Dynamics of Zika Virus Capsid Protein in Solution - ACS Publications

Apr 29, 2019 - Institute of Medical Biochemistry Leopoldo De Meis, Program of ... Centro de Desenvolvimento de Tecnologia em Saúde, Fiocruz, Rio de ...
0 downloads 0 Views 5MB Size
Download PDF
Article Cite This: Biochemistry 2019, 58, 2488−2498

pubs.acs.org/biochemistry

Dynamics of Zika Virus Capsid Protein in Solution: The Properties and Exposure of the Hydrophobic Cleft Are Controlled by the α‑Helix 1 Sequence Maria A. Morando,†,‡,§ Glauce M. Barbosa,† Christine Cruz-Oliveira,†,‡ Andrea T. Da Poian,*,† and Fabio C. L. Almeida*,†,‡

Downloaded via UNIV OF SOUTHERN INDIANA on July 17, 2019 at 14:17:15 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

†

Institute of Medical Biochemistry Leopoldo De Meis, Program of Structural Biology, Federal University of Rio de Janeiro, Rio de Janeiro 21941-902, Brazil ‡ National Center for Structural Biology and Bioimaging (CENABIO)/National Center for Nuclear Magnetic Resonance (CNRMN), Federal University of Rio de Janeiro, Rio de Janeiro 21941-902, Brazil § Centro de Desenvolvimento de Tecnologia em Saúde, Fiocruz, Rio de Janeiro 21040-361, Brazil S Supporting Information *

ABSTRACT: Zika virus (ZIKV) became an important public health concern because infection was correlated to the development of microcephaly and other neurological disorders. Although the structure of the virion has been determined by cryo-electron microscopy, information about the nucleocapsid is lacking. We used nuclear magnetic resonance to determine the solution structure and dynamics of full length ZIKV capsid protein (ZIKVC). Although most of the protein is structured as described for the capsid proteins of Dengue and West Nile viruses and for truncated ZIKVC (residues 23−98), here we show important differences in the α-helix 1 and N-terminal intrinsically disordered region (IDR). We distinguished two dynamical regions in the ZIKVC IDR, a highly flexible N-terminal end and a transitional disordered region, indicating that it contains ordered segments rather than being completely flexible. The unique size and orientation of α-helix 1 partially occlude the protein hydrophobic cleft. Measurements of the dynamics of α-helix 1, surface exposure, and thermal susceptibility of each backbone amide 1H in protein structure revealed the occlusion of the hydrophobic cleft by α1/α1′ and supported α-helix 1 positional uncertainty. On the basis of the findings described here, we propose that the dynamics of ZIKVC structural elements responds to a structure-driven regulation of interaction of the protein with intracellular hydrophobic interfaces, which would have an impact on the switches that are necessary for nucleocapsid assembly. Subtle differences in the sequence of α-helix 1 have an impact on its size and orientation and on the degree of exposure of the hydrophobic cleft, suggesting that α-helix 1 is a hot spot for evolutionary adaptation of the capsid proteins of flaviviruses.

Z

ZIKVC shows a dimeric globular core formed by eight intertwined α-helices (four per monomer) and an N-terminal intrinsically disordered region (IDR). The capsid proteins of flaviviruses are highly basic, but the charges are not symmetrically distributed on the protein surface. In one face, they display a hydrophobic cleft, shown to mediate binding of protein to intracellular hydrophobic interfaces, such as endoplasmic reticulum and/or lipid droplets, during the virus replication cycle.12,13 The opposite face is a positively charged patch, which is a possible nucleic acid binding site. For DENVC, the N-terminal IDR is also involved in association of protein with lipid droplets (LDs).14,15 DENVC−LD interaction is essential for infectious particle formation,12 but the

ika virus (ZIKV) is an arbovirus of the Flaviviridae family that recently caused a major outbreak in the Americas.1 ZIKV neurotropism has been demonstrated,2,3 and infection has been correlated to the development of microcephaly and other malformations associated with congenital ZIKV exposure4,5 and to serious neurological disorders, such as Guillain-Barré syndrome, after adult infection.6,7 The structure of ZIKV has been determined by cryoelectron microscopy (cryo-EM),8 revealing structural details of the envelope (E) and the membrane (M) proteins but lacking information about the virus nucleocapsid (NC). The ZIKV NC is formed by the 10.6 kb single-stranded positive sense genomic RNA associated with the 104-amino acid capsid protein (ZIKVC). The structure of a truncated form of ZIKVC (residues 23−98) has been previously determined by X-ray crystallography.9 As found for the described capsid proteins of Dengue virus (DENVC)10 and West Nile virus (WNVC),11 © 2019 American Chemical Society

Received: March 7, 2019 Revised: April 22, 2019 Published: April 29, 2019 2488

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry

NMR Spectroscopy. The NMR sample contained 300 μM (dimer) ZIKVC in 55 mM sodium phosphate buffer, 200 mM NaCl (pH 6.0), and 10% D2O. NMR spectra were acquired at 35 °C on Bruker Avance III 600 MHz and Avance IIIHD 900 MHz spectrometers equipped with 15N/13C/1H tripleresonance probes (Bruker TXI). NMR spectra were processed with NMRPipe19 and analyzed using CCPNmr Analysis software.20 The backbone assignment was obtained using non-uniform sampling (NUS) with multidimensional Poisson Gap scheduling. NMRPipe and iterative soft threshold (hmsIST) fast reconstruction of NMR data were used for processing.21 The protein backbone resonance assignments were achieved through analysis of the triple-resonance experiments: HNCACB, CBCA(CO)NH, HNCACO, and HNCO.22,23 The aliphatic side-chain resonances were primarily assigned using the HBHA(CO)NH24 experiment and with the help of complementary HCCH-TOCSY,25 15Nedited NOESY-HSQC, and 13C-edited NOESY-HSQC26 optimized for both aliphatic or aromatic region detection. Assignments of aromatic side chains were obtained using nuclear Overhauser effects (NOEs) between the aromatic protons and the βCH2 group in the 13C NOESY-HSQC spectrum. We chose to work at 35 °C because of the higherquality spectra, especially in the regions involved in conformational exchange. The data were processed using NMRPipe19 or TopSpin3.1 (Bruker Biospin), while resonance assignments were achieved using CCPNMR Analysis.27 Distance restraints were derived from 15N NOESY-HSQC and two 13C NOESY-HSQC spectra, for aromatics and aliphatics. For NOESY experiments, we used the NUS scheduling to increase the resolution. We ran TALOS-N28 for backbone chemical shift-based dihedral angle prediction. The predicted backbone dihedral angles φ and ψ of the residues involved in secondary structure were used as a restraint for structural calculations. Setup details of the NMR experiments are described in Table S1. Structure Calculation. Structure calculations for ZIKVC were performed iteratively using ARIA version 2.329,30 combined with CNS version 1.2,31 using 15N NOESY-HSQC and 13C NOESY-HSQC data sets as source of distance restraints. All data were converted to xml format file. The backbone φ and ψ dihedral restraints were predicted on the basis of analysis of 1HN, 15N, 1Ha, 13Ca, 13CO, and 13Cb chemical shifts using TALOS-N.28,32 Structures were calculated using CNS 1.2 using torsion angle simulated annealing. Several cycles of ARIA were performed using standard protocols. After each cycle, rejected restraints, side-chain assignments, NOEs, and dihedral violations were analyzed. Finally, 400 conformers were calculated with ARIA/ CNS, and among them, the 20 best water-refined structures with the lowest energy were selected to represent the solution structural ensemble of ZIKVC. The structural ensemble was visualized and analyzed with Chimera and Pymol. Quality validation was performed using PROCHECK33 and the Protein Structure Validation Software suite (PSVS) (http://psvs-1_5-dev.nesg.org/). NMR Relaxation Parameters. 15N backbone amide relaxation parameters ( 15 N R 1 , 15 N R 2 , and 1 H− 15 N heteronuclear NOE) were measured for 15N-labeled ZIKVC [300 μM dimer, in 55 mM sodium phosphate buffer, 200 mM NaCl, and 10% D2O (pH 6.0)] using Bruker Avance IIIHD 500 (11.74 T, operating at 500.13 MHz) and Avance IIIHD

mechanism that drives the transition from the membraneassociated C protein dimers to the assembled NC is unknown. In this work, we used nuclear magnetic resonance (NMR) to determine the structure and dynamics of the full length ZIKVC in solution. We found that most of the protein globular region is structured similarly as described for DENVC and WNVC as well as for the crystal structure of the truncated ZIKVC, but remarkable differences occur for α-helix 1 and the N-terminal IDR. Because the ZIKVC IDR corresponds to approximately one-third of the protein, spanning residues 1−35, we took advantage of a number of NMR approaches to study its dynamical properties. Indeed, a conformational equilibrium can be fully characterized only in solution, and NMR spectroscopy is particularly powerful for characterizing regions involved in dynamics.16,17 We distinguished two dynamical regions in the ZIKVC IDR, a highly flexible N-terminal end and a transitional disordered region, indicating that the ZIKVC IDR contains ordered segments rather than being completely flexible. Measurements of the surface exposure and thermal susceptibility of each backbone amide 1H in the protein structure revealed the occlusion of the hydrophobic cleft by α1/α1′ and supported the α1 positional uncertainty observed in the structural ensemble. On the basis of the findings described here, we propose that the dynamics of ZIKVC structural elements responds to a structure-driven regulation of interaction of protein with intracellular hydrophobic interfaces, which would have an impact on the switches that are necessary for NC assembly. We also suggest that α-helix 1 is a hot spot for evolutionary adaptation of the capsid proteins of flaviviruses.

■

MATERIALS AND METHODS Protein Expression and Purification. ZIKVC protein purification was adapted from the protocols used for DENV2C purification.10,14 The ZIKVC coding sequence (encoding amino acid residues 1−104, deposit AMD16557.1)18 was synthesized by Genscript and codon-optimized for Escherichia coli and subcloned into the pET3a vector using NdeI and BamHI restriction sites. Recombinant protein was expressed in E. coli BL21-DE3-pLysS overnight, at 30 °C, after induction with 0.5 mM IPTG. The uniform 15N-labeled and 15N- and 13 C-labeled proteins were expressed in M9 minimal medium supplemented with 1 g/L [15N]NH4Cl and 3 g/L [13C6]glucose with 100 μg/mL ampicillin and 34 μg/mL chloramphenicol. After protein expression, cells were harvested by centrifugation at ∼5000g for 20 min at 4 °C. The pellet was resuspended in 40 mL of buffer, containing 25 mM 2-[4-(2hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES), 0.5 M NaCl, 1 mM 2,2′,2″,2′″-(ethane-1,2-diyldinitrilo)tetraacetic acid (EDTA), 5% (v/v) glycerol (pH 7.4), and 1 mL of protease inhibitor cocktail (P8465, Sigma-Aldrich). After centrifugation and ultrasonication (model VCX 130, VibraCell), NaCl was added to the cell lysate to achieve a final concentration of 2 M. The lysate was stirred for 1 h at 4 °C. After the precipitation, the lysate was ultracentrifuged at 70400g for 50 min at 4 °C. ZIKVC was soluble in the supernatant and was purified in a HiTrap Heparin HP column coupled to an AKTA Start instrument (GE Healthcare). A step gradient was employed with an increasing NaCl concentration (1.0, 1.5, and 2 M). Fractions containing ZIKVC protein were confirmed by absorbance at 280 nm and 18% sodium dodecyl sulfate−polyacrylamide gel electrophoresis. Fractions were concentrated and stored at −20 °C. 2489

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry 600 (14.09 T, operating at 600.03 MHz) instruments at 35 °C, with 1024 and 96 complex points in F2 (1H) and F1 (15N), respectively. R1 and R2 were collected with 11 accumulations by increment, and 1H−15N NOEs with 16 accumulations by increment. R1 was measured with delays varying from 0.05 to 1 s. R2 was measured with delays varying from 17 to 170 ms. The experimental error was evaluated from the signal-to-noise ratio of the spectra displaying ∼50% of the decay in signal. The 1 H−15N NOEs were acquired with or without proton saturation for 8.0 s. The R1 and R2 values were obtained using the relaxation module of CCPNMR.20 The 1H−15N heteronuclear NOE values were determined using the ratio of the intensity with saturation spectra to the intensity without saturation spectra. The data could not be fitted in a Lipari− Szabo model free formalism34 because of the lack of a hydrodynamic model of the protein due of the long IDR. Thus, the raw data were interpreted along with the 15N CPMG relaxation dispersion experiments (CPMG-RD), also collected at two fields and two temperatures (Figure 2). 15 N Relaxation Dispersion Measurements. 15N Carr− Purcell−Meiboom−Gill relaxation dispersion (CPMG-RD) profiles for a 15N-labeled ZIKVC [300 μM dimer, in 55 mM sodium phosphate buffer, 200 mM NaCl, and 10% D2O (pH 6.0)] were recorded using Bruker Avance IIIHD 500 (11.74 T, operating at 500.13 MHz) and Avance IIIHD 600 (14.09 T, operating at 600.13 MHz) instruments, at two temperatures (30 and 35 °C), using a constant relaxation time of 30 ms in a 15 N CPMG relaxation compensation, comprising two CPMG elements of 15 ms,35 resulting in the transverse relaxation of a mix of in-phase (Nx,y) and antiphase (2Nx,yHz) coherences (R2eff = 0.5R2effin‑phase + 0.5R2effantiphase). All 15N CPMG-RD experiments were performed with a constant relaxation time period Trelax of 30 ms and with CPMG frequencies (νCPMG) ranging from 66 to 1000 Hz. CPMG-RD R2eff(νCPMG) values were calculated from peak intensities (I) in a series of twodimensional (2D) 1H−15N correlation spectra recorded in an interleaved way at different CPMG frequencies (νCPMG), using the equation R 2eff(νCPMG) = −1/Trelax ln(I /I0), where I is the signal intensity in the spectra collected at Trelax = 30 ms and I0 is the signal intensity in the reference spectrum recorded at Trelax = 0. The experimental error in R2eff rates was estimated as the signal-to-noise ratio for each resonance via the equation ΔR 2eff(νCPMG) = 1/Trelax[1/(signal/noise)] Amide Chemical Shift Temperature Coefficient. The amide 1HN chemical shift temperature coefficient of ZIKVC [200 μM dimer, in 55 mM sodium phosphate buffer, 200 mM NaCl, and 10% D2O (pH 6.0)] was determined by recording a series of two-dimensional 15N/1H HSQC spectra at 20, 25, 30, 35, and 40 °C, using a Bruker Avance 600 spectrometer (14.09 T, operating at 600.13 MHz). All of the spectra were referenced to the water signal for each temperature, processed in NMRPipe, and analyzed in CcpNmr analysis software. The water chemical shift was referenced using 3-(trimethylsilyl)propane-1-sulfonic acid (DSS). The chemical shift values (δHN) of all residues at different temperature were plotted as a function of temperature. The resulting slope (dδHN/dT) of every curve was plotted for each residue. Solvent Paramagnetic Relaxation Enhancement (sPRE). Amide 1HN/15N cross-peak intensities for each residue of ZIKVC [200 μM dimer, in 55 mM sodium phosphate buffer, 200 mM NaCl, and 10% D2O (pH 6.0)] were determined by recording a series of two-dimensional

15

N/ 1H HSQC spectra at different concentrations of gadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA) (0, 0.5, 1.0, 2.0, 3.0, and 4.5 mM), at 35 °C, using a Bruker Avance 600 MHz spectrometer (operating at 14.09 T). All of the spectra were processed in NMRPipe and analyzed in CCPNmr analysis software. The intensities of each residue 1 HN/15N cross-peak (INH) were plotted as a function of the concentrations of Gd-DTPA, and the resulting slopes (INH[Gd]−1) were plotted for each residue. To calculate an accurate I NH [Gd] −1 , the points where the intensities approached zero (lack of linearity) were not taken into consideration, according to Figure S5. Flavivirus Sequence Alignment and Sequence Clustering Analysis. Protein−protein BLAST (BLASTP) using the ZIKVC as a query was run over a database of nonredundant protein sequences of viruses from the Flaviviridae family (Taxid 11050). From the 20000 BLASTP best hits (scored using the matrix BLOSUM62), a representative sequence of the complete polyprotein of each flavivirus was selected, resulting in the following sequence IDs: AMD16557.1, YP_009227185.1, AFR66758.1, AEA72437.1, AFN43044.1, AER25364.2, ALE71321.1, AEV41145.1, ANK79133.1, AFP95929.1, AEI27244.1, AGV15509.1, AIU94743.1, AAV54504.1, ALK02500.1, ABW76844.2, AAX82481.1, NP_051124.1, AJE59927.1, YP_164264.1, YP_009350103.1, YP_009126874.1, ARR96288.1, AIU94744.1, AIJ19434.1, AIJ19433.1, AAV34157.1, YP_00932 8360.1 , AGJ840 83.1, YP_ 001040 004.1 , YP_002790882.1, AFK83757.1, ABW82078.1, AGE13481.1, and ACW82911.1. Multiple-sequence alignments of these sequences were carried out using Clustalw 2.1 (http://www. genome.jp/tools-bin/clustalw),36 with Blosum 80, 62, 45, and 30 used as the matrices. A dendogram was constructed using the simplified neighbor-joining method. The secondary structures of capsid proteins were predicted using Jpred4 at the server http://www.compbio.dundee.ac.uk/ jpred/.37

■

REMARKS ON THE INTERPRETATION OF DATA Interpretation of the Relaxation Dispersion Parameters. CPMG-RD experiments were collected at two temperatures, 30 and 35 °C. The CPMG-RD profiles did not show a typical dispersion of an intermediate exchange regime, with R2eff decreasing toward R2eff free of exchange (R2eff∞) in the range of νcp varying from 66 to 1000 s−1, as illustrated in Figures S3B and S4B. Rather, we observed an approximately constant of R2eff. For most residues, R2eff is approximately R2eff∞, as illustrated in Figures S3C,D and S4C,D. For a few residues, those illustrated in Figures S3E and S4E, R2eff > R2eff∞. These relaxation dispersion profiles are typical of conformational exchange in a fast exchange regime (Figures S3 and S4). At this exchange regime, it is not possible to obtain R2eff∞ from the CPMG-RD profile. To analyze and plot the data in Figure 2, we estimated R2eff∞ on the basis of the expected overall behavior of R2eff of a dynamic region. The values interpreted as R2eff∞ are traced as a dotted line in Figures S3A and S4A. Only the values of R2eff that significantly exceeded the dotted lines (R2eff∞) were considered as a 15NH in conformational exchange in a fast exchange regime (microsecond time scale). Interpretation of the Amide Chemical Shift Temperature Coefficient. The amide hydrogen (HN) chemical shift 2490

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry temperature coefficient (dδHN/dT) reports on the thermal susceptibility of the HN−C′ amide hydrogen bonds. dδHN/dT is correlated to the length of the hydrogen bond (rHNC′) and hydrogen bond J coupling 3hJNC′.38 In proteins, rHNC′ increases with temperature due to the increase in the extent of thermal fluctuation.39,40 An empirical simple rule stands that dδHN/dT > −5 ppb/K for HNs involved in secondary structure41 (low thermal susceptibility), and residues hydrogen bonded with water present dδ HN /dT < −5 ppb/K (high thermal susceptibility). The accepted reasoning is that the hydrogen bonding with water is weaker and more expandable than the HN−C′ bonding.40 However, dδHN/dT is considered to be poorer than HN exchange rates for depicting internal hydrogen bonds, and therefore, residues are frequently found to be involved in secondary structure with dδHN/dT < −5 ppb/ K.38,42 The reason is that dδHN/dT senses local fluctuation, providing thermodynamic information about the local environment.43 The chemical shift is a function of the local environment for each nucleus and consequently reflects all of the degrees of freedom in which one nucleus is experimenting. For this reason, the thermal coefficient is fundamentally related to the local entropy.38 Residues with high amide chemical shift thermal susceptibility (dδHN/dT < −5 ppb/K) make more expandable hydrogen bonds, which are weaker, presenting smaller hydrogen bond J coupling 3hJNC′.38 For this, an amide with dδHN/dT < −5 ppb/K may be considered a point of break of a secondary structure or, when in a loop or IDR, more exposed to a hydrogen bond with water. Residues with low amide chemical shift thermal susceptibility (dδHN/dT > −5 ppb/K) make less expandable hydrogen bonds, which are stronger, presenting larger hydrogen bond J coupling 3hJNC′.38 For this, an amide with dδHN/dT > −5 ppb/K is involved in secondary structures of hydrogen bonds. When in a loop or IDR, the amide may tend to form intramolecular hydrogen bonds and, thus, to increase order.

Table 1. Structural Statistics of ZIKVC no. of restraints total distance restraints intraresidue sequential (i, i + 1) medium-range (i, i + 2, i + 3) long-range (>i, i + 3) intersubunit intrasubunit total no. of dihedral angle restraints total no. of hydrogen bond restraints no. of violations distance restraints >0.5 Å dihedral restraints >5° hydrogen bond restraints >0.5 Å energy geometry

1883 (unambiguous) 645 564 304 150 220 102 74 0 0 0 −5112.83 ± 15 × 10−1 kcal mol−1 38.64 ± 11 kcal mol−1

distance restraints root-mean-square deviation (rmsd) in NOE restraints and idealized geometry NOE (Å) 1.80 ± bond lengths (Å) 2.86 ± bond angles (deg) 0.46 ± improper dihedral angles (deg) 1.21 ± Coordinate rmsd from Average Structure all heavy atoms (Å) all, 35−104 35−97 44−97 (α2−α4)

■

RESULTS ZIKVC resonance assignment (Figure S1) and structure determination (Table 1 and Figure 1) were performed using 13 C- and 15N-labeled protein at pH 6.0. The ZIKVC structure shows a larger IDR when compared to those of DENVC and WNVC structures.10,11 α-Helix 1 (α1) spans residues 36−39, which makes the IDR at least nine residues longer (residues 1−35). It is noteworthy that α1 shows a unique size and orientation, with α1/α1′ partially occluding the α2/α2′ hydrophobic cleft (Figure 1A). Among the four α-helices, α2/α2′, α3/α3′, and α4/α4′ form the intertwined dimeric structure dominated by quaternary contacts, while α1/α1′ displays a small positional uncertainty (Figure 1A,B). The presence of three π-stacking interactions among the conserved aromatic residues (Figure 1C,D) and six salt bridges (Table 2 and Figure 1E,F) is remarkable, two more than in DENVC and six more than in WNVC. The interchain salt bridge between R55 and D87′ (and R55′ and D87) seems to be the most important for structural stabilization. Conversely, in the crystal structure of truncated ZIKVC, a rotameric conformation of the R55 side chain makes the formation of this salt bridge unfavorable (Table 2). A salt bridge between K75 and E79, which is important for stabilizing α-helix 4, is observed only in the ZIKVC solution structure described here. The lack of a K75/E79 salt bridge in the crystal structure could be explained by the fact that this region is involved in crystal contacts.

backbone atoms (Å)

2.4 1.27 0.92 PROCHECK Ramachandran analysis

most favored regions (%) additional allowed regions (%) generously allowed regions (%) disallowed regions (%)

2.7 1.5 1.9 8.0

1.69 0.91 0.4

secondary structure

all residues

98.7 1.3 0 0

90 7.5 1.4 1.1

To describe the plasticity of the ZIKVC structure, especially in what refers to α1/α1′ and IDR structural behaviors, we analyzed its dynamics in solution. We measured the 15N relaxation parameters (Figure S2), which is typical of a dimer and allowed us to distinguish different dynamic regions: (i) a highly flexible disordered region observed for residues 1−23, (ii) a transition disordered region from residue 24 to 35, (iii) a globular ordered region from residue 36 to 98, and (iv) a flexible C-terminus (residues 99−104). We observed conformational exchange (increase in R2/R1), especially in α1 at residue G36, and in the loop between α1 and α2 (loop 1−2) at residues G40 and G42. CPMG-RD experiments (Figure 2 and Figures S3 and S4) revealed conformational exchange in a fast regime [microsecond time scale motion (Figures S3B,E and S4B,E)]. We also observed conformational exchange at the IDR, indicating that some order occurs in this region. Five residues (N3, S8, G9, G10, and N15) located in its highly flexible segment and two residues (G28 and K31) in the transition region showed increased R2eff values. In the globular domain, the most evident dynamics in the microsecond time scale occurred at residue G36 of α1 and residues G40 and G42 of loop 1−2, supporting the positional uncertainty of α1/α1′ observed in the structural ensemble. It is important to mention that, despite the conformational exchange observed for α1 and the adjacent loops, we did not observe disorder for α1. 2491

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry

Figure 1. Solution structure of the ZIKVC globular domain. (A and B) Superposition of the three low-energy structures showing the positional uncertainty of α-helix 1, in (A) frontal view and (B) top view. (C and D) Ribbon representation of the backbone of the ZIKVC showing the side chains of aromatic residues involved in the π-stacking interaction, (C) in frontal view, with the black line representing the symmetry axis, and (D) in lateral view. (E) Frontal and (F) lateral views showing the localization of the salt bridges of ZIKVC, as detailed in Table 2. Negatively charged residues are colored red, and positively charged residues blue. Salt bridges R55/D87′ and R55′/D87 connect α-helices 2 and 4 of different subunits, contributing to the stabilization of the quaternary structure. Salt bridges R45′/E76′ and R45/E76 connect α-helices 2 and 4 of the same subunit, contributing to the stabilization of the tertiary structure. Salt bridges K75′/E79′ and K75/E79 stabilize α-helix 4.

Relaxation parameters that report on high flexibility (R1, R2, and 1H−15N heteronuclear NOE) are similar for α1−α4. Similar behavior was observed for the random coil index,32 which is on average 0.81, 0.87, 0.85, and 0.87 for α1−α4, respectively. To confirm the occlusion of the hydrophobic cleft by α1/ α1′ and the occurrence of some order in the IDR, we measured the solvent exposure of each backbone amide by solvent paramagnetic relaxation enhancement (sPRE) (Figure 3A−C and Figure S5). The globular region showed the smallest sPRE effect, with the C-terminal portion of α2/α2′ (50-ILAFLRF-56) being the least exposed region. This observation supports the idea that α1′ and the IDR occlude α2, while α1 and the IDR occlude α2′ (Figure 3B,C). This partial occlusion of the hydrophobic cleft by α1/α1′ may have important biological consequences, because binding of C

protein to hydrophobic interfaces seems to mediate NC assembly.14,15 As expected, the C-terminus and the highly flexible region of the IDR (R12, V21, and A22) were the most exposed regions. In agreement with the 15N CPMG-RD experiments, the most protected residues of the IDR were the residues in conformational exchange, N3, S8, G9, G10, and N15, in the highly flexible segment, and G28 and K31, in the transition region, corroborating the existence of some degree of order in the ZIKVC IDR. We also measured the thermal susceptibility of the amide (HN) chemical shifts (dδHN/dT) (Figure 3D−F and Figure S6; see Materials and Methods for details about the interpretation of this parameter). dδHN/dT provided extra information about the relative order of the IDR and specific features of the globular region. Residues G10, G20, G28, G29, 2492

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry

bonds with water are expected for a fully disordered polypeptide chain. The thermal susceptibility measurements also enabled us to explain the smaller size of ZIKVC α-helix 1 (Figure 4). Residues K31 and L33 show weaker hydrogen bonds (dδHN/ dT < −5 ppb/K), behaving as secondary structure breakers. The P34 in the vicinity of these residues contributes to the loss of helical structure in this α-helix 1 region. With these results, we were able to map the key amino acid residues that make ZIKVC α-helix 1 different from those of the C proteins of the other flaviviruses. They also explain the unique features of ZIKVC α-helix 1, which may be the driving force for hydrophobic cleft occlusion. It is remarkable that the presence of K31, L33, and P34 as helix breakers induced a smaller αhelix 1, changing its orientation and the degree of exposure of the hydrophobic cleft. It also led to the increased IDR. In the globular region, most of the residues displaying dδHN/ dT values of greater than −5 ppb/K are in secondary structure elements (Figure 3D−F). The exceptions A49, R93, and N96 (dδHN/dT < −5 ppb/K) reflect weakness in the hydrogen bonding of α-helices. Residue G42, at loop 1−2, and residue S62, at loop 2−3, are strong secondary structure breakers (dδHN/dT < −5 ppb/K).

Table 2. Salt Bridges in ZIKV, DENV2, and WNV Capsid Proteins44 Solution ZIKVC Structure salt bridge

solvent exposure (%)

minimum/maximum distances (Å)

R45′/ 55.1/46.1 2.7/10.1 E76′ R45/E76 55.1/46.1 2.7/10.1 R55′/ 44.5/45.3 2.6/7.3 D87 R55/ 44.5/45.3 2.6/7.3 D87′ K75′/ 53.3/46.6 2.7/6.0 E79′ K75/E79 53.3/46.6 2.7/6.0 salt bridge solvent exposure (%)

average distance (Å) 5.39 5.39 4.1 4.1 4.67 4.67 distance (Å)

Crystal ZIKVC Structure (PDB entry 5YGH) R45′/E76′ 46.2/33.5 R45/E76 46.2/33.5 R55′/D87 58.9/27.2 R55/D87′ 58.9/27.2 K75′/E79′ 49.7/40.0 K75/E79 49.7/40.0 DENV2C K45′/E87′ 20.6/26.2 K45/E87 20.6/26.2 R55′/E87 40.2/26.2 R55/E87′ 40.2/26.2 WNVC none

10.9 11.3 6.1 6.3 8.4 11

■

3.7 3.7 2.8 2.8

DISCUSSION Several structural features of ZIKVC demonstrated here, namely, (i) its larger IDR displaying some order in a specific segment, (ii) the unique size and orientation of its α1/α1′, (iii) its microsecond dynamics, and (iv) the partial occlusion of its hydrophobic cleft, support structure-driven regulation of the interaction of ZIKVC with intracellular hydrophobic interfaces, which would be essential to the structural switches necessary for NC assembly. On the basis of these findings, we hypothesized that α1 sequence is a hot spot for evolutionary pressure. To explore this hypothesis, we analyzed the sequence similarity among the 20000 ZIKVC BLASTP best hits. With

L30, and R32 show dδHN/dT values of greater than −5 ppb/K, displaying less expandable hydrogen bonds (stronger), typical of HNs in secondary structure. This further supports an order in the IDR with these amides more likely to make intramolecular hydrogen bonds. Intermolecular hydrogen

Figure 2. Summary of the 15N CPMG-RD experiments. (A and D) R2eff as a function of residue number, acquired at 35 °C and 11.74 and 14.09 T, respectively. (B and E) R2eff as a function of residue number, acquired at 30 °C and 11.74 and 14.09 T, respectively. (C and F) 15N R2/R1 as a function of residue number acquired at 35 °C and 11.74 T (Figure S2 shows the complete set of relaxation parameters of ZIKVC). The labels in panels A, B, D, and E highlight residues in conformational exchange, according to the analyses shown in Figures S3 and S4. Asterisks denote overlapped residues, and plus signs prolines. 2493

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry

Figure 3. Surface exposure and thermal susceptibility of the HN. (A) NH surface exposure measured (INH[Gd]−1) as a function of residue number (data were derived from Gd titration curves shown in Figure S5). (B and C) Worm representations of ZIKVC (frontal and top views, respectively). The thickness and colors represent the degree of NH exposure, with the weakly exposed residues colored blue (thin) and the strongly exposed residues colored red (thick). The region of 50-ILAFLRF-56 is the most protected from solvent exposure. (D) Temperature dependence of the HN chemical shift (dδHN/dT) as a function of residue number. (E and F) Worm representations of ZIKVC (frontal and top views, respectively). The thickness and colors correlate to the dδHN/dT, with the low-thermal susceptibility residues colored blue (thin) and the high-thermal susceptibility residues colored red (thick). Colorless residues represent a lack of information. Asterisks denote overlapped residues, and plus signs prolines.

DENVC, and WNVC showed important differences for α1 (Figure 4). Remarkably, the predicted α1 region is conserved among all hits. ZIKVC α1 is shorter than the predicted region, with only the C-terminal segment of the predicted portion being actually helical, while DENVC α1, although also shorter than predicted, shows only the N-terminal segment of the predicted portion in a helical structure. For WNVC, the structure of α1 is very similar to that predicted. Accordingly, sequence clustering separates ZIKVC, DENVC, and WNVC in different clades (Figure 5). For ZIKVC, the solution and the crystal structures are very similar with respect to their protein α-helix 2−α-helix 4 segments. α-Helix 1 is similar in size (36-GLLL-39), but we observed important differences in the position and orientation of α-helix 1 (Figure 4D). These differences may be explained by the fact that the crystallographic arrangement of the truncated ZIKVC consists of hexamers of three dimers (Figure

this strategy, we could compare the ZIKVC sequence only with the flaviviruses that are close in evolution and structurally similar with ZIKV, leaving out hepatitis C or yellow fever capsid proteins, which are more distant in evolution. Interestingly, α2 and the 14-VNMLKR-19 motif (located in the IDR) were the regions showing the highest degree of sequence similarity (Figure S7). This is in agreement with the important role of the interactions between capsid proteins and intracellular membranes during the replication cycle of flaviviruses, because both regions were shown to mediate interaction of protein with lipid surfaces in DENVC.14,15 On the other hand, we observed only 17% similarity among α1 sequences, which indicates that this region is the most susceptible to acquiring sequence modifications that would be reflected in capsid protein structure. Indeed, the comparison between the Jpred predicted secondary structures and the experimentally determined structures of ZIKVC, 2494

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry

Figure 4. Summary of the sequences of different flaviviruses and structural analyses of ZIKVC, DENVC, and WNVC. (A) Secondary structure prediction using Jpred4 and secondary structure information obtained from the determined structures. Residues in helices are colored red. The virus names are color-coded according to the sequence clustering shown in Figure 5. (B) Superposition of ZIKVC and DENVC structures. (C) Superposition of ZIKVC and WNVC structures. Loop 1−2 is colored red, and α1 blue (ZIKVC) or cyan (DENVC or WNVC). (D) Superposition of NMR (presented here) and crystal (PDB entry 5ygh9) structures of ZIKVC. Loop 1−2 is colored red, and α1 blue (ZIKVC) or cyan (DENVC, WNVC, or truncated ZIKVC).

S8) in which the IDR region of residues 26−35 and α-helix 1 are involved in a network of crystal contacts. Thus, the apparent contradictory findings might reflect the importance of α-helix 1 plasticity to the recognition of hydrophobic surfaces.

While the NMR structure describes better ZIKVC structure and dynamics in solution, the crystal structure might be informative with respect to a conformational state that mimics interactions of protein with ligands. 2495

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry

Figure 5. Evolutionary history of the capsid proteins. The sequence clustering was inferred by the simplified neighbor-joining method, revealing three main clusters (clades): (i) ZIKV and Spondweni virus (blue), (ii) DENV (red), and (iii) a cluster that contains WNV (black).

Accession Codes

These observations have important evolutionary consequences. They show that the size and orientation of α1/α1′ may be controlled by the sequence (Figure 4), meaning that the capacity of recognizing hydrophobic interfaces would have been modified in evolution through changes in the α1 sequence. Capsid proteins of ZIKV and Spondweni virus have conserved helix breakers in the α1 N-terminal portion: K31 and L33 [dδHN/dT ≪ −5 ppb/K for ZIKVC (Figure 3D)] and P34. On the other hand, G28 and K31 are in conformational exchange and prone to switching conformations, which could favor the formation of an extended α1, as predicted. Thus, these features would control the exposure of the hydrophobic cleft to putative ligands. In DENVC, the orientation of α1 is lateral, making the hydrophobic cleft more exposed (Figure 4B), while the α1 orientation in ZIKVC and WNVC is similar (Figure 4C), both partially occluding the hydrophobic cleft. Additionally, the size of the IDR would also be modified due to the α1 feature. IDRs of capsid proteins of flaviviruses display many conserved Lys and Arg residues: six Lys and four Arg for ZIKVC and four Lys and four Arg for DENVC. Functional analysis of DENVC IDR suggests its participation in RNA recognition,45 and its size would contribute to the acceleration of the identification of the nucleic acid target site,46−48 making the intersegmental transfer between different regions of the genome more efficient.

■

The atomic coordinates, experimental restraints, and chemical shift assignments are available in the Protein Data Bank (PDB entry 6C44) and Biomagnetic Resonance Data Bank (BMRB entry 30397), and the sequence is available from NCBI/ GenBank AMD16557.1.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Fabio C. L. Almeida: 0000-0001-6046-7006 Author Contributions

M.A.M. is a structural biologist and NMR spectroscopist, who contributed with the sample preparation, experimental design, resonance assignments, NMR data collection, data analysis, and structure calculation. G.M.B. contributed to sample preparation, experimental design, and NMR data collection. C.C.-O. contributed to sample preparation, experimental design, and NMR data collection. A.T.D.P. is a virologist, an expert in Flavivirus, who contributed to the data analysis, experimental design, and funding. F.C.L.A. is a structural biologist and NMR spectroscopist, who contributed to the structure and dynamics of ZIKVC, data analysis, experimental design, and funding. All authors contributed to the preparation of the manuscript.

ASSOCIATED CONTENT

* Supporting Information S

Funding

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.9b00194.

The work presented here was funded by FAPERJ Grants 215141, 210361, 239229, and 204432, awarded to F.C.L.A., FAPERJ Grants 201167 and 201316, awarded to A.T.D.P., CNPq Grants 309564/2017-4 and 457773/2014-6, awarded to

Table S1 and Figures S1−S8 (PDF) 2496

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry

(4) Mlakar, J., Korva, M., Tul, N., Popović, M., Poljšak-Prijatelj, M., Mraz, J., Kolenc, M., Resman Rus, K., Vesnaver Vipotnik, T., Fabjan Vodušek, V., Vizjak, A., Pižem, J., Petrovec, M., and Avšič Ž upanc, T. (2016) Zika Virus Associated with Microcephaly. N. Engl. J. Med. 374, 951−958. (5) Calvet, G., Aguiar, R. S., Melo, A. S. O., Sampaio, S. A., De Filippis, I., Fabri, A., Araujo, E. S. M., De Sequeira, P. C., De Mendonça, M. C. L., De Oliveira, L., Tschoeke, D. A., Schrago, C. G., Thompson, F. L., Brasil, P., Dos Santos, F. B., Nogueira, R. M. R., Tanuri, A., and De Filippis, A. M. B. (2016) Detection and sequencing of Zika virus from amniotic fluid of fetuses with microcephaly in Brazil: a case study. Lancet Infect. Dis. 16, 653−660. (6) Parra, B., Lizarazo, J., Jiménez-Arango, J. A., Zea-Vera, A. F., González-Manrique, G., Vargas, J., Angarita, J. A., Zuñiga, G., LopezGonzalez, R., Beltran, C. L., Rizcala, K. H., Morales, M. T., Pacheco, O., Ospina, M. L., Kumar, A., Cornblath, D. R., Muñoz, L. S., Osorio, L., Barreras, P., and Pardo, C. A. (2016) Guillain-Barré Syndrome Associated with Zika Virus Infection in Colombia. N. Engl. J. Med. 375, 1513−1523. (7) Soares, C. N., Brasil, P., Carrera, R. M., Sequeira, P., De Filippis, A. B., Borges, V. A., Theophilo, F., Ellul, M. A., and Solomon, T. (2016) Fatal encephalitis associated with Zika virus infection in an adult. J. Clin. Virol. 83, 63−65. (8) Sirohi, D., Chen, Z., Sun, L., Klose, T., Pierson, T. C., Rossmann, M. G., and Kuhn, R. J. (2016) The 3. 8 Å resolution cryo-EM structure of Zika virus. Science 352, 467−470. (9) Shang, Z., Song, H., Shi, Y., Qi, J., and Gao, G. F. (2018) Crystal Structure of the Capsid Protein from Zika Virus. J. Mol. Biol. 430, 948−962. (10) Ma, L., Jones, C. T., Groesch, T. D., Kuhn, R. J., and Post, C. B. (2004) Solution structure of dengue virus capsid protein reveals another fold. Proc. Natl. Acad. Sci. U. S. A. 101, 3414−9. (11) Dokland, T., Walsh, M., Mackenzie, J. M., Khromykh, A. A., Ee, K.-H., and Wang, S. (2004) West Nile virus core protein; tetramer structure and ribbon formation. Structure 12, 1157−63. (12) Samsa, M. M., Mondotte, J. A., Iglesias, N. G., Assunçaõ Miranda, I., Barbosa-Lima, G., Da Poian, A. T., Bozza, P. T., and Gamarnik, A. V. (2009) Dengue virus capsid protein usurps lipid droplets for viral particle formation. PLoS Pathog. 5, No. e1000632. (13) Byk, L. A., and Gamarnik, A. V. (2016) Properties and Functions of the Dengue Virus Capsid Protein. Annu. Rev. Virol. 3, 263−281. (14) Martins, I. C., Gomes-Neto, F., Faustino, A. F., Carvalho, F. A., Carneiro, F. A., Bozza, P. T., Mohana-Borges, R., Castanho, M. A., Almeida, F. C. L., Santos, N. C., and Da Poian, A. T. (2012) The disordered N-terminal region of dengue virus capsid protein contains a lipid droplet-binding motif. Biochem. J. 444, 405−415. (15) Faustino, A. F., Guerra, G. M., Huber, R. G., Hollmann, A., Domingues, M. M., Barbosa, G. M., Enguita, F. J., Bond, P. J., Castanho, M. A. R. B., Da Poian, A. T., Almeida, F. C. L., Santos, N. C., and Martins, I. C. (2015) Understanding Dengue Virus Capsid Protein Disordered N-Terminus and pep14−23-Based Inhibition. ACS Chem. Biol. 10, 517−526. (16) Valente, A. P., Miyamoto, C. A., and Almeida, F. C. L. (2006) Implications of protein conformational diversity for binding and development of new biological active compounds. Curr. Med. Chem. 13, 3697−3703. (17) Neudecker, P., Lundström, P., and Kay, L. E. (2009) Relaxation dispersion NMR spectroscopy as a tool for detailed studies of protein folding. Biophys. J. 96, 2045−54. (18) Calvet, G., Aguiar, R. S., Melo, A. S. O., Sampaio, S. A., De Filippis, I., Fabri, A., Araujo, E. S. M., De Sequeira, P. C., De Mendonça, M. C. L., De Oliveira, L., Tschoeke, D. A., Schrago, C. G., Thompson, F. L., Brasil, P., Dos Santos, F. B., Nogueira, R. M. R., Tanuri, A., and De Filippis, A. M. B. (2016) Detection and sequencing of Zika virus from amniotic fluid of fetuses with microcephaly in Brazil: a case study. Lancet Infect. Dis. 16, 653−660.

F.C.L.A., CNPq Grant 306669/2013-7, awarded to A.T.D.P., and FINEP Grant 0267/16, awarded to A.T.D.P. and F.C.L.A. M.A.M. is funded by a Capes-CDTS-Fiocruz visiting research scholarship (001/2012). C.C.-O. is funded by a FAPERJ Senior Postdoc (PDS) scholarship (201.849/2017). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Prof. Francisco Prosdocimi, from the Institute of Medical Biochemistry of the Federal University of Rio de Janeiro, for discussion regarding the sequence clustering and phylogenetic history of Flavivirus. The authors also thank ́ Trambaioli, from the School of Pharmacy of the Prof. Mauricio Federal University of Rio de Janeiro, for help in the crystal symmetry and contacts analysis.

■

ABBREVIATIONS ZIKV, Zika virus; cryo-EM, cryo-electron microscopy; IDR, intrinsically disordered region; E, envelope; M, membrane; NC, nucleocapsid; ZIKVC, ZIKV capsid protein; DENV, Dengue virus; DENVC, DENV capsid protein; DENV2C, DENV serotype 2 capsid protein; WNV, West Nile virus; WNVC, WNV capsid protein; IDR, intrinsically disordered region; LDs, lipid droplets; NMR, nuclear magnetic resonance; IPTG, isopropyl β-D-1-thiogalactopyranoside; HEPES, 2-[4-(2hydroxyethyl)piperazin-1-yl]ethanesulfonic acid; EDTA, 2,2′,2″,2′″-(ethane-1,2-diyldinitrilo)tetraacetic acid; NUS, non-uniform sampling; HSQC, heteronuclear single-quantum coherence; TOCSY, total correlation spectroscopy; NOE, nuclear Overhauser effect; NOESY, NOE spectroscopy; CPMG, Carr−Purcell−Meiboom−Gill sequence; CPMG-RD, CPMG relaxation dispersion experiments; PSVS, Protein Structure Validation Software suite; DSS, 3-(trimethylsilyl)propane-1-sulfonic acid; sPRE, solvent paramagnetic relaxation enhancement; Gd-DTPA, gadolinium-diethylenetriamine pentaacetic acid.

■

REFERENCES

(1) Faria, N. R., Azevedo, R. d. S. d. S., Kraemer, M. U. G., Souza, R., Cunha, M. S., Hill, S. C., Theze, J., Bonsall, M. B., Bowden, T. A., Rissanen, I., Rocco, I. M., Nogueira, J. S., Maeda, A. Y., Vasami, F. G. d. S., Macedo, F. L. d. L., Suzuki, A., Rodrigues, S. G., Cruz, A. C. R., Nunes, B. T., Medeiros, D. B. d. A., Rodrigues, D. S. G., Nunes Queiroz, A. L., Silva, E. V. P. d., Henriques, D. F., Travassos da Rosa, E. S., De Oliveira, C. S., Martins, L. C., Vasconcelos, H. B., Casseb, L. M. N., Simith, D. d. B., Messina, J. P., Abade, L., Lourenco, J., Alcantara, L. C. J., Lima, M. M. d., Giovanetti, M., Hay, S. I., De Oliveira, R. S., Lemos, P. d. S., Oliveira, L. F. d., De Lima, C. P. S., Da Silva, S. P., Vasconcelos, J. M. d., Franco, L., Cardoso, J. F., VianezJunior, J. L. d. S. G., Mir, D., Bello, G., Delatorre, E., Khan, K., Creatore, M., Coelho, G. E., De Oliveira, W. K., Tesh, R., Pybus, O. G., Nunes, M. R. T., and Vasconcelos, P. F. C. (2016) Zika virus in the Americas: Early epidemiological and genetic findings. Science (Washington, DC, U. S.) 352, 345−349. (2) Carteaux, G., Maquart, M., Bedet, A., Contou, D., Brugières, P., Fourati, S., Cleret de Langavant, L., De Broucker, T., Brun-Buisson, C., Leparc-Goffart, I., and Mekontso Dessap, A. (2016) Zika Virus Associated with Meningoencephalitis. N. Engl. J. Med. 374, 1595− 1596. (3) Garcez, P. P., Loiola, E. C., Madeiro da Costa, R., Higa, L. M., Trindade, P., Delvecchio, R., Nascimento, J. M., Brindeiro, R., Tanuri, A., and Rehen, S. K. (2016) Zika virus impairs growth in human neurospheres and brain organoids. Science (Washington, DC, U. S.) 352, 816−818. 2497

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498

Article

Biochemistry (19) Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) Nmrpipe - a Multidimensional Spectral Processing System Based on Unix Pipes. J. Biomol. NMR 6, 277−293. (20) Vranken, W. F., Boucher, W., Stevens, T. J., Fogh, R. H., Pajon, A., Llinas, M., Ulrich, E. L., Markley, J. L., Ionides, J., and Laue, E. D. (2005) The CCPN data model for NMR spectroscopy: development of a software pipeline. Proteins: Struct., Funct., Genet. 59, 687−96. (21) Hyberts, S. G., Milbradt, A. G., Wagner, A. B., Arthanari, H., and Wagner, G. (2012) Application of iterative soft thresholding for fast reconstruction of NMR data non-uniformly sampled with multidimensional Poisson Gap scheduling. J. Biomol. NMR 52, 315−327. (22) Wittekind, M., and Mueller, L. (1993) HNCACB, a HighSensitivity 3D NMR Experiment to Correlate Amide-Proton and Nitrogen Resonances with the Alpha- and Beta-Carbon Resonances in Proteins. J. Magn. Reson., Ser. B 101, 201−205. (23) Ikura, M., Kay, L. E., and Bax, A. (1990) A novel approach for sequential assignment of proton, carbon-13, and nitrogen-15 spectra of larger proteins: heteronuclear triple-resonance three-dimensional NMR spectroscopy. Application to calmodulin. Biochemistry 29, 4659−4667. (24) Grzesiek, S., and Bax, A. (1993) Amino acid type determination in the sequential assignment procedure of uniformly 13C/15Nenriched proteins. J. Biomol. NMR 3, 185−204. (25) Kay, L. E., Xu, G. Y., Singer, A. U., Muhandiram, D. R., and Formankay, J. D. (1993) A Gradient-Enhanced HCCH-TOCSY Experiment for Recording Side-Chain 1H and 13C Correlations in H2O Samples of Proteins. J. Magn. Reson., Ser. B 101, 333−337. (26) Sattler, M. (1999) Heteronuclear multidimensional NMR experiments for the structure determination of proteins in solution employing pulsed field gradients. Prog. Nucl. Magn. Reson. Spectrosc. 34, 93−158. (27) Vranken, W. F., Boucher, W., Stevens, T. J., Fogh, R. H., Pajon, A., Llinas, M., Ulrich, E. L., Markley, J. L., Ionides, J., and Laue, E. D. (2005) The CCPN data model for NMR spectroscopy: development of a software pipeline. Proteins: Struct., Funct., Genet. 59, 687−696. (28) Shen, Y., and Bax, A. (2013) Protein backbone and sidechain torsion angles predicted from NMR chemical shifts using artificial neural networks. J. Biomol. NMR 56, 227−41. (29) Rieping, W., Habeck, M., Bardiaux, B., Bernard, A., Malliavin, T. E., and Nilges, M. (2007) ARIA2: automated NOE assignment and data integration in NMR structure calculation. Bioinformatics 23, 381−2. (30) Linge, J. P., O’Donoghue, S. I., and Nilges, M. (2001) Automated assignment of ambiguous nuclear overhauser effects with ARIA. Methods Enzymol. 339, 71−90. (31) Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr., Sect. D: Biol. Crystallogr. 54, 905−921. (32) Shen, Y., Delaglio, F., Cornilescu, G., and Bax, A. (2009) TALOS+: a hybrid method for predicting protein backbone torsion angles from NMR chemical shifts. J. Biomol. NMR 44, 213−23. (33) Laskowski, R. A., Rullmann, J. A. C., MacArthur, M. W., Kaptein, R., and Thornton, J. M. (1996) AQUA and PROCHECKNMR: Programs for checking the quality of protein structures solved by NMR. J. Biomol. NMR 8, 477−486. (34) Lipari, G., and Szabo, A. (1982) Model-free approach to the interpretation of nuclear magnetic resonance relaxation in macromolecules. 2. Analysis of experimental results. J. Am. Chem. Soc. 104, 4559−4570. (35) Loria, J. P., Rance, M., and Palmer, A. G., III (1999) A Relaxation-Compensated Carr-Purcell-Meiboom-Gill Sequence for Characterizing Chemical Exchange by NMR Spectroscopy. J. Am. Chem. Soc. 121, 2331−2332. (36) Larkin, M. A., Blackshields, G., Brown, N. P., Chenna, R., McGettigan, P. A., McWilliam, H., Valentin, F., Wallace, I. M., Wilm,

A., Lopez, R., Thompson, J. D., Gibson, T. J., and Higgins, D. G. (2007) Clustal W and Clustal X version 2.0. Bioinformatics 23, 2947− 2948. (37) Drozdetskiy, A., Cole, C., Procter, J., and Barton, G. J. (2015) JPred4: a protein secondary structure prediction server. Nucleic Acids Res. 43, W389−W394. (38) Hong, J., Jing, Q., and Yao, L. (2013) The protein amide 1 H(N) chemical shift temperature coefficient reflects thermal expansion of the N-H···O = C hydrogen bond. J. Biomol. NMR 55, 71−8. (39) Baxter, N. J., and Williamson, M. P. (1997) Temperature dependence of 1H chemical shifts in proteins. J. Biomol. NMR 9, 359− 69. (40) Tomlinson, J. H., and Williamson, M. P. (2012) Amide temperature coefficients in the protein G B1 domain. J. Biomol. NMR 52, 57−64. (41) Cierpicki, T., Zhukov, I., Byrd, R. A., and Otlewski, J. (2002) Hydrogen Bonds in Human Ubiquitin Reflected in Temperature Coefficients of Amide Protons. J. Magn. Reson. 157, 178−180. (42) Andersen, N. H., Chen, C. P., Marschner, T. M., Krystek, S. R., and Bassolino, D. A. (1992) Conformational isomerism of endothelin in acidic aqueous media: a quantitative NOESY analysis. Biochemistry 31, 1280−95. (43) Frauenfelder, H., McMahon, B. H., Austin, R. H., Chu, K., and Groves, J. T. (2001) The role of structure, energy landscape, dynamics, and allostery in the enzymatic function of myoglobin. Proc. Natl. Acad. Sci. U. S. A. 98, 2370−4. (44) Costantini, S., Colonna, G., and Facchiano, A. M. (2008) ESBRI: A web server for evaluating salt bridges in proteins. Bioinformation. 3, 137−138. (45) Samsa, M. M., Mondotte, J. A., Caramelo, J. J., and Gamarnik, A. V. (2012) Uncoupling cis-Acting RNA elements from coding sequences revealed a requirement of the N-terminal region of dengue virus capsid protein in virus particle formation. J. Virol. 86, 1046−58. (46) Khazanov, N., and Levy, Y. (2011) Sliding of p53 along DNA Can Be Modulated by Its Oligomeric State and by Cross-Talks between Its Constituent Domains. J. Mol. Biol. 408, 335−355. (47) Vuzman, D., and Levy, Y. (2012) Intrinsically disordered regions as affinity tuners in protein-DNA interactions. Mol. BioSyst. 8, 47−57. (48) Krepel, D., and Levy, Y. (2017) Intersegmental transfer of proteins between DNA regions in the presence of crowding. Phys. Chem. Chem. Phys. 19, 30562−30569.

2498

DOI: 10.1021/acs.biochem.9b00194 Biochemistry 2019, 58, 2488−2498