ISHS Acta Horticulturae 1064: VI International Symposium on Rose Research and Cultivation (2015)
MOLECULAR GENETIC STUDIES ON CONTINUOUS-FLOWERING ROSES THAT DO NOT ORIGINATE FROM ROSA CHINENSIS
T. Horibe, K. Yamada, S. Otagaki, S. Matsumoto, K. Kawamura

Abstract:
The Continuous-Flowering (CF) behavior of modern roses is considered to originate from a Chinese rose, Rosa chinensis. In R. chinensis, an insertion of a copia-like retrotransposon is present in the floral repressor gene KSN, and it blocks the maturation of transcripts of the gene and allows the rose to flower continuously. Most modern CF roses are expected to have this mutated allele of KSN (ksncopia). Because of this narrow genetic background of CF modern roses, we aim to discover the CF roses that have no ksncopia allele, which will be good breeding materials for novel CF roses. We also aim to test the possible involvement of KSN in the regulation of CF behavior in the novel CF species, R. rugosa. By genotyping of KSN we found that the CF species (R. rugosa) have only a wild KSN allele without the insertion, indicating that its CF behavior has a different origin than that of R. chinensis. Sequencing analysis did not find any obvious mutations in the KSN gene of R. rugosa, and the seasonal expression of KSN was clearly linked with the alternation of flowering/vegetative stages of shoot apices in R. rugosa. Thus, KSN is involved in the control of CF behavior of R. rugosa, although it is not simply explained by the mutation in coding region of KSN. In order to uncover the genetic determinism of CF behavior in R. rugosa, cross hybridization and Quantitative Trait Loci analysis of segregating population for flowering-behavior is necessary.

INTRODUCTION
Roses have two contrasting flowering-behavior (Fig. 1): Once-Flowering (OF) and Continuous-Flowering (CF). Most wild rose species produce flowers only in spring (OF), while many cultivars continuously produce flowers during all favorable seasons (CF). The CF behavior of modern cultivators is considered to originate from a Chinese rose, Rosa chinensis in 18th century (Hurst, 1941). The old CF rose R. chinensis is derived from a wild ancestral OF species, R. chinensis spontanea (Ogisu, 1996). Recent molecular studies have uncovered the molecular basis of CF behavior in R. chinensis (Hibrand-Saint Oyant et al., 2008; Foucher et al., 2008; Remay et al., 2009; Kawamura et al., 2011; Iwata et al., 2012; Wang et al., 2012; Randoux et al., 2012; reviewed by Bendahmane et al., 2013). The CF behavior of R. chinensis is controlled by a monogenic recessive locus called RB, and RB was shown to be a homolog of Arabidopsis TERMINAL FLOWER 1 (TFL1). The roseTFL1 gene was named KSN (or RoKSN), and KSN has a function of floral repressor in rose. In R. chinensis, an insertion of a copia-like retrotransposon is present in KSN, and it blocks the maturation of transcripts of the gene and allows the rose to flower continuously. Because this retrotransposon insertion is absent in the KSN of the wild ancestral OF species R. chinensis spontanea, the genetic origin of CF behavior in R. chinensis is supposed to be the mutation (i.e., the retrotransposon insertion in KSN). Therefore most modern cultivars with CF behavior are expected to have this mutated allele of KSN.

Seasonal Expression Analysis of RoFT and KSN
Figure 6 shows the seasonal expression patterns of RoFT and KSN in R. rugosa (CF species), R. multiflora (OF species), and ‘Hinaarare’ (CF cultivar). As reported by previous studies (Remay et al., 2009; Iwata et al., 2012), KSN expression was suppressed, and the floral activator RoFT increased gradually for all roses during the floral initiation period in early spring (Fig. 6a). RoFT expression increased dramatically during the 1st flowering period (Fig. 6b), while it decreased to zero in the 2nd shoots at young vegetative stages in all roses (Fig. 6c). In this stage, R. rugosa (CF species) showed a low level of KSN expression, which is contrast to the high level of KSN expression in the OF species R. multiflora (Fig. 6c). This different expression pattern of KSN between R. rugosa and R. multiflora may be associated with the different flowering behavior between the species. After the vegetative stages, RoFT increased again along with the second flowering of CF roses (R. rugosa and‘Hinaarare’), whereas it was not detected in the OF species R. multiflora (Fig. 6d). As expected, ‘Hinaarare’ has no detectable expression of KSN over all seasons, because the copia-like retrotransposon insertion blocks the transcript accumulation of KSN (Iwata et al., 2012).

Fig. 6. Seasonal expression of RoFT and KSN genes on shoot apices of Rosa rugosa, R. multiflora, and ‘Hinaarare’. 1SA, 1st Shoot Apices immediately after bud flush in March; 1SA2w, 2 weeks after 1SA; 1SA4w, 4 weeks after 1SA; 1FB, Floral buds formed on 1st shoot apices in April; 1BS, Balloon Stage of floral buds with visible petals; 2SAJune, 2nd shoot apices sampled on 27th June; 2SAJuly, 2nd shoot apices sampled on 11th July. (a) RoFT increased gradually, whereas KSN was suppressed during the floral initiation of 1st flowering; (b) RoFT was highly expressed in all species during the 1st flowering; (c) RoFT was suppressed during the vegetative period of 2nd shoot. KSN was also suppressed in the Continuous-Flowering (CF) rose R. rugosa, but it increased in the Once-Flowering (OF) rose R. multiflora; (d) RoFT increased again in the CF roses R. rugosa and ‘Hinaarare’ during the 2nd flowering period, while it was undetectable in R. multiflora.

CONCLUSIONS
KSN genotyping demonstrates that the continuous-flowering species R. rugosa does not have the mutated allele of KSN as is present in R. chinensis. This species can be a good breeding material for new continuous-flowering roses. The KSN genotyping of a wider range of rose species and groups will clarify which groups have different origins of continuous-flowering behavior from R. chinensis. We also showed that the KSN protein of R. rugosa has no obvious mutation and seems to keep the function of floral repressor, and the expression analysis is consistent with the involvement of this gene in the control of continuous-flowering of the species. In order to uncover the genetic determinism of continuous-flowering behavior in R. rugosa, cross hybridization experiment and QTL analysis of segregating population for flowering-behavior will be very helpful.