Wittrup: Protein conjugation with fluorophores
Directly labeled antibodies can be of great use and convenience over the use of secondary antibodies. For example, when studying rapid dissociation of an antibody, directly conjugating a fluorophore to the antibody of interest avoids the need for a second incubation step and multiple washes in between. It can also reduce background significantly, especially when the alternative is an elaborate labeling scheme (e.g. his-tagged scFv, mouse anti-his, and anti-mouse PE).
For labeling small amounts of protein (100 μg to 2 mg), the Molecular Probes protein labeling kits are very convenient. Three different kits are available: protein labeling kits, monoclonal antibody labeling kits, and microscale labeling kits. Protein labeling kits are for 1 mg (3 mg per kit) of > 40 kDa proteins. Monoclonal antibody kits are optimized for 100 μg of IgG (0.5 mg per kit), but they can be used for any protein over 30 kDa. Both of these kits offer less freedom than the microscale kits since you cannot choose the molecular ratio (MR) of dye to protein in the reaction. Microscale labeling kits are ideal for labeling small amounts (20-100 μg) of scFvs and other small proteins between 12 and 150 kDa (product specifications). Depending on the dye to protein ratio, often 2-4 times this mass of protein can be labeled (300 μg to 1 mg per kit).
The dyes are also available for purchase individually. Unfortunately, they are delivered in a single aliquot, so unless you are labeling one large sample of protein, it has to be split up. After the dye is dissolved, it can start to hydrolyze, so splitting into aliquots is not straightforward. Talking with a tech rep, he did indicate that it could be dissolved in water, split up, and dried using a roto-vac, although I have not tried this. The cost of each is roughly $40-70 (depending on reactivity, which varies) to label 100 μg of protein when ordering in bulk versus $100 to label 100 μg with the kits. Considering the loss in dividing the bulk Alexa, variability in reactivity, and additional materials required for separation, the savings is not substantial.
The protocols provided with the kits work very well. The information below can be considered a supplement to these protocols.
For labeling scFvs with the microscale labeling kit, the following degrees of labeling have been obtained:
Antibody MR DOL
4M5.3 15 ~1.7
4M5.3 30 ~2.2
sm3E 15 ~2.5
Yields are typically in the 70-90% range, unless there is a problem with the purification. The limiting reagent in the kits (if used as provided) is the separation gel. (The tubes are only able to separate 50 μL each, and there isn’t enough resin to separate 100 μL for each of the 3 labelings.) To maximize the amount of protein labeled with a kit, additional separation materials can be ordered. BioGel P-6 Gel fine (from BioRad) can be used to supplement the kits. (Dissolve 5 g dry weight gel in 43.3 mL PBS and allow to hydrate for > 4 hours at RT.) Alternate spin columns can be used, such as those from Corning (Costar Spin-X centrifuge tubes) since multiple uses can clog the filter and give poor separations (see below).
For monoclonal antibody labeling kits: For monoclonal antibody labeling kits, I’ve had better luck by incubating overnight at 4C after the 1 hr RT incubation. There is no flexibility in changing the MR, so you have to optimize the labeling amount with longer times instead of higher ratios.
For the microscale labeling kits: When purifying the reaction mixture, make sure all the buffer is out of the gel before adding the rxn mixture. Excess buffer will increase the final volume (diluting your protein), give a higher chance of free dye flow through, and possibly cause poor yield. A typical reason for excess buffer is poor flow through the filter (which often occurs when reusing the tubes, since the filters get clogged). This also means that not all the reaction mixture will flow through during the separation spin-down, resulting in more protein staying in the column.
I’ve had success labeling proteins down to 0.3 mg/mL if they weren’t concentrated enough just by increasing the MR (e.g. using 15 instead of 12).
General: Always perform TLC (thin layer chromatography) on the labeled protein. The FluorS imager is suitable for imaging both fluorescein/488 dyes and Texas red/594 dyes. The procedure only uses 1 μL of sample, takes 15 minutes (only 5 of which is hands on), and is very cheap (~$1 for running 4 samples), but it gives a clear picture of whether there is free dye in the solution, which can have a major impact on experiments.
If separation is not complete, the dye can be re-separated quickly using the spin columns. If the yield was poor or the antibody was diluted too much due to extra buffer coming through, I’ve successfully removed the remained using the Eluta tube dialysis tube provided with the kit just dialyzing at 4C overnight in PBS.
When using bleachable dyes, I’ve had better luck with fluorescein ester rather than FITC (fluorescein isothiocyanate).
Mixing during the reaction isn’t essential, but it will increase the labeling efficiency.