REVIEW in Choanoflagellates RICHARD P. TUCKER* Department of Cell Biology and Human Anatomy, University of California, Davis, Davis, California

ABSTRACT Horizontal gene transfer (HGT), also known as lateral gene transfer, results in the rapid acquisition of genes from another organism. HGT has long been known to be a driving force in speciation in prokaryotes, and there is evidence for HGT from symbiotic and infectious to metazoans, as well as from to bacteria. Recently, it has become clear that as many as a 1,000 genes in the genome of the choanoflagellate Monosiga brevicollis may have been acquired by HGT. Interestingly, these genes reportedly come from , bacteria, and other choanoflagellate prey. Some of these genes appear to have allowed an ancestral choanoflagellate to exploit nutrient-poor environments and were not passed on to metazoan descendents. However, some of these genes are also found in genomes, suggesting that HGT into a common ancestor of choanozoans and may have contributed to metazoan . J. Exp. Zool. (Mol. Dev. Evol.) 320B:1–9, 2013. © 2012 Wiley Periodicals, Inc. J. Exp. Zool. (Mol. Dev. Evol.) How to cite this article: Tucker RP. 2013. Horizontal gene transfer in choanoflagellates. J. Exp. 320B:1–9, 2013 Zool. (Mol. Dev. Evol.) 320B:1–9.

A key event that can contribute to the evolution of a is the these are intracellular that have been infecting protists acquisition of new genetic material. This can take a number of for a very long time. Nevertheless, it points to the potential for the different forms ranging from point mutations to chromosomal two-way acquisition of potentially beneficial genes across recombination and rearrangements to whole genome duplica- kingdoms. tions; the inheritance of, and selection for, the novel genetic Choanoflagellates are now widely considered to be the most material derived from pre-existing genes can lead to distinctive closely related group to metazoans (Fig. 1A; King and Carroll, traits that are the hallmark of a novel species. In prokaryotes it has 2001; Philippe et al., 2004; Abedin and King, 2008; King et al., been recognized for decades that new genetic material in the form 2008, 2009). Intriguingly, examination of choanoflagellate genes of complete, novel genes can be acquired through horizontal indicates that HGT is particularly common in these single-celled gene transfer (HGT; also referred to a lateral gene transfer; organisms, with genes appearing to arise from prokaryotic and Lawrence, '99; de la Cruz and Davies, 2000). This can be a eukaryotic prey caught in the mucous net arrayed between relatively common event. Early studies predicted that between 8% microvilli at the base of their flagellum (Fig. 1B–D). Here, I will and 18% of the Escherichia coli genome was acquired by HGT review the recent (and rapidly growing) literature supporting HGT (Ochman and Lawrence, '96), and more recent studies comparing in the choanoflagellate Monosiga and present novel the genomes of different E. coli strains suggest that HGT may even analyses of these genes taking into consideration the number of be more common (e.g., Brzuszkiewicz et al., 2006; reviewed by novel genomes that have become available since the original Dobrindt et al., 2010). research was done. HGT is not limited to prokaryotes. Recently, HGT from bacteria to and nematodes, as well as from bacteria to hydrozoans, has been described (Dunning Hotopp, 2011). For example, approximately 30% of the genome of the proteobacteria Conflict of interest: none to declare. Wolbachia, a parasitic microbe of arthropods, is found in the *Correspondence to: Richard P. Tucker, Department of Cell Biology and genome of the chinensis (Nikoh et al., 2008). Human Anatomy, University of California, Davis, Davis, CA 95616. E-mail: [email protected] There is even evidence of HGT from to bacteria: Received 20 December 2011; Revised 25 May 2012; Accepted 20 August approximately 100 genes in Legionella pneumophila, the bacteria 2012 behind Legionnaire's Disease, appear to have a origin Published online 19 September 2012 in Wiley Online Library (wiley (Lurie-Weinberger et al., 2010). The number of -derived onlinelibrary.com). genes in L. pneumophila is likely to be unusually high given that DOI: 10.1002/jez.b.22480

© 2012 WILEY PERIODICALS, INC. 2 TUCKER

Figure 1. A schematic tree of life showing plausible phylogenetic relationships between choanoflagellates, fungi, and metazoans (A). Choanoflagellates are characterized by a microvilli collar found at the apical end of the cell body (B). A flagellum creates currents that can drive algal and prokaryotic prey into a mucous net arrayed between the microvilli (C). The prey are then phagocytosed, creating the potential for horizontal gene transfer (D).

PHOSPHOFRUCTOKINASE: THE FIRST EXAMPLE OF most similar to PFK sequences encoded in prokaryotic genomes HORIZONTAL GENE TRANSFER INTO CHOANOZOANS (e.g., it shares 62% amino acid identity with PFK from the The first description of HGT between a prokaryote and a photosynthetic acidobacteria Chloracidobacterium thermophi- choanoflagellate is found in Bapteste et al. (2003). This article lum, AEP13132), and among eukaryotic sequences it is most describes the complex evolutionary history of phosphofructoki- similar to PFK from (e.g., 45% amino acid identity with nase (PFK), a highly conserved enzyme that is central to the PFK from Arabidopsis thaliana, AAL90928). Interestingly, the canonical glycolytic pathway. Construction of a phylogenetic tree PFK from M. brevicollis (XP_001742461) is more similar to the based on the PFK sequences from a number of species resulted in PFKs found in fungi and metazoa that the PFK from M. ovata. puzzling relationships, not the least of which was the grouping of When these and other representative sequences are aligned with a PFK from Monosiga ovata with PFKs from the spirochetes ClustalW (http://www.genome.jp/tools/clustalw/) (Pairwise Borrelia brugdorferi and Treponema sp. Such groupings are alignment parameters = gap open penalty:10.0; gap extension typically considered to be evidence of HGT. The authors concluded penalty:0.01; Mulitple alignment parameters = gap open penal- that this unusual relationship argued for HGT between the gram ty:10.0; gap extension penalty:0.05; no weight transition; negative bacteria and the choanoflagellate, and that this was one hydrophobic residues:GPSNDQERK) and the sequences aligning of many cases where PFK and genes encoding related enzymes with the first 243 amino acids of the M. ovata enzyme (which may have passed from one genome to another by HGT. Evidence align with few gaps) are subjected to simple phylogenetic tree for the expression of this PFK in M. ovata is supported by the analysis(UPGMA;fromthe ClustalWsite),itisclearthat thegene presence of ESTs (Table 1). in M. ovata that came from a prokaryote via HGT does not sort Reanalysis of the possible origins of PFK were undertaken by with the homolog from M. brevicollis, and that a gene related to using the default parameters of tblastn (http://blast.ncbi.nlm. the one in M. brevicollis, and not the gene that arose by HGT, was nih.gov/) and searching the NCBI nucleotide collection. As likely to be the gene that ultimately was contributed to the reported by Bapteste et al. (2003), the M. ovata PFK sequence is metazoan genome (Table 2; Fig. 2A).

J. Exp. Zool. (Mol. Dev. Evol.) HORIZONTAL GENE IN 3

Table 1. Representative genes reported to be acquired by choanoflagellates via horizontal gene transfer. Protein name RefSeq/JGI ID Putative origin EST?a Refs. Phosphofructokinase AAQ55175b Bacteria Yes Bapteste et al. (2003) Spo11-3 EDQ87716 Algae No Malik et al. (2007) Top6B XP_001744261 Algae No Malik et al. (2007) Uroporphyrin III methyltransferase XP_001742170 No Maruyama et al. (2009) Cobalamin synthesis protein XP_001746731 Cyanobacteria No Maruyama et al. (2009) Amino acid aminotransferase XP_001749475 Cyanobacteria Yes Maruyama et al. (2009) Glyoxalase I family-like protein XP_001750995 Cyanobacteria No Maruyama et al. (2009) lysA XP_001746610 Bacteria Yes Torruella et al. (2009) Ammonium transporter 1 XP_001743145 Algae No Nedelcu et al. (2009) Glutamine synthetase 32625 Algae Yes Nedelcu et al. (2009) Glutamate synthase 15256 Algae No Nedelcu et al. (2009) thrA 34304 Algae Yes Sun and Huang (2011) ilvD 32689 Algae Yes Sun and Huang (2011) ilvE 32828 Yes Sun and Huang (2011) Teneurin XP_001749414 Diatom/bacteria Yes Tucker et al. (2012)

aAs determined from reference or by tblastn of Monosiga sequence against non-human, non-mouse ESTs limited to Monosiga. bMonosiga ovata (RefSeq/JGI ID numbers not indicated by a superscript b are from Monosiga brevicollis).

Table 2. Analysis of genes reported to be acquired by choanoflagellates via horizontal gene transfer (HGT).

Likely origin HGT gene Pre-existing From Replaces Found with Into homolog into Protein name HGT Prokaryotic homolog homolog Novela metazoa metazoa Phosphofructokinaseb Yes Yes Yesf Yes Spo11-3 Yes Yes Yes Yes Top6B Yes Yes Yes Uroporphyrin III methyltransferase Yes ?c ?e Yes Cobalamin synthesis proteinb Yes Yes Yes Amino acid aminotransferase Yes Yes ? Glyoxalase I family-like protein Yes ?e lysA Yes Yes ? ? ? Ammonium transporter 1 Yes Yes Yes Yes Glutamine synthetaseb Yes Yes Yes Yes Glutamate synthase Yes Yes Yes Yes thrA Yes ? ? Yes ilvD Yes Yes Yes Yes Yes Yes ilvE Yes Yes ? Teneurinb Yes Yesd Yes Yes

aHomologous gene not found in or fungi, suggesting that gene appeared in unikonts via HGT. bPhylogenetic trees are provided elsewhere. cAnalysis inconclusive. dA fusion of domains acquired by HGT and a pre-existing transmembrane protein. ePossible origin from an . fReplaces homolog in Monosiga ovata, but pre-existing gene is found in M. brevicollis.

J. Exp. Zool. (Mol. Dev. Evol.) 4 TUCKER

Figure 2. Phylogenetic tree analysis of representative proteins proposed to have been acquired in choanozoans by horizontal gene transfer (see text for details). A phosphofructokinase (A)inMonosiga ovata (underlined) clusters with prokaryotic sequences, a hallmark of horizontal gene transfer. However, the phosphofructokinase found in M. brevicollis is found in the same as the homologous enzyme in fungi and metazoans, as one would expect if horizontal gene transfer did not take place. This demonstrates that horizontal gene transfer can often be quite specific in choanozoan lineages. Glutamine synthetase (B) is representative of this group. It appears to have been acquired from algal prey, as it clusters at the bottom of this tree with and plants. The second clade at the top of the tree contains fungi and metazoans. One possible explanation for this arrangement is that the gene was acquired by a choanoflagellate after the evolution of metazoa. M. brevicollis has two cobalamin synthesis proteins (C). One is encoded by a gene that was likely to have been acquired by horizontal gene transfer (“1”; underlined) as it is found in the same clade as proteins from a microscopic alga and a diatom. The second cobalamin synthesis protein (“2”) clusters with fungi and metazoans. In this case, the pre-existing gene was not replaced by the gene that was likely to have been acquired by horizontal gene transfer, and the pre-existing gene was transmitted into metazoa. Three genes that make up the glutamate synthase cycle have been proposed to have been acquired in choanoflagellates by horizontal gene transfer. A fourth protein in M. brevicollis, teneurin (D), has distinct properties. The C-terminal sequences cluster with YD proteins from proteobacteria. The N-terminus of the M. brevicollis teneurin has EGF repeats, which are unique to eukaryotes. Thus, the gene encoding teneurin appears to be a hybrid of prokaryotic and pre-existing genes, and the resulting gene was transmitted into metazoa. Full species names and accession numbers can be found in the text.

J. Exp. Zool. (Mol. Dev. Evol.) HORIZONTAL GENE IN CHOANOFLAGELLATES 5

REPORTS OF THE TRANSFER OF GENETIC MATERIAL so the authors speculate that the acquired genes were functionally FROM PLANTS TO CHOANOFLAGELLATES superior and may have allowed a choanoflagellate to exploit The possibility of HGT between and Monosiga was nitrogen-poor environments. The search, alignment and tree first suggested by Malik et al. (2007) who studied phylogenetic analyses described above confirm this report (Table 2; Fig. 2B). For relationships between Spo11, a protein that generates breaks in example, the M. brevicollis glutamine synthetase is most similar to DNA during , and its paralogs. They found Spo11-1 in this enzyme in the diatom (58% amino acid , plants and many prokaryotes, and Spo11-2 in plants identity, DAA12505) as well as to various plants. A homolog is and many prokaryotes. Spo11-3 and the related topoisomerase found in fungi (e.g., 54% amino acid identity, Clavispora Top6B, in contrast, are found not only in plants and prokaryotes, lusitaniae, XP_002619296) and in cnidarians, , and but also in M. brevicollis. Both proteins are missing from other (e.g., 50% amino acid identity, Hydra magnipapillata, opisthokonts. The authors conclude that HGT between algal prey XP_002161673). Phylogenetic tree analysis suggests the follow- and choanoflagellate predator is the most likely explanation for ing: a glutamine synthetase gene was acquired by an ancestor of the presence of these genes in M. brevicollis. This was confirmed M. brevicollis from a diatom, and it replaced the existing gene. by our analysis: Spo11-3 from M. brevicollis is most similar to However, it was the glutamine synthetase gene found in fungi that homologs from the green algae Volvox carteri (60% amino acid persisted in the that evolved into metazoa, not the gene identity, XP_002947507) and the marine diatom Thalassiosira acquired by HGT (Fig. 2B). pseudonana (60% amino acid identity, XP_002294481), and the Amino acid biosynthesis pathways in M. brevicollis are reported M. brevicollis Top6B is most similar to the Top6B subunits of A. to have evolved from HGT as well (Sun and Huang, 2011). An thaliana (47% amino acid identity, NP_188714) and V. carteri enzyme common to the threonine/methionine pathway, thrA, (46% amino acid identity, XP_002952053). Neither protein has appears to have an algal origin, as does ilvD, an enzyme used for homologs in amoebozoa, fungi, or metazoa (though, as noted isoleucine biosynthesis. Another isoleucine synthesis enzyme, above, the paralog Spo11-1 is found widely). ilvE, appears to have come from a diatom. Expression of all of Three enzymes found in M. brevicollis that form the glutamate these enzymes is supported by ESTs (Table 1). Our tblastn, synthase cycle—ammonium transporter (AMT), glutamine syn- alignment and phylogenetic tree analyses support these con- thetase and glutamate synthase—are all reported to be of algal clusions, with a few notable exceptions. One such exception is the origin (Fig. 3; Nedelcu et al., 2009). This is remarkable for a possible origin of the M. brevicollis thrA, which is most similar to a number of reasons. First, an entire enzymatic pathway is proposed thrA from a chloroflex bacteria (44% amino acid identity, to have been incorporated into the Monosiga genome, and the fact Caldilinea aerophila, BAL98984), and clusters with this prokary- that the genes are not clustered in any way suggests that they were otic enzyme by phylogenetic tree analysis. However, one must probably incorporated following separate HGT events. Second, note that the choanozoan thrA is also quite similar to plant such a pathway already existed in an ancestral choanoflagellate, homologs, so an origin from plants remains possible. In addition, the ilvD from M. brevicollis appears to have a prokaryotic origin: it shares 47% amino acid identity with the ilvD from Staphylococcus saprophyticus (BAE19438) and other bacteria, and phylogenetic tree analysis shows only distant relationships with plants. The ilvD from M. brevicollis is particularly remarkable in that it is most similar to metazoan sequences (e.g., 53% amino acid identity with the homolog from the Amphimedon queenslandica, XP_003384343). It appears to be a rare example of a gene acquired by HGT persisting and perhaps playing a role in the evolution of metazoa.

REPORTS OF THE TRANSFER OF GENETIC MATERIAL FROM PROKARYOTES TO CHOANOFLAGELLATES Torruella et al. (2009) make a strong argument that the lysA gene Figure 3. Key enzymes in the glutamate synthase cycle in M. brevicollis was acquired from bacteria. Lysine is an (ammonium transporter 1, glutamine synthetase and glutamine essential amino acid that must be part of the animal diet, but synthase) were all acquired by Monosiga brevicollis through some organisms have pathways for lysine biosynthesis: the DAP horizontal gene transfer from algal prey (adapted from Nedelcu pathway requires the enzyme lysA, and the AAA pathway et al., 2009). These enzymes may have functioned more efficiently requires alpha-aminoadipate reductase (AAR). The DAP pathway than innate enzymes, thus allowing an ancestral choanoflagellate is found in bacteria and plants, whereas the AAA pathway is to exploit nitrogen-poor environments. found in fungi. Torruella et al. (2009) report that M. brevicollis

J. Exp. Zool. (Mol. Dev. Evol.) 6 TUCKER has both an AAR gene and a lysA gene, and phylogenomics probably did arise by HGT, replaced the existing gene, and was suggests that the lysA genewasacquiredbyHGTfrom not transmitted into metazoa. proteobacteria. Its expression is also confirmed by ESTs. Analysis A different story emerges from the analysis of cobalamin here confirms these conclusions. The gene encoding LysA in synthesis protein. This protein in M. brevicollis is most similar to a many prokaryotes is fused with the gene encoding LysC. tblastn homolog in V. carteri (58% amino acid identity, XP_002957516), searches of the nucleotide collection with the lysA sequence from microscopic marine algae (e.g. Micromonas sp. and Ostreococcus M. brevicollis reveals that the choanozoan sequence is most tauri) and diatoms (Phaeodactylum tricornutum), suggesting an similar to the part of the fused prokaryotic protein that encodes origin not from cyanobacteria, but via HGT from algae. LysA (e.g., 46% identity, Legionella longbeachae, CBJ11456). Interestingly, limiting the tblastn search to opisthokonts reveals Phylogenetic tree analysis also shows the M. brevicollis sequence a second cobalamin synthesis protein in Monosiga brevicollis clustering with the bacterial sequences. Numerous enzymes (40% amino acid identity, XP_001746159), as well as homologs sharing structural similarity to lysA are found in fungi and in fungi and metazoa. Phylogenetic tree construction (Fig. 2C) metazoans, and the phylogenetic relationships between them are shows that this second protein has more traditional evolutionary complex (e.g., see Sandmeier et al., '94). This complicates this origins and is probably encoded on the gene that was eventually particular analysis. transmitted into animals. Note that a recent study describing phylogenomic screening of Like the cobalamin synthesis protein described above, the the M. brevicollis genome using PhyloGenie and Darkhorse has amino acid transferase of M. brevicollis also appears to have arisen clarified the possible extent of HGT in choanoflagellates (Sun by HGT from plants: it is most similar to homologs in the primitive et al., 2010). These programs predicted that 1,324 genes in M. lycophyte Selaginella moellendorffii (47% amino acid identity, brevicollis were likely to have been acquired from other XP_002982481) and the green algae V. carteri (45% amino acid organisms: 53 of these genes appear to have arisen from HGT identity, XP_002948580), and less similar to this protein in from diatoms or , and the remaining came from algae cyanobacteria (e.g., 36% amino acid identity, Cyanothece sp., and bacteria. The authors conclude that HGT in Monosiga from its ACL43948). Limiting the search to opisthokonta reveals a distant prey is widespread, and that the mechanism of prey gene homolog in fungi but no homolog in metazoa. Tree analysis integration into the genome (which is currently unknown) merits clusters the M. brevicollis protein with the plant proteins and puts further investigation. the fungal proteins into a different clade. Thus, the amino acid transferase appears to have evolved in choanozoa by HGT from a FROM CYANOBACTERIA OR PLANTS? FOUR STORIES eukaryote, and while the gene encoding this protein persisted into FROM FOUR ENZYMES fungi (and perhaps into choanozoa, only to be replaced by the An analysis of the genomes of eukaryotes lacking plastids for novel gene), it did not find its way into the metazoan genomes that genes of cyanobacterial ancestry led Maruyama et al. (2009) to are available for analysis. propose four prokaryote-derived genes in M. brevicollis: Phylogenomic analysis of the glyoxalase I family-like protein uroporphyrin III C-methyltransferase, cobalamin synthesis reveals that is it most similar to a homolog from the oomycete protein, an amino acid aminotransferase and a glyoxalase I Phytophthora infestans (48% amino acid identity, family-like protein (Table 1). These four genes were further XP_002901811), as well as to a homolog from the percolozoan analyzed here as described above with surprising results. Naegleria gruberi (46% amino acid identity, XP_002683444). For example, the M. brevicollis uroporphyrin III C-methyltrans- More dissimilar homologs are found in plants and prokaryotes, but ferase gene is most similar to a homologous gene from the no homologs are found in amoebozoa, fungi or metazoa. This oomycete Phyophthora infestans (50% amino acid identity, suggests that while the gene was likely to have been acquired by XP_002998208). Phylogenetic tree analysis shows the choano- HGT, its origins are difficult to determine. Perhaps, like zoan enzyme clustering with both the P. infestans sequence and uroporphyrin III C-methyltransferase, the gene originated from the sequences from several genera of cyanobacteria, while an oomycete or related algae. another clade contains fungal sequences (e.g., 32% amino acid identity, Schizosaccharomyces japonica, XP_002172446) and a TENEURIN: A GENE ENCODING A TRANSMEMBRANE homologous enzyme from the anthozoan Nematostella vectensis PROTEIN REPORTED TO BE A FUSION OF SEQUENCES (but not other metazoans). These complicated relationships ACQUIRED BY HGT AND AN EXISTING (Table 2) make it difficult to ascertain the origins of this enzyme GENE in M. brevicollis: did it come from HGT following a Teneurins are multi- transmembrane proteins necessary meal, or did it arise from a cyanobacterial gene that is closely for the normal development of the nervous system and pattern related to the gene in a heterokont? Analysis is further hampered formation in both protostomes and (Fig. 4A; Tucker by the challenging of these organisms. In any event, and Chiquet-Ehrismann, 2006; Young and Leamey, 2009). the uroporphyrin III C-methyltransferase gene in M. brevicollis Recently, a teneurin was identified in the genome of M. brevicollis

J. Exp. Zool. (Mol. Dev. Evol.) HORIZONTAL GENE IN CHOANOFLAGELLATES 7

Figure 4. Teneurins are transmembrane cell–cell adhesion proteins necessary for the normal development of the nervous system in a broad range of metazoans. In vertebrates, teneurins have a transmembrane domain (indicated to two short vertical bars), eight EGF repeats, a C-rich region, NHL domains, a series of YD repeats, and an RHS protein domain (A). Monosiga brevicollis has a teneurin gene encoding a protein with the same domain organization as vertebrate teneurins (B). Most of the extracellular domain of the M. brevicollis teneurin is encoded on a single huge exon. This is characteristic of a gene acquired by horizontal gene transfer. The extracellular domain of the M. brevicollis teneurin is most similar to the C-terminal half of a YD protein from the aquatic bacterium Desulfurivibrio alkaliphilus (C). Other YD proteins from proteobacteria that are also similar to teneurins have different N-terminal domains, demonstrating that the C-terminal sequences are used like a cassette to create novel genes in prokaryotes. Some of the YD proteins, like teneurins, are transmembrane proteins. The acquisition of a teneurin by an ancestral choanoflagellate may have increased its ability to catch prokaryotic prey through teneurin–YD protein interactions.

J. Exp. Zool. (Mol. Dev. Evol.) 8 TUCKER

(Fig. 4B; Tucker et al., 2012). Most of the extracellular domain of ACKNOWLEDGMENTS the choanoflagellate teneurin appears to have evolved through The author would like to thank Ruth Chiquet-Ehrismann and HGT: it is most similar to YD proteins found in the cell walls and Josephine Adams for helpful discussions, and John Hess for periplasmic space of certain aquatic bacteria, and it is encoded on reviewing the manuscript. a single, huge exon (Fig. 4B and C; Tucker et al., 2012). Since the extracellular domain of teneurins can mediate homotypic cell–cell LITERATURE CITED adhesion, the acquisition of a teneurin by an ancestral choano- Abedin M, King N. 2008. The premetazoan ancestry of cadherins. flagellate may have improved “fishing” for bacteria expressing Science 319:946–948. cell surface YD proteins. A similar hypothesis was first proposed Bapteste E, Moreira D, Philippe H. 2003. Rampant horizontal gene by Abedin and King (2008) to explain the appearance of cadherins transfer and phospho-donor change in the evolution of the in the choanoflagellate mucous net. phosphofructokinase. Gene 318:185–191. tblastn with the 2000 C-terminal amino acids of the M. Brzuszkiewicz E, Brüggemann H, Liesegang H, et al. 2006. How to brevicollis teneurin (i.e., the NHL, YD, and RHS protein domains) become a uropathogen: Comparative genomic analysis of extra- reveals that this part of the teneurin sequence is most similar to intestinal pathogenic Escherichia coli strains. Proc Natl Acad Sci strings of identical domains found in predicted YD proteins from USA 103:12879–12884. four delta-proteobacteria (30% amino acid identity, Desulfurivi- de la Cruz F, Davies J. 2000. Horizontal gene transfer and the brio alkaliphilus, ADH85039; 28% amino acid identity, Anaero- origin of species: Lessons from bacteria. Trends Microbiol 8:128– myxobacter sp., ACG72557; 27% identity, Desulfococcus 133. oleovorans, ABW66652; 27% amino acid identity, Syntropho- Dobrindt U, Chowdary MG, Krumbholz G, Hacker J. 2010. Genome bacter fumaroxidans, ABK19729) and one from a gamma- dynamics and its impact on evolution of Escherichia coli. Med proteobacterium (28% identity, Nitrosococcus halophilus, Microbiol Immunol 199:145–154. ADE15009). When the domains of these YD proteins are mapped Dunning Hotopp JC. 2011. Horizontal gene transfer between bacteria with Pfam (http://pfam.sanger.ac.uk/), it is clear that the region N- and animals. Trends Genet 27:157–163. terminal to the NHL domains is different in each of the five YD Jackson AP, Thomas GH, Parkhill J, Thomson NR. 2009. Evolutionary proteins: two, like teneurin, are type 2 transmembrane proteins, diversification of an ancient gene family (rhs) through C-terminal while the others are likely to be secreted into the periplasmic space displacement. BMC Genomics 10:584. (Fig. 4C). Thus, it appears that the conserved C-terminal sequences King N, Carroll SB. 2001. A receptor tyrosine kinase from were inserted into existing genes to create novel genes. Such an choanoflagellates: Molecular insights into early animal evolution. insertion was most likely regulated by the RHS protein domain Proc Natl Acad Sci USA 98:15032–15037. (RHS stands for “rearrangement hot spot”), as this domain has King N, Westbrook MJ, Young SL, et al. 2008. The genome of the been implicated in genome rearrangements and HGT in bacteria choanoflagellate Monosiga brevicollis and the origin of metazoans. (reviewed by Jackson et al., 2009). Thus, it appears that the Nature 451:783–788. teneurin of M. brevicollis, which persisted in metazoa where it is King N, Young SL, Abedin M, Carr M, Leadbeater BS. 2009. indispensable for normal development of the nervous system, is a The choanoflagellates: Heterotrophic nanoflagellates and sister hybrid of an RHS protein domain cassette derived from a group of the metazoa. Cold Spring Harbor Prot 2:pdb. bacterium and a pre-existing transmembrane protein that emo116. featured a string of EGF repeats. Lawrence JG. 1999. Gene transfer, speciation, and the evolution of bacterial genomes. Curr Opinion Microbiol 2:519–523. CONCLUSIONS Lurie-Weinberger MN, Gomez-Valero L, Merault N, et al. 2010. The The acquisition of new genetic material by HGT may account origins of eukaryotic-like proteins in Legionella pneumophila. Int J for over 1,000 genes in modern choanoflagellates. These genes Med Microbiol 300:470–481. originated from bacteria, algae, diatoms, and other prey species, Malik SB, Ramesh MA, Hulstrand AM, Logsdon JM Jr. 2007. Protist perhaps including . The potential impact of HGT on homologs of the meiotic Spo11 gene and topoisomerase VI reveal an speciation in these organisms is nicely illustrated by the evolutionary history of gene duplication and lineage-specific loss. acquisition of entire enzymatic pathways from algae, with Mol Biol Evol 24:2827–2841. individual components that were able to replace or supplement Maruyama S, Matsuzaki M, Misawa K, Nozaki H. 2009. Cyanobacterial pre-existing choanoflagellate genes, and by the evolution of novel contribution to the genomes of the plastid-lacking protists. BMC transmembrane proteins like teneurins. However, of the prey-to- Evol Biol 9:197. predator acquired genes examined to date, only a few appear to Nedelcu AM, Blakney AJ, Logue KD. 2009. Functional replacement of a have persisted in the metazoan descendants of choanoflagellates, primary metabolic pathway via multiple independent eukaryote-to- where they would have had the potential to contribute to the eukaryote gene transfers and selective retention. J Evol Biol evolution of multicellularity and histogenesis in animals. 22:1882–1894.

J. Exp. Zool. (Mol. Dev. Evol.) HORIZONTAL GENE IN CHOANOFLAGELLATES 9

Nikoh N, Tanaka K, Shibata F, et al. 2008. Wolbachia genome Sun G, Huang J. 2011. Horizontally acquired DAP pathway as a unit of integrated in an chromosome: Evolution and fate of laterally self-regulation. J Evol Biol 24:587–595. transferred endosymbiont genes. Genome Res 18:272–280. Sun G, Yang Z, Ishwar A, Huang J. 2010. Algal genes in the closest Ochman H, Lawrence JG. 1996. and the amelioration of relatives of animals. Mol Biol Evol 27:2879–2889. bacterial genomes. In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin Torruella G, Suga H, Riutort M, Peretó J, Ruiz-Trillo I. 2009. The ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, evolutionary history of lysine biosynthesis pathways within Umbarger HE, editors. Escherichia coli and Salmonella typhimu- eukaryotes. J Mol Evol 69:240–248. rium: cellular and molecular biology. 2nd edition. Washington, DC: Tucker RP, Beckman J, Leachman NT, Schöler J, Chiquet-Ehrismann R. American Society for . p 2627–2637. 2012. Phylogenetic analysis of the teneurins: Conserved features Philippe H, Snell EA, Bapteste E, et al. 2004. Phylogenomics of and premetazoan ancestry. Mol Biol Evol 29:1019–1029. eukaryotes: Impact of missing data on large alignments. Mol Biol Tucker RP, Chiquet-Ehrismann R. 2006. Teneurins: A conserved family Evol 21:1740–1752. of transmembrane proteins involved in intercellular signaling during Sandmeier E, Hale TI, Christen P. 1994. Multiple evolutionary origin of development. Dev Biol 290:237–245. pyridoxal-5'-phosphate-dependent amino acid decarboxylases. Eur Young TR, Leamey CA. 2009. Teneurins: Important regulators of neural J Biochem 221(3): 997–1002. circuitry. Int J Biochem Cell Biol 41:990–993.

J. Exp. Zool. (Mol. Dev. Evol.)