Stu2, the Budding Yeast XMAP215/Dis1 Homolog, Promotes Assembly of Yeast Microtubules by Increasing Growth Rate and Decreasing Catastrophe Frequency*

Background: The reported inhibition of microtubule growth by Stu2 is difficult to reconcile with its cellular phenotypes. Results: Using microscopy assays, we found that Stu2 increases the growth rate and decreases the catastrophe frequency of yeast microtubules. Conclusion: Stu2 promotes microtubule growth, with considerably higher activity on budding yeast microtubules. Significance: The biochemical properties of Stu2 reported here account for the mitotic phenotypes observed in cells. Stu2 is the budding yeast member of the XMAP215/Dis1 family of microtubule polymerases. It is essential in cell division, allowing proper spindle orientation and metaphase chromosome alignment, as well as spindle elongation during anaphase. Despite Stu2 having a phenotype that suggests it promotes microtubule growth, like the other members of the XMAP215/Dis1 family, previous studies with purified Stu2 indicate only that it antagonizes microtubule growth. One potential explanation for these contradictory findings is that the assays were performed with mammalian brain tubulin, which may not be the right substrate to test the activity of Stu2 given that yeast and brain tubulins are quite divergent in sequence and that the vertebrate tubulins are subject to many post-translational modifications. To test this possibility, we reconstituted the activity of Stu2 with purified budding yeast tubulin. We found that Stu2 accelerated microtubule growth in yeast tubulin by severalfold, similar to the acceleration reported for XMAP215 in porcine brain tubulin. Furthermore, Stu2 accelerated polymerization in yeast tubulin to a much greater extent than in porcine brain tubulin, and the concentration of Stu2 required to reach 50% maximum activity in yeast tubulin was nearly 2 orders of magnitude lower than that in porcine brain tubulin. We conclude that Stu2 is a microtubule polymerase, like its relatives, and that its activity is considerably higher in yeast tubulin compared with mammalian brain tubulin. The biochemical properties of Stu2 reported here account for many of the phenotypes of Stu2 observed in cells.

Microtubules, cytoskeletal filaments composed of ␣,␤-tubulin heterodimers, serve as tracks for molecular motors and provide structure to organelles such as mitotic spindles and axon-emes (1,2). The growth and shrinkage of microtubules are regulated by microtubule-associated proteins (MAPs) 2 that can increase or decrease the rate of polymerization or depolymerization or that can regulate the switching between growth and shrinkage (3)(4)(5). This study addresses the protein Stu2, a member of the XMAP215/Dis1 family of microtubule polymerases, which, unexpectedly and unlike other members of the family, has been reported to inhibit, rather than accelerate, polymerization.
Members of the XMAP215/Dis1 family promote microtubule growth both in cells and in biochemical assays with purified proteins (6). The founding member of this family, XMAP215, is required for rapid microtubule growth in Xenopus laevis egg extracts and accelerates elongation of the fastgrowing plus ends of microtubules by up to 8-fold in the presence of purified bovine brain tubulin (7). Tea1, the Aspergillus nidulans homolog of XMAP215, accelerates the growth of porcine brain microtubules to a similar extent as XMAP215 (8). Stu2, the budding yeast homolog of this family, is required for elongation of the mitotic spindle during anaphase (9), as well as for spindle orientation and metaphase chromosome alignment (10). Mutants lacking the TOG1 domain of Stu2 exhibit shorter cytoplasmic microtubules and shorter spindles (11), consistent with Stu2 promoting microtubule growth. However, studies with purified Stu2 have shown that it slows, rather than accelerates, elongation of microtubules grown in brain tubulin (12). Furthermore, these authors reported that Stu2 also promotes catastrophe, the transition of growing microtubules to shrinking ones, which often correlates with slower growth rates. One interpretation of these results is that not all members of the XMAP215/Dis1 family are polymerases and that the shared protein domains, notably the TOG domains (13,14), may have divergent activities.
An alternative explanation for the difference in Stu2 and XMAP215 activities is that mammalian brain tubulin may not be the right substrate to test the activity of Stu2 given that yeast * This work was supported by National Institutes of Health Grant R01 and brain tubulins are quite divergent in sequence and that the vertebrate tubulins are subject to many post-translational modifications (see "Discussion"). To test this possibility, we reconstituted the polymerization activity of Stu2 with budding yeast tubulin.
The crude lysate was clarified by centrifugation at 13,000 ϫ g for 20 min and loaded onto an SP-Sepharose column (Amersham Biosciences HiTrap SP HP). The column was washed with cation buffer (6.7 mM HEPES, 6.7 mM MES, 6.7 mM NaOAc (pH 7.2), 150 mM NaCl, 1.5 mM MgCl 2 , 1 mM DTT, 10 M MgATP, and protease inhibitors), and the protein was washed with cation buffer and then with cation buffer containing 250 mM NaCl and eluted from the column with cation buffer containing 350 mM NaCl. This fraction was loaded onto a Ni 2ϩ -Sepharose column (Amersham Biosciences HisTrap HP). The column was washed with imidazole buffer (50 mM phosphate buffer (pH 7.5), 300 mM NaCl, 10 mM imidazole, 10% glycerol, 1 mM MgCl 2 , 20 M ATP, and protease inhibitors) and then with imidazole buffer containing 70 mM imidazole, and the protein was eluted with imidazole buffer containing 300 mM imidazole. Peak fractions were pooled, aliquoted, and snap-frozen in liquid N 2 . Protein was stored at Ϫ80°C. XMAP215 was purified as described previously (24). Porcine brain tubulin was purified as described previously (15). Cycled tubulin was labeled with Alexa Fluor 488 or rhodamine (TAMRA (5(6)-carboxytetramethylrhodamine succinimidyl ester), Invitrogen) as described by Hyman et al. (16). GMP-CPP 2 -stabilized microtubules were grown as described (17).
Budding yeast tubulin was purified from strain BJ2168. The culture was collected at an exponential phase of growth (A 600 ϭ 1), and the pellet was resuspended in buffer (0.7 ml of buffer/ liter of cells) containing 30 mM HEPES (pH 7.2), 50 mM KOAc, 2 mM Mg(OAc) 2 , 1 mM EGTA, 10% glycerol, 0.2% Triton X-100, 0.2 mM PMSF, 10 g/ml pepstatin A, 10 g/ml leupeptin, and 10 g/ml aprotinin. The cells were then dropped into liquid nitrogen, giving cryo-beads. The beads were processed by grinding in a Retsch ZM 200 grinder, yielding a final powder lysate that was stored at Ϫ80°C. This lysate was used for purification of tubulin with the TOG column as described previously (18).
Polymerization Assays and Imaging-Dynamic microtubule assays using rhodamine-labeled seeds and dynamic (Alexa Fluor 488-labeled or unlabeled) extensions were performed as described previously (19). The imaging buffer consisted of BRB80 (80 mM PIPES, 1 mM MgCl 2 , 1 mM EGTA, pH 6.8) supplemented with 40 mM glucose, 40 g/ml glucose oxidase, 16 g/ml catalase, 0.1 mg/ml casein, 1% DTT, 1 mM GTP, and 50 mM KCl (unless stated otherwise). In total internal reflection fluorescence microscopy experiments, images were acquired every 5 or 10 s with an Andor iXon camera on a Zeiss inverted microscope with a Zeiss alpha FLUAR 100ϫ/1.45 oil objective. An objective heater was used to heat the sample to 35°C (brain tubulin) or 28°C (yeast tubulin). Static assays were performed in the same way as the dynamic assays, except that Alexa Fluor 488-labeled tubulin was not used in this case, and GTP was not added to the imaging buffer. Images were collected either every 10 s or every 100 ms (in streaming mode). In differential interference contrast (DIC) experiments, we used the setup described previously (20). Data were analyzed by making kymographs from acquired images using Fiji software, which were then used to measure the parameters of microtubule dynamics. Growth rate was determined by measuring the change in length of a microtubule over time. Catastrophe frequency was expressed as the number of catastrophes over the total time of microtubule growth.

Stu2 Is a Weak Polymerase on Porcine Brain Tubulin-To
investigate the activity of Stu2, we first performed experiments using porcine brain tubulin. Stable microtubules were grown from porcine brain tubulin, 25% of which was labeled with rhodamine, in the presence of GMP-CPP, a slowly hydrolyzable GTP analog (21). These stable microtubules were bound to the surface of a coverslip using anti-rhodamine antibodies and imaged by total internal reflection fluorescence microscopy ( Fig. 1A). To investigate the effect of Stu2 on microtubule growth, 7 M 10% Alexa Fluor 488-labeled porcine brain tubulin and various concentrations of unlabeled Stu2 were added. Microtubule growth and the transition between the growing and shrinking phases were quantified from the kymographs (Fig. 1B). Increasing the concentration of Stu2 increased the microtubule growth rate at the plus end by 2.4-fold from 0.448 Ϯ 0.010 m/min (mean Ϯ S.E., n ϭ 91) in control experiments with tubulin to only 1.071 Ϯ 0.034 m/min (mean Ϯ S.E., n ϭ 121) with 500 nM Stu2 (Fig. 1C, left panel). Approximately 200 nM Stu2 was required to reach 50% of this increase in growth rate. A similar acceleration of growth was observed at the minus end: Stu2 increased the growth rate from 0.131 Ϯ 0.007 m/min (mean Ϯ S.E., n ϭ 22) with tubulin alone to 0.342 Ϯ 0.011 m/min with 500 nM Stu2 (mean Ϯ S.E., n ϭ 79). Thus, Stu2 is a microtubule polymerase, like its Xenopus relative XMAP215 and its Aspergillus relative Tea1, although the acceleration of growth by Stu2 is only ϳ2-fold less than the 8-fold increase observed for the other proteins.
Stu2 had a small effect on catastrophe (Fig. 1C, right panel). In the absence of Stu2, the catastrophe rate at the plus end was 0.119 Ϯ 0.013 min Ϫ1 (mean Ϯ S.E., n ϭ 79). This rate increased slightly but significantly at intermediate Stu2 concentrations

Stu2 Is a Microtubule Polymerase
(0.162 Ϯ 0.022 min Ϫ1 at 100 nM and 0.168 Ϯ 0.022 min Ϫ1 at 200 nM) but then decreased to control levels at 500 nM. At the minus end, Stu2 slightly inhibited catastrophe. The effect of Stu2 was similar to that of XMAP215 and Tea 1, which increased catastrophe at the plus end by 2.6-and 2.7-fold, respectively, at 200 nM (8,22). Based on these effects on the growth rate and catastrophe frequency (and assuming the model of Ref. 23), Stu2 is expected to increase the mean length at which microtubules catastrophe at the plus end from 3.8 m (0 Stu2) to 8.9 m (500 nM Stu2), similar to the measured values of 3.5 Ϯ 0.4 m (0 Stu2, mean Ϯ S.E., n ϭ 79) to 8.2 Ϯ 0.9 m (500 nM Stu2, mean Ϯ S.E., n ϭ 78).
Stu2 Binds to the Microtubule Ends and Acts as a Depolymerase in the Absence of Tubulin-We used a GFP-tagged construct to directly observe Stu2 interactions with microtubules. Stu2-GFP has a weak polymerase activity with porcine brain tubulin. The addition of 100 nM Stu2-GFP caused a modest increase in microtubule growth rate in 10 M tubulin from 0.386 Ϯ 0.050 m/min (mean Ϯ S.D., n ϭ 9) to 0.485 Ϯ 0.081 m/min (mean Ϯ S.D., n ϭ 20). This rate increase was similar to that induced by the same concentration of unlabeled Stu2 (Fig. 1C). During dynamic assays in the presence of unlabeled tubulin, Stu2-GFP bound to the plus end of growing microtubules with higher affinity than the seed or the extension ( Fig.  2A). At low concentrations of Stu2 (Ͻ Ͻ1 nM), single molecules of Stu2-GFP were observed to bind to and diffuse on the microtubule surface; other Stu2 molecules were observed to bind stably near both microtubule ends (Fig. 2B). At lower salt concentrations, Stu2-GFP bound much more strongly to the microtubule lattice (Fig. 2C). In BRB80 without extra KCl, GMP-CPP-stabilized microtubules were decorated with Stu2-GFP along their whole length; furthermore, under these conditions, Stu2-GFP accelerated the depolymerization of these stable microtubules by 10-fold from the spontaneous rate of 0.0046 Ϯ 0.0023 m/min (mean Ϯ S.D., n ϭ 18) to 0.058 Ϯ 0.014 m/min (mean Ϯ S.D., n ϭ 35; p Ͻ 0.0001 by a t test).
At 7 M porcine brain tubulin, no rescues were observed; in the presence of 500 nM Stu2, the rescue rate increased to 0.54 Ϯ 0.09 min Ϫ1 (mean Ϯ S.E., n ϭ 36 rescues). However, rescues were still relatively infrequent. At 500 nM Stu2, only 46% of the   OCTOBER 10, 2014 • VOLUME 289 • NUMBER 41 shrinking plus ends rescued; this is expected to increase the microtubule length at catastrophe by a factor of Ͻ2.

Stu2 Is a Microtubule Polymerase
Stu2 Is a Strong Polymerase on Budding Yeast Tubulin-After characterizing the interaction of Stu2 with porcine brain tubulin, we studied its interaction with the conspecific budding yeast tubulin purified from strain BJ2168 (see "Experimental Procedures"). Because the yield was low (18), we had insufficient protein for fluorescent labeling. We therefore used DIC microscopy to visualize the microtubules extending from GMP-CPP-stabilized porcine tubulin seeds. We made two other changes in the assays conditions. First, we performed the experiments at 28°C, which is more appropriate for physiological conditions of budding yeast proteins. Second, we used a lower concentration of yeast tubulin than brain tubulin because yeast tubulin polymerizes at lower concentrations than mammalian tubulin, as reported previously (25,26).
Stu2 increased the elongation rate of microtubules grown in 5 M yeast tubulin by 5-fold from 0.26 Ϯ 0.04 m/min (mean Ϯ S.D., n ϭ 36) to 1.26 Ϯ 0.14 m/min (mean Ϯ S.D., n ϭ 37) in the presence of 125 nM Stu2 (Fig. 3, A and C). The Michaelis-Menten fit to these data gave a K m of 22.7 Ϯ 5.5 nM (mean Ϯ S.E.) and a V max of 1.51 Ϯ 0.10 m/min (mean Ϯ S.E.). Thus, Stu2 accelerated growth in yeast tubulin to a greater degree than in porcine brain tubulin. Furthermore, this acceleration of growth occurred at much lower Stu2 concentrations, 23 nM for half-maximal acceleration for yeast tubulin compared with ϳ200 nM for mammalian tubulin. For example, 5 nM Stu2 was sufficient to double the growth rate of yeast microtubules, whereas nearly 500 nM Stu2 was needed to double the growth rate of porcine microtubules (Fig. 1). There was also a difference between yeast and mammalian tubulins at tubulin concentrations that gave the same basal growth rates. At 9 M porcine tubulin, 125 nM Stu2 increased the growth rate by only 20% from 0.27 Ϯ 0.08 m/min (mean Ϯ S.D., n ϭ 29) to 0.32 Ϯ 0.09 m/min (mean Ϯ S.D., n ϭ 13) (Fig. 3, B and C). By comparison, 125 nM Stu2 increased the growth rate of yeast tubulin by 470% to 1.23 Ϯ 0.15 m/min (mean Ϯ S.D., n ϭ 37).
In addition to increasing growth rate, Stu2 also decreased the catastrophe frequency of yeast microtubules (Fig. 3D). In the absence of Stu2, the catastrophe frequency was 0.272 Ϯ 0.018 min Ϫ1 (mean Ϯ S.E., n ϭ 240 catastrophes), and this rate decreased to 0.116 Ϯ 0.033 min Ϫ1 (mean Ϯ S.E., n ϭ 12 catastrophes) in the presence of 125 nM Stu2. Based on these effects on the growth rate and catastrophe frequency (and again assuming the model of Ref. 23), Stu2 is expected to increase the mean length at which yeast microtubules undergo catastrophe at the plus end from 0.96 m (0 Stu2) to 10.8 m (125 nM Stu2), similar to the measured values of 1.3 Ϯ 0.2 m (0 Stu2, mean Ϯ S.E., n ϭ 28) and 11.5 Ϯ 3.3 m (125 nM Stu2, mean Ϯ S.E., n ϭ 12). Thus, Stu2 promotes the growth of yeast microtubules, increasing mean length by 10-fold through its combination of polymerase and anti-catastrophe activities. No rescues were observed.
XMAP215 Is a Strong Polymerase of Yeast Tubulin-Having found that Stu2 is a strong polymerase of its conspecific tubulin, we asked whether XMAP215 is also a stronger polymerase of vertebrate tubulin over yeast tubulin. To the contrary, we found that XMAP215 is also a strong polymerase of yeast tubulin. When we introduced increasing amounts of XMAP215 and 5 M budding yeast tubulin to GMP-CPP-stabilized seeds and visualized the formation of dynamic microtubules using DIC microscopy at 28°C (Fig. 4, A and B), we observed a large increase in yeast microtubule growth rates. Namely, the growth rate increased by 13-fold from 0.26 Ϯ 0.07 m/min (mean Ϯ S.D., n ϭ 198) observed with yeast tubulin alone to 3.45 Ϯ 0.87 m/min (mean Ϯ S.D., n ϭ 45) in the presence of 125 nM XMAP215. Consistent with previous reports (22,24), XMAP215 caused an 9-fold increase in porcine microtubule growth rates, increasing the growth rate from 0.32 Ϯ 0.08 m/min (mean Ϯ S.D., n ϭ 106) in the absence of XMAP215 to 2.87 Ϯ 1.25 m/min (mean Ϯ S.D., n ϭ 18) in the presence of 125 nM XMAP215. These results show that XMAP215 increases the polymerization rates of both yeast and porcine tubulins to similar extents. However, the Michaelis-Menten fits to these curves reveal different K m values, indicating that XMAP215 is more active on porcine tubulin than on yeast tubulin (K m for porcine brain tubulin ϭ 2.13 Ϯ 0.56 nM, which is considerably lower than that for yeast tubulin, K m ϭ 17.35 Ϯ 3.22 nM).

DISCUSSION
We found that Stu2 stimulates the growth of microtubules polymerizing in the presence of budding yeast tubulin. The maximum stimulation of 5-fold is similar to that of the evolutionarily related proteins XMAP215 and Tea1. Thus, Stu2 is a microtubule polymerase, like the other members of the XMAP215/Dis1 family. The polymerase activity of purified Stu2 occurs at concentrations similar to those found in cells (80 nM) (27) and accounts for some of the observed phenotypes during mitosis. For example, anaphase microtubule elongation is impaired in both Stu2-depleted and Stu2-⌬TOG1 mutant cells (9 -11), as expected if polymerase activity is decreased. The effects of Stu2 depletion on the dynamics of cytoplasmic microtubules (10) are not easily reconciled with our in vitro results, suggesting that interactions of Stu2 with other MAPs may be involved.
We could not confirm earlier results showing that Stu2 inhibits, rather than promotes, elongation of mammalian brain microtubules (12). Although Stu2 is a weak polymerase in the presence of porcine brain tubulin and requires high Stu2 concentrations, it nevertheless promoted microtubule growth under all conditions that we tested.
A key finding was that Stu2 has higher polymerase activity on budding yeast tubulin than on porcine brain tubulin. Although the ␣and ␤-tubulins are highly evolutionarily conserved families of proteins, there are considerable differences between budding yeast and mammalian tubulins at both the gene and protein levels that might lead to this difference in Stu2 activity. Budding yeast has two ␣-tubulin genes and one ␤-tubulin gene (28,29), whereas vertebrates have six ␣-tubulin genes and seven ␤-tubulin genes (30). Within ␣and ␤-tubulins, the sequence identity between yeast and mammalian proteins is only 70 -75%, with large differences at their C termini. Post-translational modifications of vertebrate tubulins lead to additional tubulin complexity at the protein level (31). These post-translational modifications occur mostly on the C termini of ␣and ␤-tubulins, sites that are recognized by many regulators of microtubule dynamics (31,32), including XMAP215 (24). In the case of bovine brain tubulin, these different gene products and post-translational modifications give rise to up to 21 different tubulin species that are distinguishable on an isoelectric focusing gel (33,34) Thus, there are many potential molecular mechanisms that could give rise to the differences in activity of Stu2 on the different sources of tubulin.
XMAP215 also differentiated between tubulins: although the maximum enhancement of growth was in the same range for yeast and porcine tubulins, less XMAP215 was needed to give a half-maximal promotion on porcine tubulin compared with yeast tubulin. This provides additional evidence for differences between yeast and brain tubulins. The lower concentration of XMAP215 required to accelerate growth may reflect a higher affinity of XMAP215 for the end of porcine microtubules; once at the end, however, the growth acceleration is similar. Interestingly, under the conditions that we tested, XMAP215 was a stronger polymerase than Stu2 on both tubulins, enhancing the growth rates of yeast and porcine tubulins by 13-and 9-fold, respectively, compared with enhancements of 5-and 2-fold by Stu2.
Stu2 decreased the catastrophe frequency of yeast microtubules but had little effect on the catastrophe frequency of porcine microtubules. Interestingly, XMAP215 also decreased catastrophe in yeast tubulin but increased catastrophe in porcine tubulin. This is another difference in regulation by MAPs between yeast and porcine tubulins. It is possible that these effects reflect the inherent differences in polymerization dynamics of these two tubulin types.
Stu2 shares many properties with XMAP215. Both accelerate microtubule growth and inhibit catastrophe, both preferentially bind microtubule ends over lattice, both diffuse on the lattice, and both accelerate depolymerization of stable microtubules when there is no free tubulin in solution. There are, however, some differences. As reported above, Stu2 can bind to the minus end, where it accelerates microtubule growth, which is not observed for XMAP215 (24). Another possible difference between Stu2 and XMAP215 is that the former increased the rate of transition from shrinking to growing (rescue). By contrast, rescues have not been reported for XMAP215 (and we also saw no rescues under our experimental conditions).
In summary, our results demonstrate that differences between yeast and mammalian tubulins can give rise to differential activities of interacting proteins, such as previously reported for kinesin (35) and dynein (36). In this study, we reported significant differences in protein activity when using tubulins from different sources, showing the importance of using conspecific tubulin when studying MAPs.