| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
CRLC Val d’Aurelle Paul-Lamarque, Centre de Recherche en Cancérologie, Parc Euromédecine, 34298 Montpellier Cedex 5, France.
2
CIS bio international, 91192 Gif sur Yvette Cedex,
France.
a Author for correspondence. Fax 33-4-67-63-28-73; e-mail
ecrapez{at}valdorel.fnclcc.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: An oligodeoxynucleotide array was constructed on a silicon device by copolymerization of 5'-pyrrole-labeled oligodeoxynucleotides and pyrrole. The samples to be analyzed were then amplified by PCR, and the single-stranded biotin-labeled amplified DNA was specifically hybridized to the addressed probes. Perfectly matched duplexes were detected by fluorescence microscopy using R-phycoerythrin as the detection label. The developed methodology was applied to genotype assignment of K-ras in human samples. The genotypes of 75 DNA genomic samples from colorectal cancer patients were analyzed side by side using direct DNA sequencing and a polypyrrole DNA chip.
Results: The chip method unequivocally defined all of the genotypes. Mutations present at <10% of the wild-type DNA concentration could be distinguished.
Conclusions: This probe array assay is a rapid and reliable procedure that may be used to detect mutations.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The K-ras protooncogene is altered by point mutations on codon 12, 13, or 61 in a wide variety of tumors (12). The detection of K-ras mutations enables understanding of cancer biology and pathogenesis. Pancreatic and lung cancers harbor high incidences of K-ras mutant alleles, and these mutations are early events in colorectal tumor development (13). Moreover, alterations involving this oncogene may be of clinical importance because they can provide information for early diagnosis and prognosis (14)(15)(16). Activation of the K-ras gene has been detected not only on the tumor, but also in the stools (17) and serum (18) of patients with colorectal cancer. PCR protocols have been described for the diagnosis of pancreatic adenocarcinoma from the presence of mutant K-ras DNA in pancreatic juice (19); therefore, according to the clinical implication of this gene in human tumorigenesis and its potential role as a target for novel therapeutic approaches, reliable methods are needed for the analysis of the K-ras sequence in clinical samples.
Recently, we developed a rapid, large-scale method to detect K-ras gene mutations in tumor samples (20). This approach is based on the use of biotinylated oligonucleotide probes fixed to avidin-coated tubes and a radioactive detection method. In this technology, each tube corresponds to a specific K-ras mutant allele, and to determine the K-ras codon 12 sequence, seven tubes are necessary (one tube for the wild-type sequence, three tubes corresponding to replacement of the first base of the codon, and three other tubes for the mutation on the second base of the codon).
To eliminate the use of radioactive labels and to allow the simultaneous analysis of the seven K-ras codon 12 alleles in a single hybridization, an oligonucleotide array on a silicon device allowing fluorescence detection has been developed. The construction of this DNA chip is based on the electropolymerization of 5'-pyrrole-labeled oligonucleotides and pyrrole. Here, we report the technical development and use of this assay for the detection of K-ras mutations in DNA from human colorectal carcinomas.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
GTT(Val)], pancreatic carcinoma cell
line MIA PaCa-2 [homozygous, GGT(Gly)TGT(Cys)], and colon
adenocarcinoma cell lines SW1116 [heterozygous, GGT(Gly)GCT(Ala)]
and LS 174T [heterozygous, GGT(Gly)GAT(Asp)], which display a
mutation on the K-ras codon 12 sequence, and the lymph node
metastasis of prostate carcinoma cell line LNCaP (wild-type
K-ras codon 12 sequence) were purchased from the American
Type Culture Collection.
Colorectal samples
Samples were obtained during surgery from 75 patients with
colorectal cancer (CRLC Val d’Aurelle, Montpellier, France).
Colorectal carcinoma and nontumoral mucosa were resected from
each patient. According to the Dukes classification, the tumors were
staged from A to D (3 stage A, 31 stage B, 18 stage C, and 23 stage D).
Fresh samples were immediately frozen in liquid
N2 before subsequent nucleic acid extraction.
DNA extraction
High-molecular weight DNA from the cell lines and frozen biopsies
was prepared with proteinase K digestion and phenol-chloroform
extraction as described previously (21).
Oligonucleotides
The oligonucleotide primers and probes were synthesized on
an Applied Biosystems 381 A DNA synthesizer and purified by
reversed-phase HPLC (CIS bio international). For the amplification
step, one primer was 5'-labeled with a biotin residue (22),
and the other was 5'-labeled with a phosphate residue. Eight
oligonucleotides perfectly matched to the oligonucleotides grafted onto
the microelectrodes were labeled at their 5' end with a biotin residue.
Probes to be addressed were 5'-labeled with a pyrrole residue
according to a previously reported procedure using pyrrole
phosphoramidite building blocks in the course of oligonucleotide
synthesis and a d(T)10 oligonucleotide linker
(11).
DNA sequencing
Sequence determination was done on amplified (MD1/MD2) DNA
using [
33-P]dATP in the
dideoxy-chain-termination method as described previously
(20).
Silicon chip
The chips were made using microelectronic technologies on a
silicon support (CEA/LETI) (23).
Oligonucleotide addressing
The electropolymerization was carried out as described previously
(6). Briefly, each of the 128 microelectrodes (50 µm) was
successively switched on, and the electropolymerization was performed
by cyclic voltamperometry (-0.35 to +0.85V/saturated calomel
electrode) in an aqueous 0.1 mol/L LiClO4
solution containing 20 mmol/L pyrrole and 1 µmol/L of the pyrrole
oligonucleotide to be addressed. At the end of the process, the matrix
was rinsed with water and stored at 4 °C until needed.
Detection of K-ras mutations
on the chip
Generation of the target.
The DNA was amplified by PCR. Three
primers were used for the different PCR reactions: P3ras1 (5'-GGC CTG
CTG AAA ATG ACT GAA TAT-3'), P-P3ras1, (5'phosphate-P3ras1), and
Bio-P3ras2 (5'biotin-TGT TGG ATC ATA TTC GTC CAC AAA ATG-3'). The PCR
was carried out in a total volume of 50 µL containing the following:
250 ng of extracted DNA, 200 µM dNTP, 2 U of Taq DNA polymerase
(Perkin-Elmer), 10 mM Tris-HCl (pH 8.3), 1.5 mM
MgCl2, 50 mM KCl, and 0.1 g/L gelatin. The PCR
was performed in a Touch Down Hybaid thermal cycling system with 32
cycles (94 °C for 30 s, 55 °C for 30 s, and 72 °C
for 30 s). The symmetric PCR was carried out with 12 pmol of each
PCR primer (P-P3ras1 and Bio-P3ras2). The asymmetric PCR
(P3ras1/Bio-P3ras2) was performed as described above, but with 1
pmol of P3ras1 and 50 pmol of Bio-P3ras2. Experimental conditions to
avoid PCR product carryover were applied (24).
Specific hybridization of the target.
Hybridization reactions
were carried out in the microreaction chamber (4 x 5 x 1
mm) of the chip created by the packaging. All reactions were performed
in a temperature-controlled incubator. A prehybridization step was
first conducted for 15 min at 50 °C in a buffer containing the
following: 1.8 mol/L tetramethylammonium chloride, 50 mmol/L Tris-HCl
(pH 7.4), 2 mmol/L EDTA, 1 g/L sodium dodecyl sulfate,
and 5x Denhart’s solution. Targets (5 fmol of 5'-biotinylated
complementary oligonucleotide or 4 µL of PCR products) were then
hybridized in a final volume of 10 µL in the above-described buffer
for 30 min at 50 °C. Before the symmetric (P-P3ras1/Bio-P3ras2)
amplicons were used, the product was digested with lambda exonuclease
(Amersham, Pharmacia-Biotech), whereas the asymmetric
(P3ras1/Bio-P3ras2) amplicons were directly hybridized. Two
posthybridization washes were carried out, the first wash in 2x
saline-sodium phosphate-EDTA containing 1.0 g/L sodium dodecyl
sulfate at room temperature, followed by a second wash in 5x
saline-sodium phosphate-EDTA containing 1.0 g/L sodium dodecyl sulfate
at 50 °C.
Acquisition and analysis of the fluorescent signals.
The chip
was rinsed in a detection buffer containing phosphate-buffered saline,
0.5 mol/L NaCl, and 5 mL/L Tween 20. The biotin moiety was
identified by incubation in a 50 mg/L solution of
streptavidin-R-phycoerythrin (Molecular Probes) in detection buffer.
This reaction was done at room temperature and in the dark. The DNA
chip was then rinsed with the detection buffer and placed between a
microscope slide and a microcover glass without the necessity of an
additional step. A fluorescence microscope (BH-2; Olympus) was used to
observe the fluorescence emission. The phycoerythrin was excited by a
mercury lamp, and fluorescent signals were obtained by accumulating the
photons for 1 s on a chilled CCD camera (Hamamatsu). Image
analysis software (Morphostar; Imstar) was used. For each
microelectrode, the mean and the mode of the fluorescent signals were
recorded. After detection, the chip was regenerated by a 1-min
denaturation step in 0.1 mol/L NaOH.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
) was a 10-mm2 active multiplexed device
bearing 128 octagonal electrodes with nine gold inputs/outputs. The
50-µm wide microelectrodes were arranged in a rectangular matrix of 8
rows and 16 columns. Each electrode had its own unique address and
could be selected individually. The silicon chip was integrated in a
package (Fig. 1B
) compatible with both the electrocopolymerization step
and the biological reactions. The copolymerization process allowed the
formation of a homogeneous film (20 nm thick). The amount of
hybridizable oligonucleotides grafted onto each electrode was estimated
to be 200 fmol/mm2 (6). To determine
the K-ras codon 12 sequence, a K-ras silicon chip
was designed (Fig. 2
) with a set of seven mutation-specific oligonucleotides (Table 1
) addressed to seven different gold microelectrodes.
|
|
|
Preliminary studies on oligonucleotide hybridization
The main steps of the process are described in Fig. 3
. The assay was based on the allele-specific hybridization
principle, and the aim was to find the hybridization/washing procedure
that allowed the detection of all seven possible sequences on codon 12
under the same operating conditions. Biotinylated oligonucleotides that
were fully matched to those attached to the solid support were used to
optimize the reaction conditions. Twenty-base-long oligonucleotides,
the sequence that had been optimized previously for the detection of
K-ras mutations on a tube format assay based on radioactive
detection (mutation tube assay test) (20), were
copolymerized to the gold electrodes (data not shown). A reduction of
the length of the screening oligonucleotides (probes W and M1–M6) to
14 bases was then made to increase the specificity. Hybridization of a
particular oligonucleotide under mild conditions [low temperature
(37 °C) and high salt concentration (0.5 mol/L
Na+)] produced substantial nonspecific binding
in addition to the signal produced by annealing to the fully
match sequence (Fig. 4A
). A correlation was found between the amount of
fluorescence and the number of mismatches in the duplexes formed. The positive
control, the sequence of which is outside the codon 12 region,
produced no hybridization signal regardless of the biotinylated probe
(M1–M6) used. A 16-base-long sequence for the positive control was
kept to obtain a high fluorescent signal that validated the entire
process, including PCR, hybridization, and detection. No nonspecific
binding of the biotinylated oligonucleotide was observed on the
nonaddressed gold microelectrode (Bg electrode) or on the negative
control electrode, which was composed of a homopolymer of pyrrole (PP
electrode). Better discrimination between codon 12 sequences was
obtained when the hybridization temperature was increased from 37 °C
to 50 °C, but an increase in temperature did not permit the
detection of single-nucleotide mismatches. Specific hybridization was
obtained with the introduction in the hybridization buffer of
tetramethylammonium chloride at a final concentration of 1.8 mol/L
(Fig. 4B
). Under these hybridization conditions, (i.e., combining high
temperature with an additive), a clear discrimination was made between
a fully matched hybrid and a 1-bp mismatched hybrid.
|
|
The detection limit was determined under mild hybridization conditions.
We determined this value as the amount of biotinylated N probe allowing
a mode fluorescence signal on the N electrode equal to twice the mode
value on the PP electrode. Within the above-mentioned conditions, 0.1
fmol of N probe could be detected, corresponding to a detection limit
of 
10 pmol/L.
Characterization of the assay with cell lines
To evaluate the ability of the methodology to analyze PCR
products, amplifications were performed on genomic DNA from cell lines.
A 117-bp DNA fragment was produced, and its purity was checked by
agarose gel electrophoresis. Two ways of obtaining single-stranded DNA
were evaluated: (a) asymmetric PCR using a 1:50 ratio of
amplification primers, and (b) symmetric PCR using a
5'-phosphorylated primer, followed by digestion with lambda
exonuclease. The hybridization results obtained for symmetric digested
product were better in terms of fluorescence intensity than those
obtained for the asymmetric PCR (data not shown). The operating
conditions optimized on the basis of oligonucleotide hybridization
permitted detection of homozygous (MIA-PaCa-2 and SW620 cell lines), as
well as heterozygous (LS174T and SW1116 cell lines) DNA. The
sensitivity of the method was investigated by reconstructing samples
with various dilutions of mutant DNA (MIA-PaCa-2 cell line) in
wild-type DNA (LNCaP cell line). Samples were amplified through
independent PCR reactions, and the amplicons were mixed in different
ratios. The generated symmetric amplicons were then digested with
lambda exonuclease and hybridized to the chip. A K-ras
mutated sequence was unambiguously detected even when the MIA-PaCa-2
cell line represented only 10% of the initial material: a mode value
of 42 fluorescence units (FU) on the M3 electrode (Table 2
). The background value was 23 FU.
|
Genotyping human colorectal dna samples
To demonstrate the reliability of the procedure, a blind study was
conducted on DNA from 75 patients with colorectal cancer. Table 3
shows the characteristics of human tumoral samples in terms of
exon 1 (codons 12 and 13) K-ras sequences obtained by direct
PCR-product sequencing. Thirty-two percent of patients (24 of 75) had
an exon 1 K-ras mutation, and 88% (21 of
24) of the mutations were codon 12 alterations. The only
K-ras codon 13 mutated allele found corresponded to
GGC(Gly)GAC(Asp). The genotypes scored by the polypyrrole DNA chip
assay were 100% in agreement with conventional DNA sequencing results.
All of the K-ras codon 12 mutations detected by direct
sequencing were positive by polypyrrole DNA chip analysis, and no
false-positive result was obtained with the developed methodology.
Moreover, three patients displayed a K-ras codon 13 mutation
detected by direct sequencing, and no nonspecific hybridization signal
was observed on the electrodes corresponding to the K-ras
codon 12 mutated sequences. Fig. 5
shows an example of the data obtained from direct sequencing
and DNA chip analysis for a wild-type (GGT) and a mutated (GG/AT) DNA.
In both samples, analysis of all controls gave the expected results.
Particularly, the positive control probe gave a strong positive signal,
whereas the PP control was negative. Moreover, no background was
observed on the gold nonpolymerized microelectrode Bg. The genotype
assignment of K-ras codon 12 by DNA chip analysis
distinguished homozygous and heterozygous sequences unambiguously.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The reliability of hybridization on ODN chips depends largely on both the quality of the surface-bound oligomer and the type of linkage between the ODN and the solid support. With the methodology described here, the ODNs were characterized and HPLC-purified before grafting. The functionalization of the gold electrodes by electrocopolymerization allows the formation of a controlled thin film of ODN-pyrrole that is very stable. This feature makes it possible to carry out hybridization and washing steps under stringent conditions such as the use of tetramethylammonium chloride and high temperatures. These conditions were neutral toward the polymer film and had no side effects on the packaging compounds. The stability of the ODN-polypyrrole surface is also compatible with NaOH, and a regeneration of the ODN chip after the analysis (hybridization, washing, and detection) is completed is possible without substantial loss of signal. The 5'-end d(T)10 linker makes it possible to increase the hybridization signal by improving the ODN accessibility. Moreover, the 3' end of the grafted ODN is free, and extension by polymerases of the fixed ODN could be envisaged.
With regard to the geometry of the packaging, the "T"-shaped chip was easy to handle, and the microreaction chamber created was suitable for the hybridization and detection steps. Moreover, all of the steps in the preparation of the functionalized biochip are automated.
The binding and accessibility characteristics of the silicon-grafted ODNs were similar to those observed for nylon membranes because the optimal conditions that allow perfect discrimination of mismatched duplexes were similar. The use of tetramethylammonium chloride, which acts on the melting temperature (30), is especially efficient in equalizing duplex stabilities on both supports.
The methodology described here does not require multiple steps subsequent to PCR, i.e., purification or post labeling, to obtain reliable biological material. The amplicons to be analyzed are relatively short and are able to reassociate rapidly after heat denaturation; therefore, generation of single-stranded PCR products is important to increase hybridization efficiency with the solid-phase fixed probes. Several approaches for generating single-stranded DNA have been described. Use of asymmetric PCR (31) is not the best way because only relatively small amounts of single-stranded DNA are produced under extensive optimization conditions. The use of T7 gene 6 exonuclease, which digests 5'-phosphorylated and 5'-hydroxylated DNA and is blocked by phosphorothioate (S) linkages, has been reported previously (32). Nevertheless, in our developed technology this method requires modifications in the strand of interest by incorporation of sulfur bridges in one of the amplification primers. Lambda exonuclease (33) is able to convert double-stranded amplified DNA to single strands and to enhance specific hybridization. In our assay, this alternative is attractive because (a) the modification (5' phosphate residue) is not in the hybridizable strand, and (b) the production of single strand is very efficient; moreover, this approach can be applied to multiplex PCR.
The allele-specific mutation detection method on the chip using the same process steps over the entire array was rapid. Starting from genomic DNA, the detection of point mutations can be accomplished in less than 4 h. We are researching ways to develop an automatic station for the hybridization and analysis steps.
The developed DNA chip based on electrocopolymerization on gold
electrodes gives a support with low autofluorescence, allowing the
detection of <10% of mutated sequences within a mixed
mutant/wild-type population. The sensitivity of direct sequencing
performed with radioactive incorporation was 15%, whereas
it was 5% for our previously reported mutation tube assay test
(20) using tubes as solid support and serial hybridizations.
To enable detection of all K-ras mutations, we are developing methods to screen codons 13 and 61 by starting with a multiplex PCR. Moreover, because 128 electrodes are available, we can expect to analyze other genes of interest whose mutations are implicated in the development of cancer. The unambiguous discrimination among the seven potential alleles in the same region suggests that this technology may be useful in multiallelic gene analysis. Finally, a quantitative application of the silicon device we have developed is conceivable because the amount of the grafted ODN is fully controlled.
This work was supported in part by the Ministère de l’Industrie, France (Grant No. 95.493.0183), and by support from la Ligue Départementale contre le Cancer, France. We thank Bernard Saint Aubert, Marc Ychou, and Pierre Senesse for providing colorectal tissues and Joëlle Simony-Lafontaine for the anatomopathological analysis of biopsies. We thank Dr. P. Caillat (CEA/LETI, Grenoble, France) for providing silicon devices, and we gratefully acknowledge the helpful support of the LETI/CEA. The quality of the technical work of Josette Noletti during the entire study was appreciated.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  A. Halperin, A. Buhot, and E. B. Zhulina Hybridization Isotherms of DNA Microarrays and the Quantification of Mutation Studies Clin. Chem., December 1, 2004; 50(12): 2254 - 2262. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-H. Park, I.-J. Kim, H. C. Kang, Y. Shin, H.-W. Park, S.-G. Jang, J.-L. Ku, S.-B. Lim, S.-Y. Jeong, and J.-G. Park Oligonucleotide Microarray-Based Mutation Detection of the K-ras Gene in Colorectal Cancers with Use of Competitive DNA Hybridization Clin. Chem., September 1, 2004; 50(9): 1688 - 1691. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Halperin, A. Buhot, and E.B. Zhulina Sensitivity, Specificity, and the Hybridization Isotherms of DNA Chips Biophys. J., February 1, 2004; 86(2): 718 - 730. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Prix, P. Uciechowski, B. Bockmann, M. Giesing, and A. J. Schuetz Diagnostic Biochip Array for Fast and Sensitive Detection of K-ras Mutations in Stool Clin. Chem., March 1, 2002; 48(3): 428 - 435. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I.-J. Kim, H. C. Kang, J.-H. Park, J.-L. Ku, J.-S. Lee, H.-J. Kwon, K.-A. Yoon, S. C. Heo, H.-Y. Yang, B. Y. Cho, et al. RET Oligonucleotide Microarray for the Detection of RET Mutations in Multiple Endocrine Neoplasia Type 2 Syndromes Clin. Cancer Res., February 1, 2002; 8(2): 457 - 463. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
),
the other phosphate-labeled (P)]. Seven different
oligonucleotide probes corresponding to wild-type or mutated codon 12
sequences are addressed to individual gold microelectrodes by
electrocopolymerization through their pyrrole extremities (•). The
chip is then hybridized to a fraction of the single-stranded DNA
obtained by lambda exonuclease digestion of the PCR product and washed.
DNA duplexes are detected by streptavidin-R-phycoerythrin
(
),
labeling. The fluorescence emission is then analyzed through the use of
a CCD camera connected to image software
(
excitation = 500/550 nm;
