An evolutionary study of the lamin B2 gene.

Lamins are endonuclear proteins which form the lamina, a karyoskeletal structure lining the inner side of the nuclear membrane in eukaryoles. On the basis of immunological and biochemical criteria, lamin proteins in vertebrates have been grouped into two major subfamilies: A-type lamins and B-type lamins which are differently expressed. Moreover lamins display structural homology to the cytoplasmic intermediate filaments. Our study concerns the evolution of lamin B2 whose complete or partial sequence is known in human, mouse, chicken and xenopus. We have performed a comparative analysis of lamin Bg genes in order to clarify the causes of the apparent variable rate of evolution in different portions of the coding sequences and to relate this variability to the possible functional significance of certain domains of the protein. We have at first compared the genomic sequence of mouse lamin B2, which is completely known (1), with a portion at the 3' end of the human sequence, which has been recently determined (2). This portion includes the major part of the coding region, the total 3' non-translated part, two complete introns and part of a third intron. In order to study the distribution and the amount of variation in the human and mouse coding sequence of the lamin B2 gene we have made a computer program 1o evaluate zones with statistically significant deviations from the average dissimilarity between the two sequences previously aligned. The program calculates the percentage of mutated bases within a fixed window {50 bp. long) and then, moving one base downstream, it scans all the sequence length, obtaining a distribution of dissimilarity values. These values have been compared to those obtained on average in 2000 random permutations of substitutions, allowing the same number of mutation as observed, but locating them at random along ihe sequences. A comparison of the observed distribution of dissimilarity with the behaviour of the random permutations allows us to recognize zones of higher ( or lower) variability than expected by random mutation aione. This analysis has shown a good overall conservation of the lamin B2 gene in the two species. However we could identify three distinct regions in which evolution occured in different ways, The largest region (76% of the sequence), which we cad LV, is well conserved, with rates of synonymous and nonsynonymous substitutions per site, per 109 years, of 3.4 and 0.4 respectively. These values are comparable to those characteristic of well studied proteins. There is a short region (13% of the bases), which we call IP, consisting of two separate but neighbouring segments which has undergone a much higher modification, both in bases and in type and functionality of aminoacids. The rates of synonymous and nonsynonymous substitutions in this region are not significantly different from each other and their average is 6.3 substitutions per site, per 109 years. This value, being much higher than the rate of synonymous substitution in the LV region and that in pseudogenes (estimated to be about 5), cannot be easily explained by neutral fixation of point mutations alone and suggests that deeper DNA reorganization events, such as insertions or deletions, has occurred. We also identified a third region (10% of the sequence), called MV and made of the flanking regions of IP, in which the overall evolutionary rate is intermediate between that of LV and IP regions. However, while the rate of synonymous substitutions in this region is equal to that characteristic of the IP region (6.3), the rate of nonsynonymous substitutions is equal to 0,8, a value perfectly comparable to that of the LV region and of many other proteins. These contrasting results can be explained if we suppose that the reorganization events that have originated the IP region had initially also included MV. The modifications thus produced In MV could have damaged the functionality of the protein in this region and could have generated positive selectional pressures tending to restore the sequence composition of the region to its primitive functional state. If this hypothesis is true, we have to suppose that the initial alterations, on the other hand, had to have positive functional effects in the IP region, in order to balance the negative ones produced in the MV region. Otherwise the same selective forces which tended to restore the original aminoacidic situation in MV would have sooner prevented the fixation of the initial alteration. An indication that the high variability observed in the IP region could be functionally significant derives from preliminary comparisons of the human and mouse sequences with those from chicken and xenopus, which give almost Ihe same patterns of variability, with regions corresponding to IP where most disslmiiarily is concentrated. These results seem to suggest that the event which produced the diversity in IP occurred several limes in vertebrate phylogenesis and that this focalized diversity could have a functional role in the lamin B2 gene.