calmodulin-binding proteins in brain

11
Neurochemistry International, Vot. 5, No. 2, pp, 159 to 169. 1983 0197-0186/83/020159-11 $03.00/0 Printed in Great Britain © 1983 Pergamon Press Ltd. COMMENTARY CALMODULIN-BINDING PROTEINS IN BRAIN SHIRO KAKIUCHI Department of Neurochemistry and Neuropharmacology, Institute of Higher Nervous Activity, Osaka University Medical School, Kita-ku, Osaka 530, Japan (Received 20 July 1982; accepted 24 August 1982) Abstract--It is now widely accepted that actions of intracellular Ca 2+ are mediated by a four-domain Ca 2+- binding protein, calmodulin. Brain is especially rich in calmodulin, containing about 400 mg (24#mol) of EGTA-extractable calmodulin per kg of brain. However, only a fraction of the above amount is required for the calmodulin-activated enzymes and most of the rest may be assigned to calmodulin-binding proteins, proteins which are apparently devoid of enzyme activities but undergo Ca2+-dependent associations with calmodulin. Several of such proteins have been recently discovered in brain. These include a heat-labile 80 K phosphodies- terase inhibitor protein (calcineurin), a heat-stable 70 K phosphodiesterase inhibitor protein, a 50 K protein, myelin basic protein, tubulin, microtubule z (tau) factor, a spectrin-like doublet protein (240 plus 235 K) (calspectin; fodrin) and a particle-associated 155 K protein. Functions of these calmodulin-binding proteins have not been fully elucidated yet. Some proteins may be calmodulin-regulated enzymes catalyzing yet unknown biochemical reactions, e.g. a protein phosphatase ac- tivity was found for calcineurin. Some proteins may interact with contractile elements or cytoskeleton of the cell, e.g. ~ factor and calspectin interacted with tubulin and F-actin, respectivelyand tubulin itself is a calmodu- lin-binding protein. So, interesting possibilities are the regulation of the functions of cytoskeleton by calmodu- lin through these calmodulin-binding proteins. Regulation of microtubule assembly by Ca2÷-dependent bind- ing of calmodulin to tubulin and/or r factor and possible involvement of calspectin in the mechanism regulating axonal transport of neuronal proteins have been suggested. Thus, the exploration of the regulating functions of Ca2+/calmodulin in brain depends largely upon the further study of the properties of these calmodulin-binding proteins. In 1970, two independent lines of research led to the discovery of calmodulin. In one line, Kakiuchi and Yamazaki (1970) first discovered a type of phospho- diesterase whose activity is dependent upon the pres- ence of a minute amount of Ca 2÷ (Ca2+-activated phosphodiesterase). Since they used a crude tissue extract as the enzyme preparation, certain amount of endogenous calmodulin must have been contained in the preparation. In a subsequent study, they demon- strated in a brain extract the presence of a heat-stable protein factor that is required for the activation of the enzyme activity by Ca 2+ and they isolated such a factor from the enzyme by gel filtration column chromatography of the brain extract (Kakiuchi, Yamazaki and Nakajirna, 1970). Quite independently, Cheung (1970) found that the activity of brain phos- phodiesterase decreases upon purification and the addition of an activator protein which was present initially in the crude enzyme preparation but subse- quently removed from the enzyme during the purifica- tion steps fully restores the enzyme activity. Subse- N.C.L 5/2- -A quently, he partially purified this protein activator from a brain extract (Cheung, 1971). In these studies, however, he was not aware of the role of Ca 2÷ in the enzyme activation. Since only a minute amount of Ca 2+ (~10-6M) is sufficient to activate enzyme, a contaminating Ca 2÷ in the enzyme preparation must have played the role. Later, these two independent lines of research merged when the identity of the two proteins as a Ca 2 +-binding protein was finally established (Teo and Wang, 1973; Kakiuchi, Yamazaki, Teshima and Uenishi, 1973; Lin, Liu and Cheung, 1974). Since then this protein, nowadays called calmodulin, has been shown to cause Ca2+-dependent activation of a number of enzymes (e.g. see Klee, Crouch and Rich- man, 1980). Activation of enzyme by calmodulin requires the formation of an enzyme-calmodulin complex in the presence of Ca 2 +. The first example of such complex formation was demonstrated by Tesh- ima and Kakiuchi (1974) by using brain phosphodies- terase. They separated the complex from free forms of 159

Upload: shiro

Post on 06-Jan-2017

212 views

Category:

Documents


0 download

TRANSCRIPT

Neurochemistry International, Vot. 5, No. 2, pp, 159 to 169. 1983 0197-0186/83/020159-11 $03.00/0 Printed in Great Britain © 1983 Pergamon Press Ltd.

COMMENTARY

CALMODULIN-BINDING PROTEINS IN BRAIN

SHIRO KAKIUCHI

Department of Neurochemistry and Neuropharmacology, Institute of Higher Nervous Activity, Osaka University Medical School, Kita-ku, Osaka 530, Japan

(Received 20 July 1982; accepted 24 August 1982)

Abstract--It is now widely accepted that actions of intracellular Ca 2+ are mediated by a four-domain Ca 2+- binding protein, calmodulin. Brain is especially rich in calmodulin, containing about 400 mg (24#mol) of EGTA-extractable calmodulin per kg of brain. However, only a fraction of the above amount is required for the calmodulin-activated enzymes and most of the rest may be assigned to calmodulin-binding proteins, proteins which are apparently devoid of enzyme activities but undergo Ca 2 +-dependent associations with calmodulin. Several of such proteins have been recently discovered in brain. These include a heat-labile 80 K phosphodies- terase inhibitor protein (calcineurin), a heat-stable 70 K phosphodiesterase inhibitor protein, a 50 K protein, myelin basic protein, tubulin, microtubule z (tau) factor, a spectrin-like doublet protein (240 plus 235 K) (calspectin; fodrin) and a particle-associated 155 K protein.

Functions of these calmodulin-binding proteins have not been fully elucidated yet. Some proteins may be calmodulin-regulated enzymes catalyzing yet unknown biochemical reactions, e.g. a protein phosphatase ac- tivity was found for calcineurin. Some proteins may interact with contractile elements or cytoskeleton of the cell, e.g. ~ factor and calspectin interacted with tubulin and F-actin, respectively and tubulin itself is a calmodu- lin-binding protein. So, interesting possibilities are the regulation of the functions of cytoskeleton by calmodu- lin through these calmodulin-binding proteins. Regulation of microtubule assembly by Ca2÷-dependent bind- ing of calmodulin to tubulin and/or r factor and possible involvement of calspectin in the mechanism regulating axonal transport of neuronal proteins have been suggested. Thus, the exploration of the regulating functions of Ca 2 +/calmodulin in brain depends largely upon the further study of the properties of these calmodulin-binding proteins.

In 1970, two independent lines of research led to the discovery of calmodulin. In one line, Kakiuchi and Yamazaki (1970) first discovered a type of phospho- diesterase whose activity is dependent upon the pres- ence of a minute amount of Ca 2÷ (Ca2+-activated phosphodiesterase). Since they used a crude tissue extract as the enzyme preparation, certain amount of endogenous calmodulin must have been contained in the preparation. In a subsequent study, they demon- strated in a brain extract the presence of a heat-stable protein factor that is required for the activation of the enzyme activity by Ca 2+ and they isolated such a factor from the enzyme by gel filtration column chromatography of the brain extract (Kakiuchi, Yamazaki and Nakajirna, 1970). Quite independently, Cheung (1970) found that the activity of brain phos- phodiesterase decreases upon purification and the addition of an activator protein which was present initially in the crude enzyme preparation but subse- quently removed from the enzyme during the purifica- tion steps fully restores the enzyme activity. Subse-

N.C.L 5/2- -A

quently, he partially purified this protein activator from a brain extract (Cheung, 1971). In these studies, however, he was not aware of the role of Ca 2 ÷ in the enzyme activation. Since only a minute amount of Ca 2+ (~10-6M) is sufficient to activate enzyme, a contaminating Ca 2 ÷ in the enzyme preparation must have played the role.

Later, these two independent lines of research merged when the identity of the two proteins as a Ca 2 +-binding protein was finally established (Teo and Wang, 1973; Kakiuchi, Yamazaki, Teshima and Uenishi, 1973; Lin, Liu and Cheung, 1974). Since then this protein, nowadays called calmodulin, has been shown to cause Ca2+-dependent activation of a number of enzymes (e.g. see Klee, Crouch and Rich- man, 1980). Activation of enzyme by calmodulin requires the formation of an enzyme-calmodulin complex in the presence of Ca 2 +. The first example of such complex formation was demonstrated by Tesh- ima and Kakiuchi (1974) by using brain phosphodies- terase. They separated the complex from free forms of

159

16() ~IIIR() K ~,KII ~ Ill

proteins by Sephadex G-200 gel filtration column chromatography. Thus, the meclmnism of the acti- vation of enzyme by Ca 2 and calmodt, lin was for-

mulated as follows:

CALMODULIN + C a 2+ ~ CALMODULIN*.Ca -~ ENZYME + CALMODULIN*.Ca n.

ENZYME*- CALMODULIN*.Ca2 ' (* activated form of proteinsl

This formulation was later confirmed by Lin, Liu and Cheung (1975).

Brain is especially rich in calmodulin over other tissues (Kakiuchi, Yasuda, Yamazaki, Teshima, Kanda, Kakiuchi and Sobue, 1982c)(see Table 1).

About 400 mg (24 #mol) of calmodulin per kg of brain tissue was found in the cytosol fraction when the cytosol fraction was separated from the particulate

fraction in the presence of EGTA. This amount is in excess of that required for activation of enzymes. For instance, according to Sharma, Desai, Waisman and

Wang (1979), the amount of cahnodulin required I~ brain phosphodieslerase is 0.07 ;tmol per kg of wet

weight, which is less than 1300 of the total amount of

cahnodulin found in this tissue (see Table 2}. Ih i s discrepancy was dissolved when calmodulin-binding proteins which are apparently devoid of enzyme ac- tivities but undergo Ca-" -dependen t associalions with calmodulin were discovered. The first such

example was discovered in 1976 by Wang and Desai. Since then several proteins of this category have been

reported (Table 2l. Thus, the elucidation of the roles of calmodulin in

the brain is largely dependent upon the studies on these calmodulin-binding proteins. Little is as yet known about their functions at the present time. Some proteins may be calmodulin-regulated enzymes catalyzing yet unknown biochemical reactions. Another interesting aspect of these proteins is that

some proteins may interact with contractile elements or the cytoskeleton of the cell. In these cases, calmo- dulin can regulate the fimction of the cytoskeleton

Table 1. Calmodulin concentrations in rat and bovine tissues (Kakiuchi et al.. 1982c).

Calmodulin concentration

Tissue

In particulate fraction In soluble fraction (EGTA-nonextractable formi Total

Itg Itg l~g l~g Itg "~, of

g of tissue mg prof. I~M g of tissue mg prot. total g of tissue

Rat tissue Cerebral cortex (4)* 435 ± 5t 17.8 26.1 163 ± 7 2.69 27 598 Testis (3) 427 + 14 12.1 25.6 51.6 ± 4.8 1.46 11 479 Cerebellum (3) 312 ± 27 11.5 18.7 123 ± I4 1.69 28 435 Lung (6) 95.3 ± 2.7 1.38 5.7 55.4 +_ 5.4 (/.78 37 151 Adrenal gland (3) 89.8 ± 1.4 4.65 5.4 23.0 _+ 5.2 0.67 20 113 Prostate (31 93.0 ± 4.4 3.26 5.6 19.4 ± 1.5 0.26 17 112 Liver (7) 78.6 4__ 2.2 0.89 4.7 27.6 ± 3.1 (}.31 26 106 Kidney (6) 49.5 ± 6.4 0.89 3.() 41.5 ± 3.6 0.51 46 91.0 Spleen (3) 30.2 ± 3.2 0.47 1.8 52.3 ± 5.6 1.02 63 82.5

Bovine tissue Cerebral cortex (4) 432 +_ 29 16.9 25.9 162 ± 23 2.16 27 594 Caudate nucleus (4) 281 ± 18 16.2 16.8 193 ± 13 3.31 41 474 Hippocampus (4) 295 + 17 15.2 17.7 128 ± 3 2.44 30 423 Pituitary, ant. (2) 186 4.01 I 1.2 38.6 0.32 17 225 Hypothalamus (4j 173 +_ 14 8.80 10.4 50.4 ± 7.3 0.98 23 223 Pituitary, post. (1) 168 3.60 10.1 28.4 0.44 14 196 Adrenal medulla (2) 162 3.84 9.7 32.3 [I.41 17 194 Adrenal cortex (1) 135 3.54 8.1 22.0 (}.27 14 157 Brain white matterJ/(4) 55.5 ± 3.6 5.23 3.3 ND~ ND 55.5

* Numbers of determinations are shown in parenthesis. t Values are means ± S.E.M. 1~ For brain white matter, crura cerehel lL was used. § ND = Not detectable.

Calmodulin-binding proteins in brain

Table 2. Concentrations of calmodulin-binding proteins in brain

161

Concentration

Protein (mg/kg) (/lmol/kg) Reference

Calmodulin (soluble form) 400 Soluble calmodulin-binding proteins

Calcineurin 52 70 k protein 0.2 50 k protein 140 phosphodiesterase 10

Particulate form of calmodulin-binding protein Spectrin-like protein 1000" 155 k protein 220

24 Kakiuchi et al., 1982c

0.7 Sharma et al., 1979 0.003 Sbarma et al., 1979 2.8 Maekawa and Abe, 1980 0.07 Sharma et al., 1979

4.21" Kakiuchi et al., 1982b 1.4 Kakiucbi et al., 1982a

* Sum of the amounts in both the particulate and soluble fractions. t Calculated as a monomer.

through these calmodulin-binding proteins. The third possibility is that some proteins may serve as calmo- dulin-storage systems controlling the concentration of calmodulin in the cell.

In the following sections, I shall describe in detail each of the calmodulin-binding proteins found in the brain tissue to date.

I. A HEAT-LABILE PHOSPHODIESTERASE INHIBITOR PROTEIN: CALCINEURIN

Wang and Desai (1976) demonstrated in a bovine brain extract the presence of a protein exhibiting in- hibitory activity upon the Ca 2 +- and calmodulin-acti- vated phosphodiesterase. This phosphodiesterase in- hibitor protein represents the first instance of the calmodulin-binding protein since, as clarified by the same workers in a subsequent study (Wang and Desai, 1977), the inhibitory activity of this protein is due to a complex formation between this protein and enzyme in the presence of ~a 2+. This protein also abolished the Ca 2÷- and calmodulin-dependent acti- vation of the erythrocyte membrane Ca2+-pump ATPase (Sharma e t al., 1979) and that of brain ad- enylate cyclase (Wallace, Lynch, Tallant and Cheung, 1978). The inhibitor protein was preferentially bound to the calmodulin-Sepharose column over phospho- diesterase (Klee and Krinks, 1978), indicating that the calmodulin associates with the inhibitor protein with a higher affinity than with phosphodiesterase. In the presence of cyclic AMP, the affinity of calmodulin for phosphodiesterase increased, while that for the inhibi- tor protein being unaltered (Cheung, Lynch, Wallace and Tallant, 1981). The result is interpreted to indi- cate that, if the function of the inhibitor protein is to control the activity of phosphodiesterase, the level of

cyclic AMP, the substrate of the phosphodiesterase reaction, controls the susceptibility of the Ca 2÷- dependent activation of enzyme to the inhibition by the inhibitor protein.

The inhibitor protein was purified from the brain by three independent research groups i.e. Sharma e t

al. (1979), Klee and Krinks (1978) and Wallace, Lynch, Tallant and Cheung (1979). The purified pro- tein was shown to be composed of two polypeptide chains which are designated as subunits A and B. A structure of A-B with estimated Mr of 61,000 and 15,000 for A and B, respectively, was reported by Klee, Crouch and Krinks (1979) for the native form of this protein. According to Wallace e t al. (1979), corre- sponding Mr values are 60,000 and 18,500. Sharma e t

al. (1979) proposed a slightly different structure of A-B2 with M r of 60,000 and 14,500. Subunit A is responsible for the calmodulin binding and a 1:1 complex between two proteins is formed in a Ca 2 +- dependent manner (Sharma e t al., 1979; Klee et al.,

1979). Association of subunits A and B does not require the presence of Ca 2÷ (Klee e t al., 1979).

The protein is rich in brain. According to Sharma e t al. (1979) its concentration is 52 mg (0.7/~mol) per kg of brain (Table 2). Klee e t al. (1979) could not detect this protein in other tissues by direct assay method. Because of its specific location in the brain tissue, Klee e t al. (1979) has proposed to call this protein calcineurin. The immunofluorescence localiz- ation of calcineurin revealed that it is associated with neuronal elements only. In nerve cells it was local- ized in the postsynaptic densities and dendritic micro- tubules (Wood, Wallace, Whitaker and Cheung, 1980).

The function of calcineurin has not been estab- lished. Klee e t al. (1979) found that calcineurin is also

] fi2 SHIRO K AKII ( H I

a Ca2'-binding protein. It bound 4 mol of Ca 2 ~ per mol with a K a value less than 10 ~ M (determined in the presence of l mM MgClz~. This K d value for Ca 2+ is even smaller than that of calmodulm (2 3 x 10 ~ M) measured under the same condition. Ca-" + bound to subunit B. They suggested a possible role for calcineurin as a regulator of the free Ca e ~ concentration in the nerve cell (Klee et al., 1979).

In a recent report by Stewart, Ingebristen, Mana- lan, Klee and Cohen (19821, purified calcineurin was shown to have a protein phosphatase activity with the same properties and specific activity as protein phos- phatase-2B of skeletal muscle, an enzyme which specifically dephosphorylates the ~-subunit of glyco- gen phosphorylase kinase. The result seems to be inconsistent with the observation by Klee et al. (1979) that calcineurin is specifically located in brain.

2. A HEAT-STABLE 70 K CALMODULIN-BINDING PROTEIN

Sharma, Wirch and Wang (1978a) found in brain another inhibitor protein for Ca: +- and calmodulin- activated phosphodiesterase and they subsequently purified it to homogeneity (Sharma, Desai, Thompson and Wang, 1978b). Like calcineurin this protein in- hibits the activity of the phosphodiesterase by form- ing a Ca2+-dependent complex with calmodulin. It differed from calcineurin, however, in that its activity as a calmodulin-binding protein is stable upon boiling for 5 min in either 1 N NaOH or I N HC1. A molecu- lar weight of 68,000 or 70,000 was estimated by SDS- polyacrylamide gel electrophoresis or by gel filtration column chromatography, respectively, for the purified protein (Sharma et al., 1978b). Coincidence of the molecular weight values obtained with native and SDS-treated proteins as above indicates a monomeric structure for the native form of the protein. Its con- centration in bovine brain tissue was 0.2mg (0.003 #mol) per kg of wet wt. (Sharma et al., 1979), which is about 1/50 the concentration of calcineurin on the weight basis (Table 2). Its physiological signifi- cance is unknown.

3. A 50 K CALMODULIN-BINDING PROTEIN

A calmodulin-binding protein of an estimated mol- ecular weight of 50,000 has been demonstrated in a high speed supernatant fraction of rat brain hom- ogenates and purified from it to reasonable purity by a combination of column chromatographic steps by Ma6kawa and Abe (1980). Binding of caimodulin

to this protein is Ca2'-dependent and is resistant to the presence of 6 M urea. This protein accounted for 0.4'!0 of total soluble protein by densitometric scan- ning of the fresh brain extract (Maekawa and Abe, 19801. Assuming the amount of the total soluble pro- tein to be about 40 rag/g, the above figure will give about 140mg (2.8/*mol) of 50k protein per kg of brain (Table 2). The concentration of this protein de- creased rapidly atier death and only a trace amount of this protein was detectable in a homogenate pre- pared from a brain kept at 37'C for 20rain after death. The function of this protein is yet to be eluci- dated.

4. MYELIN BASIC PROTEIN AS A CALMODULIN-BINDING PROTEIN

Grand and Perry (1979) purified a calmodulin- binding protein of molecular weight 22,000 from bovine brain. It interacted with calmodulin as well as skeletal muscle troponin C in a Ca 2 +-dependent man- ner and inhibited the activation of the brain phospho- diesterase activity by calmodulin. This protein was then identified as myelin basic protein (Grand and Perry, 1979, 1980). Although in vitro studies indicate the formation of a 1 : 1 complex between calmodulin and myelin basic protein, calmodulin may not be bound to myelin basic protein in vivo mainly because both proteins are separated from each other in the cell. With regard to the Cae+-dependency of the interaction of calmodulin and myelin basic protein, controversial results were obtained in different labor- atories. Itano, Itano and Penniston (1980) reported that the formation of a complex between both pro- teins can be explained by the nonspecific and Ca 2 +- independent interaction of the basic protein with a strongly acidic protein (calmodulin). On the other hand, according to Grand and Perry (1979) and Iwasa, Iwasa, Matsui, Higashi and Miyamoto (1981), binding of calmodulin to myelin basic protein occurred in a Ca e +-dependent fashion and may have a biological meaning.

5. ASSOCIATION OF CALMODULIN WITH COATED VESICLES

Coated vesicles, first described by Gray (1961), are involved in receptor-mediated endocytotic processes (Goldstein, Anderson and Brown, 1979). The major constituent of these vesicles is a 180,000-dalton pro- tein called clathrin (Pearse, 1975).

Linden, Dedman, Chafouleas, Means and Roth (1981) observed that calmodulin associates with

Calmodulin-binding proteins in brain 163

coated vesicles in the presence of Ca 2 +. Firstly, cal- modulin was contained in coated vesicle preparations purified in the presence of Ca 2 ÷ at a concentration 7 times more than that of the samples obtained in the absence of Ca 2 +. Secondly, 125I-calmodulin bound to coated vesicles in a Ca 2 +-dependent manner. How- ever, which of the coated vesicle components is re- sponsible for the calmodulin binding has not been made sufficiently clear. Chromatographic fraction- ation of 2 M urea-solubilized coated vesicles on a cal- modulin-Sepharose column revealed several calmo- du:lin-interacting proteins (including clathrin) that were retained on the column in the presence of Ca 2 + and eluted from the column in the absence of Ca 2 ÷. However, these proteins were also found in the flow through fraction from the column and no conclusion was drawn from the results as to which of the solubil- ized proteins are specifically responsible for the ob- served calmodulin binding to the coated vesicles (Lin- den et al., 1981).

6. CALMODULIN-BINDING PROTEINS OF BRAIN MICROTUBULES

Although cytoplasmic microtubules are mainly composed of tubulin (about 85~o), they also contain a number of other proteins (about 15~o) as 'micro- tubule-associated proteins' (MAPs). These non- tubulin accessory proteins are present in microtubules in roughly constant stoichiometry with tubulin through multiple cycles of assembly-disassembly. A possibility that these MAPs are involved in the regu- lation of microtubule assembly in vivo has deserved attention recently because when the MAPs are separ- ated from the tubulin by phosphocellulose column chromatography, the tubulin no longer assembles into microtubules and addition of the MAPs back to the tubulin fully restored the capacity of tubulin to as- semble (for a review see Timasheff and Grisham, 1980). Two MAP species have been characterized recently. These are identified on SDS-polyacrylamide gel electrophoresis as two major high molecular weight (~300,000) bands (HMW-MAPs) and a family of four closely related low molecular weight (55,000-62,000) bands. The latter polypeptides (M r = 55,000-62,000) were collectively termed tau (z) factor by Weingarten, Lockwood, Hwo and Kirschner (1975).

There is now much evidence that the concentration of Ca 2 + regulates the assembly-disassembly of micro- tubules in the cell: increase and decrease in the Ca 2 ÷ concentration causes disassembly and assembly, re- spectively (see Timasheff and Grisham, 1980). There-

fore, an interesting possibility is the involvement of calmodulin in the regulation of microtubule assembly as a mediator of the action of Ca 2 +. Indeed, Marcum, Dedman, Brinkley and Means (1978) and Nishida, Kumagai, Ohtsuki and Sakai (1979) independently observed that calmodulin both inhibits and reverses microtubule assembly in the presence of Ca 2 ÷. More- over, the immunofluorescence localization of calmo- dulin in the dividing cell revealed that it is associated with the chromosome-to-pole region of the mitotic apparatus, where the disassembly of microtubules takes place (Welsh, Dedman, Brinkley and Means, 1978, 1979). In subsequent studies using microtubules purified from brain, Kumagai and Nishida (1979, 1980) demonstrated the Ca 2 +-dependent complex for- mation between calmodulin and tubulin on 'Sephadex G-200 gel filtration column chromatography and high speed liquid column chromatography. The complex thus formed was incapable of polymerizing into microtubules. Quantitative analysis of the binding revealed that 2 mol of calmodulin can bind to 1 mol of tubulin with a dissociation constant of about 4/~M (Kumagai, Nishida and Sakai, 1982).

In an independent study, Sobue, Fujita, Muramoto and Kakiuchi (1981a) found that the tau (z) factor, one species of the microtubule associated proteins (see above), is a calmodulin-binding protein undergoing a Ca 2 +-dependent association with calmodulin. Subse- quently, they reconstituted with purified tubulin, r factor and calmodulin a Ca2+-sensitive microtubule assembly system (Kakiuchi and Sobue, 1981). In this system, raising the Ca 2 + concentration produced the inhibition of assembly by forming a calmodulin-z factor complex. Thus, r protein acted as a flip-flop switch forming complexes between calmodulin at the increased Ca 2+ level and tubulin at the decreased Ca 2÷ level thereby leading to disassembly and as- sembly, respectively, of microtubules (Fig. 1). It is unclear at present whether calmodulin, in the pres- ence of Ca 2 +, interacts in vivo with both tubulin and T factor or either of the two proteins for the micro- tubule disassembly. This point should be clarified in the future study.

Fig. 1. Ca2+-dependent and ~ factor-linked flip-flop regu- lation of microtubule assembly (Kakiuchi and Sobue,

1981). CaM, calmodulin; T, tubulin.

1h4 ~IIIRI) KAKIt i ttl

7. A SPECTRIN-LIK[. ('ALMODULIN-BINDING PROTEIN

Brain contains about 400 mg of EGTA-extractable calmodulin per kg of wet wt. (Kakiuchi t,t ,l.. 1982c: Teshima and Kakiuchi, 1978). Studying the subcellu- lar distribution of calmodulin activily in the brain tissue, Teshima and Kakiuchi {1978) found that con- siderable amounts of calmodulin (about 100 mg per kg of brain) coprecipitates with particulate fractions when the fractionation procedure of a homogenate is carried out in the presence of Ca" +. The result indi- cates the existence of particle-associated calmodulin- binding protein(s) in brain in the cell the highest binding activity was found to be associated with a microsomal fraction containing disrupted synapto- some structures (Teshima and Kakiuchi, 1978). In a subsequent study using 3H-calmodulin. this particle- associated calmodulin-binding activit S was found in all rat tissues examined with the highest concen- tration in brain followed by adrenal gland (Sobue, Muramoto. Yamazaki and Kakiuchi. 1979), (Fig. 2).

Subsequently, Kakiuchi, Sobue and Fujita (1981a) solubilized this protein from a membrane fraction of bovine brain with 6M urea and then purified to homogeneity by a combination of column chroma- tographic steps u s i n g DEAE-cellulose, calmodulin- Sepharose and Sepharose 4B. The purified protein had a molecular weight of 240,000 upon SDS-polyac- rylamide gel electrophoresis, which is identical to that of :~ subunit of spectrin, an erythrocyte protein involved in the cytoskeleton structure underlying the

plasma membrane Ifor a review, Marches), 1979). Spectrin was shown to be a Ca2 +-dependent calmo- dulm-bmding protein by the same research gruup earlier (Sobue, Fujita, Muramoto and Kakiuchi, 1980; Sobue, Muramotu, Fujita and Kakiuchi. 1981at.

Therefore, Kakiuchi et al. t l981a) suggested that the brain 240,000 dalton protein is closely related to spectrin. This view was supported by its specilic lo- cation on membranes, solubility in 6 M urea, insolu- bility in Triton X-t00, estimated molecular weight on SDS-polyacrylamide gel and calmodulin-binding ability. Then it was puzzled that tile purified protein appeared to be a monomeric form on SDS-polyacry- lamide gel electrophoresis because spectrin is known to consist of two species of polypeptide, M, = 240,000 (c~ subunit) and 220,000 (fl subunit) (for reviews, Steck, 1974: Marches), 1979). However, a later report by these workers showed a dimeric structure composed of 24(1,000 (~ subunit) and 235,000 (fl subunit) dalton polypeptides on SDS-polyacrylamide gel for this pro- tein (Kakiuchi, Sobue, Morimoto and Kanda, t982b; Kakiuchi, Sobue, Kanda, Morimoto, Tsukita, Fsuk- ita, lshikawa and Kurokawa, 1982d). It was found that the purification procedure used in their earlier study (Kakiuchi e t al., 1981al separated the :~ subunit fiom its fl counterpart: the [* subunit was retained oil and not eluted from the DEAE-cellulose column. The p,otein was termed calspectin for the calmodulm- binding spectrin-like properties (Kakiuchi et al.. 1982b).

The two subunit polypeptides of calspectin are dis- tinct from filamin or myosin (Fig. 3) IKakiuchi e t al.,

1982d). Although its ~ subunit band coincided with that of spectrin, its fl subtinit band did not (Fig. 3t. The calmodulin binding ability of the calspectin

w z w

i ~ v

a b c d

Fig. 2. Distribution of particle-associated calmodulin- binding activity in rat tissues (Sobue et al., 1979). l mM EGTA was present in the medium throughout the hom- ogenate making and subsequent centrifugations. 105,0000 pellets, obtained either in the presence (Ill or absence (11)

of 0.6 M KC1, were assayed for 3H-calmodulin binding.

Fig. 3. SDS-polyacrylamide gel electrophoresis (in 5'~o gelsl of purified proteins (Kakiuchi et al., 1982d). (a) 240 K cal- modulin-binding protein prepared by the procedure of K akiuchi et al. (1981 a); (b) calspectin prepared as described by Kakiuchi et al. (1982b, d); (c1 spectrin of human erythro- cytes; (d) filamin (upper band) and myosin heavy chain

from chicken gizzard smooth muscle.

Calmodulin-binding proteins in brain 165

appears to be attributed to its ~t subunit because the isolated ~ subunit shows Ca 2 ÷-dependent calmodulin binding (Kakiuchi et al., 1981a), while the fl subunit failed to bind labeled calmodulin by a gel overlay method (Kakiuchi et al., 1982d).

The calspectin molecule undergoes, like spectrin (Ralston, Dunbar and White, 1977; Ungewickell and Gratzer, 1978; Liu and Palek, 1980), dimer-tetramer interconversion (Kakiuchi et al., 1982d). Although both forms of protein associated with F-actin, only the tetramer was capable of cross-linking actin fila- ments to produce a viscous gel, indicating that the dimeric form has a single binding site for F-actin and the tetrameric form has two F-actin-binding sites (Kakiuchi et al., 1982d). The morphological appear- ance of calspectin visualized by low angle rotary-sha- dowing technique (Tyler and Branton, 1980) is quite similar to that of spectrin, i.e. the tetrameric form of calspectin, which is about twice as long as the dimeric form, appeared as a flexible rod-like shape consisting of two parallel strands attached side to side (Kakiuchi et al., 1982d). This shape is most simply explained as two head-to-head associated heterodimers (for a sche- matic representation, see Fig. 4). Calspectin's actin binding sites are at the tail ends of the dimers. In mammalian erythrocytes, spectrin, presumably in its tetrameric form together with actin oligomers, forms a Triton-insoluble cytoskeletal meshwork that is lining the cytoplasmic surface of the cell membrane (for a review, Branton, Cohen and Tyler, 1981). The same may be proposed for calspectin in the neuronal cell.

Several independent lines of research also led to the discovery of this protein. In one of such lines, Willard and his associates have studied intraaxonal transport of neuronal proteins by injecting intraocularly (rab- bit), a radioactive amino acid. They observed on SDS- polyacrylamide gel electrophoresis about 40 different species of labeled polypeptides that are synthesized in the retinal ganglion cells and then transported down the axones (Willard, Cowan and Vagelos, 1974). They

Fig. 4. Molecular structure of calspectin tetramer, a sche- matic representation (Kakiuchi et al., 1982d).

classified these polypeptides into five different groups based on their transport velocities. Among these uncharacterized polypeptides there was a pair of high molecular weight (M r = 250,000 and 240,000) poly- peptides that are transported at a maximum time- averaged velocity of 40mm/day. In a subsequent study, Levine and Willard (1981) prepared a specific antibody against these polypeptides and used the antibody to localize them in neurons and other cells by means of indirect immunofluorescence. The result showed that these polypeptides are highly concen- trated in the cortical cytoplasm of neurons and also nonneuronal cells that include skeletal muscle, uterus, intestinal epithelium and cultured cells. As their dis- position resembles a lining of the cell, Levine and Willard (1981) proposed a name fodrin (from Greek fodros) for the protein. More recently, Glenney, Jr., Glenney, Osborn and Weber (1982) purified this pro- tein from chicken brain to homogeneity. They also purified a similar protein with M r of 260,000 and 240,000 (TW 260/240) from the brush border of chicken intestinal epithelial cells. These proteins were characterized as Ca 2+-dependent calmodulin-binding proteins and also F-actin-binding proteins. They showed a spectrin-like appearance of these proteins by a low angle rotary-shadowing technique.

Besides above two research lines conducted by Japanese and American (and European) workers, the same protein has been obtained by two other research groups about the s~me time. Davies and Klee (1981) partially purified from a supernatant fluid of brain homogenates a CaZ+-dependent calmodulin-binding protein of doublet form (M r = 235,000 and 230,000) capable of binding to F-actin. Independently, Shimo- oka and Watanabe (1981) partially purified a F-actin- binding doublet protein (M r = 240,000 and 235,000) capable of activating actomyosin ATPase activity.

Spectrin, together with actin, is a major constituent of the cytoskeletal net work underlying the inner sur- face of the cell membrane (for a review, Branton et al., 1981). It has been thought that this type of protein is specific in erythrocytes because attempts to detect spectrin or spectrin-like proteins by immunological techniques in different cell types were without success (Painter, Sheetz and Singer, 1975; Hiller and Weber, 1977). However, the results just reviewed above now provide new evidence that proteins which share com- mon features with erythrocyte spectrin exist in many cells including neuronal cells. These features include (1) heterodimeric structures with subunit molecular weight near 240 K on SDS-polyacrylamide gel elec- trophoresis [240 and 235 K for brain calspectin (fod- rin) (Glenney, Jr. e t al., 1982; Kakiuchi et al.,

166 SmRO KAKI~ ~'H]

1982b, d), 260 and 240 K for intestinal epithelial cell brush border protein (TW 260/240) (Glenney, Jr. et

al., 19821 and 240 and 220 K for erythrocyte spectrin (Steck, 1974; Marchesi, 1979)], t2) elongate double- stranded molecular shapes by rotary-shadowing tech- nique (for calspectin, Glenney, Jr. et al.. 1982; Kakiuchi et al., 1982d); for TW 260/240 protein, Glenney, Jr. et al., 1982; for spectrin, Shotton. Burke and Branton, 19791, {3) dimer tetramer interconver- sion (for calspectin, Kakiuchi et al., 1982d: for spec- trin, Ralston et al., 1977; Ungewickell and Gratzer, 1978; Liu and Palek, 19801, (4) ability to bind to actin filament (for calspectin, Levine and Willard, 1981; Shimo-oka and Watanabe, 1981; Davies and Klee, 1981; Glenney, Jr. et al., 1982; Kakiuchi et al.,

1982b, d; for TW 260/240 protein, Glenney, Jr. et al.,

1982; for spectrin, see reviews Marchesi. 1979; Bran- ton et al., 1981), (5) ability to bind to calmodulin in a CaZ+-dependent fashion (for calspectin, Kakiuchi et

al., 1981a, 1982b, d; Davies and Klee, 1981: Glenney, Jr. et al., 1982; for TW 260/240 protein, Glenney, Jr. et al., 1982; for spectrin, Sobue et al., 1980, 1981a).

As for the physiological significance of calspectin, it may perform analogous functions as spectrin does in erythrocytes. As the major constituent of the sub- membranous cytoskeleton, spectrin is thought to be responsible for the control of cell shape and cell deformability (Marchesi, 19791. In addition, there is a good possibility that spectrin is involved in the regu- lation of the lateral movement of membrane proteins and the distribution of surface markers (Nicolson, 1976). In this respect, a recent report by Levine, Skene and Willard (1981) is of particular interest. During the cross-linking agents-induced capping of surface pro- teins of lymphocytes, they observed the redistribution of lymphocyte fodrin (revealed by anti-fodrin im- munofluorescence) to the formation of a correspond- ing subcap on the inside of the plasma membrane. They have proposed that aggregated surface proteins could become attached to a mobile lining, of which fodrin is one constituent and be moved to one pole of the cell. If this model is a reality, then, as they further suggested, axonal transport of neuronal proteins may be explained by an analogous mechanism with such the mobile lining as a carrier of the transported pro- teins. Another interesting possibility is the involve- ment of brain calspectin (fodrin), together with other contractile proteins, in the regulation of neurotrans- mitter release. The fact that the concentration of this protein is especially high in a synaptic membrane fraction (Sobue et at., 1979; Kakiuchi et al., 1982b) supports this view. Further studies on the biological functions of calspectin (fodrin) are urgently needed.

8. A 155K PARTICLE-ASSOCIATED CALMODU LIN-BINDING PROTEIN

During the course of the purification of the spec- trin-tike calmodulin-binding protein from a micro- somal fraction of bovine brain homogenates, another calmodulin-binding protein was separated from the former protein and purified to homogeneity (Kakiu- chie t al., 1982a). The purified protein, with a mol- ecular weight of 155,0(X1 on SDS-polyacrylamide gel electrophoresis, was apparently different from 160,(X)0 dalton component of neurofilament. It is unlikely that this protein is a degradation product of calspectin due to proteolysis. This is because the 155 K protein to calspectin ratio of 0.6 (protein weight basis) calculated from the elution profile from the Sepharose gel column on which 6M urea-solubilized microsomal fraction was applied agreed with that obtained by densitometric scanning of the SDS-polyacrylamide gel on which a fresh sample of the microsomal fraction was electrophoresed. Moreover. this ratio was repro- ducible with different preparations. The 155 K protein is present in bovine brain m a concentration of 220 mg or 1.4/~mol per kg (Table 2). Its function is as yet unknown.

9. PARTICLE-BOUND AND EGTA-NONEXTRACTABLE FORM

OF CALMODULIN

There are two different pools of the particle-asso- ciated calmodulin. The EGTA-nonextractable form of calmodulin should be distinguished from the calmo- dulin which is bound to the calmodulin-binding pro- teins present in the particulate fraction, i.e. calspectin and 155K protein. The latter type calmodulin is easily released into the soluble fraction by the ex- posure of the particulate fraction to a medium con- taining EGTA. The activity of the EGTA-nonextrac- table form of calmodulin was latent to a certain extent and its solubilization required the presence of nonionic detergents (Kakiuchi e t al., 1982c; Sobue et al., 1981b).

Sobue e t al. (1981b) solubilized this form of calmo- dulin from a membrane fraction of bovine brain hom- ogenates with 2~ Triton XI00 and purified it by column chromatographies. The purified protein was identical to the soluble form of calmodulin in the two principal criteria (Kakiuchi, Sobue, Yamazaki, Nagao, Umeki, Nozawa, Yazawa and Yagi, 1981b), i.e. mobilities upon polyacrylamide gel electrophoresis in the presence and absence of Ca 2 + and the acti- vation of the calmodulin-deficient phosphodiesterase

Calmodulin-binding proteins in brain 167

activity in quantitative means. The mechanism through which calmodulin is tied to the particles is unclear. The quantity of this form of calmodulin in brain is about 100 #g/g tissues, which is about 1/4 that of the total EGTA-extractable calmodulin in the cell (Kakiuchi et al., 1982c; Sobue et al., 1981b).

In this article, calmodulin-binding proteins found in brain tissue is reviewed. It is expected that the regulating functions of calmodulin as a mediator of the actions of intracellular Ca 2 + will be elucidated to a large extent by the further study of the properties of these calmodulin-binding proteins.

Acknowledgements--This work was supported in part by Medical Research Grants from the Japan Medical Associ- ation and from Japan Research Foundation for Clinical Pharmacology. I wish to thank Tomoko Nagasaka and Junko Ohtani for typing the manuscript.

REFERENCES

Branton, D., Cohen, C. M. and Tyler, J. (1981). Interaction of cytoskeletal proteins on the human erythrocyte mem- brane. Cell 24, 24-32.

Cheung, W. Y. (1970). Cyclic 3',5'-nucleotide phosphodies- terase: Demonstration of an activator. Biochem. biophys. Res. Commun. 38, 533-538.

Cheung, W. Y. (1971). Cyclic 3',5'-nucleotide phosphodies- terase: Evidence for and properties of a protein acti- vator. J. biol. Chem. 246, 2859-2869.

Cheung, W. Y., Lynch, T. J., Wallace, R. W. and Tallant, E. A. (1981). cAMP renders Ca2÷-dependent phospho- diesterase refractory to inhibition by a calmodulin-bind- ing protein (calcineurin). J. biol. Chem. 256, 4439-4443.

Davies, P. J. A. and Klee, C. B. (1981). Calmodulin-binding proteins: A high molecular weight calmodulin-binding protein from bovine brain. Biochem. Int. 3, 203-212.

Glenney, Jr., J. R., Glenney P., Osborn M. and Weber, K. (1982). An F-actin- and calmodulin-binding protein from isolated intestinal brush borders has a morphology related to spectrin. Cell 28, 843-854.

Goldstein, J L., Anderson, R. G. W. and Brown, M. S. (1979). Coated pits, coated vesicles and receptor- mediated endocytosis. Nature, New Biol. 279, 679-685.

Grand, R. J. A. and Perry, S. V. (1979). Calmodulin-bind- ing proteins from brain and other tissues. Biochem. J. 183, 285-295.

Grand, R. J. A. and Perry, S. V. (1980). The binding of calmodulin to myelin basic protein and histon H2B. Bio- chem. J. 189, 227-240.

Gray, E. G. (1961). The granule cells, mossy synapses and purkinje spine synapses of the cerebellum: Light and electron microscope observations. J. Anat. 95, 345-356.

Hiller, G. and Weber, K. (1977). Spectrin is absent in various tissue culture cells. Nature, New Biol. 266, 181-183.

Itano, T., Itano, R. and Penniston, J. T. (1980). Interaction of basic polypeptides and proteins with calmodulin. Bio- chem. J. 189, 455-459.

Iwasa, Y., Iwasa, T., Matsui, K., Higashi, K. and Miya- moto, E. (1981). Interaction of calmodulin with chroma- tin associated proteins and myelin basic protein. Life Sci. 29, 1369-1377.

Kakiuchi, S. and Sobue, K. (1981). Ca 2+- and calmodulin- dependent flip-flop mechanism in the microtubule assembly-disassembly. FEBS Lett. 132, 141-143.

Kakiuchi, S. and Yamazaki, R. (1970). Stimulation of the activity of cyclic Y,5'-nucleotide phosphodiesterase by calcium ion. Proc. Japan Acad. 46, 387-392.

Kakiuchi, S., Yamazaki, R. and Nakajima, H. (1970). Properties of a heat-stable phosphodiesterase activating factor isolated from brain extract: Studies of cyclic 3',5'-nucleotide phosphodiesterase. II. Proc. Japan Acad. 46, 587-592.

Kakiuchi, S., Yamazaki, R., Teshima, Y. and Uenishi, K. (1973). Regulation of nucleotide cyclic 3',5'-monophos- phate phosphodiesterase activity from rat brain by a modulator and Ca 2÷. Proc. natn. Acad. Sci. U.S.A. 70, 3526-3530.

Kakiuchi, S., Sobue, K. and Fujita, M. (1981a). Purification of a 240,000-dalton calmodulin-binding protein from a microsomal fraction of brain. FEBS Lett. 132, 144-148.

Kakiuchi, S., Sobue, K., Yamazaki, R., Nagao, S., Umeki, S., Nozawa, Y., Yazawa, M. and Yagi, K. (1981b). Ca 2+- dependent modulator proteins from Tetrahymena pyri- formis, sea anemone, and scallop and guanylate cyclase activation. J. biol. Chem. 256, 19-22.

Kakiuchi, S., Sobue, K., Kanda, K. and Morimoto, K. (1982a). Purification of a 155000-Mr calmodulin-binding protein from a microsomal fraction of brain. FEBS Lett. 138, 173 177.

Kakiuchi, S., Sobue, K., Morimoto, K. and Kanada, K. (1982b). A spectrin-like calmodulin-binding protein (cals- pectin) of brain. Biochem. Int. in press.

Kakiuchi, S., Yasuda, S., Yamazaki, R., Teshima, Y., Kanda, K., Kakiuchi, R. and Sobue, K. (1982c). Quanti- tative determinations of calmodulin in the supernatant and particulate fractions of mammalian tissues, d. Bio- chem. 92, 1041-1048.

Kakiuchi, S., Sobue, K., Kanda, K., Morimoto, K., Tsuk- ita, S., Tsukita, S., Ishikawa, H. and Kurokawa, M. (1982d). Correlative biochemical and morphological studies of brain calspectin: A spectrin-like calmodulin- binding protein. Biomed. Res. 3, 400-410.

Klee, C. B. and Krinks, M. H. (1978). Purification of cyclic Y,5'-nucleotide phosphodiesterase inhibitory protein by affinity chromatography on activator protein coupled to Sepharose. Biochemistry 17, 120-126.

Klee, C. B., Crouch, T. H. and Krinks, M. H. (1979). Calci- neurin: A calcium- and calmodulin-binding protein of the nervous system. Proc. natn. Acad. Sci. U.S.A. 76, 6270-6273.

Klee, C. B., Crouch, T. H. and Richman, P. G. (1980). Calmodulin. Annu. Rev. Biochem. 49, 489-515.

Kumagai, H. and Nishida, E. (1979). The interaction between calcium-dependent regulator protein of cyclic nucleotide phosphodiesterase and microtubule protein. II. Association of calcium-dependent regulator protein with tubulin dimer. J. Biochem. 85, 1267-1274.

Kumagai, H. and Nishida, E. (1980). The interactions between calcium-dependent regulator protein (calmodU- lin) and microtubule proteins. Further studies on the mechanism of microtubule assembly inhibition by cal- modulin. Biomed. Res. 1,223-229.

IGN ~Il lRO K ~,K[I ~ [11

Kumagai, H., Nishida, E. and Sakai, It. (I 982i The interac- tion between calmodulin and microtubule p~oteins. IV Quantitative analysis of the binding between calmodulin and lubulin dimer. J. Biochem. 91, 1329 1336.

Levine J. and Willard M. (1981L Fodrin: Axonally trans- ported polypeptides associated with the internal periph- ery of many cells. J. cell Biol. 9t1, 631 643.

Levine J.. Skenc P. and Willard, M. (19811. GAPs and todrin, Novel axonally transported proteins. Trends Neurosci. 4, 273 277.

Lin. Y. M.. Liu. Y. P. and (:heung, W. Y. 11974). Cyclic Y,5'-nucleotide phosphodiesterase: Purification, charac- terization, and active torm of the protein activator from bovine brain. ,I. hioL Chem. 249, 4943 4954.

Lin, Y. M., Liu. Y. P. and Cheuug, W. Y 11975). Cyclic 3',5'-nucleotide phosphodiesterase. Ca 2 ~ ..dependent formation of bovine brain enzyme-activator complex. FEBS Lett. 49, 356 360

Linden, C D., Dedman, J, R., Chafouleas. J. G.. Means, A. R. and Roth, T. F. 11981). Interactions of calmodulin with coated vesicles from brain. Pro~. mun. Acad. Sci. U.S.A. 78, 31)8 312.

Liu S.-C. and Palek J. 11980). Spectrin tetramer-dimer equi- librium and the stability of erythrocyte membrane skel- etons, Nature, New Biol. 285, 586 588.

Maekawa, S. and Abe, T. (1980}. Isolation of a new calmo- dulin-binding protein from rat brain. Biochem. hiophys. Res. Commun. 97, 621 627.

Marchesi, V. T. (1979). Spectrin: Present status of a puta- tive cytoskeletal protein of the red cell membrane. J. Membrane Biol. 51, 101 131.

Marcum, J. M.. Dedman, J. R., Brinkley, B. R. and Means, A. R. (19781. Control of microtubule assembly-disassem- bly by calcium-dependent regulator protein. Proc. natn. Acad. Sci. U.S.A. 75, 3771 3775.

Nicolson, G. L. 11976). Transmembrane control of the receptors on normal and tumor cells. I. Cytoplasmic influence over cell surface components. Bioehim. Biophy~s. Aeta 457, 57 108.

Nishida, E., Kumagai, H., Ohtsuki, I. and Sakai, H. (1979). The interaction between calcium-dependent regulator protein of cyclic nucleotide phosphodiesterase and microtubule proteins. 1. Effect of calcium-dependent regulator protein on the calcium sensitivity of micro- tubule assembly. J. Biochem. 85, 1257--1266.

Painter, R. G., Sheetz, M. and Singer, S. J. 11975). Detec- tion and ultrastructural localization of human smooth muscle myosin-like molecules in human non-muscle cells by specific antibodies. Proc. hath. Acad. Sci. U.S.A. 72, 1359 1363.

Pearse, B. M. F. 11975). Coated vesicles from pig brain: Purification and biochemical characterization. J. molec. Biol. 97, 93 98.

Ralston, G., Dunbar, J. and White, M. (19771. The temper- ature-dependent dissociation of spectrin. Biochim. hio- phys. Acta 491, 345-348.

Sharma, R. K., Wirch, E. and Wang, J. H. (1978a). Inhi- bition of Ca 2 +-activated cyclic nucleotide phosphodies- terase reaction by a heat-stable inhibitor protein from bovine brain. J. biol. Chem. 253, 3575--3580.

Sharma, R. K., Desai, R., Thompson, T. R. and Wang, J. H. (1978bL Purification of the heat-stable inhibitor protein of the Ca 2 +-activated cyclic nucleotide phosphodiester- ase by affinity chromatography. Can. J. Biochem. 56, 598 604.

Sharma. R., l)esai. R+. Walsman, 1). M. and Wang, .L H. (19791. Purification and subunit structure of bovinc braul modulator binding protein. ,I hiol . ('hem 254, 4276 4282.

Shimo-oka. T. and Watanabe. Y. 11981). Stimulation of actomyosin Mg2+-ATPase activity by a brain microtu- hule-associated protein ti-action, ttigh-motecular-wcight actin-binding protein is the stinmlaling factor. +1 Bio- chem. 90, 1297 13117

Shotton. D. M,, Burkc, B. E. and Branton, D, (19791 The molecular structure of human erythrocyte spectrin. Bio- physical and electron microscopic studies..I, molec. BioL 131,31/3 329.

Sobuc. K., Muramoto, Y., Yamazaki, R. and Kakiuctu. S. 11979). Distribution in rat tissues of modulator-binding protein of particulate nature: Studies with [SH]modula- tor protein. FEBS Lett. 105, 105 109.

Sobue, K., Fujita, M., Muramoto. Y. and Kakiuchi+ S, (1980)J Spectrin as a major modulator binding protein of erythrocyte cytoskeleton. Biochem. Int. 1,561-566.

Sobue, K.. Muramoto, Y., Fujita, M. and Kakiuchi, S. (1981aL Calmodulin-binding protein of erythrocyte cyto- skeleton. Biochem. biophvs, res. Commlol. 100, 1063 1070.

Sobuc, K.. Yamazaki, R., Yasuda, S. and Kakiuchi, S. (1981b1. Identity of the particulate form of calmodulin with soluble calmodulin. FEBS Lett. 129, 215 219.

Sobue. K., Fujita, M.. Muramoto, Y. and Kakiuchi. S. (I 981c). The calmodulin-binding protein in microtubules is TAU factor. FEBS Lett. 132, 137 140.

Steck. T. L. (19741. Organization of proteins in the human red blood cell membrane. J. cell Biol. 62, i 19.

Stewart, A. A., lngebritsen, T. S., Manalan, A., Klee. ('. B. and Cohen, P. 11982). Discovery of a Ca 2~- and calmo- dulin-dependent protein phosphatase: Probable identity with calcineurin (CaM-BP8o). FEBS Lett. 137, 80-84.

Teo, T. S. and Wang, J. H. (1973L Mechanism of activation of a cyclic adenosine 3'.5'-monophosphate phosphodies- terase from bovine heart by calcium ions. J. biol, Chem. 248, 5950 5955.

Teshima, Y. and Kakiuchi, S. 11974). Mechanism of stimu- lation of Ca 2 ~ plus Mg z ~-dependent phosphodiesterase from rat cerebral cortex by the modulator protein and Ca-" +. Biochem. biophys. Res. Commun. ~ , 489-495.

Teshima, Y. and Kakiuchi, S. 11978). Membrane-bound forms of Ca'- +-dependent protein modulator: Ca 2"- dependent and independent bindings of modulator pro- tern to the particulate fraction from brain. J. Cyclic Nu- cleotide Res. 4, 219 231

Timasheff, S. N. and Grisham, L. M. 11980). hi i'itro assembly of cytoplasmic microtubules. Amlu. Ret. Bio- chem. 49, 565 591.

Tyler, J. M. and Branton, D. (19811). Rotary shadowing of extended molecules dried from glycerol. J. Ultrastruct. Reds. 71, 95 102.

Ungewickell, E. and Gratzer, W. 11978). Self-association ot human spectrin. A thermodynamic and kinetic study. Eur. J. Biochem. 88, 379 385.

Wallace, R. W., Lynch, T. J., Tallant, E. A. and Cheung, W. Y. 11978). An endogenous inhibitor protein of brain adenylate cyclase and cyclic nucleotide phosphodiester- ase. Arch. biochem. Biophys. 187, 328 334.

Wallace, R. W., Lynch. T. J., Tallant. E. A. and Cheung, W. Y. (1979). Purification and characterization of an in- hibitor protein of brain adenylate cyclase and cyclic nu- cleotide phosphodiesterase. J. biol. Chem. 254, 377 382.

Calmodulin-binding proteins in brain 169

Wang, J. H. and Desai, R. (1976). A brain protein and its effect on the Ca 2 ÷- and protein modulator-activated cyc- lic nucleotide phosphodiesterase. Biochem. biophys. Res. Commun. 72, 926-932.

Wang, J. H. and Desai, R. (1977). Modulator binding pro- tein: bovine brain protein exhibiting the Ca2+-depen - dent association with the protein modulator of cyclic nucleotide phosphodiesterase. J. biol. Chem. 252, 4175-4184.

Weingarten, M. D., Lockwood, A. H., Hwo, S.-Y. and Kirschner, M. W. (1975). A protein factor essential for microtubule assembly. Proc. natn. Acad. Sci. U.S.A. 72, 1858-1862.

Welsh, M. J., Dedman, J. R., Brinkley, B. R. and Means, A. R. (1978). Calcium-dependent regulator protein:

Localization in mitotic apparatus of eukaryotic cells. Proc. hath. Acad. Sci. U.S.A. 75, 1867-1871.

Welsh, M. J., Dedman, J. R., Brinkley, B. R. and Means, A. R. (1979). Tubulin and calmodulin: Effects of micro- tubule and micro-filament inhibitors on localization in the mitotic apparatus, d. cell Biol. 81,624-634.

Willard, M., Cowan, W. M. and Vagelos, P. R. (1974). The polypeptide composition of intra-axonally transported proteins: Evidence for four transport velocities. Proc. natn. Acad. Sci. U.S.A. 71, 2183-2187.

Wood, J. G., Wallace, R. W., Whitaker, J. N. and Cheung, W. Y. (1980). Immunocytochemical localization of cal- modulin and a heat-labile calmodulin-binding protein (CaM-BPs0) in basal ganglia of mouse brain. J. cell Biol. 84, 66-76.