Cold Denaturation of the HIV-1 Protease Monomer - ACS Publications

Feb 7, 2017 - Biotech Research and Innovation Centre (BRIC), Faculty of Health and Medical Sciences, University of Copenhagen, Ole Maaloees...
0 downloads 0 Views 851KB Size
Rapid Report pubs.acs.org/biochemistry

Cold Denaturation of the HIV‑1 Protease Monomer Heike I. Rösner,*,†,‡ Martina Caldarini,§ Andreas Prestel,† Maria A. Vanoni,∥ Ricardo A. Broglia,§,⊥ Alessandro Aliverti,∥ Guido Tiana,§ and Birthe B. Kragelund† †

Structural Biology and NMR Laboratory (SBiNlab), Department of Biology, University of Copenhagen, Ole Maaloees Vej 5, DK-2200 Copenhagen N, Denmark ‡ Biotech Research and Innovation Centre (BRIC), Faculty of Health and Medical Sciences, University of Copenhagen, Ole Maaloees Vej 5, DK-2200 Copenhagen N, Denmark § Department of Physics, University of Milano and INFN, via Celoria 16, 20133 Milano, Italy ∥ Department of Biosciences, University of Milano, via Celoria 26, 20133 Milano, Italy ⊥ Niels Bohr Institute, Blegdamsvej 17, 2100 Copenhagen Ø, Denmark S Supporting Information *

Human immunodeficiency virus-1 (HIV-1) protease is a retroviral aspartic acid protease that plays a pivotal role in the life cycle progression of HIV. In its active form, HIV-1 protease forms a homodimer.10 Deletion of the four C-terminal residues stabilizes a monomeric, folded form11 with a stability of 1.35 kcal/mol at pH 6 and 25 °C.12 The structure of this monomeric form of HIV-1 protease, mHIV-1-PRΔ95−99, is similar to that in the dimeric state and contains seven β-sheets and a C-terminal αhelix.13 Native, non-native, and denatured states of the HIV-1protease, as well as of some of its variants, have been widely characterized both in silico14,15 and in vivo.16 However, despite its central role as a target for antiretroviral therapies,16 biochemical and biophysical data on the HIV-1-protease are limited. With the goal of exploring new targets for drugs interfering with the correct folding17 and maturation of the active protease, the denatured states of HIV-1 protease have been studied recently by nuclear magnetic resonance (NMR) in different denaturants.18−22 To further substantiate the description of the denatured state of HIV-1 protease, we discovered that the monomer of a HIV-1 protease variant undergoes cold denaturation at temperatures well above 0 °C. Using highresolution NMR spectroscopy and optical spectroscopy, we have gained unique insight into the architecture of this denatured state. We followed the far-ultraviolet (far-UV) circular dichroism (CD) spectra spanning wavelengths from 250 to 200 nm of the mixture of folded mHIV-1-PRΔ95−99 and cold-denatured mHIV1-PRΔ95−99 (Figure 1A) over the temperature range of 3−90 °C. A dominating aromatic contribution in the far-UV region23 results in an atypical CD spectrum for a β-sheet protein. However, the reason this aromatic contribution remains dominant throughout the entire temperature titration remains to be investigated. To monitor the unfolding temperature of mHIV-1-PRΔ95−99, we recorded the mean residue ellipticity at 205 nm as a function of increasing temperature from 3 to 90 °C (Figure 1B). Surprisingly, we observed three transitions. The first and most distinct occurred at low temperatures with a Tcapp of 11

ABSTRACT: The human immunodeficiency virus-1 (HIV-1) protease is a complex protein that in its active form adopts a homodimer dominated by β-sheet structures. We have discovered a cold-denatured state of the monomeric subunit of HIV-1 protease that is populated above 0 °C and therefore directly accessible to various spectroscopic approaches. Using nuclear magnetic resonance secondary chemical shifts, temperature coefficients, and protein dynamics, we suggest that the cold-denatured state populates a compact wet globule containing transient non-native-like α-helical elements. From the linearity of the temperature coefficients and the hydrodynamic radii, we propose that the overall architecture of the cold-denatured state is maintained over the temperature range studied.

T

ransient structures in the denatured state of proteins have long been assumed to contain crucial answers to the questions of protein stability and protein (mis)folding.1 Because the denatured state usually is only marginally populated under physiological conditions, chemical denaturation has become the method of choice to stabilize the denatured state for detailed structural studies. Cold denaturation is another rare but thermodynamically possible route for gaining access to a denatured state. However, because cold denaturation for most proteins occurs far below the freezing point of water, supercooled conditions through capillary effects,2,3 encapsulation in reverse micelles,4 destabilization by the addition of a denaturant, or introduction of destabilizing mutations5,6 has been applied. To date, and to the best of our knowledge, there are only three reported proteins that undergo cold denaturation above the freezing point of water in the absence of denaturants: yeast frataxin homologue 1 (Yfh1),7 the C-terminal domain of ribosomal protein L9 (CTL9),8 and the scaffold protein for iron−sulfur (Fe−S) cluster biosynthesis in Escherichia coli, IscU.9 The three proteins were found to populate compact colddenatured states when compared to their chemically denatured counterparts with significant amounts of native and non-native transient secondary structure. © XXXX American Chemical Society

Received: November 9, 2016 Revised: February 6, 2017 Published: February 7, 2017 A

DOI: 10.1021/acs.biochem.6b01141 Biochemistry XXXX, XXX, XXX−XXX

Biochemistry

Rapid Report

Figure 1. CD and NMR measurements of mHIV-1-PRΔ95−99. (A) CD spectra of mHIV-1-PRΔ95−99 as a function of temperature: black, 5 °C; light green, 24 °C; olive, 35 °C; teal, 45 °C; dark blue, 55 °C; violet, 65 °C; mid blue, 75 °C; cyan, 85 °C. (B) CD-monitored temperature transitions of mHIV-1-PRΔ95−99. Mean residue ellipticity at 205 nm of 10 μM mHIV-1-PRΔ95−99 in 20 mM sodium phosphate (pH 6) as a function of temperature. (C and D) 15N heteronuclear single-quantum coherence spectra of 200 μM 15N-labeled mHIV-1-PRΔ95−99 in 20 mM sodium phosphate (pH 6) recorded at 5 and 25 °C. Further experimental details can be found in the Supporting Information.

± 1 °C and a ΔHvH(c)app of 304 ± 87 kJ/mol corresponding to cold denaturation. The second transition, with an apparent midpoint Tmapp of approximately 51 ± 2 °C and a ΔHvH(m)app of 133 ± 20 kJ/mol, reflected heat denaturation (Figure S1). Interestingly, ΔHvH(c)app was larger than ΔHvH(m)app, suggesting a more pronounced and cooperative cold denaturation compared to heat denaturation. Successive cooling at various stages showed the onset of irreversible denaturation from temperatures above 60 °C. Thus, the third transition visible at ∼80 °C coincided with the irreversible aggregation of the protein, possibly also affecting the size of ΔHvH(m)app of the middle transition. In contrast, the process of cooling and heating the sample from 3 to 25 °C was fully reversible. As monitored by a set of five two-dimensional 15N-enhanced heteronuclear single-quantum coherence (HSQC) spectra recorded with 10 °C intervals from 5 to 45 °C, the native state was never fully populated under any conditions that were explored. Cross peaks arising from the denatured state were still visible in all spectra recorded and could be followed through the entire temperature titration (see Figure 1C,D and Figure S3). Unfortunately, this meant that the equilibrium unfolding transitions could not be fitted to a standard equilibrium transition curve, and only apparent values are reported. However, the rate of exchange between the unfolded and folded state of mHIV-1PRΔ95−99 was found to be slow on the NMR time scale (see Figure S4). As described in detail in the Supporting Information, we next produced 15N and 13C double-isotope-labeled protein and assigned the backbone resonances of the cold-denatured state to assess the secondary chemical shifts that could confirm and further characterize possible transient secondary structures. In Figure 2A−D, we report the secondary chemical shifts for the Cα, C′, N, and HN backbone atoms of the cold-denatured state. In these plots, we make use of an intrinsic reference, i.e., the reference chemical shifts obtained at a denaturant concentration of 8 M urea. Using other reference sets24,25 gave comparable results (Figure S1). We chose to use the most naiv̈ e

Figure 2. Secondary chemical shift (SCS) analysis of the cold-denatured state of mHIV-1-PRΔ95−99 at 5 °C. SCS of (A) Cα, (B) C′, (C) 15N, and (D) 1HN nuclei calculated by the use of the intrinsic random coil reference. (E) SCS calculated by combining Cα, C′, and 15N chemical shifts using the formula Δδ = [Δ(δCα) + Δ(δC′) − 0.5Δ(δN)] as described by Reed et al.26

interpretation of the chemical shifts by applying no additional statistical analysis of the data, as the currently available secondary structure prediction algorithms in some cases appear to be insufficient for unstructured states of folded proteins. From these secondary chemical shifts, the cold-denatured state of mHIV-1PRΔ95−99 was seen to dominantly sample a random coil conformation under native conditions. However, the secondary chemical shifts also revealed three distinct regions of transient αhelicity. This conclusion was supported by the coherent behavior of the sign of the secondary chemical shifts of all four atom types (Figure 2), which in the case of α-helical structure is positive for Cα and C′ and negative for N. Subsequently, we pooled these three sets of chemical shifts accordingly26 (Figure 2E). Two of the segments corresponded to non-native regions of transient helicity located in residues I15−N25 [transient helix (TH) 1] and D30−M36 (TH 2) and a third region of V82−L90 (TH 3), overlapped with the C-terminal helix of the native state. To probe for possible changes in the ensemble structure as a function of temperature, we plotted the temperature coefficient for HN for each residue (Figure 3A,B). In general, 98% of the temperature coefficients of amide protons fall between −11 and 1 ppb/K. Temperature coefficients more positive than −4.6 ppb/ K are considered a strong indicator of amide proteins involved in hydrogen bonds, and values more negative than −7 ppb/K have a