Coconuts (Cocos nucifera L.) are divided by the height into tall and dwarf types. In many plants the short phenotype was emerged by mutation of the GA20ox gene encoding the enzyme involved in gibberellin (GA) biosynthesis. Two CnGA20ox genes, CnGA20ox1 and CnGA20ox2, were cloned from tall and dwarf types coconut. The sequences, gene structures and expressions were compared. The structure of each gene comprised three exons and two introns. The CnGA20ox1 and CnGA20ox2 genes consisted of the coding region of 1110 and 1131 bp, encoding proteins of 369 and 376 amino acids, respectively. Their amino acid sequences are highly homologous to GA20ox1 and GA20ox2 genes of Elaeis guineensis, but only 57% homologous to each other. However, the characteristic amino acids two histidines and one aspartic acid which are the two iron (Fe) binding residues, and arginine and serine which are the substrate binding residues of the dioxygenase enzyme in the 20G-FeII_Oxy domain involved in GA biosynthesis, were found in the active site of both enzymes. The evolutionary relationship of their proteins revealed three clusters in vascular plants, with two subgroups in dicots and three subgroups in monocots. This result confirmed that CnGA20ox was present as multi-copy genes, and at least two groups CnGA20ox1 and CnGA20ox2 were found in coconut. The nucleotide sequences of CnGA20ox1 gene in both coconut types were identical but its expression was about three folds higher in the leaves of tall coconut than in those of dwarf type which was in good agreement with their height. In contrast, the nucleotide sequences of CnGA20ox2 gene in the two coconut types were different, but the expression of CnGA20ox2 gene could not be detected in either coconut type. The promoter region of CnGA20ox1 gene was cloned, and the core promoter sequences and various cis-elements were found. The CnGA20ox1 gene should be responsible for the height in coconut, which is different from other plants because no mutation was present in CnGA20ox1 gene of dwarf type coconut.

ABSTRACT. The fragrance gene, betaine aldehyde dehydrogenase 2 (Badh2), has been well studied in many plant species. The objectives of this study were to clone Badh2 and compare the sequences between aromatic and non-aromatic coconuts. The complete coding region was cloned from cDNA of both aromatic and non-aromatic coconuts. The nucleotide sequences were highly homologous to Badh2 genes of other plants. Badh2 consisted of a 1512-bp open reading frame encoding 503 amino acids. A single nucleotide difference between aromatic and non-aromatic coconuts resulted in the conversion of alanine (non-aromatic) to proline (aromatic) at position 442, which was the substrate binding site of BADH2. The ring side chain of proline could destabilize the structure leading to a non-functional enzyme. Badh2 genomic DNA was cloned from exon 1 to 4, and from exon 5 to 15 from the two coconut types, except for intron 4 that was very long. The intron sequences of the two coconut groups were highly homologous. No differences in Badh2 expression were found among the tissues of aromatic coconut or between aromatic and non-aromatic coconuts. The amino acid sequences of BADH2 from coconut and other plants were compared and the genetic relationship was analyzed using MEGA 7.0. The phylogenetic tree reconstructed by the Bayesian information criterion consisted of two distinct groups of monocots and dicots. Among the monocots, coconut (Cocos nucifera) and oil palm (Elaeis guineensis) were the most closely related species. A marker for coconut differentiation was developed from one-base substitution site and could be successfully used.