Characterization of Microtubule-Binding and Dimerization Activity of Giardia lamblia End-Binding 1 Protein

End-binding 1 (EB1) proteins are evolutionarily conserved components of microtubule (MT) plus-end tracking protein that regulate MT dynamics. Giardia lamblia, with two nuclei and cytoskeletal structures, requires accurate MT distribution for division. In this study, we show that a single EB1 homolog gene of G. lamblia regulates MT dynamics in mitosis. The haemagglutinin-tagged G. lamblia EB1 (GlEB1) localizes to the nuclear envelopes and median bodies, and is transiently present in mitotic spindles of dividing cells. Knockdown of GlEB1 expression using the morpholinos-based anti-EB1 oligonucleotides, resulted in a significant defect in mitosis of Giardia trophozoites. The MT-binding assays using recombinant GlEB1 (rGlEB1) proteins demonstrated that rGlEB1102–238, but not rGlEB11–184, maintains an MT-binding ability comparable with that of the full length protein, rGlEB11–238. Size exclusion chromatography showed that rGlEB1 is present as a dimer formed by its C-terminal domain and a disulfide bond. In vitro-mutagenesis of GlEB1 indicated that an intermolecular disulfide bond is made between cysteine #13 of the two monomers. Complementation assay using the BIM1 knockout mutant yeast, the yeast homolog of mammalian EB1, indicated that expression of the C13S mutant GlEB1 protein cannot rescue the mitotic defect of the BIM1 mutant yeast. These results suggest that dimerization of GlEB1 via the 13th cysteine residues plays a role during mitosis in Giardia.


Introduction
Microtubules (MTs) are dynamic polymers involved in mitosis, organelle biogenesis, and intracellular transport [1]. Growth and catastrophe of MTs are mediated by MT-associated proteins including plus-end tracking proteins (+TIPs) which preferentially bind to the growing ends of polymerized MTs in vivo [2]. Among the +TIPs, end-binding 1 (EB1) proteins are the main regulator of the protein complexes at the plus end, as shown in mammalian EB1 [3] and yeast EB1 homologs, Mal3p [4] and Bim1p [5].
The proteins belonging to the EB1 family are composed of an N-terminal calponin homology (CH) domain, a composite domain consisting of a a-helically coiled-coli and a unique EB1 homology region, and the C-terminal acidic tail ending with an EEY/F motif [6]. The CH domain is reported to be involved in MT binding [4], whereas the EB1domain play a role in dimerization as well as binding to the EB1-interacting proteins [7]. The EEY/F motif serves as a binding site for p150 Glued in mammals [8]. The EB1 protein is one of highly conserved proteins among in all eukaryotes, including Giardia lamblia, a protozoan belonging to the earliest diverging eukaryotic lineage [9].
G. lamblia has two nuclei and cytoskeletal structures including an adhesive disc, a median body, and four pairs of flagella [10].
Observations using three-dimensional deconvolution and electron microscopies indicated that two extranuclear spindles move chromosomes laterally through a polar opening in the nuclear membrane during cell division of G. lamblia [11]. Using a dominant-negative mutant, kinesin-13 was found to control MT dynamics in G. lamblia [12]. A single open reading frame (ORF) encoding an EB1-homologous protein was found in the database of G. lamblia. Using GFP-tagged EB1, G. lamblia EB1 (GlEB1) was found at the flagellar tips and median bodies [12]. In addition, the role of GlEB1 was assessed by complementation assays using a BIM1 mutant of Saccharomyces cerevisiae, in which proper positioning of the nucleus is abolished [13]. Through co-immunoprecipitation and yeast two-hybrid assays, block of proliferation 1 and cytoskeletal proteins such as band c-giardins, were identified as EB1-interacting proteins [14], [15], [16]. However, it remains to be elucidated how GlEB1 modulates MT dynamics during the cell cycle of G. lamblia.
In this study, expression and intracellular localization of GlEB1 were examined in various stages using a transgenic G. lamblia expressing haemagglutinin (HA) epitope-tagged EB1. In addition, a biochemical characterization of GlEB1 was performed by defining the domains and an amino acid residue responsible for MT binding and dimerization.
To induce encystation in vitro, G. lamblia trophozoites were transferred into an encystation medium (TYI-S-33 medium, 10 mg/mL bovine bile, pH 7.8) [18]. At various time-points after the incubation in the encystation medium, the cells were harvested by centrifugation at 3000 rpm for 15 min at 4uC. To monitor the encystation process, intracellular level of CWP1 [19] was measured in the harvested cells.

Construction of G. lamblia expressing HA Epitope-tagged GlEB1
Plasmid pLop2 and pNLop2-GItetR were a gift from Dr. Jung-Hsiang Tai [20]. To generate an HA epitope tag to the C-terminal of the eb1 gene, a 950 bp DNA fragment containing the promoter and the full ORF of the eb1 gene was amplified from G. lamblia WB genomic DNA by PCR using two primers, eb1-NcoI-F and eb1-HA-R (Table 2). NcoI and EcoRI sites, located at the ends of the resultant eb1 DNA, were used for cloning into the corresponding site of plasmid pLop2, resulting in the plasmid pLop2-eb1-HA. A 950 bp NheI/SalI fragment of pLop2-eb1-HA was cloned into the plasmid pNLop2-GItetR to yield the plasmid pNLop2-eb1-HA-GItetR, in which GlEB1 is expressed as a fused protein in frame with an HA-epitope. All constructs were verified by DNA sequencing provided by a sequencing service company (Macrogen, Seoul, Korea).
The trophozoites were grown for 72 h in normal TYI-S-33 medium. Thirty micrograms of pNLop2-eb1-HA-GItetR were transformed into 1610 7 trophozoites by electroporation under the following conditions: 350 volts, 1000 mF, and 700 V (BioRad). Trophozoites harboring pNLop2-eb1-HA-GItetR were selected by adding G418 (Life Technologies) to the TYI-S-33 medium at a final concentration of 150 mg/mL, and after 4 to 5 days of cultivation, the resistant cells were transferred to and maintained in the medium containing 600 mg/mL of G418. As a control, trophozoites carrying pNLop2-GItetR were constructed as described above.

Western Blot Analysis
Cell extracts were prepared from G. lamblia containing pNLop2-GItetR, or pNLop2-eb1-HA-GItetR in a phosphate buffered saline (PBS: 137 mM NaCl, 2.7 mM KCl, 10.1 mM Na 2 HPO 4 , and 2 mM KH 2 PO 4 , pH 7.4), separated by SDS-PAGE, and transferred onto a polyvinylidenefluoride (PVDF) membrane (Millipore). The membrane was incubated with monoclonal Following incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies, the immunoreactive protein was visualized using an enhanced chemiluminescence (ECL) system (Amersham Pharmacia). Membranes were incubated in a stripping buffer (Thermo Scientific) at room temperature for 30 min, and then reacted with polyclonal rat antibodies specific to the a-tubulin of G. lamblia (1:10000) [21]. In the case of trophozoites with pNLop2-eb1-HA-GItetR, they were prepared under various cell cycle stage: without aphidicolin treatment, 6 h-aphidicolin treatment, or released from the aphidicolin treatment every hour up to 6 h. Intracellular levels of GlEB1 were monitored in these cells by Western blot analysis using anti-HA antibodies (1:2000). As a loading control, an amount of a-tubulin was also detected in these cell extracts using anti-Gla-tubulin antibodies (1:10000).

Immunofluorescence Assay (IFA)
To examine the localization of GlEB1 in G. lamblia expressing HA-tagged GlEB1, the cells were attached to glass slides coated with L-lysine in a humidified chamber. The attached cells were fixed with chilled 100% methanol at 220uC for 10 min, and permeabilized with PBS/0.5% Triton X-100 for 10 min. After a 1 h-incubation in blocking buffer (PBS, 5% goat serum, and 3% BSA), the cells were reacted overnight with rat anti-GlEB1 polyclonal antibodies (1:400) [13] and mouse anti-HA antibodies (1:50; Sigma). Following three 5 min-washes with PBS, the cells were incubated with AlexaFluor 555-conjugated anti-rat IgG To determine the intracellular location of GlEB1 in dividing cells, trophozoites were treated with aphidicolin for 6 h, and released for 3 h. They were treated with anti-HA antibodies (1:100) or anti-Glc-tubulin antibodies (1:100). In addition, encysting cells were also reacted with anti-HA antibodies to monitor GlEB1 localization during encystation. As a control to ensure the formation of cysts via an in vitro-encystation experiment, the encysting cells were also treated with anti-GlCWP1 antibodies (1:100) [22]. These cells were then reacted with tetramethylrhodamine isothiocyanate (TRITC)-conjugated anti-rat IgG antibodies (1:200; Jackson ImmunoResearch Lab).

Synchronization of Giardia Trophozoites with Aphidicolin and Flow Cytometry Analysis
Synchronization of Giardia trophozoites was performed as described previously, with some modifications [23]. To summarize, 2610 4 cells were cultured in TYI-S-33 medium for 24230 h until they reached 70% confluence. Aphidicolin (Sigma) was added to culture medium at a final concentration of 5 mg/mL, and then incubated for 6 h. The culture media was replaced with fresh medium, and cultured up to 6 h. Since the aphidicolin added to the culture was prepared in dimethyl sulfoxide (DMSO), control Giardia cultures were treated with 0.05% DMSO, showing that DMSO did not affect G. lamblia at that concentration. Synchronized cells were analyzed for their DNA content via flow cytometry, as described previously [24]. The DMSO-treated and aphidicolin-treated cells were placed on ice for 20 min, and harvested by 10 min centrifugation at 3000 rpm. The harvested

Knockdown of GlEB1 Expression Using Morpholinos
The expression of GlEB1 was decreased by knockdown experiment as described [25]. Briefly, 25-mer morpholinos was designed to target the first 25 bases open reading frame of the GlEB1 (Table 2; Gene Tools). Another morpholinos identical to the GlEB1 morpholinos except 5 mispairs ( Table 2) was used as a control. Lyophilized morpholinos were added to 5610 6 cells in 0.3 mL of medium at a final concentration of 100 mM. As a negative control, an equal volume of sterile water was added to the cells. After electroporation, cells were grown for 48 or 72 h, and then analyzed for expression of GlEB1 by Western blot and IFA as described above.

Measuring Mitotic Index of G. lamblia Trophozoites
With G. lamblia trophozoites treated with water, mispair anti-EB1 morpholinos, or anti-EB1 morpholinos, the ratio of cells with 2 nuclei to those with 4 nuclei was determined to monitor mitotsis as described [26]. The cells were attached on coverslips were fixed with methanol for 5 min and air-dried. The cells were then mounted in VECTASHIELD anti-fade mounting medium with DAPI. The numbers of cells with 4 nuclei or 2 nuclei were counted in a total of 200 cells per each condition.

Expression and Purification of rGlEB1 Proteins
Full-length rGlEB1 was expressed in Escherichia coli and purified as described previously [13]. In addition, the DNA fragment containing the GlEB1 ORF was dissected into two parts. The 59region of eb1 (552 bp) was amplified from the genomic DNA of Giardia WB by PCR using the primers, EB1-full-ERI-F and EB1-CHD-R (Table 2), and the 39-region of eb1 (474 bp) was amplified with another set of primers, EB1-EBD-F and EB1-full-XhoI-R ( Table 2). The resultant eb1 DNA fragments were cloned into pET21b (Novagen) to generate overexpression plasmids, pET21b-EB1-CHD and pET21b-EB1-EBD, for the truncated rGlEB1 polypeptides. These rEB1 polypeptides were expressed in E. coli BL21 (DE3) with 0.5 mM IPTG at 30uC for 3 h, and purified using a TALON metal affinity chromatography as described by the manufacturer (Clontech).

Microtubule Binding Assay
The binding of rGlEB1 to polymerized MTs was observed in vitro using the microtubule binding protein spin down assay kit BK029 (Cytoskeleton). Briefly, MTs were assembled from 100 mg of pure tubulins (isolated from bovine brain; Cytoskeleton) in 20 mL of PEM [80 mM piperazine-N,N9-bis(2-ethanesulfonic acid), pH 6.8, 1 mM EGTA, and 1 mM MgCl 2 ] in the presence of 1 mM GTP and 5% glycerol at 35uC for 20 min, and immediately stabilized in 200 mL of warm PEM-20 mM taxol (Cytoskeleton). Various amount of MTs (0.522 mM tubulin) were incubated with 10 mg of rGlEB1 (0.38-0.7 mM) in a total volume of 50 mL at room temperature for 40 min. The reaction mixtures were then centrifuged through a 50% glycerol cushion-PEM-taxol mixture at 100000xg at 25uC for 40 min using an ultracentrifuge (Hitachi Koki), and the supernatant and pellet fractions were then resolved on SDS-PAGE.
Protein bands, the pellet fraction and the supernatant fraction, were quantified using Multi Gauge V3.0 software (Fujifilm) within the linear signal intensity range, which was determined using standard curves of rGlEB1. The fraction of rGlEB1 bound to MTs was plotted against the concentration of MT included in each reaction. To determine the apparent K d of rGlEB1 for MTs, the data from three independent experiments were fitted by the equation Y = (B max +X)/(K d +X) using SigmaPlot version 9.0 (Systat software). Y is the concentration of rGlEB1 partitioning to the pellet with the MTs whereas B max is the maximal fraction of rGlEB1 in rGlEB1-MTs complex. K d is the dissociation constant where X is the concentration of MT used for each reaction.

Size Exclusion Chromatography (SEC)
Full-length or truncated rGlEB1 protein (about 800 mg) was diluted with PBS on ice before loading onto a Superdex 75 10/ 300 GL (GE Healthcare). Eluted fractions were collected at a volume of 500 mL, and then separated by SDS-PAGE. Size control proteins of 75, 29, 14, and 6 kDa were loaded onto the same SEC. To examine an involvement of disulfide bond in dimerization of GlEB1, eluted proteins from SEC, were incubated with 100 mM DTT for 10 min prior to being separated upon SDS-PAGE.

Site-directed Mutagenesis of the eb1 Gene
To determine the role of the disulfide bond between the two EB1 monomers, cysteine residues at amino acid positions #13 and #46 were mutagenized to serine. Using the EB1 overexpression plasmid, pET21b-EB1-full, as a template, mutant DNA containing the mutation at amino acid position #13 was produced by PCR using primer EB-C13S-21b-F and EB1-full-XhoI-R (Table 2), resulting in a mutant EB1 DNA fragment containing the conversion of cysteine into serine at amino acid position #13. The expression plasmid pET21b-EB1-C46S carrying a mutation at cysteine position #46 of GlEB1 was made by PCR using primers EB1-full-ERI-F and EB-C46S-21b-R (Table 2), and the other mutant DNA carrying mutation at the same amino acid was also made by PCR using primers EB-C46S-21b-F and EB1-full-XhoI-R.
Complementation assays were performed as described [13]. Briefly, yeast cells were synchronized at G1/S by incubation with 0.2 M hydroxyurea (HU; Sigma) for 4 h and released by washing several times with distilled water. To induce the expression of genes under the GAL1-10 promoter, yeast cells were transferred to the SC dropout medium with 2% raffinose instead of glucose for 16 h, and then incubated with 0.2 M HU and 2% galactose/1% raffinose for 4 h. Following 1 min-fixation in 70% ethanol and staining with 1 mg/mL DAPI, the cells were observed with a 100 X objective on an Axiopvert 200 (Zeiss), and the images were captured with an Axiocam MRc camera using AxioVision software (Zeiss).

Construction of G. lamblia Trophozoites Expressing HAtagged GlEB1
End-binding 1 (EB1) protein is one of the +TIPs that connect MT spindles to the kinetochores and modulates MT stability [27]. In this study, functional analysis was performed on the only EB1 homologue in G. lamblia. An E. coli-G. lamblia shuttle plasmid was constructed to contain the DNA fragment encoding the promoter and ORF of GlEB1, from which EB1 is expressed as a fused form in frame with the HA epitope ( Figure 1A). Upon electroporation and selection by G418 resistance, G. lamblia trophozoites carrying this plasmid were established and used for further investigation. These trophozoites were then examined for the expression of a chimeric EB1 tagged with an HA epitope by Western blot using anti-HA antibodies ( Figure 1B). G. lamblia trophozoites carrying the vector plasmid without the eb1 DNA fragment were also constructed, and included in the Western blot with anti-HA antibodies. As expected, the immunoreactive protein band was present in Giardia carrying the HA-tagged GlEB1-expressing plasmid, but absent in the control trophozoites. IFAs of Giardia trophozoites carrying the HA-tagged GlEB1-expressing plasmid were performed using two different antibodies, anti-HA antibodies and anti-GlEB1 antibodies ( Figure 1C). Incubation of these trophozoites with anti-HA antibodies revealed fluorescence at the nuclear membranes and median bodies. IFA using anti-GlEB1 also demonstrated fluorescence staining at the same positions in these transgenic Giardia trophozoites as shown in IFA using anti-HA antibodies.

Determination of Expression Levels and Localization of GlEB1 at Different Stages of the Giardia Life Cycle
Limited information is available on the expression pattern of GlEB1 in G. lamblia. In a previous study, we reported that the intracellular level of GlEB1 is not altered during encystation [13]. In order to examine the expression level of GlEB1 in different stages of the cell cycle, we obtained synchronized trophozoites using Giardia carrying the GlEB1 expression plasmid (Figure 2A). Flow cytometry analysis (FACS) of their DNA content indicated that Giardia trophozoites without any treatment were comprised of cells at G1 phase (31%) and at G2 phase (63%). After 6 htreatment with aphidicolin, the majority of Giardia trophozoites (77%) were arrested at the G1 phase. The aphidicolin-treated cells were transferred to drug-free medium in order to be released into G2 phase. Released cells showed a DNA pattern that is typical of cells at G2 phase (91%).
Giardia cells expressing HA-tagged GlEB1 were harvested at various stages, trophozoite without drug treatment, aphidicolintreated cells, and cells released from growth arrest at various times, ranging from 1 to 6 h. These extracts were analyzed by Western blots to determine intracellular levels of GlEB1 at different stages of Giardia growth, indicating that GlEB1 expression does not vary significantly by cell cycle ( Figure 2B). Western blot of these extracts with anti-Gla-tubulin antibodies served as a loading control.
Giardia cells at G2 phase, released cells from aphidicolin treatment, were examined by IFAs using anti-HA antibodies ( Figure 2C). Mitotic cells were frequently observed with mitotic spindles labeled with fluorescence, indicating that GlEB1 was localized to the mitotic spindles of the mitotic cells at anaphase. IFA of these cells with anti-Glc-tubulin antibodies showed fluorescent labeling at the two basal bodies during anaphase, confirming that they are mitotic cells ( Figure 2D).
The intracellular level of GlEB1 was found to be constitutive during encystation [13]. In this study, we examined the possibility that localization of EB1 was altered during encystation by IFA of Giardia expressing HA-tagged GlEB1 with anti-HA antibodies ( Figure 3A). Up to 24 h-after induction to encystation, GlEB1 was mainly found at the nuclear membrane in the most of the Giardia cells, whereas some of the cells showed fluorescent labeling of the median bodies. In cysts at 48 h induction, fluorescence was absent in the nuclear membranes and dispersed in the cytoplasm of the cells. As a control for encystation process, intracellular location of CWP1 was also examined in trophozoites as well as encysting cells ( Figure 3B). CWP1 was barely detected in trophozoites, and later found in encystation-specific vesicles in the encysting cells. Localization of CWP1 in the cyst walls was distinct in the cysts at 48 h-after induction of encystation. These results showed that the location of GlEB1 at the nuclear membrane and median bodies was changed into a condensed pattern in the cytoplasm of cysts.

Effect of GlEB1 Knockdown in Mitosis of G. lamblia
To define the role of EB1 in G. lamblia, we designed a morpholinos, anti-EB1, which can block translation of GlEB1 mRNAs ( Table 2). As a control, another morpholinos, mispair anti-EB1, was made and used to transfect into G. lamblia trophozoites by electroporation ( Table 2). In addition, a set of G. lamblia was treated with sterile water instead of morpholinos.
Cell extracts prepared from these cells at 48 h and 72 h posttransfection were monitored for their intracellular levels of EB1 by Western blot using anti-GlEB1 antibodies ( Figure 4A). Both cells treated with mispair anti-EB1 or anti-EB1 morpholinos showed decreased amount of GlEB1 at 48 h post-transfection to 69% and 43%, respectively ( Figure 4B). Amount of GlEB1 was restored to 88% and 70% of that of cells treated with water in cells at 72 h post-transfection. Decreased level of GlEB1 was also confirmed by IFA of G. lamblia cells treated with sterile water, the mispair anti-EB1, or anti-EB1 morpholinos after 48 h ( Figure 4C). GlEB1 was found in nuclear membrane and median bodies in cells treated with water or mispair anti-EB1 morpholinos. In G. lamblia transfected with anti-EB1 morpholinos, GlEB1 was barely detected in nuclear membranes. Clear staining of median bodies were observed in G. lamblia cells treated with anti-EB1 morpholinos, but at lower levels than in cells treated with water or mispair anti-EB1 morpholinos.
Effect of low expression of GlEB1 was determined by measuring mitotic indexes, which is portions of cells with four nuclei ( Figure 5). Upon staining with DAPI, the cells were counted for those with two nuclei or four nuclei ( Figure 5A). About three percentages of cells contained 4 nuclei when they were treated with water or mispair anti-EB1 morpholinos. The percentage of cells with four nuclei was increased to 10% upon transfection with anti-EB1 morpholinos ( Figure 5B). This result indicates that GlEB1 plays some role in mitosis of G. lamblia.
The MT-binding region in GlEB1 was identified by testing the three rGlEB1 proteins in a sedimentation assay with taxolstabilized MTs ( Figure 6). Full-length rGlEB1 showed a significant affinity for the MTs, with a dissociation constant (K d ) of 0.34 mM and 24% of the added rEB1 was precipitated with the MTs. On the other hand, rGlEB1 1-184 containing the CH domain and the linker showed significant attenuation in MT-binding, resulting in a K d of 0.83 mM and a B max of 0.03. Interestingly, rGlEB1 102-238 containing the linker and the EB1 domain demonstrated a comparable MT-binding ability to that of the full-length rEB1 in this assay (K d = 0.22 mM, B max = 0.2).

Characterization of a Functional Domain for the Dimerization of GlEB1 Protein
Each of these three rGlEB1 proteins was analyzed by size exclusion chromatography (SEC) to examine whether they are present as a dimeric form ( Figure 7A). rGlEB1 1-238 eluted as two overlapping peaks earlier than the marker protein of 29 kDa, indicating the dimeric formation of rGlEB1 1-238 . rGlEB1  containing a CH domain and the linker, also eluted earlier than the 29 kDa marker protein, suggesting that this truncated protein is able to form dimers. In the case of rGlEB1 102-238 , they passed through the column as an elongated, dimeric peak, raising the possibility of conformational modification or multimeric formation of rGlEB1 102-238 . Elution of rGlEB1 1-184 in a dimeric form from SEC is interesting in that this region is involved in MT-binding, but not in the dimerization of Bim1p, an EB1 homolog of budding yeast [5].
When the eluted fractions from SEC were analyzed via SDS-PAGE, rGlEB1 1-238 appeared as a dimer and became monomer with the addition of dithiothreitol (DTT) ( Figure 7B). This result suggests that intermolecular disulfide bond(s) are involved in the formation of GlEB1 dimers. The amino acid sequence of GlEB1 revealed two cysteine residues at amino acid position #13 and #46. Based on the location of these cysteine residues, rGlEB1  was also examined, in which dimerization occurs through intermolecular disulfide bond(s). rGlEB1  was present as 40 kDa without DTT, and became 20 kDa upon incubation with DTT. As expected, rGlEB1 102-238 appeared as a monomer regardless of the presence of DTT (data not shown).
Two cysteine-mutant rGlEB1 variants, C13S and C46S, were analyzed by SEC (Figures 7C 7D, respectively). C13S-mutant rGlEB1 eluted as a single peak in dimeric form, and appeared as a 26 kDa band on SDS-PAGE, regardless of DTT. C46S-mutant rGlEB1 eluted as two overlapping peaks earlier than the marker protein of 29 kDa, and appeared as a dimer on SDS-PAGE without DTT. This experiment indicated that dimer formation of Giardia EB1 occurs via two different mechanisms: the interaction between the EB1 domains and intermolecular disulfide bond formation of two EB1 monomers. Mutant rGlEB1, containing both C13S and C46S conversions showed an elution pattern identical to that of the C13S-mutant rGlEB1 (data not shown).

Complementation of Yeast BIM1 Mutant with C13S mutant GlEB1
To examine whether the putative EB1 protein of G. lamblia plays a role in controlling MT dynamics, we used a budding yeast system that is amenable to various genetic approaches. In addition to the positive controls for complementation for giardial eb1 and the BIM1 gene of S. cerevisiae, mutant eb1 gene encoding mutant GlEB1 with C13S was cloned into the same yeast expression system and transformed into YSK1134 (bim1D mutant). The expression of these G. lamblia EB1 proteins in the yeast transformants was examined by Western blot analysis using anti-GlEB1 antibodies ( Figure 8A). The bim1D cells carrying only pRS426+P GAL1-10 (negative controls) or yeast BIM1 gene did not show any immunoreactive protein. Only the mutant yeast strains carrying wild type eb1 gene or mutant eb1 gene showed two immunoreactive bands.
We examined the nucleus position in YSK1134 (bim1D) cells expressing giardial EB1 proteins or yeast Bim1p in comparison with the control, YSK1134 carrying pRS426+P GAL1-10 . These yeast cells were grown in the presence of glucose (as a non-induced condition) or in the presence of galactose and raffinose (as an induced condition) and number of the yeast cells with bud-distal nuclei was determined under both conditions ( Figure 8B). About sixty percent of bim1D cells carrying pRS426+P GAL1-10 showed defects in nucleus positioning. Under induced conditions, the percentage of cells with a bud-distal nucleus were decreased to 38% and 40% in the bim1D cells carrying pZhu20 (wild type BIM1 gene) and pZhu19 (wild type eb1 gene). indicating a partial, but significant complementation of the bim1D yeast cells with Giardia eb1 gene. On the other hand, the bim1D cells carrying pRS426+ P GAL1-10 EB1-C13S with C13S mutant EB1 maintained a similar level of cells with a bud-distal nucleus (62%). These results demonstrate that the EB1 homolog of G. lamblia can function similarly to Bim1p in the budding yeast and cysteine residue #13 of GlEB1 plays a role in this process.

Discussion
Understanding the role of EB1 is important in revealing the mechanism of how Giardia divides with a coordinated distribution of its organelles and cytoskeletal structures. In this study, the function of GlEB1 was assessed by monitoring the expression and localization of GlEB1 at various phases of the cell cycle ( Figures 2B  and 2C, respectively) and during encystation ( Figure 3A). A study using green fluorescent protein-tagged GlEB1, demonstrated its localization at flagellar tips and median bodies, and suggested that it is also located at mitotic spindles [12]. IFA demonstrated that GlEB1 is localized to the nuclear membranes, median bodies, and axonemes of the trophozoites [13]. Giardia cells arrested at the G1 and G2 phases did not show a dramatic change in intracellular GlEB1 levels ( Figure 2B). Western blot analysis showed that the intracellular amount of GlEB1 is constant in encysting cells as well as trophozoites [13]. These observations are consistent with information on the expression levels of EB1 in other organisms in which a constant amount of EB1 is maintained at any stage of cell division and cell differentiation [28].
It is likely that EB1 localization at the correct sites is critical for its proper function. As expected, in mitotic cells, localization of GlEB1 in mitotic spindles was clearly observed ( Figure 2C). IFA using anti-Glc-tubulin antibodies ( Figure 2D) demonstrated its localization to the basal bodies of mitotic cells of Giardia at anaphase, which had been previously reported in G. lamblia [29]. Intracellular localization of GlEB1 in the encysting cells was in the nuclear membrane and median bodies ( Figure 3A). GlEB1 is localized to the cytoplasm in a more dispersed pattern in the cyst form. This observation needs to be explained through an extensive investigation of cyst structures using more cell markers.
EB1 proteins are the most conserved and ubiquitous +TIPs in eukaryotes [30]. Putative amino acid sequences of GlEB1 demonstrated the presence of a CH domain, a linker, and an EB1 domain (Figure 6 and 7). Interestingly, GlEB1 lacks the EEY/ F motif which is conserved in other eukaryotes [6]. A previous study demonstrated that rGlEB1 is precipitated with polymerized MTs in vitro [13]. In this study, the role of the two domains in GlEB1 was assessed using two truncated forms of rGlEB1, rGlEB1 1-  domain showed dramatic attenuation in its MT-binding ability. The CH domain is known to be involved in MT binding in other organisms [31]. However, budding yeast Bim1p containing only the N-terminal CH domain was also found to be defective in MT binding due to its inability to form dimers [5]. Loss of the MTbinding ability of rGlEB1 1-184 containing the CH domain and a linker did not occur because this truncated rGlEB1 1-184 is able to form dimers in an EB1 domain-independent manner ( Figure 7A). We cannot rule out the possibility that the conformation of this truncated rGlEB1 1-184 was modified as it is impertinent to MT binding. Interestingly, rGlEB1 102-238 with the EB1 domain and linker demonstrated equivalent binding to MTs to that of full-length rGlEB1. It is contrary that the N-terminal-truncated rBim1p containing the EB1 domain showed dramatic defect in MT-binding assays [5].
Using these three rGlEB1 proteins, we examined whether GlEB1 was also present as a dimer ( Figure 7A). SEC analysis indicated that both full-length rGlEB1 1-238 and rGlEB1 1-184 with the CH domain were found as dimers. rGlEB1 102-238 carrying the EB1 domain eluted earlier than 29 kDa, the expected size of the dimer. This can be explained by rGlEB1 having a conformational change or modification resulting in earlier elution in SEC analysis. Otherwise, this rGlEB1 may have a tendency to form multimer.
Interestingly, full-length rGlEB1 eluted as two peaks, an earlier smaller peak and a later major peak ( Figure 7A). SDS-PAGE of the eluted proteins yielded two protein bands larger than the monomer, which became 26 kDa after DTT treatment, indicating and indicated in each elution graph. (B) SDS-PAGE of the SEC fractions of rGlEB1 1-238 , without any treatment (a) or with DTT treatment (b). (C) SEC profile of C13S mutant rGlEB1 (a), SDS-PAGE of the SEC fraction without DTT treatment (b), and SDS-PAGE of the SEC fraction with DTT treatment (c). (D) SEC profile of C46S mutant rGlEB1 (a), SDS-PAGE of the SEC fraction without DTT treatment (b), and SDS-PAGE of the SEC fraction with DTT treatment (b). doi:10.1371/journal.pone.0097850.g007 Figure 8. Complementation of wild type or C13S mutant GlEB1 in the BIM1 knock-out mutants of S. cerevisiae. (A) Expression of wild type or C13S mutant GlEB1 in yeast. A plasmid, pRS426+P GAL1-10 , serves as a vector control, which has a galactose/raffinose-inducible promoter functioning in budding yeast. For pZhu19, the eb1 gene of G. lamblia was inserted under P GAL1-10 , whereas plasmid pZhu20 containing the BIM1 gene, a yeast homologous gene for eb1, serves as a positive control for complementation. In addition, mutant eb1 gene with C13S change was expressed in yeast. The expression of giardial EB1 protein in various yeast cells carrying a vector plasmid or one of the complementation plasmids. Extracts of various yeast cells were analyzed by a Western blot using EB1-specific antibodies. (B) The percentage of Bim1p-or wild type GlEB1complemented strains with bud-distal nucleus under induced conditions was significantly different from that of control strain carrying vector plasmid. On the other hand, the percentage of C13S mutant-complemented strain with bud-distal nucleus was not statistically different from that of control yeast carrying vector plasmid. The percentage of complemented strains with a bud-distal nucleus under inducing conditions was significantly different from that of the control strain carrying pRS426+P GAL1-10 under inducing conditions by the Student's t-test. Data with p-value of ,0.01 are indicated with an asterisk. Standard deviations were derived from three independent experiments. The percentage of abnormal cells with biased nucleus distribution was measured in triplicates for each sample. doi:10.1371/journal.pone.0097850.g008 the presence of a disulfide bond in the GlEB1 dimer. In silicoanalysis of GlEB1 amino acid sequences revealed that two cysteine residues are present at amino acid residues #13 and #46. EB1 proteins of other organisms (S. cerevisiae, Homo sapiens, and Mus musculus) have 3-4 cysteine residues; specifically 2-3 cysteines are present in the N-terminal portion of the protein. Formation of a disulfide bond between the EB1 monomers has not yet been reported in these organisms. Site-directed mutagenesis of these two cysteine residues clearly indicate that cysteine #13 is involved in intermolecular disulfide bonds between two monomer of EB1. Dimer formation of GlEB1 is mediated through an EB1 domainmediated interaction, as suggested by the other system, which appeared as a major band in SEC. In addition, the earlier minor peak is derived from disulfide bond formation as it disappeared in rGlEB1 with C13S replacement. The presence of dimeric GlEB1, mediated by an intermolecular disulfide bond in vivo form should be examined in further investigation.
A comparison of GlEB1 with other EB1 sequences revealed Cterminal tail sequences ranging from 20 bases (H. sapiens and M. musculus) to 68-70 bases (S. cerevisiae and Schizosaccharomyces pombe). Along with the EB1 domain, these C-terminal tail sequences are known to be involved in interactions with other +TIPs, such as CLIP and p150 [32], [33]. Interestingly, GlEB1 has an unusually short C-terminal tail of four bases (EYHQ). In the case of human and murine EB1, the last three amino acids are EEY, which are involved in the interaction with +TIPs [34], [35]. In the case of the EB1 of budding yeast, the last three amino acids are ETF, which have similar biochemical properties to that of EEY. However, EB1 homologues of fission yeast and plants do not have the C-terminal tail sequence as well as GlEB1. An important question would be to identify the functional difference caused form the lack of these sequences in GlEB1. The cytoskeleton-associated protein-glycanrich (CAP-Gly) domain of +TIPs is known to function in interaction with the C-terminal EEY sequence of EB1 and microtubules in mammalians and yeasts [36]. In contrast with GlEB1, two putative a-tubulins of G. lamblia (accession numbers GL50803_112079 and GL50803_103676) have a C-terminal tail sequence (MEEDDAY), which shows 4 conserved amino acids of the 6 consensus C-terminal tail, P/E/D-E/D-P/E/D-E-E/T-Y/F [36]. As interacting proteins with C-terminal EEY sequence, proteins having the CAP-Gly domain were searched in Giardia database, resulting in identification of two putative proteins, tubulin specific chaperone B (GL50803_5374) and tubulin specific chaperone E (GL50803_16535). Thus, it is likely that tubulin interacts with tubulin specific chaperon in G. lamblia, even though it should be experimentally proven.
Role of GlEB1 in mitosis was indirectly examined via complementation experiment using budding yeast system [13]. In this study, we examined the role of GlEB1 in G. lamblia via the knockdown experiments using morpholinos-based anti-sense oligonucleotides (Figures 4 and 5). At 48 h-post transfection, inhibition of EB1 expression was dramatic. Cells transfected with mispair anti-EB1 morpholinos, also showed decreased level of GlEB1. However, the cells transfected with anti-EB1 showed more dramatic reduction in GlEB1 expression as measured in Western blot analysis (Figures 4A and 4B) and IFA ( Figure 4C). Significant increase of cells with 4 nuclei among anti-EB1 morpholinostreated cells indicated that GlEB1 play an important role in mitosis of G. lamblia ( Figure 5). In addition to mitotic index, we observed that size of median bodies was reduced in Giardia treated with anti-EB1 morpholinos (data not shown). This observation needs to be documented in further investigation. The complementation assay using the BIM1 mutant S. cerevisiae indicated that GlEB1 can replace the function of Bim1p in the budding yeast and dimerization of GlEB1 via its 13rd cysteine residue is important ( Figure 8B). Functional complementation between EB1 homologues of different organisms has been reported [37]. Human EB1 can substitute for the loss of mal3 + gene product, the fission yeast EB1 homologue, as appeared suppression of its hypersensitivity to thiabendazole, the microtubule destabilizing drug.
This study provides an in vitro biochemical characterization of Giardia EB1 with respect to its MT-binding activity and dimerization. Dimer formation of GlEB1 was mediated by an interaction between the EB1 domains and an intermolecular disulfide bond. In addition, role of EB1 in division of G. lamblia was shown by the knockdown experiment.