In Search of Genomic Variation

The fairly nebulous term mutation detection addresses two fundamentally different questions: "Do any mutations-or, more broadly, polymorphisms or variations-exist in a given gene?" "How frequently does a specific mutation occur in a population?" Getting the answer to each question presents different challenges, and scientists must address each using different technologies. The first question is answered with mutation scanning or screening techniques, the second with mutation scoring, or genot

Oct 29, 2001
Laura Defrancesco
The fairly nebulous term mutation detection addresses two fundamentally different questions: "Do any mutations-or, more broadly, polymorphisms or variations-exist in a given gene?"

"How frequently does a specific mutation occur in a population?"

Getting the answer to each question presents different challenges, and scientists must address each using different technologies. The first question is answered with mutation scanning or screening techniques, the second with mutation scoring, or genotyping methods. Because mutation-scoring technologies were reviewed last year,1 screening techniques are addressed here.

A world of difference exists between mutation screening and mutation scoring, says Joe Rudolph, a senior applications scientist at Omaha, Neb.-based Transgenomic. Many companies offer products for "SNP (single nucleotide polymorphism) detection," which are really scoring technologies. To actually screen for polymorphisms requires a tremendous amount of time and resources. Such studies require genetic material from hundreds or thousands of individuals, especially when hunting for rare variations. Cost is a key consideration. A reagent that costs $2 per sample is fine for routine lab work, but it becomes prohibitively expensive when that reagent is being applied to 10,000 samples. The same is true for throughput: What works well for a small study of polymorphisms in a single exon with a dozen patients can become a real bottleneck when applied to a large population study.

Therefore, the ideal mutation screening technique should be high throughput, reliable (capable of identifying at least 95-98 percent of variations), and low-cost. A literature search of the field reveals a number of approaches, including those based on sequencing, enzymatic reactions, microarrays, and HPLC fractionation. Researchers can scan for mutations with nothing more than a garden-variety gel stand or spend tens of thousands of dollars on specialized equipment. So far, no single technique has supplanted sequencing as the method of choice for detecting mutations, and it's too soon to predict when or even if one ever will.

Because of the range of options available to them, investigators are urged to consider what they hope to achieve with a given study before selecting a screening method. For small-scale studies, scientists might be able to use some polymorphism scoring platforms and methods, such as sequencing or PCR-based assays, or even Uppsala, Sweden-based Pyrosequencing AB's Pyrosequencing™ technique, in lieu of the techniques described below.

Assay Comparisons

The melting properties of DNA fragments are sequence- dependent, and homoduplex and heteroduplex DNA fragments exhibit different denaturation characteristics. Among the simplest approaches for detecting mutations, conceptually as well as technically, are gel-based assays based on these properties. Such techniques include denaturing gradient gel electrophoresis (DGGE), temperature gradient gel electrophoresis (TGGE), and constant denaturant gel electrophoresis (CDGE).

These assays differ only in the method used to denature the DNA fragments. DGGE uses a chemical denaturant gradient, TGGE employs a temperature gradient, and CDGE uses a constant amount of denaturant. Researchers can perform all three on native DNA, but coupling them with heteroduplex formation enhances their potential, because it's easier to distinguish heteroduplexes from homoduplexes than to distinguish two different homoduplexes. With the right-sized fragment, these techniques can detect as little as a one-base difference.2

A related technique, single-strand conformation polymorphism analysis (SCCP), scans for conformation changes when single-stranded DNA is allowed to fold into secondary structures as it renatures.3 Both SCCP and DGGE work most efficiently with small (100-300 base pair [bp]) fragments. Because eukaryotic genes are typically several kilobases long, they must be analyzed in many pieces. However, scientists can multiplex these experiments to reduce the number of reactions and wells that need to be run to complete analysis of a given gene.

Typically, conformation-dependent assays are performed on PCR-amplified fragments to facilitate detection; hence, primer design is a critical step in using these approaches. DGGE, for example, works best with fragments that have only one or two melting domains. Since a partially denatured molecule's branched structure gives it its characteristic mobility, the fragments must contain a high melting domain to prevent them from completely denaturing during the run. Researchers can circumvent this problem by adding GC-rich sequences to the end of the PCR primers (GC-clamps), which create an artificially high melting domain.4

DGGE and SCCP detect only about 80 percent of mutations. The techniques often miss mutations embedded in GC-rich regions, as well as mutations that occur close to the fragment ends, which might not significantly affect mobility. However, some modifications to the techniques can improve their efficiency. Steve Sommer and colleagues at the City of Hope National Medical Center in Duarte, Calif., have made two such improvements. Combining SCCP with dideoxy fingerprinting, using just one dideoxy terminator, increases the efficiency of detection to nearly 100 percent,5 as does a technique called SSCP - DOVAM-S (for detection of virtually all mutations), in which SSCP is run under five conditions (varying temperature and gel matrices).6 Using DOVAM-S, J.R. Mendell of the Ohio State University Medical Center in Columbus, Ohio, detected mutations in the dystrophin gene in 90 percent of patients with Duchenne muscular dystrophy. By comparison, conventional technology can only detect one-third of the mutations in this complex gene.7

A new technology under development by Accelerated Genomics of San Antonio, Texas, uses an automated two-dimensional gel set-up to increase the scope and throughput of DGGE. The company's OptiScan HT Gene Discovery Platform separates by size in the first dimension and by sequence in the second (by DGGE), to generate a unique pattern of target fragments. Each gel can resolve up to 120 fragments with multicolor fluorescence, allowing mutations and polymorphisms to be detected in the context of the whole gene. Accelerated Genomics' automated instrument runs eight gels simultaneously and increases the speed of DGGE separation while maintaining high resolution. Researchers can use the company's gene-specific test kits in high-throughput processing.

In 1995 Stanford University's Peter Oefner and Peter Underhill developed a quick and automatable way to detect conformational changes in DNA fragments without gel electrophoresis. The process, called denaturing HPLC (DHPLC),8 combines the non-ionic size separation of DNA fragments with thermal denaturation, giving essentially the same information as SSCP or DGGE. While considerably more costly than electrophoresis equipment, DHPLC offers several advantages over electrophoretic approaches. For example, the technique is fast (samples are analyzed in a few minutes), sensitive, automatable, reproducible, and compares very favorably to DNA sequencing as a screening technique.

Two companies, Transgenomic and Walnut Creek, Calif.-based Varian, have commercialized DHPLC for single nucleotide polymorphism (SNP) detection in two similar instruments-the WAVE® System (Transgenomic) and the Helix System (Varian). Transgenomic holds the exclusive license to an alkylated polymer column for DHPLC developed by Guenter Bonn, Oefner, and coworkers at the University of Innsbruck. This column, the DNASep® Cartridge for the WAVE System, is designed to withstand the rigorous conditions of DHPLC. However, conventional HPLC columns can perform the reproducible, automated size separations of DNA fragments required for DHPLC analysis.

Both the Transgenomic WAVE System and the Varian Helix System work on the same principle: DNA strands containing polymorphisms or mutations form heteroduplexes when recombined with wild-type strands of the same exons. DNA heteroduplexes have lower melting temperatures than the corresponding homoduplex species. An ion-pairing reagent such as triethylammonium acetate neutralizes the DNA phosphate backbone, so separation is based on the relative hydrophobicity of the DNA species. When run on a DHPLC system at temperatures high enough to partially melt heteroduplex DNA, those samples containing mutations will migrate significantly differently from the corresponding DNA homoduplexes.

DHPLC is fast becoming an established technique for SNP detection and validation. For example, researchers at the National Cancer Institute recently used the technique to identify human polymorphisms and to genotype individuals for their presence.9 To better serve this growing need, both Transgenomic and Varian have increased their instruments' throughput. Transgenomic's new WAVE System Model 3500HT doubles the unit's sample processing capacity, thanks to a new proprietary DNASep cartridge. And Varian's Helix has a multichannel fluorescence detector that can run four samples simultaneously.

Structure-specific Enzymes

Various structure-specific enzymes have been enlisted in the hunt for mutations, including ribonucleases, cleavases, and endonucleases, all of which recognize secondary structural features to some extent. Several of these mutation-scanning techniques have been commercialized. Broadly, this class of detection technique relies on the fact that wild-type/mutant heteroduplexes exhibit a bubble or bulge at the sequence mismatch site that endonucleases can recognize.

The enzymatic mutation detection (EMD) assay represents one example of this assay type. Based on the work of Richard Cotton and Rima Youil at the Murdoch Institute in Melbourne, Australia, and Borris Kemper of the University of Cologne, this assay uses an endonuclease from bacteriophage T4 to identify sequence mismatches in heteroduplexes.10 EMD is notable in that it identifies all kinds of mutations, including point mutations, deletions, and insertions in fragments up to several kilobases in size. And unlike the conformation-dependent assays, investigators can determine approximate location, which reduces the amount of sequencing that needs to be done for mutation confirmation. Amersham Pharmacia Biotech of Piscataway, N.J., offers PASSPORT kits based on this technology.

Also available is the structure-specific endonuclease, Cleavase®, which is the central element of Madison, Wis.-based Third Wave Technologies Inc.'s CFLP™ (cleavase fragment length polymorphism) assay. This enzyme probes secondary structural elements of single-stranded DNA fragments that have been denatured and allowed to renature; no heteroduplexes are formed in this process. Cleavage at the junction between single- and double-stranded regions creates a characteristic banding pattern for each fragment; changes in the banding pattern are symptomatic of altered bases. Third Wave's scientists have made recent improvements to the assay by incorporating temperature ramping into the procedure. This modification makes it possible to use identical conditions for all fragments without optimization, thereby increasing the throughput and lowering the costs associated with this assay.11

Madison, Wis.-based Epicentre's base excision sequence scanning system (BESS-T and BESS-G), another enzymatic detection assay, can be used for both mutation discovery and screening. This system generates T- and G-lane sequence ladders from PCR fragments using excision enzymes instead of processive polymerases and DNA chain terminators. Researchers can use the resulting fragment patterns to discover and map mutations. Like CFLP analysis, this method does not rely on heteroduplex formation or specialized gel conditions, and results can be obtained using ordinary gels or automated DNA sequencing gels.

Gaithersburg, Md.-based Trevigen offers yet another enzymatic kit. Marketed under the name MIDAScan™, the kit is based upon patented technology that uses DNA mismatch repair enzymes. MIDAScan combines two enzymes, Methanobacterium thermoautotrophicum TDG, which cleaves DNA at sites of T/G and G/G mismatches, and Escherichia coli Mut Y, which cleaves at sites of A/G mismatches. Together, the two enzymes can detect any mutation. The kit includes a control DNA that generates heteroduplexes with a mismatch every 20 bases, to aid mutation localization.

Finally, Ambion of Austin, Texas, has revitalized a 16-year-old enzymatic mutation detection technique to identify RNA/DNA heteroduplexes.12 Ambion's MutationScreener™ Kit increases the reaction's efficiency by using a combination of RNase 1 and RNase T1, which together cleave more mismatches than RNase A, with less nonspecific background cleavage. In addition, the kit uses double-stranded RNA as targets, which, because of their stability, can be added to the assay directly without extensive cleanup. This new assay is called NIRCA, for Non-Isotopic RNAse Cleavage Assay, and more than 50 reports in the literature describe using the technique for mutation detection in a wide variety of clinically important genes.13 Upping the Ante With Chips

A final mutation scanning approach uses microarrays, and Santa Clara, Calif.-based Affymetrix's variant detector array (VDA) might hold the record for the size of a region scanned for mutations. In a recent collaboration with several groups, including one at Case Western Reserve University, researchers scanned more than eight megabases of DNA using high-density oligonucleotide arrays.14,15 DNA from 40 unrelated individuals of three different ethnic origins was amplified, labeled, and hybridized to arrays of probes representing genomic, coding, and regulatory regions. It took 4,500 chips, each with 400,000 different probes, to perform this experiment; as a result, more than 15,000 SNPs were identified and deposited into the public SNP repository, dbSNP (www.ncbi.nlm.nih.gov/SNP/).

Mutation detection may have once been the sole province of geneticists and statisticians, but that is no longer true. There is now widespread interest in identifying and cataloging human variations, fueled at least in part by the availability of the entire human genome sequence. To meet this need, companies will continue to develop new technologies and refine old ones. To choose which approach to use in a given study, researchers must match the goals, scope, and cost of the study with the options available to them.

Laura DeFrancesco (defrancesco1@earthlink.net) is a freelance science writer
in Pasadena, Calif.;
Jeffrey M. Perkel can be contacted at jperkel@the-scientist.com.
References
1. A. Francis, "Variations on a gene," The Scientist, 14[15]:35, July 24, 2000.

2. R. Fodde, M. Losekoot, "Mutation detection by denaturing gradient gel electrophoresis (DGGE)," Human Mutation, 3:83-94, 1994.

3. M. Orita et al., "Rapid and sensitive detection of point mutations and DNA polymorphisms using the polymerase chain reaction," Genomics, 5:874-9, 1989.

4. V.C. Sheffield et al., "Attachment of a 40-bp G+C-rich sequence (GC-clamp) to genomic DNA fragments by polymerase chain reaction results in improved detection of single-base changes," Proceedings of the National Academy of Sciences (PNAS), 86:232-6, 1989.

5. Q. Liu, S.S. Sommer, "Parameters affecting the sensitivities of dideoxy fingerprinting and SSCP," PCR Methods and Applications, 4[2]:97-108, 1994.

6. C.H. Buzin et al., "Scanning by DOVAM-S detects all unique sequence changes in blinded analyses: Evidence that the scanning conditions are generic," BioTechniques, 28:746-53, 2000.

7. J.R. Mendell et al., "Diagnosis of Duchenne dystrophy by enhanced detection of small mutations," Neurology, 57:645-50, Aug. 28, 2001.

8. C.G. Huber et al., "High-resolution liquid chromatography of DNA fragments on non-porous poly(styrene-divinylbenzene)particles," Nucleic Acid Research, 21:1061-6, 1993.

9. L.M. Schriml et al., "Use of denaturing HPLC to map human and murine genes and to validate single-nucleotide polymorphisms," BioTechniques, 28: 740-5, 2000.

10. R. Youil et al., "Screening for mutations by enzyme mismatch cleavage with T4 endonuclease VII," PNAS, 92:87-91, 1995.

11. M.C. Oldenburg, M. Siebert, "New Cleavase Fragment Polymorphism method improves mutation detection assay," BioTechniques, 28:351-7, 2000.

12. R.M. Myers et al., "Detection of single base substitutions by ribonuclease cleavage at mismatches in RNA:DNA duplexes," Science, 230:1242-6, 1985.

13. M.M. Goldrick, "RNase cleavage-based methods for mutation/SNP detection: Past and present," Human Mutation, 18:190-204, September 2001.

14. D. Cutler et al., "High-throughput variant detection and genotyping using microarrays," Genome Research, in press.

15. J. Warrington, "New developments in variation detection using high density microarrys," paper presented at VI International Symposium on Mutations in the Human Genome, Bled, Slovenia, May 3-7, 2001.



Supplemental Materials