Kinetics Study on the HIV-1 Ectodomain Protein Quaternary Structure

Subscriber access provided by University of Washington | Libraries

Article

Kinetics Study on the HIV-1 Ectodomain Protein Quaternary Structure Formation Reveals Coupling of Chain Folding and Self-Assembly in the Refolding Cascade Shu-Fang Cheng, Taiching Sung, Chung-Chieh Chang, Yun-Wei Chiang, Mei-Ju Chou, and Ding-Kwo Chang J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/jp508360k • Publication Date (Web): 21 Oct 2014 Downloaded from http://pubs.acs.org on October 26, 2014

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.

The Journal of Physical Chemistry B 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 23

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

The Journal of Physical Chemistry

Kinetics Study on the HIV-1 Ectodomain Protein Quaternary Structure Formation Reveals Coupling of Chain Folding and Self-assembly in the Refolding Cascade

Shu-Fang Cheng,a Tai-Ching Sung,b Chung-Chieh Chang,a Mei-Ju Chou,a Yun-Wei Chiangb and Ding-Kwo Chang*a

a

Institute of Chemistry, Academia Sinica, Taipei, Taiwan, Republic of China 11529.

b

Department of Chemistry and Frontier Research Center on Fundamental and Applied Sciences of Matters, National Tsing Hua University, Hsinchu, Taiwan, Republic of China 30013.

To whom correspondence should be addressed: Ding-Kwo Chang, Tel: 886-2-27898594; Fax: 886-2-27831237; E-mail: [email protected]

ACS Paragon Plus 1 Environment


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

Abstract Entry of HIV-1 into the target cell is mediated by the envelope glycoprotein consisting of noncovalently associated surface subunit gp120 and transmembrane subunit gp41. To form functional gp41 complex, the protein undergoes hairpin formation and self-assembly. The fusion event can be inhibited by gp41-derived peptides at nanomolar concentration and is highly dependent of the time-of-addition, implying a role of folding kinetics on the inhibitory action. Oligomerization of gp41 ectodomain was demonstrated by light scattering measurements. Kinetic study by stopped-flow fluorescence and absorption measurements (i) revealed multi-state folding pathway and stable intermediates; (ii) showed a dissection of fast and slow components for early and late stages of folding, respectively, with three orders of magnitude difference in the time scale; (iii) the slow process was attributed to misfolding and unzipping of the hairpin; (iv) retardation of the native hairpin formation is assumed to lead to coupling of the correctly registered hairpin and self-assembly. This coupling allows the deduction on the time scale of intra-chain folding (0.1-1 s) for the protein. The folding reaction was illustrated by a free energy profile to explain the temporal dichotomy of fast and slow steps of folding as well as effective inhibition by gp41-derived peptide.

Keywords: six helix bundle, hairpin formation, self-assembly, chevron plot, fluorescence quenching, probe stacking


Page 2 of 23

Page 3 of 23

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


Introduction To infect target cells by an enveloped virus, cellular and viral membranes must undergo fusion to allow the transfer of genetic materials (RNA or DNA and polymerases, etc.). For class 1 viruses such as human immunodeficiency virus (HIV), fusion is mediated by the fusion protein, gp160, which exists in triplex form and consists of the surface subunit (SU) gp120 and transmembrane subunit (TM) gp41. Coalescence of membranes involves dehydration of membrane surface, drawing the apposing membranes into contact, and disruption of lipid structure and organization; as such, the fusion reaction is a multi-step event that must surmount huge free energy barrier. Hence the fusion proteins have to undergo large scale conformational changes and assemble the trimeric units in a highly coordinated manner. It has been documented that the gp41-mediated fusion can be inhibited by a peptide, T20 (Fuzeon), derived from gp41depending on the time of addition, implicating that molecular dynamics plays a critical role in the fusion process 1. Additionally, it was observed 2 that the resistance genotype of the peptide-based fusion inhibition can be mapped to the mutation at the position within the gp41 core not overlapped with the sequence of the inhibitor. Interestingly, the resistance strain developed an inhibitor-dependent phenotype, which was postulated to be modulated by folding kinetics of the protein. Dynamics of folding of homo-dimer proteins has been extensively studied. Because the process involves association of monomers and folding to the native form 3, the mechanism is in general more complex and has more paths than the single domain proteins. For instance, the monomers could fold to near-native structure before associating to dimer 4 or the monomers combine to form dimer with disordered structure as an intermediate, or the formation of native dimer could be achieved by coupling of monomer association with folding. The last alternative has some bearing to the notion of induced fitting of enzyme active sites. The native structure of gp41, as determined by crystallography and NMR analysis 5 6, is a helical homotrimeric heterodimer. The inner core is formed by the N-terminal heptad repeat (NHR) region of gp41 ectodomain while the C-terminal heptad repeat (CHR) region packs against the groove formed by the trimeric core. A wide line of evidence suggests that the NHR core is a preformed stable complex, as peptide molecules derived from NHR was found to form stable helical multimers 7. In contrast, the CHR-derived peptide is disordered, but is induced into helix by complexing with the NHR peptide. Thus in considering the mechanism underlining six helix bundle (SHB) formation, it is likely that trimeric NHR core of the prehairpin conformation is formed in the initial stage of folding. Yet it is unclear that folding is via framework (i.e. the completely formed NHR core attached by disordered CHR segments) or nucleation-condensation (i.e. NHR core formation is coupled to CHR attachment) pathway. Because the helical coiled coil is a common tertiary structure motif, for example, in DNA-binding transcription factor, tropomyosin and viral fusion proteins 8 9, the study on folding dynamics of the helix bundle has been extensive. Folding time on the microsecond scale was found for a 38-residue double helix hairpin by Du and Gai 10 using infrared temperature jump technique. Remarkably, 20 µs folding time was deduced from measuring a three-helix bundle to which folding of the double helix was slowed by one order of magnitude resulting from docking of the remaining helix 11. It is likely that misdocking of helix 1 on the helix 2-turn-helix 3 core



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

and undocking to form the native contact engender a high energy barrier for the molecule to transit to the native state, thereby slowing the folding reaction 12. This phenomenon may bear profound implications in the folding study of SHB of gp41 ectodomain, requiring correct register of three monomers of hairpin configuration. Previous folding studies mainly focused on the kinetics of intramolecular events without considering the formation of homo-multimeric complex via diffusion and intermolecular collision. Coupling of these two processes confers kinetic advantages in protein functionality as exemplified in the protein-mediated signal transduction 13 14 15. Because the virus-promoted membrane fusion requires highly coordinated organization of trimeric units, in addition to large scale structural rearrangement, mechanistic understanding of formation of the fusion protein complex to drive the fusion process is of great fundamental and clinical significance. Methods such as circular dichroism and IR have been utilized to detect the secondary structure or chemical bond vibrational frequency that reflected local structural change 16. NMR and Mass spectroscopy have been used to monitor hydrogen exchange (through pulse-labeling) for more widespread detection of structural rearrangement 17 18. As bending of the viral fusion protein is thought to drive membrane fusion, the detection of end-to-end (e-t-e) distance of the molecules becomes essential. Among the techniques available, small angle X-ray scattering is perhaps most effective in measuring the pairwise atom-to-atom distance distribution and kinetics of conformational and molecular assemblage via global size and shape determination 19. Fluorescence resonance energy transfer technique has been employed to monitor the interaction between tryptophan and the fluorescent tag based on the distance change between the fluorescent pair 20. As an alternative approach to directly detect the e-t-e distance and self-assembly, we report herein a method to monitor the conformational change (via tertiary contact) through labeling at the residues near the termini, as described previously 21. Additionally, a single probe has been labeled at the middle section of the protein to probe changes exclusively in the inter-chain distance. Denaturation and renaturation of the protein were modulated by adjusting the denaturant guanidinium chloride (GdmCl) concentration to induce the refolding and unfolding processes.

UV-Vis absorption and fluorescence quenching stopped-flow technique were used to measure the dynamics of the oligomerization and conformational change. The latter technique is most sensitive the change in distance longer than that determined by the former, therefore reporting early stage of refolding cascade. Light scattering experiments revealed distribution of molecular assembly. In combination with electron spin resonance (ESR) experiments to probe the equilibrium inter-probe distance and the degree of oligomerization, the results allowed us to uncover evidence for intermediate accumulation and to distinguish between intra- vs. inter-molecular events, thus shedding light on the mechanism of folding process of the viral fusion protein. Analysis of kinetic data yielded a multi-state folding process, which was affirmed by departure from linearity of the chevron plot. The fast, millisecond formation of the NHR triplex core was implicated by the burst phase observed for the singly-labeled protein. Difference in the protein concentration dependence of kinetic rate constant from results obtained with single- vs. double-labeled protein can be accounted for by a slow folding into correctly registered hairpin coupled to assemblage of the trimer. In other word, intrachain folding to form the helical hairpin was indirectly deduced from the coupling of intra- and inter-chain processes. A working model was proposed for the mechanism of formation of native and functional structure of gp41 ACS Paragon Plus 4 Environment

Page 4 of 23

Page 5 of 23

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


ectodomain complex. The present study on the SHB formation and clustering of gp41 ectodomain protein represented a paradigm for the quaternary formation of homologous oligomeric biomolecules.

Experimental methods Protein biosynthesis and labeling Synthesis of the recombinant gp41 ectodomain (designated as gp41e; the molecular model with the sequence shown in Fig. S1 in the Supporting Information) and its I573P and N637K mutants followed the protocols described previously 21 22. Tetramethylrhodamine-5(and-6)-maleimide (5(6)-TAMRA, Anaspec, Inc., San Jose, CA) was used for Rhodamine labeling and the labeled proteins were purified by dialysis and HPLC as detailed in previous study 21. Rhodamine (Rho) was singly labeled on residue 546 (N-terminal of gp41, Rho-C546) or residue 604 (loop region of gp41, Rho-C604) to investigate the molecular assembly during protein renaturation. Double labeling of Rho on residues 546 and 660 (N- and C- terminal of gp41, respectively, Rho-C546/660) was used to monitor the e-t-e interaction of protein folding. Labeling of gp41e with 4-Maleimido-2,2,6,6-tetramethyl-1-piperidinyloxy (4-Maleimido-TEMPO, Sigma-Aldrich, St. Louis, MO) spin label (MSL) for ESR measurements was carried out by the method similar to that for rhodamine labeling, except for a ten-fold molar excess of MSL dissolved in DMSO. The synthesized recombinant protein variants were confirmed by mass spectrometry (Fig. S2). All gp41 proteins used in this study were C598S/C604S doubly mutated to ensure that the molecular assembly determined was not contributed by disulfide linkage between cysteine residues, with the exception that a single mutation C598S was made in Rho-C604 for rhodamine conjugation at residue 604. Light Scattering measurements A triple-angle light scattering detector (miniDawn by Wyatt Technology Co., Santa Barbara, CA, USA) was used to examine the absolute molecular weights of the constructed proteins. The molecular size (hydrodynamic radius) of the proteins was investigated by dynamic light scattering using WyattQELS (Wyatt Technology Co.) located at 90 degree of the miniDAWN. Gp41e or its mutant, I573P or N637K, incubated in various concentrations of GdmCl/50 mM HCOONa was filtered by a NanoFilter with 0.02 µm Anodisc membranes and then measured in batch mode by using a microcuvette. Concentrations of the filtrated proteins were determined by UV spectrophotometer (U-2001, Hitachi, Chiyoda, Tokyo, Japan) with absorption coefficient of 42390 M-1·cm-1 at 280 nm. Proper background was recorded before the protein measurement and used as the baseline for data analysis by ASTRA V (Wyatt). Note that gp41 helix bundle is morphologically a long rod with a diameter of about 4.0 nm and a length of about 10 nm according to the structure of PDB1QBZ. Yet the number of molecules in the assembly was calculated by assuming spherical shape for the assembly; thus the deduced number is considered to indicate the relative oligomerization propensity among the protein variants.



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

Stopped-flow fluorescence measurements The fluorescent dye, Rhodamine, would self-quench as the Rho-labeled protein self-associated and/or refolded from denatured state. Stopped-flow fluorescence experiments were performed on an Applied Photophysics SX.18MV stopped flow analyzer (Leatherhead, Surrey, United Kingdom) equipped with an asymmetric mixing system at a ratio of 10:1. Fluorescence was measured at 25ºC using a 530 nm cutoff filter with the excitation wavelength of 530 nm. For the refolding experiments, 0.5 µM of rhodamine-labeled protein solution was incubated in 8 M GdmCl/50 mM sodium formate solution and rapidly diluted with 10-fold (v:v) GdmCl solution of proper concentration in 50 mM sodium formate at pH 3. The final concentration of the Rho-labeled protein was ~45 nM. Sigma Plot 12.0 (San Jose, CA, USA) was used to fit the signal transients by exponential decay equation (Eq. 1 in Supporting Information). The folding rate constant under the denaturant-free condition was obtained by extrapolating the rate constant vs. GdmCl concentration plot. Stopped-flow absorbance measurements The rhodamine moiety has an absorption maximum near 555 nm and the formation of rhodamine dimers gives rise to a new absorption band at 518 nm. The ratio of absorbance at 518 nm to that at 555 nm was used to estimate the dimer/monomer composition of rhodamine 21 23 24. The ratios of 0.4 and 1.4 signify that rhodamine on the proteins are predominantly monomeric and dimeric, respectively. To investigate the protein refolding rate by monitoring the rhodamine stacking, a stopped-flow system (SFM-20 by Bio-Logic SA, Claix, France) equipped with a high power UV-VIS fiber light source (HAMAMATWU L10290 by Hamamatu photonics K.K., Shizuoka Pref., Japan) and a diode array spectrometry system (TIDAS I MMS-UV 500-3 by J&M Analitik AG, Aalen, Germany) were employed. Two syringes of 1 ml and 10 ml containing the denatured protein (dissolved in 8 M GdmCl/HCOONa 50 mM) and 0~6 M GdmCl/HCOONa 50 mM, respectively, a high density mixer and a 1 mm×10 mm cuvette (TC-100/10F) were set on the stopped-flow system. For each injection, 20 µl of denatured protein and 201 µl of refolding solution were mixed with a flow rate of 5 ml/s. Data were recorded in the range of 450-650 nm with integration time of 1 ms at 25ºC. Initially 1500 points were collected every 1 ms, followed by 2190 points every 0.3 s. Deconvolution of the time evolution of 518/555 nm peak ratio was fitted by multi-exponential function using Sigma Plot 12.0 (San Jose, CA, USA). The minimalist approach was adopted, i.e. smallest number of exponential functions was taken which can fit the data to an acceptable degree. For Rho-C546/660, tri-exponential function was employed whereas bi-exponential function was used for Rho-C604. The fitting equations and parameters were presented in Supporting Information (Eq. 2-4). The folding rates were plotted against the equilibrated denaturant concentrations and extrapolated to the native (denaturant-free) state. To evaluate the activation energy, temperature dependence of the reaction rate was

employed through Arrhenius’ equation = ∙ exp ( ), where k is the rate constant, A is the pre-factor, Ea is the activation energy, R is the universal gas constant and T is the absolute temperature. Three temperatures (25, 37 and 50ºC) were tested for Rho-C546/660 ACS Paragon Plus 6 Environment

Page 6 of 23

Page 7 of 23

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


Cw-ESR measurements A Bruker ELEXSYS E580 cw/pulsed spectrometer, equipped with a Bruker pulse ELDOR unit E580-400, a dielectric resonator (ER4118X-MD5W) and a helium gas flow system (4118CF and 4112HV), was used. The cw-ESR experiments were performed at an operating frequency of 9.4 GHz and 1.5 mW incident microwave power with swept magnetic range of 200 Gauss. Experimental setup was the same as previously described 25. Accumulation time for each set of data was about 10~12 h. The cw-ESR techniques is useful for measuring inter-spin distances within 0.5~2 nm. The spectrum of MSL attached to the gp41e complex is expected to be insensitive to the tumbling and backbone motions even at 300K, as the MSL is a large and motionally restricted probe. Thus, the spectral broadening observed for the MSL-C546/660 complex at 300 K reflects largely the dipolar broadening effected by the inter-spin distance in the range of 0.5~2 nm. For all experiments, the data of more than two independent measurements were taken for averaging.

Results Propensity of gp41 ectodomain for self-assembly is evidenced by light scattering experiments In the previous study on gp41 ectodomain protein, we have observed a dual-phase helix transition from titration by denaturant while single-phase transition at lower denaturant concentration was found for the mutant with destabilized NHR region 21. This result allowed us to deduce a preformed NHR triplex core but not CHR triplex formation in the course of folding to the native SHB. The idea was corroborated by the findings of highly stable helical NHR peptide homolog and a random coil structure for the CHR peptide in solution 26. The result also suggested a multi-state folding process and the sharp transition signified cooperativity for gp41e folding. To carry out the function, the fusion protein molecules must self-assemble in highly orchestrated manner, therefore it is imperative to have an insight into the stability and organization of gp41e clustering. To achieve these goals, we tested the molecule’s resistance to disassembly against GdmCl by light scattering experiments and compared results obtained with the wild type and point-mutational analogs implicated in the SHB stability 21 27. As is seen in Figure 1, in the absence of denaturant, both gp41e and N637K exhibit high order oligomerization while I573P mutant displays some tendency to trimerize. Increasing GdmCl concentration reduces the propensity for oligomerization for all the proteins tested. By 3 M GdmCl, I573P exists mainly as monomers. In contrast, at this denaturant concentration, the wild type protein is on average trimeric and N637K nonameric. The latter two molecules are reduced to monomers by ~5 M GdmCl, with N637K oligomers even more resistant to complete dissociation. The observed trend is in agreement with the assertion that the NHR region forms a triplex helical core, thus the disruptive mutation at position 573 within the NHR region by proline critically reduces helix stability, thereby lowering the propensity for oligomerization. Our data further indicate higher stability of N637K oligomers than the wild type protein, in support of the



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

proposition regarding the T20-dependent N637K mutant 2. Taken together with the previous study on the gp41 mutants 21, these results point to a correlation between the oligomerization propensity and helix stability of gp41 core.

Figure 1. Molecular weight and size analysis by light scattering measurements.

An early stage of fast refolding is revealed by stopped-flow fluorescence self-quenching results In the course of refolding reaction from the denaturant-induced disordered form, the interprobe, e-t-e distance, is shortened as the hairpin is formed for gp41e. Since rhodamine fluorescence self-quenching is operative in the distance longer than that for the absorption-monitored stacking, it is expected to report early event(s) of folding process. The temporal variation of the signal can be deconvoluted sufficiently well into single exponential decay for both labeled proteins studied. Fitting residuals were shown in Figures S3A-B. To investigate the mechanism of folding dynamics, a chevron plot is frequently used in which the logarithm of rate constant was plotted as a function of denaturant concentration 28 29. For a protein folding in the two-state mode with a transition barrier but no intermediate, the plot is characterized by a sloped, linear trace for the folding or unfolding branch. The analysis is based on the modulation of reaction free energy by solvent exposure and has been used to distinguish the two-state pathway from multi-state and downhill pathways 16 30. In Figure 2B the chevron plot was displayed for the two labeled variants. Notably, the rate constants extrapolated to zero denaturant concentration were estimated to be 403 s-1 and 493 s-1 for the doubly- and singly-labeled gp41e, respectively. These millisecond time constants indicated that relatively fast early kinetic events were probed by fluorescence measurements. The finding of nearly the same extrapolated rate for doubly- and singly-labeled gp41 proteins supports the contention that the fluorescence self-quenching senses the early stage of protein renaturation where little distinction can be made between intra- and inter-chain interactions, a point to be further elaborated in Discussion. The nearly flat chevron plot is a manifestation of gradual transition probed by the present fluorescence measurements, as a consequence of low transition energy barrier (hence the fast rate, see the working model illustrated in Figure 7) and little surface burial incurred during the early phase of folding reaction. The latter notion suggests that the transient structure is substantially solvent-exposed and that the early molten globule state consists of loosely organized domains. In terms of energy landscape theory 28 31, the analysis led to the following scenario: the early stage of gp41e folding can be described by the molecular chain diffusing over ACS Paragon Plus 8 Environment

Page 8 of 23

Page 9 of 23

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


free energy landscape surface with low ruggedness and hence depicted as moderately downhill transition.

Figure 2. Refolding kinetics of gp41 ectodomain proteins monitored by stopped-flow rhodamine fluorescence self-quenching.

Stacking of rhodamine moieties monitored by absorption peak ratio yields dynamics of folding at later stage The late stage of folding involves structural compaction of the molecules, hence is better detected by stacking of rhodamine moiety measured by the absorption peak ratio approach 21. The absorbance diode array data were shown in Figure S3. Decreasing and increasing absorbance at 518 nm and 555 nm, respectively, were seen as the concentration of GdmCl was raised (Fig. S4A-C). Figure S4D revealed the time trajectory of 518/555 nm ratio. Figure 3A displays the time evolution of the probe stacking for gp41e at final denaturant concentration of 0.73 M. The curve was fitted by tri- and bi-exponential functions for doublyand singly-labeled proteins, respectively. The residuals are shown in Figures S5A,B, along with the residuals obtained from other types of exponential function for comparison. The multi-exponential deconvolution of reaction kinetics points to a multi-state, rather than a two-state, process with possibility of more than one intermediate. In Figure 3B, the chevron plot exhibited two plateau regions and cooperative transitions at 3.6 and 5 M GdmCl for all the components determined. Since the transitions roughly coincide with transitions in oligomerization order shown in Figure 1A, the anomalous behavior hints a role of inter-chain interaction in the late stage of folding and coupling of intra-chain folding and inter-chain assembly to be elaborated in Discussion. The clear deviation of the plot from linearity again reaffirms a departure from the two-state mode for gp41e refolding and suggests accumulation of intermediates 32. Another manifest of substantial population of stable intermediates can be discerned from the susbstantial amplitude of burst phase which represents processes with millisecond time constant for the doubly- and singly-labeled proteins due to the ~9 ms instrumental deadtime. Since the time regime overlaps that deduced from the stopped-flow fluorescence result shown in Figure 2, the burst phase that reflects the early phase of refolding composed of inter- and intra-molecular interactions. Interestingly, the singly-labeled gp41e exhibits a burst phase amplitude smaller than the doubly-labeled gp41e. The observed burst phase for Rho-604 signifies that inter-chain, yet intra-trimer, association occurs on the order of a millisecond and argues for the rapid association



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

of three NHR domains of gp41e. The difference between the burst phase amplitudes of doublyand singly-labeled gp41e is attributed to intra-molecular interactions, namely bending of CHR against NHR to form a hairpin. Clearly, the absorbance experiments yield two distinct folding processes: the fast processes with millisecond time scale (burst phase) and slow processes spanning 0.1 to 10 seconds or longer. Activation energy (Ea) of the reaction, a measure of the barrier height to be surmounted by the reactant, can be determined from the temperature dependence of rate constant. Ea derived from Figure 3C are 48, 58 and 44 kJ/mole for k1, k2 and k3, respectively. The results demonstrate that the rate limiting step of the reaction involved breaking of considerable number of hydrogen bond, van der Waals contact, and/or hydrophobic interaction.

Figure 3. Absorbance kinetic studies.

Protein concentration dependence of late stage of folding rate reveals coupling of self-assembly and hairpin-forming bending of gp41e To further characterize kobs shown in Figure 3B and the early stage of the folding kinetics implicated by the burst phase in Figure 3A, we performed the protein concentration dependence of rate constants as shown in Figure 4. There are two notable features for these two proteins tested: a linear increase of the fast component (k1) and nearly identical k2 or k3 (slow component) values at the same concentrations for both proteins. The results point to a strong coupling of the intra- and inter-chain reactions for k1. For k3, the smaller magnitude does not permit accurate assessment, but the absence of cubic concentration dependence argues against its contribution from intermolecular processes.


Page 10 of 23

Page 11 of 23

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


Figure 4. Protein concentration dependence of folding rates obtained by absorption kinetics.

Intra-chain folding as a major source of burst phase amplitude detected in absorbance data is deduced from the protein concentration dependence of folding rate The submillisecond folding time observed in fluorescence quenching experiments (Fig. 2) prompted us to examine the nature of burst phase (due to the deadtime of ~9 ms) indicated in Figure 3A. Figure 5 displays the amplitude of burst phase as a function of the protein concentration for both doubly- and singly-labeled gp41e. The moderate variation with concentration of folding rates of Rho-C546/660 and Rho-C604 is consistent with the contention that the fast process assigned for the doubly-labeled gp41e arises primarily from intra-chain folding. This is because that, for the singly-labeled protein, raising the concentration does not increase the burst phase amplitude contributed solely by intermolecular reactions. The result also hints less distinction between intra- and inter-chain processes in the early phase of refolding.

Figure 5. Analysis of burst phase amplitude changes in absorption-monitored kinetics experiments.



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

ESR experiments confirm stable intermediates in the late stage of gp41e folding Since it was postulated that the fast, submillisecond processes deduced from the stopped-flow fluorescence measurements are the processes underlining the burst phase found in the stopped-flow absorption experiments and contributed considerably from intra-chain folding, we sought to use cw-ESR experiments to support the hypothesis. Doubly spin-labeled gp41e, designated as MSL-C546/660, was studied. Cw-ESR is useful for measuring inter-spin distances smaller than 2 nm. Figure 6A shows that, as GdmCl is elevated from 1 to 4 M, the spectra of MSL-C546/660 approach the spectrum acquired in 6 M GdmCl. Note that the spectra for 0 and 1 M GdmCl are alike (data not shown). This indicates that the inter-spin distance in the doubly-labeled gp41e increases with GdmCl and become greater than 2 nm with GdmCl greater than 4 M. Using the spectrum for 6 M GdmCl as a reference for the spectrum devoid of dipolar interactions, we obtained the distance distributions P(r) as depicted in Figure 6B by the Tikhonov-based analysis. The sensitivity of the technique to denaturing conditions is demonstrated by the difference in P(r) for 1 and 2 M [GdmCl]. A model for MSL-C546/660 is illustrated in Figure 6C for reference using the structure adapted from PDB1QBZ. The peak of P(r) at 1 M GdmCl near 1.1 nm concurs the model shown in Figure 6C for the probes located at residues 546 and 660. On the other hand, a shift of P(r) peak to ~1.25 nm at 2 M GdmCl supports the hypothesis of a stable intermediate with e-t-e distance larger than that of the native state. The ESR results are also consistent with the heterogeneous hydrodynamic radius distribution observed in the light scattering measurements (Fig. 1C).

Figure 6. Cw-ESR result confirms the hairpin structure of gp41 ectodomain protein.

Discussion Evidence for the non-two state folder for the gp41 ectodomain protein renaturation and existence of stable intermediates


Page 12 of 23

Page 13 of 23

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


The finding that both fluorescence- and absorption-monitored chevron plots exhibit departure from two-state folding — characterized by sloped, linear traces — for gp41e is striking. However, the origin of the flat chevron plot obtained with these two measurements is quite different. The self-quenching result was interpreted as downhill-like behavior for the early stage of folding. The data lend support for the view that, at the early folding stage, the decrease in enthalpy was largely compensated by a reduction in entropy, resulting in little change in free energy. Lack of substantial activation energy barrier can lead to a plateau region in the plot. The finding is also consistent with the argument that no energetically stable intermediate exists in the molten globule state in the initial phase of folding. It is of interest to note that, although thermodynamically the conformation of molecules in solution of low denaturant concentration approaches that of the native state, the kinetics monitored by Rhodamine fluorescence quenching reflects only the early phase of folding in which there is little distinction between intra- and inter-chain quenching interactions. Evidence for the latter idea comes from nearly identical profiles of chevron plot for Rho-C546/660 and Rho-C546 (Fig 2B). Deviation from linearity of the folding branch of the chevron plot observed in absorption measurements, which report the late stage of folding, implicates a multi-step kinetics and accumulation of the stable intermediate, thereby slowing down the folding 33. (Downhill transition can be excluded since the deduced rate constants from absorbance experiments are too slow.) Evidence for the existence of stable intermediates is afforded by the polydispersity in hydrodynamic radius distribution found in light scattering measurements (Fig. 1C) as well as the cw-ESR data (Fig. 6) displaying multiple e-t-e distance distributions for the protein. From the viewpoint of energy landscape theory 28 31, the anomaly of the plateau regions (Fig. 3B) in the plot as well as the sharp transition implies that the extent of solvent exposure (contact) experiences large change in the transition region whereas little alteration in the degree of solvent contact is incurred in the plateau. Apparently change in the solvent exposure is predominantly contributed by the assembly event, while insignificant change is induced by packing and unpacking of CHR against the NHR core. Moreover, nearly the same values and variation with [GdmCl] in the range wild type>I573P. As predicated in the Methods, the sizes deduced from light scattering measurements were based on treating the molecule as a sphere. Discrepancy between the mean value of molecular radius and the maximum of radius distribution indicates heterogeneity of the oligomerization of the protein. The hydrodynamic radius distributions of the three analogs shown in (C) were used to derive results in (B). The higher population of larger size for N637K and a tail in the small radius distribution for I573P argue for a critical role of NHR stability in the gp41e assembly. Figure 2. Refolding kinetics of gp41 ectodomain proteins monitored by stopped-flow rhodamine fluorescence self-quenching. (A) Evolution of Rho-C546/660 folding in 0.73 M GdmCl solution. The data of first 10 ms were shown in the inset for clarity. The refolding curves were fitted with single exponential function to obtain the observed folding rates (kobs). The predicted trend was shown by the grey line. (B) Plot of folding constants against GdmCl


Page 20 of 23

Page 21 of 23

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


concentrations. For comparison the intermolecular clustering was probed by measurements on Rho-C546, for which rhodamine is predominantly monomeric in [GdmCl] range tested in accordance with Reference 21. Figure 3. Absorbance kinetic studies. (A) Evolution of gp41e refolded in 0.73 M GdmCl probed by the 518/555 nm absorption ratio. The inset displayed expanded data of the initial 1.5 s. Proteins tested were Rho-C546/660 (left panel) and Rho-C604 (right panel). kobs values were obtained with bi-exponential and tri-exponential fittings for singly- and doubly-labelled proteins, respectively. (B) Chevron plots of devoluted rate constant for Rho-C546/660 and Rho-C604. For both labeled proteins, the plateau regions with a sharp transition largely coinciding that of oligomerization revealed in Figure 1suggest that the observed rate involves coupling between intra- and inter-chain processes. (C) Arrhenius plot for Rho-C546/660. The activation energy was determined as the slope of the linearized plot for each of the three components of rate constant and indicated at the top of the figure. Figure 4. Protein concentration dependence of folding rates obtained by absorption kinetics. Folding rates of singly- and doubly-labeled gp41e in various concentrations were plotted against the final protein concentrations. The inset depicted expanded scale of k2 and k3 as a function of protein concentration for clarity. The nearly identical rates for k1 of the singly- and doubly-labeled gp41e and their concentration dependence strongly argued for the coupling of intramolecular folding and intermolecular clustering during the transition to the native form. Figure 5. Analysis of burst phase amplitude changes in absorption-monitored kinetics experiments. The amplitude of burst phase of protein refolding in 0.73 M GdmCl was plotted against the final protein concentrations. The near independence of burst phase amplitude on protein concentration and the differential burst phase amplitude between the doubly- and singly-labeled proteins implied a substantial contribution of intramolecular NHR:CHR hairpin formation to the fast phase of refolding event (millisecond time scale). The results also hint at a fast process for the misfolded hairpin formation. Figure 6. Cw-ESR result confirms the hairpin structure of gp41 ectodomain protein. (A) Normalized cw-ESR spectra at indicated [GdmCl]. (B) The inter-spin distance distributions for MSL-C546/660. The average distances, 1.07 nm and 1.25 nm for 1 M and 2 M GdmCl, respectively, were determined using the spectrum attained with 6 M GdmCl as the non-interacting reference. The increased average inter-probe distance (from 1.07 to 1.25 nm) with the denaturant concentration between 1 and 2 M may reflect the sensitivity to the denaturant of the end fraying effect and/or the misfolded state of the hairpin. For [GdmCl] >4 M, the inter-spin distances (regardless of inter- or intra-molecular distances) are outside the sensitive regime for cw-ESR measurements (i.e.,

Kinetics Study on the HIV-1 Ectodomain Protein Quaternary Structure

Recommend Documents