Assay of both activities of the bifunctional
tRNA-modifying enzyme MnmC reveals a
kinetic basis for selective full modification
of cmnm5s2U to mnm5s2U
David Pearson* and Thomas Carell*
Center for Integrated Protein Science (CiPSM) at the Department of Chemistry, LMU Munich,
Butenandtstrasse 5–13, 81377 Munich, (Germany)
Received December 25, 2010; Revised January 25, 2011; Accepted January 26, 2011
Transfer RNA (tRNA) contains a number of complex
‘hypermodified’ nucleosides that are essential for a
number of genetic processes. Intermediate forms of
these nucleosides are rarely found in tRNA despite
the fact that modification is not generally a complete
process. We propose that the modification machin-
ery is tuned into an efficient ‘assembly line’ that
performs the modification steps at similar, or se-
quentially increasing, rates to avoid build-up of
possibly deleterious intermediates. To investigate
this concept, we measured steady-state kinetics
for the final two steps of the biosynthesis of the
mnm5s2U nucleoside in Escherichia coli tRNAGlu,
which are both catalysed by the bifunctional MnmC
enzyme. High-performance liquid chromatography-
based assays using selectively under-modified tRNA
substrates gave a Km value of 600nM and kcat
0.34s?1for the first step, and Km 70nM and kcat
0.31s?1for the second step. These values show
that the second reaction occurs faster than the first
reaction, or at a similar rate at very high substrate
modified mnm5(s2)U while avoiding build-up of the
nm5(s2)U intermediate. The assay method developed
here represents a general approach for the com-
parative analysis of tRNA-modifying enzymes.
non-canonical nucleosides, which perform a number of
biological functions including structural stabilization (1)
molecules areheavily modifiedwith many
and optimization of codon binding (2). These modifica-
tionsrange from simple
‘hypermodified’ residues involving multistep biosyntheses.
The hypermodified nucleosides are particularly interest-
ing, as the cellular machinery allocates a relatively large
amount of energy to the biosyntheses of these modifica-
tions, indicating a high importance to cellular processes.
These nucleosides are most often located in or adjacent to
the anticodon (3), and are involved in key translation
processes anticodon–stem loop stabilization (2), codon
binding and wobble base pairing (4,5). Recent studies on
the quantification of modified nucleosides show that
tRNA is not always fully modified, instead containing a
significant number of unmodified positions [(6,7) and un-
published data]. However, partially modified precursors of
hypermodified nucleosides are generally not observed,
with the exception of those present in tRNAs that
always contain the partially modified form (3,8–10).
Together, these results
pathways are organized to give complete hypermodified
nucleosides, avoiding partially modified intermediate nu-
cleosides. Biosynthetic tuning of this type would require a
coordinated ‘assembly line’ process in which sequential
modifications are performed at similar, or increasing,
rates in order to efficiently construct the final products
without a build-up of intermediates. We propose that
this process is likely to be predominantly controlled by
tuning of the activities and abundances of the tRNA-
modifying proteins. Alternatively, biosynthetic tuning
could involve selective degradation of partially modified
tRNAs (11) or control of enzyme compartmentalization
(12) in eukaryotes.
The 5-methylaminomethyl(-2-thio)uridine (mnm5(s2)U)
residue, which is present in position 34 of the anticodon of
Escherichia coli tRNAs Glu, Lys, Arg and, probably, Gln
(3), is particularly interesting due to its involvement in
wobble basepairing (4,5,13,14)
*To whom correspondence should be addressed. Tel: +49 892 1807 7849; Fax: +49 892 1807 7756; Email: email@example.com
Correspondence may also be addressed to Thomas Carell. Tel: +49 892 1807 7750; Fax: +49 892 1807 7756; Email: firstname.lastname@example.org-
Nucleic Acids Research, 2011, Vol. 39, No. 11Published online 8 February 2011
? The Author(s) 2011. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/
by-nc/2.5), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
four-step biosynthetic pathway (Scheme 1). The s2group
is inserted by the MnmA enzyme (15,16), while the mnm5
group is independently formed in a three-step sequence.
addition of glycine to a reactive intermediate (17–19).
The cmnm5(s2)U base is then converted to mnm5s2U in
two steps by the bifunctional enzyme MnmC (20–22),
which initially carries out a flavin adenine dinucleotide
2-thio)uridine (nm5(s2)U), followed by methylation to
cofactor. To test our proposal that the modification
steps are coordinated by tuning of the kinetics and abun-
dance of biosynthesis enzymes, we investigated the
activities of the two modification steps performed by
MnmC. As these two reaction steps are both performed
by a single enzyme, differences in enzyme abundance for
the individual steps are excluded. Therefore, control of
enzyme kinetics would be necessary for our proposed
assembly-line type process. To our knowledge, the
partial modification nm5(s2)U has not been reported in
normal cellular tRNA, indicating that this intermediate
is avoided by the biosynthetic machinery. The presence
or absence of under-modified cmnm5(s2)U is not clear
due to the natural occurrence of this nucleoside at
certain tRNA positions.
Kinetic assay of complex tRNA-modifying enzymes
is generally problematic due to the difficulty of obtain-
the measurement ofreaction
cmnm5(s2)U!mnm5(s2)U assays have used either total
tRNA from cells lacking MnmC (20–22) or substantially
under-modified tRNA from in vitro transcription (18). The
most accurate assay results to date indicate that the first
step is rate limiting (20), but could only measure the
methylation reaction and did not have the defined sub-
strates necessary for full kinetic analysis. In order to
obtain such substrates for reaction with MnmC, we
overexpressed tRNAGluin an E. coli expression strain
lacking MnmC, then isolated the selectively under-
anion-exchange high-performance liquid chromatography
(HPLC). The second substrate (containing nm5s2U) was
formed by reaction of the cmnm5s2U-containing tRNA
with MnmC in the absence of the SAM cofactor.
Reaction of each substrate with MnmC was assayed
progress. To further assess the importance of the
measured growth curves for an MnmC knockout E. coli
strain and its corresponding wild-type strain.
to monitor reaction
MATERIALS AND METHODS
Preparation of recombinant MnmC protein
His-tagged MnmC was cloned and expressed as previously
reported (21) in E. coli BL21 cells, and purified as follows.
All steps were performed at 0–4?C. Harvested cells were
suspended in buffer A (50mM Tris–HCl, pH 8.0, 200mM
KCl, 10mM MgCl2, 3mM b-mercaptoethanol, 10%
glycerol) with the addition of lysozyme (0.2mg/ml), then
lysed using a French press. The lysate was cleared by cen-
trifugation (24000g for 30min) and filtration (0.45mM)
and was applied to a HiScreen IMAC FF column (GE
Healthcare) charged with Ni2+, using an A¨KTApurifier
system (GE Healthcare). The column was washed with
Scheme 1. Biosynthesis of mnm5(s2)U in tRNA. X=O (in U) or S (in s2U), R=cmnm5, nm5or mnm5. Modified positions are coloured in red.
Nucleic Acids Research, 2011,Vol.39, No. 11 4819
buffer A supplemented with 5mM imidazole then MnmC
was eluted with buffer A containing 0.5M imidazole. The
protein obtained was concentrated and exchanged into
buffer B (buffer A lacking KCl) using an Amicon Ultra
30000 MWCO centrifugal filter (Millipore) then applied
to a 1ml Mono Q column (GE Healthcare) equilibrated
with the same buffer. The enzyme was eluted with a linear
gradient of buffer C (buffer A containing 500mM KCl),
concentrated then applied to a Superdex 200 GL10/300
size-exclusion column and eluted with buffer C. The
purified protein was concentrated to 10mg/ml in buffer
D (buffer A containing only 50mM KCl) and stored at
?80?C. A yield of ?10mg of purified protein was
obtained from 4l of expression culture. Protein concentra-
tion was calculated by Bradford assay.
MnmC knockout from an E. coli expression strain
The MnmC gene was removed from T7 express E. coli
(NEB) by replacement with a chloramphenicol resistance
cassette. The cassette was made by polymerase chain
reaction (PCR) amplification of the cmr gene from the
pDEST17 plasmid (Invitrogen) using Phusion polymerase
incorporating 50 bases of the MnmC gene: 50-TGAAAC
TGCCACT. Chemically competent T7 express cells were
transformed with the pSIM6 plasmid (23), kindly supplied
by Donald L. Court, then made electrocompetent and
transformed with the PCR-amplified Cam cassette. After
pregrowth in LB medium then selection on Cam plates,
several colonies were chosen and confirmed by PCR to
contain the Cam gene and no MnmC gene (primers:
CCGCTTTA CCCTTCAAC). The cells were made
electrocompetent then stored at ?80?C in 10% glycerol.
Cloning, expression and purification of
The tRNAGlugene (corresponding to RNA sequence GU
AGGGGACGCCA) was amplified by PCR from NEB5a
E. coli (NEB) using Phusion polymerase and the following
primers to give an EcoRI cleavage site and the T7
promoter 50of the tRNA sequence, and BsaI and
HindIII cleavage sites 30of the tRNA: 50-GAATTCTAA
CCAGGA and 50-AAGCTTGGTCTCATGGCGT CCC
cleaved with the EcoRI and HindIII restriction enzymes,
and ligated into the pSGAT2 plasmid. The presence of the
tRNAGlugene was confirmed by sequencing. The expres-
sion plasmid was then used to transform electrocompetent
?MnmC T7 express cells (see above). Transformed cells
were grown in LB-medium
carbenicillin (100mg/ml) to an OD600 of 0.5, then
The PCRproduct was
induced with 1mM isopropyl b-d-1-thiogalactopyranoside
(IPTG) for 6h and harvested. The cell pellets were phenol
extracted as reported previously (6) to give total cell
tRNA (?20mg per litre of E. coli culture), which was
dissolved at ?10mg/ml in buffer E (100mM Tris, pH 8,
50mM MgCl2) after ethanol precipitation.
The crude tRNA was then separated by two sequential
anion exchange HPLC purifications using a DNAPac
PA100 22?250mm column (Dionex) and a Merck
Hitachi Lachrom system. The first purification was done
in buffer F (100mM NaOAc, pH 5.0, 50mM MgCl2) with
a gradient of 205–210mM NaCl, followed by concentra-
tion of the tRNA using an Amicon Ultra 10000 MWCO
centrifugal filter (Millipore) and a further purification in
buffer E using a gradient of 205–215mM NaCl The
purified tRNA was exchanged into buffer E and stored
at ?20?C. To obtain nm5s2U-containing tRNA, the
cmnm5s2U-containing tRNA after pH 5 purification
(400mg) was incubated with MnmC (40mg) in 1ml of
buffer G (50mM Tris, 20mM NH4Cl) at 37?C for 1h
before the pH 8 purification. For mnm5s2U-containing
tRNA, the same procedure was carried out using 100mg
cmnm5s2U-tRNA and 10mg MnmC with the addition of
SAM (final conc. 800mM).
Characterization of tRNA
The purified tRNA was analysed by digestion with
RNases A or T1, followed by matrix-assisted laser desorp-
tion/ionization-mass spectrometry (MALDI-MS) (24).
For identification of pseudouridine bases, tRNA (2mM)
and acrylonitrile (1.8M) were initially dissolved in
TEA-AcOH buffer (0.98M), ethanol (36%) at a total
volume of 34ml, incubated at 70?C for 2h, then
lyophilized, redissolved in water and exchanged into
pure water using a centrifugal filter.
RNase digests were performed by incubation of tRNA
(0.2mg) with RNase A (0.1mg, Fermentas) or RNase T1
(1U, Ambion) in 3mg/ml 3-HPA, 1% acetonitrile at a
total volume of 3ml at 37?C for 2h (for RNase A) or 7h
(for RNase T1). The hydrolysed tRNA was desalted three
times using a 0.025mM drop dialysis filter (Millipore) then
0.5ml of the solution was mixed with 0.5ml of matrix
(0.6M 3-HPA, 32mM picolinic acid, 18mM diammonium
citrate, 10% acetonitrile) on a MALDI sample plate and
allowed to dry. MALDI-TOF MS was performed using a
Bruker Autoflex II in negative mode with 19kV and 16kV
ionization source voltages, 8.55kV lens voltage, 20kV re-
flector voltage and 200ns pulsed ion extraction. Masses
were externally calibrated using oligonucleotides of
known sequence, or internally calibrated using invariant
tRNA fragments from the sample (e.g. for accurate
analysis of the anticodon–stem loop fragment).
Assay of MnmC
For assay of the FAD-dependent MnmC activity, the fol-
lowing final reaction mixture was used: cmnm5s2U-
containing tRNA (0.13–2.0mM), MnmC (1.7nM), Tris
(60mM, pH 8.0), NH4Cl (20mM) KCl (2.5mM), MgCl2
(0.5%), bovine serum albumin (BSA; 5mg/ml) [note:
4820Nucleic Acids Research, 2011,Vol.39, No. 11
KCl, b-mercaptoethanol, glycerol and BSA were only
present as they were used in the enzyme storage buffer].
All components except the enzyme were assembled on ice,
then preincubated at 37?C in a water bath for 3min. The
enzyme (in buffer D plus 0.1mg/ml BSA) was then added,
and aliquots of at least 20ng tRNA were removed within
the first 5min of reaction, diluted at least 2?with buffer F
then injected onto a DNAPac PA100 4?250mm column
(Dionex) and analysed in buffer E with a gradient of
175–180mM NaCl. Reaction volumes were typically
40ml. Each assay was carried out in triplicate.
For assay of the SAM-dependent MnmC activity, the
following final reaction mixture was used: nm5s2U con-
taining tRNA (25–400nM), MnmC (42pM), SAM
(500mM), Tris (60mM, pH 8.0), NH4Cl (20mM) KCl
(5mg/ml) [note: KCl, b-mercaptoethanol, glycerol and
BSA and ethanol were only present as they were used in
the enzyme or SAM storage buffers]. The reaction was
performed as described for the FAD-dependent activity
above, using an HPLC gradient of 160–165mM NaCl
and typical reaction volumes of 160ml.
Reaction progress was calculated from the ratio of the
areas of the product and starting material UV absorbance
peaks (260nm) in the HPLC chromatograms. In the case
of the SAM-dependent reaction, the product and starting
material peaks overlapped, and were deconvoluted using
the eXPFit14 add-in (www.chem.qmul.ac.uk/software/
eXPFit.htm) for Microsoft Excel. Reaction progress was
then plotted versus time, and the linear slope of each plot
measured as the initial reaction rate at the corresponding
substrate concentration. Michaelis–Menten curves were
fitted to the initial rate versus [S] data using the EZ-Fit5
program (Perrella Scientific).
To ensure that the UV absorbance data gave a linear
response for the product and substrate peak areas, cali-
bration injections were performed at a range of concen-
trations of cmnm5s2U, nm5s2U and mnm5s2U containing
tRNA. To ensure that the SAM concentration of 500mM
was saturating, preliminary nm5s2U to mnm5s2U assays
were carried out with a range of SAM concentrations.
SAM concentrations from 100 to 1000mM were found
to give similar reaction rates, indicating that the KMfor
SAM is much <500mM.
Bacterial growth rate experiments
Bacterial strains were obtained from the Keio collection of
E. coli knockouts (obtained from the NBRP-E.coli at
NIG). Growth rates of the wild-type (BW25113) and
?MnmC (JW5380) strains were measured at 37?C in
rich (LB) medium by recording optical densities at
600nm (OD600) at various time points. Pre-cultures were
initially grown exponentially for several generations
before dilution into fresh medium for data collection.
Each strain was grown in triplicate under identical condi-
tions. Growth rates were calculated by fitting an exponen-
tial curve of the type y=Aekt.
Preparation of selectively under-modified tRNAGlu
substrates for MnmC
In order to prepare selectively under-modified tRNAGlu,
we assembled a tRNA overexpression system (25) using an
E. coli strain that lacks the MnmC gene. As the expression
construct, the E. coli tRNAGlugene was cloned into a
protein expression plasmid (pSGAT2) containing a T7
promoter for inducible expression. An additional T7
promoter was added directly 50to the tRNA gene in
order to expedite tRNA processing (as the RNA tran-
scripts from this promoter should begin at the first
position of the tRNA and therefore not require enzymatic
50-cleavage). To obtain an expression strain lacking
MnmC, we used the recombineering method (23) to
replace the MnmC gene in T7 expression cells with an
antibiotic resistance cassette. The tRNA expression
plasmid was then used to transform ?MnmC cells, from
which tRNAGluwas expressed and isolated by phenol ex-
traction (6). The major tRNA ‘modivariants’ (10)
obtained were then separated by anion-exchange HPLC
(Figure1aand b) then
MALDI-MS analysis of RNase digests (Figure 2). All
major peaks in the HPLC chromatogram were found to
(Figure 2a–c and Supplementary Figure S1), showing
that a very high level of expression was sustained by the
cells. A single modivariant (tRNA 1, Figure 1a and c) was
found to contain the desired cmnm5s2U modification
(Figure 2c) as well as the other expected mass-detectable
tRNAGlumodifications (T and m2A, Figure 2a–c). The
presence of the two expected pseudouridine (?) nucleo-
tides in this tRNA was then confirmed by reaction with
(Supplementary Figure S2) (24). The other modivariants
(Supplementary Figure S1), as labelled in Figure 1c.
The purified cmnm5s2U-containing tRNA 1 was reacted
with purified recombinant MnmC (21) in the absence or
presence of the SAM cofactor to give nm5s2U or
mnm5s2U-containing tRNA, respectively (tRNA 5 and
6, Figure 1c). These tRNA products were further
purified by anion-exchange HPLC and analysed by
MALDI-MS ofRNase T1
products. Mass peaks corresponding to the anticodon–
stem loop fragments confirmed the identity of each
tRNA (Figure 2d and e). High purity of the tRNAs 1
and 5 was determined from the lack of contaminating
under-modified fragments in MALDI-MS spectra, the
single major HPLC peak detected for each tRNA, and
the observed complete reaction with MnmC as measured
this MALDI sequencing
Development of an HPLC-based assay for both
activities of MnmC
Anion-exchange HPLC of the cmnm5s2U- mnm5s2U- and
nm5s2U-containing tRNAs showed that each elutes at a
different retention time (Figure 3a and b), allowing an
HPLC-based assay to be used to measure both activities
Nucleic Acids Research, 2011,Vol.39, No. 114821
of the enzyme. As full separation of tRNA 5 and 6 was not
obtained, a computational peak deconvolution was
employed to calculate the area of each peak (Figure 3b).
Preliminary studies confirmed the linear response of the
areas of the product and substrate peaks, and the approxi-
mate saturating concentration of the SAM cofactor
(?100mM). Subsequently, steady-state kinetic assays for
each reaction step were carried out. The conditions for
each assay were chosen to be similar in order to
compare kinetic constants, and were based on conditions
known to optimize enzyme activity (20) with the addition
of MgCl2to stabilize tRNA (26). Rate versus substrate
concentration curves are shown in Figure 3c and d.
Fitting a rectangular hyperbola to these curves gives the
Michaelis–Menten constants shown in Table 1. The Km
for the second reaction is substantially lower (?9-fold)
than that of first reaction, and the kcatconstant is not
significantly different. This shows that the enzyme binds
the second substrate (tRNA 5) tighter than the first
(tRNA 1). This clearly will result in a higher rate of
reaction for the second step, or similar rates at very high
substrate concentrations where the kcatconstant becomes
dominant. The kcat/Km value, a general indicator of
activity, is correspondingly eight times higher for the
second reactionstep, showing
activity at intermediate substrate concentrations.
To further supportour
hypermodified nucleosides negatively affect translation
processes, we measured the growth rate of a well-defined
MnmC knockout strain from the Keio collection of E. coli
knockouts (Figure 4). Exponential fits to the growth
curves of the ?MnmC and wild-type strains reveal a
larger growth constant, k, for the wild-type strain. As
the knockout strain is expected to contain the cmnm5
Figure 1. Overexpressed tRNAGlufrom a ?MnmC E. coli strain. (a) HPLC of total extracted tRNA. HPLC buffers: 100mM Tris pH 8,
50!150mM MgCl2. The four major modivariants are labelled 1–4. Each was isolated by HPLC for subsequent MS analysis of modified nucleo-
sides. (b) Representative purified tRNA (tRNA 5, see below) after anion-exchange purifications at pH 5 and 8. HPLC buffers: 100mM Tris pH 8,
50mM MgCl2, 0!500mM NaCl. [Note: The MgCl2and NaCl gradients in (a) and (b) were used interchangeably for analytical purposes] (c)
Diagrams showing the modified nucleosides present in tRNAs 1–5 and those present in tRNAs 5 and 6, the products of reaction of tRNA 1 with
MnmC (see below). Nucleosides labelled in red (or orange or yellow) were identified by MS (Figure 2 and Supplementary Figure S1). The presence of
? nucleosides labelled in grey was not confirmed. tRNA 1 contains cmnm5s2U34and all other expected modifications, while tRNA 2–4 are less fully
modified at anticodon–stem loop postions as shown. The nm5s2U and mnm5s2U modifications obtained by reaction with MnmC are shown in orange
and yellow, respectively, to differentiate from cmnm5s2U.
4822 Nucleic Acids Research, 2011,Vol.39, No. 11
Figure 2. MALDI-MS spectra of RNase digests of tRNA 1, 5 and 6. (a) RNase A digest of tRNA 1 (cmnm5s2U-tRNAGlu). (b) RNase T1 digest of
tRNA 1. (c) Expanded section of (b) showing the anticodon–stem loop fragment of tRNA 1. Extra peaks are assigned as follows. a: 3165.5,
Fragment A [AAUCCCCUAGcp (cp=cyclic phosphate)]; b: 3182.9, Fragment B (AAUCCCCUAGp); c: 3203.7, Fragment A – H+K and
Fragment B – H+Na; d: 3221.3, Fragment B – H+K; e: 3235.6, Fragment C (CCCUcmnm5s2UUCm2ACGcp); f: 3242.4, Fragment A –
2H+2K; g: 3258.4, Fragment B – 2H+2K; h: 3274.4, Fragment C – H+K and Fragment D (CCCUcmnm5s2UUCm2ACGp) – H+Na; i:
3290.9, Fragment D – H+K; j: 3296.0, Fragment B – 3H+3K. (d) Section of a RNase T1 digest of tRNA 5 (nm5s2U-tRNAGlu) showing the
nm5s2U-containing fragment. Extra peaks are assigned as follows. a: 3165.3, Fragment A (AAUCCCCUAGcp); b: 3177.8, Fragment E
(CCCUnm5s2UUCm2ACGcp), c: 3182.9, Fragment B (AAUCCCCUAGp); d: 3204.2, Fragment B – H+Na; e: 3215.8, Fragment E – H+K; f:
3221.2, Fragment B – H+K; g: 3233.5, Fragment F (CCCUnm5s2UUCm2ACGp) – H+K; h: 3243.3, Fragment B – 2H+Na+K; i: 3258.5,
Fragment B – 2H+2K; j: 3270.1, Fragment F – 2H+2K; k: 3296.1, Fragment B – 3H+3K; l: 3307.8, Fragment F – 3H+3K. (e) Section of an
RNase T1 digest of tRNA 6 (mnm5s2U-tRNAGlu) showing the mnm5s2U-containing fragment. Extra peaks are assigned as follows. a: 3165.4,
Fragment A (AAUCCCCUAGcp); b: 3182.9, Fragment B (AAUCCCCUAGp); c: 3191.0, Fragment G (CCCUmnm5s2UUCm2ACGcp); d: 3204.9,
Fragment B – H+Na; e: 3221.0, Fragment B – H+K; f: 3229.3, Fragment G – H+K; g: 3246.4, Fragment H (CCCUmnm5s2UUCm2ACGp) –
H+K; h: 3258.9, Fragment B – 2H+2K; i: 3267.3, Fragment G – 2H+2K; j: 3283.9, Fragment H – 2H+2K; k: 3296.5, Fragment B – 3H+3K; l:
3306.7, Fragment A – 3H+3K; m: 3322.0, Fragment H – 3H+3K; n: Fragment B – 4H+4K. The tRNA fragments analysed in (a) and (b) are
labelled with mass and sequence on each spectrum, and coloured in red on each corresponding tRNA diagram. The data in (c)–(e) are calibrated to
the constant fragment B (AAUCCCCUAGp, 3182.9) to allow accurate determination of the variable anticodon fragment.
Nucleic Acids Research, 2011,Vol.39, No. 114823
modification in place of mnm5, the result shows that this
cmnm5under-modification is detrimental to cell growth.
The MnmC enzyme catalyses the final two steps of the
biosynthesis of the mnm5(s2)U34 nucleoside in tRNA.
The mnm5biosynthesis pathway seems to be tuned to
give only the fully modified residue, in particular
avoiding nm5under-modification [cmnm5s2U has been
detected in vivo (8); however, it is not clear whether this
is due to under-modification or the natural presence of
cmnm5in certain tRNAs]. The biosynthetic machinery
could be regulated in a number of ways, for example by
selective degradation of partially modified tRNA (11),
controlled expression [or compartmentalization, in eu-
karyotes (12)] of particular modifying enzymes or by op-
timization of enzyme activities and specificities. This study
reveals that the two transformations catalysed by MnmC
are regulated by enzyme kinetics, as the second substrate is
bound tighter than the first. Even at relatively high sub-
strate concentrations (?3mM), the two reactions would be
performed at similar rates, resulting in a relatively low
concentration of the partially modified intermediate.
Degradation of the under-modified tRNA cannot be
ruled out as an additional mechanism; however, this
seems unlikely given the existing kinetic control and the
Figure 3. HPLCs and Michaelis–Menten plots for each MnmC catalysed reaction. (a) Representative HPLC showing complete separation of tRNA
1 from tRNA 5. HPLC gradient: 100mM Tris, 50mM MgCl2, 175!180mM NaCl over 1!20min. (b) Representative HPLC showing partial
separation of tRNA 5 from tRNA 6, and calculated peaks for each tRNA. HPLC gradient: 100mM Tris, 50mM MgCl2, 160!165mM NaCl over
1!30min. (c) Michaelis–Menten plot for the FAD-dependent cmnm5s2U!nm5s2U demodification. (d) Michaelis–Menten plot for the
SAM-dependent nm5s2U!mnm5s2U methylation. Rates in (c) and (d) represent the amount of substrate formed in a 40ml reaction per minute
per milligram enzyme.
Table 1. Michaelis–Menten constants for each MnmC catalysed
4824 Nucleic Acids Research, 2011,Vol.39, No. 11
inefficiency associated with degradation of a highly
In terms of the applicability of our in vitro study to
in vivo enzyme activity, several aspects need to be con-
sidered. It is possible that the modification reactions are
affected in vivo by other cellular components. Importantly,
4concentrations were reported previously
to affect enzyme activity. However, this is unlikely to sub-
stantially change the results reported here, as both
activities of MnmC were reported to be similarly
affected by changes in the concentrations of these ions
(20). Intracellular SAM concentrations are reported to
be similar to those used in our assay (27), so our result
is likely to be relevant in this respect. Other modifying
enzymes have been found to be dependent on the state
of tRNA processing, such as TilS, which modifies
tRNAIle2at a precursor stage (28). None of the cmnm5-
or mnm5-containing tRNA obtained in the tRNAGlu
overexpressions here and in a previous study (10) was
found to lack other modifications, implying that MnmC
acts on tRNA at a late stage of processing. Therefore, it is
likely that our substrates represent those normally
modified by MnmC.
The fact that the mnm5(s2)U pathway is tuned to give
full modification suggests that the partially modified nu-
cleoside (particularly nm5(s2)U) is particularly detrimental
to translation. Our result that ?MnmC E. coli, which
contain cmnm5(s2)U instead of mnm5(s2)U, grows signifi-
cantly slower than wild-type E. coli, supports this
proposal. It is likely that the observed slow growth of
the MnmC knockout is a result of disruption of the
reported fine tuning of A and G wobble pairing with
mnm5s2U (14). Interestingly, some species completely
lack an MnmC-type modification and instead use
cmnm5s2U in the 34 position of tRNA e.g. Bacillus
subtilis (3) and Saccharomyces cerevisiae (29). It is
unclear at this stage as to why other organisms expend
energy to further modify cmnm5s2U to mnm5s2U.
The assay method reported here is a general approach
that should be applicable to comparative assay of most
tRNA modifications. Although a number of very effective
radioactivity-based assays have been reported (20,30),
these require specific radioactive substrates and are not ap-
plicable to modifications for which no radioactive incorp-
oration can be measured (e.g. the cmnm5s2U!nm5s2U
reverse-phase HPLC of full RNase digests is also
possible (18) but is complicated by the extra digestion
step and the requirement for quantification standards.
Our analysis of intact tRNA using HPLC also allows veri-
fication of tRNA stability during reaction, which is often
not taken into account (31). We believe our general
approach will allow better comparison of the activities
of tRNA-modifying enzymes in future studies.
In summary, we reveal that the full modification of the
mnm5s2U tRNA modification is regulated by kinetic
tuning of the activities of the MnmC enzyme. Our
method represents a general approach to the comparative
analysis of tRNA-modifying enzymes, and should be ap-
plicable to the study of other modification pathways. With
the methods for selective preparation and assay of
modified tRNA developed in this article, future studies
will be carried out to further elucidate the regulation of
tRNA modification and
hypermodification on genetic processes.
ofreaction progress by
the impactof partial
Supplementary Data are available at NAR Online.
We thank Markus Mu ¨ ller for training in microbiological
Alexander von Humboldt Foundation [postdoctoral
fellowship to D.P.]; the Deutsche Forschungsgemeinschaft
[grant numbers CA275/8-4, SFB 749]; Excellence Cluster
CIPSM. Funding for open access charge: Deutsche
Forschungsgemeinschaft [SFB 749].
Conflict of interest statement. None declared.
1. Motorin,Y. and Helm,M. (2010) tRNA stabilization by modified
nucleotides. Biochemistry, 49, 4934–4944.
2. Agris,P.F. (2008) Bringing order to translation: the contributions
of transfer RNA anticodon-domain modifications. EMBO Rep., 9,
3. Ju ¨ hling,F., Mo ¨ rl,M., Hartmann,R.K., Sprinzl,M., Stadler,P.F. and
Pu ¨ tz,J. (2009) tRNAdb 2009: compilation of tRNA sequences and
tRNA genes. Nucleic Acids Res., 37, D159–D162.
4. Agris,P.F., Vendeix,F.A.P. and Graham,W.D. (2007) tRNA’s
wobble decoding of the genome: 40 years of modification.
J. Mol. Biol., 366, 1–13.
5. Takai,K. and Yokoyama,S. (2003) Roles of 5-substituents of
tRNA wobble uridines in the recognition of purine-ending
codons. Nucleic Acids Res., 31, 6383–6391.
6. Bru ¨ ckl,T., Globisch,D., Wagner,M., Mu ¨ ller,M. and Carell,T.
(2009) Parallel isotope-based quantification of modified tRNA
nucleosides. Angew. Chem. Int. Ed., 48, 7932–7934.
Figure 4. Growth curves showing the impact of an MnmC gene
knockout on E. coli growth rate. Strains are from the Keio
Collection: BW25113 (wild type) and JW5380 (?MnmC). For each
strain, the geometrical symbols on the graph (triangle, square,
diamond) each represent the data points for one independent growth
Nucleic Acids Research, 2011,Vol.39, No. 11 4825
7. Saikia,M., Fu,Y., Pavon-Eternod,M., He,C. and Pan,T. (2010) Download full-text
Genome-wide analysis of N1-methyl-adenosine modification in
human tRNAs. RNA, 16, 1317–1327.
8. Buck,M., Connick,M. and Ames,B.N. (1983) Complete analysis
of tRNA modified nucleosides by high-performance liquid
chromatography: The 29 modified nucleosides of Salmonella
typhimurium and Escherichia coli tRNA. Anal. Biochem., 129,
9. McCloskey,J.A., Graham,D.E., Zhou,S., Crain,P.F., Ibba,M.,
Konisky,J., So ¨ ll,D. and Olsen,G.J. (2001) Post-transcriptional
modification in archaeal tRNAs: identities and phylogenetic
relations of nucleotides from mesophilic and hyperthermophilic
Methanococcales. Nucleic Acids Res., 29, 4699–4706.
10. Madore,E., Florentz,C., Giege ´ ,R., Sekine,S.-i., Yokoyama,S. and
Lapointe,J. (1999) Effect of modified nucleotides on Escherichia
coli tRNAGlustructure and on its aminoacylation by
glutamyl-tRNA synthetase. Eur. J. Biochem., 266, 1128–1135.
11. Phizicky,E.M. and Alfonzo,J.D. (2010) Do all modifications
benefit all tRNAs? FEBS Lett., 584, 265–271.
12. Hopper,A.K., Pai,D.A. and Engelke,D.R. (2010) Cellular
dynamics of tRNAs and their genes. FEBS Lett., 584, 310–317.
13. Murphy,F.V. IV, Ramakrishnan,V., Malkiewicz,A. and Agris,P.F.
(2004) The role of modifications in codon discrimination by
tRNALysUUU. Nat. Struct. Mol. Biol., 11, 1186–1191.
14. Kru ¨ ger,M.K., Pedersen,S., Hagervall,T.G. and Sørensen,M.A.
(1998) The modification of the wobble base of tRNAGlu
modulates the translation rate of glutamic acid codons in vivo.
J. Mol. Biol., 284, 621–631.
15. Kambampati,R. and Lauhon,C.T. (2003) MnmA and IscS are
required for in vitro 2-thiouridine biosynthesis in Escherichia coli.
Biochemistry, 42, 1109–1117.
16. Numata,T., Ikeuchi,Y., Fukai,S., Suzuki,T. and Nureki,O. (2006)
Snapshots of tRNA sulphuration via an adenylated intermediate.
Nature, 442, 419–424.
17. Yim,L., Moukadiri,I., Bjo ¨ rk,G.R. and Armengod,M.-E. (2006)
Further insights into the tRNA modification process controlled
by proteins MnmE and GidA of Escherichia coli.
Nucleic Acids Res., 34, 5892–5905.
18. Moukadiri,I., Prado,S., Piera,J., Vela ´ zquez-Campoy,A.,
Bjo ¨ rk,G.R. and Armengod,M.E. (2009) Evolutionarily conserved
proteins MnmE and GidA catalyze the formation of two
methyluridine derivatives at tRNA wobble positions.
Nucleic Acids Res., 37, 7177–7193.
19. Shi,R., Villarroya,M., Ruiz-Partida,R., Li,Y., Proteau,A.,
Prado,S., Moukadiri,I., Benı´tez-Pa ´ ez,A., Lomas,R., Wagner,J.
et al. (2009) Structure-function analysis of E. coli MnmG (GidA),
a highly-conserved tRNA-modifying enzyme. J. Bacteriol., 191,
20. Hagervall,T.G., Edmonds,C.G., McCloskey,J.A. and Bjo ¨ rk,G.R.
(1987) Transfer RNA(5-methylaminomethyl-2-thiouridine)-
methyltransferase from Escherichia coli K-12 has two enzymatic
activities. J. Biol. Chem., 262, 8488–8495.
21. Bujnicki,J.M., Oudjama,Y., Roovers,M., Owczarek,S., Caillet,J.
and Droogmans,L. (2004) Identification of a bifunctional enzyme
MnmC involved in the biosynthesis of a hypermodified uridine in
the wobble position of tRNA. RNA, 10, 1236–1242.
22. Roovers,M., Oudjama,Y., Kaminska,K.H., Purta,E., Caillet,J.l.,
Droogmans,L. and Bujnicki,J.M. (2008) Sequence–structure–
function analysis of the bifunctional enzyme MnmC that catalyses
the last two steps in the biosynthesis of hypermodified nucleoside
mnm5s2U in tRNA. Proteins, 71, 2076–2085.
23. Sharan,S.K., Thomason,L.C., Kuznetsov,S.G. and Court,D.L.
(2009) Recombineering: a homologous recombination-based
method of genetic engineering. Nat. Protoc., 4, 206–223.
24. Mengel-Jørgensen,J. and Kirpekar,F. (2002) Detection of
pseudouridine and other modifications in tRNA by
cyanoethylation and MALDI mass spectrometry.
Nucleic Acids Res., 30, e135.
25. Sylvers,L.A., Rogers,K.C., Shimizu,M., Ohtsuka,E. and So ¨ ll,D.
(1993) A 2-thiouridine derivative in tRNAGluis a positive
determinant for aminoacylation by Escherichia coli
glutamyl-tRNA synthetase. Biochemistry, 32, 3836–3841.
26. Madore,E., Florentz,C., Giege ´ ,R. and Lapointe,J. (1999)
Magnesium-dependent alternative foldings of active and inactive
Escherichia coli tRNA(Glu) revealed by chemical probing.
Nucleic Acids Res., 27, 3583–3588.
27. Hallidaya,N.M., Hardiea,K.R., Williamsa,P., Winzera,K. and
Barrett,D.A. (2010) Quantitative liquid chromatography–tandem
mass spectrometry profiling of activated methyl cycle metabolites
involved in LuxS-dependent quorum sensing in Escherichia coli.
Anal. Biochem., 403, 20–29.
28. Nakanishi,K., Bonnefond,L., Kimura,S., Suzuki,T., Ishitani,R.
and Nureki,O. (2009) Structural basis for translational fidelity
ensured by transfer RNA lysidine synthetase. Nature, 461,
29. Wang,X., Yan,Q. and Guan,M.-X. (2010) Combination of the
loss of cmnm5U34with the lack of s2U34modifications of
tRNALys, tRNAGlu, and tRNAGlnaltered mitochondrial
biogenesis and respiration. J. Mol. Biol., 395, 1038–1048.
30. Van Lanen,S.G., Kinzie,S.D., Matthieu,S., Link,T., Culp,J. and
Iwata-Reuyl,D. (2003) tRNA modification by
J. Biol. Chem., 278, 10491–10499.
31. Uhlenbeck,O.C. (1995) Keeping RNA happy. RNA, 1, 4–6.
4826Nucleic Acids Research, 2011,Vol.39, No. 11