Physicochemical characterization and biological activity
of a glycoglycerolipid from
Mycoplasma fermentans
Klaus Brandenburg
1
, Frauke Wagner
1
, Mareike Mu¨ ller
1
, Holger Heine
1
,Jo¨ rg Andra¨
1
, Michel H. J. Koch
2
,
Ulrich Za¨ hringer
1
and Ulrich Seydel
1
1
Forschungszentrum Borstel, Center for Medicine and Biosciences, Borstel;
2
European Molecular Biology Laboratory,
Outstation Hamburg, Hamburg, Germany
We report a comprehensive physicochemical characteriza-
tion of a glycoglycerolipid from Mycoplasma fermentans,
MfGl-II, in relation to its bioactivity and compared this with
the respective behaviors of phosphatidylcholine (PC) and a
bacterial glycolipid, lipopolysaccharide (LPS) from deep
rough mutant Salmonella minnesota strain R595. The b«a

as animals, plants and humans, where they cause several
invasive or chronic diseases [1–3]. M. fermentans was first
isolated from the human urogenital tract [4], and since then
its role as pathogen and cofactor in diverse diseases has
emerged, in particular its role in the pathogenesis of
rheumatoid arthritis [5]. In recent years it was suggested
that M. fermentans is involved in triggering the develop-
ment of AIDS in HIV-positive individuals, acting as a
cofactor in pathogenesis [6]. Although little is known about
the molecular mechanisms underlying M. fermentans
pathogenicity, it is reasonable to assume that the inter-
actions with host cells are mediated by components of its
plasma membrane [7–9]. Matsuda et al.isolatedtwo
phosphocholine-containing glycoglycerolipids [10] and
elucidated the structure of one as 6¢-O-phosphocholine-
a-glucopyranosyl-(1¢-3)-1,2-diacyl-sn-glycerol (MfGl-I) [11].
Recently, we identified and characterized a major glyco-
glycerolipid from the membrane of M. fermentans which
was found to be 6¢-O-(3¢¢-phosphocholine-2¢¢-amino-
1¢¢-phospho-1¢¢,3¢¢-propanediol)-a-
D
-glucopyranosyl-(1¢-3)-
1,2-diacyl-sn-glycerol (MfGl-II) [12]. Furthermore, we
could show that MfGl-II triggers inflammatory response
in primary rat astrocytes such as activation of protein kinase
C, secretion of nitric oxide and prostaglandin E2 as well as
augmented glucose utilization and lactate formation [11].
These data were supported by others [13,14].
From these findings, the elucidation of molecular
mechanisms underlying or mediating these activities on a

transform infrared (FTIR) spectroscopy was applied to
determine the phase behavior via the analysis of the
peak position of the symmetric stretching vibration of
the methylene groups. Additionally, this technique was
applied to monitor the conformation of headgroup
moieties such as phosphate. The data obtained for
MfGl-II are related to those from LPS and phospha-
tidylcholine (PC) and show characteristic differences
between these amphiphiles. Synchrotron radiation small-
angle X-ray diffraction was applied for the determination
of the aggregate structure, and from the diffraction
patterns the existence of mixed unilamellar/nonlamellar
aggregate structures are deduced similar to those observed
for lipid A. We furthermore show by fluorescence reso-
nance energy transfer (FRET) technique that, analogously
to LPS, intercalation of MfGl-II in negatively charged
membrane systems such as liposomes made from phos-
phatidylserine (PS) can be mediated by lipopolysaccha-
ride-binding protein (LBP). In biological tests, we can
show that MfGl-II is able to induce tumor necrosis factor-
a (TNF-a) in human mononuclear cells, whereas in the
LPS-specific Limulus amebocyte lysate test no activity is
observed, which indicates that no LPS contamination is
present.
With these data, our conformational concept of endo-
toxicity [17] can for the first time be successfully applied also
to a nonlipid A structure.
Materials and methods
Growth of bacteria
Mycoplasma fermentans strain PG18 was grown in a

nonpolar lipids was achieved by HPLC on Nucleosil
column (10 · 500 mm, Nucleosil 50-7, Macherey-Nagel).
Crude lipid extracts (20 mg) were applied to the column and
eluted with a linear gradient of solvent A (chloroform/
methanol 1 : 4, v/v) and solvent B (chloroform/methanol/
water 1 : 4 : 2.5, v/v/v) starting with 0% solvent B for
30 min, then stepwise increasing to 15% B (150 min), 50%
B (10 min), holding 20 min 50% solvent B at a flow rate of
2mLÆmin
)1
(35 bar). Fractions were collected for 2 min
each and analyzed by TLC (chloroform/methanol/water
100 : 100 : 30, v/v/v). MfGl-II eluted as the last lipid,
appropriate fractions (nos 36–60) were combined,
R
f
¼ 0.17 (yield 4.16 mg).
Lipid samples
LPS from deep rough mutant Re from Salmonella
minnesota strain R595 was extracted according to PCP I:
2% phenol/5% chloroform/8%petrol ether, v/v) proce-
dure [19], purified by treatment with DNAse/RNAse and
proteinase K, and lyophilized and used in the natural salt
form. The lipopeptide palmitoyl-3-cysteine-serine-lysine-4
(Pam3CSK4) and the macrophage-activating lipopeptide-2
(MALP-2) were kind gifts of K H. Wiesmu
¨
ller
1
(Tu

(CH
2
),
which lies around 2850 cm
)1
in the gel and between
2852 cm
)1
and 2853 cm
)1
in the liquid-crystalline phase
[20,21].
Lipid headgroup conformation
For a characterization of the conformation of functional
groups within the lipid backbones such as the phosphate,
lipid suspensions were prepared as described above.
Subsequently, 10 mL were spread on a CaF
2
crystal and
allowed to stand at room temperature until all free water
was evaporated. After this, IR spectra were recorded at
room temperature and at 37 °C. Usually, the original
spectra were evaluated directly and a spectral analysis was
performed in the fingerprint region between 1800 and
900 cm
)1
. In the case of overlapping absorption bands,
either resolution enhancement techniques like Fourier self-
deconvolution [22] or a curve-fit analysis [23] were
performed.

(NBD-PE) and N-(lissamine rhodamine B sulfonyl)-phos-
phatidylethanolamine (Rh-PE) (Molecular Probes, Eugene,
OR, USA), respectively. Intercalation of unlabeled mole-
cules into the double-labeled liposomes leads to probe
dilution and with that to a decrease in the efficiency of
FRET: the emission intensity of the donor increases
and that of the acceptor decreases (for clarity, only the
donor emission intensity is shown). The double-labeled
liposomes were preincubated with unlabeled LPS and
recombinant human lipopolysaccharide-binding protein
LBP was added.
Endotoxin activity determination by the chromogenic
Limulus test
Endotoxin activity of the glycolipids was determined by a
quantitative kinetic assay [27] based on the reactivity of
Gram-negative endotoxin with Limulus amebocyte lysate
(LAL) using test kits from BioWhittaker (# 50–650 U).
Induction of tumor necrosis factor-a
For the isolation of mononuclear cells (MNC), blood was
taken from healthy donors and heparinized (20 IEÆmL
)1
).
The heparinized blood was mixed with an equal volume of
Hank’s balanced salt solution and centrifuged on a Ficoll
density gradient for 40 min (21 °C, 500 g)
2
.Theinterphase
layer of mononuclear cells was collected and washed
three times in serum-free RPMI 1640 containing 2 m
M

6b from Intex, Germany). Cell culture supernatants and
the standard (recombinant TNF, Intex) were diluted with
buffer. The plates were shaken 16–24 h at 4 °C.Forthe
removal of free antibodies, the plates were washed six
times in distilled water. Subsequently, the color reaction
was started by addition of tetramethylbenzidine in alco-
holic solution and stopped after 5–15 min by addition of
0.5
M
sulfuric acid. In the color reaction, the substrate is
cleaved enzymatically, and the product can be measured
photometrically. This was carried out on an ELISA
reader (Rainbow, Tecan, Crailsham, Germany) at a
wavelength of 450 nm, and the values were related to
the standard.
Cell lines
The CHO/CD14 reporter line, clone 3E10, is a stably
transfected CD14-positive CHO cell line that expresses
inducible membrane CD25 (Tac antigen) under transcrip-
tional control of the human E-selectin promoter (pE-
LAM.Tac [28]). The CHO/CD14/huTLR2 (3E10TLR2)
reporter cell line was constructed by stable cotransfection of
3E10 with the cDNA for human TLR2 and pcDNA3
(Invitrogen), as described [29]. CHO cell lines were grown in
Ham’s F12 medium containing 10% fetal bovine serum and
1% penicillin/streptomycin at 37 °C in a humidified 5%
CO
2
environment. Medium was supplemented with
400 UÆmL

(CH
2
) revealed a phase transition at around 30 °C for
MfGl-II similar to that of deep rough mutant LPS from
S. minnesota strain R595 (Fig. 2). The entire phase behavior
of MfGl-II and LPS is very similar except that the
wavenumber values are lower for the latter indicating a
slightly higher overall acyl chain order. In contrast, natural
PC exhibits high wavenumber values over the entire
temperature range, from which the existence of only the
unordered a-phase can be concluded.
The infrared spectrum of MfGl-II in the fingerprint
region (Fig. 3a) displays strong bands at 1739 cm
)1
corres-
ponding to the ester double bond stretching m (C ¼ O), the
lipid scissoring band d (CH
2
) at 1465 cm
)1
,theantisym-
metric and symmetric stretching vibrations of the negatively
charged phosphate groups m
as
(PO
2
–
)at1220cm
)1
and m

2
–
) 1300–
1190 cm
)1
(B)ofhydratedPC,MfGl-II,andLPSRe.
3274 K. Brandenburg et al. (Eur. J. Biochem. 270) Ó FEBS 2003
lipopolysaccharides can be excluded. The band contour of
m
as
(PO
2
–
) was analyzed after baseline subtraction (Fig. 3b)
and revealed strong differences between MfGl-II, LPS Re,
and PC (in this case dimyristoyl PC).
LPS exhibits two band components, one at higher
wavenumber (1260 cm
)1
) with relatively low bandwidth,
corresponding to phosphate with low hydration, and one
broader band component (at 1223 cm
)1
), corresponding to
higher hydration [23]. The spectrum for PC shows the
occurrence of one major band around 1225 cm
)1
in
accordance with the well-known high water-binding capa-
city of lecithin headgroups [30]. Finally, MfGl-II exhibits a

between 0.1 and 0.3/nm can be interpreted by the existence
of a unilamellar structure. The location of the four small
peaks superimposed fit the relations 17.0 ¼ 8.48 Ö2,
16.9 ¼ 6.90 Ö6, 17.0 ¼ 4.90 Ö12, 17.0 ¼ 3.47 Ö24, which
can be assigned to a cubic structure with a periodicity
a
Q
¼ (16.95 ± 0.10) nm. The space group, however, can-
not be determined due to a lack of observable reflections.
From these findings, a superposition of a main unilamellar
with a nonlamellar cubic structure can be deduced, which
would correspond to a very slight conical conformation of
the individual molecules with different cross-sections of the
hydrophobic and the hydrophilic moieties. From these data,
however, no unequivocal statement is possible which of
the moieties has a higher cross-section.
Fig. 4. NBD-donor fluorescence intensity as function of time of double-
labeled liposomes made from PS (A) or from PC (B) after the addition of
MfGl-IIorLPSReatt = 50 s and subsequent addition of LBP
(0.2 m
M
)att = 100 s in comparison to control NaCl/P
i
(phosphate
buffered saline). The concentration of the glycolipids, PC and PS was
10 m
M
each.
Fig. 5. Synchrotron radiation X-ray small-angle diffraction pattern of
MfGl-II at 40 °C and 85% water content. The diffraction pattern

, while for MfGl-II a significant response is
found down to 100 ngÆmL
)1
, thus indicating that MfGl-II,
although two orders of magnitude less active than LPS, still
induces cytokines to a significant degree.
CHO reporter system
In order to investigate the potential involvement of TLR2
and TLR4 in the recognition and signal transduction of the
glycolipids, we analyzed the stimulatory activity of MfGl-II
in a CHO cell reporter system. Upon the induction of
nuclear factor-kappa B translocation in these reporter cells,
human CD25 is expressed on the cell surface [28]. The data
(Fig. 7) clearly indicate that neither the expression of CD14
and TLR4 (3E10) nor the expression of CD14, TLR2, and
TLR4 (3E10 TLR2) is sufficient to enable the cells to
respond to MfGl-II even at the highest concentration
10 mgÆmL
)1
. As controls, stimulation of the different cell
lines with either LPS from Salmonella friedenau or the
lipopeptides from synthetic (Pam3CysSerLys4) or natural
origin (MALP-2) showed the expected phenotype, i.e. LPS
reacts essentially to TLR4, while the lipopeptide or protein
exhibit TLR2-reactivity. Thus, a possible contamination of
the MfGl-II with a lipoprotein or MALP-2 [32], which
could explain the cytokine-inducing capacity, can be
excluded.
Discussion
Mycoplasma fermentans has been reported to accompany

of LPS [35]. This refers to the phase transition behavior and
the fluidity of the glycolipid chains at 37 °C (Fig. 1), the
Fig. 6. Induction of TNF-a in human mononuclear cells by MfGl-II and
LPS Re as function of glycolipid concentration. The error bar (standard
deviation) results from the determination of TNF-a in duplicate at two
different dilutions. The data are representative of three independent
measurements.
Fig. 7. Relative activation of CHO reporter cells stimulated with MfGl-
II, the lipopeptide Pam3CysSerLys4, LPS S-form from Salmonella
friedenau, the MALP-2 (macrophage activating lipoprotein), and inter-
leukin-1. The IL-1 induced expression of NF-6B reporter signal was set
to 100%.
3276 K. Brandenburg et al. (Eur. J. Biochem. 270) Ó FEBS 2003
strong LBP-induced incorporation into negatively charged
phospholipid liposomes (Fig. 4), and a diffraction pattern,
which is consistent with the existence of a unilamellar
superimposed by a cubic structure (Fig. 5). Thus, according
to our concept of an endotoxic activity, which requires a
cubic supramolecular aggregate structure corresponding to
a conical conformation of the individual molecules and an
LBP-driven incorporation into target cell membranes, for
which a sufficiently high negative charge is needed, MfGl-II
is a candidate as immunostimulating agent. Actually, MfGl-
II induced TNF-production, although to a lower degree
than LPS (Fig. 6).
Although we presently cannot answer the question
whether the cubic structure observed is of the ÔnormalÕ
(right side out) or the inverted type, the geometry of the
molecule with its bulky headgroup is in favor of the former
type. With respect to a correlation to bioactivity, we have

which shows activation in the range ‡1mgÆmL
)1
[38]. This
glycolipid was found to stimulate human MNCs in a CD14-
independent way, and the response could not be blocked by
antagonistic lipid A part structures, therefore indicating a
completely different activation pathway [39].
In contrast, the results from the FRET measurements
(Fig. 4) indicate a signaling pathway identical to that of
LPS. This may be explained by the fact that MfGl-II as
well as lipid A, the endotoxic principle of LPS, exhibits a
high negative charge density due to the presence of two
phosphates. GSL molecules have only one negative
charge, a glucuronic acid. Whether the kind of charge,
phosphate or uronic acid, plays a role in endotoxin
signaling, cannot be answered unequivocally. We found
earlier that a lipid A analogue in which the 1-phosphate is
substituted by a carboxymethyl group (CM-506), exhibits
the same activity as natural Escherichia coli-type lipid A
or its synthetic analogue 506 [15]. In contrast, the lipid A
from Rhodospirillum fulvum, in which the 1-phosphate is
substituted by a heptose and the 4¢-phosphate by a
galacturonic acid, is biologically, i.e. agonistically as well
as antagonistically, completely inactive. The lack of
antagonistic activity may be explained by the fact that
this lipid A does not intercalate into target cell membranes
by LBP-mediated transport [35].
We have shown recently that endotoxin aggregates are
the active units, i.e. they are at least one order of magnitude
more active than monomers [40]. Furthermore, we have

unilamellar and cubic structures is not possible using the
NMR technique.
Acknowledgments
We are indepted to G. von Busse, S. Groth, and U. Diemer for
performing the IR spectroscopic, TNF-induction and LAL activity
measurements, respectively.
This work was financially supported by the Deutsche Forschungsg-
emeinschaft (SFB 367 project B8) and by the German-Israeli
foundation for Scientific Research and Development (GIF grant
I-373-169-09/94).
References
1. Lee, I M. & Davis, R.E. (1992) Mycoplasmas with infect plants
and insects. In Mycoplasmas ) Molecular Biology and Pathogen-
esis (Maniloff, J., McElhaney, R.N., Finelli, M.R. & Baseman,
J.B., eds), pp. 379–390. American Society for Microbiology,
Washington, DC.
2. Simecka, J.W., Davis, J.K., Davidson, M.K., Ross, S.E., Sta-
dtla
¨
nder, C.T.K.H. & Cassell, G.H. (1992) Mycoplasma diseases
an animals. In Mycoplasmas – Molecular Biology and Pathogenesis
(Maniloff, J., McElhaney, R.N., Finch, L.R. & Baseman, J.B.,
Ó FEBS 2003 Biophysical study of Mycoplasma glycolipid (Eur. J. Biochem. 270) 3277
eds), pp. 391–415. American Society for Microbiology, Wash-
ington, DC.
3. Krause, D.C. & Taylor-Robinson, D. (1992) Mycoplasmas which
infect humans. In Mycoplasmas – Molecular Biology and Patho-
genesis (Maniloff,J.,McElhaney,R.N.,Finch,L.R.&Baseman,
J.B., eds), pp. 417–444. American Society for Microbiology,
Washington, DC.

269, 33123–33128.
12. Za
¨
hringer, U., Wagner, F., Rietschel, E.Th, Ben-Menachem., G.,
Deutsch, J. & Rottem, S. (1997) Primary structure of a new
phosphocholine-containing glycoglycerolipid of Mycoplasma fer-
mentans. J. Biol. Chem. 272, 26262–26270.
13. Matsuda, K., Li, J L., Harasawa, R. & Yamamoto, N. (1997)
Phosphocholine-containing glycoglycerolipids (GGPL-1 and
GGPL-III) are species-specific major immunodeterminants of
Mycoplasma fermentans. Biochem. Biophys. Res. Commun. 233,
644–649.
14.Li,J L.,Matsuda,K.,Takagi,M.&Yamamoto,N.(1997)
Detection of serum antibodies against phosphocholine-containing
aminoglycoglycerolipid specific to Mycoplasma fermentans in
HIV-1 infected individuals. J. Immunol. Methods 208, 103–113.
15. Seydel, U., Oikawa, M., Fukase, K., Kusumoto, S. & Branden-
burg, K. (2000) Intrinsic conformation of lipid A is responsible for
agonistic and antagonistic activity. Eur. J. Biochem. 267, 3032–3039.
16. Brandenburg,K.,Schromm,A.B.,Koch,M.H.J.&Seydel,U.
(1995) Conformation and fluidity of endotoxins as determinants
of biological activity. In Bacterial Endotoxins: Lipopolysaccharides
from Genes to Therapy (Levin, J., Alving, C.R., Munford, R.S. &
Redl, H., eds), pp. 167–182. John Wiley & Sons, New York.
17. Schromm, A.B., Brandenburg, K., Loppnow, H., Moran, A.P.,
Koch, M.H.J., Rietschel, E.Th & Seydel, U. (2000) Biological
activities of lipopolysaccharides are determined by the shape of
their lipid A portion. Eur. J. Biochem. 267, 2008–2013.
18. Bligh, E.G. & Dyer, W.J. (1959) A rapid method of total lipid
extraction and purification. Can. J. Biochem. Physiol. 37, 911–917.

polysaccharide into phospholipid membranes. FEBS Lett. 399,
267–271.
27. Friberger, P., So
¨
rskog, L., Nilsson, K. & Kno
¨
s, M. (1987) The use
of a quantitative assay in endotoxin testing. Progr. Clin. Biol. Res.
231, 49–169.
28. Delude, R.L., Yoshimura, A., Ingalls, R.R. & Golenbock, D.T.
(1998) Construction of a lipopolysaccharide reporter cell line and
its use in identifying mutants defective in endotoxin, but not
TNFa, signal transduction. J. Immunol. 161, 3001–3009.
29. Yoshimura, A., Lien, E., Ingalls, R.R., Tuomanen, E., Dziarski,
R. & Golenbock, D. (1999) Cutting edge: recognition of Gram-
positive bacterial cell wall components by the innate immune
system occurs via Toll-like receptor 2. J. Immunol. 163, 1–5.
30. Fringeli, U.P. & Gu
¨
nthard, H.H. (1981) Infrared membrane
spectroscopy. In Membrane Spectroscopy (Grell, E., ed.), pp.
270–332. Springer-Verlag, Berlin.
31. Brandenburg, K. & Seydel, U. (2002) Vibrational spectroscopy of
carbohydrates and glycoconjugates. In Handbook of Vibrational
Spectroscopy, Vol. 5 (Chalmers, J.M. & Griffiths, P.R., eds),
pp. 3481–3507. John Wiley & Sons, Chicester.
32. Mu
¨
hlradt, P.F. & Frisch, M. (1994) Purification and partial bio-
chemical characterization of a Mycoplasma fermentans-derived

chem. 267, 3370–3377.
37. Takeuchi, O. & Akira, S. (2001) Toll-like receptors; their physio-
logical role and signal transduction system. Int. Immunopharma-
col. 1, 625–635.
38. Kawahara, K., Seydel, U., Matsuura, M., Danbara, H., Rietschel,
E.T. & Za
¨
hringer, U. (1991) Chemical structure of glyco-
sphingolipids isolated from Sphingomonas paucimobilis. FEBS
Lett. 292, 107–110.
39. Krziwon, C., Za
¨
hringer, U., Kawahara, K., Weidemann, B.,
Kusumoto, S., Rietschel, E.T., Flad, H D. & Ulmer, A.J. (1995)
Glycosphingolipids from Sphingomonas paucimobilis induce
monokine production in human mononuclear cells. Infect.
Immunol. 63, 2899–2905.
40. Mu
¨
ller, M., Scheel, O., Lindner, B., Gutsmann, T. & Seydel, U.
(2002) The role of membrane-bound LBP, endotoxin aggregates,
and the MaxiK channel in LPS-induced cell activation.
J. Endotoxin Res. 9, 181–186.
41. Gutsmann, T., Mu
¨
ller, M., Carroll, S.F., MacKenzie, R.C.,
Wiese, A. & Seydel, U. (2001) Dual role of lipopolysaccharide
(LPS)-binding protein in neutralization of LPS and enhancement
of LPS-induced activation of mononuclear cells. Infect. Immunol.
69, 6942–6950.