Six- and seven-dimensional experiments by combination of sparse random sampling and projection spectroscopy dedicated for backbone resonance assignment of intrinsically disordered proteins
Two novel six- and seven-dimensional NMR experiments are proposed. The new experiments employ non-uniform sampling that enables achieving high resolution in four indirectly detected dimensions and synchronous sampling in the additional dimensions using projection spectroscopy principle. The resulted data sets could be processed as five-dimensional data using existing software. The experiments facilitate resonance assignment of intrinsically disordered proteins. The novel experiments were successfully tested using 1 mM sample of α-synuclein on 600 and 800 MHz NMR spectrometers equipped with standard room temperature probes. The experiments allowed backbone assignment from a 1-day acquisition.
KeywordsIntrinsically disordered proteins Resonance assignment High-dimensionality NMR Non-uniform sampling Projection spectroscopy
In recent years, intrinsically disordered proteins (IDPs) have attracted widespread interest in molecular biology research (Wright and Dyson 1999; Habchi et al. 2014). Their properties like structural propensity, dynamics and interactions could be effectively studied by solution NMR spectroscopy methods. However, the intrinsic disorder results in a fast conformational dynamics which causes an effective averaging of chemical shifts. Thus, very poor peak separation makes resonance assignment difficult, even for a relatively small disordered protein fragments. As a consequence, the signal overlap in the spectra strongly limits the possibility of standard NMR experiments to provide sufficient resolution for IDPs characterization. On the other hand, the fast local dynamics significantly decreases transverse relaxation rates, enabling the application of long multidimensional pulse sequences, and making possible to achieve long evolution times in order to increase the resolution and facilitate assignment of resonances.
The projection NMR spectroscopy methods are founded on the idea of Accordion Spectroscopy (Bodenhausen and Ernst 1981), and employ synchronous incrementing of two or more evolution periods in the pulse sequence. The particularly important approach of this type is simultaneous sampling of more than one chemical shift evolution, which is referred to as radial sampling. Such an option is utilized in projection spectroscopy (Kupče and Freeman 2005; Coggins et al. 2004, 2010), and requires the algebraic decoding of peak frequencies (Kim and Szyperski 2003; Koźmiński and Zhukov 2003; Hiller et al. 2005), or the reconstruction of multidimensional spectrum (Kupče and Freeman 2003).
To date, several high dimensional strategies have been proposed for backbone resonance assignment of IDPs (Reddy and Hosur 2014; Goradia et al. 2015; Yoshimura et al. 2015), including direct 13C detection (Bermel et al. 2006, 2009; Nováček et al. 2011; Pantoja-Uceda and Santoro 2013, 2014) and HA detection (Mäntylahti et al. 2010, 2011; Yao et al. 2014) which allow for revealing of proline residues. The number of dimensions, i.e. number of simultaneously correlated frequencies, could be increased by employing automated projection spectroscopy (APSY) (Hiller et al. 2007; Narayanan et al. 2010; Yao et al. 2014), or the non-uniformly sampled (NUS) high-dimensional experiments (Kazimierczuk et al. 2010; Zawadzka-Kazimierczuk et al. 2012).
APSY approach enabled the acquisition of experiments with dimensionality up to 7, by recording a set of 2D projections of high dimensionality spectrum at different angles. The high-dimensional peak list is achieved by algebraic reconstruction from automatically collected peak coordinates in the projections. The possible disadvantage of this technique is severe peak overlap on 2D spectral projections, which may influence the precision of frequency determination.
The advantage of non-uniformly sampled high-dimensionality experiments, followed by SMFT (sparse multidimensional Fourier transform) processing (Kazimierczuk et al. 2009), is the ability for visual inspection of the spectrum of the full dimensionality. Such examination of 2D cross-sections helps to assign peaks suffering from partial overlap.
SMFT processing procedure with the output in the form of a series of 2D cross-sections, requires, however, a so-called base spectrum of dimensionality of N-2, which becomes impractical for N > 5. Therefore, we propose here new 6 and 7D experiments that feature high resolution and high dimensionality resulting from the use of non-uniform sampling in the indirectly detected dimensions combined with projection spectroscopy principle. The experiments utilize NUS in four indirectly sampled dimensions, while additional frequencies are added by synchronous incrementing with t 1. Thus, the previously described 5D SMFT processing with a 3D base spectrum could be used for the proposed experiments, resulting in a 5D spectrum with additional dimensions projected into F 1.
Since the N and CO chemical shifts were found to be best dispersed in IDPs (Piai et al. 2014; Nowakowski et al. 2015), the proposed experiments are derived from 5D (H)NCO(NCA)CONH (Zawadzka-Kazimierczuk et al. 2012), by adding HN (6D) or HN and CA (7D) chemical shifts. The experimental verification was performed on a 1 mM sample of α-synuclein, employing 600 and 800 MHz NMR spectrometers with standard RT triple resonance probes.
Materials and methods
Quadrature for t 1 dimensions is achieved using a standard States-TPPI procedure. Quadratures in coevolved dimensions require recording and storing of two data sets, cosine and sine modulated, respectively. It is obtained by 90 degree shift of phases ϕ5, and ϕ5, ϕ6 (for 6D and 7D versions respectively) (Koźmiński and Zhukov 2003; Kim and Szyperski 2003). To allow standard SMFT data processing the π/2 phase shifts for sine modulated components are compensated by simultaneous π/2 shifting of ϕ1 phase, for t 1 dimension. Therefore, such procedure of processing of spectra after SMFT approach is limited to a simple co-addition to obtain spectra differing in frequency signs in coevolved dimensions.
Data processing and inspection
All 5D spectra were processed using ToASTD (Kazimierczuk et al. 2006) and reduced (Kazimierczuk et al. 2009) programs which take peak list from 3D HNCO as the input for SMFT routine. For 3D HNCO sampling artefacts were removed from the spectra using Signal Separation Algorithm (Stanek and Koźmiński 2010), no such procedure was performed for 5D spectra. All used programs are available at http://nmr.cent3.uw.edu.pl/software. After processing, all spectra were inspected using Sparky (Goddard and Kneller 2000).
NMR sample contained 1 mM of 13C, 15N-labeled α-synuclein in 20 mM sodium phosphate buffer, pH 6.5, 200 mM NaCl. NMR spectra were recorded at 288 K on Agilent 800 MHz and 600 MHz spectrometers both equipped with room temperature probes. 7D HNCO(N)CACONH was acquired on the 800 MHz spectrometer in 31 h with evolution times set to 20, 30, 25, 52, 25, 30 ms for H i N , Ni, COi−1, CAi−1, COi, Ni+1 dimensions, respectively. 6D HNCO(NCA)CONH was acquired on both, 800 MHz and 600 MHz, spectrometers in 13 and 23 h, respectively. Evolution times for the measurement on the 800 MHz spectrometer were set to: 20, 30, 25, 25, 30 ms, for H i N , Ni, COi−1, COi, Ni+1 dimensions, respectively, and to sustain high enough resolution on the lower field evolution times were set to: 20, 40, 40, 40, 40 ms, respectively on the 600 MHz spectrometer.
In order to verify results obtained from 6D and 7D projection experiments full backbone assignment of the α-synuclein was performed using 5D HN(CA)CONH (Kazimierczuk et al. 2010), 5D (H)NCO(NCA)CONH (Zawadzka-Kazimierczuk et al. 2012) and 5D HabCabCONH (Kazimierczuk et al. 2010) experiments. To avoid any discrepancies, exactly the same sample as for 6D and 7D measurements was used.
Analysis of both recorded 6D spectra led to identification of all but two expected sequential peaks. Two missing residues were the first two residues in the protein sequence which suggest that those signals were broadened beyond detection limit due to amide proton exchange. Similar result was obtained from 7D spectrum, the same set of sequential signals was identified.
The most important advantage of presented approach is the establishment of highly robust connectivities due to the expansion of sequential peaks frequencies by addition of H i−1 N frequency. Similar approach was already proposed in the literature which involved the combination of two different 5D experiments (Kazimierczuk et al. 2013; Piai et al. 2014) having shared sequential nitrogen dimension. Here it was presented that such “HNCO-HNCO” strategy can be successfully implemented within one experiment. Additional expansion to 7D experiment giving CA chemical shift can further limit number of overlapping signals as well as provide partial information needed for identifying amino acid type.
True benefit of proposed approach is the possibility of a visual inspection of resulting spectra which allows to use the expertise and the experience of a spectroscopist in resolving most difficult, severely overlapped cases. As a result, all of non-proline residues of α-synuclein were successfully identified and assigned from a single experiment (except aforementioned first two residues).
Surprisingly, even example of a moderate size α-synuclein already justifies expanding “CON-CON” strategy. Presented approach can be especially beneficial in performing the resonance assignment of proteins of a larger size than α-synuclein, since the spectra in such a case will be more crowded. What is more, sensitivity of presented experiments is high enough to record good quality spectra in a relatively short time compared to other six- and seven-dimensional approaches (Hiller et al. 2007; Yao et al. 2014), even without the use of cryogenically cooled probe.
We have demonstrated the principles of the new experiment and data inspection protocol. Although, it could be modified in several ways. The proposed experiment is fully compatible with BEST approach (Lescop et al. 2007; Solyom et al. 2013) and co-solute paramagnetic relaxation enhancement (Oktaviani et al. 2015), which could be used for acceleration of the signal repetition rate, and thus increase the number of data points acquired in a given time. Amide hydrogen exchange process, particularly effective in the IDP molecules, may cause a severe signal loss at conditions close to physiological (temperature and pH). Therefore, the pulse sequence could be modified using the aliphatic proton excitation combined with the 13C or HA detection (Mäntylahti et al. 2010). Moreover, aliphatic protons or 13C detection would allow observation of additional cross-peaks from proline residues (Bermel et al. 2006; Hellman et al. 2014). Additionally, the amide protons exchange causes further decay of 15N–1H antiphase coherences. Hence, the refocusing of 15N–1H couplings and 1H composite pulse decoupling applied during the main part of pulse sequence is used to maximize the sensitivity.
We presented six- and seven-dimensional experiments for backbone assignment of intrinsically disordered proteins. Described techniques combine projection spectroscopy and SMFT data processing that allow access to full dimensional spectra information. 6D experiment correlates H i N , Ni, COi−1, H i−1 N , Ni−1, COi−2. 7D adds additional correlation with CAi−1 dimension. We proved that these experiments can be successfully run on a standard room-temperature probe, even on a medium field NMR spectrometer (600 MHz).
This work was supported by Polish National Science Centre with PRELUDIUM grant No. 2013/11/N/ST4/01832. The study was carried out at the Biological and Chemical Research Centre, University of Warsaw, established within the project co-financed by European Union from the European Regional Development Fund under the Operational Programme Innovative Economy, 2007–2013.
- Goddard TD, Kneller DG (2000) SPARKY 3. University of California, San FranciscoGoogle Scholar
- Hellman M, Piirainen H, Jaakola V, Permi P (2014) Bridge over troubled proline: assignment of intrinsically disordered proteins using (HCA)CON(CAN)H and (HCA)N(CA)CO(N)H experiments concomitantly with HNCO and i(HCA)CO(CA)NH. J Biomol NMR 58:49–60. doi: 10.1007/s10858-013-9804-0 CrossRefGoogle Scholar
- Nováček J, Zawadzka-Kazimierczuk A, Papoušková V, Žídek L, Šanderová H, Krásný L, Koźmiński W, Sklenář V (2011) 15D 13C-detected experiments for backbone assignment of unstructured proteins with a very low signal dispersion. J Biomol NMR 50:1–11. doi: 10.1007/s10858-011-9496-2 CrossRefGoogle Scholar
- Oktaviani NA, Risør MW, Lee Y, Megens RP, de Jong DH, Otten R, Scheek RM, Enghild JJ, Nielsen NC, Ikegami T, Mulder FAA (2015) Optimized co-solute paramagnetic relaxation enhancement for the rapid NMR analysis of a highly fibrillogenic peptide. J Biomol NMR 62:129–142. doi: 10.1007/s10858-015-9925-8 CrossRefGoogle Scholar
- Yoshimura Y, Kulminskaya NV, Mulder FAA (2015) Easy and unambiguous sequential assignments of intrinsically disordered proteins by correlating the backbone 15N or 13C′ chemical shifts of multiple contiguous residues in highly resolved 3D spectra. J Biomol NMR 61:109–121. doi: 10.1007/s10858-014-9890-7 CrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.