cerebral cortex slices by a-latrotoxin from black widow spider venom by cometjunkie57


									Proc. Nati. Acad. Sci. USA Vol. 75, No. 8, pp.. 4016-4020, August 1978


Release of neurotransmitters and depletion of synaptic vesicles in cerebral cortex slices by a-latrotoxin from black widow spider venom*
('y-aminobutyric acid/brain cortex/neurotoxin) MU-CHIN TZENG, ROCHELLE S. COHENt, AND PHILIP SIEKEVITZ
The Rockefeller University, New York, New York 10021

Contributed by Philip Siekevitz, May 3, 1978

ABSTRACT The effect of a-latrotoxin on cerebral cortex slices was studied by both biochemical and morphological methods. This toxin greatly stimulates the release of preloaded -y-amino[3H butyric acid from cortex slices. The response increases linearly with dose. The release is not dependent on the presence of extracellular Ca2+, and therefore it is not mediated by the release of other transmitters from other types of neurons. In contrast, no significant increase in the release of a nontransmitter substance a-amino[14Clisobutyric acid is observed. Since previously we have shown that a-latrotoxin stimulated the re ease of acetylcholine and norepinephrine from cortex slices, it appears that the toxin probably selectively releases all neurotransmitters. The toxin also profoundly depletes the synaptic vesicle population in boutons in the cortex slices. The results suggest that the release of neurotransmitter and the depletion of synaptic vesicle in boutons are manifestations of a single action of the toxin. Therefore, a-latrotoxin can be used as a good tool for the identification of neurotransmitters and in studies on the mechanism of neurotransmitter release.
Previous work (1) has described the fractionation of an extract of the black widow spider venom gland (BWSV) into several toxic protein fractions. One of the fractions was purified to the degree of no detectable contaminating proteins and was recently named a-latrotoxin (2). This single toxin factor was demonstrated to be responsible for all the effects of BWSV on frog and mouse neuromuscular junctions, namely, the increase in the frequency of miniature end-plate potentials, the complete depletion of synaptic vesicles in the nerve terminals (boutons), and the ultimate blockage of neuromuscular transmission (1). Since BWSV has many actions on both vertebrates and invertebrates, particularly their nervous systems (reviewed in ref. 1), and since several toxic fractions have been found in the venom, it is of interest to investigate the spectrum of action of this highly purified a-latrotoxin in a single species. In a previous communication (2) we demonstrated that a-latrotoxin caused an increase in the release of both acetylcholine and norepinephrine (NE) from cerebral cortex slices of mouse. We report here: (i) that a-latrotoxin enhances the release of another putative neurotransmitter, -y-aminobutyric acid (GABA), from mouse cortical slices, but not the release of a non-transmitter substance, a-aminoisobutyric acid (AIB, an analogue of GABA); and (ii) that, as in the case of neuromuscular junction, it also causes depletion of synaptic vesicles in cerebral cortex slices.
MATERIALS AND METHODS The purification and assay of purity of a-latrotoxin were performed as described (1); occasionally another step of Sephadex
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

G-200 chromatography was carried out after the DEAESephadex step to ensure homogeneity of the toxin. For each experiment, one male mouse (25-30 g) was killed by decapitation and its brain quickly removed. One slice was taken with a blade and blade guide from the top layer of cerebral cortex of each of the two hemispheres. The two slices (30-40 mg) were first incubated for 30 min at 370 in 1.0 ml of a modified Krebs-Ringer solution with the following composition: 120 mM NaCl, 4 mM KC1, 1.8 mM CaCl2, 0.8 mM MgSO4, 4 mM Na2HPO4/HCl buffer (pH 7.4), 15 mM NaHCO&, and 10 mM glucose. The incubation solution was equilibrated with a gas mixture of 95% 02/5% CO2 throughout the experiment. In experiments assaying for GABA release, 0.3 MAM [3H]GABA (35 Ci/mmol, New England Nuclear) was present in the incubation solution, and 0.5 mM (aminooxy)acetic acid (Eastman Co.) was used to inhibit the metabolism of GABA by GABA-glutamate transaminase (3). In the experiments with [14C]AIB, a greater concentration, 35 AM, was used because of the lower specific radioactivity (57 mCi/mmol, Amersham/Searle). After 30 min incubation, the slices were removed and washed once by immersion for 10 min in fresh solution. They were then transferred to a perfusion vessel containing 1.5 ml of solution. The vessel was immersed in a shaker maintained at 370. Fresh oxygenated modified KrebsRinger solution was perfused through the vessel at a rate of 0.5 ml/min by a peristaltic pump. The perfusate of the first 30 min was discarded, and thereafter 1-ml fractions were collected every 2 min into vials in a fraction collector. At a suitable time during the perfusion, a-latrotoxin was injected into the perfusion chamber. Radioactivity discharged into the perfusate was determined after the addition of Aquasol, and radioactivity remaining in the brain slices was determined after solubilization of the slices with Protosol. In some experiments, Ca2+ was omitted, 0.5 mM EGTA was present, and the MgSO4 concentration was increased to 3.0 mM. In order to take into account the variations among the various experiments in the uptake of the radioactive substances and to correct for the diminution of radioactivity in the tissue during the perfusion, we expressed the radioactivity in the perfusate at any given time as a percentage of the radioactivity still present in the tissue at that time (4). We also assayed, by a procedure modified from that of Nadler and Cooper (5), aliquots from some fractions to deterAbbreviations: AIB, a-aminoisobutyric acid; BWSV, black widow spider venom gland extract; NE, norepinephrine; GABA, 'y-aminobutyric acid. * This work was presented in abstract form by M. Tzeng and P. Siekovitz at the Seventh Annual Meeting of the Neuroscience Society, 1977. t Present address: Department of Anatomy, University of Illinois Medical School, Chicago, IL.


Neurobiology: Tzeng et al.
mine the degree of metabolism of [3H]GABA. The perfusates were first chromatographed through an AG 50W (BisRad) (Na+ form, pH 10) column. Neutral and acidic com s were eluted with three 1-ml H20 washes, and basic compounds were then washed out with 2 M NaOH. The H20 eluates were combined and acidified to pH < 2, and then further fractionated on an AG SOW column (NH4+ form, pH 2.5). The column was eluted in sequence by two 1-ml washes of each of the folGABA C-on

Proc. Natl. Acad. Sci. USA 75 (1978)


a.I 0

diffusion problems.

NaOH and then analyzed for GABA as described above. For studies by electron microscopy, brain slices were cut into small strips and then incubated for 1 hr at 370 in 1.0 ml of the modified Krebs-Ringer solution with 10 gg a-latrotoxin added at zero time and another 5 ,ug added at 30 min. Control samples were handled in the same manner except that no toxin was added. The strips were then fixed with 1% glutaraldehyde/1% paraformaldehyde in 0.12 M phosphate buffer, pH 7.2, and postfixed with 1% OS04 in 0.03 M barbital buffer, pH 7.4. The tissue was then stained en bloc with 0.5% uranyl acetate, dehydrated by standard procedures, and embedded in Epon, which was polymerized at 15.50 for 3 days. Thin sections were cut with a Porter-Blum MT2B microtome, stained sequentially with 8% uranyl acetate and 4% lead citrate, and examined with both Hitatchi HU-11B and Siemens-Elmiskop 101 electron microscopes. Only the outermost 20-,um layer of the strips was used for quantitative analysis, because the inner part of the tissue slices might not have been exposed to the toxin due to

M HN40H fractions. In some experiments the amount of 3H in the tissue that remained as GABA was also determined. To do this, the brain slices were homogenized in 10% (wt/vol) trichloroacetic acid and then centrifuged in a microcentrifuge for 10 min. The supernatant was treated twice with two volumes of ethyl ether, the water phase was neutralized with

lowing: H20, 0.3 M NH40H, 0.6 M NH40H, 1.0 M NH4OH, 1.5 M NH4OH, and 2.0 M NH40H (all solutions of NH40H having pH of 2.5). Al the GABA appeared in the 1.0 M and 1.5

Oi .+-O





Time, min

FIG. 1. Effect of a-latrotoxin on the release of [3H]GABA and [14C]AIB from cortical slices. The amount of radioactivity released in 2 min is expressed as a percentage of the radioactivity still in the tissue at that time. The initial amount of radioactivity in the tissue at zero time was 2.97 X 106 cpm for [3H]GABA and 1.04 X 10r cpm for [14C]AIB. The arrow indicates the time at which 10 jug of a-latrotoxin was added. (Aminooxy)acetic acid was present in both cases.

RESULTS In the various experiments, 30-40% of the [3H]GABA was taken up by the cortical slices. A typical pattern of efflux of radioactivity from these slices is illustrated in Fig. 1. The basal release rate of tritiated compounds in the absence of toxin was very slight, less than 0.1%/min. After a single application of 10 i~g of a-latrotoxin, the release rate began to increase in less than 3 min and reached a maximum of about 17 fold (average 16.2 + 1 fold in five experiments) at about 8 min, and then it gradually returned to the base line over a 30-min interval. At the end, about 85% of the radioactivity originally taken up was still in the tissue. When the perfusates were assayed by ion-exchange chromatography, it was found that about 95% of the radioactivity released into the medium and about 97% of the radioactivity remaining in the tissue at the end of the experiment was in GABA. In those experiments in which a Ca2+-free perfusion solution containing EGTA was used, 10 ,g of a-latrotoxin elicited a similar increase in the release of GABA (maximum 14.8 fold in two experiments). When another protein fraction (E) from BWSV (cf. ref. 1) was tested, no stimulation on the release of GABA was found. This finding is consistent with our proposal (2) that a-latrotoxin is perhaps the only component in BWSV that is active towards vertebrates. The 85% of [3H]GABA that was still retained in the tissue at the end of 1-hr exposure to a-latrotoxin is similar to the case of NE (2) but different from that observed with acetylcholine (2). Possible reasons for this high retention of NE and GABA are: these experiments were done by a single short-pulse application

of toxin to the perfused tissue, and the toxin in the solution was washed away quickly; therefore those boutons in the interior of the slices might not have been exposed to the toxin. However, the acetylcholine experiments were done in an incubation vessel without perfusion. Also, there are rapid uptake mechanisms for both NE and GABA, not only by boutons, but also, in the case of GABA, by glial elements (6), and this latter pool presumably would not be available for release by a-latrotoxin. Fig. 2 shows the result of an experiment in which increasing doses of a-latrotoxin were applied to one brain-slice preparation. The application of 1 ,ug of a-latrotoxin increased the efflux rate to a maximum of 2.4 fold, and increasing the dose of alatrotoxin caused an increasing response in the release of GABA. However, 10 jg of a-latrotoxin applied in this experiment increased the release rate to a maximum of only 10 fold, which is significantly less than the 17-fold increase that was observed (Fig. 1) when the same amount of toxin was applied to a tissue preparation not previously exposed to the toxin. This phenomenon is most probably a consequence of irreversible vesicle depletion by earlier stimuli.


0 v


o 0.2-



t 9g 30 40


tl0 g t4Mg 60 70 80 90 100 110 120 Time, min

FIG. 2. Effect of increasing doses of a-latrotoxin (arrows) on the release of [3HJGABA from a brain-slice preparation. Radioactivity released is expressed as given in the legend to Fig. 1.


Neurobiology: Tzeng et al.

Proc. Natl. Acad. Sci. USA 75 (1978)

A dose-response relationship of GABA efflux by a-latrotoxin is shown in Fig. 3. The data were all from experiments in which only one addition of toxin was given to the slices. The threshold dose obtained from the graph was about 0.2 ,g. The response was fairly linear up to the largest dose (10 Mug) of a-latrotoxin used. This linear dose-response relationship is in contrast to the nonlinear relationship observed electrophysiologically (1). The reason perhaps is that the electrophysiological method measures the response at one synapse, whereas the present method measures the summation of the responses at many synapses. Also, the variation among the present experiments, generally within 20%, is less than that obtained by the electrophysiological method, which had variations as large as 200-300% (1). A comparison of the release of a known nontransmitter amino acid, AIB, with that of GABA is also shown in Fig. 1. AIB has been shown to be taken up by both neurons and glia (6). About 20% of the [14C]AIB was taken up by the tissue in our experiments. When 10 Mug of a-latrotoxin was applied, only a slight increase (average 10% in three experiments) in the efflux of AIB occurred. This small increase has doubtful significance, for GABA efflux was increased 16-fold by the same amount of toxin. In the case of AIB efflux, the presence or absence of (aminooxy)acetic acid did not make any difference in the results. After a 1-hr continuous exposure to a-latrotoxin, brain slices looked swollen even to the naked eye. Electron micrographs (Fig. 4) showed that most terminals were indeed swollen, with swollen mitochondria. Some showed in the plasmalemma localized discontinuities, which may have been caused by the swelling induced by the toxin (see Discussion). It should be pointed out that the disrupted appearance of even the control tissue (Fig. 4a) is probably due to the shaking during the experiment. Many boutons with synaptic vesicles inside were prominent in the control samples, whereas in toxin-treated tissue, vesicle-containing elements were almost completely absent (Fig. 4b). A quantitative analysis of the relative population of synaptic vesicles in boutons of control and toxin-treated slices was attempted. Boutons with identifiable synaptic complexes were counted and were classified as either full, partially depleted, or completely depleted of vesicles. Examples of each

of these three classes are shown in Fig. 4; Table 1 shows the distribution of such boutons among the three categories. Most boutons (87%) in control tissue were full of vesicles, only 5% were judged to be completely depleted, and these depletions could have been caused by tissue damage. In toxin-treated slices, over 70% of the synaptic boutons were completely depleted. Treated tissue appeared to have approximately 4 times less synaptic boutons per area than control tissue. This may be explained by the enlargement of the boutons due to possible incorporation of vesicle membrane into synaptic plasma membrane and by the swelling induced by the toxin. Based on this observation, a correction factor of four was used in a part of Table 1 (see below). In many cases synaptic complexes were not observed in obvious boutons, which were identified by the presence of synaptic vesicles. Of all the control boutons observed, -%0% were without synaptic complexes. This result may be explained by the orientation of sectioning, but also by the actual absence of these specializations in certain types of boutons. Recent work (7, 8) has shown that only 5% of the aminergic boutons exhibit synaptic complexes. Since these types of boutons cannot be distinguished from other elements of the neurites when they are depleted of vesicles, a different strategy was used for analysis. The micrographs were screened for obvious vesiclecontaining boutons that also lacked observable synaptic connections, and the data are also given in Table 1. In control experiments, in an area of 3500 ,um2, about 380 boutons without synaptic complexes were found to contain full complements of vesicles, while in toxin-treated tissue, only 16 such boutons were found in 6400 Am2. When the number per um2 of boutons full of vesicles was compared after a 4-fold correction had been made in the toxin-treated sample to take into account the swelling of the tissue due to the toxin, there was a 1:11 decrease by a-latrotoxin treatment. Therefore, it appeared that a-latrotoxin caused a complete depletion of synaptic vesicles in almost all types of boutons in the cerebral cortex of the mouse, regardless of the transmitters involved. DISCUSSION Previously (2) we have shown that a-latrotoxin caused increases in the release of acetylcholine and NE from mouse cerebral cortex slices, thus demonstrating that the effect of a-latrotoxin is not limited to cholinergic neuromuscular junctions. In this report, a-latrotoxin was shown also to enhance greatly the release of another putative transmitter, GABA, from cerebral cortex, but not of a nontramsmitter, AIB. This effect on GABA release was independent of extracellular Ca2+, similar to the effect on acetylcholine and NE release. The absence of a Ca2+-dependence indicates that the enhancement of GABA release is not mediated trans-synaptically by other transmitters released from other neurons by the toxin. While this paper was in preparation, a communication (9) appeared on the release of GABA from synaptosomes by a less well characterized fraction from BWSV. Our morphological studies (2) extended the observation at the neuromuscular junction that a-latrotoxin depletes synaptic vesicles similarly in the cerebral cortex. Boutons with or without synaptic complexes were affected to a similar extent, apparently regardless of the transmitters involved. These results strongly suggest that a-latrotoxin affects all types of boutons via a common mode of action, probably by inducing the fusion of vesicles with the presynaptic membrane to release vesicle contents and thus lead to eventual vesicle depletion. Previous investigators failed to find depletion of vesicles in synaptosomes prepared from cerebral cortex (10), thalamus and basal ganglia











ot-Latrotoxin, ,ug

FIG. 3. Dose-response relationship for the increase in [3H]GABA efflux by a-latrotoxin. The data are taken from experiments such as given in Fig. 1 but with varying doses. Trhe level of GABA release increased by va-larotoxin stimidAtionnat.the-peak of the efflux curve was expressed in terms of multiples of the basal rate. The line was fitted by the least mean square method.


iZ-~ \d4I<5~ ~ ~ ~',
Neurobiology: Tzeng et al.

Proc. Natl. Acad. Sc. USA 75(1978)



:K -X -Zt



1ww}4 aciE~~J's ff'4,

/4 r'/<' * >( ^W ''t,>'J


3 k k- ' (~~~~~Ai



U4*.' ;



- ¢, :n- wn 4 ,,t,



s' v/8is' Z '4

v"-G b




' Kt -'u






< t- ) i§ KL}t e
*.- > ..,. *¾



t>-~~~~~~~~~~~~~~~. -d --A,-ri. :stN e
4n9%;_/ gi , #f, r.') i

(a,b)1 Double arrows point to boutons having distinct synaptic complexes (single arrow). Arrow heads point to boutons lacking recognizable synaptic complexes. (X14,250.) (c) Higher magnification view of some boutons with full complement of synaptic vesicles (v). (X28,500.), (d)
A partially depleted synaptic bouton. (x28,500.) (e) A synaptic bouton completely depleted of synaptic vesicles. The synaptic projections seem to be denser than usual. (x28,500.)

FIG. 4. Electron micrographs of brain slices incubated for 1 hr in modified Krebs-Ringer solution with (b, d, e) or without (a, c) a-latrotoxin.

(11), and Torpedo electric organ (12) treated with BWSV. The reason for this discrepancy between our work and these reports is not known. If our results are substantiated, a-latrotoxin could be used as a tool first to determine whether a substance is stored in vesicles, and second, if the hypothesis of transmitter storage in vesicles

is correct, to aid the identification of transmitters. One important criterion for a substance to qualify as neurotransmitter is that it should be released by direct nerve stimulation in a Ca2+-dependent manner. But in many systems, it is not possible to isolate the nerve for stimulation, and, as a substitute method, indiscriminate electrical stimulation or high concentrations of


Neurobiology: Tzeng et al.

Proc. Natl. Acad. Sci. USA 75 (1978)

Table 1. Effect of a-latrotoxin on the morphology of boutons in cortex slices
State of synaptic vesicle population Boutons with synaptic complexes Partially Completely Boutons without synaptic complexes Full, depleted, depleted, Partially Total % Full % % depleted 265 87 8 5 376 (0.11 ;m-2)* 42 (0.012 Mm-2)* 121 9 17 74 16 (0.010 Mm-2)*t 34 (0.021 Am-2)*t

Area examined,

Control a-Latrotoxin-treated

3500 6400

Mousecerebral cortex slices were incubated for 1 hr at 370 in modified Krebs-Ringer buffer, pH 7.4; to the experimental sample were added 10 jtg a-latrotoxin at zero time and 5 ,ug more at 30 min. * Figure in parentheses is the number of boutons per square micrometer. t The value in the toxin-treated case has been corrected to the original area, to take into account the swelling due to the toxin (see text).

K+ and other depolarizing agents have been used. However, nontransmitter substances are also released by high concentrations of K+ from neuronal (13, 14) as well as from non-neuronal (15) elements. Although well controlled electrical stimulation should be more specific, it has been observed that electrical stimulation is unable to release GABA from cerebral cortex slices at stimulating potentials adequate for the release of other putative transmitters. Enhanced release of GABA occurred only with applied potentials that were high enough to also enhance release of nontransmitter substances (16). Some snake toxins such as f3-bungarotoxin have also been found to cause depletion of synaptic vesicles at neuromuscular junctions (17), but it was found that nontransmitter substances were also released by this toxin (18). Because the nontransmitter, AIB, was not released by a-latrotoxin, it appears that a-latrotoxin is unique as a transmitter releasing agent. The mechanism of action of a-latrotoxin is still unknown. It has been reported (19) to increase cation premeability in artificial lipid bilayers, which suggests that the toxin molecule can insert itself into lipid bilayers. However, this would not explain the selectivity of the toxin towards its target tissue. Thus, consistent with the selective action of the toxin to neural tissue, we have demonstrated recently a very high-affinity specific binding of iodinated a-latrotoxin to protein receptor in synaptic membrane fractions from cerebral cortex but not in liver plasma membrane preparations (unpublished results). However, the possibility exists that a-latrotoxin still retains its ionophore-like activity after binding to its receptor. A cation permeability increase could be responsible for the depolarization (20) and swelling of the boutons (1, 21, 22), and for the anomalous effect of high concentrations of extracellular Ca2+ (23). Conceivably, the ionophore-like activity may stimulate transmitter release, but, since the effect of a-latrotoxin is independent of extracellular Ca2+ (this paper and refs. 1, 20) and of Na+ (24), it seems that this activity is not sufficient to account for all the effects of the toxin. Another hypothesis based on the interaction of the toxin or its receptor with presynaptic contractile elements has been proposed (25), and it is a candidate for further testing. It is not clear at present what is the relevance of transmitter release evoked by a-latrotoxin to that evoked by nerve stimulation. The latter process depends on the presence of extracellular Ca2:+, whereas the former does not. This difference would seem to argue that toxin-induced release occurs via a mechanism that is different from normal release. But, because we do not know how Ca2+ acts to release transmitters, it is possible that Ca2+ and a-latrotoxin both activate the same release mechanism but by different means. B. Ceccarelli, F. Grohovaz and W. P. Hurlbut (personal communication) have found that plasmalemmal deformations, presumably due to vesicle fusions, induced by venom in the presence of Ca2+ occur mainly near the active zones, whereas deformations induced by K+ in the

presence of Ca2+ occur all over the presynaptic face of the axolemma. Our interpretation of these findings is that the toxin does act like Ca2+ at the active zone but possibly at a receptor site different from that of Ca2+. Whatever the mechanism of action, the use of a-latrotoxin is one of the better means to clarify the relationship between synaptic vesicles and transmitter release.
We thank Mr. B. Capparella and Dr. N. Frontali for help in obtaining the spiders, Lois Lynch for aid in the electron microscopy, and Drs. W. P. Hurlbut, N. H. Chua, and A. Gorio for their editing of the manuscript. This research was supported in part by U.S. Public Health Service Grant NS 12726 to P.S. 1. Frontali, N., Ceccarelli, B., Gorio, A., Mauro, A., Siekevitz, P., Tzeng, M. & Hurlbut, W. P. (1976) J. Cell Biol. 68,462-479. 2. Tzeng, M. & Siekevitz, P. (1978) Brain Res. 139, 190-196. 3. Gelder, N. M. (1966) Biochem. Pharmacol. 15,533-539. 4. Hopkin, J. & Neal, M. J. (1971) Br. J. Pharmacol. 42, 215223. 5. Nadler, J. V. & Cooper, J. R. (1972) J. Neurochem. 19, 20912105. 6. Hamberger, A. (1971) Brain Res. 31, 169-178. 7. Descarries, L., Beaudet, A. & Watkins, K. C. (1975) Brain Res. 100,563-588. 8. Descarries, L., Watkins, K. C. & Lapierre, Y. (1977) Brain Res. 133, 197-222. 9. Grasso, A., Rutini, S. & Senni, I. (1978) FEBS Lett. 85, 241244. 10. Baba, A., Sen, I. & Cooper, J. R. (1977) Life Sci. 20, 833-842. 11. Kornguth, S. E. (1974) Rev. Neurosci. 1, 63-114. 12. Granata, F., Traina, M. E., Frontali, N. & Bertolini, B. (1974) Comp. Biochem. Physiol. 48A, 1-7. 13. Roberts, P. J. (1974) Brain Res. 67, 419-428. 14. Vargas, 0. & Orrego, F. (1976) J. Neurochem. 26,31-34. 15. Sellstrom, A. & Hamberger, A. (1977) Brain Res. 119, 189198. 16. Orrego, F. & Miranda, R. (1976) J. Neurochem. 26, 10331038. 17. Chen, I. L. & Lee, C. Y. (1970) Virchows Arch. Abt. B. Zellpath. 6, 318-325. 18. Wernicke, J. F., Vanker, A. D. & Howard, B. D. (1975) J. Neurochetn. 25, 483-496. 19. Finkelstein, A., Rubin, L. L. & Tzeng, M. (1976) Science 193, 1009-1011. 20. Longenecker, H. E., Jr., Hurlbut, W. P., Mauro, A. & Clark, A. W. (1970) Nature 225,701-703. 21. Clark, A. W., Mauro, A., Longenecker, H. E., Jr. & Hurlbut, W. P. (1970) Nature 225, 703-705. 22. Clark, A. W., Hurlbut, W. P. & Mauro, A. (1972) J. Cell. Biol. 52, 1-14. 23. Smith, J. E., Clark, A. W. & Kuster, T. A. (1977) J. Neurocytol.

24. Gorio, A., Rubin, L. L. & Mauro, A. (1978) J. Neurocytol. 7, 193-205. 25. Tzeng, M. (1978) Dissertation (The Rockefeller University, New York,


To top