Domains of rough endoplasmic reticulum (a review)

Molecular and CellularBiochemistry 87: 93-103, 1989. © 1989KluwerAcademic Publishers. Printedin the Netherlands. Review

Domains of rough endoplasmic reticulum (a review) Ian F. Pryme Department of Biochemistry, University of Bergen, Arstadveien 19, N-5009 Bergen, Norway Received15 July 1988; accepted20 December1988

Key words: endoplasmic reticulum, membranes, ribosomes, messenger RNA, cytoskeleton Abstract

The rough endoplasmic reticulum isolated from several eukaryotic cell lines can be separated into subfractions. These subfractions possess different properties indicating that they represent separate domains of the endoplasmic reticulum system.

Introduction

The endoplasmic reticulum (ER) of eukaryotic cells can be classically separated into two domains, the rough endoplasmic reticulum (RER) and the smooth endoplasmic reticulum (SER). The R E R and SER are both morphologically and biochemically quite distinct. The major morphologic difference between the two compartments of the ER system is the occurrence of ribosomes on the cytosolic surface of the R E R membranes while SER membranes are essentially free of such particles [1, 2]. Biochemically the two types of ER membranes differ in a variety of respects [3], for example the cholesterol content of SER is higher than that in RER, while the dolichol phosphate concentration is high in R E R but low in SER; R E R membranes have a higher net negative surface charge than SER membranes; the affinity of R E R for monovalent cations is much greater than that of SER membranes. The ratio between the respective amounts of R E R and SER is quite different from cell type to cell type and this variation is coupled to the primary function that individual sets of cells perform [4]. Functionally the RER, in contrast to the SER, is responsible for the synthesis of protein molecules on membrane-bound polysomes (see references in [3, 5-71).

Various pieces of evidence suggest that there is a compartmentalization of rough ER membranes into functional domains. Eriksson [8] has separated the rough microsomal membranes of rat liver into three subfractions (RI, RII, and Rill). These fractions differed with respect to number of membrane-bound ribosomes associated with the vesicles as visualized in two-dimensional pictures; the RI, RII and RIII fractions contained 5-15, 3-8 and 1-5 ribosomes/vesicle respectively. The subfractions also differed in other respects, for example, RNA, cholesterol and phospholipid content. In comparison to the ribosome-rich subfractions (RI and RII) and smooth microsomes, the Rill fraction was shown to be enriched in electron transport enzymes. Leskes et al. [9], however, using a cytochemical technique to localize glucose-6-phosphatase, a marker enzyme for RER, were unable to demonstrate any form of specific localization of this particular enzyme in R E R compartments. Using a freeze-etching technique to visualize the interior of cellular membranes [10] a structural differentiation has been observed between those rough ER membranes that are found in close association with mitochondria and those that are not [11].

94 Mueckler and Pitot [12] have provided evidence for the existence of distinct populations of membranebound polysomes. Although mRNA isolated from two tightly-bound fractions, when translated in vitro, produced identical products, individual polypeptides were synthesized in different quantities. Furthermore, the mRNA isolated from a looselybound polysome preparation synthesized several large molecular weight polypeptides which were not detected as in vitro products when mRNA from free or tightly-bound polysomes was utilized. The proteins synthesized on R E R in non-infected cells can be basically divided into two groups: various membrane proteins ultimately destined to become integrated into other cellular membrane systems, and secretory proteins. In virus-infected cells some proteins coded for by the viral genome are synthesized on R E R [13]. It is also apparent that some proteins initially synthesized on R E R are later released to the cytosolic compartment. It is now well known that membrane proteins can have quite different orientations in the membrane. For instance a membrane protein can span the lipid bilayer either once (monotopic), twice (bitopic) or several times (polytopic) [14]. The N-terminal or C-terminal end may be exposed at the cytosolic face or at the luminal side of the membrane. Some proteins are integrated into membranes such that both the N-terminus and C-terminus are exposed at the ectoplasmic side while others are orientated such that both N-, and C-termini are directed toward the lumenal space. Membrane-spanning polypeptides are inserted into R E R membranes by the process of co-translational translocation and their orientation within the membrane is presumably not later altered. The membrane proteins then diffuse in the lateral plane of the R E R toward the SER where segregation and sorting occurs, or new membrane compartments may be constructed by a combination of budding and fusion processes. Unlike membrane proteins, those proteins which are ultimately destined to be secreted from the cell pass through the entire lipid bilayer, are sequestered in the lumen of the RER, and then transported into the SER before being directed further along the secretory pathway (Golgi apparatus, secretory vesicles etc.). It is now clear that a membrane protein

must contain within its structure information, or signals, which define the final destination and properties of the molecule. An early signal in the nascent polypeptide chain is required for directing the sequence to a membrane where translocation of the polypeptide across the lipid bilayer can be accomplished. The next signal in the polypeptide chain defines the manner of incorporation into the membrane and a third would have the property as a sorter being responsible for directing the polypeptide to a so-called receiver, precursor membrane. In some instances a further signal may be required in order to assure the correct positioning or anchorage of a protein in a membrane. These signals are coded for in the gene of each individual polypeptide thus the protein molecule itself contains the necessary information required for deciding its ultimate destination. Regarding the first signal, this sequence is the first part of the mRNA molecule which is translated. Initiation of the synthesis of secretory and membrane proteins occurs on ribosomes not attached to the ER i.e. on so-called free polysomes. Codons adjacent to the initiator codon A U G code for the signal peptide. When this peptide (i.e. the signal sequence), a nascent chain 1535 amino acids in length has emerged from the 60S ribosome, binding to the R E R membrane surface is facilitated [15]. The signal recognition particle and docking protein ensure that translocation of the polypeptide chain into the ER membrane can then occur [16, 17]. Signal peptides from a great number of proteins have been sequenced and although there is no apparent sequence homology they all have a common feature in the presence of an uninterrupted stretch of hydrophobic amino acid residues of a minimum length 6--7 residues in the center of the peptide. The fact that E R membranes from a variety of sources can recognize signal sequences from unrelated species favours the existence of a single, ubiquitous receptor mechanism. The question arises, however, whether or not a mechanism exists for directing a nascent polypeptide chain to a specific domain of the R E R where translation followed by post-translational modification of the polypeptide chain can occur. Thus, does the R E R consist of a series of separate areas possessing pre-determined properties and enzyme

95 systems, each being 'reserved' for the synthesis of specific sets of proteins, or is the RER system entirely homogeneous without any division into sub-populations? If the latter situation is the one which prevails then there must be a 'free for all' among mRNA species with translation occurring on a 'first-come first-served' basis. This situation, which obviously incurs disorder, would be the simplest to explain in biochemical terms. The existence of individual RER compartments each having its own defined functions is an attractive hypothesis but would require an extremely high degree of organization. In this article accumulating evidence in favour of the existence of distinct domains of the RER is summarized.

Heavy rough and light rough endoplasmic reticulum in mouse myeloma cells Melchers [18] showed that ER membranes of MOPC-46 cells obtained from solid tumours could be separated into three fractions by discontinuous sucrose gradient centrifugation. When subjected to electron microscopy two of the fractions were shown to contain RER and the third SER. Membranes in one of the RER subfractions passed through 2.0M sucrose during ultracentrifugation and accumulated above a cushion of 2.3 M sucrose. Rough membranes of the latter type were identified in gradients following disruption of cells by Potter Elvehjem homogenization or ultrasonication either by optimal density (260 nm) measurements or by radioactivity incorporated into protein, but not after disruption in a Dounce-type tissue homogenizer. The two RER subfractions have also been identified in MPC-11 cells grown in suspension culture where identical buffer and gradient conditions were used [19]. In these experiments, however, cell disruption was performed by nitrogen cavitation. In order to distinguish between the two RER subfractions, membranes which accumulated above 2.0 and 2.3M sucrose were termed light rough (LR) and heavy rough (HR) respectively. It was observed that the yield of ER is not 100% following cell disruption by nitrogen cavitation, further RER

remaining attached to the nuclei following the homogenization procedure. This RER, which can be solubilized by treatment with detergent [19], has been termed the nuclear-associated ER (NER). Polysome profiles obtained from solubilized LR and HR membranes, NER and the cytosol (free polysomes) were quite different in appearance [20] and substantial differences between amounts of polysomes/monosomes were observed (Table 1). RNA/phospholipid ratios in the HR and LR subfractions were 0.44 and 0.38 respectively. Based on this evidence and the difference in polysome profiles it would seem likely that there is a higher density of packing of ribosomes per unit membrane in the HR than in the LR subfraction. The specific activities of UDPase and 5'nucleotidase in LR membranes were twice as high as those in HR membranes [20].

RER subfractions in other transformed cells and normal cells In addition to myeloma cells the presence of HR and LR ER membranes has been established in L-929 cells and HeLa cells grown in suspension culture, and Krebs II ascites cells propagated intraperitoneally in mice and then incubated further in culture medium [21-23]. Only the LR subfraction was identified in confluent cultures of mouse embryo C3H/10T1/2 cells. However, incubation for 24 h in the presence of the tumour promoter TPA resulted in the appearance of the HR subfraction [23]. Two possible explanations have been Table 1.

Polysomesand monosomesin variouscellfractions

Fraction

Cytosol HR LR NER

% Distribution Polysomes

Monosomes

74 84 82 71

26 16 18 29

Percentage distribution of A260 in sucrose density gradient profiles obtainedfromvarious fractionsisolatedfromMPC-1] cells. Data fromref. [20].

96 forwarded in an attempt to account for these observations: a) the tumour promoter may have resulted in the synthesis of new proteins not originally synthesized, these specific proteins being synthesized on a 'new' class of R E R membranes, i.e. H R membranes, b) the tumour promoter may have caused a reorganization of the manner by which ribosomes are attached to rough ER membranes resulting in the production of R E R of higher density. A variety of normal mouse tissues have been tested for the presence of H R and LR membranes: liver, spleen, kidney, heart, lung and bone marrow. Only the LR subfraction was identified in these tissues [21, 23]. The production of H R membranes was not stimulated either by the incubation of rat liver slices in culture medium for 2 h, or by in vivo stimulation of the rat pancreas by a combined injection of secretin and cholecystokinin/pancreozymin [21]. One can therefore draw the general conclusion from these observations that the H R subfraction of the R E R represents a specific population of membranes that is peculiar to either transformed cells or normal cells treated with a tumour promoter. Shore and Tara [24, 25] have isolated two subfractions of R E R from rat liver: rapidly sedimenting E R (RSER) and R E R of the rough microsomal fraction. When mRNA was isolated from these subfractions and the products of translation in a heterologous system were analysed by immunoprecipitation, then it was found that albumin mRNA was distributed about equally between the two fractions of R E R while that coding for mitochondrial proteins was enriched in membranes of the rough microsomes. Meier et al. [26, 27] isolated a subfraction of R E R from crude nuclear fractions of rat liver homogenates and these membranes were shown to be closely associated with mitochondria (mito-ER complexes). Their data suggest that mito-ER complexes play an important role in the synthesis and assembly of cytochrome P450. In light of these results it has been suggested that compartmentalization of cytoplasmic synthesis of mitochondrial proteins on R E R membranes closely apposed to mitochondria could well be an important aspect of the mechanism whereby the endomembrahe system differentiates between polypeptides destined for transport to different cellular loci.

Stuhne-Sekalec and Senacev [28] have studied the biosynthesis of microsomal lipids in RSER and microsomes in rat liver. Their results show convincingly that RSER contains the key enzyme system for the formation of phospholipids. The amount of phosphatidic acid found in purified RSER was considerably lower than that synthesized by microsoreal membranes, while on the other hand the opposite was the case for the synthesis of neutral lipids. These results indicate the presence of a more active phosphatidate phosphohydrolase in RSER than in microsomes. Furthermore, the formation of CDPcholine by purified RSER and its diglyceride-dependent conversion to phosphatidylcholine was significantly higher in RSER than in microsomal membranes. The RSER was also shown to be involved in the synthesis of lipids which were ultimately imported into mitochondria. Hazel and Zerba [29] have studied light and heavy microsomal membranes isolated from the liver of thermally acclimated (5 and 20 ° C) rainbow trout. Although acclimation had no effect on the distribution of marker enzymes in the two fractions, significant differences were observed between specific activities in light and heavy microsomes. Contrary to findings in rat liver the heavy microsomal fraction contained the bulk of glucose-6-phosphatase activity. It was also very rich in N A D P H cytochrome C reductase activity but contained only small amounts of 5'-nucleotidase. The light microsomal fraction possessed significant levels of N A D P H cytochrome C reductase activity, little glucose-6-phosphatase but was enriched in 5'-nucleotidase. Concerning phospholipid composition, the heavy microsomal membranes appeared to be influenced to a lower degree than light microsomes under conditions of thermal acclimation. It was observed at 5 °, in the light microsomal fraction particularly, that long-chain polyunsaturated molecular species replaced the shorter chain monoand dienoic, and occasionally saturated species, characteristic of growth at 20°. Alterations in both phosphatidylcholine and phosphatidyl-ethanolamine were highly significant. Results from both marker enzymes and phospholipid content thus indicate that the light and heavy microsomes contain different species of R E R membranes.

97 Pathak et al. [30] using UT-l-cw cells, have shown by immunological procedures that H M G CoA reductase is not associated with R E R which is free in the cell but instead with R E R of the outer nuclear envelope. They thus speculate that this latter region is specialized for the synthesis of this membrane protein. These workers suggest, furthermore, that the sorting of membrane proteins destined for specific membranous cellular compartments may occur already at the level of synthesis and insertion into ER. Recent experiments have shown that labeled phosphatidylcholine, initially synthesized in the NER, migrates to the plasma membrane within a 2 h chase [31]. Other evidence including the synthesis of glycoproteins in N E R [32], the stimulation of incorporation of 3Hcholine label into N E R by styrene oxide [33] and a TPA-stimulated release of labeled phospholipids from NER [34] indicate that important events in membrane biogenesis occur in close proximity to the nucleus (for review on NER, see ref. [35]).

RER subfractions and the cell cycle

It has been established that there is a correlation between cell cycle and comparative amounts of H R and LR membranes in MPC-11 cells [36]. The type of variation in amount of individual E R membrane subfractions in relation to phase of cell cycle is shown in Table 2. A series of experiments showed that in cell cultures with a high S/G2 + M ratio then H R membranes were more predominant than LR membranes. As the ratio fell more LR than H R was observed. Despite the variation in respective amounts of H R and LR membranes the proportion

of S membranes of total ER remains extremely constant. Cell cultures containing a large population of cells in early S phase have a particularly high content of H R membranes. Maximal light chain immunoglobulin synthesis in MPC-11 cells occurs in late G1/early S phase [37], coinciding with the peak of H R membranes. Immunoprecipitation of nascent polypeptide chains isolated from polysomes of H R and LR membranes showed a fourfold enrichment of immunoglobulin light chain in the former class of R E R membranes. There is therefore a correlation between cell cycle, amount of H R membranes and synthesis of light chain immunoglobulin. Recent experiments have verified that H R membranes are enriched in mRNA coding for light chain immunoglobulin [38]. Poly-A containing mRNA was isolated from H R and LR membrane subfractions by oligo-dT cellulose chromatography, translated in an in vitro system and the products analyzed for light chain polypeptides by immunoprecipitation. Sixty per cent of total 14C-amino acid-labeled polypeptides synthesized by m R N A isolated from H R membranes was in light chain while the corresponding percentage for LR membranes was only 15. The accumulation of MPC-11 cells in G1 phase by subjecting a culture to a 48 h period of starvation results in a major conversion of E R membranes to the S type [39]. After feeding the culture with fresh medium for 2.5 h then the percentage of S membranes decreases, that of LR increases while that of H R remains virtually unchanged (Table 3). It is thus apparent that following re-entry into the cell cycle from a G1 arrested state then the two R E R

Table 3. ER subfractions in starved and starved/fed cultures Table 2. ER subfractions and dependency on cell cycle Type of cells

Late GJearly S Early G1

Culture type

ER membrane subfractions HR

LR

S

38.8 2.5

17.1 45.1

39.0 47.8

Distribution of radioactivity in ER subfractions isolated from 3H-choline labeled MPC-11 cells. See ref. [36] for details.

Starved cells (Go) Starved/fed cells

ER membrane subfractions HR

LR

S

2.4 3.1

16.3 42.0

81.3 54.9

Distribution of radioactivity in 3H-choline labeled MPC-ll cultures. Starved cells (Go arrested) and starved cells incubated for 2.5 h with an equal volume of fresh medium. For details see ref. [39].

98 subfractions exhibit different rates of 'recovery'. A similar behaviour has been observed in Krebs II ascites cells during a period of incubation in suspension culture after propagation of cells in the mouse peritoneal cavity. After 0.5 h of incubation HR membranes account for about 11% of total ER membranes while at 18 h the value was above 50%, the corresponding values for LR membranes were about 63% and 26% respectively [22]. Taking into account these results and those obtained from MPC-11 cells it is clear that the HR membrane subfraction does not merely appear as part of a response to 'step-up' conditions promoted by the addition of fresh medium. The time-dependent behaviour of HR and LR membranes in these cell lines suggests a functional difference between these two domains of the RER.

RER membranes and the cytoskeleton The cytoskeleton consists of essentially three components: microtubules, microfilaments and intermediate filaments [40--43]. Cellular organelles are closely associated with elements of the cytoskeletal structures. ER membranes for examples, appear to interact with microfilaments. Cytochalasin B (CB) has several effects on cellular processes, one of these being prevention of micro filament formation in vivo, thus incubation of cells with this agent will result in breakdown of the micro filament system [44]. In Krebs II ascites cells the yield of total ER membranes was increased substantially after a 20min incubation with 5 ~g CB/ml culture [45]. The individual increases in yield o f E R subfractions were about 200%, 150% and 100% for the HR, LR and S subfractions respectively. These results perhaps indicate that ER subfractions differ in the manner by which they interact with the micro filament system. High concentration of CB (40/zg/ml culture) resulted in a conversion of HR to LR membranes. Thus CB at high concentration presumably results in a dissociation of polysomes initially attached to HR membranes causing a drop in density and thereby altering their migration properties upon ultracentrifugation. A conversion of R E R to S did not occur since there was no increase

in S membranes after treatment at high CB concentration. The effect of high CB concentration observed in vivo could not be mimicked in vitro such that a loss of ribosomes from HR membranes by an RNase type mechanism (see below) can be ruled out. The differences observed between HR and LR membranes regarding their susceptibility to CB is further evidence indicating that these two RER subfractions are indeed different membrane species. Recent experiments have shown that HR membranes of L-929 cells, in contrast to LR and S membranes, bind 3H-cytochalasin B very poorly [46]. It was also shown that the LR subfraction is composed of two types of membranes, one of high density, termed LR I (low CB affinity), and the other of low density, termed LR II (high CB affinity). When proteins in HR and LR membranes were separated by electrophoresis, and gels immunoblotted with monoclonal antibodies against actin, then it was observed that actin is present in LR but not in HR membranes [46]. One can thus conclude that there is no direct attachment of actincontaining microfilaments to HR membranes. When the salt concentration of the buffer used for the disruption of MPC-11 cells by nitrogen cavitation was increased from 25 to 100 mM KC1 then the yield of HR membranes was decreased by more than 90% [19]. This was in contrast to the yields of LR and S membranes which were virtually unaltered. However, when the salt concentration was reduced from 100 to 25 mM KC1 by dilution at the moment of reduction in pressure from 40 to 1 atmosphere (i.e. upon cell disruption), then a 3.4 fold increase was observed in the yield of HR membranes [47]. Other experiments indicated that when the salt concentration of the homogenization buffer was increased above 40 mM then the yield of HR membranes fell drastically. It is therefore evident that the optimal yield of HR membranes is dependent on the salt concentration and this dependency has been attributed to an interaction between HR membranes and a component of the cytoskeleton [7]. The manner by which such interaction is manifested is at present unknown,

99 27S

2sI

27S i8$

8S

1 /'i D

1.25

\

/

J\[

i

/

.\j.-.~J'

7

\

7

14

14

? E

E A

o

t61 8

._.//

I

7

H

c /

}

./ I"

/ 7

14

//\

t .~

J/ /

15

g

7 ,

i

i -~-

FRACTION NO.

Fig. 1. Cultures of MPC-11 cells were labeled with 3H-uridine [38] for 5 h, subcellular fractions were prepared following cell disruption by nitrogen cavitation and poly A-containing RNA was isolated by oligo d(T) cellulose chromatography. RNA was separated into different size classes on linear 5-20% sucrose gradients centrifuged for 4 h at 39,000 rpm (100,000 xg) in the SB 488 rotor in the International B-60 ultracentrifuge. Radioactivity in gradient fractions was determined in material precipitated by the addition of an equal volume of 10% TCA. 18S and 27S ribosomal RNA was run on parallel gradients in order to serve as markers. A: LR membranes, B: HR membranes, C: NER, D: S membranes, E: cytosol.

mRNA in subfractions of endoplasmic reticulum

Size differences among poly A-containing mRNA species in subfractions of ER, NER and cytosol

have been examined in MPC-11 cells. 3H-uridine labeled poly A containing R N A was isolated from the respective fractions by oligo(dT) cellulose chromatography. Radioactivity in poly A-containing R N A was calculated as a percentage of the total radioactivity applied to the columns. The percentages obtained for the cytosol, NER, LR, H R and S membranes were 2.6, 5.9, 6.2, 8.5 and 22.0% respectively. We have earlier observed the presence of R N A in the S membrane subfraction [20]. When solubilized S membrarre preparations were run on sucrose density gradients to display ribosomes/polysomes, then virtually no A260material was observed in gradient profiles. It was surprising, therefore, to find that the S membranes contained such a high percentage of poly A-containing RNA. This 'membrane RNA' [48] may well represent mRNA that is in transit from the nucleus to sites of translation in the cytoplasm. The R N A could perhaps be attached to the cytoskeleton since we have recently demonstrated that actin microfilaments are associated with S membranes [46]. Furthermore, evidence has been provided showing an attachment of mRNA and polysomes to the cytoskeleton [49-55]. Other experimental procedures will be required in order to determine whether the poly A-containing R N A observed in the S membranes is membranebound or bound to components of the cytoskeleton. The low percentage of poly A-containing R N A in the cytosol fraction (2.6%) was as expected taking into account the large pool of ribosomes present in this cellular compartment [20]. Poly A-containing R N A species in individual subcellular fractions were separated into size classes by centrifugation on sucrose density gradients, fractions were collected and radioactivity was determined. The results are presented in Fig. 1. In all subcellular fractions the bulk of R N A appeared in a major peak of size below 18S. The distribution of poly A-containing R N A according to size class is presented in Table 4. The S membrane fraction contained by far the largest amount of mRNA species above 27S, while the cytosol contained the lowest. It is difficult to account for these large differences, however, as mentioned above, if the m R N A present in the S membrane subfraction is associated with the cytoskeleton (possibly in transit

100 from the nucleus), then perhaps the large size observed suggests that some form of sorting mechanism for mRNA based on size has occurred either before or during release of mRNA from the nucleus. That the S membrane subfraction on the whole contains large poly A-containing species is reflected by the fact that 43.5% of the total was above 18S, while the corresponding value for the cytoplasm was 13.8%. Of the rough E R subfractions the LR membranes had the largest average size with 38.1% of poly A-containing R N A above a size class of 18S. Two possible explanations to account for the difference between H R and LR membranes have been earlier suggested [20]: either the H R contained larger m R N A species on average than the LR membranes, thus requiring a greater density of ribosomes, or the H R membrane surface can allow for the binding of larger numbers of ribosomes. Based on the results described here the latter explanation would seem to be the correct one since the average size of poly A-containing R N A was highest in the LR membrane subfraction. The percentage distribution of poly A-containing R N A among the components of the ER was 65, 23, 9 and 3% for the NER, HR, LR and S subfractions respectively. It has previously been shown that the respective amounts of H R and LR subfractions that can be isolated from MPC-11 cells is cell cycle dependent [36]. In a cell population with a major proportion of cells in late G1/early S phase then the H R subfraction accounted for

38.8% of total E R membranes while the LR and S subfractions accounted for 17.1 and 39.0% respectively. In contrast, in cultures exhibiting G1 arrest the distribution was 2.5, 45.1 and 47.8% respectively for the H R , LR and S membrane subfractions. In starved MPC-11 cells (i.e. Go) the distribution of membranes was 2.4% (HR), 16.3% (LR) and 81.3% (S) while after 2 1/2 h feeding the equivalent percentages were 3.1, 42.0 and 54.9% [39], indicating a more rapid degree of recovery of LR than H R membranes. It is thus obvious that the percentage distribution of poly A-containing RNA among ER membrane subfractions indicated above only represents those values for the experiments described here on heterogeneous cultures of cells. The percentages will otherwise vary considerably during cell cycle traverse.

Interaction between polysomes and HR and LR membranes

Since RNA/protein and RNA/phospholipid ratios are higher in H R than LR membranes this is taken as evidence indicating a greater density of ribosomes on the former type of membranes [20]. Isolation of H R and LR membranes from discontinuous sucrose gradients followed by extensive washing and re-centrifugation had no effect on the migration properties of the membrane vesicles. However, when H R and LR membranes were isolated

Table 4. Size differences in mRNA species

Subcellular fraction

Poly A-containing RNA Above 27S

Cytosol NER LR HR S

18S-27S

Below 18S

cpm

% total

cpm

%total

cpm

%total

906 1,517 629 719 563

1.0 2.4 5.5 2.9 13.7

1 l, 127 12,537 3,758 6,391 1,226

12.8 19.1 32.6 25.7 29.8

74,864 50,759 7,138 17.762 2,319

86.2 78.5 61.9 71.4 56.5

3H-uridine labeled poly A-containing RNA species in subcellular fractions of MPC-11 cells classified according to size. Radioactivity in fractions 1-6, 7-10 and 11-15 respectively for individual subcellular fractions (data from Fig. 1) has been summed and is expressed as a percentage of the total poly A-containing RNA in each subcellular fraction.

101

o•.

B. 5"

Fig. 2. Proposed types of association between polysomes and rough E R membranes. A-HR membranes: Initiation occurs on free ribosomes in the cytosol. Attachment of complex to rough E R through interaction between nascent polypeptide chain and membrane receptors. Ribosomal subunits are released from mRNA upon chain termination (signal recognition hypothesis). B-LR membranes: 60S subunits are in a long-term association with membranes and are not released upon chain termination, in contrast to 40S subunits which are released to the general cellular pool of subunits.

and then treated with RNase before re-centrifugation then differences were apparent: the HR membranes lost their ability to migrate through 2.0 M sucrose and collected at the position of LR membranes in density gradients, while the migration properties of LR membranes were unaffected by such treatment [56]. These experiments suggested that in HR membranes not all ribosomes in polysomes were membrane bound such that RNase treatment caused a loss of subunits from the membranes resulting in vesicles of lower density. These accumulated at the LR position upon re-centrifugation. In LR membranes, on the other hand, the type of association between ribosomes and membranes must be otherwise since the sedimentation characteristics of this class of RER was unaffected. If the situation in LR membranes is such that the majority of ribosomes in polysomes are membrane bound then cleavage of mRNA inbetween adjacent ribosomes would not result in the release of major amounts of ribosomes from the membrane surface. That RNase treatment of RER results in the release of ribosomes is well documented, but that release occurs from a specific membrane species is a relatively recent finding [56]. The manner of interaction between polysomes and HR and LR

Table 5. Summary of differences between LR and HR subfractions of endoplasmic reticulum Cell line

Property

LR

HR

Ref.no.

MPC-11

RNA/protein ratio RNA/phospholipid ratio Yield dependent on [salt] of homogenization buffer Yield in late Gl/early S Yield in early G1 Sensitive to RNase Sensitive to cycloheximide Ribosomes tightly bound Ribosomes loosely bound Yield in Go arrest cells Present in starved cells Order of re-appearance after feeding starved cells Order of re-appearance after period of chilling Actin present 3H-cytochalasin B binding Percentage increase in yield after cytochalasin B Order of appearance during incubation in vitro

0.34 0.38 No Low High No No Yes No High Yes First First Yes High 150 First

0.40 0.44 Yes High Low Yes Yes No Yes Low No Second Second No Low 200 Second

20 20 19 36 36 55 55 38 38 7 7 7 61 46 46 45 22

L-929

Krebs II ascites

102 membranes is envisaged in Fig. 2, and a summary of observed differences between LR and HR subfractions of RER is presented in Table 5. The initiation of synthesis of polypeptides bearing a signal sequence occurs on ribosomes which are not initially membrane associated [57-59]. After the signal sequence has emerged from the 60S ribosome binding between RER surface and ribosome is facilitated and membrane-bound polysomes are thus formed. Upon the termination of the synthesis of such proteins then the ribosomal subunits are released from the mRNA and they join the 'free' pool. The synthesis of other proteins not containing a signal sequence may occur on membrane-bound polysomes where the attachment of ribosomes to membranes could be of more permanent nature, i.e. a situation not requiring the complete loss of ribosomes from mRNA between subsequent 'rounds' of protein synthesis. Evidence that this is the case for 60S subunits associated with LR membranes has been recently provided [38]. The equilibration of these ribosomes with the general cellular 60S pool was apparently slow since the specific activity (cpm/A260) of LR 60S subunits was very much lower than that observed in the HR membranes and the cytosol. These results demonstrate a slow turnover rate of LR membrane-associated 60S subunits, and furthermore, indicate that this population is in a more permanent form of attachment with the membrane surface than the comparable population in the HR membrane fraction. Experiments performed with MPC-11 cells where cycloheximide was used to inhibit protein synthesis resulted in a selective conversion of HR to S membranes while the amount of LR membranes remained unaffected [56]. The situation could be reversed by diluting the culture such that the inhibitory effect of cycloheximide was abolished. These results, which are concomitant with those where RNase was used, clearly show that there is a difference in the type of interaction between polysomes and HR and LR membranes. The type of attachment of polysomes to HR but not LR membranes is that envisaged for those bearing a signal sequence. Taking these results into account it is of interest here to draw a parallel with the

earlier findings of Rosbash and Penman [60, 61] who demonstrated the presence of two types of membrane-bound polysomes in eukaryotic cells, namely 'tightly bound' and 'loosely bound' polysomes. Based on available evidence it seems likely that HR and LR membranes contain loosely and tightly bound polysomes respectively, but the significance of this is not yet understood. It seems likely that translocated membrane proteins are synthesized in a manner similar to that for secretory proteins [63]. In the case of membrane proteins, however, the translocation step must be arrested at some stage in order to enable integration of the molecule into a membrane site in the lipid bilayer. Certain membrane proteins which are only embedded in the membrane at the endoplasmic surface presumably do not require signal sequences, and would therefore be synthesized independent of the translocation event. The possibility thus arises that secretory proteins and translocated membrane proteins on the one hand, and embedded membrane proteins on the other, may be synthesized on different domains of the RER system.

References 1. Palade GE: J Biophys Biochem Cytol 2, 417-422, 1956 2. Birbeck MSC, Mercer EH: Nature (Lond.) 189,558-560, 1961 3. Svardal AM, Pryme IF: Subcell Biochem 7, 117-170, 1980 4. Threadgold LT: The Ultrastructure of the Animal Cell, Pergamon, Oxford, U.K. 1976 5. McIntosh PR, O'Toole K: Biochim Biophys Acta 457, 171212, 1976 6. Shore GC, Tata JR: Bioehim Biophys Acta 472, 197-236, 1977 7. Pryme IF: Mol Cell Biochem 71, 3-18, 1986 8. Eriksson LC: Acta Pathol et Microbiol Scand Suppl 239, 1973 9. Leskes A, Siekevitz P, Palade GE: J Cell Bio149,264-287, 1971 10. Malhotra SK: SubceU Biochem 1,171-177, 1972 li. TewariJP, MalhotraSK: SubcellBiochem2,287-293, 1973 12. Mueckler MM, Pitot HC: J Cell Biol 94,297-307, 1982 13. Morrison TG, Lodish HF: J Biol Chem 250, 6955-6962, 1975 14. Rapoport TA, Wiedman M: Current topics in membranes and transport 24, 1-47, 1985 15. Hortsch M, Avossa D, Meyer DI: J Cell Biol 103,241-253, 1986

103 16. 17. 18. 19. 20.

Hortsch M, Meyer DI: Int Rev Cytology 102,215-242, 1986 Siegel V, Walter P: TIBS 13. 314-316, 1988 Melchers F: Biochemistry 10,653-659, 1971 Svardal AM, Pryme IF: Anal Biochem 89,332-336, 1978 Svardal AM, Pryme IF, Dalen H: Mol Cell Biochem 34, 165-175, 1981 21. Pryme IF, Svardal AM, Skorve J: Mol Cell Biochem 34, 177-183, 1981 22. Fjose A, Pryme IF: Mol Cell Biochem 56, 131-136, 1983 23. Pryme IF, Svardal AM, Skorve J, Lillehaug JR: in Cancer Cell Organelles, (Reid E, Cook GMW, Morr6 DJ ed.) Vol. 11, pp 293-298, Ellis Horwood Limited, Chichester, U.K. 1981 24. Shore GC, Tata JR: J Cell Biol 72, 714-725, 1977 25. Shore GC, Tata JR: J Cell Biol 72, 726-743, 1977 26. Meier PJ, Spycher MA, Meyer MA: Exp Cell Res 111, 479-483, 1978 27. Meier PJ, Spycher MA, Meyer MA: Biochim Biophys Acta 646, 283-297, 1981 28. Stuhne-Sekalec L, Stanacev NZ: Can J Biochem 60, 877881, 1982 29. Hazel JR, Zerba E: J Comp Physiol B 156, 665-674, 1986 30. Pathak RK, Luskey KL, Anderson RGW: J Cell Biol 102, 2158-2168, 1986 31. Pryme IF, Quarcoo ET: Mol Cell Biochem 79, 119-124, 1988 32. Pryme IF: Biochem Internat 16, 755-764, 1988 33. Male R, Lillehaug JR, Djurhuus R, Pryme IF: Carcinogenesis 6, 1367-1370, 1985 34. Pryme IF, Lillehaug JR, Fjose A, Kleppe K: FEBS Lett 152, 17-20, 1983 35. Pryme IF: Int J Biochem (In press), 1989 36. Svardal AM, Pryme IF: Mol Cell Biochem 29, 159-171, 1980 37. Garatun-TjeldstO O, Pryme IF, Weltman J, Dowben RM: J Cell Biol 68, 232-239, 1976 38. Pryme IF: Biochim Biophys Acta 967, 218-223, 1988 39. Pryme IF: Biochem Internat 8, 647-654, 1984

40. Weber K, Osborn M: in Cellular Controls in Differentiation, (Lloyd CW, Rees DA ed.) pp 11-28, Academic Press, New York, 1981 41. Schliwa M, van Blerkom J: J Cell Biol 90,222-235, 1981 42. Wolosewick JJ, Porter KR: J Cell Biol 82, 114-139, 1979 43. Batten BE, Aalberg JJ, Anderson E: Cell 21,849-857, 1980 44. Stossel TP: Fed Proc 36, 2181-2184, 1977 45. Fjose A, Pryme IF: Cell Biochem & Function 2, 38-42, 1984 46. Pryme IF, Hesketh JE: Cell Biol Int Repts 11,615-623, 1987 47. Pryme IF: Biochem Internat 13, 28%293, 1986 48. Tata JR: Nature (UK) 213,566-569, 1967 49. Lenk R, Penman S: Cell 16, 289-301, 1979 50. Cervera M, Dreyfus G, Penman S: Cell 23,113-120, 1981 51. Van Venrooij WJ, Sillekens PTG, van Eekelen AG, Reinders RJ: Exp Cell Res 135, 79-91, 1981 52. Moon RT, Nicosia RF, Olsen C, Hille MB, Jeffery WR: Dev Biol 95,447-458, 1983 53. Jeffery WR: Dev Biol 103,482-492, 1984 54. Ornelles DA, Fey EG, Penman S: Mol Cell Biol 6, 16501662, 1986 55. Hesketh JE, Pryme IF: FEBS Lett. 231, 62-66, 1988 56. Pryme IF: in Protein Synthesis (Abraham AK, Eikhom TS, Pryme IF ed.) pp 117-129, The Humana Press, 1983 57. Blobel G, Sabatini DD: in Biomembranes, (Manson LA ed.) Vol 2, pp 193-195, Plenum Publ. Corp., New York, 1971 58. Blobel G, Dobberstein B: J Cell Biol 67,835-851, 1975 59. Blobel G, Dobberstein B: J Cell Biol 67, 852-862, 1975 60. Rosbash M, Penman S: J Mol Biol 59,227-241, 1971 61. Rosbach M, Penman S: J Mol Biol 59,243-253, 1971 62. Pryme IF: Biochem Internat 11,387-396, 1985 63. Schatz G: Nature 321,108-109, 1986

Address for offprints: I.F. Pryme, Biokjemisk institutt, Universitetet i Bergen, Arstadveien 19, N-5009 Bergen, Norway