ABSTRACT
TRPM7, a TRP channel with ion conductance and kinase activities, has emerged as an attractive drug target for immunomodulation. Reverse genetics and cell biological studies have already established a key role for TRPM7 in the inflammatory activation of macrophages. Advancing TRPM7 as a viable molecular target for immunomodulation requires selective TRPM7 inhibitors with in vivo tolerability and efficacy. Such inhibitors have the potential to interdict inflammatory cascades mediated by systemic and tissue-specialized macrophages. FTY720, an FDA-approved drug for multiple sclerosis inhibits TRPM7. However, FTY720 is a prodrug and its metabolite, FTY720-phosphate, is a potent agonist of sphingosine-1-phosphate (S1P) receptors. In this study, we test non-phosphorylatable FTY720 analogs, which are inert against S1PRs and well tolerated in vivo, for activity against TRPM7 and tissue bioavailability. Using patch clamp electrophysiology, we show that VPC01091.4 and AAL-149 block TRPM7 current at low micromolar concentrations. In culture, they act directly on macrophages to blunt LPS-induced inflammatory cytokine expression, though this likely occurrs through multiple molecular targets. We found that VPC01091.4 has significant and rapid accumulation in the brain and lungs, along with direct anti-inflammatory action on alveolar macrophages and microglia. Finally, using a mouse model of endotoxemia, we show VPC01091.4 to be an efficacious anti-inflammatory agent that arrests systemic inflammation in vivo. Together, these findings identify novel small molecule inhibitors that allow TRPM7 channel inhibition independent of S1P receptor targeting which demonstrate potent, polymodal anti-inflammatory activities ex vivo and in vivo.
Subject(s)
Fingolimod Hydrochloride , TRPM Cation Channels , Fingolimod Hydrochloride/pharmacology , Cyclopentanes , PhosphorylationABSTRACT
Insulin secretion depends on the Ca2+-regulated fusion of granules with the plasma membrane. A recent model of Ca2+-triggered exocytosis in secretory cells proposes that lipids in the plasma membrane couple the calcium sensor Syt1 to the membrane fusion machinery (Kiessling et al., 2018). Specifically, Ca2+-mediated binding of Syt1's C2 domains to the cell membrane shifts the membrane-anchored SNARE syntaxin-1a to a more fusogenic conformation, straightening its juxtamembrane linker. To test this model in live cells and extend it to insulin secretion, we enriched INS1 cells with a panel of lipids with different acyl chain compositions. Fluorescence lifetime measurements demonstrate that cells with more disordered membranes show an increase in fusion efficiency, and vice versa. Experiments with granules purified from INS1 cells and recombinant SNARE proteins reconstituted in supported membranes confirmed that lipid acyl chain composition determines SNARE conformation and that lipid disordering correlates with increased fusion. Addition of Syt1's C2AB domains significantly decreased lipid order in target membranes and increased SNARE-mediated fusion probability. Strikingly, Syt's action on both fusion and lipid order could be partially bypassed by artificially increasing unsaturated phosphatidylserines in the target membrane. Thus, plasma membrane lipids actively participate in coupling Ca2+/synaptotagmin-sensing to the SNARE fusion machinery in cells.
Subject(s)
Insulin-Secreting Cells , Membrane Fusion , Membrane Lipids/metabolism , SNARE Proteins/metabolism , Insulin-Secreting Cells/metabolism , Cell Membrane/metabolism , Synaptotagmin I/chemistry , Synaptotagmin I/metabolism , Exocytosis , Recombinant Proteins/metabolism , Calcium/metabolismABSTRACT
TRPM7, a TRP channel with ion conductance and kinase activities, has emerged as an attractive drug target for immunomodulation. Reverse genetics and cell biological studies have already established a key role for TRPM7 in the inflammatory activation of macrophages. Advancing TRPM7 as a viable molecular target for immunomodulation requires selective TRPM7 inhibitors with in vivo tolerability and efficacy. Such inhibitors have the potential to interdict inflammatory cascades mediated by systemic and tissue-specialized macrophages. FTY720, an FDA-approved drug for multiple sclerosis inhibits TRPM7. However, FTY720 is a prodrug and its metabolite, FTY720-phosphate, is a potent agonist of sphingosine 1-phosphate (S1P) receptors. In this study, we tested non-phosphorylatable FTY720 analogs, which are inert against S1PRs and well tolerated in vivo , for activity against TRPM7 and tissue bioavailability. Using patch clamp electrophysiology, we show that VPC01091.4 and AAL-149 block TRPM7 current at low micromolar concentrations. In culture, they act directly on macrophages to blunt LPS-induced inflammatory cytokine expression, an effect that is predominantly but not solely mediated by TRPM7. We found that VPC01091.4 has significant and rapid accumulation in the brain and lungs, along with direct anti-inflammatory action on alveolar macrophages and microglia. Finally, using a mouse model of endotoxemia, we show VPC01091.4 to be an efficacious anti-inflammatory agent that arrests systemic inflammation in vivo . Together, these findings identify novel small molecule inhibitors that allow TRPM7 channel inhibition independent of S1P receptor targeting. These inhibitors exhibit potent anti-inflammatory properties that are mediated by TRPM7 and likely other molecular targets that remain to be identified.
ABSTRACT
Variants of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) challenge currently available coronavirus disease 2019 vaccines and monoclonal antibody therapies through epitope change on the receptor binding domain of the viral spike glycoprotein. Hence, there is a specific urgent need for alternative antivirals that target processes less likely to be affected by mutation, such as the membrane fusion step of viral entry into the host cell. One such antiviral class includes peptide inhibitors, which block formation of the so-called heptad repeat 1 and 2 (HR1HR2) six-helix bundle of the SARS-CoV-2 spike (S) protein and thus interfere with viral membrane fusion. We performed structural studies of the HR1HR2 bundle, revealing an extended, well-folded N-terminal region of HR2 that interacts with the HR1 triple helix. Based on this structure, we designed an extended HR2 peptide that achieves single-digit nanomolar inhibition of SARS-CoV-2 in cell-based and virus-based assays without the need for modifications such as lipidation or chemical stapling. The peptide also strongly inhibits all major SARS-CoV-2 variants to date. This extended peptide is â¼100-fold more potent than all previously published short, unmodified HR2 peptides, and it has a very long inhibition lifetime after washout in virus infection assays, suggesting that it targets a prehairpin intermediate of the SARS-CoV-2 S protein. Together, these results suggest that regions outside the HR2 helical region may offer new opportunities for potent peptide-derived therapeutics for SARS-CoV-2 and its variants, and even more distantly related viruses, and provide further support for the prehairpin intermediate of the S protein.
Subject(s)
COVID-19 Drug Treatment , Spike Glycoprotein, Coronavirus , Antiviral Agents/chemistry , Antiviral Agents/pharmacology , Humans , Peptides/chemistry , Peptides/pharmacology , SARS-CoV-2/drug effectsABSTRACT
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) cell entry starts with membrane attachment and ends with spike (S) protein-catalyzed membrane fusion depending on two cleavage steps, namely, one usually by furin in producing cells and the second by TMPRSS2 on target cells. Endosomal cathepsins can carry out both. Using real-time three-dimensional single-virion tracking, we show that fusion and genome penetration require virion exposure to an acidic milieu of pH 6.2 to 6.8, even when furin and TMPRSS2 cleavages have occurred. We detect the sequential steps of S1-fragment dissociation, fusion, and content release from the cell surface in TMPRRS2-overexpressing cells only when exposed to acidic pH. We define a key role of an acidic environment for successful infection, found in endosomal compartments and at the surface of TMPRSS2-expressing cells in the acidic milieu of the nasal cavity.
Subject(s)
COVID-19 , Nasal Cavity , SARS-CoV-2 , Serine Endopeptidases , Virus Internalization , COVID-19/virology , Furin/genetics , Furin/metabolism , Humans , Hydrogen-Ion Concentration , Nasal Cavity/chemistry , Nasal Cavity/virology , SARS-CoV-2/physiology , Serine Endopeptidases/metabolism , Spike Glycoprotein, Coronavirus/metabolismABSTRACT
Variants of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) challenge currently available COVID-19 vaccines and monoclonal antibody therapies through epitope change on the receptor binding domain of the viral spike glycoprotein. Hence, there is a specific urgent need for alternative antivirals that target processes less likely to be affected by mutation, such as the membrane fusion step of viral entry into the host cell. One such antiviral class includes peptide inhibitors which block formation of the so-called HR1HR2 six-helix bundle of the SARS-CoV-2 spike (S) protein and thus interfere with viral membrane fusion. Here we performed structural studies of the HR1HR2 bundle, revealing an extended, well-folded N-terminal region of HR2 that interacts with the HR1 triple helix. Based on this structure, we designed an extended HR2 peptide that achieves single-digit nanomolar inhibition of SARS-CoV-2 in cell-based fusion, VSV-SARS-CoV-2 chimera, and authentic SARS-CoV-2 infection assays without the need for modifications such as lipidation or chemical stapling. The peptide also strongly inhibits all major SARS-CoV-2 variants to date. This extended peptide is ~100-fold more potent than all previously published short, unmodified HR2 peptides, and it has a very long inhibition lifetime after washout in virus infection assays, suggesting that it targets a pre-hairpin intermediate of the SARS-CoV-2 S protein. Together, these results suggest that regions outside the HR2 helical region may offer new opportunities for potent peptide-derived therapeutics for SARS-CoV-2 and its variants, and even more distantly related viruses, and provide further support for the pre-hairpin intermediate of the S protein. Significance Statement: SARS-CoV-2 infection requires fusion of viral and host membranes, mediated by the viral spike glycoprotein (S). Due to the importance of viral membrane fusion, S has been a popular target for developing vaccines and therapeutics. We discovered a simple peptide that inhibits infection by all major variants of SARS-CoV-2 with nanomolar efficacies. In marked contrast, widely used shorter peptides that lack a key N-terminal extension are about 100 x less potent than this peptide. Our results suggest that a simple peptide with a suitable sequence can be a potent and cost-effective therapeutic against COVID-19 and they provide new insights at the virus entry mechanism.
ABSTRACT
Efficient clearance of apoptotic cells by phagocytosis, also known as efferocytosis, is fundamental to developmental biology, organ physiology, and immunology. Macrophages use multiple mechanisms to detect and engulf apoptotic cells, but the signaling pathways that regulate the digestion of the apoptotic cell cargo, such as the dynamic Ca2+ signals, are poorly understood. Using an siRNA screen, we identify TRPM7 as a Ca2+-conducting ion channel essential for phagosome maturation during efferocytosis. Trpm7-targeted macrophages fail to fully acidify or digest their phagosomal cargo in the absence of TRPM7. Through perforated patch electrophysiology, we demonstrate that TRPM7 mediates a pH-activated cationic current necessary to sustain phagosomal acidification. Using mice expressing a genetically-encoded Ca2+ sensor, we observe that phagosome maturation requires peri-phagosomal Ca2+-signals dependent on TRPM7. Overall, we reveal TRPM7 as a central regulator of phagosome maturation during macrophage efferocytosis.
Subject(s)
Calcium Signaling , Phagocytosis , TRPM Cation Channels , Animals , Macrophages/metabolism , Mice , Phagocytosis/physiology , Phagosomes/metabolism , TRPM Cation Channels/genetics , TRPM Cation Channels/metabolismABSTRACT
SARS-CoV-2 cell entry starts with membrane attachment and ends with spike-protein (S) catalyzed membrane fusion depending on two cleavage steps, one usually by furin in producing cells and the second by TMPRSS2 on target cells. Endosomal cathepsins can carry out both. Using real-time 3D single virion tracking, we show fusion and genome penetration requires virion exposure to an acidic milieu of pH 6.2-6.8, even when furin and TMPRSS2 cleavages have occurred. We detect the sequential steps of S1-fragment dissociation, fusion, and content release from the cell surface in TMPRRS2 overexpressing cells only when exposed to acidic pH. We define a key role of an acidic environment for successful infection, found in endosomal compartments and at the surface of TMPRSS2 expressing cells in the acidic milieu of the nasal cavity. Significance Statement: Infection by SARS-CoV-2 depends upon the S large spike protein decorating the virions and is responsible for receptor engagement and subsequent fusion of viral and cellular membranes allowing release of virion contents into the cell. Using new single particle imaging tools, to visualize and track the successive steps from virion attachment to fusion, combined with chemical and genetic perturbations of the cells, we provide the first direct evidence for the cellular uptake routes of productive infection in multiple cell types and their dependence on proteolysis of S by cell surface or endosomal proteases. We show that fusion and content release always require the acidic environment from endosomes, preceded by liberation of the S1 fragment which depends on ACE2 receptor engagement. One sentence summary: Detailed molecular snapshots of the productive infectious entry pathway of SARS-CoV-2 into cells.
ABSTRACT
Activation of Pannexin 1 (PANX1) ion channels causes release of intercellular signaling molecules in a variety of (patho)physiological contexts. PANX1 can be activated by G protein-coupled receptors (GPCRs), including α1-adrenergic receptors (α1-ARs), but how receptor engagement leads to channel opening remains unclear. Here, we show that GPCR-mediated PANX1 activation can occur via channel deacetylation. We find that α1-AR-mediated activation of PANX1 channels requires Gαq but is independent of phospholipase C or intracellular calcium. Instead, α1-AR-mediated PANX1 activation involves RhoA, mammalian diaphanous (mDia)-related formin, and a cytosolic lysine deacetylase activated by mDia - histone deacetylase 6. HDAC6 associates with PANX1 and activates PANX1 channels, even in excised membrane patches, suggesting direct deacetylation of PANX1. Substitution of basally-acetylated intracellular lysine residues identified on PANX1 by mass spectrometry either prevents HDAC6-mediated activation (K140/409Q) or renders the channels constitutively active (K140R). These data define a non-canonical RhoA-mDia-HDAC6 signaling pathway for GαqPCR activation of PANX1 channels and uncover lysine acetylation-deacetylation as an ion channel silencing-activation mechanism.
Subject(s)
Connexins/metabolism , Histone Deacetylase 6/metabolism , Nerve Tissue Proteins/metabolism , Receptors, Adrenergic, alpha-1/metabolism , Acetylation , Animals , Cells, Cultured , Connexins/genetics , Connexins/physiology , HEK293 Cells , Histone Deacetylase 6/genetics , Humans , Jurkat Cells , Lysine/genetics , Lysine/metabolism , Membrane Potentials/physiology , Mice, Inbred C57BL , Mice, Knockout , Nerve Tissue Proteins/genetics , Nerve Tissue Proteins/physiology , Patch-Clamp Techniques , Receptors, Adrenergic, alpha-1/genetics , Signal Transduction/genetics , rhoA GTP-Binding Protein/genetics , rhoA GTP-Binding Protein/metabolismABSTRACT
Insulin secretion from ß-cells is reduced at the onset of type-1 and during type-2 diabetes. Although inflammation and metabolic dysfunction of ß-cells elicit secretory defects associated with type-1 or type-2 diabetes, accompanying changes to insulin granules have not been established. To address this, we performed detailed functional analyses of insulin granules purified from cells subjected to model treatments that mimic type-1 and type-2 diabetic conditions and discovered striking shifts in calcium affinities and fusion characteristics. We show that this behavior is correlated with two subpopulations of insulin granules whose relative abundance is differentially shifted depending on diabetic model condition. The two types of granules have different release characteristics, distinct lipid and protein compositions, and package different secretory contents alongside insulin. This complexity of ß-cell secretory physiology establishes a direct link between granule subpopulation and type of diabetes and leads to a revised model of secretory changes in the diabetogenic process.
Diabetes is a disease that occurs when sugar levels in the blood can no longer be controlled by a hormone called insulin. People with type 1 diabetes lose the ability to produce insulin after their immune system attacks the ß-cells in their pancreas that make this hormone. People with type 2 diabetes develop the disease when ß-cells become exhausted from increased insulin demand and stop producing insulin. ß-cells store insulin in small compartments called granules. When blood sugar levels rise, these granules fuse with the cell membrane allowing ß-cells to release large quantities of insulin at once. This fusion is disrupted early in type 1 diabetes, but later in type 2: the underlying causes of these disruptions are unclear. In the laboratory, signals that trigger inflammation and molecules called fatty acids can mimic type 1 or type 2 diabetes respectively when applied to insulin-producing cells. Kreutzberger, Kiessling et al. wanted to know whether pro-inflammatory molecules and fatty acids affect insulin granules differently at the molecular level. To do this, insulin-producing cells were grown in the lab and treated with either fatty acids or pro-inflammatory molecules. The insulin granules of these cells were then isolated. Next, the composition of the granules and how they fused to lab-made membranes that mimic the cell membrane was examined. The experiments revealed that healthy ß-cells have two types of granules, each with a different version of a protein called synaptotagmin. Cells treated with molecules mimicking type 1 diabetes lost granules with synaptotagmin-7, while granules with synaptotagmin-9 were lost in cells treated with fatty acids to imitate type 2 diabetes. Each type of granule responded differently to calcium levels in the cell and secreted different molecules, indicating that each elicits a different diabetic response in the body. These findings suggest that understanding how insulin granules are formed and regulated may help find treatments for type 1 and 2 diabetes, possibly leading to therapies that reverse the loss of different types of granules. Additionally, the molecules of these granules may also be used as markers to determine the stage of diabetes. More broadly, these results show how understanding how molecule release changes with disease in different cell types may help diagnose or stage a disease.
Subject(s)
Calcium/metabolism , Diabetes Mellitus, Type 1/metabolism , Diabetes Mellitus, Type 2/metabolism , Exocytosis , Insulin-Secreting Cells/metabolism , Insulin/metabolism , Animals , Cholesterol/metabolism , Cytokines/pharmacology , Diabetes Mellitus, Type 1/genetics , Diabetes Mellitus, Type 2/genetics , Exocytosis/drug effects , Humans , Insulin/genetics , Insulin-Secreting Cells/drug effects , PC12 Cells , Palmitates/pharmacology , Rats , SNARE Proteins/metabolism , Secretory Pathway , Sphingomyelins/metabolism , Synaptotagmins/metabolismABSTRACT
Phagocytes reallocate metabolic resources to kill engulfed pathogens, but the intracellular signals that rapidly switch the immunometabolic program necessary to fuel microbial killing are not understood. We report that macrophages use a fast two-step Ca2+ relay to meet the bioenergetic demands of phagosomal killing. Upon detection of a fungal pathogen, macrophages rapidly elevate cytosolic Ca2+ (phase 1), and by concurrently activating the mitochondrial Ca2+ (mCa2+) uniporter (MCU), they trigger a rapid influx of Ca2+ into the mitochondria (phase 2). mCa2+ signaling reprograms mitochondrial metabolism, at least in part, through the activation of pyruvate dehydrogenase (PDH). Deprived of mCa2+ signaling, Mcu-/- macrophages are deficient in phagosomal reactive oxygen species (ROS) production and defective at killing fungi. Mice lacking MCU in their myeloid cells are highly susceptible to disseminated candidiasis. In essence, this study reveals an elegant design principle that MCU-dependent Ca2+ signaling is an electrometabolic switch to fuel phagosome killing.
Subject(s)
Calcium/metabolism , Candida albicans/pathogenicity , Mitochondria/metabolism , Phagosomes/metabolism , Animals , Mice , Signal TransductionABSTRACT
Highly selective, positive allosteric modulators (PAMs) of the M1 subtype of muscarinic acetylcholine receptor have emerged as an exciting new approach to potentially improve cognitive function in patients suffering from Alzheimer's disease and schizophrenia. Discovery programs have produced a structurally diverse range of M1 receptor PAMs with distinct pharmacological properties, including different extents of agonist activity and differences in signal bias. This includes biased M1 receptor PAMs that can potentiate coupling of the receptor to activation of phospholipase C (PLC) but not phospholipase D (PLD). However, little is known about the role of PLD in M1 receptor signaling in native systems, and it is not clear whether biased M1 PAMs display differences in modulating M1-mediated responses in native tissue. Using PLD inhibitors and PLD knockout mice, we showed that PLD was necessary for the induction of M1-dependent long-term depression (LTD) in the prefrontal cortex (PFC). Furthermore, biased M1 PAMs that did not couple to PLD not only failed to potentiate orthosteric agonist-induced LTD but also blocked M1-dependent LTD in the PFC. In contrast, biased and nonbiased M1 PAMs acted similarly in potentiating M1-dependent electrophysiological responses that were PLD independent. These findings demonstrate that PLD plays a critical role in the ability of M1 PAMs to modulate certain central nervous system (CNS) functions and that biased M1 PAMs function differently in brain regions implicated in cognition.
Subject(s)
Cerebral Cortex/enzymology , Neuronal Plasticity , Phospholipase D/genetics , Phospholipase D/metabolism , Receptor, Muscarinic M1/genetics , Receptor, Muscarinic M1/metabolism , Allosteric Site , Animals , CHO Cells , Calcium/chemistry , Cognition , Cricetinae , Cricetulus , Electrophysiology , Female , Humans , Long-Term Synaptic Depression , Male , Mice , Mice, Inbred C57BL , Mice, Knockout , Prefrontal Cortex/enzymology , Signal Transduction , Type C Phospholipases/metabolismABSTRACT
In pancreatic ß-cells, insulin granule membranes are enriched in cholesterol and are both recycled and newly generated. Cholesterol's role in supporting granule membrane formation and function is poorly understood. ATP binding cassette transporters ABCG1 and ABCA1 regulate intracellular cholesterol and are important for insulin secretion. RNAi inter-ference-induced depletion in cultured pancreatic ß-cells shows that ABCG1 is needed to stabilize newly made insulin granules against lysosomal degradation; ABCA1 is also involved but to a lesser extent. Both transporters are also required for optimum glucose-stimulated insulin secretion, likely via complementary roles. Exogenous cholesterol addition rescues knockdown-induced granule loss (ABCG1) and reduced secretion (both transporters). Another cholesterol transport protein, oxysterol binding protein (OSBP), appears to act proximally as a source of endogenous cholesterol for granule formation. Its knockdown caused similar defective stability of young granules and glucose-stimulated insulin secretion, neither of which were rescued with exogenous cholesterol. Dual knockdowns of OSBP and ABC transporters support their serial function in supplying and concentrating cholesterol for granule formation. OSBP knockdown also decreased proinsulin synthesis consistent with a proximal endoplasmic reticulum defect. Thus, membrane cholesterol distribution contributes to insulin homeostasis at production, packaging, and export levels through the actions of OSBP and ABCs G1 and A1.
Subject(s)
ATP Binding Cassette Transporter 1/metabolism , ATP Binding Cassette Transporter, Subfamily G, Member 1/metabolism , Cytoplasmic Granules/metabolism , Insulin/metabolism , Receptors, Steroid/metabolism , ATP Binding Cassette Transporter 1/deficiency , ATP Binding Cassette Transporter, Subfamily G, Member 1/deficiency , Animals , Cholesterol/pharmacology , Cytoplasmic Granules/drug effects , Endoplasmic Reticulum/metabolism , Exocytosis/drug effects , Fluorescence , Gene Knockdown Techniques , Glucose/pharmacology , Green Fluorescent Proteins/metabolism , Humans , Lysosomes/drug effects , Lysosomes/metabolism , Mice , Models, Biological , RNA Interference , Rats , Secretory Pathway/drug effects , beta-Cyclodextrins/pharmacologyABSTRACT
Highly selective positive allosteric modulators (PAMs) of the M1 subtype of muscarinic acetylcholine receptor have emerged as an exciting new approach for improving cognitive function in patients suffering from Alzheimer's disease and schizophrenia. However, excessive activation of M1 is known to induce seizure activity and have actions in the prefrontal cortex (PFC) that could impair cognitive function. We now report a series of pharmacological, electrophysiological, and behavioral studies in which we find that recently reported M1 PAMs, PF-06764427 and MK-7622, have robust agonist activity in cell lines and agonist effects in the mouse PFC, and have the potential to overactivate the M1 receptor and disrupt PFC function. In contrast, structurally distinct M1 PAMs (VU0453595 and VU0550164) are devoid of agonist activity in cell lines and maintain activity dependence of M1 activation in the PFC. Consistent with the previously reported effect of PF-06764427, the ago-PAM MK-7622 induces severe behavioral convulsions in mice. In contrast, VU0453595 does not induce behavioral convulsions at doses well above those required for maximal efficacy in enhancing cognitive function. Furthermore, in contrast to the robust efficacy of VU0453595, the ago-PAM MK-7622 failed to improve novel object recognition, a rodent assay of cognitive function. These findings suggest that in vivo cognition-enhancing efficacy of M1 PAMs can be observed with PAMs lacking intrinsic agonist activity and that intrinsic agonist activity of M1 PAMs may contribute to adverse effects and reduced efficacy in improving cognitive function.