A simple, sensitive technique for classification of apolipoprotein(a) isoforms by sodium dodecyl sulphate-polyacrylamide gel electrophoresis

C&&a Chimica Acta, 207 (1992) 215-225 @ 1992 Elsevier Science Publishers B.V. All rights reserved. ~-8981/9~~05.~

215

CCA 05281

A simple, sensitive technique for classification of apolipoprotein(a) isoforms by sodium dodecyl sulphate-polyacrylamide gel electrophoresis M. Farreravc, F.L. Gameayc, PC. Adamsb, M-F. Lakera and K.G.M.M. Alberti” *Department of Clinical Biochemistry, bDepartment of Cardiology and cDepartment of Medicine, University of Newcastle upon Tyne, The Medical School, Framlington Place, Newcastle upon Tyne (UK) (Received 2 October 1991; revision received 20 February 1992; accepted 9 March 1992)

lvey wordx A~li~protein(a)

isoforrns; El~tropho~sis;

Lipoprotein(a)

Summary Lipoprotein(a) (Lp(a)) is a lipoprotein containing a unique glycoprotein, apolipoprotein(a) (ape(a)), which shows considerable heterogeneity of apparent molecular mass on sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). A unifying classication of isoforms has been lacking. A simple sensitive procedure for classifying ape(a) isoforms was developed in which the relative mobility of ape(a) on SDS-PAGE was related to that of apolipoprotein (apo) B-100 (&vs 8). After Western blotting ape(a) bands were visual&i by a sensitive double antibody technique employing commercial polyclonal antibodies (sheep antihuman Lp(a) antibody, alkaline phosphatase-linked donkey antisheep antibody). The technique was sensitive (lower limit of detection 0.02 pg ape(a)) and had good reproducibility (coefficient of variation 0.9-6.4%). Ten isoform mobilites are described (~0.35, 0.40, 0.50, 0.60, 0.70, 0.80, 1.0, 1.10, >1.15). Individuals may have single or double band phenotypes. This classification is compatible with those previously described and the method is suitable for many laboratories, as it employs standard equipment and commercially available materials.

Corres~ndence to: Dr. M. Farrer, Cardiothoracic upon Tyne NE7 7DN. UK.

Unit, Freeman Hospital, Freeman Road, Newcastle

216

Lipoprotein(a) (Lp(a)) has attracted much attention since its discovery by Berg [ 1] because of its strong association with risk for coronary artery disease [2-6). Lp(a) contains a unique glycoprotein, apolipoprotein(a) (ape(a)) linked by a disulphide bond to an apolipoprotein B-100 (apoB) molecule indistinguishable from that of LDL [7,8]. Ape(a) has amino acid 191 and cDNA sequence homology with plasminogen [lo] but, unlike plasminogen, has multiple repeats of the kringle 4 subunit and an inactive serine protease [lo]. Apoca) has a uo~id~rable size heterogeneity on SDS-PAGE [l l- 131. The number of the kringle 4 repeats varies between isoforms and this can be demonstrated to be due to diffe~n~gGnoty~s 114,151.The molecular basis of ape(a) polymorphism is unclear and an unequivocal classification of the isoforms of apofa) is also lacking. Some family studies suggest that phenotype expression (determined by isoforms present on SDS-PAGE) is wholly genetically determined [ 12,161. Others 1131suggest that phenotype expression does not fit a classical Mendelian inheritance pattern. The high carbohydrate content of the isoforms affects mobility on SDS-PAGE [12,13] and it is possible that subtle changes in mobility could reflect post translational modification. Initially 6 isoforms of ape(a) were described [12] with no more than two isoforms per subject. Many (>40%) null phenotypes were reported although this may reflect the in~nsiti~ty of early t~hniques. More sensitive methods have now been developed which identify isoform bands in all but 1% of population samples with 1 l- 12 different isoforms being visually identified [I 3f. A recent report describes 19 genotypes on the basis of different length restriction fragments which included the variable number of kringle 4 repeats [ 151. The present report describes a simple, sensitive method for identification and classification of ape(a) isoforms, based on mobility on SDS-PAGE. Methods Blood cotiection and storage Blood was collected into plain tubes, centrifuged at 4°C and the serum divide into portions which were stored frozen at -9O’% pending assay or phenotype analysis. Immunoqwntitation of apoia) Ape(a) was measured by an enzyme-linked immunosorbent assay (ELISA) technique (Biopool, Umea, Sweden). This assay employs a polyclonal goat antihuman Lp(a) antibody as both a capture and signal antibody. Standard human reference sera, commercial, (Immune GMBI-I, Vienna, Austria; Biopool) and an internal control sample from a single patient of known B mobility isoform, were included in each analysis. The performance characteristics of the antibodies are described in this report and elsewhere [17]. All samples were assayed using antibody from a single batch.

217

Electrophoretic procedure A vertical 14 lane gel system with discontinuous gel and buffer [18] was used. Running gels contained 6% acrylamide and 0.038% piperazine diacrylamide (PDA; Biorad, Richmond, CA) in 375 mmol/l Tris-HCl, pH 8.7 with 1 giI SDS. Stacking gels contained 3% acrylamide and 0.08% PDA in 125 mmol/l Tris-HCl, pH 6.8 with 1 g/l SDS. Portions of serum (lo-40 ~1, the volume depending on ape(a) concentration) were diluted to 250 ~1 with SDS-PAGE sample buffer containing 16 mmol/l Tris-HCl, 10 mmol/l Na EDTA, 100 g/l glycerol, 50 g/l SDS and saturated bromophenol blue, boiled for 10 min with mercaptoethanol (final cont. 20 g/l). Mercaptoethanol was included to disrupt the disulphide linkage between ape(a) and the apoB molecule to which it is bound. Samples were loaded to achieve between 0.06 pg and 2.0 pg ape(a) per lane. Duplicate gels were prepared. A serum sample containing an ape(a) isoform known to have a mobility identical to apo B was used as internal standard. Electrophoresis was performed at room temperature for 3 h at 35 mA/gel constant current in buffer containing 25 mmol/l Tris-HCl, 200 mmol/l glycine, 1 g/l SDS, pH 8.3. Western blotting Electrophoretic transfer of protein from gel to nitrocellulose was performed overnight at 10 V followed by 1 h at 60 V in buffer containing 200 mmol/l glycine, 25 mmol/l Tris-HCl, 20% methanol (v/v), pH 8.3. After. washing for 1 h with phosphate buffered saline (PBS: 154 mmol/l NaCl, 7 mmol/l KH,P04, 20 mmolil Na2HP04, pH 7.4 containing 5 g/l Tween 20) blots were incubated for 2 h with either polyclonal sheep antihuman Lp(a) antibody (Immuno) or the polyclonal goat antihuman ape(a) antibody (Biopool) employed in the ELISA immunoquantitation of ape(a). A 2-h incubation with the appropriate species-specific polyclonal alkaline phosphatase-linked donkey antibodies (both from Jackson Immunoresearch Labs Inc, West Grove, PA) followed. After further washing in PBS-Tween 20, ape(a) bands were using Nitro blue tetrazolium (100 mg/l in 150 mmol/l Verona1 acetate, pH 9.6) with 500 mg/l 5-bromo-4-chloro-3-indolyl phosphate in 0.5% (v/v) dimethyl formamide and 4 mmol/l MgC&. Mobilities were determined by measurement from the origin to the mid-point of a band taking 4 points across its width and expressed relative to that of the designated B mobility bands (R/ vs. apo B). For apparent molecular mass (n/i,) measurements gels were divided with one half stained with Coomassie Blue to demonstrate molecular weight markers while the other half was stained to visualise ape(a) isoforms. Neuraminidase treatment Equal volumes of serum and neuraminidase solution (5 kU/l in 0.1 mmolll sodium acetate, pH 4.5) were incubated for 2 h at 37°C. The treated samples were then prepared and loaded onto gels as described.

218

The sensitivity of the method, defined by the amount of protein detected, was 0.02 pg Lp(a) or a serum Lp(a) concentration of 7 mg/l. All patients studied in the present report showed at least one band. Faint bands visible at a position corresponding to apo B-100 were not uncommon, particularly at high serum or antibody loads and occasionally an additional third band was seen, in keeping with previous phenotyping techniques [12,33]. Band resolution was high and thickness rarely exceeded 2 mm (usually s 1 mm). In agreement with other authors we have observed single bands in 62?! of subjects and double band phenotypes in 38%. Mean mobility of apo B in the system was 2.5 cm, corresponding to an i$ of 0.23, with good reproducibility (within-gel coefficient of variation (CV) 0.9-4.9%; between-gel CV 8.9%; n = 14). To establish the reproducibility of the mobility of the different isoforms, samples were run on 4 consecutive gels and the results are presented in Table I. Figure 1 illustrates the distribution of Rf vs. B of 590 isoform bands from TABLE I Reproducibility of phenotyping method Sample

+vs.

apo B

cv

A~j~~nt

of isoform class

By duplicate

By quadruplicate

A

0.73

0.9%

0.7

0.7

B

0.80

I .5%

0.8

0.8

C

0.79 0.41

2.6% 6.4%

0.8 0.4

0.8 0.4

D

0.82

3.3%

0.8

0.8

E

0.93 0.51

1.4% 1.6%

0.9 0.5

0.9 0.5

F

0.81

4.0%

0.8

0.8

G

0.79

6.1%

0.8

0.8

0.43

2.3%

0.4

0.4

H

0.70 0.80

09% 2.2%

0.7 0.8

0.7 0.8

I

0.47

2.0%

0.5

0.5

J

0.79

3.7%

0.8

0.8

K

0.51

6.4%

0.5

0.5

219

.2

.4

.6

Mobility

.8

relative

1

1.2

1.4

1.6

to apo B

Fig. I. Relative mobilities vs. apo B of 590 Lp(a) phenotype bands from a population of patients with coronary artery disease.

a population of patients having severe coronary artery disease and indicating an almost continuous distribution within which are several high frequency mobilities. These are in agreement with the visual appearances of the Western blots. Using these data the distribution of R,vs.B has been divided into 10 classes <0.35, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1 and > 1.15 (cut point ranges ~0.35, 0.35-0.45, 0.45-0.55, 0.55-0.65, 0.65-0.75, 0.75-0.85, 0.85-0.95, 0.95-1.05, 1.05-1.15, > 1.15). Two mobilities were faster than apo B, one showed the same mobility and 7 were slower. The distance between mid points of adjacent isoform mobilities was 2.5 mm in this system. The molecular weight for apo B was 501 kDa, in good agreement with previous data [ 12,131. Reproducibility of classification was determined using duplicate measurements of 100 randomly selected samples from a large population of patients with coronary artery disease. Figure 2 shows the distribution of relative mobility differences (Rf gel 1 - Rr gel 2) for this group. Differences are normally distributed, with a mean of +O.Ol and 95% confidence interval -0.15 to +O.17. The implications of this for classification of duplicates are summarised in Table II. If the mean of the duplicate values Rf vs.B was used, 134 of 144 (93%) of bands were allocated to an isoform class. The remaining 10 could not be classified because the mean fell on a cut point between isoform class (n = 8) or because duplicates did not fall in the same or adjacent Rf vs.B classes (n = 2). From Table I there was no change in allocation of

220

E

-0.3 r 0.2

1 0.4

I 0.6

Mean

I 0.8

I 1.0

I 1.2

1 1.4

Rf vs B

Fig. 2. Difference in relative mobility of duplicate samples is plotted against mean relative mobility for the duplicate to demonstrate reproducibility of the method. Mean and 95% confidence intervals are showu for difference between duplicates.

TABLE II Comparison of isoform mobility class allocated to single and double band phenotype samples analysed on duplicate Western blots Comparison of mobility class for duplicate samples IdenticaI R, vs. B class for both blots Discrepancy of 1 R, vs. B class division between duplicate blots Discrepancy of 2 Rf vs. B class divisions between duplicate blots Both duplicates run at mobility of cut point Non-visualisation of 1 band in duplicates Total number of bands

Single band phenotypes

Double band phenotypes

46 (79%)

58 (66%)

11 (19%)

26 (30%)

1 (2%)

1 (1%)

0

1 (1%)

0 57

2 (2%) 87

221

isoforrn class by taking duplicates as opposed to quadruplicates; therefore duplicate analysis should suffice for the majority of subjects and if values cannot be assigned following this a further duplicate analysis should allow reliable allocation. This represents a repeat assay rate of approximately 7%. Ape(a) contains 25-40% carbohydrate [7,18] and heterogeneity of this component between isoforms has been described [S]. This could cause overestimation of the apparent molecular weight of the protein 1191.The contribution of siaiic acid residues to ape(a) molecular size was investigated by treatment with neuraminidase [12,13]. Treatment with ne~~~d~e resulted in changes of Rf and the class of the allocated isoform (Table III). The binding affinity of the ELISA antibody for different isoforms of ape(a) was investigated by comparing observed and expected concentrations for a varying amount of one isoform added to a constant concentration of another pure isoform of different Rf class. The Rf vs. B for the isoforms chosen were 1.0, 0.8 and 0.5. Observed versus expected concentrations were plotted with a minimum of 5 points in quadruplicate for each isoform combination. The correlation coefficients for the 6 combinations were 0.992-0.996, slopes were 0.88-0.97. Discussion A sensitive, high resolution method for identif~ng ape(a) isoforms is described. This method relies upon vertical SDS-PAGE under reducing conditions and subsequent Western blotting using a sensitive double antibody technique with an alkaline phosphatase-linked colour development. It is a simple, quick method which does not rely on the preparation of gradient gels or the use of radiolabelled antibody preparations [12,13]. Commercially available antibodies have been used and the method is appropriate to many laboratories, as standard equipment and readily available materials are employed. The antibodies have previously been characterised and are known to bind all isoforms of ape(a) described [17] (Fig. 3). In a recent publication [20] commercial antibodies were also employed to demonstrate phenotypes of

TABLE III Effect of neuraminidase treatment on Rr of apofa) isoforms and phenotype allocation Sample

A B C D E F

Rfvs.

Isoform allocation

apo B

-Neuraminidase

+Neuraminidase

-Neuraminidase

+Neuraminidase

0.73 1.0 0.92 O.% 0.78 0.67 0.90

0.8 1.13 1.0 I .06 0.86 0.76 1.02

0.7 1.0 0.9 1.0 0.8 0.7 0.9

0.8 1.1 1.0 1.1 0.9 0.8 1.0

vs

B

0.5

0.4

1

1.0

2

0.7

0.5

3

0e7

4

0.8

5

l*O

6

O-8

7

0*9

6

1.1

0*7

8

jl*15

0.9

10

J-0

11 0

Fig. 3. Western blots of ape(a) isoforms on SDS-PAGE visual&d using: (A) polyclonal sheep antihuman ape(a) antibody (Immuno); (B) polyclonal goat antihuman ape(a) antibody (Biopool). Apo B mobility isoform samples were run in lanes 2, 6 and I I as mobility markers. 0 represents the origin. All classes of isoform except ~rO.35 and 0.6 are shown. Sampies were loaded as described in the text and were adjusted for the concentration of ape(a) present in serum.

Rf

B

Lane

._

223

ape(a). In that study of a Caucasian population no phenotype band (null allele) could be detected in 12%, although all subjects appear to have ape(a) coding DNA [15]. All patients in the present report gave at least one band on Western blotting. Though in part this is likely to reflect higher concentrations of Lp(a) in a population with coronary artery disease it is also due to the greater sensitivity of alkaline phosphatase compared to horseradish peroxidase as antibody-linked signal enzyme. A unifying classification of ape(a) phenotypes is lacking. Utermann originally described 6 isoforms, 1 having lower apparent molecular weight than apo B (F: running faster than B), 1 with equal apparent molecular weight to apo B (B) and 4 with higher apparent molecular weight than apo B (S 1-S4: running slower than B). Other authors [13] have described up to 12 mobilities with 2 of lower apparent molecular weights (isoforms 1 and 2), one of the same apparent molecular weight (isoform 3) and 8 of higher apparent molecular weight (isoforms 4- 11). Our groups can be compared with and appear compatible with, the existing classifications of isoforms. Two isoforms of greater Rf than apo B were found (Rf 1.1 and Rf > 1.15; approx. M, 403 and 457 kDa, respectively) which are comparable to F of Utermann et al. [12] and isoforms 1,2 of Gaubatz [13]. One isoform has identical mobility to apo B (Rf 1.O; approx. M, 501 kDa). Seven isoforms of Rf less than apo B were determined (Rfvs. B; 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, ~0.35; approx. M, 550 to
224

glycosylation is also unknown although large differences in apparent molecular weight are unlikely to be solely due to differences in the carbohydrate content of the ape(a). All of these possible explanations for differences between phenotypes and genotypes can be resolved by a study in which both are determined simultaneously. Estimates of molecular weight of carbohydrate-rich glycoproteins in SDS-PAGE systems must be interpreted with caution [22,23] and molecular weights assigned in this system must be viewed as approximations. However, the molecular weight of apo B (501 kDa) is in agreement with previous observations. For the study of different human populations and the effect of disease states (e.g. changing diabetic control) on ape(a) the precision of isoform classification may be more important than the ability to ascribe an accurate molecular weight to each phenotype. In conclusion we describe a simple, sensitive procedure for the classification of ape(a) isoforms. Additional techniques are needed to describe not only the isoforms present in a double band phenotype but also the mass quantities of each isoform present. Such techniques, when applied to large populations, will allow the investigation of particular isoforms to see if atherogenicity is associated with the mass of ape(a) or because of an intrinsic quality of the ape(a) isoform which characterises the circulating Lp(a). Acknowledgement

We are grateful to the Wellcome Trust and The British Heart Foundation for their generous support. M. Farrer holds a Wellcome Junior Clinical Research Fellowship. F. Game holds a MRC Training Fellowship. We acknowledge the expert technical assistance of S. Finlay and J. Mitcheson. References 1

Berg K. A new serum type system in man: the Lp system. Acta Pathol Microbial 8cand 1963;59:369-382. 2 Berg K, Dahlen G, Frick MH. Lipoprotein Lp(a) and pm-B-lipoprotein in patients with coronary heart disease. Clin Genet 1965$x230-235. 3 Durrington PN, Iohola M, Hunt L, Arrol S, Bhatnagar D. Apolipoproteins(a), Al and B and parental history in men with early onset ischaemic heart disease. Lancet 1988;i:1070-1073. 4 Murai A, Miyhara T, Fujimoto M, Matsuda M, Kameyama M. Lp(a) lipoprotein as a risk factor for coronary heart disease and cerebral infarction. Atherosclerosis 198659: 199-204. 5 Kostner GM, Avogaro P, Cazaolato G, Marth E, Bittolo-Bon G, Qunci GB. Lipoprotein Lp(a) and the risk for myocardial infarction. Atherosclerosis 1981;38:59-61. 6 Rhoads GG, Dahlen G, Berg K, Morton NE, Dannenberg AL. Lp(a) lipoprotein as a risk factor for myocardial infarction. J Am Med Assoc 1986;256:2540-2544. 7 Fless GM, ZumMallen ME, 8canu AM. Physicochemical properties of apolipoprotein[a] and lipoprotein[a-] derived from the dissociation of human plasma lipoprotein[a]. J Biol Chem 1986;261:8712-8718. 8 Gaubatz JW, Chari MV, Nava ML, Guyton JR, Morrisett JD. Isolation and characterisation of the two major apoproteins in human lipoprotein[a]. J Lipid Res 1987;28:69-79. 9 Eaton DL, Fless GM, Kohr WJ, McLean JW, Xu Q-T, Miller CG, Lawn RM, 8canu AM. Partial amino acid sequence of apolipoprotein[a] shows that it is homologous to plasminogen. Proc Nat1 Acad Sci USA 1987;84:3224-3228. 10 McLean JW, Tomlinson JE, Kuang W-J, Eaton DL, Chen EY, Fless GM, 8canu AM, Lawn RM.

22s

II

12 13

14 15

16 17

18 19 20 21 22

23

cDNA sequence of human apolipoprotein[a] is homologous to plasminogen. Nature 1987;330:132-137. Fless GM, Rolih CA, Scanu AM. Heterogeneity of human plasma lipoprotein(a). J Biol Chem 1984;259:11470-11478. Utermann G, Menzel HJ, Kraft HG, Duba HC, Kremmler HG, Seitz C. Lp(a) glycoprotein phenotypes. J Clin Invest 1987:80:458-465. Gaubatz JW, Ghanem KI, Guevara J Jr, Nava ML, Patsch W, Morrisett JD. Polymorphic forms of human apolipoproteinla]: inheritance and relationship of their molecular weights to plasma ievels of lipoprotein]a]. J Lipid Res 1990,31:603-613. Gavish D, Azrolan N, Breslow JL. Plasma Lp(a) concentration is inversely correlated with the ratio of kringle IV&tingle V encoding domains in the ape(a) gene. J Clin Invest 1989;84:2021-2027. Lackner C, Boerwinkle E, Leffert CC, Rahmig T, Hobbs HH. Molecular basis of apoli~protein (a) isoform size heterogeneity as revealed by pulse-field eiectrophoresis. J Chn Invest 1991;87:2153-2161 Kratt HG, Dieplinger H, Hoye E, Utermann G. Lp(a) phenotyping by immunoblotting with polyclonal and monoclonal antibodies. Arteriosclerosis 1988;8:212-214. Farrer M, Game FL, Finlay S, Laker MF, Alberti KGMM. Assessment of commercial monoclonal and polyclonal antibodies to human serum lipoprotein Lp(a) phenotypes. Atherosclerosis 19PO;85:94 (Abstract). King J, Laemmli UK. Tail fibres of bacteriophage T4. J Mol Biol 1971;62:465-473. Seman LJ, Breckenridge WC. Isolation and partial characterisation of apolipoprotein (a) from human lipoprotein (a). Biochem Cell Biol 1986;64:999-1009. Huang CM, Kraft HG, Gregg RE. Modified i~unoblotting technique for phenotyping lipoprotein(a). Clin Chem lPP1;37:576-578. Scanu AM, Pfaffinger D. The rhesus monkey as a model for the study of li~protein(a). In: Scanu AM, ed. Lipoprotein(a). San Diego: Academic Press, 1~175-181 Poduslo JF. Glycoprotein mol~ular-wei~t determination using sodium dodecyt sulfate-pore gradient electrophoresis: Comparison of Tris-glycine and Tris-borate-EDTA bulTer systems. Anal Biochem 1981;114:131-139. Scanu AM, Fless GM. Lipoprotein(a): Heterogeneity and biological relevance. J Clin Invest 1990;85:17OP-1715.