Production of Peptide Antibodies in Rabbits and Purification of Immunoglobulin

I. INTRODUCTION
As a consequence of the completion of human genome sequencing, a huge number of new putative proteins can be derived from the DNA sequence information. The new challenge is how to discover their function and cellular localization, as well as their interaction with other proteins. The characterization of proteins is facilitated greatly by the availability of specific antibodies. There are at least two commonly used strategies to obtain antigen needed for antibody production: (1) cloning of the gene in an expression vector followed by transfection in an expression system and subsequent protein purification, often with the help of a peptide tag to facilitate detection and purification, and (2) chemical synthesis of a peptide corresponding to a short portion of the protein against which antibodies are to be raised. The latter strategy has several big advantages over the cloning expression method: (1) it is less time-consuming; (2) many potential problems, such as difficult cloning, low expression level, or difficult purification of the protein, can be avoided; the peptide antigen is easily available in large quantities and can be used for subsequent affinity purification of the antiserum; and by correct choice of the peptide sequence, antibodies of extreme specificity can be obtained and be used to detect one specific member out of a protein family with strong homologies.

The cloning expression of the entire protein would of course display many epitopes shared by other members of the family and therefore give rise to crossreacting antibodies (recognizing several members of the family or even unrelated proteins); the only solution here is to make monoclonal antibodies, which is more time-consuming and needs intense screening work. Other applications for which peptide antibodies are the better strategy include detection of one specific splice form, one specific mutation, or one specific phosphorylation site.

The only real disadvantage of antibody production using peptides as a general antibody production approach is the risk related to the choice of the region in the protein sequence. In fact, the correct choice of a suitable peptide sequence is the most crucial step in the production process and particular attention should be paid to this selection. For a given protein, it is always advised to use more than one synthetic peptide as antigen to increase the chances of getting antibodies recognizing the protein and, if possible, in its native form. As a detailed discussion of the many parameters influencing the selection of a well-suited peptide would be beyond the scope of this article, only a few hints can be given here.

Short synthetic peptides are not immunogenic by their own because of their low molecular mass (generally 1-3 kDa); in order to elicit an immune response in the animals, they must be either coupled covalently to a so-called carrier protein or synthesized on a special resin, called MAP (Tam, 1988). This article does not cover the synthesis of the peptides, as custommade peptides can be easily obtained commercially. A second important issue to consider when selecting a peptide sequence is the choice of the conjugation strategy to allow for a proper display of the entire peptide sequence for antibody production in the animals. The

conjugated peptide (or MAP peptide) is then used as antigen in rabbits as described (Huet, 1998). High-titered antiserum can be obtained within about 2-3 months and can be subsequently affinity purified using the synthetic peptide coupled to a suitable matrix. This process allows to get rid of all the unrelated antibodies present in the antiserum (endogenous antibodies from the rabbits and the antibodies produced against the carrier protein) that otherwise may give rise to background signals.

II. MATERIALS AND INSTRUMENTATION
15-25 mg of custom-made peptide (or MAP peptide) at >70% purity, stored lyophilized at 4°C under inert gas and protected from moisture

KLH (Sigma Cat. No. H7017), stored lyophilized at 4°C

Glutaraldehyde (Sigma Cat. No. G6257), 25% solution in water, stored at room temperature. Because this product is toxic, always work under a hood and wear gloves.

EDC (Pierce Cat. No. 22980)

MBS (Pierce Cat. No. 22311)

Dimethyl formamide (DMF) (Sigma Cat. No. D8654), stored at room temperature

Glycine (Sigma Cat. No. G7126), stored at room temperature

Boric acid (Vel Cat. No. 1021), phosphate-buffered saline (PBS) 10× (Sigma Cat. No. P7059), tromethamine (Tris) (Sigma Cat. No. T1503), sodium hydrogen carbonate trihydrate (NaHCO3·3H2O) (Vel Cat. No. 7958), sodium chloride (NaCl) (Vel Cat. No. 1723), disodium hydrogen phosphate dodecahydrate (Na2HPO4·12H2O) (Vel Cat. No. 1773), glacial acetic acid (Vel Cat. No. 1005), 0.1M hydrochloric acid (Sigma Cat. No. 210-4), sodium hydroxide (Sigma Cat. No. S5881), and guanidine hydrochloride (Sigma Cat. No. G4505), all stored at room temperature

PD10 column (Pharmacia Cat. No. 17-0851-01), activated CH-Sepharose 4B (Pharmacia Cat. No. 17- 0490-01), EAH-Sepharose 4B (Pharmacia Cat. No. 17-0569-01), stored at room temperature

Dialysis membrane Spectra/Por (Spectrum Cat. No. 132678), molecular weight cutoff 12-14000

pH meter MP220 (Mettler-Toledo) or similar Spectrophotometer Ultraspec 3300P20 (Pharmacia) or similar

FPLC Akta with UV detector (Pharmacia) or similar

III. PROCEDURES
A. Selection of Suitable Peptide Sequences and Definition of Coupling Strategy
1. Selection of Antigenic Peptides
Whenever possible, use a software for predicting the "'antigenicity" of the peptide. We use the Lasergene software from DNASTAR, Inc., particularly the EditSeq (sequence edition), Protean (antigenicity prediction), MegAlign (sequence alignments), and GeneMan (homology searches) modules.

Other useful free software can be used on the Expasy Molecular Biology Server of the Swiss Institute of Bioinformatics (http://www.expasy.org/); e.g., the Protscale program can be used to assess hydrophilicity and other parameters. Links to software for predicting signal peptides, transmembrane domains, potential phosphorylation, and glycosylation sites can be found there too. Another good source is the protein analysis software contained in the GCG Wisconsin package from Accelrys (http://www.accelrys.com/ products / gcg_wisconsin_package / program_list.html #Protein) and there may be many more available through the web. All these computer programs can be very helpful in selecting a well-suited peptide. A few hints are given here to facilitate the selections based on the most used algorithms.

Check in an international database if the protein has been described. There is often helpful information concerning different domains of known proteins in their description (e.g., transmembrane regions, signal peptides, potential glycosylation or phosphorylation sites, or even a link to the three-dimensional structure if this has been resolved). If the three-dimensional structure is known, use the structure information to define a surface-located region (e.g., using the Swiss Viewer software available at the Expasy server: http://www.expasy.org/spdbv/). After pasting the sequence in the EditSeq module, open it with the Protean program and check the following parameters over the whole length of the protein sequence: hydrophilicity (Hopp-Woods and Kyte-Doolittle), antigenic index (Jameson-Wolf), surface probability (Emini), flexibility (Karplus-Schulz), turn, α, β, and coil plot (Garnier-Robson), as well as positive and negative regions in terms of charge.

If the N or C terminus of the protein is hydrophilic, it is often a good choice. Some proteins contain signal sequences, which are cleaved in the mature proteins. Those signal peptides are very often hydrophobic parts of about 20-30 amino acids, which are not well suited. If doubt persists, use a software (e.g., SignalP program available at http://www.cbs.dtu.dk / services/SignalP/) to predict its cleavage site.

If internal peptides have to be used, they should be hydrophilic, present a high surface probability, and should not be predicted as strongly α helical. Short synthetic peptides will hardly adopt the same α-helical structure and therefore the antibodies will not recognize such structured regions in the native protein. Good hydrophilicity and surface probability are essential if the antibodies have to recognize the native protein (e.g., immunohistochemistry experiments).

Choose a peptide in a flexible region whenever possible. Avoid potential phosphorylation (serine, threonine, and tyrosine) and glycosylation sites (not predicted by these algorithms, but can be predicted by other software if no better information is available). Antibodies made against a simple peptide will off course not fix at the corresponding region in the protein if this contains a large sugar moiety.

Avoid peptides containing several cysteines, as they may participate to disulfide bridges impossible to mimic with a short synthetic peptide. If the protein contains transmembrane regions, avoid loops between two transmembrane domains that are shorter than 30-35 amino acids. Short loops are not well mimicked by flexible and linear peptides.

Choose several well-suited peptides of about 12-15 amino acids length and rank them from best to less good based on the aforementioned parameters. Shorter peptides (9-11 amino acids) can be used if the antibodies have to recognize only a very small region, e.g., a phosphorylation site, which should then be placed about in the middle of the peptide.

2. Homology Searches
In order to avoid unspecific or background signals, it is important to check the chosen peptide sequences for possible homologies with known proteins other than the target one. Depending on the application, interspecies cross-reactivities may not be disturbing or may even be desired.

For every chosen peptide sequence, run a BLAST search (available at the NCBI: http://www.ncbi.nlm. nih.gov/BLAST/) using the Protein Blast and the "Search for short nearly exact matches" option. A standard BLAST with a short peptide will hardly return any result, while the option here above gives matches for short parts of proteins.

Discard those peptides for which stretches of more than five identical amino acids are found in another protein(s). Because one epitope generally contains more than five amino acids, specific antibodies can be expected if less or equal than five amino acids one following the other are present. The following example illustrates these issues:

peptide: E H R T P R G K E D S S V P  
  | | | | | | | |     |   | |  
other protein: E H R T P R G K Y Q S I V P  

antibodies may cross-react (8 identical amino acids in a stretch)

peptide: E H D T D R G M I D S S V P  
  | |   |   | |       | | | |  
other protein: E H R T P R G K Y Q S S V P  

antibodies will be specific (no stretch >4 identical amino acids)

Care should also be taken with very similar amino acids, such as serine and cysteine, with the only difference being the functional group on the side chain (OH vs SH), which is only a very minor difference that may not be sufficient to avoid cross-reaction if the rest of the epitope is identical.

If specificity is the major goal (e.g., antibodies against one specific member out of a protein family), select those peptides with the lowest homologies as first choices. If the first goal is to obtain working antibodies, give a higher weight to antigenicity analysis than to the homology part.

3. Coupling Strategy
It is very important to consider the following rules for the coupling of peptides to the carrier protein. After coupling to the carrier, the peptide should mimic the original structure in the best possible way, which means that the peptide should be linked to the carrier protein at one end and not in the middle of the peptide. A conjugation in the middle of the sequence would only display two small parts useful for antibody production (considering that the antibodies containing the conjugation residue in their epitope will not recognize the protein of interest in which this conjugation reagent is not present of course).

If the chosen peptide corresponds to the real N terminus of the protein, coupling must be done on the C-terminal part of the peptide. If the peptide sequence contains lysine, aspartate, or glutamate, add a cysteine C-terminally for conjugation using MBS and synthesize the peptide as C-terminal amide (C terminus:- CONH2), which better mimics the normal amide bond and does not introduce an unnatural negative charge. If it does not contain lysine, aspartate, or glutamate, synthesize the peptide as free acid (C terminus:- COOH) and use this acid group for coupling with EDC (this reagent would also react with internal glutamates, aspartates, and, to a lesser extent, lysines and the conjugation would not be specific in that case).

If the chosen peptide corresponds to the real C terminus of the protein, coupling must be done on the Nterminal part of the peptide and the peptide must be synthesized as free acid (C terminus: COOH). If the peptide sequence contains lysine, add a cysteine at the N terminus of the peptide to allow for a specific conjugation using MBS. If no lysine is present, the peptide can be coupled specifically on the N-terminal amino group using glutaraldehyde (lysine would also crossreact with the glutaraldehyde and the conjugation would result in peptide multimers rather than specific conjugation to carrier). In that case, no additional cysteine is required.

If the chosen peptide corresponds to an internal peptide, coupling should be done on the less suited end of the peptide (the end showing the lowest hydrophilicity, surface probability, and antigenicity index). For N-terminal conjugation, a cysteine should be added in case the peptide sequence contains lysine. The peptide should be synthesized as C-terminal amide (C terminus: CONH2). For C-terminal conjugation, a cysteine should be added C-terminally if the peptide sequence contains glutamate, aspartate, or lysine for the aforementioned reasons. The peptide should be acetylated on the N terminus to avoid introduction of an unnatural positive charge at that point. If no aspartate, glutamate, and lysine are present, the C-terminal acid function of the peptide can be used for specific coupling by EDC.

Once the coupling strategies are defined, synthesize the corresponding peptides using standard Fmoc or Boc chemistry on a peptide synthesizer. Synthetic peptides are routinely obtained from commercial sources and should be synthesized in quantities of at least 10-15 mg and in purities above 70%. There is no need for very high-purity peptides (which are much more expensive), as quite low amounts of impurities will not give rise to antibody production (the injected quantities of impurities are not high enough to elicit immune responses generally).

B. Conjugation of Peptides to Carrier Proteins
1. Coupling Using Glutaraldehyde (Coupling of N-Terminal Amine of Peptide to Lysine Residues of KLH)
Buffer preparation: Borate buffer, pH 10; dissolve 6.2 g of boric acid in 1 liter of distilled water.

1M glycine: dissolve 751 mg glycine in 10ml water

Dissolve 5 mg of peptide in 1 ml of borate buffer in a 10-ml glass tube.

Dissolve 5 mg carrier protein (KLH) in 1 ml of borate buffer and add this solution to the peptide solution.

Prepare a solution of 0.3% glutaraldehyde by mixing 120µl glutaraldehyde solution (25%) in 9.88ml of water.

Add dropwise 1 ml of this solution to the stirred peptide-KLH mix. Incubate 2h at room temperature. The solution will turn yellow.

Add 0.25 ml of 1M glycine solution to block unreacted glutaraldehyde. Stir 30min at room temperature.

Dialyze twice for 4h against 5 liters of PBS buffer. Measure the absorbance at 280nm (nanometers if the peptide sequence contains tryptophan or tyrosine). The resulting solution is ready for antibody production. Measure the final volume and divide in injection aliquots considering a coupling yield of 50%. We inject generally 200µg of coupled peptide (based on peptide content), which corresponds to 1/12.5 of the total volume. Dilute the aliquot with PBS to 0.5 ml and mix with adjuvant prior to injection as described (Huet, 1998).

2. Coupling Using EDC (Coupling of C-Terminal Acid to Lysine Residues of KLH)
Dissolve 5 mg of peptide in 1 ml of water in a 10-ml glass tube.

Add 20mg EDC to the stirred peptide solution and adjust the pH to 4.5 using 0.1M hydrochloric acid solution. Incubate for 10min at room temperature.

Dissolve 5 mg carrier protein (KLH) in 1 ml of water and add this solution to the activated peptide solution.

Stir for 2h at room temperature.

Dialyze twice for 4h against 5 liters of PBS buffer. Measure the absorbance at 280nm (nanometers if the peptide sequence contains tryptophan or tyrosine). The resulting solution is ready for antibody production. Measure the final volume and divide in injection aliquots considering a coupling yield of 50%. We inject generally 200µg of coupled peptide (based on peptide content), which corresponds to 1/12.5 of the total volume. Dilute the aliquot with PBS to 0.5 ml and mix with adjuvant prior to injection as described (Huet, 1998).

3. Coupling Using MBS (Coupling of Thiol Group of Cysteine on the Peptide to Lysine Residues of KLH)
Buffer preparation: phosphate buffer 10mM, pH 7.0; dissolve 3.6g of disodium hydrogen phosphate dodecahydrate (Na2HPO4·12H2O) in 1 liter of water

Phosphate buffer 50mM, pH 6.0; dissolve 17.9g of disodium hydrogen phosphate dodecahydrate (Na2HPO4·12H2O) in 1 liter of water

Dissolve 5 mg of KLH in 2.5 ml of 10mM phosphate buffer, pH 7. Dialyze or desalt against the same buffer.

Dissolve 15 mg MBS in 1 ml of DME Use this solution within 1 h!

Add 70 µl of the MBS solution to the solution of KLH. Stir for 30min at room temperature.

Equilibrate a PD10 column by washing with 50ml 50mM phosphate buffer, pH 6.0, charge the KLH solution on the column, and add 3.5ml of 50mM phosphate buffer, pH 6.0. Recover 3.5 ml phosphate buffer containing the activated carrier.

Dissolve 5 mg of peptide in 1 ml of PBS buffer and add to the activated carrier solution. Stir for 3 h at room temperature.

Dialyze twice for 4h against 5 liters of PBS buffer. Measure the absorbance at 280nm (nanometers if the peptide sequence contains tryptophan or tyrosine). The resulting solution is ready for antibody production. Measure the final volume and divide in injection aliquots considering a coupling yield of 60%. We inject generally 200µg of coupled peptide (based on peptide content), which corresponds to 1/15 of the total volume. Dilute the aliquot with PBS to 0.5 ml and mix with adjuvant prior to injection as described (Huet, 1998).

C. Affinity Purification
In many cases, crude peptide antiserum from the rabbits can be used in different techniques to probe the protein of interest. However, background signals may be a problem, either already in the preimmune serum or only in the serum from the hyperimmunized animal. Those antibodies may be very disturbing in techniques such as immunohistochemistry or even in Western blots. The best way to get rid of those background signals is to purify the antibodies by affinity against the peptide. The fact that the synthetic peptide is generally available in large amounts (in contrast to many protein antigens) makes the affinity purification of peptide antibodies very attractive in obtaining an improved signal-to-noise ratio or getting rid of crossreacting antibodies. This technique is also best if the antibodies are to be labeled by biotin or a fluorescent compound if one wants to avoid the use of secondary antibodies.

It is important to link the peptide to the affinity matrix through the same residue used for coupling. Otherwise, the peptide is not displayed the same way and good antibodies may be lost because they would not bind to a changed environment. For this reason, we advise using different resins depending on the coupling strategy used for the conjugation to the carrier protein. We generally use 10mg of the peptide for preparation of the affinity column; in most cases, this quantity is sufficient to deplete up to 50ml antiserum and to get, depending on specific titer, 8-25 mg of pure antibodies.

1. Conjugation of Peptide to Activated CHSepharose 4B
This procedure is used for peptides that have undergone a glutaraldehyde coupling to a carrier protein.

Buffer preparation: l mM HCl; dilute 1 ml of 0.1M hydrochloric acid by water to 100ml Carbonate buffer, 0.5M salt, pH 8.5; dissolve 6.89 g of sodium hydrogen carbonate trihydrate (NaHCO3.3H2O) and 29.29 g of sodium chloride in 1 liter of water

Weigh 0.75g of Activated CH-Sepharose in a 50-ml tube and swell the resin in 40 ml of 1 mM HCl for 30 min under stirring.

Dissolve 10mg of peptide in 1 ml of carbonate buffer, 0.5M salt, pH 8.5.

Centrifuge the gel and discard the supernatant. Wash the gel three times with 40ml carbonate buffer, 0.5 M salt, pH 8.5. Centrifuge and discard the supernatant each time. Suspend the gel material in 1 ml carbonate buffer, 0.5M salt, pH 8.5.

Add the peptide solution to the gel, add 1 ml carbonate buffer, 0.5M salt, pH 8.5, and let it react for 2h at room temperature under gentle stirring.

Centrifuge and discard the supernatant. Resuspend in 1 ml carbonate buffer, 0.5 M salt, pH 8.5, centrifuge, and discard the supernatant.

Suspend the gel in 1 ml PBS buffer and transfer the gel in the column adapted to the FPLC instrument.

2. Conjugation of Peptide to EAH-Sepharose 4B Using EDC
This procedure is used for peptides that have undergone a EDC coupling to a carrier protein.

Buffer preparation: 0.1M acetate buffer, 0.5M salt, pH 4.0; dilute 57.2ml of glacial acetic acid by water to 1 liter and dissolve 29.29 g of NaCl

Phosphate buffer, 0.5 M salt, pH 7.4; dissolve 29.29 g of sodium chloride in 1 liter of PBS buffer Water, pH 4.5: adjust pH of water by adding 0.1M hydrochloric acid solution

0.1M sodium hydroxide: dissolve 4g of sodium hydroxide in 1 liter of water

Take 3ml of EAH-Sepharose 4B suspension, centrifuge, and discard the supernatant. Wash the gel three times by 2 ml water, pH 4.5. Resuspend the gel in 1ml of water and adjust pH to 4.5 by adding some 0.1M hydrochloric acid solution.

Dissolve 10mg of peptide in 1 ml of water and adjust pH to 4.5 by adding some 0.1M hydrochloric acid solution.

Add the peptide solution to the gel suspension. Add 70 mg EDC and adjust the pH again to 4.5 by adding some 0.1M hydrochloric acid solution.

Incubate for 4h at room temperature. During the first hour of reaction, the pH decreases and it is very important to readjust the pH to 4.5 regularly by adding small volumes of 0.1M sodium hydroxide solution.

Centrifuge and wash the gel with 3 ml of 0.1M acetate buffer, 0.5M salt, pH 4.0 buffer and then 3ml of phosphate buffer, 0.5M salt, pH 7.4. Repeat this washing at least three times.

After the last centrifugation, discard the supernatant, resuspend the gel in 1 ml PBS, and transfer the gel in the column adapted to the FPLC instrument.

3. Conjugation of Peptide to EAH-Sepharose 4B Using MBS
This procedure is to be used for peptides that have undergone a MBS coupling to a carrier protein.

Buffer preparation: 50mM phosphate buffer, pH 6.0; dissolve 17.9g of disodium hydrogen phosphate dodecahydrate (Na2HPO4·12H2O) in 1 liter of water

Take 3ml of EAH-Sepharose 4B suspension, centrifuge, and discard the supernatant. Wash the gel three times by 2ml of 50mM phosphate buffer, pH 6.0. Discard the supernatant each time.

Resuspend the gel in 1 ml of 50mM phosphate buffer, pH 6.0

Dissolve 15 mg MBS in 1 ml of DME Use this solution within 1 h!

Add 140µl of the MBS solution and 500µl DMF to the gel suspension and let react for 30 min at room temperature under gentle stirring.

Centrifuge and discard the supernatant. Wash with 2 ml of PBS buffer, pH 7.4, centrifuge, and discard the supernatant. Repeat this washing three times and then resuspend in 1 ml of PBS buffer.

Dissolve 10 mg of peptide in 1 ml of PBS buffer, pH 7.4, and add this solution to the gel suspension. Incubate under gentle stirring for 3h at room temperature.

Centrifuge, discard the supernatant, resuspend in 1 ml of PBS buffer, pH 7.4, and transfer the gel in the column adapted to the FPLC instrument.

4. Affinity Purification of Antiserum
Buffer preparation: PBS high salt; dissolve 20g of sodium chloride in 1 liter of PBS buffer

100 mM glycine buffer, pH 2.5; dissolve 7.5 g of glycine in 1 liter of water

1M Tris buffer, pH 9.0; dissolve 121.14 g of Tris in 1 liter of water

1.5M guanidine hydrochloride: dissolve 143.30g of guanidine hydrochloride in 1 liter of water

Thaw the serum to be purified (we use generally 50 ml serum in one run), place it in a graduated cylinder, and add 1 g of sodium chloride per milliliter of serum.

Mount the column on the FPLC instrument and wash the column with 10ml of 1.5M guanidine hydrochloride followed by 10ml of PBS high salt.

Prepare a Falcon tube for antibody recovery and place 400 µl of 1 M Tris buffer in the tube.

Charge the serum slowly on the column (maximum flow rate should be 0.5 ml/min) and take care of the counterpressures that may arise (in case the counterpressure gets too high, decrease the flow rate). Recover the flow through of the column in the original serum tube. Never discard any liquids before making sure the affinity purification has worked correctly!

Once all the serum has been charged, wash the column with PBS high salt until the absorbance (measured at 280 nm) decreases to zero. Stop the recovery in the original serum tube as soon as the absorbance starts to decrease.

Elute the antibodies by passing the glycine buffer, pH 2.5, on the column and recover the antibodies (increase in absorbance) in a Falcon tube containing Tris buffer. Stop the recovery as soon as the absorbance gets back close to zero.

Measure immediately the pH in the Falcon and adjust to pH 7-7.5 if necessary (by Tris buffer or glycine buffer). Dialyze the antibody solution against 5 liters PBS buffer for 4h at 4°C; repeat the dialysis with 5 fresh liters of PBS.

Divide the antibody solution in working aliquots and store the aliquots at -20°C. If the antibodies are to be stored a long time, it is best to add 0.1% sodium azide as a preservative (except if it is planned to label the antibodies afterward, in this case 0.1% thimerosal as a preservative is preferred).

Wash the affinity column for 5min with guanidine hydrochloride solution, check the absorbance during this process, and recover eventual peaks in a Falcon tube. Dialyze immediately (see Section IV) against 5 liters PBS buffer (two times).

Wash the column for 10min by PBS, recover the gel material, and store it at 4°C after the addition of one equivalent volume of ethanol.

Quantify the antibody solution by measuring the absorbance of an aliquot at 280nm. One optical density unit corresponds to about 0.75 mg of pure antibody.

IV. PITFALLS
Depending on the sequence, peptides are sometimes difficult to get into aqueous solution. In case a peptide does not dissolve in water, dissolve a dry quantity in a minimum amount of dimethyl sulfoxide (DMSO); once it is dissolved completely, dilute by water to the appropriate concentration. The DMSO does not influence the coupling process.

Coupling may sometimes give rise to precipitates, depending on the peptide sequence. These precipitates should not be discarded for the antibody production; vortex the solution well before aliquoting the injection amounts to include the peptide-carrier aggregates. Some antigens are known to be phagocytosed even better than soluble material. In case the aggregates are too large, crush them first with a spatula to get them as fine as possible. Never freeze the affinity gels; these are best kept at 4°C in PBS buffer containing sodium azide or PBS/50% ethanol.

It may happen that the glycine buffer does not elute some high-affinity antibodies. These antibodies will come off the column during the guanidine hydrochloride wash. Dialyze immediately the recovered solution two times against PBS as described earlier, as the antibodies will generally still be active.

References
Huet, C. (1998). Production of polyclonal antibodies in rabbits. In "Cell Biology: A Laboratory Handbook" (J. Celis, ed.), 2rd Ed., Vol. 2, pp.381-391. Academic Press, New York.

Tam, J. P. (1988). Synthetic peptide vaccine design: Synthesis and properties of a high-density multiple antigen peptide system. Proc. Natl. Acad. Sci. USA 85, 5409-5413.