A protein interaction map of Drosophila melanogaster.

Drosophila melanogaster is a proven model system for many aspects of human biology. Here we present a two-hybrid-based protein-interaction map of the fly proteome. A total of 10,623 predicted transcripts were isolated and screened against standard and normalized complementary DNA libraries to produce a draft map of 7048 proteins and 20,405 interactions. A computational method of rating two-hybrid interaction confidence was developed to refine this draft map to a higher confidence map of 4679 proteins and 4780 interactions. Statistical modeling of the network showed two levels of organization: a short-range organization, presumably corresponding to multiprotein complexes, and a more global organization, presumably corresponding to intercomplex connections. The network recapitulated known pathways, extended pathways, and uncovered previously unknown pathway components. This map serves as a starting point for a systems biology modeling of multicellular organisms, including humans.

Cloning, expression, and localization of a novel gamma-adaptin-like molecule.

We describe the cloning, expression, and localization of gamma2-adaptin, a novel isoform of gamma-adaptin. The predicted human and mouse gamma2-adaptin proteins are approximately 90 kDa and 64.4% and 61.7%) identical to gamma-adaptin, respectively. gamma2-Adaptin was localized to the Golgi, its localization distinct from gamma-adaptin. The membrane association of gamma- and gamma2-adaptin could further be distinguished by differential sensitivity to the fungal metabolite brefeldin A, gamma2-adaptin binding being insensitive to drug treatment. Together, these results suggest that gamma2-adaptin plays a role in membrane transport distinct from that played by gamma-adaptin.

ADP-Ribosylation factor 1 (ARF1) regulates recruitment of the AP-3 adaptor complex to membranes.

Small GTP-binding proteins such as ADP- ribosylation factor 1 (ARF1) and Sar1p regulate the membrane association of coat proteins involved in intracellular membrane trafficking. ARF1 controls the clathrin coat adaptor AP-1 and the nonclathrin coat COPI, whereas Sar1p controls the nonclathrin coat COPII. In this study, we demonstrate that membrane association of the recently described AP-3 adaptor is regulated by ARF1. Association of AP-3 with membranes in vitro was enhanced by GTPgammaS and inhibited by brefeldin A (BFA), an inhibitor of ARF1 guanine nucleotide exchange. In addition, recombinant myristoylated ARF1 promoted association of AP-3 with membranes. The role of ARF1 in vivo was examined by assessing AP-3 subcellular localization when the intracellular level of ARF1-GTP was altered through overexpression of dominant ARF1 mutants or ARF1- GTPase-activating protein (GAP). Lowering ARF1-GTP levels resulted in redistribution of AP-3 from punctate membrane-bound structures to the cytosol as seen by immunofluorescence microscopy. In contrast, increasing ARF1-GTP levels prevented redistribution of AP-3 to the cytosol induced by BFA or energy depletion. Similar experiments with mutants of ARF5 and ARF6 showed that these other ARF family members had little or no effect on AP-3. Taken together, our results indicate that membrane recruitment of AP-3 is promoted by ARF1-GTP. This finding suggests that ARF1 is not a regulator of specific coat proteins, but rather is a ubiquitous molecular switch that acts as a transducer of diverse signals influencing coat assembly.

Altered expression of a novel adaptin leads to defective pigment granule biogenesis in the Drosophila eye color mutant garnet.

Drosophila eye pigmentation defects have thus far been attributed to mutations in genes encoding enzymes required for biosynthesis of pigments and to ABC-type membrane transporters for pigments or their precursors. We report here that a defect in a gene encoding a putative coat adaptor protein leads to the eye color defect of garnet mutants. We first identified a human cDNA encoding delta-adaptin, a structural homolog of the alpha- and gamma-adaptin subunits of the clathrin coat adaptors AP-1 and AP-2, respectively. Biochemical analyses demonstrated that delta-adaptin is a component of the adaptor-like complex AP-3 in human cells. We then isolated a full-length cDNA encoding the Drosophila ortholog of delta-adaptin and found that transcripts specified by this cDNA are altered in garnet mutant flies. Examination by light and electron microscopy indicated that these mutant flies have reduced numbers of eye pigment granules, which correlates with decreased levels of both pteridine (red) and ommachrome (brown) pigments. Thus, the eye pigmentation defect in the Drosophila garnet mutant may be attributed to compromised function of a coat protein involved in intracellular transport processes required for biogenesis or function of pigment granules.

Beta3A-adaptin, a subunit of the adaptor-like complex AP-3.

Recent studies have described a widely expressed adaptor-like complex, named AP-3, which is likely involved in protein sorting in exocytic/endocytic pathways. The AP-3 complex is composed of four distinct subunits. Here, we report the identification of one of the subunits of this complex, which we call beta3A-adaptin. The predicted amino acid sequence of beta3A-adaptin reveals that the protein is closely related to the neuron-specific protein beta-NAP (61% overall identity) and more distantly related to the beta1- and beta2-adaptin subunits of the clathrin-associated adaptor complexes AP-1 and AP-2, respectively. Sequence comparisons also suggest that beta3A-adaptin has a domain organization similar to beta-NAP and to beta1- and beta2-adaptins. beta3A-adaptin is expressed in all tissues and cells examined. Co-purification and co-precipitation analyses demonstrate that beta3A-adaptin corresponds to the approximately 140-kDa subunit of the ubiquitous AP-3 complex, the other subunits being delta-adaptin, p47A (now called mu3A) and sigma3 (A or B). beta3A-adaptin is phosphorylated on serine residues in vivo while the other subunits of the complex are not detectably phosphorylated. beta3A-adaptin is not present in significant amounts in clathrin-coated vesicles. The characteristics of beta3A-adaptin reported here lend support to the idea that AP-3 is a structural and functional homolog of the clathrin-associated adaptors AP-1 and AP-2.

AP-3: an adaptor-like protein complex with ubiquitous expression.

We have identified two closely related human proteins (sigma3A and sigma3B) that are homologous to the small chains, sigma1 and sigma2, of clathrin-associated adaptor complexes. Northern and Western blot analyses demonstrate that the products of both the sigma3A and sigma3B genes are expressed in a wide variety of tissues and cell lines. sigma3A and sigma3B are components of a large complex, named AP-3, that also contains proteins of apparent molecular masses of 47, 140 and 160 kDa. In non-neuronal cells, the 47 kDa protein most likely corresponds to the medium chain homolog p47A, and the 140 kDa protein is a homolog of the neuron-specific protein beta-NAP. Like other members of the medium-chain family, the p47A chain is capable of interacting with the tyrosine-based sorting signal YQRL from TGN38. Immunofluorescence microscopy analyses show that the sigma3-containing complex is present both in the area of the TGN and in peripheral structures, some of which contain the transferrin receptor. These results suggest that the sigma3 chains are components of a novel, ubiquitous adaptor-like complex involved in the recognition of tyrosine-based sorting signals.

Copper-dependent degradation of the Saccharomyces cerevisiae plasma membrane copper transporter Ctr1p in the apparent absence of endocytosis.

The cell surface protein repertoire needs to be regulated in response to changes in the extracellular environment. In this study, we investigate protein turnover of the Saccharomyces cerevisiae plasma membrane copper transporter Ctr1p, in response to a change in extra-cellular copper levels. As Ctr1p mediates high affinity uptake of copper into the cell, modulation of its expression is expected to be involved in copper homeostasis. We demonstrate that Ctr1p is a stable protein when cells are grown in low concentrations of copper, but that exposure of cells to high concentrations of copper (10 microM) triggers degradation of cell surface Ctr1p. This degradation appears to be specific for Ctr1p and does not occur with another yeast plasma membrane protein tested. Internalization of some Ctr1p can be seen when cells are exposed to copper. However, yeast mutant strains defective in endocytosis (end3, end4 and chc1-ts) and vacuolar degradation (pep4) exhibit copper-dependent Ctr1p degradation, indicating that internalization and delivery to the vacuole is not the principal mechanism responsible for degradation. In addition, a variant Ctr1p with a deletion in the cytosolic tail is not internalized upon exposure of cells to copper, but is nevertheless degraded. These observations indicate that proteolysis at the plasma membrane most likely explains copper-dependent turnover of Ctr1p and point to the existence of a novel pathway in yeast for plasma membrane protein turnover.

Antibacterial proteins of granulocytes differ in interaction with endotoxin. Comparison of bactericidal/permeability-increasing protein, p15s, and defensins.

Bactericidal/permeability-increasing protein (BPI), antibacterial 15-kDa protein isoforms (p15s), and defensins (neutrophil peptides or NPs) are granule-associated antibacterial proteins of polymorphonuclear leukocytes (PMN) that have both direct and synergistic growth inhibitory activity against Gram-negative bacteria. In this study, we have compared in vitro the abilities of these antibacterial proteins, alone and in combination, to inhibit the endotoxic activity of isolated LPS and whole bacteria. All three proteins blocked endotoxin activity in: 1) the Limulus amoebocyte lysate assay, 2) priming of PMN for enhanced arachidonate release, and 3) stimulating leukocyte oxidase activity in 1% blood. However, the proteins differ markedly in both relative potency (BPI >> p15s = NP1) in the presence of the plasma LPS-binding protein and in the range of LPS chemotypes that can be inhibited. BPI potently neutralizes LPS of any chemotype, but p15s and defensins are less active against long-chain (S-type) LPS. In whole blood ex vivo, the p15s and NP1 are approximately 1000-fold less potent than BPI, but at subinhibitory doses act in synergy with BPI to inhibit the TNF-inducing activity of a serum-resistant encapsulated strain of Escherichia coli (K1/r). The anti-endotoxic effects of p15 and NP1 against E. coli K1/r in whole blood appear secondary to growth arrest, because, in marked contrast to BPI, they are not evident against nonviable bacteria (pretreated with antibiotic) nor isolated LPS. Thus, BPI stands out for its ability to inhibit isolated or bacterial LPS under physiologic conditions. However, p15s and defensins may also contribute to suppression of endotoxic signaling by Gram-negative bacteria via synergistic (with BPI) growth inhibition upon extracellular release of these proteins from PMN during inflammation.

Overexpression of wild-type and mutant ARF1 and ARF6: distinct perturbations of nonoverlapping membrane compartments.

The ARF GTP binding proteins are believed to function as regulators of membrane traffic in the secretory pathway. While the ARF1 protein has been shown in vitro to mediate the membrane interaction of the cytosolic coat proteins coatomer (COP1) and gamma-adaptin with the Golgi complex, the functions of the other ARF proteins have not been defined. Here, we show by transient transfection with epitope-tagged ARFs, that whereas ARF1 is localized to the Golgi complex and can be shown to affect predictably the assembly of COP1 and gamma-adaptin with Golgi membranes in cells, ARF6 is localized to the endosomal/plasma membrane system and has no effect on these Golgi-associated coat proteins. By immuno-electron microscopy, the wild-type ARF6 protein is observed along the plasma membrane and associated with endosomes, and overexpression of ARF6 does not appear to alter the morphology of the peripheral membrane system. In contrast, overexpression of ARF6 mutants predicted either to hydrolyze or bind GTP poorly shifts the distribution of ARF6 and affects the structure of the endocytic pathway. The GTP hydrolysis-defective mutant is localized to the plasma membrane and its overexpression results in a profound induction of extensive plasma membrane vaginations and a depletion of endosomes. Conversely, the GTP binding-defective ARF6 mutant is present exclusively in endosomal structures, and its overexpression results in a massive accumulation of coated endocytic structures.

Individual and synergistic effects of rabbit granulocyte proteins on Escherichia coli.

Affinity purification of crude acid extracts of rabbit polymorphonuclear leukocytes using Escherichia coli (J5) as adsorbent yields the bactericidal/permeability-increasing protein (BPI), two 15-kD species (p15s), and the two most potent (cationic) defensin species (neutrophil peptides [NP] -1 and -2). Tested in buffered isotonic medium, the relative antibacterial potency of these proteins against E. coli J5 is BPI (IC50 0.2 nM) > p15A (10 nM) > NP -1 (400 nM). Sublethal doses of p15A or NP-1 can synergize with BPI to decrease the dose required to inhibit the growth of E. coli by up to 50-fold. BPI and p15A display similar features of antibacterial action distinct from defensin NP-1, but NP-1 acts synergistically only with BPI and not with p15A. All aspects of the combined action of BPI and NP-1 resemble those observed with higher concentrations of BPI alone, implying that NP-1 enhances BPI potency. Neither NP-1 nor p15A alter the amount of BPI binding to E. coli but BPI enhances binding of p15A to E. coli, raising the possibility that synergy between these two proteins may occur at least partially at the level of binding. The potent synergistic actions of these proteins can also be demonstrated against serum-resistant clinical isolates of encapsulated E. coli tested in whole blood and plasma ex vivo, suggesting that such combined action may contribute to host defense in vivo.

Antibacterial 15-kDa protein isoforms (p15s) are members of a novel family of leukocyte proteins.

We have previously described the isolation and initial characterization of functionally distinct 15-kDa protein isoforms (p15s) from rabbit polymorphonuclear leukocytes (PMN) that bind with high affinity to Escherichia coli and modulate the antibacterial actions of other leukocyte proteins on this Gram-negative bacterium. We now report the cloning and sequencing of two distinct cDNAs from a rabbit bone marrow library that encode p15s differing at only 2 residues (His-3, Arg-88 versus Arg-3, Trp-88). Tryptophan-directed chemical cleavage of two isoforms purified from a single rabbit confirms the existence of multiple isoforms with distinct function and primary structure in a single rabbit. The p15 cDNAs encode putative signal sequences and studies of cellular and subcellular localization indicate that the p15s are granule-associated proteins of PMN. Both purified isoforms bind avidly to lipopolysaccharide (LPS), the major component of the Gram-negative bacterial outer membrane. Analysis of the deduced primary structures of the p15s reveals homology to three other leukocyte proteins: CAP-18, an 18-kDa LPS-binding protein from rabbit PMN, pro-indolicidin, a 16-kDa precursor of an antibacterial peptide of bovine PMN, and cathelin, an 11-kDa cysteine protease inhibitor from porcine leukocytes, suggesting the existence of a novel family of leukocyte proteins with LPS-binding, antimicrobial, and protease-inhibitory activities.

Structural characterization of BPI-modulating 15 kDa proteins from rabbit polymorphonuclear leukocytes: identification of a novel family of leukocyte proteins.

We have previously described the isolation and initial characterization of 15 kDa protein isoforms (p15s) from rabbit polymorphonuclear leukocytes (PMN) that bind to Escherichia coli and modulate the antibacterial actions of other leukocyte proteins on this gram negative organism. We now report that the p15s differ in primary structure. The cloning and sequencing of two distinct p15 cDNAs from a rabbit bone marrow library reveal that two of the isoforms are closely similar in primary structure differing at only two amino acid positions. The p15 cDNAs encode putative signal sequences suggesting a granule-associated localization for these proteins. Analysis of the derived p15 primary structures reveals homology to two leukocyte proteins: CAP-18, an 18 kD lipopolysaccharide (LPS) binding protein from rabbit PMN and cathelin, an 11 kD cysteine protease inhibitor from porcine leukocytes. This structural similarity suggests the existence of a novel family of low molecular weight leukocyte proteins with potential roles in inflammation.

Bidirectional movement of a nascent polypeptide across microsomal membranes reveals requirements for vectorial translocation of proteins.

The translocation of polypeptides across the endoplasmic reticulum is a vectorial process that occurs probably through a protein channel by a mechanism as yet undetermined. Here, we demonstrate bidirectional movement of a 221 residue nascent polypeptide across microsomal membranes and provide evidence suggesting that the retrograde movement is through the translocation channel. Retrograde movement is observed only when the polypeptide is generated from a truncated transcript; addition of a stop codon after codon 221 confers vectorial movement. Retrograde movement can also be prevented by glycosylation of the nascent polypeptide, as well as by inclusion of 32 additional amino acids that may promote folding of the translocated chain. We propose that the protein translocation channel is a passive pore that does not create a directional bias in polypeptide movement and that vectorial translocation is driven by nascent chain elongation and sustained by posttranslocation events that prevent retrograde movement.

Endotoxin-neutralizing properties of the 25 kD N-terminal fragment and a newly isolated 30 kD C-terminal fragment of the 55-60 kD bactericidal/permeability-increasing protein of human neutrophils.

The bactericidal/permeability-increasing protein (BPI) of polymorphonuclear leukocytes (PMN) is a potent cytotoxin, specific for Gram-negative bacteria, that also inhibits endotoxin activity by neutralizing isolated bacterial lipopolysaccharides (LPS). We have previously shown that an isolated 25 kD N-terminal fragment of human BPI carries all the antibacterial activities of the parent 55-60 kD molecule. In this study we have compared the LPS-neutralizing activities of human holo-BPI, the N-terminal fragment and a 30 kD C-terminal fragment that we have now isolated. We show that the N-terminal fragment also has LPS-neutralizing activity as detected by inhibition (up to 95%) of (a) activation by LPS of procoagulant proteases in Limulus amebocyte lysates, (b) LPS "priming" of PMN, and (c) LPS-mediated production of tumor necrosis factor in whole human blood. Holo-BPI and the 25 kD fragment have similar neutralizing potency (in nanomolar range) in all assays toward "smooth" LPS from Escherichia coli O111:B4 and O55:B5 (possessing long chain polysaccharide or O-antigen), and "deep rough" LPS from Salmonella minnesota Re595 mutant (possessing no O-antigen). The C-terminal fragment of BPI is devoid of antibacterial activity when tested against BPI-sensitive E. coli J5, but does have endotoxin-neutralizing activity. This activity is weak relative to holo-BPI and the 25 kD N-terminal fragment in the Limulus and PMN-priming assay, but is comparable for inhibition of TNF production in whole blood. We conclude that the principal determinants for LPS recognition and neutralization, like those for antibacterial action, reside in the N-terminal half of the BPI molecule, but that sites within the C-terminal half can also contribute to BPI-LPS interaction once LPS is detached from the bacterial envelope.

Structural and functional organization of the human neutrophil 60 kDa bactericidal/permeability-increasing protein.

We have isolated, after limited proteolysis of the bactericidal/permeability-increasing protein (BPI) of human polymorphonuclear leukocytes (PMN), two fragments representing roughly the two halves of the BPI molecule. The 25 kDa N-terminal fragment possesses all the antibacterial activities of the 60 kDa parent protein, while the ca. 30 kDa C-terminal fragment has no detectable activity. The 25 kDa fragment is as potent on a molar basis as holo-human BPI against rough Escherichia coli, is more potent than holo-BPI against more resistant smooth E. coli, and retains the specificity of BPI toward Gram-negative bacteria. The findings suggest that all of the molecular determinants of the antibacterial properties of BPI reside within the N-terminal half of the molecule, implying a novel structural/functional organization for a cytotoxic protein.

Isolation of two isoforms of a novel 15-kDa protein from rabbit polymorphonuclear leukocytes that modulate the antibacterial actions of other leukocyte proteins.

We have recently reported the use of the highly selective and reversible binding of the potent bactericidal/permeability-increasing protein (BPI) to target Gram-negative bacteria (Escherichia coli) for its isolation from crude extracts of human polymorphonuclear leukocytes (PMN). We now report the use of the same procedure for the purification from rabbit PMN of BPI and also of a novel 15-kDa species that consists of two nearly identical isoforms. These 15-kDa proteins have no demonstrable antibacterial activities by themselves. However, one isoform (p15A) potentiates strongly and the other (p15B) weakly the early antibacterial effects of both rabbit and human BPI. Both isoforms inhibit the late lethal action of BPI. Whereas the potentiating effect is specific for BPI the inhibitory effect is seen also with another antibacterial protein of PMN granules, azurocidin. Thus, we have identified in rabbit PMN a previously unrecognized 15-kDa protein species that may modulate during phagocytosis the antimicrobial effects of BPI (and other granule proteins).

Purification of a cellular (granulocyte) and an extracellular (serum) phospholipase A2 that participate in the destruction of Escherichia coli in a rabbit inflammatory exudate.

A granule-associated phospholipase A2 from rabbit polymorphonuclear leukocytes and a closely similar phospholipase A2 from rabbit serum have been purified to near homogeneity by ion-exchange and reverse-phase chromatography. The cellular (polymorphonuclear leukocyte) phospholipase A2 has been purified greater than 100,000-fold and the extracellular (serum) phospholipase A2 approximately 60,000-fold. The NH2-terminal amino acid sequence of the ascitic fluid phospholipase A2 that we have recently purified from inflammatory exudates produced in rabbits is nearly identical (15 of 16 residues) to that of the polymorphonuclear leukocyte phospholipase A2 and completely identical (19 of 19 residues) to that of the purified serum phospholipase A2. The functional properties of these three phospholipases A2 are indistinguishable. Each enzyme is active against Escherichia coli killed by the bactericidal/permeability-increasing protein of polymorphonuclear leukocyte, a property shared only by a subset of phospholipases A2. The presence of structurally and functionally very closely similar phospholipases A2 in the cellular and extracellular compartments of an inflammatory exudate is consistent with the apparent role of these enzymes in the destruction of certain microbial invaders during the acute inflammatory response.

Cloning of the cDNA of a human neutrophil bactericidal protein. Structural and functional correlations.

The bactericidal permeability increasing protein (BPI) is a 50-60-kDa membrane-associated protein isolated from granules of polymorphonuclear leukocytes. A full-length cDNA clone encoding human BPI has been isolated and the derived amino acid sequence reveals a structure that is consistent with previously determined biological properties. BPI may be organized into two domains: the amino-terminal half, previously shown to contain all known antimicrobial activity, contains a large fraction of basic and hydrophilic residues. In contrast, the carboxyl-terminal half contains more acidic than basic residues and includes several potential transmembrane regions which may anchor the holoprotein in the granule membrane. The cytotoxic action of BPI is limited to many species of Gram-negative bacteria; this specificity may be explained by a strong affinity of the very basic aminoterminal half for the negatively charged lipopolysaccharides that are unique to the Gram-negative bacterial envelope. The amino-terminal end of BPI exhibits significant similarity with the sequence of a rabbit lipopolysaccharide-binding protein, suggesting that both molecules share a similar structure for binding lipopolysaccharides.

A 25-kDa NH2-terminal fragment carries all the antibacterial activities of the human neutrophil 60-kDa bactericidal/permeability-increasing protein.

We have isolated, after limited proteolysis of the bactericidal/permeability-increasing protein (BPI) of human neutrophils, a 25-kDa fragment that possesses the bactericidal and envelope-altering activities of the 60-kDa parent protein. On a molar basis, the fragment is as potent as holo-human BPI against rough Escherichia coli, is more potent than holo-BPI against more resistant smooth E. coli, and retains the specificity of BPI toward Gram-negative bacteria. NH2-terminal amino acid sequence analysis shows that the fragment is derived from the NH2 terminus of the BPI molecule. These findings suggest that all of the molecular determinants of the antibacterial properties of BPI reside within the NH2-terminal 25-kDa segment, implying a novel structural/functional organization for a cytotoxic protein.