Journal of General Microbiology (1984)’ 130, 1883-1 892. Printed in Great Britain
1883
Mycobactins as Chemotaxonomic Characters for Some Rapidly Growing
Mycobacteria
By R I C H A R D M. HALL* A N D C O L I N R A T L E D G E
Department of Biochemistry, University of Hull, Hull HU6 7RX, UK
(Received 19 December 1983; revised 23 March 1984)
Thirty-nine strains of rapidly growing mycobacteria were examined for the production of
mycobactins (lipid-soluble, iron-binding compounds) when grown under conditions of ironlimitation on solidified medium. Different growth conditions had little effect on the structure of
individual mycobactins, indicating them to be strongly conserved molecules showing intraspecies consistency and thus suitable for use as chemotaxonomic characters of high
discriminatory power. Strains of Mycobacterium aurum, M. chitae, M. chelonae subsp. abscessus,
‘ M . diernhoferi‘, M . duvalii, M . jlavescens, M . fortuitum, M. gadium, ‘M. gallinarum’, M .
neoaurum, M . parafortuitum, ‘M.
peregrinum’, M . phlei, M . smegmatis, M . thermoresistible and
M . vaccae formed mycobactins which were readily isolated and characterized by a combination
of thin-layer and high-performance liquid chromatography. All strains of M . komssense and
‘M. kanazawa’ failed to produce a mycobactin; some strains of M. aurum, M. chelonae, M .
parafortuitum, M . thermresistible and M. vaccae were similarly negative. Mycobacteria of the
M .fortuitum complex ( M .fortuitum, M . chelonae and ‘ M .peregrinum’) formed distinctive mycobactins, as did those in the M. parafortuitum complex ( M .aurum, M . neoaurum, ‘M. diernhoferi‘,
M. uaccae and M . parafortuitum). The mycobactin from ‘M. gallinarum’ was different from those
of the related species M._Pavescens,for which four distinct mycobactin patterns were recorded.
For routine examination of mycobactins in a diagnostic laboratory with limited resources, thinlayer chromatography used alone offers a simple but adequate means of characterization and
final identification of the producing mycobacterium. High-performance liquid chromatography
is only needed in those few instances where a high degree of discrimination is required.
INTRODUCTION
The application of various chemical methods has greatly improved the classification and
identification of mycobacteria (Minnikin & Goodfellow, 1980; Goodfellow & Wayne, 1982).
Snow & White (1969) suggested a possible approach to identification at the species level based
on the variation in chemical constitution of the lipid-soluble, iron-chelating compounds, the
mycobactins (Fig. l), produced by mycobacteria under conditions of iron-limitation. These
compounds are retained within the cell envelope and function to transport iron across these
hydrophobic layers into the cell (Ratledge, 1982b). Different mycobacterial species produce
mycobactins, which comprise a family of homologues having long alkyl side chains of various
lengths (Fig. 1). Although Snow (1970) showed that species-specificdifferences could be readily
observed following thin-layer chromatography (TLC) of these compounds, their real taxonomic
potential was never proved, owing to the major difficulty of devising a suitably simple procedure
for isolating, purifying and analysing mycobactins. Only on a few occasions have they been used
to identify a ‘difficult’ species (White & Snow, 1969; Hedren, 1974).
We have recently developed an extremely simple method which promotes good accumulation
of mycobactin using a simple glycerol/asparagine medium solidified with agar which does not
need any pre-treatment for the removal of iron (Hall & Ratledge, 1982). This approach, in
conjunction with rapid methods of purification which have been developed in this laboratory
Downloaded
from www.microbiologyresearch.org by
0022-1287/84/0001-1675 $02.00
0 1984 SGM
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
1884
R . M. H A L L A N D C . R A T L E D G E
CHR5
i""
Fig. I . General structure of ferric mycobactin showing points of substitution (R,-R,) where R, is an
acyl sidechain (C, *-CI9)(Snow, 1970; Ratledge & Ewing, 1978)and R2, R3, R, and R5 are usually H or
CH3 substituents (Snow, 1970; Ratledge, 19820).
(Ratledge, 1982a), and the use of TLC and high-performance liquid chromatography (HPLC)
(Ratledge & Ewing, 1978) has now proved to be a powerful tool for the classification and
identification of mycobacteria. Our results with mycobactins from some rapidly growing
mycobacteria are presented.
METHODS
Organism and growth for mycobactinformation. The bacterial strains examined for mycobactin formation are
listed in Tables 1 and 3. They were maintained on glucosefyeastextractfagarslopes (Gordon & Mihm, 1962)or as
dense suspensions of cells frozen in 20% (v/v) glycerol at -20 "C (Wellington & Williams, 1978). Mycobacteria
were initially grown in liquid glycerol/asparagine medium (Ratledge & Hall, 1971)or glucose/yeast extract broth,
neither of which was treated for iron removal, until good growth had been established. A sample (0.2 ml) was
removed from each culture and inoculated on to glycerol/asparagine medium solidified with 2% (w/v) Lab M agar
(London Analytical and Bacteriological Media, London, UK) as previously described (Hall & Ratledge, 1982)for
the promotion of mycobactin formation. Cultures were then incubated until mycobactin was clearly visible by
characteristic apple-green fluorescence of the desferri-molecule under UV light.
Isolation of mycobactinfor characterization. Bacteria were harvested and the mycobactin extracted with calculation of final yields and partial purification carried out as previously described (Hall & Ratledge, 1982). Species
which failed to form mycobactin on glycerol/asparagine/agar (Table 3) were re-grown on glucose/yeast
extractfagar and yielded mycobactin where shown. Those strains which still failed to produce a mycobactin were
grown in 2 1 irondeficient glycerol/asparagine (Ratledge & Hall, 1971)medium dispersed in 100 ml lots in 250 ml
conical flasks, shaken at 37 "C at 200 r.p.m.; and on between 50 and 100 plates (9 cm diam.) containing
glycerol/asparagine/agar(Hall & Ratledge, 1982).
Analytical methods. TLC analyses of the mycobactins were performed, in a single dimension, employing
various adsorbents and solvents: system I, silica gel G, 20 cm x 20 cm (Analtech, Newark, USA), which was
developed with petroleum spirit (b.p. 60 to 80 "C)/ethyl acetatefn-butanol(2 :3 :3, by vol.); system 11, Kieselgel60,
10 cm x 20 cm with a 2.5 cm x 10 cm concentrating zone (Merck, BDH), with petroleum spirit/ethyl acetatelm
butanol (2 :3 :3, by vol.); system 111, high performance thin-layer plates 10 cm x 10 cm with a 2.5 cm x 10 cm
concentrating zone (Merck, BDH) with petroleum spirit/ethyl acetateln-butanol (2: 3 :3, by vol.); system IV,
alumina GF, 20 cm x 20 cm (Analtech) with cyclohexaneln-butanol (9 :1, v/v); system V, plates as for system I
but using methanol/ethyl acetate (4 :1. v/v); system VI, plates as for system I but using propan-2-01 as solvent.
HPLC was carried out using a Spectra-Physics (Santa Clara, Calif., USA) SP 8000 instrument. Optimum
separation was achieved using a reverse-phase column (4 mm x 250 mm) with packing material having a nonpolar surface activity and a particle size of 10 pm (Lichrosorb RP18, Jones Chromatography, Llanbradach, MidGlamorgan, UK). Mycobactins isolated from six strains of M . fortuitum were initially chromatographed using a
Lichrosorb RP8 column and a methanolfwater gradient (70 :30, v/v) which was changed in a linear gradient over
90 min to methanol alone (Table 2). However, a methanol/water gradient (80:20, v/v), run over 30 min to
methanol alone, was routinely employed for all other mycobactins examined. The solvent flow rate was 2 ml
min - I , with a column temperature of 50 "C. For analytical separations, 10 pl of mycobactin in methanol (about
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
1885
My cobact ins
10 mg m1-I) was applied to the column. The eluate was monitored with a UV-visible detector (Spectra-Physics
Model 770) set at 450 nm - the absorption maximum of the mycobactins. All solvents were of HPLC-grade quality.
The relative percentages of the various peak areas were calculated by the electronic integrator of the SP 8OOO.
RESULTS
Consistency of mycobactin production
In view of the likely chemotaxonomic importance of mycobactin, it was first essential to
establish that mycobactins isolated from different strains of the same species produced the same
TLC and HPLC patterns. Mycobacterium fortuitum was chosen as a representative of rapidly
growing mycobacteria, a number of different strains being available. Seven strains were grown
under conditions of iron-limitation on solid medium for 6 d. Examination of the mycobactins by
TLC (Table 1) showed that each migrated as a single component in each of four different solvent
systems. In each case, the RF values of all seven mycobactins were extremely close, if not
identical. Each mycobactin, except that of M . fortuitum NCTC 2291, was then examined by
HPLC with the RP8 column and separated into a number of components (Table 2). The
chromatographic elution patterns for the mycobactins were also similar for all strains, except for
Table 1. Thin-layer chromatography of mycobactins isolated from M . fortuitum strains
Mycobactin yields are given with RF values obtained by TLC employing different solvent systems.
M.fortuitum
NCTC strain no.
Mycobactin
yield (%)
1542
2291
2891
363 1
8697
10394
10395
3-9
3.5
3.2
3.2
3.8
3.9
4.0
r
RF values for TLC system:*
A
I
IV
V
VI
0-48
0.48
0.49
0.48
0-48
0.48
0.49
0.29
0-29
0.29
0.29
0-30
0-31
0-29
0.90
0.90
0.90
0.50
0-51
0.50
0.50
0.5 1
0.50
0.5 1
0.90
0.90
0.90
0.90
I
See Methods for details of TLC systems used.
Table 2. High-performance liquid chromatography of mycobactins isolated from M . fortuitum
strains
Separation of the mycobactins was as described in Methods, employing a Lichrosorb RP8 column and
methanol/water (70 :30, v/v) gradient over 90 min to methanol alone. All mycobactins gave a similar
number of minor components. Only those peak areas greater than 3% of final chromatogram are given,
to simplify the presentation.
Retention time (s) of mycobactins
(peak area, %)*
M.fottuitwn
NCTC strain no.
3631
8697
825
(3.6)
829
(6.8)
10395
10394
289 1
1542
830
(3.3)
832
6-91
833
(6.5)
830
(11.1)
898
(6.8)
905
(13.9)
906
(5.2)
908
(12.5)
905
(17-7)
1350
(4.2)
1436
(6-4)
1448
(5.7)
1447
(7.2)
1442
(6.3)
1445
(3.1)
1422
(10.8)
1514
(12-5)
1526
(9.5)
1533
(10-7)
1525
(9.7)
1524
(9-9)
1706
(17.1)
1816
(16.4)
1825
(17.4)
1829
(28.5)
1825
(23.7)
1827
(1 1-9)
1798
,
1914
(53-2)
1923
(42.1)
1922
(43.0)
1928.
(31.8)
1920
(44.0)
(48.0)
* The mycobactin profiles were different in form to those obtained later (Table 4; Fig. 2) using a Lichrosorb
RP18 column.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
I886
R . M . HALL A N D C. RATLEDGE
Table 3. TLC of mycobactins from rapidly growing mycobacteria
RF values with TLC system:$
Species
M.
M.
M
M.
aurum
aurum
aurum
aurum
M .chitae
M . chitae
M . chelonae subsp.
abscessus
M . chelonae subsp.
chelonae
' M . diernhoteri'
M . durialii
M . jarexens
M .flarescens
M . ,Patiescens
Yield
I
I1
111
IV
Nil
4-2t
4.07
4-2t
5.7
5.9
Trace
-
-
__
-
0.40
0.40
0.40
0.42
0.43
ND
0.28
0.28
0.27
0-31
0.33
0.89
0-26
0.26
0.26
0.44
0.45
0.86
JS78 (NCTC 946 type strain)
Nil
-
-
-
JS423 ( R . Bonicke SN 1402)
MI64 (NCTC 8645)
M410 (ATCC 14474 type strain)
M411 (M.Tsukamura, 33002)
5.1
0.40
0.41
0.37
0.7 1
0.69
0.27
0.33
0.56
0.60
0.64
0.7 1
Designation*
NCTC 10437 (type strain)
M397 (NCTC 10440)
M398 (ATCC 25801)
M399 (M. Tsukamura, 15601)
M393 (ATCC 19627 type strain)
M395 (ATCC 19629)
JS41 (NCTC 10269)
M413 (M. Tsukamura, 33004)
(04
5.4
4.7
4.9
0.72
0.76
0-28
0-30
0-75
0.76
4.0
0.66
0.63
0.38
0.79
0.68
0-40
-
-
M . jlawscens
M . fortuitum
M414 (ATCC 23008)
NCTC 10394 (type strain)
4.8
5.8
0.7 1
0.78
0.45
0.61
0.71
0.35
M .gadium
M416 (ATCC 27726 type strain)
5.4
0.70
0.30
0.68
M79 (type strain)
4.3
0.76
0.6 1
0-69
' M .gallinarum'
0.23
0.22
0.22
0.49
0.50
ND
~
0.57
0.65
0.44
0.40
0.65
0.59
0.65
-
0-70
0.36
0.70
0.72
0.58
the mycobactin from NCTC 1542 where four major peaks (those at 1350,1422,1706 and 1798 s)
were eluted some 100 s to 120 s earlier than those found in the other five strains. The reason for
this difference, which was obtained upon re-examination of the mycobactin using a different
solvent system, is not obvious. However, this minor difference in the elution profile was not
regarded as significant, as there were still sufficient points of similarity between the overall
profile given by NCTC 1542 mycobactin and the other mycobactins not to challenge the identity
of this strain. An examination by both TLC and HPLC of the mycobactins from multiple strains
of M .aurum (M397, M398, M399) and M . purufortuitum (M403, M404, M405, M406) (Table 3)
again showed the intraspecies consistency of these compounds.
Possible changes in the consistency of the mycobactins was also tested by examining the
mycobactin produced by M . smegmutis NCIB 8548 grown on both liquid and solidified media
over the entire growth period. From the earliest (2 d) to the latest (12 d) time of sampling, the
mycobactins were identical in each type of culture and were also chromatographically identical
to each other. The time of sampling and the method used to derepress mycobactin formation in a
mycobacterium are therefore of little, if any, importance. Mycobactin production by a given
organism can therefore be taken as being reliable and reproducible.
Survey of mycobactins produced by the rapidly growing mycobacteria
Having established the consistency of these compounds as chemotaxonomic markers, an
extended examination of rapidly growing mycobacteria was undertaken using both TLC (Table
3) and HPLC with the RP18 column (Table 4). From Table 3 one can see that the first-stage
examination by TLC enables most rapidly growing mycobacteria to be distinguished on the
basis of their mycobactins. Although four solvent systems were used, only two (namely I and IV)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
1887
M y cobactins
Table 3 (continued)
Species
Designation*
Yield
(%)
' M . gallinarum'
'M . kanazawa'
'M . komossense'
'M . komossense'
'M . komossense'
M . neoaurum
M . parafortuitum
M . parafortuitum
M . parafortuitum
M392 (ATCC 19710)
JS758 (ATCC 2581 1 special)
M386 (ATCC 33013)
M387 (ATCC 33014)
M388 (J. Kazda, KO 8)
JS278 (NCTC 10439)
ATCC 19686 (type strain)
NCTC 10411
M403 (ATCC 19687)
4.1
Nil
Nil
Nil
Nil
6.6
Nil
Nil
4.77
M . parafortuitum
M . parafortuitum
M . parafortuitum
'M.peregrinum'
M404 (ATCC 19686 type strain)
M405 (ATCC 19688)
M406 (M. Tsukamura, 16004)
NCTC 16575
M . phlei
M . smegmatis
M . thermoresistible
M . thermoresistible
M . thermoresistible
M . vaccae
M . vaccae
M . vaccae
M . vaccae
NCIB 8573
NCIB 8548
M407 (ATCC 19528)
M408 (ATCC 19529)
M409 (ATCC 19527 type strain)
JS R959R (Soil, Uganda)
JS R877R (Soil, Uganda)
ATCC 15483 (type strain)
M302 (NCTC 10916)
RF values with TLC system:$
f
A
I
>
I1
I11
IV
-
-
0.39
0.27
0.37
0.23
-
-
-
-
-
0.41
0.34
0.31
0.49
4.6t
4.lt
4.4t
0.40
0.39
0.41
0.46
0.26
0.35
0-34
0.36
0.34
6.6
6.4
Nil
3.2t
3.5t
Nil
Nil
Nil
Trace
0.76
0.39
0.40
0.39
0.63
0.33
0.29
0.29
0.29
049
0-43
0.77
0.45
0.47
0.45
0.46
0.37
0.32
0.60
0.20
0.31
0.30
0.30
0.28
0.48
0.46
0.75
0.60
0.66
0-62
5.0
-
-
-
-
-
-
-
ND, Not determined.
M, Cultures kindly supplied from his collection by Dr M. Goodfellow, Department of Microbiology, University of Newcastle upon Tyne, UK ; JS, Cultures kindly supplied from his collection by J. Stanford, Department of
Pathology, Middlesex Hospital, London, UK ;ATCC, American Type Culture Collection, Rockville, Md., USA;
NCTC, National Collection of Type Cultures, Central Public Health Laboratory, Colindale, London, UK; NCIB,
National Collection of Industrial Bacteria, Torry Research Station, Aberdeen, UK.
t Organisms produced mycobactin only on glucose/yeast extract/agar.
$ See Methods for details of TLC systems employed; refers to multiple spots with R, values quoted.
would be needed in any future routine examination of the mycobactin by this method. Where a
number of strains of a particular species were examined, good consistency of the RFvalues of the
mycobactins was again observed, except for the four strains of M.Jlavescens where two strains,
M410 and M411, appeared to produce mycobactins distinct from those of the other two strains.
Examination of the mycobactins by HPLC provided the most powerful method for resolving
differences between species (Table 4), giving highly characteristic elution profiles. Mycobactins
from the different species were compared on the basis of the peak shape, their number, relative
percentage and retention time. In view of the intraspecies consistency of the mycobactins only
the HPLC data for a single representative from each species is given in Table 4. However, owing
to the apparent heterogeneity of the M.Jlavescens taxon, data for all the strains are presented.
Species differences in the mycobactins were both obvious and highly reproducible, The
HPLC elution profiles of the mycobactins produced by four species of non-chromogenic mycobacteria, namely M . smegmatis NCIB 8548, M .fortuitum NCTC 10394, M . chitae M393 and M .
phlei NCIB 8573, are shown in Fig. 2. The elution profiles of each of the other mycobactins were
equally distinctive; copies of the profiles have been deposited with the British Library Lending
Division, Boston Spa, Yorkshire LS23 7BQ, UK, as Supplementary Publication No. SUP 28016
(24 pages). (Copies may be obtained from the BLLD on demand; wherever possible, requests
should be accompanied by prepaid coupons, held by many university and technical libraries and
by the British Council.)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
1888
R . M . HALL A N D C. RATLEDGE
Table 4. HPLC of mycobactins
Separation was as described in Methods employing a Lichrosorb RP18 column and methanol/
water (80: 20, v/v) gradient over 30 min to methanol alone. Only those peak areas > 39/d of final
chromatograph shown. Reproductions of mycobactin profiles for each species are deposited as
supplementary data (see Results).
Species
Retention time (s) of mycobactins (peak area, %)
M . aurum
M 397
M .chitae
825
(5-9)
315
JS41
I95
(5.1) (10.4)
1289
JS423
I084
(3.0)
(3.3)
1020
M164
831
(7.0) (71.9)
663
M410
537
(28.6) (5-0)
449
M411
299
(6.4) (5.3)
326
M413
273
(4.2) 13.6)
306
M414
273
(5-5) (8.8)
NCTC 10394 474
523
(7-1 ) ( 3-0)
M416
728
875
(5-6) 12.8)
604
M79
863
(4.2) (21.6)
1026
JS278
742
(10.3) (80-0)
M403
1340 1510
(3.5) (3.6)
NCTC 16575 389
614
(19.3) (3.4)
NCIB 8573
1083 1322
(5.8) (9 1 6)
NCIB 8548
767
1065
(5.3) (41.3)
M407
1269 1389
(36-0) (6.4)
M302
I371
(96.7)
M. chelonae
subsp. abscessus
'M. diernhqferi'
M .duvalii
M.flacescens
M .flarescens
M ..@aceseem
M . Jarescens
M . fortuitum
M . gadium
'M .
gallinarum'
M . neoaurum
M .parafortuirum
'M .
peregrinum'
M .phlei
M . smegmatis
M . thermoresistible
M . t'accae
M393
71 1
974
( 19.4) (71.4)
758
( 1 8.6)
1247
(3.1)
1028
(56-0)
400
(5.1)
1533
(85-1)
1199
(14-6)
724
(28.7)
554
(9.0)
505
(4.5)
454
(3.2)
1014
( I 1.5)
925
(42. I )
1080
(3.7)
1085
(4.2)
1712
(77.2)
877
( 1 9.8)
1289 1524
(8-6) (6.0)
926 1034 1357
(16-4) (50.0) (13.5)
1357
(4.0)
1857
91 7
(24.6) (5.3)
757
642
847
909 1021
707
(28.0) (6.9) (9.6) (23.3) (5.6) (3-9)
654
81 1
879
1023 1273
755
(8.2) (22.9) (7.3) (8.4) (21.2) (3.5)
502
763
71 1
994
1047
(5.0) (1 4.0) (23.6) (1 5.0) (20.0)
1065 1279 1320 1742
(6.2) (34.6) (31.8) (4.1 )
1068 1114 1278
(7.3) (28.4) (3-4)
1129
(67.6)
1293
(4.5)
2273
(4.6)
1131 1173
(40.8) (13.6)
1130 1343
(4.0) (31.5)
1499
(52.1)
1587
(7.5)
Examination of strains failing to produce a mycobactin under conditions of iron-limitation
As shown (Table 3), some strains failed to yield a detectable mycobactin following conventional methods of growth under iron-limiting conditions although, surprisingly, some did
produce these compounds when grown on glucose/yeast extract/agar instead of glycerol/asparagine/agar. Those strains which remained mycobactin-negative were subsequently grown in
larger quantities (see Methods). Mycobacterium chelonae subsp. abscessus JS41 and M . uaccae
M302 each yielded a trace ( - 10 pg) of mycobactin from 7-5 g and 12-5 g of dry cells,
respectively, allowing TLC and HPLC to be carried out. Of the remaining non-producers, M.
uaccae JS R877R was examined more closely. Cells (about 25 g dry wt) harvested from
glycerol/asparagine/agarfailed to yield a mycobactin which was detectable by TLC (detection
limit 0.26pg) or HPLC (detection limit about 1 ng). Additionally, cells were harvested and
extracted under conditions to exclude the introduction of iron, and the ensuing desferrimycobacDownloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
1889
M ycobact ins
(a) M . smegmatis
Y
365
.d
(b) M.fortuitum
NCTC 10394 11279
c
5
.d
a
EM
igH]
80%
MeOH
2
5
5
-
d.
1200
Time (s)
600
( c )M . chifue
2
600 1200 180
Time (s)
1800
I
NCIB 8573
Y
2
0
L
bn
L
u
1083
600
1200 1800
600 1200 1800
Time (s)
Time (s)
Fig. 2. High-performance liquid chromatograms (Spectra-Physics SP 8000) of mycobactins isolated
from (a) M.smegmatis NCIB 8548,(b) M.forruirum NCTC 10394, (c) M . chitue M393 and (d) M . phlei
NCIB 8573. Separation of the mycobactins was achieved using a reverse-phase column (250 x 4 mm
Lichrosorb RP18) employing a methanol/water gradient from 80% to 100% methanol over 30 min. A
flowrate of 2 ml min-' and column temperature of 50 "C were maintained throughout. Peaks are
labelled with their retention times in seconds; quantitative data are given in Table 4.
tin (if any) then labelled with 5sFeCl, of specific activity 4.25 pCi (pg Fe)- (157 kBq pg- l ) .
When TLC was carried out on such an extract from 2.5 g cells, radioactivity corresponding to
16 ng mycobactin (g dry wt)- was calculated to be present. If a bacterial cell weighs 2.8 x
g (Ingraham et af., 1983), it can be calculated that the amount of mycobactin present is
about three molecules per cell. Mycobactin is thus either absent from this strain and the
recovered radioactivity regarded as spurious or else mycobactin is deduced to be present but
with its synthesis almost entirely repressed. With such a low content of mycobactin, its presence
in this strain of M.vaccae must be regarded as highly dubious.
DISCUSSION
Rapidly growing mycobacteria are widespread in their distribution, and, although medically
less important than their slower-growing relatives, are known to be occasional pathogens of
man, birds and several cold-blooded species (Wolinsky, 1979). For example, over the last few
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
I890
R. M. HALL A N D C. RATLEDGE
years there has been an increase in serious human infections following open-heart surgery and
haernodialysis associated with M . fortuitum and M . chelonae strains (Jauregui et a / . , 1977).
Additionally, many fast growers such as M. caccae, M. duvalii and M . gilvum are regularly found
in clinical material, though such strains are generally considered not to cause disease. The introduction of new biochemical, chemical, genetical and numerical methods over the last two
decades has revolutionized our approach to the identification and classification of bacteria
(Goodfellow & Board, 1980). The mycobacteria have been separated from related genera by
examination of their cell wall components, particularly the structures of their mycolic acids and
other complex lipids (Minnikin & Goodfellow, 1980). Division of Mycobacterium at the species
level has been achieved mainly by biochemical and immunological tests enumerated in a series
of collaborative studies carried out under the aegis of the International Working Group on
Mycobacterial Taxonomy (I WGMT) (see Goodfellow & Wayne, 1982; Goodfellow & Cross,
1984). There is little evidence that the genus should be divided into sub-genera although, in
practical terms, the genus is divided into two major groups composed of the slow-growing and
fast-growing mycobacteria.
In this paper we have examined the mycobactins from 14 of the 16 recognized fast-growing
species (Skerman et al., 1980), the two exceptions being M. senegalense, associated with bovine
farcy (Chamoiseau, 1973), and M. gilcurn. There are, however seven or eight other species which
may be eventually approved as good species (Goodfellow & Cross, 1984); of these, we examined
only 'M.
diernhoferi', though we included in the survey several species of doubtful status: 'M.
gallinarum'. ' M . kanazawa', and 'M . peregrinum'.
The mycobactins were examined by TLC and HPLC. The former procedure separates mycobactins by virtue of their differences in structure in the R 2 , Rj,R, and RS nuclear substituents
(Fig. l), whereas HPLC separates them mainly according to length of the long alkyl chain
attached to the molecule (Fig. 1, R, position). For rapid identification by those diagnostic
laboratories wishing to adopt this technique the former procedure, using only two (I and IV) of
the four solvent systems is sufficient to characterize a mycobactin from an unidentified
organism. HPLC need only be used when further discrimination is needed. Thus a simple and
rapid method is available for mycobacteriai characterization and identification.
Examination of the mycobactins from the M. fortuitum group ( M . fortuitum, M . chefonae
subsp. abscessus and ' M . peregrinum') suggested that each was a different molecule and thus
the three taxa could be deduced to be distinct. Whilst the species status of the former two
organisms does not appear to be in doubt, the status of 'M. peregrinum' is questionable. Some
workers (Kubica et a!., 1972; see also Goodfellow & Cross, 1984) have considered 'M.
pcregrinurn' to be synonymous with M.jurtuitum, while others (Jenkins et al., 1971 ; Pattyn et al.,
1974; Baess, 1982) have continued to distinguish this organism as a separate species.
The M.parafortuitum complex, which comprises M . parafortuiturn, M. aurum, ' M .diernhoferi',
M . neoaurum and M . uaccae, has a somewhat confused taxonomy. We examined representative
strains of each of these five species and in addition three strains designated 'M. kanazaMu'
(Kanazawa et al., 1967)' which has been considered as part of this complex (Saito et al., 1977).Of
all the groups of mycobacteria we examined, this proved the most difficult in which to detect the
presence of a mycobactin. Snow (1970) recorded similar difficulties with M. aurum for which
over 10 1 of culture were needed to produce a small quantity (0-7 mg per g cell dry wt) of
mycobactin. However, a mycobactin was recovered from at least one strain of each of the five
species on the approved list; none unfortunately could be detected from the three strains of 'M.
kanazawa' examined. The five mycobactins could all be distinguished from each other by HPLC
and our conclusion that this should then indicate separation of the taxa would concur with the
previous conclusions of some other groups of workers (Tsukamura et al., 1981 ; Baess, 1982). The
co-operative study of the IWGMT on this complex of organisms (Saito et a / . , 1977) concluded
that there was a high overall similarity between strains of M. aurum, 'M. diernhoferi', M .
neoaurum and M . parafortuitum and that new discoveries were needed to allow a more definite
classification to be made. Perhaps the differences in mycobactin structures could help to clarify
the systematics of this group.
MycobacteriumJlavescens. M . phlei and M . thermoresistible are thermotolerant species which
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
M ycobactins
1891
are taxonomically distinct ; the diferences between the mycobactins from representative strains
of the three species would agree with this. However, M.Javescens is itself thought to be a heterogeneous taxon (Jenkins et al., 1972; Kubica et al., 1973). Tsukamura & Mizuno (1975)
recognized two subgroups: M. jlavescens itself and ‘ M . galZinarum’. Our findings with the
mycobactins from four strains of M.flavescens and two of ‘M. gallinarum’ would clearly indicate
heterogeneity. The mycobactins from both strains of ‘ M .gallinarum’ appeared, by both TLC and
HPLC, to be similar if not identical but these clearly differed on HPLC from those of the four M .
flavescens strains examined. In turn, the M.Juvescens strains produced mycobactins which were
obviously distinct from each other (see Table 4), at least four separate mycobactins being
recognized. Whether this warrants division of M. fzavescens into further species or subspecies
must await the results of other taxonomic studies.
Characteristic mycobactins, differentiable from each other and from other mycobactins, were
obtained from the remaining mycobacteria examined in this study: M. chime, M. smegmatis, M .
duvalii and M . gadium. Our findings would support each being a distinct taxon including M.
chitae (Goodfellow & Wayne, 1982).
The apparent absence of mycobactin in some species (see Table 3), even when grown under
conditions of strict iron-limitation, or on an alternative medium besides glycerol/asparagine/
agar, may be attributed to a number of factors. The organism may have an alternative iron
incorporation system which satisfies its requirement for the metal. Secondly, the usual habitat of
the organism may provide a ready supply of the metal and provision of a storage mechanism for
iron may thus be unnecessary. Mycobactins may not be produced by some organisms owing to
the loss or complete absence of the required biosynthetic machinery. Finally, biosynthesis of
mycobactin may be strongly repressed and mycobactin may be produced only in minute
amounts, even under conditions seemingly conducive for the derepression of mycobactin
synthesis. Massive quantities of cells grown in iron-deficient conditions would therefore be
required for their detection and isolation. Such an organism may need special conditions or
specific concentrations of iron to derepress mycobactin synthesis. The iron metabolism of such
apparent non-producing strains is currently under further examination. It is apparent, however,
that the putative absence of a mycobactin from an organism cannot be interpreted as that
organism not being a mycobacterium.
The mycobactins have an important role to play in the identification of mycobacteria at the
species level. Whether the presence of a unique mycobactin molecule in an organism should be
indicative of that organism being regarded as a separate species remains to be established. The
data presented in this paper illustrate the powerful discriminatory power of the mycobactin as a
chemotaxonomic character. If this approach can be allied with other equally modern
discriminatory techniques, the way becomes open to characterize the species in a very accurate
manner. Examination of additional strains may well show that mycobactin composition should
be included in the minimal description of certain mycobacterial species. At the very minimum,
the mycobactin should be recognized as creating distinct chemotaxa of the Mycobacterium
genus.
We wish to thank Mrs Janet Stephenson for assistance with media preparation and also colleagues who kindly
provided strains. R. M. H . gratefully acknowledges the support of the Medical Research Council.
REFERENCES
BAESS, I. (1982). Deoxyribonucleic acid relatedness
among species of rapidly-growing mycobacteria.
Acla pathologica, microbiologica et immunologica
scandinacica 90, 371 -375.
G . ( 1973). MycobacreriumfarcinogenesT
agent causal du farcin du boeuf en Afrique. Annales
de microbiologie I U A , 2 15-222.
GOODFELLOW,
M. & BOARD,R . G. (editors) (1980).
Microbiological Classification and Identification. London : Academic Press.
CmMOISEAU,
GOODFELLOW,
M. & CROSS,T. (1984). Classification.
In The Biology of the Actinomycetes, pp. 7-164.
Edited by M. Goodfellow, M. Mordarski & S. T.
Williams. London : Academic Press.
GOODFELLOW,
M. & WAYNE,L. G. (1982). Taxonomy
and
In The Biology of the Mycobacreria, pp. 471-521. Edited by C. Ratledge & J .
Stanford. London: Academic Press.
GORDON,
R. E. & MIHM,J . M. (1962). Identificationof
Nocardia curiae (Erikson) nor. comb. Annals of the
New York Academj. of’ Sciences 98, 628-636.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43
r 892
R. M. H A L L A N D C . RATLEDGE
HALL,R. M. & RATLEDGE,C. (1982). A simple method
for the production of mycobactin, the lipid soluble
siderophore, from mycobacteria. FEMS Microbiology Latters 15, 133-1 36.
HEDREN,R. W. (1974). Identification of Mycohcterium phlei ATCC 356 as a strain of Mycobacterium
smegmatis. InternationalJournal of Systematic Bacteriology 24, 491-492.
INGRAHAM, J . L., MAALBE, 0. & NEIDHARDT,F. C.
(1983). In Growth of the Bacterial Cell, pp. 38.
Massachusetts : Sinaver Associates.
JAUREGUI, L., ARBULU, A. & WILSON, F. (1977).
Osteomyelitis, mediastinitis and vasculitis due to
Mycobacterium chelonae. American Reciew of Respiratory Disease 115, 699-703.
JENKINS, P. A., MARKS,
J. &SCHAEFER, W. B. (1971).
Lipid chromatography and seroagglutination in the
classification of rapidly-growing mycobacteria.
American Review of Respiratory Disease 103, 179187.
JENKINS P. A., MARKS,J . & SCHAEFER,
W. B. (1972).
Thin-layer chromatography of mycobacterial lipids
as an aid to classification. The scotochromogenic
mycobacte ri a, including Mycobacterium scrofulaceum, M . xenopi, M . aquae, M . gordonae, M .
flar;escen.s. Tubercle 53, 118-127.
M.
KANAZAWA,
Y., KURAMATA,T. & TSUKAMURA,
(1967). Assay method for microbiologically active
isoniazid and ethambutol levels in body fluids by the
agar plate diffusion technique using M . parafortuiturn as a test organism. Kekkaku 42, 517-521.
KUBICA,G . P., B m , I., GORDON, R. E., JENKINS,
P. A., KWAPINSKI,3. B. G . , MCDURMONT, C.,
PAWN, S. R., SAITO,H . , SILCOX,
V., STANFORD,J.
L., TAKEYA, K. & TSUKAMURA,
M. (1972). A cooperative numerical analysis of rapidly growing
mycobacteria. Journal of General Microbiology 73,
55-70.
KUBICA,G . P . , SILCOX,V. A . & HALL, E. (1973).
Numerical taxonomy of selected slowly growing
mycobacteria. Journal of General Microbiology 74,
I 59- 167.
MINNIKIN,D. E. & GOODFELLOW,M. (1980). Lipid
composition in the classification and identification
of acid-fast bacteria. In Microbiological Classification
and Identification, pp. 189-256. Edited by M.
Goodfellow & R. G. Board. London: Academic
Press.
PATTYN,S. R., MAGNUSSON,M., STANFORD,J. L. &
GRANGE,J . M. (1974). A study of Mycobacterium
fortuitum (ranae).Journal of Medical Microbiology 7 ,
67-76.
RATLEDGE,
C. (1982~).Mycobactins and nocobactins.
In The CRC Handbook of Microbiology 2nd edn, vol.
IV, pp. 575-581. Edited by A. I. Laskin 8~H. A.
Lechevalier. Boca Raton, Florida : C R C Press.
RATLEDGE,C. (19826). Nutrition, growth and metabolism. In The Biology of the Mycobacteriu, vol. 1, pp.
185-271. Edited by C. Ratledge & J . L. Stanford.
London : Academic Press.
RATLEDGE,C. & EWING,D. F. (1978). The separation
of the mycobactins from Mycobacterium smegmatis
by using high-pressure liquid chromatography. Biochemical Journal 175, 853-857.
RATLEDGE,
C. & HALL,M. J . (1971). Influence of metal
ions on the formation of mycobactin and salicylic
acid in Mycobacterium smegmatis grown in static
culture. Journal of Bacteriology 108, 31 2-319.
SAITO,H., GORDON,R. E., JUHLIN, I., KAPPLER,W.,
KWAPINSKI,J . B. G . , MCDURMONT,C . , PATTYN,
S . R., RUNYON,E. H., STANFORD,J. L., TARNOK,I.,
TASAKA,H., TSUKAMURA,
M. & WEISZFEILER,
J.
(1977). Co-operative numerical analysis of rapidlygrowing mycobacteria. International Journal of Systematic Bacteriology 27, 75-85.
SKERMAN,V. B. D., MCGOWAN,V. &SNEATH,P. H. A.
(1 980). Approved lists of bacterial names. International Journal of Systematic Bacteriology 30,225420.
SNOW, G. A. (1970). Mycobactins: iron-chelating
growth factors from mycobacteria. Bacteriological
Reviews 34, 99- 125.
SNOW, G . A. & WHITE, A. J. (1969). Chemical and
biological properties of mycobactins isolated from
various mycobacteria. Biochemical Journal 115,
1031- 1045.
TSUKAMURA,
M. & MIZUNO,S. (1975). Differentiation
among mycobacterial species by thin-layer
chromatography. International Journal of’Systematic
Bacteriology 25, 27 1-280.
TSUKAMURA,
M., MIZUNO,S. & TSUKAMURA,
S. (1981).
Numerical analysis of rapidly-growing, scotochromogenic mycobacteria, including Mycobacterium
obuense sp. nov., nom. rev., Mycobacterium rhodesiae
sp. nov., nom. rev., Mycobacteriwn aichiense sp. nov.,
nom. rev., Mycobacterium tokaiense sp. nov., nom.
rev. International Journal of Systematic Bacteriology
31, 263-275.
WELLINGTON,E. M. H . & WILLIAMS,S. T. (1978).
Preservation of actinomycete inoculum in frozen
glycerol. Microbws Letters 6, 151-1 57.
WHITE,A. J. &SNOW, G . A. (1969). Isolation of mycobactins from various mycobacteria. The properties
of mycobactins S and H. Biochemical Journal 111,
785-795.
WOLINSKY,E. (1979). Nontuberculous mycobacteria
and associated diseases. American Reuiew of Respiratory Disease 119, 107-1 59.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 03:16:43