Microbio/ogy (1996), 142, 15 13-1 520
Printed in Great Britain
ExtraceIIular and surface-exposed
polysaccharides of non-tuberculous
mycobacteria
Anne Lemassu, Annick Ortalo-Magne, Fabienne Bardou, Gaby Silve,
Marie-Antoinette Laneelle and Mamadou Daff
Author for correspondence: M. Daffe. Tel: +33 61 33 59 16. Fax: +33 61 33 58 86.
e-mail: [email protected]
lnstitut Pharmacologie et
de Biologie Structurale du
CNRS and UniversitC Paul
Sabatier, 118 route de
Narbonne, 31062 Toulouse
CCdex, France
We studied the outermost constituents of the cell envelopes, which are
involved in the interaction between the bacilli and the host cells, of five
pathogenic and non-pathogenic mycobacterial species for comparison with
those we have previously characterized from M. tuberculosis. The extracellular
materials (ECMs) were isolated by ethanol precipitation and compared to the
surface-exposed materials (SXMs) extracted by mechanical means. The
materials from both sources were composed almost exclusively of
polysaccharides and proteins. Two groups of mycobacteria were clearly
distinguishable. The first group comprised the pathogenic species M. kansasii
which produced large amounts of ECM, the glycosyl composition of which was
similar to that of the SXM. The second group comprised M. avium and the nonpathogenic strains of M. gastri, M. phlei and M. smegmatis which produced
small amounts of ECM. This latter group could be subdivided into those which
produced carbohydrate-rich ECM (M. avium and M. gastrr')and those forming
protein-rich ECM (M. phlei and M. smegmatis), a classification that correlated
with the difference in the growth rate of the two subgroups. The glycosyl
composition of the ECM of a given species was qualitatively similar to that of
the SXM, except for M. avium and M. phlei whose S X M were devoid of
arabinose. In addition to glucose, mannose and arabinose, xylose was detected
in the hydrolysis products of the ECM and SXM of M. smegmatis, the SXM
of M. phlei and the ECM of some batches of M. avium. The polysaccharide
constituents of the ECM and S X M of the different mycobacteria were purified
by anion-exchange and gel-filtration chromatography; all were found to be
neutral compounds devoid of acyl substituents. The extracellular
polysaccharides consisted of high-molecular-mass glycogen-like glucans,
arabinomannans and mannans, structurally similar to the corresponding
substances previously characterized from the capsule of M. tuberculosis. The
same types of polysaccharides were characterized from the S X M of all the
strains, except M. avium and M. phlei which were devoid of arabinomannans.
This study questions the unique and universal representation of the
mycobacterial cell envelope and the existence of the so-called acidic
polysaccharide-rich outer layer.
i
Keywords : arabinomannan, cell envelope, glucan, mycobacteria, polysaccharides
INTRODUCTION
Among the infectious diseases, those resulting from
infections with Mycubacterimz species are important
sources of morbidity and mortality throughout the world
today. Almost 2 billion people have been infected with
Mycubacteriivm tdwidosis, the causative agent of tuberculosis, of whom 8 million develop active disease and
Abbreviations: AM, arabinomannan; ECM, extracellular material; SXM, surface-exposedmaterial; SmT, smooth transparent; SmD, smooth domed-opaque.
0002-0385 Q 1996 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36
1513
A. L E M A S S U and OTHERS
approximately 3 million people die each year from this
disease (Bloom & Murray, 1992; Kochi, 1991). Mycobacterial infections have also re-emerged as an important
public health problem world-wide, and this resurgence
has been accompanied by an increased incidence of
tuberculosis resistant to the standard antituberculosis
drugs. Furthermore, and as a result of the human
immunodeficiency virus epidemic, it has been learned that
HIV infection accelerates the natural history of new M.
ttlberczllusis infection (Chaisson, 1993) and that HIVinfected individuals are more susceptible to reactivation
of a previously dormant focus of the tubercle bacillus
(Chapman & Henderson, 1994), resulting in a high
incidence of tuberculosis. In addition, disseminated mycobacterial infections, notably those caused by members of
the so-called M. avitlm-intracellzllare complex, occur in a
large proportion of HIV-infected persons.
Mycobacteria are appreciably more resistant than other
bacteria to deleterious agents, such as acids, alkalis and
germicides (Youmans, 1979), a property commonly
attributed to the inner insoluble matrix of the mycobacterial cell wall which is invariable among species and is
composed of two covalently attached macromolecules,
peptidoglycan and mycoloyl arabinogalactan (Draper,
1982; McNeil & Brennan, 1991). Interestingly in the
context of the host-parasite interaction in mycobacterial
diseases, it has been observed that virulent mycobacteria
growing intracellularly are surrounded by a capsule,
observed as an electron-transparent zone by electron
microscopy (Hanks, 1961), which may be part of defence
mechanisms permitting all pathogenic mycobacteria to
resist being killed by phagocytic cells (Draper & Rees,
1970). More recently, it has been shown that in vitrugrown pathogenic species are also surrounded by a capsule
(Frihel e t al., 1988;Paul & Beveridge, 1994). Because the
outermost constituents of this protective capsule are in
direct contact with host cells, they may play an important
role in the phagocytosis process leading to the internalization of the pathogen. They may also determine, in
part, the nature and the intensity of the host response
towards the phagocytosed bacilli. As far as the chemical
nature of the mycobacterial outer layer is concerned,
however, information is generally scarce.
The outermost constituents of the bacilli are confined
around the bacterial cells when the environment happens
to be a macrophage, but may be released by actively
growing mycobacteria in artificial culture media, which
may also contain secreted substances. In the past few
years, numerous studies have been devoted to the
isolation, characterization and biological activities of the
proteins derived from early-exponential-phase culture
filtrates (Andersen & Brennan, 1994; Harboe, 1992;
Young e t al., 1992). However, little attention has been
paid to extracellular polysaccharides, despite the early
demonstration of the presence of polysaccharides in
tuberculin material derived from the culture filtrate of the
tubercle bacillus and used for skin testing (Seibert, 1949).
Because of the presence of arabinogalactan, a true cell wall
constituent (Draper, 1982; McNeil & Brennan, 1991), as
the major antigenic component of tuberculin (Daniel,
1514
1984), the impression prevails that the extracellular
polysaccharides are derived from the autolysis of bacterial
cells.
Previously, we have investigated the presence of polysaccharides in early-exponential-phase culture filtrate of
M. ttlberctlhis devoid of somatic and cell-wall-associated
compounds (Lemassu & Daffi, 1994). We showed that
the extracellular material was composed of proteins and
three types of neutral polysaccharides: a glucan, an
arabinomannan (AM) and a mannan. We also demonstrated that the same types of polysaccharides are present
at the surface of the bacterial capsule (Ortalo-Magni e t al. ,
1995), suggesting that these polysaccharides are capsular
constituents that are shed from growing cells into the
culture medium. As mycobacterial species differ in their
cell-wall-associated lipids, growth rate, drug susceptibility, solute permeability, antigenicity and pathogenicity,
the data obtained for the tubercle bacillus may not be
universal. Furthermore, we showed recently that mycobacterial species differ in their surface-exposed lipids
(Ortalo-Magni e t al., 1996b). In this study, we report the
comparison of the carbohydrate content of the extracellular materials (ECMs) and surface-exposed materials
(SXMs) of selected mycobacterial species and the chemical
characterization of the major polysaccharides that these
materials comprise.
Strains and growth conditions. M. avzum ATCC 15769 and
TMC 1468, M. gastri W471, M. Aansasii ATCC 12478, IP 890175
and IP 890370, M. pbled ATCC 11758, M. smegmatis ATCC 607
and M. tuberculosis strain Canetti (CIPT 140010059), its spontaneous rough revertant (59R, Lemassu et al., 1992) and the
Canetti-like strain CIPT 140010060 were grown on Sauton's
medium (Sauton, 1912) (100 ml per flask) as surface pellicles at
37 "C. Smooth transparent (SmT) and smooth domed-opaque
(SmD) colonial variants of M. avium strain 485 (Fattorini e t al.,
1994) were grown on Middlebrook 7H9 plus glycerol for 10 d
at 37 OC as described by Fattorini e t al. (1994). The spontaneous
reversion frequency of SmT into SmD, a common occurrence in
liquid medium, was estimated as 7 % by observing the colony
morphology on Middlebrook 7H10 agar plates.
Production of ECM during growth. Experiments were performed with bacteria from three independent batches throughout the growth phases. Cells were harvested from 100 ml culture
flasks by filtration through a Durieux filter and the corresponding extracellular materials were obtained by filtering the
culture medium through a 0 2 pm sterile Filter (Nalgene). They
were concentrated under vacuum to one-tenth of their original
volume, dialysed, lyophilized and weighed. Cells were extensively washed with distilled water, lyophilized and weighed.
Isolation of SXM. The pellicular growth allowed an easy harvest
of cells by permitting the medium to be poured off when the
pellicles remained attached to the flasks. Cells were harvested
from 10-20 flasks and gently shaken (1000 r.p.m.) for 1 min
with 10 g glass beads (2 g wet cells)-', a method known not to
affect the integrity of cells (Ortalo-Magnt e t al., 1995). The
resulting SXMs were then resuspended in distilled water (50 ml
per flask) and immediately filtered through a 0 2 pm sterile Filter
(Nalgene).
Fractionation of ECM and SXM. Filtrates (derived from the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36
Polysaccharides of non-tuberculous mycobacteria
extracellular medium and from the mechanical treatment of
cells) were concentrated under vacuum to one-tenth of the
original volumes ;chloroform and methanol were then added to
the filtrates to obtain a partition mixture comprising chloroform/
methanol/water (3 :4: 3, by vol.). The organic phases were dried
and weighed. The glucans were recovered as opalescent
solutions by extracting the interphases several times with water
(Lemassu & Daff6, 1994). The aqueous phase and the glucan
solution were concentrated and the polymers were precipitated
overnight at 4 "C with 6 vols cold ethanol. The precipitates were
collected after centrifugation at 14000 g for 1 h and dissolved in
distilled water. The polymers were then precipitated again with
ethanol, re-centrifuged and the precipitates dialysed for 3 d
against water to eliminate traces of glycerol and salts before
being 1yophilized and weighed.
Purification of the different polysaccharides. The polymers
derived from the aqueous phases were dissolved in 50 mM
NH,HCO,, treated twice with trypsin (2 %, w/w, of the protein
content) for 5 h at 37 "C and dialysed for 3 d against distilled
water. The resulting material was assessed for the presence of
protein, which usually did not exceed 2 %. When the percentage
of protein exceeded 2 % , however, the treatment with trypsin
was repeated. Then, the different polysaccharides were purified
as previously described in detail (Lemassu & Daffk, 1994).
Briefly, Ara/Man-containing polysaccharides and Glc-rich polysaccharides were chromatographed on a column of DEAETrisacryl gel and the neutral fractions were rechromatographed
on columns of Bio-Gel P-10 and Sephadex G-200, respectively.
The elution profiles of both types of gel filtration chromatography were monitored by refractive index detection and the
collected fractions were assessed for their carbohydrate content
(Dische, 1962). The glycosyl composition of pooled fractions
corresponding to the different chromatography peaks was
determined by acid hydrolysis (2 M CF,COOH at 110 "C for
2 h), followed by analysis of the trimethylsilylated sugar
derivatives by GC.
5263 Hz) and 2-73 s for 'H-NMR (spectral width : 2994 Hz) was
used. The chemical shift reference used was that of tetramethylsilane.
RESULTS
Production of ECM during growth
Growth curves and measurement of ECM derived from
100 m l culture filtrates and produced during the different
growth phases of the various mycobacterial strains were
performed on bacterial cells which were extensively
washed, dried and weighed, as all the mycobacterial
species used in the present study formed clumps when
40
looo
I
10
I
20
20
-20
-F
W
Analytical techniques. The dialysed culture filtrates and the
various chromatographic fractions were assessed for the presence of protein (Lowry method) and carbohydrate (Dische,
1962). The percentage of carbohydrate was also determined by
GC using erythritol as internal standard.
Pure polysaccharides (2 mg) were per-0-methylated four times
according to the method of Ciucanu & Kerek (1984). Portions
of the per-0-methylated products were hydrolysed with 2 M
CF,COOH at 110 OC for 2 h, reduced with NaBH, and
acetylated. The different partially 0-methylated and partially 0acetylated alditols were identified by GC-MS.
I
30
- 10
1000
z
Y
30
GC-MS was performed on a Hewlett-Packard 5890 gas chromatograph connected to a Hewlett-Packard 5989A Mass Selective
Detector. The column was a 12 m HP-1 (Hewlett-Packard).
Samples were injected in the Splitless mode. The oven
temperature was programmed to hold at 80 OC for 1 min
followed by a 15 "C min-l increase to 290 "C. The mass
spectrometer was set to scan from 50 to 600 atomic mass units.
GC was performed on aldtol acetates and trimethylsilyl
derivatives of monosaccharides using a Girdel chromatograph
(model G 30) equipped with a fused silica capillary column
(25 m length x 0-22 mm i.d.) containing WCOT OV-1 (0-3mm
film thickness, Spiral). A temperature gradient of 100-280 "C at
2 "C min-' was used.
NMR spectra were obtained in D,O with a Bruker 250 WB
instrument equipped with an Aspect 3000 computer. Onedimensional spectra were obtained with 40" pulses and an
acquisition time of 1.56 s for 13C-NMR (spectral width:
Time (d)
Fig. 1. Production of ECM (+) during the growth of M.
tuberculosis (a), M. smegmatis (b) and M. avium (c). Cells were
harvested, extensively washed, lyophilized and weighed. Cell
is plotted on a log scale. ECMs were obtained
dry weight
by concentration of 100 ml culture filtrates, followed by
extensive dialysis and lyophilization as described in Methods.
Results represent the arithmetic mean of three determinations.
(a)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36
1515
A. LEMASSU a n d O T H E R S
Table 1, Composition of ECM and SXM from different mycobacterial species
ECM and SXM were obtained by ethanol precipitation of culture filtrates and from materials extracted
by mechanical treatment of cells with glass beads, respectively, followed by dialysis and lyophilization
as described in Methods. Values are the mean of at least three independent experiments conducted on
three different batches. ND,not detected.
Strain
M. gastri
M. hnsasii
M. avium
M. smegmatis
M. phiei
Component
ECM
SXM
ECM
SXM
ECM
SXM
ECM
SXM
ECM
SXM
Yield*
1.0 f0-1
16.0 f0 2
0.9 f0.1
2.5
1.0
3.0 f0-5
Carbohydrate Protein
("/)
("/I
71
95
84
92
48
86
26
27
21
17
28
4
16
7
51
14
74
73
79
83
Sugar composition (%)
Glc
Man
Ara
Xyl
71
96
82
87
26
94
32
28
40
5
17
3
10
8
44
6
17
11
42
5
12
1
8
5
13
ND
16t
ND
ND
16
13
17
ND
ND
ND
ND
36
47
1
90
* Yield is expressed as mg of materials obtained from early-exponential-phase culture filtrates (as described
in Methods) per 100 mg of the corresponding mycobacterial dry weight fSD.
t The relative amount of Xyl may vary significantly according to the batches and the strains examined (see
text).
growing both as surface pellicles and under stirring.
Furthermore, the conventional addition of detergent to
the cultures to prevent the formation of clumps during
growth, was not suitable, as this method releases additional materials from the cells into the extracellular
medium. Thus, the culture filtrates corresponding to the
various growth phases of a slow grower (M. avium), a
rapid grower (M. smegmatis) and M . tz/bercuhis (used as
reference) were collected, submitted to ethanol precipitation, dialysed and weighed to yield the crude unfractionated ECM. As illustrated in Fig. 1, two types of curves
were obtained : while the production of ECM followed
the growth curve of M. tuberculusis (Fig. la), an invariable
and low level of ECM was observed for both M. avium and
M . smegmatis (Fig. 1b,c). To compare the ECM production
of the different species to that of M. tuberculosis (Lemassu
& Daffi, 1994), the ECMs of the different strains were
collected from early-exponential-phase growth culture
filtrates. As shown in Table 1, the ECMs of M. avium, M.
gastri, M. phlei and M. smegmatis corresponded to only
1-3 % of the dried bacterial mass. In contrast, high levels
of ECM were produced by M . kansasii, comparable to that
previously found for M. tubercdosis (Lemassu & Daffi,
1994). These data demonstrated the existence of high level
and low level producers of ECM among mycobacteria.
Chemical nature of ECM and SXM
Small amounts of SXM were recovered from the bacilli by
mechanical treatment with glass beads. They represented
10-14% of the mass of the corresponding ECM. When
chloroform/methanol/water partition was applied prior
to ethanol precipitation, the organic phases derived from
1516
the extraction of the culture filtrates of the different
mycobacterial strains contained little (< 1 %) or no lipids.
Small amounts of lipids (1-5%) were present in the
organic phases derived from the partition of SXM.
The unfractionated ECM and SXM of the various
mycobacterial strains were composed mainly ( 295 YO)
of
carbohydrates and proteins. According to the carbohydrate and protein contents of ECM and SXM (Table l),
the mycobacterial species examined may be divided into
three groups : the materials derived from M. gastri and M.
kansasii contained much more carbohydrate (up to 95 %)
than protein, whereas the reverse was true for M.phZei and
M. smegmatis. M. avium was unique in that the examined
strain contained carbohydrate-rich SXM and ECM composed of equal amounts of protein and carbohydrate
(Table 1).
The monosaccharide composition of ECM and SXM,
determined by acid hydrolysis of the polymers, was found
to be qualitatively invariable during the exponential
growth phase of all the examined strains of M. hnsasii.
The ECM and SXM of M. kansasii and those of M .
smegmatis were of similar composition but distinguished
from each other by the presence of Xyl in the materials
derived from M. smegmatis. Although the ECM and SXM
of M. gastri were composed of the same sugar constituents,
much less Ara and Man were present in SXM (Table 1).
In contrast to these mycobacterial species, the sugar
composition of the ECM of M. avium and M. phlei differed
from that of the SXM. The glycosyl composition of the
ECM of M. avitlm varied between batches. Xyl was
detected in the acid hydrolysis products of the ECM in
most of the batches of the two examined strains of M .
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36
Polysaccharides of non-tuberculous mycobacteria
Q U ~ ~ W V but
in some other batches this sugar was not
detected at all. The ECM from M. azhm differed also from
the other ECM in its content of Man, which was the major
sugar in the acid hydrolysis of all the ECM derived from
the different batches of the two strains of M. auim~.In the
case of M. #lei, a tiny amount of Xyl was present in the
ECM, whereas this sugar represented the major constituent of the SXM. The two species were also unusual in
the glycosyl composition of their SXM which was
invariably found to be devoid of Ara (Table 1). It is
interesting to note that the acid hydrolysis products of the
ECM and SXM derived from the SmT and SmD variants
of M. avitl9n, the two well-studied morphotypes, showed
similar glycosyl compositions, including the absence of
Ara in SXM (data not shown).
0.4
0.2
m
w
Purificationand characterization of the major
polysaccharides
Over 90% (by wt) of the water-soluble Ara/Mancontaining polysaccharides from the extracellular and
surface-exposed materials of the different mycobacterial
species examined were neutral compounds, since they
were eluted from the DEAE-Trisacryl column with the
buffer containing no NaC1. Chromatography of these
neutral polysaccharides on Bio-Gel P-10 columns using
1 % (v/v) acetic acid in water as eluant gave qualitatively
similar profiles for the various species, leading to the
isolation of two major fractions (Fig. 2a). The first major
peak (Ve
= 65 ml), exhibiting an apparent molecular
mass of approximately 13 kDa (using dextrans as reference
compounds), contained AMs (Lemassu & Daffi, 1994).
The second major fraction (Ve
= 90 ml) consisted of
mannans with an apparent molecular mass of 4 kDa. The
minor peak eluted in a position corresponding to the void
volume ( Vo,
molecular mass 2 20 kDa) ; acid hydrolysis
showed the presence of Ara, Man and Glc. Based on our
previous results (Lemassu & Daffk, 1994), it was concluded that the latter chromatographic peak contained a
mixture of AMs and mannans, and the small portion of
glucans that had been dragged in the aqueous phase
during the partition experiment.
The glucans from both sources were not retained on the
anion-exchange columns. When analysed in Sephadex G200 columns (Fig. 2b), the polysaccharides were eluted at
a position (Ve= 70 ml) corresponding to an apparent
molecular mass of 120 kDa.
The Xyl-containing carbohydrates, eluted in the last
fraction of the Bio-Gel P-10 column ( Ve= 130 ml) were
not investigated further.
Structural features of the major polysaccharides
Aliquots of the purified AMs from ECM and SXM were
per-O-methylated, hydrolysed, reduced, O-acetylated and
analysed by GC-MS. The arabinan segments consisted of
terminal, 2-, 3,5- and 5-linked Araf residues. The mannan
segments contained terminal, 6- and 2,6-linked pyranosides since all the mannosyl residues, after per-0methylation, contained a methoxyl group located on
10
20
Fraction no.
Fig. 2. Gel filtration profiles of the extracellular polysaccharides
of mycobacteria. (a) Neutral water-soluble polysaccharides on a
column of Bio-Gel P-10 (90 cm x 1.8 cm); (b) glucan on a column
of Sephadex G-200 (90 cm x 1.8 cm). V,, void volume. The eluant
of the gel filtration columns comprised 1 % acetic acid.
Fractions, 6 ml in (a) and 11 ml in (b), were collected, and the
carbohydrate content of each fraction was determined as
described in Methods.
carbon 4. All the AMs, except that of M. smegmatks, also
contained 2-linked Man,. In addition, the 13C-NMR
spectra of the purified AMs (data not shown) were
superimposable on that of M. tzibercdosis (Lemassu &
Daffk, 1994), suggesting that the different purified AMs
may share the major structural feature that consists of a
[ + G)a-Man,(l+ ] mannan core, substituted by arabinan
segments composed mainly of [-+ 5)a-Araf(1-+ ]
(Lemassu & Daffk, 1994; Misaki e t d., 1977). No signal
attributable to acyl functions, between 0 and 30 p.p.m.,
was observed in the 13C-NMR spectra. Likewise, fatty
acyl proton resonance signals (observable between 0 and
3 p.p.m.) were absent from the 'H-NMR spectra. These
data demonstrated that the polysaccharides under study
were devoid of fatty acyl substituents and are consistent
with the identification of Man at the reducing end of the
corresponding molecules of M. tz4bercdosi.s (Ortalo-Magni
e t al., 1996a).
To address the question of the structure of the nonreducing ends of the arabinan segments of the various
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36
1517
A. L E M A S S U a n d OTHERS
Table 2. Glycosyl linkage composition of the AMs of
M. avium and M. smegmatis
Conditions for per-0-methylation of the polysaccharides,
hydrolysis, reduction, 0-acetylation and GC-MS analysis are
given in Methods. Proof of linkage and ring form are given in
Results. t, terminal; tr, traces.
Glycosyl
residues
t- Ara,
2-Araf
5-Araf
3,5-Araf
t-Man,
2-Manp
&Man,
2,6-Man,
Glycosyl linkage
composition (mol%)
M . avium
M . smegmatis
3
7
8
8
47
8
18
40
7
22
12
2
6
tr
6
5
AMs, the glycosyl linkage composition of the polysaccharides was examined. In mycobacterial arabinogalactans (Daff6 e t al., 1990, 1993) and uncapped lipoarabinomannan (Chatterjee e t al., 1991), the molar ratio of
terminal 8-Ara, to 3,5-linked Ara, was always approximately 1. The AM of M. smegmatis, but not that derived
from the ECM of M. auitlm, (Table 2) fulfill this criterion,
suggesting some differences in the terminal region of their
arabinan segments. The combined evidence of small
amounts of terminal Araf (as compared to 3,5-linkedAra,) and the excess of terminal Man, (as compared to
2,6-linked Man,) strongly suggests the substitution of the
non-reducing end of the arabinan segment of the AM of
M.avitlm by either a terminal Man or an oligomannoside.
This hypothesis is further substantiated by the occurrence
of 2-linked Man, in the AM of M. auiutla, but not in that
of M. smegmatis (Table 2). These data were in agreement
with the finding that 2-linked Man, residues are part of
the oligomannosyls that substitute the arabinan termini of
the so-called capped lipoarabinomannans (Chatterjee e t
al., 1992). Analysis of the glycosyl linkage composition of
the per-0-methylated AMs of M.kansasii and M.gastri
(data not shown) led to the conclusion that these
molecules were as extensively capped as those of M. avitlm
and M. ttlberctllosis (Lemassu & Daffi, 1994; OrtaloMagni e t al., 1995). However, the extent of capping with
mannosyl residues may vary according to the strains and
the growth phase of the same strain.
The glucans purified from the ECM and SXM of the
different species were also per-0-methylated and their
partially 0-methylated, partially 0-acetylated alditol derivatives were analysed by GC-MS. It appeared that the vast
majority of the glucosyl residues were either 4-linked
Glc, (or 5-linked Glc,), as previously found for those of
M. tarberctllosis (Lemassu & Daffk, 1994; Ortalo-Magnk e t
al., 1995); some branched 4,6- (or 5,6-linked) glucosyl
residues were also detected. To address the question of
1518
the anomeric configuration and of the ring form of the
glucosyl units, the purified glucans were subjected to 13CNMR analysis. The spectra were superimposable on that
of M. ttlbercarlosis (Ortalo-Magnk e t al., 1995). The anomeric resonance signal was seen at 101 p.p.m. which can
only result from the resonances of C-1s of m-Glcp, the C1 resonances of 8-Glc, and of cc- and P-Glc, being expected
at lower field (104-110 p.p.m., Bradbury & Jenkins,
1984). No fatty acyl proton and carbon signal resonance
was observed in the NMR spectra. These results supported a conserved structure for the neutral, non-lipidated
mycobacterial outer layer glucans, based on a linear
[ + 4)-N-D-Glcp(1+ 1.
Extracellular materials and colony morphology
Previous studies have demonstrated that the colony
morphology of some strains of Vibrio cbolerae may be
correlated with the production of large amounts of
extracellular polysaccharides (Johnson eta!. ,1992),whereas in Clavibacter micbiganense, the molecular masses of these
polysaccharides may determine the texture of the colonies
(Henningson & Gudmestad, 1992). In mycobacteria,
however, no such data has been published. Thus, we
compared the production of extracellular materials of
smooth and rough colony-forming strains of M. tarbercarlosis and M. kansasii which consist mainly of polysaccharides (see Lemassu & Daffi, 1994, and TaMe 1,
respectively). No obvious correlation was observed
between the amount of ECM and the colony morphology.
The smooth strain 60 of M. ttlberctllosis, a Canetti-like
strain, produced no more ECM than the spontaneous
rough revertant of the Canetti strain, 59R [17-20 mg
ECM (100 mg bacteria)-'] although strain Canetti itself
produced more [about 33 mg ECM (100 mg bacteria)-'].
A smooth strain of M. kansasii, ATCC 12478, produced
only slightly more ECM [24 mg (100 mg bacteria)-'] than
two rough strains, IP 175 and IP 370, of the same species
[14-18 mg (100 mg bacteria)-']. To determine the possible influence of the structural features of the extracellular
polysaccharides on the colony morphology of the strains,
the major extracellular polysaccharides of the isogenic
strains 59 and 59R of M. ttlberctllosis, i.e. the glucans, AMs
and mannans, were purified from the culture filtrates and
their structural features were determined as described
above. Again, no difference was noted between the
analysed compounds. The polysaccharides exhibited the
same apparent molecular masses on gel filtration columns,
gave superimposable NMR spectra and contained similar
amounts of the different glycosyl residues. It followed
then that the structures and amounts of the major
mycobacterial extracellular polysaccharides are similar in
both smooth and rough colony-forming strains.
DISCUSSION
The chemical definition of the mycobacterial cell envelopes is central to understanding the pathogenesis of the
diseases caused by these bacteria. One important question
is whether there is any correlation between the cell
envelope composition and the pathogenicity and/or the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36
Polysaccharides of non-tuberculous mycobacteria
growth rate of the various mycobacterial species. Accordingly, we compared the chemical nature of the
outermost constituents of five pathogenic and nonpathogenic mycobacterial species.
Based on the relative amounts and the kinetics of
production of ECM during growth, the mycobacterial
species examined so far are not uniform and may be
divided into two main groups. The member of the first
group is the pathogen M. kansasii, (Wayne & Kubica,
1786). This species produces a high level of carbohydraterich ECM, whose polysaccharide composition is quantitatively and qualitatively similar to that of SXM. The
second group comprises the pathogenic species M. avizlm
and the non-pathogenic species M. gastri, M . phlei and M .
smegmatis. They produce small amounts of ECM, whose
carbohydrate composition may be qualitatively different
from that found on their cell surface. A subdivision of this
latter group into carbohydrate- and protein-rich SXM
was also possible. M. phlei and M. smegmatis contain
protein-rich SXM, while M. avizlm and M. gastri contain
carbohydrate-rich SXM, a classification that coincides
with the difference in the growth rate of the species. These
data suggest that the differences in the production and
composition of ECM and SXM between mycobacterial
species probably reflect differences in their cell envelope
architecture. In that connection, it is interesting to note
that differences have been demonstrated recently in the
location of the various classes of lipids on the mycobacterial cell surfaces (Ortalo-Mag& e t a/., 1776b). The
relative contribution of ECM and SXM components to
mycobacterial cell envelopes warrants further investigation. Nevertheless, it has to be noted that Man receptors
have been shown to be involved in the phagocytosis of M.
avizlm (Bermudez e t a/.,1971). The failure to detect AMs
among the SXM of 211. avizlm suggests that the mycobacterial components that interact with the Man receptors
during phagocytosis may not be the postulated lipoarabinomannans (Schlesinger e t al., 1774). It is also noteworthy that the ECM of M.avizlm and M,phlei, but not the
SXM of the same species, contained Ara (presumably
from AMs), suggesting the existence of true secreted
polysaccharides (composed at least of AMs), which, in
addition to those shed from the outermost surface of the
cells, will compose the different ECMs.
The polysaccharides of the culture filtrates and those
exposed on the cell surface consist of high molecular mass
glycogen-li ke glucans, lipid-free AMs and mannans,
structurally similar to those previously characterized in
the tubercle bacillus (Lemassu & Daffi, 1774; OrtaloMagne e t al., 1775). The only structural difference
observed between the polysaccharides of the various
species resides in the absence of additional mannosyl
residues attached to the non-reducing termini of the
arabinan segments of the AMs of M.smegmatis. As this
species, in contrast to the other strains examined herein, is
a rapid grower, the above structural difference is in
agreement with the proposal stating that capping of the
arabinan termini with mannosyl residues may be more a
feature of rapid growth than of virulence (Khoo e t al.,
1775 ; Prinzis e t al., 1773).
The polysaccharides derived from both ECM and SXM of
the mycobacterial species examined herein were almost
exclusively neutral substances that were not retained on
the ion-exchange column. This observation raises the
question of the nature of the compounds that strongly
react with ruthenium red at the cell surface of mycobacteria (Rastogi e t al., 1986). This staining method has
been used to demonstrate the existence of a true third
' polysaccharide-rich outer layer ' (POL) in mycobacteria,
as opposed to the two-layered mycobacterial cell wall
(Draper, 1782; Kolbel, 1984). Alternatively, the specificity
of the stain may be questioned.
ACKNOWLEDGEMENTS
T h e authors are grateful t o D r s L. Fattorini, G. Orefici (both
from the Istituto superiore di Sanita, Rome, Italy) for supplying
the materials from S m T and SmD of M . avium. We also thank D r
H. Montrozier (LPTF, CNRS, Toulouse, France) for skillful
advice and Miss F. Mardo (LPTF, CNRS, Toulouse, France) for
her technical help. This work was supported in part by contract
no. 920609 from the Institut National de la SantC et d e la
Recherche Medicale (INSERM, France).
REFERENCES
Andersen, A. B. & Brennan, P. 1. (1994). Proteins and antigens of
Mycobacteritlm ttlberctllosis. In Tubercdosis. Pathogenesis, Protection and
Control, pp. 307-332. Edited by B. R. Bloom. Washington, DC:
American Society for Microbiology.
Bermudez, L. E., Young, L. 5. & Enkel, H. (1991). Interaction of
Mycobacterium avium complex with human macrophages : roles of
membrane receptors and serum proteins. Infect Immtln 59,
1697-1702.
Bloom, B. R. & Murray, J. L. (1992). Tuberculosis: commentary on
a re-emergent killer. Science 257, 1055-1064.
Bradbury, 1. H. & Jenkins, G. A. (1984). Determination of the
structures of trisaccharides by I3C-NMR spectroscopy. Carbobdr
Res 126, 125-156.
Chaisson, R. (1993). Mycobacterial infections and HIV. Ctlrr Opin
Infect Dis 6, 237-243.
Chapman, 5. W. & Henderson, H. M. (1994). New emerging
pathogens - multiply resistant Mycobacteritlm ttlberctllosis. Ctlrr Opin
Infect Dis 7, 231-237.
Chatterjee, D., Bozic, C. M., McNeil, M. & Brennan, P. 1. (1991).
Structural features of the arabinan component of the lipoarabinomannan of Mycobacterium ttlberctllosis. J Biol Cbem 266,
9652-9660.
Chatterjee, D., Lowell, K., Rivoire, B., McNeil, M. R. & Brennan, P.
J. (1992). Lipoarabinomannan of Mjmbacterium tuberculosis: capping
with mannosyl residues in some strains. J Biol Cbem 267,6234-6239.
Ciucanu, 1. & Kerek, F. (1984). A simple and rapid method for
permethylation of carbohydrates. Carbobydr Res 131, 209-21 7.
Daffe, M., Brennan, P. J. & McNeil, M. (1990). Predominant
structural features of the cell wall arabinogalactan of Mycobacteritlm
tuberculosis as revealed through characterization of oligoglycosyl
alditol fragments by gas chromatography/mass spectrometry and
by 'H- and 13C-NMR analyses. J Biol Cbem 265, 6734-6743.
DaffB, M., Brennan, P. J. & McNeil, M. (1993). Major structural
features of the cell wall arabinogalactans of Mycobacteritlm,
Rhodococcus and Nocardia spp. Carbobdr Res 249, 383-398.
Daniel, T. M. (1984). Soluble mycobacterial antigens. In The
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36
1519
A. L E M A S S U and OTHERS
Mycobacteria. A Sourcebook, Part A, pp. 4 1 7 4 6 5 . Edited by G. P.
Kubica & L. G. Wayne. New York: Marcel Dekker, Inc.
Dische, 2. (1962). Color reaction of hexoses. Methods Carbobdr
Chem 1,488-494.
Draper, P. (1982). The anatomy of mycobacteria. In The Biology of
the Mycobacteria, Vol. 1, pp. 9-52. Edited by C. C. Ratledge & J. L.
Stanford, London : Academic Press.
Draper, P. & Rees, R. J. W. (1970). Electron transparent zone of
mycobacteria may be a defence mechanism. Nature 228, 860-861.
Fattorini, L, Xiao, Y., Li, B., Ippoliti, F. & Orefici, G. (1994).
Induction of IL-lB, IL-6, TNF-a, GM-CSF and G-CSF in human
macrophages by smooth transparent and smooth opaque colonial
variants of Mycobacterium auium. J Med Microbiol40, 129-133.
Frdhel, C., Rastogi, N., Bdnichou, 1. C. & Ryter, A. (1988). Do test
tube-grown mycobacteria possess a protective capsule? FEMS
Microbiol Lett 56, 225-230.
Hanks, 1. H. (1961). Significance of capsular components of
Mycobacterium leprae and other mycobacteria. Int J Lepr 29, 74-83.
Harboe, M. (1992). The significance of proteins actively secreted by
Mycobacterium tuberculosis in relation to immunity and complications
of mycobacterial diseases. Int J Lepr 60,470-476.
Henningson, P. J. & Gudmestad, N. C. (1992). Comparison of
exopolysaccharides from mucoid and nonmucoid strains of Clauibacter michiganensis subspecies sepedomkus. Can J Microbiol39,291-296.
Johnson, J. A., Panigrahi, P. & Morris, J. G. (1992). Non-01 Vibrio
cholerae NRS36S produces a polysaccharide capsule that determines
colony morphology, serum resistance and virulence in mice. Infect
Immun 60, 864-869.
Khoo, K.-H., Dell, A., Morris, H. R., Brennan, P. 1. & Chatterjee, D.
(1995). Inositol phosphate capping of the nonreducing termini of
lipoarabinomannan from rapidly growing strains of Mycobacterium.
J Biol Chem 270, 1238G12389.
Kochi, A. (1991). The global tuberculosis situation and the new
control strategy of the World Health Organization. Tubercle 72,
1-6.
KBlbel, H. K. (1984). Electron microscopy. In The Mycobacteria. A
Sourcebook, Part A, pp. 249-300. Edited by G. P. Kubica & L. G.
Wayne. New York: Marcel Dekker, Inc.
Lemassu, A. & Daffd M. (1994). Structural features of the
exocellular polysaccharides of Mjcobacterium tuberculosis. Biochem J
297, 351-357.
Lemassu, A., L&y-FrBbault, V. V., Landelle, M.-A. & DaffB, M.
(1992). Lack of correlation between colony morphology and
lipooligosaccharide content in the Mycobacterium tuberculosis complex. J Gen Microbioll38, 1 5 3 5 1 541.
McNeil, M. R. & Brennan, P. J. (1991). Structure, function and
biogenesis of the cell envelope of mycobacteria in relation to
bacterial physiology, pathogenesis and drug resistance ; some
1520
thoughts and possibilities arising from recent structural information. Res Microbioll42, 451-463.
Misaki, A., Azuma, 1. & Yamamura, Y. (1977). Structural and
immunochemical studies on D-arabino-D-mannans and D-mannans
of Mycobacterium tuberculosis and other Mycobacterium species.
J Biochem 82, 1759-1 770.
Ortalo-Magnd, A., Dupont, M. A., Lemassu, A., Andersen, A. B.,
Gounon, P. & Daff6, M. (1995). Molecular composition of the
outermost capsular material of the tubercle bacillus. Microbiology
141, 1609-1 620.
Ortalo-Magnd, A., Andersen, A. B. & Daff4, M. (1996a). The
outermost capsular arabinomannans and other mannoconjugates of
virulent and avirulent tubercle bacilli. Microbiology 142, 927-935.
Ortalo-Magnd, A., Lemassu, A., Landelle, M. A., Bardou, F., Silve,
G., Gounon, P., Marchal, G. & Daff4, M. (1996b). Identification of
the surface-exposed lipids on the cell envelopes of Mycobacterium
tuberculosisand other mycobacterial species.J Bacterioll78,456-461.
Paul, T. R. & Beveridge, T. 1. (1994). Preservation of surface lipids
and determination of ultrastructure of Mycobacterium kansasii by
freeze-substitution. Infect Immun 62, 1542-1 550.
Prinzis, S., Chatterjee, D. & Brennan, P. J. (1993). Structure and
antigenicity of lipoarabinomannan from Mycobacterium bovis BCG.
J Gen Microbioll39, 2649-2658.
Rastogi, N., FrBhel, C. & David, H. L. (1986). Triple-layered
structure of mycobacterial cell wall: evidence for the existence of a
polysaccharide-rich outer layer in 18 mycobacterial species. Curr
Microbiol 13, 237-242.
Sauton, B. (1912). Sur la nutrition minirale du bacille tuberculeux.
C R Acad Sci Ser III Sci Vie 155, 860-863.
Schlesinger, L. S., Hull, 5. R. & Kaufman, T. M. (1994). Binding of
the terminal mannosyl units of lipoarabinomannan from a virulent
strain of MjcobacteriHm tubercdosisto human macrophages. J Immunol
152,4070-4079.
Seibert, F. B. (1949). The isolation of three different proteins and
two polysaccharides from tuberculin by alcohol fractionation.
Their chemical and biological properties. A m Reu Tuberc Puim Dis
59, 86-101.
Wayne, L. G. & Kubica, G. P. (1986). Genus Mycobacterium Lehmann
and Neumann 1896, 363th. In Bergg’s Manual of Systematic Bacteriology, Vol. 2, pp. 1436-1457. Edited by P. H. A. Sneath, N. S .
Mair, M. E. Sharpe & J. G. Holt. Baltimore: Williams & Wilkins.
Youmans, G. P. (1979). Virulence of Mycobacteria. In Tuberculosis,
pp. 194-201. Edited by G. P. Youmans. Philadelphia: W. B.
Saunders Company.
Young, D. B., Kaufmann, S. H. E., Hermans, P. W. M. & Thole,
1. E. R. (1992). Mycobacterial protein antigens : a compilation. Mol
Microbiol6, 133-145.
Received 26 September 1995; revised 2 January 1996; accepted 22
January 1996.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 01:44:36