Retroviruses carry their genetic material in the form of mRNA, which must be reverse transcribed into DNA and incorporated into the host genome for expression and viral replication. Retroviral infections take advantage of this system by using a host cell to create replication-deficient virus, which may be then used to infect a different set of cells. Every cell that is infected should theoretically permanently express the viral proteins, as genomic incorporation is part of the virus life cycle.
We have several retroviral vectors in the lab: pBabe-puro, pBabe-GFP, SRa, and pMV7. The pBabe vectors and the pMV7 vectors are based on the Moloney Murine Leukemia Virus (MoMuLV), and SRa is based on the Murine Sarcoma Virus (MSV). The main difference in these vector lies in the LTR (long terminal repeats), which determine integration and act as a promoter for RNA transcription. From conversations with people in David Ron's lab and Ed Skolnik's lab, the pBabe vectors seem to be the most reliable. The SRa is also reportedly a very good vector, but I haven't used it as the pBabe has more useful cloning sites and a better map. pMV7 was obtained from Tony Brown's lab, and my experience with it has not been great. I have not been able to see expression of any construct in this vector, and it is rather large (8kb). In addition, I have observed some evidence that it is prone to rearrangements.
For the pBabe vectors, the insert is cloned into the polylinker downstream of the 5' LTR with the LTR acting as the promoter. The marker (either puromycin resistance or green fluorescent protein) is 3' to the cloning site and is driven off a SV40 promoter. This is followed by the 3' LTR. The origin of replication and amp marker are outside of the LTR's. As this vector doesn't contain any viral proteins, it is not replication competent.
Virus may be made in one of 2 ways. The pBabe vector may be transiently transfected into packaging cells permanently transfected with "helper virus", which provides the necessary viral components for creating infection competent virus (we use Bosc23 cells for this). Or pBabe and an "Ecotropic Envelope Virus" (EEV) vector can be transiently co-transfected into 293T cells. Either method provides ecotropic virus (for infecting rodent cells). The method to produce amphotropic vector (for primate or non-rodent cells) requires many more safety precautions, and we do not use this system in the lab. Helper viruses for this purpose might be available in Litman's lab.
Ed Skolnik's lab has actually combined the two methods such that they transiently co-transfect pBabe and the EEV into Bosc23 cells. This gives a very high virus titer, and this is my preferred method of producing virus. Bosc23 cells are actually a derivative of 293T cells and may be grown and transfected just like the 293Ts.
Plate the Bosc23 cells at 1:10 (approx. 9X105 cells per 100 mm plate) 24 hr before the transfection. Transiently transfect 15 ug of each vector per 100 mm plate overnight, using the standard CaPO3 method. We usually assay for protein expression app. 48 hr following transfection (by Western blotting of whole cell lysates), however, virus production is maximal after 72 hrs. To collect virus, replace the media with 5 ml of fresh DMEM (complete) and collect about every 5 hours for 2 days, starting at approx 60 to 72 hours post-transfection. For overnight, use 7 ml of media. This produces a relatively concentrated stock of virus. You should collect often as the virus is not very stable at 37û. Collected virus should be kept at 4û until the finished, as freezing and thawing also reduces the virus titer. Before infecting or freeezing, spin the media at the highest speed possible in the tabletop centrifuge to remove cells and debris, and then filter through a .45 m syringe filter -- this removes all extra cells and debris. Do not use a .22 m filter -- you will greatly reduce the viral titer, even though the Millipore sales rep claims that the virus should pass through a .22 filter easily. Any extra virus can be frozen in sterile tubes at -70û indefinitely.
The cells to be infected should usually be split at least 24 hr before infection. The best degree of confluency will have to be determined for each cell type. For example, C57MG cells are very adherent, rather hardy, grow very fast and infect well. Therefore, I split these at about 20% confluency before infecting so that they do not overgrow the plate before time to split into selective media. In contrast, PC12's grow more slowly, come off the plate easily, and infect poorly. Furthermore, I observe some cell death during the infection. These cells perform much better if infected at app. 50% confluency. Generally, I infect a 60 mm plate of cells, as this provides a decent number of cells for growing up clones and creating pools of selection resistant cells.
Replace the cell media with 1 ml of viral supernatant plus enough growth media to cover the cells (2 ml for a 60mm plate). The infection is much more efficient if you add polybrene to a concentration of 8 to 10 ug/ml. Polybrene is a cationic agent that facilitates "sticking" of virus to the cell membrane. However, this results in heavy cell death of PC12s, so I use polybrene at 3 ug/ml with this line. Without polybrene, they infect very, very poorly, so I suggest using it in all infections.
72 hr after infection, the cells should be split for selection. I split them at several different densities to compensate for variable virus titers, etc. For ring cloning, where the clones need to be well separated from each other, I generally, I put 2 plates at 1:200 and 3 at 1:2000 in 150mm plates. I split the rest of the cells between 2 X 60 mm plates with selection to create pools of un-isolated colonies. You may want to use one pool to freeze, in case you have to re-select later, and the other should be used to determine that your construct is indeed expressed in these cells from this infection. The media should be changed probably once a day the first 2 or 3 days, depending on how fast the selection occurs. After that, the cells take approximately 2 to 4 weeks (depending on cell type) to produce colonies that can be ring cloned. Depending on the infection efficiency, the pools may be ready to freeze/screen with a week or less. For pBabe, I use 2.5 ug/ml of puromycin, and I find that cell death occurs much more quickly than with G418 (generally all uninfected cells are dead within 3 days).
Construct Expression from pBabe: You can screen virus producing cells for construct expression. The viral vectors create mRNA, some of which is translated by the cells and some of which is packaged into virus. However, the best time after transfection to screen for protein expression is 48 hr., so you should do the transient transfection for each construct in 2 plates -- one for making virus, and one for screening for expression. The expression levels will probably be quite low in these cells -- not nearly as high as transient transfections using CMV based vectors.
Virus titer: This is not usually a problem, but if you are having problems getting colonies, you may wish to check you virus titer. For this, we use 3T3 cells -- just as 293Ts transfect well, 3T3s infect well with ecotropic vectors. There is a detailed method for accurately determining viral titer in Current Protocols in Molecular Biology, but if you just need to ascertain that you do indeed have virus, simply infect 3T3 cells as above and select. They can also be used to check expression of your construct, even without selection.
Selection: Include a plate of un-infected cells plated with the selective media to gauge the effectiveness of selection.
Construct expression in the cell line of choice: As mentioned above, I usually screen pools of cells from each infection, just to ensure that the construct is being expressed. Usually, I screen pools 3 to 7 days following the infection of C57MG cells, or a week to 2 weeks for PC12 cells.