Molecular Evidence for b-tubulin Neofunctionalization in
Retaria (Foraminifera and Radiolarians)
Yubo Hou,*,1 Roberto Sierra,2 David Bassen,1 Nilesh K. Banavali,1,3 Andrea Habura,1 Jan Pawlowski,2 and
Samuel S. Bowser1
1
Wadsworth Center, New York State Department of Health
Department of Genetics and Evolution, University of Geneva, Geneva, Switzerland
3
Department of Biomedical Sciences, School of Public Health, State University of New York
*Corresponding author: E-mail: [email protected]; [email protected].
Associate editor: Andrew Roger
2
Abstract
Foraminifera and radiolarians are closely related amoeboid protists (i.e., retarians) often characterized by their shells and
pseudopodia. Previous studies hypothesized that the unusual “Type 2” b-tubulin (b2) is critically involved in forming
helical filaments (HFs), a unique microtubule (MT) assembly/disassembly intermediate found in foraminiferan reticulopodia. Such noncanonical b-tubulin sequences have also been found in two radiolarian species and appear to be closely
related to the foraminiferan b2. In this study, we report 119 new b-tubulin transcript sequences from six foraminiferans,
four radiolarians, and a related non-retarian species. We found that foraminiferan and radiolarian b2-tubulins share
some of the unusual substitutions in the structurally essential and usually conserved domains. In the b-tubulin phylogeny,
retarian b2-tubulin forms a monophyletic clade, well separated from the canonical b-tubulin (b1) ubiquitous in eukaryotes. Furthermore, we found that foraminiferan and radiolarian b2-tubulin lineages were under positive selection, and
used homology models for foraminiferan a- and b-tubulin hexamers to understand the structural effect of the positively
selected substitutions. We suggest that the positively selected substitutions play key roles in the transition of MT to HF by
altering the lateral and longitudinal interactions between a- and b-tubulin heterodimers. Our results indicate that the
unusual b2-tubulin is a molecular synapomorphy of retarians, and the b-tubulin gene duplication occurred before the
divergence of Foraminifera and radiolarians. The duplicates have likely been subjected to neofunctionalization responsible for the unique MT to HF assembly/disassembly dynamics, and/or other unknown physiological processes in retarian
protists.
Key words: b-tubulin, positive selection, Foraminifera, radiolarian, microtubule modeling.
Introduction
networks of branching and anastomosing granuloreticulose
pseudopods that can rapidly extend, bend, and retract (reviewed in Travis and Bowser 1991; Bowser and Travis 2002).
Reticulopodia are involved in many life processes of forams,
especially the efficient exploration of environments for food
and other resources. Radiolarians are characterized by axopodia, which are rigid, ray-like projections stiffened by highly
patterned microtubule (MT) bundles or axonemes (reviewed
in Anderson 1983). Axopodia assist in floating and are mostly
responsible for phagocytosis by rapid retraction upon contact
with prey. Radiolarians also possess other types of pseudopodia, e.g., thread-like filopodia and branching rhizopodia; the
latter are thought to be analogous to foram reticulopodia
(reviewed in Anderson 1983; Suzuki and Aita 2011).
Cercozoans are often able to form filopodia or axopodia (in
Phaeodarea), but their pseudopodial systems are generally less
elaborated than those of retarian protists (reviewed in
Hausmann et al. 2003).
MTs play an integral role in reticulopodial and axopodial
motility and structure (Anderson 1983; Travis and Bowser
1991 and references therein). MTs are hollow cylinders
typically composed of 13 protofilaments that are linear
ß The Author 2013. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please
e-mail: [email protected]
Mol. Biol. Evol. 30(11):2487–2493 doi:10.1093/molbev/mst150 Advance Access publication September 4, 2013
2487
Article
Foraminifera (forams) and radiolarians are primarily testate or
skeletonized amoeboid protists that possess characteristic
pseudopodia. Both are widely distributed in polar, subtropical,
and tropical oceans, and forams have also been found in
freshwater and terrestrial habitats. They are major groups of
microfossils and important players in aquatic ecosystems as
primary producers, grazers, and carbon exporters. Molecular
phylogeny supports the hypothesis that forams and two main
groups of radiolarians, the strontium sulfate skeleton-bearing
Acantharea and the opaline silica-bearing Polycystinea, form
the monophyletic clade Retaria (Ishitani et al. 2011; Sierra
et al. 2013). Together with the extraordinarily diverse
Cercozoa, they comprise one of the six supergroups of eukaryotes, the Rhizaria (Moreira et al. 2007; Ishitani et al. 2011).
Recent single and multigene phylogenetic studies suggest that
the third traditional group of radiolarians, the Phaeodarea,
are members of Cercozoa but not of the Retaria (Nikolaev
et al. 2004; Yuasa et al. 2005; Sierra et al. 2013). We use the
term “radiolarians” as Acantharea and Polycystinea herein.
A characteristic structural feature of Rhizaria is their
pseudopodia. Forams possess reticulopodia, which are
Hou et al. . doi:10.1093/molbev/mst150
assemblies of a- and b-tubulin (a-b) heterodimers. They
constitute one of the major components of the cytoskeleton and are required for many essential processes in
eukaryotic cells, such as maintaining cell structure, intracellular transport, forming the mitotic spindle, and ciliary and
flagellar motility. Remarkably, in the single-chambered
foram Allogromia spp. and the freshwater naked foram
Reticulomyxa filosa, it has been found that reticulopodial
MT can transform into a unique helical filament (HF)
form—continuous ribbons of laterally connected a-b
dimers (Hauser and Schwab 1974; McGee-Russell 1974;
Rupp et al. 1986; Hauser et al. 1989; Brueker and Hauser
1997; Welnhofer and Travis 1998).
A highly unusual “Type 2” b-tubulin (Linder et al. 1997;
Habura et al. 2005), referred here as b2, has been found
co-expressed with the canonical type b-tubulin (b1) in
R. filosa (Linder et al. 1997). Recently, b2 was identified
in several other foram genomes (Habura et al. 2005), as
well as four radiolarians (one acantharean and three polycystine species) (Burki et al. 2010; Ishitani et al. 2011).
Compared with canonical b1, foram b2 is characterized
by a unique insertion in the H1-B2 loop and many unusual
substitutions in the typically conserved M loop as well as
other domains (Habura et al. 2005). It was hypothesized
that the unusual substitutions in b2 domains strengthen
the lateral contacts, relative to longitudinal interactions,
between adjacent a-b dimers, effecting the rapid transition
from conventionally understood MT structure to HF
(Habura et al. 2005).
In this study, we analyzed the multicopy b-tubulin sequences obtained from transcriptomic data of Foraminifera,
radiolarians, and Cercozoa, and other rhizarian sequences
available in GenBank, and studied the unique evolution of
b1 and b2 paralogs in retarian protists. We detected positive
selection in the b2 lineage and explored its significance in
neofunctionalization based on retarian MT and HF structural
models. Our results will lead to a better understanding of the
molecular basis of the retarian MT and HF assembly/disassembly processes.
Results
Rhizarian b-Tubulin Transcript Sequences
We report 34 new b-tubulin transcript sequences in six
forams (Ammonia sp., Brizalina sp., Bulimina marginata,
Globobulimina turgida, Nonionellina sp., and R. filosa)
(KF170429–KF170462), 79 in four radiolarians (Amphilonche
elongate, Astrolonche serrata, Collozoum sp., and Phyllostaurus
siculus) (KF170463–KF170541), and 6 in one phaeodarean
(Aulacantha
scolymantha)
(KF170542–KF170547).
Reticulomyxa filosa and G. turgida sequences are contigs of
Sanger sequences and the others are isotigs of 454 pyrosequencing data.
Rhizarian b-Tubulin Phylogeny
Rhizarian b-tubulin protein phylogenetic analysis (fig. 1)
shows a clear distinction between the b1-tubulin found in
seven forams and six radiolarians (two acantharean and four
2488
MBE
polycystine species), and the b2 in 21 forams and seven
radiolarians (three acantharean and four polycystine
species). Within a single species, b1 and b2 both occurred
in multiple polymorphic copies (except for Collozoum sp. b2
KF170531). The highest copy number (47 b1 and 13 b2
unigene copies) was found in Ast. serrata, in which the
highest total number of reads was sequenced (337,982
assembled reads) (Sierra et al. 2013). These unigene copy
numbers might be slightly overestimated due to potential
sequencing and assembly errors in next generation transcriptomic sequencing. The b1 in the major groups within
Rhizaria (i.e., Foraminifera, radiolarians, and Cercozoa)
appear to be polyphyletic, and few of the nodes above
species level are strongly supported by bootstrap analysis.
The global b-tubulin phylogeny including other groups of
eukaryotes shows that even rhizarian b1 is a polyphyletic
clade (supplementary fig. S1, Supplementary Material
online). However, the b2 sequences in Foraminifera,
Acantharea, and Polycystinea all appear in well-supported
monophyletic clades (95–98%). B2-tubulin was not found in
the phaeodarean Aul. scolymantha, Gromia sp., or other
cercozoan data sets. Two hypothetical b-tubulin sequences
derived from the acantharean P. siculus (KF170534 and
KF170535) branch between the base positions of the b1
and b2 clades (supplementary fig. S1, Supplementary
Material online). They do not contain most of the signature
substitutions found in foram b2 (as described in Habura
et al. 2005) and are only 52.9–53.7% and 53.4–54.2% identical to P. siculus b1 and b2, respectively, and thus were
excluded in this rhizarian b1 and b2 analysis.
Retarian b2-Tubulin Is Under Positive Selection
An excess of nonsynonymous substitutions relative to synonymous substitutions (dn/ds or ! > 1) in the b2 lineage is an
important indicator of positive selection (i.e., new advantageous genetic variants were preferred at certain regions).
We analyzed rhizarian b-tubulin sequences using codonsubstitution models in the program PAML4.5 (Yang 2007)
and detected that foram and radiolarian b2 lineages both
contain positively selected sites (P < 0.01). These analyses
used the alignment of all rhizarian sequences and alignments
with fewer highly similar sequences (mainly intraspecific
sequences) reduced to 95% and 90% identity levels. The likelihood ratio test (LRT) statistics were consistently significant
among multiple analyses (LRT = 8.87–8.65 and 11.07–14.15
for foram and radiolarian b2, respectively; df = 1). The positively selected amino acid sites detected on b2 proteins
(especially those with highest probabilities) were largely similar among multiple analyses (supplementary file S1,
Supplementary Material online). Interestingly, foram and
radiolarian b2 share five selected sites, but the sites with
highest probabilities are different and unique to their own
group: 184 Val in foram b2 (based on R. filosa b2 CAA65332)
and 179 Val in radiolarians (based on Sphaerozoum punctatum b2 BAK61745). These two most favored substitutions
occupy different positions in the alignments (eight amino
acids apart) and both replaced a Thr in bovine b-tubulin or
retarian b1-tubulin with a Val.
MBE
b-tubulin Neofunctionalization in Retaria . doi:10.1093/molbev/mst150
98/98/100
Larcopyle butschlii BAK61734
Collozoum sp. KF170531
Sphaerozoum punctatum BAK61745
85/78/100
Collozoum amoeboides BAK61739
90/90/99
66/./.Sphaerozoum punctatum BAK61746
100/100/100
Amphilonche elongate KF170463
Amphilonche elongate KF170465
100/100/100
100/100/100 Phyllostaurus siculus KF170532
59/34/53
97/94/100
100/100/100
Phyllostaurus siculus KF170541
81/80/. Phyllostaurus siculus KF170533
Astrolonche serrata KF170518
97/96/100
Astrolonche serrata KF170520
Astrolonche serrata KF170525
Astrolonche serrata KF170514
Astrolonche serrata KF170524
Astrolonche serrata KF170522
Astrolonche serrata KF170526
(Acantharean+Polycystine)
Astrolonche serrata KF170521
Astrolonche serrata KF170519
Astrolonche serrata KF170523
Astrolonche serrata KF170516
Astrammina rara AY818719
Crithionina delacai AAY58147
Allogromia laticollaris CAA76100
Allogromia sp. AAH2005 AY818726
Allogromia sp. NF AAY58155
100/100/100 Rhabdammina cornuta AAY58152
Rhabdammina cornuta AAY58153
100/100/100 Reticulomyxa filosa CAA65332
Reticulomyxa filosa AAY58150
78/82/. Reticulomyxa filosa KF170462
Reticulomyxa filosa AAY58149
Miliammina fusca AAY58154
79/84/100
Pyrgo peruviana AAY58148
Cornuspira antarctica AAY58151
100/ Cornuspira antarctica AY818720
100/100
Nonionellina sp. KF170456
74/73/96
Nonionellina sp. KF170454
Nonionellina sp. KF170455
71/71/98
Nonionella labradorica FR821703
100/99/100
98/94/.
Hyalinea balthica FR821697
Globobulimina turgida KF170449
Globobulimina turgida FR821699
66/60/88
Brizalina sp. KF170438
92/82/87
Brizalina sp. KF170437
Brizalina sp. KF170439
78/52/74
100/100/100
Brizalina sp. KF170440
55/42/. Brizalina sp. KF170442
81/70/99
Brizalina sp. KF170441
Rosalina sp. FR821698
Ammonia sp. FR821702
Planoglabratella opercularis BAE48508
Elphidium williamsoni FR821701
Haynesina germanica FR821700
55/49/59
100/97/100 Ammonia sp. KF170434
Radiolarian β2
46/56/100
95/98/100
65/65/100
66/62/100
Foraminiferan β2
Reticulomyxa filosa KF170458
Reticulomyxa filosa KF170460
Foraminifera
Reticulomyxa filosa CAA65331
Reticulomyxa filosa KF170459
Gromia sp. Antarctica ADK98515
Bulimina marginata KF170443
56/49/82
Globobulimina turgida KF170447
Ammonia sp. KF170429
Ammonia sp. KF170432
Foraminifera
75/61/97 Brizalina sp. KF170435
65/61/76
Globobulimina turgida KF170446
Nonionellina sp. KF170451
Nonionellina sp. KF170450
Corallomyxa tenera ABP93409
57/./77
Larcopyle butschlii BAK61733
Sphaerozoum punctatum BAK61747
Polycystinea
Collozoum amoeboides BAK61740
Collozoum sp. KF170529
Arachnula sp. ATCC50593 ACA04817
Bulimina marginata KF170444
Nonionellina sp. KF170452
Globobulimina turgida KF170448
Foraminifera
Ammonia sp. KF170433
67/64/99
Ammonia sp. KF170430
68/60/99 Ovammina opaca ADK98509
Reticulomyxa filosa KF170461
94/92/100
Aulacantha scolymantha KF170542
Aulacantha scolymantha KF170544
Phyllostaurus siculus KF170540
Astrolonche serrata KF170490
71/67/99
Astrolonche serrata KF170491
Astrolonche serrata KF170487
Acantharea
Astrolonche serrata KF170498
Astrolonche serrata KF170496
Astrolonche
serrata
KF170509
1
Astrolonche serrata KF170506
Plasmodiophora brassicae CAM98704
Capsellina sp. ATCC50039 ADK90069
100/100/100
Massisteria marina ADK90066
Aulacantha scolymantha KF170547
77/72/66
97/98/100 Aulacantha scolymantha KF170545
71/58/99
Bigelowiella natans CCMP621 AAC68509
Bigelowiella natans CCMP621 AAC68506
99/100
/100
Bigelowiella natans CCMP621 AAC68508
51/22/.
68/91/. Bigelowiella natans ABX25952
Cercomonas sp. ATCCPRA21 ABX25975
Spongomonas sp. ATCC50405 ADK90063
Cholamonas cyrtodiopsidis ATCC50325 ACR46922
Thaumatomonas seravini ABX25978
Cercomonas sp. ATCC50316 AAD55354
Heteromita globosa ADK90062
76/64/54
Bodomorpha minima ADK90061
Massisteria marina ADK90067
Massisteria marina ADK90065
Proleptomonas faecicola ADK90059
Gymnophrys sp. ATCC50923 ADK90058
60/57/81 Gymnophrys sp. ATCC50923ADK90060
90/91/99
88/82/100
100/98/100
Rhizarian β1
0.1
FIG. 1. Rhizarian b-tubulin phylogeny. The phylogenetic tree was obtained as the best maxiumum likelihood (ML) tree using LG + G model in RAxML
based on rhizarian b-tubulin protein alignment (112 sequences) using rhizarian b1-tubulin as the outgroup. Numbers on branches represent the
following: (1) 1,000 times bootstrap support values for nodes on this tree, (2) 100 times bootstrap values for the corresponding nodes on the PhyML ML
tree (LG + G + I model), and (3) the Bayesian posterior probabilities on the Bayesian tree (WAG + G + I model). Solid circles mark nodes with >50%
RAxML bootstrap values above species level. New sequences published in this study are shown in bold. Nucleotide accession numbers are shown in the
taxon labels when protein accession numbers are unavailable.
2489
Hou et al. . doi:10.1093/molbev/mst150
MBE
FIG. 2. Models for basic repeating units of R. filosa MTs and HFs. (A) A representative a- and b2-tubulin hexamer of a 13-3 MT. (B) A representative aand b2-tubulin tetramer of HF. The tubulin monomers are shown in cartoon format, with a-tubulin in blue and b2-tubulin in green. GTP, GDP, and
positively selected residues Val184, Asp297, and Ser302 are shown in sphere format. As indicated in the figure, GTP is shown in orange, GDP in yellow,
Val184 in red, Asp297 in purple, and Ser302 in ice blue.
Homology Model for the Foram Tubulin Hexamer
Longitudinal and lateral a- and b-tubulin dimer–dimer interactions in MTs in the foram R. filosa can be examined using a
hexamer homology model (fig. 2A), which was generated
using atomic models for bovine MTs based on a published
8 Å cryo-electron microscopy density map (Sui and Downing
2010). Three foram positively selected residues that are likely
implicated in dimer–dimer interactions, 184Val, 297Asp, and
302Ser, are highlighted on the MT hexamer and HF tetramer
structures (fig. 2A and B). The residue 184Val is adjacent to
two dimer–dimer interaction sites (185Val and 187Glu) and
two GTP-binding sites (186Val and 187Glu) annotated in the
NCBI Conserved Domain Database (PSSMID 100016). The
substitutions at 297Asp and 302Ser substantially alter side
chain chemical properties of these residues and are also
within 5 Å of neighboring protofilaments.
Discussion
This study shows the first convincing evidence that both
canonical b1-tubulin and highly divergent b2-tubulin are
preserved in forams and radiolarians. Since their discovery
in the foram R. filosa (Linder et al. 1997), two b-tubulin paralogs have also been found in three radiolarians (Ishitani et al.
2011), but only b2 have been recovered in other forams
(Habura et al. 2005). In this study, 21 foram b1 sequences
were revealed by whole transcriptomic sequencing of five
Rotaliida species and R. filosa, confirming that b1 had not
been lost in these forams and likely other foram lineages as
well. The global b-tubulin alignment confirms that foram and
radiolarian b1 are canonical eukaryotic b-tubulin sequences,
whereas their b2 are dramatically modified in many structurally important regions (e.g., the loops analyzed in Habura et al.
2005). In most of the modified regions (e.g., H9 and H9-B8),
radiolarian b2 share some of the signature foram-type
2490
substitutions, whereas radiolarian b2 were almost uniquely
modified in some regions (e.g., the M loop). This indicates
that radiolarian b2 might carry different or additional functionality for this group of protists. Our investigation failed to
identify any noncanonical b-tubulin in non-retarians, even
among other members of the Rhizaria. For example, Aul.
scolymantha is a member of Phaeodarea, a group previously
classified among radiolarians, but shown to be related to
Cercozoa by molecular evidence (Nikolaev et al. 2004; Sierra
et al. 2013). In this species, we only found six b1 sequences
and no b2 among over 65,000 assembled contigs. The
absence of b2 in Aul. scolymantha and other cercozoans suggests that b2 is a synapomorphic feature of Retaria. Given the
lack of candidate homologous b2 sequences in our taxon-rich
b-tubulin data set, the b-tubulin duplication event most likely
occurred within Rhizaria, before the divergence of forams and
radiolarians.
The b1-tubulin protein is poor in resolving relationships
within Rhizaria. The well-defined rhizarian groups,
Foraminifera, radiolarians, and Cercozoa, appear to be polyphyletic, and even some sequences from the same species are
polyphyletic (e.g., G. turgida, Nonionellina sp., Aul. scolymantha, Massisteria marina, Collozoum sp.). Because rhizarians are predominantly composed of yet uncultivable species,
their genomic studies rely on field-collected specimens that
may contain other coexisting organisms, e.g., prey or symbionts. Therefore, we do not exclude the possibility that some
of the polyphyletic b1 sequences in the current and previous
studies might actually come from the foreign genomes in the
samples. We attempted to build a global b-tubulin tree with
taxonomically broad groups of eukaryotes to identify foreign
sequences (supplementary fig. S1, Supplementary Material
online). However, the rhizarian b1 clade is also highly polyphyletic in this global b-tubulin phylogeny, and therefore
b-tubulin Neofunctionalization in Retaria . doi:10.1093/molbev/mst150
none of the sequences can be identified as a “contaminant”
by this approach. On the other hand, the b2-tubulin phylogeny is consistent with recent multigene phylogenies showing
the monophyly of Retaria (Burki et al. 2010; Sierra et al. 2013).
Interestingly, two divergent P. siculus sequences group neither
within b1 nor other b2 sequences (supplementary fig. S1,
Supplementary Material online), lack the characteristic
substitutions of a foram or radiolarian b2 sequence, and
possibly represent yet another variant of b-tubulin.
Therefore, it is unclear whether the rhizarian b-tubulin phylogeny supports the hypothesis that forams are derived
radiolarians, as that advanced in recent publications
(Cavalier-Smith 2003; Moreira et al. 2007; Krabberød et al.
2011; Sierra et al. 2013).
Both b1 and b2 paralogs are preserved and expressed in
Retaria, presumably due to their essential functions. Recent
cell biology studies have extended our knowledge of the
tubulin superfamily and support the “multi-tubulin
hypothesis”—each tubulin isotype, either post-translationally
modified or gene-encoded, performs a subset of roles in MT
function (McKean et al. 2001). For example, in the ciliate
Tetrahymena thermophila, it has been shown that a family
of distinct b-tubulin isotypes was used to construct subsets of
functionally different MTs (Pucciarelli et al. 2012). In this
study, by using codon-based likelihood models, we provide
molecular evidence for the positive selection in foram and
radiolarian b2 lineages, which suggests that the sequence and
structure differences of this paralog are required to carry out
unique functionality (e.g., MT–HF dynamics) for retarian protists. Negative-stain electron microscopy shows that foram HF
is composed of laterally connected a-b heterodimers (Golz
and Hauser 1986; Welnhofer and Travis 1998). Unlike the
classical MT disassembly model, which involves lateral dissociation of intact protofilaments, the transition from MT to HF
requires the internal longitudinal disassociation of the protofilaments, with the lateral connection between a-b dimers
being maintained. Here, we find that for the most favored
substitution in foram b2 (at position 184), the Thr hydroxyl in
a typical eukaryote is replaced with the Val methyl group
(fig. 2A). This replacement possibly disrupts the hydrogen
bonding that strengthens the longitudinal interaction with
the next dimer on the protofilament. We also detected two
positively selected substitutions within close contact (i.e.,
within 5 Å) of neighboring protofilaments (297Asp and
302Ser), which could alter lateral dimer interactions.
Because radiolarian b2 contains unique positively selected
residues, b2 may be responsible for other essential functions
in these protists (e.g., axopodial or rhizopodial motility).
Full-length radiolarian a- and b2-tubulin sequences are
not available for homology modeling of radiolarian MT structure, thus we could not predict what roles their positively
selected residues may play in MT function. We speculate
that the most-favored radiolarian substitution, 179Val (in
S. punctatum b2 BAK61745), plays a similar role as 184Val
does in foram b2, because the Thr in eukaryotic b1-tubulin
were replaced with the Val in both cases. Unfortunately, the
MT to HF transition has not yet been demonstrated in
radiolarians.
MBE
In summary, foram and radiolarian b-tubulins have been
subjected to neofunctionalization since their duplication
before the divergence of Retaria. Compared with the canonical eukaryotic b1-tubulin, the b2 paralog is a synapomorphic
feature of Retaria and likely carries out essential functions. In
forams, b2 may be responsible for rapid MT assembly/disassembly (Travis and Bowser 1991), as well as MT protein packaging into a readily transported form, HF paracystals (Koury
et al. 1985; Rupp et al. 1986). HF formation may allow more
rapid extension and retraction of pseudopodial networks,
which is a highly advantageous feature of forams that permits
exploration and survival in even some “extreme” environments (e.g., polar oceans). Future studies are needed to demonstrate that the MT–HF transition occurs in radiolarians, to
confirm and extend the previous observation (Linder et al.
1997) that b1 and b2 are co-localized in other retarian MTs,
and to determine their stoichiometry. In addition, in situ
mutagenesis experiments for the positively selected sites as
well as high-resolution cryo-electron microscopy structure
determination of MT and HF will help complete our understanding of MT dynamics in retarian protists.
Materials and Methods
Rhizarian b-Tubulin Transcript Sequencing
We recently sequenced the transcriptomes of six forams
(Ammonia sp., Brizalina sp., B. marginata, G. turgida,
Nonionellina sp., and R. filosa), three species of Acantharea
(Amp. elongate, Ast. serrata, P. siculus), one collodarian
(Collozoum sp.), and one phaeodarean (Aul. scolymantha).
Except for R. filosa, all specimens are unavailable in culture
and were collected from the field. Reticulomyxa filosa and G.
turgida transcriptomes were sequenced by the Sanger
method and the others by 454 Genome Sequencer FLX
with GS-FLX Titanium reagents (Roche). Specimen collection,
transcriptomic library constructions, and sequence assemblies
were described in Sierra et al. (2013). The new b-tubulin
sequences were identified by homology using the BlastX
algorithm (Altschul et al. 1990) against the transcriptomic
sequences.
Phylogenetic Analyses
All rhizarian b-tubulin sequences available in GenBank before
1 January 2013 and the new sequences reported in this study
were aligned using MUSCLE (Edgar 2004). Redundant protein
sequences were excluded, whereas their corresponding DNA
sequences were used in the codon alignment for the positive
selection analyses described later. The alignment contained
112 sequences and 275 amino acid positions. The maximum
likelihood (ML) analyses were performed using RAxML v.7.4.2
(Stamatakis 2006) with the PROTGAMMALG implementation in multiple inferences and 50 randomized parsimony
starting trees. Statistical support was evaluated with the
PROTGAMMALG setting 1,000 bootstrap replicates. The
ML phylogeny was also constructed using PhyML 3.0
(Guindon et al. 2010) LG model with a gamma-shaped distribution of the substitution rates (G) and a proportion of
invariable sites (I), and 10 random starting trees, followed by
2491
MBE
Hou et al. . doi:10.1093/molbev/mst150
100 times bootstrap analyses. This substitution model was
chosen based on the BIC criterion in ProtTest 3 (Darriba
et al. 2011). The Bayesian analyses were performed with the
WAG + G + I model using MrBayes 3.2 (Ronquist and
Huelsenbeck 2003) with one cold and two heated chains
for 5,000,000 iterations. Trees were sampled every 500 generations from the last 3,750,000 iterations (well after the chains
reached stationarity), and 7,500 trees were used for inferring
Bayesian posterior probability.
Detection of Positive Selection in Retarian b2-Tubulin
We used the branch-site model in the program CODEML in
PAML4.5 (Yang 2007) to test for positive selection on b2tubulin in the foreground lineage (forams or radiolarians)
versus the background lineage (rhizarian b1). This model
allowed ! (dn/ds ratio) to vary among sites and across lineages. Rhizarian b-tubulin codon sequence alignment data
set (175 sequences and 825 nucleotides in length) was used.
To understand how intra- and interspecific variations
contribute to the model results, two smaller data sets containing 118 and 88 sequences were also analyzed, which were
generated by reducing the highly similar (mainly intraspecific)
sequences at 95% and 90% identity levels, respectively, using
the progam CD-HIT (Huang et al. 2010). The phylogenetic
trees used in the analyses were constructed using GARLI 2.0
(Zwickl 2006) under the best-fit ML substitution model
GTR + G + I. The alternative model was the implemented
model A (model = 2 and NS sites = 2) with ! > 1 estimated
(fix_omega = 0 and omega > 1) for certain sites in b2. The
null model was model A with ! = 1 fixed (fix_omega = 1 and
omega = 1). Two models were compared by LRTs and the
significant test statistics indicate that b2-tubulin contains
one or more than one positively selected sites. The positively
selected sites were suggested by the Naive Empirical Bayes
(NEB) and Bayes Empirical Bayes (BEB) analyses implemented
in CODEML (Yang et al. 2005).
Foram MT Hexamer Homology Modeling
Protein sequences of R. filosa a- and b2-tubulin (CAA65329
and CAA65332) were aligned to bovine a- and b-tubulin
using ClustalX (Thompson et al. 1997), respectively. A homology model of R. filosa a-b2 dimer was generated based on a
template bovine a-b dimer structure (PDB ID: 1JFF; Löwe
et al. 2001) using the program MODELLER (Fiser and Sali
2003) with the DOPE and GA341 energy functions. The modeled dimer structure was then overlaid onto three neighboring dimers in bovine 13-3 MT structure derived from an 8 Å
cryo-electron microscopy map (Sui and Downing 2010) to
obtain an atomic resolution model for the basic R. filosa MT
hexamer. The detailed procedure for generation of the HF
tetramer is described in another manuscript (Bassen D, Hou
Y, Bowser SS, Banavali NK, in preparation). Briefly, R. filosa HF
models were generated from MT using a two center-of-mass
coarse-grained description of the a-b2 dimer. An assumption
that the orientation of the dimer with respect to the central
axis of the MT is maintained in the HF was used along with an
2492
intermediate coarse-grained description of the a-b2 dimer to
obtain an atomic detail HF model through simple overlay.
Supplementary Material
Supplementary file S1 and figure S1 are available at
Molecular Biology and Evolution online (http://www.mbe.
oxfordjournals.org/).
Acknowledgments
The authors thank Haixin Sui for insightful discussions regarding MT structure and function, Ingrid Hahn for helpful
comments on the manuscript, and Leila Atallahbenson for
proofreading the taxon labels. They also thank the Vital-IT
Center for high-performance computing of the Swiss Institute
of Bioinformatics (http://www.vital-it.ch, last accessed
September 18, 2013) for carrying out parts of the phylogenetic
analyses. This work was supported by the United States
National Science Foundation grant ANT 0944646 (to S.S.B.)
and the REU Site award DBI 1062963 (to S.S.B.), and the Swiss
National Science Foundation grant No. 31003A_140766 (to
R.S. and J.P.).
References
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local
alignment search tool. J Mol Biol. 215:403–410.
Anderson OR. 1983. Radiolaria. New York: Springer-Verlag.
Bowser SS, Travis JL. 2002. Reticulopodia: structural and behavioral basis
for the suprageneric placement of granuloreticulosan protists.
J Foramin Res. 32:440–447.
Brueker RN, Hauser M. 1997. Direct transformation of microtubules
into helical filaments by low salt concentrations in Reticulomyxa.
Eur J Cell Biol. 72(Suppl 43): 177.
Burki F, Kudryavtsev A, Matz MV, Aglyamova GV, Bulman S, Fiers M,
Keeling PJ, Pawlowski J. 2010. Evolution of Rhizaria: new insights
from phylogenomic analysis of uncultivated protists. BMC Evol
Biol. 10:377.
Cavalier-Smith T. 2003. Protist phylogeny and the high-level classification of Protozoa. Eur J Protistol. 39:338–348.
Darriba D, Taboada GL, Doallo R, Posada D. 2011. ProtTest 3: fast selection of best-fit models of protein evolution. Bioinformatics 27:
1164–1165.
Edgar RC. 2004. MUSCLE: multiple sequence alignment with high
accuracy and high throughput. Nucleic Acids Res. 32:1792–1797.
Fiser A, Sali A. 2003. Modeller: generation and refinement of homologybased protein structure models. Methods Enzymol. 374:461–491.
Golz R, Hauser M. 1986. Polymorphic assembly states of Allogromia
tubulin under physiological conditions. Eur J Cell Biol. 40:124–129.
Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O.
2010. New algorithms and methods to estimate maximumlikelihood phylogenies: assessing the performance of PhyML 3.0.
Syst Biol. 59:307–321.
Habura A, Wegener L, Travis JL, Bowser SS. 2005. Structural and functional implications of an unusual foraminiferal b-tubulin. Mol Biol
Evol. 22:2000–2009.
Hauser M, Schwab D. 1974. Microtubules and helical microfilaments in
the cytoplasm of the foraminifer Allogromia laticollaris, arnold.
Investigations with vinblastine and deuteriumoxide to demonstrate
a close interrelationship. Cytobiologie 9:263–279.
Hauser M, Lindernblatt J, Hülsmann N. 1989. Cytoskeleton of
Reticulomyxa filosa reticulopodia contains Glu-tubulin as a main
component. Eur J Protistol. 25:145–157.
Hausmann K, Hülsmann N, Radek R. 2003. Protistology. 3rd ed. Stuttgart
(Germany): E. Schweizerbart’sche Verlagsbuchhandlung.
b-tubulin Neofunctionalization in Retaria . doi:10.1093/molbev/mst150
Huang Y, Niu B, Gao Y, Fu L, Li W. 2010. CD-HIT Suite: a web server for
clustering and comparing biological sequences. Bioinformatics 26:
680–682.
Ishitani Y, Ishikawa SA, Inagaki Y, Tsuchiya M, Takahashi K, Takishita K.
2011. Multigene phylogenetic analyses including diverse radiolarian
species support the “Retaria” hypothesis—the sister relationship of
Radiolaria and Foraminifera. Mar Micropaleontol. 81:32–42.
Koury ST, Bowser SS, McGee-Russell SM. 1985. Ultrastructural changes
during reticulopod withdrawal in the foraminiferan protozoan
Allogromia sp., strain NF. Protoplasma 129:149–156.
Krabberød AK, Bråte J, Dolven JK, Ose RF, Klaveness D, Kristensen T,
Bjørklund KR, Shalchian-Tabrizi K. 2011. Radiolaria divided into
Polycystina and Spasmaria in combined 18S and 28S rDNA phylogeny. PLoS One 6:e23526.
Linder S, Schliwa M, Kube-Granderath E. 1997. Sequence analysis and
immunofluorescence study of a- and b-tubulins in Reticulomyxa
filosa: implications of the high degree of b2-tubulin divergence.
Cell Motil Cytoskeleton. 36:164–178.
Löwe J, Li H, Downing KH, Nogales E. 2001. Refined structure of
ab-tubulin at 3.5 Å resolution. J Mol Biol. 313:1045–1057.
McGee-Russell SM. 1974. Dynamic activities and labile microtubules in
cytoplasmic transport in the marine foraminiferan, Allogromia.
Symp Soc Exp Biol. 28:157–189.
McKean PG, Vaughan S, Gull K. 2001. The extended tubulin superfamily.
J Cell Sci. 114:2723–2733.
Moreira D, von der Heyden S, Bass D, López-Garcı́a P, Chao E,
Cavalier-Smith T. 2007. Global eukaryote phylogeny: combined
small- and large-subunit ribosomal DNA trees support
monophyly of Rhizaria, Retaria and Excavata. Mol Phylogenet Evol.
44:255–266.
Nikolaev SI, Berny C, Fahrni J, Bolivar I, Polet S, Myl’nikov AP, Aleshin VV,
Petrov NB, Pawlowski J. 2004. The twilight of Heliozoa and rise of
Rhizaria, an emerging supergroup of amoeboid eukaryotes. Proc Natl
Acad Sci U S A. 101:8066–8071.
Pucciarelli S, Ballarini P, Sparvoli S, Barchetta S, Yu T, Detrich HW 3rd,
Miceli C. 2012. Distinct functional roles of b-tubulin isotypes in
microtubule arrays of Tetrahymena thermophila, a model singlecelled organism. PLoS One 7:e39694.
Ronquist F, Huelsenbeck JP. 2003. MRBAYES 3: Bayesian phylogenetic
inference under mixed models. Bioinformatics 19:1572–1574.
MBE
Rupp G, Bowser SS, Mannella C, Rieder C. 1986. Naturally occurring
tubulin-containing paracrystals in Allogromia: immunocytochemical
identification and functional significance. Cell Motil Cytoskeleton. 6:
363–375.
Sierra R, Matz MV, Aglyamova G, Pillet L, Decelle J, Not F, de Vargas C,
Pawlowski J. 2013. Deep relationships of Rhizaria revealed by
phylogenomics: a farewell to Haeckel’s Radiolaria. Mol Phylogenet
Evol. 67:53–59.
Stamatakis A. 2006. RAxML-VI-HPC: maximum likelihood-based
phylogenetic analyses with thousands of taxa and mixed models.
Bioinformatics 22:2688–2690.
Suzuki N, Aita Y. 2011. Radiolaria: achievements and unresolved issues:
taxonomy and cytology. Plankton Benthos Res. 6:69–91.
Sui H, Downing KH. 2010. Structural basis of interprotofilament
interaction and lateral deformation of microtubules. Structure 18:
1022–1031.
Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. 1997.
The CLUSTAL_X windows interface: flexible strategies for multiple
sequence alignment aided by quality analysis tools. Nucleic Acids Res.
25:4876–4882.
Travis JT, Bowser SS. 1991. The motility of foraminifera. In: Lee JJ,
Anderson OR, editors. Biology of Foraminifera. London: Academic
Press. p. 91–155.
Welnhofer EP, Travis JL. 1998. Evidence for a direct conversion between
two tubulin polymers—microtubules and helical filaments—in the
foraminiferan, Allogromia laticollaris. Cell Motil Cytoskeleton. 41:
107–116.
Yang Z. 2007. PAML 4: a program package for phylogenetic analysis by
maximum likelihood. Mol Biol Evol. 24:1586–1591.
Yang Z, Wong WSW, Nielsen R. 2005. Bayes empirical Bayes inference
of amino acid sites under positive selection. Mol Biol Evol. 22:
1107–1118.
Yuasa T, Takahashi O, Honda D, Mayama S. 2005. Phylogenetic
analyses of the polycystine Radiolaria based on the 18S rDNA
sequences of the Spumellarida and the Nassellarida. Eur J Protistol.
41:287–298.
Zwickl DJ. 2006. Genetic algorithm approaches for the phylogenetic
analysis of large biological sequence datasets under the maximum
likelihood criterion [Ph.D. dissertation]. [Austin (TX)]: The
University of Texas at Austin.
2493