key: cord-0257077-0az58jzm
authors: Skeens, Erin; Gadzuk-Shea, Meagan; Shah, Dilip; Bhandari, Vineet; Schweppe, Devin K.; Berlow, Rebecca B.; Lisi, George P.
title: Redox-dependent Structure and Dynamics of Macrophage Migration Inhibitory Factor Reveal Sites of Latent Allostery
date: 2021-09-06
journal: bioRxiv
DOI: 10.1101/2021.09.01.458630
sha: 62713fa13347f8a5d489bc9db766c6eba5f68ba7
doc_id: 257077
cord_uid: 0az58jzm

Macrophage migration inhibitory factor (MIF) is a multifunctional immunoregulatory protein that is a key player in the innate immune response. Given its overexpression at sites of inflammation in a wide range of diseases marked by increasingly oxidative cellular environment, a comprehensive structural understanding of how cellular redox conditions may impact the structure and function of MIF is necessary. We used solution NMR spectroscopy and mass spectrometry to investigate structural and dynamic signatures of MIF under varied solution redox conditions. Our results indicate that the MIF structure is modified and becomes increasingly dynamic in an oxidative environment, which may be a means to alter the MIF functional response in a redox-dependent manner. We identified latent allosteric sites within MIF that are redox-sensitive and mutational analysis reveals that loss of redox-responsive residues attenuates activation of the coreceptor CD74. Leveraging sites of redox-sensitivity therefore reveals an avenue to modulate MIF function in its “disease state” via structure-based drug design.

its disease state to cellular redox properties, 11 most recently by investigating its cysteine 24 residues as conformational switches. 22 Mutation of either C56 or C59 in a canonical 56 An oxidative cellular environment is one of the hallmarks of inflammatory disease, 27 characterized by excess reactive oxygen species (ROS) at sites of inflammation. 25 Despite the 28 acknowledged contribution of redox imbalance to asthma, 26,27 ARDS, 28 pulmonary fibrosis, 29 and 29 cancer, 30 as well as the link between these pathologies and MIF, very little structural work has 30 directly addressed the redox behavior of MIF. We wondered whether MIF could utilize redox 31 imbalances to act as a pro-inflammatory sensor to toggle its structure based on the chemical 32 environment of the cell. Altered conformations of MIF at local sites of inflammation could then 33 modulate its function or create different modes of interaction with accessory proteins, thereby 34 controlling downstream biological responses. In a first step toward exploring such a mechanism, 35 we investigated the impact of solution redox conditions on the MIF structure and identified 36 regions of the protein that are sensitive to oxidative environments. We used solution nuclear 37 magnetic resonance (NMR) spectroscopy to assess the redox-dependent dynamics of MIF with 38 spin relaxation experiments and used mass spectrometry to identify redox-modified cysteine 39 residues. We then leveraged these findings, using further NMR structural studies and an in vivo 40 assay, to confirm redox-sensitive amino acids as latent points of functional control in MIF. 41 42

Oxidizing solution alters MIF structural stability 44

To assess the effect of solution redox potentials on the MIF structure, we performed far-45 UV circular dichroism experiments on redox-neutral MIF samples as well as MIF samples that 46 had been oxidized or reduced. CD spectra of redox-altered MIF show characteristic a-b 47 structure, with only minor differences ( Figure S1A) . These data are consistent with numerous 48 MIF structures in the Protein Data Bank (PDB) that have no obvious alterations in the presence 49 of substrates, inhibitors, or mutations. Despite similarly folded structures, temperature-50 dependent CD experiments ( Figure S1B) show that an oxidizing environment destabilizes MIF 51 to unfolding relative to reduced and redox-neutral MIF (DTm = -4.56 °C, DDG = -2.65 kcal/mol, 52 Figure S1C ). This energetic difference (expressed per trimer) is modest, but suggests a 53 loosening of the MIF structure that may affect its conformational ensemble. 54 55

We used solution NMR to pinpoint regions of the MIF structure that are sensitive to the 57 redox environment. 1 H 15 N TROSY-HSQC spectral overlays of reduced, oxidized, and redox-58 neutral MIF ( Figure S2) show that while reduced and redox-neutral samples are spectrally 59 similar, more significant spectral changes are observed for MIF under oxidizing conditions. 60

Redox-sensitive chemical shift perturbations (relative to redox-neutral MIF) are observed 61 throughout the protein, including for residues at the N-terminal enzymatic active site, solvent 62 channel, and monomer-monomer interface ( Figures 1A) . Additionally, for oxidized MIF samples, 63 duplicate resonances are observed for 22 residues (Figures 1B, S3) , suggesting that an 64 oxidizing environment modulates an equilibrium between two conformations of MIF on the NMR 65 timescale. Sites of slow exchange localize to the monomer-monomer interface (Figure 1C) , 66

indicating that the structure of this region of the protein is strongly affected by oxidizing solution, 67 consistent with the lower thermal stability observed by CD. Perturbations to MIF are not exclusive 68 to redox-active Cys residues, at times occurring 10 -20 Å from the 56 CALC 59 motif or C80. 69 However, Cys and Met residues serve as nucleation points for our observations, with a majority 70 of redox-dependent changes to the MIF structure occurring proximal to these traditionally 71 sensitive sites ( Figure S4 ). These data are consistent with an alternate oxidized MIF 72 conformation, "oxMIF," most recently probed in a synthetic peptide comprising the MIF 56 Cysteine 80 is a selectively modified redox switch 138

Prior work with a MIF epitope 22 suggested that C80 functions as a redox molecular switch. 139

However, redox sensitivity of C80 in the full-length trimeric protein is unknown. We performed 140 quantitative mass spectrometry to explore the possibility that C80 could be selectively modified greatest variance in cystine abundance, further highlighting the reactivity and sensitivity of this 147 residue to environmental conditions. The relative abundance of cystine is also sensitive to the 148 amount of oxidative stress ( Figure S8A) . Increasing the concentration of oxidant in solution from 149 1:1 to 2:1 shifts the C80 cystine relative abundance from 83.1% to 91.3%. For comparison, 150 Figure S7B shows the expected contributions to the total ion intensity of C80 modified with IAA 151 or NEM if C80 is in a fully reduced, 50% reduced, or fully oxidized state. In addition to C80, MIF 152 also contains two cysteine residues in its 56 CALC 59 motif; however, the peptide fragment 153 containing C56 and C59 is a large 34-mer, precluding quantitation by the MS methods utilized 154

here. Although we focus on C80, we cannot rule out a significant role for other cysteines in MIF 155 redox dependence. 

To test the possibility that the redox sensitivity of MIF could reveal sites of functional 166 impact that are not otherwise obvious, we created single point mutations at two residues 167 displaying redox-dependent structural and/or dynamic changes, lysine 66 (K66) and cysteine 80 168 (C80). K66 was found to have a large variation in the R1R2 parameter in oxMIF, as well as a low MIF-induced activation of the pro-inflammatory CD74 receptor. 34 The assay is indirect, using 205 chemokines released by activated macrophages to stimulate CXCR2 receptors on neutrophils 206 (that do not directly express CD74) to induce their migration to murine lungs. 35 We found that 207 wt-MIF significantly increased the total BAL fluid cell counts over saline controls, while the K66A 208 and C80A variants showed a reduced number of total BAL fluid immune cells (Figure S11) . The 209 neutrophil influx and pulmonary edema protein levels were significantly increased in the wt-MIF 210 group relative to vehicle controls (Figures 5C, 5D) . In contrast, we observed significant 211 decreases in the groups to which the MIF variants were delivered, as compared to wt-MIF. 212

Together, our data indicate that the K66A and C80A variants have a functional effect in reducing 213 the inflammatory response of MIF in the murine lung. These findings also suggest that structural 214 and/or dynamic sensitivity of residues within MIF can provide a novel route for achieving 215 functional control beyond the known MIF active sites. 216

We used solution redox potentials to modulate the MIF structure and dynamics in order 218 to reveal changes in conformational equilibria and redox-sensitive residues that may be latent 219 functional sites within its structure. This work builds upon prior studies implying redox sensitivity 220 of MIF 11,22,36,37 as well as our recent paper highlighting a series of allosteric residues linking the 221 enzymatic (tautomerase) and CD74 binding sites of MIF. 38 Here, we demonstrate that 222 examination of MIF redox sensitivity is another useful avenue for probing function beyond its 223 Here, we asked if changes in solution redox potential could reveal some of these latent 261 sites by modulating the structural and/or dynamic signatures of MIF. We experimentally altered 262 the redox potential of wt-MIF NMR samples, enabling us to identify redox-dependent structural 263 changes in residues that may be important for function. We mutated two such residues, K66 and 264 C80, that showed redox-sensitivity in wt-MIF and observed structural effects on MIF redox 265 behavior as well as altered functional responses. Having further confirmed the propensity of C80 266 for redox-driven modification by mass spectrometry, we speculate that MIF could act as a sensor 267 that toggles its conformation depending on its cellular environment to engage with varied binding 268 partners, as was previously suggested in a conformational control scheme for the 56 CALC 59 269 motif. The more extensive but partially-overlapping set of residues affected by the redox 270 environment suggests the regulatory network controlling non-overlapping MIF functions may be 271 larger than previously reported, where additional "allosteric nodes" may be revealed only when 272 MIF is exposed to stimuli, such as environmental factors associated with inflammation, serving 273 as a natural means of activating latent residues to expand MIF functions under altered cellular 274 conditions. Speculating further, targeting small molecules to redox sensitive residues may 275 prevent MIF from engaging in certain downstream cascades. Future studies of protein-protein 276 interactions with supposed binding partners of MIF, such as thioredoxin 47,48 or ribosomal protein 277 S19, 41 as well as with small molecules, can evaluate this mechanism. Another interesting avenue 278 warranting further exploration is the redox-dependent interaction of MIF with its known inhibitors 279 or with novel compounds, as it presents an opportunity to preferentially target oxMIF, a form of 280 the protein that may appear predominately within the inflammatory state. 281 282

Wild-type (wt) and/or human MIF variants cloned into a pET11b vector were grown in 285 lysogeny broth (LB) for biochemical studies and in isotopically enriched M9 minimal medium 286 supplemented with CaCl2, MgSO4, and MEM vitamins for NMR, after adapting BL21(DE3) cells 287 to D2O. Small cultures of MIF were grown overnight in LB at 37 °C and used to inoculate cultures 288 containing 50% D2O the following morning, which were grown 8 -10 hours at 37 °C and then 289 used to inoculate cultures containing 95% D2O, which were incubated at 37 °C for 12 hours. The with an exponential window function in both the direct and indirect dimensions in 338

NMRPipe. The peak intensity and volume for each pair of slow exchanging peaks were 339 calculated in SPARKY, where peak volume was calculated by integrating each peak as a 340

Lorentzian. The minor state populations were estimated by taking the volume of the minor 341 peak and dividing by the total volume of the slow exchanging pairs. 342 NMR spin relaxation experiments were performed using TROSY-based pulse 343 sequences adapted from Palmer and coworkers. 55 Relaxation data were acquired with the 1 H 344 and 15 N carriers set to the water resonance and 120 ppm, respectively. Longitudinal relaxation 345 rates (R1) were measured with T1 delays of 0, 20, 60, 100, 200, 600, 800, and 1200 ms. Relaxation dispersion profiles were generated by plotting R2 vs. 1/tcp and exchange 357 parameters were obtained from fits of these data carried out with in-house scripts. 358

Uncertainties were obtained from replicate spectra. 359

NMR spin-relaxation rates were fit to one of five semi-empirical forms of the spectral 360 density function using model-free analysis. 56,57 Fitting of motional parameters was performed in 361 RELAX. 58,59 The criteria for inclusion of residues in the diffusion tensor estimate relied on the 362 method of Tjandra and coworkers. 47 N-H bond lengths were assumed to be 1.02 Å and the 15 N 363 chemical shift anisotropy was set to -160 ppm. During the model selection process, the diffusion 364 tensor parameters were optimized simultaneously, and model selection was repeated until the 365 optimized tensor parameters and order parameter (S 2 ) did not differ from those of the previous 366 iteration. 367

Cysteine and cystine residues of purified MIF equilibrated under each of the redox 369 conditions were alkylated in a stepwise process with iodoacetamide (IAA) followed by N-370 ethylmaleimide (NEM). Parallel reaction monitoring-mass spectrometry captured the oxidation 371 states of C80 under different redox conditions. Skyline software 60 was used to calculate the 372 relative abundance of cysteine and cystine at the C80 position, indicated by binding to IAA or 373 NEM. Additional methodological details can be found in the Supporting Information. Bronchoalveolar lavage fluid (BALF) was harvested, and total cell counts in BALF were 385 determined using the TC20 automated cell counter (Bio-Rad Laboratories, Inc., Hercules, CA). 386

The differential cell counts were performed on cells cytocentrifuged onto glass slides (Fisher 387 Scientific) stained with the Hema 3 Staining System (Fisher Diagnostics, Middletown, VA) and 388 cell differential was tabulated using light microscopy. Total protein concentration in the BAL fluid 389 was measured using the Pierce TM BCA assay kit (Thermo Scientific, Rockford, IL), as previously 390 described. 6 391 392

The authors thank J. Patrick Loria for helpful discussions. This work was supported by Rhode 394

Island Foundation Grant GR5290658 and funds from the Office of the Vice President for 395

Research at Brown University (to GPL). 396 397 Competing Interests 398

MIF Family Cytokines in Innate Immunity and Homeostasis

MIF signal transduction initiated by binding to CD74

MIF is a noncognate ligand of CXC chemokine 410 receptors in inflammatory and atherogenic cell recruitment

Macrophage Migration Inhibitory Factor or its Receptor (CD74) Attenuates Growth and Invasion of DU-413

Human Macrophage Migration Inhibitory Factor (MIF)-CD74 Amtagonists via Virtual Screening

Nanosecond Dynamics Regulate the MIF-induced Activity of CD74

Role for Macrophage Migration Inhibitory Factor in 422 Asthma

Trauma Patients with Positive 424 Cultures have Higher Levels of Circulating Macrophage Migration Inhibitory Factor

Intracellular Action of the Cytokine MIF to Modulate AP-1 Activity and the Cell Cycle 429 Through JAB-1

MIF as a Glucocorticoid-Induced Modulator of Cytokine Production

Link between Macrophage Migration Inhibitory Factor and 434 Cellular Redox Regulation

A New Cytokine Link between Rheumatoid 436 Arthritis and Atherosclerosis

Gene Polymorphisms of the Macrophage 442 Migration Inhibitory Factor and Acute Pancreatitis

Regulatory Role for Macrophage Migration Inhibitory Factor in Acute Respiratory 445 Distress Syndrome

Macrophage Migration Inhibitory Factor: A Probable Link 447 between Inflammation and Cancer

Macrophage Migration Inhibitory Factor 449

Protects Cancer Cells from Immunogenic Cell Death and Impairs Anti-tumor Responses

Macrophage Migration Inhibitory Factor Involvement 452 in Breast Cancer

Macrophage Migration Inhibitory Factor Expression in Ovarian 455 Cancer

Macrophage Migration Inhibitory Factor (MIF): Biological 458 Activities and Relation with Cancer

Macrophage Migration Inhibitory Factor (MIF) Plasma

Concentration in Critically Ill COVID-19 Patients: A Prospective Observational Study

Role of the Cysteine 81 465 Residue of Macrophage Migration Inhibitory Factor as a Molecular Redox Switch

Disulfide analysis reveals a role for macrophage migration inhibitory factor 469 (MIF) as thiol-protein oxidoreductase

Specific reduction 471 of insulin disulfides by macrophage migration inhibitory factor (MIF) with glutathione and 472 dihydrolipoamide: potential role in cellular redox processes

Reactive oxygen species 474 in inflammation and tissue injury

Homeostasis is Altered in Children with Severe Asthma: Evidence for Oxidant Stress

Lung Extracellular Matrix and Redox 481 Regulation

Oxidative Stress in Pulmonary 483 Fibrosis: A Possible Role for Redox Modulatory Therapy

Tissue Redox Activity as a Hallmark of 486 Carcinogenesis: From Early to Terminal Stages of Cancer

An Effective Method for the Discrimination of 488 Motional Anisotropy and Chemical Exchange

An Analysis of MIF Structural Features that Control Functional 491 Activation of CD74

The tautomerase active site of macrophage migration inhibitory factor is a potential target for 494 discovery of novel anti-inflammatory agents

Macrophage CD74 contributes to MIF-induced pulmonary inflammation

CXC 499 chemokine receptor-2 ligands are necessary components of neutrophil-mediated host defense in invasive 500 pulmonary aspergillosis

A 16-residue peptide fragment of macrophage migration inhibitory factor, MIF-(50-65), 503 exhibits redox activity and has MIF-like biological functions

Macrophage migration inhibitory factor as a redox-sensitive cytokine in cardiac 506 myocytes

Regulation of MIF Enzymatic Activity by an Allosteric Site at the Central Solvent Channel

Macrophage migration inhibitory factor: a regulator of innate 511 immunity

Critical role of cysteine residue 81 of macrophage migration 513 inhibitory factor (MIF) in MIF-induced inhibition of p53 activity

Ribosomal protein S19 interacts with macrophage migration inhibitory factor and attenuates its pro-517 inflammatory function

Macrophage Migration Inhibitory 520 Factor is subjected to glucose modification and oxidation in Alzheimer's Disease

Evidence for the role of an altered redox state in hyporesponsiveness 523 of synovial T cells in rheumatoid arthritis

Macroglobulin from rheumatoid arthritis synovial fluid: 525 functional analysis defines a role for oxidation in inflammation

Measuring protein reduction potentials using 15N HSQC NMR 529 spectroscopy

Redox-regulated conformational 531 changes in an SH3 domain

Rotational diffusion anisotropy of human 533 ubiquitin from N-15 NMR relaxation

Direct Association of Thioredoxin-1 (TRX) with Macrophage Migration 536

Regulatory Role of TRX on MIF Internalization and Signaling Antioxid. Redox 537 Signal

Pro-1 of macrophage 539 migration inhibitory factor functions as a catalytic base in the phenylpyruvate tautomerase activity

Redox-dependent 542 stability, protonation, and reactivity of cysteine-bound heme proteins

Heat capacity of proteins. I. Partial molar heat capacity 545 of individual amino acid residues in aqueous solution: hydration effect

Contribution of hydration and non-covalent interactions 547 to the heat capacity effect on protein unfolding

NMRPipe: a 549 multidimensional spectral processing system based on UNIX pipes

A relaxation-compensated Carr

Gill sequence for characterizing chemical exchange by NMR spectroscopy

Deviations 556 from the Simple 2-Parameter Model-Free Approach to the Interpretation of N-15 Nuclear Magnetic-557 Relaxation of Proteins

Model-Free Approach to the Interpretation of Nuclear Magnetic-559

Resonance Relaxation in Macromolecules .2. Analysis of Experimental Results

Optimisation of NMR dynamic models II. A new 565 methodology for the dual optimisation of the model-free parameters and the Brownian rotational diffusion 566 tensor

Skyline: an open source document editor for creating 569 and analyzing targeted proteomics experiments

The authors declare no competing interests. 399