Using Nanoscale Substrate Curvature to Control the Dimerization of a. Surface-Bound Protein

Transcription

1 Using Nanoscale Substrate Curvature to Control the Dimerization of a Surface-Bound Protein Martin Kurylowicz, Maximiliano Giuliani and John R. Dutcher 1 Department of Physics, University of Guelph, Guelph ON, Canada N1G 2W1 1 Correspondence: Address reprint requests to John Dutcher, Department of Physics, University of Guelph, Guelph, Ontario, Canada N1G 2W1; Tel: , ext ; Fax: ; dutcher@uoguelph.ca 1

2 Supplemental Information SI1. Description of Fitting Criteria For Single Molecule Force Spectroscopy Data To obtain statistically relevant information from the single molecule force spectroscopy data, it is necessary to collect many force-distance curves and to use a consistent, robust procedure to identify and analyse those curves that are judged to be acceptable. Manual procedures for identifying and analysing large numbers of force-distance curves are time intensive and subject to bias. Instead, we have established a minimum set of criteria by which to judge the acceptability of a given force-distance curve that allowed us to implement automated processing of the curves. A force-distance curve was considered to be acceptable if: (1) it contained at least one saw-tooth peak with a maximum force that was larger than the noise background (~50 pn); (2) the detachment peak occurred at a tip-sample separation distance that was larger than 5 nm, which excluded peaks that were due to non-specific interactions between the AFM tip and the sample; (3) the data points beyond the detachment peak in the curve corresponded to a flat baseline with near-zero force; and (4) the data points within the detachment peak tracked continuously to the baseline and were well-described by the worm-like chain (WLC) function: F L p k B T = 1 " 4 1 x % $ ' # L c & x L c (S1) where F is the force experienced by the AFM tip, x is the separation distance, L c is the contour length, L p is the persistence length, k B is Boltzmann s constant and T is the absolute temperature. To perform an automated analysis of the force-distance curves, the raw data were exported in ASCII format from the data acquisition software. These files were batch processed using custom routines written in Octave ( [S1]. The raw extension data values were converted to distance from the substrate, and the raw force data values were shifted so that the average of the force values measured for the largest 20% of the distance data values was zero (see Figure SI1). A local minimum algorithm was then applied to the data to identify the detachment peak for each curve. The set of 2

3 criteria for the identification of well-defined detachment peaks described above was applied, which allowed ill-behaved local minima (such as those indicated by the green ellipses in Figure SI1) to be discarded. We note that the raw measured force values, with no smoothing of the data (as shown in Figure SI1), were used in applying the evaluation criteria. The best-fit value of the contour length L c for each detachment peak was obtained by fitting the detachment peak data to Equation S1. The fitting was performed using a Levenberg-Marquardt nonlinear regression with two fixed parameters, L p = 0.4 nm (typical value for polypeptides [S2,S3]) and F(0) = 0, and one free parameter, L c. In some situations, the fitting was sensitive to the initial guess of L c. To overcome this sensitivity, fits were performed using 20 different initial guesses, which were chosen as 20 equidistant points extending from the maximum force in the detachment peak (e.g., the red arrow on the horizontal axis in Figure SI1) to an additional distance of 50 nm away from the substrate. Using only those fits that converged, the mean of the best-fit L c values was taken as the final detachment length value. Also, for each of the force-distance curves included in the calculation of the mean value of L c, a plot of the force-distance curve together with the best-fit WLC curves was exported. These plots were examined to check if the mean L c value was consistent with the experimental points. If not, the data sets were fitted manually using the Nanoscope software (Bruker AXS) and, if the resulting L c value was also not consistent with the data, the data set was discarded. The above procedure allowed us to construct reliable histograms of the measured detachment length L c, as shown in Figure 5 of the manuscript. [S1] Eaton, J.W.; Bateman, D.; Hauberg, S., GNU Octave Manual Version 3, 2008, Network Theory Ltd., Bristol, UK. [S2] Rief, M.; Gautel, M.; Oesterhelt, F.; Fernandez, J.M.; Gaub, H.E., Reversible unfolding of individual titin immunoglobulin domains by AFM. Science 1997, 276,

4 [S3] Bouchiat, C.; Wang, M.D.; Allemand, J.; Strick, T.; Block, S.M.; Croquette, V., Estimating the persistence length of a worm-like chain molecule from force-extension measurements. Biophys. J. 1999, 76, Figure SI1: A representative force-distance curve for single molecule force spectroscopy (SMFS) measurements of a layer of β-lg molecules on a PS substrate. Only 1 of every 3 points of the raw data is shown for clarity. The data analysis routines were used to identify the detachment peak in each force-distance curve, and to discard illbehaved minima (such as those shown by the green ellipses). A well-behaved detachment peak is shown with a red curve that is the best fit of the detachment peak data to Equation S1 (Region II). The inset shows a schematic diagram of the SMFS measurement for a protein dimer. 4

5 SI2. Composite Histograms For the Bare Flat PS Surfaces In the present manuscript, histograms of the detachment length L c were presented in Figure 5 for both flat and curved PS surfaces. For completeness, we present a composite L c histogram for force-distance curves collected on the bare flat regions of all of the samples used in the present study in Figure SI2A. These composite data sets contain a very large number of detachment peaks, all of which satisfy the selection criteria described in Section SI1. The composite histogram contains very well defined peaks at L c ~ 50 nm and L c ~ 120 nm, which is consistent with each of the bare flat surface histograms shown in Figure 5. In addition, it can be seen in the histogram in Figure SI2A that the main peak centred at L c ~ 50 nm has a significant asymmetry, with a larger tail for larger L c values. In Figure SI2B, we show a corresponding composite histogram of the measured detachment force F d values collected on the bare flat regions of all of the samples used in the present study. There is a well-defined peak in this distribution at F d ~ 150 pn, with a significant tail to larger F d values. We note that the number of data points N in the L c and F d histograms is considerably larger (N > 1000) than is commonly found in the literature (N ~ ). The statistics in the composite histograms are sufficiently good that the peak shapes can be studied in detail. We are currently developing models of the AFM protein pulling experiment to attempt to elucidate the processes that are consistent with the detailed shapes of the composite L c and F d histograms. 5

6 Figure SI2: Composite histograms of SMFS data for β-lg collected on the bare flat PS regions of all of the samples used in the present study: A) detachment length L c, B) detachment force F d. The total number of data points N and the bin width are shown in the upper right of each plot. 6

7 SI3. Representative Force-Distance Curves For Different Surfaces In Figure SI3, we show representative force-distance curves for bare PS NPs of radius 30 nm, PS NPs of radius 30 nm capped with a 10 nm thick PS film, and a capped flat PS surface. Figure SI3: AFM images (upper) and representative force-distance curves for (left) bare PS NPs of radius 30 nm, (center) PS NPs of radius 30 nm capped with a 10 nm thick PS film, and (right) a capped flat PS surface. In each AFM image, a black circle indicates the spot at which force-distance curves were collected. 7