Molecular recognition of tyrosinyl adenylate analogues by prokaryotic tyrosyl tRNA synthetases

Article abstract of DOI:10.1016/S0968-0896(99)00192-3

Molecular modelling and synthetic studies have been carried out on tyrosinyl adenylate and analogues to probe the interactions seen in the active site of the X-ray crystal structure of tyrosyl tRNA synthetase from Bacillus stearothermophilus, and to search for new inhibitors of this enzyme. Micromolar and sub-micromolar inhibitors of tyrosyl tRNA synthetases from both B. stearothermophilus and Staphylococcus aureus have been synthesised. The importance of the adenine ring to the binding of tyrosinyl adenylate to the enzyme, and the importance of water-mediated hydrogen bonding interactions, have been highlighted. The inhibition data has been further supported by homology modelling with the S. aureus enzyme, and by ligand docking studies. (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(99)00192-3

Bioorganic & Medicinal Chemistry 7 (1999) 2473±2485
Molecular Recognition of Tyrosinyl Adenylate Analogues by
Prokaryotic Tyrosyl tRNA Synthetases
Pamela Brown,* Christine M. Richardson,* Lucy M. Mensah, Peter J. O'Hanlon,
Neal F. Osborne, Andrew J. Pope and Graham Walker
SmithKline Beecham Pharmaceuticals, New Frontiers Science Park, Third Avenue, Harlow, Essex, CM19 5AW, UK
Received 26 February 1999; accepted 4 June 1999
AbstractÐMolecular modelling and synthetic studies have been carried out on tyrosinyl adenylate and analogues to probe the
interactions seen in the active site of the X-ray crystal structure of tyrosyl tRNA synthetase from Bacillus stearothermophilus, and to
search for new inhibitors of this enzyme. Micromolar and sub-micromolar inhibitors of tyrosyl tRNA synthetases from both B.
stearothermophilus and Staphylococcus aureus have been synthesised. The importance of the adenine ring to the binding of tyrosinyl
adenylate to the enzyme, and the importance of water-mediated hydrogen bonding interactions, have been highlighted. The inhi-
bition data has been further supported by homology modelling with the S. aureus enzyme, and by ligand docking studies. # 1999
Elsevier Science Ltd. All rights reserved.
Introduction
Aminoacyl tRNA synthetase enzymes play a key role in
protein biosynthesis. The enzymes are responsible for
charging a tRNA molecule with its appropriate amino
acid with extremely high ®delity. Any mistakes at this level
lead to the mis-incorporation of amino acids into proteins.
We have shown that the selective inhibition of bacterial
isoleucyl tRNA synthetase by pseudomonic acid and its
synthetic analogues gives rise to potent antibacterial
activity.1 We sought to examine the inhibition of other
synthetases as a source of new potential antibacterial
agents. In this paper we describe the synthesis and
molecular modelling of novel compounds to probe the
molecular interactions seen in the active site of tyrosyl
tRNA synthetase.
Crystallographic information is available on tyrosyl
tRNA sythetase (YRS) from Bacillus stearothermophilus
both in its native form and in a complex with the inhi-
bitor, tyrosinyl adenylate (2) (Brookhaven codes 2ts1
and 3ts1).2 A number of polar interactions can be
observed in the 2.7 A resolution B. stearothermophilus
enzyme±ligand complex, which are shown in Figure 1.
The importance of the key interactions with tyrosine has
been noted elsewhere.3 The aim of the present study was
to investigate three speci®c areas of the adenylate inter-
actions; the phosphate moiety, the role of the ribose
hydroxyl groups, and the speci®city of the adenine binding
The tRNA synthetase enzymes act by a two-step
mechanism. In the ®rst stage, the enzyme recognises the
appropriate amino acid and activates this by reaction
with ATP to produce an aminoacyl adenylate; in this
case, tyrosyl adenylate (Tyr-AMP) (1), which remains
bound to the enzyme. In the second step, the enzyme
catalyses the transfer of the amino acid onto its cognate
tRNA (Scheme 1).
Key words: Amino acids and derivatives; enzyme inhibitors; molecular
modelling; X-ray crystal structures; homology modelling.
* Corresponding authors.
Scheme 1.
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00192-3

2474
P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
region. Molecular modelling and synthetic studies were
undertaken to investigate the relative importance of
these interactions and to probe the active site for other
factors important in molecular recognition. A homology
model of tyrosyl tRNA synthetase from the pathogenic
organism Staphylococcus aureus was also generated and
the inhibitory e€ects of the analogues on the enzymes
from both sources was determined.
more versatile strategy. Thus bis-CBZ tyrosinol was
converted to the cyanoethyl phosphoramidite (7) which
was stable to storage in the freezer under an argon
atmosphere. Reaction of this with the appropriate pro-
tected nucleoside analogues followed by deprotection of
the phosphotriester with methanolic ammonia gave,
after further deprotection, the 20-deoxy (8), 30-deoxy (9)
and uridine (10) analogues of tyrosinyl adenylate, as
shown in Scheme 3.
The cyanoethyl phosphoramidite (7) proved insu-
ciently reactive in some cases however, and the corre-
sponding methyl phosphoramidite was prepared and
used in situ (Scheme 4). This route was used in the
synthesis of analogues designed to probe the require-
ments around the adenine binding region, the unsub-
stituted ribose (11), naphthyl (12) and a,b-unsaturated
ester (13). Also in this sequence, use of the N-Boc pro-
tected tyrosinol (14) a€orded a simpler deprotection
strategy.
Results
Chemistry
Aminoalkyl adenylates have been prepared by con-
densation of the appropriate amino alcohol with AMP.4
However, the recently-described phosphoramidite cou-
pling approach5 appeared to provide the versatility
required for the synthesis of structural variants. Our
initial approach to the synthesis of tyrosinyl adenylate
(2) was to activate the 50 position of the protected ade-
nosine (3) by reaction with diisopropyl phosphonamidic
chloride. The resultant phosphoramidite was highly
reactive, and was coupled directly with bis-CBZ-pro-
tected l-tyrosinol followed by iodine oxidation to give
the protected phosphotriester (4) (Scheme 2). Treatment
with n-butylamine to remove the N6-benzoyl, methyl
ester and O-CBZ groups, followed by hydrogenolysis of
the amine protecting group gave the 20,30-O-iso-
propylidene derivative (5). This was deprotected with
40% aqueous TFA to give tyrosinyl adenylate (2). As a
probe to the binding region around the adenine ring, the
N6-benzoyl group was retained by deprotection of the
phosphomethyl ester with t-butylamine. Hydrogenolysis
followed by removal of the isopropylidine group gave
the derivative (6).
Tyrosyl sulfamoyl adenosine (15) was synthesised by a
procedure analogous to the reported methods6±8
(Scheme 5). As a variant of this, the acyl sulfamide (17)
was prepared from the 50-amino adenosine (18) via the
sulfamide (19) (Scheme 6).
Swern oxidation of the protected adenosine (3) gave the
aldehyde (20) which could be reacted in situ with the
Wittig reagent 21 derived from l-tyrosinol (Scheme 7),
to give after deprotection, the (E)-a,b-unsaturated ester
(22).
Modelling
Tyrosinyl adenylate (2) and analogues (5±6, 8±13, 15, 17
and 22) were modelled in the active site of the B. stear-
othermophilus tyrosyl tRNA synthetase enzyme. This
docking work was performed using the CHARMm
force ®eld and the resulting model complexes were
visualised using Quanta.9
For a more general approach to tyrosinol-derived
phosphodiesters, phosphoramidite activation of the ty-
rosinol moiety, followed by displacement by the 50-hy-
droxy group of the nucleoside analogue appeared a
Figure 1. Polar interactions from tyrosinyl adenylate to residues in the active site taken from the X-ray structure.2

P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
2475
Scheme 2.
Phosphate replacements
Ribose modi®cations
The phosphate unit of tyrosinyl adenylate lies in a solvent
exposed region of the active site (Fig. 2) and appears to
make only one polar interaction directly with the protein.
The crystal structure of the tyrosinyl adenylate±tyrosyl
tRNA synthetase complex shows that the 20 ribose hy-
droxyl makes an interaction to the side chain of Asp 194
and also hydrogen bonds to the backbone NH of Gly
192. The 30 ribose hydroxyl appears to interact with a
highly ordered water molecule, which is itself embedded
in an extended hydrogen bonding network. The 20 and
30 deoxyribose analogues (8) and (9) were modelled in the
active site. 30-Deoxyribose is known to adopt a di€erent
conformation to the 20-deoxy analogue so these com-
pounds were docked using the preferred deoxyribose
conformations taken from the Cambridge Structural
Database.11 However, it was still possible to dock the
30-deoxy variant (9) while maintaining the interaction
between the 20 sugar hydroxyl and the Asp194 side
chain. The isopropylidene derivative of tyrosinyl ade-
nylate (5) appeared too sterically demanding to bind
in the active site in a consistent manner to tyrosinyl
adenylate.
Tyrosyl sulfamoyl adenosine (15) was of interest
because sulfamates have been demonstrated to be
potent inhibitors of the seryl, alanyl and prolyl tRNA
synthetase enzymes,6±8 all of which belong to a di€erent
structural class of tRNA synthetases from that of the
tyrosyl studied here.10,27 Molecular modelling showed this
compound to ®t well within the active site of the tyrosyl
tRNA synthetase enzyme. Docking the sulfamide (17)
showed evidence of some distortion of both protein and
ligand. The a,b-unsaturated ester (22) is a signi®cant
departure from the phosphate unit of tyrosinyl adeny-
late in terms of both charge and rigidity. Compound
(22) docked well in the active site, with minimal strain in
the ligand and no signi®cant deformation of the protein
backbone around the active site.

2476
P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
Scheme 3.
Scheme 4.

P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
2477
Scheme 5.
Scheme 6.
Probing adenine recognition
uracil analogue (10) docked well into the active site,
although the ribose ring appeared in the ¯ipped orien-
tation compared to that in adenosine. Hydrogen bond-
ing to the 20 ribose hydroxyl was preserved and the
uracil group moved out to interact with lysine 228, one
of the lysines present in the mobile KMSKS loop.
Compound (13) contains the a,b-unsaturated ester group
similar to that found in mupirocin, a potent inhibitor of
isoleucyl tRNA synthetase,1 but can only be docked
into the active site at the expense of some localised
deformation of the protein backbone in the adenine
binding pocket.
The adenine ring appears to make no signi®cant polar
interactions in the tyrosinyl adenylate complex. Model-
ling studies suggested that water mediated hydrogen
bonds may be present but no highly ordered water
molecules are observed in structural data on this region
of the active site. The adenine binding pocket appeared
large enough to accommodate the N6-benzoyl group of
(6). Modelling showed the unsubstituted compound (11)
and the 1-naphthyl (12) to be well tolerated, with no
signi®cant movement within the protein backbone. The

2478
P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
Scheme 7.
Homology modelling
tyrosinyl adenylate. Thr51 does not bind to tyrosinyl
adenylate in the complex but is believed to stabilise the
transition state for the aminoacylation reaction.14
The S. aureus tyrosyl tRNA synthetase sequence12
shares 62% identity with that of the B. stear-
othermophilus (Fig. 3).
Biochemical evaluation
Homology modelling and ligand docking studies show
no signi®cant di€erences in either the shape of the active
site or in the polar interactions that can be made
between the enzyme and tyrosinyl adenylate in each
case. From the alignment, and a comparison of residues
which lie within 5 A of the bound ligand in the B.
stearothermophilus crystal structure, it can be seen that
there are only three residues which di€er between the B.
stearothermophilus and S. aureus active sites. These are
Phe37, Ala50 and Thr51, based on the B. stear-
othermophilus numbering. Phe37 is paired with Ile52 in
the Bacillus enzyme and this corresponds to a match
between Ala39 and Phe54 in the Staphylococcus
enzyme. Therefore, the overall active site shape is very
similar. Ala50 lies at the end of the adenine binding
cleft and Thr51 is located above the ribose ring of
Tyrosinyl adenylate (2) and analogues (5±6, 8±13, 15, 17
and 22) were tested as inhibitors of YRS from 2 sources.
The inhibition of YRS from B. stearothermophilus was
determined by examining the e€ect on aminoacylation
of tRNAtyr with 14C-tyrosine. The inhibition of puri®ed
recombinant S. aureus enzyme was determined by a
modi®ed PPi/ATP exchange reaction. The results are
given in Table 1.
Discussion
Tyrosinyl adenylate (2) was found to inhibit the S. aur-
eus enzyme with an IC50 of 11 nM. This is in agreement
with the level of inhibition reported against the Escher-
ichia coli enzyme.15

P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
2479
Figure 2. Tyrosinyl adenylate bound in the active site of the Bacillus stearothermophilus tyrosyl tRNA synthetase. A Connelly representation of the
enzyme surface is shown by the purple dots. The tyrosine binding pocket and the adenine binding cleft are clearly visible.
Phosphate replacements
energetically unfavourable, despite the corresponding
increase in the entropy contribution to the free energy
of the system.
The sulfamate 15 was found to be an extremely potent
inhibitor of both Bacillus and Staphylococcus enzymes.
The compounds containing other phosphate replace-
ments were poor. The modelling would suggest that this
results from a combination of increased rigidity and
unfavourable steric interactions in the case of the sulf-
amide (17). The a,b-unsaturated ester linkage of com-
pound (22) cannot make the polar interaction observed
between the phosphate of tyrosinyl adenylate and the
backbone NH of Asp38.
Adenine variants
Removal of the adenine moiety of tyrosinyl adenylate in
(11) had a substantial impact, con®rming that recognition
of the adenine ring is signi®cant for inhibitor binding.
The naphthyl variant (12) shows improved activity over
the unsubstituted compound, but is still considerably
poorer than tyrosinyl adenylate (2). This suggests that
polar interactions in this region are substantially more
important than is initially apparent from the structural
data. The uridine (10) has good activity and this may be
consistent with the polar interactions seen in the dock-
ing model. The a,b-unsaturated ester 13 showed very
poor activity. This is consistent with the backbone
movement observed in the docking model and shows
that this fragment is not a suitable adenine replacement
within this chemical series.
Ribose analogues
Protection of the ribose moiety with an isopropylidene
derivative produced an inactive compound (5), in good
agreement with the docking studies. The two deoxy-
ribose variants also showed a loss in activity. The 20-
deoxyribose (8) showed a 30-fold reduction in enzyme
inhibition, which correlates well with the loss of a
hydrogen bond between a hydroxyl group and a
charged residue.14 The 60-fold reduction in activity
caused by removal of the 30 hydroxyl group from tyro-
sinyl adenylate was greater than expected from inter-
action of this hydroxyl with a neutral NH group and a
water molecule. However, this water molecule has a
very low temperature factor and forms part of an
extended hydration network in the X-ray structure
(PDB code 3ts1). The 30 hydroxyl of tyrosinyl adenylate
replaces a second highly ordered water which is found in
the structure of YRS with no ligand (PDB code 2ts1)
and forms part of this network. Displacing this tightly
bound water molecule without compensating for the
loss of its enthalpic interaction with the protein may be
Conclusions
We have synthesised a variety of inhibitors of tyrosyl
tRNA synthetase with alterations in the adenine, ribose
and phosphate regions of tyrosinyl adenylate. While we
have found that the level of inhibition of tyrosinyl ade-
nylate can be maintained by replacement of the phos-
phate with the sulfamate moiety, other variants led to
substantially decreased potency. Homology modelling
studies have shown that the S. aureus and B. stear-
othermophilus enzymes are very closely related, sup-
porting the similar trends seen in the inhibition of the

2480
P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
Experimental
Chemistry
Infrared spectra were determined either in dichloro-
methane on a Philips PU 9706 spectrophotometer, or in
KBr on a Perkin±Elmer PE 983 spectrophotometer.
NMR spectra were recorded on a Bruker AC-250F
spectrometer. Chemical shifts are expressed in ppm (d)
relative to internal tetramethylsilane. Mass spectra were
obtained on a VG ZAB mass spectrometer.
All organic phases were dried over anhydrous MgSO4,
and the solvent removed under reduced pressure with a
Buchi rotary evaporator. Merck Kieselgel 60 (<230
mesh ASTM) was used for column chromatography of
intermediates. Tetrahydrofuran (THF) was dried by
distillation from calcium hydride followed by distilla-
tion from sodium benzophenone ketyl.
Target compounds were puri®ed on Mitsubishi Diaion
HP20SS, and the purity was con®rmed by HPLC.
HPLC was performed on a Waters Associates instru-
ment using a C18 m-Bondapak reverse-phase column
with pH 4.5 0.05 M ammonium acetate bu€er-methanol
solutions as eluant. Detection was by UV at 240 nm and
at the lmax of the test compound.
L-Tyrosinyl-20,30-O-isopropylidene adenylate (5). N6-
Benzoyl adenosine16 (820 mg, 2.0 mmol), was dissolved
in acid-free dichloromethane (8 mL), containing diiso-
propylethylamine (1.3 mL, 6.66 mmol). To this was
added N,N-diisopropylmethyl phosphonamidic chloride
(0.58 mL, 3.0 mmol) at room temperature under an
argon atmosphere. After 15 min, the solution was
transferred with 100 mL ethyl acetate (pre-washed with
saturated NaHCO3) into a separating funnel and
washed with saturated NaHCO3. The organic phase was
dried (MgSO4), and evaporated. The crude product was
dissolved in dry acetonitrile and treated with N,O-bis
benzyloxycarbonyl tyrosinol, followed by 1 H-tetrazole
(0.871 g, 10 mmol) at room temperature under an argon
atmosphere. After 15 min, the solution was treated with
a solution of iodine (490 mg, 2.0 mmol) in 1:1 THF:
water (100 mL) together with 2,6-lutidine (2.3 mL,
20 mmol). After 15 min, the solution was diluted with
ethyl acetate (200 mL), washed with 5% aqueous
sodium bisulfate, 5% citric acid, water and brine, dried
(MgSO4), and evaporated. The residue was chromato-
graphed on Kieselgel 60 eluting with 0±5% methanol in
CH2Cl2, to give the protected phosphodiester 4 as a
white foam (0.723 g, 40%), as a 1:1 mixture of diaste-
romers at the phosphorus centre; dP (CDCl3) 14.68,
14.72. m/z (ESI) 923 (MH+, 100%). This material
(380 mg, 0.41 mmol), was dissolved in methanol (10 mL)
and n-butylamine (10 mL), and the reaction stirred at
room temperature for 24 h. The solvent was evaporated
and the residue chromatographed on Kieselgel 60 elut-
ing with 20±40% methanol in CH2Cl2, to give the N6-
deprotected compound as a white foam (220 mg, 81%).
180 mg of this material (0.27 mmol) was dissolved in
50% aqueous dioxan (30 mL), acetic acid (0.15 mL)
added followed by 10% Pd/C, and the mixture hydro-
genated at room temperature and pressure for 1.5 h.
Figure 3. Alignment of the sequences of the Bacillus stearothermophilus
tyrosyl tRNA synthetase enzyme and the S. aureus tyrosyl tRNA syn-
thetase enzyme. The alignment is displayed using Genedoc.13
Table 1. Inhibition of tyrosyl tRNA synthetase (IC50 (mM))
Staphylacoccus
aureus
Bacillus
stearothermophilus
Tyrosinyl adenylate
2
0.011
0.0063
Phosphate replacement
5
17
22
0.026
0% at 100 mM
0% at 100 mM
0.0093
NTa
NT
Ribose variants
Adenine variants
5
8
9
NT
100
140
25% at 100 mM
18
36
6
NT
0.21
420
150
0.68
NT
76
5.7
NT
10
11
12
13
0% at 100 mM
a
NT, not tested.
two enzymes. In general, the modelling results are in
good agreement with the biological data, although polar
interactions do appear to be more important in adenine
recognition than immediately apparent from the crys-
tallographic data. This is at least in part due to limita-
tions imposed by the crystal structure resolution. The
study has also highlighted the importance of the highly
ordered water molecules which are present in the crys-
tallographic data and which assist in de®ning the polar
interactions in the active site.

P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
2481
The mixture was ®ltered, concentrated under reduced
pressure, and the pH adjusted to 7 with saturated
NaHCO3 and the aqueous phase chromatographed on
HP20SS eluting with 0±20% THF in water. Product-
containing fractions were freeze-dried to give 5 as a
white solid (60 mg, 41%); umax (KBr) 3353, 1647 and
1602 cm 1; lmax (H2O) 260 nm (em 12,322); dH (D2O)
1.45 (3H, s, CH3), 1.63 (3H, S, CH3), 2.55 (2H, ABx,
J=7.0, 14.1, 32.5 Hz), 3.38±3.45 (1H, m), 3.55±3.79
(2H, m, OCH2), 4.00±4.05 (2H, m, OCH2), 4.55±4.60
(1H, m), 5.10 (1H, dd, J=2.62, 6.22 Hz), 5.25 (lH, dd,
J=3.0, 6.22 Hz), 6.18 (1H, d, J=3.0 Hz, 10-H), 6.56
and 6.85 (4H, ABq, J=8.5 Hz, Ar-H), 8.10 (1H, s), 8.21
(1H, s); a2d5 13ꢀ (c1, H2O); m/z (ESI) 537 (MH+,
100%).
amine (1.4 mL, 3.3 mmol) followed by 2-cyanoethyl
N,N-diisopropylphosphoramidite (0.73 mL, 3.3 mmol),
and the mixture stirred at room temperature under
argon for 20 min. The solution was diluted with ethyl
acetate (50 mL), washed with saturated NaHCO3 and
brine, dried over anhydrous Na2SO4 and evaporated to
7; m/z (NH3+DCI) 636 (MH+, 2%).
A solution of 7 (1.4 g, 2.2 mmol), in dry acetonitrile
(10 mL) was treated with 1 H tetrazole (154 mg,
2.2 mmol) followed by 30-O-acetyl-N-benzoyl-20-deoxy-
adenosine (100 mg, 0.25 mmol) in dry acetonitrile
(3 mL). After 5 min dichloromethane (5 mL) was added
followed after a further 5 min by 50 mg 1 H tetrazole.
After stirring at room temperature for a further 20 min,
the mixture was oxidised with iodine as described in the
preparation of 2. After work up, the crude product was
dissolved in saturated methanolic ammonia (15 mL),
and the solution stirred at room temperature for 20 h.
The solvent was evaporated and the residue chromato-
graphed on Kieselgel 60 eluting with 10±50% methanol
in dichloromethane to give the protected intermediate as
a white foam (151 mg, 98%); dH (CDCl3) 2.27±2.53 (2H,
m, CH2), 2.55±2.63 (2H, m, CH2), 4.85 (2H, ABq,
J=12.5, 10.8 Hz, CH2Ar), 6.28 (1H, t, J=6.21 Hz), 6.52
and 6.85 (4H, ABq, Ar-H), 7.08±7.15 (5H, m, Ar-H),
8.08 (1H, s), 8.25 (1H, s).
L-Tyrosinyl adenylate (2). Compound
5
(40 mg,
0.075 mmol) was dissolved in 40% aqueous TFA (2 mL)
and stirred at room temperature for 4 h. The solvent
was evaporated and the residue chromatographed on a
column of HP20SS eluting with 2% THF/water. Pro-
duct-containing fractions were freeze-dried to give 2 as a
white solid (33 mg, 89%); umax (KBr) 3354, 1647 and
1610 cm 1; lmax (H2O) 261 nm (em 11,902), dH (D2O)
2.60 (2H, ABx, J=7.0, 14.8, 34.0 Hz), 3.50±3.57 (1H,
m), 3.72±3.81 (2H, m, OCH2), 4.08±4.15 (2H, m,
OCH2), 4.28±4.33 (1H, m), 4.40 (1H, t, J=5.2 Hz), 4.65
(1H, t, J=4.5 Hz), 6.01 (1H, d, J=4.3 Hz, 10-H), 6.56
and 6.86 (4H, ABq, J=7.9 Hz, Ar-H), 8.16 (1H, s), 8.28
(1H, s); m/z (ESI) 497 (MH+, 100%).
This material (145 mg, 0.24 mmol) was hydrogenated as
described in the preparation of 5. After chromato-
graphy and freeze-drying, 8 was obtained as a white
solid (82 mg, 73%); lmax (KBr) 3402, 1652 and 1606
cm 1; lmax (H2O) 260 nm (em 10,795), dH (D2O) 2.43
(2H, Abq, J=13.4, 22.3 Hz), 2.58±2.78 (2H, m, 20H2),
3.14±3.18 (1H, m, 200H), 3.53±3.71 (2H, m, 100H2), 4.08
(2H, br s, 20H2), 4.64±4.68 (1H, m, 30-H), 6.45 (1H, t,
J=6.5 Hz, 10H), 6.62 and 6.86 (4H, ABq, J=8.4 Hz,
Ar-H), 8.20 (1H, s), 8.32 (1H, s); m/z (ESI) 481 (MH+,
70%), 503 (MNa+, 100).
L-Tyrosinyl N6-benzoyl adenylate (6). Compound 4
(100 mg, 0.108 mmol) was dissolved in t-butylamine
(7 mL) and heated to re¯ux for 6 h. The mixture was
evaporated under reduced pressure to a white foam.
This was dissolved in ethanol (10 mL), 10% Pd/C
(50 mg) added and the mixture hydrogenated at room
temperature and pressure. After 3 h, a further 50 mg of
catalyst was added and hydrogenation continued for a
further 2 h. The mixture was ®ltered through Kie-
selguhr, the catalyst washed with water, and the ®ltrated
evaporated to dryness. The residue was chromato-
graphed on Kieselgel 60 eluting with 20±50% methanol
in CH2Cl2, to give the isopropylidene compound as a
white foam (40 mg, 58%). This was dissolved in 40%
aqueous TFA (4 mL) and stirred at room temperature
for 4 h. The solvent was evaporated and the residue
chromatographed on a column of HP20SS eluting with
2±10% THF/water. Product containing fractions were
freeze-dried to give 6 as a white solid (15 mg, 40%); umax
(KBr) 3336, 1702 and 1615 cm 1; lmax (H2O) 281 nm
(em 16,929), dH (D2O) 2.50 (2H, ABx, J=7.3, 14.2,
33.1 Hz), 3.47±3.53 (1H, m), 3.72±3.91 (2H, m, OCH2),
4.08±4.17 (2H, m, OCH2), 4.32±4.38 (1H, m), 4.49 (1H,
t, J=5.2 Hz), 4.76 (1H, t, J=4.8 Hz), 6.16 (1H, d,
J=4.2 Hz, 10-H), 6.50 and 6.82 (4H, ABq, J=8.4 Hz,
Ar-H), 7.50±7.72 (3H, m, Ar-H), 7.90±7.95 (2H, m, Ar-
H), 8.5 (1H, s), 8.68 (1H, s); a2d5= 4ꢀ (c1, H2O); m/z
(ESI) 601 (MH+, 100%).
L-Tyrosinyl 30-deoxy adenylate (9). Protected 20-O-ace-
tyl-N6-benzoyl-30-deoxy adenosine (50 mg, 0.125 mmol)
was coupled with 7, followed by iodine oxidation and
work up as described in the preparation of 2. Treatment
of the crude product with methanolic ammonia as pre-
viously described followed by column chromatography
a€orded the CBZ-protected product in 64% yield. This
was hydrogenated as described for 5 followed by TFA
deprotection. Chromatography on HP20SS gave 9 as a
white freeze-dried solid (24 mg, 77%); umax (KBr) 3386,
1642, and 1609 cm 1; lmax (H2O) 261 nm (em 10,794),
dH (d6-DMSO) 1.91±1.97 and 2.23±2.28 (2H, m, 30H2)
2.50 (2H, ddd, J=6.8, 13.5, 53.8 Hz, 300H), 3.01±3.04
(1H, m, 200H), 3.45±3.61 (2H, m, 100-H), 3.75±3.89 (2H,
m, 50H2), 4.41±4.45 (1H, m, 40H), 4.55±4.58 (1H, m,
20H), 5.70 (1H, br s, OH), 5.90 (1H, d, J=2.3 Hz, 10H),
6.67 and 6.98 (4H, ABq, J=8.4 Hz), 7.23 (2H, br s,
NH), 8.15 (1H, s), 8.42 (1H, s), 9.27 (1H, br s, OH); m/z
(ESI) 481 (MH+, 100%).
L-Tyrosinyl 20-deoxy adenylate (8). To a solution of bis-
CBZ tyrosinol (1.0 g, 2.2 mmol) in acid-free dichloro-
methane (8 mL), was added N-ethyl-N,N-diisopropyl-
L-Tyrosinyl uridine-50-O-phosphate (10). 20,30-Isopropyl-
idineuridine (71 mg, 0.25mmol) was coupled to 7 (0.58 g,
1.0 mmol) followed by iodine oxidation, and deprotection

2482
P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
with methanolic ammonia as described for 8. The crude
product was deprotected with 40% aqueous TFA as
described in the preparation of 2 followed by hydro-
genation to give 10 as a white solid after chromato-
graphy on HP20SS and freeze-drying (35 mg, 30%);
umax (KBr) 3417, 3185, 2924, 1698 and 1684 cm 1; lmax
(H2O) 262 nm (em 7,710), dH (D2O) 2.93 (1H, dd,
J=9.2, 14.0 Hz, CH2Ar), 3.05 (1H, dd, J=6.4, 14.0 Hz,
CH2Ar), 3.78±3.83 (1H, m), 3.90±3.96 (1H, m), 4.01±4.08
(1H, m), 4.09±4.14 (1H, m, 40-H), 4.19±4.23 (1H, m, 50-H),
4.24±4.26 (1H, m, 30-H), 4.27±4.29 (1H, m, 50H), 4.32 (1H,
t, J=4.6Hz, 20H), 5.80 (1H, d, J= 8.1 Hz), 5.90 (1H, d,
J=3.6 Hz, 10-H), 6.87 and 7.21 (4H, ABq, J=8.4 Hz),
7.78 (1H, d, J=8.1 Hz), m/z (ESI) 474 (MH+, 100%).
room temperature for 2 h and then silver oxide (90 mg,
0.4 mmol) was added. The mixture was stirred at room
temperature for a further 1 h and then ®ltered, dried
(MgSO4), evaporated and chromatographed using
methanol±dichloromethane mixtures to give the acet-
onide as an oil (400 mg, 67%). dH (CDCl3) 1.31 and 1.72
(6H, 2s, Me2), 3.8±4.1 (2H, m, 50-H2), 4.2±4.3 (1H, m,
40-H), 4.7±4.9 (2H, m, 20-H and 30-H), 5.62 (I H, d,
J=10.2 Hz, 10-H), 7.5±8.2 (7H, m, aromatics). This
material (180 mg, 0.6 mmol) was converted to the title
compound as described in the preparation of 11 giving a
12 as a white powder (70 mg, 19%). umax (KBr) 1515
cm 1; lmax (H2O) 282 nm (em 8,532); dH (D2O) 2.5±2.8
(2H, m, 3-H2), 3.4±4.0 (9H, m), 6.6 and 6.9 (4H, ABq,
tyrosyl aromatics), 7.4±8.2 (7H, m, naphthyl aromatics);
m/z (ESI+) 490 (MH+).
L-Tyrosinyl 1,4-anhydro-D-ribitol-5-O-phosphate (11).
To a solution containing N-Boc-l-(OZ)tyrosinol 14
(350 mg, 0.9 mmol), triethylamine (1.1 mL, 8.1 mmol)
and dichloromethane (5 mL) at 0ꢀC was added N,N-
diisopropylmethyl phosphonamidic chloride (0.18 mL,
0.9 mmol). After 1 h at room temperature the solution
was evaporated and the residue dissolved in a mixture
of acetonitrile (5 mL) and dichloromethane (5 mL). A
solution of 1,4-anhydro-2,3-O-isopropylidene-d-ribi-
tol18 (50 mg, 0.3 mmol) in dichloromethane (1 mL) was
then added followed by tetrazole (320 mg, 3.6 mmol),
and after 3 h at room temperature by a solution con-
taining iodine (230 mg, 0.9 mmol), 2,6-lutidine (1.0 mL,
9.0 mmol), THF (5 mL) and water (5 mL). After a fur-
ther 15 min the solution was diluted with ethyl acetate
and washed successively with 5% aqueous sodium
sul®te, 5% aqueous citric acid and brine, dried, evapo-
rated, and chromatographed using methanol±dichloro-
methane mixtures to give the fully protected product as
an oil consisting of a 1:1 mixture of diastereomers at
phosphorus (110 mg, 56%). dH (CDCl3) 1.45 (9H, s,
tBu), 1.31 and 1.68 (6H, 2s, CMe2), 2.8±3.0 (2H, m, 3-
H2), 3.8 (3H, 2d, J=7 Hz, OMe), 3.9±5.1 (10H, m), 5.22
(2H, s, ZCH2), 7.1 and 7.3 (4H, ABq, tyrosyl aro-
matics), 7.4±7.5 (5H, m, benzyl); m/z (DCI) 652 (MH+).
This material was deprotected using the following
sequence; the benzyloxycarbonyl group was ®rst
removed by treatment with t-butylamine (2 mL) and
methanol (2 mL) overnight at room temperature fol-
lowed by evaporation. The methyl ester was then
removed by treating the resulting oil with n-butylamine
(2 mL) and methanol (2 mL) for 3 days at room tem-
perature and evaporation. Finally the Boc and acet-
onide groups were removed from this second product by
treatment with tri¯uoroacetic acid (2 mL) and water
(3 mL) for 3 h at room temperature, followed by eva-
poration and chromatography using HP20SS eluting
with acetone±water mixtures to give 11 as its tri-
¯uoroacetate salt after freeze-drying (13 mg, 52%).
umax (KBr) 1680 cm 1; lmax (H2O) 274 nm (em 5,948);
dH (D2O) 2.91 (2H, d, J=8 Hz, 3-H2), 3.6±4.0 (10H, m),
6.9 and 7.2 (4H, ABq, aromatics); m/z (DCI) 364
(MH+).
L-Tyrosinyl 1-ꢀ-(E)-(3-ethoxycarbonyl-2-methylprop-2-
enyl)-1,4-anhydro-D-ribitol-5-O-phosphate (13). A mix-
ture of 5-O-t-butyldimethylsilyl-2,3-O-isopropylidene-b-
d-ribofuranose20 (3.6 g, 12 mmol), acetylmethylene-
triphenylphosphorane (4.2 g, 13 mmol) and acetonitrile
(60 mL) was heated at re¯ux overnight then evaporated
and chromatographed using hexane±ethyl acetate mix-
tures to give 1-(5-O-t-butyldimethylsilyl - 2,3 - O - iso-
propylidene-b-d-ribofuranosyl)-2-propanone as an oil
(2.8 g, 68%); dH (CDCl3) 0.08 (6H, s, SiMe2), 0.91 (9H,
s, tBu), 1.34, 1.53 and 1.57 (9H, 3s, 3ÂMe), 2.7±2.9 (2H,
m, 1-H2), 3.6±3.7 (3H, m, 40-H and 50-H2), 4.1, 4.3 and
4.7 (3H, 3m, 10-H, 20-H and 30-H).
To triethyl phosphonoacetate (4.4 g, 20 mmol) in THF
(50 mL) at 0ꢀC were added sodium hydride (800 mg,
60% in oil, 20 mmol), and, when the solution had
become clear, the above ketone (1.7 g, 5 mmol) in THF
(10 mL). After 60 h at room temperature, aqueous am-
monium chloride±ethyl acetate partition and chromato-
graphy using hexane±ethyl acetate mixtures gave the
TBDMS ether of the title compound as an inseparable
mixture of isomers (1.5 g, 3.6 mmol, 72%). This mixture
was deprotected by dissolving in THF (10 mL) and
adding a solution of tetrabutylammonium ¯uoride in
THF (5 mL, 1 M, 5.0 mmol). After 1 h at room tem-
perature, aqueous ammonium chloride±ethyl acetate
partition followed by careful chromatography using
hexane±ethyl acetate mixtures gave ethyl (E)-4-(2,3-O-
isopropylidene-b-d-ribofuranosyl)-3-methylbut-2-enoate
as an oil (430 mg, 40%). The stereochemistry of this
compound was determined by comparison with pre-
viously reported analogues.21 dH (CDCl3) 1.28 (2H, t,
J=7 Hz, ethyl), 1.35 and 1.68 (6H, 2s, CMe2), 2.24 (3H,
s, CMe), 2.54 (1H, d, J=6 Hz, H-4), 3.63 (2H, m, 5-H2),
4.1±4.2 (4H, m, 10-H, 40-H, and ethyl), 4.67 (2H, s, 20-H
and 30-H), 5.80 (10-H, d, J=1.2 Hz, H-2).
This material (150 mg, 0.5 mmol) was coupled to the
protected tyrosinol as described in the preparation of 11
to give 13 as a white powder (11 mg, 4%). umax (KBr)
1684 cm 1; lmax (H2O) 274 nm (em 1,487); dH (D2O)
1.24 (3H, t, J=7 Hz, ethyl), 2.07 (3H, s, Me), 2.4±2.5
(2H, m, 100-H2), 2.8±3.0 (2H, m, 3-H2), 3.5±4.2 (11H,
m), 5.14 (1H, s, 300-H), 6.8 and 7.2 (4H, ABq, tyrosyl
aromatics); m/z (ESI+) 490 (MH+).
L-Tyrosinyl 1-ꢀ-naphthyl-1,4-anhydro-D-ribitol-5-O-phos-
phate (12). A mixture containing 1-b-d-ribofuranosyl
naphthalene19 (520 mg, 2 mmol), p-toluenesulfonic acid
(34 mg, 0.2 mmol), and acetone (10 mL) was stirred at

P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
2483
N-Boc-L-(OZ)tyrosinol (14). Benzyloxycarbonyl chlo-
ride (2.1 mL, 15 mmol) was added to a mixture of N-
Boc-l-tyrosinol (4.0 g, 15 mmol), triethylamine (2.1 mL,
15 mmol) and dichloromethane (20 mL) at 0ꢀC. After
1 h at room temperature, the solution was washed with
aqueous NaHCO3, dried, evaporated and the resulting
residue recrystallised from 1:1 ether:hexane to give 14 as
white rhombs (3.3 g, 56%). dH (CDCl3) 1.44 (9H, s,
tBu), 2.84 (2H, d, J=7 Hz, 3-H2), 3.4±3.7 (2H, m, 1-H2),
3.0 (1H, br s, 2-H), 4.8 (1H, br d, NH), 5.78 (2H, s,
ZCH2), 7.0±7.5 (9H, m, aromatics).
dissolved in ethanol (10 mL) and hydrogenated over
10% palladium on carbon at atmospheric pressure for
3 h. Filtration, evaporation and chromatography using
methanol±dichloromethane mixtures gave the amine
(18) as a colourless oil (0.4 g, 51%). dH (CDCl3) 1.30
and 1.58 (6H, 2s, CMe2), 3.55 (2H, m, 50-H2), 4.4, 5.2,
5.4 and 6.1 (4H, 4m, 10-H, 20-H, 30-H and 40-H), 7.5 and
8.0 (5H, 2m, benzoyl aromatics), 8.35 and 8.82 (2H, 2s,
adenine aromatics); m/z (DCI) 411 (MH+).
50 -Aminosulfamoyl-50 -deoxy-2,3-O-isopropylidene-N6-
benzoyladenosine (19).
A mixture of 18 (110 mg,
50-O-(L-Tyrosylsulfamoyl) adenosine (15). A solution of
20,30-O-isopropylidene-50-O-sulfamoyladenosine17 (0.772 g,
2.0 mmol) in dry DMF (12 mL) was treated with O-
benzyl-N-Boc-tyrosine N-hydroxysuccinimide ester
(0.936, 2.0 mmol) followed by DBU, (0.3mL, 2.0 mmol),
and the mixture stirred at room temperature for 3 h.
The solvent was evaporated and the residue dissolved in
ethyl acetate and washed with saturated NaHCO3, 5%
KHSO4 and brine, dried (Na2SO4) and evaporated to
an oil. This was chromatographed on Kieselgel 60 elut-
ing with 10% methanol in dichloromethane to a€ord
the protected product (1.13 g, 76%).
0.25 mmol), sulfamoyl chloride (45 mg, 0.38 mmol),
triethylamine (0.04 mL, 0.30 mmol) and dichloro-
methane (2 mL) was stirred at room temperature for 2 h
and directly chromatographed using methanol±dichloro-
methane mixtures to give the sulfamide 19 (80 mg,
65%). dH (CDCl3) 1.40 and 1. 62 (6H, 2s, CMe2), 3.3
(2H, m, 5-H2), 4.4, 5.2, 5.5, and 6.4 (4H, 4m, 10-H, 20-H,
30-H and 40-H), 7.5 and 8.0 (5H, 2m, benzoyl aro-
matics), 8.54 and 8.72 (2H, 2s, adenine aromatics).
N-(L-Tyrosyl)-N0-(50-deoxy-50-adenosinyl)-sulfamide (17).
A mixture of 19 (500 mg, 1 mmol), BocTyr(ONSu)
(380 mg, 1 mmol), DBU (0.15 mL, 1 mmol) and DMSO
(10 mL) was stirred at room temperature for 2 h then
diluted with ethyl acetate. This was washed with water,
dried and evaporated, and the resulting residue was
chromatographed using methanol±dichloromethane
mixtures to give the acylated sulfamide (540 mg, 74%).
dH (CDCl3) 1.38 (9H, s, tBu), 1.41 and 1.58 (6H, 2s,
CMe2), 2.93 (2H, d, J=7 Hz, 3-H2), 3.3±3.5 (2H, m, 50-
H2), 3.9 (1H, m, 2-H), 4.5, 4.9, 5.3 and 6.4 (4H, 4m, 10-
H, 20-H, 30-H, and 40-H), 6.6 and 6.9 (4H, ABq, tyrosyl
aromatics), 7.5 and 8.0 (5H, 2m, benzoyl aromatics),
7.93 and 8.82 (2H, 2s, adenine). This material (0.54 g,
0.75 mmol), was dissolved in a mixture of n-butylamine
(3 mL) and methanol (3 mL) and the solution stirred at
room temperature for 1 h and then evaporated. The
resulting residue was dissolved in a mixture of tri-
¯uoroacetic acid (7 mL) and water (5 mL) and the solu-
tion stirred for 1 h at room temperature, followed by
evaporation and chromatography using HP20SS eluting
with acetone±water mixtures to give the desired com-
pound 17 as its tri¯uoroacetate salt after freeze-drying
(110 mg, 27%). umax (KBr) 1676 cm 1; lmax (H2O) 261
nm (em 12,328); dH (D2O) 2.9±3.1 (2H, m, 3-H2), 3.4 and
3.7 (2H, 2m, 50-H2) 3.9, 4.1, 4.5 and 5.9 (4H, 4m, 10-H,
20-H, 30-H, and 40-H), 4.1 (1H, m, 2-H), 6.6 and 7.0 (4H,
ABq, tyrosyl aromatics), 8.15 and 8.32 (2H, 2s, adenine
aromatics); m/z (DCI) 526 (MNH+4 ).
This material was dissolved in 98% formic acid (50 mL).
After stirring for 2.5 h at room temperature, the solvent
was evaporated and the residue azeotroped with tolu-
ene, followed by chromatography on Kieselgel 60 elut-
ing with 10±20% methanol in dichloromethane to give
the product as a white foam (432 mg, 47%). This mate-
rial (175 mg, 0.29 mmol) was dissolved in 1,4-dioxan
(5 mL), water (5 mL) and hydrogenated at room tem-
perature and pressure for 5 h over 10% Pd/C (50 mg).
The catalyst was removed by ®ltration, the ®ltrate con-
centrated and applied to a column of HP20SS, eluting
with 6% THF/water. Product-containing fractions were
combined and freeze-dried to give the title compound 15
as a white solid (60 mg, 40%); umax (KBr) 3353, 1644,
1613 and 1516 cm 1; lmax (H2O) 260 nm (em 12,134), dH
(d6-DMSO) 2.81 (1H, dd, J=8.0, 14.3 Hz, Ar CH2),
3.04 (1H, dd, J=4.8, 14.3 Hz, CH2 Ar), 3.55 (1H, dd,
J=5.2, 8.4 Hz), 4.06±4.19 (4H, m, 30, 40 and 50H), 4.60±
4.62 (1H, m, 20-H), 5.50 (1H, br s, 20-OH), 5.93 (1H, d,
J=7.8 Hz, 10-H), 7.05 and 6.68 (2Âd, J=8.4 Hz, Ar-H),
7.23 (2H, br s, NH2), 8.15 (1H, s), 8.41 (1H, s), 9.21
(1H, br s, OH). m/z (ESI) 510 (MH+, 100%). (ESI-ve)
502 ([M H] , 2%), 616 ([M.TFA H] , 20%).
50 -Amino-50 -deoxy-20,30 -O-isopropylidene-N6-benzoyl-
adenosine (18). To a solution containing 20,30-O-iso-
propylidene-N6-benzoyladenosine (1.0 g, 2.5 mmol),
triphenylphosphine (1.3 g, 5.0 mmol), sodium azide
(1.7 g, 25.0 mmol), and DMF (20 mL) was added carbon
tetrabromide (1.6 g, 5.0 mmol). After 2 h at room tem-
perature, ethyl acetate was added and the solution
washed with water, dried, evaporated and chromato-
graphed using methanol±dichloromethane mixtures to
give the azide as a yellow oil (0.9 g, 80%). dH (CDCl3)
1.41 and 1.68 (6H, 2s, CMe2), 3.60 (2H, d, J=6 Hz, 50-
H2), 4.4, 5.1, 5.4 and 6.2 (4H, 4m, 10-H, 20-H, 30-H
and 40-H), 7.6 and 8.0 (5H, 2m, benzoyl aromatics), 8.33
and 8.82 (2H, 2s, adenine aromatics). The azide was
2(S)-Amino-3-(4-hydroxyphenyl)propyl 1-(adenin-9-yl)-
(E)-5,6-dideoxy-ꢀ-D-ribohept-5-enofuranuronate hydro-
chloride (22). Compound 14 (802 mg, 2.0 mmol) was
dissolved in dry dichloromethane (5 mL) and treated
with N,N-dimethylaniline (0.254 mL, 2.0 mmol) fol-
lowed by bromoacetyl bromide (0.18 mL, 2.0 mmol), and
the resultant solution stirred at room temperature for 1 h.
The solution was diluted with dichloromethane (20 mL)
and washed with saturated NaHCO3, 2 M HCl, water
and brine, dried (MgSO4) and evaporated. The product
was puri®ed by chromatography on Kieselgel 60 to

2484
P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
a€ord the bromoacetyl ester as a white solid (0.57 g,
1
of B. stearothermophilus YRS was obtained following a
one step puri®cation on a Q Sepharose column.
Recombinant YRS activity was measured by the PPi/
ATP exchange reaction23,24 and the aminoacylation
activity of crude Bacillus YRS was assayed using mod-
i®cations to a previously described method.22
57%). umax (CH2Cl2) 3435, 1760, 1712 and 1501 cm
;
dH (CDCl3) 1.41 (9H, s, tBOC), 2.82±2.90 (2H, m,
PhCH2), 3.87 (2H, s, CH2Br), 4.10±4.19 (3H, m, CH2,
Ha to NH), 4.65 (1H, br s, NH), 5.27 (2H, s, CH2Ar),
7.05±7.45 (9H, m, Ar-H); m/z (NH3DCI) 539 (MNH+4 ,
5%), 192 (100). This was dissolved in 1,4-dioxan and
heated to 50ꢀC with triphenylphosphine (288 mg,
1.1 mmol) for 8 h. The mixture was cooled, evaporated
and azeotroped with toluene to give the crude phos-
phonium bromide. This was dissolved in chloroform
(10 mL) and stirred vigorously with saturated NaHCO3
(10 mL) for 5 min the layers were separated and the
process repeated twice. The organic phases were com-
bined, dried (MgSO4) and evaporated to give the phos-
Aminoacylation reaction. The assays were performed at
37ꢀC in a mixture containing 100 mM Tris/Cl pH 7.9,
50 mM KCl, 16 mM MgCl2, 5 mM ATP, 3 mM DTT,
4 mg/mL E. coli MRE 600 tRNA (Boehringer Man-
nheim) and 20 mM/10 mM l-Tyrosine (4 mM [U-14C] l
Tyr (Amersham, speci®c activity: 16.4 GBq/mmol),
6 mM cold l-Tyr). The concentration of inhibitor which
results in 50% inhibition (IC50) of YRS activity was
determined by preincubating enzyme with a range of
inhibitor concentrations for 10 min at room temperature
followed by the addition of pre-warmed mix for 5 min at
37ꢀC. The reaction was terminated by adding aliquots
of the reaction mix into ice-cold 7% TCA and harvest-
ing onto 0.45 mm hydrophilic Durapore ®lters (Millipore
Multiscreen 96-well plates) and counted by liquid scin-
tillation. The rate of reaction in the experiments were
linear with respect to protein and time with less than
50% tRNA acylation. The data was ®tted by a least
squares procedure using GRAFIT.25
phorane (21) as a white foam, umax (CH2Cl2) 3441,
1
1759, 1702, 1616 and 1506 cm
.
20-30-O-Isopropylidene-N6-benzoyl adenosine (206 mg,
0.5 mmol) was dissolved in toluene (3 mL) and dry
DMSO (1.5 mL) and treated with 1-ethyl-3(3-dimethyl-
aminopropyl)carbodiimide
hydrochloride
(287 mg,
1.5 mmol), pyridine (0.141 mL, 1.75 mmol) and TFA
(0.019 mL, 0.25 mmol) at room temperature under
argon. After 18 h, the crude aldehyde was treated with
the phosphorane (21) (352 mg, 0.5 mmol). After stirring
for 1 h the mixture was partitioned between ethyl ace-
tate and water, the organic layer washed with 2 M HCl
(Â2), saturated NaHCO3, water and brine, dried
(MgSO4) and evaporated. The crude product was chro-
matographed on Kieselgel 60 eluting with 50±100%
ethyl acetate/hexane to a€ord the protected derivative
as a white foam (0.217 g, 52%); umax (CH2Cl2) 3431,
1761, 1713 and 1610 cm 1; dH (D2O) (1H, d, m/z (ESI)
835 (MH+, 100%).
PPi/ATP exchange reaction. Pyrophosphate/ATP
exchange was monitored in the reverse direction at 37ꢀC
in 2 mM [32P]PPi (speci®c activity=ꢁ5Â104 cpm nmol 1),
1 mM l-Tyr, 5 mM ATP, 100 mM Tris/Cl pH 7.9,
50 mM KCl, 12 mM MgCl2 and 2 mM DTT. Reactions
were quenched after 5 min using 0.4 M pyrophosphate
in 7% perchloric acid. The reaction mixture was har-
vested by adsorption onto activated charcoal which was
subsequently collected on Millipore Multiscreen 96-well
plates. Following desorption of [32P]ATP from the
charcoal using ethanolic ammonia, the concentration of
[32P]ATP was determined by liquid scintillation count-
ing. Inhibition assays were carried out as described for
the aminoacylation reaction.
The isopropylidine and BOC groups were removed on
treatment with 40% TFA/water as described for 2, to
a€ord the deprotected product (140 mg, 80%) after
chromatography. This was dissolved in methanol
(2 mL) and treated with t-butylamine (0.5 mL). After
stirring at room temperature for 1 h, the solvent was
evaporated and the residue dissolved in 2M HCl (1 mL).
Saturated NaHCO3 was added to adjust the pH to 3.0
and the solution applied to a column of HP20SS, elut-
ing with 0±20% THF/water. Product-containing frac-
tions were combined and freeze-dried to a€ord the title
compound 22 as a white solid (27 mg, 34%); umax (KBr)
3340, 2925, 1645 and 1515 cm 1; lmax (EtOH) 259 nm
(em 15,823), dH (CD3OD) 3.47 (1H, d, J=6.5 Hz,
CH2Ar), 4.11 (2H, d, CH2O), 4.66±4.78 (1H, m), 5.01
(1H, t, J=5.3 Hz, 30-H), 5.22±5.28 (1H, m, 40-H), 5.40±
5.44 (1H, m, 20-H), 6.71 (1H, d, J=4.5 Hz, 10-H), 6.86
(1H, d, J=15.3 Hz,=H), 7.37 and 7.71 (4H, ABq,
J=8.3 Hz, Ar-H), 7.52 (1H, dd, J=5.5, 15.4 Hz). m/z
(NH3+DCI) 457 (MH+, 30%).
Modelling
The tyrosinyl adenylate ligand was removed from
the B. stearothermophilus tyrosyl tRNA synthetase
crystal structure and the enzyme was prepared by adding
hydrogens using the CHARMm force ®eld.9 Tyrosinyl
adenylate was then minimised in the active site, with
explicit solvation and relaxation of both the protein and
the ligand. The novel ligands were built from the tyrosinyl
adenylate structure and were minimised to convergence in
the B. stearothermophilus active site with explicit salva-
tion and relaxation of both the protein and the ligand.
Only conformations similar to those of the bound
tyrosinyl adenylate were considered. The resulting
complexes were analysed in terms of protein±ligand
interactions, the presence of any deformation within the
protein active site, and the strain energy of the ligand.
Biochemistry
Tyrosyl tRNA synthetase (YRS) inhibition assays. Pure
recombinant S. aureus YRS was puri®ed to near
homogeneity (ꢁ98% as judged by SDS±PAGE) using
standard puri®cation procedures. A crude preparation
In the case of the phosphate replacements, some large
scale conformational sampling of the ligands was carried
out within the constraints of the active site. This

P. Brown et al. / Bioorg. Med. Chem. 7 (1999) 2473±2485
2485
involved Monte Carlo conformational sampling in the
rigid active site in order to generate 100 structures.
Torsion increments of 30ꢀ were used, the temperature
was set to 5000 K and 50 steps of steepest descent
minimisation were performed in each case. The 20
structures of lowest energy were then minimised to
convergence within the active site, using the adopted-
basis Newton±Raphson algorithm. No constraints were
imposed on either the ligand or a 20 A sphere of the
protein around the active site. The monomeric unit of
tyrosyl tRNA synthetase was used throughout and
explicit solvation was employed. The sulfamate and
sulfamide containing compounds were modelled as
zwitterions as this is the form seen in the X-ray crystal
structure of the analogous alanyl sulfamoyl adenosine.7
4. Sandrin, E.; Boissonas, R. A. Helv. Chim. Acta 1966, 24,
273.
5. Desjardins, M.; Garneau, S.; Desgagnes, J.; Lacoste, L.;
Yang, F.; Lapointe, J.; Chenevert, R. Bioorg. Chem. 1998, 26,
1±13.
6. Belrhali, H.; Yaremchuk, A. Science 1994, 263, 1423.
7. Ueda, H.; Shoku, Y. Biochim. Biophys. Acta 1991, 180, 126.
8. Heacock, D.; Forsyth, C. J.; Shiba, K.; Musier-Forsyth, K.
Bioorg. Chem. 1996, 24, 273.
9. Molecular Simulations Inc.; 16, New England Executive
Park, Burlington, MA, 01803-S297.
10. Eriani, G.; Delarue, M.; Poch, O.; Ganglo€, J.; Moras, D.
Nature 1990, 347, 203.
11. Allen, F. H.; Bellard, S.; Brice, M. D.; Cartwright, B. A.;
Doubleday, A.; Higgs, H.; Hummelink, T; Hummelink-Peters,
B. G.; Kennard, O. et al. Acta Crystallogr., Sect. B 1979, B35,
2331.
12. Hodgson, J. E.; Lawlor, E. J. Eur. Pat. Appl., EP 785258
Al.
Sequence alignments were performed using ClustalW.26
Homology modelling was carried out within Quanta,9
with use of the rotamer libraries to optimise side chain
placement. The model was re®ned using the CHARMm
force ®eld9 with a constant dielectric and explicit solva-
tion. Tyrosinyl adenylate was docked into the active
site, both in its crystallographically observed conforma-
tion and after allowing Monte Carlo sampling of the
ligand within the enzyme active site. The results from
the two approaches were in good agreement.
13. Available from http://www.concentric.net/ꢁketchup/gene-
doc/shtml.
14. Wilkinson, A. J.; Fersht, A. R.; Blow, D. M.; Carter, P.;
Winter, G. Nature 1984, 307, 187.
15. Cassio, D.; Lemoine, F.; Waller, J.-P.; Sandrin, E.; Bois-
sonas, R. A. Biochemistry 1967, 6, 827.
16. Chladek, S.; Smrt, J. Coll. Czech. Chem. Commun. 1964,
29, 214.
17. Castro-Pichel, J. et al. Tetrahedron 1987, 43, 383.
18. Aslani-Shotorbani, G. et al. Carbohydrate Research 1985,
136, 37.
19. Ohrui, H. et al. Agr. Biol. Chem. 1972, 36, 1651.
20. Kasher, B. et al. Synthesis 1990, 1031.
21. Mann, J. et al. J. Chem. Soc. Perkin Trans. 1 1988, 433.
22. Calender, R.; Berg, P. Biochemistry 1966, 5, 168.
23. Pope, A. J.; Lapointe, J.; Mensah, L.; Benson, N.; Brown,
M. J; Moore, K. J. J. Biol. Chem. 1998, 273, 31680.
24. Pope, A. J.; Moore, K. J.; McVey, M.; Mensah, L.; Ben-
son, N.; Osborne, N.; Broom, N.; Brown, M. J. J. Biol. Chem.
1998, 273, 31691.
25. Leatherbarrow, R. J. Gra®t Version 3.0, Erithacus Soft-
ware Ltd, Staines, UK.
26. Thompson, J. D.; Higgins, D. G.; Gibron, T. J. Nucleic
Acids Research 1994, 22, 4673.
Acknowledgements
The authors thank Mary McVey and Paul Carter for
protein puri®cation.
References and Notes
1. Walker, G.; Brown, P.; Forrest, A. K.; O'Hanlon, P.; Pons,
J. E. In Recent Advances in the Chemistry of Anti-infective
Agents; Royal Society of Chemistry, London, 1993; p. 106.
2. Brick, P.; Bhat, T. N.; Blow, D. M. J. Mol. Biol. 1989, 208,
83.
3. Fersht, A. R.; Shindler, J. S.; Tsui, W.-C. Biochemistry
1980, 19, 24.
27. Arnez, J. G.; Moras, D. Trends Biochem. Sci. 1997, 22,
189.