Horizontal Gene Transfer and Fusion Spread Carotenogenesis Among Diverse Heterotrophic Protists

Abstract

Thraustochytrids (phylum: Labyrinthulomycota) are nonphotosynthetic marine protists. Some thraustochytrids have crtIBY, a trifunctional fusion gene encoding a protein capable of β-carotene biosynthesis from geranylgeranyl pyrophosphate. Here we show that crtIBY is essential in, and encodes the sole pathway for, carotenoid biosynthesis in the thraustochytrid Aurantiochytrium limacinum ATCC MYA-1381. We explore the evolutionary origins of CrtIBY and discover that the closest related protein domains are present in a small but diverse group of other heterotrophic protists, including the apusomonad Thecamonas trahens and the dinoflagellates Oxyrrhis marina and Noctiluca scintillans. Each organism within this cluster also contains one or more β-carotene 15-15' oxygenase genes (blh and rpe65), suggesting that the acquisition of β-carotene biosynthesis genes may have been related to the production of retinal. Our findings support a novel origin of eukaryotic (apo)carotenoid biosynthesis by horizontal gene transfer from Actinobacteria, Bacteroidetes, and/or Archaea. This reveals a remarkable case of parallel evolution of eukaryotic (apo)carotenogenesis in divergent protistan lineages by repeated gene transfers.

Keywords: carotenoid oxygenase; lycopene cyclase; phylogenetics; phytoene desaturase; phytoene synthase; thraustochytrids.

Fig. 1.

β-carotene biosynthesis is accomplished by orthologous enzymes across taxa. (A) Generalized flow-chart of the isoprenoid, sterol, and carotenoid biosynthesis pathways. Enzymatic coloration corresponds to orthologs outlined in (B). (B) Orthologous carotenoid biosynthesis genes in nonphotosynthetic and photosynthetic organisms (expanded from Alcaíno et al. 2016 and Sandmann 2001). Genes that are orthologous are shaded similarly; fusion genes are indicated by concatenation (e.g., crtIBY is a fusion gene of crtI, crtB, and crtYc/d).

Fig. 2.

Inactivation of crtIBY in A. limacinum. (A) Agar plate streaks of WT with natural carotenogenic pigmentation relative to the two pigment-less crtIBY KOs (32 and 33). (B) Annotated genome maps (SnapGene) generated from nanopore sequencing of the A. limacinum (WT) and two KOs (32 and 33) reveal an intact crtIBY locus in WT and disrupted, yet successful integration of shble in the crtIBY open reading frame in the two KOs. Sequences indicate a double homologous recombination event having occurred in 32, while a triple tandem repeat integration event occurred in 33. Both integration events resulted in a nonfunctional form of the CrtIBY protein.

Fig. 3.

SAHNTO phylogenetic distribution and organization of carotenoid biosynthesis domains. (A) Schematic of divergent eukaryotic lineages, modified from Charon et al. (2020). Indicated are all SAHNTO members: the thraustochytrids A. limacinum, S. aggregatum, H. fermentalgiana (Stramenopiles, Thraustochytriaceae); the dinoflagellates N. scintillans and O. marina (Alveolates, Dinophyceae); and the apusomonad T. trahens (Apusomonadidae). (B) Protein domain organization diagrams (N- to C-terminus; domain shading is consistent with fig. 1): CrtB—phytoene synthase, CrtI—phytoene desaturase, CrtYc/d—lycopene cyclase, Blh—β-carotene 15-15' oxygenase (BCD superfamily), and CrtY—phototrophic lycopene cyclase. A CrtI, CrtB, CrtYc/d trifunctional multidomain protein is found in the thraustochytrids: S. aggregatum (Schag101501), A. limacinum (Aurli_150841), H. fermentalgiana (re-annotated A0A2R5GF32; see Supplementary Material online); whereas a CrtI, CrtB, CrtYc/d, Blh quadrifunctional multidomain protein is found in T. trahens (XP_013761525.1). Two bifunctional multidomain proteins are found in N. scintillans as CrtB, CrtYc/d and CrtI, CrtYc/d (CAMPEP0194550082 and CAMPEP0194488352, respectively), whereas in O. marina CrtB, CrtI, Blh, and CrtY are found as four single-domain proteins (CAMPEP0205054184, CAMPEP0204966166, CAMPEP0205060456, and CAMPEP0204311066, respectively). Protein lengths in amino acids (aa) are included for reference.

Fig. 4.

The carotenoid biosynthesis domains from a set of diverse, taxonomically distant eukaryotes called SAHNTO (S. aggregatum, A. limacinum, H. fermentalgiana, N. scintillans, T. trahens, and O. marina) group together (⋆) in phylogenies of (A) CrtB/CrtM/HpnD, (B) CrtI/CrtH/Z-ISO/CrtISO, (C) CrtYc/d, and (D) Blh. For all phylogenies, domain sequences were aligned with multiple alignment using fast fourier transform (MAFFT), retaining positions where less than 90–99% of sequences contained gaps (see Supplementary Material online). ML phylogenies were estimated in IQ-TREE using the best-fit model (supplementary table S5, Supplementary Material online) and midpoint rooted. The CrtB/CrtM/HpnD phylogeny was truncated to remove HpnD sequences. Taxa and node support in the red boxes are magnified in figure 5, and for complete trees see supplementary figures S4–S6 and S8, Supplementary Material online. Scale bars indicate the inferred number of amino acid substitutions per site.

Fig. 5.

Phylogenetic structure surrounding SAHNTO clusters, including taxa and node support of SAHNTO (⋆) sisters in carotenoid biosynthesis phylogenies: (A) CrtB/CrtM/HpnD, (B) CrtI/CrtH/Z-ISO/CrtISO, (C) CrtYc/d, and (D) Blh. Values indicate the results of the SH-aLRT and UF of 1,000 replicates, respectively. For complete trees see supplementary figures S4–S6 and S8, Supplementary Material online. Scale bars indicate the inferred number of amino acid substitutions per site.