On the Subunit Structure of the Protein of Human Serum High Density Lipoprotein A STUDY OF SEPHADEX FRACTION IV*

Abstract Fraction IV, separated and purified from delipidated human serum high density lipoprotein of d 1.063 to 1.125 g per ml (HDL2) by gel filtration in 8 m urea, was further studied before and after reduction with β-mercaptoethanol (β-ME), or reduction and carboxymethylation (iodoacetamide). Both the reduced (R-IV) and the reduced and S-carboxymethylated (SC-IV) preparations exhibited a single band by polyacrylamide gel electrophoresis with a molecular weight of about 8,500, a figure corroborated by Sephadex G-200-8 m urea and agarose-guanidine HCl column chromatography. In the absence of the reducing agent, Fraction IV had a molecular weight of about 17,000. Treatment of R-IV (in the presence of β-ME) or of SC-IV by the bifunctional reagent dimethylsuberimidate led to the formation of four components separable by analytical sodium dodecyl sulfate (SDS)-polyacrylamide gel, with an apparent molecular weight of approximately 8,500; 17,000; 25,000; and 33,000. In turn, unreduced and suberimidate-treated IV, exhibited two bands (SDS-polyacrylamide) with an apparent molecular weight of 17,000 and 33,000, respectively. In initial studies, where partially carbamylated preparations were used, fractionation of SC-IV by DEAE-cellulose column chromatography yielded two major and three minor fractions which all reacted with antisera raised in rabbit against whole Fraction IV, and differed mainly from each other in lysine and homocitrullin content. In subsequent studies, which used noncarbamylated products, SC-IV gave a single peak by both Sephadex and DEAE-chromatography, a major component by isoelectric focusing, and exhibited a single band by 8 m urea or SDS-polyacrylamide electrophoresis. This preparation had no histidine, arginine, or tryptophan, and had glutamine as COOH-terminal (carboxypeptidase digestion and hydrazinolysis). It contained no sialic acid, showed no NH2-terminal by either dansylation or by the Edman's procedures, and gave a single precipitin line against specific antisera. It is concluded that Fraction IV, as separated and purified from apo HDL2 by Sephadex chromatography, is made up of chemically very similar, and possibly identical protomers each having the same molecular weight (about 8,500). The data also indicate that in HDL2 these protomers are paired by a single disulfide linkage into dimers of equivalent weight.


SUMMARY
Fraction IV, separated and purified from delipidated human serum high density lipoprotein of d 1.063 to 1.125 g per ml (HDL2) by gel titration in 8 M urea, was further studied before and after reduction with /?-mercaptoethanol (P-ME), or reduction and carboxymethylation (iodoacetamide). Both the reduced (R-IV) and the reduced and S-carboxymethylated (SC-IV) preparations exhibited a single band by polyacrylamide gel electrophoresis with a molecular weight of about 8,500, a figure corroborated by Sephadex G-ZOO-8 M urea and agarose-guanidine HCl column chromatography. In the absence of the reducing agent, Fraction IV had a molecular weight of about 17,000. Treatment of R-IV (in the presence of P-ME) or of SC-IV by the bifunctional reagent dimethylsuberimidate led to the formation of four components separable by analytical sodium dodecyl sulfate (SDS)-polyacrylamide gel, with an apparent molecular weight of approximately 8,500; 17,000; 25,000; and 33,000. In turn, unreduced and suberimidate-treated IV, exhibited two bands (SDS-polyacrylamide) with an apparent molecular weight of 17,000 and 33,000, respectively. In initial studies, where partially carbamylated preparations were used, fractionation of SC-IV by DEAE-cellulose column chromatography yielded two major and three minor fractions which all reacted with antisera raised in rabbit against whole Fraction IV, and differed mainly from each other in lysine and homocitrullin content.
In subsequent studies, which used noncarbamylated products, SC-IV gave a single peak by both We here report on studies dealing with the isolation of pure Fraction IV and on its partial characterization by chemical, physical, and immunological methods.
The relevance of these results as well as those on Fraction III, presented in the accompanying paper (4), will be discussed in the context of the over-all problem of I-IDL, structure.

EXPERIMENTAL PROCEDURE
The sera used for the separation of IV were those employed in the preparation of III (4 on a Gilford (Oberlin, Ohio) recording spectrophotomcter.
Fraction IV was freed of urea, concentrated in an Amicon Diaflo Ultrafiltration Cell, model 52, using UhI2 membranes and then lyophilized.
All of these operations were carried out at 4". Identification of Fraction IV was made by its position of migration in 8 M urea-or SDS-polyacrylamide gel electrophoresis (1) and by its reactivity against anti-rabbit human Fraction IV antisera.
Whenever contamination by Fraction III was crcountered, it was resolved by rechromatography of the impure fraction in Sephadex G-200 columns (100 X 2.5 cm) at 10" with ascending flow rate of the eluting buffer (0.01 M Tris, pEI 8.6, lop3 x4 ElW.1, 8 *I urea) at 10 ml per hour.
.i\ more effective separation of IV from III was achieved by addillg, to the eluting buffrr, P-LIE (concentration of 0.1 g/100 ml); this agent has been shown to significantly retard t,he elution of Fraction IV from Sephades columns (3). Alternatively, the impure prcparations of IV were reduced and carbosymethylated (see below) before column fractionation.
The puribar of Fraction IV, R-IV, or SC-IV, after l,echromatograph7-, was tested by SIX-polyacrylamide electrophoresis and reactivity against antisera prepared in the rabbit against unreduced IV, according to the scheme described for Fraction III (4). Conditions for Preparing P'raction 1 V Protomers-In a previous report (3) WC have shown that disulfide cleavage promotes the conversion of IV into its monomeric form.
The preparations directed for amino arid analysis wcrc extensively dialyzed against distilled water.
In early studies a twostep 'hi? linear gradient of 1111 8.2 was used; the first, 100 ml tot,al volume, frown 0.04 to 0.06 1~ Tris, pH 8.2; a11d the second, 500 ml total volume, between 0.06 and 0.1 M Tris, pfl 8.2. Subsequently :I single lineal gradient, 0.01 to 0.1 M Tris, p1-l 8.2, was used. All buffers contained 1OP 11 ED'I'A and 8 34 urea and were pumlxd at a flow rate of 10 ml per hour. The effluents, collccstcd in 2-ml fractions, were continuously monitored for absorbance at 280 nm and conductivity as described for Fraction III (4). The fraction obtained was freed of urea, concentratcd, alld then used for analysis either immediately or after storage ill the lyophilized state. Analytical Methods Employed in the Characterization oJ R-IV or SC-IV Subfractions Obtained by ~E.~E-chromatograpky-r~l'l~e rnetlmds of 8 M urea-and Sl Wpolyacrylamide gel elcctrophorcsis and isoelectric focusing have been reported previously (1) as have the immunological analyses (1) that were carried out with anti-ITT antisera I)rel)arerl in the rabbit.
Conditions for nrnillo 5851 acid, NH,-and COOH-terminal and sialic acid analyses were those described for Fraction III (4).

Cross-linking Experiments Using
Unreduced IV, R-IV, or SC-IV-The studies were carried out using the agent dimethyl suberimidate as described by Davies and Stark (5). After the reaction, each fraction was separated by SDS-polyacrylamide (I), and the apparent molecular weight was determined according to Weber and Osborn (6). Cleavage of the cross links was obtained by treatment with ammonia-acetic acid, 30:2 (4).

Total Protein
Determination-This was carried out by the Lowry method (7) or by absorbance at 280 nm using the extinction coefficient BjTm = 9.2, determined in the laboratory, taking into account the amino acid composition of Fraction IV. The value was not affected by reduction and carboxymethylation. Chemicals were all reagent grade or redistilled as described for Fraction III.
The precautions in utilizing urea solutions were detailed in the accompanying paper (4)  by guest on July 10, 2020 http://www.jbc.org/ Downloaded from pH 8.6, 0.01% P-ME, 10e3 M EDTA) of the impure mixture. Under such conditions SC-IV is eluted as a single peak.
Electrophoretic Properties of Iii, R-IV, and XC-IV-In initial studies, when urea-induced carbamylation was not completely controlled, unreduced or SC-IV exhibited four to six closely migrating bands by urea-polyacrylamide gel electrophoresis (Fig. 2). The fractions were readily separated by ion exchange chromatography (Fig. 3) and each gave a single line of precipitation against anti-IV antisera and exhibited a very similar amino acid composition (Table I) Fig. 2).
purified preparation of SC-IV (uncarbamylated). sis (Fig. 1lA). Unreduced IV, treated with [14C]iodoacetamide, exhibited no changes in mobility in either of the two supporting media.
Such a preparation, after extensive dialysis was free of radioactivity indicating the absence of -SH groups.
In turn, when IV was either reduced with ,&ME or the step or reduction was followed by carboxymethylation (iodoacetamide), there was a significant increase of its electrophoretic mobility by SDSpolyacrylamide as shown previously (3) with apparent molecular weight of unreduced IV and SC-IV of 17,000 and 8,500, respectively (Fig. 5). Identical results were obtained when HDLZ, first incubated with 1% SDS, was analyzed by SDS-polyacrylamide in the presence and absence of P-ME.
DEAE-column Chromatography Studies-Under the experimental conditions followed, either IV or SC-IV exhibited a single component (Fig. 6) which again gave a single band by either 8 M urea or SDS-polyacrylamide electrophoresis.
A major component, although with a minor undefined shoulder was observed by isoelectric focusing (Fig. 7). Unreduced or SC-IV gave a IOa- single line of precipitation against rabbit anti-IV antisera (Fig   8).
Chemical Xtudies on DEAE-fraction-The amino acid composition of IV is given in Table I. Characteristically, this protein had 1 mole of carboxymethylcysteine per mole of protomer (mol wt ~8500) and had virtually no histidine, arginine, or tryptophan.
The time course of hydrolysis by carboxypeptidase ,4 suggested glutamine as the COOH-terminal with threonine as the penultimate (Fig. 9). This conclusion was supported by the results with carboxypeptidase B and by the hydrasinolysis studies which failed to detect any COOH-terminal residue.
NO NH*-terminal amino acid was detected in either IV or SC-IV by either dansylation or by the Edman's procedure, suggesting a blocked terminal.
Fraction IV contained no sialic acid as assessed by both enzymatic (neuraminidase) or acid hydrolysis. SDS-polyacrylamide electrophoresis, four bands ( Fig. 10) with approximate molecular weights of 8,500; 17,000; 26,000; and 34,000 (Fig. 11) indicating a protomer, dimer, trimer, tetramer relationship and suggesting, according to Davies and Stark (5), that Fraction IV contains four protomeric units. This conclusion was also supported by the observation that unreduced IV (dimer) treated with suberimidate gave two bands on SDS-polyacrylamide with molecular weights of 16,000 and 32,000, respectively.
Studies on Single Donors-These studies showed no significant differences in results among the individuals examined.
Only in one case COOH-terminal threonine was found and this was reported in a preliminary communication (8). In subsequent studies this proved to be an exception as glutamine was the common C-terminal among the individuals studied.
It may be stressed that the present work relates to a rather homogeneous class of donors.
Whether the same results will extend to a more heterogeneous group of subjects remains to be established.

DISCUSSION
The present studies have shown, in accord with a recent report by this laboratory (3), that Fraction IV, isolated and purified from apo HDLz by Sephadex chromatography in 8 M urea (l), exists as dimers (mol wt of about 18,000) which are converted into protomers by disulfide cleavage. The protomers were shown to have the same molecular weight (about 8,500) and to contain one half-cystine per mole of protein, thus indicating that these protomers are linked as dimers by a single disulfide bridge.
The reasons for the chemical selectivity of such a pairing process are not apparent and reoxidation experiments presently underway in this laboratory are expected to help in resolving this problem.
In this context, it is important to establish whether the dimers are composed of identical or nonidentical monomers, a question which was not conclusively answered in the present studies.
Our data showed that the SC-IV protomers were indistinguishable from each other when reacted against antisera raised in the rabbit against unreduced IV.
On the other hand, their reactivity against anti-SC-IV was not tested. Thus, it may be simply concluded on the basis of these and the results in the accompanying paper (2) that HDLz contains heavy and light chains, the latter linked together as dimers. This conclusion does not account for Fraction V, which represents about 2 to 5% of the apo HDLz mass (I), and may not be considered as a true constituent of HDL2. The latter statement is based both on reassembly studies (12) and on our recent work on HDLp (d 1.125 to 1.21 g per ml) and on VHDL of d 1.21 to 1.25 g per ml which proves the existence of high density lipoprotein species having both III and IV but little (HDL3) and no (VHDL) Fraction V (unpublished).
Furthermore, there is strong evidence supporting the origin of all or most of HDL Fraction V from the process of VLDL-HDL exchange (13). The above conclusions open several questions of structural relevance.
It is possible that HDLS as well as the HDL3* class, as isolated by conventional cumulative flotational techniques may represent mixtures of subclasses varying in molecular weight, size, lipid content, and relative distribution of polypep-2 Nomenclature, Fraction IV is equivalent to R-Gln of Shore and Shore (9)