Entamoeba histolytica sirtuin EhSir2a deacetylates tubulin and regulates the number of microtubular assemblies during the cell cycle

Summary

We have discovered four sirtuin genes in Entamoeba histolytica, two of which are similar to eukaryotic sirtuins and two to bacterial and archaeal sirtuins. The eukaryotic sirtuin homolog, EhSir2a showed NAD⁺ dependent deacetylase activity and was sensitive to Class III HDAC inhibitors. Localization of EhSir2a at different cellular sites suggested that this deacetylase could have multiple targets. Using an E. histolytica cDNA library in the yeast two-hybrid genetic screen, we identified several proteins that bound to EhSir2a. These proteins included Eh α-tubulin, whose interaction with EhSir2a was validated in E. histolytica. We have shown that EhSir2a deacetylated tubulin and localised with microtubules in E. histolytica. Increased expression levels of EhSir2a in stable transformants led to reduced number of microtubular assemblies in serum synchronised cells. This effect was abrogated by mutations in the deacetylase domain of EhSir2a, showing that EhSir2a deacetylase activity affected the stability and number of microtubular assemblies during the cell cycle of E. histolytica. Our results suggest that epigenetic modification of tubulin by EhSir2a is one of the mechanisms that regulates microtubular assembly in E. histolytica.

Introduction

The human pathogen Entamoeba histolytica has an unusual mode of cell division and proliferation where known regulatory mechanisms appear to be absent. Thus cell cycle progression is not regulated by checkpoints in this protist parasite (Mukherjee et al., 2009). Several studies suggest that nuclear DNA may replicate several times before chromosome segregation and nuclear division occurred (Mukherjee et al., 2009; Mukherjee et al., 2008; Das and Lohia, 2002). This leads to the accumulation of multiple genome equivalents in a single nucleus. Segregation of duplicated or multiple copies of the genome occur on unusual microtubular structures that are either bipolar, radial monopolar or multi-polar assemblies (Mukherjee et al., 2009; Mukherjee et al., 2008; Dastidar et al., 2007). Additionally, spatial and temporal uncoupling of nuclear division from cytokinesis often leads to multi-nucleated and anucleate cells (Mukherjee et al., 2009). Both E. histolytica and E. invadens frequently use assistance from helper cells to complete cytokinesis (Mukherjee et al., 2009; Biron et al., 2001). It has also been suggested that Entamoeba cells have mechanisms to eliminate extra genome copies during differentiation or changes in growth conditions (Mukherjee et al., 2008). From all these studies, it is becoming increasingly clear that E. histolytica cells use novel modes of cell division processes that lead to genetic heterogeneity in a population of amoebae at any given time. Absence of known regulatory mechanisms and novel modes of cell cycle progression in E. histolytica has initiated a search for proteins and regulatory processes that affect cell proliferation and survival of amoeba.

Various epigenetic mechanisms are active in E. histolytica that regulate the manifestation of virulence, pathogenesis and differentiation (Ehrenkaufer et al., 2009; Huguenin et al., 2009; Lavi et al., 2008; Mirelman et al., 2008; Ehrenkaufer et al., 2007; MacFarlane and Singh, 2006; Bernes et al., 2005). Several histone modifying enzymes have been identified from the genome of E. histolytica that include homologs of Histone acetyltransferase (EhHAT) belonging to the GNAT and MYST family and Histone deacetylases (EhHDAC) from class I, II and III (Iyer et al., 2008; Ramakrishnan et al., 2004). Class I/II EhHDACs have been implicated in the regulation of encystation of trophozoites and in restricting the DNA content of trophozoites in response to short chain fatty acids in the colon (Byers and Eichinger, 2008; Byers et al., 2005).

In other organisms, Class III HDACs or Sir2 proteins are known to deacetylate both histone and non-histone proteins and consequently regulate diverse cellular processes such as gene expression, cell cycle progression and cytoskeletal activity. In humans, seven Sir2-like proteins (SIRT1-7) have been identified (Frye, 2000; Frye, 1999) which modify different acetylated substrates and are localized at different sub-cellular sites. The human SIRT2 deacetylates α-tubulin (North et al., 2003), FOXO3a (Wang et al., 2007) as well as histone H4-K16 (Vaquero et al., 2006) and its over-expression increases multi-nucleation suggesting its role in regulating cell division (North and Verdin, 2007).

In this study, we have identified four genes of class III HDAC from the genome sequence of E. histolytica. Two of these are similar to fungal and protist sirtuins while the other two are similar to archeal and bacterial homologs. The eukaryotic homolog-EhSir2a had unique sequence motifs at its N-terminal and a conserved NAD⁺ dependent deacetylase domain at its C-terminus. Using the yeast two hybrid genetic screen, we identified several proteins including Eh α-tubulin that bound to EhSir2a. In this study, we have shown that EhSir2a deacetylates tubulin, localizes on microtubules and regulates the frequency of microtubular spindles during the amoeba cell cycle.

Results

E. histolytica encodes homologs of both eukaryotic and prokaryotic sirtuins

We have identified four Sir2-like proteins from the genome of E. histolytica and named them EhSir2a, EhSir2b, EhSir2c and EhSir2d. All four Eh-sirtuins have a conserved deacetylase domain and variable lengths with unique sequences at their N and C-terminii (Fig.1A). An unrooted phylogenetic tree was constructed (Fig.1B) using the conserved deacetylase domain of Sir2 proteins from diverse organisms (Table S1). In this tree, sirtuins have been classified into five major groups, class I, II, III, IV and U. EhSir2a and EhSir2b clustered in class I with fungal, protozoan, nematodes, insect and mammalian homologs; while EhSir2c and EhSir2d clustered in class U with archaea and gram positive bacteria like Staphylococcus aureus, Thermotoga maritima, Bacillus subtilis and Clostridium acetabutylicum (Fig. 1B). Similarity with bacterial and archeal sirtuins suggests that EhSir2c and EhSir2d may have been acquired by lateral gene transfer from bacteria or archaea, or that prokaryotic sirtuins may have a common ancestor with amoebic sirtuins.

Fig. 1. Sequence analysis of EhSir2 proteins.

A. Domain analysis. The conserved domains- deacetylase (in all 4 Eh sirtuins) and Zn finger domain (present only in EhSir2a) are shown. The scale of 50 amino acids was drawn according to the length of the proteins.

B. Phylogenetic tree of sirtuins. The conserved deacetylase domain was used to identify Sir2 homologs from different organisms. The E. histolytica Sir2 homologs are underlined and the five classes (I–IV, and U) are encircled.

E. histolytica Sir2 homologs contain the conserved GAG, PXXXH, NID, FGE, and GTS, IN motifs, a four-Cys-zinc finger domain, and conserved HG residues adjacent to the zinc finger like other sirtuins (Frye, 1999). However, alignment of amino acid sequences of EhSir2a–d proteins with the eukaryotic (human) and prokaryotic (B. Subtilis) homologs showed clear differences between the deacetylase domains of the two classes (Fig. S1). EhSir2a–b, showed the conserved amino acid residues distinct from EhSir2c–d that are typically seen in eukaryotic and prokaryotic sirtuins. Except for EhSir2a that contained a zinc-finger motif at its N-terminus, the remaining Eh Sirtuins did not show any such unique domain.

EhSir2a shows NAD⁺ dependant deacetylase activity and is sensitive to nicotinamide and sirtinol but not splitomicin

Sequence analysis of EhSir2a showed the presence of the NAD⁺ dependent deacetylase conserved domain found in all known sirtuins. In order to validate its enzymatic activity, we tagged EhSir2a with His- and HA-epitopes at its C-terminus (Dastidar et al., 2007). The expression of epitope tagged, recombinant EhSir2a-HH in the stable transformants was confirmed in cell extracts and on Ni-NTA agarose beads using immuno-hybridisation with anti-Sir2a and anti-HA antibodies (Fig. 2A). Deacetylase activity of EhSir2a was analyzed after isolating EhSir2a-HH bound to Ni-NTA agarose beads using a standard flourescence assay (Fig. 2A). EhSir2a showed high levels of deacetylase activity when NAD⁺ was added. The low levels of activity detected in the assay without added NAD⁺ may be due to residual amounts of NAD⁺ from the cell extract that adhered to the beads during isolation (Fig.2A).

Fig. 2. Deacetylase activity of EhSir2a.

A. Deacetylase activity of EhSir2a. The expression of EhSir2a-HH in the cellular extract and in the Ni-NTA agarose beads is shown by western hybridization against monoclonal anti-HA. β-tubulin is used as the endogenous loading control in the cellular extract. Ni-NTA agarose beads were added to extracts of cells transformed with EhSir2a-HH (Sir2a) and control (Con). Proteins bound to the beads were analysed by SDS-PAGE and western hybridized with anti-EhSir2a and HA antibodies. The deacetylase activity was analysed fluorometrically with proteins bound to Ni-NTA beads (Control bead and Sir2a bead). Y axis shows the fluorescence intensity in arbitrary fluorescence units. The concentrations of nicotinamide (NAM), sirtinol and splitomicin were 500 µM, 100 µM and 500 µM respectively.

B, C and D. Inhibition of deacetylase activity. B, C and D show the effect of different concentrations of nicotinamide, sirtinol and splitomicin respectively on deacetylase activity of EhSir2a-HH.

E. Alignment showing the difference in amino acid sequences of ySir2 and EhSir2a. The target amino acids of Splitomicin in ySir2, is shown by downward pointed arrows. These target sites are absent in EhSir2a and is presumed to be the cause of its insensitivity to the drug.

F. Anti-proliferative activity of nicotinamide against E. histolytica. E. histolytica HM1: IMSS trophozoites were treated with varying concentrations (0–20 mM) of nicotinamide and after 20 h, the effect of drug was determined by counting the number of cells (cells ml⁻¹) in a haemocytometer. Initial inoculum was 4 × 10⁴ cells ml⁻¹ (± s.d, n=3). IC₅₀ = 4.67 mM.

Small molecules like nicotinamide, sirtinol, and splitomicin which inhibit the deacetylase activity of Sir2 in different organisms were also tested against EhSir2a (Fig. 2A). The deacetylase activity of recombinant EhSir2a was found to be inhibited by nicotinamide and sirtinol and the IC₅₀ values were calculated to be 0.79 mM and 12.47 µM (Fig.2B and 2C) respectively. Our experiments showed that EhSir2a was insensitive to splitomicin (Fig.2A, 2D). Splitomicin was initially identified as an inhibitor of Saccharomyces cerevisiae Sir2 protein (ySir2p) and later shown to require the amino acids His-286, Leu-287 and Tyr-298 residues for its inhibitory activity (Hirao et al., 2003; Bedalov et al., 2001). Sequence alignment showed that the crucial His residue responsible for drug recognition and contact was absent in EhSir2a (Fig. 2E) and this may contribute to splitomicin resistance of EhSir2a.

In growth experiments, nicotinamide showed anti-proliferative activity against E. histolytica trophozoites in a dose dependent manner (Fig. 2F) with IC₅₀ of 4.67 mM. Although this value is much higher than that observed with recombinant EhSir2a in vitro it likely reflects the multiple targets that utilise NAD⁺ and consequent lowered availability to EhSir2a within the cell. The drug, sirtinol could not be tested in growth medium due to its low solubility.

EhSir2a is expressed at multiple sub-cellular sites in E. histolytica

Sequence alignment of the four sirtuins showed overall low identity at isolated amino acids, except at the GAG domain (EhSir2a: 142–156 aa; Fig.S2). Therefore, we raised polyclonal antibodies against EhSir2a (N-terminal 1–120 aa) and (C-terminal 238–383aa). The anti-N-terminal EhSir2a antibody showed poor immunogenicity and was not used. The anti C-terminal antibody showed very high immunogenicity and specificity for EhSir2a. The antibody hybridized with a protein of ~45 kDa as expected and did not cross react with any other Sir2 proteins (MWs 39 kDa, 36 kDa and 32 kDa) in the cell extract on Western blots (Fig. 3A-a). Nuclear and cytoplasmic fractionation showed that EhSir2a was present both in the nucleus and in the cytoplasm (Fig.3A–b). Anti-Histone H2A antibody was used to identify the nuclear fraction and an anti-PGK antibody was used to show the cytoplasmic fraction. Fig.3A–d shows the immuno-localisation of EhSir2a in the nucleus and cytoplasm of amoeba cells. The expression of the recombinant epitope tagged protein EhSir2a-HH in stable transformants was confirmed on Western blots using the monoclonal anti-HA antibody (Fig. 3B). The HA antibody and the polyclonal EhSir2a antibody co-localised at the same cellular sites (Fig.3B). Expression of EhSir2a in the stable transformants was analysed both at the mRNA and protein levels using appropriate controls (Fig. 3C,D). A comparison of mRNA levels of stable transformants of EhSir2a-HH and control cells showed approximately 1.8 fold increase in the expression of EhSir2a and the protein expression was increased approximately 1.6-fold at 40 µg/ml of G418.

Fig. 3. EhSir2a is expressed in multiple sub-cellular sites.

A. Cellular extract of E. histolytica was analysed by SDS-PAGE and the western blots are hybridized against pre-immune and polyclonal anti-EhSir2a antibody (a). Cytoplasmic and nuclear fractions of E. histolytica were separated by SDS-PAGE, transferred to Western blots and hybridized with anti-EhSir2a. Monoclonal anti-PGK and polyclonal anti-Histone H2A were used as the markers of cytoplasmic and nuclear fractions respectively (b).

E. histolytica HM1: IMSS trophozoites were fixed on cover slips and hybridized with preimmune (c) or polyclonal anti-EhSir2a antisera (d) followed by TRITC-conjugated anti-rabbit antibody. DNA was stained with DAPI. Images were acquired under a 63× oil DIC objective in a Zeiss LSM 510 Meta confocal microscope. Bar represents 10 µm.

B. EhSir2a-HH transformant was stained with anti-HA and polyclonal anti-EhSir2a. Images were acquired under a 63× oil DIC objective in a Zeiss LSM 510 Meta confocal microscope. Bar represents 10 µm.

Western blots (RHS) of cell extracts from control and EhSir2a-HH transformants were hybridized with monoclonal anti-HA to detect the epitope tagged EhSir2a. Tubulin was used as loading control and detected by polyclonal anti-β-tubulin.

C and D. Comparison of mRNA and protein expression of EhSir2a in control and EhSir2a transformants. The bar diagrams show the relative expression of EhSir2a mRNA (C) and protein (D) in EhSir2a-HH transformants compared to the control transformants, ± S.D. (n=3).

EhSir2a interacts with several proteins

Protein binding motifs in EhSir2a sequence and localization of EhSir2a in different cellular compartments of E. histolytica prompted us to look for its possible interactors. We therefore set up a yeast two hybrid genetic screening in which EhSir2a was used as bait and screened against a E. histolytica cDNA library in the yeast two hybrid vector pACT2. After screening a million colonies, 21 clones were obtained that were able to grow on selective media (−LTH supplemented with 3.5 mM 3-AT). These clones were further screened on −LTA plates and for beta-galactosidase activity. Clones that were positive in all three selection tests were used for isolation of prey plasmids and re-transformed into the parental yeast strain containing the bait -EhSir2a and their positive interaction was reconfirmed. Finally, seven interactors were identified by DNA sequencing. Proteins interacting with EhSir2a included two isoforms of α-tubulin, a putative actin like protein, a hypothetical protein with homology to translation initiation factor and a few other proteins (Table.1). Identification of interactors with diverse function suggested that EhSir2a likely deacetylates multiple substrates. Tubulin has been identified as a substrate of the human SIRT2 (North et al., 2003) and more recently as a binding partner of the Leishmania sirtuin (Tavares et al., 2008). Except for tubulin, most of the EhSir2a interactors were largely uncharacterized for their function in E. histolytica.

Table 1.

Identification of proteins interacting with EhSir2a by yeast two hybrid screening

Sl. No.	Name of EhSir2a interactors	Accession number (TIGR)
1	Tubulin α-chain	249.m00082
2	Tubulin α-chain	83.m00165
3	Elongation factor 2	187.m00072
4	Ser/Thr protein phosphatase	87.m00173
5	Putative actin	27.m00265
6	Proteasome beta subunit	21.m00241
7	Hypothetical protein (Translation initiation factor)	53.m00204
8	Phospholipase B	62.m00161

EhSir2a binds tubulin in E. histolytica and deacetylates mammalian tubulin

To show physical interaction between tubulin and EhSir2a, the cellular extract of EhSir2a-HH stable transformant was incubated with Ni-NTA agarose bead and the bound proteins were separated on SDS-PAGE followed by Western blotting. EhSir2a-HH and tubulin were detected by hybridisation with monoclonal anti-HA and polyclonal anti-β-tubulin antibody respectively (Fig. 4A). It can be seen from the Western blot analysis that tubulin was not detected in Ni-NTA beads added to control cell extract (CB) but was present in the proteins isolated from Sir2-HH transformants (SB). Our anti-Eh β-tubulin and anti-Eh α-tubulin antibodies cross-reacted with both the monomers, possibly because of the high homology between the two proteins. So α- and β-tubulin could not be distinguished in the co-precipitation assays (Fig.S3). In our hands the anti-Eh β-tubulin antibody gave better results and we used it for detecting tubulin in all the subsequent experiments.

Fig. 4. EhSir2a binds tubulin in E. histolytica and deacetylates mammalian tubulin.

A. EhSir2a-HH expressed in E. histolytica binds tubulin. The cellular extract of EhSir2a-HH transformant (Sir2a) was incubated with Ni-NTA agarose beads and the proteins bound to the beads (SB) were analyzed by western hybridization against anti-Eh β-tubulin and anti-HA. C denotes the cellular extract of control transformant and CB denotes proteins isolated on Ni-NTA beads added to control cell extracts.

B. GST-EhSir2a expressed in E. coli. GST-EhSir2a and GST were expressed in BL21 DE3. The purified proteins are shown in a coomassie stained SDS-PAGE. The molecular weight marker is shown at the right.

C. GST-EhSir2a binds to tubulin GST-EhSir2a bound to glutathione sepharose 4B beads was incubated with cellular extract of E. histolytica (HM1) and the bound proteins were analyzed. Western blot hybridization with anti-Eh ²-tubulin detects tubulin in the bead coated with GST-EhSir2a (SB) and not in control beads (CB).

D. Tubulin is deacetylated by EhSir2a. Goat brain tubulin was incubated with GST alone (GST) and GST-Sir2a (Sir2a) and analyzed by western blot hybridization with anti-α-tubulin and anti-acetylated α-tubulin.

E. Co-localization of EhSir2a on microtubules. EhSir2a transformants were fixed and stained with monoclonal anti-HA and polyclonal anti-Eh ²-tubulin. In the merged image, yellow colour shows the co-localization of EhSir2a with microtubules. The corresponding bright field image with nucleus (N) is shown. Bar represents 10 µm.

F. Lower frequency of MT structures is due to over-expression of EhSir2a. MT structures were counted in stable EhSir2a transformants as described in the text. Bar diagram shows the distribution of MT structures in control and Sir2a transformants, ± S.D. (n=3, for each set 600 cells were taken).

Additionally, EhSir2a was expressed in E. coli as a GST fusion protein (Fig.4B) and bound to glutathione sepharose beads before incubating with the crude cell extract of E. histolytica HM1: IMSS. E. histolytica proteins bound to GST-EhSir2a were analysed by SDS-PAGE and immuno-hybridisation of the Western blot with Eh β-tubulin antibody (Fig. 4C). It was observed that Eh (α/β) tubulin bound to GST-EhSir2a (SB, Fig.4C) and not to GST alone (CB, Fig, 4C).

GST-EhSir2a bound to glutathione sepharose beads was used to test tubulin deacetylase activity of EhSir2a (Fig.4D). Since it was difficult to isolate the Ehα/β tubulin dimer or either of the subunits in a soluble monomeric form, we used mammalian acetylated tubulin as a substrate to test if EhSir2a deacetylates tubulin. Tubulin is easily isolated from brain tissue and is heavily acetylated (Hubbert et al., 2002; Morales and Fifkova, 1991). Using monoclonal anti-acetylated α-tubulin and anti-α-tubulin antibodies against mammalian tubulin it was shown that acetylated mammalian tubulin is deacetylated upon incubation with EhSir2a (Fig. 4D).

EhSir2a co-localizes on microtubular assemblies and regulates their frequency during the cell cycle of E. histolytica

We have earlier reported the diversity of microtubular assemblies (MT) in E. histolytica (Mukherjee et al., 2009; Dastidar et al., 2007). These MT assemblies were visualized and quantitated between 4–8 h after serum synchronization of E. histolytica. Immuno-localisation with anti-HA and anti-Eh β-tubulin antibdies showed that EhSir2a co-localized with the different bipolar, monopolar and multi-polar microtubular structures in E. histolytica. Fig.4E shows a representative localization on two bipolar spindles in a single cell. Results from three independant sets of experiments showed that the number of microtubular assemblies and spindles were significantly reduced in EhSir2a transformants (Fig. 4F). These experiments were carried out after serum synchronisation and the number of MT structures in EhSir2a transformants were compared with control cells. Clearly the 1.6 fold increase in expression of EhSir2a in stable transformants was enough to affect the number of MT assemblies during a single nuclear cycle.

Deacetylase mutants of EhSir2a restore frequency of MT assemblies during the cell cycle

In order to determine whether the observed reduction in MT structures during the amoeba cell cycle was dependent on the deacetylase activity of EhSir2a, we mutagenised and replaced histidine-247 with alanine (H247A) and asparagine-228 with alanine (N228A). Both amino acids are reported to be essential for deacetylase activity (Smith and Denu, 2006; Zhao et al., 2004; North et al., 2003). Stable transformants of EhSir2a-H247A and EhSir2a-N228A were obtained and the expression of the two mutants was compared with EhSir2a-HH transformants. It may be seen that all three transformants expressed similar mRNA (Fig. 5A) and protein levels of EhSir2a (Fig.5B). Ectopically expressed recombinant EhSir2a proteins were analyzed by determining HA expression in the transformants (Fig. 5C and 5D). All three Sir2a transformants showed comparable expression of recombinant EhSir2a (wild type and mutants). EhSir2a-H247A, EhSir2a-N228A, EhSir2a and control transformants were serum synchronized and the MT structures were counted between 2 h – 10 h after serum addition (Fig. 5E). Compared to stable transformants of EhSir2a (wt), both mutant transformants displayed higher numbers of the MT structures. In fact deacetylase mutant Sir2a transformants showed the same frequencies of MT structures as control transformants (Fig.5E). Deacetylase activity of the mutated Sir2a proteins isolated on Ni-NTA agarose beads was comparable with control Ni-NTA agarose beads and negligible compared to wild type EhSir2a on Ni-NTA beads (Fig.5F).

Fig. 5. Deacetylase mutants of EhSir2a restore wild type frequency of MT assemblies.

A and B. Two mutant transformants were generated by mutagenesis of EhSir2a at its deacetylase domain (N228A and H247A). A and B show the mRNA and protein expression levels of the different transformants respectively.

C. The expression of HA in different transformants are shown. Anti-β-tubulin was used as the endogenous control. D. Bar diagram shows the normalized comparison of HA expression.

E. The frequency of MT structures was counted in the transformants over-expressing wild type EhSir2a and its mutant forms and the results are represented as bar diagram.

F. The deacetylase activity of the transformants is shown in arbitrary fluorescence unit (AFU).

Acetylated tubulin is prevalent in the stable microtubular structures like cilia and flagella (Poole et al., 2001). Our results suggest that increased expression of EhSir2a leads to increased deacetylase activity which in turn leads to destabilisation of MT structures and consequently reduced numbers of these structures. Increased numbers of MT structures in mutant Sir2a transformants or control transformants where Sir2a is expressed at lower levels corroborates our finding that EhSir2a deacetylase activity regulates MT structure frequency.

Discussion

Sirtuins are an important group of proteins that epigenetically modify function of their targets by deacetylation. E. histolytica encodes four sirtuin genes that are distributed into two very different phylogenetic clades. Intriguingly, E. histolytica contained archeal/bacterial homologs of Sir2 proteins along with protist/fungal homologs. Horizontal gene transfer has been attributed for the presence of prokaryotic gene homologs in Entamoeba and other protists that share their environment with different microbes and indeed that may be a likely reason for the presence of prokaryotic homologs of sirtuins in amoeba.

Comparison of the amino acid sequences of the amoeba sirtuins showed that the protist/fungal homolog- EhSir2a contained a Zn-finger domain at its N-terminal that was unusual for known sirtuins. In humans, a multi-protein complex consisting of HDAC6 and SIRT2 interacts with microtubules. HDAC6 has the potential to link actin filaments and microtubules through its interaction with formin homology proteins (Destaing et al., 2005). Homologs of HDAC6 were not found in E. histolytica, but the N-terminal Zn finger domain of EhSir2a was similar to the ubiquitin-binding Zn finger domain at the C-terminus of HDAC6 (Pandey et al., 2007). Thus there was a strong possibility that the unique combination of the N-terminal Zn finger domain and deacetylase domain in EhSir2a would allow it to function like the human HDAC6 and SIRT2 protein complex.

In order to test this hypothesis, we used the yeast two hybrid genetic screen to identify amoeba proteins that interact with EhSir2a. Indeed we identified seven interactors for EhSir2a that were cytoskeletal (tubulin, actin like protein), a subunit of proteasome, phospholipase, phosphatase, transcription elongation factor and a hypothetical protein with limited homology to translation initiation factor. The interaction of EhSir2a with amoeba tubulin was validated by biochemical methods. We could not demonstrate the deacetylation of amoeba tubulin due to technical difficulties in isolating polymerized tubulin from E. histolytica. However, our in vitro experiments indicated that not only can it bind amoeba tubulin, but also it can deacetylate tubulin from a heterologous source such as mammalian tubulin. As more results are available from different organisms, tubulin may be identified as one of the conserved non-histone substrates of sirtuins.

We next observed that the nuclear localisation of EhSir2a was primarily on microtubules. In stable transformants of EhSir2a, expression levels of EhSir2a increased 1.6 fold compared to the wild type. Commensurate with this increase in EhSir2a expression we noted a significant decrease in microtubular structures in serum synchronised transformants compared to the control cells. In order to confirm if this effect was a result of increased deacetylase activity of EhSir2a in the transformants, we mutagenised crucial amino acids in the deacetylase activity domain of EhSir2a. The frequency of microtubular structures in both deacetylase mutants was comparable with control cells and much higher than that of EhSir2a transformants. It has been shown that tubulin is hyper-acetylated in stable microtubular structures such as cilia (North and Verdin, 2004; Poole et al., 2001) implying that instability of MT structure is associated with deacetylated tubulin. The destabilization of MT that likely leads to lower frequency of MT structures, may be the result of higher level of tubulin deacetylation by this Sir2 homolog in stable transformants of E. histolytica. Clearly, stability and frequency of microtubular structures are regulated by the deacetylase activity of EhSir2a. Thus apart from inherent mechanisms that regulate chromosome segregation, mitosis and cell division in E. histolytica, deacetylation of tubulin by EhSir2a serves as an additional mechanism that regulates. Based on the identity of the other proteins found to interact with EhSir2a in this study, it is likely that EhSir2a regulates cytoskeletal functions, transcription, translation, protein and lipid metabolism also in E. histolytica. In conclusion our study had identified an important epigenetic mechanism mediated by amoeba sirtuin that regulates microtubular structures in E. histolytica.

Experimental procedures

E. histolytica cell culture and maintenance

The trophozoites of E. histolytica strain HM1: IMSS were grown under axenic conditions in TYI-S33 medium (Diamond et al., 1978) at 37 °C. 24 h – 48 h grown trophozoites were used in all experiments.

Primary sequence analysis of E. histolytica Sirtuins

The amino acid sequences of E. histolytica Sir2 proteins were identified by homology searches (Altschul et al., 1997) in the E. histolytica genome database, TIGR (The Institute for Genome Research) using the amino acid sequence of Sir2 from Saccharomyces cerevisiae as a query sequence. The amino acid sequences thus identified from TIGR database were verified by non-redundant Blast in NCBI protein database. The conserved domains of Sir2 homologs were identified by Pfam and CDD (Conserved Domain Database) searches in (URL: http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml) and SMART tools (URL: http://smart.embl-heidelberg.de/) and used for domain analysis and phylogeny. Sequence alignments were performed with ClustalW (Thompson et al., 1994) using BLOSUM 62 matrix and subsequently edited by Bioedit (URL: http://www.mbio.ncsu.edu/BioEdit/bioedit.html).

Phylogenetic analysis

Amino acid sequences of the conserved domains of Sir2 proteins were aligned in MEGA (v3.1) software package using ClustalW and automatically edited in the software. The positions with gaps were not excluded. Phylogenetic tree was constructed and visualized using p-distance matrix of Neighbor-Joining (NJ) algorithms through MEGA (v3.1). To provide confidence levels for the tree topology and statistical reliability of individual nodes, bootstrap analyses were performed with 1000 replications.

Cloning, mutagenesis, bacterial expression and purification of EhSir2a

Oligonucleotide primers (5’ AAGCTTATGTCATTGCAATTTTCAG 3’ and 5’ GTCGACTGACAATTCCCACTCTAA 3’) specific for EhSir2a were used in PCR reactions to amplify EhSIR2a from genomic DNA or cDNA. PCR products of EhSIR2a were cloned in pBlueScript SK (−) and subsequently tagged with with 3 × HA and 6× His (Honey et al., 2001) at its C-terminus. This construct was sub-cloned into two E. histolytica expression vectors- pJST4 and pNeoCass (Hellberg et al., 2001). Site-directed mutagenesis of EhSir2a-HH was performed by QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). The mutagenised fragments were then sub-cloned in pNeoCass. For expression in E. coli, EhSIR2a coding sequence was cloned in-frame at the 3’ end of the Glutathione-S-transferase gene in pGEX-4T-1. The identity of all the clones and constructs was confirmed by automated DNA sequencing (Applied Biosystems, USA).

Expression of the recombinant EhSir2a was induced by 0.25 mM IPTG in mid-log phase of E. coli BL21 DE3. The culture was grown for another 16 h at 16 °C. Induced cells were collected by centrifugation (5,000 g at 4 °C) and sonicated in lysis buffer (25 mM Tris-HCl pH 7.4, 10 mM NaCl, 5 mM MgCl₂, 5 mM dithiothreitol, 1% Triton X-100, and 1 mM PMSF). After centrifugation (23,500 g at 4 °C), the lysate was incubated with pre-equilibrated Glutathione-Sepharose 4B beads (Amersham Bioscience, Germany) for 2 h. GST-EhSir2a was eluted in a buffer containing 50 mM Tris-HCl pH 8.0, 5 mM MgCl₂ and 10% glycerol in presence of 20 mM reduced glutathione.

A DNA fragment of 443 bp encoding amino acids number 238–383 of EhSir2a (Fig.S2) was cloned in pGEX-5X-2 at EcoR1 and Sal1. After induction and expression, the recombinant protein was purified and used to raise polyclonal antibody in rabbits against EhSir2a. Similarly, the full-length gene of Histone H2A was cloned in pGEX-4T-1 at BamH1 and Sal1 and the purified protein was used to raise polyclonal antibody.

Yeast Two hybrid library screens

The coding sequence of EhSir2a was cloned in pAS2-1 (GAL4 DNA-BD, CLONTECH Laboratories) and used as bait. It was transformed into pJ69-4A [MATa trp1-901 leu2-3,112 ura3-52 his3-200 gal4Δ gal80Δ LYS2::GAL1-HIS3 GAL2-ADE2 met2::GAL7-lacZ] (James et al., 1996). cDNA library of E. histolytica was custom made by Vertis Biotechnologie AG, Germany. Briefly, cDNA fragments larger than 400 bp were used for cloning into pACT2 (GAL4 AD, CLONTECH Laboratories). In primary library, the total number of clones was 1.5 × 10⁶ cfu and the titer of amplified plasmid library was 27,500 cfu/µl. The cDNA library was transformed into the stable transformants of pJ69-4A containing EhSir2a. We screened ~1 × 10⁶ primary transformants. We selected those which expressed high level of galactosidase activity and grew on selective medium like SC plates lacking leucine, tryptophan and adenine (SC-Leu-Trp-Ade or −LTA) or SC plates lacking leucine, tryptophan and histidine (SC-Leu-Trp-His or −LTH) or SC plates lacking leucine, tryptophan and histidine and containing 3.5 mM 3-aminotriazole. These clones were cured of their original bait plasmids and the prey plasmids were retransformed again to check the specificity. The clones which interacted well even after retransformation were sequenced to identify the interactors.

Cell lysis and western blot analysis

Log-phase culture from the transformants of E. histolytica were lysed in a buffer containing 25 mM HEPES-KOH, pH 7.4, 100 mM KCl, 5 mM MgCl₂, 5% glycerol supplemented with complete protease inhibitor (Roche, Germany), 1 mM E-64 (Sigma-aldrich, USA), 1 mM PMSF (Sigma-aldrich, USA) and 1% Triton X-100 (Merck, Germany) at 4 °C. The suspension was centrifuged at 12000 g for 30 min to remove the cellular debris. 30–50 µg of total protein was separated by SDS-PAGE. The western blots were hybridized overnight at 4 °C with the mouse monoclonal anti-HA (1:1000, 12CA5; Roche, Germany), anti-Phosphoglycerate kinase (PGK) (1:1000; Molecular Probes, USA) and rabbit polyclonal Eh β-tubulin (1:500), anti-EhSir2a (1:500) or anti-Histone H2A (1:200) followed by the hybridization with HRP conjugated secondary antibodies (1:5000; Sigma, USA). The signals were detected by chemi-luminescence using ECL kit (Roche, Germany).

Cell fractionation

Log phase grown E. histolytica cells were harvested and washed with PBS. The cells were then lysed in extraction buffer (10 mM HEPES-KOH (pH 7.2), 24 mM KCl, 10mM MgCl₂) supplemented with complete protease inhibitor (Roche, Germany), 1 mM E-64 (Sigma, USA), 1 mM PMSF (Sigma, USA), 2 mM DTT, and 0.1% NP-40. The lysate was centrifuged through a 0.8M sucrose cushion at 6000 g for 20 min. E. histolytica nuclei, recovered in the pellet were washed with extraction buffer supplemented with protease inhibitors. The protein samples prepared from nuclei and cytosol were separated by SDS-PAGE followed by western hybridisation with antibodies against EhSir2a, PGK (Molecular Probes, USA) and Histone H2A. PGK was used as a cytoplasmic control and Histone H2A for identifying the nuclear fraction.

Immuno-fluorescence and confocal microscopy

E. histolytica trophozoites were grown on coverslips at 37 °C in 24 well plates and fixed directly on the coverslips with warm 3.7% formaldehyde for 20–30 min, quenched with 2% Glycine and permiabilized with 0.1% Triton X-100 for 10 min. After blocking with 2% BSA, the fixed cells were stained with monoclonal anti-HA antibody, 12CA5 (1:150; Roche, Germany) followed by Alexa Fluor 488-conjugated anti-mouse secondary antibody (1:2500; Molecular Probes, USA), anti-EhSir2a antibody (1:200) followed by TRITC-conjugated anti-rabbit secondary antibody (1:200; Jackson Laboratories, USA), anti-Eh β-tubulin antibody (1:200) followed by TRITC-conjugated anti-rabbit secondary antibody (1:200; Jackson Laboratories, USA). DNA was stained with DAPI (0.1–0.2 µgml⁻¹) for 10 minutes. Kaiser’s glycerol gelatin (Merck, Germany) was used as the bleaching preservative and mounting medium for the coverslips. For confocal microscopy, images were acquired with a 63× Plan-apochromat 1.4 oil DIC objective (numerical aperture 1.4) in a Zeiss LSM 510 Meta confocal microscope equipped with 488 nm Argon laser and 543 nm He/Ne laser. Images were analyzed by LSM Meta 510 software package (Carl Zeiss, Germany).

For counting the microtubular structure, the transformants were fixed with formaldehyde and stained with polyclonal anti-Eh β-tubulin. 300 fixed cells were counted for each set (three independent sets) under 40× oil objective (numerical aperture 1.3) in a Zeiss Axiovert 200M Flourescence microscope.

In vitro binding assay

Log phase cells of EhSir2a-HH and control transformants were fixed with 1% formaldehyde for 20 min at room temperature, followed by the addition of 0.5 M glycine. Cells were chilled, harvested and, washed thoroughly with cold 1× PBS and finally resuspended in lysis buffer containing 20 mM Tris-HCl; pH 8.0, 40 mM NaCl, 5 mM MgCl₂, 10% glycerol, 0.2% Triton X-100 and protease inhibitors. After centrifugation at 13, 000 g the supernatant was allowed to bind with Ni-NTA agarose beads in presence of 1 mM GTP and 500 µM NAD⁺ for 4 h at 4 °C. The beads were washed with the lysis buffer containing 20 mM imidazole. 2× SDS PAGE sample buffer was added to the beads and boiled for 10 mins. Proteins were separated through SDS-PAGE and subsequently analysed by Western blot.

Tubulin deacetylase assay (TDAC)

Goat brain tubulin was prepared by two cycles of assembly-disassembly in PEM buffer (0.05 M Pipes, 1 mM EGTA, 0.5 mM MgCl₂, pH 6.9 at 25 °C) in presence of 1 mM GTP followed by two cycles in 1 M glutamate buffer (Hamel and Lin, 1981). GST-EhSir2a was incubated with glutathione sepharose 4B beads followed by the incubation with goat brain tubulin in deacetylase buffer (50 mM Tris-HCl pH 9.0, 4 mM MgCl₂, 0.2 mM DTT, 500 µM NAD⁺ and 1mM GTP) at room temperature for 2 h. The reaction was stopped by adding 2× sample buffer to the mixture. For the negative control, GST was used. The samples were analysed by SDS-PAGE and the corresponding western blots were hybridized overnight at 4 °C with monoclonal anti-acetylated- α-tubulin (1:1000; Sigma) and anti-a-tubulin (1:1000; Sigma) followed by incubation with HRP conjugated secondary anti-mouse antibody. The signals were detected by chemi-luminescence using ECL kit (Roche, Germany).

Deacetylase assay

Fluorescent activity assay/drug discovery kit was purchased from Biomol Research Laboratories and used for deacetylase assay of EhSir2a according to the manufacturer’s instructions. Log phase cells from the stable transformants of EhSir2a, mutants and control were, harvested, washed with cold 1× PBS and lysed in a buffer containing 50 mM Tris-HCl; pH 8.0, 40 mM NaCl, 5 mM MgCl₂, 10% glycerol, 1% Triton X-100 supplemented with 1× complete protease inhibitors, 1 mM PMSF and 1 mM E-64. After centrifugation at 13000 g, 2.5 mg of total proteins from the supernatant of each transformant was allowed to bind with 400 µl of Ni-NTA agarose beads (Qiagen) for 4 h at 4 °C. The beads were washed with the lysis buffer containing with 20 mM imidazole and with HDAC assay buffer. The corresponding beads were then used for HDAC assay. Nicotinamide (Sigma-aldrich, soluble in water), Splitomicin (Sigma-aldrich, in DMSO) and Sirtinol (Calbiochem, in DMSO) were added to the reactions at the indicated concentrations with all components of the reaction in the absence of NAD⁺ for 5 min at room temperature. The enzymatic reaction was started by adding 500 µM of NAD⁺ and incubated further for 2 h at 37 °C. The fluorescence of the reactions was measured in a microplate (POLARstar OPTIMA, BMG LABTECH, Excitation at 355 nm, Emission at 460 nm, gain = 50).

Supplementary Material

Supp Fig 1. Fig. S1. Identification of eukaryotic and prokaryotic conserved residues.

ClustalW alignment of the conserved core domains from Eh Sirtuins with that of Sir2 from Bacillus subtilis (Bs Sir2) and Human SIRT2 (Hs SIRT2). GAG, PTI, HG domains and FGE loop are shown in different boxes. The black and gray shaded residues indicate identity and similarity respectively. Dashes indicate gaps introduced for optimal alignment.

Supp Fig 2. Fig. S2. Alignment of E. histolytica Sirtuins.

ClustalW alignment of EhSir2a with EhSir2b,c and d amino acid sequences shows scattered identity of a few amino acids over the entire length except at the GAG domain. The highlighted region is 238–383 amino acids at the C-terminal against which the polyclonal antisera was raised and used in the study. Alignment data :Alignment length : 436, Identity (*) : 47 is 10.78 % , Strongly similar (:) : 46 is 10.55 % ,Weakly similar (.) : 35 is 8.03 %, Different : 308 is 70.64 %.

Supp Fig 3. Fig. S3. Polyclonal anti-α-tubulin hybridizes with β-tubulin.

A. Amino acid sequences of α- and β-tubulin were aligned. The boxes show the identical and similar amino acid sequences.

B. BL21 DE3 containing Empty vector/ α- / β-tubulin were induced by IPTG and analyzed by SDS-PAGE. Picture was generated after deleting the gaps between the three lanes. Coomassie stained gel is shown on the left and the western blot hybridized with polyclonal anti-β-tubulin on the right. Similar results were obtained with anti-α-tubulin antibody.