Summary
Factors that regulate the microtubule cytoskeleton are critical in determining cell behavior. Here we describe the function of a novel protein that we term E-like based on its sequence similarity to the tubulin-specific chaperone cofactor E. We find that upon overexpression, E-like depolymerizes microtubules by committing tubulin to proteosomal degradation. Our data suggest that this function is direct and is based on the ability of E-like to disrupt the tubulin heterodimer in vitro. Suppression of E-like expression results in an increase in the number of stable microtubules and a tight clustering of endocellular membranes around the microtubule-organizing center, while the properties of dynamic microtubules are unaffected. These observations define E-like as a novel regulator of tubulin stability, and provide a link between tubulin turnover and vesicle transport.
Introduction
Microtubules (MTs) are polarized polymers of α/β tubulin heterodimers and are responsible for a wide variety of vital cellular functions such as the formation of a bipolar spindle at mitosis, global organization of endocellular membranes and cell shape, cell motility and maintenance of cell polarity. MTs undergo alternating phases of growth and shrinkage with sudden transitions between the two. Transitions from growth to shrinkage are called catastrophes and transitions in the reverse direction are termed rescues (Desai and Mitchison, 1997; Heald and Nogales, 2002). This property is known as dynamic instability and allows depolymerizing and polymerizing MTs to coexist in the same population. In vitro, the parameters of dynamic instability (i.e. rates of polymerization, depolymerization and the frequency of catastrophe and rescue) are largely governed by the concentration of pure tubulin (Walker et al., 1988). However, in vivo, MT behavior is modulated both by local tubulin concentration and by interaction with a large number of effectors. These modulators belong to two main families of MT regulators and modulate the dynamic behavior of MTs. One consists of MT-associated proteins (MAPs), which stabilize MTs by promoting an increase in MT elongation rates and suppressing the rate of catastrophes. The other consists of different MT-destabilizing proteins termed catastrophepromoting factors because of their ability to promote transitions from growth to shrinkage (Cassimeris and Spittle, 2001). Op18/stathmin proteins and members of the Kin I kinesin subfamily of kinesins (such as XKCM1) are cell cycle regulated molecules that destabilize MTs by sequestering tubulin dimers or by promoting catastrophe at MT plus ends (Belmont and Mitchison, 1996; Desai et al., 1999; Walczak et al., 2002; Walczak et al., 1996). MTs can also be destabilized by the action of katanins, severing proteins that cut MTs at internal sites and generate an increase in the number of free ends (Quarmby, 2000).
Cofactor E is one of five tubulin-specific chaperones (cofactors A-E) which bind tubulin folding intermediates produced via interaction of de novo synthesized polypeptides with the cytosolic chaperonin CCT. The tubulin/cofactor complexes participate in the formation of assembly competent tubulin heterodimers by associating to form a multimolecular complex from which newly formed tubulin heterodimers are released upon GTP hydrolysis by β-tubulin (Tian et al., 1996; Tian et al., 1997). In addition to the de novo folding of tubulin dimers, there is compelling evidence that cofactor E also influences MT dynamics based on its ability to act both as part of a GTP-activating complex on native tubulin (in combination with cofactors C and D) (Tian et al., 1999) and to sequester α-tubulin subunits from native dimers in vitro. In the latter reaction, the heterodimer is disrupted and the free subunit decays to a non-native state. These in vitro data are in accord with the phenotype associated with overexpression of cofactor E in cultured cells, which results in the induction of massive MT depolymerization (Bhamidipati et al., 2000). Consistent with its predicted multifunctional role in the biogenesis and regulation of the tubulin heterodimer, cofactor E deletion mutants in both yeast and Arabidopsis exhibit various defects in MT stability and proper chromosome segregation, the latter presumably owing to the inability of spindle MTs to form (Grishchuk and McIntosh, 1999; Hoyt et al., 1997; Radcliffe et al., 1999; Steinborn et al., 2002). The importance of proper cofactor E functioning has been recently highlighted by the discovery of a devastating autosomal recessive disease (HRD or Sanjad-Sakati, Kenny-Caffey syndromes) in humans and of a progressive motor neuropathy in mice caused by genetic lesions in the gene encoding mammalian cofactor E (Bommel et al., 2002; Martin et al., 2002; Parvari et al., 2002). Despite the tissue-specific variability of the phenotypes associated with the human or mouse syndromes, presumably a result of differences in the penetrance of the causative mutations, it is evident that in both cases the defect can be ascribed to perturbed MT stability and polarity.
Although there is an abundance of information regarding regulatory mechanisms of MT dynamics and tubulin biogenesis in vitro and in vivo, relatively little is known about the regulation of tubulin turnover. Here we describe the identification of a protein that we term E-like based on its overall sequence similarity to human cofactor E. We find that upon overexpression in cultured cells E-like induces massive depolymerization of the MT cytoskeleton and disruption of the Golgi complex. We show that this MT depolymerization is mediated by direct misfolding of the tubulin heterodimer, which as a result is targeted to the proteosome for degradation. Notably, suppression of E-like expression by small interfering RNA (siRNA) results in an increase in the number of stable MTs and a reorganization of endocellular membranes without altering the properties of dynamic MTs. These data implicate E-like involvement in the regulation of tubulin turnover and provide a link between MT stability and vesicle distribution.
Materials and Methods
Cloning of E-like
Two Caenorhabditis elegans genes related to cofactor E were initially identified in GenBank. The more distantly related gene (u64589) was used to search human EST databases and a full-length cDNA sequence encoding E-like was assembled using the fragment assembly programs of Staden (Dear and Staden, 1991). A 5′ primer (GGTGGTATGGATCAACCTAGTGGAAGAAG) and a 3′ reverse primer (GGTGGTTTATTTTGTTTTGGATTCCACGTA) were used to amplify the 1.2 kb E-like coding region from human testis cDNA (Clontech). The human E-like sequence has been deposited in GenBank with accession number AY398644.
Plasmid construction
pGFP-E-like and pET23-E-like were constructed by insertion of the E-like coding region into pEGFP-C3 (Clontech) and pET23b (Novagen) vectors. In the case of pet23b, the PCR-generated fragment was inserted into the vector at the NdeI and SalI restriction sites; in the case of pEGFP-C3, the fragment was inserted between the SacI and SalI sites.
Northern blot analysis
Northern blot membranes containing poly(A)+ RNA from multiple human tissues (Clontech) were probed with 32P-labeled E-like cDNA, generated by random priming (High Prime, Roche). Blots were washed to a final stringency of 0.1× SSC, 68°C.
Protein expression and purification
Bovine brain tubulin, the cytosolic chaperonin CCT and cofactors B, C, D and E were purified as described previously (Tian et al., 1996; Tian et al., 1997). Untagged wild-type E-like was purified from a bacterial lysate of BL21 (DE3) cells on a Q15 anion exchange column developed using a gradient of 10-250 mM phosphate buffer, pH 7.3, containing 1 mM EGTA, 1 mM MgCl2, 10 mM KCl and 1 mM DTT. Fractions containing E-like were applied to a Superdex 200 gel filtration column run in 10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EGTA, 1 mM MgCl2 and 1 mM DTT. Bovine E-like was purified from an extract of testis tissue using the same chromatographic dimensions described for the purification of cofactor E (Tian et al., 1996). The protein was detected throughout the purification procedure by western blotting, using an anti-E-like antibody.
Generation of an E-like antiserum
Purified bacterially expressed E-like protein was used to raise antisera in rabbits (Cocalico Biologicals). The specificity of these sera was tested by their ability to recognize the cognate protein and antigen by western blotting of the recombinant protein and of whole cell extracts.
In vitro tubulin folding and refolding reactions and GTPase assay
In vitro α- and β-tubulin folding reactions, tubulin refolding reactions and tubulin-GAP assays were performed as described previously (Bartolini et al., 2002).
In vitro tubulin heterodimer disruption assay
Purified E-like or cofactor E (0.5-25.0 μM) were incubated in folding buffer (20 mM MES, pH 6.8, 50 mM KCl, 1 mM MgCl, 1 mM EGTA) with purified bovine brain tubulin (5 μ 2 M) in the presence of 1 mM ATP and GTP for 1 hour at 30°C. 35S-labeled tubulin was isolated on DEAE-Sephacel as described (Tian et al., 1997). Additional unlabeled bovine brain tubulin was added to reach the desired tubulin concentration, followed by incubation with various amounts of E-like at 30°C for 1 hour. At the end of the incubation, each sample was divided and resolved on polyacrylamide gels under both non-denaturing and denaturing conditions (Zabala and Cowan, 1992).
In vitro tubulin polymerization assay
Purified porcine brain tubulin (11 μM) was incubated at 30°C for 1 hour with purified recombinant E-like protein (2-11 μM) in PEM buffer (100 mM PIPES, pH 6.8, 2 mM EGTA and 1 mM MgSO4) supplemented with 1 mM ATP and GTP. Samples were stored briefly on ice. Microtubule assembly from axoneme fragments was examined by video-enhanced DIC microscopy 5-10 minutes after warming the samples to 37°C using methods described previously (Howell et al., 1999). The numbers of MTs present at each axoneme end were counted from the video recordings. The number of MTs at each axoneme end is steeply dependent on tubulin concentration (Walker et al., 1988).
Transfection and immunoblot analyses
293T or HeLa cells grown on coverslips or in multi-well dishes were transfected using Fugene 6 (Roche). For immunoblot analyses, cells were harvested, washed with PBS and lysed in ice-cold buffer (50 mM Tris-HCl, pH 7.5, 250 mM NaCl, 0.1% Triton X-100, 1 mM EDTA) containing protease inhibitor cocktail (Roche). Extracts were cleared by centrifugation at 14,000 g and samples containing 25 μg protein were resolved by SDS-PAGE. In experiments involving inhibition of the proteosome pathway, cells were treated 12-18 hours post-transfection with the specific inhibitor MG132 (10 μM) (Boston Biochemicals) for 8 hours at 37°C. α-tubulin and β-actin were detected on western blots using mouse monoclonal antibodies (Sigma).
Immunofluorescence
DNA- or siRNA-transfected HeLa cells were fixed with 4% paraformaldehyde in PBS for 20 minutes at room temperature, washed in PBS and permeabilised in 0.2% Triton X-100 for 5 minutes. Primary and secondary antibodies were diluted in 3% BSA, applied to fixed cells and incubated for 1 hour at 37°C. Primary antibodies included a monoclonal anti-α-tubulin and a monoclonal anti-acetylated tubulin (Sigma) for the detection of stable microtubules. Antibodies specific to giantin, calnexin and lamp-2 were used to visualize Golgi cisternae, endoplasmic reticulum membranes and lysosomal vesicles, respectively. Detection was via Texas Red or FITC-conjugated secondary antibodies (Jackson ImmunoResearch). For microtubule depolymerization experiments, cells were transfected with plasmids encoding E-like and 24 hours later incubated with 10 μM nocodazole for 1 hour prior to fixation. In nocodazole recovery experiments, siRNA transfected cells were treated 72 hours post transfection with 10 μM nocodazole for 1 hour, restored to drug-free medium, incubated for various times and fixed in 4% paraformaldehyde. In tubulin extraction experiments, siRNA-transfected cells were incubated with 0.2% Triton X-100 in MT stabilizing buffer (130 mM HEPES, pH 6.9, 2 mM MgCl2, 10 mM EGTA) for 3 minutes at 37°C, washed in the same buffer lacking detergent, incubated for 30 minutes and then fixed and processed for immunofluorescence. Slides were observed using a Zeiss Axiophot fluorescence microscope with a Plan-Neofluar 63× /1.25 objective. Images were captured using a Zeiss-Axiocam digital camera linked to Axiovision 2.0.5 software.
Microtubule binding assays
Bovine brain tubulin purified by passage through phosphocellulose (Tian et al., 1997) was incubated in the presence of increasing concentrations of taxol (0.1-10.0 μM) in tubulin buffer (80 mM PIPES, pH 6.8, 1 mM MgCl2, 1 mM EGTA, 1 mM GTP, pH 7 and 0.1 μg/μl BSA) to induce microtubule polymerization. 5μl aliquots of 35S-labeled in vitro translated GEF-H1 and E-like proteins were incubated with 50 μg taxol-polymerized tubulin for 10 minutes at 37°C, and loaded onto 50% sucrose cushions which were centrifuged at 150,000 g for 20 minutes at 37°C. Equivalent amounts of supernatant and pellet were analyzed by SDS-PAGE.
siRNA transfection
Three siRNA duplexes designed to target different portions of the mRNA encoding E-like (AACCTCCCAAGTGTACTAGTGTT, AACAGTGTCTTGTCCTTCTAT and AAGTAGAAGTCCACTTTAACG) were synthesized (Dharmacon Research; Ambion). HeLa cells were transfected using the oligofectamine transfection reagent (Invitrogen). For analysis of time points beyond 48 hours, cells were trypsinized at 48 hours post transfection and plated at lower density to avoid cell-cell contact growth inhibition.
Microtubule dynamics
HeLa cells grown on round coverslips were either transfected with pEGFP-α-tubulin (Piehl and Cassimeris, 2003) or microinjected (Piehl et al., 2004) with 50 μg/ml pEGFP-α-tubulin plasmid plus 0.1 μM E-like siRNA (needle concentration). Cells were maintained at 37°C in a POC minicultivation chamber (Zeiss) in combination with an objective heater, and imaged using a Zeiss LSM 510 Meta confocal microscope. Time-lapse image series of interphase cells were obtained every 4 seconds (typically 20 images/cell) using Metamorph imaging software (version 4.6, Universal Imaging) (Piehl and Cassimeris, 2003).
Results
A protein similar to cofactor E
We searched GenBank for sequences related to tubulin specific chaperones, and assembled a full-length cDNA sequence encoding the human homologue of a protein related to cofactor E that we termed E-like. Fig. 1A shows the alignment between human cofactor E and E-like (http://www2.igh.cnrs.fr/bin/align-guess.cgi). Sequence comparison using the SeqWeb sequence analysis program (http://ranger.med.nyu.edu/gcg-bin/seqweb.cgi) showed that the human, fly and worm genome all contain this gene (Fig. 1B). However, no discernible E-like homologues exist in the fully sequenced organisms Arabidopsis thaliana, Schizosaccharomyces pombe or Saccharomyces cerevisiae. This protein may therefore have a specific function required only in animal cells. Alternatively, analogues that have no significant sequence similarity may exist in plants, fungi or unicellular organisms. Using fold recognition algorithms (Motif at http://motif.genome.ad.jp/), we found that both proteins share leucine-rich repeats (LRR) followed by an ubiquitin-like motif (UBL) (Fig. 1C). The role of these sequence motifs in cofactor E is unknown, although they are considered to be protein-protein interaction domains (Kobe and Kajava, 2001; Upadhya and Hegde, 2003). In addition, cofactor E (but not E-like) has a CAP-gly domain (Pierre et al., 1992; Pierre et al., 1994), which it shares with cofactor B (Tian et al., 1997) and several other proteins involved in the organization of MTs and in the transport of vesicular membranes along the cytoskeleton. There is evidence that this domain contributes to α-tubulin binding (Feierbach et al., 1999; Radcliffe and Toda, 2000).
E-like, a protein related to tubulin-folding cofactor E. (A) Amino acid sequence comparison showing 23% identity between human E-like (El) and cofactor E (Cof E). (B) Evolutionary relationship between currently known cofactor E-related sequences; fly, Drosophila melanogaster; celeg, Caenorhabditis elegans; atha, Arabidopsis thaliana; yeast, Saccharomyces cerevisiae; spombe, Schizosaccharomyces pombe; leisch, Leishmania donovani. (C) Schematic of domains identified in cofactor E and E-like. CAP-gly, glycine-rich cytoskeleton-associated protein domain; LRR, leucine-rich repeat sequence; UBL, ubiquitin-like domain. (D) Northern blot showing distribution of E-like expression in various human tissues. Location of size markers is indicated on the left.
We assessed E-like tissue distribution in humans by northern blot analysis of RNA from multiple human tissues (Fig. 1D). The data show that one major E-like-specific transcript of ∼6.0 kb is abundantly expressed in testis, but is also present in several other tissues at a much lower level. This pattern of expression was confirmed by western blot analyses using a specific anti-E-like antiserum (data not shown).
E-like and cofactor E are functionally distinct
Given the sequence similarity between E-like and cofactor E we decided to compare the functions of these two proteins in vitro and in vivo. Specifically, we sought to determine whether E-like could substitute for cofactor E in the assembly of the tubulin heterodimer or enhance the GTPase activity of native tubulin triggered by the α/β tubulin/cofactor C/cofactor D complex (Tian et al., 1999). We therefore carried out in vitro folding reactions in which either labeled, unfolded α- or β-tubulin was presented to CCT by sudden dilution from denaturant in the presence of ATP, GTP and cofactors B, C and D, with or without cofactor E or E-like. The reaction products were analyzed by non-denaturing gel electrophoresis (Fig. 2A). The data show that, in contrast to folding reactions that included cofactor E, no native tubulin heterodimers were produced in reactions in which E-like was substituted for cofactor E. Similarly, we found that when native tubulin heterodimers are refolded in vitro by the action of cofactors C, D and E in the presence of radiolabeled GTP (Bartolini et al., 2002), the nucleotide was incorporated non-exchangeably into the refolded tubulin only when cofactor E was included in the reaction mix (Fig. 2B). This was the case even when a tenfold molar excess (relative to cofactor E) of E-like was included in the reactions. We conclude that E-like does not participate in the tubulin heterodimer assembly reaction. The failure of E-like to substitute for cofactor E in these reactions is not due to a defect in the recombinant protein, as the same result was obtained using E-like purified from bovine testis (data not shown).
Functional comparison between E-like and cofactor E. (A) E-like cannot substitute for cofactor E in tubulin folding reactions. 35S-labeled, unfolded α-tubulin (left panel) or β-tubulin (right panel) probes generated by expression in E. coli were presented to CCT by sudden dilution from denaturant in the presence of ATP and GTP plus the indicated cofactors (B,C,D) and either cofactor E (E) or increasing concentrations of E-like (El). 1×, 2× and 4× denote the fold molar excess of E-like compared with cofactor E. (B) E-like cannot substitute for cofactor E in tubulin refolding reactions. Tubulin refolding reactions containing various tubulin-specific chaperones were carried out in the presence of [32P]GTP as described (Bartolini et al., 2002). In A and B, reaction products were resolved by non-denaturing gel electrophoresis. Migration positions of various tubulin-containing species are shown. Note the absence of labeled native tubulin heterodimers (tub) in lanes in which E-like is substituted for E. Dβ, cofactor D/β-tubulin; Bα, cofactor B/α-tubulin. (C) E-like does not stimulate the tubulin-GAP activity of cofactors C and D. The tubulin-GAP activity of cofactors C and D was measured as described (Tian et al., 1999) either alone (open circles), in the presence of cofactor E (filled diamonds) or in the presence of E-like (crosses).
Together, cofactors C and D have the ability to stimulate GTP hydrolysis by tubulin, thereby acting as GTPase activating proteins (GAPs). This GAP activity results in the release of native tubulin from cofactors as the last step in the de novo formation of the tubulin heterodimer. There is also persuasive evidence that this reaction plays a role in the regulation of tubulin polymerization (Tian et al., 1999). Because cofactor E enhances the tubulin-GAP activity of cofactors C and D, we tested E-like in this assay. We found that in contrast to cofactor E, E-like had no effect on the tubulin GAP activity of cofactors C and D (Fig. 2C). This was true for both recombinant E-like and the authentic protein purified from tissue (data not shown).
Overexpression of E-like leads to microtubule destruction
To test the effect of overexpressing E-like in cultured cells and to analyze its subcellular distribution, we transfected HeLa cells with EGFP-tagged E-like (El-EGFP) (Fig. 3A). We found that the transgene was present in both the nuclei and cytoplasm of transfected cells. This was also the case in cells transfected with an untagged construct (data not shown). Although the significance of the appearance of a portion of overexpressed E-like in the nucleus is unclear, in all cases overexpression of El-EGFP (but not EGFP alone) resulted in a dramatic reduction in the number of MTs in the cell as assayed by α-tubulin antibody staining (Fig. 3A, panels a-d). Because the disappearance of MTs correlated with the level of expression of the transgene, it seemed likely that this phenotype reflected an effect of E-like on either tubulin or MTs. To test this idea, we performed a parallel experiment in which El-EGFP transfected cells were treated with the MT depolymerizing agent nocodazole 1 hour before fixation. We found that in the presence of the drug, E-like overexpression resulted in an almost total loss of tubulin staining, even under conditions where the level of expression of the transgene was relatively low (Fig. 3A, panels e and f). These data show that the tubulin-obliterating effect of E-like overexpression is enhanced by nocodazole-induced depolymerization. If E-like were acting only on MTs, we would not expect any change in the intensity of α-tubulin fluorescence upon nocodazole treatment; on the other hand, if E-like interacts with free tubulin dimers, treatment with nocodazole would increase the size of the tubulin pool available for interaction, resulting in a greater loss of α-tubulin staining. As the latter is what we observed, we infer that E-like affects the tubulin dimer rather than just its polymer.
E-like overexpression leads to microtubule depolymerization and Golgi membrane fragmentation. (A) Double-label immunofluorescence of HeLa cells transfected with EGFP alone (a,b), E-like-EGFP (c,d) or E-like-EGFP followed by a brief incubation with nocodazole (e,f). An anti-α-tubulin antibody recognized by a Texas Red-conjugated secondary antibody was used for the detection of microtubules and tubulin. Arrows in panels c and e highlight MT destruction in cells overexpressing E-like. (B) HeLa cells transfected with E-like-EGFP (El-EGFP) and visualized with a Texas Red-conjugated antibody recognizing an anti-β-cop antibody to detect Golgi stacks. Golgi membranes in transfected cells are arrowed. (C) E-like is not a microtubule binding protein. 35S-labeled in vitro translated hGEF-H1 (GEF) and E-like (El) were incubated in the presence of taxol-stabilized microtubules and sedimented through a sucrose cushion. Pellet (P) and supernatant (S) fractions containing equal amounts of protein were resolved by SDS-PAGE. Equivalent aliquots of the starting material are shown in the first two lanes. Molecular mass standards are shown on the left. Bar, 10 μm.
MT disruption has a devastating effect on the centrosomal localization of the Golgi complex, which in consequence undergoes a reversible dispersal throughout the cytoplasm (Cole et al., 1996). To test whether Golgi fragmentation was a consequence of E-like-induced MT depolymerization, we examined the subcellular organization of the Golgi complex. We found that in E-like expressing cells, Golgi membranes lost their organization in cisternal stacks around the nucleus, and redistributed to an ill-defined number of peripheral sites scattered throughout the cytoplasm (Fig. 3B). Taken together, these observations define E-like as a powerful MT destabilizing factor when overexpressed in cultured cells.
Regulators of MT stability often exert their function by associating directly with MTs either along their length or in close proximity to MT ends, where changes in rates of catastrophe or growth are high. Because we did not observe any clear association between overexpressed E-like and the MT cytoskeleton in HeLa cells, we decided to test whether E-like could bind to MTs or MT ends in vitro. To this end, we did a binding experiment in which 35S-labeled E-like was incubated with taxol polymerized MTs, which were then subjected to centrifugation (Fig. 3C). A reaction with GEF-H1, a MT binding protein (Krendel et al., 2002), was included as a positive control. We found that unlike GEF-H1, which mostly partitioned to the pellet, E-like was largely associated with the supernatant, suggesting no association with MTs under these conditions. These observations are consistent with a role for E-like in inducing catastrophic MT depolymerization by sequestering or degrading the pool of free tubulin available for incorporation into MTs rather than by directly destabilizing MTs.
E-like induces tubulin degradation
To assess whether levels of tubulin heterodimers were affected by E-like overexpression, we transfected 293T cells with El-EGFP or EGFP alone, and then assayed for their content of endogenous α-tubulin. 293T cells were selected for these experiments because of their very high (about 90%) transfection efficiency. Western blot analyses of matched amounts of cellular extracts from transfected cells using an α-tubulin antibody showed that levels of endogenous tubulin were substantially reduced in E-like overexpressing cells compared with controls (Fig. 4A, lanes marked –). The principal mechanism for controlled proteolysis in eukaryotic cells is the pathway of proteosomal degradation. We therefore explored the possibility that ablation of α-tubulin induced by E-like overexpression might depend on E-like-induced commitment of tubulin dimers to degradation by the proteosomal machinery. Western blot analyses of the levels of tubulin in E-like transfected cells treated with a specific proteosome inhibitor (MG132) prior to lysis showed that the level of endogenous tubulin increased in the presence of the inhibitor compared with its untreated counterpart (Fig. 4A, lanes marked +). The ability of the inhibitor to partially restore the level of tubulin suggested that the loss of MTs was a result of E-like-mediated targeting of tubulin to the proteosome for degradation. In that event, we would expect that treatment with this proteosome inhibitor would also significantly reduce the loss of tubulin staining in E-like expressing cells stained with an α-tubulin antibody. Indeed, addition of this inhibitor resulted in a visible enhancement of tubulin staining in HeLa cells expressing El-EGFP compared with their untreated counterparts (Fig. 4B). Similar results were obtained with a second proteosome inhibitor (MG101) (data not shown). These observations strongly suggest that when overexpressed E-like depolymerizes the MT cytoskeleton because of its ability to induce disruption of the tubulin heterodimer (see below).
E-like targets tubulin for degradation via the proteosome pathway. (A) 293T cells were transfected with E-like-EGFP (El-EGFP) or EGFP alone (EGFP). Twenty-four hours following transfection, cells were incubated either in the presence (+) or absence (–) of the proteosome inhibitor MG132 (P.I.). Cell lysates were analyzed by SDS-PAGE and by western blotting using a α-tubulin antibody for the detection of endogenous tubulin dimers and a β-actin antibody as a loading control. (B) HeLa cells were transfected with the identical constructs described in A. Arrows indicate tubulin staining in cells overexpressing E-like. Fixed cells were examined by indirect immunofluorescence for the detection of tubulin and microtubules using a Texas Red-conjugated anti-α-tubulin antibody. Bar, 10 μm.
E-like disrupts the tubulin heterodimer and prevents MT assembly in vitro
When native tubulin α/β heterodimers are incubated in the presence of a molar excess of cofactor E, the α-subunit is sequestered and the free β-subunit decays to a non-native state (Tian et al., 1997). To determine whether E-like has a direct effect on the stability of the tubulin heterodimer, we did an in vitro experiment in which purified cofactor E or E-like was incubated with tubulin dimers and the reaction products resolved on polyacrylamide gels under both non-denaturing and denaturing conditions (Fig. 5A). We found that in samples in which increasing amounts of cofactor E or E-like were present in the reaction mix, the levels of native tubulin were successively reduced as shown by the diminished intensity of the band corresponding to native tubulin detected under non-denaturing conditions. Because the same samples showed no reduction in the amounts of tubulin dimer when analyzed by SDS-PAGE, degradation of tubulin can be ruled out. To test whether obliteration of free native tubulin dimer was indeed a consequence of E-like-mediated tubulin misfolding, a reaction was carried out in which purified tubulin heterodimers 35S-labeled in their α-subunit were incubated with E-like and the reaction products analyzed by non-denaturing electrophoresis (Fig. 5B). We found that no product other than tubulin itself was detectable under these conditions, suggesting that no stable tubulin/E-like complexes were being formed in solution. However, with increasing concentrations of E-like, there was a dramatic reduction in the level of native tubulin dimer, with a corresponding accumulation of radioactivity at the origin of the gel, presumably reflecting E-like induced tubulin misfolding.
E-like disrupts the tubulin heterodimer in vitro. (A) E-like mediated disruption of the tubulin heterodimer in vitro. Purified E-like (El) and cofactor E (E) were incubated with depolymerized bovine brain tubulin (tub). The reaction products were resolved by non-denaturing (N.D.) gel electrophoresis (upper panel) and SDS-PAGE (lower panel). Note that the absence of a band corresponding to purified cofactor E on the native gel is a result of the migration of this protein towards the cathode under these conditions. 0.1×, 0.3×, 1×, 2× and 5× refer to relative molar concentrations of either E-like or cofactor E compared with tubulin. (B) E-like induces aggregation of tubulin in vitro. [35S]tubulin was incubated with the indicated amounts of E-like (El) and the reaction products were analyzed under non-denaturing conditions. (C) E-like prevents MT assembly in vitro. Micrographs (magnification 1300×) show MTs (arrowed) assembled from axoneme fragments 5-10 minutes after warming samples to 37°C. Pre-incubation with 2 μM E-like significantly reduced the assembly of 11 μM tubulin. MTs did not assemble after preincubation with higher concentrations of E-like. The histogram shows the number of MTs assembled per axoneme end for 11 μM tubulin preincubated for 1 hour with the indicated concentrations of E-like. Data shown are the mean±s.d. from 100 axoneme ends per condition.
To confirm that E-like disrupts tubulin dimers and renders them unable to polymerize, we incubated various concentrations of E-like protein with 11 μM purified tubulin for 1 hour and then tested whether samples could polymerize from axoneme fragments. As little as 2 μM E-like significantly reduced MT assembly from axonemes, reducing both the number and length of MTs (Fig. 5C). No MT assembly was observed at higher concentrations of E-like. By contrast, tubulin incubated for 1 hour without E-like showed robust assembly from both ends of the axonemes. We conclude that E-like alone can destroy the ability of tubulin dimers to polymerize.
E-like siRNA affects MT stability and induces membrane clustering
We examined the effect of depletion of E-like in cultured cells by introducing small interfering RNA (siRNA) duplexes. An oligonucleotide duplex with a scrambled sequence was used as a negative control. In this experiment, the level of E-like became substantially reduced (by about 70%) after 48 hours and remained low 72 and 96 hours post-transfection (Fig. 6A). We initially tested whether lowered levels of E-like affected the stability of the MT cytoskeleton by monitoring recovery following nocodazole-induced MT depolymerization (Fig. 6B, panels a-l). Stable MTs were visualized using an anti-acetylated tubulin antibody that has been shown to recognize a post-translational modification associated with long-lived MTs (Sale et al., 1988). We found that immediately following removal of nocodazole, control cells contained few or no stable MTs. By contrast, E-like depleted cells retained a large number of stable MTs, mostly concentrated in the perinuclear region. This was also true at later time points, where the majority of the MT network (with the exception of the cell periphery) was fully populated after 5 minutes of recovery. The same phenotypes were observed with several specific siRNA duplexes whose sequences derived from different regions of E-like (data not shown). We confirmed the increase in MT stability by testing the resistance of MTs to brief incubation with Triton X-100, a treatment that leads to depolymerization of dynamic MTs by extracting the soluble pool of tubulin subunits. As shown in Fig. 6B (m,n), incubation with Triton X-100 led to depolymerization of virtually all the MTs in control cells (m). By contrast, a conspicuous increase in the level of dilution-resistant MTs was detected in E-like depleted cells (Fig. 6B, n). In an attempt to characterize how E-like might affect the stability of MTs in living cells, we also examined the dynamic properties of single MTs in both E-like depleted and control cells at the cell periphery. To this end, a GFP-tubulin-encoding plasmid was microinjected into HeLa cells, either alone or in combination with an E-like-specific RNA oligoduplex and the dynamic properties of single MTs from the two samples were compared (Fig. 6C). We found no significant alteration in the dynamic parameters of single MTs in E-like depleted cells compared with their untreated controls, arguing against a role for E-like in the regulation of the dynamics of bulk MTs in vivo.
siRNA targeting of E-like leads to an increase in the stability of the microtubule cytoskeleton. (A) Depletion of E-like levels upon siRNA transfection. HeLa cells were transfected with either E-like siRNA (El) or a scrambled RNA-oligo duplex (Sc) and harvested at 48, 72 and 96 hours following transfection. Cell lysates were analyzed by immunoblotting using an E-like-specific antibody and a β-actin antibody as a control for gel loading. (B) Reduced levels of E-like increase microtubule resistance to nocodazole and to Triton X-100 extraction. a-l, HeLa cells were transfected with either scrambled (Sc) or E-like (El) oligoduplexes and treated with nocodazole to completely depolymerize the microtubule cytoskeleton. At the end of the incubation, cells were restored to drug-free medium, fixed at the time points shown (0-15 minutes), and examined by immunofluorescence using anti-α-tubulin or anti-acetylated tubulin (acet-tub) antibodies. Inset shows an area of a cell at higher magnification to more clearly illustrate a population of intact microtubules. m,n, Control and E-like silenced cells were fixed after brief extraction with Triton X-100 and examined by immunofluorescence using an anti-acetylated tubulin antibody. (C) E-like siRNA does not modify the dynamics of single microtubules in EGFP-α-tubulin expressing HeLa cells (no siRNA: 16 cells, 59 MTs; with siRNA: 26 cells, 82 MTs). Bar, 10 μm. Vrs, velocity of rapid shortening; Ve, elongation velocity.
MTs serve as tracks for vesicle movements mediated by MT-dependent ATPases belonging to the kinesin and dynein families (Lippincott-Schwartz and Cole, 1995; Lippincott-Schwartz et al., 1995), and overexpression of MT stabilizing proteins has been shown to inhibit vesicle motility and organelle trafficking (Bulinski et al., 1997). We therefore tested whether the increase in MT stability induced by suppression of E-like expression had an effect on the trafficking and organization of endocellular membranes. To this end, we examined the steady-state distribution of the endoplasmic reticulum (ER), the Golgi complex and the lysosomal compartments (each visualized using specific antibodies) in control and E-like depleted cells 72 hours after siRNA transfection (Fig. 7A). We found that about 65% of cells with reduced levels of E-like expression showed dramatic clustering of these membranous compartments in a condensed region of the cytoplasm, as opposed to a more dispersed distribution evident in about 95% of the control cells. Because this phenotype was particularly evident in the case of calnexin staining, an ER-resident protein, we examined the localization of this marker in relation to the organization of the microtubule cytoskeleton. We found that ER clustering localized in close proximity to areas of stable MTs radiating from the MTOC, as indicated by anti-acetylated tubulin staining (Fig. 7B, arrow). We interpret these data as evidence of a link between an increase in the number of stable MTs and redistribution of MT-dependent molecular motors that are responsible for vesicle trafficking. These observations implicate E-like involvement in regulating the overall stability of the MT network, and provide an indirect connection between tubulin stability and MT-dependent vesicle distribution.
Reorganization of membranes in E-like depleted cells. (A) Clustering of endocellular membranes upon E-like depletion. Control (Sc) and E-like (El) knocked down cells were stained with antibodies against calnexin, giantin or lamp-2 for the detection of the ER, the Golgi complex and lysosomal membranes, respectively. (B) The ER clusters at the MTOC in E-like depleted cells (arrow). E-like silenced cells were immunolabeled with anti-calnexin and anti-acetylated tubulin for the detection of the ER and the MTOC, respectively. Bar, 10 μm.
Discussion
We identified a novel protein that we named E-like based on its homology to the tubulin-specific chaperone cofactor E (Fig. 1). In spite of their sequence similarity, E-like and cofactor E are functionally distinct: E-like cannot substitute for cofactor E in tubulin folding or refolding reactions, nor does it enhance GTP hydrolysis by β-tubulin in vitro (Fig. 2). However, when overexpressed in cultured cells, E-like causes massive depolymerization of the MT cytoskeleton by inducing proteosomal degradation of the tubulin heterodimer (Figs 3, 4). In this respect, we note that E-like and cofactor E share a C-terminal ubiquitin-like domain (UBL). UBL domains are conserved in proteins such as rad23 and parkin in which they have been shown to mediate interactions with the proteosome (Upadhya and Hegde, 2003). Furthermore, in yeast, the cofactor E homologue PAC2 has been identified as an ATP-dependent 19S cap subunit interacting protein (Verma et al., 2000). Given our findings, we speculate that the UBL domains of cofactor E and E-like may contribute to the interaction of these proteins with the 26S proteasome, thereby linking tubulin destabilization to its degradation in vivo.
We show that induction of MT depolymerization by E-like is likely to proceed by a mechanism involving the direct disruption of the tubulin heterodimer (Fig. 5). The fact that some tubulin remains intact following incubation in the presence of an excess of E-like could reflect either a preference of this protein for certain tubulin isotypes, or a failure to disrupt dimers that have been post-translationally modified, or both. In any event, these observations are not generated as a consequence of using recombinant E-like, as we obtained the same result (not shown) in a parallel experiment done with the protein purified from testis tissue. Because E-like can disrupt the tubulin heterodimer in vitro and in vivo, this protein may normally function as a MT destabilizing factor by depleting the pool of tubulin subunits available for incorporation into the polymer. However, the ability of E-like to mediate tubulin destruction distinguishes this protein from true MT destabilizing factors, which do not regulate the abundance of tubulin in the cell. For example, Op18 is thought to function by affecting the GTP state of the MT cap and by sequestering the tubulin heterodimer (Belmont and Mitchison, 1996; Howell et al., 1999), although the relative contribution of these two mechanisms in vivo is still unclear. The kinesin-related protein XKCM1 also destabilizes MTs by binding to tubulin dimers. However, unlike Op18, XKCM-1 acts by binding to dimers at the end of the MT and inducing a conformational change in the MT lattice sufficient to cause depolymerization without affecting the nucleotide-bound state of tubulin (Desai et al., 1999).
Because E-like has the capacity to destroy the tubulin heterodimer, reducing its level in siRNA experiments might be expected to result in an increase in intracellular tubulin concentration. As E-like is present at a much lower level than tubulin and its interaction with tubulin is stoichiometric (Fig. 5), reducing the concentration of E-like would not be expected to significantly change the tubulin pool. Consistent with this notion, we found that MT plus end dynamics at the cell periphery were strikingly similar in all respects for cells expressing endogenous levels of E-like or reduced levels induced via siRNA (Fig. 6C). However, we cannot formally exclude the possibility that E-like may regulate the dynamics of a particular subset of MTs. Although suppression of E-like expression did not influence the dynamics of MT plus ends at the cell periphery, the MT population was nonetheless more stable in E-like depleted cells (Fig. 6B). Both resistance to the MT depolymerizing effect of nocodazole (Gundersen and Bulinski, 1988) and to tubulin dilution resulting from mild detergent extraction (Khawaja et al., 1988) are indicative of an increase in the overall stability of the MT cytoskeleton. We suggest that E-like depletion reduces the number of dynamic MTs and increases the number of stable MTs. If this hypothesis were correct, E-like would function by regulating the distribution of MTs in stable and dynamic pools, without directly modifying MT dynamic turnover properties. In any event, our data are consistent with a role for E-like in the regulation of MT stability independent of the dynamics at the polymer plus ends.
E-like cannot directly associate with purified MTs or MT ends, and its level of cellular expression is in large stoichiometric deficit with respect to tubulin. It is possible that an indirect mechanism exists to effectively concentrate the protein in close proximity of its substrate. In the absence of an antibody that recognizes the endogenous protein, it is currently impossible to ascribe the association of E-like to any subcellular tubulin-containing structure. We note that E-like is expressed at a particularly high level in testis (Fig. 1D), a tissue in which the MT cytoskeleton undergoes a series of drastic changes during spermatogenesis, from interphase array to spindle to manchette and flagellum (Camatini et al., 1992; Kierszenbaum, 2002). E-like, with its potential to obliterate tubulin heterodimers, may associate with any of these highly dynamic MTs and participate in cytoskeletal remodeling by regulating tubulin turnover.
When expression of E-like is suppressed, we observed a tight clustering of endocellular membranes around the MTOC (Fig. 7). MTs are tracks for membrane trafficking and act by anchoring members of the kinesin and dynein families to proteins associated with vesicles (Lippincott-Schwartz and Cole, 1995; Lippincott-Schwartz et al., 1995; Thyberg and Moskalewski, 1999). The post-translational modification of tubulins within MTs can influence motor binding to the MT lattice (Rosenbaum, 2000). A simple explanation for the MTOC clustering phenotype resulting from diminished E-like expression is that loss of E-like induces an increase in the population of stable MTs with higher affinity for minus end-directed MT-dependent motors. This preferential association would displace a significant population of plus end-directed MT motors, which would no longer be available to transport their cargo to the periphery of the cell.
Acknowledgements
We thank D. Sabatini and E. Ziff for generous gifts of anti-giantin and anti-lamp-2 antibodies. We are grateful to L. D'Adamio for helpful discussion and technical advice. This work was supported by grants (to N.J.C. and L.C.) from the National Institutes of Health and an Equipment Grant from the National Science Foundation (to L.C.).
- Accepted January 11, 2005.
- © The Company of Biologists Limited 2005
References
Article navigation
Related articles
Cited by...
More in this TOC section
Similar articles
Other journals from The Company of Biologists
2020 at The Company of Biologists
Despite the challenges of 2020, we were able to bring a number of long-term projects and new ventures to fruition. While we look forward to a new year, join us as we reflect on the triumphs of the last 12 months.
Mole – The Corona Files
"This is not going to go away, 'like a miracle.' We have to do magic. And I know we can."
Mole continues to offer his wise words to researchers on how to manage during the COVID-19 pandemic.
Cell scientist to watch – Christine Faulkner
In an interview, Christine Faulkner talks about where her interest in plant science began, how she found the transition between Australia and the UK, and shares her thoughts on virtual conferences.
Read & Publish participation extends worldwide
“The clear advantages are rapid and efficient exposure and easy access to my article around the world. I believe it is great to have this publishing option in fast-growing fields in biomedical research.”
Dr Jaceques Behmoaras (Imperial College London) shares his experience of publishing Open Access as part of our growing Read & Publish initiative. We now have over 60 institutions in 12 countries taking part – find out more and view our full list of participating institutions.
JCS and COVID-19
For more information on measures Journal of Cell Science is taking to support the community during the COVID-19 pandemic, please see here.
If you have any questions or concerns, please do not hestiate to contact the Editorial Office.