ELISA development

12 posts / 0 new
Last post
newcellbiologist
newcellbiologist's picture
ELISA development

 I have a positive control worked out for my ELISA now, so I know it's working. 

Now my issue is: human serum.

I've tried switching polyclonal and monoclonal as capture/detection - each time the detection is biotinylated and streptavidin-HRP is added later.

My basic assay is:

1. coat wells with capture Ab (10ug/mL) in PBS overnight in 4 degrees C
2. aspirate/wash three times with wash buffer
3. Block non-specific binding via 1% BSA solution, at least 1-2 hours
4. Wash
5. Add samples - positive control conditional media and serum samples in 1% BSA reagent diluent, incubate      2 hours
6. wash
7. Add 0.5ug/mL biotinylated detection (mAb if capture is polyclonal; polyclonal if mAb is capture), 2 hours
8. Wash
9. Add streptavidin-HRP, 20 minutes
10. wash
11. add 1:1 solution A:solution B (R&D systems) substrate reagent
12. once desired blue intensity is reached (aprrox 10 minutes), stop with 2N sulfuric acid (turns yellow)
13. measure the optical density at 450nm wavelength

any good ideas on how to optimize this system to detect (if it's there) an antigen of small size and possibly low concentration.

should I pretreat the human serum in some way? Any new methods you all know about?
Thanks!

- Newcellbiologist

g a
g a's picture
hi newcellbiologist

hi newcellbiologist

If you mean by small antigen.... smaller in size (MW) or you have raised the antiserum using a peptide, then couple your antigen to a different carrier protein tha the one you used to couple for immunization. For e.g If you have raise antibodies by conjugating Peptide to McKLH then use albumin for detection system.

and In other case if your antigen abundance is low then  in that case you have to increae yr source protein load to get the significant amounts of the same that can bind to the antibodies. You can enrich the protein as well using fractionation methods.

Sami Tuomivaara
Sami Tuomivaara's picture
newcellbiologist,

newcellbiologist,

Do first a dilution assay to see how low concentration of your small molecule you can detect in your positive control experiment... If that compares well to expected amounts in your sample, go ahead with your experiment. Since you are using serum as your sample, use lot of washes since serum contains biotin and potentially cross-reactive antibodies that can screw up your assay... Also, serum is a complex system with all kinds of carrier and other proteins binding to small molecules (whether metabolites or drugs or...) and potentially masking them from ELISA detection. Note also that you cannot say for sure whether your molecule "is there" or not by _any_ analytical technique, but you can come up with some lower limit based on what is the lowest concentration you can detect in your positive control. So just be cautious in evaluating the results of your experiment. I don't see any harm in increasing the incubation time of streptavidin-hrp to 1 hr or more.

I'd be interested in hearing comments about the pretreatment as well, filtering is of course a must step.

Cheers,

samm
samm's picture
Hi! I just want to add a

Hi! I just want to add a couple of points to Suola's post above (Arginine seems to have missed the point with his last sentence).
a) use 1% normal serum (human+spp of your secondary (detection) ab e.g. rabbit polyclonal or mouse mAb) to improve specificity and to address the very valid point Suola makes
b) do not filter (or "enrich the protein as well using fractionation methods.") - you'll always lose some analyte, and will have no idea about extent of loss/sample. We usually measure cytokines and small proteins such as HMGs from serum, and the best thing to do with them is dilute samples to atleast 1:10 - probably not possible for you. However, we have used serum at 1:2, and it works fine. If you have enough sample/wells, use 2 dilutions of serum that fall within the range of your standard (e.g. 1:2 and 1:4).
As Suola states, first do a dilution assay to see how low concentration of your small molecule you can detect in your positive control experiment - that will give you a working range.

Sami Tuomivaara
Sami Tuomivaara's picture
I stand corrected, thanks

I stand corrected, thanks samm... instead of filtering the sample, maybe few minutes spin with table centrifuge is in place...

Cheers

g a
g a's picture
Hi

Hi

I agree with Suola that first you determine the titre for the antibody raised and then compare it with the amount that you expect in the serum.

I did the same thing for a cellular protein not present in plasma, antibody was working with cellular extracts but not with plasma even though we identified the protein in plasma. So we did fractionated the plasma and removed albumin enriching the sample by almost 7-8 folds and Bingo....... It worked.

plasma    -ive

albumin rich fraction    -ive

Albumin depleted fraction +ive.

So I still insist if you are working with something similar then go for an enrichment step.

newcellbiologist
newcellbiologist's picture
 Thank you for all of your

 Thank you for all of your comments! I will keep working on the troubleshooting and post the outcome asap.

newcellbiologist
newcellbiologist's picture
 So, I'm having no luck with

 So, I'm having no luck with getting a signal in my ELISA assay testing human serum samples (in 1% BSA as a reagent diluent).

My options are:

1. try different samples, because maybe the protein I'm looking for is not even present or detectable in human serum. Maybe try PLASMA (as opposed to serum) vitreous fluid or saliva. My goal is to detect the extracellular part of a receptor to see if it exists in a soluble for in these extracellular spaces.

2. decrease the capture and detection antibody concentrations to lower the background produced by my 1%BSA alone, so that relative decrease in background may show my serum has a greater signal. Right now the serum's signal is lower than the reagent diluent control. My positive control for the antigen is conditional media containing the antigen.

3. Try different blocking methods. Instead of 1% BSA for both blocking at reagent diluent, maybe I should try Tween 20 or Triton X?

Any more suggestions would be appreciated? Do you think centrifuging the serum and taking certain portions of the solutions would help?

newcellbiologist
newcellbiologist's picture
 Oh yeah, I forgot to say

 Oh yeah, I forgot to say that I tried to pre-treat the serum with 8M urea for 1 hour on ice, but it made no difference.

samm
samm's picture
Can you see your protein in a

Can you see your protein in a Western? Perhaps an initial salting out all proteins with ammonium salts to concentrate, then dialyze and load.
Try loading these samples and a positive control - atleast your polyclonal should work (mAbs may be solely conformation specific)

Sami Tuomivaara
Sami Tuomivaara's picture
newcellbiologist,

newcellbiologist,

I think option 2 is useless... by lowering the antibody concentrations, you also lose sensitivity of your assay. In fact, increasing the Ab concentrations might help.

You can also try other way of blocking, since bsa is serum component, there can be some interactions that are hard to predict. I've only used BSA or non-fat milk powder, so I can't comment on Tween 20 or Triton X, but you could try milk powder buffer (1% blocking, 0.1% sample). Also, you can try to leave out the 0.1% BSA from the sample. I did some ELISAs where I left it out because I didn't have any, and ELISA turned out fine.

I wonder if there's some component in the serum that is _inhibiting_ your assay... maybe biotin that blocks your biotinylated secAb from binding.

Samm's idea of Western might work too! But, I'd choose carefully the method of concentrating the serum. Rather than losing protein in incomplete salting out and on dialysis membrane, why don't you try spin-concentrator... Uses less steps and less time...

Cheers,

samm
samm's picture
Suola, I didn't mention the

Suola, I didn't mention the CentriCon/Microcon thingies because of size issues - haven't had much luck with small prtns (~<14 kD cytokines and such). Larger proteins should be fine - and yes, ~5-30 mins is better than ~o/n minimum for salting out/dialysis! Thanks for helping newcellbiologist in this matter.

Also, newcellbiologist, note that your assay with any of the above will be more qualitative than quantitative - but before you try to figure out why your ELISA is not working, you must figure out whether the analyte is there at all in your samples (esp if your Ab pairs and positive controls/stds for ELISA work fine).

Finally, about the blocking point that Suola makes, you can perhaps add avidin during sample incubation (and wash very well). Also, in addition to NF milk powder/low Ig BSA, I vaguely remember someone in my old department using gelatin for a hormone ELISA, but that stuff is a PAIN to use.