Novel class of polytopic proteins with domains associated with putative protease activity.

A significant proportion of early onset Alzheimer's disease (AD) is caused by mutations in human genes for amyloid precursor protein (APP), presenilins 1 and 2 (PSEN1,2). AD associated mutations in PSEN1,2 genes alter the gamma-secretase cleavage activity of APP resulting in increased production of amyloidogenic Abeta42. PSEN dependent intramembrane proteolysis was described as an important step required for cleavage of Notch receptors, Notch-dependent signal transduction, and processing of other proteins. It is still unclear whether presenilins are unusual intramembrane proteases or they are necessary cofactors of gamma-secretase cleavage of APP and Notch. Identification of other proteins similar to presenilins may resolve this dilemma. We describe here the identification of novel families of genes encoding polytopic transmembrane proteins of Eukaryotes (IMPASes) and Arachaea(membrases). These proteins have a predicted structure similar to presenilins. The amino acid similarity is significant in domains carrying invariant amino acid residues, which are critical in specific presenilin-regulated endoproteolysis. Many members of the IMPAS family have protease associated domains (PA) typical of proteases. We identified and cloned five human IMPAS genes. Expression analysis of the hIMP1 gene (located on chromosome 20) was performed in human cell tissues and transfected cell cultures. The data demonstrate that a conservative class of putative protease-related polytopic proteins related to presenilins exists in multicellular eukaryotes and microorganisms.

Mechanisms underlying mismatch repair deficiencies in normal cells.

Hereditary nonpolyposis colon cancer (HNPCC) is an autosomal dominantly inherited cancer predisposition which is linked to heterozygous mutations in mismatch repair genes. HNPCC tumour cells, in which the remaining wild-type copy of the mismatch repair gene is inactivated, display instability of microsatellite markers reflecting a defect in mismatch repair. Recently, patients carrying either one of two distinct germline mutations in the MLH1 and PMS2 genes were reported to accumulate somatic mutations of microsatellites in untransformed cells. One of the mechanisms that might account for this phenomenon was a dominant negative effect of the mutant allele. To evaluate this possibility, we examined a different family carrying one of the mutations (deletion of codon 618K in the MLH1 gene) which has been suspected to induce genetic instability in untransformed cells. No mutations in dinucleotide repeat markers were observed in a large number of lymphoblast clones derived from a carrier. Evidence for the deletion of the wild-type allele in two different tumours suggested that the inactivation of both gene copies was required for tumour initiation. These results indicate that the MLH1 618K deletion mutation alone does not necessarily cause marked mismatch repair deficiency in the presence of a wild-type allele.