Anal. Chem. 2000, 72, 3138-3141
MutS-Mediated Detection of DNA Mismatches
Using Atomic Force Microscopy
Hui Bin Sun and Hiroki Yokota*
Biomedical Engineering Program, Departments of Mechanical Engineering and of Anatomy and Cell Biology, Indiana
University Purdue University at Indianapolis, Indianapolis, Indiana 46202
We have developed an atomic force microscopy-based
method for detecting DNA base-pair mismatches using
MutS protein isolated from E. coli. MutS is a biological
sensor and a locator of DNA base-pair mismatches. It
binds specifically to a mismatched DNA base pair and
initiates a process of DNA repair. To test the possibility
of visually detecting mismatched base pairs by atomic
force microscopy, we prepared DNA templates ∼500 bp
in length consisting of a single or multiple base-pair
mismatches. We demonstrate that MutS binding sites on
individual DNA molecules were readily detectable by
atomic force microscopy and that the observed positions
were in good agreement with the predicted sites of basepair mismatches at a few-nanometer resolution. The
technique described here is rapid and sensitive and is
expected to be useful in screening mutations and DNA
polymorphisms.
The completion of the human genome project in 2003 will
provide a unique opportunity for studying genetic mutations and
DNA polymorphisms in humans.1,2 DNA sequence alterations are
the primary cause for genetic disorders, and an understanding of
the relationship between sequence variations and disease risk is
critical to disease prevention and clinical treatments in the post
human genome sequence era. Furthermore, detection of sequence
alterations provides an unambiguous method for identifying
individuals in forensic situations and in questions of parenthood.3,4
To facilitate access to human DNA information, technologies are
needed that rapidly compare similarities and differences between
various DNA samples. Since many genetic disorders are caused
by a point mutations and the most common polymorphisms in
the human genome are single base-pair differences, a reliable
screening tool for detecting mismatched base pairs would be
valuable.5
* Corresponding author: Department of Anatomy and Cell Biology, Indiana
University School of Medicine, 635 Barnhill Dr., MS-504, Indianapolis, IN 46202.
Phone: 317-274-2448. Fax: 317-278-2040. E-mail: [email protected].
(1) Collins, F. S.; Patrinos, A.; Jordan, E.; Chakravarti, A.; Gesteland, R.; Walters,
L. Members of the DOE and NIH planning groups. Science 1998, 282,
682-689.
(2) Marshall, E. Science 1999, 284, 1439-1440.
(3) Dino-Simonin, N.; Brandt-Casadevall, C. Forensic Sci. Int. 1996, 81, 6172.
(4) Schneider, P. M. Forensic Sci. Int. 1997, 88, 17-22.
(5) Kruglyak, L. Nat. Genet. 1999, 22, 139-144.
3138 Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
Biological organisms possess built-in machinery for DNA
mismatch repair.6,7 In Escherichia coli, for instance, one of the DNA
repair pathways is initiated by the binding of a 97-kDa MutS
protein to a site of DNA damage, i.e., mismatched DNA base pairs.
After the binding of MutS, enzymes such as DNA helicase and
DNA polymerase work synergistically and repair the damaged
DNA. Although the binding affinity of MutS to eight different
mismatches, such as (A/A), (A/C), and (A/G), apparently varies,
the formation of MutS-DNA complexes has been reported to be
specific to all base-pair mismatches by electrophoretic mobility
shift assays or molecular images taken by electron microscopy.8-10
In this report, we describe a MutS-mediated method for
detection of DNA base-pair mismatches by using atomic force
microscopy (AFM).11 AFM is widely used to capture structural
images at nanometer resolution of biological molecules including
nucleic acids, proteins, and membranes.12-15 In analyzing DNA
sequence variations, AFM enables us to characterize individual
DNA molecules longer than the molecules commonly used for
electrophoretic mobility shift assay, DNase footprinting, or microfabricated DNA arrays. Previously we developed two AFMbased methods for straightening and immobilizing DNA and
protein-DNA complexes on an atomically flat mica surface.16-18
We applied a similar method for sample preparation and demonstrate here for the first time that MutS-DNA complexes are
detectable by AFM and the observed binding sites of MutS are
(6) Modrich P. Annu. Rev. Genet. 1991, 25, 229-253.
(7) Kolodner, R. D.; Marsischky, G. T. Curr. Opin. Genet. Dev. 1999, 9, 8996.
(8) Parsons, B. L.; Heflich, R. H. Mutat. Res. 1997, 374, 277-285.
(9) Bellanne-Chantelot, C.; Beaufils, S.; hourdel, V.; Lesage, S.; Morel, V.;
Dessinais, N.; Le Gall, I.; Cohen, D.; Dausset, J. Mutat. Res. Genomics 1997,
382, 35-43.
(10) Gotoh, M.; Hasebe, M.; Ohira, T.; Hasegawa, Y.; Shinohara, Y.; Sota, H.;
Nakao, J.; Tosu, M. Genet. Anal.: Biomol. Eng. 1997, 14, 47-50.
(11) Binnig, G.; Quate, C. F.; Gerber C. Phys. Rev. Lett. 1986, 56, 930-933.
(12) Hansma, P. K.; Elings, V. B.; Marti, O.; Bracker, C. E. Science 1988, 242,
209-242.
(13) Bustamante, C.; Vesenka, J.; Tang, C. L.; Rees, W.; Guthold, M.; Keller, R.
Biochemistry 1992, 31, 22-26.
(14) Radmacher, M,; Fritz, M.; Hansma, H. G.; Hansma, P. K. Science 1994,
265, 1577-1579.
(15) Florin, E. L.; Moy, V. T.; Gaub, H. E. Science 1994, 265, 415-417.
(16) Yokota, H.; Nickerson, D. A.; Trask, B. J.; van den Engh, G.; Hirst, M.;
Sadowski, I.; Aebersold, R. Anal. Biochem. 1998, 264, 158-164.
(17) Yokota, H.; Fung, K.; Trask, B. J.; van den Engh, G.; Sarikaya, M.; Aebersold,
R. Anal. Chem. 1999, 71, 1663-1667.
(18) Yokota, H.; Sunwoo, J.; Sarikaya, M.; van den Engh, G.; Aebersold, R. Anal.
Chem. 1999, 71, 4418-4422.
10.1021/ac991263i CCC: $19.00
© 2000 American Chemical Society
Published on Web 06/16/2000
Figure 1. The 449-bp DNA templates used in this study. Template
1 consisted of a single base-pair mismatch of (T/C) or (G/A) at a site
137 bp apart from one end. Template 2 consisted of a single basepair mismatch of (C/A) or (T/G) at a site 182 bp apart from the same
end. Template 3 consisted of two base-pair mismatches in the
templates 1 and 2.
Figure 2. AFM height image of MutS-DNA complexes. Height is
indicated by a color code with dark (0 nm) and light (5 nm). (A-C)
The MutS-DNA complexes corresponding to template 1. The specific
binding site of MutS was predicted at 0.305 (137 bp away from one
end of the 449-bp DNA template). (D-E) The MutS-DNA complexes
corresponding to template 2. The specific binding site of MutS was
predicted at 0.405 (182 bp away from one end of the 449-bp DNA
template).
in good agreement with the predicted location of mismatched
DNA base pairs.
MATERIALS AND METHODS
DNA Preparation. Six DNA templates were prepared by
synthesizing custom-made oligonucleotides, cloning them in a
plasmid, and hybridizing a pair of nucleotides consisting of one
or two base-pair mismatches (Figure 1). Briefly, three custommade oligonucleotides 100 nucleotides long were first amplified
by PCR, and the PCR products were digested with SacI and XhoI
to obtain 83-bp DNA fragments. The DNA fragments were then
inserted into the multiple cloning site of the plasmid (pBluescript
II KS, Stratagene), and the plasmids amplified in E. coli were
isolated using a plasmid isolation kit (Qiagen). After digestion of
the isolated plasmids by PvuII, the 449-bp DNA fragments used
in this study were recovered from a 1% agarose gel and purified
using a gel extraction kit (Qiagen). A pair of different DNA clones
were denatured and hybridized to form DNA templates containing
one or two base-pair mismatches. Assuming a random pairing of
449-nt DNA strands, half of the DNA duplexes were expected to
be heteroduplexes. Templates 1 and 2 were designed to consist
of a one base-pair mismatch, and template 3 was constructed to
consist of two base-pair mismatches.
Formation of MutS-Heteroduplex DNA Complexes. The
DNA duplexes, half of them consisting of one or two base-pair
mismatches, were each suspended in a 10 µL of buffer containing
10 mM Tris-HCl (pH 8.0), 5 mM KCl, and 2 mM MgCl2. To allow
the formation of MutS-heteroduplex DNA complexes, 1 µL of a
thermostable MutS suspension (Epicenter Technologies Corp.)
at 100 ng/µL was added to the DNA suspension and the mixture
was incubated at 22 °C for 30 min.
The mixture of DNA and MutS was then spread on a freshly
cleaved mica sheet (25 mm × 25 mm, Ted Pella Inc.) by using a
Figure 3. Histogram showing the normalized location of the bound
MutS on DNA templates. (A) Histogram corresponding to template
1. The mean and the standard deviation of the normalized binding
position were determined as 0.303 ( 0.018 (N ) 36), while the
predicted binding site was 0.305. (B) Histogram corresponding to
template 3. The mean and the standard deviation for two binding sites
were 0.295 ( 0.022 (N ) 15) and 0.398 ( 0.018 (N ) 32), while the
predicted sites were 0.305 and 0.405.
DNA spin-stretcher as described previously.17 Briefly, a mica sheet
was mounted horizontally on the spin-stretcher using a doublesided tape, and the mica plate was spun at ∼5000 rpm. A series
of droplets was gently dispensed on the spinning center with a
pipet, in the following order and with ∼30-s intervals: 50 µL of
H2O for prerinsing, 50 µL of 500 mM MgCl2 solution for precoating
the surface, 50 µL of H2O for rinsing, 10 µL of the sample solution,
and 50 µL of H2O for postrinsing the sample surface.
Imaging by Atomic Force Microscopy. A Nanoscope III
atomic force microscope (Digital Instruments, Inc.) was used to
capture images of MutS-heteroduplex DNA complexes immobilized on the mica surface. AFM was performed in the ambient
air at 15-20% humidity. The tapping mode was used to reduce
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
3139
Figure 4. AFM height image of single-MutS-DNA complex and double-MutS-DNA complex. To present 3-D structural features, the images
are displayed with a surface tilt of 20°. The white arrow indicates the bound MutS. (A) Single-MutS-DNA complex corresponding to template
1. (B) Double-MutS-DNA complex corresponding to template 3.
any damage to biological samples caused by physical contact with
the tip. The tapping frequency chosen was ∼200 kHz, a frequency
near the resonance of the cantilever. A scan field of view was set
to 2 µm × 2 µm with the scanning rate of 0.5-1 Hz and 512
scanning lines. The silicon tips we used had an estimated
curvature of 10-20 nm. The height bar was linear between 0 and
5 nm. Height images were flattened to remove the background
curvature of the mica surface, and the images were analyzed using
NIH Image 1.60 image analysis software. The normalized MutS
binding position was defined as d1/(d1 + d2), where d1 and d2
(d1 < d2) are the length of DNA segments bisected by MutS.
RESULTS
Detection of MutS-DNA Complex. To determine proper
biochemical conditions for detection of MutS-DNA complexes
by AFM, we investigated the effect of the DNA concentration and
the MutS concentration on the retention of MutS-DNA complexes
on the mica surface. Although a variation existed from one mica
surface to another, an increase in the MutS concentration
enhanced the formation of MutS-DNA complexes but it reduced
the retention of target molecules on the surface. In this report,
10 ng of DNA templates was incubated with 20 ng of 97-kDa MutS
in 10 µL of buffer. The ratio of MutS protein to DNA templates
was estimated to be ∼6. In a typical 2 µm × 2 µm field of view,
20-30 DNA molecules were detectable, and roughly 10% of the
DNA in the field was bound by MutS (Figure 2). Note that half of
the duplex DNA we used was expected to form a homoduplex
with no base-pair mismatch. We did not observe more than one
MutS bound to template 1 or 2, which consisted of a single basepair mismatch.
Determination of Specific MutS Binding Sites. To examine
the specificity of the binding of MutS to DNA base-pair mismatches, we analyzed the distribution of 83 binding sites on
3140
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
templates 1 and 3. A normalized site value between 0 and 0.5 was
assigned to each binding site by determining d1/(d1 + d2), where
d1 and d2 (d1 < d2) corresponded to two DNA segments bisected
by MutS. Figure 3 displays two such histograms for the templates
1 and 3, respectively. Each histogram represents the observation
made for a single incubation preparation. The mean and the
standard deviation of the binding sites on the template 1 were
determined as 0.303 ( 0.018 (N ) 36, sample number) with good
agreement with the predicted value of 0.305. For template 3
consisting of two base-pair mismatches, two local maximums
appeared in the observed distribution at 0.295 ( 0.022 (mean (
standard deviation, N ) 15) and 0.398 ( 0.018 (mean ( standard
deviation, N ) 32). The observed site values were in good
agreement with the predicted value of 0.305 and 0.405. The
normalized positional difference between the observation and the
prediction ranged from 0.002 to 0.01. Since all DNA templates
were 449 bp long, the maximum positional difference corresponded to 4-5 bp or 1-2 nm. In the same binding assay using
the template 3, MutS was bound 2.1 times as frequently to the
mismatch site of (C/A) or (T/G) as to the other site of (T/C) or
(G/A).
Detection of Double-MutS-DNA Complexes. We next
examined whether MutS proteins, tandemly bound to two point
mismatches separated by 45 bp in template 3, would be identifiable
as two proteins by AFM. We observed that ∼10% of the DNA
molecules on the mica surface formed MutS-DNA complexes,
and that ∼1% of all DNA templates exhibited a structure bound
to DNA with two globular parts (Figure 4). Compared to MutS
protein singly bound to template 1, an apparent pair of MutS
proteins on template 3 exhibited an elongated conformation along
DNA. In determining the size of the bound MutS, we defined their
outer edge using the height identical to the height of DNA. The
mean and the standard deviation of the bound MutS were
determined as 19.0 ( 0.3 × 15.0 ( 0.2 (major axis × minor axis,
N ) 20). In Figure 4, the single MutS footprint on DNA was
measured as 19 nm × 15 nm (major axis × minor axis), while
the footprint of apparent double MutS was 26 nm × 15 nm (major
axis × minor axis) with a constriction in the middle.
DISCUSSION
We describe a rapid and sensitive MutS-mediated method of
identifying the position of DNA base-pair mismatches using AFM.
For 449-bp DNA templates consisting of one or two base-pair
mismatches, the site of mismatches was determined from individual MutS binding sites within a few-nanometer deviation. The
observed binding of MutS to base-pair mismatches was specific
in our AFM-based assay. The difference between the observed
site and the prediction was less than 5 bp (2 nm) for all three
cases studied in this report, and the standard deviation was 8-9
bp (∼3 nm) in the distribution of 15-36 MutS binding sites. A
pair of MutS proteins, apparently bound tandemly at two binding
sites separated by 45 bp, was identifiable from a shape resembling
a cluster of two globular structures. The measurement was close
enough to locate DNA mismatches within a particular genomic
region, and stretching DNA molecules enabled to yield good
precision.
AFM is becoming a useful tool for analyzing genomic DNAs
mapping a recognition site of a restriction enzyme or constructing
a physical DNA map of protein-binding sites.19-21 Unlike conventional molecular tools such as electrophoretic mobility shift assays
or DNase footprinting, the AFM-based method allows us to use
large DNA templates from a minute amount of sample without
any modification or labeling of protein or DNA molecules. DNA
molecules over 100 kbp in length can easily be straightened by
the stretching apparatus we have developed and determining sites
(19) Allison, D. P.; Kerper, P. S.; Doktycz, M. J.; Spain, J. A.; Modrich, P.; Larimer,
F. W.; Thundat, T.; Warmack, R. J. Proc. Natl. Acad. Sci. U.S.A. 1996, 93,
8826-8829.
(20) Allison, D. P.; Kerper, P. S.; Doktycz, M. J.; Thundat, T.; Modrich, P.;
Larimer, F. W.; Johnson, D. K.; Hoyt, P. R.; Mucenski, M. L.; Warmack, R.
J. Genomics 1997, 41, 379-384.
(21) Wyman, C.; Rombel, I.; North, A. K.; Bustamante, C.; Kustu, S. Science 1997,
275, 1658-1661.
(22) Chee, M.; Yang, R.; Hubbell, E.; Berno, A.; Huang, X. C.; Stern, D.; Winkler,
J.; Lockhart, D. J.; Morris, M. S.; and Fodor, S. A. Science 1996, 274, 610614.
(23) Wang D. G.; et al. Science 1998, 280, 1077-1082.
of bound MutS along the uncoiled DNA is straightforward. Since
neither end of the DNA molecules was labeled in this report, the
site of MutS binding had a directional ambiguity. It is possible to
remove this ambiguity by including a reference base-pair mismatch close to one end.
Understanding variations of affinity and specificity among eight
possible mismatches will help interpret a distribution of MutSbinding sites. In template 3, for instance, four different kinds of
mismatches, such as (T/C), (G/A), (C/A), and (T/G), were
included at an equal frequency. One pair of (T/C) or (G/A)
mismatch, which was positioned at 137 bp from one end, gave a
fewer number of MutS-DNA complexes than the other pair of
(C/A) or (T/G) mismatch at 182 bp away from the same end. In
a study with the surface plasmon resonance sensor, base mismatch such as (T/C) or (T/T) was reported to exhibit lower
affinity to MutS than a mismatch such as (G/A) or (G/T).10
An AFM-based affinity study using DNA templates consisting of
one kind of mismatch only will help further characterize the
described MutS-mediated approach.
In conclusion, we have described a MutS-mediated method of
detecting DNA base-pair mismatches by AFM. In the post human
genome sequence era, identifying similarities and differences
among DNA samples is expected to be a growing task. The
described method provides a rapid and sensitive tool for identifying differences by placing a landmark at a site of base-pair
mismatches at nanometer resolution. This method is expected to
complement currently available hybridization-based methods
represented by DNA chip technologies, most of which are
designed to detect base-pair matches of relatively short oligonucleotides.22,23
ACKNOWLEDGMENT
We appreciate Satomi Ohnishi (National Institute of Materials
and Chemical Research, Japan), Helen Hansma (University of
California, Santa Barbara), Chun Li Bai (Chinese Academy of
Sciences, China), and David Mack (Indiana University) for
valuable suggestions. This work was in part supported by the
Whitaker Foundation and Sumitomo Electric Industries.
Received for review November 3, 1999. Accepted April 27,
2000.
AC991263I
Analytical Chemistry, Vol. 72, No. 14, July 15, 2000
3141