GENETIC TESTINGVolume 6, Number 1, 2002© Mary Ann Liebert, Inc.

Detection of the GRT Polymorphism at the Sp1 Binding Site of the Collagen Type Ia1 Gene by a Novel

Arms-PCR Method

LUCIO MONTANARO1,2 and CARLA RENATA ARCIOLA1

ABSTRACT

The G R T mutation at base 1 of intron 1 at the binding site of the Sp1 transcription factor of the collagentype Ia1 gene (COLIA1, GenBank accession no. AF017178) is a putative marker for low bone mineral den-sity and osteoporotic fractures. A new method for the detection of this mutation is presented, based on theamplification refractory mutation system–polymerase chain reaction (ARMS-PCR), which utilizes two sepa-rate and simultaneous PCRs to detect the normal and mutated alleles. The forward primer (positions1307–1336 of the gene) is common to both amplifications. Two reverse primers (positions 1566–1546) are used,differing in the 39 base (39-C for the normal S allele and 39-A for the mutated s allele). The former amplifi-cation uses the reverse primer specific for the S allele; the latter uses the reverse primer with the 39-base com-plementary to the mutated base of the s allele. In the SS condition, amplification occurs only in the formerreaction and in the ss condition only in the latter. Both reactions give a product in the Ss condition. DirectDNA sequencing of a COLIA1 region containing the G R T polymorphism demonstrates the validity of thisARMS-PCR method. The new method is more reliable than a previously published detection method, whichutilizes a mismatched reverse primer, introducing a restriction site in the T-substituted (s) allele. However,the restriction enzyme is costly, its digestion time long, and incomplete digestion can lead to an under-estimation of the frequency of ss homozygosity. The latter can result in incorrect conclusions about a linkagebetween osteoporosis and the COLIA1 polymorphism. In a survey of the COLIA1 polymorphism in 133 os-teoporotic subjects with femur fractures, 7 cases of ss homozygosity were consistently detected both by directDNA sequencing and by the ARMS-PCR method. This, in contradistinction to the mismatched-primer methodby which 3 of these 7 cases were inaccurately diagnosed as Ss heterozygosites.

53

INTRODUCTION

SENILE OSTEOPOROSIS is a disabling pathological condition of-ten accompanied by vertebral or femoral fractures, which

especially affects women older than 50 years. Several studieshave indicated the contribution of a genetic component to thispathology, and among the candidate genes, may be a polymor-phism of the collagen type Ia1 gene (COLIA1; gene map lo-cus: 17q21.31–q22.05, GenBank accession no. AF017178)(McGuigan et al., 2000). COLIA1 encodes the a-chain of typeI collagen, which is present in bone matrix. Mutations in thecoding sequences of this gene are known to be related to os-teogenesis imperfecta (Pope et al., 1985). The G R T mutationin the first intron of the COLIA1 gene at base 1546 corresponds

to the first base of the binding site of the Sp1 transcription fac-tor and has been associated with a low bone mineral densityand increased incidence of osteoporotic fractures. Studies havebeen reported on 299 English women (Grant et al., 1996), on180 Danish (Langdahl et al., 1998), and 341 German(McGuigon et al., 2001) osteoporotic subjects of both sexes,and on 1778 women of the Netherlands (Uitterlinden et al.,1997). On the other hand, studies on 64 Swedish women (Lidénet al., 1998), on 135 Belgian women (Aerssens et al., 2000),and on 133 Danish women (Heegard et al., 2000) did not sup-port this association. The allele of the COLIA1 gene bearingthe G R T point mutation, described in association with lowbone mineral density, has been called the s allele to distinguishit from the normal S allele (Grant et al., 1996). The s allele

1Research Laboratory on Biocompatibility of Implant Materials, Rizzoli Orthopedic Institute, Bologna, Italy.2Experimental Pathology Department, University of Bologna, Italy.

present in homozygosity may represent a marker of predispo-sition to osteoporosis.

The method originally proposed to detect this polymorphismconsists of PCR amplification of the region containing the G RT mutation using a mismatched reverse primer, which intro-duces a diallelic restriction region (Grant et al., 1996). This isfollowed by overnight digestion of the amplified product withthe restriction enzyme Bal I, which cleaves the amplified prod-uct only in the presence of the mutation. The PCR product con-taining the T-substituted allele (s allele) is cleaved by Bal I,giving a 244-bp band in comparison to the uncleaved 261-bpband of the normal S allele. In the heterozygous state (Ss), bothbands are present. However, this method presents some draw-backs, such as the long time needed for enzymatic digestionand the high cost of the restriction enzyme itself, which is par-ticularly significant when a large number of samples have tobe processed, as is the case in population linkage studies. A riskalso exists of incomplete digestion, giving rise to two bands inthe case of an ss homozygote, with consequent misidentifica-tion as a heterozygote.

Therefore, an alternative method was devised, based on theamplification refractory mutation system–polymerase chain re-action (ARMS-PCR), consisting of two separate and simulta-neous PCRs, which make possible the detection of the normaland mutated alleles. This new method, which does not requirethe introduction of a restriction site or enzymatic digestion,proved to be sensitive, rapid, and relatively low in cost. Herein,the authors report on their experience with ARMS-PCR for COLIA1 genotype detection alongside the traditional mis-matched-primer PCR method in a sample population of 133 el-derly patients with bone fractures associated with osteoporosis.

MATERIALS AND METHODS

Patients

A total of patients over 65 years of age with fractures of thefemoral neck associated with osteoporosis were selected fromthe Surgical Divisions of the Rizzoli Orthopedic Institute. Onlypatients who suffered the fractures by falling when standing orwalking were admitted to the study. Subjects with a fractureoriginating from severe trauma were excluded. A 5-ml sampleof blood was collected from each patient in an EDTA-treatedtube.

DNA extraction from whole blood

DNA was extracted following the “salting out” method,which uses high saline concentration to precipitate all the pro-teins. For the extraction, the following procedure was adopted.

To the 5-ml blood sample, 5 ml of TKM-I (10 mM Tris-HCl,pH 7.6, buffer, containing 10 mM MgCl2, 10 mM KCl, 2 mMEDTA) was added, together with 125 ml of Nonidet P-40 in a15-ml tube, with repeated stirring to keep the Nonidet in sus-pension. After centrifugation at 2,200 rpm for 10 min, the su-pernatant was discarded and the pellet, mostly comprised of nu-clei, was resuspended in 5 ml of TKM-I, with strong stirringusing a vortex mixer. A second centrifugation at 2,200 rpm for10 min followed.

After discarding the supernatant, the precipitate was sus-pended in 0.8 ml of TKM-II (10 mM Tris-HCl, pH 7.6, buffer,containing 10 mM MgCl2, 10 mM KCl, 2 mM EDTA, 0.4 MNaCl). The suspension was transferred to a 1.8-ml Eppendorftube, and 50 ml of 10% sodium dodecyl sulfate (SDS) was addedby repeated pipetting to disperse the larger aggregates. After a10-min incubation at 55°C, 0.3 ml of 6 M NaCl was added. Theprecipitated proteins were removed by centrifugation at 13,000rpm for 5 min. The supernatant was transferred to a 15-ml tube,and 2 volumes of absolute ethanol were added. The precipitatedDNA was collected by centrifugation at 13,000 rpm for 5 min,resuspended in 1 ml of 70% ethanol, and transferred to a 1.8-ml Eppendorf tube. After a second centrifugation at 13,000 rpmfor 5 min, the DNA pellet was dried in air and dissolved inH2O. DNA concentration, determined spectrophotometrically,was brought to a final concentration of 150 ng/ml.

Mismatched primer method for detection of COLIA1G R T polymorphism

The mismatched primer method was performed in all 133 cases,

MONTANARO AND ARCIOLA54

FIG. 1. Detection of G R T polymorphism in the COLIA1gene by ARMS-PCR assay. The PCR product is obtained ei-ther only with the reverse 39-G primer when the subject is SShomozygous, or with the reverse 39-T primer when the subjectis ss homozygous, or with both when the subject is Ss het-erozygous. The 346-bp band is the product of an additional pairof primers picked in a conserved region between nucleotides16,930 and 17,275 of the COLIA1 gene, used as a positive con-trol for the PCR.

TABLE 1. COLIA1 GENOTYPE DISTRIBUTION AMONG 133 PATIENTS OVER 65 YEARS OLD WITH FRACTURES ASSOCIATED WITH

OSTEOPOROSIS SCREENED BY THE MISMATCHED PRIMER METHOD, BY ARMS-PCR METHOD, AND BY DIRECT DNA SEQUENCING

COLIA1 genotype

SS Ss ss Total

Mismatched primer PCR method 89 (66.9%) 40 (30.1%) 4 (3.0%) 133 (100%)ARMS-PCR method 89 (66.9%) 37 (27.8%) 7 (5.3%) 133 (100%)Direct DNA sequencing 89 (66.9%) 37 (27.8%) 7 (5.3%) 133 (100%)

FIG

. 2.

Aut

omat

ed s

eque

nce

anal

ysis

of

the

poly

mor

phis

ms

of C

OL

IA1 .

The

rel

evan

t pa

rt o

f th

e se

quen

ce c

onta

inin

g th

e ba

se 1

546

is s

how

n fo

r th

ree

subj

ects

fou

nd S

s , s

s , a

nd S

Sby

mea

ns o

f th

e A

RM

S-P

CR

met

hod.

At

the

expe

cted

pos

itio

ns,

T i

s pr

esen

t to

geth

er w

ith

the

norm

al G

in

the

Ss s

ubje

ct(g

raph

1,

top )

onl

y T

is

pres

ent

in s

ssu

bjec

t (g

raph

2,

mid

dle )

and

, on

ly G

is

pres

ent

in t

he S

Ssu

bjec

t (g

raph

3,

bott

om).

In

each

gra

ph,

the

arro

w i

ndi-

cate

s th

e 15

46 p

osit

ion.

55

as described elsewhere (Grant et al., 1996). In brief, the mis-matched primer method uses the following primer sequences:59-TAACTTCTGG ACTATTTGCG GACTTTTTGG-39 (for-ward) and 59-GTCCAGCCCTCATCCT GGCC-39 (reverse; themismatched bases are underlined).

After PCR, the reaction product is digested overnight withBal I (5 U/mg of DNA) and analyzed by agarose gel elec-trophoresis. Because Bal I specifically recognizes the sequenceTGGQCCA, only the T-substituted allele is cleaved, givingrise to a 244-bp band instead of the 261-bp band of the G al-lele. Therefore, only one band will be present in the two ho-mozygous conditions; i.e., 261-bp in the case of SS or 244-bpin that of ss. In the heterozygous condition, both bands will bepresent.

ARMS-PCR method for detection of COLIA1 G R T polymorphism

The ARMS-PCR method consists of two PCRs in one ofwhich the 39 terminus of the reverse primer is specific for thenormal allele and in the other this 39 terminus is specific forthe mutated allele. The sequence of the forward primer (posi-tions 1307–1336 of the COLIA1 gene), which is common toboth amplifications, is 59-TAACTTCTGGACTATTTGCG-GACTTTTTGG-39. The sequences of the reverse primers (po-sitions 1566–1546) are 59-GTCCAGCCCTCATCCCGCCCC -39 (henceforth referred to as reverse 39-G primer) for the normalS allele, and 59-GTCCAGCCCTCATCCCGCCCA -39 (reverse39-T primer) for the mutated s allele.

In all 133 cases, together with the amplification of the in-vestigated polymorphic region of COLIA1, a 346-bp conservedregion of the same gene was co-amplified with an additional pairof primers as a positive control of the amplification. The follow-ing primers were used: 59-GGACCTGGGGTTCTCAGACT-3 9

(forward, corresponding to nucleotides 16,930–16,949 of theCOLIA1 gene) and 59-GGTATAAAA GGGAGCCG-39 (re-verse, corresponding to nucleotides 17,256–17,275).

All amplification reactions were carried out in the UNO IIThermocycler BIOMETRA in a volume of 50 ml containing 1unit of Taq DNA polymerase (Genenco), 200–400 ng of ex-tracted DNA, 10 mM PCR nucleotide mix (BoehringerMannheim), reaction buffer (10 mM Tris-HCl, pH 9.0, 50 mMKCl, 0.1% Triton X-100), and 2.5 mM MgCl2. The thermal stepprogram used comprised 40 cycles of 95°C for 1 min (denatu-ration), 60.5°C for 1 min (annealing), and 72°C for 1 min (extension). Following PCR, 15 ml of product was examinedby electrophoresis on a 3% agarose gel. The Marker VI(Boehringer Mannheim) was used as a molecular weight refer-ence. When a 261-bp band is obtained in both reactions, thesubject from whom DNA was extracted is heterozygous (Ss).When the band is obtained in only one of the two PCRs, thesubject is homozygous. The homozygosity is SS if the band isobtained with the reverse primer with a C residue in the 39 po-sition, and ss if the A residue is present in the 39 position ofthe reverse primer.

DNA sequencing

To validate the results obtained by the ARMS-PCR method,direct sequencing of the region containing the mutation was car-ried out in all 133 cases. Two primers were picked for the PCR-amplification of a 300-bp region between the positions 1310

and 1609 of the COLIA1 gene. This region includes, in an in-ternal position, the G R T polymorphism, which occurs at thebase 1546. The following primers were used: 59-CTTCTG-GACTATTTGCGGACTT-3 9 (forward, corresponding to nu-cleotides 1310–1331 of COLIA1 gene) and 59-GCGTGGTA-GAGACAGGAGGA-39 (reverse, corresponding to nucleotides1590–1609).

The amplification reaction was carried as described aboveexcept for the annealing temperature, which was 57°C. Fol-lowing PCR amplification, 15 ml of product was examined byelectrophoresis on 3% agarose gel. The remaining PCR prod-uct was purified by QIAquick purification kit (Qiagen). dRho-damine terminator cycle sequencing using AmpliTaq DNApolymerase FS (Perkin Elmer) was carried out in accordancewith the manufacturer’s protocol. The products of sequencingreactions were analyzed on an ABI-377 automated DNA se-quencer (Perkin Elmer).

RESULTS

Figure 1 shows the detection of the G R T polymorphismby the ARMS-PCR method; in the case of SS homozygosity aPCR product is obtained only with the reverse 39-G primer,whereas in the case of ss homozygosity the product is obtainedonly with the reverse 39-T primer. In the case of Ss heterozy-gosity, a PCR product is obtained in both reactions. The 346-bp band is produced as a positive control for successful PCRin all reactions. Table 1 shows the results of the screening forthe COLIA1 polymorphism carried out on 133 patients with os-teoporotic fractures with both the mismatched primer method(Grant et al., 1996) and with the ARMS-PCR method describedin the present paper. Also, the results of direct DNA sequenc-ing of all 133 cases are reported in Table 1.

At the expected positions, only G was present in subjectsclassified as SS by the ARMS-PCR method, T was present to-gether with the normal G in Ss subjects, and only T was pres-ent in ss subjects. The two PCR methods gave different resultsin three cases. Three subjects who appeared to be Ss by the mis-matched primer method turned out to be ss by the ARMS-PCRmethod. The results of the ARMS-PCR method were alwaysconsistent with the direct DNA sequencing of the investigatedgene region. Thus, the two PCR methods gave rise to an ap-parent discrepancy in genotype distribution, and particularly inthe frequency of ss homozygosity in the studied population(5.3% by ARMS-PCR vs. 3.0% by mismatched primer PCR),a frequency that is of importance in assessing a linkage betweenthe ss genotype and a risk of osteoporosis. The consistency ofthe results of the ARMS-PCR method with those of direct DNAsequencing validates as reliable the proposed new method. InFig. 2 the automated sequence analyses of three representativegenotypes (Ss, Ss, ss) are shown.

DISCUSSION

The presence of the T polymorphism at the Sp1 binding siteof the COLIA1 gene has recently been proposed as an impor-tant marker of low bone mineral density, with the consequentrisk of bone fractures (Garnero et al., 1998; Grant et al., 1996;

MONTANARO AND ARCIOLA56

McGuigan et al., 2000). However, the reported studies of thisputative genetic association have yielded contradictory results(Lidén et al., 1998; Heegaard et al., 2000). Therefore, exten-sive studies are still required on large patient populations to de-termine the real relevance of this proposed marker. This paperpresents an alternative method for the detection of ss homozy-gosity, which could offer considerable advantages in terms ofcosts and reliability.

The PCR method, based on a mismatched reverse primer thatintroduces a restriction site for the enzyme Bal I in the pres-ence of the T mutation, is costly and time consuming, espe-cially when large numbers of samples have to be handled. Inthe proposed ARMS-PCR method, the use of the expensive re-striction enzyme is avoided and the time required to obtain re-sults is considerably shorter. These factors could represent aconsiderable saving in both supplies and labor for screeningstudies on large populations. However, the most important ad-vantage of the proposed ARMS-PCR detection method, may beits greater reliability.

From our experience, a pitfall of the original mismatchedprimers method for the detection of the ss polymorphism is thatincomplete enzymatic digestion can make rise to true ss ho-mozygotes appear to be Ss heterozygotes. This problem is typ-ical of methods involving the use of restriction enzymes. Thus,the ARMS-PCR method has the dual advantage that it does notrequire an enzymatic digestion step and that it possesses an in-ternal positive control to confirm the amplification process. Inthe present series, three subjects were revealed to be ss ho-mozygotes only by ARMS-PCR, leading to a discrepancy inthe genotype distribution, as determined by the two detectionmethods. Direct DNA sequencing of all cases confirmed the re-sults of the ARMS-PCR method. The consistency between theresults of direct sequencing and those of the ARMS-PCRmethod allow us to conclude that the proposed method is a re-liable tool for the screening of the G R T polymorphism in theCOLIA1 gene.

Since the frequency of the T polymorphism at the Sp1 bind-ing site of the COLIA1 gene is low and only small differenceshave until now been observed between normal and osteoporoticpopulations, the reliability of the detection of ss alleles is cru-cial to assess any linkage between this polymorphism and thetendency to osteoporosis. In this context, even occasional dis-crepancies in the detection of true ss homozygotes could exertmajor effects on the determination of the allelic distribution insample populations. Indeed, even in our limited sample of 133subjects, the determination by ARMS-PCR of three additionalcases of ss homozygosity raised the frequency of such casesfrom 3.0% to 5.3%. Such discrepancies could strongly affectthe statistical significance of large-scale studies. For this rea-son, and not only for economic considerations, we suggest theadoption of the ARMS-PCR method in large-scale studies at-tempting to determine whether there is linkage between the COLIA1 polymorphism and osteoporosis.

ACKNOWLEDGMENTS

We thank Dr. Barbara Cardazzo, Institute of Clinical Medi-cine, University of Padua, Italy, and Stefania Collamati of ourLaboratory for technical help. We are also grateful to Dr.Gualtiero Gualtieri, Head of the IV Orthopedic Unit of the

Rizzoli Orthopedic Institute, for the recruiting of patiens.Thanks to Mr. Robin M.T. Cooke and to Mrs. Chiara Vescovinifor the editing of the manuscript.

REFERENCES

AERSSENS, J., DEQUEKER, J., PEETERS, J., BREEMANS, S.,BROOS, P., and BOONENE, S. (2000). Polymorphism of the VDR,ER and COLIA1 genes and osteoporotic hip fracture in elderly post-menopausal women. Osteoporosis Int. 11, 583–591.

GARNERO, P., BOREL, O., GRANT, S.F.A., RALSTON, S.H., andDELMAS, P.D. (1998). Collagen Ia1 Sp1 polymorphism, bone mass,and bone turnover in healthy French premenopausal women: theOFELY study. J. Bone Miner. Res. 13, 813–817.

GRANT, S.F.A., REID, D.M., BLAKE, G., HERD, R., FOGELMAN,I., and RALSTON, S.H. (1996). Reduced bone density and osteo-porosis associated with a polymorphic Sp1 binding site in the colla-gen type Ia1 gene. Nature Genet. 14, 203–205.

HEEGAARD, A., JORGENSEN, H.L., VESTERGAARD, A.W., HAS-SAGER, C., and RALSTON, S.H. (2000). Lack of influence of col-lagen type Ia1 Sp1 binding site polymorphism on the rate of boneloss in a cohort of postmenopausal Danish women followed for 18years. Calcif. Tissue Int. 66, 409–413.

LANGDAHL, B.L., RALSTON, S.H., GRANT, S.F.A., and ERIKSEN,E.F. (1998). An Sp1 binding site polymorphism in the COLIa1 genepredicts osteoporotic fractures in both men and women. J. BoneMiner. Res. 13, 1384–1389.

LIDÉN, M., WILÉN, B., LJUNGHALL, S., and MELHUS, H. (1998).Polymorphism at the Sp1 binding site in the collagen type Ia1 genedoes not predict bone mineral density in postmenopausal women inSweden. Calcified Tissue Int. 63, 293–295.

MCGUIGAN, F.E., REID, D.M., and RALSTON, S.H. (2000). Sus-ceptibility to osteoporotic fracture is determined by allelic variationat the Sp1 site, rather than other polymorphic sites at the COLIA1locus. Osteoporos Int. 11, 338–343.

MCGUIGAN, F.E., ARMBRECHT, G., SMITH, R., FELSENBERG,D., REID, D.M., and RALSTON, S.H. (2001). Prediction of osteo-porotic fractures by bone densitometry and COLIA1. A prospective,population-based study in men and women. Osteoporosis Int. 12,91–96.

POPE, F.M., NICHOLLS, A.C., MCPHEAT, J., TALMUD, P., andOWEN, R. (1985). Collagen genes and proteins in osteogenesis im-perfecta. J. Med. Genet. 22, 466–478.

UITTERLINDEN, A.G., BURGER, H., HUANG, Q., YUE, F.,MCGUIGAN, F.E.A., GRANT, S.F.A., HOFMAN, A., VANLEEUWEN, J.P.T.M., POLS, H.A.P., and RALSTON, S.H. (1997).Relation of alleles of the collagen type Ia1 gene to bone density andthe risk of osteoporotic fractures in postmenopausal women. N. Engl.J. Med. 338, 1016–1021.

Address reprint requests to:Dr. L. Montanaro and Dr. C.R. Arciola

Laboratorio di Biocompatibilitàdei Materiali da ImpiantoIstituti Ortopedici Rizzoli

Via di Barbiano,1/1040136 Bologna, Italy

E-mail: carlarenata.arciola@ior.itlucio.montanaro@ior.it

Received for publication May 29, 2001; accepted March 28,2002.

ARMS-PCR AND COLIA1 POLYMORPHISM 57