Article pubs.acs.org/ac
Probing the Conformational and Functional Consequences of Disulfide Bond Engineering in Growth Hormone by Hydrogen− Deuterium Exchange Mass Spectrometry Coupled to Electron Transfer Dissociation Signe T. Seger,†,‡ Jens Breinholt,*,† Johan H. Faber,† Mette D. Andersen,† Charlotte Wiberg,† Christine B. Schjødt,† and Kasper D. Rand‡ †
Global Research, Novo Nordisk A/S, Novo Nordisk Park 1, 2760 Måløv, Denmark Department of Pharmacy, University of Copenhagen, Universitetsparken 2, 2100 København Ø, Denmark
‡
S Supporting Information *
ABSTRACT: Human growth hormone (hGH), and its receptor interaction, is essential for cell growth. To stabilize a flexible loop between helices 3 and 4, while retaining affinity for the hGH receptor, we have engineered a new hGH variant (Q84C/Y143C). Here, we employ hydrogen−deuterium exchange mass spectrometry (HDX-MS) to map the impact of the new disulfide bond on the conformational dynamics of this new hGH variant. Compared to wild type hGH, the variant exhibits reduced loop dynamics, indicating a stabilizing effect of the introduced disulfide bond. Furthermore, the disulfide bond exhibits longer ranging effects, stabilizing a short α-helix quite distant from the mutation sites, but also rendering a part of the α-helical hGH core slightly more dynamic. In the regions where the hGH variant exhibits a different deuterium uptake than the wild type protein, electron transfer dissociation (ETD) fragmentation has been used to pinpoint the residues responsible for the observed differences (HDX-ETD). Finally, by use of surface plasmon resonance (SPR) measurements, we show that the new disulfide bond does not compromise receptor affinity. Our work highlight the analytical potential of HDX-ETD combined with functional assays to guide protein engineering.
Disulfide bonds play an important role for correct folding and conformational stability of many proteins.1,2 This makes the introduction of disulfide bonds an attractive choice, when engineering proteins to improve stability, thereby mimicking the stabilizing forces found in nature. There are several examples of proteins where disulfide engineering has been employed to increase stability. For instance, engineered disulfides have been used to improve the in vivo half-life of activated factor VIII,15,16 which is used for the treatment of Hemophilia A. Other examples of successful disulfide engineering include increasing the stability of T4 lysozyme,3−5,17,18 human carbonic anhydrase II,19 dihydrofolate reductase,20 subtilisin,21 trypsin inhibitor-V,22 and Arc repressor.23 While the introduction of disulfides in many cases has the desired stabilizing effect, there are also examples where the outcome has been the opposite or where no effect has been observed (e.g., subtilisin,24,25 trypsin inhibitor-V,22 and lysozyme5). In many cases, a stabilizing effect can be achieved, by introducing a disulfide bond into a flexible region of a protein,3,4 while the
hGH is a 191 amino acid single chain protein that is produced in the pituitary gland and stimulates cell growth by interaction with the transmembrane growth hormone receptor. The overall folding of human growth hormone was determined decades ago,6,7 but little is known about the loop between helices 3 and 4, as a high degree of flexibility has evaded structural analysis in this region. Figure 1 shows a crystal structure (PDB 1HWG8) of the 1:2 complex of hGH bound to the human growth hormone binding protein (hGHbp), which comprises the extracellular part the receptor. The hGH−hGHbp binding takes place sequentially: First one hGHbp molecule binds to binding site 1 of hGH, and subsequently a second hGHbp molecule binds to site 2 of hGH.9 The core of hGH helix 4 (loop3−4), respectively, while helix 2 and 3 are connected by a short loop (loop2−3). Two additional short α-helices (5 and 6 in Figure 1) are located within loop1−2. As can be seen from the hGH-hGHbp complex, helix 2 and loop3−4 are not directly involved in binding to the receptor as also shown in several studies of the interaction between hGH and hGHbp.10−14 Thus, it may be envisioned that variations in helix 2 and loop3−4 may be introduced to stabilize loop3−4, without interfering with the affinity of hGH for its receptor. Such stabilization of hGH is interesting from a pharmaceutical point of view, since recombinant hGH is used in the treatment of growth disorders. © XXXX American Chemical Society
Received: December 22, 2014 Accepted: May 15, 2015
A
DOI: 10.1021/ac504782v Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
buffer (pH 7.4, 2 mM KCl, 140 mM NaCl, 10 mM phosphate) to give a 50 μM stock solution. HDX-MS Experiments. Protein stock (1 μL, 50 pmol) was labeled in 49 μL of deuterated Tris buffer (99% D, pH 8.0 (uncorrected value), 20 mM Tris, 150 mM NaCl). Exchange times were 15 s, 1 min, 10 min, 1 h, and 3 h 20 min, and the exchange took place at room temperature (20 °C). After labeling, the protein solution was quenched in 50 μL phosphate buffer (pH 2.3, 300 mM phosphate, 50 mM TCEP) at 0 °C and immediately frozen and stored at −80 °C until use. Unlabeled control samples were prepared and stored in a similar manner from 1 μL of protein stock, 49 μL of Tris buffer (pH 8.0, 20 mM Tris, 150 mM NaCl), and 50 μL of phosphate quench buffer. 100% deuterated control samples were prepared by labeling 1 μL of protein stock in 49 μL of deuterated Tris buffer (pH 8.0 (uncorrected value), 20 mM Tris, 6 M GndCl, 150 mM NaCl) for 24 h. The samples were then quenched and frozen as described before. Labeled hGH samples were loaded onto a refrigerated nanoACQUITY UPLC system (Waters Inc.) interfaced with a Synapt G2 mass spectrometer (Waters Inc.) for online pepsin digestion (at 20 °C), chromatographic separation (0 °C) and mass analysis. Pepsin digestion (using a Poroszyme Immobilized Pepsin Cartridge) and trapping (with a Waters VanGuard Pre-Column (2.1 × 5 mm)) was performed with an isocratic flow at a rate of 200 μL/min (0.1% formic acid/CH3CN 95:5) for 3 min. Chromatographic separation was achieved with a Waters Acquity UPLC BEH C18 1.7 μm (1.0 × 100 mm) column, using a linear gradient from 92% A/8% B to 60% A/ 40% B over 7 min with a flow rate of 40 μL/min. The mobile phases consisted of A, 0.1% formic acid, and B, 0.1% formic acid in CH3CN. The ESI-MS data were acquired in positive ion mode. [Glu1]-Fibrinopetide was used as the lock mass (m/z 785.8426 for z = 2 ion) and data were collected in continuum mode. All HDX experiments were performed in triplicates to confirm the significance of detected changes in deuterium uptake (>2 × StDev). HDX-ETD Experiments. For the ETD experiments, the samples were prepared by labeling 6 μL of protein stock solution (300 pmol) in 44 μL of deuterated Tris buffer (pH 8.0 (uncorrected value), 20 mM Tris, 150 mM NaCl). The labeling took place at room temperature (20 °C) with exchange times of 15 s and 3 h 20 min. After labeling, the samples were quenched with 50 μM phosphate buffer (pH 2.3, 300 mM phosphate, 300 mM TCEP) at 0 °C and immediately frozen and stored at −80 °C until use. The ESI source was operated at predefined settings to ensure the absence of H/D scrambling as described previously.37,38 ETD was performed in the trap T-wave by reacting 1,4dicyanobenzene radical anions with quadrupole-selected z = 2 and z = 3 peptide ions. ETD experiments were performed in a targeted manner, by use of a manual acquisition list that optimized precursor selection and ETD acquisition parameters for the hGH peptides of interest based on their respective elution time and most abundant charge states observed. Data Analysis. HDX-MS raw data files were analyzed using ProteinLynx Global Server v. 2.5 (Waters Inc.), and the outputs were imported into DynamX, version 2.0.0 (Waters Inc.). HDX-ETD data were analyzed in Masslynx, version 4.1 (Waters Inc.). The measured deuterium uptake was corrected for back-exchange by normalizing against 100% deuterated controls, according to the relationship
Figure 1. Human growth hormone (hGH) in complex with the extracellular part of its receptor. hGH (gray) binds to two receptor molecules (light and dark blue) (PDB ID 1HWG8). hGH contains a four helix bundle core (helices 1−4) and two additional short α-helices (5 and 6). Helix 2 and the loop connecting helix 3 and 4 are not directly involved in the binding to the receptor.
introduction of disulfides in more rigid regions may lead to destabilization.5 However, the consequences, of introducing a new disulfide bond, for protein stability, conformation, and function are far from fully understood. A lesson learned from literature is the necessity of rational design, when introducing a new disulfide bond, and of careful assessment of the properties of the engineered variant, since the full implications of the mutations are unpredictable. Thus, analytical tools capable of monitoring the conformational response of wildtype protein to engineering are needed to exploit the full potential of such protein engineering possibilities. Measurement of the hydrogen/deuterium exchange of backbone amide groups in solution by mass spectrometry provides a sensitive technique for detecting and mapping changes in conformation and dynamics between protein states.28 Deuterium uptake can be resolved to segments of the protein sequence by pepsin proteolysis of the labeled protein sample and mass analysis of the resultant peptides. Furthermore, gas-phase fragmentation of such labeled peptides by electron transfer dissociation (ETD) mass spectrometry affords the ability to further resolve protein deuterium uptake, often down to the level of individual residues.29,30 Here, we explore the use of such high-resolution HDX-MS analyses to dissect the conformational implications of introducing a disulfide bond into human growth hormone (hGHs-s). HDX-MS analyses of hGH and hGH(s-s) show that stabling of a flexible loop in hGH to the four helix bundle core of the protein attenuates the conformational dynamics and stability in distinct regions compared to the wild type protein. These regions have subsequently been investigated in detail, by applying ETD for high-resolution HDX measurements. Furthermore, surface plasmon resonance (SPR) measurements have been used to determine the affinity of hGH(s-s) for hGHbp, and thereby correlate the measured conformational dynamics to the functional properties of the engineered hGH variant.
■
EXPERIMENTAL SECTION Protein Stocks for HDX-MS. Purified recombinant hGH and the variant hGH(s-s) were kindly provided by Novo Nordisk A/S. The lyophilized proteins were dissolved in PBS B
DOI: 10.1021/ac504782v Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry D=
Dobs Dmax D100%
focused on decreasing loop dynamics, while retaining an unaltered affinity for hGHbp, which was achieved by careful selection of the mutation sites based on functional and conformational considerations. The mutation sites were selected based on molecular loop-modeling predictions of the crystallographically unresolved loop3−4, to yield a geometrically unstrained conformation of the loop and the disulfide bridge. A model of hGH(s-s), based on the 1HWG8 crystal structure of wild type hGH, is shown in Figure 2. As depicted, hGH has
(1)
where Dobs is the observed deuterium level for the given exchange time, D100% is the observed deuterium level in the 100% control, Dmax is the theoretical maximum number of deuterium that can be incorporated at backbone amide groups in the peptide, and D is the corrected deuterium level. Dmax is taken as
Dmax = Nres − NP − 1
(2)
where Nres is the number of residues in the peptide, NP is the number of prolines, and we are disregarding the N-terminal amino acid, which is assumed unable to retain deuterium. SPR Experiments. hGHbp used for the SPR experiments was provided by Novo Nordisk A/S. It was expressed in Escherichia coli and purified using hGH-affinity and ion exchange chromatography (data not shown). The methods used for expression and purification were based on the methods described by Fuh et al.39 The anti-hGH antibody used for the SPR experiments was obtained from Fitzgerald. The analysis was carried out using a Biacore T200 instrument from GE Healthcare with 4 serially connected flow cells (Fc 1− 4). During the experiment, 8000 RU of an anti hGH antibody (anti hGH) was immobilized onto Fc 1−4 of a CM-5 chip. Subsequently, hGH or hGH(s-s) was captured in Fc 2−4, allowing Fc 1 to be used as a reference surface with only anti hGH capture antibody immobilized. The hGH capture levels varied between 20 and 50 RU in different experiments. The affinity was finally determined by injecting hGHbp as analyte in various concentrations (0−200 or 0−400 nM) to evaluate on and off rates (k a and k d , respectively). Each analyte concentration was injected twice during the analysis and buffer blank injections were included to allow for double referencing. 10 mM Glycine pH 2.5 was used to regenerate the surface between cycles. Finally, data were fitted to a 1:1 model (local Rmax) supported by the Biacore T200 software to retrieve kinetic parameters ka and kd of the interaction between hGH and hGHbp. χ2/Rmax was
