The Jeanne Manery-Fisher Memorial Lecture 1991. Maturation of reticulocytes: formation of exosomes as a mechanism for shedding membrane proteins.

REVIEW / SYNTH~SE

The Jeanne Manery-Fisher Memorial Lecture 1991

/

La conference a la memoire de Jeanne Manery-Fisher 1991

Maturation of reticulocytes: formation of exosomes as a mechanism for shedding membrane proteins

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 06/03/14 For personal use only.

R. M. JOHNSTONE Department of Biochemistry, McGill University, 3655 Drummond Street, Montrkal, Que., Canada H3G 1 Y6 Received October 15, 1991

JOHNSTONE, R. M. 1992. Maturation of reticulocytes: formation of exosomes as a mechanism for shedding membrane proteins. Biochem. Cell Biol. 70: 179-190. The transferrin receptor is a member of a group of reticulocyte surface proteins that disappear from the membranes of reticulocytes as the cells mature to the erythrocyte stage. The selective loss of membrane proteins appears to be preceded by the formation of multivesicular bodies (MVBs). At the reticulocyte stage, many species of mammalian red cells including man, and one nucleated avian species (chicken), contain these intracellular structures in both natural and induced anemias. Also characteristicof blood containing reticulocytes is the presence of circulating vesicles (exosomes), which contain proteins and lipids characteristic of the plasma membrane. These exosomes appear to arise from the contents of the MVBs, after the fusion of MVBs with the plasma membrane. The proteins in the exosomes are those frequently lost during red cell maturation (e.g., transferrin receptor). The major transmembrane proteins (such as the anion transporter) are fully retained into the mature red cell, indicating a highly selective mechanism of recognition of a specific group of proteins. The exosomes are largely devoid of soluble proteins and proteins associated with lysozomes or mitochondria. A speculative model is proposed which addresses the questions of the maturation-induced structural changes in a class of membrane proteins, their recognition and selective loss involving exosome formation, and the release of exosomes to the circulation. Key words: transferrin receptor, nucleoside transporter, reticulocyte maturation, multivesicular bodies, 70-kilodalton protein.

JOHNSTONE, R. M. 1992. Maturation of reticulocytes: formation of exosomes as a mechanism for shedding membrane proteins. Biochem. Cell Biol. 70 : 179-190. Le rkcepteur de la transferrine fait partie du groupe des protkines prksentes 11 la surface des rkticulocytes et qui disparaissent des membranes des rkticulocytes quand les cellules atteignent le stade krythrocyte. La perte sklective des protkines membranaires semble prkkdke par la formation de corps multivksiculaires (MVC). Au stade rkticulocyte, plusieurs esptces de globules rouges mammaliens, dont l'homme, et une esptce aviaire nuclkke (poulet), renferment ces structures intracellulaires dans les deux formes d'ankmie, naturelle et induite. La presence de vksicules circulantes (exosomes) renfermant des protkines et des lipides caractkristiques de la membrane plasmique est kgalement une particularitk du sang contenant des rkticulocytes. Ces exosomes proviendraient du contenu des MVC aprts leur fusion avec la membrane plasmique. Les protkines dans les exosomes sont souvent celles perdues lors de la maturation du globule rouge (e.g., rkcepteur de la transferrine). Les principales protkines transmembranaires (tel le transporteur des anions) sont entikement retenues dans le globule rouge mature, preuve d'un mkcanisme de reconnaissance hautement selectif d'un groupe spkifique de protkines. Les exosomes sont largement dkpourvus de protkines solubles et de protkines associkes a w lysosomes ou a w mitochondries. Nous proposons un modMe spkculatif qui aborde les questions des changements structuraw induits par la maturation dans une classe de protkines membranaires, leur reconnaissance et leur perte sklective impliquant la formation des exosomes et la libbation des exosomes dans la circulation. Mots clks : rkcepteur de la transferrine, transporteur de nuclkosides, maturation des rkticulocytes, corps multivksiculaires, protkine de 70 kilodaltons. [Traduit par la rkdaction]

Introduction The mammalian red cell, particularly the human cell, ranks amongst the most intensively studied of cells, probably second only to Escherichia coli. Yet, in many respects it is far from being understood. The factors underlying red cell aging are still subjects of intense investigation with few clear MVBs, multivesicular bodies; TFRs, transferABBREVIATIONS: rin receptors; HK, high potassium red cells; LK, low potassium red cells; ER, endoplasmic reticulum; EM, electron microscopy; kDa, kilodalton(s); NBMPR, nitrobenzylthioinosine; TF, transferrin; SDS-PAGE, sodium dodecyl sulfate - polyacrylamide gel electrophoresis. Printed in Canada / Imprim6 au Canada

answers. It is well known that many plasma membrane functions are diminished during the maturation process of red cells. Although the transition between the reticulocyte and mature red cell has been examined for many years, the mechanisms underlying the complete removal of some proteins during maturation, in the face of retention of others, have not been established. For example, mammalian red cells, in general, lose all their TFRs during reticulocyte maturation (Jandl and Katz 1963; Frazier et al. 1982; VanBockxmeer and Morgan 1979). These receptors are very high in number in young cells. In immature rabbit normoblasts (VanBockxmeer and Morgan 1979), 800 000 receptors per cell have been reported.

BIOCHEM. CELL BIOL. VOL. 70. 1992


180

FIG. 1. Exosomes in the circulation of (A) rabbit (phlebotomized), (B) sheep (phlebotomized), (C) rat (phenylhydrazine treated), (D) 10- to 14-day piglet (naturally anemic), (E) chicken (14 day old embryo), and (F) man (phlebotomized). Bar, 200 nm. Conversely, another prominent membrane protein, the anion exchanger, does not appear to be lost (or diminished) in any significant way during red cell maturation (Foxwell and Tanner 1981; Zanner and Galey 1985). Further, it is clear that subtle structural differences between proteins are likely to be involved in targeting the loss of a specific red cell protein during maturation of the reticulocyte. For exarnple, there are marked differences in the nature of the proteins lost between different species of mammalian red cells. Some

species (pig) lose the majority of their glucose transporters during maturation (Zeidler and Kim 1982). Human red cells, however, maintain very high levels of glucose transporters (Laris 1958) (in fact, higher than many other species). Similarly, red cells vary in their content of nucleoside transporters, with sheep red cells losing the majority of the nucleoside transporters during maturation (Jarvis and Young 1982), unlike human (Cass and Paterson 1972) and pig cells (Watts et al. 1979) where nucleoside transporters


REVIEW /

SYNTHBSE

181

FIG. 2. Presence of multivesicular sacs in reticulocytes. Thin sections of reticulocytes are from (A) embryonic chicken blood ( X 87 779, (B) hemachromatosis patient during active treatment by phlebotomy ( x 87 775). (C) anemic cat (naturally occurring anemia) ( x 38 090), and (D) anemic dog (naturally occurring anemia) ( x 99 680). Photographs C and D were courtesy of Dr. Yoshimitsu Maede DVM, Department of Veterinary Internal Medicine, Hokkaido University, Sapporo, Japan. are retained. It cannot be argued that the major difference is in the number of a particular receptor or transporter with which the cell starts out and that a constant fraction is maintained into the mature red cell. Thus as noted above, all the TFRs are lost, but none of the anion transporter. Moreover, in dog red cells (Maede and Inaba 1985), the HK reticulocytes have approximately two times more ouabainbinding sites than LK cells. A larger percentage of the total is lost from LK cells, although the number of sites lost is greater in HK cells. Thus HK cells retain about 15% of the initial number of ouabain-binding sites, whereas LK cells retain less than 5% of the number present in reticulocytes. Teleologically, both the loss of the transferrin receptor and the retention of the anion transporter make sense. Iron can be a toxic element since it can lead to the formation of free radicals from its interaction with oxygen by the wellknown Fenton reaction. In the mature red cell, which has no capacity to replace its proteins, such oxidative damage could be lethal. Thus, once the heme- and proteinsynthesizing stages have passed, the inability to take up iron

becomes a life-saving device. Similarly, since a major function of hemoglobin is to carry bicarbonate to the lungs, where it is released from the cell in exchange for C1- (for review, see Jay and Cantley 1986), the presence of the anion exchanger in full force is essential for normal red cell function and the appropriateness of retention of the anion transporter is self-evident. More difficult to explain, at present, is the selective loss of other transporters in a species-specific way. Why are both glucose and nucleoside transporters maintained to a high level in human red cells whereas the glucose, but not the nucleoside, transporter is lost in pig red cells? Similar questions apply to retention or loss of other functions, such as amino acid transporters or the Na,K-transporter. We have been interested in obtaining some understanding of the mechanism by which the mammalian red cell selectively remodels its plasma membrane during the transition between reticulocyte and erythrocyte. To this end, we have been following the fate of the transferrin receptor, which as mentioned, is universally and totally lost from

182

BIOCHEM. CELL BIOL. VOL. 70,


mammalian red cells. Our present understanding of the mechanism for the selective removal of specific proteins from the red cell membrane is outlined below. Although the major emphasis is on the transferrin receptor, our current belief is that the observations apply to many plasma membrane proteins whose function is lost or diminished during maturation.

Loss of function with red cell age The fact that very young red cells are different from older red cells has been known for many years (for review, see Rapoport 1986). Although the direct relationship between cell age and red cell density is still under dispute (Suzuki and Dale 1988), little disagreement exists that mammalian reticulocytes are (a)less dense than the corresponding mature erythrocytes and (b) become smaller as they mature. It is generally accepted that a variety of catalytic activities decline as the cell progresses from the reticulocyte to the mature cell stage. Amongst these activities are included (but differ amongst species) Na,K-ATPase (Maede and Inaba 1985; Blostein et al. 1983), insulin receptors (Thomopoulos et al. 1978; Im et al. 1984; Ginsberg and Brown 1982), glucose transporters (Zeidler and Kim 1982; Johnstone et al. 1987), nucleoside transporters (Jarvis and Young 1982; Johnstone et al. 1987), adrenergic receptors (Montandon and Porzig 1984; Limbird et al. 1980), mitochondrial activity (Rapoport 1986; Simpson and Kling 1968; Gronowicz et al. 1984; Gasko and Danon 1974), lysosomal activity (Rapoport 1986), and Golgi, ER functions, and ribosomes (Rapoport 1986). Soluble enzyme functions such as hexokinase also diminish, although this may be attributed (at least in part) to the loss of mitochondrially associated hexokinase (Rapoport 1986; Magnani et al. 1984). Although it has been known for over a quarter of a century that mammalian reticulocytes lose their capacity to bind the serum protein transferrin (and hence the capacity to take up ligated iron) (Jandl and Katz 1963), early studies suggested that the transferrin receptor may be inactivated (Leibman and Aisin 1977) by loss of carbohydrate side chains or proteolyzed by cellular proteases. Regulated loss of red cell proteins by proteolysis has been given both credence (Maede and Inaba 1985; Rapoport 1986; Blostein et al. 1983; Montandon and Porzig 1984; Blostein and Grafova 1987; Porzig et al. 1991) and importance since the discovery of ATP-dependent proteases requiring ubiquitination of proteins targeted for digestion (Goldberg and St. John 1976; Hershko and Cienchanover 1982). The proteolytic system@)showed a striking loss during reticulocyte maturation (Raviv et al. 1987), thus making the proposal of maturation-associated proteolysis of redundant proteins particularly attractive. However, it needs to be emphasized that, to date, there is no direct evidence that any plasma membrane protein in situ is digested by an ATP- and (or) ubiquitin-dependent process. The most direct attempt to show proteolysis of a plasma membrane protein was that of Maede and Inaba (1985), where purified exogenous Na,KATPase was shown to be proteolyzed by reticulocytes. The protein in situ was not analysed for proteolysis and there is no evidence that the lower plasma membrane protein concentration in erythrocytes is achieved by selective proteolysis. In fact it has been suggested that the mitochondrial membrane proteins, but not plasma membrane proteins, are the primary substrates for proteolysis in the reticulocyte

1992

stroma, along with denatured soluble proteins (e.g., phenylhydrazine oxidized) or proteins containing amino acid analogues (Rapoport 1986). Now evidence is beginning to accumulate that the loss of surface area and plasma membrane protein content during maturation may be achieved by selective removal of membrane proteins by vesiculation.

Vesicle formation as a mechanism for selective loss of proteins and surface area reduction An examination of the plasma of phlebotomized or anemic animals (characterized by a high reticulocyte level) shows the presence of a population of vesicles, 40-100 nm in diameter (depending on species), which can be harvested by centrifugation. These structures are found under conditions of naturally elevated reticulocyte levels (e.g., embryonic chickens and new born piglets), as well as in response to induced reticulocytosis caused by bleeding (Fig. 1) (Johnstone et al. 1989, 1991). The origin of these circulating vesicles appears to be the multivesicular bodies found in maturing reticulocytes (Johnstone et al. 1991;Pan et al. 1985; Harding et al. 1983, 1984). Thin sections of red cells from the same blood sample as that used to obtain the circulating vesicles show the presence of intracellular multivesicular bodies (Johnstone et al. 1989, 1991; Pan et al. 1985) (Figs. 2 and 3). Inside the sacs are vesicles which appear to be identical in size to those in the plasma, suggesting that the sac contents are the precursors of the structures in the circulation. In naturally occurring anemias of dogs and cats studied by Maede in Japan, multivesicular structures are seen by EM in thin sections of red cells (Figs. 2C and 2D). Similar multivesicular structures and circulating vesicles, which are seen in animals, appear in patients with hemochromatosis undergoing treatment by phlebotomy (Fig. 2B). In a cytological study using gold-conjugated anti-transferrin receptor antibody (Fig. 3) the in vitro maturation of sheep reticulocytes was followed. This study showed the progress of vesicle formation after the uptake of the anti-transferrin receptor antibody by endocytosis (Pan et al. 1985). To distinguish these maturation associated vesicles from other types of vesicles, we have named them exosomes. The fact that both exosomes and cellular multivesicular bodies are found in vivo and in vitro in red cells of several species (including man and nucleated chicken red cells) suggests that this process is common to all types of young reticulocytes and disappears in the course of maturation, both in vivo and in vitro. In all instances where the appropriate reagents for probing the system were available, the exosomes showed the presence of the transferrin receptor (Johnstone et al. 1989, 1991; Pan et al. 1985; Harding et al. 1983, 1984) (sheep, rat, man, and chicken) (Fig. 4), suggesting that this is a major route for the maturation-associated loss of the transferrin receptor. Moreover, studies with sheep cells showed that the externalized receptor was intact and had the same size and iodotyrosyl peptide map as the cellular receptor (Pan and Johnstone 1983, 1984). Using 125~-surface-labelled sheep reticulocytes, it could be shown that 125~-labelled receptor accumulated in the exosomes as the cellular receptor diminished (Fig. 5) (Johnstone et al. 1984). However attempts to obtain a quantitative balance sheet between receptor lost from the cells and that recovered in the vesicles were unsatisfactory, since only a small percentage (- 20-25%) of the lost receptor

183


REVIEW / SYNTHESE

FIG. 3. Progress of exosome formation: an EM study using colloid gold-conjugated antibody. After coating the sheep reticulocytes with an antibody to the transferrin receptor at 4°C. the cells were exposed to gold-conjugated secondary antibody at 4°C. The cells were washed and samples taken at intervals at 37OC. (A) Incubation at O°C, (B) 15 min at 37OC, (C) 30 min at 37OC, (D) 3 h at 37OC, and (E) 18 h at 37OC. x 87 775. Arrow indicates the formation of a fusion site. See Pan et al. (1985) for full details.

was actually recovered in the exosomes. The low recovery will be addressed below. The data to date did, however, suggest a role for exosome formation in membrane remodelling, particularly since the lipid content of the exosomes from sheep had the high sphingomyelin content character-

istic of the sheep plasma membrane lipids (Johnstone et al. 1987).

The selective protein content (Johnstone et al. 1987; Pan and Johnstone 1983, 1984) of the exosomes is worthy of comment. Exosomes from sheep (Fig. 6) showed the pres-

184

BIOCHEM. CELL BIOL. VOL. 70, 1992

proteins are lost and not a random group which would likely contain these major membrane proteins (Table 1). The latter observations are consistent with our conclusion that exosome formation is a normal maturation-associated event. Furthermore, data show that selective loss of membrane proteins and overall red cell maturation do not require the intervention of the spleen as was proposed earlier by others (Zweig et al. 1981). It is not clear, however, if the presence of the spleen would speed up the events and (or) reduce the number of circulating exosomes. In our studies with sheep, after 48 h of in vitro incubation, the cells have lost all the TFRs and the methylene blue stainable material. This too is in line with the known time frame of reticulocyte maturation in vivo.


KDa ' A ,

FIG. 4. Detection of the transferrin receptor in exosomes of various species. (A) Immunoprecipitates of exosomes from plasma of phlebotomized rats. Left, molecular mass markers; centre, immune precipitate (I); right, nonimmune precipitate (NI). (B) Immunoblots of chicken exosomes (left, NI; right I) from plasma of embryonic chicks. (C) Immunoblots of exosomes from in vitro culture of chicken reticulocytes (left, N1; right, I). (D) Immunoprecipitates of exosomes from plasma (left, NI; right I) of phlebotomized man. (E) Immunoprecipitates from sheep exosomes after in vitro culture of sheep reticulocytes for 0,6,and 24 h (lanes from left to right). The unidentified bands in this figure are the antibodies used in the immune and nonimrnune precipitations. For immunoblotting, the immune and nonimmunoprecipitates were blotted with anti-TFR antibody after immunoprecipitation with immune and nonimmune IgG, respectively. The unidentified bands are due to the species cross-reactivity of the secondary antibody with the primary anti-TFR antibody.

-

ence of two major proteins, the TFR and a 70-kDa protein which has been identified as the clathrin-uncoating ATPase (Davis et al. 1986). However as discussed below, the exosomes do contain other membrane proteins whose concentration is too low to detect by Coomassie blue staining. Just as interesting is the observation that the soluble cytosolic proteins (hexokinase, lactate dehydrogenase, glyceraldehyde3-phosphate dehydrogenase, and glucose-6-phosphate dehydrogenase) were not detectable in the exosomes (Johnstone et al. 1987). Furthermore, neither of the major membrane spanning proteins, band 3 (the anion exchanger) nor glycophorin A, were detectable by immunoblotting of exosome proteins, indicating that only a selected group of

The nature of proteins associated with the exosomes As noted in the Introduction, various plasma membrane functions, other than that of the TFR, are lost during maturation of the red cell. It thus seemed pertinent to address whether all the plasma membrane functions are associated with a common exosome or whether different populations of exosomes exist. We had shown that the material shed from sheep reticulocytes contained acetylcholinesterase, phosphatase, and the nucleoside transporter (measured by binding of the nucleoside analogue NBMPR), as well as a number of other activities (Johnstone et al. 1987). Although the shed material was devoid of mitochondrial activity (Orr et al. 1987), lysosomal enzyme activities were detectable (Johnstone et al. 1989; Orr et al. 1987). Using magnetic beads coated with anti-transferrin receptor antibody, we established that the class of exosomes containing the TFRs also contained other plasma membrane functions, but did not contain the lysosomal activities (Johnstone et al. 1989). Thus exosome formation appears to be restricted to functions originating from the plasma membrane (Fig. 7). The fact that all the TFRs but only part of the other functions, were removed by the TFR antibody-coated magnetic beads probably reflects the fact that late-appearing exosomes may have fewer TFRs than early exosomes. The reverse may be true for other functions, such as NBMPR binding, which may be released late in maturation and would contain relatively more NBMPR sites than TFR binding sites. We have in fact shown that loss of NBMPR binding activity occurs with a longer t l I 2than TFR loss (Johnstone et al. 1991), but have not yet devised satisfactory procedures to selectively remove exosomes with NBMPR binding activity. Quantification of the lost plasma membrane activities Although all these data are generally consistent with a role for exosome formation in maturation-associated membrane remodelling, the poor quantitative recovery of the lost TFRs remains a problem. It remains possible that exosome formation is a minor event. Because of the instability of TFRs during long-term storage or incubation, we began to explore other functions known to diminish during maturation. Binding of the nucleoside analogue NBMPR is a more stable function than TF binding. For example, freezing of vesicles at - 70°C overnight leads to substantial loss of TF binding (150%), but almost complete retention of NBMPR binding. In view of this, we reexamined the contribution of

185


REVIEW / SYNTHESE

reticulocytes. Sheep FIG. 5. Transfer of labelled TFR from reticulocytes to exosomes during in vitro culture of '2S~-surface-labelled reticulocytes were cultured for the times noted after surface iodination with 1 2 5 ~ At . each period, exosomes (V) were harvested. Plasma membranes were prepared from the remaining cells. All samples were immunoprecipitated prior to SDS-PAGE and radioautography. (A) Coomassie blue stains. V, material derived from exosomes (100 000 x g pellet of cell free medium): lanes 1-3, from sheep reticulocyte cultures; lanes 4-6, from erythrocyte cultures. C, control: lanes 1-3, immunoprecipitates of membranes from cultured reticulocytes; lanes 4-6, membranes from cultured erythrocytes. (B) Autoradiograph of A (from Johnstone et al. 1984). Arrows mark the positions of the monomeric and dimeric forms of TFR.

the exosome route to maturation-associated loss of NBMPR binding. using a number of small technical modifications, as well as shorter intervals to harvest exosomes (Johnstone et al. 1991), we established that over 80% of the lost nucleoside transporter could be recovered in the exosomes. If instead of measuring TF binding, the recovery of immunoprecipitatable '25~-labelledTFRs from '25~-labelledcells was followed after SDS-gel separation, recoveries of 50% of the '25~-labelledprotein could be achieved (Johnstone et al. 1991). From such data we conclude that exosome formation is a major route by which the plasma membrane is altered during the last stage of erythrocyte development (as opposed to erythrocyte aging).

-

kDa

rn

LL-d _I

-

Speculations on the mechanism for selective removal of plasma membrane proteins If one accepts the proposals advanced above that the reticulocyte has a mechanism for selective targeting of membrane proteins for externalization, major questions still remain unanswered. (a) What is the trigger for externalization? It has been shown that inhibition of protein synthesis in vitro does not alter the progress of maturation (M. Adam and R.M. Johnstone, unpublished). This suggests that the components involved in recognition already exist in the cell. Alkalinizing agents, such as chloroquine, will reduce the rate of exosome formation (Pan and Johnstone 1984), and continuous metabolic activity and sustained ATP levels (Johnstone et al. 1991; Pan and Johnstone 1983, 1984) are essential. (b) Do cellular factors accumulate which stimulate

94

=

67

.em--

;

*

-

FIG. 6. SDS-PAGE of exosomes and sheep red cell plasma membranes (from Johnstone et al. 1987). Lane 1, protein content of exosomes (V); lane 2, sheep erythrocyte membranes (M); lane 3, sheep reticulocyte membranes (R).

BIOCHEM. CELL BIOL. VOL. 70,

1992

TABLE1. Enzyme content of exosomes from sheep Function


TFR

Cellular origin PM and CV

Method of detection

Presence

Ref.

TF binding

+

aTFR

f

Harding et al. 1983. 1984; Pan and Johnstone 1983; Pan et al. 1985 Harding et al. 1983, 1984; Pan and Johnstone 1983; Pan et al. 1985 C. Yip and R.M. Johnstone, unpublished Johnstone et a/. 1987, 1989 Johnstone et al. 1987, 1989 Johnstone et al. 1987, 1989 Johnstone et al. 1987, 1989

+ + + + + + + +

Insulin receptor

PM

'2S~-labelled insulin cross-linking

Nucleoside transporter Glucose transporter Acetylcholine esterase Phosphatase

PM PM PM PM

Protein kinase /3-Adrenergic receptor Amino acid transporter Band 3

PM PM PM PM

NBMPR binding Cytochalasin B binding Thiocholine hydrolysis p-Nitrophenyl phosphate hydrolysis Phosphorylation of basic proteins [ ' ~ ] ~ i h ~ d r o a l l o ~ r e n obinding lol [3~]leucine exchange Immunoblotting

Glycophorin A

PM

Immunoblotting

-

/3-Glucuronidase

LYS

-

N-Acetyl-0-glucosaminidase

LYS

-

Johnstone et al. 1989

Glucose-6-P dehydrogenase Glyceraldehyde-3-P dehydrogenase 6-Phosphogluconic acid dehydrogenase Lactate dehydrogenase Succinate dehydrogenase Monoamine oxidase

CYT

Hydrolysis of appropriate umbelliferyl derivative Hydrolysis of appropriate umbelliferyl derivative NADPH formation

Johnstone et al. 1987 Orr et al. 1987 Johnstone et al. 1987 J. Ahn and R.M. Johnstone, unpublished J. Ahn and R.M. Johnstone, unpublished Johnstone et al. 1989

-


CYT

NAD formation

-


CYT CYT Mito Mito

NADPH formation NADH formation PMS reduction Tryptamine oxidation

-

Johnstone et al. 1987 Johnstone et al. 1987 Orr et al. 1987 Orr et al. 1987

NOTE: PM, plasma membrane; CV, cytoplasmic vesicles; LYS, lysosomes; CYT, cytoplasm; Mito,

the removal of proteins? We have found that exogenous hemin increases the rate of loss of specific proteins (Ahn and Johnstone 1989). Heme may accumulate during the maturation process (Ahn and Johnstone 1989). (c) What distinguishes a given function (e.g., nucleoside transporter) in one red cell from that in another (e.g., sheep versus pig), which leads to its near complete loss in sheep red cells in contrast to its retention in the pig red cells? (d) How does the immature red cell differentiate the proteins always lost (e.g., TFR) from the proteins always completely retained (e.g., the anion transporter, band 3)? (e) Why are some functions reduced, but not completely lost and retained to different extents, in different species (e.g., Na, K-ATPase, glucose transporters)? (f) Is there physiological significance for the variation in the types of functions lost from red cells of different species? And if so, what is it? Ready answers are not yet available to these fundamental questions. What I will attempt below are some speculations concerning the cellular events that I believe might lead to the selective loss of proteins during maturation. Firstly, we propose that the 70-kDa protein, the clathrinuncoating ATPase, plays a role in targeting proteins for externalization. In previous work, we have noted that in three different species (sheep, Johnstone et al. 1987; rat and chicken, A. Mathew, unpublished) a common nontrans-

a

mitochondria.

membrane protein is externalized, along with the membrane proteins, into the exosomes. This protein is closely related to the 70-kDa heat shock protein (Davis et al. 1986). From iodination studies, we know that the 70-kDa protein is not accessible at the exofacial surface of exosomes. Given the current studies into this class of proteins (the chaperones, Pelham 1986), the 70-kDa protein contains domains that would bind to unfolded, partly denatured proteins. This includes unfinished proteins prior to reaching their final folded conformation and (or) cellular site, or partially unfolded proteins formed after stress (Pelham 1986). Secondly, we have observed two experimental interventions that increase the rate of exosome formation: (a) the presence of hemin in the incubation medium (Table 2) and (b) heat shock (Table 3). Both interventions increase the loss of TFRs and NBMPR binding. The common feature of both these interventions is their ability to increase the rate of protein oxidation and (or) denaturation. That some alteration has occurred in the cytoplasmic domain of the TFR during maturation was earlier suggested by our studies on the phosphorylation of the TFR (Adam and Johnstone 1987). We showed that the TFR from exosomes failed to become phosphorylated by exogenous protein kinase C (Adam and Johnstone 1987). While we cannot exclude the possibility that the site was completely phosphorylated in

REVIEW / SYNTHeSE


100% -

TF

I

AchE

I

NBMPR

I

HEX-ASE

NIS I S NIB I B

FIG. 7. Multienzyme content of exosomes containing the transferrin receptor. Iron core beads were used as described in Johnstone et al. (1989), to immobilize exosomes containing the transferrin receptor. NIS, iron beads coated with nonimmune serum; supernatant

from the beads assayed for activity. IS, iron core beads coated with anti-TFR; supernatant assayed for activity. NIB, iron core beads coated with nonimmune serum; beads assayed for activity. IB, iron core beads coated with anti-TFR; beads assayed for activity. TF, 3 ~ binding ] assayed. ~ Hex-ASE, ~ ~ hexoseaminidase ~ ~ assayed. TF binding assayed. AchE, acetylcholine esterase assayed. NBMPR, [ TABLE 2. Effect of hemin on loss of membrane proteins during long-term incubation Condition of incubation

% original activity remaining

Control + 2 x lo-' M hemin Control + 2 x lo-' M hemin

76 50

Parameter measured '25~-labelled TF Binding 'H-labelled NBMPR Binding

80 64

NOTE: Sheep reticulocytes were cultured with or without hemin. The activity in the cell membranes was measured initially and after culture. The values given are representative of at least four experiments. The duration of the incubation of the cell suspension was 3-5 h for TF binding and 17-22 h for NBMPR binding. Other conditions are given in Johnstone el al. (1991).

TABLE3. Heat shock and loss of membrane proteins during long-term incubation Parameter measured 125~-labelled TF binding

'H-labelled NBMPR binding

Condition Control (2 h at 37°C) Heat shock (2 h at 43°C) Control (24 h at 37°C) Heat shock (22 h at 37°C after 2 h at 43°C) Control (2 h at 37°C) Heat shock (2 h at 43°C) Control (24 h at 37°C) Heat shock (2 h at 43°C plus 22 h at 37°C)

VO original activity remaining 67 56 16 6 85 77 53 44

NOTE: Sheep reticulocytes were incubated for 2 or 24 h, plus or minus a 2-h exposure to 43'C. The activities were measured in the plasma membranes before and after culture. A typical experiment is shown. Other conditions are given in Johnstone et al. (1991).

the exosomes and incapable of turning over, there is also the possibility that the protein conformation had been altered, as a result of mild oxidation at the cytoplasmic domain. Such a change could make the phosphorylation sites inaccessible to the kinase, even in immunoprecipitates. Thirdly, proteins to be externalized, although transmembrane, may not be firmly attached to the cytoskeleton of the red cell and may undergo internalization by endocytosis, while proteins firmly bound to the cytoskeleton are

prevented from undergoing internalization. We have recently obtained evidence for a cytoplasmic pool of nucleoside transporters (Lu and Johnstone 1992). Blostein and Grafova (1987) also deduced the existence of an intracellular pool of nucleoside transporters. An examination of the functions that appeared in exosomes from sheep (Table 1) shows that many of these, including the insulin receptor (Gavin et al. 1974), glucose transporter (Cushman and Wardzala 1980), and Na,K-ATPase (Pollock et al. 1981), have been found


A, native TFR A, unfolded TFR o,

70-kilodolton protein


70-kDo protein binds

i

6

\

vncoat lng

recycling

0

/

\

fusion

unfolding

/

budding and

l n v e r s l o n of

PM and

/

/

exocytosis of vesicles (exosomes)

FIG. 8. Scheme for targeting of obsolete proteins for externalization. The transferrin receptor is internalized into the cell from a coated pit (designated as a double membrane). Inside the cell, the clathrin is removed from the coated vesicle (double membrane). The endosome can then recycle to the plasma membranz or, if the cytoplasmic domain of the receptor is unfolded or altered, the 70-kDa protein binds to the cytoplasmic domain. This event leads to redirection of the vesicles to another pool where fusion occurs. In these fused vesicles, budding occurs, releasing the buds into the lumen and forming multivesicular bodies (MBVs). Note that the receptor is now reinverted so that extracellular domain faces the lumen of the MVBs. Fusion of MVBs with plasma membrane (PM) leads to release of the buds, named exosomes, into the circulation. in intracellular locations. No cytoplasmic pool of the anion transporter has been described to date. We propose that during maturation the membrane proteins targeted for removal undergo some subtle, denaturing change in their cytoplasmic domains (Fig. 8), perhaps via oxidation. This change in the proteins results in recognition by, and tight binding to, the 70-kDa protein. Instead of being available for recycling to the cell surface, these 70-kDa bound proteins are now segregated away from the recycling endosomes to a class of vesicles where, with the 70-kDa protein still attached, budding occurs into the lumen of the sac. In this way the 70-kDa protein is conveyed to the inside of the newly formed bud. Just as clustering of proteins occurs at the cell surface, clustering of altered proteins may occur on the membrane of the budding compartment, so that a cluster of obsolete proteins is collected into a single bud. Since the different obsolescent proteins are likely to undergo damage at different rates, the early buds may have a somewhat different balance of proteins from those appearing later. The fact that a given function (e.g., nucleoside transporter) is not uniformally externalized in all species would imply that there may be small but important differences in the amino acid sequences in these proteins of the different species. Only slight differences in a small domain of the protein could lead to significant differences in the

susceptibility to damage and hence in the rate and (or) extent of removal. If these speculations about membrane processing in maturing red cells are correct, why in a single species of red cells are some proteins lost completely and others to variable extents? Tentatively, it is proposed that the relative rate of retention is determined by the rate of denaturation of the given protein in relation to the rate of loss of the intracellular membrane structures. As is known, red cells have lost all intracellular organelles, whereas reticulocytes still contain intracellular organelles. The latter are probably involved in exosome formation. A protein with a short tl12 for externalization (like TFR where tl12 = 6-12 h) will be lost in all cases. However, if tl12 = 24-48 h (such as NBMPR binding), some activity may still be left when the processing machinery has been destroyed. We have found, for example, that the lysosomal activities, although not in the TFRexosome fraction, are lost somewhat later than the plasma membrane functions (Orr et al. 1987). Little measurable loss of the lysosomal activity is seen before 24 h and the tl12for loss is closer to 48 than 24 h. If lysosomal activity is involved in processing, a protein with a tl12 for externalization of 24 h is more likely to be partly retained at red cell maturity than one with a tl12of 6 h. Thus the fractional retention of a function into the mature red cell stage would reflect the rate of damage of that protein and its elimination in rela-


REVIEW /

tion to the rate of loss of the machinery for recognizing the altered proteins. Conclusions A speculative model is described that attempts to account for the selective mechanism of membrane remodelling during reticulocyte maturation. The principle features of the model include the following proposals. (a) Proteins to be removed have become denatured and (or) oxidized as part of the maturation event. (b) The designated proteins are not tightly bound to the cytoskeleton and are potentially capable of undergoing endocytosis. A cytoplasmic pool of these proteins is likely to be present, at least during the period of maturation. (c) The 70-kDa protein has a role in recognizing and binding to the partly denatured proteins and in directing these obsolescent proteins to a population of intracellular vesicles where exosome formation takes place. (d) Incomplete externalization of some proteins is related t o their relative resistance to damage during the time of maturation. Much work remains to be done to obtain definitive support for this model. However, despite our incomplete understanding, the recognition that reticulocytes release transferrin receptors is already leading to new approaches to detect and assess the iron status in the human population. Based on the observations first described for the sheep (Pan and Johnstone 1983) and then in man that circulating transferrin receptors are elevated when the reticulocyte level is increased (Kohgo et al. 1986, 1987), assays have been devised (Shih et al. 1990) to quantify the level of the circulating receptor. A close association has been obtained between the circulating transferrin receptor level and the degree of mild iron deficiency (Shih et al. 1990). This assay can now be used clinically and will lead to less costly, faster, and probably more accurate assessments of low grade iron deficiency, a very prevalent clinical condition. In an era when biomedical scientists are constantly being reminded to work in applied areas to provide for better health care for the here and now, it is important to point out that these new approaches originated entirely in curiosity-driven research for the then and after. Curiosity-driven research tends to be somewhat dCclasst in the present climate, which favours developmental work t o bring new tests and treatments to the market. Given the latter's importance, it still behooves us not to lose sight of the future. As a historical example, one can cite the Faraday/ Davy phenomenon. Had governments in Faraday's time been concerned with improved methods for mine safety, no doubt they would have supported Davy to build bigger and better safety lamps. Money for research into electricity would not have been considered top priority, nor to have any value for mine safety or beyond. Faraday was probably well aware of the situation. When asked by a Minister of the Crown of what use was electricity, he replied that it could one day be taxed. Far too often most of us d o not immediately see the societal implications of new discoveries. Far too often, however, the next generation cannot imagine life without the developments generated from nonmission oriented research. Acknowledgements It is a pleasure to thank my associates whose work this article reviews; my graduate students B.T. Pan, M. Adam,

L. Orr, J. Ahn, and A. Mathew; my research assistants C. Turbine and F. Nault; K. Teng, the electron microscopist; and Gisele de Souza for the preparation of the manuscript. Special thanks are due to Professor Yoshimitsu Maede DVM, Department of Veterinary Internal Medicine, Faculty of Veterinary Medicine, Hokkaido University, Sapporo, Japan, who provided Figs. 2C and 2D. Permission to publish these photographs is gratefully acknowledged. The work was made possible by continuous support from the Medical Research Council of Canada, and latterly, the National Institutes of Health (U.S.A.).

Adam, M., and Johnstone, R.M. 1987. Protein kinase C does not phosphorylate the externalized form of the transferrin receptor. Biochem. J. 242: 151-161. Ahn, J., and Johnstone, R.M. 1989. Maturation-association loss and incomplete de novo synthesis of the transferrin receptor in peripheral sheep reticulocytes. J. Cell. Physiol. 140: 107-118. Blostein, R., and Grafova, E. 1987. Characterization of membrane transport losses during reticulocyte maturation. Biochem. Cell Biol. 65: 869-875. Blostein, R., Drapeau, P., Benderoff, S., and Weigensberg, A.M. 1983. Changes in Na+-ATPase and Na,K pump during maturation of sheep reticulocytes. Can. J. Biochem. Cell Biol. 61: 23-28. Cass, C.E., and Paterson, A.R.P. 1972. Mediated transport of nucleosides in human erythrocytes. J. Biol. Chem. 247: 3314-3320.

Cushman, S.W., and Wardzala, L.J. 1980. Potential mechanism of insulin action on glucose transport in the isolated rat adipose cell. J. Biol. Chem. 255: 4758-4762. Davis, J.Q., Dansereau, D., Johnstone, R.M., and Bennett, V. 1986. Selective externalization of an ATP-binding protein structurally related to the clathrin-uncoating ATPase. J. Biol. Chem. 261: 5368-5371. Foxwell, B.M., Jr., and Tanner, M.J.A. 1981. Synthesis of the erythrocyte anion transport protein. Biochem. J. 195: 129-137. Frazier, J.L., Caskey, J.H., Joffe, M., and Seligman, P.A. 1982. Studies of the transferrin receptor on both human reticulocytes and nucleated cells in vitro. J. Clin. Invest. 69: 853-865. Gasko, 0.. and Danon, D. 1974. Endocytosis and exocytosis in membrane remodelling during reticulocyte maturation. Br. J. Haematol. 28: 463-470. Gavin, J.R., Roth, J., Neville, D.M., Jr., et al. 1974. Insulin dependent of regulation of insulin receptor concentrations. Proc. Natl. Acad. Sci. U.S.A. 71: 84-88. Ginsberg, B.H., and Brown, T.J. 1982. Regulation of insulin receptors in erythrocyte cells. Metab. Clin. Exp. 31: 728-732. Goldberg, A.L., and St. John, A.C. 1976. Intracellular protein degradation in mammalian and bacterial cells. Annu. Rev. Biochem. 45: 747-803. Gronowicz, G., Swift, H., and Steck, T.L. 1984. Maturation of reticulocytes in vitro. J. Cell Sci. 71: 177-197. Harding, C., Heuser, J., and Stahl, P. 1983. Receptor mediated endocytosis of transferrin and recycling of the transferrin receptor in rat reticulocytes. J. Cell Biol. 97: 329-339. Harding, C., Heuser, J., and Stahl, P. 1984. Endocytosis and intracellular processing of transferrin and colloidal gold transferrin. Eur. J. Cell Biol. 35: 256-263. Hershko, A., and Cienchanover, A. 1982. Mechanisms of intracellular protein breakdown. Annu. Rev. Biochem. 51: 335-364.

Im, J.H., Zeidler, R.B., Rackley, C.E., and Kim, H.D. 1984. Developmental changes in insulin receptors of pig red blood cells. Arch. Biochem. Biophys. 232: 26-37. Jandl, J.H., and Katz, J.H. 1963. The cell to plasma transferrin cycle. J. Clin. Invest. 42: 314-326. Jarvis, S.M., and Young, J.D. 1982. Nucleoside transport in sheep

190


reticulocytes and fetal erythrocytes: a proposed model for the nucleoside transporter. J. Physiol. (London), 324: 47-66. Jay, D., and Cantley, L. 1986. The anion transport protein of the red cell. Annu. Rev. Biochem. 55: 511-538. Johnstone, R.M., Adam, M., and Pan, B.T. 1984. The fate of the transferrin receptor during maturation of sheep reticulocytes. Can. J. Biochem. Cell Biol. 62: 1246-1254. Johnstone, R.M., Adam, M., Hammond, J.R., et al. 1987. Vesicle formation during reticulocyte maturation. J. Biol. Chem. 262:


9412-9420.

Johnstone, R.M., Bianchini, A., and Teng, K. 1989. Reticulocyte maturation and exosome release: transferrin receptor containing exosomes show multiple plasma membrane functions. Blood, 74: 1844-1851.

Johnstone, R.M., Mathew, A., Mason, A.B., and Teng, K. 1991. Exosome formation during maturation of mammalian and avian reticulocytes: evidence that exosome release is a major route for externalization of obsolete membrane proteins. J. Cell. Physiol. 147: 27-36. Kohgo, Y., Nishisato, T., Kondo, H., et al. 1986. Circulating transferrin receptor in human serum. Br. J. Haematol. 64: 277-281.

Kohgo, Y., Niitsu, Y., Kondo, H., et al. 1987. Serum transferrin receptor as a new index of erythropoiesis. Blood, 70: 1955-1958. Laris, P.C. 1958. Permeability and utilization of glucose in mammalian erythrocytes. J. Cell. Comp. Physiol. 51: 47-66. Leibman, A., and Aisin, P. 1977. Transferrin receptor of the rabbit reticulocyte. Biochemistry, 16: 1268-1272. Limbird, L.E., Gill, D.M., Stadel, J.M., et al. 1980. Loss of /3 adrenergic receptor - guanine nucleotide regulatory protein interactions accompanies decline in catecholamine responsiveness of adenylate cyclase in maturing rat erythrocytes. J. Biol. Chem. 255: 1854-1861. Lu, L., and Johnstone, R.M. 1992. Evidence for an internal pool of nucleoside transporters in mammalian reticulocytes. Biochim. Biophys. Acta. In press. Maede, Y., and Inaba, M. 1985. (NA-K) ATPase and ouabhin binding in reticulocytes from dogs with high K and low K erythrocytes and their changes during maturation. J. Biol. Chem. 260: 3337-3343. Magnani, M., Stocchi, V., DachA, M., and Fornaini, G. 1984. Rabbit red blood cell hexokinase: intracellular distribution during reticulocyte maturation. Mol. Cell. Biochem. 63: 59-65. Montandon, J.B., and Porzig, H. 1984. Changes in /3 adreno receptor binding properties and receptor cyclase coupling during in vitro maturation of rat reticulocytes. J. Recept. Res. 4(1-6): 91-102.

Orr, L., Adam, M., and Johnstone, R.M. 1987. Externalization of membrane-bound activities during sheep reticulocyte maturation is temperature and ATP dependent. Biochem. Cell Biol. 65: 1080-1090.

Pan, B.T., and Johnstone, R.M. 1983. Fate of the transferrin receptor during maturation of sheep reticulocytes in vitro: selective externalization of the receptor. Cell, 33: 967-977. Pan, B.T., and Johnstone, R.M. 1984. Selective externalization of the transferrin receptor by sheep reticulocytes in vitro. J. Biol. Chem. 259: 9776-9782. Pan, B.T., Teng, K., Wu, C., et al. 1985. Electron microscopic evidence for externalizationof the transferrin receptor in vesicular form in sheep reticulocytes. J. Cell Biol. 10: 942-948. Pelham, H. 1986. Speculations on the function of the major heat shock and glucose regulated proteins. Cell, 46: 959-961. Pollock, L.R., Tate, E.H., and Cook, J.S. 1981. Turnover and regulation of Na K ATPase. Am. J. Physiol. 241: C173-C183. Porzig, H., Moudry, R., and Montandon, J.B. 1991. Analysis by cell hybridization of mechanisms that regulate /3 adrenergic responses in reticulocytes and differentiating erythroid cells. J. Cell. Physiol. 147: 439-446. Rapoport, S.M. 1986. The reticulocyte. CRC Press, Boca Raton, Fla. Raviv, O., Heller, H., and Hershko, A. 1987. Alterations in components of ubiquitin-protein ligase system following maturation of reticulocytes to erythrocytes. Biochem. Biophys. Res. Commun. 145: 658-665. Shih, Y.J., Baynes, R.D., Hudson, B.G., et al. 1990. Serum transferrin receptor is a truncated form of the tissue receptor. J. Biol. Chem. 265: 19 077 - 19 081. Simpson, C.F., and Kling, J.M. 1968. The mechanism of mitochondria1 extrusion from phenylhydrazine-induced reticulocytes in the circulating blood. J. Cell Biol. 36: 103-109. Suzuki, T., and Dale, G.L. 1988. Senescent erythrocytes. Isolation of in vivo aged cells and their biochemical characteristics. Proc. Natl. Acad. Sci. U.S.A. 85: 1647-1651. Thomopoulos, P., Berthellier, M., and Laudet, M.H. 1978. Loss of insulin receptors on maturation of reticulocytes. Biochem. Biophys. Res. Commun. 85: 1460-1465. VanBockxmeer, F.M., and Morgan, E.H. 1979. Transferrin receptors during rabbit reticulocyte maturation. Biochim. Biophys. Acta, 584: 76-83. Watts, R.P., Brendel, K., Luthra, M., and Kim, H.D. 1979. Inosine as a possible energy source for pig red blood cells. Life Sci. 25: 1577-1582. Zanner, M.A., and Galey, W.R. 1985. Aged human erythrocytes exhibit increased anion exchange. Biochim. Biophys. Acta, 818: 310-315.

Zeidler, R.B., and Kim, H.D. 1982. Pig reticulocytes. IV. In vitro maturation of naturally occurring reticulocytes with permeability loss to glucose. J. Cell. Physiol. 112: 360-366. Zweig, S.E., Tokuyasu, T., and Singer, S.J. 1981. Membrane associated change during erythropoiesis. J. Suprarnol. Struct. 17: 163-181.

The 10th Max Grob Memorial Lecture--Zurich, September 5, 1991.

Looking at burn wounds: the A. B. Wallace Memorial Lecture 1991.

Exosome formation during maturation of mammalian and avian reticulocytes: evidence that exosome release is a major route for externalization of obsolete membrane proteins.

Maturation of rabbit reticulocytes: degradation of specific reticulocyte proteins.

Progress against cancer: an epidemiologic assessment. The 1991 John C. Cassel Memorial Lecture.

THE GORDON WILSON MEMORIAL LECTURE.

The Gordon Wilson Memorial Lecture.

Masahiro Iio Memorial Lecture.

THE GORDON WILSON MEMORIAL LECTURE.

Membrane proteins synthesized by rabbit reticulocytes.

The Gillies Memorial Lecture 1975.

Gordon Richards Memorial Lecture.

Review lecture. Bayliss-Starling Memorial Lecture 1976. Lymph formation by secretion or filtration?

Phosphorylation and dephosphorylation of membrane proteins as a possible mechanism for structural rearrangement of membrane components.

The Cairns memorial lecture: the Cairns tradition.

Gordon Memorial Lecture. Chicken neoplasia--a model for cancer research.

The Twenty-Seventh Long Fox Memorial Lecture: Air as a Conductor of Electricity.

Pancreatic Cancer Exosomes: Shedding Off for a Meaningful Journey.

The Long Fox Memorial Lecture: The Ætiology of Cardiac Disease.

The Long Fox Memorial Lecture: The Reality of Delusions.

The 42nd David A. Karnofsky Memorial Award Lecture: Understanding.

Berlyne Memorial Lecture Acquiring the uses of language.

Burn wound healing: James Ellsworth Laing memorial essay for 1991.

Hayes Martin Memorial Lecture. Non desperandum.