J Biomol NMR. 2009 Feb;43(2):63-78. Epub 2008 Nov 26.

De novo protein structure generation from incomplete chemical shift assignments.

Source

Laboratory of Chemical Physics, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 20892-0520, USA.

Abstract

NMR chemical shifts provide important local structural information for proteins. Consistent structure generation from NMR chemical shift data has recently become feasible for proteins with sizes of up to 130 residues, and such structures are of a quality comparable to those obtained with the standard NMR protocol. This study investigates the influence of the completeness of chemical shift assignments on structures generated from chemical shifts. The Chemical-Shift-Rosetta (CS-Rosetta) protocol was used for de novo protein structure generation with various degrees of completeness of the chemical shift assignment, simulated by omission of entries in the experimental chemical shift data previously used for the initial demonstration of the CS-Rosetta approach. In addition, a new CS-Rosetta protocol is described that improves robustness of the method for proteins with missing or erroneous NMR chemical shift input data. This strategy, which uses traditional Rosetta for pre-filtering of the fragment selection process, is demonstrated for two paramagnetic proteins and also for two proteins with solid-state NMR chemical shift assignments.

PMID:: 19034676; [PubMed - indexed for MEDLINE]
PMCID: PMC2683404

Free PMC Article

Images from this publication.See all images (7) Free text

Figure 1

Flow chart of CS-Rosetta structure generation protocol. In the hybrid fragment selection procedure, shown in red, step 1 selects 200 fragments from an initial cohort of 2000 fragments which has been extracted from the structural database by standard Rosetta methods. In the standard CS-Rosetta method, step 1 takes its fragments directly from the 2,200,000 fragments present in the structural database.

De novo protein structure generation from incomplete chemical shift assignments

J Biomol NMR. J Biomol NMR;43(2):63-78.

Figure 7

Comparison of experimental (blue) and lowest energy CS-Rosetta (red) structure for paramagnetic calbindin (A) and ferredoxin (B). Superposition is optimized for residues in secondary structure, defined in the caption to Fig. 6. The sidechains of residues involved in metal binding including their metal-ligating oxygen atoms, as well as the X-ray positions of the Ca²⁺ ions (cyan) are shown. Metal-ligating residues (atoms) include Ala¹⁴ (O), Glu¹⁷ (O), Asp¹⁹ (O), Gln²² (O), Glu²⁷ (O^ε1/O^ε2), Asp⁵⁴ (O^δ1), Asn⁵⁶ (O), Asp⁵⁸ (O^δ1), Glu⁶⁰ (O) and Glu⁶⁵ (O^ε1/O^ε2). (B) Backbone ribbon representation of the lowest-energy CS-Rosetta structure (red) superimposed on the experimental X-ray structure (blue) for ferredoxin, with superposition optimized for the residues in secondary structure (see caption to Fig. 6). The sidechain S atoms of Cys⁴², Cys⁴⁷, Cys⁵⁰ and Cys⁸², which coordinate the [2Fe-2S] cluster are marked as solid spheres. Figures made using Molmol (Koradi et al. 1996).

De novo protein structure generation from incomplete chemical shift assignments

J Biomol NMR. J Biomol NMR;43(2):63-78.

Figure 2

CS-Rosetta structure generation for TM1442 with missing chemical shift assignments for certain types of nuclei. (A, B) Plots of accuracy of fragments selected using the MFR (blue) and hybrid (red) methods with the chemical shift inputs for δ¹⁵N, δ¹H^N and δ¹³C^α (as contained in the dataset Ih). Quality of three-residue (A) and nine-residue (B) fragments is represented by the average (bold lines) and lowest (lines with dots) rmsd of 200 selected fragments relative to the experimental coordinates of the corresponding TM1442 segment. (C and D) plots of Rosetta all-atom energy, rescored by using the input chemical shifts (as contained in the dataset Ih), versus C^α rmsd relative to the experimental TM1442 structure, for CS-Rosetta models obtained using MFR (C) and hybrid (D) fragment selection methods. (E-H) CS-Rosetta fragment selections and structure generations for TM1442 using only δ¹³C^α and δ¹³C^β (as contained in the dataset Ii).

De novo protein structure generation from incomplete chemical shift assignments

J Biomol NMR. J Biomol NMR;43(2):63-78.

Figure 5

CS-Rosetta fragment selection and structure generation for GB3 (A-C) and ubiquitin (D-F), using chemical shift assignments from solid-state NMR. (A,D) Plots of the lowest (upper panel) and average (lower panel) backbone coordinate rmsds (N, C^α and C’) between query segment and two hundred 3-residue fragments, selected using the MFR (blue) and hybrid methods (red), as a function of starting position in the sequence. (B,E) same as (A,D) but for 9-residue fragments. (C,F) Plots of Rosetta all atom energy, rescored by using the experimental ssNMR chemical shifts, versus C^α rmsd relative to the experimental NMR structures of GB3 and ubiquitin for the CS-Rosetta all-atom models obtain using MFR-selected (upper panel, blue dots) and the hybrid method (lower panel, red dots) fragments. The solid black lines in (C,F) represent the normalized number of structures found at a given C^α-rmsd.

De novo protein structure generation from incomplete chemical shift assignments

J Biomol NMR. J Biomol NMR;43(2):63-78.

Figure 4

CS-Rosetta structure generation of TM1442 with chemical shift errors. (A,B) Plots of accuracy of fragments selected using the MFR (blue) and hybrid (red) methods, with the inputs swapped for the δ¹⁵N, δ¹H^N, δ¹³C^α, δ¹³C^β, δ¹³C^' and δ¹H^α assignments of dipeptides Ser⁵²-Ser⁵³ and Ser⁸²-Ser⁸³ (as contained in the dataset IIIb). For each three-residue (A) and nine-residue (B) segment of TM1442, 200 fragments were selected. Average (bold lines) and lowest (lines with dots) rmsd of these fragments relative to the experimental coordinates of the corresponding TM1442 segment are plotted with respect to the position of the first segment residue in the TM1442 sequence. The regions corresponding to the “miss-assigned” residues are shaded; secondary structure elements are displayed at the top. (C, D) Plots of Rosetta all atom energy, rescored by using the input chemical shifts (as contained in the dataset IIIb), versus C^α rmsd relative to the experimental TM1442 structure, for CS-Rosetta models obtained using MFR (C) and hybrid (D) fragment selection methods.

De novo protein structure generation from incomplete chemical shift assignments

J Biomol NMR. J Biomol NMR;43(2):63-78.

Figure 3

CS-Rosetta structure for TM1442 with missing chemical shifts. (A and B) Plots of accuracy of fragment candidates selected using the MFR (blue) and hybrid (red) methods using chemical shift values δ¹⁵N, δ¹H^N, δ¹³C^α, δ¹³C^β, δ¹³C' and δ¹H^α for residues 1–20, 30–51 and 60–120 (as contained in the dataset IIe). For each three-residue (A) and nine-residue (B) segment of TM1442, 200 fragments were selected. Average (bold lines) and lowest (lines with dots) rmsd of these fragments relative to the experimental coordinates of the corresponding TM1442 segment are plotted with respect to the position of the first segment residue in the TM1442 sequence. The regions corresponding to the “unassigned” residues are shaded; the secondary structure elements are displayed at the top. (C, D) Plots of Rosetta all atom energy, rescored by using the input chemical shifts (as contained in dataset IIe), versus C^α rmsd relative to the experimental TM1442 structure, for CS-Rosetta models obtained using MFR (C) and hybrid (D) fragment selection methods.

De novo protein structure generation from incomplete chemical shift assignments

J Biomol NMR. J Biomol NMR;43(2):63-78.

Figure 6

CS-Rosetta structure generation for paramagnetic calbindin (A-C) and ferredoxin (D-F). (A,D) Plots of the lowest (upper panel) and average (lower panel) backbone coordinate rmsds (N, C^α and C’) between query segment and two hundred 3-residue fragment candidates, selected using the MFR (blue) and hybrid methods (red), as a function of starting position in the sequence. The regions lacking chemical shift assignments are shaded. (B,E) Same as (A,D), but for 9-residue fragments. (C,F) Plots of Rosetta all-atom energy, rescored by the experimental chemical shifts, versus C^α rmsd of final al-atom models (including only residues located in elements of secondary structure) relative to the corresponding X-ray (calbindin) and NMR (ferredoxin) structure. Only results from CS-Rosetta all-atom models obtained by the hybrid fragment selection procedure are shown; when using fragments from the standard MFR method, Rosetta fails to converge. Residues included in the backbone rmsd calculation include 3–14, 25–40, 46–53 and 63–74 for calbindin, and 4–11, 15–22, 27–34, 54–56, 71–75 and 91–93 for ferredoxin.

De novo protein structure generation from incomplete chemical shift assignments

J Biomol NMR. J Biomol NMR;43(2):63-78.

Publication Types, MeSH Terms, Substances, Grant Support

Publication Types

Research Support, N.I.H., Intramural
Research Support, Non-U.S. Gov't

MeSH Terms

Computer Simulation
Crystallography, X-Ray
Isotopes/chemistry
Models, Molecular
Monte Carlo Method
Nuclear Magnetic Resonance, Biomolecular/methods*
Peptide Fragments/chemistry
Proteins/chemistry*
Software*

Substances

Isotopes
Peptide Fragments
Proteins

Grant Support

Z01 DK029046-01/DK/NIDDK NIH HHS/United States
Howard Hughes Medical Institute/United States

LinkOut - more resources

Full Text Sources

Springer
EBSCO
OhioLINK Electronic Journal Center
Ovid Technologies, Inc.
ProQuest Information and Learning
PubMed Central - Author Manuscript
PubMed Central Canada - Author Manuscript
Swets Information Services
UK PubMed Central - Author Manuscript

Other Literature Sources

COS Scholar Universe

PubMed

Display Settings:

Send to:

De novo protein structure generation from incomplete chemical shift assignments.

Source

Abstract

Publication Types, MeSH Terms, Substances, Grant Support

Publication Types

MeSH Terms

Substances

Grant Support

LinkOut - more resources

Full Text Sources

Other Literature Sources

Supplemental Content

Related citations

Cited by 20 PubMed Central articles

Related information

Recent activity

Simple NCBI Directory

Getting Started

Resources

Popular

Featured

NCBI Information