IKMTSL-PTE, a Phospholipid-Based EPR Probe ... - ACS Publications

Subscriber access provided by UNIV OF REGINA

Article

IKMTSL-PTE – a Phospholipid-Based EPR Probe for Surface Electrostatic Potential of Biological Interfaces at Neutral pH: Effects of Temperature and Effective Dielectric Constant of the Solvent Maxim A Voinov, Christina T Scheid, Igor A. Kirilyuk, Dmitrii G Trofimov, and Alex I Smirnov J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.7b00592 • Publication Date (Web): 27 Feb 2017 Downloaded from http://pubs.acs.org on March 1, 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.

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 35

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

IKMTSL-PTE – a Phospholipid-Based EPR Probe for Surface Electrostatic Potential of Biological Interfaces at Neutral pH: Effects of Temperature and Effective Dielectric Constant of the Solvent

Maxim A. Voinov,1 Christina T. Scheid,1 Igor A. Kirilyuk,2,3 Dmitrii G. Trofimov,2 Alex I. Smirnov*1

1

Department of Chemistry, North Carolina State University, 2620 Yarbrough Drive, Raleigh,

North Carolina, 27695-8204, USA; 2N. N. Vorozhtsov Novosibirsk Institute of Organic Chemistry SB RAS, Lavrentiev Ave. 9, Novosibirsk 630090, Russia; 3Novosibirsk State University, Pirogova Str. 2, Novosibirsk 630090, Russia

Corresponding Author: *[email protected]

1 ACS Paragon Plus 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

Page 2 of 35

ABSTRACT The synthesis and characterization of a lipid-like electrostatic spin probe (S)-2,3bis(palmitoyloxy)propyl 2-((4-(4-(dimethylamino)-2-ethyl-1-oxyl-5,5-dimethyl-2,5-dihydro-1Himidazol-2-yl)benzyl)disulfanyl)ethyl phosphate (IKMTSL-PTE) is being reported. The intrinsic p𝐾a0 of IKMTSL-PTE was determined by X-band (9.5 GHz) EPR titration of a water-soluble model compound 4-(dimethylamino)-2-ethyl-2-(4-(((2-hydroxyethyl)disulfanyl)methyl)phenyl)5,5-dimethyl-2,5-dihydro-1H-imidazol-1-oxyl

(IKMTSL-ME)

–

an

adduct

of

the

methanethiosulfonate spin label IKMTSL and 2-mercaptoethanol. The p𝐾a0 of IKMTSL-ME in bulk aqueous solutions was found to be significantly higher than that of 4-(((2hydroxyethyl)disulfanyl)methyl)-2,2,3,5,5-pentamethylimidazolidin-1-oxyl (IMTSL-ME) – an adduct of the corresponding methanethiosulfonate spin label IMTSL and 2-mercaptoethanol (17 C, p𝐾a0 = 6.16 ± 0.03 vs. 20 oC, p𝐾a0 = 3.33 ± 0.03, respectively). A series of EPR titration

o

experiments with IKMTSL-ME in aqueous solutions containing from 0 to 60 v/v% of isopropanol have been carried out at 17 oC and 48 oC to determine effects of temperature and bulk dielectric permittivity constant, ε, on the probe pKa. A linear relationship between the probe p𝐾a and ε has been established and found to be essentially the same at 17 oC and 48 oC.

Polarity

term contributing to pKa of IKMTSL-PTE at an uncharged lipid-like interface was determined by incorporating the probe into electrically-neutral micelles formed from non-ionic detergent Triton® X-100 and it was found, similar to IMTSL-PTE, to be negative. In negatively charged DMPG lipid bilayers IKMTSL-PTE exhibits ionization transitions with significantly higher p𝐾a values than those previously reported for IMTSL-PTE (for example, at 17 oC, p𝐾ai = 7.80 ± 0.03 vs. p𝐾a0 = 5.70 ± 0.05). The surface electrostatic potentials of DMPG lipid bilayers calculated using IKMTSL-PTE titration data were found to be somewhat lower than those


Page 3 of 35

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


calculated using IMTSL-PTE. The lower values measured by IKMTSL-PTE are the likely consequences of the structure of the linker that positions the reporter nitroxide further away from the bilayer plane into aqueous phase. Overall, the ionization transitions of IKMTSL-PTE with p𝐾a values close to the neutral pH range make this lipid-like molecule a valuable spectroscopic EPR probe for studying electrostatic phenomena at biological interfaces including lipid bilayer/membrane protein systems that could be unstable in the acidic pH range accessible by the previously available probes.



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 Surface electrostatic potential has long been recognized as one of the most fundamental biophysical properties of cellular membranes that is directly involved in regulating numerous vital functions of living cells.1 During the last decades several analytical techniques have been developed to probe surface electrostatics of cellular membranes and model lipid bilayers. Generally, these techniques include NMR spectroscopy,2-4 atomic force microscopy (AFM),5,6 interaction force measurements,7 fluorescence spectroscopy,8-11 and spin probe/ spin-labeling electron paramagnetic resonance (EPR)12-22 among few others. The pros and cons of different spin probe EPR-based methods have been discussed elsewhere.23-28 Recently, we introduced a general EPR approach for measuring surface electrostatic potentials of lipid bilayers that is based on a spectroscopic observation of a reversible protonation of a pH-sensitive nitroxide covalently attached to the lipid polar head group.25,26 The synthesis and spectroscopic characterization of one such pH-sensitive lipid - IMTSL-PTE (((S)2,3-bis(palmitoyloxy)propyl

2-(((1-oxyl-2,2,3,5,5-pentamethylimidazolidin-4-yl)methyl)-

disulfanyl) ethyl phosphate)) has also been described.23,25 EPR spectra of this electrostatic probe directly report on its ionization state through changes in nitroxide magnetic parameters and the degree of rotational averaging, thus, allowing one to extract the electrostatic contribution to the interfacial pKa of the nitroxide, and, therefore, to assess the local electrostatic potential right at the lipid bilayer interface. IMTSL-PTE is an attractive probe of the surface bilayer electrostatics because of (i) a lipid-like nature of the molecule that makes it an integral part of the lipid bilayer to minimize possible perturbations, and (ii) a covalent type of the tether that links a reporter nitroxide to the lipid’s polar head and, thus, secures its location precisely at the lipid bilayer interface. The latter feature is important for assuring that the probe is reporting on the surface


Page 4 of 35

Page 5 of 35

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


potential from a well-defined region of the bilayer interface. Despite of a relatively low intrinsic pKa of IMTSL-PTE (p𝐾a0 =3.33±0.03),23 this probe has proven itself to be very useful under conditions where the protonated form of the reporter group is stabilized by a negatively charged interface, such as one present in lipid bilayers composed of anionic and mixed anioniczwitterionic phospholipids.25 For example, in lipid bilayers composed of anionic 1-palmitoyl-2oleoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (POPG), the observed interfacial p𝐾ai of IMTSLPTE reaches 5.49±0.03 of pH units.25 However, because of the low intrinsic pKa, the sensitivity range of IMTSL-PTE could fall below pH required by biochemistry of other systems including many lipid-protein complexes that are only functional within a rather limited, near neutral, pH range. One example of the latter is provided by bacterial photosynthetic reaction center (RC) which exhibits pH-dependent rates of electron and proton transfer in the quinone acceptor complex within the range >5.5 to ≈10.0 of pH units.29 These and other applications call for developing of lipid-based probes with intrinsic p𝐾a0 ’s approaching the physiological pH range (pKa ~ 7.0). Previously, we reported on the synthesis of pH-sensitive cysteine-specific spin label S-4-(4(dimethylamino)-2-ethyl-5,5-dimethyl-1-oxyl-2,5-dihydro-1H-imidazol-2-yl)benzyl

methane-

thiosulfonate (IKMTSL) and the corresponding spin-labeled phospholipid IKMTSL-PTE (Scheme 1).23 IKMTSL contains strong basic amidino functionality as a part of the imidazoline heterocycle. The intrinsic p𝐾a0 of the amidino group in IKMTSL-PTE, even being reduced by a strong electron withdrawing effect of the nitroxide moiety,30 reaches p𝐾a0 =5.98±0.03 at 23 oC,23 thus, exceeding p𝐾a0 IMTSL-PTE by 2.65±0.05 pH units. Even though the basic chemical and EPR properties of IKMTSL-PTE have been already described,23 this spin-labeled phospholipid,



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

to the best of the authors’ knowledge, has never been employed in studies of the surface membrane electrostatics by EPR.

Scheme 1. Chemical structures of a cysteine-specific spin label S-4-(4-(dimethylamino)-2-ethyl-5,5-dimethyl-1oxyl-2,5-dihydro-1H-imidazol-2-yl)benzyl methanethiosulfonate (IKMTSL), spin-labeled phospholipid (S)-2,3bis(palmitoyloxy)propyl 2-((4-(4-(dimethylamino)-2-ethyl-1-oxyl-5,5-dimethyl-2,5-dihydro-1H-imidazol-2-yl)benzyl)disulfanyl)ethyl phosphate (IKMTSL-PTE), and its water-soluble model 4-(dimethylamino)-2-ethyl-2-(4(((2-hydroxyethyl)disulfanyl)methyl)phenyl)-5,5-dimethyl-2,5-dihydro-1H-imidazol-1-oxyl (IKMTSL-ME). Chemical structure of the previously reported (S)-2,3-bis(palmitoyloxy)propyl 2-(((1-oxyl-2,2,3,5,5pentamethylimidazolidin-4-yl)methyl)disulfanyl) ethyl phosphate (IMTSL-PTE) is shown for comparison.

Here we report on a detailed characterization of IKMTSL-PTE as an EPR electrostatic probe and its use for assessing the surface electrostatic potential of model bilayers composed of 1,2dimyristoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (DMPG). Effects of temperature and bulk dielectric permittivity constant, ε, on the probe pKa have been also determined at 17 oC and 48 o

C. The implications of our findings on molecular design of spectroscopic probes, both EPR and

fluorescent, for studying the electrostatics of the lipid bilayers are also being discussed.


Page 6 of 35

Page 7 of 35

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


EXPERIMENTAL SECTION Materials. All chemicals and solvents were purchased from VWR International (Radnor, PA) or Sigma-Aldrich (St. Louis, MO), unless otherwise indicated, and used without additional purification. 1,2-dipalmitoyl-sn-glycero-3-phosphothioethanol (PTE) and 1,2-dimyristoyl-snglycero-3-phospho-(1'-rac-glycerol) (sodium salt) (DMPG) were purchased from Avanti Polar Lipids (Alabaster, AL) as chloroform solutions (>99% pure) and used without further purification. All solvents were reagent grade and used as received. S-4-(4-(dimethylamino)-2-ethyl-5,5-dimethyl-1-oxyl-2,5-dihydro-1H-imidazol-2-yl)benzyl methanethiosulfonate (IKMTSL) was synthesized using an improved procedure described in the Supporting Information. IKMTSL-2-mercaptoethanol (IKMTSL-ME) adduct was synthesized according to the literature procedure23 and purified on a preparative TLC plate (Kieselgel 60 F254; Merck, Whitehouse, NJ) with CHCl3 containing 0.5 v/v% CH3OH as eluent. IKMTSLPTE lipid was synthesized as reported previously23 and purified on a preparative TLC plate (Kieselgel 60 F254; Merck, Whitehouse, NJ) with a mixture of CHCl3, CH3OH, and H2O (v/v% 70/15/0.5) as eluent. EPR Measurements. X-band (9.5 GHz) continuous wave (CW) EPR spectra were recorded with a Varian (Palo Alto, CA) Century Series E-109 spectrometer interfaced to a PC. For measurements of liquid samples temperature was maintained with stability better than ±0.02 oC and a gradient below 0.07 oC/cm over the sample region by a digital variable temperature accessory described previously.31 Aqueous or micellar solutions or lipid suspensions were drawn into polytetrafluoroethylene (PTFE) capillary (0.81×1.12 mm, NewAge Industries, Inc., Southampton, PA), the capillary was folded and inserted into 3×4 mm clear fused quartz EPR tube open from both ends (VWR International). Typical spectrometer settings were as follows:



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

modulation amplitude was set to a quarter or a half of the narrowest line peak-to-peak line width; time constant, 64 ms; incident microwave power, 2 mW; sweep time, 30 s; scan width, 100 G. Typically, between 10 and 50 individual scans were acquired and averaged out. EPR Titration Experiments. EPR titration experiments in an aqueous phase were carried out in 50 mM phosphate buffer solutions adjusted to a required pH. In all the experiments pH values were measured with an Orion micro-combination pH electrode 98 Series (Thermo Electron Corporation, Beverly, MA, USA) three-point calibrated using two sets of standard VWR (VWR International) buffer solutions: one at pH=1.68, 4.0, 7.0 and another at pH=4.0, 7.0, and 10.0. Before the measurements, both the samples and the standard buffer solution were equilibrated at a required temperature using a PolyScience (Niles, IL) circulating bath Model 9710 with a digital temperature controller. For EPR titrations in mixed water-organic solutions, a buffer and iso-propanol were mixed in a chosen proportion (v/v) and a nitroxide from a stock solution in CH3CN was added to yield the final concentration of ca. 0.1 mM. For these mixed solutions pH was determined by measuring pH for the buffer and then calculating the actual pH values for each sample by accounting for the dilution factor. Effects of the organic medium on the activity coefficient of the hydrogen ion were estimated to be small32 and not taken into consideration. Dielectric permittivity constants, ε, of aqueous iso-propanol solutions at specific temperature and dilution factors were recalculated from the experimental data reported in the literature.33 Specifically, available experimental data for ε of water measured from 20 to 80 oC with 10 oC or 20 oC steps were fitted to an empirical relationship 𝑙𝑛(𝜀) = 𝑎 − 𝑏 ∙ 𝑇, where T is the absolute temperature;33 and the constants a and b were determined to calculate ε at 17 oC and 48 oC. The procedure was repeated for 10, 20, 30, 40, 50, and 60 weight% of iso-propanol in aqueous


Page 8 of 35

Page 9 of 35

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


mixtures. For both 17 oC and 48 oC the dielectric permittivity constants ε of the solution were found to change linearly with weight% of iso-propanol. The latter relationships allowed for calculating ε at specific v/v% of water - iso-propanol mixtures employed at this work. EPR titrations of Triton® micelles were carried out by adding an aliquot of the micellar stock solution to a buffer with a chosen pH so that the final detergent concentration would stay above the critical micelle concentration (CMC, 0.22 to 0.24 mM for Triton ® X-100)34 yielding the final nitroxide concentration of ca. 0.1 mM. For EPR titration and pH equilibration of either MLVs or ULVs approximately 50 μL of lipid dispersion was placed into a 1.5 mL Eppendorf tube and pH of the solution was adjusted by titration with a 0.3 or 0.05 M HCl solution or a 0.1 M NaOH solution. To ensure pH equilibration inside the MLVs after adjusting pH, the dispersion was subjected to at least three consecutive freeze-thaw cycles between liquid nitrogen and a water bath maintained at the temperature of the EPR experiment.

The sample was vortexed occasionally and pH was

measured at the temperature of the EPR experiment.

RESULTS AND DISCUSSION For an ionizable molecule at a polar-apolar interface, such as IKMTSL-PTE lipid incorporated into micelles or lipid bilayer membranes, the observed interfacial p𝐾ai of the reporter nitroxide probe is mainly determined by: (i) a change in the Gibbs free energy upon transferring the probe from the bulk water phase into the environment with a different electric permittivity, ΔGpol, and (ii) a contribution from the local electric potential, Ψ, affecting the equilibrium of the charged and uncharged species, ΔGel. These contributions to the Gibbs free energy are additive and, therefore, the interfacial p𝐾ai of the ionizable probe is given by:8,10



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

pol

p𝐾ai = p𝐾a0 + Δp𝐾a

+ Δp𝐾ael

Page 10 of 35

(1)

The electrostatic term, Δp𝐾ael , is related to the surface electrostatic potential, Ψ, as Δp𝐾ael = −𝑒ψ⁄ln(10)𝑘𝑇

(2)

where e is the elementary charge, k is the Boltzman’s constant, and T is the absolute temperature.

Temperature Dependence of the Intrinsic 𝐩𝑲𝟎𝐚 of IKMTSL-PTE.

The lipid-based

IKMTSL-PTE molecule (Scheme 1) has insufficient for the EPR detection solubility in water. Therefore, a water-soluble model compound IKMTSL-ME (Scheme 1), which is the exact analogue of the polar head group of IKMTSL-PTE, has been chosen to determine the intrinsic p𝐾a0 (i.e., pKa of a molecule in an aqueous solution at a low ionic strength) of this spin-labeled lipid. The hydroxyethyl group of IKMTSL-ME is believed to mimic the inductive effects of the head group attachment of PTE lipid rather well without having the basicity of the protonatable functionality being affected.23 The initial measurements of p𝐾a0 for IKMTSL-ME at 21 oC reported previously23 were extended to the temperature interval from 17 oC to 48 oC to cover the range employed for studies of DMPG vesicles reported in the next sections.


Page 11 of 35

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


Aiso

A

8.05

B

5.86

C

D E

3.02

3370

3380

3390

3400

Magnetic Field, G Figure 1. (A), (B), and (D) are representative X-band EPR spectra of IKMTSL-2-mercaptoethanol (IKMTSL-ME) measured at 17.0 oC in a series of 50 mM buffer solutions of various pH indicated next to the spectra. Vertical dashed lines indicate approximate positions of the high-field nitrogen hyperfine coupling components corresponding to the protonated and nonprotonated forms of the nitroxide and are given as guides for an eye. Approximate magnitude of the isotropic nitrogen hyperfine coupling constant, Aiso, is shown by an arrow for the pH=8.05 spectrum. The spectrum (B) at an intermediate pH=5.86 was least squares simulated as a superposition of two components shown in (C) as green and red lines; (D) is the fit residual – a difference between the experimental and the simulated spectra.

Figure 1 shows representative X-band (9.5 GHz) EPR spectra of IKMTSL-ME that are characteristic of a nitroxide in a fast rotational motion regime. Under all the experimental conditions no splitting of the high-field nitrogen hyperfine component was observed while the splitting between the components was visibly affected by the titration (Fig. 1). The absence of the splitting is consistent with a fast or an intermediate regime for chemical exchange between the protonated, RH+, and nonprotonated, R, forms of the nitroxide: 11 ACS Paragon Plus 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

RH + ⇄ R + H +

Page 12 of 35

(3)

The equilibrium constant, 𝐾 = [H + ][R]⁄[H𝑅 + ], defines the pKa of the nitroxide probe as: [R]

p𝐾a = −log10 𝐾 = 𝑝H − log10 [RH+] ,

(4)

which is a common form of the Henderson-Hasselbalch equation.

Assuming that the

experimental isotropic nitrogen hyperfine constant, Aiso, under the conditions of fast chemical exchange is proportional to the fractions of protonated and nonprotonated nitroxides with the hyperfine constants Aiso(RH+) and Aiso(R), respectively, one arrives to the following modified Henderson-Hasselbalch equation: 𝐴iso =

𝐴iso (R)∙[R]+𝐴iso (RH+ )∙[RH+ ]

=

[R]+[RH+ ]

𝐴iso (R)∙10(pH−p𝐾a ) +𝐴iso (RH+ ) 1+10(pH−p𝐾a )

,

(5)

The single-component EPR spectra were least-squares simulated to determine the corresponding isotropic nitrogen hyperfine coupling constant, Aiso, using software described earlier.35-37 Some of the spectra recorded at pH close to the nitroxide pKa (such as the spectrum Fig.1 B measured at pH=5.86) showed a small asymmetry of the high-field nitrogen hyperfine coupling component. Such shape asymmetry is not related to a change in the tumbling rate of the probe but is rather caused by an incomplete averaging of the spectra by chemical exchange between the protonated and nonprotonated nitroxide forms. Such spectra were least-squares simulated by employing a two-component model (Fig.1 C) that fitted the experimental spectra exceptionally well as illustrated by the fit residual (a difference between the experimental and simulated spectra shown in Fig. 1D).

The resulting isotropic nitrogen hyperfine coupling

constants were averaged out proportionally to the weights of the individual components yielding an effective Aiso.23,38 Figure 2 A shows Aiso as a function of pH for all four temperatures and the corresponding least-squares fits to the modified Henderson-Hasselbalch equation (5) to yield 12 ACS Paragon Plus Environment

Page 13 of 35

experimental pKa. The obtained pKa values were accepted as intrinsic pKa’s of IKMTSL-PTE and are summarized in the Table 1 and also plotted as a function of 1/T in Figure 2B. 15.6

A

15.4

Aiso

15.2 15.0 14.8 14.6 3

4

5

6

7

8

9

pH 6.3 6.2 6.1

B

6.0

pKa

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


5.9 5.8

5.7 5.6 5.5 0.0031

0.0032

0.0033

0.0034

1/T, K-1 Figure 2. (A) Experimental room temperature EPR titration curves representing pH dependence of the average isotropic nitrogen hyperfine coupling constant Aiso of IKMTSL-ME measured in bulk aqueous solutions at ionic strength I = 0.05 M at: 17.0 oC (blue filled circles), 21.0 oC (green open triangles), 30.0 oC (magenta filled triangles), and 48.0 oC (red open filled circles). The best fits to the modified Henderson−Hasselbalch eq. 5 are shown as solid lines of the same color as the corresponding symbols. (B) Temperature dependence of the experimental pKa of IKMTSL-ME (open circles) and a linear regression with a slope 𝑑p𝐾a ⁄𝑑(1⁄𝑇 ) = 1,250 ± 250 (R=0.93).

The pKa values of IKMTSL-ME demonstrate a clear increasing trend with a decrease in temperature: such a trend is well-documented in the literature for other bases.39-43 Note that the 13 ACS Paragon Plus 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

Page 14 of 35

difference between the pKa’s of IKMTSL-ME measured at 17 and 48 oC for constant ionic strength of 0.05 M was found to be relatively large (∆p𝐾a0=0.46±0.04) compared to that reported for IMTSL-ME over the same temperature range (∆p𝐾a0=0.12±0.03).23 Typically, the temperature dependence of dissociation of molecules including pHindicators is analyzed using the van’t Hoff relationship: 44 ∆𝐻 𝜃

∆𝑆 𝜃

p𝐾ao = −log10 𝐾 = 𝑅𝑇𝑙𝑛10 − 𝑅𝑙𝑛10

, (6)

where ∆𝐻 𝜃 is the enthalpy and ∆𝑆 𝜃 is the entropy of the dissociation. Dissociation of IKMTSLME decreases with increasing temperature indicating an exothermic reaction with ∆𝐻 𝜃 =24±5 kJ/mol.

This value ∆𝐻 𝜃 is about twice as large as ∆𝐻 𝜃 =11.9±0.5 kJ/mol reported for a

fluorescent pH-probe 8-hydroxypyrene-1,3,6-trisulfonic acid.44 We note that the temperature dependence of 𝑝K 0𝑎 of IMTSL-ME also showed a trend consistent with an exothermic reaction but apparent ∆𝐻 𝜃 =6±1 kJ/mol is significantly smaller when compared with ∆𝐻 𝜃 =24±5 kJ/mol for IKMTSL-ME. The observed difference in ∆𝐻 𝜃 of dissociation of IKMTSL-ME vs. IMTSL-ME could originate from a temperature-dependent rearrangement of water molecules in the vicinity of the amidino functionality undergoing a reversible protonation. Typically, the strength of hydrogen bond for CHNO-containing moieties varies from 3 to 32 kJ/mol45 and the maximum ∆𝐻 𝜃 =24±5 kJ/mol observed for IKMTSL-ME falls within this range. An increase in temperature could weaken the hydrogen bond46 and also alter its geometry,47 thus, diminishing the stabilizing effect of the hydrogen bonding on the protonated form of the nitroxide. The latter effect is expected to decrease the probe 𝑝K 0𝑎 with temperature. The reason why this effect is more pronounced for IKMTSL-ME than for IMTSL-ME could be related to the structure of the nitroxide. Specifically, for a more hydrophobic IKMTSL-ME the stabilization of its protonated species by 14 ACS Paragon Plus Environment

Page 15 of 35

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


hydrogen bonding would play a more critical role than for a mostly hydrophilic IMTSL-ME. Another factor contributing to a higher, when compared to IMTSL-ME, enthalpy of the dissociation of IKMTSL-ME (∆𝐻 𝜃 =24±5 kJ/mol vs. ∆𝐻 𝜃 =6±1 kJ/mol), could be an intrinsic hydrophobicity of the latter molecule. Specifically, a protonation (ionization) of a hydrophobic molecule placed in an “unfriendly” polar (aqueous) environment is expected to be more favorable because it would reduce the effective Gibbs free energy due to a decrease in the solvent-solute polarity mismatch (hydrophobic effect) as the probe molecule becomes more polar with the protonation. The same phenomenon is likely to be responsible for an increase of observed pKa of the amidine nitroxides with an increase of the hydrophobicity of the substituent at the exo-cyclic nitrogen atom of the amidine functionality reported earlier.30 Thus, although intuitively the hydrophobicity of the molecular probe could be felt as a disadvantage when employed in experiments with aqueous samples, the hydrophobic effect increases the observed probe pKa. The latter increase would be particularly beneficial for nitroxide-based pH-probes as many of known nitroxides exhibit pKa only in the acidic pH range.30 The changes in the dissociation constant of IKMTSL-ME with temperature could also be related to changes in the properties of the solvent. Specifically, a decrease in the solvent dielectric constant with temperature would make the formation of the charged species more energetically unfavorable and also decrease the probe pKa. In order to investigate this effect further a series of EPR titration of IKMTSL-ME in water/isopropyl alcohol solutions of different compositions has been carried out at 17 oC and 48 oC. EPR Titration of IKMTSL-ME in Water/Isopropyl Alcohol Solutions. According to eq. pol

(1) the local dielectric constant ε is expected to contribute to the interfacial p𝐾ai by ∆p𝐾a

term.

The magnitude of this effect could be readily evaluated by carrying out a series of EPR titration



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

Page 16 of 35

experiments for the model compound IKMTSL-ME in water/isopropyl alcohol solutions of different compositions. Similar to the EPR titrations in water, all X-band (9.5 GHz) EPR spectra of the IKMTSL-ME titration series revealed three well-resolved nitrogen hyperfine coupling components indicating a fast tumbling of the nitroxide on the EPR time scale. The spectra were analyzed in a way similar to those for IKMTSL-ME in water and the results of the titration are summarized in the Table 1.

Table 1. EPR Titration Data for IKMTSL-ME in Buffer/Isopropyl Alcohol Solutions of Various Compositions. T, ˚C

[i-PrOH], v/v %

bulk a

Aiso R●, G

Aiso R●H+, G

Aiso, G

pKab

17

0

81.52

15.53±0.01

14.64±0.01

0.89±0.01

6.16±0.03

20

69.31

15.36±0.006

14.51±0.005

0.85±0.008

5.67±0.02

30

62.80

15.10±0.004

14.31±0.005

0.79±0.006

5.21±0.02

40

55.98

14.97±0.005

14.21±0.006

0.76±0.008

4.95±0.02

50

48.84

14.87± 0.003

14.13± 0.004

0.74±0.005

4.62±0.02

60

41.34

14.81±0.006

14.06±0.008

0.75±0.01

4.28±0.03

21

0

79.68

15.51±0.01

14.62±0.01

0.89±0.01

5.98±0.04

30

0

76.71

15.49±0.01

14.62±0.008

0.87±0.01

5.90±0.03

48

0

70.46

15.52±0.008

14.65±0.006

0.87±0.01

5.70±0.02

10

65.14

15.38±0.01

14.55±0.01

0.83±0.01

5.35±0.03

20

59.61

15.23± 0.006

14.43± 0.007

0.80±0.009

5.00±0.02


Page 17 of 35

30

53.83

15.10± 0.01

14.34± 0.01

0.76±0.01

4.79±0.05

40

47.78

14.99±0.007

14.23± 0.009

0.76±0.011

4.39±0.03

50

41.44

14.92±0.004

14.17±0.005

0.75±0.006

4.14±0.02

60

34.78

14.90±0.003

14.13±0.004

0.77±0.005

3.93±0.01

a

Bulk dielectric constants, ε, of the solutions were calculated from available experimental literature data 33 using procedures described in the Experimental section; bp𝐾a0 is a pKa when [i-PrOH]=0.

6

5

pKa

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


4

3 2 1 30

40

50

ε

60

70

80

Figure 3. pKa of IKMTSL-ME vs. bulk dielectric constant ε of water-i-PrOH mixtures at 17 oC (○) and 48 oC (●). The corresponding data for IMTSL-ME (20 oC) were taken from ref. 19 and shown as (■). The error bars are comparable with the size of symbols. The corresponding linear regressions are shown as solid lines with parameters given in the text as eqs. 7-9,

A decrease in the dielectric permittivity  of a solution with an increase in the isopropyl alcohol concentration is expected to destabilize the protonated species and this would results in a decrease in pKa of the probe. Indeed, the pKa values from EPR titrations of IKMTSL-ME at two temperatures, 17 oC and 48 oC, clearly demonstrate this effect (Table 1 and Figure 3). Specifically, at 17 oC the pKa values increase linearly with the bulk dielectric permittivity constant, ε, as: 17 ACS Paragon Plus 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

p𝐾a = 2.2996(±0.1033) + 0.0475(±0.0017) × 𝜀,

Page 18 of 35

𝑅 2 = 0.995,

(7)

𝑅 2 = 0.985

(8)

while at 48 oC: p𝐾a = 2.0872(±0.1509) + 0.0498(±0.0028) × 𝜀,

We note that the slopes measured at the two temperatures are identical within the experimental errors and that the difference in the offsets Δp𝐾a (𝜀 = 0) ≈ −0.21 is comparable to the experimental uncertainty 𝛿(Δp𝐾a (𝜀 = 0)) = ±0.18. These data provide a clear indication that the experimental pKa of IKMTSL-ME is primarily affected by the changes in the bulk dielectric permittivity ε of the solvent and that the linear relationship between the probe pKa and the bulk ε is temperature-independent within the biologically relevant 17-48 oC interval studied here. Thus, we propose that in absence of external electric fields IKMTSL-ME could be used as an EPR molecular probe of local effective dielectric permittivity ε. Another note we make is that any interpretation of temperature dependence of the dissociation constants of pH-probes and other ionizable molecules should also account for changes in the dielectric properties of the solvent as one of the main contributing factors. Figure 3 also shows that another EPR probe, IMTSL-ME adduct we studied previously,19 also exhibits a linear dependence of pKa vs. bulk ε: p𝐾a = 0.2461(±0.1104) + 0.0383(±0.0019) × 𝜀,

𝑅 2 = 0.991

(9)

with a slope very similar to that observed for IKMTSL-ME but with the offset shifted to the region of lower pKa values.

𝐩𝐨𝐥

Determination of the Polarity Contribution, 𝚫𝐩𝑲𝐚 , to the Interfacial 𝐩𝑲𝐚 of IKMTSLPTE in Triton X-100 Micelles. The polarity-induced pKa shift was determined from EPR titration of IKMTSL-PTE at 17 and 48 oC in micelles composed of a nonionic detergent Triton 18 ACS Paragon Plus Environment

Page 19 of 35

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


X-100. The reasons for choosing the Triton X-100 micelles as a non-polar electrically neutral reference interface (=0, Δp𝐾ael = 0, Eqs. 1, 2) formed between aqueous and lipid-like phases have been discussed by us previously.23,25 Furthermore, according to recent differential scanning calorimetry (DSC) and dielectric relaxation spectroscopy studies48 Triton X-100 micelles do not exhibit any phase transitions between 17 to 48 oC, thus, making such detergent micelles a suitable model for a non-charged lipid bilayer interface over this reasonably broad temperature range. Figure 4 shows representative X-band (9.5 GHz) EPR spectra of IKMTSL-PTE in Triton X100 micelles acquired at 17 oC and 48 oC. The spectra at 48 oC (Fig. 4A) appeared in the fast motion limit and were simulated the same way (not shown) as the ones for IKMTSL-ME. The resulting isotropic nitrogen hyperfine coupling constants, Aiso, for the neutral and the ionized nitroxide species as well as the probe pKa’s determined from the analysis of the EPR titration curves are summarized in the Table 2. In a contrast, the spectra measured at 17 oC (Fig. 4 B) appeared in an intermediate motion regime. The latter spectra were simulated as a superposition of the experimental EPR spectra from the neutral and ionized nitroxide species assuming a slow exchange regime using a least-squares procedure described earlier.25,49 The fraction of the nonprotonated form of the nitroxide was plotted as a function of pH and least-squares fitted to the Henderson-Hasselbalch equation to yield the interfacial pKa (Table 2). The polarity-induced pol

pKa shift Δp𝐾a

was calculated using Eq. 1 (=0, Δp𝐾ael = 0) and experimental titration data

obtained for IKMTSL-PTE in Triton X-100 (p𝐾ai ) and IKMTSL-ME in water (Δp𝐾a0).



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

Page 20 of 35

A

B

3340

3360

3380

3400

3420

3440

Magnetic Field, G Figure 4: Experimental X-band (9.5 GHz) EPR spectra of micellar solutions of Triton X-100 doped at 1 mol% of IKMTSL-PTE lipid at 48 oC, pH=5.14 (A) and 17 oC, pH=5.27 (B).

Similar to the polarity contribution to the interfacial pKa of IMTSL-PTE reported earlier,25 pol

the polarity-induced pKa shift Δp𝐾a

for IKMTSL-PTE is also negative (Table 2). Note that for

both detergent micelles and bulk aqueous solution the pKa values of IKMTSL-labeled molecules at 48 oC are smaller than those at 17 oC. Similar to pKa values of IKMTSL-ME measured in the aqueous solution and mixed water/isopropyl alcohol solutions discussed in the preceding sections, this effect likely originates from temperature dependence of the local dielectric constant.

Indeed, the experimental pKa=5.00±0.01 of IKMTSL-PTE in Triton X-100 at 48 oC

(Table 2) coincides exactly with pKa=5.00±0.02 of IKMTSL-ME measured at the same temperature for 20 v/v % of i-PrOH aqueous solution (Table 1). The pKa’s of the same systems but at 17 oC are also close: 5.58±0.03 and 5.67±0.02 of pH units respectively (cf. Tables 1 and 2). Thus, the observed temperature effect on pKa of IKMTSL-PTE incorporated into non-polar Triton X-100 micelles could be explained by the temperature dependence of the effective 20 ACS Paragon Plus Environment

Page 21 of 35

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


dielectric constant ε rather than changes in the location of the reporter nitroxide and ionizable moiety with respect to the micelle-water interface.

Table 2. Experimental p𝐾ai and ∆p𝐾ael of IKMTSL-PTE in Non-polar Triton X-100 Micelles and DMPG Bilayers and Corresponding Surface Electrostatic Potentials, , and Effective Dielectric Constants, ε. pol ε, effective Sample T, ˚C , mV p𝐾ai ∆p𝐾ael ∆p𝐾 a

Triton X100 Micelles DMPG 100 nm ULVs

DMPG MLVs

17a

5.58±0.03

0.58±0.04

0

0

69.1±4.4

48a

5.00±0.01

0.70±0.02

0

0

58.5±4.2

17a

7.93±0.02

0.58±0.04

2.35±0.05

135±3

69.1±4.4b

48a

7.23±0.02

0.70±0.02

2.23±0.04

142±3

58.5±4.2b

17a

7.80±0.03

0.58±0.04

2.22±0.06

128±3

69.1±4.4b

48a

6.88±0.02

0.70±0.02

1.88±0.04

120±3

58.5±4.2b

a

Fraction of the unprotonated form of the nitroxide was determined from the least-squares decomposition of the experimental slow-motion X-band EPR spectra described by Voinov and Smirnov;49 b Effective dielectric constants experienced by IKMTSL-PTE incorporated into DMPG vesicles was assumed to be the same as for IKMTSL-PTE in Triton X-100 micelles at corresponding temperatures.

We note that at 17 oC IKMTSL-PTE lipid reports on somewhat higher effective dielectric constants of ε= 69.1±4.4 at the surface of the Triton X-100 micelles than ε=60±5 measured earlier by IMTSL-PTE at 20 oC.23

A part of this effect would come from temperature

dependence of effective ε. However, if we assume that the local environment for IKMTSL-PTE is similar to that of bulk 20 v/v % i-PrOH aqueous solution, the temperature dependence would account only for Δε≈1.5 increase at 17 oC vs. 20 oC using experimental literature data.33 Thus, even after correcting for temperature dependence, IKMTSL-PTE lipid still reports on slightly higher effective dielectric constants of the Triton X-100 micelle interface vs. measurements with 21 ACS Paragon Plus 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

Page 22 of 35

IMTSL-PTE. In fact, experimental pKa≈2.52 of IMTSL-PTE/Triton X-100 at 20 oC reported earlier was similar to pKa≈2.58 of IMTSL-ME in bulk 30 v/v% i-PrOH aqueous solution.23 This trend in the effective local ε observed in Triton X-100 micelle experiments is consistent with the nitroxide reporter group of IKMTSL-PTE protruding further into the aqueous phase due to structure of the linker and despite of a somewhat higher hydrophobicity of the tethered group when compared with IMTSL-PTE.

Determination of the Interfacial pKa of IKMTSL-PTE in DMPG Vesicles and the Bilayer Surface Electrostatic Potential, . The utility of IKMTSL-PTE as an EPR probe for measurements of lipid membrane surface electrostatic potential, , was evaluated using model uni- (ULV) and multilamellar vesicles (MLV) composed of an anionic phospholipid 1,2dimyristoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (DMPG) in gel (17 oC) and fluid (48 oC) bilayer phases. EPR titration of DMPG MLVs doped with 1 mol% of IKMTSL-PTE revealed intermediate-to-slow-motion EPR spectra. Similar to a titration of IMTSL in lipid bilayers,25 a decrease in pH resulted in a progressive appearance of a more immobilized IKMTSL-PTE spectral component (not shown). The latter component was attributed an ionized fraction of the nitroxide that is expected to have a slower tumbling due to additional (vs. uncharged nitroxide) electrostatic interactions with the negatively charged bilayer interface.25 Finally, we note that the changes in the EPR lineshapes upon probe protonation at the bilayer interface are primarily caused by a slower tumbling of the nitroxide rather than changes in the nitroxide magnetic parameters. Indeed, to the best of the authors’ knowledge, the maximum reported change in the isotropic nitrogen hyperfine coupling constant Aiso for pH-sensitive nitroxides of the imidazoline series is ΔAiso≈1.12 G,50 whereas the width of the individual components of IKMTSL-PTE in


Page 23 of 35

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


DMPG lipid vesicles typically exceeds 5 G (cf. Fig. 5).

Such large linewidths make the

observation of the nitroxide protonation phenomena solely on changes in the magnetic parameters nearly impossible. For the purpose of the EPR titration experiments, rather than extracting the whole set of nitroxide motion parameters from the spectra we are only interested in measuring a fraction of the nonprotonated form of IKMTSL-PTE species 𝑓 = 𝐼𝑅 ⁄(𝐼𝑅 + 𝐼𝑅𝐻 + ), where 𝐼𝑅 and 𝐼𝑅𝐻 + are EPR intensities obtained by a double integration of the corresponding EPR spectra. Therefore, for describing experimental 𝐸(𝐵) spectra we employed a simplified slow chemical exchange model that assumes a linear superposition of the spectra from the neutral, 𝐹𝑅 (𝐵), and the ionized species, 𝐹𝑅𝐻 + (𝐵),:25,49 𝐸(𝐵) = 𝑎 ∙ 𝐹𝑅 (𝐵)+ 𝑏 ∙ 𝐹𝑅𝐻 + (𝐵)

A

E

B

F

C

G

D

H

3340

3360

3380

3400

3420 3340

3360

Magnetic Field, G

3380

(9)

3400

3420

3440

Magnetic Field, G

Figure 5. Representative experimental X-band EPR spectra of IMTSL-PTE in 100 nm unilamellar DMPG vesicles at 48 oC, pH=6.91 (A) and 17 oC, pH=8.06 and the results of least-squares decomposition into the non-protonated, (B) and (F), and protonated, (C) and (G), components. (D) and (H) are the corresponding differences between the simulated and the experimental spectra.



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

Page 24 of 35

For the decomposition procedure, the spectra 𝐹𝑅 (𝐵) and 𝐹𝑅𝐻 + (𝐵) are measured experimentally and then the coefficients a and b are derived in a course of least-squares procedure involving continuous adjustment of the spectra positions due to a shift in the resonator frequency. In addition, the phase of 𝐹𝑅 (𝐵) and 𝐹𝑅𝐻 + (𝐵) was allowed to vary to account for some admixture of an out-of-phase dispersion component that could be present in EPR spectra from liquid aqueous samples.25,37,49 Figure 5 shows examples of such spectral decompositions for 100 nm DMPG ULVs doped with IKMTSL-PTE and measured at 48 oC and 17 oC. For both samples the fit residuals, i.e., differences between the experimental and simulated spectra, show only negligible deviations, thus, demonstrating the applicability of the slow exchange model. The errors in the component intensities were estimated using the standard covariance method51 and was found to be

IKMTSL-PTE, a Phospholipid-Based EPR Probe ... - ACS Publications

Recommend Documents