I agree with Daniel on all the points he has made. I would also consider the quality of your reagents (are they freshly or recently prepared and stored as recommended?). Another big factor is the level of luciferase signal you are getting. This will depend somewhat on your system for measuring the luciferase, but if your values are extremely low, the variation you are seeing is likely random variability and your luciferase detection is below the limit of detection. Are you sure there is luciferase signal to detect in your system? A positive control such as 293 cells transfected with a strong luciferase reporter could be useful to rule out technical problems rather than problems with your luciferase system in the cells you are interested in.

What type of luminometer are you using? Are the samples and buffers dispensed by a machine or by hand?

I am using Turner Biosystems Luminometer. Samples and buffers are dispensed manually. I have tried to be as precise as possible.

The time delay between pipetting and measurement can cause major variation in luminescence readings. Try to make this as uniform as possible, or better yet, arrange for access to a machine that automatically dispenses your luminescence reagents.

High variability may be seen in poorly transfected cells. Where do you see variability in the activity of  firefly or renilla lucifrerase? 

Variation across biological duplicates is expected, but variability in technical replicates on the same lysate is not.

I agree with Daniel on all the points he has made. I would also consider the quality of your reagents (are they freshly or recently prepared and stored as recommended?). Another big factor is the level of luciferase signal you are getting. This will depend somewhat on your system for measuring the luciferase, but if your values are extremely low, the variation you are seeing is likely random variability and your luciferase detection is below the limit of detection. Are you sure there is luciferase signal to detect in your system? A positive control such as 293 cells transfected with a strong luciferase reporter could be useful to rule out technical problems rather than problems with your luciferase system in the cells you are interested in.

Thanks for suggestions. I always perform the analysis immediately after treatment with Passive lysis buffer and take the measurements immediately after adding the substraate and then the stop reagent without any delay for all samples.

The firefly luciferase activity is high and remains stable however it is the renilla luciferase signal that differs quite significantly in the duplicate or the repeated samples. One thing I have noted is that the ratios are quite reliable when I perform all the required steps on ice, though, as expected the signal intensity is much lower. When I bring everything (substrates and glo) to room temperature signal intensity is very high but the ratios differ a lot. I run the positive and negative control for quality control for each set. Other thing i would like to mention is that I directly take the lysates from the 12 well plates while measuring the luciferase activity,that is, I don,t separate the lysates from the cells. Is this responsible for such variance?

Avishekh, it seems like everything is okay with  transfection and measurement of luminescence if the activity of firefly is high and stable. You may want to increase the concentration of renilla luciferase during co-transfection, e.g from 50:1 (firefly:renilla)  to 10:1, 5:1 etc.

You're right about ice:) Should be no difference if you analyzing supernatant without separation it from cell debris.  However, if cell lysate contains high amount of floating cell fragments it is better to clear lysates.  

If you are using lysates straight from the 12 well plate, it could be an issue of how homogenized your sample is. I usually transfer to an eppie tube first, then aliquot my samples for the luciferase assay. Another factor may be that there is residual luminescence from the firefly reaction that is not being quenched. The protocol mentions that firefly is quenched 100,000-fold, thus not being a major factor, but might arise if samples are not mixed adequately prior to measurement. Since you add this manually, be sure to add straight to the bottom of the luminometer tube (I am assuming this system uses these?) and I usually flick the tube a couple times to mix. And as others have mentioned, keep your timing consistent.

Thank you Sergey and Jessica for you valuable suggestions. Next time, I will take all your advices in account.

I agree with Jessica that cell debris could be an issue, especially if improperly homogenized or incompletely lysed. If you aren't processing a ton of samples, you might consider assaying firefly and Renilla luciferase activity separately. I've found the assay to more consistent with h-Coelenterazine as the Renilla substrate than the LARII system, and it also has the added benefit of being cheaper per assay.

Thanks Daniel, if the problem persists I think I should switch to h-Coelenterazine.

Ive done a ton of these, here are my suggestions.

1.) Freeze lysates in liquid nitrogen. You must freeze samples, the reason is the luciferase lysis buffer does NOT lyse all the cells at RT, just about 50% of them cells. You can check this by lysing cells at RT and looking at them with Trypan Blue.

2.) Thaw them out, spin down full speed to pellet debris.

3.) The volumes you are using appear to be OK. For me, if I looked at the same sample but in duplicates, I got nearly identical values for Luc and Ren. Between samples from different transfections I got a coefficient of variation often as high as 20%. You really need to do the transfection 5 x to get solid averages.

In my experience, freezing the samples is neither necessary nor beneficial. See section 1C of the Promega manual (https://www.promega.com/~/media/files/resources/protocols/technical%20manuals/0/dual-luciferase%20reporter%201000%20assay%20system%20protocol.pdf). Quote excerpted below (emphasis added):

"Passive Lysis Buffer (PLB) is specifically formulated to promote rapid lysis

of cultured mammalian cells without the need for scraping adherent cells or

performing freeze-thaw cycles (active lysis)."

I think older protocols recommended this freeze-thaw step, but in general, incubating in PLB for longer periods (up to 15') works just as well if not better. There may be rare exceptions involving cell types that are particularly refractory to lysis, but I've yet to observe it, nor see this behavior reported in the literature. 

Thank you Daniel for your valuable suggestions,Today I performed the Dual luciferase assay, but, I separated the lysates into new epp tubes. Interstingly, it worked, I have datas with very low standard deviation, it seems to be working and now I have regained the lost confidence.

Thanks for your continued support since the day 1 of this post

Has anyone tried assaying for luciferase activity on previously stored but not lysed cell pellets? The cells were scraped in PBS and spun down and the pellets were stored at -80. If the pellets are sonicated in PLB buffer, do you still expect to detect luciferase activity?

Avishekh Gautam , I got the similar experience with you when I perform DLR assay. So, have you succeeded to stabilize the DLR assay to get the low variation? Please share your experience. What is your other advice to reduce this high variation?

Thank you