Conjugation of Monoclonal Antibodies: Comments on Roederer Protocols
Mario Roederer's excellent website describes almost everything you need to know about conjugation of monoclonal antibodies to fluorophores and biotin; the protocols could even be adapted to conjugate antibodies to enzymes such as HRP and alkaline phosphatase. The protocols found there are robust and well described. This page may be regarded as a commentary on that definitive site.
Contents of this page
- Description of dyes
- Description of reactive chemistries
- Alternatives to dyes listed on the Roederer site
- Desalting versus dialysis: comments on buffer exchange
Description of dyes
It is helpful to divide the fluorophores used in flow cytometery into three categories:
- Small molecule fluorophores. Fluorescein, often imprecisely referred to as FITC, is the most common small molecule fluorophore. Other small molecule fluorophores include Pacific Blue, Cascade Blue, the cyanine dyes (Cy5, Cy5.5, and Cy7), and all of the Alexa dyes, etc.; biotin can also be considered an honorary member of this group.
- Large fluorescent proteins. The most commonly used fluorescent proteins are the phycobiliproteins (see also Prozyme description) R-phycoerythrin (R-PE) and allophycocyanin (APC) (these can be obtained from Molecular Probes, or from Prozyme—at a more competitive price). Tandem dyes use small molecule fluorophores, covalently linked to phycobiliproteins, and emit light at the longer wavelength of the small molecule fluorophore, due to FRET (fluorescence resonance energy transfer); examples include PE-TR, PE-Cy5, PE-Cy5.5, PE-Cy7, APC-Cy7, etc. PerCP and it's tandems are also in this category. Until recently, PerCP was available only in the form of conjugated antibodies from Becton Dickinson; it now appears to be available from Prozyme.
- Quantum dots.
The dyes used in flow cytometry are limited to those which efficiently absorb light at the wavelengths of the common available lasers. For example, at the Emory Vaccine Center, all of our cytometers have lasers emitting light at 488 and 633 nm; beyond this, the MoFlo also has a Krypton laser that normally runs at 350 nm (it can be tuned to other wavelengths, but this takes considerable time and we don't do it often), while the new FACS Aria and LSR-II instruments have a violet laser that runs at 405 nm, but no UV laser. The UV line of the Krypton laser is only marginally useful for immunophenotyping—it's much better for measuring DNA content using DAPI and Hoechst dyes, or for measuring calcium fluxes with Indo-1—while the violet laser provides much better excitation of Cascade Blue and it's nearly identical analog, Alexa 405, which is extremely useful for immunophenotyping (and of Alexa 430, which is suboptimal due to it's broad emission spectra and weak fluorescence). It is also very useful for excitation of quantum dots.
Tandem dyes greatly extend the utility of flow cytometry by permitting efficient excitation of a large number of small molecule fluorophores, if (1) they are covalently attached to a phycobiliprotein that can be excited by an available laser (e.g. the 488 or 633 lasers), and (2) they can act as a FRET acceptor. Roederer's website describes the preparation of the tandem dyes PE-Cy5 (aka, CyChrome by Pharmingen, TriColor by Caltag, PC5 by Beckman Coulter), PE-Cy7, and APC-Cy7. All of these use the "cyanine" family of dyes, available from Amersham Biosciences as either mono-reactive or bis-reactive dyes (Roederer uses the bis-reactive dyes). A relatively more recent addition to this family of dyes, Cy5.5, is also used as both a tandem to PE and to APC, though it's use as a tandem to APC may be superseded by the utility of direct conjugates to the small molecule dyes Alexa 680 or Alexa 700. Many of the Alexa dyes can also be in tandems (e.g. PE-Alexa594, PE-Alexa647, PE-Alexa680, PE-Alexa700, PE-Alexa750, APC-Alexa750), as can Texas Red (e.g. PE-TexasRed). Preparation of the Alexa tandems uses the same principles as for Roederer's protocols for preparation of cyanine dye tandems (e.g. PE-Cy5), with only modification of the buffer used for coupling of the dye to the phycobiliprotein. This will be described below.
The structures, excitation, and emission spectra of the dyes can be obtained from the Amersham Biosciences website (PDF).
Description of reactive chemistries
Small molecule fluorophores are usually conjugated to antibodies via free amino groups (lysine side chains and amino termini) that react with amine-reactive moieties on the reactive dyes. The most common amine reactive groups are described in the following table:
Table 1. Amine Reactive Groups |
||
|---|---|---|
| Chemistry | Scheme or Structure |
Examples |
| Succinimidyl Esters |  |
Pacific Blue; Alexa 488; |
| Isothiocyanates |  |
FITC |
| Acetyl Azides |  |
Cascade Blue |
| Figures borowed from the Molecular Probes website; click on each figure to go to the original page. | ||
A given fluorophore may be available as derivatives with a number of reactive groups. Table 2 lists several derivatives of fluorescein and other dyes with similar spectral properties.
Table 2. Reactive Derivatives of Fluorescein and Dyes with Similar Spectral Properties |
|||||
|---|---|---|---|---|---|
| Name | Functional Group |
Structure |
Molecular Probes Cat# |
Price and units | Comment |
| FITC (fluorescein-5-isothiocyanate) | isothiocyanate |
 |
F1906 |
10 x 10 mg $67.00 |
The old standby |
| 5-FAM, SE (5-carboxyfluorescein, succinimidyl ester) | Succinimidyl ester |
 |
C6164 |
10 mg $140.00 |
More stable conjugates than FITC; 6-FAM and mixed isomer preps also available |
| 5-SFX (6-(fluorescein-5-carboxamido)hexanoic acid, succinimidyl ester) | Succinimidyl ester |
 |
5 mg $94.00 |
Aliphatic spacer arm | |
5-SFE (fluorescein-5-EX, succinimidyl ester) |
Succinimidyl ester |
 |
10 mg $94.00 |
More hydrophobic spacer arm | |
| Oregon Green® 488 carboxylic acid, succinimidyl ester | Succinimidyl ester |  |
O6147 | 5 mg $151.00 |
Less senstitive to pH and homoquenching than FITC |
| Alexa488-SE (Alexa Fluor® 488 carboxylic acid, succinimidyl ) | Succinimidyl Ester |  |
1 mg $205.00 |
Expensive;superior to FITC in terms of less senstivie to pH and bleaching | |
The small molecule dyes (FITC, Cascade Blue, the Alexa dyes), all have functional groups that react with amines on proteins; this means that they are quite selective for lysine side chains, as well as for the N-termini of the protein subunits. Reactivity of the amines requires that they be deprotonated, and so these reactions are done at pH values between 8.3 and 9.5 (depending upon the stability of the dye). Reactions with FITC are done at pH 9.5, while reactions with N-hydroxysuccinimidyl esters are commonly done at pH 8.3-8.4; Roederer recommends pH 9.0 for reactions with the bis-reactive Cy5 dyes. We have performed conjugations of Alexa dyes to antibodies and phycobiliproteins at pH 8.4. The variations in pH may be due to historical precedent, as opposed to careful optimization.
Phycobiliproteins are conjugated to antibodies using heterobiofunctional reagents that have one functional group that reacts with amines (an N-hydroxysucciminidyl ester) and one functional group (a maleimide) that reacts with free thiols (the reduced form of cysteine). In this method, which is sometimes called "reductive coupling", the phycobiliprotein is modified with the reagent SMCC (or a related family member), which reacts with amine on the phycobiliprotein; the unincorporated SMCC is removed, leaving a PE that now contains thiol-reactive maleimide groups. In a separate reaction, the interchain disulfide of the antibody is selectively reduced by treatement with DTT, which is rapidly removed (note that strong non-covalent interactions keep the chains of the antibody together). The reduced antibody, which contains free thiols, is then mixed with the SMCC-treated PE, and after allowing this reaction to proceed, the residual free thiols are blocked with N-ethylmaleimide (NEM).
An alternative to the reductive coupling approach is to use protein modification reagents that add a thiol to the protein; this approach would be useful for attaching phycobiliproteins to streptavidin, which does not contain available cysteines.
Alternatives to dyes listed on the Roederer site
A number of new dyes have been developed since Roederer's website was written. In particular, Molecular Probes has introduced the Alexa series of dyes. Some of these may be directly excited by common laser lines and are useful as direct conjugates to antibodies (e.g. Alexa 405, 430, 488, 647, 680, 700); others are useful in tandem with either PE (Alexa 594, 610, 647, 680, 700, 750) or APC (750) (note that APC tandems can also be prepared to Alexa 680 and Alexa 700, and might be brighter than the direct conjugates, but this is probably outweighed by the ease of preparation of the direct antibody conjugates). Some of these dyes are almost exactly spectrally equivalent to older alternatives, as in the table below.
| Older dye | Alexa equivalent | Comment |
| Cascade Blue | Alexa 405 | Same fluorophore, different reactive chemistry |
| FITC | Alexa 488 | Alexa 488 may be brighter and is less pH sensitive, but it is considerably more expensive |
| Texas Red | Alexa 594, Alexa 610 | Alexa dyes might give lower background binding, and this might vary according to cell type. |
| Cy5 | Alexa 647 | |
| Cy5.5 | Alexa 680 | |
| Cy7 | Alexa 750 |
The Alexa dyes are all available as amine-reactive NHS esters. Conditions for conjugating them to antibodies are provided in a PDF document by Molecular Probes. The recommended buffers are also appropriate for conjugation to phycobiliproteins, although we have not subjected them to optimization.
We have prepared conjugates of Alexa 647, 680, 700, and 750 to PE and of Alexa-750 to APC. We attempted to prepaer a conjugate of Alexa-610 to PE, but the coupling efficiency was poor. This observation was confirmed by Molecular Probes, who suggested that we attempt preparation of PE-Alexa594 as a substitute; we have not attempted this yet.
Desalting versus dialysis: comments on buffer exchange
The Roederer protocols frequently call for exchange of one buffer for another. For example, if you'd like to label an antibody with an amine-reactive small molecule and your antibody is stored in a buffer containing primary or secondary amines (e.g. Tris) or sodium azide, you must exchange that buffer for the amine-free carbonate buffer at pH 8.4 (or 9.5 for the FITC conjugation).
There are three principle methods for buffer exchange of a protein:
- Dialysis
- Desalting on a gel filtration column
- Extensive diafiltration using centrifugal ultrafilters
With the exception of the buffer exchange following selective DTT-mediated reduction of the interchain disulfide that is used for coupling of antibodies to maleimide-activated phycobiliproteins, any of the protocols can be used; for separation of DTT-reduced antibodies from DTT, desalting columns must be used because they are the most rapid, and speed is required to minimize oxygen-mediated reoxidation of the liberated thiols.
How does one choose from among the three techniques for buffer exchange? The relevant parameters are (1) speed, (2) convenience, (3) yield, (4) ease of processing multiple samples at one time, and (5) to a lesser extent, price. My comments below are based upon anecodotal experience and are subject to change.
Recently, we prepared 12 distinct PE tandem dyes at one time (4 dyes, at three different dye:PE ratios). At the end of the incubation of the dyes with the PE, we separated the unincorporated dye from the modified PE by desalting on PD-10 columns (our reaction volumes were approximately 1 ml, which is at the low end of what is appropriate for PD-10 columns). We observed separation between the purple tandem and the blue free dye, but it was not complete, and we were forced to take a conservative cut of the fraction containing the modified PE. This presumably contributed to a poor yield of approximately 33%.
In a subsequent preparation of PE-Alexa700, we employed dialysis for this separation, and our yield was significantly better (80-100%). Comparing the two protocols, we are willing to take the additional elapsed time required by dialysis, if it reliably provides better yields. At this point, I recommend that we use dialysis for all buffer exchange steps, except for the removal of DTT as indicated above.
It is possible that we could further optimize recovery of the gel filtration step by using slower flow rates provided by higher resins (for example, the PD-10 columns use "medium" G-25 which provides excellent speed at the expense of tighter separations; it could be replaced by "fine" or "superfine" grade G-25, but these are not available in convenient, disposable, pre-packaged columns, and must be purchased in bulk). This might be worth some investigation.
Finally, we have not fully tested the efficacy of extensive diafiltration using centrifugal ultrafilters (though we have effectively used it in a pinch). It is not clear how efficiently this process will remove small molecules, and if the actual removal approaches theoretical calculations. However, in my opinion it is the easiest method to use if you must process a large number of samples in parallel (>= 8).