Detectionof a-Thalassemiasby MultiplexPolymeraseChainReaction

CUN. CHEM.40/12, (1994) 2260-2266 #{149} Molecular Pathology Detection of a-Thalassemias by Multiplex Polymerase Chain Reaction Lemuel J. Bowie,”4 Poluru L. Reddy,’ Moolky Nagabhushan,2 Although a-thalassemia is the most common genetic abnormality in the world, there is currently no routine laboratory method to definitively identify individuals who are affected. We describea rapid and simple method that utilizes deletion-sensitive primers to amplify normal DNA sequences. Deletions involving the regions responsible for most of the a-thalassemia cases in the US prevent amplification with these primers. In tests with DNA isolated from small amounts (10 L) of whole blood, the deletion-sensitive primers gave rise to the expected 248and 375-bp (base pair) amplification products in normal individuals. These primers, along with primers designed to bind to a nonaffected control sequence from the hemoglobin p chain, could be amplified simultaneously (multiplex polymerase chain reaction). This made it possible to detect heterozygotes for a-thalassemia-2 (one a locus deleted) by determining the ratios of the 248- and 375-bp amplification products to the product of the control sequence (268 bp). The method is rapid and simple and can be performed in a routine clinical laboratory. IndexIngTerms:heritable disorders/hemoglobin defects/gene am- plification/anemia The mild a-thalassemia (deletion of one of the the most common Thalassemia can mutation in the that leads to reMore than 95% of a-tha]assemia conditions are caused by deletions of one or more a-globin genes (4); structural mutations alone rarely cause a thalassemic syndrome (4,5). The resulting imbalance between a and p chain production leads to accumulation of excess 9 chains, which are unstable and precipitate out within the cell, thus altering the membrane and causing the clinical manifestations (6, 7). In a normal individual, the two a genes along with three other pseudogenes (.1’s,,.Ira2, and .fra,) and a gene on the of undetermined function (Or) are all located short arm of chromosome 16 in the order: telomere-fi1ia-i/ia1-a-a1-O1-centromere. If a single deletion occurs on only one chromosome, the condition is clinically silent. Among black Americans, this occurs with a prevalence of 2 7.5% (8), representing a gene frequency of 0.16. Given this prevalence, >8 million black Amergenotype four genes for the hemoglobin a chain) is genetic abnormality in the world (1-3). result from either a deletion or a point a-globin gene or a controlling sequence duced synthesis of hemoglobin a chains. ‘Department of Pathology and Laboratory Medicine, Evanston Hospital, 2650 Ridge Ave., Evanston, IL 6020 1-1783, and DepartUniversity Medical School, Chiment of Pathology, Northwestern cago, IL. of Pathology, Cleveland, OH 44706. 3Pharmacia LKB Biotechnology, Inc., Uppsala, Sweden. 4Author for correspondence. Fax 708-570-2969. Received August 5, 1994; accepted September 12, 1994. 2260 CUNICAL CHEMISTRY, Vol. 40, No. 12, 1994 and Pierre Sevigny3 icans could be affected by this condition. The majority of these single deletions involve a 3.7-kilobase (kb) located primarily on the right side (-a37; deletion “rightward deletion”) of the a locus (5) [a “leftward deletion” results in the loss of a 4.2-kb segment, but this condition is rare in the overall US population (9)]. If the deletion occurs on both chromosomes, a clinically observable thalassemia occurs (a-thalassemia trait). This condition occurs with an incidence of almost 3% in American blacks (8). It is, therefore, likely to be a significant cause of microcytic anemia that does not respond to oral iron therapy in black children (4). If three loci are deleted, severe thalassemia occurs (hemoglobin H disease). In individuals from Southeast Asia and the Mediterranean areas, the most common or deletion is one that involves both a loci (- -SEA - -MED) on the same chromosome. Inheritance of this defect on both chromosomes results in no a chain production, and affected fetuses often die in utero. The prevalence of the - - SEA mutation is 3-5% in Southeast Asia (10), and that of the - -MED deletion is <1% in the Mediterranean basin (10); the frequency of these conditions in the US is unknown. The recent development of simple, sensitive, rapid DNA technologies have made it possible to develop useful prodiseases that were cedures for the diagnosis of various previously impossible or too cumbersome and costly to perform. Because most a-thalassemia mutations are due to deletions of one or more a loci on chromosome 16 (4, 5), a specific strate’ to detect these deletions (and differentiate them from nondeletional mutations) would be useful. Thus, several attempts have been made to develop a DNAbased test for diagnosing a-thalassemias (2,11-13). These methods take advantage of the sensitivity of polymerase chain reaction (PCR) techniques, but some also require the use of restriction enzyme cleavage of the amplified product (11, 12). In before analysis for adequate interpretation addition, none of those methods attempts to reliably and objectively distinguish between homozygetes, heterozygetes, and normal subjects by using quantitative PCR techniques (13, 14). Here we describe a method to detect high-frequency deletional types of a-tha]assemia (a37, -SEA, - -MED), using multiplex polymerase chain reaction (M-PCR) techniques (15,16). The method does not enzymes and provides clear require the use of restriction analytical criteria for distinguishing between various types of deletions. Materials and Methods Materials The regions of frequent gene deletions for the a1 and a2 globin loci are represented in Fig. 1. Primers were designed to be complementary to regions affected by deletions (-a37 and --SEA/-MED; Table 1). In addition, a specific pair of primers was synthesized to amplify a region surrounding the 5’ end of the f3-globin gene and was used as a positive control for amplification efficiency. Potential primer pairs were evaluated for maximum specificity, optimal GC content, and freedom from intramolecular structures by use of commercial software (OLIGO”; National Biosciences, Plymouth, MN). The a- and $-globin primers, synthesized by the phosphoroamidite method on a DNA synthesizer (Gene Assembler PlusT”; Pharmacia Biotechnology, Piscataway, NJ), were then deprotected by treatment with concentrated ammonia for 1 h at 55#{176}C and passed through “NAP 10” spin columns (Pharmacia Biotechnology) as indicated by the manufacturer. The absorbance of primer solutions was read at 260 and 280 nm, and the solutions were diluted to an absorbance of3.OA (-1.0 mol/L). PCR For PCR amplifications, we used a Gene Amp” System 9600 thermal cycler from Perkin-Elmer Cetus (Norwalk, CT). Mononucleotides (dATP, dCTP, dGTP, dTTP) for PCR amplifications were obtained from Pharmacia Biotechnology. Recombinant Taq polymerase as well as unlabeled /3-globin primers (GH2O and PCO4) Cetus; agarose was were obtained from Perkin-Elmer from FMC BioProducts (Rockland, ME). Purified human placental DNA was obtained from Oncor (Gaithersburg, MD). Residual EDTA-anticoagulated blood samples from specimens collected for hematological analyses were used for the PCR DNA amplifications. We evaluated the reliability of the M-PCR technique with DNA samples previously confirmed by Southern transfer analysis to have specific mutations (obtained from Stephen Thibodeau, Mayo Clinic, Rochester, MN). These samples were coded so that the genotypes were not known until the analysis and interpretations were completed. Procedures DNA isolation. Template DNA was prepared by a modification of the method of Lewin and StewartHaynes (17). Whole blood (10 L) was hemolyzed with 500 L of distified water and centrifuged. The pellet was then washed two times with 500 L of distilled water in a 1.5-mL microcentrifuge tube. We dissolved the DNA pellet in 10 tL of 0.2 mol/L KOH containing 0.05 mol/L this at 65#{176}C for 5 miii in a dithiothrietol, incubated 375bp 5’ a2 tIl water bath, and then added 1 1zL of 1.0 mol/L HC1. After mixing, we added 10 tL of 0.6 mol/L Tris-HC1, pH 8.3, containing 0.2 mol/L KC1, followed by 20 L of sterile distified water and heated the solution at 99.9#{176}C (on the System 9600 thermal cycler) for 8 mm to destroy any residual nucleases. Aliquots (2-4 L) of the above extract were mixed with PCR reaction mixture components to yield a total volume of 20 L for all amplifications. PCR amplification. All of the PCR amplifications were performed with stock PCR amplification buffer containing (per liter) 500 mmol of KC1, 100 mmol of Tris-HC1, 15 mmol of MgC12, and 0.1 g of gelatin (pH 8.3). The amplification reaction mixture contained PCR amplification buffer (10-fold-diluted stock buffer), 0.2 pmol of primers, 2.0 nmol of dNTP (dATP, dCTP, dGTP, and dTTP), DNA template (200-1600 ng), and 0.25 U of Taq polymerase in a total volume of 20 pL. No overlaying oil was necessary. The primers used to detect the _a&7 deletion were F37.1F and F37.3 (Table 1), the sense and antisense strands, respectively. The primers for the Southeast Asian and Mediterranean deletions were F13F and F14. The cycling sequence was as follows: an initial DNA denaturation step of 94#{176}C for 3 miii followed by 35 cycles of denaturation at 94#{176}C for 30 s, annealing at 55#{176}C for 60 s, and extension at 72#{176}C for 60 s. A final extension at 72#{176}C for 3 mm was followed by cooling at 4#{176}C. Samples were held at 4#{176}C on the thermal cycler or in a refrigerator until used for further analysis (e.g., agarose gel electrophoresis, polyacrylainide gel electrophoresis, or cycle sequencing). The amplified products resulting from amplification of a normal specimen with each of the primer pairs were 375, 248, and 268 base reaction pairs (bp) long (Table 1). A blank amplification containing all reagents but without DNA was used as a negative control. A sample of human placental DNA or known normal human leukocyte DNA was also included in each set of experiments as a positive control. Multiplex PCR amplification. Using the optimal conditions for the amplification of individual DNA fragments as a starting point, we investigated the conditions required for the optimal amplification of all three fragments in a single reaction. The optimal reaction conditions were found to be the same as for the individual PCR reactions. Therefore, we used the same cycling conditions as described above, cooling the samples to 4#{176}C after cycling. To analyze the PCR products, we used 248bp al til Fig. 1. Diagram of a-globin gene cluster on chromo- some 16. The 01locus does not code for any known functional protein; and a1 loci code for identical proteins but haveminor differences in their DNA sequences.The codingsequences are separated by noncodingsequences (introns, (exons, #{149}) 0). The most frequently occurring deletions (-a37, - -SEA, and - -MED) are indicatedbythe black bars below thegenes. Dashed lines at the endof the deletions represent the extent of the breakpoints (15). The amplification products (and their lengthsin base pairs)used for detecting these deletionsare indicated by the hashed bars above the genes. The coding region on chromosome 11 for the -globin gene andthe binding sites for the control primers GH2O and PCO4 are not shown. the III-3’ - a a 3.7jj - a 3#{149}7iii Deletion - - MED - - - Deletion Deletion Deletion SEA Deletion a2 CLINICAL CHEMISTRY, Vol. 40, No. 12, 1994 2261 Table 1. PrImer sequences. Am Am Pdmsr Pdm esqn F13F 5’-FACTGGAGAGGAGAGCGGGGC-3’ F14 5’-AGGCTGCGGGMGGACATcA-3’ length, bp 248 nans SENMED F37.1F 5’-FACcATCCCCCCAA.AAAAcAT-3’ 375 F37.3 a 5’-CTCTMCCATGACAcMGTA-3’ GH2OF 5’-FGMGAGCcMGGACAGGTAC-3’ 268 Control 5’-CAACTrCATCCACGTFCACC-3’ a The suffix“F” in the primer name indicates that itis a fluoresceinlabeled derivative. PCO4 on a 2% agarose gel (Metaphor”; FMC BioProducts) with Tris-acetate/EDTA buffer (40 and 1 mmol/L, respectively, pH 8.0) containing ethidium broof DNA. The PCR mide (0.5 mgfL) for visualization reaction mixture (6 i.i.L) was mixed with 2 L of loading buffer (2.5 gIL bromphenol blue, 2.5 gIL xylene cyanol FF, and 400 gIL sucrose in water) and run on agarose gel at 100 V for 45 miii at room temperature. The gel was viewed under ultraviolet illumination and photographed. Fragment analysis (quantification). To quantiQy relative peak areas of amplified products, we separated the M-PCR reaction products with a DNA sequencer (ALF”‘; Pharmacia Biotechnology), using a 6% polyseparation. Elecacrylamide gel for the electrophoretic trophoretic conditions were 1500 V and 34 W, and the gel was thermostated at 42#{176}C. The samples-i L of PCR reaction mixture and 1 pL of stop solution (deionized formamidefBlue Dextran; Pharmacia Biotechnology)-were denatured at 94#{176}C for 3 mm and then were loaded on the gel, which can accommodate 40 samples in each run. The fluorescent amplified products were detected with a laser power of 3 mW. Peak areas were with fragment analysis softdetermined by integration ware (Fragment Manager”‘; Pharmacia), and the ratios of the 248- and 375-bp amplicon areas to the 268-bp control amplicon area were calculated. To determine fragment sizes, we included a i00-bp DNA ladder (Pharmacia Biotechnology) in one of the lanes. The DNA ladder was labeled with fluorescein by treatment with fluorescein-labeled 12-dUTP (BoehringBiochemicals, Indianapolis, IN), as foler Mannheim lows: We added 2 L each of DNA ladder (1 g/L), fill-inbuffer (200 mmol/L Tris, pH 7.6, containing 500 mmol/L KC1, 100 mmol/L MgSO4, 40 mmol/L MgCl2, and 0.5 g/L bovine serum albumin), and fluorescein-12-dUTP (1 mmol/L) to 1 pL of Klenow fragment (Pharmacia) and 13 L of distilled water. After incubating the mixture we stopped the reaction for 90 mm at room temperature, with 2 L of 0.4 mol/L EDTA. The labeled DNA was precipitated by adding 2.5 L of 4 moIfL LmCl and 75 jtL of prechilled (-20#{176}C) absolute ethanol and then incuelectrophoresis 2262 CLINICAL CHEMISTRY, Vol. 40, No. 12, 1994 After centrifugation for 20 bated for 30 mm at - 70#{176}C. mm at 1200g, the precipitated DNA was washed with cold ethanol (700 milL) and dissolved in 50 pL of Tria/ EDTA buffer. This solution is stable for at least 6 months stored at -70#{176}C. Sequencing of PCR products. To confirm that the amplifled products were the result of the binding of the primers to the intended region, we determined the DNA sequence of each amplicon. The individual amplified products were purified from agarose gels after electrophoresis by adsorption onto glass beads (Sephaglas Band Prep Kit; Pharmacia Biotechnology). The yield and purity of the amplicons were determined qualitathe fluorescence of the ethidium tively by comparing bromide-stained bands with that of a standard DNA the ratio of marker and quantitatively by determining absorbances at 260 and 280 nm of samples and DNA standards by spectrophotometry. The DNA cycle sequencing protocol was slightly modified from that of the fmolm DNA Sequencing System (Promega, Madison, primers (1.0 WI), by substitution of fluorescein-labeled mol/L) for radiolabeled primers. DNA sequence was determined with the ALF” DNA Sequencer (Pharmacia Biotechnology). Results Figure 1 shows the relationship between the genes (a1 and a2), the most common that code for a chains deletions in the US, and the binding regions of the deletion-sensitive primers. Each a gene has three coding sequences (exons; black bars) separated by noncoding sequences (introns; white bars). The most common mutations resulting in a-thalassemia in the US give rise to deletions of -3.7 kb (-a37 deletion). This deletion results in a single hybrid a gene, composed of a 5’ a2 sequence and 3’ a1 sequence. Three subclasses of the -a37 deletion (_a3?i, _a3#{149}7ii, and _a37m) have been detected by Southern transfer hybridization (18). Individuals from Southeast Asia and the Mediterranean have a high prevalence of deletions that remove both the a2 and the a1 loci (- -SEA and - -MED, respectively). All five deletions are represented as black bars in Fig. 1, with the lines at the ends of the black bars indicating heterogeneity with regard to the exact length of these deletions. The cross-hatched double lines above the genes represent the amplification products from the binding of primers to the target sequences on the sense and anti-sense strands of regions affected by the -a37, -SEA, and - -MED deletions. The primers that bind to the right of the a2 locus (F37.1F and F37.3) generate amplification products only if the -a37 deletion is not present. Although the other primers generate a 375-bp product with all three subclasses of -a37 deletions, we did not attempt to identify the subclasses (e.g., by using restriction enzymes). The primers that bind to the right of the a1 locus (F13F and F14) generate amplification products only if the --SEA, --MED, or other similar deletions are not present. The primers GH2OF and PCO4 are known to generate a 268-bp product from the region -200bpS’ to the beginning of exon 1 of the gene for the hemoglobin /3 chain and extending through most of exon 1 (19). We used this sequence, located on chromosome 11, as a control for evaluating amplification efficiency. Figure 2 shows the PCR-amplified products resulting from the use of these primers in PCR reactions with human placental DNA (Oncor) after agarose electrophoresis (Fig. 2, left) or polyacrylamide gel electrophoresis (Fig. 2, right). Polyacrylamide gel electrophoresis on the ALF system allows for the determination of peak areas and the calculation of relative amounts from the ratios of these areas. The identity of the bands was confirmed by isolating the individual bands from the agarose gel and sequencing the DNA with a cycle sequencing protocol on the ALF (see Materials and Methods). Earlyeluting peaks (100-200 bp) represent nonspecific amplification products and do not affect the quantitation of target peak areas. When all three pairs of primers were mixed together for M-PCR, the three fragments were amplified without significant production of other products. Fig. 3 shows the three amplification products obtained from various concentrations of placental DNA (10-80 mg/L) in the M-PCR reaction mixture. As shown, all three amplillcation products are present and resolved both on agarose and polyacrylamide. However, reliable quantification of peak areas was achieved only from the polyacrylamide gels (Fig. 3, right). The ratios of amplicon peak areas (248- and 375-bp products) to the area of the control peak (268-bp product) are summarized in Table 2. The ratios of the amplified products to the control were relatively constant but the reproducibility became unacceptable (CV >10%) at DNA concentrations <20 mgIL. The presence of the background peaks at sizes other than 248,268, and 375 bp did not affect the quantification of the peaks of interest, and we have not further characterized them. Because the ratio of the products from the deletion amplicons to that of the product of the control sequence could be reliably determined for DNA concentrations >20 mg/L, these ratios were used to check for the presence of one or more of these deletions in unknown whole-blood samples. Preliminary cutoff criteria (0.10.40 for the 375/268-bp ratio) were established for the _aa7 deletion, based on replicate analyses of a presumed -a37 heterozygote. No --SEA/--MED samples were available, so we could not establish preliminary cutoffs for these deletions. Table 3 lists the ratios of the deletion amplicon areas to the control areas for 15 unknown samples whose genotypes had been determined by Southern transfer. Values above the cutoff ranges indicate that no deletion was present; values below the cutoffs indicate the presence of deletions on both chromosomes. The genotypes for the a37 locus of all 15 of the specimens were correctly identified by using a cutoff of 0.1-0.4 for the -a37 heterozygote ratio. For (e.g., -a37a/--SEA or double heterozygotes -MED), the data indicate the deletion of both deletion-sensitive loci because the - -SEA or - -MED deletions affect both the a37 primer binding region as well as the region where the SEA/MED primers bind (9). Only ii of the 15 samples had sufficient DNA for reliable quantitation of the 248/268-bp area ratios for de- 1234 Fig. 2. Electrophoresisof PCR-ampllfiedproductsobtainedfrom 40 mg/L placental DNA: (left) separationon agarose gel; (right) ALF scans after electrophoresison 6% polyacrylamidegel. Primers specific for a37, for SENMED, and for exon 1 of the -globin control regions (Table 1) were used for amplification. Amplification and electrophoretic conditions are described in Materials and Methods. Lane 1, molecular size ladder; lane 2, 248-bp amplified product for detecting the - -SEN- -MED deletion; lane 3, 375-bp amplified product for detecting the -a37 deletion; lane 4, 268-bp amplified product from hemoglobin a-chain locus serving as a control for amplification efficiency and area ratio measurements. 300 400 amplicon size (base pairs) CLINICAL CHEMISTRY, Vol. 40, No. 12, 1994 2263 12345 Fig. 3. (Left) Agarose gel electrophoresis of M-PCR-amplifled products obtained with the use of the deletion-sensitive primers shown in Table 1 and the same amplification and electrophoretic conditions as in Fig. 2 except that final placental DNA concentrations were 5, 10, 20, and 40 mg/L in the amplification reactionmixture; (right) polyacrylamideelectrophoresison ALF sequencer of same samples and as in Fig. 3 (left) except that DNA concentrations were conditions slightlyhigherfor more reliable quantitationsand peak areas were determined by using FragmentManager software. Left: Lane 1,molecular size ladder (100 bp); lane2: 5 maJL; lane 3, 10 mg/L; lane 4.20 mg/L; lane 5,40 mg/L Right: Lane 1,10 mg/L; lane 2.20 mg/L; lane 3,40 mg/L lane 4, 80 mg/L lane 5, molecular size ladder (100 bp). Table 2. Concentration dependence of area ratios for 248-, 268-, and 375-bp peaks. Mean ± SD (n = 8) conc, maJL 248/268 375/268 10 0.79 20 0.68 ± 0.03 1.00 ± 0.03 40 80 0.65 ± 0.07 0.60 ± 0.06 0.87 0.95 tecting --SEA/--MED ± 0.14 deletions. 0.92 This is because ± ± ± 0.27 0.08 0.05 the genotypes were determined by replicate analysis. The within-run mean ± SD for the 248/268-bp and the 375/ 268-bp ratios in normal subjects were 0.200 ± 0.032 and (n = 6). In -a7 heterozy0.790 ± 0.081, respectively gotes, these ratios were 0.201 ± 0.025 and 0.191 ± 0.026, respectively (n = 7). The between-run mean ± SD (5 runs over 8 weeks at 1-2-week intervals) for the 248/268-hp and the 375/268-bp ratios in normal subjects (n = were 0.303 ± 0.067 and 0.879 ± 0.163, respectively 5), and 0.268 ± 0.075 and 0.217 ± 0.059, respectively, in -a37 heterozygotes (n = 5). amplification efficiency of the SEA/MED locus is significantly less for whole-blood DNA than for purified placen- Discussion tal DNA (Fig. 4). Nevertheless, the ratios for the 11 samples were consistent with the genotypes of the samples. Figure 4 shows the actual electrophoretic separations after M-PCR amplifications performed on DNA from individuals confirmed by Southern transfer to have _aal, --SEA, and --MED deletions. As suggested above, the differences in peak heights and areas (compared with those in Fig. 3) probably reflect different amplification efficiencies at the different loci because all of the primers were in the M-PCR reaction at equal concentrations. Peak areas could be made more comparable by changing primer concentrations, but this resulted in slightly poorer precision. Therefore, we kept the primer concentrations equal while establishing cutoffs and evaluating data as shown in Table 3. The within-run and between-run reproducibility of the ratios for normal and heterozygote Southern transfer-confirmed The lack of sensitive and simple definitive tests for a-thalassemia has made the detection of carriers very difficult. As a result, its prevalence in local populations, its effect on other clinical conditions, and accurate data for genetic counseling have been absent or deficient. Therefore, we have developed a DNA-based method that provides definitive data but is simple, sensitive, and reliable enough to be used in a routine clinical laboratory or other diagnostic laboratory. To detect the most common deletions in the a locus in (-a37, --SEA, and --MED), we the US population designed primer pairs that will generate no amplification products or reduced amounts of the normal product when the deletions are present. We have also used primer pairs that amplify a region of the hemoglobin /3 locus as a control to check for amplification efficiency for each PCR reaction mixture. However, this amplified 2264 CLINICAL CHEMISTRY, Vol. 40, No. 12, 1994 Table 3. Areas and area ratios for amplified products. Product size, bp Sample no. 1 -aa/-aa 2 (aa/aa)/(-O. 248 268 QNSb 6096 1313 10133 5879 13 232 - (0.01.O.05)a 0.13 QNS 0.29 2003 504 QNS 6111 0.16 3269 0.04 QNS 816 464 12265 10261 7719 10176 12323 1778 QNS 334 12145 QNS 4 (aa/aa)/(/) 5 aaf- -SEA 6 aaa/aa 3901 ONS 2000 457 5319 - (O.1_O.4)a QNS 5032 1093 9665 3 -a37q/aa 7 (aa1--SEA)/(#{216}) 8 -a31a-a42a 9 aa/aa 10 -a37--SEA 11 -a7a/--MED 375 0.08 0.04 0.03 0.40 0.03 13221 12 -a4’2a/aa 5293 9820 13 (aa/--BRIT)/(-71) 445 13866 1184 14 aa/aa 2322 13561 9624 0.17 15 -a42a1--SEA 494 12655 3222 0.04 a Cutoff range for heterozygotes; higher values indicate no deletion, lower values indicate deletions on both chromosomes. #{176} ONS, quantity notsufficient for reliableanalysis. - 0.50 0.19 0.73 0.25 0.50 0.32 0.23 0.52 - 0.74 0.10 0.71 0.25 if the amplification product ratios are highly reproducible in an analytical gel electrophoresis system. Because the loss of one locus would result in a >50% reduction in the value of the corresponding ratio (Table 3, samples 9 and 14 vs sample 3), the reproducibility represented by within-run SDs of 0.03-0.08 (see Results) is low enough for easy detection of the deletion of one locus. Obviously, if two deletions occur at the same the corresponding band locus (one on each chromosome), will be totally absent (Fig. 4). We used the region surrounding exon 1 of the /3 locus for hemoglobin as a control for amplification efficiency; it is not the site of frequent deletions and has been used in a variety of amplification techniques and applications as a control for amplification efficiency (11, 19). Although we have not detected any deletions that have resulted in the absence of amplification of this region, mutations that might affect amplification efficiency would easily be detected by comparing amplicon ratios with those for the controls included in each run. DNA controls for the - -SEA and - -MED deletions are not included in each run because of limited supply. However, the multiplex format provides an internal control for these deletions because each - -SEA or - -MED deletion also affects the a37 amplicon area. Therefore, changes in the 248/268-bp and 375/268-bp ratios are in the same direction and of similar magnitude. Deletions that might affect the control region only would result in increased (not decreased) ratios for both the a37 and SEA/MED amplicons, compared with the ratios for control samples. As shown in Table 3, the - -BRIT deletion (9), which affects both the a37 and the SEA/MED amplicons in much the same way as the - -SEA and - -MED deletions, can also be detected with this M-PCR procedure. Although the method does not distinguish between -SEA, - -MED, and - -BRIT deletions, these conditions should be easily distinguished by family history, ethnic origin, etc. only 200 amplicon 300 size (base pairs) Fig. 4. Fragment analysis of M-PCR-amplifled products from known deletion mutations obtained with deletion-sensitive primers after polyacrylamidegel electrophoresison ALF. Lane 1,molecularsizeladder(100 bp);lane2, aa/aa; lane 3, -a37a/aa; 4, -a37--a37a; lane 5, -a37a/---SEA; lane 6, -a37a/- -MED. lane product can also be used to quantify the number of deletions occurring at the a loci. By using M-PCR that incorporates deletion-sensitive primer pairs as well as a control region, one can detect single-gene deletions involving the a locus in a single reaction. This is possible CLINICAL CHEMISTRY, Vol. 40, No. 12, 1994 2265 The method we describe is simple enough to be performed in a routine clinical laboratory. The use of M-PCR and automated fragment analysis allows at least 30-40 samples to be processed in one day. Although we describe quantification by using a large polyacrylamide gel and the ALF system, the amplicons could be quantified on any conventional fluorescence densitometer capable of scanning thin agarose or acrylamide gels. Analysis and quantification by automated capillary electrophoresis systems should also be feasible. The small quantities of sample (as little as 10 L of whole blood or 400 ng of purified DNA) make screening of large populations more convenient. Even though a-thalassemias resulting from nondeletion mutations (9,20) cannot be detected by this method, these are very for population rare (9) and would not be good candidates screening efforts. It is our hope that the availability of this rapid and definitive method for detecting a-thalassemia in US populations can lead to a better understanding of the various manifestations of a-thalassemia and how they may affect the clinical presentation of other conditions. the loan of the ALF DNA sequencer by Pharmacia in support of this work. We are grateful to Stephen Thibodeau (Mayo Clinic, Rochester, MN) for providing specimens that had been confirmed by Southern transfer analysis. We also gratefully acknowledge the assistance of William Wilkens and Tracy Roberts in the preparation of some of the figures. We appreciate References 1. Kan YW. Development of DNA analysis for human diseases: sickle cell anemia and thalassemia as a paradigm. J Am Med Assoc 1992;267:1532-6. 2. Baysal E, Huisman THJ. Detection of common deletional a-thalassemia-2 determinants by PCR. Am J Hematol 1994;4fl: 208-13. 3. Liebhaber SA. a-Thalassemia. Hemoglobin 1989;13:685-731. 4. Kazazian HH. The thalassemia syndromes: molecular basis and prenatal diagnosis in 1990. Semin Hematol 1990;27:209-28. 5. Embury SH, Miller JA, Dozy AM, Kan YW, Chan V, Todd D. Two different molecular organizations account for the single a-globin gene of the a-thalaseemia-2 genotype. J Clin Invest 1980;66:1319-25. 2266 CLINICAL CHEMISTRY, Vol. 40, No. 12, 1994 6. Higgs DR The thalassemia syndromes. Q J Med 1993;86:559- 64. 7. Schrier SL, Mohandas N. Globin-chain specificity of oxidationinduced changes in red blood cell membrane properties. Blood 1992;79:1586-92. 8. Dozy AM, Kan YM, Embuiy SH, Mentzer WC, Wang WC. Alpha globin gene organization in blacks precludes the severe form of a-thalassemia. Nature 1979;280:605-7. 9. Higgs DR, Vickers MA, Wilkie AOM, Pretorius JM, Jarman AP, Weatherafl DJ. A review of the molecular genetics of the human a-globin gene cluster. Blood 1989;73:1081-104. 10. Bowden DK, Vickers MA, Higgs DR. A PCR-based strategy to detect the common severe determinants of a-thalassemia. Br J Haematol 1992;81:104-8. 11. Dode C, Rochette J, Krishnamoorthy R. Locus assignment of human a globin mutations by selective amplification and direct sequencing. Br J Haematol 1990;76:275-81. 12. Lebo RV, Saiki RK, Swanson K, Montano MA, Erlich HA, Golbus MS. Prenatal diagnosis of a-thalassemia by polymerase chain reaction and dual restriction enzyme analysis. Hum Genet 1990;85:293-9. 13. Lundeberg J, Wahlberg J, Uhien M. Rapid colorimetric quantification of PCR-amplifled DNA. BioTechniques 1991;10:68-75. 14. Lubin MB, ElashoffjD, Wang SJ, Rotter JI, Toyoda H. Precise gene dosage determination by polymerase chain reaction: theory, methodology 307-17. and statistical approach. Mol Cell Probes 1991;5: 15. Chamberlain JS, Gibbs RA, Ranier JE, Nguyen PN, Caskey CT. Deletion screening of the Duchenne muscular dystrophy locus via multiplex DNA amplification. Nucleic Acids Res 1988;1114156. 16. Ballabio A, Ramier JE, Chamberlain JS, Zollo M, Caskey CT. Screening for steroid sulfatase (STS) gene deletions via multiplex DNA amplification. Hum Genet 1990;84:571-3. 17. Lewm HA, Stewart-Haynes JA. A simple method for DNA 1992; extraction from leukocytes for use in PCR. BioTechniques 13:522-3. iS. Higgs DR, Hill AVS, Bowden DK, Weatherall DJ, Clegg JB. Independent recombination events between duplicated human a-globin genes: implications for their concerted evolution. Nucl Acids Res 1984;12:6965-77. 19. Bauer HM, Ting Y, Greer CE, Chambers JC, Tashiro CJ, Chimera J, et al. Genital human papillomavirus infection in female university students as determined by a PCR-based method. J Am Med Assoc 199 1;265:472-7. 20. Honig GR, Shainsuddin M, Vida LM, Mompoint M, Valcourt E, Bowie U, et al. Hemoglobin Evanston (a 14 Trp - Arg) an unstable a-chain variant expressed as a-thalassemia. J Cliii Invest 1984;73:1740-9.