The new work from Shan and via distinct mechanisms, one might ex- and Shaw, D.E. (2012). Cell 149, this issue, colleagues demonstrates the expanding pect that, in combination, they would yield 860–870. reach of molecular dynamics simulations higher levels of activity, further shifting the Sharma, S.V., Bell, D.W., Settleman, J., and Haber, and sets the stage for future investiga- equilibrium toward dimerization. Finally, D.A. (2007). Nat. Rev. Cancer 7, 169–181. tions of EGFR regulation, its oncogenic the prediction that phosphorylation of Shaw, D.E., Maragakis, P., Lindorff-Larsen, K., activation, and inhibitor binding proper- Tyr845 in the EGFR activation loop acti- Piana, S., Dror, R.O., Eastwood, M.P., Bank, J.A., ties. For example, it will be interesting to vates the kinase and stabilizes the EGFR Jumper, J.M., Salmon, J.K., Shan, Y., and Wrig- know what effect inclusion of Mg2+ and dimer may motivate the field to re- gers, W. (2010). Science 330, 341–346. ATP in the molecular dynamics calcula- examine the role of this phosphorylation Sok, J.C., Coppelli, F.M., Thomas, S.M., Lango, tions has on the observed conformational event in EGFR signaling. M.N., Xi, S., Hunt, J.L., Freilino, M.L., Graner, rearrangements. Additionally, the present M.W., Wikstrand, C.J., Bigner, D.D., et al. (2006). simulations do not reveal the transition to Clin. Cancer Res. 12, 5064–5073. the inactive conformation targeted by REFERENCES Stamos, J., Sliwkowski, M.X., and Eigenbrot, C. lapatinib. Though it may be no small (2002). J. Biol. Chem. 277, 46265–46272. Corcoran, R.B., Ebi, H., Turke, A.B., Coffee, E.M., matter to extend the simulation timescale Nishino, M., Cogdill, A.P., Brown, R.D., Pelle, Wood, E.R., Truesdale, A.T., McDonald, O.B., sufficiently to observe this transition, it will P.D., Dias-Santagata, D., Hung, K.E., et al. Yuan, D., Hassell, A., Dickerson, S.H., Ellis, B., be of interest to computationally charac- (2012). Cancer Discov. 2, 227–235. Pennisi, C., Horne, E., Lackey, K., et al. (2004). terize the relative stability of the active, Cancer Res. 64, 6652–6659. Prahallad, A., Sun, C., Huang, S., Di Nicolantonio, disordered, and fully inactive states. F., Salazar, R., Zecchin, D., Beijersbergen, R.L., Yun, C.H., Boggon, T.J., Li, Y., Woo, M.S., Greu- This study also provides a few testable Bardelli, A., and Bernards, R. (2012). Nature 483, lich, H., Meyerson, M., and Eck, M.J. (2007). hypotheses that may move the field 100–103. Cancer Cell 11, 217–227. forward. For instance, if different acti- Shan, Y., Eastwood, M.P., Zhang, X., Kim, E.T., Zhang, X., Gureasko, J., Shen, K., Cole, P.A., and vating mutations stabilize the active state Arkhipov, A., Dror, R.O., Jumper, J., Kuriyan, J., Kuriyan, J. (2006). Cell 125, 1137–1149.

Sibling Rivalry among Paralogs Promotes Evolution of the Human Brain

Chris Tyler-Smith1,* and Yali Xue1 1The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SA, UK *Correspondence: [email protected] DOI 10.1016/j.cell.2012.04.020

Geneticists have long sought to identify the genetic changes that made us human, but pinpointing the functionally relevant changes has been challenging. Two papers in this issue suggest that partial duplication of SRGAP2, producing an incomplete protein that antagonizes the original, contributed to human brain evolution.

Humans differ from the other great apes in the fossil record. Brain size increased outgroup—a species external to the many ways, one of the more obvious being throughout this period, although unevenly chimpanzee-human branch, such as our larger, more complex brains. These (Figure 1A and 1B), but fossil details of gorilla (Scally et al., 2012). But there are human-specific characteristics have internal brain structure are sparse. In a around 20 million genetic changes that evolved in the last 6–7 million years since complementary way, we can catalog the are specific to humans, most of which we split from our common ancestor with genetic differences between humans probably have no functional impact what- chimpanzees and bonobos. We can and chimpanzees from their genome soever. How do we pick out the few func- catalog the phenotypic differences sequences (Chimpanzee Sequencing tionally relevant changes and link them between humans and other great apes, and Analysis Consortium, 2005) and to their phenotypic consequences? Two and in some cases, we can determine identify which differences arose on the papers in this issue of Cell now show one when they evolved, using evidence in human lineage by comparison with an way that this can be accomplished

Cell 149, May 11, 2012 ª2012 Elsevier Inc. 737 (Charrier et al., 2012; Dennis the general population, sug- et al., 2012). The authors char- gests that it is unlikely to play acterize a series of partial any important role (Figure 1E). duplications of the single Interest thus focuses on ancestral SRGAP2 gene. SRGAP2B and C, which are They make a strong case for extremely similar in sequence. the duplicates leading to Their expression patterns are a higher density of spines on also similar, but again there dendritic cells in the brain are simplifying factors: the and contributing to the neote- level of SRGAP2B transcripts nous development character- is low, and it is also absent istic of humans. from some normal indi- How do you sift through 20 viduals (Figure 1E). Thus, the million genetic differences? A main player is SRGAP2C, and good starting point is the its interaction with SRGAP2A. subset of changes that have This cannot have always major effects on protein- been true during the evolution coding genes, which includes of humans; later in this duplications and deletions. A Preview, we will discuss the previous systematic survey more important role that of such copy number changes SRGAP2B must have played in great apes had identified earlier in evolutionary history. 140 events that are specific Previous work on the mouse to the human lineage (Fortna ortholog of SRGAP2A (srGAP2 et al., 2004), including several or Srgap2) had shown that it that are implicated in neuronal Figure 1. Evolutionary Context of SRGAP2 Duplications on the induced filipodia formation in function. These should be en- Human Lineage the developing cortex through (A–D) The vertical axis shows evolutionary time from the present (0, bottom) to riched for functionally impor- 7 million years ago (top). Horizontal dotted lines represent times of SRGAP2 its F-BAR domain; in addition, tant changes. Among them duplications with their uncertainty indicated by gray shading. decreasing the levels of was SRGAP2, which was (A) Brain volume estimates of fossils and living humans, colored according to SRGAP2A reduced axonal shown in mice to regulate genus. and dendritic branching and (B) Timescales of genera thought to include human ancestors. neuronal migration and mor- (C) Sequence of duplication events of SRGAP2 copies. increased the rate of neuronal phology (Guerrier et al., (D) Inferred SRGAP2 activity as a consequence of duplication of antagonistic migration (Guerrier et al., 2009). SRGAP2 was therefore paralogs and the decay of SRGAP2B. 2009). Now, Charrier and col- (E) Levels of protein (schematic) and copy number of paralogs in modern an excellent candidate for humans. leagues (2012) use mouse more detailed investigation and cultured cell models to of the link between a genetic understand the functional change and evolution of the human brain. of the duplication events, A > B > C > D, consequences of the human-specific Investigations of gene duplicates (i.e., with C and D independently derived from SRGAP2 duplications. The authors further paralogs) can face substantial obstacles. B (Figure 1C). These findings are fully characterize the phenotypes resulting Copies recently duplicated remain similar consistent with the less detailed con- from Srgap2 knockdown or knockout in sequence—almost as similar as clusions of the accompanying study and compare them with the effects alleles—and often confuse genome (Charrier et al., 2012). Furthermore, com- of SRGAP2C expression. Srgap2 knock- assemblies; indeed, the SRGAP2 gene parisons of the SRGAP2 sequences down leads to neurons with increased family was grossly misassembled in the suggest that the duplication events densities of immature-looking dendritic human reference sequence. Now, Dennis occurred 3.4, 2.4, and 1.0 million spines in juveniles. Knockout mice are et al. (2012) present an approach to years ago. The first event duplicated only viable even as homozygotes, retaining characterize the paralogs that should 9 of the 22 exons, truncating SRGAP2 in 10% of Srgap2 expression, and be widely applicable: use DNA from its F-BAR domain; therefore, all of the show continued growth of spine heads a haploid source, a complete hydatidiform other duplicated copies are also trun- during development, with the result that mole (the product of fertilization of an cated, with key functional consequences. spine head size in adults is close to wild- enucleated oocyte by a single sperm) in The duplicates are all expressed, and, type, but spines are more numerous and which there are no allelic variants to given their similarity, sorting out their necks are longer. SRGAP2C can dimerize confound assembly, and sequence large- specific roles is challenging. The first with SRGAP2A through its truncated F- insert clones. This allows the authors simplification comes because there is an BAR domain and decrease SRGAP2A to identify four copies: the parental additional deletion within SRGAP2D, activity. Strikingly, the simple conclusion SRGAP2A and its three duplicates removing exons 2 and 3; this, together is that SRGAP2C expression closely SRGAP2B–D, and also to infer the order with its absence from some individuals in mimics the Srgap2 knockdown and

738 Cell 149, May 11, 2012 ª2012 Elsevier Inc. knockout phenotypes in almost all of the tial phenotypic effects (Figure 1D). REFERENCES characteristics examined. In summary, SRGAP2B, the progenitor of SRGAP2C, the functional studies suggest that must have been an active antagonist at Bufill, E., Agustı´, J., and Blesa, R.(2011). Am. J. Hum. Biol. 23,729–739. SRGAP2C, by reducing SRGAP2A acti- the time of its duplication 3.4 million years vity, contributes to human-like features, ago, and SRGAP2 activity would have Charrier, C., Joshi, K., Coutinho-Budd, J., Kim, J.- E., Lambert, N., de Marchena, J., Jin, W.-L., Van- including extended brain development— reached a minimum after the SRGAP2C derhaeghen, P., Ghosh, A., Sassa, T., et al. neoteny—and cell structure in the duplication 2.4 million years ago (Fig- (2012). Cell 149, this issue, 923–935. neocortex. ure 1D). These duplications would have Chimpanzee Sequencing and Analysis Consor- In humans, the phenotypes associated occurred in Australopithecus species tium. (2005). Nature 437, 69–87. with natural loss-of-function or duplica- (Figures 1B and 1C). Did they have Dart, R.A. (1925). Nature 115, 195–199. tion variants of SRGAP2A and C are of consequences for gross brain anatomy Dennis, M.Y., Nuttle, X., Sudmant, P.H., Antonacci, great interest. Particularly relevant is that might be recognized in rare fossil F., Graves, T.A., Nefedov, M., Rosenfeld, J.A., Saj- a balanced translocation disrupting one endocasts, e.g., Dart (1925) (mouse jadian, S., Malig, M., Kotkiewicz, H., et al. (2012). copy of SRGAP2A in a 5-year-old girl models might be informative here), Cell 149, this issue, 912–922. with symptoms including intellectual and did they contribute to the develop- Fortna, A., Kim, Y., MacLaren, E., Marshall, K., disability and seizures (Saitsu et al., ment and behavior of these species Hahn, G., Meltesen, L., Brenton, M., Hink, R., 2011). Loss or gain specific to SRGAP2C documented by paleontologists? Intrigu- Burgers, S., Hernandez-Boussard, T., et al. has not yet been reported, but Dennis ingly, the use of recognizable stone (2004). PLoS Biol. 2, E207. and colleagues find large duplications tools began about 2.5 million years ago Guerrier, S., Coutinho-Budd, J., Sassa, T., Gres- affecting numerous genes, including (Jobling et al., 2004), and brain size set, A., Jordan, N.V., Chen, K., Jin, W.L., Frost, SRGAP2C—predicted to increase started to increase soon after (Figure 1A), A., and Polleux, F. (2009). Cell 138, 990–1004. SRGAP2A antagonism—in both one but it is difficult to test for a direct link. Jobling, M.A., Hurles, M.E., and Tyler-Smith, C. control and three patients with intellec- Neoteny has long been recognized as (2004). Human Evolutionary Genetics: Origins, tual disability and/or autism spectrum a human characteristic (Bufill et al., Peoples and Disease (New York, Abingdon: Garland Science). disorder. Further surveys of human vari- 2011), and now we can begin to under- ants and their detailed phenotypes, stand its genetic and developmental Saitsu, H., Osaka, H., Sugiyama, S., Kurosawa, K., Mizuguchi, T., Nishiyama, K., Nishimura, A., Tsuru- particularly SRGAP2C deletions, should basis. saki, Y., Doi, H., Miyake, N., et al. (2011). Am. J. be highly informative. Med. Genet. A. 158A, 199–205. These conclusions have several impli- ACKNOWLEDGMENTS Scally, A., Dutheil, J.Y., Hillier, L.W., Jordan, G.E., cations for our thinking about human Goodhead, I., Herrero, J., Hobolth, A., Lappalai- evolution. The duplications would have Our work is supported by The Wellcome Trust nen, T., Mailund, T., Marques-Bonet, T., et al. had immediate and perhaps substan- (098051). (2012). Nature 483, 169–175.

Cell 149, May 11, 2012 ª2012 Elsevier Inc. 739