Identification of genes related to the development of bamboo rhizome bud.

Bamboo (Phyllostachys praecox) is one of the largest members of the grass family Poaceae, and is one of the most economically important crops in Asia. However, complete knowledge of bamboo development and its molecular mechanisms is still lacking. In the present study, the differences in anatomical structure among rhizome buds, rhizome shoots, and bamboo shoots were compared, and several genes related to the development of the bamboo rhizome bud were identified. The rice cross-species microarray hybridization showed a total of 318 up-regulated and 339 down-regulated genes, including those involved in regulation and signalling, metabolism, and stress, and also cell wall-related genes, in the bamboo rhizome buds versus the leaves. By referring to the functional dissection of the homologous genes from Arabidopsis and rice, the putative functions of the 52 up-regulated genes in the bamboo rhizome bud were described. Six genes related to the development of the bamboo rhizome bud were further cloned and sequenced. These show 66-90% nucleotide identity and 68-98% amino acid identity with the homologous rice genes. The expression patterns of these genes revealed significant differences in rhizome shoots, rhizome buds, bamboo shoots, leaves, and young florets. Furthermore, in situ hybridization showed that the PpRLK1 gene is expressed in the procambium and is closely related to meristem development of bamboo shoots. The PpHB1 gene is expressed at the tips of bamboo shoots and procambium, and is closely related to rhizome bud formation and procambial development. To our knowledge, this is the first report that uses rice cross-species hybridization to identify genes related to bamboo rhizome bud development, and thereby contributes to the further understanding of the molecular mechanism involved in bamboo rhizome bud development.

Introduction

Bamboo (Bambusoideae) is one of the largest members of the grass family Poaceae, which includes >1500 species which are vital to the economy of many tropical and subtropical countries of the world. Most cultured bamboos are perennial woody evergreens and are basically reproduced by the rhizome. According to the type of the rhizome, bamboos have been divided into three groups: scattered bamboos with a monopodial rhizome, caespitose bamboos with a sympodial rhizome, and pluricaespitose bamboos with a monopodial and sympodial rhizome. The rhizome bud can either develop into a bamboo shoot which will grow into a bamboo culm in a very short period, or develop into a new rhizome which will enable the sustainable production of the bamboo grove. Therefore, the development of bamboo rhizome is fairly complex and distinct from that of ordinary grasses.

Several efforts have been made to reveal the anatomical and physiological mechanisms of rhizome bud development in bamboo. The morphological and structural analyses show that the developmental course of the bamboo shoot is divided into six stages: dormancy, germination, development stage I, II, and III, and shoot stage (Zhang ). Hu observed high concentrations of gibberellin 3 (GA3), zeatin (ZT), and indole acetic acid (IAA) in the rhizome bud before bamboo shoot formation. Furthermore, precise detection by enzyme-linked immunosorbent assay (ELISA) showed that a high concentration of IAA correlated with the outgrowth of the rhizome bud before the formation of the new rhizome and bamboo shoot, while a high concentration of ZT corresponded only to the formation of the bamboo shoot from the rhizome bud (Huang ). However, the molecular bases for the development of the bamboo rhizome bud are largely unknown.

DNA microarray technology has facilitated the identification of a large number of genes involved in a particular biological process. Due to the scarce genome resources available for bamboo, it is expensive and time-consuming to develop a microarray in bamboo. Cross-species microarray hybridization is a potentially useful technique for applying model organism genomics in related species for which few functional genomics resources are available (Chismar ; Becher ; Lee ; Nieto-Díaz ). Genome sequencing and functional analysis of rice which belongs to the Poaceae have been completed. Zhu (2001) has used an oligonucleotide microarray designed for the rice genome (Oryza L. ssp. japonica) to analyse the barley (Hordeum vulgare L.) transcriptome, and the results have been confirmed by RNA gel blot analysis. Based on these promising results, in the present study the anatomical structures of the rhizome bud, rhizome shoot (early form of the rhizome), and bamboo shoot (early form of the bamboo culm), are compared and then the gene expression of the rhizome bud versus the leaf in Phyllostachys praecox is examined by using a cross-species microarray with 7500 rice unigenes. A total of 318 up-regulated and 339 down-regulated genes were observed. Among them, 52 up-regulated genes with putative functions in the rhizome bud were identified and described. Furthermore, six genes related to bamboo rhizome bud development were cloned and sequenced, and their expression patterns and potential functions were analysed.

Materials and methods

Sampling

Rhizome buds, rhizome shoots, and bamboo shoots (development stage I and shoot stage) of P. praecox (Chu and Chao), which is a typical scattered bamboo with high economic value in the east of China, were collected for anatomical comparison during early spring. Rhizome buds, bamboo shoots, rhizome shoots, leaves, and young florets were collected and prepared for gene expression analysis during early summer in the Bamboo Botanical Garden of Zhejiang Forestry Academy, PR China. All the samplings were repeated three times from different bamboos.

Histological analysis and in situ hybridization

The apical parts of the rhizome bud, rhizome shoot, and bamboo shoot were cut (∼3–5 mm) and fixed overnight in 4% paraformaldehyde in a phosphate buffer, pH 7.0, at 4 °C. The fixed tissues were dehydrated five times in a graded ethanol series, replaced twice with xylene, and embedded in paraffin. The samples were sectioned at 5 μm on a rotary microtome (Leica RM2135). Sections were stained with Ehrlich's haematoxylin and observed under a Nikon4500 digital camera.

The in situ hybridization was carried out using the method previously described by Braissant and Wahli (1998). Two templates for riboprobe syntheses were constructed by cloning the coding DNA sequence (CDS) of PpRLK1, PpTB1, and PpHB1 into pBluescript (Invitrogen). The antisense and sense RNA probes were generated by T3 and T7 RNA polymerase separately after the linearization of the plasmid. Each experiment was repeated at least three times using independent samples.

RNA isolation

Tissues were ground in liquid nitrogen and the RNA was extracted twice with TRIZOL Reagent (BBI) according to the manufacturer's instructions, and then treated with proportional DNase I at 37 °C for 30 min. The quality of the RNA was measured by both electrophoresis and optical absorbance. Only RNA samples with an A260/A280 >2.0 were used for RT-PCR. The extracted RNAs were stored in liquid nitrogen.

Detection of gene expression by using the rice gene chip

The rice gene chip (BioStar Genechip Inc., Shanghai, China) contained 10 369 spots comprising 500 bp sequences from a unigene set of 7500 clones assembled by a mixed rice cDNA library (not published). The extracted RNAs were purified with RNeasy columns according to the manufacturer's instructions. First-strand cDNAs labelled with Cy5 or Cy3 were synthesized from ∼100 mg of total RNA by reverse transcription, with an oligo(dT)24 primer, which were then further purified and fragmented by using a column.

Two samples were prepared from the leaves, designated as A1 and A2, and one from the rhizome lateral buds, designated as B1. cDNA of either A2 or B1 was labelled with Cy5, while that of A1 was labelled with Cy3. The Cy5-labelled cDNA from B1 was hybridized against the Cy3-labelled cDNA from A1. Hybridization of A2 against A1 was considered as a control parameter to check the up- or down-regulated genes in the hybridization of B1 against A1. According to the results obtained from the three replicated experiments, the genes with a Cy5/Cy3 ratio of >2.0 or <0.5 have been defined as up- and down-regulated, respectively. For statistical analysis, a quality filter was applied to all the spots as follows: (i) spots with a high background, uneven morphology, or dust specks were manually flagged as ‘bad’ and removed; and (ii) the foreground intensity must be higher than twice the background standard deviation in both the channels or four times the background standard deviation in one channel. To balance the effects of different RNA treatment and dye bias, a normalization parameter was estimated by EXP(R). Hereby, R was the average of ln(Cy5/Cy3) for effective spots. Data were normalized as the ratio Cy5 (B1)/Cy3 (A1). Arithmetic means of expression values were calculated from the three replicated experiments. The software Origin 6.1 was used to analyse the data from the gene chip.

Gene cloning and sequencing

3′ RACE (rapid amplification of cDNA ends) was carried out with 1 μg of the total RNA in a 10 μl reaction using a BD SMART™ RACE cDNA Amplification Kit according to the manufacturer's instructions. A 1 μl aliquot of the diluted cDNA mix was used as a template in the 25 μl PCR of the BD Advantage™ 2 PCR Kit. The amplification was performed for 25 cycles at 94 °C for 30 s, 68 °C for 30 s, and 72 °C for 3 min. The specific primers (Table 1) of the target genes were designed using the Omega2.0 program based on the conserved regions of rice homologous genes detected by the gene chip. The PCR products were isolated by gel electrophoresis, purified by a NucleoTrap® Gel Extraction Kit, and cloned by T-easy™. At least three positive clones, tested by PCR, were selected and sequenced using the DNA Analyzer. 5′ RACE was conducted using the same kit as above, and specific primers were designed according to the known sequences (Table 1). The sequences were assembled by ContigExpress. Sequence alignments were conducted using AlignX of Vector NTI suite 6.0 with the Dayhoff amino acids distance matrix. The phylogenetic tree of homologous genes was constructed using the Neighbor–Joining algorithm of MEGA3.1, with the multiple alignment parameters gap opening penalty 8, gap extension penalty 2, and PAM protein weight matrix (Kumar ).

Table 1.

The main primers used in this study

Gene (accession number)		Primers
PpRLK1 (DQ013804)	Conserved region	Rlkf1: 5′-GGCTACTTCAACAGCTACWCCGGTGG-3′
		Rlkr1: 5′-CCAGCTCCACCTTTGCCAATGATGT-3′
	Probe	Rlk1-f1: 5′-CAGATTTCGGGCTTGCCAAG-3′
		Rlk1-r1: 5′-AAGGAATGGTAGTAATTTAGCTTGAC-3′
PpHB1 (DQ013803)	Conserved region	Hbf1: 5′-CAGAGTTCCTCTCCAAGGCTACAGG-3′
		Hbr1: 5′-CAAGTGGCATAATGATCTGGCTCCC-3′
	Probe	Hb1-f1: 5′- GCAATACGTCCGTAGCGTTG-3′
		Hb1-r1: 5′-ACATGGATTGCGTCGGATGG-3′
PpSPY (DQ013806)	Conserved region	Spyf1: 5′-TTGTKTTGACCGAYCTTGGAACTAGC-3′
		Spyr1: 5′-CCTAACTTGTTATTTGCCGTATGACCAG-3′
	Probe	Spy-f1: 5′-GAACTACCACGACATGTGAATCTC-3′
		Spy-r1: 5′-CCCGAATCAATGGAGCATGC-3′
PpSINA (DQ013805)	Conserved region	Sinf1: 5′-CCCYGTCTGCACCAAYTCMATGTACC-3′
		Sinr1: 5′-CTGGGGTCTGCTGCTCYTTCCAGAT-3′
	Probe	Sina-f1: 5′-TCCCGTACTACAGCAAACTCAAG-3′
		Sina-r1: 5′ TCCTTCTGCCTCGTTATTCAGT-3′
PpARF1 (DQ013802)	Conserved region	Arff1: 5′-TACTTCCCKCAGGGSCACATSGAGC-3′
		Arfr1: 5′-TTTGCCAGGGAGAAACTCTTTCAGG-3′
PpHK1 (DQ013801)	Conserved region	Hkf1: 5′-CTCGGTGGAGCBTTTGATGTGGAGTC-3′
		Hkr1: 5′-TGAAACTTATYTGMCCRCCCATAAGTTC-3′
PpACT1	Probe	Act-f1: 5′-TTCATTGGTATGGAAGCTGCTG-3′
		Act-r1: 5′-GTAGCTTACATGGCAAGGACTTG-3′

Semi-quantitative RT-PCR analysis

A two-step semi-quantitative RT-PCR method (Meadus, 2003) with some modifications was used to measure the gene expression in the rhizome shoot, rhizome bud, bamboo shoot, leaf, and young floret. The first-strand cDNA was synthesized from 5 mg of RNase-free DNase-treated total RNA using oligo(dT)15 and BD PowerScript™ Reverse Transcriptase (Clontech) as described by the manufacturer. An aliquot of 1/20 of the reverse-transcribed product was used as a template in the PCR amplification. The PCRs were conducted in a total volume of 25 μl containing 1× PCR buffer, 1.5 mmol l−1 MgCl2, 200 μmol l−1 dNTPs, 1.5 U of Taq polymerase, and 40 μmol l−1 gene-specific primers under the following conditions: 94 °C for 5 min, followed by 25–28 cycles each of 94 °C for 20 s, 58 °C for 30 s, 72 °C for 1 min. The β-actin homologous gene of P. praecox, named PpACT1, was cloned and used as an internal reference to determine the relative gene expression. The primers of PpRLK1, PpHB1, PpSPY, and PpSINA genes were designed for a semi-quantitative RT-PCR (Table 1). To ensure no false-positive PCR was generated from the genomic DNA, no transcriptional RNA template was chosen as the control. The PCR products were electrophoresed on 1.5% agarose gels. After electrophoresis, the gels were stained with 10 μg ml−1 ethidium bromide and photographed under 280 nm UV light. The relative gene expression was assessed by densitometry based on the staining intensities of the corresponding bands using the ImageMaster VDS software program, version 2.0. Each experiment was repeated at least three times.

Results

The anatomical comparison of the rhizome bud, bamboo shoot, and rhizome shoot during development of P. praecox

Rhizome buds of P. praecox usually develop into bamboo culms in spring and form new rhizomes in summer. Rhizome shoots and bamboo shoots are quite different in their anatomical structure, although they both originate from rhizome buds (Fig. 1). Regular meristems were observed on the tip of the rhizome buds, and many isolated procambia were found at the front of the rhizome shoots, corresponding to their ability for unlimited growth (Fig. 1A–C). However, the meristems on the tip of bamboo shoots were irregular, and no procambia were observed at the front. Interestingly, many lateral buds appeared on the sides of the bamboo shoots, implying that the bamboo shoot was a micro-miniature of the bamboo (Fig. 1D–F). Simultaneously, a few lateral smaller buds were observed on the sides of the rhizome shoots (Fig. 1G–I). These results suggest that the rhizome bud holds a key position for the rhizome shoot and/or bamboo shoot in differentiation and development.

Fig. 1.

Anatomical comparison of the rhizome bud, bamboo shoot, and rhizome shoot during the development of Phyllostachys praecox. (A–C) Rhizome bud and its longitudinal section. (D–F) Bamboo shoot and its longitudinal section. (G–I) Rhizome shoot and its longitudinal section. Bar=100 μm.

Identification of genes related to the development of bamboo rhizome bud versus leaf by rice cross-species microarray analysis

To investigate gene expression in the bamboo rhizome buds versus leaves, rice cross-species microarray analyses were carried out. Cy5-labelled cDNA from rhizome buds was hybridized against Cy3-labelled cDNA from the leaves on the same rice gene chip with 10 369 spots. The distribution of the average normalized hybridization intensity of the three measurements showed that a total of 8795 spots were effective (Fig. 2A). The average hybridization efficiency is 84.82%. According to the criterion of the Cy5/Cy3 ratio being >2.0 or <0.5 for genes to be up- or down-regulated, respectively, a total of 318 up- and 339 down-regulated genes were identified. The reproducibility of the replicated hybridizations is 95%. By comparing the homologous genes from Arabidopsis and rice, the putative functions of these 318 up-regulated genes in the rhizome bud were classified into seven groups (Fig. 2B). Furthermore, except for the housekeeping genes, 52 of the 318 up-regulated genes with known functions in the rhizome bud are listed and described in Table 3.

Fig. 2.

Rice cross-species microarray analysis of the bamboo rhizome bud versus leaves. (A) The distribution of the average normalized hybridization intensity (Cy5, rhizome bud; Cy3, leaf) of three measurements. Grey points indicate data with a ratio >0.5 and <2, while black points indicate data that are out of the range (>2.0 for up-regulated genes and <0.5 for down-regulated genes). (B) The putative functions of 318 up-regulated genes (U, unknown genes; N, not found in Arabidopsis; H, housekeeping genes; R, regulation and signalling-related genes; M, metabolism-related genes; S, stress-related genes; C, cell wall-related genes). The number of genes is indicated.

Table 3.

Up-regulated genes excluding for housekeeping ones with possible known functions in rhizome bud

Gene_ID	Accession	Annotation	Average ratio	Function classification
R0067B07	ATY16046	Leucine-rich repeat protein (G8B7T7)	5.5	Regulation and signal transduction
R0354G06	AF193835	CLV1 receptor kinase (CLAVATA1); LRK1 protein	4.5
R0186A08	AU068629	Adaptor protein SPIKE1 (SPK1)	4.3
R0181H01	AY063945.1	Putative receptor kinase	3.9
R0201B09	AU068950	Ethylene-insensitive3 (EIN3); ethylene-insensitive-3-like protein (OsEIL1)	3.6
R0210A11	AB046871	AHK4 mRNA for histidine kinase	3.5
R0159B02	AP002819	Calcium-dependent protein kinase (CDPK6)	3.5
R0156D04	AU082872	Auxin-induced protein IAA9	3.4
R0213E09	AF141942	MRP-like ABC transporter; phytochrome C (phyC)	3.2
R0590A07	AF042196	ARF6 (ARF6)	3.1
R0008A01	AB061407	Phosphatidic acid phosphatase	3.0
R0383G08	AU056864	IAA8	3.0
R0030C04	AJ243828	Serine/threonine-specific protein kinase-like; protein phosphatase 2A regulatory A subunit (RPA1 gene)	2.9
R0200A06	AY072205.1	Putative glucose-regulated repressor protein	2.9
R0125A05	AY059850.1	Putative casein kinase I	2.8
R0440D02	ATU62135	Gibberellin signal transduction protein (SPINDLY)	2.8
R0151H01	AP002860	Putative transcriptional regulator protein	2.7
R0605C09	AY045777.1	Putative receptor serine/threonine protein kinase ARK2	2.6
R0464B02	AC024594	REVOLUTA	2.6
R0088A08	Y07748	Protein kinase (TMK1) gene, complete cds; leucine-rich repeat receptor-like kinase	2.6
R0014B03	C22395	CBL-interacting protein kinase 18 (CIPK18); protein kinase	2.5
R0099B10	AY062539.1	Calcium-dependent protein kinase	2.4
R0022G09	AC079936	Putative leucine-rich repeat transmembrane protein kinase	2.3
R0101H03	AF185577	Chromatin remodelling factor CHD3 (PKL)	2.3
R0389E08	AY059785.1	Putative kinase	2.2
R0548E04	AP002481	Ring finger E3 ligase SINAT5 (SINAT5)	2.0
R0081G03	BE607353	Putative β-glucosidase protein; β-glucosidase	8.0	Metabolism
R0272A08	L03366	Putative sucrose synthase; sucrose synthase 3 (RSs3)	6.3
R0338B11	AA750532	Starch-branching enzyme class II (sbe2-1); starch-branching enzyme I (Rbe1)	5.8
R0162H08	AY072210.1	Putative trehalose 6-phosphate synthase	4.2
R0082F06	U95816	HMG-CoA reductase (HMG1); 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR)	4.0
R0184G05	AY056285.1	Putative β-galactosidase	3.5
R0025C09	AY074343.1	Putative N-carbamyl-L-amino acid amidohydrolase	3.4
R0185F03	AY059821.1	α-Glucosidase-like protein	3.3
R0042H09	ATU79160	HMG-CoA synthase (MVA1)	3.3
R0419G09	AY064010.1	Putative trehalose-6-phosphate synthase	3.2
R0038E06	AY039932.1	Putative aldose 1-epimerase	3.0
R0215G08	AY051033.1	Putative post-transcriptional gene silencing protein	3.9	Stress
R0180E10	AC084218	Cultivar Cvi-0 RPM1 variant gene	3.3
R0193H10	AF490590.1	Na+/H+ exchanger 6 (NHX6)	2.8
R0170D08	AF322255	Magnesium transporter protein (GMN10)	2.8
R0245D03	AF213400	SLT1 protein	2.7
R0206G12	AU182068	Heat shock protein 70; LMW heat shock protein	2.5
R0026H01	ATU60445	14-3-3 protein GF14 nu (GRF7)	2.3
R0003B12	AJ276594	Putative 14-3-3 protein; 14-3-3-like protein (gf14-e gene)	2.3
R0018D10	AF272747	WRKY transcription factor 9 (WRKY9)	2.2
R0509B06	AF012660	Putative potassium transporter AtKT5p (AtKT5)	2.2
R0177C01	AY044049.1	Phytochelatin synthase (PCS2)	2.1
R0245B06	AF093674	Xyloglucan endotransglycosylase (XTR9)	13.6	Cell wall
R0160E08	AY063117.1	Putative β-1,3-glucanase (MRP15.22)	2.8
R0095C08	AF237733	Callose synthase 1 catalytic subunit (CalS1)	2.8
R0098A10	AU183649	At2g21660/F2G1.7; glycine-rich protein (OSGRP1)	2.2

The rice genes on the chip were mainly described by the homologous genes from Arabidopsis since most sequences of rice had not been published or studied yet. Accession numbers and descriptions of the corresponding rice genes are indicated in bold.

Twenty-six genes related to signalling and biosynthesis of auxin are highly expressed in the rhizome buds. Their average Cy5/Cy3 ratios ranged from 2.0 to 5.5. Surprisingly, it was found that ethylene signalling-related genes were also highly expressed in the rhizome buds. In addition, some receptor-like kinase genes (such as leucine-rich repeat receptor, CLV1 receptor kinase, and adaptor protein SPIKE1) and some meristem initiation-related transcriptional factors (such as REVOLUTA) were detected with high expression levels in the rhizome buds.

A total of 11 metabolism-related genes with comparatively high expression levels were detected in the bamboo rhizome buds in comparison with the leaves. Their average Cy5/Cy3 ratios ranged from 3.0 to 8.0. In addition, the expression of starch accumulation-related genes, such as starch-branching enzyme class II (sbe2-1) and starch-branching enzyme I (Rbe1), was notably high (average Cy5/Cy3 ratios were 5.5) in the rhizome buds.

Eleven stress response-related genes were up-regulated in the rhizome buds, most of which were involved in abiotic stress (e.g. ion homeostasis and temperature response). Expression of stress-related genes in the rhizome buds seemed to be parallel to the harsh growing conditions in early spring.

Four genes related to the synthesis and repair of the cell wall were also up-regulated in the rhizome buds. For example, the xyloglucan endotransglycosylase/hydrolase (XTH) gene was highly expressed, and the average Cy5/Cy3 ratio was 13.

Cloning and sequencing of six genes from the bamboo rhizome bud

In order to characterize selected genes related to the development of bamboo rhizome buds, genes for cloning were chosen based on three criteria: (i) cDNAs with a high signal intensity and up-regulation ratio >3.0 in the array analysis; (ii) homologous rice sequences of these 500 bp cDNAs known to be involved in meristem development; and (iii) that degenerate primers corresponding to the conserved region of homologous genes were designed. Six up-regulated genes, homologous to REVOLUTA, CLAVATA1, SPINDLY, SINAT5, ARF1, and AHK4 in rice, were cloned and sequenced from the rhizome buds, and were named as PpHB1, PpRLK1, PpSPY, PpSINA, PpARF1, and PpHK1, respectively (Table 2). The nucleotides and deduced amino acids of the six genes were compared with their homologous genes in rice, sharing 66–90% identity at the nucleotide level, and 68–98% identity at the amino acid level. Among them, PpSINA and PpHB1 displayed the highest degree of identity (97.8% and 93.2%, respectively), PpSPY and PpARF1 had a medium level of identity (87.9% and 84%, respectively), and PpRLK1 and PpHK1 shared a relatively low level of identity (74.6% and 67.7%, respectively).

Table 2.

Six genes cloned in bamboo according to gene chip detection

Gene	PpHB1	PpRLK1	PpSPY	PpSINA	PpARF1	PpHK1
Length (kb)	2.6	2.7	2.7	1.0	1.1	1.2
Identity (protein)	93.2% (OsREV)	74.6% (OsLRK1)	89.2% (OsSPY)	97.8% (OsSINA)	84.0% (OsARF1)	67.7% (OsHK1)

Identity was calculated based on the putative amino acid sequences of homologous genes from bamboo and rice.

The sequences of these six genes were further analysed by alignment and phylogeny reconstruction. Interestingly, these bamboo genes (PpHB1, PpRLK1, PpSINA, PpARF1, and PpHK1) have closer relationships with their homologous genes from rice (REVOLUTA, CLAVATA1, SINAT5, ARF, and AHK4) than those from other species (Fig. 3, Supplementary Figs S2–S6 available at JXB online).

Fig. 3.

Alignment and phylogenetic analysis of the cloned genes PpHB1 and three REVOLUTA homologues. (A) Alignment of four REVOLUTA homologues. Sequence alignment of the deduced amino acids of Phyllostachys praecox PpHB1, rice OsREV (NM_197030), maize ZmREV (RLD1, AY501430), and Arabidopsis thaliana AtREV (REVOLUTA, AF188994) was conducted with AlignX. Regions of identity (white letters on a black background), conservation (white letters on a grey background), similarity (black letters on a grey background), and differences (white) are indicated. The HD-Zip domain (*) and START domain (+) are predicted according to AtREV. (B) The phylogeny reconstruction of REVOLUTA homologues.

To find out whether the cloned genes correspond to those analysed by microarray, Southern analyses were performed against bamboo genomic DNA. Three and four major bands were obtained for PpRLK1 and PpHB1, respectively, when blots of XbaI-digested genomic DNA were hybridized with a gene-specific probe for PpRLK1 and PpHB1 (Supplementary Fig. S7 at JXB online), suggesting that the bamboo genome contains at least three and four copies of PpRLK1 and PpHB1, respectively.

Expression analysis of four genes in different tissues and organs of bamboo

To verify the microarray results, semi-quantitative RT-PCR assays were carried out for the four genes (PpRLK1, PpHB1, PpSPY, and PpSINA) in different tissues and organs of bamboo. As shown in Fig. 4a, expression of PpRLK1, PpHB1, and PpSPY was significantly higher in young florets than in other organs. In the rhizome buds, the expression of PpSINA and PpSPY was markedly higher than that of PpHB1 and PpRLK1. No difference was observed for the expression of these four genes in the rhizome shoots and bamboo shoots. In leaves, only the expression of PpSINA was detectable, but not those of the other three genes. PpHB1, PpRLK1, and PpSPY were highly expressed in the rhizome shoots, rhizome buds, bamboo shoots, and young florets. In situ hybridization further confirmed the observations that PpRLK1 was expressed in the developing procambium (Fig. 4b), but not in the tips of the bamboo shoots (Fig. 4e), suggesting a close relationship of PpRLK1 to the meristem development of bamboo shoots. However, PpHB1 was expressed not only in the developing procambium (Fig. 4d), but also in the tips of the bamboo shoots (Fig. 4f), indicating its close correlation to rhizome bud formation as well as procambial development.

Fig. 4.

Expression analysis of the rhizome bud development-related genes. (a) Semi-quantitative RT-PCR analysis of some rhizome bud expressed genes in different tissues (A, rhizome shoot; B, rhizome bud; C, bamboo shoot; D, leaf; E, young floret). PpRLK1, PpHB1, PpSPY, and PpSINA were amplified for 28 cycles. PpACT1 was amplified for 25 cycles as a control. (b) RNA in situ hybridization performed in the rhizome bud with the PpRLK1 antisense probe. (c) RNA in situ hybridization performed in the rhizome bud as control with the PpRLK1 sense probe. (d) RNA in situ hybridization performed in the rhizome bud with the PpHB1 antisense probe. (e) RNA in situ hybridization performed in the bamboo shoot with the PpRLK1 antisense probe. (f) RNA in situ hybridization performed in the bamboo shoot with the PpHB1 antisense probe. Bar=100 μm.

Discussion

Rice cross-species microarray hybridization identifies homologous genes involved in the development of rhizome buds

DNA microarray technology is a basic tool to measure genome-wide changes in gene expression, and is helpful for understanding molecular mechanisms and functions. However, it is currently restricted to a few model species such as Arabidopsis and rice, due to lack of sequence information from other species. Recently, cross-species microarray analysis has been successfully established (Becher ). This technique is promising and has been widely used in the identification of genes in non-model animal and plant species (Moody ; Hudson ; Nieto-Díaz ). In the present study, a cross-species microarray with 7500 rice cDNAs 500 bp in length was used to identify the genes related to the development of bamboo buds. The data demonstrated that cross-species analyses do not significantly affect the hybridization reproducibility of gene chips, at least between bamboo and rice which are both gramineous species in the monocotyledons. The average hybridization efficiency is 84.82%, and the reproducibility of replicate hybridizations is 95%, suggesting that the microarray chips containing rice cDNAs of 500 bp in length can be effectively applied for bamboo heterologous hybridization.

The growth and development of bamboo is fairly complex. The exploration of the molecular bases of genes in bamboo development has been intensified. Recently, Lin made great efforts to identify bamboo genes differentially expressed in a bamboo albino mutant using suppressive subtractive hybridization (SSH) and microarray analysis. They prepared microarrays (18×27 mm) carrying 960 PCR-amplified cDNA clones hand picked from the bamboo SSH cDNA library on nylon membranes. Ten differentially expressed ESTs (expressed sequence tags) were isolated and sequenced. RT-PCR analysis of the ESTs supported the results of the microarray analysis. The combined use of this SSH library with microarray analysis will provide a powerful analytical tool for future studies of the bamboo genome. In the present study, rice cross-species microarray analysis was performed in bamboo for the first time. A total of 318 up- and 339 down-regulated genes were successfully identified. Among them, six up-regulated genes in rhizome bud were successfully cloned. Sequence analysis further indicated that these genes in bamboo shared medium to high identity with those in rice (66–90% identity at the nucleotide level, and 68–98% at the amino acid level). Thus, rice cross-species hybridization provides a useful tool to identify the conserved genes involved in the development of rhizome buds.

However, the cross-species microarray does raise some concerns about signal reduction and data analysis. Moreover, hybridization with imperfectly complementary nucleic acids would lead to severe bias because of the high number of interactions among probes and templates (Pariset ). As is known, there might be several homologous cDNAs contributing to one high signal intensity in a sample. It is quite difficult to determine the specificity of cross-species hybridizations, which is a main limitation of cross-species microarray. Therefore, the candidate genes identified from the cross-species microarray require further validation. For the validation, the genes with high signal intensity and an up-regulation ratio >3.0, and the corresponding rice sequences known to be involved in meristem development, were selected. Also degenerate primers corresponding to the conserved region of the homologous genes, mainly based on sequence alignment from Oryza sativa, Zea mays, Hordeum vulgare, and Arabidopsis thaliana, were also designed in order to amplify as many homologous genes as possible. However, only one fragment was cloned with cDNA templates from bamboo rhizome buds. No other sequences were obtained despite repetitive PCR tests being performed. It is suggested that the cloned fragment was, at least, a main member among the targeted gene family expressed in rhizome buds.

It is noteworthy that the degenerate primers designed from the conserved region of homologous genes of rice and other species do not guarantee amplification of the true orthologue of bamboo. Sequence divergence and the complexity of the targeted gene family will influence the outcome of the PCR amplification. The highly variable degree of sequence similarity indicates that members of large gene families were cloned in a somewhat random fashion (Hudson, 2007). It could not be extrapolated that the cloned genes from bamboo are putative orthologues in rice without functional analysis and genomic alignment.

Rice cross-species microarray hybridization provides useful gene expression data for understanding the development of bamboo rhizome buds

Cross-species microarray hybridization has been considered as one potentially useful technique to provide gene expression data and deduce gene function for applying model organism genomics in related species for which few functional genomics resources are available (Chismar ; Becher ; Lee ; Hudson ; Nieto-Díaz ). In the present study, 26 genes related to signalling and biosynthesis of auxin are highly expressed in the rhizome buds of bamboo using rice cross-species microarray (Table 3). The results suggested that these genes might play important roles in the development of bamboo shoot. High concentrations of hormones were proved to be closely related to the formation of the bamboo shoot (Huang ). The function of ethylene in the development of the rhizome bud has not been reported to be associated with bamboo shoot formation so far. In addition, the high expression of some receptor-like kinase genes and meristem initiation-related transcriptional factors was also observed in the rhizome buds, implying that these genes might be related to the meristem characteristics during rhizome bud development.

In addition, it is reasonable for the bamboo meristem bud to accumulate plenty of starch after the cool autumn and cold winter, which could provide both energy and gravisensing for the burst in growth in the spring. The starch–statolith hypothesis, proposed by Haberlandt and Němec, was widely used to explain how plants perceive gravity, and how the rhizome buds grow upward into bamboo shoot and in a horizontal stretch into rhizome (Audus, 1962). In the present study, starch synthesis- and cell wall synthesis-related genes were found to be highly expressed in the bamboo meristem (Table 3). The high expression levels of metabolism-related genes, especially starch-branching enzyme class II (sbe2-1) and starch-branching enzyme I (Rbe1), in the bamboo rhizome buds imply high metabolism intensity in the rhizome buds. It is thus conceivable that the meristem development of the rhizome bud is different from that of the ordinary axillary bud in plants. Similarly, the high expression of the cell wall synthesis-related genes, such as the XTH gene, suggests that the expression of XTH-homologous genes in bamboo might be necessary for cell replication and growth during meristem development. Most members of the rice XTH family exhibited organ- and growth stage-specific expression patterns, although their exact functions have not been revealed yet (Yokoyama ). A recent report demonstrated that sucrose synthase isoforms play a variety of roles in bamboo shoots, directing translocated carbon towards both the polysaccharide biosynthesis and energy production necessary to sustain the remarkably rapid growth of bamboo (Chiu ).

Potential functions of the genes identified from cross-species microarray hybridization

Rhizome bud development is one of the most important problems in bamboo research. Previous studies have revealed some anatomical, physiological, and molecular mechanisms for bamboo growth (Hu ; Zhang ; Huang ; Chiu ; Hsieh ). In the present study, six genes related to rhizome bud development were successfully cloned from rhizome buds. PpSPY may affect meristem initiation indirectly according to its expression pattern and the function of its homologous gene in Arabidopsis. Previous research revealed that SPINDLY, homologous to PpSPY, encoded tetratricopeptide repeat proteins involved in gibberellin signal transduction in both monocotyledons (Robertson, 2004) and dicotyledons (Greenboim-Wainberg ). The high expression of PpSPY may be helpful in elucidating the signal transduction of gibberellins in the formation and development of bamboo shoot, although it is still not clear whether the gibberellins play an important role during the transition of the rhizome bud to bamboo shoot (Huang ). SINAT5, which is an Arabidopsis homologue of the RING-finger Drosophila protein, is expressed in the vascular tissues of mature roots and can target NAC1 to down-regulate auxin signals in the plant cells (Xie ). However, PpSINA in bamboo is expressed in all the organs including leaves (Fig. 4), suggesting that the role of PpSINA in bamboo might be different from that in Arabidopsis and rice. RT-PCR analysis showed that PpRLK1 was expressed in both the rhizome bud and bamboo shoot. In situ hybridization confirmed that PpRLK1 was expressed in the developing procambium (Fig. 4b), but not in the tips of the bamboo shoots (Fig. 4e), suggesting that the function of PpRLK1 might be necessary for the meristem development of the bamboo shoot. Previous research revealed that the homologuous gene of PpRLK1, CLV1, in Arabidopsis can determine the size of the meristem (Gross-Hardt et al., 2003), and OsLRK1 in rice plays a critical role in meristem development (Kim ). However, both RT-PCR and in situ hybridization showed that PpHB1 was strongly expressed in the developing procambium (Fig. 4d) and in the tips of the bamboo shoots (Fig. 4f), suggesting that PpHB1 might play an important role in rhizome bud formation as well as procambial development.

In conclusion, the present results demonstrate that there were significant differences in the anatomical structures among rhizome buds, rhizome shoots, and bamboo shoots, and for the first time confirm that rice cross-species microarray hybridization can be successfully applied to identify genes in bamboo. According to the rice cross-species microarray hybridization, six genes related to the development of the bamboo rhizome bud were cloned, which contributed to a further understanding of the molecular mechanism of bamboo rhizome bud development.

Supplementary data

Supplementary data are available at JXB online.

Microarray data of leaves (left) and rhizome lateral buds (right). White arrows indicate up-regulated genes and pink arrows indicate down-regulated genes in rhizome lateral buds.

Alignment of four CLAVATA homologues and their structure.

Alignment of four SPINDLY homologues and their structure.

Alignment of four SINAT5 homologues and their structure.

Alignment of the DNA-binding domain among PpARF1, OsARF1, and ARF1. The deduced amino acid sequences of OsARF1 and ARF1 were aligned with the Dayhoff amino acids distance matrix.

Phylogenetic analysis of the isolated or predicted PpHK1, OsHK1 ZmHK1, and AHK1 homologues. The deduced amino acid sequences of PpHK1, OsHK1, ZmHK1, and AHK1 were aligned with the Dayhoff amino acids distance matrix.

Southern analysis of PpRLK1 and PpHB1 in genomic DNA from Phyllostachys praecox. DNA was digested with XbaI.

The rice EST on the microarray and corresponding homologous genes for PpRLK1 cloning.

The rice ESTs on the microarray and corresponding homologous genes for PpHB1 cloning.

The rice EST on the microarray and corresponding homologous genes for PpSPY cloning.

The rice EST on the microarray and corresponding homologous genes for PpSINA cloning.

The rice EST on the microarray and corresponding homologous genes for PpARF1 cloning.

The rice EST on the microarray and corresponding homologous genes for PpHK1 cloning.