The fractionation of human liver mitochondria into inner membrane, outer membrane and matrix material is reported. Compared with rat, human liver mitochondria are more fragile. Fractionation can be achieved in only 2 steps, a digitonin treatment for removal of the outer membrane and centrifugation of the inner membrane plus matrix particles through a linear sucrose gradient resulting in purified inner membranes and matrix.

The separation of the inner and outer membranes of rat liver mitochondria has been performed by various techniques: density-gradient centrifugation following mitochondrial swelling and contraction (Parsons et al. 1967; Sottocasa, Kuylenstierna, Ernster & Bergstrand, 1967), gradient centrifugation following controlled osmotic lysis (Schnaitman, Erwin & Greenawalt, 1967; Caplan & Greenawalt, 1968) and treatment with digitonin in isotonic media followed by differential centrifugation (Schnaitman et al. 1967; Levy, Toury & Andre, 1967; Schnaitman & Greenawalt, 1968). This has permitted the systematic study of submitochondrial enzyme localization and detailed lipid analysis of the 2 mitochondrial membrane systems (Beattie, 1968; Colbeau, Nachbaur & Vignais, 1971).

Studies of isolated human liver mitochondria are scarce (Bjorntorp, Bjorkerud & Schersten, 1965; Ozawa, et al. 1972) and so far fractionation of membranes has not been reported. The study of human liver mitochondria is important both for a better knowledge of the metabolic properties of human liver and to illustrate differences in structure and function between liver mitochondria in normal and disease states.

In previous papers (Benga, Muresan, Hodarnau & Dancea, 1972; Benga & Muresan, 1974; Benga & Borza, 1975) we have described the peculiarities of human liver mitochondria that determine special conditions for their isolation. A particular lipid and amino acid composition of human liver mitochondrial membranes, related to a greater fluidity as indicated by spin-labelled fatty acid motion, has been described (Benga & Ferdinand, 1977; Benga et al. 1978). A better understanding of the interaction between lipids and proteins in membranes and its significance in the characteristic behaviour of human mitochondria requires work on purified inner and outer mitochondrial membranes. It is known that the lipid composition and protein profile of the two membranes is different (Parsons et al. 1967; Colbeau et al. 1971). The purification of the inner and outer membranes of human liver mitochondria raises, in principle, some problems taking into account the fragility of human liver mitochondria (Benga et al. 1972) as well as the small amount of tissue available for analysis. In this paper we describe the morphological and enzymic characteristics of human mitochondrial membranes isolated by a procedure, simple and rapid in comparison with that for rat liver mitochondria.

Preparation of mitochondria

Human tissue was obtained by intraoperatory biopsy from patients with abdominal diseases (duodenal or gastric ulcers, cholecystitis, uncomplicated gallstones) as previously described (Benga et al. 1972). Human and rat liver mitochondria were isolated as previously described (Benga, 1973; Benga et al. 1978). Mitochondria were tested for respiratory control and ADP/O ratios polarographically (Estabrook, 1967) and spectrophotometrically (Barzu, Muresan & Benga, 1972). For the fractionation studies, mitochondria having a respiratory control index above 4 · 5 and an ADP/O ratio of 2 · 7 or greater (with glutamate as substrate) were used. The integrity and purity of mitochondria was also checked by electron microscopy, as previously described (Benga et al. 1978). Assays of acid phosphatase and glucose-6-phosphatase showed that the contamination of rat and human mitochondria with lysosomes and microsomes was not greater than reported by others for rat liver mitochondria (Benga et al. 1978). Acid phosphatase activity and glucose-6-phosphatase activity were measured as previously described (Benga et al. 1972; Benga & Muresan, 1974).

Marker enzymes

The following markers were used to assess the purity of the fractions: monoamine oxidase (EC 1.4.3.4) for the outer membrane, adenylate kinase (EC 2.7.4.3) for material in the space between the inner and outer membrane (referred to in this paper as the intermembrane fraction), 3-hydroxybutyrate dehydrogenase (EC 1.1.1.3) and adenosine triphosphatase (EC 3.6. 1.4) for the inner membrane and glutamate dehydrogenase (EC 1.1.1.2) for matrix.

Monoamine oxidase was assayed by a modification of the method of Tabor, Tabor & Rosenthal (1955) by following the benzaldehyde spectrophotometrically at 250 nm at 25 °C in an assay system containing 2 mM benzylamine hydrochloride and 50 mM phosphate buffer (pH 7’4). Samples were activated with Lubrol (1 mg/mg of protein) for 20 min at o °C to minimize changes in optical density associated with mitochondrial swelling.

Adenylate kinase was assayed spectrophotometrically at 25 °C by following the conversion of ADP to ATP + AMP and coupling the formation of ATP to the reduction of NADP with hexokinase and glucose-6-phosphate dehydrogenase. The assay mixture contained the following in 1 ml: 0 2 mM NADP, 10 mM glucose, 10 i.u. of hexokinase, 10 i.u. of glucose-6-phosphate dehydrogenase, 0·5 mM ADP, 5 mM MgCla and 50 mM Tris-HCl buffer, pH 7 · 5. The assay mixture was allowed to incubate for about 5 min for consumption of any trace amounts of ATP present, and the reaction was initiated by the addition of enzyme.

3-IIydroxybutyrate dehydrogenase was assayed spectrophotometrically by following the reduction of NAD at 25 °C as described by Hoppel & Cooper (1969) in an assay medium (10 ml) containing 1 mM NAD, 0·1 mM KCN, 7 · 5 mM Tris-HCl (pH8-i) and 2 · 5 mM 3-hydroxybutyrate.

Adenosine triphosphatase was measured as previously described (Benga et al. 1972).

Glutamate dehydrogenase was assayed at 25 °C as described by Caplan & Greenawalt (1968) in a medium (1 · 0 ml) containing 0 · 1 mM NADH, 0 · 1 mM NH4C1, 50 mM α-oxoglutarate, 1 / μ M EDTA, 0 · 4 mM ADP, 8 · 0 mM Tris-HCl, pH 8 · 0. Samples were activated with Lubrol (1 mg/mg protein for 20 min at 0 °C).

Protein was estimated by the method of Lowry, Rosebrough, Farr & Randall (1951). Albumin was defatted according to the method of Chen (1967).

Electron microscopy

Negative staining was performed with 1 % sodium phosphotungstate. The mitochondrial suspension (about 20 mg protein/ml) was diluted in a medium containing 0 · 1 M phosphate buffer (pH 7 · 4), 1 % sodium phosphotungstate and 0 · 1 % bovine serum albumin. After 3 min a drop of the diluted suspension was placed on a grid covered with a thin film of parlodion enhanced by a thin film of evaporated carbon. The drop was blotted off with filter paper, and the grid was allowed to air-dry. The fractions of mitochondrial membranes collected from the sucrose gradients were first diluted with 20 vol. of 0 · 1 M phosphate buffer and centrifuged at 100000 g for 60 min. The pellets were suspended in a small volume of 01 M phosphate buffer and then aliquots were taken for dilution for negative staining.

For positive staining small aliquots of samples (10—20 μl suspension containing 2—15 mg protein/ml) were added to 0 · 2 ml of 2 · 5 % glutaraldehyde in 0 · 15 M phosphate buffer (pH 7’4) placed in glass microtubes. The samples were well mixed and left 1 h for prefixation at 4 °C. The microtubes were then centrifuged 5 min at 5000 g. The supernatant was decanted and replaced with phosphate buffer 0 · 15 M (pH 7 · 4). The phosphate buffer was changed 3 times and then replaced with 2 % OsO4 in 0 · 1 M phosphate buffer (pH 7 · 4) for definitive fixation. All the subsequent steps (dehydration in acetone and embedding in Westopal) were performed in the same glass microtubes. Thin sections were stained with uranyl acetate in 50 % ethanol for 60 min at 40 °C and then with lead citrate by the procedure of Reynolds (1963). They were then examined in a Hitachi UHU 11A electron microscope.

Fractionation of rat liver mitochondria

The fractionation procedure of rat liver mitochondria, derived from the method of Schnaitman et al. (1968), is described in Fig. 1. It can be seen that for the rat liver mitochondria the treatment with digitonin selectively removes the outer membrane leaving a rather intact inner membrane plus matrix fraction. This fraction has to be further treated with Lubrol in order to release matrix material. After a second centrifugation on a linear sucrose gradient one can obtain purified inner membranes and matrix. Thus 4 steps are necessary for the purification of rat liver mitochondrial membranes: the digitonin treatment for the removal of the outer membrane, the purification of the inner membrane plus matrix fraction by sucrose gradient centrifugation, the treatment with Lubrol followed by a second sucrose gradient centrifugation.

Fig. 1.

Fractionation of rat liver mitochondria. Aliquots (1–2 ml) of mitochondrial suspensions in 0 · 25 M sucrose containing about 40 mg protein/ml were placed in an ice bath, and identical aliquots of cold digitonin solution (6 mg/ml in 0 · 25 M sucrose) were added. The suspensions were incubated at 0 °C with continuous stirring for 20 min after the addition of the digitonin. The digitonin-treated suspensions were centrifuged at 9500 g for 15 min. The supernatant (I) contains the crude outer membranes and the intermembrane fraction. The pellet (I) contains the crude inner membrane plus matrix fraction. Supernatant (I) was centrifuged at 105000 g for 60 min. The resulting pellet (II) contains the crude outer membranes, while the supernatant (II) corresponds to the intermembrane fraction. Both the pellets (I) and (II) were resuspended in 0 · 25 M sucrose, 10 mM Tris-HCI and purified by centrifugation through a discontinuous sucrose gradient (1 · 2 ml of 23 · 2 % sucrose, 12 ml of 37’4 % sucrose and 1-8 ml of 51-3 % sucrose, all in 20 mM phosphate buffer, pH 7 4). From pellet (II) the purified outer membrane fraction was obtained (layer 1). From pellet (I), layer 4 corresponds to inner membrane and pellet 5 to the pure inner membrane plus matrix particles. These 2 fractions were treated with Lubrol WX (0·5 mg/mg of protein), incubated for 15 min at o °C and purified again on similar linear sucrose gradients. The resulting layers, 6 and 7, correspond to the purified inner membranes, while layer 8 corresponds to the matrix.

Fig. 1.

Fractionation of rat liver mitochondria. Aliquots (1–2 ml) of mitochondrial suspensions in 0 · 25 M sucrose containing about 40 mg protein/ml were placed in an ice bath, and identical aliquots of cold digitonin solution (6 mg/ml in 0 · 25 M sucrose) were added. The suspensions were incubated at 0 °C with continuous stirring for 20 min after the addition of the digitonin. The digitonin-treated suspensions were centrifuged at 9500 g for 15 min. The supernatant (I) contains the crude outer membranes and the intermembrane fraction. The pellet (I) contains the crude inner membrane plus matrix fraction. Supernatant (I) was centrifuged at 105000 g for 60 min. The resulting pellet (II) contains the crude outer membranes, while the supernatant (II) corresponds to the intermembrane fraction. Both the pellets (I) and (II) were resuspended in 0 · 25 M sucrose, 10 mM Tris-HCI and purified by centrifugation through a discontinuous sucrose gradient (1 · 2 ml of 23 · 2 % sucrose, 12 ml of 37’4 % sucrose and 1-8 ml of 51-3 % sucrose, all in 20 mM phosphate buffer, pH 7 4). From pellet (II) the purified outer membrane fraction was obtained (layer 1). From pellet (I), layer 4 corresponds to inner membrane and pellet 5 to the pure inner membrane plus matrix particles. These 2 fractions were treated with Lubrol WX (0·5 mg/mg of protein), incubated for 15 min at o °C and purified again on similar linear sucrose gradients. The resulting layers, 6 and 7, correspond to the purified inner membranes, while layer 8 corresponds to the matrix.

Isolation and purification of human liver mitochondrial fractions

We have tried several procedures for the fractionation of human liver mitochondria. The separation of the inner and outer membranes by high-amplitude swelling in the presence of inorganic phosphate was found not to be suitable because of the small amount of biological material available. The digitonin treatment described by Schnaitman & Greenawalt (1968) followed by Lubrol treatment, as for rat liver mitochondria, did not give good results because of the greater fragility of human liver mitochondria. If the purified inner membrane plus matrix fraction is subjected to the Lubrol treatment small vesicles containing both membranes and matrix are obtained rather than a separation of the matrix from the purified inner membranes.

The present fractionation of human liver mitochondria is described in Fig. 2. It can be seen that after treatment with digitonin (0-5 mg/mg of protein) the outer membrane is removed and can be separated from the intermembrane fraction by a sucrose gradient centrifugation. When the pellet I (crude inner membrane plus matrix) is subjected to centrifugation through the linear sucrose-gradient, then purified inner membrane and matrix fractions are obtained (layers 4 and 1, respectively, in Fig. 2). This probably means that the inner membrane plus matrix particles are fragmented during the centrifugation rendering the Lubrol treatment unnecessary. Thus the fractionation of human liver mitochondria can be achieved in just 2 steps: the digitonin treatment for the removal of the outer membrane and the centrifugation of the inner membrane plus matrix particles through a linear sucrose gradient resulting in purified inner membranes and matrix. Fig. 3 shows the striking comparison between the behaviour of rat and human liver mitochondria during gradient centrifugation of the pellet (I) obtained after digitonin treatment.

Fig. 2.

Fractionation of human liver mitochondria. Aliquots (0 · 5–1 ml) of mitochondrial suspensions in 0 · 25 M sucrose containing about 30 mg protein/ml were placed in an ice bath and identical aliquots of cold digitonin solutions (6 mg/ml in 0 · 25 M sucrose) were added. After 20 min of incubation at o °C with continuous stirring the digitonin-treated suspensions were centrifuged at 9500 g for 15 min. The supernatant (I) was centrifuged at 105000 g for 60 min for separation of intermembrane fraction (Supernatant II) from the crude outer membranes (Pellet II). From the pellet II (resuspended in 0 · 25 M sucrose, 10 HIM Tris-HCI, pH 7 · 4) purified outer membranes (layer 5) and inner membranes (layer 6) are obtained by centrifugation through a discontinuous sucrose gradient (1 · 2 ml of 23 · 2 % sucrose, 12 ml of 37 · 4 % sucrose and 18 ml of 51 · 3 % sucrose, all in 20 mM phosphate buffer, pH 7 · 4). From pellet (I) (crude inner membrane plus matrix fraction) several fractions are obtained by centrifugation through a linear sucrose gradient: matrix (layer 1), submitochondrial particles (layer 2), outer membrane (layer 3), inner membrane (layer 4) and inner membrane plus matrix particles (Pellet III).

Fig. 2.

Fractionation of human liver mitochondria. Aliquots (0 · 5–1 ml) of mitochondrial suspensions in 0 · 25 M sucrose containing about 30 mg protein/ml were placed in an ice bath and identical aliquots of cold digitonin solutions (6 mg/ml in 0 · 25 M sucrose) were added. After 20 min of incubation at o °C with continuous stirring the digitonin-treated suspensions were centrifuged at 9500 g for 15 min. The supernatant (I) was centrifuged at 105000 g for 60 min for separation of intermembrane fraction (Supernatant II) from the crude outer membranes (Pellet II). From the pellet II (resuspended in 0 · 25 M sucrose, 10 HIM Tris-HCI, pH 7 · 4) purified outer membranes (layer 5) and inner membranes (layer 6) are obtained by centrifugation through a discontinuous sucrose gradient (1 · 2 ml of 23 · 2 % sucrose, 12 ml of 37 · 4 % sucrose and 18 ml of 51 · 3 % sucrose, all in 20 mM phosphate buffer, pH 7 · 4). From pellet (I) (crude inner membrane plus matrix fraction) several fractions are obtained by centrifugation through a linear sucrose gradient: matrix (layer 1), submitochondrial particles (layer 2), outer membrane (layer 3), inner membrane (layer 4) and inner membrane plus matrix particles (Pellet III).

Fig. 3.

Pellets I (see Figs. 1, 2) from rat (A) and human (B) liver mitochondria after centrifugation through a sucrose gradient. Experimental conditions as in Figs. 1 and 2.

Fig. 3.

Pellets I (see Figs. 1, 2) from rat (A) and human (B) liver mitochondria after centrifugation through a sucrose gradient. Experimental conditions as in Figs. 1 and 2.

Enzymic characterization of the fractions

Table 1 summarizes the specific activities of marker enzymes in the fractions of rat liver mitochondria. According to the different markers, the inner membranes were enriched about 5-fold. Significant glutamate dehydrogenase activity remained associated with the inner membrane fraction, as in the case of other sources of mitochondria (Maisterrena, Comte & Gautheron, 1974; Colbeau et al. 1971). The outer membranes were purified 12-fold as shown by the monoamine oxidase activity. Contamination of this fraction by inner membranes was very little.

Table 1.

Specific activities of marker enzymes in rat liver mitochondrial fractions

Specific activities of marker enzymes in rat liver mitochondrial fractions
Specific activities of marker enzymes in rat liver mitochondrial fractions

Table 2 shows the distribution and specific activity of marker enzymes and protein in the fractions of human liver mitochondria. In this case the inner membranes were purified about 6-fold and there was little contamination with outer membranes, despite the fact that only a single centrifugation through a sucrose gradient was used. Again, significant glutamate dehydrogenase activity remained associated with the inner membrane. The outer membranes were purified about 14-fold, as indicated by monoamine oxidase activity. Practically all the adenylate kinase activity was localized in the intermembrane fraction (supernatant II), both for human and for rat liver mitochondria.

Table 2.

Distribution of protein and specific activities of marker enzymes in human liver mitochondrial fractions

Distribution of protein and specific activities of marker enzymes in human liver mitochondrial fractions
Distribution of protein and specific activities of marker enzymes in human liver mitochondrial fractions

Electron microscopy

The purification and identification of the membrane fractions was also followed by electron microscopy. The morphological aspects corroborate the results obtained with marker enzymes. In the case of rat liver membranes the electron-microscopical appearance of the fractions was that expected on the basis of enzymic characterization and similar to that described by other authors. For example, the aspect of the outer membrane fraction in Fig. 4A and Fig. 4B can be compared with outer membranes shown in fig. 6 c of the paper of Caplan & Greenawalt (1968) or the micrograph in the paper of Hayashi & Capaldi (1972).

Fig. 4.

Electron micrographs of the outer membranes of rat liver mitochondria (layer 3 in Fig. 1). This fraction comprises primarily small single membrane-limited vesicles, about 500 nm in diameter. A, positive staining; B, negative staining, × 40000.

Fig. 4.

Electron micrographs of the outer membranes of rat liver mitochondria (layer 3 in Fig. 1). This fraction comprises primarily small single membrane-limited vesicles, about 500 nm in diameter. A, positive staining; B, negative staining, × 40000.

For human liver mitochondria several typical electron micrographs have been selected. In Fig. 5 A and B the appearance of positively and negatively stained human liver mitochondria, respectively, is presented. The electron micrographs of various fractions obtained after centrifugation confirm the results of determination of marker enzymes. The negatively stained fraction 4 (see notation in Fig. 2) contains many convoluted membranous structures which, on high magnification (Fig. 5D), show the 9’0-nm knobs characteristic of inner membrane. Taking into account the appearance of the fixed thin-section of the same fraction (Fig. 5 c) the material appears to comprise mainly inner membranes. The appearance of the other inner membrane fraction (layer 6 in Fig. 2) is similar, although the membranes enclose some stainable matrix material (Fig. 5E). Fraction 3 (see notation in Fig. 2) contains primarily outer membranes (Fig. 5 F) as does fraction 5 (Fig. 5 G). The latter fraction seems to be purer than fraction 3, and contains smaller vesicles. This might be expected taking into account the effects of digitonin treatment.

With regard to the other fractions isolated by the procedure shown in Fig. 2, the pellet III showed the appearance of the inner membrane plus matrix particles, while the layer 2 was heterogenous, containing fragments of inner membrane plus matrix, as well as fragments of outer membrane.

Our major problem was to fractionate human liver mitochondria into purified inner and outer membranes. From the marker enzyme activities and electron microscopy we can conclude that our objective has been reached. The inner membranes were purified about 6-fold and were contaminated only to a minor extent by outer membranes. The specific activities of the marker enzymes were comparable to those reported in the literature for rat liver (Schnaitman et al. 1967; Schnaitman & Greenawalt, 1968). The ‘contamination’ of the inner membrane fraction by glutamate dehydrogenase was comparable to that described for rat liver by Schnaitman et al. (1967), Schnaitman & Greenawalt (1968) and the possibility of an association of some of the matrix enzymes with inner membranes has been postulated (Maisterrena et al. 1974).

The outer membranes were purified approximately 14-fold which is comparable to the best preparations of outer membranes of rat liver mitochondrial outer membranes (Sottocasa et al. 1967; Schnaitman & Greenawalt, 1968; Colbeau et al. 1971).

As far as the yield of purified membranes from human liver mitochondria is concerned, the inner membranes (layer 4 plus 6) represented about 12% and outer membranes (layers 3 and 5) about 3% of the total mitochondrial protein. From a liver biopsy of i g one can get about 25–30 mg of mitochondrial protein. If this is used for isolation of mitochondrial membranes this means that about 3 mg of inner membranes and 0 · 7 mg of outer membranes can be obtained.

A major difficulty in the fractionation of human liver mitochondria is due to their fragility, rendering the Lubrol fractionation procedure (Schnaitman & Greenawalt, 1968) unsuitable. However, as the present results show, we have found a procedure for the purification of inner and outer membranes of human liver mitochondria which takes particular advantage of their fragility. The procedure described in the present paper involves only one centrifugation through a sucrose gradient of the crude inner membrane plus matrix particles. This centrifugation is sufficient to fractionate the crude inner membrane plus matrix fraction from human liver into inner, outer membranes and matrix. The fractions obtained in this way were rather pure as assessed by the biochemical and electron-microscopical data. For both sources of mitochondria, rat and human, the treatment with digitonin strips off the outer membrane (as first described by Levy et al. 1967, for rat liver mitochondria). However, only in the case of human liver mitochondria does the digitonin treatment have a marked effect on the stability of the inner membranes plus matrix particles; this effect is shown by the fact that during a subsequent centrifugation through a sucrose gradient the inner membrane plus matrix particles of human liver mitochondria are disrupted, possibly as a result of the osmotic effects of the high concentration of sucrose in the gradient. A similar disruption of the rat liver mitochondria could be noticed only after subjecting them to osmotic shock (water washing) as shown by Schnaitman et al. (1967). If rat liver mitochondria are subjected to the digitonin treatment in isotonic solution a subsequent treatment with Lubrol is necessary in order to disrupt the inner membrane plus matrix particles.

The greater fragility of human liver mitochondria in comparison with rat liver mitochondria appears to originate in the peculiarities of their lipid and protein composition. As we have previously shown (Benga et al. 1978) human liver mitochondria contain much more lipid than those from the rat. The increase in lipid in human liver originates from a higher phospholipid content. There are no significant differences in the proportion of major phospholipid classes between the 2 types of mitochondria, but there are differences in the constituent fatty acid compositions. Human liver mitochondria contain more linoleic acid and less arachidonic acid than those of the rat (Benga et al. 1978). This pattern of fatty acid composition of mitochondrial phospholipids has not been reported for any other species. Since phospho-lipids are the characteristic lipid of cell membranes, it is clear that the lipid composition of mitochondrial membranes will be different in human and rat liver. The present results suggest that this difference is manifest by differences in fragility.

We have also shown that, in parallel with a higher content of lipid, human liver mitochondrial membranes contain a significantly higher fraction of hydrophobic amino acids. The nature of protein-lipid interactions in biological membranes is of Fig. 5. Electron micrographs of human liver mitochondria before and after fractionation. (For fraction notations see Fig. 2.) A, positive staining of intact mitochondria, × 14000. B, negative staining of intact mitochondria, × 40000. c, positive-stained appearance of fraction 4. × 40000. D, negative-stained appearance of fraction 4 (inner membrane), × 80000. E, positive-stained appearance of layer 6 (inner membrane plus matrix), × 40000. F, positive-stained appearance of fraction 3 (outer membrane), × 40000. G, positive-stained appearance of fraction 5 (outer membrane), × 36000. great importance for membrane structure and function and for membrane stability. Both electrostatic and hydrophobic interactions have been implicated. One might exp ect that in the case of human liver mitochondria the role played by the hydrophobic interactions between membrane proteins and the lipid bilayer is structurally more important than in the case of rat, where electrostatic interactions could occur to a significantly greater degree. On the other hand, the ratio of protein to lipid in the membranes of the 2 species might be important for an understanding of the greater fragility of human liver mitochondria. A parallel with other membranes might be relevant. Thus it has been shown that camel erythrocytes have an exceptionally high osmotic stability, much higher than the human erythrocyte (Livne & Kuiper, 1973). This unique stability has been explained by the peculiarities of protein-lipid interactions in the erythrocyte membranes. Camel erythrocyte membranes have a much higher protein to lipid ratio compared with the human. The membrane proteins in camel erythrocyte are also more basic than those in the human, with a higher proportion of arginine and a lower proportion of glutamic acid. Taking into account the acidic nature of phospholipids in the membranes, stronger electrostatic attractions between lipids and proteins in camel membrane have been postulated compared to the human erythrocyte membrane to explain their particular stability (Livne & Kuiper, 1973).

On a similar line of thinking, the low protein to lipid ratio and the higher proportion of hydrophobic amino acids in human liver mitochondrial membranes may be responsible for weaker protein-lipid interactions for these membranes compared to the rat liver mitochondrial membranes. Indeed a higher mobility of the membrane lipids in human liver mitochondrial membranes has been found using spin-labelled fatty acids to probe the fluidity of membrane lipids (Benga et al. 1978). The higher fluidity of human membranes might arise from a lesser immobilization of lipids by proteins in comparison with the rat liver membranes.

The procedures for purification of human liver mitochondrial membranes described in this paper will allow detailed studies of lipid composition, enzymic properties and physical properties to be made on these membranes in normal and pathological conditions.

The kind co-operation of surgeons of the Hird Surgical Clinic of the Medical and Pharmaceutical Institute, Cluj-Napoca, is acknowledged. The authors thank the Academy of Medical Sciences of Roumania, Ministry of Education and Teaching of Roumania, and the British Council for financial support for the work and scientific visits.

Barzu
,
O.
,
Muresan
,
L.
&
Benga
,
GH
. (
1972
).
Spectrophotometric method for assay of mitochondrial oxygen uptake. IV. Study of the respiratory chain in small amounts of biological materials
.
Analyt. Biochem
.
46
,
374
387
.
Beattie
,
D. S.
(
1968
).
Enzyme localisation in the inner and outer membranes of rat liver mitochondria
.
Biochem. biophys. Res. Commun
.
31
,
901
907
.
Benga
,
GH
. (
1973
).
Caracterizarea biochimica a mitocondriilor din ficatul uman
.
Stud, cercat. biochim
.
16
,
251
260
.
Benga
,
GH.
&
Borza
,
V.
(
1975
).
Differences in reactivity of cytochrome oxidase from human liver mitochondria with horse and human cytochrome c
.
Archs Biochem. Biophys
.
169
,
354
357
.
Benga
,
GH.
&
Ferdinand
,
W.
(
1977
).
Increased content of hydrophobic amino acid residues in lipid-rich
mitochondrial membranes
:
a comparison of rat and human liver mitochondria
.
Int. J. Biochem
.
8
,
17
20
.
Benga
,
GH.
,
Hodarnau
,
A.
,
Bohm
,
B.
,
Borza
,
V.
,
Tilinga
,
R.
,
Dancea
,
S.
,
Petrescu
,
I.
&
Ferdinand
,
W.
(
1978
).
Human
liver mitochondria
:
relation of a particular lipid composition to the mobility of spin-labelled lipids
.
Eur. J. Biochem
.
84
,
625
633
.
Benga
,
GH.
&
Muresan
,
L.
(
1974
).
Human liver mitochondria. III. ATPase activity as an index of mitochondrial damage
.
Biochem. Med
.
10
,
131
145
.
Benga
,
GH.
,
Muresan
,
L.
,
Hodarnau
,
A.
&
Dancea
,
S.
(
1972
).
Conditions for isolation and study of enzymic properties of human liver mitochondria
.
Biochem. Med
.
6
,
508
521
.
Bjorntorp
,
P.
,
Bjorkerud
,
S.
&
Schersten
,
T.
(
1965
).
Subcellular fractionation of human liver
.
Biochim. biophys. Acta
111
,
375
383
.
Caplan
,
A. I.
&
Greenawalt
,
J. W.
(
1968
).
The effects of osmotic lysis on the oxidative phosphorylation and compartmentation of rat liver mitochondria.f
.
Cell Biol
.
36
,
15
31
.
Chen
,
R. F.
(
1967
).
Removal of fatty acids from serum albumin by charcoal treatment
,
J. biol. Chem
.
242
,
173
181
.
Colbeau
,
A.
,
Nachbaur
,
J.
&
Vignais
,
P. M.
(
1971
).
Enzymic characterisation and lipid composition of rat liver subcellular membranes
.
Biochim. biophys. Acta
249
,
462
492
.
Estabrook
,
R. W.
(
1967
).
Mitochondrial respiratory control and the polarographic measurement of ADP:O ratios
.
Meth. Enzym
.
10
,
41
47
.
Hayashi
,
H.
&
Capaldi
,
R. A.
(
1972
).
The proteins of the outer membranes of beef heart mitochondria
.
Biochim. biophys. Acta
282
,
166
173
.
Hoppel
,
C.
&
Cooper
,
C.
(
1969
).
The action of digitonin on rat liver mitochondria
.
Biochem. J
.
107
,
367
375
.
Levy
,
M.
,
Toury
,
R.
&
Andre
,
J.
(
1967
).
Separation des membranes mitochondriales: La membrane externe
.
Biochim. biophys. Acta
135
,
599
613
.
Livne
,
A.
&
Kuiper
,
P. J. C.
(
1973
).
Unique properties of the camel erythrocyte membrane
.
Biochim. biophys. Acta
318
,
41
49
.
Lowry
,
O. H.
,
Rosebrough
,
N. J.
,
Farr
,
A. L.
&
Randall
,
R. J.
(
1951
).
Protein measurement with the Folin phenol reagent
,
J. biol. Chem
.
242
,
173
181
.
Maisterrena
,
B.
,
Comte
,
J.
&
Gautheron
,
D. C.
(
1974
).
Purification of pig heart mitochondrial membranes. Enzymatic and morphological characterisation as compared to microsomes
.
Biochim. biophys. Acta
367
,
115
126
.
Ozawa
,
K.
,
Kitamura
,
O.
,
Mizukami
,
T.
,
Yamaoka
,
Y.
,
Kamano
,
T.
,
Takeda
,
H.
,
Takasan
,
H.
&
Honjo
,
J.
(
1972
).
Human liver mitochondria
.
Clin. chim. Acta
38
,
385
393
.
Parsons
,
D. F.
,
Williams
,
G. R.
,
Thompson
,
W.
,
Wilson
,
D.
&
Chance
,
B.
(
1967
).
In Mitochondrial Structure and Compartmentation
(ed.
E.
Quagliariello
,
S.
Papa
,
E. C.
Slater
&
J. M.
Tager
), pp.
29
70
.
Bari
:
Adriatica Editrice
.
Reynolds
,
E. S.
(
1963
).
The use of lead citrate at high pH as an electron-opaque stain in electron microscopy
,
J. Cell Biol
.
17
,
208
212
.
Schnaitman
,
C.
,
Erwin
,
V. G.
&
Greenawalt
,
J. W.
(
1967
).
The submitochondrial localisation of monoamine oxidase. An enzymic marker for the outer membrane of rat liver mitochondria
.
J. Cell Biol
.
32
,
719
735
.
Schnaitman
,
C.
&
Greenawalt
,
J. W.
(
1968
).
Enzymic properties of the inner and outer membranes of rat liver mitochondria
,
J. Cell Biol
.
38
,
158
175
.
Sottocasa
,
G. L.
,
Kuylenstierna
,
B.
,
Ernster
,
L.
&
Bergstrand
,
A.
(
1967
).
An electrontransport system associated with the outer membrane of liver mitochondria. A biochemical and morphological study
,
J. Cell Biol
.
32
,
415
438
.
Tabor
,
C. W.
,
Tabor
,
H.
&
Rosenthal
,
S. M.
(
1955
).
Amine oxidase from steer plasma
.
Meth. Enzym
.
2
,
390
393
.