Eur. J. Biochem. 135, 413-423 (1983)
6 , FEBS 1983
Topography of the total protein population
from cultured cells upon fractionation by chemical extractions
Johannes A. LENSTRA and Hans BLOEMENDAL
Department of Biochemistry, University of Nijmegen
(Keceived April 27, 1083) - EJB 83 0431
1. Chemical extractions are proposed as a major tool for a fractionation of cellular proteins. As a model
system, proteins from cultured hamster lens cells have been divided by independent extractions into seven sub-
cellular fractions, corresponding to water-soluble proteins and the proteins from membranes, microfilaments (and,
other deoxycholate-soluble proteins), intermediate filaments, microtubules, polysomes and nuclei respectively.
The latter two fractions have becn subfractionated yielding ribosomal proteins, the elongation and initiation f, act o r s
of the protein-synthesis machinery, chromatin proteins and non-chromatin proteins.
2. The protein compositions of the fractions have been analyzed by one-dimensional and two-dimensional
gel electrophoresis. This resulted in an almost complete topography of the proteins detected on two-dimensional
gels of total-cell lysates.
3. Comparison of two-dimensional patterns of proteins from the total-cell lysate and proteins from hamster
erythrocytes or from liver, muscle or brain tissue showed that the different cell types have only few proteins in com-
mon. Two proteins are common to all of these cell types, namely actin and a 68-kDa protein. The latter protein
was, like actin, vimentin and the tubulin subunits, also present in most cell fractions. Evidence is presented that this
protein is identical to a 68-kDa heat-shock protein.
Subcellular fractionation of cells and tissues has been
fundamental to much of our understanding of the processes of
life. Classical fractionation procedures werc based mainly on
centrifugation techniques [I -31 and the activities of marker
enzymes to assess the purity of the separated components 141.
In contrast, recently developed methods rely more and more on
diferential resistance of cell components against various ex-
traction buffers, for instance the resistance of cytoskeletal
components against noii-ionic dctcrgents. However, most
rcports deal with the isolation of only one or two subccllular
fractions [5- 141, some of which have only an operational
definition (e.g. the Triton X-100 pellet). Further, no attempt
has becn madc to combine different methods into a complete
fractionation of cells, and to classify cellular proteins on the
base of such a fractionation.
Recently we made a classification of thc most abundant
proteins of HeLa cells on the base of the protein compositions
of intermediate filaments, mcmbranes, nuclei, ribosomes and
protein-biosynthesis factors [Is]. Here we describe a more
complete fractionation, using cultured hamster lens cells
transformed by SV40 as a model system. The procedures wcrc
based mainly on independent chemical extractions and rcsultcd
in an easy and expedient preparations of seven subcelluliii-
fractions, representing water-soluble proteins, membranes,
actin filaments together with other deoxycholate-soluble
proteins, intermediate filaments, microtubules, polysomes
and nuclei. Subfractionation of the latter two fractions
yielded ribosomes and protcin-biosynthesis factors, and
chromatin and non-chromatin (tentatively designated as
nuclear matrix) respectively.
Ahbrrviutiows. SDS, sodium dodecyl sulfatc; Tris/Mg/K buffer,
50 mM TrisiHCl, 5 mM MgC12, 25 mM KCI, pH 7.4.
En79nes. Deoxyribonuclcase (EC 126.96.36.199); ribonuclease A
Analysis by one-dimensional and two-dimensional gel
electrophoresis showed that the fractions have characteristic
protein compositions. Only a few proteins are prcscnt in more
than one fraction, leading to an almost complete classification
of the proteins detected on two-dimensional gels of total-cell
MATERIALS AND METHODS
Cell c d t u r ~ and labelling
Hamster lens epithelial cells were brought into culture,
transformed by SV40 and grown in suspension as describcd
. Cells were labelled overnight at a density of 106/ml
Eagle's minimal essential medium (containing Earle's salts
and 0.2'%, w/v NaHC03 but without methionine) by adding
5 - 10 pCi [35S]niethionine (Amersham International, about
1000 mCi/mol). Addition after 3 h of 0.1 vol. of the normal
growth medium with lo:< (v/v) newborn calf serum appeared
to enhance the labelling of the polyribosomal proteins; there-
fore, this procedure was used for the preparation of the
After labelling, thc cells were collected by centrifugation
(5 min, 600xg) and washed three times with Ca2+/Mg2+/
Tyrode solution containing 0.1 "/, (w/v) EDTA.
Cdl, f ivc t inna t ion
Unless otherwise stated, fractionations were carried out at
0 - 4 "C with buffers containing I niM phenylmethylsulfonyl
fluoridc. Sodium deoxycholate was always added from a
frcshly prepared 25'x (w/v) solution. DNase 1 (Sigma
Chemical Co.) was added from a 1 mg/ml stock solution
containing 3 niM phenylmcthylsulfonyl fluoride to inhibit
contaminating chymotrypsin .
Total-cell lysates were prepared by freezing and thawing
thc cells in lop1 Tris/Mg/K buffer/lO" cells, adding 2yg/
10"cells DNase I, leaving at room temperature for 15 min and
adding 1 vol. 2 x concentrated SDS-containing lysis buffer
. Any remaining viscosity was removed by passing the
solution repeatedly through a needle with an internal diameter
of 0.3 mm. Degrading the DNA appeared to be essential if
samples were to be analyzed by two-dimensional gel electro-
phoresis, because DNA disturbed the basic part of the pH
gradient . Further, it was found that lysing the cells in
8 M urea with 2
lines and 5 % (v/v) 2-mercaptoethanol resulted in the loss of
several proteins, as judged from two-dimensional gel electro-
Watcr-soluble proteins were prepared by lysing 1 Oh cells/ml
Tris/Mg/K buffer by 15 strokes in a Dounce homogenizer
with a B pestle. After centrifugation in a Ti50 rotor at
120000 x g for 45 min, soluble proteins were collected in the
Membrane proteins were prepared by lysing lo6 cells/ml
membrane buffer (1 iiiM NaHC03, 1 mM CaC12, pH 8.0) by
15 strokes in a Dounce homogenizer with a B pestle. After
removing nuclei and unlysed cells by centrifugation at 600 x g
for 5 min, crude meinbranes were sedimented at 5000 x g for
10 min, washed three times with the same buffer and suspended
in a sucrose solution in membrane buffer with a density of
1.18 g/ml. It was found that collecting the crude membranes
at 40000 x g for 20 min resulted in a higher yield and the same
protein pattern, but a lower enrichment in the membrane-
specific Na+/K+-stimulated ATPase. Exploiting their specific
density of 1.17 g/ml  membranes were collected after dis-
continuous-gradient centrifugation at 170000 x g for 2 h as
the interface between two sucrose solutions in membrane
buffer with densities 1.18 g/ml and 1.12 g/ml, respectively. The
purified membranes were diluted in membrane buffer, sedi-
mented for 2 h at 170000 x g and washed once with the same
buffer. Membranes prepared in this way showed an enrich-
ment of nine timcs in Na+/K +-stimulated ATPase.
The preparation of an actin filament fraction was based on
the resistance of actin filaments, intermediate-sized filaments
and nuclei against lysis in non-ionic detergents and on the
susceptibility of actin filaments to weakly ionic detergents such
as sodium deoxycholate. Cells were lysed in cytoskeleton buffer
(0.5% v/v Triton X-100, BDH Chemical Ltd, 75mM NaCl
in Tris/Mg/K buffer, 10" cells/ml) and centrifuged at 4000 x ,q
for 10 min or at 12000 x g for 5 min. The pellet, which con-
tained microfilaments, intermediate filaments and nuclei,
was washed three times with cytoskeleton buffer and incubated
briefly with 0.25 % (w/v) sodium deoxycholate in cytoskeleton
buffer. After centrifugation at 4000 x g for 10 inin or 12 000 x ,q
for 5 min, microfilament proteins and other deoxycholate-
soluble proteins were collected in the supernatant. It was
checked that although washing with cytoskeleton buffer
without deoxycholate also resulted in a preferential solubili-
zation of niicrofilanients, extraction with deoxycholate gives
a substantial, but not complete recovery of actin and other
deoxycholate-soluble proteins, and that most of the inter-
mediate-sized filament protein, vimentin, as well as the nuclear
proteins were resistant against extraction with deoxycholate.
The isolation of intermediate-sized filaments was largely
based on a method described by Franke et al. , employing
the unique resistance of these filaments against buffers con-
taining non-ionic detergents and 1.5 M KCl. We found that
contaminating microfilaments were lysed most efficiently by
a direct incubation of the cells in a buffer that contained a
(v/v) Nonidet P-40 , 2 'x (v/v) ampho-
non-ionic detergent, 1.5 M KC1 as well as 0.5 'i: (w/v) deoxy-
cholate. Thus, after lysing briefly 10" cclls/ml 0.5% (vk)
Triton X-100, 0.5'%; (w/v) sodium deoxycholate, 1.5 M KCl,
75 mM NaCl in Tris/Mg/K buffer, intermediate-sized fila-
ments were collected by centrifugation at 4000 x g for 10 min
or 12000 x g for 5 min and washed three times with the lysis
buffer. If higher concentrations of cells were lysed, it was
necessary to degrade the DNA by passing the viscous solution
repeatedly through a narrow needle and incubating with
0.04 mg/ml DNase I at room temperature.
The isolation of microtubular proteins was based on the
resistance of microtubules against non-ionic detergents at
room temperature in the presence of stabilizing agents and on
the specific sensitivity to low temperature and Ca2+ ions
(cf. [9,10]). Cells were lysed in tubulin buffer (0.5
X-100, 2 mM EGTA, 150 mM NaCI, 10% v/v dimethyl-
sulfoxide, 45 % v/v glycerol, 45 "/, Tris/Mg/K buffer, lo6 cellsj
ml) at room temperature. After centrifugation at 4000 x g for
10 min or 12000 x g for 5 min the pellet, which consisted of
microfilanients, intermediate filaments, microtubules and
nuclei, was washed three times with the same buffer, then ex-
tracted with 5 mM CaC12, 0.5% (v/v) Triton X-100, 75 rnM
NaCl in Tris/Mg/K buffer. After centrifugation at 4000 x g
for 10 min or 12000 x g for 5 min the microtubular proteins
were collected in the supernatant. It appeared essential to
include 150 mM NaCl in the tubulin buffer, since otherwise the
low ionic strength caused the coprecipitation of elongation
factor eEF-Tu, which subsequently was solubilized together
with, and in amounts about equal to, tubulin by the extraction
buffer of normal ionic strength (cf. Fig.3 from [lo]). The
same behaviour of eEF-Tu was observed with HeLa cells [I 51.
Polyribosomes were isolated essentially according a pro-
cedure of H. J. Dodemont (unpublished) from 20 x lo6 cells
labelled with 200 yCi [35S]methionine. All solutions and
glassware were made RNase-free by heat sterilization or
treatment with diethyl pyrocarbonate. Cells were lysed in 2 ml
0.25% (v/v) Nonidet P-40, 0.25% (w/v) sodium deoxy-
cholate in Tris/Mg/K buffer by 15 strokes in a Dounce homo-
genizer with a B pestle and centrifuged at 4000 x g for 30 min,
after which the supernatant was layered on 1-2 ml 2 M
sucrose in Tris/Mg/K buffer in a Ti50 centrifuge tube. After
centrifugation at 150000xg for 16 h, the pellet was washed
two times by gently shaking with water. This procedure
yielded polysomes with an A,60/A2,0 ratio of 1.8 or higher,
which were very active in a messenger-dependent cell-free
protein-synthesizing system. For our studies polysomes were
treated with pancreatic ribonuclease (1.25 pg in 25 pl, 15 min
at room temperature).
. Ribosomes and protein-biosynthesis factors were prepared
by extracting polysomes from 10 x 10' cells before RNase
treatment with 200 p 1 0.5 M NaCl in Tris/Mg/K buffer. After
centrifugation at 200000 x g for 1 h, the protein-synthesis
factors were collected in the supernatant and the ribosomes,
after washing with the same high-salt buffer, in the pellet.
Both fractions were treated with 0.5 pg RNase for 15 min at
Nuclei were prepared by lysing lo6 cells/ml nucleus buffer
(0.5% v/v Triton X-100, 0.5% w/v sodium deoxycholate in
50 "/, v/v Tris/Mg/K buffer). Small amounts of cells (5 x lo6)
were passed 5- 10 times through a needle with an internal
diameter of 0.3 mm. Larger amounts were lysed by up to
30 strokes in a Dounce homogenizer with a B pestle. Nuclei
from other cell lines may be prepared by less harsh procedures
, but the hypotonicity of the lysis buffer and the presence of
deoxycholate were essential for the removal of cytoskeletal
Fig. 1. Gel electrophoretic profiles ofjiactions isolated,fiom hamster lens epithelial cells. (A) Subcellular fractions; (B) subpolysomal fractions;
(C) subnuclear fractions. Proteins, labelled with [32S]methionine and separated by electrophoresis were detected by direct autoradiography ;
A, actin; G, eucaryotic elongation factor G (EF-2); H2B, H3 and €34, histones 2B, 3 and 4, respectively; T, c( and [j-tubulin; Tu, eukaryotic
elongation factor Tu (EF-Ia); V, vimentin
filaments (cf. [5,6]), while the presence of Triton X-100
prevented the lysis of the nuclei by deoxycholate. Nuclei were
collected by centrifugation at 600 x g for 5 min, suspended in
0.5 ml nucleus buffer and purified by centrifugation through
1 ml 1.5 M sucrose in Tris/Mg/K buffer (1.20 g/ml) at
12000 xg for 15 min. After resuspending the nuclear pellet
in 0.5 ml nuclear buffer by passing it repeatedly through nar-
row needles (internal diameter 0.6 mm, then 0.3 mm), this
procedure was repeated twice. The purified nuclei, which
were intact and free from cytoplasmic material as judged by
phase-contrast microscopy, were treated with DNase I
(I mg/mI, 10 p1/I06 nuclei) for I 5 min at room temperature.
Chromatin and non-chromatin fractions were prepared,
essentially as described , by centrifuging DNase-treated
nuclei through 1-2 ml 1 M sucrose in Tris/Mg/K buffer at
4000xg for 15 min, extracting the pellet with 0.4M
(NH4)2S04 in Tris/Mg/K buffer and centrifuging at 700 x g
for 5 min. The supernatant contained the salt-sensitive
chromatin proteins, while the pellet, after washing twice with
the extraction buffer, represented the non-chromatin fraction.
It was checked that no protein was liberated from the nuclei
by the DNase treatment.
Brain microtubules : tissue extracts
Hamster brain microtubules were purified by repeated
polymerization and depolymerization as described by She-
lanski et al. . Hamster erythrocytes were prepared as
described . Hamster liver, muscle and brain tissue were
extracted after mincing by saturation with urea and addition
of twice-concentrated SDS-containing sample buffer [l 81.
Liver tissue homogenate was, before the addition of urea,
treated with pancreatic DNase (50pg/ml, 20min at room
Analysix of’ the samples
Samples, isolated as pellets, were dissolved immediately in
SDS-containing lysis buffer [I 81; samples obtained as super-
natant were brought to 5 ”/, (w/v) trichloroacetic acid at
- 20 ”C and centrifuged, after which the pellet was washed
two times with acetone at - 20 “C, once with diethyl ether,
dried and dissolved in SDS-containing lysis buffer. All samples
were heated at 100°C for 3 min in this buffer; contrary to the
observation in  this did not result in any heterogeneity in
the first direction after two-dimensional gel electrophoresis.
One-dimensional gel electrophoresis in the presence of
SDS was carried out as described  in I-mm-thick slab gels
with a polyacrylamide gradient of 7- 18 ”/, (w/v). Two-
dimensional gel electrophoresis was carried out essentially as
described  after bringing the samples in SDS-containing
sample buffer to 9 M urea, 10
ampholines (LKB) in the pH range 3.5-10 and 5% (v/v)
2-mercaptoethanol. After isoelectric focusing in tubes with
an internal diameter of 2.8 mm with ampholines in the pH
range 3.5- 10, the gels were equilibrated for 30-60 min in
SDS-containing lysis buffer. If necessary, the gels were
stored at -20°C before or after the equilibration. In the
second dimension a gradient polyacrylamide slab gel (7 - 18
w/v, 1 mm thick) with a stacking gel of 1 cm was used.,
After electrophoresis, the proteins were fixed in 50 %
(v/v) methanol, 10% (v/v) acetic acid at 60°C for 30 min.
After rinsing briefly with water, the gels were either dried and
submitted to autoradiography or submitted to the fluoro-
graphy procedure .
(v/v) Nonidet P-40, 5 7;; (v/v)
Fig. 2. Tii,o-ditnc,n.cionaI gel puttum of'a totul-cell 1j.snte and .suhwlIulur and .~uhpol~~somul,fra~tionjrom
hamster lens epithelial cells. Proteins, labelled with [35S]methionine and separated
by isoelectric focusing (horizontal direction, the origin was at the left side) and SDS gel electrophoresis (vertical direction), were detected by fluorography. A, p and pactin:
Ann, a-actinin; Doc: sodium deoxycholate; eEF-G, eucaryotic elongation factor G (EF-2); eEF-Tu, eucaryotic elongation factor Tu (EF-IJ; eIF-4A, eucaryotic initiation
factor 4A; in eIF-3, comigrating with a component of rabbit reticulocyte initiation factor 3; i. migrating as, but not comigrating exactly with other components of eIF-3; m, major
iiicmbrane protein (see Fig. 1 a); n. typical nuclear protein; a-T. sc-tubulin; 1-T, /l-tubulin; Th, thermin, a widely distributed 68-kDa protein; V, vimentin
Gel electrophoretic analysis of the ,fractions
For the analysis of the protein contents of subcellular
fractions, one and two-dimensional polyacrylamide gel electro-
phoreses wcre used in a complementary fashion (Fig. 1 and 2,
respectively). Superior resolutions were achieved on two-
dimensional gels, but proteins with extreme isoelectric points
(e.g. the histones and most ribosomal proteins) were only
detected on one-dimensional gels.
The pattern of intense spots near and above the actin and
vimentin spots of the total-cell lysate (Fig. 2A) resembled the
patterns from other cultured cells [13,25 -281. The compo-
sition of the water-soluble fraction (Fig. 1 A, second lane;
Fig. 2 B) was relatively complex. As reported earlier for HeLa
cells [I51 the membrane fraction (Fig.lA, third lane) was
dominated by a protein of 35000 M,, which was detected as a
streaky spot in the basic part of the two-dimensional gel
(Fig.2C), indicated by the letter ‘m’. The same protein has
been found as main component in plasma membranes from
lens fiber cells (‘MP35’, ) and has been observed on gels
of plasma membranes from Chinese hamster ovary cells .
Extraction of the Triton X-100-insoluble proteins with the
weakly ionic detergent deoxycholate yielded a fraction
(Fig. 1 A, fourth lane, Fig.2D) that contained, besides the
microfilament proteins fl and pactin and cx-actinin, several
other proteins, some of which can be identified as polysomal
proteins (see below). a-Actinin was identified on the basis of
Fig. 3. Two-dimensional gel qf’u mixture o f Izumsrer brain nricrotubufes, &It.rected by staining with Coornassie hrilliunt blue, and a radioactively
labelled total-cell lysute, dererted by direct aurffra~jffgraphy.
A, actin; cc-T, a-tubulin; /3-T, /3-tubulin; Th, thermin (see Discussion)
literature data [27,31]. Lysis of the microfilaments and the
nuclei by a combination of Triton X-100, deoxycholate and
1.5 M KCI resulted in almost pure vimentin (Fig. 1 A, fifth
lane, Fig.2E), reflecting the fact that in epithelial lens cells
intermediate-sized filaments are exclusively of the vimentin
type . Extraction of HeLa cells under the same conditions
yielded, in addition to vimentin, also the prekeratin subunits
. Similar results have been reported for desmin ,
showing that the resistance against non-ionic or weakly ionic
detergents and high salt concentrations is a general feature of
Solubilization of the microtubules by Ca2+ ions in a
Triton-X-100-insoluble residue from cells, lysed under micro-
tubule-stabilizing conditions, resulted in a fraction (Fig. 1 A,
sixth lane) that contained, besides actin, M and p-tubulin,
migrating as one fuzzy band. On the two-dimensional gel
(Fig. 2F) c( and /i’-tubulin were separated, the former, under
our conditions, migrating to virtually the same position as
vimentin (see next section). During coelectrophoresis of the
cellular proteins with purified brain microtubules (Fig. 3) the
brain tubulins displace all cellular proteins, showing that
tubulins from brain and the cell line are electrophoretically
not identical. However, two other cellular proteins, marked
‘A’ (cytoplasmic actin) and ‘Th’ respectively, comigrate
completely with components of brain microtubules. These
proteins are also present in the microtubular fraction (Fig. 2F)
and correspond to NQ-MAP1 and NQ-MAP4, respectively
(‘non-quantitative microtubule-associated proteins’) observed
on two-dimensional gels of bovine brain microtubules .
Further, the 68-kDa ‘Th’ spot and the nearby 65-kDa spots in
Fig. 2F may correspond to the 68-kDa and 66-kDa proteins
found in HeLa microtubules .
The profile of the polysomal fraction (Fig. 1 A, seventh
lane) consisted of a number of bands in the molecular-weight
range of 14000 to 35000 of about equal intensities, reflecting
the equimolar protein composition of ribosomes. Extraction of
polysomes with 0.5 M NaCl resulted in a separation of the
ribosomes from a fraction with a relatively simple compo-
sition (Fig. 1 B), containing the initiation and elongation fac-
tors of protein synthesis , some of which could be identi-
fied (see ). Owing to the alkaline isoelectric points of
most ribosomal proteins, the two-dimensional pattern of the
polysomes (Fig. 2G, resembling strongly the pattern of
hamster reticulocyte polysomes; unpublished results), re-
presented only part of the polysomal proteins. Most of the
spots belong to either cytoskeletal proteins, also present in the
ribosomal fraction (Fig. 2H), or to the protein biosynthesis
factors (Fig. 2 I). The ribosomal protein pattern (Fig. 2 H)
further displayed some streaky spots, which were not present
in the polysomal protein pattern. Exactly the same was
observed with HeLa polysomes and ribosomes [I51 and was
presumably due to a specific aggregation in the polysomal
sample preventing the penetration into the gel of some of the
The nuclear fraction (Fig. 1 A, eighth lane) contained the
histones (H2B, H3 and H4 were detected; HI and H2A do not
contain methionine) and a number of proteins with higher
molecular weights. As shown in Fig. 1 C, subfractionation of
nuclei with a buffer of high ionic strength [0.4 M (NH4)2S04]
yielded a chromatin fraction, containing the histones, and a
salt-resistant non-chromatin fraction, containing no histones
but most nuclear proteins with higher molecular weights. As
a consequence of the extreme isoelectric points of nuclear
proteins, the two-dimensional gel pattern of this fraction
(Fig. 2 J) revealed only few typical nuclear proteins, indicated
by the letter ‘n’. Possibly, the intense 68-kDa protein cor-
responds to one of the nuclear lamina proteins . However,
in spite of the clearly different and complementary protein
compositions of the chromatin and non-chromatin fractions
(Fig. 1 C), two-dimensional gel electrophoresis of these frac-
tions (not shown) did not enable us to assign the typical
nuclear spots unambiguously to one of the nuclear subfrac-
tions (cf. [14,38]). From the intensities of the spots relative to
the amount of radioactivity applied and the exposure time, it
was apparent that the rather aspecific patterns corresponded
to only a small part of the proteins in the nuclear fraction and
subfractions. Conceivably other techniques, that can detect
proteins with more extreme isoelectric points, might be more
helpful for the characterization of these fractions.
Electrophoretic behaviour of’ the tubulins
On gels of other cell lines [13,25 - 281 the relative positions
of the intense spots are very similar to the patterns shown in
Fig. 2 A. However, a detailed analysis revealed a discrepancy
Fig. 4. T\vo-dimensionul gel of u mixture oJ‘ unluhelled vimentin,
isolated jiom 17 x lo6 c~t4l.s and detcctc~cl hj, stuining bvitli Coomassir
huilliant blue. and u rudioactivcly luhclleti microtuhulc prepurution
detected by direct autorarliogrup~i~..
V, vimentin. r-Tubulin has been displaced by the large amount of
a-T, a-tubulin; P-T, fi-tubulin;
regarding the position of the two spots belonging to a and
j-tubulin relative to other intense spots: in our gels a-tubulin
is at about the same position as vimentin, while in other reports
[26 - 281 both tubulins are detected above vimentin.
To detcrrnine more accurately the relativc positions of the
tubulins and vimentin under our conditions, unlabelled
vimentin and a radioactive microtubular fractions were
analyzed on the same gel (Fig.4). In accordance with the
observation of O’Farrell [I 91, that proteins that are over-
loaded on two-dimensional gels displace nearby proteins, the
radioactive a-tubulin forms a ‘halo’ left of the vimentin spot,
while p-tubulin is on a distinctly lower position.
The discrepancy between the data from the literature
126-281 and our results may be explained by the highly
irregular behaviour of a and j-tubulin in SDS-containing
gels. In one-dimensional gels, their aberrant mobilities (ap-
parent M , of 54000 and 55000, respectively versus true 1l4~ of
50000 ) is modulated by the presence of tetradecyl sulfate
or heating in urea . Apparently the conditions of two-
dimensional gel electrophoresis affect the mobilities of the
tubulins, too, resulting in apparent molecular weights of
57000 and 5.5000 (Fig.2F and 4) or even higher values
[26 - 281.
Comparison of cultured cells with extructs
, fiom erythrocytes und .speciulized tissues
Although the cell line used in this study was derived from
hamster lens epithelium, the pattern of the total protein
population resembles the pattern of other cultured cells
[25 - 281 rather than the protein pattern of lens homogenates
. To learn whether the major proteins from cultured cells
are present in other specialized cells, we analyzed the proteins
from hamster erythrocytes and from brain, liver and muscle
extracts by two-dimensional gel dectrophoresis as shown in
Fig. 5. Arrows indicate proteins that are also present in cultured
lens cells, as determined by separated coelectrophoresis
experiments with labelled total-cell lysate (cf. Fig. 3). It can
be concludcd that erythrocytes, specialized tissues and cul-
tured cells have only few proteins in common. In fact, only
two of the major proteins can be found in all five cell types,
namely actin (in muscle as the a-actin variant) and, strikingly,
a 68-kDa protein indicated by ‘Th’, which is also found in
most subcellular fractions. This protein was also found in
appreciable quantities in reticulocyte polysomcs (not shown).
Topography q f ’ cellular proteins
It has been shown that by fractionation procedures, based
mainly on independent chemical extractions, subcellular
fractions can be isolated which have characteristic protein
compositions (Fig. 1 and 2). Fig. 6 summarizes the intracellular
localizations of the proteins detected on two-dimensional gels
of the total-cell lysate. Aside from numerous, scarcely re-
producible faint spots (some of which may represent degra-
dation products), only one of the intense spots, marked with
an asterisk, could not be found in any of the subcellular
fractions. This indicates that our fractionation procedures
account for nearly all cellular proteins.
The membrane fraction has been isolated as the 1.18-
1 .I2 g/ml interface by ccll lysis in hypotonic buffer, removal
of the nuclei and discontinuous-gradient centrifugation. An
enrichment in the Na’, K+-ATPase and the predominant
presence of‘ a 35-kDa component, identical to a lenticular
plasma membrane component (MP35 ), confirmed that
this procedure, which does not deviate essentially from
published procedures, yielded a membrane preparation. How-
ever, since no attempt was made to separate plasma mem-
branes from, for example, mitochondria1 membrane fragments
or smooth endoplasmatic reticulum, our procedures as well
as those commonly used do not allow the assignment of the
membrane protein spots in Fig. 6 to specific membrane
Several reports have described the extraction of cells with
non-ionic detergents [7,26,45], resulting in a detergent-
resistant structure consisting of cytoskeletal filaments, as-
sociated with polyribosomes [46,47] and nuclei. We checked
that extraction of this structure with dcoxycholate solubilized
all proteins that did not belong to either the intermediate-
filament fraction (resistant against deoxycholate and 1.5 M
KCl) or the nuclear fraction (resistant against deoxycholate
and purified by shearing and density centrifugation), permit-
ting a complete fractionation of the ‘Triton X-100 pellet’ 
in three discrete components.
Conversely, most proteins arc typical for only one or two
subcellular fractions. However, the major cytoskeletal pro-
teins actin, vimentin, a-tubulin, p-tubulin and the 68-kDa
protein ‘Th’ occur in most fractions. In addition, the poly-
somal, microtubular and nuclear fractions have common
spots, most of which are also in the soluble (s) or deoxycholate-
soluble (a) fraction. The most likely explanation for these
common spots may be the copurification of polysoinal pro-
teins, in particular some of the protein-synthesis factors,
with the nuclear, microtubular and deoxycholate-soluble
proteins, respectively. Likcwise, without evidence 10 the
contrary, the occurrence of actin, vimentin and tubulin in most
tractions may also be considered as typical and persistent
cross-contaminations, to be expected of these abundant and
On the other hand, there is ample evidence of interactions
of actin and vimentin and membrane structures in v i w
[7,48 - 511. Further, Nelson and Traub [52,53] found that vi-
mentin from mouse ascites cells comigrated in sucrose gra-
Fig. 5. Two-dimensional gel patterns qf’ hamstpr eryt11rocyte.r. and tjss~e cwracts after stajnjnx lc,;t/l Coomus.yie brilliant blur. Arrows indicate spots that are also present in cultured
hamster lens cells, as determined by separate coelectrophoresis experiments with radioactive labelled total-cell lysate. A, actin: CA, erythrocyte carbonic anhydrase  ; CK, muscle
creatine kinase : CKe8 and CKel8, cytokeratin 8 and 18 respectively [43,44]; D, desmin; f, in cultured cells present in the protein-biospnthesis factors fraction: HSA. hamster
Serum albumin; m, in cultured cells present in the membrane fraction: MLC1, MLC:! and MLC3, myosin light chain 1, 2 and 3 respectively : a-T and 8-7, brain
respectively: Th, thermin: a-TM and B-TM, muscle a and /l-tropomyosin respectively  : V, vimentin
Fig. 6. Schematic representation of the two-dimensional gel of humstrr lens epithelial cells shown in Fix. 2A. The blackness of a spot is a measure
of its intensity on the original fluorograph. a, Present in the actin-filament (deoxycholate-soluble) fraction: ct-Ann, r-actinin: cEF-G, eucaryotic
elongation factor G (EF-2); eEF-Tu, eucaryotic elongation factor Tu (EF-la); f, present in thc protein-synthesis-factors fraction; in eIF-3,
comigrating with a component of rabbit reticulocyte initiation factor 3: eIF-4A, eucaryotic initiation factor 4A: m, membrane proteins;
n, present in the nuclear fraction; r, present in the ribosomal fraction; s, water-soluble protein; t, present in the microtubular fraction; Th,
thermin; a-Tub, a-tubulin; P-Tub, /I-tubulin; Vim, vimentin. The subcellular localizations of major protein that are present in most fractions
(the actins, the tnbulins, vimentin, thermin) have not been indicated
dients with the ribosomal subunits. On electron micrographs
[14,51] an attachment of intermediate filaments to nuclei was
observed, which also is in agreement with our fractionation
studies. Evidently subcellular fractionation may give a clue
to the interactions of subcellular structures in vivo. However,
other interactions may be lost during the lysis and extraction
procedures. This latter caveat may apply to the rather harsh
isolation of intermediate filaments.
The fractionation procedures described here may be useful
for the localization of specific molecules, e.g. viral antigens.
In this respect the localizations of the src gene product of the
Rous sarcoma virus in the membrane  and in thc cyto-
skeleton I551 or transformed chicken embryo fibroblasts
stress the importance of more complete fractionations to
localize more accurately this antigen. Further, since two-
dimensional gels are used extensively to study processes as
differentiation  or transformation [26,57], the topography
shown in Fig. 6 would help considerably in interpreting, or at
least localizing, effects of altered gene expression.
The abundant 68-kDa protein
Next to the major cytoskeletal proteins actin, vimentin and
s ( and P-tubulin, Fig.2 reveals another protein, marked ‘Th’,
that is present in the cell in about the same quantity and the
same wide distribution over the subcellular fractions. In fact,
coelectrophoresis with extracts from erythrocytes or liver,
muscle and brain tissue (Fig. 5) as well as literature data about
various cell lines (e.g. see [13,25 -28,56,57]), show that this
protein has the same wide distribution in vertebrate cells and
tissues as the cytoplasmic actins, and an even broader oc-
currence than the intermediate-filament proteins. It has been
identified as pyruvate kinase in erythrocyte extracts , but
this enzyme was located much more to the acidic side on gels
of muscle extracts .
Another clue to its identification may be based on its
occurrence in brain microtubular preparations (Fig. 3 ; ).
Strikingly, this microtubule-associated proteins is, both by
two-dimensional gel electrophoresis and by peptide mapping,
indistinguishkble from a heat-shock protein, designated
‘thermin’, that is highly conserved in vertebrate species and
tissues . Therefore, on the base of its position on two-
dimensional gels (see also ) and its occurrence in micro-
tubular preparations, we identify the 68-kDa protein found in
our fractions as thermin.
Anderson et al.  found that the synthesis of a protein
with about the same molecular weight (‘HS:I’, probably
corresponding to ‘hsp66’ from ) is increased more by heat
423 Download full-text
shock than the synthesis of thermin, so the relation between
thermin and the heat-shock response remains to be clarified
Wang et al.  proposed an association between thermin anti
intermediate filaments. Evidently its high concentration in all
vertebrate cells warrants further studies of its function and
interactions with subcellular ytructures.
J. A. L. acknowledges a fellowship of the Dutch Cancer Socicty
(Koningin Willielmina Fonds).
Hogeboom, G. H. (1955) Methods Enzymol. I , 16-19.
Fleischer, S. & Kervina, M. (1974) Metlzods Enzymol. 31, 6-41.
Reid, E. & Williamson, R. (1974) Methods Enzymol. 31, 713-
de Duve, C. (1980) Harvey Lect. 59, 49-87.
Penman, S. (1966) J. Mol. Biol. 17, 117-130.
Herman, R., Weymouth, L. & Penman, S. (1978) J. Cell Biol. 78,
663 - 674.
Franke, W. W., Weber, K., Osborn, M., Schmid, E. & Freuden-
stein, C. (1978) ESP. Cell Re.r. 116, 429-445.
Franke, W. W., Schmid, E., Weber, K. & Osborn, M. (1979)
Exp. Cell Res. 118, 95 - 109.
Rubin, R. W., Warren, R. H. & Leonardi, C. L. (1979) J. Cell
Bid. 83, pt 2, 353.
Solomon, R., Magendantz, M. & Salzman, A. (1979) Cell, 18,
Franke, W. W., Mayer, D., Schmid, E., Denk, H. & Boren-
freund, E. (1981) E,xp. Cell Res. 134, 345-365.
Smith, G. J. (1981) Anal. Biochem. 111, 97- 104.
Bravo, R., Small, J. V., Fey, S. J., Larsen, R. P. & Celis, J. E.
(1982) J. Mol. Bid. 154, 121 - 143.
Capco, D. G., Wan, K. M. & Penman, S. (1982) Cell, 29.
847 - 858.
Lenstra, J. A. & Bloemendal, H. (1983) Eur. J. Riorhem. 130,
41 9 - 426.
Bloemendal, H., Lenstra, J. A,, Dodemont, H. J., Ramaekers,
F. C. S., Groeneveld, A. A,, Dunia, I. & Benedetti, E. L. (1980)
EX^. Eye R ~ s .
Wang, D. & Moore, D. (1978) J. B i o l Chern. 153, 7216-7219.
Laemmli, U. K. & Favre, M. (1973) J. Mol. Biol. 80, 575-599.
O’Farrell, P. H. (1975) J. Bid. Chem. 250, 4007-4021.
O’Farrell, P. Z. & Goodman, H. H. (1976) Cell, 9, 289-298.
Shelanski, M. L., Gaskin, F. & Cantor, C. R. (1973) Proc. Narl
Acad. Sci. USA, 70, 765 - 768.
Edwards, J. J., Anderson, N. G., Nance, S. L. & Anderson,
N. L. (1979) Blood, 53, 1121-1132.
Wilson, D. L., Hall, M. E., Stone, G. C. & Rubin, R. W. (1977)
Anal. Biochem. 83, 33 -44.
Bonner, W. M. & Laskey, R. A. (1974) Eur. J. Biocliem. 46,
83 - 88.
O’Farrell, P. Z. & Goodman, H. H. & O’Farrell, P. H. (1977)
Cell, 12, 1133-1142.
Bravo, R. & Celis, J. E. (1980) Exp. Cell Res. 127, 249-260.
Bravo, R. & Celis, J. E. (1982) Clin. Chem. 28, 766-782.
Fey, S. J., Bravo, R., Larsen, P. M., Bellatin, J. & Celis, J. E.
(1981) Cell Biol. Int. Rep. 5, 491 -500.
29. Lenstra. J. A., Van Raaij, A. J. M. & Bloemendal, H. (1982)
FEBS Lett. 148, 263-266.
30. Horst, M. N., Baumbach, G. & Roberts, R. M. (1979) FEBS
Let/. IOU, 385 - 388.
31. Wang, C., Asai, D. J. & Lazarides, E. (1980) Proc. Nut1 Acad.
Sci. USA, 77, 1541 - 1545.
32. Ramaekers, F. C. S., Osborn, M., Schmid, E., Weber, K., Bloe-
mendal, H. & Franke, W. W. (1980) Exp. Cell Res. 127,
33. Gard, D. L., Bell, P. B. & Lazarides, E. (1979) Pvoc. Narl Acad.
Sci. USA, 76, 3894- 3898.
34. Berkowitz, S. A,, Katagiri, J., Binder, H. K. & Williams, R.
C., Jr (1977) Biochemistry, 16, 5610-5617.
35. Weatherbee, J. A., Luftig, R. B. & Weihing, R. R. (1980) Bio-
chemistry, 19, 41 16-4123.
36. Hershey, J. W. B. (1980) Cell Bid. 4, 1-68.
37. Gerace, L. & Blobel, G. (1980) Cell, 19, 277-287.
38. Peters, K. E., Okada, T. A. & Comings, D. E. (1982) Eur. J.
Biochem. 129, 221 -232.
39. Fulton, C. (1982) Nature (Lond.) 296, 308-309.
40. Best, D., Warr, P. J. & Gull, K. (1981) Anal. Biocliem. 114,
41. Lenstra, J. A., Hukkelhoven, M. W. A. C., Groeneveld, A. A,,
Smits, R. A. M. M., Weterings, P. J. J. M. & Bloemendal, H.
(1982) Exp. Eye Res. 35, 549-554.
42. Giometti, C. S., Anderson, N. G. & Anderson, N. L. (1979)
C l i n . Chem. 25, 1877 - 1884.
43. Franke, W. W., Schiller, D. L., Moll, R., Winter, S., Schinid,
E., Engelbrecht, I., Denk, H., Krepler, R. & Platzer, B. (1981)
J. Mol. Biol. 153, 933-959.
44. Moll, R., Franke, W. W., Schiller, D. L., Geiger, B. & Krepler,
R. (1982) Cell, 31, 11 -24.
45. Ben-Ze’ev, A,, Duerr, A., Solomon, F. & Penman, S. (1977)
Cell, 17, 855-8865.
46. Lenk, R., Ransom, L., Kaufman, Y. & Penman, S. (1977) Cell,
10, 67- 78.
47. van Venrooij, W. J., Sillekens, P. T. G., van Eekelen, C. A. G.
& Reinders, R. J. (1981) ESP. Cell Res. 135, 79-91.
48. Lazarides, E. (1980) Nature (Lond.) 283, 249 - 256.
49. Bloemendal, H., Benedetti, E. L., Ramaekers, F. & Dunia, I.
(1981) Mol. Biol. Rep. 7, 167-168.
50. David-Ferreira, F. L. & David-Ferreira, K. L. (1981) Eur. J.
Cell Biol. 22, 376.
51. Woodcock, L. C. F. (1980) J. Cell Biol. 85, 881-889.
52. Nelson, W. J. & Traub, P. (1981) Eur. J. Cell Bid. 23,250-257.
53. Traub, P. & Nelson, W. J. (1981) J. Cell Biol. 91, 232a.
54. Courtneidge, S. A., Levinson, A. D. & Bishop, J. M. (1980) Proc.
Nafl Acad. Sci. USA, 77, 3783-3787.
55. Burr, J. G., Dreyfuss, G., Penman, S. & Buchanan, J. M. (1980)
Proc. Nut1 Acad. Sci. USA, 77, 3484- 3488.
56. Sidhu,R.S.(1979)J. Biol. Clzem.254, 11111-11118.
57. Strand, M. & August, J. T. (1978) Cell, 13, 399-408.
58. Wang, C., Gomer, R. H. & Lazarides, E. (1981) Proc,. .h‘rrtl .4cud,
Sci. USA, 78, 3531 -3535.
59. Hickey, E. D. & Weber, L. A. (1982) Biochwnistrj.. 2 1 ~ 1513-
60. Anderson, N. L., Giometti, C. S., Gemmell, M. A,, Nance, S. L.
& Anderson, N. G. (1982) Clin. Chem. 28, 1084-1092.
J. A. Lenstra and H. Bloemendal, Laboratorium voor Biochemie, Katholieke Universiteit Nijmegen,
Geert Grooteplein Noord 21, 6525 EZ, Nijmegen, The Netherlands