360 NATURE BIOTECHNOLOGY VOL 17 APRIL 1999 http://biotech.nature.com RESEARCH Herpesviruses are large complex DNA viruses widely found in nature. Human cytomegalovirus (hCMV), a member of the b-her- pesvirus subfamily, is an important human pathogen that causes severe and even fatal disease in immunocompromised individuals. Animal CMVs, such as the mouse cytomegalovirus (mCMV), serve as model systems. The human and murine CMV genomes are the largest among the herpesviruses, comprising approximately 230 kb and more than 200 open reading frames (ORFs). Although the genomes for both viruses have been completely sequenced 1,2 , the function of most of the genes remains unknown. Fewer than 50 CMV genes are conserved among the herpesviruses, based on amino acid homology and genome location 3 . While most of the genes have no apparent homology to known viral or cellular genes, some are known to be important for virus–host interaction and immunoevasion 4 . Essential genes are of particular importance, as they may serve as targets for antiviral chemotherapy. Moreover, inactivation of at least one essential gene is required to create replica- tion-incompetent herpesvirus vectors for gene therapy, while nonessential genes can be removed to gain space for insertion of therapeutic genes 5 , or to generate attenuated vaccines. While temperature-sensitive (ts) and null mutants have played a crucial role in mapping and assigning functions of herpesviral genes 6 , working with these mutants can be arduous. For ts mutants, the mutation(s) responsible for an observed phenotype are difficult to pinpoint because the viral genome may contain other silent muta- tions. Production of null mutants by site-directed mutagenesis is laborious, requiring construction of individual mutants on a gene- by-gene basis. CMV’s large genome and slow replication make homologous recombination in cell culture tedious, requiring many rounds of selection and plaque purification 7 . Deletion of essential genes can also be time-consuming , requiring complementing cell lines that can be difficult to produce, especially if the essential virus protein is toxic to the cell. To help make herpesviruses accessible to bacterial genetic tools, we recently developed a method of cloning and mutagenizing the mCMV genome as an infectious bacterial artificial chromosome (BAC) 8 . Since then, BAC technology has been applied to the genomes of Epstein–Barr virus and herpes simplex virus (HSV) 9,10 , and other herpesvirus genomes are to follow. In this study, we adapted a mini-transposon (Tn) system for direct and random mutagenesis of the complete infectious mCMV genome. Sites of Tn insertion were pinpointed by direct sequencing of the mCMV BAC. We show that this method can be used to deter- mine rapidly whether herpesvirus genes are essential for viral growth. Results Modification of the TnMax transposon system for direct mutagen- esis of BACs. TnMax is a mini-Tn system derived from Tn1721 with a strong preference for transposition into negatively supercoiled plasmids 11–13 . As a member of the Tn3 family, it utilizes a replicative mechanism of transposition requiring transposase (tnpA), resolvase (tnpR), and a resolution site (res) located on the transposable ele- ment 11 . This method of TnMax mutagenesis 12 proved to be inefficient for mutagenesis of mCMV BAC because isolation of BACs and transformation of new bacteria were required to eliminate the TnMax donor plasmid. To avoid this step, we fused TnMax plas- mids with pST76-A, a plasmid with a ts mechanism of replication 14 . Three Tn donor plasmids were constructed (pTsTM2, pTsTM8, and pTsTM13) containing TnMax derivatives with different antibiotic resistance genes (Fig. 1). Rapid identification of essential and nonessential herpesvirus genes by direct transposon mutagenesis Wolfram Brune, Carine Ménard, Urs Hobom, Stefan Odenbreit 1 , Martin Messerle, and Ulrich H. Koszinowski* Department of Virology and 1 Department of Bacteriology, Max von Pettenkofer-Institut, Ludwig-Maximilians-Universität München, Pettenkoferstraße 9a, 80336 Munich, Germany. *Corresponding author (e-mail: koszinowski@m3401.mpk.med.uni-muenchen.de). Received 8 October 1998; accepted 9 February 1999 Herpesviruses are important pathogens in animals and humans. The large DNA genomes of several herpesviruses have been sequenced, but the function of the majority of putative genes is elusive. Determining which genes are essential for their replication is important for identifying potential chemotherapy targets, designing herpesvirus vectors, and generating attenuated vaccines. For this pur- pose, we recently reported that herpesvirus genomes can be maintained as infectious bacterial artificial chromosomes (BAC) in Escherichia coli. Here we describe a one-step procedure for random-insertion mutagenesis of a herpesvirus BAC using a Tn1721-based transposon system. Transposon insertion sites were determined by direct sequencing, and infectious virus was recovered by transfecting cultured cells with the mutant genomes. Lethal mutations were rescued by cotransfecting cells containing noninfec- tious genomes with the corresponding wild-type subgenomic fragments. We also constructed revertant genomes by allelic exchange in bacteria. These methods, which are generally applicable to any cloned herpesvirus genome, will facilitate analysis of gene function for this virus family. Keywords: BAC, transposon mutagenesis, cytomegalovirus, functional genomics © 1999 Nature America Inc. • http://biotech.nature.com © 1999 Nature America Inc. • http://biotech.nature.com