 |
Description  |
|
|
FIELD OF THE INVENTION
The invention relates generally to methods for determining the sequence of
a nucleic acid, and more particularly, to methods for determining the
sequence of a double or single stranded deoxyribonucleic acid (DNA) by
hybridization of multiple oligonucleotide probes.
BACKGROUND
The ability to determine DNA sequences is crucial for understanding the
function and control of genes and for applying many of the basic
techniques of molecular biology. Native DNA consists of two linear
polymers, or strands of nucleotides. Each strand is a chain of nucleosides
linked by phosphodiester bonds. The two strands are held together in an
antiparallel orientation by hydrogen bonds between complementary bases of
the nucleotides of the two strands: deoxyadenosine (A) pairs with
thymidine (T) and deoxyguanosine (G) pairs with deoxycytidine (C).
Presently there are two basic approaches to DNA sequence determination: the
dideoxy chain termination method, e.g. Sanger et al, Proc. Natl. Acad.
Sci., Vol. 74, pgs. 5463-5467 (1977); and the chemical degradation method,
e.g. Maxam et al, Proc. Natl. Acad. Sci., Vol. 74, pgs. 560-564 (1977).
The chain termination method has been improved in several ways, and serves
as the basis for all currently available automated DNA sequencing
machines, e.g. Sanger et al, J. Mol. Biol., Vol. 143, pgs. 161-178 (1980);
Schreier et al, J. Mol. Biol., Vol. 129, pgs. 169-172 (1979); Smith et al,
Nucleic Acids Research, Vol. 13, pgs. 2399-2412 (1985); Smith et al,
Nature, Vol. 321, pgs. 674-679 (1987); Prober et al, Science, Vol. 238,
pgs. 336-341 (1987), Section II, Meth. Enzymol., Vol. 155, pgs. 51-334
(1987), and Church et al, Science, Vol 240, pgs. 185-188 (1988).
Both the chain termination and chemical degradation methods require the
generation of one or more sets of labeled DNA fragments, each having a
common origin and each terminating with a known base. The set or sets of
fragments must then be separated by size to obtain sequence information.
In both methods, the DNA fragments are separated by high resolution gel
electrophoresis. Unfortunately, this step severely limits the size of the
DNA chain that can be sequenced at one time. Non-automated sequencing can
accommodate a DNA chain of up to about 500 bases under optimal conditions,
and automated sequencing can accommodate a chain of up to about 300 bases
under optimal conditions, Bankier et al, Meth. Enzymol., Vol. 155, pgs.
51-93 (1987); Roberts, Science, Vol. 238, pgs. 271-273 (1987); and Smith
et al, Biotechnology, Vol. 5, pgs. 933-939 (1987).
This limitation represents a major bottleneck for many important medical,
scientific, and industrial projects aimed at unraveling the molecular
structure of large regions of plant or animal genomes, such as the project
to sequence all or major portions of the human genome, Smith et al,
Biotechnology (cited above).
In addition to DNA sequencing, nucleic acid hybridization has also been a
crucial element of many techniques in molecular biology, e.g. Hames et al,
eds., Nucleic Acid Hybridization: A Practical Approach (IRL Press,
Washington, D.C., 1985). In particular, hybridization techniques have been
used to select rare cDNA or genomic clones from large libraries by way of
mixed oligonucleotide probes, e.g. Wallace et al, Nucleic Acids Research,
Vol. 6, pgs. 3543-3557 (1979), Proc. Natl. Acad. Sci., Vol. 80, pgs.
5842-5846 (1983). Nucleic acid hybridization has also been used to
determine the degree of homology between sequences, e.g. Kafatos et al,
Nucleic Acids Research, Vol. 7, pgs. 1541-1552 (1979), and to detect
consensus sequences, e.g. Oliphant et al, Meth. Enzymol., Vol. 155, pgs.
568-582 (1987). Implicit to all of these applications is the notion that
the known probe sequences contain information about the unknown target
sequences. This notion apparently has never been exploited to obtain
detailed sequence information about a target nucleic acid.
In view of the limitations of current DNA sequencing methods, it would be
advantageous for the scientific and industrial communities to have
available an alternative method for sequencing DNA which (1) did not
require gel electrophoretic separation of similarly sized DNA fragments,
(2) had the capability of providing the sequence of very long DNA chains
in a single operation, and (3) was amenable to automation.
SUMMARY OF THE INVENTION
The invention is directed to a method for determining the nucleotide
sequence of a DNA or an RNA molecule using multiple mixed oligonucleotide
probes. Sequence information is obtained by carrying out a series of
hybridizations whose results provide for each probe the number of times
the complement of the probe's sequence occurs in the RNA or DNA whose
sequence is to be determined. The nucleotide sequence of the RNA or DNA is
reconstructed from this information and from a knowledge of the probes'
sequences. The nucleic acid whose sequence is to be determined is referred
to herein as the target sequence.
The mixed oligonucleotide probes of the invention are selected from a set
whose members' sequences include every possible complementary sequence to
subsequences of a predetermined length within the target sequence. The
series of hybridizations are separately carried out such that one or more
of the probes selected from the set are combined with known quantities of
the target sequence, e.g. on a nitrocellulose filter, or like substrate,
under conditions which substantially allow only perfectly matched probe
sequences to hybridize with the target sequence. Probe sequences having
mismatched bases are substantially removed, e.g. by washing, and the
quantity of perfectly matched probe remaining hybridized to the target
sequence is determined.
In one embodiment of the invention, the set of probes comprises four
subsets. Each of the four subsets contains probes representing every
possible sequence, with respect to the size of the probe (which is
predetermined), of only one of the four bases. For example, the first
subset can contain probes where every possible sequence of G is
represented; the second subset can contain probes where every possible
sequence of T is represented; and so on for C and A. If the probes were
each 8 bases long, a member probe of the adenosine subset can be
represented as follows:
##EQU1##
Formula I
The symbol
##EQU2##
means that any of the bases C, G, or T may occupy the position where the
symbol is located. Thus, the above probe has a multiplicity, or
degeneracy, of 1.times.1.times.3.times.3.times.1.times.3.times.3.times.1
or 81. When it is clear from the context which subset is being considered,
the above notation will be simplified to AA00A00A, where A represents
deoxyadenosine and 0 represents the absence of deoxyadenosine.
Preferably, base analogs are employed in the oligonucleotide probes whose
base pairing characteristics permit one to reduce the multiplicity of the
probe. For example, in the probe of Formula II, because deoxyinosine (I)
forms nearly equally strong base pairs with A and C, but forms only a weak
or destabilizing base pair with either G or T, deoxyinosine can replace G
and T in the probe, Martin et al, Nucleic Acids Research, Vol. 13, pgs.
8927-8938 (1985). Thus, a probe equivalent to that of Formula I, but which
has a much lower multiplicity (i.e. only 16) can be represented as
follows:
##EQU3##
Formula II
Generally, base analogs are preferred which form strong base pairs (i.e.,
comparable in binding energy to the natural base pairs) with two or three
of the four natural bases, and a weak or destabilizing base pair with the
complement of a fixed base (defined below). Such base analogs are referred
to herein as degeneracy-reducing analogs.
It is not critical that the probes all have the same length, although it is
important that they have known lengths and that their sequences be
predetermined. Generally, the probes will be fixed at a predetermined
number of positions with known bases (not necessarily of the same kind),
e.g. as the A in Formula I, and the remaining positions will each be
filled by a base randomly selected from a predetermined set, e.g. T, G,
and C as in Formula I, or I and C as in Formula II. The positions in a
probe which are nondegenerate in their base pairing, i.e. have only a
single natural base, are referred to herein as fixed positions. The bases
occupying fixed positions are referred to herein as fixed bases. For
example, the fixed bases in the probes of Formulas I and II are
deoxyadenosine at positions one, two, five, and eight with respect to the
3' end of the probe.
Generally, sets and/or subsets of the invention each contain at least one
probe having a sequence of fixed and non-fixed positions equivalent to
that of each permutation of a plurality of fixed and non-fixed positions
less than or equal to the length of the probe. That is, an important
feature of the invention is that the probes collectively contain
subsequences of fixed and non-fixed positions (which may be the total
length of the probe, as is the case for the probe sequences of Appendix I)
which correspond to every possible permutation of fixed and non-fixed
positions of each of a plurality of combinations of fixed and non-fixed
positions, the plurality including combinations containing from zero to
all fixed positions. For example, consider a subset of probes of the
invention that consists of 8-mer probes whose fixed positions contain only
deoxyadenosine and whose initial (i.e., 3'-most) position is fixed. The
probes of Formulas I and II are members of such a subset. Within such a
subset, there is a leas: one probe having a subsequence of fixed and
non-fixed positions in positions 2 through 8 which corresponds to each
possible permutation of fixed and non-fixed positions for subsequences
having no fixed positions (one such permutation: A0000000), one fixed
position (seven such permutations, e.g. A000A000), two fixed positions
(twenty-one such permutations, e.g. A00AA000), three fixed positions
(thirty-five such permutations, e.g. A0000AAA), four fixed positions
(thirty-five such permutations, e.g. A0AAAA00), five fixed positions
(twenty-one such permutations, e.g. AAA00AAA), six fixed positions (seven
such permutations, e.g. AAAA0AAA), and seven fixed positions (one such
permutation: AAAAAAAA). Thus, the subset has at least
1+7+21+35+35+21+7+1=128 members.
The presence of one or more predetermined known sequence regions in the
target sequence facilitates the reconstruction of the target sequence.
Accordingly, in a preferred embodiment, the target sequence contains one
or more regions of known sequence, these regions being referred to herein
as known sequence regions. More preferably, the target sequence contains a
first and a second known sequence region, the first and second known
sequence regions being positioned on opposite ends of the region of the
target sequence containing the unknown sequence of nucleotides. This
unknown sequence of nucleotides is referred to herein as the unknown
sequence region. Most preferably, the first and second known sequence
regions are at least the length of the longest probe sequence.
Besides a method of sequencing nucleic acids, the invention also includes
kits and compositions including vectors and mixtures of oligonucleotides
for constructing probes for use in accordance with the method of the
invention. Oligonucleotide mixtures of the invention have sequences of
fixed and non-fixed bases. Preferably, such mixtures comprise sequences
represented by the formula:
X.sub.i.sub.1 X.sub.i.sub.2 X.sub.i.sub.3. . . X.sub.i.sub.n
wherein:
i.sub.1 through i.sub.n are separately either 0 or 1;
X.sub.0 represents a non-fixed base;
X.sub.1 represents a fixed base; and
n is in the range of from 6 to 10.
More preferably, whenever X.sub.1 represents A, X.sub.0 is selected from
the group consisting of C, G, T, and degeneracy-reducing analogs thereof;
whenever X.sub.1 represents C, X.sub.0 is selected from the group
consisting of A, G, T, and degeneracy-reducing analogs thereof; whenever
X.sub.1 represents G, X.sub.0 is selected from the group consisting of A,
C, T, and degeneracy-reducing analogs thereof; and whenever X.sub.1
represents T, X.sub.0 is selected from the group consisting of A, C, G,
and degeneracy-reducing analogs thereof.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 illustrates the general problem of sequence reconstruction by
showing how a 21-mer sequence can be reconstructed by four subsets of
4-mer probes.
FIG. 2 is a flow chart diagrammatically illustrating a preferred
reconstruction algorithm.
DETAILED DESCRIPTION OF THE INVENTION
The present invention provides a method of nucleotide sequence
determination which takes advantage of the extremely specific
hybridization behavior of oligonucleotides. Easily synthesized
oligonucleotide probes from a predetermined set are separately hybridized
to known quantities of a target sequence so that the number of copies of
perfectly matched probe on the target sequence can be determined. Once
such a determination is made for each probe of the predetermined set, the
target sequence can be reconstructed by a mathematical algorithm.
I. Composition and Labeling of the Probes
Mixed oligonucleotide probes for the invention are preferably synthesized
using an automated DNA synthesizer, e.g. Applied Biosystems (Foster City,
CA) models 381A or 380B, or like instrument. At nonfixed positions
mixtures of the appropriate nucleotide precursors are reacted with the
growing oligonucleotide chain so that oligonucleotides having different
bases at that position are synthesized simultaneously, e.g. as disclosed
by Wallace et al, Nucleic Acids Research, Vol. 6, pgs. 3543-3557 (1979),
and Oliphant et al, Meth. Enzymol., Vol. 155, pgs. 568-582 (1987).
The probes may be synthesized by way of any of the available chemistries,
e.g. phosphite triester, Beaucage et al, Tetrahedron Letters, Vol. 22,
pgs. 1859-1862 (1981); Caruthers et al, U.S. Pat. Nos. 4,415,723,
4,458,066, and 4,500,707; phosphotriester, Itakura, U.S. Pat. No.
4,401,796; hydrogen phosphonate, e.g. Froehler et al, Nucleic Acids
Research, Vol. 14, pgs. 5399-5407 (1986); or the like. It is not critical
that the bases of the probes be linked by phosphate esters. Analogs of the
natural phosphate ester linkages can be used, e.g. Jones, Int. J. Biol.
Macromol., Vol. 1, pgs. 194-207 (1979), reviews synthetic analogs of
phosphate esters; and Ts'o et al, U.S. Pat. No. 4,469,863 describe the
synthesis of alkyl and aryl phosphonate analogs. Probes of the invention
can also comprise oligoribonucleotides, e.g. Ebe et al., DNA, Vol. 6, pgs.
497-504 (1987).
Once synthesized, the oligonucleotides are purified for labeling by well
known techniques, usually HPLC or gel electrophoresis, e.g. Applied
Biosystems DNA Synthesizer Users Bulletin, Issue No. 13-Revised (Apr. 1,
1987); or Gait, ed., Oligonucleotide Synthesis: A Practical Approach (IRL
Press, Washington, D.C., 1984). Alternatively, the oligonucleotides for
the probe can be purified by affinity chromatography using
oligonucleotides complementary to the probe sequences covalently attached
to a suitable solid phase support, e.g. Gilham, J. Amer. Chem. Soc., Vol.
86, pg. 4982 (1964); and Kremsky et al, Nucleic Acids Research, Vol. 15,
pgs. 3131-3139 (1987).
Selecting the lengths of the probes is an important aspect of the
invention. Several factors influence the choice of length for a given
application, including (1) the ease with which hybridization conditions
can be manipulated for preferentially hybridizing probes perfectly matched
to the target sequence, (2) the ability to distinguish between roughly
integral amounts of perfectly matched probe hybidized to the target
sequence (e.g. if the probe is relatively long so that the expected
frequency of probe sequences perfectly complementary to the target is low,
one may be required to distinguish (for example) between amounts of probe
in the range of 10, 20, or 30 picomoles--to infer that 1, 2, or 3 copies
of the probe are present on the target; if the probe is relatively short
so that the expected frequency of probe sequences perfectly complementary
to the target is high, one may be required to distinguish (for example)
between amounts of probe in the range of 110, 120, or 130 picomoles--to
infer that 11, 12, or 13 copies of the probe are present on the target;
since the fractional differences between the latter quantities are small,
there may be less confidence in the inferred copy number); (3) whether
probe multiplicity permits hybridization with reasonable Cot values
(longer probes are more degenerate than shorter probes, and require higher
Cot values for hybridization, the converse is true of shorter probes); (4)
the practicality of carrying out separate hybridizations for each type of
probe (longer probes give rise to larger sets of probes, e.g. probes
having permutations of fixed and non-fixed bases in seven positions give
rise to a set of 2.sup.7 -1, or 127 probes, for each kind of base, for
eight positions this number is 2.sup.8 -1, or 255, and so on); and (5) the
tractability of the sequence reconstruction problem (the greater the
number of copies of each probe type on the target sequence--which is the
tendency if shorter probes are employed, the more difficult the
reconstruction problem). Probe sizes in the range of 7 to 11 bases are
preferred. More preferably, probes sizes are in the range of 8 to 10
bases, and most preferably, probe sizes are in the range of 8 to 9 bases.
Preferably, degeneracy-reducing analogs are employed at the non-fixed
positions of the probes to reduce probe multiplicity, or degeneracy. Many
synthetic and natural nucleoside and nucleotide analogs are available for
this purpose, e.g. Scheit, Nucleotide Analogs (John Wiley & Sons, N.Y.,
1980). For example, degeneracy-reducing analogs include deoxyinosine for
use in cytosine or adenosine probes to replace G and T at non-fixed
positions, 2-aminopurine for use in cytosine or guanosine probes to
replace A and T at non-fixed positions, and N.sup.4 -methoxydeoxycytidine,
N.sup.4 -aminodeoxycytidine, or 5-fluorodeoxyuridine for use in adenosine
or guanosine probes to replace T and C. Use of deoxyinosine in
oligonucleotide probes is disclosed by Martin et al (cited above); Seela
et al, Nucleic Acids Research, Vol. 14, pgs. 1825-1844 (1986); Kawase et
al, Nucleic Acids Research, Vol. 14, pgs. 7727-7737 (1986); Ohtsuka et al,
J. Biol. Chem., Vol. 260, pgs. 2605-2608 (1985); and Takahashi et al,
Proc. Natl. Acad. Sci., Vol. 82. pgs. 1931-1935 (1985). Accordingly, these
references are incorporated by reference. Deoxyinosine phosphoramidite
precursors for automated DNA synthesis are available commercially, e.g.
Applied Biosystems (Foster City, Calif.). The synthesis of N.sup.4
-methoxycytidine and its incorporation into oligonucleotide probes is
disclosed by Anand et al, Nucleic Acids Research, Vol. 15, pgs. 8167-8176
(1987). The synthesis of 2-aminopurine and its incorporation into
oligonucleotide probes is disclosed by Eritja et al, Nucleic Acids
Research, Vol. 14, pgs. 5869-5884 (1986). The synthesis of
5-fluorodeoxyuridine and its incorporation into oligonucleotide probes is
disclosed by Habener et al, Proc. Natl. Acad. Sci., Vol. 85, pgs.
1735-1739 (1988). And the preparation of N.sup.4 -aminodeoxycytidine is
disclosed by Negishi et al, Nucleic Acids Research, Vol. 11, pgs.
5223-5233 (1983). Accordingly, these latter four references are
incorporated by reference.
Nucleoside analogs are also employed in the invention to reduce the
differences in binding energies between the various complementary bases.
In particular, 2-aminoadenine can replace thymine in either the probe or
target sequences to reduce the binding energy differences between A-T
nucleoside pairs and G-C nucleoside pairs, e.g. Kirnos et al (cited
above), Chollet et al (cited above), and Cheong et al, Nucleic Acids
Research, Vol. 16, pgs. 5115-5122 (1988). Procedures for synthesizing
oligonucleotides containing 2-aminoadenine are disclosed by Chollet et al
(cited above); Gaffney et al, Tetrahedron, Vol. 40, pgs. 3-13 (1984), and
Chollet et al Chemica Scripta, Vol. 26, pgs. 37-40 (1986). Likewise,
2-amino-2'-deoxyadenosine can replace deoxyadenosine to increase the
binding Dihn et al, Proc. Natl. Acad. Sci., Vol. 82, pgs. 7510-7514
(1985).
In some embodiments, it may be preferable to replace a more degenerate
probe with several less degenerate probes which collectively are capable
of obtaining the same information about the target sequence. For example,
consider the 9-mer probe A00000000. This probe can be replaced by the
three less degenerate probes A0000000C, A0000000G, and A0000000T. Thus, at
the cost of two additional hybridizations, the degeneracy of the most
degenerate probe in the set is reduced from 256 to 128 (assuming the use
of deoxyinosine at non-fixed positions).
The oligonucleotides of the invention can be labeled in a variety of ways
to form probes, including the direct or indirect attachment of radioactive
moieties, fluorescent moieties, electron dense moieties, and the like. It
is only important that each sequence within a probe be capable of
generating a signal of the same magnitude, so that quantitative
measurements of probe number can be made. There are several means
available for derivatizing oligonucleotides with reactive functionalities
which, permit the addition of a label. For example, several approaches are
available for biotinylating an oligonucleotide so that fluorescent,
enzymatic, or electron density labels can be attached via avidin: Broken
et al, Nucleic Acids Research, Vol. 5, pgs. 363-384 (1978), disclose the
use of ferritin-avidin-biotin labels; Chollet et al, Nucleic Acids
Research, Vol. 13, pgs. 1529-1541 (1985), disclose biotinylation of the 5'
termini of oligonucleotides via an aminoalkylphosphoramide linker arm; and
Bayer et al, Methods of Biochemical Analysis, Vol. 26, pgs. 1-45 (1980)
provide a general review of the use of the avidin-biotin complex. Several
methods are also available for synthesizing amino-derivatized
oligonucleotides which are readily labeled by fluorescent or other types
of compounds derivatized by amino-reactive groups, such isothiocyanate,
N-hydroxysuccinimide, or the like, e.g. Connolly, Nucleic Acids Research,
Vol. 15, pgs. 3131-3139 (1987); Gibson et al, Nucleic Acids Research, Vol.
15, pgs. 6455-6467 (1987); Miyoshi et al, U.S. Pat. No. 4,605,735. Methods
are also available for synthesizing sulfhydryl-derivatized
oligonucleotides which can be reacted with thiol-specific labels, e.g.
Connolly, Nucleic Acids Research, Vol. 13, pgs. 4485-4502 (1985); and
Spoat et al, Nucleic Acids Research, Vol. 15, pgs. 4837-4848 (1987). A
comprehensive review of methodologies for labelling DNA probes is provided
by Matthews et al, Anal. Biochem., Vol 169, pgs. 1-25 (1988).
Alternatively, probes of the invention can generate a signal indirectly
using a detection scheme disclosed by Sutherland et al, European patent
application publication number 245206. In this embodiment, probe sequences
are anchored to the end of a fiber optic so that they are free to
hybridize to target sequences, which have been broken up into segments and
denatured, e.g. by random endonuclease digestion and boiling, to
facilitate hybridization. (In this embodiment the probe sequences, not the
target sequences, are anchored to a solid phase substrate). Upon
hybridization of the probe and target sequences the duplex is detected by
fluorescence generated by an intercalating dye, e.g. ethidium bromide,
which is transmitted along the fiber optic to a photodetector.
In one preferred embodiment of the invention, the probes are biotinylated,
e.g. as taught by Chollet et al (cited above), and after hybridization and
washing, the probes are reacted with avidin or streptavidin conjugated to
a fluorescent dye or one or more enzyme molecules for generating a
colorimetric signal. Preferred fluorescent dyes include fluorescein,
tetramethylrhodamine, rhodamine, phycorythrin, Texas Red, and the like.
In another preferred embodiment, the oligonucleotides of the invention are
radioactively labeled with .sup.32 P using standard protocols, e.g.
Maniatis et al, Molecular Cloning: A Laboratory Manual (Cold Spring Harbor
Laboratory, N.Y., 1982); Current Protocols in Molecular Biology, Unit 6.4
(John Wiley & Sons, N.Y., 1987); or Maxim and Gilbert, Meth. Enzymol.,
Vol. 65, pgs. 499-560 (1980). .sup.32 P-labeled probes of the invention
are preferably applied to tangent DNAs anchored to nitrocellulose, nylon,
or the like, at a concentration in the range of about 1-10 ng/ml, and more
preferably, in the range of about 1-5 ng/ml. The specific activities of
the probe are preferably in the range of about 1-5.times.10.sup.6 cpm/ml.
II. Hybridization
The hybridizations of the probes to the target sequence are carried out in
a manner which allows mismatched probe sequences and nonspecifically bound
probe sequences to be separated from the duplexes formed between the
perfectly matched probe sequences and the target sequence. Usually the
separation is carried out by a washing step. Preferably, the first step in
the hybridizations is to anchor the target sequence so that washes and
other treatments can take place with minimal loss of the target sequences.
The method selected for anchoring the target sequence depends on several
factors, including the length of the target, the method used to prepare
copies of the target, and the like. For example, DNA sequences less than
200-300 bases in length do not bind efficiently to nitrocellulose,
Meinkoth et al, Anal. Biochem., Vol. 138, pgs. 267-284 (1984); thus, for
shorter target sequences alternatives to nitrocellulose may be preferable,
such as covalent binding to diazotized paper, e.g. Alwine et al, Proc.
Natl. Acad. Sci., Vol. 74, pgs. 5350-5354 (1977); or Rabbani, U.S. Pat.
No. 4,139,346 (which patent is incorporated by reference); covalent
attachment to derivatized microspheres, e.g. Kremsky et al, Nucleic Acids
Research, Vol. 15, pgs. 2891-2909 (1987); or attachment to a derivatized
nylon substrate, such as Zeta Probe (a trademarked product of Bio-Rad) or
GeneScreen (a trademarked product of New England Nuclear). When targets
are about 1.5 kilobases or larger they can be anchored in a gel matrix,
e.g. as disclosed by Miyada et al, Meth. Enzymol., Vol. 154, pgs. 94-107
(1987).
The selection of appropriate prehybridization, hybridization, and wash
conditions for particular embodiments depends on several factors,
including the length of the probe, the degeneracy of the probe, the nature
of the bases and/or base analogs making up the probe, the means employed
for anchoring the target sequence, and the like. A great deal of
literature is available for guidance in selecting prehybridization,
hybridization, and washing conditions in view of the above factors for
particular embodiments, e.g. Wetmur et al, J. Mol. Biol., Vol. 31, pg. 349
(1968); Meinkoth et al, Anal. Biochem., Vol. 138, pgs. 267-284 (1984);
Eritja et al, Nucleic Acids Research, Vol 14, pgs. 8135-8153 (1986);
Aboul-ela et al, Nucleic Acids Research, Vol. 13, pgs. 4811-4824 (1985);
Millican et al, Nulceic Acids Research, Vol. 12, pgs. 7435-7452 (1987);
Ikuta et al, Nucleic Acids Research, Vol. 15, pgs. 797-811 (1987); Kawase
et al, Nucleic Acids Research, Vol. 14, pgs. 7727-7736 (1985); Seela et
al, Nucleic Acids Research, Vol 14, pgs. 1825-1844 (1986); Anand et al,
Nucleic Acids Research, Vol. 8167-8177 (1987); Gingeras et al, Nucleic
Acids Research, Vol. 15, pgs. 5373-5390 (1987); Miyada et al, Meth.
Enzymol., Vol. 154, pgs. 94-107 (1987); Ohtsuka et al, J. Biol. Chem.,
Vol. 260, pgs. 2605-2608 (1985); Kafatos et al, Nucleic Acids Research,
Vol. 7, pgs. 1541-1552 (1979); Wood et al, Proc. Natl. Acad. Sci., Vol.
82, pgs. 1585-1588 (1985); and Wallace et al, Nucleic Acids Research, Vol.
6, pgs. 3543-3557 (1979). Accordingly, these references are incorporated
by reference.
When the target sequence is in the range of about 400-1500 bases in length,
the preferred anchoring means is nitrocellulose or derivatized nylon, such
as GeneScreen, ZetaProbe, or the like. When the target sequence is in the
range of about 100 bases to several kilobases in length, a preferred
anchoring means is a nylon substrate, such as GeneScreen or GeneScreen
Plus, trademarked products of New England Nuclear (Boston, Mass.),
disclosed in U.S. Pat. No. 4,455,370, which is incorporated by reference.
A known quantity of single or double stranded copies of the target
sequence is anchored to the substrate, or solid phase support. As used
herein "known quantity" means amounts from which integral numbers of
perfectly matched probes can be determined. In some embodiments this means
known gram or molar quantities of the target sequence. In other
embodiments, it can mean equal amounts of target sequence on the plurality
of solid phase supports, so that signal is corresponding to integral
numbers of probes can be discerned by comparing signals from the plurality
of supports, or by comparing signals to specially provided standards.
Preferably, the anchoring means is loaded to capacity with the target
sequence so that maximal signals are produced after hybridization. The
target sequence can be prepared in double stranded form, denatured, and
then applied to the anchoring means, which is preferably a solid phase
support, such as nitrocellulose, GeneScreen, or the like. When the target
sequence is prepared in double stranded form, it is preferably excised
from its cloning vector with one or more endonucleases which leave blunt
ended fragments, e.g. Eco RV,Alu I, Bal I, Dra I, Nae I, Sma I, or the
like. In this case, both the coding, or sense, strand and the noncoding,
or antisense, strand are sequenced simultaneously. Because of sequence
complementarity, the reconstruction problem is no more difficult than in
the single stranded case. In some cases, an independent test or internal
standard may be required to determine which of the two reconstructed
sequences is the coding strand. An internal standard can comprise the
inclusion of a known sequence of nucleotides within the two strands that
are sequenced.
Suitable vectors for preparing double stranded target sequences are those
of the pUC series, e.g. Yanisch-Perron et al, Gene, Vol. 33, pgs. 103-119
(1985). These vectors are readily modified by adding unique restriction
sites to their polylinker regions. The new unique restriction sites are
selected from restriction endonucleases that leave flush-ended fragments
after digestion. For example, chemically synthesized fragments containing
such sites can be inserted into the Hind III and Eco RI sites of pUC18 or
pUC19. For these vectors such sites include Bal I, Eco RV, Hpa I, Nae I,
Nru I, Stu I, Sna BI, and Xca 1. With the modified pUC, the precursor of
the target sequence (i.e. the unknown sequence region) can be inserted
into a preexisting polylinker site, e.g. Bam HI; the vector can be
amplified and isolated; and the target sequence can be excised via the
restriction endonucleases that leave flush-ended fragments. The fragments
of the polylinker region excised along with the unknown sequence region
then become the known sequence regions of the target sequence.
Single stranded target sequences can be prepared in the form of RNA using
an SP6 promoter system, or similar system, as disclosed by Melton et al,
Nucleic Acids Research, Vol. 12, pgs. 7035-7056 (1984). SP6-based systems,
and related systems, are commercially available from Promega Corporation
(Madison, Wis.) as Riboprobe vectors, and United States Biochemical
Corporation (Cleveland, Ohio) as GeneScribe vectors. RNA target sequences
can be prepared with the SP6-like vectors in several ways. For example,
the double stranded DNA form of a target sequence can be inserted into a
cloning vector and amplified. The vector is then isolated and linearized
by digesting with the enzyme corresponding to the restriction site at the
3' end of the target sequence insert. RNA polymerase is used to generate
RNA copies of the coding strand of the target sequence. Alternatively, the
target sequence can be inserted into the SP6 (or related) vector in tandem
with a transcription termination sequence, which blocks RNA polymerase.
Thus, the linearization step is not required. Appropriate transcription
terminatiors are disclosed by Belagaje et al, U.S. Pat. No. 4,710,464.
Accordingly, that patent is incorporated by reference.
Copies of the target sequence can also be generated by way of a polymerase
chain reaction, e.g. Saiki et al, Science, Vol. 230, pgs. 1350-1354
(1985); Wrischnik et al, Nucleic Acids Research, Vol. 15, pgs. 529-542
(1987); U.S. Pat. Nos. 4,683,195 and 4,683,202. Accordingly, these
references are incorporated by reference. In particular, single stranded
copies can be obtained by providing a first primer in the standard two
primer amplification scheme (1) that is longer than the second primer and
(2) that contains a 5' region noncomplementary to its target strand. This
results in so-called "short products" of unequal length which can be
denatured and separated by size after amplification (e.g. as discussed
below).
Single stranded target sequences can also be prepared by separating the
complementary strands of the double stranded DNA by RPC-5 chromatography
under denaturing conditions, e.g. Eshaghpour et al, Nucleic Acids
Research, Vol. 5, pgs. 13-21 (1978), and Wells et al, Meth. Enzymol., Vol.
65, pgs. 327-347 (1980). Accordingly, these references are incorporated by
reference.
The preferred method of anchoring DNA to nitrocellulose filters is
essentially that described by Kafatos et al (cited above). Up to about 1
ug of target sequence is applied per square millimeter of the filter.
Before application the DNA is denatured, preferably in 0.3 to 0.4 N NaOH
for about 10 minutes, after which it is chilled with an equal volume of
cold water, or optionally cold 2 M ammonium acetate, to a concentration of
about 16 ug/ml. Known quantities of the denatured target are spotted onto
the filter by carefully controlling the volume of liquid deposited. After
each sample is spotted (approximately 1.5 minutes), the filter can
optionally be rinsed through with a drop of 1 M ammonium acetate
containing about 0.02-0.2 N NaOH, pH 7.8-9.0. Filters may also be washed
with 4.times.SSC (defined below), e.g. about 200 ml. The filters are air
dried, shaken in 2.times. Denhardt's solution (defined below) for at least
1 hour, drained and air dried again, and baked under vacuum at 80.degree.
C. for about 2 hours.
Hybridization of the probes to the target sequence usually comprises three
steps: a prehybridization treatment, application of the probe, and
washing. The purpose of the prehybridization treatment is to reduce
nonspecific binding of the probe to the anchoring means and non-target
nucleic acids. This is usually accomplished by blocking potential
nonspecific binding sites with blocking agents such as proteins, e.g.
serum albumin (a major ingredient of Denhardt's solution). The commonly
used blocking agent, denatured sonicated salmon sperm DNA, is usually not
appropriate with the present invention. Such a blocking agent could be a
source of specific and nonspecific binding sites because of the short
length of the probes. Any blocking agent comprising natural DNA has a high
probability of containing some complementary sequences to some probes.
However, non-labelled oligonucleotide mixtures specially selected to be
maximally noncomplementary to the probe sequence can be prepared for
subsets of the probes for use as blocking agents.
For example, probes of the form
##EQU4##
would be maximally noncomplementary to a probe of the form 3'-
##EQU5##
For target sequences anchored to nitrocellulose or nylon (e.g. GeneScreen),
prehybridization treatment can comprise treatment with 5-10x Denhardt's
solution, with 2-6x SSC preferably containing a mild detergent, e.g. 0.5%
sodium dodecylsulfate (SDS), for 15 min. to 1 hr. at a temperature in the
range of about 25.degree. to 60.degree.. Denhardt's solution, disclosed in
Biochem. Biophys. Res. Commun., Vol. 23, pgs. 641-645 (1966), consists at
10x concentration of 0.2% bovine serum albumin, 0.2% polyvinylpyrolidone,
and 0.2% Ficoll. SSC, another standard reagent in the hybridization art,
consists at 1x of 0.15 M NaCl, 0.015 M sodium citrate, at pH 7.0.
Preferred treatment times, temperatures, and formulations may vary with
the particular embodiment.
Preferably, the probe is applied to the anchored DNA at a concentration in
the range of about 1-10 ng/ml in a solution substantially the same as the
prehybridization solution, e.g. 5-10x Denhardt's solution with 2-6x SSC
and a mild detergent, e.g. 0.5% SDS. More preferably, the probe
concentration is in the range of about 1-5 ng/ml. Preferably, the
hybridization is carried out at a temperature 10-20.degree. C. below the
expected 50% dissociation temperature, Td, between the probe and the
target. That is, a temperature is selected at which a high proportion,
e.g. greater than 80-90%, of all the perfectly matched probes form stable
duplexes. For 8-mer probes the preferred hybridization temperature is in
the range of about 10-18.degree. C. Hybridization times are preferably in
the range of about 3-16 hours. Different hybridization times may be
selected for probes of different degrees of degeneracy, because the
effective consentrations of particular sequences within a highly
degenerate probe, e.g. 0A000000, are considerably less than those of
particular sequences in a low degeneracy probe, e.g. AAA0AAAA. Thus,
higher Cot values (which usually means longer hybridization times) may be
required for more degenerate probes to attain a sufficient degree of
binding of perfectly matche | | |
