Properties of calcium binding by Myxicola axoplasmic protein

Cdl cakilan (lwo)1I, 45Q-407 CDLongman Group UK !_td1900

0143-4160/90/0011MS10.00

Properties of calcium binding by Myxicola axoplasmic protein N.F. AL-BALDAWI and RF. ABERCROMBIE

Department of Physiology, Emory Universi?v School of Medicine, Atlanta, Georgia, USA Abstract - The 45Ca2t bindlng properties of axoplasmic protein from the Myxicda giant axon have been investigated using a centriigallconcentration-dialysis technique. Scatchard plot analysis of these binding data suggestthat Ca2’ is attached to a site with an equilibrium dissociation constant of 7.7 f 0.5 uM and a capacity of 4.4 + 0.2 povg ax!plasmic protein (n = 11). Addition of other cations - C&+, Mn2+,A13+,Cu2+, Ba ‘, and Zn - at concentrations up to 10 fl did not displace 0.2 fl ‘%a2+ from lts binding site, probably because of buffering of these cations by amino acid residues within the protein solutions. The protein could be stored at 4°C for up to 16 days with no appreciable change in the number of calcium sites. Ca2+bindlng equilibriumtook place in less than 36 min of incubation. increasing the incubationtemperaturefrom 4°C to 37°C reduced the number of Ca2+ sites. The binding capacity was reduced by one-half when the protein was dialyzed with 4 M urea or high ionic strength KCI (2 M). Calcium binding was examined as a function of pH. When the protein was dialyzed overnightat different pH values and all the binding was done at pH 7.0, the apparent number of Ca2’ sites decreased as the pH of the dialysis medium was increased. When the protein was dialyzed overnightat pH 7.0 and the binding was done at different pH values, the apparent binding capacity increased as pH increased. When the protein was dialyzed overnightat the same pH at which the binding was done, the apparent binding capacity reached a maximum at pH 7.0. The histldine-specific reagent diethyl pyrocarbonate (DEP) reduced the apparent binding capacity. The ability of axoplasm from giant axons of the marine annelid Myxicola injiutdibulum to bind calcium was first described by Baker and Schlaepfer [l]. Recently [2-41, we have suggested that the neurofilaments of this preparation are capable of binding stoichiometric amounts of calcium and that neurofilaments may be responsible for more than half of the so-called ‘energy-independent’ calcium binding of the axoplasm [4]. Myxicoh neurofilaments are composed of two major proteins having molecular masses close to 150-160 kD [S,

61. Porcine neurofilaments am composed of a triplet of proteins having apparent molecular masses of 217, 173, and 73 kD [7] and am similar to most other mammalian neurofilaments. Lefebvre and Mushynski [8] have shown #at porcine neurofilaments bind calcium. As a basis for our study, we examined the effects ,of extraction media, incubation time, storage time, calcium concentration, and other cations on calcium binding. Using the effects of pH, temperature, ionic strength, and agents that modify

459

CELL CALCIUM

460

Table 1 Buffer solutions Bu@r 1

Contents (mh4) 4000 Urea, 1% 2-mercaptoethanol, 10 HEPES, 5 MgClz, pH 7.5

2

6 samples taken from 2 to 3 preparations. Each preparation of protein represents 10 to 15 worms.

500 KCl, 180 glycine, 116 cysteic acid, 75 D.L-asparticacid pH 7.5

3

300 KCl, 10 MES, 5 MgClz, pH 5

4

300 KCI, 10 PIPES, 5 MgC12, PH 6

5

300 KCl, 10 HEPD, 5 MgC12, PH 7-8

6

300 KCl, 10 Tris, 5 MgC12, pH 9

7

2000 KCl, 10 HEPES, 5 MgClz, pH 7.5

amino acid residues, we examined some of the Ca2’ binding properties of this (presumed) neurofihunent protein.

Analysis of data A plot of bound/free (B/F) versus bound (B) calcium was either fitted to a straight line (Scatchard plot analysis), assuming a single binding site to give an apparent affinity and capacity, or, using an assumption of two sites, a nonlinear fit of the data was made to determine an approximate affinity and capacity of each site. The Scatchard-type equation for a two site model was derived from the Michaelis-Menten equation B’ =

Preparation Myxicola were collected, maintained, and the axoplasmic gel was extracted aud solubilized as previously described 141. SDS-PAGE of the resulting extract revealed that the neurofilament protein represents more than 50% of the total extract. This protein had a ‘purity’ comparable to that of protein we prepared by sucrose gradient [6], gel filtration [9], or ion exchange chromatography [ 101. None of these methods yielded pure 150-160 kD protein. Binding experiments A centrifugal concentration dialysis procedure was used for examining Ca2’ attachment to the protein [4]. In the studies of putative effector substances, the test agent was added and mixed with the solution before the protein was added. Except where indicated, the data presented in this paper represent the mean value f the standard error of 4 to

+ B2/(F+K2)]

where Bl, B2, Kl, and K2 are the maximum capacities and equilibrium dissociation constants of the two sites, respectively; B and F am the bound and free concentrations of calcium, respectively. After solving the quadratic equation for F and using this in an expression for B/F: B/F

Materials and Methods

F[Bl/(F+Kl)

=

(2Ba)/{-b + d(b2 - 4 ac)}

wherea=B-Bl-B250 b = K2(B - Bl) + Kl(B - B2) IO andc=Kl K2B. The data were then fitted to this equation using a nonlinear least squares method to give the best values for the different parameters.

Results and Discussion Measurements of calcium activity in buffer solutions We found it necessary to perform the calcium binding experiments in the absence of EGTA chelators as calcium binding to EGTA would overwhelm our measurement of calcium binding to the protein. Several conditions are met to reliably adjust free calcium without chelators. First, calcium contamination was removed by passing the solutions through a Chelex 100 ion exchange column. Also, the buffer constituents used in this study must not bind appreciable calcium [ll]. Finally, calcium activity was measured to assure that the expected activity was achieved. This was done with calcium-

CALCIUMBINDING BY MYXKXXAAXOPLASMICPROTEIN Table 2 Measured calcium activity of buffers Added total calcium

Buffer 3w

10 w

pCa

acaW

PCs

QW

3(MEspH5)

6.07

0.85

5.62

2.4

4 (PIPES pH 6)

6.07

0.85

5.62

2.4

5 (HEPES pH 7)

6.07

0.85

5.62

2.4

5 (HBPES pH 8)

5.97

1.07

5.65

2.2

6 (Tris pH 9)

6.1

0.79

5.65

2.2

specific electrodes which were made, calibrated [12] and used to measure the calcium activity of the different buffer solutions containing 3 pM and 10 pM total added calcium. In this concentration range, the sensitivity of the electrodes was 28 mV per decade change. The results, shown in Table 2, demonstrate that the free calcium activity is 0.22 to 0.36 of the total calcium concentration, which is within the expected range of the calcium activity coefficient in the buffer solutions [ 133. Extraction media

Gilbert [14] has shown that Myxicola neurofilament gel structure can be dispersed without proteolysis by solutions containing 0.5 M KC1or 0.5 M urea. We have used this property to liquefy Myxicola axoplasm and study the calcium binding properties

461

of solutions of the 150-160 kD axoplasmic protein in vitro. Axoplasmic samples were dispersed with Buffer 2 (Table 1) for 1 h at 10°C. Following centrifugation, the solutions were allowed to dialyze for 24 h at 10°C in one of the following solutions: Disassembly (Buffer 2, Table 1); KCl, pH 7.5 (Buffer 5, Table 1); or 4 M urea (Buffer 1, Table 1). The binding assay was carried out in Buffer 5, pH 7.5 (Table 1) as described in Materials and Methods. Figure 1 shows that the number of Ca2’ binding sites in the protein was reduced by the Urea Buffer compared with that of Disassembly Buffer or 0.3 M KCl. Incubation time

Samples of axoplasmic protein in solutions of 0.3 M KC1 were incubated, as described in Materials and Methods, for 0 (no incubation period), 15, 30, or 60 min, with 0.2, 3.2, or 10.2 pM 45CaC12, to determine whether equilibrium binding was reached in the ,course of these measurements. After the incubation, samples were concentrated [4]; the concentration/dialysisprocedure takes about 20 min. Figure 2 shows that equilibration takes place almost immediately; however, for convenience, an equilibrationtime of 30 min was chosen. Storage

Samples of axoplasmic protein in solutions of 0.3 M KCl, prepared as in Materials and Methods, were

0

NORMALIZED

BOUND

CALCIUM

(pmol/g)

a

b

C

DISASSEMBLY

0.3 M KCI

4 M UREA

EXTRACTION

tj

MEDIA

Fig. 1 Extraction media. Axoplasmic samples were prepared as described in the text. The solutions were dialyzed in either (a) Disassembly Buffer (Buffer 2, Table 1) (filled circles); (b) KCl, pH 7.5 (Buffer 5, Table 1) (open circles); or (c) 4 M urea (Buffer 1, Table 1) (open triangles). a: B, - 4.7 f 0.3, b: B, - 3.9 + 0.1, c: B, - 2.1 * 0.3. (B, tits are pmol/g.)

CELL CALCIUM

“.”

1.U

NORMALIZED

2.0

3.0

BOUND CALCIUM

4.0

5.0

0

30

20

10

(pmol/g)

50

40

INCUBATION

60

70

TIME (min)

Fig. 2 Time of incubation. Samples were prepared BSdescribed in Materials ,pnd Methods. Ca2’ binding was carried out in Buffer 5, pH 7.5 (Table 1). Aliquots were incubated for different periods of time before the concentration/dialysis pmcess: A: 0 min (open triangle+ 15 min (filled circles); 30 min (open circles); and 60 min (filled uiaugles). The average capacity of the combined data was 3.6 f 0.4 panel/g B--

stored for up to 16 days at 4°C. Aliquots were taken after 0, 4, 8, 12 and 16 days of storage and incubated for 30 min as described in Materials and Methods. Although there were no appreciable differences in the capacity of the binding site(s) after these periods of storage (Fig. 3), we always used the protein immediately after preparation unless otherwise stated.

calcium concentrations used in these experiments, suggests a component of low affiiity or non-specific calcium binding that causes a small overestimation of Kin and Bmaxof the high-affiity binding. If the capacity of the high-affinity component is in the range of 2 to 5 pmopg, as was the case for our measurements, then this method gives a reliable estimate of the capacity of the high affinity binding site (Fig. 4). This method is not sufficiently

Ca2’ concentration

A wide range of Ca2’ concentrations were used to examine the calcium-binding elements in Myxicoia axoplasm. Figure 4 shows how the radioactivity was displaced from the protein; i.e., the ratio of bound to free calcium dropped as the free calcium was increased from 0.2 p.Mto 100 uM. A nonlinear least squares fit of all the data to a two-site model is shown as the curved line on the graph. The dashed straight line was drawn using only data with calcium concentrations up to 10.2 p.M. Using the two-site model, parameters of the high affinity site are Bmax = 2.4 -t 0.8 j.unoYg, Ktn = 8.7 f 2.7 pM. Parameters of the low affinity site are Bmax= 44 f 20 pmol/g, Kin = 1 mM. Using the single site model and calcium concentrations only up to 10.2 @!l, the parameters of the site are Bmax= 3.2 IL0.5 pmoYg, KIT = 9.5 + 2 @VI. The long tail of the Scatchard plot of Figure 4. revealed by the range of

‘;;; 2 3 z 2 3 I $ u g 2m

6 5 -,0--O 0 4 3_

\A I\$

2*

,

o

I 0

2

4

6 STORAGE

6

10 TIME

12

I

I

14

16

(day)

Fig. 3 ca2+ binding capacity at different periods of storage. Extracts from a single batch of protein were prepared as described in Materials and Methods and stored at 4’C. Aliquots weretested for their Ca”’ binding activity after 0. 4. 8, 12., and 16 days of storage. The average capacity of the combined data was B, - 4.0 f 0.4 pmovg

CALCIUM BINDING BY MYXICUU AXOPLASMIC PROTEIN

463

is small. We have, therefore, based our conclusions on large changes in Bmax which can be distinguished by this procedure, These Scatchard plots are consistent with there being one main class of high affinity calcium-binding sites with an equilibrium dissociation constant of 3 to 9 pM and a capacity of 2.5 to 4.0 pmol/g protein. The large capacity low affinity (-1 mM) binding may represent attachment of calcium to acidic amino acid residues within the protein. Metal ions 0

2

4

NORMALIZED

6

BOUND

8

10

CALCIUM

12

(pmol/g)

Fig. 4 Scatchard plot for a wide range of calcium concentrations: 0.2, 1.2, 2.2, 3.2, 5.2, 10.2, 30.2, 60.2, and 100.2 pM Ca”. The curved line assumes two sites: LX - 2.4, Km - 8.7 for the high affinity site; and B, - 44, Ktn - 1000 for the low affiiky site. The dotted line assumes a single site with Bmx - 3.2 and KM - 9.5. (B, units are pmol/g, KIQ units are pM.)

z

0

2 METAL IiN

6 CONCENTRATION

6 (j4)

10

Fig. 5 Metal ion competition with Ca2+. The ability of six multivalent metal cations to compete with Ca2+ was examined. When added as the Cl- salt Cd”, I&” A13’, Cu”, Ba2’, or Zn2+ at concentrations up to’ 10 pM did dot change the ratio of bound/free 4sCaz+ (0.2 @I %a2+). The binding mixtures were incubated for 30 min at 4°C as described in the text

accurate for determining small changes in KID or Bmax,nor is it adequate for determining an accurate value of Kin if the high affinity binding component

In an attempt to examine the selectivity of this binding site(s) for calcium over other cations, a series of experiments was done substituting Cd2+, Mu2+ A13+ Cu2’ Ba2’, or Zn2+ for Ca2+ previously Hhown’thatMg2+ in high concen~tk~ (5 mM) does not compete with Ca ’ for the site [4]. Buffer’ 5 (Table 1) was made up with 0.2 pM 45Ca2tand either 3 or 10 pM of the chloride salt of each of the cations listed above. The ratio of bound over tree calcium dropped only when the calcium was increased (Fig. 5). Ca2’, Cd2’ and Mn2’ all have similar crystal ionic radii; therefore, it seems improbable that neither Cd2’ nor Mn2+could substitute for Ca2’ on this site. In fact, it was previously shown, using different procedures, that Cd2’ and Mn2’ were able to displace 45Ca2+ from binding sites in intact samples of axoplasm [15]. Other reported evidence shows that the 200 kD humau and bovine neurofilament ~rotein.~ are metal-binding proteins which bind Al , Cu , and 2n2+ as well as Ca2’ [161. We assume, therefore, that in our experiments these ions failed to displace calcium from its site for technical reasons. Many of these ions including Zn2+, A13+,Cu2+, and Cd2+ have a much higher affinity for amino acids than does calcium [17]. Our interpretation of these results is that in our unbuffered solutions, these ions are effectively buffered by the amino acid residues of the axoplasmic protein. In other words, these other ions have a relatively high affinity for nearly all proteins, and the buffering action of the protein mixtures masks any effect that the ions might have on the calcium site.

CELL CALCIUM

464

Ionic strengthand temperature

BINDING KC1 (M)

Ca2’ binding was tested at low (0.3 M) and high (2 M) ionic strengths of KCl. Two types of effects of high or low ionic strength were examined: those occurring during the binding equilibration and those occurring during a period of dialysis before the Ca2’ binding had taken place. After extraction, as described in Materials and Methods, the protein was divided into four fractions and each treated as follows: The first fraction was dialyzed for 48 h in 0.3 M KCl, pH 7.5 (Buffer 5, Table 1) and the binding experiment done with 0.3 M @H 7.5) and 2 M KCl buffers (Buffers 5 and 7, Table 1). The second traction was dialyzed for 48 h with 2 M KC1 (Buffer 7, Table 1) and the binding carried out in Buffers 5 @H 7.5) and 7 (Table 1). The third fraction was dialyzed for 48 h in 2 M KC1followed by dialysis in 0.3 M KC1@H 7.5) for 48 h (Buffers 7 and 5, Table l), then the binding was tested with Buffers 5 @H 7.5) and 7. The fourth fraction was dialyzed for 48 h in 0.3 M KC1 (pH 7.5) followed by dialysis in 2 M KC1 for 48 h, then the binding was carried out in Buffers 5 (pH 7.5) and 7 (Table 1). Prolonged dialysis of the protein for up to 144 h in Buffers 5 @H 7.5) or 7 (Table 1) did not alter the results. When the protein was dialyzed in high ionic strength KC1 (2 M), binding capacity was reduced by 25-30% compared with matched controls in which the protein was dialyzed in low ionic strength

1.o NORMALIZED

0.0

7 examined

Incubation

temperature

afIer 30 min

concentration/dialysis

2.0 3.0 BOUND CALCIUM

dependence.

incubation

4.0 (pmol/g)

Protein

6

Ca”

of protein

KCI (M)

binding at different ionic strength of KCl. were divided

into four fractions

The third fraction (E and F). The fourth fraction B,

- 4.1 rt 0.1; B: B,

D: BG:

B,,

-1.3fO.l;E:

- 2.09 f 0.05; C

in the

- 2.7 f 0.4; I-k B-

(G and H). A:

B-

B--4.4fO.Z.F:

- 3.2 f 0.3;

B,,ux-1.4iO.l;

- 1.4 + 0.3.

(B,

units are

PmoVg.)

KC1 (0.3 M) (Fig. 6). When the protein was dialyzed at high ionic strength KCl, followed by low ionic strength overnight dialysis, most of the binding capacity was regained, suggesting that the effect of high ionic strength KC1 on calcium binding is slowly reversible. When the protein was dialyzed with low ionic strength KC1followed by high ionic strength KCl, the binding was lower compared with

4

37

TEMPERATURE

were prepared

as in Materials

(B-

units are poVg.)

( %)

and Methods

Bmar - 4.7 zk 0.2, or 37°C (filled circles),

was carried out at the indicated temperature.

Extracts

as described

text: The first fraction (A and B). The second fraction (C and D).

5.0

aliquots

at 4°C (open circles),

DIALYSIS Fig.

and their Ca” &

binding

- 2.4 f 0.2.

The

CALCIUh4

BINDING BY hfYXKX%A AXOPLASMIC

PROTEIN

the matched controls of high ionic strength followed by low ionic strength KCl. Thus, the calcium-binding capacity depended not only on the KC1 concentration of the binding media, but also on the KC1 concentration of the media with which the protein had been previously dialyzed. One effect of increasing the ionic strength of the binding media is to alter the ionic environment of calcium ions in solution as well as the charged groups on the protein with which calcium interacts. Such effects would be expected to occur quickly after the ionic strength is changed, to aher the affinity rather than the capacity of binding, and to be completely and immediately reversible. However, high concentrations of KC1 also influence the interaction of neurofilament subunits with one another, as seen when neurofilaments change from a gel to a liquid state at high strength KCl. Our results may be viewed as a reversible denaturation of the Ca2+-binding site by high ionic strength. Reversing this ‘denaturation’ of the protein appears to require some time (Fig. 6). Calcium binding is easily detectable at low temperature (4”C), which is near the ambient temperature in the Bay of Fundy where the animals are collected. Increasing the temperature to 37°C reduced the number of Ca2+ sites by one-half (Fig. 7). Apparently, at higher temperatures, the thermal motion of the protein changes the protein conformation in such a way that a fraction of the sites no longer bind calcium.

465

\ \ \

’

‘.6

..

3

b

b - (y I

Fig. 8 pH-dependence

/

b,

7’.

1

l

of Ca”

binding.

7A

Extmcts were dialyzed

with Buffers of pH range 5.0 to 9.0 and the Ca2’ binding

was done in Buffer 5, pH 7.0; [B]: with Buffer 5, pH 7.0 and the Ca2+ binding

Calcium binding to axoplasmic protein was examined over a pH range of 5.0 to 9.0. MISS (10 mM) was used for pH 5.0 (Buffer 3, Table 1); PIPES (10 mM), for pH 6.0 (Buffer 4, Table 1); HEPES (10 n&I), for pH 7.0 to 8.0 (Buffer 5, Table 1); and Tris (10 mM), for pH 9.0 (Buffer 6, Table 1). Our results (Fig. 8) showed that binding of calcium to the axoplasmic protein was influenced by pH; however, the specific results depended on whether the pH was varied in the dialysis media, the binding media, or both. When the overnight dialysis was carried out at different pH and the calcium binding ail done at pH 7.0, the calcium capacity decreased as the pH of the

0

.\

\0'

C

[A]:

Effect of pH

A

6‘\

with btiers

was done in buffers of pH range 5.0 to 9.0; [Cl: of pH range 5.0 to 9.0 and the Ca2’ binding

done in buffers

of pH range

plotted is the maximum plots. Bl&c:

5.0 to 9.0.

capacity

At pH 5, A: B,

hm

Bmax-3.9f1.4;C: f 0.5; B: B,x

2.6 rt 0.6. B,

Bm, - 3.2 rt 0.7; B: B,, (Bman units are mol/g.)

- 1.3 f 0.4; c:

- 4.1 + 0.3; B: B,,

At pH 7, A: B,

- 5.5 f

&nu - 6.3 f 0.7: B:

- 6.2 f 0.7. At pH 8, A: B,

= 5.0 f 1.8; c:

was

calcium

a linear fit to Scatchard

- 10.3 f 0.7; B: B,

- 2.6 f 1. At pH 6, A: B, Bm,=

The bound

= 4.6

- 3.7 f 1.4. At pH 9, A:

- 5.8 f 1.6; c:

Bnnx - 2.8 f 0.7.

dialysis medium was raised from 5.0 to 9.0 (Fig. 8A). (Higher pH of the dialysis solution resulted in lower calcium capacity.) However, when the overnight dialysis for all the samples occurred at pH

CJZL CALCIUM

466

7.0 and the calcium binding was done at different pH, the binding capacity was lower at pH 5.0 but relatively unaffected by pH over the range from 6.0 to 9.0 (Fig. 8B). (Lower pH resulted in lower calcium capacity.) When the protein was dialyzed overnight at different pH and the binding carried out at the same dialysis pH, a maximum near pH 7.0 was seen in the calcium capacity (Fig. 8C). The pH-dependence of calcium binding (Fig. 8) suggests that the effects of pH am of two hinds. One effect of pH, i.e., that occurring in the acidic range between 5.0 and 7.0, is seen only when the pH of the binding buffer is varied. This could be the direct result of the protonation of ionizable groups on the calcium binding protein. An amino acid residue with a pK near 6.0 is histidine, and the possibility that a histidine residue(s) may be near the calcium binding site is discussed below. A second pH effect, i.e., that occurring in the alkaline range, was seen when the protein was dialyzed overnight in an alkaline solution. This effect is likely to be more indirect and more complex. Possibilities ate that high pH affects protein-protein interactions and/or, like high ionic strength, denatures the Ca2+-binding site of the protein. Inhibitors

DEP suggest that histidine residues am part of, or near, the Ca2+-bindingsite(s) of the protein. W7 and W5 am similar compounds; however, W7 has a chloride substituted on the aromatic ring, and it inhibits calmodulin-dependent processes at lower concentrations. Cahnodulin is 90% inhibited by 10m4M W7 but only 20% inhibited by lOA M W5 [19]. W7 and W5 both had only a small effect in reducing calcium binding to MyxicoZaaxoplasmic protein (Fig. 9).

Conclusions

We conclude that Myxicola axoplasmic protein behaves as though there is one major class of calcium-binding sites, although there may be other sites with both higher and lower affinities for calcium. This apparent site is affected by pH and ionic strength; and the binding is reduced by histidine- but not by sulfhydryl-specific reagents. It is interesting to note that the published amino acid sequence of the carboxy-terminal (tail domain b) of porcine spinal cord neurofilament protein [20] contains a single histidine residue, no cysteine residues, and 47/106 glutamic acid residues. We suggest that the effects of temperature, high pH, and high ionic strength KC1 or urea reflect the

A series of substances which are considered to modify proteins, inhibit the activity of proteins, or influence calcium metabolism were examined for m their effects on calcium binding to Myxicola axoplasmic protein. These substances include diethyl pyrocarbonate (DEP, 10m3,10e4, low5 M); p-chloromercuribenzyl-sulfonate (pCMBS, lOA, $ 109, lO+j M).g-hydroxymercuribenzoate @HMB, 9 0.2 10”‘. lo-‘, 10 M); N-(6-aminohexyl)-5-chloro-lnaphthalene sulfonamide (W7, 10V3,10m4,10m5M); oo I I and N-(6-aminohex{l)-1-naphtbalene sulfonamide i.OE-7 1.OE-4 1 .OE-6 1 .OE-5 1.OE-3 1 .OE-2 (W5, 10-3, lOA. lo- M). INHIBITOR CONCENTRATION (M) The effects of DEP. W5, W7, pCMBS and pHMB are plotted (Fig. 9). suggesting that DEP had a greater effect than W5 or W7. pCMBS and Fig. 9 Dose dependence of inhibition by DEP (filled circles), W5 (open circles), W7 (open squares), pCMBS (filled squares) pHMB had no reproducible effect. and pHMB (open triangles). Samples of protein were incubated Diethyl pyrocarbonate (DEP) has been shown to in binding mixtures containing increasing concentrations of the react with his&line residues [18]. The effect of inhibitors. The bound calcium plot-ted is the maximum capacity acidic pH on Ca2’ bind@ and the inhibition by from linear ctig to Scatchard plots

467

CALCIUM BINDING BY MYXICOLAAXOPLASMIC PROTEIN

flexibility or ‘softness’ of the protein near the These relatively mild calcium binding site. conditions are apparently able to denature the Ca2’ binding region of the protein. Speculation about the physiological significance of this Ca2’ binding site requires au accurate value for the calcium dissociation constant. It is probable that we have overestimated the dissociation constant (underestimated the affinity) of this site for technical reasons having to do with the presence of low affinity calcium binding sites (see above). The dissociation constant is, nonetheless, probably higher than the resting free calcium concentration in the axon [4, 121. The physiological role of this site could be to help buffer calcium during periods of elevated calcium or to confine changes in free calcium to local regions within the neuron.

Acknowledgements We thank Dr Otto Froehlich for valuable comments; MS Janice Abemrombie for carefully reading the manuscript. This work was supported by MH NS 19194.

References 1. Baker PF. Schlaepfer WW. (1978) Uptake and binding of calcium by axoplasm isolated from giant axons of Loligo and Myxicola. J. Physiol., 276, 103-125. 2. Abercmmbie RF. Gammeltoft K. Jackson J. Young L. (1986) An inttacellular calcium binding site on neurofilament proteins of Myxicola giant axon. J. Gen. Physiol., 88, 9a 3. Abercrombie RF. Gammeltoft K. Young L. (1986) Ca-binding site on a neurofilament protein of the Myxicola giant axon. Biophys. J., 49, 116a. 4. Abetcrombie RF. Al-Baldawi NF. Jackson J. (1990) Calcium binding by Myxicoh axoplasmic protein. Cell Calcium, 11,361-370. 5. Gilbert DS. Newby BJ. Anderton BH. (1975) Nemofilament disguise, destruction and discipline. Nature, 256,586-589. 6. Lasek RJ. Krishnan N. Kaiserman-Abramof IR. (1979) Identification of the subunit proteins of IO-nm neurofilaments isolated from axoplasm of squid and

Myxiwla giantaxons.J. Cell Biol., 82,336-346. 7. Shaw G. Debus E. Weber K. (1984) The immunological relatedness of neurofilament proteins of higher vertebrates. Eur. J. Cell Biol., 34, 130-138. 8. Lefebvre S. Mushy&i WE. (1987) Calcium binding to untreated and dephosphotylated porcine neurofilaments. B&hem. Biophys. Res. Commun., 145,1006-1011. 9. Chiu F. Norton W. (1982) Bulk preparation of CNS cytoskeleton and the separation of individual neurofilament proteins by gel filtration: Dye-binding chatacteristics and amino acid compositions. J. Nemo&em., 39, 1252-1260. 10. Karlsson J. Rosengren L. Haglid K. (1987) A rapid HPLC method to separate the triplet proteins of nemofilament. J. Neurochem., 49, 1375-1378. 11. Good NE. Winget GD. Winter W. Connolly TN. Izawa S. Singh RM. (1966) Hydrogen ion buffets for biological research. Biochemistry, 5, 467-477. 12. Abercmmbie RF. Gammeltofi K. (1987) High cytosolic pH inhibits Ca uptake by Myxicole axon mitochondria. Am. J. Physiol, 252 (Cell Physiology 21), C68C76. 13. Thomas MV. (1982) Techniques in Calcium Research. London, Academic Press. 14. Gilbert DS. (1975) Axoplasm chemical composition in Myxicola and solubility properties of its structural proteins. J. Physiol., 253,303-319. 15. Aberctombie RF. Masukawa LM. &din RA. Livermood D. (1981) Uptake and release of 45Ca by Myxicola axoilasm. J. Gen. Physiol., 78, 413-429. 16. Pierson K. Evenson M. (1988) 200 KD neurorilament protein binds Al, Cu and Zn. Biochem. Biophys. Res. Commun., 152, 598-604. 17. Martell AE. Smith RM. (1976) Critical Stability Constants. Vol. 1. New York, Plenum Press. 18. Tsurushiin S. Hiramatsu A. Inamasu M. Yasunobu K. (1975) The essential histidine residues of bovine plasma amino oxidase. Biochim. Biophys. Acta, 400,451-460. 19. Hidaka H. Sasaki Y. Tanaka T. et al. (1981) N-(6-aminohexyl)-5-chloro-l-naphthalene-sulfonamide, a calmodulin antagonist, inhibits cell proliferation. Pxoc. Nail. Acad. Sci. USA, 78,43564357. 20. Geisler N. Kaufmann E. Fischer S. Plessman U. Weber K. (1983) Neurofilament architecture combines structural principles of intetmediate filaments with cwboxy-terminal extensions increasing in size between triplet proteins. EMBO J., 2, 1295-1302.

Received : 10 April 1990 Revised : 4 June 1990 Accepted : 29 June 1990 Please send reprint requests to : Dr Ronald F. Abercrombie, Department of Physiology, Emory University School of Medicine, Atlanta, GA 30322, USA