Pseudogenes and their formation - Human Gene Evolution - page 276

conversion has occurred between highly homologous and closely linked gene
sequences (Cooper and Krawczak, 1993). Examples of gene conversion that have
occurred during primate evolution are discussed in Chapter 9, section 9.5. Gene
conversion has also served as a mutational mechanism causing human inherited
disease; probable examples involve the genes and pseudogenes for steroid 21-
hydroxylase ( CYP21 ; 6p21.3; Tusié-Luna and White, 1995), polycystic kidney
disease ( PKD1 ; 16p13), neutrophil cytosolic factor p47- phox ( NCF1 ; 7q11.2;
Görlach et al ., 1997), immunoglobulin l-like polypeptide 1 ( IGLL1 ; 22q11;
Minegishi et al ., 1998), glucocerebrosidase ( GBA ; 1q21; Eyal et al ., 1990), von
Willebrand factor ( VWF ; 12p13; Eikenboom et al ., 1994), and phosphomanno-
mutase ( PMM2 ; 16p13; Schollen et al ., 1998). These gene/pseudogene pairs are all
closely linked with the exception of the VWF gene (12p13) and its pseudogene
(22q11-q13), and the PMM2 gene (16p13) and its pseudogene (18p). Together,
these two exceptions would seem to establish a precedent for the occurrence of
gene conversion between unlinked loci in the human genome.
Could pseudogenes (both the highly dispersed processed variety and/or the
closely linked duplication-derived pseudogenes) also have templated advanta-
geous changes in their single copy functional source genes over evolutionary
time? Certainly the converse is true since sequence changes in pseudogenes can be
templated by their functional homologues (DeBry, 1998). If pseudogene-tem-
plated changes in functional genes were not deleterious, they could eventually
have become fixed in populations or even species. Thus, in principle, pseudo-
genes, whether processed or non-processed, could act as a reservoir of sequence
variation which could at some stage be transferred back to the functional gene. In
this way, different mutational combinations might be put together within the
pseudogene, all the time being shielded from selective pressure, and then func-
tionally tested after simultaneous transfer to the expressed gene copy. For any one
gene, the contribution of pseudogene-mediated gene conversion to the process of
evolutionary change might be expected to depend on:
(i) The number of homologous pseudogenes (processed or unprocessed) in the genome that
are homologous to the gene in question .
As we have seen in Section 6.1.2, this can vary by at least two orders of mag-
nitude.
(ii) Whether the pseudogenes are linked or unlinked to the functional gene .
Gene conversion does occur between unlinked homologous sequences
(Fitzgerald et al ., 1996; Murti et al ., 1994). Gene conversion between multi-
copy, dispersed retrotransposons such as Alu sequences (Kass et al ., 1995) and
LINE elements (Burton et al ., 1991) has also been reported. However, inter-
chromosomal gene conversion events are predicted to be much less frequent
than intra-chromosomal events (Liao et al ., 1997).
(iii) The length of sequence involved in the gene conversion event . Were this to be too
large, inactivating mutations present in the pseudogene would be more likely
to be transferred to the functional gene thereby inactivating it rather than
altering it. In practice, however, gene conversion events are often quite local-
ized (Kim et al ., 1993; Pamilo and Bianchi, 1993; Zhou and Lee, 1996), one
example being the HLA-DQB1 gene (6p21.3) in which gene conversion
events have been confined to exon 2 without extending into the adjacent