Gene transfers shaped the evolution of de novo NAD+ biosynthesis in eukaryotes

Abstract

NAD(+) is an essential molecule for life, present in each living cell. It can function as an electron carrier or cofactor in redox biochemistry and energetics, and serves as substrate to generate the secondary messenger cyclic ADP ribose and nicotinic acid adenine dinucleotide phosphate. Although de novo NAD(+) biosynthesis is essential, different metabolic pathways exist in different eukaryotic clades. The kynurenine pathway starting with tryptophan was most likely present in the last common ancestor of all eukaryotes, and is active in fungi and animals. The aspartate pathway, detected in most photosynthetic eukaryotes, was probably acquired from the cyanobacterial endosymbiont that gave rise to chloroplasts. An evolutionary analysis of enzymes catalyzing de novo NAD(+) biosynthesis resulted in evolutionary trees incongruent with established organismal phylogeny, indicating numerous gene transfers. Endosymbiotic gene transfers probably introduced the aspartate pathway into eukaryotes and may have distributed it among different photosynthetic clades. In addition, several horizontal gene transfers substituted eukaryotic genes with bacterial orthologs. Although horizontal gene transfer is accepted as a key mechanism in prokaryotic evolution, it is supposed to be rare in eukaryotic evolution. The essential metabolic pathway of de novo NAD(+) biosynthesis in eukaryotes was shaped by numerous gene transfers.

Keywords: NAD+ biosynthesis; endosymbiotic gene transfer; horizontal gene transfer; metabolism.

Fig. 1.—

De novo NAD⁺ biosynthesis. The two different pathways for de novo NAD⁺ biosynthesis starting with either tryptophan for the kynurenine pathway or aspartate for the aspartate pathway are shown. Both pathways converge at quinolinate. Structural formulas for intermediates and all substrate and enzyme names according to KEGG (Kanehisa et al. 2014).

Fig. 2.—

Phylogenetic distribution of de novo NAD⁺ biosynthesis pathways in eukaryotes shows a patchy distribution. Shown is the organismal phylogeny (Burki, Okamoto, et al. 2012; Keeling and Palmer 2008) for the major groups of life with emphasis on plastid bearing eukaryotes (star symbols; yellow, most species photoautotroph; white, some species photoautotroph; Phaeo’, Phaeophyceae or brown algae). For each clade, the presence of enzymes catalyzing the aspartate (closed symbols; AO, aspartate oxidase; QS, quinolinate synthase) or the kynurenine pathway (open symbols; TDO/IDO, tryptophan-/indoleamine 2,3-dioxygenase; AFM, arylformamidase; KMO, kynurenine 3-monooxygenase; KYU, kynureninase; HAO, 3-hydroxyanthranilate 3,4-dioxygenase) is indicated. Gray open symbols indicate that the IDO identified in this work, based on sequence similarity, might not actually catalyze the first reaction of the kynurenine pathway (see text). Most bacteria use the aspartate pathway, whereas some have the kynurenine pathway (see text). Very few bacteria, if any, seem to use both pathways in parallel.

Fig. 3.—

An evolutionary tree for l-aspartate oxidase indicates separate origins of eukaryotic sequences. The unrooted Bayesian tree shows posterior probabilities above the branches and PhyML bootstrap values displayed as percentages below the branches. Thickened horizontal lines represent 1.0 Bayesian posterior probability. Trees with significantly lower log-likelihoods were obtained when monophyly was enforced for all sequences from Eukaryota, from any two of the three eukaryotic branches (Viridiplantae, Glaucophyta, Stramenopiles), or from Bacteria (supplementary table S1, Supplementary Material online). Larger clades have been collapsed for presentation and size is indicative of the number of taxa within the clade (full trees are available at TreeBase; http://purl.org/phylo/treebase/phylows/study/TB2:S15669, last accessed September 5, 2014). Scale bar represents 0.2 substitutions per site.

Fig. 4.—

An evolutionary tree for quinolinate synthase indicates separate origins of eukaryotic sequences. The unrooted Bayesian tree shows posterior probabilities above the branches and PhyML bootstrap values below the branches. Thickened horizontal lines represent 1.0 Bayesian posterior probability. Trees with significantly lower log-likelihoods were obtained when monophyly was enforced for all sequences from Eukaryota, from Viridiplantae plus Cyanophora paradoxa, from Bacteria, or from Archaea (supplementary table S1, Supplementary Material online).

Fig. 5.—

An evolutionary tree for nicotinate-nucleotide pyrophosphorylase indicates that some photosynthetic eukaryotes acquired genes through transfers. The unrooted Bayesian tree shows posterior probabilities above the branches and PhyML bootstrap values below the branches. Thickened horizontal lines represent 1.0 Bayesian posterior probability. A tree with significantly lower log-likelihood was obtained when monophyly was enforced for all sequences from Eukaryota (supplementary table S1, Supplementary Material online). A tree with practically identical log-likelihood was obtained when monophyly was enforced for all sequences from Viridiplantae plus Bigelowiella natans (form a monophyletic clade within Bacteroidetes in PhyML tree). The sequence from Cyanophora paradoxa was omitted from an evolutionary analysis because it is incomplete. The best BLASTp hit for this incomplete sequence in the nr database from NCBI was with a nicotinate-nucleotide pyrophosphorylase sequence from Capsaspora owczarzaki (167 score, 67% coverage), and all top 100 BLASTp hits were with sequences from nonphotosynthetic eukaryotes, indicating possible descent of nicotinate-nucleotide pyrophosphorylase in Glaucophyta from the last common ancestor of eukaryotes.

Fig. 6.—

An evolutionary tree for kynurenine-3-monooxygenase indicates that some unicellular eukaryotes acquired genes through horizontal transfers. The unrooted Bayesian tree shows posterior probabilities above the branches and PhyML bootstrap values below the branches. Thickened horizontal lines represent 1.0 Bayesian posterior probability. Trees with significantly lower log-likelihood were obtained when monophyly was enforced for all sequences from eukaryotes, from Amoebozoa plus other Unikonta (without Sphaeroforma arctica), from S. arctica plus other Opisthokonta, or from Emiliania huxleyi, Aureococcus anophagefferens, and Perkinsus marinus plus Labyrinthulomycetes and Chlorarachniophyta (i.e., SAR and Haptophyta; fig. 2 and supplementary table S1, Supplementary Material online).

Fig. 7.—

Gene transfers in the evolution of de novo NAD⁺ synthesis. Shown is an organismal phylogeny (Burki, Okamoto, et al. 2012; Keeling and Palmer 2008) of the major eukaryotic lineages with emphasis on plastid bearing clades. Horizontal arrows indicate putative horizontal gene transfers from Bacteria or Archaea, with the most likely “donor clade” indicated. The horizontal gene transfer of a KMO gene from a proteobacterium into Sphaeroforma arctica (Ichthyosporea, Opisthokonta) is omitted for clarity. Abbreviated enzymes names are AFM, arylformamidase; AO, aspartate oxidase; KMO, kynurenine 3-monooxygenase; NADS, NAD⁺ synthase; QPRT, nicotinate-nucleotide pyrophosphorylase (quinolinate phosphoribosyltransferase); QS, quinolinate synthase; SAR, Stramenopiles, Alveolates, and Rhizaria; TDO, tryptophan dioxygenase. Black indicates kynurenine (and converged) pathway, gray indicates aspartate pathway.