The PHYTOCHROME C photoreceptor gene mediates natural variation in flowering and growth responses of Arabidopsis thaliana

Abstract

Light has an important role in modulating seedling growth and flowering time. We show that allelic variation at the PHYTOCHROME C (PHYC) photoreceptor locus affects both traits in natural populations of A. thaliana. Two functionally distinct PHYC haplotype groups are distributed in a latitudinal cline dependent on FRIGIDA, a locus that together with FLOWERING LOCUS C explains a large portion of the variation in A. thaliana flowering time. In a genome-wide scan for association of 65 loci with latitude, there was an excess of significant P values, indicative of population structure. Nevertheless, PHYC was the most strongly associated locus across 163 strains, suggesting that PHYC alleles are under diversifying selection in A. thaliana. Our work, together with previous findings, suggests that photoreceptor genes are major agents of natural variation in plant flowering and growth response.

Figure 1

Identification of a defective PHYC allele in Fr-2. (a) Expression profiles of genes that are differentially expressed between Fr-2 and Col across 34 wild strains. The genomic interval that co-segregates with early flowering of Fr-2 were analyzed for differentially expressed genes using the AtGenExpress dataset of expression data from 34 strains. Of 112 genes in this interval represented on the ATH1 microarray, 11 were differentially expressed between Col-0 and Fr-2, and are shown here. (b) Expression profile of PHYC (At5g38540) in the 34 strains shown in (a). (c) Red light response of Fr-2, Col-0 and phyB-9 in Col-0. The response of Fr-2 is similar to that of phyB-9, indicating reduced red light sensitivity. (d) Flowering under short days of different parental lines and their F₁ progeny.

Figure 2

Quantitative complementation analysis with different parental lines. Col-0 was used as the wild type and the null allele phyC-2 in the Col-0 background as a tester for the crosses. An ANOVA was performed with the following model: TLN~ Line + Cross + Line x Cross. The Line x Cross interaction was significant in all three combinations (p < 0.0001). However, the proportion of total variance accounted for by the Line x Cross interaction varies (shown as R²), which is consistent with the Ler allele being intermediate in activity between the Col-0 and Fr-2 alleles.

Figure 3

PHYC haplotypes. A representative unrooted phylogenetic tree generated from PHYC coding sequences is shown on top. Numbers indicate bootstrap values above 60. The amino acid changes that delineate the two haplotype groups are given below. Unique changes compared to the outgroup A. lyrata are found in both haplotypes. Amino acids that are conserved in other phytochromes (Q or E at position 735, and E at position 822), but changed in either the Col-0- or Ler-type haplotype, are boxed. See Supplementary Table 6 online for provenance of strains.

Figure 4

Latitudinal cline of PHYC alleles. (a) Proportion of Ler-type (black) and Col-0-type (white) PHYC alleles at different latitudes among apparently FRI functional strains. The absolute numbers for each of the classes is given on top of the histograms. (b) Distribution of p-values of a nominal logistic regression model with latitude as a factor and genotypes as response. Allele information of 65 random SNP markers with similar allele frequency as that of PHYC was available in a set of 163 strains. This information was used as a response. Note genome-wide skew towards small p-values. (c) Distribution of p-values for interaction of a given random marker with FRI in an interaction model with latitude as the response and FRI and marker genotypes as factors with interaction.

Supplementary Figure 1

Markers used for mapping of the early flowering phenotype in Fr- 2. About 80 early plants were genotyped for all markers and linkage of the phenotype to MSat 5.22 and SO191 was confirmed with 60 additional early plants.

Supplementary Figure 2

Flowering behavior of populations segregating for different PHYC alleles. (a) Distribution of flowering time in an F₂ population derived from Ler x Fr-2 cross showing a continuous distribution, in contrast to a bimodal distribution observed in an F₂ population derived from the Col x Fr-2 cross (see Lempe, J. et al. Diversity of flowering responses in wild Arabidopsis thaliana strains. PLoS Genet. 1, e6 [2005]). (b) Average flowering times of plants with different allelic combinations at PHYC in F₂ populations derived from Ler x Fr-2 and Col x Fr-2. A single copy of the Col-0 allele delays flowering much more than a single Ler allele. (c) Genetic complementation analysis with Ler.Flowering times of F₁ progeny in short days along with parental lines are shown.

Supplementary Figure 3

Latitudinal cline of PHYC alleles. (a) Distribution of p-values of a nominal logistic regression model with latitude as a factor and genotype as response. Allele information of 69 random SNP markers with similar allele frequency as that of PHYC was available in a set of 56 strains. This information was used as a response. Note genome-wide skew towards small p-values. (c) Distribution of p-values for interaction of a given random marker with FRI in an interaction model with latitude as the response and FRI and marker genotypes as factors with interaction. PHYC falls in the top bin for both associations and the p-values for PHYC are smaller than those for FLC.