Reduced Flavin: NMR investigation of N(5)-H exchange mechanism, estimation of ionisation constants and assessment of properties as biological catalyst
© Macheroux et al; licensee BioMed Central Ltd. 2005
Received: 27 June 2005
Accepted: 25 November 2005
Published: 25 November 2005
The flavin in its FMN and FAD forms is a versatile cofactor that is involved in catalysis of most disparate types of biological reactions. These include redox reactions such as dehydrogenations, activation of dioxygen, electron transfer, bioluminescence, blue light reception, photobiochemistry (as in photolyases), redox signaling etc. Recently, hitherto unrecognized types of biological reactions have been uncovered that do not involve redox shuffles, and might involve the reduced form of the flavin as a catalyst. The present work addresses properties of reduced flavin relevant in this context.
N(5)-H exchange reactions of the flavin reduced form and its pH dependence were studied using the 15N-NMR-signals of 15N-enriched, reduced flavin in the pH range from 5 to 12. The chemical shifts of the N(3) and N(5) resonances are not affected to a relevant extent in this pH range. This contrasts with the multiplicity of the N(5)-resonance, which strongly depends on pH. It is a doublet between pH 8.45 and 10.25 that coalesces into a singlet at lower and higher pH values. From the line width of the 15N(5) signal the pH-dependent rate of hydrogen exchange was deduced. The multiplicity of the 15N(5) signal and the proton exchange rates are little dependent on the buffer system used.
The exchange rates allow an estimation of the pKa value of N(5)-H deprotonation in reduced flavin to be ≥ 20. This value imposes specific constraints for mechanisms of flavoprotein catalysis based on this process. On the other hand the pK ≈ 4 for N(5)-H protonation (to form N(5)+-H2) would be consistent with a role of N(5)-H as a base.
The isoalloxazine ring system is the redox active moiety of the coenzyme forms (FMN or FAD) present in flavoenzymes. These are involved in a variety of biological processes, spanning a wide spectrum with regard to the underlying chemical reaction mechanisms. These range from the classical (de)hydrogenation, the uptake, release and transport of electrons, the production of light (bioluminescence), photochemistry of the reduced form (as in photolyases), light signal transduction (as in blue light receptors), activation of oxygen and redox sensing, to name only the most prominent ones . In addition to these functions, others have emerged that appear to require hitherto unrecognized roles of reduced flavin in chemical catalysis, such as reactions that are redox-neutral (for a review see ). These will be addressed briefly below.
In reduced flavin, N(5) is crucial for the functioning of the isoalloxazine system as it is the locus involved in uptake/release of redox equivalents and is in general in contact with reacting ligands. In the intermediate pH range reduced flavin N(5) exists in its neutral, N(5)-H form, it can be protonated to yield N(5)H2+ at low pH and might exist in the anionic form N(5)- at very high pH (cf Fig. 5, below). Based on kinetic arguments  Bruice and co-workers have estimated the pKa for deprotonation of this group as being around 20 in free flavin. Urban and Lederer, on the other hand, imply a value around 15 for flavocytochrome b2  and it has been postulated that this pK might even be lower than 7 . A method useful for the assessment of the properties of reduced flavin, and specifically of N(5) is the nuclear magnetic resonance (NMR) spectroscopy. It has been used to study the interactions of apoprotein and flavin, and the perturbations of these interactions induced by binding of substrate/ligands. In these studies, we have observed that the N(5)H group in most two-electron reduced flavoproteins appears as a doublet in the 15N-NMR spectrum due to the N(5)-H coupling, while free flavin exhibits a singlet in the pH range 5–8 due to fast proton exchange. Therefore, the doublets observed in reduced flavoproteins have been interpreted as resulting from the absence or from slow proton exchange on the NMR time scale. This can result from inaccessibility of the N(5)H group to bulk water. However, no systematic study on the basic mode of N(5)-H exchange in free reduced flavin in aqueous solution is available. The great variety of chemical reactions mentioned above raises the question about the physical interactions between the apoprotein and the coenzyme that are responsible for the tuning necessary to catalyze particular reactions.
A detailed knowledge of this process and of its mechanism would provide several insights into flavoenzyme structure and function: a) It would provide a basis for the interpretation of 15N-NMR spectra of reduced flavoproteins. b) It could help understand a variety of exchange processes of substrate/product-linked hydrogens in various dehydrogenation reactions involving flavoproteins. c) It might help clarify the possible role of reduced flavin in the β-elimination of halide from β-halogenated substrates catalyzed by several flavoproteins. X-ray structural information has shown that there are no basic amino acid residues at the active centers of e.g. (oxidized) D-amino acid oxidase [7–9]. The hypothesis has thus been put forward that N(5) of the reduced enzyme flavin is the base that is involved in enzyme catalyzed elimination . In addition labeled hydrogen is abstracted from the α position and is partially incorporated into the β position of product . These data would also be compatible with the proposed role of reduced flavin N(5)-H being involved in label transposition/exchange [7–9, 12]. d) Chorismate synthase catalyzes a redox-neutral anti-1,4-elimination of a hydrogen and a phosphate group and has an absolute requirement for a reduced FMN cofactor. Again, it has been assumed that an amino acid functional group serves as a base in the elimination of the hydrogen but the recent elucidation of the 3D-structure has revealed that the only functional group that could be invoked in this process is the N(5) position of the flavin [12, 13]. e) Based on recent studies it has been proposed that in the dehydrogenation reaction catalyzed by monoamine oxidase, the flavin N(5) of the postulated C(4a)-flavin substrate adduct (which is isoelectronic with the flavin hydroquinone) acts as a base in hydrogen abstraction . Thus, a critical functional role of the flavin N(5) nitrogen in its reduced 1,5-dihydro, or its isoelectronic 4a,5-dihydro forms appears to emerge. The present study was undertaken to increase our understanding of processes involving this position and specifically the pH-dependent proton exchange and to gain information on the basicity/nucleophilicity of N(5) in reduced flavin, which provides basic information on possible mechanistic roles in flavoenzyme catalysis.
Results and Discussion
The pH-dependence of the 15N chemical shifts of reduced free flavin (N(l), N(3), N(5), N(10)) has been previously investigated at pH 5.2 – 8 , a pH range where most flavoproteins are active and stable. This study revealed that the chemical shift of the N(1) atom in reduced free flavin is strongly pH-dependent due to its ionization. A pKa value of 6.8 was calculated for this process . Also the signal of the N(10) atom shows a similar pH-dependence; the chemical shift difference between that of the neutral and of the anionic species, however, is smaller than that of the N(1) atom. This observation was rationalized previously as resulting from a change of sp2 hybridization of the N(10) atom [15, 16]. The chemical shifts of the N(3) and the N(5) atom are practically independent of pH in the range studied. In aqueous solution only a singlet was observed for the N(5)H and the N(3)H groups. The signal of the N(3)H group remains as a sharp, narrow line over the pH range studied, whereas that of the N(5)H group exhibits a sharp line at pH 5.4 and a broadened one at pH 7 and higher [15, 16].
In the present study the pH-dependence of the 15N chemical shifts of reduced free flavin was extended to the pH range 4.0 to 12.3. The 15N chemical shifts at pH 4.0 are identical, or almost identical to those observed earlier at pH 5.0 [15, 16], with the exception that the chemical shift of the N(1) atom is shifted upfield by 1.4 ppm. Also at pH 12.3 the 15N chemical shifts of the N(5) and N(10) atoms are identical with those reported previously for pH 8.0 [15, 16]. The 15N chemical shift of the N(3) atom shows a slight downfield shift (+1.9 ppm) at pH >11 indicating the onset of deprotonation of the N(3)H group. Also a small downfield shift (+0.7 ppm) is observed for the N(1) atom in the high pH region, probably related to the (partial) ionization of the N(3)H group.
Signal structure of the N(5)-H resonance
The 1J15 N(5)-H coupling constant can be determined from the spectra obtained in the slow exchange region (Fig. 1, spectra 7 and 8) and was found to be 85 Hz. This value agrees well with that determined previously in a CHCl3 solution (87.5 Hz) . The coupling constant of reduced flavin is similar to those reported for eneamines and aniline derivatives . 1J15N-H coupling constants are mainly governed by the hybridization of the nitrogen atom. A low s-character gives rise to small coupling constant (the coupling constant for tetrahedral ammonia is 73 Hz) whereas a high s-character results in a high coupling constant (for linear nitriles coupling constants as high as 135 Hz have been found ). The observed coupling constant of 85 Hz, which is similar to the value found for pyrrole (96.5 Hz ), indicates that the N(5) atom possesses approximately 31 % s-character, i.e. it is highly sp2 hybridized. The pH-dependence of the shape of the N(5)-resonance was not affected by buffer systems like Tris, phosphate and borate.
Determination of N(5)-Hydrogen exchange rates
k = 1/τe = 4 π pA pB Δν2/(Δν1/2 - Δν1/2°) (1)
k = 1/τe = π (Δν1/2 - Δν1/2°) (2)
with k as the exchange rate; τe the lifetime of each state, Δν1/2 the half line width of the resonance line, Δν1/2° the half line width of the proton decoupled signal which should be equal to the so-called natural line width (4 Hz, see above) under conditions of fast proton exchange. Δν is the difference of the resonance frequencies of the exchanging species under the condition of slow exchange, i.e. equal to the coupling constant 1JN-H, pA and pB are the molar fractions of the exchanging species in state A and B, respectively. The exchange rates have also been determined in phosphate and borate buffers at a few selected pH values; they are in the range of those determined in Tris buffer. The signal structure at pH 10.5 (transition from doublet to singlet) indicates that the coalescence point is close to this pH value. This is supported by the fact that the exchange rates at this pH, calculated according to eqs. 1 and 2, yield very similar values, i.e. 191 s-1 and 186 s-1, respectively.
Possible mechanisms of exchange
Line widths of the 15N resonance signal of the reduced FMN N(5) atom as function of pH and estimation of the N(5)-H pKa.
Line width (Δν1/2-Δν1/2°) (Hz)
Exchange rate of N(5)-H (s-1)
Estimation of ionisation constants for the reduced flavin N(5)
The pKa of the flavin N(5)-H corresponding to the following exchange process:
can be estimated based on the general rates of proton transfer (k1 and k-1, as indicated in equation 3) between two exchanging species connected to the equilibrium constants Kflavin-N(5) and KH2O as in equation 4:
Although a doublet for N(5)H of reduced free flavin can only be observed in the narrow pH range 8.5 – 10.5, it has been documented for reduced flavoproteins at pH values as low as 5 [15, 27, 28]. This suggests that in specific, reduced flavoproteins N(5)-H exchange is slow and that access of bulk solvent to this position is hindered. Occurrence of a doublet was observed in reduced thioredoxin reductase in which the N(1) atom is protonated . Conversely, in reduced flavodoxins and many other flavoproteins the flavin N(1) position is not accessible for protonation  the flavin existing in the anionic form as long as the protein does not unfold (pH-dependent process). It thus appears that the rate of exchange at N(5)-H is not dependent on the ionisation state at N(1)-H but will be dictated by the environment and accessibility of solvent of the specific protein. However, no such conclusion can be drawn for the pH range 8.5 – 10.5 where the doublet of the N(5)-H is an inherent property of the reduced flavin. Unlike the N(5)H resonance in reduced free flavin the N(3)H resonance shows a narrow singlet over the whole pH range studied, indicating that proton exchange is fast. Therefore, the appearance of a doublet for this resonance in flavoproteins is compatible with hindered accessibility of this position for bulk water as previously interpreted .
The estimation of the pKa value of ≥ 20 for the N(5) group in reduced flavin might be an useful parameter for the formulation of mechanisms involving it. For instance Urban and Lederer [4, 5] have postulated a deprotonated N(5) with a pKa of about 15 or even as low as 7  as a catalytically relevant species in flavocytochrome b2. While our data cannot exclude such a pKa, it imposes specific energetic restraints for its formulation. On the other hand the reduced flavin N(5)-H with an estimated pKb, ≥ 4 for the free molecule possesses sufficient basicity to act as a base according to mechanism 3b (Fig. 3). Clearly, this property can be modulated by the protein environment to suit specific purposes. The conclusion is thus that the reduced flavin N(5)-H has properties that qualify it as a base catalyst for biochemical reactions. This could be realized in several cases: The recent elucidation of the structure of chorismate synthase [12, 13] suggests that N(5)-H of the reduced flavin cofactor is involved in the abstraction of the hydrogen in the C(6) pro-R position of the substrate 5-enolpyruvylshikimate 3-phosphate. A similar mechanism, abstraction of a proton from the α-CH2 (pro-R) of an amine by monoamine oxidase , seems to be emerging, replacing the previously proposed single electron transfer mechanism. Likewise, N(5) appears to be involved in the elimination of halide from β-Cl-alanine catalyzed by D-amino acid oxidase . At the active sites of these enzymes there is no amino acid functional group that could participate in the required base catalysis. The data presented in this paper thus sustain the notion that N(5) of the reduced flavin cofactor can play a direct role in the catalysis of some flavin-dependent enzymes, an involvement that has not yet been addressed in sufficient detail. Our studies towards a characterization of the hydrogen exchange processes at this position therefore lay the foundation for a critical assessment of this role.
The synthesis and purification of [1,3,5,10-15 N4]-7-methy1-10- ribitylisoalloxazine-5'-phosphate was described previously [31, 32]. 15N-NMR measurements were performed at 15°C, if not otherwise stated, with a Bruker AM 500 NMR instrument equipped with an Aspect 3000 and a temperature control unit. NMR-spectra were recorded with 12 μs pulses (= 30° flip angle) and a relaxation delay of 2 sec. In a typical experiment 10 mm Wilmad precision NMR tubes contained 1.8 ml of a 3 to 8 mM solution of flavin (or otherwise indicated in the text) in 250 mM Tris/100 mM NaCl-buffer and 0.2 ml D2O for field frequency lock. Reduction of the flavin solution was achieved by flushing the sealed NMR tube with argon before a two to three fold excess of a concentrated dithionite solution was added with a syringe. The pH of the resulting reduced solution was measured after the NMR experiment with a pH meter from Radiometer (Copenhagen, Sweden) equipped with a glass electrode from Ingold (Frankfurt, Germany). All 15N chemical shifts are expressed relative to liquid ammonia at 25°C and are corrected for bulk volume susceptibility. Neat CH3 15NO3 (™(CH3NO3) - ™(NH3) = 381.9 ppm) was used as an external standard according to Witanowski et al . The proton exchange rate (1/τe) was calculated according to Gutowsky et al  in the case of a fast exchange reaction, and according to Grunwald et al  in the case of a slow exchange reaction.
- Ghisla S, Massey V: Mechanisms of Flavoprotein-catalyzed Reactions. Eur J Biochem. 1989, 181: 1-17. 10.1111/j.1432-1033.1989.tb14688.x.View ArticlePubMedGoogle Scholar
- Bornemann S: Flavoenzymes that catalyse reactions with no net redox change. Nat Prod Rep. 2002, 19 (6): 761-772. 10.1039/b108916c.View ArticlePubMedGoogle Scholar
- Venkataram UV, Bruice TC: On the Mechanism of Flavin-Catalyzed Dehydrogenation a,β to an Acyl Function. The Mechanism of 1,5-Dihydroflavin Reduction of Maleimides. J Am Chem Soc. 1984, 106: 5703-5709. 10.1021/ja00331a047.View ArticleGoogle Scholar
- Urban P, Lederer F: Intermolecular hydrogen transfer catalyzed by a flavodehydrogenase, bakers' yeast flavocytochrome b2. J Biol Chem. 1985, 260 (20): 11115-11122.PubMedGoogle Scholar
- Lederer F: Extreme pKa displacements at the active sites of FMN-dependent alpha-hydroxy acid-oxidizing enzymes. Protein Sci. 1992, 1 (4): 540-548.PubMed CentralView ArticlePubMedGoogle Scholar
- Müller F: Nuclear magnetic resonance studies on flavoproteins. Chemistry and Biochemistry of Flavoproteins. Edited by: Müller F. 1992, Boca Raton, Florida: CRC Press, III: 558-595.Google Scholar
- Umhau S, Pollegioni L, Molla G, Diederichs K, Welte W, Pilone MS, Ghisla S: The x-ray structure of D-amino acid oxidase at very high resolution identifies the chemical mechanism of flavin-dependent substrate dehydrogenation. Proc Natl Acad Sci USA. 2000, 97 (23): 12463-12468. 10.1073/pnas.97.23.12463.PubMed CentralView ArticlePubMedGoogle Scholar
- Mattevi A, Vanoni MA, Curti B: Structure of D-amino acid oxidase: new insights from an old enzyme. Current Opinion in Structural Biology. 1997, 7 (6): 804-810. 10.1016/S0959-440X(97)80150-9.View ArticlePubMedGoogle Scholar
- Molla G, Pilone MS, Pollegioni L, Ghisla S: Studies on the elimination reaction of Rhodotorula gracilis D-amino acid oxidase with β-chloro-D-alanine. Flavins and Flavoproteins 2002: Proc 14th Int Symp, Cambridge, UK, 2002. Edited by: Chapman S, Perham R, Scrutton N. 2002, Agency for Scientific Publ., Berlin, 2002: 299-304.Google Scholar
- Ghisla S, Pollegioni L: D-Amino Acid Oxidase: Still New Lessons from a Seventy Year Old Enzyme. Flavins and Flavoproteins, Proc 15th Int Symp. 2005,Google Scholar
- Walsh CT, Krodel E, Massey V, Abeles RH: Studies on the elimination reaction of D-amino acid oxidase with α-amino-β-chlorobutyrate. Further evidence for abstraction of substrate α-hydrogen as a proton. J Biol Chem. 1973, 248 (6): 1946-1955.PubMedGoogle Scholar
- Maclean J, Ali S: The structure of chorismate synthase reveals a novel flavin binding site fundamental to a unique chemical reaction. Structure (Camb). 2003, 11 (12): 1499-1511. 10.1016/j.str.2003.11.005.View ArticleGoogle Scholar
- Lauhon CT, Bartlett PA: Substrate analogs as mechanistic probes for the bifunctional chorismate synthase from Neurospora crassa. Biochemistry. 1994, 33 (47): 14100-14108. 10.1021/bi00251a019.View ArticlePubMedGoogle Scholar
- Edmondson DE, Mattevi A, Binda C, Li M, Hubalek F: Structure and mechanism of monoamine oxidase. Curr Med Chem. 2004, 11 (15): 1983-1993.View ArticlePubMedGoogle Scholar
- Franken HD, Rüterjans H, Müller F: Nuclear-magnetic-resonance investigation of 15N- labeled flavins, free and bound to Megasphaera elsdenii apoflavodoxin. Eur J Biochem. 1984, 138 (3): 481-489. 10.1111/j.1432-1033.1984.tb07942.x.View ArticlePubMedGoogle Scholar
- Moonen CT, Vervoort J, Müller F: Reinvestigation of the structure of oxidized and reduced flavin: carbon-13 and nitrogen-15 nuclear magnetic resonance study. Biochemistry. 1984, 23 (21): 4859-4867. 10.1021/bi00316a007.View ArticlePubMedGoogle Scholar
- Rüterjans H, Kaun E, Hull WE, Limbach HH: Evidence for tautomerism in nucleic acid base pairs. 1 H NMR study of 15N labeled tRNA. Nucleic Acids Res. 1982, 10 (21): 7027-7039.PubMed CentralView ArticlePubMedGoogle Scholar
- Kaplan JI, Fraenkel G: NMR of chemically exchanging systems. 1980, New York: Academic pressGoogle Scholar
- Martin G, Martin ML, Gouesnard JP: 15N-NMR Spectroscopy. Edited by: Diel P, Fluck E, Korfeld R. 1981, Berlin: Springer Verlag, 18: 187-326.View ArticleGoogle Scholar
- Binsch G, Lambert JB, Roberts BW, Roberts JD: Nitrogen-15 magnetic resonance spectroscopy II. Coupling constants. J Am Chem Soc. 1964, 86: 5564-5570. 10.1021/ja01078a033.View ArticleGoogle Scholar
- Gunter H: NMR-Spectroscopy. 1983, Stuttgart: Thieme-VerlagGoogle Scholar
- Blomberg F, Maurer W, Rüterjans H: 15N nuclear magnetic resonance investigations on amino acids. Proc Natl Acad Sci USA. 1976, 73 (5): 1409-1413.PubMed CentralView ArticlePubMedGoogle Scholar
- Dixon M, Webb EC: Enzymes. 1979, New York: Longman/Academic PressGoogle Scholar
- Ghisla S, Macheroux P, Sanner C, Rüterjans H, Müller F: Ionization properties of reduced 1,5-dihydroflavin, rates of N(5)-H exchange with solvent. Flavins and Flavoproteins. Edited by: Curti B, Ronchi S, Zanetti G. 1991, Berlin, New York: Walter de Gruyter, 27-32.Google Scholar
- Adamson AW: A Textbook of Physical Chemistry. 1973, New York: Academic PressGoogle Scholar
- Ghisla S, Hartmann U, Hemmerich P, Müller F: Synthese, Struktur und Reaktivität von Dihydroflavinen. Liebig's Ann Chem. 1973, 1973: 1388-1415.View ArticleGoogle Scholar
- Yalloway GN, Mayhew SG, Boren SJ, Vervoort J: Effects of pH on the 13C and 15N NMR spectra of the hydroquinone of Desulfovibrio vulgaris flavodoxin and its G61A mutant. Flavins and Flavoproteins. Edited by: Ghisla S, Macheroux P, Sund H. 1999, Berlin: Agency for Scientific Publications, 187-190.Google Scholar
- Vervoort J, Müller F, LeGall J, Bacher A, Sedlmaier H: Carbon-13 and nitrogen-15 nuclear-magnetic-resonance investigation on Desulfovibrio vulgaris flavodoxin. Eur J Biochem. 1985, 151 (1): 49-57. 10.1111/j.1432-1033.1985.tb09067.x.View ArticlePubMedGoogle Scholar
- Eisenreich W, Kemter K, Bacher A, Mulrooney SB, Williams CH, Müller F: 13C-, 15N- and 31P-NMR studies of oxidized and reduced low molecular mass thioredoxin reductase and some mutant proteins. Eur J Biochem. 2004, 271 (8): 1437-1452. 10.1111/j.1432-1033.2004.04043.x.View ArticlePubMedGoogle Scholar
- Ludwig ML, Schopfer LM, Metzger AL, Partridge KA, Massey V: Structure and oxidation-reduction behavior of 1-deaza-FMN flavodoxins: modulation of redox potentials in flavodoxins. Biochemistry. 1990, 29 (45): 10364-10375. 10.1021/bi00497a011.View ArticlePubMedGoogle Scholar
- van Schagen CG, Müller F: A 13C nuclear-magnetic-resonance study on free flavins and Megasphaera elsdenii and Azotobacter vinelandii flavodoxin. 13C-enriched flavins as probes for the study of flavoprotein active sites. Eur J Biochem. 1981, 120 (1): 33-39. 10.1111/j.1432-1033.1981.tb05666.x.View ArticlePubMedGoogle Scholar
- Müller F, Vervoort J, Lee J, Horowitz M, Carreira LA: Coherent anti-Stokes Raman spectra of isoalloxazines. J Ramn Spectrosc. 1983, 14: 106-117. 10.1002/jrs.1250140211.View ArticleGoogle Scholar
- Witanowski M, Stefaniak L, Webb GA: Annu Rep NMR Spectros. 1981, 11B: 1-148.Google Scholar
- Gutowsky HS, McCall DW, Slichter CP: Nuclear magnetic resonance multiplets in liquids. J Chem Phys. 1953, 21 (2): 279-292. 10.1063/1.1698874.View ArticleGoogle Scholar
- Grunwald E, Loewenstein A, Meiboom S: Rates and mechanisms of protolysis of methylammonium ion in aqueous solution studied by proton magnetic resonance. J Chem Phys. 1957, 27 (3): 630-640. 10.1063/1.1743802.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.