The study of the E-class SEPALLATA3-like MADS-box genes in wild-type and mutant flowers of cultivated saffron crocus (Crocus sativus L.) and its putative progenitors

To further understand flowering and flower organ formation in the monocot crop saffron crocus (Crocus sativus L.), we cloned four MIKC^c type II MADS-box cDNA sequences of the E-class SEPALLATA3 (SEP3) subfamily designated CsatSEP3a/b/c/c_as as well as the three respective genomic sequences. Sequence analysis showed that cDNA sequences of CsatSEP3 c and c_as are the products of alternative splicing of the CsatSEP3c gene. Bioinformatics analysis with putative orthologous sequences from various plant species suggested that all four cDNA sequences encode for SEP3-like proteins with characteristic motifs and amino acids, and highlighted intriguing sequence features. Phylogenetically, the isolated sequences were closest to the SEP3-like genes from monocots such as Asparagus virgatus, Oryza sativa, Zea mays, and the dicot Arabidopsis SEP3 gene. All four isolated C. sativus sequences were strongly expressed in flowers and in all flower organs: whorl1 tepals, whorl2 tepals, stamens and carpels, but not in leaves. Expression of CsatSEP3a/b/c/c_as cDNAs was compared in wild-type and mutant flowers. Expression of the isolatedCsatSEP3-like genes in whorl1 tepals together with E-class CsatAP1/FUL subfamily and B-class CsatAP3 and CsatPI subfamilies of genes, fits the ABCE “quartet model,” an extended form of the original ABC model proposed to explain the homeotic transformation of whorl1 sepals into whorl1 tepals in Liliales and Asparagales plants such as C. sativus. This conclusion was also supported by the interaction of the CsatSEP3b protein with CsatAP1/FUL and CsatAP3 proteins. In contrast, expression of both B-class CsatAP3 and CsatPI genes and the C-class CsatAGAMOUS genes together with E-class CsatSEP3-like genes in carpels, without any phenotypic effects on carpels, raises questions about the role of these gene classes in carpel formation in this non-grass monocot and requires further experimentation. Finally, taking advantage of the size and sequence differences in amplified genomic sequences of the triploid C. sativus and comparing them with the respective sequences from C. tomasii, C. hadriaticus and C. cartwrightianus, three putative wild-type diploid progenitor species, we examined the origin of CsatSEP3a sequence.