Tài liệu Báo cáo Y học: Trehalose-based oligosaccharides isolated from the cytoplasm of Mycobacterium smegmatis Relation to trehalose-based oligosaccharides attached to lipid docx

Trehalose-based oligosaccharides isolated from the cytoplasm
of
Mycobacterium smegmatis
Relation to trehalose-based oligosaccharides attached to lipid
Masaya Ohta
1
, Y. T. Pan
2
, Roger A. Laine
3
and Alan D. Elbein
2
1
Department of Biochemistry, Fukuyuma University, Japan;
2
Department of Biochemistry and Molecular Biology,
University of Arkansas for Medical Sciences, Little Rock, AR 72205, USA;
3
Departments of Biological Sciences and Chemistry,
Louisiana State University, Baton Rouge, LA, USA
A series of trehalose-based oligosaccharides were isolated
from the cytoplasmic fraction of Mycobacterium smegmatis
and purified by gel-filtration and paper chromatography and
TLC. Their structures were determined by HPLC and GLC
to determine sugar composition and ratios, MALDI-TOF
MS to measure molecular mass, methylation analysis to
determine linkages,
1
H-NMR to obtain anomeric configu-
rations of glycosidic linkages, and exoglycosidase digestions
followed by TLC to determine sequences and anomeric

configurations of the monosaccharides. Six different oligo-
saccharides were identified all with trehalose as the basic
structure and additional glucose or galactose residues
attached in various linkages. One of these oligosaccharides is
the disaccharide trehalose (Glca1–1aGlc), which is present
in substantial amounts in these cells and also in other
mycobacteria. Two other oligosaccharides, the tetrasaccha-
rides Glca1–4Glca1–1aGlc6–1aGal and Gala1–6Gala1–
6Glca1–1aGlc, have not previously been isolated from
natural sources or synthesized chemically. The fourth
oligosaccharide, Glcb1–6Glcb1–6Glca1–1aGlc, has been
isolated from corynebacteria, but not reported in other
organisms. Two other oligosaccharides, Glca1–4Glca1–
1aGlc, which has been synthesized chemically and isolated
from insects but not previously reported in mycobacteria,
and Glcb1–6Glca1–1aGlc, which was previously isolated
from Mycobacterium fortuitum and yeast, were also charac-
terized. Another trisaccharide found in the cytosol has been
partially characterized as arabinosyl-1–4trehalose, but nei-
ther the anomeric configuration nor the
D
or
L
configuration
of the arabinose is known. In analogy with sucrose and its
higher homologs, raffinose and stachyose, which may act as
protective agents during maturation drying in plants, these
trehalose homologs may also have a protective role in
mycobacteria, perhaps during latency.
Keywords: Mycobacteria, oligosaccharides; trehalose.

The mycobacterial cell wall is an exceedingly complex
network of interacting molecules and includes various
polysaccharides such as mycolic acid–arabinogalactan and
lipoarabinomannan, as well as a variety of complex glycoli-
pids [1]. Among the interesting and important glycolipds are
a number of trehalose-containing lipids, such as trehalose
monomycolate and dimycolate, and other acylated trehalose
compounds [2]. In addition to its role as a structural
component, trehalose dimycolate has also been implicated
as a donor of mycolic acids to the arabinogalactan [3,4].
Furthermore, in yeast, bacteria and various other organisms,
free trehalose has been shown to have a protective role as a
stabilizer of proteins and membranes during dehydration,
dessication, heat shock and other adverse conditions [5,6].
Trehalose (Glca1–1aGlc) is not found in any vertebrates,
nor is it synthesized in these organisms. Thus, the synthesis
and acylation of trehalose represent excellent target sites for
the design of new drugs against tuberculosis. Therefore, we
have purified the trehalose phosphate synthase [7] and
trehalose phosphate phosphatase (S. Klutts, I. Pastuszak,
D. Carroll, Y. T. Pan & A. D. Elbein, unpublished results)
from Mycobacterium smegmatis and examined cytosolic
and lipid extracts of M. smegmatis for possible analogs of
trehalose. In this paper, we report on the isolation from the
cytoplasm and characterization of trehalose and five other
trehalose-based oligosaccharides. Two of these oligosaccha-
rides are newly described tetrasaccharides, so far only
reported here in M. smegmatis. Another trehalose oligosac-
charide has been synthesized chemically but not previously
isolated from any living organisms, and two others have

been demonstrated in other cells, but not in M. smegmatis.
One other oligosaccharide, with an arabinose linked to
trehalose, is also new but its complete structure is not
known.
At this point, we do not know the function of these
trehalose oligosaccharides. However, in analogy with the
plant disaccharide sucrose and its higher homologs raffinose
and stachyose, these higher derivatives of trehalose may also
stabilize M. smegmatis and other mycobacteria during stress
and perhaps latency. In plants, sucrose, raffinose and
stachyose have been implicated as protective agents during
maturation drying [9]. Sucrose is a nonreducing disac-
charide of glucose and fructose, whereas trehalose is a
Correspondence to A. D. Elbein, Department of Biochemistry and
Molecular Biology, University of Arkansas for Medical Sciences,
Little Rock, AR 72205, USA.
Fax: 501 686 8169, Tel. 501 686 5176,
E-mail:
Abbreviation: MALDI-TOF, matrix-assisted laser desorption
ionization time-of-flight.
(Received 11 January 2002, revised 27 March 2002,
accepted 30 April 2002)
Eur. J. Biochem. 269, 3142–3149 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02971.x
nonreducing disaccharide of two glucoses. Thus, the sucrose
and the trehalose series may have analogous functions.
EXPERIMENTAL PROCEDURES
Materials
M. smegmatis was obtained from the American Type
Culture Collection. Trypticase Soy broth was from Fisher
Chemical Co. a-Glucosidase and a-galactosidase were

purchased from Sigma Chemical Co. Whatman 3
MM
paper
was from Whatman Co., and silica gel thin-layer plates were
from Analtech, Inc. All other chemicals were from reliable
chemical companies and were of the best grade available.
Growth of
M. smegmatis
and isolation
of oligosaccharides
M. smegmatis was grown in trypticase soy broth for
24–48 h at 37 °C. A 125-mL Erlenmeyer flask containing
25 mL medium was inoculated from a slant of the
organism, and grown overnight. This culture was then
used as the starter to inoculate a number of large culture
flasks (2-L conical flasks containing 1 litre of medium).
Cells were grown for 36–48 h, harvested by centrifugation,
and stored as a paste in aluminium foil at )20 °C until used.
M. smegmatis cells were suspended in water (100 g cell
paste in 200 mL water) and disrupted by sonication. The
sonicate was centrifuged at  37 000 g for 30 min, and the
supernatant, representing the cytosolic fraction of the cell,
was removed and saved. To this ice-cold supernatant, was
added trichloroacetic acid to a final concentration of 5% to
precipitate the protein. After removal of the protein by
centrifugation, the trichloroacetic acid was removed by
extraction with diethyl ether, and then alcohol was added to
the supernatant to a final concentration of 70% to
precipitate glycogen and any other trichloroacetate-soluble
polymers present in the cytoplasm.

The alcohol mixture was allowed to stand overnight in
the cold, and any precipitate was removed by centrifugation.
The supernatant from this centrifugation was concentrated
to a small volume and deionized by treatment with mixed-
bed ion-exchange resin [equal mixture of Dowex-1 (CO
3
2–
)
and Dowex-50 (H
+
)]. The deionized solution was applied to
a Bio-Gel P-2 column (1.5 · 200 cm) and eluted with 1%
acetic acid. Fractions were collected, and an aliquot of each
was analyzed with the anthrone reagent [10] for the presence
of hexose. A number of anthrone-positive peaks were
identified which corresponded to monosaccharide, disac-
charide, trisaccharide and tetrasaccharide (Fig. 1). A large
anthrone-positive peak emerged early in the elution and
may represent polymeric material. Each of these peaks was
pooled and subjected to additional gel filtrations for further
purification. In some cases, the peaks from the Bio-Gel
column were also subjected to paper chromatography in
various solvents to separate the oligosaccharides further.
Sugars and oligosaccharides were visualized with silver
nitrate reagent [11]. Oligosaccharides were also separated by
TLC on both analytical plates and, in some cases, prepar-
ative plates and by preparative HPLC. As trehalose is a
nonreducing sugar, each of the oligosaccharides was tested
with the reducing sugar test [12] to determine whether it was
also a nonreducing sugar.

Paper chromatography and TLC
Oligosaccharides were streaked on 9 inch sheets of What-
man 3
MM
paper (22 inches long) and chromatographed by
descending chromatography in the following solvent sys-
tems: butan-1-ol/pyridine/0.1
M
HCl (5:3:2, by vol.);
propan-2-ol/butan-1-ol/water (140 : 20 : 40, by vol.); pro-
panol/ethyl acetate/water (140 : 20 : 40, by vol.). TLC of
sugars and oligosaccharides was performed on silica-gel
plates in acetonitrile/water (4 : 1). Usually the thin-layer
plates were subjected to multiple developments in the
acetonitrile/water solvent with complete drying between
each run.
HPLC
HPLC analysis was performed with a Shimadzu model
LC-10A liquid chromatograph (Shimadzu Co.) equipped
with a TSK-gel Amide-80 column (0.46 · 250 mm;
TOSOHO) at 40 °C. Elution was isocratic with acetonitrile/
water (56 : 42, v/v) at a flow rate of 0.5 mLÆmin
)1
.Elution
of oligosaccharides was monitored with a refractive index
detector [13].
Matrix-assisted laser desorption ionization time-of-flight
(MALDI-TOF) MS characterization
The molecular mass of each oligosaccharide was determined
by MALDI-TOF MS (Voyager DE STR, Perseptive

Biosystems) in a positive-ion reflector mode using 20 kV
total accelerating voltage. Desorption/ionization was via a
nitrogen laser operating at 337 mm. The matrix solution
contained 10 mgÆmL
)1
2,5-dihydroxybenzoic acid in 10%
ethanol. One microlitre of the matrix solution and 1 lLof
the peracetylated oligosaccharides in chloroform were
mixed on the sample plate directly and dried to avoid
Fig. 1. Separation of the cytosolic oligosaccharides on columns of Bio-
Gel P-2. A1.5· 200 cm column of Bio-Gel P-2 was prepared and
washed with 1% acetic acid. The samples were applied to the column
and eluted with the same solution. Fractions were collected and an
aliquot of each was removed and assayed for its hexose content by the
anthrone reaction. Letters shown at the top indicate the elution posi-
tions of the sugar standards: G ¼ glucose, T ¼ trehalose, R ¼
raffinose, S ¼ stachyose.
Ó FEBS 2002 Cytoplasmic trehalose oligosaccharides (Eur. J. Biochem. 269) 3143
crystallization. The analyzed substances were detected as
their pseudomolecular ions ([M + Na]
+
). Oligosaccharides
were peracetylated by suspending lyophilized samples in
50 lL acetic anhydride and incubating at 80 °C for 2 h. The
reaction mixture was evaporated to dryness under a stream
of N
2.
Methylation analysis
Oligosaccharides were permethylated by the method of
Ciucanu & Kerek [14]. Methylated samples were hydro-

lyzed in acid, reduced and acetylated as described by Liang
et al. [15]. The sugar acetates were chromatographed on a
silica-gel capillary column (DB-5MS; 0.25 mm · 30 m; J
and W Scientific Co., Folsom, CA, USA) in a Finnigan
GCQ (Trace GC 2000 coupled with Polaris MSD). The
column was operated at 50 °C for 1 min, increased to
170 °Catarateof20° per min, and then increased to
260 °Cat3° per min. It was then held at this final
temperature for 4 min. The column was calibrated with
various methylated glucose standards.
Exoglycosidase digestions
To determine the sequence of sugars in these oligosaccha-
rides, as well as the anomeric configurations of the linkages,
various oligosaccharides were subjected to digestion by
specific exoglycosidases that remove sugars from the
nonreducing ends of the oligosaccharides. The following
enzymes and incubation conditions were used: (a) 1 unit
a-glucosidase (from Bacillus stearothermophilus)wasincu-
bated with substrates in 0.1
M
sodium phosphate buffer,
pH 6.8; (b) 0.5 unit a-galactosidase (from Aspergillus niger)
was incubated with substrates in 50 m
M
sodium acetate
buffer, pH 4.0.
All incubations were performed at 37 °C for 24 h, and
were stopped by placing the incubation mixtures in a boiling
water bath for 3 min. The mixtures were diluted in water
and treated with mixed-bed ion-exchange resin to remove

salt and other charged compounds. The digestion products
were then analyzed by TLC.
RESULTS
Isolation of cytosolic oligosaccharides
M. smegmatis cells were disrupted by sonication and the
cytosolic fraction was isolated by high-speed centrifugation
to remove cell walls and membranes. The supernatant was
treated with trichloroacetic acid and ethanol to remove
protein and glycogen as described, and after treatment with
mixed-bed ion-exchange resin, it was applied to a column of
Bio-Gel P-2 to roughly separate the various oligosaccha-
rides into their different sizes, i.e. mono, di, tri, and tetra.
Figure 1 shows a profile of the elution pattern of sugars
(i.e. anthrone-positive material) from this column. A
number of anthrone-positive peaks were eluted from the
column, with the three major peaks corresponding to
glucose (fractions 111–120), trehalose (fractions 101–110)
and high-molecular-mass material (fractions 50–62). How-
ever, a number of other peaks containing smaller amounts
of anthrone-positive material were also detected and each of
these peaks was designated as follows: CO7, CO6, CO5,
CO4, CO3, CO2, CO1 (see Fig. 1 for positions of each
peak). CO1 was the monosaccharide peak containing
mostly glucose, and CO7 was composed of several higher-
molecular-mass oligosaccharides, which were not further
characterized because of limiting amounts of material (see
Fig. 2 for TLC profiles). A large peak was eluted very early
(fractions 50–65), which probably represents polysaccharide
material. This peak has not yet been characterized.
Each of the peaks labeled CO2 through CO6 were pooled

and analyzed to determine the amount of oligosaccharide
present. Several were also analyzed by the reducing sugar
test to compare the amount of hexose present with the
amount of reducing sugar. This was to determine whether
nonreducing oligosaccharides of the trehalose class were
present in these fractions. That is, for a reducing tetrasac-
charide, one would expect one reducing sugar for every four
hexoses, whereas for a trisaccharide there should be one
reducing sugar for every three hexoses. Table 1 shows the
results from the gel-filtration column, indicating the position
of elution of each peak (fraction numbers), the amount of
hexose based on anthrone determinations [10] and reducing
sugar based on the Nelson method [12] compared with a
glucose and maltose standard. These standards gave hexose/
reducing sugar values of 1 : 1 and 2 : 1 as expected. The
ratio of total hexose to reducing sugar of several of the
trehalose oligosaccharides are shown in Table 1. The
amount of each oligosaccharide was as follows (total lmol
glucose/100 g cell paste): CO1, 90; CO2, 150; CO3, 3; CO4,
14; CO5, 7; CO6, 2. It can be seen that these cells contained
substantial amounts of trehalose (CO2) and free glucose
(CO1), and considerably smaller amounts of the various
higher oligosaccharides. However, most of these peaks
appeared to be composed mostly of nonreducing oligosac-
charides, presumably of the trehalose class, because the ratio
of number of hexoses/reducing sugar was greater than 8 and
as high as 50.
Fig. 2. TLC of the various oligosaccharides after isolation by gel fil-
tration and paper chromatography. Samples were applied to the plate
and it was developed twice in acetonitrile/water (4 : 1, v/v). Carbo-

hydrates were detected by spraying the dried plates with 50% con-
centrated sulfuric acid in ethanol followed by heating at 110 °C
for 5–10 min. Lane 1, various sugar standards (G ¼ glucose, T ¼
trehalose, R ¼ raffinose, S ¼ stachyose); lanes 2–8, the isolated
oligosaccharides designated CO1 through CO7.
3144 M. Ohta et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Each of the peaks shown in Table 1 was rerun on the Bio-
Gel column for further purification, and then each peak
from these columns was streaked on sheets of Whatman
3
MM
paper and subjected to chromatography in one or
more of the solvents described in Experimental Procedures.
Areas corresponding to various oligosaccharides were cut
from the papers, and the oligosaccharides eluted with water
and rechromatographed until they appeared to be homo-
geneous.
Figure 2 shows a thin-layer chromatogram of the indi-
vidual oligosaccharides characterized in this study. Oligo-
saccharide CO3 was found to be composed of two
compounds, on the basis of Amide-80 HPLC (Fig. 3),
and these were separated and designated CO3-1 and CO3-2.
Likewise, CO4 was separated into four fractions by HPLC
(Fig. 3), and these were designated CO4-1, CO4-2, CO4-3
and CO4-4.
Structures of the oligosaccharides
Fractions CO1 and CO2. The mobility on TLC (Fig. 2),
the methylation data (Table 2), and the
1
H-NMR analysis

revealed that CO-1 was glucose and CO-2 (referred to as
oligosaccharide CD2, see Table 4) was trehalose.
Fraction CO3. CO3 was composed of two different
oligosaccharides as shown by TLC (Fig. 2). These two
components, designated as CO3-1 and CO3-2, were
separated in sufficient amounts by HPLC (Fig. 3) to allow
chemical characterization of each. CO3-1 was identified as
trehalose by TLC as well as by MALDI-TOF MS analysis
of the products obtained by permethylation and hydrolysis,
andalsoby
1
H-NMR.
Oligosaccharide CO3-2 appeared to be a trisaccharide, as
MALDI-TOF MS analysis of the peracetylated oligosac-
charide gave a pseudomolecular ion [M + Na]
+
at m/z
989, suggesting that it was composed of three hexoses.
Methylation, followed by hydrolysis and reduction, of
this oligosaccharide, as shown in Table 2, gave 1 mol 2,3,6-
tri-O-methylglucitol acetate and 2 mol 2,3,4,6-tetra-O-
methylglucitol acetate. These results establish that the
oligosaccharide has the structure Glc1–4Glc1–1Glc. The
1
H-NMR analysis, presented in Fig. 4, shows three ano-
meric proton signals at 5.21 p.p.m. (J
12
¼ 4.0), 5.21 p.p.m.
(J
12

¼ 4.0), and 5.41 p.p.m. (J
12
¼ 3.6), attributable to all
a-glycosidic linkages. The overlapping a-linked signals at
5.21 p.p.m. were suggestive of an a,a¢-linked disaccharide,
i.e. a,a-trehalose [16]. Moreover, these values for the
anomeric proton signals are the same as seen for bemisiose
isolated from Bemisia honeydew [17]. On the basis of these
results, oligosaccharide CO3-2 is identified as Glca1–
4Glca1–1aGlc and is termed CT3A (see Table 4).
Fraction CO4. The migration of CO4 on TLC plates, as
seen in Fig. 2, indicated that it was composed of several
poorly separated components. Subsequent fractionation by
Amide-80 HPLC (Fig. 3) resulted in separation of four
peaks, designated CO4-1, CO4-2, CO4-3 and CO4-4, in the
proportions 11%, 45%, 23%, and 21%, respectively. Each
oligosaccharide was subjected to structural characterization
by methylation analysis, MALDI-TOF MS and
1
H-NMR.
CO4-1, when peracetylated and subjected to MALDI-
TOF MS, gave a single peak at m/z 917, indicating that this
oligosaccharide was a trisaccharide consisting of 1 mol
pentose and 2 mol hexose. Further investigation of the
permethylated alditol acetates from this oligosaccharide gave
Fig. 3. Amide-HPLC separataion of trehalose oligosaccharide fractions
CO3 and CO4. An Amide89 column, 4.6 · 250 mm, was run with
isocratic 58% acetonitrile in water at a flow rate of 0.5 mLÆmin
)1
.

Oligosaccharides were detected by refractive index. Standard sugars
wererunonthiscolumnandemergedinpositionsshownatthetopof
the profile, i.e. T ¼ trehalose, R ¼ raffinose, S ¼ stachyose.
Table 1. Amount of oligosaccharides isolated from 100g cells.
Peak designation Fractions from P-2 column lmol Glucose
a
Ratio hexose/reducing sugar
Polymer 51–60 164 –
CO7 61–70 5 –
CO6 73–79 2 8.3 (5 : 0.6)
CO5 77–85 7 69 (6.9 : 0.1)
CO4 85–92 14 –
CO3 97–103 3 8 (3 : 0.4)
CO2 100–105 150
CO1 111–114 90
a
Based on anthrone assay.
Ó FEBS 2002 Cytoplasmic trehalose oligosaccharides (Eur. J. Biochem. 269) 3145
three peaks corresponding to terminal arabinose, terminal
glucose, and 4-linked glucose (Table 2). These results
suggested that the arabinose was attached to position 4 of
one of the glucoses of trehalose. Because of limited amounts
of this oligosaccharide, further analysis was not possible.
This peak is most likely Ara-1,4Glca1–1aGlc, but the
anomeric configuration and the
D
or
L
form of the arabi-
nose have not been determined. No arabinose-containing

oligosaccharides of trehalose have previously been reported.
The most prominent peak, CO4-2, gave the same results
as CO3-2 when subjected to MALDI-TOF MS, methyla-
tion analysis and
1
H-NMR. Thus CO4-2 appears to be the
trisaccharide, Glca1–4Glca1–1aGlc, and was referred to as
CT3A (Table 4).
In oligosaccharide CO4-3, a major pseudomolecular ion
[M + Na]
+
signal at m/z 989 was observed corresponding
to a trisaccharide composed of three hexoses. A minor
pseudomolecular ion [M + Na]
+
at m/z 1073 was also
detected, which probably corresponds to a compound
composed of two hexoses and one hexitol. Table 2 shows
that methylation of CO4-3 gave two peaks, indicating the
presence of terminal glucose and 6-linked glucose in the
ratio 2 : 1.
1
H-NMR of the CO4-3 fraction is presented in
Fig. 4 and showed the presence of three anomeric proton
signals at 5.20 p.p.m. (J
12
¼ 4.0), 5.20 p.p.m. (J
12
¼ 4.0),
and 4.41 p.p.m. (J

12
¼ 7.6). The signals at 5.20 were
assigned to two a-linkages and the signal at 4.41 p.p.m.
was assigned to a b-linkage.
The b-linked signal (4.41 p.p.m) suggested that the
terminal glucose residue was attached to C-6 of an internal
glucose [18]. From these results, the structure of the major
oligosaccharide in the CO4-3 peak is Glcb1–6Glca1–1aGlc
(termed CT3b, Table 4).
MALDI-TOF MS analysis of the peracetylated oligo-
saccharide in CO4-4 gave a pseudomolecular ion
[M + Na]
+
at m/z 1277, suggesting a tetrasaccharide
structure composed of four hexoses. Methylation analysis,
as seen in Table 2, gave an almost equal amount of 2,3,4,6-
tetramethylglucitol acetate and 2,3,4-trimethylglucitol acet-
ate, indicating that the structure of this tetrasaccharide was
either Glc1–6Glc1–1Glc6–1Glc or Glc1–6Glc1–6Glc1–
1Glc. The
1
H-NMR spectrum showed four anomeric
proton signals at 5.21 p.p.m. (J
12
¼ 4.0), 5.19 p.p.m.
(J
12
¼ 4.0), 4.41 p.p.m. (J
12
¼ 7.6) and 4.51 p.p.m.

(J
12
¼ 8.0), attributable to two a-glycosidic and two
b-glycosidic linkages (Fig. 4). The two b-linked glucose
signals at 4.41 p.p.m and 4.51 p.p.m. are consistent with a
Fig. 4.
1
H-NMR spectra of the cytoplasmic oligosaccharides, CO3-2,
CO4-3 and CO4-4. The methodology is described in Experimental
Procedures. All oligosaccharides obtained in sufficient quantity were
subjected to NMR as described here and gave the results described in
the text.
Table 2. Methylation analysis of cytoplasmic oligosaccharides from M. smegmatis.
Residues
Oligosaccharides (molar ratio)
CO1 CO2 CO3–1 CO3–2 CO4-1 CO4–2 CO4-3 CO4-4 CO5 CO6
Arabinitol
2,3,5-Tri-O-methyl- – – – – 0.8 – – – – –
Glucitol
2,3,4,6-Tetra-O-methyl- 1.0
a
2.0
b
2.0
b
2.0
b
1.0
a
2.0

b
2.0
b
2.0
b
1.0
a
0.7
2,3,6-Tri-O-methyl- – – 0.8 0.7 0.9 – – 1.2 –
2,3,4-Tri-O-methyl- – – – – – – 0.8 1.8 0.9 1.1
Galactitol
2,3,4,6-Tetra-O-methyl- – – – – – – – +
c
0.9 1.0
d
2,3,4-Tri-O-methyl- – – – – – – – – – 1.1
a
Values normalized to one residue of 2,3,4,6-tetra-O-methylglucitol.
b
Values normalized to two residues of 2,3,4,6-tetra-O-methylglucitol.
c
+ means < 0.1.
d
Values normalized to one residue of 2,3,4,6-tetra-O-methylgalactiol.
3146 M. Ohta et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Glcb1–6Glcb1–6 disaccharide linked to a trehalose core
[19]. Therefore the most likely structure of this tetrasaccha-
ride is Glcb1–6Glcb1–6Glca1–1aGlc. It was designated
CT4AasindicatedinTable4.
Fraction CO5. MALDI-TOF MS analysis of the peracet-

ylated CO5 showed a pseudomolecular ion at 1277,
characteristic of a tetrasaccharide of four hexoses. Table 2
presents the results of methylation analysis of this tetrasac-
charide. This method gave the following four methylated
sugars in almost equal amounts: 2,3,4,6-tetramethylglucitol
acetate, 2,3,6-trimethylglucitol acetate, 2,3,4-trimethylgluc-
itol acetate and 2,3,4,6-tetramethylgalactitol acetate, as seen
in Fig. 5. The presence of 2,3,4-trimethylglucitol acetate and
2,3,6-trimethylglucitol acetate indicated that these two
glucoses were linked together in a 1–1 linkage (i.e. as
trehalose) with a terminal galactose linked to one of the
glucoses of the trehalose in either a 1–6 or 1–4 linkage, and
the nonreducing glucose linked to the other trehalose
glucose in either a 1–6 or a 1–4 linkage.
To determine the sugar sequence of this tetrasaccharide, it
was subjected to exoglycosidase digestions with specific
glycosidases, and the products of the digestions were
isolated and characterized. When CO5 was incubated with
purified a-galactosidase, it was converted into a trisaccha-
ride that migrated on TLC plates with Glca1–4Glca1–
1aGlc as shown in Fig. 6. When the trisaccharide product
was subjected to methylation analysis, 2 mol 2,3,4,6-tetra-
methylglucitol acetate and 1 mol 2,3,6-trimethylglucitol
acetate were found (see Table 3 and Fig. 5). This indicated
that the galactose residue was linked in an a1–6 linkage to
one of the glucoses of trehalose, and the glucose was linked
to the other trehalose glucose in an a1–4 linkage.
The trisaccharide resulting from the a-galactosidase
digestion was susceptible to digestion by a-glucosidase with
the release of glucose. When the original tetrasaccharide was

treated with a-glucosidase, a trisaccharide product was
obtained. As shown in Table 3, methylation of this trisac-
charide gave 1 mol 2,3,4,6-tetramethylglucitol acetate and
1 mol 2,3,4-trimethylglucitol acetate and 1 mol 2,3,4,6-
tetramethyl galactitol acetate. These results show that this
tetrasaccharide has the structure Glca1–4Glca1–1aGlc6–
1aGal. This oligosaccharide is designated CT4b (Table 4).
Fig. 5. Methylation analysis of the various oligosaccharides isolated as described. The methylation results of CO5 and CO6 are shown here before
and after glycosidase treatment, but similar results and similar methodology was used with other oligosaccharides.
Fig. 6. Determination of the sugar sequence and the anomeric confi-
guration using specific exoglycosidases. Oligosaccharides were subjected
to digestion with a-glucosidase, a-galactosidase, or both and the
products determined by TLC in acetonitrile/water (4 : 1, v/v).
Lane 1, sugar standards (G ¼ glucose, T ¼ trehalose, R ¼ raffinose,
S ¼ stachyose); lane 2, a-galactosidase treatment of CO5; lane 3,
a-galactosidase treatment of CO6; lane 4, a-glucosidase digestion of
CO5; lane 5, a-glucosidase digestion of CO6. After chromatography
and drying of the plates, carbohydrates were detected with sulfuric acid
and heating.
Ó FEBS 2002 Cytoplasmic trehalose oligosaccharides (Eur. J. Biochem. 269) 3147
Fraction CO6. The peracetylated CO6 gave only a single
peak at m/z 1277 as a pseudomolecular ion [M + Na]
+
,
indicating that this oligosaccharide was also a tetrasaccha-
ride composed of four hexose residues. When this oligosac-
charide was subjected to methylation analysis (see Table 2),
four different methylated products, all present in about
equal amounts, were identified as follows: 2,3,4-trimethyl-
glucitol acetate, 2,3,4,6-tetramethylglucitol acetate, 2,3,

4-trimethylgalactitol acetate and 2,3,4,6-tetramethylgalacti-
tol acetate. As the amount of CO6 available for character-
ization was too low for NMR analysis, additional charac-
terization was performed using various exoglycosidases to
remove terminal sugars, followed by analysis of the
products of these digestions. The tetrasaccharide was
resistant to hydrolysis by a-glucosidase (Fig. 6, lane 5),
but was susceptible to digestion by a-galactosidase (Fig. 6,
lane 3). Treatment of CO6 with a-galactosidase resulted in
the formation of trehalose, as identified by TLC (Fig. 6).
The disaccharide resulting from treatment with a-galactosi-
dase was further converted into free glucose by digestion
with a-glucosidase (data not shown). In addition, methyla-
tion of the disaccharide resulting from a-galactosidase
digestion gave only 2,3,4,6-tetramethylglucitol acetate
(Table 3). This product could only arise from a disaccharide
linked in a 1–1-glycosidic bond. These results indicate that
this tetrasaccharide has the structure Gala1–6Gala1–
6Glca1–1aGlc,andisreferredtoasCT4c(Table4).
DISCUSSION
The isolation and characterization of a number of trehalose-
based oligosaccharides from the cytoplasmic fraction of
M. smegmatis are described. Besides trehalose, five other
nonreducing oligosaccharides were isolated and character-
ized. Two of these, the tetrasaccharides CT4b (Glca1–
4Glca1–1aGlc6–1aGal) and CT4c (Gala1–6Gala1–
6Glca1–1aGlc) are newly described and have not been
previously reported from any natural sources, nor are there
any reports on their chemical synthesis. On the other hand,
a number of trehalose-containing oligosaccharides have

been synthesized chemically, some of which are related to
those described here. For example, trehalose with a
b-galactose linked to the 6-hydroxy group of each glucose
of the trehalose (i.e. Galb1–6Glca1–1aGlc6–1bGal) has
been synthesized chemically [20], as well as a trehalose with
an a1,6-linked disaccharide of glucose linked to the 6
position of one of the trehalose glucoses (i.e. Glca1–
6Glca1–6Glca1–1aGlc) [21]. A number of other tetrasac-
charides of trehalose have been chemically synthesized with
either galactose or glucose linked in a or b linkages.
The third tetrasaccharide isolated from M. smegmatis,
CT4a (Glcb1–6Glcb1–6Glca1–1aGlc), has been reported in
corynebacteria, but has not previously been described in
mycobacteria [19]. However, a pentasaccharide, with the
same structure as CT4a except for an additional glucose
linked b to the 3-hydroxy group of the second glucose of
trehalose, has been isolated from the cell wall lipids of
Mycobacterium kansasii [1]. That study strongly suggests
that some of the oligosaccharides described here, and
isolated from the cytoplasm of M. smegmatis, will also be
found as part of the cell wall components of this organism.
All trehalose-containing structures reported from mycobac-
terial glycolipids thus far contain b-linked additional sugars.
The cytoplasm of M. smegmatis also contains at least two
trisaccharides that have trehalose as the base. One of these
compounds identified in both CO3-2 and CO4-2 is the
trisaccharide, Glca1–4Glca1–1aGlc, i.e. CT3a. This com-
pound has been synthesized chemically [22] and also isolated
from insects and yeast. It is interesting to note that this
trisaccharide is also related to the new tetrasaccharide

reported here as CT4b and may be a precursor in the
synthesis of that tetrasaccharide. It is important to deter-
mine whether there are galactosyltransferases and glucosyl-
transferases in these cells that can convert trehalose into
these various oligosaccharides, because some of these
enzymes could be target sites for chemotherapy. The other
trisaccharide found in fraction CO4-3 has the structure
Glcb1–6Glca1–1aGlc. This compound has been reported in
Aspergillus sydowi [23], and it may also be a precursor of
CT4a.
Table 4. Proposed structures of cytoplasmic oligosaccharides from
M. smegmatis.
Oligosaccharide Structure Fraction
CD2 G1ca1-1aG1c CO2, CO3-1
CT3a G1ca1- 4G1ca1-1aG1c CO3-2,CO4-2
CT3b G1cb1- 6G1ca1-1aG1c CO4-3
CT4a G1cb1-6 G1cb1-6G1ca1-1aG1c CO4-4
CT4b G1ca1-4G1ca1-1aG1c6-1aGal CO5
CT4c Ga1a1-6 Gala1-6G1ca1-1aG1c CO6
Table 3. Methylation analysis of exo-glycosidase digestion products.
Residues
CO5 CO6
Native
a-Galactosidase
digestion
a-Glucosidase
digestion Native
a-Galactosidase
digestion
Glucitol

2,3,4,6-Tetra-O-methyl 1.0 2.1 2.2
a
0.7 1.9
2,3,6-Tri-O-methyl- 1.2 1.0
___
2,3,4-Tri-O-methyl- 0.9
_
0.9 1.1
_
Galactitol
2,3,4,6-Tetra-O-methyl- 0.9 0.9
b
0.9 1.0 2.1
b
2,3,4-Tri-O-methyl-
__ _
1.1
_
a
Values include the amount of the released glucose residue.
b
Values indicate the amount of the released galactose residue.
3148 M. Ohta et al.(Eur. J. Biochem. 269) Ó FEBS 2002
The origin of these cytosolic oligosaccharides is not
known. They could be synthesized in the cytosol by
glycosyltransferases which add glucosyl or galactosyl resi-
dues to the free trehalose that arises from trehalose
phosphate by the action of the specific trehalose phosphate
phosphatase (S. Klutts, I. Pastuszak, D. Carroll, Y. T. Pan &
A. D. Elbein, unpublished results). Thus the synthesis of the

higher homologs of trehalose may be analogous to synthesis
of raffinose and stachyose from sucrose [9]. In that case the
donor of galactose to sucrose to form raffinose is galactinol,
i.e. galactosyl myoinositol, rather than the expected UDP-
galactose.
As trehalose and its higher homologs may have key
roles as protectants or stabilizers of mycobacteria and
may also be important structural components, any of the
reactions involved in their biosynthesis represent excellent
potential target sites for chemotherapeutic intervention in
tuberculosis. However, in mycobacteria there may be as
many as three different pathways for the formation of
trehalose. The well-established pathway involves forma-
tion of trehalose 6-phosphate by transfer of glucose from
UDP-glucose to glucose 6-phosphate by the trehalose
phosphate synthase [7], and then removal of the phos-
phate by a highly specific trehalose phosphate phospha-
tase to give free trehalose (S. Klutts, I. Pastuszak, D.
Carroll, Y. T. Pan & A. D. Elbein, unpublished results).
Both of these enzymes have been purified and character-
ized from M. smegmatis [7,8] and other mycobacteria.
Two other pathways have been proposed on the basis of
sequence homologies between established enzymes and the
sequences obtained from the Mycobacterium tuberculosis
genome [24]. One of these pathways involves the direct
rearrangement of the a1,4-glucosidic linkage of maltose to
the a,a1,1 linkage of trehalose, and this enzyme has been
demonstrated in Pimelobacter species [25]. A third path-
way involves the generation of trehalose from glycogen,
and this pathway was reported in Arthrobacter species

[26]. These alternative pathways have not been demon-
strated in mycobacteria, but, if they do exist, they could
represent different ways to make trehalose for different
functions in the cell. That is, one pathway could provide
trehalose for a structural role and another pathway could
produce it and its oligosaccharides to act as stabilizers.
ACKNOWLEDGEMENTS
This research was supported in part by NIH RO3 AI43292 to A.D.E.
REFERENCES
1. Brennan, P.J. & Nikaido, H. (1995) The envelope of Myco-
bacteria. Annu. Rev. Biochem. 64, 29–63.
2. Kato, M. & Maeda, J. (1974) Isolation and biochemical activities
of trehalose-6- monomycolate of Mycobacterium tuberculosis.
Infect. Immun. 9, 8–14.
3. Crick, D.C., Mahapatra, S. & Brennan, P.J. (2001) Biosynthesis
of arabinogalactan-peptidoglycan comples of Mycobacterium
tuberculosis. Glycobiology 11, 107–118.
4. Belanger, A.E., Bersa, G.S., Ford, M.E., Mikusova, K., Belisle,
J.T., Brennan, P.J. & Inamine, J.M. (1996) The embAB genes
encode an arabinosyl transferase involved in cell wall arabinan
biosynthesis that is the target for the antimycobacterial drug
ethambutol. Proc. Natl. Acad. Sci. USA 93, 11919–11924.
5. Crowe, J.H., Crowe, L.M. & Chapman, D. (1984) Preservation of
membranes in anhydrobiotic organisms: role of trehalose. Science
223, 701–703.
6. Singer, M.A. & Lindquist, S. (1998) Multiple effects of trehalose
on protein folding in vitro and in vivo. Mol. Cell 1, 639–648.
7. Pan, Y.T., Drake, R.R. & Elbein, A.D. (1996) Trehalose-P syn-
thase of mycobacteria: its substrate specificity is affected by
polyanions. Glycobiology 6, 453–461.

8. Reference withdrawn.
9. Peterbauer, T., Mucha, J., Mach, L. & Richter, A. (2002)
Chain elongation of raffinose in pea seeds. J. Biol. Chem. 277, 194–
200.
10. Loewus, F. (1952) Improvement of anthrone method for deter-
mination of carbohydrates. Anal Chem. 24, 219–224.
11. Trevelyan, W.E., Proctor, D.P. & Harrison, J.S. (1950) Detection
of sugars on paper chromatograms. Nature (London) 166, 444–
445.
12. Nelson, N. (1944) A photometric adaptation of the Somogyi
method for the determination of glucose. J. Biol. Chem. 153, 375–
380.
13. Ohta, M., Matsura, F., Kobayashi, Y., Shigeta, S., Ono, K. &
Oka, S. (1991) Further characterization of allergenically active
oligosaccharitols isolated from sea squirt H antigen. Arch. Bio-
chem. Biophys. 290, 474–483.
14. Ciucanu, I. & Kerek, F. (1984) A simple and rapid method for the
permethylation of carbohydrates. Carbohydr. Res. 131, 209–217.
15. Liang, C.J., Yamashita, K., Muellenberg, C.G., Shichi, H. &
Kobata, A. (1979) Structure of the carbohydrate moieties of
bovine rhodopsin. J. Biol. Chem. 254, 6414–6418.
16. Hunter,S.W.,Murphy,R.C.,Clay,K.,Goren,M.B.&Brennan,
P.J. (1983) Trehalose-containing lipooligosaccharides: a new class
of species specific antigens from mycobacteria. J. Biol. Chem. 258,
10481–10487.
17. Hendrix, D.L. & Wei, Y.A. (1994) Beminose: an unusual
trisaccharide in Bemisia honeydew. Carbohydr. Res. 253, 329–
334.
18. Bersa, G.S., McNeil, M.R. & Brennan, P.J. (1992) Characteriza-
tion of the specific antigenicity of Mycobacterium fortuitum. Bio-

chemistry 31, 6504–6509.
19. Powalla, M., Lang, S. & Wray, V. (1989) Penta- and disaccharide
lipid formation by Nocardia corynebacteroides grown on
n-alkanes. Appl. Microbiol. Biotechnol. 31, 473–479.
20. Nair, V.G., Joseph, J.P. & Bernstein, S. (1982) O-b-D (and O-b-a)
multigalactopyranosyl, xylopyranosyl, and glucopyranosyl sulfate
salts. US Patent Application 1–21 CCSD accession number 13555.
21. Nakada, T., Maruta, K., Tsusaki, K., Kubota, M., Chaen, H.,
Sugimoto, T., Kurimoto, M. & Tsujisaka, Y. (1995) Purification
and properties of a novel enzyme, maltooligosyl trehalose synthase
from Arthrobaacter species. Biosci. Biotechnol. Biochem. 59, 2210–
2214.
22. Ajisaka, K. & Fujimoto, H. (1990) Regioselective synthesis of
trehalose-containing trisaccharide using various glycohydrolases.
Carbohydr. Res. 199, 227–234.
23. Muramatsu, M. & Nakakkuki, T. (1995) Enzymatic synthesis of
novel fructosyl and oligofructosyl trehaloses in Arthrobacter spe-
cies. Biosci. Biotechnol. Biochem. 59, 208–212.
24. De Smet, K.A.L., Weston, A., Brown, I.N., Young, D.B. &
Robertson, B.D. (2000) Three pathways for trehalose biosynthesis
in mycobacteria. Microbiology 146, 199–208.
25. Tsusaki, K., Nishimoto, T., Nakada, T., Kubota, M., Chaen, H.,
Sugimoto, T. & Kurimoto, M. (1996) Cloning and sequencing of
trehalose synthase gene. Biochim. Biophys. Acta 1290, 1–3.
26. Murata, K., Mitsuzumi, H., Nakada, T., Kubota, M., Chaen, H.,
Fukuda, S., Sugimoto, T. & Kurimoto, M. (1996) Cloning and
sequencing of a cluster of genes encoding enzymes of trehalose
biosynthesis. Biochim. Biophys. Acta 1291, 177–181.
Ó FEBS 2002 Cytoplasmic trehalose oligosaccharides (Eur. J. Biochem. 269) 3149