Crystal structure of an enzyme displaying both inositol-polyphosphate-1-phosphatase and 3'-phosphoadenosine-5'-phosphate phosphatase activities: a novel target of lithium therapy.

Lithium cations exert profound and selective psychopharmacological effects on ameliorate manic-depressive psychosis. Although lithium is an effective drug for both treatment and prophylaxis of bipolar disorder, the precise mechanism of action is not well understood. Lithium acts as both an uncompetitive and non-competitive inhibitor of several lithium- sensitive phosphatases with regard to substrate and magnesium cofactor, respectively. In this work, we report the crystal structure and reaction mechanism of Rattus norvegicus 3'-phosphoadenosine 5'-phosphate and inositol 1,4-bisphosphate phosphatase (RnPIP), a recently identified target of lithium therapy. This Li(+)-sensitive enzyme plays a crucial role in several cellular processes, such as RNA processing, sulphation reactions and probably inositol recycling. RnPIP specifically removes the 3'-phosphate group of 3'-phosphoadenosine 5'-phosphate (PAP) and the 1'-phosphate group of inositol 1,4-bisphosphate (I(1),(4)P(2)) producing AMP and inositol 4'-phosphate, respectively. The crystal structure of RnPIP complexed with AMP, Pi and magnesium ions at 1.69 A resolution provides insight into the reaction mechanism of the hydrolysis of PAP. The core fold of the enzyme is equivalent to that found in other Li(+)-sensitive phosphatases, such as inositol monophosphatase, but molecular modelling of I(1),(4)P(2) in the RnPIP active site reveals important structural determinants that accommodate this additional substrate. RnPIP is potently inhibited by lithium and, as the accumulation of PAP inhibits a variety of proteins, including sulphotransferases and RNA processing enzymes, this dual specificity enzyme represents a potential target of lithium action, in addition to inositol monophosphatases.

The X-ray structure of the FMN-binding protein AtHal3 provides the structural basis for the activity of a regulatory subunit involved in signal transduction.

BACKGROUND: The Arabidopsis thaliana HAL3 gene product encodes for an FMN-binding protein (AtHal3) that is related to plant growth and salt and osmotic tolerance. AtHal3 shows sequence homology to ScHal3, a regulatory subunit of the Saccharomyces cerevisae serine/threonine phosphatase PPz1. It has been proposed that AtHal3 and ScHal3 have similar roles in cellular physiology, as Arabidopsis transgenic plants that overexpress AtHal3 and yeast cells that overexpress ScHal3 display similar phenotypes of improved salt tolerance. The enzymatic activity of AtHal3 has not been investigated. However, the AtHal3 sequence is homologous to that of EpiD, a flavoprotein from Staphylococcus epidermidis that recognizes a peptidic substrate and subsequently catalyzes the alpha, beta-dehydrogenation of its C-terminal cysteine residue. RESULTS: The X-ray structure of AtHal3 at 2 A resolution reveals that the biological unit is a trimer. Each protomer adopts an alpha/beta Rossmann fold consisting of a six-stranded parallel beta sheet flanked by two layers of alpha helices. The FMN-binding site of AtHal3 contains all the structural requirements of the flavoenzymes that catalyze dehydrogenation reactions. Comparison of the amino acid sequences of AtHal3, ScHal3 and EpiD reveals that a significant number of residues involved in trimer formation, the active site, and FMN binding are conserved. This observation suggests that ScHal3 and EpiD might also be trimers, having a similar structure and function to AtHal3. CONCLUSIONS: Structural comparisons of AtHal3 with other FMN-binding proteins show that AtHal3 defines a new subgroup of this protein family that is involved in signal transduction. Analysis of the structure of AtHal3 indicates that this protein is designed to interact with another cellular component and to subsequently catalyze the alpha,beta-dehydrogenation of a peptidyl cysteine. Structural data from AtHal3, together with physiological and biochemical information from ScHal3 and EpiD, allow us to propose a model for the recognition and regulation of AtHal3/ScHal3 cellular partners.

The c-Jun NH(2)-terminal kinase promotes insulin resistance during association with insulin receptor substrate-1 and phosphorylation of Ser(307).

Tumor necrosis factor alpha (TNFalpha) inhibits insulin action, in part, through serine phosphorylation of IRS proteins; however, the phosphorylation sites that mediate the inhibition are unknown. TNFalpha promotes multipotential signal transduction cascades, including the activation of the Jun NH(2)-terminal kinase (JNK). Endogenous JNK associates with IRS-1 in Chinese hamster ovary cells. Anisomycin, a strong activator of JNK in these cells, stimulates the activity of JNK bound to IRS-1 and inhibits the insulin-stimulated tyrosine phosphorylation of IRS-1. Serine 307 is a major site of JNK phosphorylation in IRS-1. Mutation of serine 307 to alanine eliminates phosphorylation of IRS-1 by JNK and abrogates the inhibitory effect of TNFalpha on insulin-stimulated tyrosine phosphorylation of IRS-1. These results suggest that phosphorylation of serine 307 might mediate, at least partially, the inhibitory effect of proinflammatory cytokines like TNFalpha on IRS-1 function.

A novel target of lithium therapy.

Phosphatases converting 3'-phosphoadenosine 5'-phosphate (PAP) into adenosine 5'-phosphate are of fundamental importance in living cells as the accumulation of PAP is toxic to several cellular systems. These enzymes are lithium-sensitive and we have characterized a human PAP phosphatase as a potential target of lithium therapy. A cDNA encoding a human enzyme was identified by data base screening, expressed in Escherichia coli and the 33 kDa protein purified to homogeneity. The enzyme exhibits high affinity for PAP (K(m)<1 microM) and is sensitive to subtherapeutic concentrations of lithium (IC(50)=0.3 mM). The human enzyme also hydrolyzes inositol-1, 4-bisphosphate with high affinity (K(m)=0.4 microM), therefore it can be considered as a dual specificity enzyme with high affinity (microM range) for both PAP and inositol-1,4-bisphosphate. Hydrolysis of inositol-1,4-bisphosphate was also inhibited by lithium (IC(50)=0.6 mM). Thus, we present experimental evidence for a novel target of lithium therapy, which could explain some of the side effects of this therapy.

X-ray structure of yeast Hal2p, a major target of lithium and sodium toxicity, and identification of framework interactions determining cation sensitivity.

The product of the yeast HAL2 gene (Hal2p) is an in vivo target of sodium and lithium toxicity and its overexpression improves salt tolerance in yeast and plants. Hal2p is a metabolic phosphatase which catalyses the hydrolysis of 3'-phosphoadenosine-5'-phosphate (PAP) to AMP. It is, the prototype of an evolutionarily conserved family of PAP phosphatases and the engineering of sodium insensitive enzymes of this group may contribute to the generation of salt-tolerant crops. We have solved the crystal structure of Hal2p in complex with magnesium, lithium and the two products of PAP hydrolysis, AMP and Pi, at 1.6 A resolution. A functional screening of random mutations of the HAL2 gene in growing yeast generated forms of the enzyme with reduced cation sensitivity. Analysis of these mutants defined a salt bridge (Glu238 ellipsis Arg152) and a hydrophobic bond (Va170 ellipsis Trp293) as important framework interactions determining cation sensitivity. Hal2p belongs to a larger superfamily of lithium-sensitive phosphatases which includes inositol monophosphatase. The hydrophobic interaction mutated in Hal2p is conserved in this superfamily and its disruption in human inositol monophosphatase also resulted in reduced cation sensitivity.

The pleckstrin homology and phosphotyrosine binding domains of insulin receptor substrate 1 mediate inhibition of apoptosis by insulin.

Insulin and insulin-like growth factor 1 (IGF-1) evoke diverse biological effects through receptor-mediated tyrosine phosphorylation of insulin receptor substrate (IRS) proteins. We investigated the elements of IRS-1 signaling that inhibit apoptosis of interleukin 3 (IL-3)-deprived 32D myeloid progenitor cells. 32D cells have few insulin receptors and no IRS proteins; therefore, insulin failed to inhibit apoptosis during IL-3 withdrawal. Insulin stimulated mitogen-activated protein kinase in 32D cells expressing insulin receptors (32DIR) but failed to activate the phosphatidylinositol 3 (PI 3)-kinase cascade or to inhibit apoptosis. By contrast, insulin stimulated the PI 3-kinase cascade, inhibited apoptosis, and promoted replication of 32DIR cells expressing IRS-1. As expected, insulin did not stimulate PI 3-kinase in 32DIR cells, which expressed a truncated IRS-1 protein lacking the tail of tyrosine phosphorylation sites. However, this truncated IRS-1 protein, which retained the NH2-terminal pleckstrin homology (PH) and phosphotyrosine binding (PTB) domains, mediated phosphorylation of PKB/akt, inhibition of apoptosis, and replication of 32DIR cells during insulin stimulation. These results suggest that a phosphotyrosine-independent mechanism mediated by the PH and PTB domains promoted antiapoptotic and growth actions of insulin. Although PI 3-kinase was not activated, its phospholipid products were required, since LY294002 inhibited these responses. Without IRS-1, a chimeric insulin receptor containing a tail of tyrosine phosphorylation sites derived from IRS-1 activated the PI 3-kinase cascade but failed to inhibit apoptosis. Thus, phosphotyrosine-independent IRS-1-linked pathways may be critical for survival and growth of IL-3-deprived 32D cells during insulin stimulation.

Heterologous pleckstrin homology domains do not couple IRS-1 to the insulin receptor.

Pleckstrin homology (PH) domains occur in many signaling proteins, including substrates for the insulin receptor tyrosine kinase (IRS proteins). Based on the hypothesis that PH domains may have a common function such as membrane targeting we tested the ability of PH domains from other signaling molecules to link IRS-1 to the insulin receptor. Chimeric IRS-1 proteins containing a homologous PH domain derived from other IRS proteins (IRS-2 or Gab-1) were tyrosine phosphorylated normally in response to insulin. In contrast, heterologous PH domains from the beta-adrenergic receptor kinase, phospholipase Cgamma, or spectrin failed to mediate tyrosine phosphorylation of chimeric IRS-1 proteins, even in cells expressing high levels of insulin receptor. Moreover, IRS-1 proteins containing heterologous PH domains did not bind phosphorylated NPEY motifs derived from the insulin receptor, suggesting that the presence of these structures interfered with the function of the adjacent PTB binding domain. Thus, tyrosine phosphorylation of IRS-1 by the insulin receptor specifically requires a PH domain derived from IRS proteins.

The 60 kDa insulin receptor substrate functions like an IRS protein (pp60IRS3) in adipose cells.

The 60 kDa insulin receptor substrate in rat adipocytes that binds to the PI-3 kinase displays several functional characteristics in common with the IRS proteins; so we propose the name pp60(IRS3) to distinguish it from other tyrosine phosphorylated proteins of similar size. During insulin stimulation, p85 associated with pp60(IRS3) more rapidly than with IRS-1 or IRS-2. In mice lacking IRS-1, p85 associated more strongly with pp60(IRS3) than with IRS-2, suggesting that pp60(IRS3) provides an alternate pathway in these cells. Synthetic peptides containing two phosphorylated YMPM motifs displace pp60(IRS3) and IRS-1 from alphap85 immune complexes, suggesting that pp60(IRS3), like IRS-1, engages both SH2 domains in p85. Moreover, pp60(IRS3) binds to immobilized peptides containing a phosphorylated NPXY motif, suggesting that it contains a PTB domain with similar specificity to that in IRS-1. The cloning of pp60(IRS3) will reveal a new member of the IRS protein family which mediates insulin receptor signals in a narrow range of tissues.

The IRS-signalling system during insulin and cytokine action.

The discovery of the first intracellular substrate for insulin, IRS-1, redirected the field of diabetes research and has led to many important advances in our understanding of insulin action. Detailed analysis of IRS-1 demonstrates structure/function relationships for this modular docking molecule, including mechanisms of substrate recognition and signal propagation. Recent work has also identified other structurally similar molecules, including IRS-2, the Drosophila protein, DOS, and the Grb2-binding protein, Gab1, suggesting that this intracellular signalling strategy is conserved evolutionarily and is utilized by an expanding number of receptor systems. In fact, IRS-1 itself has been shown to be important in other growth factor and cytokine signalling systems, including growth hormone and several interleukins. Analysis of mice lacking IRS-1 confirms an important physiological role for this protein in glucose metabolism and general cell growth in the intact animal. Disregulation of the signalling pathways integrated by the IRS proteins may contribute to the pathophysiology of non-insulin-dependent diabetes mellitus or other diseases.

The IRS-2 gene on murine chromosome 8 encodes a unique signaling adapter for insulin and cytokine action.

Signal transduction by insulin and IGF-1, several interleukins (IL-2, IL-4, IL-9, IL-13), interferons, GH, and other cytokines involves IRS proteins, which link the receptors for these factors to signaling molecules with Src homology-2 domains (SH2-proteins). We recently reported the amino acid sequence of murine IRS-2; in order to examine a potential genetic role for this molecule in disease, we isolated the murine IRS-2 gene and compared the expression pattern of IRS-2 against IRS-1. Like IRS-1, IRS-2 is encoded by a single exon. Whereas IRS-1 is located on murine chromosome 1, IRS-2 is located on murine chromosome 8 near the insulin receptor. IRS-2 is expressed together with IRS-1 in many cells and tissues; however, IRS-2 predominates in murine hematopoietic cells where it may be essential for cytokine signaling; IRS-1 predominates in adipocytes and differentiated 3T3-L1 cells where it contributes to the normal insulin response. In 32D cells, IRS-1 and IRS-2 undergo differential tyrosine phosphorylation during insulin or IL-4 stimulation, as assessed indirectly by interaction with various recombinant SH2 domains. Thus, signaling specificity through the IRS proteins may be accomplished by specific expression patterns and distinct phosphorylation patterns during interaction with various activated receptors.

The pleckstrin homology domain is the principal link between the insulin receptor and IRS-1.

Interaction domains located in the NH2 terminus of IRS-1 mediate its recognition by the insulin receptor. Alignment of IRS-1 and IRS-2 reveals two homology regions: the IH1(PH) contains a pleckstrin homology (PH) domain, and the IH2(PTB) contains a phosphotyrosine binding (PTB) domain. A third region in IRS-1 called SAIN was proposed to contain another functional PTB domain. Peptide competition experiments demonstrated that the IH2(PTB) in IRS-2, like the corresponding domain in IRS-1, binds directly to peptides containing NPXY motifs. In contrast, these peptides do not bind to IH1(PH) or the SAIN regions. In 32D cells the IH1(PH) was essential for insulin-stimulated tyrosine phosphorylation of IRS-1 and insulin-stimulated phosphatidylinositol 3-kinase activity and p70(s6k) phosphorylation. In contrast, the IH2(PTB) and the SAIN regions were not required for these insulin actions; however, the IH2(PTB) improved the coupling between IRS-1 and the insulin receptor. Overexpression of the insulin receptor in 32DIR cells increased IRS-1 tyrosine phosphorylation and mediated insulin-stimulated DNA synthesis. The sensitivity of these responses was partially reduced by deletion of either the IH1(PH) or the IH2(PTB) and significantly reduced when both regions were deleted together. Thus, the PH and PTB domains equally couple IRS-1 to high levels of insulin receptor normally expressed in most cells, whereas at low levels of insulin receptors the PTB domain is inefficient and the PH domain is essential for a productive interaction.

YMXM motifs and signaling by an insulin receptor substrate 1 molecule without tyrosine phosphorylation sites.

Tyrosine phosphorylation of insulin receptor substrate 1 (IRS-1) by the activated receptors for insulin, IGF-1, and various cytokines creates binding sites for signaling proteins with Src homology 2 domains (SH2 proteins). Determining the role of specific SH2 proteins during insulin signaling has been difficult because IRS-1 possesses as many as 18 potential tyrosine phosphorylation sites, several of which contain redundant motifs. Using 32D cells, which contain no endogenous IRS proteins, we compared the signaling ability of an IRS-1 molecule in which 18 potential tyrosine phosphorylation sites were replaced by phenylalanine (IRS-1(F18)) with two derivative molecules which retained three YMXM motifs (IRS-1(3YMXM)) or the two COOH-terminal SHP2-Fyn binding sites (IRS-1(YCT)). During insulin stimulation, IRS-1(F18) failed to undergo tyrosine phosphorylation or mediate activation of the phosphotidylinositol (PI) 3'-kinase or p70(s6k); IRS-1(YCT) was tyrosine phosphorylated but also failed to mediate these signaling events. Neither IRS-1(3YMXM) nor IRS-1(YCT) mediated activation of mitogen-activated protein kinases. IRS-1(F18) and IRS-1(YCT) partially mediated similar levels of insulin-stimulated mitogenesis at high insulin concentrations, however, suggesting that IRS-1 contains phosphotyrosine-independent elements which effect mitogenic signals, and that the sites in IRS-l(YCT) do not augment this signal. IRS-1(3YMXM) mediated the maximal mitogenic response to insulin, although the response to insulin was more sensitive with wild-type IRS-1. By contrast, the association of IRS-1(3YMXM) with PI 3'-kinase was more sensitive to insulin than the association with IRS-1. Thus, the binding of SH2 proteins (such as PI 3'-kinase) by YMXM motifs in IRS-1 is an important element in the mitogenic response, but other elements are essential for full mitogenic sensitivity.

The Drosophila insulin receptor activates multiple signaling pathways but requires insulin receptor substrate proteins for DNA synthesis.

The Drosophila insulin receptor (DIR) contains a 368-amino-acid COOH-terminal extension that contains several tyrosine phosphorylation sites in YXXM motifs. This extension is absent from the human insulin receptor but resembles a region in insulin receptor substrate (IRS) proteins which binds to the phosphatidylinositol (PI) 3-kinase and mediates mitogenesis. The function of a chimeric DIR containing the human insulin receptor binding domain (hDIR) was investigated in 32D cells, which contain few insulin receptors and no IRS proteins. Insulin stimulated tyrosine autophosphorylation of the human insulin receptor and hDIR, and both receptors mediated tyrosine phosphorylation of Shc and activated mitogen-activated protein kinase. IRS-1 was required by the human insulin receptor to activate PI 3-kinase and p70s6k, whereas hDIR associated with PI 3-kinase and activated p70s6k without IRS-1. However, both receptors required IRS-1 to mediate insulin-stimulated mitogenesis. These data demonstrate that the DIR possesses additional signaling capabilities compared with its mammalian counterpart but still requires IRS-1 for the complete insulin response in mammalian cells.

Role of IRS-2 in insulin and cytokine signalling.

The protein IRS-1 acts as an interface between signalling proteins with Src-homology-2 domains (SH2 proteins) and the receptors for insulin, IGF-1, growth hormone, several interleukins (IL-4, IL-9, IL-13) and other cytokines. It regulates gene expression and stimulates mitogenesis, and appears to mediate insulin/IGF-1-stimulated glucose transport. Thus, survival of the IRS-1-/- mouse with only mild resistance to insulin was surprising. This dilemma is provisionally resolved with our discovery of a second IRS-signalling protein. We purified and cloned a likely candidate called 4PS from myeloid progenitor cells and, because of its resemblance to IRS-1, we designate it IRS-2. Alignment of the sequences of IRS-2 and IRS-1 revealed a highly conserved amino terminus containing a pleckstrin-homology domain and a phosphotyrosine-binding domain, and a poorly conserved carboxy terminus containing several tyrosine phosphorylation motifs. IRS-2 is expressed in many cells, including tissues from IRS-1-/- mice, and may be essential for signalling by several receptor systems.

Interferon-alpha engages the insulin receptor substrate-1 to associate with the phosphatidylinositol 3'-kinase.

Interferon-alpha (IFN alpha) induces rapid tyrosine phosphorylation of the insulin receptor substrate-1 (IRS-1), a docking protein with multiple tyrosine phosphorylation sites that bind to the Src homology 2 (SH2) domains of various signaling proteins. During IFN alpha stimulation, the p85 regulatory subunit of the phosphatidylinositol 3'-kinase binds via its SH2 domains to tyrosine-phosphorylated IRS-1, and phosphatidylinositol 3'-kinase activity is detected in association with IRS-1. Thus, IFN alpha responses occur by activation of the IRS signaling system, which it shares with insulin, insulin-like growth factor-1, and interleukin-4.

DNA-binding and trans-activation properties of Drosophila E2F and DP proteins.

The temporal activation of E2F transcriptional activity appears to be an important component of the mechanisms that prepare mammalian cells for DNA replication. Regulation of E2F activity appears to be a highly complex process, and the dissection of the E2F pathway will be greatly facilitated by the ability to use genetic approaches. We report the isolation of two Drosophila genes that can stimulate E2F-dependent transcription in Drosophila cells. One of these genes, dE2F, contains three domains that are highly conserved in the human homologs E2F-1, E2F-2, and E2F-3. Interestingly, one of these domains is highly homologous to the retinoblastoma protein (RB)-binding sequences of human E2F genes. The other gene, dDP, is closely related to the human DP-1 and DP-2 genes. We demonstrate that dDP and dE2F interact and cooperate to give sequence-specific DNA binding and optimal trans-activation. These features suggest that endogenous Drosophila E2F, like human E2F, may be composed of heterodimers and may be regulated by RB-like proteins. The isolation of these genes will provide important reagents for the genetic analysis of the E2F pathway.

Role of IRS-1-GRB-2 complexes in insulin signaling.

GRB-2 is a small SH2- and SH3 domain-containing adapter protein that associates with the mammalian SOS homolog to regulate p21ras during growth factor signaling. During insulin stimulation, GRB-2 binds to the phosphorylated Y895VNI motif of IRS-1. Substitution of Tyr-895 with phenylalanine (IRS-1F-895) prevented the IRS-1-GRB-2 association in vivo and in vitro. The myeloid progenitor cell line, 32-D, is insensitive to insulin because it contains few insulin receptors and no IRS-1. Coexpression of IRS-1 or IRS-1F-895 with the insulin receptor was required for insulin-stimulated mitogenesis in 32-D cells, while expression of the insulin receptor alone was sufficient to mediate insulin-stimulated tyrosine phosphorylation of Shc and activation of p21ras and mitogen-activated protein (MAP) kinase. The Shc-GRB-2 complex formed during insulin stimulation is a possible mediator of p21ras and MAP kinase activation in IRS-1-deficient 32-D cells. Interestingly, IRS-1, but not IRS-1F-895, enhanced the stimulation of MAP kinase by insulin in 32-D cells expressing insulin receptors. Thus, IRS-1 contributes to the stimulation of MAP kinase by insulin, probably through formation of the IRS-1-GRB-2 complex at Tyr-895. Our results suggest that the Shc-GRB-2 complex and the activation of p21ras-dependent signaling pathways, including MAP kinase, are insufficient for insulin-stimulated mitogenesis and that the essential function(s) of IRS-1 in proliferative signaling is largely unrelated to IRS-1-GRB-2 complex formation.

Functional domains of the insulin receptor responsible for chemotactic signaling.

The insulin receptor mediates a variety of cellular responses to insulin, including glucose transport, endocytosis, and cell proliferation. The role of the insulin receptor in mediating cellular motility has not, however, been extensively investigated. In this report, we demonstrate that chinese hamster ovary (CHO) cells that normally have low concentrations of insulin receptor display chemotaxis toward insulin after overexpression of the wild type human insulin receptor. Chemotaxis toward insulin proceeded through a pertussis toxin-sensitive pathway and required both tyrosine kinase activity and tyrosine autophosphorylation of the regulatory region of the beta-subunit. In contrast, the autophosphorylation sites in the carboxyl terminus of the receptor were not required for chemotactic activity. A mutation in the juxtamembrane region, which disabled tyrosine phosphorylation of the insulin receptor substrate-1 (IRS-1), also prevented the chemotactic response, suggesting a possible role for IRS-1 in chemotactic signaling. In the absence of insulin receptor, however, the presence of excess transfected IRS-1 was not sufficient to mediate chemotaxis toward insulin. These results demonstrate that the intact insulin receptor can stimulate a chemotactic signaling pathway and that this initial pathway more closely correlates with that for insulin-stimulated cell proliferation than for insulin-stimulated receptor endocytosis.