- Email: [email protected]
A homologous domain between the amyloid protein of Alzheimer's disease and the neurofilament subunits
Biochimie 71 (1989) 853-856 (~) Soci6t6 de Chimie biologique/Elsevier, Paris
853
A homologous domain between the amyloid protein of Alzheimer's disease and the neurofilament subunits Christian D E L A M A R C H E *
INSERM U298, Centre Hospitalier R~gional Universitaire, 49033 Angers Cedex, France (Received 1-6-1989, accepted 7-6-1989)
Summary m Polypeptide (A4), which is derived from a larger precursor membrane protein (/3APP), is an important component of brain amyloid in Alzheimer's disease. The physiological function and the processing of this precursor are largely unknown. In order to elucidate the actual role of t3APP, we searched for domain homology with other proteins. The present study reveals the presence of a highly conserved region between the amyloid precursor and the 3 neurofilament subunits. Identical amino acids are present at about 45% of the positions aligned between the 4 sequences. These observations are discussed in terms of a possible involvment of positive ions in the maturation of these proteins and are in favour the implication of aluminium in Alzheimer's disease. Alzheimer's disease / amyioid / neurofilaments / aluminium
Introduction
In the last 2 years, 4 laboratories have reported the nucleofide sequence of the/3amyloid protein precursor (/3APP) and the localization of its gene on chromosome 21. These results have marked the entry of the recombinant DNA technology in the field of "Alzheimerology". Four/3APP mRNAs encoding proteins of 695, 751, 770 and 843 amino acids have now been isolated with the cDNA technique [1-8]. All these predicted sequences contain the A4 polypeptide characteristic of the cerebral amyloid but they differ by the variable p~esence of a domain highly homologous to the reactive center of some protease inhibitors. The longest published cDNA has an intriguing 5' end extension of 73 residues long. The/3APP was first described as a cell-surface receptor [1, 9]. Recently, Schubert et al. [10] suggested that the /3APP is a heparan sulfate proteoglycan core protein and may function in
cell-cell interactions. Experimental results by Shivers et al. suggest a role in cell contact [11]. Glenner [12] proposed that the precursor protein is processed to release a peptide ligand as in the case of the experimental growth factor. To further clarify the role of the/3APP, the sequence homology to known proteins has been examined. The Microgenie alignment program (Beckman, Pain Alto, USA) was applied to the National Biochemical Research Foundation library (release No. 13) and to hand-entered sequences. Results
Among the 4525 sequences of the data bank only 27 proteins showed a level of homology higher than 40% for at least 40 consecutive amino acids. Surprisingly the strongest identity was not observed with the sequences of protease inhibitors as expected from previous reports but with the
*Present address: CNRS URA 256 Campus de Beaulieu, 35042 Rennes Cedex, France.
854
C. Delamarche
neurofilament triplet L protein from pig and ~ith tile micro glutamic acid-rich protein from ox. This protein arises in brain tissue by restricted proteolysis of the small neurofilamcnt subunit [13]. An analogous sequence containing successive tracts of glutamic acid residues is found in the 3 neurofilament proteins (NFs). NFs are the major cytoskeletal elements in neurons and have been detected with monoclona! antibodies in Alzheimer's disease neurofibrillary tangles [14]. The homologous region including the residues 222 to 264 of the BAPP shows a peculiar amino acid composition: there are only 8 amino acid types, and glutamic/aspartic acids account for 67% of the total. This sequence is located in the putative extracellular domain/3APP but could be associated with senile plaques as demonstrated independently by Palmert [15] and Perry [16]. Figure 1 shows the relationship between the /3APP sequence [2] from 222 to 264 and the corresponding region of the human neurofilament triplet proteins (NFs) [17-19]. The sequences are aligned by the computer program for optimal homology and boxed residues indicate the presence of amino acids identical to the /3APP. Table I summarizes the same analysis expressed as percent identity within t h e / 3 A P P (222 to 264) and various neurofilament proteins from man (H), mouse [20] (M) and pig [21] (P). These results show that the BAPP contains many amino acids in consensus with a part of the unusual carboxy-terminal tail of the NFs preceding its multiphosphorilation domain. The flAPP domain is much closer to each NF
sequence than the NFs themselves. A 12 amino acid correspondance is observed in all 4 sequences in comparison with only 14 conservative positions between the 3 neurofilaments subunits.
Discussion Senile plaques and neurofibrillary tangles, which are the 2 major lesions of Alzhcimer's disease contain insol,tble, proteinase-resistant and aggregating filaments. The presence of acidic sequences m the composition of brain deposits may contribute to their aggregating properties and explain the anomalous mobility of poly-
Table 1. Percent identity within the [3APP (222-264) and various neurofilament proteins from man (H), mouse (M) and pig (P). APP APP
NFM(H)NFL(H) NFH(H)NFM(M)NFL(M) NFL(P)
100
NFM(H) 43.5 lO0 NFL{H) 50
38.8 11~t
NFq(H) 47.8
-'1"~r_.5
/,1"7 1O0 ~,il
NFM(M) 32
59.6
48
35,4
100
NFL(M) .14.7
34.7
59.6
35.4
33.3
100
NFL(P) 44.5
39.6
52
38.3
30.6
41.7
Protein APP
NFN
(H)
Matches
IA,EEEv.lv
~D ~V V i
E E E E P E[~E
""IEIE°'E
E T-ZE
E IlO00 0' 0 011VEE
E V A~IK K S P V KIAIT A p EIV KIE E - E GIEE'ICIIEE E V[~EIE[~I~-iAIK E E EIG KIEE E...~GI E E
-EE-Ik E E E EA~-IE
E E[A "IA~K e ",~JS E,..L E A" KEI1
N' '
1~)
Length H/L (~)
E- E- -
20
46
43.5
E- -
2~
44
~7.8
ETK
23
46
50
Fig. 1. The relationship between the/3APP sequence from 222 to 264 and the corresponding region of the human neurofilament triplet proteins (NFs).
Hypothesis on A l z h e i m e r ' s disease
peptides in SDS gels. The reduced SDS binding by the /3APP was previously postulated by Beyreuther et al. [9]. The existence of a conservative domain between the /3APP and the 3 neurofilament subunits suggest the existence of a comparable and specific physiological function: this could I:,rovide a highly charged trap able to bind positive ions or proteins. Some observations reveal a role for positive ions in biological mechanisms of protein maturation: a calcium deoendent processing of the neurofilaments has been observed in relation to the axonal transport [13]. The cadherins are a class of calcium dependant cell-cell adhesion glycoproteins. Ca 2+ protects these molecules from protease digestion and appears as an essential ion in multiple processes of cell adhesion [22]. The acidic domain described in this paper could be a putative Ca-'* binding motif. Thus, the present observations agree with some experimental results [11] on the putative functions o f / 3 A P P in the interprotein cross-bridges of the extracellular matrix, perhaps in cell-cell interactions and information transport between neurons. A stretch of consecutive acidic amino acids is also present in tubulin [23], the main protein of microtubules; and in a ubiquinol-cytochrome c reductase subunit [24], a proposed "hinge" mwtochondriai protein for interaction between cytochromes c and cl. In view of the unusual n.~ iaItK Lni kl rIKA[n 1
'~,1~
n'~t.~wf I
tlghi o! ¢v,l.1o. j ' ~ . . ¢
cd p'1~1~n~~ nlk.,l[o n'1~¢p1o1 c' 1 ~ " l ~
~
f1 nIL..~~ r1 t L| ai oi ri,.~
i
c. 1 ti~. 'li..~ n r~ l l io'~..c
tl
- a r o}
B I
required in looking for them in brain deposits, especially since excess numbers of mitochondria were previously described in senile plaques by light microscopy and as the microtubule binding protein tau is a major component of the paired helical filaments [25] The accumulation of acidic debris in Alzheimer brains could be also correlated with the role of aluminium in the disease. The presence of high concentrations of aluminium has been observed in neuronal tangles and in senile plaques [26]. Acidic sequences may serve as specific targets for aluminium binding. Aluminium is an aggregating agent for proteins and also shows antiprotease activity in vitro (Delamarche C. & Wion D., unpublished observations). Thus, aluminium could be a competitor for Ca 2+ or other ions and could interfere in the maturation or functional activity of t h e / 3 A P P and neurofilament subunits. This argues for the possible direct or indirect implication of mineral deposits to the genesis or maturation of neuritic plaques.
g55
References 1 Goldgaber D.. Lerman M.I.. McBride O W.. Saffiotti U. & Gajdusek D C. (1987) Science 235. 877- 880 2 Kang J., Lemaire H.G.. Unterbeck A.. Saibaum J.M. & Masters C.L. (1987) Nature 325, 733-736 3 Robakis N.K.. Ramakrishna N., Wolfe G. & Wisniewski H.M. (1987) Proc. Natl. Acad. Sci. USA 84, 4190-4194 4 Tanzi R.E., Gu~ella J.F., Watkins P.C., Bruns G.A.P. & St. George-Hyslop P.H. (!987) Science 235,880- 884 5 Ponte P., Gonzalez-De Whitt P.. Schilling J.. ,~,liller J. & Hsu D. (1988) Nature 331,525-527 6 Tanzi R.E., McClatchey A.I., Lamperti E.D.. Villa-Komaroff L.. Gusella J.F. & Neve R,L. (1988) Nature 331,528-530 7 Kitaguchi N., Takahashi Y., Tokushima Y., Shiojiri S. & Ito H. (1988) Nature 331,53(I-532 8 Mita S., Sadlock J., Herbert J. & Schon E.A. (1988) Nucleic Acids Res. 16, 9351 9 Dyrks T., Weidemann A., Multhaup G., Salbaum J.M. & Lemaire H.G. (1988)EMBOJ. 7. 949-957 10 Schubert D., Schroeder R., LaCorbiere M., Saitoh T. & Cole G. (1988) Science 241,223-226 11 Shivers B.D., Hilbich C., Multhaup G., Salbaum M., Beyreuther K. & Seeburg P.H. (1988) EMBO J. 7. 1365-1370 12 Allsc,p D., Wong C.W., Ikeda S.. Landon M., Kidd M. & Glenner G.G. (1988) Proc. Natl. Z.I'JU--~,I"J-~ ,'waa. Sci. [_,r . ) 2-1 13 Ixobe T. & Okuyama T. (1985) FEBS Lett. 182, 389-392 14 Lee V.M.-Y.. Otvos Jr L., Schmidt M.L. & Trojanowski J.Q. (1988)Proc. Natl. Acad. Sci. USA 85, 7384- 7388 15 Palmert M.R., Podlisny M.B., Witker D.S.. Oltersdorf T. & Younkin L.H. (1988) Biochem Biophys. Res. Commun. 156,432-437 16 Perry G., Lipphardt S., MuMhill P.. Kancherla M. & Mijares M. (1988) Lancet 746 17 Myers M.W., Lazzarini R.A., Lee V.M.-Y, Schlaepfer W.W. & Nelson D.L. (1987) EMBO J. 6, 1617-1626 18 Lees J.F., Shneidman P.S.. Skuntz S.F.. Carden M.J. & Lazzarini R.A. (1988) EMBO J. 7. 1947-1955 19 Julien J.P., Grosveld F., Yazdanbaksh K., Flayell D., Meijcr D. & Mushynski W. (1987) Biochim. Biophvs. Acta 9()9~ 10-211
20 Julien J.P.,Meyer D.. Flavell D., Hurst J. & Grosveld F. (1986~ Mol. Brain Res, 1,243-250 21 Geisler N.. Piessmann U. & Weber K. (1985) FEBS Lett. 182,475- 478 22 Takeichi M. (1988) Development 102,639-655 23 Ponstingl H., Little M., Krauhs E. & Kempf T.
&%
C. Delat,~arche
(1979) Nature 282,423-424 24 Van Loon A.P.G.M., De Groot R.J., De Haan M., Dekker A. & Grivell L.A. (1984) EMBO J. 3, 1039-1043
25 Lee G., Cowan N. & Kirschner M. (1988) Science 239,285-288 26 Candy J.M., Oakley A.E., Klinowski J., Carpenter T.A. & Perry R.H. (1986) Lancet 354-356
853
A homologous domain between the amyloid protein of Alzheimer's disease and the neurofilament subunits Christian D E L A M A R C H E *
INSERM U298, Centre Hospitalier R~gional Universitaire, 49033 Angers Cedex, France (Received 1-6-1989, accepted 7-6-1989)
Summary m Polypeptide (A4), which is derived from a larger precursor membrane protein (/3APP), is an important component of brain amyloid in Alzheimer's disease. The physiological function and the processing of this precursor are largely unknown. In order to elucidate the actual role of t3APP, we searched for domain homology with other proteins. The present study reveals the presence of a highly conserved region between the amyloid precursor and the 3 neurofilament subunits. Identical amino acids are present at about 45% of the positions aligned between the 4 sequences. These observations are discussed in terms of a possible involvment of positive ions in the maturation of these proteins and are in favour the implication of aluminium in Alzheimer's disease. Alzheimer's disease / amyioid / neurofilaments / aluminium
Introduction
In the last 2 years, 4 laboratories have reported the nucleofide sequence of the/3amyloid protein precursor (/3APP) and the localization of its gene on chromosome 21. These results have marked the entry of the recombinant DNA technology in the field of "Alzheimerology". Four/3APP mRNAs encoding proteins of 695, 751, 770 and 843 amino acids have now been isolated with the cDNA technique [1-8]. All these predicted sequences contain the A4 polypeptide characteristic of the cerebral amyloid but they differ by the variable p~esence of a domain highly homologous to the reactive center of some protease inhibitors. The longest published cDNA has an intriguing 5' end extension of 73 residues long. The/3APP was first described as a cell-surface receptor [1, 9]. Recently, Schubert et al. [10] suggested that the /3APP is a heparan sulfate proteoglycan core protein and may function in
cell-cell interactions. Experimental results by Shivers et al. suggest a role in cell contact [11]. Glenner [12] proposed that the precursor protein is processed to release a peptide ligand as in the case of the experimental growth factor. To further clarify the role of the/3APP, the sequence homology to known proteins has been examined. The Microgenie alignment program (Beckman, Pain Alto, USA) was applied to the National Biochemical Research Foundation library (release No. 13) and to hand-entered sequences. Results
Among the 4525 sequences of the data bank only 27 proteins showed a level of homology higher than 40% for at least 40 consecutive amino acids. Surprisingly the strongest identity was not observed with the sequences of protease inhibitors as expected from previous reports but with the
*Present address: CNRS URA 256 Campus de Beaulieu, 35042 Rennes Cedex, France.
854
C. Delamarche
neurofilament triplet L protein from pig and ~ith tile micro glutamic acid-rich protein from ox. This protein arises in brain tissue by restricted proteolysis of the small neurofilamcnt subunit [13]. An analogous sequence containing successive tracts of glutamic acid residues is found in the 3 neurofilament proteins (NFs). NFs are the major cytoskeletal elements in neurons and have been detected with monoclona! antibodies in Alzheimer's disease neurofibrillary tangles [14]. The homologous region including the residues 222 to 264 of the BAPP shows a peculiar amino acid composition: there are only 8 amino acid types, and glutamic/aspartic acids account for 67% of the total. This sequence is located in the putative extracellular domain/3APP but could be associated with senile plaques as demonstrated independently by Palmert [15] and Perry [16]. Figure 1 shows the relationship between the /3APP sequence [2] from 222 to 264 and the corresponding region of the human neurofilament triplet proteins (NFs) [17-19]. The sequences are aligned by the computer program for optimal homology and boxed residues indicate the presence of amino acids identical to the /3APP. Table I summarizes the same analysis expressed as percent identity within t h e / 3 A P P (222 to 264) and various neurofilament proteins from man (H), mouse [20] (M) and pig [21] (P). These results show that the BAPP contains many amino acids in consensus with a part of the unusual carboxy-terminal tail of the NFs preceding its multiphosphorilation domain. The flAPP domain is much closer to each NF
sequence than the NFs themselves. A 12 amino acid correspondance is observed in all 4 sequences in comparison with only 14 conservative positions between the 3 neurofilaments subunits.
Discussion Senile plaques and neurofibrillary tangles, which are the 2 major lesions of Alzhcimer's disease contain insol,tble, proteinase-resistant and aggregating filaments. The presence of acidic sequences m the composition of brain deposits may contribute to their aggregating properties and explain the anomalous mobility of poly-
Table 1. Percent identity within the [3APP (222-264) and various neurofilament proteins from man (H), mouse (M) and pig (P). APP APP
NFM(H)NFL(H) NFH(H)NFM(M)NFL(M) NFL(P)
100
NFM(H) 43.5 lO0 NFL{H) 50
38.8 11~t
NFq(H) 47.8
-'1"~r_.5
/,1"7 1O0 ~,il
NFM(M) 32
59.6
48
35,4
100
NFL(M) .14.7
34.7
59.6
35.4
33.3
100
NFL(P) 44.5
39.6
52
38.3
30.6
41.7
Protein APP
NFN
(H)
Matches
IA,EEEv.lv
~D ~V V i
E E E E P E[~E
""IEIE°'E
E T-ZE
E IlO00 0' 0 011VEE
E V A~IK K S P V KIAIT A p EIV KIE E - E GIEE'ICIIEE E V[~EIE[~I~-iAIK E E EIG KIEE E...~GI E E
-EE-Ik E E E EA~-IE
E E[A "IA~K e ",~JS E,..L E A" KEI1
N' '
1~)
Length H/L (~)
E- E- -
20
46
43.5
E- -
2~
44
~7.8
ETK
23
46
50
Fig. 1. The relationship between the/3APP sequence from 222 to 264 and the corresponding region of the human neurofilament triplet proteins (NFs).
Hypothesis on A l z h e i m e r ' s disease
peptides in SDS gels. The reduced SDS binding by the /3APP was previously postulated by Beyreuther et al. [9]. The existence of a conservative domain between the /3APP and the 3 neurofilament subunits suggest the existence of a comparable and specific physiological function: this could I:,rovide a highly charged trap able to bind positive ions or proteins. Some observations reveal a role for positive ions in biological mechanisms of protein maturation: a calcium deoendent processing of the neurofilaments has been observed in relation to the axonal transport [13]. The cadherins are a class of calcium dependant cell-cell adhesion glycoproteins. Ca 2+ protects these molecules from protease digestion and appears as an essential ion in multiple processes of cell adhesion [22]. The acidic domain described in this paper could be a putative Ca-'* binding motif. Thus, the present observations agree with some experimental results [11] on the putative functions o f / 3 A P P in the interprotein cross-bridges of the extracellular matrix, perhaps in cell-cell interactions and information transport between neurons. A stretch of consecutive acidic amino acids is also present in tubulin [23], the main protein of microtubules; and in a ubiquinol-cytochrome c reductase subunit [24], a proposed "hinge" mwtochondriai protein for interaction between cytochromes c and cl. In view of the unusual n.~ iaItK Lni kl rIKA[n 1
'~,1~
n'~t.~wf I
tlghi o! ¢v,l.1o. j ' ~ . . ¢
cd p'1~1~n~~ nlk.,l[o n'1~¢p1o1 c' 1 ~ " l ~
~
f1 nIL..~~ r1 t L| ai oi ri,.~
i
c. 1 ti~. 'li..~ n r~ l l io'~..c
tl
- a r o}
B I
required in looking for them in brain deposits, especially since excess numbers of mitochondria were previously described in senile plaques by light microscopy and as the microtubule binding protein tau is a major component of the paired helical filaments [25] The accumulation of acidic debris in Alzheimer brains could be also correlated with the role of aluminium in the disease. The presence of high concentrations of aluminium has been observed in neuronal tangles and in senile plaques [26]. Acidic sequences may serve as specific targets for aluminium binding. Aluminium is an aggregating agent for proteins and also shows antiprotease activity in vitro (Delamarche C. & Wion D., unpublished observations). Thus, aluminium could be a competitor for Ca 2+ or other ions and could interfere in the maturation or functional activity of t h e / 3 A P P and neurofilament subunits. This argues for the possible direct or indirect implication of mineral deposits to the genesis or maturation of neuritic plaques.
g55
References 1 Goldgaber D.. Lerman M.I.. McBride O W.. Saffiotti U. & Gajdusek D C. (1987) Science 235. 877- 880 2 Kang J., Lemaire H.G.. Unterbeck A.. Saibaum J.M. & Masters C.L. (1987) Nature 325, 733-736 3 Robakis N.K.. Ramakrishna N., Wolfe G. & Wisniewski H.M. (1987) Proc. Natl. Acad. Sci. USA 84, 4190-4194 4 Tanzi R.E., Gu~ella J.F., Watkins P.C., Bruns G.A.P. & St. George-Hyslop P.H. (!987) Science 235,880- 884 5 Ponte P., Gonzalez-De Whitt P.. Schilling J.. ,~,liller J. & Hsu D. (1988) Nature 331,525-527 6 Tanzi R.E., McClatchey A.I., Lamperti E.D.. Villa-Komaroff L.. Gusella J.F. & Neve R,L. (1988) Nature 331,528-530 7 Kitaguchi N., Takahashi Y., Tokushima Y., Shiojiri S. & Ito H. (1988) Nature 331,53(I-532 8 Mita S., Sadlock J., Herbert J. & Schon E.A. (1988) Nucleic Acids Res. 16, 9351 9 Dyrks T., Weidemann A., Multhaup G., Salbaum J.M. & Lemaire H.G. (1988)EMBOJ. 7. 949-957 10 Schubert D., Schroeder R., LaCorbiere M., Saitoh T. & Cole G. (1988) Science 241,223-226 11 Shivers B.D., Hilbich C., Multhaup G., Salbaum M., Beyreuther K. & Seeburg P.H. (1988) EMBO J. 7. 1365-1370 12 Allsc,p D., Wong C.W., Ikeda S.. Landon M., Kidd M. & Glenner G.G. (1988) Proc. Natl. Z.I'JU--~,I"J-~ ,'waa. Sci. [_,r . ) 2-1 13 Ixobe T. & Okuyama T. (1985) FEBS Lett. 182, 389-392 14 Lee V.M.-Y.. Otvos Jr L., Schmidt M.L. & Trojanowski J.Q. (1988)Proc. Natl. Acad. Sci. USA 85, 7384- 7388 15 Palmert M.R., Podlisny M.B., Witker D.S.. Oltersdorf T. & Younkin L.H. (1988) Biochem Biophys. Res. Commun. 156,432-437 16 Perry G., Lipphardt S., MuMhill P.. Kancherla M. & Mijares M. (1988) Lancet 746 17 Myers M.W., Lazzarini R.A., Lee V.M.-Y, Schlaepfer W.W. & Nelson D.L. (1987) EMBO J. 6, 1617-1626 18 Lees J.F., Shneidman P.S.. Skuntz S.F.. Carden M.J. & Lazzarini R.A. (1988) EMBO J. 7. 1947-1955 19 Julien J.P., Grosveld F., Yazdanbaksh K., Flayell D., Meijcr D. & Mushynski W. (1987) Biochim. Biophvs. Acta 9()9~ 10-211
20 Julien J.P.,Meyer D.. Flavell D., Hurst J. & Grosveld F. (1986~ Mol. Brain Res, 1,243-250 21 Geisler N.. Piessmann U. & Weber K. (1985) FEBS Lett. 182,475- 478 22 Takeichi M. (1988) Development 102,639-655 23 Ponstingl H., Little M., Krauhs E. & Kempf T.
&%
C. Delat,~arche
(1979) Nature 282,423-424 24 Van Loon A.P.G.M., De Groot R.J., De Haan M., Dekker A. & Grivell L.A. (1984) EMBO J. 3, 1039-1043
25 Lee G., Cowan N. & Kirschner M. (1988) Science 239,285-288 26 Candy J.M., Oakley A.E., Klinowski J., Carpenter T.A. & Perry R.H. (1986) Lancet 354-356