Printable Version

The Translation Post Vol.2 Issue 1

Featured Article - EpiShot^TM: a Rational Approach for Epitope Prediction

 

By John Mountzouris, Ph.D., Director Product Support

Antibody Production:  Starting out Right

Antibodies are powerful tools to profile protein function, interaction, cellular location, expression, and post-translational modification, and as such are indispensable to modern biological research.  The alternative for antibody production consists in making the decision of using the type of immunogen: protein or peptide. The ability to produce an immune response in an animal to a peptide conjugated to an appropriate carrier protein circumvents in many cases the time and expense associated with isolating or expressing and purifying native protein.  Additionally, the use of a peptide permits one to produce epitope-specific polyclonal antibodies, an extra measure of selectivity not possible when immunizing with whole protein. The substantial reduction in cost and time are compelling incentives that have fueled the rise of anti-peptide antibodies in both basic and applied research. That said there are a number of conditions that go into designing the peptide antigen and producing the antibody. Understanding those issues while working with the antibody production team enhances the likelihood of success. The EpiShot^TM epitope prediction procedure has been established at Abgent over the past years by compiling the expertise of different scientific fields within the same team. This multi-field expertise has been offered to our customers ever since, and is now the objective of this review.

Defining the Right Expectations for the Project

Whether the antigen is a peptide or a protein, the end user must choose whether a polyclonal or monoclonal antibody best matches their specific applications and requirements for their experiments.

Polyclonal antibodies will recognize multiple epitopes of an antigen. Since the minimum length of an epitope is thought to be 4-6 amino acids, even a peptide may contain several epitopes. Polyclonal antibodies are more likely to maintain recognition of antigenic epitopes even when modest changes in conformation or aggregation occur. Polyclonal antibodies are also capable of recognizing different epitopes with different affinities. Because of this ability, polyclonal antibodies have a broader range of potential applications than monoclonal antibodies.

Monoclonal antibodies recognize only one epitope of the antigen and are highly specific to that particular antigen. They generally yield much lower background because of this specificity. Monoclonal antibodies are also highly homogenous and allow for unlimited production with sustained specificity.

Whether producing a polyclonal or monoclonal antibody, much effort goes into selecting a peptide with high antigenicity, meaning that the particular amino acid composition and sequence strongly stimulate the immune system to produce antibodies that will bind to the peptide. This is a crucial first step for the project, and the step over which the most control can be exerted. Greater detail will follow regarding specific methods for this step. However, it is important to understand that with the skillful and experienced application of design principles to peptide selection, an anti-peptide antibody that recognizes the native protein may be obtained for approximately 75% of projects [1]. The failures may be assigned to a number of reasons. The cognate peptide sequence may not be readily accessible to antibodies, because it is buried in the protein structure. The target sequence within the protein may be anchored within secondary structural elements that render it highly inflexible. A flexible free peptide may generate antibodies to structures that do not occur within the protein. It is also possible that anti-peptide antibodies will not bind to denatured forms of the protein that render the target sequence inaccessible. Finally, it is important to remember that peptides represent linear B-cell epitopes. Vaccine research combined with three-dimensional structural studies has shown that many epitopes are discontinuous, i.e. conformational in nature, defined by the proximity in space of amino residues that are far apart in primary sequence. This library of epitopes is inaccessible to the anti-peptide antibody approach.

Biophysical Parameters for Epitope Prediction

The selection of candidate peptides for antibody production from the full-length protein sequence has been subjected to a dizzying array of predictive algorithms, all of which work from different facets of an underlying principle: The immune system produces antibodies against "features" that are present on the "surface" of the protein. Amino acids can be ranked, based on previous structural knowledge, according to their respective probabilities in promoting attractive structural features for antibody recognition (i.e. proline's role in promoting β-turns) and in promoting surface exposure (i.e. the tendency of highly hydrophilic residues such as arginine and lysine positioned between protein and the aqueous environment). Among applied amino acids scales to predict antigenicity are the following [from 2]:

• Parker [3]
• Hopp and Woods [4]

Protein hydrophobicity algorithms:

• Kyte and Doolittle [5]
• Janin [6]
• Sweet and Eisenberg [7]
• Manavalan [8]

Protein flexibility prediction algorithm:

• Bhaskaran [9]

Protein secondary structure prediction algorithms:

• Deleage and Roux [10]

• Chou and Fasman [11]

• Levitt [12]

Protein surface exposure algorithm:

• Rose [13]

Generally, the average values for these physicochemical parameters are plotted over a window of 6-8 amino acids moving along the protein sequence. Regions that show a consensus among the different parameters are then further evaluated against other criteria for final selection of the peptide (Figure 1).

Figure 1: Antigen scan for a 150-amino acid protein by using the EpiShot^TM epitope prediction procedure. Antigenic regions considered for antibody production are highlighted in black line.
 

Antibody Production: The EpiShot^TM Epitope Prediction Procedure

With the EpiShot^TM epitope prediction procedure, Abgent provides the customers with a report that visually summarizes the biophysical information, and makes accompanying recommendations about the best candidates for antibody production. Such criteria include:

1.       Length

For example, the length of the peptide should be between 12 and 25 amino acids according to Abgent experience. Longer peptides may adopt non-native structures that give rise to antibodies that do not recognize the native protein. Shorter peptides do not provide sufficient epitopic diversity, thereby producing a constricted range of antibodies, limiting the opportunity to generate antibodies that recognize the protein. In the case of extremely short peptides selected from the interior of the protein sequence, for example, the overwhelming majority of antibodies may be against the charged N or C terminal residue carrying the respective amide or carboxyl functional group, which would impair recognition by antibodies of the sequence in a native protein where the residue is linked to its flanking neighbors by a peptide bond. The exception to this length requirement occurs in the case of antibodies to phosphor- or other modification-specific sites. Phosphorylation of a single residue produces minor conformational changes in an epitope. It is desirable to "focus" antibodies on this site, so peptides of 8-12 amino acids with the modification centrally located are selected to enhance the production of phospho-specific antibodies. Abgent is experienced in special synthesis and purification protocols for antibodies specific to a range of post-translational modifications, including phosphorylation, methylation, acetylation, sumoylation, and others.

2.       Post-translational modifications and topology

Unless one is interested in generating antibodies that are modification specific, peptide candidates must be carefully screened to avoid those incorporating known modification sites, since the modification may block recognition by the antibodies of the native site if it is modified. Cysteine disulfide bonds can introduce unique structural features, bulky glycosylation sterically hinders access by antibodies, and a host of more than 200 known protein modifications may introduce unknown perturbations in the structure of either the peptide or the protein that may impair the transfer of peptide recognition to protein recognition by an anti-peptide antibody. Online tools for screening peptides for these modifications are available [14].

When designing peptides for antibodies against transmembrane proteins such as G-protein coupled receptors, it is absolutely critical to avoid those regions that span the membrane, since the antibody does not have the ability to access them in the native protein. The membrane spanning regions may have been described in publication; alternatively, online tools are available to assist in defining these regions [15]. Whether to choose a peptide from the extracellular or cytoplasmic loops will depend upon the applications and the objectives defined by the researcher.

With EpiShot procedure, Abgent provides a specialized report for transmembrane proteins to assist in selecting a peptide (Figure 2).

Figure 2: Generic example of membrane protein topology and epitope candidates from the EpiShot™ procedure.

3.       Conjugation

Peptides less than 25 amino acids in length are usually conjugated to keyhole limpet cyanin (KLH) or some other immunogenic large molecular weight protein. Conjugation is typically done through a single cysteine present the N- or C-terminus of the peptide. Internal cysteines are not desirable conjugation sites because the full peptide containing all epitopes will be partially occluded by the carrier protein and therefore will restrict generation of a full complement of antibodies. Given these requirements, three scenarios are possible:

a.      A cysteine flanks a desirable sequence in the native protein.

The peptide sequence is extended by a few amino acid residues to incorporate the terminal cysteine.

b.      No cysteine flanks a desirable sequence in the native protein.

A non-native cysteine is placed at either the N- or C-terminus of the peptide to provide for conjugation. The addition of this single residue does measurably influence the production of antibodies, particularly since it is hidden close to the surface of the conjugating protein.

c.       There is a cysteine within a desirable sequence in the native protein.

 Where it is impossible to avoid an internal cysteine because the target peptide is otherwise an optimal antigen, a short artificial linker can be attached to a terminal residue on the peptide to provide for conjugation, while the cysteine is blocked.  The linker should be small, stable, and not distort peptide structure. Abgent is a leader in the application of Hydralink^TM technology (see our Couplage^TM Kits for chemistry information) to these special cases.

4.       Ease of Peptide Synthesis and Solubility

The chemical, physical, and structural properties of a peptide depend on its unique amino acid sequence and composition. For this reason, peptide sequences can be challenging to synthesize, purify, and/or solubilize. In general, peptide sequences that are rich in hydrophobic amino acids can be difficult to dissolve in aqueous solutions, and may be unsuitable for use in biological systems. The following tips will be useful in modifying your peptide sequence to assist in solubility, synthesis, and/or purification:

• The number of difficult amino acids such as Cys, Met, Arg, or Trp should be minimized in the sequence.

• The number of hydrophobic amino acids should be minimized, or a stretch should be broken up with a polar amino acid.

• Glutamic acid residues at the N-terminus should be avoided because of potential formation of a pyroglutamine by cyclization.

• Multiple glutamine residues should be avoided, since they may cause insolubility by forming hydrogen bonds between peptides.

• A single cysteine in the selected sequence is useful for conjugation to a carrier protein. The cysteine should be at the N- or C- terminal end for optimal conjugation. An inter- or intra- peptide disulfide bonds may be formed due to the presence of additional cysteine residue in the sequence, thus leading to insolubility and structural alteration of the peptide.

• Proline and tyrosine residues enhance the immunogenicity of a peptide as they induce conformational change of peptide that mimics the native structure. One or more proline, arginine and histidine in the middle of the sequence will greatly reduce peptide aggregation during synthesis.

5.       Sequence Homology

To enhance specificity, peptide sequences must be tested for homology against other proteins from the target species. This assures that peptides holding epitopes in common with proteins other than the target protein are removed from consideration. Other cases where homology may come into play include a requirement that an antibody work in more than one species, or that an antibody be isoform-specific. Candidate peptide sequences can be blasted against either the NCBI or Swiss-Prot protein databanks, and the results easily surveyed to determine which candidate best meets these criteria.

6.       Structural information

It is often the case that a structure for a target protein is known from X-ray crystallography or NMR study, and the coordinates are available from a public databank such as PDB.  There are also a number of free prediction programs online that predict structures from submitted sequences based on homology modeling [16]. Viewing 3-D structures with color-coded epitopes can help in either identifying or confirming that peptide sequences are positioned on the surface of the protein, where they have the highest likelihood of producing an antibody (Figure 3).

Figure 3:  The sequence of a peptide candidate (highlighted in yellow) can be refined by 3D-viewing during the EpiShot^TM epitope prediction procedure. 

At Last: Peptide Purity and Quantity

The quantity of peptide required will depend upon the immunization and boosting schedule, and the number of animals to be immunized.  While as little as 0.1 mg per animal has been used with special protocols, it is generally recommended to plan for 1-2 mg of peptide per animal. If peptide affinity purification of the antibody is required, an additional 5 mg of peptide should be allotted.   

Peptide purity >70% is generally sufficient for production of good anti-peptide antibodies. Peptides that were truncated during synthesis are included in these impurities, and contain virtually the same epitopes as the parent peptide.  Additionally, the concentrations of individual truncated peptides along with peptides containing deletion residues are at least an order of magnitude less than the parent peptide.  While HPLC purification of peptides to >90% or >95% may add 5-20% to the cost of the project depending upon the peptide, such stringency may be called for where the antibody must discriminate between proteins differing in only a few residues within the epitope. The standard production of reagent antibodies employs either protein-G purification or saturated ammonium sulfate precipitation of the antisera. In those cases where the user wishes to follow up with purification of the antibody against the peptide itself, a higher level of purity for the peptide may reduce the number of contaminant antibodies that co-elute.  Peptide-affinity purification costs somewhat more and reduces the yield of antibody (IgG fraction) by ~95%, and therefore may not be suitable for everyday applications.  

Some researchers seek reduce project costs by co-immunizing animals with multiple peptides. This two-for-one approach is intended to enrich the diversity of antibodies that are produced and to short-circuit potential failure if a single antigen does not produce an immune response. In some cases a good polyclonal antibody that can capture the protein from multiple epitopes is produced. However, in practice all antigens are not equal in immunizing potential. Very often the immunodominant peptide will produce the overwhelming majority of antibodies, and little ground is gained for the added expense. Consultation with the peptide and antibody production teams prior to choosing the strategy will help in making the best decision regarding this option. 

In Closing 

The article has reviewed key aspects of peptide design for the production of antibodies.  With so many aspects to consider and literally thousands of peptide sequences from a typical full-length protein, epitope prediction programs have come to the fore as a powerful means to winnow down the choices to a handful for final consideration. Although such automated computational methods are becoming indispensable to high-throughput production of commercial scale antibody collections, the field still demands the art of a practiced team to consistently produce successful outcomes. For custom antibody production, the need for a strong collaboration and effective communication with the vendor remain as essential as ever. It is advisable to inspect the company's track record by analyzing the scope and quality of data for their product lines, the assembled experience of the team, to search the literature for the company's name to find antibodies have been cited, and to contact others who have used the company's services and can provide an independent review. 

Abgent and Custom Antibody Production

Since inception in 2001, Abgent has generated over 5,000 monoclonal and polyclonal antibodies, mostly by using peptides as antigens. Abgent has currently over 2,000 catalog antibodies, and has been a major provider for custom antibody production to the scientific community. The EpiShot^TM epitope prediction procedure consists in using the expertise of a multi-scientific background production team to generate the best reagents for our customers.

EpiShot^TM is a trademark of Abgent. 

References 

[1] Abgent unpublished data based on 4,000 polyclonal antibody production projects

[2] ProtScale bioinformatics tool from ExPASy website: http://tw.expasy.org/cgi-bin/protscale.pl. Gasteiger et al. Protein Identification and Analysis Tools on the ExPASy Server; (In) John M. Walker (ed): The Proteomics Protocols Handbook, Humana Press (2005). pp. 571-607

[3] Parker et al.  New hydrophilicity scale derived from high-performance liquid chromatography peptide retention data: correlation of predicted surface residues with antigenicity and X-ray-derived accessible sites. Biochemistry 25: 5425-5431 (1986).

[4] Hopp et al. Prediction of protein antigenic determinants from amino acid sequences. Proc. Natl. Acad. Sci. U.S.A. 78: 3824-3828 (1981).

[5] Kyte et al. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157: 105-132 (1982).

[6] Janin. Surface and inside volumes in globular proteins. Nature 277: 491-492 (1979).

[7] Sweet et al. Correlation of sequence hydrophobicities measures similarity in three-dimensional protein structure. J. Mol. Biol. 171: 479-488 (1983).

[8] Manavalan et al. Hydrophobic character of amino acid residues in globular proteins. Nature 275: 673-674 (1978).

[9] Bhaskaran et al. Dynamics of amino acid residues in globular proteins. Int J Pept Protein Res. 1984 Aug; 24(2): 180-91.

[10] Deleage et al. An algorithm for protein secondary structure prediction based on class prediction. Protein Engineering 1: 289-294 (1987).

[11] Chou et al. Prediction of the secondary structure of proteins from their amino acid sequence. Adv. Enzym. 47: 45-148 (1978).

[12] Levitt. Conformational preferences of amino acids in globular proteins. Biochemistry 17: 4277-4285 (1978).

[13] Rose et al. Hydrophobicity of amino acid residues in globular proteins. Science 229: 834-838 (1985).

[14] Post-translational modification prediction tools from the ExPASy website: http://tw.expasy.org/tools/#ptm

[15] Topology prediction tools from the ExPASy website: http://tw.expasy.org/tools/#topology

[16] Tertiary structure prediction tools from the ExPASy website: http://tw.expasy.org/tools/#tertiary