At A Glance
Name: Xiaolian Gao
Position: Assistant professor, department of chemistry, University of Houston PRIOR Experience: Co-founder of Xeotron, postdoc in Dinshaw Patel’s lab at Columbia University
In a paper published this month in Nature Biotechnology (see Research Reports, p. 6) Xiaolian Gao describes her approach to synthesizing arrays of peptide fragments for use in peptide-protein and other interactions. In an interview, she provided her thoughts on her background and the technology’s future:
Were you always interested in doing this type of chemistry?
The initial driver was to do things faster, rather than look at a single experiment at a time. Actually my research field is in three-dimensional structure [determination] by NMR. Usually in NMR you make one molecule and then you look at whether it’s going to form a structure, or whether it’s going to form a complex. You do it one by one, which is very slow. That of course drives [things] in two directions: can we do anything faster, and can we do anything better? Microarrays seemed to be one way to go, but there was really no easy ready way to make them [at the time].
We started working with DNA, [because] you might want to look at some DNA complexes with drug molecules. You can buy several hundred thousand oligos, and then look at this system. That drove us into looking into how arrays were made at the time, [because] we were not satisfied with the technologies. Then we realized that in situ synthesis was the way to go. [But] you need a gating step if you do parallel synthesis. That’s how we thought about this.
Rather than using photoprotective groups, which would require making all different monomers to [generate a peptide] sequence, we used light as a gating step. That makes the whole process much easier because the basic reactions only require conventional reagents. So that’s how we got started.
Then we also came in with digital photolithography, rather than using photomasks. This was also to simplify the process, so we could actually implement it in a lab.
Why synthesize peptides for use with arrays, as opposed to attaching recombinant proteins?
When you’re [developing] arrays for proteomics, there will be many answers to many questions. I don’t think that in any way peptide arrays [will substitute for] protein arrays. It has its applications, for example, when you ask a question about a protein interaction. In those areas there seem to be some applications, because sometimes the [interactions occur] in a linear region of the protein, and other times protein-protein interactions involve C-terminal peptides. There are examples where people have shown that you can make a C-terminal peptide to look at interactions with a protein. In cases like this the peptide array is useful because sometimes it’s not possible to isolate the full-length protein. This allows you take a quick look to get some information out. Sometimes you can get important information; sometimes you may be limited by the length of the peptide.
But there are other applications of peptide arrays. For example, you can make phosphorylated peptides, or you can use the peptide [as the substrate] for a protease or kinase. These are classic assays. By adding the microarray format you can condense hundreds of such assays on a very small area, to screen kinases or proteases for their functions or inhibitors.
Also, our method of chemistry allows us to make non-natural peptides. Now you can broaden into the area of making synthetic analogs, and make quite a few of them, because you can buy those amino acid analogs to form sequences. Then you get into the area of using the peptides or peptide mimetics in screening for drug binding or protein-binding molecules as potential drugs.
You mentioned the unique reaction chemistry. How do you describe your particular approach?
Essentially on the surface [of the array] you have thousands of spots, and you want certain spots to react during the first reaction cycle, and certain spots not to react. Let’s say you want to add in monomer A. You want monomer A only added to certain sites, so all those sites will be A; all the other sites will have nothing. In the second step you add in B, and you want B only on certain sites. You [want to] end up with a surface where some sites have the A molecule, some sites have the B molecule, some sites have an AB dimer, and some sites have nothing.
Essentially you want to run parallel reactions, and at each step you want some sites to react and some sites not to react. In order for [this to happen], you need to be able to gate each site. This is what the photochemistry allow us to do.
Is the photochemistry unique to your lab?
Yes, what’s unique is our chemistry. The photogenerated reagents, especially the photogenerated acids, are known. The compounds are not a novel discovery of ours; they’re used a lot in the solid phase to make computer chips. What’s new in [our lab] is that we’re the first ones to demonstrate that the chemistry works in solution for parallel synthesis, as a way for combinatorial synthesis. We’re the first to demonstrate that we can use such a reaction in solution, not in polymer phase, and actually run conventional chemistry to make combinatorial molecules. This is where [we added] unique aspects.
How long can you make these peptide sequences? Are there constraints on what kind of peptide arrays you can make?
Basically the restrictions will be whatever [restrictions apply to] conventional [t-butyloxycarbonyl or 9-fluorenylmethoxycarbonyl] chemistry. So in theory we can make sequences of up to 27 or 30 mers. But because we consider all 20 amino acids at each step, a practical consideration would be the number of steps [required to build the peptide]. For DNA synthesis there are only four nucleotides, so at each step you cycle through only four. For peptides, if you want to build a 16-mer sequence with 20 different amino acids, you quickly have a very long synthesis cycle. So in practice, anything less than a 15-mer array would be practical. Anything longer should probably [use fewer amino acids], so the overall steps are not tremendous.
Are there particular benefits to using synthetic peptides? Is folding an issue when probing an interaction between a peptide and a protein?
[Folding] issues are always going to exist. People have debated for a long time [the merits of using an assay] on a solid surface [because] it’s harder to know what’s exactly going on. Certain motifs are going to have a folding-type issue. But because with this approach arrays can be made very quickly, the problem can be identified very quickly. So in a sense in terms of larger scale screening, people will just have to go where the information is useful and meaningful, and over time, solve some of these issues.
You mentioned earlier that the main applications of these peptide arrays are in looking at protein-protein interactions. Is the technology suited for other types of applications?
Compared to DNA chips, peptide synthesis is not a natural market, but it’s going to grow. The reason [the market is not] there right now is because there’s nobody who can produce peptide chips on a very small size and also at the density that a DNA chip can achieve. We’re at a stage where we’re getting the peptide technology to where we can make peptide chips readily available. [Once we do], I’m sure there will be a lot of applications that will grow from there. Right now, I can see that some of the conventional assays can be moved to this miniature format, so the process can be handled more efficiently, thereby saving time and cost.
Also, by identifying the right ligand for a substrate you can profile kinases and proteases. This is going to be a huge area. Others would be in identifying therapeutic lead compounds. I see it as an emerging market: it’s going to grow. First you have to provide the tools to people. In the next few months we’ll try to actually commercialize the peptide chips. Once the tool is there, I think people are going to develop applications besides what we can recognize at this time.
What is your relationship with Xeotron?
The technology development contributions came from three parties. My lab at the University of Houston, where we developed the chemistry; Xeotron; and Erdogan Gulari’s lab at the University of Michigan. Their expertise is in MEMS and microfabrication. The combination of digital photolithography and microfabrication make this approach very simple and efficient, and make it possible to develop peptide microchips.
I’m a co-founder of [Xeotron] but my major research activities are at the University of Houston. Gulari and I both have close collaborations with Xeotron to commercialize the technology.
Xeotron is interested in coming out with a product?
Yes, Xeotron is now trying to commercialize the DNA chip; that’s their first-generation product. Because the technology platform is very broadly applicable to a variety of different molecules, the second generation will be peptide arrays, and perhaps later on other types of arrays, for example, RNA arrays.
So you can apply the same technology to building these other types of arrays?
Our DNA process is of very high quality, so we know we can make DNA analogs and RNA molecules. Now [in the Nature Biotechnology paper] we show that peptides with [t-butyloxycarbonyl] chemistry can be used in an array. We are also working on the [9-fluorenylmethoxycarbonyl] chemistries. Given the time and need, we’ll get into other [classes of molecules]. We’ve also devoted some effort to making small molecule libraries. We see [our approach] as a general way of performing combinatorial synthesis.