A simplified in vitro ligation approach to clone an E1B55k-deleted double-targeted conditionally-replicative adenovirus
- 3.7k Downloads
Construction of conditionally-replicative Adenovirus (CRAd) is complex and time-consuming. While homologous recombination (HR) using a two-plasmid system in bacteria is commonly used to generate CRAds, alternative methods may be required when HR fails. Previously, in vitro ligation has been suggested to facilitate construction of E1/E3-deleted, replication-incompetent Ad vectors. However, in vitro ligation has only rarely been used to generate CRAds and may be a complex procedure for molecular biologists who are not experts in the field.
Methods and Results
A modified in vitro ligation approach was developed to construct a double-targeted, E1B55k-deleted CRAd. The method allowed the incorporation of a tumor-specific promoter, e.g. the heat-shock protein 70 (hsp70) promoter, upstream of E1a, deletion of the E1B55k gene, and HR-free cloning of the recombined E1Δ55k gene into the Ad genome. The genetic structure of the CRAd was confirmed using restriction analysis and PCR. The replication rate of the hsp70E1Δ55k CRAd was 1.5–2% of Ad without E1Δ55k deletion.
A 3-step cloning approach can generate a double-targeted, E1B55k-deleted CRAd using a straight-forward, modified in vitro ligation procedure.
KeywordsHomologous Recombination Hsp70 Promoter Packaging Cell Line Shuttle Plasmid Unique Restriction Site
multiplicity of infection
heat shock protein
multiple cloning site
inverted terminal repeat
tumor specific promoter.
Adenoviruses (Ad) are excellent gene transfer vectors and are extensively used for high-level transgene expression in vitro and in vivo. One of the attributes rendering Ad vectors particularly well suited for this purpose is the capacity to manipulate their genome. Typically, Ad vectors are converted into mammalian gene transfer vectors by replacing the E1 gene with the foreign gee of interest. E1 deletion serves two goals, i.e. to increase the cloning capacity to 5 kb (and to 8 kb if the E3 region is also deleted) and to render Ad vectors replication-incompetent.
Two approaches to have been traditionally used to construct recombinant Ad vectors. First, homologous recombination (HR) in either packaging cell lines, bacteria or yeast. Second, in vitro ligation using standard molecular biology procedures. These two methods rely on the fact that purified, linear Ad DNA is infectious, resulting in recombinant Ad virions after transfection into packaging cell lines. In vitro ligation was the first approach to manipulate the Ad genome. However, it was abandoned for many years because the large, 35.9 kb Ad genome that encodes more than 50 gene products, generally lacks sufficient unique restriction sites. To overcome this limitation HR was developed, initially using ClaI digested viral DNA. Next, the viral DNA was transfected into the packaging cell line to undergo spontaneous, albeit inefficient HR. To enhance HR efficiency in mammalian cells, a two-plasmid rescue system was developed [1, 2]. According to this method, two non-infectious plasmids with two overlapping Ad sequences were co-transfected into 293 cells, yielding an infectious recombinant Ad. The first use of the two plasmid system to construct recombinant Ad vectors incorporating foreign transgenes was developed by Graham's lab . The manipulated plasmid encoding the transgene has been termed the shuttle plasmid and was co-transfected into 293 cells with along with a large plasmid containing the Ad backbone, incorporating a 2.2 kb ampicillin resistance gene in the XbaI site. Because neither plasmid is infectious alone, recombinant E1-deleted Ad vectors were generated as a result of HR between the overlapping regions. Further modifications were made by Bett who removed a large fragment from the E3 region and introduced a unique PacI site immediately adjacent to the Ad inverted terminal repeat (ITR), allowing linearization of the backbone plasmid . HR thus employs two plasmids with overlapping sequences to be recombined in E1-compelmenting 293 or 911 cells.
To construct Ad vectors encoding heterologous genes, the smaller plasmid, termed the shuttle plasmid and containing the left ITR, a packaging signal and a sequence overlapping with the larger Ad backbone plasmid, was engineered to encode the gene of interest. The larger plasmid, contained almost all the entire Ad genome devoid of the packaging signal and the E1 and E3 genes, was co-transfected into 293 cells where HR was to generate the recombinant Ad genome. However, this procedure required plaque purification of the recombinant Ad vector by screening individual clones. In addition, HR in 293 cells occurs in low frequency, is a time-consuming procedure and recombinant Ad vector progeny may be contaminated with the wild-type Ad virus. Thus, to obtain a pure Ad preparation at least two rounds of plaque purification assays may be required. To facilitate vector selection in cells, screening for the recombinant virus has been facilitated by using counter-selection methods, extensive fragmentation of Ad DNA complexed with the terminal protein or by using Cre-lox-mediated recombination . These methods had the advantage that copies of the recombinant viral DNA were purified from clones and could therefore generate homogenous Ad preparations.
Optimization of two-plasmid HR was further reported in bacterial systems , using reconstitution of the sequence of the recombinant Ad in BJ5183 E. coli strains before transfection into 293 cells. Another advantage was preparation of the recombinant Ad genome in large quantities before transfection into 293 cells. The BJ5183 bacterial strain is recombination-proficient because i) it lacks the RecBCD enzyme and ii) it contains the RecF enzyme allowing DNA strand exchange between two linear DNA molecules that share at least 50 homologous bp at each end. One of these DNA molecules is the linearized shuttle plasmid previously subcloned with the heterologous gene and containing a bacterial origin of replication, a kanamycin-resistance gene and the left and right inverted terminal repeats (ITR) segments of the Ad genome.
While HR in bacteria has been generally accepted, it requires multiple steps in different E. coli stains. In addition, while in mammalian cells, the rate-limiting recombination step produced only the correct viral product of recombination, constructing the Ad genome by recombination in E. coli or yeast does not guarantee a correct, infectious recombinant Ad DNA. Unpredicted recombination events can occur, especially in yeast , requiring a thorough analysis of the recombinants to exclude clones that are unable to generate virus.
Similarly, Souza and Armentano employed the same logic for the construction of Ad serotype 2 vectors using I-CeuI and SnaBI sites . However, their pAdvantage™ system has a substantially lower reported cloning efficiency than the method reported by Mizuguchi and Kay [8, 9].
Danthinne also developed a HR-free system ('AdenoQuick™') for preparation of replication-deficient Ad, using a relatively complex system requiring encapsidation of the recombinant Ad genome into λ phage particles .
In the context of oncolytic Ad, re-introduction of modified E1 cassettes is required to construct conditionally-replicative Ad (CRAd). The superiority of CRAds over replication-deficient Ad vectors for cancer includes selective amplification of the cancer cell-killing capacity while relatively sparing normal cells. The attributes of CRAds include lysis of tumor cells that are resistant to standard therapy, selective cancer cell killing and induction of cell-mediated anti-tumor immunity. Because there is currently no established effective therapy for disseminated Ad infection, CRAd replication should be stringently restricted to permissive cancer cells. To this end, two genetic approaches have been reported involving either insertion of a tumor-specific promoter (TSP) upstream of E1a and E4 or partial deletion of E1 genes, e.g. E1B55k or the E1A conserved region 2 (CR2) . Infrequently, double-targeted CRAds may combine both types of genetic modifications.
In contrast to generation of replication-deficient Ad vectors, construction of CRAds is not straight-forward and may be time and labor-consuming. Two molecular strategies to construct CRAds have been described. First, and by far more popular, HR has been used by many groups to recombine an E1-encoding shuttle plasmid with the pAdEasy backbone plasmid [13, 14, 15, 16, 17, 18, 19, 20, 21, 22]. The second strategy to generate CRAds involves in vitro ligation [23, 24, 25, 26, 27, 28]. The setbacks of HR in the context of CRAds include the unpredicted chance of in vivo HR in BJ5183 bacteria and the lack of sufficient plasmid DNA production in BJ5183 bacteria required to allow diagnostic restriction analysis, thereby requiring transfer to candidate plasmids into other E. coli strains, such as DH5α.
Despite the utility of HR, potentially targeting virtually the entire Ad genome, when HR fails, alternative approaches to construct CRAds may be required. The two reported approaches to generate CRAds using in vitro ligation differ in their methodologies [23, 24, 25, 26, 27]. First, Hernandez-Alcoceba et al have reintroduced E1a and E4 into the Ad genome [23, 24]. To generate a CRAd, direct cloning of TSPs upstream of E1a or E4 within the Ad backbone is feasible using unique restriction sites flanking the promoter regions of E1A (Bst BI sites) and E4 (I-Ce uI and Swa I sites) [23, 24].
The setbacks of this approach may include the complete lack of the two E1B genes and the large Ad backbone containing the E1 and E4 genes, rendering manipulation of E1/E4 within the large Ad genome less efficient than in smaller shuttle plasmids which are readily amenable for genetic modifications.
The second in vitro ligation approach to generate CRAds without HR was developed by Danthinne [25, 26, 27, 28]. This method is based on the AdenoQuick™ system, originally developed to construct replication-deficient E1-deleted Ad vectors . To construct CRAds, this strategy involves re-introduction of modified E1 genes into a shuttle plasmid containing the Ad left arm and the Ad packaging signal. While several CRAds were reported using this system [25, 26, 27, 28], the process is rather complex and may require multi-order fragment ligation [25, 27] and a cos site next to the gene of interest for subsequent in vitro packaging into phage λ after cloning the recombinant Ad backbone . Thus, this in vitro methodology for CRAd construction may be too labor and time-consuming to investigators who are not experts in the field.
In the current study we report a simplified, HR-free in vitro ligation approach to construct a double-targeted, TSP-regulated, ΔE1B55k CRAd involving a direct, 3-step cloning strategy. This modified in vitro ligation system is a simple and efficient method to construct recombinant ΔE1B55k CRAds.
Results and discussion
As a proof of principle to construct a double-targeted CRAd via in vitro ligation, CRAd-hsp70E1Δ55k was generated using two genetic modifications, e.g. cloning the heat-shock 70 promoter (hsp70) TSP upstream of the E1A gene and deletion of the E1B55kD gene (Fig. 1). Overall, we combined components from the AdEasy  and the Adeno-X [8, 9] systems. An E1-encoding plasmid, constructed on the basis of the pShuttle of the AdEasy system , was employed to subclone the hsp70 promoter upstream of E1 (Fig. 1b). However, instead of using this shuttle plasmid for HR, we further subcloned an E1B55k-deleted, TSP-regulated E1 construct from this shuttle plasmid into the shuttle vector of the Adeno-X system (Fig. 1d). The modified E1 construct was then directly cloned into an E1/E3 deleted Ad backbone genome using in vitro ligation instead of HR (Fig. 1e, f).
Of note, CRAd-hsp70E1Δ55k manifested a decreased cytopathic effect relative to AdCMVE1 (not shown), deriving from attenuated viral replication. The relative viral DNA replication rate of CRAd-hsp70E1Δ55k after 36 hours was 1.98% and 1.52% of Ad viruses driven by the CMV promoter (AdCMVE1) or the native E1a promoter (Ad5luc3), respectively (Fig. 4b). The fidelity of the hsp70 promoter in CRAd hsp70E1Δ55k was not maintained during heat induction, i.e. there was no heat-shock induction of CRAd hsp70E1Δ55k DNA replication (not shown). This finding is in accordance with previous reports of hsp70 promoter in replicating Ad [20, 33]. Thus, while heat-shock enhances the oncolytic effect of CRAds in general , the hsp70 promoter does not provide tight control in the context of CRAd-hsp70E1Δ55k.
CRAd-hsp70E1Δ55k differs from two previously-reported CRAds with hsp70 promoters. The first hsp70 CRAd contained a ΔE1B55k deletion, but the hsp70 promoter did not drive E1a expression but rather a cytosine deaminase/thymidine kinase (CD/TK) suicide gene . The second hsp70 CRAd combined both hsp70 and mouse tyrosinase promoters to drive expression of E4, without tumor-targeted partial E1a or E1B deletions . Of note, both these hsp70-CRAds, as the vast majority of other CRAds, were produced using HR.
Deletion of the E1B55k gene was the genetic modification in the first CRAd, dl 1520 (Onyx-015) . The deletion of a viral gene such as E1B55k that inactivates cellular regulatory proteins, e.g. p53, was suggested to restrict CRAd replication to cancer cells with specific genetic mutations thought to complement the viral genetic loss of function . The Ad E1B gene encodes two major species of mRNAs. One mRNA codes for a 19-kDa polypeptide (E1B19k) and the other codes for a 55-kDa protein (E1B55k). The two proteins are encoded by alternative reading frames and share no sequence homology. During Ad infection the E1B proteins inhibit apoptosis to allow Ad protein production and viral DNA replication. While the E1B19k protein has an anti-apoptotic effect mimicking the cellular bcl-2, The E1B55k protein exerts its anti-apoptotic effect via inhibition of p53-induced transcription. In addition, E1B55K induces a cellular environment conductive for viral protein synthesis via a complex with the Ad E4 ORF6 protein. This complex inhibits the transport of host cellular mRNA from the nucleus to the cytoplasm while selectively stabilizing and transporting viral mRNA. Because approximately half of human cancer types are mutated for p53, deletion of the E1B55k gene has been suggested as a tumor-targeted approach suggesting conditional Ad replication only in p53-mutant cells . This approach has been subsequently disputed , but E1B55k deletion is still considered one of the genetic approaches to mitigate Ad replication in normal cells , probably via loss of E1B55k-mediated late viral RNA export rather than p53-inactivation . The cloning strategy reported here allows the deletion of E1B55k via digestion of the E1-encoding vectors pMKE1 or phsp70E1 at the kpn I site. A prior step of subcloning a TSP upstream of the E1a gene allows the construction of a double-targeted E1 gene for subsequent cloning into the Ad genome. Thus, in view of the relatively complex construction of CRAds using HR [13, 14, 15, 16, 17, 18, 19, 20, 21, 22] and previously-reported in vitro ligation approaches [23, 24, 25, 26, 27, 28], the method we propose has the advantage of only 3 standard cloning steps required to produce a double-targeted CRAd.
The drawbacks of this approach include limitation to E1-encoding expression cassettes free of Cla I, Swa I and Pac I, Bst BI and Psp1406I sites. In addition, in vitro ligation is generally limited to recombining the ends of the Ad genome at the E1 and E4 regions, while HR can use linearized vectors to target virtually any region in the Ad backbone to allow point mutations and small deletions.
In conclusion, HR and in vitro ligation have been frequently and rarely employed to generate CRAds, respectively. The simplified in vitro ligation system described herein may be efficiently employed to construct TSP-driven, E1Δ55k-deleted CRAds.
Materials and methods
Cells, plasmids and reagents
293 and A549 cells were from ATCC (Manassas, VA) and maintained in DMEM supplemented with 10% fetal calf serum, 100 units/ml of penicillin and 100 μg/ml of streptomycin in 5%CO2. The SPV-110 plasmid (StressGen, Victoria, Canada) contains the 383 bp human hsp70B promoter (GenBank accession no. X13229), flanked by XhoI and SalI restriction sites. pMKE1 was previously described . Briefly, it was derived from the shuttle plasmid of the pAdeasy system where the multi-cloning site and right arm in the pAdeasy pShuttle were replaced by the 2.6 kb midkine (MK) promoter and the complete Ad E1 region and protein IX genes, devoid of the native Ad E1a promoter region (Δ324–488) . pShuttle and pAdenoX backbone were from Clontech, San Jose, CA (currently BD, Franklin Lakes, NJ). All restriction enzymes and ligase were from New England BioLabs except for I-CeuI and PI-Sce-I (Clontech). All DNA purificaiotn and cell lysis kits were from Qiagen (Valencia, CA).
Standard molecular biology procedures for in vitro CRAd construction
Screening for recombinants was with miniprep plasmid DNA prepared by standard alkaline lysis kit (Qiagen), restriction analysis and 0.8% agarose gel electrophoresis with ethidium bromide staining. Large scale Ad backbone plasmid and recombinant Ad plasmids were purified by phenol-chloroform extraction and ethanol precipitation or Maxi-prep kit (Qiagen). Digested hsp70 promoter and hsp70E1Δ55k fragments were purified by gel extraction after agarose gel electrophorsesis. To clone the hsp70E1Δ55k fragment from phsp70E1 into pShuttle (Fig. 1c, d), a ratio of 3:1 (insert: digested pShuttle) was optimal. Ligation of the I-Ceu/I-PI-Sce-I flanked hsp70E1Δ55k fragment (after digesting 1 μg of phsp70E1Δ55k for 3-hrs) into 750 ng of the pAdeno-X backbone was performed with T4 DNA ligase overnight at 16°C. Cloning efficiency was enhanced by heat inactivation of the T4 DNA ligase and by digesting with SwaI that recognizes a unique site between the I-Ceu-I/PI-Sce-I sites of the Adeno-X backbone (but not within the I-Ceu-I/PI-Sce-I fragment from phsp70E1Δ55k), thereby preventing religation of the E1-deleted backbone. The recombinant pAdhsp70E1Δ55k plasmid was used to transform 50 μl chemical-competent DH5α bacteria via 60 sec heat-shock and selection with ampicillin (100 μg/ml).
To rescue CRAd-hsp70E1Δ55k, 5 μg of pAdhsp70E1Δ55k were linearized with PacI to expose both ITRs and DNA was purified by phenol-chloroform. Next, to transfect 293 cells, 4 μg of the digested DNA was mixed at room temperature with 20 μl Lipofectamine (Life Technologies), diluted in 500 μl of OptiMEM (Life) and added to the DMEM media for 4-hrs. Media was replaced to growth media and 293 cultured for 12 days until a cytopathic effect was observed. After three cycles of viral propagation, CRAd-hsp70E1Δ55k was purified using double CsCl density gradient centrifugation and tittered using both DNA optical density at 260 nm (viral particle [vp] concentration) and plaque assays in HEK 293 cells (infectious units [iu]). The vp/iu ratio was 100.
Validation of constructs using PCR
The oligonucleotide primer corresponding to the human hsp70B promoter was sense strand 5'-AGCTAGAACCTTCCCCGCG-3'. The oligonucleotide primers corresponding to the Ad E1a were sense strand 5'-TCTTGAGTGCCAGCGAGTAG-3' and the antisense strand was 5'-AAGTCCAAAGGTTGCCCAGG-3'. The oligonucleotide primers corresponding to the Ad E1B were sense strand 5'-TTTTCTGCTGTGCGTAACTT-3' and the antisense strand was 5'-ATCTTCATCGCTAGAGCCAA-3. For whole E1 PCR the forward primer was the sense sequence of E1a and the reverse prime was the antisense sequence of E1B. The oligonucleotide primer corresponding to the forward I-Ceu-I sense strand was 5'-TAACTATAACGGTCCTAAGGTA-'3 and the reverse PI-Sce-I primer was 5'-TTTCTCCGCACCCGACATAGA-'3. To analyze viral DNA with standard PCR, 1 μl of purified CRAd-hsp70E1Δ55k was directly applied into a 50 μl PCR reaction with 1 μl of each primer, at 30 cycles of 95°C 1-min denaturation, 50°C 2-min annealing and 72°C 2.5-min extension.
Analysis of CRAd replication with real-time PCR
One hundred and fifty thousand A549 cells were infected in triplicates with 10 MOI of CRAd hsp70E1Δ55k, AdCMVE1 or Ad5luc3. Thirty-six hours later (to allow one cycle of complete Ad DNA replication) cellular DNA was extracted using a Blood DNA purification kit (Qiagen). Real-time PCR conditions were 35 cycles of (94°C, 20 s→ 55°C, 20 s → 72°C, 30 s). Ad backbone plasmid pTG3602 (Transgene, Strasbourg, France) was available for plotting a standard curve for the E4 copy number. E4 copy numbers were normalized to the β-actin DNA copy number.
One μl of eluted DNA sample was analyzed by real-time PCR amplification to measure Ad E4 copy number as an indicator of CRAd replication and E1B copy number to confirm E1B55k deletion in CRAd-hsp70E1Δ55k and absence of wild type Ad contamination. The oligonucleotide primers corresponding to the E1B sense strand were 5'-GACAGGGCCTCTCAGATGCT-3' (Ad genome 3075–3094) and the antisense strand was 5'-5'-TGGCTACGTGAATGGTCTTCAG-3' (Ad genome 3144–3123). The oligonucleotide primers corresponding to the E4 sense strand were 5'-TGACACGCATACTCGGAGCTA-3' (Ad genome 34885–34905) and the antisense strand was 5'-TTTGAGCAGCACCTTGCATT-3 (Ad genome 34977–34958). The TaqMAn probe was 5'-CGCCGCCCATGCAACAAGCTT-3' (Ad genome 34930–34951).
This study was performed during a research fellowship at the Division of Human Gene Therapy, University of Alabama at Birmingham, Birmingham, AL, USA and was funded by the by the Birmingham Jewish Federation. The author wishes to thank David T. Curiel, Jerry L. Blackwell, Yasuo Adachi, Hongju Wu and Koichi Takayama for their support, helpful discussions and valuable reagents. Ming Wang performed the real-time PCR. Hui Li is acknowledged for technical support. Current research is funded by the German-Israeli Foundation grant no. 817/2004, Israel Science Foundation grant no. 573/03, the Israeli Ministry of Health, Chief Scientist Office, and the Hadassah Women's health fund.
- 15.Takayama K, Reynolds PN, Adachi Y, Kaliberova L, Uchino J, Nakanishi Y, Curiel DT: Vascular endothelial growth factor promoter-based conditionally replicative adenoviruses for pan-carcinoma application. Cancer Gene Ther 2007,14(1):105-16. 10.1038/sj.cgt.7700991PubMedCentralCrossRefPubMedGoogle Scholar
- 17.Bauerschmitz GJ, Guse K, Kanerva A, Menzel A, Herrmann I, Desmond RA, Yamamoto M, Nettelbeck DM, Hakkarainen T, Dall P, Curiel DT, Hemminki A: Triple-targeted oncolytic adenoviruses featuring the cox2 promoter, E1A transcomplementation, and serotype chimerism for enhanced selectivity for ovarian cancer cells. Mol Ther 2006,14(2):164-74. 10.1016/j.ymthe.2006.01.010CrossRefPubMedGoogle Scholar
- 19.Sauthoff H, Pipiya T, Heitner S, Chen S, Norman RG, Rom WN, Hay JG: Late expression of p53 from a replicating adenovirus improves tumor cell killing and is more tumor cell specific than expression of the adenoviral death protein. Hum Gene Ther 2002,13(15):1859-71. 10.1089/104303402760372954CrossRefPubMedGoogle Scholar
- 24.Cuevas Y, Hernández-Alcoceba R, Aragones J, Naranjo-Suárez S, Castellanos MC, Esteban MA, Martín-Puig S, Landazuri MO, del Peso L: Specific oncolytic effect of a new hypoxia-inducible factor-dependent replicative adenovirus on von Hippel-Lindau-defective renal cell carcinomas. Cancer Research 2003, 63: 6877-6884.PubMedGoogle Scholar
- 26.Li X, Zhang J, Gao H, Vieth E, Bae KH, Zhang YP, Lee SJ, Raikwar S, Gardner TA, Hutchins GD, VanderPutten D, Kao C, Jeng MH: Transcriptional targeting modalities in breast cancer gene therapy using adenovirus vectors controlled-lactalbumin promoter. Mol Cancer Ther 2005, 4: 1850-1859. 10.1158/1535-7163.MCT-05-0167CrossRefPubMedGoogle Scholar
- 36.O'Shea CC, Johnson L, Bagus B, Choi S, Nicholas C, Shen A, Boyle L, Pandey K, Soria C, Kunich J, Shen Y, Habets G, Ginzinger D, McCormick F: Late viral RNA export, rather than p53 inactivation, determines ONYX-015 tumor selectivity. Cancer Cell 2005,6(6):611-623. 10.1016/j.ccr.2004.11.012CrossRefGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.