Dynamic allostery in substrate binding by human thymidylate synthase

The parameters for the slow (A↔B) process are fixed based on the values obtained in a global 2-state fit of the CEST data alone. This script requires a fitting function (Source code 5), which is where the calculation of the residuals is carried out.

The parameters for the slow (A↔B) process are fixed based on the values obtained in a global 2-state fit of the CEST data alone. This script requires a fitting function (Source code 6).

For one of the methyl groups, a B↔A↔C model is used, while for the other a 2-state model is used. The faster process (A↔C) parameters are shared between the two methyl groups, while the slow process (A↔B) parameters are fixed on the values obtained in a global 2-state fit of the CEST data alone. This script requires a fitting function (Source code 7).

This script requires a fitting function (Source code 9 and 10).

The parameters obtained by this fit are used to fix the slow process parameters in the 3-state fits of the methyl groups involved in the global motion. This script requires a fitting function (Source code 13).

This script requires a fitting function (Source code 15).

(a) values from global fit with $p_b\cong 1.3\%$, $k_{ex,ab}=240s^{-1}$

(b) population and exchange rate shown for slower of two processes. Unlike the B↔A↔C model, which gives very similar $p_b$ and $k_{ex,ab}$ for individual fits of these probes, the A↔B↔C model yields heterogeneous values, primarily for the exchange rate (c) ${\chi }_{red}^2$ is calculated as $\frac{\sum _i{\left(\frac{residua{l}_i}{erro{r}_i}\right)}^2}{numdatapoints-numfitparam}$

Probes involved in the concerted process are given the purple background. If a probe is fit to a 2-state model ($A\leftrightarrow C$), the parameters for the $A\leftrightarrow B$ process are left blank. Cases where a ¹H $∆\omega$ is fixed at 0 are indicated by the ‘-‘. Parameter values are reported as median ±standard deviation of fit values from 200 Monte Carlo simulations of the data. For probes involved in the concerted process, the error in $p_b$ is the sum of the errors from the 2-state fit of the CEST data alone and from the global fit of the CPMG and CEST data with the slow process parameters fixed. Only the ¹³C $∆\omega$’s for the slow, concerted process have sign information from the CEST data; all other $∆\omega$’s should be interpreted only as the magnitude of the chemical shift difference. The background of each parameter value is color-coded based on the magnitude of the error relative to the parameter value, where blue indicates low error, yellow indicates medium error, and red indicates large error. In some cases, particularly for the populations, physical constraints on the parameter value (i.e. population cannot be less than 0) lead to highly skewed distributions of the parameter values in our Monte Carlo simulations. This can lead to nonsensical values when reported in our typical manner, for example the $0\pm 6\%p_b$ given for L279 despite the fact that the population cannot be less than 0. In these cases, we have also listed (5% quantile, mode, 95% quantile) to provide greater insight into the distribution of values seen in the Monte Carlo simulations. For L269, marked with the asterisk, the CEST data was not included in the fit. Probes possessing $R_{ex}$ which weren’t analyzed include L67, V164, and V313. Refer to the legend of Supplementary file 7 for a description of the ‘met 1’ and ‘met 2’ labels.

All probes are fit to a 2-state model. Color-coding and values in parenthesis are as described in Supplementary file 2. L192 met 2, marked by the asterisk, is presumably the same methyl group analyzed in apo hTS given the similarity in the $∆\omega$’s, but assignment of this signal is complicated by a large chemical shift perturbation between apo and dUMP bound states.

All probes are fit to a 2-state model. Color-coding and values in parenthesis are as described in Supplementary file 2.

Rotational correlation times for various hTS-bound states determined by ¹⁵N relaxation (Source data 1–19).

For LV methyls, the labels ‘met1’ and ‘met2’ are given, where ‘met1’ has the larger ¹³C chemical shift, as stereospecific assignments have not been made. For all other states, the ‘met1’ and ‘met2’ labels are given based on chemical shift similarity to the apo state. The only exception is L192, where $∆\omega$’s obtained from our dispersion fits guided the assignment.