Superoxide dismutases in trypanosomatids
1
The Presence of Four Iron-Containing Superoxide Dismutase Isozymes in
2
Trypanosomatidae: Characterization, Subcellular Localization and Phylogenetic Origin
3
in Trypanosoma brucei
4
5
Fabienne Duferneza, Cédric Yernauxb, Delphine Gerbodb, Christophe Noëla,c, Mélanie
6
Chauveneta, René Wintjensd, Virginia P. Edgcombe, Monique Caprona, Fred R. Opperdoesb,
7
and Eric Viscogliosia,*
8
9
a
Institut Pasteur, Inserm U547, 1 Rue du Professeur Calmette, B. P. 245, F-59019 Lille cedex,
10
France
11
b
12
Christian de Duve Institute of Cellular Pathology (ICP)and Catholic University of Louvain
13
(UCL), Avenue Hippocrate 74-75, B-1200 Brussels, Belgium
14
c
15
Newcastle upon Tyne, NE1 7RU, UK
16
d
17
de la Plaine, Boulevard du Triomphe, B-1050 Bruxelles, Belgium
18
e
19
MA 02543, USA
Research Unit for Tropical Diseases (TROP) and Laboratory of Biochemistry (BCHM),
School of Biology, Institute for Research on Environment and Sustainability, University of
Université Libre de Bruxelles, Institut de Pharmacie, Chimie Générale, CP 206/04, Campus
Department of Marine Chemistry and Geochemistry, 203 McLean Lab, MS#8, Woods Hole,
20
21
*Corresponding
22
(+33) 320877888.
author. E-mail eric.viscogliosi@pasteur-lille.fr; Tel. (+33) 320877961 ; Fax
23
24
Running title: Superoxide dismutases in trypanosomatids
25
26
1
Superoxide dismutases in trypanosomatids
27
28
Abstract
Metalloenzymes such as the superoxide dismutases (SODs) form part of a defense
29
mechanism that helps protect obligate and facultative aerobic organisms from oxygen toxicity
30
and damage. Here, we report the presence in the trypanosomatid genomes of four SOD genes:
31
soda, sodb1 and sodb2 and a newly identified sodc. All four genes of Trypanosoma brucei
32
have been cloned (Tbsods), sequenced and overexpressed in Escherichia coli and shown to
33
encode active dimeric FeSOD isozymes. Homology modelling of the structures of all four
34
enzymes using available X-ray crystal structures of homologs showed that the four TbSOD
35
structures were nearly identical. Subcellular localization using GFP-fusion proteins in
36
procyclic insect trypomastigotes shows that TbSODB1 is mainly cytosolic, with a minor
37
glycosomal component, TbSODB2 is mainly glycosomal with some activity in the cytosol
38
and TbSODA and TbSODC are both mitochondrial isozymes. Phylogenetic studies of all
39
available trypanosomatid SODs and 106 dimeric FeSODs and closely related cambialistic
40
dimeric SOD sequences suggest that the trypanosomatid SODs have all been acquired by
41
more than one event of horizontal gene transfer, followed by events of gene duplication.
42
43
44
45
Keywords
Antioxidant enzymes; Evolution; Structural models; Subcellular localization;
Superoxide dismutase; Trypanosoma
46
2
Superoxide dismutases in trypanosomatids
47
48
Introduction
Superoxide dismutases (SODs; EC 1.15.1.1) are a group of metalloenzymes that
49
eliminate superoxide radicals by dismutation into hydrogen peroxide and molecular oxygen
50
[1]. In concert with catalase, SODs have strong anti-oxidant properties and have been shown
51
to protect normal cells as well as a number of pathogens from reactive oxygen species (ROS).
52
According to their metal cofactor, SODs can be classified into three isoform types:
53
copper/zinc-, manganese- and iron-containing enzymes (Cu/ZnSOD, MnSOD and FeSOD).
54
Some classes of bacteria are also known to possess activity with either iron or manganese
55
incorporated in the same protein moiety (cambialistic SODs). FeSODs and MnSODs appear
56
as homodimers or homotetramers and exhibit a high degree of sequence and structure
57
similarity [2,3], strongly suggesting that these enzymes have a common ancestry. Recently,
58
most of the residues potentially involved in the mode of oligomerization and metal ion
59
specificity of the Fe/MnSOD family have been identified [4]. All obligate and facultative
60
aerobic organisms possess SOD. Typically, eukaryotes including mammals have a Cu/ZnSOD
61
in their cytosol, and MnSOD in the mitochondrial matrix whereas FeSODs have been found
62
in prokaryotes, protozoans, and chloroplasts of plants and algae.
63
Trypanosomatids, including the causative agents of African sleeping sickness
64
(Trypanosoma brucei gambiense, Trypanosoma brucei rhodesiense), Chagas disease
65
(Trypanosoma cruzi) and the different manifestations of leishmaniasis, all have SODs to
66
protect themselves against oxidative stress. However, in these parasites, so far, only evidence
67
has been found for the presence of FeSOD. MnSOD and Cu/ZnSOD could not be detected.
68
Interestingly, catalase, the enzyme that catalyzes the dismutation of hydrogen peroxide to
69
oxygen and water and which is present in most aerobic organisms has not been detected in
70
any of the pathogenic trypanosomatids. The known sensitivity of trypanosomatids towards
71
ROS and the absence of catalase, renders the other enzymes of the oxidant stress protection
72
system such as FeSOD promising targets for the development of parasite-specific drugs. Thus
3
Superoxide dismutases in trypanosomatids
73
far several members of the FeSOD family and their corresponding genes have been described
74
for T. cruzi, T. brucei and Leishmania spp. In T. brucei, SOD activity has been detected in
75
long slender forms; however, whether it is present in other developmental stages of the
76
parasite is not known [5,6]. In bloodstream forms, SOD activity has been shown to be present
77
in the cytoplasm, in the mitochondrion and in the glycosomes [5]. For T. cruzi two SOD
78
genes, FeSODA and FeSODB, have been cloned and characterized [7,8]. A FeSODB gene of
79
T. brucei encoding a protein of 198 residues has been cloned and its gene product functionally
80
characterized as an iron-containing SOD [6]. This FeSOD was shown to be developmentally
81
regulated with the highest level of expression in rapidly dividing cells but its subcellular
82
location was not described. Two closely related FeSODs (SODB1 and SODB2) have also
83
been identified in Leishmania chagasi and these two isozymes were localized within the
84
glycosomes [9,10]. Leishmania chagasi parasites in which one allele of the Lcsodb1 gene had
85
been knocked out exhibited a reduction in growth when endogenous superoxide levels were
86
increased with paraquat in culture and a reduction in survival within human macrophages
87
suggesting that FeSODB plays an important role in survival and growth of Leishmania. Also
88
FeSODC (C-type according to the present study)-depleted Leishmania tropica parasites
89
showed enhanced sensitivity to menadione and hydrogen peroxide in axenic culture, and a
90
markedly reduced survival in mouse macrophages [11].
91
Here, we describe the presence in T. brucei of four SOD genes: Tbsoda, Tbsodb1 and
92
Tbsodb2 and a newly identified Tbsodc. All genes encode active dimeric FeSODs, which in
93
procyclic insect trypomastigotes are localized to the cytosolic, glycosomal and mitochondrial
94
counterparts, respectively. Their phylogenetic relationship suggests that these SODs have
95
been acquired by more than one events of horizontal gene transfer followed by an event of
96
gene duplication.
97
98
Materials and methods
4
Superoxide dismutases in trypanosomatids
99
100
Cloning and sequencing of Tbsod genes
SOD sequences were retrieved by BLAST searches [12] at the T. brucei strain 927/4
101
GeneDB web site (http://www.genedb.org/genedb/tryp/index.jsp) using the publicly available
102
SODB sequence from T. brucei strain TC221 (Accession number: AF364812). Four distinct
103
SOD genes called soda, sodb1, sodb2, and sodc, respectively, were identified. From these
104
sequences, oligonucleotides flanking the coding region of each gene were designed (sense
105
primer Tryp1F and antisense primer Tryp2R, respectively; the list of all the primers used in
106
this study is provided in supplementary information) and used to amplify the homologous
107
sequences from T. brucei strain 427 (Tbsods). PCRs were carried out according to standard
108
conditions for Platinum Taq DNA polymerase High Fidelity (Invitrogen) using genomic DNA
109
of T. brucei strain 427 as template. After the denaturation step at 94°C for 5 min, 40 cycles of
110
amplification were performed with a GeneAmp PCR System 9700 apparatus (Applied
111
Biosystems) as follows: 1 min at 94°C, 1 min at 60°C and 1 min at 72°C. The final extension
112
step was continued for 15 min. The PCR products were separated by agarose gel
113
electrophoresis, and the bands of the expected sizes were purified using the QIAEX II Gel
114
Extraction Kit (Qiagen). Purified PCR products were cloned in the T-vector, pCR 2.1-TOPO
115
(Invitrogen) and amplified in E. coli TOP10 competent cells. Minipreparations of plasmid
116
DNA were done using the QIAprep Spin Miniprep Kit (Qiagen). Plasmids containing inserts
117
(pCR 2.1-Tbsods) were sequenced on both strands by primer walking using the Big Dye
118
Terminator Cycle Sequencing Kit (Applied Biosystems) and an automated ABI PRISM 377
119
DNA Sequencer (Applied Biosystems). The Tbsod gene sequences obtained in this study have
120
been deposited in GenBank under the accession numbers AY894557 to AY894560.
121
122
123
124
Computer-based analysis of proteins
Various physical and chemical parameters of TbSODs were computed with ProtParam
(http://us.expasy.org/cgi-bin/protparam). Targeting signals of proteins and signal peptide
5
Superoxide dismutases in trypanosomatids
125
cleavage sites were predicted by SignalP v3.0 (http://www.cbs.dtu.dk/services/SignalP/) [13],
126
TargetP v1.01 (http://www.cbs.dtu.dk/services/TargetP/) [14,15], iPSORT (http://hc.ims.u-
127
tokyo.ac.jp/iPSORT/) [16], MitoProt II v1.0a4 (http://ihg.gsf.de/ihg/mitoprot.html) [17],
128
Predotar v1.03 (http://genoplante-info.infobiogen.fr/predotar/), and Phobius
129
(http://phobius.cgb.ki.se/) [18].
130
131
Homology modelling
132
Three-dimensional models of the TbSODs were built with the automated comparative
133
modelling program Modeller 6.2 [19] using as homologous protein templates highly-resolved
134
X-ray structures of dimeric FeSODs from E. coli (Protein Data Bank (PDB) code: 1isa; x-ray
135
resolution 1.8 Å) and Pseudomonas ovalis (1dt0; 2.1 Å) and dimeric cambialistic SOD from
136
Porphyromonas gingivalis (1qnn; 1.8 Å). For each TbSOD sequence, five dimeric models
137
were built and the best one was kept according to the modeller objective function. The iron
138
ions and the water molecules of the active sites were taking into account in the molecular
139
modelling procedure. Note that as no homologous residues were available for the amino-
140
terminal 35 residues of TbSODA and the amino-terminal 97 residues of TbSODC, these both
141
parts were therefore omitted in our analysis. The final models had in disallowed regions of
142
Ramachandran map only a single residue, no residue, no residue and two residues for
143
TbSODA, TbSODB1, TbSODB2, and TbSODC, respectively. The equivalent resolutions
144
according to this criteria computed by PROCHECK-NMR program [20] were 1.4 Å, 1.0 Å,
145
1.0 Å and 1.7 Å for TbSODA, TbSODB1, TbSODB2, and TbSODC, respectively. Secondary
146
structures were defined by DSSP program [21] and models were analyzed by PROMOTIF
147
[22]. Pictures of the monomers of TbSODs were obtained using a combination of Molscript
148
[23] and Raster 3D [24].
149
150
Protein expression and SOD activity assays
6
Superoxide dismutases in trypanosomatids
151
Overproduction of TbSODs was done using a two-plasmid system [25]. 150 ng of
152
purified plasmid pCR 2.1-Tbsod was used as template to amplify the complete open reading
153
frame of Tbsod genes. Amplifications consisted of 25 cycles with steps identical to those of
154
the PCR described above. The sense oligonucleotide Tryp3F created a NdeI site at the
155
position of the ATG codon of the SOD coding sequences whereas the antisense primer
156
Tryp4R created a PstI site downstream from the stop codon. In the case of the TbSODC, two
157
distinct Tryp3F primers were used, each including one ATG codon corresponding to Met1 and
158
Met40, respectively (numbering as in the TbSODC amino acid sequence; see supplementary
159
material and Table 1), found in the long amino-terminal extension of this polypeptide. PCR
160
products were digested with NdeI and PstI (Invitrogen) then ligated in-frame with the
161
similarly digested pT7-7 expression vector using standard techniques. All PCR-derived clones
162
were sequenced and found to match the parent clones pCR 2.1-Tbsods. Following sequence
163
verification, the pT7-7 vector containing the introduced gene was used to transform
164
competent E. coli double mutant strain K38, lacking genes for both MnSOD and FeSOD [26]
165
and containing the pGP1-2 plasmid. Expression of the introduced genes was initiated by heat
166
induction as previously described [27]. As a control, competent E. coli strain K38/pGP1-2
167
transformed with the pT7-7 vector without insert, was used. After centrifugation of bacteria,
168
the pellet was resuspended in lysis buffer (0.3 mg/ml lysozyme; 20 mM Tris-HCl, pH 7.2; 2
169
mM EDTA) for 2 h on ice. Then, 0.1 mg/ml DNase and 10 mM MgCl2 were added to the
170
sample and the incubation was performed for 2 h on ice. Lysed cells were centrifuged for 30
171
min at 4°C and the supernatant was stored at –30°C until use for SOD activity assays. The
172
concentration of proteins in the supernatant was determined using the BCA Assay Kit
173
(Ultima). SOD activity of lysates was determined by the pyrogallol autoxidation method [28]
174
in which 1 U of SOD activity corresponds to 50% inhibition of the pyrogallol autoxidation
175
observed in the control in the absence of the enzyme. In the present method, 1 U corresponds
176
to the activity of 100 ng/ml of bovine Cu/ZnSOD.
7
Superoxide dismutases in trypanosomatids
177
178
Intracellular localization constructs
Vector pHD1336 (kindly provided by C. Clayton, Heidelberg, Germany) was used as
179
starting material for the construction of a plasmid suitable for the expression of GFP-fusion
180
proteins in trypanosomes. It contains the PARP promoter under the control of the tetracycline
181
operator, 5’- and 3’-untranslated regions of the actin gene, and the blasticidin resistance gene.
182
The complete open reading frame of GFP combined with the adjacent multiple cloning site
183
(MCS) were amplified by PCR from the plasmids pEGFP-N1 and pEGFP-C1 (Clontech).
184
Plasmid pHD1336 was doubly digested by HindIII and BamHI and the 5’ overhanging ends
185
were filled by incubation with Taq DNA polymerase (TaKaRa) in appropriate buffer supplied
186
with 0.2 mM dNTPs. The resulting blunt ended vectors were used for ligation with the GFP-
187
MCS PCR products, yielding the plasmids pGC1 and pGN1, respectively, for carboxy- and
188
amino-terminal fusion with the GFP. This strategy left the restriction sites BamHI and HindIII
189
available for cloning. The coding region of the Tbsoda, Tbsodb1, and Tbsodb2 genes was
190
amplified by PCR using a sense primer Tryp5F which created a HindIII site and an antisense
191
primer Tryp6R which created a BamHI site. The coding region of the Tbsodc gene was
192
amplified using sense (Tryp7F) and antisense (Tryp6R) primers that both created a BamHI
193
site. Primers Tryp6R-SODA and Tryp6R-SODC also mutated one nucleotide position
194
allowing the suppression of the stop codon. All amplifications consisted of 35 cycles of 1 min
195
at 94°C, 1 min at 50°C, and 1 min at 72°C using 150 ng of purified plasmid pCR2.1-Tbsod as
196
template. The PCR-amplified fragment of the Tbsoda gene was doubly digested with HindIII
197
and BamHI and cloned in the similarly digested plasmid pGN1 using standard methods. PCR
198
products of the Tbsodb1 and Tbsodb2 genes were similarly cloned in the pGC1 vector. The
199
pGC1 construct containing the Tbsodb2 gene was further mutated by site-directed
200
mutagenesis to suppress the restriction site NotI found in the nucleotide sequence, using the
201
QuickChange Site-Directed Mutagenesis Kit according to the protocol of the manufacturer
202
(Stratagene). Briefly, 10 ng of purified plasmid pGC1-Tbsodb2 and the sense Tryp8F and
8
Superoxide dismutases in trypanosomatids
203
antisense Tryp9R primers were used. Both primers changed C for A in position 129
204
(numbering as in the nucleotide sequence of the Tbsodb2 coding region) without any mutation
205
at the amino acid level. The Tbsodc gene amplicon was digested with HindIII then ligated in-
206
frame in the pGN1 vector after dephosporylation by Alcaline Phosphatase (Amersham). The
207
pGC1 and pGN1 plasmids containing the introduced genes were used to transform competent
208
E. coli TOP10 cells. Following sequence verification, positive clones were directly used in
209
transfection assays.
210
211
Parasite culture, transfection, and immunofluorescence
212
Procyclics of T. brucei strain 427 (cell line 449, constitutively expressing the
213
tetracyclin repressor) were cultured in SDM-79 medium [29] supplemented with 15% foetal
214
bovine serum (Gibco BRL) and 1 µg/ml phleomycin. Transfection was done as described
215
[30]. Briefly, 2-3 107 cells were centrifuged, washed once in ice-cold Zimmerman Post-
216
Fusion Medium (ZPFM) and resuspended in 500 µl ZPFM. Ten µg of pGC1 and pGN1
217
plasmids containing the introduced Tbsod genes was linearized overnight with NotI
218
(Fermentas), ethanol precipitated and resuspended in 20 µl of water. DNA and cells were
219
incubated together for 10 min on ice, in a 0.4 cm electroporation cuvette. Cells were then
220
subjected to a single pulse by a BTX ECM 630 electroporator set for a peak discharge of 1.8
221
kV, 25 Ω and 50 µF and directly diluted in 4.5 ml of SDM-79 medium. Selection was applied
222
the day after by addition of 4.5 ml SDM-79 with 20 µg/ml blasticidin. Induction of expression
223
was done by addition of 2 µg/ml tetracycline. Trypanosomes were allowed to grow overnight
224
before immunofluorescence analysis. Cells were fixed with 4% formaldehyde in PBS,
225
permeabilized with 1% Triton X-100 and settled on poly-L-lysine coated slides. Cells were
226
then incubated for 45 min in PBS/BSA 5%, followed by incubation in PBS/BSA 2% with the
227
primary antibody (rabbit polyclonal anti-T. brucei aldolase as glycosomal marker [31]). After
228
washing with PBS, cells were allowed to react with 5 µg/ml anti-rabbit antibodies-Alexa 568
9
Superoxide dismutases in trypanosomatids
229
(Molecular Probes), washed again and mounted in Mowiol. Mitochondria staining was
230
performed using Mitotracker-Red (Molecular Probes) according to the instructions of the
231
manufacturer. Cells were visualised using a Zeiss Axiovert microscope coupled to an MRC-
232
1024 confocal scanning laser imaging system (BioRad).
233
234
235
Western blot
Cell fractionation of parasites and isopycnic centrifugation of resulting fractions using
236
linear density sucrose gradients were done as previously described [32] and 15 fractions per
237
gradient were collected. Activities of marker enzymes for cytosol (6-phosphogluconate
238
dehydrogenase and alanine aminotransferase), mitochondria (NADP-isocitrate
239
dehydrogenase), and glycosomes (hexokinase, phosphoglycerate kinase, and glycerol-3-
240
phosphate dehydrogenase) were determined in each fraction as already described [32]. Protein
241
concentration in each fraction was determined by the method of Lowry et al. [33]. For
242
Western blots, 7 g/lane of T. brucei soluble protein of each fraction was separated on a 15%
243
SDS-PAGE and subsequently blotted onto nitrocellulose [34]. The blots were probed with
244
primary rabbit anti-T. brucei SODB polyclonal antibody [6] raised against the synthetic
245
peptide TTKKLKVFQTHDAGC used at 1:2000. After washing, the nitrocellulose membrane
246
was incubated with a peroxidase-conjugated goat anti-rabbit antibody at 1:10000 (Rockland),
247
and bound antibodies were visualized using the ECL Western blotting system (Amersham)
248
according to instructions of the manufacturer.
249
250
Phylogenetic analyses
251
A published protein alignment containing 81 dimeric FeSODs and closely related
252
cambialistic dimeric SOD sequences was available [4]. BLAST searches and retrieval of
253
additional protein sequences were performed using the NCBI web interface
254
(http://www.ncbi.nlm.nih.gov). Some sequences were also obtained from genome sequencing
10
Superoxide dismutases in trypanosomatids
255
programs and available on The Institute for Genomic Research (http://www.tigr.org/) and The
256
Wellcome Trust Sanger Institute (http://www.sanger.ac.uk/) web sites. Only full-length
257
sequences assigned as dimeric FeSODs and closely related cambialistic dimeric SODs
258
according to sequence and structure characteristics [4] were extracted. Additional SOD
259
sequences from trypanosomatids were identified by searching the Trypanosoma brucei
260
gambiense (http://www.sanger.ac.uk/Projects/T_b_gambiense/), Trypanosoma cruzi strain CL
261
Brener (http://www.genedb.org/genedb/tcruzi/index.jsp), Trypanosoma congolense
262
(http://www.sanger.ac.uk/Projects/T_congolense/), Trypanosoma vivax
263
(http://www.sanger.ac.uk/Projects/T_vivax/), Leishmania infantum clone JPCM5
264
(http://www.genedb.org/genedb/linfantum/index.jsp), and Leishmania major strain
265
MHOM/IL/80/Friedlin (http://www.genedb.org/genedb/leish/index.jsp) genome projects. All
266
in all, TbSOD sequences obtained in this study were added to a large data set including 162
267
non-trypanosomatid and 46 other trypanosomatid sequences belonging to the genera
268
Trypanosoma and Leishmania including a partial sequence from L. tropica (Accession
269
number: AY161306 [11]). Amino acid sequences from trypanosomatids and other organisms
270
were aligned with the use of the BioEdit v7.0.1 package
271
(http://www.mbio.ncsu.edu:BioEdit/bioedit.html) and only unambiguously alignable sites
272
were chosen for phylogenetic inference. Full-length alignments and sites used in analyses are
273
available upon request to the corresponding author. Removal of indels from the first
274
trypanosomatid data set including 49 full-length sequences yielded 188 sites for analysis. A
275
second trypanosomatid data set included the same sampling plus the partial SOD sequence
276
from L. tropica and allowed the analysis of 156 alignable positions. Phylogenetic analyses of
277
these two data sets were carried out using MrBAYES v3_0b4 [35]. Bayesian analyses were
278
performed using the Jones-Taylor-Thornton (JTT) amino acid replacement model [36]. In
279
both Bayesian analyses, starting trees were random, four simultaneous Markov chains were
280
run for 1 million generations, burn-in values were set at 35,000 generations (based on
11
Superoxide dismutases in trypanosomatids
281
empirical values of stabilizing likelihoods), and trees were sampled every 100 generations.
282
Bayesian posterior probabilities (BPP) were calculated using a Markov chain Monte Carlo
283
(MCMC) sampling approach [37] implemented in MrBAYES v3_0b4. Trees have been
284
rooted using the Midpoint rooting method with RETREE implemented in the PHYLIP
285
package v3.62 [38]. A broadest taxonomic sampling included the four TbSOD sequences
286
obtained in this study and the 162 non-trypanosomatid SOD sequences from prokaryotes and
287
eukaryotes extracted from databases as described above. When ambiguous sites were
288
removed for this sampling, a dataset of 168 positions was left for analysis. To reduce
289
computer time and to remove redundant sequences (uninformative closed sequences from
290
phylogenetically-related species and species-specific duplicated genes) for further tree
291
reconstructions, this data set was first analysed with the Protein Maximum Likelihood
292
(ProML) program (PHYLIP package v3.62) using the JTT probability model of change
293
between amino acids. Analysis of this latter tree allowed the selection of a reduced data set
294
(110 sequences and 168 positions) that was analysed using MrBAYES v3_0b4 as described
295
above. For the same data set, TREE-PUZZLE v5.1 [39] was also used to estimate the model
296
of amino acid transition, proportion of invariant sites and among-site-rate variation categories
297
under gamma or gamma plus invariant models. The model was VT [40] and the estimated
298
shape parameter  and the fraction of invariable sites were 1.27 and 0.05, respectively. ML
299
corrected distance matrices were calculated using TREE-PUZZLE v5.1 and the shell script
300
PUZZLEBOOT (by Holder and Roger: http://www.tree-puzzle.de) and then analyzed with
301
NEIGHBOR (PHYLIP package v3.62). Bootstrap values (BV) for the distance tree were
302
obtained from 1,000 replicates with SEQBOOT implemented in the PHYLIP package v3.62.
303
In one case where an interesting polyphyly was observed in the second dataset, the 10,000
304
trees generated by this second MrBAYES analysis were imported into PAUP* v4.0b10 [41]
305
where the 350 burn-in trees were removed. PAUP was then used to define a constraint group
12
Superoxide dismutases in trypanosomatids
306
of these taxa and to filter out those trees from the dataset that had held that group
307
monophyletic to indicate the percentage of total trees with that topological feature.
308
309
Results
310
Protein sequence and structure analyses of TbSOD enzymes
311
A search of the T. brucei strain 927/4 GeneDB web site with the protein sequence of a
312
previously identified T. brucei strain TC221 SODB [6] yielded four hits for putative SODs. It
313
included homologous proteins of SODB1 (Temporary accession: Tb11.01.6660), SODB2
314
(Tb11.01.7550), and SODA (Tb05.27M3.490) according to recent SOD gene assignments in
315
Leishmania species [10] and an as-yet unidentified new type of SOD (Tb11.01.7480) in
316
trypanosomatids that we call SODC in this study. In T. brucei strain 927/4, the sodb1, sodc,
317
and sodb2 genes are situated on the same chromosome 11 (proceeding from 5’ to 3’).
318
Genomic domains between coding regions of sodb1 and sodc and between those of sodc and
319
sodb2 are around 189 and 25 kB, respectively. On the other hand, the soda gene is located on
320
chromosome 5. A similar SOD gene grouping was also observed in L. major strain
321
MHOM/IL/80/Friedlin (sodb1, sodb2, and sodc together on chromosome 32 and soda on
322
chromosome 8) while sodb1 and sodb2 genes are organized in tandem in both L. chagasi and
323
L. donovani [10]. On chromosome 6, a fifth SOD-like gene (Tb06.4F7.290) was identified in
324
T. brucei strain 927/4. This gene is well conserved within the three trypanosomatids for which
325
the genome is being sequenced and the corresponding protein shares between 24-34% of its
326
residues with the other FeSODs and has a size that resembles the other SODs. However, none
327
of the residues predicted to be involved in the binding of the metal ion were conserved.
328
Therefore, the predicted protein cannot possibly function as a SOD and its real function
329
remains to be established.
330
The coding regions of the four types of bona-fide SODs in our strain T. brucei 427
331
were obtained by PCR using primers flanking the homologous protein sequences (Fig. 1).
13
Superoxide dismutases in trypanosomatids
332
They were called Tbsodb1, Tbsodb2, Tbsoda, and Tbsodc. At the amino acid level, Tbsods of
333
strain 427 exhibited a few differences with their counterparts of T. brucei strain 927/4 as
334
available in the genome database. The full open reading frames of Tbsodb1 and Tbsodb2
335
encoded proteins of 198 and 208 amino acids with predicted molecular masses of 22,048 and
336
23,266 Da, and predicted isoelectric points of 5.71 and 6.49, respectively. The amino acid
337
sequences of these two enzymes were 91% identical (see supplementary material) and
338
differed primarily by the presence of a 10-amino-acid extension found at the carboxyl
339
terminus of TbSODB2. In contrast to the TbSODB1 (-LKS), we noted that the last three
340
residues of TbSODB2 (-SDL) resembled the typical carboxy-terminal tripeptide SKL named
341
PTS-1 (peroxisome-targeting signal of type 1) [42] which constitutes the targeting signal of
342
the majority of glycosomal enzymes. However, this carboxy-terminal signal is known to be
343
highly degenerate [43] and it has been shown that the variants SDL and SQL allowed the
344
targeting of L. chagasi SODB1 and SODB2, respectively, to the glycosomes [10], suggesting
345
a possible localization of TbSODB2 in these organelles.
346
Analysis of the Tbsoda and Tbsodc genes revealed open reading frames which could
347
be translated to 238 and 309 amino acids, with molecular masses of 26,873 and 36,031 Da,
348
and predicted isoelectric points of 8.65 and 6.86, respectively. Moreover, in comparison to
349
TbSODBs, both TbSODA and TbSODC possessed an amino-terminal extension composed of
350
35 and 97 residues, respectively. Since such extensions could represent putative
351
mitochondrial targeting signals, we analysed their targeting potential using several prediction
352
algorithms (Table 1). A mitochondrial location of TbSODA was supported by three
353
algorithms, while only one supported such a location for TbSODC. Also the predicted length
354
of the identified mitochondrial transit peptides varied according to the algorithms used.
355
Amino acid sequences of TbSODs were aligned to those of 12 dimeric and tetrameric
356
FeSODs and MnSODs and cambialistic SODs from various organisms (Fig. 1). For each of
357
these 12 SOD sequences, a high-resolution X-ray structure is available allowing the
14
Superoxide dismutases in trypanosomatids
358
optimization of the alignment on the basis of the sequence and structure similarities [4]. From
359
the common part of our alignment (180 shared amino acid residues), it is apparent that the
360
most divergent protein of the four TbSODs is that of TbSODC (38% identity) in comparison
361
with TbSODA (43 and 44% identity with TbSODB1 and TbSODB2, respectively) while
362
TbSODC and TbSODA protein sequences exhibited 43% identity (see supplementary
363
material). Moreover, all four TbSODs exhibited a higher degree of identity to the dimeric
364
FeSOD from E. coli (41 to 54%) and P. ovalis (36 to 56%), and that of the cambialistic
365
dimeric SOD from P. gingivalis (36 to 44%) than to the dimeric MnSODs and tetrameric
366
MnSODs and FeSODs (22 to 42%) suggesting that TbSODs all belong to the class of dimeric
367
FeSODs. This was confirmed by the analysis of the ensemble of residues previously identified
368
as ensuring the metal specificity and/or oligomeric state of SOD enzymes [4]. In addition to
369
the conserved residues found in most of the SOD sequences (Figs 1 and 2) including those
370
that ligand the metal ion (His26, His73, Asp156, and His160; numbering as in the E. coli FeSOD
371
protein sequence), TbSOD protein sequences display almost all sequence features of dimeric
372
FeSODs. This included i) the residues systematically encountered in dimers and never in
373
tetramers such as Asn65, Phe118, and Pro144 and ii) the iron dimer-specific residues Phe64 (with
374
the exception of TbSODC exhibiting a conservative substitution Phe64-Tyr64 which is
375
common in dimeric FeSOD sequences from Amoebozoa and Apicomplexa as in some
376
bacteria), Ala68, Gln69, Phe75, and Ala141. These data allowed a very strong prediction of the
377
oligomeric state and nature of the metal co-factor of T. brucei strain 427 enzymes as dimeric
378
FeSODs.
379
Structural models of TbSODs were obtained by homology modelling using publicly
380
available X-ray structures of dimeric FeSODs. The four TbSOD structures were nearly
381
identical (Fig. 2). All contained the typical fold of dimeric FeSOD, consisting of 9 alpha-
382
helices and 3 beta-strands (Fig. 2A), and all exhibited the dimeric packing which closely bring
383
together the two active sites (Fig. 2B), although no cooperativity in function has been
15
Superoxide dismutases in trypanosomatids
384
reported. As mentioned above, TbSODs encompassed the characteristic side chains of both
385
SOD proteins (shown in Fig. 2D) and dimeric FeSODs (see Fig. 2C). We also noted that most
386
of these conserved side chains were located around the metal ion/active site (Fig. 2D). The
387
only differences between TbSOD three-dimensional structures were the amino-terminal 310
388
helices before the H1a helix encountered in TbSODB1 and TbSODB2 models (depicted in
389
blue on Figs 2A and 2B), and the 5-residue long loop insertion between H2a and H2b helices
390
found in TbSODA and TbSODC models (in blue on Figs 2C and 2D).
391
392
393
Expression of active TbSODs
Mutant E. coli strain K38/pGP1-2, deficient in both MnSOD and FeSOD genes, was
394
transformed with the pT7-7 vector, harbouring the Tbsod genes and heat-induced. Protein
395
concentration and SOD activity assays were performed on the lysates of induced bacteria. The
396
supernatants analyzed gave specific activities of 43, 50, and 144 U mg-1 protein for the
397
recombinant enzymes TbSODB1, TbSODB2, and TbSODA, respectively. No SOD activity
398
was detected in the supernatant of the lysed bacteria containing plasmid without insert
399
(control) and plasmid with the full-length TbSODC sequence (residues 1 to 309). This lack of
400
SOD activity of the recombinant form TbSODC1-309 could be explained by its insolubility
401
since a protein band of 36 kDa was only identified on the electrophoretic profile of the
402
bacterial pellet after lysis on a denaturing polyacrylamide gel (data not shown). An incorrect
403
folding of the native TbSODC protein and its subsequent precipitation at least in the bacterial
404
strain used could be linked to the presence of a very long amino terminal extension. Indeed, a
405
parallel could be drawn with the failure to recombinantly express in an active form the
406
mitochondrial SOD2 of Plasmodium falciparum which also exhibits an unusually long amino-
407
terminal extension [44]. Moreover, the recombinant form TbSODC40-309 corresponding to the
408
enzyme deleted for a part of its amino-terminal extension was soluble and showed a specific
409
SOD activity of 22 U mg-1.
16
Superoxide dismutases in trypanosomatids
410
Intracellular localization of TbSODs
411
As hypothesized from sequence analyses, TbSODB1 and TbSODB2 could represent
412
cytosolic and glycosomal enzymes, respectively, whereas TbSODA and TbSODC could be
413
imported into the mitochondrion. To confirm these possibilities, the plasmids pGC1 and
414
pGN1 were created to allow, respectively, carboxy- and amino-terminal fusion of proteins
415
with the GFP in transfection experiments. Since Tbsoda and Tbsodc encode amino-terminal
416
extensions with possible mitochondrial targeting information, the coding regions of both
417
genes were cloned in frame in the pGC1 vector, while Tbsodb1 and Tbsodb2, which could
418
carry glycosomal targeting information in their carboxy-terminal PTS-1, were cloned in frame
419
in the pGN1 vector. All four constructs were transfected into procyclic trypanosomes and
420
examined for fluorescence (Fig. 3). To accurately determine the localization of the TbSODs in
421
organelles, transfected parasites were stained both with an anti-glycosomal aldolase
422
polyclonal antibody [31] and Mitotracker-red, a dye that preferentially accumulates in the
423
mitochondria [45]. As shown in Fig. 3A, parasites expressing TbSODB1 fused to GFP
424
showed fluorescence throughout the cell, a pattern consistent with cytosolic localization. No
425
clear colocalization of GFP fluorescence and anti-glycosomal aldolase antibody staining was
426
observed. In the case of the TbSODB2, the subcellular pattern was more complex since this
427
enzyme was localized in both the cytosol and glycosomes (Fig. 3B). Indeed, although the
428
GFP fluorescence was uniformly distributed in the cytoplasm, we also noted the
429
colocalization of GFP fluorescence and anti-glycosomal aldolase antibody staining strongly
430
suggesting the partial targeting of TbSODB2 into the glycosomes. Both transfected parasites,
431
expressing TbSODB1- and TbSODB2-GFP, did not show colocalization of fluorescence and
432
Mitotracker staining (data not shown). In contrast to the previous constructs, mitochondrial
433
localization was clearly evident for transfected parasites expressing TbSODA- (Fig. 3C) and
434
TbSODC-GFP (Fig. 3D), where GFP fluorescence colocalized with Mitotracker staining. This
435
strongly suggests the targeting of both the A and C isozymes into the mitochondrion. No
17
Superoxide dismutases in trypanosomatids
436
colocalization of GFP fluorescence and anti-glycosomal aldolase antibody staining was
437
observed (data not shown).
438
In parallel, a polyclonal antibody raised against a synthetic peptide of the SODB of T.
439
brucei strain TC221 [6] was used for detection of homologous proteins in subcellular
440
fractions of T. brucei strain 427 by immunoblotting. Only the labelling of the most enriched
441
cytosolic, glycosomal, and mitochondrial fractions (according to the enzymatic activities of
442
the markers of these cellular compartments) were shown (Fig. 4). The peptide used as
443
immunogen for the production of the antibody (see Fig. 1) perfectly matched with both
444
TbSODB1 and TbSODB2 but showed very low identity with the homologous domains of
445
TbSODA and TbSODC. Thus, this antibody allowed us to detect specifically both TbSODBs
446
by Western blot. As shown in Fig. 4, the antibody did not recognize any protein in the
447
mitochondrial fraction but labelled two bands of unequal intensity of around 22 and 24 kDa,
448
respectively, in both the cytosolic and the glycosomal fraction, the lower band being the
449
TbSODB1 and the upper band the TbSODB2 according to their respective molecular mass.
450
The TbSODB1 was found predominantly in the cytosol with a small glycosomal component
451
and we estimate that around 40% of the TbSODB2 was targeted into glycosomes.
452
453
454
Overall phylogeny of FeSODs and the search for the origin of TbSODs
Systematic investigation of conserved sequence and structure characteristics of SODs
455
allowed us to determine the oligomeric state and metal specificity of these enzymes [4].
456
According to this study, TbSODs clearly belong to the family of dimeric FeSODs. Thus, all
457
the dimeric FeSODs and closely related cambialistic dimeric SODs available in databases
458
were extracted and aligned with TbSODs. This first dataset including 162 non-trypanosomatid
459
sequences was analyzed with the ProML program (data not shown) that allowed us to remove
460
phylogenetically uninformative redundant sequences. Subsequently, Bayesian and distance
461
unrooted trees (Fig. 5) were constructed with a reduced dataset including the four TbSODs
18
Superoxide dismutases in trypanosomatids
462
and 106 non-trypanosomatid sequences. The emergence of most of the lineages was not well
463
supported in view of the low BV or BPP on the corresponding nodes. This lack of
464
phylogenetic resolution was likely due to the low number of sites analyzed (168 shared
465
positions). Among the bacterial groups, we noted the emergence of two protozoan clades. The
466
first one not supported by BV and BPP values included dimeric FeSODs from Alveolates
467
(Plasmodium, Theileria, Cryptosporidium, Babesia, Toxoplasma, Neospora, Perkinsus),
468
Amoebozoa (Entamoeba and Phreatamoeba), and the TbSODB1 and TbSODB2 sequences
469
which showed a very discrete relationship with the mitochondrial enzyme from Perkinsus
470
marinus. The second protozoan clade only included the TbSODA and TbSODC sequences
471
and the trichomonad (Trichomonas and Tritrichomonas) enzymes. This grouping was well
472
supported by BV and BPP values (75 and 99%, respectively). Moreover, these four protozoan
473
sequences were related to the -proteobacteria homologues from Helicobacter,
474
Campylobacter, and Wolinella but this clustering was only strongly supported by BPP of 82%
475
but not by BV. In our tree, the TbSODA and TbSODC clustered together with high support
476
(BV and BPP of 78 and 99%, respectively) as did TbSODB1 and TbSODB2 (both BV and
477
BPP of 100%).
478
As emphasized in previous studies [44], FeSODs from protozoans including
479
Trypanosoma are clearly prokaryotic in nature according to the sequence, structure, and
480
phylogenetic data. Moreover, this prokaryotic nature and the presence of two distinct clades
481
of TbSODs are suggestive of more than one event of lateral gene transfer which gave rise to
482
the appearance of the multiple SODs in T. brucei. To further investigate this possibility, we
483
imported into PAUP the tree file containing all 10,000 trees from the Bayesian analysis that
484
were used to construct the consensus tree. After removing the first 350 trees representing the
485
burn-in trees in the Bayesian analysis, PAUP was used to set a constraint holding the four
486
TbSOD sequences together as a monophyletic group and then to search among the remaining
487
trees in the tree file for those that showed that monophyletic group. None of the trees had the
19
Superoxide dismutases in trypanosomatids
488
four taxa confined together as a monophyletic group, offering further support for the idea that
489
they have two separate origins. If burn-in is increased to 1,000 trees to be even more
490
conservative, the consensus tree groups TbSODA and TbSODC 100% of the time, and
491
TbSODB1 and TbSODB2 100% of the time, but the two groups are still always separated.
492
Phylogenetic evaluation of the trypanosomatid SOD sequences
493
In an attempt to retrace more precisely the evolution of the SOD genes within the
494
trypanosomatids and to study phylogenetic relationships among these parasites, all the
495
available SOD sequences found in databases from these protozoans were extracted so
496
allowing the phylogenetic analysis of 49 full-length and one partial SOD sequences. All
497
SODs showed the characteristics of dimeric FeSODs and could be included in one of the four
498
SOD classes here described for T. brucei. Interestingly, although the genome sequencing of
499
some of the trypanosomatid species analysed here is still only partially complete, the four
500
types of SOD here identified are all present in T. gambiense, T. vivax, T. congolense, T. cruzi,
501
L. infantum, and L. major suggesting that these four SODs are common to all
502
trypanosomatids. A first Bayesian analysis including the 49 full-length FeSOD sequences and
503
188 sites was performed and the tree is shown in Fig. 6A. A salient point is the highly
504
supported divergence (BPP of 100%) observed between trypanosomatid SODA and SODC
505
sequences suggesting an earlier gene duplication which occurred before the Trypanosoma and
506
Leishmania lineages separated. Within both SODA and SODC clusters, we noted two strong
507
dichotomies between i) the Trypanosoma and Leishmania sequences, and ii) the T. brucei and
508
T. cruzi homologues as shown in recent SSU rRNA- and GAPDH-based trees [46]. A second
509
Bayesian analysis restricted to the 156 positions shared with the partial and as-yet unclassified
510
SOD sequence from L. tropica (see partial tree Fig. 6B) showed that the latter sequence
511
clustered together with other Leishmania SODC sequences with very high support (BPP of
512
100%). From our analyses, we showed that Trypanosoma SODB1 and SODB2 sequences
513
emerged in strongly supported species-specific clusters (Fig. 6A). Although such a topology
20
Superoxide dismutases in trypanosomatids
514
can be interpreted as the result of recent gene duplications which occurred independently in
515
each Trypanosoma species, such an evolutionary scenario is unlikely and therefore we
516
interprete this to indicate that an early gene duplication leading to the formation of highly
517
similar sodb1 and sodb2 genes in a trypanosomatid ancestor has subsequently undergone gene
518
conversions and homologous recombinations which led to frequent homogenization of the
519
two sodb sequences. This scenario is supported by the fact that in T. brucei the sodb1 and
520
sodb2 genes are closely together on the same chromosome, or as in the genomes of L. major,
521
L. chagasi and L. donovani, tandemly arranged [10] and that they are, only with the exception
522
of the carboxy-terminal extension found in the SODB2 sequence, still > 90% identical. It is
523
well known that in T. brucei genes have high tendency to undergo homologous recombination
524
necessary for the high rate of antigenic variation of these organisms. Moreover a similar
525
example of homogenization of homologous sequences has been described for the PGK
526
isozymes of T. brucei [47-49]. In order to study this in more detail, the method of split
527
decomposition was used allowing the visualization of complex evolutionary processes like
528
recombination and horizontal gene transfer [50]. The sequences of the Leishmania SODB1/2
529
clade revealed networked evolution, indicative of such recombination, while those of all other
530
clades, including the Trypanosoma SODB1/2 clade, were devoid of any network structure.
531
Thus apart from the B-type SODs, the other SODs have not been subject to significant
532
recombinational events, while due to the much higher rate of homologous recombination
533
amongst the Trypanosoma sodb species any such information must have been erased (data not
534
shown). The splitstree is available from the authors on request. The fact that the
535
corresponding SODB2 of Leishmania also carries a carboxy-terminal extension with a similar
536
PTS-1, supports the idea that the acquisition by trypanosomatids of the glycosomal isozyme
537
predates the separation of Leishmania and Trypanosoma species. In contrast to the single
538
SODB1/SODB2 clade for Trypanosoma, the Leishmania SODB1 and SODB2 isozymes
539
constitute separate clades. Apparently events of gene conversion and homologous
21
Superoxide dismutases in trypanosomatids
540
recombination in the Leishmania spp. have been less frequent than in the Trypanosoma spp.
541
This may be related to the fact that contrary to the genus Trypanosoma, Leishmania is not
542
capable of antigenic variation.
543
544
545
Discussion
In all Trypanosomatidae analyzed, there are 4 iron-containing SOD isozymes,
546
FeSODA, FeSODB1, FeSODB2 and FeSODC. We have over-expressed in E. coli all four
547
TbSODs and shown that they are active enzymes. TbSODA and TbSODC in procyclic insect
548
stages are located in the mitochondrion. Both carry N-terminal extensions which were
549
predicted by various algorithms to encode a mitochondrial transit peptide. These predictions
550
were confirmed by us in subcellular localization experiments using GFP-fusion proteins.
551
Although it was not possible to more precisely localize TbSODA and TbSODC within the
552
mitochondrion, the length of their N-terminal extensions suggests that they may represent,
553
respectively, the mitochondrial matrix and intermembrane-space isozymes. The same type of
554
subcellular localization experiments revealed that the two TbSODB isozymes are distributed
555
over both the cytoplasm and the glycosomes, with the B2 isozyme predominantly in the
556
glycosome and the B1 isozyme mainly in the cytoplasm. This subcellular distribution clearly
557
differs from that observed for the SODB1 and SODB2 isozymes identified in L. chagasi [10],
558
in which both were targeted exclusively to the glycosomes. In L. chagasi both isozymes carry
559
at their C-termini a tripeptide reminiscent of a PTS-1 glycosomal targeting signal (–SQL and
560
–SDL, respectively), and these two tripeptides have been shown to be responsible and
561
sufficient for the targeting of reporter proteins into glycosomes. Moreover, the reporter
562
proteins were found exclusively inside glycosomes. This contrasts with the situation in T.
563
brucei where both enzymes distribute over both the cytosolic and the glycosomal
564
compartments and the B2 isozyme carries a –SDL at its C terminus, identical to the
565
Leishmania B2 isozyme, while the B1 isozyme bears no recognizable PTS-1 (–LKS). Why in
22
Superoxide dismutases in trypanosomatids
566
T. brucei both B-type isozymes distribute over two cellular compartments rather than each
567
individual enzyme being targeted to a separate compartment as in L. chagasi is not entirely
568
clear. However, a search in the trypanosome genome databases has allowed us to identify
569
many genes coding for potential glycosomal proteins present in multiple copies, some of
570
which predict C-termini reminiscent of a PTS-1, while others do not conform to the consensus
571
pattern, because one or more of the three predicted C-terminal residues has been mutated into
572
a non-accepted variant. In a diploid organism such as T. brucei, a single nucleotide
573
polymorphism in the region encoding the PTS-1 could lead to the creation of enzyme subunits
574
both with and without functional PTS-1. This could have occurred by mere change, but it
575
could also serve a biological function. In the case of multimeric proteins, this would lead to
576
the formation of hetero-oligomers where some of the subunit(s) would bear a PTS-1 while
577
others do not. Contrary to the import of proteins into the endoplasmatic reticulum,
578
chloroplasts and mitochondria, protein unfolding does not seem to be a prerequisite for
579
protein import into peroxisomes. It has been demonstrated that both peroxisomes and
580
glycosomes can import completely folded and even oligomeric proteins [51-54]. For instance,
581
in the case of the dimeric FeSODBs, various combinations of the FeSODB1 and FeSODB2
582
subunits could lead to homo- and heterodimers bearing either zero, one, or two PTS-1 signals
583
and in this way both cytosolic and glycosomal isoenzymes would be created. If just one PTS-
584
1 per oligomer would be sufficient for the import of the native protein, differential expression
585
of two slightly different genes coding for an oligomeric enzyme would determine the relative
586
abundance of the three isozymes, and thus their respective distribution over the cytosol and
587
the glycosomes. This represents a novel mechanism of regulation of the relative contribution
588
of enzymes to the cytosolic and peroxisomal/glycosomal compartments.
589
In bloodstream forms SOD activity has been experimentally localized to the soluble,
590
mitochondrial and glycosomal fractions [5] but no information was available on the isozymes
591
responsible for these activities. The only isozyme which was identified in bloodstream forms
23
Superoxide dismutases in trypanosomatids
592
is a B-type SOD of which the activity increased with the rate of growth, but its subcellular
593
location was not studied [6]. Our studies show that TbSODB1/2 is present both in glycosomes
594
and in the cytosol. In actively dividing long slender bloodstream forms, the glycosome is
595
metabolically very active being responsible for the consumption of large amounts glucose via
596
an aerobic type of glycolysis which involves the consumption of molecular oxygen at the
597
impressive rate of 100 nanomole. min-1. mg protein-1. Most likely this high rate of
598
consumption via a cyanide-insensitive trypanosome alternative oxidase (TAO) imposes an
599
important degree of oxidative stress upon the organism [55]. None of the glycolytic enzymes
600
are directly involved in the consumption of molecular oxygen or have as cofactor FAD which
601
may be responsible for the production of superoxide radicals. Other reactions catalysed by
602
glycosomes may do so. Glycosomes of procyclics are involved in fatty acid oxidation [56],
603
and although a hydrogen-peroxide producing acyl CoA oxidase has not yet been detected,
604
such an enzyme would constitute a potential source of superoxide. Moreover, the presence of
605
a PTS-1 in other enzymes involved in the protection against activated oxygen species, such as
606
glutathione/trypanothione peroxidase and peroxiredoxin suggests that also glycosomes require
607
an effective protection against superoxide radicals.
608
Since Cyanobacteria and -proteobacteria both possess dimeric FeSODs, TbSODs
609
could have a chloroplastic or mitochondrial endosymbiotic origin. Although the TbSODA
610
(and consequently the related TbSODC) has recently been suggested to derive from a plastid
611
[57], the greater taxonomic sample used in this study does not provide evidence for a
612
chloroplastic origin of these genes. Neither of the enzymes cluster with cyanobacterial,
613
chloroplastic plant or chlorophyte homologues. Also there is no direct evidence for an -
614
proteobacterial affiliation of the TbSODs which could be interpreted as a mitochondrial origin
615
of TbSODs . In fact, protozoan FeSODs did not exhibit clear relationship to any bacterial
616
group with the exception of the clade including the TbSODA, TbSODC and trichomonad
617
sequences, However they cluster with -proteobacterial homologues rather than with -
24
Superoxide dismutases in trypanosomatids
618
proteobacteria. The most likely scenario for the origin of four bacterial-type FeSOD genes in
619
the Trypanosomatidae is by invoking two separate events of lateral gene transfer (LGT). The
620
precise order of these events and the identity of the respective donor organisms cannot be
621
retraced due to the lack of resolution in the bacterial SOD phylogenetic tree. The fact that
622
both events of LGT are being shared with the SODs of other protozoans indicates that these
623
events must have taken place very early in the evolution of protists as previously suggested
624
for the dimeric FeSOD of Entamoeba [58]. One event of LGT must have led to the acquisiton
625
of the B-type isozyme, which probably first only functioned as a cytosolic enzyme. After gene
626
duplication and the acquisition of a C-terminally located PTS-1 by one of them, this gene
627
became targeted to the glycosome. A second event of LGT must have led to the acquisition of
628
a mitochondrial isozyme which, after gene duplication, gave rise to both the A and C
629
isozymes now found in the trypanosomatids.
630
631
632
633
Acknowledgements
This work was supported by Interuniversity Attraction Pole programme of the Belgian
634
Government P5/29 (to F.R.O.), the Institut National de la Santé et de la Recherche Médicale,
635
the Institut Pasteur de Lille, and the Centre National de la Recherche Scientifique (to E.V.).
636
F.D. was supported by a grant from the Ministère Français de l’Education Nationale, de la
637
Recherche et de la Technologie. D.G. was supported by an ICP postdoctoral fellowship. We
638
thank J. Goldstone (WHOI, Woods Hole, USA) for his expertise in the phylogenetic analyses
639
and D. Steverding (School of Biological Sciences, University of Bristol) for the anti-SODB
640
antibody.
641
642
List of Abbreviations
25
Superoxide dismutases in trypanosomatids
643
BCA, bicinchoninic acid; BPP, Bayesian posterior probability; BSA, bovine serum
644
albumin;
BV,
bootstrap
value;
Cu/ZnSOD,
copper/zinc-containing
SOD;
EDTA,
645
ethylenediaminetetraacetic acid; FeSOD, iron-containing SOD; GAPDH, glyceraldehyde-3-
646
phosphate dehydrogenase; GFP, green fluorescent protein; JTT model, Jones-Taylor-
647
Thornton model; LGT, lateral gene transfer; MCS, multiple cloning site; ML, maximum
648
likelihood; MnSOD, manganese-containing SOD; PBS, phosphate buffered saline; PCR,
649
polymerase chain reaction; PTS-1, peroxisome-targeting signal of type 1; SDS-PAGE,
650
sodium dodecyl sulphate polyacrylamide gel electrophoresis; SOD, superoxide dismutase;
651
SSU rRNA, small subunit rRNA; TAO, trypanosome alternative oxidase; TbSOD, T. brucei
652
SOD protein; Tbsod, T. brucei SOD gene; ZPFM, Zimmerman post-fusion medium.
653
654
References
655
[1] McCord, J. M.; Fridovich, I. Superoxide dismutase: the first twenty years (1968-1988).
656
Free Radic. Biol. Med. 5:363-369; 1988.
657
[2] Parker, M. W.; Blake, C. C. F.; Barra, D.; Bossa, F.; Schinina, M. E.; Bannister, W. H.;
658
Bannister, J. V. Structural identity between the iron- and manganese-containing superoxide
659
dismutases. Protein Eng. 1:393-397; 1987.
660
[3] Jackson, S. M. J.; Cooper, J. B. An analysis of structural similarity in the iron and
661
manganese superoxide dismutases based on known structures and sequences. BioMetals
662
11:159-173: 1998.
663
[4] Wintjens, R.; Noël, C.; May, A. C. W.; Gerbod, D.; Dufernez, F.; Capron, M.; Viscogliosi,
664
E.; Rooman, M. Specificity and phenetic relationships of iron- and manganese-containing
665
superoxide dismutases on the basis of structure and sequence comparisons. J. Biol. Chem.
666
279:9248-9254; 2004.
667
[5] Opperdoes, F. R.; Borst, P.; Bakker, S.; Leene, W. Localization of glycerol-3-phosphate
668
oxidase in the mitochondrion and particulate NAD+-linked glycerol-3-phosphate
26
Superoxide dismutases in trypanosomatids
669
dehydrogenase in the microbodies of the bloodstream form to Trypanosoma brucei. Eur. J.
670
Biochem. 76:29-39; 1977.
671
[6] Kabiri, M.; Steverding, D. Identification of a developmentally regulated iron superoxide
672
dismutase of Trypanosoma brucei. Biochem. J. 360:173-177; 2001.
673
[7] Temperton, N. J.; Wilkinson, S. R.; Kelly, J. M. Cloning of an Fe-superoxide dismutase
674
gene homologue from Trypanosoma cruzi. Mol. Biochem. Parasitol. 76:339-343; 1996.
675
[8] Ismail, S. O.; Paramchuk, W.; Skeiky, Y. A. W.; Reed, S. G.; Bhatia, A.; Gedamu, L.
676
Molecular cloning and characterization of two iron superoxide dismutase cDNAs from
677
Trypanosoma cruzi. Mol. Biochem. Parasitol. 86:187-197; 1997.
678
[9] Paramchuk, W. J.; Ismail, S. O.; Bhatia, A.; Gedamu, L. Cloning, characterization and
679
overexpression of two iron superoxide dismutase cDNAs from Leishmania chagasi: role in
680
pathogenesis. Mol. Biochem. Parasitol. 90:203-221; 1997.
681
[10] Plewes, K. A.; Barr, S. D.; Gedamu, L. Iron superoxide dismutase targeted to the
682
glycosomes of Leishmania chagasi are important for survival. Infect. Immun. 71:5910-5920;
683
2003.
684
[11] Ghosh, S.; Goswami, S.; Adhya, S. Role of superoxide dismutase in survival of
685
Leishmania within the macrophage. Biochem. J. 369:447-452; 2003.
686
[12] Altschul, S. F.; Gish, W.; Miller, W.; Myers, E. W.; Lipman, D. J. Basic local alignment
687
search tool. J. Mol. Biol. 215:403-410; 1990.
688
[13] Bendtsen, J. D.; Nielsen, H.; von Heijne, G.; Brunak, S. Improved prediction of signal
689
peptides: SignalP 3.0. J. Mol. Biol. 340:783-795; 2004.
690
[14] Nielsen, H.; Engelbrecht, J.; Brunak, S.; von Heijne, G. Identification of prokaryotic and
691
eukaryotic signal peptides and prediction of their cleavage sites. Prot. Eng. 10:1-6; 1997.
692
[15] Emanuelsson, O.; Nielsen, H.; Brunak, S.; von Heijne, G. Predicting subcellular
693
localization of proteins based on their N-terminal amino acid sequence. J. Mol. Biol.
694
300:1005-1016; 2000.
27
Superoxide dismutases in trypanosomatids
695
[16] Bannai, H.; Tamada, Y.; Maruyama, O.; Nakai, K.; Miyano, S. Extensive feature
696
detection of N-terminal protein sorting signals. Bioinformatics 18:298-305; 2002.
697
[17] Claros, M. G.; Vincens, P. Computational method to predict mitochondrially imported
698
proteins and their targeting sequences. Eur. J. Biochem. 241:779-786; 1996.
699
[18] Käll, L.; Krogh, A.; Sonnhammer, E. L. L. A combined transmembrane topology and
700
signal peptide prediction method. J. Mol. Biol. 338:1027-1036; 2004.
701
[19] Sali, A.; Blundell, T. L. Comparative protein modelling by satisfaction of spatial
702
restraints. J. Mol. Biol. 234:779-815; 1993.
703
[20] Laskowski, R. A.; Rullmann, J. A.; MacArthur, M. W.; Kaptein, R.; Thornton, J. M.
704
AQUA and PROCHECK-NMR: programs for checking the quality of protein structures
705
solved by NMR. J. Biomol. NMR 8:477-486; 1996.
706
[21] Kabsch, W.; Sander, C. Dictionary of protein secondary structure, pattern recognition of
707
hydrogen-bonded and geometrical features. Biopolymers 22:2577-2637; 1983.
708
[22] Hutchinson, E. G.; Thornton, J. M. PROMOTIF - A program to identify and analyze
709
structural motifs in proteins. Protein Sci. 5:212-220; 1996.
710
[23] Kraulis, P. J. Molscript: a program to produce both detailed and schematic plots of
711
protein structures. J. Appl. Crystallogr. 24:946-950; 1991.
712
[24] Merrit, E. A.; Murphy, M. E. P. Raster 3D version 2.0. A program for photorealistic
713
molecular graphics. Acta Crystallogr. 50:869-873; 1994.
714
[25] Tabor, S.; Richardson, C. C. A bacteriophage T7 RNA polymerase / promoter system for
715
controlled exclusive expression of specific genes. Proc. Natl. Acad. Sci. USA 84:1074-1078;
716
1985.
717
[26] Carlioz, A.; Touati, D. Isolation and superoxide dismutase mutants in E. coli: Is
718
superoxide dismutase necessary for aerobic life? EMBO J. 5:623-630; 1986.
719
[27] Gratepanche, S.; Ménage, S.; Touati, D.; Wintjens, R.; Delplace, P.; Fontecave, M.;
720
Masset, A.; Camus, D.; Dive, D. Biochemical and electron paramagnetic resonance study of
28
Superoxide dismutases in trypanosomatids
721
the iron superoxide dismutase from Plasmodium falciparum. Mol. Biochem. Parasitol.
722
120:237-246; 2002.
723
[28] Marklund, S.; Marklund, G. Involvement of the superoxide anion radical in the
724
autoxidation of pyrogallol and a convenient assay for superoxide dismutase. Eur. J. Biochem.
725
47:469-474; 1974.
726
[29] Brun, R.; Schönenberger, M. Cultivation and in vitro cloning of procyclic culture forms
727
of Trypanosoma brucei in a semi-defined medium. Acta Trop. 36:289-292; 1979.
728
[30] Biebinger, S.; Rettenmaier, S.; Flaspohler, J.; Hartmann, C.; Pena-Diaz, J.; Wirtz, L. E.;
729
Hotz, H. R.; Barry, J. D.; Clayton C. The PARP promoter of Trypanosoma brucei is
730
developmentally regulated in a chromosomal context. Nucleic Acids Res. 24:1202-1211;
731
1996.
732
[31] Chevalier, N.; Callens, M.; Michels, P. A. M. High-level expression of Trypanosoma
733
brucei fructose-1,6-biphosphate aldolase in Escherichia coli and purification of the enzyme.
734
Protein Expr. Purif. 6:39-44; 1995.
735
[32] Opperdoes, F. R.; Markos, A.; Steiger, R. F. Localization of malate dehydrogenase,
736
adenylate kinase and glycolytic enzymes in glycosomes and the threonine pathway in the
737
mitochondrion of cultured procyclic trypomastigotes of Trypanosoma brucei. Mol. Biochem.
738
Parasitol. 4:291-309; 1981.
739
[33] Lowry, O. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. Protein measurement with
740
the Folin phenol reagent. J. Biol. Chem. 193:265-275; 1951.
741
[34] Towbin, H.; Staehelin, T.; Gordon, J. Electrophoretic transfer of proteins from
742
acrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad.
743
Sci. USA 76:4350-4354; 1979.
744
[35] Huelsenbeck, J. P.; Ronquist, F. MrBAYES: Bayesian inference of phylogenetic trees.
745
Bioinformatics 17:754-755; 2001.
29
Superoxide dismutases in trypanosomatids
746
[36] Jones, D. T.; Taylor, W. R.; Thornton, J. M. The rapid generation of mutation data
747
matrices from protein sequences. CABIOS 8:275-282; 1992.
748
[37] Green, P. J. Reversible jump Markov chain Monte Carlo computation and Bayesian
749
model determination. Biometrika 82:711-732; 1995.
750
[38] Felsenstein, J. PHYLIP (Phylogeny Inference Package) version 3.62. Distributed by the
751
author. Department of Genetics, University of Washington, Seattle. 1995
752
[39] Strimmer, K.; von Haeseler, A. Quartet puzzling: a quartet maximum likelihood method
753
for reconstructing tree topologies. Mol. Biol. Evol. 13:964-969; 1996.
754
[40] Muller, T.; Vingron, M. Modeling amino acid replacement. J. Comput. Biol. 7:761-776;
755
2000.
756
[41] Swofford, D. L. PAUP*. Phylogenetic analysis using parsimony (*and other methods),
757
Version 4. Sinauer Associates, Sunderland, MA. 1998
758
[42] Subramani, S. Protein translocation into peroxisomes. J. Biol. Chem. 271:32483-32486;
759
1996.
760
[43] Sommer, J. M.; Cheng, Q. L.; Keller, G. A.; Wang, C. C. In vivo import of firely
761
luciferase into the glycosomes of Trypanosoma brucei and mutational analysis of the C-
762
terminal targeting signal. Mol. Biol. Cell 3:749-759; 1992.
763
[44] Sienkiewicz, N.; Daher, W.; Dive, D.; Wrenger, C.; Viscogliosi, E.; Wintjens, R.; Jouin,
764
H.; Capron, M.; Müller, S.; Khalife, J. Identification of a mitochondrial superoxide dismutase
765
with an unusual targeting sequence in Plasmodium falciparum. Mol. Biochem. Parasitol.
766
137:121-132; 2004.
767
[45] Poot, M.; Zhang, Y. Z.; Kramer, J. A.; Wells, K. S.; Jones, L. J.; Hanzel, D. K.; Lugade,
768
A. G.; Singer, V. L.; Haugland, R. P. Analysis of mitochondrial morphology and function
769
with novel fixable fluorescent stains. J. Histochem. Cytochem. 44:1363-1372; 1996.
30
Superoxide dismutases in trypanosomatids
770
[46] Hamilton, P. B.; Stevens, J. R.; Gaunt, M. W.; Gidley, J.; Gibson, W. C. Trypanosomes
771
are monophyletic: evidence from genes for glyceraldehydes phosphate dehydrogenase and
772
small subunit ribosomal RNA. Int. J. Parasitol. 34:1393-1404; 2004.
773
[47] Le Blancq, S. M.; Swinkels, B. W.; Gibson, W. C.; Borst, P. Evidence for gene
774
conversion between the phosphoglycerate kinase genes of Trypanosoma brucei. J. Mol. Biol.
775
200:439-447; 1988.
776
[48] Parker, H. L.; Hill, T.; Alexander, K.; Murphy, N. B.; Fish, W. R.; Parsons, M. Three
777
genes and two isozymes: gene conversion and the compartmentalization and expression of the
778
phosphoglycerate kinases of Trypanosoma (Nannomonas) congolense. Mol. Biochem.
779
Parasitol. 69:269-279; 1995.
780
[49] Adjé, C. A.; Opperdoes, F. R.; Michels, P. A. M. Molecular analysis of phosphoglycerate
781
kinase in Trypanoplasma borreli and the evolution of this enzyme in Kinetoplastida. Gene
782
217:91-99; 1998.
783
[50] Huson, D. H. SplitsTree: analyzing and visualizing evolutionary data. Bioinformatics
784
14:68-73; 1998.
785
[51] Glover, J. R.; Andrews, D. W.; Subramani, S.; Rachubinski, R. A. Mutagenesis of the
786
amino targeting signal of Saccharomyces cerevisiae 3-ketoacyl-CoA thiolase reveals
787
conserved amino acids required for import into peroxisomes in vivo. J. Biol. Chem. 269:7558-
788
7563; 1994.
789
[52] McNew, J. A.; Goodman, J. M. An oligomeric protein is imported into peroxisomes in
790
vivo. J. Cell Biol. 127:1245-1257; 1994.
791
[53] Häusler, T.; Stierhof, Y. D.; Wirtz, E.; Clayton, C. Import of a DHFR hybrid protein into
792
glycosomes in vivo is not inhibited by the folate-analogue aminopterin. J. Cell Biol. 132:311-
793
324; 1996.
31
Superoxide dismutases in trypanosomatids
794
[54] Titorenko, V. I.; Nicaud, J. M.; Wang, H.; Chan, H.; Rachubinski, R. A. Acyl-CoA
795
oxidase is imported as a heteropentameric, cofactor-containing complex into peroxisomes of
796
Yarrowia lipolytica. J. Cell Biol. 156:481-494; 2002.
797
[55] Fang, J.; Beattie, D. S. Alternative oxidase present in procyclic Trypanosoma brucei may
798
act to lower the mitochondrial production of superoxide. Arch. Biochem. Biophys. 414:294-
799
302; 2003.
800
[56] Wiemer, E. A.; Ijlst, L.; van Roy, J.; Wanders, R. J.; Opperdoes, F. R. Identification of 2-
801
enoyl coenzyme A hydratase and NADP(+)-dependent 3-hydroxyacyl-CoA dehydrogenase
802
activity in glycosomes of procyclic Trypanosoma brucei. Mol. Biochem. Parasitol. 82:107-
803
111; 1996.
804
[57] Hannaert, V.; Saavedra, E.; Duffieux, F.; Szikora, J. -P.; Rigden, D. J.; Michels, P. A.
805
M.; Opperdoes, F. R. Plant-like traits associated with metabolism of Trypanosoma parasites.
806
Proc. Natl. Acad. Sci. USA 100:1067-1071; 2003.
807
[58] Smith, M. W.; Feng, D. -F.; Doolittle, R. F. Evolution by acquisition: the case for
808
horizontal gene transfers. Trends Biochem. Sci. 17:489-493; 1992.
809
810
32
Superoxide dismutases in trypanosomatids
811
Figure Legends
812
Fig. 1. Sequence alignment of the four TbSODs on 12 known SOD structures. Structure
813
alignments of SOD proteins were taken from [4]. The SOD structures are labeled by their
814
protein code with annotations about the oligomeric state and the metal cofactor specificity.
815
The 1mng SOD is an atypical tetramer (designed 4’mer) that has the sequence specificity of
816
dimers. The first line contains the PDB sequence numbering from the 1isa PDB file. Full
817
species names are as follows: E. coli, Escherichia coli; P. ovalis, Pseudomonas ovalis; P.
818
gingivalis, Porphyromonas gingivalis; S. acidocaldarius, Sulfolobus acidocaldarius; S.
819
solfataricus, Sulfolobus solfataricus; A. pyrophilus, Aquifex pyrophilus; M. tuberculosis,
820
Mycobacterium tuberculosis; P. shermanii, Propionibacterium shermanii; H. sapiens, Homo
821
sapiens; A. fumigatus, Aspergillus fumigatus; T. thermophilus, Thermus thermophilus. The
822
four TbSOD sequences are aligned below the 12 SOD structures, with consensus secondary
823
structures showed by colored boxes and labeled (alpha-helix and beta-sheet in red and green,
824
respectively). Secondary structure elements according to the DSSP program [21] are also
825
underlined in the TbSOD sequences. Sites of specific residue conservation in 261 SOD
826
sequences [4] are colored in all sequences; yellow background, residues conserved in at least
827
90% of 261 SOD sequences; turquoise background, residues specific to dimers; blue
828
background, tetramer-specific; purple background, iron-specific; orange background,
829
manganese-specific; violet background, iron dimer-specific; violet letters, specific for all but
830
iron dimers; dark khaki background, manganese dimer-specific; dark khaki letters, specific
831
for all but manganese dimers; green background, iron tetramer-specific; green letters, specific
832
for all but iron tetramers. More information can be found in [4]. Blue letters represent
833
residues which either introduce a short 310 helix in both TbSODB1 and TbSODB2 or
834
constitute a loop insertion in both TbSODA and TbSODC. The forward arrows (>) indicate
835
the omitted amino-terminal extensions of the TbSODA and TbSODC (35 and 97 residues,
33
Superoxide dismutases in trypanosomatids
836
respectively; see Table 1). The last line indicates the location of the synthetic peptide used as
837
immunogen for the production of an anti- T. brucei SODB polyclonal antibody [6].
838
839
Fig. 2. Ribbon representations of TbSOD models. Alpha-helices and beta-strands are colored in
840
red and in green, respectively. The amino- and carboxyl-termini are indicated. The iron ion and
841
active site water molecule are depicted in magenta and blue, respectively. (A) TbSODB1 model.
842
The secondary structure elements are labeled as in Fig. 1. The amino-terminal 310 helix is
843
represented in blue. (B) TbSODB2 model. Ternary conformation of the TbSODB2 dimer is
844
depicted. The particular 310 helix is showed in blue. (C) TbSODA model. Side chains specific to
845
dimers (Asn65, Phe118, Pro144; numbering according to the E. coli FeSOD protein sequence) and to
846
iron-dimer (Phe64, Ala68, Gln69, Phe75, Ala141) are labeled and colored in turquoise and in magenta,
847
respectively. The 5-residue long turn insertion is showed in blue. (D) TbSODC model. Amino acid
848
side chains conserved (>90%) in SOD sequences are depicted in yellow and labeled. There are
849
Leu7, Pro16, His26, His30, His31, Tyr34, Asn39, His73, Trp77, Ser120, Trp122, Pro151, Asp156, Trp158,
850
Glu159, His160, Ala161, Tyr162, Tyr163, Asn168, and Trp183. The loop insertion is displayed in blue.
851
852
Fig. 3. Subcellular localization of TbSODs. Parasites were transfected with plasmids containing
853
Tbsodb1 (A), Tbsodb2 (B), Tbsoda (C), and Tbsodc (D) genes coupled to GFP, stained with both
854
Mitotracker and anti-glycosomal aldolase antibody, and analyzed by fluorescence microscopy.
855
Images show GFP fluorescence only in the cytosol for the TbSODB1 (A) and in both cytosol and
856
glycosomes for TbSODB2 (B) as demonstrated by co-localization with the anti-aldolase antibody
857
staining. About TbSODA (C) and TbSODC (D), both enzymes are mitochondrial as shown by co-
858
localization of the GFP signals with Mitotracker staining. Scale bar, 10 m.
859
860
Fig. 4. Immunodetection of TbSODBs. Equal amounts of proteins (7 g/lane) from the most
861
enriched glycosomal (G, lane 1), cytosolic (C, lane 2), and mitochondrial (M, lane 3) fractions
34
Superoxide dismutases in trypanosomatids
862
were blotted onto nitrocellulose and probed with polyclonal antibody raised against SODB of
863
T. brucei strain TC221 [6]. This antibody recognizes two bands of unequal intensity of
864
approximately 22 and 24 kDa corresponding to the expected molecular mass of TbSODB1
865
and TbSODB2 monomers, respectively, in glycosomal and cytosolic fractions whereas no
866
protein is recognized by the same antibody in the mitochondrial fraction.
867
868
Fig. 5. Neighbor-joining tree based on SOD protein sequences. Accession numbers or
869
references of the sequences retrieved from genome sequencing projects are given in
870
parentheses. Sequences obtained in this study from T. brucei strain 427 are indicated in grey
871
whereas sequences from -proteobacteria, -proteobacteria, -proteobacteria, and
872
cyanobacteria are shown in blue, orange, red, and green, respectively. Sequences from
873
protozoans are underlined. If experimentally determined, the subcellular localization of
874
protozoan SODs is indicated in parentheses after sequence references (ND: not determined).
875
Numbers near the individual nodes indicate bootstrap values (left of the slash) and Bayesian
876
posterior probabilities (right of the slash) given as percentages by the two different tree
877
reconstruction methods (corrected distances/MrBAYES). Asterisks designate nodes with
878
values below 50%.
879
880
Fig. 6. Consensus Bayesian tree of trypanosomatids inferred from SOD protein sequences.
881
The root position was inferred by midpoint rooting. (A) Analysis of the data set including 49
882
trypanosomatid full-length SOD sequences (188 positions). (B) Partial view of the
883
phylogenetic tree obtained from the same data set plus the partial SOD sequence from L.
884
tropica (156 positions). This second tree exhibits the same topology as the previous one and
885
only the branch leading to the SODC from trypanosomatids is shown and indicated by a dot in
886
(A). Note the clustering of the boxed SOD sequence from L. tropica with other SODC
887
sequences from trypanosomatids. Accession numbers or references of the sequences retrieved
35
Superoxide dismutases in trypanosomatids
888
from genome sequencing projects are given in parentheses. Bayesian posterior probabilities
889
are given as percentages near the individual nodes. The scale bar indicate 0.1 substitution
890
(corrected) per site.
36