Disordered nanostructure in Huntingtin interacting protein K acts as

21 hours ago - ... compared to its non-aggregating counterpart. Overall, our study provides evidence for the existence of disordered nanostructure in ...
2 downloads 7 Views 3MB Size
Download PDF
Subscriber access provided by FLORIDA STATE UNIV

Article

Disordered nanostructure in Huntingtin interacting protein K acts as stabilizing switch to prevent protein aggregation Debasish Kumar Ghosh, Ajit Roy, and Akash Ranjan Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.7b00776 • Publication Date (Web): 30 Jan 2018 Downloaded from http://pubs.acs.org on January 30, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Disordered nanostructure in Huntingtin interacting protein K acts as stabilizing switch to prevent protein aggregation Debasish Kumar Ghosh1,2, Ajit Roy1,3 and Akash Ranjan1,* 1

Computational and Functional Genomics Group Centre for DNA Fingerprinting and Diagnostics Nampally, Hyderabad 500001. INDIA

2

Graduate Studies, Manipal University, Manipal, Karnataka 576104. INDIA

3

Present address: National Institute of Child Health and Human Development National Institute of Health Building 6A, 6 Center drive Bethesda, MD 20892-2790 United States of America * Corresponding Author

1 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

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 nano-disordered region plays a significant modulatory role in preventing its own and other protein’s aggregation. HYPK’s C-terminal hydrophobic regions lead to annular oligomerization and inter-molecular charge-interactions among the residues of low complexity region (LCR) generate amorphous aggregates. The N-terminal disordered nanostructure loops towards 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 (Huntingtin97Qexon1 and Huntingtin25Qexon1). A sliding interaction of the specific N-terminal segment of HYPK along the extended poly-glutamine region of Huntingtin-exon1 is responsible for HYPK’s stronger affinity towards aggregation-prone Huntingtin compared to its non-aggregating counterpart. Overall, our study provides evidence for the existence of disordered nanostructure in HYPK protein that mechanistically plays a decisive role in preventing both self and non-self protein aggregation.

TABLE OF CONTENT GRAPHIC

2 ACS Paragon Plus Environment

Page 3 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

INTRODUCTION Protein aggregates represent supramolecular structures that have escaped the cellular quality-control mechanisms (1, 2). These protein aggregates result from various means, like mutations in amino acid sequence(s) (3), aberrant translation failure associated with compromised chaperone functioning (4), generation of transient unstable forms (like unfolded or molten globule-like states) due to changes in biochemical micro-environments (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 granules (10) which 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 alpha helices result in the collapse of structures (12) due to hydrophobic associations (13); some proteins aggregate due to beta sheet mediated complementary structure/shape based association (14). Aggregationprone regions in proteins are often termed as amyloidogenic regions (15). Aggregation is known for many proteins like Huntingtin, α-Synuclein, Tau, Amyloid beta, Prions etc., which cause various degenerative diseases (16). Huntington’s disease, Alzheimer’s disease, and Parkinson’s disease are some common neurodegenerative diseases manifested by abnormal protein aggregations (16). Even though containing aggregation-prone patches, most proteins have evolved to contain stabilizing or aggregation preventing regions. For example, unstructured regions spanning between structural domains are known to create steric hindrance zones which 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). In the sequence perspective, these nanostructures can also be low complexity stretches which can assemble into fibrillar structures (22). Though functions of the disordered regions in maintaining protein stability are reported in various studies, mechanistic understanding of such regulations are 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 nano-dimension and it acts as stabilizing switch to prevent its own as well as other protein’s aggregation. Huntingtin Interacting Protein K (HYPK) is a small (129 amino acid, 14.66 kDa) aggregation-prone 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-acetyl transferase (Nat) proteins (25) and eEF1A at ribosome (26). These interactions help HYPK in modulating nascent polypeptide acetylation function as well as assisting in ribosomal chaperone’s activity. HYPK associates with many other cytoplasmic and nuclear proteins to help in cell survival process (26). Our other works suggest that HYPK has a high tendency to co-accumulate with several other aggregation-prone proteins like α-Synuclein-A53T and Super oxide 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 H-granules, which form sequestration complex platform for different aggregates of Huntingtin97Qexon1, αSynuclein-A53T and Super oxide dismutase1-G93A. In Caenorhabditis elegans, HYPK has been shown to directly influence 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 3 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 27

spectroscopy, dynamic light scattering), microscopy (like atomic force microscopy, confocal fluorescence microscopy) and computational approaches (like molecular modeling, simulation), we have been able to demonstrate the essential function of HYPK’s nano-disordered region towards prevention of both of its own and other protein’s aggregations. Inhibitory action of disordered nanostructure is observable against both annular and amorphous aggregates of HYPK. 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 molten globule state. Strong holding of this charge-interaction is effective against HYPK’s self-association. Not only this disordered nanostructure prevents HYPK’s selfaggregation, but it also helps in high-affinity binding with aggregation-prone Huntingtin by multiple sliding interactions along the expanded poly-glutamine region. The later interaction finally leads to enhanced stabilization of Huntingtin-exon1.

METHODS All experiments were done in triplicate. Cloning To clone HYPK and its different deletion mutants, cDNA of HYPK was generated by RT-PCR from HeLa cell extracted total RNA. Full-length ORF of HYPK and its deletion mutants were PCR amplified 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 products and empty plasmid vectors were restriction digested by specific restriction enzymes followed by ligation of each vector with cognate PCR product and finally transformation of ligation products in ultra-competent 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 Research Support Service Group of CDFD. Htt97Qexon1-GFP and Htt25Qexon1-GFP constructs were kindly given by Prof. Ron R. Kopito. Htt97Qexon1 and Htt25Qexon1 were re-cloned in pET21b by the above described method. Recombinant protein production Different recombinant proteins were expressed and purified from Escherichia coli BL21DE3 strain [fhuA2 [lon] ompT gal (λ DE3) [dcm] ∆hsdSλ DE3 = λ sBamHIo ∆EcoRI-B int::(lacI::PlacUV5::T7 gene1) i21 ∆nin5] by 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: 50 mM NaH2PO4 (pH-8.0), 300 mM NaCl, 10 mM Imidazole, 1 mM PMSF] and lysates were subjected to protein purification by metal ion [Ni+2] affinity column chromatography. After washing [wash Buffer: 50 mM NaH2PO4 (pH-8.0), 300 mM NaCl, 40 mM Imidazole], proteins were eluted in elution buffer [elution Buffer: 50 mM NaH2PO4 (pH-8.0), 300 mM NaCl, 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 concentrating of proteins were done in centrifugal filter units (of various cutoff sizes) using 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 freshly-cut mica sheet followed by uniform spreading and air-drying in 4 ACS Paragon Plus Environment

Page 5 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

dust-free chamber at room temperature. Protein concentrations varied in different experiments as mentioned in the respective results. Samples were washed twice with Milli-Q water followed by airdrying and final drying using a gentle stream of nitrogen gas. Images were captured in Bruker Dimension Icon atomic force microscope (ex situ tapping mode). Following AFM parameters were used – scan rate: 0.653 – 0.712 Hz, amplitude set point: 260mV – 315mV, drive amplitude: 2.3V – 3.1V. Image processing was done in Nanoscope analysis and Gwyddion (30) software. Cell culture and confocal laser scanning microscopy HeLa cells were procured and validated from National Centre for Cell Sciences (NCCS, India). Cells were maintained in DMEM medium complemented with 10% FBS, 2 mM glutamine and antibiotic-antimycotic cocktail. Cells were cultured at 37oC, 5% CO2 level in humified incubator. Transfections of clones in cells were done by Lipofectamine 2000 (Thermofischer Scientific) and opti-MEM (Gibco) following manufacturer’s protocol. Fluorescence imaging was performed by usual method of fixation and preparation. Briefly, cells were washed with PBS followed by fixation with 4% paraformaldehyde [pH – 7.4]. Fixed cells were visualized for GFP and mCherry fluorescence. Images were acquired in LSM700 META (Zeiss) confocal laser scanning microscope with 63X Plan Apo/1.4 NA oil immersion objectives. Image processing was done in Zen-lite software. Dynamic light scattering For dynamic light scattering, HYPK protein was kept in 50 mM NaH2PO4 (pH – 7.5), 50 mM NaCl buffer and different protein concentrations were used in different experiments (mentioned in respective results). Particle size/diameters were measured by Rayleigh light scattering in Malvern particle size analyzer (ZEN 3690 ZETASIZER NANO ZS 90, version 7.03). Data were analyzed in 21CFR Part11 software. ANS titration and ANS binding assays Fluorescence spectroscopy was used for ANS (1-Anilinonaphthalene 8-Sulfonic acid) binding/titration assays. Proteins [HYPK, HYPK N-terminal 60 residue region (HYPK-N60) and C-terminal 69 residue region (HYPK-C69)] were kept in 20 mM potassium phosphate buffer (pH-7.0). 20 mM ANS stock solution was prepared in 20 mM potassium phosphate buffer (pH-7.0) and required volumes from this stock were added in reactions to make final concentrations in 0-240 μM range. In every reaction, the final concentrations of proteins 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 ANS concentration in the final reaction was 100 μM. ANS fluorescence spectra were recorded (in 10mm quartz cell) at 25˚C within emission range 450500nm and 450-600nm (excitation wavelength 380nm) in Hitachi F-7000 Fluorescence Spectrophotometer. Fluorescence intensity values at λemission = 470nm and λemission = 450-600nm were used for calculations. All the readings were taken in triplicate immediately after ANS addition. ANS fluorescence intensity due to protein binding was calculated by 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 and Fbuffer is the fluorescence intensity of buffer. ‘C’ is inner filter correction factor which was calculated by equation, C = 10

AANS/2

5 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 27

AANS is absorbance of solution in 1cm quartz cell measured in Umchem UV-Vis scanning spectrophotometer. Corrected fluorescence is calculated by equation, Fcorr = F × C Isothermal titration calorimetry In isothermal titration calorimetry experiments, both interacting proteins (HYPK and Htt97Qexon1/ Htt25Qexon1) were kept in PBS buffer (pH – 7.4). All the thermal titrations for bindings were performed at 25oC in MicroCal iTC 200 (GE Healthcare) instrument fitted with MicroCal Thermovac temperature control thermostat system. Following instrument parameters were used in the experiments – total injection/ binding analysis: 20, reference power (µcal/sec): 10, initial delay: 60 seconds, syringe protein (HYPK) concentration: 100 µM, cell protein (Htt97Qexon1 or Htt25Qexon1) concentration: 10 µM, syringe speed: 300rpm. Surface plasmon resonance Surface plasmon resonance (SPR) experiments were done to understand the binding interactions of HYPK N-terminus or its mutants (HYPK-N60, HYPK-N60 E-D/A, HYPK-N60 E/D; see results for the description of mutants) with the C-terminal region of HYPK (HYPK-C69). SPR interaction studies were conducted in Biacore 3000 system. For this, 20 µg of HYPK-C69 was immobilized on CM5 chip through NHS/EDC coupling reagent. HYPK N-terminal region or its mutants were used as analytes and their concentrations were kept in the range of 40nM to 25µM. In control runs, sham immobilized surface in other channel was monitored to determine the reference index change by non-specific bindings. Analytes were kept in HBS-EP buffer. Structure modeling and molecular dynamics simulation Advanced homology modeling program in Modeller (31) was used to create the model structure of HYPK using archaeal nascent polypeptide associated complex [PDB: 1TR8 chain A (32)] as the template. Quality of model structure (Ramachandran plot) was analyzed in Procheck program (33). Addition of missing hydrogen atoms/bonds and bond order assignments were done in protein preparation wizard of Maestro 9.2 (Schrödinger Inc.) [OPLS_2005 force field and convergence heavy atom to RMSD 0.30Å]. Structures were minimized by Molecular Modeling Tool kit (MMTK) program (34) with Amber configuration (35) and 103 iterations of steepest descent followed by conjugate gradient minimization. Huntingtin structures (Huntingtin N-terminal 17 amino acid region with 13Q or 23Q) were taken from PDB database (PDB: 3IO4 and 4FE8 respectively). Unconstrained docking of HYPK on Huntingtin structures were performed in Piper program (36) of Bioluminate suite (Maestro 9.2, Schrödinger Inc.). Among the docked structures, best posed outputs were curated and further prepared as described for 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 put in 40 Å side length cuboidal virtual water environment [TIP3P model (37)] and 150 mM Na+/Cl- salt concentration using system builder set up (OPLS_2005 force field, minimization of simulation volume and protein’s charge neutralization by opposite ions). Structures were then minimized by 1Kcal/mol/Å convergence threshold steepest descent minimization (2000 iterations). Complete MDS was divided into two stages. The brief pre-simulation or the equilibration simulation periods were meant to complete the relaxation of 6 ACS Paragon Plus Environment

Page 7 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

structures and they consisted 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 -2 ns equilibrating simulations were adequate for intrinsically unstructured proteins). The complete 300 ns simulation of HYPK-Huntingtin complexes and 50 ns full simulation of HYPK were unrestrained with following defined conditions: constant particle-temperature-pressure ensemble class, 25oC temperature, thermostat [Nose-Hoover chain method (38) with 1 ps relaxation time], barostat [Martyna-Tobias-Klein method (39)] to keep 1.01325 bar pressure with 2 ps relaxation time and isotropic coupling, RESPA integrator (6fs far, 2fs near and 2fs bonded) (40, 41). Randomized velocity coupled OPLS_2005 force field was used and interactions were regulated by 1 nm short range cutoff interaction/smooth particle mesh Ewald long range (42) columbic interaction (e-09 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 done in Pymol (43) 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 carried out in Maestro 9.2 (Schrödinger Inc.). Secondary structure analyses of native and molten globule state of HYPK were conducted using algorithms of 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), 50 mM NaF buffer. Measurements were done in JASCO 810 Spectropolarimeter with quartz cuvette of 10 mm path length. For secondary structure analysis, CD spectra were taken in far UV region of light (190-260nm) at room temperature. In slow thermal denaturation experiments (20oC – 90oC, 5oC temperature interval), proteins were stabilized at different discrete temperatures followed by the acquisition of CD spectra at far UV wavelength (190-260nm). To determine the melting temperature (Tm) of different proteins, they were continuously heated (1oC/min) and Δε values at λ=222nm were collected [Δε: change in ellipticity]. Tm values were determined by plotting dΔε/dT [dθ/dT] versus temperature [T = temperature in Kelvin scale]. In pH dependent CD experiments, protein was kept in citrate phosphate buffer of different pH (2.6 – 8.0) and protein’s secondary structure contents were analyzed by taking CD spectra in far UV range at room temperature. Deconvolutions of spectral sets were done by Convex Constraint Algorithm (47). Hydropathy plot analysis Hydropathy plot was generated in ExPASy ProtScale server (http://web.expasy.org/protscale/). Statistical analysis Statistical evaluations for the level of significance were done by paired t-test.

7 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 27

RESULTS Aggregation propensity of HYPK is prevented by its own N-terminal region Huntingtin Interacting Protein K (HYPK) is an aggregation-prone protein. In our other studies, we had found that the oligomerization of HYPK resulted from prion-like seed nucleation mechanism. In this process, protein monomers associated to form small oligomers which were termed as seeds. Seeds subsequently nucleated and associated with other seeds to extend the oligomerization, finally resulting in the aggregates. At lower concentration (10 nM), HYPK typically remained as lower oligomeric forms like monomer, dimer, trimer etc. (Figure 1A). With the increase in concentration, HYPK initially tended to form annular oligomeric complexes from seeds (concentration: 1-5 µM). This was followed by collapse of these annular complexes to form amorphous aggregates (concentration > 5 µM) (Figure 1A). The C-terminal Ubiquitin associated (UBA)-like domain of HYPK (C-terminal 45 residue region [HYPKC45]) was found to be responsible for formation of annular structures. The surface exposed hydrophobic residues in this region played an important role in the annular oligomerization (Figure 1A, 1B). The intermolecular hydrophobic interactions between HYPK molecules led to the deposition of concentric rim-like arrays of HYPK seeds during formation of annular structures. Upstream of the UBA-like domain, a small charge-rich low complexity region (LCR) was responsible for collapse of annular complexes to finally form amorphous aggregates. Since the LCR contained high 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, 1B). The N-terminal region (HYPK-N60) lacked any potential towards positive contributions in HYPK aggregation (Figure 1A, 1B). Though the endogenous expression level of HYPK in vivo did not form aggregates in significant number of cells, ectopic expression of HYPK showed an increase in 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 C-terminal region of HYPK was highly aggregation-prone, we asked the question “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 N-terminus had preventive properties against HYPK’s self-aggregation. Aggregation potential of HYPK-C69 was reduced when it was co-incubated with equimolar HYPK-N60 (Figure 1C, 1D). HYPK-C69 could form large soluble oligomers/aggregates in size (diameter) range of 100 nm – 1 µm even in low concentration (100-500 nM). The population of such large and globular aggregates drastically went down when mixed with HYPK-N60. In this case, soluble assemblies were small-granular and they were mostly in size range of 10 nm – 100 nm (Figure 1C, 1E). In vivo, when HYPK-C69 was co-expressed with HYPK-N60, it did not produce highly clumped aggregates as it was used to form when expressed alone. Instead, it produced small scattered aggregates in very few numbers of cells (Figure 1C). N-terminus of HYPK is 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 N-terminus in order to understand the mechanism of aggregation modulation. HYPK had bipartite structure. Structural studies by segmental deletion of HYPK along with CD spectroscopy revealed that the C-terminus of HYPK was mostly structured (alpha helical) whereas the N-terminus was disordered (Figure 2A). Molecular modeling of HYPK also showed the unstructured nature of the N-terminus (Figure 2B). Unrestrained molecular dynamics simulation of HYPK model structure in aqueous environment showed sporadic loss of secondary structure and increase in the proportion of unstructured region during simulation time (Figure 2C). However, the structure 8 ACS Paragon Plus Environment

Page 9 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

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 [HYPKC45] forms annular oligomers, C-terminal 69 residue region [HYPK-C69] produces amorphous aggregates and N-terminal 60 residue region [HYPK-N60] does not form aggregate. (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 aggregate. (C) HYPK-N60 prevents large spheroid type aggregate formation by HYPK-C69. (D) AFM studies show that particle sizes [heights] of HYPK-C69 do not increase over time when mixed with equimolar HYPK-N60 [concentration of both HYPK-C69 and HYPK-N60 are 200nM]. (E) Dynamic light scattering shows inhibition in formation of large particles of HYPK-C69 when incubated with equimolar HYPK-N60.

remained fairly stable throughout the 50ns simulation time window. This was reflected by the stabilization of RMSD and lower RMSF values of the protein backbone (Figure 2D). During simulation, this stability was maintained by keeping cumulatively nearly-equal numbers of intra and inter-molecular 9 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

Figure 2. N-terminus of HYPK is disordered and it has nanostructure type features. (A) CD spectroscopy shows that HYPK is structurally divided into two regions. The C-terminus is mostly alpha helical structured and N-terminus is disordered. (B) Molecular model of HYPK shows a prolonged unstructured region in the N-terminus. (C) 50ns molecular dynamics simulation of HYPK shows sporadic loss of secondary structure composition along simulation time course [red arrows]. (D) RMSD and RMSF values of HYPK backbone represent stabilization of structure with time. (E) Number of hydrogen bonds formed by HYPK and radius of gyration of HYPK at different time points of 50ns molecular dynamics simulation. (F) AFM of HYPK-N60 shows that it has nano-dimension. Most of the HYPK-N60 particle’s sizes range below 10nm.

10 ACS Paragon Plus Environment

Page 11 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

hydrogen bonding (Figure 2E). Since this unstructured region contained only 50-60 amino acids, we believed that this region could assume nano-structure. Atomic force microscopy (AFM) study actually showed the nano-dimension of HYPK-N60 (Figure 2F), with most of the particles showing height less than 10nm (Figure 2F). Model structure of HYPK also showed that the span of this disordered region remained in nanometer scale (Figure 2B). Both disorder and nano-structure of N-terminus were believed to play significant roles in preventing HYPK’s uncontrolled self-association. Stabilization of the C-terminal LCR is done by a negative charge rich patch of N-terminal disordered nanostructure Despite the fact that C-terminus of HYPK had high intrinsic ability to form aggregates, surprisingly, HYPK did not form aggregates of considerably higher size and number in the majority of cells under normal endogenous expression level (Figure 3C). In order to understand the precise mechanism and sequence stretches that stabilized intra-cellular HYPK and prevented its own aggregation, we constructed various deletions and multiple-point mutant constructs of HYPK (Figure 3A). We studied the aggregate forming abilities of HYPK and its deletion/point- mutant constructs. While endogenous expression of HYPK led to a very low level of aggregation, over expression of HYPK caused an increase of aggregation (Figure 3B, 3C). The C-terminal 69 residue region (HYPK-C69) induced significant aggregation. The disordered Nterminal 60 residue region (HYPK-N60) did not form intracellular aggregates (Figure 3B, 3C). AFM studies revealed that the C-terminal charge-rich LCR (residue 70-87) caused aggressiveness in terms of formation of HYPK’s amorphous aggregates. The question was how exactly this aggregation was prevented inside the cell by the activity of N-terminus? The close examination of N-terminus region identified a patch of negatively charged residues. To examine if this negatively charged rich region in the N-terminus (residue 12-22) had any influence on prevention of aggregation, the five glutamate and one aspartate residues were mutated to alanine [HYPK E-D/A]. Mutation of these residues elicited high aggregation of HYPK (Figure 3B, 3C). However, when the glutamate residues were mutated to similar charged aspartate residues [HYPK E/D], aggregation induction significantly dropped (Figure 3B, 3C). Further, multiple point mutations of the positively charged residues (arginine and lysine) to alanine in LCR (HYPK R-K/A) also showed significant reduction of aggregation (Figure 3B, 3C). These observations suggested that the negatively charged patch in the N-terminus might be involved in electrostatically interacting with C-terminal LCR. This interaction prevented LCR mediated aggregation propagation in physiological condition (Figure 3D). Binding studies of N-terminal 60 residue region (HYPK-N60) and its multiple point mutant variants (HYPK-N60 E-D/A and HYPK-N60 E/D) with the C-terminal 69 residue 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 intra-molecular charge-interactions inhibited aggregation of HYPK, we expected that altered pH conditions would perturb such interactions, leading to accumulation of larger sized particles. The shift of pH from 7.5 to 3 actually showed a gradual increase in particle size of HYPK (Figure 3F). This finally proved that intra-molecular charge-interactions, mediated by protein’s disordered nanostructure, not only acted as a molecular stabilizer, but also as an anti-aggregation factor. Loss of charge-interactions between the LCR and N-terminal disordered nano-region leads to formation of HYPK’s molten globule-like state and subsequent aggregation 11 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 27

Figure 3. HYPK’s N-terminal glutamate/aspartate rich region stabilizes the C-terminal charge rich low complexity region and prevents HYPK’s self-aggregation. (A) Different deletion and multiple point mutant constructs of HYPK which are used in this study. [E/D = aspartate/glutamate rich region, LCR = low complexity region, HR = hydrophobic region; EGFP = enhanced green fluorescent protein.] (B) Aggregates of HYPK and its different mutants in HeLa cell. (C) Quantitative evaluation of aggregate forming potential of HYPK and its different mutants in HeLa cells. Charge neutralizations by alanine mutation of N-terminal negatively charged residues [HYPK E-D/A] cause higher

12 ACS Paragon Plus Environment

Page 13 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

propensity of aggregation [*HYPK end/HYPK ectopic: paired t-test, n=6, df=4, p