Physical Scientists May Be Key To Speedup of Gene Sequencing

Chemist Lloyd M. Smith's entry into the world of gene sequencing came about during his postdoc days, in Leroy Hood's molecular biology lab at the California Institute of Technology in Pasadena. With his background in fluorescence chemistry and instrumentation, Smith saw a way to speed the tedious process of reading DNA sequences off gels. The result was the development of the first automated DNA sequencer (L.M. Smith, et al., "Fluorescence Detection in Automated DNA Sequence Analysis," Nature, 321:674-79, 1986). Smith, now at the University of Wisconsin, Madison has been working on gene sequencing ever since, coming up even with faster ways of sequencing DNA. For Smith and other physical scientists, the grail of the government's Human Genome Project (HGP) is not the sequence and analysis of our 3 billion DNA bases, as it is for the biologists, but the technology that will actually do it. Armed with an array of analytical hardware such as mass spectrometers, scanning tunneling microscopes, and flow cytometers, these physicists and chemists are seeking new and improved systems that can read the order of the four nucleotide bases from the stretch of DNA. While many of the ideas for new sequencing technologies have been around since the mid-1980s, the HGP in the last two years has spurred the pace of their development. It has also created a niche for collaborations among physical scientists and molecular biologists, says Jane Peterson, chief of research centers for the National Institutes of Health's HGP effort. Developing new sequencing technologies "takes interdisciplinary work," she says. "It really takes people in other disciplines to come up with these new technologies....[But} the most successful projects coming through are those that are in touch with the biology." As it's currently done in most laboratories in the world, sequencing involves cloning a fragment of DNA, running it through four base-identifying reactions, separating the pieces from each reaction by gel electrophoresis, then reading each piece to determine the sequence. Even though most of the procedure has been automated, it is still far too slow, tedious, and expensive to be used on anything the size of the human genome. The HGP program set a goal of five years, beginning last October, to bring the speed and efficiency of sequencing up and the cost down. This year the genome programs at the Department of Energy and NIH are spending about $17 million (out of a total HGP budget of $135 million) toward that end. Much of the effort is focused on squeezing as much performance as possible out of the current, gel-based technology. But the HGP, primarily through DOE with some support from NIH, is also supporting about a dozen research groups investigating the use of other instrument-based methodologies as high-speed alternatives that eliminate the use of gels. "You can increase the rapidity, efficiency, and cost-effectiveness of state-of-the-art sequencing capabilities only so much, but you can increase them by factors of two or four by automation and by improving one aspect or another,"says Benjamin J. Barnhart, manager of DOE's part of the HGP. "What we're really looking for are increases by factors of 10 to 100." The use of mass spectrometers to directly analyze DNA has emerged as one promising means of achieving those objectives. These powerful tools have long been used to analyze everything from meteorites in the research lab to chemical weapons on the battlefield, although until recently they were limited to detecting and identifying relatively small molecules, like lead. But in the last few years, chemists have gone after larger quarry, namely proteins, and have succeeded in nailing down their molecular weights and amino acid sequences. The success in sequencing proteins has encouraged chemists to train mass spectrometers on DNA. "The extrapolation, in a simplistic way, in going from proteins to DNA seems reasonable,"says Smith. "It's a rapidly moving field. Nobody ever imagined even two or three years ago that you could do what we are doing now." These instruments read the mass of a molecule, and hence its identity, from the time it takes to move through an electronic or magnetic field. But mass spectrometry has its limitations. First the molecule must be ionized, or given a charge, and volatized, and large biological molecules like DNA fall apart under the usual ionizing conditions-say, shooting it full of electrons. Researchers are applying some of the tricks used on proteins to get around the problem. One way involves matrix-assisted laser desorption, a mass spectrometry technique developed by German researchers (M.Karas and F.Hillenkamp, "Laser Desorption Ionization of Proteins with Molecular Weights Exceeding 10,000 Daltons," Analytical Chemistry, 60:2299-2301, 1988). "This technique is amazing," says Smith, who is collaborating with chemist Brian T. Chait of New York's Rockefeller University to apply it to DNA sequencing. A similar version of the technique is being developed by Peter M. Williams at Arizona State University, among others. The matrix is a lightweight substance that holds the DNA. The mixture is inserted in the mass spectrometer and zapped with a laser pulse at a wavelength that destroys the matrix but leaves the DNA intact, airborne, and ready for analysis. This process would take seconds, rather than the hours required to run a gel. Although the method is still at the feasibility-demonstration stage, its developers predict it could churn out raw data for several hundred thousand bases a day, compared with the 10,000 or so bases a day that existing procedures manage. A second recently developed trick of mass spectrometry, electrospray ionization (J.B. Fenn, et al., "Electospray Ionization for Mass Spectrometry of Large Biomolecules," Science, 246:64-71, 1989), is also being Battelle Pacific Northwest Laboratories (PNL) in Richland, Wash., and here. With this approach , a sample containing the DNA is forced into droplets that get ionized, evaporated, and launched into the spectrometer. PNL mass spectroscopist Richard D. Smith is using that technique as the starting point for one of the more radical sequencing schemes that have been floated. The idea is to trap a single molecule of electospray-ionized DNA in an ion cyclotron mass spectrometer, which would measure what remains of the DNA each time a single base gets clipped off. The main hitch: finding a way to fragment the DNA inside the vacuum of device one base at a time. A team of physicists, chemists, and biologists at Los Alamos National Laboratory in New Mexico, one of nine HGP centers in the United States that do sequencing and mapping research, is also stalking a way to cleave and read one base at a time,but with different instrumentation. In this case, the bases of a very long piece of DNA get tagged with four different fluorescent dyes and run through a flow cyctometer, an instrument that usually identifies cell types as they pass through a laser beam. But in this case, the bases are enzymatically clipped off one by beam that lights up the dye, indicating which base it is. There several major hurdles to this project, but investigators are attacking all fronts. In 1989, the researchers succeeded in demonstrating that single molecules could be detected, the basis for this scheme. And in late March, the lab signed a three year cooperative research and development agreement with Life Technologies Inc. of Gaithersburg, Md., to develop the enzymes involved in tagging and cleaving the individual bases. "It looks promising," says DOE'd Barnhart. If all the pieces can put together, it will "by far be the most rapid and cost-effectiveness approach to sequencing that anyone can possibly dream up at this point." The maximum rate envisioned for this technique: raw data on several million bases a day. Scanning tunneling microscopy (STM) has also tantalized researchers with the possibility of rapid and flows between the scanning probe and the surface of a molecule. Changes in the surface affect the current and the probe, producing information that gets processed into an image. If the resolution is improved, STM could conceivably distinguish the four DNA bases. Many laboratories began pursuing STM for sequencing after researchers at Lawrence Berkeley and Lawrence Livermore national laboratories in California published a high-detail STM image of the DNA double helix in 1989 (Thomas Beebe, Jr., et al., "Direct Observation of Native DNA Structures with the Scanning Tunneling Microscope," Science, 243:370-72). Just recently, the researchers at Lawrence Livermore have been able to make out the differences between two bases, adenine and thymine, in images taken of the separate bases. But they still don't know if the technique will work for large-scale DNA sequencing. There are problems with preparing the DNA, obtaining reproducible results, and interpreting the images. And many of the details of the tunneling process are still not understood. "It's still more of an art than a science,"says Rodney L. Balhorn, a physical chemist at Lawrence Livermore. Yet another gel-less technique being supported by Energy Department's genome iniative and the Yugoslavian government is sequencing by hybridization. Yugoslavian molecular biologists Radomir Crkvenjakov and Redoje Drmanac, who designed the scheme, are spending a year at Argonne National Laboratory in Illinois to test its feasibility. The method involves applying many short probes-strands of a few known bases-to DNA fragments wherever the bases can form pairs. The sequence of the genome could than be pieced together by looking at where the probes overlap. About five other groups are pursuing variations of this technique. All of these novel techniques need another two to five years to prove whether they will work for large-scale sequencing at an acceptable cost. But speedier sequencers alone won't put the HGP on the fast track. It will also take compatible automated front-end technology-preparing the DNA to be sequenced-and computer software t process the mass of data the sequencers generate. In the meantime, several incremental improvements to gel-based sequencing are much closer to commercial reality. Molecular geneticist Raymond F. Gesteland at the University of Utah, chemists Norman J. Dovichi of the University of Alberta, Barry Karger of Northeastern University, and Wisconsin's Smith are putting together systems based on capillary gel electrophoresis. These systems use gels poured into ultra-thin capillaries or slabs that, because of their high surface-area-to-volume ratio, won't heat up like ordinary gels when higher voltages are applied. Higher voltages mean faster separation. Thomas Brennan of Genomyx Inc. in South San Francisco, Calif., has hooked a capillary gel to mass spectrometer that reads the bases labeled with four different sulfer isotopes. Researchers at Oak Ridge National Laboratory in Tennessee are using other isotopes for detection by mass spectrometry. Others are working on automating the gel multiplex approach to sequencing. This approach, developed by Harvard University's George M. Church, gains speed by electrophoresing 20 or more DNA fragments en masse, then picking out their sequences using labeled probes (George M. Church and Stephen Keiffer-Higgens, "Multiplex DNA Sequencing," Science, 240:185-8, 1988. Even if none of the new methods is used as the HGP's next sequencer of choice, "it's good to have people working on pushing technology," says Wisconsin's Smith. There is always the promise of spin-off applications in biotechnology, structural biology, and other fields, proponents say. "That's the real payoff of the genome project," says Barnhart. Christine Mlot is a Wisconsin-based freelance science writer.

Physical Scientists May Be Key To Speedup of Gene Sequencing

Interested in reading more?

Become a Member of