Virus Genes

, Volume 36, Issue 1, pp 55–66

Evaluation of the T-REx™ transcription switch for conditional expression and regulation of HSV-1 vectors


    • Forschungsschwerpunkt Angewandte TumorvirologieDeutsches Krebsforschungszentrum, DKFZ - ATV
  • Oksana Zavidij
    • Forschungsschwerpunkt Angewandte TumorvirologieDeutsches Krebsforschungszentrum, DKFZ - ATV
  • Ingeborg Rezuchova
    • Institute of VirologySAS
  • Július Rajčáni
    • Institute of VirologySAS

DOI: 10.1007/s11262-007-0178-9

Cite this article as:
Knopf, C.W., Zavidij, O., Rezuchova, I. et al. Virus Genes (2008) 36: 55. doi:10.1007/s11262-007-0178-9


Herpes simplex virus 1 (HSV-1) strain ANG and ANGpath were cloned as bacterial artificial chromosome (BAC). Two different types of BAC genomes were obtained. BAC genomes of type I contained the BAC replicon at the intended target region between the genes of UL48 and UL49. In BAC genomes of type II, the BAC sequences were found to be aberrantly fused between the termini of the HSV-1 genome. Both the BAC types were used to establish a conditional gene expression system for HSV-1 by Flp recombinase-mediated insertion of expression vectors that were modified to respond to the T-REx™ tetracycline (Tet)-inducible transcription switch. During BAC cloning and mutagenesis in E. coli, not only deletions but also defined mutations of the HSV-1 genome were observed. Successful virus reconstitution from BACs with large inserts demonstrated that HSV-1 has a packaging capacity for foreign sequences of at least 8.1% of its genome size. Targets for Tet-regulated gene expression were the viral DNA polymerase gene (pol) and a reporter gene of glycoprotein B fused to enhanced green fluorescent protein (gBGFP). Results with the pol gene as target showed that virus plaque production could not be significantly controlled by the T-REx™ gene switch using vectors encoding one copy of the tetR gene. In contrast, an efficient Tet-response was achieved with the gBGFP reporter, which was optimal in a Tet repressor (TetR)-producing cell line, demonstrating that the TetR concentration provided by the virus was not sufficient for a tight control of Tet-regulated gene expression.


HSV-1 ANGANGpathBAC cloningRed/ET recombinationTetracyclineT-REx™ transcription switchGene expression


Herpes simplex virus 1 (HSV-1) has number of properties that make it an attractive platform for new vector developments with the central nervous system as target [1]. The advantage is its pronounced neurotropism, leading to the establishment of latent infections. Its large genome persists as an episome in the neuronal cells, and has a great carrier capacity for therapeutic genes. As a basis for oncolytic vector design and study of viral pathogenicity, our labs have cloned the HSV-1 genomes of strain ANG and its pathogenic derivative (ANGpath) as bacterial artificial chromosome (BAC) by homologous recombination in eukaryotic cells, using the technique exemplified for the human cytomegalovirus (hCMV) genome [2]. In order to construct replication-competent vectors without sequence deletions, the gene junction between UL48 and UL49 was selected for the integration of the BAC replicon. In addition, it was mandatory for the therapeutic application of the recombinant vectors to remove the BAC replicon from the vector genome. Therefore, we applied a BAC replicon that was flanked by loxP sequences, and could be excised by Cre recombinase [3]. The established HSV-1 BACs were employed for the construction of therapeutic vectors with suicide phenotype, in which both expression as well as replication should be controlled by a tetracycline (Tet)-controlled transcription switch, following the strategy exemplified for CMV [4]. The Tet-regulatable expression vectors contained (i) the Tet repressor gene (tetR) under the control of the hCMV major immediate-early enhancer–promoter (PCMV), and (ii) the gene of interest under the control of the Tet-inducible hybrid hCMV promoter (PCMV2tetO) bearing two minimal tet operator sequences of type 2 (2tetO) inserted 10 bp downstream of the TATA element [5]. The regulation vectors were recombined with HSV-1 BAC using Flp recombinase-mediated insertion [6]. In the present study, we have chosen the viral DNA polymerase gene (pol) as target for controlling virus replication by applying a recently introduced Tet-inducible transcription switch (T-REx™; Invitrogen, Karlsruhe, Germany) [5], which was successfully tested in a replication-defective HSV-1 vector [7]. Using Red/ET and Flp recombinase-mediated recombination for targeted mutagenesis [6, 8], we generated several pol (UL30) knockout (K.O.) mutants, in which PCMV2tetO or only 2tetO elements were inserted at the 5′-control region of the gene. In addition, for in situ visualization and optimization of conditional gene expression, we have constructed a reporter gene that encoded a fusion protein (gBGFP) of HSV-1 glycoprotein B (gB) and the enhanced green fluorescent protein (EGFP). The gBGFP fusion gene under the control of PCMV2tetO was inserted together with the tetR gene at two different locations on the HSV-1 genome: (i) into the ribonucleotide reductase (RR) gene locus (UL39, UL40) of wildtype ANG BAC, and (ii) in the terminal repeat sequences of the large unique region (TRL) of a ANGpath BAC clone, in which the BAC replicon was aberrantly fused between the genome termini. Our results show that the applied T-REx™ transcription switch in a replication-competent HSV-1 vector, encoding only one copy of the tetR gene and with the pol gene under the control of a 2tetO-bearing hCMV promoter, is inappropriate for controlling virus replication as determined by virus plaque formation. In contrast, an apparently tight control of gene expression was achieved with virus vectors, in which the gBGFP reporter was set under the control of T-REx™ Tet-regulation, but only in cells constitutively expressing TetR, manifesting that the TetR concentration is the critical factor for an optimal functioning of the T-REx™ transcriptional switch.

Materials and methods

Lipofectamin, Plus reagent, doxycycline, Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum (FBS), penicillin, streptomycin, and Opti-MEM I were purchased from Gibco/Invitrogen (Karlsruhe, Germany). Luria broth (LB) without additions was used for bacterial growth throughout. l-arabinose was purchased from ICN Biomedicals (Meckenheim, Germany). Restriction enzymes and reaction buffers were routinely used from New England BioLabs (Frankfurt a.M., Germany), Roche (Mannheim, Germany), and Fermentas (St. Leon-Rot, Germany). HyperLadder I (Bioline, Luckenwalde, Germany) was used as DNA size standard. DNA oligonucleotide synthesis and DNA sequencing were kindly performed by Wolfgang Weinig and Andreas Hunzicker, respectively (DKFZ Core Facility, Heidelberg, Germany).

Cells and viruses

The baby hamster kidney cell line BHK-21 (clone 13) was a gift of Prof. J. Subak-Sharpe (Medical Research Council, Glasgow), and was cultured in DMEM supplemented with 5% FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml) (supplemented DMEM/5% FBS). Standard HSV-1 strain ANG and ANGpath virus were propagated as described previously [9]. E. coli strains DH5α, DH10B, and PIR1 were obtained from Invitrogen (Karlsruhe, Germany). E. coli K12 DH5α strain BT340 containing pCP20 was a kind gift of Wilfried Wackernagel (Carl-von-Ossietzky University, Oldenburg, Germany). Mary K. B. Berlyn (E. coli Genetic Stock Center, Yale University, New Haven, CT 06520-8103) kindly provided E.coli strain MG1655 containing pKD46.


The nucleotide positions of the HSV-1 genome provided here refer to McGeoch et al. [10]. Design of plasmids and primers, and sequence handling were performed with Clone Manager 7, version 7.11, Professional Suite (Scientific & Educational Software Durham, NC 2772). Plasmid constructions were routinely confirmed by DNA restriction as well as sequence analysis. All plasmids containing the R6Kγ origin (ori) were propagated in E. coli strain PIR1, which provides the phage lambda Π protein. Plasmid pKD46 (Apr, repA101ts) carries the phage lambda recombination genes red α, β, and γ under the control of the araB (arabinose) promoter [11]. Plasmid pCP20 (Apr, Cmr, ts repA101ts) carries the yeast Flp recombinase gene [6]. The following plasmids were a kind gift of Eva Borst and Martin Messerle (Hannover Medical School, Hannover, Germany): pSLFRTKn [12], containing the kanamycin resistance (kanr) gene from transposon Tn903, and flanked by two identically orientated Flp recombinase recognition targets (FRT); pHA1, a derivative of pK18, containing a 1.5-kbp fragment from murine gammaherpesvirus 68 and a PacI fragment including the BAC replicon and a gpt (guanosine phosphoribosyl transferase) gene [13]; pOri6K-Kan1 (1,568 bp) [14], carrying the kanr gene with one flanking FRT signal site, and the Π protein-dependent R6Kγ ori [15]. Plasmid p6KFRT was constructed from pSLFRTKn by inserting between its NcoI and EcoRV sites the 450-bp EcoRI blunt-end/BspHI R6Kγ ori fragment of pOri6K-Kan1. Plasmid p11 derived from pUC19 containing a 11-kbp EcoRI joint fragment of ANGpath ΔBAC with sequences from nt 5,732 to 12,591 of the HSV-1 genome. The vectors pcDNA3.1, pcDNA4/TO as source for PCMV2tetO, and pcDNA6/TR as source for the tetR gene, controlled by PCMV, were obtained from Invitrogen (Karlsruhe, Germany). pcDKpn-nR derived from pcDNA3.1 clone containing the KpnI n fragment of wildtype (wt) HSV-1 ANG from nt 52,732 to 57,433. pcD4gB (5,690 bp) derived from pcDNA4/TO containing between EcoRV and XbaI sites the NruI/XbaI fragment from pcDKpn-nR comprising the UL27 gene from nt 52,732 to 56,097. pgBP77K79 derived from pUC19, containing an EcoRI/HindIII fragment generated by PCR with the gB gene of wt HSV-1 ANG spanning nt 53,020–56,086, in which the SanDI site was changed into a BamHI site (G- > A) at nt 55,597. Plasmid pEGFP derived from pIRES-EGFP (Clontech-Takara, Saint-Germain-en-Laye, France) by BamHI collapse. pgBGFP (6,181 bp) originated from pgBP77K79 containing the blunt-ended 1,757-bp NruI/XhoI fragment from pEGFP between the blunt-ended BamHI and SnaBI sites. Plasmid pgBGFP0 (5,442 bp) derived from pgBGFP by BamHI collapse. pgBGFP2 (4,607 bp) was constructed from pgBGFP0 by BsrGI collapse. pcD4gBG2 (7,351 bp) was obtained by inserting the 1,662-bp BsiWI/BstEII fragment of pgBGFP2 into pcD4gB. pcD4gBG05 (7,173 bp) derived from pcD4gBG2 by HindIII/BsiWI collapse. pR6KTR (3,738 bp) was constructed by inserting the blunt-ended SpeI/NotI fragment, encoding the tetR gene, from pcDNA6/TR into the EcoRV site of pOri6K-Kan.

Construction of Tet-regulated expression vector pRTRgBG

Plasmid pRTRgBG (6,658 bp) was constructed by inserting the 2,915-bp NruI/EcoRV fragment of pcDgBG05, containing the gBGFP gene, into the blunt-ended BstEII site of pR6KTR.

Construction of Tet-regulated expression vectors pRTRPOL and pRTRPOLi

Plasmid pPOLKZ served as source for the HSV-1 ANG pol gene with a Kozak initiation codon [16], and was constructed from pSKIILORF [17] by exchanging the HindIII/SnaBI fragment against a 204-bp HindIII/SnaBI-cleaved PCR fragment, generated with primer PATG 5′-ACC AAC CTT GCC ACC ATG TTT TCC GGT GGC GGC GGC CC, and primer PSNAB 5′-TCA TCG CAT TCG CTA TAG TAC G. pRTRPOL (8,735 bp) was constructed by inserting the 3,874-bp ClaI/blunt-ended XbaI pol gene fragment from pPOLKZ between the FspI and AccI sites of pRTRgBG (configuration a). pRTRPOLi (8,735 bp) was identically constructed as described for pRTRPOLi using pRTRgBG (configuration b).

Construction of Tet-regulated expression vector pRTRPOLII

Plasmid pRTRPOLII (7,609 bp) with a promoterless pol gene was constructed by inserting the 3,867-bp blunt-ended HindIII/XbaI pol fragment from pPOLKZ into the XmnI site of pR6KTR.

Construction of HSV-1 ANG BAC and ΔBAC

HSV-1 ANG BACs were constructed following the protocol of Borst et al. [2]. For homologous recombination the PacI fragment of pHA1 were flanked by subcloning in pUC19 with UL48 sequences, from nt 104,155 to 105,345 and with UL49 sequences, from nt 105,346 to 106,754. Gel purified 9,974-bp HindIII/SacI BAC replicon fragments were cotransfected with CsCl density gradient-purified wt or ANGpath DNA into BHK-21 cells, that were cultured in the presence of 100 μM mycophenolic acid and 25 μM xanthine. After four rounds of selection, Hirt DNA [18] was prepared as described previously [19] and electroporated into DH10B cells. Positive clones were characterized by restriction enzyme digestion.

BAC mutagenesis

HSV-1 BACs with gene knockouts were constructed in E. coli DH10B by the Red/ET mutagenesis method [8] that is based on homologous recombination between a linear fragment and the HSV-1 BAC. The recombination function was provided by plasmid pKD46. The linear fragment containing a kanr gene flanked by FRT signal sites was generated by PCR using p6KFRT as template, and the primers listed in Table 1. The 5′ end of each primer consisted of 50 nt, homologous to the target region of the BAC genome, and was followed by 25 nt that were specific for the kanr gene. Forward primers, TETO1 and TETO2, contained in addition 34 nt encoding two copies of tetO. PCR was performed with the hot-start Immolase DNA polymerase under the reaction conditions recommended by the manufacturer (Bioline, Luckenwalde, Germany).
Table 1

Primers for gene knockouts used in this study













The sequences of the 2tetO element are underlined

Transformation of E. coli DH10B

Transformation of E. coli with plasmids was routinely performed with the ROTI® Transform Kit and protocol (Roth, Karlsruhe, Germany). Electroporation was used for BAC DNA and DNA fragments. E. coli DH10B cells were made competent for electroporation as follows. Microfuge tubes (2 ml) containing 1.4 ml LB and proper antibiotics were inoculated with 30 μl of a fresh overnight culture, and grown at the required temperature by shaking at 1,000 rpm using a thermomixer. After 3 h, bacteria were collected by brief centrifugation (10,000 ×g, 30 s, 4°C) and resuspended in 2 ml ice-cold and sterile distilled water. After two further rounds of centrifugation and resuspension, cell pellets were resuspended in 45 μl ice-cold and sterile distilled water, 5 μl (0.5 μg) of DNA was added, and the cell/DNA mixture was transferred into electroporation cuvettes (0.2 cm Ø) and kept on ice until used. Electroporation was performed with a Micropulser (Biorad, München, Germany) at 1.35 kV and 5 ms pulses for DNA fragments, and at 2.5 kV for BAC DNA. Electroporated cells were taken up in 1-ml prewarmed LB and incubated at the selected temperature with 1,000 rpm for 1 h. Aliquots were plated on selection agar.

Generation of recombinant viral BACs for conditional gene expression

Escherichia coli DH10B cells, containing the recombinant KFRT BACs, were transformed with the temperature-sensitive Flp recombinase expression plasmid pCP20, and selected by overnight growth at 30°C on agar plates containing 17 μg Cm/ml and 100 μg Ap/ml. Kns/Apr/Cmr transformants, containing both pCP20 and FRT BAC, were obtained by replica plating on agar plates with 100 μg Ap/ml at 30°C, 30 μg Kn/ml at 30°C, and 17 μg Cm/ml at 37°C, respectively. Individual colonies were grown up, made competent with ROTI® Transform, and transformed with the Tet-regulated expression vectors. After 70-min incubation at 30°C, selection was carried by overnight growth on agar plates containing 17 μg Cm/ml and 30 μg Kn/ml at 42°C. Three to five colonies were grown up on a small scale for DNA analysis.

Analysis of BAC DNA

For restriction analysis, BAC DNA was isolated from 2-ml overnight E. coli cultures by the alkaline lysis procedure [20], and the BAC DNA suspended into 50-μl 10 mM Tris-HCl/1 mM EDTA pH 7.5 (TE). For large preparations, BAC DNA was isolated from 12× 2-ml bacterial cultures as follows. After neutralization and centrifugation (step 3 of the alkaline lysis procedure), supernatants were combined in a 15-ml centrifuge tube, and incubated for 60 min at 37°C in the presence of 25 mM EDTA and 50 μg/ml RNAse A (DNase free) (Roche, Mannheim, Germany). BAC DNA was extracted with one volume of phenol/chloroform/isoamyl alcohol (ratio 25:24:1) and separated by centrifugation (3,000× g/10 min/4°C) using a swing-out rotor. BAC DNA was precipitated by adding one volume of 2-propanol and carefully mixing, and collected by centrifugation (3,000× g/30 min/4°C). The DNA pellet was taken up into 80% ethanol and transferred into 2-ml reaction tubes. After centrifugation at 15,000× g and 4°C for 5 min in a tabletop centrifuge, the pellet was air-dried and resuspended in 300 μl TE.

Restriction enzyme digestions were carried out following the protocols of the manufacturers. Incubations using BamHI and EcoRI were routinely performed for 60 s at 650 W in a microwave oven with a turntable.

Reconstitution of recombinant viruses from BACs

For reconstitution of recombinant virus, HSV-1 BAC DNA was transfected into BHK-21 cells. When cytopathic effect (CPE) was complete, cells were subjected to three cycles of freezing and thawing, and virus pools prepared from the supernatants of a low speed centrifugation (500× g for 10 min at room temperature). To enrich for infectious virus, half-confluent BHK-21 cell monolayers grown on six-well plates were infected with aliquots of the supernatant virus, and viral growth continued until CPE was complete. Then virus stocks were prepared as stated above.

Transfection of HSV-1 BACs

HSV-1 BAC DNA was transfected into BHK-21 cells using Lipofectamine™ and Plus™ reagent according to the protocol of the manufacturer (Invitrogen, Karlsruhe, Germany). In brief, for transfection in six-well plates, 1 μg BAC DNA was diluted in 245 μl Opti-MEM and complexed with 5 μl Plus™ reagent, and in a separate reaction tube 245 μl Opti-MEM was mixed with 5 μl Lipofectamine™. After 20-min incubation at room temperature, both mixtures were combined, and kept for another 30 min at room temperature. The DNA-lipofectin complexes were applied dropwise to the BHK-21 monolayers which were kept in serum-free DMEM. After 4 h of incubation at 37°C with 5% CO2, 1.5 ml DMEM with 10% FBS was added; then incubation continued for 2–3 days until the CPE was complete. For fusion protein induction, 5 μg Dox/ml was added at the time of medium replenishment.

Growth kinetics of reconstituted BAC virus

Half-confluent BHK-21 cell monolayers grown in a series of 35-mm dishes were inoculated with 0.05 plaque-forming units (PFU) per cell for 1 h at 37°C. Then virus was removed and 4 ml of supplemented DMEM/5% FBS was added. At the stated times of infection, viral growth was terminated by placing the dishes in a −70°C freezer. At the end of the kinetics, all dishes were subjected to three cycles of freezing and thawing, and then virus pools were prepared from the supernatants by sedimentation at 500× g, and titrated.

Virus titration

From serial dilutions of virus stocks made in serum-free DMEM, 0.333 ml aliquots were inoculated per well of 12-well dishes containing the specified half-confluent cell monolayers. After 1 h of incubation at 37°C, each inoculum was removed, and virus growth continued until plaque formation in supplemented DMEM/5% FBS containing 1.5% carboxymethyl cellulose (low viscosity; Sigma, Taufkirchen, Germany) and the indicated Dox concentrations. Virus plaques were determined by direct microscopic visualization as well as from cell monolayers fixed with 5% formaldehyde, and stained with 1% crystal violet.


Phase contrast and fluorescence microscopy was applied to determine virus plaque formation and for in situ gBGFP expression using a Leica DMIL fluorescence microscope with a C Plan L20×/0.30 objective. Images were collected with a Leica DFC350 FX black-and-white camera using the Leica Software FireCam 1.2.0 (Leitz, Wetzlar, Germany).

Experimental approach

BACs with the genomes of the HSV-1 ANG standard virus and the pathogenic variant ANGpath were established using the technique exemplified for hCMV [2]. To achieve conditional Tet-regulated transcription for controlling viral replication and expression, we have adopted the system previously introduced for conditional CMV replication [4]. It consists of the insertion of a Tet-regulated expression vector into the BAC genome via FRT-mediated recombination. Selected gene K.O. mutants with a single FRT signal site were prepared using Red/ET and Flp recombinase-mediated recombination [6, 8]. The latter technique was also used to integrate a T-REx™ regulation vector, encoding the target gene under control of the 2tetO-bearing PCMV2tet0 or the endogenous promoter, and the tetR gene with PCMV. The BAC recombinants were then transfected into eukaryotic cells to generate infectious virus. Supernatant virus was collected and used to produce larger virus stocks. The Tet response of the reconstituted BAC virus was determined by the capability of viral plaque formation or the induction of gBGFP expression. In addition, a TetR-expressing cell line was applied for virus infection, in order to assess the effect of TetR concentration on Tet-regulation.


BAC cloning of HSV-1 ANG genomes resulted in two distinct BAC types

For BAC generation via homologous recombination, a floxed BAC-gpt gene cassette was constructed with flanking homologies for insertion into the intergenic region of UL48 and UL49 genes as described in “Materials and methods”. Examination of the resulting BAC clones yielded two different BAC types. In BAC of type I, representing the majority of ANG and ANGpath clones, the BAC replicon, as shown by BamHI restriction, was correctly integrated within the intergenic region of UL48 and UL49 (Fig. 1). As a consequence of this integration the BamHI fragment f was altered, and two novel fragments of 8,164 bp and 7,189 bp arose. Due to the genome circularization, the terminal BamHI fragments were absent in the BAC pattern. Interestingly, the BamHI fragment u was not found at its expected position in all HSV-1 ANG BAC clones examined so far. Its detection was facilitated, because in strain ANG the BamHI fragment u is separable from fragment v, due to a tandem duplication of oriL [21]. The size of the wildtype (wt) and ANGpath BACs, containing the oriL deletion, was calculated to be 159.4 kbp, which is 4.7% larger than the HSV-1 strain 17 standard genome [10]. From BAC of type II only one clone with an ANGpath genome was obtained, and designated ΔBAC. Sequence analysis as well as EcoRI restriction (Fig. 1) revealed that in type II the floxed BAC-gpt gene cassette was fused between the termini of the HSV-1 genome. In virtue of the novel junction in ΔBAC, three novel EcoRI joint fragments of 10,862 bp, 7,023 bp, and 5,898 bp were obtained. From the TRL sequences 5,732 bp were deleted and the remainder fused to 1,385 bp of E. coli sequences. The E. coli sequences were identified to be derived from a HindIII fragment (GeneBank accession no.U00096.2, pos. 19406-20791, Escherichia coli K12 MG1655, complete genome) that was apparently present as a contaminant in the gel-purified BAC recombination fragment. From the genome structure of ΔBAC, a hypothetical size of 157.9 kbp was calculated, which is 3.7% larger than the HSV-1 standard genome. After successful reconstitution of infectious virus, we have chosen ΔBAC and the wt BAC of type I, shown in Fig. 1, as starting material for the vector constructions.
Fig. 1

Genome organization of HSV-1 ANG BACs of type I and type II. In type I, the floxed BAC-gpt fragment was integrated between UL48 and UL49 genes as described in “Materials and methods”. BamHI patterns of the standard ANG virus DNA (ST), and of representative BACs containing the wildtype (WT; clone 6–4) and ANGpath (PATH; clone 1–3) genomes are shown together with a schematic illustration of the integration site and a BamHI restriction map. In type II, designated ΔBAC, the floxed BAC-gpt fragment was fused to the termini of the ANGpath genome together with a HindIII fragment of E. coli (E. c.) as shown schematically, and by the appearance of three novel fragments (highlighted by arrows) in the EcoRI profile of ΔBAC. Positions of BamHI fragment u, and EcoRI fragments l, m, n, and o are indicated

BAC mutants for Tet-inducible expression of DNA polymerase

In the first step of construction, we introduced a FRT signal site into selected regions of both the HSV-1 BAC types by Red/ET and Flp recombination, as described in “Materials and methods”. Three different pol gene knockouts, eliminating 94% of the pol-coding sequences, were prepared from wt BAC (Fig. 3A). In POLFRT, a single FRT signal site was placed immediately adjacent to a short open reading frame (SORF) that lies in front of the pol gene. In TETO1FRT, a FRT signal site flanked 5′-upstream by a 2tetO element was inserted at the major mRNA start site of the pol gene, and SORF was eliminated. In TETO2FRT, the FRT signal site and the 2tetO element were inserted right after SORF. POLFRT was used for the integration of the Tet-regulated expression vectors, pRTRPOL and pRTRPOLi, differing only in the orientation of the pol gene (Fig. 2), as described in “Materials and methods”. The resulting recombinant BACs were designated POL and POLi, and the BamHI restriction analysis from small-scale DNA preparations is presented in Fig. 3C. In these BACs, the synthesis of DNA polymerase (Pol) was expected to be regulated by PCMV2tetO.
Fig. 2

Tet-regulated expression vectors. Each vector is equipped with one FRT signal site, kanr gene for selection, Π-protein dependent R6Kγ origin (R6K), tetR gene with rabbit ß-globin intron II (IVS) and SV40 polyadenylation signal, (A)SV40, under control of hCMV major immediate-early promoter (Pcmv). In vectors, pRTRPOL and pRTRPOLi, the pol gene is regulated from the 2tetO-bearing PCMV2tetO. The pol gene in pRTRPOLII has no promoter. pRTRgBG contains the gBGFP fusion gene (gB/EGFP/gB) under the control of PCMV2tetO
Fig. 3

Construction and characterization of Pol BACs. (A) Scheme of pol gene (UL30) region with oriL palindromes, major mRNA start site, short open reading frame (S), and with the position of primers for generating the indicated pol gene knockouts (boxed) containing a single FRT signal site with and without 2tetO element. (B) BamHI maps of the pol region of wildtype BAC (WTBAC) and Pol BACs after integration of Pol Tet-regulation vectors. Novel BamHI fragments in bp used for the characterization shown in C are presented in bold. B, BamHI; for other abbreviations refer to the legend of Fig. 2. (C) BamHI profiles of wildtype BAC (WTBAC) and Pol BACs with highlighted novel bands in bp, and positions of fragments q, deleted u (Δu), and w. In all BAC constructs, due to the deletion of the oriL palindrome, fragment u runs in the position of fragment w. (D) Alterations in band mobility. In one out of three examined TETO1POL clones, and two out of three TETO2POL clones, differences in the mobility of BamHI fragments a and b were observed

Tet-regulated expression vector pRTRPOLII, containing a promoterless pol gene (Fig. 2), was inserted into TETO1FRT and TETO2FRT, yielding the recombinant BACs TETO1POL and TETO2POL (Fig. 3C). These constructions were performed to test the effect of the 2tetO positioning in the context of the endogenous promoter. In all the three vector constructs, the 5′-upstream sequence of the translation initiation codon of the pol gene was modified according to the Kozak consensus (AAGCTT GCCACC ATG) [16].

POL and POLi were calculated to have a size of 164.7 kbp, and with the size increase of 8.1%, these BAC recombinants carried the largest DNA insertion. For TETO1POL and TETO2POL theoretical sizes of 163.4 kbp and 163.6 kbp were calculated.

It should be noted that during the initial cloning of HSV-1 BACs, we consistently observed variations in the BamHI pattern, affecting predominantly the size of DNA fragments with known palindromic or repetitive sequence content, such as BamHI fragments u and b. Variations of fragment b that embraces internal repeat sequences of the large unique region, were also noted in the BamHI profile of TETO2POL clones (Fig. 3D). In two independent settings of the pol vector integration experiments, affecting one out of three progeny clones, and three out of three progeny clones, a larger BamHI fragment a was identified (Fig. 3D) that could arise only by the fusion of fragments a and c’ (Fig. 1, BamHI map).

BAC mutants for Tet-inducible expression of gBGFP

The RR locus in wt BAC and the terminal sequences in ΔBAC were selected as non-essential gene targets to test the Tet-inducible gene expression. In RR K.O. mutant RRFRT, in which both RR genes were completely removed (Fig. 4A), the Tet-regulated expression vector pRTRgBG (Fig. 2) was inserted by Flp recombinase-mediated insertion as described in “Materials and methods”. The latter vector was designed for in situ monitoring of the Tet-inducible gBGFP gene expression that was controlled by PCMV2tetO. The resulting gBGFP fusion protein consisted of the 5′-terminal 71 amino acid residues of gB, followed by 238 residues of EGFP, and terminated by 215 residues of gB. It also contained the trans-membrane anchorage signal and the syn3 mutation (V855A) of gB [22].
Fig. 4

Construction and characterization of RR K.O. mutants. (A) Scheme of the ribonucleotide reductase (RR) gene (UL39, UL40) region with the position of primers for generating the RR gene knockout, RRFRT, containing a single FRT signal site. (B) BamHI map of the RR gene region of wildtype BAC (WTBAC) and RRgBG. Novel fragments in bp, arising by the integration of Tet-regulation vector pRTRgBG and used for the characterization, are given in bold. For abbreviations, refer to the legend of Fig. 2. (C) BamHI profile of WTBAC and RRgBG. BamHI fragments h and o, and the novel fragments in bp are indicated

As a consequence of the insertion of pRTRgBG in RRgBG the original BamHI fragments h, h’, and o were altered, and four new fragments were identified in the BamHI profile of a representative clone, as depicted in Fig. 4C. The theoretical size of RRgBG was calculated to be 161.3 kbp.

Alternatively, regulation vector pRTRgBG was integrated into ΔBAC as illustrated in Fig. 5. For this purpose, p11-6, comprising the terminal 10,862-bp EcoRI fragment (Fig. 1) was opened with PmlI, and ligated with a 1,075-bp blunt-ended EcoRI fragment from pSLFRTKn [12] containing a FRT-flanked kanamycin resistance (kanr) gene. From the latter subclone the 4,256-bp NcoI/NheI fragment was used for Red/ET recombination (Fig. 5A). After excision of the kanr gene by Flp recombinase, ΔFRT was obtained, in which 1,904 bp of TRL, together with the E. coli and UL48 sequences, were replaced by a single FRT signal site. ΔgBG with a theoretical size of 160.1 kbp was successfully generated by inserting pRTRgBG into ΔFRT as shown by the EcoRI restriction analysis presented in Fig. 5C.
Fig. 5

Construction and characterization of ΔgBG. (A) Scheme of the 10,862-bp EcoRI fragment from ΔBAC, subcloned in p11-6, in which a FRT-flanked kanr gene cassette was inserted. The indicated 4,256-bp NcoI/NheI fragment was used for generation of ΔBAC mutants with a single FRT signal site. (B) EcoRI map of ΔBAC, and ΔgBG with integrated Tet-regulation vector pRTRgBG. Fragments (in bp) printed in bold were used for the characterization. For abbreviations, refer to the legend of Fig. 2. (C) BamHI profile of ΔBAC and ΔgBG. EcoRI fragments l, m, n, o, and specific fragments are indicated

Reconstitution of infectious virus from Pol BAC

For generation of infectious virus, BHK-21 cell monolayers were transfected with Pol BAC and cultured in the presence of Dox as described in “Materials and methods”. Microscopic analyses, performed 2 days after transfection, clearly showed in all BAC-transfected monolayers the formation of more or less extended syncytial plaques, a plaque type typical for strain ANG. According to the progress of infection, culture supernatants were collected between 3 and 5 days after transfection, and used to prepare virus stocks, as described in “Materials and methods”. Aliquots of virus stocks were used for another round of infection in the presence of Dox, and analyzed for plaque formation. Results showed that supernatant virus, rescued from transfections with POL, TETO1POL, and/or TETO2POL, produced syncytial plaques. In contrast, no plaques were detected in POLi-infected monolayers. The latter finding was repeatedly obtained with POLi constructs from wt ANG and ANGpath BAC genomes (data not shown), and is currently under study.

Since Pol is the major essential constituent of the viral replication machinery, we decided to assess the Tet-inducible Pol expression via syncytial plaque formation. To document the Dox response of Pol mutants, the reconstituted virus fractions were subjected to virus titration in the absence and in the presence of Dox using BHK-21 and LMTK tetR cells. The LMTK tetR cell line was established by transformation with pCDNA6/TR, and constitutively expressed TetR. The latter cell line was used to circumvent potential problems due to limited expression of TetR. The titration results, summarized in Table 2, showed that indeed no virus was detected in both POLi-infected cell lines at the lowest dilution step examined. All other reconstituted Pol BAC viruses yielded titers between 104 and 106 PFU/ml with BHK-21 cells, and 2- to 10-fold lower titers with LMTK TetR cells. Compared to TETO1POL, the titer obtained by virus rescued from TETO2POL was about two log units greater. However, and this was repeatedly confirmed, no significant Dox response was obtained.
Table 2

Titer of reconstituted BAC virus in the presence and absence of doxycycline on BHK-21 and LMTK tetR cells

BAC virus

Titer (PFU/ml)


MTK tetR






2.3 × 105

2.4 × 105

1.4 × 105

1.0 × 105







1.1 × 104

1.1 × 104

3.4 × 103

4.0 × 103


1.6 × 106

1.2 × 106

1.2 × 105

1.0 × 105

Since the reconstituted POL virus carried the largest DNA insert, and was the first engineered herpesvirus that contained a pol gene under the control of PCMV2tetO, the replication capability of this virus was examined by performing a growth kinetic as described in “Materials and methods”. While wt virus titers plateaued at 24 h after infection with 108 PFU/ml, the POL virus titer also leveled off at 24 h but remained three log units lower at 2 days after infection, indicating that POL virus replication was by far less efficient than that of the wt virus.

Dox response of gBGFP expression in cells infected with reconstituted virus of RRgBG and ΔgBG

RRgBG and ΔgBG were successfully used for reconstitution of infectious virus, reaching titers hat were comparable to those of reconstituted virus from Pol BACs in the absence and in the presence of Dox. For in situ monitoring of the Dox response of gBGFP expression, BHK-21 cells as well as LMTK tetR cells were infected with 50 PFU of reconstituted virus per monolayer. The virus replication proceeded in the presence and in the absence of Dox until plaque formation, then fluorescence microscopy was performed as described in “Materials and methods”. Figure 6 shows the phase contrast and fluorescence microscope images collected 2 days after infection. BHK-21 cells infected with reconstituted virus stocks of RRgBG or ΔgBG displayed a significantly greater green fluorescence in the presence of Dox than in its absence. Especially with RRgBG-infected monolayers, gBGFP expression was stronger and correlated well with the regions of syncytial plaque formation. In monolayers grown without Dox, the fluorescent staining was considerably reduced. The Tet-depending gBGFP expression was clearly improved with LMTK tetR cells. In the absence of Dox, gBGFP expression was below background fluorescence in both RRgBG- and ΔgBG-infected LMTK tetR cells indicating an efficient regulation by the T-REx™ Tet-inducible transcription switch. In the presence of Dox, gBGFP expression was significantly stronger in RRgB-infected monolayers in which all virus plaques and in addition many individual cells exhibited strong green fluorescent staining (Fig. 6).
Fig. 6

Dox response of gBGFP expression by ΔgBG- and RRgBG-infected cells. BHK-21 and LMTK tetR cell monolayers were infected with the indicated reconstituted BAC virus in the presence or in the absence of 5 μg/ml Dox. Phase-contrast and fluorescence images were collected (20× objective) as described in “Materials and methods”


Generation of BAC mutants for conditional gene expression

In this report, we analyzed whether or not a previously published conditional viral gene expression system is applicable for HSV-1 [4]. For this purpose, the genomes of wt HSV-1 strain ANG and of its pathogenic variant, ANGpath, were cloned as BACs following the protocol of Borst et al. [23]. Using Red/ET and Flp recombinase-mediated recombination [6, 8, 11], we could establish very quickly the described BAC mutants for Tet-controlled pol and gBGFP gene expression starting from FRT-bearing knockout mutants of HSV-1 BACs of type I, and of ΔBAC. With the exception of POLi BAC, infectious virus was successfully reconstituted from all BAC constructs, exhibiting calculated sizes between 157.9 kbp and 164.7 kbp. From POL BAC with the largest DNA insert, and its successful reconstitution and titration can be derived that mature virions of HSV-1 ANG have a packaging capacity for foreign sequences that extends the standard genome by at least 8.1%. Wade-Martins et al. [24] reported for HSV-1 amplicons a maximal packaging size of 153 kbp. This size discrepancy may be related to the different genomes. It should be mentioned that, in the context of this article, we have not examined whether the virus yield can be improved by removing the floxed BAC replicon from POL BAC.

Sequence deletions as well as mutations occur during BAC cloning in E. coli

During BAC cloning in E. coli, we have noted that the oriL sequences were deleted in all established HSV-1 BACs (Figs. 1 and 3). Deletions of the oriL sequences were previously reported to occur during cloning via plasmids in E. coli [2527], but could be prevented by using a different vector or another E. coli strain [2830]. Mutations in oriL that abolish the initiation of viral DNA synthesis have little effect on viral replication in cultured cells, but reduce pathogenesis during acute infection of mice and impair reactivation from latency [31]. It remains to be shown whether virus reconstituted from BAC genomes with deleted oriL shares the same characteristics, that could have a considerable drawback of its in vivo application as oncolytic vectors.

For BAC propagation in E. coli, it is known that tandemly repeated and repetitive sequences cannot be stably maintained [13, 32]. So far, there is no report that mutations are occurring during BAC cloning in E. coli. In the course of our cloning experiments, we found subclones with a larger BamHI fragment a (Fig. 3D). Since sequence analysis reveals that the location of the deleted BamHI site between fragment c’ and a in strain ANG (BamHI map in Fig. 1) is not embedded in a palindromic or repetitive sequence structure, the observed alteration can be explained only as a result of a sequence mutation affecting the respective BamHI site. In a more recent study, during cloning of gB K.O. BAC mutants, one out of five examined clones from a recombination experiment carried the same mutation (N. Strempel, unpublished observation). Presently, we have no clue how this mutation was generated and how frequently mutations in general occur during BAC cloning. Nevertheless, one should bear in mind that the E. coli DH10B cell might represent an error prone environment for the maintenance of viral genomes.

Characterization of BAC mutants for conditional gene expression

The choice of the viral Pol as target for Tet-controlled HSV-1 replication and its monitoring by virus plaque formation has turned out to be a very sensitive measurement. With several alternative constructions, in which the T-REx™ transciptional switch with PCMV2tetO [5] and the endogenous pol promoter with differently placed 2tetO elements were applied, no convincing Tet-controlled Pol expression was achieved as assessed by virus titration (Table 2). No improvement was gained with a host cell that constitutively produced TetR. Recent experiments were also unsuccessful, in which the tetR gene was replaced by the gene of the Tet-controlled transcriptional silencer [33] suggesting that the latter is not functioning in an HSV-1 environment (N. Strempel and C.W. Knopf, unpublished data). It may be argued that the positioning of the 2tetO elements in TETO1 and TETO2 was not optimal, while this should not play a role in POL, in which pol gene expression is controlled by PCMV2tetO. The finding that Tet regulation of the pol gene with the applied system was negative with the Pol mutants even by using a TetR producing cell line may be explained by the fact that the escape of only a few Pol copies is sufficient to replicate the viral genome.

More successful were the results for Tet-inducible gBGFP expression by RRgBG and ΔBAC (Fig. 6). The reason for this could be that, in contrast to Pol, the gBGFP fusion protein represents a rather insensitive probe for the in situ visualization, because in order to detect EGFP by conventional fluorescence microscopy about 10,000 molecules in single living cells are required [34]. In the same context belong recent positive results of the regulation of galactosidase expression with the T-REx™ transcription switch system using replication-defective HSV-1 vectors [7]. Whereas a less tight Tet-regulated gBGFP expression was achieved with BHK-21 cells that were infected with reconstituted virus from RRgBG and ΔgBG, a considerable improvement of the Tet-regulated transcription switch was gained when LMTK tetR cells, constitutively expressing TetR, were used as host for the infection. This result suggested that the TetR concentration supplied from the vector, encoding one copy of the tetR gene, was apparently not sufficient to warrant a stringent Tet-regulation. It remains to be shown, whether the introduction of two copies of the tetR gene into the vector as suggested [7] could overcome the TetR shortage, and possibly enable the construction of HSV-1 vectors for conditional expression and replication.


This research was supported by a BMBF grant for the scientific and technical collaboration with the Slovakian Republic, Project SVK No. 00/006.

Copyright information

© Springer Science+Business Media, LLC 2008