IAI Accepted Manuscript Posted Online 20 July 2015 Infect. Immun. doi:10.1128/IAI.00168-15 Copyright © 2015, American Society for Microbiology. All Rights Reserved.
1 2 3
Expression, Purification and Biological Characterization of Babesia microti Apical Membrane Antigen 1
4 5 6
Prasun Moitra1, Hong Zheng1, Vivek Anantharaman2, Rajdeep Banerjee1, Kazuyo
7
Takeda3, Yukiko Kozakai1, Timothy Lepore4, Peter J. Krause5, L. Aravind2 and Sanjai
8
Kumar1*
9 10 11
Running title: BIOLOGICAL CHARACTERIZATION OF B. MICROTI AMA1
12 13
Laboratory of Emerging Pathogens, Division of Emerging and Transfusion Transmitted Diseases, Office of Blood
14
Research and Review, Center for Biologics Evaluation and Research, Food and Drug Administration, Silver
15
Spring, Maryland1, National Center for Biotechnology Information, National Library of Medicine, National
16
Institutes of Health, Bethesda, Maryland2, Microscopy and Imaging Core Facility, Center for Biologics
17
Evaluation and Research, Food and Drug Administration, Silver Spring, Maryland3, Nantucket Cottage Hospital,
18
Nantucket, Massachusetts4, Yale School of Public Health and Yale School of Medicine, New Haven, Connecticut5
19 20 21
Keywords: Babesia microti, Apical Membrane Antigen 1, RBC Binding, Antigen Polymorphism
22 23
*Corresponding author: Laboratory of Emerging Pathogens, DETTD, OBRR, CBER, Food and Drug
24
Administration, Silver Spring, Maryland 20993. Phone: 240-402-9652; E-mail:
[email protected].
25 1
26 27
The intraerythrocytic apicomplexan Babesia microti, the primary causative agent of human
28
babesiosis, is a major public health concern in the United States and elsewhere. Apicomplexans
29
utilize a multiprotein complex that includes a type I membrane protein called apical membrane
30
antigen 1 (AMA1), to invade host cells. We have isolated the full length B. microti AMA1
31
(BmAMA1) gene and determined its nucleotide sequence, as well as the amino acid sequence of the
32
AMA1 protein. This protein contains an N-terminal signal sequence, an extracellular region, a
33
transmembrane region and a short conserved cytoplasmic tail. It shows the same domain
34
organization as the AMA1 orthologs from piroplasm, coccidian and haemosporidian
35
apicomplexans but differs from all other currently known piroplasmida, including other Babesia
36
and Theileria species, in lacking two conserved cysteines in the highly variable domain III of the
37
extracellular region. Minimal polymorphism was detected in BmAMA1 gene sequences of parasite
38
isolates from six babesiosis patients from Nantucket. Immunofluorescence microscopy studies
39
showed that BmAMA1 is localized on the cell surface and cytoplasm near the apical end of the
40
parasite. Native BmAMA1 from parasite lysate and refolded recombinant BmAMA1 (rBmAMA1)
41
expressed in E. coli reacted with a mouse anti-BmAMA1 antibody using Western blot. In vitro
42
binding studies showed that both native and rBmAMA1 bind to human red blood cells (RBC).
43
This binding is trypsin and chymotrypsin treatment sensitive but neuraminidase independent.
44
Incubation of B. microti parasites in human RBCs with a mouse anti-BmAMA1 antibody inhibited
45
parasite growth by 80% in a 24 hour assay. Based on its antigenically conserved nature, and
46
potential role in RBC invasion, BmAMA1 should be evaluated as a vaccine candidate.
47
2
48
INTRODUCTION
49 50
Babesiosis, is caused by tick borne intra-erythrocytic apicomplexan parasites of the genus
51
Babesia that infect a wide variety of wild and domesticated animals (1). Human cases have been
52
reported throughout the world, including the United States where it is endemic in the Northeast and
53
upper Midwest; Europe, Asia and Australia (2), (3). Babesia usually are transmitted by Ixodes ticks but
54
also may be transmitted by blood transfusion and transplacentally (4-6). Babesia microti is the primary
55
cause of babesiosis with an increasing incidence in many areas of the United States with up to 4 to 20-
56
fold in the last decade. To address this growing public health threat, the Centers for Disease Control and
57
Prevention declared babesiosis a nationally notifiable disease in 2011 and consequently, expanded
58
surveillance from 18 states in 2011 to 33 states in 2013(7). B. microti infections in young and healthy
59
adults generally cause a mild viral-like infection but may be asymptomatic. More severe disease occurs
60
primarily in neonates, the elderly, and those who are immunocompromised with mortality rates as high
61
as 20% (8, 9).
62
One of the most salient features of Babesia and other Apicomplexan parasites (e.g.,
63
Plasmodium, Babesia and Toxoplasma) is a complex of specialized secretory organelles that include
64
micronemes, rhoptries and dense granules, which secrete molecules necessary for host cell invasion in a
65
coordinated manner (10). One such molecule is the conserved apical membrane antigen 1 (AMA1), a
66
type I membrane protein (11). It is localized in micronemes and is subsequently expressed in a
67
processed form circumferentially around merozoites (12) where it forms a multi-protein complex with
68
other proteins in the moving junction (13). The precise biological role of AMA1 is not known, but
69
evidence based on its sub-cellular localization and data from genetic knock out and trans-species
70
expression studies suggest that it facilitates invasion of RBCs and further growth of intraerythrocytic
71
Plasmodium parasites (14, 15). Crystallization of complete and truncated forms of the extracellular
3
72
region from P. vivax, P. falciparum and T. gondii (16-18) revealed that it contains three domains.
73
Domains I and II are homologous to the PAN domains (plasminogen, apple and nematode), which
74
facilitate protein-protein and protein–carbohydrate interactions among a class of adhesion molecules
75
(19). AMA1 is a major malaria vaccine candidate and its efficacy against asexual stage parasites is being
76
evaluated in clinical studies (20).
77
Despite its growing public health importance, very limited efforts have been made to understand
78
the process of invasion of RBCs by B. microti parasites and the molecules that are associated with this
79
process. Recently X-ray crystallography data characterizing AMA1 from Babesia divergens and
80
Neospora canium was published (21). Here, we report on the gene cloning, recombinant expression,
81
genetic and biological characterization and natural polymorphism in the AMA1 of B. microti
82
(BmAMA1).
83 84
4
85 86
MATERIALS AND METHODS
87 88
B. microti propagation in mice. Babesia microti (Franca) Reichenow Peabody strain (22) was
89
obtained
from the American Type Culture Collection (Manassas, VA). The Peabody strain was
90
originally isolated in 1973 from a Nantucket woman and since has been adapted for growth in hamsters
91
and mice. B. microti was injected into DBA/2NCr mice and parasites were isolated when 10-20% of
92
RBCs were infected, as determined by Giemsa-stained thin-blood films. Mice were maintained at the
93
Center for Biologics Evaluation and Research (CBER) animal care facility and studies were conducted
94
under an Animal Study Protocol approved by the CBER Animal Care and Use Committee.
95 96
B. microti from human patients. Human B. microti samples were obtained from six babesiosis
97
residents of Nantucket in 2009. They were diagnosed with B. microti infection, based on typical
98
symptoms and identification of Babesia on thin blood smears and/or amplification of B. microti DNA
99
using PCR. B. microti infected blood samples were stored at -80ºC until used.
100 101
RNA, Genomic DNA and cDNA isolation. Total RNA was prepared by lysing B. microti
102
infected RBC with Trizol reagent (Life Technologies, Grand Island, NY), followed by chloroform
103
extraction and precipitation with iso-propyl alcohol and ethanol. Genomic DNA (gDNA) was prepared
104
from B. microti-infected human blood and B. microti infected mouse blood using the QIAamp DNA
105
blood mini kit (Qiagen, Valencia, CA). cDNA was prepared from the RNA isolated from B. microti
106
parasites grown in mice with the use of SuperScript Kit (Life Technologies) following the instructions
107
provided by the company.
108
5
109
Gene cloning and nucleotide sequencing of BmAMA1. At the time when this research was
110
conducted, the genome sequence of B. microti was not published. To isolate the full length BmAMA1
111
gene, the following approach was used to design the degenerate sequencing primers. Nucleotide
112
sequences of AMA1 of B. bovis, B. gibsoni, B. bigemina, and B. divergens (GenBank accession nos.
113
AY486101, DQ368061, GQ257738, and EU486539 respectively) were aligned and several sets of
114
degenerate forward and reverse primers were synthesized and used to amplify partial sequences of
115
AMA1 gene from cDNA and gDNA. A set of degenerate primers F1 5' GGW AAR TGC CCA GTT 3'
116
and R1 5' CTC SAR TGG WGA ACC 3', (corresponding to the amino acid sequence GKCPV and
117
GSPLE), gave a band with an amplicon size of 921bp and was cloned into pCR2.1TOPO, sequenced
118
and aligned to the published Babesia AMA1 amino acid sequences. It was found to be ~ 39.7% similar
119
with B. bovis AMA1, 34.8% with B.divergens AMA1, 37.4% with B.gibsoni AMA1 and ~29% with P.
120
falciparum AMA1. Several forward and reverse gene specific primers were synthesized to clone the full
121
length cDNA. Briefly, RACE ready c-DNA was obtained from total parasite RNA following the
122
instructions provided in the GeneRacer Kit (Life Technologies). 5' rapid amplification of c-DNA ends
123
(RACE) reaction were performed with the 5' GeneRacer primers and 3' gene specific primers SK687 and
124
SK686 (parent and nested oligos). 3'RACE was performed with 5' gene specific primers SK670 and SK
125
671 and 3' GeneRacer primers (parent and nested oligonucleotides) (Table 1). Finally, nucleotide
126
sequences obtained from 5' and 3' RACE products were overlaid to obtain the full transcript sequence
127
and then full length gene was amplified using cDNA as a template.
128 129
Recombinant expression, purification and refolding of BmAMA1. The gene fragment
130
encoding for the BmAMA1 extracellular region corresponding to amino acids A41-G529 was PCR-
131
amplified by using the primer sets listed in Table 1. The amplified gene fragment was cloned into
132
pCR2.1TOPO vector (Invitrogen), the nucleotide sequence was determined and plasmid was maintained
6
133
in TOP10 E. coli competent cells. To construct the recombinant expression plasmid, the extracellular
134
region of the gene was cloned into pET101/D-TOPO (Champion™ pET101 Directional TOPO®
135
Expression Kit, Invitrogen) which has a 6x His tag to facilitate purification, and then transformed into
136
BL21 Star™ (DE3) One Shot® Chemically Competent E. coli cells. For recombinant expression, 500
137
ml LB medium was inoculated with 5 ml overnight culture, grown at 37ºC to an A600 0.5 and then
138
induced with 0.5mM IPTG. After 4 h, cells were pelleted and stored at -80ºC until use.
139
The procedure for on-column refolding and purification was performed essentially as described by Saini
140
et al. (23). Briefly, the cell pellet was resuspended in sonication buffer (20 mM Tris - HC1, pH 8.0, 300
141
mM NaC1, 0.5mMDTT, 1X protease inhibitors –EDTA free [Roche] and lysed on ice with a sonicator.
142
The inclusion bodies were resuspended in solubilization buffer (20 mM Tris-Hcl, pH 8.0, 300mM NaCl,
143
6M GdnHcl) and incubated at room temperature (RT) for 2 h with shaking, followed by centrifugation at
144
15,000g for 30 min. The supernatant was loaded onto a pre- equilibrated Ni2+–NTA agarose column and
145
incubated for 3 h at room temperature. The column was washed with 3 bed volumes of solubilization
146
buffer and then washed with decreasing concentrations of GdnHcl (6 to 0 M) in refolding buffer (20 mM
147
Tris-HC1, pH 8.0, 300 mM NaC1, 10% glycerol, 20 mM imidazole, 0.5 mM oxidized glutathione and 5
148
mM reduced glutathione). The bound protein was eluted with refolding buffer containing 200 mM
149
imidazole in fractions of 1 ml, fractions checked by SDS-PAGE, and the peak fractions pooled and
150
dialyzed against PBS and again separated by gel-filtration chromatography using a Superdex 75
151
(10/300) column (GE Healthcare, Piscataway, NJ). Fractions containing recombinant BmAMA1 were
152
pooled, quantitated, aliquoted and stored in -80ºC. NH2-terminal amino acid sequencing was performed
153
by the Edman degradation method to confirm the identity of rBmAMA1. The purified recombinant
154
protein has been named rBmAMA1.
155
7
156
BmAMA1 DNA plasmid vaccine. The DNA vaccine plasmid VR1020 has a strong
157
cytomegalovirus promoter and a tissue plasminogen activator (TPA) signal sequence, which promotes
158
the secretion of the cloned gene. The gene fragment encoding for BmAMA1 extracellular region (G148-
159
E454) was PCR-amplified using the primer sets SK682 and SK683 (see Table 1) and cloned into the
160
BamH1 restriction site of the mammalian expression plasmid VR1020 (24). Recombinant plasmid
161
containing the BmAMA1 extracellular region gene was identified by restriction mapping and the
162
nucleotide sequence was determined. The DNA plasmid for mouse immunizations was purified by
163
Mega Plasmid preparation kit (Qiagen).
164 165
Nucleotide sequencing of BmAMA1 from clinical and field isolates. Blood samples were
166
obtained from six individuals infected with B. microti from Nantucket, MA. B. microti gDNA was
167
extracted and the BmAMA1 extracellular region gene fragments (nucleotides 442-1362 corresponding
168
to amino acids G148-E454) were PCR-amplified from each of the human six samples and from the B.
169
microti Peabody-strain
170
BmAMA1 extracellular region gene fragments were then cloned into TOPO vector (Invitrogen) and
171
their nucleotide sequences determined.
parasites grown in mice using primers SK670F and SK675R (Table 1). The
172 173
Sequence analysis. Iterative sequence searches were performed using the PSI-BLAST programs
174
against the NCBI non-redundant (NR) protein database (25). Multiple sequence alignments were built
175
using Kalign2 or Muscle, followed by manual adjustments based on profile-profile comparisons,
176
secondary structure, and structural alignments. Similarity-based clustering for both classification and
177
culling of nearly identical sequences was performed using the BLASTCLUST program
178
(ftp://ftp.ncbi.nih.gov/blast/documents/blastclust.html). Structural visualization and manipulations were
179
performed using the PyMol program (http://www.pymol.org). Signal peptide and transmembrane helices
8
180
were predicted using SignalP 3.0 and TMHMM v1.0 programs (CBSA, Technical University of
181
Denmark, Denmark), respectively. From multiple alignments of each set of AMA1 sequences, the
182
Shannon entropy (H) for each position was calculated using a custom script with the equation:
183
H = − Pi log 2 Pi
M
i =1
184
where Pi is the fraction of residues of amino acid type i and M is the number of amino acid types.
185 186
Generation of anti-BmAMA1 antibody. Six- to 8- weeks old female Swiss out-bred mice were
187
purchased from Jackson Laboratories (Bar Harbor, Maine) and were housed, fed and used in the
188
experiments in accordance with the guidelines set forth in the National Institutes of Health manual
189
‘Guide for the Care and Use of Laboratory Animals’. Mice were immunized (4 per group) with
190
BmAMA1 extracellular region DNA plasmid by three intradermal injections using 50 µg of DNA
191
plasmid delivered at the base of the tail at 4 week intervals. Control mice were immunized with a
192
recombinant plasmid containing a gene insert encoding for B. microti cysteine protease protein or mice
193
that did not receive any immunization. Non-heparinized blood was collected from the tail vein of all the
194
experimental and control mice at the time of each immunization and 10 days after the last booster.
195
Serum samples were collected and stored at -20°C until use.
196 197
ECL-Western blot. Enhanced Chemiluminescence-Western blot analysis was performed as
198
described earlier (26). Briefly, purified rBmAMA1 was subjected to a NuPAGE® 4-12% Bis-Tris SDS-
199
PAGE (Invitrogen) and transferred to a PVDF membrane. Membranes were blocked in 2% I-Block (Life
200
Technologies) and incubated with a mouse anti-Bm-AMA1 antibody (1:2000 dilution) for 1h. Following
201
washings, the membrane was incubated with the ECL-goat AP-conjugated anti-mouse IgG (1:5000
202
dilution) for 1h. Following additional washings, the membrane was reacted with ECL substrate solution
9
203
(Life Technologies, Western-Star™ Immunodetection System, Grand Island, NY), exposed to KODAK
204
X-OMAT AR (XAR) autoradiography film, (VWR Scientific, Radnor, PA) and developed using a
205
Kodak developer (X-OMAT 1000A).
206 207
Immunofluorescence Antibody Assay (IFA). B. microti infected mouse RBC were washed and the
208
RBC membrane was stained by a PKH26 red fluorescent cell membrane labeling kit (Sigma-aldrich, MO,
209
USA) according to the manufacturer’s protocol, fixed in 3% paraformaldehyde plus 10mM piperazine-N,N′-
210
bis (2-ethanesulfonic acid) buffer (PIPES) pH 6.4, in PBS for 30 minutes at RT, smeared on poly-L-Lysine
211
coated 22mm2 coverslips, permeabilized with 0.25% Triton X-100 in PBS, washed three times in PBS, and
212
blocked in 5% bovine serum albumin (BSA)/PBS for 30 min (27). Cover slips were incubated in immune
213
mouse serum diluted 1:20 in BSA/PBS for 1h, washed three times in PBS and incubated in goat anti-mouse
214
Alexa Fluor 488 (Invitrogen) for 1 h. Cover slips were washed three times with PBS, briefly swirled in
215
distilled water and mounted in antifade reagent Vectashield containing 6-diamidino-2-phenylindole (DAPI)
216
over a glass slide and sealed with nail polish. The samples were imaged with Leica TCS_SP8 DMI6000
217
confocal microscope system (Leica Microsystems, Mannheim, Germany) with a 100x 1.4 NA oil objective
218
lens. Images were acquired at 1024 by 1024 pixel resolution and stored in lif format for further analysis.
219
Huygens (Scientific Volume Imaging, Hilversum, Netherlands) and Imaris (Bitplane AG, Zurich,
220
Switzerland) software were used for deconvolution and image analysis (28).
221 222
Enzymatic treatment and RBC ghost preparation. Whole blood from random volunteers (O+ ve)
223
was obtained from the NIH Blood Bank. After leukoreduction, RBC were washed three times with PBS and
224
treated with 0.01 U/ml neuraminidase (Sigma), 1mg/ml trypsin (Sigma), and 1 mg/ml concentration of
225
chymotrypsin (Sigma) for 1 hr. After three washes, trypsin treated RBCs were further subjected to treatment
226
with 0.5 mg/ml soybean trypsin inhibitor for 30 min at RT (29). RBC were washed three times with PBS and
10
227
RBC membranes (white ghosts) were prepared by lysing with cold 5mM NaP, pH 8.0 (30, 31) for 30 min on
228
ice, centrifuged at 15000 x g and washed three times to obtain white ghosts. These white ghosts were used for
229
binding assays.
230 231
RBC-binding assay. The ability of BmAMA1 to bind to human red blood cells (RBC) was
232
determined with a RBC-binding assay. Both native (parasite) and recombinant (E. coli) expressed BmAMA1
233
were used in the assay.
234 235
RBC binding using native Bm AMA1. B. microti parasites for use in the RBC binding assay were
236
prepared from infected RBCs collected from DBA/2 mice - when parasitemia was 25%-30%. The RBCs were
237
washed three times in PBS, subjected to lysis using a protocol described by Martin et al. (32) and then washed
238
three times with cold PBS. The binding assay was performed as described by Yokoyama et al. (33). Briefly,
239
the pellets containing approximately 109 parasites were suspended in 1 ml of a lysis buffer (50 mM Tris-HCl
240
[pH 7.5], 0.1% Triton X-100, 150 mM NaCl, 20% glycerol, 0.5 mM EDTA, 1 mM phenylmethylsulfonyl
241
fluoride [PMSF], 1 mM dithiothreitol [DTT], and 1x protease inhibitor (Roche), incubated on ice for 30 min,
242
and then centrifuged at 18,000 g for 30 min at 4°C. These clarified lysates were dialyzed overnight against
243
dialysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 20% glycerol, 0.5 mM EDTA, 1 mM PMSF, 1 mM
244
DTT) with three changes. The dialysates were centrifuged at 18,000g for 30 min at 4°C again, and the
245
supernatants were stored at -80°C until use. The cell extracts were equally divided and the aliquots were
246
incubated with 20 µl of human RBC ghosts for 6 h at 4°C. RBC ghosts were precipitated by centrifugation at
247
10,000 g for 10 min at 4°C, washed in 1 ml of a dialysis buffer three times, suspended in 50 µl of an 2x SDS
248
sample buffer, and heated at 100°C for 10 min. After centrifugation at 13,000 g for 5 min at room
249
temperature, the supernatants were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) with 4-12%
250
polyacrylamide gel. Western blot analysis was carried out with the anti-AMA1 antibody.
251 11
252
RBC binding using rBmAMA1. RBC binding assay using the rBmAMA1 was performed by
253
modification of a procedures described earlier (34). RBC ghosts were incubated with 25 μg of
254
recombinant BmAMA1 for 2 h at RT, and washed extensively in PBS to remove the unbound protein.
255
The pellet was resuspended in 4X SDS sample loading buffer, boiled for 5 min at 100◦C, and subjected
256
to SDS –PAGE and ECL-Western blot analysis to determine binding affinity of untreated and
257
enzymatically treated RBC membranes with the native or recombinant BmAMA1. To obtain a semi-
258
quantitative estimation of bound native and rBmAMA1, the band intensity of individual bands were
259
measured as pixel units in a defined area by using the ImageJ program (26) and expressed as integrated
260
optical density (IOD). The relative percentage binding of native or rBmAMA1 to RBC was determined
261
as follows: IOD of enzymatically treated RBC/IOD of untreated RBC x 100.
262
To demonstrate successful removal of sialic acids and to confirm effective enzymatic digestion, RBCs
263
were treated with neuraminidase and RBC ghosts were prepared as described earlier. An independent assay
264
was conducted where untreated and neuraminidase treated RBC ghosts were stained with The Thermo
265
Scientific Pierce Glycoprotein Staining Kit using the periodic acid-Schiff (PAS) method (supplementary
266
material). For RBC binding assay, untreated and neuraminidase treated membranes were incubated with
267
parasite extracts , subjected to SDS-PAGE and western blot was carried out with anti-AMA1 antibody
268
(supplementary material).
269 270 271
Growth inhibition assay. This assay was performed using a procedure described earlier by Sun
272
et al. (35) with slight modification. B. microti infected RBCs from DBA/2 mice was collected when
273
parasitemia reached 25%-30%, washed three times in PBS, and subjected to lysis using a protocol
274
described by Martin et al. (32). Isolated B. microti parasites (merozoite and intracellular forms) were
275
washed and resuspended in RPMI1640 with 10% FCS. The number of B. microti parasites/µl of solution
12
276
was determined by Count Bright Absolute Counting Beads according to the manufacturer’s protocol
277
(Invitrogen). The total RBC count/µl of the suspension was determined by using a hemocytometer.
278
Parasites suspended in RPMI1640 were incubated with pooled heat-inactivated mouse anti-BmAMA1
279
sera or pre-immune sera at a final dilution of 1:15 for 1 h at room temperature. The parasites were then
280
added to normal human RBCs (2% hematocrit) equivalent to a final parasitemia of 1% in 96 well tissue
281
culture plates in triplicate. After 8 and 24 h of incubation at 37ºC, RBCs were collected, centrifuged at
282
300 g for 5 min, and washed three times with PBS to remove the unbound parasites. Giemsa-stained thin
283
blood-films from each well were prepared and the percentage of B. microti infected RBCs was
284
determined by counting approximately 1000 cells for each sample using oil immersion light microscopy.
285
The experiment was repeated three times. The percentage of inhibition of merozoite invasion was
286
determined using the formula: [1 - % parasitemia with test sera / % control (pre-immune sera)
287
parasitemia] x 100%.
288 289
Statistical methods. A comparison of the relative percent inhibition of growth of B. microti
290
parasites in human RBCs was made between anti-BmAMA1 sera and pre-immune sera after 8 h and 24
291
h cultivation.
292
percent of growth between the two test sera. All data were evaluated using two-sided hypothesis test. A
293
p value of < 0.05 indicated statistical significance. All statistical analyses were implemented by “R”
294
statistical software.
Mann-Whitney nonparametric test was used to evaluate the difference in the relative
295 296
Nucleotide sequence accession number. The full length BmAMA1 cDNA sequence described
297
in this paper has been deposited in the GenBank database under accession number JX488467 (GI
298
number 490386734).
299
13
300 301
RESULTS
302 303
BmAMA1 gene cloning and nucleotide sequence. Sets of forward and reverse degenerate
304
primers were initially designed from consensus sequences of aligned amino acids of AMA1 of four
305
Babesia species (B. bovis, B.gibsoni, B.bigemina, and B. divergens) that corresponded to conserved
306
GKCPV and GSPLE sequences. A PCR product of 921 bp encoding for 307 amino acids (aa) was
307
obtained from gDNA of B. microti that was cloned into TOPO 2.1 vector and the nucleotide sequence
308
was determined. To obtain the full length BmAMA1, gene-specific forward and reverse primer sets
309
were synthesized from this sequence and primers from GeneRacer kit (Invitrogen) were used to derive
310
the N- and C- terminal sequences from a B. microti cDNA. This approach allowed assembly of the
311
complete BmAMA1 cDNA that corresponded to approximately 2100 bp, which was sub cloned into the
312
TOPO TA vector and the nucleotide sequence determined. The deduced BmAMA1 sequence was
313
aligned with four published Babesia AMA1sequences mentioned above and was used to establish the
314
start of translation and the translated open reading frame (ORF). The analysis of nucleotide sequence of
315
BmAMA1 cDNA revealed an ORF of 1845 bp preceded by a 172 bp 5' non coding region preceding the
316
first methionine codon. A 3' untranslated region of 102 bp contained a 25 bp long poly (A) tail. A small
317
intron of 23 bp located after 134 nucleotides from the initial ATG codon was identified in the gDNA
318
sequence.
319
We next sought the copy number of BmAMA1 in the parasite genome. To accomplish this, B.
320
microti gDNA was digested with restriction enzymes and then subjected to Southern blot analysis and
321
hybridized under high stringent condition with the BmAMA1 extracellular region probe of 921bp.
322
Analysis of the pattern of hybridization revealed that BmAMA1 is present as a single copy gene (data
14
323
not shown). This result is further corroborated by the recently published B. microti genome sequence
324
(36) that suggests a single copy for BmAMA1 gene on chromosome 3.
325 326
Analysis of BmAMA1 and comparison with the other apicomplexan AMA1 proteins. The
327
ORF encoded the BmAMA1 protein of 614 amino acids with a predicted molecular weight of 69010.02
328
Daltons and an isoelectric point of 6.245. The SignalP 3.0 predicted a 35aa signal sequence of
329
BmAMA1 with a putative cleavage site at position 38 and 39: VSA-AL. The TMHMM version 1.0
330
program predicts an extracellular region from A41- G529, a hydrophobic C-terminal transmembrane helix
331
from F530-L552 and a small cytoplasmic tail from R553-Y614, characteristic of conserved type I integral
332
membrane proteins, which are present in all apicomplexan AMA1 proteins (Fig. 1A). The sequence of
333
BmAMA1 derived from the B. microti genome sequence has been erroneously predicted (Genbank gi:
334
399217547; BBM_III01160). As a consequence, the N-terminal-most parts, including the signal peptide,
335
have been fused to the adjacent gene. This work presents for the first time a corrected version of the
336
BmAMA1 sequence. Examination of the multiple alignments of the sequences of AMA1 reveals that the
337
basic pattern of three domains in the extracellular region (domains I-III) is preserved across all
338
apicomplexans (Fig.1B). The first two of these are PAN domains, which are observed in bacterial and
339
eukaryotic extracellular proteins. Within the two PAN domains there are 10 cysteines, which are almost
340
universally conserved across all species. The C-terminal domain (III) has a core sheet formed by 4
341
strands, which contains 4 conserved cysteines largely preserved across all species. However, B. microti
342
is unique (differing even from other piroplasmida) in that it lacks two of the cysteines forming a
343
disulfide bond in the C-terminal domain. Thus, the extracellular region of B. microti contains 6 disulfide
344
bridges as opposed to 7 in other piroplasmida. The domain III from most coccidian and haemosporidian
345
versions contains an additional cysteine dyad suggesting that these AMA1s form a total of 8 disulfide
346
bonds.
15
347
BLAST searches with the BmAMA1 helped identify its orthologs from other piroplasmida
348
(Theileria and other Babesia species), haemosporidian (Plasmodium species), and coccidian
349
(Toxoplasma, Eimeria and Neospora) apicomplexans. However, no AMA1 orthologs were detected in
350
more basal apicomplexan lineages such as Cryptosporidium. A phylogenetic tree with the complete
351
sequences of AMA1 from the organisms listed above showed that the coccidian versions form the most
352
basal clade with the two vertebrate blood-infecting clades (piroplasmida and haemosporidia) grouping
353
together (Fig. 2). Interestingly, while all coccidians have at least 3 paralogs of AMA1 in their genome,
354
both the blood-infecting clades have only a single ortholog. Within the piroplasma, the version from B.
355
microti groups outside of the clade unites other Babesia species (Babesia bigemina, Babesia divergens,
356
Babesia gibsoni and Babesia ovata) and the Theileria group (Theileria parva, Theileria annulata,
357
Theileria orientalis and Theileria/Babesia equi). These findings are consistent with previous
358
phylogenetic analyses which suggest that B. microti is a distinct lineage of piroplasmida outside of the
359
classical Theileria and Babesia genera (37).
360 361
Minimal polymorphism in BmAMA1 gene in B. microti human isolates and comparison to
362
other apicomplexans. We wanted to know the extent of natural polymorphism in BmAMA1. To
363
accomplish this, we isolated the gene fragments for the extracellular region (G148-E454) of BmAMA1
364
from six B. microti infected individuals from Nantucket in 2009, determined their nucleotide sequences
365
and then compared them to the laboratory adapted B. microti Peabody strain that was isolated from the
366
Nantucket in 1973 for nucleotide and amino acid polymorphisms. There was 99.9% homology at the
367
amino acid level and 99.98 % at the nucleotide level between these six BmAMA1 sequences. The
368
extracellular region BmAMA1 sequences from the three Nantucket clinical isolates were identical to the
369
laboratory adapted B. microti Peabody strain. Each of the other three clinical B. microti isolates from
370
Nantucket
revealed a single amino acid substitution at the following positions: Aspartate368 to
16
371
Glycine368, Lysine350 to Glutamate350 and Asparagine416 to Isoleucine416. These results demonstrate that
372
BmAMA1 is highly conserved in B. microti parasites transmitted on Nantucket.
373
Despite the lack of variability in our current sample set we sought to better understand the
374
general tendencies in variability in piroplasma AMA1 with respect to other apicomplexans which
375
possess orthologs of this protein. Given the rapid evolution of the AMA1 genes across apicomplexans
376
with multiple amino acid substitutions at aligned sites, they are not amenable to a conventional positive
377
selection analysis using Dn/Ds ratios. Hence, we used a parallel approach based on Shannon entropy
378
plots to study the per-site amino acid variation in the fungal TET/JBP proteins(38). In this analysis we
379
compared the variability within the piroplasma, which include B. microti as the out group to the clade
380
uniting the Theilerias and other Babesias. This allowed us to analyze local regions of high variability
381
within piroplasma (i.e. regions with high per-site entropy). In general we observed that in the
382
extracellular region, specific domain III showed the greatest variability throughout most of its length
383
barring the conserved cysteines. In both domains I and II (the PAN domains) the maximum variability
384
was concentrated in certain exposed loops, which are regions that are likely to be exposed to immune
385
attack. We next compared these to the similar variability plots in Plasmodium and across all
386
apicomplexa with AMA1. Regions of high variability in piroplasma (at least in the two PAN domains)
387
showed some correlation to the variability across Apicomplexa (Pearson correlation=0.55-0.67). Thus, it
388
appears that across Apicomplexa some of the equivalent exposed regions are attacked by the host
389
immune response and thereby prone to rapid evolution. However, the lower correlation between the
390
genus Plasmodium and piroplasma (Pearson correlation=0.25) suggests that there might also be lineage-
391
specific pressures that shape the variability in the extracellular region.
392 393
Antigenic characterization of native and rBmAMA1. Native BmAMA1 was identified in the
394
B. microti parasite lysate based on reactivity with mouse anti-BmAMA1 antibody generated by
17
395
immunization with BmAMA1 DNA plasmid vaccine. Native protein migrated as a band at ~78 kDa
396
(p78) and, a possible proteolytic cleaved fragment of 37 kDa (p37) was observed in ECL-Western Blot
397
(Fig. 3A). The predicted molecular weight of BmAMA1 is 69 kDa (p69). Thus, while p78 may represent
398
the full length AMA1 expressed by B. microti, the nature of p37 remains unclear.
399
The extracellular BmAMA1 encoding the amino acids A41-G529 was cloned in pET101/D-TOPO
400
with a C-term His tag and viral V5 epitope and expressed in E.coli. The recombinant protein was present
401
in the inclusion bodies, which were solubilized in 6M GdnHcl, refolded on Ni2+–NTA agarose column,
402
eluted with 200mM imidazole and again purified on Superdex 75 10/300 size-exclusion column. Identity
403
of rBmAMA1 was determined by electrophoretic migration in PAGE, ECL-Western Blot and NH2-
404
terminal amino acid sequencing. Though the predicted molecular weight of His6-tagged extracellular
405
region is 60 kDa (p60), under the non-reducing conditions, it migrated at a molecular weight of 67 kDa
406
(p67) in 4-12% SDS –PAGE (Fig. 3B, Lane 1) that exhibited slower migration under reduced conditions
407
(Lane 2). NH2-terminal amino acid sequencing confirmed the first twelve amino acids of p67 BmAMA1
408
(data not shown). A second band at approximately 54 kDa also was observed (Fig. 3B) which was a
409
proteolytic fragment of the full length BmAMA1 as confirmed by NH2-terminal amino acid sequencing
410
(cleaved at position S205). ECL-Western Blot revealed that both rBmAMA1 protein bands were reactive
411
with the mouse anti-BmAMA1 antibody (Fig. 3C).
412 413
Localization
of
BmAMA1
in
intraerythrocytic
B.
microti
merozoites
by
414
immunofluorescence. The localization of BmAMA1 in intraerythrocytic B. microti parasites was
415
determined by immunofluorescence analysis using a mouse anti-BmAMA1 antibody. Following staining
416
of RBC membrane with PKH26 and incubation with the anti-BmAMA1 antibody, intraerythrocytic
417
parasites were counter-stained with an anti-mouse IgG-conjugated to Alexa-488 and the nucleus was
418
stained with DAPI. The presence of the DAPI-stained parasite nucleus is shown in blue while a specific
18
419
reactivity of anti-BmAMA1 antibody shown in green was detected on B. microti merozoites (Fig. 4). We
420
observed that about 26% of B. microti infected RBC had a fluorescence pattern that was consistent with
421
reactivity with anti-BmAMA1 antibody. BmAMA1 appeared to be localized peripherally with a
422
concentration near the apical end (based on the apparent shape of the parasite) along with the regions of
423
condensed peripheral cytoplasmic localization (Fig. 4, merged color image). Parasites labeled with pre-
424
immune serum lacked any specific fluorescent signal.
425 426
BmAMA1 binds to human RBC and the RBC receptor for BmAMA1 is sensitive to treatment with
427
trypsin and chymotrypsin. We wanted to determine whether BmAMA1 binds to human RBCs and plays a
428
role in B. microti invasion of RBCs. To accomplish this, we used a standardized RBC binding assay where
429
native (parasite lysate) and rBmAMA1 were incubated with untreated and enzymatically treated RBC
430
membranes. We also used anti-BmAMA1 antibody to detect the bound AMA1 in a semi-quantitative ECL-
431
Western Blot. The band density representing the rBmAMA1 bound to RBC in ECL-Western Blot was
432
measured as integrated optical density (IOD). RBCs were treated with neuraminidase, trypsin and
433
chymotrypsin that selectively cleave different moieties of membrane proteins.
434
PAS staining of neuraminidase treated ghost RBC did not show staining of glycoproteins
435
whereas the glycoproteins of untreated ghost RBC were effectively stained with PAS (Fig. S1). These
436
results suggested that neuraminidase treatment did effectively remove the sialic acid residues on ghost
437
RBC.
438
The RBC binding assay with parasite lysate revealed that native BmAMA1 binds to untreated
439
and neuraminidase treated human RBC ghost membranes (Fig. 5A, Lanes 1 and 2), but not following
440
trypsin or chymotrypsin treatment (Lanes 4 and 5). RBC membranes treated with only trypsin inhibitor
441
had no effect on the binding (Lane 3). No bands were observed when untreated membranes were
442
incubated with lysate from non-infected mouse RBC (Fig. 5A Lane 6). IOD measurements of the
19
443
protein bands showed that the highest BmAMA1 binding occurred in untreated RBCs (IOD 214.6 K
444
where K is 1000; 100% binding) neuraminidase treated RBCs (IOD 180.8 K; 15.8% reduction) and
445
Trypsin Inhibitor treated RBCs (208.6 K; 2.8% reduction), suggesting that this binding is neuraminidase
446
independent (Fig. 5A). Because neuraminidase primarily removes sialic acid residues, this data suggests
447
that BmAMA1 binding is not mediated through α2–3-linked sialic acid residue containing proteins on
448
the surface of RBC. On the other hand, RBC treatment with trypsin (IOD 25.7 K; 88.1% reduction) and
449
chymotrypsin (IOD 5.0 K; 97.7% reduction) almost completely abolished the binding (Fig. 5A),
450
suggesting the protein nature of receptor moiety on RBC. While trypsin treatment of RBC membranes
451
removes glycophorin A and glycophorin C (39), chymotrypsin cleaves the integral membrane protein
452
Band 3 of RBC (40).
453
We have also observed a similar binding trend of rBmAMA1 with human RBCs. The highest
454
rBmAMA1 binding was detected in untreated RBCs (IOD 269.3 K where K is 1000; 100% binding) and
455
in neuraminidase treated RBCs (IOD 261.5 K; 2.9% reduction), indicating that this binding is
456
neuraminidase independent (Fig. 5B). RBC treatment with trypsin (IOD 10.6 K; 96.1% reduction) and
457
chymotrypsin (IOD 40 K; 85.2% reduction) abolished the binding (Fig. 5B), Thus, the ligand binding of
458
BmAMA1 to RBC membranes can be attributed to a trypsin and chymotrypsin sensitive receptor. No
459
binding of RBC membrane was observed with purified, E. coli expressed Plasmodium falciparum
460
circumsporozoite protein that served as a negative control (Fig. 5C).
461 462
Inhibition of in vitro B. microti growth in human RBC by anti-BmAMA1 antibody. Among
463
apicomplexans, AMA1 proteins are secreted from micronemes and are thought to be involved in host
464
cell invasion (41). Antibodies raised against AMA1 have been shown to effectively block the invasion
465
of P. falciparum (42, 43), T. gondii (44), B. bovis (45, 46) and B. divergens (47) parasites into host cells.
466
Accordingly, we determined the inhibitory effect of a mouse anti--BmAMA1 polyclonal antibody
20
467
directed against rBmAMA1 domains I and II on the growth of B. microti parasites in human RBC after 8
468
h and 24 h cultivation in three independent experiments. Compared to the medium control (PBS) and
469
pre-immune groups, anti-BmAMA1 sera exhibited very strong growth inhibitory activity. At 8 hr. post-
470
cultivation parasitemias were 4.8% (medium control), 5.2% (un-immunized) and 1.2% (anti-BmAMA1)
471
respectively. A similar level of inhibitory effect was also evident after 24 h of cultivation with
472
parasitemias of 5.4% (medium control), 5.2% (un-immunized) and 1.0% (anti-BmAMA1) (Fig. 6).
473
Overall, anti-BmAMA1 antibody had highly significant inhibitory effects that corresponded to 82% and
474
80% (p
