Enzyme Activation with a Synthetic Catalytic Co-enzyme Intermediate: Nucleotide Methylation by Flavoenzymes

Article abstract of DOI:10.1002/anie.201706219

To facilitate production of functional enzymes and to study their mechanisms, especially in the complex cases of coenzyme-dependent systems, activation of an inactive apoenzyme preparation with a catalytically competent coenzyme intermediate is an attractive strategy. This is illustrated with the simple chemical synthesis of a flavin-methylene iminium compound previously proposed as a key intermediate in the catalytic cycle of several important flavoenzymes involved in nucleic acid metabolism. Reconstitution of both flavin-dependent RNA methyltransferase and thymidylate synthase apoproteins with this synthetic compound led to active enzymes for the C5-uracil methylation within their respective transfer RNA and dUMP substrate. This strategy is expected to be of general application in enzymology.

Full text of DOI:10.1002/anie.201706219

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Enzyme activation with a synthetic catalytic coenzyme
intermediate: Nucleotide methylation by new flavoenzymes
Authors: Charles Bou-Nader, Cornu David, Vincent Guerineau, Marc
Fontecave, and Djemel Hamdane
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201706219
Angew. Chem. 10.1002/ange.201706219
Link to VoR: http://dx.doi.org/10.1002/anie.201706219
http://dx.doi.org/10.1002/ange.201706219

10.1002/anie.201706219
Angewandte Chemie International Edition
Enzyme activation with a synthetic catalytic coenzyme intermediate:
Nucleotides methylation by new flavoenzymes
Charles Bou-Nader1, David Cornu 2, Vincent Guerineau3, Thibault Fogeron1, Marc Fontecave*1 and
Djemel Hamdane*1
1Laboratoire de Chimie des Processus Biologiques, CNRS-UMR 8229, Collège De France, Université
Pierre et Marie Curie, 11 place Marcelin Berthelot, 75231 Paris Cedex 05, France
2Institut de Biologie Intégrative de la Cellule, CNRS, CEA, Université Paris-Saclay, Gif-sur-Yvette,
France.
3Institut de Chimie des Substances Naturelles, Centre de Recherche de Gif, CNRS, 1 avenue de la
Terrasse, 91198 Gif-sur-Yvette, France
Corresponding author
Correspondence to: djemel.hamdane@college-de-france.fr
or marc.fontecave@college-de-france.fr
Abstract: To facilitate production of functional enzymes and to study their mechanisms,
especially in the complex cases of coenzyme-dependent systems, activation of an inactive
apoenzyme preparation with a catalytically competent coenzyme intermediate is an attractive
strategy. This is illustrated with the simple chemical synthesis of a flavin-methylene iminium
compound previously proposed as a key intermediate in the catalytic cycle of several important
flavoenzymes involved in nucleic acids metabolism. Reconstitution of both flavin-dependent
RNA methyltransferase and thymidylate synthase apoproteins with this synthetic compound led
to active enzymes for the C5-uracil methylation within their respective transfer RNA and dUMP
substrate. This strategy is expected to be of general application in enzymology.
Structural and functional characterization as well as industrial exploitation of enzymes often
require efficient maturation when they depend on complex metallic or bioorganic coenzymes. To
bypass costly maturation steps, it can be advantageous to produce an apoprotein version and to
subsequently introduce the coenzyme, produced synthetically, for final activation. An exquisite
example is the synthetic maturation of hydrogenases, which are key enzymes for the development of
new energy storage technologies[1]. In that case not only the reaction led to full activation of the
enzyme but also definitively established the structure of the natural cofactor, under controversy[2].
Here, we develop further this concept in which the target apoenzyme is activated by reaction with a
synthetic catalytically competent coenzyme intermediate, involved in the catalytic cycle. While
challenging since intermediates are highly reactive in general, this approach is particularly appealing
when addressing mechanistic issues. Few previous reports have shown the great potential of that
approach [3]. We applied this concept on a new class of flavoenzymes that use a flavin as a coenzyme
in nucleotide C5-uracil-methylation (Figure 1). Bacterial thymidylate synthase ThyX converts dUMP
to dTMP, then incorporated into DNA [4], while the transfer RNA and ribosomal RNA
methyltransferases, TrmFO and RlmFO, catalyze the conversion of uridine to 5-methyl-uridine (m5U)
within their RNA substrate [5]. They all use 5,10-methylenetetrahydrofolate (CH2THF) derivative as
the methylene donor and a non-covalently bound flavin adenine dinucleotide hydroquinone (FADH-,
compound 2) issued from the reduction of FAD (compound 1) by NAD(P)H. Recently, 2 has been
proposed to mediate methylene transfer from the folate carrier to the nucleotide acceptor [6]. Via an
addition-elimination reaction (Figure 2A), 2 is converted into a transient reactive flavin-methylene
iminium (compound 4), which would act as an unprecedented nucleotide methylating agent (Figure
S1), expanding further the chemical repertoire of this versatile coenzyme[6c, 7]. Although the existence
of such an intermediate has never been proved directly, its involvement in the reaction has been
supported by several consistent data. We herein synthesized 4 separately for the first time and used it
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
to activate Bacillus subtilis TrmFO and Thermotoga maritima ThyX apoproteins for tRNA and dUMP
methylation, respectively, thus unambiguously establishing this flavin iminium as the genuine
methylating transfer agent in nucleotide metabolism.
To do so, we first reacted 2 with a 100-fold excess of chloro-dimethyl-sulfide (ClCH2SCH3)
under anaerobic conditions. We anticipated the methylenic carbon of ClCH2SCH3 to be electrophilic
enough for nucleophilic attack by N5 from 2 (Figure 2B) and the intermediate 3’ to decompose into
methylthiol and 4. Addition of dichloromethane to the reaction mixture led to separation of a
yellowish aqueous phase (AQ) most likely containing the flavin species from the organic phase
containing the unreacted ClCH2SCH3. Absence of the latter in AQ was confirmed by ESI-MS. The
absorption spectrum of AQ recorded under anaerobic conditions exhibited an unresolved broad
absorbance band with a maximum at around 350-354 nm (gray line), different from that of 1 (orange)
and 2 (dark green) (Figure 3A) but reminiscent of N5-alkylated flavin species such as that found in
TrmFO [6a]. This indicates full conversion of 2 into an alkylated flavin during reaction with
ClCH2SCH3.
AQ was then analyzed by positive mode ESI-MS and the resulting MS and MS/MS spectra are
reported in Figure 3C-F. Control 1 had a peak at m/z 786.25, which fragmented into m/z 348.1 and m/z
439.13 assigned to the adenosine monophosphate and the ribityl monophosphate isoalloxazine,
respectively (Figure 3C & 3D). In contrast, AQ showed a m/z 800.27 corresponding to 4
([FADH(N5)=CH2]+), as well as Na+ adducts (m/z 822.276 and m/z 844.24) (Figure 3E). The
presence of peaks of 1 in AQ suggested some degradation of 4 within the MS measurement chamber.
Fragmentation of 4 gave m/z 453 assigned to [FMNH(N5)=CH2]+ species (Figure 3F), which could be
further fragmented into isoalloxazine bearing an N5 methylene (m/z 257.14) (Figure S2). We could not
detect a peak corresponding to ClCH2SCH3 further confirming its absence in AQ. Thus, these data
established 4 as the only product of the alkylation reaction while no evidence for the presence of
compound 3’ could be obtained. To determine whether 3’ to 4 conversion occurred during mass
measurement or in solution, AQ was incubated with NaBH4 directly in the glove box and the resulting
mixture was subsequently analyzed by MS. This procedure has been applied to trap an iminium flavin
intermediate in UDP-galactopyranose mutase[8]. In addition to the peaks of 1 and its respective Na+
adduct species, new peaks at m/z 802.3, 824.28 and 846.26 were observed and assigned to
[FADH(N5)CH3]H+ and the corresponding mono and di-sodium adducts (Figure S3A). MS/MS also
confirmed that the -CH3 group was carried by the isoalloxazine and not by the adenosine group of
FAD (Figure S3B). The 3’ to 4 reaction is likely favored by sulfur protonation of 3’, as recently
supported by DFT calculations [6b]. Interestingly, a precedent for an unstable alkylated flavin leading to
the formation of free flavin iminium was reported in the literature. The N5 of reduced flavin was
shown to nucleophilically attack carbonyl groups of small molecules to generate a carbinolamine,
which was subsequently converted into an imine [9].
A distinctive feature between 2 and its N5-alkyl adducts is their different intrinsic stability in
the presence of air. It is known that 2 oxidizes within seconds upon reaction with oxygen[10] whereas
alkylated flavins are generally quite resistant [6b, 11]. We thus investigated the kinetics of oxidation of
the synthetically modified flavin by monitoring the absorbance at 446 nm. Upon exposure to air, the
absorbance increased slowly consistent with a slow 4 to 1 conversion (Figure S4). The kinetics were
fitted to a double exponential function with a fast phase representing only 4 % of the total variation of
the signal and a slow phase (~ 96 %) proceeding with an observed rate constant of 0.003 ± 4.10-4 min-1
and a half-life of ~ 3.75 hours. In contrast, oxidation of 2 occurred in less than 10 sec. The high
stability of the synthetically modified flavin to air provides further support for N5-alkylation of 2 by
ClCH2SCH3 as anticipated. Altogether, the data show that 4 can be prepared quantitatively from
reaction of 2 with ClCH2SCH3 and it displays a remarkable stability in solution.
Next, we examined whether such a synthetic 4 could bind and activate the apoTrmFO and
apoThyX proteins (Figure 2C). The latter was obtained according to the procedure described by
Mishanina et al, which relies on a classical chemical-induced denaturation of the holoprotein followed
by a renaturation process leading to an apoThyX as supported from its absorption spectrum (Figure
3B, blue and green for holo- and apoThyX, respectively). However, preparation of apoTrmFO form
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
via classical methods such as dialysis under high concentration of potassium iodide or chaotrope-
induced denaturation/renaturation was not possible due to protein instability[12]. To circumvent these
difficulties, Y346A TrmFO variant was chosen as a receptor since it was previously purified without
flavin and was easily reloaded with free FAD[12b]. Accordingly, the LC-MS and absorption spectrum
of the mutant confirmed the absence of flavin (Figure S5 and S6). Incubation of apoY346A TrmFO
(from now referred as TrmFO) or apoThyX with the synthetic 4 followed by a purification step on a
gel filtration column was performed in the glove box. The resulting proteins exhibited a very light
yellowish color suggesting that 4 was likely bound to the enzymes. UV-visible spectra of both
preparations (Figure 3A, purple curve for TrmFO and Figure 3B, violet curve for ThyX) confirmed the
presence of synthetic 4 retained within the respective enzymes.
To test whether TrmFO-4 complex was able to achieve methylation of tRNA, the protein was
incubated with a 3-molar excess of a tRNA transcript (tRNAPhe) at 37°C during 60 min. After the
reaction, the tRNA was phenol-extracted and digested with RNAseT1 and the resulting nucleotidic
fragments were analyzed on MALDI-MS and compared to those issued from a free tRNAPhe control
(Figure S7). As shown in figure 4E, a new fragment at m/z 1295.22 appeared during the reaction,
corresponding to the addition of a methyl group on the U54UCG57 nucleotide sequence. This fragment
was subjected to a MS fragmentation and the resulting MS/MS spectrum (Figure S8) showed that a1,
a3 and c3 bore a + 14 Da mass increase, thus providing evidence that the methyl group was grafted on
U54. When control activity assays were carried out with 4 alone, apoTrmFO alone or with apoTrmFO
reconstituted with either 1 or 2, the m/z 1295.22 fragment could not be detected. We also performed
tRNA methylation assays using apoTrmFO reconstituted with D-labelled 4 synthesized from
ClCD2SCD3. As shown in figure 4B, a new fragment m/z 1297.13 appeared in the MS spectrum of the
resulting tRNA fragments that we assigned to [7,7’-D]m5U54UCG57. Hence, the 7-methylenic group
incorporated into m5U54 tRNAPhe by TrmFO-4 derives from ClCH2SCH3 exclusively. Moreover, an
accurate LC-ESI-MS/MS analysis of the proteolytic peptide mixture issued from trypsinolysis of
TrmFO-4 did not reveal the presence of additional CH2 or CH3 group on active site residues located in
the immediate vicinity of FAD (Figure S9). Therefore, the tRNA methylation reaction can be assigned
exclusively to the activity of the TrmFO-4 complex, which retains natural substrate specificity (i.e
methylation of U54).
We also tested whether the reconstituted ThyX-4 complex was able to synthesize dTMP from
dUMP by itself. As shown in the LC profiles (Figure 4C), the thymidylate synthase activity was only
observed in the sample wherein dUMP has been mixed with ThyX-4, likewise indicating that this
latter complex is responsible for dTMP synthesis.
The first outcome of this study is the direct demonstration that a flavin-methylene iminium is a
genuine methylating agent and the active species in enzymes of nucleotides metabolism, thus
supporting the proposed mechanisms shown in Figure S1. Furthermore and more generally, we
illustrate and validate experimentally the strategy that complex enzymes can be readily activated with
a synthetic active cofactor intermediate. In the specific case reported here, the methodology is
facilitated because the target intermediate, 4, proved synthetically accessible, stable alone in solution
but in contrast very reactive within the apoenzyme to be activated. Such a situation is encountered in
many enzymes[13] implying a much more general application of that approach.
Experimental Section
Methods are in Supporting Information.
Acknowledgements
This work was supported by the Centre National de la Recherche Scientifique, the university Pierre et
Marie Curie as well as the French State Program ‘Investissements d’Avenir’ (ANR-15-CE11-0004-01
MIGRATE and Labex program DYNAMO, ANR-11-LABX-0011].
Keywords: Flavoenzyme mechanism, intermediate, synthetic chemistry, artificial enzyme, nucleotide
methylation
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
References
[1]
[2]
V. Artero, G. Berggren, M. Atta, G. Caserta, S. Roy, L. Pecqueur, M. Fontecave, Accounts of
chemical research 2015, 48, 2380-2387.
G. Berggren, A. Adamska, C. Lambertz, T. R. Simmons, J. Esselborn, M. Atta, S. Gambarelli,
J. M. Mouesca, E. Reijerse, W. Lubitz, T. Happe, V. Artero, M. Fontecave, Nature 2013, 499,
66-69.
[3]
[4]
S. Zhang, M. Liu, Y. Yan, Z. Zhang, F. Jordan, J Biol Chem 2004, 279, 54312-54318.
H. Myllykallio, G. Lipowski, D. Leduc, J. Filee, P. Forterre, U. Liebl, Science 2002, 297, 105-
107.
[5]
aJ. Urbonavicius, S. Skouloubris, H. Myllykallio, H. Grosjean, Nucleic Acids Res 2005, 33,
3955-3964; bC. Lartigue, A. Lebaudy, A. Blanchard, B. El Yacoubi, S. Rose, H. Grosjean, S.
Douthwaite, Nucleic Acids Res 2014, 42, 8073-8082; cH. Nishimasu, R. Ishitani, K.
Yamashita, C. Iwashita, A. Hirata, H. Hori, O. Nureki, Proc Natl Acad Sci U S A 2009, 106,
8180-8185.
[6]
aD. Hamdane, M. Argentini, D. Cornu, B. Golinelli-Pimpaneau, M. Fontecave, J Am Chem
Soc 2012, 134, 19739-19745; bD. Hamdane, E. Bruch, S. Un, M. Field, M. Fontecave,
Biochemistry 2013, 52, 8949-8956; cD. Hamdane, H. Grosjean, M. Fontecave, J Mol Biol
2016, 428, 4867-4881; dT. V. Mishanina, L. Yu, K. Karunaratne, D. Mondal, J. M. Corcoran,
M. A. Choi, A. Kohen, Science 2016, 351, 507-510.
[7]
[8]
aD. Leys, N. S. Scrutton, Current opinion in structural biology 2016, 41, 19-26; bV. Piano, B.
A. Palfey, A. Mattevi, Trends Biochem Sci 2017.
M. Soltero-Higgin, E. E. Carlson, T. D. Gruber, L. L. Kiessling, Nature structural &
molecular biology 2004, 11, 539-543.
[9]
B. T. C. Williams Robert F., J Am Chem Soc 1976, 98, 7752-7768.
[10]
[11]
V. Massey, J Biol Chem 1994, 269, 22459-22462.
aS. C. Rothman, J. B. Johnston, S. Lee, J. R. Walker, C. D. Poulter, J Am Chem Soc 2008,
130, 4906-4913; bG. Gadda, R. D. Edmondson, D. H. Russell, P. F. Fitzpatrick, J Biol Chem
1997, 272, 5563-5570; cS. Ghisla, V. Massey, Y. S. Choong, Journal of Biological Chemistry
1979, 254, 662-669.
[12]
[13]
aD. Hamdane, C. Velours, D. Cornu, M. Nicaise, M. Lombard, M. Fontecave, Physical
chemistry chemical physics : PCCP 2016, 18, 20410-20421; bD. Hamdane, C. Bou-Nader, D.
Cornu, G. Hui-Bon-Hoa, M. Fontecave, Biochemistry 2015, 54, 4354-4364.
P. A. a. H. Frey, Adrian D., Enzymatic reaction mechanisms. New York ; Oxford : Oxford
University Press, c2007, 2007.
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
Figure 1. Flavin and folate-dependent C5-uracil methylation reaction catalyzed by ThyX,
TrmFO and RlmFO. The methylene moiety being transferred from the folate carrier (5,10
methylenetetrahydrofolate, CH2=THF), to the substrate, is in red while the hydride from the reduced
flavin (FADH-) used to reduce CH2 into the final exocyclic CH3 group is in blue.
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
Figure 2. Strategy for the activation of the inactive tRNA methyltransferase and thymidylate
synthase apoproteins by a synthetic catalytic intermediate (A) Postulated mechanism for formation
of the flavin iminium methylating agent FAD(N5)=CH2 (4) common to ThyX, TrmFO and RlmFO[6a,
6c, 6d]. R group on the isoalloxazine represents the adenosine-5’-pyrophosphate-ribityl. (B) Proposed
chemical synthetic pathway leading to 4 from the reaction of free FADH-, obtained by the reduction of
FAD with sodium dithionite, and chloromethyl-methylsulfide. (C) Biochemical reconstitution strategy
for testing the methylating activity of synthetic 4 using the flavin-dependent tRNA m5U54-
methyltransferase Bacillus subtilis Y346A apoTrmFO variant (colored in blue) and the wild type
thymidylate synthase ThyX from Thermotoga maritima (colored in violet) as the protein carrier
models. ThyX is a homotetrameric enzyme but, here, for the sake of clarity, we represented only one
protomer. A zoom view on the crystal structure of holoTrmFO active site is shown in the blue box.
FAD is in yellow, Y346 in red and C53 in purple. The Y346 phenol group stacks against the FAD
isoalloxazine moiety whilst its side chain OH group hydrogen bonds with the NH amide of C53. Note
that the residues are numbered according to the Bacillus subtilis TrmFO sequence.
C
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
Figure 3. Characterization of synthetic 4 and reconstitution with TrmFO and ThyX apoproteins.
(A) Light absorption spectrum, recorded under anaerobic conditions, of 1 (orange), 2 (dark yellow),
the aqueous phase after dichloromethane extraction from the reaction of 2 with ClCH2SCH3 (gray),
Y346A TrmFO reconstituted with the alkylated flavin contained in the aqueous phase (purple). (B)
Light absorption spectrum, recorded under anaerobic conditions, of holoThyX (blue), apoThyX
(green) and ApoThyX reconstituted with the alkylated flavin contained in the aqueous phase (violet).
(C) and (D) are the ESI MS and MS/MS spectra of commercial oxidized FAD, respectively. Peaks at
m/z 348.1, and 439.13 correspond to adenosine monophosphate, and ribityl monophosphate
isoalloxazine, respectively. (E) and (F) are the ESI MS and MS/MS spectra of the aqueous phase
containing the alkylated flavin derivative, respectively. Peak at m/z 453.18 corresponds to the
methylene-FAD ribityl monophosphate isoalloxazine. The weakly intense peak at m/z 257.11 on the
MS/MS corresponds to the methylene-isoalloxazine.
E
F
C
D
A
B
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
Figure 4. Functional activation of apo-TrmFO and apo-ThyX. (A) MALDI-MS analysis of
tRNAPhe before and after reaction with the TrmFO-4 complex. MALDI mass spectra of tRNA control,
tRNA + synthetic 4 alone, tRNA + apoTrmFO alone, tRNA + apoTrmFO-4 and tRNA + apo-TrmFO-
2. The inserts correspond to the spectral region around the UUCG fragment containing the U54
substrate of TrmFO showing the peaks for the unmethylated (m/z 1281.4) and the corresponding
monomethylated (m/z 1295.22) ions. The secondary structure together with the sequence of the
tRNAPhe is also reported. (B) Activation of apo-TrmFO with synthetic D-labeled 4 synthesized from
the deuterated ClCD2SCD3 alkylating agent. The figure corresponds to the spectral region around the
U54UCG57 fragment showing the peaks for the unmethylated (m/z 1281.4) and the corresponding
[7,7’-D] m5U54UCG57 monomethylated (m/z 1297.22) ions. (C) Activation of apoThyX by synthetic 4
for conversion of dUMP into dTMP. HPLC profiles of: a) dUMP + 4, b) dUMP + ThyX reconstituted
with 4, c) dUMP + ThyX reconstituted with 2, d) dUMP standard and e) dTMP standard. The retention
times for dUMP and dTMP were 2.05 and 3.07 min, respectively.
A
tRNAPhe
C
B
This article is protected by copyright. All rights reserved.

10.1002/anie.201706219
Angewandte Chemie International Edition
Entry for the table of Contents (please choose one layout)
Charles Bou-Nader, David Cornu, Vincent
Guerineau, Thibault Fogeron, Marc Fontecave
and Djemel Hamdane
Page N0.- Page N0.
Enzyme activation with a synthetic catalytic
coenzyme intermediate: Nucleotides
methylation by new flavoenzymes
Artificial enzyme engineering for uncovering cofactor-dependent complex mechanistic issues:
Chemical synthetized N5-flavin iminium catalytic intermediate can be used to artificially activate the
apoprotein version of ThyX and TrmFO for dUMP and tRNA methylation, respectively. This procedure
bypasses the natural biosynthetic pathway and establishes this flavin iminium species as the bona fide
nucleotide methylating agent.
This article is protected by copyright. All rights reserved.