INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Developmental responses of plants to environmental light are mediated by at least four different photoreceptor systems: phytochromes (Sage, 1992), cryptochromes (Ahmad and Cashmore, 1996a), UV-A receptors (Young et al., 1992), and UV-B receptors (Christie and Jenkins, 1996). The red/far-red light-sensing phytochromes are the best characterized of these photoreceptors (Quail et al., 1995). Phytochrome molecules exist in two spectrally distinct, photointerconvertible forms: a red light-absorbing form (Pr) and a far-red light-absorbing form (Pfr). Although it was formerly believed that Pr was biologically inactive and Pfr active, spectrophotometrically detectable levels of Pfr were found to be inconsistent with physiological responses in most studies performed during the past four decades (Furuya, 1993).

Phytochrome apoproteins are now known to be encoded by five different genes in Arabidopsis: PHYA, PHYB, PHYC (Sharrock and Quail, 1989), PHYD, and PHYE (Clack et al., 1994). Investigations of the biological functions of individual phytochromes have been helped by studies of phytochrome A-deficient (phyA) mutants (Dehesh et al., 1993; Nagatani et al., 1993; Parks and Quail, 1993; Whitelam et al., 1993), phytochrome B-deficient (phyB) mutants (Koornneef et al., 1980; Reed et al., 1993), and phytochrome D-deficient (phyD) mutants (Aukerman et al., 1997) of Arabidopsis. Recent physiological studies of these phytochrome-deficient mutants have allowed the identification of individual phytochromes that are responsible for different phytochrome effects. For example, in the photoinduction of seed germination, phytochrome A (PhyA) mediates the photo-irreversible response to very-low-fluence (VLF) light at 300 to 800 nm, while phytochrome B (PhyB) controls the low-fluence (LF) response that exhibits red/far-red reversibility (Shinomura et al., 1996). Consequently, differences in the functions of the various phytochromes are associated with differences in requirements for both fluence and wavelength.

In the past two decades, the regulation by light of gene expression in higher plants has been the focus of extensive studies, and various nuclear genes have been identified as being light regulated (Tobin and Silverthorne, 1985; Thompson and White, 1991). The regulation by phytochrome of the expression of the genes (Lhcb) for the light-harvesting chlorophyll a/b-binding protein of photosystem II (PSII) has been analyzed in detail (Silverthorne and Tobin, 1987; Tobin and Kehoe, 1994). In several plant species, expression of Lhcb genes can be induced via either a VLF or a LF response pathway (Kaufman et al., 1984; Horwitz et al., 1988; Wehmeyer et al., 1990). A recent study of the fluence and wavelength requirements for expression of the Lhcb gene in Arabidopsis demonstrated clearly that PhyA photo-irreversibly mediates the VLF response, while PhyB and the other phytochromes photoreversibly regulate the LF response that leads to induction of expression of the Lhcb gene (Hamazato et al., 1997). However, we know very little about the roles of each phytochrome in the expression of genes other than Lhcb (Reed et al., 1994; Hamazato et al., 1997), Athb-2 (Carabelli et al., 1996), and tub1 (Leu et al., 1995). Clearly, the identification of phytochromes that control the expression of particular genes is essential if we are to fully understand phytochrome-regulated gene expression. To define the phytochromes involved in gene expression, we attempted to identify entire groups of PhyA-regulated genes instead of investigating previously characterized "light-response" genes on a gene-by-gene basis.

The differential display of mRNA is a powerful tool for studying differential gene expression (Liang and Pardee, 1992). This technique has several advantages over conventional methods such as subtractive hybridization and differential hybridization. The main advantages are that multiple samples of RNA can be compared simultaneously and that both up-regulated and down-regulated genes can be detected. Nevertheless, the utility of earlier versions of the differential display technique was limited by the low reproducibility of fingerprints and the high frequency (>70%) of false positive clones (Debouck, 1995). The more recently developed fluorescent differential display (FDD) technique overcomes these problems (Ito et al., 1994). The use of modified anchor primers with a fluorescent label and an automated fluorescent DNA sequencer for detection results in high reproducibility, high throughput, and operational safety. With these tools, we were able to perform a rapid and large-scale screening for differentially expressed genes.

Using the FDD technique, we attempted a large-scale screen for Phy-regulated genes. In this study, the FDD signals obtained from phyA mutant seedlings exposed to far-red light were compared with those obtained from wild-type seedlings to identify a collection of PhyA-responsive genes. We identified 15 genes whose expression was regulated by PhyA through the VLF response. In addition, we showed that the expression of these 15 genes was also regulated by PhyB and other phytochromes through the LF response.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant Materials

The wild type and the phyA-201 (Nagatani et al., 1993) and phyA-201phyB-1 mutants (Hamazato et al., 1997) of Arabidopsis Landsberg erecta were used in all analyses. Seeds were surface-sterilized and plated on filter paper on top of 0.2× Murashige-Skoog agar medium (Murashige and Skoog, 1962) supplemented with 0.6% (w/v) Suc. Immediately after plating, seeds of the wild type and the phyA-201 were exposed to far-red light (25 µmol m² s¹) for 10 min to inhibit dark germination. After subsequent imbibition for 16 h in darkness at 23°C ± 1°C in a temperature-controlled chamber (Koito-toron FR-6113W, Koito, Tokyo), seeds were exposed to red light (30 µmol m² s¹) for 8 h to induce germination and then kept in total darkness at 23°C ± 1°C for 5 d. Seeds of the phyA-201phyB-1 double mutant were sown and incubated under continuous white light (12 W m²) for 60 h to induce germination. The germinated seedlings were kept in total darkness at 23°C ± 1°C for 3.5 d.

Light Treatment

Etiolated 6-d-old seedlings were irradiated with a pulse of far-red light or red light and then returned immediately to darkness. Four hours after the light irradiation, seedlings were harvested and quickly frozen in liquid nitrogen. Far-red light was obtained by filtering light from far-red fluorescent tubes (FL20S.FR-74, Toshiba, Tokyo) through a 3-mm-thick far-red acrylic plate (Deraglass 102, Asahikasei, Tokyo). Red light was obtained by filtering light from white fluorescent tubes (FL20SSW/18[G], Hitachi, Tokyo) through a 3-mm-thick red acrylic plate (Shinkolite A102, Mitsubishi Rayon, Tokyo). Fluence rates were measured with an optical power meter (1830-C, Newport, Irvine, CA).

Isolation of RNA and Northern-Blot Analysis

Total RNA was isolated from etiolated seedlings by the phenol/SDS/LiCl method (Verworerd et al., 1989). For northern-blot analysis, total RNA (10 µg) was subjected to electrophoresis on a 1.5% (w/v) agarose gel that contained 0.66 M formaldehyde in 20 mM MOPS buffer, and transferred to a nylon membrane (Hybond-N, Amersham-Pharmacia Biotech, Uppsala). Northern blots were hybridized with random-prime ³²P-labeled probes that corresponded to fragments of cloned cDNAs for 16 h at 45°C in hybridization buffer (50% [v/v] formamide; 5× SSPE; 5× Denhardt's solution; 50 µg/mL denatured salmon-sperm DNA; and 0.5% [w/v] SDS). After hybridization, filters were washed with 2× SSC and 0.1% (w/v) SDS for 15 min at room temperature, twice with 2× SSC and 0.1% (w/v) SDS at 55°C for 15 min, and with 0.2× SSC and 0.1% (w/v) SDS at 55°C for 30 min. Hybridization levels were quantified with an image analyzer (BAS 1000, Fujifilm, Tokyo). The amount of total RNA loaded in each lane was normalized by reference to results of hybridization with a gene for 18S rRNA from pea (Jorgensen et al., 1987). The sizes of transcripts were determined by comparison of mobilities with those of RNA standards (0.16- to 1.77-kb RNA Ladder, GIBCO-BRL, Rockville, MD). Northern-blot analysis was performed at least twice to confirm the reproducibility of the results.

FDD

FDD was performed as described previously (protocol II from Ito et al., 1994) with some modifications. Total RNA was treated with RNase-free DNase (Ambion, Austin, TX) for 30 min to remove contaminating genomic DNA. First-strand cDNAs were synthesized from each total RNA (2.5 µg) using three different Texas Red-labeled 3'-anchored oligo(dT) primers (5'-Texas Red-GT₁₅N-3', n = G, C, or A, Yukigouseikagaku, Tokyo) and a SuperScript Preamplification System (GIBCO-BRL). cDNAs produced from 25 ng of total RNA were amplified by PCR using combinations of Texas Red-labeled anchored and arbitrary 10-mer primers (kit B, D, F, and X, Operon Technologies, Alameda, CA). The conditions for PCR were as follows: 94°C for 3 min, 40°C for 5 min, and 72°C for 5 min, followed by 24 cycles of 94°C for 15 s, 40°C for 2 min, and 72°C for 1 min, with an additional extension step at 72°C for 5 min. Electrophoresis and detection of the PCR products were performed with a automated fluorescent DNA sequencer (SQ5500, Hitachi, Tokyo).

Cloning of the cDNAs of Interest

Preparative electrophoresis was performed to isolate the cDNAs of interest. The differential display patterns visualized by the fluorescent image analyzer (FMBIO II Multi-View, Takara, Shiga, Japan) were printed true to size. The gel on the lower of the two glass plates was laid over the printed image and the band of interest was excised. After excision of bands, the gel was scanned again to confirm the excision of each band. The cDNAs were eluted into distilled water by several rounds of freezing and thawing, and then reamplified by PCR with the appropriate pairs of primers. Products of reamplification were purified with a PCR purification kit (Qiagen, Chatsworth, CA) and subcloned into the pGEM-T vector (Promega, Madison, WI). For each reamplified fragment, several independent Escherichia coli colonies were chosen, and inserted fragments from these colonies were amplified by PCR. The sizes of inserted fragments were determined by comparison of mobilities with the isolated band of the original FDD samples using a fluorescent DNA sequencer. Several independent clones with inserts of the expected size were selected and sequenced. To identify contaminating cDNAs (false positive clones) that possess the same sizes but distinct sequences, individual cloned fragments were used as probes for Southern-blot analysis of the original products from FDD. The cDNA clones with the same banding patterns as those obtained by FDD analysis were selected.

Reverse Transcription (RT)-PCR Analysis

RT-PCR analysis of clones 9, 13, and 15 was performed using gene-specific primers. PCR primers used to detect mRNAs of the clones were as follows: clone 9-forward (F), 5'-GCGAAAGCTCCACAACATTCATA-3'; clone 9-reverse (R), 5'-GAAGGGATCGATCGATAAACAAT-3'; clone 13-F, 5'-TGTTTTGATTGATAATATTACACA-3'; clone 13-R, 5'-CTTTCGGACAAATCGGATT-3'; clone 15-F, 5'-ACGT- AAAGATACAAGGAGATTGA-3'; clone 15-R, 5'-GCTTT- GATGATGATGAGGAAG-3'. Total RNA was treated with RNase-free DNase (Ambion) and reverse transcribed to first-strand cDNAs using oligo(dT) primer (SuperScript Preamplification System, GIBCO-BRL). cDNAs produced from 25 ng of total RNA were used as templates in 20 µL of PCR mixture. The conditions for PCR were as follows: 20 cycles (for clone 9 and 13) or 23 cycles (for clone 15) of 94°C for 15 s, 50°C for 30 s, and 72°C for 1 min, with an additional extension step at 72°C for 5 min. The Arabidopsis actin 8 gene (ACT8) was used as a positive internal control (An et al., 1996). PCR primers for detection of ACT8 mRNAs were 5'-ATGAAGATTAAGGTCGTGGC-3' and 5'-TCCGAGTT- TGAAGAGGCTAC-3' (Aida et al., 1997). The amplified PCR products (10 µL) were electrophoresed on a 3% (w/v) agarose gel (NuSieve 3:1, FMC BioProducts, Rockland, ME), stained with ethidium bromide, and scanned using the fluorescent image analyzer.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Screening of PhyA-Regulated Genes

Six-day-old etiolated wild-type and phyA-201 mutant seedlings were irradiated with 10 mmol m² far-red light for 5 min and then kept in darkness for 4 h. Total RNA was isolated from both far-red light-treated and untreated samples, and then mRNA fingerprints were generated by PCR with arbitrary primers. Comparison of fingerprints from wild type and the phyA mutant identified bands that were differentially expressed in a PhyA-dependent manner; an example of such fingerprints is shown in Figure 1. In this figure, each lane contains more than 150 bands of cDNAs that range in size from 100 to 800 bp. The majority of bands had the same intensities in all samples (Fig. 1, lanes D/W, D/A, FR/W, and FR/A). We PCR-amplified with three different anchored oligo(dT) primers, each in combination with 80 arbitrary 10-mer primers, and screened approximately 30,000 cDNA bands. The FDD analysis of candidate PhyA-regulated cDNA bands was repeated at least twice using independently prepared samples of total RNA to confirm the differentially expressed banding patterns. The cDNA bands that were consistently associated with differential expression were isolated and analyzed in further detail.

View larger version (114K): [in this window] [in a new window]   Figure 1.   Image of a typical FDD gel. cDNAs that had been reverse transcribed from mRNAs isolated from wild-type and phyA mutant (phyA-201) etiolated seedlings of Arabidopsis with or without a prior pulse of far-red light were compared. Total RNA isolated from 6-d-old wild-type (W) and phyA201 (A) etiolated seedlings that had been kept in total darkness (D) or exposed to 10 mmol m2 far-red light (FR) was subjected to FDD analysis using 10 different primer sets (numbers above each group). The DNA size markers were included for the estimation of the length of cDNAs of interest. A differentially expressed band is indicated by the oval.

We identified a total of 24 cDNA bands that were differentially expressed between wild type and the phyA mutant (Fig. 1; Table I). Twenty-one bands increased in intensity upon far-red irradiation of wild-type but not of phyA mutant seedlings. Three bands decreased in intensity upon far-red irradiation of the wild type but not of the phyA mutant. All 24 candidate cDNAs were recovered from gels and re-amplified with the same primer set as used for the initial amplification. The re-amplified fragments were subcloned into the TA cloning vector and sequenced. The comparative sequence analysis of the cloned cDNA fragments revealed that the 24 cDNAs represent 20 distinct mRNA species. Two different mRNA species were amplified twice with different sets of primers. One mRNA species was detected as three cDNA fragments with different sizes, each amplified with the same set of primers (Table II). All cloned cDNA fragments contained sequences that corresponded to the particular oligo(dT) primers and the arbitrary 10-mer primers used for PCR (data not shown). Thus, we isolated 24 cDNA fragments by FDD, and those 24 cDNAs represent 20 distinct mRNA species.

View this table: [in this window] [in a new window]   Table I.   Results of FDD screening for PhyA-regulated genes in etiolated seedlings of Arabidopsis mRNA expression patterns of 6-d-old etiolated seedlings of wild type (WT) and phyA-201 mutant (phyA) were compared by FDD. The etiolated seedlings were kept in darkness (D) or exposed to 10 mmol m2 far-red light (FR).

PhyA-Dependent Expression of the Candidate Clones

To confirm that the cloned cDNA fragments represented the PhyA-regulated genes, we used the 20 representative clones as probes in northern-blot analysis of the total RNA that had been subjected to the initial FDD screening (Fig. 2A). RT-PCR analysis using gene-specific primers was performed for the detection of the eight cloned cDNAs that did not yield signals on northern blots (Fig. 2B). As shown in Figure 2, the abundance of mRNAs that corresponded to up-regulated cDNAs (clones 1-13) was 1.5- to 2.0-fold higher in far-red light-treated wild-type seedlings (lane 2, WT/FR) than in untreated control seedlings (lane 1, WT/D), whereas the abundance did not differ significantly between far-red light-treated samples and untreated samples in the phyA mutant (lanes 3 and 4; phyA/D and phyA/FR).

View larger version (43K): [in this window] [in a new window]   Figure 2.   Northern-blot analysis (A) and RT-PCR analysis (B) confirming the PhyA-regulated expression of the genes that correspond to 14 cloned cDNAs. A, Total RNA (10 µg per lane) from the wild type (WT) and the phyA mutant (phyA) that had been subjected to FDD was hybridized with cDNA probes generated from each cloned cDNA. The sizes of transcripts were estimated from mobilities of RNA size markers. The signal for hybridized 18S rRNA is included for normalization of results. B, cDNAs corresponding to 25 ng of total RNA were used as templates for PCR. The ACT8 gene was used as a positive internal control (An et al., 1996). The amplified PCR products were electrophoresed on a 3% (w/v) agarose gel, stained with ethidium bromide, and scanned with the fluorescent image analyzer.

The abundance of mRNAs that corresponded to two down-regulated cDNAs declined approximately 2-fold (clone 14) and 1.2-fold (clone 15), respectively (Fig. 2). The patterns of expression of mRNAs observed by northern-blot or RT-PCR analysis faithfully reflected the FDD banding patterns (Fig. 1). The PhyA-dependent induction of expression of clone 7, which was identical to the Lhcb1*3 gene (formerly named cab1) that had been analyzed previously (Hamazato et al., 1997). The results clearly demonstrate that the expression of the mRNAs that corresponded to the 15 cloned cDNAs was mediated by PhyA upon far-red irradiation of etiolated Arabidopsis seedlings. Five cDNAs that could be detected only by RT-PCR analysis did not show differential expression in the wild type compared with the phyA mutant. Finally, we identified 15 genes that were regulated by PhyA; of these, 13 were up-regulated and two were down-regulated. These 15 genes were represented by 19 cloned FDD cDNAs from etiolated seedlings of Arabidopsis (Table II).

The Up-Regulated Genes by PhyA

The sequences of the 15 isolated clones were used to search for similar sequences in the EMBL/GenBank databases. The results are summarized in Table II. The 13 up-regulated clones included nine known genes of Arabidopsis, three homologs of plant genes from other organisms, and one unknown gene. Nine known genes of Arabidopsis included such six genes for photosynthetic-proteins as Lhca1*1 (Jensen et al., 1992), Lhcb1*3 (Leutwiler et al., 1986), glyceraldehyde-3-P dehydrogenase subunit A (gapA; Shih et al., 1991), subunit K of photosystem I (PSI) (psaK; Ikeuchi et al., 1990), subunit G of PSI (psaG; Okkels et al., 1992), subunit P of PSII (psbP; Kochhar, 1996), two genes for enzymes involved in biosynthesis of chlorophyll; geranylgeranyl reductase (Keller et al., 1998), Mg chelatase (CHL H; Gibson et al., 1996), and one gene for DNA damage repair/toleration-related protein (Drt112; Pang et al., 1993).

Three other cloned cDNAs represent Arabidopsis homologs of nuclear genes for photosynthetic proteins in other plants. The 5' region (nucleotides 1-475) of clone 4 was 72.8% similar to the spinach gene (psbS) that encodes a 22-kD protein in PSII. The identity at the deduced amino acid level was 89.0% in this region (Fig. 3A; Wedel et al., 1992). The 5' region (nucleotides 1-113) of clone 5 (accession no. AB015860) was 78.7% similar to the barley gene for subunit E of PSI (psaE) at the nucleotide level and 84.2% similar at the deduced amino acid level (Fig. 3B; Okkels et al., 1988). Clone 6 (accession no. AB015861) was 75.4% similar at the nucleotide level and 86.8% similar at the deduced amino acid level to the spinach gene for subunit L of PSI (psaL; Fig. 3C; Fliger et al., 1993). Clone 13 (accession no. AB015862) was very similar (95.5%) to an Arabidopsis cDNA clone (GenBank accession no. T45262) that exhibited no significant similarity to any known genes in the database, suggesting that this clone represents a novel gene. Thus, the majority of up-regulated genes were identical to the nuclear genes for plastid proteins involved in photosynthesis and the biosynthesis of chlorophyll.

View larger version (32K): [in this window] [in a new window]   Figure 3.   The deduced amino acid sequence of three PhyA-regulated genes. A comparison of clones 4, 5, and 6 with the C-terminal regions of spinach psbS (A), barley psaE (B), and spinach psaL (C) is shown. The dashes () represent identical amino acids, the asterisk indicates a stop codon, and the numbering is as previously described (Okkels et al., 1988; Wedel et al., 1992; Fliger et al., 1993).

The Down-Regulated Genes by PhyA

Two down-regulated clones showed significant similarity to Arabidopsis genes. Clone 14 showed 98.9% similarity to a xyloglucan endotransglycosylase-related gene (XTR7; Xu et al., 1996). The sequence of clone 15 was 95.9% identical to a part of a deduced open reading frame in the Arabidopsis genome database (GenBank accession no. AB005234). This open reading frame is predicted to be a novel member (ASK3) of the ASK protein kinase family (Park et al., 1993). In contrast to the up-regulated genes, the two down-regulated genes encode proteins functionally distinct from plastid proteins.

Overlapping Effects of Phytochromes on Expression of mRNAs

To examine whether phytochromes other than PhyA might be involved in regulation of the expression of the above-described 15 distinct genes, we irradiated 6-d-old etiolated seedlings of the wild type, the phyA mutant, and a phyAphyB double mutant with 1 mmol m² red light, with 1 mmol m² red light followed by 3 mmol m² far-red light, or with 3 mmol m² far-red light. Total RNA was isolated from the seedlings 4 h after the light treatments, and the expression of mRNAs that correspond to the 15 isolated genes was investigated by northern-blot or RT-PCR analysis. Representative results of the expression of nine selected genes (Lhca1*1, Drt112, gapA, psbS, psaE, psaL XTR7, clone 13, and ASK3) are shown in Figure 4.

View larger version (49K): [in this window] [in a new window]   Figure 4.   Northern-blot analysis (A) and RT-PCR analysis (B) of red/far-red reversible effects on expression of genes that correspond to nine selected genes in the wild-type, the phyA mutant, and phyAphyB double mutant seedlings. Etiolated seedlings of the wild type (WT), the phyA-201 mutant (phyA), and the phyA-201phyB-1 double mutant (phyAphyB) were kept in total darkness (D) or exposed to red light (R; 1 mmol m2), or to red light followed by far-red light (R/FR; 1 mmol m2/3 mmol m2), or to far-red light only (FR; 3 mmol m2). A, Total RNA (10 µg per lane) was allowed to hybridize with cDNA probes generated from each cloned cDNA. The signal for hybridized 18S rRNA is included for normalization of results. B, cDNAs corresponding to 25 ng of total RNA were used as templates for PCR. The ACT8 gene was used as a positive internal control (An et al., 1996). The amplified PCR products were electrophoresed on a 3% (w/v) agarose gel, stained with ethidium bromide, and scanned with the fluorescent image analyzer.

In the phyA mutant, the expression of the selected genes was induced or repressed by 1 mmol m² red light (Fig. 4, lanes phyA, D and R) and the extent of induction (1.6- to 3-fold) or repression (4-fold for XTR7, 1.25-fold for ASK3) was reduced by subsequent irradiation with far-red light (Fig. 4, lane phyA, R/FR). The results indicate that, in addition to PhyA, other phytochromes photoreversibly regulate the expression of these mRNAs. Moreover, changes in expression of the selected genes by red light and red/far-red reversibility of those changes were also observed when we examined the phyAphyB double mutant (Fig. 4, lanes phyAphyB, D, R, and R/FR). The extent of changes of induction (1.2- to 1.5-fold) or repression (1.3-fold for XTR7, 1.2-fold for ASK3) by red light in the phyAphyB double mutant (Fig. 4, lanes phyAphyB, R) was lower than that in the phyA mutant (Fig. 4, lanes phyA, R). The five remaining up-regulated genes (geranylgeranyl reductase, CHL H, psaK, psaG, and psbP) also showed similar changes of expression (data not shown). The function of phytochromes other than PhyA on induction of expression of the Lhcb1*3 gene (clone 7) has been well studied previously (Hamazato et al., 1997). The results presented here indicate that PhyB and phytochromes other than PhyA and PhyB also photoreversibly regulate the expression of the 15 genes from this study in etiolated seedlings of Arabidopsis.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Identification of PhyA-Regulated Genes by Large-Scale Screening

In the present study, we screened approximately 30,000 cDNAs by FDD using 240 combinations of primers. If the total number of genes in Arabidopsis is somewhere between 16,000 and 33,000 (Gibson and Somerville, 1993; Meyerowitz, 1994) and if about 15% of all of the genes are expressed at a particular time in a eukaryotic cell (Liang and Pardee, 1992), we would expect 2,500 to 5,000 mRNAs to be transcribed at any particular time in any particular Arabidopsis tissue. For the display of all mRNAs in the present FDD system, about 60 combinations of primers should be sufficient for detection of each Arabidopsis mRNA, at least once with 95% probability (Bauer et al., 1993). We used four times the number of primer sets than should be required to display all expressed species of mRNA in Arabidopsis. The sensitivity of FDD is similar to that of the original differential display technique (Ito et al., 1994). Wan et al. demonstrated that both abundant and rare mRNAs are identified by differential display (Wan et al., 1996). Therefore, we probably screened the majority of mRNA species that are expressed in etiolated Arabidopsis seedlings 4 h after irradiation with far-red light, since both abundant and rare mRNAs should have been displayed in our FDD screening.

We identified 19 differentially displayed bands of cDNA that correspond to 15 PhyA-regulated genes, 13 are up-regulated and two are down-regulated (Table II). The 19 differentially expressed bands represent a very small fraction of the total number of bands displayed, indicating that the present FDD method is a highly reliable technique for detection of rare differentially expressed genes among genes with only a few differences in their profiles of expression. In addition, the Lhcb1*3 gene (clone 7; Hamazato et al., 1997) and the gapA gene (clone 3; Dewdney et al., 1993), which have already been well characterized as VLF response genes in Arabidopsis, were identified as PhyA-regulated gene by our FDD screening. In spite of using 240 primer sets in present study, however, we failed to detect the Fd gene that was previously reported as a VLF-inducible gene (Caspar and Quail, 1993). Therefore, the present large-scale FDD screening would not cover all of the expressed mRNA in etiolated Arabidopsis seedlings.

Assignment of Phytochromes to the Control of Expression of Specific Light-Dependent Genes

The expression of PhyA-regulated genes such as Lhcb1*3 is induced upon far-red irradiation of etiolated wild-type plants but not of phyA mutant plants (Hamazato et al., 1997). We compared levels of RNAs in samples of wild-type and phyA mutant seedlings that had or had not been irradiated with far-red light (Fig. 1; Table I). We were able to eliminate from consideration those faint bands that were generated sporadically by PCR, and confirmed the differentially expressed bands of cDNA as representative of PhyA-regulated genes.

The high irradiance response to far-red light (FR-HIR) is known to be a PhyA-mediated response, as is the VLF response (Smith and Whitelam, 1990). Accumulation of Lhcb1*3 transcripts in etiolated Arabidopsis seedlings requires a fluence of far-red light from 1 µmol m² to 10 mmol m² (Hamazato et al., 1997). Therefore, we used high-fluence far-red light (10 mmol m²) for light treatment to induce maximum changes in the abundance of mRNAs of candidate genes. However, the fluence and the duration of far-red irradiation that we used were insufficient to induce the FR-HIR. Thus, the isolated genes were exclusively VLF response genes that are regulated by PhyA in etiolated seedlings of Arabidopsis. The combination of the FDD technique with utilization of the phyA mutant and far-red light irradiation was the key to the successful large-scale screening for PhyA-regulated genes.

The Role of PhyA in Gene Expression in Etiolated Seedlings

Twelve of the 13 up-regulated genes were identical to known nuclear genes for photosynthetic or chloroplast proteins (Table II), even though the expression of only three of the genes, Arabidopsis Lhcb1*3, Arabidopsis gapA, and spinach psbS, had previously been reported to be regulated through VLF response pathway (Dewdney et al., 1993; Adamska et al., 1996; Hamazato et al., 1997). Regulation by light of the expression of the psaL gene of cucumber and the geranylgeranyl reductase gene of Arabidopsis has been reported, but it is not yet clear whether phytochromes are involved in those phenomena (Toyama et al., 1996; Keller et al., 1998). In addition, regulation by phytochrome of the seven other genes in Arabidopsis (Lhca1*1, Drt112, psaE, CHL H, psaK, psaG, and psbP) has not previously been reported. Accordingly, the present study allows, for the first time to our knowledge, the assignment of PhyA to the regulation of the induction of nine genes for photosynthetic or chloroplast proteins in etiolated seedlings of Arabidopsis.

To date, most genes for plastid proteins have been shown to exhibit the LF response, while a limited number of genes exhibit the VLF response in etiolated tissues (Thompson and White, 1991). Among previously reported VLF-responsive genes other than the Lhcb genes, only the gene for the early light-inducible protein (ELIP; Adamska et al., 1995) and the psbS gene encode plastid proteins (Adamska et al., 1996). However, our results show clearly that the expression of genes for other plastid proteins, in addition to the Lhcb1*1 genes, is also controlled by the PhyA-mediated VLF response in etiolated seedlings of Arabidopsis.

The expression of the Drt112 gene, whose product may be involved in ctDNA resistance to photodamage (Pang et al., 1993), was also induced by PhyA through the VLF response. It has been proposed that the PsbS protein, which is also the product of a VLF response gene, plays a significant role in protection against light stress (Adamska et al., 1995). These results suggest that the genes not only for photosynthetic proteins but also for photodamage-related proteins are included among the genes for plastid proteins that exhibit the VLF response.

Induction of expression of the genes for photosynthetic proteins and photodamage-related proteins by PhyA through the VLF response suggests the biological significance of PhyA in etiolated tissue. VLF light conditions might be present below the surface of the soil during seed germination. Plants need to rapidly produce both the components of photosystems for the effective absorption of light energy and the proteins required for protection against light stress prior to emergence of the seedling from the soil. Thus, the elevated photosensitivity of PhyA would be advantageous for the survival of germinating seeds or young seedlings.

Although most of the genes reported previously as being phytochrome regulated showed up-regulation of expression, a few genes, including genes for PHYA (Lissemore et al., 1988), NPR1 (Okubara et al., 1991), TUB1 (Leu et al., 1995), and ATHB2 (Carabelli et al., 1996), are down-regulated by phytochromes. Photoregulated expression of the XTR7 gene has also been reported, but the photoreceptor(s) involved in this regulation remains unknown (Xu et al., 1996). Expression analysis in this study clearly demonstrated that down-regulated expression of the XTR7 gene is controlled by PhyA through the VLF response in etiolated seedlings of Arabidopsis. Although five of the seven members of XTR gene family show down-regulated expression by light in green plants of Arabidopsis (Xu et al., 1996), only XTR7 was identified as down-regulated gene by FDD screening in etiolated seedlings. This result suggests that XTR7 might be a major xyloglucan endotransglycosylase mediating cell wall alterations required for elongation of etiolated plants.

Clone 15 corresponds to a novel member (ASK3) of the Arabidopsis ASK kinase family (ASK1 and ASK2) which contain highly acidic domains at the C terminus (Park et al., 1993). Expression of genes for ASK1 and ASK2 kinases was induced by light and was highest in the leaf (Park et al., 1993). In contrast to these two genes, clone 15 showed down-regulated expression that was mediated by PhyA. Although the tissue specific expression of clone 15 was not investigated in this study, the difference in the regulation of expression by light might suggest specific roles of this novel ASK kinase in etiolated plants. The identification of genes which had been previously unreported as PhyA-regulated genes such as XTR7, ASK3, and the novel gene (clone 13) suggests a new aspect of the role of PhyA in etiolated tissue.

Overlapping Effects of Different Phytochromes

The expression of all 15 PhyA-regulated genes characterized in this study was also regulated by PhyB and phytochromes other than PhyA and PhyB (Fig. 4). In other words, none of the identified genes was regulated exclusively in a PhyA-specific manner. Although photoperception by PhyA and PhyB is quite different in terms of both fluence and wavelength, both regulated expression of the same genes. This result is consistent with results of previous studies of the phytochrome regulation of Lhcb1*3 gene expression (Reed et al., 1994; Anderson et al., 1997; Hamazato et al., 1997). Moreover, Cerdán et al. (1997) reported that a 146-bp fragment of the tobacco Lhcb1*2 promoter responds to all three modes of phytochrome action (VLF response, LF response, and HIR) in etiolated transgenic tobacco seedlings. Overlapping effects of phytochromes have also been observed in several phytochrome-related responses, such as regulation of the photoinduction of seed germination (Shinomura et al., 1996), hypocotyl elongation, promotion of cotyledon expansion, and flowering (Reed et al., 1994). Thus, the overlapping effects of phytochromes might be a general phenomenon, suggesting that the various signal transduction pathways converge at the same places.

Some recently isolated mutants exhibit specific alterations in response to either PhyA or PhyB. For example, fhy1, fhy3 (Whitelam et al., 1993; Barnes et al., 1996), spa1 (Hoecker et al., 1998), vlf1, and vlf2 (Yanovsky et al., 1997), in the PhyA response, and pef2, pef3 (Ahmad and Cashmore, 1996b), and red1 (Wagner et al., 1997), in the case of the PhyB response. The existence of such mutants suggests the existence of PhyA- and PhyB-specific signal transduction pathways. Thus, at least some events in PhyA and PhyB signal transduction appear to be distinct and specific to either PhyA or PhyB. To uncover the mechanisms of the overlapping effects of each phytochrome, further analysis of the signal transduction pathways of phytochrome-regulated responses is obviously necessary. The identification of a group of PhyA-regulated genes provides useful molecular markers for the study of signal transduction pathways from photoperception by phytochromes to the regulation of gene expression.