We describe a technology developed for the site-specific correction of a single base carried on an episome or chromosome in prokaryotic and eukaryotic cells. Critical to the development of this technology as a therapeutic device for treating genetic disorders, like [[alpha].sub.1]-antitrypsin deficiency, is the establishment of a standardized assay to study its mode of action and structure-activity relationships (SARs). To this end, a positive-selection system in Escherichia coli has been developed to assess RNA/DNA oligonucleotide (RDO)-directed repair activity. We demonstrate that RDO-directed repair requires the concerted action of the two following repair proteins: the pairing protein RecA; and the mismatch recognition protein, MutS. SAR studies demonstrate that the RDO molecule is functionally asymmetric. The RNA-containing strand enables strand-pairing and stabilization of the molecule, and the DNA-containing strand confers the information transfer.
Abbreviations: A1AD = [[alpha].sub.1]-antitrypsin deficiency; [[alpha].sub.1]AT = [[alpha].sub.1]1-antitrypsin; IPTG = isopropyl-[beta]-D-thiogalactopyranoside; kan = kanamycin; neo = neomycin; PI = protease inhibitor; RDO = RNA/ DNA oligonucleotide; SAR = structure-activity relationship
[[alpha].sub.1]-Antitrypsin ([[alpha].sub.1]AT) is a major plasma serum protease inhibitor (PI) that is expressed predominantly in hepatocytes that plays a critical role in controlling tissue damage following inflammation. [[alpha].sub.1]AT deficiency (A1AD) is a rare autosomal-recessive disorder that is characterized by low serum levels of [[alpha].sub.1]AT and, subsequently, by an imbalance of PI in the lung. Decreased PI levels in the lung cause the destruction of its alveolar wall and the early onset pulmonary emphysema. (1) There are > 60 rare allele variants of [alpha]1AT, of which the most common for A1AD is the PI*Z allele. PI ZZ homozygotes have 15 to 20% of the normal plasma levels of [[alpha].sub.1]AT. The Z allele contains a single nucleotide base change resulting in an amino acid substitution at position 342 (Glu 342-Lys). (2) The PiZ variant protein is poorly secreted from the liver cell, leading to [[alpha].sub.1]AT protein self-aggregation and to the subsequent formation of inclusion bodies. An additional consequence of PIZ patients, therefore, is cirrhosis of the liver, affecting young children and 12 to 15% of adult patients. It is believed that chronic liver disease associated with A1AD is a consequence of the toxicity associated with the build up of aggregated [[alpha].sub.1]AT protein in PiZ patients. (3)
The clinical consequences of A1AD vary from severe to mild. As is the ease with non-A1AD emphysema, there is no curative medical treatment for A1AD emphysema. For A1AD patients who have received diagnoses early and have not yet developed chronic lung disease, the primary recommendation is to keep the airways open and free from inflammation, which is in accordance with similar therapeutic strategies for asthma. IV augmentation therapy with purified human [[alpha].sub.1]AT in patients with A1AD may offer some protection against a rapid loss of lung function. (4) Gene-addition therapy, whereby a normal [[alpha].sub.1]AT gene is delivered to the muscle by a recombinant adeno-associated viral vector, may raise the levels of serum P1 protein. (5) Although IV protein and gene augmentation therapies may offer some lung-protective benefit to A1AD patients by transiently raising serum levels of [[alpha].sub.1]AT, these therapies are expensive, require continuous dosing, and show little promise for treating the chronic liver disease often associated with PiZ A1AD. Liver transplantation is an option for patients severely affected with A1AD but cannot be offered as a general therapy. The most desirable approach for treating A1AD patients, as well as patients with other genetic diseases, is gene repair, that is, the site-specific correction of mutant genes to restore function. Gene repair allows the expression of the corrected genes to remain under their endogenous physiologic controls. (6) Correcting the PIZ allele would raise [[alpha].sub.1]AT serum levels and would decrease the hepatotoxic effect of the [[alpha].sub.1]ATZ retained in the endoplasmic reticulum of liver cells. Homologous recombination is a logical way to achieve such an objective and has been successfully used to generate a wide variety of mouse models with specific mutated genes. However, it is a relatively inefficient process in differentiated mammalian cells and has not been practical to date for gene therapy applications. (7)
In the past few years, a novel approach using chimeric RNA/DNA oligonucleotides (RDOs) to introduce site-specific, single-DNA base-pair alterations has been developed and successfully used to modify both extrachromosomal DNA (8) and chromosomal DNA (9-16) in a variety of cell types. The technology, termed Genoplasty (Valigen; Lawrenceville, NJ), is based on the observation that RDOs containing complementary RNA/DNA hybrid regions are more active than duplex DNA in homologous pairing reactions in vitro. The chimeric molecules are designed with a short homologous targeting sequence composed of deoxynucleotides flanked by blocks of 2'-O-methylated RNA residues, a complementary all-DNA strand, thymidine hairpin caps, a single-strand break, and a double-stranded clamp region (Fig 1). The homologous regions of the RDO are entirely complementary with a genomic target sequence, except for the designated single base-pair mismatch, which is thought to be recognized and corrected by the cell's endogenous DNA repair systems. (17) The relative efficiency of RDO-directed gene targeting, which is several-fold more efficient than traditional homologous recombination, and the ease of obtaining the targeting oligonucleotide suggest numerous potential applications for this technology. By introducing site-specific nucleotide changes, Genoplasty can be used to characterize novel genes, can modify cell lines, can create novel animal and plant models, and can be used to develop therapies for monogenic diseases.
[FIGURE 1 OMITTED]
Still a development-stage technology, the Genoplasty literature reports inconsistent conversion frequencies among different investigators and cell types. (18) Alexeev and Yoon (19) have reported conversion frequencies at the tyrosinase locus in an albino melanocyte cell line that varied from 0.01 to 15%, using the same RDO and transfection conditions. The reasons for this variability are currently under investigation in this and other laboratories. Possible parameters affecting conversion activity include the sequence, structure, quality, and delivery of the RDO, cell growth conditions, and detection methods. Standardizing the RDO technology, therefore, is of paramount importance to the development of therapeutic applications.
In this report, we describe an experimental system designed to measure gene conversion in a rapid and reproducible manner. We used this system to investigate RDO structure activity relationships, the mechanism of RDO-directed repair, and its potential therapeutic application to patients with A1AD.
MATERIALS AND METHODS
Plasmid Constructs
Mutants carrying a nucleotide substitution or deletion were generated by polymerase chain reaction mutagenesis. (20) pKanS contained a single T [right arrow] G point mutation at position 4021 in pWE15 (Stratagene; La Jolla, CA). pAC[DELTA]TrecA and pAC[DELTA]TmutS, which were derived from pACYC184 (New England BioLabs; Beverly, MA), lack the tetracycline region between AvaI and XbaI and contain the Lac operator of pUC19 (21) and the coding regions of wild-type RecA or MutS (which were cloned from MC1061 genomic DNA) in frame with the first 12 amino acids of the LacZ gene, respectively. RecA and MutS expression were induced with 1 mM isopropyl-[beta]-D-thiogalactopyranoside (IPTG).
Bacterial Strains, Media, and Growth Conditions
Cells were grown in LB media. (22) Where appropriate, cells were grown in the presence of kanamycin (kan; 50 [micro]g/mL), ampicillin (50 [micro]g/mL), IPTG (1 mM), or chloramphenicol (20 [micro]g/mL). MC1061mutS was generated by generalized transduction using a P1 phage produced from BMH71-18 cells (mutS). (23) All rec[A.sup.+] and rec[A.sup.-] bacteria were tested for RecA activity by determining their resistance to ultraviolet irradiation, as previously described, (22) in some cases following induction by 1 mM IPTG. Electrocompetent Escherichia coli containing pKanS was transfected with 0.01 to 1 [micro]g RDO using an apparatus (GenePulser; Bio-Rad; Hercules, CA) under standard electroporation conditions (ie, 2.5 kV, 25 [micro]F (microFarad), and 200 ohms). Immediately following, cells were incubated for 1 h in the presence of 1 mL SOC media at 37 [degrees] C with moderate shaking. Bacteria then were plated on LB agar containing ampicillin or kan and were incubated overnight at 37 [degrees] C.
Oligonucleotide Synthesis
Chimeric RDOs were synthesized as previously described, (7) using DNA and 2'O-methyl RNA phosphoramidite monomers on an oligonucleotide synthesizer, and were purified by liquid chromatography. Purity was determined to be approximately 90% using ion-exchange high-performance liquid chromatography analysis.
Analysis of Plasmid DNA
Plasmid DNA was isolated by anion-exchange chromatography (Qiagen; Chatsworth, CA). For pKan, the plasmid DNAs were digested with BfaI (New England BioLabs). DH5[alpha] bacteria were transformed with plasmid DNA from the primary converted colonies to obtain a pure population of plasmid for DNA sequence analysis using an ABI 373 (Applied Biosystems, Foster City, CA) automated sequencer.
RESULTS
Design of the Detection Systems
We employed a positive selection strategy to detect chimeric oligonucleotide-mediated DNA repair in bacteria (Fig 1). The chimeric oligonucleotides were designed as described previously. (8) The gene-specific region is mismatched with the target sequence at a single position where a nucleotide substitution or insertion is to be made. The test system was created by introducing a site-directed mutation in the neomycin (neo) resistance gene using conventional in vitro methods. The mutant neo gene contains a T-to-G change at the third position in codon 22, resulting in the substitution of tyrosine (TAT) with a termination codon (TAG) and the concomitant creation of a diagnostic Bfa1 restriction enzyme recognition site. A number of RDOs (Fig 1) of varying lengths and structures were designed to cause a G-to-C substitution, replacing the stop codon with that of an alternative tyrosine (TAC). This permits the converted sequence (TAC) to be distinguished from the original (TAT) sequence and makes it possible to control for contamination. The following two control oligonucleotides were used: an all-DNA oligonucleotide, Kan1DNA, having the same sequence and structure as that of Kan1; and a nonspecific RDO (SC1, targeting human [beta]-globin (9)).
The frequency of correction was calculated as the number of kan-resistant colonies relative to the total number of ampicillin-resistant bacteria. The RDOs with 25-nucleotide and 35-nucleotide homology regions were found to be approximately 40-fold and approximately 500-fold, respectively, more active than that containing 15 nucleotides of homology (Fig 1). The frequency of correction for either control oligonucleotide, SC1 or Kan1DNA, was no more active than the background. Importantly, the failure of the oligonucleotide lacking modified RNA (ie, kan1DNA) to effect correction suggests an essential role for RNA in the oligonucleotide. DNA sequencing confirmed that the RDO-directed correction resulted in sequence changes specified by the oligonucleotides (TAG to TAC for Kan1). In contrast, the rare, spontaneously occurring, kan-resistant colonies contained either a TGG or a TAT genotype, and never the TAC genotype (data not shown). These data demonstrate that the chimeric oligonucleotides are capable of directing a site-specific nucleotide substitution and that, by increasing the region of homology around the intended mismatch, the efficiency of repair is increased.
Since the original RDO molecule was empirically designed, its structure-activity relationships (SARs) are poorly understood. In order to assess the functional significance of the intervening DNA sequence of the RNA strand, we replaced the DNA with modified RNA (Kan/ RNA), creating one all-RNA strand and one all-DNA strand. The Kan/RNA oligonucleotide had an approximately twofold greater conversion frequency than the RDO Kan1. The increased RNA content, which increases the thermal melting point and base pairing (data not shown), may stabilize the mismatched oligonucleotide/ target DNA complex in vivo, leading to an improved repair reaction.
In order to determine whether one or both of the RDO strands are responsible for site-specific nucleotide alteration, we synthesized two asymmetric RDOs targeted to the mutant neo genes, KanCC and KanGG. KanCC contained the correcting mismatch on the 5'-proximal all-DNA strand, and the chimeric strand (RNA-DNA) was perfectly complementary to the mutant target neo sequence. KanGG contained only the correcting mismatch on the 5'-distal, chimeric strand (Fig 1). The frequency of repair for KanCC was approximately fivefold higher than that observed with the Kan1. Surprisingly, the efficiency of repair of KanGG was significantly lower (approximately 10-fold) than that of the Kan1. Sequence analysis of the corrected plasmids revealed that base substitution catalyzed by Kan1 and KanCC RDOs was of high fidelity (ie, only the TAC sequence was observed from several clones picked from three different experiments). In contrast, only 3 of 20 of the KanGG-treated colonies contained the intended TAC change (data not shown). Codons, TAT, and TGG were observed the most, suggesting that the 2'O-methyl RNA-containing strand could not support accurate nucleotide repair. The combined loss in activity and fidelity that was observed when the mismatch was carried on the chimeric RNA/DNA strand of the KanGG supports the notion that the 5'-proximal DNA strand is the active strand for the nucleotide alteration. Together, these data demonstrate that, even though RNA is essential, the role of the RNA-containing strand may be to confer higher nuclease and base-pairing stability, providing a scaffold for the efficient pairing of the DNA strand, which actively participates in the nucleotide repair.
Role of RecA and MutS in RDO-Directed Correction
Since the RDO design strategy is based on homologous pairing, we hypothesized that the ubiquitous pairing protein, RecA, may be required for the process. As such, we tested the Kan1 chimera in isogenic E coli strains, MC1061 and WM1100, which contained a functional or a mutant RecA gene, respectively. As expected, RDO-directed repair was observed only in MC1061 which contained a functional RecA gene (Fig 2). Moreover, when the wild-type RecA gene was expressed exogenously in WM1100 using an inducible expression vector, pAC[DELTA]TRecA, RDO-directed correction was recovered (Fig 2). In a similar fashion, we explored the requirement of the mismatch repair enzyme, mutS, in RDO-directed repair. For this purpose, an isogenic mutS strain, MC1061mutS, was generated by generalized transduction. The RDO-directed kan gene correction was not detected above the background in cells lacking mutS expression but was observed following the induction of mutS at levels similar to those seen with MC1061. Sequence analysis demonstrated that the background-resistant clones from untreated samples from the mutagenic strain, MC1061mutS, in the absence of MutS expression contained mostly TAT and TGG codons at position 22, while the analyzed RDO-treated, kan-resistant colonies observed in the presence of MutS were all TAC (data not shown).
[FIGURE 2 OMITTED]
The data thus suggest that RDO-directed repair operates by a mechanism that is different from homologous recombination, requiring both RecA and MutS. The requirement of RecA and the MutS mismatch repair pathway further suggests the following two-step mechanism: RecA-mediated pairing followed by a MutS-mediated repair (Fig 3).
[FIGURE 3 OMITTED]
DISCUSSION
The current treatment for patients with A1AD is limited primarily to keeping the lung airways open and free from inflammation. While controversial and expensive, protein-augmentation therapy may provide some benefit for preventing a rapid loss of lung function. Although progress is being made toward additional therapies, including gene-addition therapy, the promise of a cure is far from a reality. Given the complexities of this disease, the ideal therapy would be to repair the single-nucleotide defects in the P1Z allele and in other mutant alleles. A novel DNA repair technology that makes site-directed, single-nucleotide alterations, Genoplasty may be used to convert the PIZ mutant allele to normal. The strategy for targeting the PiZ allele is shown in Figure 4.
[FIGURE 4 OMITTED]
The Genoplasty technology is based on the observation that RDOs containing complementary RNA/DNA hybrid regions are more active than duplex DNA in homologous pairing reactions in vitro. Several laboratories have shown (8-17) that genoplasty induces the DNA repair machinery to make site-specific DNA alterations. Still, many technical challenges remain. Inconsistent or variable conversion efficiencies have been described, (18) requiring a standardized system for understanding the mechanism and parameters affecting the activity of the RDO. Such a system will, we believe, streamline the development of a robust and highly active compound for gene correction (conversion).
Toward this end, we developed a standardized phenotypic conversion system, which is described in this review, to assess RDO structure-activity parameters and to further elucidate the mechanism of RDO-directed gene repair. Specifically, our E coli system has shown the following: the activity of the RDO is related to its structure and that the modified RNA component is essential for its activity; the DNA strand confers information transfer to the target sequence; an RDO with a 25-nucleotide to 35-nucleotide region of homology with the target is sufficient to introduce single-nucleotide substitutions in a plasmid at a frequency approaching 1% of the bacteria; and at least two components, RecA and MutS, of independent DNA repair pathways are required for this process. Our data support a model whereby the chimeric oligonucleotides affect gene correction in vivo by a RecA-mediated pairing event, followed by a mismatch repair-directed transfer of information. We have furthered these studies by comparing the RDO-directed repair activity observed in E coli to that of the episomal and integrated Kans targets in human cell lines, such as the human liver HUH7 cell line (data not shown). These data indicate that the SAR parameters affecting RDO activity in E coli also affect activity in human cells. This system thus allows for the rapid assessment of RDO function, which should further aid in the development of the RDO technology toward a therapeutic application.
As the RDO field comes closer to a clinical reality, several issues need to be considered before treatments can proceed. First, the number of cells in the A1AD patient's liver that need to be corrected to provide a therapeutic benefit or protection from rapid loss of lung function is still unknown. Previous studies (2) have shown that PiZ heterozygotes, with an A1AT serum concentration of 80 mg/dL, are protected against early-onset pulmonary dysfunction, suggesting that as many as 50% of the PiZ alleles would have to be corrected before a therapeutic benefit may be realized. One may speculate that since the correction of the PiZ allele may offer some growth advantage to the hepatocyte by reducing the hepatotoxic affect of the aggregated [[alpha].sub.1]AT-Z protein, the correction of fewer genes (followed by a partial hepatectomy) may preferentially enrich the liver with those cells carrying the corrected allele. Alternatively, genoplast therapy may be sequentially applied to produce therapeutically sufficient [[alpha].sub.1]AT levels.
Advancements in RDO design are necessary for most clinical applications. Optimizing RDO delivery to the liver remains a top priority as well. Significantly, Kren et al (11) have reported a highly efficient uptake and gene conversion with the RDO in the rat liver. In this study, targeted delivery was mediated by a neutral liposome possessing a ligand for a hepatocyte-specific receptor. More recently, efficient gene correction in the liver was observed following injections of an RDO complexed with polyethylenimine (ExGen 500; Fermentas; Hanover, MD) into the peritoneal cavity of a mouse. (16)
Our laboratory is developing similar strategies for liver-directed targeting as well. We have encapsulated the RDO with liposomes composed of dioleoylphosphatidylcholine, cholesterol, 1,2-dioleoyl-3-(dimethylamino)propane, and N-lactosyl-dioleoylphosphatidylethanolamine to target the asialoglycoprotein receptor of the hepatocyte. Alternatively, using a high-volume, high-pressure IV administration (Zhang et al (24)) of "naked" fluorescein-labeled RDOs, we have observed that as many as 10% of the hepatocytes contain RDOs. Currently, we are assessing liver-specific gene targeting using these systems.
The rapid phenotypic conversion models described in this review will prove critical to continued progress in RDO design and delivery, and will move Genoplasty closer to a clinical application for AiAD and other monogenic disorders.
ACKNOWLEDGMENT: The authors acknowledge John Lapolla for his suggestions and critical review of the manuscript.
REFERENCES
(1) Perlmutter DH. Alpha-1 antitrypsin deficiency. In: Schiff ER, Sorrell MF, Maddrey WC, eds. Diseases of the liver. Philadelphia, PA: Lippincott, 1999; 1131-1150
(2) Cox DW. A-1-Antitrypsin deficiency. In: Schriver CR, Beaudet AL, Sly WS, et al, eds. The metabolic and molecular bases of inherited disease. New York, NY: McGraw-Hill, 1995; 4125-4156
(3) Carrell RW. Alpha 1-Antitrypsin: molecular pathology, leukocytes, and tissue damage. J Clin Invest 1986; 78:1427-1431
(4) Wencker M, Fuhrmann B, Banik N, et al. Longitudinal follow-up of patients with [alpha]1-protease inhibitor deficiency before and during therapy with IV [[alpha]1-protease inhibitor. Chest 2001; 119:737-744
(5) Song S, Morgan M, Ellis T, et al. Sustained secretion of human alpha-1-antitrypsin from murine muscle transduced with adeno-associated cirus vectors. Proc Natl Acad Sci U S A 1998; 95:14383-14388
(6) Kren BT, Metz R, Kumar R, et al. Gene repair using chimeric RNA/DNA oligonucleotides. Semin Liver Dis 1999; 19:93-104
(7) Yanez RJ, Porter ACG. Therapeutic gene targeting. Gene Ther 1998; 5:149-159
(8) Yoon K, Cole-Strauss A, Kmiec EB. Targeted gene correction of episomal DNA in mammalian cells mediated by a chimeric RNA/DNA oligonucleotide. Proc Natl Acad Sci U S A 1996; 93:2071-2076
(9) Cole-Strauss A, Yoon K, Xiang Y, et al. Correction of the mutation responsible for sickle cell anemia by an RNA-DNA oligonucleotide. Science 1996; 273:1386-1389
(10) Kren BT, Cole-Strauss A, Kmiec EB, et al. Targeted nucleotide exchange in the alkaline phosphatase gene of HuH-7 cells mediated by a chimeric RNA/DNA oligonucleotide. Hepatology 1997; 25:1462-1468
(11) Kren BT, Bandyopadhyay P, Steer CJ. In vivo site-directed mutagenesis of the factor IX gene by chimeric RNA/DNA oligonucleotides. Nat Med 1998; 4:285-290
(12) Kren BT, Parashar B, Bandyopadhyay P, et al. Correction of the UDP-glucuronosyltransferase gene defect in the Gunn rat model of Crigler-Najjar syndrome type I with a chimeric oligonucleotide. Proc Natl Acad Sci U S A 1999; 96:10349-10354
(13) Alexeev V, Igoucheva O, Domashenko A, et al. Localized in vivo genotypic and phenotypic correction of the albino mutation in skin by RNA-DNA oligonucleotide. Nat Biotechnol 2000; 18:43-47
(14) Rando TA, Disatnik M-H, Zhou L Z-H. Rescue of dystrophin expression in mdx mouse muscle by RNA/DNA oligonucleotides. Proc Natl Acad Sci U S A 2000; 97:5363-5368
(15) Bartlett RJ, Stockinger S, Denis MM, et al. In vivo targeted repair of a point mutation in the canine dystrophin gene by a chimeric RNA/DNA oligonucleotide. Nat Biotechnol 2000; 18:615-622
(16) Tagalakis AD, Graham IR, Riddell DR, et al. Gene correction of the apolipoprotein (Apo) E2 phenotype to wild-type ApoE3 by in situ chimeraplasty. J Biol Chem 2001; 276: 13226-13230
(17) Gamper HB Jr, Cole-Strauss A, Metz R, et al. A plausible mechanism for gene correction by chimeric oligonucleotides. Biochemistry 2000; 39:5808-5816
(18) Van der Steege G, Schuilenga-Hut H, Buys CHCM, et al. Persistent failures in gene repair [letter]. Nat Biotechnol 2001; 19:305-306
(19) Alexeev V, Yoon K. Stable and inheritable changes in genotype and phenotype of albino melanocytes induced by an RNA-DNA oligonucleotide. Nat Biotechnol 1998; 16:1343-1346
(20) Saiki RK, Gelfand DH, Stoffel S, et al. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 1988; 239:487-491
(21) Yanisch-Perron C, Vieira J, Messing J. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 1985; 33:103-119
(22) Ausubel FM, Brent RB, Kingston RE, et al. In: Current protocols in molecular biology. New York, NY: John Wiley & Sons, 2001
(23) Miller JH. In: Experiments in molecular genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1972
(24) Zhang G, Budker V, Wolff JA. High levels of foreign gene expression in hepatocytes after tail vein injections of naked plasmid DNA. Hum Gene Ther 1999; 10:1735-1737
* From ValiGen, Inc, Princeton, NJ.
Correspondence to: Richard Metz, 37 Winthrop Rd, Lawrenceville, NJ 08648; e-mail: janislmetz@aol.com
COPYRIGHT 2002 American College of Chest Physicians
COPYRIGHT 2002 Gale Group