Elucidation of a carotenoid biosynthesis gene cluster encoding a novel enzyme, 2,2'-beta-hydroxylase, from Brevundimonas sp. strain SD212 and combinatorial biosynthesis of new or rare xanthophylls

Abstract

A carotenoid biosynthesis gene cluster mediating the production of 2-hydroxyastaxanthin was isolated from the marine bacterium Brevundimonas sp. strain SD212 by using a common crtI sequence as the probe DNA. A sequence analysis revealed this cluster to contain 12 open reading frames (ORFs), including the 7 known genes, crtW, crtY, crtI, crtB, crtE, idi, and crtZ. The individual ORFs were functionally analyzed by complementation studies using Escherichia coli that accumulated various carotenoid precursors due to the presence of other bacterial crt genes. In addition to functionally identifying the known crt genes, we found that one (ORF11, named crtG) coded for a novel enzyme, carotenoid 2,2'-beta-hydroxylase, which showed intriguingly partial homology with animal sterol-C5-desaturase. When this crtG gene was introduced into E. coli accumulating zeaxanthin and canthaxanthin, the resulting transformants produced their 2-hydroxylated and 2,2'-dihydroxylated products which were structurally novel or rare xanthophylls, as determined by their nuclear magnetic resonance and high-performance liquid chromatography/photodiode array detector/atmospheric pressure chemical ionization mass spectrometry spectral data. The new carotenoid produced was suggested to have a strong inhibitory effect on lipid peroxidation.

FIG. 1.

Organization of the carotenoid biosynthesis gene clusters isolated from Brevundimonas sp. strain SD212 and from other cyclic carotenoid-generating bacteria. (a) Organization of the Brevundimonas sp. strain SD212 carotenoid biosynthesis gene cluster composed of an 11,991-bp EcoRI fragment. Numbers represent nucleotide positions numbered from the 5′ end of the EcoRI site. The numbers over and under the gene map, respectively, represent the positions of the start and termination of individual genes or ORFs. ORF11 was identified as a 2,2′-β-hydroxylase gene in this study and is designated crtG. (b) Comparison of the organization of the cyclic carotenoid gene clusters of Paracoccus sp. strain N81106 (GenBank accession number D58420) (31), Paracoccus zeaxanthinifaciens (U62808) (37), Xanthobacter autotrophicus Py2 (AF408848) (23), Bradyrhizobium sp. strain ORS278 (AF218415) (17), Pantoea ananatis (formerly classified as Erwinia uredovora 20D3, D90087) (30), and Pantoea agglomerans (M87280) (19). Arrows represent the orientations of genes. The values shown below the genes indicate the percentages of amino acid identity compared with Brevundimonas sp. strain SD212, which were calculated by using Clustal W (18) and GeneDoc (35). The zeaxanthin-hyperproducing bacterium that was formerly classified as Flavobacterium sp. strain R1534 has recently been reported to be the mutant of P. zeaxanthinifaciens ATCC 21588, as taxonomically reevaluated by Berry et al. (6). Note that all of the other bacteria, except for the genus Pantoea of the γ-Proteobacteria, belong to the α-Proteobacteria.

FIG. 2.

Carotenoid products from E. coli carrying pUCBre-O11 and pAC-Asta (a), pUCBre-O11 and pACCAR25ΔcrtX (b), and pUCBre-O11 and pAC-Cantha (c). HPLC/PDA was performed using a TSK gel ODS-80Ts column, as described in Materials and Methods. Detection was carried out at 470 nm. The numbers of carotenoids correspond to those discussed in “Spectral data for individual carotenoids” in Materials and Methods.

FIG. 3.

Chemical structures of the carotenoid pigments produced by Brevundimonas sp. strain SD212 (55) and proposed functions of the carotenoid biosynthesis gene products. The functions of CrtE, CrtB, CrtI, and CrtY are shown in Fig. 4. FPP, farnesyl diphosphate (pyrophosphate).

FIG. 4.

Combinatorial biosynthesis of novel and rare carotenoids by using E. coli carrying plasmids pUCBre-O11 and pACCAR25ΔcrtX or pAC-Cantha.

FIG. 5.

Alignment of the deduced amino acid sequences of CrtG from Brevundimonas sp. strain SD212 and sterol-C5-desaturases (C5D) derived from human (Homo sapiens) (GenBank accession no. AB016247), mouse (Mus musculus) (GenBank accession no. BC024132), and rat (Rattus norvegicus) (GenBank accession no. AB052846) origins. Asterisks and dots indicate amino acid residues identical and similar, respectively, for all of the aligned sequences. Identical or similar residues among all aligned sequences are shown with a black background. A gray background indicates partially aligned sequences. Arrows indicate the homologous region (the amino acids shown for CrtG are 51 to 233).