Biosynthesis of glycosylphosphatidylinositol membrane anchors

Document Sample
Biosynthesis of glycosylphosphatidylinositol membrane anchors Powered By Docstoc
					Biochem. J.   (1995) 310, 361-370 (Printed in Great Britain)                                                                                                     361

Biosynthesis of glycosylphosphatidylinositol membrane anchors
Victoria L. STEVENS
Division of Cancer Biology, Department of Radiation Oncology, and Department of Biochemistry, Emory University School of Medicine, Atlanta, GA 30335, U.S.A.

INTRODUCTION                                                                                   and structural elucidation of candidate GPI precursors [17-20]
                                                                                               as well as studies with isolated trypanosome membranes have
Covalent modification of proteins with phosphatidylinositol (PI)                               established that the glycolipid is synthesized by the sequential
was first suggested as a novel mechanism of membrane attach-                                   addition of the predicted carbohydrates to PI [21-23].
ment when it was found that PI-specific phospholipase C                                           Although GPI-anchored proteins are not as abundant in yeast
(PI-PLC) could release alkaline phosphatase from tissue slices                                 and mammalian cells as in parasites, the availability of mutants
[1-5]. Subsequent compositional analysis of Trypanosoma brucei                                 defective in GPI biosynthesis has made these species invaluable
variant surface glycoprotein (VSG) [6,7], Torpedo acetyl-                                      for the study of this pathway. These mutants are listed in Table
cholinesterase [8], erythrocyte acetylcholinesterase [9] and rat                               1. Particularly useful have been a set of lymphoma cell mutants
brain Thy-l [10] revealed that carbohydrates and ethanolamine                                  deficient in the surface expression of the GPI-anchored T-cell
were also components of the membrane anchor. These findings led                                antigen Thy- 1. These mutants, which were isolated using
to the conclusion that these proteins were attached to the cell                                complement-mediated lysis by Dr. Robert Hyman [24,25], fall
surface by a glycosylphosphatidylinositol (GPI) anchor. Since                                  into six complementation groups. The Thy- I protein is produced
then, over 100 proteins have been found to be GPI-anchored                                     normally by these cells but is not expressed on the cell surface
(see [11-13] for reviews). These proteins exist in all eukaryotic                              because of defects in either an enzyme or another component
species including parasites, Drosophila, yeast and mammals. The                                required for biosynthesis of the GPI anchor [26,27].
functions of the GPI-anchored proteins are also diverse, ranging                                  Our current understanding of the biosynthetic pathway for
from enzymes, coat proteins and cell-adhesion molecules to                                     GPI membrane anchors obtained from studies in parasites, yeast
receptors.                                                                                     and mammalian cells will be discussed in this review. The aspects
   The initial studies of both GPI structure and biosynthesis were                             of this process which are particular for only one or two of these
done with African trypanosomes. Because GPI-anchored VSG                                       species will also be considered. Because structure and biosynthesis
represents approximately 10% of the total cellular protein in this                             are intimately related, the former will also be discussed. However,
parasite [14], GPI biosynthesis is very active and GPI precursors                              the coverage of this topic is not meant to be comprehensive and
and mature anchors are abundant. In vivo radiolabelling studies                                the reader is referred to recent reviews by McConville and
in which GPI was found associated with VSG within 1 min after                                  Ferguson [12], Ferguson [28,29], Englund [30] and Lublin [31] for
translation of the protein established that the glycolipid core is                             further information on GPI structure as well as other topics such
pre-assembled and transferred en bloc to newly synthesized                                     as the attachment of the membrane anchor to protein and its
proteins in the endoplasmic reticulum (ER) [15,16]. The isolation                              function in various cellular processes.

Table 1 Mammalian cell mutants defective in GPI biosynthesis

 Complementation                                                                                           Protein predicted
 group                       Cell line                             Defect                                  from cDNA                               Refs

 A                           S49(Thy-1 -a), JY-5                   GIcNAc-PI synthesis                     484 amino acids; putative C-terminal    58,60,70-74
                                                                                                            transmembrane span
 B                           SIA(Thy-1 -b)                         Addition of the third mannose           554 amino acids                         38,60,97
 C                           IMI(Thy-1-c)                          GIcNAc-PI synthesis                     Not cloned                              58,60
 E                           BW5147(Thy-1 -e), M38/20              Dol-P-Man synthesis                     Dol-P-Man synthase                      91,95
 F                           EL4(Thy-1-f)                          Addition of the terminal                219 amino acids; very hydrophobic       60,66
 H                           S49(Thy-1-h), Ltk-                    GIcNAc-PI synthesis                     188 amino acids; hydrophobic core       58,60,74-76
 J                           K562.1B, K562.lVD                     GIcNAc-PI deacetylation                 Not cloned                              85,86
 ?                           G9PLAP.85                             GIcNAc-PI deactylation                  Not cloned
      V. L. Stevens, H. Zhang and M. Harreman, unpublished work.

   Abbreviations used: GPI, glycosylphosphatidylinositol; VSG, variant surface glycoprotein; ER, endoplasmic reticulum; PI, phosphatidylinositol; Pl-
 PLC, PI-specific phospholipase C; LPG, lipophosphoglycan; GIPL, glycoinositolphospholipid; Dol-P-Man, dolichol-phosphomannose; PARP, procyclic
 acid repeat protein; PNH, paroxysmal nocturnal haemoglobinuria.
   Address correspondence to: Department of Radiation Oncology, Loughlin Radiation Oncology Center, 145 Edgewood Ave., S.E., Atlanta, GA 30335,
362               V. L. Stevens

GPI STRUCTURE                                                                                        galactosylate a hydrophobic thioglycoside [Manal-6Manal-S-
Despite the diversity in the eukaryotic organisms in which they                                      (CH2)7-CH3] that is homologous to a portion of the mature GPI
are found and the function of the proteins which they modify, all
                                                                                                     anchor [35]. Glycosylated GPI precursors (galactosylated in
GPI membrane anchors have a common core structure which is                                           trypanosomes [36] and GlcNAc-modified in Toxoplasma gondii
illustrated in Figure 1. A myo-inositol-containing lipid, usually                                    [37]) that arise from modification of the GPI core rather than
PI, anchors the GPI in the cell membrane. Unacetylated glucos-                                       from degradation of protein-attached GPI anchors have also
amine (GlcN), which occurs infrequently in eukaryotic bio-                                           been found. Therefore the possibility that at least some of the
molecules, is attached in an al-6 linkage to the inositol. Three                                     accessory carbohydrates are added to GPI before transfer to
mannoses (Man) are then attached through different linkages to
                                                                                                     protein cannot be ruled out. The addition of extra ethanolamine.
form a linear trisaccharide attached to the GlcN residue in an                                       phosphates occurs during the biosynthesis of the GPI core in
al-4 linkage. Finally, the phosphate of phosphoethanolamine                                          mammalian cells [38-42]. Whether the resulting GPI precursors
is attached to the 6-position of the third mannose. This glycan,                                     containing two or three ethanolamine phosphates can be trans-
                                                                                                     ferred to protein is not known. However, it seems likely that
consisting of ethanolamine-PO4-6Mana l-2Manal-6Manacl-                                               these modifications are retained in cases where the protein-
4GlcNa.l-6myo-inositol, is attached to the C-terminus of a                                           attached GPI anchors have accessory ethanolamine phosphates.
protein by a peptide bond involving the amine of the                                                    Other variations in the GPI structure are found in the inositol
phosphoethanolamine.                                                                                 lipid portion of the molecule. These are in the nature of the lipid
  The GPI core is decorated with accessory carbohydrates and
                                                                                                     backbone (glycerol versus ceramide), the linkages used for
phosphoethanolamines that extend from the mannoses and vary                                          attachment of the hydrocarbon chains (ester versus ether), the
with the protein to which it is attached and the tissue or organism
in which it is expressed. The extra carbohydrates identified in
                                                                                                     number and length of the hydrocarbon chains and whether the
                                                                                                     inositol ring is modified by an acyl chain. Some of these
protein-linked GPI anchors include mannose, glucose, galactose                                       differences appear to be species-specific, while others change with
(Gal), N-acetylgalactose, N-acetylneuraminic acid and N-acetyl-                                      the developmental stage of the organism. Most species utilize a
glucosamine (GlcNAc). Comparison of the kinetics of addition                                         glycerolipid backbone exclusively in the GPI core. However, a
of the GPI anchor (t2 of < 1 min) [15,16] and at least some of the
                                                                                                     ceramide backbone has been found in the GPI anchors of yeast
galactoses (t1 of 15 min) [32] with trypanosome VSG suggests                                         [43] and Dictyostelium [44]. The fatty acids in Trypanosoma
that the side-chain carbohydrates are added after transfer of the
GPI anchor to the protein [33]. Consistent with this conclusion,
                                                                                                     brucei VSG and Torpedo acetylcholinesterase [45] are attached to
                                                                                                     the glycerol backbone by ester linkages, which are found in most
endogenous VSG has been shown to be a substrate for a-                                               phospholipids. In contrast, an ether linkage is found in the sn-l
galactosylation by UDP-['4C]Gal in a cell-free trypanosomal                                          position of the glycerol backbone of GPI anchors on many other
system [34]. However, these membranes have also been shown to                                        parasitic and mammalian proteins [46-48]. In many of the
                                                                                                     parasitic proteins, the GPI anchor is a lyso-lipid with only the
                                                                                                     alkyl chain in the sn-l position and no acyl chain in the sn-2
                                                                                                     position [49,50]. In most species, there appear to be no rigid
                                                                                                     requirements for specific fatty acids in GPI. However, both acyl
                                                                                                     chains in the mature GPI anchor of Trypanosoma brucei VSG are
                                                                                                     exclusively myristate [45]. The occurrence of the acyl chain on
                                                                                                     the inositol ring {esterified to the 2-hydroxy group ([51], A.
                                                                                                     Treumann, M. R. Lifely, P. Schneider and M. A. J. Ferguson,
                                                                                                     unpublished work)} is variable. This modification, which renders
                                                                                                     the GPI anchor resistant to cleavage by PI-PLC [51-53], is found
                                                                                                     on some mammalian GPI anchors [19,47,52,54,55] and on
                                                                                                     procyclic acidic repetitive protein (PARP) of the procyclic
                                                                                                     Trypanosome brucei [56].
                                                                                                        Glycolipids that are structurally related to GPI but are not
                                                                                                     used as protein membrane anchors have been found in
                                                                                                     Leishmania parasites. These are lipophosphoglycan (LPG) and
                                                                                                     glycoinositolphospholipids (GIPLs). Although the early steps in
                                                                                                     the biosynthetic pathways appear to be shared by GPI anchors,
                                                                                                     LPG and GIPLs, there is divergence in the later reactions. Since
                                                                                                     this review will deal only with the biosynthesis of GPI membrane
                                                                                                     anchors, the reader is referred to studies by McConville et al.
                                                                                                     (summarized in [57]) for further information about LPG and
                                                                                                     GIPL biosynthesis.

                                                                                                     BIOSYNTHESIS OF N-ACETYLGLUCOSAMINYL-PI
                                                                                                     In the first reaction of GPI biosynthesis, which is shown in
                                                                                                     Figure 2, GlcNAc is transferred from UDP-GlcNAc to PI to
                                                                                                     form N-acetylglucosaminyl-PI (GlcNAc-PI). This reaction has
Figure 1 The GPI core                                                                                been studied with cell-free systems using trypanosome
The conserved GPI core found in all eukaryotic organisms is shown. No structural detail is
                                                                                                     membranes [21-23], mammalian cell lysates [58,59] and
provided for the lipid portion of the glycolipid because of the variability noted in the text (see   microsomal fractions [60,61], yeast membranes [62] and
'GPI Structure' section). Abbreviations: Etn, ethanolamine; Ins, inositol.                           Toxoplasma gondii membranes [37]. In these assays, radiolabelled
                                                                                                Biosynthesis of glycosylphosphatidylinositol membrane anchors      363

                            Pi                                                                     thermore, microsomes prepared from EL4 (Thy- 1-f) cells, which
      IUDP-GIcNAC -                                                                                are unable to make alkyl lipids because of a defect in ether lipid
                                                                                                   biosynthesis [67], produce normal levels of GlcNAc-PI and other
                       GIcNAc-PI                                                                   GPI precursors (V. L. Stevens, unpublished work). After cor-
  G DP-Man                                                                         G DP-Man
                                      OAc                                                          rection of the defect in the later step of GPI biosynthesis in these
                                  S                                                                cells by transfection with the PIG-F gene [66], GPI anchoring of
 jDOI-P-Man|            GIcN-PI                       G IcN Pikac-yi,              DoI-P-Man]      Thy-1 was recovered but the ether lipid deficiency was not
                                                                                                   corrected (V. L. Stevens and T. Kinoshita, unpublished work).
                      ManGldNPII                   ManGIlcN-Pk acyl                                Therefore mammalian cells can make GPI anchors having only
                                                                                                   acyl-linked fatty acids without any deleterious consequences.
              11    Man GIcN: Pll1                                                                 The final type of specificity seen in the PI portion of GPI is in the
                                                                                                   fatty acid composition of the VSG membrane anchor. Detailed
                    MansGlcN-PI                   Man-GlcN-P!iacyl'                                analysis (see 'Fatty acid remodelling' section below) has es-
                            I.:                                 -   -
                                                                                                   tablished that trypanosomes synthesize GlcNAc-PI and other
                                              E'n P Man^1GleN           P4(NCaev                   GPI precursors using PI with a variety of fatty acids [68] and
                                                                                                   introduce myristate into the anchor after complete assembly of
                                                                                                   the core carbohydrates.
Figure 2 Pathways for GPI biosynthesis in trypanosomes, yeast and
mammalian cells                                                                                     Enzymology
The reactions common to all three species are shown in black. Those specific to trypanosomes        Analysis of the GlcNAc-PI biosynthetic capacity of the Thy-l-
are on the left in grey-shaded boxes, and those specific for yeast and mammalian cells are on       deficient lymphoma mutants revealed that three of the
the right in red. The endogenous donors of the GPI core components are shown in pink-shaded
boxes. Abbreviations: PE, phosphatidylethanolamine; OAc, acetate; Etn, ethanolamine.                complementation groups (classes A, C and H) were defective in
                                                                                                    this reaction [58,60]. Therefore three separate gene products
                                                                                                    are required for this reaction in mammalian cells. Three
GlcNAc (3H or 'IC) is transferred from a water-soluble donor                                        complementing yeast mutants in this reaction have also been
(UDP-GlcNAc) to a chloroform-soluble product (GlcNAc-PI).                                           isolated ([69] and P. Orlean, unpublished work), indicating that
The second substrate in the reaction, PI, is supplied by the                                        three gene products are also involved in GlcNAc-PI biosynthesis
membranes or cellular fractions which also contain the enzymic                                      in this species. While there is no reason to expect any difference
activity which catalyses this reaction. Subcellular localization of                                 in the requirements for this reaction in trypanosomes, the number
the GlcNAc-PI biosynthetic activity in mammalian cells and                                          of gene products involved has not been established in this
trypanosomes has established that it resides in the ER [63,64].                                     parasite.
In mammalian cells there is significant metabolism of UDP-                                             cDNAs codings for class A and H genes have been expression
GlcNAc to non-GPI-related products [61] that decreases the                                          cloned using the mutants in these components. The class A
amount of this substrate available for GlcNAc-PI biosynthesis.                                      cDNA, which has been designated PIG-A, was cloned from a
This competition does not occur with trypanosome membranes                                          HeLa library using the class A mutant human B lymphoblastoid
[23]. These degradatory activities can be effectively inhibited by                                  cell line JY-5 [70]. The protein encoded by this cDNA is 484
including ATP and a chelator (EDTA or dimercaptopropanol,                                           amino acids in length with a predicted molecular mass of
depending on the cell type) in the reaction mixture [61].                                           62.6 kDa and no homology to any known protein. The PIG-A
                                                                                                    protein contains no apparent signal peptide sequence and a
                                                                                                    hydrophobic sequence of 27 residues that is 42 amino acids away
Substrate specificity                                                                               from the C-terminus of the protein. This hydrophobic sequence
The high levels of inositol-containing phospholipids (particularly                                  may serve as a transmembrane domain which anchors the PIG-
PI) relative to the amount of GlcNAc-PI made by the membranes                                       A gene product in the ER membrane. The amino acid sequence
have precluded in vitro radiolabelling of this GPI precursor with                                   deduced from the murine PIG-A cDNA is 88 % identical to that
[3H]inositol. However, it has been shown that the addition of                                       of the human gene product, indicating that this protein is highly
exogenous PI can increase the amount of GlcNAc-PI made by                                           conserved [71]. The genomic PIG-A gene has been cloned and
mammalian cell lysates [58]. Despite the apparent specificity for                                   found to be at least 17 kb in length and to have six exons [72,73].
variations in structure of the inositol lipid portion of the GPI                                    Chromosomal localization studies have established that the PIG-
anchor in many species, the available evidence suggests that this                                   A gene is on the X chromosome [72,74]. To date, no biochemical
is not determined by the selection of particular inositol                                           studies have been carried out to investigate the role of PIG-A in
phospholipids for incorporation into GlcNAc-PI. In yeast,                                           GlcNAc-PI synthesis.
GlcNAc-PI and the other early intermediates in GPI biosynthesis                                        A cDNA which corrects the class H defect in GPI biosynthesis,
are alkaline-labile [62,65], indicating that they contain a glycerol-                               designated GPI-H, has been cloned from a human placental
based inositol phospholipid. Therefore the ceramide backbone                                        library using the Ltk- cell line [75], which belongs to this
present in many of the yeast protein-attached GPI anchors is not                                    complementation group [76]. The protein encoded by this cDNA
introduced during the early steps in the biosynthesis of this                                       is 188 amino acids in length and is predicted to have a molecular
glycolipid [66]. In mammalian cells, alkylacyl-PI represents less                                   mass of 21 kDa. Like the PIG-A protein, GPI-H is not related by
than 5 0% of the total cellular PI [67]. Therefore tremendous                                       sequence to any other protein. GPI-H, which contains hydro-
selectivity would be required to introduce the alkyl linkage found                                  philic sequences at the N- and C-termini (30 amino acids each)
in many mammalian GPI anchors during GlcNAc-PI bio-                                                 and a relatively hydrophobic core, is proposed to be a cytosolic
synthesis. That there is no such selectivity is suggested by the                                    protein because there is no signal peptide sequence in the cDNA.
finding that only 15-20 % of the GlcNAc-PI synthesized by                                            However, the activity provided by the class H gene product
lysates of a normal lymphoma cell line contains the sn- alkyl                                       clearly stays associated with the ER membrane upon subcellular
chain as judged by sensitivity to alkaline hydrolysis [58]. Fur-                                     fractionation [61,63]. If the class H protein does not associate
364          V. L. Stevens

with the membrane itself (which has not been ruled out), it may        function, the one active PIG-A gene can be affected with a single
interact with the class A and/or the class C protein in such a way     mutation. The consequence of this is to increase the probability
that it behaves as a membrane-bound protein. The genomic               of a mutation which results in a defect in GPI biosynthesis by
location of GPI-H is on chromosome 12 in the mouse [74].               several orders of magnitude.
Because many of the genes on mouse chromosome 12 are located
on human chromosome 14, Ware et al. [74] suggest that human            DEACETYLATION OF GIcNAc-PI TO GlcN-PI
GPI-H is located on chromosome 14. As with the class A gene
product, no information is available regarding the function of         In the second reaction in GPI biosynthesis, the N-acetyl group of
the GPI-H gene product in GPI biosynthesis. Therefore, even            GlcNAc-PI is removed to form GlcN-PI. Because no additional
though studies of the N-ethylmaleimide inhibition of GlcNAc-PI         substrates are required, deacetylation of varying amounts of
biosynthesis suggest that there is a cysteine residue at or close to   GlcNAc-PI occurs spontaneously in the parasite, yeast and
the UDP-GlcNAc binding site [77], it is not known which of the         mammalian cell-free systems. As with the other steps in the
three gene products contains the affected amino acid or the            pathway, the deacetylation of GlcNAc-PI was first demonstrated
substrate binding site.                                                in trypanosome membranes [22]. That GlcN-PI was derived from
   Despite the cloning of two of the three genes required for          GlcNAc-PI and not from the direct transfer of GlcN from UDP-
GlcNAc-PI biosynthesis in mammalian cells, no insight as to            GlcN to PI was established by showing that the N-acetylated
why three separate proteins are needed for this reaction has been      intermediate arose first and that the addition of high levels of
gained. The finding that GlcNAc-PI biosynthesis occurs on the          UDP-GlcN did not affect the level of synthesis of mannose-
cytosolic face of the ER [63] eliminates the possibility that one of   containing GPI precursors. Furthermore, direct deacetylation of
the gene products is required for the transport of UDP-GlcNAc          exogenously supplied GlcNAc-PI by the trypanosome mem-
into the lumen of this organelle. Mixing experiments in which          branes was also demonstrated.
GlcNAc-PI biosynthesis was observed when microsomes from                  Mammalian cell lines defective in the second step of GPI
each of the mutants was combined with microsomes from normal           biosynthesis have not been available until recently. Two mutants
cells ruled out the possibility that any of the mutants possessed      reported to be affected in GlcNAc-PI deacetylation have been
an inhibitory activity [58]. One possibility that has not been         isolated by fluorescence-activated cell sorting of K562 cells
addressed is that a heterotrimeric enzyme catalyses the bio-           deficient in the surface expression of decay accelerating factor
synthesis of GlcNAc-PI. Biochemical complementation using              [85,86]. The first of these mutants, designated IB, deacetylates
microsomes for three mutants has been unsuccessful, even in the        GlcNAc-PI with 10 % of the efficiency of wild-type K562 cells
presence of detergent [using levels which do not inhibit this          [85]. The other mutant, IVD, appears to be completely defective
reaction in normal microsomes (H. Zhang and V. L. Stevens,             in this activity [86]. Because these two cell lines show different
unpublished work)]. This finding is consistent with some as-           levels of mutation and were generated using different chemical
sociation between the three gene products being required that          mutagens, it is proposed that they have different genetic lesions
cannot be achieved in vitro. Perhaps this association occurs co-       [86]. However, whether separate genes are affected in these
translationally and is dependent on the proteins having certain        mutants is not known, because the complementation analysis has
conformations that are not present when they are expressed             not been reported. A third cell line which is completely blocked
separately. Another possibility is that one or both of the             in the GlcNAc-PI deacetylase reaction has been isolated from
remaining gene products is rapidly degraded when the third is          CHO cells expressing alkaline phosphatase by complement-
not synthesized. Both of these explanations are compatible with        mediated lysis (V. L. Stevens, H. Zhang and M. Harreman,
the three gene products forming a heterotrimeric enzyme and can        unpublished work). The relationship of this mutant to the two
be tested once antibodies against the two cloned proteins are          K562 cell lines has not been investigated. These mutants are
made.                                                                  expected to provide the system required for expression cloning of
                                                                       the GlcNAc-PI deacetylase.
Paroxysmal nocturnal haemoglobinuria (PNH)                             Enzymology
PNH is an acquired disease in which patients suffer from               The GlcNAc-PI deacetylase is the only enzyme in GPI bio-
haemolytic anaemia because several proteins that protect their         synthesis which has been partially purified and characterized
blood cells from lysis by the complement system (e.g. decay            biochemically. Using trypanosome membranes, Milne et al. [87]
accelerating factor and CD59) are not properly expressed [78,79].      achieved a 140-fold purification of this enzyme. However, the
The underlying defect in PNH is an inability to synthesize the         GlcNAc-PI deacetylase became very unstable upon isolation and
GPI anchor which normally attaches these proteins to the cell          only 200o of the expected activity was recovered after four
surface [79,80]. Biochemical studies of T-cells [81] and lympho-       purification steps. Therefore substrate specificity and inhibitor
blastoid cell lines [82,83] established that the PNH defect in GPI     studies were done using stable deacetylase preparations obtained
biosynthesis was in GlcNAc-PI formation. Complementation               after one purification step (11-fold purified). Using chemically
analysis of these cells with the Thy-l-deficient lymphoma cell         prepared GlcNAc-PI, the deacetylase was found to have a K,,, of
mutants revealed that the defect resides in the class A gene           approx. 1.5 ItM for this substrate. Substrate analogues lacking
[81,82]. To date, the PIG-A gene has been examined and found           the fatty acids were deacetylated poorly and were ineffective
to be mutated in 31 Japanese patients ([72,84] and N. Yamada,          inhibitors of the enzyme, indicating that the lipid portion of
T. Miyata, K. Maeda, T. Kitani, J. Takeda and T. Kinoshita,            GlcNAc-PI is important in either substrate recognition or
unpublished work) and 14 Thai patients (T. Kinoshita, unpub-           presentation. GlcNAc-PI analogues without the glycero-
lished work). Therefore it appears that mutations in PIG-A are         phosphate portion of the molecule were neither substrates nor
responsible for PNH in all patients. The reason that PNH has           inhibitors of the deacetylase. Therefore this part of the glycolipid
not been found to be caused by defects in the other genes              appears to be essential for enzyme recognition. The enzyme
required for the biosynthesis of GPI appears to be related to the      appears to have no specificity for the configurations of the C-2 to
localization of PIG-A on the X chromosome. Unlike genes on             C-5 hydroxy groups on the inositol ring, because the D- and L-
autosomes, where two copies must be mutated to affect the              myo-inositol-containing substrates are used equally well. GlcN-
                                                                        Biosynthesis of glycosylphosphatidylinositol membrane anchors   365

PI was found to be only a weak inhibitor of GlcNAc-PI                      acids that is not homologous to any known protein. No
deacetylation, suggesting that end-product inhibition of this              hydrophobic N-terminal signal sequence was found in the
reaction is not likely to play a role in the regulation GPI                pwwrotein. While it is assumed that this protein is a mannosyl-
biosynthesis in trypanosomes.                                              transferase, this has not been confirmed biochemically.
   The nature of the active site of the GlcNAc-PI deacetylase                 A number of studies have investigated the ability of
remains elusive. Other studies have shown that this reaction is            mannosamine to affect GPI biosynthesis. This amino sugar
not inhibited by serine-peptidase and serine-esterase inhibitors           inhibits this pathway in mammalian cells [98,99], trypanosomes
[88] or by thiol alkylating agents [77]. Preliminary studies indicate      [98,100] and Leishmania parasites [101]. Some conflicting reports
that Mg2+ is required for GlcNAc-PI deacetylation in mammalian             exist about the mechanism of inhibition [102], but most evidence
cells (V. L. Stevens and H. Zhang, unpublished work), but                  indicates that mannosamine is incorporated into the GPI core
whether this is true in other systems is not clear.                        predominantly in the position of the second mannose [99,100].
                                                                           Because the position 2 substituent in this mannose analogue is an
                                                                           amine that cannot be used for the addition of the third mannose,
MANNOSE ADDITIONS                                                          the mannosamine-containing glycolipid acts as a competitive
After deacetylation of GIcNAc-PI to GlcN-PI, the GPI core is               inhibitor of this reaction and blocks GPI biosynthesis by a chain
extended by the addition of three mannoses. In trypanosomes,               termination type of mechanism [100].
GlcN-PI is the substrate for these additions [21,23]. In yeast [62]
and mammalian cells [89,90], an acyl group must be added to the
inositol ring of GlcN-PI before the mannoses can be added (see              PHOSPHOETHANOLAMINE ADDITION
'Inositol acylation' sections below). Therefore GlcN-PI(acyl) is            The biosynthesis of the GPI core is completed by the addition of
the substrate for mannose additions in these species. The first             phophoethanolamine to the 6-hydroxy group of the third man-
mannose is added in an Lxl-4 linkage to the GlcN. The second                nose residue. The endogenous donor of the phosphoethanol-
mannose is then linked cl-6 to the first, and the third is linked           amine in yeast and trypanosomes is the phospholipid
xl-2 to the second. In trypanosomes, the pathway for GPI                    phosphatidylethanolamine. This was established in yeast by
biosynthesis is as expected, with ManGlcN-PI being the substrate            demonstrating that mutants that could not transfer exogenous
for the second mannose addition and Man2GlcN-PI being the                   [3H]ethanolamine into phosphatidylethanolamine because of
substrate for the third mannose addition. In mammalian cells,               disruption of the ethanolaminephosphotransferase gene also did
the pathway appears to be more complicated. Although the                    not incorporate this radiolabelled compound into GPI-anchored
sequence of reactions has not yet been established, it appears that         proteins [103]. In trypanosomes, experiments in which GPIs were
an ethanolamine-phosphate residue is added to the first mannose             radiolabelled with [3H]serine via phosphatidylserine decarboxyl-
before addition of the second and third mannoses (see 'Phos-                ation established the role of phosphatidylethanolamine as the
phoethanolamine addition' section below). Therefore the actual              donor [104]. This study also demonstrated that the phospho-
substrates for the second and third mannose additions in                    ethanolamine is acquired en bloc in a single reaction. Similar
mammalian cells have not been established conclusively.                     [3H]serine radiolabelling experiments in murine lymphoma cells
   The mannoses of the GPI core can be radiolabelled by adding              indicate that phosphatidylethanolamine is the phosphoethanol-
GDP-[3H]mannose to trypanosome membranes [21,23]. How-                      amine donor in mammalian cells too (M. Dhalla and A. K.
ever, the endogenous donor for the three mannoses is dolichol-              Menon, unpublished work).
phosphomannose (Dol-P-Man). The first evidence that Dol-P-
Man was the source of at least one of the core mannoses came
from studies of mutants defective in GPI biosynthesis. Chapman              Enzymology
et al. [91] found that the class E Thy-l-deficient lymphoma                 A GPI intermediate with three mannoses accumulates in the
mutant was unable to synthesize Dol-P-Man. Other mammalian                  class F Thy-1-deficient lymphoma mutant, indicating that this
cell [92,93] and yeast [94] mutants with defects in Dol-P-Man               cell type is unable to add the terminal phosphoethanolamine to
production or utilization are also deficient in GPI biosynthesis.           the GPI core [60]. A cDNA that corrects the defect in GPI
Correction of the defect in the class E mutant by transfection              biosynthesis was expression cloned and found to code for a
with a yeast Dol-P-Man synthase gene resulted in recovery of                protein of 219 amino acids [66]. Like the other cDNAs encoding
GPI biosynthesis [95]. That all three mannoses were donated to              proteins required for GPI biosynthesis, PIG-F is not homologous
the GPI core by Dol-P-Man was established by experiments with               to any known proteins. The sequence is very hydrophobic, as
trypanosomes in which Menon et al. [96] demonstrated that each              might be expected for an enzyme which transfers phospho-
of the mannoses became radiolabelled when trypanosome                       ethanolamine from a membrane-bound donor to a membrane-
membranes were incubated with Dol-P-[3H]Man.                                bound acceptor. Although the gene for PIG-F has not yet been
                                                                            cloned, it has been localized to chromosome 17 in the mouse [74].
                                                                               The addition of the terminal phosphoethanolamine to
Enzymology                                                                  Man3GlcN-PI is inhibited both in vivo and in vitro in
Because different linkages and substrates are used for each of the          trypanosomes by PMSF [88]. This finding suggests that the
three mannose additions, it seems likely that three different               phosphoethanolamine transferase has an active-site serine residue
enzymes catalyse these reactions. However, the only one of these            which may be used to form a serine-phosphoethanolamine
enzymes about which any information is available is the third               intermediate during the transfer reaction.
mannosyltransferase. Analysis of the GPI precursors which
accumulate upon metabolic radiolabelling with either
[3H]mannose or [3H]inositol has shown that the class B Thy-l-               UNIQUE ASPECTS OF GPI BIOSYNTHESIS IN TRYPANOSOMES
deficient lymphoma cell mutant is defective in this reaction                Two modifications of GPI anchors occur in the bloodstream
[38,60]. A cDNA which corrects the GPI biosynthetic defect in               form of Trypanosoma brucei that either do not occur or happen
these cells has been expression cloned with the SIA(Thy-l-b)                differently in the procyclic form of this parasite and in other
mutant [97]. The PIG-B cDNA encodes a protein of 554 amino                  organisms. These are replacement of the fatty acids in the GPI
366               V. L. Stevens

Figure 3 Fatty acid remodelling in bloodstream trypanosomes
The step-wise exchange of the fatty acids of the newly synthesized GPI core for myristate (shown in red), which occurs in trypanosomes, is illustrated.

with myristic acid and the reversible acylation of the inositol                                   'proofreading' mechanism to ensure that the Cu4o fatty acid is
ring.                                                                                             in the GPI anchor.

Fatty acid remodelling                                                                            Reversible inositol acylation
The GPI anchor of VSG has exclusively myristic acid attached in                                   The GPI anchor of Trypanosoma brucei VSG is completely
ester linkages in the sn-I and sn-2 positions of the glycerol                                     sensitive to cleavage by PI-PLC [7], indicating that it is not
backbone [45]. However, GPI is initially synthesized by                                           inositol-acylated. However, acylated GPI precursors have been
trypanosomes with stearic acid in the sn- I position and a mixture                                detected and characterized in this parasite [19,21,23,111]. The
of fatty acids including 18:0, 18:1, 20:4 and 22:6 in the sn-2                                    major acylated precursor (referred to as either P3 or glycolipid
position [68]. Therefore the fatty acids must be removed and                                      C) contains all components of the mature GPI core and has been
replaced with myristic acid. This process, which has been termed                                  shown to be competent for transfer to VSG in a trypanosomal
fatty acid remodelling, was dissected by Masterson et al. [105]                                   cell-free system [112]. When first characterized, the role of this
using a trypanosomal membrane cell-free system. As shown in                                       acylated precursor was not clear. Initially it was suggested to be
Figure 3, remodelling occurs by the sequential replacement of                                     an intermediate in the fatty acid remodelling of the trypanosome
first the sn-2 fatty acid and then the sn-I fatty acid with myristic                              GPI [105]. However, recent evidence indicates that glycolipid C
acid. The fatty acid donor for both of these reacylation reactions                                is not an obligate intermediate in the biosynthesis of GPI in this
is myristoyl-CoA. During remodelling, the GPI precursor retains                                   parasite [106]. Guther et al. [106] have shown that the mature
at least one acyl chain at all times and can remain membrane-                                     unacylated GPI core continued to be made (both in vivo and in
associated. The glycolipid which is normally the substrate for                                    vitro) when synthesis of the acylated precursor was blocked by
fatty acid remodelling is the phosphoethanolamine-containing                                      PMSF. Despite inhibition of its synthesis, glycolipid C was
GPI core [105]. However, when the addition of the phospho-                                        deacylated to form the unacylated core. Therefore these authors
ethanolamine is blocked with PMSF, the Man3GlcN-PI which                                          speculate that glycolipid C is used as a reservoir for excess
accumulates is found to undergo this process in vivo [106].                                       mature GPI core that could be mobilized for addition to protein
   The reason why trypanosomes require myristic acid in their                                     by deacylation if needed.
VSG GPI anchors is not known. However, because the hosts of                                          The enzyme which catalyses inositol acylation is sensitive to
these parasites do not have this requirement and bloodstream                                      inhibition by PMSF [106] and acylates predominantly the
trypanosomes utilize exogenous myristic acid for fatty acid                                       phosphoethanolamine-modified GPI precursor (acylated forms
remodelling [107], this process is a potential target for anti-                                   of earlier precursors have also been detected [21,23]). The
trypanosomal drugs. In fact, several myristic acid analogues                                      acylating enzyme is not responsible for the deacylation of
have been shown to be toxic to trypanosomes [108,109].                                            glycolipid C, as this process was not inhibited by PMSF [106].
   A process similar to fatty acid remodelling, called myristate                                  Radiolabelling studies have shown that the inositol acyl group is
exchange, has been described in trypanosomes [110]. During                                        predominantly palmitic acid [19]. Field et al. [113] showed by
myristate exchange the fatty acids of VSG-attached GPI (which                                     pulse-chase analysis that acylation of the GPI anchor of PARP
are already myristic acid) are replaced with myristate donated                                    in procyclic trypanosomes increased during the chase despite
from myristoyl-CoA. This exchange is thought to occur in a                                        removal of [3H]palmitate, indicating that the acyl group was
different cellular compartment than fatty acid remodelling and to                                 derived from a large metabolic pool. Therefore, rather than
have some different characteristics. While the function of this                                   being transferred directly from an acyl-CoA, the inositol acyl
exchange of fatty acids is unclear, it has been proposed to be a                                   group appears to be derived from a phospholipid precursor.
                                                                      Biosynthesis of glycosylphosphatidylinositol membrane anchors        367

UNIQUE ASPECTS OF GPI BIOSYNTHESIS IN YEAST                              Inositol acylation
The biosynthesis of yeast GPI differs from that of trypanosome           The occurrence of an inositol acyl group on the GPI anchor of
and mammalian membrane anchors in at least two ways. First,              mammalian proteins is variable with cell type [55]. However, the
like some precursors found in the other species, many of the GPI         majority of the mannose-containing GP1 precursors detected
intermediates in yeast are acylated on the inositol ring [65].           either by metabolically radiolabelling cells or in cell-free systems
However, the mechanism involved in this modification appears             have this extra acyl group [38-42]. As in yeast, the finding that
to be different in this organism. Second, many of the GPI                GlcN-PI(acyl) accumulated in a mutant lymphoma cell line
anchors on yeast proteins are ceramide-based rather than                 (class E) unable to make mannose additions to the GPI core led
glycerolipid-based. This variation in structure has also been            to the proposal that the acylation of GlcN-PI is the obligatory
found in Dictyostelium [44].                                             third step in GPI biosynthesis in these cells [89]. That GlcN-
                                                                         PI(acyl) is a substrate for mannose additions has been
                                                                         demonstrated      by pulse-chase analysis         of   lymphoma   cell
Inositol acylation                                                       microsomes in which the radiolabel in this glycolipid was seen to
There is no evidence that the GPI anchors of yeast proteins are          decrease and appear in mannose-containing GPI precursors
resistant to PI-PLC cleavage. However, inositol-acylated GPI             upon addition of GDP-mannose to the incubation [90]. The
intermediates have been detected in yeast [65] and the inositol          inositol acylation reaction has been studied in vitro and found to
acylation reaction has been studied using a cell-free system [62].       be stimulated by the addition of either CoA or an acyl-CoA [90].
Unlike trypanosomes, in yeast the acyl group is added to GlcN-           When synthesized in a coupled reaction in which GlcN-PI must
PI, the second intermediate in the pathway (see Figure 2). This          also be made in the same incubation, the acylation of this
reaction was found to be stimulated when both ATP and CoA                glycolipid is strictly dependent on GTP. This nucleotide
were added to the incubation, suggesting that an acyl-CoA is the         triphosphate probably does not affect this reaction directly but is
donor of the inositol acyl group. Further support for this was           necessary for making the substrate available for acylation (see
provided by the demonstration that [14C]palmitate could be               'Regulation of GPI biosynthesis' section below for more details
transferred directly from [14C]palmitoyl-CoA to GlcN-PI.                 of GTP effects). The stimulation of GlcN-PI acylation by CoA
Although mannose addition to GlcN-PI(acyl) by yeast has not              occurred in the absence of ATP, an essential cofactor for acyl-
been shown directly, the finding that this acylated GPI inter-           CoA synthesis, indicating that free CoA is the endogenous
mediate accumulates in yeast defective in Dol-P-Man synthesis            effector of this reaction and that the acyl group is most likely
[94] supports the proposal that it is an obligatory intermediate in      derived from a phospholipid. Therefore the third step in mam-
the pathway [62].                                                        malian GPI biosynthesis has been proposed to be catalysed by a
                                                                         CoA-dependent transacylase [90].
                                                                            Inositol acylation involves the same substrate (GlcN-PI) but
Glycerolipid to sphingolipid exchange                                    different acyl donors in yeast and mammalian cells. In
Most, but not all, of the protein-attached GPI anchors in yeast          trypanosomes and mammalian cells, the substrates for acylation
are alkaline-resistant because of the presence of a ceramide             differ while the acyl donors may be the same. However, inositol
backbone [114]. As described above (see 'Biosynthesis of                 acylation in the parasite is inhibited by PMSF while in mam-
N-acetylglucosaminyl-PI' section), GlcNAc-PI and other GPI               malian cells it is not [90], indicating a further difference in the
precursors synthesized by this organism are completely alkaline-         systems responsible for this reaction.
sensitive [62,65]. Therefore an exchange reaction in which the              The role of the inositol acyl group during GPI biosynthesis in
diacylglycerol backbone is replaced by a ceramide backbone has           all species remains unclear. Despite the fact that many mam-
been proposed to generate the profile of mature GPI anchors              malian GPI anchors are inositol-acylated, analysis of the Pl-
seen attached to protein. The switch in alkaline-sensitivity does        PLC-sensitivity of newly synthesized alkaline phosphatase has
not occur in secretion mutants unable to transport proteins out          shown that the anchor behaves as if it is unacylated at the earliest
of the ER [114]. Therefore the exchange reaction must happen at          time after GPI addition [116]. PI-PLC resistance only appeared
some place other than the ER, the site of GPI biosynthesis.              at later times, suggesting that the GPI transferred to protein was
Because only alkaline-sensitive GPI precursors corresponding to          unacylated. Therefore the mature GPI core made by mammalian
the complete core have been found, the exchange reaction has             cells must be deacylated prior to addition to protein.
been proposed to occur after the transfer of the glycolipid anchor       Alternatively, there could be either independent pathways or free
to protein [65], making this process different than the fatty acid       interconversion of acylated and non-acylated GPI precursors
remodelling that occurs in trypanosomes. As in the parasite, the         which has not been seen because most of the characterization of
reason for the exchange of the glycerolipid backbone to a                these intermediates has been done in mutant cells which ac-
ceramide base in yeast is unknown. However, studies using                cumulate these glycolipids [38,39,41], perhaps in non-physio-
inhibitors of sphingolipid biosynthesis in which yeast are found         logical ways. Until precursor-product relationships are estab-
to make only alkaline-sensitive GPI anchors suggests that the            lished between the acylated and non-acylated intermediates,
need for ceramide in this glycolipid can be overcome [115].              which of these mechanisms occurs will not be known.

UNIQUE ASPECTS OF GPI BIOSYNTHESIS IN MAMMALIAN CELLS                     Addition of extra phosphoethanolamines
Like in yeast, inositol acylation occurs early in GPI biosynthesis        The structural analysis of GPI precursors in normal and Thy-l-
in mammalian cells. However, the current evidence suggests that           deficient cell lines has revealed that many of the mannose-
different types of enzymes catalyse this reaction in the two              containing intermediates have extra phosphoethanolamines at-
species. Another modification of the GPI core which occurs                tached to the core [38-42]. The glycolipids characterized include
during its assembly in mammalian cells which has not been                 Etn-P-ManGlcN-PI(acyl), Man(Etn-P)ManGlcN-PI, Man-
described in other species is the addition of extra phospho-              Man(Etn-P)ManGlcN-PI, Etn-P-ManMan(Etn-P)ManGlcN-Pl
ethanolamine to the first and second mannose residues.                    and Etn-P-Man(Etn-P)Man(Etn-P)ManGlcN-Pl. The various
368              V. L. Stevens

                                       G Ic N PI acy

                               ManGIcN-Pl(acylb                               >             ManGIcN-PI(acylM

                         Man-ManGIcN-PI(acyl                                          Marn-Man G lcN-PI (acyl)

                  Man-Ma n-ManG IcN-PI (acyl)              ------------           Manl Man ManGIcN Plhacyl)
                                                                                            P-Etn                                                                ~~~~~~~~~~~~~~~~~~~~~~4..........

          _Etn-P-Man-Man-ManGI              P                             Eti P Man - Mari- M IanGIcN-PI(acyl)
                                                                                                                                        Etn-P-Man-Marn-M anGIcN-PI(acyl)
                                                                                                                                                  P-Etn P-Etr
Figure 4 Possible pathways by which mammalian cells may synthesize GPI
The most probable pathway for synthesis of the GPI core containing extra phosphoethanolamines by mammalian cells is shown in red. The intermediates in black have been found in cells, while
those in grey-shaded boxes have not been detected.

ways that these intermediates could arise are shown in Figure 4.                                 by P1-PLC sensitivity [64]. Concanavalin A binding of the
Although there is no evidence that any of these potential GPI                                    intermediates with three mannose residues confirmed this
precursors are actually intermediates in GPI biosynthesis, the                                   localization. Pulse-chase analysis showed that the PI-PLC-
most likely pathway for assembly of the membrane anchor based                                    insensitive GPI precursors were not metabolized into subsequent
on the occurrence of the different glycolipids is shown in red in                                intermediates, consistent with the idea that these intermediates
Figure 4. In this pathway, a phosphoethanolamine is added to                                     had been translocated to the ER lumen and were no longer
ManGlcN-PI(acyl) before addition of the next two mannoses.                                       accessible to the cytosolicly oriented biosynthetic enzymes. The
The phosphoethanolamine attached to the 6-position of the third                                  finding that expression of a Trypanosoma brucei GPI-specific
mannose and which is used for linkage to protein is added next. A                                phospholipase D on the cytosolic face of the ER in Leishmania
third phosphoethanolamine is then added to the second mannose                                    major resulted in the loss of ethanolamine-containing GPI
to form a three-mannose, three-phosphoethanolamine GPI pre-                                      precursors led to the conclusion that the later reactions have a
cursor. What the linkages of the extra phosphoethanolamines to                                   similar topology in this parasite as in trypanosomes [118]. While
the mannoses are, and whether the two- or three-phospho-                                         it is apparent that the GPI precursor must be translocated to the
ethanolamine-containing precursors are used as protein anchors,                                  lumenal face of the ER before transfer to protein, nothing is
is not known.                                                                                    known about this process.
   The donor of the two phosphoethanolamines attached to the
first two mannoses is not known. The finding that the class F
Thy- 1-deficient mutant synthesizes the Etn-P-ManGlcN-PI(acyl)                                   REGULATION OF GPI BIOSYNTHESIS
precursor demonstrates that different enzymes catalyse the ad-
dition of the phosphoethanolamines on the first and third                                        Very little is known about the regulation of GPI biosynthesis in
mannoses [38,39]. Identification of the endogenous donors of the                                 any organism. In bloodstream trypanosomes, inhibition of GPI
extra phosphoethanolamines and the enzymes involved in these                                     biosynthesis by up to 900O did not affect the transfer of this
reactions will require further study in mammalian cell-free                                      glycolipid to VSG [88,100]. Therefore GPI appears to be made
systems.                                                                                         in a 5-10-fold excess of what is needed for protein anchoring. In
                                                                                                 this case, GPI biosynthesis may not be regulated.
                                                                                                    In mammalian cells, the levels of both GPI precursors and
TOPOLOGY OF GPI BIOSYNTHESIS                                                                     GPI-anchored proteins are much lower. That GPI biosynthesis is
The biosynthesis of GPI and the transfer of this glycolipid to                                   regulated in these cells is suggested by the results reported by
protein occur in the ER [63,64]. Protease protection experiments                                 Schubert et al. [119], who found that treatment of normal
have established that the later reaction takes place in the lumen                                peripheral blood lymphocytes with phorbol esters and the Ca 2'
of this organelle [64]. Investigation into the topology of GPI                                   ionophore A23187 increased the synthesis of GPI precursors.
biosynthesis in mammalian cells [63], Trypanosoma brucei                                         Because the effect of this stimulation on the level of GPI-
[64,117] and Leishmania major [1 18] has shown that either part or                               anchored proteins in these cells is not known, whether there is a
all of this pathway, depending on the species, occurs on the                                     limiting amount of GPI precursor available for transfer to
cytosolic face of the ER. The first two GPI intermediates,                                       protein in unstimulated cells cannot be determined from this
GlcNAc-PI and GlcN-PI, were > 70 0O sensitive to hydrolysis by                                   study. Other evidence, however, suggests that GPI precursors
PI-PLC in mammalian cell microsomes and Streptolysin 0-                                          may be made in excess of the need for protein attachment in
permeabilized cells [63]. Therefore the first two reactions in the                               mammalian cells. In characterizing a murine T-cell hybridoma
pathway in mammalian cells are primarily localized in the                                        mutant which produced < 1 00 of the wild-type level of mature
cytoplasmic leaflet of the ER. The resistance to P1-PLC cleavage                                 GPI precursor, Thomas et al. [120] found that the level of
conferred by the inositol acyl group on all the subsequent                                       expression of two GPI-anchored proteins was decreased to 17 0,
intermediates in mammalian cells has prevented the localization                                  (Thy- 1) and 2.4 00 (Ly-6A) of the normal levels. If GPI were the
of the later reactions with this methodology. In trypanosomes,                                   limiting component for membrane anchoring, then a 99 ",,
where inositol acylation usually does not occur until biosynthesis                               decrease in the level of the two proteins would be expected.
of the core is complete, all the GPI intermediates were found to                                  Further evidence that GPI precursors are made in excess of the
be predominantly on the cytosolic face of microsomal vesicles                                    cell's needs is provided by Puoti and Conzelmann [42]. These
                                                                                           Biosynthesis of glycosylphosphatidylinositol membrane anchors                             369

investigators quantified inositol-labelled GPI precursors and                                   8 Futerman, A. H., Low, M. G., Ackermann, K. E., Sherman, W. R. and Silman, I. (1985)
proteins in SIA lymphoma cells and found that the amount of                                       Biochem. Biophys. Res. Commun. 129, 312-317
radiolabel present in protein was only 16-28 % of that found in                                 9 Roberts, W. L. and Rosenberry, T. L. (1985) Biochem. Biophys. Res. Commun. 133,
the GPI precursors. However, these results may be misleading                                      621-627
                                                                                               10 Tse, A. G. D., Barclay, A. N., Watts, A. and Williams, A. F. (1985) Science 230,
because it is not known whether the precursors identified in                                      1003-1 008
either of these studies represented the glycolipid which is actually                           11 Low, M. G. (1987) Biochem. J. 244, 1-13
transferred to protein.                                                                        12 McConville, M. J. and Ferguson, M. A. J. (1993) Biochem. J. 294, 305-324
   The second reaction in GPI biosynthesis, the deacetylation of                               13 Cross, G. A. M. (1990) Annu. Rev. Cell Biol. 6, 1-39
GlcNAc-PI, has been found to be stimulated in vitro in mam-                                    14 Cross, G. A. M. (1975) Parasitology 71, 393-417
malian cells by GTP [61]. This stimulation is specific for the                                 15 Bangs, J. D., Hereld, D., Krakow, J. L., Hart, G. W. and Englund, P. T. (1985) Proc.
guanosine-containing nucleotide triphosphate and results in an                                    Natl. Acad. Sci. U.S.A. 82, 3207-3211
                                                                                               16 Ferguson, M. A. J., Duszenko, M., Lamont, G. S., Overath, P. and Cross, G. A. M.
approx. 5-fold acceleration in the initial rate of GlcNAc-PI                                      (1986) J. Biol. Chem. 261, 356-362
deacetylation by lymphoma cell microsomes. Guanosine 5'-[y-                                    17 Krakow, J. L., Hereld, D., Bangs, J. D., Hart, G. W. and Englund, P. T. (1986) J. Biol.
thio]triphosphate completely blocks the stimulation by GTP,                                       Chem. 261, 12147-12153
indicating that the nucleotide triphosphate must be hydrolysed                                 18 Menon, A. K., Mayor, S., Ferguson, M. A. J., Duszenko, M. and Cross, G. A. M.
to affect this reaction. Although the deacetylation of GlcNAc-PI                                  (1988) J. Biol. Chem. 263, 1970-1977
is not expected to require energy, the possibility that GTP is                                 19 Mayor, S., Menon, A. K. and Cross, G. A. M. (1990) J. Biol. Chem. 265, 617446181
hydrolysed for this purpose has not been ruled out. Alternatively,                             20 Mayor, S., Menon, A. K., Cross, G. A, M., Ferguson, M. A. J., Dwek, R. A. and
the stimulation by GTP could reflect a regulatory mechanism                                       Rademacher, T. W. (1990) J. Biol. Chem. 265, 61644173
                                                                                               21 Masterson, W. J., Doering, T. L., Hart, G. W. and Englund, P. T. (1989) Cell 56,
that involves a GTP-binding protein and affects the level of GPI                                    793-800
precursors synthesized by mammalian cells.                                                     22 Doering, T. L, Masterson, W. J., Englund, P. T. and Hart, G. W. (1989) J. Biol.
                                                                                                  Chem. 264, 11168-11173
CONCLUSIONS                                                                                    23 Menon, A. K., Schwarz, R. T., Mayor, S. and Cross, G. A. M. (1990) J. Biol. Chem.
                                                                                                  265, 9033-9042
Our understanding of GPI biosynthesis has expanded tremen-                                     24 Hyman, R. (1973) J. Natl. Cancer Inst. 50, 415-422
dously in the last 6 years since the first studies of this pathway                             25 Hyman, R. and Stallings, V. (1974) J. Natl. Cancer Inst. 52, 429-436
were reported. The sequence of reactions and most of the                                       26 Fatemi, S. H. and Tartakoff, A. M. (1988) J. Biol. Chem. 263,1288-1294
endogenous donors of the anchor components are now fairly                                      27 Conzelmann, A., Spiazzi, A., Bron, C. and Hyman, R. (1988) Mol. Cell. Biol. 8,
well defined in most systems. As the focus of investigation has                                     674-478
shifted to specific reactions in the pathway, it has become                                    28   Ferguson, M. A. J. (1992) Biochem. Soc. Trans. 20, 243-256
                                                                                               29   Ferguson, M. A. J. (1991) Curr. Opin. Struct. Biol. 1, 522-529
apparent that there are significant differences in GPI biosynthesis                            30   Englund, P. T. (1993) Annu. Rev. Biochem. 62, 121-138
in different species. These differences are well defined between                               31   Lublin, D. M. (1992) Curr. Top. Microbiol. Immunol. 178,141-162
trypanosomes and mammalian cells and have been suggested                                       32   Bangs, J. D., Doering, T. L., Englund, P. T. and Hart, G. W. (1988) J. Biol. Chem.
between yeast and mammalian cells. Therefore specific steps in the                                263, 17697-17705
GPI biosynthetic pathway are potential targets for anti-parasitic                              33 Bangs, J. D., Andrews, N. W., Hart, G. W. and Englund, P. T. (1986) J. Cell Biol.
and anti-fungal agents. Future investigation of the enzymes                                       103, 255-263
                                                                                               34 Pingel, S. and Duszenko, M. (1992) Biochem. J. 263, 479-485
and reaction mechanisms of GPI biosynthesis will facilitate the                                35 Pingel, S., Field, R. A., Guther, M. L. S., Duszenko, M. and Ferguson, M. A. J. (1995)
development of inhibitors which could be used as drugs. These                                     Biochem. J. 309, 877-882
inhibitors may also be useful for elucidating the function of the                              36 Mayor, S., Menon, A. K. and Cross, G. A. M. (1992) J. Biol. Chem. 267, 754-761
GPI anchor for various proteins in different species or cell types.                            37 Tomavo, S., Dubremetz, J.-F. and Schwarz, R. T. (1992) J. Biol. Chem. 267,
   The regulation of GPI biosynthesis is another process for                                      21 446-21 458
which a greater understanding would provide insight into the                                   38 Puoti, A., Desponds, C., Fankhauser, C. and Conzelmann, A. (1991) J. Biol. Chem.
significance of GPI anchoring in different cells. Important                                       266, 21051-21059
                                                                                               39 Kamitani, T., Menon, A. K., Hallaq, Y., Warren, C. D. and Yeh, E. T. H. (1992) J. Biol.
questions to be addressed are: (1) are the synthesis of GPI and                                   Chem. 267, 24611-24619
the proteins which become GPI-anchored co-ordinately                                           40 Hirose, S., Prince, G. M., Sevlever, D., Ravi, L., Rosenberry, T. L., Ueda, E. and
regulated?; (2) is the regulation of GPI biosynthesis coupled to                                  Medof, M. E. (1992) J. Biol. Chem. 267,16968-16974
cell growth and, if so, what signal transduction pathways are                                  41 Puoti, A. and Conzelmann, A. (1993) J. Biol. Chem. 268, 7215-7224
involved?; and (3) if the amount of GPI synthesized is limited, is                             42 Puoti, A. and Conzelmann, A. (1992) J. Biol. Chem. 267, 22673-22680
there prioritized addition to some proteins in the cell over                                   43 Fankhauser, C., Homans, S. W., Thomas-Oates, J. E., McConville, M. J., Desponds,
others? Hopefully, the next few years of study will provide                                       C., Conzelmann, A. and Ferguson, M. A. J. (1993) J. Biol. Chem. 268, 26365-26374
answers to these and other questions about GPI biosynthesis.                                   44 Stadler, J., Keenan, T. W., Bauer, G. and Gerisch, G. (1989) EMBO J. 8, 371-377
                                                                                               45 Ferguson, M. A. J. and Cross, G. A. M. (1984) J. Biol. Chem. 259, 3011-3015
  would like to thank Dr. Taroh Kinohita, Dr. Michael Ferguson, Dr. Peter Orlean and           46 Roberts, W. L., Myher, J. J., Kuksis, A. and Rosenberry, T. L. (1988) Biochem.
Dr. Anant Menon for sharing their unpublished results. This work was supported by                 Biophys. Res. Commun. 150, 271-277
the American Cancer Society Grant BE-1 29.                                                     47 Luhrs, C. A. and Slomiamy, B. L. (1989) J. Biol. Chem. 264, 21446-21449
                                                                                               48 Schneider, P., Ferguson, M. A. J., McConville, M. J., Mehlert, A., Homans, S. W. and
                                                                                                   Bordier, C. (1990) J. Biol. Chem. 265,16955-16964
REFERENCES                                                                                     49 Rosen, G., Pahlsson, P., Londner, M. V., Westerman, M. E. and Nilsson, B. (1989)
 1 Low, M. G. and Finean, J. B. (1977) Biochem. J. 167, 281-284                                   J. Biol. Chem. 264,10457-10463
 2 lkezawa, H., Yamanegi, M., Taguchi, R., Miyashita, T. and Ohyabu, T. (1976)                 50 Turco, S. J., Orlandi, P. A., Jr., Homans, S. W., Ferguson, M. A. J., Dwek, R. A. and
   Biochim. Biophys. Acta 450,154-164                                                             Rademacher, T. W. (1989) J. Biol. Chem. 264, 6711-6715
 3 Taguchi, R. and Ikezawa, H. (1978) Arch. Biochem. Biophys. 186, 196-201                     51 Ferguson, M. A. J. (1992) Biochem. J. 284, 297-300
 4 Low, M. G. and Finean, J. B. (1978) Biochim. Biophys. Acta 508, 565-570                     52 Roberts, W. L., Myher, J. J., Kuksis, A., Low, M. G. and Rosenberry, T. L. (1988)
 5 Low, M. G. and Zilversmit, D. B. (1980) Biochemistry 19, 3913-3918                             J. Biol. Chem. 263,18766-18775
 6 Alvarez, J. F., Cabello, M. A., Feliu, J. E. and Mato, J. M. (1987) Biochem. Biophys.       53 Roberts, W. L., Kim, B. H. and Rosenberry, T. L. (1987) Proc. Natl. Acad. Sci. U.S.A.
   Res. Commun. 147, 765-771                                                                      84, 7817-7821
 7 Ferguson, M. A. J., Low, M Gand Cross, G. A. M. (1985) J. Biol. Chem. 260,                  54 Walter, E. I., Roberts, W. L, Rosenberry, T. L, Ratnoff, W. D. and Medof, M. E.
   14547-1 4555                                                                                   (1990) J. Immunol. 1", 1030-1036
370               V. L. Stevens

55 Low, M. G., Stiernberg, J., Waneck, G. L., Flavell, R. A. and Kincade, P. W. (1988)            88 Masterson, W. J. and Ferguson, M. A. J. (1991) EMBO J. 10, 2041-2045
    J. Immunol. Methods 113, 101-111                                                              89 Urakaze, M., Kamitani, T., DeGasperi, R., Sugiyama, E., Chang, H. and Yeh, E. T. H.
56 Field, M. C., Menon, A. K. and Cross, G. A. M. (1991) J. Biol. Chem. 266,                         (1992) J. Biol. Chem. 267, 6459-6462
    8392-8400                                                                                     90 Stevens, V. L and Zhang, H. (1994) J. Biol. Chem. 269, 31397-31403
57 McConville, M. J., Schneider, P., Proudfoot, L., Masterson, C. and Ferguson, M. A. J.          91 Chapman, A., Fujimoto, K. and Kornfeld, S. (1980) J. Biol. Chem. 255, 4441-4446
    (1994) Braz. J. Med. Biol. Res. 27, 139-144                                                   92 Singh, N. and-Tartakoff, A. M. (1991) Mol. Cell. Biol. 11, 391-400.
58 Stevens, V. L. and Raetz, C. R. H. (1991) J. Biol. Chem. 266,10039-10042                       93 Camp, L. A., Chauhan, P., Farrar, J. D. and Lehrman, M. A. (1993) J. Biol. Chem.
59 Hirose, S., Ravi, L., Hazra, S. and Medof, M. E. (1991) Proc. Natl. Acad. Sci. U.S.A.             268, 6721-6728
    88, 3762-3766                                                                                 94 Orlean, P. (1990) Mol. Cell. Biol. 10, 5796-5805
60 Sugiyama, E., DeGasperi, R., Urakaze, M., Chang, H., Thomas, L. J., Hyman, R.,                 95 DeGasperi, T., Thomas, L. J., Sugiyama, E., Chang, H. M., Beck, P. J., Orlean, P.,
    Warren, C. D. and Yeh, E. T. H. (1991) J. Biol. Chem. 266, 12119-12122                           Albright, C., Waneck, G., Sambrook, J. F., Warren, C. D. and Yeh, E. T. H. (1990)
61 Stevens, V. L. (1993) J. Biol. Chem. 268, 9718-9724                                               Science 250, 988-991
62 Costello, L. C. and Orlean, P. (1992) J. Biol. Chem. 267, 8599-8603                            96 Menon, A. K., Mayor, S. and Schwarz, R. T. (1990) EMBO J. 9, 4249-4258
63 Vidugiriene, J. and Menon, A. K. (1993) J. Cell Biol. 121, 987-996                             97 Kinoshita, T., Takahashi, M., Inoue, N., Miyata, T. and Takeda, J. (1994) Braz. J.
64 Vidugiriene, J. and Menon, A. K. (1994) J. Cell Biol. 127, 333-341                                Med. Biol. Res. 27,127-132
65 Sipos, G., Puoti, A. and Conzelmann, A. (1994) EMBO J. 13, 2789-2796                           98 Lisanti, M. P., Field, M. C., Caras, I. W., Menon, A. K. and Rodriguez-Boulan, E.
66 Inoue, N., Kinoshita, T., Orii, T. and Takeda, J. (1993) J. Biol. Chem. 268,                      (1991) EMBO J. 10, 1969-1977
    6882-6885                                                                                     99 Pan, Y.-T., Kamitani, T., Bhuvaneswaran, C., Hallaq, Y., Warren, C. D., Yeh, E. T. H.
67 Stevens, V. L. and Raetz, C. R. H. (1990) J. Biol. Chem. 265, 15653-15658                         and Elbein, A. D. (1992) J. Biol. Chem. 267, 21250-21255
68 Doering, T. L., Pessin, M. S., Hart, G. W., Raben, D. M. and Englund, P. T. (1994)            100 Ralton, J. E., Milne, K. G., Guther, M. L. S., Field, R. A. and Ferguson, M. A. J.
     Biochem. J. 299, 741-746                                                                        (1993) J. Biol. Chem. 268, 24183-24189
69 Leidich, S. D., Drapp, D. A. and Orlean, P. (1994) J. Biol. Chem. 269,10193-10196             101 Field, M. C., Medina-Acosta, E. and Cross, G. A. M. (1993) J. Biol. Chem. 268,
70 Miyata, T., Takeda, J., lida, Y., Yamada, N., Inoue, N., Takahashi, M., Maeda, K.,                9570-9577
     Kitani, T. and Kinoshita, T. (1993) Science 259, 1318-1320                                  102 Sevlever, D. and Rosenberry, T. L. (1993) J. Biol. Chem. 268, 10938-10945
71 Kawagoe, K., Takeda, J., Endo, Y. and Kinoshita, T. (1994) Genomics 23,                       103 Menon, A. K. and Stevens, V. L. (1992) J. Biol. Chem. 267, 15277-15280
     566-574                                                                                     104 Menon, A. K., Eppinger, M., Mayor, S. and Schwarz, R. T. (1993) EMBO J. 12,
72 Takeda, J., Miyata, T., Kawagoe, K., lida, Y., Endo, Y., Fujita, T., Takahashi, M.,               1907-1914
     Kitani, T. and Kinoshita, T. (1993) Cell 73, 703-711                                        105 Masterson, W. J., Raper, J., Doering, T. L., Hart, G. W. and Englund, P. T. (1990)
73 lida, Y., Takeda, J., Miyata, T., Inoue, N., Nishimura, J., Kitani, T., Maeda, K. and             Cell 62, 73-80
     Kinoshita, T. (1994) Blood 83, 3126-3131                                                    106 Guther, M. L. S., Masterson, W. J. and Ferguson, M. A. J. (1994) J. Biol. Chem.
74 Ware, R. E., Howard, T. A., Kamitani, T., Chang, H.-M., Yeh, E. T. H. and Seldin,                 269, 18694-18701
     M. F. (1994) Blood 83, 3753-3757                                                            107 Doering, T. L., Pessin, M. S., Hoff, E. F., Hart, G. W., Raben, D. M. and Englund,
75 Kamitani, T., Chang, H.-M., Rollins, C., Waneck, G. L. and Yeh, E. T. H. (1993)                    P. T. (1993) J. Biol. Chem. 268, 9215-9222
     J. Biol. Chem. 268, 20733-20736                                                             108 Doering, T. L., Raper, J., Buxbaum, L. U., Adams, S. P., Gordon, J. I., Hart, G. W.
76 Singh, N., Singleton, D. and Tartakoff, A. M. (1991) Mol. Cell. Biol. 11, 2362-2374               and Englund, P. T. (1991) Science 252, 1851-1854
77 Milne, K. G., Ferguson, M. A. J. and Masterson, W. J. (1992) Eur. J. Biochem. 208,            109 Doering, T. L., Lu, T., Werbovetz, K. A., Gokel, G. W., Hart, G. W., Gordon, J. I. and
     309-314                                                                                          Englund, P. T. (1994) Proc. Natl. Acad. Sci. U.S.A. 91, 9735-9739
 78 Rosse, W. F. (1990) Annu. Rev. Med. 41, 431-436                                              110 Buxbaum, L. U., Raper, J., Opperdoes, F. R. and Englund, P. T. (1994) J. Biol.
 79 Rosse, W. F. (1990) Blood 75, 1595-1601                                                           Chem. 269, 30212-30220
80 Hansch, G. M., Weller, P. F. and Nicholson-Weller, A. (1988) Blood 72, 1089-1092              111 Krakow, J. L., Doering, T. L., Masterson, W. J., Hart, G. W. and Englund, P. T.
 81 Armstrong, C., Schuber, J., Ueda, E., Knez, J. J., Gelperin, D., Hirose, S., Silber, R.,          (1989) Mol. Biochem. Parasitol. 36, 263-270
     Holland, S., Schmidt, R. E. and Medof, M. E. (1992) J. Biol. Chem. 267,                     112 Major, S., Menon, A. K. and Cross, G. A. M. (1991) J. Cell Biol. 114, 61-71
     25347-25351                                                                                 113 Field, M. C., Menon, A. K. and Cross, G. A. M. (1991) EMBO J. 10, 2731-2739
 82 Takahashi, M., Takeda, J., Hirose, S., Hyman, R., Inoue, N., Miyata, T., Ueda, E.,           114 Conzelmann, A., Puoti, A., Lester, R. L. and Desponds, C. (1992) EMBO J. 11,
     Kitani, T., Medof, M. E. and Kinoshita, T. (1993) J. Exp. Med. 177, 517-521                        457-466
 83 Hillmen, P., Bessler, M., Mason, P. J., Watkins, W. M. and Luzzatto, L. (1993) Proc.         115 Riezman, H., Horvath, A., Manning-Krieg, U. and Mowa, R. (1994) Braz. J. Med.
      Natl. Acad. Sci. U.S.A. 90, 5272-5276                                                          Biol. Res. 27, 323-326
 84 Miyata, T., Yamada, N., lida, Y., Nishimura, J., Takeda, J., Kitani, T. and Kinoshita, T.    116 Wong, Y. W. and Low, M. G. (1994) Biochem. J. 301, 205-209
      (1994) N. Engl. J. Med. 330, 249-255                                                       117 Menon, A. K. and Vidugiriene, J. (1994) Braz. J. Med. Biol. Res. 27, 167-175
 85 Hirose, S., Mohney, R. P., Mutka, S. C., Ravi, L., Singleton, D. R., Perry, G., Tartakoff,   118 Mensa-Wilmot, K., LeBowitz, J. H., Chang, K.-P., Al-Qahtani, A., McGuire, B. S.,
     A. M. and Medof, M. E. (1992) J. Biol. Chem. 267, 5272-5278                                     Tucker, S. and Morris, J. C. (1994) J. Cell Biol. 124, 935-947
 86 Mohney, R. P., Knez, J. J., Ravi, L., Sevlever, D., Rosenberry, T. L., Hirose, S. and        119 Schubert, J., Schmidt, R. E. and Medof, M. E. (1993) J. Biol. Chem. 268,
      Medof, M. E. (1994) J. Biol. Chem. 269, 6536-6542                                              6281-6287
 87 Milne, K. G., Field, R. A., Masterson, W. J., Cottaz, S., Brimacombe, J. S. and              120 Thomas, L. J., Urakaze, M., DeGasperi, R., Kamitani, T., Sugiyama, E., Chang, H.,
      Ferguson, M. A. J. (1994) J. Biol. Chem. 269, 16403-16408                                      Warren, C. D. and Yeh, E. T. H. (1992) J. Clin. Invest. 89, 1172-1177

Shared By: