SEARCH

SEARCH BY CITATION

Keywords:

  • Brassicaceae;
  • Cardamine;
  • cleistogamy;
  • cross-species microarray;
  • ecological genomics;
  • environmental conditions;
  • evolutionary mechanism;
  • gene expression;
  • heterologous microarray;
  • phenotypic plasticity

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • 1
    Cleistogamy is a mating system found in approximately 300 species of flowering plants. Cleistogamous plants produce closed (cleistogamous, CL) flowers that require obligate self-pollination and open (chasmogamous, CH) flowers that allow for outcross-pollination. CL and CH flowers are induced by various environmental factors; this appears to be an adaptive mating strategy in unpredictable environments.
  • 2
    We examined the molecular basis of CL and CH flowering in Cardamine kokaiensis, which is closely related to the model flowering plant Arabidopsis thaliana. CL and CH flowering should be regulated by gene expression that is dependent on environmental conditions. By elucidating the molecular basis of CL and CH flowering, we can determine the changes in gene regulatory networks involved in the transition from CH to CL flowering. Furthermore, these results may help clarify the molecular evolutionary mechanisms leading to cleistogamy.
  • 3
    We regulated CL and CH flowering of C. kokaiensis using chilling treatments in a growth chamber. In a control treatment without chilling, C. kokaiensis produced CH and intermediate (INT) flowers. Long chilling of seedlings led to INT flowers, while long chilling of seeds induced the formation of CL flowers. Chilling seeds, and to a lesser extent of seedlings, induced early flowering and small plant size at flowering.
  • 4
    We also conducted a cross-species microarray analysis to compare gene expression patterns between CL and CH flowers using genomic DNA-based probe-selection strategy in an A. thaliana microarray. In this result, 69 genes, including genes related to floral development, auxin, flowering time, cold-stress, and drought-stress, were differentially expressed between CL and CH flowers.
  • 5
    Synthesis. This is the first report on the molecular basis of cleistogamy. We hypothesize that the interaction between the genetic network of the chilling response and that of floral development has been important in the evolution of cleistogamy in C. kokaiensis. Our results help to clarify the molecular basis for the evolution of plant mating systems that depend on environmental conditions.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Cleistogamy is a mating system in flowering plants that has evolved in approximately 300 species in 60 families (Lord 1981; Campbell et al. 1983). Cleistogamous plants produce two types of flowers: closed (cleistogamous, CL) flowers that require obligate self-pollination and open (chasmogamous, CH) flowers that allow for outcross-pollination. Cleistogamy functions in reproductive compensation (Schemske 1978), in the avoidance of sibling competition as proposed in the near and far dispersal hypothesis (Schoen & Lloyd 1984), and in the avoidance of geitonogamous selfing (Masuda et al. 2001). These functions are attributed to CL and CH flowering dependent on environmental conditions.

Cleistogamous (CL) and CH flowers can be induced by various environmental factors, including light intensity (Simpson et al. 1985; Bell & Quinn 1987; Trapp & Hendrix 1988; Kawano et al. 1990; Le Corff 1993; Mattila & Salonen 1995), photoperiod (Lord 1982; Kenworthy et al. 1989), water availability (Bell & Quinn 1987), and nutrient availability (Le Corff 1993; Mattila & Salonen 1995). For example, under high nutrient and high light conditions, Calathea micans plants have more CH flowers than under the opposite conditions; this appears to be an adaptive mating strategy under unpredictable environmental conditions (Le Corff 1993).

What is the molecular basis of cleistogamy? CL and CH flowering, an example of phenotypic plasticity and polyphenism, should be regulated by changes in gene expression that are dependent on environmental conditions. By elucidating the molecular basis of CL and CH flowering, we can examine changes in gene regulatory networks involved in the transition from CH to CL flowering. Furthermore, such studies will clarify the molecular evolutionary mechanisms of cleistogamy. However, the molecular basis of CL and CH flowering has not been reported, and the clarification of flowering mechanisms under different environmental conditions is challenging. Thus, to elucidate the molecular basis of cleistogamy, differences in gene expression patterns relating to CL and CH flowering under different environmental conditions should be examined.

Microarray analysis is a feasible technique to detect differences in gene expression patterns. It has been used for ecological and evolutionary studies (reviewed by Lee & Mitchell-Olds 2006; Ranz & Machado 2006; van Straalen & Roelofs 2006; Kammenga et al. 2007; Ouborg & Vriezen 2007) such as competition among plants of Solanum nigrum (Schmidt & Baldwin 2006), host–parasite interactions (e.g. host Carpodacus mexicanus and parasite Mycoplasma gallisepticum; Wang et al. 2006), and hybrid speciation in Helianthus (Lai et al. 2006). Here, we used a cross-species microarray hybridization of target RNA and probes from different species to investigate the molecular basis of CL and CH flowering. While changes in gene expression and DNA sequences with low signal intensity were, until recently, difficult to distinguish, this problem has been solved through the development of a calibration method that uses genomic DNA (gDNA) hybridization (Bar-Or et al. 2007; Buckley 2007). Hammond et al. (2005, 2006) hybridized gDNA from Brassica oleraceae and two Thlaspi species to an Arabidopsis thaliana Affymetrix high-density oligonucleotide microarray (GeneChip ATH1), selected a probe that enabled gene expression analysis, and analysed patterns caused by phosphorus stress and zinc hyperaccumulation, respectively.

Cardamine (Brassicaceae) contains at least 200 species distributed in various environments worldwide (Lihováet al. 2004, 2006). The genus is closely related to the genomic model species A. thaliana and diverged from a lineage containing A. thaliana 13–19 million years ago (Koch et al. 2001). The A. thaliana genome has been sequenced almost completely (The Arabidopsis Genome Initiative 2000, The Arabidopsis Information Resource <http://www.arabidopsis.org/>), and A. thaliana microarray systems are commercially available. Furthermore, nucleotide sequences of previously studied genes are well conserved between Cardamine and A. thaliana (Koch et al. 2000, 2001; Hay & Tsiantis 2006). Therefore, A. thaliana microarray systems are suitable for a cross-species microarray for expression analysis in Cardamine.

The annual cleistogamous herb Cardamine kokaiensis Yahara is an endemic plant along the Kokai River in Japan. We regulated its CL and CH flowering under specific environmental conditions. Here, we report the molecular basis of CL and CH flowering in response to environmental conditions. We examined (i) the effects of chilling on CL and CH flowering and (ii) the differences in gene expression patterns between CL and CH flowering by cross-species microarray analysis using an A. thaliana Affymetrix high-density oligonucleotide microarray (GeneChip ATH1). We discuss the molecular basis for the evolution of cleistogamy. Our results help to clarify the molecular basis for the evolution of plant mating systems that depend on environmental conditions.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

plant materials

Cardamine kokaiensis Yahara (Brassicaceae) is an annual cleistogamous herb that grows only near the Kokai River, Ibaraki Prefecture, Japan (36°1′58″N, 140°0′27″E). This plant was discovered by T. Yahara (Kyushu University, Japan) and will be described under this species name elsewhere (Yahara, personal communication). The water level of the Kokai River increases from autumn to early winter and decreases from mid-winter to spring. Thus, seeds are soaked in water and begin to germinate when the water level decreases. Individuals that germinate early (i.e. in January and February) have short soaking periods and long growth periods, become larger plants, and produce CL and CH flowers. Individuals that germinate later (i.e. in March and April) have long soaking periods and short growth periods, become smaller plants, and produce only CL flowers. Seeds are likely vernalized by soaking. CH flowers have four sepals, four petals, two lateral stamens, four medial stamens, and one pistil (Fig. S1 in Supplementary Material). Lateral stamens are shorter than medial stamens. These flowers, which are morphologically similar to A. thaliana flowers, have the capacity for both self- and outcross-pollination. In contrast, CL flowers lack petals and lateral stamens (Supplementary Fig. S1) and are obligatory self-pollinated.

In April 2003, C. kokaiensis plants were collected from the Kokai River. After the plants were self-pollinated and cultivated for four generations, single descendant seeds were used in experiments. Self-pollinated seeds from both CH and CL flowers were fertile (Morinaga et al. unpubl. data).

chilling experiments

Because cleistogamy is often related to plant size, we focused on the environmental conditions that influence changes in plant size. Cardamine flexuosa is a closely related species to C. kokaiensis. The size of C. flexuosa plants is affected by different chilling treatments, which was interpreted as a vernalization effect (Kudoh et al. 1995, 1996).

Therefore, we conducted chilling treatments to examine the phenotypic responses of floral traits and plant size of C. kokaiensis under seven growth conditions: chilling treatments after germination for 14, 28, or 56 days (0–14, 0–28, and 0–56 treatments, respectively), chilling treatments before germination for 14, 28, or 56 days (14–0, 28–0, and 56–0 treatments, respectively), and no chilling (0–0 control).

Plants were cultivated in a growth chamber at 22 °C under 16 h light and 8 h darkness. Chilling was performed at 4 °C under 24 h darkness. Five plants that germinated from single descendant seeds self-pollinated for four generations were used for each experiment. The seeds were sown on medium containing 0.46% Murashige and Skoog Plant Salt Mixture (Wako Pure Chemical Industries, Osaka, Japan), 0.1% Gamborg's vitamin solution 1000× (Sigma-Aldrich, St. Louis, MO, USA), 1% sucrose, and 0.8% agar. Each seedling was transplanted to a 125-mL pot containing a 2 : 1 mixture of vermiculite and perlite. The plants were watered and fertilized (HYPONeX, Hyponex, Marysville, OH, USA) every 3–4 days.

The first and fifteenth flowers of the first flowering inflorescence of each plant were sampled. One sample of sepal, petal, lateral stamen, medial stamen, and pistil was randomly collected from each flower, and their length measured to the nearest 0.1 mm under a stereomicroscope. The number of days to flowering without chilling treatment was recorded. The dry weight of each plant was measured to the nearest 1 mg after drying in an oven at 60 °C for 72 h when its fifteenth flower had opened. We examined differences in the length of floral organs, days to flowering, and plant dry mass between treatments using one-way analysis of variance (anova) and post hoc Tukey's test. These statistical analyses were conducted using stat view 5.0 (SAS Institute Inc., Cary, NC, USA).

genomic dna hybridization and probe selection

We used a genomic DNA (gDNA)-based probe-selection strategy (Hammond et al. 2005) to check for the quality of the cross-species microarray. Cardamine kokaiensis gDNA was hybridized to a GeneChip A. thaliana ATH1 array (Affymetrix, Santa Clara, CA, USA), in which 22 810 probe-sets were scanned and at least 11 probe-pairs (i.e. perfect-match and mismatch probes) were allocated for each probe-set. Probe-pairs were selected for gene expression analysis based on C. kokaiensis gDNA hybridization intensity thresholds ranging from 0 (no probe selection) to 300 according to Hammond et al. (2005) (Supplementary Appendix S1). In the resulting optimal probe-selection strategy for gene expression analysis of C. kokaiensis, we used gDNA hybridization intensity thresholds of 50, 75, and 100 (Supplementary Appendix S1).

cross-species microarray experiments

We investigated the gene expression patterns among CH, intermediate (INT), and CL flowers. Three chilling treatments were performed at 4 °C in darkness: before germination for 14 days or after germination for 14 or 28 days (14–0, 0–14, and 0–28 treatments, respectively). The first flowers of 14–0, 0–14, and 0–28 plants produced CH, INT, and CL flowers, respectively. We performed two biological replications per treatment. Biological replications were technically replicated once per treatment. For each replication, RNA was extracted from 160 plants that were germinated from single descendant seeds self-pollinated for four generations. The first inflorescence with flower primordia was collected from each plant when the first flower bud was visible. We placed the 160 inflorescence of the 160 plants in a same microtube. Total RNA was extracted from six samples (i.e. two biological replications of three treatments) using Isogen (Nippon Gene, Tokyo, Japan). The quality and quantity of total RNA were checked using an RNA Nano Labchip and 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA).

Double-strand cDNA synthesis, cRNA synthesis, labelling, hybridization, and scanning for Affymetrix ATH1 arrays were performed according to the manufacturer's instructions <http://www.affymetrix.com/support/technical/manual/expression_manual.affx>. CEL files (CL1, CL2, INT1, INT2, CH1, and CH2.cel, accession no. GSE9799 in GEO) were generated using a Microarray Analysis Suite. This experiment was performed on two different days: the CL1 and CH1.cel were generated on 26 November 2004 and the others were generated on 22 March 2005. CEL files were loaded into GeneSpring (Agilent Technologies, USA) analysis software package using the robust multichip average (RMA) pre-normalization algorithm (Irizarry et al. 2003). During CEL file loading, calculation of intensities for each probe-set, and pre-normalization, CEL files were interpreted using data files generated from gDNA hybridization intensity thresholds of 50, 75, and 100 that were considered an optimal range for probe-selection for gene expression analysis based on gDNA hybridization experiment (Appendix S1). We performed per-chip normalization, that is, the intensity of each probe-set was divided by the median intensity of all genes in the array. Genes with different signal intensities between CL and CH flowers were analysed with one-way anova and fold-change using GeneSpring. Furthermore, when selecting for high and low values of differentially expressed genes in CL and CH flowers, we used strict fold-change to consider the effects of experiment day (26 November 2004 and 22 March 2005) in signal intensity. The strict fold-change was considered the lowest value of either ratio obtained from both the days.

semi-quantitative reverse transcription polymerase chain reaction (rt-pcr)

The double-stranded cDNA performed in the cross-species microarray experiments was also used for semi-quantitative RT-PCR to validate the microarray results. Primers for RT-PCR of 13 genes differentially and not differentially expressed were designed based on exon regions that were well conserved among A. thaliana and other organisms using Primer 3 <http://biotools.umassmed.edu/bioapps/primer3_www.cgi>. Identities of the amplified PCR products of seven genes were confirmed by sequencing using an ABI Prism 3130xl sequencer (Applied Biosystems, Foster City, CA, USA). Cardamine kokaiensis orthologues of A. thaliana DRM1, SPL5, AT4G29190, HSP81-4, NMT1, RD21, and ACTIN2 genes were amplified by primer sets (Table 1). Each semi-quantitative RT-PCR was carried out using a GeneAmp PCR system 9700 (Applied Biosystems) and EX Taq HS (TaKaRa, Shiga, Japan) in a 25-µL total volume containing 40 ng of double-strand cDNA. The PCR program consisted of 1 cycle at 94 °C for 5 min followed by 27 (ACTIN2 ortholog gene) or 29 (other genes) cycles of 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s. PCR products were not saturated under these conditions. PCR products were electrophoresed in 1% agarose gel and detected by ethidium bromide staining. Analyses of expression intensities were performed using Gel Analyzer in ImageJ <http://rsb.info.nih.gov/ij/>. The relative gene expression was calculated as:

Table 1.  Primer sets for RT-PCR in C. kokaiensis
Orthologous genesForward primerReverse primer
DRM1ACGACTCCAGGATCGGTGACCGAACAGTAGAATTATTATTACAAGA
SPL5CACAATGAAAGGAGGAGGAAGCATAAGTCTTTTACATTTGCGGATA
AT4G29190GCGTTTCAGCTCTCGATTTCGGGTACGGGTCGGATTTTAT
HSP81-4CCTTGGGTTTGTCAAGGGTACCTTCCTTGGTTGCAGAGAC
NMT1CAATGGTCATCGCATCATGTCTCCCGTCACGATCTTGTTT
RD21AGCATTCCGCTGATCTGACTTGGAATAGAAACGGGGTTCA
ACTIN2TTGTTGGTAGGCCAAGACATCGGCATGAGGAAGAGAGAAACC

Relative gene expression (%) = (objective gene expression)/(ACTIN2 ortholog gene expression). The relative gene expression in the cross-species microarray was also calculated based on the gDNA hybridization intensity threshold of 100.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

effects of chilling on cl and ch flowering

CL and CH flowering was affected by chilling (Fig. 1; Table 2). Of the first flowers, the control (0–0) and the shorter chilling treatments after germination (0–14 and 0–28) induced complete CH flowers (Fig. 1a). In contrast, longer chilling treatments before germination (28–0 and 56–0) induced complete CL flowers (Fig. 1b). Furthermore, the shortest chilling treatments before germination (14–0) and longest chilling treatments after germination (0–56) treatments induced intermediate (INT) flowers (Fig. 1a,b) that did not open and formed incomplete petals and lateral stamens. Of the fifteenth flowers, the longer chilling treatments before germination (28–0 and 56–0) induced complete CL flowers (Fig. 1d), whereas the other treatments (0–0, 0–14, 0–28, 0–56, and 14–0) induced INT flowers (Fig. 1c,d).

image

Figure 1. Lengths of floral organs (mean ± SE), including sepals, petals, lateral stamens, medial stamens, and pistils, under different chilling treatments. Results for the first and fifteenth flowers are provided under the different chilling treatments after germination (a and c, respectively) and before germination (b and d, respectively). Results of statistical analyses are reported in Table 2.

Download figure to PowerPoint

Table 2.  Differences in length of floral organs (mm, mean ± SE) among chilling treatments with one-way anova and post hoc Tukey's test
 First flower
Chilling treatment after germinationChilling treatment before germination
0–00–140–280–56F valueP value0–014–028–056–0F valueP value
Sepal3.9 ± 0.22a3.5 ± 0.02ab3.5 ± 0.07ab3.4 ± 0.08b 3.2470.04963.9 ± 0.22ab3.9 ± 0.11a3.3 ± 0.18bc3.1 ± 0.06c  6.9680.0033
Petal4.8 ± 0.38a4.6 ± 0.09a4.2 ± 0.13a1.2 ± 0.79b14.4220.00014.8 ± 0.38a2.4 ± 1.00b0.0 ± 0.00c0.0 ± 0.00c 18.3590.0001
Lateral stamen3.0 ± 0.25a2.9 ± 0.04a2.6 ± 0.04a1.0 ± 0.58b 8.9760.00103.0 ± 0.25a2.6 ± 0.1.0ab0.0 ± 0.00b0.0 ± 0.00b142.4850.0001
Medial stamen4.1 ± 0.32a3.7 ± 0.04ab3.6 ± 0.08ab3.3 ± 0.07b 3.7710.03204.1 ± 0.32a3.9 ± 0.11ab3.1 ± 0.16bc3.0 ± 0.06c  8.1290.0016
Pistil4.1 ± 0.31a3.8 ± 0.06ab3.6 ± 0.06ab3.3 ± 0.09b 3.3270.04644.1 ± 0.31a3.9 ± 0.11ab3.2 ± 0.18bc3.0 ± 0.07c  7.1210.0030
 Fifteenth flower
Chilling treatment after germinationChilling treatment before germination
0–00–140–280–56F valueP value0–014–028–056–0F valueP value
  1. Different letters indicate significant differences among treatments in each organ.

Sepal3.4 ± 0.123.3 ± 0.033.2 ± 0.053.3 ± 0.081.1990.34193.4 ± 0.12a3.3 ± 0.14a3.2 ± 0.03a2.8 ± 0.05b 6.8080.0036
Petal2.1 ± 0.67ab3.1 ± 0.37a1.0 ± 0.63bc0.0 ± 0.00c7.2590.00272.1 ± 0.67a1.0 ± 0.70ab0.0 ± 0.00b0.0 ± 0.00b 4.3670.0199
Lateral stamen1.8 ± 0.33ab2.6 ± 0.04a2.0 ± 0.04ab0.9 ± 0.52b5.3590.00951.8 ± 0.33a1.6 ± 0.44a0.0 ± 0.00b0.0 ± 0.00b13.2940.0001
Medial stamen3.3 ± 0.123.3 ± 0.033.0 ± 0.063.2 ± 0.082.7340.07803.3 ± 0.12a3.2 ± 0.13a3.1 ± 0.03ab2.7 ± 0.05b 7.7040.0021
Pistil3.4 ± 0.133.3 ± 0.023.1 ± 0.043.2 ± 0.082.6200.08653.4 ± 0.13a3.2 ± 0.13a3.2 ± 0.03ab2.8 ± 0.03b 7.4130.0025

The days to flowering were also affected by chilling (Fig. 2). Chilling also induced early flowering. Chilling treatments before germination affected the days to flowering more strongly than did chilling treatments after germination. Early flowering resulted in low plant dry mass (Fig. 3).

image

Figure 2. Number of days to flowering without chilling periods (mean ± SE) under different chilling treatments after germination (a) and before germination (b). Different letters indicate significant differences among treatments by one-way anova and post hoc Tukey's tests.

Download figure to PowerPoint

image

Figure 3. Plant dry mass (mean ± SE) under different chilling treatments after germination (a) and before germination (b). Different letters indicate significant differences among treatments by one-way anova and post hoc Tukey's tests.

Download figure to PowerPoint

gene expression patterns in cl and ch flowering

We analysed the gene expression patterns in CL and CH flowering C. kokaiensis based on the results of the gDNA hybridization, in which the optimal probe-selection ranged from an intensity threshold of 50 to 100. Hence, the criterion for our analysis was based on a 1.5-fold-change due to the low number of genes that were more than twofold differentially expressed. The number of genes differentially expressed in at least 2 of the 50, 75, and 100 thresholds of gDNA hybridization intensity surpassed the number of genes that showed greater than 1.5-fold-change by 69 (0.30% of 22,746 genes); this was considered an effect of two sets of experimental days (Tables 3, 4). Expression patterns of the 69 genes among CL, INT, and CH flowers are represented in Fig. 4. Although the fold-change in gene expression in INT flowers was not analysed, most of the signal intensities of these flowers were within the range of the CL and CH flowers (Friedman-tests, P < 0.0001; Wilcoxon-tests, P < 0.0001 of all combinations; Fig. 5a,b). Most of the 69 genes consistently responded throughout the different experimental days (Tables 3, 4). In contrast, the gene expression patterns of ZIP4 and IRT3 orthologues greatly varied between the two experiments (Table 4).

Table 3.  Genes differentially expressed more than 1.5-fold less in CL flowers than in CH flowers, which were included in at least two of 50, 75, and 100 thresholds
AGI code*GeneChip IDSignal intensity (mean ± SD)Fold-changeDescription§**Threshold
CLINTCH
  • *

    Arabidopsis Genome Initiative code (The Arabidopsis Genome Initiative 2000).

  • Normalized signal intensities of CL, INT, and CH flowers in largest threshold in which each was included. See ).

  • Fold-change between CL and CH flowers was defined the lower value of either two sets of experiments.

  • §

    Description of gene name and putative function in Arabidopsis thaliana genes by Affymetrix, NCBI, and TAIR are given for the listed genes.

  • Threshold in which each gene was included.

  • **

    Description of rubric ‘expressed’ genes in floral organ in A. thaliana according to ATTED-II (Obayashi et al. 2007; http://www.atted.bio.titech.ac.jp/) are shown in the bracket.

AT1G28330245668_at3.29 ± 0.046.12 ± 1.4211.72 ± 0.063.56Auxin-associated, dormancy-associated (DRM1)100, 75, 50
AT5G10140250476_at1.67 ± 0.277.75 ± 0.165.06 ± 0.773.04FLOWERING LOCUS C (FLC) [young flower]100, 75, 50
AT2G46830266719_at0.68 ± 0.271.81 ± 0.021.71 ± 0.832.30MYB-related transcription factor (CCA1)100, 75
AT1G69490256300_at0.44 ± 0.020.67 ± 0.021.00 ± 0.082.22NAC-LIKE, ACTIVATED BY AP3/PI (NAP) [petal, stamen]100, 75
AT4G35750253163_at1.38 ± 0.223.11 ± 0.402.85 ± 0.472.06Rho-GTPase-activating protein-related100, 75, 50
AT2G19810266695_at11.76 ± 1.1922.39 ± 2.9123.96 ± 0.552.04CCCH-type zinc-finger transcription factor [inflorescence]100, 75, 50
AT1G56220256225_at17.21 ± 4.8438.61 ± 1.4234.90 ± 8.852.03Auxin-associated100, 75, 50
AT1G80920261901_at18.58 ± 4.9626.08 ± 0.1536.98 ± 8.431.99J8-like protein similar to DnaJ100, 75, 50
AT4G04630255285_at4.75 ± 0.116.32 ± 0.559.93 ± 1.001.97unknown [inflorescence]100, 75
AT4G10940254945_at1.35 ± 0.621.46 ± 0.042.64 ± 1.191.95PHD finger family100, 75
AT1G52250257504_at5.91 ± 1.719.22 ± 0.6011.75 ± 2.951.94Dynein light chain type 1 family100, 75
AT4G29190253722_at10.99 ± 0.2519.03 ± 0.3121.26 ± 2.221.93CCCH-type zinc-finger transcription factor100, 75, 50
AT1G58290256020_at0.95 ± 0.511.82 ± 1.001.89 ± 0.781.87Glutamyl-tRNA reductase (HEMA1)100, 75
AT1G56280256226_at24.01 ± 0.9242.22 ± 10.7544.69 ± 5.911.86Drought-induced 19100, 75, 50
AT1G20880262804_at1.40 ± 0.192.56 ± 0.422.71 ± 0.211.86RNA recognition motif containing [petal, stamen]100, 50
AT3G15270257051_at1.20 ± 0.092.23 ± 0.092.23 ± 0.191.85SQUAMOSA promoter binding protein-like 5 (SPL5) [petal, stamen]100, 75, 50
AT1G19830261137_at0.85 ± 0.171.84 ± 0.131.84 ± 0.091.83Auxin-associated [young flower, inflorescence]100, 75
AT5G04340245711_at0.46 ± 0.210.45 ± 0.150.93 ± 0.251.82C2H2-type zinc-finger transcription factor [petal, stamen]100, 75
AT1G48100260727_at4.11 ± 0.316.68 ± 0.997.80 ± 0.111.82Glycoside hydrolase family 28 [inflorescence, flower]100, 75
AT4G10610254990_at1.31 ± 0.101.82 ± 0.092.40 ± 0.111.79RNA-BINDING PROTEIN 37 (RBP37)100, 75
AT2G16580263238_at4.85 ± 0.178.59 ± 3.0012.09 ± 4.631.77Auxin-responsive [inflorescence]75, 50
AT5G54870248115_at0.18 ± 0.010.19 ± 0.050.35 ± 0.061.75Unknown [petal, stamen]100, 75
AT4G35770253161_at0.28 ± 0.000.46 ± 0.140.55 ± 0.081.73SENESCENCE ASSOCIATED GENE 1 (SEN1) [petal, stamen]100, 75
AT1G19540260662_at2.15 ± 0.092.72 ± 0.063.81 ± 0.081.70Isoflavone reductase100, 75
AT1G19660261144_s_at7.46 ± 1.239.01 ± 0.4512.12 ± 1.851.62Wound-responsive family100, 75, 50
AT3G45290252572_at0.60 ± 0.471.19 ± 0.121.16 ± 0.501.62MILDEW RESISTANCE LOCUS O 3 (MLO3)100, 75
AT4G30430253632_at1.03 ± 0.131.33 ± 0.011.77 ± 0.041.60TETRASPANIN9100, 75
AT1G67820245194_at0.19 ± 0.040.27 ± 0.050.40 ± 0.061.60Phosphatase 2C100, 75
AT1G42570256470_at0.19 ± 0.000.24 ± 0.030.33 ± 0.041.59Unknown100, 75
AT5G42750249190_at4.33 ± 0.055.68 ± 3.487.47 ± 0.891.57BRI1 KINASE INHIBITOR 1(BKI1) [petal, stamen]100, 75
AT5G20030246155_at0.21 ± 0.070.30 ± 0.050.34 ± 0.101.55Agenet domain-containing100, 75
AT4G27130253900_at12.25 ± 0.2012.65 ± 2.7219.01 ± 0.281.55Eukaryotic translation initiation factor SUI1100, 75, 50
AT3G57210251671_at0.26 ± 0.030.36 ± 0.040.41 ± 0.031.53Unknown100, 75
AT5G16110246506_at3.48 ± 0.826.60 ± 1.077.88 ± 2.331.53Unknown100, 75
AT1G50350262462_at1.08 ± 0.061.31 ± 0.501.67 ± 0.091.53Unknown100, 75
AT4G04955255310_at5.00 ± 0.277.71 ± 0.697.95 ± 0.021.53ARABIDOPSIS ALLANTOINASE (ATALN) [petal, stamen]100, 75
AT5G10100250467_at0.34 ± 0.090.39 ± 0.020.53 ± 0.111.50Trehalose-6-phosphate phosphatase100, 75
Table 4.  Genes differentially expressed more than 1.5-fold greater in CL flowers than in CH flowers, which were included in at least two of 50, 75, and 100 thresholds
AGI code*GeneChip IDSignal intensity (mean ± SD)Fold-changeDescription§**Threshold
CLINTCH
  • *

    ,

  • †,

  • ‡,

  • §

    §, and

  • , see Table 3 for details.

  • **

    Description of rubric ‘suppressed’ genes in floral organ and ‘expressed’ genes under stress in A. thaliana according to ATTED-II are shown in brackets.

AT1G03870265066_at0.92 ± 0.540.26 ± 0.030.37 ± 0.222.49Asciclin-like arabinogalactan-protein 9 (Fla9)100, 75
AT3G18000258218_at5.35 ± 2.443.73 ± 1.191.85 ± 1.642.35N-METHYLTRANSFERASE 1 (NMT1) [petal, stamen]100, 75, 50
AT4G13410254773_at1.34 ± 0.271.06 ± 0.130.64 ± 0.162.05CELLULOSE SYNTHASE-LIKE A15 (CSLA15)100, 75, 50
AT1G60960259723_at18.69 ± 21.362.14 ± 0.231.95 ± 0.212.00IRON REGULATED TRANSPORTER 3 (IRT3)100, 75
AT2G34810267425_at1.95 ± 0.071.42 ± 0.420.96 ± 0.041.92FAD-binding domain-containing100, 75, 50
AT5G24780245928_s_at16.80 ± 3.2716.44 ± 1.798.18 ± 2.701.89VEGETATIVE STORAGE PROTEIN 1 (VSP1) [drought stress]100, 75
AT5G56000248043_s_at7.43 ± 2.533.49 ± 0.353.34 ± 0.261.78Hsp81-4, Hsp90 family [cold stress]100, 75, 50
AT1G12090264371_at31.26 ± 12.3016.90 ± 5.1617.06 ± 6.221.78EXTENSIN-LIKE PROTEIN (ELP) [stamen]75, 50
AT3G09440258979_at13.49 ± 5.027.69 ± 0.106.95 ± 1.761.74Hsc70-3, Hsp70 family [cold stress]100, 75, 50
AT4G31210253566_at3.77 ± 1.113.60 ± 0.162.03 ± 0.861.73DNA topoisomerase family [petal, stamen]100, 75
AT4G24140254202_at0.78 ± 0.080.63 ± 0.070.41 ± 0.001.73Hydrolase, alpha/beta fold family75, 50
AT4G15210245275_at3.29 ± 1.913.80 ± 0.201.88 ± 1.161.72AT-BETA-AMY (BMY1) [petal, stamen, drought stress]100, 75
psbT244973_at9.33 ± 3.045.49 ± 0.695.21 ± 1.371.70PSII T protein encoded in a chloroplast genome100, 75, 50
AT1G20010261230_at0.87 ± 0.230.74 ± 0.010.52 ± 0.151.67TUBULIN BETA-5 CHAIN (TUB5)100, 75
AT1G50150262471_at0.79 ± 0.050.67 ± 0.280.40 ± 0.081.65Unknown100, 75
AT3G23730257203_at0.55 ± 0.290.23 ± 0.050.28 ± 0.101.65XYLOGLUCAN ENDOTRANSGLYCOSYLASE 16 (XTH16) [petal, stamen]100, 75, 50
AT5G55120248091_at0.95 ± 0.510.75 ± 0.110.56 ± 0.341.63Unknown100, 75
AT1G36000260187_at2.10 ± 0.221.98 ± 0.151.24 ± 0.061.63LOB DOMAIN-CONTAINING PROTEIN 5 (LBD5)100, 75
AT1G63730260270_at1.10 ± 0.750.60 ± 0.020.60 ± 0.581.61Disease resistance100, 75
psbB244972_at7.97 ± 3.785.20 ± 1.634.31 ± 1.441.61PSII 47KDa protein encoded in a chloroplast genome100, 75, 50
AT3G24530258134_at2.21 ± 0.461.65 ± 0.161.06 ± 0.161.61AAA-type ATPase family [cold stress]100, 75
AT2G39310266988_at0.45 ± 0.130.34 ± 0.000.25 ± 0.031.59Jacalin lectin family100, 50
AT3G12110256275_at2.16 ± 0.571.30 ± 0.041.25 ± 0.201.58Actin11100, 75, 50
AT4G16985245318_at4.20 ± 1.912.71 ± 0.162.63 ± 1.241.58Unknown100, 75, 50
AT4G13890254740_s_at9.04 ± 0.976.65 ± 0.655.16 ± 0.191.58SERINE HYDROXYMETHYLTRANSFERASE 5 (SHM5)100, 75
AT1G10970260462_at23.43 ± 27.363.68 ± 0.162.93 ± 0.421.55ZRT/IRT-LIKE PROTEIN 4 (ZIP4)100, 75, 50
AT1G28600262745_at1.94 ± 0.162.44 ± 0.561.13 ± 0.281.54Lipase100, 75
AT5G08260250517_at5.04 ± 0.864.37 ± 0.043.17 ± 0.721.54SERINE CARBOXYPEPTIDASE-LIKE 35 (SCPL35)100, 75
AT3G27060257809_at28.14 ± 3.2222.09 ± 0.6718.07 ± 2.471.533 ribonucleotide reductase (RNR) small subunit [petal]100, 75
AT4G09890255028_at0.28 ± 0.060.21 ± 0.080.16 ± 0.011.52Unknown100, 75
AT2G33880267453_at9.53 ± 0.208.36 ± 0.295.81 ± 0.501.52STIMPY100, 50
AT1G14700262830_at0.40 ± 0.150.42 ± 0.100.26 ± 0.101.51PURPLE ACID PHOSPHATASE 3 (ATPAP3) [young flower]100, 75
image

Figure 4. Graphic representation of gene expression among cleistogamous (CL), intermediate (INT), and chasmogamous (CH) flowers with greater than 1.5-fold smaller (a) and larger (b) signal intensities in CL flowers than in CH flowers according to Tables 3 and 4. Grey lines represent changes in signal intensity per gene. Black lines represent the mean ± SD of the signal intensity per gene. Relative gene expression = (objective gene expression in flower type)/(average objective gene expression among three flower types). Letters indicate significant differences among treatments with Wilcoxon tests and Bonfferoni multiple corrections.

Download figure to PowerPoint

image

Figure 5. Relative gene expression (mean ± SE) of six orthologues with a gDNA hybridization intensity threshold of 100 in the cross-species microarray (black bars) and semi-quantitative RT-PCR (white bars) among cleistogamous (CL), intermediate (INT), and chasmogamous (CH) flowers. Each bar represents N = 2. Relative gene expression = (objective gene expression)/(ACTIN2 ortholog gene expression).

Download figure to PowerPoint

Among the 69 genes, signal intensities of 37 genes were lower for CL flowers than for CH flowers (Table 3), whereas 32 genes were higher for CL flowers than for CH flowers (Table 4). Some gene families, such as CCCH-type zinc-finger transcription factor, auxin-associated (Table 3), and heat shock proteins (Table 4), were among these genes.

To validate the microarray analysis, semi-quantitative RT-PCR was performed for six amplified genes from 6 selected genes (Fig. 5): three genes (DRM1, SPL5, and AT4G29190 orthologues in C. kokaiensis) expressed lower values for CL flowers than for CH flowers (Table 3), two genes (HSP81-4 and NMT1) expressed higher values for CL flowers than for CH flowers (Table 4), and one gene (RD21) was not differentially expressed between CL and CH flowers. Semi-quantitative RT-PCR showed similar patterns to those of the microarray analysis.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

effects of chilling on cl and ch flowering

The flower form (Fig. 1; Table 2), days to flowering (Fig. 2), and plant dry mass (Fig. 3) of C. kokaiensis were all affected by chilling treatments. While previous studies have reported several environmental factors that induce CL flowers (Lord 1982; Simpson et al. 1985; Bell & Quinn 1987; Trapp & Hendrix 1988; Kenworthy et al. 1989; Kawano et al. 1990; Le Corff 1993; Mattila & Salonen 1995), this is the first report on the affect of chilling on CL flowering.

CL and CH flowering was correlated with plant size, as reported previously (Clay 1982; Antlfiger et al. 1985; Schnee & Waller 1986; Jasieniuk & Lechowicz 1987; Trapp & Hendrix 1988; Kawano et al. 1990; Berg & Redbo-Torstensson 1998; Diaz & Macnair 1998; Lu 2002). However, the correlation between CL or CH flower determination and plant size varies among plant species. For example, larger Oxalis acetosella plants have more CL flowers than CH flowers, the adaptive significance of which involves an optimizing response to environmental variation that affects resource availability (Berg & Redbo-Torstensson 1998). In our study, the effect of chilling on CL and CH flowering depended on plant dry mass – smaller plants produced CL flowers, INT flowers, or both, while larger plants produced CH and INT flowers (Fig. 1). For future studies, we will examine the adaptive significance of the size-dependent CL and CH flowering strategy in C. kokaiensis.

gene expression patterns in cl and ch flowering

We examined gene expression patterns among CL, INT, and CH flowers using the gDNA-based probe-selection strategy. We found 69 genes based on fold-changes greater than 1.5 in at least 2 of 50, 75, and 100 thresholds of the gDNA hybridization intensity (Tables 3, 4). Most signal intensities in INT flowers, which were not analysed for fold-change, were intermediate between those of CL and CH flowers; this agree with the INT flower morphology (Fig. 4). The results of the cross-species microarray analyses agreed with those of the semi-quantitative RT-PCR (Fig. 5). Furthermore, we partially sequenced 13 A. thaliana orthologous genes in C. kokaiensis (accession nos AB372085–AB372097 in DDBJ). High similarities (92.5 ± 0.73%, mean ± SE) were found to those of A. thaliana orthologous genes. These results suggest that the cross-species microarray hybridization and subsequent analyses were appropriate.

Thirty-seven of the 69 differentially expressed genes were down-regulated in CL flowers (Table 3). SPL5 (Cardon et al. 1999), NAP (Sablowski & Meyerowitz 1998), and AT5G04340 orthologues were expressed in petals and stamens, suggesting their possible involvement in differences observed between CL and CH flowers (Table 3). Auxin-related genes, such as AT1G19830 and AT2G16580 orthologues, are also involved in CL flower formation because floral organ development requires a proper auxin gradient and gene regulation by auxin (Benkováet al. 2003; Aloni et al. 2006). Since early flowering was observed in CL plants exposed to a long chilling treatment (Fig. 2), we suggest that the FLC ortholog, which acts as a repressor for flowering (Michaels & Amasino 1999), was repressed in CL plants rather than the CH plants that flowered late (Table 3).

Thirty-two of the 69 differently expressed genes were up-regulated in CL flowers (Table 4). Two heat-shock protein genes, HSP81-4 and HSC70-3 (Krishna et al. 1995; Li et al. 1999; Sung et al. 2001), and AT3G24530 were up-regulated in CL flowers following induction by cold-stress (Table 4). Furthermore, BMY1 and VSP1 were up-regulated due to induction by drought-stress. We suggest that genetic pathways involved in cold- and drought-stress may have some interaction at the molecular level (reviewed by Shinozaki & Yamaguchi-Shinozaki 2000). Furthermore, down-regulated genes pertaining to petals and stamens in A. thaliana (Table 4) may also be involved in the differences between CL and CH flowers.

In contrast, orthologues of other known genes related to petal and stamen development such as MADS-box genes (reviewed by Krizek & Fletcher 2005), SUPERMAN (Sakai et al. 1995), HUA1 (Li et al. 2001), RABBIT EARS (Takeda et al. 2004), PETAL LOSS (Brewer et al. 2004), and ROXY1 (Xing et al. 2005), did not differ in gene expression between CL and CH flowers.

ecological and evolutionary implications of cleistogamy

This study is the first report on the molecular basis of cleistogamy. We consider that not only floral developmental genes, but also other genes (e.g. those involved in the chilling response) are important in determining CL and CH flowering. Because the determination of CL and CH flowering occurred at the time of chilling treatments (Fig. 1; Table 2), down and up-regulated genes in CL flowers induced by the long chilling treatment are likely to be down-stream genes in the chilling response network. Consequently, we cannot distinguish between the ‘cause’ genes and the ‘consequence’ genes for CL and CH flowering in this microarray study. Both CL and CH flowering occurred during chilling (Fig. 1; Table 2), and, as such, down- and up-regulated genes of CL flowers induced by a long chilling treatment are likely to act downstream from a genetic network pertaining to the cold-stress response. As a result, we cannot distinguish between genes that are expressed due to induction from those that are expressed as part of a cascading event in CL and CH flowering.

The evolutionary argument for cleistogamy, such as reproductive compensation (e.g. Schemske 1978), assumes that the energetic cost of producing a CL flower is lower than that of a CH flower. In our results, however, many stress-response genes were expressed in CL flowers of C. kokaiensis. Therefore, the energetic cost, especially in terms of gene expression, may be considerably higher for the production of CL flowers than that which has been reported previously.

Stress-response genes expressed in CL flowers of C. kokaiensis (Table 4) were induced by environmental stress (i.e. chilling; Fig. 1; Table 2). In several cleistogamous plants, CL and CH flowering is also affected by light, water, and nutrient stress (Bell & Quinn 1987; Trapp & Hendrix 1988; Kawano et al. 1990; Le Corff 1993; Mattila & Salonen 1995). Therefore, although the types of environmental stress and response genes may vary between cleistogamous species, the interaction between genetic networks in response to environmental stress and floral development may be well conserved. We suggest that the genetic interactions that occur during environmental stresses are involved in the evolution of CL and CH flowering. Our results help to clarify the molecular basis for the evolution of plant mating systems depending on environmental conditions.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank N.J. Ouborg, Y. Hiwatashi, N. Aono, and two anonymous referees for their valuable comments on this manuscript; T. Yahara, H. Tsukaya, and C. Kikuchi for their information on C. kokaiensis; T. Fujita, N. Sumikawa, and M. Kitani for their technical support and advice; and C. Nanba in Research Support Facilities of NIBB for support with plant cultivation. This study was supported in part by Japan Society for the Promotion of Science (JSPS) Research Fellowships for Young Scientists to S.-I.M. and by a Grant-in-Aid for Science Research in a Priority Area from the Ministry of Education, Culture, Sports, Science, and Technology of Japan to M.H.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Appendix S1. gDNA hybridization.

Fig. S1. Flower of C. kokaiensis.

Fig. S2. Probe selection.

Fig. S3. Volcano plot.

Fig. S4. Number of genes differentially expressed.

Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
JEC_1392_sm_AppendixS1.doc37KSupporting info item
JEC_1392_sm_FigS1.pdf4112KSupporting info item
JEC_1392_sm_FigS2.pdf131KSupporting info item
JEC_1392_sm_FigS3.pdf2393KSupporting info item
JEC_1392_sm_FigS4.pdf114KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.