Professor, Departments of Chemistry and Molecular Biology
Scripps Research Institute
Name: Reza Ghadiri
Position: Professor, Departments of Chemistry and Molecular Biology, Scripps Research Institute
Experience and Education:
— Postdoctoral Fellow with Emil Kaiser, Rockefeller University, 1987-1989
— PhD in chemistry, University of Wisconsin-Madison, 1987
— BA in chemistry, University of Wisconsin-Milwaukee, 1982
Reza Ghadiri’s group at the Scripps Research Institute has been exploring nanopores for DNA sequencing for several years. After initially toying with self-assembling peptide nanotubes, he and his colleagues teamed up with Hagan Bayley, a professor of chemical biology at the University of Oxford in England, to study protein nanopores.
In 2005, Ghadiri and Bayley won a five-year, $4.2 million grant under the National Human Genome Research Institute’s “$1,000 Genome” technology program to explore ways of using engineered alpha-hemolysin pores to sequence DNA.
Last month, Ghadiri’s group published an article in the Journal of the American Chemical Society describing a nanopore device that detects DNA polymerase activity with single-nucleotide resolution.
In Sequence spoke with Ghadiri last week to find out about this work, and how it paves the way for nanopore sequencing.
Can you briefly describe the work in your recent JACS paper, and what you have been able to show?
Let me give you a little background. The idea of using nanopores for DNA sequencing, of course, has been around for some time now. However, in order to do so, you need to be able to have some form of recognition of a nucleobase, and you have to have single-base resolution. In order to determine [if we can do so], we developed a system in our previous publications that [allows us to] mechanically hold a single molecule of DNA inside the alpha-hemolysin pore through formation of single-molecule rotaxanes and pseudorotaxanes. [Rotaxanes consist of a dumbbell-shaped molecule threaded through a circular molecule.]
By holding the molecule inside the pore in a presumably specific configuration, we asked the question, can you have signals that can detect different nucleobases, as well as a particular nucleobase located at a precise position somewhere on the DNA that is threaded? This early work showed that there is considerable promise proceeding with this project.
The [JACS] paper we published builds on the devices that we made before, the rotaxanes. We put [a hybrid of] two types of polymers [through the pore, consisting of] a nucleic acid polymer, the DNA, and a polyethylene glycol phosphate molecule.
One of these highly evolved machines that duplicates DNA is DNA polymerase, and it does it with high fidelity and quite a high speed rate. One possibility to read single-molecule DNA is to detect the action of DNA polymerase without any labeling. [We used a] primed DNA [in our device], and if DNA polymerase works and extends the primer according to the template sequence, then the [double-stranded stretch of DNA] would grow longer. We [first] designed a series of primers of different lengths and showed that you could detect a different signal for each of those. That would suggest that if DNA polymerase operates on this, then you might pick it up.
That’s precisely what we did. The system shows successive incorporation of single nucleotides, with a spatial resolution of approximately 2.7 angstroms, less than a base pair stack of 3.4 angstroms. So the system shows a very high resolution, high enough for looking at single nucleotide incorporation.
[This] really provides encouragement to this field, that it is possible to do this. Prior to our work, single nucleotide incorporation at the molecule level with the aid of DNA polymerase had not been established.
What do you still need to solve to develop this further for DNA sequencing?
The goal of the next phase of our program is to develop new methods that can read out the sequence, to identify what dNTP is coming in. That would be the key, the one last remaining objective.
[The initial experiment] was designed to see whether you can see single-nucleotide incorporations, because without that, you would not be able to read any sequence. It establishes the basis that you can have a single-nucleotide readout. But then, there is another series of designs that has to be implemented that could potentially identify what’s coming in.
Can you give me an idea of how you would identify them?
We have several ideas, let’s see which one works.
What else do you need to work on to make this practical?
I think if one can do what one hopes [to do], then the path to DNA sequencing would require just some technological development of how to stabilize the pores in various device configurations, for example in lipid bilayers or solid pores, so one can run [many] of them simultaneously. I think there are many advances in this field currently ongoing that will all converge, hopefully, and be able to assist each other in developing a practical device. I’m hopeful that if the basic science part of it works, technology advances will allow practical applications.
What about the read length – how long could you make the DNA that’s tethered in the pore?
In principle, you can have very large read lengths. Once the DNA is captured inside the pore, it’s not going to let go. It’s also the issue of how processive the enzyme is that one is using, how often does it let go or not.
Why are you using protein nanopores rather than solid-state nanopores?
[One reason is] the very elegant work of Hagan Bayley in protein nanopore engineering. [He and his colleagues have introduced] organic molecules and adapters and so forth. One has great leeway, great capacity to change and modify the protein pore characteristics to achieve some of the desired physical properties for DNA capturing, sequencing, and readout. We very much like the protein pore because it gives us those designable characteristics. However, one should expect great things also from solid-state pores as that technology develops, and I don’t think they are mutually exclusive.
Why have you chosen alpha-hemolysin over other protein nanpores?
There are lots of things one can use, but the alpha-hemolysins are very stable [and] the channel is always open. [There is also] a high-resolution X-ray structure available and its biophysical characteristics and conductance have been studied, in addition to the numerous studies, through site-directed mutagenesis and chemical modification, by Hagan Bayley. There is rich data and understanding that one can exploit for further design. There are other [protein nanopores], and people work on other things.
The abstract for your NHGRI grant with Hagan Bayley lists a number of approaches. Are there others, besides the one you just published, that look promising?
We are going after this on multiple fronts. This is basically very early research; we are trying to do something that has not yet been achieved. One should follow leads and reasonable prospects and find out for each the scope and limitations, to be able to go forward in a rational way.
For example, Hagan Bayley in his lab has shown that using a cyclodextrin adaptor, you can detect different nucleotides. Coupled with an exonuclease, which chops off one [nucleotide] at a time [which] can be fed into a channel, he already has a readout. So that’s another approach which looks extremely promising.
In addition, we have a number of chemical modification approaches that we are testing.
A year ago, we covered Oxford NanoLabs (see In Sequence 2/27/2007), a spinout from the University of Oxford that is helping Hagan Bayley commercialize his technology. Do you also work with them?
I have spoken with them, and we are in the early stages of figuring out if we can interact with each other. I’m looking forward to that if that possibility arises.
A few years from now, what do you think a nanopore sequencer will look like, and what do you think it will be capable of doing?
It will look like any piece of equipment attached to a computer. I think it would have hundreds of nanopores running in parallel. That is obvious — if you can get it to work for one, you should run multiple things in parallel just because you have a much higher throughput.
And then you will have the potentially long read length that makes sequence assembly much easier. The whole promise of this is that it’s going to be very long reads. It depends on how long a fragment you are going to capture in the pore — you can capture any size, basically, up to a point. It could be much longer than current technology, for sure. But then there are many other questions, [for example] how accurate your reads are, how many times you have to read it. Certainly one read is not enough; you have to re-read the same fragment several times to get accurate data. All those things have to be determined.
But [long reads are] just one benefit. The reason there is so much interest in the field is to decrease the cost of genome sequencing, or resequencing, so that everybody can have their genes sequenced for a very small price.
There will be, no doubt, multiple single-molecule methods [that will be] able to sequence DNA. And they are all going to be based on fantastic science. The question then, at the end, would be which one of them — based on cost and practicality — will lead the field.
I think in the near term, there are many techniques that are maturing that will be very useful for considerably decreasing the cost from where we are right now. And then we will find out whether there is any technology — nanopore or not — that is going to be able to bring it to a level that would [make sequencing] available for an entire nation. That’s a few years away from that, but I feel the field is going ahead strongly, and there are many basic science advances on multiple fronts that show promise.