An in vitro system in which bacteriophage T7 DNA is replicated and efficiently packaged into procapsids to form viable phage has been used to examine mutagenesis. The fidelity of replication was assayed both by measuring-reversion of an amber mutation in an essential gene and by generation of temperature-sensitive mutants among the phage produced in vitro. Under standard reaction conditions, the fidelity of DNA replication is about equal to that normally found in vivo. However, when O6_ methyldeoxyguanosine triphosphate is included in the reaction, 06-methylguanine is incorporated into newly synthesized DNA and the mutation frequencies increase 10-to 70-fold over the control. These experiments demonstrate in vitr mutagenesis with the T7 DNA replication-packaging system and provide more direct evidence for the premutagenic role of 06-methylguanine.Studies on the fidelity ofin vitro DNA synthesis have frequently been carried out with synthetic DNA templates and purified DNA polymerases (1). These measurements provide useful information but represent a significant departure from the in vivo situation. Experiments using natural DNA templates and welldefined enzyme systems have allowed valuable extensions of these studies (1). Even closer adherence to the in vivo situation is now possible by. using an in vitro system that replicates intact chromosomes of duplex T7 DNA with efficiency sufficient to replicate 2-5 times as much product DNA as the amount of exogenous template provided (2, 3). Also, biological activity of T7 DNA can be effectively monitored with an in vitro system able to encapsulate T7 DNA into proheads with good efficiency (4, 5). Thus, the ability to synthesize T7 DNA in vitro and to package that DNA into phage particles means that the entire process ofT7 self-replication can be carried out in vitro (6). In this paper, we report'mutagenesis with this in vitro system in which bacteriophage T7 DNA is replicated and then packaged into capsids to form viable phage. Using this system, we have examined the consequences of incorporating _6-methylguanine (m6Gua) into T7 DNA.Only indirect evidence supports the original proposal (7) that m6Gua is mutagenic because it may mispair with thymine and thus cause C -+ T transitions. For example, in vivo experiments with Escherichia coli show that a majority of the mutations induced in a particular gene by methylnitronitrosognanidine are of the G-C -) A-T type (8), and in vitro measurements suggest that, during DNA synthesis (9) or transcription (10), m6Gua can base pair with thymine (or uridine). Furthermore, studies ofin vitro DNA synthesis with prokaryotic DNA polymerases indicate that the presence of m6Gua in the template significantly favors incorporation of thymine rather than cytosine in the product (unpublished observation). Also, the mutagenicity of alkylating agents appears to correlate with their tendency to react at the oxygen sites in DNA, including the O6 of guanine (11). Collectively, these observations argue that direct alkylation ofDNA can prod...

DNA sequence analysis of genetic deletions in bacteriophage T7 has shown that these chromosomal rearrangements frequently occur between directly repeated DNA sequences. To study this type of spontaneous deletion in more quantitative detail synthetic fragments of DNA, made by hybridizing two complementary oligonucleotides, were introduced into the non-essential T7 gene 1.3 which codes for T7 DNA ligase. This insert blocked synthesis of functional ligase and made the phage that carried an insert unable to form plaques on a host strain deficient in bacterial ligase. The sequence of the insert was designed so that after it is put into the T7 genome the insert is bracketed by direct repeats. Perfect deletion of the insert between the directly repeated sequences results in a wild-type phage. It was found that these deletion events are highly sensitive to the length of the direct repeats at their ends. In the case of 5 bp direct repeats excision from the genome occurred at a frequency of less than 10(-10), while this value for an almost identical insert bracketed by 10 bp direct repeats was approximately 10(-6). The deletion events were independent of a host recA mutation.

When DNA from bacteriophage T7 is irradiated with UV light, the efficiency with which this DNA can be packaged in vitro to form viable phage particles is reduced. A comparison between irradiated DNA packaged in vitro and irradiated intact phage particles shows almost identical survival as a function of UV dose when Escherichia coli wild type or polA or uvrA mutants are used as the host. Although uvrA mutants perform less host cell reactivation, the polA strains are identical with wild type in their ability to support the growth of irradiated T7 phage or irradiated T7 DNA packaged in vitro into complete phage. An examination of in vitro repair performed by extracts of T7-infected E. coli suggests that T7 DNA polymerase may substitute for E. coli DNA polymerase I in the resynthesis step of excision repair. Also tested was the ability of a similar in vitro repair system that used extracts from uninfected cells to restore biological activity of irradiated DNA. When T7 DNA damaged by UV irradiation was treated with an endonuclease from Micrococcus luteus that is specific for pyrimidine dimers and then was incubated with an extract of uninfected E. coli capable ofremoving pyrimidine dimers and restoring the DNA ofits original (whole genome size) molecular weight, this DNA showed a higher packaging efficiency than untreated DNA, thus demonstrating that the in vitro repair system partially restored the biological activity of UV-damaged DNA.

An in vitro system based on extracts of Escherichia coli infected with bacteriophage T7 was used to study genetic deletions between directly repeated sequences. The frequency of deletion was highest under conditions in which the DNA was actively replicating. Deletion frequency increased markedly with the length of the direct repeat both in vitro and in vivo. When a T7 gene was interrupted by 93 bp of nonsense sequence flanked by 20-bp direct repeats, the region between the repeats was deleted in about 1 out of every 1,600 genomes during each round of replication. Very similar values were found for deletion frequency in vivo and in vitro. The deletion frequency was essentially unaffected by a recA mutation in the host. When a double-strand break was placed between the repeats, repair of this strand break was often accompanied by the deletion of the DNA between the direct repeats, suggesting that break rejoining could contribute to deletion during in vitro DNA replication.Genetic deletions often occur between directly repeated sequences of DNA (1, 2, 5, 8). The presence of homologous sequences from 6 to 20 bp long at the ends of deleted regions brings to mind mechanisms that cause deletion by annealing complementary regions from direct repeats on opposite ends of the deleted sequence. Various models have been proposed to account for deletion mechanisms (2,5,8,11,18,24,31). Most of these can be classified into mechanisms that involve a form of copy choice during DNA synthesis (16,17,34,39,40), recombination between the direct repeats (2,3,8,10,19,30,33), or incorrect rejoining of DNA molecules which are broken as part of normal DNA metabolic activities (4,8,11,18,31,32,38). Rearrangements caused by switching of subunits between gyrase molecules resident at each of the direct repeats can be included in this latter category (13). The presence of nearby ends on DNA molecules undergoing deletion has been suggested as a common feature among these putative mechanisms (9). Achieving a clear understanding of deletion mechanisms has been difficult because of the complexity of biological events involved and the relatively infrequent occurrence of deletion. Nonetheless, the important roles of deletion in human inherited diseases, evolution, and biotechnology, as well as the potential insight into DNA replication and recombination mechanisms that may be reached by gaining a better understanding of how deletions occur, provide strong incentives for studying deletion mechanisms. In this study, we tested whether double-strand ends can stimulate deletion between direct repeats.We have been investigating deletions between direct repeats in bacteriophage T7 (14,25,26,28,29). The relative simplicity of its DNA replication process (27, 36) and the abundant biochemical and genetic information regarding T7 that has already been amassed (7,27,35,36) make T7 an especially interesting system with which to study deletions. We have adapted an in vitro DNA replication-packaging system that can