This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Dual-band selective double cross polarization for heteronuclear polarizationtransfer between dilute spins in solid-state MAS NMR

Zhengfeng Zhang a, Yimin Miao b,c, Xiaoli Liu a, Jun Yang a,⇑, Conggang Li a, Feng Deng a, Riqiang Fu c,⇑a Wuhan Center for Magnetic Resonance, State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Institute of Physics and Mathematics,Chinese Academy of Science, Wuhan 430071, PR Chinab Department of Chemistry and Biochemistry, Florida State University, Tallahassee, FL 32306, USAc National High Magnet Field Lab, Tallahassee, FL 32310, USA

a r t i c l e i n f o

Article history:Received 9 December 2011Revised 21 February 2012Available online 5 March 2012

Keywords:Double cross polarization (DCP)NCA/NCOSolid-state MAS NMRHeteronuclear polarization transfer

a b s t r a c t

A sinusoidal modulation scheme is described for selective heteronuclear polarization transfer betweentwo dilute spins in double cross polarization magic-angle-spinning nuclear magnetic resonance spectros-copy. During the second N ? C cross polarization, the 13C RF amplitude is modulated sinusoidally whilethe 15N RF amplitude is tangent. This modulation induces an effective spin-lock field in two selective fre-quency bands in either side of the 13C RF carrier frequency, allowing for simultaneous polarization trans-fers from 15N to 13C in those two selective frequency bands. It is shown by experiments and simulationsthat this sinusoidal modulation allows one to selectively polarize from 15N to its covalently bonded 13Caand 13C’ carbons in neighboring peptide planes simultaneously, which is useful for establishing the back-bone connectivity between two sequential residues in protein structural elucidation. The selectivity andefficiency were experimentally demonstrated on a uniformly 13C,15N-labeled b1 immunoglobulin bindingdomain of protein G (GB1).

� 2012 Elsevier Inc. All rights reserved.

1. Introduction

In the past decade, solid-state magic-angle-spinning (MAS) NMRhas rapidly emerged as a powerful technique for studying insolubleproteins such as membrane proteins, amyloid proteins, and supra-molecules. Resonance assignment is an essential step in the struc-tural determination of uniformly 15N and 13C labeled proteins. Inthis regard, sequential assignments have been commonly used toextract structural constraints [1–12] via through-spaced dipolarcouplings in well designed multi-dimensional experiments basedon heteronuclear [13,14] and homonuclear dipolar recoupling[15–25] techniques in solid-state MAS NMR. While the heteronu-clear correlation between 13C and 15N is one of the most importantexperiments in such sequential assignments, due to the fact thateach backbone nitrogen covalently bonds with the Ca and carbonyl(C0) carbons in neighboring peptide planes, thus providing usefulthrough-bond links between two sequential residues.

The heteronuclear polarization transfer between 13C and 15Ncan be achieved by the so-called double cross polarization, i.e.DCP [13,26,27], in which 15N is first polarized from the abundant1H spins and then transferred to 13C via a second cross polarizationfrom 15N to 13C. Whereas the 1H–15N or 1H–13C CP are relatively

mature, most of the issues regarding DCP are primarily focusedon the second N ? C CP involving polarization transfer, in termsof selectivity, efficiency, and the influence of RF instability, be-tween two low-c nuclear species [28–42]. For example, SPECIFICCP [31], a DCP experiment with carefully adjusted offsets and RFamplitudes on both the 13C and 15N channels, leads to selectivepolarization transfer from 15N to one of its bonded 13Cs (either13Ca or 13C0), known as the two-dimensional (2D) NCA and NCOexperiments, respectively. With the two NCA and NCO spectra,the connectivity between the Ca[i] and C0[i�1] in the adjacent pep-tide planes can thus be established through N[i]. Alternatively, inthe CANCO experiments [43], the 13Ca[i] is first selectively polar-ized from 1H, and then sequentially transferred to 13C’ [i�1]through two-step selective polarization transfers (i.e. from 13Ca[i] ? 15N[i] and then from 15N[i] ? 13C’[i�1]). However, suchthree-dimensional (3D) experiments (if including the 15N chemicalshift dimension) suffer from very low sensitivity due to the multi-step polarization transfers among the dilute spins. The efficiencyfor each heteronuclear polarization transfer between 13C and 15Nis around 30–40%. Therefore, there is only 10–15% polarizationtransfer efficiency from 13Ca[i] to 13C0[i�1] [43].

In this work, we propose a dual-band selective double crosspolarization to simultaneously polarize the 13Ca[i] and 13C0[i�1]from 15N[i], which allows one to establish the backbone13Ca[i]–15N[i]–13C’[i�1] connectivity in a single 2D 15N–13Ccorrelation spectrum, rather than in the two NCA and NCO spectra

1090-7807/$ - see front matter � 2012 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.jmr.2012.02.020

⇑ Corresponding authors. Address: 1800 E. Paul Dirac Drive, Tallahassee, FL32310, USA. Fax: +1 850 644 1366 (R. Fu).

E-mail addresses: yangjun@wipm.ac.cn (J. Yang), rfu@magnet.fsu.edu (R. Fu).

Journal of Magnetic Resonance 217 (2012) 92–99

Contents lists available at SciVerse ScienceDirect

Journal of Magnetic Resonance

journal homepage: www.elsevier .com/locate / jmr

Author's personal copy

or in a low-sensitive CANCO spectrum, provided that there is a suf-ficient resolution in the 15N dimension. In this new scheme, the RFamplitude of the 13C channel is modulated sinusoidally while thatof the 15N channel is kept constant or tangent-modulated duringthe second N ? C CP. This sinusoidal modulation [44,45] directlyapplies the RF amplitude onto the Ca and C0 chemical shift regionssimultaneously, even if the carrier frequency is away from thoseregions. The modulation period needs not to be commensuratewith the period of sample rotation. We refer to this new methodas sine-modulated double cross polarization (smDCP). It will beshown by experiments and simulations that smDCP can be effec-tively used for the dual-band selective heteronuclear polarizationtransfer using low RF amplitudes, thus improving selectivity andrelaxing the demand on 1H decoupling power. The advantages ofthis new scheme will be illustrated by using a uniformly 13C,15N-labeled b1 immunoglobulin binding domain of protein G (GB1) inthe following sections.

2. Materials and experiments

The 56-residue b1 immunoglobulin binding domain of proteinG (GB1) was expressed in E. coli BL21 (DE3) and minimal media(1.0 g/L 15N NH4Cl, 2.0 g/L 13C glucose), induced with 1.0 mM iso-propyl b-D-thiogalactoside (IPTG) for 4 h. The cell pellet was resus-pended in PBS buffer (50 mM Phosphate, 150 mM NaCl, pH 8.0)and incubated at 80 �C for 5 min. After centrifugation at10,000 rpm for 30 min, the supernatant was purified by ion ex-change (AKTA purifier 10.0), gelfiltration (Sephadex 100) and fol-lowing desalination (AKTA purifier 10.0) with phosphate buffer(pH 8.0). The precipitation of GB1 was incubated over night at4 �C in precipitation solution of 33% isopropanol (IPA) and 67% 2-methylpentane-2,4-diol (MPD). The final microcrystals were cen-trifuged at 5000 rpm for 2 min and transferred into a 4 mm rotor.

All the NMR experiments were performed on a wide-bore Var-ian 600 MHz NMR spectrometer of VNMRS system, equipped witha 4 mm triple-resonance T3-HXY MAS probe, at a temperature of283 K (calibrated in separate experiments using the lead nitratesample [46]). The sample spinning rate was set to 11 kHz ± 2 Hz.Fig. 1 shows the pulse sequences used in our experiments. The

1H p/2 pulse was 3.3 ls, while the 13C p pulse length of 7.2 lswas used for decoupling in the 15N chemical shift dimension (i.e.,t1 dimension). During the first 1H–15N CP with a contact time of1.2 ms, the 15N RF amplitude was about 37 kHz and the 1H RF-fieldwas ramped from 90% to 110% of 57 kHz. The contact time for thesecond N ? C CP was set to 5 ms, while the 15N and 13C RF ampli-tudes were varied depending upon the experimental schemes andwill be described in their corresponding figures. The SPINAL-64[47] sequence with 71 kHz RF amplitude and a 7.9 ls flip pulsewas applied during both t1 and t2 dimension for 1H decoupling,while an 83 kHz CW decoupling was used during the 15N–13C con-tact. The 13C chemical shifts were referenced to adamantane at40.48 ppm of the downfield signal [48] and the 15N referencewas calculated based on the gyromagnetic ratio as recommendedby IUPAC [49]. The RF amplitudes were calculated based on severalmeasured amplitudes with the GB1 sample and the linearity ofVarian linear amplifiers.

Fig. 1(I) shows the standard N ? C CP scheme where the 15N RFamplitude is tangent while the 13C RF amplitude remains constant,or vice versa, to enhance the CP efficiency [30,50]. For comparison,we use the rectangular 13C RF amplitude in our standard DCPexperiments and refer it as ‘rectangular DCP’ in the following dis-cussions. While in the smDCP scheme, the 15N RF amplitude is tan-gent while the 13C RF amplitude x1C is sinusoidally modulated, asshown in Fig. 1(II):

x1CðtÞ ¼ xmax1C sinðxmtÞ; ð1Þ

where, xmax1C and xm represent the maximum RF amplitude and the

modulation frequency, respectively. Mathematically, such a sine-wave modulation induces a spin-lock field in two frequency bandspositioning at distances of ±xm away from the carrier frequency,which allows for polarization transfers from 15N to 13C in thesetwo frequency bands. While the effective RF amplitude applied onthese two frequency bands is just about half of the maximum RFamplitude xmax

1C .

3. Results and discussion

Fig. 2 shows the one-dimensional (1D) 13C spectra of the GB1sample spinning at 11 kHz recorded under different N ? C polari-zation conditions. With the direct CP from 1H to 13C, all 13C reso-nances were observed, as shown in Fig. 2A. Fig. 2B and C showthe SPECIFIC CP spectra using the rectangular DCP scheme (e.g.Fig. 1I) with the selective polarization pathway from 15N to 13Caand 13C0, respectively. In these experiments, the 13C carrier fre-quency was set closely to either Ca or C’ region and its RF ampli-tude was chosen to be low in order to achieve the requiredselectivity. The efficiency for these SPECIFIC CP experiments was�50% as compared to the direct CP from 1H to 13C. If the 13C carrierfrequency was set to the mid way between the Ca and C0 regions,the rectangular DCP experiment could also select both the 13Caand 13C0 region simultaneously with a carefully adjusted RF ampli-tude [13,30,51], as shown in Fig. 2D. However, both simulations(not shown) and experiments indicate that higher 13C RF ampli-tude is needed for such a dual-band selectivity. Unfortunately,using the higher RF amplitude compromises the selectivity. Ascan be seen in Fig. 2D, small aliphatic carbon resonances (from15 to 40 ppm), as well as the spinning sidebands at around100 ppm, appear to be visible. Moreover, for the glycine 13Ca re-gion (�42 ppm), the signal intensities seem to be much smallerthan that in the SPECIFIC CP spectrum. The efficiency for thisdual-band selectivity was �34%, as compared to the spectrum inFig. 2A.

On the other hand, for the sine modulation scheme, even if the13C carrier is set far away from the targeted region, the modulation

CP Dec DecDec1H

t213C

t115N

π

π2

13C

15N

N

...

(I) (II)

N->C CP

Fig. 1. Pulse sequences used in 2D 15N–13C heteronuclear correlation experimentsin solid-state MAS NMR. (I) The standard 15N–13C CP where the 15N RF amplitude istangent but the 13C RF amplitude remains constant. (II) The sine modulated schemewhere the 15N RF amplitude is tangent while the 13C RF amplitude is sinusoidallymodulated at a given modulation frequency xm.

Z. Zhang et al. / Journal of Magnetic Resonance 217 (2012) 92–99 93

Author's personal copy

effectively compensates the carrier offset, such that low RF ampli-tude can be used to achieve the needed selectivity. As shown inFig. 2E and F, the 13C carrier frequency was set to ±16.667 kHzaway from the Ca and C0 regions, respectively. With the sine mod-ulation frequency of 16.667 kHz and a low 13C RF amplitude of6.0 kHz, the Ca and C0 resonances were selectively polarized. Asshown in the figures, the smDCP spectra are almost identical totheir respective SPECIFIC CP spectra, with an efficiency of �50%as compared to the spectrum in Fig. 2A. This implies that both fre-quency bands induced by the sine modulation in either side of thecarrier frequency can be used for the efficient N ? C polarizationtransfer. When the 13C carrier was set to the mid way betweenthe Ca and C0 regions and the modulation frequency matched8.889 kHz (half of the distance between the Ca and C0 regions),both the 13Ca and 13C0 regions were selected simultaneously withthe efficiency of 37%, as shown in Fig. 2G. Apparently, the selectiv-ity was comparable with that in the SPECIFIC NCO and NCA exper-iments, while the unwanted signals appearing in Fig. 2D was notobserved in the spectrum, owing to the use of the low 13C RFamplitude.

Fig. 3 shows the 2D 15N–13C heteronuclear correlation spectra ofthe uniformly 13C,15N-labeled GB1 under different polarizationconditions. Fig. 3A shows the SPECIFIC NCO (left) and NCA (right)spectra recorded in the separate experiments where the 13C carrierfrequency was set closely to the C0 and Ca regions, respectively.Since our sample was prepared at pH 8.0, different from pH 5.5used in the GB1 sample preparation in the literature [52], the pro-tein’s nano-crystalline form might be different. Thus, it is not sur-prising that the GB1 samples at different pHs gave rise to slightlydifferent resonance patterns in the NCO and NCA spectra. However,their spectral resolution in both 13C and 15N dimensions are com-parable [52], confirming that our sample was well prepared. It isworth noting that the slight difference in the resonance patternsresulting from the sample preparations should not affect any con-clusion derived in this paper. Fig. 3A shows some partial assign-ments. Fortunately many resonances could be recognized basedon the GB1 assignments as reported in the literature[53], despitethe fact that there exist some additional peaks, such as the onesnear the T49N-Ca and T49N-A48CO resonances. Fig. 3B showsthe dual-band 15N–13C’/13Ca rectangular DCP spectrum using therectangular 13C RF field during the N ? C contact. Overall, the15N–13C resonance patterns are similar to their respective SPECIFICNCO and NCA spectra in Fig. 3A. However, some resonances such asthe Gly 15N–13Ca resonances (G9N–Ca, G14N–Ca, G38N–Ca, andG41N–Ca) become very weak, while other resonances that didnot appear in the SPECIFIC CP spectra started to show up in thespectrum (such as the peak in the left side of the T25N–Ca reso-nance), implying that the selectivity over the dual bandwidths ishardly uniform, which was also seen in 1D spectrum (e.g.Fig. 2D). It is worth noting that additional cross-peaks around(115 ppm, 103 ppm) and (141 ppm, 41 ppm) (not shown in thespectrum) should belong to the sidebands of 13C0 and 15N and tothe folded-in aliphatic 13C and 15N in sidechains. Thus, the selectiv-ity for the dual-band rectangular DCP is poor owing to the use ofthe required large 13C RF amplitude during the N ? C contact.

Fig. 3C shows the single-band 15N–13C0 (left) and 15N–13Ca(right) smDCP spectra using the sine modulated 13C RF amplitudeduring the N ? C contact. It can be clearly noted that these spectraare almost identical to those in Fig. 3A, except for the one appear-ing at (112 ppm and 179 ppm) (below the T11 N-K10CO reso-nance), again implying that the frequency bandwidths inducedby the sine modulation in both sides of the carrier frequency canbe used for selective N ? C polarization transfer. Fig. 3D showsthe dual-band selective 15N–13C0/13Ca smDCP spectrum when thetwo frequency bandwidths induced by the sine modulation fellinto the C0 and Ca regions simultaneously. Clearly, the resonancepatterns in both the 15N–13C0 and 15N–13Ca regions are almostthe same as in the single-band selective 15N–13C0 and 15N–13Caspectra as shown in Fig. 3A and C, despite the fact that a few addi-tional peaks, such as (130 ppm and 180 ppm), (130 ppm and61 ppm) and (118 ppm and 68 ppm), start to appear or gain inintensity. However, the Gly resonances, which almost disappearedin Fig. 3B, were as intensive as in the single-band selective 15N–13C0

and 15N–13Ca spectra as shown in Fig. 3A and C. On the other hand,those unwanted resonances appearing at around (115 ppm,103 ppm) and (141 ppm, 41 ppm) in Fig. 3B completely disap-peared in the dual-band selective 15N–13C0/13Ca smDCP spectrum.Thus, it is concluded that this sine modulation scheme provides anexcellent selectivity.

Fig. 4 shows the three 1D slices taken along T49N (Top), N37N(Middle), and V29N (Bottom) from the 2D SPECIFIC NCO/NCA spec-tra in Fig. 3A and from the smDCP NC spectrum in Fig. 3D, respec-tively. From their relative intensities, it can be noticed that thesignal-to-noise (S/N) ratio for the dual-band selective smDCP isabout 68–74% of that for the SPECIFIC NCO/NCA, which is consis-tent with the observation of the efficiency in the 1D spectra (c.f.

13C (ppm) 200 180 160 140 120 100 80 60 40 20 0

(D)

(C)

(B)

(A)

(E)

(F)

(G)

*

Fig. 2. 1D 13C CP/MAS spectra of uniformly 13C,15N-labeled GB1 protein: (A) 13Cspectrum directly polarized from 1H. The asterisk indicates the signal of a spacerused to provide a better sealing of the rotor. (B–G) Double cross polarized 13Cspectra using different N ? C polarization conditions, normalized with the spec-trum in (A) to illustrate their polarization efficiency. The arrows in the spectraindicate the 13C carrier position in the experiments. The single-band selectiveSPECIFIC CP NCA (B) using x1N = 9.0 kHz with 10% tangent ramp and x1C = 3.0 kHz,and NCO (C) spectra using x1N = 7.0 kHz with 15% tangent ramp and x1C = 4.4 kHz.The dual-band rectangular DCP 13C spectrum (D) using x1N = 7.0 kHz with 11%tangent ramp and x1C = 16.0 kHz. The NCA (E) and NCO (F) spectra using the sinemodulated 13C RF field with the amplitude of xmax

1C ¼ 6:0 kHz and xm = 16.667 kHzand x1N = 9.0 kHz with 10% and 17% tangent ramp, respectively. The dual-bandselective 13C smDCP spectrum (G) using the sine modulated 13C RF field with theamplitude of xmax

1C ¼ 6:0 kHz and xm = 8.889 kHz and x1N = 9.0 kHz with 13%tangent ramp. The nominal RF amplitudes were calculated assuming a fine linearityof the Varian linear amplifier, which should have an accuracy of 0.5 kHz. In allexperiments, 64 transients were used to accumulate the signals with a recycle delayof 3.5 s.

94 Z. Zhang et al. / Journal of Magnetic Resonance 217 (2012) 92–99

Author's personal copy

Fig. 2). Therefore, the S/N for a single dual-banded smDCP spec-trum (in the same overall time by doubling the number of scans)should be comparable to the two separate SPECIFIC NCO/NCA spec-tra. Our experimental result is beyond what is expected from thefact that the S/N for the dual-banded smDCP is considered to behalf of that for SPECIFIC CP since the 15N magnetization is trans-

ferred to both carbon sites in the former scheme as compared toonly one site in the latter scheme. However, one has to realize thatsuch an expectation holds only when the 15N magnetization iscompletely transferred to the 13C and/or there exhibits the samespin dynamics during the N ? C CP. Due to the fact that the C–Ndipolar coupling is very small (� a few kHz), the polarization trans-

180 175 170

ω2 - 13C (ppm)

ω1

- 15 N

(ppm

)ω

1 - 1

5 N (p

pm)

ω1

- 15 N

(ppm

)ω

1 - 1

5 N (p

pm)

ω1

- 15 N

(ppm

)ω

1 - 1

5 N (p

pm)

130

120

110

100

65 60 55 50 45

130

120

110

100

130

120

110

100

130

120

110

100 ~ ~

~ ~

~ ~

~ ~

(A)

(B)

(C)

(D)

130

120

110

100G14N-Ca

G38N-Ca G41N-Ca

G9N-Ca

T49N-Ca

T11N-Ca T44N-Ca

N37N-Ca

D22N-Ca

V29N-Ca

T25N-Ca

D46N-Ca

N8N-CaA20N-CaD40N-Ca

E56N-Ca

T49N-A48CO

G14N-K13CO

T44N-W43CO

T11N-K10CO

G38N-N37CO

T53N-F52COT17N-T16CO

E15N-G14CO

L5N-K4COE56N-T55COD40N-V39CO

130

120

110

100

Fig. 3. 2D 15N–13C heteronuclear correlation spectra of uniformly 13C,15N-labeled GB1 protein using different N ? C polarization conditions. (A) Single-band SPECIFIC15N–13Ca (Right) and 15N–13C0 (Left) spectra recorded in two separate experiments with the same experimental parameters as in Fig. 2B and C, respectively. (B) Dual-bandrectangular DCP 15N–13Ca/13C0 spectrum using the same parameters as in Fig. 2D. (C) Single-band selective 15N–13Ca (Right) and 15N–13C0 (Left) smDCP spectra recorded intwo separate experiments with the same experimental conditions as in Fig. 2E and F, respectively. (D) Dual-band selective 15N–13Ca/13C’ smDCP spectrum using the sameparameters as in Fig. 2G. The spectra were processed by nmrPipe [54] and plotted with Sparky [55]. In all experiments, 40 transients were used to accumulate the signals foreach t1 increment with a recycle delay of 3.5 s.

Z. Zhang et al. / Journal of Magnetic Resonance 217 (2012) 92–99 95

Author's personal copy

fer rate from N to C is rather slow. A few millisecond CP contacttime (e.g. 5 ms in our experiments) can hardly transfer the 15Nmagnetization completely to 13C. On the other hand, the spindynamics during the N ? C CP may be different in these twoschemes since they exhibit different spin systems during CP (N–Cfor SPECIFIC CP and C–N–C for the dual-banded smDCP). Such a dif-ference might lead to different spin dynamics, as been observed indifferent 1H–S systems [56–58].

In the DCP experiments, low 13C RF amplitude in the N ? Cpolarization transfer will relax the demand on the 1H decouplingfield during the N ? C contact. In our experiments, the maximum13C RF amplitude xmax

1C used in the dual-band selective smDCP

experiments (e.g. Fig. 2G) was about one-third of that used in thedual-band rectangular DCP experiments (e.g. Fig. 2D). Fig. 5 showsthe dependence of the N ? C polarization efficiency as a functionof the applied 1H decoupling power during the N ? C contact. Asthe 1H decoupling power decreasing from 85 to about 71 kHz,the polarization efficiency from N ? Ca dropped to about 60%and 80% for the rectangular and sine modulated DCP schemes,respectively, while the efficiency from N ? C0 decreased to �75%and 90% in their respective spectra. It appears that the 1H decou-pling power is more influential on the Ca carbons, probably owingto the strength of dipolar interactions between 1H and 13C as wellas the difference between the applied 1H decoupling and the 13C RF

185 180 175 70 65 60 55 50 4513C (ppm)13C (ppm)13C (ppm)

170 40

~ ~~ ~

185 180 175 70 65 60 55 50 45170 40

V29N

N35N~ ~

T49N

SPECIFIC NCO

SPECIFIC NCA

smDCP NC

~ ~

Fig. 4. 1D slices taken from the SPECIFIC NCO/NCA spectra in Fig. 3A and from the smDCP NC spectrum in Fig. 3D along T49 N (Top), N37N (Middle), and V29N (Bottom).

85 55 457080 751H Decoupling Field (kHz)

85 55 457080 751H Decoupling Field (kHz)

(A)

(B)

Cnoiger ‘C α region

Fig. 5. Dependence of the N ? C polarization efficiency as a function of the 1H decoupling amplitude during the N ? C contact time in different N ? C polarization schemes.(A) The dual-band rectangular DCP using the parameters in Fig. 2D. (B) The dual-band smDCP using the parameters in Fig. 2G.

96 Z. Zhang et al. / Journal of Magnetic Resonance 217 (2012) 92–99

Author's personal copy

amplitude during the N ? C contact. A large difference betweenthe 1H decoupling and 13C RF amplitude during the N ? C contactis needed to provide sufficient mismatch between the 1H and 13Cspins to avoid potential spin exchange between the two spins,especially when they possess a stronger dipolar coupling. There-fore, this might be the reason why the 15N ? 13Ca polarization effi-ciency for the Gly residues was much worse in the dual-bandrectangular DCP (e.g. Fig. 3B) than in the dual-band smDCP exper-iments (e.g. Fig. 3D). Indeed, the former experiment required fivetimes higher 13C RF amplitude than the latter one under the same1H decoupling power of 83 kHz during the N ? C contact.

It had been reported that the parameters in the DCP experi-ments were sensitive to experimental settings, such as MAS speed,frequency offset and RF fields [33,38,39]. The influence of instabil-ity of 13C RF-field on the N ? C polarization efficiency is also inves-tigated for the dual-band polarization transfers, as shown in Fig. 6.It can be seen in Fig. 6A that the 15N ? 13Ca and 15N ? 13C0 polar-ization efficiencies for the rectangular DCP scheme suffer severelyfrom the variation of 13C RF amplitude and decrease to about 30–50% with about 10% variation. On the other hand, for the smDCPscheme the efficiency from 15N ? 13Ca was just slightly affectedby the 13C RF variation, while about 70% efficiency from15N ? 13C0 was preserved. Thus, there is an advantage to use thissine modulation scheme in long duration experiments.

Fig. 7 shows the simulated selectivity for the sine modulationscheme (e.g. Fig. 1II) as a function of offset and modulation fre-quency xm. In the simulations, the experimental parameters ofx1N = 9 kHz, xmax

1C ¼ 6 kHz, and the MAS rate of 11 kHz were used.Clearly, the frequency selective bandwidth (i.e. the offset rangewithin which the polarization efficiency is above 60%) was aboutxm ± 2 kHz, apparently independent of the spinning rate. In otherwords, the selective bandwidth is about ±2 kHz at distances of±xm away from the carrier frequency. This bandwidth is indepen-dent of spinning rate but depends on the RF amplitude xmax

1C usedin the experiments. For a better selectivity, low-power RF ampli-tude is always preferable. Our simulations (not shown here) sug-gest that the dual-banded smDCP scheme should be effective atthe magnetic field B0 up to 18.8 T with the spinning rate of25 kHz. However, at very high fields where the chemical shiftanisotropy (CSA) interactions become significant, there is a poten-

tial that the efficiency and selectivity could be compromised whenusing low-power RF amplitude which becomes unable to spin-lockthe magnetization during the N ? C CP. It has been demonstratedrecently [60] that low-power symmetry-based band-selective se-quences appear to be an effective approach for applications in highfields and fast sample spinning.

4. Conclusion

We have demonstrated that sinusoidal modulation of the 13C RFamplitude during the second N ? C cross polarization in double

(A)

(B)

C‘ region Cα region

Variation of matched 13C RF amplitude +10% 0 -10% +10% 0 -10%

Fig. 6. Instability effect of 13C RF field on the N ? C polarization efficiency in different schemes. The 13C RF amplitude was varied by ±10% of the optimized values (set to 0) inFig. 2D and G.

0

0.2

0.4

0.6

0.8

1

0

5

10

15

20

25

30

5 10 15 20 25ωm (kHz)

offs

et (k

Hz)

Fig. 7. Simulated selectivity of the sine modulated scheme (e.g. Fig. 1II) as afunction of offset and modulation frequency xm. The SPINEVOLUTION [59] wasused for the simulations. For simplicity, a constant 15N RF-field was utilized, and thespin system was 15N–13C, with a distance of rC–N = 1.448 Å. The chemical shiftparameters for 13C were diso = 0 ppm, daniso = 24 ppm, g = 0.92; for 15N, diso = 0 ppm,daniso = 100 ppm, g = 0.8. In the simulations, x1N = 9 kHz, xmax

1C ¼ 6 kHz, and theMAS rate of 11 kHz were used. The maximum intensity in this plot was normalizedto 1.

Z. Zhang et al. / Journal of Magnetic Resonance 217 (2012) 92–99 97

Author's personal copy

cross polarization (DCP) MAS NMR spectroscopy provides an effec-tive spin-lock field in two selective frequency bands in either sideof the 13C RF carrier frequency, allowing for simultaneous polariza-tion transfers from 15N to 13C in those two selective frequencybands. When the two frequency bands fall into the C0 and Ca re-gions, the dual-band selective heteronuclear polarization transferfrom 15N to both 13C0 and 13Ca can be achieved simultaneouslyin one experiment, rather than in two separate SPECIFIC NCO andNCA experiments or in a low-sensitive CANCO experiment, whichis useful for establishing the backbone connectivity between twosequential residues in protein structural elucidation. It should berecognized that this information about the sequential Ca[i]–N[i]–C’[i�1] topology alone is not sufficient to obtain sequential reso-nance assignments. It has to be combined with other 2D or 3Dexperiments such as NCACX/NCOCX and/or 13C–13C correlationexperiments. With the ability of selecting both the backbone Caand C0 carbons, some elaborate 2D or 3D experiments could be de-signed to simplify crowded resonances typically observed in13C–13C correlation spectra of biological samples being studiedand to help assigning resonances, especially when combined withspecific labeling approaches such as glycerol labeling strategy. Forthe dual-band selective polarization transfer, this new schemegreatly relaxes the demand on both 13C and 1H decoupling ampli-tudes. As compared to standard DCP experiments, this sine modu-lated scheme is also less sensitive to the instability of the 13C RFfield. Moreover, such a modulation could also be applicable fordual-band selective heteronuclear polarization transfer betweenother spin pairs, as mentioned by Baldus et al. [31], such as 19F,31P and even 1H in paramagnetic systems, with the use of a mod-erate RF amplitude when applied to the spin that has a large chem-ical shift distribution, especially under fast MAS condition [45].

Acknowledgments

This work was supported in part by the CAS/SAFEA Interna-tional Partnership Program for Creative Research Teams and byGrants from the National Natural Science Foundation of China(21075133, 21173259, 20921004) and the National Basic ResearchProgram of China (2009CB918600). Y.M. and R.F. are grateful forsupport from the National Institute of Health (NIH) Grant AI23007.

References

[1] K. Takegoshi, S. Nakamura, T. Terao, C-13–H-1 dipolar-assisted rotationalresonance in magic-angle spinning NMR, Chem. Phys. Lett. 344 (2001) 631–637.

[2] F. Castellani, B. van Rossum, A. Diehl, M. Schubert, K. Rehbein, H. Oschkinat,Structure of a protein determined by solid-state magic-angle-spinning NMRspectroscopy, Nature 420 (2002) 98–102.

[3] C.P. Jaroniec, C. Filip, R.G. Griffin, 3D TEDOR NMR experiments for thesimultaneous measurement of multiple carbon–nitrogen distances inuniformly C-13, N-15-labeled solids, J. Am. Chem. Soc. 124 (2002) 10728–10742.

[4] C.R. Morcombe, V. Gaponenko, R.A. Byrd, K.W. Zilm, Diluting abundant spins byisotope edited radio frequency field assisted diffusion, J. Am. Chem. Soc. 126(2004) 7196–7197.

[5] J.R. Lewandowski, G. De Paepe, R.G. Griffin, Proton assisted insensitive nucleicross polarization, J. Am. Chem. Soc. 129 (2007) 728–729.

[6] G. De Paepe, J.R. Lewandowski, A. Loquet, A. Bockmann, R.G. Griffin, Protonassisted recoupling and protein structure determination, J. Chem. Phys. 129(2008).

[7] I. Scholz, M. Huber, T. Manolikas, B.H. Meier, M. Ernst, MIRROR recoupling andits application to spin diffusion under fast magic-angle spinning, Chem. Phys.Lett. 460 (2008) 278–283.

[8] M. Weingarth, D.E. Demco, G. Bodenhausen, P. Tekely, Improvedmagnetization transfer in solid-state NMR with fast magic angle spinning,Chem. Phys. Lett. 469 (2009) 342–348.

[9] I. Scholz, B.H. Meier, M. Ernst, NMR polarization transfer by second-orderresonant recoupling: RESORT, Chem. Phys. Lett. 485 (2010) 335–342.

[10] M. Weingarth, G. Bodenhausen, P. Tekely, Broadband magnetization transferusing moderate radio-frequency fields for NMR with very high static fields andspinning speeds, Chem. Phys. Lett. 488 (2010) 10–16.

[11] G.J. Hou, S. Yan, S.J. Sun, Y. Han, I.J.L. Byeon, J. Ahn, J. Concel, A. Samoson, A.M.Gronenborn, T. Polenova, Spin diffusion driven by R-symmetry sequences:applications to homonuclear correlation spectroscopy in MAS NMR ofbiological and organic solids, J. Am. Chem. Soc. 133 (2011) 3943–3953.

[12] B.W. Hu, O. Lafon, J. Trebosc, Q. Chen, J.P. Amoureux, Broad-band homo-nuclear correlations assisted by (1)H irradiation for bio-molecules in very highmagnetic field at fast and ultra-fast MAS frequencies, J. Magn. Reson. 212(2011) 320–329.

[13] J. Schaefer, R.A. McKay, E.O. Stejskal, Double-cross-polarization NMR of solids,J. Magn. Reson. 34 (1979) 443–447.

[14] A.W. Hing, S. Vega, J. Schaefer, Transferred-echo double-resonance NMR, J.Magn. Reson. 96 (1992) 205–209.

[15] D.P. Raleigh, M.H. Levitt, R.G. Griffin, Rotational resonance in solid-state NMR,Chem. Phys. Lett. 146 (1988) 71–76.

[16] R. Tycko, G. Dabbagh, Measurement of nuclear magnetic dipole–dipolecouplings in magic angle spinning NMR, Chem. Phys. Lett. 173 (1990) 461–465.

[17] A.E. Bennett, J.H. Ok, R.G. Griffin, S. Vega, Chemical shift correlationspectroscopy in rotating solids: radio frequency-driven dipolar recouplingand longitudinal exchange, J. Chem. Phys. 96 (1992) 8624–8627.

[18] R. Tycko, S.O. Smith, Symmetry principles in the design of pulse sequences forstructural measurements in magic angle spinning nuclear-magnetic-resonance, J. Chem. Phys. 98 (1993) 932–943.

[19] N.C. Nielsen, H. Bildsoe, H.J. Jakobsen, M.H. Levitt, Double-quantumhomonuclear rotary resonance: efficient dipolar recovery in magic-angle-spinning nuclear-magnetic-resonance, J. Chem. Phys. 101 (1994)1805–1812.

[20] D.M. Gregory, D.J. Mitchell, J.A. Stringer, S. Kiihne, J.C. Shiels, J. Callahan, M.A.Mehta, G.P. Drobny, Windowless dipolar recoupling: the detection of weakdipolar couplings between spin-1/2 nuclei with large chemical-shiftanisotropies, Chem. Phys. Lett. 246 (1995) 654–663.

[21] Y.K. Lee, N.D. Kurur, M. Helmle, O.G. Johannessen, N.C. Nielsen, M.H. Levitt,Efficient dipolar recoupling in the NMR of rotating solids: a sevenfoldsymmetrical radiofrequency pulse sequence, Chem. Phys. Lett. 242 (1995)304–309.

[22] B.Q. Sun, P.R. Costa, D. Kocisko, P.T. Lansbury, R.G. Griffin, Internucleardistance measurements in solid state nuclear magnetic resonance: dipolarrecoupling via rotor synchronized spin locking, J. Chem. Phys. 102 (1995) 702–707.

[23] M. Hohwy, H.J. Jakobsen, M. Eden, M.H. Levitt, N.C. Nielsen, Broadband dipolarrecoupling in the nuclear magnetic resonance of rotating solids: acompensated C7 pulse sequence, J. Chem. Phys. 108 (1998) 2686–2694.

[24] R. Verel, M. Baldus, M. Ernst, B.H. Meier, A homonuclear spin-pair filter forsolid-state NMR based on adiabatic-passage techniques, Chem. Phys. Lett. 287(1998) 421–428.

[25] M. Hohwy, C.M. Rienstra, C.P. Jaroniec, R.G. Griffin, Fivefold symmetrichomonuclear dipolar recoupling in rotating solids: application to doublequantum spectroscopy, J. Chem. Phys. 110 (1999) 7983–7992.

[26] J. Schaefer, E.O. Stejskal, J.R. Garbow, R.A. McKay, Quantitative determinationof the concentrations of C-13–N-15 chemical bonds by double cross-polarization NMR, J. Magn. Reson. 59 (1984) 150–156.

[27] E.O. Stejskal, J. Schaefer, R.A. McKay, Analysis of double cross-polarizationrates in solid proteins, J. Magn. Reson. 57 (1984) 471–485.

[28] X.L. Wu, K.W. Zilm, Cross-polarization with high-speed magic-angle-spinning,J. Magn. Reson., Ser. A 104 (1993) 154–165.

[29] B.Q. Sun, P.R. Costa, R.G. Griffin, Heteronuclear polarization transfer byradiofrequency-driven dipolar recoupling under magic-angle-spinning, J.Magn. Reson., Ser. A 112 (1995) 191–198.

[30] M. Baldus, D.G. Geurts, S. Hediger, B.H. Meier, Efficient N-15–C-13 polarizationtransfer by adiabatic-passage Hartmann–Hahn cross polarization, J. Magn.Reson., Ser. A 118 (1996) 140–144.

[31] M. Baldus, A.T. Petkova, J. Herzfeld, R.G. Griffin, Cross polarization in the tiltedframe: assignment and spectral simplification in heteronuclear spin systems,Mol. Phys. 95 (1998) 1197–1207.

[32] A. Brinkmann, M.H. Levitt, Symmetry principles in the nuclear magneticresonance of spinning solids: heteronuclear recoupling by generalizedHartmann–Hahn sequences, J. Chem. Phys. 115 (2001) 357–384.

[33] M. Bjerring, N.C. Nielsen, Solid-state NMR heteronuclear coherence transferusing phase and amplitude modulated rf irradiation at the Hartmann–Hahnsideband conditions, Chem. Phys. Lett. 382 (2003) 671–678.

[34] C.T. Kehlet, A.C. Sivertsen, M. Bjerring, T.O. Reiss, N. Khaneja, S.J. Glaser, N.C.Nielsen, Improving solid-state NMR dipolar recoupling by optimal control, J.Am. Chem. Soc. 126 (2004) 10202–10203.

[35] N. Khaneja, Sensitivity enhanced recoupling experiments in solid-state NMRby gamma preparation, J. Magn. Reson. 183 (2006) 242–251.

[36] N. Khaneja, C. Kehlet, S.J. Glaser, N.C. Nielsen, Composite dipolar recoupling:anisotropy compensated coherence transfer in solid-state nuclear magneticresonance, J. Chem. Phys. 124 (2006) 114503.

[37] C. Kehlet, M. Bjerring, A.C. Sivertsen, T. Kristensen, J.J. Enghild, S.J. Glaser, N.Khaneja, N.C. Nielsen, Optimal control based NCO and NCA experiments forspectral assignment in biological solid-state NMR spectroscopy, J. Magn.Reson. 188 (2007) 216–230.

[38] J.H. Yang, F.C. Chou, D.L.M. Tzou, Coherence selection in double CP MAS NMRspectroscopy, J. Magn. Reson. 195 (2008) 116–120.

[39] J. Lin, M.J. Bayro, R.G. Griffin, N. Khaneja, Dipolar recoupling in solid state NMRby phase alternating pulse sequences, J. Magn. Reson. 197 (2009) 145–152.

98 Z. Zhang et al. / Journal of Magnetic Resonance 217 (2012) 92–99

Author's personal copy

[40] M. Bjerring, A.B. Nielsen, Z. Tosner, N.C. Nielsen, Broadband heteronucleardipolar recoupling without (1)H decoupling in solid-state NMR using simplecross-polarization methods, Chem. Phys. Lett. 494 (2010) 326–330.

[41] A.B. Nielsen, L.A. Straaso, A.J. Nieuwkoop, C.M. Rienstra, M. Bjerring, N.C.Nielsen, Broadband heteronuclear solid-state NMR experiments byexponentially modulated dipolar recoupling without decoupling, J. Phys.Chem. Lett. 1 (2010) 1952–1956.

[42] J. Lin, R.G. Griffin, N. Khaneja, Recoupling in solid state NMR using gammaprepared states and phase matching, J. Magn. Reson. 212 (2011) 402–411.

[43] Y. Li, D.A. Berthold, H.L. Frericks, R.B. Gennis, C.M. Rienstra, Partial C-13 and N-15 chemical-shift assignments of the disulfide-bond-forming enzyme DsbB by3D magic-angle spinning NMR spectroscopy, ChemBioChem 8 (2007) 434–442.

[44] D. Artemov, Z.M. Bhujwalla, J.D. Glickson, Band-selective heteronuclear crosspolarization in liquids, J. Magn. Reson., Ser. B 107 (1995) 286–288.

[45] J.P. Demers, V. Vijayan, S. Becker, A. Lange, Tailored low-power cross-polarization under fast magic-angle spinning, J. Magn. Reson. 205 (2010)216–223.

[46] G. Neue, C. Dybowski, Determining temperature in a magic-angle spinningprobe using the temperature dependence of the isotropic chemical shift of leadnitrate, Solid State Nucl. Magn. Reson. 7 (1997) 333–336.

[47] B.M. Fung, A.K. Khitrin, K. Ermolaev, An improved broadband decouplingsequence for liquid crystals and solids, J. Magn. Reson. 142 (2000) 97–101.

[48] C.R. Morcombe, K.W. Zilm, Chemical shift referencing in MAS solid state NMR,J. Magn. Reson. 162 (2003) 479–486.

[49] R.K. Harris, E.D. Becker, S.M.C. De Menezes, R. Goodfellow, P. Granger, NMRnomenclature. Nuclear spin properties and conventions for chemical shifts –(IUPAC recommendations, 2001), Pure Appl. Chem. 73 (2001) 1795–1818.

[50] S. Hediger, B.H. Meier, R.R. Ernst, Adiabatic passage Hartmann–Hahn cross-polarization in NMR under magic-angle sample-spinning, Chem. Phys. Lett.240 (1995) 449–456.

[51] C.M. Rienstra, M. Hohwy, M. Hong, R.G. Griffin, 2D and 3D N-15–C-13–C-13NMR chemical shift correlation spectroscopy of solids: assignment of MASspectra of peptides, J. Am. Chem. Soc. 122 (2000) 10979–10990.

[52] W.T. Franks, K.D. Kloepper, B.J. Wylie, C.M. Rienstra, Four-dimensionalheteronuclear correlation experiments for chemical shift assignment of solidproteins, J. Biomol. NMR 39 (2007) 107–131.

[53] W.T. Franks, D.H. Zhou, B.J. Wylie, B.G. Money, D.T. Graesser, H.L. Frericks, G.Sahota, C.M. Rienstra, Magic-angle spinning solid-state NMR spectroscopy ofthe beta 1 immunoglobulin binding domain of protein G (GB1): N-15 and C-13chemical shift assignments and conformational analysis, J. Am. Chem. Soc. 127(2005) 12291–12305.

[54] F. Delaglio, S. Grzesiek, G.W. Vuister, G. Zhu, J. Pfeifer, A. Bax, NMRPIPE: amultidimensional spectral processing system based on UNIX pipes, J. Biomol.NMR 6 (1995) 277–293.

[55] T.D. Goddard, D.G. Kneller, Sparky 3, University of California, San Francisco.[56] M. Mehring, Principles of High Resolution NMR in Solids, second ed., Springer-

Verlag, New York, 1983.[57] L.B. Alemany, D.M. Grant, T.D. Alger, R.J. Pugmire, Cross polarization and magic

angle sample spinning NMR spectra of model organic compounds. 3. Effect ofthe 13C–1H dipolar interaction on cross polarization and carbon-protondephasing, J. Am. Chem. Soc. 105 (1983) 6697–6704.

[58] R. Fu, J. Hu, T.A. Cross, Towards quantitative measurements in solid-stateCPMAS NMR: a Lee–Goldburg frequency modulated cross polarization scheme,J. Magn. Reson. 168 (2004) 8–17.

[59] M. Veshtort, R.G. Griffin, SPINEVOLUTION: a powerful tool for the simulationof solid and liquid state NMR experiments, J. Magn. Reson. 178 (2006) 248–282.

[60] C. Herbst, J. Herbst, J. Leppert, O. Ohlenschlaeger, M. Gorlach, R.Ramachandran, Chemical shift correlation at high MAS frequenciesemploying low-power symmetry-based mixing schemes, J. Biomol. NMR 50(2011) 277–284.

Z. Zhang et al. / Journal of Magnetic Resonance 217 (2012) 92–99 99