Human Serum Amyloid A Protein

THEJOURNAL OF BIOLOGICAL CHEMISTRY Q 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 264, No. 31, issue of November 5, pp. 18363-18373 1989 Printed in d.S.A.

Human Serum Amyloid A Protein THE ASSIGN~ENT OF THE SIX MAJOR SOF FORMS TO THREE P U ~ L I S ~ EGENE D S E Q ~ N ~ E AND S EVIDENCE FOR TWO GENETIC LOCI* (Received for publication, April 28,1989)

IF.

Alistair F. StrachanS, Wolf BrandtB, Patricia Wooll, Deneys R. van der Westhuyzen, Gerhard A. Coetzee, Maria C. de Beer$, Enid G. ShephardS, and FrederickC. de BeerSJJ From the $Department of Internal Medicine, University of Stelknbosch Medical School, Tygerberg7505, Republic of South Africa, the 8Department of Biochemistry, University of Cape Town, Rondebosch 7700, Republic of South Africa, the YMolecular RheumatologySection, Clinical Research Centre, Harrow HA1 3UJ, United Kingdom, and the University of Cape To~n/MedicalResearch Council Unit for the Cell Biology ofAtherosclerosis, Department of Medical Biochemistry, University of c South Africu Cape Town Medical School, Observatory 7925, R e p ~ l i of

Serum amyloid A protein (apo-SAA) is an acutephase reactant and an apolipoprotein of high density lipoproteins (HDL). Six major isoforms ofapo-SAA occur in humans (PI6.0,6.4, 7.0, 7,4,7.5,8.0). In this report we have rationalized the phenotypic expression of apo-SAA isoforms with published apo-SAA structures predicted from apo-SAA cDNA’s pAl and pSAA82 and the genomic DNA SAAg9. The six apoSAA isoforms fall intothree pairs, PI 6.016.4, 7.017.5, and 7.418.0, which are products of cDNA p A l , cDNA pSAA82, and genomic DNA SAAg9, respectively. The second of each isoform pair (Le. PI 6.4,7.5, and 8.0) is the ”primary”product: a 104-residue peptide with the NH2-terminal sequence Arg-Ser-Phe-Phe. Each primary product is processed either to a major 103-residue peptide with the NHz-terminal sequence Ser-PhePhe orprocessed to a minor 102-residue product which results from the loss of both an Arg and a Ser residue from the NH2 termini. These “secondary” products have the lower PI values of 6.0, 7.0, and 7.4, respectively. The isoelectric points of the SAAg9 products were confirmed by expression of SAAg9 in transfected mouse L-cells. Both the PI 8.0 and 7.4 isoforms were present in cellular extracts, suggesting that post-translational modi~cationof apo-SAA may occur intracellularly. However, the greater relative abundance of the PI 7.4 isoform extracellularly suggests that the major conversion may occur after secretion. Whereas the gene corresponding to the pAlcDNA sequence does not show allelic variation, the segregation characteristics of the PI 7.017.6 and 7.418.0 isoform pairs amongst individuals suggests that these isoforms are the products of genes (with sequences corresponding to pSAA82 and SAAg9, respectively) which are allelic variants at a single locus distinct from that for thePI 6.016.4 isoform pair.

* This study was financially supported by the SouthAfrican Medical Research Council, the South African Arthritis and Rheumatism Association, and theUniversity of Cape Town. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 11 To whom correspondence should he addressed Dept. of Internal Medicine, P. 0. Box 63, Tygerberg 7505, Republic of South Africa. Tel.: 021-931.3131 (ext. 464). Fax: 021-931.7810.Telex: 526226 SA.

Serum amyloid A protein (apo-SAA)’ is the putative precursor of the 76-residue amyloidA protein (protein AA) which aggregates to form the fibrillar deposits characteristic of reactive systemic amyloidosis. Apo-SAAis an apolipoprotein of the high density lipoprotein (HDL) complex and plasma apoSAA concentrations are greatly elevated in individuals mounting an acute-phase response (to infection, trauma, etc.) (1-3). Moreover, apo-SAA is a polymorphic protein (4) and three phenotypic patterns have been reported ( 5 ) . In each pattern, isoforms with pl 6.0 and 6.4 occur. In addition, individuals may express either isoforms with PI 7.0 and 7.5 (pattern 2) or isoforms with PI 7.4 and 8.0 (pattern 3) or isoforms with pf 7.0,7.4,7.5, and 8.0 (pattern 1).The basis for these phenotypic expressions has notbeen established, but theobserved isoform patterns would be compatible with the transcription of various apo-SAA genes. Data from recent reports support this postulate (6-8) and suggest that three different human apo-SAA genes exist, corresponding to cDNA pAl (9), cDNA pSAA82 (lo), and SAAg9 genomic DNA (7). We have performed NHf-terminal sequence analysis and amino acid composition analysis on each of the major apoSAA isoforms. Using these data and by directly confirming the isoelectric pattern of the SAAg9 gene product (by analyzing the apo-SAA synthesized by mouse L-cells transfected with the SAAg9 gene), it was possible to assign each of the major apo-SAA isoforms to one of the three known human gene sequences. Furthermore, the isoform patterns observed in a group of 18 individuals suggest that the pSAA82 and SAAgQgene sequences are allelic variants and occur at a different locus from the pAl gene sequence. MATERIALS AND METHODS

Preparation of HDL-HDL was prepared using methods previously described from pooled plasma obtained from patients in “acutephase” (5). Informed consent and Ethical Committee approval were obtained. Preparation of Apo-SAA Isoform-200-pg portions of HDL (see above) were freeze-dried and each was delipidated with 0.5 ml of chloroform/methanol(21, v/v) (11).The pellets were resuspended in 20 pl of 1% (w/v) sodium decyl sulfate, 7 M urea, 5% (v/v) 2mercapt~thanol.A total of 36 portions were electrofocused on 0.3 mm polyacrylamide gels containing ampholytes (Pharmacia LKB

* The abbreviations used are: apo-SAA, serum amyloid A protein; apo, apolipoprotein; HDL, high density lipoproteins; SDS-PAGE, polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate; kb, kilobase(s); DMEM, Dulhecco’s modified Eagle’s medium; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonicacid; CHAPS,3-[(3-cholamidopropyl)dimethylammonio]-l-propanesu~fonic acid; PTH, phenylthiohydantoin.

18368

18369 Biotechnology Inc.) in the following proportions: pH 3-10, 20% (vi v);pH 4-6.5, 40% (v/v); pH 7-9, 40% (v/v). Full details of the methodology have been described (5). Coomassie Blue-stained bands of the major apo-SAA isoforms (PI 6.0,6.4, 7.0, 7.4, 7.5, and 8.0) (5) wereexcised from the gel and soaked in 10 mM N-cetyl-N,N,Ntrimethylammonium bromide, 10 mM cysteamine, 0.5 M acetic acid for 1 h. The protein together with the Coomassie Blue was electroeluted at a field strength of 10 V/cm for 3 h in the presence of an electrode solution of 0.5 M acetic acid. The electroelution chamber contained the detergent buffer (see above). The electroelution device comprised two vertical chambers partitioned from the electrode buffer by two dialysis membranes. A collecting chamber was formed by inserting a cellulose acetate filter(pore size = 0.2 pm). All membranes were sealed by silicone O-rings and access to the collecting chamber was gained via a small port. The electroeluted protein solution (300 pl), which also contained Coomassie Blue stain, was freeze-dried in Eppendorf vials. The dried material was redissolved in 100 pl methanol to which 10 pl HCl and 2 ml of acetone was added, they were then left to stand for at least 1 h at -20 "C. The precipitate was centrifuged and washed twice with acetone and dried in a desiccator. The protein was redissolved in water and subjected to amino acid sequence and composition analysis. NHZ-terminal Sequence Analysisand Amino Acid Content-NHzterminal sequence analysis was performed using a gas-phase sequenator as described previously (12). The amino acid content of the isoforms was determined by the method of Bidlingmeyer et aL (13). Proteins were hydrolyzed in 5.7 M HC1, 0.3% (v/v) thioglycollic acid by the gas-phase procedure for 90 min a t 150 "C. Prediction of Isoelectric Points-The isoelectric points of the apoSAA protein sequences derived from the predicted nucleotide sequences of cDNA pA1,cDNA pSAA82, and genomic DNA SAAg9 (7, 9, lo), were predicted as follows, where the charge on an amino acid can be calculated at any given pH using the acid dissociation constant. For basic amino acids:

For acid amino acids: A-

16. KO,+ H'

where nx corresponds to the number of amino acid X and KO,to its acid dissociation constant. The single-letter code for amino acids has been used. As the netcharge on a protein iszero a t its isoelectric point, given the dissociation constants, the PI can be readily calculated using a computer by substituting different proton concentrations until the net charge is zero. Conditioned Medium-Mononuclear cells (15 X 106/60-mm dish containing 5 ml of RPMI, 70% (v/v) autologous human serum), were incubated at 37 "C for 2 h and the non-adherent cells removed. The adherent cells were washed twice with RPMI and incubated in dishes containing 5 ml of RPMI (minus serum) containing 10 pg/ml lipopolysaccharide, M dexamethasone, and 0.02 units/ml insulin, for 24 h at 37 "C. The medium was harvested and centrifuged to remove non-adherent cells, and passed through a 0.2 pm filter unit. Apo-SAA from Transfected Cells-A human apo-SAA genomic clone (SAAg9)was previously transfected into mouse L-cells and stable lines obtained (7). This clone is in the X phage vector L47.1, and contains 5 kb of 5'- and approximately 10 kb of 3'-flanking sequences. The transfected cells were seeded at a density of 0.3 X IO6 cells/dish (60 mm diameter) containing Dulbecco's modified Eagle's medium, 10% (v/v) fetal bovine calf serum, 13.6 pg/mlhypoxanthine, 0.17 pg/ml aminopterin, 3.9 pg/ml thymidine, and 4 mM glutamine

(DMEM-HAT). The medium was replaced after 72 h and 96 h. 24 h later the dishes were subjected to two separate treatments: 4 ml of DMEM-HAT was applied to some dishes and 2 ml of DMEM-HAT plus 2 ml of monocyte-conditioned medium (see above) was applied to theother dishes. The dishes were incubated overnight a t 37 "Cand then washed twice with phosphate-buffered saline before the addition of 1.5 ml/dish minimal Eagle's medium (methionine-free), 5%.(v/v) lipoprotein-deficient serum containing [35S]methionine(100 pCl/rnl). The dishes were incubated at 37 "C for 3 h. The medium from each dish was centrifuged in a Microfuge for 10 min and the supernatant stored at -70 "C. The cells were washed twice with 2 mi of 10 mM HEPES, 150 mM NaCI, 2 mM CaC12,pH 7.4, and then lysed by the addition of two batches of 200 p1 of 10 mM HEPES, 200 mM NaCl, 2 1mM phenylmethylsulfonyl fluoride (0.5% mM CaCL,, 2.5 mM MgC12, (v/v) in dimethyl sulfoxide), 0.1 mM leupeptin, 1%(v/v) Triton X100, pH 7.4, per dish and harvested by scraping with a Teflon policeman. Lysates were centrifuged in a microcentrifuge for 10 min. 320 pl of each supernatant was removedand stored a t -70 "C. Preformed immune complexes containing specific apo-SAA antibody were used to isolate apo-SAA in the cell lysates and culture medium. The methods used were essentially as described by Tolleshuang et ai. (13, 14) for the immunoisolation of the low density lipoprotein receptor from cultured cells. 300 pg of immunopurified rabbit anti-human apo-SAA in 75 pi of phosphate-buffered saline (containing 10 mM sodium azide), 3 mg (300 pl) of swine anti-rabbit IgG (immunoglobulin fraction, DAKO, Denmark), and 300 pl of immune complex buffer (50 mM Tris-HC1, pH 8.0, containing 200 mM NaCl, 1 mM EDTA, 0.5% (v/v) Nonidet P-40) were incubated a t 4 "C overnight. The tubes were then microcentrifuged for 30 s (at 4 "C),the pellets resuspended in 1 mi of immune complex buffer, and the tubes centrifuged again. The pellets (immune complexes) were resuspended in 1 ml of immune complex buffer and stored at 4 "C. 40 pl of immune complexes was added to each 320-111 aliquot of cell harvest or 700-p1 aliquot of medium and themixtures incubated on a rocking platform for 1 h at room temperature. The tubes were microcentrifuged for 5 min (4 "C) and the pellets dissolved in 200 pl of 10 mM HEPES, pH7.4, containing 200 mM NaCI, 2 mM CaC&,2.5 mMMgC12, 1 mM phenylmethylsulfonyl fluoride (0.5% (v/v) in dimethyl sulfoxide), 0.1 mM leupeptin, 1% (v/v) Triton X-100 before loading on top of a 4-step (750 pl each step) sucrose gradient. Layer 1 (top): 10% (w/v) sucrose, 0.5% (v/v) CHAPS. Layer 2: 20% (w/v) sucrose, 0.5 mM NaCl, 0.2% (v/v) Nonidet P-40. Layer 3: 30% (w/v) sucrose, 0.1% (w/v) SDS, 0.2% (v/v) Nonidet P-40. Layer 4 (bottom): 40% (w/v) sucrose. All sucrose solutions contained 10 mM Tris-HCl, pH 8, 1 mM phenylmethylsulfonyl fluoride. The gradients were centrifuged a t 2000 rpm (Beckman TJ-6) for 30 min and the supernatants poured off. The pellets were resuspended in 200 pl of 50 mM TrisHC1, pH 8, 2 mM CaCl,, microcentrifuged for 20 s and dissolved in either: (i) 30 PI of 4 M urea, 100 mM dithiothreitol, 10% (v/v) glycerol, 5% (v/v) 2-mercaptoethanol, 2.4% (w/v) SDS, 7.5 mM Tris-HC1, pH 6.8, and boiled for 3 min prior to separation by 5-20% (w/v) SDS/ PAGE; or (ii) 30 pl of 7 M urea, 0.1% (w/v) sodium decyl sulfate, 5% (v/v) 2-mercaptoethanol for electrofocusing. SDS-PAGE gels were stained, destained, washed in distilled water for 30 min and soaked in 1 M sodium salicylate for 30 in before being dried, sprayed with Amplify (Amersham International) and exposed to Kodak XAR-5 film at -70 "C. Electrofocusing gels werefixed in isopropanol/water/ acetic acid (25:65:10,by volume) for 10 min, immersed in destain solution (35% (v/v) ethanol, 10% (v/v) acetic acid solution) for 5 min, immersed in staining solution (0.5% (w/v) Coomassie Blue in destain solution) for 5 min, destained for two 5-min cycles, immersed in water for 5 min, soaked in Amplify (Amersham International) for 20 min rinsed with distilled water, dried, and exposed to Kodak XAR5 film a t -70 "C. RESULTS

~ ~ * - Sequence ~ e Analysis-Each ~ ~ ~ isoform ~ Z was subjected to seven cycles of automatic gas-phase isothiocyanate degradation and from the NHz-terminal sequences it is apparent that the isoforms fall into two groups (Table I). The PI 8.0, 7.5,and 6.4 isoforms possess the same NHZ-terminal sequence, Arg-Ser-Phe-Phe-Ser-Phe-Leu. The PI 747.0, and 6.0 isoforms in each case consist of twosequences. The predominant sequence is Ser-Phe-Phe-Ser-Phe-Leu-Gly (80%),with a minor component Phe-Phe-Ser-Phe-Leu-Gly-

18370 TABLE r PTH-derivatives recovered on degrading electroeluted apo-SAA isoforms Yield of the PTH-derivatives in picomoles after subjecting the various electroeluted apo-SAA samples to gas-phase isothiocyanate degradation. APO-SAA

pH 6.4 Cycle

PTH

Arg Ser Phe Phe Ser Phe Leu

pmol

Cycle

2nd sequence

PTH

pmol

PTH

pmol

404 733 960 400 653 680 591

Phe Phe Ser Phe Leu Gly Glu

178 200" 44 222 182 133 89

Phe Phe Ser Phe Leu Gly Glu

111 120" 22 102 124 53 44

422 422 858 924 347 782 711

1 2 3 4 5 6 7

Ser Phe Phe Ser Phe Leu Glv

22 53 120 120 22 82 63

1 2 3 4 5 6 7

Ser Phe Phe Ser Phe Leu Gly

pH 7.0

133 284 284 84 231 222

156

~-

~

1st sequence

PH 7.4

pH 8.0

Arg Ser2 Phe3 Phe

Apo-SAA

pH 6.0

pH 7.5

Arg Ser Pbe Phe Ser Phe Leu

TABLEI1 Amino acid compositionof the three up-SAA isoforms deduced from the DNA sequence(Fig. 1) The number of amino acid residues/molecule and the number of aminoacids/100residues (mol %) havebeencalculated from the amino acid sequences (Fig. 1). Differences in the composition in the three proteins are indicatedbv asterisks. NC, not calculated. Amino acid

Asx Thr

Ser Glx Pro G~Y Ala CYS Val Met Ile Leu TY~ Phe Trp His LYS Arg

PA1 Residue mol

14 0 7 9

SAAg9 %

14.4

o.o* 7.2* 9.3*

NC 12 16 0

1 2 3 3 5 8

12.4 16.5 0.0 1.0 2.1 3.1 3.1* 5.2 8.2*

NC 3 4 10

pSAA82

Residue

mol %

Residue mol

14 1 7 8 NC 12 16 0 1 2 3 4 5 6

14.4

14 1 7

LO* 3.2* 8.2* 12.4 16.5 0.0 1.0 2.1 3.1 4.1* 5.2 6.2*

NC 3.1* 4.1 10.3*

2 4 12

8 NC 12 16 0 1 2 3 4 5 6

% '

14.4

LO* 7.2' 8.2* 12.4 16.5 0.0

1.0 2.1 3.1 4.1* 5.2 6.2*

NC 2.1* 4.1 12.4'

3 4 11

3.1* 4.1 11.3*

Ser 102 71 Phe Total 98 97 100 100 98 100 Phe 150 80" Phe 9 Phe Ser 178 8 9 8 4 Ser 17 Phe 44 ASP 6 7 Glu 6 Ser 5 Phe 151 36 Leu Phe 6 Leu 164 36 GIY Basic 18 17 I8 Leu 7 GIu 27 77 GlY 16 Acid __ 14 14 8 Glu Ala 32 102 9 9 Ala Phe 27 116 As Phe occurs in both sequencesthe relative amounts have been absence of Thr, an increased Glx content, and a lower Leu content. However, discrepancies between the expected (Table estimated based onthe yield of Phe in cycle 1. 1 2 3 4 5 6 7 8

22 22 138 124 18 133 124

1

11) and the observed (Table 111) amino acid compositions (for the pAlgene products) did occur (e.g. with Ser, Glx, Gly, and apo-SAA ~ s F F ~ F L G E A F D G - ~ Y s D ~ - ~ G s D ~Arg). ~ ~ The ~ ~ Gpi8.0 ~ ~and ~ 7.4 isoforms closely resembledthe PI 7.5 (PA1) * and 7.0 isoforms, respectively, and both pairs of isoforms apo-SAA R S F F S F L G E R F D G A R D ~ Y S D ~ ~ I G S D K P F H A R G closely resembled the SAA@/pSAA82-derived compositions (SAW91 apo-SAA R S F F S F L G ~ D G A R D M S D ~ ~ I G S D X ~ ~ G ~ D ~ G P G G A n A (although discrepancies between the observed and expected (PSAA82) c compositions were evident). The PI 8.0, 7.5, and 6.4 isoforms 55 fo $5 ?a 7.5 y o ?5 10 6s I ~ O each had one more Arg residue than the PI 7.4, 7.0, and 6.0 apo-sm ~E~IS~AREN~QRFF~XGAE~S~~-~GRSG~P~~RP~~PEXY isoforms, respectively, reflecting the generation of the latter * * * * ( PA1 1 apo-SAA AEVISNARENIPRLTGRGAEDSLiWQAANKWGRSGRDPNHFRPAGLPEXY isoforms by trimming of the NH2-terminalArg residue of the ** * * * (SAA99) * 104-residueisoforms. apo-Sa AEVISNARENIQRLTGHGAEDSLADQAMXWGPSGRDPN (psnnsz) * * ** t * * pl Prediction-Given the limitations in amino acid comRe. 1. Derived-amino acid sequence from three published position analysis (see above), we proceeded to perform calcuhuman apo-SAA cDNA/genomic DNA sequences (Refs. 7-9). lations of the isoelectric points of the isoforms (based on the Differences are indicated by an asterisk. amino acid sequences predicted by pA1, pSAA82,and SAAg9) by computer assisted analysis of the pK, values of the conGlu (20%). All the isoforms are closely related to one another stituent residues. Table IV, top, displays a calculation of the charge distriand the second group (PI 7.4, 7.0, and 6.0) could have arisen from the first by removal of an NHp-terminalArg residue or, bution at pH 6.4 of the 104-residueapo-SAA isoform deduced from the pAl sequence. The calculated and experimentally less frequently of an Arg residue plus a Ser residue. determined isoelectric points of the apo-SAA isoforms (104 A m ~ ~Acid o C o m ~ s i ~ i o ~ - T h r eapo-SAA e codingDNA sequences have been reported (7, 9, 10). The correspm\ding and 103 residues) derived from the three gene sequences,and amino acid sequences are shown in Fig. 1 and the d e d u 4 the calculated and reported isoelectric points of certain standamino acid compositions in Table 11. The compositions of the ard proteins(16) are shown in Table IV, bottom. The isoforms proteins are closely related but there are nevertheless signif- showed a deviation of determined and predicted PI values at icant differences. When our six apo-SAA isoforms were sub- higher pH (Table IV, bottom). A similar deviation occurred jected to amino acid composition analysis (Table 111) it be- with standard proteins (16). When their PI values were calcame apparent that, despite the limited accuracy of this culated by the above method and compared with the reported technique, the PI 6.4 and 6.0 isoforms resembled the pAl isoelectric points (16), the calculated and experimentally dederived composition and differed from the other isoforms by termined values showed an increasing divergence as pH inhaving two additional Phe residues, a lower Arg content, the creased. Thus the PI calculations performed here were only f

5

I?

:1

*t

2p

27

3:

37

4p

47

5;

Apo-SAA Isoforms

18371

TABLE 111 Amino acid composition of the electroeluted apo-SAA isoforms obtained after electrofocwing ND, not determined; the number in brackets corresponds to the most likely number of residues (res) in the protein. A~o-SAA acid Amino

pH 6.4

pH 6.0

pH 8.0

pH 7.4

pH 7.5

pH 7.0

mole % (Res)

mole % (Res)

mole % (Res)

mole 5% (Res)

mole % (Res)

mole % (Res)

Asx Thr Ser Glx Pro G~Y Ala CYS Val Met Ile Leu TYr Phe Trp His LYs Arg

14.7 0.1 6.7 9.9

(14)

14.7 (14) 0.1 ( 0 ) 6.5 (8) 10.1 (10)

(0) (6) (10)

ND

ND

12.9 (13) 17.0 (16) 0.0 (0) 1.1 (1) 2.2 (2) 3.2 (3) 3.2 (3) 5.1 (5) 8.0 (8)

(14) (1) (7) (9)

15.1 (15) 1.2 (1) 7.0 (7) 9.3 (9)

ND

12.9 (13) 16.9 (16) 0.0 ( 0 ) 1.2 (1) 2.2 (2) 3.3 (3) 3.4 (3) 4.1 (5) 8.0 (8)

14.5 1.2 6.9 9.1

12.6 (12) 16.0 (15)

0.0 (0)

0.0 (0)

2.2 1.9 3.1 5.6 3.9 5.2

1.6 2.1 3.4 4.7 4.8 5.8

(2) (2) (4) (5) (4) (5)

ND

(14) (1) (7) (9)

15.1 (15) 1.2 (1) 6.6 (6) 9.2 (9)

ND

ND

12.5 (12) 13.5 (13)

ND

ND

Total

14.5 1.3 7.7 9.6

ND 12.9 (13) 16.8 (16) 0.0 ( 0 ) 1.2 (1) 2.1 (2) 3.4 (3) 4.4 (4) 4.2 (4) 5.9 (6)

12.5 (12) 15.8 (15) 0.0 ( 0 ) 1.3 (1) 2.1 (2) 3.1 (3) 4.5 (4) 5.3 (5) 5.9 (6)

(2) (2) (3) (5) (5) (6)

ND

ND

ND

3.0 (3) 3.9 (4) 9.0 (9)

3.1 (3) 4.1 (4) 8.6 (8)

2.1 (2) 4.8 (5) 11.4 (11)

2.2 (2) 4.1 (4) 10.3 (10)

2.8 (3) 4.5 (4) 10.5 (10)

3.2 (3) 4.1 (4) 9.7 (9)

100.0 (97)

100.0 (96)

100.0 (96)

100.0 (98)

100.0 (96)

100.0 (96)

TABLEIV Charge distribution on apo-SAA (pal) at pH 6.4 (PI) (top); predicted and observed PI values of apo-SAA isoforms and standard proteins (bottom) Calc., calculated from the amino acid sequence. Deter., determined experimentally. Diff., difference between Calc. and Deter. (see text). Charge on residues Amino acid

pK.

Residues (n)

Basics (H'I(K.

X

Asp Glu Tyr Cys COOH His Lys Arg NH,

4.4 4.3 10.3 8.3 3.3

9 7 5

6.7 10.5 12.5 8.5

3 4 10 1

+ H'))

(KJH

+ K.))

X

-0.99 -0.99 -0.00 -0.02 -1.00

1

0.61 1.00 1.00 0.99

n (X) -8.9 -6.9 -0.0 -0.0 -1.0

0.0 Calc. Residues

103 104 103 104 103 104

-68 -45

Ape-A-1 -

1.8 4.0 10.0 1.0

Total Insulin (bovine) Myoglobin (bovine) Myoglobin (sperm whale) Ribonuclease (bovine) APO-SAA(pAl) APO-SAA(pAl) APO-SAA(pSAA82) APO-SAA(pSAA82) APO-SAA(SAAgS) ADO-SAA(SAAnS)

- 96

-27

All Acidics

0

Protein

Mr(~lO-~l

Deter.

Diff.

5.5 7.8 9.1 9.9

5.4 7.3 8.2 8.8

0.1 0.5 0.9 1.1

6.0 6.5 7.8 8.9 8.9 9.5

6.0 6.4 7.0 7.5 7.4 8.0

0.0 0.1 0.8 1.4 1.5 1.5

able to predict the relative isoelectric points of the isoforms. As can be seen (Table IV, bottom),the 104-residueproduct of SAAg9 has a higher PI value than that of pSAA82 which is, in turn, higher than that of pAl. The 103-residuederivative of each 104-residue isoform acquires a more acidic PI upon the loss of the NH2-terminal Arg residue. Because of the unexplained divergence of calculated and experimentally de-

Apo-SAA APO-A-D(monomer)

,

-18 -14 2

3

4

5

6

RG.2. Autoradiograph of 5-209'0 (w/v) SDS/PAGE of immunoprecipitated harvests from transfected cells. Lane 1, acute-phase HDL-positions of apoA-I, apo-SAA, and apoA-I1 (monomer) are shown. Lane 2, harvests from cells exposed to DMEMHAT only. Lane 3, harvests from cells exposed to DMEM-HAT and monocyte-conditioned medium. Lane 4, medium collected from cells exposed to DMEM-HAT only. Lane 5, medium collected from cells exposed to DMEM-HAT and monocyte-conditioned medium. Lane 6,molecular mass markers of 96,68,45,27, 18, and 14-kDa positions of Coomassie-stained bands are shown (see "Materialsand Methods" for full details).

termined PI values at higher pH, absolute PI assignments were impossible. Apo-SAA from Transfected Cells-The PI of the SAAg9 products was directly confirmed by the isolation of [35S] methionine-labeled apo-SAA from transfected mouse L-cells (containing genomic SAAg9) by immunoprecipitation and analysis by5-20% (w/v) SDS-PAGE and electrofocusing. SDS-PAGE analysis (Fig. 2) revealed the immunoprecipitation (from cells and medium) of an [35S]methionine-labeled protein with an apparent molecular weightsimilar to thatof apo-SAA.Clearlymoreapo-SAAwas present both in the medium and in the cells in the dishes that had been exposed to conditioned medium than in untreated controls. A high degree of constitutive expression was nevertheless apparent as reported previously (7). In all cases electrofocusing of medium and cellular extracts showed the presence of the PI

18372

Apo-SAA Isoforms

further NH2-terminaltrimming may impair the ability of apoSAA to bind to HDL or other lipid complexes such that it would not be isolated by our techniques. Dwulet et al. (6) recently described the structures of multiple forms of apo-SAA from a single individual. They partially resolved the apo-SAA isoforms using anion-exchange I chromatography and reverse-phase high performance liquid chromatography. Two of the isoforms characterized (SAAI and SAA, des-Arg) correspond to thePI 6.4 and 6.0 isoforms of our study. They did not observe a 102-residue isoform in the PI 6.0 population. Although unable to resolve the isoforms with PI 7.0, these authors reported the presence of two very 1 2 3 4 5 6 similar sequences within a singlepeakwhichdiffered by FIG.3. Electrofocusing of immunoprecipitated harvests having either arginine or histidine at position 71. (Each of from transfected cells. An autoradiograph of the electrofocusing these isoforms also had a 103-residue des-Arg form.) They gel (lanes 1-5) was contact-printed onto photographic paper. Lane 1, cell harvests after treatment of cells with DMEM-HAT only. Lune 2, labeled the isoform with histidine at position 71, SAA 2a and cell harvests after treatment of cells with DMEM-HAT and mono- the isoform with arginine at position 71, SAA 28. The SAA cyte-conditioned medium. Lune 3, delipidated acute-phase HDL apo- 2a/SAA 2a des-Arg corresponds to thepSAA82 cDNA prodlipoproteins (not radioactively labeled). Lane 4, medium from cells uct and hence the PI 7.5/7.0 pair described herein and SAA treated with DMEM-HAT only. Lune 5, medium from cells treated 2p/SAA 28 des-Arg to the SAAg9 gene product and the PI with DMEM-HAT and monocyte-conditioned medium. Lune 6,Coo- 8.0/7.4 pair described herein. The authors did not identify the massie Blue-stained lane 3. The isoelectric points of the six major apo-SAA isoforms are indicated (see “Materials and Methods” for corresponding 102-residueapo-SAA isoforms ineither case. It is noteworthy that while the pAl sequence differs from full details). that of pSAA82 at 7 positions (residues 52, 57, 60,68,69,84, 8.0 and smaller amounts of an isoform of apo-SAA in the PI and 90) and from SAAg9 at 8 positions (residues 52, 57, 60, 68, 69, 71, 84, and go), the pSAA82 and SAAg9 sequences 7.4-7.5 region (Fig. 3) differ at only one position (Fig. 1).At this position (residue 71)only a single base change (CAU + CGU) wouldbe DISCUSSION required to encode the different residue. In this study we have attempted to assign the major apoKluve-Beckerman et al. (8) reported the isolation of three SAA isoforms separated by electrofocusing to the published mRNAs from a single individual and postulated a minimum gene sequences. From our data we can make the following of two human apo-SAA genes.They proposed that each apoassignments. SAA gene sequence couldeither exist at a separate locus or, (i) The PI6.0 and 6.4 isoforms are the products of an apo- alternatively, that two of the three sequences could bealleles SAA gene correspondingto cDNA pAl(9). ThePI 6.0 isoform at a single locus. We have previously reported (5) that three (103 residues), whichlacks the NH,-terminal Arg residue, distinct patterns of apo-SAA isoforms were observed when results from proteolytic processing of the PI6.4 isoform. The the apolipoproteins from HDL of 18 patients in acute phase observed values of PI 6.0 and PI 6.4 for this pair of isoforms were analyzedby electrofocusing.The PI 6.0/6.4 isoform pair correspond to those predicted. About20% of the PI6.0 isoform was present in all these individuals and, asdiscussed aboveis population has beenproteolyzed further to a 102-residue the product of a gene corresponding to the pAl cDNA seprotein. Expectedly, the loss of the Ser residue doesnot alter quence (8). Although no allelic variation for this gene was the PI of this subpopulation of the PI 6.0 isoform. detected on the basis of electrofocusing, it remains a theoret(ii) A second gene corresponding to the pSAA82cDNA ical possibilitythat allelic variation does exist.In contrast the sequence produces a 104-residue apo-SAA isoformwith a PI PI 7.0/7.5 and 7.4/8.0 pairs were found to segregate in the 7.5. The prediction gives this isoform a higher PI than the sense of coexisting in 33% of individuals or of appearing pAl product. Proteolytic processing produces a 103-residue exclusively at the expense of each other (PI 7.0/7.5 in 61% of isoform lacking an NH2-terminal Arg, which has a PI of 7.0. individuals and PI 7.4/8.0 in 6% of individuals). Considered About 20% of the PI 7.0 population consists of an isoform together, all our results strongly suggest that these apo-SAA which has lost the NHZ-terminal Arg and Ser residues. isoforms are the produds of allelic variants at a single locus (iii) The SAAg9 gene produces isoformswith PI 8.0 and 7.4 distinct from that coding forthe PI 6.0/6.4 isoforms. having 104 and 103 residues, respectively. The gene product is structurally very similar to thepSAA82 product. However, Acknowledgments-Prof. W. Gevers is thanked for critically reada His + Arg substitution at position 71 results in the isoelec- ing the manuscript and B. Henstock for excellent secretarial assisttric points of the PI 8.0/7.4 pair being more basic than the ance. Dr. Rachel Saunders kindly assisted in the preparation of pSAA82 products. The isoelectric points of the SAAg9 gene monocyte-conditioned medium. products were confirmed directly by the expression of the REFERENCES SAAg9 gene in mouse L-cells. Analysisof SAAg9 gene prod1. Strachan, A. F., De Beer, F. C., Coetzee, G. A., Hoppe, H. C., ucts in transfected L-cell cultures revealed the presence of Jeenah, M. S., and Van der Westhuyzen, D. R. (1986)Protides both the PI 8.0 and 7.4 isoforms incellular extracts as well as Biol. Fluids P m . Collog. 34,359-362 in the culture medium. This suggests that thepartial proteo2. Coetzee, G . A., Strachan, A. F., Van der Westhuyzen, D.R., lytic conversionof the PI 8.0 isoform occursintracellularly. Hoppe, H. C., Jeenah, M. S., and De Beer, F. C. (1986)J. Biol. Chem. 261,9644-9651 We have not identified any apo-SAA isoforms in which 3. De Beer, F. C., Mallaya, R. K., Fagan, E. A., Lanham, J. G., NHz-terminal trimming has proceededbeyond position 2 Hughes, G. R. V., and Pepys, M.B. (1982)Lancet 11,231-233 (Ser). It is possible that aromatic amino acids at positions 3 4. Bausserman, L. L., Saritelli, A. L., Herbert, P. N., McAdam, K. and 4 are resistant to attack by this exopeptidase. AlternaP. W. J., and Schulman, R. S. (1982)Biochirn. Biophys. Acta tively, since an a-helical lipid binding domain of apo-SAA 704,556-559 5. Strachan, A. F.,De Beer, F. C., Van der Westhuyzen, D. R., and beginning at residue 3 can be predicted, it is possible that

18373 Coetzee, G. A. (1988) Biochem. J. 2 5 0 , 203-207 11. 6. Dwulet, F. E., Wallace, D. K., and Benson, M. D. (1988) Biochemistry 27,1677-1682 7. Woo, P., Sipe, J., Dinarello, C. A., and Colten, H. R. (1988) J. Biol. Chem. 262,15790-15795 K1uve-Beckerman’B’’ F’ E’’and M’ (1988) ‘’ Clin. Invest. 82,1670-1675 9. SiPe, J. Dv Colten, Ha R.9 Goldberg, D.9 Edge, D.9 Tack, B. F,, Cohen, A. S., and Whitehead, A. S. (1985) Biochemistry 24, 2931-2936 10. Kluve-Beckerman, B.,Long, G . L., and Benson, M. D. (1986) Biochem. Genet. 24,795-803

12. 13. 14.

15. 16.

Folch, J., Lees, M., and Sloane-Stanley, G . H. (1957) J. Biol. Chem. 225,497-509 Brandt, W. F., Alk, H., Chauchan, M., and Von Holt, C. (1984) FEBS Lett. 174,228-232 Bidlingmeyer, B. A., Cohen, S. A., and Tanvin, T. L. (1984) J. Chromato~r.336,93-104 Tolleshaug, H.,Goldstein, J. L., Schneider, W. J., and Brown, M. S. (1982) Celt 30,715-724 To1leshaug, H,, Hobgood, K. K., Brown, M. s.,and &)&&in, J. L. (1983) Cell 32,341-351 Andrews, A. T. (1981) in E ~ c t r o p ~ r eTheory, s ~ : tech^^^^ and Biochemical and Clinical Applications, p. 203, Oxford University Press, Oxford