Identification of novel inhibitors of bacterial surface enzyme Staphylococcus aureus Sortase A

Article abstract of DOI:10.1016/j.bmcl.2007.10.051

In-silico virtual screening of bacterial surface enzyme Staphylococcus aureus Sortase A against commercial compound libraries using FlexX software package has led to the identification of novel inhibitors. Inhibition of enzyme catalytic activity was deter

Full text of DOI:10.1016/j.bmcl.2007.10.051

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 380–385
Identification of novel inhibitors of bacterial surface enzyme
Staphylococcus aureus Sortase A
Bala Chandra Chenna,a Bidhan A. Shinkre,a Jason R. King,a Aaron L. Lucius,a
Sthanam V. L. Narayanab and Sadanandan E. Velua,*
aDepartment of Chemistry, University of Alabama at Birmingham, 901 14th Street South, Birmingham, AL 35294, USA
bCenter for Biophysical Sciences and Engineering, School of Optometry, University of Alabama at Birmingham,
1025 18th Street South, Birmingham, AL 35294, USA
Received 12 September 2007; revised 12 October 2007; accepted 15 October 2007
Available online 18 October 2007
Abstract—In-silico virtual screening of bacterial surface enzyme Staphylococcus aureus Sortase A against commercial compound
libraries using FlexX software package has led to the identification of novel inhibitors. Inhibition of enzyme catalytic activity
was determined by monitoring the steady state cleavage of a model peptide substrate. Preliminary structure activity relationship
studies on the lead compound resulted in the identification of compounds with improved activity. The most active compound
has an IC50 value of 58 lM against the enzyme.
Ó 2007 Elsevier Ltd. All rights reserved.
Many Gram-positive bacteria are human pathogens.
Staphylococci are responsible for more than one million
hospital-acquired bacterial infections every year. Com-
pounding their importance is the recent, steep increase
in multi-drug resistance found in bacteria and the poten-
tial of ‘drug-resistant bacteria’ to be used in biowarfare
and bioterrorism.1,2 Hospital strains of Staphylococcus
aureus are usually resistant to a variety of different anti-
biotics. A few strains are resistant to all clinically useful
antibiotics except vancomycin, and vancomycin-resis-
tant strains are increasingly reported.3–5 Methicillin-
resistant S. aureus (MRSA) is widespread, and most
methicillin-resistant bacterial strains are also resistant
against multiple antibiotics.6–8 This necessitates the need
for identifying new therapeutic agents, which do not
promote the emergence of drug-resistant microbes.
sophisticated arsenal of virulence factors, many of which
are surface-bound or secreted. Gram-positive bacteria
such as staphylococci are endowed with a multitude of
cell-wall anchored proteins that serve as an interface be-
tween the microbe and its host. Bacterial sortases are
cysteine transpeptidases that participate in secretion
and anchoring of many cell-wall proteins by a mecha-
nism conserved in almost the entire class of Gram-posi-
tive bacteria. Sortases, because of their control over the
cellular location of multiple virulence factors, are an
attractive potential target for interrupting virulence ver-
sus bacterial growth.
Surface proteins can be attached to the bacterial surface
in one of several fashions.9,10 Proteins that are cova-
lently attached to the cell-wall share conserved regions
known as the ‘sorting signal’ or cell-wall anchors.11,12
The sorting signal includes a conserved amino acid mo-
tif, usually LPXTG. Precursor proteins are directed into
a secretory pathway by their N-terminal signal peptides.
They are translocated across the membrane and the sig-
nal peptide is cleaved.12,13 Then, the C-terminal sorting
signal retains the protein in the secretory pathway. The
enzyme sortase acts at this point to cleave the protein
between the threonine (T) and the glycine (G) of the
LPXTG motif.14 The carboxyl group of the Thr is then
amide-linked to the amino group of a ‘cross-bridge’ pep-
tide in the lipid II precursor for cell-wall synthesis.15
New approaches for the prevention and treatment of
bacterial infections require greater understanding of
the molecular structure and mechanisms of the chosen
intervention targets and of the pathogenic role played
by the target in the infection process. Bacterial infections
are very complex and involve the action of a large,
Keywords: Staphylococcus aureus; Sortase A; Inhibitors; Virtual
screening.
*
Corresponding author. Tel.: +1 205 975 2478; fax: +1 205 934
2543; e-mail: svelu@uab.edu
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.10.051

B. C. Chenna et al. / Bioorg. Med. Chem. Lett. 18 (2008) 380–385
381
Sortase-defective strains of various pathogens were
O
C
N
shown to be faulty in the display of surface proteins
and are less virulent.16,17 In a number of studies, individ-
ual sortase genes have been deleted and the loss of sor-
tase function resulted in less virulence in several animal
models of the disease.16,18–22
S
O
A
B
OH
N
O
H
Due to their universal presence in Gram-positive bacte-
ria, sortases are attractive pharmacotherapeutic tar-
gets.10 Currently there have only been a few reports of
specific sortase inhibitors.23–26 Recently, Oh et al.27
identified a small-molecule reversible inhibitor with a
low micromolar IC50 value by structurally modifying
a lead compound identified by random screening of a
group of small molecules. We have recently determined
Inhibitor 1, IC50 = 75 4.1 μM
Figure 1. Structure and IC50 value of inhibitor 1.
A FlexX docking model of inhibitor 1 revealed several
critical interactions of the inhibitor within the active site
(Fig. 2).
the high resolution crystal structure of recombinant S.
28
The morpholine ring oxygen atom has a hydrogen bond-
ing interaction with amide backbone NH of Trp194.
aureus SrtAD59
,
which is a fully active variant of SrtA
with 59 residues shorter from the NH-terminus. Utiliz-
ing our SrtAD59 crystal structure, we initiated structure
based inhibitor design. This manuscript describes the
discovery of micromolar inhibitors of Sortase A identi-
fied by in-silico virtual screening and preliminary struc-
ture activity relationship (SAR) studies.
˚
The cysteine residue (Cys184) is at 2.94 A distance from
the morpholine ring carbon atom next to the oxygen.
The carboxylic acid on the middle phenyl ring has a di-
rect salt bridge interaction with the guanidine side chain
of Arg197. The amide NH group of the linker has a
hydrogen bonding interaction with the carboxylic acid
side chain of Glu105, and the amide carbonyl oxygen
of the linker has a hydrogen bonding interaction with
the OH side chain of Ser116.
We have conducted virtual screening of commercial
small-molecule libraries from Maybridge and Chem-
Bridge (ꢀ150,000 compounds) against S. aureus SrtAD59
active site using the FlexX software package integrated in
SYBYL 7.0.29–31 FlexX is a fast, flexible docking pro-
gram that uses an incremental construction algorithm
to place ligands into the active site of the receptor. In
FlexX docking, the enzyme active site environment is
identified by a receptor description file (RDF). We cre-
ated the RDF for SrtAD59 using the chain A of SrtAD59
PDB file, the ‘active site file’ which is constituted of all res-
The thiophene ring (ring A) of inhibitor 1 is in close
˚
˚
proximity to Gln172 (3.2 A) and Asn114 (3.1 A).
Appropriate substitution to the thiophene ring may re-
sult in additional hydrogen bonding interactions with
Gln172 and Asn114. In addition, a wide range of hydro-
phobic residues (Val201, Phe200, Ile199, Leu169, Ile158,
Val168, Val164, Leu179, Ile115, Leu181, Ile182, Ile117,
Leu97, and Phe103) surround the middle phenyl ring B
on either side.
´
˚
idues within 6.5 A distance from the center of the active
site and a ‘pocket file’ which is constituted of a few
hand-picked residues that are directly interacting with
the substrate peptide. 2D compound data sets from com-
mercial sources were transformed into three-dimensional
molecular structures using the Concord program. All
compounds were generated in the protonated state that
can be assumed under physiological conditions. The fol-
lowing limits were applied for the virtual screening:
Molecular weight >150 <500, ClogP < 5. In addition,
compounds bearing metal ions were omitted from the
data set. The structures of top scoring compounds were
manually examined for their drug like properties based
on Lipinski’s rule32 and for synthetic versatility. 108 best
scoring compounds were purchased and screened against
SrtAD59 using a previously reported fluorescence reso-
nance energy transfer (FRET) enzymatic assay.27,33,34
In this assay, the IC50 values were determined by moni-
toring the effect of the compounds on the steady state
cleavage of a model substrate peptide, Dabcyl-QALPET-
GEE-EDANS. Eight out of 108 compounds exhibited
in vitro inhibition catalytic activity of SrtAD59 with IC50
values ranging from 75 to 400 lM. The most active com-
pound identified from this virtual screening is compound
1 (Fig. 1), which had an IC50 value of 75 4.1 lM.
Inhibitor 1 was initially obtained from a commercial
source in milligram quantities. For SAR studies we
needed this compound in larger quantities. So, a general
method for the synthesis of inhibitor 1 was developed in
our laboratory. The synthetic procedures for all analogs
were modeled around this synthesis. Chemistry em-
ployed for the synthesis of inhibitor 1 is outlined in
Scheme 1.
Commercially available methyl 2-(N-morpholino)-5-
nitrobenzoate (9) was reduced using hydrogen in the
presence of Pd/C in anhydrous ethyl acetate to afford
the corresponding amino compound 10. The compound
10 was coupled with commercially available trans-3-(thi-
ophene-2-yl)acrylic acid (11a) in the presence of
ethyl(N,N-dimethylaminopropyl) carbodiimide (EDAC)
and N,N-dimethylaminopyridine (DMAP) in 1,2-dichlo-
roethane to form the amide compound 12a.35 Basic
hydrolysis of the ester methyl group present in com-
pound 12a afforded the inhibitor 1 as a white crystalline
solid.
We have synthesized a few derivatives of inhibitor 1 in
order to derive preliminary structure activity relation-
ship data. Their activities against SrtAD59 were deter-
IC50 values of compounds 2–8 and their structures are
given in Table 1.

382
B. C. Chenna et al. / Bioorg. Med. Chem. Lett. 18 (2008) 380–385
Table 1. S. aureus SrtAD59 inhibitors identified from the FlexX virtual screening and their IC50 values
a
Compound
Structure
IC50 (lM)
O
O
CN
2
97.3 7.6
150 4.8
226 16.7
275 11.8
400 6.5
400 41.1
125 3.6
S
S
S
N
O2N
NO2
NO2
O
O
CH3
O
S
3
4
5
6
7
8
N
CN
O
N
H
H
O
N
N
OH
O
S
O
O
N
CH3
H
N
NH2
NH2
N
N
N
NH2
H2N
O
O
CH3
S
S
NO2
N
CN
O
N
O
NH
N
NH
N
H2N
Cl
O
HN
O
OH
NH
O
OH
a IC50 values were determined by fluorescence resonance energy transfer (FRET) assay. Measurements were carried out in triplicate and an average
value with standard deviation is reported.
COOCH3
COOCH3
H2,
O2N
N
O
H2N
N
O
Pd/C
EtOAc
10
9
O
O
N
X
X
OH
11b, X = O
11a, X = S
COOCH3
EDAC, DMAP
ClCH2CH2Cl
N
H
12a, X = S
12b, X = O
O
O
N
Figure 2. FlexX docking model showing the interactions of inhibitor 1
with amino acid residues in SrtAD59 active site.
X
COOH
NaOH
CH3OH
THF
N
H
1, X = S
13, X = O
mined using the FRET assay. Structures and IC50 values
of the newly synthesized derivatives of inhibitor 1 are gi-
ven in Table 2.
O
Scheme 1. Synthesis of inhibitor 1 and its furan analog 13.
Compound 13 is a furan analog of inhibitor 1. Design of
this structure is inspired by the comparison of activities
of thiophene and furan compounds 3 and 6 (Table 1).
These two compounds have identical structures except
for the fact that one contained a furan ring and the other
contained a thiophene ring. Furan compound, 3, is more
than twofold more active than the thiophene compound,

B. C. Chenna et al. / Bioorg. Med. Chem. Lett. 18 (2008) 380–385
Table 2. S. aureus SrtAD59 inhibition of synthesized derivatives of inhibitor 1
383
O
N
X
R
A B
N
O
H
No.
X
A–B
R
IC50 (lM)
12a
12b
13
S
CH@CH
CH@CH
CH@CH
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH@CH
CH@CH
CH@CH
CH@CH
CH@CH
COOCH3
COOCH3
COOH
71 2.5
58 4.9
181 14
>600
O
O
S
14
15
COOH
COOCH3
COOH
S
>600
>600
16
17
O
O
S
COOCH3
CH2OH
CH2OH
CHO
>600
73
18a
18b
19a
19b
20
2
O
S
111 5.7
77 6.5
O
S
CHO
CONH2
107 8.5
105 3.7
6. So, we expected that a furan substitution of the thio-
phene ring of inhibitor 1 may increase its activity. Com-
pound 13 was synthesized by a similar procedure as
described for the synthesis of inhibitor 1 (Scheme 1).
We have used trans-3-(furan-2-yl) acrylic acid (11b) in-
stead of trans-3-(thiophene-2-yl) acrylic acid (11a) in this
procedure. Compound 10 was coupled with commer-
cially available trans-3-(furan-2-yl) acrylic acid (11b) in
the presence of EDAC and DMAP in 1,2-dichloroeth-
ane to form the amide compound 12b. Basic hydrolysis
of the ester methyl group present in compound 12b
afforded compound 13. Compound 13 along with the
synthetic intermediate methyl esters 12a and 12b were
evaluated for their enzyme inhibitory activity (Table
2). Compound 13 did not show an enhancement of
activity. Instead, it was found to be less active compared
to inhibitor 1 in the enzymatic assay (IC50 = 181 lM).
But, the methyl ester intermediates (12a and 12b)
showed improved inhibition as compared to the corre-
sponding acid derivatives 1 and 13. The methyl ester
derivative of the thiophene compound (12a) showed an
IC50 value of 71 lM and methyl ester derivative of the
furan compound (12b) showed a much improved inhibi-
tion at an IC50 value of 58 lM. Although this increase in
activity of methyl ester analogs is counter intuitive of the
FlexX model which shows a possible salt bridge interac-
tion between the carboxylic acid and Arg197 side chain,
the increased activity may be explained using the FlexX
model of the methyl ester in the active site. This model
shows a similar fit of the molecule in the active site with
the ester methyl group resting in a small-hydrophobic
pocket alongside the central phenyl ring without com-
pletely destroying the electrostatic interaction between
Arg197 side chain and ester group O atoms of the inhib-
itor. The increased activity could be a result of these
additional hydrophobic interactions.
1 and 13 for their activity. Compounds 14 and 16 are
the saturated analogs of the carboxylic acid derivatives
1 and 13. Compounds 15 and 17 are the saturated ana-
logs of the methyl ester derivatives 12a and 12b. Our as-
say showed that all four saturated compounds 14–17
were inactive up to a concentration of 600 lM, showing
that the double bond is necessary for the activity of these
inhibitors (Table 2). Synthesis of compounds 14–17 is
outlined in Scheme 2. Hydrogenation of the thiophene
derivatives 1 and 12a using Pd/C as catalyst did not
work, possibly due to the catalyst poisoning effect of
the thiophene ring present in these molecules. Use of a
stronger catalyst like Pd black in the presence of ammo-
nium formate resulted in the hydrogenation of 1 and 12a
to corresponding hydrogenated compounds 14 and 15.
Hydrogenation of 13 and 12b was carried out under
H2 in the presence of Pd/C catalyst to afford compounds
16 and 17.
We have also made a few derivatives of inhibitor 1 by
incorporating different substituents such as a –CH2OH
O
O
H2 or
HCOONH4
N
N
X
O
X
O
Catalyst/
Y Solvent
O
O
Y
N
H
N
H
O
O
Catalyst
/Solvent
Yield
(%)
Reactant
X
Y
Product
1
12a
13
12b
S
S
O
O
H
CH3
H
CH3
Pd black/MeOH
Pd black/MeOH
Pd/C/EtOAc
14
15
16
17
69
82
88
Pd/C/EtOAc
77
Compounds 14–17 were synthesized to evaluate the
importance of the double bond present in compounds
Scheme 2. Synthesis of compounds 14–17.

384
B. C. Chenna et al. / Bioorg. Med. Chem. Lett. 18 (2008) 380–385
2 orders of magnitude tighter than the LPXTG peptide
and thus should be effective at blocking the enzymes’
activity in vivo.
O
O
N
N
X
X
O
DIBAL
In conclusion, we have discovered a novel class of
small-molecule inhibitors of S. aureus SrtA using in-
silico virtual screening. We utilized the software pack-
age FlexX integrated in SYBYL 7.0 to carry out the
virtual screening of commercial compound libraries
against the SrtA active site. Inhibitors were screened
for their activity against SrtA using a previously re-
ported FRET assay. Micromolar inhibitors of the en-
zyme are identified. We have carried out preliminary
structure activity relationship studies that have re-
sulted in the identification of inhibitors with improved
activity. Further SAR studies and attempts to obtain
high resolution inhibitor/SrtA complex co-crystal
structures are currently in progress.
THF
OH
OMe
CH2Cl2
N
H
N
H
O
O
12a, X = S
12b, X = O
18a, X = S
18b, X = O
O
N
X
O
H
PCC
THF
N
H
O
19a, X = S
19b, X = O
Scheme 3. Synthesis of compounds 18a,b and 19a,b.
Acknowledgments
Authors acknowledge the generous financial support
from the Department of Chemistry, University of Ala-
bama at Birmingham, and a faculty development grant
from UAB-FDGP to S.E.V.
O
O
N
N
S
O
S
O
1. SOCl2
2. NH3
NH2
OH
N
H
N
H
References and notes
20
O
O
1
1. Pappas, G.; Panagopoulou, P.; Christou, L.; Akritidis, N.
Infect. Dis. Clin. North Am. 2006, 20, 395.
2. Clarke, S. C. Br. J. Biomed. Sci. 2005, 62, 40.
3. Mohr, J. F.; Murray, B. E. Clin. Infect. Dis. 2007, 44,
1536.
4. Webster, D.; Rennie, R. P.; Brosnikoff, C. L.; Chui,
L.; Brown, C. Diagn. Microbiol. Infect. Dis. 2007, 57,
177.
5. Weigel, L. M.; Donlan, R. M.; Shin, D. H.; Jensen, B.;
Clark, N. C.; McDougal, L. K.; Zhu, W.; Musser, K. A.;
Thompson, J.; Kohlerschmidt, D.; Dumas, N.; Limberger,
R. J.; Patel, J. B. Antimicrob. Agents Chemother. 2007, 51,
231.
6. Baggett, H. C.; Hennessy, T. W.; Rudolph, K.; Bruden,
D.; Reasonover, A.; Parkinson, A.; Sparks, R.; Donlan,
R. M.; Martinez, P.; Mongkolrattanothai, K.; Butler, J. C.
J. Infect. Dis. 2004, 189, 1565.
Scheme 4. Synthesis of compound 20.
(18a), –CHO (19a), or –CONH2 (20) in the place of the
carboxylic acid group. Substitution with –CH2OH and –
CHO groups did not result in a major change in the
activity (18a, IC50 = 73 lM and 19a, IC50 = 77 lM),
while substitution with –CONH2 group resulted in a de-
crease in activity as compared to inhibitor 1 (20,
IC50 = 105 lM) (Table 2). Derivatives of Furan com-
pound 13 incorporating substituents such as a –CH2OH
(18b) and –CHO (19b) in the place of the carboxylic acid
group were also made. These compounds showed
improved inhibition as compared to the parent furan
compound, 13 (Table 2). Compound 18b showed an
IC50 value of 111 lM, and compound 19b showed an
IC50 value of 107 lM. Synthesis of compounds 18a,b
and 19a,b is outlined in Scheme 3.
7. Deresinski, S. Clin. Infect. Dis. 2005, 40, 562.
8. Salerno, D.; Vahid, B.; Marik, P. E. Ann. Thorac. Surg.
2007, 83, 1888.
9. Navarre, W. W.; Schneewind, O. Microbiol. Mol. Biol.
Rev. 1999, 63, 174.
10. Cossart, P.; Jonquieres, R. Proc. Natl. Acad. Sci. U.S.A.
2000, 97, 5013.
11. Fischetti, V. A.; Pancholi, V.; Schneewind, O. Mol.
Microbiol. 1990, 4, 1603.
12. Schneewind, O.; Model, P.; Fischetti, V. A. Cell 1992, 70,
267.
13. Schneewind, O.; Mihaylova-Petkov, D.; Model, P. EMBO
J. 1993, 12, 4803.
14. Navarre, W. W.; Schneewind, O. Mol. Microbiol. 1994, 14,
115.
Compounds 12a,b were reduced using DIBAL in a mix-
ture of anhydrous dichloromethane and THF to afford
the alcohol derivatives 18a,b. Oxidation of alcohols
using PCC in anhydrous THF afforded the aldehydes,
19a,b. Synthesis of amide 20 is outlined in Scheme 4.
Compound 20 was prepared from the inhibitor 1 by
treatment with SOCl2, followed by the treatment of
the acid chloride produced with ammonia.
Strikingly, the IC50 values determined for all of the ac-
tive compounds are well below the previously measured
Km value of 5.5 mM for SrtA binding to the LPXTG
peptide.33 This suggests that these inhibitors bind ꢀ1–
15. Perry, A. M.; Ton-That, H.; Mazmanian, S. K.; Schnee-
wind, O. J. Biol. Chem. 2002, 277, 16241.
16. Jonsson, I. M.; Mazmanian, S. K.; Schneewind, O.;
Bremell, T.; Tarkowski, A. Microbes Infect. 2003, 5, 775.

B. C. Chenna et al. / Bioorg. Med. Chem. Lett. 18 (2008) 380–385
385
17. Mazmanian, S. K.; Liu, G.; Jensen, E. R.; Lenoy, E.;
Schneewind, O. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
5510.
18. Barnett, T. C.; Scott, J. R. J. Bacteriol. 2002, 184, 2181.
19. Bierne, H.; Mazmanian, S. K.; Trost, M.; Pucciarelli,
M. G.; Liu, G.; Dehoux, P.; Jansch, L.; Garcia-del
Portillo, F.; Schneewind, O.; Cossart, P. Mol. Microbiol.
2002, 43, 869.
25. Kim, S. W.; Chang, I. M.; Oh, K. B. Biosci., Biotechnol.,
Biochem. 2002, 66, 2751.
26. Scott, C. J.; McDowell, A.; Martin, S. L.; Lynas, J. F.;
Vandenbroeck, K.; Walker, B. Biochem. J. 2002, 366, 953.
27. Oh, K. B.; Kim, S. H.; Lee, J.; Cho, W. J.; Lee, T.; Kim, S.
J. Med. Chem. 2004, 47, 2418.
28. Zong, Y.; Bice, T. W.; Ton-That, H.; Schneewind, O.;
Narayana, S. V. J. Biol. Chem. 2004, 279, 31383.
29. Kramer, B.; Rarey, M.; Lengauer, T. Proteins 1999, 37, 228.
30. Rarey, M.; Kramer, B.; Lengauer, T. Bioinformatics 1999,
15, 243.
20. Bolken, T. C.; Franke, C. A.; Jones, K. F.; Zeller, G. O.;
Jones, C. H.; Dutton, E. K.; Hruby, D. E. Infect. Immun.
2001, 69, 75.
21. Garandeau, C.; Reglier-Poupet, H.; Dubail, I.; Beretti, J.
L.; Berche, P.; Charbit, A. Infect. Immun. 2002, 70, 1382.
22. Jonsson, I. M.; Mazmanian, S. K.; Schneewind, O.;
Verdrengh, M.; Bremell, T.; Tarkowski, A. J. Infect. Dis.
2002, 185, 1417.
23. Kang, S. S.; Kim, J. G.; Lee, T. H.; Oh, K. B. Biol. Pharm.
Bull. 2006, 29, 1751.
24. Kim, S. H.; Shin, D. S.; Oh, M. N.; Chung, S. C.; Lee, J.
S.; Chang, I. M.; Oh, K. B. Biosci., Biotechnol., Biochem.
2003, 67, 2477.
31. Rarey, M.; Kramer, B.; Lengauer, T.; Klebe, G. J. Mol.
Biol. 1996, 261, 470.
32. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P.
J. Adv. Drug Deliv. Rev. 1997, 23, 3.
33. Kruger, R. G.; Dostal, P.; McCafferty, D. G. Anal.
Biochem. 2004, 326, 42.
34. Ton-That, H.; Mazmanian, S. K.; Faull, K. F.; Schnee-
wind, O. J. Biol. Chem. 2000, 275, 9876.
35. Skotnicki, J. S.; Kearney, R. M.; Smith, A. L. Tetrahedron
Lett. 1994, 35, 197.