Those of us working in mouse or human immunology are spoiled for choice when it comes to fluorescently conjugated antibodies for flow cytometry. We have well validated clones and generally a wide range of fluorophore options to choose from. This isn’t always the case for people who work with other species or in other fields. Moreover, when you push the boundaries of the field, you will run into situations where you’d love to do flow but can’t find a commercially available conjugated antibody, at least not from the major vendors. In these situations, a good option is to conjugate the antibody yourself, and today I’m going to provide some tips for how to go about this.
In short, the best option for most people is going to be to use a small molecule fluorophore that you just mix with the antibody. This will give the most consistent results.
First, a word of caution: above I illustrate a hypothetical scenario in which you can’t find an antibody from your preferred vendors. In many cases, a web search will reveal an antibody from a handful of other vendors, oftentimes with indications that the antibody will work in every conceivable assay and can be provided in a dozen colours (usually variations on the Alexa Fluors). These are not usually good choices. From experience, I can say that you will spend a lot of time optimizing your conditions in order to get any specific signal (so you will absolutely need knockout controls to validate specificity), and that you can achieve better conjugation yourself using the tips below.
Right, on to conjugation. When we conjugate an antibody to a fluorophore, we are performing a chemical reaction, creating a covalent bond between the fluorophore and the antibody. Usually this occurs at primary amino groups via an amide bond. There are not many such sites on IgG molecules that are not masked by glycosylation. In an ideal situation, we would attach the fluorophore to the Fc region of the antibody, leaving the variable domain untouched to continue binding to our antigen. With most kits, particularly those that are easiest to use, we will have no control over where the fluorophore attaches. This means that it is possible to overconjugate the antibody, attaching so many fluorophores that it no longer functions.
Take for example this conjugation of 1.5ug of anti-mouse CD4 to Alexa Fluor 660 or mFluor Red 780 at variable amounts of fluorophore (indicated as volumes of a 10mM stock in DMSO). The resulting conjugates have been cleaned up and used for staining splenocytes. With Alexa Fluor 660, there is a decent range of fluorophore:antibody ratios that works, with both too little and too much producing inferior separation. With the Red 780, there's only one ratio that's giving a nice, clean separation. For each antibody there will be an optimal ratio to use for the conjugation reaction and, unfortunately, this will not always be the same. Oftentimes, we can get good results with the ratio around 10,000:1 fluorophore:Ab (0.6 to 2ul of 10mM fluorophore per 100ug antibody). That's a lot of excess fluorophore.
Okay, down to brass tacks. What works? For best results, you will need the following:
Antibody at >0.5mg/ml in PBS.
Standard concentrations of azide or thimerosal are okay.
Gelatin, BSA, other carrier proteins or glycerol need to be removed prior to conjugation. Gelatin is more difficult, so I recommend finding an antibody without gelatin in it if possible.
Centrifugal filters with a molecular weight cut-off between 10 and 100kDa, depending on what you need to remove.
1M Sodium bicarbonate.
Reactive fluorophore or biotin, preferably as a succinimidyl ester.
Sodium azide (1% stock solution) for long-term storage.
Protocol:
Check your antibody spec sheet for the buffer composition and concentration. If the antibody contains BSA or other proteins, remove the contaminating protein using an Amicon-Ultra centrifugal filter with an appropriate molecular weight cut-off. IgG is approximately 150kDa. BSA is about 66kDa, so a 100kDa filter would work. Pipette 100ug of antibody into the upper compartment of the filter, top up to 500ul with clean PBS and spin for 5min at ~12,000g. Repeat at least one more time. Each spin will remove up to 90% of the BSA while retaining around 95% of the antibody. The BSA concentration needs to be substantially lower than the antibody. Recover the antibody from the upper reservoir, rinse with a small amount of PBS to optimize recovery, and bring the volume up to 100ul (~1mg/ml) or 50ul (~2mg/ml).
Gelatin is a mix of proteins, some with high molecular weight. To remove gelatin, you can use the Melon Gel kit from ThermoFisher, but you will lose some antibody.
If your antibody concentration is below 0.5mg/ml, perform as many spins as needed to bring the concentration to around 1mg/ml. For this, you can use a filter with a cut-off as low as 10kDa if there is no protein to remove. Lower molecular weight cut-offs will slightly enhance the recovery, but will take a bit longer to concentrate down.
For each 100ul of antibody, add 10ul of sodium bicarbonate. This brings the pH up into a more effective range for the reaction. If your antibody is an IgM, do not add bicarbonate as it will denature the IgM. Instead, keep the antibody in PBS and leave the reaction overnight.
Add the reactive fluorophore and mix well. Suggestion: use 2 ul of 10mM stock of succinimidyl ester fluorophore per 100ug antibody.
Incubate at least 1hr at room temperature in the dark. The reaction can be left overnight will no adverse effects; it is unlikely that you will observe increased conjugation after a couple of hours. I usually leave it for about 4hrs.
Optional: add glutamine to neutralize any unreacted fluorophore. This is unnecessary if you either remove the fluorophore (next step) or wait a couple of days prior to using the conjugate.
Remove the excess unconjugated fluorophore. This step can be skipped if you are planning to only use the antibody for short surface staining protocols on lymphocytes. I recommend performing this step to get better staining, though. Place the reaction mixture in the upper compartment of the centrifugal filter (anything between 10 and 100kDa is fine). Top up to 500ul with clean PBS. Spin, discard flow-through. Repeat. Recover the antibody from the upper reservoir and rinse the upper reservoir with PBS to enhance recovery.
Adjust the antibody concentration with clean PBS. I usually assume the recovery is around 90% and store the antibody at 0.2mg/ml.
Add sodium azide (0.1% final) for long-term storage if the antibody will not be used in cell culture. For long-term storage, a tinted tube may be better.
Label the conjugate with the date, clone, catalogue number, fluorophore and any other relevant details.
Which options work really well?
Alexa Fluors. There are both succinimidyl esters (SE) and protein/antibody labeling kits available. The labeling kits are very easy to use, but don’t use the size exclusion columns to clean-up—it’s hard to get good recovery with those. Use the Amicon Ultra filters instead.
Biotium CF dyes. These are great, and appear to form the basis of many of the “new” commercial fluorophores coming onto the market. They are essentially improved versions of the Alexa Fluors, but in many cases the spectra are better for spectral flow cytometry. They are provided as easy-to-use Mix-n-Stain kits. Follow the protocol for those, but use the Amicon Ultra filters instead; the ones in the Mix-n-Stain kits will fail under high g-load.
Biotin. Great option. You can get biotin on flexible linker chains and it’s small enough that you can attach many biotins per antibody with little risk of damage. This then allows you to use any fluorescently conjugated streptavidin on the market, amplifying your signal. And yes, biotininylated antibodies work fine with overnight intracellular staining.
AAT Bioquest’s mFluors. Available as succinimidyl esters in many colours. Some are rather dim, including the UV excitable options. I particularly like mFluor Vio 610, which is pretty bright and gives you an option like BV605.
What is not so great?
PE, PerCP and APC-based conjugation kits. These protein fluorophores are larger than IgG molecules, so they cannot be easily removed using molecular weight filters. As such, you will always have free fluorophore in your antibody, increasing the background, particularly with longer staining or on myeloid cells. Also, you can only usually get one fluorophore attaching to each antibody due to the size; as a result, you’re much more likely to have unconjugated antibodies present, which will block binding of the fluorescently tagged ones. These will work, but they will not be optimal. If you want PE-level signal, you’re better off using biotin and PE-streptavidin.
Tandems of the above. Kits based on PE or APC tandems are, in my experience, very susceptible to tandem breakdown. Again, they’ll work, just not the best.
Most kits trying to target a very large fluorophore specifically to the Fc region of the antibody. Examples include certain PE kits, some using quantum nanocrystals, and others using DNA-based fluorophores. These are more difficult to use reliably. In my experience, they either produce great results, or a completely useless antibody that sticks non-specifically to everything. I don’t like taking that risk, as it’s at least a third of the time that I got complete failure, even with technical assistance from the company specialists. There is a reason the companies charge a lot for antibody conjugates--it’s not always easy.
This kit, for instance, produced a non-functional conjugate. The staining on human PBMCs gave a background shift and a no separation.
Testing the antibody on compensation beads (Ultra Comp Plus) revealed that the antibody was indeed conjugated and bound to the positive beads. My guess is that the antibody was denatured or otherwise damaged during the conjugation process.
Looking to the future, we can expect a new class of bright, clean fluorophores to become available for use in conjugation kits. Really excited for this, but I don't think I'm allowed to say yet which these are.