Identification of Critical AZF Genes
The recurrent AZF deletions previously discussed account for most of those understood at a molecular level, but there are exceptions for all three regions. Partial deletions of AZFa have been described (32); out of a sample of 11 deletions removing AZFb, 2 could not be explained by homologous recombination (30); and 1 out of 48 AZFc deletions was smaller than the common size shared by the other 47 (19). Although these findings require further investigation and interpretation in the light of the large variety of possible deletion substrates, they do suggest that there is a low level of more diverse and sometimes nonhomologous deletion. Although of less clinical significance because of their rarity, they may be highly informative in pinpointing the functionally important genes in the large regions identified by the recurrent mutations.
An alternative way of identifying the critical genes is offered by some of the variant structures. Despite searches, point mutations have not so far been found in AZFb or AZFc, and one possible explanation for this failure is that inactivation of a single copy of any of the genes present, which are all multicopy in the reference sequence, would be insufficient to lead to spermatogenic failure. In some lineages, however, such as N, the genes BPY2 and CDY1 are present in single copies, and DAZ is reduced to two copies (Fig. 3). Men belonging to lineage N might therefore experience spermatogenic failure as a result of inactivation of the last copy of one of these genes by a point mutation, so a screening program focusing on this lineage might prove fruitful.
Understanding the Significance of Partial AZF Deletions
Independent replication of the association study between gr/gr deletions and spermatogenic failure is needed, and further work should identify the deletion breakpoints and thus the individual members of gene families that are lost. In addition, more thorough investigation of the effect of the common partial deletions within AZFc on spermatogenesis is desirable, taking into account both the phylogenetic background and any associated duplications. Lineages Db2 and N would be of particular interest. Population genetic approaches might also provide information: have these lineages experienced negative selection and expanded less rapidly than expected for a neutral lineage? Methods that compare the age of a lineage with its frequency and determine whether the observed pattern is compatible with neutrality (33,34) could be used to address this question. If some partial deletions are associated with fitness costs, these might indicate the underlying selective pressures that have operated to favor multicopy genes on the Y chromosome.
Was this article helpful?