Murine leukemia virus (MLV) replication monitored with fluorescent proteins
© Sliva et al. 2004
Received: 26 November 2004
Accepted: 20 December 2004
Published: 20 December 2004
Skip to main content
© Sliva et al. 2004
Received: 26 November 2004
Accepted: 20 December 2004
Published: 20 December 2004
Cancer gene therapy will benefit from vectors that are able to replicate in tumor tissue and cause a bystander effect. Replication-competent murine leukemia virus (MLV) has been described to have potential as cancer therapeutics, however, MLV infection does not cause a cytopathic effect in the infected cell and viral replication can only be studied by immunostaining or measurement of reverse transcriptase activity.
We inserted the coding sequences for green fluorescent protein (GFP) into the proline-rich region (PRR) of the ecotropic envelope protein (Env) and were able to fluorescently label MLV. This allowed us to directly monitor viral replication and attachment to target cells by flow cytometry. We used this method to study viral replication of recombinant MLVs and split viral genomes, which were generated by replacement of the MLV env gene with the red fluorescent protein (RFP) and separately cloning GFP-Env into a retroviral vector. Co-transfection of both plasmids into target cells resulted in the generation of semi-replicative vectors, and the two color labeling allowed to determine the distribution of the individual genomes in the target cells and was indicative for the occurrence of recombination events.
Fluorescently labeled MLVs are excellent tools for the study of factors that influence viral replication and can be used to optimize MLV-based replication-competent viruses or vectors for gene therapy.
Efficient and long-lasting gene delivery is the major challenge in the development of vectors for gene therapy. Replication-competent retroviruses (RCRs) encoding suicide genes linked via an internal ribosome entry site (IRES) offer a significant advantage over replication-deficient vectors in cancer gene therapy, since they are able to spread efficiently in vivo [1–4]. Uncontrolled virus spread is, however, associated with serious risk of adverse events due to viral-integration mutagenesis. Therefore, for a therapeutic application, RCRs have to be equipped with additional safety features, e.g. transcription controllable by exogenous agents or viral entry restricted to the diseased cells. The selective delivery of a therapeutic gene by targeting retroviral entry would immensely reduce unfavorable side effects and ease the clinical application of gene therapy. The ecotropic MLV envelope protein does not recognizes receptors on human cells. An obvious challenge has been to extend the host range of vectors carrying the ecotropic envelope glycoprotein to a predetermined human cell type. This change in host range requires the inclusion of a novel attachment site and the induction of fusion via a novel receptor interaction. It has been shown before that it is possible to modify ecotropic Env and change its binding specificity, however, the efficient triggering of the membrane fusion or the escape from endosomes of viral particles targeted to e.g. epidermal growth factor (EGF)-receptor is still missing [5, 6]. The further development of such targeted vectors requires the understanding of the mechanisms that are involved in adsorption and internalization of retroviruses.
Investigating murine leukemia virus (MLV) replication is technically inconvenient because MLV infection does not cause a cytopathic effect in the infected cell. Viral replication can only be studied by immunostaining, measurement of reverse transcriptase activity or syncytia formation. We have developed a tool to simplify these analyses. We generated an MLV tagged with a fluorescent envelope protein, which allows viral replication and Env attachment to target cells to be followed by flow cytometry. This method will be useful for optimizing RCRs or retroviral vectors for gene therapy.
Sequestering of EGF-Env-containing viral particles has been described before [8, 9]. Viral particles containing EGF-Env were rapidly trafficked to endosomes and became degraded. This effect was dominant over the normal entry pathway, because mouse cells expressing the ecotropic receptor and the EGF-receptor showed a severely decreased infectivity of EGF-Env containing vectors . We were interested, if replication competent GFP-EMO might be useful to select viral variants able to escape the degradation in the endosomes. Transfection of GFP-EMO into cells expressing only the EGF-receptor (A431, COS-7) did not result in viral replication (data not shown). Therefore, GFP-EMO and GFP-MOV were transfected into FLY-Jet cells , which express the human EGF-receptor and the receptor for ecotropic MLV. Viral replication of GFP-EMO could be observed in FLY-Jet cells, although strongly delayed, after 10 days only 7.4 % of the cells were GFP-positive. After 38 days, all cells were GFP-positive and the N-terminus of the Env gene was analyzed by PCR amplification of genomic DNA isolated from infected cells. Predominantly a band migrating faster than the GFP-EMO fragment was amplified (Figure 1C), which was verified by sequence analysis to contain wt Env sequences. The less abundant, slower migrating fragments still contained the EGF sequences in Env. This confirms the sequestering of EGF-Env containing retroviral particles via the EGF-receptor. The selection of viruses able to escape the endosomal degradation was not possible and shows that degradation of viral particles in the endosomes favors the selection of wt Env-containing MLV, which escapes the sequestering by EGF-receptor.
The specificity of cell staining by supernatants containing GFP-MOV was further examined using chronically Mo-MLV-infected NIH3T3 cells (NIH3T3i-MLV). These cells have only negligible numbers of mCAT molecules on the cell surface, because Env expression leads to their retention within the cell (receptor interference). As expected, NIH3T3i-MLV cells produced no shift when incubated with GFP-MOV supernatants (Fig. 2B). Furthermore, binding of GFP-MOV supernatants could be inhibited by preincubation of NIH3T3 target cells with a soluble Env fragment containing the receptor binding domain (sRBD) derived from the ecotropic Env , but not with the equivalent sRBD derived from the amphotropic Env , which binds to a different receptor (Fig. 2B). This shows that GFP-tagging can be used to investigate Env-binding properties by flow cytometry.
The size of a retroviral genome is limited to roughly 11 kb. The capacity for the insertion of a therapeutic gene for gene therapy is, however, increased by the use of semi-replicative retroviral vectors (SRRVs), where the gag/pol and env genes are split between two viral genomes. We constructed split viral genomes and used fluorescent proteins to monitor the replication of the resulting SRRVs.
In addition, we examined the stability of the GFP-tagged Env in the split genome approach. As shown in Figure 5B, PCR analysis with primers flanking the GFP sequences in Env (p3, p4; Fig. 3) clearly demonstrated that GFP-Env is stable and the GFP sequences were not deleted from the viral genome after 32 days of culture (Fig. 5B, lanes 5, 6 and 7).
Our data demonstrate that labeling the MLV Env with a fluorescent protein is an easy method of monitoring MLV replication and the attachment of Env to target cells. This is especially useful for the development of novel cancer gene therapies that use replication-competent MLV encoding a cytotoxic gene . Labeling Env with GFP in the PRR leaves the 3' untranslated region at the Env boundary available for the insertion of IRES-linked therapeutic genes . These recombinant viruses could be monitored by GFP expression and would allow the study of replication kinetics in vitro and in vivo. The biodistribution of replication-competent viruses in animal models and their safety for cancer treatment could, thereby, be assessed.
A further improvement of replication-competent viruses would be tumor cell-specific entry. The inclusion of tumor-specific ligands into Env is one option to potentially expand the ecotropic host range of MLV to human tumor cells [6, 5]. Ecotropic MLV containing GFP-tagged Env can be used to analyze the receptor-dependent binding of the viral Env proteins to target cells. Labeling Env in the PRR leaves the N-terminus or the receptor binding site  available for further insertions of ligands to target tumor cell specific receptors. The use of GFP-tagged Env to determine receptor binding is very simple and in addition GFP-tagged Envs are helpful for the identification of recombinant viruses from retroviral library screens. GFP-Env fusions will therefore be very useful for the development of targeted vectors and as a screening system for retroviral-receptor antagonists. However, selecting EGF-Env containing MLV on cells that express both receptors (EGF- and ecotropic receptor) did not permit the isolation of a virus with an EGF-receptor specific tropism. EGF sequences were deleted from the viral genome in this setting. EGF sequences in Env, however, did not alter the replication kinetics in mouse fibroblasts (Fig. 1), which further indicates that targeting retroviruses to membrane spanning receptor tyrosine kinases inactivates retroviral particles.
In our experiments using semi-replicative retroviral vectors, we found that a rapid increase in GFP-positive cells correlated with the appearance of recombinations and the formation of full-length MLV genomes. This indicates that semi-replicative vectors have to be improved to avoid intergenomic recombination before they can be considered to be used for gene therapy. The recombinants did contain the GFP-Env gene, providing further proof that insertion of GFP into the proline-rich region of Env did not interfere with viral fitness.
Fluorescently labeled MLVs are excellent tools for the study of factors that influence viral replication and can be used to optimize MLV-based vectors or viruses for gene therapy. This method is not limited to ecotropic Env, but can be extended to amphotropic MLV, since it has been shown recently that the amphotropic MLV Env can also be tagged with GFP .
NIH3T3, A431, 293T and COS-7 cells were grown in Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% fetal calf serum, 4 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C in 10% CO2.
The construction of GFP-EMO has been described previously . GFP-MOV was generated by replacing a Pfl MI fragment of pGFP-EMO with wt MLV sequences using standard cloning procedures . GAG/POL-RFP was generated starting with the genomic MLV clone, pKAΔenv-egfp, which contains a 30 nucleotide-linker with an Sfi I-site introduced at position 5893 (all positions according to GenBank Accession No. J02255) and an additional Sfi I-site at position 5389 removed by mutation. The start codon of MLV env (position 5777) was deleted to allow translation to start at the inserted GFP sequence . We replaced GFP with RFP, which was introduced as a Sfi I-Cla I fragment. GFP and RFP sequences were derived from vectors purchased from Clontech (BD Biosciences Clontech, Heidelberg, Germany)
Plasmids encoding the MLV genomes or soluble receptor binding fragments (sRBDs)  were transfected using the calcium phosphate procedure . For the sRBDs, supernatant was collected two days after transfection, filtered through a 0.45 μm pore filter (Millipore, Eschborn, Germany) and 1 ml was used per binding assay.
Supernatants of tissue culture cells were collected, filtered through a 0.45 μm pore filter (Millipore, Eschborn, Germany) and added to target cells. After 5 min. at room temperature, the cells were spun down, redispersed in PBS and immediately monitored by fluorescence-activated cell sorting (FACScan, Becton Dickinson, Heidelberg) using the Cellquest software.
Green fluorescence protein (GFP) expression was monitored by a shift to green fluorescence (FL-1) and red fluorescent protein (RFP) by a shift to red (FL-2). FACS analysis was performed with FACScan (Becton Dickinson, Heidelberg) using the Cellquest software.
Genomic DNA was isolated after proteinase K digestion and phenol/chloroform extraction. PCR was performed using the manufacturers protocol (Qiagen, Hilden, Germany).
N-terminal EGF-Env sequences were analyzed using the primers BS-5: 5'-TCT GAG TCG GAT CCC AAA TGT AAG and MLV-5'-Env: 5'-TAA CCC GCG AGG CCC CCT AAT CC, which amplified a 899 bp fragment from GFP-EMO and a 726 bp fragment from wt MLV. The generation of full-length genomes was analyzed using the primers p1: 5'-GAA TAG AAC CAT CAA GGA GAC and p2: 5'-CTC GAG AAG CTT AGT ACT GA, which amplify a 600 bp fragment from full-length MLV. No fragment should be amplified from the semi-replicative vectors, because the primers bind to genes on separate constructs. The stability of the GFP-Env fusion gene was analyzed using the primers p3: 5'-GTC AGT AAG CTT CTC GA and p4: 5'-GGT TTT GTC AGG ACT GGT GAG, which amplify a 1.5 kb fragment from gfp-env and an 800 bp fragment form wt env.
We thank C. Haynes for helpful discussions and critically reading the manuscript. We are grateful to D. Lindemann, K. Cichutek and F.-L. Cosset for kindly providing the plasmids pczCFG5 IEGZ, pKAΔenv-egfp, E-sRBD and A-sRBD.
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.