| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cell Biology and Metabolism Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892-5430, USA
1Correspondence: CBMB, NICHD, National Institutes of Health, Bldg. 18T Room 101, 18 Library Dr. MSC 5430, Bethesda, MD 20892-5430. USA. E-mail: juan{at}helix.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONCHARACTERISTICS OF LYSOSOME...GENETIC DISORDERS OF LYSOSOME...BIOGENESIS OF LYSOSOME-RELATED...CONCLUDING REMARKSREFERENCES |
|---|
Key Words: melanosome • lytic granule • MIIC • platelet-dense granule • basophil granule • azurophil granule • pigment granule • Chediak-Higashi syndrome • Griscelli syndrome • Hermansky-Pudlak syndrome • protein trafficking
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF LYSOSOME...GENETIC DISORDERS OF LYSOSOME...BIOGENESIS OF LYSOSOME-RELATED...CONCLUDING REMARKSREFERENCES |
|---|
2
3)
. Both endogenous and exogenous macromolecules can
be delivered to lysosomes through the biosynthetic and endocytic
pathways, respectively. In addition, lysosomes can degrade proteins
transported from the cytosol (4)
. The degradative function
of these organelles is carried out by more than 50 acid-dependent
hydrolases (e.g., proteases, lipases, glycosidases) contained within
its lumen (1)
. The limiting membrane of the lysosome
contains a set of highly glycosylated, lysosomal-associated membrane
proteins (LAMPs) such as LAMP-1, LAMP-2, and CD63/LAMP-3, the functions
of which are still unclear. Additional lysosomal membrane proteins
mediate transport of ions, amino acids, and other solutes across the
lysosomal membrane and contribute to the maintenance of an acidic
luminal pH in the range of 4.6–5.0 (5)
.
Lysosomes are morphologically heterogeneous, often resembling other
organelles of the endocytic and secretory pathways. Therefore, they are
currently distinguished from other organelles on the basis of an
operational definition, which describes them as membrane-bound acidic
organelles that contain mature acid-dependent hydrolases and LAMPs but
lack mannose 6-phosphate receptors (MPRs) (2)
. Most or all
of these characteristics are shared with a group of cell type-specific
organelles that includes melanosomes, lytic granules, major
histocompatibility complex (MHC) class II compartments (MIICs),
platelet-dense granules, basophil granules, and neutrophil azurophil
granules (Table 1
). These shared traits suggest that these specialized organelles may be
biogenetically related to lysosomes, a relationship that has been
further illuminated by recent studies on multiorganellar genetic
disorders such as the Chediak-Higashi and Hermansky-Pudlak syndromes.
|
In this study, we review the properties of lysosome-related organelles, genetic evidence that supports a common organellar lineage with lysosomes, and current ideas about their biogenesis.
|
CHARACTERISTICS OF LYSOSOME-RELATED ORGANELLES |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF LYSOSOME...GENETIC DISORDERS OF LYSOSOME...BIOGENESIS OF LYSOSOME-RELATED...CONCLUDING REMARKSREFERENCES |
|---|
) are the site of synthesis and storage of a group of
related pigments known as melanins. In mammals, these pigments are the
primary determinant of skin and hair color, as well as a major source
of protection from the harmful effects of ultraviolet radiation
(8)
. Melanosomes are generated within melanocytes and
retinal pigment epithelial cells, which are specialized cells derived
from the embryonic neural tube. Morphological studies have led to a
classification of melanosomes into four maturation stages (I-IV), with
stages I and IV representing the most immature and mature forms,
respectively (8)
. Mature melanosomes can either remain in
the melanocyte or be transferred to skin keratinocytes by an as yet
poorly understood mechanism (9)
.
|
Melanosomes share a number of features with lysosomes (Table 1)
,
including the presence of both soluble and transmembrane lysosomal
proteins, an acidic luminal environment, accessibility to endocytic
tracers, and the ability to fuse with phagosomes (8)
. In
addition, they contain several melanosome-specific transmembrane
glycoproteins that are directly involved in melanin biosynthesis,
namely, the enzymes tyrosinase (10)
, tyrosinase-related
protein 1 (TRP-1) (11)
and tyrosinase-related protein 2
(TRP-2) (12)
. Another melanosomal protein, the
pmel17/silver gene product, probably contributes to the
formation of an striated luminal matrix that is characteristic of
melanosomes (13)
. Finally, the products of the
OCA1 and OCA2/pink-eyed dilution genes
have also been localized to the melanosomal membrane
(14
15
16)
.
Lytic granules
The targeted secretion of macromolecules from specialized
organelles known as lytic granules (Fig. 1C
) is a major
mechanism by which cytotoxic T lymphocytes and natural killer cells
destroy virus-infected or tumor cells. These electron-dense organelles
maintain an acidic pH, are accessible to endocytic tracers, and contain
both soluble and transmembrane lysosomal proteins (Table 1)
(17)
. In addition, they accumulate specific luminal
proteins involved in cell lysis, namely, the pore-forming protein
perforin and a group of serine proteases known as granzymes, which have
caspase-like or trypsin-like specificities (e.g., granzymes A, B, H,
and M). Specific transmembrane proteins of lytic granules include
CTLA-4 and Fas ligand, which are modulators of the immune response
(18
, 19)
, and GMP-17, a multi-spanning membrane protein of
unknown function (20)
.
MIICs
Professional antigen-presenting cells (i.e., macrophages,
dendritic epithelial cells, and B lymphoblasts) express surface MHC
class II-peptide complexes, which enable activation of
CD4+ helper T cells. Binding of peptides to MHC
class II molecules is thought to occur in pleiomorphic organelles
termed MIICs (Fig. 1D
). MIICs are accessible through both
the endocytic pathway, which provides peptides derived from the
degradation of internalized antigens, and the biosynthetic pathway,
which provides the newly synthesized MHC class II molecules (21
, 22)
. These organelles have several characteristics in common
with conventional lysosomes (Table 1)
, including the presence of both
soluble and transmembrane lysosomal proteins and the absence of MPRs
(23)
. In addition to MHC class II molecules and lysosomal
proteins, MIICs are enriched in HLA-DM (24)
and HLA-DO
(25)
, two proteins that regulate association of
antigen-derived peptides to the MHC class II molecules
(26)
. Dendritic-cell MIICs also transiently contain
DC-LAMP, a LAMP-like glycoprotein of unknown function
(27)
.
Platelet dense granules
Blood platelets are anucleate bodies that are derived from bone
marrow megakaryocytes and play a central role in hemostasis and
thrombosis. They contain three main types of secretory granules:
-granules, dense granules, and lysosomes (28)
. Dense
granules (also called 
-granules or dense bodies) have a highly
condensed core that consists of serotonin, calcium, ATP, ADP, and
pyrophosphate and allows these granules to be readily detected by
whole-mount electron microscopy (Fig. 1E
). Secretion of
dense granules is a critical event in the formation of the hemostatic
plug.
Although the lumen of platelet-dense granules is less acidic than
conventional lysosomes and appears to be devoid of lysosomal hydrolases
(Table 1)
, there is growing evidence that these organelles belong to
the lysosomal lineage. Indeed, the dense granule membrane is enriched
in the lysosomal proteins CD63/LAMP-3 and LAMP-2 (29
, 30
;
but see also ref 31
). Moreover, genetic disorders that
affect the biogenesis of melanosomes and lysosomes also result in
platelet-dense granule deficiency (Table 1)
.
In addition to the LAMPs, the dense granule membrane contains a
specific H+ pump (32)
and a
serotonin transporter (33)
, both of which are involved in
serotonin uptake. Other transmembrane proteins found in platelet-dense
granules—P-selectin (29)
, GPIb, and
IIb/ß3 integrin
(34)
—are thought to act as receptors for adhesive
proteins and mediate platelet aggregation (28)
.
Basophil granules
Basophils, together with mast cells, act as the major cellular
mediators of inflammation associated with allergic disease. These cells
have secretory granules that contain various effectors of inflammation
including histamine, serotonin, heparin, and the neutral proteases
tryptase and chymase (35)
. In addition, basophil granules
contain lysosomal hydrolases (36
, 37)
and lysosomal
membrane proteins such as LAMP-1, LAMP-2, CD63/LAMP-3, and LIMP IV/5G10
antigen (38
39
40)
. Activation of high-affinity receptors
for immunoglobulin E or chemotactic receptors results in exocytic
release of the contents of these granules (35)
, leading to
the exposure of lysosomal membrane proteins on the surface of the
activated cells (38
39
40)
. These surface-exposed lysosomal
membrane proteins are rapidly internalized via clathrin-coated pits and
partly recycled to secretory granules (38
, 39)
.
Neutrophil azurophil granules
Neutrophils are phagocytic cells that circulate in peripheral
blood and play a central role in defense against invading bacteria.
They contain a number of secretory granules that are generally
classified into azurophil, specific, and gelatinase granules
(41)
. Azurophil granules, also known as primary granules,
contain microbicidal polypeptides such as myeloperoxidase, bactericidal
permeability-increasing protein, defensins, and azurocidin
(41)
. These organelles have long been thought to represent
the neutrophil lysosomes, as they are the major cellular reservoir of
lysosomal acid-dependent hydrolases (42)
. In addition,
azurophil granules have been shown to contain the lysosomal membrane
proteins CD63/LAMP-3 (43)
and CD68 (44)
and
to be accessible to endocytosed fluid phase markers under conditions of
cellular stimulation (45)
. However, the idea that these
granules are closely related to lysosomes has been challenged by recent
studies demonstrating that azurophil granules are not enriched in
LAMP-1 or LAMP-2 (46
, 47)
. Still, azurophil granules are
abnormal in patients with Chediak-Higashi syndrome (48)
, a
disorder that affects both lysosomes and lysosome-related organelles
(Table 1)
.
Drosophila pigment cell granules
In the fruit fly, Drosophila melanogaster, the eye
consists of ~800 functional units termed ommatidia, with each
ommatidium comprising a group of eight photoreceptor cells surrounded
by a sheath of pigment cells (49)
. Pigment cells provide
optical isolation of each ommatidium and, similarly to mammalian
melanocytes, store pigments in membrane-bound compartments (i.e.,
pigment granules; Fig. 1B
). The pigments accumulated in
these granules are chemically unrelated to melanins and fall into two
major classes: drosopterins (red) and ommochromes (brown). Little is
known about the protein composition of pigment cell granules.
Presumably, they contain the products of genes involved in either the
uptake or processing of pigment precursors (e.g., white,
purple, vermilion; ref 50
). Evidence
that Drosophila pigment granules are related to lysosomes
has been largely obtained from genetic studies (discussed below).
|
GENETIC DISORDERS OF LYSOSOME-RELATED ORGANELLES |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF LYSOSOME...GENETIC DISORDERS OF LYSOSOME...BIOGENESIS OF LYSOSOME-RELATED...CONCLUDING REMARKSREFERENCES |
|---|
) and of mutations that affect eye pigmentation in
Drosophila. Among the first group are autosomal, recessive
disorders that affect the biogenesis and/or function of lysosomes,
melanosomes, and one or more of the remaining lysosome-related
organelles. In humans, these disorders include the Chediak-Higashi and
Hermansky-Pudlak syndromes. The following sections summarize the main
characteristics of these disorders and recent advances in our
understanding of their molecular bases. A subsequent section is devoted
to Drosophila pigmentation genes of the so-called ‘granule
group’ and their relationship to lysosomal protein transport.
|
Chediak-Higashi syndrome
In humans, Chediak-Higashi syndrome is characterized clinically by
variable hypopigmentation of skin, hair and eyes, bleeding diathesis
(due to platelet-dense granule deficiency), progressive neurological
dysfunction, and severe immunological deficiency (9
, 51)
.
The immunological deficiency is observed in ~90% of the cases, is
associated with neutropenia and defective natural killer cell function,
and leads to an accelerated lymphoproliferative phase that is often
fatal. The remaining ~10% of the cases have few or no infections,
but slowly develop serious neurological symptoms such as mental
deficiency, seizures, and tremor (51)
.
A characteristic feature of Chediak-Higashi syndrome at the cellular
level is the presence of giant lysosomes, melanosomes, MIICs, lytic
granules, and azurophil granules (48
, 51
, 52)
.
Platelet-dense granules, on the other hand, are either absent or fewer
in number in circulating blood platelets (53)
. A gene
associated with Chediak-Higashi syndrome in humans (CHS1)
has been cloned and found to encode a novel ~430-kDa protein
(54
, 55)
. The murine ortholog of CHS1 is
mutated in the beige mouse strain (54
, 56)
,
which has long been regarded as a model for the syndrome (Table 2)
. The
protein encoded by this gene, named LYST or Beige, has been detected in
a wide variety of cell types and is localized to the cytosol
(57)
, although association with microtubules has also been
reported (52)
.
The biological function(s) of the LYST/Beige protein remains
incompletely understood. The ‘giant-lysosome’ phenotype, also
observed in beige mouse fibroblasts, could be reverted on
expression of the wild-type LYST/Beige protein in these cells
(57)
. Notably, overexpression of LYST/Beige in normal
fibroblasts resulted in abnormally small lysosomes (57)
.
Together, these results argue for a role of this protein in
fusion/fission events that determine the size of lysosomes and related
organelles. Another study has shown abnormal trafficking of
lysosomal/MIIC membrane proteins in B-lymphoblasts derived from
patients with Chediak-Higashi syndrome, thus leading to the proposal
that LYST/Beige functions in protein transport to late endosomal
compartments (52)
. As both sets of observations may
reflect different aspects of LYST/Beige function, it is clear that
further experiments will be required to elucidate its role in organelle
biogenesis.
Griscelli syndrome and related disorders
At least three additional recessive disorders resembling
Chediak-Higashi syndrome have been described in humans: Griscelli
syndrome (58)
, neuroectodermal melanolysosomal disease
(59)
, and partial albinism and immunodeficiency (PAID)
syndrome (60)
. All three disorders involve
hypopigmentation of the skin and silvery gray hair, presumably due to
impaired transfer of melanosomes from melanocytes to keratinocytes. On
the other hand, none of these disorders is associated with a deficiency
in platelet-dense granules. Griscelli and PAID syndromes also exhibit
cellular immunodeficiency. In Griscelli syndrome, the immunodeficiency
is likely to result, at least in part, from a dysfunction of secretory
lysosomes (e.g., lytic granules) and can lead to an accelerated
lymphoproliferative phase similar to that observed in Chediak-Higashi
syndrome (51)
. Patients with PAID syndrome or
neuroectodermal melanolysosomal disease, but not with Griscelli
syndrome, also display progressive dysfunction of the central nervous
system. Abnormal organelle morphology has only been reported for
neuroectodermal melanolysosomal disease (59)
.
Although these three disorders have been described as distinct clinical
entities, circumstantial evidence suggests that they may be allelic. A
common candidate gene is MYO5A, which encodes a
nonconventional myosin (myosin Va) and is the ortholog of the gene
associated with the dilute mutation in mouse (Table 2)
(61)
. Dilute mutant mice exhibit
hypopigmentation due to impaired accumulation of mature melanosomes at
the dendritic tips of melanocytes (from where the melanosomes are
transferred to keratinocytes), thus suggesting a role for myosin Va in
intracellular melanosome translocation (62)
. Mutations in
MYO5A have been detected in a human patient suffering from
Griscelli syndrome, as well as in a patient with symptoms that could
correspond to Griscelli syndrome, neuroectodermal melanolysosomal
disease, or PAID syndrome (63
, but see also ref
64
).
Hermansky-Pudlak syndrome
Hermansky-Pudlak syndrome (65)
is an autosomal,
recessive disorder characterized by oculocutaneous albinism and
prolonged bleeding, due to abnormal melanosomes and apparent absence of
dense granules from blood platelets, respectively. In addition,
patients suffering from this syndrome display progressive accumulation
of partially degraded proteolipids in lysosomes (i.e., ceroid
lipofuscinosis), which can eventually lead to death due to
complications such as restrictive pulmonary fibrosis (66)
.
Hermansky-Pudlak syndrome thus may arise from mutations that affect the
biogenesis and/or function of lysosomes and at least two
lysosome-related organelles, namely, melanosomes and platelet-dense
granules (Table 1)
. The mutations that cause Hermansky-Pudlak syndrome
have only recently begun to be identified. It is now clear that there
are different variants of the disorder that are associated with
distinct gene loci. Described below are the variants for which the
affected genes have been identified and characterized.
Hermansky-Pudlak syndrome type 1
The gene responsible for this form of the disease,
HPS1, was identified by positional cloning and found to
encode a ubiquitously expressed protein with no homology to any known
protein (67)
. The murine ortholog of HPS1 was
subsequently found to be mutated in the pale ear mutant
mouse (68
, 69)
, which is a model for the disease (Table 2)
. Recent biochemical analyses have established that HPS1p is a ~79
kDa cytosolic protein capable of associating peripherally with
membranes (70
, 71)
.
While giant lysosomes and melanosomes are considered a cellular
hallmark for Chediak-Higashi syndrome, enlarged melanosomes have also
been described for pale ear melanocytes (68)
.
It is not clear, however, whether the same holds true for
HPS1p-deficient melanocytes derived from patients with Hermansky-Pudlak
syndrome type 1 (72)
. Another interesting observation is
that ammonia-induced secretion of lysosomal hydrolases is defective in
pale ear fibroblasts (73)
. Although the exact
biological function of HPS1p remains elusive, all of the observations
described above suggest that it plays a role in the biogenesis of both
lysosomes and lysosome-related organelles.
Hermansky-Pudlak syndrome type 2
Soon after the identification of HPS1, it became
apparent that a subset of patients with Hermansky-Pudlak syndrome do
not bear mutations on this gene (74
, 75)
. Using a
candidate gene approach, a second gene associated with this disorder
was identified (76)
. The gene, referred to as
ADTB3A, encodes the ß3A subunit of the heterotetrameric
protein complex AP-3, which belongs to a family of adaptor protein (AP)
complexes involved in protein trafficking (77
, 78)
. The
murine ortholog of ADTB3A was found to be mutated in the
pearl mouse strain (Table 2)
, which displays a phenotype
similar to Hermansky-Pudlak syndrome (79)
. A yeast
counterpart for AP-3 has been identified and implicated in protein
transport to the vacuole (i.e., to the yeast equivalent of the
mammalian lysosome) (80
, 81)
. Mammalian AP-3 has been
shown to interact in vitro with the vesicle-forming protein
clathrin (82)
and with both tyrosine-based (83
, 84)
and dileucine-based sorting signals (85)
.
Consistent with a role for AP-3 in protein trafficking, AP-3-deficient
cells displayed abnormally enhanced trafficking of lysosomal membrane
proteins through the plasma membrane (70
, 76
, 86)
.
Therefore, AP-3 may mediate the trafficking of a subset of integral
membrane proteins from an intracellular site to lysosomes and,
presumably, to melanosomes and platelet-dense granules. However, not
all integral membrane proteins use AP-3 for targeting to
lysosome-related organelles, as suggested by the apparently normal
trafficking of MHC class II molecules and the associated invariant
chain to MHCs in AP-3-deficient cells (87)
.
Additional mouse models of Hermansky-Pudlak syndrome
At least 14 different murine gene loci have been associated with
phenotypes related to Hermansky-Pudlak syndrome. As mentioned above,
two of them, responsible for the pale ear and
pearl mutations, represent the murine counterparts of
HPS1 and ADTB3A, respectively (Table 2)
. Another
locus, responsible for the mocha mutation, encodes the
subunit of murine AP-3 (88)
. Consistent with the
trafficking phenotype reported for AP-3 ß3A-mutant cells,
mocha fibroblasts exhibit increased trafficking of lysosomal
membrane proteins via the cell surface (70)
. In addition
to the melanosome and platelet-dense granule defects, mocha
mice display neurological abnormalities such as hyperactivity and
seizures.
Another recently identified gene is mutated in the pallid
mouse strain (89)
. This gene encodes a soluble ~25-kDa
protein, named pallidin, which exhibits no significant homology to any
previously characterized protein (note that this recently described
pallidin protein is distinct from pallidin/protein 4.2, the product of
a previous candidate gene; ref 90
). Like the
mocha mice, pallid mice display a marked
neurological phenotype in addition to the pigmentation and
platelet-dense granule defects. A possible role for pallidin in vesicle
trafficking has been suggested on the basis of its interaction with
syntaxin 13 (89)
, a known component of the machinery that
mediates membrane fusion (91)
.
Finally, the gunmetal mouse strain has recently been shown
to carry a mutation in the subunit of Rab geranylgeranyl
transferase, an enzyme that adds prenyl groups to the carboxyl termini
of Rab GTPases (92)
. The mutation results in decreased
prenylation and membrane association of Rab27a (92)
. Since
Rabs are involved in protein trafficking (93)
, decreased
association of Rab27a or other Rabs with membranes could result in
impaired transport of proteins to lysosome-related organelles.
Mutations in Drosophila pigment granule genes
Over 80 mutations affecting Drosophila eye color have
been described and classified into four major groups: 1)
those affecting the biosynthesis of drosopterin pigments, 2)
those affecting the biosynthesis of ommochrome pigments, 3)
those associated with ABC transporters involved in transport of pigment
precursors, and 4) those affecting genes of the so-called
‘granule group’ (50)
. The latter group of mutations
causes a reduction in both types of Drosophila pigments, as
well as additional phenotypes not restricted to eye color. Strikingly,
the recent identification of seven granule group genes has revealed a
close relationship to genes involved in vacuolar trafficking in the
yeast, Saccharomyces cerevisiae (Table 3
), and lysosomal trafficking in other organisms.
|
Before AP-3 subunit mutations were identified in mammals, the
Drosophila garnet gene was found to encode the
subunit of AP-3 (94
, 95)
. Subsequently, additional
genes of the granule group, namely, ruby,
carmine, and orange, were found to encode the
remaining ß3, µ3, and 
3 subunits of AP-3, respectively
(96
, 96a)
. As discussed above, both the mammalian and
yeast AP-3 complexes mediate trafficking of integral membrane proteins
to lysosomal organelles. As Drosophila AP-3 is also likely
to be involved in lysosomal transport, the pigmentation phenotype of
AP-3 mutants supports the idea that, like mammalian melanosomes,
Drosophila pigment granules are related to lysosomes.
In addition to the genes encoding AP-3 subunits, three other genes of
the granule group encode Drosophila homologs of yeast
vacuolar protein sorting (VPS) gene products (Table 3)
. The product of
the Drosophila light gene (97)
is an
ortholog of yeast Vps41p, which has recently been shown to interact
with the subunit of AP-3 and to be required for the formation of
AP-3-coated carrier vesicles (98)
. In addition, the
products of the Drosophila deep orange and
carnation genes are highly similar to yeast Vps18p and
Vps33p, respectively (99
, 100)
. In yeast, Vps18p and
Vps33p associate into a multisubunit protein complex that also contains
Vps11p and Vps16p (101)
. Mutations in any of these
subunits results in accumulation of prevacuolar multivesicular bodies
and impaired transport to the vacuole from both the biosynthetic and
endocytic pathways (101)
. Similarly, Sevrioukov et al.
(100)
have shown that the products of deep
orange and carnation are associated into a large
complex, and that mutations in deep orange result in
accumulation of multivesicular bodies and defective trafficking of an
internalized ligand to lysosomes. Therefore, the pathways for protein
trafficking to the vacuole in yeast, and to lysosomes and pigment
granules in Drosophila, appear to be highly conserved.
|
BIOGENESIS OF LYSOSOME-RELATED ORGANELLES |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF LYSOSOME...GENETIC DISORDERS OF LYSOSOME...BIOGENESIS OF LYSOSOME-RELATED...CONCLUDING REMARKSREFERENCES |
|---|
Do lysosomes and lysosome-related organelles coexist in the same
cell?
Before attempting an explanation of how lysosome-related
organelles are formed, it is pertinent to ask whether cells that
contain lysosome-related organelles also contain a population of
conventional lysosomes. This question is crucial to explain the
biogenesis of lysosome-related organelles, since if lysosome-related
organelles were nothing but the lysosomes of particular cell types,
then no special sorting events would be needed to generate them. In
such a scenario, lysosome-related organelles would acquire their
special properties by expression of cell type-specific proteins that
are targeted to the organelles by means of conventional lysosomal
trafficking pathways. On the other hand, if a cell generated separate
populations of lysosomes and lysosome-related organelles, then sorting
mechanisms should exist to ensure differential targeting of their
corresponding resident proteins.
The answer to this question appears to vary depending on the cell type
and the corresponding lysosome-related organelle. In cytotoxic T
lymphocytes and natural killer cells, for example, lytic granules have
been shown to constitute the vast majority of lysosomal organelles
(6
, 102)
. In antigen-presenting cells, MIICs were
originally described as compartments distinct from lysosomes (23
, 103
, 104)
. However, recent morphological studies suggest that
MIICs are either identical to lysosomes or correspond to a normal
subpopulation of lysosomes referred to as ‘early lysosomes’ (7
, 105)
. In contrast, platelets have distinct populations of
platelet-dense granules and lysosomes (28)
. Similarly,
melanocyte- or melanoma-derived cell lines appear to have overlapping
but distinct populations of melanosomes and lysosomes (106
, 107)
. A similar situation occurs in a basophilic cell line, in
which histamine-containing granules can be separated from a population
of apparently conventional lysosomes (108)
. These
observations suggest that, although the pathways leading to the
formation of lysosomes and lysosome-related organelles may be shared to
a large extent, there must be a divergence in some cell types, such
that the two types of organelle become partially or totally segregated.
Sorting of luminal proteins to lysosome-related organelles
Sorting of acid-dependent hydrolases to lysosomes is
mediated by mannose 6-phosphate residues on the hydrolases that bind to
either of two membrane-bound MPRs in the trans-Golgi network
(TGN) (2)
. The hydrolases are transported by the receptors
to early endosomes (Fig. 2
, step 1) or late endosomes (Fig. 2
, steps 2 or 3), where they
dissociate from the receptors. The hydrolases are subsequently
delivered to lysosomes (Fig. 2
, step 4) whereas the MPRs are returned
to the TGN for additional rounds of transport (Fig. 2
, step 5). This
mechanism is also likely responsible for the targeting of
acid-dependent hydrolases to lysosome-related organelles. Moreover,
some cell-specific luminal components of lysosome-related
organelles might be sorted by the same mechanism. Examples of this
are the lytic granule-specific proteins granzyme A and granzyme B,
which contain mannose 6-phosphate residues (109)
. These
residues appear to be important for sorting to lytic granules since
cells from patients with I-cell disease, who fail to add mannose
6-phosphate residues, secrete large amounts of granzyme A and granzyme
B into the extracellular medium (109)
.
|
Other cell-specific luminal components are not modified with
mannose 6-phosphate residues and must therefore be sorted by mechanisms
that do not involve the MPRs. Conceivably, these proteins could bear
other sorting signals. In this regard, it is worth mentioning that a
fraction of lysosomal enzymes can be transported to lysosomes
independently of the MPRs, which probably explains why some cell types
in I-cell disease patients (e.g., lymphocytes) store a normal
complement of lysosomal enzymes in lysosomes (110)
. The
nature of this alternative mechanism has not been elucidated. Another
possible mechanism is sorting by aggregation in the TGN, which appears
to be a major means for sorting of hormones, neuropeptides, and enzymes
to nonlysosomal secretory organelles such as endocrine and exocrine
granules (111)
. For instance, the lytic granule protein,
perforin, has been shown to form complexes with sulfated proteoglycans
(112)
, which have been implicated in sorting by
aggregation of other secretory granule contents. This mechanism has
also been proposed to account for the sorting of cell-specific proteins
to azurophil granules (41)
.
Sorting of integral membrane proteins to lysosome-related
organelles
Two pathways for lysosomal targeting have been described
(3)
. The first pathway, referred to as ‘direct’,
involves transport from the TGN to endosomes (Fig. 2
, steps 1 and 3, or
step 2), followed by transport to lysosomes (Fig. 2
, step 4). The
second pathway, referred to as ‘indirect’, entails transport from the
TGN to the plasma membrane (Fig. 2
, step 6), from which the proteins
are internalized into early endosomes (Fig. 2
, step 7) and successively
delivered to late endosomes (Fig. 2
, step 3) and lysosomes (Fig. 2
,
step 4). Different lysosomal membrane proteins may utilize each pathway
to different extents. LAMP-1 and lysosomal acid phosphatase are
examples of proteins believed to use preferentially the direct and
indirect pathways, respectively.
It is now well established that targeting of most integral membrane
proteins to lysosomes is mediated by signals contained within the
cytosolic tails of the proteins (3
, 113
114
115)
. Two types
of lysosomal targeting signals have been well characterized
(Table 4
). The first type, referred to as ‘tyrosine-based signals’, conform to
the motif YXX
(Y is tyrosine, X is any amino acid, and is a
bulky hydrophobic amino acid). The second type, referred to as
‘dileucine-based signals’, contain critical leucine-leucine or
leucine-isoleucine pairs and often an acidic residue four or five
positions upstream of the pair. These signals interact with different
affinities with subunits of AP complexes, which mediate vesicle budding
and cargo recruitment at different stages of lysosomal-targeting
pathways (77
, 78
, 115)
.
|
A growing body of data suggests that some membrane proteins targeted to
lysosome-related organelles use signals and pathways similar to those
described above for lysosomal integral membrane proteins. First,
expression of these proteins in cells that contain lysosomes but not
lysosome-related organelles (e.g., fibroblasts) almost always results
in localization of the expressed proteins to late endosomes or
lysosomes. This has been documented for the melanosomal proteins
tyrosinase (116
, 117)
and TRP-1 (118)
, the
MIIC protein HLA-DM (119)
, and the lytic granule protein
CTLA4 (18)
. Second, the cytosolic tails of these proteins
are sufficient to effect targeting to late endosomes and lysosomes
(116
117
118
119)
. Finally, some of these tails have sequence
motifs resembling tyrosine-based or dileucine-based signals involved in
lysosomal targeting (Table 4)
. For instance, a typical tyrosine-based
sorting signal is involved in targeting HLA-DM to late endosomes and
lysosomes in fibroblasts and to MIICs in antigen-presenting cells
(119
, 120)
. Similarly, a dileucine-based signal in the
cytosolic tail of the melanosomal protein TRP-1 has been shown to
mediate targeting to the endosomal-lysosomal system (118)
.
A dileucine-based signal in the cytosolic tail of tyrosinase also
contributes to its targeting to endosomes and lysosomes in nonpigmented
cells (116
, 117)
. Other melanosomal proteins such as
TRP-2, Pmel17, and the P-protein likewise contain potential
tyrosine-based and/or dileucine-based signals within their cytosolic
tails (118)
, although the importance of these signals for
sorting has not yet been assessed.
Other evidence, however, suggests that melanosomal sorting signals are
not identical to lysosomal targeting signals. For example, the
dileucine-based signal of tyrosinase has been shown to compete with
tyrosine-based signals and to induce enlargement of endosomal
structures, unlike dileucine-based signals from lysosomal proteins,
which do not elicit these effects (116)
. In addition,
TRP-1 and tyrosinase colocalize only partially with LAMP-1 in some
melanoma cells (107)
. These observations suggest that
melanosomal membrane proteins might have additional sorting information
directing transport to melanosomes. Indeed, the sequence similarities
in the cytosolic tails of melanosomal membrane proteins extend beyond
the canonical tyrosine- or dileucine-based signals (116
, 118)
. Comparison of the cytosolic tails of tyrosinase, TRP-1,
and TRP-2 from various animal species has revealed a similar
arrangement of four segments comprising 1) basic residues,
2) a canonical tyrosine- or dileucine-based signal,
3) an acidic/spacer sequence, and 4) a sequence
containing tyrosine and/or acidic residues (116)
. Whether
these additional segments harbor specific melanosomal sorting signals
and how these putative signals might be recognized are issues that need
to be investigated further. We should add that some cell-specific
proteins could be transported to lysosome-related organelles by signals
that do not conform to known consensus motifs, as is the case for
P-selectin (121)
.
On the basis of the observations described above, it is reasonable to
hypothesize that many cell-specific membrane proteins are transported
to lysosome-related organelles along lysosomal targeting pathways. Most
are thought to follow the direct pathway. However, there is evidence
that some of these proteins are expressed at the cell surface, where
they can be internalized and delivered to late endosomes and lysosomes
following the indirect pathway (116)
. The indirect pathway
could allow retrieval of membrane proteins that are exposed on the cell
surface on exocytic release of the organelle contents (38
, 39
, 43)
. At some stage of these lysosomal targeting pathways,
though, some cell-specific membrane proteins may be diverted away from
the path to lysosomes and directed to lysosome-related organelles.
Models for the biogenesis of lysosome-related organelles
Early models of the biogenesis of lysosome-related organelles
postulated that precursors of some of these organelles budded directly
from the TGN (Fig. 2
, step 10). This could be a major mechanism for the
generation of organelle precursors containing proteins that are sorted
by aggregation in the TGN (e.g., perforin, azurophil granule proteins).
The budded vesicles could also carry lysosomal hydrolases and membrane
proteins, as has been shown to occur for immature regulated secretory
granules (122)
. Coated vesicles shuttling between these
precursors and other organelles could either remove or add proteins,
thus contributing to the maturation of the precursors.
The relationship of lysosome-related organelles to lysosomes, however,
suggests an alternative model in which the endosomal system plays a key
role in their biogenesis. As discussed in the previous section, several
proteins that reside in lysosome-related organelles have signals that
allow them to traffic through either the direct (Fig. 2
, steps 1, 3,
and 4, or steps 2 and 4) or indirect (Fig. 2
, steps 6, 7, 3, and 4)
lysosomal targeting pathways. For lysosome-related organelles that
represent the only lysosomal compartment in a particular cell type, as
may be the case for lytic granules, late endosomes containing
cell-specific proteins would simply mature into mixed lytic
granule-lysosomal structures. For other lysosome-related organelles, on
the other hand, specific proteins would need to be segregated from
those of conventional lysosomes, presumably in a late endosomal
compartment (Fig. 2
, step 8). This process could rely on specific
sorting signals and recognition molecules, as described in the previous
section. Alternatively, cell-specific proteins could modify the late
endosomal environment in such a way that physical segregation would
ensue. For example, melanin synthesis likely starts at a premelanosomal
stage (8)
. Since melanosomal proteins are known to
interact with melanin to form insoluble aggregates (8)
, it
is conceivable that vesicle domains containing these aggregates could
be segregated from a precursor compartment to generate more
differentiated melanosomes. Such a segregation process could be
mediated by a specific machinery involved in vesicle fission, of which
the LYST/Beige protein might be a component (57)
. It is
also possible that this segregation would not occur from endosomes but
from lysosomes (Fig. 2
, step 9).
The two basic models discussed above are not mutually exclusive as the mature organelles could arise from fusion of TGN-derived vesicles carrying proteins sorted by aggregation with endosomal/lysosomal structures carrying proteins sorted by specific signals.
Translocation of lysosome-related organelles and fusion with the
plasma membrane
Concurrent with their maturation, some lysosome-related organelles
must acquire specific machineries that allow them to move to their
sites of action. Melanosomes, for example, undergo long-range movement
from the juxtanuclear area to dendrites and dendritic tips of
melanocytes. This movement is mediated by association of the
melanosomes with microtubules (62)
. Once in the cell
periphery, melanosomes are ‘captured’ by interactions mediated by
myosin Va and F-actin, which prevent them from returning to the cell
body (62)
. The melanosomes are then transferred to
keratinocytes. It is currently unknown whether this process involves
phagocytic engulfment of the melanocyte dendrites by keratinocytes or
secretion into the extracellular space followed by uptake.
Maturation of other lysosome-related organelles must also involve
acquisition of a specific machinery that allows them to fuse with the
plasma membrane on stimulation of secretion (Fig. 2
, step 11). Such is
the case for lytic granules, platelet-dense granules, and basophil
granules, which has prompted their classification as ‘secretory
lysosomes’ (102)
. Even MIICs have been shown to undergo
exocytic fusion with the plasma membrane, suggesting that this may be
one pathway by which peptide-loaded MHC class II molecules are deployed
at the cell surface (123)
. By comparison, most
conventional lysosomes appear to be unable to fuse with the plasma
membrane, although a small population of them can be induced to
exocytose their contents by treatment of cells with calcium ionophores
(124)
.
|
CONCLUDING REMARKS |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF LYSOSOME...GENETIC DISORDERS OF LYSOSOME...BIOGENESIS OF LYSOSOME-RELATED...CONCLUDING REMARKSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
and
Mickey Marks (University of Pennsylvania) for critical review of the
manuscript. We apologize to those authors whose work we failed to cite
due to space limitations or oversight. C. H. is supported by a National
Research Council Associateship and S. C. by a fellowship from the Human
Frontiers Science Program Organization. |
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF LYSOSOME...GENETIC DISORDERS OF LYSOSOME...BIOGENESIS OF LYSOSOME-RELATED...CONCLUDING REMARKSREFERENCES |
|---|
lin, C. (1995) The lysosomal membrane glycoproteins Lamp-1 and Lamp-2 are present in mobilizable organelles, but are absent from the azurophil granules of human neutrophils. Biochem. J. 311,667-674
