If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Most of the human genome has now been sequenced and about 30,000 potential open reading frames have been identified, indicating that we use these 30,000 genes to functionally organize our biologic activities. However, functions of many genes are still unknown despite intensive efforts using bioinformatics as well as transgenic and knockout mice. Retrovirus-mediated gene transfer is a powerful tool that can be used to understand gene functions. We have developed a variety of retrovirus vectors and efficient packaging cell lines that have facilitated the development of efficient functional expression cloning methods. In this review, we describe retrovirus-mediated strategies used for investigation of gene functions and function-based screening strategies.
Function-based gene cloning
It was only 30 years ago that recombinant DNA technology was initiated [
]. For less abundant mRNAs, the cDNAs were frequently isolated based on the amino acid sequences of purified proteins. In late 1970s, a method called hybrid selection was developed. The principle of this method was to detect proteins translated from the mRNA hybridized to a particular pool of subdivided cDNA library fixed on nitrocellulose membranes, thereby identifying a pool that contains a cDNA of interest. Xenopus oocyte was used for production of proteins. This type of experiment is called “expression cloning,” which means “cloning of cDNA by detection of proteins expressed from cDNA libraries.” This strategy is suitable for isolation of rare cDNAs, such as cDNAs for cytokines and cytokine receptors. Levels of the protein expression are low, but a small amount of protein is enough to exert biologic functions.
A variety of expression cloning strategies has been established and utilized for cloning of cDNAs based on the biologic functions of their protein products. One of the early expression cloning methods used the Escherichia coli expression system for expression of cDNAs followed by detection by antibodies. In the early 1980s, the hybrid selection method was used for identification of cDNAs for cytokines using growth stimulation as a screening method. This method was later modified to directly transcribe cDNAs in Xenopus oocyte using the SP6 promoter. Alternatively, genomic DNAs were used to functionally clone cDNAs. The famous oncogene screening method, the focus-forming assay, using NIH3T3 cells falls into this category. It was notable that the invention of COS cells, in which plasmids can be amplified for the first time in mammalian cells, enabled expression cloning using mammalian cells [
]. A variety of cDNAs were isolated by the COS cells-based functional cloning method. However, this strategy depended on specific cells such as COS7 cells where the SV40 large T antigens are expressed to enable amplification of SV40 origin-bearing plasmids [
]. Therefore, only transient assays can be used for the screening.
To overcome the limitations of the conventional expression cloning system using COS cells, we and others turned to the idea of harnessing the power of retrovirus gene transfer to develop function-based screening of cDNAs.
Retrovirus-mediated expression screening; rationale and application
Retrovirus-mediated expression cloning was developed in mid 1990s [
]. One can generate either uni-directional or bi-directional cDNA libraries depending on the intended application. Complimentary DNAs are generated using either oligo-dT primers or random hexamer primers. The library is kept as DNA solution, and is converted to retroviruses by using packaging cell lines. To generate retroviruses that represent and cover a high complexity of cDNA libraries, it is recommended to use 293-based packaging cell lines that are efficient in transient packaging [
]. The virus stock containing high-titer retroviruses is used to infect target cells, and the infected cells are selected for the phenotype of interest. The integrated cDNA then is recovered by genomic polymerase chain reaction (PCR) or reverse transcriptase (RT)-PCR to determine which cDNA is responsible for the phenotype and is subjected to the sequence.
The retrovirus-mediated expression cloning method is efficient because the number of the provirus integrations in each cell is limited. Therefore, it is not necessary to recover and reintroduce the plasmid from, and into, the cells repeatedly, as in the conventional method using COS7 cells. In retrovirus-mediated expression cloning, the infection efficiencies should be controlled between 10% and 30% to avoid multiple integration in a cell as much as possible. Alternatively, one can recover the integrated retroviruses by transfecting a helper construct harboring gag-pol and env genes into the isolated clone that has acquired a phenotype of interest after transduction of the cDNA library. In this case, the recovered retroviruses are infected to the target cells to determine which integration was responsible for the phenotype.
The most important advantage over the conventional method is that any functional assay can be applied to identify cDNAs by their functions because, once integrated, the expression of the retrovirally transduced cDNA usually is stable.
Retrovirus-mediated expression cloning: some examples
A variety of functional assays can be utilized in retrovirus-mediated expression cloning. For instance, cellular receptors for various viruses were identified based on infectability of the viruses. Infection-resistant cells transduced with the library derived from infectible cells are screened by infection of the virus vector harboring a reporter gene such as GFP. The cDNA recovered from reporter gene-positive cell (i.e., infectible cell) should encode a receptor for a virus of interest. Co-receptors for human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) were identified in this way from cDNA libraries derived from human T cells [
Tumor necrosis factor (TNF) and Fas induce apoptosis through activation of downstream signaling pathways. After introduction of cDNA library into the cells, some cells may become resistant to TNF and Fas stimulation by expressing a retrovirally introduced cDNA. Using this strategy, a novel transcription factor BSAC and an adhesion molecule ICAM-2 have been identified as inhibitors of TNF-induced and TNF- and Fas-induced apoptosis, respectively [
]. Thus, identification of proteins based on their functions sometimes leads to unexpected and important findings. Another example of the unexpected result was encountered in our study. Mouse leukemic M1 cells differentiate into macrophages and undergo apoptosis upon interleukin-6 (IL-6) stimulation. Starr et al. [
] identified an inhibitor of the cytokine signal, SOCS-1, by isolating an IL-6–resistant M1 clone after transducing a cDNA library to M1 cells via retrovirus infection. By the same strategy, we identified A1, which is an anti-apoptotic protein of the bcl-2 family and protected M1 cells from IL-6–induced apoptosis [
]. We also discovered a novel GAP MgcRacGAP in the anti-sense orientation from an IL-6–resistant clone; expression of anti-sense MgcRacGAP protected M1 cells from IL-6–induced differentiation and apoptosis. On the other hand, overexpression of MgcRacGAP induced differentiation into macrophages in HL60 cells. An unexpected finding brought by the subsequent study was that MgcRacGAP is required for cell division, especially for cytokinesis [
]. It would be interesting to investigate the link between cytokinesis and cell differentiation. Thus, functional identification of protein is a powerful tool in cell biology.
Structure/function analysis using retrovirus-mediated screening combined with PCR-driven random mutagenesis
In addition to applications for a variety of expression cloning strategies, the retrovirus-mediated expression system can be used to identify a mutant molecule with altered functions. For instance, we identified constitutively active forms of a cytokine receptor MPL [
], and the ligated DNA was amplified in E. coli. Thus, the resulting plasmid DNA represented a mutation library of the protein of interest. This library was transiently transfected into a 293-based packaging cell to generate the retrovirus stock representing the mutation library of a particular molecule. It then was infected to target cells, and the infected cells were selected for a phenotype of interest, followed by retrieving and sequencing the integrated retrovirus of the selected clones. In our experiments, we used mouse IL-3–dependent Ba/F3 cells as targets, and selected the library-transduced Ba/F3 cells in the absence of IL-3 to isolate factor-independent clones. In that manner, we were able to identify constitutively active mutants of MPL [
] that induced factor-independent growth in Ba/F3 cells as well as other IL-3–dependent cell lines. The constitutively active mutants of various signaling molecules will be useful for analyzing signaling pathways. In addition to isolation of active mutants of various molecules, this strategy will be applicable in generating various mutants with acquired functions. For instance, one may want to generate restriction enzymes that recognize altered restriction sites, or cytokines with stronger activities or with less side effect in vivo.
SST-REX, an efficient signal sequence trap based on retrovirus-mediated gene transfer
Sorting of the protein between the cellular components is regulated by various sorting signals in the proteins, such as the nuclear localization signal and the mitochondrial targeting sequence. Signal sequence is one such sorting signal found in type I transmembrane proteins and secreted proteins. Tashiro et al. [
] developed an elegant method called signal sequence trap (SST) by which signal sequence-harboring cDNAs are specifically isolated. The rationale of the method is to search for a cDNA fragment that contains a signal sequence and directs a signal sequence-defective CD25 to the cell surface by the fusion. This method, however, is time consuming and leads to frequent isolation of false-positive clones. Klein et al. [
] invented a modified SST method using growth of a yeast mutant YT455 as a screening method. We applied retrovirus-mediated gene transfer and developed an efficient and accurate method SST-REX (signal sequence trap–retrovirus-mediated expression screening) using mammalian cells [
] in developing SST-REX, as illustrated in Figure 1. In brief, we construct a cDNA library in a retrovirus vector in which cDNA fragments are fused to an extracellular deletion mutant of the constitutively active MPL. The fusion library then is transduced into IL-3–dependent Ba/F3 cells via retrovirus infection. When the inserted cDNA fragment contains a signal sequence, it directs the mutant MPL on the cell surface and confers Ba/F3 cells factor independence. We then recover the integrated cDNA from the factor-independent Ba/F3 clones. Ten milliliters of the virus supernatant gives rise to isolation of about 1000 to 2000 signal sequences with 95 to 100% accuracy. Whereas previous methods collected shorter cDNA fragments for SST, we use cDNA fragments in the range from 0.5 to 6 kbp. In the results, the average length of cDNA fragments isolated in SST-REX is about 1 kbp. In SST, generally speaking, shorter cDNA fragments (shorter than 100–200 bp) will give false-positive results more frequently (unpublished results), and we believe this is a reason why SST-REX achieves higher accuracy than other methods.
FL-REX, a method by which cDNA is identified based on subcellular localization of its protein product
We also developed a novel expression screening method (FL-REX: fluorescence localization–retrovirus-mediated expression screening) in which a cDNA can be isolated based on subcellular localization of the protein [
]. Briefly, we express cDNAs as GFP-fusion proteins in NIH3T3 cells via a retrovirus vector, and GFP-fused proteins are identified by subcellular localization through fluorescence microscopy. With this method, it is possible to clone cDNAs for proteins that specifically localize in the nucleus, nucleoli, Golgi apparatus, cell surface, and mitochondria. If computerized fluorescence microscopy is combined with FL-REX, it also would be possible to isolate cDNAs whose protein products shuttle between different cellular compartments in response to various stimuli, such as cytokines, ultraviolet irradiation, and heat shock.
Genetic approaches in a mammalian system
Genetic approaches using complementation is another field for retrovirus-mediated expression cloning. It is possible to complement the deficiency of the mutant cells from patients with genetic disorders by introducing cDNA libraries from normal cells. If efficient assay systems were available, it would be easy to isolate the causative genes of various genetic diseases using retrovirus-mediated expression cloning.
It also is possible to first establish the mutant by chemical mutagenesis or irradiation and then complement the defect of the mutant using a cDNA library derived from normal cells to search for a missing gene in the mutant. In that way, one can identify a series of molecules responsible for a particular function. For example, Yamaoka et al. [
] identified BRF1 as a regulator essential for ARE (AU-rich element)-dependent mRNA decay. In this experiment, a mutagenized cell line (slowC) that failed to degrade cytokine mRNA was used. When a GFP reporter construct with ARE was introduced into slowC mutant, unlike in normal cells, mRNA for GFP was not degraded because of a missing factor that was responsible for the ARE-dependent degradation of the mRNA, thereby maintaining the high GFP expression. A cDNA library was introduced into the slowC mutant to identify the revertant based on the reduced expression of GFP, and BRF1 was found to be a regulator that degraded GFP mRNA in an ARE-dependent manner.
We also used a genetic approach combined with retrovirus-mediated expression cloning [
] to search for bone marrow stroma cell-derived growth factors (Fig. 2). We used the Ba/F3 cell line, which is unique because of its dependence on only mouse IL-3 and not other cytokines. Most bone marrow stroma cell lines do not produce IL-3 and therefore cannot maintain the growth of Ba/F3 cells. Our strategy was to establish stroma-dependent Ba/F3 mutants. Most of such mutants were found to be dependent on granulocyte-macrophage colony-stimulating factor (GM-CSF) or stem cell factor (SCF), thus indicating that they ectopically expressed receptors for GM-CSF and SCF through chemical mutagenesis. Some mutants did not respond to known factors, and we chose one such clone (S21), which can grow on ST2 stroma cells but not on other several lines such as MS10. Subdivided pools of an ST2-derived cDNA library are transduced via retrovirus infection into MS10 cells to identify the pool that contains a cDNA encoding the ST2-derived factor responsible for induction of S21 growth. After isolation of a pool expected to contain such a cDNA, the pool is further divided to eventually identify a single clone (sibling). In this way, we were able to identify a membrane integral protein with six transmembrane domains. This molecule had been originally identified as a soluble factor called immune suppressor factor (ISF), but curiously it turned out to be a subunit of the vacuolar-type ATP-associated pump. This experiment never would have been successful by the conventional expression cloning method utilizing COS7 cells because cell–cell interaction was required for the induction of cell growth by ISF. Intriguingly, when ISF was overexpressed in MS10 cells, the S21 Ba/F3 mutant as well as bone marrow progenitor cells proliferated on and underneath the stroma cells, and formed a cobblestone-like area, which is a hallmark of proliferation of hemopoietic progenitor cells.
Advantages of retrovirus-mediated gene transfer in investigating gene functions
In addition to the function-based screening of cDNA libraries, efficient retrovirus-mediated gene transfer is useful to investigate functions of genes, particularly those with inhibitory functions in the control of cell growth because it is difficult to establish stable transformants that express such genes. One may want to use inducible expression systems including the tet-repressor system. However, it turned out to be difficult to establish stable transformants expected to inducibly express genes such as SOCS-1, probably because even leaky expression of such genes hampers establishment of a stable transformant (unpublished results).
If the infection efficiency is more than 30 to 50% in a particular cell type, we should be able to readily see the gene function by observing the effect of gene expression in a bulk culture transduced with the gene of interest via retrovirus infection. It also is possible to use a bicistronic retrovirus vector that harbors the internal ribosomal entry site (IRES) and simultaneously expresses a gene of interest and a reporter gene from one mRNA. Alternatively, retrovirus vectors can be designed to express a dominant-negative form of a particular protein, the anti-sense cDNA, or the RNAi construct to investigate functions of a particular gene in a particular cell type. Using our efficient retrovirus-mediated gene transfer system described in the following, we are able to investigate gene functions even in primary cultured cells such as T cells and mast cells.
Improvement of retrovirus-mediated gene transfer
Improved retroviral systems are being developed to facilitate applications to expression cloning and functional genomics. Our own efforts to enhance both packaging and vector components are briefly reviewed in the following to illustrate possible approaches.
Improvement of retrovirus packaging cells
Retrovirus vectors lack viral structural genes and require packaging cells to generate viral particles [
] established the unique packaging cell line BOSC23; they stably introduced gag-pol and env expression constructs into 293T cells well known for high efficiency in transient transfection. Establishment of BOSC23 brought a revolution in retrovirus-mediated gene transfer. High-titer retroviruses (∼106 IU/mL) can be readily prepared in a couple of days by transiently transfecting retrovirus vectors into BOSC23 cells. However, it was difficult to maintain the potential of the BOSC23 cells to produce high-titer retroviruses by transient transfection during long-term culture.
To generate more stable and efficient packaging cell lines, we developed unique packaging constructs to express gag-pol and env genes (Fig. 3). The packaging constructs used the EF1α promoter for efficient expression in 293T cells, the IRES to simultaneously express viral structural genes and drug selection markers from one mRNA, and the Kozak sequence for efficient translation. These packaging constructs were stably transfected into 293T cells, and an efficient packaging cell line PLAT-E was obtained [
]. PLAT-E cells produce 1×107 IU/mL by transient transfection of pMX even after long-term culture if the cells are maintained in the presence of blasticidin and puromycin. We also used env genes derived from amphotropic retrovirus (4070A) and feline endogenous retrovirus (RD114) to generate PLAT-A and PLAT-F cell lines. The titers of retroviruses produced from PLAT-A and PLAT-F are 1×106 IU/mL, one tenth of those produced from PLAT-E (unpublished results). The RD114 Env derived from feline endogenous viruses was reported to facilitate gene transfer into human hematopoietic stem cells [
]. Another packaging cell line PLAT-gp expressing only the gag-pol gene is available for generating pseudotype retroviruses with differing envelop proteins such as VSV-G.
New series of retrovirus vectors that avoid expression of Gag and Gag-fusion proteins
A variety of retrovirus vectors are available for many different purposes. Comparison of these vectors is not the purpose of this review. Here we introduce mainly our retrovirus vectors designed for expression cloning and efficient gene transfer. We have been using the pMX vector [
]. The pMX vector harbors 5′ long terminal repeat (LTR) and the extended packaging signal derived from MFG followed by a multi-cloning site (MSC) suitable for cDNA library construction and 3′ LTR of MMLV. The resulting vector pMX in combination with PLAT-E cells produces, on average, 1×107 IU/mL. With high-titer retroviruses, one can efficiently introduce the gene of interest into most mouse cell lines as well as primary cultured cells such as T cells, mast cells, and neuronal cells with an infection efficiency ranging from 20 to 60% [
]. To further increase infection efficiency, either Retronectin (Takara, Kyoto, Japan) or concentration of retroviruses by medium-speed centrifuge (8000g at 4° C for 16 hours) can be applied, and the infection efficiencies in these cells can reach 80% [
Unlike most other retrovirus vectors such as LXSN and MSCV, the pMX vector harbors the splicing donor and acceptor sites and produces two types of transcripts, the full-length genome RNA and the subgenomic RNA (Fig. 4A). Therefore, in addition to the correct proteins produced from the subgenomic RNA, Gag-fusion proteins could be expressed from the full-length genome RNA if the reading frame of gag and that of the inserted gene match and there is no stop codon between them. To generate improved vectors free of the Gag and the Gag-fusion proteins, we disrupted the ATG start codon of gag, inserted a stop codon downstream of the CTG start codon of glyco gag, and inserted triple stop codons in three different reading frames just before MCS. The improved vectors are termed pMXs. A variety of pMXs series are depicted in Fig. 4B, including pMXs-puro, pMXs-neo, pMXs-IG (IRES-GFP), pMXs-IN (IRES-neo), and pMXs-IP (IRES-puro).
MMLV-based vectors usually are silenced in immature cells, including embryonic carcinoma (EC) cells and embryonic stem (ES) cells, and possibly hematopoietic stem cells. Indeed, pMX vectors are quickly silenced in EC and ES cells. MESV and PCMV are mutants of MMLV and can stably express genes in immature cells [
]. These include FMEV (a hybrid between FMCF and MESV) and MPEV (a hybrid between MPSV and MESV) vectors, which can express genes in EC and ES cells. These vectors were modified by Hawley's group in Canada [
], and the resulting vector was called MSCV (mouse stem cell virus). However, these vectors will not give optimum titers in transient packaging because the U3 region of 5′ LTR lacks one of the 75-bp direct repeats in the enhancer region. In the transient transfection, the promoter of 5′ LTR drives expression of transfected vectors, thus determining the retrovirus titers. To increase the promoter activity in transient transfection, we replaced the U3 region of 5′ LTR of MPEV and FMEV with that of MMLV using the KpnI restriction site (Fig. 4C) and inserted the pMXs-derived extended packaging signal and MCS before 3′ LTR. We named these vectors pMYs and pMZs. The pMYs vector is based on MPEV, and its 3′ LTR consists of U3 of MPSV and U5 of MMLV. The pMZs vector is based on FMEV, and its 3′ LTR consists of U3 of SFFV and U5 of MMLV. As expected, pMYs and pMZs vectors produced high-titer retroviruses in transient transfection, and these viruses can efficiently express genes in EC cells, most ES cells, and hemopoietic progenitor cells.
We also have developed pMCs vector that uses PCMV LTRs for both 5′ and 3′ ends like MSCV (T. Mizutani, H. Iba, and T. Kitamura, unpublished data). The U3 region of 5′ LTR has been replaced by MuLV LTR to increase the efficiency of transient transfection. Thus, the pMCs vector produces proviruses similar to those produced by MSCV. However, pMCs is supposed to produce higher titers of retroviruses because of its higher efficiency in transient transfection and is expected to produce higher amounts of proteins because of the presence of the splice donor and acceptor sites in pMCs that was derived from MFG vector and is present in pMXs, pMYs, and pMZs.
In the present review, we describe the efficient retrovirus expression system and its applications in a variety of experiments. Although we mainly introduced our works, there are almost infinite other potential applications for retrovirus-mediated expression screening. For example, short peptide libraries can be screened using the retrovirus expression system [
]. It would be useful to perform two-hybrid screening in mammalian cells, in which molecular interactions will be more physiologic. In the past, insertional mutagenesis using retrovirus vectors was performed extensively; however, it was difficult to identify the responsive gene induced by the retrovirus because retrovirus integration could alter transcriptional activities of the genes within 20 to 30 kbp from the integration site. Recent genome information and expression profiling using gene chips and DNA microarrays have made it much easier to identify the responsive gene in insertional mutagenesis. Recently, even in vivo mammalian genetic screenings have been performed successfully using retrovirus-mediated insertional mutagenesis in the mouse [
Most of the human genomes have now been sequenced, and about 30,000 genes have been identified. However, there is still much to learn about the human genome. We believe that retrovirus-mediated gene transfer and expression cloning will continue to be important tools to understanding the genome. As we introduced some examples in the present review, isolation of proteins based on their functions sometimes opens up insights into cell biology. In addition to expression cloning, retrovirus-mediated gene transfer also is useful for investigating gene functions and can be applied in high-throughput analysis of gene product in a variety of cells. We sincerely hope that this review will invite more researchers to retrovirus-based technologies.
This work was supported by grants from the Ministry of Education, Science, Sports and Culture and the Ministry of Health and Welfare, Japan. The division of Hematopoietic Factors is supported by Chugai Pharmaceutical Co., Ltd. The authors thank Drs. Carol Stocking and Wolfram Ostertag for MPEV and FMEV vectors and valuable discussions, Drs. Hirofumi Hamada and Hideo Iba for the valuable discussions, and M. Ohara for excellent language assistance.