Clinical Chemistry 50:3
481– 489 (2004)
Rapid Detection of K-ras Mutations in Bile by
Peptide Nucleic Acid-mediated PCR Clamping
and Melting Curve Analysis: Comparison
with Restriction Fragment Length
Polymorphism Analysis
Chiung-Yu Chen,1 Shu-Chu Shiesh,2* and Sheu-Jen Wu3
Background: Current methods for detection of K-ras
gene mutations are time-consuming. We aimed to de-
velop a one-step PCR technique using fluorescent hy-
bridization probes and competing peptide nucleic acid
oligomers to detect K-ras mutations in bile and to
compare the efficacy with restriction fragment length
polymorphism (RFLP) analysis.
Methods: Bile samples were obtained from 116 patients
with biliary obstruction, including gallstones (n ⴝ 64),
benign biliary strictures (n ⴝ 6), pancreatic cancer (n ⴝ
20), and cholangiocarcinoma (n ⴝ 26). The DNA was
extracted and subjected to K-ras mutation analysis by
real-time PCR and RFLP analysis. Mutations were con-
firmed by direct sequencing. The sensitivity and speci-
ficity were calculated according to the clinical results.
Results: The analysis time for real-time PCR was <1 h,
whereas RFLP analysis took more than 2 days. With the
sensor probe designed for the GAT (G12D) mutant in
codon 12 of the K-ras gene, the real-time PCR method
also detected the GTT (G12V) mutant. In contrast, a
specific sensor probe for the TGT (G12C) mutant de-
tected GAT (G12D), AGT (G12S), and GTT (G12V)
mutants in addition to the TGT mutant. The real-time
PCR assay allowed the detection of mutation in a
3000-fold excess of wild-type bile DNA. In bile, K-ras
codon 12 mutations were detected in 16 of 46 malignant
Departments of 1 Internal Medicine and 2 Medical Technology, College of
Medicine, National Cheng Kung University, Tainan, Taiwan.
3
Clinical Biochemistry Research Laboratory, Veterans General Hospital,
Taipei, Taiwan.
*Address correspondence to this author at: Department of Medical Tech-
nology, National Cheng Kung University, 1 University Road, Tainan 701,
Taiwan. Fax 886-6-236-3956; e-mail hsieh@mail.ncku.edu.tw.
Received July 16, 2003; accepted December 16, 2003.
Previously published online at DOI: 10.1373/clinchem.2003.024505
481
Molecular Diagnostics
and Genetics
cases by real-time PCR with the TGT probe and 15 by
RFLP analysis. All benign cases were wild type.
Conclusion: Real-time PCR with a cysteine-specific
(TGT) sensor probe can rapidly detect K-ras gene mu-
tations in bile and diagnose malignant biliary obstruc-
tion with high specificity.
© 2004 American Association for Clinical Chemistry
It is well documented that malignant transformations are
a result of the accumulation of carcinogenic steps corre-
sponding to activation of oncogenes and inactivation of
tumor suppressor genes (1 ). Among the available candi-
dates, the K-ras protooncogene is the most well-studied
cellular gene whose alterations seem to have an important
role in the pathogenesis of human cancer. The gene,
which encodes 21-kDa GTP-binding protein, controls the
mechanisms of cell growth and differentiation. Point
mutations in the K-ras gene lead to uncontrolled stimula-
tion of ras-related functions by the altering p21 ras pro-
tein-related pathway, locking it in the “on” position for
signal transduction (2 ).
The presence of K-ras gene mutations in bile, duodenal
juice, and pancreatic juice has been used to distinguish
benign from malignant biliary strictures (3– 8 ). Although
K-ras mutations have been detected in some patients with
benign biliary strictures, such as chronic pancreatitis and
gallstones (3, 6 ), the detection of such mutations appears
to be specific and is considered to be of help in differen-
tiating benign from malignant strictures (3–5 ). Caldas et
al. (9 ) noted that activating point mutations in codon 12 of
the K-ras gene occur in approximately one-half of non-
papillary duct lesions but were present in the majority of
papillary, more advanced pancreatic cancers. Cholangio-
carcinoma also carries K-ras mutations, but the reported
frequencies, as assessed in tissue DNA, vary considerably,
from 20 –100% in cases in Western countries (10 –12 ) to
482 Chen et al.: Rapid Detection of K-ras Mutations in Bile
5–58% in Japanese cases (13 ). In general, the frequencies
of K-ras mutations in biliary duct carcinoma were re-
ported to be 8 –100% in tissue DNA, 5–30% in bile DNA
(3, 4, 10 ), and 42% in brushing fluids (14 ). The point
mutations reside mainly in the first two nucleotides of
codon 12. Sequencing analysis has revealed that the
single-nucleotide substitutions of K-ras codon 12 are
mainly GAT, GTT, and CGT, compared with the wild-type
codon GGT (10 ) for patients with pancreatic cancer and
GAT, GTT, and TGT/AGT/GCT for patients with cholan-
giocarcinoma (15 ).
In addition to differences in the geographic etiologies
and carcinogenesis of cholangiocarcinoma, variations in
the sensitivities of the methods used for detection of the
K-ras mutation might explain some of the differences in
K-ras mutation frequencies among studies. K-ras muta-
tions are detected mainly by mutation-specific oligonucle-
otide hybridization, PCR with restriction fragment length
polymorphism (RFLP) analysis, single-strand conforma-
tion polymorphism analysis, mutant allelic-specific am-
plification, or direct sequencing. These procedures in-
volve multiple steps and are time-consuming; they
therefore are impractical for routine clinical use. With
recent innovations in molecular technology, a new high-
speed thermal cycler with online detection has been
developed to detect N-ras gene mutations in leukemia
(16 ). Such real-time systems offer many advantages, in-
cluding speed, reduced risk of contamination, and higher
accuracy (17 ). To detect a minimal amount of mutant in
clinical samples, peptide nucleic acid (PNA) oligomers
have been used to improve the allele-specific PCR by
suppressing the amplification of the background wild
type (18 ). This PNA-mediated PCR clamping technique,
followed by ethidium bromide staining (19 ), a PCR-RFLP
step (20 ), or mass spectrophotometry (21 ), has been
applied to K-ras mutations in tumor samples. Combining
PNA-mediated PCR clamping and use of a real-time
thermal cycler, Sotlar et al. (22 ) detected the c-kit point
mutation D816V in skin biopsy samples from patients
with urticaria pigmentosa.
We conducted this study to establish a one-step real-
time PCR method that combines a competing PNA oli-
gomer (clamped probe assay) with fluorescent hybridiza-
tion probes and melting curve analysis for the detection of
K-ras gene codon 12 mutations in bile and to validate their
clinical use in the differential diagnosis of biliary obstruc-
tion.
Materials and Methods
collection of bile
Fasting bile was obtained from patients with biliary
obstruction via an endoscopic nasal biliary drainage cath-
eter or a percutaneous transhepatic biliary drainage cath-
eter on the day of biliary drainage. The bile was cen-
trifuged at 6290g for 10 min, and the pellets and
supernatants were stored at ⫺80 °C. The causes of biliary
obstruction included hepatolithiasis (n ⫽ 17), choledocho-
lithasis (n ⫽ 47), benign biliary stricture (n ⫽ 6), pancre-
atic cancer (n ⫽ 20), and cholangiocarcinoma (n ⫽ 26).
Patients were either diagnosed based on pathology as
having biliary or pancreatic cancer after surgery or bi-
opsy, or by imaging studies and a long-term follow-up
that showed the progress of index lesions.
rflp analysis of K-ras codon 12 mutations
DNA was extracted from the pellet by proteinase K
treatment followed by the Qiagen DNA Isolation Kit
(Qiagen). An enriched PCR-RFLP was performed for the
detection of K-ras codon 12 mutations as described previ-
ously (10, 23 ) with minor modifications. In brief, the PCR
was conducted in a GeneAmp 9700 thermocycler (Ap-
plied Biosystems) using K1A (5⬘-ACTGAATATAAACTT-
GTGGTAGTTGGACCT-3⬘) and K1B (5⬘-TCAAAGAATG-
GTCCTGGACC-3⬘) as primers (23 ). The primers
contained a mismatch near the 3⬘ ends (underlined base)
to generate a BstNI site right upstream of codon 12 in
cases of wild type, but not in K-ras mutants. The PCR
mixture contained 100 ng of DNA template, 1 ␮M each of
the primers, 200 ␮M deoxynucleotide triphosphates, 1.5
mM MgCl2, and 2 U of Taq polymerase in a total volume
of 50 ␮L. The conditions for the PCR were 94 °C for 5 min,
followed by 94 °C for 1 min, 55 °C for 90 s, and 72 °C for
90 s in each cycle for 20 cycles. After PCR, 5 ␮L of the first
PCR product was digested with 2 U of BstNI (Biolabs) at
60 °C for 8 h. We used 1/1000 of the first digest as a
template for the second PCR with 30 cycles using K1A
and K1C (5⬘-GCATATTAAAACAAGATTTAC-3⬘) as
primers. The second PCR product was digested with
BstNI again. After digestion, the DNA products were
separated by electrophoresis on 3% agarose gels at 100 V
for 1 h and stained with ethidium bromide (Sigma). To
rule out PCR- and/or incomplete digestion-generated
artifacts, all cases with K-ras mutations or equivocal
results were confirmed by repeating PCR-RFLP analysis
with different dilutions of the first digest.
clamped probe assay for analysis of K-ras
mutations
A set of primers was chosen to amplify a specific 164-bp
genomic fragment from K-ras exon 1. Hybridization
probes were designed complementary to specific mutant
types in K-ras codon 12; the anchor probe was 3⬘-labeled
with fluorescein, and the sensor probe 5⬘-labeled with
LC-Red dye and 3⬘-phosphorylated. Two different sensor
probes were selected to cover the mutation region: (a)
5⬘-LC-Red 640-labeled oligonucleotide probe to bind the
K-ras GAT (G12D) mutation (i.e., the 12Asp sensor probe),
and (b) 5⬘-LC-Red 705-labeled probe to bind the TGT
(G12C) mutation of K-ras gene (12Cys sensor probe). The
different dyes labeled on the probes allowed us to per-
form these two assays in the same run. The competing
wild-type PNA oligomer covered codons 10 –14. The
antisense strand was chosen for the PNA oligomer and
the detection probes because of its lower purine content
 Clinical Chemistry 50, No. 3, 2004
Table 1. DNA sequences of primers, PNA oligomer, and probe for detecting K-ras mutations.
Name DNA (5ⴕ–3ⴕ) or PNA (NH2–CONH2) sequencea
c
K-ras F AGGCCTGCTGAAAATGACTG
K-ras R GGTCCTGCACCAGTAATATGCA
Anchor CGTCCACAAAATGATTCTGAATTAGCTGTATCGTCAAGGCACT-Flu
12Asp LC-Red 640-CCTACGCCATCAGCTCCA-P
12Cys LC-Red 705-TTGCCTACGCCACAAGCTCCAA-P
PNA CCTACGCCACCAGCTCC
a
LC-Red 640 and LC-Red 705 are fluorophores. The 3⬘ end of the mutation probe was phosphorylated to prevent probe elongation by Taq polymerase during PCR.
Bold, codon 12; underlined, mutated base.
b
The base numbering is according to GenBank accession no. L00045.
c
F, forward; R, reverse; Flu, fluorescein; LC-Red 640, LightCycler-Red 640; P, phosphorylated; LC-Red 705, LightCycler-Red 705.
and, therefore, more precise hybridization results (24 ).
The primers, probes, and the PNA oligomer were from
TIB MOLBIOL. The sequences of the primers, probes, and
the PNA oligomer are listed in Table 1.
For amplification with the 12Asp sensor probe, we
mixed the LightCycler DNA Master Hybridization Mix-
ture (Taq polymerase, PCR buffer, and deoxynucleotide
triphosphate mixture; Roche Diagnostics), 5 mM MgCl2,
0.3 ␮M anchor probe, 0.15 ␮M sensor probe, 0.15 ␮M PNA
oligomer, and 0.5 ␮M each of primers K-ras F and K-ras R
with 10 ng of genomic DNA and added water to a final
volume of 20 ␮L/capillary. PCR was performed with the
Roche LightCycler System (Roche Diagnostics). The con-
ditions for the 12Cys sensor probe were the same as for
the 12Asp sensor probe, except that we used 50 ng of
genomic DNA, 1.5 ␮M PNA oligomer, and 3 mM MgCl2.
The PCR protocol consisted of 95 °C for 10 min for initial
denaturation, 50 cycles of 2 s at 95 °C for denaturation,
10 s at 70 °C for PNA oligomer binding, 7 s at 60 °C for
primer annealing and probe binding, and 15 s at 72 °C for
extension.
controls
DNA from the colon cancer cell line SW480, which
harbors a homozygous GGT-to-GTT mutation at codon 12
of K-ras, was used as positive control in each run. One
negative control (DNA from colon cancer cell line HT29
with wild-type K-ras) and a water negative control (as con-
trol for contamination) were processed in parallel with each
batch of samples. Mutation analysis for each bile sample
was performed at least twice to confirm reproducibility.
dilution experiments
We evaluated the sensitivities of the methods for K-ras
mutation detection by diluting SW480 cell line DNA,
which carries a homozygous G12V mutation, or a bile
DNA sample carrying a heterozygous G12C mutation
(confirmed by sequencing) in wild-type human bile DNA.
The 1:10 mixture contained 1 ng of mutant DNA and 10
ng of wild-type DNA.
melting curve analysis
After amplification, melting analysis was performed by
holding the denaturation reaction at 95 °C for 20 s and
483
Position,b nt
83–102
246–225
186–144
139–122
142–121
139–123
hybridization at 40 °C for 20 s, followed by heating from
40 °C to 85 °C at 0.3 °C/s with continuous monitoring of
fluorescence at 640 nm (for LC-Red 640 in channel 2) or
705 nm (for LC-Red 705 in channel 3). The change of
fluorescence was converted to a melting peak (Tm) by
plotting the negative derivative of the fluorescent signal
corresponding to the temperature (⫺dF/dT) with the
software. The K-ras gene mutation was identified by
comparing the Tm of each patient’s result with that of the
DNA positive control.
To optimize the PCR condition, we adjusted the Mg2⫹,
PNA oligomer, and DNA concentrations. Mg2⫹ titration
experiments were done with a patient sample harboring
the K-ras codon 12 TGT (G12C) mutant as determined by
the clamped probe assay and direct sequencing. The
clamped probe assay and melting curve analysis were
performed as described above except that the MgCl2
concentration used was varied from 1 to 7 mM (total of
seven concentrations).
dna sequencing
The DNA from those samples that gave positive results
either by RFLP analysis or the clamped probe assay was
further analyzed by direct sequencing. The second PCR
product, visualized as a 129-bp band on the electrophore-
sis gel, was purified by use of a QIAquick Gel Extraction
Kit (Qiagen), and concentrated (from 250 ␮L to 15 ␮L) in
a vacuum centrifuge dryer (Hetovac; VR-1). Bidirectional
DNA sequencing of PCR products was carried out on an
ABI PRISM 377 sequencer (Applied Biosystems) with a
Big Dye Terminator Cycle Sequencing Ready Reaction
Kit.
statistical analysis
All data are expressed as the mean (SD). The statistical
analysis was performed with SPSS 8.0 for Windows
software (SPSS).
Results
demographic data for patients
We had bile from 116 patients for analysis of K-ras codon
12 mutations (Table 2). The diagnosis of cholangiocarci-
noma (n ⫽ 26) or pancreatic cancer (n ⫽ 20) was con-
firmed pathologically in 34 patients (74%). Patients with-
484 Chen et al.: Rapid Detection of K-ras Mutations in Bile
Table 2. Demographic data for patients.
Mean (SD) Mean (SD)
Group M:F age, years follow-up, months
Pancreatic cancer 12:8 64.1 (13.4) 8.9 (13.3)
Cholangiocarcinoma 13:13 67.0 (9.5) 6.2 (7.2)
Hepatolithiasis 7:10 55.8 (15.6) 18.4 (17.6)a,b
Choledocholithasis 29:18 67.3 (15.1) 13.3 (13.1)b
Benign biliary stricture 4:2 57.8 (17.5) 20.6 (15.4)b
a
P ⬍0.05 between this group and pancreatic cancer group.
b
P ⬍0.05 between this group and the cholangiocarcinoma group by one-way
ANOVA.
out a pathology diagnosis were diagnosed based on the
severe and lethal course of the disease, which accounts for
the shorter follow-up period in these patients than for
patients with benign biliary obstruction. The patients who
were diagnosed with benign biliary stricture included
those with chronic pancreatitis (n ⫽ 2), pancreatic tuber-
culosis (n ⫽ 1), papillary stenosis (n ⫽ 2), and serous
adenoma of the pancreas (n ⫽ 1). None of the patients
with gallstones or benign biliary obstructions developed
biliary malignancies during a follow-up period of 15.2
(14.6) months.
analysis of mutant and wild-type K-ras by
clamped probe assay
Real-time PCR with the clamped probe assay accurately
identified the K-ras mutation from the wild-type genome.
In the absence of PNA oligomer, we observed a melting
peak at 66.3 °C for the wild type with the 12Cys sensor
probe. The addition of PNA oligomer successfully sup-
pressed the amplification of the wild type. For the 12Cys
sensor probe, Mg2⫹ titration experiments showed that the
reaction with 3 mM MgCl2 provided the highest melting
peak for bile DNA with the K-ras codon 12TGT mutant
(Fig. 1 in the Data Supplement that accompanies the
online version of this article at http://www.clinchem.
org/content/vol50/issue3/). The effects of various
amounts of PNA oligomer, Mg2⫹, and DNA template on
the melting curves obtained with the 12Cys probe are
shown in Fig. 2 in the online Data Supplement. In the
presence of 1.5 ␮M PNA oligomer, 3 mM Mg2⫹, and 50 ng
of DNA, the clamped probe assay showed the least
wild-type background and best mutant signal and was
applied to the rest of the reactions. Table 3 and Figs. 3 and
4 provide the results of testing for K-ras codon 12 muta-
tions. The melting curve analysis showed that the 12Cys
sensor probe can detect four kinds of K-ras codon 12
mutations—GAT, GTT, AGT, and TGT mutants—with
mean (SD) melting temperatures of 65.1 (0.18), 66.5 (0.29),
67.8 (0.17), and 70.6 (0.22) °C, respectively (Fig. 1 and
Table 3). With the 12Asp sensor probe, we observed a
melting peak at 62.5 °C for the wild type in the absence of
PNA oligomer. The 12Asp sensor probe can detect only
two kinds of point mutation, however: the GTT and GAT
mutations with melting temperatures of 61.9 (0.45) and
67.1 (0.40) °C, respectively (Fig. 2).
Table 3. Melting temperatures of mutants by real-time PCR
with PNA and codon 12 mutation-specific sensor probes.
Mean (SD) Tm, °C
Mutant No. of cases 12Cys probe 12Asp probe
G12D (GAT) 8 65.14 (0.18) 67.1 (0.40)
G12C (TGT) 2 70.65 (0.22) ⫺a
G12S (AGT) 1 67.8 (0.17)b ⫺a
G12V (GTT) 5 66.5 (0.29) 61.9 (0.45)c
a
No melting peaks detected.
b
Obtained by repeating the measurement in five different runs.
c
One case was not detected by 12Asp probe.
rflp analysis of mutant and wild-type K-ras
Typical RFLP results are shown in Fig. 3. PCR with
primers K1A and K1B gave rise to a 157-bp fragment
containing two BstNI restriction sites if codon 12 was wild
type and one BstNI site if codon 12 contained a mutation
in either of its first two bases. Hence, BstNI cut the
amplified wild-type fragment twice to yield 29-, 114-, and
14-bp fragments, but cut K-ras codon 12 mutant once to
generate 143- and 14-bp fragments. When a second PCR
was performed with primers K1A and K1C, the mutant
K-ras fragment after amplification was 129 bp, which was
no longer cut by BstNI, whereas the wild-type fragment
was further cleaved by BstNI to a 100-bp fragment. If we
counted those with a single band of 129 bp or heterozy-
gous bands of 129 and 100 bp as mutants, the frequency of
K-ras mutations in bile would be 40% (28 of 70) and 59%
(27 of 46) in benign and malignant cases, respectively. To
avoid possible errors from PCR or incomplete digestion,
mutations were counted only in those with a single band
of 129 bp or a predominant band of 129 bp (signal of
129-bp band greater than that of 100-bp band) on the
second PCR product after digestion.
comparison of different methods for
detecting K-ras mutations
The clamped probe assay detected K-ras codon 12 mu-
tants within 1 h, whereas the RFLP analysis took 3 days to
complete. Overall, 15 of 116 cases tested positive for a
mutation in codon 12 of K-ras with the RFLP method,
whereas 16 had mutations with the clamped probe assay
using a 12Cys sensor probe and 12 had mutations with the
12Asp sensor probe (Table 4). The positive samples de-
tected by RFLP analysis or clamped probe assay were
subjected to direct sequencing. In 17 cases, sequence
results showed 15 mutations. In malignant cases, there
was one K-ras mutation detected by the clamped probe
assay (using a 12Cys or 12Asp probe) but not by RFLP
analysis or sequencing, one detected by the clamped
probe assay and sequencing but not by RFLP analysis, one
detected by the clamped probe assay and RFLP analysis
but not by sequencing, and one detected by RFLP analysis
and sequencing but not by the clamped probe assay.
We assessed the sensitivities of these methods for
 Clinical Chemistry 50, No. 3, 2004 485

Fig. 1. Representative melting curves obtained with the clamped probe assay with the 12Cys sensor probe.
Results show GAT (G12D), GTT (G12V), AGT (G12S), and TGT (G12C) mutations at codon 12 of the K-ras gene.

detection of K-ras mutations by dilution experiments
using various amounts of homozygous G12V mutant
SW480 DNA or heterozygous G12C mutant bile DNA in
wild-type bile DNA. In RFLP analysis, the mutation was
detectable in the 1:100 mixture of SW480 and wild-type
DNA and the 1:10 mixture of G12C mutant and wild-type
bile DNA (Figs. 3, B and C). In direct sequencing of
concentrated PCR products, the mutation was detectable
in the 1:20 mixture of SW480 and wild-type DNA and the
1:10 mixture of G12C mutant and wild-type bile DNA.
The clamped probe assay detected SW480 K-ras DNA
mutations in a 3000-fold excess of wild-type bile DNA
when we used the 12Cys sensor probe and 1000-fold
when we used the 12Asp sensor probe. However, for
detection of G12C mutant bile DNA, the mutation-specific
melting peak was detected only down to a 20-fold excess
of wild-type bile DNA (Fig. 4)
clinical performance of different methods for
the diagnosis of pancreaticobiliary
malignancies
In bile, K-ras codon 12 mutations were detected in 16 of 46
(35%) patients with malignancy by the clamped probe
assay using the 12Cys sensor probe, 12 (26%) by the assay
using the 12Asp sensor probe, and 15 (33%) by RFLP
analysis. All benign cases were wild type by these meth-
ods.
Discussion
Mutations in the ras oncogene are frequent in malignant
pancreatic disease (10, 25, 26 ) and variable in cholangio-
carcinoma (3, 27–29 ). K-ras codon 12 mutations have been
detected in 80 –95% of infiltrating pancreatic ductal carci-
noma cases (30 ) and in 67% of cases with perihilar and
50% of cases with intrahepatic cholangiocarcinoma (29 ).
486 Chen et al.: Rapid Detection of K-ras Mutations in Bile

Fig. 2. Representative melting curves obtained with the clamped probe assay with the 12Asp sensor probe.
Curves represent GAT (G12D) and GTT (G12V) mutations at codon 12 of the K-ras gene.

However, lower frequencies of K-ras gene mutations
assessed from bile samples were found. In this study, we
developed a one-step real-time PCR and detected K-ras
mutations in approximately one-third of patients with a
malignant etiology for the biliary stenosis, whereas mu-
tations were not present in any bile from patients with
benign obstructions. The mutation frequency in bile was
comparable to the frequencies reported in previous stud-
ies (3, 4, 8, 31 ). This low frequency may be explained by
the scirrhous nature of cholangiocarcinoma; the extraduc-
tal growth of pancreatic cancer, which makes it difficult
for the cancer cells to be exfoliated into the bile; and/or
the PCR inhibitors present in bile. Nevertheless, the high
specificity of K-ras gene analysis in diagnosing biliary
malignancies remains valuable for clinical use.
The ideal method for K-ras gene analysis in clinical
practice would be rapid and accurate enough to allow for
high-throughput screening. The results of the present
study indicate that real-time PCR using fluorescence
hybridization probes with a PNA clamping reaction and
melting curve analysis may offer a promising method to
accurately identify K-ras gene mutations in patients with
biliary obstruction. The absence of post-PCR manipula-
tion reduces the risk of carryover contamination and
false-positive results. This detection procedure also elim-
inates the need for the restriction enzyme digestion,
electrophoresis, and staining steps required in the RFLP and
single-strand conformation polymorphism methods. Be-
cause rapid temperature changes are possible in the closed
glass capillary, PCR occurs under rapid cycling conditions
and the entire process can be completed within 1 h. To-
gether, these advantages make this assay feasible and prom-
ising for screening larger numbers of clinical samples.
In this study, we found a high frequency of G-to-A
transitions (n ⫽ 9) and G-to-T transversions (n ⫽ 7) within
codon 12, which is in agreement with the literature
(3, 13, 15 ). Using a mutant-specific sensor probe and
melting curve analysis, we were able to identify different
K-ras mutations. The different affinities between the mu-
tant-specific sensor probe and K-ras mutations lead to
changes in DNA thermal stability, which in turn lead to
different melting temperatures. The technique has also
 Clinical Chemistry 50, No. 3, 2004 487

Fig. 3. PCR-RFLP analysis of K-ras mutations.

(A), the gel on the left (mutant result) shows a single 129-bp
band for P2 (the second PCR product) and D2 (after enzyme
digestion of P2). The gel on the right (wild-type result) shows
a single 129-bp band for P2 (the second PCR product) and
100-bp band for D2 (after enzyme digestion of P2). (B),
dilution experiment using mixtures of DNA from cell line
SW480, which harbors a homozygous K-ras codon 12 muta-
tion (G12V), and DNA from a bile sample with wild-type K-ras.
Lane pairs: 1, 1:2 mixture; 2, 1:10 mixture; 3, 1:20 mixture;
4, 1:50 mixture; 5, 1:100 mixture; 6, 1:1000 mixture. The
band on the left in each lane pair represents the second PCR
product; the band on the right represents the result of
enzyme digestion of the second PCR product. (C), dilution
experiment using mixtures of DNA from a bile sample
harboring a heterozygous K-ras codon 12 mutation (G12C)
and DNA from a bile sample with wild-type K-ras. Lane M,
DNA marker.

been applied successfully in the point mutation of c-kit (22 ) the 12Cys sensor probe detected more kinds of K-ras
and the genotyping of hepatitis C virus (32 ). The present codon 12 mutations with higher sensitivities than the
study demonstrates that the 12Cys sensor probe plus melt- 12Asp sensor probe. Because most of the K-ras codon 12
ing curve analysis can be used to identify at least four types mutations in our study were G-to-A transitions (GAT)
of K-ras mutations, including the frequently encountered and G-to-T transversions (GTT), both of which were
GAT and GTT mutants, with high specificity. This assay readily detected by both sensor probes, the negative
therefore possesses another advantage: that of eliminating predictive value was similar regardless of whether the
the need for expensive direct DNA sequencing. 12Cys or the 12Asp sensor probe was used. The advan-
The positive rate for K-ras mutations in bile from tage of the 12Cys sensor probe, however, may be demon-
patients with pancreaticobiliary malignancies was 35% strated in other clinical settings when larger numbers of
(16 of 46) by clamped probe assay with 12Cys sensor mutants other than GAT and GTT are present.
probe and 33% (15 of 46) by RFLP analysis. The positive The PNA oligomer used to suppress the amplification
samples were confirmed by sequencing, which showed of the wild-type DNA was designed to bind complemen-
mutations in 14 of 16 cases identified by the clamped tary sequences with higher thermal stability than DNA or
probe assay. Because we concentrated the PCR product RNA and cannot be extended by DNA polymerase (18 ).
⬃17-fold before sequencing, the sensitivity of direct se- Use of the PNA oligomer molecule, which allowed sensi-
quencing was then comparable to that of the clamped tive detection of a mutation in a minor fraction of tumor
probe assay, indicating the superior sensitivity of the cells in bile, was therefore key to our method. Our results
clamped probe assay. Our study also demonstrated that showed that the addition of 1.5 ␮M PNA oligomer com-
pletely suppressed the wild-type melting signal in 50 ng
of bile DNA and allowed detection of the mutation in a
Table 4. Frequency of codon 12 K-ras mutations in bile 3000-fold excess of wild-type bile DNA. The ratio of PNA
from patients with biliary obstructions detected by RFLP oligomer and target DNA we used was different from that
analysis and clamped probe assay. used by Sotlar et al. (22 ); they used 0.75 ␮M PNA oligomer
Clamped probe in 100 ng of template DNA to detect c-kit mutations with a
assay
detection limit of 1:2000 (22 ). This may be attributable to the
Group Patients, n RFLP 12Asp 12Cys different types of samples and target genes used in their
Pancreatic cancer 20 6 4 6 study and the present study. Furthermore, the specific
Cholangiocarcinoma 26 9 8 10 melting peak for wild type in our study was similar to that
Hepatolithiasis 17 0 0 0 for the G12V mutation; therefore, a high PNA oligomer
Choledocholithasis 47 0 0 0 concentration is needed to completely suppress the amplifi-
Benign biliary stricture 6 0 0 0
cation of wild-type DNA and thus to prevent false positives.
Total 116 15 12 16
However, increases in the PNA oligomer concentration may
488 Chen et al.: Rapid Detection of K-ras Mutations in Bile

Fig. 4. Serial dilution experiment showing the sensitivity of the clamped probe assay.
(A), DNA from cell line SW480, which harbors a homozygous K-ras codon 12 mutation (G12V), was diluted into DNA from a bile sample with wild-type (WT) K-ras. (B),
DNA from a bile sample harboring a heterozygous K-ras codon 12 mutation (G12C) was diluted into DNA from bile with wild-type K-ras.

lower the amplification of the mutant alleles, reducing the
sensitivity of the method. Thus, the ratio of PNA oligomer to
template DNA is critical for the detection of mutations with
good specificity and sensitivity. When Sun et al. (21 ) used
PNA oligomer and matrix-assisted laser-desorption/ioniza-
tion time-of-light mass spectrometry for enrichment of K-ras
and p53 mutants in the presence of excess amounts of
wild-type DNA, the detection limit was reported to be as
few as 3 mutant alleles in the presence of 10 000-fold excess
of wild-type alleles. These results indicate the potential of
this technique in high-throughput screening applications.
The clamped probe assay, however, had a diagnostic
sensitivity similar to that of RFLP analysis. It could only
detect the mutation in a 1:20 mixture of heterozygous
mutant bile DNA and wild-type bile DNA. This may be
attributable to the presence of PCR inhibitors in bile
samples. Iron-containing proteins and their breakdown
products, such as bilirubin, and bile salts have been
identified as major inhibitors in PCR of blood samples,
including real-time DNA amplification with the Light-
Cycler Instrument (33, 34 ). The underlying mechanisms
may involve the inactivation of thermostable DNA poly-
merase and/or degradation of the nucleic acids. Because
bile contains higher concentrations of bile salts and bilirubin
than blood samples, the different extents of inhibitory effects
of these substances in RFLP analysis and the clamped probe
assay may explain the different detection limits for the
SW480 cell line and mutant bile DNA. Although bovine
serum albumin was reported to be the most efficient ampli-
fication facilitator (34 ), we found no improvement in the
sensitivity with the addition of 4 g/L bovine serum
albumin in the clamped probe assay. Further studies are
needed to determine ways to minimize the inhibitory
effects of bile components in PCR and thus improve the
detection sensitivity of the clamped probe assay.
In conclusion, we present a practical method for rapid
analysis of K-ras mutations in bile. Because K-ras gene
mutations are frequently detected in other gastrointestinal
malignancies, such as colon cancer, this method may be
promising for the analysis of other body fluids (such as
stool) as a mass-screening tool for gastrointestinal cancers.
 Clinical Chemistry 50, No. 3, 2004
We thank Hsiu-Chen Chiu and Ping-Hong Chen for
excellent technical assistance. This work was supported
by National Science Council Grants NSC90-2320-B006-
056, NSC91-2323-B006-011, and NSC92-2314-B006-044.
References
1. Bishop JM. Molecular themes in oncogenesis. Cell 1991;64:235–
48.
2. Tahey M, McCormick F. A cytoplasmic protein stimulates normal
N-ras p21 GTPase, but does not affect oncogenic mutants.
Science 1987;238:542–5.
3. Muller P, Ostwald C, Puschel K, Brinkmann B, Plath F, Kroger J, et
al. Low frequency of p53 and ras mutations in bile of patients with
hepato-biliary disease: a prospective study in more than 100
patients. Eur J Clin Invest 2001;31:240 –7.
4. Saurin JC, Joly-Pharaboz MO, Pernas P, Henry L, Ponchon T,
Madjar JJ. Detection of K-ras point mutations in bile specimens for
the differential diagnosis of malignant and benign biliary stric-
tures. Gut 2000;47:357– 61.
5. Lee JG, Leung JW, Cotton PB, Layfield LJ, Mannon PJ. Diagnostic
ultility of K-ras mutational analysis on bile obtained by endoscopic
retrograde cholangiopancreatography. Gastrointest Endosc 1995;
42:317–20.
6. Uehara H, Nakaizumi A, Tatsuta M, Baba M, Takenaka A, Uedo N,
et al. Diagnosis of pancreatic cancer by detecting telomerase
activity in pancreatic juice: comparison with K-ras mutations. Am J
Gastroenterol 1999;94:2513– 8.
7. Berthelemy P, Bouisson M, Escourrou J, Vaysse N, Rumeau JL,
Pradayrol L. Identification of K-ras mutations in pancreatic juice in
the early diagnosis of pancreatic cancer. Ann Intern Med 1995;
123:188 –91.
8. Iguchi H, Sugano K, Fukayama N, Ohkura H, Sadamoto K, Ohkoshi
K, et al. Analysis of Ki-ras codon 12 mutations in the duodenal
juice of patients with pancreatic cancer. Gastroenterology 1996;
110:221– 6.
9. Caldas C, Hahn SA, Hruban RH, Redston MS, Yeo CJ, Kern SE.
Detection of K-ras mutations in the stool of patients with pancre-
atic adenocarcinoma and pancreatic ductal hyperplasia. Cancer
Res 1994;54:3568 –73.
10. Hruban RH, van Mansfeld ADM, Offenhaus GJA, van Weering DH,
Allison DC, Goodman SN, et al. K-ras oncogene activation in
adenocarcinoma of the human pancreas: a study of 82 carcino-
mas using a combination of mutant enriched polymerase chain
reaction analysis and allele specific oligonucleotide hybridization.
Am J Pathol 1993;143:545–54.
11. Momoi H, Itoh T, Nozaki Y, Arima Y, Okabe H, Satoh S, et al.
Microsatellite instability and alternative genetic pathway in intra-
hepatic cholangiocarcinoma. J Hepatol 2001;35:235– 44.
12. Hidaka E, Yanagisawa A, Seki M, Takano K, Setoguchi T, Kato Y.
High frequency of K-ras mutations in biliary duct carcinomas of
cases with a long common channel in the papilla of vater. Cancer
Res 2000;60:522– 4.
13. Kubicka S, Kuhnel F, Flemming P, Hain B, Kezmic N, Rudolph KL,
et al. K-ras mutations in the bile of patients with primary scleros-
ing cholangitis. Gut 2001;48:403– 8.
14. Sturm PD, Rauws EA, Hruban RH, Caspers E, Ramsorekh TB,
Huibregtse K, et al. Clinical value of K-ras codon 12 analysis and
endobiliary brush cytology for the diagnosis of malignant extrahe-
patic bile duct stenosis. Clin Cancer Res 1999;5:629 –35.
15. Boberg KM, Schrumpf E, Bergquist A, Broome U, Pares A, Remotti
H, et al. Cholangiocarcinoma in primary sclerosing cholangitis:
K-ras mutations and Tp53 dysfunction are implicated in the
neoplastic development. J Hepatol 2000;32:374 – 80.
489
16. Nakao M, Janssen JWG, Seriu T, Bartram CR. Rapid and reliable
detection of N-ras mutations in acute lymphoblastic leukemia by
melting curve analysis using LightCycler technology. Leukemia
2000;14:312–5.
17. Fox CA, Parkes HC. Emerging homogeneous DNA-based technol-
ogies in the clinical laboratories. Clin Chem 2001;47:990 –1000.
18. Ørum H, Nielsen PE, Egholm M, Berg RH, Burchardt O, Stanley C.
Single base pair mutation analysis by PNA directed PCR-clamping.
Nucleic Acids Res 1993;21:5332– 6.
19. Thiede C, Bayerdorffer E, Blasczyk R, Wittig B, Neubauer A. Simple
and sensitive detection of mutations in the ras proto-oncogenes
using PNA-mediated PCR clamping. Nucleic Acids Res 1996;24:
983– 4.
20. Behn M, Thiede, Neubauer A, Pankow W, Schuermann M. Facili-
tated detection of oncogene mutations from exfoliated tissue
material by a PNA-mediated “enriched PCR” protocol. J Pathol
2000;190:69 –75.
21. Sun X, Jung K, Wu L, Sidransky D, Guo B. Detection of tumor
mutations in the presence of the excess amounts of normal DNA.
Nat Biotechnol 2002;20:186 –9.
22. Sotlar K, Escribano L, Landt O, Mohrle S, Lass U, Horny HP, et al.
One-step detection of c-kit point mutations using peptide nucleic
acid-mediated polymerase chain reaction clamping and hybridiza-
tion probes. Am J Pathol 2003;162:737– 46.
23. Su WC, Shiesh SC, Liu HS, Chen CY, Chow NH, Lin XZ. Expression
of oncogene products HER2/New and Ras and fibrosis-related
growth factors bFGF, TGF-␤, and PDGF in bile from biliary malig-
nancies and inflammatory disorders. Dig Dis Sci 2001;46:1387–92.
24. Li Y, Zon G, Wilson WD. NMR and molecular modeling evidence for
a G. A. mismatch base pair in a purine-rich DNA duplex. Proc Natl
Acad Sci U S A 1991;88:26 –30.
25. Tada M, Omata M, Ohto M. Clinical application of ras gene
mutation for diagnosis of pancreatic adenocarcinoma. Gastroen-
terology 1991;100:233– 8.
26. Scarpa A, Capelli P, Villanueva A, Zamboni G, Lluis F, Accolla R, et
al. Pancreatic cancer in Europe: Ki-ras mutation pattern shows
geographical differences. Int J Cancer 1994;57:167–71.
27. Tada M, Omata M, Ohto M. High incidence of ras gene mutation in
intrahepatic cholangiocarcinoma. Cancer 1992;69:1115– 8.
28. Watanabe M, Asaka M, Tanaka J, Kurosawa M, Kasai M, Miyazaki
T. Point mutation of K-ras gene codon 12 in biliary tract tumors.
Gastroenterology 1994;107:1147–53.
29. Ohashi K, Nakajima Y, Kanehiro H, Tsutsumi M, Taki J, Aomatsu
Y, et al. Ki-ras mutations and p53 protein expression in intrahe-
patic cholangiocarcinomas: relation to gross tumor morphology.
Gastroenterology 1995;109:1612–7.
30. Inoue S, Tezel E, Nakao A. Molecular diagnosis of pancreatic
cancer. Hepatogastroenterology 2001;48:933– 8.
31. Tannapfel A, Benicke M, Katalinic A, Uhlmann D, Kockerling F,
Hauss J, et al. Frequency of p16INK4A alterations and K-ras
mutations in intrahepatic cholangiocarcinoma of the liver. Gut
2000;47:721–7.
32. Bullock GC, Bruns DE, Haverstick DM. Hepatitis C genotype
determination by melting curve analysis with a single set of
fluorescence resonance energy transfer probes. Clin Chem 2002;
48:2147–54.
33. Akane A, Matsubara K, Nakamura H, Takahashi S, Kimura K.
Identification of heme compound copurified with deoxyribonucleic
acid (DNA) from bloodstains, a major inhibitor of polymerase chain
reaction (PCR) amplification. J Forensic Sci 1994;39:362–72.
34. Abu Al-Soud W, Radstrom P. Purification of characterization of
PCR-inhibitory components in blood cells. J Clin Microbiol 2001;
39:485–93.