Structure and absolute configuration of mycobactin J

Article abstract of DOI:10.1016/S0040-4039(01)00531-7

Mycobactin J 1 is a commercially available siderophore isolated from Mycobacterium avium subsp. paratuberculosis. There are discrepancies between previous reports of its structure and none have addressed its absolute configuration. We report here the complete structure and stereochemistry of mycobactin J, along with methodology to enable the determination of the absolute configuration of other mycobactins on a small scale.

Full text of DOI:10.1016/S0040-4039(01)00531-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 3653–3655
Structure and absolute configuration of mycobactin J
Brett D. Schwartz and James J. De Voss*
Department of Chemistry, University of Queensland, Brisbane, Queensland 4072, Australia
Received 20 March 2001; revised 22 March 2001; accepted 29 March 2001
Abstract—Mycobactin J 1 is a commercially available siderophore isolated from Mycobacterium avium subsp. paratuberculosis.
There are discrepancies between previous reports of its structure and none have addressed its absolute configuration. We report
here the complete structure and stereochemistry of mycobactin J, along with methodology to enable the determination of the
absolute configuration of other mycobactins on a small scale. © 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
The major point of contention among the published
mycobactin J structures is the identity of the b-hydroxy-
acid that links the two lysine derived moieties of the
siderophore. This was originally identified as being
derived from the unusual 2,4-dimethyl-3-hydroxypen-
tanoic acid.5 Later, it was suggested that it was a
2-methyl-3-hydroxy carboxylic acid with a chain length
of approximately 13 carbons.6 Finally, on the basis of
mass spectral evidence, it was recently assigned as the
more common 2-methyl-3-hydroxypentanoic acid.7
Additionally, the aromatic moiety has been suggested
to be derived from either 6-methylsalicylate or salicylate
and the oxazoline has been reported both with and
without a methyl substituent.5–7
Iron is the fourth most abundant element in the earth’s
crust and is an essential nutrient for most microorgan-
isms. However, because the solubility of Fe(III) in
aqueous media is so low (10−18 M at neutral pH) many
bacteria have developed specialised mechanisms for its
acquisition. In mycobacteria, iron uptake is generally
mediated by siderophores (iron chelators) which fall
broadly into two structural classes, the mycobactins
and the exochelins.1 The mycobactins contain a hydroxy-
phenyloxazoline moiety and were the first siderophores
to be isolated from mycobacteria. Structurally, they are
characterised by a core, which varies only slightly
between different mycobactins, and by one to three
very variable fatty acid derived alkyl chains. Myco-
bactins from any given species are usually isolated as
a family of compounds differing in their alkyl substi-
tution patterns. They were originally identified as a
required growth factor for the in vitro culture of
Mycobacterium paratuberculosis (now reclassified M.
avium subsp. paratuberculosis2) the causative agents of
Johne’s disease in cattle.3 Ironically, strains of M.
paratuberculosis adapted to growth on laboratory
2. Results
Our analysis of the 400 MHz proton NMR spectrum of
1 clearly reveals the four protons expected for the
unsubstituted hydroxyphenyl moiety shown in 1 (Fig.
1). The TOCSY spectrum reveals a spin system consist-
ing of a methyl doublet at l 1.6 (J 6 Hz), a multiplet at
media were subsequently shown to produce
a
H
H
CH3
CH3
mycobactin, mycobactin J,4 which is now the only
commercially available mycobactin. The structure of
mycobactin J has been the subject of several reports
which have arrived at slightly differing conclusions and
none of which have addressed the stereochemistry of
the system.5–7 We report here the structure and absolute
configuration of mycobactin J, along with methodology
for rapidly determining the configuration of
mycobactins on a small scale.
O
O H2C
O
O
H
N
H
H
H
N
N
N
H
OH
OH
O
CH3
O
O
N
OH
(CH2)12CH3
1
Figure 1. Structure and stereochemistry of the major compo-
* Corresponding author. E-mail: devoss@chemistry.uq.edu.au
nent of mycobactin J.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00531-7

3654
B. D. Schwartz, J. J. De Voss / Tetrahedron Letters 42 (2001) 3653–3655
HO
OH
O
OH
O
i
ii
n-C12H25
O
n-C13H27
5
Z-3
HO
NH
(CH2)4
O
O
C13H27
Cbz
HO
N
H
N
O
(CH2)4
4
O
Cbz
N
H
iii
O
Z-2
Scheme 1. Synthesis of model compound Z-2. (i) n-BuLi, HMPA, THF, n-C13H27Br, 65%; (ii) H2 Lindlar catalyst, n-hexane,
72%; (iii) EDC, 18%.
l 5.1 and a doublet at l 4.5 (J 8 Hz) consistent with a
methyl substituted oxazoline. Analysis of the TOCSY
spectrum also suggested that the hydroxyacid was a
2-methyl-3-hydroxyacid as shown in 1, although the
number of carbons in the chain could not be deter-
mined. The geometry of the double bond of the long
chain fatty acid moiety was difficult to determine from
the NMR as the vicinal coupling constant cannot be
measured because the vinyl protons appear as broad
multiplets. We therefore synthesised and characterised
E- and Z-2 as model compounds. The E isomer was
readily available from EDC mediated coupling of the
known8 acid E-3 with the protected No-hydroxylysine
4.9 The cis acid Z-3 was available from the partial
hydrogenation of the corresponding acetylene 5, which
in turn came from alkylation of the dianion of propiolic
acid with 1-bromotridecane (Scheme 1). The chemical
shifts of the vinylic and allylic protons seen in 3 and
mycobactin J are given in Table 1, clearly indicating
that the double bond geometry is Z as expected.
methyl ester.9 Mild acid hydrolysis converted the
nitrone to the free hydroxylamine which was immedi-
ately reacted with pentafluoropropionic anhydride
(PFPA). The Cbz groups of the resulting amides were
then removed by hydrogenolysis and the amines treated
with PFPA to yield the triacyl derivative of No-hydroxy-
+
’
lysine 6 as a single product by GC–MS (M , 614). This
compound was found to be somewhat unstable, slowly
converting to the diacyl compound 7 (GC–MS M+
’
468) at room temperature. Treatment with PFPA
regenerated the triacyl derivative 6. Both the - and
DL-forms of the PFPA protected No-hydroxylysine
were synthesised from - and DL-Cbz-lysine methyl
L
L
ester, respectively. Analysis of the protected No-hydroxy-
lysines employing the Chirasil-Val® column showed
that while the enantiomers of the diacyl derivative 7 did
not separate well, those of the triacyl derivative 6 were
clearly resolved. Thus, we felt that we could determine
the stereochemistry of all the amino acid derived cen-
tres in mycobactin J by direct analysis of the hydrolysis
products.
In order to establish the absolute configuration of
mycobactin J we planned to employ the two step
degradation procedure pioneered by Snow in the
1960s.10 Base catalysed hydrolysis cleaves the only ester
bond in the mycobactin and yields two fragments, a
cobactin and a mycobactic acid that can be separated
on the basis of their solubility.
O
RO
N
C2F5
O
O
(CH2)4
HO
O
O
C2F5
N
H
O
(2S,3R)-8
6 R = COC2F5
7 R = H
Acid catalysed hydrolysis of the amide bonds then
yields the constituent acids from each molecule. This
route has the advantage of differentiating the two No-
hydroxylysines in the mycobactin, enabling their
configuration to be determined separately. We wished
to develop methodology for stereochemical determina-
tion that would be compatible with the small amount of
other mycobactins usually available and thus explored
GC analysis employing a chiral column. A Chirasil-
Val® column is known to separate all of the naturally
occurring amino acids when they are protected with
pentafluoropropionyl groups,11 but its resolving power
with respect to No-hydroxylysine was unknown.
Hydrolysis of mycobactin J under basic conditions
yielded the corresponding mycobactic acid and
cobactin. Acid hydrolysis of the mycobactic acid was
followed by esterification and PFPA derivatisation.
GC–MS analysis utilising the Chirasil-Val® column
Table 1. Chemical shifts (400 MHz, CDCl3) of character-
istic protons in 1 and 2
Proton
E-2
Z-2
Mycobactin J 1
H2%
H3%
H4%
5.99
6.95
2.17
5.85
6.05
2.49
5.79
5.97
2.40
No-Hydroxylysine is somewhat unstable, but protected
equivalents of it are readily available through Miller’s
dimethyldioxirane mediated oxidation of Cbz-lysine

B. D. Schwartz, J. J. De Voss / Tetrahedron Letters 42 (2001) 3653–3655
3655
indicated that the only amino acids present in mycobac-
tic acid were -threonine and
-No-hydroxylysine. In
for the mechanism of mycobactin biosynthesis. For
example, retention of the 3R configuration of L-
L
L
addition, methyl salicylate was the only aromatic acid
identified in the diazomethane esterified crude acid
hydrolysate.† The mycobactic acid fragment also con-
tains the fatty acyl substituent which is the source of
much of the structural variation observed in
mycobactins. As expected therefore, GC–MS analysis
revealed a range of fatty acids that included the methyl
esters of 2-hexadecenoic, hexadecanoic and octade-
canoic acids as the major components with traces of
2-octadecenoic and eicosanoic acids. Comparison with
synthetic E- and Z-methyl-2-hexenoate revealed that
the E isomer, rather than the expected Z form was
produced in the degradation of mycobactin J. The ease
of isomerisation of a,b-unsaturated fatty acids
prompted us to investigate the periodate catalysed
cleavage of the fatty acid moiety from the correspond-
ing mycobactic acid.10 As previously reported, no iso-
merisation occurs under these conditions and we
identified only the Z-methyl-2-hexenoate by compari-
son with authentic material (GC coelution, MS
fragmentation).
threonine in the oxazoline implies that cyclisation
occurs via hydroxyl attack on the adjacent amide car-
bonyl rather than the reverse; the latter mechanism is
well precedented in the synthesis of these heterocycles.13
The gross structure of mycobactin J is the same as that
recently proposed for mycobactin Av from M. avium as
might be expected given the reclassification of M.
paratuberculosis as a subspecies of M. avium.2 No stere-
ochemical information is yet available for mycobactin
Av but it appears likely that mycobactin Av and J will
prove identical. The development of GC protocols for
determining the absolute stereochemistry of the
mycobactins will be compatible with the small amounts
of naturally occurring mycobactins available and be
invaluable for comparing the recently identified water
soluble mycobactins7,14 with the cell associated ones.
Acknowledgements
This work was supported by NH&MRC Grant 143031
to J.D.V.
Similarly, analysis of the appropriately protected acid
hydrolysate of cobactin yielded only
L
-No-hydroxy-
lysine and methyl 2-methyl-3-hydroxypentanoate 8. All
four stereoisomers of the latter were synthesised via
double methylation of the dianion derived from methyl
acetoacetate followed by selective reduction of the
ketone of the ketoester with sodium borohydride. The
known 2S,3R enantiomer of 812 was available from
propionic acid by utilisation of Evan’s methodology.
We found that all four stereoisomers could be sepa-
rated by GC on a g-cyclodextrin column and thus
showed that the hydroxyacid derived from mycobactin
J had the 2S,3R configuration. The erythro configura-
tion of these two centres is consistent with that deter-
mined for all other mycobactins to date and the
absolute configuration is the same as that seen in
mycobactin R from M. terrae.10
References
1. De Voss, J. J.; Rutter, K.; Schroeder, B. G.; Barry, C. E.
J. Bacteriol. 1999, 181, 4443–4451.
2. Thorel, M. F.; Krichevsky, M.; Levyfrebault, V. V. Int. J.
Syst. Bacteriol. 1990, 40, 254–260.
3. Francis, J.; Macturk, H. M.; Madinaveitia, J.; Snow, G.
A. Biochem. J. 1953, 55, 596–607.
4. Merkal, R. S.; McCullough, W. G. Curr. Microbiol. 1982,
7, 333–335.
5. McCullough, W. G.; Merkal, R. S. Curr. Microbiol. 1982,
7, 337–341.
6. Barclay, R.; Ewing, D. F.; Ratledge, C. J. Bacteriol.
1985, 164, 896–903.
7. Gobin, J.; Moore, C. H.; Reeve, Jr., J. R.; Wong, D. K.;
Gibson, B. W.; Horwitz, M. A. Proc. Natl. Acad. Sci.
USA 1995, 92, 5189–5193.
3. Conclusion
8. Hino, T.; Nakakyama, K.; Taniguchi, M.; Nakagawa, M.
J. Chem. Soc., Perkin Trans. 1 1986, 1687–1690.
9. Hu, J. D.; Miller, M. J. J. Org. Chem. 1994, 59, 4858–
4861.
10. Snow, G. A.; White, A. J. Biochem. J. 1969, 115, 1031–
1045.
We have clarified the structure of mycobactin J as that
shown in 1, in which the hydroxyacid is derived from
(2S,3R)-3-hydroxy-2-methylpentanoate. In addition, we
have also shown that the amino acid derived stereocen-
tres all have the configuration found in the naturally
occurring
L-isomers. This has important implications
11. Frank, H.; Nicholson, G. J.; Bayer, E. J. Chromatogr.
Sci. 1977, 15, 174–176.
12. Cane, D. E.; Yang, C. C. J. Am. Chem. Soc. 1987, 109,
1255–1257.
13. Wipf, P.; Miller, C. P. Tetrahedron Lett. 1992, 33, 907–
910.
14. Lane, S. J.; Marshall, P. S.; Upton, R. J.; Ratledge, C.;
Ewing, M. Tetrahedron Lett. 1995, 36, 4129–4132.
† The standard methanol/HCl and PFPA derivatisation procedures
resulted in the loss of the more volatile fatty esters. These were
examined by GC–MS after esterification with ethereal dia-
zomethane.
.