The Use of Mass Spectrometry and Molecular
Modeling to Design Structural Model of Mouse NKR-P1 Proteins
Daniel Rozbesky1, 2; Petr Man1,
2; Zdenek Kukacka1, 2; Zofie Sovova3, 4; Rudiger
Ettrich3, 4; Julien Marcoux5; Carol V. Robinson5;
Petr Novak1, 2
1Institute
of Microbiology, Prague, Czech Republic;
2Faculty of Sciences, Charles University, Prague, Czech Republic;
3Institute of Nanobiology and Structural
Biology, Nove Hrady, Czech
Republic;
4Faculty of Sciences, University of South Bohemia, Nove Hrady, Czech Republic;
5Department of Chemistry, University of Oxford, Oxford, United
Kingdom
Introduction
Determination of protein conformation has traditionally been realized by X-ray
crystallography and NMR spectroscopy. Although these techniques provide high
resolution atomic data, they have some limitations. Both NMR and X-ray require
large amounts of pure analyte and are time-consuming techniques. Mass
spectrometry combined with chemical cross-linking offers alternative approach
to identify the protein fold. This method is fast and uses small amounts of
material. Our aim was to gain insight into low-resolution structure of NKR-P1C
receptors. NKR-P1C is an activating immune receptor expressed on the surface of
mouse natural killer cells. Using distance constraints derived from chemical
cross-linking and disulfide arrangement in combination with computational
methods, protein conformation was designed. The validation of structural model
was addressed using ion-mobility mass spectrometry.
Methods
In order to design structural model of
NKR-P1C, protein was cross-linked using homobifunctional cross-linking reagents
disuccinimidyl suberate (DSS) and disuccinimidyl glutarate (DSG). After
cross-linking reaction, SDS-PAGE of cross-linking reaction mixture was performed.
Also, disulfide bound arrangement was determined after non-reducing
SDS-PAGE. The band of cross-linked or non-reduced protein was excised and
subjected to in gel digestion by Asp-N and trypsin. The peptide mixtures from
the enzymatic digest were analyzed by LC/ESI-FT-ICR MS. Cross-links and
disulfide-linked products were identified using Links software. These distance
constraints were used for molecular modeling. Homology modeling followed by a
short steepest descent minimization was performed using the MODELLER 9v7
package. To verify the fold of NKR-P1C, native mass spectrometry with ion
mobility measurements were performed.
Preliminary Data
Restraint-based computational modeling was
used to generate a model that represents the experimentally determined
constraints with a minimum of violations. Molecular dynamics was used to refine
the model and to describe the most populated protein conformers in solution. In
addition to the positional constraints obtained from the disulfide mapping and
from cross-linking experiments the model needs to preserve the overall C-type
lectin-like fold in these simulations, as the protein core is strongly
conserved, and the template and modeled structure share a sequence identity of
88%. However, as crystal structures are rigid contrary to protein dissolved in
solution we allowed the side chains in the core to be more flexible and adapt
to the given experimentally determined constraints. Specific attention was paid
to the extended loop region proposed to be involved in protein-ligand
interactions and ligand specificity. The only crystal structure published to
date for the entire NKR-P1 family, mouse NKR-P1A, shows this extended loop
pointing away from the protein core, in a conformation in which the loop would
be fully exposed to the solvent. Such a conformation could be clearly excluded
from the cross-links of the protein in solution. This was further supported by
IM-MS measurements corresponding to the compact form of the molecule based on
the experimentally derived collisional cross section. Therefore, in the most
populated conformation in solution, NKR-P1C most likely adopts the conformation
similar to the solution structure of NKR-P1A. Our model enables us to describe
this conformation on an atomic scale.
This work has been financially supported by the Grant Agency of the Czech Republic (GACR P207/10/1040), the Ministry of Education, Youth and Sports of the Czech Republic (Centre for Microbiology CZ.1.07/2.3.00/20.0055), and by the Institutional Research Project of the Institute of Microbiology (RVO61388971).