Dec 14, 1989 - and in PC 12 cells (Leonard et al., 1988; Thompson and Ziff, ...... a-internexin and NF-M. Mary Ann Gawinowicz performed the automated.
The EMBO Journal vol.9 no.3 pp.749 - 755, 1990
The predicted amino acid sequence of ao-internexin is that of a novel neuronal intermediate filament protein
K.H.Fliegner1, G.Y.Ching and R.K.H.Liem Departments of Pathology, and Anatomy and Cell Biology, Columbia University College of Physicians and Surgeons, New York, NY 10032 and 'Department of Cell Biology, New York University School of Medicine, New York, NY 10016, USA Communicated by B.Pernis
Our laboratory recently isolated and began to characterize a 66 kd rat brain cytoskeletal protein, dubbed a-internexin for its interactions in vitro with several other cytoskeletal proteins. Although a-internexin bore several of the characteristics of intermediate filament (IF) proteins, including the recognition by an antibody reactive with all IF proteins, it did not polymerize into 10 nm rldaments under the conditions tested. Here we show that the predicted amino acid sequence of a cDNA encoding a-internexin shows the latter to be an IF protein, probably most closely related to the neurofilament proteins. Northern blotting shows that a-internexin expression is brain specific, and that rat brain a-internexin mRNA levels are maximal prior to birth and decline into adulthood, while the converse is seen for NF-L, the low molecular weight neurofilament subunit, suggesting that these two proteins play different roles in the developing brain. Key words: a-internexin/cDNA sequence/intermediate filament/neurofilaments
Introduction The cytoskeleton of most eukaryotic cells includes a fibrillar network consisting of three distinct components: actin-based microfilaments, tubulin-based microtubules and intermediate filaments (IF). Though this last group comprises a heterogeneous family of proteins exhibiting pronounced tissue specificity, in most cell types-apart from the lamins of the nuclear karyoskeleton-a single type of cytosolic IF protein predominates. Thus, although epithelial cells each contain roughly equal amounts of two types of keratin protein, vimentin predominates in most mesenchymally-derived cells, glial fibrillary acidic protein (GFAP) in astrocytes and desmin in muscle cells (Steinert and Roop, 1988). The IF network in neurons is, however, more complicated. We have known for some time that neurofilaments (NFs) are composed of three proteins-designated NF-L, NF-M and NF-H for low-, middle- and high-molecular weight subunits (Hoffman and Lasek, 1975; Liem et al., 1978). A further complexity was added with the recent discovery of a 57 kd IF protein (peripherin) abundant in peripheral neurons and some central neuron populations (Portier et al., 1984; Parysek and Goldman, 1987; Landon et al., 1989) and in PC 12 cells (Leonard et al., 1988; Thompson and Ziff,
1989). (©) Oxford University Press
Protein and cDNA sequence data have shown all the IF proteins to be members of a large multigene family and to share a number of structural motifs, the most notable of which is a long (310-356 amino acids), a-helical 'rod' domain (Steinert and Roop, 1988). This highly conserved region is flanked by amino- and carboxy-terminal domains of variable length and sequence, and is subdivided by non-oa-helical 'linker' sequences of generally conserved length and location into two segments of roughly equal size (coils 1 and 2). A characteristic of the rod domain thought to be critical to the formation of higher order structures by IF proteins is the presence throughout the ca-helical regions of heptad repeats, in which the amino acids occupying the 'a' and 'd' positions usually bear hydrophobic side chains, while the other residues are frequently polar or charged. Additional conserved features of all known IF proteins include a reverse in phase of the heptad repeats in coil 2 and the presence of the intermediate filament antigen (IFA) epitope (Pruss et al., 1981), which Geisler et al. (1983) have mapped to the last (highly conserved) 20 amino acids of the rod region. Our laboratory has recently isolated and begun to characterize a 66 kd protein that co-purifies with NFs when rat optic nerve or spinal cord axons are extracted with 1 % Triton X-100 and 0.1 M NaCl (Pachter and Liem, 1985). Dubbed 'a-intemexin' for its presumed cytoskeletal interactions in vivo, this highly insoluble protein is separable from NFs on ion exchange chromatography in urea, and was found to have a pI of -5.8 when subjected to twodimensional gel electrophoresis. 1251-labeled ca-intemexin on blot overlays associated strongly with vimentin, GFAP and NF-L, but bound only weakly to NF-M and tubulin, and not at all to NF-H and several non-cytoskeletal proteins. Like the NFs, this protein is axonally transported in slow component 'a' (Monaco et al., 1985; Pachter and Liem,
1985). While recognized by the IFA antibody, ca-intemexin nevertheless did not polymerize into 10 nm filaments under conditions that reassemble the NF proteins and GFA into their respective polymers (Pachter and Liem, 1985). a-internexin was therefore classified as an intermediate filament-associated protein (IFAP; cf. Steinert and Roop, 1988). Here we show that the predicted amino acid sequence of a cDNA encoding c-internexin bears all the characteristics of an IF sequence. Moreover, the sequences shared by a-internexin and the 'novel 66 kd NF subunit' described by Chiu et al. (1989) demonstrate the identity of the two proteins.
Results Figure 1 shows the N-terminal sequences of peptides derived from the chymotryptic digestion of purified a-internexin. 749
K.H.Fliegner, G.Y.Ching and R.K.H.Liem 10
CAGCCCCACCCGGCCCCACCCCGGCAC ATG AMC TTC OGA TCA CA CAC TAC TTO TGC TCC OCC H S F C S E H Y L C S A TCC TCC TAC S S Y
OCA 0CC
W-1-s" !"IT .::;,
0
:::::-
A
TCC S
30 20 CCC AAM 0TC TOC 000 GAT aCc TCC CCC CTC TCT OCC CCC CTC 7CC 000 cCC 0 C 0 S R L S A R L S C P R X V F 40 50 COC AMC TT C AC 7CG CM TCC CTC TCC CCC AMC AAT CTG 0CC TCC ACM 0CC G S F R S Q 2 L S R S N V A S T A
amasliL
OCT A GAC
W"
7CC 7CC TCC 0CC TCC TCT C
S
AAC
N
A
*
S
60 CTC L
aCc C
CTC AGC CAA OCA O Q n ..~L A A S
GAM AAM GMA I K K
70
CTC QCC C L
CTC 0CC TAC CGC CCC CTC CCG CCC TCC A L
A
Y
R
A
R
T
N
E
Y
A
-mo
UV
CAC H
GccGAG A
e
CCG P
x
i
T
110 CAG CTG CAG COC CTC AAC GAC CCC TTC 0CC 070 TTC ATC GAA AAM 0T7 a L L N D K F A V F T F K V 120
Q . L....
S
L_ L. .
100
CAC CM 070 GM AC CM AM CC QCC CTC GM 0CC GM0CTC .
R
90 CCC ACT OCCC CCC aCc AAC CAC TAC AM ATC ATC I
so
aCc CTC CAC
.J... 1..
amm
S
T
Q
N
TCC CCC S R
R
A
140
CTC V
CCC C
L
GhM E
E
A
E
L
A
0CC 070 A
L
130 CCC R
CMG CC 0
R
150
CTC TTC L
F
CM CCC QG CT0 C AC GM CT0 CCC QCC Q
R
E
L
R
E
L
R
C
170
160
7CC CCC CQC 0CCCM ACC 070 00 GMG CCC GM CSC 070 0CC GMG C0C0 GM GM OCC M QC L K K A S S A R A 0 A L L K R D C L A Z 160 190 CCC CAA acc 0Cc GMA GMA CT0 CAM 0CA CTC CCC CCC CCC 7GC GMG GG GM C CAC Q L A R C E K E S R C R E C A V R R E E
GAM CCC 0Cc CT0 AAM E
R
A
L
K
200
CCC A
CAM CAM CCC GAC 070 OAC Q
R
Q
220
CT0 GMA AAM AAM C07 GMA TCC C0CCT07 L
K
X
X
V
S
E
L
L
240 GM 0C CCC MG C CAC GAM GMc CA C V L A Q A E E E
OCT A
Fig. 1. (A) Chymotrypsin digestion of purified ax-intemexin produced the fragments presented schematically in (B). Lanes 1-2: mol. wt markers; lane 3: cx-chymotrypsin; lane 4: 1 Ag purified, undigested ax-internexin; lanes 5-8: 5 /Ag ca-internexin treated for 5 min at 37°C with a-chymotrypsin in enzyme:substrate ratios (w/w) of, respectively, 0.02, 0.002, 0.001 and 0.0002. (B) A schematic of the fragments shown in (A) as suggested by their N-terminal peptide sequences. A scaled-up digestion of the kind shown in (A) was subjected to SDS-PAGE and transferred to polyvinylidene difluoride membrane. The desired Coomassie blue-stained bands were excised and sequenced. The amino acid sequence derived from peptide 'b' was compared to those of various IF proteins; the resulting percent homologies are shown.
AAA CCA GAC K P D
OCT AAM A
X
modification. The presence of an in-frame stop codon following Met5OO suggests that this cDNA is fully-encoding. The presumed initiating methionine is encoded by an ATG that lies in a context typical of eukaryotic translation initiation sites (Kozak, 1989). The molecular weight of the predicted protein is 55 kd and the pI -5.9. Further evidence that our cDNA encodes a complete a-intemexin polypeptide is presented in Figure 3, which shows that the 35S-labeled in vitro translation product of the 750
CT0 L
CAG 07 cCCCOC A S O A
L
MC e
CAM OCT
CG AMC CAG K R Q
ATC TC L I
R
ACC S
ATO 0G0 I
C
0
ACT T
TCG S
CAG TCG CC S A Q
CAG CTC 0 L
CCC
A
TCC
CAA
S
S
Q
A0C GMG
G CTG L
E
E
7 K
CCC A
ATC T
QCC
0CC
C
A
ACT CT0 L T
R
K
N
Y
CCA 0CC R A
320 CT ACC ATT GAM ATA GAM K I R T I E
CAG Q
ATC I
X
A
A
CGC aCC R TCC
S
A
C
CTC L
V
CAG TAT Q Y
AAM TTC CCC K F A
AC CCA GAM V _ N
CCC
E
GMA K
L
E
ACT S
GAC
CTC
AC
AA
D
L
R
N
R
230
0T7
V
R
C
C T
AMA AMC GAM X
S
E
CAC GAC H D
OCT
A
CCC A
CT0
V
A
GMG
270 TCC
TTO
GCC
E
S
L
A
AA
CAG
TAC
CCC
290
AAC
N
CT0 L
H
310
I
AT D
L
D
ATC CAC
CCC GCA CCC
C"A C07 GMA GMA CCC CAC AMC OCT GMG 0T7 K L K K R H S A K V
CT00 GM
A
250
A
340
360
210
0CC CCCC07
CCC 0CC cCC GMA CT0 GMC 0T7
260 GAM GAM TOO TAC AAM 300
AC GAA
C
AC CGM C0C0CC TTC 070 CGC CAM L A F D V R Q E
TCT
260 A
V
H
GMG E
330 AAT CAA N
CCC C
ATG GCC N A
E
V
v
R
Y
7CC CT0 S
L
350
TAC CAC GAT D Y Q
370
CCC R
CAC H
CT7 L
360
390 CCC GAA TAC CAG GAC TOO CTC AAT OTC AAA ATG OCT CTT GAC ATT GCM ATA OCA OCT TAC R E L N I I A A Q D L N V A L D Y Y X E
AC AMG
Peptide 'a' yielded a discontinuous sequence of eight amino acids, while peptides 'b' and 'c' had N-termini identical to each other but different from that of 'a'. Peptide 'b' yielded the longest sequence: 30 residues-27 of them contiguousbearing the homologies shown in Figure lb to other IF proteins; the strongest resemblance is to the NF proteins. The sequence of a cDNA cloned from a rat brain library by hybridization to a BstI -XhoI fragment of the 5' region of the a-intemexin gene is presented in Figure 2. The predicted amino acid sequence, also shown, consists of 500 residues and contains within it the peptide sequences derived from chymotrypsin-treated ca-internexin, confirming that this cDNA encodes c-internexin. Note that the stretch of 14 dATP residues at the 3' end is not preceded by a true polyadenylation signal (AATAAA), suggesting that this poly(A) sequence is genomic, and not a post-transcriptional
AAC N
CT0
D
R
X
CTG L
CT0
GAA
L
E
GCC C
400
GAA E
GAM E
ACA QCA TTT AGC ACC F T R S T
AMC S
OG G
410 TTA AGC ATC TCA OGG 0 S S L I
420
430
CT0 AAT CCA CTT CCC AAT CCC AMT TAT TTO CTC CCT CCT AMA ATC CTT ACT TCT ACA ACC L
N
P
L
P
N
P
S
Y
L
L
P
P
I
R
L
S
S
T
T
450 TCC AAA CTC TCA TCC OCT CC0 CT0 TCC C0C AMA AAC GMG GG GG GA GM CMA GAG CM 440
S
X
V
S
S
A
G
L
S
L
X
X
K
E
E
E
K
K
E
E
460
470 CMA CCC OCT TCT AAM CMAA 0TC ACT AAM AAA ACA TCC AAM CTA Gcc0AA ACT TTC CA CA K K C S T F A S X e v T X K V S G K e E
480 490 ACA CTC GMA GAA ACC 0T7 CTA TCT ACC AM AM ACC CA AAM TCA MT ATA CMlA CAM ATC K K L S K T E T X s I I V V T K T K T E ACC ACC ? T
AMT TCC AMC S
S
S
500 CAA AAA ATC Q
X
N
TMA 7CC07TC07TT a
AAA
Fig. 2. The nucleotide sequence of a--1.6 and its predicted amino acid sequence. The cDNA includes an open reading frame of 1500 nucleotides. Overlapping exonuclease III deletion clones were used to sequence both strands. The protein sequences of Figure lB are underlined, as is the sequence published by Chiu et al. (1989; beginning at Trp282).
cDNA co-localizes in two dimensions with the a-intemexin present in a rat brain cytoskeletal preparation when the two are mixed and subjected to two-dimensional gel electro-
phoresis. Comparison of the predicted amino acid sequence of a-internexin with those of four IF proteins-rat NF-M, NF-L and peripherin, and hamster vimentin (Figure 4)demonstrates that ae-internexin is also an IF protein, with the characteristic head, rod and tail regions well-established
a-lnternexin is a novel neuronal IF protein
A
F:
Fig. 3. Two-dimensional gel electrophoresis of in vitro translated a-internexin and rat brain Triton X-100 insoluble proteins. (A) Coomassie bluestained gel. (B) Autoradiogram of same gel showing the position of the in vitro translated ca-intemexin. I, a-intemexin; L, NF-L; H, hsc 70 (,Bintemexin); M, NF-M. The position of the labeled protein exactly matches that of a-intemexin in the Coomassie blue-stained gel.
for this class of molecules. Especially noteworthy for its similarity to the corresponding region of the NFs is the tail domain, which is highly glutamic acid-, lysine-, serine- and threonine-rich. Conspicuously absent from this region of a-internexin is a skein of residues-underlined in peripherin and vimentin-that is present in all the type I IFs-desmin, GFAP, peripherin and vimentin-but absent from all other IF proteins. Also underlined are a stretch of 11 residues largely shared by the head domains of a-internexin and NF-M, and another in the tail domain that is 68 % homologous between the two proteins and equidistant from the respective carboxy-termini. The residues occyping the 'a' and 'd' positions of the rod domain heptad repeats are marked in Figure 4. a-Internexin exhibits the same heptad repeat patterns as the other proteins in the figure: most of the 'a' and 'd' residues are hydrophobic; where not, they are of the same character (e.g. basic)-often, indeed, the same residue (e.g. arginine)-as in the other proteins. Finally, this heptad repeat changes phase at His343, the identical position in coil 2 as in the other IFs depicted. Table I shows the percent homology between each region of a-internexin and the corresponding domain of the other IF proteins of Figure 4. Readily seen are the low head and tail domain homologies and relatively high homology throughout the rod-especially in coil la and coil 2-typical among the IF proteins. a-internexin exhibited almost no variation in its overall homology to these four proteins. However, considerable homology between a-internexin and NF-L is seen in the linker segments. Figure 5 demonstrates that 32P-labeled a-internexin cDNA hybridizes most strongly to a band of 3.5 kb on a Northern blot of total RNA from rat brain. Although no a-internexin expression was detected in whole embryo total RNA on days 8 and 10 of gestation even after long exposure (we were unable to isolate the brains from these tiny embryos), total RNA from day E13 rat brain revealed levels -
of expression easily detectable upon overnight exposure. Maximum levels of cx-internexin mRNA are seen on day E16, with a subsequent decline into adulthood. By contrast, while NF-L mRNA is also detectable by day E13, its expression is lower on E16, with an ensuing increase into adulthood. a-internexin mRNA was not detected in adult heart, kidney, liver, lung, muscle, placenta, salivary gland and spleen (not shown). Thus this protein apears to be brainspecific. Actin mRNA amounts were fairly constant throughout the embryonic samples.
Discussion We have isolated a cDNA encoding a-internexin and found its predicted amino acid sequence to bear the characteristics of an IF protein. This protein consequently no longer belongs in the category of IF associated proteins. Peptide sequences derived from chymotryptic fragments of purified cx-internexin are identical to sequences encoded by the clone, confirming its identity. This cDNA encodes the complete protein, as is indicated both by the presence of an in-frame stop codon and by the two-dimensional co-migration of the in vitro translated protein with rat brain a-internexin. The disparity in size between the molecular weight calculated for a-internexin (55.5 kd) and that derived from SDS -PAGE of the rat protein (66 kd; Pachter and Liem, 1985) is typical of NF proteins; the calculated molecular weight of rat NF-M, for example, is only 65% of the apparent weight on SDS -PAGE (Napolitano et al., 1987). Previous work in our laboratory had suggested the possibility that a-internexin is an IF protein, because of its insolubility and its recognition by an antibody (Pruss et al., 1981) reactive with all IFs. Despite its failure to polymerize into 10 nm filaments under the conditions tested, we reasoned that partial chymotrypsin digestion-by releasing a relatively chymotrypsin-resistant rod domain of - 40 kd (as in Geisler et al., 1983)-might reveal whether a-internexin is an IF.
751
K.H.Fliegner, G.Y.Ching and R.K.H.Liem HEAD I M L P V
1 1 1 1 1
I M L P V
60 59 53 68 56
MSFGSEHYLCSASSYRKVFGDGSRLSARLSGPGASaQRSaLSRSNVASTAACSSASS
----
---------MSYTLDSLGNPSAYRRVPTETRSSFSRVSGSPSEGGFRSOSWSRGSPSTVSSSYKRSA
---------MSSFSYEPYFSTSYKRRYVETPRVHISSVRSGYSTARSAYSS------YSAPVSSSLS MPSSASTSHHSSGRRSSISSTSYRRTFGPPPSLSPGAFSYSSSSRFSSSRLLGSGSPSSSARLGSFR ------------MSTRSVSSSSYRRMFGGPGTSNRQSSNRSYVTTSTRTYSLGSLRPSTSRSLYSSS ROD
LGLGLAYR------------LAPRLAYSSAML --------VRRSYSSSSGSLM-------APRAGAL-------------PGGAYVTRSSAVRLRSSMPGV
COIL la
RLPASDGLDLSQAAARTNEY----- KI IRTNXEKQLQGLNDRFAV# - SSAESSLDFSQSSSLLNGGSGGDY KLSRSRZEEQLQGLLNDRFAG --PSLENLDLSQV-AISNDL----- KSIRTQEZKAQLQDLNDRFAS RLPS-ERLDFSMAEALNQEF----- LATRSNZXQELQELNDRFAN RLLQ-DSVDFSLADAINTEF ------ KNTRTNZEVELQELNDRFAN
I M L P V
COIL lb Ll 108 FIEKVHQLEZTQNRALEAZLAAL RQRHAEPSRVGEL FQRZLRELRGELEEASSARAQALLERDGLAE 115 YIZKVHYLEQQNKEIEAEIHAL RQKQAGHAQLGDA YDQZIRELRATLEMVNHEKAQVQLDSDHLE E 103 F IRVHELEQQNKVLEAELLVL RQKHSEPSRFRAL YEQZIRDLRLAAEDATNEKQALQGEREGLEE 114 FIEKVRFLZQQNAALRGELSQA RG--QEPARADQL CQQZLRE LRRELZLLGRSAYRVQVERDGLAE 116 YIDKVRFLFQQNKILLAELEQL KG--QGKSRLGDL YEEZMRELRRQVDQLTNDKARVEVERDNLAE
I M L P V
D)T nrvvvrver-aT nWT 'ArFnarXrnMZM'VrTl re7 174 181 D IHRLKERFZERARLRDDTZAAIRAVRKD IEESSMVKVELDKKVQSLQDEVAFLRSNHEZEVAD L LA 169 TLRNLQARYZZZVLSREDAEGRLMEARKGADEAALARAELEKRIDSLMDZIAFLKKVREEEIAEL QA 178 DLGALKQRLEEZTRKREDAZHNLVLFRKDVDDATLSRLELERKIESLMDZIEELKKLEEZZLRDLI Q\ V 180 DIMRLREKLQEZXMLQREEAZSTLQSFRQDVDNASLARLDLERKVESLQEEIAFLKKLBEDZEIQEL IQA
I M L P V
241 248 236 245 247
L1-2 SSQAAA- - - -ZVDVAVAEPDLT QIQASHITV-ERKDY-LKTDIS QIQYAQISV-ZMDVS-SKPDLS SVESQQVQQVEVEAT-VXPELT QIQEQHV-QIDVDV--SKtPDLT
COIL 2 SALRE IRAQYESLAAKNLQSAEEVYKSKFANLNEQAARSTEAI RA
TALKEIRSQLZCHSDQNMHQAZENFKCRYAKLTEAAEQNKEAItRS AALKD IRAQYZKLAAKNMQNAZNFECSRFTVLTESAAKNTDAVRA AALRD IRAQYZNIAAKNLQEAZNhYKSKYADLSDAANRNHEALRQ
AALRDVRQQYESVAAKNLQEAZZNDKSKFADLSEAANRNNDALRQ
I 304 SREZIHZYRRQLQARTIEIEGLRGANESLERQILELEERHSAEVAGYQDSIGQLZSDLRNTKS EMARH M 313 AKEZIAZYRRQLQSKSIELESVRGTKZSLERQLSDIZERHNHDLSSYQDTIQQLZNELRGTRWZMARH L 301 AKDEVSZSRRLLKAKTLEIEACRGMNEALEKQLQELEDKQNADISAMQDTINKLZNELRSTKSEMARY P 311 AKQEMNESRQRIQSLTCRVDGLRtGTNEALLROLRELZEQFALEAGGYQAGAARLZEELRQLKEEMARH V 311 AKQZSNZYRRQVQSLTCEVDAL KGTNZSLERQMREMZENFALEAANYQDTIGRLQDEIQNMAEEMARH I M L P V
372 381 369 379 379
LREYQDLLNVKMALDIZIAAYRELLZGEZTRFS LRZYQDLLNVCMALDIZIAAYRKLLZGZZTRFS LKZYQDLLNVKMALDIEIAAYRKLLZGZZTRLS LREYQELLNVKMALDIZIAT YRKLLEGXSRI S LREYQDLLNVKMALDIZIAT YRLLGZZSRI S
TAIL TSGLSISSGLNPLPNPSYLLPPRILSTTSKVSS
TFSGSITGPLYTHRQPSVTISSKIQKTKVEAPK FTSVGSITSGYSQSSQVFGRSAYSGLQSSSYLM
VPVHSFASLSLKTTVPEVEPPQDSHSRKMVLm LPLPNFSSLNLRETNLESLPLVDTHSKRTLLLK
I 438 AGLSLKKEEEEEEEEEGASKEVTKKTSKVGESFEETLEETVVSTKKTEKSTIEEITTSSSQKM M 447 LKVQHKFVEEIIEETK... (809) TOKVEE-HEETFEEKLVSTKKVEKVTSHAIVKEVTQGD L 435 SARAFPAYYTSHVQEEQSEVEETIEATKAEEAKDEPPSEGEAEEEEKEKEEGEEEEGAEEEEA...
P 445 IETRQGEKVVTESQKEQHSELDKSSIHSY V 445 TVTRDGQVINETSQHHDDLE
Fig. 4. The predicted amino acid sequence of ax-internexin (I) manually aligned with those of rat NF-M (M; Napolitano et al., 1987), rat NF-L (L; Chin and Liem, 1989), rat peripherin (P; Leonard et al., 1988) and hamster vimentin (V; Quax et al., 1983). The sequence shared by a-internexin and NF-M in the head domain is underlined, as is a region of the tail domain shared between peripherin and vimentin. Residues identical in all five proteins or in the four neuronal proteins are emboldened and italicized. Amino acids occupying the 'a' and 'd' positions of the heptad repeats are marked underneath with a (e) according to (Geisler and Weber, 1983). Dashes are inserted to optimize alignment.
Table I. Homology of ca-intemexin to other IF proteins Head Rod Coil la L 1 Coil lb L 1-2 Coil 2 Tail Overall M L P V
33 12 24 22
54 58 56 57
69 71 60 62
31 62 45 36
44 44 51 49
30 55 40 55
65 65 64 65
18 14 14 13
44 43 46 47
Various regions of a-intemexin were compared to the corresponding regions of the four other IF proteins depicted in Figure 4. Alignments used were those of Figure 4. In each case the number of identical residues was divided by the length of the shorter of the two sequences being compared.
The sequences derived from the N-termini of such peptides indeed displayed strong homology to known IF proteins (Figure 1). The predicted sequence of a-internexin reveals little about 752
its evolutionary (or, for that matter, functional) relationship IFs, if only sequence homology is considered (Table I). However, the composition of its tail domain rich in lysine, threonine, serine and especially glutamic acid-fairly demands its classification with the NFs, the tails of which are similarly composed. Moreover, the presence in the head domains of ai-internexin and NF-M of a stretch of 11 residues, 10 of which are identical in the two proteins, and a region of 68 % homology in their tail domains (Figure 4), further suggests their close relationship, as well as an explanation for the cloning of the a-internexin gene (see Materials and methods). Thus cv-internexin is probably a type IV IF, unlike peripherin, the recently discovered type III IF found primarily in cells of the peripheral nervous system, but also in select CNS cell populations (Leonard et al., 1988; Brody et al., 1989). Whether the two proteins exhibit complementary spatial patterns of expression remains to be to the other
c-lnternexin is a novel neuronal IF protein
1
2
a
4
5
Table H. Comparison of amino acid composition of the 66 kd NF (Chiu et al., 1989) with the predicted composition of a-intemexin
6
Residue
-. .
4a
ON
B
c Fig. 5. Northern blot of embryonic rat brain total RNA. (A) internexin; (B) NF-L; (C) Actin. Lane 1, embryonic day 8 (E8); lane 2, ElO; lane 3, E13; lane 4, E16; lane 5, post-natal day 0 (PO); lane 6, P4; lane 7, adult. Arrowheads, 28S and 16S rRNA. ax-internexin mRNA peaks on day E16 and declines into adulthood. NF-L is low on day E16 and thereafter increases. Neither is detectable on days E8 and ElO (whole embryo), unlike actin, which is relatively constant from E8 to adulthood. a-
seen. Elucidation of the position of a-internexin in the phylogeny of intermediate filaments depends ultimately on the characterization of its genomic organization. The finding that a-intemexin is an IF protein explains some of the properties earlier ascribed to it, such as its Triton X-100 insolubility, recognition by the IFA antibody, and its binding on blot overlays to vimentin, GFAP, NF-L and NF-M (Pachter and Liem, 1985). This last interaction was presumably mediated by the rod domains of each protein binding to that of a-internexin. Why the latter also bound to tubulin and not to NF-H are matters for speculation, but may have been artifactual. Northern blotting of rat brain total RNA demonstrates that the prenatal developmental regulation of a-internexin and NF-L are different. While the amount of a-intemexin mRNA appears to peak on or about day 16 of gestation, that of NF-L decreases at the same time, and then increases into adulthood as a-internexin declines. Whether these changes are due to alterations in mRNA stability or in transcriptional rates remains unknown. Neither NF-L nor a-internexin was detected on days 8 or 10-even after long exposure-either because there is none, or because these mRNA species represent at that time an immeasureably small fraction of whole embryo total RNA (an issue we hope to resolve by in situ hybridization). Actin mRNA levels were fairly constant throughout embryonic life; thus, the changes seen in the amounts of a-internexin and NF-L mRNA are not gel loading artifacts. The major band recognized by the a-internexin cDNA in Figure 5 had a molecular weight approximately 2 kb larger than the apparently fully-encoding cDNA. Moreover, the poly(A) region at the 3' end of the cDNA (from which reverse transcription was presumably primed) is not preceded
AlaA Val V LeuL Ile I Pro P Met M Phe F Trp W Gly G Ser S Thr T Cys C Tyr Y Asn N Asp D Asx Gln Q Glu E Glx Lys K Arg R His H
Number 57 21 59 16 9 4 10 1 26 58 22 3 11 13 16 29 25 72 97 25 45 7
ca-internexin
66 kd NF
(%)
(%)
11.4 4.2 11.8 3.2 1.8 0.8 2.0 0.2 5.2 11.6 4.4 0.6 2.2 2.6 3.2 5.9
11.3 4.5 11.6 3.2 2.3 0.83 2.1 0.26 5.83 9.67 4.5 0.68 2.3
5.0
-
14.4 19.4 5.0 9.0 1.4
-
-
6.97 -
19.63 5.5 8.0 1.7
Asx = Asn + Asp; Glx = Gin + Glu.
by a classic polyadenylation signal, and is therefore presumably exonic and not a post-transcriptional modification. Thus the major mRNA species encoding cx-internexin apparently contains a 3' untranslated region larger than the open reading frame. Such large untranslated regions have precedent (cf. Lewis et al., 1984) and may contain sequences involved in the regulation of mRNA stability [e.g. Shaw and Kamen (1986) and Casey et al. (1988)]. Chiu et al. (1989) have recently described the isolation from rat brain of a 66 kd neuronal intermediate filament protein. Although their protein and c-intemexin both migrate on two-dimensional gels to approximately the same isoelectric point as actin [compare Figure Id in Chiu et al. (1989) with Figure 8 in Pachter and Liem (1985)], Chiu et al. (1989) argue against the identity of the two proteins, based on the different pIs ascribed to actin in the two papers (5.4 versus 5.8), and on their ability to demonstrate polymerization of their protein. In addition, while recognizing that the antibody used in early immunohistochemical staining studies on a-internexin was cross-reactive with the ubiquitous (3-internexin (Napolitano et al., 1985; Green and Liem, 1989), Chiu et al. (1989) cite the differences between those studies and their own findingswhich employed a different antibody to demonstrate brain specificity-as further evidence of the distinction between the two proteins. However, the 20 amino acid peptide sequence Chiu et al. (1989) derived from their 66 kd molecule is, with the exception of a single amino acid, wholly subsumed within our predicted a-internexin amino acid sequence (the only difference being the first residue in their sequence). Moreover, the amino acid composition presented by Chiu et al. (1989) is nearly identical to our predicted composition (Table II). Thus the two proteins are almost certainly the same molecule. By deduction, then, az-internexin, like peripherin and the NFs, can polymerize, in
753
K.H.Fliegner, G.Y.Ching and R.K.H.Liem
congruence with its predicted structure. The earlier failure property (Pachter and Liem, 1985) may have resulted from irreversible denaturation and/or degradation, either of which may have been reduced by the more rapid purification scheme developed by Chiu et al. (1989). If the functions of IFs generally remain mysterious, the necessity for five neuronal IF proteins is even less clear. However, if one function of the three NF proteins is to influence axonal caliber, as has been argued (Hoffman et al., 1987), a-internexin seems unlikely to fulfill this role, since its abundance relative to the NFs diminishes during neural development. a-internexin, therefore, more likely subserves a function in the developing neuron-perhaps in axonal growth. A detailed analysis of the developmental timing of a-intemexin expression in different neuronal populations, perhaps by in situ hybridization, might allow more conclusive statements regarding its presumably special function in the developing brain. to detect this
Materials and methods Protein
sequencing ca-internexin purified according to Pachter and Liem (1985) was subjected to chymotryptic digestion at 37°C for 5 min at varying enzyme: substrate ratios, yielding several relatively stable sets of fragments. Reactions were stopped by the addition of SDS sample buffer and boiling; the products were then applied to a 7.5% polyacrylamide gel (Laemmli, 1970). Following electrophoretic transfer to a polyvinylidene difluoride (PVDF) membrane (Matsudaira, 1987; 'Immobilon', Waters) the desired Coomassie blue-stained bands were excised and subjected to an Applied Biosystems 470A gas phase sequencer equipped with on-line PTH analyzer. Isolation of clones
Approximately 2 x I05 clones from an oligo(dT)-primed X gtl 1 adult rat brain library (generously provided by A.Furley and D.R.Colman) were screened by standard procedures (Huynh et al., 1985) with a 580 bp BstXI-XhoI fragment of the cloned rat cx-intemexin gene (G.Y.Ching and R.K.H.Liem, unpublished), which had been isolated by hybridization to a 5' portion of a cDNA encoding rat NF-M. The probe was labeled with [32P]dCTP (Dupont) by nick translation (Meinkoth and Wahl, 1987). Seven positively hybridizing clones were isolated and analyzed by Southern blotting (Wahl et al., 1987). Six consisted of a single 1.6 kb EcoRI insert that hybridized to the genomic probe. One of these (c- 1.6) was chosen for sequence analysis. The seventh appeared to be an artifact of cloning. cDNA sequencing Miniprepped -i-1.6 DNA was glass powder-purified (Bio 101) from an agarose gel following EcoRI digestion. The 1.6 kb EcoRI fragment was ligated into the unique EcoRI site in the pGEM7-Zf(+) vector (Promega) and the resultant construct used to transform JM109 Escherichia coli. Restriction mapping of miniprep plasmid DNA demonstrated the presence of insert in either orientation, allowing the squencing of both strands. Samples of 20 jig of each of the oppositely oriented constructs were treated with ApaI and XbaI to produce 5' and 3' overhangs at either end of the linearized plasmid DNA. Progressive deletions were then created with Exonuclease mIusing the Erase-a-Base system (Promega). Both strands were sequenced by dideoxynucleotide chain termination (Sanger et al., 1980) according to Sequenase (United States Biochemicals) protocols, using [35S]dATP (Dupont). Sequenced samples were separated on wedge-shaped 6% or 8% polyacrylamide gels, which were fixed, dried and exposed to Kodak XAR-5 film.
Nothern
blotting Samples of 20 yg of total RNA prepared from rat brain or whole embryo (at days E8 and EIO, when brain was technically too difficult to isolate) by extraction with guanididium isothiocyanate buffer (MacDonald et al., 1987) were subjected to agarose gel electrophoresis in the presence of formaldehyde (Davis et al., 1986). The samples were then transferred and UV cross-linked to a GeneScreen (Dupont) nylon membrane. Successive hybridizations with nick-translated, 32P-labeled cDNAs fora-internexin, NF-L (Chin and Liem, 1989) and actin (Leonard et al., 1988) followed, with prehybridization, hybridization, and wash steps carried out according 754
to GeneScreen protocols (Dupont) at high stringency. Hybridized probe was detected by exposure of the membrane to Kodak XAR-5 film at -70°C. Between hybridizations, the membrane was stripped of probe in boiling 10 mM Tris, pH 7.5, 1 mM EDTA, 1% SDS for 30 min.
Two-dimensional gel electrophoresis The pGEM-x-1 .6 plasmid was employed for in vitro transcription (Melton et al., 1984) by T7 RNA polymerase (Promega Biotech.). The resultant capped mRNA was used for in vitro translation in the presence of [3 S]methionine (ICN Pharmaceuticals) using a rabbit reticulocyte lysate (Dupont). The translation products were subjected to two-dimensional gel electrophoresis (O'Farrell, 1975) in the presence of a Triton-insoluble cytoskeletal preparation from rat brain (Pachter and Liem, 1985). A 4: 1 ratio of pH 5-8 to pH 3.5- 10 ampholines was used in the isoelectric focusing step. The Coomassie blue-stained gel was dried and exposed to Kodak XAR-5 film to yield the autoradiogram shown in Figure 3.
Acknowledgements We thank Drs Andrew Furley and David Colman for their generous gift of a cDNA library. Thanks also to Steven Chin, David Green, and Akiko Iwaki for their helpful advice, and to Ray Manson for photography. Dr Gerry Shaw graciously indicated the tail domain homology between a-internexin and NF-M. Mary Ann Gawinowicz performed the automated peptide sequencing and analyzed the resultant chromatograms; we thank her very much. This work was supported by National Institutes of Health grant Ns15182 (Jacob Javits Neuroscience Investigator Award). KHF is a Medical Scientist Training Program trainee at New York University School of Medicine, 1985 -present.
References Brody,B.A., Ley,C.A. and Parysek,L.M. (1989) J. Neurosci., 9, 2391-2401. Casey,J.L., Hentze,M.W., Koeller,D.M., Caughman,S.W., Rouault,T.A., Klausner,R.D. and Harford,J.B. (1988) Science, 240, 924-928. Chin,S.S.M. and Liem,R.K.H. (1989) Eur. J. Cell Biol., 50, 475-490. Chiu,F.-C., Barnes,E.A., Das,K., Haley,J., Socolow,P., Macaluso,F.P. and Fant,J. (1989) Neuron, 2, 1435-1445. Davis,L.G., Dibner,M.D. and Battey,J.F. (1986) Basic Methods in Molecular Biology. Elsevier, New York. Geisler,N. and Weber,K. (1983) EMBO J., 2, 2059-2063. Geisler,N., Kaufmann,E., Fischer,S., Plessmann,U. and Weber,K. (1983) EMBO J., 2, 1295-1302. Green,L.A.D. and Liem,R.K.H. (1989) J. Biol. Chem., 264, 15210-15215. Hoffman,P.N. and Lasek,R.J. (1975) J. Cell Biol., 66, 351-366. Hoffman,P.N., Cleveland,D.W., Griffin,J.W., Landes,P.W., Cowan,N.J. and Price,D.L. (1987) Proc. Natl. Acad. Sci. USA, 84, 3472-3476. Huynh,T.V., Young,R.A. and Davis,R.W. (1985) In Glover,D.M. (ed.), DNA Cloning: A Practical Approach. IRL Press, Oxford, Vol. I, pp. 49-78. Kozak,M. (1989) J. Cell Biol., 108, 229-241. Laemmli,U.K. (1970) Nature, 227, 680-685. Landon,F., Lemonnier,M., Benarous,R., Huc,C., Fiszman,M., Gros,F. and Portier,M.-M. (1989) EMBO J., 8, 1719-1726. Leonard,D.G.B., Gorham,J.D., Cole,P., Greene,L.A. and Ziff,E.B. (1988) J. Cell Biol., 106, 181-193. Lewis,S.A., Balcarek,J.M., Krek,V., Shelanski,M. and Cowan,N.J. (1984) Proc. Natl. Acad. Sci. USA, 81, 2743-2746. Liem,R.K.H., Yen,S.-H., Salomon,G.D. and Shelanski,M.L. (1978) J. Cell Biol., 79, 637 -645. MacDonald,R.J., Swift,G.H., Przybyla,A.E. and Chirgwin,J.M. (1987) Methods Enzymol., 152, 220-227. Matsudaira,P. (1987) J. Biol. Chem., 262, 10035-10038. Meinkoth,J.L. and Wahl,G.M. (1987) Methods ,Enzymol:, 152, 91-94. Melton,D.A., Kreig,P.A., Rebagliati,M.R., Maniatis,T., Zinn,K. and Green,M.R. (1984) Nucleic Acids Res., 12, 7035-7056. Monaco,S., Autilio-Gambetti,L., Zabel,D. and Gambetti,P. (1985) Proc. Natl. Acad. Sci. USA, 82, 920-924. Napolitano,E.W., Chin,S.S.M., Colman,D.R. and Liem,R.K.H. (1987) J. Neurosci., 7, 2590-2599. O'Farrell,P.H. (1975) I. Cell Biol., 250, 4007-4021. Pachter,J.S. and Liem,R.K.H. (1985) J. Cell. Biol., 101, 1316-1322. Parysek,L.M. and Goldman,R. (1987) J. Neurosci., 7, 781 -791.
a-lnternexin is a novel neuronal IF protein Portier,M.-M., deNechaud,B. and Gros,F. (1984) Dev. Neurosci., 6, 335 -344. Pruss,R.M., Mirsky,R.M., Raff,C., Thorpe,R., Dowding,A.J. and Anderton,B.H. (1981) Cell, 27, 419-428. Quax,W., Egberts,W.-V., Hendriks,W., Quax-Jeuken,Y. and Bloemendal,H. (1983) Cell, 35, 215-223. Sanger,F., Coulson,A.R., Barrell,B.G., Smith,A.J.H. and Roe,B.A. (1980) J. Mol. Biol., 143, 161-178. Shaw,G. and Kamen,R. (1986) Cell, 46, 659-667. Steinert,P.M. and Roop,D.R. (1988) Annu. Rev. Biochem., 57, 593-625. Thompson,M.A. and Ziff,E.B. (1989) Neuron, 2, 1043-1053. Wahl,G.M., Meinkoth,J.L. and Kimmel,A.R. (1987) Methods Enzvmol., 152, 572-581. Received on October 27, 1989; revised on December 14, 1989
755
