122 Biochimica et Biophysica Acta, 957 (1988) 122-130 Elsevier

BBA 33244

The separation and characterization of two forms of Torpedo electric organ caldectrin

Ulrich Fritsche a, Annette von Kieckebusch a , , , Mart in Potschka b,

Victor P. Whit taker a and Veit Wi tzemann a

" Abteilung Neurochemie and h Abteilung Bioci~¢mie, Max.Planck-h:stitut f ~ biophysikalische Chemie, G~ttingen (£R. G.)

(Received 19 May 1988)

Key words: Calelectrin; Electric organ; (T. marmorata)

Two methods for extracting calelectrin, a CaZ+-regulated membrane-binding protein from the electric organ of Torpedo marmorata, have been compared and the more promising one was modified to increase the yield to 7-8 mg. kg- t wet weight of tissue, that is 4-5-times greater than the original method. The calelectrin so obtained could be resolved into a minor component (designated L-calelectrin) eluted from an anion-ex- change column at relatively low ionic strength (100 mM NaCi) and a major component (H-calelectrin) eluted at higher ionic strength (300 mM NAG). The two forms were also separated by chromatography on a hydrophobic resin. Electmphoresis on cellulose acetate indicated that L-calelectrin had a lower mean isoelectric point that the H-form and polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate showed that under reducing conditions (presence of 5%/3-mercaptoethanol) both forms migrated as single species, the L-form having a lower apparent relative molecular mass (Mr 32000) than the H-form (34000). Under non-reducing conditions, there was no change in the migration of L-calelectrin but the H-form was resolved into two components of Mr 34000 and 32000. The addition of 2 mM Ca z+ had no effect on the migration of either form. Both forms were equally recognized by an anti-calelectrin antiserum and were microheterogeneous with respect to their isoelectric points (pH 4.3-5.5) in two-dimensional gel electrophoresis. Physical measurements were carried out on the major H-form. The Stokes radius was estimated to be 3 nm, c¢, responding to an apparent M r of 44000. It was unaffected by changes in ionic strength, pH or Ca z+ concentration. Analytical ultracentrifngation gave a sedimentation constant of 2.9 S and an apparent Mr of 36000. Measurements of circular dichroism indicated that 78% of the molecule was in the a-belix configuration and 22% in random coil. Ca 2+ had no significant effect on the conformation.

* Present address: Labor fOr Biochemie II, Eidgenbssische Technische Hochschule, ZUrich, Switzerland.

Abbreviations: Bistris, 2-[bis(2-hydroxyethyl)amino]-2-(hy- droxymethyi)propane-l,3.diol; EGTA, ethylene glycol bis(/3- aminoethyl ether)-N,N,N',N'-tetraacetic acid; FPLC, fast-per- formance liquid chromatography; SDS-PAGE, sodium dodecyl sulphate-polyacrylamide gel electrophoresis; TBS, Tris-buffered saline.

Correspondence: V.P. Whittaker, Arbeitsgruppe Neurochemie, Max-Planck-lnstitut f'~ biophysikalische Chemie, Postfach 2841, D-3400 Gbttingen, F.R.G.

Introduction

The electric organ of Torpedo marmorata has been extensively used as a model system for cholinergic neurotransmission and has facilitated the purification and characterization of synapse- associated components. Ca 2+ is essential for syn- aptic transmission and therefore attempts have been made to identify proteins in this specialized tissue that interact selectively with this ion. This resulted in the discovery [1] of a novel protein of

0167-4838/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

this type, subsequently named calelectrin [2], that did not conform in its properties to any of the then known Ca2+-regulated proteins.

Calelectrin is a protein of Mr 34000 that binds to membranes in the presence of Ca 2+. Im- munochemically and biochemically similar pro- teins have been identified in and isolated from cells, membranes and organelles of diverse origin, including chromaffin granules [3-5], lymphocytes [6,7], other types of leucocytes [7], intestinal epi- thelial cells [8], smooth muscle membranes [9,10], liver [11,12], adrenal medulla [3,11,12], brain [11,12], kidney, intestine, pancreas [11] and mammary epithelial cells [13]. A structural re- lationship between these proteins was first sug- gested when an antiserum to the Torpedo electric organ calelectrin was found to cross-react with several of them [11].

Although the precise function and cellular lo- calization of these proteins (for which the name annexin has been proposed [14]) remains to be elucidated, they are now generally agreed to repre- sent members of a new Ca 2+- and phospholipid- dependent group which is distinct from the calmodulin and calmodulin-related Ca2+-binding proteins [14,15]. Their ability to interact with Ca 2+ and phospholipids may be linked in some cases to a specific interaction with membrane-bound en- zymes or substrates for such enzymes. Two mem- bers of the class, proteins p35 and p36, serve as substrates for retrovirally coded protein kinases [8,16,17], suggesting that they are involved in some way in intracellular chemical signalling.

Torpedo electric organ calelectrin was first de- tected and partially purified by extracting corn- minuted electric organ membranes with EGTA following extensive washing of the membranes with Ca2+-containing solutions [1]. Later, it was prepared by releasing it from membranes by an initial extraction with EGTA, and purified by a cycle of Ca2+-induced resorption to membranes and EGTA-induced release [18]. The biochemical and biophysical characterization of the protein and the raising of monospecific antibodies to it require its complete purification. In this paper we describe an improved purification procedure and some characteristics of the purified calelectrin. The work forms part of two recent theses [19,20].

123

Materials and Methods

Extraction of calelectrin from electric organ Tissue. T. marmorata, supplied by the Station

Biologique d'Arcachon, France, were kept in tanks of circulating artificial sea water at 15-18°C until required. The electric organs were removed from fish anaesthetized by immersion in sea water con- taining 0.05% Tricaine methanesulphonate and were stored frozen in liquid nitrogen. The extrac- tion procedures for calelectrin were based on Refs. 1 and 18.

Extraction method I. Frozen electric organs were crushed to a fine powder in the presence of liquid nitrogen. Tris-buffered saline iso-osmotic with Torpedo body fluids (0.4 M NaCl, 10 mM Tris-HCl (pH 7.4); TBS) containing 10 mM CaC! 2 was added to the tissue thawed to 0°C in a ratio of 2 : 1 (v/w). The resultant slurry was homogenized in a Waring blender for three 15-s periods at low speed and two 15-s periods at high speed. The homogenate was passed through four layers of cheese cloth and centrifuged at 2000 × g for 15 rain. The resulting supernatant was again centri- fuged at 130000×g for 30 min. The pelleted membrane fraction was resuspended using a glass-Teflon homogenizer in TBS contairdng I mM CaCI 2 and pelleted again at 130000 × g for 30 min. The membranes were washed a further three times using the same buffer and centrifugation conditions before being extracted for 15-30 min with TBS containing 10 mM EGTA. The solubi- lized calelectrin was recovered in the supernatant obtained after removal of extracted membranes by centrifugation at 130000 x g for 60 min and was dialysed twice against 5 1 of 10 mM Tris-HCl (pH 7.4). After recentrifugation at the same speed to remove a small precipitate, the solubilized cale- lectrin was dialysed against 5 1 of 50 mM im- idazole (pH 6.0) containing 5 mM fl-mercapto- ethanol.

Extraction method II. The crushed electric organ powder was suspended in TBS containing 1 mM EGTA instead of CaCI 2 to release all membrane- bound calelectrin. The supernatant remaining after centrifugation at 2000 x g for 15 min contained membrane fragments to which calelectrin bound when CaCl 2 was added to a final concentration of

124

10 mM. The membranes together with their bound calelectrin were sedimented at 100000 × g for 30 rain and washed three times with 0.05 M NaC1, 10 mM Tris-HC1 (pH 7.4) containing 3 mM CaCl2. The calelectrin was then resolubilized by resus- pending the membranes in TBS adjusted to pH 9.2-9.4 and containing 10 mM EDTA. Dialysis was performed as in method I, but the final dialy- sis was against 20 mM Bistris buffer (Sigma, De- isenhofen, F.R.G.) (pH 6.0) and 5 mM fl- mercaptoethanol.

Chromatography and electrophoresis Chromatography. This was performed with the

Pharmacia (Freiburg, F.R.G.) LCC 500 FPLC sys- tem using stationary phases of appropriate com- position. Anion-exchange chromatography was done on FPLC-MonoQ anion exchange resin. Calelectrin prepared by method I or II in the appropriate dialysis buffer was applied to columns previously equilibrated with the buffer and eluted in a 0-500 mM NaCI gradient generated by mix- ing, in continuously varying proportions, the ap- propriate dialysis buffer with the same buffer con- taining 1 M NaCI. Gel filtration was carried out on 10 x 300 mm FPLC-Superose 12 TM and 6 TM

columns with exclusion limits of 1000-300000 and 5000-5 000 000 respectively. The columns were calibrated with appropriate standards. Partition chromatography was carried out on FPLC-phen- yI-Sepharose. The column was equilibrated with 20 mM Bistris buffer (pH 7.0) containing 2 mM Ca 2+, 0.5 mM dithiothrettol and 2 M (NH4)2SO4. Calelectrin was applied in the same solution and eluted in a 2-0 M (NH4)2SO4 gradient obtained by mixing the solution with (NH4)2SOefree di- alysis buffer.

Gel electrophoresis. Cellulose acetate electro- phoresis was performed in 50 mM Tris-HCl (pH 7.0), 20 mM KCI and 0.1 mM EGTA at about 3-5 rnA. Calelectrin was rendered visible by staining with Amido-Schwarz or anti-calelectrin antibodies and horseradish peroxidase-conjugated second an- tibody. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) was carried out according to Laemmli [21]. Tw~-d~lensional electrophoresis was performed according to O'FarreU [221.

Immunochemistry Two polyclonal antisera and a monoclonal anti-

body were used to investigate the immunochemi- cal relationship between different forms of cale- lectrin. One polyclonal antiserum (I) was the one originally prepared by Dr J.H. Walker [1]; a sec- ond (II) was raised in a rabbit to purified native H-calelectrin by injecting 200 /tg of protein ini- tially and three further amounts of 200 #g at 3, 4 and 5 weeks thereafter. The monoclonal antibody (designated B25) to H-calelectrin was prepared by conventional methods. Immunostaining was per- formed on blots of cellulose acetate electrophero- grams.

Physical methods Analytical ultracentrifugation. This was per-

formed in a Beckman model E centrifuge, follow- ing standard procedures [23]. Sedimentation equi- librium experiments were done with 0.4~ (w/v) calelectrin in 40 mM potassium borate buffer (pH 7.0) containing 1 mM CaCI2 and 95 mM KCI to give an ionic strength of 100 mM.

Spectroscopic measurements. UV absorption was measured in a Hitachi 220 spectrophotometer. The circular dichroism of calelectrin was determined in 20 mM cacodylate buffer (pH 8.0) to which either 2 mM Ca 2+ or 0.1 mM EGTA had been added, using a Jobin Yvon Dichrograph Mark V in ttle range 31-200 nm. The resulting spectra were re- corded and analysed using the Jobin Yvon Dichro 7/80 program. Estimates of secondary structures were obtained using a program of Provencher and GlSckner [24].

Fluorescence measurements were performed in a MPF 4 Perkin-Elmer spectrofluorimeter. For all spectroscopic measurements calelectdn was sub- mitted to gel filtration to ascertain its purity, concentration and buffer concentration.

Protein concentrations. These were determined spectroscopically [25] using e2~Snm = 0.67 M -1. cm -1, or colorimetrically according to Lowry et al. [26] or Bradford [27].

Results

Isolation of two forms of calelectrin

Extraction from tissue To optimize the initial extraction procedure a

number of parameters were varied. Following the

125

original extraction procedure [1] (here referred to as method I) the frozen electric organ was ho- mogenized in a high ionic strength buffer in the presence of 10 mM Ca 2+, at pH 7.4. Coarse cell debris was removed by low-speed centrifugation and a membrane fraction containing calelectrin was obtained upon high-speed centrifugation. The resulting pellet was washed in the presence of 1 mM Ca 2+. Calelectrin bound to the pelleted mem- branes was solubilized by removing Ca 2+ with EGTA. Changing the pH of the homogenization buffer to 6.0 or 8.0 had little effect on the yield of calelectrin which was about 17 rag. kg -~ tissue (wet weight). Significantly higher yields were ob- tained by means of a later extraction procedure [18] which makes better use of the Ca2+-depen - dent binding of calelectrin to membranes. Cale- lectrin is solubilized during the first homogeniza- tion step by means of a buffer containing 1 mM EGTA. Call debris is removed and Ca 2+ is added to a final concentration of 10 mM to allow cale- lectrin to rebind to membranes. Sedimentation and washing of the resulting pellets in the pres- ence of 3 mM Ca 2+ yields a membrane-bound calelectrin fraction, from which calelectrin is

,.,,.,,fe. containing solubilized by suspending it in t...r . 10 mM EGTA. The yield of calelectrin is 2-3-times greater (45 nag. kg -1 wet weight of tissue) than that of method I. We have now further improved the yield (method II) by changing the ionic strength and pH during the isolation procedure on the basis of an analysis of the interaction of cale- lectrins with lipids [30]; the EGTA-solubilized calelectrin is rebound to membranes at low ionic strength and eluted at high ionic strength and pH 9.2-9.4. This increases the yield of calelectrin still further, to 70-80 mg. kg-~ wet weight of tissue or 4-5-times greater than the original method [1,28]. If the protein content of electric organ is assumed to be 10 g. kg-~, calelectrin must constitute over 0.8% of the total protein.

Purification by ion-exchange chromatography Calelectrin prepared by method I and chro-

matographed in imidazole buffer emerged as a single peak at 300 mM NaCI (Fig. la) and be- haved as a homogeneous polypeptide of the ex- pected M r in SDS-PAGE (Fig. lb). One major and several minor impurities were removed by this

0.8

0.6

< 0/.

0.2

0

o)

" I I I FRACTIONS

bl

I

.,=D 9/- [0.6 ~ 67

~' Q , ~ ,w -'-3 0/. ~ ,m. 30

z ~ 21

0.2 o 1/.

0 21

Fig. 1. (a) Separation of calelectrin, prepared by method I, by FPLC on a column of MonoQ anion-exchange resin. A NaCI gradient (0-1.0 M; broken line) in imidazole buffer (50 raM, pH 6.0) containing 0.5 mM EGTA and 5 mM ,8- mercaptoethanol was used for ehition. Calelectrin eluted as a sharp peak of UV absorbance at 280 nm (A28 o. continuous line) when the NaCI concentration (broken line) reached 300 mM (fractions 21 and 22). (b) Fractions 21 and 22, eluting at 300 mM NaCI, contain a polypeptide of Mr 34000 which migrates as a single band in SDS-PAGE in a 10% gel. The right-hand lane contains standard proteins of the Mr values

( × 10 -3) indicated.

procedure. However, when calelectrin prepared by method II was chromatographed in a Bistris buffer, the elution pattern was more complex (Fig. 2a) and in addition tu Ul~ mare peak eluting at ~0u mM NaC1, there was a smaller, but prominent peak at 100 mM together with other components, one of which was a shoulder on the high-ionic- strength side of the main peak at 300 mM NaCI. Only the prominent peaks at 300 and 100 mM NaC1 contained calelectrin as judged by their im- munochemical reactions and ability to bind to lipids in a Ca 2 +-dependent manner.

The two main peaks, at 300 and 100 mM NaCI respectively, when submitted to further analysis by means of SDS-PAGE, proved to contain poly- pept ides with appa ren t M r values of 32000-34000; both components also responded to an anti-calelectrin antiserum. It is concluded that calelectrin, at least as prepared by method II, exists in two forms, which we refer to as the high-salt (H) and low-salt (L) forms respectively. Estimates based on the areas under the FPLC peaks indicated that the H-form accounts for about 9070 of the total extractable electric organ cale- lectrin. This was confirmed by the relative inten- sities of the immunochemically stained bands cor- responding to the two forms of calelectrin after

126

electrophoretic sep~.ration on cellulose acetate of crude calelectrin pr~.pared by method I. Accord- ingly, this form v'as used for physical and bio- chemical characterization.

For characterization, H-calelectrin was freed from buffer salts by gel filtration on FPLC-Super- ose 12 TM. The resulting preparation contained no detectable contaminants in SDS-PAGE and had a defined, reproducible UV absorption spectrum (see below).

Separation of the two forms on phenyl-Sepharose Calelectrin extracts prepared according to

method II were adjusted by dia!ysis and additions to pH 7.0, 2 mM Ca 2 +, 0.5 mM dithiothreitol and 2 M (NH4)2SO 4 and applied to the column previ- ously equilibrated with the same buffer. Fractions containing protein washed through by the buffer

a}

b}

o.s

3.1 0

I t t I

2 h,,,o,..../cC ¢1

9t~ 67 - ~ 43 -,,-

3t, .--- 30 - "

21 "-~

i '

0.5 -- sr

0.3 o z

15 6 7 6 9 10, ,19 20 21 22 23 ,

FRACTION No.

Fig. 2. Separation similar to that of Fig. 1 except that cale- lectrin was prepared by method IL (a) Separation by FPLC resolved calelcctrin, as extracted, into three or four compo- nents; only the two largest of these, oluting at 100 and 300 mM NaC! respectively, had the immunochemical (Fig. 4) and Ca2+-r©gulated lipid-binding properties (Fig. 1, Ref. 30) char- acteristio of calolectrin. Analysis by SDS-PAGE (b, c) showed that both these peaks contained a single pcptide of apparent

Mr in the expected range for calelectrin of 32000-34000.

o} 0.5 "

, <

0 . . . . . . . ~'~

Fi . . . . . . . . . "~ 1 ° I i

, ~ ~ I I

e~

' 1.0

!

I

, 2 0

FRACTIONS

b)

O2

0,1 <

FRACTIONS

c}

;0 % ['" 11°

r 0,5 olI lo. + , o

FRACTIONS

Fig. 3. (a) Chromatography on phenyl-sepharose e~ a prepara- tion of calelectrin extracted from electric organ by method IL Continuous line, absorbanco at 280 nm (A2,o); broken line, (NH4)2SO 4 gradient. For other experimental details, see text. (b, c) Separation, on MonoQ anion exchange resin, of (b) protein not retained by phenyl-Sepharose, (c) protein removed as the concentrations of (NH4)2SO + and Ca 2+ in the eluent foil to zero. Note that the main protein peak in (b) emerges at a NaCI concentration of 0-100 mM allowing the accompanying calelectrin to be identified as the L-form, while that in (c) emerges at 300 mM and is identified as consisting of H-cale-

lectrin.

and those containing protein eluted as the (NH4)2SO 4 concentration fell towards zero (Fig. 3a) were separately pooled and submitted to an- ion-exchange chromatography on a MonoQ col- umn (Figs. 3b, c). Fractionation in this way showed that L-calelectrin had not been retained by the column, but that the H-form had been and was only recovered from the hydrophobic matrix when the (NH4)2SO4 concentration fell to zero. L-cale- lectrin forms only a small part of the initial, incompletely resolved peaks seen in Fig. 3b since it is accompanied by other components that also have little or no affinity for the column. Fig. 3a shows that there is little separation of the material retained by the column, elution occurring only when the (NH4)2SO 4 (and Ca 2+) concentrations fall to zero. Thus a selective fractionation due to

differences in Ca2+-dependent hydrophobicities of the bound proteins appears unlikely. However, the fact that H-calelectrin was retained by the column under the conditions under which it was applied, whereas L-calelectrin was not indicates that the overall charge and hydrophobicity of these two related peptides differ significantly.

Electrophoretic properties H- and L.calelectrin obtained by separating a

method-II extract on MonoQ were compared by electrophoresis on cellulose acetate (Fig. 4a, b), SDS-PAGE (Fig. 4c) and two-dimensional electro- phoresis (Fig. 4d). Electrophoresis on cellulose acetate (Fig. 4a) showed that L-calelectrin had a

lower mean isoelectric point than ferritin, i.e. less than 4.5, whereas the H-form had a higher one than ferritin, i.e. greater than 4.5. Both forms were equally immunoreactive to an anti-calelectrin anti- serum (Fig. 4b). In SDS-PAGE (Fig. 4c) under reducing conditions (presence of 5~ fl-mercapto- ethanol, lanes 1 and 5) both forms migrated as single species, the L-form having a lower apparent M r (32000) than the H-form (34000). Under non-reducing conditions (lanes 2, 4, 6, 8) there was no change in the migration of L-calelectrin (compare lane 6 with 5 and 8 with 7) but the H-form was resolved into two components of M r 34000 and 32000 (compare lane 2 with 1 and 4 with 3). The addition of Ca 2÷ (2 raM) had no effect on mobility under either condition.

In two-dimensional electrophoresis, both cale- lectrin forms showed microheterogeneity in re- spect of their isoelectric points which ranged from pH 4.3 to 5.5. The mean value for H-calelectrin was, as in cellulose acetate electrophoresis, higher than that for the L-form.

Physical parameters of H-calelectrin

The results of measurements of the Stokes radius, sedimentation coefficient relative to mole- cule mass, molar absorption coefficient and a- helical content of H-calelectdn are summarized in Figs. 5 and 6 and Tables I and II.

Fig. 5 shows determinations of Stokes radius by a comparison of the elution volume of H-cale- lectrin with those of standard proteins after gel filtration through FPLC-Superose 6 TM. It will be

127

Ferr i t in

o)

i 1 2 3

H - CuE " - ~

L - CoE - - ~

bl

/, 5

c~ 9~

67

~3

30

1 2 3 4 5 6 7 8 9 H-CoE L-CaE

dl

H-CoE L-CoE

Fig. 4. (a) Electrophoresis of L- (lane 2) and H- (lane 3) calelectrin prepared as shown in Fig. 2. The gel was calibrated with ferritin (lane 1), isoelectric point approx. 4.5. It will be seen that the two forms differ in isoelectric point. (b) in spite of this difference, the two forms both recognize an anti-H- calelectrin antiserum (II). Similar results were obtained with antiserum I (see Materials and Methods). (c) SDS-PAGE of the H- (lanes 1-4) and L- (lanes 5-8) forms of calelectrin, Equal amounts (15 #g) of protein were applied to the gel. after treatment with fl-mercaptoethanol (5~ v/v) and EGTA (2 raM) (lanes 1, 5), under non-reducing conditions (2 mM EGTA only) (lanes 2, 6), or in the presence of Ca 2+ (2 raM) under reducing (5~ v /v fl-mercaptoethanol, lanes 3, 7) or non-reduc- ing (lanes 4. 8) conditions. Lane 9 contained marker proteins with the M r values (× 10 -3) indicated. (d) Two-dimensional gel electrophoresis of H- and L- calelectrin indicating micro-

heterogeneity in isoelectric points.

seen that the Stokes radius is independent of the buffer used (borate-KF or cacodylate), the ionic strength (5-100 mM) and the pH (7 or 8) over the ranges tested. Similar results (not shown) were

128

TABLE i

SUMMARY OF HYDRODYNAMIC PROPERTIES OF H-CALELECTRIN

Method Parameter lneasured Value obtained Unit Mr estimated

Gel filtration Pokes radius 3.1 + 0.3 (13) nm 44000 Ultracentrifugation Sedimentation coefficient 2.9 S 36000 SDS-PAGE - - - 34000 Amino-acid analysis Amino-acid composition

of hydrolysate - - 32 000

obtained in experiments with Sepharose 12 TM and in which ionic strengths up to 700 mM and pH values down to 6 were tested. Ca :+ (2 raM) had no effect and did not produce self-aggregation. The Mr for a spherical protein of this Stokes radius is 44000 (Table I).

o)

10

5

3

2

0 Z~ G Thyroglobulin

\ ... g IgO

i -

""0 r~ Myoglobin

/ / I I I I , 0.4 O.6 0.8 1.0

REV Fig, 5, The Stokes radii (rstok,) Of four reference proteins plotted as ordinates against their relative elution volumes (REV) as abscissae when subjected to gel filtration through FPLC-Sepbarose 6 TM under varying conditions. Triangles and dotted line, cacodylate buffer (pH 7.0) of ionic strength 20 mM; squares and solid line, borate-KF buffer (pH 8.0), ionic strength 100 raM; circles and broken line, same buffer at ionic strength $ raM. The fiUed symbols indicate the relative elution volumes of H,calelectrin under the three conditions, projected onto the calibration lines. The point with the bar is the mean of six determinations at an ionic strength of 20 mM+ S.E. The mean of all values for the Stokes radius of caielectdn+ S.E, is 3,1+0,3 (13), Duplicate runs with and without 2 mM Ca 2+ showed no effect of this ion on the Stokes radius. The depar- ture of the points at 5 mM ionic strength from linearity indicates some degree of non-specific interaction with the column matrix at this low ionic strength. The magnitude of the

Stokes radius is not, however, affected.

04

03

< 02_ , , I < 0.1

0 I 240 250 260 270 280 290 300 310 320

300 320 340 360 380 / ,00 /,20

c) 10

5

0

~. -t0 ,~-15

-20

-25 , t I I I I 210 220 230 2/,0 250 260

Wovelength Into) 270

Fig. 6. (a) UV absorption, (b) fluorescence emission spectra of H-calelectdn. in (b), almost identical spectra were obtained in (1) the absence, (2, 3) the presence of Ca 2+ at (2) 5 mM, (3) 10 mM concentration. (c) Circular dichroism of H-calelectrin in cacodylate buffer (pH 8.0) of ionic strength 20 mM, in the presence of 0.1 mM EGTA (broken line) or 2 mM Ca 2+ (continuous line). Ordinates: molar ellipticity at 222 nm (0222) in deg.cm2.dmol - I x l 0 -3. In neither (b) nor (c) does the

presence of Ca 2+ have any significant effect.

TABLE I!

MOLAR ELLIPTICITY AND SECONDARY STRUCTURE OF H-CALELECTRIN

Measurements were made at pH 7.0 and ionic strength 20 mM.

Ca 2 + 0222 Structure (~) (mM) (deg.cm 2.dmol- t)

a-hefix B-fold random sheet coil

0 -19275 79 0 21 2.0 -19280 77 0 23

Velocity and equilibrium sedimentation in an analytical ultracentrifuge indicated a sedimenta- tion coefficient of 2.9 S and apparent Mr of 36000. Ca 2+ had no effect on the sedimentation properties of H-calelectrin and no evidence for Ca 2 +-induced self-aggregation was obtained.

Spectroscopic measurements are shown in Fig. 6. The UV absorption (Fig. 6a) shows a maximum at 278 mm; the molar absorption coefficient is 2.28- 104 M- ~- cm- ~ assuming a molecular weight of 34000. In Fig. 6b, the fluorescence emission spectrum of H-calelectrin is shown. Tyrosine and tryptophan residues are responsible for the emis- sion maximum at 303 nm and the additional sig- nals at 328 and 348 nm. Ca 2+ at 5 or 10 mM concentrations caused no significant change in fluorescence.

Circular dichroism in the far UV has been measured to determine whether Ca 2 + could induce conformational changes that could affect the bio- logical properties of calelectrin. The measure- ments (Fig. 6c and Table II) were made in cacodylate buffer, whose excellent optical proper- ties reduced the noise to signal ratio which is increasingly troublesome at wavelengths below 210 nm. However, similar results were obtained in Tris buffer, in which Ca2+-regulated binding to phos- phatidylserine has been demonstrated [30] despite the less favourable signal to noise ratio. In spite of these precautions, some variation (less than 10%) is observed between different preparations. Mea- surements with and without 2 mM Ca 2+ revealed no significant changes in the proportion of a-helix and random coil configurations in H-calelectrin. Calelectrin is characterized by a high content of a-helix and little, if any B-configuration (Table II).

129

Discussion

The improved isolation procedure described here in which the extraction procedure of Siidhof et al. [18] has been optimized and followed by FPLC anion-exchange chromatography has re- sulted in a substantial improvement in the yield of calelectrin and its resolution into two forms, a major, H-form, eluting from the anion exchange resin and binding to a hydrophobic resin at rela- tively high ionic strengths and a minor, L-form, eluting from the anion exchange resin at low ionic strength and showing little affinity for the hydro- phobic resin. It is estimated that 90~ of the cale- lectrin of the electric organ is in the H-form and that the two forms constitute about 1 ~ of the total protein of the organ.

The two forms of calelectrin show similar im- munoreactivity and behave similarly but not iden- tically in gel electrophoresis and isoelectric focus- ing. As previously reported for total calelectrin, though with slightly different M r values [1,28], the H-form was resolved into two bands of apparent M r 34000 and 32000 under non-reducing condi- tions but migrated as a single band of M r 34000 in the presence of 5~ fl-mercaptoethanol. L-cale- lectrin migrated as a single band of M r 32000 under both conditions. H-calelectrin had a lower isoelectric point than the L-form, but both showed charge microheterogeneity in gel isoelectric focus- ing. This may result from varying amounts of phosphorylation since other proteins of the cale- lectrin class are known to possess phosphorylation sites [29].

The UV absorption and fluorescence emission spectra of H-calelectdn were similar to those re- ported earlier [18] and were consistent with the presence of tryptophan and tyrosine residues in the polypeptide chain. A determination of the Stokes radius by gel filtration on Superose 6 TM

and 12 TM (not shown) gave a value of 3.1 nm which, for a typical globular protein of spherical shape implies a Mr of 44000. This is 22~ higher than the value obtained from measurements of the sedimentation coefficient and 23~ higher than that obtaine0 from SDS-PAGE, which were in good agreement with each other and with a value derived from amino-acid analysis [20~30]. Values fairly close to ours (Stokes radius 2.9 nm, sedi-

130

mentation coefficient 3.5 S) have recently been obtained [28] for calelectrin prepared according to the original method [1]. Measurements of circular dichroism revealed that a considerable portion of the molecule is in a random-coil configuration which would tend to increase hydrodynamic drag and might go some way to accounting for the discrepancies between the different estimates of the M r of H-calelectrin.

An important, if negative finding is the lack of effect of Ca 2+ o n any of the physical properties of calelectrin examined. The small effect (Table II) on the secondary structure is not considered to be significant. This is in marked contrast to the be- haviour of calmodulin and other proteins of the 'EF-hand' group, and clearly distinguishes it from the latter. We were unable to confirm the previ- ously reported self-aggregation of calelectrin [2,4] with our purified preparation; however, our find- ings are consistent with the previously reported lack of any effect of Ca 2+ o n the ultraviolet and fluorescence spectra of calelectrin and failure to detect Ca2+-induced conformational changes with hydrophobic fluorescent probes [18].

Judged by the change in electrophoretic mobil- ity induced by reducing conditions, the conforma- tion of native H-calelectrin may be influenced by the redox state of internal disulphide bridges.

The function of calelectrin in the electric organ is not known thought immunocytochemical evi- dence [31] points to a role in stabilizing fila- mentous structures in the electrocyte and the elec- tromotor axons in their relation to membranes. It is interesting to speculate that H- and L-cale- lectrin may have cell-specific locations in the tis- sue, the more abundant H-form being associated with the electrocytes and the less abundant in the presynaptic electromotor nerve terminals or asso- ciated glial processes.

Acknowledgements

U.F. and A.v.K. were supported by stipends from the Max-Planck-Gesellschaft.

References

1 Walker, J.H. (1982) J. Neurochem. 39, 815-823. 2 Siidhof, T.C., Walker, J.H. and Ohrocki, J. (1982) EMBO J.

l, 1167-1170.

3 Geisow, M.J. and Burgoyne, R.D. (1982) J. Neurochem. 38, 1735-1741.

4 Walker, J.H., Obrocld, J. and SUdhof, T.C. (1983) J. Neuro- chem. 41, 139-145.

5 Creutz, C.E., Dowling, L.G., Sando, J.J., Villa-Palasi, C., Whipple, J.H. and Zaks, W.J. (1983) J. Biol. Chem. 258, 14664-14674.

6 Owens, RJ. and Crumpton, M.J. (1984) Biochem. J. 219, 309-316.

7 Siidhof, T.C., Zimmermann, C.W. and Walker, J.H. (1983) Eur. J. Cell. Biol. 30, 214-218.

8 Gerke, V. and Weber, K. (1984) EMBO J. 3, 227-233. 9 Raeymakers, L., Wuytack, F. and Castells, R. (1985) Bio-

chem. Biophys. Res. Commun. 132, 526-532. 10 Moore, P.M. and Dedman, J.R. (1982) J. Biol. Chem. 257,

9663-9667. 11 Sfidhof, T.C., Ebbecke, M., Walker, J.H., Fritsche, U. and

Boustead, C. (1984) Biochemistry 23, 1103-109. 12 Geisow, M., Childs, J., Dash, B., Harris, A., Panayotou, G.,

Siidhof, T. and Walker, J.H. (1984) EMBO J. 3, 2969-2974. 13 Braslau, D.L., Ringo, D.L. and Rocha, V. (1984) Exp. Cell

Res. 155, 213-221. 14 Geisow, M.J. (1986) FEBS Lett. 203, 99-103. 15 Geisow, M.J. and Walker, J.H. (1986) Trends Biochem. SCi.

11, 420-422. 16 Fava, R.A. and Cohen, S. (1984) J. Biol. Chem. 259,

2636-2645. 17 Johnsson, N., Vanderkerckhove, J., Van Damme, J. and

Weber, K. (1986) FEBS Left. 198, 361-364. 18 Sildhof, T.C., Walker, J.H. and Fritsche, U. (1985) J. Neu-

rochem. 44, 1302-1307. 19 Fritsche, U. (1986) Reinigung und Charakterisierung yon

Calelectrin, einem Calcium- und Membran-bindenden Pro- tein, isoliert aus Torpedo marmoram. Dissertation, Marburg University.

20 Von Kieckebusch, A. (1986) Calcium-abhiingige Eigen- schaften yon Calelectrin aus dem elektrischen Organ yon Torpedo marmorata. Diplomarbeit, GSttingen University.

21 Laemmli, U.K. (1970) Nature 227, 680-685. 22 O'Farrell, P.H. (1975) J. Biol. Chem. 250, 4007-4021. 23 Schachman, M.K. (1959) Ultracentrifugation in Biochem-

istry, Academic Press, New York. 24 Provencher, S.W. and Gl~kner, J. (1981) Biochemistry 20,

33-37. 25 Scopes, R.K. (1974) Anal. Biochem. 59, 277-282. 26 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall,

R.J. (1951) J. Biol. Chem. 193, 265-275. 27 Bradford, M.M. (1976) Anal. Biochem. 72, 248-254. 28 Saito, T. and Miret, O. (1987) J. Neurochem. 48, 745-751. 29 Weber, K., Johnsson, N., Plessmann, U., Van, P.N., SSling,

H.-D., Ampe, C. and Vanderkerckhove, J. (1987) EMBO J. 6, 1599-1604.

30 Von Kieckebusch, A., Fritsche, U., Vogel, V., Witzemann, V. and Whittaker, V.P. (1988) Biochim. Biophys. Acta 957, 131-137.

31 Fiedler, W. and Walker, J.H. (1985) Eur. J. Cell. Biol. 38, 34-41.