Cs//Ca~dum(1880)11,
W-370
@LongmanGrcupUKLtdlWO
Ca*+ binding by Myxkola neurofilament proteins R.F. ABERCROMBIE, N.F. AL-BALDAWI and J. JACKSON Department of Physiology, Emory University School of Medicine, Atlanta, Georgia, USA Abstract - Titrimetric, ‘%a dialysis, and autoradiographic methods were used to examine how axopiasmic proteins from the giant neuron of the marine anneiid Myxicola infundibulum bind calcium. Following the autoradiographic method of Maruyama et al. [I], the 150-160 kD neurofliament subunits were identified as prominent intracellular Ca-binding peptides. Using equilibrium dialysis, extracts of axopiasmic proteins (MO% neurofiiament subunits) were examined in 300 mM KCI at different concentrations of free Ca and Mg, and at different pH. Axoviasmic proteins showed a high affinity Ca binding site (KIT 3-6 fl, capacity 3-7 mole g- protein) at pH 6.6 or pH 7.6. Changing the Mg concentration from 0 to 5 mM had no effect on the Ca binding. Elevating the dialysis pH from 7.0 to 9.0 reduced the apparent number of binding sites for Ca. Using microelectrodes to record the free Ca, microtitrations of axopiasmic proteins were completed by adding small amounts of CaCi2 to 100 u.l volumes of protein solutions. in a medium containing ionic constituents closely resembling those of the Myxicola axon, a Ca binding capacity of 5.0 pole g-’ protein and a Kin of -1 j.N were measured. The Myxicola giant axon has been described by Gilbert [23 as a nearly ideal source of neurofilaments since these filaments comprise a relatively large component of the total protein of the axon. Previous studies using Myxicolu have shown that neurofdaments are composed of two major proteins with apparent molecular mass close to 160 D and 150 kD [3-51. The ability of neurofdaments to bind Ca with high affinity has been described only recently [6,7]. Lefebvre and Mushynske [8] have presented data Abbreviations used TRIS, Tris(hydroxymethyl)amiminomethane; EDTA, Ethylenediaminetetraacetic acid; HEPES, N-[2-hydroxyethyll-lpiperazine-N’-[2-ethanesulfonic acid]; Kin, MichaelisMenten constant for half-saturation; NTA, Nitrilotriacetic acid, PAGE, Polyacrylamide gel electrophoresis; PIPES, Piperazine-N,N’-bis[2-ethanesulfonic acid]; SDS, Sodium dodecylsulfate
describing the calcium-binding properties of untreated and in vitro dephosphorylated neurofilaments from porcine spinal cord. Their binding studies were done at low ionic strength (20 mM HEPIS, 4 mM diethanolamine, pH 6.8). The b&iing studies of the present work have been done under physiological ionic strength using three different technical approaches and at different pH and Mg concentrations in or&r to better characterize this novel Ca binding site. Myxicola giant axons contain energy-dependent calcium binding systems involving mitochondria [9] and endoplasmic reticulum [lo]) as well as an unidentified energy-independent system [9, 111. The main observation of this report is that significant neurofilaments may be a energy-independent Ca buffer in the axon. The capacity attributed to neurofilament proteins in this study (-4 pm01 g-’ protein) corresponds to l-2 Ca 361
362
CELL CALCIUM
bound per 150-160 kD neurof~ament monomer. Based on the molar concentration of neurofilament protein in axoplasm, these sites could hold a maximum of 100-200 pmole Ca per kg of Myxicolu axoplasm at elevated Cai. At normal levels of Cai (-200 nM), these sites would hold one-tenth or less of this amount of calcium.
Materials and Methods Preparation Myxicola were collected in the Bay of Fundy by Mr Robert Bosien, Deer Island, New Brunswick Canada EOG-lR0, shipped by air to Atlanta, and maintained in a refrigerated (1O’C) seawater aquarium for up to eight weeks. Axons 700-1.000 pm in diameter were dissected, bathed for ten min in Ca-free artificial seawater, blotted, and cut open. Using forceps, the axoplasmic gel was extracted from 5-10 giant axons. This gel was solubilized [2. 121 by placing it in a dialysis bag (Spectralpor2, mol. wt. cutoff 12,000) and dialyzing it in the disassembly buffer F (see Table 1) for 1 h at 10°C. All solutions used with the axoplasm were passed through a Chelex 100 ion exchange column in the K form (Bio-Rad Labs, Richmond, CA, USA) to lower [Ca2+] to below 0.1 @I. The liquefied axoplasm was centrifuged at 100,000 g for 1 h to remove organelles and other particulate matter and then dialyzed overnight at 10°C in various low ionic media used for strength before being experimentation. The resulting material contained 3-10 p.g protein plII-‘.
At low or normal ionic strength for a period of weeks, filaments will reform from these soluble proteins. This process of filament reformation, which occurs spontaneously but slowly, can be hastened by freezing and thawing the solutions of liquefied axoplasm 1131. When this protein solution was frozen and thawed, a gel was formed and was sedimented to a pellet by centrifugation at 100,000 g and 10°C for 1 h. The proteins which remained in solution after this procedure will be referred to as the supernatant, and the pellet will be referred to as the gel pellet. SDS-electrophoresisand autoradiography The procedure of Mamyama et al. [l] was used to identify calcium-binding proteins within the axoplasm. The liquefied axoplasm, supematant, or gel pellet was resolved by SDS-PAGE [14] (7.5-15% acrylamide gradient), and electrophoretically transferred to a nitrocellulose membrane [15]. The nitrocellulose containing the resolved proteins was incubated with buffer D (Table 1 and either 0.4, 1.0, or 2.0 p.M CaClz with 2 tracer 4 Ca (0.3-1.4 pCi/ml) for 20 min at room temperature and then rinsed with distilled water [l]. After air drying, the nitrocellulose paper was exposed to Kodak XAR-5 film for 48 h at -70°C while the film and nitrocellulose were clamped between two plexiglass sheets. Film was developed by an automatic processor. Care was taken to conduct each experiment with uniformity and to avoid contamination of the incubation solutions with calcium. Densitometric tracings were made from X-ray film and from the nitrocellulose paper
Table 1 Solutions*
KC1 (a) A B C D
300 300 300 200
E F
500
HEPES hw
PIPES hw
TRIS
10 10 10 1
PH
hw 7.0-8.0 6.8 9.0 6.8 7.5 7.5
cac12
Mg
(Pw
tmW
O-10 O-10 O-10 0.4,1,2
f5 f5 f5 2
GUY Wf)
180 180
K-Asp
C)si
hw
hw
75 75
116 116
*Sohhm A, B andC usedin dialysisexperiments D usedto incubate x&cc&lose transfers beforeautoradiogaphy; E, for &e titration experiments; F, for liquefying theaxoplasm
363
Ca INTERACTlON WITH NEUROFKAMENTS
using a non-linear least squares method, Scatchard plots were made of the data.
and
Concentrationdialysis ‘bar*A second dialysis procedure for examining 45Ca attachment to the proteins was as follows: The protein solution (10-15 pl at -2 p&l) was added to 0.5 ml of buffer containing different concentrations of CaClz (0.2 to 10 @I) with tracer 45Ca. After 30 min incubation at 4°C. the buffer was filtered and the protein concentrated in centrifugal-concentrator/ tiltrator tubes (Bio-Molecular Dynamics, Beaver-ton, Fig. 1 The dialysis chamber consisted of upper and lower OR, USA). The filtrate and the concentrated protein compartments separated by a’ dialysis membrane. -he upper solution were analyzed for 45Ca radioactivity. The The lower compartment contained the protein solution. free 45Ca activity was subtracted from the total 45Ca compartment was filled with constantly circulating dialysis of the protein solution to determine the bound 45Ca, solution. Both compartments were stirred with magnetic stir bars and the data were normahz& to protein 7 mm in length. The dialysis medium was solution A, B, or C concentration. (Table 1) containing the amido-stained proteins using a Bio-Rad model 620 video densitometer (P.O. Box 708, 220 Maple Avenue, Rockville Center, NY 11571, USA). Dial,ysis The dialysis chamber illustrated in Figure 1 and based on the flow dialysis chamber of Colowick and Womack [16] was used for equilibrium dialysis measurements. The chamber consisted of upper and lower compartments separated by a dialysis membrane. The upper compartment contained the protein solution. The lower compartment was filled dialysis medium. Both with circulating compartments were constantly stirred. The dialysis solution contained 300 mM KCI, a pH buffer, and tracer 45Ca (0.02 pCifml). At intervals of eight to twenty-four h, samples were taken from the upper chamber to determine the amount of radioactive calcium in the protein solution. This was compared with the radioactivity in an equal volume of the The amount of Ca bound to circulating bath. protein was determined by subtracting free calcium from the total calcium content of the protein Data were normalized to the protein solution. content of the solution [17]. Plots of bound versus tree calcium were fitted to saturation binding curves
Titrationsand Ca electrodes Measured volumes (100 pl) of liquefied axoplasm or of buffer solution E (Table 1) were transferred to clean 0.5 ml Eppendorf tubes at 10°C. Protein content was determined by the Bradford assay 1171, and total Ca was measured by atomic absorption at the beginning and end of the titration. Ca was added with a 1 p.l Hamilton syringe. Free Ca was using Ca-specific determined calibrated microelectrodes constructed by imbeddmg the neutral carrier ETHlOOl [18] in a PVC matrix [19]. The reference electrode had 1 M!L! resistance with filling solution of 3 M KCl. The calcium electrodes were calibrated in solution A containing 10m3 M Ca2’, 10e5 M Ca2’ (0.344 mM CaClti1.0 mM NTA, pH 8.0), or 1.75 x 10e7 M Ca2’ (0.715 mM CaCld1.0 mM EDTA, pH 7.0). These conditions for calibration were near those of the measurements. Atomicabsorption After dilution of 5 pl samples in 95 NO.1 N HNOs, 10 pl aliquots were analyzed in a graphite furnace attachment to a model 460 Perkin-Elmer atomic (Perkin-Elmer, spectrophotometer absorption Norwalk, CT, USA). Instrument parameters were 20 s drying at lOO’C, 20 s charring at 1000°C. and
CELL CALCIUM
364
LAX
SUP
ReSUlt.9
GEL
-200 kDa -116 kDa -66 kDa
-45 kDa
-31 kDa
-22 kDa -14 kDa Fig. 2 Electmphoretic sepamtion of protein constituents of axoplasm. SDS-PAGE was performed with 7.545% acryltide. gradient as described in Materials and Methods. The proteins were stained with coomassie blue and the molecular weights determined with molecular weight standards. The axoplasm was liquefied by dialysis for 1 h at 10°C in Solution F (Table 1) and centrifuged at 100,000 g and WC for 1 h to remove particulate matter. Afkr overnight dialysis at 10°C in solution E (Table l), the sample (-5 p.g protein pl-‘) remaimxl in the form of liquefied axoplasm (LAX). The LAX was then fkzen, thawed, and then centrifuged to form the gel pellet (GEL) and supematant (SUP) hXionS
5 s atomizing at 2,700”C. Light was detected at 422 nm through a 0.7 run slit Calibration of 1, 2, 10, and 15 nmoYml were made from Fisher atomic absorption standards (Fisher Scientific). Atomic absorption measurements were made in conjunction with titration experiments and calcium electrode experiments as follows. The calcium content of protein solutions was determined by atomic absorption; known amounts of calcium were then added to these protein solutions while recording free calcium with microelectrodes. At the end of the experiment, the calcium content of the solutions was again measured by atomic absorption. This value agreed with the original measurement plus the known added calcium witbin 20%.
Neurojilament content and separation of liquefied axoplasminto fractions The 10 nm neurofilaments from Myxicofa giant axons are thought to consist almost entirely of protein subunits with apparent molecular mass of Figure 2 shows an 150 and 160 kD [4]. electrophoretic analysis of three fractions of axoplasmic proteins from Myxicola. At the left of the figure, the lane labeled ‘LAX’ contained axoplasmic proteins liquefied by dialysis in disassembly buffer F (Table 1) as in Materials and The particulate matter, including Methods. mitochondria and endoplasmic reticulum, had been removed by centrifugation at 100,000 g and 10°C for 1 h. The lane on the right of the figure (labeled GEL) contained proteins from a gel pellet formed by freezing, thawing, and centrifuging the liquefied axoplasm as described in Materials and Methods. The lane in the center contained the supernatant proteins which did not form into a gel after freezing and thawing. The supematant fraction contained little of the 150-160 kD neurofilament proteins [3]. The gel fraction contained a large amount of these proteins, as did the sample of liquefied axoplasm. Autoradiography The autoradiographic procedures of Maruyama et al. [l] provide an ideal method for testing whether different axoplasmic proteins bind calcium. A simple demonstration that neurotilament proteins specifically bind calcium is shown in Figure 3. The exposed X-ray film shown on the upper panel reveals proteins to which 45Ca remained bound after soaking in the 45Ca medium and rinsing in distilled water (see Materials and Methods). The panel below shows the same strip of nitrocellulose after staining with amid0 black to reveal all proteins that were transferred to the nitrocellulose. It should be emphasized that the autoradiographs give no information about the relative proportions of the different Ca-binding proteins in the axon. This is because the electrophomtic mobilities of the different proteins are not the same, and proteins are not transferred quantitatively to the nitrocelhrlose.
365
200 Fig. 3
Autoradiographic
electrophoretically
Ca-binding
transferred
(Table 1). and 1 ph4 CaCh location of &binding using a Bio-Rad
kDa
assay.
The protein constitue.nts
to nitmceUulose with tracer “Ca
proteins.
paper.
calcium
were separated the proteins
water, dried, and covered in this assay.
by SDS-PAGE,
and
was soaked in solution D
with X-ray film to reveal the
The densitometric
tracings
paper containing
were made
amid0 stained
mode was used with the autoradiographs
Dialysis
examine
axoplasm
containing
14kDa
kDa
mode was used with the nitrccellulose
We found the high molecular weight axoplasmic proteins to be incompletely transferred; therefore, the total calcium binding capacity of neurofilament proteins is under-represented on this autoradiograph. The low molecular weight calcium binding protein could be calmodulin as a protein of apparent molecular mass 13-15 kD is present in the axoplasm of invertebrates [20].
To
membrane
kD proteins appeared prominently The reflection
45
of the liquefied
The nitmcellulose
for 20 min, rinsed in deionized
The 150-160
model 620 video densitometer.
proteins, and the transmission
66kDa
116kDa
binding
properties
of
axoplasmic proteins in another way, dialysis experiments were done using the ColowickWomack chamber described in Materials and Methods. This chamber made it possible to follow, in time, the attachment of 45Ca to the axoplasmic proteins. Binding increased toward a maximum as the chamber and protein equilibrated with 4sCa. Radioactive Ca could then. be displaced from the axoplasmic proteins by adding low concentrations of unlabeled CaCh to the dialysis bath (Fig. 4). The data in this and subsequent experiments were normalized to the protein concentration.
CELL CALCIUM
366
0
20
40
60
Fig. 4 “Ca proteins.
attachment
displaced
bath:
140
160
160
from, axoplasmic
pM “CaCl2.
3.05 pM
The ratio drops, i.e., “Ca
as unlabeled or
10.05
pM
is
CaClz is added to the CaCl2.
The dialysis
solution B (open circles) or solution A (closed
0.004 g% Na Azide, and 5 mM MgCl2 (Table 1). The
ascending
curves
determined points.
to, and displacement
from the protein
medium contained circles),
120
(h)
The ratio of bound to free ‘%a is shown as a function
of time for 0.05 dialysis
100
80 TIME
physiological range, the Ca binding site is unaffected by protons. Although calcium binding to the liquefied axoplasm was unaffected by pH between 6.8 and 7.5 (Fig. 6); dialyzing the protein for 24 h at pH 9.0 altered the binding in an unexpected way. The Scatchard plot of paired samples from the same batch of protein (Fig. 7) suggests that dialysis at pH 9.0 decreases the number of calcium binding sites This behavior within the axoplasmic protein. suggests that the interaction between the binding site and protons is an indirect one and does not involve a competition between Ca2’ and H+. Such a simple interaction would result in an increased affinity for calcium with no change in the concentration of sites. This effect of pH on the Ca binding activity is reversible as shown in Figure 7. This experiment was done using the concentration/filtration method as described in Materials and Methods because smaller amounts of the protein could be used with this method, and the duration of the experiment was much shorter.
were drawn
by a non-linear
Descending
according least-squares
to single exponentials curve fit to the data
curves were drawn by hand.
Experiments
carried out at 1o’C
Effect ofpH and h4g
Ca titrationof nativeaxoplasmicproteins
Using equilibrium dialysis, calcium binding to the liquefied axoplasm and the supematant fi-actions were examined at pH 6.8 and pH 7.5, with and without 5 mM Mg. Figure 5 shows no effect on calcium binding of adding Mg to the liquefied axoplasm (pH 6.8), suggesting that Mg cannot compete for this calcium binding site. The binding isotherm of Figure 5 suggests a single class of calcium binding sites in the extract autoradiographs, on the other hand, show more than one type of calcium binding protein. Our of these results is ‘that the interpretation neurofilament protein makes up a larger fraction of the total protein in the extract than it represents in the nitrocellulose transfers (compare Figs 2 and 3). and that other calcium binding proteins, if they are of a different type, are present in small amounts in the extract. In PaireJl experiments from the same batch of axoplasmic protein, we detected no difference between Ca binding to the liquefied axoplasm at pH 7.5 or at pH 6.8. Figure 6 is a plot of such an This suggest& tit b this nmw experiment.
The ionic medium used in the dialysis experiments of Figures 5,6, and 7 (300 mM KCl, 5 mM MgCl2 and 10 mM pH buffer, Table 1) differed in composition from the native ionic medium of the axon. The amino acids and organic anions present 6r pH
6.6
5-
2 2 i $
3
0
_
A
Mg free 5mMMg
2’
o 0.0
2.0
4.0
6.0
6.0
10.0
p M[co**] Ng. 5 Efkt Data obtained
of Mg on calcium binding to extracts of axoplasm. by the equilibrium
with and without 5 n&l Mgclz
accordingto V-
a least-square
dialysis
method
in solution
B
(Table 1). The curve was drawn
fit to the Michaelis-Menten
- 5.8f0.25. k/a - 1.9f0.3
curve.
Ca INTERACTION
6
367
WITH NEUROFILAMENTS
supematant fraction (2 f. 1 pmoles g-l), supporting the notion that the fraction containing relatively more neurofilament proteins also had relatively more calcium binding capacity.
-
5_
Discussion Ca site on new-ofilarnents I
0.0
2.0
4.0
6.0
8.0
10.0
p M [co2*]
Ng. 6 Effect of pH on calcium Data obtained
by the equilibrium
binding to extracts of axoplasm. dialysis
and B with 5 mM MgCl2 (Table 1). V,,
method in solutions
A
- 2.8 f 0.3 ~I&,
Kin - 2.8 f 0.7 w
in the axon were not incIuded in the dialysis solutions because their finite affinity for calcium would complicate the analysis. Potassium chloride was used in the dialysis experiments; however, at high concentration, KC1 disrupts neurofilaments [12]. In order to examine calcium binding under conditions in which organic anions were present and KC1 reduced, a series of Ca titration experiments were done using liquefied axoplasm [2] that had been dialyzed overnight in a medinm similar to that found in the axon (solution E, Table 1). Protein solution (100 pl) was placed in a 0.5 ml micro-centrifuge tube and calcium was added from a 1 mM stock solution with a 1 pl Hamilton syringe. The free calcium concentration in the protein solution was determined with calibrated Ca ion microelectrodes and the total calcium was determined by atomic absorption and from the Control known amount of added calcium. experiments were done in which the ionic solution without protein was titrated. Results from such an experiment arc shown in Figure 8A. The normalized difference between the total calcium in solutions with and without protein as a function of free Ca concentration is plotted in Figure 8B. The calcium capacity determined in this manner was similar to that determined by the dialysis method. The estimated Ca dissociation constant was -1 p.M. The calcium capacity of the liquefied axoplasm fraction determined by the titration method (5.0 f 0.3 ]umoles g-’ protein) was greater than that of the
Three lines of evidence suggest that the 150 and 160 kD axoplasmic proteins are authentic calcium binding proteins: (a) The autoradiographic studies show that when the proteins are separated in SDS-polyacrylamide and transferred to nitrocellulose, the 150-160 kD proteins bind 45Ca in a specific manner (Fig. 3); (b) the equilibrium dialysis results show that axoplasmic gsmteins have one main high affinity binding site for Ca (Fig. 4); and (c) titration experiments suggest a calcium binding component in axoplasmic proteins containing neurofilaments but less capacity in fractions of axoplasmic proteins partiaIly depleted of neurofilaments (Fig. 8).
0.5
0.0
1 .o
CALCIUM Fig.
7
Reversibility
of pH 9.0 effect.
concentration/filtration single
batch
fraction
method.
of protein
was dialysed
1.5
BOUND/FREE
(pl/pg)
Data obtained
The extract
was divided
(180
into three fractions:
for 24 h at 10°C with solution
(Table 1) (open circles); Vmax - 2.3 f 0.1 wol/g. 0.1 pM; one was dialysed (Table 1) (open triangles);
by the
@) from
A, pH 7.0
Kin - 2.2 f
for 24 h at 10°C with solution Vm, - 0.5 f 0.1 pal/g,
C
Kin - 0.5 f
0.2 w
and one was dialysed
followed
by solution A for 24 h (Table 1) (filled circle); the pH
effect was reversible. A (Table l), V,
at 10°C with solution
a
one
For the combined
- 2.8 f 0.4 pal/g,
C for 24 h
data at pH 7.0, solution
KUZ - 2.6 f 0.7 @I
368
CELL CALCIUM
TITRATION EXPERIMENT
If 80% of this were the 150-160 kD neurofilament protein, then the gel would contain a total of 0.8 x 23.5 g kg-’ A60 kD = 118 pm01 neurofilament protein kg-’ gel. This compares favorably with the total calcium content measured in reformed gels of 100-200 pm01 calcium kg-’ gel (see Table 2).
I
pH and Mg dependence
-7
-6
-5 log[Ca
-4
“1
89
.
COLLECTEDAND NDRhMllZED DATA
Our interpretation is that a large part of the 45Ca binding to the axoplasmic proteins is the result of Ca attachment to neurofilament proteins. Magnesium at concentrations of 5 mM does not interfere with the binding of ph4 concentrations of calcium (Fig. 5). This site is, therefore, not an indiscriminate divalent ion-binding site. Increasing the dialysis pH from 7.0 to 9.0 reduced the binding (Fig. 7). This effect of pH cannot be ascribed to simple competition between Ca2’ and Ht as such an interaction would be expected to i&ease the Ca2’ affinity as pH is increased to 9.0. The effect of pH on the high affinity calcium binding seems to Titratable groups near the be more complex. binding site may have an influence on protein protein-protein conformation or influence interactions involving calcium and the neurotilament proteins.
1::Q&G?&8 8--
B
T
- -0
1
**
w
’ 1 5uP O+
4
-&-
-7
1
-6
-5
-4
log[Ca “1 8
Calcium
determined error
bars
normalizing nonlinear curve.
binding
in
liquefied
axoplasmic
proteins
by the titration method in a solution E (Table 1). The represent
SEM
of
5 to
13 measurements
to the protein concentration. least squares
after
The line represents
a
fit of the points to a Michaelis-Menten
The best fit parameters
were Vmax - 5.0 f 0.3 pmol/g,
Ku2 = 0.4 f 0.1 pM for the LAX, and V,
- 2 f 1 pmol/g, Ku-z
Batch-to-batch variability
- 0.3 f 0.3 for the SUP
Stoichiometry
The calcium binding capacity observed for the axoplasmic proteins can be qualitatively accounted for by one or two binding site(s) per 160-150 kD protein, If we assume conservatively that half of the axoplasmic protein in fresh preparations consists of the 150-160 kD peptides, and one site per peptide, then the Ca-binding capacity should be OS/160 kD = 3.1 pm01 g-’ axoplasmi~protein, which is nearly equal to the 4.2 pm01 g observed in liquefied Thus, one site per peptide seems axoplasm. sufficient to account for nearly all of the Ca binding. In addition, the calcium content of reformed gel pellets (Table 2) is consistent with a stoichiometry of one or two sites per protein. The protein content of the gel pellet was 23.5 + 2.6 g protein kg-’ gel.
One observation which we have not yet been able to satisfactorily explain is the variability in the calcium Table 2 Total calcium
content
of reformed axoplasmic
gels Reference
[Ca2”1
Total Ca
of IAX before
of suprnatant
GEL
gel formed
after gel formed
(WJ
(W)
[CP]
hmolelkg)
6-26-85
8.0
7-6-85
0.3
0.5
82
7-13-85
0.5
0.6
104
7-22-85
0.5
0.8
135
7-29-85
0.3
0.3
7-29-85
50.0
8-6-85
0.3
0.4
45
8-6-85b
6.3
6.3
176
200
66 230
Ca INTERACTION WITH NBUROFILAMENTS
binding parameters from one batch of protein to another. The binding capacity, for example, ranged from -3 pmoles/g protein to above 6 pmoles/g protein in different batches. Since each batch of protein was taken from at least five worms, it seems unlikely that this variability is the result of differences among worms. However, it is possible that the variability among batches is due to differences in the health or physiological state of batches of worms. When binding experiments were done on samples of protein from the same batch, very consistent results were obtained. Yet when we tested the protein composition of different batches that had different binding capacities, we were not able to detect obvious differences in the SDS electrophomtic pattern. It is known that treatment of Myxicolu neurofilament proteins with extraction media such as urea can change the appearance of the protein as seen in electron micrographs [21]. We do not know yet whether these treatments also change the Ca-binding properties of the neurofilaments or whether small differences in the extraction procedures can alter the neurofilament proteins and their interaction with calcium. Relutive significance of neurojilament Ca buflering
The apparent, dissociation constant of the presumed neurofilament Ca binding site as determined in these studies was -3-6 p.M (free calcium concentration). The significance of this protein as a buffer for intracellular calcium will depend on whether the free calcium in the axon ever approaches these concentrations. Our measurements of calcium activity in extracted Myxicolu axoplasm using calcium selective microelectrodes have generally been higher (0.2 @I) [22] than those seen in other invertebrate preparations (0.02 to 0.05 l&I) (e.g., see [23]). In fact, we have recorded stable calcium activities as high as 0.6 j&I in intact axons of this Assuming a calcium activity preparation. coefficient of 0.3 in the axoplasm (24), the free calcium concentration might reach values as high as 2 p,M. If this were to occur, then one-fourth to one-half of the Ca binding sites would be occupied, accounting for 25 to 100 pm01 intracellular calcium/liter of axoplasm. Previous measurements
369
of the total analytical Ca of fresh axoplasm from MyxicoZu gave an average of 580 pmol/l axoplasm
[22]. The analytical calcium determined in three worms from a single batch gave an average value of 467 + 136 pmol/l axoplasm. We conclude from these estimates and determinations that this protein is a significant calcium buffer. Although it is clearly not the major source of total Ca in the axon, it probably accounts for much of the energy-independent calcium-binding capacity (6-60 pm01 Cakg axoplasm) that was determined by Baker and Schlaepfer [9]. Function of neurojilament Ca site
The function of Ca binding sites on neurofilaments is not known. They might function simply as calcium pools or calcium buffers. Thus, sites would be available throughout the axon to reduce any phasic changes in cytosolic calcium or confine such changes to specific regions of the cell. Additional, mom speculative functions could be the regulation of filament formation and hence, axon size [WI, or the regulation of the interaction of other cytosolic substances with the nemofilaments. In motorneurons, for example, there is a relationship between activity and size 1261, the molecular basis for which is unknown. Regulatory calcium binding sites on neurofilaments might have roles in the formation of particular neuronal geometries during development and/or in their response to physiological stimuli.
Acknowledgements We wish to thank Drs Leona Young and Karsten Gammeltoft for help with some of the early experiments, Dr John Wood for use of the densitometer, and Mrs Janice Abercrombie for reading the manuscripts. This work was supported by NM NS19194.
References 1. Maruyama K. h%kawa T. Ebashi E. (1984) Detection of calcium binding by “Ca autoradiography on nitrocellulose membrane after sodium dodecyl sulfate gel electmphomsis. J. B&hem., 95,511-519. 2. Gilbert DS. (1975) Axoplasm amhitccture and physical properties as seen in the Myxicola giant axon. J. Physiol.,
370 253,257-301. 3. Gilbert DS. Newbv BJ. Anderton BH. (1975) Neutolilament disguise destruction, and‘discipline. Natum, 256, 586-589. 4 Lasek RJ. Krishnan N. Kaisetman-Abramof IR. (1979) Identification of the subunit proteins of IO-MI neurofilaments isolated f&m axoplasm of squid and Myxicola giant axons. J. Cell Biol., 82, 336-346. 5 Eagles AM. Gilbert DS. Maggs A. (1981) The polypeptide composition of axoplasm and the nemofilaments from the marine worm Myxicola inhtndibulum. B&hem. J., 199, 89- 100. 6 Abercmmbie RF. Gammeltoft K. Young L. (1986) Ca-binding site on a neuroflament protein of the Myxicoh giant axon. Biophys. J., 49, 116a. 7. Abemrombie RF. Gammeltoft K. Jackson J. Young L. (1986) An intracellular calcium binding site on neurofilament proteins of Myxicola giant axon. J. Gen. Physiol., 88, 9a. 8. Lefebvre S. Mushynski WE. (1987) Calcium binding to untreated and dephosphorylated porcine neurotilaments. B&hem. Biophys. Res. Commun., 145.1006-1011. 9. Baker PF. Schlaepfer WW. (1978) Uptake. and binding of calcium by axoplasm isolated from giant axons of J-.&go and Myxicola. J. Physiol., 276, 103-125. 10. Ortiz OE. Sjodin RA. Boyne A. (1985) ATP-dependent calcium accumulation by non-mitochondrial organelles of axoplasm isolated from Myxicola giantaxons.Biochim. Biophys. Acta, 814, 13-22. 11. Abercrombie RF. Masukawa LM. Sjodin R4. Livengood D. (1981) Uptake and release of “Ca by Myxicola axoplasm. J. Gen. Physiol., 78, 413-429. 12. Gilbert DS . (1975) Axoplaam chemical composition in Myxicola and solubility properties of its structural proteins. J. Physiol., 253,303-319. 13. Huneeus FC. Davison PF. (1970) Fibrillar proteins from squid sxons I. Neurofilament protein. J. Mol. Biol., 52, 415-428. 14. Laemmli UK. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227,680-685. 15. Towbin H. Staehelin T. Gordon J. (1979) Electrophomtic transfer of proteins from polyacrylamide gels to nitmcellulose sheets: Ptocedure and some applications.
CELL CALCIUM Proc. Natl. Acad. Sci. USA, 76,4350-4354. 16. Colowick SP. Womack FC. (1969) Binding of diffusible molecules by macromolecules: Rapid measurement by rate of dialysis. J. Biol. Chem. 244,774777. 17. Bradford MM. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. B&hem., 72, 248-254. 18. Lamer F. Steiner RA. Ammann D. Simon W. (1982) Critical evaluation of tbe applicability of neutral carrier-based calcium selective microelectrodes. Anal. Chim. Acta, 135,51-59. 19. Tsien RY. Rink TJ. (1980) Neutral canier ion-selective microelectrodes for measurement of intracellular free calcium. Biochim. Biophys. Acta, 599, 623-638. 20. Alema S. Calissano P. Rusca G. Guiditta A. (1973) Identification of a calcium-binding, brain specific protein in the axoplasm of squid giant axons. J. Nemo&em., 20, 681-689. 21. Krishnan N. Kaiserman-Abramof lR. Lasek RJ. (1979) Helical substructure of nemofilaments isolated from Myxicoia and squid giant axons. J. Cell Biol., 82, 323-335. 22. Abercrombie RF. Hart CE. (1986) Calcium and proton buffering and diffusion in isolated cytoplasm from Myxicola axons. Am. J. Physiol., 250, C391-C405. 23. Dipole R. Requena J. Brinley FJ. Mullins LJ. Scarpa A. Tiffert T. (1976) Ionized calcium concentrations in squid axons. J. Gen. Physiol., 67, 433-467. 24. Thomas MV. (1982) Techniques in Calcium Research Academic Press. 25. Lasek RJ. Phillips L. Katz, MJ. Autilio-Gambetti L. (1985) Function and evolution of nemofilament proteins. AM. NY. Acad. Sci., 455,462-478. 26. Hemreman E. Clamam HP. Gillies JD. Skinner RD. (1974) Rank-order of motoneurons within a pool: Law of combination. J. Neurophysiol., 37, 1338. Please send reprint requests to : Dr Ronald F. Abercrombie, Department of Physiology, Emory University School of Medicine, Atlanta, GA 30322, USA. Received Revised Accepted
: 8 November 1989 : 8 December 1989 : 7 January 1990
W-370
@LongmanGrcupUKLtdlWO
Ca*+ binding by Myxkola neurofilament proteins R.F. ABERCROMBIE, N.F. AL-BALDAWI and J. JACKSON Department of Physiology, Emory University School of Medicine, Atlanta, Georgia, USA Abstract - Titrimetric, ‘%a dialysis, and autoradiographic methods were used to examine how axopiasmic proteins from the giant neuron of the marine anneiid Myxicola infundibulum bind calcium. Following the autoradiographic method of Maruyama et al. [I], the 150-160 kD neurofliament subunits were identified as prominent intracellular Ca-binding peptides. Using equilibrium dialysis, extracts of axopiasmic proteins (MO% neurofiiament subunits) were examined in 300 mM KCI at different concentrations of free Ca and Mg, and at different pH. Axoviasmic proteins showed a high affinity Ca binding site (KIT 3-6 fl, capacity 3-7 mole g- protein) at pH 6.6 or pH 7.6. Changing the Mg concentration from 0 to 5 mM had no effect on the Ca binding. Elevating the dialysis pH from 7.0 to 9.0 reduced the apparent number of binding sites for Ca. Using microelectrodes to record the free Ca, microtitrations of axopiasmic proteins were completed by adding small amounts of CaCi2 to 100 u.l volumes of protein solutions. in a medium containing ionic constituents closely resembling those of the Myxicola axon, a Ca binding capacity of 5.0 pole g-’ protein and a Kin of -1 j.N were measured. The Myxicola giant axon has been described by Gilbert [23 as a nearly ideal source of neurofilaments since these filaments comprise a relatively large component of the total protein of the axon. Previous studies using Myxicolu have shown that neurofdaments are composed of two major proteins with apparent molecular mass close to 160 D and 150 kD [3-51. The ability of neurofdaments to bind Ca with high affinity has been described only recently [6,7]. Lefebvre and Mushynske [8] have presented data Abbreviations used TRIS, Tris(hydroxymethyl)amiminomethane; EDTA, Ethylenediaminetetraacetic acid; HEPES, N-[2-hydroxyethyll-lpiperazine-N’-[2-ethanesulfonic acid]; Kin, MichaelisMenten constant for half-saturation; NTA, Nitrilotriacetic acid, PAGE, Polyacrylamide gel electrophoresis; PIPES, Piperazine-N,N’-bis[2-ethanesulfonic acid]; SDS, Sodium dodecylsulfate
describing the calcium-binding properties of untreated and in vitro dephosphorylated neurofilaments from porcine spinal cord. Their binding studies were done at low ionic strength (20 mM HEPIS, 4 mM diethanolamine, pH 6.8). The b&iing studies of the present work have been done under physiological ionic strength using three different technical approaches and at different pH and Mg concentrations in or&r to better characterize this novel Ca binding site. Myxicola giant axons contain energy-dependent calcium binding systems involving mitochondria [9] and endoplasmic reticulum [lo]) as well as an unidentified energy-independent system [9, 111. The main observation of this report is that significant neurofilaments may be a energy-independent Ca buffer in the axon. The capacity attributed to neurofilament proteins in this study (-4 pm01 g-’ protein) corresponds to l-2 Ca 361
362
CELL CALCIUM
bound per 150-160 kD neurof~ament monomer. Based on the molar concentration of neurofilament protein in axoplasm, these sites could hold a maximum of 100-200 pmole Ca per kg of Myxicolu axoplasm at elevated Cai. At normal levels of Cai (-200 nM), these sites would hold one-tenth or less of this amount of calcium.
Materials and Methods Preparation Myxicola were collected in the Bay of Fundy by Mr Robert Bosien, Deer Island, New Brunswick Canada EOG-lR0, shipped by air to Atlanta, and maintained in a refrigerated (1O’C) seawater aquarium for up to eight weeks. Axons 700-1.000 pm in diameter were dissected, bathed for ten min in Ca-free artificial seawater, blotted, and cut open. Using forceps, the axoplasmic gel was extracted from 5-10 giant axons. This gel was solubilized [2. 121 by placing it in a dialysis bag (Spectralpor2, mol. wt. cutoff 12,000) and dialyzing it in the disassembly buffer F (see Table 1) for 1 h at 10°C. All solutions used with the axoplasm were passed through a Chelex 100 ion exchange column in the K form (Bio-Rad Labs, Richmond, CA, USA) to lower [Ca2+] to below 0.1 @I. The liquefied axoplasm was centrifuged at 100,000 g for 1 h to remove organelles and other particulate matter and then dialyzed overnight at 10°C in various low ionic media used for strength before being experimentation. The resulting material contained 3-10 p.g protein plII-‘.
At low or normal ionic strength for a period of weeks, filaments will reform from these soluble proteins. This process of filament reformation, which occurs spontaneously but slowly, can be hastened by freezing and thawing the solutions of liquefied axoplasm 1131. When this protein solution was frozen and thawed, a gel was formed and was sedimented to a pellet by centrifugation at 100,000 g and 10°C for 1 h. The proteins which remained in solution after this procedure will be referred to as the supernatant, and the pellet will be referred to as the gel pellet. SDS-electrophoresisand autoradiography The procedure of Mamyama et al. [l] was used to identify calcium-binding proteins within the axoplasm. The liquefied axoplasm, supematant, or gel pellet was resolved by SDS-PAGE [14] (7.5-15% acrylamide gradient), and electrophoretically transferred to a nitrocellulose membrane [15]. The nitrocellulose containing the resolved proteins was incubated with buffer D (Table 1 and either 0.4, 1.0, or 2.0 p.M CaClz with 2 tracer 4 Ca (0.3-1.4 pCi/ml) for 20 min at room temperature and then rinsed with distilled water [l]. After air drying, the nitrocellulose paper was exposed to Kodak XAR-5 film for 48 h at -70°C while the film and nitrocellulose were clamped between two plexiglass sheets. Film was developed by an automatic processor. Care was taken to conduct each experiment with uniformity and to avoid contamination of the incubation solutions with calcium. Densitometric tracings were made from X-ray film and from the nitrocellulose paper
Table 1 Solutions*
KC1 (a) A B C D
300 300 300 200
E F
500
HEPES hw
PIPES hw
TRIS
10 10 10 1
PH
hw 7.0-8.0 6.8 9.0 6.8 7.5 7.5
cac12
Mg
(Pw
tmW
O-10 O-10 O-10 0.4,1,2
f5 f5 f5 2
GUY Wf)
180 180
K-Asp
C)si
hw
hw
75 75
116 116
*Sohhm A, B andC usedin dialysisexperiments D usedto incubate x&cc&lose transfers beforeautoradiogaphy; E, for &e titration experiments; F, for liquefying theaxoplasm
363
Ca INTERACTlON WITH NEUROFKAMENTS
using a non-linear least squares method, Scatchard plots were made of the data.
and
Concentrationdialysis ‘bar*A second dialysis procedure for examining 45Ca attachment to the proteins was as follows: The protein solution (10-15 pl at -2 p&l) was added to 0.5 ml of buffer containing different concentrations of CaClz (0.2 to 10 @I) with tracer 45Ca. After 30 min incubation at 4°C. the buffer was filtered and the protein concentrated in centrifugal-concentrator/ tiltrator tubes (Bio-Molecular Dynamics, Beaver-ton, Fig. 1 The dialysis chamber consisted of upper and lower OR, USA). The filtrate and the concentrated protein compartments separated by a’ dialysis membrane. -he upper solution were analyzed for 45Ca radioactivity. The The lower compartment contained the protein solution. free 45Ca activity was subtracted from the total 45Ca compartment was filled with constantly circulating dialysis of the protein solution to determine the bound 45Ca, solution. Both compartments were stirred with magnetic stir bars and the data were normahz& to protein 7 mm in length. The dialysis medium was solution A, B, or C concentration. (Table 1) containing the amido-stained proteins using a Bio-Rad model 620 video densitometer (P.O. Box 708, 220 Maple Avenue, Rockville Center, NY 11571, USA). Dial,ysis The dialysis chamber illustrated in Figure 1 and based on the flow dialysis chamber of Colowick and Womack [16] was used for equilibrium dialysis measurements. The chamber consisted of upper and lower compartments separated by a dialysis membrane. The upper compartment contained the protein solution. The lower compartment was filled dialysis medium. Both with circulating compartments were constantly stirred. The dialysis solution contained 300 mM KCI, a pH buffer, and tracer 45Ca (0.02 pCifml). At intervals of eight to twenty-four h, samples were taken from the upper chamber to determine the amount of radioactive calcium in the protein solution. This was compared with the radioactivity in an equal volume of the The amount of Ca bound to circulating bath. protein was determined by subtracting free calcium from the total calcium content of the protein Data were normalized to the protein solution. content of the solution [17]. Plots of bound versus tree calcium were fitted to saturation binding curves
Titrationsand Ca electrodes Measured volumes (100 pl) of liquefied axoplasm or of buffer solution E (Table 1) were transferred to clean 0.5 ml Eppendorf tubes at 10°C. Protein content was determined by the Bradford assay 1171, and total Ca was measured by atomic absorption at the beginning and end of the titration. Ca was added with a 1 p.l Hamilton syringe. Free Ca was using Ca-specific determined calibrated microelectrodes constructed by imbeddmg the neutral carrier ETHlOOl [18] in a PVC matrix [19]. The reference electrode had 1 M!L! resistance with filling solution of 3 M KCl. The calcium electrodes were calibrated in solution A containing 10m3 M Ca2’, 10e5 M Ca2’ (0.344 mM CaClti1.0 mM NTA, pH 8.0), or 1.75 x 10e7 M Ca2’ (0.715 mM CaCld1.0 mM EDTA, pH 7.0). These conditions for calibration were near those of the measurements. Atomicabsorption After dilution of 5 pl samples in 95 NO.1 N HNOs, 10 pl aliquots were analyzed in a graphite furnace attachment to a model 460 Perkin-Elmer atomic (Perkin-Elmer, spectrophotometer absorption Norwalk, CT, USA). Instrument parameters were 20 s drying at lOO’C, 20 s charring at 1000°C. and
CELL CALCIUM
364
LAX
SUP
ReSUlt.9
GEL
-200 kDa -116 kDa -66 kDa
-45 kDa
-31 kDa
-22 kDa -14 kDa Fig. 2 Electmphoretic sepamtion of protein constituents of axoplasm. SDS-PAGE was performed with 7.545% acryltide. gradient as described in Materials and Methods. The proteins were stained with coomassie blue and the molecular weights determined with molecular weight standards. The axoplasm was liquefied by dialysis for 1 h at 10°C in Solution F (Table 1) and centrifuged at 100,000 g and WC for 1 h to remove particulate matter. Afkr overnight dialysis at 10°C in solution E (Table l), the sample (-5 p.g protein pl-‘) remaimxl in the form of liquefied axoplasm (LAX). The LAX was then fkzen, thawed, and then centrifuged to form the gel pellet (GEL) and supematant (SUP) hXionS
5 s atomizing at 2,700”C. Light was detected at 422 nm through a 0.7 run slit Calibration of 1, 2, 10, and 15 nmoYml were made from Fisher atomic absorption standards (Fisher Scientific). Atomic absorption measurements were made in conjunction with titration experiments and calcium electrode experiments as follows. The calcium content of protein solutions was determined by atomic absorption; known amounts of calcium were then added to these protein solutions while recording free calcium with microelectrodes. At the end of the experiment, the calcium content of the solutions was again measured by atomic absorption. This value agreed with the original measurement plus the known added calcium witbin 20%.
Neurojilament content and separation of liquefied axoplasminto fractions The 10 nm neurofilaments from Myxicofa giant axons are thought to consist almost entirely of protein subunits with apparent molecular mass of Figure 2 shows an 150 and 160 kD [4]. electrophoretic analysis of three fractions of axoplasmic proteins from Myxicola. At the left of the figure, the lane labeled ‘LAX’ contained axoplasmic proteins liquefied by dialysis in disassembly buffer F (Table 1) as in Materials and The particulate matter, including Methods. mitochondria and endoplasmic reticulum, had been removed by centrifugation at 100,000 g and 10°C for 1 h. The lane on the right of the figure (labeled GEL) contained proteins from a gel pellet formed by freezing, thawing, and centrifuging the liquefied axoplasm as described in Materials and Methods. The lane in the center contained the supernatant proteins which did not form into a gel after freezing and thawing. The supematant fraction contained little of the 150-160 kD neurofilament proteins [3]. The gel fraction contained a large amount of these proteins, as did the sample of liquefied axoplasm. Autoradiography The autoradiographic procedures of Maruyama et al. [l] provide an ideal method for testing whether different axoplasmic proteins bind calcium. A simple demonstration that neurotilament proteins specifically bind calcium is shown in Figure 3. The exposed X-ray film shown on the upper panel reveals proteins to which 45Ca remained bound after soaking in the 45Ca medium and rinsing in distilled water (see Materials and Methods). The panel below shows the same strip of nitrocellulose after staining with amid0 black to reveal all proteins that were transferred to the nitrocellulose. It should be emphasized that the autoradiographs give no information about the relative proportions of the different Ca-binding proteins in the axon. This is because the electrophomtic mobilities of the different proteins are not the same, and proteins are not transferred quantitatively to the nitrocelhrlose.
365
200 Fig. 3
Autoradiographic
electrophoretically
Ca-binding
transferred
(Table 1). and 1 ph4 CaCh location of &binding using a Bio-Rad
kDa
assay.
The protein constitue.nts
to nitmceUulose with tracer “Ca
proteins.
paper.
calcium
were separated the proteins
water, dried, and covered in this assay.
by SDS-PAGE,
and
was soaked in solution D
with X-ray film to reveal the
The densitometric
tracings
paper containing
were made
amid0 stained
mode was used with the autoradiographs
Dialysis
examine
axoplasm
containing
14kDa
kDa
mode was used with the nitrccellulose
We found the high molecular weight axoplasmic proteins to be incompletely transferred; therefore, the total calcium binding capacity of neurofilament proteins is under-represented on this autoradiograph. The low molecular weight calcium binding protein could be calmodulin as a protein of apparent molecular mass 13-15 kD is present in the axoplasm of invertebrates [20].
To
membrane
kD proteins appeared prominently The reflection
45
of the liquefied
The nitmcellulose
for 20 min, rinsed in deionized
The 150-160
model 620 video densitometer.
proteins, and the transmission
66kDa
116kDa
binding
properties
of
axoplasmic proteins in another way, dialysis experiments were done using the ColowickWomack chamber described in Materials and Methods. This chamber made it possible to follow, in time, the attachment of 45Ca to the axoplasmic proteins. Binding increased toward a maximum as the chamber and protein equilibrated with 4sCa. Radioactive Ca could then. be displaced from the axoplasmic proteins by adding low concentrations of unlabeled CaCh to the dialysis bath (Fig. 4). The data in this and subsequent experiments were normalized to the protein concentration.
CELL CALCIUM
366
0
20
40
60
Fig. 4 “Ca proteins.
attachment
displaced
bath:
140
160
160
from, axoplasmic
pM “CaCl2.
3.05 pM
The ratio drops, i.e., “Ca
as unlabeled or
10.05
pM
is
CaClz is added to the CaCl2.
The dialysis
solution B (open circles) or solution A (closed
0.004 g% Na Azide, and 5 mM MgCl2 (Table 1). The
ascending
curves
determined points.
to, and displacement
from the protein
medium contained circles),
120
(h)
The ratio of bound to free ‘%a is shown as a function
of time for 0.05 dialysis
100
80 TIME
physiological range, the Ca binding site is unaffected by protons. Although calcium binding to the liquefied axoplasm was unaffected by pH between 6.8 and 7.5 (Fig. 6); dialyzing the protein for 24 h at pH 9.0 altered the binding in an unexpected way. The Scatchard plot of paired samples from the same batch of protein (Fig. 7) suggests that dialysis at pH 9.0 decreases the number of calcium binding sites This behavior within the axoplasmic protein. suggests that the interaction between the binding site and protons is an indirect one and does not involve a competition between Ca2’ and H+. Such a simple interaction would result in an increased affinity for calcium with no change in the concentration of sites. This effect of pH on the Ca binding activity is reversible as shown in Figure 7. This experiment was done using the concentration/filtration method as described in Materials and Methods because smaller amounts of the protein could be used with this method, and the duration of the experiment was much shorter.
were drawn
by a non-linear
Descending
according least-squares
to single exponentials curve fit to the data
curves were drawn by hand.
Experiments
carried out at 1o’C
Effect ofpH and h4g
Ca titrationof nativeaxoplasmicproteins
Using equilibrium dialysis, calcium binding to the liquefied axoplasm and the supematant fi-actions were examined at pH 6.8 and pH 7.5, with and without 5 mM Mg. Figure 5 shows no effect on calcium binding of adding Mg to the liquefied axoplasm (pH 6.8), suggesting that Mg cannot compete for this calcium binding site. The binding isotherm of Figure 5 suggests a single class of calcium binding sites in the extract autoradiographs, on the other hand, show more than one type of calcium binding protein. Our of these results is ‘that the interpretation neurofilament protein makes up a larger fraction of the total protein in the extract than it represents in the nitrocellulose transfers (compare Figs 2 and 3). and that other calcium binding proteins, if they are of a different type, are present in small amounts in the extract. In PaireJl experiments from the same batch of axoplasmic protein, we detected no difference between Ca binding to the liquefied axoplasm at pH 7.5 or at pH 6.8. Figure 6 is a plot of such an This suggest& tit b this nmw experiment.
The ionic medium used in the dialysis experiments of Figures 5,6, and 7 (300 mM KCl, 5 mM MgCl2 and 10 mM pH buffer, Table 1) differed in composition from the native ionic medium of the axon. The amino acids and organic anions present 6r pH
6.6
5-
2 2 i $
3
0
_
A
Mg free 5mMMg
2’
o 0.0
2.0
4.0
6.0
6.0
10.0
p M[co**] Ng. 5 Efkt Data obtained
of Mg on calcium binding to extracts of axoplasm. by the equilibrium
with and without 5 n&l Mgclz
accordingto V-
a least-square
dialysis
method
in solution
B
(Table 1). The curve was drawn
fit to the Michaelis-Menten
- 5.8f0.25. k/a - 1.9f0.3
curve.
Ca INTERACTION
6
367
WITH NEUROFILAMENTS
supematant fraction (2 f. 1 pmoles g-l), supporting the notion that the fraction containing relatively more neurofilament proteins also had relatively more calcium binding capacity.
-
5_
Discussion Ca site on new-ofilarnents I
0.0
2.0
4.0
6.0
8.0
10.0
p M [co2*]
Ng. 6 Effect of pH on calcium Data obtained
by the equilibrium
binding to extracts of axoplasm. dialysis
and B with 5 mM MgCl2 (Table 1). V,,
method in solutions
A
- 2.8 f 0.3 ~I&,
Kin - 2.8 f 0.7 w
in the axon were not incIuded in the dialysis solutions because their finite affinity for calcium would complicate the analysis. Potassium chloride was used in the dialysis experiments; however, at high concentration, KC1 disrupts neurofilaments [12]. In order to examine calcium binding under conditions in which organic anions were present and KC1 reduced, a series of Ca titration experiments were done using liquefied axoplasm [2] that had been dialyzed overnight in a medinm similar to that found in the axon (solution E, Table 1). Protein solution (100 pl) was placed in a 0.5 ml micro-centrifuge tube and calcium was added from a 1 mM stock solution with a 1 pl Hamilton syringe. The free calcium concentration in the protein solution was determined with calibrated Ca ion microelectrodes and the total calcium was determined by atomic absorption and from the Control known amount of added calcium. experiments were done in which the ionic solution without protein was titrated. Results from such an experiment arc shown in Figure 8A. The normalized difference between the total calcium in solutions with and without protein as a function of free Ca concentration is plotted in Figure 8B. The calcium capacity determined in this manner was similar to that determined by the dialysis method. The estimated Ca dissociation constant was -1 p.M. The calcium capacity of the liquefied axoplasm fraction determined by the titration method (5.0 f 0.3 ]umoles g-’ protein) was greater than that of the
Three lines of evidence suggest that the 150 and 160 kD axoplasmic proteins are authentic calcium binding proteins: (a) The autoradiographic studies show that when the proteins are separated in SDS-polyacrylamide and transferred to nitrocellulose, the 150-160 kD proteins bind 45Ca in a specific manner (Fig. 3); (b) the equilibrium dialysis results show that axoplasmic gsmteins have one main high affinity binding site for Ca (Fig. 4); and (c) titration experiments suggest a calcium binding component in axoplasmic proteins containing neurofilaments but less capacity in fractions of axoplasmic proteins partiaIly depleted of neurofilaments (Fig. 8).
0.5
0.0
1 .o
CALCIUM Fig.
7
Reversibility
of pH 9.0 effect.
concentration/filtration single
batch
fraction
method.
of protein
was dialysed
1.5
BOUND/FREE
(pl/pg)
Data obtained
The extract
was divided
(180
into three fractions:
for 24 h at 10°C with solution
(Table 1) (open circles); Vmax - 2.3 f 0.1 wol/g. 0.1 pM; one was dialysed (Table 1) (open triangles);
by the
@) from
A, pH 7.0
Kin - 2.2 f
for 24 h at 10°C with solution Vm, - 0.5 f 0.1 pal/g,
C
Kin - 0.5 f
0.2 w
and one was dialysed
followed
by solution A for 24 h (Table 1) (filled circle); the pH
effect was reversible. A (Table l), V,
at 10°C with solution
a
one
For the combined
- 2.8 f 0.4 pal/g,
C for 24 h
data at pH 7.0, solution
KUZ - 2.6 f 0.7 @I
368
CELL CALCIUM
TITRATION EXPERIMENT
If 80% of this were the 150-160 kD neurofilament protein, then the gel would contain a total of 0.8 x 23.5 g kg-’ A60 kD = 118 pm01 neurofilament protein kg-’ gel. This compares favorably with the total calcium content measured in reformed gels of 100-200 pm01 calcium kg-’ gel (see Table 2).
I
pH and Mg dependence
-7
-6
-5 log[Ca
-4
“1
89
.
COLLECTEDAND NDRhMllZED DATA
Our interpretation is that a large part of the 45Ca binding to the axoplasmic proteins is the result of Ca attachment to neurofilament proteins. Magnesium at concentrations of 5 mM does not interfere with the binding of ph4 concentrations of calcium (Fig. 5). This site is, therefore, not an indiscriminate divalent ion-binding site. Increasing the dialysis pH from 7.0 to 9.0 reduced the binding (Fig. 7). This effect of pH cannot be ascribed to simple competition between Ca2’ and Ht as such an interaction would be expected to i&ease the Ca2’ affinity as pH is increased to 9.0. The effect of pH on the high affinity calcium binding seems to Titratable groups near the be more complex. binding site may have an influence on protein protein-protein conformation or influence interactions involving calcium and the neurotilament proteins.
1::Q&G?&8 8--
B
T
- -0
1
**
w
’ 1 5uP O+
4
-&-
-7
1
-6
-5
-4
log[Ca “1 8
Calcium
determined error
bars
normalizing nonlinear curve.
binding
in
liquefied
axoplasmic
proteins
by the titration method in a solution E (Table 1). The represent
SEM
of
5 to
13 measurements
to the protein concentration. least squares
after
The line represents
a
fit of the points to a Michaelis-Menten
The best fit parameters
were Vmax - 5.0 f 0.3 pmol/g,
Ku2 = 0.4 f 0.1 pM for the LAX, and V,
- 2 f 1 pmol/g, Ku-z
Batch-to-batch variability
- 0.3 f 0.3 for the SUP
Stoichiometry
The calcium binding capacity observed for the axoplasmic proteins can be qualitatively accounted for by one or two binding site(s) per 160-150 kD protein, If we assume conservatively that half of the axoplasmic protein in fresh preparations consists of the 150-160 kD peptides, and one site per peptide, then the Ca-binding capacity should be OS/160 kD = 3.1 pm01 g-’ axoplasmi~protein, which is nearly equal to the 4.2 pm01 g observed in liquefied Thus, one site per peptide seems axoplasm. sufficient to account for nearly all of the Ca binding. In addition, the calcium content of reformed gel pellets (Table 2) is consistent with a stoichiometry of one or two sites per protein. The protein content of the gel pellet was 23.5 + 2.6 g protein kg-’ gel.
One observation which we have not yet been able to satisfactorily explain is the variability in the calcium Table 2 Total calcium
content
of reformed axoplasmic
gels Reference
[Ca2”1
Total Ca
of IAX before
of suprnatant
GEL
gel formed
after gel formed
(WJ
(W)
[CP]
hmolelkg)
6-26-85
8.0
7-6-85
0.3
0.5
82
7-13-85
0.5
0.6
104
7-22-85
0.5
0.8
135
7-29-85
0.3
0.3
7-29-85
50.0
8-6-85
0.3
0.4
45
8-6-85b
6.3
6.3
176
200
66 230
Ca INTERACTION WITH NBUROFILAMENTS
binding parameters from one batch of protein to another. The binding capacity, for example, ranged from -3 pmoles/g protein to above 6 pmoles/g protein in different batches. Since each batch of protein was taken from at least five worms, it seems unlikely that this variability is the result of differences among worms. However, it is possible that the variability among batches is due to differences in the health or physiological state of batches of worms. When binding experiments were done on samples of protein from the same batch, very consistent results were obtained. Yet when we tested the protein composition of different batches that had different binding capacities, we were not able to detect obvious differences in the SDS electrophomtic pattern. It is known that treatment of Myxicolu neurofilament proteins with extraction media such as urea can change the appearance of the protein as seen in electron micrographs [21]. We do not know yet whether these treatments also change the Ca-binding properties of the neurofilaments or whether small differences in the extraction procedures can alter the neurofilament proteins and their interaction with calcium. Relutive significance of neurojilament Ca buflering
The apparent, dissociation constant of the presumed neurofilament Ca binding site as determined in these studies was -3-6 p.M (free calcium concentration). The significance of this protein as a buffer for intracellular calcium will depend on whether the free calcium in the axon ever approaches these concentrations. Our measurements of calcium activity in extracted Myxicolu axoplasm using calcium selective microelectrodes have generally been higher (0.2 @I) [22] than those seen in other invertebrate preparations (0.02 to 0.05 l&I) (e.g., see [23]). In fact, we have recorded stable calcium activities as high as 0.6 j&I in intact axons of this Assuming a calcium activity preparation. coefficient of 0.3 in the axoplasm (24), the free calcium concentration might reach values as high as 2 p,M. If this were to occur, then one-fourth to one-half of the Ca binding sites would be occupied, accounting for 25 to 100 pm01 intracellular calcium/liter of axoplasm. Previous measurements
369
of the total analytical Ca of fresh axoplasm from MyxicoZu gave an average of 580 pmol/l axoplasm
[22]. The analytical calcium determined in three worms from a single batch gave an average value of 467 + 136 pmol/l axoplasm. We conclude from these estimates and determinations that this protein is a significant calcium buffer. Although it is clearly not the major source of total Ca in the axon, it probably accounts for much of the energy-independent calcium-binding capacity (6-60 pm01 Cakg axoplasm) that was determined by Baker and Schlaepfer [9]. Function of neurojilament Ca site
The function of Ca binding sites on neurofilaments is not known. They might function simply as calcium pools or calcium buffers. Thus, sites would be available throughout the axon to reduce any phasic changes in cytosolic calcium or confine such changes to specific regions of the cell. Additional, mom speculative functions could be the regulation of filament formation and hence, axon size [WI, or the regulation of the interaction of other cytosolic substances with the nemofilaments. In motorneurons, for example, there is a relationship between activity and size 1261, the molecular basis for which is unknown. Regulatory calcium binding sites on neurofilaments might have roles in the formation of particular neuronal geometries during development and/or in their response to physiological stimuli.
Acknowledgements We wish to thank Drs Leona Young and Karsten Gammeltoft for help with some of the early experiments, Dr John Wood for use of the densitometer, and Mrs Janice Abercrombie for reading the manuscripts. This work was supported by NM NS19194.
References 1. Maruyama K. h%kawa T. Ebashi E. (1984) Detection of calcium binding by “Ca autoradiography on nitrocellulose membrane after sodium dodecyl sulfate gel electmphomsis. J. B&hem., 95,511-519. 2. Gilbert DS. (1975) Axoplasm amhitccture and physical properties as seen in the Myxicola giant axon. J. Physiol.,
370 253,257-301. 3. Gilbert DS. Newbv BJ. Anderton BH. (1975) Neutolilament disguise destruction, and‘discipline. Natum, 256, 586-589. 4 Lasek RJ. Krishnan N. Kaisetman-Abramof IR. (1979) Identification of the subunit proteins of IO-MI neurofilaments isolated f&m axoplasm of squid and Myxicola giant axons. J. Cell Biol., 82, 336-346. 5 Eagles AM. Gilbert DS. Maggs A. (1981) The polypeptide composition of axoplasm and the nemofilaments from the marine worm Myxicola inhtndibulum. B&hem. J., 199, 89- 100. 6 Abercmmbie RF. Gammeltoft K. Young L. (1986) Ca-binding site on a neuroflament protein of the Myxicoh giant axon. Biophys. J., 49, 116a. 7. Abemrombie RF. Gammeltoft K. Jackson J. Young L. (1986) An intracellular calcium binding site on neurofilament proteins of Myxicola giant axon. J. Gen. Physiol., 88, 9a. 8. Lefebvre S. Mushynski WE. (1987) Calcium binding to untreated and dephosphorylated porcine neurotilaments. B&hem. Biophys. Res. Commun., 145.1006-1011. 9. Baker PF. Schlaepfer WW. (1978) Uptake. and binding of calcium by axoplasm isolated from giant axons of J-.&go and Myxicola. J. Physiol., 276, 103-125. 10. Ortiz OE. Sjodin RA. Boyne A. (1985) ATP-dependent calcium accumulation by non-mitochondrial organelles of axoplasm isolated from Myxicola giantaxons.Biochim. Biophys. Acta, 814, 13-22. 11. Abercrombie RF. Masukawa LM. Sjodin R4. Livengood D. (1981) Uptake and release of “Ca by Myxicola axoplasm. J. Gen. Physiol., 78, 413-429. 12. Gilbert DS . (1975) Axoplaam chemical composition in Myxicola and solubility properties of its structural proteins. J. Physiol., 253,303-319. 13. Huneeus FC. Davison PF. (1970) Fibrillar proteins from squid sxons I. Neurofilament protein. J. Mol. Biol., 52, 415-428. 14. Laemmli UK. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227,680-685. 15. Towbin H. Staehelin T. Gordon J. (1979) Electrophomtic transfer of proteins from polyacrylamide gels to nitmcellulose sheets: Ptocedure and some applications.
CELL CALCIUM Proc. Natl. Acad. Sci. USA, 76,4350-4354. 16. Colowick SP. Womack FC. (1969) Binding of diffusible molecules by macromolecules: Rapid measurement by rate of dialysis. J. Biol. Chem. 244,774777. 17. Bradford MM. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. B&hem., 72, 248-254. 18. Lamer F. Steiner RA. Ammann D. Simon W. (1982) Critical evaluation of tbe applicability of neutral carrier-based calcium selective microelectrodes. Anal. Chim. Acta, 135,51-59. 19. Tsien RY. Rink TJ. (1980) Neutral canier ion-selective microelectrodes for measurement of intracellular free calcium. Biochim. Biophys. Acta, 599, 623-638. 20. Alema S. Calissano P. Rusca G. Guiditta A. (1973) Identification of a calcium-binding, brain specific protein in the axoplasm of squid giant axons. J. Nemo&em., 20, 681-689. 21. Krishnan N. Kaiserman-Abramof lR. Lasek RJ. (1979) Helical substructure of nemofilaments isolated from Myxicoia and squid giant axons. J. Cell Biol., 82, 323-335. 22. Abercrombie RF. Hart CE. (1986) Calcium and proton buffering and diffusion in isolated cytoplasm from Myxicola axons. Am. J. Physiol., 250, C391-C405. 23. Dipole R. Requena J. Brinley FJ. Mullins LJ. Scarpa A. Tiffert T. (1976) Ionized calcium concentrations in squid axons. J. Gen. Physiol., 67, 433-467. 24. Thomas MV. (1982) Techniques in Calcium Research Academic Press. 25. Lasek RJ. Phillips L. Katz, MJ. Autilio-Gambetti L. (1985) Function and evolution of nemofilament proteins. AM. NY. Acad. Sci., 455,462-478. 26. Hemreman E. Clamam HP. Gillies JD. Skinner RD. (1974) Rank-order of motoneurons within a pool: Law of combination. J. Neurophysiol., 37, 1338. Please send reprint requests to : Dr Ronald F. Abercrombie, Department of Physiology, Emory University School of Medicine, Atlanta, GA 30322, USA. Received Revised Accepted
: 8 November 1989 : 8 December 1989 : 7 January 1990
