© 2001 Macmillan Magazines LtdNATURE REVIEWS | GENETICS VOLUME 2 | DECEMBER 2001 | 917

Up to 10% of couples suffer from infertility problems.Although environmental factors might contribute tothis high number, understanding the genetic basis ofinfertility has been the focus of active research overthe past few years. Two new genetic causes of maleinfertility are now reported in Nature Genetics.Kuroda-Kawaguchi et al. show that deletions in theAZF region of the Y chromosome result from illegiti-mate homologous recombination and are a conse-quence of the highly repetitive structure of the locus.In the second study, Rovio et al. show that an unusualallele of a mitochondrial DNA polymerase (POLG) isassociated with defective sperm function.

The Y-linked AZFa, b and c are three non-contiguous regions that are well known to thosestudying male infertility deletions in regions aand c have been associated with male infertility, butthe AZF region proved difficult to dissect because ofits repetitive nature. Using a combination of carefulsequencing and analysis of subtle differencesbetween closely related but non-allelic sequences,Kuroda-Kawaguchi et al. succeeded in assembling a4.5-Mb BAC contig spanning the AZFc region. Theiranalysis revealed repeats and rearrangements on anunprecedented scale. 93% of the region consists ofamplicons — families of massive, sometimes almostidentical, repeats — that are arranged into directand mostly palindromic inverted repeats. By map-ping the AZFc deletion ends in a group of infertilemen the authors found that these deletions ariseowing to homologous recombination betweendirect repeats that lie up to 3.5 Mb apart. Gene orga-nization within the complex turned out to be as sur-prising as the genomic structure. It seems that theregion has been under peculiar selection pressurebecause, unlike the rest of the Y chromosome, in

which only half of the genes are testis specific, mostof the 27 transcription units in this region areexpressed exclusively in the testis.

The link between POLG and male infertility wasthe focus of the second study. Rovio et al. previouslyshowed that the most common allele of POLG,which contains a ten-copy CAG microsatelliterepeat, is present in human populations at a veryhigh frequency, perhaps owing to selection. Theyfound that the common allele is absent from menwith certain sperm defects and that the POLGmicrosatellite repeat that they carry occurs in vary-ing lengths. As spermatozoa are highly motile, it hasbeen proposed that mitochondrial defects couldcause male infertility if they reduce energy levelsavailable to sperm cells. Rovio et al. suggest that sub-optimal mitochondrial polymerase might cause anaccumulation of mitochondrial mutations that inturn might affect sperm function.

Although we are some way from understandingthe genetic causes of male infertility, these studiesrepresent an important step forward — Kuroda-Kawaguchi et al. have discovered an unprecedentedexample of genomic organization, and one offshootof the Rovio et al. study could be the use of POLGin genetic screening, offering early detection ofinfertility to some men.

Magdalena SkipperReferences and links

ORIGINAL RESEARCH PAPERS Kuroda-Kawaguchi, T. et al. TheAZFc region of the Y chromosome features massive palindromes anduniform recurrent deletions in infertile men. Nature Genet. 29, 279–286(2001) | Rovio, A. T. et al. Mutations at the mitochondrial DNApolymerase (POLG) locus associated with male infertility. Nature Genet.29, 261–262 (2001)WEB SITEDavid Page’s lab: http://www.hhmi.org/research/investigators/page.html

Close-up on male infertility

H U M A N G E N E T I C S

frequency of a DIA4 variant — whichleads to loss of protection against thetoxic effects of quinones — differedbetween populations from China andPapua New Guinea, which wereassigned to different clusters.Normally, these populations wouldhave been grouped together as Asians,and this important difference wouldhave been missed.

Although these results might notcome as a complete surprise to thosewho study the complexities of popula-tion history, they do provide evidenceof shortcomings in current drug trials.They also show how genetic classifica-tion might be a valuable tool for thefuture design of clinical trials that willhave benefits both in terms of cost sav-ings and human health.

Mark Patterson

References and linksORIGINAL RESEARCH PAPER Wilson, J. F. etal. Population genetic structure or variable drugresponse. Nature Genet. 29, 265–269 (2001)WEB SITEDavid Goldstein’s lab:http://www.ucl.ac.uk/biology/goldstein/Gold.htm