In addition, de novo biosynthesis of glycosphingolipids was shown by metabolic incorporation of [14C]palmitic acid and [14C]glucose in the three intraerythrocytic stages of the parasite.
Trang 1Glycosphingolipids in Plasmodium falciparum
Presence of an active glucosylceramide synthase
Alicia S Couto1, Carolina Caffaro1, M Laura Uhrig1, Emilia Kimura2, Valnice J Peres2, Emilio F Merino2, Alejandro M Katzin2, Masae Nishioka3, Hiroshi Nonami3and Rosa Erra-Balsells1
1
CIHIDECAR, Departamento de Quı´mica Orga´nica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Argentina;2Departamento de Parasitologia, Instituto de Cieˆncias Biome´dicas, Universidade de Sa˜o Paulo, Brazil;3College of Agriculture, Ehime University, Matsuyama, Japan
Malaria remains a major health problem especially in
trop-ical and subtroptrop-ical regions of the world, and therefore
developing new antimalarial drugs constitutes an urgent
challenge Lipid metabolism has been attracting a lot of
attention as an application for malarial chemotherapeutic
purposes in recent years However, little is known about
glycosphingolipid biosynthesis in Plasmodium falciparum
In this report we describe for the first time the presence of
an active glucosylceramide synthase in the intraerythrocytic
stages of the parasite Two different experiments, using
UDP-[14C]glucose as donor with ceramides as acceptors, or
UDP-glucose as donor and fluorescent ceramides as
accep-tors, were performed In both cases, we found that the
parasitic enzyme was able to glycosylate only
dihydrocera-mide The enzyme activity could be inhibited in vitro with
low concentrations of D,L
-threo-phenyl-2-palmitoylamino-3-morpholino-1-propanol (PPMP) In addition, de novo
biosynthesis of glycosphingolipids was shown by metabolic incorporation of [14C]palmitic acid and [14C]glucose in the three intraerythrocytic stages of the parasite The structure
of the ceramide, monohexosylceramide, trihexosylceramide and tetrahexosylceramide fractions was analysed by UV-MALDI-TOF mass spectrometry When PPMP was added
to parasite cultures, a correlation between arrest of parasite growth and inhibition of glycosphingolipid biosynthesis was observed The particular substrate specificity of the malarial glucosylceramide synthase must be added to the already known unique and amazing features of P falciparum lipid metabolism; therefore this enzyme might represent a new attractive target for malarial chemotherapy
Keywords: dihydroceramide; glucosylceramide synthase; glycosphingolipids; malaria; Plasmodium falciparum
Malaria is the most serious and widespread parasitic disease
in humans Each year, approximately 300 million people
become infected and 2–3 million people die as a result In
addition there is considerable morbidity associated with this
disease [1]
The glycobiology of Plasmodium falciparum has been
causing an increasing amount of interest in recent years The
presence of N-linked glycoproteins in relation to schizogony
of the intraerythocytic stages [2] and
glycosylphosphatidyl-inositols as the major carbohydrate protein modification
have been described in the human malaria parasite [3–6] In addition, lipid metabolism has also been attracting a lot
of attention with respect to basic biology and applications for malarial chemoterapeutic purposes [7] However, little
is known about glycosphingolipids (GSLs), a group of ceramide-based lipids that in other systems regulate inter-actions of the cell with its environment and play a role in cell signalling [8,9] The first evidence of the presence of GSLs in
P falciparum was obtained by metabolic incorporation
of [3H]serine and [3H]glucosamine After labeling with the carbohydrate precursor, hydrophilic glycosphingolipids migrating slower than the penta-glycosylated ceramide standard were detected [10] More recently, the synthesis
of chloroplast galactolipids in apicomplexan parasites was reported [11]
Biosynthesis of complex GSLs in mammalian cells involves sequential glycosyltransferase reactions, starting with the formation of glucosylceramide (GlcCer), and it has been assumed that the various transferases used are functionally organized within the Golgi [12,13] It is known that the key step involves the transfer of glucose to ceramide from UDP-glucose, catalyzed by the action of a glucosyl-ceramide transferase [EC 2.4.1.80: glucosylglucosyl-ceramide syn-thase (GCS)] With regards to localization, as far as it is known, GlcCer is special because it is the only glyco-sphingolipid synthesized on the cytosolic leaflet in the early Golgi but it is used for the synthesis of higher sphingolipids
Correspondence to A S Couto, CIHIDECAR, Departamento de
Quı´mica Orga´nica, Pabello´n II, Facultad de Ciencias Exactas y
Nat-urales, Universidad de Buenos Aires, Buenos Aires, 1428, Argentina.
Fax/Tel.: + 54 11 4576 3346, E-mail: acouto@qo.fcen.uba.ar
Abbreviations: GSLs, glycosphingolipids; GlcCer, glucosylceramide;
GCS, glucosylceramide synthase; DHCer,
BODIPY-dihydroceramide; BODIPY-Cer, BODIPY-ceramide;
D , L -threo-PPMP, D , L
-threo-phenyl-2-palmitoylamino-3-morpholino-1-propanol, d18:0, 4-hydroxysphinganine; d20:0,
4-hydroxyicosa-sphinganine; C10:0, etc., decanoic acid, etc.; C10h:0, etc.,
hydroxydecanoic acid; C10-2h:0, dihydroxydecanoic acid;
C10-3h:0, trihydroxydecanoic acid, etc.
Enzyme: glucosylceramide synthase (EC 2.4.1.80).
(Received 4 December 2003, revised 26 February 2004,
accepted 6 April 2004)
Trang 2in the lumenal leaflet [14] Because glucosylceramide is a
pivotal precursor of numerous GSLs, this enzyme is
extremely important for understanding GSL function
In this report, we describe for the first time the presence
of an active glucosylceramide synthase in the
intraerythro-cytic stages of P falciparum Two different experiments,
using UDP-[14C]glucose as donor or fluorescent ceramides
as acceptors were performed In both cases, the enzyme
showed specificity for dihydroceramide as substrate
The enzyme activity could be inhibited in vitro with
D,L
-threo-phenyl-2-palmitoylamino-3-morpholino-1-pro-panol (PPMP) In addition, GSLs were shown by
meta-bolic incorporation of [14C]palmitic acid and [14C]glucose
in the three intraerythrocytic stages of the parasite
UV-MALDI-TOF mass spectrometry proved that four
fractions analysed corresponded to ceramides,
monohexo-sylceramides, trihexosylceramides and
tetrahexosylcera-mides, respectively When PPMP was added to parasite
cultures, a correlation between arrest of parasite growth
and inhibition of GSL biosynthesis was shown The
particular substrate specificity of the malarial GCS suggests
that this enzyme might represent a new attractive target for
malarial chemotherapy
Materials and methods
Materials
Lipid standards and BSA were purchased from Sigma
AlbuMax I was obtained from Gibco BRL Life
Tech-nologies (New York, NY, USA) All solvents were of
analytical or HPLC grade Percoll was purchased from
Pharmacia Chemicals (Uppsala, Sweden).D,L
-threo-phenyl-2-palmitoylamino-3-morpholino-1-propanol (PPMP) was
from Matreya (Pleasant Gap, PA, USA) and ceramide
glycanase from GlyKo, BODIPY-sphingolipids used were
from Molecular Probes Polyclonal antibodies against
human GCS were a kind gift of D L Marks and R E
Pagano, Mayo Clinic Foundation, Rochester, MN, USA
TLC was performed on silica gel 60 precoated plates
(Merck) using the following solvent systems: (a)
chloro-form/methanol/water (65 : 25 : 3, v/v/v); (b) chloroform/
methanol/0.25% KCl (80 : 30 : 2, v/v/v); (b), chloroform/
methanol/1M NH4OH (40 : 10 : 1, v/v/v); (d)
chloro-form/methanol/water (65 : 25 : 3, v/v/v); (e) chloroform/
methanol/water (80 : 20 : 2, v/v/v) In all cases, radioactive
samples were located by fluorography at )70 C using
EN3HANCE (NEN) and KodakX-OMAT AR films
Cultures ofP falciparum and metabolic labeling
An isolate (S20) of P falciparum obtained from a patient
living in Porto Velho (Rondoˆnia, Brazil) was used [15]
Parasite cultures of P falciparum were performed as
described [2]
[U-14C]Palmitic acid (Amersham 822 mCiÆmmol)1,
2.91 mCiÆmg)1) originally supplied in toluene was dried
under nitrogen, redissolved in ethanol, and coupled with
defatted BSA at a 1 : 1 (v/v) molar ratio The final labeling
medium contained 6.25 lCiÆmL)1of the radioactive
pre-cursor, 0.5% (w/v) Albumax and 0.05% (w/v) BSA
Parasites were labeled for 18 h
D-[U-14C]glucose (Amersham, 291 mCiÆmmol)1, 1,54 mCiÆmg)1) was incorporated at a concentration of 6.25 lCiÆmL)1 in RPMI 1640 medium without addition
of 11 mM of glucose Parasites (5.9% ring forms, 5.4% trophozoites, 3.7% schizonts) were labeled for 18 h
D-[U-14C]galactose (Amersham, 306 mCiÆmmol)1, 1,61 mCiÆmg)1) was incorporated at a concentration of 3.2 lCiÆmL)1in RPMI 1640 medium without addition of
11 mM of glucose Parasites (9.8% ring forms, 3.0% trophozoites, 1.6% schizonts) were labeled for 18 h The viability of the parasites was verified by microscopic evaluation of Giemsa stained smears Each stage was purified on a 40/70/80% (w/v) discontinuous Percoll gradient (15 000 g, 30 min, 25C) This procedure yielded
an upper band (40%) containing schizonts, another band with trophozoites (70–80% interface) and a pellet of ring forms [2]
A control containing a similar number of uninfected erythrocytes was incubated with the different radioactive precursors and further processed under the same conditions
In order to evaluate biosynthesis of proteins, synchronous
P falciparum ring-stage cultures, with a parasitemia of around 5%, untreated or treated with 5 lMPPMP for 48 h, were labeled with 25 lCiÆmL)1 of L-[35S]methionine (> 1000 CiÆmmol)1) (Amersham) in 10 lM methionine-deficient RPMI medium, at the beginning or after 24 h
of treatment Aliquots were collected at different times (0–48 h), precipitated with 12% (w/v) trichloroacetic acid, and radioactivity was measured with a Beckman 5000 b-counter
Treatment of parasites withD,L -threo-phenyl-2-palmitoylamino-3-morpholino-1-propanol Parasite cultures (6.4% ring forms, 2.4% trophozoites, 1.2% schizonts) were incubated with 5 lM D,L -threo-phenyl-2-palmitoylamino-3-morpholino-1-propanol (D,L -threo-PPMP) After 24 h of treatment, parasites were labeled with [14C]palmitic acid or [14C]glucose for 18 h in the presence of the drug After the labeling period, each stage was purified on a Percoll gradient as described above and freeze-dried prior to lipid extraction The effect
on parasite development was monitored by microscopy of Giemsa-stained blood smears in two independent experi-ments
In all cases, control cultures without the inhibitor and
a similar amount of uninfected erythrocytes were labeled under the same conditions
Isolation and purification of glycosphingolipids Each intraerythrocytic stage of P falciparum was extracted with chloroform/methanol 1 : 1 (3· 1 mL) Each extract was fractionated by anionic exchange chromatography on a DEAE-Sephadex A-25 (acetate form) column, which was eluted with chloroform/methanol/water (30 : 60 : 8, v/v/v)
to recover neutral GSLs and zwitterionic lipids Anionic lipids were bulkeluted with chloroform/methanol/0.8M
NaAcO (30 : 60 : 8, v/v/v) The unbound fraction was evaporated to dryness and treated with 0.1M NaOH in methanol (500 lL), at 37C for 3 h The mixture was neutralized with HCl 1 in the presence of 1 phosphate
Trang 3buffer pH 7 (50 lL) to avoid over-acidification After
evaporation, salts were removed by reverse-phase
chro-matography, using a Sep-PackC-18 cartridge (Worldwire
Monitoring, Horsham, PA, USA) Acidic lipids were also
concentrated through a Sep-Packcartridge Purification of
neutral glycosphingolipids was achieved by
chromatogra-phy on silicic acid The sample was dissolved in chloroform
and loaded into a column of Unisil (7· 50 mm) which was
eluted with chloroform (20 mL), chloroform/methanol
(98 : 2, v/v, 20 mL) and chloroform/methanol (1 : 3, v/v,
25 mL) [16]
In another experiment, total lipids from schizont forms
were extracted and purified as described above The purified
neutral GSL fraction was analysed in parallel with an
analogous fraction obtained from [U-14C]palmitic acid
labeled parasites by TLC in solvent B Spots corresponding
to the ceramide fraction (I), monohexosylceramide fraction
(II), trihexosylceramide fraction (IV) and
tetrahexosylcera-mide fraction (V) were extracted from the plate and
analysed by UV-MALDI-TOF MS
Acid methanolysis and methylation
The sample was hydrolysed for 18 h at 80C with 12M
HCl/MeOH/water (3 : 29 : 4, v/v/v) The hydrolysate was
dried and the acid eliminated by several evaporations with
addition of water Methylation of fatty acids was carried out
with BF3/MeOH in dry toluene under nitrogen at 80C for
90 min [17]
Ceramide glycanase digestion
Samples were dissolved in 250 mMphosphate buffer pH 5.0
(100 lL) containing 1% (w/v) sodium cholate Ceramide
glycanase (from Macrobdella decora) (0.3 mU) was added
and digestion was performed at 37C for 18 h Lipids
were extracted with chloroform/methanol (1 : 1, v/v) and
analysed by TLC
Glucosylceramide synthase assay
Parasite homogenates were prepared in 0.1M sodium
phosphate buffer (pH 7.4) containing 5 mM MgCl2,
25 mM KCl, 1 mMphenylmethanesulfonyl fluoride, 1 mM
N-a-tosyl-L-lisine chloromethyl ketone hydrochloride
(TLCK) and 10 lgÆmL)1 leupeptine by probe sonication
three times with 10 pulses while on ice Liposomal substrate
was performed with dipalmitoylphosphatidylcholine
and ceramide (palmitoyldihydrosphingosine or
palmitoyl-sphingosine) (10 : 1, v/v) containing 0.1 nmol of ceramide
The constituent lipids were dissolved in
chloroform/meth-anol (1 : 1, v/v), vortexed and dried under nitrogen Lipids
were dispersed in 0.1M sodium phosphate buffer pH 7.4
by sonication at 0C
The reaction mixture consisted of UDP-[14C]glucose
(1 lCi, 319 mCiÆmmol)1, Amersham), 2 mM b-NAD and
the liposomal substrate (600 nmol lipid phosphorous) in
0.1Msodium phosphate buffer (pH 7.4) The cell
homo-genate (50–100 lg protein per tube) was added making a
total volume of 15 lL The mixture was incubated at 37C
for 5 h with shaking Incubations were stopped by freezing
and the mixtures were cleaned by passage through C18
cartridges Lipids were eluted with chloroform/methanol (1 : 1, v/v) and further analysed by TLC When the inhibition test was performed, PPMP (5 lM) was added to the reaction mixture Spots were quantified using a phos-phoimager (Molecular Analyst, Bio-Rad) withMOLECULAR ANALYSTsoftware
In another experiment, BODIPY analogues of ceramides were used BODIPY-dihydroceramide (BODIPY-DHCer) was synthesized from dihydrosphingosine and BODIPY acid according to Kok& Hoekstra [18] The enzyme assay was performed in tubes precoated with dipalmitoylphos-phatidylcholine (15 nmoles added in chloroform and dried down under nitrogen) by adding the fluorescent ceramide (300 ng per tube) precoupled to BSA, the parasite lysates (50–100 lg protein) and the assay buffer (0.1M sodium phosphate buffer pH 7.4 containing 25 mM KCl, 5 mM
MgCl2, 2.5 mMUDP-glucose and 2 mMb-NAD) in a final volume of 150 lL The reaction was incubated at 37C for
5 h with shaking The mixture was extracted with chloro-form/methanol (1 : 1, v/v) and analysed by TLC Spots were visualized using a Fuji LAS1000 densitometer equipped with
IMAGE GAUGE3.122, software, Fuji Film, Japan
All protein determinations were performed using Brad-ford’s method [19]
Immunoprecipitation Parasite lysates (1–2 mg protein) were incubated with GCS 1.2 antibody (which recognizes a region near the GCS C-terminus) [20] in buffer Tris/HCl pH 8.0 containing
150 mMNaCl, 0.5% (w/v) sodium deoxycholate and 0.1% (w/v) SDS, for 2 h at 5C Protein A-Sepharose (10% in the same buffer, 100 lL) was added and it was incubated for a further 60 min The mixture was centrifuged at 10 000 g and the immunoprecipitate was washed (3· 100 lL) The immunoprecipitates were dissolved in sample buffer and subjected to SDS/PAGE in 10% gels Western blot to poly(vinylidene difluoride) membrane was performed and blots were probed with anti-peptide polyclonal antibodies GS-5.1 (1/1500) which recognizes a region near the GCS N-terminus [20] followed by an anti-rabbit horseradish peroxidase secondary antibody and visualized using ECL (Amersham) enhanced chemiluminescence reagent UV-MALDI-TOF MS analysis
Matrices for UV-MALDI-TOF MS The b-carboline (9H-pyrido[3,4-b]indole), nor-harmane and 2,5-dihydroxy-benzoic acid were obtained from Aldrich Chemical Co Calibrating chemicals for UV-MALDI-TOF analysis a-Cyclodextrin (cyclohexaamylose, Mr972.9), b-cyclodex-trin (cycloheptaamylose, Mr1135.0), c-cyclodextrin (cyclo-octaamylose, Mr 1297.1), angiotensin I (Mr 1296.49), neurotensin (Mr1672.96) and bovine insulin (Mr5733.5) were purchased from Sigma-Aldrich
Solvents Methanol, ethanol, acetonitrile (Sigma-Aldrich HPLC grade) and trifluoroacetic acid (Merck) were used as purchased without further purification Water of very low conductivity (Milli Q grade; 56–59 nSÆcm)1with PURIC-S (ORUGANO Co., Ltd, Tokyo, Japan) was used
Trang 4UV-MALDI-TOF-MS experiments Measurements were
performed using a Shimadzu Kratos, Kompact MALDI 4
(pulsed extraction) laser-desorption time-of-flight mass
spectrometer (Shimadzu, Kyoto, Japan) equipped with a
pulsed nitrogen laser (kem¼ 337 nm; pulse width ¼ 3 ns),
tunable pulse delay extraction (PDE), post source decay
(PSD) (MS/MS device) and a secondary electron multiplier
Experiments were first performed using the full range
setting for laser firing position in order to select the optimal
position for data collection, and secondly fixing the laser
firing position in the sample sweet spots The samples were
irradiated just above the threshold laser power for obtaining
molecular ions and with higher laser power for studying
cluster formation Thus, the irradiation used for producing
a mass spectrum was analyte-dependent with an
acceler-ation voltage of 20 kV Usually 50 spectra were
accumu-lated
All samples were measured in the linear and the reflectron
modes, in both the positive- and the negative-ion mode
The stainless steel polished surface 2 sample-slides were
purchased from Shimadzu Co., Japan (P/N 670-19109-01)
Polished surface slides were used in order to get better
images for morphological analysis with a stereoscopic
microscope (NIKON Optiphot, Tokyo, Japan;
magnifica-tion·400) and with a high-resolution digital microscope
(Keyence VH-6300, Osaka, Japan; magnification·800)
Sample preparation Matrix stocksolutions were made by
dissolving 1 mg of the selected compound in 0.5 mL of 1 : 1
(v/v) methanol/water Analyte solutions were freshly
pre-pared by dissolving the samples (0.05 mg) in chloroform/
methanol, 1 : 1 (v/v) (0.025 mL)
To prepare the analyte-matrix deposits two methods were
used Method A; thin-film layer method (sandwich
method) Typically 0.5 lL of the matrix solution was placed
on the sample probe tip, and the solvent removed by
blowing air at room temperature Subsequently, 0.5 lL of
the analyte solution was placed on the same probe tip
covering the matrix and partially dissolving it, and the
solvent was removed by blowing air Then, two additional
portions (0.5· 2 lL) of the matrix solution were deposited
on the same sample probe tip, producing a partial
disso-lution of the previously deposited thin-film matrix and
analyte layers The matrix to analyte ratio was 3 : 1 (v/v)
and the matrix and analyte solution loading sequence was:
(a) matrix, (b) analyte, (c) matrix and (d) matrix Method B;
mixture method The analyte stocksolution was mixed with
the matrix solution in 1 : 1–1 : 12 (v/v) ratio A 0.5 lL
aliquot of this analyte-matrix solution was deposited onto
the stainless steel probe tip and dried with a stream of forced
room temperature air Then, an additional portion of 0.5 lL was applied to the dried solid layer on the probe, causing it to redissolve partially, and the solvent was removed by blowing air
The resulting solid partially crystalline layers were found to be relatively homogeneous in both cases nor-Harmane and 2,5-dihydroxybenzoic acid as matrices showed signals of higher quality by using Method A Thus, the results shown and discussed in the present article are those obtained using this sample preparation method for each analyte, in the optimum experimental conditions
Spectra were calibrated using external calibration rea-gents: (a) commercial proteins (neurotensin; angiotensin I; bovine insulin) and (b) a-, b- and c-cyclodextrins with nor-harmane as matrix, in positive- and in negative-ion mode TheKRATOS KOMPACTcalibration program was used
Results
Metabolic labeling of GSLs Cultures of P falciparum with parasitemia around 10% (4.5% ring forms, 2.7% trophozoites and 1.6% schizonts) were metabolically labeled with [14C]palmitic acid for 18 h The different stages were purified on a Percoll gradient and extracted with chloroform/methanol (1 : 1, v/v) A control
of uninfected erythrocytes was analysed in parallel (Table 1) The different extracts were further fractionated
by DEAE-Sephadex A-25 (ACO–) column chromatography into neutral and acidic lipids TLC analysis of the unbound fraction showed that the radioactive precursor was mainly incorporated into diacyl-phospholipids (phosphatidylcho-line, phosphatidylethanolamine and their lyso-derivatives)
as reported previously [21] (not shown) The acidic fraction corresponding to the schizont stage showed a significantly high incorporation in comparison with ring and trophozoite stages (Table 1); thus similar amounts of radioactivity of each fraction was applied to the TLC plate Acidic lipids analysed in solvent A showed main spots corresponding
to phosphatidylinositol, phosphatidic acid and fatty acids (Fig 1A)
The unbound fraction of each stage was treated with 0.1MNaOH in methanol to hydrolyse non ceramide-based lipids and after purification, the samples were analysed
by TLC in solvent B (Fig 1B) These lipidic components migrated close to standards of GSLs A spot with the mobility similar to a standard of sphingomyelin was also shown Even though the sample of control erythrocytes used was enhanced, only faint bands were observed In
Table 1 Incorporation of radioactive precursors (CPM per 108parasites) in the different fractions of lipids obtained fromthe three intraerytrocytic stages of P falciparum C, control uninfected erythrocytes; R, rings; T, trophozoites; S, schizonts.
[ 14 C]palmitic acid [ 14 C]glucose
Total lipids Neutral lipids Acidic lipids Sphingolipids Total lipids Neutral lipids Acidic lipids Sphingolipids
S 6269500 5192900 73800 115900 87854 71200 2500 3800
Trang 5order to ensure that the labeled components corresponded
to ceramide-based lipids, the spot comigrating with the
standard of GbOse3Cer (IV) from the schizont fraction
(Fig 1B, lane 4), was eluted from the plate and digested
with ceramide glycanase As expected, hydrolysis was not
complete, however, a new spot comigrating with a standard
of ceramide was obtained (Fig 2A)
A sample of the saponified neutral lipids obtained from schizont stages, was further purified by Unisil column chromatography Three fractions (CHCl3, CHCl3/MeOH (98 : 2, v/v), and CHCl3/MeOH (1 : 3, v/v)) were eluted The latter, containing the glycosphingolipids, was hydro-lysed with HCl/MeOH/water (3 : 29 : 4, v/v/v), treated with BF3/MeOH to methylate the rest of the fatty acids that could interfere, and analysed by TLC in solvent C (Fig 2B) Two spots, migrating in the region where long chain bases are resolved, were detected One of them (RF 0.33) with the mobility of an authentic standard of C18 -sphinganine, the other one (RF 0.38) migrating slightly above, would correspond to C20-sphinganine A similar result was obtained when the spot comigrating with GbOse3Cer was analysed under the same conditions (not shown)
In another experiment a [14C]glucose incorporation was tried The three sugar labeled stages were extracted as above and the extracts were fractionated by DEAE-Sephadex column chromatography and saponified Although recov-eries in ring and trophozoite forms extracts were low, the purified glycosphingolipid fraction obtained from the schizont stage showed a significant higher incorporation
of the radiolabeled sugar than uninfected erythrocytes (Table 1) This extract was analysed by TLC in solvent B
in comparison with an analogous [14C]palmitic acid labeled fraction (Fig 3A) Four spots with RF similar to those obtained by [14C]palmitic acid labeling were detected Cerebroside (RF 0.85) was clearly resolved in the sugar labeled sample (Fig 3A, lane 2) When a similar experiment was performed using [14C]galactose as precursor, a faint band corresponding to galactosylceramide was also detected (Fig 3B)
In order to further analyse each GSL fraction, extracts obtained from schizont stages were fractionated as above and the neutral GSL fraction was subjected to TLC in parallel with an analogous [14C]palmitic acid labeled
Fig 1 Incorporation of [14C]palmitic acid into lipids of Plasmodium falciparum (A) TLC analysis in chloroform/methanol/water (65 : 25 : 3, v/v/v) of the acidic lipids Samples obtained from 4.8 · 10 7 ring forms (lane 1), 2.06 · 10 7 trophozoites (lane 2) and 4.38 · 10 6
schizonts (lane 3) were spotted in order to apply similar amounts of radioacti-vity PtdGr, phosphatidyl glycerol; PtdH, phosphatidic acid; PtdIns, phosphatidylinosi-tol; PtdSer, phosphatidylserine; lysoPtdIns, lysophosphatidylinositol (B) The unbound fractions of the DEAE-Sephadex column were saponified and analysed by TLC Samples corresponding to: 2.4 · 10 8 ring forms (lane 1); 1.0 · 10 8
trophozoites (lane 2); 0.4 · 10 8
schizonts (lane 3); 7.0 · 10 8
non-infected erythrocytes (lane 4) were analysed in chloroform/methanol/0.25% KCl (80 : 30 : 2, v/v/v); I, ceramide; II, glucosylceramide; III, lactosylceramide; IV, globotriaosylceramide;
V, globotetraosylceramide; VI, sphingo-myelin.
Fig 2 Hydrolysis of the GSLs The spot comigrating with GbOse 3 Cer
(IV) from Fig 1B (lane 4) was incubated with ceramide glycanase for
18 h, extracted with choroform/methanol and further subjected to
TLC in chloroform/metanol/0.25% KCl (80 : 30 : 2, v/v/v) Cer,
cer-amide; GbOse 3 Cer, globotriosylceramide The glycosphingolipid
fraction metabolically labeled with [14C]palmitic acid was subjected to
methanolysis, further treated with BF 3 /methanol and analysed by TLC
in chloroform/metanol/1 M NH 4 OH (40 : 10 : 1, v/v/v).C 18 -Sph, C 18
-sphingosine; C -sSph, C - dihydrosphingosine.
Trang 6fraction Spots corresponding to the ceramide fraction (I),
monohexosylceramide fraction (II), trihexosylceramide
fraction (IV) and tetrahexosylceramide fraction (V) were
extracted from the plate and analysed by UV-MALDI-TOF
MS (Fig 4) Table 2 shows the m/z values (mass numbers)
and possible sphingoid-fatty acid-sugar combinations of
ceramides for the signals obtained from each fraction,
taking into account the results obtained by TLC analysis of
the long chain bases
Presence of an active glucosylceramide synthase
In a first approach, the activity of the enzyme in a parasite
lysate was examined using UDP-[14C]glucose as donor and
two different ceramides, palmitoyldihydrosphingosine and
palmitoylsphingosine as acceptors (Fig 5A) A spot
comi-grating with glucosylceramide was obtained but,
interest-ingly, although dihydroceramides are poor substrates for
the mammalian enzymes [22], P falciparum enzyme seemed
to be active only with the saturated compound (Fig 5A,
lane 2) In order to confirm the substrate specificity of the
plasmodial enzyme, the enzymatic assay was performed
using fluorescent ceramides and UDP-glucose as donor
Lysates from each stage were assayed and the products were analysed by TLC As expected, in the three stages, only parasite lysates incubated with BODIPY-DHCer as accep-tor synthesized fluorescent glucosylceramide (Fig 5B, lanes 1–3) No fluorescent product was obtained when the unsaturated ceramide was used (Fig 5B, lanes 4–6) The special substrate specificity assures the parasite origin of the detected enzyme activity
In order to show the presence of the GCS, immunopre-cipitation of parasite lysates from each stage was performed using polyclonal GCS 1.2 antibody [20] The immunopre-cipitates were subjected to SDS/PAGE and electrotrans-ferred to poly(vinylidene difluoride) membranes When the membranes were developed with the GCS 5.1 antibody, a band at a molecular mass of 48 kDa was detected in the three intraerythrocytic stages (Fig 5C, lanes 2–4) Never-theless, the possibility that the 48 kDa band can be due to a cross-reacting parasite protein not related with the GCS cannot be ruled out
In mammalian cells, low concentrations of D,L -threo-PPMP have no effect on sphingomyelin synthase but can inhibit the synthesis of glucosylceramides [23–26] In order
to establish if the plasmodial enzyme activity was affected, the experiment was performed in the presence of 5 lM
PPMP (Fig 5D) TLC analysis revealed that the presence of threo-PPMP efficiently inhibited the synthesis of glucosyl-ceramides (52%) Additionally, the primary effect of PPMP seemed to be specifically on GSLs, because no difference in bulkprotein synthesis was seen when comparing whole [35S]methionine labeled precipitate of identical number of parasites that were left untreated or treated with 5 lM
PPMP (Fig 5D)
Previous reports showed that treatment of parasite cultures with PPMP resulted in a potent inhibition of the intraerythrocytic development of P falciparum [27–30]
In order to determine the effect of the inhibitor in GSLs synthesis, treatment of parasite cultures with threo-PPMP for 24 h was performed followed by incorporation of [14C]glucose or [14C]palmitic acid in the presence of the drug Parasite development was monitored by microscopy
of Giemsa-stained blood smears (Table 3) As expected, treatment with threo-PPMP showed inhibition of the intraerythrocytic development at the ring stage as described
by Haldar et al [26] Each labeled stage from treated and nontreated parasites was purified by Percoll gradient and further fractionated as above to achieve purified GSLs Comparison of the incorporation of [14C]glucose in the same number of treated and nontreated parasites, showed a clear reduction in the ring stage (Fig 6A) As a result of the arrest on development, a low amount of treated trophozoite and schizont stages were obtained, this fact joined to a low incorporation of the sugar precursor precluded further analysis of these stages Fractions corresponding to the same number of [14C]glucose-labeled ring forms were analysed by TLC in solvent A (Fig 6B) While the fraction obtained from nontreated ring forms showed spots corres-ponding to the labeled GSLs, no spots were detected in the fraction obtained from PPMP-treated ring forms
As regards the palmitic acid labeled parasites, the same analysis was carried out In accordance, when the incor-poration of palmitic acid was compared in treated and nontreated parasites, inhibition of the precursor
incorpor-Fig 3 Incorporation of [ 14 C]glucose and [ 14 C]galactose into
glyco-sphingolipids of Plasmodium falciparum (A) TLC analysis in
chloro-form/methanol/0.25% KCl (80 : 30 : 2, v/v/v) of the unbound
fractions after mild alkaline treatment Lane 1, [14C]glucose labeled
control erythrocytes (7 · 10 8
cells); lane 2, [14C]glucose labeled gly-cosphingolipids from schizonts (4 · 10 8 cells); lane 3, [ 14 C]palmitic
acid labeled glycosphingolipids from schizonts (0.4 · 10 8
cells).
I, ceramide; II, glucosylceramide; III, lactosylceramide; IV,
globo-triaosylceramide; V, globotetraosylceramide; VI, sphingomyelin (B)
TLC analysis in chloroform/methanol/0.25% KCl (80 : 30 : 2, v/v/v)
of the glycosphingolipid fraction purified from schizonts (1.4 · 10 8
parasites) after metabolic incorporation of [ 14 C]galactose GalCer,
galactosylceramide.
Trang 7ation in the three stages was also observed (Fig 6C) The lipidic precursor is incorporated more efficiently probably
as a result of the development of the membrane network Consequently, even when a low amount of schizont stages was obtained, the comparison on TLC could be carried out (Fig 6D) Interestingly, treated schizonts showed no gly-cosphingolipid components in contrast with the nontreated samples
Discussion
Glycosphingolipids seem to be a general feature of eukary-otic cells However, the physiological functions of these glycolipids have only been documented in mammalian cells, whereas very little information is available of their roles in other systems [31] In this report we show for the first time the presence of an active glucosylceramide synthase in the intraerythrocytic stages of P falciparum Incorporation of [14C]palmitic acid and [14C]glucose allowed the analysis of purified glycosphingolipids When the long chain base component of these GSLs was investigated, using [14 C]pal-mitic acid as a precursor, labeled sphinganine was obtained (Fig 3) in contrast with the major long chain base present in erythrocytes, indicating clearly the parasite origin of the detected compound Degradation of host sphingomyelin to produce ceramide for parasite growth has been suggested, supported by the existence of sphingomyelinase in P falci-parum[30,32,33] However, although the amount of
cera-Fig 4 UV-MALDI-TOF mass spectra in positive ion mode of the different GSLs frac-tions Values indicate m/z of sodium adducted molecular ions, [M + Na] + , in nominal mass Posible ceramide species are listed in Table 2 (A) UV-MALDI-TOF MS of cera-mides (fraction I) in reflectron mode; matrix: nor-harmane; (B) UV-MALDI-TOF MS of monohexosylceramides (fraction II) in linear mode; matrix: nor-harmane; (C) UV-MALDI-TOF MS of globotriaosylceramides (frac-tion IV) in reflectron mode; matrix: 2,5-di-hydroxybenzoic acid; (D) UV-MALDI-TOF
MS of globotetraosylceramides (fraction V) in reflectron mode; matrix: 2,5-dihydroxybenzoic acid.
Table 2 Mass numbers and possible sphingoid-fatty acid-sugar
combi-nations of ceramides in the different fractions of GSLs obtained from
schizont forms after TLC analysis Molecular related ions, [M+Na]+
are expressed as nominal mass Listed ceramide species were deduced
from UV-MALDI-TOF MS spectra (Fig 4) Spectra are shown in
Fig 6 m/z, Data from UV-MALDI-TOF MS.
Spectra m/z Proposed structures
A 494.9 d18:0-C10h:0
522.9 d20:0-C10h:0 d18:0-C12h:0
537.2 d20:0-C10-2h:0 d18:0-C12-2h:0
550.8 d18:0-C14h:0 d20:0-C12h:0
553.2 d20:0-C10-3h:0 d18:0-C12-3h:0
569.0 d18:0-C14-2h:0 d20:0-C12-2h:0
B 522.5 d20:0-C10h:0 d18:0-C12h:0
550.7 d18:0-C14h:0 d20:0-C12h:0
568.0 d18:0-C14-2h:0 d20:0-C12-2h:0
656.6 d18:0-C10h:0(+hex)
686.3 d20:0-C10h:0 (+hex) d18:0-C12h:0 (+hex)
C 1036.5 d18:0-C14h:0(+3hex) d20:0-C12h:0(+3hex)
1052.3 d18:0-C14-2h:0(+3hex) d20:0-C12-2h:0(+3hex)
1068.3 d18:0-C14-3h:0(+3hex) d20:0-C12-3h:0(+3hex)
D 1038.2 d18:0-C14h:0(+3hex) d20:0-C12h:0(+3hex)
1052.8 d18:0-C14-2h:0(+3hex) d20:0-C12-2h:0(+3hex)
1185.0 d18:0-C10h:0
(3hex+hexNAc)
Trang 8mide produced is low, our results confirm that the de novo
biosynthetic pathway of ceramides is active in this parasite
Nevertheless, the last step of the ceramide biosynthesis,
involving the dehydrogenation of N-acylsphinganine to
N-acylsphingenine would be absent in P falciparum
Addi-tionally, [14C]galactose incorporation showed the presence
of galactosylceramide as reported recently [11]
The sphingolipid structure of the different products obtained in the GSL fraction was proven by UV-MALDI-TOF mass spectrometry The spectra showed that the predominant components of Fraction I were ceramides involving long chain bases d18:0 or d20:0 and hydroxy fatty acids C10:0, C12:0 and C14:0 bearing one, two or three hydroxy residues (Fig 4A, Table 2) This
Fig 5 Glucosylceramide synthase analysis (A) The enzymatic assay was performed using UDP-[14C]glucose as marker in 0.1 M sodium phosphate buffer (pH 7.4), 2 m M b-NAD and a liposomal substrate consisting in dipalmitoylphosphatidylcholine and ceramide (10 : 1) (containing 0.1 nmol
of ceramide) The mixture was purified and analysed by TLC in chloroform/methanol/water (65 : 25 : 2, v/v/v) Lane 1, palmitoylceramide; lane 2, palmitoyldihydroceramide (B) The enzyme assay was performed using the fluorescent ceramide precoupled to BSA, UDP-glucose (2.5 m M ) and
2 m M b-NAD The parasite lysates (50–100 lg protein) in 0.1 M sodium phosphate buffer (pH 7.4) The mixture was extracted with chloroform/ methanol (1 : 1) and analysed by TLC in chloroform/methanol/water (65 : 25 : 2, v/v/v) Lanes 1, 2 and 3 are rings, trophozoites and schizonts, respectively, using BODIPY-dihydroceramide; lanes 4, 5 and 6, the same using BODIPY-ceramide (C) Immunoprecipitation of parasite lysates performed using polyclonal GCS 1.2 antibody The immunoprecipitates were subjected to SDS/PAGE and electrotransferred to poly(vinylidene difluoride) membranes The membranes were developed with the GCS 5.1 antibody followed by ECL 1, control erythrocytes; 2, ring forms; 3, trophozoites; 4, schizonts Molecular mass of markers is indicated (in kDa) at the right side of the figure The arrow at the left side shows the band at
48 kDa (D) The enzymatic assay was performed using schizonts as enzymatic source as in (A), without (lane 1) or with (lane 2) 5 l M PPMP as inhibitor GlcCer, glucosylceramide; R, ring forms; T, trophozoites; S, schizonts (E) Incorporation of L -[ 35 S]methionine in proteins obtained from parasites treated or nontreated with 5 l M PPMP Aliquots were collected at different times (0–48 h), precipitated with 12% (w/v) trichloroacetic acid and radioactivity was measured.
Trang 9finding is in accordance with a previous report showing
that the de novo biosynthetic pathway of fatty acids in
P falciparum involved C10:0 to C14:0, some of them
hydroxylated [34] Fraction II migrated very near Frac-tion I and was shown to be a mixture of ceramides and monohexosyl ceramides, not very well resolved The latter were mainly monohexosylceramides of d18:0 and d20:0 acylated with C10h:0 and C12h:0 (Fig 4B, Table 2) Spectrum C (Fig 4) showed that the predominant com-ponent of Fraction IV is a trihexosylceramide (m/z 1036.5) with a possible sphingoid-fatty acid combination d18:0-C14h:0 (or d20:0-C12h:0) On the other hand, Fraction V (Fig 4, spectrum D) showed less intense signals than the others However, it was very interesting
to detect a component of m/z 1185.0 corresponding to a tetrahexosylceramide bearing an N-acetylhexosamine resi-due This result agrees with the fact that incorporation of tritiated glucosamine led to the preferential detection of GSLs migrating as highly glycosylated species [10] Biosynthesis of GSLs in P falciparum pointed to the presence of an active glucosylceramide transferase When the enzyme activity was searched in parasite lysates using UDP-[14C]glucose as marker as well as using fluorescent ceramides, activity was found only when the dihydrocera-mide was used as substrate This is in good agreement with the result described above and would explain earlier reports showing that the parasites were not competent to the formation of glucosylceramide when using unsaturated ceramides [28]
GCS from different eukaryotic kingdoms have been cloned; remarkably their sequences present only a few conserved amino acids and the overall similarity between the enzymes from species with remote evolutionary rela-tionship is rather low [31] In particular for P falciparum,
we were unable to find any sequences related to GCS This is not rare as it has been suggested that enzymes are more difficult to identify in P falciparum by sequence similarity methods The difficulty has been attributed either to the great evolutionary distance between P falciparum and other well studied organisms or to the high A + T content
of the genome [35] Nevertheless, we detected a potential gene for GCS (GenBankTM, accession number NP_701286) with conserved domains for glycosyltransferases [36] However, in an attempt to detect the presence of the plasmodial enzyme in a parasite lysate, immunoprecipita-tion with polyclonal antibodies against the human GCS was tried A band of molecular mass near 48 kDa was recognized in the three stages of the parasite This band
Fig 6 Inhibition of GSL synthesis by threo-PPMP treatment GSLs
were purified from threo-PPMP treated and nontreated parasites and
were further analysed by TLC in chloroform/methanol/water
(65 : 25 : 3, v/v/v) (A) Comparison of the radioactivity recovered in
the GSL fractions obtained from treated (unfilled bars) and nontreated
(filled bars) [14C]glucose incorporated parasites R, ring forms;
T, trophozoites; S, schizonts; C, control uninfected erythrocytes (B)
Lane 1, control [ 14 C]glucose labeled ring stage (4.2 · 10 8 parasites);
lane 2, threo-PPMP-treated [ 14 C]glucose labeled ring stage (5.5 · 10 8
parasites) (C) Comparison of the radioactivity recovered in the GSL
fractions obtained from treated (unfilled bars) and nontreated (filled
bars) [ 14 C]palmitic acid incorporated parasites R, ring forms;
T, trophozoites; S, schizonts; C, control uninfected erythrocytes (D)
Lane 1, control [14C]palmitic acid labeled schizonts (0.13 · 10 8
para-sites); lane 2, threo-PPMP-treated [ 14 C]palmitic acid labeled schizonts
(0.16 · 10 8
parasites) In all cases, at each stage, a similar number of
parasites was compared.
Table 3 Effect of 5 l M threo-PPMP on parasite development Parasite cultures were incubated with 5 l M DL -threo-PPMP After 24 h of treatment, parasites were labeled with [ 14 C]palmitic acid or [ 14 C]glu-cose for 18 h in the presence of the drug Control cultures without the inhibitor were labeled under the same conditions The effect on para-site development was monitored by microscopy of Giemsa-stained blood smears in two independent experiments R, ring forms;
T, trophozoites; S, schizonts.
[ 14 C]glucose [ 14 C]palmitic acid
Control Treated Control Treated
Trang 10was absent in the control erythrocytes The apparent Mr
resembles the predicted molecular mass of the human and
rat GCS polypeptides although the empirical molecular
mass described is 38 kDa [20]
In mammalian cells, low concentrations (1–5 lM) ofD,L
-threo-PPMP have no effect on sphingomyelin synthase but
can inhibit the synthesis of glucosylceramides In P
falcipa-rum, PPMP has been described as a potent inhibitor of the
intraerythrocytic maturation leading to an arrest of the
parasites at ring stage Rings formed in the presence of
the drug contain no tubular structures On the contrary,
mature trophozoites and schizonts that contain a fully
extended tubular networkwere not affected by the drug
[26,27,29,30] When we tried the action of PPMP in vitro on
the GCS, using UDP-[14C]glucose as marker, the enzyme
activity which resulted was clearly reduced (Fig 5D) In
another experiment, when the inhibitor was added in parasite
cultures, we observed an arrest on parasite development
Parasites collected at the ring stage had been treated with
PPMP at the trophozoite stage ( 40 h before), and resulted
unaffected (Table 3) On the contrary, parasites collected at
the schizont stage that had received the inhibitor at the ring
stage, were not able to evolve and died When the [14
C]glu-cose labeled GSL fraction purified from PPMP treated and
from control parasites collected at the ring stage were
compared by TLC, disappearance of GSLs was shown
(Fig 6B) This fact indicates that although parasites are able
to evolve to the ring stage, no new GSLs are biosynthesized
Using [14C]palmitic acid as precursor, the analysis could
also be performed with the schizont stage Likewise,
para-sites treated with PPMP showed disappearance of GSLs
(Fig 6D) In this case two hypotheses may be postulated:
PPMP is also acting on the glucosyltransferase and although
there is de novo synthesis of ceramides, the glycosylating step
is blocked; or, parasites that overcome treatment are so
stressed that the tubovesicular membrane networkis not able
to import the lipidic precursor Anyway, the possibility of
both events taking place simultaneously must be considered
In conclusion, we have isolated and characterized the
major GSL structures present in the intraerythrocytic forms
of P falciparum by UV-MALDI-TOF mass spectrometry
A glucosylceramide synthase activity with specificity for
saturated ceramides which can be inhibited by low
concen-trations of PPMP was identified for the first time The
inhibitor, used in cultures, arrests parasite development with
a concomitant depletion of GSLs The special feature
presented by the plasmodial GCS, joined to the expanding
number of cellular functions that may be glycosphingolipid
dependent, makes this enzyme a promising target for
antimalarial drug development Studies are underway for
the characterization of the enzyme and its intracellular
location in P falciparum
Acknowledgements
This workwas supported by grants from: CONICET, Universidad de
Buenos Aires and Agencia Nacional de Promocio´n Cientı´fica y
Tecnolo´gica (Pict 06-06545), Argentina FAPESP, CNPq, PRONEX,
Brazil, UNDP/World Bank/WHO (TDR) A S C and R E.-B are
members of Research Council CONICET (Argentina) and C C.,
ANPCyT fellow Mass spectrometry was performed as part of the
Academic Agreement between R E.-B and H N with the facilities of
the High Resolution Liquid Chromatography-integrated Mass Spec-trometer System Laboratory of the United Graduate School of Agricultural Sciences (Ehime University, Japan) and partially suppor-ted by Heiwa Nakajima Foundation.
References
1 Cooke, B.M (2000) Molecular approaches to Malaria: seeking the whole picture Parasitol Today 16, 407–408.
2 Kimura, E.A., Couto, A.S., Peres, V.J., Casal, O.L & Katzin, A.M (1996) N-linked glycoproteins are related to schizogony of the intraerythrocytic stage in Plasmodium falciparum J Biol Chem 271, 14452–14461.
3 Gerold, P., Dieckmann-Schuppert, A & Schwarz, R.T (1994) Glycosylphosphatidylinositols synthesized by asexual erythrocytic stages of the malaria parasite, Plasmodium falciparum Candidate for plasmodial glycosyl-phosphatidylinositol membrane anchor precursors and pathogenicity factors J Biol Chem 269, 2597– 2606.
4 Gerold, P., Schofield, L., Blackman, M.J., Holder, A.A & Schwarz, R.T (1996) Structural analysis of the glycosyl-phos-phatidylinositol membrane anchors of the merozoite surface proteins-1 and -2 of Plasmodium falciparum Mol Biochem Parasitol 75, 131–143.
5 Gowda, D.C., Gupta, P & Davidson, E.A (1997) Glycosyl-phosphatidylinositol anchors represent the major carbohydrate modification in proteins of intraerythrocytic stage of Plasmodium falciparum J Biol Chem 272, 6428–6439.
6 Naik, R.S., Krishnegowda, G & Gowda, D.C (2003) Glucosa-mine inhibits inositol acylation of the glycosylphosphatidylinositol anchors in intraerythrocytic Plasmodium falciparum J Biol Chem 278, 2036–2042.
7 Mitamura, T & Palacpac, N.M.Q (2003) Lipid metabolism in Plasmodium falciparum-infected erythrocytes: possible new targets for malaria chemotherapy Microbes Infect 5, 545–552.
8 Spiegel, S., Forster, D & Kolesnick, R (1996) Signal traduction through lipid second messengers Curr Op Cell Biol 8, 159–167.
9 Hakomori, S., Handa, K., Iwabuchi, K., Yamamura, S & Pri-netti, A (1998) New insights in glycosphingolipid function: gly-cosignaling domain, a cell surface assembly of glycosphingolipids with signal transducer molecules involved in cell adhesion coupled with signaling Glycobiology 8, xi–xix.
10 Gerold, P & Schwarz, R.T (2001) Biosynthesis of glyco-sphingolipids de-novo by the human malaria parasite Plasmodium falciparum Mol Biochem Parasitol 112, 29–37.
11 Marechal, E., Azzouz, N., Santos de Macedo, C., Block, M.A., Feagin, J.E., Schwarz, R.T & Joyard, J (2002) Synthesis of chloroplast galactolipids in Apicomplexan parasites Eukaryotic Cell 1, 653–656.
12 Futerman, A.H & Pagano, R.E (1991) Determination of the intracellular sites and topology of glucosylceramide synthesis in rat liver Biochem J 280, 295–302.
13 Trinchera, M., Fabbri, M & Ghidoni, R (1991) Topography of glycosyltransferases involved in the initial glycosylations of gangliosides J Biol Chem 266, 20907–20912.
14 Paul, P., Kamisaka, Y., Marks, D.L & Pagano, R.E (1996) Purification and characterization of UDP-glucose: Ceramide glucosyltransferase from rat liver Golgi membranes J Biol Chem.
271, 2287–2293.
15 Katzin, A.M., Kimura, E.A.S., Alexandre, C.O.P & Ramos, A.M (1991) Detection of antigens in urine of patients with acute falciparum and vivax malaria infections Am J Trop Med Hyg.
45, 453–462.
16 Schnaar, R.L (1994) Guide to techniques in Glycobiology In Methods in Enzymology 230 (Lennarz, W.J & Hart, G.W., eds),
pp 348–384, Academic Press, New York.