Regulation of Microtubule Dynamic Instability in Vitro by Differentially Phosphorylated Stathmin*

Stathmin is an important regulator of microtubule polymerization and dynamics. When unphosphorylated it destabilizes microtubules in two ways, by reducing the microtubule polymer mass through sequestration of soluble tubulin into an assembly-incompetent T2S complex (two α:β tubulin dimers per molecule of stathmin), and by increasing the switching frequency (catastrophe frequency) from growth to shortening at plus and minus ends by binding directly to the microtubules. Phosphorylation of stathmin on one or more of its four serine residues (Ser16, Ser25, Ser38, and Ser63) reduces its microtubule-destabilizing activity. However, the effects of phosphorylation of the individual serine residues of stathmin on microtubule dynamic instability have not been investigated systematically. Here we analyzed the effects of stathmin singly phosphorylated at Ser16 or Ser63, and doubly phosphorylated at Ser25 and Ser38, on its ability to modulate microtubule dynamic instability at steady-state in vitro. Phosphorylation at either Ser16 or Ser63 strongly reduced or abolished the ability of stathmin to bind to and sequester soluble tubulin and its ability to act as a catastrophe factor by directly binding to the microtubules. In contrast, double phosphorylation of Ser25 and Ser38 did not affect the binding of stathmin to tubulin or microtubules or its catastrophe-promoting activity. Our results indicate that the effects of stathmin on dynamic instability are strongly but differently attenuated by phosphorylation at Ser16 and Ser63 and support the hypothesis that selective targeting by Ser16-specific or Ser63-specific kinases provides complimentary mechanisms for regulating microtubule function.

Stathmin is an important regulator of microtubule polymerization and dynamics. When unphosphorylated it destabilizes microtubules in two ways, by reducing the microtubule polymer mass through sequestration of soluble tubulin into an assemblyincompetent T2S complex (two ␣:␤ tubulin dimers per molecule of stathmin), and by increasing the switching frequency (catastrophe frequency) from growth to shortening at plus and minus ends by binding directly to the microtubules. Phosphorylation of stathmin on one or more of its four serine residues (Ser 16 , Ser 25 , Ser 38 , and Ser 63 ) reduces its microtubule-destabilizing activity. However, the effects of phosphorylation of the individual serine residues of stathmin on microtubule dynamic instability have not been investigated systematically. Here we analyzed the effects of stathmin singly phosphorylated at Ser 16 or Ser 63 , and doubly phosphorylated at Ser 25 and Ser 38 , on its ability to modulate microtubule dynamic instability at steadystate in vitro. Phosphorylation at either Ser 16 or Ser 63 strongly reduced or abolished the ability of stathmin to bind to and sequester soluble tubulin and its ability to act as a catastrophe factor by directly binding to the microtubules. In contrast, double phosphorylation of Ser 25 and Ser 38 did not affect the binding of stathmin to tubulin or microtubules or its catastrophe-promoting activity. Our results indicate that the effects of stathmin on dynamic instability are strongly but differently attenuated by phosphorylation at Ser 16 and Ser 63 and support the hypothesis that selective targeting by Ser 16 -specific or Ser 63 -specific kinases provides complimentary mechanisms for regulating microtubule function.
Stathmin is known to destabilize microtubules in two ways. One is by binding to soluble tubulin and forming a stable complex that cannot polymerize into microtubules, consisting of one molecule of stathmin and two molecules of tubulin (T2S complex) (17)(18)(19)(20)(21)(22)(23)(24). Addition of stathmin to microtubules in equilibrium with soluble tubulin results in sequestration of the tubulin and a reduction in the level of microtubule polymer (17-18, 22, 25-28). In addition to reducing the amount of assembled polymer, tubulin sequestration by stathmin has been shown to increase the switching frequency at microtubule plus ends from growth to shortening (called the catastrophe frequency) as the microtubules relax to a new steady state (17,29). The second way is by binding directly to microtubules (27)(28)(29)(30). The direct binding of stathmin to microtubules increases the catastrophe frequency at both ends of the microtubules and considerably more strongly at minus ends than at plus ends (27). Consistent with its strong catastrophe-promoting activity at minus ends, stathmin increases the treadmilling rate of steady-state microtubules in vitro (27). These results have led to the suggestion that stathmin might be an important cellular regulator of minus-end microtubule dynamics (27).
Phosphorylation of stathmin diminishes its ability to regulate microtubule polymerization (3,14,(25)(26). Phosphorylation of Ser 16 or Ser 63 appears to be more critical than phosphorylation of Ser 25 and Ser 38 for the ability of stathmin to bind to soluble tubulin and to inhibit microtubule assembly in vitro (3,25). Inhibition of stathmin phosphorylation induces defects in spindle assembly and organization (3,14) suggesting that not only soluble tubulin-microtubule levels are regulated by phosphorylation of stathmin, but the dynamics of microtubules could also be regulated in a phosphorylation-dependent manner.
It is not known how phosphorylation at any of the four serine residues of stathmin affects its ability to regulate microtubule dynamics and, specifically, its ability to increase the catastrophe frequency at plus and minus ends due to its direct interaction with microtubules. Thus, we determined the effects of stathmin individually phosphorylated at either Ser 16 or Ser 63 and doubly phosphorylated at both Ser 25 and Ser 38 on dynamic instability at plus and minus ends in vitro at microtubule polymer steady state and physiological pH (pH 7.2). We find that phosphorylation of Ser 16 strongly reduces the direct catastrophe-promoting activity of stathmin at plus ends and abolishes it at minus ends, whereas phosphorylation of Ser 63 abolishes the activity at both ends. The effects of phosphorylation of individual serines correlated well with stathmin's reduced abilities to form stable T2S complexes, to inhibit microtubule polymerization, and to bind to microtubules. In contrast, double phosphorylation of Ser 25 and Ser 38 did not alter the ability of stathmin to modulate dynamic instability at the microtubule ends, its ability to form a stable T2S complex, or its ability to bind to microtubules. The data further support the hypotheses that phosphorylation of stathmin on either Ser 16 or Ser 63 plays a critical role in regulating microtubule polymerization and dynamics in cells.

EXPERIMENTAL PROCEDURES
Protein Preparation-For the production of specifically phosphorylated forms of stathmin, the serine residues to remain unphosphorylated were first converted to alanines. Specifically, Ser 25 (25). Recombinant stathmin proteins were bacterially expressed and purified as previously described (23,25). Phosphorylation of the stathmin constructs was then carried out as described previously (25). Briefly, protein kinase A (cAMP-dependent protein kinase) was used to phosphorylate Ser 16 and Ser 63 and a mixture of mitogen-activated protein kinase (MAPK) and Cdc2 was used to phosphorylate Ser 25 and Ser 38 . Proteins were incubated with the kinases (2.8, 2.0, and 0.5 units of cAMP-dependent protein kinase, MAPK, and/or Cdc2, respectively, per microgram of stathmin) in 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 5 mM EGTA, 2 mM dithiothreitol, 500 mM ATP for 6 -8 h at 30°C. Product formation was assessed by native-PAGE. Kinases were heat-inactivated at 75°C for 10 min; because stathmin displays a fully reversible thermal unfolding/refolding (21), this heatinactivation step does not impair the activity of stathmin. Phosphoisoforms (Ͼ93% phosphorylation) were purified to homogeneity by anion exchange chromatography. Identities of the stathmin proteins were assessed by mass spectral analyses. We have previously demonstrated that both the tubulin-binding properties and the tubulin-sequestering activities of p16 stathmin and p63 stathmin are very similar to p16,p25,p38 and p25,p38,p63 stathmins, respectively (25). We thus decided to work with the single phosphorylated versions of p16 and p63 stathmin. It should also be noted that high quality production of large amounts of triply phosphorylated stathmin is much more difficult compared with single phosphorylation. Purified bovine brain tubulin was prepared as described previously (27). Protein concentrations throughout this work were determined by the method of Bradford with bovine serum albumin as the standard (31).
Analysis of Microtubule Dynamics-Purified tubulin (20 M) was assembled onto the ends of sea urchin (Strongylocentrotus purpuratus) axoneme seeds at 30°C in PMME buffer (87 mM Pipes, 36 mM Mes, 1.4 mM MgCl 2 , 1 mM EGTA) at pH 7.2 in the presence or absence of unphosphorylated or phosphorylated stathmin isoforms and incubated at 30°C for 40 min to reach steady state (confirmed by light scattering at 350 nm). Analysis of dynamic instability at plus and minus ends was carried out by video-enhanced differential interference contrast microscopy as previously described (27,32). Briefly, reaction mixtures containing tubulin and stathmin proteins at the desired concentrations were injected into a chamber made of a glass coverslip on a glass slide that was pre-saturated with axoneme seeds in PMME buffer. Microtubules were polymerized to steady state at both ends of the axoneme seeds, and real-time images of the microtubules were recorded. The plus ends of microtubules were identified and distinguished from minus ends on the basis of their fast growth rates, the number of microtubules that grew at the ends, and the relative lengths of the microtubules. The growth rates, shortening rates, and transition frequencies were determined as previously described (27). We considered microtubules to be growing if they increased in length Ͼ0.3 m at a rate Ͼ0.3 m/min. Shortening events were identified by a Ͼ1 m length change at a rate of Ͼ2 m/min. Microtubules that changed Ͻ0.3 m/min over a duration of 4 data points were considered to be in an attenuated (paused) state. Between 30 and 50 microtubules were analyzed for each condition. Because microtubule minus ends are relatively stable and rarely undergo shortening or growth events and only a few microtubules out of many that are tracked undergo catastrophe or rescue events, the number of events recordable at the minus ends was lower than the at the plus ends.
Gel Filtration-Mixtures of tubulin (25 M) and unphosphorylated or phosphorylated stathmin (5 M) were incubated for 20 min at room temperature and then loaded into a Superose 12 GF 24 ϫ 1.0 cm fast-protein liquid chromatography column. The elution profiles of unbound and stathmin-bound tubulin (T2S complex) were obtained by monitoring absorbance at 280 nm as described (18). The column was equilibrated with PEM buffer (100 mM Pipes, 1 mM EGTA, and 0.5 mM MgCl 2 ) at pH 7.2 containing 2 mM GTP and developed at a flow rate of 0.3 ml/min. Total volume (V t ) of the gel bed and the void volume (V o ) were determined as previously described (32). The column was calibrated using a gel filtration molecular weight calibration kit (Amersham Biosciences).
Determination of Microtubule Polymer Mass-Microtubule suspensions containing 20 M tubulin and unphosphorylated or phosphorylated stathmin isoforms (4 M) were polymerized at 37°C for 1 h in PMME buffer (pH 7.2) with 1.5 mM GTP in the presence of 1% glycerol-stabilized nucleating microtubule seeds (32,33). Seeds were prepared by assembling purified tubulin (2.0 mg/ml) in PEM buffer, pH 6.8, plus 10% glycerol and 1 mM GTP, then shearing the microtubules six times through a 25-guage needle. Seed suspensions were used at a 1:10 dilution. The amount of protein in microtubule pellets was determined after sedimenting and re-suspending the microtubules in ice-cold buffer (32). The final amounts of protein polymerized into microtubules were determined by subtracting the amount of protein present in the control seeds (10% v/v) from the total amount of protein in the sedimented pellets.

Binding of Stathmin Isoforms to Pre-assembled Microtubules-
We analyzed the binding of stathmin to microtubules by first polymerizing purified tubulin to steady state (40 M) with 4% glycerol-induced microtubule seeds for 1 h at 37°C. Stathmin isoforms (4A stathmin, p63 stathmin, p16 stathmin, or p25,38 stathmin) were added to the preassembled microtubules, and incubation continued for an additional 30 min. Microtubules were sedimented onto coverslips through 30% sucrose cushions containing 1 M taxol and treated with a mixture of primary antibodies against ␣-tubulin (DM1A mouse monoclonal (1:500), Sigma) and stathmin (rabbit antiserum (1:200), Calbiochem) followed by addition of secondary antibodies (donkey anti-mouse Cy3 and goat-anti-rabbit fluorescein) (27). We confirmed by Western blot analysis that the primary antibody against stathmin, which was specific to the C-terminal region of stathmin, has similar affinities for unphosphorylated wild-type stathmin, 4A stathmin, and the phosphorylated stathmin isoforms (not shown). Images were obtained with a Nikon Eclipse E800 Immunofluorescence microscope with Metamorph 4.6 software.

Effect of Phosphorylation of Ser 16 and Ser 63 of Stathmin on Microtubule Polymerization and on Sequestration of Soluble
Tubulin into Inactive T2S Complexes-Specifically phosphorylated stathmin isoforms were constructed by mutating all serine residues into alanines except for the specific phosphoserine to be studied ("Experimental Procedures"). As a control for the possibility that conversion of serines to alanines at any of the four positions might modulate the activity of stathmin, we also prepared and analyzed the activity of stathmin with all four serines converted to alanines (4A stathmin). Mixtures of the unphosphorylated and 4A stathmin at a molar ratio of stathmin isoform to tubulin (20 M) of 1:5 were assembled to steady state at 37°C by addition of nucleating microtubule seeds, and the quantity of polymer was determined in the microtubule pellets after sedimentation ("Experimental Procedures"). As shown in Fig. 1A, unphosphorylated stathmin (un-phos) reduced the percent tubulin in microtubules by ϳ53%, and 4A stathmin reduced polymerization by a similar extent of ϳ57%. Stathmin reduces the amount of assembled microtubules by sequestering soluble tubulin dimers into a T2S complex consisting of one molecule of stathmin stably bound to 2 molecules of tubulin dimer (17)(18)(19)(20)(21)(22)(23)(24)(25)(26).
Next we wanted to ensure that 4A stathmin sequestered tubulin as well as unphosphorylated stathmin. Mixture of tubulin (25 M) and wild-type stathmin were incubated for 20 min at room temperature, and the amount of T2S complex was determined after separating the complex from unbound tubulin by size exclusion chromatography ("Experimental Procedures"). Consistent with previous studies (17)(18)(19)(20)(21)(22)(23)32), wild-type stathmin formed a stable T2S complex as visualized by the ratio of the quantity of protein in the high molecular weight peak at 10.4 ml to that in the lower molecular weight peak (11.6 ml) which corresponds to free tubulin dimers (32) (Fig. 1B). Similarly, 4A stathmin formed a similar ratio of T2S complex with free tubulin (Fig. 1B). The slight difference in the relative absorbance between unphosphorylated and 4A stathmin could be due to differences in hydrophobicity. Conversion of all four serines to alanines did not affect the ability of stathmin to inhibit polymerization or to sequester tubulin, consistent with previous studies (25). Thus, any effects of the specific phosphoisoforms on microtubule polymerization could not be due to mutation of the serine residues to alanines.
We next determined how stathmin specifically phosphorylated at Ser 16 (p16 stathmin) or Ser 63 (p63 stathmin) modulates the amount of microtubule polymer and the ability of stathmin to form T2S complexes. As shown in Fig. 1A, p16 stathmin reduced polymerization by only ϳ12% (Fig. 1A). In addition, p16 stathmin sequestered soluble tubulin poorly (Fig. 1B). These data are consistent with those of Honnappa et al. (25) who demonstrated, using high sensitivity isothermal titration calorimetry, that p16 binds to two tubulin subunits although at a lower affinity as compared with unphosphorylated stathmin. Similarly, p63 stathmin did not reduce microtubule polymerization and did not form detectable T2S complexes, indicating that, at physiological pH, phosphorylation of either Ser 16 or Ser 63 strongly reduces the ability of stathmin to inhibit microtubule polymerization and to sequester soluble tubulin into assembly-incompetent T2S complexes.

Effects of P16 and P63 Stathmin on Dynamic Instability at
Polymer Mass Steady State-We wanted to analyze the effects of phosphorylation on the ability of stathmin to modulate microtubule dynamic instability at steady state. Once steady state is achieved, the mass of assembled microtubule polymer and the concentration of assembly-competent soluble tubulin remain constant, so the effects of stathmin on dynamics due to a direct interaction with the microtubules can be analyzed in the absence changes in the level of assembly-competent soluble tubulin (27). We used a pH of 7.2 (the average pH of the cytoplasm in most mammalian cells, 34 -44). Although stathmin sequesters soluble tubulin more strongly at pH 6.8 than at pH 7.2 (17,27,29), it modulates dynamic instability by a direct action on the microtubules considerably more strongly at pH 7.2 than at pH 6.8 (27,29). At pH 7.2, unphosphorylated stathmin increases the catastrophe frequency strongly (conversion of a microtubule end from a growing or attenuated state to rapid shortening) with its effects at minus ends considerably stronger than at plus ends (27). In addition, unphosphorylated stathmin does not exert significant effects on other dynamic parameters except for an increase in the percentage of time the microtubules shorten and a slight compensatory increase in the growth rate (27).
To determine how phosphorylation of Ser 16 or Ser 63 changes the ability of stathmin to regulate microtubule dynamics, we analyzed the effects of p16 and p63 stathmin at three different ratios of stathmin to tubulin on the plus and minus ends of individual steady-state microtubules by video microscopy. As shown in Table 1 and Fig. 2A, phosphorylation of serine 16 essentially abolished the ability of stathmin to increase the catastrophe frequency at plus ends. The apparent increase in the catastrophe frequency from 0.16 per min to 0.22 per min at a 1: 5 molar ratio of p16 stathmin to tubulin was not statistically significant (Student's t test). At the same ratio, p16 stathmin had no detectable effect on the catastrophe frequency at minus ends (Table 1 and Fig. 2B). Consistent with previous results (27), a similar ratio of unphosphorylated stathmin to tubulin increased the catastrophe frequency ϳ2.5-fold at plus ends and ϳ8-fold at minus ends (Table 1 and Fig. 2, A and B). Unlike the effects of unphosphorylated stathmin on the increased percentage of time that microtubules shortened at plus and minus ends, p16 stathmin did not significantly change these parameters.   JUNE 5, 2009 • VOLUME 284 • NUMBER 23

JOURNAL OF BIOLOGICAL CHEMISTRY 15643
Phosphorylation of stathmin at Ser 63 also abolished the ability of stathmin to increase the catastrophe frequency at both ends (Table 2 and Fig. 3, A and B). Like p16 stathmin, p63 stathmin had no significant effect on total shortening time or on the rate of growth or shortening at either end (Table 2). We also prepared a stathmin isoform with all four serine residues, Ser 16 , Ser 25 , Ser 38 , and Ser 63 , phosphorylated. This isoform, as expected, behaved exactly as p63 stathmin, inducing no detectable change in the catastrophe frequency or any other dynamic instability parameter at either microtubule end (data not shown). To further eliminate any possibility that the inabilities of p16 or p63 stathmin to increase the catastrophe frequency at plus or minus ends were due to mutations of the remaining serines to alanines, we analyzed the ability of 4A stathmin to modulate dynamic instability and found, as expected, that it increased the catastrophe frequency at plus and minus ends in the same way and to the same extent as unphosphorylated stathmin (Fig. 2, A and B). 25 and Ser 38 occur through the action of cyclin-dependent kinase 1, a different kinase than that which phosphorylates Ser 16 and Ser 63 (3,13,14). Phosphorylation of these two serine residues has been thought to play a minor role in modulating the microtubule regulatory activity of stathmin in cells (3,14). Thus, we wanted to determine the effects of stathmin doubly phosphorylated on these two residues (p25p38 stathmin, see "Experimental Procedures") on microtubule assembly, on formation of T2S complexes, and on steady-state dynamic instability. As shown in Fig. 4A, p25p38 stathmin reduced the steady-state   microtubule polymer mass as efficiently as unphosphorylated and 4A stathmin. Specifically, a 1:5 molar ratio of p25p38 stathmin to tubulin reduced microtubule polymerization by ϳ50% (Fig. 4A), which is similar to the extent to which 4A stathmin (Fig. 4A), and unphosphorylated stathmin inhibit polymerization (data not shown). Consistent with its strong ability to reduce microtubule polymerization, a 1:5 molar ratio of p25p38 stathmin to tubulin also formed a stable T2S complex as efficiently as 4A stathmin (Fig. 4B) or unphosphorylated wild-type stathmin (data not shown). Thus, phosphorylation at Ser 25 and Ser 38 together has no significant effect on the ability of stathmin to sequester soluble tubulin or to inhibit microtubule polymerization.

Effects of Stathmin Doubly Phosphorylated at Ser 25 and Ser 38 on Microtubule Polymerization, Sequestration of Soluble Tubulin, and Dynamic Instability-Phosphorylation of Ser
We also analyzed the effects of phosphorylation at Ser 25 and Ser 38 on the ability of stathmin to modulate dynamic instability at steady state in vitro. As shown in Table 3, p25p38 stathmin increased the catastrophe frequency at both plus and minus ends as strongly as unphosphorylated stathmin. A 1:5 molar ratio of p25p38 stathmin to tubulin increased the plus end catastrophe frequency by ϳ2.2-fold (Table 3 and Fig. 5A). At minus ends, a similar ratio of p25p38 stathmin to tubulin increased the catastrophe frequency by ϳ9-fold (Table 3 and Fig. 5B). Similar to the effects of unphosphorylated stathmin, p25p38 stathmin significantly increased the fraction of time the microtubules shortened at both plus and minus ends and moderately increased the growth rates at plus ends (compare Tables  1 and 3).
Binding of p16 Stathmin, p63 Stathmin, and p25p38 Stathmin to Microtubules-Consistent with its ability to increase the catastrophe frequency at the ends of microtubules at steady state, stathmin can also bind directly to purified microtubules (27). To determine whether the reduced effects of the phosphorylated stathmin isoforms on the steady-state catastrophe frequency at microtubule ends correlates with an inability to bind to microtubules, we analyzed the binding of p16 stathmin, p63 stathmin, and p25p38 stathmin to microtubules in vitro by using an immunofluorescence-based method (27). Mixtures of 40 M tubulin and 8 M p16 stathmin, p63 stathmin, or p25p38 stathmin, respectively, were assembled to steady state and sedimented through sucrose cushions to avoid any nonspecific binding of stathmin to microtubules. Stathmin binding to the microtubules was visualized by immunofluorescence microscopy with antibodies specific for tubulin and stathmin. Consistent with previously published reports, unphosphorylated stathmin and 4A stathmin bound to microtubules along their lengths as determined by co-staining of the microtubules with  tubulin and stathmin antibodies (only data with 4A stathmin are shown (Fig. 6, C and D). However, no binding of p63 stathmin or p16 stathmin to microtubules could be detected as shown by the lack of stathmin staining on the microtubules (Fig. 6, G-J). In contrast, p25p38 stathmin did bind to the microtubules, consistent with its ability to increase the steadystate catastrophe frequency in a manner similar to unphosphorylated stathmin (Fig. 6, E and F).

DISCUSSION
In cells, the growth and shortening dynamics of individual microtubules, not just the quantity of microtubule polymer present, is critical for microtubule function, as for example, during mitosis when rapid dynamics are essential for the precise, rapid, and time-sensitive assembly and function of the mitotic spindle (45)(46)(47). Stathmin, one of the best known regulators of microtubule dynamics, destabilizes microtubules, functioning as a catastrophe promoting factor (increasing the frequency of switching from growth to shortening at microtubule ends). It acts in two ways, indirectly by sequestering tubulin into an assembly incompetent soluble T2S complex, and directly by binding to the microtubules (17,27,29). Sequestration of tubulin into T2S complexes destabilizes microtubules by reducing the quantity of assembly-competent tubulin and by increasing indirectly the catastrophe frequency as the microtubules depolymerize to a new steady state. The main goal of the present work was to determine how phosphorylation of each of stathmin's four phosphorylatable serine residues, Ser 16 , Ser 25 , Ser 38 , and Ser 63 , affect the ability of stathmin to regulate dynamic instability through a direct action on the microtubules. We analyzed the effects of stathmin on dynamic instability after the microtubules attained steady state; a condition at which a stable steady-state equilibrium between polymerized microtubules and soluble tubulin is maintained. At such an equilibrium condition, modulation of microtubule dynamic instability will be due to a direct action of either stathmin or of the T2S complex on the microtubules, and not to reduction of the level of assembly-competent tubulin (27,29). Ser 16 and Ser 63 are phosphorylated by different kinase systems than are Ser 25 and Ser 38 indicating that phosphorylation of stathmin at these two residues may have different effects on microtubule dynamics (3,25). We found that phosphorylation of stathmin individually at either Ser 16 or at Ser 63 strongly reduced the binding of stathmin to tubulin and microtubules, and formation of stable T2S complexes (Figs. 1 and 5). Consistent with the loss of tubulin and microtubule-binding ability, phosphorylation of either of these residues strongly reduced the ability of stathmin to increase the catastrophe frequency at plus and minus ends (Tables 1 and 2 and Figs. 2 and 3). In contrast, double phosphorylation of stathmin at Ser 25 and Ser 38 , the two other remaining phosphorylatable serines, had no effect on the binding of stathmin to tubulin, formation of T2S complexes, or on the ability of stathmin to bind directly to microtubules and act as a catastrophe factor (Table 3 and Fig. 4).
How Might Phosphorylation of Stathmin at Ser 16 or Ser 63 Abolish the Catastrophe-promoting Activity of Stathmin?-The inability of p16 stathmin to increase the steady-state catastrophe frequency at either microtubule end together with the lack of detectable binding to the microtubules indicates that phosphorylation of Ser 16 impairs the interaction of stathmin with tubulin in microtubules. As described in detail in Fig. 5A of Honnappa et al. (25) (see also Fig. 7), Ser 16 of stathmin is located within the tight turn between two ␤ strands that make a ␤-hairpin structure, the only structural part of the N terminus of stathmin, which is involved in the interaction with Asn 356 ␣1 tubulin of the T2S complex. ␣-Tubulin is exposed at the minus ends of microtubules, and if stathmin binds tubulin at minus ends in a manner similar to the way it binds to tubulin in the T2S complex, it is reasonable to conclude that Ser 16 of stathmin plays a critical role in regulating the catastrophe frequency at these ends ( Fig. 1 and Table 1). It is possible that phosphorylation of Ser 16 disfavors either the formation of the tight hairpin "cap" structure (21,25,48) or inhibits interaction of that structural region with the ␣1 tubulin at minus ends. Unlike unphosphorylated Ser 16 , whose side chain points toward the ␣1-tubulin surface in the T2S complex (25), phosphorylation of Ser 16 is expected to cause a steric problem and thus, the interaction of the ␤-hairpin at stathmin's N terminus with the exposed ␣-tubulin subunit at the minus end would be disrupted.
Stathmin binds to the walls of microtubules along their entire lengths ( Fig. 6) (27). It is not known whether its binding along the lengths of the microtubules increases the catastrophe-promoting activity at the ends. It is possible that binding to tubulin in the walls might produce strain within the microtubule lattice causing an increased catastrophe frequency at the ends. Phosphorylation of Ser 16 , however, reduced the binding of stathmin along the length of microtubules to an undetectable level by immunofluorescence microscopy (Fig. 6), suggesting that the phosphoryl group at this position specifically hinders the binding of stathmin to tubulin along the entire microtubule surface including at or near the ends. Interestingly, phosphorylation of Ser 16 strongly reduced but did not completely abolish stathmin's steady-state plus end catastrophe frequency (Table 1 and Fig. 2). It is possible that p16 stathmin interacts with tubulin at plus ends with a very low affinity and thus is undetectable by immunofluorescence microscopy. The ability of p16 stathmin to weakly sequester free tubulin dimers supports this possibility (Fig. 1B). The N terminus of stathmin has been implicated as being involved in stimulating the catastrophe frequency at plus ends (29), although the mechanism for how it could do so is still unclear. A reasonable way stathmin might increase the catastrophe frequency at plus ends is that its N terminus may somehow nestle between the two tubulin dimers (between the ␣1 subunit of outer exposed tubulin dimer and the ␤2 subunit of the adjacent tubulin dimer) along individual protofilaments to attain more structural stability or a more energetically favorable conformation, thus breaking the tubulin-tubulin dimer contacts at the ends. Because Ser 16 phosphorylation interferes with tubulin binding (24,25), the alternation may reduce its ability to nestle between the adjacent tubulin dimers thus reducing its catastrophe-promoting activity.
The complete loss of the tubulin binding and sequestering activity of stathmin (Fig. 1) and its steady-state catastrophepromoting activity at both microtubule ends, together with the loss of its ability to bind to microtubules when Ser 63 is phosphorylated (Fig. 6), indicates that phosphorylation of Ser 63 completely abolishes the functional interactions of stathmin both with soluble tubulin and with tubulin in microtubules. These results are consistent with studies indicating that phosphorylation of Ser 63 significantly changes the secondary structure of stathmin, reducing the overall structural organization of its helical backbone and thus hindering alignment of the neighboring amino acid residues that maintain the tight interaction of stathmin with ␣1 tubulin in the T2S complex ( Fig. 7) (21,29,49).
Possible Role of Ser 25 and Ser 38 Phosphorylation on Stathmin Activity-p25p38 stathmin increases the steady-state catastrophe frequency at both ends as efficiently as unphosphorylated stathmin (Table 3 and Fig. 4). It also binds to the microtubules (Fig. 6). These data indicate that Ser 25 and Ser 38 by themselves are minimally involved in the interaction of stathmin with soluble tubulin or with tubulin in microtubules. Analysis of the structure of the T2S complex indicates that Ser 25 and Ser 38 are located at a poorly ordered loop region of the complex and are not in contact with tubulin ( Fig. 7) (19,20,25). Phosphorylation of Ser 25 or Ser 38 also does not significantly alter the secondary structure of stathmin as determined by CD spectroscopy (25). This raises the question of how phosphorylation of Ser 25 or Ser 38 might be involved in controlling cell cycle progression through an action on microtubules without affecting the microtubule destabilizing activity of stathmin? Phosphorylation of Ser 25 and Ser 38 appears to be a prerequisite for phosphorylation of Ser 16 and Ser 63 during mitosis (3,14), and one possibility in the case of mitosis is that phosphorylation of these residues may facilitate the activity of kinases that phosphorylate Ser 16 and Ser 63 .
Implications for Regulation of Minus End Dynamics by Stathmin-Stathmin increases the catastrophe frequency considerably more strongly at minus ends than at plus ends (27). Minus ends do not grow in cells, but rather, they either remain stable or they depolymerize (50 -53). For example, during metaphase and anaphase of mitosis, spindle microtubules in vertebrate cells depolymerize from their minus ends while remaining tethered at the spindle poles in a process called poleward flux that appears to be crucial for chromosome segregation and separation of the spindle poles (53)(54)(55). Cellular factors that modulate microtubule shortening at minus ends have been suggested to regulate poleward flux (27,(53)(54)(55)(56)(57). Stathmin localizes at the minus ends of spindle microtubules at spindle poles in mitotic HeLa cells (6). Thus, stathmin might increase the flux rate by destabilizing minus ends at the poles. Although stathmin is predominantly phosphorylated during mitosis, a fraction of stathmin remains unphosphorylated even in cells blocked in mitosis for prolonged periods (58). The small amount of remaining unphosphorylated stathmin could represent a pool of active stathmin that can bind directly to microtubule minus ends to exquisitely regulate minus-end dynamics. Stathmin might even facilitate flux at minus ends when Ser 25 and Ser 38 are phosphorylated, because phosphorylation of these serine residues does not diminish its minus-end catastrophe-promoting activity (Table 3 and Fig. 5B).

Regulation of Microtubule Dynamics by Stathmin Phosphorylation-It
is not yet understood how stathmin's abilities to destabilize microtubules and to function as a catastrophe-promoting factor are regulated by phosphorylation at its different kinase-specific phosphorylation sites in cells.
Our data indicate that the catastrophe-promoting activity of stathmin can be inactivated in similar fashion by phosphorylation of either Ser 16 or Ser 63 , and that its catastrophepromoting activity is not regulated in the absence of other factors by phosphorylation of Ser 25 and Ser 38 . Based upon our data, it is reasonable to hypothesize that in cells, the catastrophe-promoting activity of stathmin can be controlled in a similar fashion by phosphorylation of either Ser 16 or Ser 63 . Because the kinases acting at these residues are distinct and might function through different pathways, it is reasonable to think that a similar loss of function can be achieved by multiple pathways, thus providing the cell with a finely tunable mechanism for controlling microtubule assembly and dynamics in relation to its needs.