High-Affinity Interactions of Beryllium(2+) with Phosphatidylserine

Sep 5, 2017 - Beryllium has multiple industrial applications, but its manufacture is associated with a serious occupational risk of developing chronic...
1 downloads 6 Views 2MB Size
Download PDF
Subscriber access provided by University of Sussex Library

Article

High-affinity Interactions of Beryllium (2+) with Phosphatidylserine Result in a Cross-linking Effect Reducing Surface Recognition of the Lipid Yuri Ermakov, Kishore Kamaraju, Antonina Dunina-Barkovskaya, Khava S. Vishnyakova, Yegor E. Yegorov, Andriy Anishkin, and Sergei Sukharev Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.7b00644 • Publication Date (Web): 05 Sep 2017 Downloaded from http://pubs.acs.org on September 12, 2017

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 30

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

High-affinity Interactions of Beryllium (2+) with Phosphatidylserine Result in a Cross-linking Effect Reducing Surface Recognition of the Lipid Yuri A. Ermakov2†, Kishore Kamaraju1†, Antonina Dunina-Barkovskaya3†, Khava S. Vishnyakova4, Yegor E. Yegorov4, Andriy Anishkin1 and Sergei Sukharev1* 1

Department of Biology, University of Maryland, College Park, MD 20742 Frumkin Institute of Physical Chemistry and Electrochemistry, Russian Academy of Sciences, Leninsky Prosp. 31, Moscow 117071, Russia 3 Belozersky Institute of Physico-chemical Biology, Moscow State University, 119899, Moscow, Russia 4 Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Vavilov Street 32, Moscow 119991, Russia † - equally contributed 2

Correspondence:

[email protected], ph. 301-405-6923

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

ABSTRACT Beryllium has multiple industrial applications, but its manufacture is associated with serious occupational risk of developing chronic inflammation in lungs known as berylliosis, or chronic beryllium disease (CBD). Although the Be2+-induced abnormal immune responses have recently been linked to a specific MHCII allele, the nature of long-lasting granulomas is not fully understood. Here we show that Be2+ binds with a micromolar affinity to phosphatidylserine (PS), the major surface marker of apoptotic cells. Isothermal titration calorimetry (ITC) indicates that, similar to Ca2+, binding of Be2+ to PS liposomes is largely entropically driven, likely by massive desolvation. Be2+ exerts a compacting effect on PS monolayers, suggesting cross-linking through coordination by both phosphates and carboxyls in multiple configurations, which were visualized in molecular dynamics simulations. Electrostatic modification of PS membranes by Be2+ includes complete neutralization of surface charges at ~30 µM, accompanied by an increase of the boundary dipole potential. The data suggests that Be2+ can displace Ca2+ from the surface of PS and, being coordinated in a tight shell of four oxygens, it can mask headgroups from Ca2+mediated recognition by PS receptors. Indeed, 48 µM Be2+ added to IC-21 cultured macrophages specifically suppresses binding and engulfment of PS-coated silica beads or aged erythrocytes. We propose that Be2+ adsorption at the surface of apoptotic cells may potentially prevent normal phagocytosis thus causing accumulation of secondary necrotic foci and the resulting chronic inflammation.

Keywords: Beryllium, phosphatidylserine, ion coordination, cross-linking, phagocytosis.

2 ACS Paragon Plus Environment

Page 2 of 30

Page 3 of 30

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

Beryllium, a hard and light-weight non-magnetic metal has multiple industrial, military and athletic applications, but its manufacture, machining and use are associated with serious occupational risks. Aside from the acute toxicity associated with soluble and particulate Be2+ 1, inhaled particulate Be or BeO dust exerts its toxic effect in a delayed manner, causing persisting granulomas in lungs known as berylliosis, or chronic beryllium disease (CBD) 2. Non-resolving granulomatous inflammation can develop and persist in sensitive individuals long after the exposure to Be compounds. Genetic analysis of predisposition to CBD indicated that a specific allele of the MHC-II complex is responsible for making Be2+ highly immunogenic 3-7. Be2+ directly binds antigen-presenting receptors and changes their surface 8, which contributes to inflammation, but it does not fully explain the formation of primary foci of inflammation and the persisting character of the disease, both of which can potentially be linked to the impaired scavenging function of macrophages upon exposure to Be2+. It was also well-documented that Be2+ by itself induces apoptosis in macrophages 9-11. The inability of macrophages to ingest apoptotic cells was linked to the observed secondary necrosis within the granulomatous follicles causing intracellular pro-inflammatory factors to spill out, thus contributing to the non-resolving character of inflammation. Why and how Be2+ damages macrophages and prevents them from fulfilling their function of non-inflammatory removal of apoptotic bodies remains unclear. Early attempts to find sites/targets for Be2+ binding indicated nuclear proteins 12, lysosomes 13 as well as ferritin 14, but the possibility that Be2+ can also modify membrane surfaces has not been fully explored. Be2+ is a small di-cation with a strong polarizing capacity. In aqueous solution, Be2+ is tightly coordinated by a tetrahedral shell of four waters 15 with ion-oxygen distances of only 1.63 Å. At neutral pH, Be-water complexes liberate protons and form a variety of hydroxide-water clusters 16 . Be2+ was shown to be able to substitute for H+ in strong hydrogen bonds acting as a ‘tetrahedral’ proton 17. It can be coordinated by carboxylic oxygens with moderate affinity but binds stronger when it has a chance to deprotonate aliphatic hydroxyls 17. Be2+ binds especially tightly to phosphate 18. Even in the presence of strong chelating ligands such as DHBA, beryllium precipitates in the presence of phosphate 19. Our previous data 20 have suggested that binding sites for Be2+ can be created by clusters of phospholipids, especially phosphatidylserine (PS) headgroups carrying a phosphate, an amine and a carboxylate. The question is whether binding of Be2+ to PS and potential displacement of other ions can be consequential for the lipid recognition. PS, a negatively charged phospholipid, normally resides in the inner leaflet of the plasma membrane, and this asymmetric distribution is maintained by ATP-dependent aminophospholipid translocase 21 and opposed by scramblase 22. In the inner leaflet, PS electrostatically attracts components of intracellular signaling cascades through conserved domains 23-26 whose binding to phospholipids is often mediated by Ca2+. PS also serves as a cofactor in multiple Ca-dependent reactions with membrane-bound and soluble proteins 27-29. Appearance of PS on the outer surface is the primary signal of apoptosis, which is perceived by phagocytes as an “eat me” tag and is usually followed by a non-inflammatory removal of apoptotic bodies 30. Failure to contain intracellular factors during cell death results in necrosis typically associated with inflammation 29 and autoantibody production 31. PS that becomes exposed at the early stages of apoptosis is recognized by scavenger receptors on macrophages 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

and suppresses the immune and inflammatory responses 32, 33. The question is whether Be2+, with its different coordination and generally higher affinity, can interfere with Ca2+-mediated recognition processes at the cell surfaces. In the present work, by utilizing liposomes, planar bilayers, and Langmuir monolayers, we show that Be2+ binds to phosphatidylserine (PS) with micromolar affinity, which can lead to neutralization of membrane surface and condensation of lipids. To visualize possible types of Be2+ coordination, we run molecular dynamics simulations suggesting binding of ions by multiple PS headgroups, resulting in cross-linking and compaction effects. We also show that Be2+ specifically suppresses phagocytosis in cultured macrophages. The data strongly suggest the mechanism of interference of Be2+ with Ca2+-dependent mechanisms of PS recognition by macrophage receptors.

MATERIALS AND METHODS Experimental systems. Experiments were performed on three lipid model systems: liposomes, monolayers and planar bilayers. In most experiments we used mono-unsaturated DOPS (1,2dioleoyl-sn-glycero-3-phospho-L-serine) to always be above Tm and avoid possible phase transitions. For comparison, some monolayer experiments were performed with saturated DMPS (1,2-dimyristoyl-sn-glycero-3-phospho-L-serine). To reveal full electrostatic effects of divalent ions with minimal interference from other ionic species, we uniformly used 10 mM KCl buffered with 2 mM Tris-HCl (pH 7.3) as a background electrolyte. A similar strategy of studying Ca2+PS interactions in Na-based buffers at relatively low ionic strength, was used previously 34. Na+, however, is known to have higher binding affinity to PS than K+ 35 and a stronger screening capacity 34. We therefore assumed that K+ buffer would be more indifferent and would maximize the range of measured surface (zeta) potentials. Phosphate was excluded from all systems to avoid precipitation with Be2+. Langmuir monolayer were performed at 22oC to avoid evaporation from the trough in the course of slow compression. Measurements on planar lipid bilayers were also carried out at 22oC since membranes are unstable at elevated temperatures under asymmetric ionic conditions. All experiments with cultured macrophages were performed in more physiological 150 mM NaCl buffers, at 37oC. Isothermal titration calorimetry, the most precise means of determining affinities, were performed on DOPS liposomes at 30oC, halfway between 22 and 37oC. Isothermal titration calorimetry (ITC). Liposomes of mono-unsaturated DOPS from Avanti were prepared by a standard extrusion method. 4 mg of lipid was dried from chloroform solution at the bottom of a disposable glass tube under a stream of nitrogen and the residual solvent was removed under vacuum (4-12 hrs). The lipid film was rehydrated in the common buffer used for all experiments with lipid systems (10 mM KCl, 2 mM Tris-HCl, pH 7.3). Vortexed multilayer liposomes were manually extruded 3 times through a 100 nm Nucleopore filter. The size distribution estimated from dynamic light scattering was 80-110 nm. ITC experiments were carried out on a MicroCal VIP microcalorimeter. Freshly prepared DOPS liposomes (typically at 0.125-0.25 mM) were loaded in the syringe and the BeSO4 or CaCl2 solutions were prepared in the same buffer as was in the chamber. 10-15 µl injections of liposomes into the buffer with 4 ACS Paragon Plus Environment

Page 4 of 30

Page 5 of 30

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

divalents were separated by 250-300 s time intervals. Titrations were carried out at 30oC. Fitting of the integrated thermograms was done with either one-site or two-site fitting routines built into the MicroCal Origin software. Because phosphatidylserine headgroups can form a variety of binding sites for cations, the values of ∆H, ∆S, and K represent estimates for the effective parameters of the population. Langmuir monolayer experiments. All monolayer experiments were carried out on a 30 x 20 cm single-barrier 550 cm2 Teflon trough (Nima, Coventry, U.K.) housed in an Air-Clean enclosure. Registration of surface pressure was done with a standard Wilhelmy method using 10.1 mm filter paper strips attached to a pre-calibrated force sensor (NIMA model 601). Lipids were spread in a chloroform solution 15 min prior to experiment. All experiments were carried out at 22oC, with a compression speed of 10-20 cm2/min. The subphase buffer was 10 mM KCl, 2 mM Tris-HCl, pH 7.3 supplemented with different concentrations of BeSO4, MgCl2 or CaCl2 as indicated. Electrophoresis. Electrokinetic experiments were conducted on suspensions of multilayer DOPS liposomes (1 mg/ml) using the dynamic light scattering technique on a Zetasizer-II particle analyzer (Malvern Instr., UK) and custom-written analysis protocols. The details of the procedure were described previously 36. Liposomes were prepared in the same buffer as in Langmuir experiments. Treatment of the experimental data was done with the classical GouyChapman-Stern model. Due to high-concentration of lipid in the mixture, free Be2+ concentrations were re-calculated, accounting for the depletion effect 20. Intramembrane Field Compensation (IFC) measurements. The IFC method developed by Sokolov 36, 37 measures the difference between total boundary potentials φb on two sides of a planar lipid bilayer. The boundary potential φb represents the sum of the surface potential φs, i.e. the diffuse part of the electrical double layer and the inner dipole component φd. The method is based on membrane electrostriction caused by the intramembrane potential across the hydrophobic core of the bilayer (∆φin), which influences membrane capacitance. The non-linear response of the membrane to a sine wave stimulus is monitored by the amplitude of the second harmonic of the capacitive current, which reports on ∆φin. In a symmetrical case when φb on both sides of the membrane are equal, ∆φin = 0 and no electric bias is applied, membrane capacitance is minimal and the second harmonic is zero. Upon asymmetric ion binding on one side of the membrane, the difference in boundary potentials is non-zero and under short-circuit conditions ∆φb exactly translates into ∆φin generating the second harmonic. Using an externally applied electric bias one can compensate ∆φin by minimizing the second harmonic signal. The absolute value of externally applied voltage in this case is equal to ∆φb. The change of φb on one side in response to asymmetric adsorption of a polyvalent ion is compared against the other side of the bilayer serving as a control. See 20, 36 for more details. Planar lipid bilayers were formed on the 1 mm2 aperture in a Teflon chamber in a symmetric electrolyte (10 mM KCl, 3 mM Tris-HCl, pH 7.3) from DOPS solutions in n-decane (20 mg/ml). An external (±30 mV, 246 Hz) ac signal was superimposed with an automatically adjusted dc bias. The second harmonic was isolated using a DSP Lock-in-Amplifier (Stanford Research Instruments, model SR830). Molecular Dynamics simulations. The DOPS lipid bilayer system (64 molecules per leaflet) was assembled using PSFGEN plugin for VMD 38 from a pre-equilibrated DOPC bilayer as a 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

template solvated by ~10,800 TIP3P water molecules 39. After conversion of the head groups to negatively charged PS, electroneutrality was maintained by adding 256 Na+ and 128 Cl- ions as a background electrolyte. The system was equilibrated at the fixed area of 65.3 Å2 per lipid for 10 ns and then unrestrained in a flexible square simulation cell (X=Y) for 100 ns. It is known that the CHARMM 36 force field is not well adjusted for PS to maintain the experimental area per molecule in the presence of ions 40. To avoid spontaneous compaction of the bilayer and to keep area per lipid at 65 Å2 41, simulations were performed under constant tension of 28 mN/m (see explanation in Supplemental Fig. S1). We ran the equilibration of DOPS at this tension for 100 ns and used final frames of this system as the starting configuration for the following 100 and 300 ns production runs with Na+, Ca2+, and Be2+, all at the same tension. The system with divalent ions was assembled by conversion of 64 randomly chosen Na+ ions into either Ca2+ or Be2+, and other 32 to Cl- to maintain system neutrality. The simulations were performed with CHARMM36 parameters set 42, 43. Since an entry for Be2+ is absent from the original CHARMM36 parameter set, the Lennard-Jones parameters for Be2+ were adjusted to satisfy the experimentally observed Be2+-O bond distances, with the ionic radius of 0.35Å and calcium-like epsilon value of -0.120 kcal/mol. Because these parameters well reproduce the coordination number of oxygen ligands (4x) but do not reproduce the experimental residence time of water in the first hydration shell around this small ion, we will call simulated Be2+ ion “smallCa2+”. All MD simulations were performed as the NPT ensemble using NAMD 44. In simulations with Ca2+ or smallCa2+ ions, the medium contained 160 Na+, 160 Cl-, and 64 smallCa2+ or Ca2+ ions, respectively. Langevin Dynamics 45, 46 was used to maintain constant pressure (1 atm) and constant temperature (303.15K). The system was simulated with periodic boundary conditions and the particle mesh Ewald method 47. We used a real space potential-based cutoff distance of 10 Å for the potential energy of non-bonded interactions and a grid width of 1.00 Å. Energy minimizations of the initial systems and after substitution of ions in the system equilibrated with the background electrolyte were performed using the conjugate gradient method for 1000 steps. The distribution of coordination numbers for each cation was calculated as the number of oxygen atoms near the ion in the limits of the first minimum of the ion-oxygen radial distribution function (rdf), computed using “gofr” function in VMD. The rdf minma at 3.2 Å for Na+, 3.1 Å for Ca2+, and 2.1 Å for smallCa2+ were averages over all ions during a 50 ns period of simulation. Contacts of water molecules with DOPS lipids were estimated in similar way, within the cutoff distances based on the first minimum in the radial density functions for particular groups of DOPS (3.5 Å nitrogen, 3.2 Å serine carboxyl oxygens, 4.6 Å phosphorus, 3.2 Å phosphate oxygens, 3.3 Å fatty acid oxygens 3.1 Å all the heavy lipid atoms, including carbons). To quantify the exchange of ion-coordinating groups, the immediate contacts were counted within a 5 Å distance between the centers of ion and heavy lipid atoms. The exchange events by atomic groups in the coordination shell and changes of ion’s position were scored within a 50 ns period. Cells, particle preparation and estimation of phagocytic activity. Experiments were carried out on cultures of IC-21 cells (ATCC number TIB-186™ ) derived from mouse peritoneal macrophages 48. The cells were cultured in Dulbecco’s modified Eagle medium (DMEM, PanEco, Russia) supplemented with 10% of fetal bovine serum (Biolot, Russia), 2 mM Lglutamine (PanEko), and (40 U/ml) gentamycin (PanEko) in an atmosphere containing 5% CO2

6 ACS Paragon Plus Environment

Page 6 of 30

Page 7 of 30

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

at 37°С. The cells were replated twice a week. Two days before the experiments, the cells were seeded on cell culture flasks (25 cm2, Costar) at a density of 1-2×105 per flask. Five-micron silica beads (Polysciences, Inc.) were coated with a mixture of brain phosphatidylserine and phosphatidylcholine (35% PS + 65% PC) or with pure brain phosphatidylcholine (all from Avanti Polar Lipids). In control experiments, uncoated beads were tested. To prepare the lipid-coated beads, lipids in a required proportion were dissolved in 500 µl of chloroform. Beads were then added to the lipid mixture, vortexed and vacuum-dried for 24-48 hours. Immediately prior to the experiment, the beads were resuspended and washed in a physiological saline (0.9% NaCl) to a final concentration of 2.5–5 × 108 beads/ml. Large bead aggregates were removed by gravity sedimentation. The presence of lipid on the surface was independently controlled by measuring the electrophoretic mobility of beads in 10 mM KCl on a Zeta Sizer II particle analyzer. The mobility of beads corresponded to the surface charge density of 5±1 µC/cm2, which reflects the expected surface density of charged PS molecules (one per 200 A2) in the mixture with zwitterionic PC. Erythrocytes were obtained from the capillary blood of healthy volunteers using a glucometer needle. A ~50-µl droplet of blood was diluted in 1.5 ml of saline and immediately centrifuged at 800 rpm for 5 min. The supernatant was discarded and the erythrocyte sediment was resuspended in 1.5 ml PBS (Ca, Mg-containing sterile phosphate-buffered saline) and left at 4°С for 10 days. On the day of the experiment, the supernatant was discarded and the pellet of aged erythrocytes, further called ‘senescent RBCs’ 49, were resuspended in 600–800 µl of 154 mM NaCl saline supplemented with 0.5 mM Ca2+ (further ‘minimal saline’). Effect of Be2+ on phagocytic activity was evaluated by the number of beads or erythrocytes tightly bound to or engulfed by IC-21 cells, according to the protocol described in 50, 51 with some modifications. The cells in 25 cm2 flasks were incubated in 5 ml minimal saline in the presence of either ~107 silica beads or ~2×106 erythrocytes per flask, with or without Be2+ at 37°С for 50 min. Although in different preparations the concentrations of target particles could vary (±30%), in all cases, the concentrations were the same in the paired control and test flasks exposed to various Be2+ concentrations. After incubation, the cells were washed twice with minimal saline to remove unattached and loosely bound beads or erythrocytes and then examined using a Nikon Diavert microscope under phase contrast. A Nikon D5000 camera and Image J software (Scion) were used to acquire and process the images. Twenty to thirty randomly chosen viewfields on each flask were captured using a 20× objective, and the number of cell-associated beads/erythrocytes was counted in the images. At least 300 cells per flask were imaged and analyzed; the effect of Be2+ was assessed in no less than 5 independent experiments. Note that in order to avoid precipitation of Be2+ (10-100 µM), we excluded nutrients, organic ions, and phosphate or carbonate buffers from the medium in which the cells were incubated for 50 min with target particles in minimal saline. In control samples without Be2+, we observed that under such conditions the process of phagocytosis was initiated by many cells but could not be completed. For this reason, besides complete engulfment of the particle, we also counted the instances of tight association with cell surface, which was resistant to multiple washes, and partial engulfment as phagocytosis initiation events. The activity of phagocytosis was quantified by phagocytosis percentage (PP) 51, which is the fraction of active phagocytes: PP% = 100%×number of cells associated with target particles/total number of cells. Statistical analysis (calculations of mean values and SE, tests for differences between the groups and t-tests for paired samples) was performed using MS Excel or

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

MicroCal Origin (Malvern). For the comparison of the PP values obtained in different experiments, normalized values were used (PP in Be2+-free medium was taken as 100%).

8 ACS Paragon Plus Environment

Page 8 of 30

Page 9 of 30

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

RESULTS To characterize the direct binding of Be2+ to pure phosphatidylserine, we first performed a set of experiments in model systems, i.e. liposomes, planar bilayers and Langmuir monolayers. All experiments in model systems were carried out at a relatively low ionic strength (10 mM KCl) to observe maximal electrostatic effects and compare divalent ions without strong competition from background monovalents. Isothermal titration calorimetry. Fig. 1 shows the results of ITC performed on unsaturated DOPS liposomes with Ca2+ or Be2+. The fitting parameters are summarized in Table 1. Since injections of concentrated CaCl2 or BeSO4 into the control buffer produced substantial heat of dilution, the system was inverted such that liposome suspension was injected into 0.02-0.05 mM CaCl2 or BeSO4 prepared in the same buffer. The heat of liposome dilution observed in this configuration was negligibly small compared to the signal in actual experiments. The thermograms on Fig. 1A and B show that binding of Ca2+ or Be2+ to unsaturated DOPS produces positive heat, i.e. the reaction is endothermic. The curve for Ca2+ was well fit with a one-site model and produced K ~3·104 M-1. The inflected Be2+ titration curve fit with a two-site binding model predicts a smaller fraction of relatively high-affinity binding sites (K~3 ·105 M-1) with a positive ∆H of 0.6 kcal/mole of injected DOPS and a substantial positive entropy, signifying disordering. The larger fraction of lower-affinity binding sites (K~1.3 ·104 M-1) is associated with a smaller heat and entropy increase. We need to emphasize that these measurements and one- or two-site fitting procedure provide some ‘effective’ parameters for a non-uniform population of binding sites (see the MD simulation part below). The observation of a range of affinities is consistent with the hypothesis that clusters of flexible phosphatidylserine headgroups can furnish a variety of binding sites, whose affinity and occupancy can be interdependent. DOPS titrations with Ca2+ and Be2+ performed in the presence of a higher ionic strength (140 mM KCl) gave more variation. On average, they produced a similar positive ∆S of 16-20 cal M1 -1 K , a considerably lower ∆H (80-110 cal M-1), and one order of magnitude lower association constants (~3·103 M-1 for Ca2+ and ~1·104 M-1 for Be2+). All the following experiments in model systems were conducted in the presence of 10 mM KCl. Based on titrations above, the effective DOPS affinity for Be2+ is only one order of magnitude higher than for Ca2+, but with a fourfold higher positive enthalpy. The endothermic character of these reactions accompanied by a positive entropy change suggests that binding of solvated ions to solvated headgroups can be driven entropically through liberation of water from both components. To test whether Be2+ is able to displace pre-bound Ca2+ from DOPS, we prepared 2 ml of DOPS liposomes (1.25 mM) in the presence of 10 mM KCl, 2 mM CaCl2 and dialyzed them against the same buffer overnight. The purpose of this step was to equilibrate liposomes with a large volume of buffer that can be used for titration. The liposomes with pre-bound Ca2+ (at ~66 Kd) were then titrated into the dialysis buffer supplemented with 0.2 mM BeSO4. The result was unexpected as we consistently observed a large negative enthalpy (-700-800 cal M-1) of the reaction (Fig. 1C and Table 1). The inversion of the heat effect precluded fitting of the data with standard equations developed for displacement ITC 52, 53. Yet, the experiment clearly indicated that Be2+ easily outcompetes Ca2+ present in a 10-fold higher concentration. The standard competitive binding equation B1 = K1L1/(1+K1L1+K2L2), predicting the occupancy of the binding site (B1) with L1 in the presence of its competitor L2, perfectly describes the 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

observations. The reduction of effective Be2+ affinity to DOPS by two orders of magnitude (change of apparent K from 3.2×105 to 5.3×103 M-1) exactly accounts for the presence of a direct competitor with K=3.1×104 M-1 at a concentration corresponding to 66 Kd (2 mM). The striking inversion of the heat effect in the presence of Ca2+ will require further studies; however, we may speculate that the opposite heat of the displacement reaction can be due to the prior liberation of large amount of solvation water from the headgroups by the pre-bound Ca2+. This circumstance reduces the capacity of the system to increase its entropy and the negative heat effect (∆H1200 cells in total. PP(norm) = 100% ×[PP(X Be)/PP(0 Be)], where X is either 0 or 48 µM Be; the difference between PP for PS- and PC-coated beads is significant (p