Disordered nanostructure in Huntingtin interacting protein K acts as

2 Graduate Studies, Manipal University, Manipal, Karnataka 576104. INDIA. 3 Present address: National Institute of Child Health and Human Development...
0 downloads 0 Views 7MB Size
Download PDF
Article Cite This: Biochemistry 2018, 57, 2009−2023

pubs.acs.org/biochemistry

Disordered Nanostructure in Huntingtin Interacting Protein K Acts as a Stabilizing Switch To Prevent Protein Aggregation Debasish Kumar Ghosh,†,‡ Ajit Roy,†,§ and Akash Ranjan*,† †

Computational and Functional Genomics Group, Centre for DNA Fingerprinting and Diagnostics, Nampally, Hyderabad 500001, India ‡ Graduate Studies, Manipal University, Manipal, Karnataka 576104, India S Supporting Information *

ABSTRACT: Protein misfolding due to mutation(s) and/or generation of unstable intermediate state(s) can be the cause of aberrant aggregations, leading to cellular degeneration. While molecular signatures like amyloidogenic regions cause aggregation, other features in proteins, like disorder and unique complexity regions, regulate and restrict such adhesive accumulation processes. Huntingtin interacting protein K (HYPK) is an aggregation-prone protein. Using various biophysical, microscopy, and computational techniques, we have deciphered how HYPK’s N-terminal nanodisordered region plays a significant modulatory role in preventing its own aggregation and that of other proteins. HYPK’s C-terminal hydrophobic regions lead to annular oligomerization and intermolecular charge interactions among the residues of low-complexity region (LCR) generate amorphous aggregates. The Nterminal disordered nanostructure loops toward the C-terminus, and a negative charge-rich patch in this region interacts with the LCR to shield LCR’s positive charges. This interaction is required to prevent HYPK aggregation. Loss of this interaction causes partial unfolding of the structured C-terminus, resulting in HYPK’s molten globule-like state and rapid annular oligomerization. The N-terminus also determines the specificity to mediate the differential bindings with aggregation-prone and wild type Huntingtin-exon1 proteins (Huntingtin97Q-exon1 and Huntingtin25Q-exon1). A sliding interaction of the specific N-terminal segment of HYPK along the extended polyglutamine region of Huntingtin-exon1 is responsible for HYPK’s higher affinity for aggregation-prone Huntingtin than for its non-aggregating counterpart. Overall, our study provides evidence of the existence of disordered nanostructure in HYPK protein that mechanistically plays a decisive role in preventing both self and non-self protein aggregation. occur with many proteins like Huntingtin, α-Synuclein, Tau, amyloid β, prions, etc., which cause various degenerative diseases.16 Huntington’s disease, Alzheimer’s disease, and Parkinson’s disease are some common neurodegenerative diseases caused by abnormal protein aggregation.16 Even though they contain aggregation-prone patches, most proteins have evolved to contain stabilizing or aggregationpreventing regions. For example, unstructured regions spanning structural domains are known to create steric hindrance zones that prevent aggregation.17 Different complexity regions in proteins are also known to reduce aggregation propensity.18 Disordered structures in proteins can acquire the form of nanostructure.19,20 Different molecular folding states and interaction modules define these nanostructures in disordered regions.21 From the perspective of sequence, these nanostructures can also be low-complexity stretches that can assemble into fibrillar structures.22 Though functions of the disordered

P

rotein aggregates represent supramolecular structures that have escaped the cellular quality-control mechanisms.1,2 These protein aggregates result from various processes, like mutations in amino acid sequence(s),3 aberrant translation failure associated with compromised chaperone functioning,4 and generation of transient unstable forms (like unfolded or molten globule-like states) due to changes in biochemical microenvironments.5 Morphologically, protein aggregates can assume different structural forms, like annular oligomers,6 fibrils,7 and amorphous aggregates.8 Association of these aggregates can constitute specific cellular bodies like plaques,9 inclusion bodies,9 and stress granules10 that serve as the hubs for sequestration of various other cellular proteins to form heterogeneous co-aggregates of proteins.11 The underlying mechanism of protein aggregation relies on the stability of its secondary structures. While in many proteins, the excess content of α helices results in the collapse of structures12 due to hydrophobic associations,13 some proteins aggregate due to β sheet-mediated complementary structure- and shape-based association.14 Aggregation-prone regions in proteins are often termed as amyloidogenic regions.15 Aggregation is known to © 2018 American Chemical Society

Received: August 11, 2017 Revised: January 28, 2018 Published: January 30, 2018 2009

DOI: 10.1021/acs.biochem.7b00776 Biochemistry 2018, 57, 2009−2023

Article

Biochemistry

products and empty plasmid vectors were restriction digested by specific restriction enzymes followed by ligation of each vector with the cognate PCR product and finally transformation of ligation products in the ultracompetent DH5α Escherichia coli strain [F− Φ80lacZΔM15 Δ(lacZYA-argF) U169 recA1 endA1 hsdR17 (rK−, mK+) phoA supE44 λ− thi-1 gyrA96 relA1]. Positive clones were screened by colony PCR, and all clones were sequenced at the Research Support Service Group of CDFD. Htt97Qexon1-GFP and Htt25Qexon1-GFP constructs were kindly provided by R. R. Kopito. Htt97Qexon1 and Htt25Qexon1 were recloned in pET21b by the method described above. Recombinant Protein Production. Different recombinant proteins were expressed and purified from the E. coli BL21DE3 strain [fhuA2 [lon] ompT gal (λ DE3) [dcm] ΔhsdSλ DE3 = λ sBamHIo ΔEcoRI-B int::(lacI::PlacUV5::T7 gene1) i21 Δnin5] by a T7-based expression system. Clone-transformed cells were grown in LB/ampicillin medium as consecutive primary and secondary cultures using conventional methods.29 Cells were then lysed [lysis buffer consisting of 50 mM NaH2PO4 (pH 8.0), 300 mM NaCl, 10 mM imidazole, and 1 mM phenylmethanesulfonyl fluoride], and lysates were subjected to protein purification by metal ion (Ni2+) affinity column chromatography. After being washed [wash buffer consisting of 50 mM NaH2PO4 (pH 8.0), 300 mM NaCl, and 40 mM imidazole], proteins were eluted in elution buffer [50 mM NaH2PO4 (pH 8.0), 300 mM NaCl, and 300 mM imidazole]. Protein concentrations were measured by Bradford assays. Eluted proteins were dialyzed in various dialysis buffers according to downstream experimental requirements. Dialysis and concentration of proteins were performed in centrifugal filter units (of various cutoff sizes) using the repeated buffer exchange method. Atomic Force Microscopy. In atomic force microscopy (AFM), purified proteins were kept in 20 mM NaH2PO4 buffer (pH 7.5). Proteins were drop-casted over a freshly cut mica sheet followed by uniform spreading and air-drying in a dustfree chamber at room temperature. Protein concentrations varied in different experiments as mentioned in the respective sections of results. Samples were washed twice with Milli-Q water followed by air-drying and final drying using a gentle stream of nitrogen gas. Images were captured in a Bruker Dimension Icon atomic force microscope (ex situ tapping mode). The following AFM parameters were used: scan rate of 0.653−0.712 Hz, amplitude set point of 260−315 mV, and drive amplitude of 2.3−3.1 V. Image processing was performed in Nanoscope analysis and Gwyddion30 software. Cell Culture and Confocal Laser Scanning Microscopy. HeLa cells were procured and validated from the National Centre for Cell Science (NCCS, Pune, India). Cells were maintained in DMEM complemented with 10% FBS, 2 mM glutamine, and an antibiotic/antimycotic cocktail. Cells were cultured at 37 °C and 5% CO2 level in a humified incubator. Transfections of clones in cells were performed with Lipofectamine 2000 (Thermo Fischer Scientific) and opti-MEM (Gibco) following the manufacturer’s protocol. Fluorescence imaging was performed by the usual method of fixation and preparation. Briefly, cells were washed with phosphate-buffered saline (PBS) followed by fixation with 4% paraformaldehyde (pH 7.4). Fixed cells were visualized for GFP and mCherry fluorescence. Images were acquired with an LSM700 META (Zeiss) confocal laser scanning microscope

regions in maintaining protein stability are reported in various studies, mechanistic understanding of such regulations is still considered to be both intriguing and necessary. In this study, we have elucidated that the N-terminal intrinsically unstructured region in Huntingtin interacting protein K (HYPK) has nanodimensions and acts as stabilizing switch to prevent its own aggregation as well as that of other proteins. HYPK is a small (129-amino acid, 14.66 kDa) aggregationprone protein. It is known to interact with the N-terminus of Huntingtin protein,23 resulting in the reduction of Huntingtin aggregates in the cell.24 HYPK is also reported to interact with N-acetyltransferase (Nat) proteins25 and eEF1A at the ribosome.26 These interactions help HYPK to modulate nascent polypeptide acetylation function as well as assisting in the ribosomal chaperone’s activity. HYPK associates with many other cytoplasmic and nuclear proteins to help in the cell survival process.26 Our other works suggest that HYPK has a strong tendency to co-accumulate with several other aggregation-prone proteins like α-Synuclein-A53T and Superoxide dismutase1-G93A. HYPK’s co-aggregates can cause proteasomal blockades.27 In different studies, we have found that HYPK forms specific annular structures, known as Hgranules, which form a sequestration complex platform for different aggregates of Huntingtin97Q-exon1, α-synucleinA53T, and superoxide dismutase1-G93A. In Caenorhabditis elegans, HYPK has been shown to directly influence the cellular aging process.28 In this study, we have characterized the N-terminus of HYPK as a disordered nanostructure. This nanoregion can prevent HYPK’s self-aggregation, which is caused by C-terminal hydrophobic and low-complexity stretches. With the help of various biophysical techniques (like circular dichroism spectroscopy and dynamic light scattering), microscopy (like atomic force microscopy and confocal fluorescence microscopy), and computational approaches (like molecular modeling and simulation), we have been able to demonstrate the essential function of HYPK’s nanodisordered region in the prevention of both of its own aggregation and that of other proteins. The inhibitory action of disordered nanostructure can be observed against both annular and amorphous aggregates of HYPK. The N-terminal region of HYPK contains a negative charge-rich patch that interacts with the C-terminal low-complexity region. This interaction prevents HYPK’s partial unfolding to the molten globule state. The strong hold of this charge interaction is effective against HYPK’s self-association. This disordered nanostructure prevents HYPK’s self-aggregation and helps in high-affinity binding with aggregation-prone Huntingtin via multiple sliding interactions along the expanded polyglutamine region. The latter interaction finally leads to enhanced stabilization of Huntingtin-exon1.

■

METHODS All experiments were performed in triplicate. Cloning. To clone HYPK and its different deletion mutants, cDNA of HYPK was generated by reverse transcription polymease chain reaction (RT-PCR) from HeLa cell-extracted total RNA. The full length open reading frame (ORF) of HYPK and its deletion mutants were amplified by PCR from HYPK cDNA by specific primer sets (primer sequences are available on request). PCR products were cloned in different bacterial and mammalian cell line expression vectors, like pET21b(+) (Novagen), pEGFPN1 (CLONTECH Laboratories Inc.), and pcDNA3.1(+) (Life Technologies). PCR 2010

DOI: 10.1021/acs.biochem.7b00776 Biochemistry 2018, 57, 2009−2023

Article

Biochemistry

the description of mutants) with the C-terminal region of HYPK (HYPK-C69). SPR interaction studies were conducted in a Biacore 3000 system. For this, 20 μg of HYPK-C69 was immobilized on a CM5 chip through NHS/EDC coupling reagent. The N-terminal region of HYPK or its mutants were used as analytes, and their concentrations were kept in the range of 40 nM to 25 μM. In control runs, the shamimmobilized surface in the other channel was monitored to determine the reference index change by nonspecific binding. Analytes were kept in HBS-EP buffer. Structure Modeling and Molecular Dynamics Simulation. The advanced homology modeling program in Modeler31 was used to create the model structure of HYPK using the archaeal nascent polypeptide-associated complex [Protein Data Bank (PDB) entry 1TR8, chain A32] as the template. The quality of the model structure (Ramachandran plot) was analyzed with Procheck.33 Addition of missing hydrogen atoms and/or bonds and bond order assignments were performed in the protein preparation wizard of Maestro 9.2 (Schrödinger Inc.) [OPLS_2005 force field and convergence heavy atom root-mean-square deviation (RMSD) of 0.30 Å]. Structures were minimized with the Molecular Modeling Tool Kit (MMTK)34 with the Amber configuration35 and 103 iterations of steepest descent followed by conjugate gradient minimization. Huntingtin structures (Huntingtin Nterminal 17-amino acid region with 13Q or 23Q) were taken from the PDB (entries 3IO4 and 4FE8, respectively). Unconstrained docking of HYPK on Huntingtin structures were performed in the Piper program36 of the Bioluminate suite (Maestro 9.2, Schrödinger Inc.). Among the docked structures, the best posed outputs were curated and further prepared as described for the HYPK monomer. Molecular dynamics simulations (MDS) of HYPK and HYPK−Huntingtin complexes were performed in Desmond (Maestro 9.2, Schrödinger Inc.). Prepared protein structures were placed in a 40 Å side length cuboidal virtual water environment (TIP3P model37) and a 150 mM Na+/Cl− salt concentration using the system builder setup (OPLS_2005 force field, minimization of simulation volume and protein’s charge neutralization by opposite ions). Structures were then minimized by a 1 kcal mol−1 Å−1 convergence threshold steepest descent minimization (2000 iterations). Complete MDS was divided into two stages. The brief presimulations or the equilibration simulation periods were meant to complete the relaxation of structures, and they consisted of the following steps: restrained minimization of solutes, unrestrained minimization, Berendsen constant particle−volume−temperature (NVT) simulation with solute heavy atom restraints, Berendsen constant particle−pressure−temperature (NPT) simulation with solute heavy atom restraints, and Berendsen unrestrained simulation. Each of these equilibrating simulations was run for 2 ns (100 ps to 2 ns equilibrating simulations were adequate for intrinsically unstructured proteins). The complete 300 ns simulation of HYPK−Huntingtin complexes and the 50 ns full simulation of HYPK were unrestrained with the following defined conditions: constant particle−temperature−pressure ensemble class, 25 °C temperature, thermostat (Nose-Hoover chain method38 with a 1 ps relaxation time), barostat (Martyna−Tobias−Klein method39 to maintain a 1.01325 bar pressure with a 2 ps relaxation time and isotropic coupling), and RESPA integrator (6 fs far, 2 fs near, and 2 fs bonded).40,41 A randomized velocity-coupled OPLS_2005 force field was used, and interactions were regulated by 1 nm short-range

with 63× Plan Apo/1.4 NA oil immersion objectives. Image processing was performed in Zen-lite software. Dynamic Light Scattering. For dynamic light scattering, HYPK protein was kept in 50 mM NaH2PO4 (pH 7.5) and 50 mM NaCl buffer, and different protein concentrations were used in different experiments (mentioned in the respective sections of results). Particle sizes and diameters were measured by Rayleigh light scattering in a Malvern particle size analyzer (ZEN 3690 ZETASIZER NANO ZS 90, version 7.03). Data were analyzed with 21CFR Part11 software. 1-Anilinonaphthalene-8-sulfonic Acid (ANS) Titration and ANS Binding Assays. Fluorescence spectroscopy was used for ANS binding and titration assays. Proteins [HYPK, HYPK N-terminal 60-residue region (HYPK-N60), and HYPK C-terminal 69-residue region (HYPK-C69)] were kept in 20 mM potassium phosphate buffer (pH 7.0). A 20 mM ANS stock solution was prepared in 20 mM potassium phosphate buffer (pH 7.0), and required volumes of this stock were added to reaction mixtures to achieve final concentrations in the range of 0−240 μM. In every reaction, the final protein concentrations were kept at 15 μM. In pH-dependent studies, HYPK (1 μM) was kept in phosphate-citrate buffer of varying pH (2.6−8.0) and the ANS concentration in the final reaction mixture was 100 μM. ANS fluorescence spectra were recorded (in a 10 mm quartz cell) at 25 °C within the emission ranges of 450−500 and 450−600 nm (excitation wavelength of 380 nm) in a Hitachi F-7000 fluorescence spectrophotometer. Fluorescence intensity values at λemission values of 470 and 450−600 nm were used for calculations. All the readings were taken in triplicate immediately after the addition of ANS. The ANS fluorescence intensity due to protein binding was calculated with the equation F = (Fprotein/ANS − FANS)C − (Fprotein − Fbuffer)

where Fprotein/ANS is the fluorescence intensity of protein-bound ANS, FANS is the fluorescence intensity of free ANS in buffer, Fprotein is the fluorescence intensity of free protein, Fbuffer is the fluorescence intensity of buffer, and C is the inner filter correction factor that was calculated with the equation

C = 10 AANS /2 where AANS is the absorbance of the solution in a 1 cm quartz cell measured in an Umchem ultraviolet−visible scanning spectrophotometer. The corrected fluorescence is calculated with the equation Fcorr = FC

Isothermal Titration Calorimetry. In isothermal titration calorimetry experiments, both interacting proteins (HYPK and Htt97Qexon1/Htt25Qexon1) were kept in PBS (pH 7.4). All the thermal titrations for bindings were performed at 25 °C in a MicroCal iTC 200 (GE Healthcare) instrument fitted with a MicroCal Thermovac temperature-control thermostat system. The following instrument parameters were used in the experiments: 20 total injection/binding analyses, reference power of 10 μcal/s, initial delay of 60 s, syringe protein (HYPK) concentration of 100 μM, cell protein (Htt97Qexon1 or Htt25Qexon1) concentration of 10 μM, and syringe speed of 300 rpm Surface Plasmon Resonance. Surface plasmon resonance (SPR) experiments were performed to understand the binding interactions of the N-terminus of HYPK or its mutants (HYPKN60, HYPK-N60 E-D/A, and HYPK-N60 E/D; see Results for 2011

DOI: 10.1021/acs.biochem.7b00776 Biochemistry 2018, 57, 2009−2023

Article

Biochemistry

Figure 1. HYPK is aggregation-prone. The N-terminus of HYPK prevents its C-terminus mediated aggregation in vitro and in vivo. (A) Atomic force microscopy images of HYPK and its different deletion mutants. Depending upon concentration, HYPK forms annular and amorphous aggregates. HYPK’s C-terminal 45-residue region (HYPK-C45) forms annular oligomers; its C-terminal 69-residue region (HYPK-C69) produces amorphous aggregates, and its N-terminal 60-residue region (HYPK-N60) does not form aggregates. (B) In vivo ectopic expression of HYPK and its C-terminal regions (in HeLa cells). Though HYPK does not form aggregates in the majority of cells, some cells show small scattered aggregates. HYPK-C45 leads to globular type aggregates, and HYPK-C69 forms large clumped aggregates. HYPK-N60 does not produce aggregates. (C) HYPK-N60 prevents the formation of large spheroid type aggregates by HYPK-C69. (D) AFM studies show that particle sizes (heights) of HYPK-C69 do not increase over time when it is mixed with equimolar HYPK-N60 (concentrations of both HYPK-C69 and HYPK-N60 are 200 nM). (E) Dynamic light scattering shows inhibition of formation of large particles of HYPK-C69 upon incubation with equimolar HYPK-N60.

cutoff interaction/smooth particle mesh Ewald long-range42 coulombic interaction (1 × 10−9 Ewald tolerance). In every direction, periodic boundary conditions were used. Simulation interaction diagrams were generated in Maestro 9.2 (Schrödinger Inc.), and structure visualization/trajectory analysis was performed in Pymol43 and VMD.44 Analysis of water-exposed surfaces areas (solvent accessible surface area and hydrophobic surface area), folding free energy calculations, and H-bond identifications of different structures were performed in Maestro 9.2 (Schrödinger Inc.). Secondary structure analyses of the native state and molten globule state of HYPK were conducted using algorithms of the 2struc server.45 Conformation entropies of residues were calculated using PLOPS.46 Circular Dichroism Spectroscopy. In circular dichroism (CD) spectroscopy, proteins were kept in 50 mM NaH2PO4 (pH 7.5) and 50 mM NaF buffer. Measurements were taken in

a JASCO 810 spectropolarimeter with a quartz cuvette of 10 mm path length. For secondary structure analysis, CD spectra were recorded in the far-ultraviolet (far-UV) region of light (190−260 nm) at room temperature. In slow thermal denaturation experiments (20−90 °C, 5 °C temperature interval), proteins were stabilized at different discrete temperatures followed by the acquisition of CD spectra at the far-UV wavelength (190−260 nm). To determine the melting temperatures (Tm) of different proteins, they were continuously heated (1 °C/min) and Δε values at a λ = 222 nm were collected (Δε is the change in ellipticity). Tm values were determined by plotting dΔε/dT (dθ/dT) versus temperature (T is the temperature in kelvin). In pH-dependent CD experiments, protein was kept in citrate phosphate buffer of different pH values (2.6−8.0), and the protein’s secondary structure contents were analyzed by recording CD spectra in the far-UV range at room temperature. Deconvolutions of 2012

DOI: 10.1021/acs.biochem.7b00776 Biochemistry 2018, 57, 2009−2023

Article

Biochemistry

expressed with HYPK-N60, it did not produce highly clumped aggregates as it did when it was expressed alone. Instead, it produced small scattered aggregates in very few cells (Figure 1C). The N-Terminus of HYPK Is a Disordered Nanostructure. Taking note of the fact that the N-terminus of HYPK prevented the self-association of HYPK, we investigated the structural features within the N-terminus to understand the mechanism of aggregation modulation. HYPK had a bipartite structure. Structural studies by segmental deletion of HYPK along with CD spectroscopy revealed that the C-terminus of HYPK was mostly structured (α helical) whereas the Nterminus was disordered (Figure 2A). Molecular modeling of HYPK also showed the unstructured nature of the N-terminus (Figure 2B). Unrestrained molecular dynamics simulation of the HYPK model structure in an aqueous environment showed sporadic loss of secondary structure and an increase in the proportion of the unstructured region during the simulation time (Figure 2C). However, the structure remained fairly stable throughout the 50 ns simulation time window. This was reflected by the stabilization of the RMSD and lower RMSF values of the protein backbone (Figure 2D). During the simulation, this stability was maintained by keeping cumulatively nearly equal numbers of intra- and intermolecular hydrogen bonds (Figure 2E). Because this unstructured region contained only 50−60 amino acids, we believed that this region could assume nanostructure. An AFM study actually showed the nanodimension of HYPK-N60 (Figure 2F), with most of the particles showing heights of 5 μM) (Figure 1A). The Cterminal ubiquitin-associated (UBA)-like domain of HYPK [Cterminal 45-residue region (HYPK-C45)] was found to be responsible for the formation of annular structures. The surface-exposed hydrophobic residues in this region played an important role in annular oligomerization (Figure 1A,B). The intermolecular hydrophobic interactions between HYPK molecules led to the deposition of concentric rimlike arrays of HYPK seeds during formation of annular structures. Upstream of the UBA-like domain, a small charge-rich lowcomplexity region (LCR) was responsible for the collapse of annular complexes that finally led to the formation of amorphous aggregates. Because the LCR contained a large number of charged residues, it was apparent that electrostatic charge interactions between intermolecular LCRs caused random association of HYPK’s C-terminal LCR-containing 69-residue region (HYPK-C69) (Figure 1A,B). The N-terminal region (HYPK-N60) lacked any potential to make positive contributions to HYPK aggregation (Figure 1A,B). Though the endogenous expression level of HYPK in vivo did not lead to aggregates in a significant number of cells, ectopic expression of HYPK showed an increase in the number of cells with aggregates (10−15%). However, the higher proportion of cells (85−90%) still showed no induction of aggregates (Figure 1B). Considering the fact that the C-terminal region of HYPK was strongly prone to aggregation, we asked what molecular features in HYPK stabilized it and prevented it from undergoing aggregation. We tested if the N-terminus of HYPK has intrinsic properties to suppress HYPK aggregation. Indeed, it appeared that the Nterminus had preventive properties against HYPK’s selfaggregation. The aggregation potential of HYPK-C69 was reduced when it was co-incubated with equimolar HYPK-N60 (Figure 1C,D). HYPK-C69 could form large soluble oligomers and/or aggregates in the size (diameter) range of 100 nm to 1 μm even at low concentrations (100−500 nM). The population of such large and globular aggregates drastically decreased when they were mixed with HYPK-N60. In this case, soluble assemblies were small granular and mostly in size range of 10−100 nm (Figure 1C,E). In vivo, when HYPK-C69 was co2013

DOI: 10.1021/acs.biochem.7b00776 Biochemistry 2018, 57, 2009−2023

Article

Biochemistry

positively charged residues (arginine and lysine) to alanine in the LCR (HYPK R-K/A) also showed a significant decrease in the extent of aggregation (Figure 3B,C). These observations suggested that the negatively charged patch at the N-terminus might be involved in electrostatically interacting with the Cterminal LCR. This interaction prevented LCR-mediated aggregation propagation under physiological conditions (Figure 3D). Studies of the binding of the N-terminal 60-residue region (HYPK-N60) and its multiple-point mutant variants (HYPKN60 E-D/A and HYPK-N60 E/D) with the C-terminal 69residue region (HYPK C-69) showed specific interactions of HYPK-N60 and HYPK-N60 E/D with HYPK C-69, but no interaction was observed between HYPK-N60 E-D/A and HYPK-C69 (Figure 3E). This confirmed that specific charge interactions existed between N-terminal negatively charged residues and the (basic amino acids of) C-terminal LCR. We further tested if this interaction was responsible for prevention of aggressive HYPK oligomerization. Had these intramolecular charge interactions inhibited aggregation of HYPK, we expected that altered pH conditions would perturb such interactions, leading to the accumulation of larger particles. The shift of pH from 7.5 to 3 actually showed a gradual increase in HYPK particle size (Figure 3F). This finally proved that intramolecular charge interactions, mediated by the protein’s disordered nanostructure, acted not only as a molecular stabilizer but also as an anti-aggregation factor. Loss of Charge Interactions between the LCR and the N-Terminal Disordered Nanoregion Leads to Formation of HYPK’s Molten Globule-like State and Subsequent Aggregation. To analyze if HYPK existed in any conformational intermediate that drove its oligomerization, slow thermal denaturation experiments were performed. Steady state thermodynamic unfolding experiments were performed by gradual thermal denaturation of native HYPK protein, HYPKN60, and HYPK-C69 and by taking continuous CD spectra in far-UV range (190−260 nm) at different discrete temperatures (20−90 °C, 5 °C temperature interval). This was followed by deconvoluting sets of spectra by the convex constraint algorithm (CCA). Deconvolution of spectral sets of HYPK and HYPK-C69 into three component basic sets showed the existence of a (partially) folded intermediate in both of them (Figure 4A,B). However, deconvolution of the HYPK-N60 spectral set resulted in three curves; two of them were nearly identical (merged), implying the non-existence of folding intermediate(s) arising due to this region (Figure 4C). These results proved the existence of a C-terminally mediated molten globule-like intermediate state in HYPK. The high content of charged amino acids in HYPK (E, 18.6%; D, 5.4%; R, 10.1%; K, 7.0%) encouraged us to investigate if the previously described intramolecular charge interactions played any modulatory role in the formation of the molten globule state. Under varying pH conditions, CD spectra of HYPK showed changes in secondary structure. CCA threecomponent deconvolution of the set of pH-regulated CD spectra showed the specific existence of a molten globule state within the pH range of 3−5 (Figure 4D). Because the change in the medium’s pH altered the protein’s polar (ionic and Hbond) interactions, we surmised that the molten globule state of HYPK had partially unfolded structure that arose due to modulation of intramolecular charge interactions in HYPK. Alterations in pH changed the charge interactions between the LCR-residing positive charges and N-terminal negative charges.

Figure 2. N-Terminus of HYPK is disordered and has nanostructure type features. (A) CD spectroscopy shows that HYPK is structurally divided into two regions. The C-terminus is mostly α helically structured, and the N-terminus is disordered. (B) The molecular model of HYPK shows a prolonged unstructured region at the Nterminus. (C) A 50 ns molecular dynamics simulation of HYPK shows sporadic loss of secondary structure composition along the simulation time course (red arrows). (D) RMSD and RMSF values of the HYPK backbone represent stabilization of structure with time. (E) Numbers of hydrogen bonds formed by HYPK and radii of gyration of HYPK at different time points of the 50 ns molecular dynamics simulation. (F) AFM of HYPK-N60 shows that it has nanodimension. Most of the HYPK-N60 particle’s sizes are