1074
BIOCHEMICAL SOCIETY TRANSACTIONS
toxin (1 pg/ml) or dibutyryl cyclic AMP (0.2m) to the culture medium in which the
cells are propagated.
The altered distribution of glycosylation in cells exposed to cholera toxin and
dibutyryl cyclic AMP, which is less evident in cells exposed to conditions that restrict
the expression of transformation, is also manifested in immune-precipitation reactions,
in which cellular glycoproteins are allowed to react with antisera versus murine
leukaemia virus components. In such experiments, immune serum detects a differential
effect of cholera toxin on glycosylation, which appears to depend on conditions that
allow or restrict the expression of malignant transformation in vivo.
Guerrant, R. L., Brunton, L. L., Schnaitman,T. C., Rebhun, L. 1. &Gilrnar, A. G. (1974)Infect.
Immun. 10,320-327
Hollenberg, M. D., Fishman, P. H., Bennett, V. & Cuatrecasas,P. (1974) Proc. Natl. Acad. Sci.
U.S.A. 71,4224228
Rieber, M. & Irwin, J. C. (1974) Cancer Res. 34,3469-3473
Rieber, M., Bacalao, J. &, Alonso, G. (197%) Cancer Res. 35,2104-2108
Rieber, M., Bacalao, J. & Alonso (19756) Cancer Res. 35,3009-3013
Liver Sinusoid Surface Membranes and Glycoprotein Secretion
MILOSLAV DOBROTA, FAIZ S. ISSA, BARBARA M. MULLOCK and
RICHARD H. HINTON
Worfson Bioanalytical Centre, University of Surrey, Guildford GU2 5XH, Surrey, U.K.
The liver is the source of most serum proteins (Miller & John, 1970). These are synthesized in the rough endoplasmic reticulum, transported to the Golgi apparatus and there
packaged into secretion granules. The same route is apparently followed by cell-coat
proteins (Keenen & M o d , 1975). In both cases, the secretion granules fuse with the
plasma membrane and discharge their contents, although serum proteins appear to be
temporarily associated with the plasma membrane and may be detected immunologically (Riordan et al., 1974). Export proteins will presumably be transported to the sinusoidal surface of the cell, but this will not necessarily be the case with cell-coat proteins.
The sheets of plasma membrane, which sediment with the crude nuclear fraction,
arederived principally from the bilecanalicular face of the cell (Wisher &Evans, 1975) so
one might expect that sinusoid membrane fragments would be found principally in the
microsomal fraction. This is supported by the concentration in microsomal plasma
membrane of glucagon-activated adenylate cyclase (Wisher & Evans, 1975).
However, the plasma-membrane fragments found in the microsomal fraction are
themselves heterogeneous (Norris e f al., 1974), and it is not clear whether all thevesicles
are derived from the sinusoid membranes. One approach to this problem is to label the
sinusoid membrane in situ. The dye SITS (4-acetamido-4-isothiocyanatostilbene2,2’-disulphonic acid), which bindscovalently to surface membranes without penetrating
into the cells (Knauf & Rothstein, 1971), appears to be suitable.
In the present experiments, sinusoid membranes were labelled by perfusion of the
livers with ~ O ~ M - S I(BDH
T S Chemicals, Poole, Dorset, U.K.). In separate experiments
newly made glycoproteins were labelled by the intravenous injection of ~-[l-’~C]fucose
(The Radiochemical Centre, Amersham, Bucks., U.K.) 15min before the death of the
animal (Riordan et al., 1974). Plasma-membrane sheets were separated from the crude
nuclear fraction of the liver homogenates as described by Hinton e f al. (1970); the microsoma1 preparations were subfractionated as described by Norris et al. (1974). Marker
enzymes were assayed as described by Prosper0 et al. (1973). The distribution of newly
made glycoprotein was determined by precipitation with 5 % (w/v) trichloroacetic acid,
collection on glass-fibre filters, and dispersing and counting for radioactivity as described
by Scherrer (1969). SITS fluorescence was measured at 483nm with excitation at 350nm
by using a Perkin-Elmer (Beaconsfield, Bucks., U.K.) spectrofluorimeter. KOH was
added to a concentration of 15 % (w/v) immediately before taking the reading.
1976
1075
565th MEETING, STIRLING
I .3
h
1.2
3
1.1
p
Q&
zj
n
I .o
Fig. 1 . Pattern obtained after fractionation, by centrifugation in an A-XZZ zonal rotor-of a
crude nuclear fraction prepared from the liver of ratsperfused with SITS dye
The rats were anaesthetized with diethyl ether and perfused through the portal vein with
0.25~-sucrose/5mhl-Tris,
pH 8.0,until the livers were completely blanched. Each liver
was then perfused with 30ml of SITS dye ( 5 0 , ~in~0.25~-sucrose/~mM-Tris,
pH8.0)
and then extensivelyperfused with 0.25 ~-sucrose/5
mM-Tris, pH 8.0, to remove unbound
SITS dye. Homogenization and centrifugation were done as described by Hinton et al.
(1970). The results were corrected for endogenous fluorescence by comparison with a
control experiment in which the SITS dye perfusion was omitted. Fluorescence,
enzyme activitiesand protein concentrationsare in arbitrary units. 0,SITS fluorescence;
...., protein; 0 , 5’-nucleotidase activity; -.-. , density (20°C).
4
2
10
20
30
0
Fraction no.
Fig. 2. Subfractionation of microsomal preparation from a rat labelled by intravenous
injection of ~ - [ ~ ~ C ] f u c15o min
s e before death
After homogenization of the livers, large particulates were removed by centrifugation
for 15 min at 11 500g.The microsomal fraction was collected by centrifugation for 1h
at 122000g resuspended in 2~-sucroseand layered under a linear sucrose gradient in a
B-XIV zonal rotor. Enzyme activities are given in arbitrary units. 0, 14Cradioactivity;
0,5‘-nucleotidase activity; A, alkaline phosphodiesterase activity; ... ., protein;-,
density.
Fluorescence microscopy of livers perfused with SITS showed that labelling was restricted to blood vessels and sinusoids. Difficulties were, however, encountered in localizing the SITS label in thesubcellular fractions owingto thehighendogenousfluorescence.
After correction, it was clear that the distribution of fluorescence among subfractions
Vol. 4
1076
BIOCHEMICAL SOCIETY TRANSACTIONS
separated from the nuclear fraction was generally similar to the distribution of 5’nucleotidase (Fig. l), but that there was proportionately more label in plasma-membrane
fragments of a microsomal size (fractions 4-6) than in plasma-membrane sheets (fractions 21-23). The fluorescence/5’-nucleotidaseratio in the microsomal fraction was
similar to that of the crude nuclear microsomal fraction. The high endogenous fluorescence, however, prevented clear localization of the SITS dye in submicrosomal fractions,
although there were some indications that the label was concentrated in the lowerdensity material.
The distribution of rapidly labelled glycoproteins among particles in the crude
nuclear fraction was extremely similar to the distribution of SITS fluorescence. Subfractionation of microsomal preparation (Fig. 2) showed that newly made glycoproteins
were concentrated in very-low-density vesicles presumably derived from the Golgi
apparatus (fractions 1-5) and in the lower-density plasma-membrane vesicles (fractions
11-14). As it was thought possible that this apparent localization could be due to secretion granules, the vesicles were treated with 0.15M-NaC1, to remove loosely bound protein, and0.2~-NaHCO~,
pH9.0, to remove vesicle contents ( h a & Hinton, 1974). The
results showed that about 45 % of the labelled glycoprotein could be solubilized from
the plasma-membrane vesicles as against 30% of the label in plasma-membrane sheets
from the crude nuclear fraction and over 60% of the material from the low-density
‘Golgi’ vesicles. Solubilization of the extracted membranes by 1% sodium deoxycholate
and 0.5% Lubrol W (Blomberg & Perlman, 1971) and analysis by immunodiffusion
against anti-(liver plasma membrane) confirmed that newly made membrane glycoproteins were concentrated in plasma-membranevesicles banding at a density of 1.14g/cm3
(fractions 11-14, Fig. 2).
The results presented above show tha: newly made serum and cell-coat glycoproteins
appear first in plasma-membranefragments, which appear to be derived from the sinusoidal surface of the cell. This suggests that there is no difference in packaging in the
Golgi apparatus between proteins destined for release into the serum and those that
will be incorporated into the cell coat. Proteins destined for eventual release into the
serum are probably associated with the plasma membrane, so explaining the immunological cross-reaction between well-washed plasma-membrane preparations and anti(rat serum)(Issa & Hinton, 1974). The subsequent release of the proteins into the serum
would then beexplained by their low affinity for the membrane as compared with the
proteins that form a permanent part of the cell coat.
We thank Mr. P. Trumper for performing the intravenous injections and our colleagues in the
Department of Biochemistry for use of the spectrofluorimeter. Financial support was provided
by the Medical Research Council and the Wellcome Trust.
Blomberg, F. & Perlmann, P. (1971) Biochim. Biophys. Actu 233,5340
Hinton, R. H., Dobrota, M., Fitzsimons, J. T. R. &Reid,E. (1970)Eur. J. Biochem. 12,349-361
Ism, F. S. & Hinton, R. H. (1974) Biochem. Soc. Truns.2,483-486
Kecnan. T . W. & Morr6, D. J. (1975) FEBSLett. 55,8-13
Knauf, P. A. & Rothstein, A. (1971)J. Gen. Physiol. 58,190-210
Miller, L. L. & John, D. W. (1970) in Plusmu Protein Metubohm (Rothschild, M . A. &
Waldmann, T., eds.), pp. 207-222, Academic Press, New York
Norris, K. A., Dobrota, M., Issa, F. S.,Hinton, R. H. & Reid, E. (1974) Biochem. J. 142,66767 1
Prospero, T. D., Burge, M. L. E., Norris, K. A., Hinton, R. H. & Reid, E. (1973) Biochem. J.
132,449-458
Riordan, J. R., Mitranic, M., Slavik, M. & Moscarello, M. A. (1974) FEBSLett. 47,248-251
Scherrer, K. (1969) in Fundamental Techniques in Virology (Hubel, K. & Saleman, H., eds.),
p. 315, Academic Press, New York
Wisher, M. H. & Evans, W. H. (1975) Biochem. J. 146,375-388
1976