Dear Omar:

I am sorry to ask a silly question, but what about the concentration of your protein in the sample? Although ELISA is a very sensitive technique, the low amount of your protein may be another issue to take into account.

Which is your detection reagent? Have you tried out chemoluminiscense?

Which is the specie of your mAb? Is it the same as in your biotinilated-pAb? If not, you may consider a two-step detection (a specific pAb first, and then a biotinilated pAb againts this reagent); this may increase your background signal, though.

Good luck!

Dear Matías,

Thanks a lot for your suggestions. I am using a range from 1000 to 7,5 pg/ml in a two-fold serial dilution, so I believe it should more that enough in terms od sensitivity, am I wrong? The reagent is TMB and I am using different species for the different antibodies, just in order to obtain as less crossreactivity as I can.

Thanks for your advice!!

Dear Omar:

The amount of protein you are employing should be enough to be detected by ELISA.

I hope that you find the solution you are looking for.

Good luck!

Dear Omar, Is it mouse monoclonal antibodies? If so, try to coat wells with RAMIg overnight in 4°C, than wash wells with PBS and than coat wells with your monoclonal antibodies (2h, room temperature).

Dear Omar,

I had a good experience using the ELAST amplification system (PerkinElmer) to get a higher signal/noise-ratio. For reference take a look at our paper:

https://www.researchgate.net/publication/250926614_A_simple_set_of_validation_steps_identifies_and_removes_false_results_in_a_sandwich_enzyme-linked_immunosorbent_assay_caused_by_anti-animal_IgG_antibodies_in_plasma_from_arthritis_patients

Good luck, Tue

Article Simple set of validation steps identifies and removes false ...

Is there a specific problem e.g. no signal at all or no difference between signal and noise? I would find the optimal coating conc. of pAb using bicarbonate buffer pH 9.6 to begin then run a checkerboard assay with protein and mAb from high to low conc,. What buffer are you using for mAb? If low signal then you could use a biotinylated mAb.

Dear Omar,

if you have purified material of your HIV protein, I suggest direct coating in order to check whether the capture and detection ab actually recognize your test sample. This would be my first trouble shooting measure.

Moreover, ensure that your blocking solution does not contain traces of biotin. As you may know, this is a typical pitfall when working with biotinylated detection reagents

All the best,

Bodo.

Hi, how long is your peptide and what conformation/sequence? It may be that the monoclonal can't bind properly - are you getting any signal at all?

Dear Omar,

Do you use commercial antibodies, or they are of your own production? Whether these antibodies used in the sandwich before your current work?

Dear Omar, what type of streptavidin-HRP reagent are you using, to bind to the biotonylated antibiody and to react with the TMB? Streptavidin-poly-HRP may increase sensitivity. It should be used with a special dilution buffer, see e g http://www.piercenet.com/browse.cfm?fldID=08020404

Best regards,

Maria

Dear Omar, the protocols for binding and running the test seem to be okay. But are you sure that both of your antibodies, capture and tracer, recognize your protein? In order to find out, I would immobilize the protein on the plate and test the antibodies or, if you don't have the protein in purified form, do a detection of the separated proteins by Western blot. When you are sure that both antibodies work in this way, you can go back to the sandwich format and optimize on the conditions.

As you can read above, there is a consensus that you should test each antibody independently to confirm that it will in fact engage the antigen. Also, are the antibody epitopes well separated to ensure that you are not competing for binding? One last thing, you might try the ELISA without Tween in your wash buffer as some mabs are too sensitive and will be washed away. Good luck!

You could try using different plates eg greiner high binding plates or if your monoclonal was produced in mouse you can buy anti mouse strips eg Delfia anti mouse strips

Omar, What is your sample. If you are trying to detect HIV-1 directly from blood you may have considerable interference from pre-existing antibody. Especially if you are trying to detect p24 or GP120. Maybe the reference below can help

Development of an HIV-1 nucleocapsid protein capture assay.

Goebel PB; Drummond J; Waters D; Bess J; Sowder R; Henderson L; Arthur

December 30, 1995

Natl Conf Hum Retroviruses Relat Infect (1st). 1993 Dec 12-16;:110.

Long stretch of protein sequences sometimes leads masking of the epitopes hence shorter peptides should be considered, bicarbonate wash is preferred. Moreover the concentration and purity of the protein is also to be looked into for proper functioning of the assay.

First of all, the MAb, PAb should be tested for the specificity against your interest HIV-1 protein. For ELISA, I would suggest to incubate the coating at 4degreeC overnight for even and better monolayer coating. For optimisation, you'll need to perform checkerboard for the correct ratio of MAb and PAb, as well as the blocking percentage, incubation time. Every optimization can have only one parameter to be tested once optimized, move on to the next parameter and fixed the first parameter.

Do you have a positive control? If you are not detecting the protein in your samples (dont know what they are) and your positive control is giving correct signals, this means that your experiment is working and it is possible that you are looking for something that is not there??? What do you think???

You can get kits for detection of HIV Protein. I found this link on the net for this ELISA...it seems simple enough..you can try it out..

I would use the polyclonal ab to capture the antigen, Carbonate buffer and PBS buffer not more than one week old. You can biotinlayte the monoclonal Ab, You can purchase kits quite easily. Are your Abs work in other assays? western blot?

Linda Steward have reason. Perhaps your plate not is the better. Revised the kind of Maxisorb used. you know the characteristics of the protein, hydrophobicity, hidrofilicidade, isoelectric point and others. These can to help you choose a plate that fits those characteristics.

Thanks a lot for your help!!! The protein I am trying to detect is p55, so its wight is 55 kDa. I have tried the ELAST kit but no significant differences were found. In addition, the main problem is that no differences in absorbance were found between background and the controls. All the antibodies used are commercially available.

I am really happy for your interest.

Thanks!!!

Test your Abs to see if they can bind to the Ag in question. Do not rust companies always check for yourself.

Up to date, I have not used jet any clinical samples, which is the final aim, I am just using different concentrations of p55 complete protein. In terms of the PAb, I am using an already biotinilated Ab, in order to avoid any option of biotin interference.

When you test your antibody with purified protein, as above mentioned, I suggest you take care very carefully of the pH of your carbonate or carbonate/bicarbonate buffer and try with different pH levels.

If the MoAb is mixed with e.g. fetal calf serum, the proteins of added serum can inhibit binding of MoAb to the plate.

As said above it could be that your Abs aren't binding or you need to increase their concentration. It could also be your buffer blocking too much if ODs are low.

First make sure that your antibodies didn't recognize the same epitope. (Label on of the antibodies with biotin, bind your p55 to the solid phase microtiter plate [2 µg/mL in 0,1 mol/L carbonate buffer pH 9.6 overnight at 4 °C], wash the plate 3 times, incubate them with a mixture of both antibodies of choice, one biotinylated on unbiotinylated. Start at 1 µg/mL (in PBS 0.3 mol/L NaCl, 0.05% v/v Tween 20. dilute one antibody leave the other constant and v.v.).

Don't take care on the microtiter brand. so called high binding plates (maxisorb, ...) creating only background. The detection limit of your later assay will depend on the affinity of the chosen antibody. The antibody at the solid phase should be the one with the highest affinity.

Avoid incubations at higher (or lower) temperatures than room temperature. This will create edge effects. High affinity antibodies will work also at room temperature.

You biotinylated anti-species antibody will react with both monoclonals, the one at the solid phase and the detection antibody. You will have always better results for direct labeling, e.i. biotinylation of the detection antibody followed by streptavidin-HRP-reaction.

Some other suggestions:

You could try to bind staphylococcal Protein A to the plate first to capture the mAbs at the Fc end. Protein A has been used to make affinity chromatography columns. I would bind the capture Abs overnight at 4–8 °C in all circumstances. Do not forget that maximal Ab saturation of the wells may not result in maximal binding of the target because of steric hindrance.

Are you certain the p55 complete protein has the same conformation as it does in the complete virion? I have tried mAbs developed to isolated Ags that bound the whole cell weakly or not at all.

I recommend blocking with BLOCK ACE (purified casein) because the smaller proteins seem to fill in the gaps better than do milk or BSA.

Hows many washes do you perform in each washing step? Some kits recommend four washes, but sometimes three give better results. I assume you're using a plate washer.

Is your primary detection Ab biotinylated, or are you using a biotinylated secondary Ab? I got better results with a biotinylated 2° Ab when I doubled the recommended concentration of that Ab.

Stephan Kiessig has some great advice. You definitly need to put the p55 directly on the plate and test your antibodies. However, you should know that p24 does not stick well to any microtiter plate that I am aware of. I am not sure how p55 will fare.

As for the plates, until you get a decent signal I would stick with generic microtiter plates. I agree with Stephan that high binding plates tend to give high background. Better to have a decent assay that you can improve then kill yourself trying more esoteric things first.

Yes I also think you should increase the concentration of your coating antibody. I have tried 2-5ug/ml and they work well. You also mentioned that no differences were found between background and the controls. If I understand you correctly, there could be non-specific reactivity. Try adding a blocking step with a protein of non-interest to your purpose such as skimmed milk, Bovine serum albumin (BSA), Casein or gelatin at concentrations of 1 to 5% after the coating step with 15 min to 1 hr incubation time. If the non-specific reactivities are still high, try adding 0.01 to 1% of your blocking agent to the washing buffer for the washing steps. In a worse scenario, you may also add 0.01 to 0.1% of the blocking agent to the diluent for your test samples and secondary antibodies.

Hi Irene,

Forget all the blocking staff. The capacity of 1 cm² polystyrole is at about 100 - 400 pg protein binding. Therefore higher concentrations are only required for test optimization. Protein binding to those surfaces is realized via hydrophobic interactions. The styrole ring will interact with the hydrophobic amino acids in your protein (Tyr, Tre, Ala, ...). If this binding is once established (and thats done with a very high affinity), you can destroy your protein only to get ths from the surface removed. Other proteins (as used in blocking solutions) are able to compete this hydrophobic interactions. Therefore a washing step using any neutral buffer (the cheapest is PBS) with some (non-ionic) detergent (mostly used Tween 20, 40, 60, 80) at 0.1% v/v will sufficiently block any free binding sites at the polystyrole. Blocking is only required in commercial assays which should have a very long shelf live at room temperature. For this purpose you need also some sugars to make sure that there is still a sufficient amount of water.

What are the blocking proteins are doing? They build a second layer on your primary coat. Until you didn't know anything about the potential interactions between your coating protein and the blocking protein or the blocking protein an proteins from the sample ... be careful. Just a simple example: allergies against cow milk are frequently present in the population. Some of them will have also IgG-Antibodies against the milk proteins you are used for blocking. Rheumatic patients can have antibodies against collagen. Gelatine is denatured collagen. Do you all diagnoses of all your the patients? Thats why the detergent is the key.

With all due respect to Stephan Kiessig, the blocking proteins bind to unbound sites on the plate stratum or to molecules that have no specific binding. I have never seen blocking destroy signal. This cuts down tremendously on background. If there is a problem with anti-milk-protein IgG (which I understand; I couldn't drink cow's milk as a toddler), then use another small protein molecule. The point of blocking is to prevent nonspecific binding of Ag or Ab.

I am referring to the earlier recommendation of Wales Nematollahi who suggested that protein A should be used for coating, in order to capture the capture antibody. While this works well in affinity purification, it can't be used in antigen EIAs, as it will most likely also bind the tracer antibody and result in nonspecific signal, i. e. , high background.

Hi Omar. Sorry I'm coming rather late in this question. Seems to me that your first step is to confirm reactivity between your protein and the antibodies. An easy way is a fast dot blot, locating your target protein on paper and testing each antibody individually. If both the monoclonal and the polyclonal give a positive signal, I would use the monoclonal for coating and the polyclonal (biotin-labeled) for detection. The polyclonal is more likely to work, because of detecting several epitopes. The monoclonal may pose more problems, it may react with a conformational epitope that is lost in your test samples. This is specially important when testing glycosylated membrane proteins. You may try several forms of your target protein until you're sure the system is working. As Stephen Kiessing advised above, incubation at 37oC will give edge effects.

Hi Maria and everyone,

First of all, thanks for your answers. During this time, I have been perfoming several assays in order to test the reactivity of both antibodies with the protein, and the results were positive. Thus, I am a bit concern about whether the problems could be in the coating step. I am using the monoclonal antibody for coating and the polyclonal as secondary antibody. Any suggestions???

Thanks a lot!!!!

Hi Wales,

blocking is really not required in ELISA. Why:

1. you are coating with an huge excess of protein (normally at a concentration of 1 ... 10 µg/mL), The binding capacity of 1 cm² polystyrole is at about 400 ... 800 ng.

2. the affinity of the coating protein depends on the number of available hydrophobic binding sites on the surface of the protein. Therefore large proteins will bind faster and with a higher affinity than small ones (or even peptides). Thats why peptides should be constructed for ELISAs in a way to define the binding site differs from the epitope.

Large proteins (or an high excess of proteins can compete with the proteins at the solid phase, slowly but surely)

3. It's could be possible that there are free binding sites. These binding sites will be covered during the first washing with a detergent containing buffer. That's why plates should never be washed (with a detergent containing buffer) before storage. They should be washed the first time just before the tests starts.

4. Proteins, protein mixtures like milk proteins can interact with your insolubilized protein (uncontrolled). Competition was already discussed, double layer formation is often a reason for high variation coefficients and imprecisicion.

5. Protein mixtures have to be free on any protease activity. If you need really a blocking, use denatured proteins (like Gelatinepolysuccinat or Albumin). (It's not that easy to find some)

6. Blocking is required in commercial assays, but it's done for an completely other purpose: The tests should be stored for 12 ... 36 month at room temperature with the coated plate. If the coat is getting dry, the protein will denatured, so you need to make sure that some water (humidity) is available. That's done by adding defined sugars (1 ... 3 % w/v glocuse or mannose). Also the proteins used in this "blocking buffer" will bind some water and will keep the proteins structure "alive".

Until you can store your coated plates frozen (and this would never work in a commercial assay) you didn't need any blocking. That#s what I found in more than 400 ELISA I've validated for lab use and more than 50 assays for commercial use.

Omar, sounds like you have some progress in your question. In most of my 'home' assays, the coating has been optimal with a monoclonal at 1-2 ug/ml in carbonate buffer (pH 9.6). In few cases coating worked better in PBS (pH 7.4). We always used blocking, usually with albumin (Ig-free). Milk powder is not recommended because it contains some biotin. Never worked with gelatine, may be worth trying. Next is to try some 3 concentrations of your protein (high-low and something in between), and a good confirmed positive control. If the control works, then you can focus on why your protein doesn't show reactivity. On this track, if the monoclonal is binding your protein only weakly, you may need to try a different capturing antibody.

Hi, Stefan,

I just became aware of your reply. My experience with ELISA is that I always get a higher background if I don't block. Whether I add 1° or 2° Ab in PBS with block or whether I block after binding, I get better results with block. I've found this with validating novel capture (1°) Ab, using both commercial plates with capture Ab and capture plates we made ourselves.

1. My decrease in background has often been correlated with the amount of block I've used. For example, I've often gotten a better result with 5% BLOTTO than with 1% . Nevertheless, I do try to optimize the blocking if I have the time to do so. Regardless of the binding capacity of polystyrene, I use the block that gives me the clearest background.

2. You seem to be looking at direct or indirect assays vs. sandwich assays here. In such assays, I try to saturate the well with the Ag. However, if I am testing a novel detector 1° AB, I prefer to preform sandwich assays with a 2° Ab.

3. Can you supply a reference that has data supporting the idea that wash buffer (with detergent, of course) alone blocks binding sites? My washes have removed enough protein, including capture Ab, that I've had to cut down on the number of washes a few times to get better results. How many washes do you perform after each Ag or Ab step? Admittedly, one can say that the detector (1°) Ab was binding weakly, but I still got a larger difference between positives and background when I removed a wash or two. If the detergent were binding to open sites, I should have seen a much higher background after reducing washes, but I never have seen that.

4. I prefer to use BLOCK ACE (commercial name; purified casein) to milk, but I do still use milk sometimes. If I had any consistent double layer formation as you suggest, should that not have blocked binding by either 1° Ab, 2° Ab, or both?

5. I haven't tried to control for proteases in milk, except by using BLOCK ACE, where I was really looking to use a single small protein to cover unbound sites. What data have you that show measurable effects of proteases on signal strength in ELISAs? Validated commercial assays in the USA typically come with protocols that recommend 5% milk in PBS.

6. Perhaps I misunderstand you here. The pre-coated commercial assays I've used came dry _with silica gel packets_, i.e. the capture Ab was bound to the plate then the plate was dried. If kept at about 4 °C, these assays are valid for about 8–12 months. As far as I know, no commercial supplier in the USA sends wet pre-coated plates.

Hi Omar,

The coating step is often a problem. First, I recommend binding capture Ab overnight at 4 °C for best binding. Second, I still recommend trying Protein A or something like it to bind your capture Ab. In that set-up, incubate the plate at 4 °C with Protein A overnight, then try to bind the capture Ab with Protein A at 35–37 °C for one hour.

The commercial plates pre-coated directly with capture Ab usually have a proprietary "stabilizer" that is supposed to enhance the stability of the capture Ab. I also suspect there are steps left out that they don't always tell us.

For blocking, I recommend you use BLOCK ACE (commercial name; purified casein) because of the smaller size of the casein. You may want to experiment with other small proteins to cover all the unbound sites as well.

When you refer to the polyclonal as "secondary antibody", do you mean the 2° Ab is the detector Ab? Are you using a 2nd Ab with tag for detection of signal, or is your tag on the detector Ab? I am also curious as to why you wish to use the biotin system.

I really would not use protein A for coating unless your tracer antibody does not react with it!

Hi Jörg,

Just curious, since I haven't heard this specific criticism of Protein A before. Have you tried to use it in ELISA, and were you unable to saturate the binding sites on the Protein A in the wells with your capture Ab? Was the binding strength sufficiently weak that some capture Ab came off during washes?

The criticism of Protein A that one commercial supplier of pre-coated ELISA plates told us was that they got just as good or better coating with direct application of capture Ab.

Try first coating the ELISA plate with serum or plasma samples containing your target HIV antigen and with antigen free control samples followed by blocking with BSA, monoclonal (or polyclonal antibody), conjugate and substrate steps. If you get significant positive result with the samples containing the target antigen in comparison with the control samples, the sandwich ELISA will be achieved.

Hi Wales,

the binding strength of protein A to different Ig varies depending on animal species or Ig isotype. In the individual situation, you can thus not exclude the possibility that other antibodies, namely the tracer antibody, will compete with the capture antibody for PA binding. Therefore you cannot exclude nonspecific binding, i.e. binding that is not mediated by the antigen in quest. You would have to prove that these theoretical concerns do not apply to your test.

Hi Jörg,

Can not the theoretical concern be tested empirically? I understand and respect your concern, but the point I am trying to make is that I don't believe one should exclude Protein A out of hand. Unless Omar is in a time crunch, I believe he should consider use of it as an option. Apparently one _can_ saturate the Protein A binding sites with capture Ab, because one of my colleagues and I were each told independently about use of Protein A to bind capture Ab years ago.

The problem of Omar is that he gets no reaction at all. The chance that this is due to not using protein A is null. He first has to get at least some positive signal, maybe by exchanging the antibodies. Although he has demonstrated that his antibodies bind to the antigen when this is immobilized, proof that they will also bind the soluble Ag has not been obtained. He could label the antigen (biotinylation) and check whether the antibody(ies) bound to the plate can capture the biotinylated antigen (make visible by peroxidase-coupled avidin). If both antibodies produce a signal in this way, there is only one possibility why they don't work in combination: they recognize the same idiotope. If either of them doesn't: that's why the test doesn't work.

Protein A could possibly be used for optimizing a test that already works, but as Wales suggests (and I have mentioned before), one would first have to make sure that the the tracer antibody does not bind to the plate coated with PA and the capture Ab.