Long-Lived Hydrated FMN Radicals. EPR Characterization and

Subscriber access provided by University of Sunderland

Article

Long-Lived Hydrated FMN Radicals. EPR Characterization and Implications for Catalytic Variability in Flavoproteins Arpad Rostas, Christopher Einholz, Boris Illarionov, Lorenz Heidinger, Tarek Al Said, Anna Bauss, Markus Fischer, Adelbert Bacher, Stefan Weber, and Erik Schleicher J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 09 Nov 2018 Downloaded from http://pubs.acs.org on November 9, 2018

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 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 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.

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 13 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

Journal of the American Chemical Society

Long-Lived Hydrated FMN Radicals. EPR Characterization and Implications for Catalytic Variability in Flavoproteins Arpad Rostas,† Christopher Einholz,† Boris Illarionov,‡ Lorenz Heidinger,† Tarek Al Said,† Anna Bauss,† Markus Fischer,‡ Adelbert Bacher,§ Stefan Weber† and Erik Schleicher*,† Institut für Physikalische Chemie, Albert-Ludwigs-Universität Freiburg, Germany, ‡Universität Hamburg, Institut für Lebensmittelchemie, Germany, and §Department of Chemistry, Technical University of Munich, Germany. †

KEYWORDS: FMN radical / EPR spectroscopy / ENDOR spectroscopy / electronic structure / protein–cofactor interactions

ABSTRACT: Up to now, FMN/FAD radicals could not be stabilized in aqueous solution or other protic solvents, because of rapid and efficient dismutation reactions. In this contribution, a novel system for stabilizing flavin radicals in aqueous solution is reported. Subsequent to trapping FMN in an agarose matrix, light-generated FMN radicals could be produced that were stable for days even under aerobic conditions, and their concentrations were high enough for an extensive EPR characterization. All large hyperfine couplings could be extracted by using a combination of continuouswave EPR and low-temperature ENDOR spectroscopy. To map differences in the electronic structure of flavin radicals, two exemplary proton hyperfine couplings were compared with published values from various neutral and anionic flavoprotein radicals: C(6)H and C(8)H3. It turned out that FMN•– in aqueous environment shows the largest hyperfine couplings, whereas for FMNH• under similar conditions, hyperfine couplings are at the lower end, and the values of both vary up to 30%. This finding demonstrates that protein-cofactor interactions in neutral and anionic flavoprotein radicals can alter their electron spin density in different directions. With this aqueous system that allows characterization of flavin radicals without protein interactions, and that can be extended by using selective isotope labeling, a powerful tool is now at hand to quantify interactions in flavin radicals that modulate reactivity in different flavoproteins.

Introduction Derivatives of riboflavin, FMN and FAD, are among the most frequently utilized organic cofactors in nature.1 Their redox-active part is a 7,8-dimethyl isoalloxazine moiety that is able to adopt three different oxidation states: fully reduced, semiquinone, and fully oxidized, with various states of protonation (see Supporting Scheme 1).2 Flavoproteins not only engage in hydridetransfer reactions, but also in one-electron-transfer reactions;3-5 the latter entail flavin semiquinone states as reaction intermediates. Protein-bound flavin radicals have been investigated for decades and special emphasis was given to unravel the electronic structure of the isoalloxazine moiety. This

is because knowledge on the spin density distribution is considered essential to fully understand the chemical reactivity of the isoalloxazine ring and its catalytic versatility in different protein environments. Amino acid residues surrounding a flavin cofactor alter its chemical reactivity by providing an environment required to bind and act upon a specific substrate, and by adjusting the electronic properties of the cofactor through weak interactions such as hydrogen-bonding, dipole–dipole, and π-stacking interactions. A number of chemically modified flavin derivatives in organic solvents (see, e.g., Ref. 6-12), and protein-bound neutral/anionic flavin radicals have been characterized using EPR spectroscopy (see, e.g., Ref. 13-15); however, a full quantification of the influence of the protein

1 ACS Paragon Plus Environment

Journal of the American Chemical Society 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

environment is only possible if a flavin radical without a specific protein environment, i.e., in aqueous solution, can be characterized at physiologically relevant pH values. Unfortunately, this could not be achieved up to now as FMN/FAD radicals cannot be stabilized in aqueous solution or other protic solvents due to the following reasons: fast reactions of flavin radicals with molecular oxygen occur if performed under aerobic conditions,16 and more importantly, rapid disproportionation reactions between individual flavin radicals inhibit their accumulation.9 To overcome these obstacles, we applied a novel strategy to trap FMN molecules within an aqueous environment without the necessity to alter the chemical structure of the cofactor. In short, we used an agarosegel matrix to separate the individual flavin radicals from each other in order to inhibit intermolecular electron transfer, and to minimize diffusion even at room temperature. Subsequently, neutral and anionic flavin radicals were generated by blue-light illumination and characterized by continuous-wave (cw) EPR and pulsedENDOR spectroscopy. Experimental Section Generation of FMN radicals in agarose-gel matrices. The solvent (10 mM Tris in H2O or D2O) was bubbled for 10 minutes with argon to remove oxygen. 25 mM FMN (Sigma Aldrich) and 2.5–3% (w/v) agarose (peqGold Universal Agarose) were added under continuous argon flow and stirring. The solution was heated to 323 K, argon flow was stopped and the solution was warmed up to 353 K in the dark. The hot solution was transferred into the respective EPR sample tubes in the dark and cooled to room temperature. Irradiation was performed by using continuous-wave laser light (445 nm) (Elliot Martock DHOM-M-445-600 mW) at ~1 cm distance until the color of the sample turned greenish (approximately 5–10 s). Samples were frozen and stored under liquid nitrogen unless otherwise noted. Selectively labeled FMN was synthesized by published procedures.17-18 X-band EPR spectroscopy. X-Band EPR experiments were carried out using a Bruker EMX spectrometer equipped with a Bruker X-SHQ 4119HS-W1 X-band resonator and a digital temperature control system Bruker ER 4131VT. Experimental settings were: receiver gain 6·103, modulation amplitude 0.05 mT, time constant 81.92 ms, conversion time 327.68 ms, and microwave power 20 mW. Spin counting. Spin counting was performed using a standard protocol and a nitroxide spin label as standard. As only a part of the EPR sample could be illuminated with our setup, the actual volume that contained FMN

Page 2 of 13

radicals was estimated based on determining the volume that changed its color. Q-band pulsed ENDOR spectroscopy. All pulse EPR experiments were carried out using a Bruker Elexsys E580 spectrometer equipped with a Bruker Q-FT EN 5107D2 Q-band-ENDOR resonator. For ENDOR experiments also a radio-frequency (RF) amplifier (Amplifier Research 250A250), a Tektronix TDS 684A oscilloscope and a helium temperature control system ESR900 were used. For Davies-type ENDOR spectroscopy, a π/2 pulse of 160 ns, a separation time t of 45 µs and an RF pulse of 13 µs starting 0.5 µs after the first microwave pulse were used. The pulse spacing was 500 ns in all experiments. ENDOR experiments were performed at 4 K or 80 K, and data were recorded at the maximum of the 2-pulse electron-spin echo detected EPR signal unless otherwise noted. Data analysis. Spectral simulations were carried out using the Matlab (The MathWorks, Natick, MA) package EasySpin (with its “garlic”, “salt” and “pepper” simulation routines);19 fitting was performed using self-written Matlab scripts. DFT calculations. DFT calculations were carried out with the ORCA calculation routine. A Def2-tzvp basis set and a BP86 functional were used.20 All obtained hfcs are summarized in Supporting Table 3. Atomic force mictroscopy. Measurements were carried out using a Multimode-V atomic force microscope and a Nanospcope-V microscope controller (Bruker). Rectangular silicon cantilevers (Olympus AC 240 TS) with a nominal resonance frequency of 70 kHz and a spring constant of 2.0 N/m were used. The topography pictures were acquired in dynamic mode at a constant excitation frequency (amplitude modulation mode). The cantilever oscillation was tuned to about 5% above the resonance frequency and the amplitude set point was set to 60% to 80% of the free oscillation amplitude. The acquired topography pictures were processed by firstorder flattening using the Nanoscope software. Agarose gels were produced in the same way as the EPR samples; however, protruding water was removed before imaging. Results Generation of FMN radicals. FMN radicals were generated by blue light illumination inside an EPR sample tube, but outside of an EPR spectrometer, as described in detail in the Materials section. A color change from orange to greenish-yellow was indicative of FMN reduction. As no exogenous electron donor was present in the sample, oxidation of agarose to oxyagarose, similar to the starch-to-oxystarch


Page 3 of 13 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


oxidation,21 most likely occurs, and thus, agarose serves as electron donor for flavin reduction. In order to investigate both biologically relevant protonation states, FMNH• and FMN•–, measurements were conducted at neutral pH values and at pH 9. Riboflavin can be used instead of FMN; however, its poor solubility in water leads to a significantly lower radical concentration. Using the aforementioned protocol, FMN radical concentrations of up to 25 µM as determined by spin counting could be obtained. The assumption that the radical concentration within the irradiated volume is homogeneous in space is unlikely and thus, the error margins of these numbers are rather large. Characterization of the neutral FMN radical. The flavin radicals embedded in agarose gels were first investigated using cw-EPR spectroscopy at room temperature. Whereas agarose gels without FMN did not show any EPR signal even after prolonged irradiation (data not shown), gels containing photo-bleached FMN revealed a single EPR line, centered at g~2.003, with a pronounced and complex hyperfine structure (Figure 1A). The radicals turned out to be quite long-lived: EPR measurements revealed a half life (τ) of around 24 days under aerobic conditions at room temperature (Supporting Figure 1). As motional restrictions of the FMN radicals could account for their prolonged life time, a temperature series between 250 K and 331 K was carried out (Supporting Figure 2). Two observations could be made: First, the hyperfine pattern but not the EPR signal disappeared at temperatures below the freezing point of water. Second, the intensity of the hyperfine structure increased significantly above 326 K although the hyperfine pattern itself remained identical (Supporting Figure 2 and Figure 1A, upper two panels). The upper temperature limit is 331 K as the agarose-gel matrix melts at higher temperatures, whereby radical accumulation is inhibited. The observed temperature behavior agrees well with a model in which the translational diffusion of FMN molecules is restricted by the agarose matrix, whereas at least the isoalloxazine moiety is able to rotate freely. Hence, all subsequent cwEPR experiments were performed at 331 K. To disentangle the EPR parameters including the complex hyperfine pattern, various additional experiments were carried out. First, a W-band cw-EPR spectrum was measured at 80 K to extract precise g principal values of 2.0040(5), 2.0033(5) and 2.0021(5) (Supporting Figure 3). Second, a sample in D2O afforded a spectrum without the hyperfine couplings (hfcs) of the exchangeable protons H(3) and H(5) (Figure 1A, panel three). As the spin density at H(3) is expected to be weak,22 the observed altered hyperfine pattern clearly

indicates that one large exchangeable hfc, namely H(5) is missing, which agrees with a neutral FMN radical present in this sample. Using an isotropic value of –3.7 MHz for D(5), the agreement between experiment and spectral simulation is remarkably good. Third, selectively isotope-labeled FMNs (Figure 1, panels four to eight) were used to assign the complex hyperfine pattern, in conjunction with spectral simulations and DFT calculations. Specifically, DFT calculation of FMNH• suggest that only hfcs of the atoms H(1´)(1), H(5), H(6), H(8α), N(5) and N(10) are expected to be larger than the experimental linewidth of ≈ 1.4 MHz,23 and can contribute to the observed hyperfine pattern. On the other hand, weak hfcs of, e.g., N(1), C(2) and N(3) are not expected to contribute to the hyperfine pattern. Cw-EPR spectra of [1,3-15N2]-FMNH• and [2-13C1]-FMNH• (Figure 1A) confirm this assumption as the hyperfine pattern of both spectra are identical, and both are similar to those obtained with the unlabeled sample. It is obvious that the depths of the patterns are less pronounced in the labeled samples, which can be explained by small amounts of impurities. - please insert Figure 1 around hereConsequently, only by labeling atoms that carry a significant amount of spin density, the complexity of the hyperfine pattern can be reduced. As starting point for the analysis of the hyperfine pattern, [6,8α-D4]-FMNH• is an important sample as two large proton hfcs are removed, and as the gyromagnetic ratio of γn(H)/γn(D) is ~+6.5, all deuterium hfcs are expected to be below the spectral linewidth. As a consequence, the remaining hyperfine pattern consists predominantly of signals arising from H(1´)(1), H(5), N(5) and N(10). Because of the fast motional averaging of the isoalloxazine moiety at 331 K, the hyperfine coupling originating from the two protons attached to C(1´) is expected to merge in one coupling of Aiso ~10 MHz, see also below. Using these boundary conditions, spectral simulation leads to isotropic hyperfine values of 16 and 10.3 MHz for the two nitrogen atoms N(5) and N(10), respectively (Table 1). Proton pulsed ENDOR spectroscopy using the Davies sequence at Q-band frequencies was conducted at low temperatures (Figure 2). In line with published ENDOR assignment from flavin radicals (e.g., 15, 22, 24), an ENDOR spectrum of the neutral flavin radical mainly shows hfcs of protons harboring high electron-spin density, such as H(1'), H(5), H(6), and the three H(8α) protons. On the other hand, smaller hfcs from protons with lower



electron-spin densities strongly overlap in the matrix region and cannot be read out easily. From a mere visual inspection of Figure 2A, the spectrum of FMNH• in agarose matrix is similar to published ENDOR spectra of protein-bound flavin radicals;15 however, the resonances of some of the hfcs are clearly broadened-out, specifically at radio frequencies of ~44 and 57 MHz, respectively. This can be explained by weak and/or unspecific interactions of isoalloxazine protons with solvent molecules or residues in the agarose matrix: their increased conformational freedom leads to a significant hyperfine strain. Resonances arising from H(8α) protons can be easily identified by (i), their axially symmetric shape and strong intensity at 80 K and, (ii), their significantly decreased intensity below 10 K due to slowed-down methyl-group rotation (Figure 2B).24 Moreover, hyperfine resonances arising from proton H(5) can be identified using a D2O-exchanged sample (Figure 2C) as H(5) is an exchangeable proton and consequently, its ENDOR signal is absent in this sample. - please insert Figure 2 around hereUsing these pieces of information, spectral simulations were carried out and afforded accurate values for all large proton hfcs (see Table 1 and Figure 2A). Specifically, isotropic hfcs of 10.6 MHz, ~2 MHz, – 20.0 MHz, 5.75 MHz and 7.7 MHz were obtained for protons H(1´)(1), H(1´)(2), H(5), H(6) and H(8α), respectively. - please insert Table 1 around hereCharacterization of the anionic FMN radical. FMN radicals generated at pH 9 were explored in a similar way as the aforementioned samples using both cw-EPR at elevated temperatures and low-temperature ENDOR spectroscopy. The cw-EPR spectrum of the sample at pH 9 is only slightly shifted as compared to the one at pH 7, but its hyperfine coupling pattern is considerably altered (Figure 1B). This is because N(5) is deprotonated at pH ≥ 9 and thus, one strong proton hyperfine coupling is missing in the hyperfine pattern. Moreover, the electron spin distribution is somewhat different as compared to the one of FMNH•; more specifically, a build-up of spin density on the xylene ring has been proposed in FMN•–.25 Again, the strategy for assigning hfcs includes spectral simulations of cw-EPR/ENDOR spectra and DFT calculations. Q-band proton pulsed-ENDOR spectroscopy was performed at 80 K (Figure 2D). In this sample, only H(1´), H(6) and H(8α) are expected to harbor enough electron

Page 4 of 13

spin density to be visible in the spectrum. It is apparent from a qualitative inspection of the ENDOR spectrum that signals of H(5) are indeed missing. H(8α) hfcs are increased in strength and now strongly overlap with hfcs from H(1´) and H(6). In order to assign the individual hfcs, a second ENDOR spectrum was recorded at the upper magnetic-field edge corresponding to gz. Here, only the Az components of the respective proton hyperfine tensors should be in resonance, thus reducing the complexity of the ENDOR spectrum to some extent (see Supporting Figure 4). The strategy for spectral simulations of the ENDOR spectrum was to use the manually extracted isotropic hyperfine couplings of H(6) and H(8α) as starting parameters, and to assume similar shapes of all hyperfine tensors as those obtained from DFT calculations (Supporting Table 3). As outcome, precise values for all large proton hfcs were obtained (Table 1). In detail, isotropic hfcs of 6.0 MHz, ~3.5 MHz, 8.5 MHz and 10.3 MHz were obtained for the two H(1´) protons, H(6) and H(8α), respectively. The proton hfcs were used as input parameters for the simulation of the cw-EPR spectrum in order to extract values for N(5) and N(10) (Figure 2D and Table 1). When comparing the obtained values with the ones from FMNH• it is obvious that upon N(5) deprotonation spin density is indeed transferred to the xylene ring as the isotropic hfcs of H(6) and H(8α) are significantly larger. On the other hand, spin density is reduced at the pyrazine ring, which is illustrated by the decreased values for H(1´) and N(10). Discussion The concept of this study was to create polar, nonaromatic cavities surrounding single flavin radicals, thus mimicking a flavin radical in solvent without specific interactions. Water is the obvious solvent of choice to compare our model system with protein-bound flavin radicals. As the system has been established, we are now able to shed light on the variability of the electron spin density distribution of flavin radicals with and without specific (protein) interactions. A polysaccharide agarose matrix fulfilled the requirements because agarose gels form pores, whose average size can be modified by varying the agarose concentration. In principle, the FMN molecules could either be located inside the cavities or, if the ribityl side chain strongly interacts with the agarose matrix, at or embedded into the agarose walls. Independent of where the actual location is, we have strong evidence suggesting that at least the isoalloxazine moiety is able to tumble freely. Published studies suggest that a pore radius in the range of ~100–300 nm can be assumed in our experiments.26-27 With the FMN concentration used in


Page 5 of 13 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


our experiments, ~106 FMN molecules on average can be estimated per cavity, which would be contrary to the assumption that disproportionation reactions are inhibited. To gain more information about the pore size and the actual position of the FMN molecules, we performed atomic force microscopy experiments of our samples. Whereas the surface of the agarose gel without FMN (Supporting Figure 7, left panel) is highly similar to published data,27 the surface is less smooth and the pore size is decreased significantly, when an agarose/FMN mixture is measured (Supporting Figure 7, middle panel). However, no signs of clustered FMN can be detected. These finding indicates that the pores of FMN/agarose gel mixtures harbor fewer molecules than anticipated, and that water molecules are displaced by FMN. Even more severe alterations of the agarose surface can be detected, if the gel was illuminated prior measurement, presumably due to heating effects and proton release concomitant to the oxidation of agarose (Supporting Figure 7, right panel). These first results suggest that the pore sizes in our system are different to the ones published so far, and require further investigations. Additional benefits arise from using agarose: Agarose efficiently acts as electron donor,21 which excludes the necessity for additional electron donors, and requires only short illumination times due to its transparency to blue light. It has to be mentioned in this context that the unexpectedly efficient oxidation of polysaccharides by light-activated flavins could have an impact on food stability and preservation. Finally, the oxygen diffusion in agarose appears to be drastically reduced as the flavin radical concentrations decrease only very slowly over time even under aerobic conditions. Thus, the observed properties of flavin radicals in agarose matrix indicate that their translational diffusion is effectively inhibited as the disproportionation reaction between two flavin radicals is considerably reduced. Moreover, the cw-EPR temperature behavior suggests that the space provided by the agarose matrix is large enough for a free tumbling of at least the isoalloxazine moieties, perhaps of the entire FMN molecules. The overall yield of FMN radicals could in principle be further increased, most likely because the majority of FMN molecules are reduced to their fully reduced redox state and do not contribute to the EPR signals. Titration with an electron donor in solution could help to reduce the amount of fully reduced FMN and in turn, increase the yield of FMN radicals. However, the long-term stability and its fast and easy preparation makes the system perfectly practical for spectroscopic investigations. In order to put the obtained hfcs in perspective to other flavoprotein radicals, we compared them with a

number of published hfcs of non-covalently bound FMN or FAD cofactor radicals in both protonation states. Most of these studies, however, only report proton-hfcs, in particular values from H(5), H(6), and H(8α). In detail, values from E. coli photolyase,28-29 Aspergillus niger glucose oxidase,25 Vibrio cholerae NADH:quinone oxidoreductase,30 Arabidopsis thaliana cryptochrome 1,15, 31 Avena sativa LOV2,24, 32 chorismate synthase,33 Photobacterium leiognathi lumazine protein,34 Bacillus megaterium CbiY,35 Rhodotorula gracilis D-amino acid oxidase,36 and Anabaena flavodoxin37 have been extracted and are summarized in Figure 3. The scheme makes no claim for completeness, but we tried to show at least one example per folding-type and protein family. It turned out that all extracted values for the isotropic H(5) hfcs are not reliable, most probably because only two principal components can be extracted from ENDOR data and therefore, the large uncertainty of the respective third principal component contributes to the uncertainty of the isotropic H(5) value. As a consequence, all extracted values are identical within the error range (see Supporting Figure 5) and have been excluded from further analyses. Isotropic H(6) and H(8α) hfcs of flavoproteins show significant variability (Figure 3, all H(8α) principal values are summarized in Supporting Figure 6), which is analyzed below. In general, it is apparent that both hfcs are considerably larger in anionic radicals than in neutral radicals, which again documents that decreased electron spin density at the xylene ring is common upon N(5) protonation. Whereas all H(6) hfcs are around ~8.5 MHz in anionic radicals, H(8α) hfcs differ slightly from ~9.7 MHz to ~10.7 MHz. Remarkably, aqueous FMN•– in agarose matrix shows the largest hfcs (and anisotropy) of all samples. - please insert Figure 3 around hereIn case of neutral flavin radicals, H(6) values appear to be only moderately affected by different protein surroundings, and can be subdivided into two classes: The first has values of ~5 MHz and contains two larger proteins with the cofactor buried deeply inside, and the flavin radical in agarose matrix. The second class contains all other neutral flavin radical proteins with an H(6) values of Aiso ~6 MHz. Aiso values of H(8α) in neutral radicals, on the other hand, differ to about 25% and have values from ~7.5 MHz to ~9 MHz. Although an exact analysis of the electron-spin density distribution is not possible with only two proton hfcs at hand, a few trends can be identified: first a more surface-exposed xylene



Page 6 of 13

ring results in an increase of H(8α) hfcs. Examples for this type of cofactor location are lumazine protein, flavodoxin and NADH-oxidoreductase. Second, large proteins with the cofactor buried deeply lead to a smaller H(8α) coupling as exemplified by, e.g., DNA photolyase and glucose oxidase. One has to be aware, however, that a single H-bond can modify the electron spin density significantly: DNA photolyase and plant cryptochromes for example are highly conserved in their flavin-binding motif. A single altered amino acid opposite to N(5), an AsnAsp exchange, results in a stronger H-bridge to N(5), which obviously affects both proton hfcs. In summary, it is apparent from Figure 3 that aqueous FMNH• harbors only low electron spin densities at their H(6) and H(8α) positions. In turn, protein surrounding of neutral radicals increases the spin density at H(6) and in particular at H(8α), hence shifts electron spin density towards the xylene ring. The opposite behavior, a reduced spin density at H(8α), is observed in anionic flavin radicals, if aqueous solution and protein environment are compared. At this stage, the analysis is restricted to a qualitative comparison of electron spin densities at a few protons. For a more quantitative analysis of the bandwidth of flavin radicals in solution and protein-bound, more hfcs from an in-depth EPR/ENDOR characterization including those from all nitrogens and carbons are required. To this end, various stable-isotope labeled FMNs need to be synthesized, incorporated into agarose gels or into proteins, and analyzed spectroscopically. As outcome, the influence of individual protein-cofactor interactions could be directly correlated with altered electron spin densities of the isoalloxazine moiety, which in turn sheds light on the catalytic variability in flavoproteins. On the other hand, flavin radicals in solution in combination with functionalized amphiphilic block copolymers could be used as redox-active gels, which might have potentially a broad range of applications.38 Conclusions In this contribution, a novel system for spectroscopy of flavin radicals in aqueous solution was established. EPR spectroscopy of neutral and anionic flavin radicals could be performed, and all major hfcs could be extracted. Having these values at hand, trends between different flavoprotein radicals could be unraveled. In addition to EPR/ENDOR spectroscopy used in this study, a number of other spectroscopic methods such as infrared or Raman spectroscopy could benefit from our system as for the first time, signals from flavin radicals without protein surroundings become accessible.


Page 7 of 13 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


FIGURES

Figure 2. Pulsed ENDOR spectra (black) of FMNH• in protonated buffer recorded at 80 K, at 10 K, and in deuterated buffer (A). Spectral simulations are depicted in red, the respective hyperfine tensors are depicted in the first panel. The spectrum (black) of FMN•– and its spectral simulation (red) is shown in (B).

Figure 1. Cw X-band EPR spectra (black) of FMNH• (A) and FMN•– (B) in agarose matrix, and their spectral simulations (red). Unless otherwise noted, all spectra were recorded at 331 K. Background signals are marked with an asterisk. In detail, FMNH• in H2O recorded at 298 K, FMNH• in H2O, FMNH• in D2O, selectively labeled FMNH• samples ([6,8αD4]-FMNH•, [1,3-15N2]-FMNH•, [2-13C1]-FMNH•, [4a-13C1]FMNH•, [4,10a-13C2]-FMNH•).



Page 8 of 13

Figure 3. Isotropic H(8α) and H(6) proton hfcs obtained from published data of various flavoproteins 15, 24-25, 28-35 in comparison with data published in this work. Proteins are sorted from larger to smaller H(8α) values. Anionic flavin radicals are shown in red, neutral radicals in blue. Colored boxes are used as eye-catchers. For details, see text.


Page 9 of 13 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


Table 1. Hyperfine couplings of FMNH• and FMN•– obtained from spectral simulations of cw-EPR and pulsed ENDOR data. The signs of hyperfine couplings have not been determined experimentally, but were taken from theoretical calculations. Error ranges are indicated individually for all hfcs, dependent on the accuracy of the respective spectral simulations. *The uncertainty of the Ax H(5) component was between 0–6 MHz. Hyperfine couplings / MHz

FMNH• [A1 A2 A3] H(1´)(1) H(1´)(2)

Aiso

[A1 A2 A3]

Aiso

–20.0 ± 0.5

---

ENDOR

D(5)

–3.7 ± 0.1

---

H(6)

–5.75 ± 0.1

N(5) N(10)

[7.1 7.1 9.0] ± 0.2

ACKNOWLEDGMENT This work was supported by the Hans-Fischer-Gesellschaft (to E.S.) e.V. S.W. and E.S. thank the SIBW/DFG for financing EPR instrumentation that is operated within the MagRes Center of the University of Freiburg

ABBREVIATIONS

Obtaine d from ENDOR DFT

H(8α)

E-mail:

FMN•–

6.0± 0.5 ~ 3.5 ± 1

–[0–6* 35 25] ± 0.5

2036204;

Cw, continuous-wave; hfcs, hyperfine couplings.

10.6 ± 0.3 ~2±1

H(5)

(0)761 2036222; Tel: +49 (0)761 [email protected]

7.7 ± 0.2 16.0 ± 0.3 10.3 ± 0.2

[9.5 10.0 11.5] ± 0.2

–8.5 ± 0.2

cw-EPR (D2O) ENDOR

10.3 ± 0.2

ENDOR

17.0 ± 0.2 9.0 ± 0.2

cw-EPR cw-EPR

ASSOCIATED CONTENT Supporting Information. - decay kinetics of the FMN radical, - additional EPR/ENDOR spectra, - analysis of published ENDOR data - atomic force microscopy analysis - EPR, ENDOR parameters from spectral simulations - hyperfine couplings from DFT calculations This material is available free of charge via the Internet at http://pubs.acs.org.”

AUTHOR INFORMATION Corresponding Author * Institut für Physikalische Chemie, Albert-LudwigsUniversität Freiburg, 79104, Freiburg, Germany. Fax: +49



REFERENCES

1. Macheroux, P.; Kappes, B.; Ealick, S. E., Flavogenomics – a genomic and structural view of flavin-dependent proteins. FEBS J. 2011, 278, 2625-2634. 2. Heelis, P. F., The photophysical and photochemical properties of flavins (isoalloxazines). Chem. Soc. Rev. 1982, 11, 1539. 3. Massey, V., The chemical and biological versatility of riboflavin. Biochem. Soc. Trans. 2000, 28, 283-296. 4. Joosten, V.; van Berkel, W. J. H., Flavoenzymes. Curr. Opinion Chem. Biol. 2007, 11, 195-202. 5. Fraaije, M. W.; Mattevi, A., Flavoenzymes: diverse catalysts with recurrent features. Trends Biochem. Sci. 2000, 25, 126132. 6. Walker, W. H.; Ehrenberg, A.; Lhoste, J. M., 13C studies on flavin free radicals. Biochim. Biophys. Acta 1970, 215, 166-175. 7. Walker, W. H.; Ehrenberg, A., Optical and ESR spectra of flavin neutral and cation radicals in unpolar solvents. FEBS Lett. 1969, 3, 315-318. 8. Tan, S. L. J.; Webster, R. D., Electrochemically induced chemically reversible proton-coupled electron transfer reactions of riboflavin (vitamin B2). J. Am. Chem. Soc. 2012, 134, 5954-5964. 9. Ehrenberg, A.; Müller, F.; Hemmerich, P., Basicity, visible spectra, and electron spin resonance of flavosemiquinone anions. Eur. J. Biochem. 1967, 2, 286-293. 10. Kurreck, H.; Bock, M.; Bretz, N.; Elsner, M.; Kraus, H.; Lubitz, W.; Müller, F.; Geissler, J.; Kroneck, P. M. H., Fluid solution and solidstate electron nuclear double resonance studies of flavin model compounds and flavoenzymes. J. Am. Chem. Soc. 1984, 106, 737-746. 11. Weilbacher, E.; Helle, N.; Elsner, M.; Kurreck, H.; Müller, F.; Allendoerfer, R. D., 1H, 2H, 19F, 14N ENDOR and TRIPLE resonance investigations of substituted flavin radicals in

Page 10 of 13

their different protonation states. Magn. Reson. Chem. 1988, 26, 64-72. 12. Breinlinger, E.; Niemz, A.; Rotello, V. M., Model systems for flavoenzyme activity. Stabilization of the flavin radical anion through specific hydrogen bond interactions. J. Am. Chem. Soc. 1995, 117, 5379-5380. 13. Barquera, B.; Morgan, J. E.; Lukoyanov, D.; Scholes, C. P.; Gennis, R. B.; Nilges, M. J., X- and W-band EPR and Q-band ENDOR studies of the flavin radical in the Na+-translocating NADH:quinone oxidoreductase from Vibrio cholerae. J. Am. Chem. Soc. 2003, 125, 265-275. 14. Medina, M.; Cammack, R., ENDOR and related EMR methods applied to flavoprotein radicals. Appl. Magn. Reson. 2007, 31, 457-470. 15. Schleicher, E.; Wenzel, R.; Ahmad, M.; Batschauer, A.; Essen, L.-O.; Hitomi, K.; Getzoff, E. D.; Bittl, R.; Weber, S.; Okafuji, A., The electronic state of flavoproteins: investigations with proton electron–nuclear double resonance. Appl. Magn. Reson. 2010, 37, 339-352. 16. Penzer, G. R., The chemistry of flavins and flavoproteins: aerobic photochemistry. Biochem. J. 1970, 116, 733-743. 17. Sedlmaier, H.; Müller, F.; Keller, P. J.; Bacher, A., Enzymatic-synthesis of riboflavin and FMN specifically labeled with 13C in the xylene ring. Z. Naturforsch. C 1987, 42, 425-429. 18. Illarionov, B.; Zhu, F.; Eisenreich, W.; Bacher, A.; Weber, S.; Fischer, M., Preparation of flavocoenzyme isotopologues by biotransformation of purines. J. Org. Chem. 2015, 80, 2539-2544. 19. Stoll, S.; Schweiger, A., EasySpin, a comprehensive software package for spectral simulation and analysis in EPR. J. Magn. Reson. 2006, 178, 42-55. 20. Neese, F. ORCA. An ab initio, DFT and semiempirical electronic structure package, Version 2.9; Max-Planck-Institut für Bioanorganische Chemie: Mülheim (Germany). 21. Penzkofer, A., Reduction–oxidation photocycle dynamics of flavins in starch films. International Journal of Molecular Sciences 2012, 13, 9157-9183. 10

ACS Paragon Plus Environment

Page 11 of 13 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


22. Schleicher, E.; Weber, S., Radicals in flavoproteins. Top. Curr. Chem. 2012, 321, 4166. 23. Weber, S.; Möbius, K.; Richter, G.; Kay, C. W. M., The electronic structure of the flavin cofactor in DNA photolyase. J. Am. Chem. Soc. 2001, 123, 3790-3798. 24. Brosi, R.; Illarionov, B.; Mathes, T.; Fischer, M.; Joshi, M.; Bacher, A.; Hegemann, P.; Bittl, R.; Weber, S.; Schleicher, E., Hindered rotation of a cofactor methyl group as a probe for protein–cofactor interaction. J. Am. Chem. Soc. 2010, 132, 8935-8944. 25. Okafuji, A.; Schnegg, A.; Schleicher, E.; Möbius, K.; Weber, S., G-tensors of the flavin adenine dinucleotide radicals in glucose oxidase: a comparative multifrequency electron paramagnetic resonance and electron–nuclear double resonance study. J. Phys. Chem. B 2008, 112, 3568-3574. 26. Narayanan, J.; Xiong, J.-Y.; Liu, X.-Y., Determination of agarose gel pore size: Absorbance measurements vis a vis other techniques. Journal of Physics: Conference Series 2006, 28, 83-86. 27. Pernodet, N.; Maaloum, M.; Tinnland, B., Pore size of agarose gels by atomic force microscopy. Electrophoresis 1997, 18, 55-58. 28. Fuchs, M.; Schleicher, E.; Schnegg, A.; Kay, C. W. M.; Törring, J. T.; Bittl, R.; Bacher, A.; Richter, G.; Möbius, K.; Weber, S., The gtensor of the neutral flavin radical cofactor of DNA photolyase revealed by 360-GHz electron paramagnetic resonance spectroscopy. J. Phys. Chem. B 2002, 106, 8885-8890. 29. Kay, C. W. M.; Feicht, R.; Schulz, K.; Sadewater, P.; Sancar, A.; Bacher, A.; Möbius, K.; Richter, G.; Weber, S., EPR, ENDOR and TRIPLE resonance spectroscopy on the neutral flavin radical in Escherichia coli DNA photolyase. Biochemistry 1999, 38, 16740-16748. 30. Barquera, B.; Ramirez-Silva, L.; Morgan, J. E.; Nilges, M. J., A new flavin radical signal in the Na+-pumping NADH:quinone oxidoreductase from Vibrio cholerae. An EPR/electron nuclear double resonance investigation of the role of the covalently bound flavins in subunits B and C. J. Biol. Chem. 2006, 281, 36482-36491.

31. Bouly, J.-P.; Schleicher, E.; DionisioSese, M.; Vandenbussche, F.; Van der Straeten, D.; Bakrim, N.; Meier, S.; Batschauer, A.; Galland, P.; Bittl, R.; Ahmad, M., Cryptochrome blue-light photoreceptors are activated through interconversion of flavin redox states. J. Biol. Chem. 2007, 282, 9383-9391. 32. Kay, C. W. M.; Schleicher, E.; Kuppig, A.; Hofner, H.; Rüdiger, W.; Schleicher, M.; Fischer, M.; Bacher, A.; Weber, S.; Richter, G., Blue light perception in plants. Detection and characterization of a light-induced neutral flavin radical in a C450A mutant of phototropin. J. Biol. Chem. 2003, 278, 10973-10982. 33. Macheroux, P.; Petersen, J.; Bornemann, S.; Lowe, D. J.; Thorneley, R. N. F., Binding of the oxidized, reduced, and radical flavin species to chorismate synthase. An investigation by spectrophotometry, fluorimetry, and electron paramagnetic resonance and electron nuclear double resonance. Biochemistry 1996, 35, 16431652. 34. Paulus, B.; Illarionov, B.; Nohr, D.; Roellinger, G.; Kacprzak, S.; Fischer, M.; Weber, S.; Bacher, A.; Schleicher, E., One protein, two chromophores: comparative spectroscopic characterization of 6,7-dimethyl-8ribityllumazine and riboflavin bound to lumazine protein. J. Phys. Chem. B 2014, 118, 1309213105. 35. Collins, H. F.; Biedendieck, R.; Leech, H. K.; Gray, M.; Escalante-Semerena, J. C.; McLean, K. J.; Munro, A. W.; Rigby, S. E. J.; Warren, M. J.; Lawrence, A. D., Bacillus megaterium has both a functional BluB protein required for DMB synthesis and a related flavoprotein that froms a stable radical species. PLoS ONE 2013, 8, Art.-No. e55708. 36. Kay, C. W. M.; El Mkami, H.; Molla, G.; Pollegioni, L.; Ramsay, R. R., Characterization of the covalently bound anionic flavin radical in monoamine oxidase A by electron paramagnetic resonance. J. Am. Chem. Soc. 2007, 129, 1609116097. 37. Martínez, J. I.; Alonso, P. J.; Medina, M., The electronic structure of the neutral isoalloxazine semiquinone within Anabaena flavodoxin: new insights from HYSCORE 11



Page 12 of 13

experiments. J. Magn. Reson. 2012, 218, 153162. 38. Fitzpatrick, B.; Creran, B.; Cooke, G.; Rotello, V. M., Flavin-Functionalized Amphiphilic Block Copolymer Gels. Macromolecular Chemistry and Physics 2012, 213, 1758−1767.


Page 13 of 13 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


TOC Figure


13

Long-Lived Hydrated FMN Radicals. EPR Characterization and

Recommend Documents