Genotyping and characterization of PERK-/- cells

Genotyping and characterization of PERK-/- cells

Dear Colleague,

We have made every effort to insure that the PERK-/- cells you have received (or will receive) are of the correct genotype. However, we recognize that it is important for you to validate this. In the letter below we should like to make several points related to this issue:

The cells in question are SV40 large T-antigen transformed PERK-/- MEFs; we are currently unable to send primary MEFs. Please note that the immortalized cells have undergone lengthy adaptation to tissue culture and some phenotypic features such as the hypersensitivity to ER stress may have been attenuated. Nonetheless the knock-out cells have no PERK protein and are a useful loss-of-function model for those proximal signaling events that require this kinase and cannot be easily compensated for, by adaptation in function of other genes.

The cells we send out are portions of a large batch (lot) of cells that were frozen into individual vials on a single day. The QC on our end has included genotyping the cells by PCR for presence of the mutant allele and absence of the wildtype and analysis of PERK expression by immunoblot. However, we do not recommend genotyping the cells by immunoblot. PERK is a very low abundance protein which can be detected reliably only by IP-western in most cell types (including fibroblasts). We do not know of any commercial antibodies that detect both the non-phosphorylated and phosphorylated forms of the protein; however, we have successfully used the Cell Signaling Technologies anti-phospho-PERK antibody (cat#3191) to detect phosphorylated PERK by IP-western.

We have also noted that the pancreatic acinar cell line AR42J expresses ~10 times more PERK than any other cell line we have tested and only in these cells have we been able to reliably detect PERK by western of whole cell extracts. We suggest using this cell line as a positive control for studies in which the PERK protein is to be detected.

Please note that the predicted MW of PERK is 125kDa, but, in our hands, the protein runs between ~150kDa (non-phosphorylated) and 170kDa (hyper-phosphorylated), depending on the level of phosphorylation. There are many phosporylation sites in the PERK protein, therefore, the activated form often runs as a heterogenous population (a smear) in slightly stressed cells. We suggest comparing the size of the protein detectable by IP western from wildtype untreated cells and ER stress treated cells (e.g. 400nm thapsigargin treated 1hr, 2mM DTT treated 30 min-1hr, or 2.5ug/ml tunicamycin, for 4 hr). With thapsigargin or DTT treatment we typically see a 100% shift from the faster migrating to the slower migrating form. This should allow you to judge the quality of the antibodies you are using. Once you have validated the antibodies, we expect you should be able to confirm that the cells we sent you are mutant.

If you are using phospho-eIF2alpha as an indicator of the cell genotype, please note that there are 3 other eIF2alpha kinases (HRI, PKR, and GCN2) at least two of which are expressed in PERK-/- cells (PKR & GCN2). We do occasionally see increased eIF2alpha phosphorylation in PERK-/- cells exposed to severe ER stress (e.g. > 4 hours thapsigargin) and we believe that is most likely mediated by one of these other eIF2alpha kinases. Nevertheless we have consistently observed a severe to complete defect in eIF2alpha phosphorylation in PERK-/- cells in response to ER stress during the first 1- 4 hours of treatment.

Protocols for IP western detection of PERK are available on our web site (also download pdf).

You may also want to read our notes on the phenotype of the cells. Finally, please beware that PERK-/- and other ISR defective cells are susceptible to contamination with wildtype cells (that may rapidly overtake the mutant cells because of their growth advantage). However, this problem can be eliminated by selecting in G418 (250mcg/ml), as the PERK-/- cells are NEO resistant by virtue of the targeting cassette whereas any contaminating wildtype cells are NEO sensitive.

We hope these comments will prove useful to you.

Heather P. Harding

October 21, 2004

(updated, March 6, 2006 by David Ron)