Apolipoprotein E genotyping using the Investigators have shown thatapproximately 50% of
the variability in normal serum cholesterol levels is due to
polymerase chain reaction and allele-specific
genetic differences amongindividuals (5). It was esti-
oligonucleotide primers mated that as much as 16% of the genetic variance of low
density lipoprotein (LDL) cholesterol was due to allelic
Blair F. Main,* Peter J. H. Jones,',* Ross T. A.
MacGillivray,t and David K. Banfieldt differences at the apoE gene locus (5). Subjects with the
E312 phenotype have about 20% lower, and E314 subjects
Division o Human
f Nutrition*and Department o f have on average 10% higher, levels of LDL cholesterol
Biochemistv, t University .f British Columbia, Vancouveq than subjects possessing the E313 phenotype (6, 7). Due
B. C., Canada V6T lW5
tothe associatiorl between LDL cholesterol levels and
atherosclerosis, it has been suggested that apoE polymor-
Summary A method for apolipoprotein (apo) E genotyping was phism may play a role in determining therisk of coronary
developed using the polymerase chain reaction (PCR) with artery disease (8).
allele-specific oligonucleotide primers (ASP). Synthetic oli- In this report, a rapid, simple, noninvasive technique
gonucleotides with base-pair mismatches at the 3' terminus were for apoE genotyping is described. This method identifies
used as primers to amplify the apoE gene in subjects previously
phenotyped using isoelectric focusing (IEF).Complementary the six common apoE genotypes with high specificity and
primer-allele combinations were specifically amplified by PCR, would be appropriate for the detection of other, less com-
together with a control pair of primers specific to the human mon apoE mutation alleles. The procedure takes advan-
prothrombin gene. Identification of genotype by PCR using tage of the polymerase chain reaction (PCR) in
ASP was consistent with the phenotypes that were determined conjunction with four allele-specific oligonucleotide
by IEF for 14 healthy normolipidemic subjects. These results
primers; each primer is specific for the single base change
Downloaded from www.jlr.org by on February 9, 2010
were achieved using DNA isolated from buccal epithelial cells
obtained from a mouthwash or DNA extracted from leukocytes. that results in either Cys or Arg at positions 112 and 158.
Genotype identification required analysis of the PCR products
on an ethidium-stained agarose gel, yielding results 3 h after
DNA extraction. In comparison with othercurrent methods, MATERIALS AND METHODS
PCR using ASP is suggested as a rapid and simple noninvasive
technique for determining population apoE allelic distribu- Oligonucleotides
tion. -Main, B F , P J. H.
. . . Jones, R. T. A. MacGillivray, and
Allele-specific oligonucleotide primers were synthesized
D. K. Banfield. Apolipoprotein E genotyping using the
polymerase chain reaction and allele-specific oligonucleotide on an Applied Biosystems 391A - PCR Mate synthesizer.
primers. J Lipid Res. 1991. 32: 183-187. Primers were designed with the nucleotide change at the
3' end. The primer sequences and their location within
Supplementary key words polymorphism genotype identification the apoE gene are presented in Fig. 1.
ApolipoproteinE(apoE),a 34,000 mol wt glycopro- Individuals genotyped using allele-specific primers
tein, is a normal component of plasma chylomicrons, very were previously phenotyped by IEF or immunoblot ana-
low density lipoproteins (VLDL), and high density lipo- lysis. This group included 4 E312, 7 E313, 2 E314, and l
proteins (HDL) (1). Mature apoE is a 299 amino acid E4/4. After verification of the methodology, 30 additional
polypeptide that mediates the uptake of apoE-containing subjects were genotyped indicating 22 E313, 6 E314, and
lipoproteins by receptor-mediated endocytosis. 2 E312 genotypes.
The stuctural gene for apoE is polymorphic, coding for
three common isoforms of apoE (€2, €3, and €4) with the Genomic DNA preparation
three alleles producingthree homozygous (E212, E313, DNA was isolated from buccal epithelial cells obtained
E414) and three heterozygous (E213, E2/4, E314) geno- from a mouthwash with sterile water or from leukocytes.
.types (2). The common isoforms differ by amino acid
substitutions at one or each of two sites, residues 112 and
158. The E2 isoform differs from the most common E3 Abbreviations: PCR, polymerase chain reaction; ASP, allele-specific
primers; IEF, isoelectric focusing; VLDL, very low density lipoprotein;
isoform by a substitution of an Arg for a Cys at amino LDL, low density lipoprotein; HDL, high density lipoprotein; dATP,
acid site 158 while E4 differs from E3 due to substitution deoxyadenosine triphosphate; dCTP, deoxycytidine triphosphate;
of Cys for Arg at site 112 (3). As the E4 isoform has one dTTP, deoxythymidine triphosphate; dGTP, deoxyguanosine triphos-
more positive charge thanE3 while E2 has one less phate; DNA, deoxyribosyl nucleic acid.
'To whom reprint requests should be addressed at: Division of Human
positive charge E3, be
isoform can distin- Nutrition, 2205 East Mall, University of British Columbia, Vancouver,
guished by isoelectric focusing (IEF) ( ) 4. B.C., Canada V6T 1W5.
Journal of Lipid Research Volume 32, 1991 183
( H ) 5' AAGGAGTTGAAGGCCTAWAT 3'
cR P P R H P
R A G G S H G G A
Gene 5' AAGGAGTTGAAGGCCTACAAATCG..........GCGCGGCCGCCTGGTGCAGTACC
.............................. GCGCCTGGCAGTGTACCAGGCCG 3'
Sequence 3' TTCCTCAACTTCCGGATGTTTAGC ..........C G C G C C C 5'
3' GCTCCGGCGGACCACGTCAT 5' 3' GCTGACCGTCACATGGTCCG 5' Primer
3' ACTCCGGCGGACCACGTCAT 5'
[F) 3' ACTGACCGTCACATGGTCCG 5'
Fig. 1. Allele-specificoligonucleotide primers with 3' base pair mismatch. (A) Orientation of allele-specific oligonucleotide primers on exon 4 coding
for the 299 amino acid apoE allele. (B) Specific oligonucleotide primers. The gene sequence is representative of an apoE4 allele; - denotes the base
that determines the amino acid Cys or Arg; * denotes the G-T deliberate mismatch. Primer D is used in conjunction with the common primer H
to give amplification of the €4 allele. Primer G is used similarly to detect the presence of the €2 or €3 allele.
Sufficient DNA for 10 PCR reactions was obtained from as internal controls. Samples were overlaid with mineral
a pin-prick method for obtaining blood. The cells were oil (Sigma) and the DNA was denatured for 10 min at
Downloaded from www.jlr.org by on February 9, 2010
first lysed with 0.5 m1 of 0.23 M sucrose, 10 mM Tris-HC1, 96OC. Four units of Thermus aquaticus (Taq) polymerase
p H 7.5, 1% Triton X-100, and then digested for a mini- was added to each sampleprior to the PCR. Samples
mum of 2 h in 10 mM Tris, p H 8.0, 2 mM EDTA, pH 8.0, underwent 25 cycles with each cycle consisting of a 10-sec
10 mM NaCl, 1% SDS, 0.4 mg/ml proteinase K, and 8 denaturation at 96OC, then 30 sec for annealing at 58OC,
(DTT). The digest was then ex- and an extension of 1 min at 65°C.
tracted with phenol-chloroform followed by a
ethanolDNAprecipitation. The DNA was briefly air RESULTS
dried, then resuspended in 10p1of water.
In brief, the mouthwash method required the subject to Primer design
wash hidher mouthwith 10 m1 sterile water; the cells were Initial oligonucleotide primers (Table 1, primers A-C)
then suspended in the above-described sucrose lysis solu- were designed with a single base pair mismatch at the 3'
tion. The solution was then centrifuged at 1500g, the end of the sequence. Results indicated nonspecific
supernatant was poured off, and the lysate was resuspend- amplification unless very stringentannealingtempera-
ed in 100 p1 water. After boiling for 2 min and centrifuga- tures were implemented. At this point various other
tion (Eppendorf) for 5 min, the supernatant was removed buffers were tried and thedNTP concentration decreased,
and used directly in the PCR. yielding inconsistent products. The shifting of the mis-
match 1 base pair in from the 3' end yielded the same
Amplification of genomic DNA nonspecific results. These results indicated that a single
Alleles were determined using the primers described in mismatch was not sufficient to prevent amplification by
Fig. 1. PCR amplification was performed in an auto- the PCR. The introduction of an additional mismatch in-
mated thermocycler (Perkin-Elmer Cetus) as follows. creases the specificity of the primers allowing for specific
Reactions were carried out in atotal volume of 50 p1 con- allele amplification.
taining 1 pg DNA, deoxyadenosine triphosphate (dATP),
deoxycytidine triphosphate (dCTP), deoxythymidine Genotypic identification
triphosphate (dTTP), and deoxyguanosine triphosphate Each PCR reaction contained commonthe primer
(dGTP), each at 0.2 mM. The dNTP stock solution was (primer H)and one of the allele-specific primersde-
prepared by adjusting the magnesiumchloride concentra- scribed in Fig. l. Therefore, four reaction mixtures are re-
tion to 20mM in 5 mM dNTP. The 10 x buffer consisted quiredper subject for genotype diagnosis. Appropriate
of 670 mM Tris-HC1, 15mM magnesium sulfate, 166 mM annealing conditions were determined using an em-
ammonium sulfate, and 100 mM P-mercaptoethanol. pirical, temperature curve approach. The specific primers
Also added were 10% dimethyl sulfoxide (DMSO) and20 were 19 nucleotides in length and the common primer was
pmol of each primer. In additiontoapoEprimers, all 22 nucleotides. An attempt was made to combine two
tubes contained prothrombin primers (see Table 1) used different allele-specific primers in one reaction mixture to
184 Journal of Lipid Research Volume 32, 1991
TABLE 1. Oligonucleotide primer sequences Initial experiments indicated that the internal control
Primer Sequence (5' -.
(Exon 13 of the human prothrombin gene) primers would
not amplify under the stringent annealing conditions
A desribed. The annealing time was increased to 3 min and
B TACTGCACCAGGCGGCCGCA the number of cycles increased from 25 to 35. This gave
D TACTGCACCAGGCGGCCTCG positive results but increased the occurrence of false pro-
E TACTGCACCAGGCGGCCTCA ducts due to nonspecific annealing. To eliminate false an-
F GCCTGGTACACTGCCAGTCG nealing the reactions were seeded with a small aliquot,
H AAGGAGTTGAAGGCCTACAAAT approximately 5 to 10 ng, of cloned human prothrombin
I ACAGAATTCCTGGGCTATGAGCTATGCTC and conditions were returned to 25 cycleswitha 30-sec an-
J ACACTGCAGATAATTCTTTCACGGGCTTG nealing time. Results indicated specific amplification of
both the human prothrombin segment and the polymor-
Primers A and B contain the differentiating nucleotides at the 3' end phic sites of apoE.
while primer C has the mismatch one nucleotide in from the 3' end. The 14 subjects used to verify this procedure were
Primers D and G are designed similarly to A and B but have a second
deliberate mismatch (G-T) three nucleotides in from the 3' end. Primer previously phenotyped using either the isoelectric focus-
H is the common primer. Primers I and J are the human prothrombin ing method or immunoblot technique. Genotypes iden-
internal controls. Oligonucleotide K is a 29-nucleotide primer similar to tified wereE414 [l], E3/4 , E3/3 , E3/2 . The E3/3
genotype yielding a 145 base pair product with primer E,
indicating the presence of Cys 112, and a 277 base pair
product with primer F indicating thepresence of Arg 158.
allow for genotype diagnosis in two reaction mixtures in- The heterozygote E3/4 produces appropriately sized pro-
Downloaded from www.jlr.org by on February 9, 2010
stead of four. The products could be differentiated easily ducts with primers E and F identifying the E3 allele, but
because each set of primers produces a different sized pro- also produces a 145 base pair product with primer D in-
duct. Uneven yieldsof the two product concentrations dicating Arg 112. Homozygous E414 subjects react only
were observed and mayhave been due to each primer with primers D and F marking the presence of Arg 112
competing for the same common primer; thus thesmaller andArg 158, while E212 homozygotes react only with
product always out-competed the larger product. Fig. 2 primers E and G indicating Cys 112 and Cys 158. In all
depicts subject possessing the apoE 313 and E312 geno- reactions a 500 base pair fragment was observed in-
types. dicating a successful amplification of the internal control.
Fig. 2. Apolipoprotein E genotyping using the polymerase chain reaction and allelespecificoligonucleotide
primers. Ethidium bromide-stained agarose gel showing the products from the PCR reactions, as described in
Methods, of apoE 3/3 and apoE 3/2 genotypes. Lane l is negative (primer D), lane 2 is positive (primer F , lane
3 is positive (primer E), and lane 4 is negative (primer G) indicating the homozygous E 313 genotype. Lanes 5
through 8 are the reactions for subject 2. Lane 5 (primer D) is negativeindicating the absence of the e4 allele; lanes
6 through 8 are positive with primers E, F, and G. indicating the presence of E3 and E2 yielding a genotype of
Mitchell cl al. Lovastatin and biliary lipid metabolism 185
DISCUSSION Neubauer, Newbauer, and Edison (21) used this differen-
tial PCR method to detect allelic loss of the p-interferon
Despite the identification of apoE as a genetic factor gene. Human immunodeficiency virus studies have in-
contributing significantly to within-population variation vestigated the effects of different primer-template mis-
in plasma cholesterol level, methods for apoE phenotyp- matches on DNA amplification efficiency (22). Kwok and
ing remain time-consuming and labor-intensive. Isoelec- co-workers (22) found that some mismatches (A:G, G:A,
tric focusing (IEF) on a single polyacrylamide cylindrical and C:C) reduced PCR product yield drastically while
gel has been the method of choice in apoE phenotyping mismatches involving a T amplified efficiently.
in the past (9). This method requires a prolonged ultra- Specificity of a primer for thetemplate may be
centrifugation step and is often unreliable due to variation modulated by at least threemeans:incorporation of a
in isoform staining intensity and distance of migration base-pair altering
deliberate mismatch, annealing
during the electrophoresis procedure.moreA recent temperature, or by decreasing dNTP concentration. We
alternativeapproachtodiagnosingapoEphenotypes is originally tried to differentiate theapoE alleles using
use of a combination of isoelectric focusing of delipidated primers with only a single mismatch at the 3' end and
serum and immunoblotting (10, 11). This methodiden- results from the agarose gel indicated a large amount of
tifies apoE phenotypes immunologically using a double nonspecific amplification. New oligonucleotides were syn-
antibody reaction and does not require the long ultracen- thesized with the mismatch 1 base pair from the 3' end;
trifugationstep used in IEF (11). However, a highly however, a significant amount of nonspecific amplification
specific antibody preparation is required. was observed. The addition of further a deliberate
More recently, a DNA amplification technique has mismatch enhanced the specificity greatly. We have ob-
been developed using the polymerase chain reaction with tained optimal results using a more stringent annealing
Downloaded from www.jlr.org by on February 9, 2010
allele-specific oligonucleotide probes (12-16). These temperature of 58OC because lower temperatures also
probes are used in conjunction with the amplified DNA resulted in nonspecific amplification. To allow for
product from the PCR to detect cysteine-arginine inter- amplification of the larger human prothrombin segment,
changes at residues 112 and 158 of the apoE allele that dis- the reactions are seeded with a cloned DNA sequence of
tinguish the variousisoforms. Specific amplification of the the human prothrombin gene, allowing amplification of
apoE allele using the is hy-
PCR followed the internal control indicating each PCR mixture con-
bridization with radioactively labeled allele-specific oli- tained the appropriate reagents. The restriction sites in-
gonucleotide probes, which is multi-step and requires corporated into the internal control primerswere used for
radioisotope. Restriction isotyping of apoE is another other purposes than apoE genotyping.
methodthat uses oligonucleotides to amplify theapoE Initially, amplification of the various apoE alleles yield-
gene (17). The PCR products are digested with a restric- ed inconsistent amounts of product. Primer K was syn-
tionenzyme andthen separated by electrophoresis on thesized in an attempt to alleviate this problem by
polyacrylamide gels. The restriction isotyping method re- the
increasing to 29 nucleotides. Aspecific
quires an incubation step which extends this technique's amplification was observed at a variety of different an-
diagnosis time. nealing temperatures andbuffer conditions. Winship (23)
To develop a simple, specific method for identification observed that G-C-rich sequences were difficult to
of apoE genotypes, we have modified the previously amplify and added DMSO to the PCR mixture. This is
reported PCR technique. The modifications eliminate the thought to reduce the secondary structure of DNA result-
need for radiolabeled probes as well as all incubation, ing in amplification of G-C-rich sequences. The addition
hybridization, and washing procedures. This new tech- of 10% DMSO to our PCR mixtures increased the yield
nique requires only an agarose gel for an assay and can of certain apoE segmentsspecifically even though DMSO
diagnose an apoE genotype within 3 h after DNA isola- is thought to reduce Taq polymerase activity by approx-
tion. imately 50% (24).
Diagnosis of individuals with alpha l-antitrypsin defi- Presently, this novel PCR-allele-specific primer
ciency or cystic fibrosis has used similar approaches to method of apoE genotyping offers advantages of simplici-
identify point mutations (18-20). In each application, the ty and cost compared with othermethods. Obtaining
identical principal is used; oligonucleotides with a 3' DNA from a mouthwash also obviates the need for veni-
mismatched base pair will not function as primers in the puncture (25). the
In addition, method is based on
PCR mixture. In some cases a single mismatch is not nucleotide sequence variation; therefore, by using other
sufficient to prevent nonspecific amplification and a se- specific primers,uncommonstructural mutants of the
cond deliberate mismatch must be incorporated into the apoE allele could be analyzed. Using past methods such
oligonucleotide (18). Investigators have used this tech- as IEF and immunoblotting, these alleles cannot be de-
nique to detectdifferent mutations in an allele selectively. tected. This new procedure may enable large populations
186 Journal of Lipid Research Volume 32, 1991
to be screened rapidly and accurately, thus contributing 12. Weisgraber, K. H., Y. M. Newhouse, and R. W. Mahley.
to our understanding apoE’s allelic distribution and its
of 1988. ApoE genotyping using the polymerase chain reac-
regulation of plasma cholesterol levels. I tion and allele-specific oligonucleotide probes. Bzochem.
Biophys. Commun. 57: 1212-1217.
13. Houlston, R. S., C. Snowden, F. Green, K. G. M. M.
The authors appreciate the assistance and insight ofB. Chow, Alberti, and S. E. Humphries. 1989. ApoE genotypes by
W. Funk, and B. Ritchie. Also greatly appreciated are the efforts polymerase chain reaction and allele-specific oli-
ofJ. Frohlich, J. Hill, H. Pritchard, and R. Roe for phenotyping gonucleotide probes: no detectable linkage disequilibrium
subjects by isoelectric focusing and immunoblotting. This between apoE and apoC-11. Hum. Genet. 83: 364-368.
research was supported by a grant from the British Columbia 14. Emi, M,, L. L. Wu, M. A. Robertson, R. L. Myers, R. A.
Health Care Research Foundation. Hegele, R.R. Williams, R. White, and J. M. Lalouel.
1988. Genotyping and sequence analysis of apolipoprotein
Manuscript received 22 June 1990 and in revisedform 10 September 1990.
E isoforms. Genomics. 3: 373-379.
15. Smeets, H. J. M,, J. Poddighe, P. M. J. Stuyt, A. F. H.
Stalenhoef, H. H. Ropers, and B. Wieringa. 1988. Iden-
REFERENCES tification of apoE polymorphism by using synthetic oli-
gonucleotides. J. Lipid Res. 29: 1231-1237.
1. Havel, R. J., N. Yamada, and D. M. Shames. 1987.Role 16. Funke, R., S. Rust,and G. Assman. 1986. Detection of
of apolipoprotein E in lipoprotein metabolism. Am. HeartJ. apoE variants by an oligonucleotide “melting” procedure.
2: 470-474. Clin.Chem. 32: 1285-1289.
2. Zannis, V. I., P.W. Just, and J. L. Breslow.1981. Human 17. Hixson, J. E., and D. T. Vernier. 1990. Restriction isotyp-
apoE isoprotein subclasses are genetically determined. Am. ing of human apolipoprotein E by gene amplification and
J. Hum. Genet. 33: 1033-1041. cleavage with HhaI J Lipid Res. 31: 545-548.
3. Rall, S. C., K. H. Weisgraber, and R. W. Mahley. 1982. 18. Ballabio, A., R. A. Gibbs, and C. T. Caskey. 1990. PCR
Human apoE: the complete amino acid sequence. J. Biol. test for cystic fibrosis deletion. Nature. 343: 220.
Downloaded from www.jlr.org by on February 9, 2010
C h a . 257: 4171-4178. 19. Newton, C. R., A. Graham, L. E. Heptinstall, S. J. Powell,
4. Wardell, M. R., P. A. Suckling, and E. D. Janus. 1982. C. Summers, N. Kalsheker, J. C. Smith, and A. F.
Genetic variation in human apoE. J. Lipid Res. 23: Markham. 1989. Analysis of any point mutation in DNA.
1174-1182. The amplification refractory mutation system (ARMS).
5. Sing, C. F., and J. Davignon. 1985. Role of the apoE poly- NucleicAcidsRes. 17: 2503-2516.
morphism in determining normalplasma lipid and lipopro- 20. Okayama H., T. Curiel, M. L. Brantly, M. D. Holmes,
tein variation. Am. J. Hum. Genet. 37: 268-285. and R. G. Crystal. 1989. Rapid, nonradioactive detection
6. Utermann, G. 1986. The apoE system: genetic control of of mutations in the human genome by allele-specific
plasma lipoprotein concentration. In AdvancesEx- in amplification. J. Lab.Clin.Med. 114: 105-113.
perimental Medicine. A.Angel and J. Frohlich, editors. 21. Neubauer, A., B. Neubauer, and L. Edison. 1990.
Plenum Press, NewYork.261-273. Polymerase chain reaction-based assay to detect allelic loss
7. Schaefer, E. J., R. E. Gregg, and E. Ghiselli. 1986. Familial in human DNA: loss of beta-interferon gene in chronic
apoE deficiency.J. Clin. Invest. 78: 1206-1219. myelogenous leukemia. NucleicAcidsRes. 18: 993-998.
8. Kuusi, T., M. S. Nieminem, C. Ehnholm, H. Y. Jarvinen, 22. Kwok, S., D. E. Kellogg,N. McKinney, D. Spasic, L.
M. Valle, E. A. Nikkila, and M. R. Taskinen. 1989. ApoE Goda,C. Levenson, and J. J. Sninsky. 1990. of
polymorphism and coronary artery disease. Arteriosclerosis. primer-template mismatches on the polymerase chain reac-
9: 237-241. tion: human immunodeficiency virus type 1 model studies.
9. Warnick, G. R., C. Mayfield, J. J. Albers, and W. R. NucleicAcids Res. 18: 999-1005.
Hazzard. 1979. Gel isoelectric focusing method for specific 23. Winship, P. R. 1989. An improved method for directly se-
diagnosis of familial hyperlipoproteinemia type 111. Clin. quencing PCR amplified material using dimethyl sulphox-
Chem. 25: 279-284. ide. NucleicAcids Res. 17: 1266.
10. Havekes, L. M., P. de Knijff, U. Beisiegel, J. Havinga, 24. Lawyer, F. C., S. Stoffel, R. K. Saiki, K. Myambo, R.
M. Smit, and E. Klasen. 1987. A rapid micromethod for Drummond, and D. H. Gelfand. 1989. Isolation,
apolipoprotein E phenotyping directly in serum. J. Lipid characterization, and expression in E. coli of the DNA
Res. 28: 455-463. polymerase gene from Thermus aquaticus. J Biol. Chem. 264:
11. Kamboh, M. I., R.E. Ferrell, and B. Kottke. 1988. Genetic 6427-6437.
studies of human apolipoproteins. V. A novel rapid proce- 25. Lench, N., P. Stanier, and R. Williamson. 1988. Simple
dure to screen apoE polymorphism. J. Lipid Res. 29: noninvasive method to obtain DNA for gene analysis.
1535-1543. Lancet. 1: 1356-1358.
Mitchell et al. biliary
Lovastatin and lipid metabolism 187