research papers\(\def\hfill{\hskip 5em}\def\hfil{\hskip 3em}\def\eqno#1{\hfil {#1}}\)

IUCrJ
ISSN: 2052-2525

Binding structures of SERF1a with NT17-polyQ peptides of huntingtin exon 1 revealed by SEC-SWAXS, NMR and molecular simulation

crossmark logo

aDepartment of Chemical Engineering, National Tsing Hua University, Hsinchu 300044, Taiwan, bNational Synchrotron Radiation Research Center, Hsinchu 300092, Taiwan, cGenomics Research Center, Academia Sinica, Taipei 115024, Taiwan, dChemical Biology and Molecular Biophysics Program, Taiwan International Graduate Program, Institute of Biological Chemistry, Academia Sinica, Taipei 115024, Taiwan, eInstitute of Biochemical Sciences, National Taiwan University, Taipei 106319, Taiwan, and fCollege of Semiconductor Research, National Tsing Hua University, Hsinchu 300044, Taiwan
*Correspondence e-mail: yrchen@gate.sinica.edu.tw, usjeng@nsrrc.org.tw

Edited by M. Takata, SPring-8, Japan (Received 31 December 2023; accepted 27 June 2024; online 8 August 2024)

The aberrant fibrillization of huntingtin exon 1 (Httex1) characterized by an expanded polyglutamine (polyQ) tract is a defining feature of Huntington's disease, a neurodegenerative disorder. Recent investigations underscore the involvement of a small EDRK-rich factor 1a (SERF1a) in promoting Httex1 fibrillization through interactions with its N terminus. By establishing an integrated approach with size-exclusion-column-based small- and wide-angle X-ray scattering (SEC-SWAXS), NMR, and molecular simulations using Rosetta, the analysis here reveals a tight binding of two NT17 fragments of Httex1 (comprising the initial 17 amino acids at the N terminus) to the N-terminal region of SERF1a. In contrast, examination of the complex structure of SERF1a with a coiled NT17-polyQ peptide (33 amino acids in total) indicates sparse contacts of the NT17 and polyQ segments with the N-terminal side of SERF1a. Furthermore, the integrated SEC-SWAXS and molecular-simulation analysis suggests that the coiled NT17 segment can transform into a helical conformation when associated with a polyQ segment exhibiting high helical content. Intriguingly, NT17-polyQ peptides with enhanced secondary structures display diminished interactions with SERF1a. This insight into the conformation-dependent binding of NT17 provides clues to a catalytic association mechanism underlying SERF1a's facilitation of Httext1 fibrillization.

1. Introduction

Huntington's disease, a neurodegenerative disorder, is caused by CAG (cytosine, adenine, guanine) expansion in exon 1 of the huntingtin (Htt) gene, leading to progressive brain dysfunction with signs of psychiatric disorder, muscle coordination impairment and cognitive decline (Matlahov & van der Wel, 2019[Matlahov, I. & van der Wel, P. C. (2019). Exp. Biol. Med. (Maywood), 244, 1584-1595.]). The abnormal expansion translates to mutant Htt proteins (huntingtin exon 1 or Httex1) characterized by an expanded polyglutamine (polyQ) (more than 35–40 glutamine) tract flanked by N-terminal 17 amino acids (NT17) and a C-terminal polyproline domain. The polyQ segments are prone to misfold and form α-helical packing. Subsequently, the helical packing converts to β-sheets for fibrillization conversion and deposition on neurons, leading to neuro­degenerative diseases (Arndt et al., 2015[Arndt, J. R., Chaibva, M. & Legleiter, J. (2015). Biomol. Conc. 6, 33-46.]). Previous studies showed that the amyloid fibrillization process could be enhanced by human small EDRK-rich factor 1a (SERF1a), through α-helix interactions with the NT17 domain of Httex1 (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]; Falsone et al., 2012[Falsone, S. F., Meyer, N. H., Schrank, E., Leitinger, G., Pham, C. L. L., Fodero-Tavoletti, M. T., Holmberg, M., Dulle, M., Scicluna, B., Gesslbauer, B., Rückert, H.-M., Wagner, G. E., Merle, D. A., Nollen, E. A., Kungl, A. J., Hill, A. F., Cappai, R. & Zangger, K. (2012). Cell. Rep. 2, 358-371.]). The interaction was revealed mainly from NMR observation that the α-helical regions of SERF1a were significantly influenced upon binding with Httex1-39Q protein (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]). Moreover, the binding ratio of SERF1a with the synthesized NT17 peptide, simulating the N-terminal first 17 amino acids of Httex 1, was shown to be 1:2 on the basis of isothermal titration calorimetry (ITC) results. The binding was found to be weakened when NT17 was attached to α-helical polyQ. Alterations in the radius of gyration (Rg) of SERF1a were noted following its binding with NT17 and with NT17-polyQ, as observed through small-angle X-ray scattering (SAXS) analysis (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]). Although the NMR and SAXS data have provided sufficient evidence and structural features of SERF1a–polyQ binding, the specific molecular structures that could better characterize this binding mechanism remain unresolved. A deeper understanding of how the polyQ segment structure influences the binding of NT17-polyQ peptides to SERF1a continues to be a significant area of interest.

Probing solution binding structures of the complex of SERF1a and NT17-polyQ peptides could benefit greatly from integrated methodologies of complementary structural sensitivities. Recent reviews (Schroer & Svergun, 2018[Schroer, M. A. & Svergun, D. I. (2018). Emerg. Top. Life Sci. 2, 69-79.]; Jeffries et al., 2021[Jeffries, C. M., Ilavsky, J., Martel, A., Hinrichs, S., Meyer, A., Pedersen, J. S., Sokolova, A. V. & Svergun, D. I. (2021). Nat. Rev. Methods Primers, 1, 70.]; Bizien et al., 2016[Bizien, T., Durand, D., Roblina, P., Thureau, A., Vachette, P. & Pérez, J. (2016). Protein Pept. Lett. 23, 217-231.]; Trewhella et al., 2017[Trewhella, J., Duff, A. P., Durand, D., Gabel, F., Guss, J. M., Hendrickson, W. A., Hura, G. L., Jacques, D. A., Kirby, N. M., Kwan, A. H., Pérez, J., Pollack, L., Ryan, T. M., Sali, A., Schneidman-Duhovny, D., Schwede, T., Svergun, D. I., Sugiyama, M., Tainer, J. A., Vachette, P., Westbrook, J. & Whitten, A. E. (2017). Acta Cryst. D73, 710-728.]) illustrate a development trend of SAXS with online size-exclusion chromatography (SEC) for complex conformations of biomolecules in solution. Further incorporated with optical measurements of UV–Vis absorption and differential refractive index (dRI) (Blanchet et al., 2015[Blanchet, C. E., Spilotros, A., Schwemmer, F., Graewert, M. A., Kikhney, A., Jeffries, C. M., Franke, D., Mark, D., Zengerle, R., Cipriani, F., Fiedler, S., Roessle, M. & Svergun, D. I. (2015). J. Appl. Cryst. 48, 431-443.]), SEC-SAXS can reveal conformation and composition of a biomolecular binding complex in one sample elution (Shih et al., 2023[Shih, O., Yeh, Y., Liao, K., Li, K., Tsai, J., Li, C., Chiang, Y., Heenan, R. K., Sun, Y. & Jeng, U. (2023). Mater. Chem. Phys. 308, 128253.], 2017[Shih, O., Yeh, Y.-Q., Liao, K.-F., Sung, T.-C., Chiang, Y.-W. & Jeng, U.-S. (2017). Phys. Chem. Chem. Phys. 19, 7947-7954.]; Yeh et al., 2017[Yeh, Y.-Q., Liao, K.-F., Shih, O., Shiu, Y.-J., Wu, W.-R., Su, C.-J., Lin, P.-C. & Jeng, U.-S. (2017). J. Phys. Chem. Lett. 8, 470-477.]). Very recently, SEC-SAXS was further extended to wide-angle X-ray scattering (WAXS) for SEC-SWAXS, allowing better model discrimination and molecular model construction (Shih et al., 2022[Shih, O., Liao, K.-F., Yeh, Y.-Q., Su, C.-J., Wang, C.-A., Chang, J.-W., Wu, W.-R., Liang, C.-C., Lin, C.-Y., Lee, T.-H., Chang, C.-H., Chiang, L.-C., Chang, C.-F., Liu, D.-G., Lee, M.-H., Liu, C.-Y., Hsu, T.-W., Mansel, B., Ho, M.-C., Shu, C.-Y., Lee, F., Yen, E., Lin, T.-C. & Jeng, U. (2022). J. Appl. Cryst. 55, 340-352.]; Bizien et al., 2016[Bizien, T., Durand, D., Roblina, P., Thureau, A., Vachette, P. & Pérez, J. (2016). Protein Pept. Lett. 23, 217-231.]; Schroer & Svergun, 2018[Schroer, M. A. & Svergun, D. I. (2018). Emerg. Top. Life Sci. 2, 69-79.]). With SEC-SWAXS structural sensitivity, hydration water structures near the surface of biomolecules become relevant in the SWAXS data analysis, as demonstrated recently (Hub, 2018[Hub, J. S. (2018). Curr. Opin. Struct. Biol. 49, 18-26.]; Manalastas-Cantos et al., 2021[Manalastas-Cantos, K., Konarev, P. V., Hajizadeh, N. R., Kikhney, A. G., Petoukhov, M. V., Molodenskiy, D. S., Panjkovich, A., Mertens, H. D. T., Gruzinov, A., Borges, C., Jeffries, C. M., Svergun, D. I. & Franke, D. (2021). J. Appl. Cryst. 54, 343-355.]).

In this study, we establish a combined analysis of SEC-SWAXS, NMR and molecular simulation to construct the molecular models of SERF1a, NT17, NT17-polyQ peptides and their complexes in solution. The UV–Vis absorption and dRI data concomitantly measured along the SEC-SWAXS sample elution provide complementary the composition information, which is crucial in the constructions of the complex structures. These constructed molecular structures provide hints to the association mechanism of SERF1a with the NT17-polyQ peptides.

2. Materials and methods

2.1. Sample preparation

SERF1a was purified from the human SERF1a gene as reported previously (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]). NT17 containing the N-terminal first 17 amino acids of the Httex1 protein and NT17-polyQ peptides with 33 amino acids of Htt-0, Htt-1 and Htt-3 (cf. Table 1[link]) were synthesized by the peptide synthesis core in Genomics Research Center, Academia Sinica. Htt-0 is of 14 successive glutamine amino acids attached to NT17, simulating the native polyQ structure of successive glutamines. Htt-1 has four leucine residues replacing residues of position a/d in the polyQ region of Htt-0, to enhance α-helix formation (Fiumara et al., 2010[Fiumara, F., Fioriti, L., Kandel, E. R. & Hendrickson, W. A. (2010). Cell, 143, 1121-1135.]); whereas, for Htt-3, the positions a/d in the polyQ region are substituted by four prolines to suppress formation of α-helix structure (Table 1[link], see italic letters for the replaced residues). Sample solutions of SERF1a (4.1 mg ml−1), NT17 (1.6 mg ml−1) and Htt-0, Htt-1 and Htt-3 (all of 10 mg ml−1) were prepared with sodium phosphate buffer (PB) solution (containing 480 µl of 10 mM PB, pH 7.4, 16.5 µl of 100 mM NaOH and 10 µl of 1% TFA). To avoid sedimentation due to strong binding, concentration-reduced solutions of 0.255 mg ml−1 SERF1a (61 µM) and of 0.445 mg ml−1 NT17 (129 µM) were prepared individually then mixed for a SERF1a:NT17 mixture of 1:2 molar ratio. A mixture of SERF1a:Htt3 with 1:2 molar ratio was prepared from sample solutions of 0.7 mg ml−1 SERF1a and 0.8 mg ml−1 Htt-3. All the sample solutions were filtered (0.22 µm pore size) prior to SWAXS measurements.

Table 1
The sequence of amino acids of NT17 and NT17-polyQ peptides of Htt-0, Htt-1 and Htt-3

The amino acids of SERF1a are marked in bold for the helical segments determined by NMR backbone assignments.

Sample Sequence
NT17 MATLEKLMKAFESLKSF
Htt-0 MATLEKLMKAFESLKSFQQQQQQQQQQQQQQYK
Htt-1 MATLEKLMKAFESLKSFLQQLQQQLQQLQQQYK
Htt-3 MATLEKLMKAFESLKSFPQQPQQQPQQPQQQYK
SERF1a MARGNQRELARQKNMKKTQEISKGKRKEDSLTASQRKQRDSEIMQEKQKAANEKKSMQTREK
  1------8---12-------------------33----------------51---------62

2.2. SEC-SWAXS

SWAXS data were measured at the Taiwan Photon Source (TPS) 13A BioSWAXS endstation of the National Synchrotron Radiation Research Center (NSRRC), using a 15 keV X-ray beam and two in-vacuum Eiger X 9M and 1M detectors for simultaneous SAXS and WAXS measurements, as detailed in previous reports (Shih et al., 2022[Shih, O., Liao, K.-F., Yeh, Y.-Q., Su, C.-J., Wang, C.-A., Chang, J.-W., Wu, W.-R., Liang, C.-C., Lin, C.-Y., Lee, T.-H., Chang, C.-H., Chiang, L.-C., Chang, C.-F., Liu, D.-G., Lee, M.-H., Liu, C.-Y., Hsu, T.-W., Mansel, B., Ho, M.-C., Shu, C.-Y., Lee, F., Yen, E., Lin, T.-C. & Jeng, U. (2022). J. Appl. Cryst. 55, 340-352.]; Liu et al., 2021[Liu, D.-G., Chang, C.-H., Chiang, L.-C., Lee, M.-H., Chang, C.-F., Lin, C.-Y., Liang, C.-C., Lee, T.-H., Lin, S.-W., Liu, C.-Y., Hwang, C.-S., Huang, J.-C., Kuan, C.-K., Wang, H.-S., Liu, Y.-C., Tseng, F.-H., Chuang, J.-Y., Liao, W.-R., Li, H.-C., Su, C.-J., Liao, K.-F., Yeh, Y.-Q., Shih, O., Wu, W.-R., Wang, C.-A. & Jeng, U. (2021). J. Synchrotron Rad. 28, 1954-1965.]). Sample solutions of 50–100 µl were injected into the SEC-SWAXS system, coupled with an in-line high-performance liquid chromatography (HPLC) unit (Agilent 1260 series) and UV–Vis absorption/refractive index spectrometers. With a flow rate of 0.35 ml min−1, the sample solution (sandwiched by the buffer solution, with a dilution factor of ca. 2–4) was directed to a quartz capillary (of 2 mm diameter and a wall thickness of 10 µm) thermostated at 283 K for simultaneous X-ray and UV–Vis exposures, followed by dRI measurements at ca. 40 cm downstream. Elution peaks and peak widths of the measured UV–Vis and dRI profiles were aligned through a profile broadening and alignment process by the software package VISION (WYATT Technology Corporation, Santa Barbara, California, USA) (including VOYAGER and ASTRA) for determination of the complex compositions. SWAXS data were collected continuously with 2 s per data frame over the elution peak. With the TPS 13A SWAXS Data Reduction Kit (DRK) Version 3.6, the frame data of well overlapped SAXS profiles were averaged and subtracted with buffer scattering (measured before or after the elution peak), followed by normalizations of the incoming X-ray flux and sample thickness, then rescaled to absolute scattering intensity in units of cm−1 via comparing with the absolute scattering intensity of water. All the presented SAXS data are not normalized by sample concentrations. Each frame of the elution SAXS data was routinely examined by our own data-evaluation package, which incorporates several functions of ATSAS, including (1) checking the Guinier region quality with AUTORG, (2) examining the Kratky plot for background-subtraction quality and structural compactness, (3) overlapping the selected SAXS data frames along the elution for a lack of concentration-dependency (interparticle interaction) check, and (4) generating distance distribution functions p(r) and executing DAMMIF for shape evaluation with an ab initio model. All these were detailed in our previous report (Shih et al., 2022[Shih, O., Liao, K.-F., Yeh, Y.-Q., Su, C.-J., Wang, C.-A., Chang, J.-W., Wu, W.-R., Liang, C.-C., Lin, C.-Y., Lee, T.-H., Chang, C.-H., Chiang, L.-C., Chang, C.-F., Liu, D.-G., Lee, M.-H., Liu, C.-Y., Hsu, T.-W., Mansel, B., Ho, M.-C., Shu, C.-Y., Lee, F., Yen, E., Lin, T.-C. & Jeng, U. (2022). J. Appl. Cryst. 55, 340-352.]). The WAXS data frames corresponding to these processed SAXS data were then processed by the DRK. The two sets of data were merged for integrated SWAXS data, covering a wide scattering vector q range over 1.0 Å−1, where q = 4πλ−1sin θ is defined by the X-ray wavelength λ and scattering angle 2θ.

2.3. SWAXS data analysis with I-TASSER, CRYSOL and Rosetta-fastsaxs

Five initial three-dimensional atomic models with relatively low free energy of SERF1a, NT17 or the polyQ peptides were constructed using the I-TASSER protein structure and function prediction server (Roy et al., 2010[Roy, A., Kucukural, A. & Zhang, Y. (2010). Nat. Protoc. 5, 725-738. ]; Yang & Zhang, 2015a[Yang, J. & Zhang, Y. (2015a). CP Bioinformatics, 52, 5. ],b[Yang, J. & Zhang, Y. (2015b). Nucleic Acids Res., 43, W174-W181.]; Zhang, 2008[Zhang, Y. (2008). BMC Bioinformatics, 9, 40. ]), from multiple threading alignments and iterative structural-assembly simulations. The initial structures were built from scratch by ab initio modeling (Wu et al., 2007[Wu, S., Skolnick, J. & Zhang, Y. (2007). BMC Biol. 5, 17. ]), as the PDB library had no structurally related homologues for the coiled SERF1a and NT17-polyQ peptides. Subsequently, theoretical SWAXS profiles of the five output models were computed using the all-atom-calculation program CRYSOL (Svergun et al., 1995[Svergun, D., Barberato, C. & Koch, M. H. J. (1995). J. Appl. Cryst. 28, 768-773. ]), which fits the model-calculated SWAXS profiles to the experimental data using only two parameters of the average displaced solvent volume per atomic group and the contrast of a hydration layer. The models with reasonably low least-square fitting (χ2, goodness of fit) were selected for further energy minimization with SAXS data fitting χ2 using the Rosetta modeling suite (Bender et al., 2016[Bender, B. J., Cisneros, A. III, Duran, A. M., Finn, J. A., Fu, D., Lokits, A. D., Mueller, B. K., Sangha, A. K., Sauer, M. F., Sevy, A. M., Sliwoski, G., Sheehan, J. H., DiMaio, F., Meiler, J. & Moretti, R. (2016). Biochemistry, 55, 4748-4763. ]; Das & Baker, 2008[Das, R. & Baker, D. (2008). Annu. Rev. Biochem. 77, 363-382. ]; Kaufmann et al., 2010[Kaufmann, K. W., Lemmon, G. H., DeLuca, S. L., Sheehan, J. H. & Meiler, J. (2010). Biochemistry, 49, 2987-2998. ]), with a two-step free-energy minimization of backbones followed by side chains. The SAXS data fitting constraint was applied by adding the fastsaxs term in the Rosetta default scoring function Talaris2014 (O'Meara et al., 2015[O'Meara, M. J., Leaver-Fay, A., Tyka, M. D., Stein, A., Houlihan, K., DiMaio, F., Bradley, P., Kortemme, T., Baker, D., Snoeyink, J. & Kuhlman, B. (2015). J. Chem. Theory Comput., 11, 388-398.]; Stovgaard et al., 2010[Stovgaard, K., Andreetta, C., Ferkinghoff-Borg, J. & Hamelryck, T. (2010). BMC Bioinformatics, 11, 429. ]). The fastsaxs term contributes to the total Rosetta score as an effective energy score and was scaled from the SWAXS data fitting χ2 (up to q = 0.7 Å−1) by a weighting factor. The weighting factor of χ2 was set between 1 and 500 according to the scores of the other free-energy terms, to ensure that a low χ2 value below ca. 2–3 could be reached in the model search with the total Rosetta score minimization process (χ2 of SWAXS data fitting in a wider q range is generally higher than that of the SAXS data fitting alone). In the structure-optimization process of SERF1a, the NMR-determined α-helix and relevant residue structures were used as the model constraint. Each optimized model of SERF1a, NT17 or NT17-polyQ is a representative one selected from the ten best models of independent runs of the Rosetta-fastsaxs modeling. The degree of model convergence is represented by the normalized spatial discrepancy (NSD) of the three-dimensional points of the ten models, as a quantitative measure of conformation similarity; NSD = 0 indicates perfectly overlapped models, whereas NSD < 1 is a general criterion for converged resembled models (Konarev et al., 2016[Konarev, P. V., Petoukhov, M. V. & Svergun, D. I. (2016). J. Appl. Cryst. 49, 953-960.]; Prior et al., 2020[Prior, C., Davies, O. R., Bruce, D. & Pohl, E. (2020). J. Chem. Theory Comput. 16, 1985-2001.]).

2.4. Composition determination of the complex

The number density no of a two-component complex of SERF1a (A) with NT17 (B) (or NT17-polyQ peptides) in solution with complexing numbers NA and NB can be determined from the SEC-SWAXS zero-angle scattering intensity I0-SAXS (or I0), together with the sample concentrations determined from the simultaneously measured optical density IUV of UV–Vis absorption and differential refractive index ΔndRI, over the sample elution region (Shih et al., 2018[Shih, O., Yeh, Y.-Q., Liao, K.-F., Su, C.-J., Wu, P.-H., Heenan, R.-K., Yu, T.-Y. & Jeng, U. (2018). J. Phys. Chem. Lett. 9, 4287-4293.], 2023[Shih, O., Yeh, Y., Liao, K., Li, K., Tsai, J., Li, C., Chiang, Y., Heenan, R. K., Sun, Y. & Jeng, U. (2023). Mater. Chem. Phys. 308, 128253.]; Lin et al., 2009[Lin, J.-M., Lin, T.-L., Jeng, U.-S., Huang, Z.-H. & Huang, Y.-S. (2009). Soft Matter, 5, 3913-3919.]). The three parameters of no, NA and NB can be resolved analytically using the following set of equations:

[\Delta n_{\rm d{\rm RI}} = C\left [{{\varphi _{\rm A}}{{\left({{{{\rm d}n} \over {{\rm d}c}}} \right)}_{\rm A}} + {\varphi _{\rm B}}{{\left({{{{\rm d}n} \over {{\rm d}c}}} \right)}_{\rm B}}} \right], \eqno (1)]

[I_{\rm UV} = CL(\varphi_{\rm A}\varepsilon_{\rm A} + \varphi_{\rm B} \varepsilon_{\rm B}) \eqno (2)]

and

[I_{0{\hbox{-}}{\rm SAXS}} = {n_{\rm o}}{\left({{\,f_{\rm t}} - {\rho _w}{V_{\rm t}}} \right)^2}. \eqno (3)]

In equation (1)[link], C is the weight concentration of the complex in solution, and the mass fractions of components A and B of the complex are φA = NAMA/Mt and φB = NBMB/Mt, with the complex molar weight Mt = NAMA + NBMB; (dn/dc)A and (dn/dc)B are the corresponding specific refractive index increments, and ɛA and ɛB are the UV–Vis light extinction coefficients. NB/NA defines the association ratio χBA. In equation (2)[link], L is the sample path length defined by the sample capillary diameter. We note that I0-SAXS was extrapolated from the SAXS data using the Guinier approximation, with data in the q range of qRg < 1.3. In equation (3)[link], no = CNA0/Mt, where NA0 is Avogadro's number and ρw is the scattering length density of the buffer. The total X-ray scattering length ft = (NAEA + NBEB)fe is contributed to by the number of electrons EA and EB of A and B with the X-ray scattering length of electrons fe = 2.8179 × 10−5 Å. The complex volume Vt = NAVA + NBVB is contributed to by two components, with VA = 9719 Å3 of SERF1a and VB = 3190 Å3 of NT17 calculated on the basis of their sequences (Jacrot, 1976[Jacrot, B. (1976). Rep. Prog. Phys. 39, 911-953.]). The dn/dc values measured individually for SERF1a and NT17 are 0.1847 and 0.1870 ml g−1, respectively; the corresponding UV–Vis molar absorption coefficients ε214 at 214 nm are 64 310 and 27 715 M−1cm−1, respectively (214 nm data were used for the UV–Vis absorption as the commonly used 280 nm is low for either NT17 or SERF1a). For Htt-0, Htt-1 and Htt-3, ε214 values are 49887, 49499 and 60019 M−1cm−1, respectively, as summarized in Table S1 of the supporting information.

3. Results and discussion

3.1. SERF1a

Fig. 1[link](a) shows SEC-SWAXS elution profiles of radius of gyration Rg, zero-angle scattering intensity I0 and deduced aggregation number N of SERF1a, revealing largely monodisperse monomers (N = 1) of Rg = 23.5 ± 1.0 Å. The SWAXS data are compared with a theoretical profile [Fig. 1[link](b)] calculated from a SERF1a pure NMR structure model. The NMR structure model was selected from 10 000 models built from Chemical-Shift-Rosetta (CS-Rosetta) (Lange et al., 2012[Lange, O. F., Rossi, P., Sgourakis, N. S., Song, Y., Lee, H.-W., Aramini, J. M., Ertekin, A., Xiao, R., Acton, T. B., Montelione, G. T. & Baker, D. (2012). Proc. Natl Acad. Sci. USA, 109, 10873-10878.]) using the NMR backbone chemical shift data reported previously (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]), on the basis of a lowest free energy. The apparent deviation of the calculated model profile from the SWAXS data in the low-q region indicates a deficiency of the model constructed merely from considerations of the NMR backbone structure with energy minimization. Subsequently, we used the NMR-derived model as an initial structure in the Rosetta-fastsaxs modeling. Whereby, all the NMR-determined helical segments (cf. Table 1[link]) and relevant amino residues were used as the modeling constraints. As a result, an optimize model with a relatively elongated conformation incorporating all of the NMR-determined helical structure can fit nearly the full q range of the SWAXS data (Fig. 1[link]). This model is representative of one of the ten best-fitted models obtained from ten independent runs of the Rosetta-fastsaxs modeling, with comparable free-energy scores and χ2 values as shown in Fig. S1 of the supporting information. The modest NSD value of 1.3, calculated from ten models with the same NMR-determined helical features and similar Rg values, suggests that SERF1a may exhibit a relatively converged conformation, despite some variability in the deployment of local non-rigid segments. In our previous report (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]), we analyzed SAXS data of SERF1a using Rosetta-fastsaxs without NMR constraints. However, the model [shown in Fig. 1[link](b)] constructed from the SWAXS data with NMR constraints now provides a more refined orientation of SERF1a's segmented helices.

[Figure 1]
Figure 1
(a) Elution profiles of Rg, zero-angle intensity I0 and non-aggregation number values near N = 1, deduced from the integrated analysis of SAXS I0 and optical data of SERF1a. (b) SWAXS data of SERF1a are fitted using an NMR model (inset) and the SWAXS-NMR model (inset, scaled with a bar for 10 Å size). The data and model are deposited in the SASBDB database with the code SASDVL5 (Kikhney et al., 2020[Kikhney, A. G., Borges, C. R., Molodenskiy, D. S., Jeffries, C. M. & Svergun, D. I. (2020). Protein Sci. 29, 66-75. ])

3.2. NT17 conformation and transition via polyQ

Shown in Fig. 2[link](a) are the SWAXS data for the NT17-polyQ peptide Htt-3, with Rg = 16.7 Å extracted from the Guinier approximation. On the basis of the model searching process of I-TASSER and CRYSOL, followed by Rosetta-fastsaxs detailed above, a coil model (shown as an inset in the figure plot) was built to adequately fit the SWAXS data. This coil feature is consistent with that reported previously for a similar NT17-polyQ peptide at pH = 7 on the basis of NMR data analysis (Baias et al., 2017[Baias, M., Smith, P. E. S., Shen, K., Joachimiak, L. A., Żerko, S., Koźmiński, W., Frydman, J. & Frydman, L. (2017). J. Am. Chem. Soc. 139, 1168-1176.]). Furthermore, a parallel SWAXS data analysis for the segment of NT17 [Fig. 2[link](b)] consistently reveals a loose NT17 coil of Rg = 11.6 ± 0.5 Å. The revealed coil structures of NT17 and Htt-3 are consistent with their circular-dichroism (CD) spectra, i.e. both with no pronounced secondary structure measured (Fig. S2).

[Figure 2]
Figure 2
SWAXS data of (a) Htt-3 (SASDVM5) and (b) NT17 (SASDVN5) are fitted using the represented models shown in the insets (with a scale bar of 10 Å). The corresponding NSD values calculated from ten Rosetta models of repeated runs are 0.18 for Htt-3 and 0 for NT17.

In contrast, SWAXS data analysis for Htt-0 [Fig. 3[link](a)] indicates that the polyQ region of 14 successive glutamine amino acids attached to NT17 could form a more extended structure, with an Rg value of 19.8 Å significantly larger than that (16.7 Å) of the coiled Htt-3. The slightly richer helical feature of the Rosetta-fastsaxs model is consistent with the higher α-helical content revealed from the CD spectrum measured (Fig. S2). In the case of Htt-1 of a specially designed polyQ sequence with leucines for enhancing α-helical structure, the CD spectrum indeed reveals a dominant α-helical feature (Fig. S2); our molecular-simulation result also consistently suggests a highly helical structure of Htt-1, including the NT17 and the helical-enhancing polyQ regions. Namely, NT17 could be directed by the helical polyQ segment to form a helical conformation. The SWAXS data measured [Fig. 3[link](b)] reveal a dimer conformation for Htt-1, as determined from the zero-angle scattering intensity I0 value and the sample concentration measured from UV–Vis absorption. We, therefore, fitted the SWAXS data with two helical Htt-1 placed close to each other as an initial model. As a result, the best dimer conformation shown in Fig. 3[link](b) can largely fit the data in the lower-q region, supporting the dimer conformation. Nevertheless, the local structures of the dimer model could not adequately describe the broad hump centered at q ≃ 0.45 Å−1; this hump is attributed to a feature of helix–helix correlation of the two Htt-1 in a dimer form that could not be observed in the SWAXS profiles of Htt-0 and Htt-1 monomers (Fig. 3[link]). We also tried the SASREF algorithm of the ATSAS package (Cowieson et al., 2020[Cowieson, N. P., Edwards-Gayle, C. J. C., Inoue, K., Khunti, N. S., Doutch, J., Williams, E., Daniels, S., Preece, G., Krumpa, N. A., Sutter, J. P., Tully, M. D., Terrill, N. J. & Rambo, R. P. (2020). J. Synchrotron Rad. 27, 1438-1446.]) for a rigid body refinement with two Htt-1 monomers, based on a whole-helix structure of Htt-1 built from I-TASSER. However, the best-fitted model [shown in Fig. 3[link](b)] also could not describe the helix–helix correlation hump at q ≃ 0.45 Å−1. Nevertheless, the formation of helical Htt-1 dimers revealed from CD and SEC-SWAXS (I0 and optical data) suggests an association mechanism of the polyQ peptides through interactions of helical segments.

[Figure 3]
Figure 3
(a) SWAXS data of Htt-0 (SASDVP5) are fitted using the Rosetta model shown in the inset (the scale bar in the inset is for 10 Å). (b) SWAXS data of Htt-1SASDVQ5 are fitted using the dimer models optimized by Rosetta-fastsaxs and SASREF, as shown in the insets. Also shown is the profile calculated with a helical monomer model of Htt-1.

The low-q data of Htt-0 [Fig. 3[link](a)] deviate slightly, ∼10% lower in intensity, from the fitting profile, which may suggest an effect of interparticle interactions. However, the corresponding UV–Vis intensity and Rg value evolutions across the SEC-SWAXS elution peak (Fig. S5) demonstrate relatively stable Rg values, with small fluctuations within 1–2 Å range, despite varying sample concentrations. This consistency suggests negligible interparticle effects in our data fitting, as further shown by the moderate χ2 value of 1.9.

Previous ITC results (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]) indicate that the binding affinity of SERF1a to NT17 alone is fourfold of that with Htt-3 and it has little or no binding affinity with Htt-0 or Htt-1. Combining the above structural results for NT17 and the NT17-polyQ of different α-helical contents (Fig. S2), we suggest that the helical conformation of the polyQ region could affect NT17 for adopting helical conformation. Moreover, the highly helical conformation of the NT17-polyQ peptide Htt-1 favors self-association into dimer conformation, compared with interaction with SERF1a. Previous studies have already shown that polyQ peptides could influence the NT17 structure in certain conditions, thereby altering the aggregation behavior of the NT17-polyQ (Thakur et al., 2009[Thakur, A. K., Jayaraman, M., Mishra, R., Thakur, M., Chellgren, V. M., Byeon, I. J., Anjum, D. H., Kodali, R., Creamer, T. P., Conway, J. F., Gronenborn, A. M. & Wetzel, R. (2009). Nat. Struct. Mol. Biol. 16, 380-389.]; Urbanek et al., 2020[Urbanek, A., Popovic, M., Morató, A., Estaña, A., Elena-Real, C. A., Mier, P., Fournet, A., Allemand, F., Delbecq, S., Andrade-Navarro, M. A., Cortés, J., Sibille, N. & Bernadó, P. (2020). Structure, 28, 733-746.e5.]). The feature of NT17 conformation-dependent interaction with SERF1a revealed here suggests a possible catalytic role for the latter in its facilitating aggregation of Httex1. Presumably, when the polyQ segment of Httex1 begins to form an α-helical structure through SERF1a-facilitated aggregation, this helical configuration could encourage a similar helical transformation in NT17, promoting self-aggregation and thereby reducing interactions with SERF1a. Namely, the increased helical structure within polyQ enhances its self-aggregation, consequently diminishing interactions with SERF1a. Indeed, a previous report by Tsai et al. (2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]) supports this hypothesis, indicating minimal presence of SERF1a in the Httex1 fibril aggregates mediated by SERF1a, which underscores its proposed catalytic role.

3.3. Complexing of SERF1a with NT17

In the following, we focus on two complex structures of SERF1a – with NT17 and with Htt-3. Fig. 4[link](a) shows the evolutions of Rg (ca. 20 Å) and individual concentrations of NT17 and SERF1a of the complex determined from the optical data of UV–Vis absorption and dRI collected over the SEC-SWAXS elution. The result suggests a stable concentration ratio M of [NT17]/[SERF1a] ≃ 2 over the elution. Following the method detailed previously, we deduce the aggregation numbers of the complex to be two NT17 with one SERF1a from the I0 value and the sample component concentrations; this composition is consistent with the binding ratio 2:1 determined from ITC (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]). On the basis of the composition analysis, the SWAXS data of the complex are fitted with an initial model of two NT17 placed close on the N-terminal side of SERF1a in the Rosetta-fastsaxs protocol. This spatial arrangement is inspired by the previous NMR result that SERF1a interacts with Httex1-39Q mainly with its helical regions (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]).

[Figure 4]
Figure 4
(a) Evolutions of Rg, individual concentrations C of SERF1a and NT17, and molar ratio M, deduced from the SEC-SWAXS elution of the mixture. (b) The SWAXS data of the NT17–SERF1a complex (SASDVR5), merged from the data frames shadowed in (a), are fitted using Rosetta-fastsaxs with the model shown in (c). (c) The representative Rosetta-fastsaxs model with two NT17 (represented by a ribbon and green dots) binding to the N-terminal side of SERF1a (in red). (d) A zoom-in view for the binding of the N side of NT17 (colored coil) on the short helical segment (Site-1) and the long coil region (Site-2) on the N-terminal side of SERF1a (cf. Table 1[link]).

Shown in Fig. 4[link](b) are the SAXS data fitted using the model shown in Fig. 4[link](c). The model has a relatively minimized free-energy score and a reasonably small χ2 value of 2.4 among the ten models constructed by ten independent runs of Rosetta-fastsaxs (Fig. S3). The model illustrates that one NT17 tightly entangles with the N terminus of SERF1a, while the other more extended NT17 associates (via its first five amino acids Met1 to Glu5, and C-terminal Leu14 and Phe17) with the short helical segments of Arg7 to Arg11 near the N terminus [Site-1 in Fig. 4[link](d)] and Thr32 and Arg36 in the central coil region [Site-2 in Fig. 4[link](d)] of SERF1a. Although the local structural features proposed by the Rosetta model may not be unique, the Rosetta model (of complementary optimized free-energy scores and SWAXS data fitting χ2) could elucidate a reliable global complex conformation and likely local structural features of the SERF1a–NT17 complex as a basis for further structural verification. We observe a minor interparticle effect in the low-q data [Fig. 4[link](b)], indicated by a slight (10–15%) intensity deviation from the fitting profile. This effect may contribute to the minor fluctuations observed in the composition profile [see Fig. 4[link](a)], which is derived from the concentration-normalized zero-angle scattering intensity, I(0)/C, throughout the SEC-SWAXS elution.

Parallel analysis of the SEC-SWAXS result for the complex of SERF1a with Htt-3 is shown in Fig. 5[link], with the evolutions of Rg and decomposed concentrations of Htt-3 and SERF1a [Fig. 5[link](a)] of the complex over the SEC-SWAXS elution. The molar ratio M of Htt-3 over SERF1a over the elution peak is ∼1. Determined from the absolute intensity I0 and the sample concentration, the complex comprises one NT17 and SERF1a for a 1:1 binding ratio. This binding ratio is significantly lower than the 1:2 binding ratio with NT17 segments alone. Consistently, previous ITC results also revealed a significantly lower binding affinity of SERF1a–Htt-3 than that of SERF1a–NT17, but still with a binding ratio of 2:1 for Htt-3:SERF1a (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]). The lower binding ratio of 1:1 deduced here might be attributed to a fast sample mixing process used in the SEC-SWAXS measurement, compared with the slow Htt-3 titration process into the SERF1a solution in the ITC measurement. Nevertheless, according to the 1:1 binding ratio deduced from the SEC-SWAXS sample elution, the SWAXS data were fitted with an initial model of one Htt-3 with one SERF1a in the Rosetta-fastsaxs fitting protocol. Shown in Fig. 5[link](b) are the SWAXS data fitted with an optimized model, with relatively minimized free-energy score and χ2 = 2.19 (Fig. S4). The model reveals two major interactions sites of Thr3 (NT17 segment) and Pro28 (polyQ segment) of Htt-3 with Asn5 and Lys23 of the coil segments of SERF1a, respectively, via hydrogen bonding. Furthermore, the model suggests that although the long helical segments and the nearby amino residues of SERF1a do not directly interact with Htt-3, the conformation is also influenced, which is consistent with the NMR observation reported previously (Tsai et al., 2023[Tsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.]). The local interaction sites are determined mainly by minimization of the Rosetta free-energy score and would fluctuate somewhat over the models of comparable free-energy scores and χ2 (as shown in Fig. S4). However, these models collectively propose that the NT17-polyQ of Htt-3 loosely attaches to the N-terminal side of SERF1a, via only a few binding sites. This result is consistent with the significantly weaker interactions revealed by ITC, when compared with the robust binding observed between the short NT17 segments alone and SERF1a (cf. Fig. 4[link]).

[Figure 5]
Figure 5
(a) Evolutions of Rg, decomposed concentrations of SERF1a and Htt-3 in µM, and the molar ratio M of the complex of Htt-3/SERF1a, over the SEC-SAXS elution peak. (b) The SEC-SAXS data, merged from the data frames shadowed in (a), are fitted (solid curve with χ2 = 2.19) using a representative Rosetta model (inset; SASDVS5) comprising one SERF1a and one Htt-3. Furthermore, (c) and (d) are zoom-in views of the two binding sites of Htt-3 (via Thr3 and Pro28 of the NT17-polyQ) with the N-terminal side (Asn5 and Lys23 amino acids) of SERF1a.

3.4. Model uncertainty and heterogeneities

We have developed representative models for the non-compact SERF1a, NT17 and NT17-polyQ in solution using size-exclusion SWAXS instrumentation, combined with data analysis from SWAXS, NMR, optical spectroscopy and molecular simulation. These models incorporate all relevant structural features. In our analysis of the SWAXS data, we assumed a single conformation based on selecting data with the same Rg value from the SEC-SWAXS elution profile (refer to Fig. S5). However, these models exhibit certain local structural uncertainties, as indicated by the non-zero values of the NSD, shown in Table 2[link]. The NSD values were calculated from each set of the ten best-fitted models with comparable free-energy scores and χ2 values. Such uncertainties, primarily in local structures, are partly due to the limited q range available in the SWAXS data. Furthermore, the current SEC-SWAXS measurement resolution, with frame times of 0.5 or 1 frame per second, does not allow for the differentiation of conformational fluctuations within this time resolution. Expanding the q range in SWAXS data and improving the time resolution in SEC-SWAXS data collection are recommended to reduce structural uncertainties and achieve better model convergence in SWAXS-directed molecular dynamics simulations (Hub, 2018[Hub, J. S. (2018). Curr. Opin. Struct. Biol. 49, 18-26.]; Manalastas-Cantos et al., 2021[Manalastas-Cantos, K., Konarev, P. V., Hajizadeh, N. R., Kikhney, A. G., Petoukhov, M. V., Molodenskiy, D. S., Panjkovich, A., Mertens, H. D. T., Gruzinov, A., Borges, C., Jeffries, C. M., Svergun, D. I. & Franke, D. (2021). J. Appl. Cryst. 54, 343-355.]; Sønderby et al., 2017[Sønderby, P., Rinnan, Å., Madsen, J. J., Harris, P., Bukrinski, J. T. & Peters, G. (2017). J. Chem. Inf. Model. 57, 2463-2475.]).

Table 2
NSD from ten best-fitted models (of comparable χ2) of independent runs of the Rosetta-fastsaxs modeling for the samples indicated

Sample NSD
SERF1a 1.29
NT17 0
Htt-0 0.69
Htt-3 0.41
SERF1a/NT17 (1:2) 0.61
SERF1a/Htt-3 (1:1) 0.63

Beyond modeling uncertainties, the intrinsic structural ensembles of the SERF1a–polyQ complex are revealed through their SEC-SWAXS elution profiles, which show varying Rg values [see Figs. 4[link](a) and 5[link](a)]. In this study, we have selectively analyzed data from a finite range of single Rg values. Data showing slowly changing Rg values along the elution profiles, which indicate stable compositions and reveal conformational polymorphism, were not analyzed. However, the minor changes in Rg suggest that conformational inhomogeneities are not significant relative to the representative conformations derived from the data of stable Rg regions. We propose that conformational ensembles of a non-compact biomolecule are better represented using a complete set of models constructed from the corresponding SEC-SWAXS data with progressively changing Rg values along the elution profile. In such analyses, maintaining a constant I(0)/C value (concentration-normalized zero-angle scattering intensity) is crucial to ensure there is no aggregation or decomposition, as demonstrated by the composition profiles in Figs. 4[link](a) and 5[link](a). The advantages of using SEC-SWAXS in conjunction with concurrent measurements of UV–Vis light absorption and dRI are particularly evident when analyzing the conformation ensembles of non-compact biomolecules in solution.

4. Conclusions

In this investigation, we establish a comprehensive protocol for SEC-SWAXS data analysis, integrating methodologies from I-TASSER, CRYSOL and Rosetta-fastsaxs. This approach enables the construction of molecular models for the coiled protein and peptides of SERF1a, NT17 fragments, NT17-polyQ peptides, and their complexes. The study further highlights the synergistic benefits of integrating UV–Vis absorption and dRI data with SAXS absolute scattering intensity to determine the composition of the complexes. Accurate composition determination is established as a crucial prerequisite for constructing binding structural models, such as SERF1a with two NT17 fragments and one Htt-3. The molecular models generated through this protocol encapsulate all structural features disclosed by complementary SWAXS and NMR data, supplemented by reduced free-energy considerations, although they may not represent ultimate structures. Consequently, these models serve as valuable references for understanding the global and local structures of coiled proteins and peptides in solution. Based on these models, the study reveals that NT17 fragments exhibit robust binding to both the coil and helical segments on the N-terminal side of SERF1a. Interactions between NT17 and SERF1a diminish as the helical content increases in the NT17-polyQ peptides. This conformation-dependent binding affinity suggests a dynamic association of SERF1a with Httext1 during fibrillization, involving a disorder-to-order transition of the polyQ segments from coil to α-helix.

5. Related literature

The following references are only cited in the supporting information for this article: Goddard & Kneller (2001[Goddard, T. D. & Kneller, D. G. (2001). SPARKY 3. University of California, San Francisco, California, USA.]).

Footnotes

These authors made equal contributions.

Acknowledgements

We thank Dr J.-W. Chang for the help in SWAXS measurements.

Funding information

This work was supported by the National Science and Technology Council, Taiwan (grant No. NSTC 111-2112-M-213-014).

References

First citationArndt, J. R., Chaibva, M. & Legleiter, J. (2015). Biomol. Conc. 6, 33–46.  CrossRef CAS Google Scholar
First citationBaias, M., Smith, P. E. S., Shen, K., Joachimiak, L. A., Żerko, S., Koźmiński, W., Frydman, J. & Frydman, L. (2017). J. Am. Chem. Soc. 139, 1168–1176.  Web of Science CrossRef CAS PubMed Google Scholar
First citationBender, B. J., Cisneros, A. III, Duran, A. M., Finn, J. A., Fu, D., Lokits, A. D., Mueller, B. K., Sangha, A. K., Sauer, M. F., Sevy, A. M., Sliwoski, G., Sheehan, J. H., DiMaio, F., Meiler, J. & Moretti, R. (2016). Biochemistry, 55, 4748–4763.   Web of Science CrossRef CAS PubMed Google Scholar
First citationBizien, T., Durand, D., Roblina, P., Thureau, A., Vachette, P. & Pérez, J. (2016). Protein Pept. Lett. 23, 217–231.  Web of Science CrossRef CAS PubMed Google Scholar
First citationBlanchet, C. E., Spilotros, A., Schwemmer, F., Graewert, M. A., Kikhney, A., Jeffries, C. M., Franke, D., Mark, D., Zengerle, R., Cipriani, F., Fiedler, S., Roessle, M. & Svergun, D. I. (2015). J. Appl. Cryst. 48, 431–443.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationCowieson, N. P., Edwards-Gayle, C. J. C., Inoue, K., Khunti, N. S., Doutch, J., Williams, E., Daniels, S., Preece, G., Krumpa, N. A., Sutter, J. P., Tully, M. D., Terrill, N. J. & Rambo, R. P. (2020). J. Synchrotron Rad. 27, 1438–1446.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationDas, R. & Baker, D. (2008). Annu. Rev. Biochem. 77, 363–382.   Web of Science CrossRef PubMed CAS Google Scholar
First citationFalsone, S. F., Meyer, N. H., Schrank, E., Leitinger, G., Pham, C. L. L., Fodero-Tavoletti, M. T., Holmberg, M., Dulle, M., Scicluna, B., Gesslbauer, B., Rückert, H.-M., Wagner, G. E., Merle, D. A., Nollen, E. A., Kungl, A. J., Hill, A. F., Cappai, R. & Zangger, K. (2012). Cell. Rep. 2, 358–371.  Web of Science CrossRef CAS PubMed Google Scholar
First citationFiumara, F., Fioriti, L., Kandel, E. R. & Hendrickson, W. A. (2010). Cell, 143, 1121–1135.  Web of Science CrossRef CAS PubMed Google Scholar
First citationGoddard, T. D. & Kneller, D. G. (2001). SPARKY 3. University of California, San Francisco, California, USA.  Google Scholar
First citationHub, J. S. (2018). Curr. Opin. Struct. Biol. 49, 18–26.  Web of Science CrossRef CAS PubMed Google Scholar
First citationJacrot, B. (1976). Rep. Prog. Phys. 39, 911–953.  CrossRef CAS Web of Science Google Scholar
First citationJeffries, C. M., Ilavsky, J., Martel, A., Hinrichs, S., Meyer, A., Pedersen, J. S., Sokolova, A. V. & Svergun, D. I. (2021). Nat. Rev. Methods Primers, 1, 70.  Web of Science CrossRef Google Scholar
First citationKaufmann, K. W., Lemmon, G. H., DeLuca, S. L., Sheehan, J. H. & Meiler, J. (2010). Biochemistry, 49, 2987–2998.   Web of Science CrossRef CAS PubMed Google Scholar
First citationKikhney, A. G., Borges, C. R., Molodenskiy, D. S., Jeffries, C. M. & Svergun, D. I. (2020). Protein Sci. 29, 66–75.   Web of Science CrossRef CAS PubMed Google Scholar
First citationKonarev, P. V., Petoukhov, M. V. & Svergun, D. I. (2016). J. Appl. Cryst. 49, 953–960.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationLange, O. F., Rossi, P., Sgourakis, N. S., Song, Y., Lee, H.-W., Aramini, J. M., Ertekin, A., Xiao, R., Acton, T. B., Montelione, G. T. & Baker, D. (2012). Proc. Natl Acad. Sci. USA, 109, 10873–10878.  Web of Science CrossRef CAS PubMed Google Scholar
First citationLin, J.-M., Lin, T.-L., Jeng, U.-S., Huang, Z.-H. & Huang, Y.-S. (2009). Soft Matter, 5, 3913–3919.  Web of Science CrossRef CAS Google Scholar
First citationLiu, D.-G., Chang, C.-H., Chiang, L.-C., Lee, M.-H., Chang, C.-F., Lin, C.-Y., Liang, C.-C., Lee, T.-H., Lin, S.-W., Liu, C.-Y., Hwang, C.-S., Huang, J.-C., Kuan, C.-K., Wang, H.-S., Liu, Y.-C., Tseng, F.-H., Chuang, J.-Y., Liao, W.-R., Li, H.-C., Su, C.-J., Liao, K.-F., Yeh, Y.-Q., Shih, O., Wu, W.-R., Wang, C.-A. & Jeng, U. (2021). J. Synchrotron Rad. 28, 1954–1965.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationManalastas-Cantos, K., Konarev, P. V., Hajizadeh, N. R., Kikhney, A. G., Petoukhov, M. V., Molodenskiy, D. S., Panjkovich, A., Mertens, H. D. T., Gruzinov, A., Borges, C., Jeffries, C. M., Svergun, D. I. & Franke, D. (2021). J. Appl. Cryst. 54, 343–355.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationMatlahov, I. & van der Wel, P. C. (2019). Exp. Biol. Med. (Maywood), 244, 1584–1595.  Web of Science CrossRef CAS PubMed Google Scholar
First citationO'Meara, M. J., Leaver-Fay, A., Tyka, M. D., Stein, A., Houlihan, K., DiMaio, F., Bradley, P., Kortemme, T., Baker, D., Snoeyink, J. & Kuhlman, B. (2015). J. Chem. Theory Comput., 11, 388–398.  Google Scholar
First citationPrior, C., Davies, O. R., Bruce, D. & Pohl, E. (2020). J. Chem. Theory Comput. 16, 1985–2001.  Web of Science CrossRef CAS PubMed Google Scholar
First citationRoy, A., Kucukural, A. & Zhang, Y. (2010). Nat. Protoc. 5, 725–738.   Web of Science CrossRef CAS PubMed Google Scholar
First citationSchroer, M. A. & Svergun, D. I. (2018). Emerg. Top. Life Sci. 2, 69–79.  Web of Science CrossRef CAS PubMed Google Scholar
First citationShih, O., Liao, K.-F., Yeh, Y.-Q., Su, C.-J., Wang, C.-A., Chang, J.-W., Wu, W.-R., Liang, C.-C., Lin, C.-Y., Lee, T.-H., Chang, C.-H., Chiang, L.-C., Chang, C.-F., Liu, D.-G., Lee, M.-H., Liu, C.-Y., Hsu, T.-W., Mansel, B., Ho, M.-C., Shu, C.-Y., Lee, F., Yen, E., Lin, T.-C. & Jeng, U. (2022). J. Appl. Cryst. 55, 340–352.  Web of Science CrossRef CAS IUCr Journals Google Scholar
First citationShih, O., Yeh, Y., Liao, K., Li, K., Tsai, J., Li, C., Chiang, Y., Heenan, R. K., Sun, Y. & Jeng, U. (2023). Mater. Chem. Phys. 308, 128253.  Web of Science CrossRef Google Scholar
First citationShih, O., Yeh, Y.-Q., Liao, K.-F., Su, C.-J., Wu, P.-H., Heenan, R.-K., Yu, T.-Y. & Jeng, U. (2018). J. Phys. Chem. Lett. 9, 4287–4293.  Web of Science CrossRef CAS PubMed Google Scholar
First citationShih, O., Yeh, Y.-Q., Liao, K.-F., Sung, T.-C., Chiang, Y.-W. & Jeng, U.-S. (2017). Phys. Chem. Chem. Phys. 19, 7947–7954.  Web of Science CrossRef CAS PubMed Google Scholar
First citationSønderby, P., Rinnan, Å., Madsen, J. J., Harris, P., Bukrinski, J. T. & Peters, G. (2017). J. Chem. Inf. Model. 57, 2463–2475.  Web of Science PubMed Google Scholar
First citationStovgaard, K., Andreetta, C., Ferkinghoff-Borg, J. & Hamelryck, T. (2010). BMC Bioinformatics, 11, 429.   Google Scholar
First citationSvergun, D., Barberato, C. & Koch, M. H. J. (1995). J. Appl. Cryst. 28, 768–773.   CrossRef CAS Web of Science IUCr Journals Google Scholar
First citationThakur, A. K., Jayaraman, M., Mishra, R., Thakur, M., Chellgren, V. M., Byeon, I. J., Anjum, D. H., Kodali, R., Creamer, T. P., Conway, J. F., Gronenborn, A. M. & Wetzel, R. (2009). Nat. Struct. Mol. Biol. 16, 380–389.  Web of Science CrossRef PubMed CAS Google Scholar
First citationTrewhella, J., Duff, A. P., Durand, D., Gabel, F., Guss, J. M., Hendrickson, W. A., Hura, G. L., Jacques, D. A., Kirby, N. M., Kwan, A. H., Pérez, J., Pollack, L., Ryan, T. M., Sali, A., Schneidman-Duhovny, D., Schwede, T., Svergun, D. I., Sugiyama, M., Tainer, J. A., Vachette, P., Westbrook, J. & Whitten, A. E. (2017). Acta Cryst. D73, 710–728.  Web of Science CrossRef IUCr Journals Google Scholar
First citationTsai, T.-Y., Chen, C.-Y., Lin, T.-W., Lin, T.-C., Chiu, F.-L., Shih, O., Chang, M. Y., Lin, Y. C., Su, A. C., Chen, C. M., Jeng, U. S., Kuo, H. C., Chang, C. F. & Chen, Y. R. (2023). Commun. Biol. 6, 767.  Web of Science CrossRef PubMed Google Scholar
First citationUrbanek, A., Popovic, M., Morató, A., Estaña, A., Elena-Real, C. A., Mier, P., Fournet, A., Allemand, F., Delbecq, S., Andrade-Navarro, M. A., Cortés, J., Sibille, N. & Bernadó, P. (2020). Structure, 28, 733–746.e5.  Web of Science CrossRef CAS PubMed Google Scholar
First citationWu, S., Skolnick, J. & Zhang, Y. (2007). BMC Biol. 5, 17.   Google Scholar
First citationYang, J. & Zhang, Y. (2015a). CP Bioinformatics, 52, 5.   Google Scholar
First citationYang, J. & Zhang, Y. (2015b). Nucleic Acids Res., 43, W174–W181.  Web of Science CrossRef CAS PubMed Google Scholar
First citationYeh, Y.-Q., Liao, K.-F., Shih, O., Shiu, Y.-J., Wu, W.-R., Su, C.-J., Lin, P.-C. & Jeng, U.-S. (2017). J. Phys. Chem. Lett. 8, 470–477.  Web of Science CrossRef CAS PubMed Google Scholar
First citationZhang, Y. (2008). BMC Bioinformatics, 9, 40.   Google Scholar

This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.

IUCrJ
ISSN: 2052-2525