|
Tsinghua Science and Technology, Vol. 6, No. 3, August 2001 pp. 285-288 Thermal Transition of Ribonuclease A Observed Using Proton Nuclear Magnetic Resonance* YAN Yongbin 1,2 , LUO Xuechun 1,2 , ZHOU Haimeng 1 , ZHANG Riqing 1,2 1. Department of Biological Sciences and Biotechnology, Tsinghua University,
Beijing 100084, China; * Supported by the National Key Basic Research Special Funds of China (No.G1999075607), the National Key Science and Technology Program of China (No.96-900-09-03) and THSJZ of Tsinghua University, P.R.China. Received: 2000-11-01; revised: 2000-12-18 Code Number: ts01084 Abstract: The thermal transition of bovine pancreatic ribonuclease A (RNase A) was investigated using proton nuclear magnetic resonance (NMR). Significant resonance overlap in the large native protein limits accurate assignments in the 1 H NMR spectrum. This study proposes extending the investigation of large proteins by dynamic analysis. Comparison of the traditional method and the correlation coefficient method suggests successful application of spectrum image analysis in dynamic protein studies by NMR. Key words: correlation coefficient; nuclear magnetic resonance (NMR); ribonuclease A; thermal transition Introduction Bovine pancreatic ribonuclease A (RNase A), a single-domain protein that has four disulfide bonds, has played a crucial role in studies of protein structure, folding and enzyme catalysis. We propose extending these studies by image analysis through the use of nuclear magnetic resonance (NMR) spectroscopy, which has the advantage of direct, continuous, dynamic, nondestructive, and quantitative monitoring. Absorbance measurements at 286 nm/287 nm (UV difference spectroscopy) have been traditionally used to get the thermal transition curves and the thermodynamic parameters[1] Tm, DHo(Tm), and DSo(Tm). Several new methods have been introduced to monitor the structural changes of the protein caused by thermal denaturation. Fourier transform infrared (FTIR) difference spectroscopy[2] , NMR hydrogen/deuterium exchange rate, fluorescence, circular dichroism (CD), optical rotatory dispersion (ORD)[1] and 2-D NMR spectroscopy[3] have also been used to investigate the kinetics of the active site and the secondary structure changes during thermal unfolding/refolding. The results show that RNase A unfolds reversibly along a two-state mechanism with different melting temperatures Tm in various redox systems. Though NMR can provide extensive conformation and structural information, few dynamic NMR measurements had been reported for native large proteins because of the limits imposed by the significant overlap of resonance peaks in the 1 H NMR spectra. The primary goal of this study is to investigate a new method for dynamic study of the thermal transition properties of RNase A. Later reports will provide more results about the unfolding/refolding progress with denaturants and other conditions. 1 Materials and Methods 1.1 Sample preparation Highly purified lyophilized ribonuclease A (type XII-A) from bovine pancreas (RNase A) was purchased from Sigma Chemical Co. and used without further purification. NMR samples of native RNase A were dissolved in 0.5 mL 100 mmol/L phosphate sodium solvent (PBS; containing 10% D2O to provide a signal for the lock), pH 8.0 (no corrections for isotope effects) at 1.0-1.1 mmol/L (14 mg/mL). 1.2 NMR spectroscopy All 1 H NMR experiments were carried out on a Bruker AM500 superconductor spectrometer at Tsinghua University. The 90° pulse width was 6.5 ms, the sweep width was 8333 Hz and each FID had 16 kbit data points. Solvent suppression was carried out by presaturation at all times except during acquisition and the frequency of presaturation was adjusted to the H2O/HOD resonance frequency at the current temperature. The thermal transition measurements were carried out by increasing the temperature in one degree increments from 40 °C to 60 °C . Approximately 5 min was allowed for thermal equilibration at each measured temperature. 1.3 Data analysis Fourier transform was proceeded after getting the FID signal. The phase correction parameters were the same as those for the first spectrum with a small adjustment due to the temperature changes. The data sets were transferred to an SGI workstation using UXNMR software. 1.3.1 General analysisSTWTHT The assignments were based on the well-resolved resonance of CH3 of Val63 in the 1 H NMR spectrum. Assignments were done by referring to the works of Robertson and his coworkers[4] and were confirmed by 2-D NOE (NOESY) spectrum. 1.3.2 Image analysis The 1-D acquisition and processed data were stored in the files consisting of a sequence of 32 bit integers in binary format in the SGI workstation. For image analysis, these files were converted and stored as text files based on the standard JCAMP-DX data file format (FIX form)[5] . This plain text file format was then converted to MATLAB file format for further analysis. The correlation coefficient describing the correlation between different images was calculated basing MATLAB programs developed in house. The correlation coefficient was the traditional coefficient that describes the correlation between two systems. Here we use the definition and program in MATLAB for the correlation coefficient C(X,Y) between X and Y given by where SXY is the covariance between X and Y, SXX and SYY are the variances of X and Y. The correlation coefficient is zero if X and Y are statistically independent and is 1 if X and Y are the most significant correlated. Two regions of each spectrum were selected instead of the whole spectrum to remove the great influence of the solvent resonance of H2O/HDO. The upfield region was defined as 0-4.6x10-6 , which contains the main resonances of all the aliphatic CH indicating the backbone structure. The downfield region was defined as 6.3x10-6 -10.0x10-6 , which contains the main resonances of all the amide protons both on the backbone and on the side chains, with the ring protons of the aromatic and the indole of the amino acid residues. 2 Results and Discussion 2.1 Thermal transition monitored by 1 H NMR spectrum Figure 1 shows the stack plot of a series of 1 H NMR spectra of RNase A between 0 and 4.6x10-6 (Fig.1(a)) and between 6.3 and 10x10-6 (Fig.1(b)) during irreversible thermal denaturation experiments. Above 50 °C most of the structured NH resonances become smaller until they disappear. Then the resolution becomes worse and the integral area of the ring region becomes smaller. This indicates that the RNase A in the PBS system begins to aggregate. The aliphatic region gives a similar result that the structured numerous peaks aggregate in several unstructured regions. The thermal transition dynamics can be analyzed by observing specific resonance split phenomena, chemical shift changes and the integral areas. The resonance at 0.4x10-6 was resolved as CH3 of Val63 and its thermodynamic curve is shown in Fig.2. The thermal transition is a two-state process. The calculated Tm and the van't Hoff's enthalpy change for denaturation at Tm are listed in Table 1. DHm values are quite consistent with those obtained in previous studies[6, 7]. 2.2 Thermal transition characterized by correlation coefficient The image parameter, the correlation coefficient, calculated to characterize the 1 H NMR spectra of the thermal unfolding process of RNase A, are shown in Fig.3. The thermal transition curves characterized by the correlation coefficient are a flat between 40 °C and 44 °C , which suggests thermal stability. The significant increase of the correlation coefficient above 50 °C in the downfield region was due to protein aggregation and bad resolution. The conformational change of RNase A characterized by the correlation coefficient was quite similar to that characterized by CH3 of Val63 from 44 °C to 50 °C . The analysis also yielded a similar Tm (47.5±0.2)°C and Hm to those listed in Table 1.
In common, the proton resonances in a spectrum must be assigned for dynamic protein investigations based on NMR. However, significant overlap in the spectra of large native proteins limits the application of NMR. This study proposes extending the investigations of large proteins by dynamic analysis. Comparison of the traditional method and the correlation coefficient method suggests successful application of spectrum image analysis. Acknowledgements The authors thank Mr. Jiang Bo and Mr. Li Yuheng for their expert technical assistance. References
Copyright 2001 - Tsinghua Science and Technology The following images related to this document are available:Photo images[ts01084f3.jpg] [ts01084f2.jpg] [ts01084t1.jpg] [ts01084f1.jpg] |
|