Skip to main content
Premium Trial:

Request an Annual Quote

PNNL s Michael Feldhaus Discusses a New Approach to Antibodies


At A Glance

Name: Michael Feldhaus

Age: 35

Position: Senior research scientist, Pacific Northwest National Laboratory, Richland, Wash., since 2001.

Background: PhD, Experimental Pathology, University of Utah, 1996. Used antibodies to study signal transduction in neutrophils.

Senior research scientist, Acaris, Salt Lake City, 1996-2000.

Published a paper online in Nature Biotechnology last week entitled “Flow-cytometric isolation of human antibodies from a nonimmune Saccharomyces cerevisiae surface display library.”


What led you to develop new antibody technology?

When I was a graduate student working with Tom MacIntyre at the University of Utah, I found antibodies to be very useful reagents to induce as well as block signaling. After I finished my graduate studies, I took a position with a startup biotech company in Salt Lake City, Acaris. That company was trying to activate or block signaling pathways using peptides. The company was not successful, it fell out with the lack of VC funding, and so I moved up here to Pacific Northwest National Lab about two years ago. Here, I have been creating a library of antibodies for proteomics endeavors.

Where did the idea for the yeast surface display library come from?

Back in about 1997, Dane Wittrup [at MIT, who is the last author on the Nature Biotechnology paper] developed the yeast display technology platform, which can display many different types of proteins on the surface of yeast. One of the proteins he displayed was a single chain antibody. You can determine what the affinity of that antibody for the antigen is, and you can confirm this affinity using other types of analysis, such as surface plasmon resonance.

He built on this further by taking that antibody and doing molecular evolution on it. This basically induces random point mutations to create a small sub-library based on the original clone, which is then screened using flow cytometry, and you can identify clones that are of higher affinity.

Using this approach, he increased the affinity of the antibody for its antigen by five orders of magnitude, from a nanomolar to a femtomolar affinity.

How long did it take you to make the library?

The library was made in E. coli. To build it in the E. coli vectors took about two months, and then to transform it into yeast took about another three months, and that was just a lot of brute force effort. This resulted in 109 different clones.

How do you screen for antigen-antibody binding?

We use a combination of two approaches: highly parallel magnetic bead sorting based on the Miltenyi [Biotec] magnetic sorting technology, followed by flow cytometry cell sorting. In these approaches, we add the purified and biotinylated antigen. Then we use either streptavidin magnetic particles to pull out the cells that bind the biotinylated antigen, or we use streptavidin linked to a fluorochrome, which we can see on a flow cytometer, to sort out those clones that are binding the biotinylated antigen. What this allows us to do in about seven days is to go from a diversity of 109 different clones to a single clone that binds your antigen. To identify a monoclonal antibody for a new protein by conventional monoclonal antibody technology would take you probably 3 to 6 months.

What are the particular strengths or weaknesses of this yeast-based system?

Number one, you have removed the immune system from the animal, so if you want to make antibodies to the animal’s own proteins, you can. Second, this library that we have created can be used over and over again, whereas a mouse is used one time. Compared to polyclonals, which provide a finite amount of material, with our system, you have it as long as you want — similar to a monoclonal antibody. We can also put our antibodies through molecular evolution and rapidly evolve it to maybe a higher affinity, or we can even put it into a different type of context so you have different epitope tags, or you can just simply cut and paste it and make it into looking like a very simple monoclonal antibody. For comparing it to phage display and ribosomal display, I believe that these are more technically demanding, and the libraries themselves, for phage display, are not [capable of being propagated]. The yeast library is a tightly regulated expression of the single chain antibody, so when you are not expressing it, you can grow the library up in yeast and expand it. If you tried growing a phage display library, you would lose clones, therefore losing diversity of the library. There are uses for phage display or ribosomal display that are unique to them. Phage display you can use to do tissue panning. You can’t do that with the yeast display system.

A large benefit of our technology is that we monitor our selection process in real time using a flow cytometer. Probably one of the very important features is, this is a human monoclonal antibody library.

What are you currently using the library for?

We have identified antibodies to receptors on the surface of cells, intracellular proteins, as well as phosphorylated peptides. Now we are characterizing the antibody interactions with cell surface receptors by such things as flow cytometry, and we look at using some of the antibodies for an ELISA, and for immunoprecipitations. With our phosphospecific antibodies, we are looking at the phosphorylation status of proteins in cancer vs. normal cell lines with our collaborator Ettore Appella, at the NCI. With Dick Smith here at PNNL we are identifying antibodies to proteins that are present in Shewanella oenidensis, which is an organism of interest to the Department of Energy in the context of bioremediation. [This will be funded under the DOE’s Genomes to Life program.] We will be using the antibodies to do immunoprecipitations, and Dick Smith is going to be running them on his mass specs to identify the proteins that are in those complexes.

What about using the antibodies in microarrays?

Currently we are funded through the chemical and biological weapons non-proliferation program. What we are planning on doing is to create panels of antibodies to organisms, proteins, or chemicals of interest to the biological weapons detection program. What we want to do is make a series of antibodies that bind the same epitope but differ only in their affinity, maybe by five logs. What this allows us to do is look at a much larger dynamic range for the concentration of that antigen. It may be present normally in the environment, but when it’s a thousand times higher, that’s when it’s a concern. We are looking at several types of potential microarrays, from the standard glass slide microarrays to some of the soluble microparticle arrays that are available.

Can your library create specific antibodies quickly?

In the paper it’s demonstrated that we are able to do a multiplex selection. What that means is, when we screen our library, we can throw in 10, 20, we believe up to 50 or 100 different antigens at one time, and seven to ten days later have an individual single-chain antibody for each one of those antigens. That gives you speed, which also brings down cost. Growing yeast and growing our library in a standard laboratory is incredibly inexpensive. The cost of media is very cheap. No one is up in arms about killing yeast, as they are about killing mice or other animals. We are also working with several companies right now to automate the process of selection as well as characterization. One company that we are looking at is DakoCytomation. They have a high throughput cell sorter with front- and back-end automation that would make our sorting as well as our analysis much faster, less people-intensive.

Who owns the intellectual property to your system?

I would like to say ‘who doesn’t.’ A couple of groups that do would be Cambridge Antibody Technology, Abbott Laboratories, and, I believe, Dyax. The yeast display system is a patent originally generated by Dane Wittrup at the University of Illinois, which has now been licensed to Abbott.

So you are not planning to exploit this technology commercially?

We are not trying to commercialize it. My understanding is that we are making it freely available for non-profit universities and government institutes. But we are not licensing it to companies.

What antibody initiative would you prefer — cataloging existing antibodies or creating new ones?

I think to gather up all the antibodies that are presently available, to characterize them in a similar fashion and [put] them into a central database, is probably an enormous amount of work, maybe more costly than starting over. The ability to create and identify an antibody is actually the simple part. Characterizing it and producing it, those are the hard parts.

Our yeast system is a simple way to make research agents. You could actually hold just the plasmid DNA instead of the yeast clone. Other researchers could take on the financial responsibility of producing their own antibody that they are interested in. This would lower the overall cost for a central organization to produce and maintain these reagents.

I think monoclonal antibodies are a robust technology; they have been around for years; there are many very useful reagents out there. But creating this huge repertoire of affinity reagents is a 20th century problem, and we are in the 21st century. Why look at older technology, such as monoclonal antibodies, when we have all the molecular biology and all the high-throughput technology to create these reagents very quickly?

The Scan

Missed Early Cases

A retrospective analysis of blood samples suggests early SARS-CoV-2 infections may have been missed in the US, the New York Times reports.

Limited Journal Editor Diversity

A survey finds low diversity among scientific and medical journal editors, according to The Scientist.

How Much of a Threat?

Science writes that need for a provision aimed at shoring up genomic data security within a new US bill is being questioned.

PNAS Papers on Historic Helicobacter Spread, Brain Development, C. difficile RNAs

In PNAS this week: Helicobacter genetic diversity gives insight into human migrations, gene expression patterns of brain development, and more.