Integration site preferences of endogenous retroviruses.

Retroviruses have the ability to integrate into the genome of their host, in many cases with little apparent sequence or site specificity. However, relatively few studies have addressed more general features of chromosomal integration. In this study we directly visualized the chromosomal organization of three representative endogenous retroviruses by in situ hybridization. Because there are 50-1000 copies of each of these retroviruses in the genome, it was possible to evaluate repeated integration events. Each retroviral sequence exhibited a unique and markedly different integration pattern. In order to characterize more precisely the chromosomal domains targeted by each retrovirus, later replicating domains were differentially labeled. Additionally, prototypic SINES and LINES (short and long interspersed reiterated sequences), which are inhomogeneously distributed on chromosome arms, were simultaneously detected. Retroviral copies of greater than or equal to 2 kb were found (i) exclusively in a discrete set of later replicating domains, most of which have the staining characteristics of constitutive heterochromatin, (ii) widely represented in disparate types of chromosome domains, or (iii) almost completely confined to CpG Alu-rich regions that are known to be early replicating. Retroviral elements in Alu-rich domains would be expected to be actively transcribed in all cells. Surprisingly, hybridization to blots of brain RNA showed an approximately 25 fold lower level of transcripts from these Alu associated elements than from retroviral sequences restricted to later replicating, heterochromatic domains. Retroviral insertions may subvert more typical transcriptional characteristics of a domain. The present results indicate that there are highly specific integration patterns for each endogenous retrovirus that do not readily relate to their sequence or particle classification. Each host genome may utilize these elements for contrary, and possibly beneficial functions.