Anal. Chem. 2005, 77, 3709-3714
Correspondence
Viscoelastic Modeling of Template-Directed DNA Synthesis Gudrun Stengel,*,†,‡ Fredrik Ho 1o 1 k,*,§ and Wolfgang Knoll†,‡
Max-Planck-Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany, Department of Materials Science and Chemistry, National University of Singapore, 10 Science Drive 4, 117543 Singapore, and Division of Solid State Physics, Lund University, S-22100 Lund, Sweden
In the present study, we have used the QCM-D technology to study the replication of surface attached oligonucleotide template strands using Escherichia coli DNA polymerase I (Klenow fragment, KF). Changes in resonance frequency (F) and energy dissipation (D) for DNA hybridization and polymerization were recorded at multiple harmonics. Formation of the polymerase/DNA complex led to a significant decrease in energy dissipation, which is consistent with a conformational change induced upon enzyme binding. This interpretation was further strengthened by a data analysis using a Voigt-based viscoelastic model. The analysis revealed a significant increase in shear viscosity and shear modulus during KF binding, whereas the viscoelastic properties of single- and doublestranded templates were almost identical. During the actual DNA synthesis, an initial increase in rigidity (shear viscosity) was followed by a gradual decrease that has two components corresponding to the release of enzyme and to the presence of the catalytically active enzyme/substrate complex. The corresponding decrease in surface concentration was found to underestimate the rate of enzyme release due to viscously coupled water that compensates for the loss in enzyme mass. Furthermore, the modeling elucidates that significant changes in both F and D originate from variations in the viscoelastic properties, which means that changes in F alone should be used with care for estimations of coupled mass and kinetics. Therefore, the modeled temporal variation in effective thickness, being proportional to coupled mass and, thus, independent of structural changes, was used to estimate the catalytic constants of the polymerization reaction. The reported work is the first example providing this type of structural information for the catalytic action of an enzyme, thereby demonstrating the potential of the technique for advanced analysis of complex biological reactions, including proper analysis of enzyme kinetics. The accurate replication of genetic information is an indispensable process for every living organism. In cells, DNA polymerases * Corresponding authors. G.S. current address: The Scripps Research Institute, MB-19, 10550 North Torrey Pines Road, La Jolla, CA 92037. Phone: (858) 784-9046. Fax: (858) 784-9067. E-mail:
[email protected]. (F.H.) Phone: +46-(46) 2221494. Fax: +46-(46) 2223637. E-mail:
[email protected]. † Max-Planck-Institute for Polymer Research. ‡ National University of Singapore. § Lund University. 10.1021/ac048302x CCC: $30.25 Published on Web 04/29/2005
© 2005 American Chemical Society
are the enzymes in charge of template-directed DNA synthesis. Extension of the primer DNA is achieved by the stepwise addition of the appropriate deoxynucleoside triphosphates (dNTPs) to the 3′-OH terminus. As a result, the DNA chain grows in the 5′-to-3′ direction.1 The structure of prokaryotic DNA polymerases resembles a human right hand with the polymerization site in the palm, the dNTP binding region in the fingers and the pocessivity factors in the thumb domains. Accordingly, these enzymes enclose DNA in the bound state, thereby imposing a number of steric restrictions on DNA.1,2 In the present study, we reproduced the process of DNA polymerization in a biosensor format using the Klenow fragment (KF) of Escherichia coli DNA polymerase I. Recently, surface-based analytical tools, in particular, surface plasmon spectroscopy (SPR) and quartz crystal microbalance (QCM), have been proven useful for studies of DNA polymerase binding3 and catalytic activity.4,5 The merit of surface analytical techniques is the possibility to monitor, in real time, a sequence of reaction steps, such as enzyme binding, extension along the template, and enzyme release, using the same device. However, separation of the single steps is not trivial in a reaction, in which polymerase binding and DNA growth add mass, whereas mass is lost during the release of enzyme after complete DNA synthesis. For QCM measurements in liquid, a further complication arises from surface-coupled water that is sensed as additional mass.6,20 Matsuno et al.4,5 measured the kinetic parameters (kcat and KM) (1) Patel, P. H.; Loeb, L. A. Nat. Struct. Biol. 2001, 8, 656-659. (2) Steitz, T. A. Nature 1998, 391, 231-232. (3) Tsoi, P. Y.; Yang, J.; Sun, Y. T.; Sui, S. F.; Yang, M. S. Langmuir 2000, 16, 6590-6596. (4) Matsuno, H.; Niikura, K.; Okahata, Y. Chem.sEur. J. 2001, 7, 33053312. (5) Niikura, K.; Matsuno, H.; Okahata, Y. J. Am. Chem. Soc. 1998, 120, 85378538. (6) Ho ¨o ¨k, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796-5804. (7) Larsson, C.; Rodahl, M.; Ho ¨o ¨k, F. Anal. Chem. 2003, 75, 5080-5087. (8) Rodahl, M.; Kasemo, B. Rev. Sci. Instrum. 1996, 67, 3238-3241. (9) Ho ¨o ¨k, F.; Rodahl, M.; Brzezinski, P.; Kasemo, B. Langmuir 1998, 14, 729734. (10) Knoll, W.; Zizlsperger, M.; Liebermann, T.; Arnold, S.; Badia, A.; Liley, M.; Piscevic, D.; Schmitt, F. J.; Spinke, J. Colloids Surf. A 2000, 161, 115137. (11) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391-396. (12) The streptavidin layer was excluded during the modeling because it can be considered perfectly elastic.
Analytical Chemistry, Vol. 77, No. 11, June 1, 2005 3709
and the affinity constants (KA) for KF acting at surface-bound oligonucleotides. They quantified the absolute amount of synthesized DNA by calibrating the system with known mass loads, assuming a linear relationship between mass uptake (∆m) and decrease in frequency (∆F). Although misinterpretations caused by the influence of coupled water are reduced in that way, this approach ignores the viscoelastic nature of DNA films and withholds information obtainable from an exhaustive analysis using a viscoelastic model.7 The application of such a model requires the knowledge of the frequency shifts and of the corresponding damping the quartz resonator experiences, quantified by the energy dissipation (∆D). These data can be acquired simultaneously using the quartz crystal microbalance with dissipation monitoring (QCM-D) technology (see refs 8, 9 for technical details). The merits of piezoelectric mass-sensing devices in biosensor application are reviewed in ref 22. In this work, we used the QCM-D technique to follow the enzymatic extension of surface-attached primer oligonucleotides. We demonstrate that the QCM-D response can be efficiently represented with a Voigt-model-based viscoelastic representation.11 Using modeling, the QCM-D response was translated into effective viscoelastic parameters of the probed film(s), which are here being discussed on the basis of structural changes that the DNA polymerase imposes on DNA upon binding. In addition, we present a brief kinetic analysis of the enzymatic reaction, demonstrating that even the presence of low amounts of enzyme are sufficient to affect the simple linear relationship between frequency and mass changes. The experiments were conducted by attachment of 5′-biotinylated 15-mer primer oligonucleotides (P15) to the gold electrodes of a quartz crystal via standard streptavidin/biotin coupling,10 as also schematically illustrated in Figure 1. Hybridization to a 50mer template (T50) produced a DNA substrate with a 5′-overhang of 35 single-stranded nucleotides. A complete DNA duplex was created either by enzymatic DNA synthesis or by hybridization of an oligonucleotide (T35) being complementary to the singlestranded portion of the template strand. MATERIALS AND METHODS Surface Preparation. Prior to surface modification, the gold electrodes of the quartz were cleaned by immersion in a 1:1:6 mixture of H2O2 (30%, Merck), NH3 (25%, Merck), and MilliQ water (Millipore) for 20 min at 60 °C. A streptavidin layer was immobilized on the gold electrodes using a self-assembled (13) Note that ∆mVoigt represents the sum of adsorbed molecules and water entrapped in the film. By SPR and QCM data for the present system, a water content of 90% and an effective density of F ) 1.06 g/cm3 was estimated, of which the latter value was used to estimate dVoigt. (14) Li, Y.; Korolev, S.; Waksman, G. EMBO J. 1998, 17, 7514-7525. (15) Furrey, S.; Joyce, CM.; Osborne, M.; Klenermann, D.; Peliska, J.; Balasubramanian, S. Biochemistry 1998, 37, 2979-2990. (16) Datta, K.; LiCata, V. J. J. Biol. Chem. 2003, 278, 5694-5701. (17) Kuchta, R. D.; Mizrahi, V.; Benkovic, P. A.; Johnson, K. A.; Benkovic, S. J. Biochemistry 1987, 26, 8410-8417. (18) Polesky, A. H.; Steitz, T. A.; Grindley, N. D.; Joyce, C. M. J. Biol. Chem. 1990, 265, 14579-14591. (19) Singh, K.; Modak, M. J. J. Biol. Chem. 2003, 278, 11289-11302. (20) Reimhult, R.; Larsson, C.; Kasemo, B.; Ho¨o ¨k, F. Anal. Chem. 2004, 76, 7211-7220. (21) Cho, Y. K.; Kim, S.; Lim, G.; Granik, S. Langmuir 2001, 17, 7732-7734. (22) Janshoff A.; Galla H. J.; Steinem C. Angew. Chem. Int. Ed. 2000, 39(22), 4004-4032.
3710 Analytical Chemistry, Vol. 77, No. 11, June 1, 2005
Figure 1. Schematic representation of the surface architecture based on attachment of biotinylated 15-mer primer oligonucleotides (P15) to a streptavidin-modified gold electrode of a quartz crystal followed by hybridization to a 50-mer template (T50). This produces a DNA substrate with a 5′-overhang of 35 single-stranded nucleotides, and extension of primer DNA is achieved by addition of polymerase and dNTPs, either stepwise or mixed. Also shown are the physical parameters used to represent the viscoelastic properties of the layers.
monolayer (SAM) of thiols (formed ex situ), consisting of 11-mercapto-(8-biotinamido-4,7,dioxaoctyl)-undecanoylamide and 1-mercaptoundecanole as lateral spacer. For a detailed description of the SAM/streptavidin system, see ref 10. The streptavidin layer functioned as the platform for the immobilization of biotinylated oligonucleotides. Primer and template oligonucleotides (MWG Biotech, Germany) of the following sequences were used: 5′biotin-ACG TCA GTC TCA CCC-3′ (P15), 5′-AGT TAC AGA GGT AGT AGT GGC TGA GTG AAT ATT GT G GGT GAG ACT GAC GT-3′ (T50), and 5′-ACA ATA TTC ACT CAG CCA CTA CTA CCT CTG TAA CT-3′ (T35). Enzymatic Reaction. DNA extension assays were carried out using an exonuclease-free mutant of the Klenow fragment of DNA polymerase I from E. coli (30 nM) in the presence of a 10 µM mixture of all four dNTPs (both purchased from Amersham Pharmacia, Sweden). All reaction steps were conducted in HSM buffer (10 mM HEPES, 150 mM NaCl, 10 mM MgSO4, pH 7.4). Data Acquisition. QCM-D measurements were performed using the Q-Sense D300 system (from Q-Sense AB, Go¨teborg) using polished AT-cut quartz crystals with fundamental frequencies of ∼5 MHz (from Q-Sense AB, Go¨teborg). All measurements were carried out at a constant flow rate of 0.25 mL/min in a temperature-stabilized chamber (T ) 21(0.2 °C). Changes in ∆F and ∆D were recorded up to the fourth harmonics. For details about the modeling, see Supporting Information. RESULTS AND DISCUSSIONS QCM-D Response at Multiple Harmonics. Figure 2 shows a typical QCM-D measurement (∆D and ∆F/n versus time at two harmonics: n ) 5 and 7). After formation of the (rigid: ∆ ∼ -22 Hz and ∆D ∼ 0) streptavidin layer (not shown), the sensor surface was exposed to the P15 strand, which was subsequently hybridized with template T50. In agreement with previous studies on DNA hybridization,7 F decreases while D increases in both reaction steps. The so-created DNA duplex exhibits a specific binding site for KF, which motivated three different routes for the continuation of the experiment: simultaneous addition of KF and dNTPs (1); hybridization of T35 (2); simulating the response expected for successful primer extension (cf. route 1); and binding
Figure 2. ∆F (top) and ∆D (bottom) versus time at n ) 5 (red) and 7 (blue) for the formation of duplex (P15 and T50) followed by its enzymatic extension (1). (Data were offset after immobilization of the streptavidin layer.) Also shown are superimposed data obtained upon T35 hybridization (2) at n ) 5 and KF binding in the absence of dNTPs (3) after formation of the P15/T50 template. The best fit between the data and the Voigt viscoelastic model is shown as open circles (n ) 5) and open squares (n ) 7) for 1.
of KF in the absence of dNTPs (3). Clearly, curve 1 is a superposition of curves 2 and 3: Bare KF binding (3) induces a monotonic decrease in both D and F, whereas hybridization (2) induces an increase and decrease in D and F, respectively. Under conditions enabling DNA polymerization (1) there is an initial decrease in F and D followed by an increase of both parameters. Note that the responses at saturation are in good agreement for the two cases in which complete DNA duplex formation is expected (1 and 2). Viscoelastic Modeling. The QCM-D response was further analyzed using a Voigt-based viscoelastic model,11 which allows for translating the response in F and D into three effective physical parameters: mass () thickness (dVoigt) × density (Feffective)), shear viscosity (ηVoigt), and shear elastic modulus (µVoigt).7,6 The best fits are depicted together with the raw data in Figure 2, demonstrating a deviation of
