HUMAN MUTATION 18:157�162 (2001)

© 2001 WILEY-LISS, INC.

METHODS

A New PCR Assay Useful for Screening of FRAXE/FMR2 Mental Impairment Among MalesC.B. Santos, M.A. Costa Lima, and M.M.G. Pimentel*Universidade do Estado do Rio de Janeiro, Rio de Janeiro, Brazil

Communicated by Paolo Fortina

FRAXE (FMR2) is a fragile site associated with mental impairment located in Xq28, 600 kb distalto FRAXA (FMR1), the fragile X syndrome fragile site. The FRAXE mutation is an expansion ofa CCG repeat that results in methylation of a nearby CpG island. FRAXE alleles could be dividedinto four categories: normal (6–30 CCG repeats), intermediate (31–60 CCG repeats), premutation(61–200 CCG repeats), and full mutation (over 200 repeats). We have developed a non-isotopicpolymerase chain reaction (PCR)-based assay for the identification of FRAXE full mutation alle-les among mentally impaired men. In this novel PCR test for the FRAXE locus, we used threeprimers to permit an amplification of a 223 bp monomorphic internal control fragment in additionto the amplification of a 419 bp (CCG)16 FRAXE locus band. A linear series of 93 male patientsreferred for FRAXE testing but found to be negative for the (CCG)n expansion in the FMR2 geneby Southern blotting analysis were retested by our PCR technique. In addition, we analyzed twopositive controls consisting of a FRAXE fully mutated male and one male with a Xq terminaldeletion. The developed PCR test showed accuracy of 100% in the normal individuals retested byPCR analysis, as well as in the two positive control samples utilized, in which the strategy ofmultiplex amplification worked as expected. Although not suitable for medical diagnosis of fe-males and mosaics, it constitutes an important strategy for PCR typing and for FRAXE populationscreening. Hum Mutat 18:157–162, 2001. © 2001 Wiley-Liss, Inc.

KEY WORDS: FRAXE; FMR2; mental retardation, X-linked; triplet repeat; fragile sites; PCR diagnosis

DATABASES:

FMR1 – OMIM: 309550; GDB: 129038; Genbank: L29074; HGMD: FMR1FMR2 – OMIM: 309548; GDB: 141566; Genbank: U48436; HGMD: FMR2http:// www.citi2.fr/cgi-bin/detgen?SYMB=FMR2&DISD=0 (GenAtlas Databases)

INTRODUCTION

As molecular techniques became available forthe fragile X syndrome diagnosis, it was discoveredthat a proportion of the mentally retarded indi-viduals who displayed a folate-sensitive fragile siteat Xq27-q28 in cytogenetic tests did not have anexpansion of the CGG trinucleotide repeat at theFRAXA locus (FMR1; MIM# 309550). Usingfluorescence in situ hybridization (FISH) and finemapping, a new fragile site, designated FRAXE(FMR2; MIM# 309548), was found to be locatedat Xq28 [Sutherland and Baker, 1992; Flynn et al.,1993], 600 Kb distal to the FRAXA locus [Knightet al., 1993]. The FRAXE phenotype is quite vari-able, but tends to include slow learning, hyperac-tivity, and language delay [Flynn et al., 1993;Knight et al., 1993, 1994; Hamel et al., 1994].

The FRAXE mutation belongs to the categoryof dynamic mutations [Richards and Sutherland,1992] and seems to behave very similarly toFRAXA, since the mutational mechanism is rep-resented by an inherited pathologic expansionof unstable CCG triplet repeats located at the 5′

Received 25 January 2001; accepted revised manuscript 12April 2001.

*Correspondence to: Dr. M.M.G. Pimentel, Serviço deGenética Humana, Departamento de Biologia Celular e Genética,Instituto de Biologia Roberto Alcântara Gomes, Universidade doEstado do Rio de Janeiro, Rua São Francisco Xavier 524, PHLC—sala 205, Maracanã, 20550-013, Rio de Janeiro, RJ, Brazil.E-mail: cbs@alternex.com.br

Contract grant sponsor: Coordenação de Aperfeiçoamentode Pessoal de Nível Superior; Contract grant sponsor: Fundaçãode Amparo à Pesquisa do Estado do Rio de Janeiro; Contractgrant number: E-26/170.858/99.

158 SANTOS ET AL.

end of the FMR2 gene, which causes methyla-tion of a CpG island nearby [Gu et al., 1996;Gecz et al., 1996]. The FRAXE (CCG)n micro-satellite, which is naturally polymorphic in thegeneral population, is a pure repeat with nounique interspersed sequences [Zhong et al.,1996]. Normal CCG alleles are 6–30 repeats andare stable on transmission, whereas alleles in theintermediate range (31–60 repeats) may varyslightly between parents and their offspring[Murray et al., 1997]. Alleles ranging from 61to 200 repeats may expand to a full mutation(over 200 repeats) leading to the methylation ofthe promoter region of the FMR2 gene and, con-sequently, causing the mental impairment ob-served in the affected individuals. The thresholdbetween the normal and the premutation alle-les size is still poorly defined, because the FRAXEmutation is understudied and rare. Accordingto Knight et al. [1996], 38 FRAXE families fromdifferent countries have been identified, and fouradditional families have been subsequently re-ported [Biancalana et al., 1996].

The prevalence of FRAXE syndrome was es-timated to be 1/50,000 males, which is 14-foldless than the incidence of FRAXA syndrome[Brown, 1996]. However, Youings et al. [2000]found an estimated prevalence of FRAXE fullmutations of 1/23,423 males. Since the FRAXEphenotype seems to be relatively mild and lacksdysmorphic features, this prevalence could beunderestimated. In view of this fact, it is of greatimportance to undertake routine FRAXE screen-ing. Acknowledgment of FRAXE positive indi-viduals has therefore been made either throughcytogenetic techniques, molecular diagnosis us-ing RFLPs techniques or, more recently, poly-merase chain reaction (PCR) methodology.Nevertheless, conventional cytogenetic analy-sis does not discern FRAXE from others fragilesites nearby it and has been replaced by moreprecise and less time consuming methods of de-tection, such as RFLPs. Although RFLPs analy-sis is an accurate method, making possible theinvestigation of both the presence of CCG ex-pansions and the methylation status, it has thedisadvantage of requiring the slow and labori-ous technique of Southern blotting. Whole, PCRhas became a good tool to perform routinescreening, since it defines the real size of thenormal, intermediate, and premutated CCG al-leles. Few PCR protocols were described to in-

vestigate the FRAXE locus [Knight et al., 1993;Wang et al., 1995; Brown, 1996; Zhong et al.,1996; Strelnikov et al., 1999].

We have developed a non-isotopic PCR-basedtechnique for the identification of FRAXE fullmutation alleles among mentally impaired men.In this novel PCR test for the FRAXE locus, weused three primers to permit an amplification of amonomorphic internal control fragment in addi-tion to the amplification of the FRAXE locus band.The method is useful for identification of affectedmales in population screening studies of theFRAXE mutation and allele sizing of normal, in-termediate, and smaller premutated individuals.

MATERIALS AND METHODS

Patients

A linear series of 93 male patients, referredfor FRAXA and FRAXE testing but found to benegative for the (CCG)n expansion in the FMR2gene by Southern blotting analysis, were retestedby our PCR technique. Blood samples from thepatients were collected, after parental consent,in vacuntainer tubes containing EDTA, andDNA was extracted according to standard meth-ods [Miller et al., 1988]. Alternatively, DNAsamples were obtained by using the GFX Ge-nomic Kit (Gibco BRL, USA), according tomanufacturer’s instructions.

Southern Blotting Analysis

For the Southern blotting analysis, EcoRI/EagIdouble digests were performed on 10µg of ge-nomic DNA, electrophoresed in 0.8% agarosegels overnight at 35V and blotted onto HybondN membranes (Amersham Pharmacia Biotech,UK). The intragenic OxE18 probe was radiola-beled with [32P] dCTP by a random priming kit(Amersham Pharmacia Biotech, UK). Overnighthybridization was carried out at 42°C, filters werewashed and exposed to Kodak X-OMAT film at–70°C for 7 d.

PCR Analysis

For PCR analysis, primers XE1 (5′ GGCGG-GCGGCCCAGCCCGCCTGAGCC 3′) andXE2 (5′ GCCCTCCCGCCCAGCTAAAAGT-GTCCGGG 3′) were designed using the soft-ware OMIGA 1.1.3 (Table 1) to have effectivepriming temperatures similar to primer 603[Knight et al., 1993].

A total of 150 ng of genomic DNA was am-

FRAXE/FMR2 MENTAL RETARDATION DIAGNOSIS 159

plified by the PCR technique in a total volume of25 µl in 0.2 ml polypropylene microtubes. All thereactions were performed in a PTC-100 (MJ Re-search, USA) thermocycler. Final concentrationsof reagents in each reaction were: 50 mM KCl;10 mM Tris-HCl; 1.6 mM MgCl2; 10% dimeth-ylsulfoxide (DMSO); 200 µM of each dATP,dTTP, dCTP; 50 µM of dGTP; 150 µM 7-deaza-dGTP (Amersham Pharmacia Biotech, UK); 0.2µM primer XE1; 0.6 µM primer XE2; and 0.4µM primer 603 [Knight et al., 1996]. The spe-cific PCR profile was carried out immediatelyand consisted of 10 min of “hot starting” at 95°C,5–10 min at 72°C for addition of 2 U of Taq DNApolymerase (Amersham Pharmacia Biotech, UK),followed by 30 cycles of 1.5 min of denaturation at95°C, 1 min of annealing at 65°C, and 2 min ofextension at 72°C, with a final extension step at72°C for 10 min. At the end of the program, thesamples were held at 4°C in the thermal cycleruntil removed for further analysis. Then, 7 µl ofthe amplification products were mixed with 2 µlof loading buffer (5X TBE, 25% glycerol, 1% SDS,0.1% bromophenol blue, 0.1% xylene cyanol, and100mM EDTA) and electrophoresed in 6% non-denaturing polyacrylamide gels. The gel was silverstained and the band patterns were analyzed incomparison with a 1 Kb ladder molecular sizemarker (Gibco BRL, USA).

Control Patients

To test the efficiency of the developed PCRtest, we analyzed a positive control consistingof a DNA sample from a FRAXE fully mutatedmale harboring 315 CCG methylated repeats(individual II.8/family C) [Biancalana et al.,1996]. As another positive control of our ex-periment, we used a DNA from a male with aXq terminal deletion directed to our service.

RESULTS AND DISCUSSION

This novel PCR assay was developed basedon the idea that FMR2 mutations have long

CCG-expanded alleles that fail to amplify un-der certain experimental conditions. In this way,we designed two primers (XE1 and XE2) to am-plify together with a third already reportedprimer (603) described by Knight et al. [1993].The primer XE2 amplifies together with theprimer 603, a 419 bp polymorphic fragment fora (CCG)16 allele with 81% CG content. For aninternal control for the PCR reaction we usedthe primer XE1 that amplifies with the primerXE2, a monomorphic band of 223 bp (77% CGcontent), out of the CCG repeats region (Fig.1). The fact that the primer XE2 amplifies si-multaneously with two others provides a com-petition among the upper primers for completionof the desired product. Since it has smaller size,

TABLE 1. Properties of Primers XE1 and XE2 Designedfor FRAXE Syndrome Screening

Annealing Effective Effective primingPrimer sitea length (Ln)

b temperature (Tp)b

XE1 151–176 49 93XE2 345–373 48 93aAnnealing site within the FMR2 gene sequence fromGenebank database (U-48436).bAccording to Wu et al. [1991].

FIGURE 1. Scheme of the developed strategy for PCRscreening of FRAXE full mutation alleles among mentallyimpaired men. The box indicates the (CCG)n region withinthe physical region of the FMR2 gene. The primers usedfor the PCR are shown as arrows pointing from the 5′ tothe 3′ direction. Primers 603 [Knight et al., 1993] andXE2 amplify a DNA segment that includes the polymor-phic (CCG)n repeat. Primers XE1 and XE2 amplify a mono-morphic fragment (223 bp) that constitutes the internalcontrol. PCR products for normal individuals, normaltransmitting males (premutation), and affected individu-als are shown. Note the absence of the polymorphic locusband in the affected individuals.

160 SANTOS ET AL.

the internal fragment should be preferentiallyamplified. For this reason, extensive experimentswere needed to obtain appropriate primers con-centration that could yield a reproducible andstrong signal for the monomorphic internal con-trol without losing intensity of the polymorphicproducts.

Once the target fragments we were amplify-ing had a high percentage of CG content, ap-propriate conditions were employed to allowreproducible amplification. The use of co-sol-vents (like dimethyl sulfoxide) and high Tmprimers improved both product yield and speci-ficity of PCR. Amplifications performed withdGTP or 7-deaza dGTP only were less efficientthan reactions containing a mixture of 7-deazadGTP and dGTP in a ratio of 3:1. Pre-amplifi-cation heating (“hot starting”) promoted strin-gent primer annealing and minimized thepresence of nonspecific bands.

The use of 0.2 ml thin-walled tubes and totalreaction volumes as low as 25 µl enhances heattransfer and reduces thermal lag.

The silver stain, employed to visualize DNA(Fig. 2), offers advantages of sensitivity overethidium bromide and was routinely utilized forassessment of the microsatellites allelic bands.

For the optimization of the PCR technique

developed, we initially performed a test involv-ing the DNA of 93 male probands, which werepreviously identified as normal subjects by theSouthern blotting technique. All of them werecorrectly assigned by PCR, in which both themonomorphic 223 bp internal control and thepolymorphic CCG band were presented, show-ing an accuracy of 100%.

To prove the efficiency of the developed PCRtest, we analyzed two positive controls consist-ing of DNA samples of a FRAXE fully mutatedman and of a male with a Xq terminal deletion.In the latter, as expected, there was no amplifi-cation of both fragments (internal control bandand FRAXE locus band). In the other DNAsample, only the monomorphic 223 bp internalcontrol could be identified, owing to completefailure of the amplification of the diagnostic poly-morphic fragment, as predicted.

Few PCR protocols have been described toinvestigate the FRAXE locus [e.g., Knight et al.,1993; Wang et al., 1995; Brown, 1996; Zhong etal., 1996; Strelnikov et al., 1999]. Generally,those protocols identify affected men by the ab-sence of amplification of the full mutation frag-ment. However, the failure of amplification couldbe due to others factors like poor quality DNA,inefficiency of the polymerase to read throughthe repeats, problems with one or more reagents,and non-uniform temperatures across the block.In those cases, it is impossible to know if theamplification failed because the individual is re-ally FRAXE positive or there was another arti-fact acting in the absence of the amplifiedproduct. We resolved this ambiguous questionby developing a novel multiplex PCR techniquein which a third primer provides a monomor-phic control fragment that must be present inall the amplicons if the reaction really happened.An individual will be considered as positive whenonly the control band is present and, in thosecases, the Southern blotting technique is neededto investigate the methylation pattern. If bothbands are absent, it is clear that the PCR reac-tion failed due to an external factor. The samestrategy was used by Haddad et al. [1996] forthe FRAXA locus and, currently, is widely em-ployed for Fragile X syndrome screeningamong males.

In females, the amplification of only a singleallele referred to the FRAXE locus band in ad-dition to the control band will mostly be due to

FIGURE 2. Multiplex PCR amplification test for the FRAXEregion using oligonucleotides 603, XE1, and XE2. Lane1: size standards from a 1 Kb ladder (Gibco BRL, USA)with molecular sizes, from the bottom up, of 201, 220,298, 344, 396, 506, 517, and 1018 bp; Lane 2: negativecontrol (PCR reaction containing water in the place ofDNA template); Lane 3: normal control (note the pres-ence of both bands); Lane 4: positive control of a FRAXEaffected male (individual II.8/family C) [Biancalana et al.,1996] (note the presence of only the internal control band);Lane 5: patient with a Xq terminal deletion (note the ab-sence of both bands); Lanes 6 to 8: normal patients (notethe presence of both bands). The upper band has differ-ent sizes due to the CCG polymorphism.

FRAXE/FMR2 MENTAL RETARDATION DIAGNOSIS 161

homozygosity for allele size. However, the PCRanalysis cannot distinguish a homozygote femalefrom a fully mutated one, since in both cases wewill be able to detect only one amplification frag-ment for the FRAXE locus in addition to thecontrol band. It happens due to a preferentiallyamplification of the smaller size FRAXE allele,allowing a false-negative result in affected fe-males. For the same reason, the PCR exclusionapproach will not differentiate between normaland mosaic (males or females) where there arecells with normal repeat copy numbers. How-ever, mosaic patients are apparently rare andmight not be a significant cause of error in popu-lation screenings.

Another problem common to the PCR de-tection methods, especially with trinucleotiderepeats, is the presence of additional (“stutter”)bands beside the microsatellite band, differingby few base pairs. The major mechanism postu-lated for this is “slipped strand mispairing” (akinto slippage synthesis mechanisms for micro-satellite size expansion and sequence evolution),in which the presence of multiple repeats per-mit slippage of the copied strand on the tem-plate, producing fragments with few-nucleotidespacing.

However, the use of PCR to analyze the CCGrepeat expansions has many advantages: 1) al-lows many samples to be screened at low cost;2) avoids the use of radioisotopes; 3) times con-servatively quoted for a result from Southernblotting analysis could be two–three weeks com-pared with two days for PCR; 4) small amountsof DNA are needed and screening studies couldbe performed through other less invasive sourcesof DNA, such as buccal swabs; 5) PCR analysiscan be helpful to distinguish alleles within thenormal polymorphic range from those at thelower end of the premutation range.

Although the FRAXE mutation is consideredrare, some recent studies [Youings et al., 2000]reported a FRAXE prevalence higher than theestimated prevalence of 1/50,000 males previ-ous established by Brown [1996]. It is postulatedthat the FRAXE mutation could be under-diagnosed, given that its associated phenotypeis, in some cases, so mild that positive individu-als would not normally be tested. On these bases,it is necessary to perform routine screening ofFRAXE mental retardation in patients referredfor Fragile X syndrome diagnosis to define the

real incidence of the FMR2 mutation. InFRAXA/FRAXE screening programs, in whichthe samples investigated are numerous, the useof Southern blotting analysis is impracticable dueto the high cost and slowness. So, the developedPCR test could be an important and useful toolin this kind of investigation. Also, it is question-able whether the FRAXE intermediate and/orpremutated alleles, like the FRAXA premutatedalleles, are associated with a cognitive disabilityphenotype [Murray et al., 1996; Crawford et al.,1999]. Nevertheless, the investigation of thenumber of CCG repeats in normal, intermedi-ate, or premutated alleles only could be per-formed through PCR analysis. Southern blottingtechniques discriminate only larger premutatedand fully mutated alleles.

In conclusion, although the PCR test we de-veloped is not suitable for the medical diagnosisof females and mosaics, it constitutes an impor-tant tool for FRAXE screening in a populationcomprised mainly of special education-needsmales.

ACKNOWLEDGMENTS

We thank Dr. Valérie Biancalana (Institutede Génétique et de Biologie Moleculaire,INSERM, Strasbourg, France) for kindly provid-ing the DNA sample of the FRAXE positive in-dividual and Dr. K.E. Davies (Department ofBiochemistry, University of Oxford) for the re-lease of the OxE18 probe.

REFERENCES

Biancalana V, Taine L, Bouix JC, Finck S, Chauvin A, DeVerneuil H, Knight SJL, Stoll C. 1996. Expansion andmethylation status at FRAXE can be detected on EcoRIblots used for FRAXA diagnosis: analysis of four FRAXEfamilies with mild mental retardation in males. Am J HumGenet 59:847–854.

Brown WT. 1996. The FRAXE syndrome: is it time for rou-tine screening? Am J Hum Genet 58:903–905.

Crawford DC, Meadows KL, Newman JL, Taft LF, Pettay DL,Gold LB, Hersey SJ, Hinkle EF, Stanfield ML, HolmgreenP, Yeargin-Allsopp M, Boyle C, Sherman SL. 1999. Preva-lence and phenotype consequence of FRAXA and FRAXEalleles in a large, ethnically diverse, special education-needspopulation. Am J Hum Genet 64:495–507.

Flynn GA, Hirst MC, Knight SJL, Macpherson JN, BarberJC, Flannery AV, Davies KE, Buckle VJ. 1993. Identifica-tion of the FRAXE fragile site in two families ascertainedfor X linked mental retardation. J Med Genet 30:97–100.

Gecz J, Gedeon AK, Sutherland GR, Mulley JC. 1996. Clon-

162 SANTOS ET AL.

ing of FMR2: a gene associated with FRAXE mental re-tardation. Nat Genet 13:105–108.

Gu Y, Shen Y, Gibbs RA, Nelson DL. 1996. Identification ofFMR2, a novel gene associated with the FRAXE CCG re-peat and CpG island. Nat Genet 13:109–113.

Haddad LA, Mingroni-Netto RC, Vianna-Morgante AM,Pena SDJ. 1996. A PCR-based test suitable for screeningfor fragile X syndrome among mentally retarded males.Hum Genet 97:808–812.

Hamel BC, Smits AP, Graaf ED, Smeets DF, Schoute F, EussenBHJ, Knight SJ, Davies KE, Assman-Hulsmans CF, OostraBA. 1994. Segregation of FRAXE in a large family: clini-cal, psychometric, cytogenetic and molecular data. Am JHum Genet 55:923–931.

Knight SJL, Flannery AV, Hirst MC, Campbell L, Christo-doulou Z, Phelps SR, Pointon J, Middleton-Price HR,Barnicoat A, Pembrey ME, Holland J, Oostra BA, BobrowM, Davies KE. 1993. Trinucleotide repeat amplificationand hypermethylation of a CpG island in FRAXE mentalretardation. Cell 74:127–134.

Knight SJL, Voelckel MA, Hirst MC, Flannery AV, MonciaA, Davies KE. 1994. Triplet repeat expansion at theFRAXE locus and X-linked mild mental handicap. Am JHum Genet 55:81–86.

Knight SJL, Ritchie RJ, Chakrabarti L, Cross G, Taylor GR,Mueller RF, Hurst J, Paterson J, Yates JR, Dow DJ, DaviesKE. 1996. A study of FRAXE in mentally retarded indi-viduals referred for fragile X syndrome (FRAXA) testingin the United Kingdom. Am J Med Genet 58:906–913.

Miller AS, Dykes DD, Polesky HF. 1988. A simple salting outprocedure for extracting DNA from human nucleated cells.Nucleic Acids Res 16:1215.

Murray A, Youings S, Dennis N, Latsky L, Linehan P,

McKechnie N, Macpherson J, Pound M, Jacobs P. 1996.Population screening at the FRAXA and FRAXE loci:molecular analyses of boys with learning difficulties andtheir mothers. Hum Mol Genet 5:727–735.

Murray A, Macpherson JN, Pound MC, Sharrock A, YouingsSA, Dennis NR, McKechnie N, Linehan P, Morton NE,Jacobs PA. 1997. The role of size, sequence and haplotypein the stability of FRAXA and FRAXE alleles during trans-mission. Hum Mol Genet 5:727–735.

Richards RI, Sutherland GR. 1992. Fragile X syndrome: themolecular picture comes into focus. TIG 8:249–255.

Strelnikov V, Nemtsova M, Chesnokova G, Kuleshov N,Zaletayev D. 1999. A simple multiplex FRAXA, FRAXEand FRAXF PCR assay convenient for wide screening pro-grams. Hum Mutat 13:166–169.

Sutherland GR, Baker E. 1992. Characterization of a new rarefragile site easily confused with the fragile X. Hum MolGenet 1:111–113.

Wang Q, Green E, Bobrow M, Mathew CG. 1995. A rapid,non-radioactive screening test for fragile X mutations atthe FRAXA and FRAXE loci. J Med Genet 32:170–173.

Wu DY, Ugozzoli L, Pal BK, Qian J. 1991. The effect of tem-perature and oligonucleotide primer length on the speci-ficity and efficacy of amplification by the polymerase chainreaction. DNA Cell Biol 10:233–238.

Youings SA, Murray A, Dennis N, Ennis S, Lewis C,McKechnie N, Pound M, Sharrock A, Jacobs P. 2000.FRAXA and FRAXE: the results of a five year survey. JMed Genet 37:415–421.

Zhong N, Ju W, Curley D, Wang D, Pietrofesa J, Wu G, ShenY, Pang C, Poon P, Liu X, Gou S, Kajanoja E, Ryynanen M,Dobkin C, Brown WT. 1996. A survey of FRAXE allelesizes in three populations. Am J Med Genet 64:415–419.