On the use of hydrogen/deuterium exchange mass spectrometry data to improve de novo protein structure prediction.
On the use of hydrogen/deuterium exchange mass spectrometry data to improve de novo protein structure prediction.
复制标题
关于使用氢/氘交换质谱数据来改进从头蛋白质结构预测。
DOI：
10.1002/rcm.3882
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
Goodlett,DavidR
中科院分区：
文献类型：
--
作者：
Malmstroem,Lars;Hou,Liming;Atkins,WilliamM;Goodlett,DavidR
De novo structure prediction (PSP) has improved in recent years to the point where models may be generated at atomic resolution1, 2 and even, given time constraints, for full proteomes, such as S. cerevisiae, 3 at albeit lower resolution. Such structure models can serve as a basis for hypothetical generation for biologists to predict molecular function. 4 These PSP methods then can play an important role in systems biology, 5–9 a reductionist approach used to propose hypotheses unbiased by prior findings and to build predictive models of cellular behavior. Currently, most PSP technologies produce many false positives which make sorting the most likely correct protein structure models from incorrect predictions difficult. 2 Mass spectrometry (MS) has seen a widespread adoption across biological fields in the last decade enabling researchers to both identify proteins and covalent alterations (eg posttranslational modifications) at amino acid residue or atomic resolution. Given the speed and sensitivity with which MS data may be obtained we asked whether we might use MS-obtained physical constraints on a purified protein to better sort the many correct from incorrect in silico predicted protein structures. Hydrogen/deuterium exchange mass spectrometry (DXMS) denotes a technology in which the hydrogens of protein amide bonds along the backbone, but not the imide bond of proline, are exchanged for deuterium in a tertiary structure dependent manner during limited solvation of the protein in D2O. 10, 11 Because DXMS data should be global in nature reflecting the overall structure of a protein in solution, we chose to test whether DXMS analysis on a protein of known structure could be used to sort correct from incorrect in silico predicted structures. In this study, we investigated the extent to which DXMS data obtained on purified glutathione-S-transferases (GST) could be used to improve the fidelity of de novo protein structure prediction.To this end we expressed two close homologs of known tertiary structure, GST-A1-1 and GST-A4-4. The GST-A1-1 and GST-A4-4 proteins were expressed in Escherichia coli using the BL21 strain and purified via GSH-agarose and size-exclusion chromatography as previously described. 12, 13 DXMS experiments were performed on the free native enzymes by wellestablished experimental and data analysis methods. 14–17 The deuterium exchange was initiated by a 20-fold dilution of the enzyme with 10mM Tris-DCl/D2O buffer (pH 7) at 58C. The protein in D2O was incubated at 58C for 5 s, immediately followed by quenching with an equal volume of ice-cold 50 mM potassium phosphate/sodium citrate buffer (pH 2.2). The quenched protein solution was of pH 2.5. For the back-exchange control, the proteins were 20-fold diluted in 2 M per-deuterated urea solution in D2O, incubated for 8h at 408C and quenched in the same way. 18 The quenched mixture was allowed to equilibrate on ice for 30s, after which 1 mL of ice-cold pepsin solution (3.660 units/mg solid, 24 mg/mL in H2O) was added to 20 mL of the protein solution where the final concentration of protein was 0.5 mg/mL. The digestion proceeded for 2min on ice, and the resulting peptides were analyzed immediately by high-performance liquid chromatography (HPLC)/MS where the HPLC column and the entire injector assembly were packed in ice to minimize backexchange during analysis. Electrospray mass spectra were recorded on a quadrupole/time-offlight mass spectrometer (Micromass,
