Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Rice tillering is an important agronomic trait for grain production. The HIGH-TILLERING DWARF1 (HTD1) gene encodes an ortholog of Arabidopsis MAX3. Complementation analyses for HTD1 confirm that the defect in HTD1 is responsible for both high-tillering and dwarf phenotypes in the htd1 mutant. The rescue of the Arabidopsis max3 mutant phenotype by the introduction of Pro35S:HTD1 indicates HTD1 is a carotenoid cleavage dioxygenase that has the same function as MAX3 in synthesis of a carotenoid-derived signal molecule. The HTD1 gene is expressed in both shoot and root tissues. By evaluating ProHTD1:GUS expression, we found that the HTD1 gene is mainly expressed in vascular bundle tissues throughout the plant. Auxin induction of HTD1 expression suggests that auxin may regulate rice tillering partly through upregulation of HTD1 gene transcription. Restoration of dwarf phenotype after the removal of axillary buds indicates that the dwarfism of the htd1 mutant may be a consequence of excessive tiller production. In addition, the expression of HTD1, D3 and OsCCD8a in the htd1 and d3 mutants suggests a feedback mechanism may exist for the synthesis and perception of the carotenoid-derived signal in rice. Characterization of MAX genes in Arabidopsis, and identification of their orthologs in pea, petunia and rice indicates the existence of a conserved mechanism for shoot-branching regulation in both monocots and dicots.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Shoot branching plays an important role in the formation of distinct plant architectures. Shoot branches arise from axillary meristems, which form in the axils of leaves on the primary shoot axils. The activity of axillary meristems is regulated by a wide range of genetic, developmental and environmental factors (Beveridge et al., 2003; Leyser, 2003; Shimizu-Sato and Mori, 2001). In many plant species, the growth of an axillary bud is initially suppressed by the shoot tip, a phenomenon known as apical dominance (Thimann, 1933). Several lines of evidence suggest that auxin, synthesized at the shoot apex, is responsible for this phenomenon (Klee et al., 1987; Panigrahi, 1966; Romano et al., 1991); however, other lines of evidence strongly support the idea that auxin indirectly inhibits lateral branching. This possible indirect effect suggests the existence of a second messenger that carries the auxin signal into the bud (reviewed by Leyser, 2003). Cytokinin has been proposed as the strongest candidate for this second messenger. Direct application of cytokinin to buds promotes their outgrowth, and cytokinin levels rise in buds as they are activated (Cline, 1991; Turnbull et al., 1997). The function of cytokinin in outgrowth of axillary buds has also been supported by a positive correlation between cytokinin levels and bushy phenotype in the corresponding mutants or transgenic plants (Catterou et al., 2002; Chaudhury et al., 1993; Helliwell et al., 2001; Medford et al., 1989; Tantikanjana et al., 2001; Zubko et al., 2002).
Monocots and dicots share many similarities in branching during vegetative development (McSteen and Leyser, 2005). During rice development, axillary buds produced from the basal nodes of the plant grow as tillers during vegetative growth, while those produced after transition to the reproductive phase grow as panicle branches and spikelets. The rice tillering process involves two developmental stages: the formation of axillary meristems in the leaf axils and their subsequent outgrowth (Hanada, 1993). Due to the agronomic importance of tillering in rice breeding, many rice tillering mutants have been reported. The rice MOC1 gene, a putative GRAS family member, was found to control the formation of axillary buds (Li et al., 2003). It is interesting to note that high tillering often accompanies dwarfism in rice (Iwata et al., 1995). For instance, in fc1 (Ostb1) and dwarf3 (d3) mutants, high tillering and dwarfism are exhibited simultaneously. Both OsTB1 and D3 were identified as negative regulators of tillering that inhibit the outgrowth of axillary buds but do not affect the formation of axillary meristems (Ishikawa et al., 2005; Takeda et al., 2003).
Previously, we characterized the abnormal phenotypes of a rice high-tillering and dwarf mutant (htd1), and mapped the HTD1 locus to a 30 kb region on chromosome 4 by positional cloning (Zou et al., 2005). In this study, HTD1 was confirmed to be OsCCD7 by a complementation test, and the gene function was indirectly examined by expressing HTD1 in the Arabidopsis max3 mutant. In addition, the interaction between tiller number and plant height was analyzed.
High tillering of the htd1 mutant results from the release of axillary buds
As previously reported, the htd1 mutant exhibited excessive tillers, due to either the release of axillary buds or an increased number of axillary buds (Zou et al., 2005). To examine the role of HTD1 in rice tillering, we further characterized the htd1 mutant at the anatomical level. The htd1 mutant and its wild-type Nanjing6 (NJ6) were grown both in half-strength Murishige and Skoog (MS) medium and in soil. At the third-leaf stage, the axillary buds of NJ6 seedlings were too small to be observed directly, but could be observed by light microscopy (Figure 1a). However, the axillary buds of htd1 seedlings were clearly visible (Figure 1b). By the fourth-leaf stage, each htd1 plant had more than two tillers, while only one tiller was observed in each of the NJ6 plants (Figure 1c,d). Both the htd1 and NJ6 plants had a single axillary bud in each leaf axil, indicating it is the release of axillary buds and not the number of axillary buds that is responsible for the greater tiller number in htd1 plants.
The dwarf phenotype of the htd1 mutant could be rescued by removing axillary buds
To further investigate the possible interaction between tiller number and plant height, we observed the change in plant height after removal of axillary buds in htd1 and NJ6 plants. Removal of all the buds reduced the tiller number (36.0 ± 3.2, n = 4) to one, and resulted in a significant increase in htd1 plant height, from 52.4 ± 2.3 cm (n = 4) to 88.8 ± 4.4 cm (n = 4; Figure 2a). No difference was seen in plant height between the NJ6 plants from which all axillary buds had been removed and control plants (120.5 ± 8.1 cm versus 119.1 ± 4.0 cm, n = 4) under the same experimental conditions, except that the flag leaves of NJ6 plants from which the axillary buds had been removed grew larger than those of control plants (Figure 2b,c). In addition, the panicles and leaves of the htd1 plants increased in size, and were similar to those of NJ6 plants following removal of axillary buds (Figure 2b,c). These changes suggest that the dwarf plant and the decreased size of the panicle and leaves of the htd1 mutant may at least partly result from the outgrowth of excessive tillers.
Complementation test with OsCCD7 in the htd1 mutant
In our previous study, the HTD1 gene was fine-mapped to a 30 kb DNA region of chromosome 4. By DNA sequence comparison, we found that a nucleotide substitution existed in OsCCD7, the ortholog of MAX3. By comparison analysis of OsCCD7 in the htd1 mutant and other rice varieties, we identified OsCCD7 as the candidate gene for HTD1 (Zou et al., 2005). To further confirm that OsCCD7 corresponds to the HTD1 locus, a complementation test was conducted in the htd1 mutant. A 6.9 kb genomic DNA fragment containing the entire OsCCD7 gene plus a 3.7 kb upstream region was inserted into a binary vector, pCAMBIA1301. This construct was introduced into the htd1 mutant by Agrobacterium-mediated transformation. Three independent transgenic lines (T0) with the wild-type phenotype were achieved. Investigation of tillering and plant height in T1 progeny plants showed that all transgenic plants containing the exogenous OsCCD7 gene exhibited normal tiller number (11.7 ± 1.0) and plant height (124.8 ± 4.3 cm) compared with wild-type plants (Figure 3). However, transgenic plants containing an empty vector exhibited more tillers (77.6 ± 5.2) and greater dwarfism (63.2 ± 2.4 cm) (Figure 3), and appeared to be htd1 plants. This result suggests that the htd1 mutant phenotype is caused by loss-of-function of the OsCCD7 gene. DNA analysis of T1 progeny from a self-pollinated transgenic line showed co-segregation of the transgene and the tillering and dwarf phenotypes, providing further evidence that OsCCD7 is the HTD1 gene (data not shown).
Molecular characterization of the HTD1 gene
Based on HTD1 cDNA sequence data (accession number AK109771), the structure of the genomic HTD1 gene was determined. It contains seven exons (Figure 4a) and encodes a protein of 609 amino acids belonging to a family of CCD proteins (Tan et al., 2003), including Arabidopsis proteins MAX3/CCD7 and MAX4/CCD8. In this study, we found that, in the htd1 mutant, the gene carried a single point mutation (C to T), resulting in the proline at position 596 of HTD1 being replaced by leucine (corresponding to the residue at 599 of HTD1 in Nipponbare). In addition, DNA blot analysis indicated that HTD1 is a single-copy gene in the rice genome (data not shown). The amino acid sequences of HTD1 and MAX3 show 50.79% identity (Figure 4b), suggesting that HTD1 may be the ortholog of MAX3, a CCD identified by two independent groups (Booker et al., 2004; Schwartz et al., 2004).
Phenotypic rescue of the max3 mutant by the introduction of HTD1 suggests that HTD1 has the same function as MAX3 in synthesis of the carotenoid-derived signal molecule
To characterize the function of HTD1, a construct of Pro35S:HTD1 was introduced into the max3-9 mutant. Of 30 transgenic lines, 25 showed over-expression of HTD1 and exhibited similar phenotypes to those of the wild-type (Col), and the other five lines, which lacked HTD1 transcription, were indistinguishable from the max3-9 mutant (Figure 5). The phenotypic rescue of max3 mutants by Pro35S:HTD1 indicates that HTD1 is the ortholog of MAX3 in rice; therefore, HTD1 should have the same biochemical function as MAX3 in synthesis of the carotenoid-derived signal controlling shoot branching.
Expression pattern of the HTD1 gene
RNA gel blot analysis showed that the HTD1 gene was expressed in all tissues examined. The levels of expression were generally higher in aerial tissues (leaf, stem and panicle) than in root tissue (Figure 6a). To confirm the expression patterns, a binary vector containing the uidA gene (β-glucuronidase, GUS) driven by the HTD1 promoter (ProHTD1:GUS) was constructed, and 11 independent ProHTD1:GUS transgenic lines were generated by Agrobacterium-mediated transformation. Five independent ProHTD1:GUS transgenic lines were used to observe GUS expression. The transgenic plants showed GUS expression in leaf, sheath, stem, panicle and root tissue (Figure 6b–g). The strongest intensity of GUS expression was found at the node of the stem where axillary meristems initiate (Figure 6e). Furthermore, cross-sections of sheath and stem tissue showed that GUS expression was primarily found in the vascular-associated tissues (Figure 6f,g), demonstrating that the HTD1 gene is mainly expressed in vascular bundles throughout the plant. No GUS expression was detected in the axillary buds themselves. In Arabidopsis, MAX3, the ortholog of HTD1, was detected both in shoot and root tissues, with the highest levels occurring in root tissue (Booker et al., 2004). In this study, ProHTD1:GUS expression was also detected in shoot and root tissues, but with a lower level of expression in root tissue. Further study of this apparent difference is necessary.
HTD1 expression is induced by auxin
To examine whether HTD1 transcription is regulated by auxin, HTD1 expression was investigated in NJ6 plants. Rice seedlings (the third-leaf stage) were sprayed with 10 μm exogenous auxin (NAA), and RNA was extracted from the whole plant at specific time intervals between 0 and 24 h after application. Using RT-PCR, we found that HTD1 expression was higher in auxin-treated plants than in normal plants at the 3 h time point. Increased HTD1 transcription was maintained for 6 h after auxin treatment; however, after 12 h, HTD1 expression decreased to the basal level (Figure 7). HTD1 expression did not exhibit a significant increase in control seedlings treated with water (data not shown).
Expression of other tillering-related genes in htd1 and d3 mutants
To investigate whether the expression of other tillering-related genes is affected in htd1 plants, expression of D3 and OsCCD8a was monitored by RT-PCR. The expression of D3 was significantly increased in the htd1 mutant compared to NJ6 plants, while the expression of OsCCD8a in the htd1 mutant was only slightly higher than in NJ6 plants (Figure 8a). However, the fact that HTD1 did not exhibit increased expression in the htd1 mutant might be attributed to its unstable transcript level in htd1 plants. HTD1 and OsCCD8a were analyzed in the d3 mutant, and, contrary to the elevated expression of D3 and OsCCD8a in the htd1 mutant, both HTD1 and OsCCD8a exhibited lower expression in the d3 mutant than in the wild-type plants (Figure 8b).
Due to the agronomic importance of dwarfism mutants in rice breeding, many dwarf mutants have been reported and characterized. One group of dwarf mutants, the high-tillering dwarfs, exhibit characteristics of both dwarf and bushy plants (Iwata et al., 1995). In our previous study, a high-tillering dwarf (htd1) mutant was described in detail (Zou et al., 2005). The bushy and dwarf phenotypes of the htd1 mutant were found to be independent of the effects of the known plant hormones IAA, ABA and GA. By map-based cloning, the HTD1 gene was mapped to a 30 kb DNA region on chromosome 4, and OsCCD7 was proposed as the candidate gene. In this study, OsCCD7 was confirmed to be the HTD1 gene, and HTD1 was shown to have the same biochemical function as MAX3 by complementation of HTD1 in the max3 mutant.
The interaction between tiller number and plant height
The outgrowth of axillary buds in rice is also affected by various environmental factors, including nutrients, planting density, light, temperature and water supply. Several studies have reported that assimilate supply and light quality could affect tillering in grass crops (Jitla et al., 1997; Lafarge et al., 2002). Because the htd1 mutant is unable to effectively regulate shoot branching, the emergence of excessive tillers may exhaust photosynthetic products (such as sugar), which may in turn affect the cell cycle by adjustment of the expression of CycD2 and CycD3 (Riou-Khamlichi et al., 1999, 2000; Soni et al., 1995). Therefore, we speculated that the apical dominance of the primary shoot would be compromised if axillary buds were not inhibited.
Mutation of HTD1 results in a high-tillering phenotype
In this study, we have shown that mutation of HTD1 is responsible for the high-tillering and dwarf phenotype of the htd1 mutant. HTD1 encodes a member of the 9-cis epoxycarotenoid dioxygenase family (Tan et al., 2003), which can be grouped into multiple distinct clades by phylogenetic analysis (Snowden et al., 2005; Sorefan et al., 2003; Tan et al., 2003). An ABA biosynthesis enzyme, VP14 (maize) falls within one large clade within NCEDs, while HTD1 and its Arabidopsis ortholog, MAX3 form a distinct clade. This phylogenetic data and our previous quantitative measurement of ABA in the htd1 mutant (Zou et al., 2005) suggest that HTD1 is not involved in the biosynthesis of ABA. Within the HTD1 clade, MAX3 has been identified as a dioxygenase that catalyzes a specific 9–10 cleavage of β-carotene to produce 10-apo-β-carotenal (C27) and β-ionoe (C13), while MAX4 further catalyzes the MAX3-derived C27 intermediate to produce 13-apo-β-carotenone (C18) (Booker et al., 2004; Schwartz et al., 2004). The phenotypic rescue of max3 by the introduction of Pro35S:HTD1 indicates that HTD1 is a CCD that has the same function as MAX3 in the synthesis of a carotenoid-derived signal molecule.
A conserved branch-regulating mechanism shared by both monocots and dicots
In Arabidopsis, MAX1–MAX4 have been shown to function in a single pathway that is required for the production and perception of a novel carotenoid-derived signal controlling shoot branching (Booker et al., 2004; Sorefan et al., 2003; Stirnberg et al., 2002). In Arabidopsis, pea and petunia, three orthologous genes (MAX4, RMS1 and DAD1) and their similar phenotypes indicate that a conserved mechanism probably exists that regulates the branch development of dicots. In rice, corresponding orthologs of MAX genes have also been reported (Ishikawa et al., 2005; Tan et al., 2003). In particular, as mentioned above, phenotypic rescue of the max3 mutant by HTD1 demonstrates that MAX3 and HTD1 have the same function in the synthesis of a novel carotenoid-derived signal, suggesting a conservative branch-regulating mechanism may exist in both monocots and dicots. However, contrary to the generally low expression pattern of MAX3 in Arabidopsis, with the highest expression levels in root tissue (Booker et al., 2004), HTD1 in rice exhibited higher expression levels in the aerial tissues and relatively low levels in the root tissue.
Auxin may regulate rice tillering partially through upregulation of HTD1 transcription
Apically derived auxin is known to regulate shoot branching by inhibiting the activity of axillary meristems in Arabidopsis, and many other plants (Chatfield et al., 2000), but its mechanism of action is complex and indirect (reviewed by Leyser, 2003). Xu et al. (2005) reported that, in OsPIN1 RNAi transgenic plants, the increase in tiller number was associated with the reduction of OsPIN1 gene expression, which is involved in polar auxin transport in rice. Based on the role of auxin in many plants, we speculated that auxin might also be the key regulator in the development of rice tillers. However, we did not find significant differences in the levels of auxin (IAA) between htd1 and wild-type plants in a previous study (Zou et al., 2005). RT-PCR was used to test HTD1 transcription when NAA was applied, to see whether HTD1 acts downstream of auxin. As expected, HTD1 expression was quickly induced. This suggests that auxin may regulate rice tillering partly through upregulation of HTD1 transcription. Similar experiments with MAX4/RMS1 indicated that auxin could induce the transcript of MAX4/RMS1, although with different patterns in pea and Arabidopsis (Sorefan et al., 2003). Recently, the MAX-dependent hormone was identified as a novel regulator of auxin transport (Bennett et al., 2006). The capacity of auxin transport is increased in max primary stems by increasing the expression of PIN auxin efflux facilitators. As the ortholog of MAX3, HTD1 may have a similar function in regulating auxin transport. It should be noted that, as MAX1,HTD1 is also mainly expressed in vascular tissues, in which the PIN family members of auxin efflux facilitators are localized in the cells of the xylem parenchyma and mediate directional movement of auxin down the stem (Galweiler et al., 1998; Okada et al., 1991).
A feedback mechanism may exist for synthesis and perception of the carotenoid-derived signal in rice
In pea, a long-distance feedback signal was proposed by analysis of the RMS1 transcript level in rms mutants, and particularly the rms4 mutant, which exhibited the highest transcript level of RMS1, compared with wild-type plants and other rms mutants (Foo et al., 2005). Consistent with the upregulation of RMS1 in rms4 plants, MAX4 expression was slightly elevated in the hypocotyls of the Arabidopsis max2 mutant (Bainbridge et al., 2005). In addition, feedback upregulation of DAD1 (PhMAX4) expression in petunia was also observed (Snowden et al., 2005). Recently, RMS4 has been reported as an ortholog of MAX2 and D3 (Beveridge, 2006). Although some branching-related orthologous genes and their similar phenotypes have been reported in different plant species, little is known about how these orthologous genes affect branch-regulating mechanisms in different plant species. To examine whether a similar feedback mechanism exists in the synthesis and perception of a carotenoid-derived signal in rice, the expression of HTD1, D3, and OsCCD8 was analyzed in htd1 and d3 mutants and wild-type plants. In this study, OsCCD8a and OsCCD8b were both analyzed at first, but we could not detect the transcript of OsCCD8b by RT-PCR. Using whole-plant-tissue RNAs for the RT-PCR analysis may have weakened or hidden the exact expression patterns of these tillering-related genes. Therefore, more precise experiments that use specific tissues should be performed to identify the expression patterns of these genes and study the feedback mechanisms. However, as a first insight, the gene expression analysis did show that both D3 and OsCCD8a exhibited increased expression, particularly a significantly elevated transcription of D3 in the htd1 mutant. Meanwhile, we detected no obvious difference in HTD1 expression between the htd1 and NJ6 plants (data not shown). On the other hand, both HTD1 and OsCCD8a exhibited lower expression in the d3 mutant, in contrast to the increased expression of RMS1 in the rms4 mutant. The increased expression of D3 and OsCCD8a in the htd1 mutant and the reduced expression of HTD1 and OsCCD8a in the d3 mutant are indicative of a feedback control mechanism that is responding to changes in the synthesis and perception of the carotenoid-derived branch-inhibiting signal molecule in rice. The different expression patterns of RMS1, MAX4 and OsCCD8a genes in pea, Arabidopsis and rice, respectively, suggest that feedback mechanisms of rice may be different from those of pea and Arabidopsis, although a conserved branching-regulating mechanism is probably shared by both monocots and dicots.
Plant material and growth conditions
The htd1 mutant and NJ6 and Nipponbare wild-type varieties were analyzed. Detailed genetic information on these plants is provided by in our previous study (Zou et al., 2005). The d3 mutant was provided by Dr Qian Qian (China National Rice Research Institute). Rice seeds were grown on half-strength MS medium under a 12 h light/12 h dark regimen at 30°C for 2 weeks. Rice plants were grown in the greenhouse at 24–30°C (12 h light/12 h dark). For analysis of HTD1 induction expression by exogenous auxin, RNA was extracted from NJ6 seedlings at the third-leaf stage at specific time intervals (3, 6, 12 and 24 h) after spraying with 10 μm NAA. To study the effect of removing axillary buds, the htd1 mutants, NJ6 and Nanjing11 (a semi-dwarf cultivar containing sd1), were analyzed and their axillary buds were removed from the leaf axil. In addition, for the analysis of the HTD1 expression pattern, Nipponbare (japonica) was used as transformation material to monitor the reporter gene ProHTD1:GUS. For the analysis of other tillering-related gene expression, RNA used in RT-PCR was derived from whole rice plant tissues. The Arabidopsis max3-9 mutant and Columbia (wild-type) were provided by Dr Ottoline Leyser (University of York, UK). These plants were grown under 16 h light/8 h dark conditions at 22°C, and the vector containing Pro35S:HTD1 was introduced into the max3-9 plants by Agrobacterium-mediated method.
Vector construction and plant complementation
A DNA fragment containing a full-length genomic HTD1 gene was obtained by digesting BAC clone AL663000 with EcoRI. The digested fragment was inserted into a binary vector pCAMBIA1301 harboring a hygromycin-resistant gene. The fragment was introduced into the htd1 mutant by an Agrobacterium-mediated transformation method (Hiei et al., 1994). An empty pCAMBIA1301 vector harboring a hygromycin resistance marker was transformed into the htd1 mutant as a negative control.
The construct for HTD1 overexpression (Pro35S:HTD1) was generated by introducing HTD1 cDNA into the plasmid pzh01 (Xiao et al., 2003) after the GUS gene had been removed from the plasmid by digestion with XbaI and KpnI. The HTD1 cDNA was amplified using cDNA clone AK109771 as the template. Transgenic Arabidopsis plants (Pro35S:HTD1) were generated by the floral dip method (Clough and Bent, 1998), and screened on solid plates containing 25 μg ml−1 hygromycin. The construct ProHTD1:GUS was created by introducing the 3 kb promoter sequence of HTD1 into pCAMBIA1301 after the 35S promoter had been removed by digestion with XbaI and NcoI. The promoter regions (3 kb) of HTD1 were amplified by PCR. Nipponbare was used as the plant material to obtain transgenic lines containing ProHTD1:GUS by an Agrobacterium-mediated transformation method.
RNA isolation and RT-PCR analysis
RNA was isolated from plant material using Trizol reagent according to the manufacturer's manual (Invitrogen, Carlsbad, CA, USA). In RNA blot analysis, the RNAs, 20 μg in each lane, were separated on a 1.0% denaturing gel and transferred to a Hybond N+ filter membrane (Amersham Pharmacia Biotech, Uppsala, Sweden). The membranes were hybridized with a probe of PCR product amplified from a template of HTD1 cDNA fragments. The extracted RNA was treated with DNaseI during RNA purification to eliminate genomic DNA contamination; the DNaseI treatment was performed according to the protocols recommended by the manufacturer (Invitrogen). First-strand synthesis of HTD1 cDNA was performed on 1 μg of total RNA using a reverse transcription system (catalog number A3500, Promega, Madison, WI, USA). In the analysis of HTD1 expression, two HTD1 gene-specific primers, 5′-GAGGATGGTGGCTATGTTCTTCT-3′ and 5′-AGTAGTTATTTGGTTCCCCTGAT-3′ were used to amplify 321 bp cDNA fragments of the HTD1 cDNA. Two gene-specific primers, 5′-CTGGCCTCTAGAAGGAGTAGATTAGGTAG-3′ and 5′-CTGATGAAGAGAAACCAGGGAAAAC-3′ were designed to amplify 342 bp cDNA fragments of D3/OsMAX2; primers 5′-CCTCGGCAGGAAGTACCAGTATG-3′ and 5′-TGATCCTAGTCTTCTCGGCTACAG-3′ were used to amplify 382 bp cDNA fragments of OsCCD8a (Accession no. AP003296.3). The conditions used during PCR were 94°C for 5 min, 94°C for 30 sec, 53°C for 30 sec and 72°C for 30 sec in 25–40 cycles. In the analysis of the induction of HTD1 expression, Actin-specific primers, 5′-TCATGAAGATCCTGACGGAG-3′ and 5′-AACAGCTCCTCTTGGCTTAG-3′, were amplified as the control.
GUS staining was performed according to the method described by Jefferson (1989). Various organs of ProHTD1:GUS transgenic seedlings were incubated in a solution containing 50 mm NaP buffer, pH 7.0, 5 mm K3Fe(CN)6, 5 mm K4Fe(CN)6, 0.1% Triton X-100 and 1 mm X-Gluc, and incubated at 37°C for 12 h. All samples were vacuum-treated for 5 min before staining.
We thank Dr Ottoline Leyser of the University of York, UK, for providing the Arabidopsis max3-9 mutant. We thank Dr Qian Qian of the China National Rice Research Institute for providing the rice d3 mutant. This research was supported by grants from the National Science Foundation of China (90208001 and 30550005), the Chinese Academy of Sciences (KSCX2-SW-306), and the Plant Gene Research Centre (Beijing).