paralogies with the region on Hsa8 (Fig. 1D). According to our LETTER model (Fig. 1E), an ancestral Asip/Agrp precursor was duplicated two times during the two rounds of vertebrate ge- The teleost agouti-related protein 2 nome duplication (R1 and R2). After R2, Asip and Agrp were gene is an ohnolog gone missing retained in all bony vertebrates, but a third ohnolog (4) went missing in the tetrapod lineage. This gene was retained, however, from the tetrapod genome in the rayfin fish lineage, giving rise to agrp2 and asip2 after the TSGD. A central question in evolutionary genomics is how morpholog- Our results support a revision of agouti family nomenclature ical differences between taxa arise from differences in gene (Fig. 1E). Importantly, our model suggests that the genetic repertoire. Answers require rigorous orthology assignments. basis of teleost background adaptation may not root in the agouti Recent work (1) illuminated the mechanism of background ad- family expansion during the TSGD but to R2, far deeper in aptation in teleost fishes by investigating the function of agouti- vertebrate evolution (Fig. 1E). It is unlikely that zebrafish agrp2 related protein 2 (agrp2), a gene so far found only in teleosts. (new name: asip2b) newly evolved its single pineal-specific Agouti family proteins antagonize melanocortin receptors, function after the TSGD (as suggested by ref. 1), because asip2 function in energy homeostasis, and regulate pigmentation. (asip2a) is absent in zebrafish but present in other teleosts. The authors (1) proposed a neurocrine axis in which zebrafish The interesting findings of Zhang et al. (1) should, thus, spur Agrp2 is expressed exclusively in the pineal gland, where it research on the function of asip2a/b in other teleosts and binds to the melanocortin 1 receptor and thereby, controls ex- agouti in basal rayfin fishes, basal lobefins like lungfish pression of melanin-concentrating hormones; this causes mela- and coelacanth, background-adapting tetrapods, and more nosome contraction in melanophores and adaptation to light basal vertebrates in the light of our rigorous phylogenetic backgrounds (1). These findings intriguingly implied that the framework (Fig. 1E). acquisition of agrp2 in teleosts enabled the evolution of back- ground adaptation by regulation through the pineal. ACKNOWLEDGMENTS. This work was supported by a Feodor Lynen Fellowship of the Alexander von Humboldt Society (to I.B.) and National Tetrapods have two agouti family members [Agouti signaling Center for Research Resources at National Institutes of Health Grant protein (Asip) and Agrp], but teleosts have four (asip1, asip2, R01RR020833 (to J.P.). agrp1, and agrp2). Published studies (1, 2) vaguely suggest that Ingo Braasch1 and John H. Postlethwait fi agrp2 is a paralog of agrp1 generated during the teleost-speci c Institute of Neuroscience, University of Oregon, Eugene, OR 97403- genome duplication (TSGD). Phylogenetic analyses, however, 1254 remain ambiguous regarding vertebrate agouti gene relation- 1. Zhang C, et al. (2010) Pineal-specific agouti protein regulates teleost background ad- ships (2) (Fig. 1A). aptation. Proc Natl Acad Sci USA 107:20164–20171. Our investigation of this gene family using synteny data 2. Kurokawa T, Murashita K, Uji S (2006) Characterization and tissue distribution of multiple agouti-family genes in pufferfish, Takifugu rubripes. Peptides 27:3165–3175. clearly reveals that the teleost agrp2 chromosomal region shares 3. Nakatani Y, Takeda H, Kohara Y, Morishita S (2007) Reconstruction of the vertebrate syntenies neither with the teleost agrp1 region nor with the ancestral genome reveals dynamic genome reorganization in early vertebrates. Ge- – Agrp B nome Res 17:1254 1265. tetrapod region (Fig. 1 ), contradicting the current gene 4. Wolfe K (2000) Robustness—it’s not where you think it is. Nat Genet 25:3–4. nomenclature (2). In contrast, teleost agrp2 and asip2 regions 5. Catchen JM, Conery JS, Postlethwait JH (2009) Automated identification of conserved – show conserved synteny to a region on human 8 synteny after whole-genome duplication. Genome Res 19:1497 1505. (Hsa8) (Fig. 1C). Furthermore, we find at least three regions of paralogy in the (Fig. 1D)—Hsa16 (AGRP Author contributions: I.B. and J.H.P. designed research; I.B. performed research; I.B. and region), Hsa8, and Hsa20 (ASIP region)—most likely derived J.H.P. analyzed data; and I.B. and J.H.P. wrote the paper. from a single Asip/Agrp region on ancestral vertebrate proto- The authors declare no conflict of interest. chromosome B (3). The ASIP region on Hsa20 shares more 1To whom correspondence should be addressed. E-mail: [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1101594108 PNAS | March 29, 2011 | vol. 108 | no. 13 | E47–E48 Downloaded by guest on September 27, 2021 A 99 Fugu scaffold446 NM_001098666 E 99 Tetraodon chrUn ENSTNIG00000001575 Stickleback chrXVII contig7384 R1 R2 TSGD 58 Asip1 Zebrafish chr6 ENSDARG00000077858

Xenopus scaffold38 100 teleost asip1 asip1 99 Human chr20 NM_001672

Mouse chr2 NM_015770 loss Asip 71 Chicken chr20 NM_001115079 tetrapod Asip Asip1 98 Fugu scaffold427 NM_001098654

95 Tetraodon chrUn ENSTNIG00000001285 teleost asip2 asip2a Stickleback chrXXI contig5615 Asip2 61 Medaka chr20 BJ487471 teleost agrp2 asip2b

96 Fugu scaffold166 NM_001098656 52 tetrapod OGM Tetraodon chr15 ENSTNIG00000007827 proto- 98 Medaka chr17 AM300511 Asip/Agrp teleost agrp1 agrp

Stickleback chrIII contig5451 Agrp2 89 loss Zebrafish chr2 XM_001334910

99 Fugu scaffold174 NM_001098655 tetrapod Agrp Agrp 91 Tetraodon chr5 ENSTNIG00000010225 95 Stickleback chrII ENSGACG00000017661 68 Medaka chr3 ENSORLG00000000398 Agrp1

Zebrafish chr6 ENSDARG00000069089 bony vertebrate OGM

Xenopus scaffold283 XM_002937321 94 98 Chicken chr11 NM_001031457

Human chr16 NM_001138 Agrp 86 Mouse chr8 NM_007427

0.5

B CDX 1 agrp2 1 2 2 Y 3 3 1 4 4 2 5 5 3 6 6 4 7 7 5 agrp1 8 8 6 9 9 7 10 10 8 11 11 9 ‘ASIP2 region’ 12 12 10 13 13 11 14 14 12 15 15

Medaka 13 16 Human Chromosomes Zebrafish Chromosomes 16 agrp2 17 14 17 18 ripk2 15 18 19 16 19 asip2 20 17 AGRP 20 21 snx16 wwp1 18 21 22 19 22 23 23 20 24 ASIP 24 21 25 SNX16 WWP1 Hsa8 22 Hsa16 RIPK2 AGRP 30Mb 40Mb 60Mb 70Mb 60Mb 80Mb 100Mb Hsa20 Paralogs Hsa16 Orthologs Hsa8 Orthologs

Fig. 1. Evolution of vertebrate agouti genes. (A) Maximum likelihood phylogeny of vertebrate agouti family proteins with current names (JTT+G+I model; 100 bootstrap replicates). (B–D) Conserved synteny dot plots derived from the Synteny Database (5). (B) The zebrafish agrp2 region on Dre2 (red box) shares conserved syntenies neither with the zebrafish agrp1 region (Dre7) nor with the human AGRP region (Hsa16). (C) The agrp2 and asip2 regions in medaka and other teleosts share conserved synteny with each other and with a region on human Hsa8, including several agrp2- and asip2-neighboring genes. (D) Analysis of the human genome shows that the ASIP region on Hsa20 shows more paralogous connections to the inferred ASIP2 region on Hsa8 than to the AGRP region on Hsa16 (43 vs. 25 genes, respectively). (E) Model for the evolution of the vertebrate agouti family by three rounds of genome duplication, including a revised gene nomenclature. The pineal-specific function of teleost agrp2 (new name: asip2b) could have evolved anywhere along the thick red line. OGM, ohnolog gone missing.

E48 | www.pnas.org/cgi/doi/10.1073/pnas.1101594108 Braasch and H. Postlethwait Downloaded by guest on September 27, 2021