Calcium-independent, pH-regulated Effects of S- 100 Proteins on Assembly-Disassembly of Brain Microtubule Protein in Vitro*

At alkaline pH, Ca" is no longer required for S- 100 proteins to inhibit the assembly and to promote the disassembly of brain microtubules in vitro, though the presence of Caa+ significantly favors the S-100 effects. These effects are inversely related to the microtubule protein concentration and directly related to the 5-100 concentration and the pH. Ca'+-independent, pH-reg-ulated inhibition of assembly of phosphocellulose-pu- rified tubulin by 5-100 is also described. The microtubule disassembling effect of s- 100 is additive to that of alkali (used to raise the pH), and 5-100 further disassembles microtubules after alkalinization. Thus the larger inhibitory effect of S-100 on microtubule assembly at alkaline uersus acid pH depends on both a decrease in the assembly rate and an increase in the disassembly rate. Together with previous data on this topic, the present findings indicate that 5-100 proteins act on microtubule protein in vitro primarily by bind- ing to tubulin, this event being Ca2+-regulated at a given pH, and pH-regulated at a given free Ca2+ concentration.

may play a role in the system regulating the state of assembly of MTs (9,12).
s-100 is a protein fraction originally purified from bovine brain, comprising three small ( M , = 21,000), acidic (pHi 4.3), Ca2+-binding isoforms, which are structurally related to calmodulin and other Ca2+-binding proteins and are modified in their secondary structure by Ca2+ to nearly the same extent (for a recent review see Ref. 13). Thus, the S-100 functions are expected to be mediated by Caz+. Bovine brain S-100, which contains all three isoforms (S-loo%, S-100a, and S-100b) and rat brain S-100, which is 2 95% S-100b (14), were recently shown to be involved in the control of the assembly-disassembly of brain MTP and to inhibit the assembly of purified tubulin in a Ca2+-mediated way in vitro (15)(16)(17)(18)(19). Individual S-100 isoforms appear to exert equipotent effects on MTP and purified tubulin assembly-disassembly (19). Turbidimetric and ultrastructural studies have shown that s-100 interferes with both the nucleation and the elongation of MTs in the presence of Ca2+ (18,20) by binding to tubulin (21). Also, some competition between S-100 and MAPs for binding to tubulin has been reported (21). Finally, S-100 proteins have been shown to increase the Ca2+ sensitivity of MTP, this effect being dependent on pH and on the KC1 concentration (22).
To obtain more detailed information on the role of pH in S-100 effects on MTP assembly-disassembly, we studied the assembly-disassembly of MTP and of purified tubulin in the absence and in the presence of S-100 at various pH values, in the presence and in the absence of Ca2+. Data obtained in the presence of Ca2+ are consistent with the idea that S-100 increasingly interferes with the nucleation and the elongation of MTs in uitro as the pH rises, at a given S-lOO/tubulin dimer molar ratio (23). This is most evident at pH 7.5, a condition where the MT number concentration and the mean MT length decrease in the presence of S-100 (23). The data obtained in the absence of Caz+ are presented here. We show that, as the pH rises, Ca2+ is no longer required for S-100 to inhibit the assembly of either MTP or purified tubulin, and to promote the disassembly of steady state MTs.

MATERIALS AND METHODS
Purification of Protein-S-100 was purified from bovine and rat brain as described (18). Bovine and rat S-100 proteins were checked for purity by polyacrylamide gel electrophoresis (PAGE) both in the presence (Fig. 1, lanes A and B ) and in the absence (Fig. 1, lanes C-E ) of sodium dodecyl sulfate (SDS). The S-100 fraction obtained from bovine brain was a mixture of S-100a and S-1OOb (Fig. 11, which was fractionated no further. Its concentration was calculated by UV spectroscopy using the value of E&% = 0.344 (24) since the mixture used here contained nearly equal amounts of S-100a and S-100b by densitometric scanning of the gel (not shown). Rat brain S-100 was -100% S-100b (Fig. 1). Its concentration was calculated using the value of E " & % = 0.185 (14). MTP was purified from adult rat brain by three cycles of assembly-disassembly at pH 6.7 (25). Tubulin was separated from MAPS by phosphocellulose chromatography (2) of thrice-cycled MTP. The concentration of MTP and phosphocellulosepurified tubulin (PC-tubulin) was measured by the method of Lowry et al. (26). SDS-PAGE revealed that tubulin represented -80% of MTP (Fig. 1, lane F ) by densitometric scanning of the gel (not shown), that PC-tubulin was essentially pure (Fig. 1, lane G ) , and that MAPs consisted of high molecular weight polypeptides and of a minor percentage of low molecular weight polypeptides ( Fig. 1, lnne H). The tubulin dimer concentration of MTP preparations was calculated on the assumption that 1 mg of MTP contained 0.8 mg of tubulin, based on data in Fig. 1, lane F.
Assay of MTP Assembly-Disassembly-MTP assembly-disassembly and PC-tubulin assembly were continuously monitored spectrophotometrically at 37 "C as the change in absorbance at 350 nm (A350) (5). Details are given in the legends to figures and tables.
Other Procedures-Electrophoretic analyses were performed by the method of Laemmli (27) in the presence of SDS, and by the method of Isobe et al. (28) in the absence of SDS.

Effect of S-100 on MTP Assembly at Various pH Values in
the Absence of Ca2+-MTP assembles in the absence of Ca2+ (1 mM EGTA) with the same lag phase but at an ever slower rate and to an ever smaller extent as the pH rises from 6.7 to 7.5 ( Fig. 2), in accordance with previous data (9). In the presence of S-100 (from either bovine or rat brain), used at a molar ratio of 2.0 with respect to the tubulin dimer, a small but significant increase in the lag phase and a decrease in both the rate and the extent of assembly are registered at pH 6.7 ( Fig. 2 and Table I). As the pH rises to 7.1 and 7.5, the efficacy of S-100 as an inhibitor of MTP assembly increases linearly with the pH (Fig. 2 and Table I). When Ca2+ to 1 mM (free Ca2+) is added to MTP assembled to steady state, MTs formed in the presence of S-100 disassemble at a slower rate but to a larger extent than those formed in its absence at all pH values tested (Fig. 2). A comparative analysis of S-100 effects on MTP assembly in the absence and presence of Ca2+ is presented in Table I. Ca2+ mostly favors the S-100 effect at pH 6.7. As the pH rises, Ca2+ is no longer a prerequisite for S-100 to increase the lag phase and to decrease the rate and the extent of assembly, though its presence significantly potentiates the S-100 effects.  is inversely related to the MTP concentration ( Fig. 3A) and that the critical concentration for MTP assembly in the presence of S-100 increases, under these conditions, from 3.5 p M tubulin dimer to 4.5 pM at pH 6.7, from 5.5 to 9.0 p M at pH 7.1, and from 7.5 to 13.0 p~ at pH 7.5. proportional to the S-100 concentration and that, as the pH rises, an ever smaller S-100 concentration is required for a given percent inhibition of assembly to occur. In the presence of S-100, the lag phase increases and the rate of assembly decreases, particularly at pH 7.1 and 7.5 (not shown).

Effect of S-100 on Assembly of MTP Supplemented with MT Fragments at Various pH Values in the Absence of
Cu*+-The value of AS5,, at steady state is a measure of the MT mass, i.e. of the MT number concentration times the average MT length (29,30). In the absence of Ca2+, S-100 reduces the rate (not shown) and the extent (Fig. 3C) of the assembly of MTP supplemented with MT fragments. The effect is small at pH 6.7 and increases linearly with the pH. Practically, no lag phase is registered in these experiments, irrespective of the pH and of the absence or presence of S-100 (not shown).
Effect of S-100 on Assembly of PC-tubulin at Various pH Values in the Absence of Cu2'-S-100 from either bovine or rat brain also inhibits the assembly of PC-tubulin in a Caz+independent, pH-regulated way (Fig. 4). As observed with whole MTP, the inhibitory effect of S-100 on PC-tubulin assembly is small at pH 6.7 and most evident at pH 7.5, at a fixed S-lOO/tubulin dimer molar ratio, and, at the pH values tested, the effect is more intense on the initial rate than on the extent of assembly.

S-100-and Alknli-induced Disassembly of MTs at Various pH Values in the Absence of Ca2+-S-100
, added to steady state MTs at various pH values in the absence of Ca2+, produces an ever larger disassembly at an ever higher rate as the pH at which assembly was carried out rises (Table 11).

TABLE I1 Effect of S-100 on MTP assembled at various p H values in the
absence of Ca2+ MTP corresponding to 15.72 pM tubulin dimer was assembled to steady state in the absence of Ca2+ ( 1 mM EGTA) at the pH values indicated, after which S-100 was added to 20 p~ in a small volume. Changes in A360 were followed for 10 min after adding S-100. bly curves and expressed as the percent decrease in Axa/min.

TABLE I11 Combined effect of S-100 and alkali on MTP assembled at various pH values in the absence of Ca"
MTP corresponding to 18.23 pM tubulin dimer was assembled to steady state in the absence of Ca2+ ( 1 m M EGTA) at the pH values indicated, after which either NaOH (from a concentrated solution to raise the pH to 7.5) or NaOH plus S-100 (to a final concentration of 20 p~) was added in small volumes. Changes in A3m were followed for 10 min after additions, i.e. until a new steady state was attained.
In separate experiments, we standardized the minimal volume of the concentrated solution of NaOH to be added to microtubule suspensions in order to obtain the desired rise in pH. When added along with NaOH (used to raise the pH) to steady state MTs (Table 111), S-100 potentiates the effect of alkali, particularly if MTs were assembled at alkaline pH (7.1). Again, S-100 has a larger disassembling effect on MTs assembled at alkaline (7.1) than at acidic (6.7) pH, in agreement with data in Table 11. Moreover, S-100 further reduces the ASm value when it is added to a MT suspension after alkalinization (not shown). Thus, the strong inhibitory effect of S-100 on the extent of assembly at alkaline pH seems to depend on both a decrease in the assembly rate and on an increase in the disassembly rate.

DISCUSSION
The results presented in this paper show that 53-100 from either bovine brain (S-100a plus S-100b) or rat brain (S-100b) inhibits the assembly and promotes the disassembly of whole MTP and PC-tubulin in vitro in a Ca2+-independent, pHregulated way. S-100 increases the lag phase and decreases the rate and the extent of assembly pH dependently. These observations, together with the progressive increase in the critical concentration for MTP assembly, indicate that S-100 increasingly interferes with the nucleation as the pH rises, in analogy with what occurs under the same conditions in the presence of Ca2+ (18,23). Since the addition of preformed nuclei of MTs to a MTP solution does not abolish the S-100dependent inhibition of assembly under these conditions, it is conceivable that S-100 may also interfere with the elongation reaction in a pH-regulated way. On the other hand, data in Tables I1 and I11 suggest that the S-100 effect on the assembly of MTP supplemented with MT fragments at alkaline pH may not be entirely due to interference with the addition of tubulin dimers onto the growing MT polymers, but may also depend on S-100-induced disassembly of added fragments. Yet, adding S-100 to MTP previously assembled to steady state at various pH values and then sheared to increase the MT number concentration, while keeping the MT mass constant, results in the same disassembly as observed with undisturbed MTs (not shown), and diluting a suspension of MTs 10 times at various pH values produces identical disassembly curves in the absence and in the presence of S-100 (22). Thus, S-100-induced disassembly of added fragments does not contribute to data presented in Fig. 3C to a significant extent.
The extent of MTP assembly, measured as the A,,, value at steady state, is the net result of the rate at which MTP assembles and of the rate at which MTs disassemble. Data in Tables I1 and 111 show that the MT disassembling effect of S-100 increases with the pH. This explains why experimental points reported in Fig. 3B fall on lines rather than on curves, at least in the S-100 concentration range tested.
The present data further stress the fact that S-100 acts on tubulin (18,21,31). Quantitative analyses indicate that the stoichiometry of S-100 binding to PC-tubulin increases with increasing free Ca2+ concentrations at pH 6.7, and with the pH (21). Also, at alkaline pH (7.4), Ca2+ is no longer required for S-100 binding to PC-tubulin (21). This Ca2+-independent binding of S-100 to tubulin at alkaline pH explains the results presented here.
On the contrary, calmodulin has no effect on assembly of purified tubulin (32) and acts on MTP by binding to T and MAP2 (32,33). Of course, the present experiments do not exclude the possibility that S-100 may also interfere with the activity of MAPs. However, data presented elsewhere indicate that S-100-tubulin complexes can be recovered after gel chromatography (Sephadex G-200) and after precipitation with ammonium sulfate, while the same does not occur when S-100 is reacted with MAPs (21). This indicates that the affinity of S-100 for tubulin is higher than that for MAPs. Recently, it has been reported that tubulin and, to a lesser extent, T factors bind to Sepharose-immobilized S-100 (34). Binding of S-100 to T factors in the turbidity experiments presented here might play a role in the overall effect of S-100 on assemblydisassembly of whole MTP. However, the observation that S-100 inhibits the assembly of PC-tubulin indicates that binding to T and to other MAPs is not essential for S-100 to affect the MT assembly-disassembly. Also, in view of the fact that S-100 is not present in axons (Ref. 12) and T factors appear to be confined to axons (see Ref. 35 and references cited therein), the interaction between S-100 and T factors appears of questionable biological significance.
In spite of the structural relatedness of S-100 to calmodulin and other Ca2+-binding proteins (36,37), the affinity of S-100 for Caz+ is not high, particularly in the presence of physiologic K+ concentrations, with K' antagonizing the binding of Ca2+ to high affinity sites on S-100 (38). Within the cell, $100 is thus expected not to be capable of amplifying the Ca2+ signal. Yet, S-100 affects the MTP assembly both in the absence and in the presence of K+ (18,22). The increased lability of MTs in the presence of K+ (39) cannot explain this effect of S-100 completely (22). In addition, since p M Ca2+ levels do not saturate the Ca2+-binding sites on 5-100 (38), one is forced to postulate a Ca2+-independent ability of S-100 to affect the MTP assembly, though Ca2+ significantly increases the inhibitory effect of S-100 (Table I). Precedent exists for Ca2+independent association of S-100, as well as of calmodulin, with target proteins (40)(41)(42)(43)(44)(45)(46)(47). Particularly, data have been presented suggestive of Ca2+-independent association of calmodulin with MTs (44).
At alkaline pH, the tubulin-tubulin interactions are expected to be reduced since the tubulin molecules become more negatively charged. This is in line with the decrease in the ability of purified tubulin to polymerize at alkaline pH (48) (see also Fig. 4) and with the observation that the enzymatic cleavage of a small fragment from the C-terminal end of both a and @ subunits of tubulin, which is negatively charged, results in an increase in the ability of purified tubulin to polymerize (49, 50). It is not clear at present how the rise in pH, which also induces an increase in the net negative charge of S-100, results in the potentiation of the S-100 inhibitory effect on MTP and PC-tubulin assembly. MAPs were reported to coassemble with tubulin at a constant stoichiometry irrespective of pH (9), and S-100 seems to interact with tubulin in a pH-dependent way (21, 31), whereas some uncertainty still remains on S-100 binding to MAPs (21, 34). The critical concentration for MTP assembly increases dramatically with the pH in the presence of S-100 and absence of Ca2+ (Fig.  3A), indicating that an increasingly larger proportion of MTP does not take part in the formation of MTs. Since the ability of S-100 to disassemble steady state MTs also increases with the pH, one can tentatively conclude that the steady state tubulin flux or treadmilling is increased by the presence of S-100. In this respect, since the treadmilling increases with increasing pH (9), the effect of S-100 seems to add to that of pH. Accordingly, while barely detectable differences in the MT number concentration and in the mean MT length are observed at pH 6.7 in the presence of S-100, a decrease in both parameters is registered at pH 7.5 in the presence of S-100 (not shown). Similar experiments performed in the presence of 10 ~C I M free Ca2+ gave qualitatively similar results at pH 7.5, while, due to the presence of Ca2+, a significant decrease in the MT number concentration and an increase in the mean MT length were registered at pH 6.7 (23). The latter finding is due to the small MT disassembling effect of S-100 and to the larger effect of S-100 on the nucleation than on the elongation at pH 6.7 (20, 23).
On the basis of kinetic data and on counts of MTs formed in the absence and presence of S-100 we proposed that S-100 interferes with both the nucleation and the elongation of MTs by sequestering unassembled tubulin (18, 23). The effect on elongation is most evident at pH 7.5, both in the presence (23) and in the absence (Fig. 3C) of Ca2+. However, we cannot exclude the possibility that S-100 may act directly on MTs. Actually, preliminary data suggest that S-100 interacts with steady state MTs at pH 6.7 even under conditions where it does not produce disassembly, i.e. in the absence of high Ca2+ concentrations (51). Moreover, the rapid and nearly complete disassembly produced by the simultaneous addition of S-100 and 1 mM Caz+ to steady state MTs (15, 17) and the S-100induced increase in the rate of dilution-dependent disassembly of MTs in the presence of Ca2+ (22) cannot be exclusively due to sequestration of unassembled tubulin. Thus, the effect of tubulin on MTP assembly-disassembly appears to be a complex event where the key step is the binding to tubulin. The latter is regulated by Ca2+ at a given pH and by pH at a given free Ca2+ concentration (21). The available evidence indicates that, owing to the formation of S-100-tubulin complexes, a smaller percentage of tubulin molecules is available for MT formation and that the nearly complete disassembly of steady state MTs on addition of S-100 plus Ca2+ might result from S-100 binding to MTs with consequent polymer destabilization.
As mentioned above, the pH seems to regulate the state of assembly of MTs both in vitro (9) and in vivo (12). We show here that Ca2+ is not a prerequisite for S-100 to affect the MTP assembly-disassembly, particularly at alkaline, but still physiologic, pH. This suggests the possibility that, even in the absence of waves of free Ca", S-100 may intervene in the control of MTP assembly-disassembly in a pH-mediated way.