J. Cell Set. yt, 99-108 (1978)Printed in Great Britain © Company of Biologists Limited IQJ8

MACROTUBULES INDUCED BY

HALOTHANE: IN VITRO ASSEMBLY

ROBERT E. HINKLEY, JHDepartments of Anatomy and Anesthesiology, University of Miami School of Medicine,Miami, Florida

SUMMARY

The formation of macrotubules by the volatile anaesthetic halothane was investigated in vitrousing mjcrotubule-enriched fractions of crayfish nerve cords. Sequential studies showed thatmacrotubules assemble from helical ribbons of 18—20 laterally associated micro tubule proto-filaments which fold upon themselves to form intact macrotubules averaging 48 run in diameter.The initial rate of macrotubule assembly is dependent on the concentration of halothaneemployed and is stimulated by calcium. Glycerol pretreatment blocked macrotubule formationby halothane and caused preformed macrotubules to reassemble rapidly into typical micro-tubules. These experiments show that microtubules and macrotubules require different con-ditions for assembly and support the contention that macrotubule formation by halothane isdue to a direct interaction between the anaesthetic molecule and the microtubule subunit.

INTRODUCTION

The conversion of microtubules into larger tubular forms has been reported inseveral cell types following exposure to a number of experimental treatments (Tyson &Bulger, 1973). As a class, these enlarged tubular forms or 'macrotubules' measure3O-o-6o-o run in diameter and have wall thicknesses in the 5io-J7io nm range. Althoughthere is good evidence that macrotubules represent alternative assembly forms ofmicrotubule protein, little is known about their structural formation or conditions ofassembly.

Halothane, a widely used volatile anaesthetic, rapidly transforms axonal micro-tubules of crayfish ventral nerve cords into macrotubules averaging 46-0 nm indiameter (Hinkley & Samson, 1972). The substructure of halothane-induced macro-tubules and the sequence of structural events by which they assemble were recentlydemonstrated in experiments using tannic acid and negative-staining techniques(Hinkley, 1976). This tubular transformation has now been examined in vitro usingmicrotubule-enriched preparations of crayfish nerve cords. The results show thatmacrotubule formation by halothane will occur in vitro and that macrotubule formationis accelerated by calcium and can be blocked or reversed by glycerol.

Author's mailing address: Departmentof Anatomy, University of Miami, School of Medicine,P.O. Box 520875, Miami, FL. 33152, U.S.A.

ioo R. E. Hinkley, Jr

MATERIALS AND METHODS

The procedure used to isolate crayfish tubulin was developed by Pierson & Burton (1975).Crayfish ventral nerve cords were isolated in o-i M MES (2-(N-morpholino) ethane-sulphonicacid) containing 1 mM GTP, 1 mM EGTA, 0-5 mM MgCla, and 0-047 M NaCl, pH 6-4 at 4 °Cin a ratio of 10 cords/o-i ml bufFer. The cords were diced into short segments, thoroughlyhomogenized, and centrifuged at 80000 g for 45 min. The resulting supernatant (S-i) con-tained about 8-10 mg protein/ml and was used as the starting material for the in vitro assemblyof both microtubules and macrotubules.

Crayfish microtubules were assembled in vitro by diluting the cold S-i fraction with an equalvolume of polymerization buffer (identical to the isolation buffer but containing 5 mM GTP)and incubating the mixture at 30-32 CC for up to 1 h (Pierson & Burton, 1975).

To induce macrotubules in vitro, halothane (2-bromo, 2-chloro, 1:1: i-trifluoroethane,Ayerst Laboratories, New York) was dissolved in cold polymerization buffer at initial con-centrations of either 10 or 20 mM. Equal volumes of halothane-containing polymerizationbuffer and S-i supernatant were then mixed and incubated in gas-tight vials at 30—32 °C forup to 1 h.

For electron microscopy, solutions containing microtubules or macrotubules were centri-fuged at 80000 g, for 30-35 min and the resulting pellets fixed in 6% glutaraldehyde-o-i Mcacodylate buffer (pH 70) for 30 min. After a short buffer wash, the pellets were postfixedin 1 % osmium tetroxide-cacodylate buffer, dehydrated in ethanol through propylene oxide,and embedded in Epon 812. Thin sections were stained with ethanolic uranyl acetate and leadcitrate and examined in an Hitachi 11F electron microscope.

Sequential stages of macrotubule assembly were examined by negative-staining techniques.The S-i fraction was sampled at 1, 2, 5, 10 and 30 min after initiating macrotubule assemblywith halothane. All samples were allowed to settle on Formvar-carbon-coated 200-mesh gridsfor about 15s before rinsing in rapid succession with 3 drops each of 1 % aqueous cytochromec, distilled water, and 5 % uranyl acetate.

The effects of Ca*"1", halothane, and GTP on the rate of macrotubule assembly in vitro weredetermined spectrophotometrically and confirmed by electron microscopy. To test the effectof Ca8"1" on macrotubule assembly, a series of solutions were prepared by adding 1, 2, 5 or10 mM CaCls to cold o-i M MES buffer (pH 6-4) containing 1 mM GTP, 0-5 mM MgCls, and0-047 M NaCl. Just before use, each Ca1+ test solution was equilibrated with 20 mM halothane.Each test solution was mixed with an equal volume of cold S-i supernatant in a cuvette andthe resulting increase in turbidity measured at 350 nm in a Gilford 2400-S recording spectro-photometer equilibrated to 32 °C. Halothane test solutions were prepared by dissolving 1, 4,10 or 20 mM halothane in o-i M MES buffer (pH 6-4) containing 1 mM GTP, 0-5 mM MgCls,0-047 M NaCl, and 5 mM CaCla. GTP test solutions were prepared by dissolving 1, 2, 4, or10 mM GTP (Sigma, type II-S) ino-i M MES buffer containing 0-5 mM MgClj, 0-047 M NaCl,and s mM CaCl,. Just before use, each GTP test solution was equilibrated with 20 mM halo-thane. Each halothane or GTP test solution was mixed 1:1 with the S-i supernatant and moni-tored for increases in turbidity.

To test the reversibility of macrotubule formation, macrotubules were assembled in vitro inthe presence of halothane (and Ca1+) and pelleted by centrifugation. Part of the pellet wasfixed for electron microscopy to confirm the presence of macrotubules. The remainder of thepellet was resuspended in fresh microtubule-polymerizing buffer and incubated at 32 °C for30 min-i h. This solution was then recentrifuged and the resulting pellet fixed for electronmicroscopy. In parallel experiments, the macrotubule pellet was resuspended in microtubule-polymerizing buffer containing 4 M glycerol.

Macrotubule assembly in vitro 'iff

RESULTS

Numerous microtubules appeared in the pellets and negatively stained samples ofthe S-i fraction polymerized at 32 °C (Figs. 1, 2). These microtubules possess materialadhering to their exterior surfaces and are constructed of protofilaments orientedapproximately longitudinally within the wall lattice.

v .

*a?o,« •* » . u

jki'.i.-

wFig. 1. Crayfish microtubules assembled from the S-i supernatant at 32 °C for30 min. Note adhering material on exterior surfaces of the microtubules. x 114100.Fig. 2. Negatively stained microtubule assembled from S-i supernatant. Protofila-ments are oriented lengthwise within the wall matrix, x 234 000.

If the S-i fraction is polymerized in the presence of 5-10 mM halothane, assemblingand intact macrotubules are seen in the resulting samples (Figs. 3-7). In cross-sectionmacrotubules measure 48-2 ± 3-2 run (n = 35) in diameter, have walls about 6-o-6-5 run thick, and usually contain a 'dot' of electron-dense material in their lumens.The wall structure of in vitro assembled macrotubules is shown in Figs. 6, 7. Theprotofilaments comprising the wall lattice are inclined at an angle of about 350 relativeto the longitudinal axis of the macrotubule. In some regions the protofilaments areclearly organized into pairs. These doublet or bifilar protofilaments average 9-0-10-0run in width with a centre-to-centre spacing between protofilament pairs of aboutn-o-13'5 run.

As expected, the initial rate of halothane-induced macrotubule assembly in vitro isdependent on the concentration of halothane employed (Table 1). Ca2+ blockedmicrotubule assembly but stimulated macrotubule assembly in a dose-dependentfashion over the range of concentrations used (Table 2). However, Ca2+ alone was

102 R. E. Hinkley, Jr

Fig. 3. Macrotubules polymerized from the S-i supernatant in the presence of i DIMCaClj and 5 IBM halothane, for 30 min at 32 °C. x 117 000.

Macrotubule assembly in vitro 103

Figs. 4-7. Sequential stages of macrotubule assembly examined by negative staining.Twisted ribbons of microtubule protofilaments such as those seen in Figs. 4, 5 werepresent after 1-5 rnin of halothane exposure. After 5-10 min, macrotubules in variousstages of assembly predominated in the samples. Fig. 6 shows a macrotubule with 2'breaks' in its wall structure. An intact macrotubule segment is shown in Fig. 7. Tiltpage to visualize the bifilar protofilament construction of ribbons and macrotubules.All x 146800.

104 R. E. Hinkley, Jr

insufficient to cause macrotubule formation, even when tested over a range of con-ditions, including changes in pH. Changes in GTP concentrations had no appreciableeffect on the initial rate of macrotubule assembly in vitro (Table 3).

Although macrotubules rapidly dissociated when resuspended in fresh polymeri-zation buffer, few intact microtubules reappeared. On the other hand, if macrotubuleswere resuspended in polymerization buffer containing 4 M glycerol, the reversetransformation to microtubules occurred rapidly. Crayfish microtubule proteinexposed to glycerol during isolation did not form macrotubules when exposed tohaJothane and Ca2+ in the polymerization buffer. Instead, typical microtubules wereseen in the resulting pellets.

Table 1. The effects of halothane on the initial rate of macrotubule assembly

[Halothane], mM. . . 0-5 2-0 5-0 io-o

A O.D.jio/io min/ml S-i. . . 0-2255 0-5250 10500 2-5500

See Methods for description of test solutions and conditions for spectrophotometric assay.

Table 2. The effects of Ca2+ on the initial rate of macrotubule assembly

CaClj, mM

Control 0-5 i-o 2-5 5-0

A O.D.3J0/10 min/ml S - i . . . 08625 °-975° 1*3125 2-2500 4-3500

See Methods for description of test solutions and conditions for spectrophotometric assay.

Table 3. The effects of GTP on the initial rate of macrotubule assembly

GTP,

Control 025 0-50 i-o 2-5

A O.D.JM/IO min/ml S- i . . . 3-0560 27700 2-7750 2-7740 2-6180

See Methods for description of test solutions and conditions for spectrophotometric assay.

DISCUSSION

The term 'macrotubule' has been used to describe experimentally induced tubularstructures measuring 3O-o-6o-o nm in diameter which are constructed of microtubuleprotein. Among the agents reported to induce macrotubules are low temperatures(Tilney & Porter, 1967), colchicine (Tilney, 1968), vinblastine (Tyson & Bulger, 1973;Warfield & Bouck, 1974, 1975), digitonin (Hanzely & Olah, 1973), hyaluronidase(Burton & Fernandez, 1973), isopropyl Ar-phenyl carbamate (Brown & Bouck, 1974),and a number of volatile anaesthetics, including halothane (Allison etal. 1970; Hinkley,1976). The pharmacological diversity of these agents has made it difficult to establish

Macrotubule assembly in vitro 105

unifying principles concerning macrotubule formation or conditions of macrotubuleassembly. Moreover, none of these agents routinely induces macrotubules in all systemsand opportunities to study macrotubule formation and substructure are corre-spondingly rare. Previous studies on macrotubules induced by halothane in crayfishnerve cords demonstrated the tubulin-like composition of the macrotubules as well astheir wall substructure and mechanism of assembly. These studies have now beenexpanded using an in vitro microtubule polymerization system to define conditionswhich regulate macrotubule assembly by halothane.

Macrotubules induced by halothane in vitro assemble by a set of structural eventsidentical to those reported to occur within crayfish axons. Essentially, macrotubuleformation is initiated by the appearance of filaments that are twisted or folded atintervals of about 0-2 fim along their lengths. Evidence obtained in earlier studiessuggested that these twisted filaments could be formed directly from disassemblingcrayfish microtubules. The formation of twisted filaments in vitro could occur by thereassociation of halothane-modified subunits and perhaps, by a direct conversion ofhalothane-modified ring structures which are present in cold nerve cord extracts.Regardless of their exact mechanism(s) of assembly, the twisted filaments formedin vitro rapidly associate laterally to produce structures resembling 'twisted ribbons'.These twisted ribbons are identical to ribbons observed in negatively stained axo-plasmic samples of nerve cords exposed to halothane (Hinkley, 1976) and resemblethose reported elsewhere during 'normal' microtubule assembly m vitro (Erickson,1976; Kirschner & Williams, 1974; Kirschner, Honig &Williams, 1975; Penningroth,Cleveland & Kirschner, 1976). Twisted ribbons such as those shown in Figs. 4 and5, however, were never encountered during the assembly of crayfish microtubules.Most ribbons rapidly acquire a helical appearance and range in width to a maximumof about 18-20 protofilaments. Finally, intact macrotubules are thought to form whenthe twisted ribbons reach an optimum width of about ten protofilament pairs andthe subunits lining the edges of adjacent ribbon gyres come into contact and join.

In earlier reports, haJothane was suggested to promote macrotubule formation inone of two possible ways: (1) directly by altering the conformation of either the micro-tubule subunits or microtubule-associated proteins (MAPs); or (2) indirectly byadversely affecting the microtubule environment (Hinkley, 1976; Hinkley & Samson,1972). The in vitro polymerization experiments reported here support the contentionthat macrotubule formation by halothane involves a direct interaction between theanaesthetic molecule and the tubulin subunit or associated proteins. Although theexact nature of the halothane-tubulin (or -MAP) interaction is open to speculation,the ability of fluorinated hydrocarbon anaesthetic agents to bind to globular proteinsand produce reversible conformational changes has been demonstrated in severalmodel systems (Balasubramanian & Wetlaufer, 1966; DiPaolo & Sandorfy, 1974;Laasberg & Hedley-Whyte, 1971; Leuwenkroon-Strosberg, Laasberg & Hedley-Whyte, 1973; Wetlaufer & Lovrien, 1964). An analogous conformational change couldoccur within the crayfish tubulin subunit or perhaps associated proteins, as suggestedelsewhere (Hinkley, 1976). Regardless of the exact nature of the interaction, the netstructural effect appears to be a change in the bonding angles between assembling

106 R.E.Hinkley,Jr

subunits which results in the formation of twisted protofilaments that act as templatesfor the assembly of twisted ribbons and ultimately, macrotubules.

The in vitro assembly of crayfish microtubules and halothane-induced macrotubulesare regulated by different conditions. As expected, the initial rate of macrotubuleformation by halothane in vitro depends on the halothane concentration used. How-ever, turbidimetric assays and electron-microscope studies show that Ca2+ has a dose-dependent stimulatory effect on macrotubule assembly. These results are difficult tointerpret in view of the well known Ca24" sensitivity of microtubule assembly in vitro(Lee, Samson, Houston & Himes, 1974; Olmsted & Borisy, 1975; Weisenberg, 1972).Recent studies show that Ca2+ binds to tubulin and that low concentrations of Ca2+

actually may be required for microtubule assembly (Borisy, Johnson & Marcum,1976; Hayashi & Matsumura, 1975; Solomon, 1976). Concentrations of Ca2+ up to30 mM have been reported to promote the reversible formation of tubulin aggregatesor precipitates (Weisenberg & Timasheff, 1970; Wilson, Bryan & Mazia, 1970). Inthis context, it should be noted that Ca2+ concentrations which stimulate macrotubuleassembly generally exceed levels which block microtubule assembly in vitro anddepolymerize preformed microtubules. However, under no conditions did Ca2+

alone induce the formation of macrotubules in this system. Accordingly, macrotubuleformation in this system requires the presence of halothane and appears to be morecomplicated than Ca2+-pH effects, which have been reported to cause the formationof twisted ribbons and 'super-macrotubules' during tubulin assembly in vitro(Matsumura & Hayashi, 1976; Langford, 1977).

Glycerol stabilizes microtubule protein and enhances microtubule assembly in vitro(Shelanski, Gaskin & Cantor, 1973). The present study shows that glycerol protectsmicrotubule protein against alteration by halothane. For example, pretreatment withglycerol completely blocked macrotubule formation in crayfish nerve cord extractsexposed to halothane. Moreover, glycerol appears able to reverse the action ofhalothane in this system since macrotubules reverted to microtubules when re-suspended in polymerization buffer containing glycerol. Since glycerol is thought tostabilize microtubule protein by binding tightly to the tubulin dimer (Detrich,Berkowitz, Kim & Williams, 1976), it is tempting to speculate that the preferentialbinding of glycerol by crayfish tubulin may be responsible for the prevention andreversibility of macrotubule formation by halothane. This view, however, dependson the existence of 'halothane-alterable sites' and, at present, there is no directevidence to support the existence of such sites on the tubulin molecule.

In general, macrotubule induction has been viewed as an unpredictable event,depending on unique combinations of specific agents and microtubule systems. Thestructural transformation of crayfish microtubules into macrotubules has beendescribed along with conditions which regulate the rate and reversibility of macro-tubule assembly in vitro. Comparative experiments with agents reported to inducemacrotubules in other systems show that vinblastine and isopropyl iV-phenyl carba-mate induce macrotubule forms in crayfish microtubule preparations. Digitonin andhyaluronidase blocked microtubule assembly but did not cause macrotubule formation.On the basis of the present experiments, several mechanisms of macrotubule induction

Macrotubule assembly in vitro 107

probably exist and experiments are now under way to compare the in vitro formationof macrotubules by vinblastine, isopropyl iV-phenyl carbamate and halothane incrayfish tubulin preparations.

This research was supported by NIH Grant GM 19813 and NIH Research Career Develop-ment Award 1 KO4-GM00173.

REFERENCESALLISON, A. C, HULAKDS, G. H., NUNN, J. F., RETCHING, J. A. & MACDONALD, A. C. (1970).

The effect of inhalational anaesthetics on the microtubular system in Actinosphaermmnucleofilum. J. Cell Sci. 7, 483-499.

BALASUBRAMANIAN, D. & WETLAUFER, D. B. (1966). Reversible alteration of the structure ofglobular proteins by anaesthetic agents. Proc. natn. Acad. Sci. U.S.A. 55, 762-765.

BORISY, G. G., JOHNSON, K. A. & MARCUM, J. M. (1976). Self-assembly and site-initiatedassembly of microtubules. In Cell Motility, Book C (ed. C. R. Goldman, T. Pollard &J. Rosenbaum), pp. 1093-1108. New York: Cold Spring Harbor Laboratories.

BROWN, D. L. & BOUCK, G. B. (1974). Microtubule biogenesis and cell shape in Ochroinonas.III. Effects of the herbicidal mitotic inhibitor isopropyl i\f-phenylcarbamate on shape andftagellum regeneration. J. Cell Biol. 61, 514-536.

BURTON, P. R. & FERNANDEZ, H. L. (1973). Delineation by lanthanum staining of filamentouselements associated with the surfaces of axonal microtubules. J. Cell Set. la, 567-583.

DETRICH, H. W., BERKOWITZ, S. A., KIM, H. & WILLIAMS, R. C. (1976). Binding of glycerolby microtubule protein. Biochem. biophys. Res. Cotnmun. 68, 961—968.

DIPAOLO, T. & SANDORFY, C. (1974). Fluorocarbon anaesthetics break hydrogen bonds.Nature, Lond. 254, 471-472.

ERICKSON, H. P. (1976). Facilitation of microtubule assembly by polycations. In Cell Motility,Book C (ed. R. Goldman, T. Pollard & J. Rosenbaum), pp. 1069-1080. New York: ColdSpring Harbor Laboratories.

HANZELY, L. & OLAH, L. V. (1973). Effect of digitonin on cellular division. VI. Inducementof oversized tubular elements in treated AUium cells. Cytologia 38, 623-633.

HAYASHI, M. & MATSUMURA, F. (1975). Calcium binding to bovine brain tubulin. FEBSLetters, Amsterdam 58, 222-225.

HINKLEY, R. E. (1976). Microtubule-macrotubule transformation induced by volatile anesthe-tics. Mechanism of macrotubule assembly. J. Ultrastruct. Res. 57, 237-250.

HINKLEY, R. E. & SAMSON, F. E. (1972). Anesthetic-induced transformation of axonal micro-tubules. J. Cell Biol. S3, 258-263.

KIRSCHNER, M. W., HONIG, L. S. & WILLIAMS, R. C. (1975). Quantitative electron microscopyof microtubule assembly in vitro. J. molec. Biol. 99, 263-276.

KIRSCHNER, M. W. & WILLIAMS, R. C. (1974). The mechanism of microtubule assembly invitro. J. supramolec. Struct, a, 412-428.

LAASBERG, L. H. & HEDLEY-WHYTE, J. (1971). Optical rotatory dispersion of hemoglobin andpolypeptides. Effect of halothane J. biol. Chem. 246, 4886—4893.

LANGFORD, G. M. (1977). Calcium induction of coiled sheet polymers of dogfish brain tubulin.Fedn Proc. Fedn Am. Socs exp. Biol. 36, 899.

LEE, Y. C, SAMSON, F. E., HOUSTON, L. L. & HIMES, R. H. (1974). The in vitro polymerizationof tubulin from beef brain. J. Neurobiol. 5, 317-330.

LEUWENKROON-STROSBERG, E., LAASBERG, L. H. & HEDLEY-WHYTE, J. (1973). Myosin con-formation and enzymatic activity: Effect of chloroform, diethyl ether, and halothane onoptical rotatory dispersion and ATPase. Biochim. biophys. Acta 295, 178-187.

MATSUMURA, F. & HAYASHI, M. (1976). Polymorphism of tubulin assembly in in vitro formationof sheet, twisted ribbon, and microtubule. Biochim. biophys. Acta 453, 162-175.

OLMSTED, J. B. & BORISY, G. G. (1975). Ionic and nucleotide requirements for microtubulepolymerization in vitro. Biochemistry, N.Y. 14, 2996-3005.

io8 R. E. Hinkley, Jr

PENNINGROTH, S. M., CLEVELAND, D. W. & KIRSCHNER, M. W. (1976). In vitro studies of theregulation of microtubule assembly. In Cell Motility, Book C (ed. R. Goldman, T. Pollard &J. Rosenbaum), pp. 1233-1257. New York: Cold Spring Harbor Laboratories.

PIERSON, G. & BURTON, P. R. (1975). Isolation and repolymerization of tubulin from crayfishnerve cord into microtubules with 11, 12 and 13 protonlaments. J. Cell Biol. 67, 336a.

SHELANSKI, M. L., GASKIN, F. & CANTOR, C. R. (1973). Microtubule assembly in the absenceof added nucleotides. Proc. natn. Acad. Sci. U.S.A. 70, 765—768.

SOLOMON, F. (1976). Characterization of the calcium binding activity of tubulin. In CellMotility, Book C (ed. R. Goldman, T. Pollard & J. Rosenbaum), pp. 1139-1148. New York:Cold Spring Harbor Laboratories.

TILNEY, L. G. (1968). Studies on the microtubules in heliozoa. IV. The effect of colchicine onthe formation and maintenance of the axopodia and the redevelopment of pattern in Actino-sphaerium nucleofilum (Barrett). J. Cell Sci. 3, 549—562.

TILNEY, L. G. & PORTER, K. R. (1967). Studies on the microtubules in heliozoa. II. Theeffect of low temperature on these structures in the formation and maintenance of theaxopodia. J. Cell Biol. 34, 327-343.

TYSON, G. E. & BULGER, R. E. (1973). Vinblastine-induced paracrystals and unusually largemicrotubules (macrotubules) in rat renal cells. Z. Zellforsch. mikrosk. Anat. 141, 443-458.

WARFIELD, R. K. N. & BOUCK, G. B. (1974). Microtubule-macrotubule transitions: Inter-mediates after exposure to the mitotic inhibitor vinblastine. Science, N.Y. 186, 1219—1221.

WARFIELD, R. K. N. & BOUCK, G. B. (1975). On macrotubule structure. J. molec. Biol. 93,117-120.

WEISENBERG, R. C. (1972). Microtubule formation in vitro in solutions containing low calciumconcentrations. Science, N.Y. 177, 1104—1105.

WEISENBERG, R. C. & TIMASHEFF, S. N. (1970). Aggregation of microtubule subunit protein.Effects of divalent cations, colchicine, and vinblastine. Biochemistry, N.Y. 9, 4110—4116.

WETLAUFER, D. B. & LOVRIEN, R. (1964). Induction of reversible structural changes in proteinsby nonpolar substances. J. biol. Chem. 239, 596-603.

WILSON, L., BRYAN, J. & MAZIA, D. (1970). Precipitation of proteins by vinblastine and calciumions. Proc. natn. Acad. Sci. U.S.A. 66, 807-814.

(Received 10 January 1978)