Design of allele-specific protein methyltransferase inhibitors

Article abstract of DOI:10.1021/ja011423j

Protein arginine methyltransferases, which catalyze the transfer of methyl groups from S-adenosylmethionine (SAM) to arginine side chains in target proteins, regulate transcription, RNA processing, and receptor-mediated signaling. To specifically address the functional role of the individual members of this family, we took a "bump-and-hole" approach and designed a series of N⁶-substituted S-adenosylhomocysteine (SAH) analogues that are targeted toward a yeast protein methyltransferase RMT1. A point mutation was identified (E117G) in Rmt1 that renders the enzyme susceptible to selective inhibition by the SAH analogues. A mass spectrometry based enzymatic assay revealed that two compounds, N⁶-benzyl- and N⁶-naphthylmethyl-SAH, can inhibit the mutant enzyme over the wild-type with the selectivity greater than 20. When the E117G mutation was introduced into the Saccharomyces cerevisiae chromosome, the methylation of Np13p, a known in vivo Rmt1 substrate, could be moderately reduced by N⁶-naphthylmethyl-SAH in the resulting allele. In addition, an N⁶-benzyl-SAM analogue was found to serve as an orthogonal SAM cofactor. This analogue is preferentially utilized by the mutant methyltransferase relative to the wild-type enzyme with a selectivity greater than 67. This specific enzyme/inhibitor and enzyme/substrate design should be applicable to other members of this protein family and facilitate the characterization of protein methyltransferase function in vivo when combined with RNA expression analysis.

Full text of DOI:10.1021/ja011423j

11608
J. Am. Chem. Soc. 2001, 123, 11608-11613
Design of Allele-Specific Protein Methyltransferase Inhibitors
Qing Lin,† Fanyi Jiang,† Peter G. Schultz,*,†,‡ and Nathanael S. Gray*,‡
Contribution from the Department of Chemistry and Skaggs Institute for Chemical Biology, The Scripps
Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, and The Genomic Institute
of the NoVartis Research Foundation, 3115 Merryfield Row, Suite 200, San Diego, California 92121
ReceiVed June 11, 2001
Abstract: Protein arginine methyltransferases, which catalyze the transfer of methyl groups from S-
adenosylmethionine (SAM) to arginine side chains in target proteins, regulate transcription, RNA processing,
and receptor-mediated signaling. To specifically address the functional role of the individual members of this
family, we took a “bump-and-hole” approach and designed a series of N6-substituted S-adenosylhomocysteine
(SAH) analogues that are targeted toward a yeast protein methyltransferase RMT1. A point mutation was
identified (E117G) in Rmt1 that renders the enzyme susceptible to selective inhibition by the SAH analogues.
A mass spectrometry based enzymatic assay revealed that two compounds, N6-benzyl- and N6-naphthylmethyl-
SAH, can inhibit the mutant enzyme over the wild-type with the selectivity greater than 20. When the E117G
mutation was introduced into the Saccharomyces cereVisiae chromosome, the methylation of Npl3p, a known
in vivo Rmt1 substrate, could be moderately reduced by N6-naphthylmethyl-SAH in the resulting allele. In
addition, an N6-benzyl-SAM analogue was found to serve as an orthogonal SAM cofactor. This analogue is
preferentially utilized by the mutant methyltransferase relative to the wild-type enzyme with a selectivity greater
than 67. This specific enzyme/inhibitor and enzyme/substrate design should be applicable to other members
of this protein family and facilitate the characterization of protein methyltransferase function in vivo when
combined with RNA expression analysis.
Introduction
Despite the large number of bioactive compounds that have
been identified from screening synthetic and natural product
collections, it remains challenging to identify potent and
selective inhibitors of a given protein of interest. The problem
of improving inhibitor selectivity is especially acute in large
protein families with highly conserved active sites. One solution
to this problem is through a “bump-and-hole” approach in which
a mutation is introduced to the protein active site that renders
the enzyme (E*) susceptible to selective inhibition by a small
molecule (L*) (Figure 1).1,2 This approach has been successfully
implemented to kinases,3 kinesin and myosin motors,4 nuclear
hormone receptors,5 and human growth hormone receptors.6 The
major advantage of this approach over a traditional genetic
strategy is that greater control over the time and extent of protein
inhibition can be achieved.
Figure 1. A schematic representation of a “bump-and-hole” approach.
E, E*, L, and L* denote wild-type enzyme, engineered enzyme, natural
ligand (inhibitor or substrate), and “bumped” ligand, respectively.
regulation, RNA processing, and receptor-mediated signaling
in eukaryotic organisms.7 Several arginine methyltransferases
have recently received considerable attention including the yeast
Rmt1/Hmt18 and Hsl7,9 the rat PRMT110 and PRMT3,11 and
the human CARM112 and JBP1,13 yet their in vivo functions
are still far from clear. The predominant protein arginine
methyltransferase in S. cereVisiae, RMT1/HMT1 represents an
ideal entry point into the study of this important gene family.
For example, RMT1 is responsible for the synthesis of 85% of
S-Adenosylmethionine (SAM) dependent protein arginine
methyltransferases (PRMTs) are involved in transcriptional
* To whom correspondence should be addressed. E-mail: schultz@
scripps.edu; gray@gnf.org.
† The Scripps Research Institute.
‡ The Genomic Institute of the Novartis Research Foundation.
(1) Hwang, Y.-W.; Miller, D. L. J. Biol. Chem. 1987, 262, 13081-13085.
(2) For a review, see: Alaimo, P. J.; Shogren-Knaak, M. A.; Shokat, K.
M. Curr. Opin. Chem. Biol. 2001, 5, 360-367.
(7) (a) McBride, A. E.; Silver, P. A. Cell 2001, 106, 5-8. (b) Gary, J.
D.; Clarke, S. Prog. Nucleic Acid Res. Mol. Biol. 1998, 23, 89-91. (c)
Aletta, J. Trends Biochem. Sci. 1998, 23, 89-91.
(8) Gary, J. D.; Lin, W.-J.; Yang, M. C.; Herschman, H. R.; Clarke, S.
J. Biol. Chem. 1996, 271, 12585-12594.
(9) Lee, J.-H.; Cook, J. R.; Pollack, B. P.; Kinzy, T. G.; Norris, D.; Pestka,
S. Biochem. Biophy. Res. Commun. 2000, 274, 105-111.
(10) Lin, W.-J.; Gary, J. D.; Yang, M. C.; Clarke, S.; Herschman, H. R.
J. Biol. Chem. 1996, 271, 15034-15044.
(11) Tang, J.; Gary, J. D.; Clarke, S.; Herschman, H. R. J. Biol. Chem.
1998, 273, 16935-16945.
(3) (a) Bishop, A. C.; Ubersax, J. A.; Petsch, D. T.; Matheos, D. P.;
Gray, N. S.; Blethrow, J.; Shimizu, E.; Tsien, J. Z.; Schultz, P. G.; Rose,
M. D.; Wood, J. L.; Morgan, D. O.; Shokat, K. M. Nature 2000, 407, 395-
401. (b) Bishop, A. C.; Shah, K.; Liu, Y.; Witucki, L.; Kung, C.; Shokat,
K. M. Curr. Biol. 1998, 8, 257-166. (c) Liu, Y.; Shah, K.; Yang, F.;
Witucki, L.; Shokat, K. M. Chem. Biol. 1998, 5, 91-101.
(4) Kapoor, T. M.; Mitchison, T. J. Proc. Natl. Acad. Sci. U.S.A. 1999,
96, 9106-9111.
(12) Chen, D.; Ma, H.; Koh, S. S.; Huang, S. M.; Schurter, B. T.; Aswad,
D. W.; Stallcup, M. R. Science 1999, 284, 2174-2177.
(13) Pollack, B. P.; Kotenko, S. V.; He, W.; Izotova, L. S.; Barnoski, B.
L.; Pestka, S. J. Biol. Chem. 1999, 274, 31531-31542.
(5) (a) Ye, H. F.; O’Reilly, K. E.; Koh, J. T. J. Am. Chem. Soc. 2001,
123, 1521-1522. (b) Shi, Y.; Koh, J. T. Chem. Biol. 2001, 12, 243-254.
(6) Guo, Z.; Zhou, D.; Schultz, P. G. Science 2000, 288, 2042-2045.
10.1021/ja011423j CCC: $20.00 © 2001 American Chemical Society
Published on Web 11/01/2001

Design of Allele-Specific Protein Methyltransferase Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11609
Scheme 1
of the post I region, the main chain NH of I310, and a water
molecule, respectively. It is expected that specific enzyme/
substrate or enzyme/inhibitor pairs can be designed by modify-
ing the adenosine-E311 interactions in the adenosine-binding
pocket (Figure 2B). Since this site is remote from the reactive
sulfonium group of the SAM molecule, it should not adversely
affect the methyl-transfer ability of the resulting engineered
enzyme. Thus, a “bump-and-hole” strategy consisting of two
steps was employed in the design of Rmt1 selective inhibitors/
substrates. First, a single amino acid point mutation (E117G,
corresponding to E311G in Prmt3, Figure 2A)17 was made to
Rmt1 to create a “hole” around the N6-position of the adenosine
ring (E* in Figure 1). Second, a bulky substitutant was placed
on the N6-position of the adenosine ring to create a “bump” on
the SAM or SAH ligands (L* in Figure 1). It was anticipated
that this mutant enzyme, but not the wild-type transferase or
its homologues, would accept a “bumped” SAM analogue such
as N6-benzyl-S-adenosylmethionine (16) (Scheme 1) as a
cofactor. As a result, only the substrates of the mutant enzyme
should be methylated in the presence of an isotopically labeled
N6-benzyl SAM analogue. In addition, the mutant enzyme may
be selectively inhibited by the “bumped” SAH analogues in the
presence of the wild-type or its homologous enzymes.
A series of N6-substituted SAH analogues (1-15) were
synthesized18 in three steps starting from 6-chloroadenosine
(Scheme 1). A variety of groups were introduced into the N6-
position of the adenosine ring through the nucleophilic substitu-
tion of chloro by alkyl-, cycloalkyl-, benzyl-, and arylamines.
Screening the SAH analogues with the substituents of different
size, shape, and rigidity should yield a complementary ligand
(L*) for the engineered hole in the active site (E*). Due to the
resulting sulfonium chiral center,19 a diastereomeric mixture of
N6-benzyl-S-adenosylmethionine (16) was synthesized by alky-
lating the sulfur atom of N6-benzyl-S-adenosyl-homocysteine
(12) with excess methyl iodide in the presence of AgClO4. It
should be noted that only one N6-benzyl-SAM diastereomer is
active as the methyl donor in the methyltransferase catalyzed
reactions.20
Figure 2. (A) The sequence alignment of three methyltransferase
fragments where the conserved Glu that make contacts with SAH were
underlined. (B) The crystal structure of the active site of the Prm3-
SAH complex. E311 forms a hydrogen bonding with N6 of SAH.
NG,NG-dimethylarginine found inside the cell. It methylates an
mRNA binding protein Npl3p both in vitro and in vivo, and
the methylation is important for the export of Npl3p from
nucleus.14 The discovery of cell permeable selective substrates/
inhibitors of Rmt1 should provide a useful tool for in vivo
studies of this enzyme, including identification of its downstream
substrates. Here, we present an approach to the discovery of
allele-specific inhibitors of protein arginine methyltransferases
and preliminary functional studies of these inhibitors in Sac-
charomyces cereVisiae in vivo.
Results
Selective Substrate/Inhibitor Design and Synthesis. All
PRMTs utilize S-adenosylmethionine (SAM) as a cofactor in
the methyl-transfer reaction, and share very high homology in
the SAM binding domains (Figure 2A).15 The product of the
methyltransferase reaction, S-adenosylhomocysteine (SAH),
functions as a nonselective inhibitor. Structural studies have
indicated that there are four conserved motifs (designated as I,
post I, II, and III) in the SAM-binding domains of all protein
arginine methyltransferases. The structural module for SAM
binding is formed mainly by motifs I and post I. A crystal
structure16 of a rat methyltransferase PRMT3 catalytic core-
SAH complex has been solved recently (Figure 2B), providing
the first glimpse into the cofactor-binding geometry and the
possible underlying methyl-transfer mechanism. The residues
that interact with SAH are located largely in the three N-terminal
helices of the protein. Furthermore, nitrogens N6, N1, N7 of the
adenosine ring interact with the side chain of E311 in the loop
(14) (a) Henry, M. F.; Silver, P. A. Mol. Cell Biol. 1996, 16, 3668-
3678. (b) Shen, E. C.; Henry, M. F.; Weiss, V. H.; Valentini, S. R.; Silver,
P. A.; Lee, M. S. Genes DeV. 1998, 12, 679-691. (c) McBride, A. E.;
Weiss, V. H.; Kim, H. K.; Hogle, J. M.; Silver, P. A. J. Biol. Chem. 2000,
275, 3128-3136.
(17) The Rmt1 structure was reported recently. The SAM-binding domain
of Rmt1 is virtually superimposable to that of PRMT3. See: Weiss, V. H.;
McBride, A. E.; Soriano, M. A.; Filman, D. J.; Silver, P. A.; Hogle, J. M.
Nat. Struct. Biol. 2000, 7, 1165-1171.
(15) Cheng, X.; Blumenthal, R. M., Eds. S-Adenosylmethionine-depend-
ent methyltransferases: Structures and functions; World Scientific: Sin-
gapore, 1999.
(18) Borchardt, R. T.; Huber, J. A.; Wu, Y. S. J. Org. Chem. 1976, 41,
565-567.
(19) Samejima, K.; Nakazawa, Y.; Matsunaga, I. Chem. Pharm. Bull.
1978, 26, 1480-1485.
(16) Zhang, X.; Zhou, L.; Cheng, X. EMBO J. 2000, 19, 3509-3519.

11610 J. Am. Chem. Soc., Vol. 123, No. 47, 2001
Lin et al.
Table 1. Methyltransferase Activity of Wild-Type and Mutant
Enzyme
Table 2. Inhibition of the Wild-Type and Mutant Rmt1 Enzymes
by SAH Analogues
methyl-
transferase
Km
(µM)
kcat
kcat/Km
Ki (µM)a
cofactor
SAM
(min-1
)
(M-1min-1
)
analogue
WT
1.6
0.6
>100
E117G
selectivityb
WT
E117G
WT
1.1
3.2
>100
18.0
6.7
4.9
<0.02
0.23
6.1 × 106
1.5 × 106
<200
SAH
sinefungin
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
7.2
3.7
350
80
100
53
110
350
530
510
240
50
100
4.4
80
0.22
0.16
Bzl-SAM (16)
>0.29
>1.2
>1.0
>1.9
>0.91
>0.29
>0.19
>0.20
>0.42
>2.0
>1.0
>23
E117G
1.3 × 104
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
Kinetic Studies and Inhibitor Screening. The catalytic
activities of the Glu117fGly mutant and wild-type enzymes were
investigated by using a mass spectrometry based assay with a
synthetic peptide (H-GGFGGRGGFG-NH2) as the substrate.21
The relative rates of methylation for wild-type or mutant enzyme
catalyzed reactions were followed with either SAM or Bzl-SAM
(16) as the cofactor. The corresponding kcat and Km values for
each individual enzyme/substrate pair were derived from curve-
fitting the plots with the Michaelis-Menten equation and listed
in Table 1. The kcat/Km of the E117G mutant was determined
to be 1.5 × 106 M-1 min-1, about 4-fold less active than the
wild-type enzyme. However, when the “bumped” SAM ana-
logue, 16, was used as the cofactor, the methylation reaction
catalyzed by the wild-type enzyme could not be observed.22 In
contrast, the E117G mutant enzyme catalyzed the methyl-
transfer reaction in the presence of analogue 16 with a kcat/Km
only 115-fold lower than that for SAM. This differential catalytic
activity is similar to that found in other orthogonal enzyme-
substrate pairs. For example, in the work on engineering of the
orthogonal kinase/ATP pair the kcat/Km of the engineered Src
(V323A, I338A)/N6-cyclopentyl-ATP pair is about 50-fold lower
than that of the natural Src/ATP pair.23 Thus, the “bumped”
cofactor (16) is a viable substrate for the mutant enzyme.
To determine whether the E117G mutant enzyme could be
selectively inhibited by “bumped” SAH analogues, an in vitro
inhibition assay was carried out for compounds 1-15 in the
presence of either wild-type or mutant enzyme. Two moderately
selective inhibitors, benzyl-SAH (12) and naphthylmethyl-SAH
(15), were identified (Table 2). This suggests that the substit-
uents with a rigid aromatic ring and a flexible methylene linker
may provide the optimum fit for the engineered “hole” in the
active site. Both compounds preferentially inhibit the mutant
enzyme with Ki values of 4.4 and 5.0 µM, respectively, and
selectivity factors greater than 20 relative to the wild-type
enzyme.24 For comparison, two known methyltransferase inhibi-
tors, SAH and sinefungin, preferentially inhibit the wild-type
enzyme over the mutant enzyme with selectivity of 0.22 and
0.16, respectively (Table 2). A radioactive inhibition assay was
also carried out for 15 with [14C]-SAM as the methyl donor
and an in vivo Rmt1 substrate Npl3p in yeast cell extracts as
the methyl acceptor. After separation by gel electrophoresis,
the extent of Npl3p methylation was quantified by the incor-
poration of 14C in its corresponding protein bands on the gel.
Again, analogue 15 was able to preferentially inhibit the E117G
mutant catalyzed reaction over the wild-type with an IC50 of
100 µM (Figure 3).
>1.3
ND
>20
>100
5.0
a Ki values were calculated from Ki ) IC50/(1 + ([S]/Km)); b Se-
lectivity was defined as Ki(wt)/ Ki(mt); ND, not determined.
Figure 3. Inhibition of Npl3p methylation by 15 in the cell extracts.
Cellular Effects of the Selective Inhibitors. To demonstrate
that the N6-substituted SAH analogues can exert their inhibitory
effect on the methylation process catalyzed by Rmt1 in vivo,
we constructed a yeast strain rmt1-as1 in which the wild-type
RMT1 gene on the chromosome was replaced by an RMT1-
E117G mutant gene through a homology recombination tech-
nique.25 Briefly, the RMT1-E117G gene was placed on a URA3-
containing integration vector pRS306. The vector was linearized
and integration was achieved through homologous recombina-
tion and a uracil selection. The removal of the vector from the
chromosome was achieved through self-arrangement and selec-
tion with 5-FOA. The strains that carry the RMT1-E117G alleles
were identified through sequencing of the PCR products of the
genomic DNA.
Treatment of both yeast strains with 10 µM to 1 mM inhibitor
15 had no apparent effect on yeast growth, although the mutant
strain grows a bit slower than the wild-type (Figure 4). This is
consistent with the previous observation that Rmt1 is a
nonessential gene. The RMT1 null mutant has no apparent
defects in either mRNA nuclear export or nuclear protein
localization.14 However, when cells treated with 1 mM 15 were
lysed and the lysates were probed with an antibody that
selectively recognizes the methylated form of an Rmt1 substrate,
Npl3p, the methylation of Npl3p in the mutant allele was found
to be reduced by more than 2-fold26 (lane 8 vs lane 5 in Figure
5). The methylation of Npl3p in the wild-type allele was
essentially unaffected (lanes 1 to 4 in Figure 5). As a control,
the protein expression levels of Npl3p in both wild-type and
(20) Haba, G.; Jamieson, G. A.; Mudd, A. H.; Richard, H. H. J. Am.
Chem. Soc. 1959, 81, 3975-3980.
(21) Kim, S.; Merrill, B. M.; Rajpurohit, R.; Kumar, A.; Stone, K. L.;
Papov, V. V.; Schneiders, J. M.; Szer, W.; Wilson, S. H.; Paik, W. K.;
Williams, K. R. Biochemistry 1997, 36, 3185-3192.
(22) The maximum concentration of 16 tested was 100 µM. The values
for Km and kcat were estimated based on 0.1% methyl-transfer (detection
limit) over the 12 min reaction course.
(23) Shah, K.; Liu, Y.; Deirmengian, C.; Shokat, K. M. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 9565-9570.
(24) The full inhibition of Rmt1 could not be observed at the saturated
concentrations of 12 and 15.
(25) Sikorski, R. S.; Hieter, P. Genetics 1989, 122, 19-27.
(26) The extent of methylation was estimated by the sum of all pixels in
methylation bands (top panel) divided by the sum of all pixels in the
corresponding protein spot (lower panel) using NIH Image software.

Design of Allele-Specific Protein Methyltransferase Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11611
Table 3. Gene List for the Mutant Rmt1-as1 Strain Treated with Compound 12 (|FC| > 2)a
category
gene name
fold change
primary role
cellular communication
cellular organization
CMK2
DHH1
HSP104
YPS6
-2.1
2.1
-2.5
-2.1
calmodulin-dependent protein kinase (intracellular communication)
putative RNA helicase of DEAD/DEAH box family (nuclear organization)
heat shock protein 104 (stress response)
GPI-anchored aspartic protease (extracellular/secretion proteins)
metabolism
GSC2
-3.4
catalytic component of 1,3-â-D-glucan synthase
(C-compound and carbohydrate metabolism)
MET1
MET10
GAC1
-3.1
-2.6
-2.6
siroheme synthase (biosynthesis of vitamins, cofactors, and prosthetic groups)
subunit of assimilatory sulfite reductase (nitrogen and sulfur utilization)
regulatory subunit for Glc7p (regulation of C-compound and
carbohydrate utilization)
MET2
CTR1
AFR1
-2.2
-2.3
-2.1
homoserine O-trans-acetylase (amino acid biosynthesis)
high affinity copper transporter, probable integral membrane protein
coordinates regulation of R-factor receptor signaling and induction
of morphogenesis during conjugation
miscellaneous
PMC1
FRE1
-2
-2
putative vacuolar Ca2 + ATPase, ion transporters
ferric (and cupric) reductase (homeostasis of metal ions)
a Only genes that have been annotated were listed.
to be downregulated by 2.1-fold upon treatment. This finding
suggests that protein phosphorylation might in some way be
related to the protein methylation. Indeed, the phosphorylation
of Ser411 of Npl3p by a kinase Sky1p has been reported
recently, and this phosphorylation regulates nuclear import of
Npl3p by modulating the interaction of RGG motif (the
methylation site) with its nuclear import receptor Mtr10p.28
Further profiling experiments with more potent methyltrans-
ferase inhibitors should give us a clearer picture on the Rmt1-
mediated processes.
Discussion
For allele-specific substrates or inhibitors to be useful in gene
family functional studies, the cross-reactivity between E and
L* should be minimized (diagonal arrow in Figure 1). In the
case of orthogonal substrates, the wild-type enzyme or its
homologues should not accept the “bumped” substrates. At the
same time, the mutant enzyme should preferentially accept the
“bumped” substrate over the natural substrate. As expected, in
our study the wild-type enzyme could not catalyze the methyl-
transfer reaction with the “bumped” SAM analogue 16. The
mutant enzyme could utilize natural SAM (with only 4-fold
reduction in efficiency compared to wild-type enzyme), and
could also utilize 16 with a kcat/Km roughly 450-fold lower than
that of the wild-type enzyme for natural SAM (Table 1). This
reduction of activity may also account for the slower growth
rate observed for the mutant alleles than for the wild-type strain
in the yeast growth experiment (Figure 4), as well as the reduced
methylation of Npl3p (lane 5 vs lane 1 in Figure 5).
The naphthylmethyl-SAH analogue (15) was able to selec-
tively inhibit the methyltransferase activity in enzymatic assays
with a Ki of 5 µM (Table 2). In cell extracts, however, the
inhibition of methylation of Npl3p by 15 was modest with an
IC50 of 100 µM (Figure 3). Furthermore, the inhibitory activity
of 15 was even lower when the mutant alleles were treated with
15 in cell culture and the methylation product was monitored
by western blot (IC50 ∼ 100 µM to 1 mM). Several factors may
contribute to the observed low in vivo efficacy in yeast: (a)
poor cell membrane permeability; (b) poor bioavailabilty; and
(c) a rather long half-life (∼4-5 h) for the substrate Npl3p used
in this study. Similar poor membrane permeability has also been
Figure 4. The yeast growth curve of wild-type or mutant alleles in
the presence of varying concentrations of 15.
Figure 5. In vivo activity of 15 determined by western blot. Yeast
strain BY4743 was used as the WT allele; Yeast strain W303 was used
for the construction of the MT allele, rmt1-as1. The estimated extents
of methylation are given in the bottom plot.
mutant alleles were shown not to alter upon treatment of 15
when a methylation independent anti-Npl3p antibody was used
in the same western blot (lower panel Western in Figure 5).
In preliminary experiments the in vivo effects of the SAH
analogues on the mutant alleles have been evaluated by mRNA
expression profiling experiments with oligonucleotide microar-
ray.27 The rmt1-as1 yeast cells were treated with 500 µM
compound 12, and the cells were harvested after 2 h. The mRNA
expression was subject to gene chip analysis. Among 6200 yeast
genes examined, we found only 30 genes (0.5%) having fold
changes greater than 2.0 relative to untreated cells (Table 3).
In addition to genes related to metabolism and cellular organiza-
tion, a calmodulin-dependent protein kinase, CMK2, was found
(27) Gray, N. S.; Wodicka, L.; Thunnissen, A. M.; Norman, T. C.;
Kwon, S.; Espinoza, F. H.; Morgan, D. O.; Barnes, G.; LeClerc, S.; Meijer,
L.; Kim, S. H.; Lockhart, D. J.; Schultz, P. G. Science 1998, 281, 533-
538.
(28) Yun, C. Y.; Fu, X. D. J. Cell Biol. 2000, 150, 707-718.

11612 J. Am. Chem. Soc., Vol. 123, No. 47, 2001
Lin et al.
observed in the N6-substituted ADP analogues.29 Inhibitor
designs based on alternative scaffolds, such as triazines and
pyrimidines,30 may yield more potent and selective methyl-
transferase inhibitors. Alternatively, the use of yeast lacking key
drug pumps may increase the intracellular levels of inhibitors.
Preliminary profiling experiments with one of the compounds
exemplified the potential utility of this type of selective inhibitor
in the analysis of gene function. The downregulation of CMK2
kinase expression may indicate the existence of a regulatory
mechanism in which the kinase activity and methyltransferase
activity are interdependent. It was reported recently that Npl3p
also undergoes phosphoration catalyzed by Sky1p kinase both
in vitro and in vivo, and this phosphorylation affects Npl3p
nucleocytoplasmic transport.28 More potent methyltransferase
inhibitors and experiments with synchronized yeast should
provide additional insights into the cellular effects of the
inhibitors.
7.31 (m, 4H), 7.21 (m, 1H), 5.90 (d, J ) 2.9 Hz, 1H), 5.52 (s, 1H),
5.44 (s, b, 1H), 4.76 (s, 1H), 4.71(s, 1H), 4.15 (s, 1H), 4.02 (s, 1H),
2.93 (m, 2H), 2.80 (m, 2H), 1.99 (m, 1H), 1.80 (m, 1H); ES-MS calcd
for C21H27N6O5S [M + H]+ m/z 475.2, found 475.2.
N6-Naphthylmethyl-S-adenosylhomocysteine (15): white powder;
1H NMR (400 MHz, DMSO-d6) δ 8.45 (s, b, 1H), 8.38 (s, 1H), 8.25
(m, 4H), 8.01 (t, J ) 8.1 Hz, 1H), 7.95 (dd, J ) 7.4, 2.0 Hz, 1H), 7.82
(d, J ) 9.1 Hz, 1H), 7.68-7.52 (m, 4H), 7.46-7.41 (m, 2H), 5.91 (d,
J ) 5.7 Hz, 1H), 5.14 (s, 1H), 4.78 (s, 1H), 4.55 (m, 1H), 4.16 (m,
1H), 4.04-3.97 (m, 2H), 2.94 (dd, J ) 13.8, 5.8 Hz, 1H), 2.83 (dd, J
) 13.7, 7.5 Hz, 1H), 2.65 (m, 2H), 2.06-1.91 (m, 1H); ES-MS calcd
for C25H29N6O5S [M + H]+ m/z 525.2, found 525.2.
N6-Benzyl-S-adenosylmethionine‚TFA salt (16): white powder;
1H NMR (400 MHz, D2O) δ 8.33 (s, 2H), 7.35 (m, 5H), 6.09 (d, J )
4.0 Hz, 1H), 4.70 (m, 2H), 4.54-4.46 (m, 2H), 3.96-3.77 (m, 4H),
3.62-3.57 (m, 1H), 3.40-3.37(m, 1H), 2.92 (s, 3H), 2.31-2.23 (m,
2H); ES-MS calcd for C22H29N6O5S [M]+ m/z 489.2, found 489.2.
Construction and Expression of GST-Rmt1, GST-Rmt1-E117G.
The Glu117fGly mutant was generated by the QuickChange site-
directed mutagenesis method (Stratagene, La Lolla, CA) with a
pGEX-Rmt1 construct (a gift from Dr. Steve Clarke, UCLA) according
to the manufacturer’s instruction with the following primers:
5′-CCTTGCTAAGAGGCAAGTTGGSGGACGTTCATTTACCCTTTCC-
3′ and 5′-GGAAAGGGTAAATGAACGTCCSCCAACTTGCC-
TCTTAGCAAGG-3′. The resulting mutant plasmid was confirmed
by DNA sequencing. The GST fusion proteins were expressed in
Escherichia Coli BL21cells (Stratagene, La Lolla, CA) upon induction
with a final concentration of 1.0 mM isopropyl-â-D-thiogalactopyra-
noside. Cells were washed and resuspended in 10 mL of lysis buffer
(50 mM K-HEPES, pH 7.4; 250 mM KCl; 1% NP40; 1 mM EDTA;
1 mM DTT) and lysed by four 30 s sonicator pulses (40%) on ice
with a dismembraner (Fisher Scientific). The resulting lysate was
centrifuged at 23 000 g for 10 min at 4 °C. The fusion proteins were
then purified from soluble extracts by binding to glutathione-Sepharose
4B beads (Amersham Pharmacia Biotech) according to the manufac-
turer’s instructions. Proteins were eluted with a 10 mM glutathione
elution buffer (50 mM Tris, 50 mM NaCl, pH 7.5). The protein
concentrations were determined with use of the BioRad protein assay
solution.
The greatest advantage of the “bump-and-hole” approach is
that an entire gene family can be studied in a parallel and
systematic manner. Once a potent small-molecule inhibitor is
identified, it can be quickly adapted to the study of all family
members. For example, a “bumped” version of a very potent
kinase inhibitor, 4-amino-1-tert-butyl-3-(p-methylphenyl)-pyra-
zolo[3,4-d]pyrimidine (PP-1), has been successfully applied in
the functional analysis of a variety of tyrosine kinases.3
Conclusions
Many proteins are methylated in vivo including histone H431
and STAT1.32 The identification of new methyltransferase
substrates has proven important in the elucidation of the role
of protein methylation in various processes. We have demon-
strated that an orthogonal methyltransferase/inhibitor pair can
be designed through the synergistic use of protein engineering
and ligand design. The designed SAH analogue 15 can
selectively inhibit the activity of mutant methyltransferase over
the wild-type enzyme both in vitro and in vivo. Our approach
should be generizable to all members of the protein methyl-
transferse family since all have a conserved Glu residue around
the N6 position of the adenosine ring in the active site. An
orthogonal SAM cofactor (16) was also obtained with the same
approach. This naphthylmethyl-SAM (16) is a selective substrate
for the mutant methyltransferase. We have now begun to use
the selective substrate and inhibitors described here to identify
specific Rmt1 substrates in vivo.
Kinetic Analysis. For kinetic analysis, 5 µL of R1 peptide
(H-GGFGGRGGFG-NH2, 1 mM), 10 µL of 10X assay buffer (25 mM
Tris-HCl, 1 mM sodium EDTA, 1 mM sodium EGTA, pH 7.5), 83.5
µL of H2O, and 1 µL of varying concentration SAM (final 0.2-10
µM) were mixed well and 0.5 µL of Rmt1 solution (0.213 µg) was
added to the mixture. The reactions were incubated at 30 °C for 30
min, and then quenched by adding 40 µL of 1% TFA/H2O. Aliquots
of 10 µL sample solution were injected sequentially into an electrospray
LC-MS instrument (Agilent Technologies). The mass peaks (m/z) of
441.2 ( 0.2 (M - CH3 + H)2+ (product) and 434.2 ( 0.2 (M + H)2+
(starting material) were integrated to give ion intensities. The relative
methylation velocity was defined as eq 1 and measured by following
Experimental Section
All chemicals were purchased from Sigma/Aldrich (St. Louis, MO)
and used without further purification. 6-Chloroadenosine was purchased
from General Intermediates of Canada (Edmonton, Alberta). The SAH
analogues were synthesized according to ref 18 and the SAM analogue
16 was synthesized according to ref 19. All final compounds were
purified by reverse phase preparative HPLC (C18 column, 10-90%
acetonitrile/H2O linear gradient in 30 min, 0.1% TFA added) and
characterized by 400 MHz Bruker NMR and ES-MS.
A
(M-CH3+H)2+
d
(
)
A
+ A(M+H)2+
(M-CH3+H)2+
Vrel
)
(1)
dt
N6-Benzyl-S-adenosylhomocysteine (12): white powder; 1H NMR
(400 MHz, DMSO-d6) δ 8.39 (m, 2H), 8.22 (s, 1H), 7.60 (s, b, 2H),
the product formation over 12 min reaction courses. The reaction
velocities were plotted against the cofactor concentration.
(29) Gillespie, P. G.; Gillespie, S. K. H.; Mercer, J. A.; Shah, K.; Shokat,
K. M. J. Biol. Chem. 1999, 274, 31373-31381.
(30) Hajduk, P. J.; Dinges, J.; Schkeryantz, J. M.; Janowick, D.;
Kaminski, M.; Tufano, M.; Augeri, D. J.; Petros, A.; Nienaber, V.; Zhong,
P.; Hammond, R.; Coen, M.; Beutel, B.; Katz, L.; Fesik, S. W. J. Med.
Chem. 1999, 42, 3852-3859.
(31) Wang, H.; Huang, Z.-Q.; Xia, L.; Erdjument-Bromage, H.; Strahl,
B. D.; Briggs, S. D.; Allis, C. D.; Wong, J.; Tempst, P.; Zhang, Y. Science
2001, 293, 853-857.
(32) Mowen, K. A.; Tang, J.; Zhu, W.; Schurter, B. T.; Shuai, K.;
Herschman, H. R.; David, M. Cell 2001, 104, 731-741.
The values of Vrel,max and K were derived through fitting the curves
with eq 2. The kcat values were derived from eq 3. For inhibition studies,
[S]
V
)
(2)
(3)
Vmax
Km + [S]
Vmax
kcat
)
[E]t

Design of Allele-Specific Protein Methyltransferase Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11613
1 µL of R1 peptide (1 mM), 1 µL of SAM (200 µM), 2 µL of 10X
assay buffer, 14 µL of H2O, and 1 µL of varying concentrations of
inhibitor (100 nM to 10 mM) were mixed well and 1 µL of Rmt1
solution (0.425 µg) was added to the mixture. Methylation was
monitored in the same way as in the kinetic analysis. The IC50 values
for each compound were determined by examining the concentration-
dependent inhibition curves. The Ki was calculated according to eq 4,
Single resistant colonies were picked and further selected on 5-FOA
plates for the presence of a single copy RMT1 gene. The strains that
carry RMT1-E117G alleles were identified by sequencing the PCR
products of the genomic DNA, using primers with identical sequences
to the RMT1 gene up- and downstream regions.
Western Blots. The wild-type (BY4743, Research Genetics, Hunts-
ville, AL) and mutant yeast cells (W303, Mat a, bar1:hisG) were
cultured in YPD media overnight. The cultures were then diluted into
4 portions for each strain (∼5 mL culture each). The cells were treated
with either 50 µL of DMSO or compound 15 at varying concentrations.
The cells were harvested after 5 h, and lysed by using a standard glass
bead method. The supernatants were collected and the protein concen-
trations were determined by using a BioRad protein assay solution.
Proteins (5 µg) were loaded onto a Tris-Glycine gel (12%, Invitrogen)
and separated by SDS-PAGE. The proteins were then transferred onto
a nitrocellulose membrane. The western blot detection was carried out
by using the ECL kit (Amersham Pharmacia Biotech). For the methyl-
specific antibodies, 1:250 ratio was used followed by 1:1000 anti-mouse
HRP secondary antibodies. The antibodies were stripped according to
the ECL instruction. To probe with anti-Npl3p antibodies, a 1:1000
ratio was used followed by 1:1000 anti-rabbit HRP secondary antibod-
ies.
IC50
Ki )
(4)
1 + [S]/Km
where [S] is the concentration of SAM, and Km is the Michaelis-
Menten constant for either wild-type or mutant enzyme.
Cell Extracts Methylation Assay. The Rmt1∆ yeast cells (Research
Genetics, Huntsville, AL) were cultured in 250 mL of YPD media and
harvested at OD ) 2.4. The cells were spun down and the pellets were
collected. The cells were lysed by using a standard glass bead method
(Current Protocol in Molecular Biology 13.13.4). The supernatants were
collected and the protein concentrations were determined by using a
BioRad protein assay solution. Ten microliters of yeast extracts (57
µg) were incubated in 4.0 µL of 14C-SAM (0.1 µCi, Amersham
Pharmacia Biotech), 2.0 µL of Rmt1 (0.85 µg) or mutant enzyme (0.72
µg), 1 µL of varying concentration 15 (final 10-1000 µM) or DMSO
control, and a buffer of 25 mM Tris-HCl, 1 mM sodium EDTA, and
1 mM sodium EGTA at pH 7.5 in a final volume of 25 µL. After 30
min at 30 °C, the reactions were stopped by adding 5 µL of 6X SDS-
gel electrophoresis sample buffers. The samples were then separated
by SDS-gel electrophoresis (12% acrylamide) and the gel fluorographed
after 6 days of exposure at -80 °C. The 14C intensities at the
fluorographic bands of 55 KDa corresponding to Npl3p were integrated
to give rise to the specific methylation extents of the cell extracts.
Construction of Yeast Rmt1-E117G Mutant Allele. The RMT1-
E117G was subcloned into the integration plasmid pRS306 through
the SalI and NotI sites. The plasmid was linearized by using the SwaI
restriction enzyme and transformed into yeast strain W303. The
transformants were plated on a synthetic dropout (SD)-Ura3 plate.
Acknowledgment. We thank Dr. X. Cheng for the coordi-
nates of the PRMT3 structure, Dr. M. Swanson for the methyl-
specific antibody 1E4, Dr. X. Fu for the anti-Npl3p antibody,
and Dr. S. Clarke for the pGEX-Rmt1 plasmid. Q.L. is supported
in part by the Cancer Research Fund of the Damon Runyon-
Walter Winchell Foundation Fellowship, DRG-1612.
Note Added after ASAP: There were errors in Scheme 1
in the version posted ASAP November 1, 2001; the corrected
version was posted November 9, 2001.
Supporting Information Available: Details of kinetics
determination and inhibition assays (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA011423J