Limits...
Deconvolution of Complex 1D NMR Spectra Using Objective Model Selection.

Hughes TS, Wilson HD, de Vera IM, Kojetin DJ - PLoS ONE (2015)

Bottom Line: Fluorine (19F) NMR has emerged as a useful tool for characterization of slow dynamics in 19F-labeled proteins.In current methods, determination of the deconvolution model best supported by the data is done manually through comparison of residual error values, which can be time consuming and requires model selection by the user.In contrast, the BIC method used by decond1d provides a quantitative method for model comparison that penalizes for model complexity helping to prevent over-fitting of the data and allows identification of the most parsimonious model.

View Article: PubMed Central - PubMed

Affiliation: Department of Molecular Therapeutics, The Scripps Research Institute, Scripps Florida, Jupiter, Florida, 33458, United States of America.

ABSTRACT
Fluorine (19F) NMR has emerged as a useful tool for characterization of slow dynamics in 19F-labeled proteins. One-dimensional (1D) 19F NMR spectra of proteins can be broad, irregular and complex, due to exchange of probe nuclei between distinct electrostatic environments; and therefore cannot be deconvoluted and analyzed in an objective way using currently available software. We have developed a Python-based deconvolution program, decon1d, which uses Bayesian information criteria (BIC) to objectively determine which model (number of peaks) would most likely produce the experimentally obtained data. The method also allows for fitting of intermediate exchange spectra, which is not supported by current software in the absence of a specific kinetic model. In current methods, determination of the deconvolution model best supported by the data is done manually through comparison of residual error values, which can be time consuming and requires model selection by the user. In contrast, the BIC method used by decond1d provides a quantitative method for model comparison that penalizes for model complexity helping to prevent over-fitting of the data and allows identification of the most parsimonious model. The decon1d program is freely available as a downloadable Python script at the project website (https://github.com/hughests/decon1d/).

No MeSH data available.


Related in: MedlinePlus

Fit of real experimental data.The ligand binding domain of PPARγ C285S/K474C was treated with BTFA and then NMR was performed at 298K. Deconvolution of the 19F NMR signal was carried out using the indicated programs. In each case the difference between the data and the fit (residual error) is shown in grey and the sum of individual fitted peaks is shown in green.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4524620&req=5

pone.0134474.g006: Fit of real experimental data.The ligand binding domain of PPARγ C285S/K474C was treated with BTFA and then NMR was performed at 298K. Deconvolution of the 19F NMR signal was carried out using the indicated programs. In each case the difference between the data and the fit (residual error) is shown in grey and the sum of individual fitted peaks is shown in green.

Mentions: We recently reported 2D and 3D heteronuclear NMR studies of the apo-, or ligand-free, ligand binding domain (LBD) of PPARγ—a ligand-responsive transcription factor and the target of FDA-approved insulin sensitizing drugs [15]. These studies revealed that an important region for protein-protein interaction, which includes the C-terminal “helix 12”, is dynamic on the intermediate exchange time scale [16, 17]. Thus peaks corresponding to residues in helix 12 were not visible, suggesting that this region of PPARγ samples at least two conformations in the absence of a ligand. To probe the conformations accessed by helix 12, we attached 3-bromo-1,1,1-trifluoroacetone (BTFA) to a PPARγ LBD double mutant (C285S, K474C), where a native cysteine was removed and a new cysteine was introduced in helix 12 (PPARγ-BTFA). A 19F NMR signal of PPARγ-BTFA is detectable, but very broad with a total width of ~170 Hz at the half maximum of the signal (FWHM). Fitting two separate NMR spectra (acquired at ~376 MHz) with two separate protein samples using decon1d yields a prediction of five or six peaks, with 90±8% of the total area attributable to two main signals; these peaks have FWHM values of 89±2 Hz (61±3% of area) and 65±4 Hz (29±5% of total area) and are separated by 0.175±0.035 ppm (66±13 Hz) with slightly opposing phase (Fig 6, S9 Fig). In two-site exchange signal coalescence is expected when kex/Δδ ~ 4.44 [18]. In this case with resolved peaks separated by 66±13 Hz, coalescence would be expected at a kex (= k-1+k1) of 290±58 s-1. Therefore these data suggest that PPARγ-BTFA helix 12 is switching between two long-lived conformations at an exchange rate slower than ~250 s-1 and that in addition minor populations of PPARγ exist with helix 12 in at least two other minor conformations It should be noted that the intrinsic linewidth of a non-exchanging fluorine label attached to PPARγ could be on the order of tens of hertz. This non-exchange linewidth adds to the uncertainty in chemical shift differences between the two peaks and thus the exchange rate. Two other minor conformations (total of 5–8% of the total signal fractional population) of 50±21 Hz and 11±1 Hz FWHM are predicted at very consistent chemical shifts (-84.40±0.01 ppm and 84.21±0.00 ppm). A wide small peak (Fig 6) was predicted for only one of the repeated experimental datasets. We also fit these spectra using various commercially available programs and found that decon1d produced the lowest residual value and was unique in consistently predicting a narrower linewidth for the right peak centered at -84.1ppm (Fig 6, S9 Fig).


Deconvolution of Complex 1D NMR Spectra Using Objective Model Selection.

Hughes TS, Wilson HD, de Vera IM, Kojetin DJ - PLoS ONE (2015)

Fit of real experimental data.The ligand binding domain of PPARγ C285S/K474C was treated with BTFA and then NMR was performed at 298K. Deconvolution of the 19F NMR signal was carried out using the indicated programs. In each case the difference between the data and the fit (residual error) is shown in grey and the sum of individual fitted peaks is shown in green.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4524620&req=5

pone.0134474.g006: Fit of real experimental data.The ligand binding domain of PPARγ C285S/K474C was treated with BTFA and then NMR was performed at 298K. Deconvolution of the 19F NMR signal was carried out using the indicated programs. In each case the difference between the data and the fit (residual error) is shown in grey and the sum of individual fitted peaks is shown in green.
Mentions: We recently reported 2D and 3D heteronuclear NMR studies of the apo-, or ligand-free, ligand binding domain (LBD) of PPARγ—a ligand-responsive transcription factor and the target of FDA-approved insulin sensitizing drugs [15]. These studies revealed that an important region for protein-protein interaction, which includes the C-terminal “helix 12”, is dynamic on the intermediate exchange time scale [16, 17]. Thus peaks corresponding to residues in helix 12 were not visible, suggesting that this region of PPARγ samples at least two conformations in the absence of a ligand. To probe the conformations accessed by helix 12, we attached 3-bromo-1,1,1-trifluoroacetone (BTFA) to a PPARγ LBD double mutant (C285S, K474C), where a native cysteine was removed and a new cysteine was introduced in helix 12 (PPARγ-BTFA). A 19F NMR signal of PPARγ-BTFA is detectable, but very broad with a total width of ~170 Hz at the half maximum of the signal (FWHM). Fitting two separate NMR spectra (acquired at ~376 MHz) with two separate protein samples using decon1d yields a prediction of five or six peaks, with 90±8% of the total area attributable to two main signals; these peaks have FWHM values of 89±2 Hz (61±3% of area) and 65±4 Hz (29±5% of total area) and are separated by 0.175±0.035 ppm (66±13 Hz) with slightly opposing phase (Fig 6, S9 Fig). In two-site exchange signal coalescence is expected when kex/Δδ ~ 4.44 [18]. In this case with resolved peaks separated by 66±13 Hz, coalescence would be expected at a kex (= k-1+k1) of 290±58 s-1. Therefore these data suggest that PPARγ-BTFA helix 12 is switching between two long-lived conformations at an exchange rate slower than ~250 s-1 and that in addition minor populations of PPARγ exist with helix 12 in at least two other minor conformations It should be noted that the intrinsic linewidth of a non-exchanging fluorine label attached to PPARγ could be on the order of tens of hertz. This non-exchange linewidth adds to the uncertainty in chemical shift differences between the two peaks and thus the exchange rate. Two other minor conformations (total of 5–8% of the total signal fractional population) of 50±21 Hz and 11±1 Hz FWHM are predicted at very consistent chemical shifts (-84.40±0.01 ppm and 84.21±0.00 ppm). A wide small peak (Fig 6) was predicted for only one of the repeated experimental datasets. We also fit these spectra using various commercially available programs and found that decon1d produced the lowest residual value and was unique in consistently predicting a narrower linewidth for the right peak centered at -84.1ppm (Fig 6, S9 Fig).

Bottom Line: Fluorine (19F) NMR has emerged as a useful tool for characterization of slow dynamics in 19F-labeled proteins.In current methods, determination of the deconvolution model best supported by the data is done manually through comparison of residual error values, which can be time consuming and requires model selection by the user.In contrast, the BIC method used by decond1d provides a quantitative method for model comparison that penalizes for model complexity helping to prevent over-fitting of the data and allows identification of the most parsimonious model.

View Article: PubMed Central - PubMed

Affiliation: Department of Molecular Therapeutics, The Scripps Research Institute, Scripps Florida, Jupiter, Florida, 33458, United States of America.

ABSTRACT
Fluorine (19F) NMR has emerged as a useful tool for characterization of slow dynamics in 19F-labeled proteins. One-dimensional (1D) 19F NMR spectra of proteins can be broad, irregular and complex, due to exchange of probe nuclei between distinct electrostatic environments; and therefore cannot be deconvoluted and analyzed in an objective way using currently available software. We have developed a Python-based deconvolution program, decon1d, which uses Bayesian information criteria (BIC) to objectively determine which model (number of peaks) would most likely produce the experimentally obtained data. The method also allows for fitting of intermediate exchange spectra, which is not supported by current software in the absence of a specific kinetic model. In current methods, determination of the deconvolution model best supported by the data is done manually through comparison of residual error values, which can be time consuming and requires model selection by the user. In contrast, the BIC method used by decond1d provides a quantitative method for model comparison that penalizes for model complexity helping to prevent over-fitting of the data and allows identification of the most parsimonious model. The decon1d program is freely available as a downloadable Python script at the project website (https://github.com/hughests/decon1d/).

No MeSH data available.


Related in: MedlinePlus