Total synthesis and semi-synthetic approaches to analogues of antibacterial natural product althiomycin

Article abstract of DOI:10.1016/S0960-894X(01)00802-2

Numerous analogues of the naturally occurring antibiotic althiomycin have been synthesised exploiting both total- and semi-synthetic methodologies. The antibacterial activity of these derivatives has been determined in whole cell assays and indicates the natural product exhibits a restricted SAR.

Full text of DOI:10.1016/S0960-894X(01)00802-2

Bioorganic & Medicinal Chemistry Letters 12 (2002) 561–565
Total Synthesis and Semi-Synthetic Approaches to
Analogues of Antibacterial Natural Product Althiomycin
Paola Zarantonello,a,* Colin P. Leslie,a,* Rafael Ferrittoa,y
and Wieslaw M. Kazmierskib
aGlaxoSmithKline SpA, Medicines Research Centre, Via Fleming 4, 37135 Verona, Italy
bGlaxoSmithKline, Five Moore Drive, Research Triangle Park, NC 27709, USA
Received 14September 2001; revised 22 November 2001; accepted 26 November 2001
Abstract—Numerous analogues of the naturally occurring antibiotic althiomycin have been synthesised exploiting both total- and
semi-synthetic methodologies. The antibacterial activity of these derivatives has been determined in whole cell assays and indicates
the natural product exhibits a restricted SAR. # 2002 Elsevier Science Ltd. All rights reserved.
Althiomycin (1) is an antibiotic, isolated in 1957 from
Streptomyces althioticus,1 whose biological action is
believed to derive from its ability to inhibit protein
synthesis at the peptidyltransferase stage.2 It has been
characterised as a broad spectrum agent endowed with
low cytotoxicity and good selectivity towards pro-
karyots.3 The structure of althiomycin was determined
independently by various groups4 and has been con-
firmed by total synthesis.5 As depicted in Figure 1, the
cysteine embedded in the thiazoline moiety has an (S)-
configuration whereas the stereochemistry of the asym-
metric centre on the serine residue is still ambiguous: it
may either exist in racemic form or it might undergo
racemisation during the isolation procedure.
SAR that has emerged. A small number of analogues
have already been reported,6 but in some cases anti-
microbial activities were not evaluated and there is,
however, insufficient information to define a usable
SAR.
We simplified chemistry by starting with the core 2 pre-
pared from glycine (Fig. 2, X=H), which also allowed
us to eliminate ambiguities with respect to the stereo-
chemistry. This modification was not expected to have a
detrimental effect on activity as dehydroxymethyl-
althiomycin is reported to be as potent an antibiotic as
althiomycin itself.6a
The potency and selectivity profile of althiomycin along
with the relative simplicity of its structure makes it an
attractive medicinal chemistry target. The molecule,
which can be retrosynthetically disconnected at both
amide bonds, appears well suited to a parallel synthesis
strategy based around a functionalised thiazoline core 2
(Fig. 2).
Figure 1. Structure of althiomycin.
In this letter, we wish to present the results of our initial
explorations directed towards establishing chemical
routes to althiomycin analogues, as well as preliminary
*Corresponding authors. Fax: +39-045-9218196; e-mail: pz29728@
gsk.com; col44577@gsk.com
yCurrent address: Eli-Lilly SA., Arda de la Industria 30, 28108 Alco-
bendas, Madrid, Spain.
Figure 2. Envisaged synthetic pathway.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00802-2

562
P. Zarantonello et al. / Bioorg. Med. Chem. Lett. 12 (2002) 561–565
Scheme 1. (a) DCC, TEA, DCM, rt, 70–80%; (b) Lawesson’s reagent, xylene, reflux, 40–50%; (c) pyrrolidine neat, rt, 1 h, 90%; (d) 4 N HCl/
dioxane, rt, 2 h, quantitative yield; (e) 8a–e, PyBOP, TEA, DCM, rt, 3–18 h, 50% to quantitative yield; (e0) 8f–n, EDC, HOBT, TEA, DCM, rt, 18 h,
yield >70%; (f) TEA, PyBOP, DMAP, rt, thiazole-2-carboxylic acid, 18 h, 50%; (g) HNR1R2 9a–f, neat, 50–75%.
Chemistry
reagent. Fortuitously we found that the thioacyl serine
formed in situ cyclised directly under the reaction con-
ditions to give the thiazoline, providing a short and
extremely convenient protocol for the preparation of 3.
Subsequent hydrolysis of the methyl ester proved capri-
cious under various conditions and the desired acid,
when detected, was always accompanied by substantial
amounts of a thiazoline ring opened product. This
unforeseen setback restricted the possibilities for scaf-
fold decoration, but nonetheless direct amidation with
pyrrolidine proved successful, furnishing the desired
amide in 90% yield. Amine deprotection using HCl
gave 4, which was coupled to a series of substituted
thiazole carboxylic acids 8a–n (Fig. 3) to produce a
small array of dehydroxymethylalthiomycin analogues
5a–n.
The key thiazoline intermediate 3 (Scheme 1) was init-
ially prepared by condensation of Cys-OMe with the
ethyl imidate of Boc-Gly, adapting the chemistry used
by Shiba in his total synthesis of althiomycin,5 but we
found that this procedure did not readily lend itself to
large scale preparations.7 The thiazoline formation
strategy utilised in the only other reported synthesis of
althiomycin3a was deemed to be unacceptably low
yielding for our purposes, so an alternative methodol-
ogy was sought. After consulting the literature we deci-
ded that the dehydration of a thioacyl serine using
Burgess’ reagent8 was the most promising alternative
for merit of its mildness and high yield. To this end Boc-
Gly-Ser-OMe was prepared and heated with Lawesson’s
Couplings were initially effected using PyBOP, which
gave the desired amides in excellent yield. Unfortu-
nately, purification was complicated due to the presence
of byproducts arising from the reagent itself that proved
difficult to separate chromatographically. Coupling
conditions were therefore revised to use EDC in order
to facilitate product isolation. Following this synthetic
route the oxime-substituted thiazole moiety of althio-
mycin can be introduced in either protected or non-
protected form (Fig. 3).
The key intermediate 3 was also deprotected and cou-
pled with thiazole-2-carboxylic acid to afford compound
6. Selective hydrolysis of the methyl ester was again
found to be impractical, but an additional small array
of dehydroxymethylalthiomycin analogues 7a–f was
obtained by direct treatment of the ester with different
commercially available amines 9a–f (Fig. 4). Some more
functionalised and more hindered amines failed in this
direct amidation procedure due to decomposition and
unreactivity, respectively.
Figure 3. Thiazole acids 8a–n used in the synthesis of amides 5a–n
(Scheme 1).
As mentioned above, the lability of the thiazoline ring in
methyl esters 3 and 6 towards hydrolysis prevents access
to a more activated acylating species, thereby seriously
limiting the derivatives that may be prepared via our
original reaction pathway — in particular, the poorly
nucleophilic pyrrolinone present in althiomycin cannot
be introduced by direct displacement of the methyl
ester. To circumvent these limitations we turned to an
Figure 4. Amines 9a–f used in the synthesis of amides 7a–f (Scheme 1).

P. Zarantonello et al. / Bioorg. Med. Chem. Lett. 12 (2002) 561–565
563
Scheme 2. (a) (i) p-nitrophenylchloroformate, TEA, CH2Cl2, 0ꢀC, 15 min; (ii) DMAP, 0ꢀC, 30 min, 100%; (b) BuLi, 4-methoxy-3-pyrrolin-2-one,
THF, rt, 2 h, 89%; (c) DBU, MeCN, 40ꢀC, 1 h, 97%; (d) [2-(6-nitro-benzotriazol-1-yl)-2-thioxo-ethyl]-carbamic acid tert-butyl ester, THF, rt, 24h,
61%; (e) 1 M HCl/Et2O, DMS, 100%; (f) 18a–f, EDC, HOBt, K2CO3, 35–90%; (g) TFA, DMS; (h) Burgess’ reagent, THF, rt to 50ꢀC, 30–90%; (i)
.
TBAF, THF, rt, 1 h (56%, 17b); (j) H2NOH HCl, TEA, MeOH, rt, 1 h (72%, 17e; 48%, 17f).
alternative pathway in which the sensitive thiazoline
ring is constructed at the last possible moment. In this
case it is not possible to exploit the ‘one-pot’ thiazoline
formation used previously due to the presence of a sec-
ond amide group, which would react with Lawesson’s
reagent. Hence, we reverted to our original plan, pre-
paring the thioacyl serine, via coupling of serine to an
activated thioacyl species, and subsequently dehyd-
rating with Burgess’ reagent. As hoped this sequence
turned out to be simple and high yielding with the only
drawback of having to protect other moieties, such as
an oxime, that are subject to dehydration.
A small array of derivatives was prepared using the
acids 18a–f depicted in Figure 5.
TBAF deprotection of 16b gave synthetic dehydroxy-
methylalthiomycin 17b0 and aldehydes 16e,f were con-
verted into their oximes 17e0,f0.
Semi-synthetic approach
Concurrently with the work described above, fermenta-
tions were set up to provide a supply of althiomycin to
allow both further biological characterisation and the
preparation of analogues via semi-synthetic strategy.
The range of potential products available from such a
strategy is somewhat limited, but those that are may be
of particular importance in determining the roles played
by the oxime and hydroxyl groups.
This second route proved considerably more versatile
and should be suitable for the preparation of almost any
desired derivative but it is penalised by the number of
chemical transformations involved, meaning that the
time required to accumulate enough data for an SAR is
greatly extended.
Initial experiments reacting 1 (Scheme 3) with amines,
which were expected to substitute the pyrrolinone, were
successful using pyrrolidine and phenylpiperazine
Fmoc-Ser(tBu)-OH was activated as its p-nitrophenyl
ester 10 and coupled to 4-methoxy-3-pyrrolin-2-one to
furnish 11. Amine deprotection and subsequent thio-
amide formation using Rapoport’s thioacylbenzotriazole
procedure9 gave thioamide 13. The procedure developed
to complete the synthesis consisted of selective depro-
tection of the Boc group using HCl followed by cou-
pling to an activated acid to give 14. TFA deprotection
t
of the Bu ether and final thiazoline formation using
Burgess’ reagent then gave dehydroxymethylalthiomy-
cin analogues 16.
This procedure was subsequently changed after encoun-
tering problems of stability of the protected oxime 14b
towards TFA treatment. The revised procedure calls for
t
global deprotection of the Boc and Bu ether groups of
13, followed by amide coupling and cyclisation; the
presence of an unprotected serine residue during the
amide coupling turned out to be inconsequential.
Figure 5. Acids used in the synthesis of amides 14a,b and 15b–f
(Scheme 2).

564
P. Zarantonello et al. / Bioorg. Med. Chem. Lett. 12 (2002) 561–565
Scheme 3. (a) Amine, rt, o/n, 90%; (b) benzyltrichloroacetimidate, TFA, CH2Cl2, rt, 70%; (c) 21a, P=H, P1=Me: TMSCHN2, MeOH, rt, 30 min,
48%; 21b, P=P1=TBDMS: TBDMSOTf, DIPEA, DCM, rt, 46%; (d) Ac2O, Py, DMAP, 80 ꢀC, 30%; (e) TMSCHN2, MeOH, rt, 20 h, 35%.
Table 1. Minimum inhibitory concentrationa (MIC) results for compounds 1 and 17b0
Compds
S. aureus
853E
S. aureus
col
B. subtilis
6633b
E. faecalis
850E
S. pneum
3512
E. coli
1852E PM
E. coli
1852E+3 mg/mL PMBN
S. cerevisiae
NCYC 81
Ceftriaxone
1
17b0
4
16
32
>32
16
32
0.25
16
16
16
16
32
0.01
41
16
0.06
16
0.01
8
>32
>32
>32
1
aValues expressed in mg/mL.
(1!19a,b), while methylbenzylamine and m-anisidine
did not react. Attempted benzylation using benzyl-
trichloroacetimidate in the presence of trifluoro-
acetic acid induced dehydration without alkylating the
oxime (1!20). Methylation of the oxime was achieved
using trimethylsilyldiazomethane in methanol (1!21a),
with subsequent methanolysis of the imide being
observed upon extended reaction time (1!23). The
doubly silylated product 21b was prepared using
TBSOTf. Heating the oxime acetate gave the nitrile
derivative 22.
Moreover, the semi-synthetic products have, as hoped,
been extremely valuable in helping to determine impor-
tant pharmacophoric points. The complete loss of anti-
microbial activity associated with minor structural
changes indicates the essential roles played by the oxime
(cf., the methylated derivative 23 to 1), the hydroxyl
group (compare the dehydrated product 20 or the
dehydroxymethyl compound 17b0 to 1) and the meth-
oxypyrrolinone moiety (19a,b vs 1). It is premature to
draw conclusions as to the roles that these groups may
play, but the next step in defining what appears to be a
very restricted SAR could be the preparation of deriva-
tives in which each pharmacophoric element is selec-
tively replaced by carefully chosen alternatives. Metal
chelating motifs could be considered as oxime replace-
ments, as H-bond donor/acceptors could be used to
substitute the hydroxyl group. Another possibility to
further simplify the chemistry and explore the SAR could
rest in replacing the chemically labile thiazoline moiety
with an alternative heterocycle to determine whether
this ring functions as a spacer or if it also plays an
important part in the biological action of althiomycin.
Results and Discussion
All final products and ‘full length’ precursors, plus a
sample of natural althiomycin, were evaluated in a
standard MIC assay10 using a panel of Gram-positive
and Gram-negative bacteria. The reported activity of
althiomycin was confirmed, although its potency
towards a number of clinically significant Gram-positive
strains leaves much to be desired.
Unfortunately, only one of the ꢁ50 synthetic molecules
tested, dehydroxymethylalthiomycin 17b0, showed an
appreciable (<32 mg/mL) antibacterial activity (Table 1).
Conclusions
In summary, we have described novel synthetic routes
to synthesise althiomycin analogues. Products prepared
using these approaches, along with those available semi-
synthetically, have enabled us to identify several impor-
tant pharmacophoric points present in althiomycin.
Further studies, which are necessary in order to under-
stand the true potential of althiomycin-like compounds
as novel antibacterial drugs, will be the subject of future
publications.
Interestingly, the MICs exhibited by 17b0 in our assay
are considerably lower than those reported by others.6a
Although this result casts a shadow on the validity of
our initial decision to work around the dehydroxy-
methylalthiomycin core, we have succeeded in setting up
a versatile synthetic route that opens the way for the
preparation of more complex derivatives bearing greater
structural similarity to althiomycin itself.

P. Zarantonello et al. / Bioorg. Med. Chem. Lett. 12 (2002) 561–565
565
Acknowledgements
4. (a) Nakamura, H.; Itaka, Y.; Sakakibara, H.; Umezawa, H.
J. Antibiot. 1974, 27, 894. (b) Sakakibara, H.; Naganawa, H.;
Ohno, M.; Maeda, K.; Umezawa, H. J. Antibiot. 1974, 27,
897. (c) Kirst, H. A.; Szymanski, E. F.; Dorman, D. E.;
Occolowitz, J. L.; Jones, N. D.; Chaney, M. O.; Hamill, M. L.;
Hoehn, M. M. J. Antibiot. 1975, 286. (d) Shiba, T.; Inami, K.;
Sawada, K.; Hirotsu, Y. Heterocycles 1979, 13, 175. (e)
Bycroft, B.; Pinchin, R. J. Chem. Soc., Chem. Commun. 1975,
121.
The authors wish to express their gratitude to the
Microbiology and Analytical Departments, Glaxo-
SmithKline SpA, Verona, Italy and the Bioprocessing
Department, GlaxoSmithKline, Stevenage, UK for their
support in this work.
5. Inami, K.; Shiba, T. Tetrahedron Lett. 1984, 25, 2009.
6. (a) Inami, I.; Shiba, T. Bull. Chem. Soc. Jpn. 1986, 59, 2185.
(b) Cram, D. J.; Theander, O.; Jager, H.; Stanfield, M. K. J.
Am. Chem. Soc. 1963, 85, 1430.
References and Notes
1. Yamaguchi, H.; Nakayama, Y.; Takeda, K.; Tawara, K.;
Maeda, K.; Takeuchi, T.; Umezawa, H. J. Antibiot. 1957,
195.
2. (a) Fujimoto, H.; Kinoshita, T.; Suzuki, H.; Umezawa, H.
J. Antibiot. 1970, 23, 271. (b) Pestka, S.; Brot, N. J. Biol.
Chem. 1971, 246, 7715. (c) Burns, D. J.; Cundliffe, E. Eur. J.
Biochem. 1973, 37, 570.
3. (a) Toogood, P. L.; Hollenbeck, J. J.; Lam, H. M.; Li, L.
Bioorg. Med. Chem. Lett. 1996, 6, 1543. (b) Inami, K.; Shiba,
T. Bull. Chem. Soc. Jpn. 1985, 58, 352.
7. Yields were found be extremely variable with oxidation of
cysteine to cystine or rearrangement of the ethyl imidate to
ethyl amide variably becoming important side reactions.
8. Wipf, P.; Fritch, P. Tetrahedron Lett. 1994, 35, 5397.
9. Shalaby, M. A.; Grote, C. W.; Rapoport, H. J. Org. Chem.
1996, 61, 9045; we were unable to prepare Brain’s thio-
acylphthalimide reagent11 starting from Boc-Gly.
10. NCCLS M7-A4.
11. Brain, C. T.; Hallett, A.; Ko, S. Y. J. Org. Chem. 1997, 62,
3808.