Deletion mutants of minute virus of mice arising during a single high-multiplicity passage and after serial undiluted passage have been isolated, and the incomplete viral genomes contained therein have been analyzed. The DNA isolated from incomplete virions derived from a single high-multiplicity passage was heterogeneous, ranging in size from 15 to 70% of the intact viral genome, with an average molecular length of approximately 2,000 nucleotides. Two distinct types of molecules, designated as type I D-DNA and type II D-DNA, could be distinguished on the basis of their degree of secondary structure, and these were present in roughly equal amounts. Type I D-DNAs were predominantly singlestranded, recombinant molecules in which the self-complementary sequences

A partially purified preparation of DNA polymerase alpha, obtained from the cytosol of Ehrlich ascites tumour cells, has been found to catalyze the conversion of MVM parvovirus, SS DNA (5 kilobases) to RF in vitro. The reaction initiates at a natural 55 base pair hairpin which exists at the 3' terminus of MVM SS DNA. The SS leads to RF conversion is sensitive to aphidicolin, resistant to ddTTP and is promoted by purine ribonucleoside 5' triphosphates, a phenomenon which could not be explained simply by stabilization effects on the in vitro deoxynucleotide precursor pool. In the absence of rNTPs, nascent complementary strands frequently terminate prematurely at a preferred location, between 1300 and 1700 nucleotides from the initiating 3' hairpin terminus. This in vitro system, involving self-primed parvovirus DNA synthesis, provides a convenient assay for those components of the mammalian replicative DNA polymerase complex which are required for the elongation of nascent DNA chains.

, J. Virol. 32:276-292, 1979) revealed a remarkably high rate of spontaneous deletion in viral DNA during lytic infection of cultured murine cells with minute virus of mice (MVM), an autonomous parvovirus. In the present study, we have isolated plasmid and phage recombinants containing MVM DNA inserts bearing deletions and we have determined the DNA sequence spanning three deletion junctions. The deletions, which average 3 kilobases in length, occur between pairs of perfectly homologous 4-to 10-base-pair direct repeats, such that one copy of the repeated sequence is lost, whereas the other remains behind at the deletion junction. When compared, the three sets of direct repeats exhibit no apparent sequence homology and have an A+T content of between 50 and 80%. These results indicate that 4-to 10-base-pair homologies mediate spontaneous deletion formation in the MVM genome and highlight parvoviruses as novel model systems for studies of this ubiquitous pathway of genetic variation.Spontaneous deletions account for many independent mutations within the lacd gene (1, 6), the tryptophan operon (25), and the transposable element IS1 of Escherichia coli (18) as well as the rII region of bacteriophage T4 (14) and the early region of bacteriophage T7 (20). In addition, the precise excision of certain procaryotic transposable elements such as TnJO may be viewed as analogous to spontaneous deletion formation (1, 9). In eucaryotes, spontaneous deletions, rather than base substitutions, are largely responsible for divergence between members of the 3-globin gene family in humans (5) and also have been found within integrated viral sequences present in the polyoma-transformed cell line 82-Rat (15). High-frequency mutation at the adenine phosphoribosyltransferase locus in Chinese hamster ovary cells is also due to deletion of the gene (17). In most of these cases, the deletions are associated with short, directly repeated DNA sequences that flank the deleted DNA segments, and a model that invokes slipped mispairing during DNA replication has been proposed to explain how such direct repeats might participate in the deletion process (1,5,19). We have extended these observations in the present study of high-frequency deletion mutations that arise spontaneously in the DNA genome of the parvovirus minute virus of mice (MVM) (8).MVM is a member of a large group of mammalian parvoviruses that have proven to be useful model systems for the study of mammalian DNA replication (4, 23). The MVM genome is a linear single-stranded DNA molecule consisting of 5,081 nucleotide residues with terminal self-complementary sequences that exist in the form of hairpin duplexes (3). Viral DNA replication occurs by a self-priming mechanism that leads to the accumulation of linear monomeric and concatemeric duplex viral replicative forms (2, 7) and is * Corresponding author. dependent upon cellular functions expressed transiently in the late S or early G2 phase of the cell cycle (22,24).Of the progeny virus resulting from a single high-multiplicity infe...

The mechanism of nonhomologous recombination in murine cells infected with the parvovirus minute virus of mice (MVM) has been investigated by analysis of DNA sequences at recombination junctions in naturally occurring deletion variants of the virus. We report here that nonhomologous recombination in the MVM chromosome is characterized by short homologies, by insertion at recombination junctions of foreign DNA sequences that are enriched for preferred eucaryotic too,isomerase I cleavage sites, and by an association with a common DNA sequence motif of the type 5'-CT#TTY 3'. Additional analyses of broken MVM chromosomes provided evidence for specific enzymatic cleavage within 5'-CTTATC-3' and 5'-CTATTC-3' sequences. The results indicate that the 5'-CT TT-3' motif is an important genetic element for nonhomologous recombination in the parvovirus chromosome.In mammalian cells, nonhomologous recombination is important for evolutionary variation in gene families (13), chromosome translocations (16), gene amplification (27), movement of retroviruses (29), and integration of pseudogenes (6). Short homologies that are often associated with nonhomologous recombination events are believed either to stabilize intermediates that arise by slippage (13, 28) or to facilitate end joining of broken DNA molecules (22). Although the molecular mechanism remains obscure, eucaryotic topoisomerase I has been implicated in some aspects of nonhomologous recombination on the basis of an association of the preferred topoisomerase I cleavage sites CTT and GTT with excisional recombination crossover points (7,8).High-frequency deletion in the 5.0-kilobase linear singlestranded parvovirus chromosome proceeds via a nonhomologous recombination pathway that is mediated by 4-to 10-base-pair homologies (18). The deletions are nonrandom and range in size from approximately 2.0 to 4.5 kilobases (15,26). Despite this extensive deletion, the cis-dominant genetic elements necessary for viral DNA replication and encapsidation are selectively conserved in or near palindromic regions located at the termini of the parvovirus chromosome (9, 15; E. A. Faust and A. Hogan, in P. Tijssen, ed., Handbook of Parvoviruses, in press). Deletion variants are therefore readily obtained in encapsidated form and are separable from standard infectious virus by equilibrium density gradient centrifugation in CsCl (9, 15). The DNA can be recovered from purified virus particles and cloned in plasmid (pBR322) or phage (M13) vectors, and the DNA sequences at the deletion junctions can be determined (18). Our analysis of three recombinants has shown that deletions averaging 3 kilobases in length occur between pairs of perfectly homologous 4-to 10-base-pair direct repeats such that one copy of the repeated sequence remains at the recombination junction (18). In the present study we analyzed eight inde-* Corresponding author. t Present address:

The distribution of termination and initiation sites in a 5081-nucleotide minute virus of mice DNA template being copied by a highly purified mouse DNA polymerase a-DNA primase complex in the presence of GTP has been examined. The 3'-hydroxyl termini (17 in all) were clustered at six sites that were located 2-14 nucleotides upstream of C2A2C2, C2AC3, or C2A2T2 sequences. When either [a-32p] Highly purified preparations of the eukaryotic replicative enzyme DNA polymerase a contain a tightly bound oligoribonucleotide polymerase that is generally referred to as primase (1,2). Primase is distinguished from RNA polymerases by the fact that the length of the oligoribonucleotide synthesized is restricted (4-10 nucleotides) and by its ability to synthesize mixed oligoribo-deoxyribonucleotide chains (3-5). Primase purified from Drosophila melanogaster embryos appears to comprise polypeptides of Mr 50,000 and 60,000 that are tightly bound to a Mr 182,000 DNA polymerase catalytic subunit (6). An apparently homogeneous primase preparation from mouse hybridoma cells, free of DNA polymerase activity, consists exclusively of polypeptides of Mr 46,000 and 56,000 (7). Immunological evidence has been obtained for a tight association of primase activity with DNA polymerase a from (murine) Ehrlich ascites cells (8) and from KB cells (9). These physical and enzymological properties of DNA polymerase a-primase are entirely consistent with a role for this enzyme in the discontinuous RNA-primed DNA synthesis that occurs predominantly, but not exclusively, on the lagging strand of the DNA replication fork in mammalian cells (10, 11).The ability of highly purified preparations of mouse DNA polymerase a to copy the 5081-nucleotide linear singlestranded DNA genome of minute virus of mice (MVM), an autonomous parvovirus, has been described (12,13 Incubations were at 37°C for 2 hr and reactions were terminated by the addition of 1/5th vol of agarose gel electrophoresis sample buffer (50 mM EDTA/2 M urea/2% NaDodSO4/1 M sucrose/0.1% bromophenol blue). 4023The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.