Cfr and RlmN contain a single [4Fe-4S] cluster, which directs two distinct reactivities for s -adenosylmethionine: Methyl transfer by SN2 displacement and radical generation

Article abstract of DOI:10.1021/ja207327v

The radical SAM (RS) proteins RlmN and Cfr catalyze methylation of carbons 2 and 8, respectively, of adenosine 2503 in 23S rRNA. Both reactions are similar in scope, entailing the synthesis of a methyl group partially derived from S-adenosylmethionine (SAM) onto electrophilic sp²-hybridized carbon atoms via the intermediacy of a protein S-methylcysteinyl (mCys) residue. Both proteins contain five conserved Cys residues, each required for turnover. Three cysteines lie in a canonical RS CxxxCxxC motif and coordinate a [4Fe-4S]-cluster cofactor; the remaining two are at opposite ends of the polypeptide. Here we show that each protein contains only the one "radical SAM" [4Fe-4S] cluster and the two remaining conserved cysteines do not coordinate additional iron-containing species. In addition, we show that, while wild-type RlmN bears the C355 mCys residue in its as-isolated state, RlmN that is either engineered to lack the [4Fe-4S] cluster by substitution of the coordinating cysteines or isolated from Escherichia coli cultured under iron-limiting conditions does not bear a C355 mCys residue. Reconstitution of the [4Fe-4S] cluster on wild-type apo RlmN followed by addition of SAM results in rapid production of S-adenosylhomocysteine (SAH) and the mCys residue, while treatment of apo RlmN with SAM affords no observable reaction. These results indicate that in Cfr and RlmN, SAM bound to the unique iron of the [4Fe-4S] cluster displays two reactivities. It serves to methylate C355 of RlmN (C338 of Cfr), or to generate the 5′-deoxyadenosyl 5′-radical, required for substrate-dependent methyl synthase activity.

Full text of DOI:10.1021/ja207327v

COMMUNICATION
pubs.acs.org/JACS
Cfr and RlmN Contain a Single [4Fe-4S] Cluster, which Directs Two
Distinct Reactivities for S-Adenosylmethionine: Methyl Transfer by
SN2 Displacement and Radical Generation
Tyler L. Grove,† Matthew I. Radle,† Carsten Krebs,*,†,‡ and Squire J. Booker*,†,‡
†Department of Chemistry and ‡Department of Biochemistry and Molecular Biology, The Pennsylvania State University,
University Park, Pennsylvania, 16802, United States
S Supporting Information
b
Scheme 1. Reactions Catalyzed by RlmN and Cfr in Vivo
ABSTRACT: The radical SAM (RS) proteins RlmN and
Cfr catalyze methylation of carbons 2 and 8, respectively, of
adenosine 2503 in 23S rRNA. Both reactions are similar in
scope, entailing the synthesis of a methyl group partially
derived from S-adenosylmethionine (SAM) onto electro-
philic sp2-hybridized carbon atoms via the intermediacy of a
protein S-methylcysteinyl (mCys) residue. Both proteins
contain five conserved Cys residues, each required for turnover.
Three cysteines lie in a canonical RS CxxxCxxC motif and
coordinate a [4Fe-4S]-cluster cofactor; the remaining two are
at opposite ends of the polypeptide. Here we show that each
protein contains only the one “radical SAM” [4Fe-4S]
cluster and the two remaining conserved cysteines do not
coordinate additional iron-containing species. In addition,
we show that, while wild-type RlmN bears the C355 mCys
residue in its as-isolated state, RlmN that is either engi-
neered to lack the [4Fe-4S] cluster by substitution of the
coordinating cysteines or isolated from Escherichia coli cultured
under iron-limiting conditions does not bear a C355 mCys
residue. Reconstitution of the [4Fe-4S] cluster on wild-type
apo RlmN followed by addition of SAM results in rapid
production of S-adenosylhomocysteine (SAH) and the mCys
residue, while treatment of apo RlmN with SAM affords no
observable reaction. These results indicate that in Cfr and RlmN,
SAM bound to the unique iron of the [4Fe-4S] cluster displays
two reactivities. It serves to methylate C355 of RlmN (C338 of
Cfr), or to generate the 50-deoxyadenosyl 50-radical, required for
substrate-dependent methyl synthase activity.
Escherichia coli (Ec) RlmN and Staphylococcus aureus (Sa) Cfr
share 33% sequence identity and modify the same nucleotide in
23S rRNA, adenosine 2503 (A2503). Although RlmN and Cfr
act preferentially on naked rRNA,9 A2503 resides ultimately in
the peptidyltransferase center of the 50S subunit of the bacterial
ribosome.10ꢀ13 Methylation of C2 by RlmN is found throughout
eubacteria (Scheme 1).14 Although this activity is nonessential,
Ec mutants that lack it lose to wild-type (wt) Ec in cogrowth
competition experiments.15 Cfr, which is evolutionarily related to
RlmN, also can catalyze methylation of C2 of A2503; however,
C8 is its preferred target (Scheme 1).16 Methylation of C8
confers bacterial resistance to multiple classes of antibiotics,
including phenicols, lincosamides, oxazolidinones, pleuromuti-
lins, and streptogramin A, as well as the macrolides josamycin
and spiramycin.17
Cfr and RlmN are members of the radical SAM (RS) super-
family, enzymes that cleave SAM reductively to a 50-deoxyade-
nosyl 50-radical (50-dA•).18ꢀ21 The common feature of RS enzymes
is the use of the 50-dA• to abstract key substrate hydrogen atoms.
Cleavage of SAM requires an electron, which is supplied by a
reduced [4Fe-4S]+ cluster. The Cys residues that coordinate this
essential [4Fe-4S] cluster typically reside in a CxxxCxxC motif,21
although exceptions have been reported.22ꢀ24
he addition of a methyl group to an acceptor molecule is among
Our recent mechanistic studies of RlmN and Cfr have unraveled
an unprecedented strategy for SAM-dependent methylation of
an sp2-hybridized carbon atom.7 When RlmN or Cfr was incu-
bated with S-adenosyl-L-[methyl-d3]-methionine (d3-SAM) or
unlabeled SAM under single-turnover conditions, the isotopic
composition of the methyl group incorporated at C2 (RlmN) or
C8 (Cfr) of the target adenosine nucleotide did not always reflect
that of the methyl donor in the reaction, but rather the isotopic
composition of L-methionine in the growth media of the Ec used
for overproducing the protein. Consistent with this observation,
analysis of RlmN by high-resolution mass spectrometry7 and
X-ray crystallography25 showed the as-isolated (AI) protein to
T
the most critical of cellular reactions. Although methyl groups
can emanate from several complex cofactors (e.g., methylcoba-
lamin and methylene- and methyltetrahydrofolate), the vast majority
derive from the more understated molecule, S-adenosyl-L-
methionine (SAM).1ꢀ3 Seminal model and enzymatic studies
on SAM-dependent methyl-transfer reactions argue for a polar
process, in which an appropriate nucleophile attacks the methyl
group of SAM with concomitant elimination of S-adenosyl-L-
homocysteine (SAH).4ꢀ6 Until recently, a direct SN2 displace-
ment was the only means by which methyl groups derived from
SAM were thought to be appended onto acceptor atoms.7,8
However, RlmN and Cfr catalyze methyl transfer as well as
methyl synthesis via radical-dependent methylene transfer to
electrophilic carbon atoms.7,8
Received: August 4, 2011
Published: September 14, 2011
r
2011 American Chemical Society
19586
dx.doi.org/10.1021/ja207327v J. Am. Chem. Soc. 2011, 133, 19586–19589
|

Journal of the American Chemical Society
COMMUNICATION
bear a methylcysteinyl (mCys) residue at C355. Moreover,
methylation always involved incorporation of one hydrogen atom
from the mCys residue into 50-dA, indicating that the 50-dA•
activates the methyl group for radicaladdition rather than abstracts
a hydrogen atom from the nucleotide to be modified. Cleavage of
an ensuingmethylene-bridgedproteinꢀnucleicacidcross-link was
proposed to be catalyzed by disulfide-bond formation—with
participation of a second absolutely conserved Cys residue
(C118 in RlmN; C105 in Cfr), as shown in Scheme S1.7
A related study by Yan et al. also found that the 50-dA• radical
does not abstract a hydrogen atom from the substrate nucleotide.
In that study, however, the authors were unaware of the initial
transfer of a methyl group from SAM to C355 of RlmN (C338 of
Cfr), and therefore proposed that the 50-dA• abstracts a hydro-
gen atom from the methyl group of a second, simultaneously
bound, SAM molecule, activating it for radical addition to the
target nucleotide.8
Figure 1. 4.2-K/53-mT M€ossbauer spectra of RlmN and Cfr (vertical
bars) and quadrupole doublet simulations with parameters quoted in the
text (solid lines).
of the total intensity. Together with the ratio of 4.0 Fe per
polypeptide, determined by quantitative analyses of acid-labile iron
and sulfide (Table S2), the M€ossbauer spectrum reveals a stoichi-
ometry of 0.9 [4Fe-4S] clusters per AI RlmNwt. Reconstitution of
AI RlmNwt with additional Fe and sulfide (RCN RlmNwt) does not
result in an increase of Fe and sulfide associated with the protein
(4.1 Fe per polypeptide). Consistent with this observation, the 4.2-
K/53-mT M€ossbauer spectrum of RCN RlmNwt (Figure 1B) is
virtually identical to that of AI RlmNwt (95% of total intensity
attributable to the above quadrupole doublet) and reveals that
RCN RlmNwt harbors 1.0 [4Fe-4S] cluster, which is presumably
coordinated by the cysteines of the canonical CxxxCxxC RS motif:
C125, C129, and C132. Consistent with this hypothesis, the
variant, in which C125, C129, and C132 have been replaced with
non-coordinating Ala residues, harbors less than 0.02 [4Fe-4S]
clusters per protein (Figures S3ꢀS5).
The recent X-ray crystal structure of RlmN with SAM bound
(2.05 Å)25 is consistent with the mechanism proposed by Grove
et al.7 The structure shows C355 to reside in a flexible loop
containing residues 350ꢀ358, which are strictly conserved. This
loop is disordered in the absence of SAM but is visible in the
RlmN+SAM structure. Only one SAM-binding site was identi-
fied in the structure, wherein SAM is coordinated to the unique
iron of the [4Fe-4S] cluster via its α-amino and α-carboxy
groups, as is observed for other structurally characterized RS
proteins.26 Although it was speculated that the related protein,
Cfr, contains two Fe/S clusters,27 this stoichiometry is not
supported by the structure of RlmN. The structure shows
C355 to be S-methylated, and to be located ∼6 Å from the side
chain of C118 and C50 of SAM. This arrangement is consistent
with disulfide-bond formation between C355 and C118 to
resolve the proteinꢀnucleic acid covalent adduct as well as
abstraction of a hydrogen atom from the C355 mCys by the
50-dA• as proposed by Grove et al.7 The absence of any other
obvious SAM binding site led to the proposal that SAM
coordinated to the [4Fe-4S] must exhibit dual functionality; it
acts as both a methylating agent and as the source of the 50-dA•
intermediate.25 In this work we use M€ossbauer spectroscopy in
concert with quantitative analyses of iron and sulfide to show that
both RlmN and Cfr contain only one [4Fe-4S] cluster, which is
coordinated by cysteines in the canonical CxxxCxxC motif. This
finding implies that the two remaining conserved Cys residues do
not coordinate Fe/S species and are free to participate in other
modes of catalysis. In addition, we overproduce RlmN in Ec
under iron-limiting conditions and show that the AI protein is
not S-methylated at C355 and does not catalyze methylation of
C355 when incubated with SAM. Reconstitution of the protein
with iron and sulfide followed by addition of SAM results in rapid
methylation of C355, consistent with the proposal that SAM
coordinated to the [4Fe-4S] cluster has a dual function.
In addition to C125, C129, and C132, RlmN has two additional,
strictly conserved, cysteines that are absolutely required for com-
plete turnover: C118 and C355.28 To show that they do not ligate
Fe/S species, CysfAla variants of those residues (RlmNC118A
and RlmNC355A) were engineered and studied with a combina-
tion of biochemical and spectroscopic methods. RlmNC118A and
RlmNC355A contain 4.0ꢀ4.6 Fe per polypeptide in both their AI
and RCN forms (Table S2), and display spectroscopic features and
cluster stoichiometries that are virtually identical to those of RlmNwt,
supporting the presence of one [4Fe-4S] cluster per polypeptide
(Figures S6 and S7) ligated by C125, C129, and C132.
The 4.2-K/53-mT spectrum of AI Cfrwt (Figure 1C) is domi-
nated by a broad quadrupole doublet indicative of [4Fe-4S]2+
clusters (δ = 0.44 mm/s and ΔEQ = 1.10 mm/s), which accounts
for 86% of the total intensity. The broad and poorly resolved
absorption extending from -2 to 2 mm/s (15% of total Fe) is
assigned to unspecifically bound iron, because an identical
sample does not display EPR features associated with Fe/S
clusters with S = 1/2 ground state (Figure S5). Together with
the ratio of 4.5 Fe per polypeptide (Table S2), the M€ossbauer
spectrum reveals a stoichiometry of 1.0 [4Fe-4S] cluster per AI
Cfrwt. Reconstitution of AI Cfrwt with additional Fe and sulfide
followed by purification by anaerobic molecular-sieve chromato-
graphy (RCN Cfrwt) results in a slight decrease of Fe. The 4.2-K/
53-mT M€ossbauer spectrum of RCN Cfrwt (Figure 1D) can be
simulated with a quadrupole doublet with identical parameters,
and accounts for 98% of the total intensity. Thus, the results
suggest that RCN Cfrwt harbors 1.0 [4Fe-4S] cluster, which is
coordinated by C112, C116, and C119. We therefore con-
clude, that in contrast to radical SAM proteins that catalyze
To determine the stoichiometry and configuration of iron-
containing species associated with Cfr and RlmN, both proteins
were purified from Ec cultured in minimal media supplemented
with 57Fe, and then analyzed by M€ossbauer spectroscopy. The
UVꢀvisible spectra of both AI proteins are consistent with the
presence of [4Fe-4S] clusters (Figures S1 and S2), in agreement
with previous studies.9,27 The 4.2-K/53-mT spectrum of as-isolated
(AI) wild-type (wt) RlmN (RlmNwt) (Figure 1A) is dominated
by a broad quadrupole doublet with parameters typical of
[4Fe-4S]2+ clusters [isomer shift (δ) of 0.44 mm/s and quadrupole
splitting parameter (ΔEQ) of 1.14 mm/s], and accounts for 93%
19587
dx.doi.org/10.1021/ja207327v |J. Am. Chem. Soc. 2011, 133, 19586–19589

Journal of the American Chemical Society
COMMUNICATION
Figure 3. (A) UVꢀvis traces of apo RlmNwt (5 μM, solid black line)
and apo RlmNwtfRCN (12 μM, solid red line). (B) Methyl transfer
catalyzed by apo RlmNwt (150 μM) or apo RlmNwtfRCN (150 μM).
Production of SAH by apo RlmNwt (purple diamonds) or apo
RlmNwtfRCN (red circles) in the presence of 1.5 mM SAM; or apo
RlmNwtfRCN (150 μM) in the presence of 1.5 mM d3-SAM (black
triangles). Error bars indicate one standard deviation from the average of
three assays. (C) Q-Tof MS analysis of tryptic peptides from apo
RlmNwtfRCN after incubation in the presence of SAM (black trace) or
d3-SAM (red trace). Indicated m/z values correspond to the +2 charge state.
Figure 2. Extracted ion chromatograms (EIC) for the C355-containing
peptide from trypsin digests of RlmN. Black traces are EIC of m/z 931.0
[M + 2H], indicative of carbamidomethyl modification to C355. Red
traces are EIC of m/z 909.6 [M + 2H], indicative of a methyl
modification of C355. Traces correspond to (A) AI RlmNwt; (B) AI
RlmNC125A-C129A-C132A; (C) AI RlmNC118A; (D) apo RlmNwt, inclusion
bodies; (E) apo RlmNwt, soluble fraction; (F) apo RlmNwtfRCN after
addition of 1.5 mM SAM. Spectral pair intensities are normalized to the
most abundant EIC.
insoluble protein, apo RlmNwt inclusion bodies were also analyzed
by LCꢀMS; fragments obtained from trypsin digests of the inclusion
body fraction did not bear a C355 mCys modification (Figure 2D).
Reconstitution of apo RlmNwt followed by further purification
by size-exclusion chromatography afforded a protein containing
2.2 irons and 1.2 sulfides, respectively, per polypeptide (Table S2).
This stoichiometry implies that 50% of the RCN protein con-
tained a [4Fe-4S] cluster, given that its UVꢀvis spectrum does
not contain telltale signatures of [2Fe-2S] clusters, and most, if
not all, of the adventitiously bound iron was removed by
molecular-sieve chromatography. When SAM was added to
apo RlmNwt, methyl transfer to C355 did not occur, as evidenced
by no significant production of SAH (Figure 3B, purple dia-
monds). By contrast, when SAM or d3-SAM was added to apo
RlmNwtfRCN, ∼0.5 equiv of SAH was produced within the first
time point (15 s). Subsequent time points did not show increased
amounts of SAH, suggesting that 50% of apo RlmNwt was
reconstituted to its holo form (Figure 3B, red circles). Impor-
tantly, the quantity of SAH produced was directly proportional to
the RlmN cluster stoichiometry, providing further evidence that
the [4Fe-4S] cluster is required for methyl transfer. In addition,
apo RlmNwtfRCN was capable of catalyzing time-dependent
methylation of the target nucleotide in a 771-base synthetic
RNA substrate (Figure S10). Analysis of apo RlmNwtfRCN by
LCꢀMS after incubating it with SAM reveals a peak exhibiting a
mass-to-charge ratio (m/z) of 909.6 (+2 charge state) for the
C355-containing peptide fragment obtained after trypsin diges-
tion, indicating that it bears a methyl group (Figure 2F, red and
black traces). Moreover, the peak at m/z = 909.6 shifts to 911.2
(+2 charge state; thus the shift is ∼1.5 units instead of the expected
3 for the +1 charge state) when apo RlmNwtfRCN is incubated with
d3-SAM and similarly analyzed (Figure 3C red trace). This
modification is specific to C355 of RlmN, given that MS/MS
analysis of the peptides shows that the y+ ion series is shifted by +15
for fragments that bear C355 (y11+1 and above) (Figure S8).7
The effect of C118, another Cys residue that plays a key role in
RlmN catalysis, on methylation of C355 was also assessed. As
observed in Figure 2C, a large fraction of the AI C118A variant
contains a mCys modification, consistent with this residue
playing an insignificant role in SAM binding and methylation
of C355. In addition, the inclusion of the flavodoxin/flavodoxin
reductase/NADPH reducing system had no effect on C355
methylation, indicating that methyl transfer to C355 does not
require a reduced [4Feꢀ4S] cluster (Figure S9), and therefore
methylthiolation, which contain two distinct Fe/S clusters,18,29,30
RlmN and Cfr only contain the cluster housed in the canonical
CxxxCxxC motif.
While previous studies have firmly established one role for the
[4Fe-4S] cluster in the unique methylation reactions catalyzed
by RlmN and Cfr7ꢀ9,25 (reductive cleavage of SAM to generate
the 50-dA• intermediate), the role of the [4Fe-4S] cluster, if any,
in the first step of the reaction, methylation of a conserved
cysteine (C355 in RlmN and C338 in Cfr), has not been
addressed. Interestingly, when the RlmNC125A-C129A-C132A triple
variant was analyzed by ESI+-MS after alkylation with iodoace-
tamide and digestion with trypsin, the peptide containing C355
was found exclusively carbamidomethylated, indicating that it did
not bear a mCys residue (Figure 2B). This observation contrasts
with results obtained for RlmNwt, wherein almost 90% of the AI
protein contained a mCys residue, and very little of it was
alkylated by iodoacetamide during preparation for analysis by
mass spectrometry (Figure 2A).7 To assess the effect of the
ironꢀsulfur (Fe/S) cluster on generation of the mCys residue in
RlmN, apo RlmNwt was produced by expression of the Ec yfgB
gene, which encodes RlmN, in a modified M9 minimal medium
containing 75 μM o-phenanthroline to chelate available iron (see
Supporting Information for experimental details).31 A large
fraction of RlmN was produced in inclusion bodies when
expression was carried out in this manner. The soluble fraction
was isolated anaerobically under conditions identical to those for
isolation of holo RlmN, and a portion was set aside for
reconstitution of its [4Fe-4S] cluster with Fe and sulfide by
standard methods.32 Figure 3A shows UVꢀvis spectra of apo
RlmNwt isolated from Ec cultured in o-phenanthroline-contain-
ing medium (solid black line), and RlmNwt reconstituted with
iron and sulfide (apo RlmNwtfRCN) (solid red line). As can be
observed, the telltale features of [4Fe-4S] clusters are signifi-
cantly diminished in apo RlmNwt, consistent with the finding of
0.04 Fe and 0.07 sulfides per polypeptide (Table S2), but
reappear in the apo RlmNwtfRCN sample. Analysis of trypsin-
digested apo RlmNwt by LCꢀMS showed that it did not bear a
C355 mCys modification, consistent with the premise that SAM
bound to the [4Fe-4S] cluster acts as a methylating agent
(Figure 2E). To ensure that in vivo methylation of apo RlmNwt
was not the factor that distinguished soluble protein from
19588
dx.doi.org/10.1021/ja207327v |J. Am. Chem. Soc. 2011, 133, 19586–19589

Journal of the American Chemical Society
COMMUNICATION
Scheme 2. SAM Bound to the [4Fe-4S] Cluster of RlmN or
Cfr Is Activated toward Two Distinct Reactivities
(2) Markham, G. D. S-Adenosylmethionine. In Encyclopedia of Life
Sciences; John Wiley & Sons, Inc.: 2010.
(3) Silverman, R. B. Academic Press: New York, 2002.
(4) Hegazi, M. F.; Borchardt, R. T.; Schowen, R. L. J. Am. Chem. Soc.
1979, 101, 4359–4365.
(5) Iwig, D. F.; Grippe, A. T.; McIntyre, T. A.; Booker, S. J.
Biochemistry 2004, 43, 13510–13524.
(6) Woodard, R. W.; Tsai, M.-D.; Floss, H. G.; Crooks, P. A.;
Coward, J. K. J. Biol. Chem. 1980, 255, 9124–9127.
(7) Grove, T. L.; Benner, J. S.; Radle, M. I.; Ahlum, J. H.; Landgraf,
B. J.; Krebs, C.; Booker, S. J. Science 2011, 332, 604–607.
(8) Yan, F.; Fujimori, D. G. Proc. Natl. Acad. Sci. U.S.A. 2011,
108, 3930–3934.
(9) Yan, F.; LaMarre, J. M.; R€ohrich, R.; Wiesner, J.; Jomaa, H.;
Mankin, A. S.; Galoníc Fujimori, D. J. Am. Chem. Soc. 2010,
132, 3953–3964.
(10) Ban, N.; Freborn, B.; Nissen, P.; Penczek, P.; Grassucci, R. A.;
Sweet, R.; Frank, J.; Moore, P. B.; Steitz, T. A. Cell 1998, 93, 1105–1115.
(11) Harms, J.; Schluenzen, F.; Zarivach, R.; Bashan, A.; Gat, S.;
Agmon, I.; Bartels, H.; Francheschi, F.; Yonath, A. Cell 2001,
107, 679–688.
(12) Schuwirth, B. S.; Borovinskaya, M. A.; Hau, C. W.; Zhang, W.;
Vila-Sanjurjo, A.; Hoton, J. M.; Cate, J. H. D. Science 2005, 310, 827–834.
(13) Selmer, M.; Dunham, C. M.; Murphy, F. V., IV; Weixlbaumer,
A.; Petry, S.; Kelley, A. C.; Weir, J. R.; Ramakrishnan, V. Science 2006,
313, 1935–1942.
(14) The_UniProt_Consortium, The Universal Protein Resource
( UniProt). Nucleic Acids Res. 2007, 35 (Suppl. 1), D193ꢀD197.
(15) Toh, S.-M.; Xiong, L.; Bae, T.; Mankin, A. S. RNA 2008,
14, 98–106.
(16) Giessing, A. M. B.; Jensen, S. S.; Rasmussen, A.; Hansen, L. H.;
Gondela, A.; Long, K. S.; Vester, B.; Kirpekar, F. RNA 2009, 15, 327–336.
(17) Smith, L. K.; Mankin, A. S. Antimicrob. Agents Chemother. 2008,
52, 1703–1712.
(18) Booker, S. J. Curr. Opin. Chem. Biol. 2009, 13, 58–73.
(19) Frey, P. A.; Booker, S. J. Adv. Protein Chem. 2001, 58, 1–45.
(20) Frey, P. A.; Hegeman, A. D.; Ruzicka, F. J. Crit. Rev. Biochem.
Mol. Biol. 2008, 43, 63–88.
(21) Sofia, H. J.; Chen, G.; Hetzler, B. G.; Reyes-Spindola, J. F.;
Miller, N. E. Nucleic Acids Res. 2001, 29, 1097–1106.
(22) Chatterjee, A.; Li, Y.; Zhang, Y.; Grove, T. L.; Lee, M.; Krebs,
C.; Booker, S. J.; Begley, T. P.; Ealick, S. E. Nat. Chem. Biol. 2008,
4, 758–765.
most likely proceeds by a polar nucleophilic displacement, as do
all characterized SAM-dependent methyltransferases.1
Herein we have shown that RlmN and Cfr catalyze two separate
and distinct reactions, exploiting both polar and radical reactiv-
ities of SAM within a single polypeptide (Scheme 2). Indeed,
these proteins both possess methyltransferase and methylsynthase
activities. They contain only one [4Fe-4S] cluster (ligated by
cysteines in CxxxCxxC motifs) to which SAM binds via its α-amino
and carboxy groups.25 In this conformation, methyl transfer to
unmodified C355 of RlmN (C338 of Cfr) takes place rapidly.
Release of the product, SAH, permits binding of a second molecule
of SAM to the exact same site; however, methyl transfer to the
reactive cysteine is now blocked. Upon binding of the RNA
substrate and the addition of an electron to the [4Fe-4S]2+
cluster, SAM is now induced to fragment into methionine and
the 50-dA•, the latter initiating catalysis by abstracting a hydrogen
atom from the mCys residue. Our studies show that methyl
transfer to the target Cys residue does not require the presence of
substrate. Whether substrate binding accelerates this step is yet to
be determined. However, our previous studies showed that
substrate binding does indeed effect radical formation when
the physiological reducing system is used to supply the requisite
electron.7 Excitingly, this mode of catalysis is a clever twist on the
“principle of economy in the evolution of binding sites,” wherein
Nature evolves only a single substrate-binding site for reactions
that involve two or more substrates with similar structural elements,
and then mediates transfer of a group from one to the other by means
of an intermediate enzyme functional group covalent adduct.33
The Cfr and RlmN reactions are the first recognized instances in
which a single substrate-binding site activates one molecule both
for polar and radical-based chemistry.
(23) MartinezꢀGomez, N. C.; Downs, D. M. Biochemistry 2008,
47, 9054–9056.
(24) McGlynn, S. E.; Boyd, E. S.; Shepard, E. M.; Lange, R. K.;
Gerlach, R.; Broderick, J. B.; Peters, J. W. J. Bacteriol. 2010,
192, 595–598.
(25) Boal, A. K.; Grove, T. L.; McLaughlin, M. I.; Yennawar, N. H.;
Booker, S. J.; Rosenzweig, A. C. Science 2011, 332, 1089–1092.
(26) Vey, J. L.; Drennan, C. L. Chem. Rev. 2011, 111, 2487–2506.
(27) Booth, M. P. S.; Challand, M. R.; Emery, D. C.; Roach, P. L.;
Spencer, J. Protein Expr. Purif. 2010, 74, 204–210.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental details and char-
b
acterization data. This material is available free of charge via the
Internet at http://pubs.acs.org.
(28) Kaminska, K. H.; Purta, E.; Hansen, L. H.; Bujnicki, J. M.;
Vester, B.; Long, K. S. Nucleic Acids Res. 2010, 38, 1652–1663.
(29) Atta, M.; Mulliez, E.; Arragain, S.; Forouhar, F.; Hunt, J. F.;
Fontecave, M. Curr. Opin. Struct. Biol. 2010, 20, 1–9.
(30) Booker, S. J.; Cicchillo, R. M.; Grove, T. L. Curr. Opin. Chem.
Biol. 2007, 11, 543–552.
(31) Parkin, S. E.; Chen, S.; Ley, B. A.; Mangravite, L.; Edmondson,
D. E.; Huynh, B. H.; Bollinger, J. M., Jr. Biochemistry 1998,
37, 1124–1130.
(32) Cicchillo, R. M.; Lee, K.-H.; Baleanu-Gogonea, C.; Nesbitt,
N. M.; Krebs, C.; Booker, S. J. Biochemistry 2004, 43, 11770–11781.
(33) Frey, P. A., Nucleotidyltransferases and Phosphotransferases:
Stereochemistry and Covalent Intermediates. In The Enzymes, 3rd ed.;
Sigman, D. S., Ed.; Academic Press, Inc.: San Diego, 1992; Vol. 20.
’ AUTHOR INFORMATION
Corresponding Author
ckrebs@psu.edu; squire@psu.edu
’ ACKNOWLEDGMENT
This work was supported by the National Institutes of Health
(GM-63847 to S.J.B.). We thank Tatiana N. Laremore of the
PSU MS facility for collecting high-resolution MS data.
’ REFERENCES
(1) Frey, P. A.; Hegeman, A. D. Oxford University Press: New York,
2007.
19589
dx.doi.org/10.1021/ja207327v |J. Am. Chem. Soc. 2011, 133, 19586–19589