If, following an inductive treatment of 2 d of continuous darkness, shoot apices of Pharbitis nil are cultured 1 d later on White's medium supplemented with 2% sucrose, they cannot form carpels, but they can if they are cultured on 2% glucose. It was hypothesized that the differential effect of these sugars was because of differential expression of carpel-specific genes. Partial cDNA homologues to the Arabidopsis genes, LEAFY (PnLFY), AGAMOUS (PnAG1/2), and CRABS CLAWS (PnCRC1/2) were cloned. PnLFY was expressed in the shoot apex 1 d following the start of induction and remained higher than in non-induced apices for a further 6 d before exhibiting a major peak of expression on day 7. Peaks of expression of PnAG1 and PnAG2 spanned days 7–11, coinciding with the appearance of stamens and then carpels. The Pharbitis ‘PnCRC2’ showed greatest homology to Arabidopsis YABBY2 (PnYABBY). Its expression peaked on day 8 when the carpels first appeared. ‘PnCRC1’ showed greatest homology to Arabidopsis FILAMENTOUS (PnFIL). Its expression was approximately the same in inductive and non-inductive treatments. Apart from PnFIL these partial cDNAs could be used as markers to test the hypothesis concerning differential effects of sucrose and glucose. Cultured shoot apices from induced plants were sampled at weekly intervals. All four genes were expressed more strongly in the glucose compared with the sucrose treatment, most notably at day 17. A more intensive sampling (days 15–19) indicated that PnLFY and PnYABBY exhibited much higher expression on glucose compared with sucrose, most notably on days 15–16 and days 18–19.
This paper is available online free of all access charges (see http://jxb.oupjournals.org/open_access.html for further details)
Florally induced shoot apical meristems (SAMs) are determined if they form flowers, even when returned to non-inductive conditions (McDaniel, 1978). To test for floral determination in the short day (SD) plant, Pharbitis nil, it was exposed to an inductive treatment, and the shoot apex was cultured under non-inductive conditions for 28 d and its floral state recorded (Durdan et al., 2000). Surprisingly, different carbon sources added to the culture medium elicited differential effects on whorl determination; SAMs cultured on 2% sucrose did not form carpels, but they did on 2% glucose, suggesting that glucose is required for carpel-specific gene expression (Durdan et al., 2000).
Studies on Arabidopsis thaliana and Antirrhinum majus have resulted in substantial progress in identifying genes that are crucial to the change from a vegetative to a floral shoot apical meristem. The published data for SAMs of these species, portray a cascade of protein–DNA interactions with meristem identity genes at the top, and organ identity genes some way down which, in turn, must activate substrates that, in most cases, have yet to be resolved (Araki, 2001). A consensus view is that the meristem identity genes, LEAFY (LFY) in Arabidopsis and FLORICAULA (FLO) in Antirrhinum are crucial for committing the SAM to floral morphogenesis (Coen et al., 1990; Weigel et al., 1992; Schultz and Haughn, 1993). LFY expression has been detected in the earliest cells destined to become part of the floral meristem but not in the inflorescence meristem (Weigel et al., 1992). However, LFY expression does occur in the apices of young vegetative meristems, dropping to lower levels over time until there is a strong increase during floral induction (Hempel et al., 1997), or gradually increasing in young vegetative meristems until a level conferring floral competence is reached (Blázquez et al., 1997). As the floral meristem grows, LFY expression is confined first to sepal primordia and then to petal, stamen, and carpel primordia during the early growth of these organs (Weigel et al., 1992).
LFY protein activates both meristem and organ identity genes. One of the earliest targets for LFY is AGAMOUS, a so-called ‘C’ class gene. AG RNA is first detected in cells destined to become stamens and carpels in early floral organ development and remains restricted to these organs (Yanofsky et al., 1990; Drews et al., 1991). AG expression is evenly distributed throughout stamens and carpels until late in organ development. It is then restricted to the filament, connective tissue, and anther walls of stamens, the ovule endothelium and stigmatic tissue of carpels and the nectaries. In addition to AG, which is necessary for both stamen and carpel development, the homeotic genes SUPERMAN (SUP), AGAMOUS-LIKE 5 (AGL5), CRABS CLAW (CRC), and SPATULA (SPT) contribute specifically to carpel determination (Bowman et al., 1992; Savidge et al., 1995; Alvarez and Smyth, 1999). The CRC gene was cloned using chromosome walking. It is a member of the YABBY family of genes. There are at least five other YABBY family genes in A. thaliana and all have a zinc finger domain and a helix-loop-helix domain in line with a role as transcription factors. In situ hybridization shows that CRC expression is confined to flowers, with strong expression in carpels and nectaries (Bowman and Smyth, 1999; Siegfried et al., 1999; Heisler et al., 2001).
In Arabidopsis, the LFY protein functions not only in the switch to flowering, but it also works in co-operation with the protein encoded by the stem cell identity gene, WUCHSEL, to localize the expression domain of AG. This occurs until AG protein initiates a negative feedback loop which down-regulates WUS and this signals the completion of floral morphogenesis (Lohmann et al., 2001). AG expression is necessary for carpel wall differentiation and style growth and SPT for transmitting tract development (Alvarez and Smyth, 1997). Both AG and SPT are involved in producing curvature of the carpel wall. However, while CRC expression controls the growth of the carpel wall, SPT promotes growth at the point of carpel fusion (Alvarez and Smyth, 1997).
In P. nil, the differential effect of sucrose and glucose in specifying the carpel whorl (Durdan et al., 2000) led to a search for a LFY homologue and homologues to the C class gene, AG, together with the carpel-specific AGL5 and CRABS CLAW. Unlike Arabidopsis, Pharbitis nil has a qualitative photoperiodic requirement with a critical dark period of >11 h (Vince-Prue and Gressel, 1985). Previous data of Durdan et al. (2000) were consistent with the transport of a floral stimulus from the cotyledons to the SAM 12–16 h following the start of induction. Given this precise timing, the first aim of the work reported here was to test the extent to which the expression of putative floral genes in P. nil is tightly regulated temporally. It is also known exactly when each floral whorl is morphologically visible in this variety of P. nil at specific days following the start of induction: bracts (day 4), sepals (day 6) petals/stamens (day 7), and carpels (day 8) (Durdan, 1998). Hence a second aim was to examine the expression pattern of each gene in vivo in relation to the morphogenesis of each whorl. The third aim was to test the hypothesis that the effect of glucose, but not sucrose, in specifying carpel formation in culture (Durdan et al., 2000), is through differential expression of carpel-specific genes.
Here, the cloning of partial cDNAs from P. nil that are homologous to part of the open reading frames of the A. thaliana genes: LFY, AG, and CRC are reported. The data are consistent with a tightly-regulated temporal expression of first PnLFY and later, PnAG1/2 coinciding with the appearance of stamens and then carpels whereas PnCRC2 (renamed PnYABBY) expression coincides exclusively with carpel morphogenesis. Moreover, compared with 2% sucrose, 2% glucose does indeed stimulate a substantially higher level of expression of PnLFY, Pn AG1/2, and PnYABBY in cultured shoot apices from induced plants.
Materials and methods
Seeds of Pharbitis nil Choisy cv. Violet (Muritane Seed Company, Kyoto) were selected for uniformity, stirred in 16 M sulphuric acid for 30 min, washed in running water at 27±3 °C (temperature used throughout unless stated otherwise) for 2 h, imbibed in aerated distilled water for 24 h and allowed to germinate in darkness between damp filter papers for 24 h. Seedlings with radicles of ≥1.5 cm were selected and planted in John Innes No.1 compost and exposed to 5 d of continuous light (CL) provided by cool white fluorescent tubes (∼250 μmol m−2 s−1).
For in vivo analysis, P. nil seeds were germinated and the seedlings grown in CL for 5 d followed by continuous darkness (CD) for 2 d (Fig. 1). In this system, 2 d of CD is necessary to induce flowering in the terminal apex and all axillary buds. Control seedlings were left in non-inductive CL throughout. Every 24 h from the start of the dark inductive period for 14 d, 20 shoot apices were excised from the induced plants and 20 from non-induced plants (Fig. 1). Each apex comprised the SAM, 4–5 unexpanded primordia, and 10 mm of basal tissue. Samples were flash-frozen in liquid nitrogen and stored at −80 °C. Two separate crops of samples were prepared.
For in vitro analysis further induced apices and controls were excised 3 d following the start of induction and grown in White's (1943) medium supplemented with 111 mM (2%) glucose or 58 mM (2%) sucrose. At weekly intervals from the start of culture, for 4 weeks, 20 apices from induced plants and 20 from non-induced plants were sampled from the glucose and sucrose treatments, flash-frozen in liquid nitrogen, and stored at −80 °C (Fig. 1). For reasons explained below, a follow-up in vitro experiment was undertaken whereby apices were sampled from the sucrose and glucose treatments on days 15–19 following the start of induction (Fig. 1).
Genomic DNA extraction, RNA extraction, and cDNA synthesis
Genomic DNA was extracted from young leaves according to Edwards et al. (1991). For RNA extraction, 20 apices (0.2 g), were disrupted by grinding in liquid nitrogen, and extracted using 2 ml of TRI REAGENT (Sigma) as per the manufacturer's instructions. The final pellet was resuspended in 50 μl of sterile water and stored at −80 °C. Total RNA was DNAse-treated with RQ1 DNase kit (Promega) and first-strand cDNA was synthesized using M-MLV RNase H− Reverse Transcriptase (Promega). The complete removal of genomic DNA was verified by RT-PCR using intron-spanning primers to β-tubulin: PnTUB1 F: CAATGAAGCACTGTATGATA and PnTUB1 R: GTGTAAACGCGGGAAGG derived from a cloned PnTUB PCR fragment, obtained by RT-PCR (see below). A band of 184 bp would be expected from cDNA, while an additional band at 292 bp would indicate contamination by genomic DNA.
Cloning of partial cDNAs from P. nil using degenerate PCR primers
A 343 bp portion of a P. nil β-tubulin gene was amplified using degenerate tubulin primers: TUBGEN F: GAATGCHGAYGAGTGYATG and TUBGEN R: CGGCGCRAABCCSACCAT PCR and HotStar Taq DNA Polymerase (QIAGEN) from P. nil genomic DNA. PCR conditions were: 1 cycle of 95 °C (15 min), 55 °C (1 min), 72 °C (1 min), followed by 33 cycles of 95 °C (1 min), 55 °C (1 min), 72 °C (1 min), and 1 cycle of 95 °C (1 min), 55 °C (1 min), and 72 °C (5 min).
Amplified products were gel purified using a QIAquick PCR purification kit (QIAGEN), ligated into pGEM-T Easy Vector (Promega), and sequenced in both directions using an automated ABI3100 sequencer. The nucleotide sequences reported here will appear in the EMBL, GenBank, and DDBJ Nucleotide Sequence Databases under the following Accession numbers: PnLFY: (AJ633042), PnAG1 (AJ633043), PnAG2 (AJ633044), PnFIL (AJ633045), PnYABBY (AJ633046), and PnTUB1 (AJ633047).
Degenerate primers to amplify partial P. nil cDNAs homologous to floral homeotic genes were designed by alignment of amino acid sequences from available plant species and checked using PrimerSelect (DNASTAR inc.). Primers PnAG F: CGACAAGTNACHTTYTGCAA, and PnAG R: GTGACGGTBMRATGYGG designed to AG/AGL5, amplified a 233 bp portion of AG/AGL5. PnCRC F: GTGACG GTBMRATGYGG and PnCRC R: GCCCAATTCTTNGCNGC amplify a 368 bp region from CRC. Existing degenerate primers G1201: CGGAATTCATGMGIC AYTAYGTICATYGYTAYGC and G1204: TTGGATCCIYKIGTIGGIACR TACCAWAT (Pouteau et al., 1997) were used to amplify a 233 bp portion of LFY. Partial cDNAs were amplified as above from mature carpel or mature stamen cDNA. Intron-spanning β-tubulin primers and minus-RT controls were used to ensure that the cDNA was free from genomic DNA contamination. PCR products were cloned and sequenced as above and sequences compared to the databases using BLAST. Sequences were compared with each other using MEGALIGN CLUSTAL (DNASTAR Inc.).
Semi-quantitative RT-PCR analyses
Specific P. nil primers were designed from two divergent PnAG clones: PnAG1 (PnAG1 F: CGCCGTAATGGCCTTCTC and PnAG1 R: TTTGTGGAATCAGAATGATGT), PnAG2 (PnAG2 F: CGCAGAAATGGGTTGCTG and PnAG2 R: AGTGGAATCTGAGCATGCC), two divergent PnCRC clones: PnCRC1 (PnCRC1 F: CAACCAGCTTCACCTTCC and PnCRC1 R: CTTGATGAAGCGGTTGTAAG), PnCRC2 (PnCRC2 F: GAAGCAACAACAGTCCAACA and PnCRC2 R: GCTTCCCGGTGGCTAATAT), and PnLFY1 (PnLFY1 F: CTGCATTGCCTTGACGAG and PnLFY1 R: GGGGGTGGGCATTGAAG). The specificity of each of the five primer pairs was confirmed by cloning and sequencing their PCR products.
The optimum number of PCR cycles for semi-quantitative RT-PCR was determined for each of the five specific primer pairs by amplifying a set of standards over a range of cycles. In preliminary experiments, standards were produced from dilutions of a cDNA which produced high levels of product with each of the primer sets. Product quantity was estimated from ethidium-bromide stained agarose gels using the Gene Genius Bioimaging System (Syngene Ltd.) to measure band intensity. Optimum cycles gave a steep linear plot of band strength against concentration, with an intercept close to zero and were: PnAG1 F, PnAG1 R and PnAG2 F, PnAG2 R: 34 cycles, PnCRC1 F, PnCRC1 R and PnCRC2 F, PnCRC2 R: 32 cycles, and PnLFY1 F, PnLFY1 R and PnTUB1 F, PnTUB1 R: 30 cycles.
All PCRs were repeated to provide duplicate results and a set of the standards were added in each PCR run. The raw measurement of gene expression, gel band strength, was converted to a percentage of the full strength standard and further normalized either by assuming that the β-tubulin gene or total RNA was expressed equally in all samples. In other words, each point on Fig. 3 is the mean of eight nested replicates (crop×2, pcrs×2, standardization×2).
A partial cDNA for a P. nil β-tubulin cDNA was cloned and used for normalization of expression
PCR products of ∼350 bp were produced from genomic DNA using degenerate TUBGEN primers. The portion cloned includes an intron of 188 bp in the same position as found in other plant β-tubulin genes. The coding region at the DNA level was 76% homologous to β-tubulin in Daucus carota and 74% homologous to Zinnia elegans confirming that a portion of a P. nil β-tubulin gene had been cloned. These sequences were used to design specific P. nil primers (PnTUB1).
The in vivo expression of PnTUB1 was generally similar in induced and non-induced apices, with a drift downwards in level of expression with time, although there was a peak in expression in the induced apices at day 4 (Fig. 2).
Isolation of partial cDNA clones of LEAFY, AGAMOUS and CRABS CLAW
A ∼200 bp PCR product was obtained from P. nil apex cDNA with G1201/1204 (LEAFY) primers. The region amplified corresponds to approximately half of the C2 conserved region in LFY and was most similar to the tobacco (U16174) LFY homologue (94% at the amino acid level) strongly indicating that PnLFY corresponds to a homologue of LFY.
The AG degenerate primers were designed to amplify the last 40 of the 56 amino acid residues of the MADS domain, the complete 34 residue I region and the first three residues of the K region of A. thaliana AG and AGL5 proteins. Sequences from cloned PnAG PCR products also fell into one of two groups designated PnAG1 and PnAG2, based on sequence homology. There was at least a 97% homology within the PnAG1 and PnAG2 groups at the DNA level so that they are much more similar to each other than the PnCRC groups (81% similar at the DNA level and 93% similar at the protein level). PnAG1 was most similar to the petunia AG homologue (BAB79434) (96% homology) and PnAG2 was most similar to the tobacco AG homologue (Q43585 ) (96% homology). An alignment of the PnAG sequences with the equivalent region from the entire Arabidopsis thaliana gene family, clearly shows that the Pharbitis sequences are most closely homologous to AG itself, AGL5, and AGL1 (data not shown). Specific primers to the two groups were designed, checked for specificity, and used for separate analysis of expression. However, due to the high levels of homology in this region it was not possible to assign homology to a specific gene family member without extensive analysis of the P. nil gene family structure.
Sequences from the originally designated PnCRC cloned PCR products fell into one of two groups originally designated PnCRC1 and PnCRC2. The two groups were only 46% homologous at the DNA level, 54–55% at the protein level, while within groups the homology was over 97%. This suggests that the two groups may represent different members of the CRC gene family and thus their expression was studied separately by designing specific primers which were checked by cloning and sequencing PCR products.
The CRC degenerate primers were designed to amplify the last eight residues of the CRC zinc finger domain, the whole of the central region, and most of the helix-loop-helix domain. Overall the homology to Arabidopsis CRC sequences was moderate; the closest homologue to PnCRC1 was FIL (62% homology) and the closest homologue to PnCRC2 was YABBY2 (58% homology). On this basis, and from this point on, PnCRC1 has been redesignated as PnFIL and PnCRC2 as PnYABBY.
PnLFY is expressed soon after the start of induction and peaks before expression of putative carpel-specific genes
RT-PCRs on RNA extracted from the apices of intact induced and non-induced plants were undertaken over the 14 d following the start of induction. PnLFY1 exhibited a clear peak of expression at day 7 following the start of induction. This represented a 4-fold increase compared with non-induced controls in CL (Fig. 3a). The data also show a significant 2-fold increase in PnLFY1 expression compared with controls on days 1–6 but not days 8–14 (Fig. 3a). In other words, in induced meristems, increased PnLFY1 expression started early and very soon after the known arrival of the floral stimulus at the SAM (see Introduction). This pattern occurred at similar magnitudes, in two different batches, all replicates and whichever correction method was applied. However, although PnLFY is clearly up-regulated in response to floral induction there is a constant, albeit low level, of expression in non-induced vegetative apices (Fig. 3a).
Peaks of PnAG1, PnAG2, and PnYABBY expression are tightly regulated temporally and occur after the peak of PnLFY expression
Having identified primers amplifying putative P. nil carpel identity genes, it was now possible to determine whether the timing of expression of these genes in vivo was consistent with their putative function. This would underpin their use as markers to explore the differential effects of different carbohydrate sources in vitro.
At 8 and 9 d following the start of induction, PnAG1 expression in induced apices was approximately 5–8-fold higher than expression in non-induced apices; the peak spanned 7–10 d (Fig. 3b). The expression level was consistently low in induced and non-induced apices both before day 7 and after day 10. Expression patterns were similar in all replicates and whichever correction method was applied. The pattern of expression of PnAG2 was similar to that of PnAG1, except that it peaked on day 10 (4-fold higher than in non-induced apices) and spanned days 7–11 (Fig. 3c).
The expression of the two PnCRC genes, was markedly different. There was little difference in the expression pattern of PnFIL between induced and non-induced apices (data not shown), suggesting that this gene is not a major player in the transcriptional regulation of floral development of the apex. However, PnYABBY expression in induced apices clearly peaked on day 8 being 3-fold higher in the induced compared with the non-induced plants (Fig. 3d). Expression levels were similar in induced and non-induced apices until day 7 and from day 10 onwards. Again expression patterns were similar in both crops, in all replicates and whichever correction method was applied.
In summary, three of the putative floral gene homologues in vivo, PnAG1, PnAG2, and PnYABBY, all peaked in expression after the LFY homologue, as would be expected if they are homologous to meristem/organ identity genes characterized in Arabidopsis and other species.
Floral genes were more highly expressed when apices were cultured on glucose
As stated earlier, shoot apices of P. nil, cultured on 2% sucrose instead of 2% glucose do not form carpels in culture. Hence the hypothesis that glucose, but not sucrose, induces the expression of carpel-specific genes, becomes eminently testable given the strong temporal relationship between the expression of PnAG1/2, PnYABBY, and the appearance of carpels in vivo (Fig. 3; Durdan, 1998).
As apices grow more slowly in culture it was difficult to decide which times corresponded to when the floral genes exhibited peaks of expression in vivo (between days 0 and 12 following start of induction). However, previous work on cultured apices on 2% sucrose established that 20, 40, and 60% of apices exhibited the outer three whorls following 2, 3, and 4 weeks in culture, respectively (Durdan, 1998). It was therefore decided to sample apices 1, 2, 3, or 4 weeks following the start of culture which corresponded to days 10, 17, 24, and 31 following the start of induction (Fig. 1).
Expression levels for all five gene homologues were followed using semi-quantitative RT-PCR. Non-induced apices cultured on sucrose or glucose did not exhibit major peaks of floral gene expression and their expression levels were consistently lower than for apices cultured from induced plants (data not shown, see Parfitt, 2003). Each bar in Fig. 4 is the mean of two replicates, which precluded a statistical analyses. Instead, the interpretation of clear differences in expression between treatments is made only when error bars do not overlap.
PnLFY1 expression on glucose was 2-fold higher than expression on sucrose on day 31 (4 weeks in culture) (Fig. 4a). The lack of overlap between error bars on day 10 and day 17 may also represent differences between treatments, as might the somewhat higher bar in the sucrose treatment on day 24 (Fig. 4a). PnAG1 expression was higher in induced apices in the glucose than the sucrose treatment on days 10, 17, and 31, the most substantial increase (4-fold) occurring on day 17 (Fig. 4b). A remarkably similar pattern of differences between treatments was evident for PnAG2 (Fig. 4c). PnYABBY expression was 2-fold higher in the glucose than in the sucrose treatment on days 10 and 17, but not on days 24 and 31 (Fig. 4d).
As in the in vivo experiments, there was little difference between expression levels of PnFIL in any of the treatments in vitro (data not shown). Thus the expression of this gene is consistent in vitro with its in vivo expression, acting as a useful experimental control for observed differences in expression of the other genes.
Clearly, the once per week sampling intervals were much less frequent than for the in vivo expression data. Given that in vivo, peaks of expression lasted only between 1 and 8 d, closer sampling times were called for. The in vitro data gave peaks of expression after 2 weeks in culture (day 17 following the start of induction) for PnAG1, PnYABBY, and on day 17 PnAG2 and PnLFY exhibited their second highest peak of expression on glucose compared with sucrose treatments. Thus it was decided to repeat the experiment, but to have daily sampling times before, during, and after day 17. Here, the expression pattern of PnLFY to mark the onset of flowering is reported together with PnYABBY to mark the formation of carpels. Given the small number of replicates (n=3) in these experiments, a P value of 0.1 is taken as the cut-off for significance; values of ≥0.05 to ≤0.1 are interpreted as ‘marginally’ significant differences and ≤0.05 as a significant difference between treatments. Expression of PnLFY was between 2–6-fold higher in the glucose compared with the sucrose treatment on day 15 (P=0.029) and day 16 (P=0.060) (Fig. 5a). However, the data were not significantly different between treatments on day 17 (P=0.30). On day 18 there was a significant difference in favour of the glucose treatment (P=0.020) which was reversed on day 19 in favour of sucrose (P=0.019) (Fig. 5a). PnYABBY expression was significantly higher in the glucose treatment on day 15 (P=0.03) and 19 (P=0.050). The data for days 16 and 17 indicated a marginally significant increase in the glucose treatment (P=0.088) and a reversal of this on day 18 (P=0.76) (Fig. 5b). However, there are some discrepancies between the two sets of in vitro data. For example, PnLFY expression is significantly higher in the glucose compared with the sucrose treatment on day 17 of the first in vitro experiment (Fig. 4); there is no difference in the second in vitro experiment (Fig. 5a). Given that apices grow much more slowly in vitro, it seems most likely that there were differential differences in growth/development of cultured apices in the first compared with the second in vitro experiment. Hence, it was difference in expression levels on glucose compared with sucrose that was being looked for and the data demonstrate significantly higher levels of expression on glucose compared with sucrose on two or three days out of five days for both PnLFY (Fig. 5a) and PnYABBY (Fig. 5b).
Five partial sequences homologous to A. thaliana floral homeotic genes were cloned in P. nil, four of which showed a good fit with both onset of flowering at the SAM and with subsequent initiation and appearance of the stamen and carpel whorls in vivo (Fig. 6).
The initial increase in PnLFY1 on day 1 occurs within 8–12 h of the arrival of the floral stimulus at the SAM (Vince-Prue and Gressel, 1985; Durdan et al., 2000) and the continued high expression between day 1 and 6 occurs during the subsequent switch to floral morphogenesis. Notably the major peak in PnLFY expression occurred on day 7, following the completion of the sepal whorl and coincided with the appearance of petals and stamens, but immediately before the morphological appearance of carpels (Fig. 6). Thus the peak in PnLFY1 expression on day 7 may represent a prompt for subsequent peaks in expression of PnAG1, PnAG2, and PnYABBY and the formation of carpels in the way that the LFY protein regulates the transcription of AG in Arabidopsis (Huala and Sussex, 1992). Note that the peaks of expression for PnAG1/2 span day 7 through to day 11, coinciding with the appearance of stamens and then carpels (Fig. 6). This pattern would be consistent with the function of a ‘C’ type organ identity gene with overlapping functions in the two inner whorls (Coen and Meyerowitz, 1991). By contrast, the sharp peak of PnYABBY expression on day 8 coincides with the first appearance of carpels, which would be consistent with the activity of a carpel-specific gene.
Although expression levels of PnAG2, PnYABBY, and PnLFY1 were much higher in induced apices, their expression was detected by RT-PCR in non-induced ones albeit at very low levels. Thus, if their expression is transcriptionally controlled during floral development, then their effects on flowering must be quantitative/threshold, rather than qualitative, effects. In this respect, the expression data are more comparable to what is known in Arabidopsis than Antirrhinum (Weigel et al., 1992; Hempel et al., 1997; Blazquez et al., 1997; Bowman and Smyth, 1999; Yanofsky et al., 1990; Drews et al., 1991; Smyth et al., 1990). However, in these papers, A. thaliana levels of gene expression have only been reported as ‘high’, ‘medium’, ‘low’ or none. Later, high expression levels of AG, AGL5, and CRC are restricted to certain cell types. The A. thaliana developmental timings are reported in terms of age of flower, because the timing of the arrival of the floral stimulus is generally not known as precisely as it is in P. nil. The expression data for P. nil fit into a tight temporal sequence (Fig. 6): commitment to floral morphogenesis prior to the appearance of floral organs, otherwise known as floral evocation (Evans, 1971) (days 1–6 PnLFY↑,); and a possible cue for expression of PnAG1/2 and PnYABBY (day 7 PnLFY↑), onset of stamen and carpel morphogenesis (days 7–10, PnAG1↑, PnAG2↑), and onset of carpel morphogenesis (day 8, PnYABBY↑).
Apices cultured on 2% glucose and sampled at weekly intervals first exhibited the carpel whorl 3 weeks following the start of culture (Parfitt, 2003) which means that this whorl began to be initiated at least a few days before. Although the timing of whorl morphogenesis was much slower than in vivo, PnLFY1, PnAG1, PnAG2, and PnYABBY, were all expressed more strongly in induced apices in the glucose compared with the sucrose treatments most notably on day 17, four days before the first sighting of carpels in vitro.
The patterns of gene expression found at weekly intervals inevitably left open the validity of peaks over comparatively long sampling times. Hence, the second in vitro experiment reported here comprised more intensive daily sampling between days 15–19, that encompassed the peak of activity of PnAG1 and PnYABBY, and to a lesser extent, PnAG2 and PnLFY on day 17 of the original culture experiment. The data show that PnLFY and PnYABBY were more highly expressed on apices cultured on glucose compared with sucrose on at least two or three of the five samplings. Note that the significant, albeit occasional, higher level of PnLFY expression in the sucrose treatment in both in vitro experiments may fit because of LFY's role in activating other floral genes (Weigel and Nilsson, 1995).
Shoot apices cultured from induced plants could not form carpels if cultured on 2% sucrose even if the apices were left for 10 weeks in culture (Durdan, 1998). This may be attributed, at least partly, to low levels expression of PnLFY and PnYABBY and, based on less frequent sampling, for the putative C class genes, PnAG1/2. However, carpels do form on apices cultured on 2% glucose (Durdan, 1998; Durdan et al., 2000; Parfitt, 2003) and this is consistent with much higher expression levels of these putative floral genes on glucose, most notably, PnYABBY.
In conclusion, the data are the first of their kind for floral gene expression in shoot apices of Pharbitis nil. They show an expression pattern for PnLFY that fits with the timing of floral evocation and, subsequently, a peak which could be a cue for the expression of putative class C (PnAG1, PnAG2) and carpel-specific genes (PnYABBY). Finally, an earlier discovery in P. nil of the induction of carpel morphogenesis in cultured shoot apices by glucose, but not sucrose (Durdan et al. 2000), can now be interpreted in relation to the increased expression levels of PnLFY and PnYABBY, and based on fewer data, PnAG1 and PnAG2 in apices cultured on glucose instead of sucrose. Indeed, glucose may be participating in a signalling mechanism that up-regulates the transcription of LFY and carpel-specific genes.
D Parfitt thanks the University of Wales, and University College Worcester for a research studentship.