THE JOURNAL OF CHEMISTRY Vol. 266, No. 23, Issue of pp ... › content › 266 › 23 › 15035.full.pdf · THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 266, No. 23, Issue of August 15,

THE J O U R N A L OF BIOLOGICAL CHEMISTRY Vol. 266, No. 23, Issue of August 15, pp. 15035-15041, 1991 Printed in U. S. A.

Squid Low Molecular Weight Neurofilament Proteins Are a Novel Class of Neurofilament Protein A NUCLEAR LAMIN-LIKE CORE AND MULTIPLE DISTINCT PROTEINS FORMED BY ALTERNATIVE RNA PROCESSING*

(Received for publication, March 11, 1991)

Ben G . SzaroS, Harish C. Pant, James Way, and James Batteyg From the Laboratory of Neurochemistry, National Institute of Neurological Disorders and Stroke, Bethesda, Maryland 20892 and the Marine Biological Laboratory, Woods Hole, Massachusetts 02543

The primary structure of the major 60-kDa squid low molecular mass neurofilament protein (NF60) and a related 70-kDa neurofilament protein have been de- termined from cDNA clones isolated from a squid brain cDNA library. Structural analysis suggests that the squid NF60 and NF70 neurofilament genes and pro- teins are remarkably distinct from vertebrate neuronal intermediate filaments characterized previously. Both proteins are encoded on mRNAs generated by alter- native RNA processing of the primary transcript of a single gene. Among the known intermediate filament proteins, NF60 and NF70 neurofilament proteins show highest similarity to an epithelial intermediate fila- ment protein from Helixpomatia, a gastropod mollusk, and are less similar to vertebrate neurofilaments. The length of the a-helical rod domain in the NF60 and NF70 proteins was reminiscent of the vertebrate nu- clear lamins, 6 heptads longer than is found in all known vertebrate cytoplasmic intermediate filaments, in particular the vertebrate neurofilaments. These dis- tinct structural properties suggest that the vertebrate and invertebrate low molecular weight neurofilaments evolved independently from primordial intermediate filament proteins.

Neurofilament proteins are the major component of the axonal intermediate filaments of all vertebrate and many invertebrate neurons. They are members of a larger superfam- ily of cytoskeletal proteins referred to collectively as the intermediate filament proteins. Vertebrate neurofilament pro- teins and the other vertebrate cytoplasmic intermediate fila- ment proteins possess an cy-helical rod domain defined by regions of quasiheptad repeats of amino acids in which the 1st and 4th residues of the heptad are usually hydrophobic or nonpolar. Although these rod domains contain the same num- ber and distribution of heptad repeats, their individual amino acid sequences vary extensively between separate classes of intermediate filament proteins, but less so within a class (1,

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence($ reported in thispaper has been submitted

M64 71 7 and M64 718. to the GenBankTM/EMBL Data Bank with accession number(s)

$ Current address: Dept. of Biological Sciences, University at Al- bany, State University of New York, Albany, NY 12222.

dressed: Bldg. 36, Rm. 4D-20, NIH, Bethesda, MD 20892. Tel.: 301- 5 To whom correspondence and reprint requests should be ad-

496-9917.

2). Vertebrate neurofilament proteins are found in two sepa- rate classes of intermediate filament proteins. One class (Type IV intermediate filament proteins) consists of the three most abundant neurofilament proteins: a low molecular weight protein (NF-L)’ and two higher molecular weight forms (NF- M and NF-H). Because of their amino acid sequences and the placement of their introns (3), they are considered separate from neurofilament proteins like peripherin, which are con- sidered Type I11 (4, 5 ) . Some non-neuronal intermediate filament proteins like vimentin and desmin are also Type I11 (reviewed in Refs. 1 and 2), whereas the remaining vertebrate cytoplasmic intermediate filament proteins are Types I and I1 (acid and basic cytokeratins).

Cytoplasmic intermediate filament proteins of invertebrate cells do not necessarily resemble the intermediate filament proteins of analogous vertebrate cell types. For example, two intermediate filament proteins from epithelial cells of the gastropod mollusk, Helix pomatia, do not resemble cytokera- tins ( 6 ) , nor do the intermediate filament proteins of the muscle of the nematode, Ascaris lumbricoides, resemble des- mins (7). Instead, the number of heptad repeats within their rod domains is the same as for lamins, which are found in the nucleus. Nuclear lamins constitute a fifth class of intermedi- ate filament proteins, distinct from all vertebrate cytoplasmic intermediate filament proteins because their rod domain con- tains 6 extra heptads (8).

The question remains whether the neurofilament proteins of invertebrates are similar at the amino acid level to verte- brate neurofilament proteins. Arguments in favor of a com- mon derivation for all neurofilament proteins are based partly on observations that invertebrate neurofilament proteins dif- fer from invertebrate non-neuronal intermediate filament proteins (6, 9, 10) by molecular weights and antibody cross- reactivities; whereas the same non-neuronal intermediate fil- ament proteins can be found in a variety of invertebrate cell types (6, 7). Furthermore, invertebrate and vertebrate neuro- filament proteins share many physical and biochemical prop- erties (10-12) that are distinct from non-neuronal interme- diate filament proteins. Thus, information on the amino acid sequences of invertebrate neurofilaments is needed.

The squid, Loligo pealei, is especially suitable for studying invertebrate neurofilament proteins, because its giant axon contains enough axoplasm to permit the characterization of

The abbreviations used are: NF-L, mammalian low molecular weight neurofilament protein; NF-M, mammalian middle molecular weight neurofilament protein; NF-H, mammalian high molecular weight neurofilament protein; GFAP, glial fibrillary acidic protein; SDS, sodium dodecyl sulfate; kb, kilobase(s); kbp, kilobase pair(s); bp, base pair(s).

15035

15036 Squid Neurofilament Proteins

specifically neuronal intermediate filament proteins. The most abundant neurofilament protein in the squid giant axon is a 60-kDa protein (as estimated by SDS-polyacrylamide gel electrophoresis; reviewed in Ref. 10). In addition, the giant axon contains several less abundant neurofilament proteins (11, 13, 14). Using a cDNA probe (15) from mouse NF-L for low stringency hybridization screenings of a squid brain cDNA library, we found clones representing the 60-kDa (NF60) and a separate 70-kDa neurofilament protein (NF70). At the amino acid level, these proteins were more distantly related to mammalian NF-L than are Type I11 non-neuronal vertebrate intermediate filament proteins and were more closely related to the gastropod epithelial intermediate fila- ment protein. Moreover, these two squid neurofilament pro- teins were identical in their N-terminal and rod domains, but differed extensively in their C-terminal domains. Unlike ver- tebrate neurofilament proteins, which are encoded by separate genes, these two squid neurofilament proteins were encoded by a single gene. Thus, multiple distinct squid neurofilament proteins result from alternatively spliced mRNA encoded from a single gene.

EXPERIMENTAL PROCEDURES AND RESULTS*

DISCUSSION

In this paper, we describe the cloning and sequencing of cDNAs representing two neurofilament proteins from cDNA libraries made from the optic lobe of the squid, Loligo pealei. One of these (NF60) was identified as the well known 60-kDa squid low molecular weight neurofilament protein. The other (NF70) was an additional, less abundant, neurofilament pro- tein generated from the same gene as NF60 by alternative RNA processing. Despite expectations from biochemical and immunological studies, apart from being intermediate fila- ment proteins, these two proteins were otherwise unrelated to all previously characterized vertebrate neurofilament pro- teins.

Relationship of NF60 and NF70 to Vertebrate Neurofila- ment Proteins-Several considerations led us to conclude that the squid NF60 and NF70 proteins are unrelated to the neurofilament proteins of vertebrates. First, dot-plot analyses (not illustrated) revealed that the amino and carboxyl ends of the rod domain were the only regions of significant simi- larity between the two squid NF proteins and vertebrate intermediate filament proteins, including representative ex- amples of both neurofilaments and non-neuronal intermedi- ate filament proteins. These regions were the same as those typically found conserved between intermediate filament pro- teins of different classes (2). Second, an analysis of amino acid identity (Table 1) indicated that the squid proteins were equivalently distantly related to representatives of the verte- brate intermediate filament classes. This was also true when N- and C-terminal domains were compared. Thus, by these structural criteria, the squid NF60 and NF70 proteins could not be assigned to any one class of vertebrate intermediate filament proteins. Furthermore, the l b coil of the rod domain of NF60 and NF70 was 6 heptads longer than all known vertebrate cytoplasmic intermediate filament proteins, but the same length as for lamins. Longer lamin-like l b coil domains are also found in the cytoplasmic intermediate fila- ment proteins of two other invertebrates, i.e. the epithelial

Portions of this paper (including “Experimental Procedures,” “Results,” Table 1, and Figs. 1-9) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

intermediate filament proteins of the mollusk Helix pomatia (6) and the muscle intermediate filament proteins of the nematode Ascaris lumbricoides (7). It has been suggested that nuclear lamins and cytoplasmic intermediate filament pro- teins arose from a common lamin-like primordial precursor (7). The similarities between lamin, the squid NF60 and NF70 proteins, and other invertebrate non-neuronal intermediate filament proteins further suggest that the invertebrate cyto- plasmic intermediate filament proteins evolved along a path- way separate from their vertebrate counterparts. This conclu- sion is supported by the observations that the N-terminal and rod domains of the Helix intermediate filament proteins and squid NF60/70 proteins were more closely related by amino acid identity to one another than the squid proteins were to any of the vertebrate cytoplasmic intermediate filament pro- teins. Moreover, in the rod domain, the amino acid identity between the Helix intermediate filament and squid NF60/70 proteins (37.0%) was much less than that seen between mam- malian neurofilament proteins (54.5% between NF-L and NF- M) or even between Type I11 non-neuronal intermediate filament proteins and mammalian neurofilament proteins (49.3% between GFAP and NF-L). Thus, these data do not support a common origin for the vertebrate and invertebrate low molecular weight neurofilament proteins. Instead, these two squid neurofilament proteins appear to represent a sepa- rate class of neurofilament protein.

Additional sequences of invertebrate intermediate filament proteins and analyses of the structures of the squid NF60/70 and other invertebrate intermediate filament genes (e.g. in- tron locations) will be necessary in order to arrive at a clearer understanding of the evolution of intermediate filament pro- teins. For example, the 220-kDa squid neurofilament protein also shares numerous biochemical properties with NF-M and NF-H in mammals (10, 12). It will be important to determine its sequence in order to establish its relationship to squid NF60 and to mammalian neurofilament proteins. Further- more, it will be important to sequence some non-neuronal intermediate filament proteins from the squid to determine if the amino acid differences between the squid NF60/70 pro- teins and the Helix epithelial intermediate filament proteins are representative of differences between invertebrate neu- ronal and non-neuronal intermediate filament proteins or of more general differences between intermediate filament pro- teins in cephalopods versus gastropods. Examination of the intermediate filament proteins and genes of invertebrate phyla that are more closely related to the vertebrates than are the mollusks (i.e. echinoderms) may also provide impor- tant clues to the origins of the shorter rod domain of verte- brate cytoplasmic intermediate filament proteins.

Both mammalian NF-L and the squid 60-kDa neurofila- ment protein bind Bodian’s silver stain (10,34). In the mam- mal, Bodian’s silver (44) binds to the glutamate-rich, acidic region of the C-terminal domain common to NF-L, -M, and -H (45). However, there is no such glutamate-rich region in the C-terminal domain of NF60. Since in the squid, Bodian’s silver stain reacts with the 60-kDa neurofilament protein, but not the 70-kDa one (12, 34), a comparison of the C-terminal domains of these two proteins could provide some clues re- garding the potential binding sites for the Bodian stain. The closest approximation to a contiguous sequence of glutamates in the C-terminal domain of squid NF60 is a pair of glutamates at nucleotides 1594-1599. However, a pair of glutamates also occurs in NF70, which does not bind Bodian’s stain (NF70 nucleotides 1678-1683). Pairs of glutamates also occur in the C-terminal domains of a Xenopus Type I cytokeratin listed in Table 1 (amino acids 471-472) and of human lamin C

Squid Neurofilament Proteins 15037

(amino acids 36 and 37), which also do not bind Bodian’s stain. A more likely possibility for an acidic domain that could serve as a potential Bodian’s binding site is the unusual amino acid sequence “TTTTTTSSQE,” unique to the C-terminal domain of NF60 (nucleotides 1549-1588). If these amino acids were phosphorylated in vivo, they would represent a highly negatively charged domain, comparable to the glutamate-rich domains found in mammalian NF-L, -M, and -H.

The low molecular weight neurofilament proteins of mam- mals and the squid appear to have been generated by com- pletely different choices of amino acids and seem to have evolved along separate evolutionary pathways. Thus, it may be more important for the axon to possess cytoskeletal com- ponents with the physicochemical characteristics of neurofil- aments rather than a highly conserved sequence of amino acids. Such a situation is not unprecedented in nature. A more extreme example of this principle is the case of the lens crystallins, in which essentially unrelated proteins have been recruited in different phyla for the same function (46).

Alternative Splicing of Squid NF mRNAs from a Single Gene-In mammalian neurons, in addition to NF-L, two other low molecular weight neuronal intermediate filament proteins are encoded by separate genes: peripherin (4,5) and a-inter- nexin (47, 48). Multiple low molecular weight neurofilament proteins are also found in numerous lower vertebrates (34,49, 50, 51), also apparently derived from separate genes. Thus, our finding that NF60 and NF70 in the squid are generated by alternative splicing suggests a fundamentally different means of regulating the expression of these cytoplasmic in- termediate filament proteins and further underscores the difference between the genetic organization of the squid and vertebrate neurofilament systems. Invertebrate genomes are typically an order of magnitude lower in complexity than vertebrate genomes. The squid may have compensated for this by producing multiple neurofilament proteins from the same gene. It has been suggested that the duplication of primordial intermediate filament genes provided the tem- plates for subsequent intermediate filament protein diversi- fication during evolution (12). In the squid, diversification of the low molecular weight neurofilament proteins has been achieved, at least in part, by yet a different mechanism involving alternative splicing of mRNAs.

All of the major alternative forms in the squid involved variations in the C-terminal domain (Fig. 5). Although NF- L, NF-M, and NF-H are encoded by separate genes in verte- brates, they differ most extensively in the length and organi- zation of their C-terminal domains, which are thought to regulate interactions between neurofilaments and other com- ponents of the cell (10, 52). The rod domain of intermediate filament proteins is generally accepted to form the filament itself (2). Thus, the divergent C-terminal domains of these squid neurofilament proteins would alter only those portions of the protein involved in regulating interactions with neigh- boring cellular components while preserving the structure responsible for filament assembly. These data further suggest that the templates for the highly variant C-terminal domains of vertebrate intermediate filament proteins may have already been in place when primordial intermediate filament genes duplicated.

Acknowledgments-We thank Dr. Harold Gainer for his interest and advice during the project and his helpful comments on the manuscript. In addition, we thank Dr. Peter Steinert for his helpful comments on the manuscript. We also thank Dr. Philip Grant for his help in dissecting the squid and for his helpful tutelage in squid neuroanatomy.

1.

2.

3.

4.

5. 6.

7.

8.

9.

10.

11.

12.

13.

14. 15. 16.

17.

18.

19.

20.

21.

22.

23.

24

25.

26.

27.

28.

29.

30.

31.

32.

33. 34.

35.

36.

37.

38.

REFERENCES Steinert, P. M., and Parry, D. A. D. (1985) Annu. Reu. Cell Biol.

Steinert, P. M., and Roop, D. R. (1988) Annu. Reu. Biochem. 5 7 ,

Lewis, S. A., and Cowan, N. J. (1986) Mol. Cell Biol. 6, 1529-

Parysek, L. M., Chisholm, R. L., Ley, C. A., and Goldman, R. D.

Thompson, M. A,, and Ziff, E. B. (1989) Neuron 2 , 1043-1053 Weber, K., Plessmann, U., Dodemont, H., and Kossmagk-Ste-

Weber, K., Plessmann, U., and Ulrich, W. (1989) EMBO J. 8,

McKeon, F. D., Kirschner, M. W., and Caput, D. (1986) Nature

Cohen, R. S., Pant, H. C., House, S., and Gainer, H. (1987) J. Neurosci. 7, 2056-2074

Eagles, P. A. M., Pant, H. C., and Gainer, H. (1990) in Cellular and Molecular Biology of Intermediate Filaments (Goldman, R. D., and Steinert, P. M., eds) pp. 37-94, Plenum Publishing Corp., New York

Pant, H. C., Shecket, G., Gainer, H., and Lasek, R. J. (1978) J. Cell Biol. 7 8 , R23-R27

Lasek, R. J., Phillips, L., Katz, M. J., and Autilio-Gambetti, L. (1985) Ann. N. Y. Acad. Sci. 455,462-478

Zackroff, R. V., and Goldman, R. D. (1980) Science 2 0 8 , 1152- 1155

Zackroff, R. V. (1986) Methods Enzymol. 134 , 371-379 Lewis, S. A., and Cowan, N. J . (1985) J. Cell Biol. 100,843-850 Davis, L., Dibner, M., and Battey, J. (1986) Basic Methods in

Molecular Biology, Elsevier Science Publishing Co., New York Frohman, M. A., Dush, M. K., and Martin, G. R. (1988) Proc.

Natl. Acad. Sci. U. S. A. 85, 8998-9002 Devereaux, J., Haekerli, P., and Smithies, 0. (1984) Nucleic Acids

Res. 11,524-545 Needleman, S. B., and Wunsch, C. D. (1970) J. Mol. Biol. 48,

443-453 Maizel, J. V., and Lenk, R. P. (1981) Proc. Nutl. Acad. Sci. U. S. A.

78,7665-7669 Wilbur, W. J., and Lipman, D. J. (1983) Proc. Natl. Acad. Sci.

Hoffman, W., Sterrer, S., and Konigstorfer, A. (1988) FEBS L

Franz, J. K., and Franke, W. W. (1986) Proc. Nutl. Acad. Sci.

Ferrari, S., Battini, R., Racsmarek, L., Kittling, S., Calabretta, B., de Riel, J. K., Philiponis, V., Wei, J.-F., and Baserga, R. (1986) Mol. Cell. Biol. 6, 3614-3620

Lewis, S. A., Balcarek, J. M., Krek, V., Shelanski, M., and Cowan, N. J. (1984) Proc. Natl. Acad. Sci. U. S. A. 8 1 , 2743-2746

Julien, J.-P., Meyer, D., Flavell, D., Hurst, J., and Grosveld, F. (1988) Mol. Brain Res. 1, 243-250

Napolitano, E. W., Chin, S. S. M., Colman, D. R., and Liem, R. K. H. (1987) J. Neurosci. 7, 2590-2599

Shneidman, P. S., Carden, M. J., Lees, J. F., and Lazzarini, R. A. (1988) Mol. Brain Res. 4 , 217-231

Fisher, D. Z., Chaudhary, N., and Blobel, G. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 6450-6454

Gruenbaum, Y., Landesman, Y., Drees, B., Bare, J. W., Saum- weber, H., Paddy, M. E., Sedat, J. W., Smith, D. E., Benton, B. M., and Fisher, P. A. (1988) J. Cell Biol. 106, 585-596

Voller, A., Bidwell, D. E., and Bartlett, A. (1976) Bull. WHO 53,

Pruss, R. M., Mirsky, R., Raff, M. C., Thorpe, R., Dowding, A. J., and Anderton, B. H. (1981) Cell 27,419-428

Pruss, R. M. (1985) J. Neuroimmunol. 8 , 293-299 Phillips, L. L., Autilio-Gambetti, L., and Lasek, R. J. (1983)

Gallant, P. E., Pant, H. C., Pruss, R. M., and Gainer, H. (1986)

Tytell, M., Zackroff, R. V., and Hill, W. D. (1988) Cell Motil.

Tytell, M., Pant, H. C., Gainer, H., and Hill, W. D. (1990) J.

Neville, D. M., Jr. (1971) J. Biol. Chem. 246, 6328-6334

1,41-65

593-625

1534

(1988) Neuron 1,395-401

phan, K. (1988) EMBO J. 7,2995-3001

3221-3227

319,463-468

U. S. A. 80, 726-730

237, 178-182

U. S. A. 83,6475-6479

55-65

Brain Res. 278, 219-223

J. Neurochem. 46, 1573-1581

Cytoskeleton 9 , 349-360

Neurosci. Res. 25, 153-161

15038 Squid Neurofilament Proteins 39. Towbin, H., Staehlin, T., and Gordon, J. (1979) Proc. Natl. Acad.

40. Szaro, B. G., and Gainer, H. (1988) J. Comp. Neurol. 273, 344-

41. Humason, G. (1962) Animal Tissue Techniques, W. H. Freeman

42. Szaro, B. G., Whitnall, M. H., and Gainer, H. (1990) J. Comp.

43. Myers, M. W., Lazzarini, R. A,, Lee, V. M.-Y., Schlaepfer, W.

44. Bodian, D. (1936) Anat. Rec. 65,89-97 45. Autilio-Gambetti, L., Crane, R., and Gambetti, P. (1986) J.

Sci. U. S. A. 76, 4350-4354

358

and Co., San Francisco

Neurol. 302,220-235

W., and Nelson, D. L. (1987) EMBO J. 6, 1617-1626

Neurochem. 46, 366-370

46. Piatigorsky, J., and Wistow, G. J. (1989) Cell 57, 197-199 47. Chiu, F.-C., Barnes, E. A., Das, K., Haley, J., Socolow, P.,

48. Kaplan, M. P., Chin, S. S. M., Fliegner, K. H., and Liem, R. K.

49. Sharpe, C. R., Pluck, A,, and Gurdon, J. B. (1989) Deuelopment

50. Quitschke, W., Jones, P. S., and Schechter, N. (1985) J. Neuro-

51. Quitschke, W., and Schechter, N. (1986) J. Neurochem. 46,545-

52. Geisler, N., Kaufmann, E., Fischer, S., Plessmann, U., and Weber,

Macaluso, F. P., and Fant, J. (1989) Neuron 2, 1435-1445

H. (1990) J. Neurosci. 10, 2735-2748

107,701-714

chem. 44, 1465-1476

555

K. (1983) EMBO J . 2, 1295-1302

W U I O LOW MOLECULAR WEIGHT NEUROFILAMENT PROTEINS ARE ii NOVEL CLASS OF SUPPLEMENTAL MATERIAL TO:

NEUROFILAMENT PROTEIN

EXPERIMENTAL PROCEDURLS =DNA llbrarg obtained for all exoeriments at

construction in lambda gtlO are described fully elsewhere (16). -7OOC. The details of random hexami?rloliqo dT irlmed CDNA library

Screenlnq o f k c D N A libraries and assemblv of full length clones

m o u s e NF-L (15) labeled With 32P by nlck translation. Optimal Approxlmately 1 x 10' plaques were screened with a cDNA probe from

hybridization conditions were prev~ou~ly determined for this probe on RNA blots made from souid antic lobe. Double lifts with nitrocellulose filters were made from each plate. 'These filters were prehybridized for severa l hours and then hybridlzed overnight at 17'C I" a solution designed for short Oligonucleotide probes (6xSSC. 2 0 % formamide. 200mM Tris. 50 mM sodlum phosphate at pH 1 . 0 , 10% dextran sulfate. 2 . 5 Y Denhardt's. and 1 mq denatured herring Sperm DNA; probe concentration at 7.5-10 X IO5 Cpmlml; lXSRC 1s 150 mM NaC1/15 mM sodium citrate). iilters were then washed twice at room temperaturp ~n 2xSSCI0.1%SDS. followed by another low stringency wash in 2xSSC/O.1% SDS a t 48OC. Doubly positive clones were then piaque purified, subcloned into MI1 and sequenced by standard dideoxynucieatide chain termination methods 1161. Two overlapolng independent clones. 550

. . were made from each plate. 'These filters were prehybridized for severa l hours and then hybridlzed overnight at 17'C I" a solution designed for short Oligonucleotide probes (6xSSC. 2 0 % formamide. 200mM Tris. 50 mM sodlum phosphate at pH 1 . 0 , 10% dextran sulfate. 2 . 5 Y Denhardt's. and 1 mq denatured herring Sperm DNA; probe concentration at 7.5-10 X IO5 Cpmlml; lXSRC 1s 150 mM NaC1/15 mM sodium citrate). iilters were then washed twice at room temperaturp ~n 2xSSCI0.1%SDS. followed by another low stringency wash in 2xSSC/O.1% SDS a t 48OC. Doubly positive clones were then piaque purified, subcloned into MI1 and sequenced by standard dideoxynucieatide chain termination methods 1161. Two overlapolng independent clones. 550

. .

and 600 bp in length, were found to ha;; <equenc& Wlth properties characteristic of the 2 coil region of the rad domain of intermediate filaments.

=DNA insert fraqments were then labeled with I2P by random primer synthesis and used to &reen IO5 plaques from a n amplified ?base size.= 1 . 2

overnlqht at 37'C in a higher Stringency hybridization solutim (40% x l o 5 ) , > 2 kbp insert size library. Filter hybrldization was performed

formamide. SXSSC, 20 m M Tris. pH 7.4, 1 x Denhardt's 10% dextran sulfate.

were washed twice ~n 2xSSC/O.1% SDS. and then at 60'C in O.lxSSCIO.1%SDS. 20 pglml denatured herring Sperm DNA; probe at 5 x'105 CpmInIl). Filters

From this screening. 19 positive 'plaques were identlfied. The three longest (appraxlaately 2 kbp in length) Were Subcloned into PGEM4 and sequenced in both directions by standard double stranded dideoxynucleotide chain termination methods (SeqUrnaseTM klf; U . S . BiochPm. rorp.. Cleveland. OH). Sequences were detrrmlned for DNA Strands i n both

spaced 250-300 bp apart were Used to prlme successive Sequencing reactions. orlentatlons for all sequences reported here. Synthetic oligonucleotldes

These initial clones a l l began at the same 5 ' position (an

positions. TO Obtaln the 5 ' end of the codlng reqlon we used a 5"anchored incompletely methylated Eco R 1 Site). but terminated at dlfferent 3 ,

polymerase chaln reaction (17). In this method. an antisense. sequence Specific primer (CTGAGCIITTCCTGGCTTCCRC: nucleotldes 1221-1201. Fiq. 1 ) located near the 5 ' end of the available sequence was used to prime first strand cDNA synthesis from squid optic lobe polyadenylated mRNA. This single Stranded cDNA was then end-talled with dATP, and another antisense

together With 2 additional primers (GACTCGAGTCGACATCG-d'l' '

primer (GIITCAGCIIAGTCTGTTACG; nucleatides 1183-1201) nested 5' to the first,

CACTCGAGTCGACATCG). were used to prime the polymerase chain Feact\:A through 4 0 cycles. The resultant DNA was purified, size fractionated by PDlyPcrylamide gel electrophoresis and cloned into MI]. Fifteen overlapping independent clones were ,Solated by screening with a third

sequenced. Probes prepared from these sequences were used to Obtain 1 oligonucleotide probe (ATTTCACTCRTATTACTCTT; nucleotides 1181-1162; and

averlapplnq clones from the CDNA library. These clones were then subcloned

they spanned 1969 bp (NF6Oi. and included a single long open reading frame. Into PGEM 4 and sequenced in both orientations as above. When assernbled

1536 bp long. TO further verify that these assembled averlappinq sequences

represented a species of mRNA from the squid optic lobe, first strand cDNA synthesis Was primed from squid optlc lobe palyadenylated mRNA with a Sequence-speclflc antisense primer (TATCTACAGTCCCATTGACA; nucleotides 1960- 1941. F l g . 1) from the 1' untranslated reqlon of the NF60 sequence. Two additlonal primers that delimited the i'(sense; AATITCTCGCCTAATAGTTAACA; nucleotides 70-901 and 3 ' (antisense; TCCCCTTCTAATCCGATGI; nucleotides 1633-1614) ends of the coding region were Used to prime a polymerase chain

region precisely, confirming that our assembled full lenqth clone was reaction. The resultant cDNR matched the predicted length of the coding

actuallv renresentative of a mRNA soecies from souid ontic lobe.

independent clones. of lower abundance than those of the NFBO, were DLrinS multiple screenings 'of the CDNR' libraries, 2 additional

discovered. These clones (NF70) identically matched the sequence of NF60 Over the 5 ' ends through nucleotide 1 1 5 4 2 (riq.l), where they diverged (Fig. 2 ) .

RNA blot ad~alvsis of NF60170 exeresson RNA blots were used to verify that the 2 forms of NF clones Were

present in optic lobe and stellate ganglia RNA's. and to examine the tissue specificity of their expression. T o t a l RNA was prppared from optic lobes, stellate oanalia. mantle muscle. and tentacle anoaratus of freshlv caUOht

Pant, James Way, and James Battey

The followinq double stranded DNA probes were Constructed With the polymerase chain reaction and sequence specific primers. l ) ~ probe specific for the N-terminal domain shared by NF60 and NF70 (NF60170.N; nucleotLdes 131-300)- 2)A probe covering the l a coil of the rod domain. and extending into the linker between la and Ib (Nf60170,la coil; nucleotides 281-521). 1)A probe consisting of sequences extending from the

probe speclfic for the 2 coil domaln of the rod of NF60 and NF70 (NF60170. linker region into the lb coll (Nf60/70,Ib coil; nucleotides 5 0 2 - 5 8 7 ) . 4)A

the last feu nucleotides Of the C-terminal domain of NF-60 into the I' 2 coil; nucleotides 1093-1430). 5 ) A NFLO-specific DNA probe extending from

untranslated region Of Nf60 (NF60, 3 , UT; NF60 nucleotides 1602-1969). 6)A NF70-specific DNA probe directed against the divergent C-terminal coding

probes labeled by a kinase reaction were also used. 1)A probe directed domain Of NF70 (NF70.C; NF7O nucleotides 1543-2011). N o oligonucleotide

agalnst a portion of the C-terminal domain shared by NF60 and NF70 lNF60170.C: nucleotldes colnolelnentarv to 1 4 8 4 - 1 5 3 2 1 . 21A NF60-sneCifiC probe' directed against th'e 3 ' &d of its codinq 'region (ilF6O.C; complementary to NF60 nucleotides 1582 to 1620). A 1 1 fragment probes Were labelled by random prlming. and oligonucleotide probes were end-labelled using T1 polynucleotlde kinase and [qamma-'ZP]ATP.

m ~ a r a t i p anqlvsis TO determine that both NF60 and NF70 were encoded by the same gene W e

prepared genomic DNA from frozen squid optic lobes. One optic lobe per tube vas thawed and gently minced in 700 111 50 mM TIis, pH 8.0. 100 mM EDTA. 1 0 0 mI4 NaC1, 1% SDS. Proteinase K (50 pl of 10 mg/ml) was then added and tissues were digested Overnight at 55'C. DNase-free. RNase ( 3 0 P l , 20 uglml) was added and tubes incubated at 37-c for an additional hour.

hand for 10 min, followed by centrifugation in a inicrofuge for 10 min. The Samples were then mixed with an equal Volume of phenol and shaken gently by

aqueous phase was then extracted repeatedly with phena1:chloroform until the interface material was no longer visible. DNA was then precipitated With an equal volume of isopropanol. which was decanted afterward, and the DNA rinsed repeatedly with several Washes of ethanol. This DNA was then dissolved overnight in 100 pl of 10 mM Tris, pH 7.4/0.1 mI4 EUTA. The concentration of genomic DNA Was measured spectrophotometrlCally and adjusted to 0.5 mg/ ml.

An aliquot of 20 ug of squid genomic DNA was digested by Eco R1. The

denatured and transferred to nitrocellulose by capillary blotting ( 1 6 ) . sample was then split and separated on 2 lanes of an 0.8% agarose gel,

One filter vas hybridized with the NFBO/?O.N probe and the Other with the NF60170.2 coil probe. Both were washed at high stringency (60'C. 0.1XSSC) in the manner described earlier.

~ m . 3 l p k ~ filament nroteins

programs Of the University of Wisconsin Genetics Conputee Group (18). Nucleotide and amino acid sequences were analyzed using the computer

Peptide sequences of entire proteins and for the N-terminal, rod, and C- terminal donains. separately, were aligned with the GAP program (algorithm of Needleman and WunSCh (19); gap weight at 1.0, length weight at 0.10). Dotplot comparisons between nucleotide sequences were made with a window size of 21 and a stringency of 14 ( 2 0 ) . When only peptide sequences were available (i.e.. invertebrate intermediate filament proteins). dotplot

window sire of 7 (stringencies 4 and 5; 20). comparisons were made with word sires Of 2,1, and 4 (21). and also With a

The followina intermediate filament Drotein SeauenceS Were selected

W , 51 koa cytokeratin (22); Type I1 cytokeratin from oocytes Of for cornparison ta'the squid NF60 and NFiO sequencis. Type I &lX?EU2

XSwpl43 L?L!~&Z (23); Type 111: 1lrat peripherin ( 5 ) . 2lhuman vinentin (24); 3)nouse glial fibrillary acidic protein (GFAP; 2 5 ) ; Type IV: 1)mouse NF-L

2)Drasophila lamin (lo); Invertebrate cytoplasmc intermediate fllament (26); 2)rat NF-M (27). 3)moUse NF-H (28); Type V : 1)human lanin C (29);

proteins: Epithellal mtermediate filament protein A from Helix nomatia (6.7).

Antlbodv nroduction To identlfy and localize the proteins encoded by the NF60 and NFlO

CDNA c lones . antibodies were generated against a synthetic peptide (SQNF peptide; Peptide Technologies, Washington, DC) found in the C-terminal domain common to bath NF60 and Nf70 (amino acid sequence: EAEVLSTILTRSEGC). The peptide was glutaraldehyde-coupled to bovine serum albumin and injected into 3 rabbits by the Berkeley Antibody Company (Richmond, CAI, accorainq to their standard commercial protocol. Sera from each of 4 bleeds were checked for activity aqalnst the synthetic peptide by a standard ELISA immunoassay (31).

neUrOfilament proteins in the Squid were also used. The mouse lnOnoclona1 TWO previously characterized antibodies that are known to react With

antibody. alFA. is directed against an epitope common to nearly all

alpha-helical rod domain (12, 13). This antibody reacts with the 60 kDa interrnedlate filament proteins. which is located at the Carboxyl end Of the

low molecular mass squid nevrafilarnent pra,tein and several. less abundant

axoplasm (9 3 4 15 36 37). The mouse rnonocional antibody. BNFP, is intermediate filament proteins present ~n the stellate ganglion and

specific fok a bhosdhor;lated eOitoOe on the 220 kDa Squid neurofilament protein (91.

hybridization conditions Were the same as ?or low strinsency screeninq of the CDNA library, and washes in 2XSSCIO.IP SDS Yere 15-C below the predicted melting temperature (16).

Squid Neurofilament Proteins 15039 L n a l v s ~ s of intermedlate fllament DreparAtions bv Western blotting

Intermediate filament preparatlons Were made from squid OptlC lobes and from the axoplasm of the qiant axon by 2 cycles of a reversible assembly-dlsassembly procedure a5 described in Zackroff ( 1 4 ) and Stored at -2O'c. Approximately 300 pg of each intermediate filament preparation was diluted to a final volume of 900 p1 in upper gel buffer ( 3 8 ) containing 6 M urea. 1.9% SDS, 3 . 7 5 % D-mercaptoethanol, and denatured at 40°C. SDS polyacrylanide gel electrophoresls vas performed on 10 cm long Vertlcal slab gels with an acrylamide concentration of 7.5% and the buffer system of Nevllle ( 3 8 ) . Samples were loaded into a 1 2 cn vide well, adjacent to a reference w e l l containing molecular weight standards. After electrophoresis. proteins yere transferred (39) overnight (6oV. l o ° C ) Onto nitrocellulose paper ( 0 . 4 5 urn pore sire, Schleicher and Schuell. Keene, NHI. Blots Were cut into 3-5 mm strios and stained bv antibodies at

tris, pll 1.6) to dilutions of 1/500, 1 j 1 0 0 0 . 1/2000, and 1/4000. Preimnune sera taken from the rabbits prior to immunization were Used as negative COntrOlS

ImmunocYtochemlstry optic lobes and stellate ganglia were dissected intact from live

adult squid and fixed by immersion in Bouin's fixative. Tissues were then dehydrated ~n alcohols, embedded in paraffin, cut into 15 urn thlck SestlonS and mounted onto slides with album>" according to Standard histologlcal procedures (dl). After the paraffin vas removed by xylenes and the sections rehydrated. slides were processed far immunoperaxidase histochemistry exactly a s described elsewhere (42). Primary antisera from each of the 3 rabbits immunized with the SQNF peptide were diluted into blacking 501ut10n at l/lOOO. 1/2000. and 1/4000. Preimmune sera taken from each of the rabbits prior to immunization yere similarly diluted and used a 5 negative controls. The bluish black immUnOperOXldaSe reaction product was produced by a coprecipitant of diaminobenzldine and nickel chloride.

RESULTS

Identification of clones encodinq NF60 and NF70 PrevlouS information On the primary Structure of squid neurafilalnenf

prOteinS Was unavailable. but lmmunological studies using the aIFA antibody (9,121 suqqested that at least the 2 coil reqion of the rod domain would share similarities with mammalian neurofilament proteins. Thus, Ye used cDNA clones coverlnq this domain of severa l mammalian intermediate filament proteins as probes~ at very law stringencies on Northern blots of squid Optlc lobe RNA. These included a <DNA probe against mouse NF-L ( 1 5 1 . a mouse GFAP clone ( 2 5 ) . and 3 cDNA'S from human NF-M (HNF1,4,11)(43). The mouse NF-L probe qave the best signal Over background ( 3 barely visible bands between 5 kb and 9 kb; data nit shown). plaques of an unamplified library sire selected for inserts ,550 bp) with

Low stringency screening of a squid optic lobe CDNA library (IO5

the mouse NF-L =DNA probe yielded 2 overlapping, independent clones 550 and 600 bp in lenqth (SQN8-1; SQNL-I). Probes made from these two clones were used at high stringency to obtain additional clones. There clones in conjunction With overlapping CDNA clones found by using probes generated by an anchored polymerase chain reaction (Materials and Methods) encoded a =DNA approximately 3.1 kbp in length. The first 1964 nucleotides, given for the coding strand, are presented in Figure 1. The longest open reading frame extended from nucleotides X 9 1 to X1626, and encoded a protein

molecular weight of 59,160. During the library screenings, 2 additional (designated as the NF60 protein) with 511 amino acids and a predicted

independent clones. of lower abundance than those of the NF60, were

nucleotide 1 1 5 4 2 . and conLained a n open reading frame to nucleotide $1948. isolated. These clones identically overlapped the sequence of NF60 until

This sequence accounted for a second protein with 615 amino acids and a predicted molecular weight of 7 1 , 0 0 2 (NF70, Fig. 2 ) . A schematic drawing Of the assembled sequences of N860 and NF7O is shown in Figure 3 .

Figure 1. Nucleotide and predicted amino acid Sequence of the squid NF60 protein. A long open readlng frame was compiled from the sequences of two overlapDlnq =DNA's and confirmed bv additlonal indenendent clones (see

divergent C-terminal domain of the squid N F l O protein. The complete Figure 2 . The nucleotide and predicted amino acid sequence of the

sequence of NF70 was identical to Nr'60 up to the arrowhead, which marks thelr palnt of divergence (nucleotides 1542 and 1 5 4 3 ) .

SONF wp,,ae

Figure 3 . Schematic d r d + l n g of the CONi's ercoding the N F 6 0 aqd NF70 proteins. Whose Sequences were presented in Figures 1 and 2 . Codlng regions are indicated by rectangles, untranslated regions by stralght lines. The locations of the N-termlnal dornaln. the Central rod Core with

are indicated. Sequences encodlng the divergent C-terminal domains of the its different quasi-heptad repeat coiled domalns. and the C-terminal domain

proteins are indicated by the shaded regions to the riqht of the large arrowhead. 'The small arrowheads indicate the position of the SQNF peptide aqalnst which antibodies were made.

The sequences of NF60 and NF70 DredLGted two intermediate filament DrDteinS vith a lamin-like rather than neurofUanent-like rod domain

contalned motifs that helDed to establish the identities of these proteins The predicted amino acid sequences of the NF60 and NF70 proteins

as intermediate filament proteins, but demonstrated they were distinct from vertebrate neurofilament protelns and from all other Vertebrate cytoplasmic intermediate filament proteins as well . The NF60 and NF70 proteins shared a region with the characteristics of the rod domain Of intermediate filament proteins ( 2 ) . which helped to identify them as members of the intermediate filament oratein familv. However. like the rod domains of the

clearly intemebiate filament like, because it contained three regions of quasl-heptad repeats (abcdefg)" in which the a and d positions were Occupied by hydrophobic and non-polar amino acids at greater than 75a of the positions. TO illustrate this. W e have aligned the sequence Of the NF60 with human lamin C , with the aid af the "GAP" program ( 1 8 ) . Only 12 charged residues dlsrupted this heptad pattern (11%). but of these 12, 7 of them were aligned Wlth charged residues in human lamin C. In addition to this heDtad oattern. onlv two other features found in vertebrate cytoplasmic intermediaie fiiament protein sequences were also found in the NF60/70 Proteins. First. two Teaions of amino acid similaritv that define

. .

the .ami& and carboxyl termlni -of the Core domain were evident (double underlined Sequences. Flg. I ) . The carboxyl terminus of the rod domain contalned the canonical Intermediate filament protein sequence. KLLECEE. This Sequence is part of the epitope for the aIFA antibody, which recognizes intermediate filament proteins of both vertebrates and invertebrates 112, 32). This resian and the less similar domain definina the amino end of the rod dona-in, however, were the only reqions 0; significant similarity between the squid NF60J70 sequences and mouse NF-L

Proteins. Vertebrate CVtoDlalmiC intermediate filament Droteins of the (not shown) as well as Other vertebrate Cytoplasmic intermediate filament

and 392, Fig. 4.1. One additional region of amino acld similarity was seen between lamin C and the squid NF proteins at the end of the Ib coil domain

intermediate fllament proteins, including neurofilament proteins. (Fig. 4 ) . This region has been deleted from all vertebrate cytoplasmic

. .

Methodij. ~ Translation Was begun 'at the first ii phase methionine. Sequences encoding amino acids corresponding to the quasi-heptad repeat containing seqmenta of the rod domain have been underlined. The sequence corresponding to the SQNF Synthetic peptide, present in both NF60 and NF70

of the NF60 and NF7O sequences (nucleotides 1542-41) is marked by a n and used to make antibodies, is double-underlined. The paint of divergence

15040 Squid Neurofilament Proteins

51 FRSSMGSNARYTRSYOFSYGAT~PGRYAN~SSTCVNHVKANREREKQD~ loo NF60 I I l l I l l I l l I I1 I

1 .. METPSQRRATR~ ..... GAQA . . . . . . . . SSTPLSPTRITRLQEKEPL 3 5 L a m i n C . .. .. 101 R D L N E R F A N Y I E K V R F L E A Q ~ K K ~ R G E ~ E E ~ K S K : G K E T S A I K E M ~ E T E L 150

. . I I I I I1 I1 I 1 I I I I I I1 I I I I

36 O E L N a R S R E V S G I K A R Y E B E L 85

151 EEARKL~DA~NKEIKIT.SDVR:TE.LIDQ~ER~QKO?~EESRTYHQIDQEQIA 200 l a cod A db

86 nT)PIRKTI.DSVRKERARLOI.RI.RKVRRRFKF.I.KI 135 I l l I I1 I

coil 201 R ~ N ~ ~ ~ A O ~ E G E ~ S " : R R S ~ E S ~ E K E ~ M R ~ S N I ~ A K ~ D E ~ K ~ ~ ~ N N 250

I l l I I I I1 136 > 185

251 ?TIN~L&NR$QT?EEE~EF~KDVHAQELKELA. . .ALAYRDTTAENRE 297

lE6 """""""""_ 298 F.WRNEL:QA;ROI~QE~DAK:DQ~RGD~EA~NLK:QE:RTG~TK~N~: 347

A A

I I I I I I I I I l l I l l I I I KRRHETRLVEIDNGKQR 235

2 3 6 E F E S R L P 285 I I I I I I I

2 c o i l 348 T R : K E E f T K o , K s N b E f R N R t A D t E A 8 N . A O t E R T ~ Q o ~ L ~ ~ E E ~ R Q : E 3 9 1

I1 286 8 335

I I I I I I I

398 L E S C Q ~ K E E ~ T K ~ R G E ~ E S ~ L K E ~ Q ~ ~ ~ ~ I K L S ~ E ~ E I ~ Y R K ~ L E G E E S 4 4 1 e . * A

I I I I I I I I I I I I I I I I I I I I I I I I 3 3 6 e 385

448 R:GMKQIV ............... EQWGARPNEAEVLSTILTRSEGGYEA 482

I I I I I I 386 BLRLSPSPTSQRSRGRASSHSSQT~GGS~KKRKLESTESRSSFSQHAR 435

~ i q u r e 4 . Characteristic reatures of the rod domain or intermediate

of the predicted amino acid sequCnCF of the rod domaLn of squid NFGU with filament proteins 11Iustrated for the squid NF60170 proteins by aliqnment

human Lamin c. Amino acids identical between squtd NF6O and human lamin C are connccted by a Vortical har. Reqions correspondinq to the hellcal domains of human I m i n c (McKcon, et a l . , 1986) are underlined and labeled. The positions O f uncharged amino acids 01 the quasi-heptad repaat .ire indicated above tho sequ~nse of NF60 by lillrd circLcs. ThP rilled trianqles show the locations Of Chsrqcd residues in tho heptad-repeat positions t h a t were also occupied by charged residues in Inmin. Unfilled

conserved. The asterisks (NF6O amino aclds 328 and 392) mark the phase triangles mark charged amino acids in heptad positions that YET* not

changes in the 2 coil domain. which are conserved anonq all intermediate

carbouy trrmlni of the rod domain are doubly underlined. An additional filament proteins. The regions of SOnsErVEd sequence at the amino and

of the Ib coil that has been deleted from vertebrate FYtOplaSmiC reqion or hiqh similarity between squid NF6O and human lamin C. a t the end

intermediate rilsnent roteins. is underlined by a dashed line.

that the squid NF6O and NF70 proteins were only distantly related to By comparison of total amino acid identities, it vas further Seen

rcpresentst~vo examples of a l l classes or vertebrate intermediate filament proteins, including mammalian neuroriloments (Tahlr I). In addition, by total amino acid identity the squid neurorilaacnt proteins were not very closely related to nuclear lamin c despite the similarity in rod domain Icnqths. and the additional isolated reqion Of similarity In the I b Coil.

amino acid identity, the rod domain of the squid NF60/70 proteins most EYosely resembled another molluscan Cytoplasmic intermediate filament protein. the &Lis epithelial intermediate filament protein ( 3 7 % identity). IIOWEVE~. even this similarity vas not as close as that found between vertebrate Type 111 nonneuronsl and Type IV ncuronnl intermediate filamrnt proteins (GFAP t o NF-L, 49.3: identity).

28.2 11.8 1 . 4 3 ) 27.0 3 0 . 7 29.8 ( 1 4 2 ) 25.4 21.7 32.8 ( * ) - I ) 2 5 . 4

WE%.! Iturnan leain c 24.2 32.8 ( - 3 ) Drosophiia lamin

1 4 . 3 27.3 29.1 1 -3) 27.0

VembEBfe IF comxss_r&ows Mouse NF-LlMouse GFAP 30.0 4 9 . 3 ( 12) MOUSC NF-LIUnt NF-M 3 3 . 3 54.4 ( -1) Mouse GFAPlIlunan vimentin 4 1 . 5 63.8 ( 0 ) Droaophila/lluman lamin C 48.5 37.7 ( 0)

35.9 44.6 19.0 35.3

29.2

28.9

16.1

26.2 23.2 10.1

21.1 20.2 1 9 . 1

28.5 26.4

NF60/70 NF60 NF70

a- b- c-

d-

N la coil Ib coil 2coil C 3'UT C

with domain-specific =DNA probes of NF6O and NF70. Total RNA. 10 Irq per Fiqure 5. RNA blot analysis Of Optic lobs ( 0 ) and Stellate ganglia (SI UNA

lane, was resolved on formaldehyde-agarose gels, transferred to nltrocellulose and hybridized with probes specific for dlfferent regions of the NF6O and NFlO proteins (see Materials and Methods for a complete description). Prohe abbreviatlons. lelt to riqht: N (NFSOl70, N-terminal domain): l a coil (NF60/70. 117 coil); Ib coil (NF60170. Lb coil); ? coil INF60170. 2 c o i l ) ; C (NF60170. overlapping C-terminal domain); INFLO) 3' UT (NF6O specific 3' untranslated reqlonl; (tlF70) c rNF70, C-terminal domain). PrObCS spvclric ror the N-terminal ond rod domains or both NFLO and NFlO hybridized ulth all the major bands: a. at 8.5 kb; b, at 7 kb; C , at 5 kb; and d. n t 3kb. The 3 kh hand was present only in optic lobe UNA. Probes specilic for domains 3 , to the rod domain hybridized with subsets of thP5e four. The NF60/70,C-terninal probr hybridized vlth hands a and c. The NF60.3' UT probe was Specific lor h and 5 . The NF70.C terminal probe hyhridizfd vlth a and d.

uscd 2 prOhCS directed uqainst dilferent ECO U1 Fraqments. One probe Was fron thc NFhOl70 N-terminal domain and the other from the 2 coll domain (see Flq. which are Separated by two ECO R 1 sltes located in the codinq doma?) of the NFGO and NF70 clones at nucleotides 898 and 1013. Alfhouqh on Northdrn blots both probes had produced identical multibanded patterns, on the qenomic DNA b l o t . each probe hybridlzed with a s l n q l e separate band. The N-terminal probe hybridized with B sinqle band at 6.8 kbp, and the 2 coil probe hybridized With a hand at 1.5 kbp.

23.1 - 9.4 - 6.6

* 4.4

- 2 . 3 - 2.0

N 2 coil

Squid Neurofilament Proteins 15041

Optic Lobe Axoplasm

NF220-

NF70 c

NF60 *

C 8 oNFP dFA oSQNF PI oNFP olFA e S W F PI

Fiqurr n . Western blot analysls Of intermediate filament preparations made from the squid Optic lobe lleft) and the axoplasm of the qlant axon lrlqht). Intermediate filamncnt prepererions Were separated by SDS.polyncrylenide qel electrophoresis (7.5%), transferred to nitrocellulose and stained hy the aSONF pcptlde antibody. ArrOYheadS

Lva squid oneurorilmmt proteins, rcspcrtively. Abbreviations: cn. labolrd NF ? 2 0 , -70, and -60 point to tho pOS~tionS Of the 220. 70, and 60

C00moS51F bluc; aNFP. anCi-2?0 koa nfurofllancnt proteln lnonaclonal

uSONF, anti-SONF peptide antibody: PI. prrimnune serum lr9m tho ral,llrl antibody; nlFA, pan-sprciflc anti-inCernrdx.lta riloaent protcin antibody;

immonirrd w i t h the SONF peptide.