Some cases of hereditary nonpolyposis colorectal cancer (HNPCC) are due to alterations in a mutS-related mismatch repair gene. A search of a large database of expressed sequence tags derived from random complementary DNA clones revealed three additional human mismatch repair genes, all related to the bacterial mutL gene. One of these genes (hMLH1) resides on chromosome 3p21, within 1 centimorgan of markers previously linked to cancer susceptibility in HNPCC kindreds. Mutations of hMLH1 that would disrupt the gene product were identified in such kindreds, demonstrating that this gene is responsible for the disease. These results suggest that defects in any of several mismatch repair genes can cause HNPCC.

Hereditary nonpolyposis colorectal cancer (HNPCC) is one of man's commonest hereditary diseases. Several studies have implicated a defect in DNA mismatch repair in the pathogenesis of this disease. In particular, hMSH2 and hMLH1 homologues of the bacterial DNA mismatch repair genes mutS and mutL, respectively, were shown to be mutated in a subset of HNPCC cases. Here we report the nucleotide sequence, chromosome localization and mutational analysis of hPMS1 and hPMS2, two additional homologues of the prokaryotic mutL gene. Both hPMS1 and hPMS2 were found to be mutated in the germline of HNPCC patients. This doubles the number of genes implicated in HNPCC and may help explain the relatively high incidence of this disease.

Members of the double-stranded RNA-(dsRNA) specific adenosine deaminase gene family convert adenosine residues into inosines in dsRNA and are involved in A-to-I RNA editing of transcripts of glutamate receptor (GluR) subunits and serotonin receptor subtype 2C (5-HT 2C R). We have isolated hADAR3, the third member of this class of human enzyme and investigated its editing site selectivity using in vitro RNA editing assay systems. As originally reported for rat ADAR3 or RED2, purified ADAR3 proteins could not edit GluR-B RNA at the "Q/R" site, the "R/G" site, and the intronic "hot spot" site. In addition, ADAR3 did not edit any of five sites discovered recently within the intracellular loop II region of 5-HT 2C R RNAs, confirming its total lack of editing activity for currently known substrate RNAs. Filter-binding analyses revealed that ADAR3 is capable of binding not only to dsRNA but also to singlestranded RNA (ssRNA). Deletion mutagenesis identified a region rich in arginine residues located in the N-terminus that is responsible for binding of ADAR3 to ssRNA. The presence of this ssRNA-binding domain as well as its expression in restricted brain regions and postmitotic neurons make ADAR3 distinct from the other two ADAR gene family members, editing competent ADAR1 and ADAR2. ADAR3 inhibited in vitro the activities of RNA editing enzymes of the ADAR gene family, raising the possibility of a regulatory role in RNA editing.

Nucleotide excision repair, which is defective in xeroderma pigmentosum (XP), involves incision of a DNA strand on each side of a lesion. We isolated a human gene homologous to yeast Rad1 and found that it corrects the repair defects of XP group F as well as rodent groups 4 and 11. Causative mutations and strongly reduced levels of encoded protein were identified in XP-F patients. The XPF protein was purified from mammalian cells in a tight complex with ERCC1. This complex is a structure-specific endonuclease responsible for the 5' incision during repair. These results demonstrate that the XPF, ERCC4, and ERCC11 genes are equivalent, complete the isolation of the XP genes that form the core nucleotide excision repair system, and solve the catalytic function of the XPF-containing complex.

Abstract. The functional organization of the nucleus was studied using a fluorescence microscopy approach which allowed integration of positional information for RNA, DNA, and proteins. In cells from sea urchin to human, nuclear poly(A) RNA was found concentrated primarily within several discrete "transcript domains" which often surrounded nucleoli . Concentrations of poly(A) RNA were coincident with snRNP antigen clusters, providing evidence for the localization of premRNA splicing at these sites . The spatial relationship of transcript domains with respect to various classes of DNA was established, in that the poly(A) RNA-rich regions coincided with discrete regions of low DNA density and were non-randomly distributed with respect T HE cell nucleus performs numerous complex tasks which include packaging the enormous length of DNA, high fidelity replication ofthis DNA, transcription, processing, and transport of different classes of RNA, protein import and export, and precise redistribution of nuclear components during mitosis . Additionally, many of these are done in a cell-type specific manner. It has often been proposed that the extremely dense structure of the nucleus may be spatially compartmentalized as a means to efficiently carry out its multiple functions (see Comings, 1968Comings, , 1980 . However, evidence for such physical partitioning has been modest . For example, while great advances have been made in understanding the biochemical steps involved in pre-mRNA transcription and processing, the subnuclear location ofthese events and the extent to which they are compartmentalized is unknown . In this paper we investigate the potential compartmentalization of these and other nuclear events by analyzing the distribution of poly(A) RNA in relation to several nuclear constituents of known functional significance .Unlike the cytoplasm, there are no lipid membranes in the nucleus, but the confinement to the nucleolus ofrRNA genes originating on several different chromosomes shows that elegant spatial and functional regionalization is possible. Strong evidence supporting the early suggestions that interphase centromeres and telomeres are specifically positioned (Rabl, 1885) and individual chromosomes occupy distinct territories (Boveri, 1909) has been presented recently by several