|Home | About | Journals | Submit | Contact Us | Français|
Natural killer (NK) cell lines are difficult to transfect using standard techniques, which limits the ability to establish long-term knockdown of proteins with short hairpin (sh)RNAs. We have a developed a method to stably knockdown protein expression in human NK-like lines by introducing shRNAs in retroviral vectors. After a single transduction with retrovirus, shRNA-containing cells can be selected with drug treatment or sorted for enhanced green fluorescent protein (EGFP) expression. With this method, protein expression can be stably decreased to less than 10% of wildtype levels.
Primary NK cells are difficult to transfect with standard vectors under a variety of conditions proven successful in other cell types (2). These cells are also not generally amenable to viral or retroviral infection (1), although recent successes with lentivirus (see protocols by Kung and Savan/Young in this volume) are overcoming this technical hurdle. In contrast, NK-like cell lines are more permissive to gene transfer by transfection (3, 4), and especially by retroviral transduction (1), thereby allowing for overexpression and knockdown of genes of interest.
The technique of RNA interference (RNAi) has revolutionized modern cell biology by enabling researchers to selectively eliminate the expression of specific mRNAs. To achieve short-term knockdown of the target mRNA expression (and subsequent protein expression), double-stranded short interfering (si)RNAs of 21–23 nucleotides (nt) matching specific sequences in the target mRNA are introduced into cells by electroporation or lipofection. Alternatively, short hairpin (sh)RNAs can be expressed in cells (see below) and processed by the Dicer endonuclease complex to generate sustained expression of double-stranded 21–23 nt siRNAs, achieving long-term, stable knockdown of target mRNAs (5). The siRNAs hybridize with target mRNAs to tag them for degradation by the RNAi-induced silencing complex (RISC), which contains an endoribonuclease (6).
The stable expression of shRNAs in cell lines can be achieved by transfection with certain vectors or transduction with retroviral or lentiviral vectors. Constructs cloned into these vectors generally consist of a 60 nt oligo that, when expressed in cells, is processed to generate a 19-nt siRNA with uridine overhangs (Figure 1). Expression of the shRNA construct is driven by the polymerase III HI promoter, which produces a small RNA transcript, lacking a poly-A tail, and can be processed into a standard siRNA molecule.
Here, we describe a method for knocking down proteins of interest in NK-like cell lines by the co-expression of two different retroviral vector-based shRNAs. This protocol can be modified to include co-expression of up to three distinct shRNAs simultaneously in the same NK cell line.
There are numerous free programs available for designing shRNAs (Oligoengine, Dharmacon, etc). For this protocol, Oligoengine software was used, since the interface was user-friendly, GeneBank sequences could be uploaded directly and both the secondary structure and nucleotide usage were considered in the selection algorithm. We suggest designing a minimum of four shRNAs for each gene of interest, since we encountered an approximate 50% success rate. For knocking down SHP-2 phosphatase in the human NK-like cell line, KHYG-1, the expression of a single shRNA resulted in ~50% decrease in wildtype protein levels, while co-expression of two shRNAs decreased levels by >90% (7). Avoid designing shRNAs that target common domains within gene families (e.g. phosphatase domain), as these will be less specific for the gene of interest and could non-specifically suppress other members of the gene family. The final sequence should be tested for homology with other mRNAs with the BLAST program (http://blast.ncbi.nlm.nih.gov/Blast.cgi), and sequences with high homology to other known sequences should be abandoned. In addition, do not limit shRNAs to just one region of the mRNA. Finally, generate scrambled versions of each shRNA (having the same nucleotides as a shRNA targeting the gene of interest, in a randomly scrambled order) for use as controls. Scrambled control sequences should also not cross-react with known mRNAs in the targeted species, as assessed by a BLAST search. Alternatively, predesigned/pretested shRNAs of many genes cloned into retroviral vectors can be purchased from a commercial provider (e.g. Santa Cruz Biotechnology Inc, Santa Cruz, CA; Sigma-Aldrich, St. Louis, MO; Invitrogen, etc.).
The procedure for generating shRNA constructs was adapted from the Oligoengine pSuperior protocol. The protocol can be adapted for other specific vector and restriction enzyme combinations.
It is important for both the Phoenix-amphotropic and NK-like cell lines to be freshly passed a day prior to transfection and transduction, respectively. This protocol works well for KHYG-1, NKL, NK3.3 and NK-92 cell lines (7).
Selection scheme will depend upon which vectors (puromycin, neomycin or gfp) were used.
We would like to thank Drs. Lauren O’Donnell and S.M. Shahjahan Miah for helpful comments on the manuscript. Supported by National Institutes of Health grants R01-CA083859, R01-CA100226 (K.S.C.), T32-CA009035 (A.K.P.), and Centers of Research Excellence grant CA06927 (FCCC). The research was also supported in part by an appropriation from the Commonwealth of Pennsylvania. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute.
1Depending upon your selection scheme (selection with antibiotics versus GFP expression), it is important to choose the appropriate retroviral vector(s). With this system, one can express and select for up to three separate shRNAs at the same time (one shRNA in puromycin, neomycin and gfp vectors). We have had success using each shRNA-containing vector singly or all together.
2It is easiest to use a heat block for this step, turning it off after the 95°C incubation and allowing the block to cool to RT.
3It is recommended to have a 1:3 ratio of vector:insert for the ligation reaction. Some PCR and gel extraction kit elution buffers disrupt accurate determinations of the DNA concentration by absorbance spectroscopy. To avoid this issue, measure the concentration of the insert and digested vector on an agarose gel using a DNA ladder for the concentration control.
4To decrease the number of false positive colonies, digest the ligation reaction with BglII for 1 hr at 37°C. The BglII site is destroyed upon successful cloning of the shRNA pair, therefore vectors containing the shRNA insert will not be cut.
5To prevent recombination at retroviral LTRs, always use recombination-defective bacteria (e.g. Stbl2) when working with retroviral vectors, and always culture bacteria at 30°C.
6Grow colonies for longer times to compensate for the reduced growth temperature.
7False positives that stem from re-ligation of empty vector will have a ~1 kb band in the Oligoengine system.
8Phoenix cells should be ~80% confluent for optimal transfection. Never allow the cells to reach full confluence or transduction efficiency will suffer.
9Several varieties of the Phoenix packaging cells are available (e.g. amphotropic, ecotropic, polytropic), which differ primarily in the expression of the viral envelope proteins that mediate viral entry into target cells (8). Use Phoenix-amphotropic cells when generating virus for transduction of human cells, Phoenix-ecotropic for murine cells.
10Phoenix cells are semi-adherent and can easily detach from the plate. Exercise caution when manipulating the cells or when changing the medium. Add medium very slowly to the side of the 6-well plate, keeping the pipet tip horizontal to the plane of the cells. Never add medium directly on top of the cells.
11Although the retrovirus is replication defective and should be non-infectious, be sure to UV-treat all contaminated glass and plasticware for at least 1 hr before discarding. The use of retroviral technology requires standard class BSL2 biohazard safety precautions and approval by the local biohazard safety committee in your institution.
12Retrovirus can be used immediately or can be stored at −80°C for up to 4 months with an estimated potency loss of ~50%. We have had success using virus that was frozen to co-transduce KHYG-1 cells with two vectors simultaneously when subsequently selected in medium containing the appropriate combination of antibiotics.
13To increase viability, let cells detach from the 12-well plate overnight instead of forcing cell detachment with pipeting.
14Transduction efficiency varies widely among known NK-like cells lines. For example, NK-92 cells had ~4% transduction efficiency with this protocol, while KHYG-1 had almost 30% efficiency for single vector transductions.
15To improve drug selection, expand cells to multiple wells if overgrown.
16Determine the optimal concentration of drug needed to kill each NK-like cell line being used. For transduction of KHYG-1, NK-92 and NKL cells, puromycin was used at 2.5 µg/ml and G418 at 1.25 mg/ml.
17To ensure proper drug selection, include non-transduced cells during selection.
18Puromycin-induced death was observed by days 1–2 of drug treatment and days 3–5 for neomycin.