The Primary Structure of Amyloid Fibril Protein ... - Wiley Online Library

3 downloads 0 Views 444KB Size Report
The Primary Structure of Amyloid Fibril Protein AA in Endotoxin-Induced Amyloidosis of the Mink. Kristian WAALEN, Knut SLETTEN, Gunnar HUSBY, and Knut ...
Eur. J. Biochem. 104;407-412

(1980)

The Primary Structure of Amyloid Fibril Protein AA in Endotoxin-Induced Amyloidosis of the Mink Kristian WAALEN, Knut SLETTEN, Gunnar HUSBY, and Knut NORDSTOGA Department of Biochemistry, University of Oslo, Department of Rheumatology, University of Tromsd, and National Veterinary Institute, Oslo (Received July 5, 1979)

Two AA proteins were isolated from the same amyloid fibril preparation from the liver of a mink, in which amyloidosis had been induced by injections with endotoxin. The two proteins were of different size, one containing 53 amino acid residues and the other 64 residues. The amino acid sequence was otherwise found to be identical. Both proteins revealed pyrrolidone carboxylic acid as the Nterminal amino acid. Sequence homologies with protein AA from other species were very striking. However, no antigenic cross-reaction was seen between mink protein AA and antisera to protein AA from human, mouse or rabbit sources. Amyloid protein AA is a major component of the amyloid fibrils in secondary amyloidosis in man [ 1- 51, while monoclonal immunoglobulin light chains or fragments thereof play a similar role in primary amyloidosis. When amyloidosis is induced experimentally in different animals by repeated injections with various antigens, the resulting amyloid fibrils have consistantly been shown to be of the protein AA type [6 - 81. This also appeared to be the case with amyloidosis of the mink induced with endotoxin. Thus, the primary structure of a peptide containing 13 amino acid residues obtained from mink protein AA showed marked homology with human protein AA [7]. However, the N-terminal amino acid residue of mink protein AA was shown to be blocked, and direct sequencing could not be performed. When protein AA from mink was tested with antiserum to human protein AA, no cross-reaction was seen [7]. In addition to being an essential component of secondary amyloid in man, protein AA is also found as a major amyloid fibril subunit in spontaneous [9 111 and experimental [6-81 amyloidosis in an increasing number of different animals. Abbreviation. BNPS-skatole, bromine adduct of 2-(2-nitrophenylsulfenyl)-3-methylindole. Enzymes. Trypsin (EC 3.4.21.4); pepsin (EC 3.4.23.1); thermolysin (EC 3.4.24.4). Nomenclature and Symbols. The nomenclature for amyloid proteins is described in [26]. The peptides obtained from digestion with trypsin, pepsin, thermolysin, cyanogen bromide and BNPS-skatole are numbered T-I. . . , P-1.. ., Th-1.. ., CB-1, CB-2, CB-3, and BNPS-skatole-2, BNPS-skatole-3 respectively, in the order of the presented sequence. Symbols : residues identified by Edman degradation, ?, are so indicated.

In the present communication we report the complete amino acid sequence of mink protein AA. Two AA proteins were isolated from the same amyloid fibril preparation obtained from the liver of a mink (Mustela vision) in which amyloidosis had been induced by injections with endotoxin. The two proteins were different only with respect to size while they were otherwise identical in their primary structure. Marked homologies with protein AA from man as well as from various animals were observed. MATERIALS AND METHODS Amyloidosis was induced in minks using subcutaneous injections with Eschericha coli 026 : B6 endotoxin (Difco Laboratories) [7,12,13]. One animal in particular developed a large liver (76 g) containing rich amounts of amyloid substance, and this liver was chosen for the preparation of amyloid fibrils. The weight of the remaining five livers varied from 19 g to 38 g. Amyloid fibrils were extracted with water after repeated washings of homogenized tissue with physiological saline [14,15], degraded with 6 M guanidine/ 0.55 M Tris-HC1 containing 0.1 M dithiothreitol, and fractionated on a Sephadex G-100 column (88 x 2.5 cm) equilibrated with 5 M guanidine/l M acetic acid [15,16]. For further purification, rechromatography of selected fractions was performed on a column (2.5 x 88.5 cm) with Sephadex G-75 Superfine, equilibrated with 10% formic acid. Double diffusion in agarose against antiserum to mink amyloid

408

protein AA [7] was used for antigenic characterization of the proteins. The homogeneity and molecular weight of AA proteins were checked h! sodium dodecylsulfate gel electrophoresis [17]. The amino acid composition was carried out on acid hydrolysates as described [5]. The protein was cleaved by cyanogen bromide and the resulting fragments (CB) were purified by precipitation in 10% formic acid, and gel-filtration on Sephadex G-50 Fine. Cleavage of the protein with 2-(2-nitrophenylsulfenyl)-3-methyl3'-bromoindolenine (BNPS-skatole) was performed [IS] and the fragments were purified by gel filtration on Sephadex G-25 Fine and by thin-layer chromatography. Dilute acid cleavage was carried out [5] and the peptides purified by thin-layer chromatography. The protein as well as the CB-1 fragment was also enzymatically degraded by trypsin, thermolysin and pepsin and the peptides obtained from the different digests were purified by ion-exchange chromatography, thin-layer chromatography and thin-layer electrophoresis [5]. Peptide mapping of tryptic digested protein AA1 and AA2 was performed on Polygram SIL G/UV254; first dimension : chloroform/methanol/ 25 ammonia (2 : 2 :1) ; second dimension : electrophoresis in pyridine/acetic acid, pH 3.5 at 1000 V for 45 min. The N-terminal cyanogenbromide fragment (CB-1) was deblocked by methanolysis using 1 M HC1 in 2-chloroethanol to solubilize the sample and incubated for 65 h at 23 'C [19]. The deblocking was tested by dansylation [20,21]. Edman degradation was performed manually and with a JEOL-JAS-47K Sequence Analyzer [22]. The phenylthiohydantion amino acids (PTH-AA) were identified by gas chromatography, thin-layer chromatography and amino acid analysis after hydrolysis of the derivatives in 6 M HC1 for 18 h [22].

The Primary Structure of Amyloid Fibril Protein AA from Mink

100

150

200 Elution volume (ml)

250

300

Fig. 1. Elution diagram for degraded umyloid fibrils of mink on Sephadex G-100 column (2.5 x 88 em)

0.4 0.5

CI

5 1

2 % 0.3

AAI

I

AA2

02

c m

e

:: 01

4 Q

n 180

240 Elution volume (mi)

300

Fig. 2. Elution diagramjor umyloidprotein AAfraction on a Sephadex G-75 Superfine column (2.5x88.5 em). Effluent: loo/, formic acid

+

RESULTS AND DISCUSSION The yield of lyophilized amyloid fibrils from 76 g (wet weight) of liver tissue was 2.38 g. When degraded amyloid fibrils were subjected to gel filtration on Sephadex G-100, two incompletely separated protein peaks were obtained at the position where protein AA was expected to be eluted [7]. The amount of protein AA was 53 of the total proteins eluted as estimated by the areas of the protein peaks. Fractions were collected as indicated in Fig. 1, dialyzed against distilled water and lyophilized. Following rechromatography on Sephadex G-75 Superfine the final amounts of AAI and AA2 were 71 mg and 118 mg (dry weight) respectively, when fractions were pooled as shown in Fig. 2. Immunological studies showed that both AAI and AA2 reacted with the anti-mink-protein-At serum showing precipitation lines of antigenic identity with

Fig. 3. Sodium dodecJ;l.tulfatelpolyacrJ;lumid~, gel rlc~c~irophoresis of protein A A 1 and AAz. 50 pg protein was applied. The gels stained with Coomassie blue are shown at the top, and scans of the gels at 600 nm are shown below. The band close to the anode is the bromophenol blue marker. The electrophoresis was carried out in 12.5 polyacrylamide gel according to Swank and Munkres [I71

each other and with the amyloid preparation used for immunization. However, no cross-reaction was observed when mink proteins AA1 and AA2 were tested against antisera to protein AA from human, mouse or rabbit sources. Both AAI and AA2 gave one single band on dodecylsulfate/polyacrylamide gel electrophoresis when tested separately (Fig. 3), and migrated accord-

409

K. Waalen, K. Sletten, G. Husby, and K. Nordstoga Table 1. Amino acid composition of mink amyloid protein A A , and AA2 and of fragments obtained from cyanogen bromide cleavage All values are averages of two or more acid hydrolysates Amino acid

AA, whole protein

AA2

CB-1

CB-2

CB-3

whole protein

CB-I

CB-2

CB-3

1.19 (1)

1.04 (1)

6.23 (6)

7.14 (7)

1.18 (1)

0.98 (I)

5.16 (5)

0.81 (1) 3.27 (3)

0.98 (1)

2.48 3.98 1.0 3.94 7.89 0.80

1.13 (1) 3.33 ( 3 )

0.83 (1)

0.99 2.13 0.95 4.20 6.61

(1) (2) (1) (4) (6-7)

0.54 (1)

3.16 (3) 5.72 (5) 1.09 (1) 6.09 (6) 9.33 (9) 1.26 (1) 2.00 (2) 0.18

3.08 1.12 0.97 2.78 5.06

6.31 3.20 1.04 2.24 4.70

2.54 0.97 1.09 2.07 3.19 1.0

(1) (1) (2) (3) (1)

residues/molecule

8.05 (8) 0.12 4.27 (4) 6.91 (7) 1.0 (1) 6.01 (6) 11.0 (11) 1.84 (2) 2.00 (2) 0.77 (1) 0.11 5.90 (6) 3.12 (3) 0.98 (1) 2.89 (3) 6.14 (6) (3)

ASP Thr Ser Glu Pro GlY Ala Val Met Ile Leu TYr Phe His LYs Arg TrP Yield (%) a

2.05 (2) 2.22 (2) 0.71 (1) 0.40"(1)

1.01 (1) 2.01 (2)

1.05 (1) 0.85"(1)

1.81 (2)

0.91 (1) 2.07 (2) 74

(2) (4) (1) (4) (8) (1)

57

(3) (1) (1) (3) (5) (1)

62

(6) (3) (1) (2) (4) (3)

1.87 (2) 2.05 (2) 1.11 (1) 0.47"(1)

0.54 (1) 1.78 (2)

0.99 (1) 0.76"(1)

1.62 (2)

1.02 (1) 1.8 (2) 25

61

(3)

68

Determined as homoserine and hemoserine lactone

ing to a molecular weight of about 6000. No difference between the molecular weight of AAt and AA2 could be distinguished by this method [37]. N-terminal analysis by Edman degradation and by dansylation confirmed previous observations [7] that the N-terminal amino acid was blocked, this being the case for both AA1 and AA2. Cyanogen bromide cleavage of both proteins in 70 % formic acid followed by dilution and gel filtration on a column with Sephadex G-50 Fine, resulted in 3 peptides from each protein. The amino acid composition of protein AA1 and AA2 as well as of the cyanogen bromide fragments obtained (Table l), revealed that protein AAI consisted of 64 amino acid residues and protein AA2 of 53 residues. Two peptides from each protein, called CB-1 and CB-2, had an identical amino acid composition and contained homoserine or homoserine lactone. As peptide CB-1 from both proteins was found to have a blocked N-terminal, this peptide had to be the N-terminal fragment, while CB-2 accordingly was the middle one. BNPS-skatole cleavage of protein AA2 followed by gel filtration and thin-layer chromatography resulted in free alanine and three peptides. One of these, a dipeptide, contained glutamic acid and tryptophan, gave a negative ninhydrin staining and thus indicating pyrrolidone carboxylic acid as the N-terminal. Methanolysis of peptide CB-1 for 65 h at 2 3 T , followed by dansylation, gave glutamic acid as N-terminal. Edman degradation of the peptide revealed the amino

acid sequence of residues 1 through 11 (Fig. 4). Edman degradation of BNPS-skatole peptides 2 and 3 elucidated the amino acid sequence up to position 30. These sequence data was verified by peptides P-1, P-2 and P-3 obtained from a peptic digest of peptide T-1, and further by peptides Th-1, Th-2, Th-3, Th-4 and Th-5 obtained from a thermolysin digest of CB-1. Edman degradation of CB-2 and CB-3 elucidated the sequence from position 17 to 57 in AAI (Fig.4). The total amino acid sequence of AA1 and AA2 was elucidated by peptide mapping of tryptic peptides. Peptide maps and amino acid analysis showed clearly that peptides T-1 through T-6 were identical from both proteins (Fig. 5). Peptide T-7 from AA1 (Fig. 5A) contained one residue more of alanine and Iysine than a similar peptide from AA2 (Fig. 5 B). Both peptides were Ehrlich positive, which would conclude that protein AA2 consisted of 53 amino acid residues. Peptides T-8 and T-9 could be distinguished as separate spots on peptide maps of AA1. Tryptic peptide T-8 was cleaved by dilute acid and the two resulting peptides were isolated and characterized (Fig. 4). A tryptic dipeptide, T-9, containing glutamic acid and arginine, was obtained in a lower yield. The amino acid sequence of this peptide was based upon the specificity of trypsin. According to the amino acid composition of CB-3 from AAt, the fourth residue of glutamic acid or glutamine was placed in position 64. According to the yield of tryptic peptides from AA1, it seemed that a certain amount

410

The Primary Structure of Amyloid Fibril Protein AA from Mink

. . . . . . . . . . . . . . . . . . . . .

. >

d--S 7

1 -3

c-s-2-3

7

Fig. 4.The amino acidsequence ofmink amyloidprotein A A , andAAz.The peptides are indicated by double-headed arrows. Amino acid residues ( - - - 7 ) are so indicated. Abbreviations used are: CB, cyanogen bromide fragments; T, tryptic pepidentified by Edman degradation (-) fragments; P, peptic peptides; Th, peptides obtained from tides; BNPS-skatole 2-(2-nitrophenylsulfenyl)-3-methyl-3’-bromoindolenine thermolysin digest; S, peptides obtained from dilute acid cleavage

m o o +

B

2

a,$ : Q

Solvent front

0

0 7

Q5 0

2

0 0

0 4

0 4

+

@

+

.

6 0 3

0 4

0 4

Gly-Trp Ref.

3

-2)

@

GIy - Tr p

+ Electrophoresis

Ref.

-

+

-2) Electrophoresis

Fig.5. Peptide mapping oftryptic-digestedprotein A A , ( A ) and AA2 ( B ) . The chromatograms were stained with ninhydrin and with a chlorbenzidin stain. Peptide T-1 precipitated out during the digestion and was not applied. The peptides were numbered in the order of the presented sequence. T-3 and T-4 appeared as double spots

of AA1 also was slightly shorter. Two of the tryptic peptides, T-3 and T-4, appeared on peptide maps of both proteins as double spots, with different electrophoretic mobilities. This differences are most probably due to deamidation of the asparagine residues in positions 27 and 30. It was interesting to note that this protein AA1 from mink consisted of the same number of amino acid residues as a protein AA obtained from the liver of a patient with ankylosing spondylitis previously reported by us [23].

The amino acid sequence showed a rather regular distribution of hydrophobic amino acids throughout the polypeptide, except for the N-terminal part. The amino acid sequence from position 16 to 28 in proteins AA1 and AA2 was identical to that of a fragment of another preparation of mink protein AA [7], except for a tyrosine/arginine substitution in position 24. The mink protein AA is, to our knowledge, the first case of such amyloid proteins to have a blocked N-terminus. However, Eriksen et al. [6] reported that

41 1

K. Waalen, K. Sletten, G. Husby, and K. Nordstoga 1 5 10 15 Human (5) NH 2-Arg-Ser-Phe-Phe-Ser-Phe-Leu-Gly-Glu-Ala-Phe-Asp-Gly-Ala-Arg-Asp-Met -Trp-TyrMonkey (10) NH2-Asp-Asn-Pro-Phe-Thr-Arg-Gly-Gly-Arg-Phe-Val-Asp-Ala-Gly- D -TrpDuck (9) -Phe-Val-Gln-TrpMink PCA-Trp-Tyr-

Human (5) Monkey (10) Duck ( 9 ) Mink

Human ( 5 ) Monkey (10) Duck ( 9 ) Mink

Human ( 5 ) Monkey (10) Duck (9) Mink

20 25 30 35 -Trp-Arg-Ala-Tyr-Ser-Asn-Met-Arg-Glu-Ala-Asn-Tyr-Ile-Gly-Ser-Asp-Lys-Tyr-Phe-His-Ala-Arg-Asp-LyS-Lys-Asn-Tyr-lle-Ala -Leu-Arg-AspHis Val -Tyr-Asp-Lys-Asn-

40 45 50 55 60 Gly-Asn-Tyr-Asp-Ala-Ala-Lys-Arg-Gly-Pro-Gly-Gly-Val-Trp-Ala-Ala-Glu-Ala-Ile-Ser-Asp-Ala-Val-Gln-

-Arg-

-Gln-

-Ala-Ala-

-Arg-Val-Lys-Val-

-A1a-AsnSer

65 70 75 Arg-Glu-Asn-Ile-Gln-Arg-Phe-Phe-Gly-His-Gly-Ala-Glu-Asn-Ser-COOH -Lys-Leu-Leu-Thr-COOH -Trp-Gly-Gly-Val-Ser-Arg(Gly,Ala,Glu,Asp,Arg/Asp,Thr,Glu,Gly,Ala,Argl -Arg-Glx-COOH

Fig. 6. Comparison of the primary structure ofprotein A A from different species. The continuous amino acid sequence is that of human protein AA [5]. Only the variant residues of proteins from other species are placed in the figure

approximately 50% of the N-terminal residue of mouse protein AA was blocked, while the unblocked was glycine. The biological significance of proteins having a blocked N-terminal residue is not known. When comparing mink protein AA with the corresponding protein from man [ 5 ] , monkey [lo] and duck [9], marked differences are seen in the N-terminal and C-terminal parts of the polypeptide while the other parts have an almost identical sequence (Fig. 6). In spite of these marked homologies, no antigenic cross-reactivity could be observed between the mink AA and the corresponding proteins from other species. This indicated that the antigenic determinant(s) may be located to areas of the protein where substitutions occur between different species. It was interesting that when comparing protein AA from different species, one hydrophobic amino acid was in many cases substituted by another hydrophobic group. This indicates that amphipathic helix structures are a consistent feature of AA proteins [24]. This lipid binding property of protein AA is interesting in view of the recent report by Benditt et al. [25] showing evidence that the AA-related serum protein SAA is found in the lipoprotein fraction high-density lipoprotein 3 (HDL3) in serum. This work was supported by The Norwegian Research Council for Science and the Humanities, The Norwegian Women's Public Health Association and The Norwegian Rheumatism Council. The technical assistance of Mrs J. Juul is greatly appreciated.

REFERENCES 1. Benditt, E. P., Eriksen, N., Hermodson, M. A. & Ericsson, L. H. (1971) FEBS Lett. 19, 169- 173. 2. Ein, D., Kimura, S., Terry, W. D., Magnotta, J. & Glenner, G. G. (1972) J . Biol. Chem. 247, 5653-5655. 3. Levin, M., Franklin, E. C., Frangione, B. & Pras, M. (1972) J . Clin. Invest. 51, 2773-2776. 4. Husby, G., Sletten, K., Michaelsen, T. E. & Natvig, J. B. (1973) Scand. J. Immunol. 2, 395-404. 5. Sletten, K. & Husby, G. (1974) Eur. J . Biochem. 41, 117-125. 6. Eriksen, N., Ericsson, L. H., Pearsall, N., Lagunoff, D. & Benditt, E. P. (1976) Proc. Nutl Acad. Sci. U.S.A. 73, 964-967. 7. Husby, G., Natvig, J. B., Sletten, K., Nordstoga, K. & Anders, R. F. (1975) Scand. J . Immunol. 4, 811-816. 8. Skinner, M., Cathcart, E. S., Cohen, A. S. & Benson, M. D. (1974) J . Exp. Med. 140, 871 -876. 9. Gorevic, P. D., Greenwald, M., Frangione, B., Pras, M. & Franklin, E. C. (1977) J . Immunol. 118, 1113-1118. 10. Hermodson, M. A., Kuhn, R. W., Walsh, K. A,, Neurath, H., Eriksen, N. & Benditt, E. P. (1972) Biochemistry, 11, 29342938. 11. Westermark, P., Sletten, K., Naeser, P. & Natvig, J. B. (1979) Scand. J . Immunol. 9, 193- 196. 12. Nordstoga, K. (1972) Acta Path. Microbiol. Scand. A , 80, 159168. 13. Anders, R. F., Natvig, J. B., Sletten, K., Husby, G. & Nordstoga, K. (1977) J . Immunol. 118, 229-234. 14. Pras, M., Schubert, M., Zucker-Franklin, D., Rimon, A. & Franklin, E. C. (1968) J . Clin. Invest. 47, 924-933. 15. Husby, G., Sletten, K., Michaelsen, T. E. &Natvig, J. B. (1972) Scand. J . Immunol. 1, 393 - 400. 16. Harada, M., Isersky, C., Cuatrecasas, P., Page, D., Bladen, H. A., Eanes, E. D., Keiser, H. R. & Glenner, G. G. (1971) J . Histochem. Cytochem. 19, 1- 15.

422

K. Waalen, K. Sletten, G. Husby, and K. Nordstoga: The Primary Structure of Amyloid Fibril Protein AA from Mink

17. Swank, R. T. & Munkres, K. D. (1971) Anal. Biochem. 39, 462 -477. 18. Omenn, G . S., Fontana, A. & Anfinsen, C. B. (1970) J . Biol. Chem. 245, 1895- 1902. 19. Kawasaki, I. & Itano, H. A. (1972) Anal. Biochem. 48, 546556. 20. Hartley, B. S. (1970) Biochem. J. 119, 805-822. 21. Gros, C. & Labouesse, B. (1969) Eur. J . Biochem. 7, 463-470. 22. Otnaess, A,-B., Little, C., Sletten, K., Wallin, R., Johnsen, S., Flengsrud, R. & Prydz, H. (1977) Eur. J . Biochem. 79, 459468.

23. Sletten, K., Husby, G. & Natvig, J. B. (1976) Biochem. Biophys. Res. Commun. 69, 19 - 25. 24. Segrest, J. P., Pownall, H. J., Jackson, R. L., Glenner, G. G. & Pollock, P. S. (1976) Biochemistry, 15, 3187-3191. 25. Benditt, E. P. & Eriksen, N. (1977) Proc. Natl, Acad. Sci. U . S . A . 74, 4025-4028. 26. Wegelius, 0. & Pasternack, A,, eds (1976) Amyloidosis, p. ix, Academic Press, New York.

K. Waalen and K . Sletten, Biokjemisk institutt, Universitetet i Oslo, Postboks 1041, Blindern, Oslo 3, Norway G . Husby, Revmatologisk avdeling, Universitetet i TromsB, N-9000 Troms4, Norway

K. Nordstoga, Norges VeterinaerhByskole, Oslo 4,Norway