The additional optimization of other ELISA factors may be necessary after determining the best concentration of antigen/enzyme conjugate to use (see Note 8). Three parameters may be varied to make improvement. Each one should be adjusted independently and an optimum dilution factor or value determined for routine use. Molecules of interest may be masked by components of the test fluid leading to unexpectedly low signals. This may be remedied by mixing the sample with PBS/Tween supplemented with 10% fetal bovine serum (see Note 9).
1. Vary the pH of carbonate coating buffer between pH 7.0 and 10.0 in 0.5 pH unit steps. Also try PBS at pH 7.4. Some antibodies, particularly monoclonals will bind better at a specific pH.
2. Make a dilution of the capture antibody in coating buffer from 0.5 to 10 Mg/mL and use to coat ELISA wells. The binding capacity of ELISA wells may be reduced if the concentration of antibody is not optimal. Overdilution, and paradoxically underdilution may lead to poor well binding. In some competition ELISA assays up to 100 Mg/mL may be required.
3. Make dilutions of streptavidin-HRP conjugate between 1:1000 and 1:10,000 and use to determine the optimum concentration for this reagent.
1. The quality of the antibodies used is perhaps the most important aspect in setting up a good ELISA. Antibodies should have a high affinity for the sample to be measured.
2. The pH of the coating buffer can affect the antibody, which will bind to the plate. Basically, a higher pH will result in more antibody binding but may have a detrimental effect upon its immunoactivity. Thus, a pH must be found which is suitable for the antibody in question. When beginning optimization of the assay, test a range of carbonate buffers from pH 7.0 to 10.0 as well as PBS pH 7.4. A carbonate buffer pH of 9.5 gives good results for binding antibody to a plate. In some cases, commercially available coating antibodies recommended by the manufacturer to be adsorbed onto the plate at pH 7.4 can be far more effective at higher pH's - improving the lower detection limit of sensitivity by up to 1000%, and allowing the coating concentration to be reduced.
3. There can be a significant difference between the protein binding capability of different makes, and even batches of microtitre plate. Some appear to be extremely good for binding antibodies, whilst others more useful for other proteins. The only real way to choose a suitable plate is by trial and error, or by using a recommended type known to bind the protein with which you intend to coat.
4. Although the HRP/TMB system is usually a good, reliable, and sensitive combination, HRP has a number of alternative substrates, which can be used such as o-phenylene diamine. There are also number of options for the enzyme used other than HRP, such as alkaline phosphatase, which can be used in combination with the substrate ^-nitrophenyl phosphate. It is important to note that if alkaline phosphatase is used, the wash buffer must not contain phosphate. Usually in this case a Tris-buffered rather than phosphate-buffered wash buffer is used. The choice of enzyme-substrate system depends on a number of factors, including price, sensitivity and whether a spectrophotometer filter is available for the substrate specific wavelength to be measured.
5. In some cases optimal results may be achieved by using the detergent Tween-20 only during the wash after incubation with the final reactant. The reason for this, though, is not well-understood (3).
Although Tween-20 is suitable for most applications, in some cases a more gentle detergent such as CHAPS may be more suitable. This may be the case for example, when using a surface such as the Nunc Polysorp brand (3).
6. Occasionally, an ELISA can give rise to unexpected results in wells close to the edge of the microplate. The phenomenon is often referred to as the "edge effect" whereby unexpectedly low or high OD readings are observed. In some cases this is a result of light or heat sources. For example, if the plate is being incubated near a strong light source such as a window any reaction that is photosensitive (such as the substrate reaction) may give rise to elevated OD levels in those wells closest to the light. More commonly, temperature difference causes the edge effect. This is particularly common when plates are stacked one upon the other or where a solution is taken straight from the fridge and the incubation step is performed at 37°C for a short period of time. In such cases, the outside wells are the first to heat up and thus the reaction rate is greater there, most likely owing to the faster rate of movement of molecules at this higher temperature.
To avoid these complications, plates should ideally be separated from each other during incubation periods and liquids should be adjusted to the temperature that the incubation step is to be performed at. Overnight incubation steps at 4°C tend to give better results than 37°C with respect to lowering the edge effect. It is also advisable to perform incubation in the dark or in subdued light. Sealing the plates or using a 100% humidified environment can also help.
7. The use of biotinylated secondary antibodies in conjunction with enzyme-conjugated streptavidin (or avidin, extravidin) both increases sensitivity and saves time in that a further step is eliminated from the assay and therefore another step of optimization is not required.
8. If the major aim of the ELISA is to obtain quantification of substances present in extremely low concentrations there are a number of adaptations to the technique that can be used. Such techniques often use alkaline phosphatase enzyme systems, which can be used, for example, to lock into a circular redox cycle producing an end product such as red formazan which, is hugely amplified in comparison to standard amplification methods (4). Chemiluminescent amplified ELISA principles have also been shown to give very high sensitivity (5) and can be optimized to measure as little as 1 zeptomole (about 350 molecules!) of alkaline phosphatase (6,7). Although extremely sensitive, such techniques are extremely time consuming to set up and optimize, and are far more expensive than the simple colormetric ELISAs described in this chapter.
9. In some cases, molecules present in a sample are masked by the solution that they are in. This problem can sometimes be solved by diluting the samples in PBS/Tween/10% fetal bovine serum. If this is performed, remember to make similar adjustments to the solution used for the standards. Possible interference molecules within samples such as soluble receptors for the antigen can also cause a problem.
Was this article helpful?