 |
Description  |
|
|
FIELD OF THE INVENTION
The present invention relates to a method, a system, and a kit useful for
the identification of molecules that specifically bind to defined nucleic
acid sequences.
REFERENCES
Ausubel, F. M., et al., Current Protocols in Molecular Biology. John Wiley
and Sons, Inc., Media Pa.
Chang, H.-K,, et al., Mol. Cell. Biol. November: 5189-5197 (1989).
Chaiet, L., et al., Arch. Biochem. Biophys. 106:1 (1964).
Chen, K-X., et al., J. Biomol. Struct. Dyn. 3:445-466 (1985).
Courtois, G., et al., Proc. Natl. Acad. Sci. USA 85:7937-7941 (1988).
Elias, P., et al., Proc. Natl. Acad. Sci. USA 85:2959-2963 (1988).
Fried, M. G., et al., Nuc. Acid. Res. 9:6505 (1981).
Galas, D., et al., Nuc. Acid. Res. 5:3157-3170 (1981).
Garner, M. M., et al., Nuc. Acid. Res. 9;3047(1981).
Gessner, R. V., et al., Biochemistry 24:237-240 (1985).
Gilbert, D. F., et al., Proc. Natl. Acad. Sci. USA 86:3006 (1988).
Gilman, A. G., et al., eds., The pharmacologically Basis of Therapeutics,
Eighth Edition, Pergamon Press (1990).
Goldin, A. L., et al., J. Virol. 38:5-58 (1981).
Green, N. M., Adv. Protein Chem. 29:85 (1975).
Harlow, E., et al., Antibodies: A Laboratory Manual. Cold Spring Harbor
Laboratory Press (1988).
Jain, S. C., et al., J. Mol. Biol. 68:1-20 (1972).
Kadonaga, J. T., PNAS 83:5889-5893 (1986).
Koff, A., et al., J. Virol. 62:4096-4103 (1988).
Luck, G., et al., Nucl. Acids Res. 1:503 (1974).
Luckow, V. A., et al., Virology 170:31 (1989).
Maniatis, T., et al. Molecular Cloning: A Laboratory Manual, Cold Spring
Harbor Laboratory (1982). McGeoch, D. J., et al., J. Virology 62(2):444
(1988).
Olivo, P. D., et al., Proc. Natl. Acad. Sci. USA 85:5414-5418 (1988).
Olivo, P. D., et al., J. Virology 3:196-204 (1989).
Polinksy, B., et al., PNAS 72:3310-4 (1975).
Quigley, G. J., et al., Science 232:1255-1258 (1986).
Sambrook, J., et al., In Molecular Cloning: A Laboratory Manual. Cold
Spring Harbor Laboratory Press, Vol. 2 (1989).
Sherman, S. E., et al., Chem. Rev. 87:1153 (1987).
Siebenlist, U., et al., Proc. Natl. Acad. Sci. USA 77:122-126 (1980).
Smith, D. B., et al., Gene 67:31 (1988).
Sobell, H. M., et al., J. Mol. Biol. 68:21-34 (1972).
Sobell, H. M., Prof. Nucl. Acid. Res. Mol. Biol. 13:153-190 (1973).
Stow, N. D., et al., Virology 130:427-438 (1983).
Summers, M. D., et al., A manual of Methods for Baculovirus Vectors and
Insect Cell Culture Procedures, Texas Agricultural Experimental Station
Bulletin, No. 1555 (1987).
Tullius, T. D., Ann. Rev. Biophys. Biochem. 18:213-237 (1989).
Wartel, R. M., et al., J. Biol. Chem. 15:285-318 (1975).
Weir, H. M., et al., Nucl. Acids Res. 17:1409-1425 (1989).
Woodbury, C. P., et al., Biochemistry 22(20):4730-4737 (1983).
Wu C. A., et al., J. Virol. 62:435-443 (1988).
Zein, N., et al., Science 240:1198 (1988).
Zimmer, C., Pros. Nucl. Acid Res. Mol. Biol. 15:285-318 (1975).
BACKGROUND OF THE INVENTION
Several classes of small molecules that interact with double-stranded DNA
have been identified. Many of these small molecules have profound
biological effects. For example, many aminoacridines and polycyclic
hydrocarbons bind DNA and are mutagenic, teratogenic, or carcinogenic.
Other small molecules that bind DNA include: biological metabolites, some
of which have applications as antibiotics and antitumor agents including
actinomycin D, echinomycin, distamycin, and calicheamicin; planar dyes,
such as ethidium and acridine orange; and molecules that contain heavy
metals, such as cisplatin, a potent antitumor drug.
Most known DNA-binding molecules do not have a known sequence binding
preference. However, there are a few small DNA-binding molecules that
preferentially recognize specific nucleotide sequences, for example:
echinomycin preferentially binds the sequence [(A/T)CGT]/[ACG(A/T)]
(Gilbert et al.); cisplatin covalently cross-links a platinum molecule
between the N7 atoms of two adjacent deoxyguanosines (Sherman et al.); and
calicheamicin preferentially binds and cleaves the sequence TCCT/AGGA
(Zein et al.).
The biological response elicited by most therapeutic DNA-binding molecules
is toxicity, specific only in that these molecules may preferentially
affect cells that are more actively replicating or transcribing DNA than
other cells. Targeting specific sites may significantly decrease toxicity
simply by reducing the number of potential binding sites in the DNA. As
specificity for longer sequences is acquired, the nonspecific toxic
effects due to DNA-binding may decrease. Many therapeutic DNA-binding
molecules initially identified based on their therapeutic activity in a
biological screen have been later determined to bind DNA.
Experiments performed in support of the present invention have identified
an in vitro assay useful to screen for DNA-binding molecules. The assay
also allows the discrimination of sequence binding preferences of such
molecules. The potential therapeutic applications for molecules that bind
to specific DNA sequences are widespread.
SUMMARY OF THE INVENTION
The present invention provides a method for screening molecules or
compounds capable of binding to a selected test sequence in a duplex DNA.
The method involves adding a molecule to be screened, or a mixture
containing the molecule, to a test system. The test system includes a DNA
binding protein that is effective to bind to a screening sequence, i.e.
the DNA binding protein's cognate binding site, in a duplex DNA with a
binding affinity that is substantially independent of the sequences
adjacent the binding sequence--these adjacent sequences are referred to as
test sequences. But, the DNA binding protein is sensitive to binding of
molecules to such test sequence, when the test sequence is adjacent the
screening sequence. The test system further includes a duplex DNA having
the screening and test sequences adjacent one another. Also, the binding
protein is present in an amount that saturates the screening sequence in
the duplex DNA. The molecule is incubated in contact with the test system
for a period sufficient to permit binding of the molecule being tested to
the test sequence in the duplex DNA. The amount of binding protein bound
to the duplex DNA is compared before and after the addition of the test
molecule or mixture.
Candidates for the screening sequence/binding protein may be selected from
the following group: EBV origin of replication/EBNA, HSV origin of
replication/UL9, VZV origin of replication/UL9-like, HPV origin of
replication/E2, interleukin 2 enhancer/NFAT-1, HIV-LTR/NFAT-1,
HIV-LTR/NFkB, HBV enhancer/HNF-1, fibrinogen promoter/HNF-1, lambda
o.sub.L -o.sub.R /cro, and other known DNA:protein interactions.
A preferred embodiment of the present invention utilizes the UL9 protein,
or DNA-binding proteins derived therefrom, and its cognate binding
sequence SEQ ID NO:1, SEQ ID NO:2, or SEQ ID NO:15.
The test sequences can be any combination of sequences of interest. The
sequences may be randomly generated for shot-gun approach screening or
specific sequences may be chosen. Some specific sequences of medical
interest include the following sequences involved in DNA:protein
interactions: EBV origin of replication, HSV origin of replication, VZV
origin of replication, HPV origin of replication, interleukin 2 enhancer,
HIV-LTR, HBV enhancer, and fibrinogen promoter.
In the above method, comparison of protein-bound to free DNA can be
accomplished using either a gel band-shift assay, a filter-binding assay,
or a capture/detection assay.
In one embodiment of the DNA capture/detection assay the capture system
involves the biotinylation of a nucleotide within the screening sequence
(i) that does not eliminate the protein's ability to bind to the screening
sequence, (ii) that is capable of binding streptavidin, and (iii) where
the biotin moiety is protected from interactions with streptavidin when
the protein is bound to the screening sequence. The capture/detection
assay also involves the detection of the captured DNA.
The present invention also includes a screening system for identifying
molecules that are capable of binding to a test sequence in a duplex DNA
sequence. The system includes a DNA binding protein that is effective to
bind to a screening sequence in a duplex DNA with a binding affinity that
is substantially independent of a test sequence adjacent the screening
sequence. The binding of the DNA protein is, however, sensitive to binding
of molecules to the test sequence when the test sequence is adjacent the
screening sequence. The system includes a duplex DNA having the screening
and test sequences adjacent one another. Typically, the binding protein is
present in an amount that saturates the screening sequence in the duplex
DNA. The system also includes means for detecting the amount of binding
protein bound to the DNA.
As described above the test sequences can be any number of sequences of
interest.
The screening sequence/binding protein can be selected from known
DNA:protein interactions using the criteria and guidance of the present
disclosure.
A preferred embodiment of the screening system of the present invention
includes the UL9 protein, or DNA-binding protein derived therefrom (e.g.,
the truncated UL9 protein designated UL9-COOH). In this embodiment the
duplex DNA has (i) a screening sequence selected from the group consisting
of SEQ ID NO:1, SEQ ID NO:2, and SEQ ID NO:15, and (ii) a test sequence
adjacent the screening sequence, where UL9 is present in an amount that
saturates the screening sequence. The system further includes means for
detecting the amount of UL9 bound to the DNA, including, band-shift
assays, filter-binding assays, and capture/detection assays.
The present disclosure describes the procedures needed to test DNA:protein
interactions for their suitability for use in the screening assay of the
present invention.
The present invention further defines a DNA capture system and detection
system. In the first part of this system, the cognate DNA binding site of
the DNA binding protein is modified with a detection moiety, such as
biotin or digoxigenin. The modification must be made to the site in such a
manner that (i) it does not eliminate the protein's ability to bind to the
cognate binding sequence, (ii) the moiety is accessible to the capturing
agent (e.g., in the case of biotin the agent is streptavidin) in DNA that
is not bound to protein, and (iii) where the moiety is protected from
interactions with the capture agent when the protein is bound to the
screening sequence.
In the second part of this system, the target oligonucleotide is labelled
to allow detection. Labelling of the target oligonucleotide can be
accomplished by standard techniques such as radiolabelling. Alternatively,
a moiety such as digoxigenin can be incorporated in the target
oligonucleotide and this moiety can then be detected after capture.
Two embodiments of the capture/detection system described by the present
disclosure are as follows:
(i) the target oligonucleotide (containing, for example, the screening and
test sequences)--modification of the cognate binding site with biotin and
incorporation of digoxigenin; capture of the target oligonucleotide using
streptavidin attached to a solid support; and detection of the target
oligonucleotide using a tagged anti-digoxigenin antibody.
(ii) the target oligonucleotide--modification of the cognate binding site
with digoxigenin and incorporation of biotin; capture of the target
oligonucleotide using an anti-digoxigenin antibody attached to a solid
support; and detection of the target oligonucleotide using tagged
streptavidin.
BRIEF DESCRIPTION OF THE FIGURES
FIG. 1A illustrates how a DNA-binding protein is displaced by a small
molecule because of steric hinderance. FIG. 1B illustrates how a
DNA-binding protein is displaced because of conformational changes induced
in the DNA by a small molecule.
FIG. 2 illustrates an assay for detecting inhibitory molecules based on
their ability to preferentially hinder the binding of a DNA-binding
protein to its binding site. Protein (O) is displaced from DNA (/) in the
presence of inhibitor (X).
FIG. 3 shows a DNA-binding protein that is able to protect a biotin moiety,
covalently attached to the oligonucleotide sequence, from being recognized
by the streptavidin when the protein is bound to the DNA.
FIG. 4 shows the incorporation of biotin and digoxigenin into a typical
oligonucleotide molecule for use in the assay of the present invention.
The oligonucleotide contains the binding sequence (i.e., the screening
sequence) of the UL9 protein, which is underlined, and test sequences
flanking the screening sequence.
FIG. 5 shows a series of sequences that have been tested in the assay of
the present invention for the binding of sequence-specific small
molecules.
FIG. 6 outlines the cloning of a truncated form of the UL9 protein, which
retains its sequence-specific DNA-binding ability (UL9-COOH), into an
expression vector.
FIG. 7 shows the pVL1393 baculovirus vector containing the full length UL9
protein coding sequence.
FIG. 8 is a photograph of a SDS-polyacrylamide gel showing (i) the purified
UL9-COOH/glutathione-S-transferase fusion protein and (ii) the UL9-COOH
polypeptide.
FIG. 9 shows the effect on UL9-COOH binding of alterations in the test
sequences that flank the UL9 screening sequence.
FIG. 10A shows the effect of the addition of several concentrations of
Distamycin A to DNA:protein assay reactions utilizing different test
sequences. FIG. 10B shows the effect of the addition of Actinomycin D to
DNA:protein assay reactions utilizing different test sequences. FIG. 10C
shows the effect of the addition of Doxorubicin to DNA:protein assay
reactions utilizing different test sequences.
FIG. 11A illustrates a DNA capture system of the present invention
utilizing biotin and streptavidin coated magnetic beads. The presence of
the DNA is detected using an alkaline-phosphatase substrate that yields a
chemiluminescent product. FIG. 11B shows a similar reaction using biotin
coated agarose beads that are conjugated to streptavidin, that in turn is
conjugated to the captured DNA.
FIG. 12 demonstrates a test matrix based on DNA:protein-binding data.
DETAILED DESCRIPTION OF THE INVENTION
Definitions
Small molecules are desirable as therapeutics for several reasons related
to drug delivery: (i) they are commonly less than 10 K molecular weight;
(ii) they are more likely to be permeable to cells; (iii) unlike peptides
or oligonucleotides, they are less susceptible to degradation by many
cellular mechanisms; and, (iv) they are not as apt to elicit an immune
response. Many pharmaceutical companies have extensive libraries of
chemical and/or biological mixtures, often fungal, bacterial, or algal
extracts, that would be desirable to screen with the assay of the present
invention. Small molecules may be either biological or synthetic organic
compounds, or even inorganic compounds (i.e., cisplatin).
Dissociation is the process by which two molecules cease to interact: the
process occurs at a fixed average rate under specific physical conditions.
Functional binding is the noncovalent association of a protein or small
molecule to the DNA molecule. In the assay of the present invention the
functional binding of the protein to the screening sequence (i.e., its
cognate DNA binding site) has been evaluated using filter binding or gel
band-shift experiments.
On-rate is herein defined as the time required for two molecules to reach
steady state association: for example, the DNA:protein complex.
Off-rate is herein defined as the time required for one-half of the
associated complexes, e.g., DNA:protein complexes, to dissociate.
Sequence-specific binding refers to DNA binding molecules which have a
strong DNA sequence binding preference. For example, restriction enzymes
and the proteins listed in Table I demonstrate typical sequence-specific
DNA-binding.
Sequence-preferential binding refers to DNA binding molecules that
generally bind DNA but that show preference for binding to some DNA
sequences over others. Sequence-preferential binding is typified by
several of the small molecules tested in the present disclosure, e.g.,
distamycin. Sequence-preferential and sequence-specific binding can be
evaluated using a test matrix such as is presented in FIG. 12.
Screening sequence is the DNA sequence that defines the cognate binding
site for the DNA binding protein: in the case of UL9 the screening
sequence can, for example, be SEQ ID NO:1.
Test sequence is a DNA sequence adjacent the screening sequence. The assay
of the present invention screens for molecules that, when bound to the
test sequence, affect the interaction of the DNA-binding protein with its
cognate binding site (i.e., the screening sequence). Test sequences can be
placed adjacent either or both ends of the screening sequence. Typically
binding of molecules to the test sequence interfere with the binding of
the DNA-binding protein to the screening sequence. However, some molecules
binding to these sequences may have the reverse effect, causing an
increased binding affinity of the DNA-binding protein to the screening
sequence.
I. The Assay
One feature of the present invention is that it provides an assay to screen
libraries of synthetic or biological compounds, including small molecules
and proteins, for their ability to bind DNA in a sequence-preferential
manner.
A. General Considerations
The assay of the present invention has been designed for detecting test
molecules or compounds that affect the rate of transfer of a specific DNA
molecule from one protein molecule to another identical protein in
solution.
A mixture of DNA and protein is prepared in solution. The concentration of
protein is in excess to the concentration of the DNA so that virtually all
of the DNA is found in DNA:protein complexes. The DNA is a double-stranded
oligonucleotide that contains the recognition sequence for a specific
DNA-binding protein (i.e., the screening sequence). The protein used in
the assay contains a DNA-binding domain that is specific for binding to
the sequence within the oligonucleotide. The physical conditions of the
solution (e.g., pH, salt concentration, temperature) are adjusted such
that the half-life of the complex is amenable to performing the assay
(optimally a half-life of 5-30 minutes).
As one DNA:protein complex dissociates, the released DNA rapidly reforms a
complex with another protein in solution. Since the protein is in excess
to the DNA, dissociations of one complex always result in the rapid
reassociation of the DNA into another DNA:protein complex. At equilibrium,
very few DNA molecules will be unbound. The minimum background of the
assay is the amount of unbound DNA observed during any given measurable
time period. The brevity of the observation period and the sensitivity of
the detection system define the lower limits of background DNA.
FIG. 1A illustrates how such a protein can be displaced from its cognate
binding site by steric hinderance of a small molecule. Alternatively, a
molecule may interfere with a DNA:protein binding interaction by inducing
a conformational change in the DNA (FIG. 1B). In either event, if a test
molecule that binds the oligonucleotide hinders binding of the protein,
the rate of transfer of DNA from one protein to another will be decreased.
This will result in a net increase in the amount of unbound DNA. In other
words, an increase in the amount of unbound DNA indicates the presence of
an inhibitor.
Alternatively, molecules may be isolated that, when bound to the DNA, cause
an increased affinity of the DNA-binding protein for its cognate binding
site. In this case the amount of unbound DNA (observed during a given
measurable time period after the addition of the molecule) will decrease
in the reaction mixture as detected by the capture/detection system
described in Section II.
B. Choosing a DNA:protein Complex
There are several approaches that could be taken to look for small
molecules that specifically inhibit the interaction of a given DNA-binding
protein with its binding sequence (cognate site). One approach would be to
test biological or chemical compounds for their ability to preferentially
block the binding of one specific DNA:protein interaction but not the
others. Such an assay would depend on the development of at least two,
preferably three, DNA:protein interaction systems in order to establish
controls for distinguishing between general DNA-binding molecules
(polycations like heparin or intercalating agents like ethidium) and
DNA-binding molecules having sequence binding preferences that would
affect protein/cognate binding site interactions in one system but not the
other(s).
One illustration of how this system could be used is as follows. Each
cognate site could be placed 5' to a reporter gene (such as genes encoding
.beta.-galactoside or luciferase) such that binding of the protein to the
cognate site would block transcription of the reporter gene. In the case
where multiple protein/cognate binding sites are used for screening, a
competitive inhibitor that blocks one interaction but not the others could
be identified by the lack of transcription of a reporter gene in a
transfected cell line or in an in vitro assay. Only one such DNA-binding
sequence, specific for the protein of interest, could be screened with
each assay system. This approach has a number of limitations including
limited testing capability and the need to construct the appropriate
reporter system for each different protein/cognate site of interest.
Experiments performed in support of the present invention have defined a
second approach for identifying molecules having sequence-preferential
DNA-binding. In this approach small molecules binding to sequences
adjacent the cognate binding sequence can inhibit the protein/cognate DNA
interaction. This assay has been designed to use a single DNA:protein
interaction to screen for sequence-specific or sequence-preferential
DNA-binding molecules that recognize virtually any sequence.
While DNA-binding recognition sites are usually quite small (4-17 bp), the
sequence that is protected by the binding protein is larger (usually 5 bp
or more on either side of the recognition sequence--as detected by DNAase
I protection (Galas et al.) or methylation interference (Siebenlist et
al.). Experiments performed in support of the present invention
demonstrated that a single protein and its cognate DNA-binding sequence
can be used to assay virtually any DNA sequence by placing a sequence of
interest adjacent to the cognate site: a small molecule bound to the
adjacent site can be detected by the dissociation of the protein from its
cognate site. Such dissociation might occur by either steric hindrance or
induced conformational changes in the recognition sequence for the
protein.
There are several considerations involved in choosing DNA:protein complexes
that can be employed in the assay of the present invention including:
The off-rate (i.e., the time period from contact of the protein with the
DNA site until disassociation) should be fast enough to accomplish the
assay in a reasonable amount of time. The interactions of some proteins
with cognate sites in DNA can be measured in days not minutes: such
tightly bound complexes would inconveniently lengthen the period of time
it takes to perform the assay.
2) The off-rate should be slow enough to allow the measurement of unbound
DNA in a reasonable amount of time. The level of background free DNA is
dictated by the ratio between the time needed to measure free DNA and the
amount of free DNA that occurs naturally due to the off-rate during the
measurement time period.
In view of the above two considerations, practical useful DNA:protein
off-rates fall in the range of approximately two minutes to several days.
3) A further consideration is that the kinetic interactions of the
DNA:protein complex is insensitive to the nucleotide sequences flanking
the recognition sequence. The affinity of many DNA-binding proteins is
affected by differences in the sequences adjacent to the recognition
sequence. The most obvious example of this phenomenon is the preferential
binding and cleavage of restriction enzymes given a choice of several
identical recognition sequences with different flanking sequences
(Polinsky et al.). If the off-rates are affected by flanking sequences the
analysis of comparative binding data between different flanking
oligonucleotide sequences becomes difficult but is not impossible.
Experiments performed in support of the present invention have identified a
DNA:protein interaction that is particularly useful for the above
described assay: the Herpes Simplex Virus (HSV) UL9 protein that binds the
HSV origin of replication (oriS). The UL9 protein has fairly stringent
sequence specificity. There appear to be two binding sites for UL9 in
oriS, SEQ ID NO:1 and SEQ ID NO:2 (Stow et al.). One sequence (SEQ ID
NO:1) binds with at least 10-fold higher affinity than the second sequence
(SEQ ID NO:2): the embodiments described below use the higher affinity
binding site (SEQ ID NO:1).
DNA:protein association reactions are performed in solution. The
DNA:protein complexes can be separated from free probe by any of several
methods. One particularly useful method for the initial study of
DNA:protein interactions has been visualization of binding results using
band shift gels (Example 3A). In this method DNA:protein binding reactions
containing both labelled complexes and free DNA are separated
electrophoretically on polyacrylamide/TBE gels. These gels are fixed,
dried, and exposed to X-ray film. The resulting autoradiograms are
examined for the amount of free probe that is migrating separately from
the DNA:protein complex. These assays include (i) a lane containing only
free labeled probe, and (ii) a lane where the sample is labeled probe in
the presence of a large excess of binding protein. The band shift assays
allow visualization of the ratios between DNA:protein complexes and free
probe. However, they are less accurate than filter binding assays for
rate-determining experiments due to the lag time between loading the gel
and electrophoretic separation of the components.
The filter binding method is particularly useful in determining the
off-rates for protein:oligonucleotide complexes (Example 3B). In the
filter binding assay, DNA:protein complexes are retained on a filter while
free DNA passes through the filter. This assay method is more accurate for
off-rate determinations because the separation of DNA:protein complexes
from free probe is very rapid. The disadvantage of filter binding is that
the nature of the DNA:protein complex cannot be directly visualized. So
if, for example, two proteins are used with a single DNA molecule, filter
binding assays cannot differentiate between the binding of the two
proteins nor yield information about whether one or both proteins are
binding.
There are many known DNA:protein interactions that may be useful in the
practice of the present invention, including (i) the DNA protein
interactions listed in Table I, (ii) bacterial, yeast, and phage systems
such as lambda o.sub.L -o.sub.R /cro, and (iii) modified restriction
enzyme systems (e.g., protein binding in the absence of divalent cations).
Any protein that binds to a specific recognition sequence may be useful in
the present invention. The major constraining factor is the effect of the
immediately adjacent sequences (the test sequences) on the affinity of the
protein for its recognition sequence. DNA:protein interactions in which
there is little or no effect of the test sequences on the affinity of the
protein for its cognate site are preferable for use in the described
assay; however, DNA:protein interactions that exhibit (test
sequence-dependent) differential binding may still be useful if algorithms
are applied to the analysis of data that compensate for the differential
affinity. The present disclosure provides methods and guidance for testing
the usefulness of such DNA:protein interactions, i.e., other than UL9, in
the screening assay.
C. Preparation of Full Length UL9 and UL9-COOH Polypeptides
UL-9 protein has been prepared by a number of recombinant techniques
(Example 2). The full length UL9 protein has been prepared from
baculovirus infected insect cultures (Example 3A, B, and C). Further, a
portion of the UL9 protein that contains the DNA-binding domain (UL9-COOH)
has been cloned into a bacterial expression vector and produced by
bacterial cells (Example 3D and E). The DNA-binding domain of UL9 is
contained within the C-terminal 317 amino acids of the protein (Weir et
al.). The UL9-COOH polypeptide was inserted into the expression vector
in-frame with the glutathione-S-transferase (gst) protein. The gst/UL9
fusion protein was purified using affinity chomatography (Example 3E). The
vector also contained a thrombin cleavage site at the junction of the two
polypeptides. Therefore, once the fusion protein was isolated (FIG. 8,
lane 2) it was treated with thrombin, cleaving the UL9-COOH/ gst fusion
protein from the gst polypeptide (FIG. 8, lane 3). The UL9-COOH-gst fusion
polypeptide was obtained at a protein purity of greater than 95% as
determined using Coomaisie staining.
Other hybrid proteins can be utilized to prepare DNA-binding proteins of
interest. For example, fusing a DNA-binding protein coding sequence
in-frame with a sequence encoding the thrombin site and also in-frame with
the .beta.-galactoside coding sequence. Such hybrid proteins can be
isolated by affinity or immunoaffinity columns (Maniatis et al.; Pierce,
Rockford IL). Further, DNA-binding proteins can be isolated by affinity
chromatography based on their ability to interact with their cognate DNA
binding site. For example, the UL9 DNA-binding site (SEQ ID NO:1) can be
covalently linked to a solid support (e.g., CnBr-activated Sepharose 4B
beads, Pharmacia, Piscataway N.J.), extracts passed over the support, the
support washed, and the DNA-binding then isolated from the support with a
salt gradient (Kadonaga).
The results presented below in regard to the DNA-binding ability of the
truncated UL9 protein suggest that full length DNA-binding proteins are
not required for the DNA:protein assay of the present invention: only a
portion of the protein containing the cognate site recognition function
may be required. The portion of a DNA-binding protein required for
DNA-binding can be evaluated using a functional binding assay (Example
4A). The rate of dissociation can be evaluated (Example 4B) and compared
to that of the full length DNA-binding protein.
D. Functional Binding and Rate of Dissociation.
The full length UL9 and purified UL9-COOH proteins were tested for
functional activity in "band shift" assays (see Example 4A). The buffer
conditions were optimized for DNA:protein-binding (Example 4C) using the
UL9-COOH polypeptide. These DNA-binding conditions also worked well for
the full-length UL9 protein. Radiolabelled oligonucleotides (SEQ ID NO:14)
that contained the 11 bp UL9 DNA-binding recognition sequence (SEQ ID
NO:1) were mixed with each UL9 protein in appropriate binding buffer. The
reactions were incubated at room temperature for 10 minutes (binding
occurs in less than 2 minutes) and the products were separated
electrophoretically on non-denaturing polyacrylamide gels (Example 4A).
The degree of DNA:protein-binding could be determined from the ratio of
labeled probe present in DNA:protein complexes versus that present as free
probe. This ratio was typically determined by optical scanning of
autoradiograms and comparison of band intensities. Other standard methods
may be used as well for this determination, such as scintillation counting
of excised bands. The UL9-COOH polypeptide and the full length UL9
polypeptide, in their respective buffer conditions, bound the target
oligonucleotide equally well.
The rate of dissociation was determined using competition assays. An excess
of unlabelled oligonucleotide that contained the UL9 binding site was
added to each reaction. This unlabelled oligonucleotide acts as a specific
inhibitor, capturing the UL9 protein as it dissociates from the labelled
oligonucleotide (Example 4B). The dissociation rate, as determined by a
band-shift assay, for both full length UL9 and UL9-COOH was approximately
4 hours at 4.degree. C. Neither non-specific oligonucleotides (a
10,000-fold excess) nor sheared herring sperm DNA (a 100,000-fold excess)
competed for binding with the oligonucleotide containing the UL9 binding
site.
E. oriS Flanking Sequence Variation.
As mentioned above, one feature of a DNA:protein-binding system for use in
the assay of the present invention is that the DNA:protein interaction is
not affected by the nucleotide sequence of the regions adjacent the
DNA-binding site. The sensitivity of any DNA:protein-binding reaction to
the composition of the flanking sequences can be evaluated by the
functional binding assay and dissociation assay described above.
To test the effect of flanking sequence variation on UL9 binding to the
oriS SEQ ID NO:1 sequences oligonucleotides were constructed with 20-30
different sequences (i.e., the test sequences) flanking the 5' and 3'
sides of the UL9 binding site. Further, oligonucleotides were constructed
with point mutations at several positions within the UL9 binding site.
Most point mutations within the binding site destroyed recognition.
Several changes did not destroy recognition and these include variations
at sites that differ between the two UL9 binding sites (SEQ ID NO:1 and
SEQ ID NO:2): the second UL9 binding site (SEQ ID NO:2) shows a ten-fold
decrease in UL9:DNA binding affinity (Elias et al.) relative to the first
(SEQ ID NO:1). On the other hand, sequence variation at the test site
(also called the test sequence), adjacent to the screening site (FIG. 5,
Example 5), had virtually no effect on binding or the rate of
dissociation.
Taken together the above experiments support that the UL9-COOH polypeptide
binds the SEQ ID NO:1 sequence with (i) appropriate strength, (ii) an
acceptable disassociation time, and (iii) indifference to the nucleotide
sequences flanking the assay (binding) site. These features suggested that
the UL9/oriS system could provide a versatile assay for detection of small
molecule/DNA-binding involving any number of specific nucleotide
sequences.
The above-described experiment can be used to screen other DNA:protein
interactions to determine their usefulness in the present assay.
F. Small Molecules as Sequence-Specific Competitive Inhibitors
To test the utility of the present assay system several small molecules
that have sequence preferences (i.e., a preference for AT-rich versus
GC-rich sequences) have been tested.
Distamycin A binds relatively weakly to DNA (K.sub.A =2.times.10.sup.5
M.sup.-1) with a preference for non-alternating AT-rich sequences (Jain et
al.; Sobell; Sobell et al.). Actinomycin D binds DNA more strongly
(K.sub.A =7.6.times.10.sup.-7 M.sup.-1) than Distamycin A and has a
relatively strong preference for the dinucleotide sequence dGdC (Luck et
al.; Zimmer; Wartel). Each of these molecules poses a stringent test for
the assay. Distamycin A tests the sensitivity of the assay because of its
relatively weak binding. Actinomycin D challenges the ability to utilize
flanking sequences since the UL9 recognition sequence contains a dGdC
dinucleotide: therefore, it might be anticipated that all of the
oligonucleotides, regardless of the test sequence flanking the assay site,
might be equally affected by actinomycin D.
In addition, Doxorubicin, a known anti-cancer agent that binds DNA in a
sequence-preferential manner (Chen, K-X, et al.), has been tested for
preferential DNA sequence binding using the assay of the present
invention.
Actinomycin D, Distamycin A, and Doxorubicin have been tested for their
ability to preferentially inhibit the binding of UL9 to oligonucleotides
containing different sequences flanking the UL9 binding site (Example 6,
FIG. 5). Binding assays were performed as described in Example 5. These
studies were completed under conditions in which UL9 is in excess of the
DNA (i.e., most of the DNA is in complex).
Distamycin A was tested with 5 different test sequences flanking the UL9
screening sequence: SEQ ID NO:5 to SEQ ID NO:9. The results shown in FIG.
10A demonstrate that distamycin A | | |
