- Email: [email protected]
Gene Targeting for Gene Therapy: Prospects
Molecular Genetics and Metabolism 68, 276 –282 (1999) Article ID mgme.1999.2910, available online at http://www.idealibrary.com on
MINIREVIEW Gene Targeting for Gene Therapy: Prospects Vladislav A. Lanzov Division of Molecular and Radiation Biophysics, Petersburg Nuclear Physics Institute, Russian Academy of Sciences, Gatchina/St. Petersburg 188350, Russia Received July 7, 1999; and in revised form July 28, 1999
One of the basic findings of molecular medicine is that not only inherited but also most acquired diseases have a genetic component. The purpose of gene therapy is to modify this component and, at a minimum, to alter the course of disease or, at a maximum, to arrest the disease at its source (1,2). Every gene therapy procedure includes three consecutive steps: the delivery of therapeutic DNA in vivo or ex vivo; the integration (or lack thereof) of this DNA into a host genome; and the expression of the transgene formed. Different therapeutic tasks dictate the various strategies of gene therapy. The use of nonintegrative, integrative, or replacement vectors for DNA delivery (3) will result in a transient, long-term, or permanent transgene expression. For example, some applications of gene therapy for cancer treatment (such as antisense therapy, immunotherapy, prodrug therapy, etc. (4)) using the therapeutic DNA to produce pharmaceutical agents may respond to the transient expression of transgene products; treatment of diseases caused by recessive mutations, on the other hand, requires a prolonged expression coupled with either a random (gene complementation) or a precise (gene targeting) integration of the transferred DNA. Finally, diseases resulting from dominant mutations can be treated effectively only via gene targeting. In fact, the precise replacement of damage is an ideal variant of gene therapy, since this strategy overcomes many obstacles and allows control of the duration, the level, and the place of transgene expression. Gene targeting is the homologous pairing of extrachromosomal and chromosomal DNA molecules. It
Ideally, gene therapy involves the correction of genetic defects through the natural means of gene targeting. This therapy possesses a number of conceptual advantages. However, a major obstacle to successful gene therapy is the relative inefficiency of the targeting process in mammalian cells. Gene targeting may be accomplished by two different mechanisms: the homologous recombination and the mismatch correction of DNA heteroduplexes. Based on the model of homologous recombination for the wellstudied prokaryotic and the less studied eukaryotic systems, three approaches have been employed to improve the efficiency and accuracy of homologous recombination events. These are: (1) artificial doublestrand breaks in both the exogenous and the chromosomal DNA, (2) a contiguous long homology between the exogenous and chromosomal DNA, and (3) a transient overproduction of an active recombinase, the bacterial RecA or mammalian RecA-like proteins, in mammalian cell nuclei. Combining these approaches can result in more effective gene targeting protocols. The second mechanism has been improved based on recent observations of recombinogenic activity of oligonucleotides and, especially, specifically designed chimeric RNA/DNA oligonucleotides. The use of RecA-like proteins to stimulate searching for homology and forming stable DNA heteroduplexes between oligonucleotides and chromosomal DNA remains an attractive idea for additional improvement of gene targeting events. © 1999 Academic Press
Key Words: chimeric RNA/DNA oligonucleotides; homologous (targeting) recombination; gene therapy; mismatch correction systems; RecA/Rad51/ Rec2-like proteins. 276 1096-7192/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
GENE TARGETING FOR GENE THERAPY
277
results from two different mechanisms of DNA repair: the homologous recombination (5) and the mismatch correction of DNA heteroduplexes (6,7). This minireview addresses both mechanisms, as well as recently discovered ways of improving gene targeting. Multiple aspects of gene targeting therapy have been previously reviewed (7–13). GENE TARGETING BY HOMOLOGOUS RECOMBINATION Homologous recombination (HR) signifies the exchange of genetic material between two homologous DNA molecules that is essential to all organisms. HR participates in maintaining genome integrity, generating genetic diversity, and segregating chromosomes properly. Our understanding of HR, both the genetics and the biochemistry of the process, is based primarily on observations made in Escherichia coli. To date, more than 20 E. coli proteins involved in recombination have been described (14) and subdivided according to their functions in the consecutive steps of this process (15). The initial phase of HR is DNA end-dependent. Figure 1 demonstrates this introductory step (step 1) with doublestrand breaks (DSBs) which are well-known inducers of HR in higher eukaryotes (17,18). Each end of DSBs formed either in donor or (as can be seen in the figure) in recipient DNA is resected by a 59exonuclease (probably acting together with a helicase) to produce long 39-terminal single-stranded DNA (ssDNA) overhangs (step 2) which are covered by a RecA-like protein to form the recombinationally active nucleoprotein filament. The latter triggers a search for homology with partner double-stranded DNA (dsDNA) and switching of pairing between donor and recipient helices. This results in strand exchange and heteroduplex DNA formation (step 3) followed by DNA gap repair and formation of recombination intermediates known as Holliday structures (step 4). These structures can be extended and/or migrated by branch migration proteins (step 5) and finally cleaved by an X-shaped DNA recognizing protein, a DNA resolvase. Two possible modes of DNA cleavage with resolvase produce two types of recombinant DNA products (step 6). Note that both products contain small regions of heteroduplex DNA which, as shown in Fig. 1, can be the regions of imperfect homology and, therefore, become a substrate of mismatch correction systems to be repaired to donor or recipient nucleotide sequence (19). Thus,
FIG. 1. Homologous recombination coupled to the mismatch correction of DNA heteroduplexes. The scheme is based on the double-strand break repair model (16). The donor DNA fragment and recipient DNA are shown by thick and thin lines, respectively. Open and filled circles indicate, respectively, donor and recipient markers located in DNA heteroduplex regions. Numbers from 1 to 7 show consecutive steps of the process (for details, see text). Arrows and double arrows indicate two ways for resolution of recombinant structures into two configurations.
the correction systems play a positive or negative role in recombinant product fixation (step 7). The counterparts (structural homologs or functional analogs) of E. coli proteins participating in steps 1 to 4 have been revealed in Saccharomyces cerevisiae (15). These give evidence of conceptual similarity of the HR process in eubacteria and eukaryotes. Two differences have, however, become clear: first, the number of homologs and complexity of analogs appear to be much higher in S. cerevisiae; second, there is complex genetic control of the initial step in S. cerevisiae (17,18). In mammalian cells the situation appears more complicated still. Unlike
278
VLADISLAV A. LANZOV
yeast cells, which commonly integrate DNA in chromosomal sites with targeted events outnumbering random events by more than 10 to 1, this ratio is on average 1 to 1000 in mammalian cells (reviewed in Ref. 20). This discrepancy results mainly from the end-joining mechanism of mammalian cells that sticks DNA ends together, arising primarily from errors in DNA metabolism, regardless of the terminal sequences (20). This mechanism is significantly less common in yeast cells and virtually unknown in bacteria. Although HR is a multistep process, for the purposes of gene targeting improvement only several initial steps seem to be crucial. These include, first, directed formation of DSB both in the donor and in the targeted chromosomal site; second, the length and homology perfection of donor DNA; and third, the amount and recombinogenic ability of RecA-like proteins promoting the pairing and strand exchange. DSBs DSBs created by rare-cutting endonucleases have been shown to stimulate mitotic homologous recombination in mammalian cells (21,22). In principle, a similar approach had been developed earlier for two site-specific recombination systems, the P1 bacteriophage Cre-Lox and the yeast FLP systems. The Cre protein catalyzes integration of P1 phage into chromosomal lox sites. In mammalian cells, a pure Cre protein introduced by lipofection causes both the excision of DNA sequences flanked by lox sites and targeted insertion of the lox-containing vector at a chromosomal lox site (23). The yeast recombinase Flp stimulates intrachromosomal recombination between FLP target sites integrated in the human b-globin locus in MEL cells (24). The most impressive effect of DSBs induced in vivo with the S. cerevisiae mitochondrial endonuclease I-SceI has been recently shown in male mouse embryonic stem (ES) cells (22). This work involved targeting recombination between an extrachromosomal segment of the hprt gene and the entire chromosomal hprt gene. This gene is a classical model for the analysis of gene targeting events (25) because it is on the X chromosome and thus exists as a single copy in XY ES cells. In addition, negative–positive selections facilitate observation of both the loss and the restoration of the hprt function. The frequency of gene targeting in clones with an I-SceI site in the presence of the endonuclease appears 5000 times
higher than that in clones without this site. Moreover, this frequency decreases with distance from the DSB site. Together with the observation that introduction of DSBs into targeting vectors significantly increases the accuracy of gene targeting (26,27), these data support the idea of artificial DSB production as potential recombinogenic sites for directed gene targeting events. Length and Perfection of DNA Homology In wild-type E. coli, HR requires a minimum length of homology of about 23 to 27 bp, and single mismatches in 31-bp substrates cause reductions in recombinant frequencies from 2- to 12-fold (28). In addition, the efficiency of HR has been found to be proportional to the length of homology between the donor and acceptor DNA in the range between 175 and 1200 bp (29). Thus, in bacteria, the efficiency of HR depends on both the extent and the perfection of homology between recombining partners. Similar conclusions have been reached regarding gene targeting in mammalian cells. The targeted inactivation of the endogenous hprt gene in male mouse ES cells shows that the targeting frequency is strongly dependent on the length of homology between the targeting vector and target locus (30). In the range between 2 and 10 kb, this dependence is exponential, yielding a 100-fold increase in gene targeting frequency and reaching saturation with respect to length of homology at about 14 kb. In addition, targeting efficiency appears dependent on the perfection of homology between donor and recipient DNA; vectors prepared from isogenic DNA appear to be 4 to 5 times more efficient than corresponding vectors carrying nonisogenic DNA (30). The obvious explanation for this last finding is the interference between HR and the mismatch repair (MMR) system of the cell. The recombination between nonisogenic DNAs creates many mismatch bases in heteroduplex regions. Because the correction of recombination heteroduplexes by the MMR system is not a strand-directed process, simultaneous exonuclease degradation proceeding on both DNA strands must result in DNA destruction and cell death (31). RecA-like Proteins Figure 1 demonstrates the DNA pairing and strand transfer forming the central step of HR. In bacteria, this step is controlled by RecA and several helper proteins. The RecA protein binds ATP, and then, through a ssDNA binding (SSB) protein which
GENE TARGETING FOR GENE THERAPY
removes ssDNA secondary structures, polymerizes on ssDNA to form continuous spiral presynaptic filament(s) (14). The complex of RecO and RecR proteins facilitates this polymerization while the complex of RecF and RecR proteins halts the filament extension (32). These filaments attach to dsDNA partner molecules, form paranemic three-stranded DNA structures, switch pairing within the structures, and initiate processing of DNA heteroduplexes which result in Holliday structures. Figure 1 is an in vivo picture based mainly on an analysis of the in vitro strand transfer reaction visualizing the transfer of one strand of bacteriophage M13 (or fX174) linear dsDNA to a closed circular ssDNA of the same phage. The RecA protein is universal in nature (15,33). Rad51 proteins of lower and higher eukaryotes including human HsRad51 were shown to be the structural and functional homologs of RecA (34). Interestingly, an SSB protein analog in eukaryotes, the RPA complex, can be effectively substituted by bacterial SSB in some strand exchange reactions (35). These findings show the universal nature of enzymes performing the central step of HR and address the question of whether the introduction of taxonomically distinct RecA-like proteins into mammalian cells can improve gene targeting. The following three observations suggest the feasibility of such an approach: first, a correlation between the level of HR and the extent of HsRAD51 gene transcription in immortal human cells (36); second, the 20-fold stimulation of spontaneous intrachromosomal HR in CHO cells by a 2- to 3-fold overproduction of the CgRad51 protein from Chinese hamster Cricetulus griseus (37); and third, the 10-fold stimulation of HR in transgenic tobacco cells producing a nucleus-targeted E. coli RecA (38). Last, recent research using the system for production of a functional E. coli RecA protein (modified with a nuclear location signal) in nuclei of mammalian cells (39) has shown a 10-fold stimulation of HR within the hprt gene of male mouse F9 cells (40). Among different bacterial RecA proteins studied to date, the Pseudomonas aeruginosa RecA manifests a hyperrecombination activity, initiating recombination 6 to 7 times more frequently than the E. coli RecA (41). Moreover, some chimeric RecA proteins, structural hybrids between P. aeruginosa and E. coli RecAs, appear still more aggressive in initiating HR (IV Bakhlanova, T Ogawa, VA Lanzov, manuscript in preparation). Experimentation is now underway to test whether these unusual proteins
279
maintain their hyperrecombinogenic potential in mammalian cell nuclei. Recently, a new human HsREC2 gene has been isolated. It is a structure–function homolog of the REC2 gene from the corn smut, Ustilago maydis (42). Unlike the latter, HsREC2 is structurally similar to human HsRAD51 and is expressed in a wide range of tissues. In addition, its production is induced by ionizing radiation. Because this protein obviously takes part in the mechanism of DSB repair, its overproduction can improve the cell enzymatic machinery promoting gene targeting events (43). Other Factors A new protocol of ex vivo gene targeting in mouse ES cells has been proposed (44). It allows an increase in the absolute targeting frequency at the hprt gene from 10 25 to 10 21. This impressive result has been achieved at an unusually high plating density of electroporated cells followed by the delay of colony selection for 60 h. If this method can be reproduced in other experimental systems, it opens new possibilities for gene targeting without the use of positive–negative selection. GENE TARGETING BY MISMATCH CORRECTION Mismatches arising as errors in DNA metabolism or among intermediates (heteroduplexes) of HR are repaired by several mechanisms found ubiquitously (19). In bacteria, the most substantial is a methylation-directed long patch MMR operated by the MutHLS enzymatic complex (45). In fact, all 8 base– base mismatches (except possibly «C–C») can be corrected by this system. The system has been reconstructed in vitro and its sequential steps have been well documented. Normally, DNA in E. coli is methylated at GATC sites by a special Dam methylase. After DNA replication, the daughter strand appears transiently unmethylated and thus becomes a substrate for the repair reaction. The latter is initiated by binding of MutS protein to a mismatch followed by binding of MutL to stabilize the MutSL complex and activate MutH endonuclease which nicks the unmethylated strand of hemimethylated DNA at a GATC site located up to 1–2 kb from the mismatch. Consecutive steps of DNA excision and resynthesis result in producing a long oligomer patch in the corrected DNA strand. It is noteworthy that such a
280
VLADISLAV A. LANZOV
gram-positive bacterium as Streptococcus pneumoniae has the Hex system that is similar to MutHLS but is directed by a nick in the daughter DNA and thus has no necessity in a MutH-like endonuclease. Structure–function homologs of the MutHLS systems have been found both in yeast and in human cells (6,46). The human MMR system can repair not only 8 mismatches, among which «C–C» is the weakest substrate, but also small insertions and deletions, from 8 to 12 nucleotides in length. Like the bacterial Hex, the mammalian MMR has no MutH analog and probably performs a nick-directed correction (6). As expected, eukaryotic systems appear much more complex. The human MMR contains two heterodimer complexes, MutSa (MSH2 1 MSH6) and MutSb (MSH2 1 MSH3), of the MutS protein family and heterodimers MutLa (MLH1 1 PMS2) and MutLb (MLH1 1 PMS1) of the MutL family. Combinations of these MutS and MutL complexes possess a definite substrate specificity (6). Nucleotide excision repair (NER) is another universal repair mechanism responsible for removing damaged bases (adducts, fusion, etc.) (47). NER produces two coordinated incisions around the damaged base to excise a small oligomer of approximately 28 –29 nucleotides and then correctly resynthesizes the gap formed. As has been recently proved (48), the short patch NER system of Schizosaccharomyces pombe can correct C–C mismatches and, to a lesser extent, other base– base mismatches, efficiently and independently of MMR. In light of findings of cooperative interactions between some proteins of NER and MMR systems in mammalian cells in order to remove large loops (reviewed in Ref. (46)), it becomes clear that mammalian cells have an efficient protection from various types of mismatches in the heteroduplex DNA formed during either DNA replication or recombination. Gene targeting via mismatch correction implies the introduction of a designed short DNA fragment or oligonucleotide into mammalian cells, pairing of the fragment with recipient DNA to form a heteroduplex region, and correction of the recipient sequence to a designed one by one of the mismatch correction systems. RecA-like Proteins The human genome contains at least a 10 9-fold excess of nonhomologous to homologous sites, so that a search for the homologous site must proceed
in a rapid manner. Evidence exists (49) that the rate of synaptic complex formation between short oligonucleotides (24 –27 nucleotides) and linear (57 bp) or even supercoiled (2.7 kb) DNA in the presence of a sufficient amount of the E. coli RecA protein exibits second-order kinetics. In other words, this rate depends on the concentration of both oligonucleotides and targeted duplex DNA. Notably, the extent of complexity of heteroduplex DNA has no effect on the reaction rate discussed. Thus, if targeted cells contain an appropriate amount of the RecA-like protein and absorb a sufficient number of extrachromosomal DNA molecules, the search for homology is not a rate-limiting step in gene targeting. Another advantage of RecA protein-assisted pairing that is useful for gene targeting is that RecA has the ability to decrease the fidelity of homology recognition in synaptic complexes between an oligonucleotide and a supercoiled plasmid (50). In principle, such a property should facilitate heteroduplex DNA formation. A Small Fragment Homology Replacement (SFHR) Gene targeting by SFHR has recently been tested in the correction of the mutant cystic fibrosis transmembrane conductance regulator (CFTR) gene to its wild-type sequence (51). The 491-bp fragments of genomic CFTR DNA encapsulated in liposome or polyamidoamine vehicles have been introduced into transformed epithelial cells containing the CFTR gene with the most commonly distributed mutation, DF508. About 1 in each 100 cells restored the defect in Cl 2 ion transport giving evidence of successful gene targeting. Unfortunately, a preliminary coating of DNA fragments by E. coli RecA protein does not enhance the homologous replacement. The literature suggests that this may occur because the conditions for RecAassistant reaction have not been optimized (51). Chimeric RNA/DNA Oligonucleotides A new strategy for gene targeting by oligonucleotides has recently been developed (7). This new strategy is based on the discovery of high recombinogenic activity of the RNA stretch within RNA/ DNA small chimeric molecules designed as a continuous RNA/DNA complementary duplex constructed in a double-hairpin configuration (52). Relative to analogous DNA/DNA molecules, RNA/DNA constructs bind the targeted DNA sequence packaged
GENE TARGETING FOR GENE THERAPY
into nucleosomal structures 20-fold more efficiently (7,43). This strategy has been applied to correct (with relatively high frequency) a single base alteration in the alkaline phosphatase gene of HuH-7 human hepatoma cells (53); to induce in vivo a point mutation in the rat factor IX gene that resulted in a significant reduction of this factor coagulant activity (54); to correct the mutation in the hemoglobin b s allele for sickle cell anemia in lymphoblastoid cells homozygous for this mutation (55); and to induce a sickle cell anemia mutation in normal lymphoblastoid cells (43). FINAL REMARKS The past decade has provided remarkable progress in the development of new gene targeting strategies. However, our current knowledge of homologous recombination mechanisms in higher eukaryotes is only the first sketch for a much more complicated picture of recombinational events coupled with different modes of the DNA heteroduplex correction. In fact, homologous recombination in prokaryotes is not limited by RecA-dependent pathways. Both lambda-type bacteriophages and some E. coli bacteria possess, respectively, Redab and RecET protein systems combining activities of Reda or RecE exonucleases with Redb or RecT ATP-independent annealase to promote strand exchange (56). Significantly, both Redab and RecET systems also promote DNA double-strand break repair (57) and thus serve as prototypes for analogous systems in higher eukaryotes (reviewed in Ref. (56)). Finding such systems in human cells can open new perspectives for improvement of therapeutic gene targeting. ACKNOWLEDGMENTS I am grateful to Dr. Eugene Schwartz (PNPI) for critical reading of the manuscript, Yuri Kil (PNPI) for help with figure preparation, and Natasha Blinkova for English correction. I apologize to those researchers whose contributions were not cited because of space limitation in this minireview. This work was supported by an International Research Scholar’s award from the Howard Hughes Medical Institute, Grant 75195-546101.
281
3.
Kay MA, Liu D, Hoogerbrugge: Gene Therapy. Proc Natl Acad Sci USA 94:12744 –12746, 1997.
4.
Blaese RM: Gene therapy for cancer. Sci Am June:91–95, 1997.
5.
Capecchi, MR: Altering the genome by homologous recombination. Science 244:1288 –1292, 1989.
6.
Modrich P: Strand-specific mismatch repair in mammalian cells. J Biol Chem 272:24727–24730, 1997.
7.
Kmiec EB: Genomic targeting and genetic conversion in cancer therapy. Semin Oncol 23:188 –193, 1996.
8.
Waldman AS: Targeted homologous recombination in mammalian cells. Crit Rev Oncol/Hematol 12:49 – 64, 1992.
9.
Doi S, Campbell C, Kucherlapati R: Directed modifications of genes by homologous recombination in mammalian cells. In Transgenic Animals (Grosveld F, Kollias G, Eds.) London: Academic Press, pp 27– 46, 1992.
10.
Lanzov VA: Ideal gene therapy: approaches and horizons. Mol Biol 28:321–327, 1994.
11.
Kowalczykowski SC, Zarling DA: Homologous recombination proteins and their potential applications in gene targeting technology. In Gene Targeting (Vega MA, Ed.). Boca Raton: CRC Press, pp 167–210, 1995.
12.
Vega MA: Gene targeting in human gene therapy. In Gene Targeting (Vega MA, Ed.). Boca Raton: CRC Press, pp 211– 229, 1995.
13.
Yanez RJ, Porter ACG: Therapeutic gene targeting. Gene Ther 5:149 –159, 1998.
14.
Kowalczykowski SC, Dixon DA, Egglestone AK, Lauder SD, Rehrauer WM: Biochemistry of homologous recombination in Escherichia coli. Microbiol Rev 58:401– 465, 1994.
15.
Bianco PR, Trasy RB, Kowalczykowski SC: DNA strand exchange proteins: a biochemical and physical comparison. Frontiers Bio Sci 3:570 – 603, 1998.
16.
Szostak JW, Orr-Weaver TL, Rothstein RJ, Stahl FW: The double-strand break repair model for recombination. Cell 33:25–35, 1983.
17.
Shinohara A, Ogawa T: Homologous recombination and the roles of double-strand breaks. Trends Biochem Sci 20:387– 391, 1995.
18.
Haber JE: A super new twist on the initiation of meiotic recombination. Cell 89:163–166, 1997.
19.
Freidberg EC, Walker GC, Siege W: DNA Repair and Mutagenesis. Washington, DC: ASM Press, pp 367– 405, 1995.
20.
Roth D, Wilson J: Illegitimate recombination in mammalian cells. In Genetic Recombination (Kucherlapati R, Smith GR, Eds.). Washington, DC: ASM Press, pp 621– 653, 1988.
21.
Jasin M: Genetic manipulation of genomes with rare-cutting endonucleases. Trends Genet 12:224 –228, 1996.
22.
Donoho G, Jasin M, Berg P: Analysis of gene targeting and intrachromosomal homologous recombination stimulated by genomic double-strand breaks in mouse embryonic stem cells. Mol Cell Biol 18:4070 – 4078, 1998.
REFERENCES 1.
Anderson WF: Prospects for human gene therapy. Science 226:401– 409, 1984.
23.
Boubonis W, Saucer B: Genomic targeting with purified Cre recombinase. Nucleic Acids Res 21:2025–2029, 1993.
2.
Verma IM, Somia N: Gene therapy: promises, problems and prospects. Nature 389:239 –242, 1997.
24.
Fiering S, Kim CG, Epner EM, Groudin M. An “in– out” strategy using gene targeting and FLP recombinase for the functional dissection of complex DNA regulatory elements:
282
25.
26.
27.
28.
29.
30.
31.
32. 33.
34.
35.
36.
37.
38.
39.
40.
41.
VLADISLAV A. LANZOV analysis of the b-globine locus control region. Proc Natl Acad Sci USA 90:8469 – 8473, 1993. Thomas KR, Capecchi MR: Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51:503–512, 1987. Subramani S, Seaton BL: Homologous recombination in mitotically dividing mammalian cells. In Genetic Recombination (Kucherlapati R, Smith GR, Eds.). Washington, DC: ASM Press, pp 549 –573, 1988. Valancius V, Smithies O: Double-strand gap repair model in a mammalian gene targeting reaction. Mol Cell Biol 11: 4389 – 4397, 1991. Shen P, Huang HV: Homologous recombination in Escherichia coli: dependence on substrate length and homology. Genetics 112:441– 457, 1996. Ogawa T, Shinohara A, Ogawa H, Tomizawa J: Functional structures of the RecA protein found by chimera analysis. J Mol Biol 226:651– 660, 1992. Thomas KR, Deng C, Capecchi MR: High-fidelity gene targeting in embryonic stem cells by using sequence replacement vectors. Mol Cell Biol 12:2919 –2923, 1992. Rayssiguier C, Thaler DS, Radman M: The barrier to recombination between E. coli and S. typhimurium is disrupted in mismatch-repair mutants. Nature 342:396 – 401, 1989. Cox MM: A broadening view of recombinational DNA repair in bacteria. Genes Cells 3:65–78, 1998. Roca AI, Cox MM: RecA protein: structure, function, and role in recombinational DNA repair. Prog Nucleic Acid Res Mol Biol 56:129 –223, 1997. Brendel V, Brocchieri L, Sandler SJ, Clark AJ, Karlin S: Evolutionary comparisons of RecA-like proteins across all major kingdoms of living organisms. Mol Evol 44:528 –541, 1997. Sugiyama T, Zaitseva EM, Kowalczykowski SC: A singlestranded DNA-binding protein is needed for efficient presynaptic complex formation by the Saccharomyces cerevisiae Rad51 protein. J Biol Chem 272:7940 –7945, 1997. Xia SJ, Shammas MA, Reis RJS: Elevated recombination in immortal human cells is mediated by HsRAD51 recombinase. Mol Cell Biol 17:7151–7158, 1997. Vispe S, Cazaux C, Lesca C, Defais M: Overexpression of Rad51 protein stimulates homologous recombination and increases resistance of mammalian cells to ionizing radiation. Nucleic Acids Res 26:2859 –2864, 1998. Reiss B, Klemm M, Kosak H, Schell J: RecA protein stimulates homologous recombination in plants. Proc Natl Acad Sci USA 93:3094 –3098, 1996. Kido M, Yoneda Y, Nakanishi N, Uchida T, Okada Y: Escherichia coli RecA protein modified with a nuclear location signal binds to chromosomes in living mammalian cells. Exp Cell Res 198:107–114, 1992. Scherbakova OG, Ogawa H, Lanzov VA, Filatov MV: Targeting recombination in somatic mammalian cells: stimulating effect of bacterial RecA protein. Doklady Rus Acad Sci, in press, 1999. Namsaraev EA, Baitin DM, Bakhlanova IV, Alexseyev AA,
Ogawa H, Lanzov VA: Biochemical basis of hyper-recombinogenic activity of Pseudomonas aeruginosa RecA protein in Escherichia coli cells. Mol Microbiol 27:727–738, 1998. 42.
Rice MC, Smith ST, Bullrich F, Havre P, Kmiec EB: Isolation of human and mouse genes based on homology to REC2, a recombinational repair gene from the fungus Ustilago maydis. Proc Natl Acad Sci USA 94:7417–7422, 1997.
43.
Cole-Strauss A, Noe A, Kmiec EB: Recombinational repair of genetic mutations. Antisense Nucleic Acid Drug Dev 7:2111– 2116, 1997.
44.
Templeton NS, Roberts DD, Safer B: Efficient gene targeting in mouse embrionic stem cells. Gene Ther 4:700 –709, 1997.
45.
Kolodner RD: Mismatch repair: mechanisms and relationship to cancer susceptibility. Trends Biochem Sci 20:397– 401, 1995.
46.
Sancar A: Excision repair invades the territory of mismatch repair. Nature Genet 21:247–249, 1999.
47.
Sancar A: Excision repair in mammalian cells. J Biol Chem 27:15915–15918, 1995.
48.
Fleck O, Lehmann E, Schar P, Kohli J: Involvement of nucleotide-excision repair in msh2 pms1-independent mismatch repair. Nature Genet 21:314 –317, 1999.
49.
Yancey-Wrona JE, Camerini-Otero RD: The search for DNA homology does not limit stable homologous pairing promoted by RecA protein. Curr Biol 5:1149 –1158, 1995.
50.
Malkov VA, Sastry L, Camerini-Otero R: RecA protein assisted selection reveals a low fidelity of recognition of homology in a duplex DNA by an oligonucleotide. J Mol Biol 271:168 –177, 1997.
51.
Kunzelmann K, Legendre J-Y, Knoell DL, Escobar LC, Xu Z, Gruenert DC: Gene targeting of CFTR DNA in CF epithelial cells. Gene Ther 3:859 – 867, 1996.
52.
Kotani H, Germann H, Andrus A, Vinayak R, Mullah B, Kmiec EB: RNA facilitates RecA-mediated DNA pairing and strand transfer between molecules bearing limited regions of homology. Mol Gen Genet 250:626 – 634, 1996.
53.
Kren BT, Cole-Strauss, Kmiec EB, Steer CJ: Targeted nucleotide exchange in the alkaline phosphatase gene of HuH-7 cells mediated by a chimeric RNA/DNA oligonucleotide. Hepatology 25:1462–1468, 1997.
54.
Kren BT, Bandyopadhyay P, Steer CJ: In vivo site-directed mutagenesis of the factor IX gene by chimeric RNA/DNA oligonucleotides. Nature Med 4:285–290, 1998.
55.
Cole-Strauss A, Yoon K, Xiang Y, Byrne BC, Rice MC, Gryn J, Holloman WK, Kmiec EB: Correction of the mutation responsible for sickle cell anemia by an RNA–DNA oligonucleotide. Science 273:1386 –1389, 1996.
56.
Hall SD, Kolodner RD: Homologous pairing and strand exchange promoted by the Escherichia coli. Proc Natl Acad Sci USA 91:3205–3209, 1994.
57.
Yokochi T, Kusano K, Kobayashi I: Evidence for conservative (two progeny) DNA double-strand break repair. Genetics 139:5–17, 1995.
MINIREVIEW Gene Targeting for Gene Therapy: Prospects Vladislav A. Lanzov Division of Molecular and Radiation Biophysics, Petersburg Nuclear Physics Institute, Russian Academy of Sciences, Gatchina/St. Petersburg 188350, Russia Received July 7, 1999; and in revised form July 28, 1999
One of the basic findings of molecular medicine is that not only inherited but also most acquired diseases have a genetic component. The purpose of gene therapy is to modify this component and, at a minimum, to alter the course of disease or, at a maximum, to arrest the disease at its source (1,2). Every gene therapy procedure includes three consecutive steps: the delivery of therapeutic DNA in vivo or ex vivo; the integration (or lack thereof) of this DNA into a host genome; and the expression of the transgene formed. Different therapeutic tasks dictate the various strategies of gene therapy. The use of nonintegrative, integrative, or replacement vectors for DNA delivery (3) will result in a transient, long-term, or permanent transgene expression. For example, some applications of gene therapy for cancer treatment (such as antisense therapy, immunotherapy, prodrug therapy, etc. (4)) using the therapeutic DNA to produce pharmaceutical agents may respond to the transient expression of transgene products; treatment of diseases caused by recessive mutations, on the other hand, requires a prolonged expression coupled with either a random (gene complementation) or a precise (gene targeting) integration of the transferred DNA. Finally, diseases resulting from dominant mutations can be treated effectively only via gene targeting. In fact, the precise replacement of damage is an ideal variant of gene therapy, since this strategy overcomes many obstacles and allows control of the duration, the level, and the place of transgene expression. Gene targeting is the homologous pairing of extrachromosomal and chromosomal DNA molecules. It
Ideally, gene therapy involves the correction of genetic defects through the natural means of gene targeting. This therapy possesses a number of conceptual advantages. However, a major obstacle to successful gene therapy is the relative inefficiency of the targeting process in mammalian cells. Gene targeting may be accomplished by two different mechanisms: the homologous recombination and the mismatch correction of DNA heteroduplexes. Based on the model of homologous recombination for the wellstudied prokaryotic and the less studied eukaryotic systems, three approaches have been employed to improve the efficiency and accuracy of homologous recombination events. These are: (1) artificial doublestrand breaks in both the exogenous and the chromosomal DNA, (2) a contiguous long homology between the exogenous and chromosomal DNA, and (3) a transient overproduction of an active recombinase, the bacterial RecA or mammalian RecA-like proteins, in mammalian cell nuclei. Combining these approaches can result in more effective gene targeting protocols. The second mechanism has been improved based on recent observations of recombinogenic activity of oligonucleotides and, especially, specifically designed chimeric RNA/DNA oligonucleotides. The use of RecA-like proteins to stimulate searching for homology and forming stable DNA heteroduplexes between oligonucleotides and chromosomal DNA remains an attractive idea for additional improvement of gene targeting events. © 1999 Academic Press
Key Words: chimeric RNA/DNA oligonucleotides; homologous (targeting) recombination; gene therapy; mismatch correction systems; RecA/Rad51/ Rec2-like proteins. 276 1096-7192/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
GENE TARGETING FOR GENE THERAPY
277
results from two different mechanisms of DNA repair: the homologous recombination (5) and the mismatch correction of DNA heteroduplexes (6,7). This minireview addresses both mechanisms, as well as recently discovered ways of improving gene targeting. Multiple aspects of gene targeting therapy have been previously reviewed (7–13). GENE TARGETING BY HOMOLOGOUS RECOMBINATION Homologous recombination (HR) signifies the exchange of genetic material between two homologous DNA molecules that is essential to all organisms. HR participates in maintaining genome integrity, generating genetic diversity, and segregating chromosomes properly. Our understanding of HR, both the genetics and the biochemistry of the process, is based primarily on observations made in Escherichia coli. To date, more than 20 E. coli proteins involved in recombination have been described (14) and subdivided according to their functions in the consecutive steps of this process (15). The initial phase of HR is DNA end-dependent. Figure 1 demonstrates this introductory step (step 1) with doublestrand breaks (DSBs) which are well-known inducers of HR in higher eukaryotes (17,18). Each end of DSBs formed either in donor or (as can be seen in the figure) in recipient DNA is resected by a 59exonuclease (probably acting together with a helicase) to produce long 39-terminal single-stranded DNA (ssDNA) overhangs (step 2) which are covered by a RecA-like protein to form the recombinationally active nucleoprotein filament. The latter triggers a search for homology with partner double-stranded DNA (dsDNA) and switching of pairing between donor and recipient helices. This results in strand exchange and heteroduplex DNA formation (step 3) followed by DNA gap repair and formation of recombination intermediates known as Holliday structures (step 4). These structures can be extended and/or migrated by branch migration proteins (step 5) and finally cleaved by an X-shaped DNA recognizing protein, a DNA resolvase. Two possible modes of DNA cleavage with resolvase produce two types of recombinant DNA products (step 6). Note that both products contain small regions of heteroduplex DNA which, as shown in Fig. 1, can be the regions of imperfect homology and, therefore, become a substrate of mismatch correction systems to be repaired to donor or recipient nucleotide sequence (19). Thus,
FIG. 1. Homologous recombination coupled to the mismatch correction of DNA heteroduplexes. The scheme is based on the double-strand break repair model (16). The donor DNA fragment and recipient DNA are shown by thick and thin lines, respectively. Open and filled circles indicate, respectively, donor and recipient markers located in DNA heteroduplex regions. Numbers from 1 to 7 show consecutive steps of the process (for details, see text). Arrows and double arrows indicate two ways for resolution of recombinant structures into two configurations.
the correction systems play a positive or negative role in recombinant product fixation (step 7). The counterparts (structural homologs or functional analogs) of E. coli proteins participating in steps 1 to 4 have been revealed in Saccharomyces cerevisiae (15). These give evidence of conceptual similarity of the HR process in eubacteria and eukaryotes. Two differences have, however, become clear: first, the number of homologs and complexity of analogs appear to be much higher in S. cerevisiae; second, there is complex genetic control of the initial step in S. cerevisiae (17,18). In mammalian cells the situation appears more complicated still. Unlike
278
VLADISLAV A. LANZOV
yeast cells, which commonly integrate DNA in chromosomal sites with targeted events outnumbering random events by more than 10 to 1, this ratio is on average 1 to 1000 in mammalian cells (reviewed in Ref. 20). This discrepancy results mainly from the end-joining mechanism of mammalian cells that sticks DNA ends together, arising primarily from errors in DNA metabolism, regardless of the terminal sequences (20). This mechanism is significantly less common in yeast cells and virtually unknown in bacteria. Although HR is a multistep process, for the purposes of gene targeting improvement only several initial steps seem to be crucial. These include, first, directed formation of DSB both in the donor and in the targeted chromosomal site; second, the length and homology perfection of donor DNA; and third, the amount and recombinogenic ability of RecA-like proteins promoting the pairing and strand exchange. DSBs DSBs created by rare-cutting endonucleases have been shown to stimulate mitotic homologous recombination in mammalian cells (21,22). In principle, a similar approach had been developed earlier for two site-specific recombination systems, the P1 bacteriophage Cre-Lox and the yeast FLP systems. The Cre protein catalyzes integration of P1 phage into chromosomal lox sites. In mammalian cells, a pure Cre protein introduced by lipofection causes both the excision of DNA sequences flanked by lox sites and targeted insertion of the lox-containing vector at a chromosomal lox site (23). The yeast recombinase Flp stimulates intrachromosomal recombination between FLP target sites integrated in the human b-globin locus in MEL cells (24). The most impressive effect of DSBs induced in vivo with the S. cerevisiae mitochondrial endonuclease I-SceI has been recently shown in male mouse embryonic stem (ES) cells (22). This work involved targeting recombination between an extrachromosomal segment of the hprt gene and the entire chromosomal hprt gene. This gene is a classical model for the analysis of gene targeting events (25) because it is on the X chromosome and thus exists as a single copy in XY ES cells. In addition, negative–positive selections facilitate observation of both the loss and the restoration of the hprt function. The frequency of gene targeting in clones with an I-SceI site in the presence of the endonuclease appears 5000 times
higher than that in clones without this site. Moreover, this frequency decreases with distance from the DSB site. Together with the observation that introduction of DSBs into targeting vectors significantly increases the accuracy of gene targeting (26,27), these data support the idea of artificial DSB production as potential recombinogenic sites for directed gene targeting events. Length and Perfection of DNA Homology In wild-type E. coli, HR requires a minimum length of homology of about 23 to 27 bp, and single mismatches in 31-bp substrates cause reductions in recombinant frequencies from 2- to 12-fold (28). In addition, the efficiency of HR has been found to be proportional to the length of homology between the donor and acceptor DNA in the range between 175 and 1200 bp (29). Thus, in bacteria, the efficiency of HR depends on both the extent and the perfection of homology between recombining partners. Similar conclusions have been reached regarding gene targeting in mammalian cells. The targeted inactivation of the endogenous hprt gene in male mouse ES cells shows that the targeting frequency is strongly dependent on the length of homology between the targeting vector and target locus (30). In the range between 2 and 10 kb, this dependence is exponential, yielding a 100-fold increase in gene targeting frequency and reaching saturation with respect to length of homology at about 14 kb. In addition, targeting efficiency appears dependent on the perfection of homology between donor and recipient DNA; vectors prepared from isogenic DNA appear to be 4 to 5 times more efficient than corresponding vectors carrying nonisogenic DNA (30). The obvious explanation for this last finding is the interference between HR and the mismatch repair (MMR) system of the cell. The recombination between nonisogenic DNAs creates many mismatch bases in heteroduplex regions. Because the correction of recombination heteroduplexes by the MMR system is not a strand-directed process, simultaneous exonuclease degradation proceeding on both DNA strands must result in DNA destruction and cell death (31). RecA-like Proteins Figure 1 demonstrates the DNA pairing and strand transfer forming the central step of HR. In bacteria, this step is controlled by RecA and several helper proteins. The RecA protein binds ATP, and then, through a ssDNA binding (SSB) protein which
GENE TARGETING FOR GENE THERAPY
removes ssDNA secondary structures, polymerizes on ssDNA to form continuous spiral presynaptic filament(s) (14). The complex of RecO and RecR proteins facilitates this polymerization while the complex of RecF and RecR proteins halts the filament extension (32). These filaments attach to dsDNA partner molecules, form paranemic three-stranded DNA structures, switch pairing within the structures, and initiate processing of DNA heteroduplexes which result in Holliday structures. Figure 1 is an in vivo picture based mainly on an analysis of the in vitro strand transfer reaction visualizing the transfer of one strand of bacteriophage M13 (or fX174) linear dsDNA to a closed circular ssDNA of the same phage. The RecA protein is universal in nature (15,33). Rad51 proteins of lower and higher eukaryotes including human HsRad51 were shown to be the structural and functional homologs of RecA (34). Interestingly, an SSB protein analog in eukaryotes, the RPA complex, can be effectively substituted by bacterial SSB in some strand exchange reactions (35). These findings show the universal nature of enzymes performing the central step of HR and address the question of whether the introduction of taxonomically distinct RecA-like proteins into mammalian cells can improve gene targeting. The following three observations suggest the feasibility of such an approach: first, a correlation between the level of HR and the extent of HsRAD51 gene transcription in immortal human cells (36); second, the 20-fold stimulation of spontaneous intrachromosomal HR in CHO cells by a 2- to 3-fold overproduction of the CgRad51 protein from Chinese hamster Cricetulus griseus (37); and third, the 10-fold stimulation of HR in transgenic tobacco cells producing a nucleus-targeted E. coli RecA (38). Last, recent research using the system for production of a functional E. coli RecA protein (modified with a nuclear location signal) in nuclei of mammalian cells (39) has shown a 10-fold stimulation of HR within the hprt gene of male mouse F9 cells (40). Among different bacterial RecA proteins studied to date, the Pseudomonas aeruginosa RecA manifests a hyperrecombination activity, initiating recombination 6 to 7 times more frequently than the E. coli RecA (41). Moreover, some chimeric RecA proteins, structural hybrids between P. aeruginosa and E. coli RecAs, appear still more aggressive in initiating HR (IV Bakhlanova, T Ogawa, VA Lanzov, manuscript in preparation). Experimentation is now underway to test whether these unusual proteins
279
maintain their hyperrecombinogenic potential in mammalian cell nuclei. Recently, a new human HsREC2 gene has been isolated. It is a structure–function homolog of the REC2 gene from the corn smut, Ustilago maydis (42). Unlike the latter, HsREC2 is structurally similar to human HsRAD51 and is expressed in a wide range of tissues. In addition, its production is induced by ionizing radiation. Because this protein obviously takes part in the mechanism of DSB repair, its overproduction can improve the cell enzymatic machinery promoting gene targeting events (43). Other Factors A new protocol of ex vivo gene targeting in mouse ES cells has been proposed (44). It allows an increase in the absolute targeting frequency at the hprt gene from 10 25 to 10 21. This impressive result has been achieved at an unusually high plating density of electroporated cells followed by the delay of colony selection for 60 h. If this method can be reproduced in other experimental systems, it opens new possibilities for gene targeting without the use of positive–negative selection. GENE TARGETING BY MISMATCH CORRECTION Mismatches arising as errors in DNA metabolism or among intermediates (heteroduplexes) of HR are repaired by several mechanisms found ubiquitously (19). In bacteria, the most substantial is a methylation-directed long patch MMR operated by the MutHLS enzymatic complex (45). In fact, all 8 base– base mismatches (except possibly «C–C») can be corrected by this system. The system has been reconstructed in vitro and its sequential steps have been well documented. Normally, DNA in E. coli is methylated at GATC sites by a special Dam methylase. After DNA replication, the daughter strand appears transiently unmethylated and thus becomes a substrate for the repair reaction. The latter is initiated by binding of MutS protein to a mismatch followed by binding of MutL to stabilize the MutSL complex and activate MutH endonuclease which nicks the unmethylated strand of hemimethylated DNA at a GATC site located up to 1–2 kb from the mismatch. Consecutive steps of DNA excision and resynthesis result in producing a long oligomer patch in the corrected DNA strand. It is noteworthy that such a
280
VLADISLAV A. LANZOV
gram-positive bacterium as Streptococcus pneumoniae has the Hex system that is similar to MutHLS but is directed by a nick in the daughter DNA and thus has no necessity in a MutH-like endonuclease. Structure–function homologs of the MutHLS systems have been found both in yeast and in human cells (6,46). The human MMR system can repair not only 8 mismatches, among which «C–C» is the weakest substrate, but also small insertions and deletions, from 8 to 12 nucleotides in length. Like the bacterial Hex, the mammalian MMR has no MutH analog and probably performs a nick-directed correction (6). As expected, eukaryotic systems appear much more complex. The human MMR contains two heterodimer complexes, MutSa (MSH2 1 MSH6) and MutSb (MSH2 1 MSH3), of the MutS protein family and heterodimers MutLa (MLH1 1 PMS2) and MutLb (MLH1 1 PMS1) of the MutL family. Combinations of these MutS and MutL complexes possess a definite substrate specificity (6). Nucleotide excision repair (NER) is another universal repair mechanism responsible for removing damaged bases (adducts, fusion, etc.) (47). NER produces two coordinated incisions around the damaged base to excise a small oligomer of approximately 28 –29 nucleotides and then correctly resynthesizes the gap formed. As has been recently proved (48), the short patch NER system of Schizosaccharomyces pombe can correct C–C mismatches and, to a lesser extent, other base– base mismatches, efficiently and independently of MMR. In light of findings of cooperative interactions between some proteins of NER and MMR systems in mammalian cells in order to remove large loops (reviewed in Ref. (46)), it becomes clear that mammalian cells have an efficient protection from various types of mismatches in the heteroduplex DNA formed during either DNA replication or recombination. Gene targeting via mismatch correction implies the introduction of a designed short DNA fragment or oligonucleotide into mammalian cells, pairing of the fragment with recipient DNA to form a heteroduplex region, and correction of the recipient sequence to a designed one by one of the mismatch correction systems. RecA-like Proteins The human genome contains at least a 10 9-fold excess of nonhomologous to homologous sites, so that a search for the homologous site must proceed
in a rapid manner. Evidence exists (49) that the rate of synaptic complex formation between short oligonucleotides (24 –27 nucleotides) and linear (57 bp) or even supercoiled (2.7 kb) DNA in the presence of a sufficient amount of the E. coli RecA protein exibits second-order kinetics. In other words, this rate depends on the concentration of both oligonucleotides and targeted duplex DNA. Notably, the extent of complexity of heteroduplex DNA has no effect on the reaction rate discussed. Thus, if targeted cells contain an appropriate amount of the RecA-like protein and absorb a sufficient number of extrachromosomal DNA molecules, the search for homology is not a rate-limiting step in gene targeting. Another advantage of RecA protein-assisted pairing that is useful for gene targeting is that RecA has the ability to decrease the fidelity of homology recognition in synaptic complexes between an oligonucleotide and a supercoiled plasmid (50). In principle, such a property should facilitate heteroduplex DNA formation. A Small Fragment Homology Replacement (SFHR) Gene targeting by SFHR has recently been tested in the correction of the mutant cystic fibrosis transmembrane conductance regulator (CFTR) gene to its wild-type sequence (51). The 491-bp fragments of genomic CFTR DNA encapsulated in liposome or polyamidoamine vehicles have been introduced into transformed epithelial cells containing the CFTR gene with the most commonly distributed mutation, DF508. About 1 in each 100 cells restored the defect in Cl 2 ion transport giving evidence of successful gene targeting. Unfortunately, a preliminary coating of DNA fragments by E. coli RecA protein does not enhance the homologous replacement. The literature suggests that this may occur because the conditions for RecAassistant reaction have not been optimized (51). Chimeric RNA/DNA Oligonucleotides A new strategy for gene targeting by oligonucleotides has recently been developed (7). This new strategy is based on the discovery of high recombinogenic activity of the RNA stretch within RNA/ DNA small chimeric molecules designed as a continuous RNA/DNA complementary duplex constructed in a double-hairpin configuration (52). Relative to analogous DNA/DNA molecules, RNA/DNA constructs bind the targeted DNA sequence packaged
GENE TARGETING FOR GENE THERAPY
into nucleosomal structures 20-fold more efficiently (7,43). This strategy has been applied to correct (with relatively high frequency) a single base alteration in the alkaline phosphatase gene of HuH-7 human hepatoma cells (53); to induce in vivo a point mutation in the rat factor IX gene that resulted in a significant reduction of this factor coagulant activity (54); to correct the mutation in the hemoglobin b s allele for sickle cell anemia in lymphoblastoid cells homozygous for this mutation (55); and to induce a sickle cell anemia mutation in normal lymphoblastoid cells (43). FINAL REMARKS The past decade has provided remarkable progress in the development of new gene targeting strategies. However, our current knowledge of homologous recombination mechanisms in higher eukaryotes is only the first sketch for a much more complicated picture of recombinational events coupled with different modes of the DNA heteroduplex correction. In fact, homologous recombination in prokaryotes is not limited by RecA-dependent pathways. Both lambda-type bacteriophages and some E. coli bacteria possess, respectively, Redab and RecET protein systems combining activities of Reda or RecE exonucleases with Redb or RecT ATP-independent annealase to promote strand exchange (56). Significantly, both Redab and RecET systems also promote DNA double-strand break repair (57) and thus serve as prototypes for analogous systems in higher eukaryotes (reviewed in Ref. (56)). Finding such systems in human cells can open new perspectives for improvement of therapeutic gene targeting. ACKNOWLEDGMENTS I am grateful to Dr. Eugene Schwartz (PNPI) for critical reading of the manuscript, Yuri Kil (PNPI) for help with figure preparation, and Natasha Blinkova for English correction. I apologize to those researchers whose contributions were not cited because of space limitation in this minireview. This work was supported by an International Research Scholar’s award from the Howard Hughes Medical Institute, Grant 75195-546101.
281
3.
Kay MA, Liu D, Hoogerbrugge: Gene Therapy. Proc Natl Acad Sci USA 94:12744 –12746, 1997.
4.
Blaese RM: Gene therapy for cancer. Sci Am June:91–95, 1997.
5.
Capecchi, MR: Altering the genome by homologous recombination. Science 244:1288 –1292, 1989.
6.
Modrich P: Strand-specific mismatch repair in mammalian cells. J Biol Chem 272:24727–24730, 1997.
7.
Kmiec EB: Genomic targeting and genetic conversion in cancer therapy. Semin Oncol 23:188 –193, 1996.
8.
Waldman AS: Targeted homologous recombination in mammalian cells. Crit Rev Oncol/Hematol 12:49 – 64, 1992.
9.
Doi S, Campbell C, Kucherlapati R: Directed modifications of genes by homologous recombination in mammalian cells. In Transgenic Animals (Grosveld F, Kollias G, Eds.) London: Academic Press, pp 27– 46, 1992.
10.
Lanzov VA: Ideal gene therapy: approaches and horizons. Mol Biol 28:321–327, 1994.
11.
Kowalczykowski SC, Zarling DA: Homologous recombination proteins and their potential applications in gene targeting technology. In Gene Targeting (Vega MA, Ed.). Boca Raton: CRC Press, pp 167–210, 1995.
12.
Vega MA: Gene targeting in human gene therapy. In Gene Targeting (Vega MA, Ed.). Boca Raton: CRC Press, pp 211– 229, 1995.
13.
Yanez RJ, Porter ACG: Therapeutic gene targeting. Gene Ther 5:149 –159, 1998.
14.
Kowalczykowski SC, Dixon DA, Egglestone AK, Lauder SD, Rehrauer WM: Biochemistry of homologous recombination in Escherichia coli. Microbiol Rev 58:401– 465, 1994.
15.
Bianco PR, Trasy RB, Kowalczykowski SC: DNA strand exchange proteins: a biochemical and physical comparison. Frontiers Bio Sci 3:570 – 603, 1998.
16.
Szostak JW, Orr-Weaver TL, Rothstein RJ, Stahl FW: The double-strand break repair model for recombination. Cell 33:25–35, 1983.
17.
Shinohara A, Ogawa T: Homologous recombination and the roles of double-strand breaks. Trends Biochem Sci 20:387– 391, 1995.
18.
Haber JE: A super new twist on the initiation of meiotic recombination. Cell 89:163–166, 1997.
19.
Freidberg EC, Walker GC, Siege W: DNA Repair and Mutagenesis. Washington, DC: ASM Press, pp 367– 405, 1995.
20.
Roth D, Wilson J: Illegitimate recombination in mammalian cells. In Genetic Recombination (Kucherlapati R, Smith GR, Eds.). Washington, DC: ASM Press, pp 621– 653, 1988.
21.
Jasin M: Genetic manipulation of genomes with rare-cutting endonucleases. Trends Genet 12:224 –228, 1996.
22.
Donoho G, Jasin M, Berg P: Analysis of gene targeting and intrachromosomal homologous recombination stimulated by genomic double-strand breaks in mouse embryonic stem cells. Mol Cell Biol 18:4070 – 4078, 1998.
REFERENCES 1.
Anderson WF: Prospects for human gene therapy. Science 226:401– 409, 1984.
23.
Boubonis W, Saucer B: Genomic targeting with purified Cre recombinase. Nucleic Acids Res 21:2025–2029, 1993.
2.
Verma IM, Somia N: Gene therapy: promises, problems and prospects. Nature 389:239 –242, 1997.
24.
Fiering S, Kim CG, Epner EM, Groudin M. An “in– out” strategy using gene targeting and FLP recombinase for the functional dissection of complex DNA regulatory elements:
282
25.
26.
27.
28.
29.
30.
31.
32. 33.
34.
35.
36.
37.
38.
39.
40.
41.
VLADISLAV A. LANZOV analysis of the b-globine locus control region. Proc Natl Acad Sci USA 90:8469 – 8473, 1993. Thomas KR, Capecchi MR: Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51:503–512, 1987. Subramani S, Seaton BL: Homologous recombination in mitotically dividing mammalian cells. In Genetic Recombination (Kucherlapati R, Smith GR, Eds.). Washington, DC: ASM Press, pp 549 –573, 1988. Valancius V, Smithies O: Double-strand gap repair model in a mammalian gene targeting reaction. Mol Cell Biol 11: 4389 – 4397, 1991. Shen P, Huang HV: Homologous recombination in Escherichia coli: dependence on substrate length and homology. Genetics 112:441– 457, 1996. Ogawa T, Shinohara A, Ogawa H, Tomizawa J: Functional structures of the RecA protein found by chimera analysis. J Mol Biol 226:651– 660, 1992. Thomas KR, Deng C, Capecchi MR: High-fidelity gene targeting in embryonic stem cells by using sequence replacement vectors. Mol Cell Biol 12:2919 –2923, 1992. Rayssiguier C, Thaler DS, Radman M: The barrier to recombination between E. coli and S. typhimurium is disrupted in mismatch-repair mutants. Nature 342:396 – 401, 1989. Cox MM: A broadening view of recombinational DNA repair in bacteria. Genes Cells 3:65–78, 1998. Roca AI, Cox MM: RecA protein: structure, function, and role in recombinational DNA repair. Prog Nucleic Acid Res Mol Biol 56:129 –223, 1997. Brendel V, Brocchieri L, Sandler SJ, Clark AJ, Karlin S: Evolutionary comparisons of RecA-like proteins across all major kingdoms of living organisms. Mol Evol 44:528 –541, 1997. Sugiyama T, Zaitseva EM, Kowalczykowski SC: A singlestranded DNA-binding protein is needed for efficient presynaptic complex formation by the Saccharomyces cerevisiae Rad51 protein. J Biol Chem 272:7940 –7945, 1997. Xia SJ, Shammas MA, Reis RJS: Elevated recombination in immortal human cells is mediated by HsRAD51 recombinase. Mol Cell Biol 17:7151–7158, 1997. Vispe S, Cazaux C, Lesca C, Defais M: Overexpression of Rad51 protein stimulates homologous recombination and increases resistance of mammalian cells to ionizing radiation. Nucleic Acids Res 26:2859 –2864, 1998. Reiss B, Klemm M, Kosak H, Schell J: RecA protein stimulates homologous recombination in plants. Proc Natl Acad Sci USA 93:3094 –3098, 1996. Kido M, Yoneda Y, Nakanishi N, Uchida T, Okada Y: Escherichia coli RecA protein modified with a nuclear location signal binds to chromosomes in living mammalian cells. Exp Cell Res 198:107–114, 1992. Scherbakova OG, Ogawa H, Lanzov VA, Filatov MV: Targeting recombination in somatic mammalian cells: stimulating effect of bacterial RecA protein. Doklady Rus Acad Sci, in press, 1999. Namsaraev EA, Baitin DM, Bakhlanova IV, Alexseyev AA,
Ogawa H, Lanzov VA: Biochemical basis of hyper-recombinogenic activity of Pseudomonas aeruginosa RecA protein in Escherichia coli cells. Mol Microbiol 27:727–738, 1998. 42.
Rice MC, Smith ST, Bullrich F, Havre P, Kmiec EB: Isolation of human and mouse genes based on homology to REC2, a recombinational repair gene from the fungus Ustilago maydis. Proc Natl Acad Sci USA 94:7417–7422, 1997.
43.
Cole-Strauss A, Noe A, Kmiec EB: Recombinational repair of genetic mutations. Antisense Nucleic Acid Drug Dev 7:2111– 2116, 1997.
44.
Templeton NS, Roberts DD, Safer B: Efficient gene targeting in mouse embrionic stem cells. Gene Ther 4:700 –709, 1997.
45.
Kolodner RD: Mismatch repair: mechanisms and relationship to cancer susceptibility. Trends Biochem Sci 20:397– 401, 1995.
46.
Sancar A: Excision repair invades the territory of mismatch repair. Nature Genet 21:247–249, 1999.
47.
Sancar A: Excision repair in mammalian cells. J Biol Chem 27:15915–15918, 1995.
48.
Fleck O, Lehmann E, Schar P, Kohli J: Involvement of nucleotide-excision repair in msh2 pms1-independent mismatch repair. Nature Genet 21:314 –317, 1999.
49.
Yancey-Wrona JE, Camerini-Otero RD: The search for DNA homology does not limit stable homologous pairing promoted by RecA protein. Curr Biol 5:1149 –1158, 1995.
50.
Malkov VA, Sastry L, Camerini-Otero R: RecA protein assisted selection reveals a low fidelity of recognition of homology in a duplex DNA by an oligonucleotide. J Mol Biol 271:168 –177, 1997.
51.
Kunzelmann K, Legendre J-Y, Knoell DL, Escobar LC, Xu Z, Gruenert DC: Gene targeting of CFTR DNA in CF epithelial cells. Gene Ther 3:859 – 867, 1996.
52.
Kotani H, Germann H, Andrus A, Vinayak R, Mullah B, Kmiec EB: RNA facilitates RecA-mediated DNA pairing and strand transfer between molecules bearing limited regions of homology. Mol Gen Genet 250:626 – 634, 1996.
53.
Kren BT, Cole-Strauss, Kmiec EB, Steer CJ: Targeted nucleotide exchange in the alkaline phosphatase gene of HuH-7 cells mediated by a chimeric RNA/DNA oligonucleotide. Hepatology 25:1462–1468, 1997.
54.
Kren BT, Bandyopadhyay P, Steer CJ: In vivo site-directed mutagenesis of the factor IX gene by chimeric RNA/DNA oligonucleotides. Nature Med 4:285–290, 1998.
55.
Cole-Strauss A, Yoon K, Xiang Y, Byrne BC, Rice MC, Gryn J, Holloman WK, Kmiec EB: Correction of the mutation responsible for sickle cell anemia by an RNA–DNA oligonucleotide. Science 273:1386 –1389, 1996.
56.
Hall SD, Kolodner RD: Homologous pairing and strand exchange promoted by the Escherichia coli. Proc Natl Acad Sci USA 91:3205–3209, 1994.
57.
Yokochi T, Kusano K, Kobayashi I: Evidence for conservative (two progeny) DNA double-strand break repair. Genetics 139:5–17, 1995.