All Is Not Quiet on the Western Front

A grab bag of advances is making Western blots faster, more sensitive, and more reliable.

By | May 1, 2015

COURTESY OF ADVANSTA INC.If you don’t look too closely, Western blots are seemingly the same slog as when they were first described in 1979. You separate proteins by size (or charge) using gel electrophoresis, transfer them to a membrane, and probe the membrane using antibodies to your particular protein. Westerns have been a staple in protein research for many years, allowing researchers to identify, and sometimes semiquantify, proteins within tissues and cell cultures. It’s easy to take them for granted. At the same time, the familiar blots are often the subject of scrutiny, fodder for fraud, and the reason behind some research retractions.

Western blotting is moving in the right direction, though. In recent years, new reagents and instruments have made Westerns more sensitive and have condensed some of the main steps, namely electrophoresis, blotting, and visualization. Although many researchers still use chemiluminescence detection and X-ray film, the introduction of digital fluorescence imaging has boosted the technique’s sensitivity and reliability and pushed it into the “quantitative” rather than semiquantitative realm. New advances are also making it possible to study proteins in single cells, or to probe limited or precious samples—and to automate some parts of the process. And users are taking greater care in addressing the issue of reliable controls.

The Scientist spoke with researchers and companies working to make Westerns a little easier and more reliable. Here’s what we found.

Background: The past few years have seen great strides in measuring DNA and RNA in individual cells. By comparison, protein measurements have lagged behind. Poor antibody quality is usually the culprit, limiting how accurately researchers can measure protein levels in single cells using flow cytometry and immunocytochemistry. And separating the proteins, as done in Western blotting, is a must, but hasn’t been possible to do using single cells until recently.

Approach: Amy Herr’s bioengineering group at the University of California, Berkeley, has developed a microfluidic device they call the scWestern (for single-cell Western) that can perform Westerns on about 2,000 individual cells in less than four hours. The scWestern consists of thousands of microwells (20 μm wide by 30 μm deep) etched into a thin polyacrylamide gel coating atop a glass microscope slide.

BASED ON NAT METHODS, 11:749-55, 2014, REDRAWN WITH PERMISSION FROM AMY HERRThe cells settle into individual microwells, where the scientists lyse them and apply an electric field. Lysates migrate through the walls of the microwell and into the photoactive gel, where the proteins separate by size. They are locked into the gel using ultraviolet light and probed with antibodies directly. Herr’s group is able to analyze several proteins, in part by stripping the gels and applying different probes—up to 15 times.

Herr sees this technology as particularly important for measuring different protein isoforms and proteins that form complexes with other proteins or molecules.

Getting started: Although the technique is new, scientists outside of Herr’s lab have, with some guidance, replicated the entire process, all the way from fabricating the devices to running the assays (Nat Methods, 11:749-55, 2014). New graduate students in Herr’s lab were able to get it down within a month, she says.

Herr’s technology has led to the creation of the spinout Zephyrus Biosciences, which is developing a benchtop instrument that automates much of the handling. The company plans to launch by the end of next year.

Considerations: For now, the method has two limitations. One is that after the cells are lysed, proteins spill out of the microwells, resulting in losses of up to 40 percent. The second is that the separation resolution is subpar; researchers cannot separate proteins that are less than 50 percent different in weight. Herr’s group is working on solving both of these problems. “I think we have some exciting results,” she says.

Microwestern arrays, including Herr’s “μWesterns” (a precursor to her scWestern that uses larger wells), offer a gain in separation resolution for researchers who are willing to work with cell populations rather than single cells. The microwestern approach, which prints cell lysates on 96 tiny gel blocks arranged in the dimensions of a 96-well plate, is documented in a detailed series of YouTube videos ( at the University of Chicago’s Microwestern Array Core (MWAC) Facility website. The technique is able to resolve proteins that are at least 75 kDa different in size, if both proteins are more than 200 kDa. The technique requires the use of a piezoelectric dispenser, which can be quite pricey, says director Sam Bettis.

Costs: The price of the benchtop instrument is TBD; fabricating your own scWestern costs a few thousand dollars if you purchase most materials fully prepared. The microwestern array alternative costs $1,800 per run (examining six cell lysates with 96 antibodies) through the University of Chicago’s MWAC.

COURTESY OF JILLIAN SILVABackground: From start to finish, a traditional Western blot takes about two and a half days, which makes having to re-run it frustrating, especially if you have a lot of samples. “I do a lot of [Westerns], and trying to get through all of the samples was exhausting,” says Jillian Silva, a postdoctoral researcher who studies melanoma signaling pathways in Martin McMahon’s group at the University of California, San Francisco.

Approaches: Silva modified three steps in her Western protocol to cut her total run time down to one day. First, she replaced her SDS-PAGE gels with a Bis-Tris gel system, which takes 35 minutes to run (as opposed to 1.5–2 hours for SDS-PAGE) and has the benefit of boosting sensitivity. Second, she purchased a rapid blotting machine (iBlot Dry Blotting System, Life Technologies), which transfers protein to the membrane in seven minutes. Using the old method, this step took one to two hours.

The third—and most important—step, Silva says, was to move from chemiluminescence to fluorescence detection. LI-COR’s Odyssey series uses lasers that excite the membrane in two infrared wavelengths, 700 nm and 800 nm, allowing Silva to measure two different proteins on the same membrane. “That helped a lot because I could run all these different proteins at one time, whereas [before] I was running a gel for each of those proteins,” she says.

Getting started: Silva has taught several other labs her protocol, which is available in the Journal of Visualized Experiments (e51149, 2014). She finds that people are initially hesitant to try fluorescence imaging, but once they do, they don’t go back. Getting up and running takes some tweaking of, for example, blocking buffer amounts and antibody incubation times, but she says it’s well worth it for the higher speed and sensitivity.

Costs: $14.50 for a Bis-Tris pre-cast gel; $1,996 USD for the iBlot 2 Gel Transfer Device; $30–60K for a LI-COR Odyssey Infrared Imaging System. Prices vary by region.

Considerations: The time savings come with an added cost, however. Silva estimates about $20 extra per experiment for the Bis-Tris gel and iBlot consumables. But you can account for the time saved, either in personnel costs, animal maintenance costs, or both. Also, if you opt for fluorescence and digital imaging, you won’t have to pay for chemical developer and fixer, vendor maintenance of film processing machines (about $324/year for a user in Silva’s department), and, of course, film.

TAKE THE WHOLE LANE: Researchers are realizing that individual housekeeping proteins, typically used as controls because they were thought to be consistently expressed, can vary in their amounts across different tissues and disease states. Analysis of the total protein concentration within a sample (highlighted in the yellow box) is a better reference for quantification of proteins.IMAGE COURTESY OF SAMANTHA EATON, THE WISHART LABORATORY, THE ROSLIN INSTITUTE, U.K.Background: To help quantify proteins of interest, researchers typically compare them to single proteins also present in the lysate, such as GADPH, β-actin, or tubulin, which are thought to occur at consistent levels across cells and tissues and over time.

However, recent evidence suggests that levels of these so-called loading controls can vary across disease states and in different places within the same tissue, and may skew results by up to 20 percent (PLOS ONE, 8:e72457, 2013). “Many of the proteins that people commonly use as loading controls . . . are not stable at all, and they actually change in a broad range of different neurodegenerative conditions,” says Thomas Wishart, a research fellow at the University of Edinburgh’s Roslin Institute in the U.K.

Approach: Wishart’s group instead compares its proteins of interest to the total protein levels in the same lane, and has found that to be a more reliable method for normalization. That involves running two gels in parallel, and soaking one of them in Coomassie blue dye. Then, using digital imaging software, the researchers measure protein signal along the length of the entire lane.

Getting started: Wishart’s protocol, with troubleshooting tips, is available online (J Vis Exp, e52099, 2014). Having to load a second gel can introduce another potential source of error, but as long you can pipette single-digit μL volumes consistently, it shouldn’t be a problem, Wishart says. It helps to cross-check your Coomassie results with protein concentration in a BCA or other protein assay.

Wishart’s approach is one of many subtle variations on doing total-protein normalization. Bio-Rad, for example, makes “stain-free” gels that you can scan for total protein before transferring and probing for your protein of interest, though this approach requires special equipment for visualization.

Costs: Wishart’s protocol adds about 20 percent to your costs to run and stain that extra gel with Coomassie. “[But] realistically, you’re probably saving money in the long run because you’re doing it right from the start,” he says.

Considerations: If you insist on using individual proteins as loading controls, look through published proteomic data sets (using mass spectrometry) for proteins whose expression levels don’t seem to vary and are therefore more likely to be reliable.

Background: The multiple steps of traditional Western blots have remained much the same since the technique’s initial development. Although the steps may feel automatic to some scientists, the more handling and incubation steps there are, the more likely something will to go wrong. Researchers recognize the results as semiquantitative at best.

Approach: The San Jose, California, company ProteinSimple has come up with a different way of doing Western blots by using a matrix-filled capillary rather than a gel slab to separate proteins based on size, charge, or both. The subsequent steps of immobilizing the proteins, probing them with antibodies, and visualization are fully automated, making the firm’s instruments the most automated of the Western blot systems available.

LOOK, NO HANDS: ProteinSimple’s Simple Western systems use matrix-filled capillaries to resolve proteins based on size, charge, or both. The automated system shines ultraviolet light on the capillaries to immobilize resolved proteins, probes with primary and labeled secondary antibodies, and then quantifies the resulting signals.In 2011, the company launched Simon, the first in its line of Simple Western systems. The newest instrument, Wes, launched in 2014, is marketed for individual labs and analyzes up to 25 samples in less than three hours. Sally Sue and Peggy Sue are higher throughput, analyzing 96 samples per run with as little as 0.2 μg/μL of cell lysate.

“Besides improved reproducibility and consistency of the results, the biggest advantage, and this is especially significant for charge separation, is the ability to assay very quantity-limited samples,” says Joanna Liliental, director of the Translational Applications Service Center at the Stanford School of Medicine, which uses a Peggy Sue and its charge-separating precursor NanoPro 1000. “For a charge-separation assay, depending on the abundance of target protein in the cell, it is possible to quantify signal from only 25 cells. You’d never be able to do that with a [standard] Western.”

Costs: Although prices vary by region, Wes is priced “similar to a high-end imaging system,” says Simple Western product manager Patricia Piatti. Simple Western’s consumables are more expensive than if you made everything from scratch for a traditional Western; Piatti equates the costs to those of an “average” conventional Western.

Getting started: More cores are adding Simple Western instruments. Some of them, such as Liliental’s TASC, serve external users for at-cost pricing; it might make sense to consult with a core before considering the purchase of an instrument.

Of TASC’s 70-some users, only a few are trained on the instrument; mastery requires a four hour training session and a couple of weeks of regular use to become comfortable manipulating the machine. The software can take a bit longer to learn, depending on how complex the experiment is, Liliental says. With these systems, most of the work falls into planning the experiments to get the most out of each run and interpreting the data, she adds.

Considerations: As with any Western blot method, the quality of your antibody can be a limiting factor. ProteinSimple and its sister companies are working to validate antibodies for use with Simple Western instruments. “This is an ongoing project, and we’ll continue to add more certified antibodies to our already long list,” Piatti says. 

Correction: A previous version of the article incorrectly identified Bio-Techne as the sister company of ProteinSimple. In fact, ProteinSimple was acquired by Bio-Techne. The Scientist regrets the error.

Add a Comment

Avatar of: You



Sign In with your LabX Media Group Passport to leave a comment

Not a member? Register Now!

LabX Media Group Passport Logo


Avatar of: bcomer


Posts: 1

May 8, 2015

The Wishart groups inclusion of a total protein normalizer is a great idea that my previous lab shifted towards as well using a slightly different method that eliminates the need for a second gel.  In our case we used a reversible membrane stain from G-Biosciences (Blot-Fast Stain).  After transfering proteins to the membranes, prior to blocking, simply stain membranes with Blot-Fast stain, scan with near-infrared imager (Licor Odyssey)or image with camera-based system, destain in warm water, then proceed to membrane blocking and immunodetection as normal.  In our experience, this was linear to at least ~20-30 mcg, if I recall correctly, and it eliminates the need for a separate gel as well as controls for transfer variability in different regions of the membrane.  We never published results in which we normalized to the total protein directly but we did use the total protein stain images for visual confirmation of equal protein loading.  For example images see PMIDs: 25217662 and 25180620.

Avatar of: Leonid Schneider

Leonid Schneider

Posts: 4

May 18, 2015

Why do some people feel such dislike about proper loading controls for western blots?  
Thomas Wishart stains one single gel with Coumassie, and runs and probes all follow-ups without ever checking if transfer went just as well or the actual loading was indeed the same. Measuring protein concentration can certainly not replace loading controls. 

Indeed, the community has to trust his word or that of his student or technician that exactly the same volume/amount of protein was loaded on every gel, instead of being shown a loading control for each gel.
Avatar of: Leonid Schneider

Leonid Schneider

Posts: 4

Replied to a comment from bcomer made on May 8, 2015

May 18, 2015

Your method is actually very appropriate, unlike that of Wishart. You stain the very same membrane you later probe with antibodies (similar to Ponseau S method). His lab stains one gel for loading control and loads another with same samples (and yet another, and yet another, supposedly), which, according to the text above, are never checked for equal loading. I don't understand how Wishart's method actually passed the peer review. 

Popular Now

  1. A Potential Remedy for the Aging Brain
    The Scientist A Potential Remedy for the Aging Brain

    In mice, injected fragments of a naturally occurring protein boost memory in young and old animals and improve cognition and mobility in a model of neurodegenerative disease. 

  2. The Sleeping Brain Can Learn
    Daily News The Sleeping Brain Can Learn

    Humans can remember new sensory information presented during REM sleep, but this ability is suppressed during deep, slow-wave slumber.

  3. USDA Emails: Don’t Use “Climate Change”
  4. Nature Index Identifies Top Contributors to Innovation