Eur. J. Biochem. 55, 583-592 (1975)

The Distribution of J Blood-Group Activity on Lipoprotein and Protein Fractions of Bovine Serum Antonin RADAS and Otto Wolfgang THIELE Physiologisch-Chemisches Institut, Universitat Gottingen (Received February 11 / April 10, 1975)

There is a diversity of carriers of the J blood-group activity of bovine serum. The qualitative and quantitative distribution of the J activity on different carriers was studied, using various fractionation procedures. Approximately one third of J activity was found in the total lipids extracted from serum, two thirds in the lipid-free residue precipitated by lipid extraction. One third of the lipid J substance was found to be bound to the very low density lipoprotein, two thirds to the low density lipoprotein, while the high density lipoprotein was completely free of J activity. All non-lipidic J substance was present in the lipid-free protein. There was no J activity in the low molecular weight mucoproteins of serum and in the apoproteins of the lipoprotein fractions. The lipoprotein fractions were prepared by ultracentrifugation at different solvent densities. The lipoprotein fractions were characterized by chemical analyses and physical properties. The lower total cholesterol concentration of bovine serum, as compared to human serum, is reflected in a lower concentration of low density lipoprotein. The results obtained by ultracentrifugation coincide with the results obtained by precipitation of “P-lipoproteins” with dextran sulfate and calcium chloride and with results obtained by gel filtration of bovine serum. The “P-lipoprotein” fraction contains lipoproteins of very low and low density, and probably chylomicrons and a variety of other proteins, however no high density lipoprotein. In contrast to other bovine blood-group substances, the J substance is primarily dissolved in the blood-plasma (or serum) [l] and in various body fluids [2].It was demonstrated in a previous study [3] that the J substance of serum is found in the total lipid fraction and in the protein-containing residue precipitated by the lipid extraction. The J substance is absorbed from the plasma onto the red cells during a postnatal period. This “coating” of red cells can also be demonstrated in vitro by incubating J-negative bovine red cells with J-containing bovine serum [4]. Before studying the coating phenomenon in more detail, it is necessary to know more about the distribution of the J substance on its various carriers in bovine serum. This paper describes the results of such investigations. MATERIALS AND METHODS Mattirials Bovine blood was obtained from the Veterinarian Institute (Tierarztliches Institut), University of GottinEur. J Biochem. 55 (1975)

gen, and from the local slaughter house. The blood samples were classified J-negative (j”)or J-positive (J”), respectively, by testing small pilot samples of blood (containing an anticoagulent) for J activity with anti-J serum which had been previously checked in international comparison tests. Serum was obtained from freshly drawn blood by allowing it to stand at room temperature overnight and by centrifuging it twice at 1000 x g, for 10 min.

Extraction of Total Lipids Total lipids were extracted from serum with chloroform/methanol (1 : 1, v/v), then purified by passing them through a column of Sephadex G-25 fine, evaporated to dryness, weighed and stored as described earlier [5].Extraction of lipids from fractions obtained by precipitation with dextran sulfate or by ultracentrifugation was carried out in a similar way. In some cases, aliquots of total lipids were dried in a vacuum oven (35 - 40 “C, 4 h) and weighed as described earlier [5].

584

The residues precipitated by the lipid extraction procedure were dried and stored at a low temperature ( - 20 ‘C) for further tests for J activity.

Exfrurfionqf’ Mucoproteins Lyophilized bovine serum was extracted by a phenoliwater mixture at 5 ‘C and purified by precipitation with ethanol as described earlier [ 3 ] . The preparation thus obtained was taken as mucoprotein. The approximate yield of mucoprotein was reported in a preceding paper [3]. Fvac t ionu t ion I?!.Gel Fil trution Fractionation of bovine serum by gel filtration with Sephadex G-200 was carried out according to the method of H. E. Muller [6] which is based on the procedure described for human serum by Fireman et al. [7]. A column of 100 cm in length and 0.9 cm in diameter was packed with Sephadex G-200 readily swollen in a 0.9‘>,,saline solution. A 2-ml volume of serum was applied to the column and eluted with 0.9% saline. The flow rate was 0.66 ml/min. The eluates were collected in 6.5-ml volumes using a fraction collector (model Ultralac 7000). Transmissions of the eluates were measured at 280 and 254 nm in a spectrophotometer (Uvicord 11). All procedures were performed in a cold room (4 “C).Solutions in appropriate concentrations of lipoproteins, obtained by precipitation with dextran sulfate, and of mucoproteins, were also passed through the Sephadex column in a similar way. The column was calibrated by filtrating commercially available substances (No. 18706, Serva, Heidelberg, Germany) as molecular weight standards.

J Rlood-Group Activity of Bovine-Serunh Lipoproteins

Ultracen trifugution The major lipoprotein classes of bovine serum were prepared as described for pig serum by Janado et al. [13]. However, the following modifications were done: Serum (density 1.031 g/ml at room temperature) was centrifuged at 113OOOxg (Spinco Ti 60 rotor) for 32 h at a density of 1.060 g/ml. The floating lipoproteins of very low and low density were removed with a syringe, dialyzed against diluted (0.2 saline, adjusted to a density of 1.007 g/ml, and centrifuged at 143000 x g for 18 h to recover very low density lipoprotein as the upper phase, and low density lipoprotein as the lower phase. The lower phase obtained by centrifugation of the original serum was adjusted to a density of I. 170 g/ml and centrifuged at 113 000 x g for 45 h to recover high density lipoprotein. Density adjustment to 1.060 and 1.007 g/ml was done by adding solid NaCl, and to 1.170 g/ml, by adding solid NaBr. Aliquots of the three lipoprotein fractions were dialyzed against distilled water, then lyophilized and dried in a vacuum oven to a constant weight.

x)

Chemical Analyses

Procedures of chemical analyses were the same as previously described [3,5]. In addition, total and unesterified cholesterol was assayed by the method of Courchaine et al. [14]. Because of the low glycosphingolipid level in bovine serum, lipid sugar and lipid hexosamine were assayed in a fraction eluted with acetone/methanol(9: 1, viv), which was obtained through chromatographic separation of serum total lipids on a silicic acid column. Electrophoretic Methods

Pvecipitution by Polyunions Bovine serum was treated with dextran sulfate and calcium chloride as described for human serum by Kritchevsky et ul. [8]. The supernatant thus obtained was taken ah “x-lipoproteins”, and the precipitated material, after removal of the excess of calcium, as “0-lipoproteins”. The dextran sulfate was, in some cases. removed from the “b-lipoproteins” by complexing with protamine sulfate [9], to prevent any influence on the electrophoretic mobility or on the tests for J activity. The complete absence of dextran sulfate was then checked by the assay method of Walton and Ricketts [lo], which is based on the phenomenon of “metachromasia” of toluidine blue by the presence of acidic polysaccharides. Analysis of “0-lipoproteins” by precipitation with heparin and calcium chloride [ l l ] is not applicable to bovine serum. as we have reported elsewhere [12].

Electrophoretic mobility of bovine serum mucoprotein was measured in an electrophoresis apparatus (Elphor, Bender & Hobein, Munchen, Germany) using strips of cellulose acetate and acetate buffer solutions (0.04 M) of various pH values (from 4.4 up to 5.6). Isoelectric focusing of mucoprotein was carried out by the method of Zech and Zurcher [16]. Electrophoresis of lipoproteins was run on slides of agarose gel by the method of Rapp and Kahlke [17]; lipids were stained with Sudanschwarz B (Serva, Heidelberg, Germany). Electrophoresis of lipoproteins was timed by marking a reference slide using vitamine B,, and bromophenol blue plus albumin as tracking samples. Lipoproteins prepared by ultracentrifugation and by dextran sulfate precipitation were subjected to polyacrylamide gel electrophoresis [ 181 in order to test for contaminating proteins. Finally, lipoproteins obtained by dextran sulfate precipitation were bur. J . Biochem. 55 (1975)

585

A. RadaS and 0. W. Thiele Table 1. Protein and lipid levels of bovine serum. Mean f S.E.M. is given. n The range of values is given in parantheses Protein

n

Total lipid

n

g I IJU nil

my, 1 uo

6.96 k 0.41 12 (5.5.5- 10.50)

450.2 k 19.3 10 (320.4-558.8)

a

1111

Total cholesterol

n

Free cholesterol

n

ing; 100 ml

mgl100 in1

125.5 k 7.9 26 (43.0- 192.0)

20.0 k 1.7 (6.0- 33.5)

20

=

number of animals. S.E.M.

c 1/----

(S - s)?

=

n (n- 1)

n Lipid hexosamine

Lipid n phosphorus

Lipid sugar"

mg, 100 ml

mg, I00 1111

mg, 100 1111

1.47 k 0.063 3

0.30

5.78 k 0.35 5 (5.19 - 7.13)

n

0.043 3

As galactose.

subjected to immunoelectrophoresis [19], using antibovine rabbit serum as the antiserum. Staining was done with Coomassie brilliant blue R-250 and Sudanschwarz B (Serva, Heidelberg, Germany).

Table 2. Serum lipid level at different times of day Animal

Time when blood was drawn

Other Methods

Ultraviolet spectra were recorded with a Beckman spectrophotometer DB equipped with a 10-inch recorder. Whenever necessary, aqueous protein (or lipoprotein) solutions were concentrated by passing them through collodion filters (No. 13200, Sartorius Membranfilter, Gottingen, Germany) at 4 "C with saline (isotonic or other) as the exterior solution. Periodate oxidation was carried out as described by Coffin et al. [20] and Bigley et al. [21]. Excess of periodate was removed by adding sufficient glucose and subsequent dialysis. All analyses and other experiments were performed at least twice, in most cases three or four times.

Total cholesterol

Lipid phosphorus

mg/100 ml

Quantification of J Activity

J activity of J-positive serum or serum fractions was detected by immunological hemolysis-inhibition tests. Anti-J sera used for these tests had been checked in international comparison tests. Quantification of tests for J activity was achieved by determining the amount of the sample that gives rise to a 50 % hemolysis of J-positive erythrocytes under standard conditions. The detailed procedure was reported previciusly [5].

Total lipids

Lactating cow

8 a.m. 1 p.m. 6 p.m.

402.7 396.5 399.0

69.1 69.3 67.8

5.81 5.59 5.29

8 a.m. 1p . m . 6p.m.

440.7 439.7 429.5

59.4 59.4 58.8

5.41 5.31 5.73

_ _ _ _ _ ~

Heifer

in a number of cattle raised in this region. It was not recorded whether cows, lactating cows, heifers, or oxen served as the blood donors. All cattle were of the German low land breeds. Calves were not used. The results are summarized in Table 1. The great variability of the total cholesterol concentration in serum is remarkable. Variations in serum lipids due to different diets have been reported [24]. In our own studies, such long-term variations seem to be less important than short-term variations that might occur in the course of one day due to ingestion of food. Since ruminants probably digest continuously both day and night, in contrast to man, daily variations of serum lipid levels are expected to be minimal. This is actually the case, as shown in Table 2. J Activity in Serum Total Lipids

REiSULTS Protein and Lipid Level of Serum

Lipid levels of bovine sera are seemingly dependent on the genotypes and certain environmental conditions [22,23]. Therefore, we determined those parameters Eur. J. Biochem. 55 (1975)

Quantitative determination of J activity in serum total lipids, in comparison to that in original serum, confirms data of a previous paper [3]. Accordingly, we found about one third of the total serum J activity to be present in the total lipids (the average in 5 different sera is 32 & 1.2 %) and two thirds in the proteincontaining residue obtained after lipid extraction.

586

J Blood-Group Activity of Bovine-Serum Lipoproteins 106 0 M , .

10

- 20 -

8 30

2 m .ar

g 40 ._

L

.;

3

m

50

105

f

v)

5 I-

2-

60 70 80

90 100 20

t

1o4

30

v, 130

50

40

60

70

Tube number 390

260

Volume of eluate (ml) Fig. 1. Celfi'ltrution qf'horino Jpositive serum. Sephadex G-200 was used. Transmission was measured in 1-cm cuvettes at 280 nm (0-0) and at 254 nm (0----O). (A) Peak of elution of molecular weight standards: 1 = cytochrome c (mol. wt 12400), 2 = egg albumin (mol. wt 45000), 3 = bovine serum albumin (mol. wt

67000), 4 = human y-globulin (mol. wt 160000). 5 ==dextran blue (mol. wt 2 x lo6). V, = void volume. H = high molecular weight peak, M = medium molecular weight peak, L = low molecular weight peak. Hatched areas: fractions being combined and tested for J activity

Fractionation by Gel Filtration

The protein elution pattern of bovine J-positive serum following Sephadex (3-200 filtration (Fig. 1) resembles that of human serum [6,25]. The elution pattern of bovine J-negative serum is the same. According to the calibration of the column, the three peaks of the elution diagram can be attributed to high, medium, and low molecular weight classes. Eluted fractions were combined within each main peak as indicated in Fig. 1, so that no overlapping of adjacent peak material occurred. The combined fractions of both the high molecular weight peak and the medium molecular weight peak showed J activity (roughly to equal extents), while the low molecular weight peak was devoid of any J activity. Lipoprotein electrophoresis on agarose gel of bovine serum revealed two lipoprotein bands that could also be detected in the electropherogram of the high molecular weight peak. No lipoprotein bands were seen in the electropherograms of the other peaks obtained by gel filtration. Serum Mucopwteins

Yield and analytical data of mucoproteins extracted from bovine serum by phenol-water were similar to those reported earlier [3], the high percentage of the carbohydrate moiety is remarkable. The ultraviolet spectrum of an aqueous solution exhibits a band in the region about 280nm (A;:,,, = 12.0, based on the protein contents as determined by the

10

20

30

40

50

Tube number 130

260

Volume of eluate (ml) Fig. 2. Gel filtration of' bovine serum niucuiiwteins e.rtrac.ted hy phenol/wuter. Sephadex (3-200 was used. Transmission was measured in 1-cm cuvettes at 280 nm. ( s - - O ) Mucoproteins from 100 ml serum; (o----o) original serum (2 ml) for comparison. Other designations are the same as in Fig. 1

Lowry method) which is characteristic for almost all proteins. The mucoproteins are not precipitated by perchloric acid, but are precipitated by ammonium sulfate or by molybdic phosphoric acid. Moreover, it is shown by gel filtration (Fig.2) that the mucoproteins are eluted with the low molecular weight fraction. The zone electrophoresis revealed two major bands, the isoelectric points of which were shown Eur. .I.Biocheni. S5. (1975)

587

A. RadaS and 0. W. Thiele

to be about 4.3 and 4.6. Besides those major bands, a series of weakly stained bands with vague outlines were to be seen. Isoelectric focusing of the mucoproteins showed a similar result. The J activity of the mucoproteins was determined and compared with that of the original J-positive serum, the total lipids, and the protein-containing residue prepared from the same serum. We found only 0.4 2; of the original total serum J activity to be present in the mucoproteins, while the total lipids contained about one third and the residue precipitated by the lipid extraction procedure about two thirds of the original serum activity; i.e. the bulk of the nonlipidic J activity must be present in a protein other than a mucoprotein. Since mucoproteins are eluted with the low molecular weight fraction following gel filtration of serum, and since no J activity has been found in this fraction, this result might be a consequence of the low concentration of mucoproteins. However, mucoproteins isolated by the phenol-water procedure and subsequen tly eluted with the low molecular weight fraction were inactive when tested in the bovine J system. As shown in Fig. 2, however, the medium and high molecular weight fractions following gel filtration of isolated mucoproteins exhibited slight J activities, even though no protein could be detected in those fractions. This fact is probably due to the high sensitivity of the serological tests for J activity as compared to the sensitivity of measuring proteins by their absorbance at 280nm. The serum mucoproteins per se do not seem to contain any J determinant. The low J activity of mucoproteins isolated by the phenol/water procedure seems to be due to contaminating proteins of medium and high molecular weights.

Precipitation by Polyanions Addition of dextran sulfate to bovine serum in the presence of calcium ions produces a precipitate described as “P-lipoprotein”, the supernatant as “a-lipoprotein”. Gel filtration of those two fractions indicates (Fig. 3 ) that the P-lipoprotein fraction predominantly contains high molecular weight substances, while the elution pattern of the a-lipoprotein fraction does not markedly differ from the elution pattern of original serum. This result was expected because, as judged from the results of precipitation of human serum with polyanions under similar conditions, the a-lipoprotein fraction should contain predominantly high molecular weight j-lipoproteins ; the a-lipoprotein fraction, all other proteins of serum including a-lipoproteins. Results of chemical analyses of the two fractions of two different sera are summarized in Table 3. As expected, it shows that lipids, particularly cholesterol, are concentrated in the /$lipoprotein fraction while Eur. J. Biochem. 55 (1975)

M 10 -

20 -

- 30.-6

40 -

.-$ 506070-

80 90 100

!

!

10

20

30

50

40

Tube number 130 260 Volume of eluate (ml)

10 Ol I

20

I

-

I

9

oz3 0 -

I

v

.-s

40.2 50-

I0

E

$

‘

I

60-

4

70-

I 1 0

90I W

,

0

80 -

I

I

10

20

30 Tube number

40

50

130 260 Volume of eluate (ml)

Fig.3. Gel filtration o j bovine serum dipuprotein ( A ) and 8-lipoprotein j i m t i o n s ( B ) prepared bJ!use of dextran suvate and CaCl,. Sephadex (3-200 was used. Transmission was measured at 280 nm. (0-0) “a-lipoproteins” from 2 ml serum or “8-lipoproteins’‘ from 20 ml serum, respectively; (O----O)original serum (2 ml each) for comparison. Other designations are the same as in Fig. 1

the protein level of this fraction is about 1/10 of the total serum protein level. The sum of the data of the two fractions agrees adequately with the corresponding data of total serum. The absorption spectrum of the two freshly prepared lipoprotein fractions are given in Fig. 4. It can be seen that the substances responsible for the absorption of total serum in the region of 460 nm are greatly concentrated in the P-lipoprotein fraction. Lipoprotein electrophoresis of the P-lipoprotein fraction freed from dextran sulfate showed two bands visualized by lipid staining. These bands, as compared with the electropherograms of the lipoproteins

588

J Blood-Group Activity of Bovine-Serum Lipoproteins

Tablc 3 . Anu/yricu/ tlurtr of'.wrurnand of lipoproteins Prepared with the aid of dextran sulfate and CaCI,. Values are expressed as mgj100 ml serum

Samplc

Animdk n urn her

Protcin

Sum or protein and total lipid

Total lipid

Cholesterol

Lipid

~

~

total

5Ugdl"

Lipid phosphorus

5 09

free

mg/100 ml serum _____

~~

J-po\itiLc scruin

. a-Lipoproteiii "[b Lipoprotci 11"

Sum of "x"

+ "/I"

~

~

~~~~

~

1 6

6050 0 6538 5

320 4 470 3

6370 4 7008 8

116 4 110 8

27 9 16 8

1 24

1

5470 0 5970 0

100 2 147 1

5570 2 6117 1

27 6 34 5

40 60

0 91

6

1 6

568.0 585.0

220.2 315.2

788.2 900.2

87.8 78.0

21.5 10.0

0.33

1 6

6038.0 6555.0

320.4 462.3

6358.4 7017.3

115.4 112.5

25.5 16.0

1.24

4.85

-

-

-

~

2 41 -

~

~

2.44 ~

As galactosc.

A;: = 0.91 at 476nm

OJ 200

300

400

500

Wavelength (nrn)

t i g . 4. A / ~ . s o r ~ ~ r. iso/ rJ ~( ' ~ / I x of I ~ wholr horirrc~seruni f S ) , cc-lipoprotein f i a c f i o ~i r ) , a,,cll(-/ipo/,rorcinfraclion ( p ) . Absorption is calculated on thc basii of 1 protein concentration, as determined by the Lowry method (',)

obtained by ultracentrifugation, can most probably be described as lipoproteins of very low and low density or, in terms of electrophoretic characterization, as pre-/?- and P-lipoproteins, respectively. There is also a small band at the origin, probably representing chylomicrons. This result agrees with that obtained with human serum [26] where both lipoproteins of low or very low density are precipitated by dextran sulfate and CaCl,. In addition, polyacrylamide gel electrophoresis of the a- and /?-lipoprotein fractions showed that these fractions also contained proteins other than the respective lipoproteins. The a-lipoprotein fraction had been expected to contain all serum proteins except j-lipoproteins. The /?-lipoprotein fraction also turned out to be contaminated with a variety of proteins other than lipoproteins. This is consistent

with similar experiences in the preparations of human serum lipoproteins; according to Margolis [27], simple precipitation of lipoproteins with polyanions does not provide a pure preparation. On the other hand, it could be demonstrated by immunoelectrophoresis that the a-lipoprotein fraction was free of P-lipoproteins and that the P-lipoprotein fraction was free of x-lipoproteins. The fractions obtained by precipitation were tested for J activity. One difficulty arouse in the course of such tests. When testing negative controls, it was seen that dextran sulfate mimicked J activity. This may be due to contaminations of the dextran sulfate with carbohydrates containing terminal carbohydrate sequences, being similar or even identical to the terminal carbohydrate sequence of the J antigen. The influence of dextran sulfate on the tests for J activity in various fractions was studied by using J negative serum. Thus, it turned out that both a- and P-lipoprotein fractions gave false positive results when tested for J activity. If, however, the lipids had been extracted from those fractions, and purified by passing them through Sephadex G-25 columns, both the purified lipids and the lipid-free residues became free of false positive activity, all contaminants obviously having been extracted with the total lipids and subsequently retained by the Sephadex column. Alternatively, correct results of J activity could be obtained after purifying the a- and P-lipoprotein fractions with the aid of protamine sulfate as described under Methods. As indicated in Fig.5, all lipid-bound J activity was present in the lipids of the p-lipoprotein fraction, while the nonlipid J substance was found in the protein part of both fractions. b u r J . Hiochem 5.i (1975)

A. RadaS and 0. W. Thiele

589

Table 4. Composition of lipoproteins of’ bovine serum fractionated by ultrucentrijugation All values are expressed as mg/ 100 ml serum Fraction

Protein

Total lipids

Lipoprotein calculated”

Lipoprotein found

Cholesterol ~

total

Lipid phosphorus

free

mg/100 ml serum ~

J-positive serum

~~

~~

7313

393.0

-

- .~

~

-

121.9

.~

21.0

5.34

7.2

58.8

66.0

67.8

13.3

3.35

0.67

Low density lipoprotein

37.7

117.9

155.6

158.0

48.3

5.20

1.24

High density lipoprotein

130.4

166.1

296.5

302.3

39.6

6.65

2.09

Sum of lipoprotein

175.3

342.8

518.1

528.1

101.4

15.20

4.00

Very low density lipoprotein

Calculated as the sum of protein and lipid Found by weighing the dried lipoproteins.

a

Table 5 Analytical data of lipoproteins of bovine herun?fractionated bj ultracentrfugation Values calculated from those of Table 4 Lipoprotein density

Values based on the respective lipoprotein class ~

~

protein

~

total lipid

cholesterol ~

phospholipid”

-

~

total

free

204 310 134

51 33 22

‘ib ~

Very low Low High

109 242 440

-~

~

89 1 758 560

_. ~~

-

254 199 176 ~

~~

Values based on the respective lipid fraction of total serum Very low

-

Low

-

High

-

a

14.8 29.8 42.0

11.1 39.6 32.5

16.0 24.8 31.7

12.5 23.2 39.1

Calculated as %, lipid phosphorus x 25

Ultracentrifugation Ultracentrifugation of serum gave four fractions ; very low, low, and high density lipoproteins, and lipoprotein-free residue, as characterized by their flotations at different densities. Results of analyses of the three lipoprotein fractions are summarized in Tables 4 and 5. Of the three lipoproteins, it is shown that only 2.4 % of the total serum proteins are present as apoproteins. The lipoprotein fractions obtained by precipitation with dextran sulfate, however, contained lipid-free proteins in addition to the lipoproteins. This is particularly true for the a-lipoprotein fraction, as outlined in the previous paragraph. If one takes the total sum of the lipids of the three ultracentrifugal lipoprotein fractions (Table 4), there is a deficit of Eur. J. Biochem. 55 (1975)

13 % as compared with the total lipid concentration in the original serum. This is probably due to losses of material that could not be avoided when removing the various fractions from the ultracentrifuge tubes. The three lipoprotein fractions were subjected to various electrophoretic studies. Accordingly, the very low density lipoprotein fraction probably contained chylomicrons, as did also the P-lipoprotein fraction. This has been concluded from the presence of a band at the startingline in the lipoprotein electropherograms. Quantitative determination of J activity in the lipids and lipid-free residues subsequent to lipid extraction of the lipoproteins and in the lipoproteinfree residue revealed the absence of J activity in the apoproteins of lipoproteins of low, very low and high density and in the lipids of high density lipoprotein, as shown in Fig.5. The lipid-bound J activity was distributed on the very low density lipoprotein (up to one third) and on the low density lipoprotein fractions (up to two thirds). The lipoprotein-free residue actually proved to be lipid-free when checked by an additional extraction procedure. It contained 65% of the total serum J activity, which disappears completely after periodate oxidation.

DISCUSSION It can be concluded from the above results that one third of the J substance dissolved in serum is bound to a lipid and that two thirds of the J substance are present as lipid-free protein, thus confirming previous investigations [ 3 ] . This latter protein probably is a glycoprotein, since the determinant of the J lipid was shown to be the carbohydrate moiety of a glycosphingolipid [28] and because the J activity of the lipid-free protein disappeared by periodate oxidation.

590

J Blood-Group Activity of Bovine-Serum Lipoproteins

Ool0 J activity

14OlOJ activity

Lipid extract ion

Lipid extraction

Lipid extraction

rn lipoprotein

lipoprotein

0.!'lo protein

0.5'10 protein

lipoprotein 1.8'10 protein

residue

Ultracentrifugation 100°/o protein

Lipid extraction

100°lo lipid

by dextran sulfate

"p- Lipoproteins"

"a-Lipoproteins" 9lolo protein

1 "a-Lipids"

Ool0 J activity

9% protein

Lipid extraction

"a- Proteins" 35%

J activity

"a- Lipids" 29'l0 J activity

"p- Proteins" 28%

J activity

Fig. 5. Lcve1.r of prorc,iir. fotul lipid and J ucfivifyof serum,fractions. All values based on arbitrary serum levels of IOO";, each

However, it has not been excluded the possibility that the J active substance, not being extractable by the lipid extraction procedure, is a glycosphingolipid of unusual complexity similar to that of human erythrocyte membrane described recently by Gardas and KoBcielak [28a]. Furthermore, it has been shown that the mucoproteins extracted from serum by a phenol1 water procedure exhibit only very little J activity, which is probably due to contaminants, while the pure mucoproteins of low molecular weight are devoid of J activity. Previous assumptions about the mucoprotein nature of the nonlipidic J substance [3,29,30] should therefore be abandoned. One-third of the J lipid has been found in the very low density lipoprotein and two thirds in the low density lipoprotein of J-positive serum, while the apoproteins of all lipoprotein fractions are J inactive. The present investigation includes assays of various lipid fractions (Table 1) of bovine serum and thus makes a comparison possible with corresponding values of thc literature. The bovine serum lipid level is distinctly lower than that of humans and proves to be at a similar level as reported in cattle by Kirkeby [31]. O'Kelly [22- 24.32,33], however, reported lower values in certain breed types of cattle raised in Australia. The low concentration of total and free cholesterol in the serum of various mammalian species as compared to human serum is well known. It is also

confirmed in the above study. We also found the phospholipid level of bovine serum to be in the same range as reported by O'Kelly [22-24,32,33]. Apart from the ganglioside level [33a], the total glycolipid level of bovine serum, to the best of our knowledge, has never been reported. Studies on the composition and structure of serum lipoproteins have been mainly confined to humans and rats. In particular, serum lipoproteins of cattle have rarely been studied systematically. Jensen [34] found "2, possibly 3 lipoproteins" by electrophoretic procedures. Glascock et al. [35] precipitated "p-lipoproteins" from bovine serum and found electrophoretically one band migrating close to the P-globulin and another remaining at the origin with diffuse lipid staining in between. They found 42 "/, of the total lipids of the "/3-lipoproteins" consisting of sterols, 24':/, of phospholipids ; these figures are close to ours, calculated from Table 3 (animal No. 1, 40"< cholesterol. 28 % phospholipids). Gotz et al. [36] found one lipoprotein band in the cxl- and another in the cx2/l-globulin region when studying bovine whole serum by paper electrophoresis. Kirkeby [31] has performed comparative lipoprotein determinations in various mammalian species by a paper electrophoretic method and found 313 mg lipids/100 ml serum in the cx-lipoproteins and 170 mg/100 ml in the P-lipoproteins (mean values each), i.e. 6 5 % of the serum total lipids were transEur. J . B1ochem. 55 (1975)

59 1

A. RatlaS and 0. W. Thiele

ported with the a-lipoprotein fraction, 35% with the P-lipoprotein fraction, while in humans he found only 30 % (in males) or 34 % (in females) of the serum total lipids to be transported with the a-lipoproteins. Also Alexander et al. [37] found more lipids transported with the a-lipoproteins than with the P-lipoproteins by evaluating agarose gel electropherograms of bovine whole serum. Jonas [38] has studied the high density lipoprotein of bovine serum isolated by ultracentrifugation and found 80% of the total lipoproteins to belong to the high density lipoprotein fraction, while the respective percentage in human serum is merely about 40 %. The above results of characterization of the lipoprotein fractions isolated by ultracentrifugation agree with those of human serum lipoproteins on the whole. Thus, we found a protein moiety of roughly 11 % in bovine lipoprotein of very low density, 24 % in that of low density, 44 % in that of high density. However, the protein moiety of the P-lipoprotein fraction, that corresponds to the sum of lipoproteins of low and very low density, is, in two samples, as high as 72% and 65 %., respectively (calculated from data of Table 3), instead of the expected value of 35%. This deviation is due to the presence of a variety of proteins other than lipoproteins in the P-lipoprotein fraction. As in humans, the low density lipoprotein fraction is the: richest in cholesterol, the high density fraction the richest in phospholipids. The major difference between human and bovine serum lipoproteins is the lower concentration of low density lipoprotein in bovine serum. This fact is consistent with the lower concentration of total cholesterol in bovine serum. It is reflected in a higher ratio high density/total lipoproteins in bovine serum than in human serum; this has previously been observed by Kirkeby [31] and Jonas [38]. The absorption spectrum of the P-lipoprotein fraction shows an absorption pattern (Fig. 4) being very similar to that of human P-lipoproteins [39]. This type of absorption pattern is believed to be mainly due to the presence of carotenoids bound to the P-lipoproteins [39]. The absorption at the peak is higher in our results from cattle (At:,,, = 0.91) than in humans (A:?,,, approx. 0.55), as reported by Gurd [39]. Bovine serum obviously contains a larger portion of carotenoids than human serum does; this is comprehensible in a herbivore. A11 lipoproteins are probably eluted with the high molecular weight fraction in gel filtration. Nevertheless, J activity was not only found in this fraction, but a.lso in the medium fraction. This is probably due to the presence of J active glycoproteins. Such lipidfree J-active glycoproteins were also present in the /%lipoprotein fraction so that the J activity found in Eur. J. Biochem. 55 (1975)

the protein-containing residue obtained after lipid extraction does not imply the presence of J active apoproteins. As for the mucoproteins, there is a considerable confusion in their classification and definition. Most properties of the mucoproteins of bovine serum prepared by extraction with phenollwater are in agreement with those of the Ba-a2-glucoprotein of human serum described by Schmid and Burgi [40] and of the a,-mucoid described by Schultze et al. [41]. This investigation has been carried out with financial aid of Forschungsmittel des Landes Niedersachsen. Part of the laboratory equipment used has been borrowed from the Deutsche Forschungsgemeinschuft. We are indebted to Dr J. Koch, Tierarztliches Institut der Universitat, Gottingen, for the supply of bovine blood and of J antiserum. Technical assistance by Mrs E. Bodden is gratefully acknowledged.

REFERENCES 1. Stormont, C. (1949) Proc. NatlAcad. Sci.U.S.A. 124,232-237. 2. Thiele, 0. W., Froneberg, B. & Koch, J. (1975) h i m . Blood Grps Biochem. Genet. 5, 215 - 224. 3. Schroffel, J., RadaB, A,, Thiele, 0. W. & Koch, J. (1971) Eur. J . Biochem. 22,396-399. 4. Blakeslee, D. & Stone, W. H. (1971) Vox Sang. 21, 269-283. 5. Schroffel, J., Thiele, 0. W. & Koch, J. (1971) Eur. J . Biochem. 22,294- 300. 6. Miiller, H. E. (1968) Med. Klinik, 63, 125- 128. 7. Fireman, P., Vannier, W. E. & Goodman, H. C. (1963) J . Exp. Med. 117,603-619, 8. Kritchevsky, D., Tepper, S. A,, Alaupovic, P. & Furman, R. H. (1963) Proc. Soc. Exp. Biol. 112, 259-262. 9. Cornwell, D. G. & Kruger, F. A. (1963) J . Lipid Res. 2, 110- 134. 10. Walton, K . W. & Ricketts, C. R. (1954) Brit. J . Exp. Pathol. 35, 227 - 240. 11. Prellwitz, W. & Kottgen, E. (1969) Arztl. Lah. 15, 20-27. 12. Thiele, 0. W. & RadaS, A. (1975) Zentralbl. Veterinamed. Reihe A , in press. 13 Janado, M., Martin, W. G. & Cook, W. H. (1966) Can. 1. Biochem. 44,1201 - 1209. 14 Courchaine, A. J., Miller, W. H. & Stein, D. B. (1959) Clin. Chem. 5, 609-614. 15 Vance, D. E. & Sweeley, C. C. (1967) J . LipidRes. 8,621 -630. 16 Zech, R. & Ziircher, K. (1973) Life Sci. 13, 383-389. 17 Rapp, W. & Kahlke, W. (1968) Clin. Chem. Acta, 19,493-498. 18 Stegemann, H. (1970) Z . Anal. Chem. 252, 165-169. 19 Cawley, L. P. (1969) Electrophoresis and lmmunoelectrophoresis, Little, Brown & Co., Boston, Mass. 20 Coffiin, S. F. &Pickles, M. M. (1953) J. Immunol. 71,177- 182. 21 Bigley, N . Y., Chandler, R. W. & Dodd, M. T. (1958) J . Immunol. 80,85-93. 22 O’Kelly, J. C. (1968) Aust. J . Biol. Sci. 21, 1013- 1024. 23. O’Kelly, J. C. (1973) Comp. Biochem. Physiol. 44, 313-320. 24. O’Kelly, J. C. (1973) Comp. Biochem. Physiol. 44, 303-312. 25. Fireman, P., Vannier, W. E. & Goodman, H. C. (1964) Proc. Soc. Exp. Biol. 115, 845 - 849. 26. Burstein, M., Scholnick, H. R. & Morfin, R. (1970) J . Lipid Res. 11,583-595, 27. Margolis, S. (1969) in Structural and Functional Aspects of Lipoproteins in Living Systems (Tria, E. & Scanu, A. M., eds) p. 373, Academic Press, London, New York.

593

A. Radag and 0. W. Thiele: J Blood-Group Activity of Bovine-Serum Lipoproteins

28. Thiele. 0. W. & Koch. J. (1970) Eur. J . Biochem. 14, 379-386. 28a. Gardas, A . & KoScielak, J. (1974) FEBS Lett. 42, 101 104. 29. Thiele, 0. W. (1972) Bhrt, Z . Gesomte Blutfbrsch. 314-320. 30. Thiele. 0. W. & Koch. J. (1973) VO.YSong. 25, 317-326. 31. Kirkeby, K. (1966) Suind. J . Clin. Lab. Invest. 18; 437-442. 32. O’Kelly. J . C. (1968) 4u.st. J . Biol. Sci. 21, 1025-1032. 33. O’Kelly. .I C. (1972) Corzp. Biockem. Phj,siol. 43, 283-294. 33a. Yu. R . K . & Ledcen. R. W. (1972) J . LipidRes. 13, 680-686. 34. Jensen. I. K.(1963) A c f a Vei. S ~ ~ o n4, d . 64-84. 35. Glascock. R. F., Welch. V. A,, Bishop, C., Davies, T., Wright, E. W. & Noble, R. C. (1966) Biochem. J . Y8, 149-156. -

36. Gotz, H. & Heinbrodt. A. (1969) Zerztrolld. Ci~trrinomc.tl,Reihe A , 16,691 - 702. 37. Alexander, C . & Day, C. E. (1973) Comnp. Biochem. P/ry,siol. 46, 295-312. 38. Jonas, A. (1972) J . Biol. Chem. 23, 7767-7772; 7173-7778. 38a. Jonas, A. (1973) Bioc.licnzislv)~,12; 4503 -4507. 39. Gurd, F. R. N . ( 1960) i n Lipid C h m i s t q . (Hanahan, D. J., ed.). pp. 260- 325, Wiley, New York, London. 40. Schmid, K. & Biirgi, W. (1961) Biochirn. Bioplrys. .4cta. 47, 440-453. 41. Schultze, H. E., Heide, K. & Haupt. H. (1962) Naturn~i.ssenschqfien, 49, 15.

A. RadaS and 0 . W . Thiele. Physiologisch-Chemisches Institut der Georg-August-Universitit zu Gottingen, D-3400 Giittingen, Humboldtallee 7. Federal Republic of Germany

Eur. I . Biochem. 55 (1975)