Gene Transfer from Bacteria and Archaea Facilitated Evolution of an Extremophilic Eukaryote

See allHide authors and affiliations

Science  08 Mar 2013:
Vol. 339, Issue 6124, pp. 1207-1210
DOI: 10.1126/science.1231707

Hot, Toxic Eukaryote

Unusually, the single-celled eukaryote red alga, Galdieria sulphuraria, can thrive in hot, acidic springs. This organism is endowed with extraordinary metabolic talents and can consume a variety of strange carbohydrates, as well as turn on photosynthesis when the food runs out. Schönknecht et al. (p. 1207; see the Perspective by Rocha) discerned from phylogenetic analysis of its genome that during its evolution, G. sulphuraria appears to have commandeered at least 75 bacterial and archaeal genes by horizontal gene transfer and then applied gene expansion to boost its metabolic repertoire.


Some microbial eukaryotes, such as the extremophilic red alga Galdieria sulphuraria, live in hot, toxic metal-rich, acidic environments. To elucidate the underlying molecular mechanisms of adaptation, we sequenced the 13.7-megabase genome of G. sulphuraria. This alga shows an enormous metabolic flexibility, growing either photoautotrophically or heterotrophically on more than 50 carbon sources. Environmental adaptation seems to have been facilitated by horizontal gene transfer from various bacteria and archaea, often followed by gene family expansion. At least 5% of protein-coding genes of G. sulphuraria were probably acquired horizontally. These proteins are involved in ecologically important processes ranging from heavy-metal detoxification to glycerol uptake and metabolism. Thus, our findings show that a pan-domain gene pool has facilitated environmental adaptation in this unicellular eukaryote.

Although bacteria and archaea usually dominate extreme environments, hot and extremely acidic habitats are typically devoid of photosynthetic bacteria. Instead, eukaryotic unicellular red algae of the Cyanidiophyceae are the principal photosynthetic organisms in these ecological niches (1). Cyanidiophyceae can grow at pH 0 to 4 and temperatures up to 56°C, close to the upper temperature limit for eukaryotic life (2). Galdieria sulphuraria is a unique member of the Cyanidiophyceae, displaying high salt and metal tolerance and exhibiting extensive metabolic versatility (3, 4). G. sulphuraria naturally inhabits volcanic hot sulfur springs, solfatara soils, and anthropogenic hostile environments. In habitats with high concentrations of arsenic, aluminum, cadmium, mercury, and other toxic metals, G. sulphuraria frequently represents up to 90% of total biomass and almost all the eukaryotic biomass (1, 5).

To understand the molecular mechanisms underlying G. sulphuraria's extremophilic and metabolically flexible lifestyle (Fig. 1), we determined its genome sequence (13.7 Mb; table S1) (6). The only member of the Cyanidiophyceae for which a genome sequence was previously available, Cyanidioschyzon merolae (7), diverged from G. sulphuraria about 1 billion years ago, which approximates the evolutionary distance between fruit flies and humans (see fig. S1 and supplementary materials). C. merolae maintains a strictly photoautotrophic lifestyle and does not tolerate high salt or metal concentrations; it differs markedly from G. sulphuraria in ecology, cell biology, and physiology. Accordingly, we find orthologs for only 42% of the 6623 G. sulphuraria proteins in C. merolae, and only 25% of both genomes constitute syntenic blocks (fig. S2). Coding sequences make up 77.5% of the G. sulphuraria genome, resulting in a median intergenic distance of 20 base pairs (bp) (fig. S3). Protein-coding genes contain on average two introns (fig. S4), with median lengths of 55 bp (fig. S5). Thus, the G. sulphuraria genome is highly condensed by comparison with that of C. merolae and most other eukaryotes.

Fig. 1

Photoautotrophic (left) and heterotrophic (right) G. sulphuraria cells. Cell cultures (top) and light microscopic images (bottom; bar represents 10 μm) of G. sulphuraria cells grown under continuous illumination in the absence of glucose (left) or in darkness in the presence of 200 mM glucose (right).

Eukaryotic innovations usually arise through gene duplications and neofunctionalizations, which lead to expansion of existing gene families (8). In contrast, archaea and bacteria commonly adapt through horizontal gene transfer (HGT) from other lineages (9). HGT has also been observed in some unicellular eukaryotes (10); however, to our knowledge, horizontally acquired genes have not been linked to fitness-relevant traits in free-living eukaryotes (11). Phylogenetic analyses of G. sulphuraria genes using highly stringent criteria indicate at least 75 separate gene acquisitions from archaea and bacteria (supplementary materials). The origin of these G. sulphuraria genes from HGT is supported by the finding that compared to the genomic average, they have significantly fewer introns (mean 0.8 versus 2.06, P = 0.0012, Mann-Whitney test; fig. S6), slightly higher GC content (40.6% versus 39.9%, P = 0.0030, Student's t test; fig. S7), and deviating oligonucleotide usage (P = 0.00034, Mann-Whitney test; fig. S7). Gene transfers can be traced to a broad range of donor taxa (fig. S8 and table S4), with a significant enrichment from extremophile bacteria (P = 7.8 × 10−9, Fisher's exact test). The genome of G. sulphuraria thus shows notable contributions from a pan-domain gene pool.

The two largest G. sulphuraria protein families form a monophyletic branch within the so-called archaeal ATPases (adenosine triphosphatases) (Fig. 2A). These soluble ATPases have not been observed in other eukaryotes. Phylogenetic analyses indicate that G. sulphuraria acquired an ancestral ATPase gene from archaea, followed by duplications and diversification into separate families (fig. S11). Genes encoding the different families cluster in pairs on the G. sulphuraria genome (Fig. 2B). Pairs of homologous ATPase genes that are transcribed together are observed in archaeal genomes (12). Although their physiological function is unknown (13), archaeal ATPases may contribute to heat tolerance. We found a correlation between ATPase gene copy number and optimal growth temperature across thermophilic and hyperthermophilic archaea (Fig. 2C). These findings suggest that G. sulphuraria's adaptation to heat may have been facilitated by the acquisition and subsequent expansion of an archaeal ATPase gene family.

Fig. 2

Archaeal ATPases in G. sulphuraria. (A) The phylogeny of archaeal ATPases indicates HGT into G. sulphuraria. The unrooted Bayesian tree is shown with posterior probabilities (full tree and details in fig. S9, alignment of ATPase domains in fig. S10, and detailed tree of families #1a, #1b, #2, and #19 in fig. S11). Major phylogenetic groups are color-coded (Archaea, gray; Cyanobacteria, teal; other Bacteria, olive; Rhodophyta, red; Amoebozoa, purple) and labeled according to Leipe et al. (20). (B) Genes encoding archaeal ATPases in G. sulphuraria form unidirectional clusters in specific combinations; pairs are always formed from different (sub)families, with #1b genes always at the 5′ end. (C) Copy number of archaeal ATPase genes in genomes of thermophilic and hyperthermophilic archaea is correlated with optimum growth temperature (for details, see fig. S12).

G. sulphuraria's tolerance to high salinity is also likely to have been facilitated by HGT. Resistance to salt stress requires the removal of Na+ from the cytosol and an increase of osmolarity using compatible solutes in the cytosol. In addition to several Na+:H+ antiporters of eukaryotic origin, G. sulphuraria encodes two monovalent cation:proton antiporters that appear to have been acquired from bacteria (fig. S13). Furthermore, genes encoding sarcosine dimethylglycine methyltransferase (SDMT) appear to originate from halophilic cyanobacteria (fig. S14). These enzymes allow the production of the compatible solute betaine from glycine (14), which indeed accumulates in G. sulphuraria under salt stress (fig. S15).

How does G. sulphuraria maintain near-neutral cytosolic pH against a 106-fold H+ gradient across its plasma membrane (15, 16)? There is no evidence for an enhanced capacity to pump H+ out of the cytosol (which would be an energetically intense strategy). Yet, there are indications of a reduced H+ permeability of the plasma membrane. For G. sulphuraria, one voltage-gated ion channel gene was identified in the genome, compared to three in C. merolae (table S3) and 16 or more in other unicellular algae. Voltage-gated ion channels allow single-file diffusion of water, and therefore have a very high conductance for protons. A plasma membrane devoid of proton-conducting voltage-gated ion channels probably has a low H+ permeability, preventing the acidification of the cell interior even at very high external H+ concentrations. G. sulphuraria shows an expansion of both, the intracellular chloride channel (CLIC) family and the chloride carrier/channel (ClC) family (table S3), which do not conduct protons.

G. sulphuraria copes with toxic metals typically found in volcanic areas and acid mine drainage by a variety of metal transporters (table S3). Different plasma membrane uptake systems for divalent metal cations permit the selective uptake of essential metals (such as iron or copper) at high concentrations of toxic metals (such as aluminum or cadmium). G. sulphuraria can also neutralize biohazardous metals, making it potentially useful in biotechnological applications. For example, G. sulphuraria arsenite methyltransferases can biotransform arsenic, which is often found at very high concentrations in geothermal environments, into less toxic and possibly gaseous methyl derivatives (17). In addition, two intronless G. sulphuraria genes encode the bacterial arsenical membrane protein pump, ArsB. The sequences most similar to G. sulphuraria ArsB are from thermoacidophilic bacteria (Fig. 3), again indicating a central role for HGT in extremophilic adaptation. Mercury is found at concentrations up to 200 μg/g in soils from which G. sulphuraria has been isolated. G. sulphuraria can reduce cytotoxic Hg2+ into less toxic metallic mercury. The enzyme responsible, mercuric reductase, was also most likely acquired horizontally from Proteobacteria (fig. S17).

Fig. 3

The phylogeny (Bayesian tree) of bacterial arsenical resistance efflux pumps (ArsB) indicates HGT into G. sulphuraria (full tree and details in fig. S16). The curly bracket marks G. sulphuraria and four thermophilic and/or acidophilic bacteria (in olive), which live in the same environment as G. sulphuraria and from which taxa the algal ArsB may have derived.

In total, 5.2% of G. sulphuraria genes encode membrane transport proteins (mostly metabolite transporters), which is more than has been discovered in most other eukaryotes (fig. S18). Gene family expansions (table S3) are found, for example, in sugar porters (fig. S19), amino acid/auxin permeases (fig. S20), and putative acetate transporters (fig. S21); these three protein families are among the 20 largest in G. sulphuraria (table S2). Further major expansions were found in amino acid–polyamine–organocation transporters (fig. S22), glycoside-pentoside-hexuronide:cation symporters, and glycerol uptake facilitators (fig. S23). Together, these metabolite transporters are likely to be essential for the exceptional ability of G. sulphuraria to grow heterotrophically on many different metabolites. In contrast, most other unicellular algae, including C. merolae, are strictly (or almost strictly) photoautotrophic.

Again, phylogenetic analyses indicate that HGT contributed to G. sulphuraria's enormous metabolic flexibility. All acetate permeases (fig. S21) appear to originate from bacteria, whereas some of the amino acid–polyamine–organocation transporters (fig. S22) seem to stem from thermoacidophilic archaea. G. sulphuraria can grow heterotrophically on glycerol as sole carbon source (3) using a family of five glycerol uptake facilitators and a family of three glycerol dehydrogenases; both families apparently originate from HGT (figs. S23 and S24).

Some of G. sulphuraria's metabolic pathways appear to be conserved from a heterotrophic last common eukaryotic ancestor, but subsequently were lost from other photosynthetic eukaryotes. For example, animals (metazoa) use the methylmalonyl–coenzyme A pathway for the degradation of odd-numbered chain fatty acids and leucine. Whereas green plants and C. merolae lack this pathway, it is present in G. sulphuraria, as well as in diatoms and brown algae. Furthermore, green plants, diatoms, and brown algae synthesize the essential nicotinamide adenine dinucleotide precursor quinolinate from aspartate. In contrast, G. sulphuraria and C. merolae produce quinolinate from tryptophan by way of kynurenine, a pathway common to animals and fungi.

G. sulphuraria contains a number of metabolite transporter families that group with fungi and show less similarity to metabolite transporter families from more closely related organisms (figs. S19, S20, and S25). This unexpected phylogenetic trace could be explained by the loss of genes present in the common eukaryotic ancestor from other clades, by HGT from unsequenced bacteria or archaea into both eukaryotic lineages, or by HGT between fungi and G. sulphuraria. We currently cannot distinguish between these alternative hypotheses owing to limited taxon representation in sequence databases.

G. sulphuraria can also survive saprophytically by excreting catabolic enzymes that decompose extracellular organic polymers into small metabolites for uptake by plasma membrane transporters. Proteomics (18) and/or bioinformatics analyses (supplementary materials) indicate excretion of aspartyl proteases, which cleave peptide bonds at acidic pH; β-galactosidases (fig. S26) and glucoamylases, which degrade polysaccharides; and acid phosphatases (fig. S27), which remove phosphate groups from organic molecules. These excreted enzymes lack orthologs in other photosynthetic eukaryotes, but show homology to excreted enzymes encoded by fungi or bacteria. In particular, G. sulphuraria may have acquired acid phosphatases (fig. S27) and some β-galactosidases (fig. S26) from bacteria.

Extensive gene transfer appears to have been key to the genomic evolution of a metabolically versatile, extremophilic, red alga. Numerous proteins acquired through HGT interact with G. sulphuraria's physico-chemical and metabolic environment. Protein families acquired horizontally by G. sulphuraria are 3-fold enriched in membrane transporters (10.5% for HGT families, P = 0.010, Fisher's exact test) and 14-fold enriched in protein families also found in extremophilic bacteria or archaea (86.8% for HGT families, P = 1.5 × 10−22). These findings for G. sulphuraria mirror the results of a previous systematic study, which showed that proteobacterial adaptation relies on the horizontal acquisition of genes that function at the bacteria's interface to the environment (19). Whereas the importance of HGT for evolution of Bacteria and Archaea is well established, adaptation of a eukaryotic extremophile by gene transfer from Bacteria and Archaea is unexpected and shines a new light on the evolution of unicellular eukaryotes.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S27

Tables S1 to S3

Additional Data Table S4

References (2179)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: This work was made possible by NSF grant EF 0332882 (to A.P.M.W.). Partial support came from the Deutsche Forschungsgemeinschaft (DFG) (CRC TR1, IRTG 1525/1 and WE2231/7-1 to A.P.M.W., EXC 1028 to M.J.L. and A.P.M.W., CRC 680 to M.J.L., and a DFG Mercator Fellowship to G.S.). G.S. appreciates support from NSF (MCB 0925298) and the College of Arts and Sciences, Oklahoma State University (OSU). We thank M. Hanikenne and E. Koonin for helpful discussion, and A. Doust and P. Pelser for advice on phylogenetic analyses. We are grateful to B. Sears for introduction to and assistance with CsCl purification of bisbenzamide-treated nuclear DNA. Some of the computing for this project was performed at the OSU High Performance Computing Center. Sequence data have been deposited at DNA Data Bank of Japan, EMBL, and GenBank under accession ADNM00000000 (SRA012465). The version described in this paper is the first version, ADNM01000000.
View Abstract

Navigate This Article