Proc. Nati. Acad. Sci. USAVol. 78, No. 5, pp. 2757-2761, May 1981Biochemistry

Complete amino acid sequence of a-tubulin from porcine brain(sequence microheterogeneity/homology with muscle proteins)

H. PONSTINGL, E. KRAUHS, M. LITTLE, AND T. KEMPFInstitute of Cell and Tumor Biology, German Cancer Research Center, D-6900 Heidelberg, Germany

Communicated by Hans Neurath, January 26, 1981

ABSTRACT The amino acid sequence of a-tubulin from por-cine brain was determined by automated and manual Edman deg-radation of eight sets of overlapping peptides. It comprises 450residues plus a COOH-terminal tyrosine that is present only in15% of the material. A region of 40 residues at the COOH-ter-minus is highly acidic, mainly due to 16 glutamyl residues. Thishigh concentration ofnegative charge suggests a region for bindingcations. At least six positions, most of them around position 270,are occupied by two amino acid residues each. Several of theseexchange sites were assigned to specific peptides by analysis of thepurified corresponding fragments. These data indicate four a-tu-bulins in porcine brain. Although a-tubulin on the whole is un-related to other proteins, there are regions that can be correlatedto sequences of the myosin head, to actin, to tropomyosin, and totroponins C and T.

Tubulins occur in all eukaryotic cells as the constituents of mi-crotubules, which participate in cell division, intracellulartransport and secretion processes, ciliary and flagellar move-ment, morphogenesis, and cell orientation. Tubulins fromwidely differing species and cell types appear to be remarkablysimilar regarding composition, molecular weight, binding ofcytostatic and psychopharmacological drugs, immunologicalcrossreactivity, and capacity to copolymerize. Yet even withinone cell, there are several types of microtubules that have dif-fering stabilities and assemble into distinct organelles at varioustimes. Knowledge of the primary structure should clarifywhether there is just one tubulin for all functions or whetherthere exists a family of similar proteins. It will also facilitatemapping of binding sites for various ligands, production of an-tibodies to well-defined antigenic sites, matching of proteinstructure with that of messengers and genes, and investigationof functionally defective tubulin mutants. Comparison of thestructure with those of known proteins may give hints for ex-periments regarding tubulin function.

Tubulin in solution is assumed to exist as a heterodimer oftwo chains, a and f, each with a molecular weight4of 50,000,and very similar amino acid compositions. Yet functional dif-ferences have been reported. For example, only a-tubulin(from blood platelets) binds cyclic AMP (1) and only /3-tubulinbinds exchangeable GTP (2). Here we present the sequence ofthe a-chain from porcine brain and report on the general strat-egy used.

MATERIALS AND METHODSWe have purified tubulin from porcine brain by a modificationof the methods used by Eipper (3) and by Luduena et al. (4).The 100,000 X g brain supernatant in 0.05 M sodium pyro-phosphate buffer (pH 7.0) was incubated with 0.1 mM colchi-cine for 15 min at 37°C before chromatography on DEAE-cel-

lulose with a linear gradient of 0.1-0.3 M sodium chloride.Tubulin was identified by the fluorescence of its complex withcolchicine (5). The preparation was reduced, alkylated with io-doacetic acid, and assayed for protein impurities by disc gelelectrophoresis in the system of Yang and Criddle (6) using 8%polyacrylamide gels. The gels were stained with Coomassie blueand scanned in a Vernon scanner. Only tubulin of more than95% purity was processed further.

For separation of a- and f-chains, the protein was chroma-tographed on hydroxyapatite in 0.1% NaDodSO4with a lineargradient of 0.2-0.4 M sodium phosphate (7). Fractions wereassayed for purity by gel electrophoresis as above. Only a-chainof at least 95% purity was used for sequence determination asdescribed (7, 8).

To remove NaDodSO4 the protein was extensively dialyzedagainst 1 mM ammonium bicarbonate; the solution was thenconcentrated by vacuum evaporation, brought to pH 5.5 withacetic acid, and treated with 9 vol of ice-cold acetone. The su-pernatant was discarded after 2 hr at -20TC, and the precipitatewas dissolved in dilute ammonium hydroxide and dialyzedagainst 0.01 M ammonium bicarbonate for enzymatic digestion,which, in all cases, was done at pH 8.0 with 1-4 mg a-tubulinper ml and, usually, an enzyme/substrate ratio of 1:100 at 370C.a-Tubulin (50-100 mg) was digested with either thrombin(Sigma); affinity-purified trypsin (a gift from K.-D. Jany, Stutt-gart) (9); chymotrypsin (Merck); or protease from Staphylococ-cus aureus (Miles) (EC 3.4.21.19), from Astacus leptodactylusEsch. (EC 3.4.99.6), donated by R. Zwilling (Heidelberg) (10),from Pseudomonasfragi (EC 3.4.24) (a gift from G. Drapeau,Montreal) (11), or from mouse submaxillary glands (EC 3.4.21)(Boehringer Mannheim). Cleavage times and exceptions fromthe general schedule were chymotrypsin, 3 hr; trypsin, 7 hr;thrombin, 7 hr; submaxillary protease, 24 hr; staphylococcalprotease/0.2 M ammonium bicarbonate at an enzyme/sub-strate ratio of 1:50, 24 hr; Astacus protease/0. 1 M ammoniumbicarbonate at 20'C and an enzyme/substrate ratio of 1:50, 2hr; protease ofa Pseudomonasfragi mutant/0.01 M ammoniumbicarbonate/2 M urea, 24 hr. For cleavage with cyanogen bro-mide (Serva, Heidelberg) the acetone precipitate was evapo-rated under reduced pressure, and the residue was dissolvedin pure formic acid, diluted to 70%, and cleaved with a 150-foldexcess of CNBr over methionyl residues for 24 hr in the dark.The product was lyophilized.

The digests were fractionated on Sephadex G-50 and G-100in 8 M urea/0. 1 M ammonium bicarbonate, and the fractionswere desalted on Sephadex G-10. Peptides were further sep-arated by chromatography on DEAE-cellulose, Dowex 1 x 2and 50 x 2, cellulose thin layers, and, more recently, by re-versed-phase high-pressure liquid chromatography with a DuPont 850 liquid chromatograph on a Zorbax C-8 column, using0.05 M ammonium bicarbonate brought to pH 7.5 with aceticacid and 0-60% acetonitrile gradients at 400C.Amino acid analyses were performed on a Durrum D-500

analyzer. Automated Edman degradations used the Beckman2757

The publication costs ofthis article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

June

22,

202

1

2758 Biochemistry: Ponstingl et al.

890 C sequencer with 0. 1 M quadrol as buffer and a single cleav-age program adapted from Brauer et al. (12). To reduce peptidelosses by extraction, 3 mg of Polybrene and 200 jig of glycyl-glycine were applied to the cup and subjected to three cyclesof degradation prior to analyzing the sample (13). Phenylthio-hydantoin derivatives of amino acids were identified by high-pressure liquid chromatography (14) and, in some cases, by ad-ditional thin-layer chromatography (15); both methods alsoserved for the assignment of amides. Manual Edman degra-dation plus dansylation was performed as described (16).

RESULTS AND DISCUSSIONThe sequence of the 450 amino acid residues of porcine braina-tubulin (Mr =50,000, depending on the variant), is given inFig. 1. It is consistent with the amino acid composition and wasestablished from the eight sets of peptides generated by cyan-ogen bromide, trypsin, chymotrypsin, staphylococcal protease,the less-frequently used thrombin and mouse submaxillaris pro-tease, and by two enzymes that may not have been used beforein sequence studies, one recently isolated from a mutant ofPseudomonasfragi, which specifically cleaves at the NH2-ter-minal side of aspartyl groups, and the other a protease from thedigestive tract of the crayfish Astacus leptodactylus Esch.,cleaving preferentially at the NH2-terminal side of alanine, gly-cine, threonine, and serine.

Tubulin peptides strongly aggregate in Solution, as does theparent molecule; hence it was necessary to include 8 M urea

in all peptide separations. This limited the types of separationmethods that could be used, resulted in loss of insoluble andsmall peptides on desalting by gel filtration, and led to partialblockage of a- and E-amino groups by cyanate from decompos-ing urea, producing heterogeneous fragments in low yields inany further digestion or purification. Therefore, we abandonedsubdigestions of peptides and chose to work with a larger num-ber of overlapping primary fragments. A summary of the frag-ments generated for sequence analysis is given in Fig. 2.A striking feature of the sequence is the COOH-terminal

region, which we already have discussed in detail (17). The last66 residues are entirely devoid of asparagine, glutamine, thre-onine, cysteine, proline, and isoleucine, and the last 40 posi-tions have 47% acidic side chains, 16 glutamic and three as-partic, rendering this segment one of the most acidic known.Its high content of glutamyl residues suggests that it may beresponsible for binding cations, for instance Ca2 , or for thebasic microtubule-associated proteins, which play antagonisticroles in microtubule assembly in vitro (for review, see ref. 18).This part is predicted to have a helical structure, quite differentfrom the rest of the chain.

In agreement with x-ray data, no indications were found fora sterical organization in domains-e.g., there are no major se-quence repeats and the 12 cysteines are spaced unevenly, fourcysteinyl together with two methionyl residues forming a prom-inent "sulfur" cluster at residues 295-316. Some other aminoacids also show a highly asymmetric distribution: although po-sitions 55-135 and 288-378 are devoid of serine with the ex-

25MET-ARG-GLU-CYS-ILE-SER-ILE-HIS-VAL-GLY-GLN-ALA-GLY-VAL-GLN-ILE-GLY-ASN-ALA-CYS-TRP-GLU-LEU-TYR-CYS-

50LEU-GLU-HIS-GLY-ILE-GLN-PRO-ASP-GLY-GLN-MET-PRO-SER-ASP-LYS-THR-ILE-GLY-GLY-GLY-ASP-ASP-SER-PHE-ASN-

75THR-PHE-PHE-SER-GLU-THR-GLY-ALA-GLY-LYS-HIS-VAL-PRO-AXG-ALA-VAL-PHE-VAL-ASP-LEU-GLU-PRO-THR-VAL-ILE-

100ASP-GLU-VAL-ARG-THR-GLY-THR-TYR-ARG-GLN-LEU-PHE-HIS-PRO-GLU-GLN-LEU-ILE-THR-GLY-LYS-GLU-ASP-ALA-ALA-

125ASN-ASN-TYR-ALA-ARG-GLY-HIS-TYR-THR-ILE-GLY-LYS-GLU-ILE-ILE-ASP-LEU-VAL-LEU-ASP-ARG-ILE-ARG-LYS-LEU-

150ALA-ASP-GLN-CYS-THR-GLY-LEU-GLN-GLY-PHE-SER-VAL-PHE-HIS-SER-PHE-GLY-GLY-GLY-THR-GLY-SER-GLY-PHE-THR-

175SER-LEU-LEU-MET-GLU-ARG-LEU-SER-VAL-ASP-TYR-GLY-LYS-LYS-SER-LYS-LEU-GLU-PHE-SER-ILE-TYR-PRO-ALA-PRO-

200GLN-VAL-SER-THR-ALA-VAL-VAL-GLU-PRO-TYR-ASN-SER-ILE-LEU-THR-THR-HIS-THR-THR-LEU-GLU-HIS-SER-ASP-CYS-

225ALA-PHE-MET-VAL-ASP-ASN-GLU-ALA-ILE-TYR-ASP-ILE-CYS-ARG-ARG-ASN-LEU-ASP-ILE-GLU-ARG-PRO-THR-TYR-THR-

250ASN-LEU-ASN-ARG-LEU-ILE-GLY-GLN-ILE-VAL-SER-SER-ILE-THR-ALA-SER-LEU-ARG-PHE-ASP-GLY-ALA-LEU-ASN-VAL-

ILE HIS TH'Y-L- 275ASP-LEU-THR-GLU-PHE-GLN-THR-ASN-LEU-VAL-PRO-TYR-PRO-ARG-ALA- IE PHE-PRO-LEU-ALA- HR-TYR ALA.PRO-VAL-GLYILE ~~ARG PHE ASXGLY ~~~~~~~~~300

ILE-SER-ALA-GLU-LYS-ALA-TYR-HIS-GLU-GLN-LEU-SER-VAL-ALA-GLU-ILE-THR-ASN-ALA-CYS-PHE-GLU-PRO-ALA-ASN-325

GLN-MET-VAL-LYS-CYS-ASP-PRO-ARG-HIS-GLY-LYS-TYR-MET-ALA-CYS-CYS-LEU-LEU-TYR-ARG-GLY-ASP-VAL-VAL-PRO-350

LYS-ASP-VAL-ASN-ALA-ALA-ILE-ALA-THR-ILE-LYS-THR-LYS-ARG- ILE-GLN-PHE-VAL-ASP-TRP-CYS-PRO-THR-GLY-SER375

PHE-LYS-VAL-GLY-ILE-ASN-TYR-GLU-PRO-PRO-THR-VAL-VAL-PRO-GLY-GLY-ASP-LEU-ALA-LYS-VAL-GLN-ARG-ALA-VAL-400

CYS-MET-LEU-SER-ASN-THR-THR-ALA-ILE-ALA-GLU-ALA-TRP-ALA-ARG-LEU-ASP-HIS-LYS-PHE-ASP-LEU-MET-TYR-ALA-425

LYS-ARG-ALA-PHE-VAL-HIS-TRP-TYR-VAL-GLY-GLU-GLY-MET-GLU-GLU-GLY-GLU-PHE-SER-GLU-ALA-ARG-GLU-ASP-MET-450

ALA-ALA-LEU-GLU-LYS-ASP-TYR-GLU-GLU-VAL-GLY-VAL-ASP-SER-VAL-GLU-GLY-GLU-GLY-GLU-GLU-GLU-GLY-GLU-GLU-(TYR)

FIG. 1. Amino acid sequence of a-tubulin from porcine brain. Positions 265, 266, 271-273, and 340 are heterogeneous. The COOH-terminaltyrosine is present in only 15% of the material.

Proc. Natl. Acad. Sci. USA 78 (1981)

Dow

nloa

ded

by g

uest

on

June

22,

202

1

Proc. Natl. Acad. Sci. USA 78 (1981) 2759

RESIDUE NUMBER 100 200

T a 0 W

B E

V L- Iu Vi///IILz D

300

m miffyOMMMM/ FMS //F/ //g N//

m~ em 6EI::]

D_ ,,,,,,, I n Th.V~~~ ~ ~ ~~~~~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~If VZX^, ]

l u MI

E/7/Y1//1, INTACT CHAIN

FIG. 2. Summary of fragments generated for the sequence analysis of a-tubulin. The hatched section of each bar indicates the portion of thesequence determined. Peptides were generated by trypsin (T); cyanogen bromide (B); protease from Staphylococcus aureus V8 (V), from mouse sub-maxillary glands (S), chymotrypsin (CH), protease from Astacus (A), thrombin (TO), and a protease from a mutant ofPseudomonas fragi (F).

ception of a single residue in a variant, 10 serine residues arerather regularly spaced between positions 136 and 198. Posi-tions 43-45 and 142-148 carry clusters of glycines, suggestingthat these areas may be flexible regions, while positions 163-231and 247-309 are devoid of this frequent amino acid in tubulin,whose abundance may be responsible for the low amount ofsecondary structure.

Although tubulin has been reported to be present in orclosely associated with membranes (19, 20), there are no regionsof the sequence that are predominantly hydrophobic. Also tu-bulin, its isolated a-chain, and most of the peptides obtainedfrom digests were fairly soluble in aqueous solution at pH =7.5.Thus, the tendency for tubulin and its fragments to aggregateand the interaction of tubulin with membranes may be due toionic forces.One possible way ofregulating tubulin assembly is posttrans-

lational modification of side chains. So far, however, we havenot detected any modified amino acids. An additional COOH-terminal tyrosine is present in 15% of our material (17) and,recently, a ligase has been isolated from porcine brain (21),which specifically adds this residue to the COOH-terminalglutamate.

There is no evidence for a carbohydrate moiety, nor for y-carboxyglutamic acid. Also, not having any radioactive label inour material, we did not detect any phosphorylated residues.

Microheterogeneity. The establishment ofthe sequence wasimpeded by microheterogeneity in several positions. Althoughthe electrophoretic homogeneity of the starting material madethe presence ofimpurities unlikely, it was nevertheless possiblethat the preparation contained similar peptides derived fromdifferent regions of the protein or that incomplete degradationhad resulted in the presence of more than one residue in a po-sition. The first possibility could be excluded by extensive over-lapping and the second by separating variant peptides by high-pressure liquid chromatography and analyzing the homogene-ous fractions. A total of at least six positions carry amino acidexchanges (Fig. 3) and most of them are concentrated in a "hotspot" around position 270. Analyses of homogeneous fractionsof these variant peptides allowed unambiguous identificationof the residues at several exchange sites. Other peptides in thesame area and around residue 160, however, were found to yieldheterogeneous degradation products at one position. Discus-sion of them is omitted from this paper. As a rule, we found a

mixture of two amino acid residues in a given position. How-ever, position 265 appears to have three-isoleucine, glycine,and alanine-in four different linkage groups. Hence at leastfour different a-chains may be present in our preparation.

Most of these exchanges can be explained by a single basesubstitution in the codon except that the isoleucine to glycine,and isoleucine to alanine at 265 and isoleucine to histidine at266 each require two base changes.

Although this heterogeneity might be due to alleles, it mayalso reflect the presence of different tubulins in different celltypes of the brain-e.g., nerve and glia cells. An organ-specific/3-tubulin has already been described in Drosophila (22). Al-ternatively, more than one a-tubulin may be required evenwithin one cell.

Secondary Structure Prediction. We have tried to predictthe secondary structure of a-tubulin according to Chou andFasman (23). a-Tubulin appears to be rather irregularly folded:only 26% of the chain is predicted to be helical and 33% is pre-dicted to have a ,3-sheet conformation, which is similar to resultsofearlier circular dichroism studies with native a- and ,3-tubulinfrom pig (22% a, 30% /3) (24) and calfbrain (26% a, 47% ,3) (25).

All major helix potentials reside in the COOH-terminal halfof the chain around residues 275-291 and the three helices atthe COOH-terminus already reported in an earlier paper (17),residues 383-403, 413-435, and 440-450. Major /8-sheet re-gions are expected at positions 49-94 (five strands), 169-195(three strands), and 223-239 and 340-378 (four strands). A seriesof overlapping turns are predicted at positions 31-49 and139-149.

In these regions, there is only one position with well-docu-

270ILE-HIS-PHE-PRO-LEU-ALA-THR-TYR-ALA

GLY-HIS-PHE-PRO-LEU-ALA-THR-TYR

GLY-ILE-PHE-PRO-LEU-ALA-ARG-PHE-ASX

ALA-HIS-PHE-PRO-LEU-ALA- X -PHE-ASX

FIG. 3. Assignment ofsequence variants around position 270. Sep-arate peptide fractions were degraded and yielded homogeneoussequences.

S

CH

400

A E

TO-

F

V/v ml I I

Q 0 mm, E0I --I

F//7//7///7//-

Biochemistry: Ponstingl et al.

OR ROMP mmff-_. 0 0rl- rg ra 0m 0m

El EM 0 gm 0 M E:::]V/Z//- M-/l EJ 0 % rg

Dow

nloa

ded

by g

uest

on

June

22,

202

1

2760 Biochemistry: Ponstingl et al.

mented microheterogeneity-the threonine to serine exchangeat position 340-presumably at the beginning of a strand of f3-sheet that is not likely to be greatly influenced by this substi-tution. The other amino acid substitutions are located in areaswith less clearcut structural potentials; thus, their effect on therespective structures would be difficult to predict.

The known lability of the tubulin molecule, as measured byits capacity to polymerize and to bind colchicine, may be ex-plained by its low level of secondary structures, as predictedby this model.

Tubulin Peptides from Other Sources. Some sequence in-formation for tubulin peptides from other sources has been re-ported: Comparison with previous data on the NH2-terminal25 residues ofchicken brain a-tubulin (26) shows six differences.Residues 10, 13, and 17 have been identified as threonine inchicken brain a-tubulin and are glycine in the porcine protein.Cysteine residues are present in porcine a-tubulin at positions4, 20, and 25 whereas, in the chicken protein, the residues atthese positions have been tentatively identified as serine. Res-idues 22-36, 303-313, 389-398, and 414-425 correspond tounlocated fragments and residue 426-450 corresponds to theCOOH-terminal cyanogen bromide peptide isolated from calfbrain and sequenced by Lu and Elzinga (27). However, twoother peptides designated a by these authors have no counter-part in our a-sequence but resemble porcine /3-tubulin (un-published observations).

Homology to Other Proteins. Several conflicting hypotheseshave been advanced to explain microtubule function in intra-cellular transport. It has been suggested that microtubules arepassive skeletons or pointers for directional movement, providescaffolds to which force-generating molecules are attached, oreven actively function as motors ofmovement. Although knowl-

edge of a structure alone does not explain function, it forms abasis on which to tackle functional problems, and comparisonof protein structures may suggest further experiments.

a-Tubulin on the whole is unrelated to any other known pro-tein, but some parts of the sequence appear to be variations ofknown motifs. Because several regions of a-tubulin resembleareas of various proteins, we can not assume a genetic relation-ship. More likely, the similarities indicate that a given func-tion-for example, binding and hydrolysis of a nucleotide-canbe performed by a limited number of similar structures. Belowwe give a few examples (Fig. 4).

Four regions of a-tubulin are similar to actin sequences (28)and, with one exception, they are in the same order in bothproteins. Between 32% and 70% ofthe residues in these regionsare identical, comparable with the similarity of a- and 13-tu-bulin. The first of these segments includes a thrombic cleavagesite in a-tubulin and in actin.A particularly interesting relationship exists between a-tu-

bulin positions 192-238 and a fragment from the globular headofmyosin (29). This segment ofthe myosin heavy chain includestwo cysteines, whose alkylation modifies the ATPase activity ofmyosin (31). The head ofmyosin can form a crossbridge betweenthe thick and thin filaments by attaching to an actin molecule.The reaction between actin and myosin is cyclical, and eachcycle includes the hydrolysis of one molecule ofATP. Residues1-46 of this fragment appear to be similar to positions 192-238in a-tubulin. In particular, the cysteine SH(1), which can bealkylated in the absence of bound nucleotides with the resultthat the Ca2+-ATPase activity is stimulated, resembles the cys-teine-213, and the SH(2), which can be alkylated in the presenceof ADP, occupies a position comparable with cysteine-200 ina-tubulin. Myosin alkylated at both sulfhydryl groups is devoid

a 57- 68 G A G K H V P R A V F VACTIN 21- 32 F A G D D A P R A V F P

a 95-127 G K E D A 'A N Y A R G H Y[T I G K E I I D L V L D RIR K L A DACTIN 289-321 R K D L Y A N N V M S G G T T M Y P G T A D R M Q K E I T A L A P

a 239-254 T H S L F D IG A L N V D L T EACTIN 173-187 H A I ML L L A G R - D L T D

a 299-312 A iQ M V K C DlP R H GHKYACTIN 278-291 Y N S I M K C D I D I R K D

a 191-215 T H T T L E H S - D[ A F M V D Nf A I Y D I C R RMYOSIN 1- 23 E H E L V L H Q L RU- - N G V LWE GW- RLI CRL KHEAD

a 216-239 N L D I E R P T T N L-N[ L I G N I V S SHTMYOSIN 24- 47 G F P - S[R I LIJ A D F K Q Y K V L N A S A II PHEAD

a 430-448 K Y|E E VIG V D S VGEG E E E|GTNT 1- 16 s - |E E V - - E HV E E E AIE E EIA

FIG. 4. Homology of a-tubulin with actin (28), a fragment of the myosin head (29), and troponin T (30) from rabbit skeletal muscle. A, Ala; B,Asx; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y,Tyr; Z, Glx.

Proc. Natl. Acad. Sci - USA 78 (1981)

Dow

nloa

ded

by g

uest

on

June

22,

202

1

Proc. Natl. Acad. Sci. USA 78 (1981) 2761

of ATPase activity. Although direct participation of this regionin actin or nucleotide binding has not been proven, the evidencesuggests that they are at or near the catalytic site for myosinATPase. No secondary structure could be predicted for the re-gion between SH(2) and SH(1), which is also the case for thecorresponding tubulin sequence. Sulfhydryls are essential fortubulin polymerization, and blockage ofas few as one or two SHgroups inhibits the assembly of microtubules by an as yet un-determined mechanism (32, 33).The highly acidic COOH-terminal part of a-tubulin resem-

bles the NH2-terminal sequence of troponin T (see Fig. 4) (30).This protein, as a component of the troponin complex, partic-ipates in the Ca2+ regulation ofactin-myosin contacts. One mayspeculate that these similar structures ofa-tubulin and troponinT perform analogous physiological functions that, in view of theclusters of glutamyl residues, could involve cation binding.

In addition to the sequence similarity to troponin T, somesimilarities have been observed to the structures of a-tropomy-osin and troponin C. Also a tripeptide, His-Gly-Lys, that hasbeen isolated from cat spinal cord and reported to impair firingof neurones in the dorsal horn (34) is present at positions309-311 of a-tubulin.

As these proteins are quite unrelated, it is difficult to explainthe similarities on the basis ofevolutionary relationship. We feelthat a more useful approach would be to evaluate the sequencesimilarities strictly on the basis of structure-function criteria.Thus, one would expect that two unrelated proteins or regionsof proteins that perform analogous functions should also havesimilar amino acid sequences regardless of the evolutionaryrelationship.Note Added in Proof. After we had communicated this article, the nu-cleotide sequence ofcDNA from chicken brain tubulin messengers waspublished by Valenzuela et al. (35). From these data, an amino acidsequence for chicken brain a-tubulin was deduced, corresponding toresidues 41-451 of our sequence and differing only in residues 175 (ar-ginine), 295 (tyrosine), and 358 (glutamate) from one of our variants.We wish to thank Mr. Jurgen Kretschmer, Mrs. Ch. Orlando, and

Miss Herta Scherer for their skillful technical assistance; Dr. G. Os-terburg for programming the method ofsecondary structure prediction,and Drs. G. Schulz and R. Woodbury for discussions. This work wassupported by the Deutsche Forschungsgemeinschaft.1. Steiner, M. (1978) Nature (London) 272, 834-835.2. Geahlen, R. L. & Haley, B. E. (1979) J. Biol. Chem. 254,

11982-11987.3. Eipper, B. A. (1972) Proc. Natl. Acad. Sci. USA 69, 2283-2287.4. Luduena, R. F., Shooter, E. M. & Wilson, L. (1977), J. Biol.

Chem. 252, 7006-7014.5. Bhattacharyya, B. & Wolff, J. (1974), Proc. Natl. Acad. Sci. USA

71, 2627-2631.

6. Yang, S. & Criddle, R. S. (1970) Biochemistry 9, 3063-3072.7. Little, M. (1979) FEBS Lett. 108, 283-286.8. Ponstingl, H., Krauhs, E., Little, M., Kempf, T. & Hofer-War-

binek, R. (1980) in Methods in Peptide and Protein SequenceAnalysis, ed. Birr, C. (Elsevier/North-Holland, Amsterdam),pp. 225-234.

9. Jany, K. D., Keil, W., Meyer, H. & Kiltz, H. H. (1976) Biochim.Biophys. Acta 453, 62-66.

10. Sonneborn, H. H., Zwilling, R. & Pfleiderer, G. (1969) Hoppe-Seyler's Z. Physiol. Chem. 350, 1097-1102.

11. Drapeau, G. R. (1980)J. Biol. Chem. 255, 839-840.12. Brauer, A. W., Margolies, M. N. & Haber, E. (1975) Biochem-

istry 14, 3029-3035.13. Tarr, G. E., Beecher, J. F., Bell, M. & McKean, D. J. (1978)

Anal. Biochem. 84, 622-627.14. Zimmerman, C. L., Appella, E. & Pisano, J. J. (1977) Anal.

Biochem. 77, 569-573.15. Edman, P. & Henschen, A. (1975), in Protein Sequence Deter-

mination, ed. Needleman, S. B. (Springer, New York), pp.232-279.

16. Ponstingl, H., Nieto, A. & Beato, M. (1978) Biochemistry 17,3908-3912.

17. Ponstingl, H., Little, M., Krauhs, E. & Kempf, T. (1979) Nature(London) 282, 423-424.

18. Kirschner, M. W. (1978) Int. Rev. Cytol. 54, 1-71.19. Zenner, H. P. & Pfeuffer, T. (1976) Eur. J. Biochem. 71,

177-184.20. Kelly, P. T. & Cotman, C. W. (1978)J. Cell Biol. 79, 173-183.21. Murofushi, H. (1980)1. Biochem. 87, 979-984.22. Kemphues, K. J., Raff, R. A., Kaufman, T. C. & Raff, E. C.

(1979) Proc. Natl. Acad. Sci. USA 76, 3991-3995.23. Chou, P. Y. & Fasman, G. D. (1978) Annu. Rev. Biochem. 47,

251-276.24. Ventilla, M., Cantor, C. R. & Shelanski, M. (1972) Biochemistry

11, 1554-1561.25. Lee, J. C., Corfman, D., Frigon, R. P. & Timasheff, S. N. (1978)

Arch. Biochem. Biophys. 185, 4-14.26. Luduena, R. F. & Woodward, D. 0. (1973) Proc. Natl. Acad.

Sci. USA 70, 3594-3598.27. Lu, R. C. & Elzinga, M. (1978) Biochim. Biophys. Acta 537,

320-328.28. Collins, J. H. & Elzinga, M. (1975) J. Biol. Chem. 250,

5915-5920.29. Elzinga, M. & Collins, J. H. (1977) Proc. Natl. Acad. Sci. USA

74, 4281-4284.30. Pearlstone, J. R., Johnson, P., Carpenter, M. R. & Smillie; L. B.

(1977) J. Biol. Chem. 252, 983-989.31. Yamashita, T., Soma, Y., Kobayashi, S. & Sekine, T. (1974) J.

Biochem. 75, 447453.'32. Kuriyama, R. & Sakai, H. (1974)J. Biochem. 76, 651-654.33. Nishida, E. & Kobayashi, T. (1977)J. Biochem. 81, 343-347.34. Lote, C. J., Gent, J. P., Wolstencroft, J. H. & Szelke, M. (1976)

Nature (London) 264, 188-189.35. Valenzuela, P., Auiroga, M., Zaldivar, J., Rutter, W. J., Kirsh-

ner, M. W. & Cleveland, D. W. (1981) Nature (London) 289,650-655.

Biochemistry: Ponstingl et al.

Dow

nloa

ded

by g

uest

on

June

22,

202

1