Britten R J
Division of Biology, California Institute of Technology, Corona del Mar 92625.
Proc Natl Acad Sci U S A. 1994 Jun 21;91(13):6148-50. doi: 10.1073/pnas.91.13.6148.
The primate Alu interspersed repeats can be subdivided into classes on the basis of shared nucleotides at a set of diagnostic positions. Each of the classes of Alu sequences is apparently the result of past retrotransposition of transcripts of highly conserved class-specific source genes that differed from each other at the diagnostic positions. The nucleotides at the majority of positions are identical among the source genes and therefore were identical among all of the Alu sequences at the time of their insertion. These CONSBI (conserved before insertion) positions are useful because the changes that have occurred after insertion are recognizable and the divergence resulting from nucleotide substitutions, insertions, and deletions is informative. The divergence of Alu sequences at the CONSBI positions is a measure of the time since a class was inserted. The greatest majority of Alu sequences are in one class (identified as class II), and it is particularly suitable for such examination, since nearly full-length sequences are now known for nearly a thousand members of this class. The average divergence of class II members indicates that the class has an average age of about 40 million years. The distribution in divergence of class II accurately fits a sum of two Poisson distributions. The implication is that class II Alu sequences were derived from two massive past events of insertion of many Alu sequences. In this model the younger subset of class II sequences (corresponding to about 300,000 copies in the genome) has an average divergence of 5% at CONSBI positions. The older set of class II sequences (corresponding to about 150,000 genomic copies) has a 9% average divergence. Based on the drift rate of primate DNA sequences, the events of insertion probably occurred 30-50 million years ago. The goodness of the fit to the Poisson distribution indicates that no significant number of members of class II have been inserted since 30 million years ago.
灵长类动物的Alu散布重复序列可以根据一组诊断位置上共享的核苷酸细分为不同类别。每个Alu序列类别显然是过去高度保守的类别特异性源基因转录本逆转录转座的结果，这些源基因在诊断位置上彼此不同。大多数位置上的核苷酸在源基因之间是相同的，因此在插入时所有Alu序列之间也是相同的。这些插入前保守（CONSBI）位置很有用，因为插入后发生的变化是可识别的，并且核苷酸替换、插入和缺失导致的差异是有信息价值的。Alu序列在CONSBI位置的差异是衡量一个类别插入后时间的指标。绝大多数Alu序列属于一个类别（被鉴定为II类），它特别适合进行此类检查，因为现在已知该类别的近一千个成员的几乎全长序列。II类成员的平均差异表明该类别平均年龄约为4000万年。II类差异的分布精确拟合两个泊松分布的总和。这意味着II类Alu序列源自过去两次大量插入许多Alu序列的事件。在这个模型中，II类序列的较年轻子集（对应于基因组中约300,000个拷贝）在CONSBI位置的平均差异为5%。II类序列的较老集合（对应于约150,000个基因组拷贝）平均差异为9%。根据灵长类动物DNA序列的漂移率，插入事件可能发生在3000万至5000万年前。泊松分布的拟合优度表明，自3000万年前以来，没有大量II类成员被插入。
