Corrections

COMMENTARY. For the article ‘‘The drifting ,’’ by BIOCHEMISTRY. For the article ‘‘Crystal structure and functional Jianzhi Zhang, which appeared in issue 51, December 18, 2007, analysis of tetracenomycin ARO/CYC: Implications for cycliza- of Proc Natl Acad Sci USA (104:20147–20148; first published tion specificity of aromatic polyketides,’’ by Brian Douglas Ames, December 10, 2007; 10.1073͞pnas.0710524105), the authors note Tyler Paz Korman, Wenjun Zhang, Peter Smith, Thanh Vu, Yi that, due to a printer’s error, the DOI appeared incorrectly. The Tang, and Shiou-Chuan Tsai, which appeared in issue 14, April DOI 10.1073/pnas.0710524105 should have appeared as 10.1073/ 8, 2008, of Proc Natl Acad Sci USA (105:5349–5354; first pnas.0710524104. The online version has been corrected. published April 3, 2008; 10.1073͞pnas.0709223105), the authors note that on page 5353, left column, in Biological Significance and ͞ ͞ ͞ ͞ www.pnas.org cgi doi 10.1073 pnas.0804195105 Proposed Mechanism, line 5, ‘‘intercalculating’’ should have appeared as ‘‘intercalating.’’ Additionally, on page 5354, left column, in Expression and Protein Purification, in the last BIOCHEMISTRY. For the article ‘‘Heterogeneity in EGF-binding line, ‘‘10 mM Hepes (pH 7.0) and NaOH’’ should instead read: affinities arises from negative cooperativity in an aggregating ‘‘10 mM Hepes (pH 7.0)/NaOH.’’ Finally, in the main article text, system,’’ by Jennifer L. Macdonald and Linda J. Pike, which the numbering and links to some supporting information (SI) appeared in issue 1, January 8, 2008, of Proc Natl Acad Sci USA ͞ figures and tables were incorrect and have been corrected. These (105:112–117; first published December 28, 2007; 10.1073 errors were printer’s errors and do not affect the conclusions of pnas.0707080105), the authors note that in Fig. 3, in the param- the article. eter list inset at lower right, the labels K21 and L20 are reversed. The correct values for all four parameters are K11 ϭ 4.6 Ϯ 0.6 ϫ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0804020105 9 11 9 10 , L20 ϭ 5.3 Ϯ 2.2 ϫ 10 , K21 ϭ 5.3 Ϯ 0.4 ϫ 10 , and K22 ϭ 3.4 Ϯ 1.1 ϫ 108. This error does not affect the conclusions of the article. The corrected figure and its legend appear below. BIOCHEMISTRY. For the article ‘‘EB1 promotes Aurora-B kinase activity through blocking its inactivation by protein phosphatase 2A,’’ by Lei Sun, Jinmin Gao, Xin Dong, Min Liu, Dengwen Li, Xingjuan Shi, Jin-Tang Dong, Xianyu Lu, Chunyong Liu, and Jun Zhou, which appeared in issue 20, May 20, 2008, of Proc Natl Acad Sci USA (105:7153–7158; first published May 13, 2008; 10.1073͞pnas.0710018105), the authors note that Jinmin Gao (J.G.), and not Jun Zhou (J.Z.), should have been included among those credited with contributing equally to this work. The corrected author contributions footnote appears below. In ad- dition, in the Acknowledgments, a grant number was inadver- tently omitted. ‘‘This work was supported by grants from the

National Key Scientific Program of China (2007CB914503 and CORRECTION 2007CB914802),’’ should instead read: ‘‘This work was sup- ported by grants from the National Key Scientific Program of China (2006CB910103, 2007CB914503, and 2007CB914802).’’

Author contributions: L.S., J.G., and X.D. contributed equally to this work; J.Z. designed research; L.S., J.G., X.D., M.L., D.L., X.S., and J.Z. performed research; J.-T.D., X.L., and C.L. contributed new reagents/analytic tools; L.S., J.G., and J.Z. analyzed data; and J.Z. wrote the Fig. 3. Binding of EGF to cells expressing increasing levels of wild-type EGF paper. receptors. CHO-K1 tet-on EGFR cells were induced to express EGF receptors by www.pnas.org͞cgi͞doi͞10.1073͞pnas.0804731105 using increasing doses of doxycycline. 125I-EGF-binding isotherms were gen- erated from each set of cells, and all six isotherms were globally fit to Eq. 1, with only the value of R0 varying among curves. Data points represent the mean Ϯ SD of triplicate determinations. Solid lines represent the fitted curve through the data points of the same color.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0804457105

www.pnas.org PNAS ͉ July 1, 2008 ͉ vol. 105 ͉ no. 26 ͉ 9129–9130 Downloaded by guest on September 23, 2021 COMMENTARY

The drifting human genome

Jianzhi Zhang* Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI 48109

round the time of the comple- more, they found a normal distribution and Old World monkeys because of the tion of the draft human ge- of gene copy number among human in- inactivation of TRPC2, a crucial channel nome sequence, biologists dividuals, with a standard deviation of protein gene, in the common ancestor of heatedly debated the number 54. Between any two individuals, the hominoids and Old World monkeys (9). Aof contained in the human ge- average gene copy number difference is Nozawa et al. (2) further extended nome. In 2003, GeneSweep, an informal as high as 61.5. When Africans, Asians, their analysis from within populations to gene-count betting pool that began at and Europeans were analyzed sepa- between populations and between spe- the 2000 Cold Spring Harbor Labora- rately, Africans showed greater variation cies. This type of analysis is in principle tory Genome Meeting, announced Lee than Asians and Europeans, consistent similar to the Hudson–Kreitman– Rowen of the Institute of Systems Biol- with the out-of-Africa model of modern Aguade´ (HKA) test of neutrality of sin- ogy in Seattle to be the winner. His bet human origins. gle nucleotide changes (10). The HKA (25,946) was the lowest and was closest Their subsequent analysis focused on test compares the ratios of the intraspe- to what many computer programs pre- three families of chemosensory genes cific polymorphism level and the inter- dicted from the then draft human ge- that encode odorant receptors (ORs), specific divergence level between two nome sequence (1). But, in this issue of type 1 vomeronasal (pheromone) recep- loci, which are expected to be identical PNAS, Nozawa, Kawahara, and Nei (2) tors (V1Rs), and bitter taste receptors under neutrality. When one of the loci tell us that asking about the exact num- (T2Rs), respectively, because chemosen- is a functional gene and the other is a ber of human genes is meaningless, be- sory genes have particularly high levels cause different human individuals have pseudogene, rejection of the null hy- different numbers of genes, and it is not pothesis by the HKA test suggests non- uncommon for one person to have 100 It is not uncommon for neutral evolution of the functional gene, more gene copies than another person. caused by positive or negative selection. They argue that this large variation re- one person to have 100 Nozawa et al. measured the intraspecific flects genomic drift, random changes of polymorphism of OR gene copy number gene copy number in evolution. more gene copies than by its standard deviation among human Genetic variation among humans individuals and measured the interspe- takes many forms. In the past two de- another person. cific divergence of OR gene copy num- cades, most studies have focused on ber by the absolute difference in OR single-nucleotide polymorphisms. For gene number between the reference ge- example, the International HapMap of CNV (2, 4, 8). It is important to note nome sequences of humans and chim- Consortium recently published HapMap panzees. Similar measures were used for Ͼ that chemosensation is apparently im- II, which characterizes 3 million portant and highly refined in mammals, OR pseudogenes. However, they did not single-nucleotide polymorphisms geno- as Ͼ5% of all genes in a mammalian formally test the neutral hypothesis, pre- typed in 270 individuals from four hu- genome are devoted to the detection of sumably because it is not possible to man populations of African, Asian, and odorants, pheromones, and tastants. The compare polymorphism and divergence European origins (3). In the last few most interesting analysis Nozawa et al. statistically when one is measured by years, multiple groups reported a high (2) conducted is a comparison of CNV standard deviation and the other is mea- prevalence of copy number variation between functional and nonfunctional sured by difference. Inspired by their (CNV) of DNA segments ranging in size attempt, I designed a neutrality test in from 103 to 106 nucleotides in humans chemosensory genes. They found that the proportion of genes that have CNV which polymorphism is measured by the (4, 5), chimpanzees (6), and mice (7). number of OR loci that show CNV and CNV is generated by segmental duplica- is not significantly different between functional and nonfunctional OR genes. divergence is measured by the difference tion and deletion mutations that are ap- in OR gene number between the refer- parently quite frequent. Thanks to the Furthermore, the distribution of OR ence genome sequences of humans and rapid advance of genomic technologies, gene copy number among human indi- chimpanzees. Although some informa- genome-wide human CNV data recently viduals is normal for both functional tion about the intraspecific variation is became available, but studies of CNV genes and pseudogenes, with almost have been largely descriptive. The work identical standard deviations. Because lost in this treatment as the quantitative of Nozawa et al. (2) puts CNV into the pseudogenes are not subject to natural level of CNV at each locus is not con- framework of population and evolution- selection, the simplest explanation of sidered, I can now compare polymor- ary and is thus an exciting ad- this observation is that the CNV of phism and divergence directly. As shown vance in our understanding of CNV in functional OR genes is also neutral. in Table 1, the null hypothesis that OR the human genome. Similar results were found for T2Rs, but gene number variation is neutral cannot 2 Nozawa et al. (2) analyzed a large the correspondence between functional be rejected (␹ ϭ 2.06, P ϭ 0.15), al- dataset of human CNV that was based genes and pseudogenes is less impressive though the functional gene copy number on a survey of the 270 HapMap individ- because the T2R family is much smaller uals (4). The dataset included 1,447 than the OR family. For V1Rs, the au- CNV regions across the human genome thors rightly treated all intact and Author contributions: J.Z. wrote the paper. (4), containing 3,144 annotated genes disrupted genes as nonfunctional gene The author declares no conflict of interest. (2). Nozawa et al. found that on aver- relics because the vomeronasal signal See companion article on page 20421. age, individuals differ from a reference transduction pathway was destroyed in *E-mail: [email protected]. individual at 277 CNV loci. Further- all hominoids (i.e., humans and apes) © 2007 by The National Academy of Sciences of the USA www.pnas.org͞cgi͞doi͞10.1073͞pnas.0710524104 PNAS ͉ December 18, 2007 ͉ vol. 104 ͉ no. 51 ͉ 20147–20148 Table 1. Testing the hypothesis of neutral variation of the number of odorant tion of the interspecific divergence in receptor genes sensory gene number reflects adapta- No. of genes with CNV Gene no. difference between tion, whereas the remaining fraction is Genes in humans humans and chimpanzees neutral. Although it is difficult to assess the size of each fraction at this time, the Functional 116 16 neutral fraction is likely much higher in Pseudogenes 143 11 humans than in some other mammals Data are from ref. 2. ␹2 ϭ 2.06, P ϭ 0.15. that rely heavily on chemosensation. Beyond the focus on CNV at chemo- sensory loci, the high level of CNV for difference between species appears to vious study (12) showed that purifying many genes across the human genome be larger than the neutral expectation. selection on human T2R bitter receptor demonstrates the neutrality or near neu- The findings of Nozawa et al. (2) and genes appears to have been completely trality of CNV at many loci, and the above neutrality test suggest that relaxed. Thus, even if the various copies genomic drift apparently occurs. How- both intraspecific variation and interspe- of a given chemosensory gene are ever, both disease-causing CNV (5) and cific divergence of gene copy number in slightly different in function, the fitness beneficial CNV (21) have been re- human OR and other chemosensory re- consequence in humans is expectedly to ported, so this type of genetic variation ceptor families are simply due to be extremely small. is not entirely neutral in humans. Under chance. For two reasons, this result is Would the result be different had the background of neutral CNV, identi- not unexpected. First, for intraspecific Nozawa et al. (2) studied other mam- fication of such nonneutral human CNV CNV, the proteins encoded by various mals such as mice? The answer is proba- will enhance our understanding of the copies of the same gene are likely to be bly yes. Many studies showed that the genetic basis of disease as well as adap- extremely similar in sequence and func- numbers of T2R, OR, V1R, and V2R tation. Although genome-wide identifi- tion because of very recent divergences (type 2 vomeronasal receptor) genes cation of CNV is rather crude at of the different copies. Because dosage vary dramatically across mammals and present, because the breakpoints of each is unlikely to be important to chemosen- that this variation is among the largest copy and the copy number difference sory receptors, the actual copy number of all mammalian gene families exam- between two individuals at each locus of each sensory gene should not affect ined (13–19). For example, of those are not precisely determined, the impli- the fitness of an individual as long as at mammals that have an intact vomerona- cations of CNV polymorphism and least one copy is present in the individ- sal system, platypus has 270 functional ual. Second, olfaction is believed to play V1R genes, whereas the dog has only 8 divergence can be reconfirmed and a less important role in humans than in (15). Furthermore, the number of func- refined when better data become other primates, as reflected by a de- tional V1R genes correlates well with available. creased number of functional OR genes the morphological complexity of the Although it has been some time since in the human genome compared with vomeronasal organ (14, 15). It is thus the completion of the draft human ge- those in many other primates (11). likely that the interspecific differences nome sequence, we continue to discover As mentioned, V1Rs are not used in of chemosensory receptor gene numbers and explore new types of genomic varia- humans because of the loss of the vome- are not entirely neutral. However, it has tion. Fifteen years ago, when I first ronasal signal transduction pathway. Hu- also been suggested that in mice and heard about the Human Genome mans also rely less on bitter taste than rats, V1R gene duplication is mediated Project as a senior undergraduate stu- other apes do, because (i) bitter taste is by L1 repetitive elements as these ele- dent, I asked my professor ‘‘Whose ge- primarily used to detect toxic food, but ments densely populate the genomic re- nome is going to be sequenced?’’ He humans eat less plant tissues and more gions harboring V1R genes (20). Thus, replied that it does not matter because animal tissues and, thus, encounter variation in sensory gene number could all human genomes are essentially the fewer toxins than other apes do, and (ii) arise from differential activities of repet- same. Now I can tell him, ‘‘No, I may humans cook food, which is an effective itive elements in different species. have a hundred more genes than you method of detoxification. In fact, a pre- Taken together, it is likely that a frac- have!’’

1. Pennisi E (2003) Science 300:1484. 7. Li J, Jiang T, Mao JH, Balmain A, Peterson L, Harris C, 15. Grus WE, Shi P, Zhang J (2007) Mol Biol Evol 24:2153– 2. Nozawa M, Kawahara Y, Nei M (2007) Proc Natl Acad Rao PH, Havlak P, Gibbs R, Cai WW (2004) Nat Genet 2157. Sci USA 104:20421–20426. 36:952–954. 16. Shi P, Zhang J (2007) Genome Res 17:166– 3. Frazer KA, Ballinger DG, Cox DR, Hinds DA, Stuve 8. Nguyen DQ, Webber C, Ponting CP (2006) PLoS Genet 174. LL, Gibbs RA, Belmont JW, Boudreau A, Hardenbol 2:e20. 17. Shi P, Zhang J (2006) Mol Biol Evol 23:292– P, Leal SM, et al. (2007) 449:851–861. 9. Zhang J, Webb DM (2003) Proc Natl Acad Sci USA 300. 4. Redon R, Ishikawa S, Fitch KR, Feuk L, Perry GH, An- 100:8337–8341. 18. Niimura Y, Nei M (2006) J Hum Genet 51:505– 10. Hudson RR, Kreitman M, Aguade´ , M (1987) Genetics drews TD, Fiegler H, Shapero MH, Carson AR, Chen W, 517. 116:153–159. et al. (2006) Nature 444:444–454. 19. Young JM, Trask BJ (2007) Trends Genet 23:212– 11. Gilad Y, Przeworski M, Lancet D (2004) PLoS Biol 2:E5. 5. Cooper GM, Nickerson DA, Eichler EE (2007) Nat Genet 215. 12. Wang X, Thomas SD, Zhang J (2004) Hum Mol Genet 39:S22–S29. 13:2671–2678. 20. Lane RP, Cutforth T, Axel R, Hood L, Trask BJ (2002) Proc 6. Perry GH, Tchinda J, McGrath SD, Zhang J, Picker SR, 13. Young JM, Kambere M, Trask BJ, Lane RP (2005) Ge- Natl Acad Sci USA 99:291–296. Caceres AM, Iafrate AJ, Tyler-Smith C, Scherer SW, nome Res 15:231–240. 21. Perry GH, Dominy NJ, Claw KG, Lee AS, Fiegler H, Eichler EE, et al. (2006) Proc Natl Acad Sci USA 14. Grus WE, Shi P, Zhang YP, Zhang J (2005) Proc Natl Acad Redon R, Werner J, Villanea FA, Mountain JL, Misra 103:8006–8011. Sci USA 102:5767–5772. R, et al. (2007) Nat Genet 39:1256–1260.

20148 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0710524104 Zhang