2D relayed anisotropy correlation NMR: Characterization of the 13C′ chemical shift tensor orientation in the peptide plane of the dipeptide AibAib

An approach to the determination of the orientation of the carbonyl chemical shift (CS) tensor in a ¹³C-¹⁵N-¹H dipolar coupled spin network is proposed. The method involves the measurement of the Euler angles of the ¹³C-¹⁵N and ¹⁵N-¹H dipolar vectors in the ¹³C CS tensor principal axes system, respectively, via a ¹³C-¹⁵N REDOR experiment and by a 2D relayed anisotropy correlation of the ¹³C CSA (₂) and ¹⁵N-¹H dipolar interaction (₁). Via numerical simulations the sensitivity of the ₁ cross sections of the 2D spectrum to the Euler angles of the ¹⁵N-¹H bond vector in the ¹³C CSA frame is shown. Employing the procedure outlined in this work, we have determined the orientation of the ¹³C CS tensor in the peptide plane of the dipeptide AibAib-NH₂ (Aib = -aminoisobutyric acid). The Euler angles are found to be (_CN, _CN) = (34° ± 2°, 88° ± 2° ) and (_NH, _NH) = (90° ± 10°, 80° ± 10° ). From the measured Euler angles it is seen that the ₃₃ and ₂₂ components of the ¹³C CS tensor approximately lie in the peptide plane.     

Determination of <Superscript>13</Superscript>C CSA Tensors: Extension of the Model-independent Approach to an RNA Kissing Complex Undergoing Anisotropic Rotational Diffusion in Solution

Chemical shift anisotropy (CSA) tensor parameters have been determined for the protonated carbons of the purine bases in an RNA kissing complex in solution by extending the model-independent approach [Fushman, D., Cowburn, D. (1998) J. Am. Chem. Soc. 120, 7109–7110]. A strategy for determining CSA tensor parameters of heteronuclei in isolated X–H two-spin systems (X = ¹³C or ¹⁵N) in molecules undergoing anisotropic rotational diffusion is presented. The original method relies on the fact that the ratio κ₂=R₂^auto/R₂^cross of the transverse auto- and cross-correlated relaxation rates involving the X CSA and the X–H dipolar interaction is independent of parameters related to molecular motion, provided rotational diffusion is isotropic. However, if the overall motion is anisotropic κ₂ depends on the anisotropy D_||/D_⊥ of rotational diffusion. In this paper, the field dependence of both κ₂ and its longitudinal counterpart κ₁=R₁^auto/R₁^cross are determined. For anisotropic rotational diffusion, our calculations show that the average κ_av = 1/2 (κ₁+κ₂), of the ratios is largely independent of the anisotropy parameter D_||/D_⊥. The field dependence of the average ratio κ_av may thus be utilized to determine CSA tensor parameters by a generalized model-independent approach in the case of molecules with an overall motion described by an axially symmetric rotational diffusion tensor.     

Sensitivity enhancement in (HCA)CONH experiments

A novel sensitivity-enhancement technique is proposed for experiments which correlate protein backbone resonances and start with magnetization from ¹³C-¹H groups. The technique is based on replenishing magnetization lost by dipole-CSA cross-correlated relaxation of the ¹³C spin with ¹³C steady state magnetization. The principle is demonstrated for the (HCA)CONH experiment, resulting in 1.6-fold sensitivity enhancement compared to the HN(CA)CO experiment. Furthermore, other versions of the (HCA)CONH experiment were evaluated, including a novel experiment with spin-locking of transverse ¹³C-¹H two-spin coherence, and a cross-correlation compensated (CA)CONH experiment which starts from ¹³C rather than ¹H magnetization.     

Three-dimensional 1H-TOCSY-relayed ct-[13C,1H]-HMQC for aromatic spin system identification in uniformly 13C-labeled proteins

Summary Three-dimensional ¹H-TOCSY-relayed ct-[¹³C,¹H]-HMQC is a novel experiment for aromatic spin system identification in uniformly ¹³C-labeled proteins, which is implemented so that it correlates the chemical shift of a given aromatic proton with those of the directly attached carbon and all vicinal protons. The ct-HMQC scheme is used both for overlay of the indirect ¹H and ¹³C chemical shift evolution periods and for the generation of ¹H-¹H antiphase magnetization to accelerate the ¹H-TOCSY magnetization transfer at short mixing times. As an illustration, data recorded for the 18 kDa protein cyclophilin A are presented. Since transverse relaxation of ¹³C-¹H zero-quantum and double-quantum coherences is to first order insensitive to ¹³C-¹H heteronuclear dipolar relaxation, the new experiment should work also for proteins with molecular weights above 20 kDa.     

Measurement of methyl 13C-1H cross-correlation in uniformly 13C-, 15N-, labeled proteins

An understanding of side chain motions in protein is of great interest since side chains often play an important role in protein folding and intermolecular interactions. A novel method for measuring the dynamics of methyl groups in uniformly ¹³C-, ¹⁵N-labeled proteins has been developed by our group. The method relies on the difference in peak intensities of ¹³C quartet components of methyl groups, in a spectrum recording the free evolution of ¹³C under proton coupling in a constant-time period. Cross-correlated relaxation rates between ¹³C-¹H dipoles can be easily measured from the intensities of the multiplet components. The degree of the methyl restrictions (S ²) can be estimated from the cross-correlated relaxation rate. The method is demonstrated on a sample of human fatty acid binding protein in the absence of fatty acid. We obtained relaxation data for 33 out of 46 residues having methyl groups in apo-IFABP. It has been found that the magnitude of the CSA tensor of spin ¹³C in a methyl group could be estimated from the intensities of the ¹³C multiplet components.     

13C-labeled oligodeoxyribonucleotides: A solution study of a CCAAT-containing sequence at the nuclear factor I recognition site of human adenovirus

The solution behavior of the single-stranded CCAAT-containing octamer 1 , d(AGCCAATA), that comprises part of the nuclear factor I (NF-I) recognition site at the origin of replication of human adenovirus has been studied by nmr spectroscopy at 500 and 600 MHz. Proton resonance assignments for 1 were aided by selective ¹³C enrichment at C1′ of A₁ or A₅. High-resolution ¹³C-¹H heteronuclear multiple-bond coherence spectra of the ¹³C-labeled oligomers permitted the selective detection of furanosyl ring protons within each labeled residue due to short- and long-range ¹³C-¹H couplings to the enriched C1′. The resulting assignments provided firm starting points in the interpretation of double quantum filtered correlated spectra, yielding information supplemented by total correlated spectroscopy (TOCSY) and rotating frame nuclear Overhauser effect spectroscopic data to completely assign the ¹H-nmr spectrum of 1 and extract ³J_HH values for furanose con-formational analysis. Several ¹³C-¹H spin-coupling constants within the ¹³C-enriched A₁ or A₅ residues were measured from cross-peak shifts in TOCSY spectra, and their signs determined by inspection of the relative orientations of these shifts. ¹H-²-H and ¹³C-¹H spin-couplings both indicate a preference (> 75%) for south (C2′-endo) conformations by the furanosyl rings of 1 . © 1994 John Wiley & Sons, Inc.     

Influence of the O-phosphorylation of serine, threonine and tyrosine in proteins on the amidic 15N chemical shielding anisotropy tensors

Density functional theory was employed to study the influence of O-phosphorylation of serine, threonine, and tyrosine on the amidic ¹⁵N chemical shielding anisotropy (CSA) tensor in the context of the complex chemical environments of protein structures. Our results indicate that the amidic ¹⁵N CSA tensor has sensitive responses to the introduction of the phosphate group and the phosphorylation-promoted rearrangement of solvent molecules and hydrogen bonding networks in the vicinity of the phosphorylated site. Yet, the calculated ¹⁵N CSA tensors in phosphorylated model peptides were in range of values experimentally observed for non-phosphorylated proteins. The extent of the phosphorylation induced changes suggests that the amidic ¹⁵N CSA tensor in phosphorylated proteins could be reasonably well approximated with averaged CSA tensor values experimentally determined for non-phosphorylated amino acids in practical NMR applications, where chemical surrounding of the phosphorylated site is not known a priori in majority of cases. Our calculations provide estimates of relative errors to be associated with the averaged CSA tensor values in interpretations of NMR data from phosphorylated proteins.     

The structural properties of the transmembrane segment of the integral membrane protein phospholamban utilizing C CPMAS, H, and REDOR solid-state NMR spectroscopy

Solid-state NMR spectroscopic techniques were used to investigate the secondary structure of the transmembrane peptide phospholamban (TM-PLB), a sarcoplasmic Ca²⁺ regulator. ¹³C cross-polarization magic angle spinning spectra of ¹³C carbonyl-labeled Leu39 of TM-PLB exhibited two peaks in a pure 1-palmitoyl-2-oleoyl-phosphocholine (POPC) bilayer, each due to a different structural conformation of phospholamban as characterized by the corresponding ¹³C chemical shift. The addition of a negatively charged phospholipid (1-palmitoyl-2-oleoylphosphatidylglycerol (POPG)) to the POPC bilayer stabilized TM-PLB to an α-helical conformation as monitored by an enhancement of the α-helical carbonyl ¹³C resonance in the corresponding NMR spectrum. ¹³C-¹⁵N REDOR solid-state NMR spectroscopic experiments revealed the distance between the ¹³C carbonyl carbon of Leu39 and the ¹⁵N amide nitrogen of Leu42 to be 4.2 ± 0.2Å indicating an α-helical conformation of TM-PLB with a slight deviation from an ideal 3.6 amino acid per turn helix. Finally, the quadrupolar splittings of three ²H labeled leucines (Leu28, Leu39, and Leu51) incorporated in mechanically aligned DOPE/DOPC bilayers yielded an 11° ± 5° tilt of TM-PLB with respect to the bilayer normal. In addition to elucidating valuable TM-PLB secondary structure information, the solid-state NMR spectroscopic data indicates that the type of phospholipids and the water content play a crucial role in the secondary structure and folding of TM-PLB in a phospholipid bilayer.     

Order Parameters of a Transmembrane Helix in a Fluid Bilayer: Case Study of a WALP Peptide

A new solid-state NMR-based strategy is established for the precise and efficient analysis of orientation and dynamics of transmembrane peptides in fluid bilayers. For this purpose, several dynamically averaged anisotropic constraints, including ¹³C and ¹⁵N chemical shift anisotropies and ¹³C-¹⁵N dipolar couplings, were determined from two different triple-isotope-labeled WALP23 peptides (²H, ¹³C, and ¹⁵N) and combined with previously published quadrupolar splittings of the same peptide. Chemical shift anisotropy tensor orientations were determined with quantum chemistry. The complete set of experimental constraints was analyzed using a generalized, four-parameter dynamic model of the peptide motion, including tilt and rotation angle and two associated order parameters. A tilt angle of 21° was determined for WALP23 in dimyristoylphosphatidylcholine, which is much larger than the tilt angle of 5.5° previously determined from ²H NMR experiments. This approach provided a realistic value for the tilt angle of WALP23 peptide in the presence of hydrophobic mismatch, and can be applied to any transmembrane helical peptide. The influence of the experimental data set on the solution space is discussed, as are potential sources of error.     

High resolution 13C NMR spectra on oriented lipid bilayers: from quantifying the various sources of line broadening to performing 2D experiments with 0.2-0.3 ppm resolution in the carbon dimension

¹³C NMR spectra routinely performed on oriented lipid bilayers display linewidth of 1–2 ppm, although T₂ measurements indicate that 0.1–0.2 ppm could be obtained. We have prepared a DMPC – ¹³C₄-cholesterol (7/3) sample, and oriented the lipid bilayers between glass plates so that the bilayer normal makes an angle of 90° (or of the magic angle) with B₀. We have measured T₂s, CSAs, and linewidths for the choline ¹³C--methyl, the cholesterol-C₄ carbons and the lipid head group phosphorus, at both angles and 313 K. The magnetic field distribution within the sample was calculated using the surface current formalism. The line shapes were simulated as a function of B₀ field inhomogeneities and sample mosaic spread. Both effects contribute to the experimental linewidth. Using three signals of different CSA, we have quantified both contributions and measured the mosaic spread accurately. Direct shimming on a sample signal is essential to obtain sharp resonances and ¹³C labelled choline methyl resonance of DMPC is a good candidate for this task. After optimisation of the important parameters (shimming on the choline resonance, mosaic spread of ±0.30° ), ¹³C linewidth of 0.2–0.3 ppm have been obtained. This newly achieved resolution on bilayers oriented at 90°, has allowed to perform two 2D experiments, with a good sensitivity: 2D PELF (correlation of carbon chemical shifts and C-H dipolar couplings) and 2D D-resolved experiment (correlation of carbon chemical shifts and C-C dipolar couplings). A C-C dipolar coupling of 35 ± 2 Hz between the choline methyl carbons was determined.     

Two-and three-dimensional HCN experiments for correlating base and sugar resonances in 15N, 13C-labeled RNA oligonucleotides

Summary New 2D and 3D ¹H-¹³C-¹⁵N triple resonance experiments are presented which allow unambiguous assignments of intranucleotide H1'-H8(H6) connectivities in ¹³C-and ¹⁵N-labeled RNA oligonucleotides. Two slightly different experiments employing double INEPT forward and back coherence transfers are optimized to obtain the H1'-C1'-N9/N1 and H8/H6-C8/C6-N9/N1 connectivities, respectively. The correlation of H1' protons to glycosidic nitrogens N9/N1 is obtained in a nonselective fashion. To correlate H8/H6 with their respective glycosidic nitrogens, selective ¹³C-refocusing and ¹⁵N-inversion pulses are applied to optimize the magnetization transfers along the desired pathway. The approach employs the heteronuclear one-bond spin-spin interactions and allows the 2D ¹H-¹⁵N and 3D¹H-¹³C-¹⁵N chemical shift correlation of nuclei along and adjacent to the glycosidic bond. Since the intranucleotide correlations obtained are based exclusively on through-bond scalar interactions, these experiments resolve the ambiguity of intra-and internucleotide H1'-H8(H6) assignments obtained from the 2D NOESY spectra. These experiments are applied to a 30-base RNA oligonucleotide which contains the binding site for Rev protein from HIV.     

Robust refocusing of 13C magnetization in multidimensional NMR experiments by adiabatic fast passage pulses

We show that adiabatic fast passage (AFP) pulses are robust refocusing elements of transverse ¹³C magnetization in multidimensional NMR experiments. A pair of identical AFP pulses can refocus selected parts or a complete ¹³ C chemical shift range in ¹³C spectra. In the constant time ¹³C-¹H HSQC, replacement of attenuated rectangular pulses by selective AFP pulses results in a sensitivity enhancement of up to a factor of 1.8. In the 3D CBCA(CO)NH the signal-to-noise ratio is increased by a factor of up to 1.6.     

Determination of the residue-specific 15N CSA tensor principal components using multiple alignment media

The individual components of the backbone ¹⁵N CSA tensor, σ₁₁, σ₂₂, σ₃₃, and the orientation of σ₁₁ relative to the NH bond described by the angle β have been determined for uniformly labeled ¹⁵N, ¹³C ubiquitin from partial alignment in phospholipid bicelles, Pf1 phage, and poly(ethylene glycol) by measuring the residue-specific residual dipolar couplings and chemical shift deviations. No strong correlation between any of the CSA tensor components is observed with any single structural feature. However, the experimentally determined tensor components agree with the previously determined average CSA principal components [Cornilescu and Bax (2000) J. Am. Chem. Soc. 122, 10143–10154]. Significant deviations from the averages coincide with residues in β-strand or extended regions, while α-helical residue tensor components cluster close to the average values.Electronic Supplementary Material Supplementary material is available to authorised users in the online version of this article at .     

HNCA-TOCSY-CANH experiments with alternate <Superscript>13</Superscript>C-<Superscript>12</Superscript>C labeling: a set of 3D experiment with unique supra-sequential information for mainchain resonance assignment

Described here is a set of three-dimensional (3D) NMR experiments that rely on CACA-TOCSY magnetization transfer via the weak ³ \textJ_{\textCa\textCa} ^{ 3} {\text{J}}_{{{\text{C}}\alpha {\text{C}}\alpha }} coupling. These pulse sequences, which resemble recently described ¹³C detected CACA-TOCSY (Takeuchi et al. 2010) experiments, are recorded in ¹H₂O, and use ¹H excitation and detection. These experiments require alternate ¹³C-¹²C labeling together with perdeuteration, which allows utilizing the small ³ \textJ_{\textCa\textCa} ^{ 3} {\text{J}}_{{{\text{C}}\alpha {\text{C}}\alpha }} scalar coupling that is otherwise masked by the stronger ¹J_CC couplings in uniformly ¹³C labeled samples. These new experiments provide a unique assignment ladder-mark that yields bidirectional supra-sequential information and can readily straddle proline residues. Unlike the conventional HNCA experiment, which contains only sequential information to the ^{1 3} \textC^a ^{ 1 3} {\text{C}}^{\alpha } of the preceding residue, the 3D hnCA-TOCSY-caNH experiment can yield sequential correlations to alpha carbons in positions i−1, i + 1 and i−2. Furthermore, the 3D hNca-TOCSY-caNH and Hnca-TOCSY-caNH experiments, which share the same magnetization pathway but use a different chemical shift encoding, directly couple the ¹⁵N-¹H spin pair of residue i to adjacent amide protons and nitrogens at positions i−2, i−1, i + 1 and i + 2, respectively. These new experimental features make protein backbone assignments more robust by reducing the degeneracy problem associated with the conventional 3D NMR experiments.     

Biosynthetically directed fractional 13C labeling facilitates identification of Phe and Tyr aromatic signals in proteins

Analysis of 2D [¹³C,¹H]-HSQC spectra of biosynthetic fractionally ¹³C labeled proteins is a reliable, straightforward means to obtain stereospecific assignments of Val and Leu methyl sites in proteins. Herein we show that the same fractionally labeled protein sample facilitates observation and identification of Phe and Tyr aromatic signals. This is the case, in part, because the fractional ¹³C labeling yields aromatic rings in which some of the ¹³C-¹³C J-couplings, present in uniformly labeled samples, are absent. Also, the number of homonuclear J-coupling partners differs for the -, - and -carbons. This enabled us to vary their signal intensities in distinctly different ways by appropriately setting the ¹³C constant-time period in 2D [¹³C,¹H]-HSQC spectra. We illustrate the application of this approach to an 18 kDa protein, c-VIAF, a modulator of apoptosis. In addition, we show that cancellation of the aromatic ¹³C CSA and ¹³C-¹H dipolar interactions can be fruitfully utilized in the case of the fractionally labeled sample to obtain high resolution ¹³C constant-time spectra with good sensitivity.     

Thermodynamic analysis of hydration in human serum heme-albumin

Ferric human serum heme-albumin (heme-HSA) shows a peculiar nuclear magnetic relaxation dispersion (NMRD) behavior that allows to investigate structural and functional properties. Here, we report a thermodynamic analysis of NMRD profiles of heme-HSA between 20 and 60 °C to characterize its hydration. NMRD profiles, all showing two Lorentzian dispersions at 0.3 and 60 MHz, were analyzed in terms of modulation of the zero field splitting tensor for the S = ⁵/₂ manifold. Values of correlation times for tensor fluctuation (τ_v) and chemical exchange of water molecules (τ_M) show the expected temperature dependence, with activation enthalpies of −1.94 and −2.46 ± 0.2 kJ mol⁻¹, respectively. The cluster of water molecules located in the close proximity of the heme is progressively reduced in size by increasing the temperature, with ΔH = 68 ± 28 kJ mol⁻¹ and ΔS = 200 ± 80 J mol⁻¹ K⁻¹. These results highlight the role of the water solvent in heme-HSA structure-function relationships.     

Estimates of methyl 13C and 1H CSA values (Δσ) in proteins from cross-correlated spin relaxation

Simple pulse schemes are presented for the measurement of methyl ¹³C and ¹H CSA values from ¹H–¹³C dipole/¹³C CSA and ¹H–¹³C dipole/¹H CSA cross-correlated relaxation. The methodology is applied to protein L and malate synthase G. Average ¹³C CSA values are considerably smaller for Ile than Leu/Val (17 vs 25 ppm) and are in good agreement with previous solid state NMR studies of powders of amino acids and dipeptides and in reasonable agreement with quantum-chemical DFT calculations of methyl carbon CSA values in peptide fragments. Small averaged ¹H CSA values on the order of 1 ppm are measured, consistent with a solid state NMR determination of the methyl group ¹H CSA in dimethylmalonic acid.     

Editing and diagonal peak suppression in three-dimensional HCCH protein NMR correlation experiments

A novel three-dimensional (3D) HCCH NMR experiment is introduced. It involves ¹³C-¹³C COSY or TOCSY coherence transfer plus two independent editing steps according to the number of protons attached to the individual carbons before and after the ¹³C-¹³C homonuclear mixing. This double editing leads to simplification of HCCH protein side chain spectra that otherwise are prone to spectral overlap. Another interesting feature is amino acid selectivity, i.e. that the presence of certain correlations in a doubly edited HCCH subspectrum gives a clue as to assignment to a particular subgroup of amino acids or segments thereof. Finally, the selection of two different multiplicities in the two editing steps leads to diagonal peak suppression in the ¹H-¹H (3D spectrum recorded with two ¹H and one ¹³C dimension) or the ¹³C-¹³C (3D spectrum recorded with one ¹H and two ¹³C dimensions) two-dimensional projection. The new experiment is demonstrated using a ¹³C,¹⁵N-labeled protein sample, chymotrypsin inhibitor 2, at 500 MHz.     

13C-2H correlation NMR spectroscopy studies of the in vivo transformations of natural products from Artemisia annua

¹³C-²H correlation NMR spectroscopy (¹³C-²H COSY) permits the identification of ¹³C and ²H nuclei which are connected to one another by a single chemical bond via the sizeable ¹J_CD coupling constant. The practical development of this technique is described using a ¹³C-²H COSY pulse sequence which is derived from the classical ¹³C-¹H correlation experiment. An example is given of the application of ¹³C-²H COSY to the study of the biogenesis of natural products from the anti-malarial plant Artemisia annua, using a doubly-labelled precursor molecule. Although the biogenesis of artemisinin, the anti-malarial principle from this species, has been extensively studied over the past twenty years there is still no consensus as to the true biosynthetic route to this important natural product – indeed, some published experimental results are directly contradictory. One possible reason for this confusion may be the ease with which some of the metabolites from A. annua undergo spontaneous autoxidation, as exemplified by our recent in vitro studies of the spontaneous autoxidation of dihydroartemisinic acid, and the application of ¹³C-²H COSY to this biosynthetic problem has been important in helping to mitigate against such processes. In this in vivo application of ¹³C-²H COSY, [15-¹³C²H₃]-dihydroartemisinic acid (the doubly-labelled analogue of the natural product from this species which was obtained through synthesis) was fed to A. annua plants and was shown to be converted into several natural products which have been described previously, including artemisinin. It is proposed that all of these transformations occurred via a tertiary hydroperoxide intermediate, which is derived from dihyroartemisinic acid. This intermediate was observed directly in this feeding experiment by the ¹³C-²H COSY technique; its observation by more traditional procedures (e.g., chromatographic separation, followed by spectroscopic analysis of the purified product) would have been difficult owing to the instability of the hydroperoxide group (as had been established previously by our in vitro studies of the spontaneous autoxidation of dihydroartemisinic acid). This same hydroperoxide has been reported as the initial product of the spontaneous autoxidation of dihydroartemisinic acid in our previous in vitro studies. Its observation in this feeding experiment by the ¹³C-²H COSY technique, a procedure which requires the minimum of sample manipulation in order to achieve a reliable identification of metabolites (based on both ¹³C and ²H chemical shifts at the 15-position), provides the best possible evidence for its status as a genuine biosynthetic intermediate, rather than merely as an artifact of the experimental procedure.     

Refinement of protein structure against non-redundant carbonyl <Superscript>13</Superscript>C NMR relaxation

Carbonyl ¹³C′ relaxation is dominated by the contribution from the ¹³C′ chemical shift anisotropy (CSA). The relaxation rates provide useful and non-redundant structural information in addition to dynamic parameters. It is straightforward to acquire, and offers complimentary structural information to the ¹⁵N relaxation data. Furthermore, the non-axial nature of the ¹³C′ CSA tensor results in a T₁/T₂ value that depends on an additional angular variable even when the diffusion tensor of the protein molecule is axially symmetric. This dependence on an extra degree of freedom provides new geometrical information that is not available from the NH dipolar relaxation. A protocol that incorporates such structural restraints into NMR structure calculation was developed within the program Xplor-NIH. Its application was illustrated with the yeast Fis1 NMR structure. Refinement against the ¹³C′ T₁/T₂ improved the overall quality of the structure, as evaluated by cross-validation against the residual dipolar coupling as well as the ¹⁵N relaxation data. In addition, possible variations of the CSA tensor were addressed. Electronic Supplementary Material The online version of this article (doi:) contains supplementary material, which is available to authorized users.