, II and III) in Arabidopsis thaliana (Locke et al., 2006; Ueda, 2006; Zeilinger et al., 2006). In this model system, loop I couples together the evening oscillator (loop II) and the morning oscillator (loop III). Loop I, as the center of the three loops, consists of two morning-expressed genes, Late Elongated Hypocotyl (AtLHY) and Circadian Clock Associated 1 (AtCCA1), and an evening-expressed gene, Pseudo-response regulator 1/Timing of CAB2 Expression 1 (AtPRR1/TOC1). AtLHY and AtCCA1 are paralogous genes and have a partial redundant function to generate robust circadian rhythms in various environments (Schaffer et al., 1998; Gould et al., 2006). The morning expressions of AtLHY and AtCCA1 result from a direct activation by light and an indirect activation by the partner, AtPRR1/TOC1 (Wang & Tobin, 1998; Alabadíet al., 2001; Kim et al., 2003). In addition, AtLHY/CCA1 proteins directly bind to the evening element on the promoter region of AtPRR1/TOC1, resulting in repression of its transcription during the daytime (Alabadíet al., 2001). Thus, it is clear that AtLHY and AtCCA1 play a key role in the entrainment of environmental cues and the regulation of the clock system itself in the main loop of the clock system.
Homologs of Arabidopsis LHY/CCA1 genes are conserved not only in eudicotyledonous plants, but also in monocotyledonous plants. In the former, LHY/CCA1 genes have been isolated from Phaseolus vulgaris (Kaldis et al., 2003) and Castanea sativa (Ramos et al., 2005) in addition to Arabidopsis. In monocotyledonous plants, LHY/CCA1 genes have been isolated from Oryza sativa (Izawa et al., 2002), Lemna gibba and Lemna paucicostata (Miwa et al., 2006). The proteins encoded by these genes have a conserved Myb DNA-binding domain at their N-terminus. Furthermore, these genes exhibit rhythmicity, with peak expression around dawn, which is consistent with the expression patterns of Arabidopsis LHY and CCA1 (Schaffer et al., 1998; Wang & Tobin, 1998). These observations imply that functions of LHY/CCA1 genes in the plant clock system are highly conserved among angiosperm species.
Unlike the conservation of their function, the number of LHY/CCA1 genes per genome varies in plant species. In eudicots, one copy of the LHY/CCA1 gene is found in P. vulgaris (eurosids I) and two copies exist in A. thaliana (eurosids II) (Schaffer et al., 1998; Kaldis et al., 2003). In addition, two copies of LHY/CCA1 gene are annotated as predicted genes in the available genomic sequence database of Populus trichocarpa (eurosids I; Tuskan et al., 2006; see http://genome.jgi-psf.org/Poptr1_1/Poptr1_1.home.html). In monocots, one copy of the LHY/CCA1 gene has been isolated from O. sativa and two copies have been isolated from L. gibba and L. paucicostata (Miwa et al., 2006; Murakami et al., 2007). Thus, the evolutionary process of the LHY/CCA1 gene family appears to be complicated in both eudicots and monocots.
Gene duplication can result from unequal crossing-over, retroposition, and chromosomal or whole-genome duplications (reviewed in Zhang, 2003). In the genome of A. thaliana, three polyploidy events (so-called α, β and γ) are assumed to have occurred in angiosperm evolutionary lineages (Bowers et al., 2003; Blanc & Wolfe, 2004; De Bodt et al., 2005). Although the correct timing of α and β polyploidy events had been under dispute, recent completion of the draft genome sequence of Carica papaya revealed that these polyploidy events had arisen after divergence of Arabidopsis and Carica in eurosids II (Tang et al., 2008). On the other hand, the γ polyploidy event is believed to have occurred in eudicot lineages after divergence of monocots and eudicots, although the correct timing is still under debate (Jaillon et al., 2007). In the Populus lineage of eurosids I, the Salicoid polyploidy event occurred within Salicaceae after divergence between Fabales and Malpighiales, (Tuskan et al., 2006). Since the conservation of gene orders on the duplicated chromosomes results from the chromosomal duplication events (Adams & Wendel, 2005), comparisons of the gene orders around duplicated genes provide molecular evolutionary information for understanding the phylogenetic relationships (Sampedro et al., 2005; Bocock et al., 2008).
To clarify the evolutionary relationships of angiosperm LHY/CCA1 genes, here we first isolated two full-length LHY/CCA1 genes from the genus Populus and characterized their genomic structures. We next compared the exon–intron structures of LHY/CCA1 genes, reconstructed a phylogenetic tree using sequence data of angiosperm LHY/CCA1 genes and examined syntenic relationships in the neighboring regions of LHY/CCA1 genes across plant species. Furthermore, we analyzed the molecular evolutionary rates, the diurnal expression patterns and the expression levels of these genes to verify the biological function of two LHY genes in Populus plants. This study uncovered not only the evolutionary processes of the LHY/CCA1 gene family in angiosperms but also the differential expression patterns of two LHY genes in Populus plants.
Materials and Methods
Poplar (Populus nigra var. italic) plants were grown aseptically in agar medium containing Murashige and Skoog basal salt (Murashige & Skoog, 1962), Murashige and Skoog vitamin, 20 mm MES-KOH (pH 5.8), 0.5 mg l−1 indole-3-butyric acid, 3% (w/v) sucrose and 0.8% (w/v) agar at 22°C under 16 : 8 light : dark conditions (100 µmol m−2 s−1).
Isolation of full-length cDNA
The Populus LHY/CCA1 genes were isolated from a full-length enriched cDNA library constructed from mRNA of P. nigra (Nanjo et al., 2007). The LHY/CCA1 genes were subjected to dideoxy-nucleotide sequencing using a primer walking method, and nucleotide sequences were assembled by ATSQ software (Genetyx, Tokyo, Japan).
Isolation of genomic DNA encoding LHYs
To determine the exon–intron boundaries of Populus LHYs, the full-length genomic regions were isolated from P. nigra genomic DNA using PCR. Genomic DNA was extracted by the CTAB (hexadecyltrimethyl-ammonium bromide) method from the leaves of 1-month-old poplar plants maintained on agar medium (Murray & Thompson, 1980). PCR was performed by Takara LA Taq polymerase (Takara Bio, Shiga, Japan) according to the manufacturer's instructions using the primer sets 5′-TTGGCTTTCTCTTCTCACTGCC-3′ and 5′-CCATGCAAGGCCAATTCAATAC-3′ for PnLHY1 and 5′-GATGGAGTGTGTCTAACTGGT-3′ and 5′-CCGTGGAAGGCCAATTCAATACT-3′ for PnLHY2. The PCR condition was: 94°C for 1 min; 30 cycles of 98°C for 15 s, 68°C for 10 min; and 72°C for 10 min. The amplified PCR fragments were subcloned into pGEM-T Easy vector (Promega, Madison, WI, USA) and subjected to dideoxy-nucleotide sequencing using a primer walking method. The nucleotide sequences were assembled by ATSQ software (Genetyx).
Amino acid sequences were deduced from cDNA sequences of LHY/CCA1 genes and aligned using the ClustalW program. The numbers of amino acid substitutions between each pair of LHY/CCA1 proteins were estimated by the Jones–Taylor–Thornton (JTT) model (Jones et al., 1992) with the complete-deletion option. From estimated numbers of amino acid substitutions, a phylogenetic tree was reconstructed using the neighbor-joining (NJ) method (Saitou & Nei, 1987). The bootstrap values were calculated with 1000 replications (Felsenstein, 1985). These procedures were performed using MEGA4 software (http://www.megasoftware.net/index.html) (Tamura et al., 2007). We also reconstructed a phylogenetic tree by the maximum-likelihood (ML) method using PhyML (http://atgc.lirmm.fr/phyml/) (Guindon et al., 2005) applying the JTT model for amino acid substitutions. The bootstrap values for this phylogenetic tree were calculated with 100 replications. Rates of nonsynonymous (dN) and synonymous (dS) substitutions were calculated using the modified Nei–Gojobori method with the transition/transversition ratio equal to 1.2 and the Jukes–Cantor correction. This analysis was performed using MEGA4 software (Tamura et al., 2007).
To identify the syntenic relationship of the genomic regions containing PtLHY1 and PtLHY2 in the Populus genome (http://genome.jgi-psf.org/Poptr1_1/Poptr1_1.home.html), we first performed a bidirectional TBLASTN search against the genomic sequence of Populus using predicted genes located in neighboring regions of the PtLHY1 and PtLHY2 genes as queries. If a gene(s) showing high sequence similarity to a query sequence (E-value lower than 10−5) was found near the LHY/CCA1 gene, the TBLASTN search was further performed in the reverse direction using the best-hit gene as a query. When the best-hit gene in the second TBLASTN search showed high similarity (E-value lower than 10−5) to the gene used as the query in the first TBLASTN search, we considered this gene pair as orthologs.
To investigate the expression pattern of LHY/CCA1 genes in poplar, leaves and stems of P. nigra growing on the campus of Iwate University were collected at 3 h intervals from 09:00 h on 7 July 2006 to 09:00 h on 9 July 2006. The samples were immediately frozen in liquid nitrogen and stored at −80°C until use. During sampling, the natural mean day length and temperature were 14 h 52 min and 20.3°C, respectively.
Total RNA was isolated from samples using a NucleoSpin® RNA Plant kit (Macherey-Nagel, Düren, Germany) with in-column DNase I digestion. First-strand cDNA was synthesized using ReverTra Ace-α (TOYOBO, Osaka, Japan) according to the manufacturer's instructions. Real-time PCRs were performed using a Thermal Cycler Dice Real Time System (Takara Bio) according to the manufacturer's instructions. The gene-specific primers for real-time PCR were as follows: PnLHY1 (forward, 5′-GTGAGTTTTCATGTGAGTTTCCGG-3′; reverse, 5′-CTACCAATAAGCCGTCGTCTTG-3′), PnLHY2 (forward, 5′-CTCCATTGAGCTGCCTGAAACA-3′; reverse, 5′-CGACCGCATAGACTCCAATTC-3′) and ubiquitin 11 (UBQ) (forward, 5′-GGTTGATTTTTGCTGGGAAGC-3′; reverse, 5′-GATCTTGGCCTTCACGTTGT-3′). The UBQ gene was used as a normalization control. Each RNA sample was assayed in triplicate. RNAs were assayed from two biological replicates.
Real-time PCRs were also performed to examine expression levels of PnLHY1 and PnLHY2. cDNA fragments of PnLHY1, PnLHY2 and UBQ were amplified using Takara Ex Taq polymerase by the gene-specific primer pairs shown earlier in this section. The amplified fragment of PnLHY1 was subcloned into pGEM-T Easy vector (Promega) and those of PnLHY2 and UBQ were subcloned into pTAC-1 vector (Biodynamics Laboratory, Tokyo, Japan). These fragments were subjected to dideoxy-nucleotide sequencing. The vectors containing PnLHY2 and UBQ were digested with SalI and XhoI to cut out DNA fragments containing PnLHY2 and UBQ, which were subsequently introduced into the SalI site of pGEM-T Easy vector harboring PnLHY1. The pGEM-T easy vector containing the fragments of PnLHY1, PnLHY2 and UBQ was used to generate a standard curve of real-time PCR amplification. Transcript levels of PnLHY1 and PnLHY2 were normalized with that of UBQ. Each RNA sample was assayed in triplicate. RNAs were assayed from two biological replicates.
Characterization of Populus LHY/CCA1 genes
In the genomic sequence database of P. trichocarpa, two LHY/CCA1 homologs, PtLHY1 and PtLHY2, were predicted (Table 1). With this information, we isolated cDNA clones of two LHY/CCA1 genes (PnLHY1 and PnLHY2) from a full-length enriched cDNA library of P. nigra (Nanjo et al., 2007). PnLHY1 and PnLHY2 exhibited 97.0 and 96.1% homology at nucleotide sequence level to predicted PtLHY1 and PtLHY2, respectively. Although PtLHY1 has been annotated as two partial LHY/CCA1 genes in its genomic region, our data confirmed that a single, complete LHY/CCA1 gene was encoded in this region.
Table 1. Genomic location of LHY/CCA1-like genes in genus Populus
Nucleotide homology with P. trichocarpa ortholog (%)
We determined the exon–intron boundaries of PnLHY1 and PnLHY2 on the genome of P. nigra. Both PnLHY1 and PnLHY2 contained seven introns and eight exons, with the coding sequences (CDSs) lying from the third to eighth exons (designed as coding regions I to IV; Fig. 1a). Comparison of the exon–intron structures of LHY/CCA1 genes among angiosperm species revealed that the CDS of OsCCA1 consist of six exons similar to PnLHYs (Fig. 1a). By contrast, both AtLHY and AtCCA1 have an additional intron within the predicted coding region V of PnLHYs and OsCCA1 (Fig. 1a).
Further comparisons of PnLHYs, AtLHY/AtCCA1, and OsCCA1 at the amino acid sequence level revealed that these genes have a conserved Myb DNA-binding domain at the N-terminal region (Fig. 1b). In addition, their exon–intron boundaries were conserved among PnLHYs, AtLHY/AtCCA1 and OsCCA1, except for the sites of the additional intron of AtLHY/AtCCA1. The nucleotide lengths of coding regions I, II and III, which encode the Myb DNA-binding domain, were the same in all three species (39, 112 and 62 base pairs, respectively; Fig. 1a).
Phylogenetic analysis of LHY/CCA1 gene family in angiosperms
To infer evolutionary relationships of angiosperm LHY/CCA1 genes, phylogenetic trees were constructed with the NJ and ML methods using 12 genes from monocots, O. sativa and Sorghum bicolor; core eudicots, Mesembryanthemum crystallinum; rosids, Vitis vinifera; eurosids I, Phaseolus vulgaris, Castanea sativa, P. nigra, and P. trichocarpa; and eurosids II, A. thaliana (Table 2). Among these genes, OsCCA1 and SbCCA1 were used as an outgroup of the phylogenetic tree to place a root because the divergence between monocots and eudicots has been established from various studies (Angiosperm Phylogeny Group, 2003; Soltis et al., 2005).
Table 2. LHY/CCA1-like genes used in the phylogenetic analysis
The reconstructed phylogenetic trees confirmed that the topologies of these trees obtained by the two different tree-building methods were essentially the same (Fig. 2). The phylogenetic tree revealed that Arabidopsis CCA1 and LHY were distantly related to each other and diverged earlier from remaining eudicotyledonous LHY/CCA1 genes. In addition, eudicotyledonous LHY/CCA1 genes were more closely related to Arabidopsis LHY than to Arabidopsis CCA1. On the other hand, Populus LHY1 and LHY2 were more closely related than other LHY/CCA1 genes, indicating that the gene duplication event that produced Populus LHY1 and LHY2 occurred after the divergence of Populus and the other eurosids I (P. vulgaris and C. sativa). Consequently, the topology of the phylogenetic tree implies that the duplication event of Arabidopsis LHY/CCA1 does not coincide with that of Populus LHYs.
Chromosome syntenies among the genomes of Populus, Arabidopsis and Oryza
To obtain further information on evolutionary relationships of angiosperm LHY/CCA1 genes, we next investigated chromosomal syntenies of P. trichocarpa, A. thaliana and O. sativa by examining the physical positions of the orthologous genes surrounding the LHY/CCA1 genes. In the Arabidopsis genome, LHY and CCA1 are located on chromosomes 1 and 2, respectively (Murakami et al., 2007). The neighboring genes of LHY shared a syntenic relationship with those of CCA1 (Fig. 3; Bowers et al., 2003; Blanc & Wolfe, 2004). Furthermore, the flanking region of LHY and CCA1 retained a chromosomal synteny with a partial region of chromosomes 4 and 3, respectively. However, the regions of these two chromosomes did not contain LHY/CCA1 genes. In the Populus genome, LHY1 and LHY2 of P. trichocarpa are located on chromosomes 2 and 14, respectively (Table 1). Comparison of the gene organizations around PtLHY1 and PtLHY2 revealed that the flanking region of PtLHY1 showed chromosomal synteny with that of PtLHY2 (Fig. 3). Furthermore, the physical positions of the orthologous genes surrounding Populus LHY1/LHY2 and Arabidopsis LHY/CCA1 were relatively well conserved across plant species as a result of integration of the chromosomal syntenies within Populus or Arabidopsis genomes. These results suggest that gene duplications of Arabidopsis LHY/CCA1 and Populus LHYs have derived from ancient chromosomal duplication events.
We next compared the gene organizations surrounding Populus LHY1/LHY2 and Arabidopsis LHY/CCA1 against the Oryza genomic sequence. OsCCA1 is located on chromosome 8 (Murakami et al., 2007). We found that four genes located adjacent to OsCCA1 showed homologous relationships to the genes that resided near the LHY/CCA1 genes in the Populus and Arabidopsis chromosomes (Fig. 3). However, we did not find an extensive syntenic relationship around the LHY/CCA1 genes between the Oryza and Populus/Arabidopsis genomes. In addition, the neighboring genes of OsCCA1 had no syntenic regions with other chromosomes in the Oryza genome (Yu et al., 2005; Salse et al., 2008).
Molecular evolutionary rates of Populus LHYs
To examine the selection forces of Populus LHY1/LHY2 and the other eurosids LHY/CCA1 genes, we estimated ratios of nucleotide substitution rate in nonsynonymous (dN) versus synonymous (dS) mutations between LHY/CCA1 of rosids (V. vinifera) and eurosids (eurosids I, P. nigra, C. sativa and P. vulgaris; eurosids II, A. thaliana; Table 2). The dN/dS ratios of these genes were smaller than 0.4 and similar among LHY/CCA1s of eurosids (Table 3). This result implies that not only Populus LHYs, but also other LHY/CCA1 genes in eurosids, are under purifying selection.
Table 3. Rates of nonsynonymous and synonymous substitutions among eudicot LHY/CCA1 genes
P. nigra LHY1
P. nigra LHY2
C. sativa LHY
P. vulgaris LHY
A. thaliana LHY
A. thaliana CCA1
Rates of nonsynonymous (dN) and synonymous (dS) substitutions were calculated by the modified Nei–Gojobori method with the transition/transversition ratio equal to 1.2 and Jukes–Cantor correction. dN/dS ratios are indicated in parentheses.
V. vinifera LHY
Expression patterns of Populus LHYs
To reveal the functional conservation of Populus LHYs in a clock system, we determined expression patterns of PnLHY1 and PnLHY2 under field conditions in summer using real-time PCR. PnLHY1 and PnLHY2 showed typical diurnal expressions in both leaves and stems of P. nigra (Fig. 4). The transcripts of these genes began to increase gradually at midnight and reached peaks of diurnal rhythms around dawn.
We next carried out quantitative analysis of PnLHY1 and PnLHY2 expressions around their peak expression time (at 09:00 h on 8 July 2006) to determine whether there are differences in the expression levels of these genes. Interestingly, the transcripts of PnLHY2 were at least five times more abundant than those of PnLHY1 in both leaf and stem tissues (Fig. 5). Furthermore, the expression levels of PnLHYs were five- to sevenfold higher in leaves than in stems. These findings are consistent with expression data in the poplar eFP Browser that Populus LHY2 is expressed in more abundance than LHY1 in all tissues (mature leaf, young leaf, root, dark-grown seedling, continuous light-grown seedling, female catkins, male catkins and xylem) of Populus plants (Supporting Information, Fig. S1). Thus, these results suggest that Populus LHY2 but not LHY1 is the predominant gene expressed in Populus plants.
It is well known that the LHY/CCA1 gene family plays a key role in the angiosperm circadian clock system (reviewed in Yakir et al., 2007). However, the phylogenetic relationship of the LHY/CCA1 gene family among eudicots and monocots remains to be determined. In the present study, we isolated two LHYs from the poplar tree and then elucidated the evolutionary process of the LHY/CCA1 genes in angiosperms using three approaches: comparison of exon–intron structures, conventional phylogenetic reconstruction, and examination of syntenic relationships.
In eudicots (Populus and Arabidopsis) and monocots (Oryza), the exon–intron structures of their LHY/CCA1 genes were well conserved within their CDSs (Fig. 1a). Five exon–intron boundaries were shared in all of the LHY/CCA1 genes examined (Fig. 1b). The similarity in the exon–intron organization of Populus LHYs, Arabidopsis LHY/CCA1 and Oryza CCA1 implies that there is a common ancestral gene of LHY/CCA1 in eudicots and monocots.
Phylogenetic trees of angiosperm LHY/CCA1 genes reconstructed by the NJ and ML methods exhibited a distant relationship between Arabidopsis LHY/CCA1 and other eudicotyledonous genes (Fig. 2). This evolutionary relationship in regard to the divergence of Arabidopsis LHY and CCA1 has been shown in previous studies (Boxall et al., 2005; Miwa et al., 2006). In this study, by reconstructing the phylogenetic tree, we revealed a close relationship of Populus LHY1 and LHY2. These results indicate that gene duplication of Arabidopsis LHY and CCA1 would not coincide with that of Populus LHY1 and LHY2.
To obtain further details of the phylogenetic relationships of LHY/CCA1 genes in angiosperms, we analyzed the chromosomal syntenies among three model plants, P. trichocarpa, A. thaliana and O. sativa. Chromosomal syntenies that are conserved across plant species are a powerful tool for studying the evolutionary process of a gene family (Sampedro et al., 2005; Bocock et al., 2008). In the Arabidopsis genome, the flanking regions of AtLHY and AtCCA1 retained the synteny that was derived from the β polyploidy event (Figs 3, 6; Bowers et al., 2003; described as ‘old’ duplication in Blanc et al., 2003; Blanc & Wolfe, 2004; De Bodt et al., 2005). In addition, completion of the draft genome sequence of C. papaya suggests that the β polyploidy event would have taken place after divergence of Arabidopsis and Carica within Brassicales of eurosids II (Tang et al., 2008). Therefore, we speculate that the ancestral LHY/CCA1 gene was duplicated into LHY and CCA1 in the lineage leading to Arabidopsis but not in that leading to Carica. This evolutionary footprint in Brassicales is consistent with the results of a recent study that Carica retains only one copy of LHY/CCA1 gene in its genome (Ming et al., 2008). Our results also demonstrate that other chromosomal syntenies were found in Arabidopsis between chromosomes 1 and 4 and between chromosomes 2 and 3 (Fig. 3). These syntenies are assigned to the α polyploidy event that has arisen after the β polyploidy event (Bowers et al., 2003; described as ‘recent’ duplication in Blanc et al., 2003; De Bodt et al., 2005). However, since there are no LHY/CCA1 genes within the syntenic regions of chromosomes 3 and 4, duplicated LHY and CCA1 produced by the α polyploidy event may have been lost from the ancient Arabidopsis genome during the evolutionary process (Fig. 6).
The conserved syntenic relationships within the Populus genome lead us to hypothesize that Populus LHY1 and LHY2 were duplicated in the Salicoid duplication event that is believed to have occurred after the divergence of Fabaceae and Salicaceae within eurosids I (Figs 3, 6; Sterck et al., 2005; Tuskan et al., 2006). This hypothesis is consistent with the topology of the phylogenetic trees; the gene duplication of Populus LHYs occurred in a lineage of Populus (Fig. 2). Furthermore, the syntenic relationships of the Populus chromosomes are shared with four Arabidopsis chromosomes (Fig. 3). The present study showed that the chromosome duplication events that produced the syntenic relationships in the Populus and Arabidopsis genomes had occurred independently in each lineage (Fig. 6). Thus, we propose that a common ancestral LHY/CCA1 gene of eurosids I and II had been located on the common ancestral chromosome that was subsequently duplicated into two Populus chromosomes and four Arabidopsis chromosomes.
The evolutionary process of angiosperm LHY/CCA1 genes that was deduced by the syntenic relationships differs from the topology of the phylogenetic tree with regard to the timing of the gene duplication of Arabidopsis LHY/CCA1 (Figs 2, 6). Although the syntenic relationships indicated that the ancestral LHY/CCA1 gene was duplicated into LHY and CCA1 after divergence of Arabidopsis and Carica in eurosids II, the phylogenetic trees implied that the gene duplication had occurred before the divergence of eurosids I and II. Several studies have recently shown that the substitution rate among paralogous genes was accelerated in the Arabidopsis genome compared with the Populus genome, which could affect reconstruction of the phylogenetic tree (Van de Peer et al., 1996; Tuskan et al., 2006). Estimation of the synonymous substitution rates in the LHY/CCA1 genes in rosids (V. vinifera) and eurosids (eurosids I, P. nigra, C. sativa and P. vulgaris; eurosids II, A. thaliana) indicated that Arabidopsis LHY and CCA1 had a higher synonymous substitution rate than that of LHY/CCA1 genes in eurosids I (dS values in Table 3). Thus, the difference in synonymous substitution rates would affect the topology of the phylogenetic tree, resulting in inconsistency in the timing of duplication events of Arabidopsis LHY and CCA1 genes estimated by phylogenetic tree and syntenic relationships.
Unfortunately, the evolutionary process of the LHY/CCA1 gene family in monocots still remains unclear. The chromosome synteny analyses revealed a reduced degree of conserved synteny in the flanking regions of the LHY/CCA1 genes between the Oryza and eurosids genomes (Fig. 3). A reduced degree of conserved synteny between the Arabidopsis and the Oryza genomes has been shown previously in genome-wide surveys (Salse et al., 2002; Vandepoele et al., 2002). Two Lemna plants have two LHY/CCA1 genes (LHYH1 and LHYH2) that show typical morning expressions (Miwa et al., 2006). Our analysis indicated that the gene duplication event of Lemna LHYHs did not coincide with that of Arabidopsis LHY/CCA1 and Populus LHYs (data not shown). This is because Arabidopsis LHY/CCA1 and Populus LHYs were derived from the β polyploidy and Salicoid polyploidy events, respectively, and both polyploidy events occurred only in eudicot lineages (Bowers et al., 2003; Blanc & Wolfe, 2004; Tuskan et al., 2006). Furthermore, phylogenetic analysis suggests that the gene duplication of Lemna LHYHs occurred in a common ancestor of Oryza and Lemna (Miwa et al., 2006; our unpublished data).
Alterations of a promoter region can contribute to differential expression of duplicated genes (reviewed in Zhang, 2003). In the promoter regions of PnLHY1 and PnLHY2, a difference in the composition of cis-regulatory elements was found. This is presumably the result of nucleotide substitutions, insertions and deletions (Fig. S2). It has been reported that the duplicated genes derived from the Salicoid polyploidy event showed differential expression patterns in Populus plants depending on the evolutionary changes of their promoter regions (Ohmiya et al., 2003; Y. Ohmiya, pers. comm.). Thus, it is possible that differences in the promoter regions of Populus LHYs may affect the transcript levels between Populus LHY1 and LHY2.
In addition, diversity of duplicated genes can result from alterations of a protein coding region (reviewed in Zhang, 2003). Both of the Populus LHY1 and LHY2 genes are assumed to be functional in Populus plants because these genes are under purifying selection (Table 3). However, our analysis of specific phosphorylation sites that would be required for the physical function of LHY/CCA1 protein in angiosperms (Daniel et al., 2004) revealed that the serine residue located upstream of the Myb DNA-binding domain is highly conserved among angiosperm LHY/CCA1 proteins, except for the Populus LHY1 proteins (Fig. S3). Wang et al. (2005) have proposed that, in some genes, the loss of a phosphorylation site in a protein would contribute to acquisition of a modified function of the protein during gene and species evolution. Thus, the mutation of the phosphorylation site in the Populus LHY1 protein may result in the divergence of these duplicated genes in poplar tree.
Collectively, alterations of protein coding regions and promoter regions of Populus LHY1 and Populus LHY2 indicate that these genes have been subjected to a different evolutionary fate after gene duplication. The conservation of phosphorylation sites and higher expression of Populus LHY2 may suggest a major role of this gene in the Populus clock system. However, it is not clear at the moment whether Populus LHY1 plays a role that is redundant in relation to, or different from, Populus LHY2 in Populus tree.
In summary, we have demonstrated that Populus has two LHYs produced by the Salicoid polyploidy event and that the two genes share a common ancestor with Arabidopsis LHY and CCA1. The Salicoid polyploidy event affected nearly 92% of the Populus genome, and nearly 8000 pairs of the Salicoid duplicated genes were identified out of 45 555 genes predicted in the present Populus genome (Tuskan et al., 2006). They also revealed that the Salicoid duplicated genes were under purifying selection, which is similar to our results for LHY1 and LHY2 (Table 3). On the other hand, they found, using whole-genome microarray analyses, that 5% of duplicated genes from the Salicoid polyploidy event (nearly 400 pairs of genes) showed differential expression patterns in Populus plants (Tuskan et al., 2006). The present study also elucidated that the transcripts of LHY2 were more abundant than those of LHY1 in Populus plants, but, interestingly, both LHYs conserved typical diurnal expressions in leaf and stem tissues (Figs 4, 5). Further studies are clearly needed to understand functional differences or redundancies between LHY1 and LHY2 in Populus plants.
We thank Professor Malcolm M. Campbell (University of Toronto) for helpful advice on poplar microarray data, and Dr Abidur Rahman (Iwate University) for critical reading of the manuscript. This work was supported in part by a Grant-in-Aid for the 21st Century COE Program from the Ministry of Education, Culture, Sports, Science and Technology of Japan (K-3 to MU) and by a grant from the Japan Society for the Promotion of Science (19.9498 to NT).