|
.... Polyclonal Antibody : Antigen Design
|
|
First of all, we believe that it is important to emphasize that antigen design (and its related,
epitope prediction) is not merely to find a peptide that can generate strong immuno-response
from animals. It is to find a peptide to produce an antibody that can specific react with its target
protein. Antigen design is a very difficult area that has been subject to extensive study, but
still far from being fully understood.
In general, a good peptide antigen has the following properties: protein surface location,
flexible (usually loop) rather than helical structure, complicate and unique sequence, not
a post-translational modification site or functional site (unless the antibody is intended to
recognize such a site) and easy for synthesis. But the quantitative relationship between
these properties and antigenic strength is not clear.
Antigen design is usually done by two different approaches: predicting peptide's physical-
chemistry properties or making prediction based on statistic results. Both approaches have
their limitations. Physical chemistry properties such as peptide location in a protein or its
secondary structure, particular turn, are difficult to be predicted with high degree of accuracy
This is because the problem themselves are parts of one of the most challenge areas of
modern science ---- protein folding. Another problem is that in most cases, an isolated
peptide in solution can not maintain its native conformation found in the protein. This is the
major reason that antigens designed based on known 3D structures are also often failed.
Statistic methods that base on frequencies of amino acids in known epitopes fare even
worse. The predicted results are marginally better than randomly picked sequences,
according to a recent study (Blythe M. and Flower D. (2005) Protein Sci., 14:246-248).
The method by Kolaskar and Tongaonkar combines both approaches and is one of the
most popular epitope prediction methods. They claim that its accuracy is up to 75%
(Kolaskar AS & Tongaonkar PC (1990), FEBS Lett. 276: 172-4)
Antigen design can be further complicated by the application of an antibody. The above
methods are for antibodies intended to recognize the protein in its native folded
conformation, which is suitable for experiments such as immuno-precipitation, cell image
and other structure and function studies as well as vaccines. It may not work for SDS
PAGE based Western blot. The conformation of a SDS-denatured protein is different
from its native one. It is thought to become cigarette-like shape with hydrophobic residues
being exposed to the outside. SDS molecules bind to the hydrophobic surface. Exactly
how the structure looks like is not clear so far. Attempting to obtain such structure has
not been successful due to technical difficulties.
The antigen design service provided by EZBiolab is consist of three parts:
1. Use several popular publicly available programs
2. Narrow and refine the results from part 1 with our own algorithm ( the
principle is based on Kolaskar and Tongaonkar but with incorporation
of some parameters that are derived from our own experiences)
3. Blast search to filter out similar sequences
EZBiolab is more than glad to provide antigen design assistance to its customers. At
the same time, we strongly encourage our customers to design their own antigens
because knowledge about the structure, function, bioinformatics and other properties
of each protein is essential for finding a good antigen. Customers know their own
proteins much better than we do. If a customer is not familiar with this subject, we
can provide several potential peptides based on the general principles to allow him/her
make the final decision.
Again, we like to emphasize that antigen design is not to find a peptide that can generate
strong immuno-reaction from animals. It is to find a peptide that generates an antibody to
specifically react with its target protein. In fact, over emphasis ELISA titer to the peptide can
generate the opposite results. An example is trans-membrane domain peptides. These
peptides usually cause strong immuno-responses but they are obviously poor antigens
from the standpoint of protein-binding. For this reason, EZBiolab does not provide full
guarantee to our antibody service by promising high ELISA titers to peptide antigens. We
believe such guarantees do not mean much and to a large extent, is misleading. By no
means a strong anti-peptide antibody can be assured to also bind to its protein strongly and
specifically.
|
|
|
|
|
