Note: Descriptions are shown in the official language in which they were submitted.
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
METHOD FOR IDENTIFYING MISMATCH REPAIR GLYCOSYLASE REACTIVE
SITES, COMPOUNDS AND USES THEREOF.
STATEMENT AS TO FEDERALLY SPONSORED RESEARCH
This invention was made with Government support under grants R29
CA63334, K04 CA69296, and RO1 CA72046, awarded by the USPHS. The
Government may have certain rights in the invention.
FIELD OF THE INVENTION
The present invention is directed to methods for identifying and
labeling glycosylase-recognizable sites on nucleic acids, and to novel
compounds that bind to such glycosylase-recognizable sites on nucleic
acids. In a preferred embodiment the method can be used to identify
mutations and/or polymorphisms on a nucleic acid segment, or in an
arbitrary mixture of nucleic acid segments or genes.
BACKGROUND OF THE INVENTION
The detection of mutations has been an area of great interest in recent
years. For example, mutations in certain genes have been associated with a
variety of disorders - ranging from blood disorders to cancers. Genetic tests
are thus becoming an increasingly important facet of medical care.
Consequently, there has been an emphasis on the ability to rapidly and
efficiently detect mutations and polymorphisms.
Many electrophoreti.c techniques have been developed to rapidly
screen DNAs for sequence differences by which such mutations can be
detected. Denaturing Gradient Get Electrophoresis (DGGE) [Myers, R.M.,
Maniatis, T. and Lerman, L., Methods in Enzymology, 155, 501-527 (1987)J,
Constant Denaturant Gel Electrophoresis (CDGE) [Borresen, A.L., et al.,
Proc. Nat. Acad. Sci. USA, 88, 8405 (1991)], Single Strand Conformation
Polymorphism (SSCP) [Orrita, M., et al., Proc. Nat. Acad. Sci. USA, 86,
2766-2770 (1989)], Heteroduplex Analysis (HA) [Nagamine, C.M., et al., Am.
-1-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
J. Hum, Genet., 45,377-399 ( 19?9)] and Protein Truncation Test (PTT) [Roest,
P.A.M., et al., Hum. Molec. Genet., 2,1719-1721 (1993)] are frequently used
methods. Many labs use combinations of these methods to maximize
mutation detection efficiency. All these methods require gel electrophoresis.
Methods that do not require gel electrophoresis also exist. For example,
selective hybridization on immobilized target sequences allows screening for
rare known mutations [Zafiropoulos, A., et al., Biotechniques 223,
1104-1109 (1997)], while mass-spectrometry has been used to detect
mutations by analyzing molecular weight of proteins [Lewis, J.K., et al.,
Biotechniques 24, 102-110 ( 1998)].
A fundamental problem with currently existing mutation and
polymorphism detection methods is that they only screen for mutations on a
single gene at a time (i.e. the method is directed to looking at a 'gene of
interest', that is suspected of having a mutation). Given that the human
genome has 50,000-100,000 genes, this is a severe limitation. It is likely
that unknown mutations and polymorphisms in several other genes both
known and unknown, exist simultaneously with mutations/polymorphisms
in the 'gene of interest. However, mutations in those other genes would
likely not be identified. Therefore a method that can perform
'mutation/polymorphism scanning' in for a wide array of genes
simultaneously, without the initial need far identifying the gene one is
screening would be useful. Gel-electrophoresis - based methods are
essentially restricted to examining mutations in a single gene at a time.
Attempts have been made to devise non-gel electrophoretic methods to
identify mutations, that would not be restricted to a single gene [Cotton et
al., Proc. Natl. Acad. Sci. USA vol. 85, pp 4397-4401, ( 1988)] [Nelson, S.F.
et
al., Nature Genetics, 4, 11-8, (1993 May)] [Modrich, P., et al., Methods for
Mapping Genetic Mutations. US PATENT 5459039, (1995)]. These methods,
however, have had limited success [Nollau P and Wagener C., Clinical
Chemistry 43: 1114-1128 ( 199?)] since they are complicated, typically
requiring several enzymatic steps and they result in a large number of false
positives, i.e. they frequently score mutations and polymorphisms in normal
DNA. It would be desirable to have a method that allows highly sensitive
and specific identification and rapid purification of sites that contain
mutation/polymorphism over large spans of the genome.
-2-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
Although DNA arrays and methodologies that can simultaneously
scan a large set of DNA fragments for gene expression (e.g. the 'repertoire'
and amount of genes expressed in normal vs. cancer cells) are known
[Wodicka L, Nature Biotechnology 15:1359-1367 (1997); Lockhart, DJ,
Nature Biotechnology 14: 1675-1680 (1996).; Schena, M., Trends Biotecnnol
16:301-306, (1998); Yang, T.T., Biotechniques 18:498-503, (1995)J, the
ability to scan a large set of random DNA fragments for unknown mutations
is a much more demanding process on which the technology is lagging
[Ginot F., Human Mutation 10: 1-10 ( 1997)]. Thus far DNA array - based
methods to scan for polymorphisms (SNPs) and mutations has been
restricted to specific genes [Lipshutz, R.J., Biotechniques 19: 442-447
(1995);
Wang, D.G., Science 280: 1077-1082 (1998)]. Whereas detection of
unknown mutations over several genes requires a selectivity and sensitivity
not currently achievable by present arrays [Ginot F., Human Mutation 10: 1-
10 (1997)]. For example, when it comes to unknown mutation detection,
even a single gene with a coding sequence of the size of APC (8.5 kb) is
difficult to screen in a single experiment, especially when an excess normal
alleles is simultaneously present [Sidransky D., Science 278: 1054-1058
(1997)]. A method that permits identification of mismatches over large
spans of the genome would be desirable.
The process of mismatch repair of nucleic acids has also received
considerable attention in recent years with the elucidation of systems in
microorganisms such as E. coli, and more recently, mammals including
humans. For example, continuous cellular damages occur to nucleic acids
during the cell life cycle; for example damage resulting from exposure to
radiation, or to alkylating and oxidative agents, spontaneous hydrolysis and
errors during replication. Such damages must be repaired prior to cell
division. There are a number of different cellular repair systems and a
variety of components that participate in these systems. One component is
represented by the class of DNA repair enzymes known as mismatch repair
glycosylases. These enzymes convert mismatches in DNA to aldehyde-
containing abasic sites. These abasic sites can also occur by other means.
For example, they can occur spontaneously, or following deamination of
cytosine to uracil and subsequent removal of uracil by uracil glycosylase
[Lindahl and Myberg, 1972; Lindahl, 1982 & 1994; Demple and Harrison,
-3-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
1994; von Sonntag 1987; Loeb and Preston, 1986]. It has been estimated
that almost 10,000 abasic sites are generated per cell per day [Lindahl and
Nyberg, 1972]. Finally abasic sites are generated by DNA damaging agents
such as ionizing radiation [von Sonntag, 1987), reactive oxygen
intermediates [Ljungman and Hanawalt, 1992; Lindahl, 1994], antibiotics
[bleomycin-iron complexes, neocarzinostatin, Povirk and Houlgrave, 1988],
or alkylation agents [rnethylmethanesulfonate, dimethylsulfate etc., Loeb
and Preston-1986]. Unrepaired abasic sites can be lethal or promutagenic
lesions since during DNA replication DNA polymerases insert primarily
adenines opposite them [Kunkel et al. - 1983; Loeb and Preston - 1986].
Closely - spaced abasic sites generated within a few base pairs of each other
by damaging agents may be a particularly significant set of lesions, as they
may hinder repair [Chaudhry and Weinfeld, 1995a, 1997; Harrison et al.,
1998], or they can be enzymatically converted to double strand breaks or
other complex multiply - damaged sites [Dianov et al., 1991]. It has been
postulated that such complex forms of DNA damage may be particularly
difficult for cells to overcome [Ward 1985, 1988; Wailace, 1988; Goodhead,
1994; Chaudhry and Weinfeld, 1995a and b, and 199?; Rydberg, 1996;
Hodgkins et al., 1996; Nikjoo et al., 1998; Harrison et al., 1998].
Quantification of the overall number of abasic sites directed to looking at
abasic sites resulting from DNA damage has been reported [Futcher and
Morgan, 1979; Talpaert-Borle and Liuzzi, 1983; Weinfeld and Soderlind,
1991; Ide et al., 1993; Chen et al., 1992; Kubo et al., 1992]. The binding
efficiency of such systems has been relatively low.
SUMMARY OF INVENTION
We have now discovered a method that permits the rapid
identification of mutations in a DNA segment or in any mixture of DNA
segments (genes). This method comprises identifying mismatches that occur
when a target nucleic acid strand is hybridized to a control nucleic acid
sequence. The method comprises (a) isolating the nucleic acid, e.g., DNA, to
be screened for mutations (referred to as the target DNA), and hybridizing it
with control DNA, to create mismatches. Preferably the nucleic acid has
been digested so that it is about 50-500, more preferably 50-300 base pairs.
These mismatches occur at the exact positions of mutations or
-4-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
polymorphisms; in one embodiment PCR primers can be added in order to
subsequently amplify the mismatch containing fragments (b) removing any
pre-existing, spontaneous aldehydes by, for example, treating the DNA with
hydroxylamine; (c) using mismatch repair glycosylase enzymes (Mutt and
TDG) to convert the mismatches to reactive sites, namely, aldehyde-
containing abasic sites (these enzymes recognize mismatches and will 'cut'
the nucleic acid base, e.g., adenine at that site to create a reactive site);
(d)
using compounds (e.g. ligands) with functional groups that at one site can
covalently bind to the reactive sites on the DNA, and that at a second site
contain unique moieties that can be detected; (e) binding antibodies or
avidin to the detectable second sites of the DNA-bound ligands. These
antibodies or avidin may carry chemiluminescent or other indicators, so that
the total reactive sites on the nucleic acid, e.g., DNA segments) tested is
quantified, e.g. by chemiluminescence; (f) purifying the segments where a
reactive site is present (e.g. by immunoprecipitation, or by ELISA-microplate-
based techniques, or by microsphere-based techniques). The rest of the
nucleic acid, e.g., DNA that does not contain mismatches can then be
discarded; or (f~ directly using the sample containing mismatches and non-
mismatches; (g) in one embodiment (1) amplifying the remaining, mismatch
containing nucleic acid, e.g., DNA, by PCR using the primers added in the
first step; and (h) analyzing that - ( 1) purified nucleic acid, e.g., DNA by
standard; or (h~ analyzing the chip for labeled fragment gene-detection
methods (e.g., hybridization) containing the sample from (f') without PCR
amplification in order to find which gene each identified mismatch belongs
to. Thereafter, by known techniques determining whether that mismatch is
a mutation that either causes the disorder or is associated with the disorder
or simply an allelic variation, i.e. a polymorphism.
More specifically, the present invention permits biochemical
approaches for chemically modifying mutations in a target nucleic acid
sequence. The mutations are converted to mismatches following
hybridization with control nucleic acid sequence. The mismatches in the
hybrid nucleic acid, e.g. DNA can then be converted to aldehydes by
mismatch repair enzymes, covalently bound by a ligand molecule, and then
identified by a detectable moiety. Subsequently the mismatch-containing
-5-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
DNA can be purified by known means such as irnmunoprecipitation and the
mutation-containing genes detected.
The target nucleic acid can be cDNA or genomic DNA. For example,
the DNA can be any mixture containing one or various sizes of DNA, such as
cDNA synthesized from the whole mRNA collected from cells that need to be
screened for mutation/polymorphisrn; or fractions thereof; or the whole
genomic DNA collected from cells that need to be screened for
mutation/polymorphism; or fractions thereof; or any combination of the
above digested into smaller pieces by enzymes. The use of cDNA is
preferable.
The control will be a wild-type DNA fraction similar to the target
nucleic acid. This wild-type DNA likely will have no mutations. In some
instances the control DNA will be from a corresponding cell from the same
individual not displaying the abnormality being screened for. In many cases
the control DNA will be from a corresponding cell from a different individual
than the target nucleic acid is from. And in other cases differences within
the two alleles in a single cell type will be screened, one allele acting as a
control and the second allele acting as target DNA.
The target nucleic acid such as DNA is mixed and hybridized with
wild-type DNA to create mismatches at the positions of differences, which
are expected to be mutations/polymorphisms. In one embodiment generic
PCR primers are added to the nucleic acids, in order to amplify the
preparation at a later stage. The mismatches are then recognized and
converted to aldehyde-containing reactive sites by enzymes such as a
glycosylase mismatch repair enzyme such as the E. coli Mutt, or the thymine
DNA glycosylase (TDG) from HeLa, cells or from E. coli. A unique feature of
these enzymes is that they are highly specific, i.e, they act only on
mismatches while they leave non-mismatch containing DNA completely
intact.
These reactive sites are identified by using a compound containing an
aldehyde - binding moiety such as -O-NH2 (-hydroxylamine), or -NHNH2 (-
hydrazine) or -NH2 (-amine) and also having a second moiety that reacts
with a detectable entity (e.g. fluorescein, biotin, digoxigenin, which
respectively react with antifluorescein antibody, avidin, and antidigoxigenin
antibody. The antibodies may have chemiluminescence tags on them and
-6-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
thereby are detected) . A unique feature of the present approach is that the
aldehyde - binding moiety binds covalently to the enzyme-generated reactive
sites. Combined with the specificity of the mismatch - repair enzymes, the
use of covalently bound ligands to the position of mutations results in a
sensitivity and specificity which is unparalleled by other methods for
detection of mutations and polymorphisms.
The bi-functional compounds that bind covalently reactive sites have
the general formula:
X-Z-Y,
wherein X is a detectable moiety, preferably X is NH2, SH, NHNH2, a
fluorescein derivative, a hydroxycoumarin derivative, a rhodamine derivative,
a BODIPY derivative, a digoxigenin derivative or a biotin derivative;
Y is NHNH2, O-NH2 or NH2, preferably Y is O-NH2; and
Z is a hydrocarbon, alkylhydroxyl, alkylethoxy, alkylester, alkylether,
alkylamide or alkylamine. The hydrocarbon chain of Z may contain a
cleavable group (e.g. an S-S disulfide bond). Z may also be substituted or
unsubstituted. The reactive groups, X and Y, are used for covalent binding
to the resulting aldehydes of damaged DNA (Y) and detection by a detecting
group (X).
We have also found a method that permits one to overcome
resolutions and other limitations existing in current DNA chip technology
and utilize the existing DNA chip technology for mutations scanning over
hundreds or thousands of genes simultaneously. In one embodiment this
method comprises first identifying a DNA segment containing a mismatch.
Those mismatches can either be caused by a single nucleotide polymorphism
(SNP) or by a base substitution mutation. Thereafter, one selects a DNA
segment of from about 50-300 nucleotides containing a mismatch. Those
DNA segments can be amplified by PCR and then screened on the DNA chip.
Alternatively, one can take th sample treated the X-Z-Y compounds, which
creates the DNA containing labeled aldehydryde sites where a mismatch is
present, denature the DNA fragments, and directly apply the sample on the
DNA chips. The chip is washed to remove unhydridized DNA or unbound
label, e.g. FARP. The DNA chip is then scanned for the label via the
3S appropriate device. For example, where fluorescence is being scanned, a
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
scanning laser is used. Those elements that display the label, e.g.
fluorescence, correspond to gene fragments containing a mismatch such as
a mutation.
Accordingly, by these methods instead of selecting a single gene at a
time and examining whether it contains mutations, the present methodology
first scans DNA to identify and isolate mismatch-containing and thereby
mutation-containing DNA fragments (genotypic selection), and then
determines which genes these DNA fragments belong to, by using available
DNA arrays. Thus, the search for mutations is transformed to the easier
task of searching for genes on a DNA array to identify on which gene and
gene segment the mismatch occurs. Accordingly, DNA arrays currently used
for multiplexed gene expression scanning [Wodicka L, Nature Biotechnology
15: 1359-1367 ( 1997); Lockhart, DJ, Nature Biotechnology 14: 1675-1680
(1996).; Schena, M., ?5~ends Biotecnnol 16:301-306, (1998); Yang, T.T.,
I S Biotechniques 18:498-503, ( 1995)J can be used directly or with minor
modifications known to the artisan based upon this disclosure to scan for
mutation.
A preferred embodiment of the invention has a general formula;
X~....~CH2~CH2_W~CH2~Y,
n 'n '
wherein X' is NHNH2 or NH2, preferably NH2;
Y' is O-NH2 or NH2, preferably O-NH2;
W is -NHC(O)-, -NHC(OH)-, -C(OH)-, -NH-, C-O-, -O-, -S-, -S-S-, -
OC(O)-, or C(O)O-;
n is and integer from 0 to 12, preferably 4-7 and more preferably 6;
n' is an integer from 0 to 12, preferably 4-7, and more preferable 6,
and
n" is an integer from 1 to 4, preferably 1-2, and more preferably 1.
Preferably, the compound has a molecular weight between 100 - 500, more
preferably 100 - 300, still more preferably 150 - 200
A preferred compound is 2-(aminoacetylamino) ethylenediamine, or
AED (NH2CH2CH2NHC(O)CH20NH2).
-g-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
Other compounds are: a fluoresceinated hydroxylamine-containing (-
O-NH2) compound (e.g. FARP); a biotinylated hydroxylamine-containing (-O-
NH2) compound (BARD, Kubo K, Ide H, Wallace SS, and Kow. Biochemistry
31:3703-3708 (1992)); or hydrazine-containing (-NH-NH2) compounds (e.g.
biotin hydrazide; biotin-LC-hydrazide).
When Y = NH2 (amine), in order to remain covalently bound to the
aldehyde on DNA, an additional chemical reduction step is required. The
conditions for this reduction are well known, e.g. at pH 5-7, in the presence
of reducing agent (borohydride).
When X=NH2 (amine), in order for the covalently-bound ligand to be
recognizable by an antibody, the free -NHS group is first covalently linked to
an amine-binding compound with a recognizable group (e.g. a
succinimidylester compound such as biotin-LC-succinimidyl ester; biotin-
LC-SS-succinimidyl ester (Pierce); fluorescein-succinimidyl ester; etc.). The
reaction and purification conditions of such succinimidyl esters with -NH2
containing compounds are well known.
When X=SH (sulfhydryl), in order for the covalently-bound ligand to
be recognizable by an antibody, the free -SH group is first covalently linked
to a sulfhydryl-binding compound with a recognizable group (e.g. a
maleimide compound such as biotin-LC- maleimide; biotin-LC-SS-
maleimide (Fierce); fluorescein - maleimide; etc.). The reaction and
purification conditions of such maleimides with -SH containing compounds
are well known.
Once these compounds are covalently bound to the reactive sites,
their reaction with a detectable entity such as antibodies (e.g. avidin,
antifluorescein etc.) and their subsequent detection (e.g. chemiluminescence)
and purification (e.g. immunoprecipitation, or avidin-coated microplates, or
avidin-coated microspheres) are well known in the art.
BRIEF DESCRIPTION OF THE FIGURES
Figure 1 shows a schematic of how the present technology is applied
for identification of mutations in a complex mixture of genes, e.g. screening
for C to A transversions over hundreds or thousands of genes
simultaneously.
_g_
CA 02321119 2000-08-15
WO 99/42622 PC'f/US99/03821
Figure 2 shows the sensitivity of chemiluminescence detection of
alkaline phosphatase with a cooled ICCD camera. The inset shows time-
dependent buildup of chemiluminescence following addition of
chemiluminescent substrate plus enhancer.
Figure 3 shows chemiluminescence detection of aldehyde-containing
apurinic/apyrimidinic (AP) sites generated in plasmid DNA following
depurination in sodium citrate, pH 3.5 at 38° for up to 60 seconds. The
inset depicts fluorescence detection when extensive depurination under
identical conditions is applied. Data in the inset (from us (Makrigiorgos GM,
Chakrabarti S and Mahmood S. Int J Radiat Biol, 74:99-109 ( 1998)) were
used to convert chemiluminescence units to AP sites (right axis, see text).
Figures 4A and 4B shows sensitive detection of AP sites using FARP.
Figure 4A shows detection of AP sites in genomic calf thymus DNA
depurinated for 15 seconds, without treatment (barl) or following treatment
(bar 2) with hyrdoxylamine. Figure 4B shows detection of spontaneously
generated AP sites in hydroxylamine - treated genomic calf thymus DNA at
pH=7.0, at a temperature of 37°C (curvel) or 4°C (curve 2).
Figures SA-5C shows gel electrophoresis of Mutt-treated DNA,
examined on denaturing gels. Figure SA shows 49-mer double stranded
oligonucleotides that are Mutt-treated and visualized on polyacrylamide gels
following SYBR GOLD staining. Lane 1, No mismatch, no Mutt. Lane 2, no
mismatch, plus Mutt. Lane 3, A/G mismatch, no Mutt. Lane 4, A/G
mismatch, plus Mutt. Figure SB shows double standard homoduplex
mixtures (DNA ladder, 27-500 base pairs) are Mutt-treated and visualized
on polyacrylamide gels following SYBR GOLD staining. Lane 1, no Mutt.
Lane 2, plus Mutt. Figure 5C shows single stranded M 13 DNA (7,249 bases)
are enzymatically - treated and visualized on agarose gels following ethidium
staining. Lane 1, M 13 DNA, no Mutt. Lane 2, M 13 DNA, plus Mutt. Lanes
3-6, molecular weight markers.
Figure 6 shows FARP-based chemiluminescence detection of MutY-
treated DNA of a single length: 49-mer oligonucleotides are enzymatically
treated, FARP-labeled and captured on microplates. Bar 1, A/G mismatch,
no Mutt. Bar 2, A/G mismatch, plus Mutt. Bar 3, No mismatch, no Mutt.
Bar 4, no mismatch, plus Mutt.
- 10-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
Figure 7 shows FARP-based chemiluminescence detection of MutY-
treated DNA fragments of varying length: Single stranded M 13 DNA (7249
bases) and double stranded homoduplex mixtures (DNA ladder, 27-500 base
pairs) are enzymatically - treated, FARP-labeled and captured on
microplates. Bar 1, M 13 DNA, no Mutt. Bar 2, M 13 DNA, plus Mutt. Bar
3, ladder DNA, no Mutt. Bar 4, ladder DNA, plus Mutt.
Figures 8A and 8B show BARD-based chemiluminescence detection of
Mutt-treated DNA fragments of varying length: Figure 8A shows
chemiluminescence from single stranded M 13 DNA (that forms ~3
mismatches over 7249 bases) and double stranded homoduplex M 13 DNA
(no mismatches) enzymatically - treated by Mutt, BARD-labeled and
captured on microplates. Bar 1, s.s. M13 DNA, no Mutt. Bar 2, s.s. M13
DNA, plus Mutt. Bar 3, d.s. M 13 DNA, no Mutt. Bar 4, d.s. M 13 DNA,
plus Mutt. Figure 8B shows gel electrophoresis of the same DNA, and
demonstrates that, in agreement with the chemiluminescence results in
Figure 8A, only single stranded M 13 plus Mutt demonstrate DNA digestion
(see bands in Lane 2).
Figures 9A and 9B show detection of a mutation. Figure 9A shows
chemiluminescence detection of a single mutation (A-to-C transversion)
engineered in a p53 gene which is incorporated in a 7091 base pair plasmid.
Plasmids containing the mutation were first digested into smaller DNA
fragments (400-2,00 bp) by exposure to RSAI enzyme. These were then
melted and hybridized with normal plasmids to form mismatches at the
position of the mutation. The DNA was then enzymatically - treated with
Mutt to convert mismatches to aldehydes, BARP-labeled and captured on
microplates. Bar 1, plasmid with mismatch, no Mutt. Bar 2, plasmid with
mismatch plus Mutt. Bar 3, normal plasmid, no Mutt. Bar 4, normal
plasmid, plus Mutt. Figure 9B shows the variation of the
chemiluminescence signal obtained when different amounts of mismatch-
containing plasmid treated by Mutt and BARP are applied on microplates.
Figures l0A and lOB compare DNA binding by different compounds.
Figure l0A demonstrates the binding of the compound, AED, (2-
(aminoacetylamino) ethylenediamine) to reactive sites generated at position
of mismatches in DNA by the enzyme Mutt. The figure shows samples of
M 13 DNA containing mismatches, treated with enzyme and various
-11-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
compounds, stained with ethidium bromide and examined via gel
electrophoresis. A sample of M 13 DNA without enzymatic treatment shows a
single bright band in lane A. The sample of plasmid DNA treated with the
enzyme Mutt shows multiple bands, demonstrating the expected recognition
and cutting of mismatched bases by Mutt in lane B. Lane C In Lanes C, D
and E, the Mutt treatments are carried out in the presence of 5 mM
methoxyaxnine (C) or in presence of the novel compound AED (D, 5 mM and
E, 10 mM AED respectively). The disappearance of the bands in lanes C, D
and E is an indication of covalent high labeling of DNA by methoxyamine or
by AED, at the positions of reactive sites generated by Mutt. In Lane F, the
treatment of DNA was as in Lanes E and D, but another aldehyde reactive
compound (BARP) was used instead of AED. Lane F still shows the same
multiple bands as those generated in the absence of compound (see Lane B),
indicating an inefficient labeling of aldehyde sites by BARP.
Figure lOB demonstrates the superior DNA binding of AED over BARP
or FARP when reactive sites are generated at position of mismatches in DNA
by the enzyme TDG. Lanes 1 and 2, G/T mismatch-containing
oligonucleotide, no enzyme. Lane 3, G/T oligonucleotide with TDG enzyme.
Lane 4, G/T oligonucleotide with TDG enzyme in the presence of 5 mM
methoxyamine. Lane 5, G/T oligonucleotide with TDG enzyme in the
presence of 5 mM BARP. Lane 6, G/T oligonucleotide with TDG enzyme in
the presence of 5 mM AED. Lane 7, G/T oligonucleotide with TDG enzyme in
the presence of 0.5 mM FARP.
Figure I 1. AED-based chemiluminescence detection of mismatches
obtained when mismatch-containing s. s. M 13 DNA is Mutt treated in the
presence of 5 mM AED. Bar 1, M 13 DNA without Mutt enzyme. Bar 2, M 13
DNA with Mutt enzyme.
Figure 12 is a schematic showing how the method of mismatch
identification can be used with a DNA chip.
Figure 13 is a schematic showing how one embodiment of the method
of mismatch identification can be used with a DNA chip to detect inherited
and acquired predisposition to cancer.
DETAILED DESCRIPTION OF THE INVENTION
- 12-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
As described above, we have now developed a novel method that
identifies genetic differences between two nucleic acid strands, thereby
permitting the rapid identification of mutations in nucleic acids, e.g. DNA,
or
DNA segment(s). This method comprises (a) isolating the nucleic acid, e.g.,
DNA, to be screened for mutations (referred to as the target DNA), and
hybridizing it with control DNA, to create mismatches. Preferably the
nucleic acid has been digested to fragments of 50-500, more preferably 50-
300 base pairs. These mismatches occur at the exact positions of mutations
or polymorphisms; (b) removing any pre-existing, spontaneous aldehydes by,
IO for example, treating the DNA with hydroxylamine; (c) using repair
glycosylase enzymes to convert the mismatches to reactive sites, namely,
aldehyde-containing abasic sites (these enzymes recognize mismatches and
will 'cut' the nucleic acid base, e.g., adenine at that site to create a
reactive
site); (d) using compounds (e.g. ligands) with functional groups that at one
site can covalently bind to the reactive sites on the DNA, and that at a
second site contain unique moieties that can be detected; (e) binding
antibodies or avidin to the detectable second sites of the DNA-bound ligands.
These antibodies or avidin may carry chemiluminescent or other indicators,
so that the total reactive sites on the nucleic acid, e.g., DNA segments)
tested is quantified, e. g. by chemiluminescence; (f) either ( 1 j purifying
the
segments where a reactive site is present (e.g. by immunoprecipitation, or by
ELISA-microplate-based techniques, or by microsphere-based techniques).
The rest of the nucleic acid, e.g., DNA that does not contain mutations can
then be discarded; (g) amplifying the remaining, mutation-containing nucleic
acid, e.g., DNA, by PCR; and (h) analyzing that purified nucleic acid, e.g.,
DNA by standard gene-detection methods (e.g., hybridization), in order to
find which gene each identified mismatch belongs to; or (f~ directly adding
the sample to a substrate such as a DNA chip without isolation of
mismatches by purification; (g~ wash the chip carefully to remove
unhybridized DNA and/or unbound label (e.g., fluorescence) and read the
chip for detection of the label by the appropriate device. For example where
the label is fluorescence, a scanning laser can be used; (h~ thereafter one
looks at the chip to determine where in a gene a label indicating a mismatch
has been detected. Thereafter, by known techniques determining whether
- 13-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
that mismatch is a mutation that either causes the disorder or is associated
with the disorder or simply an allelic variation, i.e. a polymorphism.
The present method will recognize mismatches formed upon
hybridization of the target DNA and the control (wild-type) DNA. Those
S skilled in the art are aware that mismatches may appear as a result of
inherited or acquired genetic alterations. Also, that not every mismatch is
the result of mutation but that some mismatches simply represent
polymorphisms that occur naturally in populations. Both the inherited and
the acquired genetic alterations in DNA will cause a mismatch.
Furthermore, those skilled in the art are aware that because all
eukaryotic cells contain two copies of each chromosome, one paternal and
one maternal, differences between the two alleles of each gene may also
cause mismatches. In this case one gene copy (e.g. the paternal) will act as
a control DNA and the second gene copy (the maternal) will act as the target
DNA, and the mismatches will form upon hybridization of maternal and
paternal DNA (i.e. simply by self hybridization of DNA present in cells).
These inherited differences can represent either polymorphisms or
mutations.
There are a number of ways known in the art to distinguish whether a
particular mismatch is an inherited polymorphism or mutation, or an
acquired mutation.
One method that can be used to identify acquired mutations is to
have the control DNA come from the same individual. For example, when
screening a malignant cell the control DNA can be obtained from the
corresponding non-malignant cell. By screening first the non-malignant cell
alone and then the malignant cell (or a mixture of malignant and non-
malignant cells) a comparison of detected mismatches in the two cases can
be made. Differences that appear solely on the malignant cell and not on the
normal cell comprise acquired mutations which may have lead to the
malignancy.
When inherited (genetic) mutations/polymorphisms (i.e. where the
alteration from the wild-type is present at birth and in every cell of the
body)
need to be ident~ed, only normal cells need to be examined. As explained,
inherited differences between the two alleles will cause mismatches upon
- 14-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
self hybridization. Detection of these mismatches will indicate the positions
of inherited polymorphisms or mutations.
Thereafter, one standard method to discriminate inherited polymorphisms
from inherited mutations is to screen kindred and to determine whether or
not the mismatch is present in normal kindred (i.e. a benign polymorphism)
or only present in kindred showing a particular abnormality (i.e. a
debilitating mutation).
The use of databases categorizing mutations and polymorphisms has
also been increasingly popular. Thus, comparison of an identified genetic
variation with those contained in a database can in many instances be used
to determine whether the detected mismatch in DNA is due to a mutation or
due to a polymorphism. One can also look at whether the mismatch causes
truncation in the expressed protein.
Finally, another method that can be used to discriminate among
mutations and polymorphisms is by the use of in-vivo assays. Thus, one
can substitute a gene with at least one engineered base substitution
mutation for the wild-type gene in an assay to determine whether or not the
gene with the mutation can functionally replace a wild-type normal gene. If
a gene can replace a wild-type normal gene in an assay and exhibit almost
normal function that gene is not considered a mutation, but an allelic
variation (i.e, polymorphism). If it cannot that gene will be considered a
mutation.
One of the advantages of the present approach as opposed to
mutation-detection methods presently being used is the ability to identify
numerous mutations at diverse places in the genome. This permits one to
determine if certain genes not presently associated with a particular
abnormality may also have some relationship with that abnormality. For
example, with hereditary non-polyposis colorectal cancer (HNPCC),
mutations in the MSH2 and MLH 1 genes are believed to be responsible for
approximately 90% of the cases. A number of other genes have been
identified as being responsible for the other 10% of the cases. However, in
view of the cost of screening one typically looks primarily at MSH2 and
MLH 1. It may tum out when an array of genes are looked at the same time,
that mutations in other genes also play a major role, in an individual with a
particular condition. These other mutations may be associated with severity
- 15-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
of the condition. By monitoring these additional genes and looking at
disease state and recovery, one can develop a better idea of prognosis and
treatment regimes than is currently available.
When using genomic DNA the skilled artisan is aware that numerous
mismatches can and will occur in non-coding genetic regions. Looking at
non-coding regions can permit the identification of mutations that affect
expression and levels of expression. On the other hand when one is
interested in looking for mutations in the expressed proteins it is preferable
to use the mRNA to generate cDNA, and then form mismatches that can be
detected by the present approach.
The present method permits biochemical approaches for chemically
identifying the mismatch sites in, for example, the target DNA sequence.
The target DNA can be identified by a detectable moiety and subsequently
directly detected on a DNA chip or hybridization array, or purified by
immunoprecipitation, microplates or microsphere technologies.
Subsequently, the purified mutation - containing DNA fragments can be
used in single-step screening of these mismatches by a wide variety of
hybridization techniques (DNA chips, large-scale hybridization arrays, etc.)
For example, in trying to detect unknown mutations it has thus far
proven difficult to screen for a single gene of about 8.5 kb (such as APC) in
a
single experiment, especially when an excess of normal alleles is
simultaneously present [Sidransky, D. Science 278: 1054-1058 (1997)). By
contrast, the present method can screen several genes at once, and selects
and isolates only those fragments containing a mutation/polymorphism.
These mismatch-containing segments can be determined by looking for the
label. In certain embodiments they can be amplified by PCR and used, for
example, in a DNA array to simply search for the matching genes) in the
array to identify which genes these mutation-containing fragments belong to.
Consequently, existing arrays for multiplexed gene expression scanning such
as known in the art can be used. For example, Affymetrix Hu6800 DNA
Chip, or the arrays described in [Wodicka, L. et al, Nature Biotechnology: 15:
1359-1367 ( 1997); Lockart, D.J. et al, Nature Biotechnology 14: 1675-1680
( 1996); Schena, M. Trends Biotechical 16:301-306 ( 1998): Yang, T.T. et al.
Biotechniques 18: 498-503 (1995); Ginot F. Human Mutation 10: 1-10
- 16-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
( 1997)]. These arrays can also be used when the sample is directly added
without a purification or amplification step.
In the present approach, in order to increase resolution (i.e. definition
of the gene segment containing the mutation/polymorphism) the fragment
should be smaller. However, in order to effectively prepare large amounts of
mismatch-containing fragments by standard techniques such as PCR, the
fragments should be at least about 50 bases. In some instances for ease of
operation, a loss in resolution can be tolerated and larger fragments used.
Preferably the 'mismatch-containing fragment is 50 - 300 bases, more
preferably 50 - 200 bases, still more preferably 50 - 100 bases and most
preferably about 50 bases.
The nucleotides on the array (gene elements) should be between 8-
300 bases preferably no larger than the size of the DNA of the mismatch-
containing fragments. For improved resolution, smaller sizes should be
used. For example, 50 bases or less, more preferably 8-25 bases. Many
arrays presently available use nucleotide fragments of about 25 bases.
Typically, these nucleotide segments are selected to be close to the 3'
portion
of the transcript.
However, other DNA arrays as discussed, infra, can also be used.
Such arrays, which contain fragments that span the whole length of the gene
(i.e. from both the 5' end of the gene as well as the 3' end) are preferred.
The preferred target nucleic acid is DNA. The DNA can be any
mixture containing one or various sizes of DNA, such as cDNA synthesized
from the whole mRNA collected from cells that need to be screened for
mutation/polymorphism; or fractions thereof; or the whole genomic DNA
collected from cells that need to be screened for mutation/polymorphism; or
fractions thereof; or any combination of the above digested into smaller
pieces by enzymes.
The control will be a wild-type fraction similar to the target. This
wild-type likely will have no mutations. The control nucleic acid can be
selected depending upon the intent of the test. For example, where acquired
mutations in cancer cells are being screened, the control nucleic acid can
come from a "normal" cell from the same individual. In other instances, for
example, where an inherited (genetic) component may be involved the control
DNA would come from a different subject than providing the nucleic acid; or
- 17-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
simply differences among the paternal and maternal alleles can be examined
by a self hybridization of the DNA of the examined individual.
Following DNA isolation, the DNA is fragmented to reduce its size to
the desired 50-300 base pairs, and generic PCR primers are added to the
nucleic acids, in order to amplify the preparation at a later stage.
Thereafter in one embodiment, the target DNA is mixed and
hybridized with wild-type DNA to create mismatches at the positions of
differences, which are expected to be mutations/polymorphisms. The
mixture is preferably treated with a compound such as hydroxylamine to
remove any spontaneous aidehydes. Thereafter, the mismatches that
occurred are recognized and converted to reactive sites (aldehydes) by
enzymes such as a glycosylase repair enzyme such as Mutt, arid thymine
DNA glycosylase (TDG) (e.g., from Hela cells or E, coh~. A unique feature of
these enzymes is that they are highly specific, i.e. they act only on
mismatches while they leave non-mismatch containing DNA completely
intact.
These reactive sites are identified by using a compound containing an
aldehyde - binding moiety such as -O-NH2 (-hydroxylamine), or -NHNH2 (-
hydrazine) or -NH2 (-amine) and also having a second moiety that reacts
with a detectable entity (e.g. fluorescein, biotin, digoxigenin, which
respectively react with antifluorescein antibody, avidin, and antidigoxigenin
antibody. The antibodies may have chemiluminescence tags on them and
thereby are detected). A unique feature of the present approach is that the
aldehyde - binding moiety binds covalently to the enzyme-generated reactive
sites. Combined with the specificity of the mismatch - repair enzymes, the
use of covalently bound ligands to the position of mutations results in a
sensitivity and specificity which is unparalleled by other methods for
detection of mutations and polymorphisms.
The compounds have the general formula:
X-Z-Y,
wherein X is a detectable moiety, preferably X is NH2, SH, NHNHa, a
fluorescein derivative, a hydroxycoumarin derivative, a rhodamine derivative;
a BODIPY derivative a digoxigenin derivative or a biotin derivative;
- 18-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
Y is NHNH2, O-NHz or NH2, preferably Y is NH2,
Z is a hydrocarbon, alkylhydroxyl, alkylethoxy, alkylester, alkylether,
alkylamide or allcylamine. Z may contain a cleavable group (e.g. S-S). Z may
be substituted or unsubstituted.
These reactive sites are identified by using a compound containing an
aldehyde - binding moiety (Y) such as -O-NH2 (-hydroxylamine), or -NHNH2 (-
hydrazine) or -NH2 (-amine) and also having a second moiety (X) that reacts
with a detectable entity (e.g. fluorescein, biotin, digoxigenin, which
respectively react with antifluorescein antibody, avidin, and antidigoxigenin
antibody. The antibodies may have chemiluminescence tags on them and
thereby are detected) . The aldehyde - binding moiety binds covalently to the
enzyme-generated reactive sites. Combined with the specificity of the
mismatch - repair enzymes, the use of covalently bound ligands to the
position of mutations results in a high sensitivity and specificity.
One preferred embodiment of the invention has a general formula;
X~-~CH2~CHz_W~CH2~Y~
n n
wherein X' is NHNH2 or NH2, preferably NH2;
Y' is O-NHa or NH2, preferably O-NH2;
W is -NHC(O)-, -NHC(OH)-, -C(OH)-, -NH-, C-O-, -O-, -S-, -S-S-,
-OC(O)-, or C(O)O-;
n is and integer from 0 to 12, preferably 4-7 and more preferably 6;
n' is an integer from 0 to 12, preferably 4-7, and more preferable 6,
and
n" is an integer from 1 to 4, preferably 1-2, and more preferably 1.
Preferably, the compound has a molecular weight between 100 - 500,
more preferably 100 - 300, still more preferably 150 - 200.
Z and W can be substituted with groups that enhance the solubility of
the resultant compound. Preferably the compounds of the formula
X-(CH2)n-(CH2-W)n"-(CH2),;-Y are overall soluble in the solvent used.
- 19-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
A preferred embodiment has the formula;
X'~--~CH2~-N~CH2~.-.Y.,
O
wherein X", Y", n, and n' are as described as above.
A more preferred compound is 2-(aminoacetylamino) ethylenediamine (AED),
(NHaCH2CH2NHC(O)CH20NH2).
O
H2N~H~0\NHZ
2-(aminoacetylamino) ethylenediamine (AED)
In another prefered embodiment, the DNA reactive site recognized by
enzymes such as glycosylases are identified by using compounds that
contain a hydroxylamine reactive group. Examples of hydroxylamine
compounds include FARP and FARPhc, both of which are fluorescent. FARP
is a novel hydroxylamine containing derivative of fluorescein and FARPhc is
a novel hydroxylamine containing derivative of hydroxy-coumarin.
These compounds have the general formula;
X,"~CHZ~...~CH2_W~CH2~-Y~.,
n n n'
wherein Y"' is O-NH2;
X"' is a fluorescent molecule, a fluorescein derivative or a
hydroxy-coumarin derivative.
W, n, n', n" and n"' are defined as above.
More preferred compounds includes fluorescein aldehyde reactive
probe, FARP, and fluorescent reactive probe hydroxycoumarin, FARPhc.
-20-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99103821
HO / O / O
\ ~ / /
OH
\ O O
H II
HN~S~N~H~~~NH2
1OI I'O
HO / O O
H O H
/ N..~H2~N~N~O~NH2
O H ~(O
FARP FARPh~
DNA samples containing mismatches that are prepared and treated
with DNA-glycosylase enzymes as described above, will form covalent oxime
bonds to FARP and FARPhc.
In an alternative embodiment, the DNA reactive sites recognized by
enzymes such as glycosylases are identified by using compounds that
contain a hydrazine reactive group. An example of this class of compounds
includes biotin hydrazine. The present invention allows using hydrazine
compounds to label reactive sites generated by the DNA-glycosylase
enzymes. In yet still another alternative embodiment, the compound is a
biotin aldehyde reactive probe, such as BARP, a biotinylated derivative of
hydroxylamine [BARD, Kubo K, Ide H, Wallace SS, and Kow, Biochemistry,
31:3703-3708, (1992)].
These biotinylated hydroxylamine or hydrazine compounds have the
general formula;
X,~~~~CH2~CH2-W~CH2~-y".
n n n'
wherein Y"' is O-NH2 or NHNH2;
X"" is a detectable molecule, biotin or biotin derivative.
W, n, n' and n" are defined as above.
-21-
CA 02321119 2000-08-15
PCT/US99/03821
WO 99/42622
For example, a Y moiety such as an amine should react with the aldehyde on
for example the DNA, while the X group remains free for further modification
and detection.
More preferred compounds includes biotin aldehyde reactive probe,
BARP (BARP, Kubo K, Ide H, Wallace SS, and Kow, Biochemistry, 31:3703-
3708 ( 1992) and biotin hydrazide:
O O
HN~NH HN~NH
H OII H
S N.H~O.NH2 S N.NH2
O or O
BARP
It was discovered that, following the recognition of mismatches by
glycosylases such as Mutt or TDG, and the resulting conversion to
aldehyde-containing reactive sites, the enzyme has to be kept inactive,
otherwise it interferes with the subsequent covalent binding of the ligand
compounds. As a result, the conditions for reaction of hydroxylamines,
hydrazines or amines with the enzymatically - generated aldehyde-
containing reactive sites are at temperature of 4°C-15°C and at
pH 6-7. (In
the specific case of Y=NHa (amines), the presence of a reducing agent such as
borohydride, 4°C-15°C for 1-3 hours is also required during
binding to
reactive sites). Following covalent attachment of the ligand compounds to
reactive sites, the enzyme is then inactivated via heating at 70°C, for
10
minutes. Alternatively, to remove the enzyme a standard phenol-chloroform
extraction, or treatment with proteinase K can be adopted.
When X=NH2 (amine), in order for the covalently-bound ligand to be
recognizable by an antibody, the free -NH2 group is first covalently linked to
an amine-binding compound with a recognizable group (e.g. a
succinimidylester compound such as biotin-LC-succinimidyl ester; biotin-
LC-SS-succinimidyl ester [Pierce]; fluorescein-succinimidyl ester; etc.). The
-22-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
reaction and purification conditions of such succinimidyl esters with -NH2
containing compounds are well known.
When X=SH (sulfhydryl), in order for the covalently-bound ligand to
be recognizable by an antibody, the free -SH group is first covalently linked
to a sulfhydryl-binding compound with a recognizable group (e.g. a
maleimide compound such as biotin-LC- maleimide; biotin-LC-SS-
maleimide [Pierce]; fluorescein - aleimide; etc.). The reaction and
purification conditions of such maleimides with -SH containing compounds
are well known.
It was discovered that binding to reactive sites becomes much more
efficient when small hydroxylamines (such as AED) are used. Therefore, the
use of small compounds of the formula X'-(CH2)n-(CHz-W)n"-(CH2)n'-Y , and
of molecular weight less than 200 is preferred. These compounds are water
soluble, can be incubated with DNA at a high molarity (e.g. 10 mM), and are
able to diffuse fast enough to bind to reactive sites at a much higher level
of
efficiency than the other compounds (e.g. FARP, BARP) that have higher
molecular weights and are less water soluble.
A maj or additional advantage of this invention is that the
identification of the mismatch - containing DNA relies in the utilization of
aldehydes as the recognition sites for mismatches combined with covalent
bonding of the marker molecule to these aldehydes. Therefore, the presence
of contaminating nucleases that cleave DNA and create 3' hydroxyl groups -
containing strand breaks (-a common problem in similar assays-) do not
generate binding sites for the marker molecules. Since the present method
does not require the use of gel electrophoresis which compares DNA strand
by their length or size, the generation of false positives from strand breaks
generated by contaminating nucleases is thereby avoided. The method of
the invention only detects labeled DNA following covalent binding of such
aldehydes with ligand compounds. This sample can thereafter be
immobilized on a solid support, e.g., a hybridization array, a DNA chip,
microplates. In addition, the length and diversity of DNA fragments are
irrelevant to the assay, which is another advantage over gel-electrophoretic
methods.
Once these compounds are covalently bound to the reactive sites,
their reaction with a detectable group such as antibodies (e.g. avidin,
-23-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
antiffuorescein etc.) and their subsequent detection (e.g. by
chemiluminescence) and isolation (e.g. immunoprecipitation, avidin-coated
microplates or microspheres, are well known in the art. For example, when
X=NH2, direct immobilization and purification of the mismatch - containing
DNA is possible on microplates coated with activated succinimidyl ester
[Costar] or malefic anhydrite [Pierce] which covalently bind the NH2 group on
the DNA-bound linker. When X=fluorescein, direct immobilization and
isolation is achieved via antifluorescein - coated microplates [BoehringerJ.
And when X=biotin, direct immobilization and isolation is achieved via
streptavidin - coated microplates (Fierce). In all cases, the immobilized DNA
can be detected via alkaline-phosphatase or peroxidase - based
chemiluminescence assays (see paper submitted to Nucleic Acid Research].
Those of ordinary skill in the art will recognize that a large variety of
other possible detectable moieties can also be coupled to antibodies used to
I5 bind the DNA-coupled linkers at the positions of mismatches in this
invention. Thereby providing additional methods to , detect the
antibody-bound mismatches on DNA. See, for example, "Conjugate
Vaccines", Contributions to Microbiology and Immunology, J.M. Cruse and
R.E. Lewis, Jr. (eds), Carger Press, New York, (1989).
The term "substituted," as used herein refers to single or multiple
substitutions of a molecule with a moiety or moieties distinct from the core
molecule. Substituents include, without limitation, halogens, hetero atoms,
(i.e. 0, S and N), nitro moieties, alkyl (preferably C1 - C6), amine moieties,
nitrite moieties, hydroxy moieties, alkoxy moieties, phenoxy moieties, other
aliphatic or aromatic moieties. Preferably the aliphatic or aromatic moieties
are lower aliphatic or aromatic moieties, i.e. 12 or less carbons, more
preferably 6 or less carbon atoms. Substituted compounds may be referred
to as derivatives of the core structure.
Antibodies of the present invention can be detected by appropriate
assays, such as the direct binding assay and by other conventional types of
immunoassays. For example, a sandwich assay can be performed in which
the receptor or fragment thereof is affixed to a solid phase. Incubation is
maintained for a sufficient period of time to allow the antibody in the sample
to bind to the immobilized labeled DNA on the solid phase. After this first
incubation, the solid phase is separated from the sample. The solid phase is
-24-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
washed to remove unbound materials and interfering substances such as
non-specific proteins which may also be present in the sample. The solid
phase containing the antibody of interest bound to the immobilized labeled
DNA of the present invention is subsequently incubated with labeled
S antibody or antibody bound to a coupling agent such as biotin or avidin.
Labels for antibodies are well-known in the art and include radionucleotides,
enzymes (e.g. maleate dehydrogenase, horseradish peroxidase, glucose
oxidase, catalase), fluorophores (fluorescein isothiocyanate, rhodamine,
phycocyanin, fluorescamine), biotin, and the like. The labeled antibodies are
incubated with the solid and the label bound to the solid phase is measured,
the amount of the label detected serving as a measure of, for example, the
amount of anti-FARP antibody present in the sample. These and other
immunoassays can be easily performed by those of ordinary skill in the art.
The present method allows for extremely sensitive mismatch -
scanning in diverse DNA fragments, thereby resulting in sensitive and high
throughput mutation screening over several hundreds or thousands of genes
at once. For example, it becomes possible for the screening and discovery of
novel mutations in tumor samples which is instrumental to establish the
pathogenesis of cancer and to establish new relations between mutations
and cancer or other diseases. The new compounds and methods described
above are also useful in analysis of the genetic background (polymorphisms,
mutations) of any individual. These new compounds and methods may also
be used for high throughput genotyping and genotypic selection.
One can use DNA chips to identify the gene where the mismatch is
present. For example, the Affymetrix Inc. (San Diego, CA) HU6800 DNA
chip; the Clontech AtlasTM DNA array (Palo Alto, CA); the Telechem
International array (San Jose, CA); the Genetix Ltd. array (Dorset, UK); and
the BioRobotics Ltd. array (Cambridge, UK). The chip such as the Affymetrix
DNA chip contains densely-packed DNA or RNA elements. For highest
resolution the oligomers on the chip should be small. Preferably 8-50
nucleotides, more preferably 8-25 nucleotides. This will provide the highest
resolution. However, the DNA or mRNA on the chips can be as large as the
mismatch-containing DNA fragments, e.g. 50-300 nucleotides.
For example, using a conventional array, (e.g., the Affymetrix chip for
detecting gene expression) the array will have multiple DNA or RNA elements
-25-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
densely packed, each comprising 25-mer oligonucleotides immobilized on a
solid support. For each of the 6,800 genes which are represented on the
chip, there are 20 elements each containing 25-mer oligonucleotides with a
distinct portion of the mRNA sequence. Thereby the 20 elements 'sample'
the mRNA sequence of the gene. In the current chip version, the
immobilized probes are biased towards the 3' end of the mRNA, thus
sequences towards the 5' end are not well represented. To use the array for
detecting gene expression, users generate cDNA from the genes to be
screened in the test sample (typically l:g) and then perform in-vitro
transcription to collect cRNA and biotinylate it (-50:g), 12 :g of which are
hybridized on the chip (alternatively, cDNA can directly be applied on the
chip without in-vitro transcription). If a gene is present in the test sample,
then it hybridizes to an appropriate array element. Because the array is
constructed to contain known gene sequences at known positions, all the
transcribed genes are detected in a single step. The detection process
utilizes addition of a marker-identifier such as a fluorescent scanner. The
magnitude of the signal from each element signifies the degree of gene
expression for the specific gene.
It should be noted that the present invention can use the same arrays
as currently being used, but not anymore to detect gene expression (i.e.
difference in signal among array elements), but mutations, which requires
only detection of presence or absence of signal (indicating
polymorphism/mutation in the specific gene fragment which was captured),
thereby making the detection task much simpler.
Inherited single nucleotide polymorphisms (SNPs) and mutations can
define a genetic predisposition towards several diseases, including cancer,
cardiovascular, neurodegenerative and others. Indeed, acquired SNPs,
mutations and loss of heterozygocity are particularly pertinent to cancer
development, and early cancer detection. All of the above can be
simultaneously detected in a single step by the above-described methodology
(See e.g., Fig. 12).
For example, cDNA for tumor and normal issue of a single individual
is prepared. (See Figures 12 and 13) Because inherited polymorphisms is a
frequent event (average 1 SNP per 1000 bases), several genes will have more
-26-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
than one SNP. Also, the tumor genes will contain one or more inherited
SNPs as well as occasional acquired SNPs/mutations. Next, the cDNA is
digested by enzymes down to small fragments (-100-200 bp), thereby
generating fragments that are likely to contain only one -or none- genetic
alternations. Then, each sample is melted and self hybridized, to generate
mismatches at positions of SNPs/mutations. The above-described
methodology using an X-Z-Y compound is applied as described above, to
identify only the mismatch-containing cDNA. In one embodiment the sample
can be added directly to the chip. Alternatively, the sample can also be
purified to isolate the mismatch-containing DNA.
The mismatch-containing cDNA can be PCR-amplified, labeled, e.g.
biotinylated, and applied on a chip such as the Affymetrix chip. Each
mismatch-containing fragment will hybridize to its complementary
oligonucleotide on the array, thereby revealing which gene and which gene
region (to within 100-200 base pairs) the SNP/mutation belongs to. When
the sample is added without the isolation step, the mismatch containing
fragments are directly analyzed by reading the chip for the label. By
comparing arrays A and B, both the inherited and the acquired
SNPs/mutations can be derived. Loss of heterozygocity may occur when an
acquired SNP/mutation occurs in the same gene with an inherited
SNP/mutation. Such genes can readily be identified by comparing A and B.
Current arrays, including the Affymetrix chips, because they are
intended for detection of gene expression, they utilize immobilized
oligonucleotides which are biased toward the 3' end of mRNA. Accordingly,
the 5' end of the gene is underrepresented and therefore will miss all the
mutations that are towards the 5' end of the genes. Therefore, although the
combination of the present methodology with the existing DNA chips allows
mutation scanning over several sections of the genome (which is currently
impossible by other methods), the mutation scanning is restricted towards
the 3' end of the genes. By contrast, our methodology combined with new
DNA chips (infra) makes it possible to identify mismatches over complete
sections of the genome.
A preferred Mutation Scanning Array should contain immobilized
oligonucleotides, preferably 8-25 bases long, which span the whole mRNA
sequence of each gene represented on the array, and not biased toward one
-27-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
or the other mRNA end. As mentioned, the oligonucleotides can be larger,
but by increasing size, resolution is lost. The oligonucleotides should
sample the mRNA in intervals not bigger than the DNA fragments isolated by
present method preferably 50-100 bases but capable of ranging from 20 -
300 bases In this manner the mismatch-containing fragment will be
assured of finding a complementary sequence on the array. When
immobilized oligonucleotides on the array are arranged to sample the mRNA
at small intervals (e.g. 20 bases) there will be redundancy of information
upon hybridization of the mutant fragments to the DNA chip, as each
fragment may simultaneously hybridize to two or more immobilized
oligonucleotides. In this case, by using the combined information from all
array elements, a better resolution of the position of the mutation will be
achieved.
This Mutation Scanning Array can be constructed using the same
technologies as for the current arrays. The described modification will allow
SNPs/mutation detection over the whole length of the immobilized genes to
be identified. The immobilized genes can be either the whole genomic cDNA
library, or an arbitrary fraction of that, or a specific collection of genes
that
are known to be related to a specific disease (i.e. disease spec arrays).
A major advantage of the present mutation scanning chip technology
is that it can detect SNPs/mutations in the presence of an excess of normal
alleles in the initial sample because the methodology first isolates the
mutants, and the array subsequently identified the gene. This is currently
impossible to do with existing technology.
A preferred kit will comprise reagents to isolate mRNA from tissues,
synthesize cDNA, fragment DNA to 100-200mers and add PCR primers, form
heteroduplexes, use Mutt and TDG enzymes to cut the mismatches, remove
spontaneous aldehydes, apply the X-Z-Y compounds e.g., FARP/BARP/AED,
to detect mismatches, and isolate mutant fragments by immobilization on
microplates, recover and PCR mutants, and finally apply on an array to
detect SNPs/mutations at specific genomic positions.
The kit can be used to screen an individual for inherited susceptibility
to cancer, cardiovascular disease, neurodegenerative disorders, etc. by
mapping positions of heterozygocities and SNPs in the whole genome or in
selected fractions of the genes.
-28-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
The present methodology also permits one to detect early onset of
cancer (acquired SNPs/mutations) from tissue biopsies or excretions. The
present technology also permits research labs to detect new mutations and
correlate them to other diseases.
The ligand compounds described demonstrated excellent detection of
DNA mismatch - repair recognition sites. In addition, based on our
discovery that small (MW<200-250) compound allow high binding efficiency
(>50%) to DNA reactive sites, new compounds (like AED) were designed,
synthesized and tested. These were shown to bind reactive sites generated
by Mutt much more efficient than compounds of higher (>250) molecular
weight. These new compounds are unique in that they are small, water
soluble, do not encounter significant steric interactions with DNA and can
diffuse fast to the enzyrnatically - generated reactive sites on DNA. This
class of new bifunctional compounds is also uniquely designed to retain
their water solubility as the chain length is extended. The simultaneous
addition of internal polar functional groups along with methylene groups
maintains the water solubility of these compounds in spite of the increased
length of the molecule. Care must be taken however to retain a low overall
molecular weight for the final compound. Useful polar functional groups
include; alcohols, esters, ethers, thioethers, amines and amides. This allows
users of this method the flexibility to tailor the chain length of the
compounds to suit their specific needs with out the loss of water solubility,
which is essential.
In one method which aims to map base substitution mutations in
tumor samples, mRNA is isolated from a malignant cell. The corresponding
mRNA from a healthy or normal tissue sample is also isolated. The mRNA
from the normal tissue will serve as the wild-type control. A cDNA library
can be made for each mRNA sample, the cancerous and wild-type. The two
cDNA libraries are added together, for example in a 1:1 ratio and hybridized.
(See Figure 1) The hybridization produces a mixture of double stranded
DNA. The double strands of DNA that consist of cDNA from the malignant
cell hybridized with a strand of wild-type DNA will now typically contain
some mismatches that are associated with the malignancy.
The mixture of hybridized cDNA is then treated with hydroxylamine to
remove any spontaneous aldehydes, and then the hydroxylamine is removed
-29-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
via G25 filter centrifugation of the samples. The double stranded cDNA
which is now void of pre-existing aldehydes, is then treated with a mismatch
- repair glycosylase, such as Mutt or TDG. Mutt is a DNA-repair enzyme
that recognizes mismatched adenosine nucleotides, and TDG recognizes
mismatched thymines. Upon recognition, Mutt or TDG remove the base by
cleavage at the point of attachment to the deoxyribose sugar. Removal of the
base by this method of cleavage results in the opening of the deoxyribose
ring with formation of an aldehyde. Since pre-existing aldehydes were
removed by hydroxylamine treatment, the only aldehydes are those
generated at positions of mutations.
The resulting strands of cDNA now contain an aldehyde located at
each point of mismatch. These resulting aldehydes are then treated with
one of the compounds, e.g. the 2-(aminoacetylamino)ethylenediamine (AED)
or one of its analogues, at low temperature so that further activity of the
MutY/TDG enzymes is suppressed. The DNA labeled with AED can then be
immobilized on DNA chips, arrays, microplates as described earlier in this
text. The chips and arrays can be scanned for the identification of the
mismatch-containing fragment. The label can also be used to selectively
immobilize the fragment. In such an embodiment the unlabeled DNA is then
washed away leaving behind only AED labeled DNA attached to the
microplate. The DNA with the labeled mutations, while immobilized on the
microplates is then biotinylated and the mutations can be detected, for
example, via chemiluminescence. Mutation-containing DNA can then be
recovered from microplates for identification of the genes involved via PCR
and large-scale hybridization techniques which are established in the field of
molecular biology. Consequently, all mismatch containing genes are
captured at once and the number of genes that can be simultaneously
screened is only limited by the total genes the DNA array can handle. To
verify and identify the exact position of the mismatches) on each particular
gene identified by the present invention, conventional procedures such as
sequencing can be used.
In another embodiment, the fluoresceinated compound FARP can be
used instead of AED. FARP-labeled DNA is immobilized on microplates,
isolated from unlabeled DNA and the total number of mismatches may be
detected by a sensitive photon-detecting technique, e.g. fluorescence or
-30-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
chemiluminescence. The mismatch containing DNA is subsequently
recovered from the microplates for identification of the genes containing
mismatches. This may be performed in a single step by large-scale
hybridization procedures on DNA arrays.
This method can also be used to detect a variety of other DNA lesions
that are converted to reactive sites by glycosylase enzymes or by chemical
means (e.g. clustered DNA-damaged sites); abasic sites; carcinogen-DNA
adducts; damaged DNA bases). In these embodiments, mixing of for
example the target DNA with wild-type DNA to create mismatches is not
needed. Enzymes will recognize damage and will generate reactive sites
directly in the target DNA. Such enzymes include all known glycosylases,
such as endonuclease III, T4 endonuclease V, 3-methyladenine DNA
glycosylase, 3- or 7- methylguanine DNA glycosylase, hydroxymethyluracile
DNA glycosylase, FaPy-DNA glycosylase, M. Luteus UV-DNA glycosylase.
Also, chemical agents such as bleomycin, alkylation agents or simple acid
hydrolysis can generate reactive sites automatically in target DNA without
any enzyme. The crucial step however is again the same, i.e. covalent
addition of compound to the reactive site of the DNA lesion, which allows
subsequent sensitive detection.
The described technology can be used for mutation screening and for
research. For example, the use of solid supports at every stage of the assay
will substantially shorten the time required to screen tumor samples,
improve its cost-effectiveness in terms of man-power as well as its
reliability
and reproducibility.
For instance, magnetic microsphere technology can be utilized to
immobilize heteroduplexes at an early stage of the assay. Following mRNA
extraction from e.g., a host cell such as cancerous and normal samples,
cDNA for e.g. 588 genes can be generated. Thereafter PCR primers that
contain a cleavable (S-S) biotin are added. Hybridization of the cancerous
cDNA with wild-type alleles generates heteroduplexes at the positions of base
substitution mutations, and the DNA sample is immobilized on, for example,
the streptavidin - coated magnetic microspheres (available from Dynal Inc.).
From this point onwards, all subsequent steps of the ALBUMS assay can be
conducted on the solid support.
-31-
CA 02321119 2000-08-15
WO 99/42622
PCT/US99/03821
The microspheres allow chemical/enzymatic treatment of the
immobilized DNA and efficient, rapid separation of chemicals from DNA via
magnetic immobilization of the microspheres during washing. For example,
in one embodiment the assay uses hydroxylamine treatment to remove
traces of aldehydes and subsequent complete removal of hydroxylamine via
repeated (x3) ultracentrifugation through G25 filters. This can be time-
consuming and result in an inevitable loss of sample, which can be
important when tissue samples are limited. In contrast, by immobilizing the
DNA magnetic microspheres, all subsequent steps become faster, easier and
without DNA loss: Hydroxylamine treatment and removal, enzymatic
treatment and washing, X-Z-Y treatment and washing, binding
antifluorescein-AP to e.g. AED-trapped mismatches and washing, and finally
chemiluminescent detection of mismatches are performed on the magnetic
microsphere format.
Alternatively, to recover the DNA from magnetic microspheres and
isolate the X-Z-Y, e.g. FARP, containing DNA, instead of adding
antifluorescein-AP the immobilized DNA can be recovered by cleaving the
disulfide (S-S) bond on the biotin by mild exposure to a reducing reagent
(DTT, 50 mM, ~10 min, 25°C).
To construct primers end-labeled with a cleavable moiety such as
biotin, oligonucleotides containing a terminal aliphatic amine are ordered,
and reacted with e.g. a biotin -S-S- succinimidyl ester (available from
Pierce).
Reactions of succinimidyl ester with amino-oligonucleotides and subsequent
purification by reverse C18 column chromatography are standard
procedures on which our group has had prior experience.
Following removal of DNA samples from the magnetic microspheres,
the samples will be applied on e.g. antifluorescein-microplates to isolate
e.g.,
FARP-containing heteroduplexes which subsequently will be recovered, PCR
amplified and screened on the Clontech DNA hybridization array. Using the
above procedures, base substitution mutations can be isolated via ALBUMS,
amplified by PCR if desired and screened on the DNA array in less than 24
hours. Thus, this technique results in a standardized procedure with easy
access to researchers and clinicians for cost - effective, large - scale
mutation screening of a target sample, such as cancer samples.
-32-
CA 02321119 2000-08-15
WO 99/42622
PCTNS99/03$21
In one embodiment, kits for carrying out the identification of these
DNA mismatches can be sold. The kits would include the repair glycosylase,
an X-Z-Y compound and preferably instructions. These materials can be in
any vial. The materials can be in lyophilized form.
In a preferred embodiment, PCR primers would also be included.
In one preferred embodiment the following kit materials and
instructions can be included:
Kit Formulation:
1. Isolate target and control cDNA. Fragment DNA to 100-200mers by
standard enzymes.
2. Add PCR primers that contain a cleavable biotin at the end.
3. Mix target with control, cross-hybridize.
4. Bind sample to streptravidin - coated magnetic microspheres.
(alternatively, streptavidin - coated microplates can be used).
5. With the sample immobilized on solid support, perform:
hydroxylamine treatment/washing; MutY/TDG treatment(s)/washing;
FARP/BARB/AED labeling/washing. Antibody labeling/washing;
Chemiluminescence detection of mismatches. All these steps are very easy
and convenient to perform with the DNA immobilized.
6. To recover sample and isolate the mutation - containing DNA, add
DTT (see below) to break the S-S bond on the cleavable biotin.
?. Now apply the preparation on an appropriate solid support for the
ligand compound chosen: (antifluorescein, streptavidin, succinimidyl - ester
-coated plates for FARP, BARD and AED respectively). Remove unbound
DNA, capture only mutated DNA.
8. Now collect mutated DNA from microplates. This can be done by
several methods; e.g. adding 1 M of hydroxylamine to break the bond
between the ligand and the DNA; or raising the temperature to denature
captured DNA and collect the unmodified strand; or, in the case of cleavable
-S-S- containing probes, simply add DTT to break the bond to the
microplate.
9. Apply PCR using the primers inserted in step 2.
10. Detect mutated genes using hybridization techniques.
-33-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
All documents mentioned herein are incorporated herein by reference.
The following examples are illustrative of the invention and are not
limitations thereon.
EXAMPLE 1 METHOD FOR LARGE-SCALE DETECTION OF
BASE-SUBSTITUTION MUTATIONS IN CANCEROUS SAMPLES, USING ONE
OF THE X-Z-Y COMPOUNDS, THE FARP MARKER MOLECULE (See Figure
1)
Isolated mRNA from a cancerous tissue is transcribed into cDNA.
Primers can be added to DNA at this stage for PCR amplification at a later
stage (see Figure 1). The sample is then hybridized with a corresponding
wild-type sample of DNA to generate mismatch pairing at the positions of
mutations. The hybridized DNA is treated with hydroxylamine to remove
any aldehydes that may have formed spontaneously. The hybridized DNA
sample is then treated with the Mutt enzyme. Enzyme treatment recognizes
A/G mismatches and upon recognition, depurinates the DNA and
simultaneously generates an aldehyde at the site of mismatch. The DNA is
then treated with the labeling compound AED or FARP or BARP to generate
a covalent oxime bond at the position of the mismatch. Upon labeling, the
DNA is immobilized on microplates appropriate for the specific labeling
compound and excess, unlabeled DNA is washed away. The DNA labeled at
mismatch sites can now be analyzed by a variety of methods including
detection of total mutations by chemiluminescence or identification of
labeled genes via DNA arrays.
MATERIALS AND METHODS
1) DNA, oligomers and chemicals: FARP [5-(((2-(carbohydrazino)-
methyl)thio)acetyl)-aminofluorescein, aminoxyacetyl hydrazide, Fluorescent
Aldehyde Reactive Probe) was synthesized as described (Makrigiorgos GM,
Chakrabarti S and Mahmood S., Int J Radiat Biol, 74:99-109 (1998)). High
purity genomic calf thymus DNA and double stranded ladder (pUC 18 Msp I
digest, 27-500 base pairs) was purchased from Sigma Chemical and used
without further purification. Single stranded (+strand) M 13 DNA was
purchased from Pharmacia Biotech and pGXIs 14 plasmid DNA, a gift from
-34-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
via spontaneous depurination at 37°C, pH=7.0, over a period of days,
and
these were monitored with the present assay. Prior to incubation at
37°C,
the DNA was treated with 5mM hydroxylamine for 1 hour at room
temperature to remove traces of existing aldehydes from the pool of potential
FARP-binding sites. The hydroxylamine was then removed via G25
ultracentrifugation and the sample was resuspended in sodium phosphate
buffer, pH 7.
3) FARP-trapping of aldehydes and subsequent DNA biotinylation.
To covalently trap open-chain aldehydes generated in DNA at the
position of AP sites, 500 :M FARP was reacted with 0.05-2.5:g of DNA in 40
mM sodium citrate pH 7.0 at 15-22°C, for 30 minutes. Non-covalently
bound FARP was removed by G25 uitracentrifugation. FARP-labeled DNA
was either used on the same day or stored at 4°C or -20°C for a
few days,
prior to further experiments. To immobilize FARP-labeled DNA on
neutravidin microplates, the DNA was exposed for one hour to a
commercially available biotinylation reagent (Biotin Label IT'E'M reagent,
1:1;
reagent per :g DNA, in MOPS buffer , pH 7.5 at 37°). Excess reagent was
them removed by G25 ultracentrifugation. The samples were either used
immediately or stored at 4°C for a few days, prior to chemiluminescent
studies.
4) Chemiluminescence measurement of FARP-trapped aldehydes in calf
thymus or plasmid DNA.
Double stranded DNA, doubly labeled with FARP and biotin, was
immobilized on neutravidin - coated microplate strips in the presence of 5
nM antiF-AP. 30-50 ng of doubly labeled DNA plus 5 nM antiF-AP in a total
of 50 :1 was incubated at room temperature for one hour in TE pH 7.5.
Unbound sample and antiF-AP were removed by pipeting and washing with
TE at least four times. The microplate strips were then transferred in to 50
ml polypropylene tubes and washed four times in 30m1- 50 ml of TE buffer
with constant agitation for 10 minutes. The chemiluminescent
substrates(CDP-Star plus Emerald II enhancer) were then added in 0.1 M
diethanolamine, pH 8.5 and the anti-F-AP-catalyzed reaction was carried out
at room temperature for 1 hour, after which maximum light generation was
achieved. In separate experiments, to quantitate the fraction of biotinylated
-36-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
Professor MacLeod, MD, Anderson Cancer Center, was isolated from the host
bacteria as described earlier (Makrigiorgos GM, Chakrabarti S and Mahmood
S., Int JRadiat Biol., 74:99-109 (1998)). Both agarose gel electrophoresis
and the absorbance ratio at 260 nm to 280 nm were performed to determine
the purity of the plasmid. Gel-purified 49-mer oligonucleotides representing
the TFIIIA transcription factor - binding sequence of the Xenopus rRNA gene
(enumerated in Table 1, at the end of this Example) were supplied by Oligos
Etc Inc. Enzyme Mutt (E. colt) was purchased from Trevigen Inc. and stored
as recommended by the manufacturers. Hydroxylamine purchased from
Sigma Chemical was already freshly made prior to the experiments. GTG
agarose was obtained from FMC Bioproducts, polyacrylamide gel
electrophoresis reagents were from National Diagnostics while SYBR GOLD
nucleic acid gel stain and Picogreen9 DNA quantitation dye was supplied by
Molecular Probes. For chemiluminescence studies, Reacti-Bind NeutrAvidin
coated polystyrene plates (pre-blocked with Bovine Serum Albumin) were
supplied by Pierce. Anti-fluorescein-Fab fragments {Sheep) - alkaline
phosphatase conjugate (antiF-AP) was purchased from Boehringer
Mannheim. CDP-Star, a 1, 2 dioxetane chemiluminescent enzyme substrate
and Emerald-II8 enhancer used with CDP-star was purchased from TROPIX.
Micro Bio-Spin G25 chromatography columns were obtained from Bio-Rad
laboratories. Label IT9 Nucleic Acid biotinylation kit was purchased from
PanVera Inc. All reagents and buffers were of analytical grade and made
with ultrapure water ( 1800 Mohm m-1 resistivity) delivered by an Alpha-Q
system (Millipore).
2) Acidic or physiological depurination of calf thymus DNA.
Treatment with hydroxylarnine.
Aldehyde containing apurinic/apyrimidinic {AP) sites were chemically
induced in calf thymus or plasmid DNA by a short exposure (0-60 seconds)
to acidic conditions (pH=3.5) over a set time period at a temperature of
38°C,
as described (Makrigiorgos GM, Chakrabarti S and Mahmood S., Int J Radiat
Biol., 74:99-109 (1998)). The reaction was halted by placing the sample
quickly on ice and adding a neutralization solution ( 10% of 3M sodium
acetate and 1 M potassium phosphate buffer at pH 7 and 7.5 respectively), 'to
final volume of 50 :1. AP sites were also slowly generated in calf thymus DNA
-35-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
DNA captured on microplates Picogreen9dye was used to measure double
stranded DNA just prior and after its removal from neutravidin-coated
plates.
5) Chemiluminescence Instrumentation
The low light from the chemiluminescence reaction was detected
using an intensified charged coupled device (ICCD) system (Princeton
Instruments). This ICCD camera utilizes a proximity focused microchannel
plate (MCP) image intensifier, fiber-optically coupled to the CCD array. The
entire area of the ICCD is capable of light detection, giving a total of 576 x
384 pixels on a Pentium~ PC computer screen. Both the intensifier and
CCD are cooled to -35°C thermoelectrically and the dark current is
less than
50 counts per minute. The ICCD was used to detect total light generation
from each cell of the microplate strip. Cells were individually placed in a
reproducible geometry at -2mm distance from the ICCD and the total light
output per second measured. The background chemiluminescence (signal
measured when FARP was omitted from the procedure) was subtracted from
all samples. All measurements were repeated at least three times.
6) Formation of homoduplex and heteroduplex oligonucleotides.
49-mer oligonucleotides and their complementary strands with or
without a centrally located T-to-G base substitution were synthesized. In
another synthesis of the same oligomers, 5' biotinlyated 49-mers and their
complementary unbiotinylated strands were synthesized (Table 1 ) . For
hybridization, equimolar amounts (~0.5 :g) of each oligonucleotide were
annealed in 40 mM Tris-HCl (pH 7.5), 20 mM MgCI2 and 50 mM NaCl to
form duplex oligonucleotides. The mixture was first heated to 95°C for
2
minutes, then allowed to hybridize at 65°C for 3 hours and cooled
slowly to
room temperature. Following hybridization, the double stranded 49-mers
were treated with hydroxylamine (5mM in citrate pH 7.0, for 30 minutes,
25°C) to remove traces of spontaneously or heat--generated aldehydes
from
the pool of FARP-reactive sites.
7) Treatment of M 13 DNA, ladder DNA and duplex oligonucleotides with
Mutt and TDG and gel electrophoresis:
50 ng of the test DNA (single stranded M 13, ladder DNA, or duplex
oligonucleotide were incubated for 1 hour, 37°C with 1.0 unit Mutt in
40
-37-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
mM Na-citrate buffer (pH 7.0) and then alkali treated to concert positions of
missing adenine to strand breaks. Analysis of cleavage products for single
stranded M i 3 DNA was done by agarose gel electrophoresis 0.9% agarose,
run overnight at 20 V in 1X TBE buffer and stained with l:g/ml ethidium
bromide). Fragment analysis for ladder DNA and oligonucleotides was done
by 16% denaturing polyacrylamide gel electrophoresis in the presence of
7.5M urea at 20 V/cm. The DNA fragments were detected by SYBR Gold dye
or by ethidium staining and photographs taken by Eagle EyeTM Still Video
(Stratagene) .
8) Chemiluminescence measurement of FARP-trapped mismatches in
oligonucleotides, ladder and M 13 DNA
M 13 DNA, ladder DNA, or 5'-biotinylated oligonucleotide duplexes,
hydroxylamine-treated, were exposed to Mutt, FARP-labeled biotinylated
with the protocols described above. The biotinylation step was omitted for
the oligonucleotides since these were pre-biotinylated. In some experiments,
samples were kept at 70°C for 8 minutes to inactivate the enzyme at
this
stage. Typically 50 ng from the doubly (biotin plus FARP) labeled nucleic
acids were applied on neutravidin - coated microplates and their
chemiluminescence measured.
RESULTS
1) Dual labeling of DNA and chemiluminescence detection using the
present protocol
Figure 2 shows chemiluminescence obtained with the present setup
when serial dilutions of free alkaline phosphatase were added to CDP-Star~
substrate and Emerald II enhancer and measured using the cooled ICCD.
The chemiluminescence detection limit of this set up is less than 0.01
attomoles alkaline phosphatase. Examination of the buildup of alkaline
phosphatase chemiluminescent signal in solution following mixing with
substrate plus enhancer at room temperature, demonstrates that after 60
minutes a relatively constant value is achieved (Figure 2, inset). Therefore
all measurements reported were conducted 60 - 80 minutes following
addition of the substrate. To estimate the fraction of biotinylated DNA
captured on the neutravidin-coated microplates, biotinylated DNA was
-38-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
quantitated using the fluorescence of Picogreen TM dye prior to its
application
and immediately following removal of unbound DNA from microplates (not
shown). 49-mer oligonucleotides resulted in approximately 10% capturing
on the plates while of the 50-100 ng high molecular weight calf thymus DNA
less than 2% was immobilized on the plates, possibly due to secondary
structures and associated steric hindrances.
2) Ultrasensitive detection of aldehydes in DNA
Chemiluminescence detection of aldehyde-containing AP sites
generated in 100 ng plasmid DNA following depurination in sodium citrate,
pH 3.5 at 38°C for up to 60 seconds and trapping of AP sites by FARP is
depicted in Figure 3. The induction of luminescence is linear with respect to
depurination exposure. The inset, from an earlier work (Makrigiorgos GM,
Chakrabarti S and Mahmood S., Int JRadiat Biol., 74:99-109 (1998)),
demonstrated detection of fluorescence following FARP-labeling of this same
plasmid exposed under identical conditions to higher depurination times (0-
60 minutes). The fluorescence-based approach is less sensitive than the
present method, however, it allows direct quantitation of the number of
FARP molecules per DNA base pair. Five minutes depurination under the
same protocol yields approximately 1 AP site per 34,000 bases (Makrigiorgos
GM, Chakrabarti S and Mahmood S., Int J Radiat Biol., 74:99-109, 1998).
Assuming a linear decrease of AP sites for lower depurination exposures, the
15 second exposure in Figure 3 corresponds to approximately 1 AP site per 7
x 105 bases. The amount of microplate-captured DNA generating this signal
is approximately 1-2 ng. Therefore the absolute number of AP sites recorded
following 15 seconds depurination is approximately 5 attomole (see right axis
in Figure 3 ) .
To estimate the lowest number of AP sites detectable, hydroxylamine
treatment of genomic calf thymus DNA was first employed in order to remove
traces of spontaneously-generated AP sites (e.g. AP sites expected to be
present in genomic DNA from mammalian cells prior to DNA extraction plus
AP sites generated during handling). Hydroxylamine is a small molecule and
is expected to react rapidly with aldehydes, as previously demonstrated for
methaxyamine (Talpaert-Borle M, and Liuzzi M., Biochimica Biophysica Acta,
740:410-416 ( 1983)), thereby prohibiting subsequently added FARP to react
at the same positions. Figure 4A depicts the decrease in the
-39-
CA 02321119 2000-08-15
PCT/US99/03821
WO 99142622
chemiluminescence signal obtained following hydroxylamine treatment of
genomic calf thymus DNA depurinated for 15 seconds. Following
hydroxylamine removal and reaction with FARP, the chemiluminescence was
reduced to almost background levels. When hydroxylamine-treated calf
thymus DNA was kept at 37°C, phosphate buffer pH=7, and assayed for AP
sites via FARP as a function of time, a linear increase in spontaneously-
generated aldehydic AP sites was detected (Figure 4B). DNA kept at 4°C
under similar conditions did not display any luminescence signal (Figure
4B). According to Figure 4B, the limit of detection by the present
microplate-based method is ~0.2 attomole AP sites, or 1 AP site per 2x10
bases, using a starting DNA material of about 100 ng.
3) Gel electrophoresis of Mutt-treated oligonucleotides and single
stranded M 13 DNA.
49-mer oligomers engineered to form a double stranded structure,
with or without a centrally located A/G mismatch upon hybridization, were
exposed to Mutt, alkali treated and examined upon denaturing gel
electrophoresis. Generation of the two expected fragments was observed for
the heteroduplex oligomers, while no cutting as present in the homoduplexes
(Figure 5A). Under the conditions applied, the fragmented DNA appears to
be less than 50% of the total DNA per lane, which would result if all A/ G
mismatches were reacted upon by Mutt. The homoduplex-containing double
stranded DNA ladder (27-500 base pair fragments) did not demonstrate
additional fragmentation following enzymatic treatment (Figure 5B). In
contrast, Mutt treatment of the 7249 base-long M 13 single stranded DNA
resulted in the generation of approximately 6 fragments, the largest of which
is about 1000 bases long, as demonstrated in lane 5, Figure 4C. Generation
of Mutt-recognized sites in the single stranded high molecular weight DNA is
attributed to sequence self-complementation generating transient
mismatches. It can be inferred that, to generate 6 discrete fragments, and
assuming a less than 100% efficiency of Mutt in cutting each site, an
average of 3 Mutt-recognized cutting sites are generated per each 7249
base-long M 13 molecule.
4) FARP-based chemiluminescence detection of mismatches in high and
low molecular weight DNA.
-40-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
Starting with 100 ng of biotinylated 49-mer homoduplexes or
heteroduplexes,
the nucleic acid was treated successively with hydroxylamine, Mutt, then
FARP and applied on neutravidin microplates for chemiluminescence
detection of mismatches. A strong signal was obtained for A/G mismatch-
containing oligonucleotides (Figure 6), while no signal was obtained when
Mutt was omitted, or when oligonucleotides without mismatch were MutY-
treated. A mixture of double stranded homoduplexes (DNA ladder) treated in
the same way also demonstrated absence of chemiluminescence signals
(Figure 7). In contrast, single-stranded M 13 demonstrated a
chemiluminescence signal of about 100 times the signal obtained without
Mutt indicating the generation of FARP-reactive sites following Mutt
treatment (Figure 7). The chemiluminescence results agree with the
fragmentation results obtained by gel electrophoresis (Figure 5).
Table-1: Sequences of the synthesized oligonucleotides
1. B-5'-GTC TCC CAT CCA AGT ACT AAC CAG GCC CGA CCC
TGC TTG GCT TCC GAT T-3' (SEQ ID NO:1)
2. B-5'-AAT CGG AAG CCA AGC AGG GTA GGG CCT GGT TAG
TAC TTG GAT GGG AGA C-3' (SEQ ID N0:2)
3. B-5'-AAT CGG AAG CCA AGC AGG GTA GGG CCT GGG TAG
TAC TTG GAT GGG AGA C-3' (SEQ ID N0:3)
1 and 2 are complementary and form a homoduplex. 1 and 3 form a
heteroduplex with an A/G mismatch at position 20. On a separate set of
oligonucleotides, a biotin molecule (B) was incorporated at 5' end during
synthesis.
EXAMPLE 2
BARD - BASED DETECTION OF MISMATCHES FORMED VIA
SELF-COMPLEMENTATION OF SINGLE - STRANDED M 13 DNA.
Samples of M 13 single stranded DNA that contain approximately 1
Mutt-recognizable mismatch per 2,500 bases were treated with Mutt to
generate aldehyde -containing reactive sites appropriate for reaction with
BARP. Nominal gel electrophoretic studies as well as BARP-based
-41-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
chemiluminescent studies were then preformed. Control samples used were:
Single stranded M 13 without enzymatic treatment; Double stranded M 13
DNA without any mismatches and no enzyme treatment; and double
stranded M 13 DNA without mismatches and enzyme. Figure 8 (A and B)
shows the results of both methods of detection. Figure 8A (luminescence
studies) show that only when mismatches are present (single stranded M 13)
and Mutt is used is there a chemiluminescence signal. In agreement, gel
electrophoresis (Figure 8B) shows cuts in M 13 are only generated under the
same conditions. It can be seen that there is good agreement among the two
methods. As described, the method is highly specific for mismatch -
containing DNA, i.e. DNA without mismatches, or DNA with mismatches but
no Mutt generate no signals.
EXAMPLE 3
DETECTION AND ISOLATION OF DNA CONTAINING BASE - SUBSTITUTION
MUTATIONS: DETECTION OF A SINGLE A-TO-C TRANSVERSION
ENGINEERED IN A P53 GENE WITHIN A 7091 - LONG PLASMID.
The ability of the present technology (A.L.B.U.M.S) to detect base
mismatches (demonstrated in previous examples) is directly applicable to
detection of base substitution mutations. For example, a standard procedure
to generate mismatches at the positions of mutations in DNA, is to mix
mutation - containing DNA with wild - type DNA. Upon heating and re-
hybridization of the mixture, heteroduplexes with mismatches are generated
at the positions of mutations (Figure 1), which can then be detected with
high sensitivity and specificity as demonstrated in example 1.
To isolate mutation - containing DNA from normal DNA, following
BARP - labeling of the generated aldehydes at positions of mismatches
(Figure 1) the DNA is immobilized on neutravidin-coated microplates,
followed by exhaustive washing to remove the homoduplex DNA. As a result,
only BARD-containing DNA is retained on the plates, thereby isolating
mutant DNA.
To recover the purified mutation - containing DNA from the
microplate, the samples can be either heated 2 min at 96°C or treated 1
min
with NaOH to denature the DNA and recover the non-covalently modified
- 42 -
CA 02321119 2000-08-15
PCTNS99/03821
WO 99/42622
strand, which is then used for amplification via PCR. The following section
detail the procedure.
A 7,091 by long plasmid that incorporates the full-length human
cDNA p53 sequence (1,691 bp) was engineered to contain base subsitutions,
via site-specific mutagenesis. The present technology was used to detect a
known A-to-C base substitution mutation engineered in codon 378 within
the plasmid-incorporated p53. Circular plamids ( 1 llg) containing mutant
p53 genes were treated with a 5'- CG/CG-3' cutting enzyme (BstU I, Sigma,
1 unit, 1 h, 37°C) to generate linear fragments 0400 to 2,500 bp),
followed
by a 10 minute, 70°C treatment to inactivate the enzyme. The mutant-
containing sample ( 1 fig) was mixed ( 1:1 ) with a similarly treated normal
p53-containing sample, heated (96°C, 2 minutes) and hybridized
overnight,
at 65°C to generate A/G (25 %), and T/C (25 %) mismatches at p53 codon
378, as well as homoduplex p53 and plasmid fragments.
To detect the presence of the mutation via ALBUMS, 100 ng of the
mismatch-containing DNA mixture (p53 plus plasmid fragments) was treated
exactly as described for the M 13 treatment in example 2: (a) hydroxylamine
treatment and removal, (b) Mutt treatment and BARD-binding, (c)
fluoresceination and (d) binding to neutravidin plates and
chemiluminescence detection. Figure 9A demonstrates that strong signals
are observed when the mutation is present, while background signals are
obtained from normal p53-containing plasmid (i.e. complete lack of false
positives). Figure 9B shows variation of signals versus DNA amount applied
on microplates. These data represent an average of 4 independent
experiments.
In conclusion, the present technology (A.L.B.U.M.S) allows a sensitive
and specific detection of 1 base substitution mutation within a 7,091 bp-
long, p53-containing plasmid with a virtual absence of false positives
(defined as signal when no mismatch is present, Figure 9A). Unequivocal
detection of a single base substitution within a 7,091 -long plasmid cannot
easily be conducted with any of the existing methodologies (Nollau P and
Wagener C. Clinical Chemistry 43: 1114-1128 (1997)). The present method
on the other hand can detect the mutation on a microplate with minimal
sample (< 100 ng) and effort involved. Following formation of heteroduplexes,
the procedure is currently completed in 6 hours, requires no special
-43-
CA 02321119 2000-08-15
WO 99/42622 PCT/US99/03821
equipment or laborious handling and can be automated on microplates so
that 96 samples can be examined at once. To achieve a similar result using
conventional sequencing would not be possible (Primrose SB, Principles of
Genorne Analysis, Chapter 5, Sequencing Methods and Strategies, p 125,
Second Edition, Blackwell Science Ltd., Oxford, UK).
EXAMPLE 4
COMPARISON OF SMALL VERSUS LARGE LIGAND COMPOUNDS IN
BINDING TO Mutt - OR TDG- GENERATED REACTIVE SITES IN DNA:
IO SYNTHESIS AND ADVANTAGE OF AED VERSUS BARP AND FARP.
CHEMILUMINESCENCE SIGNALS BY AED.
(a) To synthesize AED, O-(Carboxymethyl)hydroxylamine
hydrochloride was conjugated to ethylenediamine (Aldrich) in distilled water
using 1-Ethyl- 3-(3-(dimethylamino)propyl] carbodiimide (EDAC) as the
coupling reagent. An 100-fold excess of ethylenediamine over
O-(Carboxymethyl)hydroxylamine hydrochloride was utilized during the
reaction to allow preferential coupling of ethylenediamine to the carboxyl
groups. The conditions for the catalysis of this reaction by EDAC is well
known to those skilled in the art. TLC analysis and purification on silica gel
with CHCIs:CHsOH:CHaCOOH in a 70:20:5 ratio indicated the product at an
Rc of 0.2-0.25. The certificate of analysis provided 1H NMR data consistent
with the AED structure provided earlier.
(b) The ability of hydroxylamine - based compounds (e.g. FARP,
AED, BARP, or methoxyamine) to bind reactive sites in DNA can be tested
with a simple experiment. It is well known that, if hydroxylamine -
compounds (such as methoxyamine) are covalently bound to aldehyde -
containing abasic sites in DNA, then treatment with alkali (NaOH) cannot
generate a strand break at the position of base loss (-otherwise a cut is
generated). This simple observation allows direct testing of ligand binding to
DNA following Mutt - treatment of the nucleic acid (Figure l0A) or TDG -
treatment of nucleic acid (Figure lOB). Mismatch - containing single
stranded M 13 DNA was subjected to Mutt to generate aldehyde containing
abasic sites, and then alkali - treated to generate fragments at the positions
of mismatches. Lane 2, in Figure l0A (agarose gel stained with ethidium
bromide and photographed under UV light) demonstrates the generated
-44-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
fragments. In lanes 3, 4, 5, and 6, during Mutt incubation the following
ligand compounds were also included: 5 mM methoxyamine, 5 mM AED, 10
mM AED or 5 mM BARP respectively. As expected, the very low molecular
weight compound methoxyamine prevents formation of any fragments,
indicating a 100% binding to all reactive sites formed. Also, AED (bands D
and E) demonstrates an almost complete binding to the reactive sites,
especially when 10 mM is used (Lane E). In contrast, BARD can only prevent
to a very small degree the formation of bands, indicating a very low (<5%)
binding affinity to the reactive sites.
Similarly, in Figure lOB, the TDG enzyme was used (TDG recognizes
mismatched thymine and generates an aldehyde at that position following
excision of thymine). Oligonucleotides with a G/T mismatch were
synthesized (lanes 1, 2, oligos alone) and exposed to TDG in the absence
(lanes 3) or in the presence of 5 mM methoxyamine (lane 4), 5 mM BARP
(lane 5j, 5 mM AED (lane 6) or 0.5 mM FARP (lane 7). It can be seen that
the cuts generated by TDG (lane 3 lower band) are not present when
methoxyamine (lane 4) or AED (lane 6) are included in the reaction,
demonstrating the binding of these compounds to the mismatches. BARD
and FARP on the other hand (lanes 5 and 7) demonstrate significantly lower
bindirzg, since the lower band is present.
In conclusion: (a) AED is almost as efficient as methoxyamine ( 100%)
in binding the Mutt - generated reactive sites. (Methoxyamine itself however
cannot be used in the present application because, unlike AED, following
binding it allows no further derivatization as it has no secondary binding
site
available for antibody binding). (b) BARP only shows little (<5%) binding;
despite that, and because the present method is extremely sensitive, high
chemiluminescence signals are still generated with BARP when mismatches
are present, as shown in the previous example. The same is valid for FARP.
The ability of DNA-bound AED to be recognized by a secondary ligand
and then by an antibody, as described in the Detailed Description section of
this invention was demonstrated by the following. The free primary amine (-
NH2 group) of AED was covalently bound to biotin by addition of 1 mM
biotin-LC-succinimidyl ester (Pierce) in 0.1 M sodium bicarbonate, pH=8.5
for 2h. The conjugate was purified by ultracentrifugation through 2 G25
filters (Pharmacia), fluoresceinated by using the Mirus fluoresceination
-45-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/0382t
reagent (Panvera inc, see example 1 ) and then applied on neutravidin
microplates. Addition of antifluorescein-AP antibody generated a strong
chemiluminescence signal (Figure 12) in the sample treated with Mutt
enzyme (i.e. aldehydes were generated), but not in the sample not-treated
with Mutt (aldehydes not generated).
EXAMPLE 5
LABELING OF MISMATCHES WITH FARP, BARP or AED: INACTIVATION OF
ENZYMATIC ACTION DURING LABELING.
A DNA sample containing mismatches is dissolved in a buffered
solution and treated with a repair glycosylase, either Mutt or TDG ( 1 unit
enzyme per ug DNA). The reaction is incubated at 37°C for 1 hour. Upon
completion of the reaction with Mutt or TDG, the solution is cooled to
15°C,
to arrest enzymatic activity. FARP is added to the sample and allowed to
react for 30 minutes at 15°C. At the end of the 30 minute incubation
with
FARP, the reaction solution is suddenly heated to 70°C for two
minutes to
inactivate the enzyme. The sample of DNA is now ready for purification and
detection as previously described. Alternatively, instead of heating to
70°C
the enzyme can be solubilized and removed via a standard phenol-
chloroform extraction, or via addition of Proteinase K (0.1 mg/ml, 2h,
37°C).
EXAMPLE 6
STRATEGY TO UTILIZE DNA CHIPS FOR DETECTION OF BOTH INHERITED
POLYMORPHISMS AND MUTATIONS, AS WELL AS ACQUIRED MUTATIONS
FROM CANCER SAMPLES.
The ability to derive both inherited and acquired genetic alterations in
a single step over 6800 genes with the present procedure, using the
Affymetrix array as an example, is described below.
Inherited single nucleotide polymorphisms (SNPs) are estimated to be
present in the two alleles of each gene with a frequency of -1:1000 bases.
When an SNP in the coding sequence causes a debilitating change in the
protein, heterozygous mutations arise which could result to early onset of
cancer (e.g. the Li-Fraumeni syndrome). When cDNA from normal cells is
melted and self hybridized, mismatches will occur at positions of
heterozygocities and SNPs, whenever both alleles are expressed, which will
-46-
CA 02321119 2000-08-15
WO 99/42b22 PCT/US99/03821
be detectable by the present technology (A.L.B.U.M.S) and would display
positive on the DNA arrays. Because SNPs among alleles occur at a high
frequency (-1:1000 bp) it is possible that within every single gene
(average~2,000 bp) there is one or more SNPs. Therefore, if both paternal
and maternal alleles are transcribed, self hybridizing cDNA from whole genes
would be expected to result in one or more mismatches per gene, as a result
of allelic cross-hybridization. All array elements would then display
positive,
resulting to trivial information. By digesting the cDNA to ~ 100-200 by
pieces prior to ALBUMS genotypic selection (as described in example 3) the
problem is avoided: Most fragments are likely to contain none, or
occasionally one inherited SNP. ALBUMS will select mismatch - containing
fragments, and array elements that score positive will be only those
capturing a 100-200-mer gene fragment with an SNP.
Acquired mutations can be detected by following the same strategy,
and by using cancer samples from the same individual as the normal
sample. Again, by self hybridizing cDNA from cancer samples and
fragmenting to 100-200-mers, it is likely that most fragments will contain
none, or occasionally one inherited SNP, or very occasionally one acquired
mutation. Array elements that score positive will be those corresponding to
genes that contain either inherited or acquired mutations, but rarely both.
An example of using the high resolution Affymetrix array (described
earlier) to detect genetic alterations in parallel normal and cancer samples
is
displayed in Figure 12. cDNA from normal tissue is melted and self
hybridized to generate mismatches (Figure 12), then digested with
appropriate enzymes to generate 100-200-mers and add primers; then the
present technology, (ALBUMS), utilizing one of the probes (FARP, AED or
BARD) selects the mismatches, PCR amplifies them and these are applied on
the Affymetrix array: The mutation-containing 200-mers isolated via
ALBUMS will cause certain 25-mer array elements to display positive,
thereby identifying both the gene and the approximate (t 100-200 bp)
location of an inherited polymorphism among the two alleles.
Next, cDNA from the cancer sample is melted, self hybridized and
processed similarly. Acquired mutations will show up as positive array
elements that are negative on the normal tissue array. Acquired mutations
scored on the same gene as an inherited mutation provide candidate genes
-47-
CA 02321119 2000-08-15
WO 99/42622 PCTNS99/03821
to be examined for loss of heterozygocity, using existing methodologies.
Finally, cDNA from cancerous cells will be cross-hybridized to cDNA from
normal cells and the procedure will be repeated (not illustrated in Figure
12).
This will detect acquired mutations in those genes that express a single
allele in their mRNA, which would not be detected by self hybridization
alone.
The use of the Clontech array will provide similar information to the
Affymetrix array. However, this array would be used with fewer genes and
with smaller 'resolution', since the array elements contain 500 bases-long
cDNA and it is possible that certain elements will capture both inherited
SNPs and acquired mutations, thereby providing unclear information. On
the other hand these arrays are simpler to use and do not require the
fluorescent laser scanner, hence they are currently more accessible to users.
-48-
