Detection of Point Mutation and Insertion Mutations in DNA Using a

Detection of Point Mutation and Insertion. Mutations in DNA Using a Quartz Crystal. Microbalance and MutS, a Mismatch Binding. Protein. Xiaodi Su,*,â€...
0 downloads 0 Views 155KB Size
Anal. Chem. 2004, 76, 489-494

Detection of Point Mutation and Insertion Mutations in DNA Using a Quartz Crystal Microbalance and MutS, a Mismatch Binding Protein Xiaodi Su,*,† Rudolf Robelek,‡ Yingju Wu,† Guangyu Wang,† and Wolfgang Knoll‡,§

Institute of Materials Research and Engineering, 3 Research Link, Singapore 117602, Department of Chemistry and Materials Science, National University of Singapore, Kent Ridge Road, Singapore 117542, and Max-Planck-Institut fu¨r Polymerforschung, Ackermannweg 10, 55128 Mainz, Germany

MutS protein is a mismatch binding protein that recognizes mispaired and unpaired base(s) in DNA. In this study, we incorporate the MutS protein-based mutation recognition into quartz crystal microbalance (QCM) measurements for DNA single-base substitution mutation and 1-4 base(s) insertion (or deletion) mutation detection. The method involves the immobilization of single-stranded probe DNA on a QCM surface, the hybridization of target DNA to form homoduplex or heteroduplex DNA, and finally the application of MutS protein for the mutation recognition. By measuring the MutS binding signal, DNA containing a T:G mismatch or unpaired base(s) is(are) discriminated against perfectly matched DNA at target concentrations ranging from 1nM to 5 µM. Furthermore, the QCM damping behavior upon MutS-DNA complex formation is studied using a Network Analyzer. The measured motional resistance changes per coupled MutS unit mass (∆R/∆f) are found to be indicative of the viscoelastic or structural properties of the bound protein, corresponding to different binding mechanisms. In addition, the ∆R/∆f values vary remarkably when the MutS protein binds at different distances away from the QCM surface. Thus, these values can be used as a “fingerprint” for MutS mismatch recognition and also used to quantitatively locate the mutation site. In human molecular and medical genetics, the vast majority of mutations and sequence polymorphisms in DNA result from single-base substitutions and small insertions or deletions of bases in a genome. Conventional methods for small genetic alternation detection include denaturing gradient gel electrophoresis, enzyme or chemical mismatch cleavage, and direct sequencing of polymerase chain reaction products.1 These methods typically require gel electrophoresis, which renders most of them unsuitable for automatic and rapid screening for polymorphisms. * To whom correspondence should be addressed. Tel: 65-68748420. Fax: 6568720785. E-mail: [email protected]. † Institute of Materials Research and Engineering. ‡ National University of Singapore. § Max-Planck-Institut fu ¨ r Polymerforschung. (1) Taylor, G. R. Laboratory Methods for the Detection of Mutations and Polymorphisms in DNA; CRC Press: Boca Raton, FL, 1997. 10.1021/ac035175g CCC: $27.50 Published on Web 12/13/2003

© 2004 American Chemical Society

DNA mismatch repair systems of a broad variety of organisms employ a family of proteins including those that recognize and bind mismatch-containing DNA. The MutS protein has been identified as such a component of the Escherichia coli mismatch repair system.2 Purified MutS protein binds DNA containing mispaired and unpaired bases but does not bind equally well to DNA without mismatches or single-stranded DNA. The use of immobilized MutS protein for single-nucleotide polymorphism detection was first reported by Wagner et al.3,4 Their nitrocellulose membrane-based chemiluminescence method does not require gel electrophoresis but involves multiple steps of blotting, blocking, washing, and staining. By this method, labeled DNA is detected at a detection limit of 100 pM. The use of chip or sensor methods to monitor MutS-DNA interactions has been reported in recent years. Begrensdorf et al. and Bi et al. each developed a DNA chip5 and a MutS protein chip6 for rapid screening of single-nucleotide polymorphism. The specific binding of dye-labeled MutS protein with surface-bound DNA or dye-labeled DNA with surface-bound MutS protein is revealed by the obtained fluorescence images. Using atomic force microscopy in combination with a DNA stretching technique, others visualized MutS-DNA complexes spread on mica sheets.7,8 By measuring the distance from the MutS binding site to DNA ends, one can locate the position of a mutation site. Being a labelfree, surface-sensitive technique, surface plasmon resonance (SPR) devices have been used to study MutS interactions with all eight possible mismatches.5,9-11 The binding reactions are monitored (2) Grilley, M.; Holmes, J.; Yashar, B.; Modrich, P. Mutat. Res. 1990, 236, 253-267. (3) Wagner, R. E.; Radman, M. Companion Methods Enzymol. 1995, 7, 199208. (4) Wagner, R. E.; Debbie, P.; Radman, M. Nucleic Acids Res. 1995, 23, 39443948. (5) Behrensdorf, H. A.; Pignot, M.; Windhab, N.; Kappel, A. Nucleic Acids Res. 2002, 30, e64. (6) Bi, L. J.; Zhou, Y. F.; Zhang, X. E.; Deng, J. Y.; Zhang, Z. P.; Xie, B.; Zhang, C. G. Anal. Chem. 2003, 75, 4113-4119. (7) Zhang, Y.; Lu, Y.; Hu, J.; Kong, X.; Li, B.; Zhao, G.; Li, M. Surf. Interface Anal. 2002, 33, 122-125. (8) Sun, H. B.; Yokota, H. Anal. Chem. 2000, 72, 3138-3141. (9) Gotoh, M.; Hasebe, M.; Ohira, T.; Hasegawa, Y.; Shinohara, Y.; Sota, J.; Nakao, H.; Tosu, M. Genet. Anal.: Biomol. Eng. 1997, 14, 47-50. (10) Blackwell, L. J.; Bjornson, K. P.; Allen, D. J.; Modrich, P. J. Biol. Chem. 2001, 276, 34339-34347.

Analytical Chemistry, Vol. 76, No. 2, January 15, 2004 489

Table 1. Nucleotide Sequences of the Probe and Target ssDNAs Used for the Assay type homoduplex T:G mismatch unpaired C unpaired CA unpaired CAG unpaired CAGG

nucleotide sequence

SEQ ID

5′-GCACCTGACTCCTGTGGAGAAGTCTGCCGT-3′ 3′-CGTGGACTGAGGACACCTCTTCAGACGGCA-5′ 5′-GCACCTGACTCCT GTGGAGAAGTCTGCCGT-3′ 3′-CGTGGACTGAGGACGCCTCTTCAGACGGCA-5′ 5′-ACGGCAGACT TCTCC CCAGGAGTCAGGTGC-3′ 3′-TGCCGTCTGAAGAGGCGGTCCTCAGTCCACG-5′ 5 -ACGGCAGACTTC TCC CCAGGAGTCAGGTGC-3′ 3′-TGCCGTCTGAAGAGGACGGTCCTCAGTCCACG-5′ 5′-ACGGCAGACTTC TCC CCAGGAGTCAGGTGC-3′ 3′-TGCCGTCTGAAGAGGGACGGTCCTCAGTCCACG-5′ 5′-ACGGCAGACTTC TCC CCAGGAGTCAGGTGC-3′ 3′-TGCCGTCTGAAGAGGGGACGGTCCTCAGTCCACG-5′

1 2 1 3 4 5 4 6 4 7 4 8

in real time without using labeled materials. All these methods do not require time-consuming blotting and staining procedures, and each has specific advantages. In this study, we combine for the first time the MutS-based mismatch recognition with quartz crystal microbalance (QCM) measurements for the detection of single T:G mismatch mutations and 1-4 base(s) insertion mutations in 30-bp DNA. This involves the immobilization of single-stranded probe DNA on a QCM sensor surface, the hybridization of single-stranded target DNA to form heteroduplex or homoduplex DNA, and finally the application of MutS protein for the mutation recognition. Throughout the binding reactions, QCM frequency and damping parameters are measured. Similar to the SPR sensors, QCM provides label-free, real-time measurements of interfacial binding reactions. Moreover, the QCM detection principle provides additional information about the structural and viscoelastic properties of adsorbed molecules. We believe that this advantage would be particularly useful for revealing MutS-DNA binding mechanisms and for locating mutation sites. We will demonstrate these possibilities in this study.

Table 1. Both the SEQ ID 1 and 4 were prepared with a biotin label at the 5′ end, serving as probe 1 and probe 4, respectively. The SEQ ID 2 (target 2) is fully complementary to probe 1, and the SEQ ID 3 (target 3) contains a single-base substitution (A f G) at the center and forms a T:G mismatch with probe 1. SEQ ID 5-8 (targets 5-8) are complementary to probe 4 but contain one to four additional bases between the 15th and 16th bases. Apparatus and Methods. AT-cut, 10-MHz quartz crystals with polished gold electrodes on both sides were purchased from International Crystal Manufacturers Inc. (Oklahoma City, OK). For in situ frequency measurement in liquid, the crystals were fixed into two Plexiglas blocks with Neoprene O-ring seals. The upper face of the crystals was exposed to 50 µL of sample solution. The frequency response was measured with PzTools hardware and software (Universal Sensors, Inc. Metairie, LA). Frequency stability is 1 Hz. For impedance analysis, the same setup was connected to a S&A 250B Network Analyzer (Saunders & Associates, Inc.), which recorded the frequency and damping simultaneously. The noise levels of the frequency and motional resistance were 4 Hz and 0.1 Ω, respectively.

EXPERIMENTAL SECTION Materials. Thermostable MutS protein derived from the thermophilic bacterium Thermus aquaticus was purchased from Epicentre Technologies (Madison, WI). This protein (2 µg/µL) is supplied in a 50% glycerol solution containing 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 1 mM dithiothreitol, 0.1 mM EDTA, and 0.1% Triton X-100. Streptavidin was purchased from Sigma-Aldrich (St. Louis, MO). The biotinylated and OH-terminated thiols used for preparation of the QCM surfaces were synthesized in our laboratory at the Max-Planck-Institute for Polymer Research (Mainz, Germany). Three buffer systems were involved in this study. (1) MutSDNA binding buffer was a 50 mM Tris-HCl (pH 7.5) solution containing 100 mM NaCl, 1 mM dithiothreitol, 0.1 mM EDTA, and 5 mM MgCl2; (2) HEPES buffer (50 mM HEPES, 25 mM NaOH, and 75 mM NaCl) provided an appropriate ionic strength and pH for the probe DNA immobilization and target DNA hybridization; (3) PBS buffer was used for streptavidin immobilization. Highly purified salt free oligonucleotides were obtained from MWG. The sequences of the oligonucleotides are summarized in

RESULTS AND DISCUSSION Figure 1a shows a schematic illustration of the QCM surface architecture and the assay procedures of probe immobilization, target hybridization, and MutS recognition of mismatched DNA. A biotin-strepavidin-biotin bridge chemistry was used for DNA probe immobilization. The gold electrodes of the QCM were first treated with a binary mixed-thiol compound consisting of 10% biotin-terminated thiol and 90% OH-terminated thiol (Figure 1b). This surface ensures maximal streptavidin binding and minimal nonspecific protein adsorption.12 Figure 2 shows the frequency response of such a QCM to the binding reactions showed in the scheme, starting from PBS calibration of the baseline. For better clarity, the assay procedures are described in five sequential steps. In step 1, streptavidin immobilization (0.2 mg/mL in PBS buffer) is observed as an irreversible, steady frequency drop, indicating an accumulation of the protein molecules at the interface over time. A subsequent BSA application (5 mg/mL in PBS) was intended to block the exposure of bare gold that might be present due to a low thiol coverage. Step 2 shows the response of the streptavidin/BSA surface to nonspecific MutS adsorption. For this purpose, a solution of 100

(11) Babic, I.; Andrew, S. E.; Jirik, F. R. Mutat. Res.: Fundam. Mol. Mech. Mutagen. 1996, 372, 87-96.

(12) Knoll, W.; Zizlsperger, M.; Liebermann, T.; Arnold, S.; Badia, A.; Liley, M.; Piscevic, D.; Schmitt, F.-J. Colloids Surf., A 2000, 161, 115-137.

490 Analytical Chemistry, Vol. 76, No. 2, January 15, 2004

Figure 1. (a) Schematic illustration of the surface architecture and the assay procedures of probe DNA immobilization, target DNA hybridization, and MutS recognition of mispaired/unpaired DNA. (b) Structural formulas of the thiols used for QCM surface preparation.

Figure 2. Frequency response of the binary thiol-modified QCM to the binding reactions shown in Figure 1a, starting from PBS calibration of the surface for steptavidin immobilization.

nM MutS protein (1:200 diluted from the stock solution using MutS-DNA binding buffer) was applied to the surface after setting the baseline with the same buffer system. This protein solution results in a ∼10-Hz baseline drop, which was entirely reversible upon rinsing the cell with the binding buffer. This signal reflects the viscosity change of the liquid medium. Since the MutS protein is supplied in Tris-HCl buffer containing 50% glycerol and 0.1% Triton X-100, it is believed that the remaining glycerol and Triton X-100 content in the 1:200 diluted protein solution is the cause of the bulk effect. Barely detectable MutS accumulation over time indicates the absence of nonspecific protein adsorption. In step 3, the carrier buffer was changed to HEPES for probe DNA immobilization and target DNA hybridization. Application of probe 1 (1 µM) results in a frequency drop of 32 Hz. The subsequent target 3 hybridization (5 µM) results in a frequency drop of 36 Hz. The resulting double-helix DNA contains a T:G

mismatch and is referred to as mismatch 1 (MM1) DNA. From the relative frequency changes, we estimated that the hybridization efficiency is nearly 100%. Switching between HEPES and Tris-HCl binding buffer was repeated in step 4 to check the stability of the double-helix DNA in these buffer systems. No loss of target DNA was found upon the buffer exchanges, indicating that these buffer solutions have appropriate salt concentration and pH to stabilize the double helix. Thus, we concluded that it was safe to use the Tris-HCl buffer for the MutS bindings. In step 5, to eliminate the bulk viscosity effect of the MutS protein, a MutS buffer blank solution (featuring exactly the same composition as the MutS stock solution except for the protein) was prepared. Prior to the MutS application, the frequency baseline was reset (∼10-Hz baseline shift) using this buffer blank at the same dilution. The subsequent MutS protein application Analytical Chemistry, Vol. 76, No. 2, January 15, 2004

491

Figure 3. Frequency responses of streptavidin/BSA-treated QCMs to probe 1 (1 µM) immobilization (arrows 1), MM0 target (upper curve) or MM1 target (lower curve) DNA hybridization at 5 µM (arrows 2), changing buffer system from HEPES to MutS binding buffer (TrisHCl) at arrows 3, resetting the surface by MutS buffer blank (1/200 diluted) at arrows 4, application of MutS protein (100 nM) at arrows 5, and finally rinsing the surface with the diluted MutS buffer blank at arrows 6.

results in a steady frequency decrease that can be solely attributed to MutS protein adsorption due to mismatch recognition. Rinsing the surface with the diluted buffer blank leads to a slight frequency increase, indicating a weak disassociation of the MutS-DNA complex. Using the above protocol (but omitting steps 2 and 4), experiments were conducted to investigate whether the MutS binding signals can provide discrimination between MM1 DNA and perfectly matched DNA (MM0 DNA). Figure 3 shows the frequency responses of strepavidin/BSA-treated QCM sensors to probe 1 immobilization, MM1 target (target 3) or MM0 target (target 2) hybridization at 5 µM, and MutS binding. At the given buffer and temperature conditions, the hybridization signals for the MM1 and MM0 DNA are very similar although a careful check of kon and koff can show different binding affinity.13 The subsequent MutS bindings, however, are remarkably different both in the amount bound and in the binding stability. The MutS interaction with MM0 DNA is not surprising as the clamp domains in the MutS proteins have certain contacts with the DNA backbone in a sequence-independent-manner.14 However, our results show that this contact is weaker than the hydrogen bonds formed in the heteroduplex DNA-MutS complex. Upon rinsing with the 1:200 diluted buffer blank, the frequency tends to rise toward its original level, indicating the loss of the MutS protein. From six experiments (three MM1 and three MM0), the MutS binding signals measured after rinsing (∆fMutS-MM1 and ∆fMutS-MM0) show a discrimination between the MM1 and MM0 DNA by a factor of 2.9 ( 0.3. For a T:G mismatch and perfectly matched DNA, the published SPR5,9-11 reported discrimination ratios of over 10. The low ∆fMutS-MM1 to ∆fMutS-MM0 ratio obtained here may be due to the fact that the QCM frequency is determined by both a mass effect and a viscoelasticity effect. For a protein film, water entrapped or hydrodynamically coupled between protein molecules makes the (13) Liebermann, T.; Knoll, W.; Sluka, P.; Herrmann, R., Colloids Surf., A 2000, 169, 337-350. (14) Lamers, M. H.; Perrakis, A.; Enzlin, J. H.; Winterwerp, H. H. K.; Wind, N.; Sixma, T. Nature 2000, 407, 711-717.

492

Analytical Chemistry, Vol. 76, No. 2, January 15, 2004

film soft and flexible. The soft and flexible film can induce dissipations detected as additional frequency decreases.15 Based on our understanding of MutS-DNA binding mechanisms, we predicted that the MutS protein interacts with the MM0 DNA in such a manner that the resulting protein layer is more dissipative than that with the MM1 DNA. The damping behavior of the QCM measured as motional resistance change supports this interpretation (see below). In a titration experiment, we found that a QCM sensor with immobilized probe 1 can sense MM1 and MM0 target hybridization down to concentrations of ∼5 nM. MutS binding reactions were measured in the concentration range of 5 nM to 5 µM. Figure 4 shows MutS-MM1 DNA interactions at MM1 target concentrations of 20 nM, 100 nM and 1 µM. In the first part of this figure, probe 1 immobilization is followed by MM1 target hybridization at 20 nM. This target causes a slow frequency decrease of 12 Hz at saturation. The subsequent MutS binding is detected as a frequency decrease of 89 Hz (after rinsing). For a MM0 DNA at this concentration, the MutS binding signal is only 34/18 Hz (before rinsing/after rinsing, curve not shown). The discrimination ratio between MM1 and MM0 is 5. At the beginning of the second part of Figure 4, the bound MutS protein was entirely removed by rinsing the surface with a buffer containing no dithiothreitol (DTT) and MgCl2 (HEPES in this case). The absence of DTT as antioxidant agent probably leads to a formation of intra- and intermolecular cysteine bridges and thereby causes conformational changes in the protein structure. Thus, the MutS protein cannot bind to DNA anymore and is released from the surface. By referring to the frequency level in the HEPES buffer, one can see that the bound probe/target DNA is fully exposed and available for either another cycle of MutS binding (∆fMutS-MM1 ) 89 Hz) or further hybridization with targets at higher concentrations (100 nM followed by 1 µM), and subsequently MutS binding (∆fMutS-MM1 ) 96 Hz). Up to now, we found that the MutS binding signals with MM1 DNA (∆fMutS-MM1) are less dependent on the density of the surfacebound double-stranded DNA. For MM1 target at a concentration of 20 nM, for example, we assume that only ∼30% of the immobilized probe DNA is occupied (by referring to the ∆f of 12 and 36 Hz for the target at 20 nM and 5 µM, respectively). The ∆fMutS-MM1, however, remains 77% of that obtained with the probes being fully occupied by the target at 5 µM (115 Hz). This is very likely due to the fact that the protein molecules (MW ) 98 000) are much bigger than the surface-bound double-stranded 30-bp DNA molecules (MW ≈ 18 000). Even though the theoretical binding ratio between MutS protein and a single-base-mismatched DNA in solution is 1:1, on the QCM surface there is not enough space for the MutS protein to bind every DNA molecules when the DNA density is high. In contrast, we found that the ∆fMutS-MM0 values at low MM0 target concentrations are much lower compared to that at 5 µM. Figure 5 summarizes the MM1 target hybridization signals at concentrations ranging from 1 nM to 5 µM, and the corresponding ∆fMutS-MM1. With the reference of the ∆fMutS-MM0 values that are also included in the figure, we conclude that single-basemismatched DNA can be distinguished from fully matched DNA (15) Hook, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796-5804.

Figure 4. Frequency response of a streptavidin/BSA-treated QCM to probe 1 immobilization (1 µM), MM1 target hybridization at 20 nM, and MutS binding (100 nM) upon resetting the surface with diluted MutS buffer blank. After MutS binding, a DTT- and MgCl2-free buffer solution (HEPES in this case) is applied to remove the bound MutS protein. The exposure of the bound probe/target DNA allows a new cycle of MutS binding or further hybridization at higher target concentrations.

Figure 5. Summaries of the frequency signals of MM1 target hybridization at concentrations ranging from 1 nM to 5 µM and the corresponding MutS binding signals (∆fMutS-MM1). The MutS binding signals on MM0 DNA (∆fMutS-MM0) in the same concentration range is also included.

in a wide concentration range with the discrimination ratios (∆fMutS-MM1/∆fMutS-MM0) of 3-5. It is worth mentioning that, even for target DNA concentrations lower than the limit for hybridization detection, 1 nM for example, the MutS-MM1 DNA binding signal is still remarkable. This detection limit (1 nM) is several tens of times lower than the one published for a label-free SPR method.9 Through a further optimization of the reaction conditions (i.e., MutS concentration and DNA probe density etc), it may be possible to further lower the detection limit toward that of the membrane-based chemiluninescent method (100 pM4). MutS binding reactions with DNA containing unpaired base(s) are detected in the following experiments. First, probe 4 was immobilized on the QCM crystal. Targets 5-8 were then allowed to anneal to form heteroduplex DNA oligonucleotides containing unpaired C, CA, CAG, or CAGG loops. MutS recognition and binding to these DNA strands results in frequency changes of 90 (unpaired C), 92 (unpaired CA), 76 (unpaired CAG), and 70 Hz (unpaired CAGG), respectively. This result corresponds well with the previous findings that MutS protein recognizes heteroduplexes containing one to four unpaired bases, but not all four equally well.3,6 The order of the detection sensitivity appeared to be two unpaired bases > (or )) one unpaired base > three unpaired bases > four unpaired bases.

It is well known that the QCM damping properties can be evaluated by measuring the energy dissipation factor (D) from a nondriven crystal16 or the so-called motional resistance (R) derived in a QCM equivalent circuit.17 Changes of these parameters are reflective of the mechanical/structural and viscoelastic properties of the adsorbed molecules. From the induced energy dissipation per coupled unit mass (∆D/∆f or ∆R/∆f ), one can estimate the flexibility of a deposited film.16-19 Typically, larger ∆D/∆f or ∆R/ ∆f values indicate the formation of flexible, dissipative films. In contrast, lower ∆D/∆f or ∆R/∆f values indicate mass additions without a significant dissipation increase. In this study, a Network Analyzer was used to measure MutS binding reactions with surface-immobilized MM1 dsDNA, MM0 dsDNA, and dsDNA containing an unpaired CAG loop. The frequency and resistance responses were recorded simultaneously (Figure 6). The application of the diluted MutS buffer blank results in a R increase together with a frequency decrease, both reflecting the increase in liquid density and viscosity.15,17 The subsequent MutS protein application results in an irreversible R increases of 0.77 ( 0.11 (MM1), 1.33 ( 0.20 (MM0), and 1.70 ( 0.21 Ω (unpaired CAG, curve not shown) at saturation (average from three duplicate measurements), respectively. The ratios between the ∆R and their corresponding ∆f, are 6.7 ( 0.5 mΩ/Hz (MM1), 26.3 ( 3.2 mΩ/Hz (MM0), and 17.0 ( 2.1 Ω/Hz (unpaired CAG), respectively. It has been reported that the MutS binding to mismatch containing DNA, a T:G mismatch for example, is based on hydrogen bond formation between the mismatch binding domain of the protein and the guanine/thymine base pair and, additionally, on a wedging of the protein into the flexible backbone of the mismatched or looped-out bases. The extremely low ∆R/ ∆f value for the MutS binding with MM1 DNA (T:G mismatch) measured here is a strong indication of the strong binding, that makes the protein molecule layer less flexible. In comparison, the ∼4 times higher ∆R/∆f value measured for the MutS-MM0 DNA (16) Rodahl, M.; Hook, F.; Fredriksson, A.; Krozer, C. A.; Keller, P.; Brzezinski, P.; Voinova, M.; Kasemo, B., Faraday Discuss. 1997, 107, 229-246. (17) Calvo, E. J.; Danilowicz, R.; Etchenique, R. J. Chem. Soc., Faraday Trans. 1995, 91, 4083-4091. (18) Hook, F.; Ray, A.; Norden, B.; Kasemo, B. Langmuir 2001, 17, 83058312. (19) Xie, Q.; Wang, J.; Zhou, A.; Zhang Y.; Zhang, H.; Liu, Z.; Xu, Y.; Yuan, M.; Deng, S.; Yao, S. Anal. Chem. 1999, 71, 4649-4656.

Analytical Chemistry, Vol. 76, No. 2, January 15, 2004

493

that MutS protein molecules binding to the DNA strands far away from the QCM surface would induce a higher energy loss as the resulting protein-DNA complexes are difficult to couple with the oscillation. For this study, a 30-mer target DNA (5′-ACGGCGGACTTCTCCACAGGAGTCAGGTGC-3′) was hybridized to the immobilized probe 1 to form a MM1 double-stranded DNA containing a T:G mismatch at 25 bases away from the surface. The subsequent MutS protein application results in a frequency change of 98 ( 5 Hz at saturation (curve not shown), which is similar to the T:G mismatch at the center (15 bases away from the surface). However, the induced ∆R are higher (1.05 ( 0.12 Ω), leading to a larger ∆R/∆f ratio of 11.0 ( 1.3 mΩ/Hz. This result would be reasonable because the 30-bp DNA is rigid enough to provide a higher distance for MutS binding 10 bases further away.

Figure 6. Simultaneous frequency (9) and motional resistance ([) responses measured from a Network Analyzer to MutS binding reactions with MM1 dsDNA (A) and MM0 dsDNA (B). The initial baselines are obtained in Tris-HCl buffer. The solid arrows indicate the time of injection of the diluted MutS buffer blank, and the dashed arrows the injection of MutS protein (100 nM).

reflects the loose contact of the protein with the DNA backbone. The bound protein molecules tend to be more flexible and induce more viscoelastic contribution to the frequency signal in addition to their mass contribution. These results explain the low discrimination (∆fMutS-MM1/∆fMutS-MM0 ratios) for the MutS binding on MM1 and MM0 DNA observed by QCM method. In comparison to MM1 and MM0 DNA, the intermediate ∆R/∆f value for the MutS-unpaired CAG DNA may imply a different binding mechanism that involves mainly wedging of the protein into the DNA strands, without hydrogen bonds formation. The distinct indication of the ∆R/∆f values to QCM damping caused by MutS-DNA interactions encouraged us to try to use this value to locate the position of a mutation site. One could image

494

Analytical Chemistry, Vol. 76, No. 2, January 15, 2004

CONCLUSION This is the first demonstration of using of QCM sensors to detect MutS-DNA binding reactions. By measuring the MutS binding signals, DNA containing a single T:G mismatch and one to four unpaired base(s) has been discriminated from perfectly matched DNA down to target concentrations of 1nM. This detection limit is several tens of times lower than the one reported for the label-free SPR method. QCM damping measured as motional resistance changes is distinctly indicative of the viscoelastic property of the bound MutS protein attributed to different binding mechanisms. The induced energy dissipation per coupled unit mass (∆R/∆f values) can be used as a “fingerprint” for mutation recognition and also used to locate the position of a mutation site. Further studies will focus on the use of the QCM method to screen single nucleotide mismatches, and small deletions (or insertions) in wide regions of DNA, which is essential for the detection of genomic polymorphism and for the identification of unknown structural alternations in DNA. In addition, using a reversed reaction scheme, that is, immobilizing MutS protein and passing through DNA, would be also interesting for SNP screening in large DNA fragments.

Received for review October 3, 2003. Accepted November 10, 2003. AC035175G