- Email: [email protected]
Properties of two pig low density lipoproteins prepared by zonal ultracentrifugation
Atherosclerosis,
22 (1975) 583-599
583
0 Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
PROPERTIES ZONAL
OF TWO PIG LOW DENSITY
LIPOPROTEINS
PREPARED
BY
ULTRACENTRIFUGATION
G. D. CALVERT
AND
P. J. SCOTT
Department of Medicine (P.J.S.), University land), and Department of Clinical Chemistry Birmingham B15 2TH (Great Britain)
of Auckland School of Medicine, Auckland (New Zea(G. D.C.), Queen Elizabeth Medical Centre, Edgbaston,
(Revised, received March 14th, 1975) (Accepted April 2nd, 1975)
SUMMARY
Pig plasma lipoproteins were separated into four density density, two low density and high density lipoproteins, VLDL,
classes (very low LDLl, LDLz and
HDL respectively) from 670 ml plasma by ultracentrifugation in a continuous density gradient using the Spinco Ti15 zonal rotor. LDLl and LDLz were partly characterised. LDLl and LDLz are p-migrating lipoproteins of different size and hydrated density; they are similar to human LDL2 and LDL3 respectively. Pig plasma contains about twice mean LDLz g/ml.
as much LDLl as LDL2. LDLl migrates at Sf 4.9 (modal value), and has a diameter of 217 A and a modal density of 1.035 g/ml (range 1.03-1.04 g/ml). migrates at Sf 1.8 and has a mean diameter of 195 8, and a density of 1.050 Both lipoproteins are precipitated by heparin and Mn++ or by dextran sulphate
and Ca++. The apoproteins of LDLl and LDLB are both largely insoluble in 8 M urea solution. When dissolved in 1% sodium dodecyl sulphate solution and electrophoresed on polyacrylamide gel at pH 7.0, the apoproteins of LDLl and LDL2 formed
a pattern
of multiple
bands of high molecular
weight similar to that obtained
from with The that and
the apoprotein of human LDL. Both LDLl and LDL2 share a major antigen each other and with VLDL; in this respect again they resemble human LDL. amino acid compositions of LDLl and LDL2 are very similar. We concluded the apoprotein moieties of pig plasma LDLl and LDL2 are probably identical, similar to apoprotein B in human serum. Zonal ultracentifugation has proved to be a rapid and effective method for isolating large quantities of these two lipoprotein classes for further metabolic studies.
This work was supported by grants from the New Zealand Medical Research Council, the National Heart Foundation of New Zealand, the Life Insurance Medical Fund of Australia and New Zealand and the West Midlands Regional Health Authority.
584
G. D. CALVERT,
P. J. SCOTT
This method allows rapid bulk preparation of lipoproteins, and provides of their distribution and quantity in a continuous density gradient.
Key words:
a record
Beta lipoproteins - Low-density lipoproteins - Swine - Zonal ultracentrifigation
INTRODUCTIOK
The domestic pig is a suitable model for the experimental study of atherosclerosis because of its many anatomical and physiological similarities to man, its susceptibility to atherosclerosis, size, diet, availability and costr. Pig plasma lipoproteins in many respects are similar to those in human plasma2-5. The pig has two major plasma lipoproteins with beta migration on paper electrophoresis, one similar to human low density lipoprotein (LDLl) and the other of slightly higher density (LDLs). Because their densities overlap, previously used standard flotation procedures in fixed angle rotors2 have proved inconsistent and unsatisfactory as a means for separating these two lipoproteins. We wished to study their metabolism and role in atherogenesis,
and a new approach
was required
for their bulk preparation.
ultracentrifugation has proved a satisfactory technique for isolating large quantities of these two lipoproteins for partial characterisation studies.
Zonal
and preparing and metabolic
METHODS
Pig plasma Blood was collected from three or four Large White pigs killed by exsanguination at a local abattoir. The pigs had been fasted 18-24 hr before death. Disodium EDTA was added fuged, and about gently at 4°C to precipitated fibrin
to a concentration of about 1 mg/ml, the blood promptly centri800 ml plasma separated. The plasma was then mixed with NaBr give a solution of the desired density. (NaCl was not used as it from pig plasma containing EDTA.)
Zonal ultracentrifugation A Beckman Spinco L2-65B preparative ultracentrifuge was used with a Spinco Ti15 zonal rotor (capacity 1665 ml). The rotor was loaded with a Beckman Model 141 gradient pump (using driver gear No. 9, idler gear No. 14 and driven gear No. 3, so that the total programme delivered about 1 litre). The programme cam was cut so that a density gradient (linear with regard to volume) of 700 ml was delivered. Different solution density gradients for the zonal rotor were investigated, adding solid NaCl or NaBr to adjust densities. Solution densities were checked at 20°C with either an Abbe High Performance refractometer (using tables of the densities
PIG LOW-DENSITY
with refractive Westphal contained
585
LIPOPROTEINS
indices
of aqueous
solution@,
interpolating
values graphically)
or a
specific gravity densitometer (Griffin and George, London). All solutions 0.02 % disodium EDTA, to prevent oxidative degradation of lipoproteins.
The following density gradient was used in the zonal rotor to isolate very low density lipoprotein (VLDL), LDLr and LDLz (modifications were used if high density lipoprotein (HDL) was required). Steep gradients did not adequately separate LDLl and LDL2, and very shallow gradients were not stable; that reported here was a compromise, giving optimum separation compatible with maximum stability. Three hundred ml NaCl solution at a density of 1.006 g/ml was pumped into the spinning rotor. This was followed by 700 ml of NaCl solution in a linear density gradient, from a density of 1.006 g/ml to a density of 1.095 g/ml, pumped in slowly to the periphery.
About
670 ml plasma
at a density
of 1.12 g/ml was then pumped
in also to
the periphery. The rotor was spun at 32,000 rpm at 4°C for 26-30 hr. The contents were then pumped out slowly, using dense salt solution to displace the contents centrally through an Isco model UA2 ultraviolet analyser; absorbance was recorded at 280 nm. 12-15 ml fractions were collected, and the lipoproteins concentrated and washed by flotation through at least an equal volume of NaCl or NaCl-NaBr solution of the appropriate density (NaCl-NaBr solutions were used if a salt solution density of 1.20 or 1.21 g/ml was desired). After this step, immunoelectrophoresis showed no albumin contamination. For studies requiring immunological techniques or amino acid analysis the lipoprotein was washed again by ultracentrifugal flotation through a NaCl or NaCI-NaBr solution. Analytical ultracentrifigation After washing and concentration lipoprotein solutions were dialysed background salt solution density required in at least 500 volumes of NaCl or NaBr solution at 4°C for 48 hr. All dialysis membranes had previously been in 0.02% disodium EDTA and dialysis solutions contained 0.02% disodium
to the NaClboiled EDTA
and 0.02 % sodium azide. Lipoprotein solutions at protein concentrations of O.l0.2% were then placed in a double cell rotor and spun at 52,640 rpm and 26°C in a Beckman Spinco Model E analytical ultracentrifuge fitted with a schlieren optical system. The rate of rotor acceleration was constant and up-to-speed time included one third of the rotor acceleration time (for calculation of sedimentation or flotation coefficients). Photographic plates of the schlieren patterns were then enlarged 5 times, measurements made from at least 5 frames, and calculations of sedimentation or flotation coefficients made by standard techniques7 using a simple desk-top calculator program. No adjustment was made for concentrations or JohnstonOgstong effects. As the lipoprotein concentrations were under 0.2 % and plots of log distance of the schlieren peak from the rotor centre against time were linear to inspection, correction was considered unnecessary. Preparative ultracentrifugation Preparative ultracentrifugation
was performed
at 4°C using a Spinco
type 40
586
G. D. CALVERT,
P. J. SCOTT
rotor in a Beckman Model L or L2-65B ultracentrifuge at 40,000 rpm for 20-24 hr. Lipoprotein solutions were concentrated and washed as described above. After ultracentrifugation under these conditions at a salt density of 1.063 g/ml both LDLl and LDLa were entirely in the supernatant. Lipoproteins were aspirated from the top 1 cm of the centrifuge
tubes with a fine Pasteur
pipette.
When
quantitative
re-
covery was important the inner parts of the tube caps were washed in a small amount of salt solution of the appropriate density. Electron microscopy Lipoproteins were studied by negative staining electron microscopy in an AEI electron microscope, using the method of Hamilton et aE.10. Particle size was calculated from the mean diameter of 20 lipoprotein particles at a magnification of times 120,000 using a scale graduated
in units of IO-am under suitable
magnification.
Amino acid analysis Lipoprotein samples, after washing twice by ultracentrifugation, were dialysed for 72 hr against at least 100 volumes 0.01 % EDTA solution (pH 7.0) at 4°C. Aliquots containing about 50-100 mg protein were lyophilised, then extracted with ethanoldiethyl ether (3 : 1, v/v) for 16 hr at 4°C with constant shaking. The insoluble protein was centrifuged, the solvent decanted, and the procedure repeated twice for 4 hr to remove all traces of lipid. The procedure was a minor modification of a previously published methodtl. Delipidated apoprotein samples were lyophilised and stored in sealed glass ampoules under nitrogen at -20°C until analysis. Because it seemed possible that some protein constituents may have been lost during this amino acid analysis on LDLi and LDLa was repeated on twice-washed 0.15 M NaCl solution. Results were identical to those obtained after the plex procedure. A weighed amount of sample was dissolved in 1 ml of 6 this was added a known quantity of norleucine as an internal standard.
procedure, samples in more comM HCl. To The sample
was degassed by repeated freezing and thawing in a vacuum. The sample tube was sealed, and the sample hydrolysed for 24 hr at the temperature of boiling toluene (110.6”C). After hydrolysis the sample was iyophilised and dissolved in pH 2.2 buffer. Analysis was performed on a Locarte automatic amino acid analyser. Separation was achieved on a single 25 cm column, by stepwise elution at pH 3.25, 4.25 and 6.65. The amino acids were detected photometrically after a ninhydrin reaction at 570 nm (440 nm for proline). Limits of detection were approximately 1 nmole. Electrophoresis Purified LDLr and LDLa fractions were dialysed in 0.15 A4 NaCl pH 7.4 before electrophoresis. Polyacrylamide disc gel electrophoresis was performed as described by Frings, Foster and Cohenr2. After prestaining with Sudan Black B the lipoprotein sample was resolved in a discontinuous pH system consisting of a sample gel, concentrating gel and separating gel. Paper electrophoresis was performed by the method of Lees and Hatch13, using Oil Red 0 as a lipid stain. Lipoproteins were applied with
PIG LOW-DENSITY
an automatic
587
LIPOPROTEINS
sampler.
Immunoelectrophoresis
was performed
on gelatinised
cellulose
acetate with a Beckman microzone apparatus using rabbit anti-pig serum to detect pig albumin. Delipidated apoproteins were electrophoresed in a sodium dodecyl sulphate
(SDS)-polyacrylamide gel 14,15 (5% acrylamide) at pH 7.0 after dissolving the samples in 1% SDS solution containing 1% mercaptoethanol at 100°C and, in one experiment, 8 A4 urea. Marker proteins were used to ascertain molecular weights. In another
experiment
electrophoresed 8% separation massie Blueis.
the delipidated
apoproteins
were dissolved
in a polyacrylamide gel (3 % polyacrylamide gel pH 9.4) containing 8 M urea 16. Proteins
in 8 A4 urea and
stacking gel, pH 6.74; were stained with Coo-
Lipoprotein precipitation Low density lipoproteins were precipitated by each of two methods. (a) Dextran sulphate and Ca++ (ref. l7). D ex t ran sulphate M.W. 500,000 (Pharmacia) 5 mg/ml in 0.1 M calcium acetate was used to precipitate LDL. After dissolving and reprecipitating LDL the dextran sulphate was removed with protamine sulphate 1 g/l00 ml (Boots Ltd., U.K.) as in the original paper. Other chemicals were BDH Analar grade. (b) Heparin and Mn++ (ref. Is; Method II). Mucous heparin was obtained from Pabyrn Laboratories, Greenford, Middlesex, U.K. Other chemicals were BDH Analar grade. After precipitation and resuspension, lipoprotein solutions were dialysed exhaustively against salt solutions of the appropriate density for analytical ultracentrifugation. Immunological
methods
Lipoprotein
fractions
used as antigens
were isolated
by zonal ultracentrifuga-
tion then purified by ultracentrifugal flotation twice through NaCl solution. The LDLi fraction (with a density of 1.03-1.04 g/ml) was floated up through a solution of density
1.055 g/ml, and the LDLs fraction
(with a density
of 1.05 g/ml) was floated
through a solution of density 1.070 g/ml. About 20 mg protein in 1 ml 0.15 M NaCl were emulsified with approximately 1 ml complete Freund’s adjuvant and injected intramuscularly into the buttocks of a New Zealand White rabbit. The lipoprotein solution was stored in a stoppered flask with 0.02 % sodium azide at 4°C. The injection in Freund’s adjuvant was repeated after 3 weeks and an injection of protein in saline was given subcutaneously at 4-5 weeks. Serum was harvested 1 week later. Antibody-antigen reactions were studied using the agar gel double diffusion method of Soothillis, and by immunoelectrophoresis on gelatinised cellulose acetate. Protein determination Protein concentrations were determined by the method of Lowry et alzO, with bovine serum albumin as a standard. At the low protein concentrations used (up to 30 ,ug in a cuvette) turbidity was not a problem. Phospholipid+, triglyceridesz2 and total cholesterols3 were measured using standard techniques.
588
G. D. CALVERT,P.J. SCOTT
RESULTS
The VLDL, LDLl and LDLz were separated by zonal ultracentrifugation (Fig. I). The VLDL was milky, LDLI opalescent, LDL2 translucent, and a mixture of HDL:! and HDLs a light yellow. The latter colour was shown not to be due to carotenoidss”. Purified LDLz in solution (0.15 M NaCl and 0.02 “/od&odium EDTA) showed a strong tendency to aggregate which made ultracentrifugation studies difficult.
VLDL
A w-
o,‘I’,’ 0
,,,,,,~,,,(‘,,,,,,, f0
20
30
40
56
60
70
60
90
100 110 120
FRACnoN (lo.3 mri NO.
FRACTION (9.8ml.l NO.
Naa I.20 NSICI
PIG LOW-DENSITY
LIPOPROTEINS
589
FRACTION (15~0ml.) NO.
Fig. 1. Zonal ultracentrifuge separation of plasma lipoproteins, using three different gradients. Conditions described in the text. The absorbance at 280 nm is plotted. a: the gradient routinely used for LDLr and LDLz preparation; details in text. b: a gradient separating LDLl and LDLz, satisfactory for some purposes. The NaCl gradient was 1.006g/ml to 1.18 g/ml, and plasma was 1.20 g/ml. c: a steeper gradient, in which LDLr and LDLa were not separated. The NaCl gradient was 1.006 g/ml to 1.21 g/ml, and plasma was 1.4 g/ml. It can be seen that the shallower gradient best separates LDLl and LDLz, though the fractions are more dilute. Under the conditions of the separation isopyknic equilibration was not achieved.
Fig. 2. Representative analytical ultracentrifuge schlieren patterns, showing LDLI and LDLz. Photographs were taken at 8 min intervals after a full speed of 52,000 rpm was reached at 26” C. The lower pattern shows NaCl solution, density 1.21 g/ml, and the upper flotation pattern shows this solution containing pig plasma lipoproteins of density less than 1.21 g/ml. The major faster component moving from left to right is LDLI, and the minor slower component is LDLz. A small amount of a low density lipoprotein slightly less dense than LDLl can also be seen.
590
G. D. CALVERT, P. J. SCOTT
Fig. 3. Electron micrographs of negatively stained pig plasma LDLI (a) and LDLz (b), magnification x 200,000. At higher magnification the spherical structures appeared to have a poorly defined granular microstructure.
PIG LOW-DENSITY
591
LIPOPROTEINS
TABLE 1 AMINO
ACID
ANALYSIS
OF DELIPIDATED
LIPOPROTEINS
@moles g-r, expressed as % of glutamic acid)
asp thre ser glu pro gly ala val met ileu leu tyr phe his lys arg
LDL (Fidge j)
HDL
LDLI
LDLz
85 57 67
84 57 69
91
51
56 73
20 26
53 21 31
100
100
100
100
100
34 42 50 47 12 46 99 23 41 23 64 31
31 42 41 46 9 44 98 29 39 22 67 30
28 46 48 45 9 46 95 15
22 26 54 28 9 9 61
20 26 52 24 2 12 65
15 18 5 50 31
13 17 5 47 28
34 13 57 26
HDLz ( Fidge5)
LDLl estimations are the mean of three separate estimations, two on delipidated LDLI and one on whole LDLI. LDLa estimations are the mean of two separate estimations on whole LDL2. HDL estimations are the means of two separate estimations on delipidated HDL, density 1.063-1.21 g/ml. All lipoproteins were twice washed by ultracentrifugation after isolation. The figures in the 3rd and 5th columns are taken from Fidge 5, It is apparent that the amino acid analyses of LDLI, LDL2 and Fidge’s LDL are similar, but different from the two analyses of the HDL fractions.
Analytical ultracentrifugation LDLl was shown to migrate in the range Sf 4.0-6.9 (mode Sf 4.9). LDLs (Sf 1.8) did not produce a clear deviation of the schlieren baseline at density 1.063 g/ml. Sf values were derived for LDLl and LDLs over a range of solution densities (Fig. 2). The modal density of LDLr was 1.035 g/ml and that of LDLs 1.050 g/ml. LDLa was clearly differentiated from HDLs at a salt solution density of 1.20 g/ml (at which density HDLa had a A.20 value of 2.2). No LDLl was seen when a LDLa fraction from the zonal ultracentrifuge was subjected to analytical ultracentrifugation, nor was LDLs seen when LDLr was analysed; two peaks were seen when the two fractions were recombined. The proportion of LDLl : LDLs judged by the schlieren patterns and by absorbance at 280 nm varied in different pigs, but was about 2 : 1. Analytical ultracentrifugal analyses of plasma lipoproteins from 14 pigs were made; all showed LDLl and LDLz to be present. Electron microscopy LDLl particles
were seen as spheres
of variable
size on electron
microscopy
592
G. D. CALVERT, P. J. SCOTT
PIG LOW-DENSITY
593
LIPOPROTEINS
‘150,000 8100,000
’
50,000
C Fig. 4. Electrophoresis of pig plasma LDL. a: paper electrophoresis showing LDLl (centre), and LDLz (left and right). b: polyacrylamide gel electrophoresis, showing LDLl (left), LDL2 and whole serum (right). c: SDS-polyacrylamide gel electrophoresis of LDLr apoprotein (left) and LDL2 apoprotein (right). This sample of LDL2 was washed once by ultracentrifugation after the initial isolation, and shows a probable albumin component (M.W. 70-72,000) which largely disappeared after a further ultracentrifugal wash. The molecular weights shown are calculated from marker proteins (gamma globulin, bovine serum albumin, ovalbumin, cytochrome C, with and without mercaptoethanol).
(mean diameter 217 A, range 210-225 A). LDLz particles were much more homogeneous spheres, of mean diameter 195 8, (range 193-200 A) -- see Fig. 3. Amino acid analysis Mean values of duplicate amino acid analyses of LDLi, LDLs and HDL were expressed as moles per unit weight on a scale in which glutamic acid, the most abundant amino acid, was given the arbitrary value of 100. This was done because persistent traces of inorganic material made results expressed as nmoles of amino acid per unit weight of sample unreliable, and also to allow comparison with analyses of the protein in intact lipoproteins. These values were compared with those obtained by Fidges, similarly adjusted (Table 1). It can be seen that there is close agreement between our values for LDL and HDL and those obtained by Fidge and the amino
594
G.
D. CALVERT,
P. J. SCOTT
Fig. 5. Immunodiffusion in agar. a: the upper 3 central wells contained (from left to right) washed pig LDLI, washed pig VLDL and washed pig LDLz. The lower central well contained washed pig LDLi. The small wells around the upper 3 central wells contained various dilutions of rabbit antiserum against pig LDLi, while those around the lower central well contained rabbit anti-pig albumin antiserum. b: as in 5a, except that the small wells around the upper 3 central wells contained various dilutions of rabbit antiserum against pig LDLz, and the lower central well contained washed pig LDLz. A reaction of identity (not shown) was obtained between pig VLDL, LDLl and LDLs using antiserum to pig LDLr. acid compositions we have obtained for LDLl and LDLz are very similar. These results were obtained after protein hydrolysis for 24 hr, during which there was a loss of about 20% methionine and about 10% threonine and serine. No attempt was made to correct for losses during hydrolysis.
Electrophoresis On paper electrophoresis LDLz consistently migrated in the anodic (leading) edge of the /I band, while LDLl occupied the middle and trailing edge. On polyacrylamide gel electrophoresis LDLr and LDLa both appeared as single bands, and LDLs moved a little faster towards the anode than LDLl (Fig. 4). Whereas the two
PIG LOW-DENSITY
forms
of LDL
LIPOPROTEINS
could
not
be separated
by paper
electrophoresis,
separation
was
consistent with polyacrylamide gel. Delipidated LDLi and LDLz could not be completely dissolved in 8 A4 urea. Most of the LDLi and LDLa apoprotein remained in the sample polyacrylamide gel, though some was dispersed through the stacking gel (4% polyacrylamide in 8 M urea) and in the interface between the stacking and running gel. It did not enter the running gel (8 % polyacrylamide in 8 M urea), though there was a small amount of diffuse dye staining in the running gel near the interface with the stacking gel which was not characterised further. VLDL and HDL patterns were similar to those obtained for human VLDL and HDL25. LDLi and LDLB apoproteins dissolved completely in 8 M urea solution containing 1% SDS and 1% mercapthoethanol maintained at 100°C for 2 min. The apoproteins formed a series of bands near the origin of a basic SDS-polyacrylamide gel consistent with molecular weights of more than 180,000 daltons (Fig. 4). A similar banded pattern was found with delipidated human LDL (S, O-20) in this system. This has been reported by others26 who consider the bands to represent degradation products of apoprotein B formed as a result of enzymatic action. LDLa inconsistently
G. D. CALVERT,
P. J. SCOTT
TABLE 2 ANALYSIS
OF
LDLl
AND
LDLe
(Expressed as % on a weight basis; mean values and range given) LDLl (n = 3) Protein Cholesterol Phospholipid Triglyceride
24.8 (23.4-25.6) 47.0 (46.6-47.2) 24.7 (24.4-25.6) 3.5 ( 2.4-5.0)
LDLz (n = 4) 29.5 (29.4-29.5) 43.2 (39.1-47.2) 26.5 (26.1-26.8) 5.1 ( 4.8-5.3)
also displayed a very faint diffuse band of staining suggesting a protein of molecular weight 70,000-72,000 daltons, which may have been due to albumin contamination, and which disappeared
with further
ultracentrifugal
purification.
Lipoprotein precipitation LDLl and LDLa were both precipitated from serum by dextran sulphate in the lipopresence of Ca++ and by heparin in the presence of Mn++. The precipitated proteins were redissolved, dialysed to a known solution density, and the identity of the proteins confirmed by analytical ultracentrifugation. Immunological studies Analysis of twice-washed lipoprotein preparations rotor runs on Ouchterloney plates (over a wide range
from two separate zonal of dilutions) and immuno-
electrophoresis showed that LDLI, LDLs and VLDL all shared a strong antigen that was not found in pig serum albumin nor in HDL (Fig. 5). This antigen is presumably analogous to the B antigen in human plasma lipoproteins. In addition VLDL shared a different antigen with HDL; this antibody gave only a trace reaction against LDL. Chemical composition Preliminary results
are shown in Table 2.
DISCUSSIOK
In 1966 Janado, Martin and Cooke described two low density lipoproteins in pig serum, one with a mean Sf of 5.1 and a molecular weight of 2.7 x 106 and the second with a mean Sf of 1.9 and a molecular weight of 2.0 x 106 which they called LDLl and LDLz respectively. They found the mean density of LDLl to be 1.03-1.04 g/ml and that of LDLB to be 1.05 g/ml. The proportion of lipid in LDLl was about 7% more than that in LDLa. These authors were obliged to use a rather slow and cumbersome differential rate flotation system in a fixed angle rotor to separate the two lipoproteins, as their densities overlapped. We have found their system unsatisfactory for our purposes, as we have needed relatively large quantities of pure LDLi
PIG LOW-DENSITY
597
LIPOPROTEINS
and LDLs for metabolic
studies.
isolate these large quantities ly by their relative flotation
Zonal
ultracentrifugation
allowed
us not only to
rapidly but also to identify lipoprotein fractions rates. In this respect zonal ultracentrifugation
accurateis both a
preparative and an analytical procedure. Fidge5 did not describe a double LDL pattern in his pigs. This difference from our findings and those of Janado, Martin and Cooks may be due to methodological details.
Unless differential
tion procedures
rate methods
are employed,
of flotation
lipoproteins
or isopyknic
with closely related
equilibrium physical
separa-
characteris-
tics will not be observed as separate entities. As we have shown in metabolic isolation of these two proteins permits separate identification of their
studiess’, metabolic
characteristics. In an attempt to produce a separation of the type we wished to achieve, Wilcox, Davis and Heimberg2s described a rate zonal method for isolating lipoproteins from
human
plasma.
The procedure
we describe
differs from theirs
in two major
respects; the amount of plasma processed in our method is larger and the gradient used is much shallower. A shallow gradient was found necessary to produce consistent separation of LDLi and LDLa. LDLr isolated by the zonal method seemed identical with
LDLl
isolated
by more
orthodox
preparative
ultracentrifugal
methods.
We
have shown that the most sensitive index of denaturation, the in viva half-life of labelled lipoprotein, has proved to be identical in both LDLl isolated by orthodox preparative ultracentrifugation and LDLl isolated in the zonal rotors7. We have called the Sf 1.8 lipoprotein in pig serum LDLs (rather than HDLr as in human serums+31). Our results from amino acid analysis, immunological studies and polyacrylamide electrophoresis of delipidated lipoprotein suggest that the protein component of porcine serum LDLi and LDLs (which is also found in VLDL) is the porcine equivalent of human apoprotein B. Porcine LDLs is quite unlike the Sf O-3 lipoprotein described by Puppione et aZ.32 in a seal, a cow and a human; these were spherical particles with a sub-unit structure phoresis, protein
ranging from 140-160 A in diameter. All had a-mobility on paper electroand in the human subject the protein cross-reacted with a- and p-lipoantisera.
Albers et al.33 have demonstrated three lipoproteins in the subclass Sf O-2 in human serum (HDLr, LDLa and a lipoprotein LDL-a-1, possibly identical to Lp(a)). The molecular weight of LDLa (which had the electrophoretic and immunological properties of LDL) was about 1.8 x 106. Their HDLr reacted with “anti-RThr” antiserum, (i.e. anti-ApoA-I, which has glutamine as the C-terminal amino acida4pss) but not with anti-a-l (probably anti-Lp(a)) or anti-LDL antisera. On polyacrylamide electrophoresis HDLr was shown to contain the polypeptides of HDLs, and to have similar electrophoretic mobility. Porcine LDL2, with many characteristics of porcine LDLi, is, therefore, probably homologous to human LDLa, but it is present in porcine serum in much greater quantities than the trace amounts of LDLa in human serum.
598
G. D. CALVERT, P. J. SCOTT
The presence studying
of two closely related
the effect of molecular
LDL species provides
size on the metabolism
an opportunity
for
of lipoproteins.
ACKNOWLEDGEMENTS
The authors
are grateful
for assistance
in this work from
Dr. D. N. Sharpe,
Miss A. Webb and Mrs. A. Hinchco for technical help, Dr. C. Morris for electron microscopy, Dr. J. Fox for amino acid analysis, Miss M. A. Howdle for art work, and Miss G. Dyson
and Mrs. B. Norris
for secretarial
assistance.
REFERENCES athero1 RATCLIFFE, H. L. AND LUGINBUHL, H., The domestic pig - A model for experimental sclerosis, Atherosclerosis, 13 (1971) 133. 2 JANADO, M., MARTIN, W. G. AND COOK, W. H., Separation and properties of pig-serum lipoprotein, Can. J. Eiochem., 44 (1966) 1201. 3 MILLS, G. L. AND TAYLAUR, C. E., The distribution and composition of serum lipoproteins in eighteen animals, Comp. Biochem. Physiol., 40B (1971) 489. 4 HAVEL, R. J., EDER, H. A. AND BRAGDON, J. H., The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum, J. C/in. Invest., 34 (1955) 1345. 5 FIDGE, N., The isolation and properties of pig plasma lipoproteins and partial characterisation of their apoproteins, Biochim. Biophys. Acta, 295 (1973) 258. 6 Handbook of Chemistry and Physics, 50th edition, edited by WEAST, R. C., The Chemical Rubber Co., Cleveland, Ohio, 1969. 7 DE LALLA, 0. F. AND GOFMAN, J. W., Ultracentrifugal analysis of serum lipoproteins, Methods
Biochem. Anal., 1 (1954) 459. 8 ONCLEY, J. L., WALTON, K. W. AND CORNWELL, D. G., A rapid method for the bulk isolation of fi-lipoproteins from human plasma, J. Amer. Chem. SW., 79 (1957) 4666. 9 JOHNSTON, J. P. AND OGSTON, A. G., A boundary anomaly found in the ultracentrifugal separation of mixtures, Trans. Faraday Sot., 42 (1946) 789. IO HAMILTON, R. L., HAVEL, R. J., KANE, J. P., BLAUROCK, A. E. AND SATA, T., Cholestasis: lamellar structure of the abnormal human serum lipoprotein, Science, 172 (1971) 475. 11 BROWN, W. V., LEVY, R. I. AND FREDRICKSON, D. S., Studies of the proteins in human plasma very low density lipoproteins, J. Biol. Chem., 244 (1969) 5687. 12 FRINGS, C. S., FOSTER, L. B. AND COHEN, P. S., Electrophoretic separation of serum lipoproteins in polyacrylamide gel, Clin. Chem., 17 (1971) Ill. 13 LEES, R. S. AND HATCH, F. T., Sharper separation of lipoprotein species by paper electrophoresis in albumin-containing buffer, J. Lab. Clin. Med., 61 (1963) 518. 14 WEBER, K. AND OSBORN, M., The reliability of molecular weight determinations by dodecyl sulphate-polyacrylamide gel electrophoresis, J. Biol. Chem., 244 (1969) 4406. 15 WEBER, K., PRINGLE, J. R., AND OSBORN, M., Measurement of molecular weights by electrophoresis on SDS-acrylamide gel. In: C. H. W. HIRS AND S. N. TIMA~HEFF (Eds.), Methods in Enzymology, Vol. 26, Part C, Academic Press, New York, London, 1972, Chap. 1, pp. 3-27. 16 REISFELD, R. A. AND SMALL, P. A., Electrophoretic heterogeneity of polypeptide chains of specific antibodies, Science, 152 (1966) 1253. 17 WALTON, K. W. AND SCOTT, P. J., Estimation of the low-density (beta) lipoproteins of serum in health and disease using large molecular weight dextran sulphate, J. C/in. Pathol., 17 (1964) 627. 18 BURSTEIN, M., SCHOLNICK, H. R. AND MORFIN, R., Rapid method for the isolation of lipoproteins from human serum by precipitation with polyanions, J. Lipid Res., 11 (1970) 583 19 SOOTHILL, J F., Estimation of eight serum proteins by a gel diffusion precipitin technique, J.
Lab. Clin. Med., 59 (1962) 859. 20 LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L. AND RANDALL, R. J., Protein with the Folin phenol reagent, J. Biol. Chem., 193 (1951) 263.
measurement
PIG LOW-DENSITY
LIPOPROTEINS
2 1 SIMONSEN,D. G., WERTMAN,M., WESTOVER, L. M. AND MEHL, J. W., The determination
22
23 24 25
26 27 28 29 30 31
32
33 34 35
599
of serum phosphate by the molybdivanadate method, J. Biol. Chen?., 166 (1946) 747. AutoAnalyzer method N78. In: AutoAnalyzer Manual, Technicon Instruments Corp., Tarrytown, N.Y. Based on G. KESSLERAND H. LEDERER,Fluorometric measurement of triglycerides. In: L. T. SKBGGS,JR. (Ed.), Automation in Analytical Chemistry, Technicon Symposium, 1965, p. 341. AutoAnalyzer method 24a. In: AutoAnalyzer Manual, Technicon Instruments Corp., Tarrytown, N.Y. SOBEL,A. E. AND SNOW, S. D., The estimation of serum vitamin A with activated glycerol dichlorohydrin, J. Biol. Chem., 171 (1947) 617. ALAUPOVIC,P. M., LEE, P. AND MCCONATHY,W. J., Studies on the composition and structure of plasma lipoproteins. Distribution of lipoprotein families in major density classes of normal human plasma lipoproteins, Biochim. Biophys. Acta, 260 (1972) 689. KRISHNAIAH,K. V. AND WIEGANDT,H., Protease activity in human serum low density lipoprotein preparations, Conference on Serum Lipoproteins, Graz, 1973, Abstract B3. CALVERT,G. D., SCOTT,P. J. AND SHARPE,D. N., The plasma and tissue turnover and distribution of two radio-iodine-labelled pig plasma low density lipoproteins, Atherosclerosis, 22 (1975) 601. WILCOX,H. G., DAVIS, D. C. AND HEIMBERG,M., The isolation of lipoproteins from human plasma by ultracentrifugation in zonal rotors, J. Lipid Res., 12 (1971) 160. NICHOLS, A. V., Human serum lipoproteins and their interrelationships, Advan. Biol. Med. Phys., I1 (1967) 109. SCANU, A. M., Human plasma high density lipoproteins. In: R. M. S. SMELLIE,(Ed.), Plasma Lipoproteins, Academic Press, London, 197 I. SHORE,B. AND SHORE,V., Some physical and chemical properties of the lipoproteins produced by lipolysis of human serum & 20-400 lipoproteins by post-heparin plasma, J. Atheroscler. Res., 2 (1962) 104. PUPPIONE, D. L., FORTE, G. M., NICHOLS,A. V. AND STRISOWER,E. H., Partial characterization of serum lipoproteins in the density interval 1.04-1.06 g/ml, Biochim. Biophys. Acta, 202 (1970) 392. ALBERS, J. J., CHEN, C.-H. AND ALADJEM,F., Human serum lipoproteins. Evidence for three classes of lipoproteins in Sf O-2, Biochemistry, 11 (1972) 57. HERBERT,P., LEVY, R. I. AND FREDRICKSON,D. S., Correction of COOH-terminal amino acids of human plasma very low density apolipoproteins, J. Biol. Chem., 246 (1971) 7068. KOSTNER,G. AND ALAUPOVIC,P., Studies of the composition and structure of plasma lipoproteins. C- and N-terminal amino acids of the two nonidentical polypeptides of human plasma apolipoprotein A, FEBS Lett., 15 (1971) 320.
22 (1975) 583-599
583
0 Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands
PROPERTIES ZONAL
OF TWO PIG LOW DENSITY
LIPOPROTEINS
PREPARED
BY
ULTRACENTRIFUGATION
G. D. CALVERT
AND
P. J. SCOTT
Department of Medicine (P.J.S.), University land), and Department of Clinical Chemistry Birmingham B15 2TH (Great Britain)
of Auckland School of Medicine, Auckland (New Zea(G. D.C.), Queen Elizabeth Medical Centre, Edgbaston,
(Revised, received March 14th, 1975) (Accepted April 2nd, 1975)
SUMMARY
Pig plasma lipoproteins were separated into four density density, two low density and high density lipoproteins, VLDL,
classes (very low LDLl, LDLz and
HDL respectively) from 670 ml plasma by ultracentrifugation in a continuous density gradient using the Spinco Ti15 zonal rotor. LDLl and LDLz were partly characterised. LDLl and LDLz are p-migrating lipoproteins of different size and hydrated density; they are similar to human LDL2 and LDL3 respectively. Pig plasma contains about twice mean LDLz g/ml.
as much LDLl as LDL2. LDLl migrates at Sf 4.9 (modal value), and has a diameter of 217 A and a modal density of 1.035 g/ml (range 1.03-1.04 g/ml). migrates at Sf 1.8 and has a mean diameter of 195 8, and a density of 1.050 Both lipoproteins are precipitated by heparin and Mn++ or by dextran sulphate
and Ca++. The apoproteins of LDLl and LDLB are both largely insoluble in 8 M urea solution. When dissolved in 1% sodium dodecyl sulphate solution and electrophoresed on polyacrylamide gel at pH 7.0, the apoproteins of LDLl and LDL2 formed
a pattern
of multiple
bands of high molecular
weight similar to that obtained
from with The that and
the apoprotein of human LDL. Both LDLl and LDL2 share a major antigen each other and with VLDL; in this respect again they resemble human LDL. amino acid compositions of LDLl and LDL2 are very similar. We concluded the apoprotein moieties of pig plasma LDLl and LDL2 are probably identical, similar to apoprotein B in human serum. Zonal ultracentifugation has proved to be a rapid and effective method for isolating large quantities of these two lipoprotein classes for further metabolic studies.
This work was supported by grants from the New Zealand Medical Research Council, the National Heart Foundation of New Zealand, the Life Insurance Medical Fund of Australia and New Zealand and the West Midlands Regional Health Authority.
584
G. D. CALVERT,
P. J. SCOTT
This method allows rapid bulk preparation of lipoproteins, and provides of their distribution and quantity in a continuous density gradient.
Key words:
a record
Beta lipoproteins - Low-density lipoproteins - Swine - Zonal ultracentrifigation
INTRODUCTIOK
The domestic pig is a suitable model for the experimental study of atherosclerosis because of its many anatomical and physiological similarities to man, its susceptibility to atherosclerosis, size, diet, availability and costr. Pig plasma lipoproteins in many respects are similar to those in human plasma2-5. The pig has two major plasma lipoproteins with beta migration on paper electrophoresis, one similar to human low density lipoprotein (LDLl) and the other of slightly higher density (LDLs). Because their densities overlap, previously used standard flotation procedures in fixed angle rotors2 have proved inconsistent and unsatisfactory as a means for separating these two lipoproteins. We wished to study their metabolism and role in atherogenesis,
and a new approach
was required
for their bulk preparation.
ultracentrifugation has proved a satisfactory technique for isolating large quantities of these two lipoproteins for partial characterisation studies.
Zonal
and preparing and metabolic
METHODS
Pig plasma Blood was collected from three or four Large White pigs killed by exsanguination at a local abattoir. The pigs had been fasted 18-24 hr before death. Disodium EDTA was added fuged, and about gently at 4°C to precipitated fibrin
to a concentration of about 1 mg/ml, the blood promptly centri800 ml plasma separated. The plasma was then mixed with NaBr give a solution of the desired density. (NaCl was not used as it from pig plasma containing EDTA.)
Zonal ultracentrifugation A Beckman Spinco L2-65B preparative ultracentrifuge was used with a Spinco Ti15 zonal rotor (capacity 1665 ml). The rotor was loaded with a Beckman Model 141 gradient pump (using driver gear No. 9, idler gear No. 14 and driven gear No. 3, so that the total programme delivered about 1 litre). The programme cam was cut so that a density gradient (linear with regard to volume) of 700 ml was delivered. Different solution density gradients for the zonal rotor were investigated, adding solid NaCl or NaBr to adjust densities. Solution densities were checked at 20°C with either an Abbe High Performance refractometer (using tables of the densities
PIG LOW-DENSITY
with refractive Westphal contained
585
LIPOPROTEINS
indices
of aqueous
solution@,
interpolating
values graphically)
or a
specific gravity densitometer (Griffin and George, London). All solutions 0.02 % disodium EDTA, to prevent oxidative degradation of lipoproteins.
The following density gradient was used in the zonal rotor to isolate very low density lipoprotein (VLDL), LDLr and LDLz (modifications were used if high density lipoprotein (HDL) was required). Steep gradients did not adequately separate LDLl and LDL2, and very shallow gradients were not stable; that reported here was a compromise, giving optimum separation compatible with maximum stability. Three hundred ml NaCl solution at a density of 1.006 g/ml was pumped into the spinning rotor. This was followed by 700 ml of NaCl solution in a linear density gradient, from a density of 1.006 g/ml to a density of 1.095 g/ml, pumped in slowly to the periphery.
About
670 ml plasma
at a density
of 1.12 g/ml was then pumped
in also to
the periphery. The rotor was spun at 32,000 rpm at 4°C for 26-30 hr. The contents were then pumped out slowly, using dense salt solution to displace the contents centrally through an Isco model UA2 ultraviolet analyser; absorbance was recorded at 280 nm. 12-15 ml fractions were collected, and the lipoproteins concentrated and washed by flotation through at least an equal volume of NaCl or NaCl-NaBr solution of the appropriate density (NaCl-NaBr solutions were used if a salt solution density of 1.20 or 1.21 g/ml was desired). After this step, immunoelectrophoresis showed no albumin contamination. For studies requiring immunological techniques or amino acid analysis the lipoprotein was washed again by ultracentrifugal flotation through a NaCl or NaCI-NaBr solution. Analytical ultracentrifigation After washing and concentration lipoprotein solutions were dialysed background salt solution density required in at least 500 volumes of NaCl or NaBr solution at 4°C for 48 hr. All dialysis membranes had previously been in 0.02% disodium EDTA and dialysis solutions contained 0.02% disodium
to the NaClboiled EDTA
and 0.02 % sodium azide. Lipoprotein solutions at protein concentrations of O.l0.2% were then placed in a double cell rotor and spun at 52,640 rpm and 26°C in a Beckman Spinco Model E analytical ultracentrifuge fitted with a schlieren optical system. The rate of rotor acceleration was constant and up-to-speed time included one third of the rotor acceleration time (for calculation of sedimentation or flotation coefficients). Photographic plates of the schlieren patterns were then enlarged 5 times, measurements made from at least 5 frames, and calculations of sedimentation or flotation coefficients made by standard techniques7 using a simple desk-top calculator program. No adjustment was made for concentrations or JohnstonOgstong effects. As the lipoprotein concentrations were under 0.2 % and plots of log distance of the schlieren peak from the rotor centre against time were linear to inspection, correction was considered unnecessary. Preparative ultracentrifugation Preparative ultracentrifugation
was performed
at 4°C using a Spinco
type 40
586
G. D. CALVERT,
P. J. SCOTT
rotor in a Beckman Model L or L2-65B ultracentrifuge at 40,000 rpm for 20-24 hr. Lipoprotein solutions were concentrated and washed as described above. After ultracentrifugation under these conditions at a salt density of 1.063 g/ml both LDLl and LDLa were entirely in the supernatant. Lipoproteins were aspirated from the top 1 cm of the centrifuge
tubes with a fine Pasteur
pipette.
When
quantitative
re-
covery was important the inner parts of the tube caps were washed in a small amount of salt solution of the appropriate density. Electron microscopy Lipoproteins were studied by negative staining electron microscopy in an AEI electron microscope, using the method of Hamilton et aE.10. Particle size was calculated from the mean diameter of 20 lipoprotein particles at a magnification of times 120,000 using a scale graduated
in units of IO-am under suitable
magnification.
Amino acid analysis Lipoprotein samples, after washing twice by ultracentrifugation, were dialysed for 72 hr against at least 100 volumes 0.01 % EDTA solution (pH 7.0) at 4°C. Aliquots containing about 50-100 mg protein were lyophilised, then extracted with ethanoldiethyl ether (3 : 1, v/v) for 16 hr at 4°C with constant shaking. The insoluble protein was centrifuged, the solvent decanted, and the procedure repeated twice for 4 hr to remove all traces of lipid. The procedure was a minor modification of a previously published methodtl. Delipidated apoprotein samples were lyophilised and stored in sealed glass ampoules under nitrogen at -20°C until analysis. Because it seemed possible that some protein constituents may have been lost during this amino acid analysis on LDLi and LDLa was repeated on twice-washed 0.15 M NaCl solution. Results were identical to those obtained after the plex procedure. A weighed amount of sample was dissolved in 1 ml of 6 this was added a known quantity of norleucine as an internal standard.
procedure, samples in more comM HCl. To The sample
was degassed by repeated freezing and thawing in a vacuum. The sample tube was sealed, and the sample hydrolysed for 24 hr at the temperature of boiling toluene (110.6”C). After hydrolysis the sample was iyophilised and dissolved in pH 2.2 buffer. Analysis was performed on a Locarte automatic amino acid analyser. Separation was achieved on a single 25 cm column, by stepwise elution at pH 3.25, 4.25 and 6.65. The amino acids were detected photometrically after a ninhydrin reaction at 570 nm (440 nm for proline). Limits of detection were approximately 1 nmole. Electrophoresis Purified LDLr and LDLa fractions were dialysed in 0.15 A4 NaCl pH 7.4 before electrophoresis. Polyacrylamide disc gel electrophoresis was performed as described by Frings, Foster and Cohenr2. After prestaining with Sudan Black B the lipoprotein sample was resolved in a discontinuous pH system consisting of a sample gel, concentrating gel and separating gel. Paper electrophoresis was performed by the method of Lees and Hatch13, using Oil Red 0 as a lipid stain. Lipoproteins were applied with
PIG LOW-DENSITY
an automatic
587
LIPOPROTEINS
sampler.
Immunoelectrophoresis
was performed
on gelatinised
cellulose
acetate with a Beckman microzone apparatus using rabbit anti-pig serum to detect pig albumin. Delipidated apoproteins were electrophoresed in a sodium dodecyl sulphate
(SDS)-polyacrylamide gel 14,15 (5% acrylamide) at pH 7.0 after dissolving the samples in 1% SDS solution containing 1% mercaptoethanol at 100°C and, in one experiment, 8 A4 urea. Marker proteins were used to ascertain molecular weights. In another
experiment
electrophoresed 8% separation massie Blueis.
the delipidated
apoproteins
were dissolved
in a polyacrylamide gel (3 % polyacrylamide gel pH 9.4) containing 8 M urea 16. Proteins
in 8 A4 urea and
stacking gel, pH 6.74; were stained with Coo-
Lipoprotein precipitation Low density lipoproteins were precipitated by each of two methods. (a) Dextran sulphate and Ca++ (ref. l7). D ex t ran sulphate M.W. 500,000 (Pharmacia) 5 mg/ml in 0.1 M calcium acetate was used to precipitate LDL. After dissolving and reprecipitating LDL the dextran sulphate was removed with protamine sulphate 1 g/l00 ml (Boots Ltd., U.K.) as in the original paper. Other chemicals were BDH Analar grade. (b) Heparin and Mn++ (ref. Is; Method II). Mucous heparin was obtained from Pabyrn Laboratories, Greenford, Middlesex, U.K. Other chemicals were BDH Analar grade. After precipitation and resuspension, lipoprotein solutions were dialysed exhaustively against salt solutions of the appropriate density for analytical ultracentrifugation. Immunological
methods
Lipoprotein
fractions
used as antigens
were isolated
by zonal ultracentrifuga-
tion then purified by ultracentrifugal flotation twice through NaCl solution. The LDLi fraction (with a density of 1.03-1.04 g/ml) was floated up through a solution of density
1.055 g/ml, and the LDLs fraction
(with a density
of 1.05 g/ml) was floated
through a solution of density 1.070 g/ml. About 20 mg protein in 1 ml 0.15 M NaCl were emulsified with approximately 1 ml complete Freund’s adjuvant and injected intramuscularly into the buttocks of a New Zealand White rabbit. The lipoprotein solution was stored in a stoppered flask with 0.02 % sodium azide at 4°C. The injection in Freund’s adjuvant was repeated after 3 weeks and an injection of protein in saline was given subcutaneously at 4-5 weeks. Serum was harvested 1 week later. Antibody-antigen reactions were studied using the agar gel double diffusion method of Soothillis, and by immunoelectrophoresis on gelatinised cellulose acetate. Protein determination Protein concentrations were determined by the method of Lowry et alzO, with bovine serum albumin as a standard. At the low protein concentrations used (up to 30 ,ug in a cuvette) turbidity was not a problem. Phospholipid+, triglyceridesz2 and total cholesterols3 were measured using standard techniques.
588
G. D. CALVERT,P.J. SCOTT
RESULTS
The VLDL, LDLl and LDLz were separated by zonal ultracentrifugation (Fig. I). The VLDL was milky, LDLI opalescent, LDL2 translucent, and a mixture of HDL:! and HDLs a light yellow. The latter colour was shown not to be due to carotenoidss”. Purified LDLz in solution (0.15 M NaCl and 0.02 “/od&odium EDTA) showed a strong tendency to aggregate which made ultracentrifugation studies difficult.
VLDL
A w-
o,‘I’,’ 0
,,,,,,~,,,(‘,,,,,,, f0
20
30
40
56
60
70
60
90
100 110 120
FRACnoN (lo.3 mri NO.
FRACTION (9.8ml.l NO.
Naa I.20 NSICI
PIG LOW-DENSITY
LIPOPROTEINS
589
FRACTION (15~0ml.) NO.
Fig. 1. Zonal ultracentrifuge separation of plasma lipoproteins, using three different gradients. Conditions described in the text. The absorbance at 280 nm is plotted. a: the gradient routinely used for LDLr and LDLz preparation; details in text. b: a gradient separating LDLl and LDLz, satisfactory for some purposes. The NaCl gradient was 1.006g/ml to 1.18 g/ml, and plasma was 1.20 g/ml. c: a steeper gradient, in which LDLr and LDLa were not separated. The NaCl gradient was 1.006 g/ml to 1.21 g/ml, and plasma was 1.4 g/ml. It can be seen that the shallower gradient best separates LDLl and LDLz, though the fractions are more dilute. Under the conditions of the separation isopyknic equilibration was not achieved.
Fig. 2. Representative analytical ultracentrifuge schlieren patterns, showing LDLI and LDLz. Photographs were taken at 8 min intervals after a full speed of 52,000 rpm was reached at 26” C. The lower pattern shows NaCl solution, density 1.21 g/ml, and the upper flotation pattern shows this solution containing pig plasma lipoproteins of density less than 1.21 g/ml. The major faster component moving from left to right is LDLI, and the minor slower component is LDLz. A small amount of a low density lipoprotein slightly less dense than LDLl can also be seen.
590
G. D. CALVERT, P. J. SCOTT
Fig. 3. Electron micrographs of negatively stained pig plasma LDLI (a) and LDLz (b), magnification x 200,000. At higher magnification the spherical structures appeared to have a poorly defined granular microstructure.
PIG LOW-DENSITY
591
LIPOPROTEINS
TABLE 1 AMINO
ACID
ANALYSIS
OF DELIPIDATED
LIPOPROTEINS
@moles g-r, expressed as % of glutamic acid)
asp thre ser glu pro gly ala val met ileu leu tyr phe his lys arg
LDL (Fidge j)
HDL
LDLI
LDLz
85 57 67
84 57 69
91
51
56 73
20 26
53 21 31
100
100
100
100
100
34 42 50 47 12 46 99 23 41 23 64 31
31 42 41 46 9 44 98 29 39 22 67 30
28 46 48 45 9 46 95 15
22 26 54 28 9 9 61
20 26 52 24 2 12 65
15 18 5 50 31
13 17 5 47 28
34 13 57 26
HDLz ( Fidge5)
LDLl estimations are the mean of three separate estimations, two on delipidated LDLI and one on whole LDLI. LDLa estimations are the mean of two separate estimations on whole LDL2. HDL estimations are the means of two separate estimations on delipidated HDL, density 1.063-1.21 g/ml. All lipoproteins were twice washed by ultracentrifugation after isolation. The figures in the 3rd and 5th columns are taken from Fidge 5, It is apparent that the amino acid analyses of LDLI, LDL2 and Fidge’s LDL are similar, but different from the two analyses of the HDL fractions.
Analytical ultracentrifugation LDLl was shown to migrate in the range Sf 4.0-6.9 (mode Sf 4.9). LDLs (Sf 1.8) did not produce a clear deviation of the schlieren baseline at density 1.063 g/ml. Sf values were derived for LDLl and LDLs over a range of solution densities (Fig. 2). The modal density of LDLr was 1.035 g/ml and that of LDLs 1.050 g/ml. LDLa was clearly differentiated from HDLs at a salt solution density of 1.20 g/ml (at which density HDLa had a A.20 value of 2.2). No LDLl was seen when a LDLa fraction from the zonal ultracentrifuge was subjected to analytical ultracentrifugation, nor was LDLs seen when LDLr was analysed; two peaks were seen when the two fractions were recombined. The proportion of LDLl : LDLs judged by the schlieren patterns and by absorbance at 280 nm varied in different pigs, but was about 2 : 1. Analytical ultracentrifugal analyses of plasma lipoproteins from 14 pigs were made; all showed LDLl and LDLz to be present. Electron microscopy LDLl particles
were seen as spheres
of variable
size on electron
microscopy
592
G. D. CALVERT, P. J. SCOTT
PIG LOW-DENSITY
593
LIPOPROTEINS
‘150,000 8100,000
’
50,000
C Fig. 4. Electrophoresis of pig plasma LDL. a: paper electrophoresis showing LDLl (centre), and LDLz (left and right). b: polyacrylamide gel electrophoresis, showing LDLl (left), LDL2 and whole serum (right). c: SDS-polyacrylamide gel electrophoresis of LDLr apoprotein (left) and LDL2 apoprotein (right). This sample of LDL2 was washed once by ultracentrifugation after the initial isolation, and shows a probable albumin component (M.W. 70-72,000) which largely disappeared after a further ultracentrifugal wash. The molecular weights shown are calculated from marker proteins (gamma globulin, bovine serum albumin, ovalbumin, cytochrome C, with and without mercaptoethanol).
(mean diameter 217 A, range 210-225 A). LDLz particles were much more homogeneous spheres, of mean diameter 195 8, (range 193-200 A) -- see Fig. 3. Amino acid analysis Mean values of duplicate amino acid analyses of LDLi, LDLs and HDL were expressed as moles per unit weight on a scale in which glutamic acid, the most abundant amino acid, was given the arbitrary value of 100. This was done because persistent traces of inorganic material made results expressed as nmoles of amino acid per unit weight of sample unreliable, and also to allow comparison with analyses of the protein in intact lipoproteins. These values were compared with those obtained by Fidges, similarly adjusted (Table 1). It can be seen that there is close agreement between our values for LDL and HDL and those obtained by Fidge and the amino
594
G.
D. CALVERT,
P. J. SCOTT
Fig. 5. Immunodiffusion in agar. a: the upper 3 central wells contained (from left to right) washed pig LDLI, washed pig VLDL and washed pig LDLz. The lower central well contained washed pig LDLi. The small wells around the upper 3 central wells contained various dilutions of rabbit antiserum against pig LDLi, while those around the lower central well contained rabbit anti-pig albumin antiserum. b: as in 5a, except that the small wells around the upper 3 central wells contained various dilutions of rabbit antiserum against pig LDLz, and the lower central well contained washed pig LDLz. A reaction of identity (not shown) was obtained between pig VLDL, LDLl and LDLs using antiserum to pig LDLr. acid compositions we have obtained for LDLl and LDLz are very similar. These results were obtained after protein hydrolysis for 24 hr, during which there was a loss of about 20% methionine and about 10% threonine and serine. No attempt was made to correct for losses during hydrolysis.
Electrophoresis On paper electrophoresis LDLz consistently migrated in the anodic (leading) edge of the /I band, while LDLl occupied the middle and trailing edge. On polyacrylamide gel electrophoresis LDLr and LDLa both appeared as single bands, and LDLs moved a little faster towards the anode than LDLl (Fig. 4). Whereas the two
PIG LOW-DENSITY
forms
of LDL
LIPOPROTEINS
could
not
be separated
by paper
electrophoresis,
separation
was
consistent with polyacrylamide gel. Delipidated LDLi and LDLz could not be completely dissolved in 8 A4 urea. Most of the LDLi and LDLa apoprotein remained in the sample polyacrylamide gel, though some was dispersed through the stacking gel (4% polyacrylamide in 8 M urea) and in the interface between the stacking and running gel. It did not enter the running gel (8 % polyacrylamide in 8 M urea), though there was a small amount of diffuse dye staining in the running gel near the interface with the stacking gel which was not characterised further. VLDL and HDL patterns were similar to those obtained for human VLDL and HDL25. LDLi and LDLB apoproteins dissolved completely in 8 M urea solution containing 1% SDS and 1% mercapthoethanol maintained at 100°C for 2 min. The apoproteins formed a series of bands near the origin of a basic SDS-polyacrylamide gel consistent with molecular weights of more than 180,000 daltons (Fig. 4). A similar banded pattern was found with delipidated human LDL (S, O-20) in this system. This has been reported by others26 who consider the bands to represent degradation products of apoprotein B formed as a result of enzymatic action. LDLa inconsistently
G. D. CALVERT,
P. J. SCOTT
TABLE 2 ANALYSIS
OF
LDLl
AND
LDLe
(Expressed as % on a weight basis; mean values and range given) LDLl (n = 3) Protein Cholesterol Phospholipid Triglyceride
24.8 (23.4-25.6) 47.0 (46.6-47.2) 24.7 (24.4-25.6) 3.5 ( 2.4-5.0)
LDLz (n = 4) 29.5 (29.4-29.5) 43.2 (39.1-47.2) 26.5 (26.1-26.8) 5.1 ( 4.8-5.3)
also displayed a very faint diffuse band of staining suggesting a protein of molecular weight 70,000-72,000 daltons, which may have been due to albumin contamination, and which disappeared
with further
ultracentrifugal
purification.
Lipoprotein precipitation LDLl and LDLa were both precipitated from serum by dextran sulphate in the lipopresence of Ca++ and by heparin in the presence of Mn++. The precipitated proteins were redissolved, dialysed to a known solution density, and the identity of the proteins confirmed by analytical ultracentrifugation. Immunological studies Analysis of twice-washed lipoprotein preparations rotor runs on Ouchterloney plates (over a wide range
from two separate zonal of dilutions) and immuno-
electrophoresis showed that LDLI, LDLs and VLDL all shared a strong antigen that was not found in pig serum albumin nor in HDL (Fig. 5). This antigen is presumably analogous to the B antigen in human plasma lipoproteins. In addition VLDL shared a different antigen with HDL; this antibody gave only a trace reaction against LDL. Chemical composition Preliminary results
are shown in Table 2.
DISCUSSIOK
In 1966 Janado, Martin and Cooke described two low density lipoproteins in pig serum, one with a mean Sf of 5.1 and a molecular weight of 2.7 x 106 and the second with a mean Sf of 1.9 and a molecular weight of 2.0 x 106 which they called LDLl and LDLz respectively. They found the mean density of LDLl to be 1.03-1.04 g/ml and that of LDLB to be 1.05 g/ml. The proportion of lipid in LDLl was about 7% more than that in LDLa. These authors were obliged to use a rather slow and cumbersome differential rate flotation system in a fixed angle rotor to separate the two lipoproteins, as their densities overlapped. We have found their system unsatisfactory for our purposes, as we have needed relatively large quantities of pure LDLi
PIG LOW-DENSITY
597
LIPOPROTEINS
and LDLs for metabolic
studies.
isolate these large quantities ly by their relative flotation
Zonal
ultracentrifugation
allowed
us not only to
rapidly but also to identify lipoprotein fractions rates. In this respect zonal ultracentrifugation
accurateis both a
preparative and an analytical procedure. Fidge5 did not describe a double LDL pattern in his pigs. This difference from our findings and those of Janado, Martin and Cooks may be due to methodological details.
Unless differential
tion procedures
rate methods
are employed,
of flotation
lipoproteins
or isopyknic
with closely related
equilibrium physical
separa-
characteris-
tics will not be observed as separate entities. As we have shown in metabolic isolation of these two proteins permits separate identification of their
studiess’, metabolic
characteristics. In an attempt to produce a separation of the type we wished to achieve, Wilcox, Davis and Heimberg2s described a rate zonal method for isolating lipoproteins from
human
plasma.
The procedure
we describe
differs from theirs
in two major
respects; the amount of plasma processed in our method is larger and the gradient used is much shallower. A shallow gradient was found necessary to produce consistent separation of LDLi and LDLa. LDLr isolated by the zonal method seemed identical with
LDLl
isolated
by more
orthodox
preparative
ultracentrifugal
methods.
We
have shown that the most sensitive index of denaturation, the in viva half-life of labelled lipoprotein, has proved to be identical in both LDLl isolated by orthodox preparative ultracentrifugation and LDLl isolated in the zonal rotors7. We have called the Sf 1.8 lipoprotein in pig serum LDLs (rather than HDLr as in human serums+31). Our results from amino acid analysis, immunological studies and polyacrylamide electrophoresis of delipidated lipoprotein suggest that the protein component of porcine serum LDLi and LDLs (which is also found in VLDL) is the porcine equivalent of human apoprotein B. Porcine LDLs is quite unlike the Sf O-3 lipoprotein described by Puppione et aZ.32 in a seal, a cow and a human; these were spherical particles with a sub-unit structure phoresis, protein
ranging from 140-160 A in diameter. All had a-mobility on paper electroand in the human subject the protein cross-reacted with a- and p-lipoantisera.
Albers et al.33 have demonstrated three lipoproteins in the subclass Sf O-2 in human serum (HDLr, LDLa and a lipoprotein LDL-a-1, possibly identical to Lp(a)). The molecular weight of LDLa (which had the electrophoretic and immunological properties of LDL) was about 1.8 x 106. Their HDLr reacted with “anti-RThr” antiserum, (i.e. anti-ApoA-I, which has glutamine as the C-terminal amino acida4pss) but not with anti-a-l (probably anti-Lp(a)) or anti-LDL antisera. On polyacrylamide electrophoresis HDLr was shown to contain the polypeptides of HDLs, and to have similar electrophoretic mobility. Porcine LDL2, with many characteristics of porcine LDLi, is, therefore, probably homologous to human LDLa, but it is present in porcine serum in much greater quantities than the trace amounts of LDLa in human serum.
598
G. D. CALVERT, P. J. SCOTT
The presence studying
of two closely related
the effect of molecular
LDL species provides
size on the metabolism
an opportunity
for
of lipoproteins.
ACKNOWLEDGEMENTS
The authors
are grateful
for assistance
in this work from
Dr. D. N. Sharpe,
Miss A. Webb and Mrs. A. Hinchco for technical help, Dr. C. Morris for electron microscopy, Dr. J. Fox for amino acid analysis, Miss M. A. Howdle for art work, and Miss G. Dyson
and Mrs. B. Norris
for secretarial
assistance.
REFERENCES athero1 RATCLIFFE, H. L. AND LUGINBUHL, H., The domestic pig - A model for experimental sclerosis, Atherosclerosis, 13 (1971) 133. 2 JANADO, M., MARTIN, W. G. AND COOK, W. H., Separation and properties of pig-serum lipoprotein, Can. J. Eiochem., 44 (1966) 1201. 3 MILLS, G. L. AND TAYLAUR, C. E., The distribution and composition of serum lipoproteins in eighteen animals, Comp. Biochem. Physiol., 40B (1971) 489. 4 HAVEL, R. J., EDER, H. A. AND BRAGDON, J. H., The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum, J. C/in. Invest., 34 (1955) 1345. 5 FIDGE, N., The isolation and properties of pig plasma lipoproteins and partial characterisation of their apoproteins, Biochim. Biophys. Acta, 295 (1973) 258. 6 Handbook of Chemistry and Physics, 50th edition, edited by WEAST, R. C., The Chemical Rubber Co., Cleveland, Ohio, 1969. 7 DE LALLA, 0. F. AND GOFMAN, J. W., Ultracentrifugal analysis of serum lipoproteins, Methods
Biochem. Anal., 1 (1954) 459. 8 ONCLEY, J. L., WALTON, K. W. AND CORNWELL, D. G., A rapid method for the bulk isolation of fi-lipoproteins from human plasma, J. Amer. Chem. SW., 79 (1957) 4666. 9 JOHNSTON, J. P. AND OGSTON, A. G., A boundary anomaly found in the ultracentrifugal separation of mixtures, Trans. Faraday Sot., 42 (1946) 789. IO HAMILTON, R. L., HAVEL, R. J., KANE, J. P., BLAUROCK, A. E. AND SATA, T., Cholestasis: lamellar structure of the abnormal human serum lipoprotein, Science, 172 (1971) 475. 11 BROWN, W. V., LEVY, R. I. AND FREDRICKSON, D. S., Studies of the proteins in human plasma very low density lipoproteins, J. Biol. Chem., 244 (1969) 5687. 12 FRINGS, C. S., FOSTER, L. B. AND COHEN, P. S., Electrophoretic separation of serum lipoproteins in polyacrylamide gel, Clin. Chem., 17 (1971) Ill. 13 LEES, R. S. AND HATCH, F. T., Sharper separation of lipoprotein species by paper electrophoresis in albumin-containing buffer, J. Lab. Clin. Med., 61 (1963) 518. 14 WEBER, K. AND OSBORN, M., The reliability of molecular weight determinations by dodecyl sulphate-polyacrylamide gel electrophoresis, J. Biol. Chem., 244 (1969) 4406. 15 WEBER, K., PRINGLE, J. R., AND OSBORN, M., Measurement of molecular weights by electrophoresis on SDS-acrylamide gel. In: C. H. W. HIRS AND S. N. TIMA~HEFF (Eds.), Methods in Enzymology, Vol. 26, Part C, Academic Press, New York, London, 1972, Chap. 1, pp. 3-27. 16 REISFELD, R. A. AND SMALL, P. A., Electrophoretic heterogeneity of polypeptide chains of specific antibodies, Science, 152 (1966) 1253. 17 WALTON, K. W. AND SCOTT, P. J., Estimation of the low-density (beta) lipoproteins of serum in health and disease using large molecular weight dextran sulphate, J. C/in. Pathol., 17 (1964) 627. 18 BURSTEIN, M., SCHOLNICK, H. R. AND MORFIN, R., Rapid method for the isolation of lipoproteins from human serum by precipitation with polyanions, J. Lipid Res., 11 (1970) 583 19 SOOTHILL, J F., Estimation of eight serum proteins by a gel diffusion precipitin technique, J.
Lab. Clin. Med., 59 (1962) 859. 20 LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L. AND RANDALL, R. J., Protein with the Folin phenol reagent, J. Biol. Chem., 193 (1951) 263.
measurement
PIG LOW-DENSITY
LIPOPROTEINS
2 1 SIMONSEN,D. G., WERTMAN,M., WESTOVER, L. M. AND MEHL, J. W., The determination
22
23 24 25
26 27 28 29 30 31
32
33 34 35
599
of serum phosphate by the molybdivanadate method, J. Biol. Chen?., 166 (1946) 747. AutoAnalyzer method N78. In: AutoAnalyzer Manual, Technicon Instruments Corp., Tarrytown, N.Y. Based on G. KESSLERAND H. LEDERER,Fluorometric measurement of triglycerides. In: L. T. SKBGGS,JR. (Ed.), Automation in Analytical Chemistry, Technicon Symposium, 1965, p. 341. AutoAnalyzer method 24a. In: AutoAnalyzer Manual, Technicon Instruments Corp., Tarrytown, N.Y. SOBEL,A. E. AND SNOW, S. D., The estimation of serum vitamin A with activated glycerol dichlorohydrin, J. Biol. Chem., 171 (1947) 617. ALAUPOVIC,P. M., LEE, P. AND MCCONATHY,W. J., Studies on the composition and structure of plasma lipoproteins. Distribution of lipoprotein families in major density classes of normal human plasma lipoproteins, Biochim. Biophys. Acta, 260 (1972) 689. KRISHNAIAH,K. V. AND WIEGANDT,H., Protease activity in human serum low density lipoprotein preparations, Conference on Serum Lipoproteins, Graz, 1973, Abstract B3. CALVERT,G. D., SCOTT,P. J. AND SHARPE,D. N., The plasma and tissue turnover and distribution of two radio-iodine-labelled pig plasma low density lipoproteins, Atherosclerosis, 22 (1975) 601. WILCOX,H. G., DAVIS, D. C. AND HEIMBERG,M., The isolation of lipoproteins from human plasma by ultracentrifugation in zonal rotors, J. Lipid Res., 12 (1971) 160. NICHOLS, A. V., Human serum lipoproteins and their interrelationships, Advan. Biol. Med. Phys., I1 (1967) 109. SCANU, A. M., Human plasma high density lipoproteins. In: R. M. S. SMELLIE,(Ed.), Plasma Lipoproteins, Academic Press, London, 197 I. SHORE,B. AND SHORE,V., Some physical and chemical properties of the lipoproteins produced by lipolysis of human serum & 20-400 lipoproteins by post-heparin plasma, J. Atheroscler. Res., 2 (1962) 104. PUPPIONE, D. L., FORTE, G. M., NICHOLS,A. V. AND STRISOWER,E. H., Partial characterization of serum lipoproteins in the density interval 1.04-1.06 g/ml, Biochim. Biophys. Acta, 202 (1970) 392. ALBERS, J. J., CHEN, C.-H. AND ALADJEM,F., Human serum lipoproteins. Evidence for three classes of lipoproteins in Sf O-2, Biochemistry, 11 (1972) 57. HERBERT,P., LEVY, R. I. AND FREDRICKSON,D. S., Correction of COOH-terminal amino acids of human plasma very low density apolipoproteins, J. Biol. Chem., 246 (1971) 7068. KOSTNER,G. AND ALAUPOVIC,P., Studies of the composition and structure of plasma lipoproteins. C- and N-terminal amino acids of the two nonidentical polypeptides of human plasma apolipoprotein A, FEBS Lett., 15 (1971) 320.