' 1993 Nature Publishing Group http://www.nature.com/naturegenetics
`
`review
`
`The rapid detection of unknown
`mutations in nucleic acids
`
`Markus Grompe
`
`The task of identifying mutations in nucleic acid sequences is a vital component of
`research in mammalian genetics. With the advent of the polymerase chain reaction,
`several useful mutation detection techniques have evolved in recent years. The different
`methods have complementing strengths and a suitable procedure for virtually any
`experimental situation is now available.
`
`Department of
`Molecular and
`Medical Genetics
`and Department of
`Pediatrics, Oregon
`Health Sciences
`University, 3181
`SW Sam Jackson
`Park Road,
`Portland, Oregon
`97201, USA
`
`Techniques for the detection of naturally occurring
`mutations that disrupt genes are of use to biologists in
`many fields. Inhuman genetics, these methods are used to
`determine whether a candidate gene is causally related to
`a phenotype and also to identify new alleles at known loci
`for diagnostic, population genetics and structure/function
`studies. Procedures for mutation detection can be
`separated into two distinct groups. The first consists of
`techniques which efflcientlyidentify known disease alleles,
`such as population screening for carriers of the common
`cystic fibrosis (CF) mutations’, 1. The second group consists
`of methods to scan sequences for unknown mutations
`and will be the subject of this review. While several useful
`technologies for the detection of sequence heterogeneity
`exist, no single method is applicable for all situations. The
`most appropriate screening technology is influenced by
`the expected nature of the mutation, size and structure of
`the locus in question, availability of mRNA, degree of
`sensitivity required and resources available.
`The spectrum of mutations ranges from cytogenetically
`visible chromosome rearrangements to micro-deletions
`and insertions and finally single base alterations. Some
`loci, such as the steroid suiphatase gene 3, are very prone to
`deletions, whereas others, such as the CF gene 4, are
`associated with hundreds of point mutations. Candidate
`loci can be as complex in structure as the dystrophin gene,
`with a 14 kilobase (kb) peptide coding region and 80
`exons5 or as simple as the SRY (male sex determination)
`locus with only one large exon6. In recessive autosomal
`and X-linked diseases the use of mRNA enables larger
`stretches of coding sequence to be examined together.
`Many genes of interest, however, may not be expressed in
`easily accessible samples such as blood or fibroblasts,
`nature genetics volume 5 october 1993 (cid:9)
`
`leaving analysis of genomic DNA as the only option.
`Finally, automation and fluorescence detection technology
`have many advantages, but the high cost limits their
`availability.
`
`Detection of large gene alterations
`Large gene alterations can be contrasted to single base
`changes and are mutations in which substantial portions
`(>500 basepairs (bp)) of a gene are deleted, duplicated or
`otherwise rearranged.
`
`Cytogenetic techniques
`Extremely large (> 1 megabase (Mb)) deletions and
`insertions can be detected by high resolution cytogenetics 7 .
`Recently, the power of cytogenetic analysis has been
`enhanced by the use of fluorescent in situ hybridization
`(FISH) 89. FISH uses fluorescently labelled DNA probes,
`which are hybridized to chromosome spreads, to detect
`not only the presence or absence of a chromosomal region
`corresponding to the probe, but also its position in the
`genome. Multiple probes labeled with different fluorescent
`dyes can be hybridized and analyzed simultaneously.
`FISH is particularly suited for the detection ofaneuploidy,
`micro-deletions or duplications and complex
`rearrangements, and has found awide range ofapplications
`ranging from the detection of loss of heterozygosity in
`tumour samples", microdeletion syndromes’ or
`duplications".
`
`Southern blot hybridization
`The Southern blot 13 remains one of the fastest methods to
`screen quickly for mutations, and offers a good first step
`in mutation analysis. The main advantage ofthe technique
`111
`
`GeneDX 1008, pg. 1
`
`(cid:9)
`

`

`review
`
`ey ' 1993 Nature Publishing Group http://www.nature.com/naturegenetics
`
`
`
`mutant DNA
`
`b (cid:9)
`
`2 3 4 5 (cid:9)
`
`a
`wild-type DNA
`
`*
`*
`
`denature (cid:9)
`
`single-stranded j
`confirmation
`
`- *
`
`*
`
`differential migration -I
`of single strands in a
`non-denaturing gel
`
`Wt (cid:9)
`
`In
`
`is that no detailed knowledge of the structure and sequence
`of a gene is required and a preliminary screen can be
`carried out with a probe of interest immediately after its
`isolation. Large deletions and insertions may be detected
`by the presence ofj unction fragments or changes in band
`intensities in blots (assuming equal amounts of DNA are
`present in all lanes). Point mutations may also be
`detectable, if they alter restriction sites. The restriction
`endonucleases MspI and TaqI have 4 base recognition
`sites (CCGG and TCGA respectively) that contain the
`mutation-prone dinucleotide CpG 14. These enzymes are
`therefore especially useful in scanning for mutations by
`Southern blot.
`
`PCR-based detection methods
`The ability to amplify DNA enzymatically via the
`polymerase chain reaction" represents the most important
`advance in mutation detection technology since the
`development ofDNA sequencing’ 6"7. All of the procedures
`described below rely on PCR amplification of sample
`DNA prior to analysis.
`
`Multiplex PCR for the detection of deletions
`Ifa locus ofinterest is prone to deletions and if its genomic
`sequences are known, the simultaneous PCR amplification
`of several sequences throughout the gene (multiplex
`amplification) is the most rapid and practical method for
`their detection. Deletions are indicated by the absence of
`some of the bands in the multiplex pattern. This method
`has found wide use in the diagnosis of Duchenne muscular
`dystrophy’s, in which 60% of cases represent deletions",
`
`112 (cid:9)
`
`Fig. 1 The principal of SSG analysis. a,
`Schematic representation. Wild-type
`(wt) and mutant (m) PCR products are
`shown at the top. The presence of a
`point mutation is represented by a dot
`on the DNA strands. Different single-
`stranded conformation lead to
`differential mobility in the gel. b, SSC
`analysis of exon 4 of the ornithine
`aminotransferase gene. Lanes 1 and
`5, normal controls: lanes 2-4,
`demonstrate band shifts indicating the
`presence of mutations.
`
`MONO (cid:9)
`
`WN (cid:9)
`
`but has been employed for other disorders" ,". In
`heterozygotes, deletions are seen as 50% reductions of
`band intensities in a quantitative analysis of the multiplex
`PCR reaction 22.
`The simultaneous amplification of several sequences of
`interest can also be useful in the search for single-base
`alterations (see below).
`
`Detection of single base changes
`Single-base alterations are the most common type of
`mutation at most loci. A number of methods may be used
`for detecting these subtle changes.
`
`Single-stranded conformational (SSC) analysis
`In 1989, Orita et al. first reported the use of single-
`stranded conformational polymorphisms (SSCP) to
`detect mutations". SSC analysis has since become the
`most widely used of the scanning technologies. Wild-
`type and mutant target DNA are amplified by PCR,
`denatured and then electrophoresed side by side through
`a (usually glycerol containing) non-denaturing
`polyacrylamide gel. The two single-stranded DNA
`molecules from each denatured PCR product assume a
`three-dimensional conformation which is dependent on
`the primary sequence. Ifa sequence difference (mutation)
`exists between wild-type and mutant DNA, this may
`result in differential migration of one or both of the
`mutant strands (Fig. I). PCR products with altered
`migration patterns can then be analysed by DNA
`sequencing to determine the exact nature of the alteration.
`The main advantage of SSC analysis is its simplicity and
`relative sensitivity. No additional steps are required after
`PCR and the gel system is relatively straightforward. The
`method also lends itself to the simultaneous examination
`of multiple samples. In most published studies the
`amplification products are rendered radioactive by the
`addition of 32P or 35S labelled dCTP to the PCR reaction.
`However, non-radioactive detection by silver-staining 24
`and ethidium bromide staining 11 have been successfully
`used. SSC analysis detects mutations but does not localize
`them within the fragment.
`SSC analysis detects 70-95% of mutations in PCR
`products of 200 bp or less 26’27. The sensitivity of the
`method decreases with the size of the PCR product and is
`less than 50% when fragments of >400 bp are analysed.
`Some of the size limitations of the method may be
`overcome by the simultaneous amplification and analysis
`of several fragments in one lane (multiplexing) or by
`restriction digestion of a larger amplification product
`prior to electrophoresis". The use of RNA generated by in
`
`nature genetics volume 5 october 1993
`
`GeneDX 1008, pg. 2
`
`

`

`' 1993 Nature Publishing Group http://www.nature.com/naturegenetics
`
`
`
`review
`
`vitrotranscription of PCRproducts also appears to improve
`detection of mutations in larger fragments 29 .
`
`Denaturing gradient gel electrophoresis (DGGE)
`DGGE also relies on differential electrophoretic migration
`of wild-type and mutant DNA for the detection of
`mutations. Double-stranded DNA (dsDNA) generated
`by PCR is electrophoresed through agradient ofincreasing
`concentration ofa denaturing agent (urea andformamide).
`As the DNA migrates, the strands progressively dissociate
`in discrete sequence-dependent domains of low melting
`temperature. This partial "melting" ofthe dsDNA leads to
`an abrupt decrease of mobility (Fig. 2). DNA molecules
`differing by only a single base substitution have been
`shown to exhibit differential mobility in such gels". The
`sensitivity of DGGE is greatly enhanced, however, if
`heteroduplex DNA between wild-type and mutant
`sequences is used for the analysis". Mutations can be
`found most reliably when the sequence heterogeneity lies
`within a domain of relatively low melting temperature.
`This can be achieved for virtually any sequence of interest
`by the use of computer programs to predict theoretical
`melting profiles and design PCRprimers". The attachment
`of a (30-50 bp) high melting temperature GC-clamp""
`to one PCR primer is critical to ensure that the amplified
`
`wild-type DNA
`
`mutant DNA
`
`denature and mix
`
`tero- and
`homoduplex DNA
`
`I (cid:9)
`
`electrophoresis in
`denaturing gradient
`
`differential
`melting behaviour
`
`low denaturant
`
`Wt wt+m
`= =
`
`- (cid:151)
`
`high denaturant
`
`Fig. 2 The principal of DGGE. Wild-type (wt) and mutant
`(m) PCR products are denatured and reannealed to
`generate 4 species of hetero- and homoduplexes.
`Differential melting behavior leads to altered migration in a
`gradient of denaturing agents compared to wild-type
`homoduplex.
`nature geneiws volume 5 october 1993 (cid:9)
`
`sequence has a low dissociation temperature.
`Once the appropriate PCR primers and denaturant
`conditions have been developed for a specific region,
`DGGE is a highly reliable and rapid method for mutation
`detection. Single base differences can be detected with
`(cid:150)95% accuracy in PCR products of up to 600 bp in length.
`Detection is usually carried out by non-radioactive
`means""’. As in SSC analysis, while the presence of a
`sequence difference is detected, its location within the
`fragment remains unknown and has to be determined by
`sequencing.
`The denaturing gradient can also be generated by
`temperature and this method is termed temperature
`gradient gel electrophoresis (TGGE)".
`
`Heteroduplex analysis (HA)
`If mutant and wild-type template sequences are present
`simultaneously in a PCR reaction, heteroduplexes between
`the two different DNA species can be formed during the
`late cycles. Heteroduplex molecules with a single bp
`variance may show differential mobility from
`homoduplexes in regular polyacrylamide gels’. This
`phenomenon is thought to be caused by sequence
`dependent conformational changes in the dsDNA.
`Recently, new gel matrices (Hydrolinkfiand MDETm from
`AT Biochem) have become available which markedly
`enhance the ability to detect mutation induced mobility
`shifts in heteroduplex molecules. These gel matrices are
`poured between glass-plates, analogous to polyacrylamide
`gels used for DNA sequencing. Mutant and wild-type
`PCR products are electrophoresed side by side and their
`mobility compared. Recent studies indicate a level of
`sensitivity similar to SSC analysis (80-90%) in small DNA
`fragments (<300 bp) 19 ’40. Both isotopic’ and non-isotopic’ 9
`detection methods have been reported. This technique is
`attractive because of its simplicity and has been applied
`successfully to the study of a number of human genetic
`disorders"".
`
`RNase A cleavage
`The principal of heteroduplex mismatch analysis
`applicable to both RNase and chemical mismatch cleavage
`(see below) is illustrated in Fig. 3. In RNase A cleavage, an
`RNA-DNA heteroduplexbetween a radioactive wild-type
`riboprobe and mutant DNA, generated by PCR, is
`subjected to cleavage by RNase A
`6. This enzyme will
`recognize and cleave single-stranded RNA at the points of
`mismatch. The reaction is analysed by electrophoresis
`and autoradiography. The presence and location of a
`mutation are indicated by a cleavage band of a given size.
`This method is limited by the need for radioactively
`labelled RNA and the fact that RNase A can only detect
`(cid:150)50% of mismatches. Ithas therefore been largely replaced
`by the similar technique chemical mismatch cleavage.
`
`Chemical mismatch cleavage (CIVIC)
`In CMC, a heteroduplex between a radiolabeled wild-
`type DNA molecule and mutant DNA (or RNA) is created
`by boiling and reannealing. Maxam-Gilbert sequencing
`chemistry ’ 7 is then used to chemically modify mismatched
`bases at the sites of mutations in the DNA(cid:150)DNA or DNA(cid:150)
`RNA heteroduplexes. The labelled DNA is cleaved by
`piperidine at the site of the modification, and this is
`followed by denaturing PAGE and autoradiography 47 .
`Osmium tetroxide is used for the modification ofmispaired
`
`113
`
`GeneDX 1008, pg. 3
`
`(cid:9)
`

`

`' 1993 Nature Publishing Group http://www.nature.com/naturegenetics
`
`
`
`review (cid:9)
`
`a
`
`9’
`
`heteroduplex between
`labeled wt and (cid:9)
`labeled mutant strand
`
`________
`
`mismatch
`
`electrophoresis
`detects cleaved
`fragment
`wild-type (cid:9) mutant
`
`chemical modification of
`mismatched bases by
`bydroxylamine or
`osmium tetroxide
`HA (cid:9)
`Oso-
`
`/
`mutation (mismatch)
`
`id
`
`piperidine cleaves
`modified mismatches
`
`(cid:151)
`
`(cid:9) - (cid:9)
`
`full length probe
`
`cleavage product
`
`mismatch
`
`4 (cid:9)
`___ (cid:9)
`
`5 (cid:9)
`
`6 (cid:9)
`
`7 (cid:9)
`
`8
`
`~40 4W
`
`b (cid:9)
`
`1 (cid:9)
`
`2 (cid:9)
`
`3 (cid:9)
`
`1196 - (cid:151)
`1025 -
`975 -
`
`775 -
`
`6W -
`
`445 -
`
`Fig. 3 The
`principal of
`heteroduplex
`mismatch
`cleavage, a,
`Schematic
`representation of
`chemical
`mismatch
`cleavage. A
`heteroduplex
`between labelled
`wild-type DNA
`(top strand) and
`unlabelled
`mutant DNA is
`shown. The
`mismatched
`base at the site
`of the mutation is
`chemically
`modified and
`subsequently
`cleaved by
`pipendine. b, 0MG analysis of the human omithine transcarbamylase cDNA 70.
`Cleavage products indicating the presence of mutations can be seen in all lanes
`except lane 4, which represents a normal control.
`
`Table I Advantages and disadvantages of mutation scanning methods
`
`Size Sensitivity Localizes
`
`Toxic
`chemicals
`
`Exon
`mRNA
`scanning scanning
`
`Best
`ref.
`
`250
`SSCP
`600
`DGGE
`1700
`0MG
`PGR DS 500
`300
`HA
`
`80% (cid:9)
`95% (cid:9)
`>95% (cid:9)
`>99% (cid:9)
`80% (cid:9)
`
`no
`no
`yes
`yes
`no
`
`none
`formamide
`yes
`none
`none
`
`.i-++
`++
`+
`++
`++
`
`+
`++
`+++
`++
`+
`
`26
`67
`68
`69
`40
`
`The size of fragments which can be analysed is given in bp. Sensitivity, the
`approximate % of mutations detectable with the method is given; localizes,
`indicates whether a method provides the exact location of the mutation within
`the examined fragment; +++, indicates a high, ++ moderate and + limited
`usefulness of the method for this application.
`
`114 (cid:9)
`
`thymines and hydroxylamine for mismatched cytosines.
`Adenosine and guanosine mismatches of the wild-type
`sense strand are detected by also labelling the anti-sense
`strand of wild-type DNA in the heteroduplex. CMC is
`very sersitive, detecting >95% of mismatches when only
`the wil4-type DNA is labelled and 100%, when both wild-
`type and mutant DNA are labelled. CMC has the lowest
`size constraint of all the mutation scanning methods.
`PCR products of up to 1.7 kb in length have been examined
`successfully49, permitting efficient screening of amplified
`mRNAs. As well as detecting the sequence alteration
`reliably, the precise localization and nature of the change
`is also indicated by the size of the cleavage band and the
`cleaving reagent. Non-isotopic labelling techniques’ and
`multiplexing" have been reported for this technique.
`A highly sensitive variation of CIVIC has been described
`by Ganguly and Prockopt 2. Chemical modification of the
`mismatched base(s) is carried Out with carbodiimide
`(CDI), and this is followed by primer extension using a
`radiolabeled oligonucleotide. Taq polymerase terminates
`extension at modified bases and the shortened extension
`product is detected byautoradiograpby. This method has
`been applied primarily for the detection of mutations in
`collagen genes.
`
`Direct sequencing
`All the methods described above are capable of detecting
`mutations with varying efficiencies, but none defines
`preciselythe nature ofthe change. DNA sequencing defines
`the location and nature of the a change and therefore is a
`necessary final step of any mutation detection method. If
`DNA sequencing can be made rapid, accurate and efficient,
`it represents the ideal mutation "scanning" technique.
`Direct sequencing (DS) refers to the direct sequence
`analysis of PCR products without prior subcloning into
`sequencing vectors and can be the primary method of
`mutation detection. As this methodbecomes more efficient
`with the use of automation and new fluorescence detection
`technology, it is likely to become the primary method of
`mutation detection. Direct sequencing is based on the
`Sanger dideoxy chain termination method’ 6, and both
`isotopic as well as fluorescence detection techniques have
`been described. Fluorescence based sequencing methods
`have many advantages, but require expensive specialized
`equipment.
`In order to sequence PCR products successfully by the
`conventional dideoxy termination protocol, it is essential
`to convert the double-stranded PCR product into a single-
`stranded sequencing template. Several methods have been
`described to achieve this. In the first technique, termed
`asymmetric PCR53, the PCR product is reamplified in a
`second reaction in which one of the oligonucleotide
`primers is in >100:1 excess of the other. This creates an
`abundance of single-stranded PCR product for sequencing.
`In the second method one of the PCR primers is
`biotinylated. After the reaction, the double-stranded PCR
`product is captured on an avidin-coated magnetic bead" ’64.
`The non-biotinylated strand is melted away with NaOH
`and the sequencing reactions are carried out on the
`immobilized single-stranded template. Inathird approach,
`termed "genomic amplification with transcript
`sequencing" (GAWTS) the original PCRprimers carry T7
`RNA polymerase binding sites". In vitro transcription is
`then used to generate single-stranded RNA template for
`sequencing.
`
`nature genelzcs volume 5 october 1993
`
`GeneDX 1008, pg. 4
`
`(cid:9)
`

`

`' 1993 Nature Publishing Group http://www.nature.com/naturegenetics
`
`
`
`review
`
`techniques over the past few years has been rapid, and new
`and promising advances are being tested in many
`laboratories. Three of these new methods will be
`mentioned here. The first is termed dideoxy fingerprinting
`(ddF). ddF represents a combination of SSC analysis and
`direct sequencing". After amplification, the PCR product
`is sequenced with dideoxyCTP (ddCTP) to generate a "C-
`ladder" (fingerprint) of bands. The sequencing reactions
`from wild-type and mutant samples are then
`electrophoresed through a SSC (non-denaturing
`polyacrylamide) gel. Mutations are detectable as shifts of
`individual bands inthe ladder. This method, when applied
`to the factor IX gene, was highly sensitive (84/84 alleles
`detected) and also provided some information about the
`localization of the sequence alteration.
`The second new technique utilizes a computer software
`variation in the analysis of fluorescent cycle sequencing
`data60. Fluorescent sequencing is carried out in routine
`fashion, and all 4 termination reactions are loaded into a
`single gel lane. A set of 4 adjacent lanes is then used for the
`parallel analysis of the sequence ladder generated by a
`single dideoxy terminator (single color analysis). This
`permits the detection of additional peaks which indicate
`mutations. Extra peaks are detectable if up to 4 individual
`sequencing reactions are loaded into a single lane, thus
`permitting the simultaneous analysis of 16 samples. This
`method appears especially useful when high throughput
`automated sequencing is available.
`Recently, E. coli mismatch repair enzymes have been
`used for the detection ofknown mutations in oncogenes 61 .
`If these enzymes are able to detect mismatches in larger
`heteroduplex molecules containing unknown mutations,
`they may find wide application in mutation detection.
`
`r fT (cid:9)
`L (cid:9)
`.cj (cid:9)
`
`fli (cid:9)
`
`" fl (cid:9)
`
`fl (cid:9)
`
`r
`
`( (cid:9) (cid:149)1 (cid:9)
`i.i (cid:9)
`U (cid:9)
`L (cid:9)
`11 C)
`r (cid:9)
`
`(
`L
`
`r
`
`Fig. 4 Direct sequencing with fluorescent primers. A
`portion of the human CD4 genomic locus was sequenced.
`In the position marked by an N, a G and I peak appear in
`the same position indicating a sequence alteration on one
`chromosome.
`
`A recent promising modification of dideoxy chain
`termination sequencing technology is cycle sequencing,
`also termed amplification sequencing’"’. In this method,
`the template is simultaneously amplified and sequenced
`by the addition of dideoxy terminators to a PCR reaction.
`This technique has the advantage of requiring onlyminute
`amounts of template. When isotopic detection methods
`or conventional fluorescence sequencing protocols are
`used, however, cycle sequencing requires that the
`sequencing primer be end-labeled (either by an isotope or
`by a fluorochrome). Recently, a new protocol based on
`cycle sequencing and fluorescence detection technology
`has been developed 58. Fluorescently labeled dideoxy
`terminators (dye terminators) are used with a different
`fluorescent dye coupled to each of the four ddNTPs. The
`extension product is labelled at the site of the dideoxy
`termination rather than at the sequencing primer end. In
`this direct sequencing protocol, neither specially modified
`PCR or sequencing primers are needed.
`An example of a result obtained by fluorescent
`technology direct sequencing is shown in Fig. 4.
`
`Future methods
`The pace of improvements in mutation detection
`
`Mutation versus polymorphism
`The methods described above will detect nucleic acids
`alterations but do not necessarily define their biological
`significance. Some sequence changes, such as those causing
`frame-shifts or chain terminations, are obviously
`functionally deleterious to a protein. Other alterations,
`such as missense mutations leading to amino acid
`substitutions or base changes in introns and untranslated
`regions, can be either functionally relevant or represent
`sequence polymorphisms. Additional information
`including family studies, population
`studies or functional assays after in
`vitro expression may be needed to
`Table 2 Genomic DNA versus mRNA as starting material for mutation analysis
`determine the significance of a
`… sequence alteration.
`
`Genomic DNA
`
`mRN
`
`Advantages (cid:9)
`
`Easily accessible (blood)
`In autosomal loci both alleles
`are equally represented
`Mutations in the promoter and
`intronic splice junctions can be
`detected
`First choice in autosomal
`dominant traits
`
`Disadvantages Genomic sequence and gene structure
`information are needed
`Only small segments of coding region
`(exons) are analysed
`More PCR reactions
`
`nauregenehco volume 5 october 1993 (cid:9)
`
`Long segments of peptide coding
`region can be analysed
`Gene structure information not
`needed
`Fewer PCR reactions
`Aberrant mRNA sizes can be seen
`First choice in X-linked traits
`
`Gene may not be expressed in
`accessible specimens
`In autosomal loci only one allele
`may be represented
`Mutations in the promoter and
`intronic splice junctions are not
`detected
`
`Comparison of methods
`The relative usefulness of the
`techniques mentioned above for the
`detection of single base alterations
`depends upon a number of criteria.
`Table 1 summarizes the advantages
`and disadvantages of the different
`procedures. Obviously, the level of
`expertise for a certain procedure, and
`the equipment available, are other
`determining factors.
`
`Size factors: mutant mRNA
`versus genomic DNA
`The two principal materials on which
`
`115
`
`GeneDX 1008, pg. 5
`
`

`

`review (cid:9)
`
`' 1993 Nature Publishing Group http://www.nature.com/naturegenetics
`
`mutation analysis can be carried out are the mRNA and
`genomic DNA. Several considerations are important in
`deciding which material to use for mutation detection
`(Table 2). If mRNA is readily available (from cultured
`cells or pathological specimens), it is usually the most
`practicalstarting material because larger stretches ofcoding
`sequence can be scanned. The use of specialized PCR
`protocols may permit the amplification of mRNA from
`tissues in which the gene is expressed at only very low
`levels (illegitimate transcription)62’ 63(cid:149) The mode of
`inheritance of a phenotype for which mutations are being
`sought also has an impact on the choice ofstarting material.
`mRNA analysis is less desirable for dominant conditions,
`because the mutant allele may not be represented in
`mRNA (null alleles).
`In mRNA analysis, reverse transcription is used to
`generate first strand cDNA, and this is followed by PCR
`amplification to generate large (1-2 kb) fragments from
`cDNAM. When genomic DNA is used, smaller segments
`of coding sequence (exons) usually are amplified. The
`different mutation detection techniques are not equally
`useful for both approaches. SSC and HA are particularly
`advantageous for the study of small DNA fragments
`(<300 bp), notably in the exon analysis of genomic DNA.
`Direct sequencing and DGGE are sensitive up to 500-600
`bp, and CMC can be used for fragments up to 1700 bp.
`The methods capable of scanning large segments of
`DNA are also useful in the study of organisms with few
`introns such as bacteria and viruses 6
`
`.
`
`Sensitivity and accuracy
`As mentioned above, the sensitivity of the different
`mutation detection procedures is dependent on the size of
`the fragment to be analysed. Even when scanning small
`DNA fragments, however, SSC analysis and HA are not as
`sensitive as DGGE, CMC and direct sequencing.
`Sequencing and CMC are the onlymethods that accurately
`localize the mutation within the examined fragment.
`
`Candidate gene versus known defective gene
`If a gene is known to be defective (in a patient with a
`genetic disease), the requirements in mutation analysis
`are different than when a gene is a candidate, but not
`already known to carry a mutation. Known defective
`
`genes are often examined in a diagnostic setting, in
`which a high degree of accuracy is needed. Frequently
`the same locus is analysed repetitively in large numbers
`of patients or in populations for carrier detection. In
`these situations, it may become practical to design
`specialized oligonucleotides for direct sequencing
`(biotinylated or fluorescent) or DGGE (GC-clamps).
`Direct sequencing and DGGE are more sensitive than
`SSC analysis or heteroduplex analysis and are therefore
`frequently used in this setting, even for the testing of
`small DNA fragments. The specialized set-up required is
`worthwhile because of the large sample numbers and
`because of the degree of accuracy required. This is in
`contrast to a candidate gene analysis, in which any given
`locus may or may not have mutations. Specialized
`reagents and equipment are less practical in this scenario,
`and it is less critical to detect 100% of mutations, if
`enough patient samples are available. Often details of the
`structure of the candidate locus are not known and
`examination of mRNA (only coding sequence needed) is
`desirable. In this context, the size of the sequence to be
`examined usually determines the method of choice (see
`above).
`
`Radioactivity and toxicity
`In all procedures described, the detection methodology
`can be either radioactive or non-radioactive. SSC, DGGE
`and heteroduplex analysis can be easily carried out non-
`isotopically with simple ethidium bromide staining of
`DNA. This is less simple for CMC and direct sequencing,
`for which fluorescence detection is needed. Of all the
`techniques, CMC clearly has the most problems with the
`use of toxic chemicals. Both osmium tetroxide and
`piperidine are highly noxious and not suitable for use in
`a clinical laboratory environment.
`
`Simplicity and speed
`SSCP, HA and DGGE (once set up) are clear front runners
`for simplicity, in that the analysis of the PCR product can
`be carried out immediately after the reaction. CMC and
`sequencing both require several manipulations after PCR
`and are more time-consuming. Automation and dye
`terminator cycle-sequencing, however, are making direct
`sequencing a very rapid approach.
`
`116 (cid:9)
`
`naturegeneflcc volume october1993
`
`GeneDX 1008, pg. 6
`
`

`

`' 1993 Nature Publishing Group http://www.nature.com/naturegenetics
`
`
`
`review
`
`I would like to
`thank R.A. Gibbs
`for providing Figure
`4 and for his critical
`review of the
`manuscript. Figure
`lb was used with
`the kind permission
`of D. Valle.
`
`rain.
`
`1. Ghettab, F.F. & Wail, J. Detection of multiple cystic fibrosis mutations by
`reveraedot blot hybridization: atectrnoiogyforcanier screwing. Hum. Garret.
`89,163--168 (1992).
`2. Ng, I$. etal. Methods for analysis of multiple cystic fibrosis mutations. Hum.
`Garret. 87, 813-617 (1991),
`3. Ballablo, A. of at Molecular heterogeneity of steroid sulfatase deficiency: a
`multicenter study on 57 unrelated patients, at DNA and protein lassie.
`Genom/cs 4, 36-40 (1989).
`4. Tsui, L.C.Thespectrumotoyatic fibrosis mutations. Trends Genet. 8,392-398
`(1992).
`5. Koenig, M., Monaco, A.P.&Kunkel, L.M.Thecompletesequenceofdystrophln
`predicts a rod-shaped cytosiceletal protein. Cell 53,219-226(1988).
`6. Sinclair, A.H.etal. A gene from the human sex-determining region encodes a
`protein with homology to a conserved DNA-binding motif. Nature 348,240-
`244 (1990).
`7. Rooney, D.E.&Czepxilkowsk B.H. Human cytogerietics:apractica’approach
`RL Press, Oxford, New York, 1992).
`6. Zhang, P.R., 1-leilig, ft. Thomas, G. & Aurias, A. A one-step efficient and
`specific non-radioactive non-fluorescent method for in situ hybridization of
`banded chromosomes. Chromosome 99,436-439(1990).
`9. Bsum,ann, H., Chrif, D. & Berger, R. Interphase cytogenetics by fluorescent
`In situ hybridization (FISH) for characterization of monosorny-7-associated
`myeloid disorders, Leukemia 7, 384-391 (1993).
`10. Taylor, C. atal. Diagnosis of Ewing’ssarcomaand penpheralneciroectoderrnal
`tumour based on the detection of 1(1 122) using fluorescence in situ
`hybridization. Br. J. Cancer 67, 128-133 (1993).
`11, Kuwano, A., Ledbetter, S.A., Dobyns, W.B., Emanuel, B.S. & Ledbetter, D.H.
`Detection of deletions and cryptic trarrslocations in Miller-Dlekersyndrome by
`In situ hybridization. Am. J. hum. Garret. 49, 707-714 (1991).
`12. Lupsis, J.R. at at DNA duplication associated with Charcot-Marie-Tooth
`disease type 1A. Cell. 66,219-232(1991).
`13. Southern, E.M. Detection of specific sequences among DNA fragments
`separated by gel electrophoresis. J. mo/ac. Biol. 98,503-517 (1 975).
`14. Cooper, D.N. & Youssoufian, H. The GPO dinucleotide and human genetic
`disease. Hum. Garret. 78,151-155 (1988).
`15. Mullis, KB.&Faioona, F.A.Specific synthesisof DNA in vitrovias polymerase-
`Catalyzed chain reaction. Math. Ertzymol. 155,335-350(1987).
`16. Sanger, F., Mcklen,S.&Couleon,A,R. DhtA