Gene duplication in the carotenoid biosynthetic pathway preceded evolution of the grasses

Abstract

Despite ongoing research on carotenoid biosynthesis in model organisms, there is a paucity of information on pathway regulation operating in the grasses (Poaceae), which include plants of world-wide agronomic importance. As a result, efforts to either breed for or metabolically engineer improvements in carotenoid content or composition in cereal crops have led to unexpected results. In comparison to maize (Zea mays), rice (Oryza sativa) accumulates no endosperm carotenoids, despite having a functional pathway in chloroplasts. To better understand why these two related grasses differ in endosperm carotenoid content, we began to characterize genes encoding phytoene synthase (PSY), since this nuclear-encoded enzyme appeared to catalyze a rate-controlling step in the plastid-localized biosynthetic pathway. The enzyme had been previously associated with the maize Y1 locus thought to be the only functional gene controlling PSY accumulation, though function of the Y1 gene product had never been demonstrated. We show that both maize and rice possess and express products from duplicate PSY genes, PSY1 (Y1) and PSY2; PSY1 transcript accumulation correlates with carotenoid-containing endosperm. Using a heterologous bacterial system, we demonstrate enzyme function of PSY1 and PSY2 that are largely conserved in sequence except for N- and C-terminal domains. By database mining and use of ortholog-specific universal PCR primers, we found that the PSY duplication is prevalent in at least eight subfamilies of the Poaceae, suggesting that this duplication event preceded evolution of the Poaceae. These findings will impact study of grass phylogeny and breeding of enhanced carotenoid content in an entire taxonomic group of plant crops critical for global food security.

Figure 1.

Gene structures of maize and rice PSY1 and PSY2. Maize PSY1 (GenBank AY324431); rice PSY1 (GenBank AP005750); maize PSY2 (GenBank AY325302); rice PSY2 (GenBank AL831803). Arrows and thin bars indicate exons (bold numbers) and introns, respectively, sizes for which are in bp.

Figure 2.

Functional complementation of PSY1 and PSY2. E. coli cells were transformed with: (A) pACCAR25; (B) pACCAR25ΔcrtB + pET23a (empty vector); (C) pACCAR25ΔcrtB + pEMPSY1-1; (D) pACCAR25ΔcrtB + pEMPSY2-1; and (E) pACCAR25ΔcrtB + pERPSY2-1 and extracted pigments analyzed by HPLC chromatograms shown at 450 nm. The spectral fine spectrum for the pathway end products, zeaxanthin β-D diglucoside (peak 1), zeaxanthin mono glucoside (peak 2), and zeaxanthin (peak 3) are shown in sections F, G, and H with retention times of 7.2 min, 8.1 min, and 13.1 min, respectively.

Figure 3.

Leaf and endosperm transcript profiles for PSY1 and PSY2 in maize and rice tested by RT-PCR. A, Specificity of species- and gene-specific primers tested for both maize and rice. Left column, PSY1 primers amplify only PSY1 template and not PSY2 template; right column, PSY2 primers amplify only PSY2 template and not PSY1 template. B, Specific transcripts for genes indicated were amplified by RT-PCR from leaf or endosperm mRNA of maize or rice. Yellow and white correspond to the endosperm phenotype of maize B73 and the maize y1 mutant, respectively.

Figure 4.

Phylogenic analysis of PSY amino acid sequences. SwissProt numbers, in bold, and GenBank accessions are in parentheses. Lycopersicon esculentum (AAA34187, L23424.1; J. Hirschberg, unpublished data), L. esculentum (AAA34153.1, M84744), Cucumis melo (CAA85775.1, Z37543), Arabidopsis (AAAN17427, BT000450.1), Citrus unshiu (BAB18514, AB037975.1), Hordeum vulgare subsp. vulgare (BI955682), Triticum aestivum (BM137086), Z. mays (AY325302), O. sativa (japonica, AL831803), H. vulgare (BE421261), O. sativa (japonica, AP005750), Z. mays (AY324431), T. aestivum (CD862515), and Narcissus pseudonarcissus (CAA55391, X78814).

Figure 5.

Poaceae subfamily genomes tested and found to possess the PSY gene duplication. A, Testing of the ortholog-specific universal primers indicated at the top using as template, M (maize B73) or R (rice IR36) genomic DNA. B, Species representing 8 of the 12 subfamilies of the Poaceae family found to have both the PSY1 and PSY2 genes based on amplification using the ortholog-specific universal primers tested in A and further confirmed by alignment of the amplified exonic DNA sequences.