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.
Present address: Boyce Thompson Institute at Cornell University, Tower Road, Ithaca, NY 14850, USA.
These authors contributed equally to the paper.
Overexpression of a salicylic-acid (SA)-inducible Arabidopsis DNA binding with one finger (Dof) transcription factor, called OBF-binding protein 3 (OBP3; AtDof3.6), has previously been shown to result in growth defects. In this study, suppressive subtraction hybridization (SSH) was used to isolate genes induced in an OBP3-overexpression line and several putative clones, called OBP3-responsive genes (ORGs), were isolated. The link with the induced expression levels of these genes and OBP3 overexpression was confirmed by analysing additional OBP3-overexpression lines. ORG1 through ORG4 are novel genes, while ORG5 is an extensin gene, AtExt1. While ORG4 has no similarity with other proteins in the database, ORG1 has weak similarity in different regions of the predicted protein with CDC2 and fibrillin. ORG2 and ORG3 share 80% overall identity in their deduced amino acid sequences and contain a basic helix-loop-helix DNA-binding domain, suggesting that ORG2 and ORG3 may be transcription factors. The expression of the ORG1, ORG2 and ORG3 genes was co-regulated under all conditions examined including upregulation by SA and downregulation by jasmonic acid (JA). Fifteen OBP3-silenced lines were generated to further explore the function of OBP3. Although there were no visible phenotypic changes in any of these lines, the expression of ORG1, ORG2 and ORG3 was reduced. Among the ORG genes, ORG1, ORG2 and ORG3 contained the highest number of potential Dof-binding sites in the promoter region, and their expression was significantly increased within 3 h after induction of OBP3 expression using an inducible promoter system, and closely reflected the expression levels of the exogenous OBP3 protein. The results from the gain-of-function and loss-of-function experiments suggest that the ORG1, ORG2 and ORG3 genes are direct target genes of OBP3.
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.
Although there is considerable similarity in the regulation of transcription between animals and plants, there is growing evidence for transcription factor families that are unique to a particular kingdom. One class of transcription factors unique to plants is a novel type of zinc finger protein called DNA binding with one finger (Dof) proteins. Dof proteins contain a highly conserved 52-amino acid region called the Dof domain, which contains a single CX2CX21CX2C-type zinc finger motif (reviewed in Yanagisawa, 2002). The zinc finger in the Dof domain is essential for DNA binding as specific mutations in cysteine residues within the zinc finger abolish DNA binding as do zinc chelaters (see Yanagisawa, 2002 and references within). Binding sites for Dof proteins typically contain (A/T)AAAG or its complementary sequence (Yanagisawa and Schmidt, 1999). Some Dof proteins have also been shown to interact specifically with other types of transcription factors including bZIP proteins (Kang and Singh, 2000; Vincente-Carbajosa et al., 1997; Zhang et al., 1995), HMG1 (Yanagisawa, 1997), Myb proteins (Diaz et al., 2002) and other Dof proteins (Yanagisawa, 1997).
We have focused on a subgroup of the 37 Arabidopsis Dof proteins called OBF-binding proteins (OBPs; Kang and Singh, 2000; Zhang et al., 1995). The first OBP protein (OBP1) was isolated as a protein that interacts with an ocs-element-binding factor, OBF4 (Zhang et al., 1995). ocs elements are found in both pathogen and plant promoters, where they function as stress-responsive elements (see Singh et al., 2002 and references within). While the OBP proteins showed similar in vitro DNA binding, protein–protein interaction and transcription-activation properties, they had distinct RNA-expression patterns with some family members showing a strong correlation with the activity of the ocs element (Kang and Singh, 2000). A range of phenotypes was observed in plants overexpressing OBP3, also called AtDof3.6, based on its chromosomal position (Yanagisawa, 2002). These phenotypes included retarded growth in both roots and aerial parts of the plant that, in the most severe cases, led to death (Kang and Singh, 2000).
To gain insight into the function of OBP3, we isolated genes induced in an OBP3-overexpression background using suppressive subtraction hybridization (SSH; Diatchenko et al., 1996). Given that OBP3 is a DNA-binding protein with transcriptional activation activity, we reasoned that the phenotypes observed from OBP3-overexpression lines might result from constitutively expressing OBP3 target genes. Using SSH, five OBP3-responsive genes (ORGs) were isolated. Expression of three of these genes –ORG1, which encodes a protein of unknown function, and ORG2 and ORG3, which encode closely related basic helix-loop-helix domain (bHLH) proteins – was co-regulated under all conditions examined, including upregulation by salicylic acid (SA) and downregulation by jasmonic acid (JA). To initiate loss-of-function experiments with OBP3, we generated fifteen OBP3-silenced lines using intron-spliced hairpin (ihp) RNA. The OBP3-silenced lines had no visible phenotypic changes in contrast to what was seen with the OBP3-overexpression lines. However, the expression of the ORG1, ORG2 and ORG3 genes was reduced in the OBP3 silenced lines. ORG1, ORG2 and ORG3 expression was also significantly increased within 3 h after induction of OBP3 expression using an inducible promoter system, and closely reflected the expression levels of the exogenous OBP3 protein. Our results strongly suggest that the ORG1, ORG2 and ORG3 genes are direct target genes for OBP3.
Screening for OBP3-responsive genes using suppressive subtraction hybridization
In an effort to identify induced genes in the OBP3-overexpressing plants, we used SSH. SSH was chosen for the following reasons: (i) SSH does not require very much plant tissue. This is important because the OBP3-overexpression lines are small in size and are difficult to propagate because of fertility problems. (ii) Low-abundance transcripts can be isolated with SSH. Total RNA from an OBP3-overexpression line and a wild-type line was isolated, and a cDNA library enriched for genes upregulated in an OBP3-overexpressing line was prepared. Differential screening of this library, which consisted of 198 clones with subtracted probes representing the OBP3-overexpressing line or the wild-type line, revealed a number of genes whose expression was enhanced in the OBP3 line. We initially focused on eight clones whose expression appeared to be upregulated the strongest in the OBP3 line. As shown in Table 1, these clones were derived from five independent genes (ORG1–ORG5).
Table 1. Isolation of OBP3-responsive genes (ORGs)
OBP-binding motifs (%)
This Table includes the frequency with which independent clones were isolated through suppressive subtraction hybridization (SSH), putative function of genes, gene locus, and number of OBP-binding motifs (CTTT (A/T) and (A/T)AAAG; Yanagisawa and Schmidt, 1999) located within 1-kb upstream of cDNA start site. Percentages in parentheses compare the number of OBP-binding motifs in the ORG promoter regions with the number of sites that is statistically expected from a 1-kb random sequence (less than four OBP-binding sites).
OBP3-responsive gene expression correlates with the level of OBP3 expression in OBP3-overexpression lines under constitutive or inducible promoters
The expression characteristics of the five genes were initially analysed using a virtual Northern blot analysis, and all the clones showed enhanced expression in the OBP3-overexpression background. The ORG gene induction by OBP3 was tested in three additional OBP3-overexpression lines – a strongly overexpressing OBP3 line (#1) similar to the original line used for the SSH screen (#5), an intermediate OBP3 line (#7) and a weakly overexpressing OBP3 line (#3) – using RT-PCR, as shown in Figure 1(a). All the ORG genes were induced to different degrees in the OBP3-overexpression lines that had high or intermediate levels of OBP3 transgene expression, with ORG2 and ORG4 showing the largest and the smallest induction, respectively. In contrast, in line #3 in which OBP3 RNA levels were similar to that in the wild type, and where there were no significant effects on plant growth and development, there was also no change in ORG gene expression.
The OBP3-overexpression lines used in Figure 1(a) employed the constitutive cauliflower mosaic virus (CaMV) 35S promoter to express OBP3. We also checked the expression pattern of the ORG genes following an increase in OBP3 expression from a glucocorticoid-inducible promoter system (Aoyama and Chua, 1997; Kang et al., 1999). The system consists of two components: a glucocorticoid-regulated chimeric transcription factor with the DNA-binding domain of the yeast transcription factor GAL4 called GVG, and the gene of interest transcribed from a promoter containing GAL4 DNA-binding sites. Transgenic plants containing the system only express the gene of interest following glucocorticoid treatment, as GVG is only activated upon binding glucocorticoid (for further details, see Aoyama and Chua, 1997). In Figure 1(b), expression of the ORG and OBP genes were analysed 6 h after the expression of OBP3 had been induced following treatment with dexamethasone (Dex), a chemical inducer for this system. We also examined ORG expression in a control line that contained only the empty vector. ORG1, ORG2 and ORG3 showed a large increase in expression following induction of OBP3, whereas ORG4 and ORG5 had little to no change in expression. In the empty vector line, we also observed a small induction in ORG1, ORG2 and ORG3 expression, although the level of induction was considerably less than that observed in the OBP3-inducible line.
To determine if OBP3 expression was enhanced before the increase in ORG expression, we examined earlier time points following Dex treatment. As shown in Figure 1(c), OBP3 RNA levels had already increased significantly by 1 h, continued to increase at 3 h and appeared to have peaked by 6 h. In contrast, there was no change in ORG expression at 1 h, with an increase in expression of the three ORG genes first being observed at 3 h. ORG expression increased further at 6 h, although the induction appeared to be levelling off. The OBP3 construct had been tagged with a hemagglutinin (HA) epitope and so changes in the levels of the exogenous OBP3 protein following Dex treatment could be analysed. As shown in Figure 1(d), the increase in OBP3 protein levels lagged a little behind the increase in RNA levels as expected, with little change at 1 h, but a significant increase by 3 h, when ORG RNA levels were first observed to have increased. Thus, the increase in expression of the ORG1, ORG2 and ORG3 genes follows soon after and parallels the increase in expression of OBP3, strongly suggesting that they are direct targets of OBP3.
The promoter regions of the ORG genes extending 1 kb from the cDNA start site were examined to see if they contain potential Dof-binding sites. As shown in Table 1, all five ORG genes contain multiple Dof-binding sites in this region, ranging from 10 for ORG2 and ORG3 to 5 for ORG5. The frequency of these Dof-binding sites was higher than the number that is statistically expected from a 1-kb random sequence, which would be less than 4. Interestingly, the three ORG genes (1, 2 and 3) that were significantly induced within 6 h when OBP3 was expressed off an inducible promoter, also contained the highest number of potential Dof-binding sites in the promoter region.
Classification of the OBP3-responsive genes
One of the ORG genes had been previously characterized. ORG5 is AtExt1, an extensin gene that codes for a class of hydroxyproline-rich glycoproteins that are abundant in plant cell walls (Merkouropoulos et al., 1999). ORG1–ORG4 are novel genes. While ORG1, ORG2 and ORG3 share some similarity with known genes, ORG4 does not have any similarity with genes in the database. The ORG1 gene comprises 13 exons and 12 introns and codes for a predicted protein of 670 amino acids. The ORG1 protein appears to be novel, although database searches revealed weak similarity in different regions of the predicted ORG1 protein with fibrillin proteins and some cdc2 proteins. ORG2 and ORG3 were closely related to each other as shown in Figure 2(a), and have 79.8% overall identity in their deduced amino acid sequences. Both proteins contain a bHLH domain, which is known to be a DNA-binding domain present in plant and animal transcription factors. The alignment of the bHLH domains with other bHLH proteins from different organisms is shown in Figure 2(b). The deduced amino acid sequence of the bHLH domain in the ORG2 and ORG3 genes are diverged from other plant Myc proteins. In particular, three amino acids were inserted in the second helix region, which resembles animal Myc proteins more (compare BMAL1B and MYC2 with ORG2 and ORG3 in Figure 2b).
Salicylic and jasmonic acid responsiveness of the OBP3-responsive genes
As OBP3 is an SA-responsive gene (Kang and Singh, 2000), we tested whether the novel ORG genes can also be induced by SA treatment. The SA-responsiveness of the ORG genes was analysed in aerial parts and roots separately. As shown in Figure 3, ORG1, ORG2 and ORG3, like OBP3, were SA-inducible in roots, while ORG4 did not respond to SA within the time period tested. None of the ORG genes showed significant SA responses in aerial tissue (data not shown), similar to what was seen with OBP3 (Kang and Singh, 2000).
The RNA expression patterns of OBP3 and the SA-inducible ORG genes were analysed in two different Arabidopsis mutants, cpr5-2 (Bowling et al., 1997) and cep1 (Silva et al., 1999), which constitutively express plant-defence responses. The CPR5 gene encodes a novel, putative transmembrane protein (Kirik et al., 2002), and cpr5-2 plants have elevated levels of SA, enhanced levels of PR-1, increased PDF1.2 gene expression and enhanced resistance to virulent strains of Pseudomonas syringae and Peronospora parasitica (Bowling et al., 1997). The cep1 mutant has higher levels of SA as well as higher levels of PR and PDF1.2 gene expression (Silva et al., 1999; Oñate-Sánchez and Singh, 2002). As shown in Figure 4(a), there was no difference in the expression levels of OBP3 or ORG1, ORG2 and ORG3 between the wild type and the cpr5-2 mutant, while there was a clear induction of PR5 and PDF1.2 expression, which are SA- and JA-responsive genes, respectively. While there was again no difference in the expression of OBP3 between the wild-type plants and the cep1 mutant, there was a significant reduction in ORG1, ORG2 and ORG3 expression in this mutant background.
As the high levels of PR5 and PDF1.2 expression in the cpr5-2 and cep1 mutants suggest activated SA- and JA-signalling pathways, we analysed the expression of OBP3, and ORG1, ORG2 and ORG3 after treatment with MeJA in a JA-resistant mutant, jar1-1 (Staswick et al., 1998). JAR1 appears to belong to the acyl adenylate-forming firefly luciferase superfamily (Staswick et al., 2002) and jar1-1 plants have decreased sensitivity to JA inhibition of root elongation and are more susceptible to the soil fungus Pythium irregulare (Staswick et al., 1992, 1998). As shown in Figure 4(b), the expression of both OBP3 and the ORG1, ORG2 and ORG3 genes were repressed in response to MeJA in the wild type and the jar1-1 mutant. PDF1.2 gene expression was used as a control for the MeJA treatment, and its expression was induced both in the wild type and in the jar1-1 mutant as seen previously by Oñate-Sánchez and Singh (2002).
ORG1 to ORG3 RNA levels are decreased in OBP3-silenced lines
We initiated loss-of-function experiments using gene constructs encoding ihpRNA (Wesley et al., 2001) against OBP3 to further analyse the function of OBP3 and the role it plays in regulating the expression of the ORG1, ORG2 and ORG3 genes. The ihpOBP3 construct contained 525 bp of OBP3-specific sequence as determined by blast search against the Arabidopsis genome. Fifteen independent lines were generated, and none of these lines had any visible phenotype. Seven of these lines contained single T-DNA-insertion sites as determined by segregation analysis, and we focused on these lines. T3 lines were further characterized by RT-PCR. As shown in Figure 5, the ihpOBP3 lines contained on average approximately 30% OBP3 RNA compared to the wild type. These ihpOBP3 lines also had reduced levels of ORG1, ORG2 and ORG3. Chitinase B was used as a control to show that the reduction in ORG1, ORG2 and ORG3 expression was a specific effect and not a general stress response. These results are consistent with the results obtained with the OBP3-overexpressing lines under constitutive or inducible promoters, and show that the expression of ORG1, ORG2 and ORG3 is strongly correlated with OBP3 expression.
The other three ORG genes encode proteins with similarity to other proteins in the database. ORG1 has weak similarity in different regions of the protein to fibrillin proteins, which are major components of the extracellular microfibrils, and some cdc2 proteins, which are involved in cell-cycle control. However, no clones with similarity to both cdc2 and fibrillin genes were present in the database, indicating that ORG1 may represent a novel class of genes. Although the region of highest similarity between ORG1 (230–302 amino acids) and the rice CDC2 protein (102–174 amino acids) scores 49% similarity with 30% amino acids being identical, ORG1 lacks the signature motifs of the cdc2 protein family: the PSTAIR motif (EGVPSTAIREISLLKE; 42–57 amino acids in the yeast CDC2), which binds to a cyclin, and the GEGTYG motif (GEGTYG: 11–16 amino acids in the yeast CDC2), which is in the amino-terminal ATP-binding domain. Therefore, it is not clear whether ORG1 has a cell cycle regulatory function merely based on the amino acid similarity, and further work will be required to address this question.
ORG2 and ORG3 both encode proteins that contain a bHLH domain, share approximately 80% overall identity, are located next to each other in the genome and are responsive to SA with similar kinetics to OBP3. Myc proteins with a bHLH domain have been shown to be potent inducers of both cell proliferation and apoptosis in animals (see Eisenman, 2001 and Pelengaris et al., 2000 for reviews). Plant Myc homologs appear to function in a number of different areas. For example, in maize, the Myc proteins encoded by the R-gene family including the R, Lc, B and Sn proteins regulate the anthocyanin biosynthesis pathway. Another plant Myc-like protein is the Arabidopsis RD22BP1 protein, which is induced by ABA, a stress signal in plants, and by dehydration stress (Abe et al., 1997). It is noteworthy that the deduced amino acid sequence of the bHLH domain in ORG2 and ORG3 are diverged from other plant Myc-like proteins, suggesting that the ORG2 and ORG3 DNA-binding profiles and function may be different from these proteins.
The expression patterns of the ORG1, ORG2 and ORG3 genes were very similar in all the situations examined. All three genes were coordinately induced in an OBP3-inducible line within 3 h following treatment with the chemical inducer Dex, although the absolute levels of induction differed. The increase in ORG1, ORG2 and ORG3 expression occurred shortly after OBP3 RNA levels started to increase, and closely reflected the expression levels of the exogenous OBP3 protein. The ORG1, ORG2 and ORG3 genes also showed a small induction in an empty vector line following treatment with Dex, although the levels of induction were small compared to the OBP3-inducible line. The expression of some stress-responsive genes (PDF1.2 and PR-5) has also been shown to be induced in empty vector lines following Dex treatment, although in these cases, induction was observed later, 24 or 48 h after Dex treatment (Kang et al., 1999).
The ORG1, ORG2 and ORG3 genes were also inducible by SA and downregulated by JA. The interaction between the SA and JA signalling pathways is known to be complex with both antagonistic and synergistic interactions reported (reviewed in Glazebrook, 2001), for example, a SA-responsive PR1 gene was repressed by JA (Rao et al., 2000). In another study, 8 out of 2300 genes were found to be induced by SA and repressed by JA using microarray hybridizations (Schenk et al., 2000). However, antagonistic response to these hormones is clearly not universal, as Schenk et al. (2000) also found that a significantly higher number of genes were co-induced or co-repressed by SA and JA. Interestingly, the expression of the ORG1, ORG2 and ORG3 genes was not affected in the cpr5-2 mutant and even reduced in the cep1 mutant, although both mutants are known to have elevated SA levels (Bowling et al., 1997; Silva et al., 1999).
OBP3 was also often expressed in a similar manner to the ORG1, ORG2 and ORG3 genes. One exception was in cep1, where the expression of OBP3 was not affected, while the expression of the ORG1, ORG2 and ORG3 genes was significantly reduced. Moreover, while the induction of OBP3 by SA is not reduced by cycloheximide (Kang and Singh, 2000), this is not the case with ORG1 (our unpublished observation). These results demonstrate that OBP3 is not co-regulated with the ORG1, ORG2 and ORG3 genes and are consistent with OBP3 being upstream of ORG regulation, and that other unknown factors also influence ORG-expression levels.
The transgenic lines overexpressing OBP3 under the control of the CaMV 35S promoter had a range of phenotypes, the severity of which correlated with the levels of OBP3 expression (Kang and Singh, 2000). These phenotypes included growth defects involving retarded growth in both roots and aerial parts of the plant. In the most severe cases, these growth defects led to death. At this stage, it remains unclear whether one or more of the ORGs isolated in this study contributed to any of these phenotypes. ORG1 and ORG4 encode proteins of unknown function, and ORG2 and ORG3 encode potential transcription factors but where the identity of target genes is unknown. In the case of ORG5, overexpression or antisense repression of other extensin genes did not give rise to any visible phenotypes (Bucher et al., 2002; Memelink et al., 1993).
In contrast to the OBP3-overexpressing lines, the OBP3-silenced lines had no obvious visible phenotypes, although they did lead to specific changes in gene expression. One possibility is that other proteins such as OBP1 and/or OBP2, which are also responsive to SA and have similar DNA-binding and protein–protein interaction properties as OBP3 (Kang and Singh, 2000), may be able to compensate for the reduction in OBP3 expression. It will be interesting to examine if the ORG genes are affected in other OBP-mutant backgrounds. Another possibility is that complete silencing of OBP3 was not achieved with the ihpRNA approach (Wesley et al., 2001). Although this silencing approach was shown to result in well over 80% suppression of target genes tested (Wesley et al., 2001), we have failed to obtain similarly efficient silencing lines not only for OBP3, but also for other transcription factors including TGA2 (Kang et al., unpublished results). The difficulty to get efficient silencing lines for OBP3 may be as a result of lethality caused from the loss of OBP3.
The promoter regions of the ORG genes were examined, and were found to contain a higher frequency of potential Dof-binding sites than statistically expected for a random sequence of the same size. For example, the ORG2 and ORG3 promoter regions each had ten Dof-binding sites and the ORG1 promoter region had nine sites, whereas a random sequence would be predicted to have less than four. Other gene promoters have also been shown to contain multiple Dof-binding sites, including the maize cyppdk1 promoter (Yanagisawa, 2000), the ArabidopsisGST6 promoter (Chen et al., 1996) and the potato KST1 promoter (Plesch et al., 2001). With whole-genome sequences available for some model systems, it was proposed that genes regulated by a transcription factor could be identified by a cluster of the corresponding binding sites (Wagner, 1999). Computational methods looking for a cluster of transcription-factor-binding sites more frequent than expected by chance alone revealed downstream genes for several transcription factors in Drosophila (Berman et al., 2002; Rebeiz et al., 2002). Moreover, microarray analysis of systemic acquired resistance (SAR)-induced genes in Arabidopsis showed that W-boxes, binding sites for WRKY transcription factors, are highly enriched in the promoter regions (Maleck et al., 2000). These results suggest that the clustering of binding sites for some transcription factors could be a common scheme for gene regulation. Interestingly, all three ORG genes whose expression was significantly increased in the OBP3-inducible lines, also had the highest frequency of Dof-binding sites in the promoter region. However, which of these Dof-binding sites in the ORG promoters are functional remains to be determined.
The combined results of the gain-of-function and loss-of-function experiments strongly suggest that the ORG1, ORG2 and ORG3 genes are direct target genes of OBP3. Thus, some of the direct target genes of OBP3 may encode regulatory factors. Further studies will be required to determine the function of the ORG1, ORG2 and ORG3 genes and the role they play in the growth defects observed in OBP3-overexpressing lines.
Plant materials, growth conditions and treatments
Arabidopsis thaliana (ecotype Columbia) seeds were surface sterilized with 70% ethanol for 15 min, dried on autoclaved paper, stratified at 4°C for 2 days and grown on MS (Sigma, St Louis, MO, USA) medium or on autoclaved soil. Plants were grown at 22°C with a day length of 16 h. For treatments with MeJA and SA, Arabidopsis seedlings were grown on MS solid medium for 8 days. The plants were treated in MS liquid medium containing 100 μm SA (Sigma) or 100 μm MeJA (Aldrich, Milw, WI, USA). Total seedlings or roots of the plants were collected and frozen in liquid nitrogen prior to RNA isolation. For Dex treatment, Arabidopsis seeds were geminated on MS solid medium with 20 μg ml−1 hygromycin B to select transgenic lines. After 1 week on the MS medium, seedlings were transferred to autoclaved soil and grown for two more weeks. 30-μm Dex solution in 0.01% Tween-20 was sprayed onto seedlings, which were then subjected to a vacuum treatment for 2 min. The plants were collected and frozen in liquid nitrogen prior to RNA isolation.
Preparation of double-strand cDNAs for RACE (rapid amplification of cDNA ends), virtual Northern blot, and suppression subtractive hybridization
One microgram of cDNA was denatured with 10 pmol of the following primers in 5 μl of total volume at 72°C for 2 min and cooled at 4°C for 2 min: cDNA synthesis primers (5′-AAGCAGTGGTAACAACGCAGAGTACT30N-1N-3′) and SMARTII oligonucleotide (5′-AAGACGTGGTAACAACGCAGATGTACGCGGG-3′). A mixture containing 2 μl of the supplied buffer (Gibco BRL, Rockville, MD, USA), 20 nmol of DTT, 10 nmol of dNTP and 200 units of MMLV reverse transcriptase (Gibco) in 10 μl final volume was added and incubated at 42°C for 1 h, followed by addition of 40 μl of TE (10 mm Tris (pH 7.6), 1 mm EDTA). The synthesized double-strand cDNA was then heat-denatured at 72°C for 7 min and stored at −80°C.
One microlitre of double-stranded cDNA, synthesized as described in the previous section, was amplified in a final volume of 20 μl, which contained 10 nmol of dNTP, 1× Advantage2 polymerase mix (Clontech, Palo Alto, CA, USA), the supplied buffer (Clontech) and the following primers synthesized specific to the gene of interest. This primer was designed to bind to the junction of two consecutive exons (this suppresses an amplification from genomic DNA).
In the case of 3′ RACE, 10 pmol of oligo dT18 was used and for 5′ RACE, a mixture of two primers (0.2 pmol of 5′-CTAATACGACTCACTATAGGGCAAGCAGTGGTAACAACGCAGAGT-3′ and 1 pmol of 5′-CTAATACGACTCACTATAGGGC-3′) was used. The following PCR programme was used: 5 cycles of 94°C for 5 sec and 72°C for 3 min, followed by 5 cycles of 94°C for 5 sec, 70°C for 10 sec and 72°C for 3 min, and again followed by 27 cycles of 94°C for 5 sec, 68°C for 10 sec and 72°C for 3 min. In the case of 3′ RACE, the PCR products were then subcloned into an appropriate vector and sequenced. In the case of 5′ RACE, 0.1 μl of the PCR products was subsequently amplified with a secondary PCR in a final volume of 20 μl that contained 10 nmol of dNTP, 1× Advantage2 polymerase mix (Clontech), the provided buffer (Clontech), 10 pmol of the nested primer for 5′ cDNA end (5′-AAGCAGTGGTAACAACGCAGAGT-3′) and 10 pmol of the nested primer for the gene of interest. The following PCR programme was used: 25 cycles of 94°C for 5 sec, 68°C for 10 sec and 72°C for 3 min. The PCR product was then cloned and sequenced.
One microliter of double-strand cDNA was amplified with a mixture containing 1× Advantage 2 polymerase mix (Clontech), the supplied buffer (Clontech), 20 nmol of dNTP, 20 pmol of PCR primer (5′-AAGCAGTGGTAACAACGCAGAGT-3′) in a final volume of 100 μl using the following PCR programme: 95°C for 1 min, followed by 20–25 cycles of 95°C for 5 sec, 65°C for 5 sec and 68°C for 6 min (for SSH, two identical reactions were performed to obtain enough material for the subsequent step). The amplification was stopped before the PCR reached the saturation phase: this was checked by aliquotting PCR samples at different cycles. 2 μl of 0.5M EDTA was added to terminate the PCR reactions. 10 μl of the amplified products were then loaded on a 1.5% agarose gel for a virtual Northern analysis, followed by the standard procedure for Southern blot analysis (Sambrook et al., 1989).
Two hundred microliters of amplified products from a virtual Northern blot analysis was purified using Concert Rapid PCR Purification System (Gibco) according to the manufacturer's instruction, resulting in a 50-μl TE solution. The solution was then digested with 30 units of RsaI in 300 μl final volume at 37°C for 4 h, followed by purification using Concert Rapid PCR Purification System (Gibco) according to the manufacturer's instruction, resulting in a 50-μl TE solution. The solution was then precipitated with 2 volumes of 100% ethanol and 0.1 volume of 3 m sodium acetate (pH 5.2) and re-suspended to 0.3 μg μl−1 with water. The digested double-stranded cDNA was then used for SSH, using the PCR-Select cDNA Subtraction Kit (Clontech) according to the manufacturer's instruction. The resulting products from the PCR-Select cDNA Subtraction Kit (Clontech) were subcloned into the pT-Adv vector (Advantage PCR cloning kit, Clontech). Plasmid DNAs of pT-Adv containing each product from the hybridization were prepared and used in a dot-blot analysis to screen for differentially expressed clones using a standard protocol (Sambrook et al., 1989).
Generation of hpOBP3 lines
To produce hairpin constructs, a 525-bp OBP3 fragment was amplified using the following primers: 5′-GGGGATCGATGGTACCTGATAATACTACTAGTA-3′ and 5′-GGGGTCTAGACTCGAGGCCTGTTGAGTTGTTGG-3′. The PCR product was digested with the appropriate restriction enzymes, purified and ligated into the XbaI/ClaI site and the XhoI/KpnI site of pHANNIBAL (Wesley et al., 2001). The NotI cassette was then transferred to the binary vector pART27. Arabidopsis was transformed with this construct using standard infiltration techniques and selection for kanamycin-resistance.
RNA Isolation, Northern and RT-PCR analysis
Total RNA was isolated from Arabidopsis, using the Purescript reagent (Gentra Systems, Minneapolis, MN, USA) according to the manufacturer's instruction, except for the protein precipitation step in which RNA samples were centrifuged twice instead of once. Purified RNA for RT-PCR analysis was subject to further purifications: Purified RNA was treated with 2 units of RNase-free DNase (Promega, Madison, WI, USA) at 37°C for 60 min and was re-purified again with the Purescript reagent (Gentra Systems).
Northern blot analysis and synthesis of random primed probes were performed as described earlier (Shah et al., 1999). Random primed probes were synthesized from PCR fragments from Arabidopsis cDNA with oligonucleotides used for RT-PCR. Northern gel blot hybridization was performed as described previously by Kachroo et al. (1995).
For cDNA synthesis, 1 μg of RNA in 5 μl of the total volume was denatured at 65°C for 10 min and mixed with 0.5 μg of oligo dT18, 12 nmol of dNTP, 2 units of AMV reverse transcriptase (Promega) and 10 units of RNase inhibitor (Promega) in 25 μl of the supplied buffer. The mixture was then incubated at 37°C for 1 h, denatured at 65°C for 10 min and stored at −80°C until the PCR reactions were performed.
All PCR reactions were performed with an annealing temperature of 60°C, 0.5 unit of Taq polymerase (Invitrogen, Carlsbad, CA, USA), 4 nmol of dNTP and a pair of primers (10 pmol each) in a final volume of 20 μl of the buffer recommended by the supplier. For each primer condition, the cycle was chosen as two cycles prior to visualization on an ethidium bromide-stained agarose gel. The following are the specific primer pairs used for the RT-PCR: Numbers in parentheses indicate the number of PCR cycles completed for each target gene.
The PCR products were run on a 1.5–2% agarose gel, transferred to a nylon membrane (Bio-Rad, Hercules, CA, USA) and probed with the corresponding random-primed 32P-labelled probe. Amplification products were imaged using a Storm 840 phosphoimager (Molecular Dynamics, Sunnyvale, CA, USA) or Cyclone phosphoimager (Packard, Meriden, CT, USA). imagequant software (Molecular Dynamics) with the default parameters was used for quantizing RT-PCR products.
Immunoblot was analysed as described previously using an enhanced chemiluminescence method (Amersham, Pharmacia Biotech, Piscataway, NJ, USA; Lin et al., 1995) with the following modification. The blot, a PVDF membrane (Millipore, Bedford, MA, USA), was probed with anti-HA antibody (Sigma, diluted 1/1000) and then with anti-mouse antibody conjugated to horseradish peroxidase (Sigma, diluted 1/5000).
We thank Louise Thatcher and John Klingler for helpful comments on the manuscript, Elaine Smith and Pradeep Kachroo for excellent technical help and other members of the Singh and Lin labs for helpful discussions. We also thank the Nottingham Arabidopsis Stock Centre for providing the seeds for cep1, cpr5-2 and jar1-1. H.K. was supported in part by a pre-doctoral fellowship from the Korean Ministry of Education. This work was supported in part by a grant from the US Department of Agriculture to K.B.S. (96-35100-3165).