endothelin receptor antagonist br Experimental design materi

Experimental design, materials and methods
Acknowledgments This work was supported by Colon Cancer Canada and support for Dr. V. Ho (postdoctoral fellow) was provided by the Canadian Institutes of Health Research Postdoctoral Fellowship.
Data These data were collected as described in detail in [1]. The data include small angle X-ray scattering curves from two sets of a two-fold dilution series containing five sample dilutions. One series was the spermidine N-acetyltransferase SpeG in the absence of spermine, and the other was SpeG in the presence of spremine. The data were collected on two CCD based area detectors, which were then radially integrated to provide intensity versus momentum transfer vector scattering curves (I vs ). These were corrected for incident beam and transmission variation by division of the measured relative transmitted beam intensity on a diode in the beam stop. Next the data were scaled to absolute units by comparison to the small angle scattering of water in the same system.
Experimental design, materials and methods It is common practice to remove the contribution of a concentration dependent structure factor from the small angle scattering of a more concentrated sample by extrapolating data collected on a dilution series to infinite dilution [2,3]. These extrapolated data points are then used to replace the low q data (q<0.1Å−1) of the concentrated data set. This in effect treats the scattering curve or the concentrated sample as the product of an undetermined concentration dependent structure factor and the form factor of the particle in ideal solution. Alternatively the low q data can be taken from the more dilute samples, where there is no evidence of a structure factor, and combined with the higher q data from the more concentrated sample, where the signal to noise ratio is more favorable. Both methods yield a single form factor curve to be used in analysis with a signal to noise in the low q data equivalent to the most dilute sample. Another method for extrapolating to infinite dilution was used for the data from the SpeG solution in the presence of spermine; where a repulsive structure factor was evident as a endothelin receptor antagonist in the low q data, but otherwise the concentration corrected curves nearly superimposed (Fig. 1a). This method involved fitting a function representing the concentration dependent structure factor of the sample and then dividing it from each of the dilution scattering curves. By using this method a structure factor free set of data for each dilution of the SpeG in presence of spermine was calculated prior to determining the volume fractions of each oligomeric state [1]. Comparatively, for the dilution series of SpeG in the absence of spermine, a structure factor could not be calculated and hence the volume fractions of the two most concentrated samples were likely skewed to favor lower molecular weight oligomers. To determine a concentration dependent structure factor, each curve in the dilution series first had its relative concentration scaled to correct for small pipetting errors. This was done by taking the ratio each scattering curve in the dilution series (buffer subtracted but not divided by concentration) to the scattering curve of the most concentrated sample. A horizontal line was then fit to the higher angle data between 2/Rg and 3/Rg (about 0.06Å−1 and 0.08Å−1) to yield a correction factor for each concentration relative to the most concentrated (Fig. 1b). Then each of the concentration corrected curves were divided by the most dilute sample (Fig. 2a) to yield structure factor curves for each of the four dilutions above the most dilute. Initially R (hard sphere radius) and η (hard sphere volume fraction) were computed using a hard-sphere form factor function [4]whereandfit by nonlinear least-squares to the structure factor curve of the most concentrated sample between 0.007Å−1 and 0.06Å−1. This yielded an R of 42.4Å and an η of 1.18×10−2 with a reduced χ2 of 0.061 (Fig. 2a). Next η was treated as a function of the sample concentration: η(cn)=cnη0; and the ratio of SF(q,cn)/SF(q,c1) was fit for four structure factors (n=2 to 5, where c1 is the most dilute) simultaneously giving the same results with an R of 42.4Å and an η0 of 1.62×10−3 (c5=7.2mg/mL) with a reduced χ2 of 0.085. Finally R was fit as the product of an initial value R0 and the concentration with the small fractional exponent Rm: R(cη)=Ru(cη)Rm. This allows the hard sphere radius to vary according to concentration in a manner similar to angiosperms observed in other protein structure factor studies [5–7]. The final results had a reduced χ2 of 0.081 with R0=53.3Å, η0 of 1.59×10−3, and Rm=−0.13 (Fig. 2b). While these results are not incompatible with the possible intermolecular interactions in these samples, no physical interpretation can be assigned based on the obvious over-fitting of a simple hard-sphere model for these dilutions. Instead the fitted SF(q,c), can now applied as the divisor to the scattering curve from each dilution (c1 through c5) to produce redundant form factor data sets (Figs. 2c and 3)