Editorial Article
Wednesday February 10, 2010
by Jeffrey M. Perkel
Among molecular biologists and protein chemists, few techniques get as much play as the Western blot. Used to assess the presence, abundance, or modification state of one or more proteins in a complex mixture, the technique is the protein equivalent of Southern blotting, and an essential component of the modern laboratory toolbox.
The classic Western blotting protocol comprises seven steps: gel electrophoresis, blotting to a membrane, blocking, primary antibody incubation, secondary antibody incubation, and development/imaging. In broad strokes, it's simple enough that even the newest lab tech can pull it off, and for most people, standard protocols suffice (see, for instance, Abcam's "Western blotting – a beginner's guide." Yet as with everything else in the lab, there are no absolutes, and what works for one lab or researcher may not work for another.
1. Strategy
Before you can begin your Western, you need to think about your detection strategy, as this will determine what equipment and reagents you need downstream.
There are three basic approaches: colorimetric, chemiluminescent, and fluorescent, the choice of which will determine what kind of secondary antibody you use. Colorimetric and chemiluminescent methods both employ enzyme-conjugated secondary antibodies; with the former, the enzyme facilitates the precipitation of visible bands onto the blot itself, with the latter, visible (but transient) light is produced. Fluorescent methods use fluorescently tagged secondary antibodies.
Colorimetric methods require no special equipment, but are also the least sensitive. The advantage of fluorescence imaging is its stability. "It's as bright the minute you turn it on as it is an hour later, a month later, or a year later," says Sally Weldon, senior molecular biologist at LI-COR. "Chemiluminescence is done usually within 30 minutes." On the other hand, fluorescence also requires a dedicated imager, whereas chemiluminescence can be imaged with x-ray film.
2. The membrane
Western blotting membranes come in two basic flavors, nitrocellulose (e.g., GE Healthcare's Hybond ECL) and polyvinylidene difluoride (PVDF) (e.g., Millipore's Immobilon membranes). For most applications, the two are equivalent, though Sahar Sibani, Western blotting product manager at Millipore, notes that PVDF has a higher binding capacity than nitrocellulose, "which means you need less protein in the gel to get it bound to the membrane." (PVDF, being somewhat hydrophobic, also must be pre-wet in methanol, whereas nitrocellulose may be wet in water.)
Membranes typically are available with 0.2 micron and 0.45 micron pores. The larger-pore filters are sufficient for most applications, but for small proteins (less than 20 kD), use a small-pore membrane, advises David Chimento, antibody validation group leader at Rockland Immunochemicals.
Chimento, whose team runs up to 150 Westerns per week, also suggests that, for very small proteins such as cytokines, users stack three or four membranes on the gel to ensure that at least one will contain the protein of interest. Alternatively, decrease transfer time, or change the transfer buffer (less methanol reduces pore size, he says). For very large proteins (>350 kD), Chimento advises skipping the normal pre-transfer soak in transfer buffer (which removes SDS), because the SDS can actually help those proteins transfer more efficiently.
Another factor to consider is the downstream detection method. Both nitrocellulose and PVDF membranes are compatible with colorimetric, chemiluminescent, and fluorescent detection, but some PVDF membranes autofluoresce (limiting their utility for fluorescent detection). Millipore's Immobilon FL and GE Healthcare's Hybond LFP are designed specifically for fluorescent applications. When in doubt, says Weldon, "just take a piece of filter, wet it, and scan it to see what kind of background it has."
3. Blocking
Following transfer, membrane regions that don't contain protein must be "blocked" to reduce background.
Blocking reagents include everything from store-bought non-fat dry milk (5% in phosphate- or tris-buffered saline plus detergent), to purified proteins like bovine serum albumin or casein (e.g., LI-COR's Casein Blocking Buffer), to synthetic reagents (e.g., Millipore's Bløk Noise Canceling Reagent).
Though generally interchangeable, blocking reagent selection does sometimes matter (see #8, below). "If you are sure your Western blot should work and you tried one blocking condition, an easy thing to do is try a different blocking reagent, and that often will solve the problem," Chimento says.
Some milk-based blockers, for instance, contain IgG and biotin that can cross-react with goat-derived secondary antibodies, says Weldon. Other commercial formulations contain the preservative azide, which can inhibit horseradish peroxidase (HRP), a common tag on secondary antibodies for chemiluminescent detection.
Blocking choice is especially important when probing with phospho-specific antibodies, Chimento says. For instance, dry milk preparations sometimes contain proteins that can either increase background or dephosphorylate the proteins on the blot, he says. (And of course, when using phospho-specific antibodies, use tris-buffered saline, rather than phosphate-buffered saline solutions, as the latter can increase your background.)
LI-COR's application note, "Good Westerns Gone Bad" dramatically illustrates the difference that blocking reagents can produce.
4. Primary antibody incubation
The primary antibody is the reagent that actually detects your particular protein of interest; the secondary then detects the primary.
Commercial primary antibodies are commonly sold at 1 mg/ml concentrations. But as each antibody is different, the conditions under which you use them should be individually optimized (see #8, below). Those conditions include antibody concentration, detergent concentration, and incubation time.
A good starting point, says Chimento, is a 1:1000 dilution for polyclonals, and 1:5000 for monoclonals, in blocking buffer containing 0.5% to 1% Tween-20 (a detergent). Tween, he says, blocks hydrophobic interactions; "in the absence of that, you get random, non-specific binding."
One hour incubation at room temperature is standard, Chimento says, but others prefer overnight at 4°C. The latter conditions are preferred if antibody affinity is poor, and if the one-hour room-temperature incubation doesn't work, "there's nothing to lose" by switching to overnight at 4°C, he says. "The downside is that you run the risk of increasing background because the antibody has more time to form weaker interactions."
5. Secondary antibody incubation
Wash before and after the secondary antibody incubation. Standard conditions are three, five-minutes rinses in PBS or TBS containing detergent. Again, Chimento recommends tweaking the Tween concentration. "You want more Tween than during blocking," says Chimento, "you are trying to drive off non-specific interactions." Start with 1% Tween-20, he advises; if background is high, increase to 2% or even higher.
Just as every primary antibody should be optimized, so too should every secondary antibody (see #8, below). Each antibody has a different affinity for its target, and each batch may contain different amounts of conjugated detection reagents (such as fluorophores or HRP).
A good starting point, Chimento says, is a 1:20,000 dilution for 30 minutes at room temperature (though he typically dilutes to 1:40,000 or even 1:80,000 to decrease background). Incubation time can be increased to an hour for very low abundance proteins.
Rinse briefly in buffer without detergent just prior to blot development.
6. Development and imaging
Of the three detection modalities (colorimetric, chemiluminescent, and fluorescent), says Ryan Short, imaging market manager at Bio-Rad Laboratories, chemiluminescent "is by far the most common application."
Yet chemiluminescent detection reagents are not identical, Chimento notes. Some produce a high initial burst that rapidly decays; others have much longer half-lives, which is useful for low abundance proteins that require longer exposure times, or simply if the walk from the lab to the darkroom is lengthy.
Reagents also differ in sensitivity. GE Healthcare's ECL Plus Reagent is more sensitive than the standard ECL Reagent itself, and ECL Advanced offers even greater sensitivity for especially expensive antibodies. Millipore's soon-to-be-released Luminata HRP reagents (which will come premixed, rather than in the usual two-bottle formulation) will be available in three flavors, according to Sibani: Luminata Classico for 10-20 pmol; Luminata Crescendo for 1-10 pmol; and Luminata Forte for mid-femtomole range proteins.
Fluorescent detection is useful for detecting two proteins on one blot, especially if those proteins are approximately the same size (for instance, total Akt protein and phosphorylated Akt). In that case, each target protein must be targeted with a primary antibody from a different animal species (such as mouse, rat, or rabbit), so that the secondary antibodies can distinguish them.
LI-COR offers IRDye-labeled secondary antibodies in each of two near-infrared channels (680 or 700 nm, and 800 nm), while GE Healthcare provides its ECL Plex reagents in both Cy3- and Cy5-labeled forms (570 nm and 670 nm, respectively). Rockland Immunochemicals offers DyLight™ secondary antibodies available in UV to near IR (488, 549, 649 nm).
If you opt for chemiluminescent or fluorescent detection, you'll need to image your results. Fluorescent detection requires an imager, such as LI-COR's Odyssey, Bio-Rad's PharosFX, or GE's Typhoon FLA-9000; chemiluminescence, though, offers a choice: imager or film.
Film, Short says, is fast and sensitive, but also expensive. More to the point, film is a poor choice for quantification because it saturates very quickly (i.e., it has a small linear dynamic range of ~1.5). In other words, it doesn't take much light for film to overexpose, at which point, the difference between bands becomes impossible to measure. Imagers, on the other hand, have much wider dynamic ranges (~3-5 orders of magnitude). With accompanying software, he says, users can "easily quantify a low- and high-expressing signal in the same image on the same blot."
"Generally, camera-based systems are very good at chemiluminescent detection, but laser-based systems are not," says Short. That's because chemiluminescent detection requires signal integration over time, something laser-based scanners, with their constantly moving scan heads, cannot do. Chemiluminescence-capable imagers include GE's ImageQuant LAS 4000 and Bio-Rad's ChemiDoc XRS+.
7. Stripping
Sometimes it's necessary to probe a blot more than once – say, to detect two variants of the same protein, to conserve reagents, or to compare one protein's abundance with another, control protein.
Standard stripping methods are sufficiently harsh that they can remove the transferred proteins as well as the antibodies from the blot, which can add variability to an experiment, observes Åsa Hagner-McWhirter, application research scientist at GE Healthcare. Indeed, Abcam's stripping protocol warns, "It is not advisable to make quantitative comparisons of targets probed before and after stripping since the procedure removes some sample protein from the membrane. For the same reason, a stripped membrane should not be probed to demonstrate the absence of a protein."
Though Hagner-McWhirter advises circumventing the issue altogether with multiplexing fluorescent antibodies, some users have no choice. In that case, a number of relatively gentle commercial stripping reagents are available (e.g., Millipore's ReBlot, available in both mild and strong formulations).
8. Optimization
For successful Western blotting, optimization is key. One simple choice is to test conditions using dot blots. Hagner-McWhirter, however, cautions against that. "If you do that, you don't see the non-specific signal because everything is in one spot, and you don't see the background," she says.
Instead, Hagner-McWhirter suggests loading each well of a gel with identical control samples, blotting the gel to a membrane, cutting the membrane into strips, and treating each with a different antibody or antibody concentration, buffer, blocking reagent, and so on. "I would recommend doing this for every antibody," she says. "You can get something reasonable if you don't, but if you want the best results, you should do it."
Alternatively, you can use a multiplexing system such as LI-COR's MPX Blotting System , which generates 24 independent incubation channels on a single blot, each of which requires just 150 ul of reagent.
Whichever route you choose -- and it's clear, there are many to choose from -- with a little careful planning and attention, you can arrive with a clean, informative and optimal Western blot.