Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Aug 1.
Published in final edited form as: Adv Biol Regul. 2019 Jul 30;73:100637. doi: 10.1016/j.jbior.2019.100637

A synthetic biological approach to reconstitution of inositide signaling pathways in bacteria

Bradley P Clarke 1, Brandon L Logeman 2, Andrew T Hale 1, Zigmund Luka 1, John D York 1,2,*
PMCID: PMC7216163  NIHMSID: NIHMS1536693  PMID: 31378699

Abstract

Inositide lipid (PIP) and soluble (IP) signaling pathways produce essential cellular codes conserved in eukaryotes. In many cases, deconvoluting metabolic and functional aspects of individual pathways are confounded by promiscuity and multiplicity of PIP and IP kinases and phosphatases. We report a molecular genetic approach that reconstitutes eukaryotic inositide lipid and soluble pathways in a prokaryotic cell which inherently lack inositide kinases and phosphatases in their genome. By expressing synthetic cassettes of eukaryotic genes, we have reconstructed the heterologous formation of a range of inositide lipids, including PI(3)P, PI(4,5)P2 and PIP3. In addition, we report the reconstruction of lipid-dependent production of inositol hexakisphosphate (IP6). Our synthetic system is scalable, reduces confounding metabolic issues, for example it is devoid of inositide phosphatases and orthologous kinases, and enables accurate characterization gene product enzymatic activity and substrate selectivity. This genetically engineered tool is designed to help interpret metabolic pathways and may facilitate in vivo testing of regulators and small molecule inhibitors. In summary, heterologous expression of inositide pathways in bacteria provide a malleable experimental platform for aiding signaling biologists and offers new insights into metabolism of these essential pathways.

Keywords: Phosphatidylinositol, Inositol Phosphate, Cell Signaling, Synthetic Biology

Introduction

Myo-Inositol, a small cyclohexanol, is the precursor for a family of inositide molecules that control many diverse cellular functions (Berridge et al., 1983; Majerus et al., 1988; Raucher et al., 2000; Otto et al., 2007). The first major class of inositol derivatives are the lipid phosphoinositide phosphates (PIPs), composed of phosphatidylinositol (PI) and its 7 phosphorylated derivatives. PIPs are created and destroyed by families of kinases and phosphatases that play important roles in various processes required for eukaryotic life (Majerus and York, 2009; Balla, 2013). PIP signaling is initiated through production of PI, formed by Phosphatidylinositol synthase (Pis1) from cytidine diphosphate diacylglycerol (CDP-DAG) and myo-inositol. PI can then be further phosphorylated to form 3 PIPs: PI(3)P, PI(4)P and PI(5)P. These PIPs can then form the PI(4,5)P2, the most abundant PIP2, PI(3,4)P2, and PI(3,5)P2. These can then further be modified, including phosphorylation of PI(4,5)P2 by p110 (class-I PI3K in humans) to form PI(3,4,5)P3, a key signal for cell growth and survival. Numerous studies of the enzyme regulation of inositide lipids has led to field changing discoveries and highlights there roles in uni-and multi-cellular eukaryotes (Fruman, Meyers and Cantley, 1998; Cantley, 2002; Irvine, 2016).

In addition to lipid signaling, PI(4,5)P2 is cleaved by phospholipase C (Plc) to form diacylglycerol and IP3, which act as second messengers for protein kinase C activation and calcium release (Streb et al., 1983; Ando et al., 2018). Subsequently, IP3 may be further phosphorylated to produce IP4, IP5, and IP6, as well as pyro-phosphorylated to produce IP7 and IP8 among others. These soluble inositol derivatives are conserved throughout eukaryotic organisms and are required for ion channel regulation, phosphate sensing, transcription, mRNA export, embryonic development, and act as structural cofactors (Hatch and York, 2010). A wide range of cellular and organismal roles have been defined for the collection of dozens of water soluble inositide messengers (Alcázar-Román and Wente, 2008; Monserrate and York, 2010; Wilson, Livermore and Saiardi, 2013). However, the intricate nature by which soluble inositol phosphate kinases and phosphatases construct the “IP code” and the transient nature by which IP species are made and destroyed have made characterization of IP regulatory enzymes difficult.

In many eukaryotic cells, the complexity of inositide metabolism as well as redundancies in both kinase and phosphatase gene product families has clouded interpretation of enzyme specificity, regulation and inositide product function. Some of the kinases involved have both kinase dependent and kinase independent functions (Hatch, Odom and York, 2017; Shears and Wang, 2019), or have both kinase and phosphatases present within the same protein (Mulugu et al., 2007; Shears et al., 2017), which can also make interpretation more difficult. Additionally, in vitro analyses, especially for lipid metabolizing enzymes, are challenging as they require recapitulation of the complex membrane, intermembrane and cofactor properties. As a means to address some of these issues, we initiated studies in bacteria because they lack endogenous or orthologous inositide signaling gene products. Our goal was to recapitulate simplified versions of both inositide lipid and soluble metabolic pathways. A previous study of heterologous expression of yeast phosphatidylinositol (PI) synthase in bacteria suggested the production of PI in prokaryotes was possible (Nikawa, Kodaki and Yamashita, 1988); however, expression of a more complete array of the full inositide signaling pathway has yet to be reported. Here, we reconstruct many components of inositide metabolism in a controlled, cell based system through a synthetic biology approach of introducing eukaryotic inositide lipase and kinase gene products into E. coli. Our system provides a greatly simplified platform that enables detailed and scalable cell-based characterization of inositide kinase, phosphatase, lipase, regulatory and effector proteins. Studies reported here may help provide new insights and clarity into enzyme specificity and function.

Materials and methods

Cloning and Bacterial Labeling methods

Standard cloning methods were used to introduce genes into the Duet vector system as previously described (Tolia and Joshua-Tor, 2006). Plasmids used in this study are listed in Table 1. Briefly, plasmids in a variety of combinations were transformed into chemically competent BL21 DE3 cells. These bacterial strains were grown and labeled as follows. First, cultures were grown in M9 minimal salts with 1 mM Thiamine, 0.4% glycerol, 0.2% casamino acids, 2 mM MgSO4 0.1 mM CaCl2 plus appropriate antibiotic selection at 37°C unti l they reached OD600 0.4. Then, 100 μCi of 3H-inositol and 1mM Isopropyl b-D-1-thiogalactopyranoside (IPTG) were added, and cells were grown at 30°C for 36 hours, with the addition of fresh antibiotics every 12 hours. Cells were then harvested by centrifugation and the cell pellets were washed with M9 minimal salts and stored at −80°C until use.

Table 1.

List of plasmids used in this study

Plasmid Insert Source
pET-duet scPisI Saccharomyces cerevisiae Phosphoinositide Synthase (sc Pis1) This study
pET-duet scPisI btPikl (sc Pis1), and Bos tarus Phosphoinositide 4 kinase beta (bt Pik1) This study
pET-duet scPisI scVPS34HE LCAT (sc Pis1), and Saccharomyces cerevisiae Vacuolar protein sorting HELical and CATalytic subunit (sc Vps34 HELCAT) This study
pACYC-duet scMss4 Saccharomyces cerevisiae Multicopy suppressor of Stt4 mutation (sc Mss4, a PI4P 5 kinase) This study
pACYC-duet scMss4 mmPLCd1 (sc Mss4), and Mus musculus Phospholipase C delta 1 (mm Plcδ1) This study
pACYC-duet scMss4 hp110 (sc Mss4), and Homo sapiens Phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit alpha (hs p110) This study
pCOLA-duet atIpk2 Arabidopsis thaliana Inositol phosphate kinase 2 (atIpk2) This study
pCOLA-duet atIpk2 scIpk1 (at Ipk2), and Saccharomyces cerevisiae inositol phosphate kinase (sc Ipk1) This study
cup1-PLC1 yeast Expression vector with scPlc1, copper inducible (Stevenson-Paulik et al., 2006)
pET24a PIP2 Operon Operon expression system for scPis1 btPik1 and scMss4 This study
Open in a new tab

Design and construction of an operon containing PIP Kinases

In addition to utilizing the Duet vector system we also designed a synthetic operon to express the genes responsible for synthesis of PI(4,5)P2, Saccharomyces cerevisiae Pis1, Bos taurus Pik1 and Saccharomyces cerevisiae Mss4 (Figure 4A). The operon consists of sequences for each of the three genes, optimized for expression in E. coli, each preceded by a restriction enzyme site, a spacer sequence 5’-AATAATTTTGTTTAAC-3’, the ribosome binding site 5’-AGGAGGTATATATA-3’, and followed by another restriction enzyme binding site. After the final gene, Mss4, a multicloning consisting of the NheI, PstI, SacI and XhoI restriction enzyme sites was introduced to allow for future expansion of the synthetic operon. The entire sequence was synthesized by Biomatik and cloned into the standard bacterial expression vector pET24a between the XbaI and XhoI restriction enzyme sites. The entire operon is under the control of the T7promoter and Lac operator and is followed by the T7 terminator.

Figure 4. A Synthetic Operon to Express the Genes required for Synthesis of PI(4,5)P2.

Figure 4.

Open in a new tab

Design of the operon (A). Codon optimized sequence for scPis1, btPik1, and scMss4 were each flanked by restriction enzyme sites and preceded by the spacer sequence 5 AATAATTTTGTTTAAC-3’ and the RBS 5’-AGGAGGTATATATA-3’ and the entire construct was inserted between the XhoI and XbaI sites in pET24a. All three genes in the operon are under the control of the T7/Lac promoter and the T7 terminator. Oxalate TLC separation (B) of extracts from BL21 Rosetta cells expressing the PIP operon demonstrates the production of PI, PI(4)P, and PI(4,5)P2. HPLC separation of the aqueous phase from extracts prepared from bacteria expressing the PIP operon without (C) or with Chloramphenicol (D). In bacteria expressing the PIP operon without Chloramphenicol, there are several peaks present in the aqueous phase (C). These may be due to the activity of bacterial phospholipases on PI(4,5)P2 that are in inhibited by Chloramphenicol since these peaks are absence with the PIP operon is expressed with Chloramphenicol.

Labeling of Yeast and Mammalian PIP and IP standards

W303 S. cerevisiae were grown in Complete Synthetic Media (CSM) +50 μCi of 3H-inositol starting with 10 μl of overnight culture per ml of media. Cells were grown overnight at 30°C, harvested by centrifugation, washed in PBS, the stored at −80°C until use. For hyperosmotic shock, yeast were grown as above, but before harvesting were subjected to osmotic shock as previously described (Bonangelino et al., 2002). For the Mammalian standards, Mouse embryonic fibroblasts were grown as previously described (Frederick et al., 2005). Briefly, cells were grown in inositol-free DMEM supplemented with 10% FBS prior to labeling with 50 μCi of 3H-inositol for 3 days. Cells were washed with PBS, then harvested in 100 μl .5M HCl, and Inositides were immediately extracted.

Extraction and Deacylation of PIPs

To extract Inositol Phosphates (IPs) and Phosphatidylinositol Phosphates (PIPs) cell pellets were resuspended in 100μl 0.5 M HCl, then 372 μl CHCl3:MeOH 1:2, and 100 μl glass beads were added and cells were lysed with bead-beating. Then, 125μl of each CHCl3 and 2M KCl was added and subjected to additional bead-beating. After centrifugation at 14,000 RPM, the upper aqueous phase contained soluble IPs, and the lower organic phase contained the lipid PIPs. IPs in the upper phase were immediately analyzed with High Performance Liquid Chromatography (HPLC), while the lower organic phase containing PIPs were analyzed by thin—layer chromatography (TLC) or dried down with 10μl 100mg/ml bovine brain extract in CHCl3:MeOH 50:1 carrier lipids, followed by deacylation prior to HPLC analysis separating gro-PIPs. Deacylation was conducted by adding 300μl 33% methylamine in ethanol: water:1-butanol 10:3:1 and incubated with rigorous agitation at 53°C for 1hr, followed by the addition of 150 μl of ice-cold 2-propanol and dried with a SpeedVac at 80°C. The dried pellet was resuspended in 200 μl water and extracted 3 times with 300 μl of 1-butanol:petroleum-ether:ethyl-formate (20:4:1), keeping the lower organic phase containing deacylated PIPs for HPLC analysis.

High Performance Liquid Chromatography

Both IPs and deacylated PIPs were loaded on to a Partisphere Sax 5 4.6 × 125mm column and eluted at 1 ml/min with ammonium phosphate (pH 3.5) gradients from 10 mM (buffer A) to 1.7 M (buffer B). For IPs separation was conducted as described in (Otto and York, 2010), and for deacylated PIPs the following gradient was used: 0% B for 5 minutes, then to 7% B over 6 minutes, followed by isocratic elution at 7% for 5 minutes, then a linear increase to 14% B over 15 minutes, to 60% B over 10 minutes, and finally to isocratic elution for 5 minutes. For IPs, 1ml fractions were collected and mixed with 6 ml each UlitimaFlowAP scintillation fluid and radioactivity was measured with a Perkin Elmer liquid scintillation analyzer. For deacylated PIPs, radioactivity was measured with an in-line βRam detector, IN-US, with MonoFlow4 scintillation fluid from National Diagnostics at a 3:1 mixing ratio and using a 1 ml flow cell.

Thin Layer Chromatography

Borate thin layer Chromatography analysis of PIPs was conducted as described (Walsh, Caldwell and Majerus, 1991). Briefly, Silica gel 60 HPTLC plates were prepared with immersion in a CDTA solution for 10 min followed by baking at 100°C for 20 min. Dried PIP samples were resuspended in CHCl3:MeOH 2:1, spotted on the plates, then developed with methanol (75 mL), chloroform (60 mL), pyridine (45 mL), boric acid (12 g), water (7.5 mL), 88% formic acid (3 mL), BHT (0.375 g), and ethoxyquin (75 μl). Radioactivity was detected using a Typhoon 9000 phosphor-imager. Representative spots for select PIPs were scraped from the silica gel plate were scraped off with a razor blade and were deacylated as described above. The deacylated PIPs were analyzed with HPLC as described above with fraction collection and scintillation counting. Oxalate thin layer chromatography was conducted as described previously (Stolz et al., 1998). Briefly, Silica gel 60 HPTLC plates were immersed in 54 mM potassium oxalate, 2mM EDTA in 47.5% ethanol, followed by drying for 1hr at 100°C. Dried PIP samples were resuspended in CHCl3:MeOH 2:1, spotted on the plates, then developed with CHCl3/Acetone/MeOH/Acetic acid/water (80:30:26:24:14) until the solvent reached the top of the plate, and radioactivity was detected as described above.

Results

Reconstitution of Lipid Phosphoinositide Phosphates in E. coli

The initial goals of our experiments were to use bacteria, which lack inositide signaling gene products, to heterologously reconstitute abridged components of yeast and metazoan inositide lipid and soluble metabolic pathways (Figure 1). First, we introduced phosphoinositide synthase 1 (Pis1), which catalyzes the formation of phosphatidylinositol (PI) from CDP-DAG and myo-inositol. In bacterial cells labeled with 3H-inositol, we observed the formation of PI, identified by comigration with standards on oxalate thin-layer chromatography (TLC) (Figure 2A lane 1). Next, as a means to generate phosphatidylinositol 4-phosphate (PI(4)P), we expressed a PI 4-kinase, Pik1, and observed an additional lipid species consistent with PI(4)P (Figure 2A lane 2). Similarly, to initiate production of PI(3)P, we co-expressed Pis1 and Vps34, a PI-specific 3-kinase. Analysis of radiolabeled lipid products revealed enriched and robust accumulation of PIP and relative depletion of PI as compared to Pis1 alone (Figure 2A lane 3 vs. lane 1). Of note, phosphorylation of PI by Vps34 was in the absence of Vps15, which is required for activation of Vps34 activity in yeast (Stack et al., 1993). The robust activity of Vps34 as compared to Pik1, and its ability to phosphorylate PI in the absence of Vps15, are consistent with published reports of the increased function of a truncated form of VPs34 used, which contains only the HELical and CATalytic domains known as HELCAT (Miller et al., 2010).

Figure 1. Synthetic approach to lipid derived inositide pathways in E. coli.

Figure 1.

Open in a new tab

The synthetic pathway used in this study to produce inositides in E. coli. Abbreviations: Myo-Inositol (Ins); Phosphatidyl Inositol (PI); Phosphatidyl Inositol 3-Phosphate (PI3P); Phosphatidyl Inositol 4-Phosphate (PI4P); Phosphatidyl Inositol 4, 5 Bisphosphate (PI(4,5)P2); Phosphatidyl 3, 4, 5 Triphosphate (PIP3), Inositol 1, 4, 5 Triphosphate (IP3); Inositol 1, 3, 4, 5, 6 Pentakisphosphate (IP5); Inositol Hexakisphosphate (IP6); Saccharomyces cerevisiae Phosphoinositide Synthase (sc Pis1); Bos tarus Phosphoinositide 4 kinase beta (bt Pik1); Saccharomyces cerevisiae Multicopy suppressor of Stt4 mutation (sc Mss4, a PI4P 5 kinase); Mus musculus Phospholipase C delta 1 (mm Plcδ1); Arabidopsis thaliana Inositol phosphate kinase 2 (at Ipk2); Saccharomyces cerevisiae inositol phosphate kinase (sc Ipk1); Saccharomyces cerevisiae Vacuolar protein sorting HELical and CATalytic subunit (sc Vps34 HELCAT); Homo sapiens Phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit α (hs p110).

Figure 2. Thin layer chromatography analysis of reconstituted lipid inositide synthesis.

Figure 2.

Open in a new tab

Thin layer chromatography (TLC) analysis of 3H-Inositol labeled lipids. (A) Oxalate TLC showing Pis1 bacteria produce PI and Lyso-PI (dashed circle), Pis1-Pik1 bacteria produce PI and PIP, Pis1-Vps34 bacteria produce PI and PIP, and Pis1-Mss4 bacteria produce PI and Lyso-PI (dashed circle). E. coli expressing PI synthase Pis1 can produce PI, but also have some Lyso-PI, presumably from the activity of bacterial phospholipase A. (B) Borate TLC resolving PI(4)P and PI(3)P production by Pis1-Pik1 and Pis1-Vps34 bacteria, respectively, as well as resolution of lyso-PI from PIPs; (C) Oxalate TLC demonstrates that Pis1-Vps34-Mss4 bacteria produce PI, PIP, PIP2 and PIP3; Pis1-Pik1-Mss4 bacteria produce PI, PIP, PIP2, and trace amounts of PIP3; Pis1-Pik1-Mss4-p110 (PIK3CA)-expressing bacteria produce PI, PIP, PIP2 and PIP3.

As a presumptive negative control for phosphorylation of PI, we expressed Mss4, a PI(4)P 5-Kinase that is not reported to utilize PI as a substrate, along with Pis1 and did not observe significant changes in lipid profiles as compared to Pis1 alone (Figure 2A, lane 4). The lipid species observed were further characterized scraping the region of silica and deacylation to their corresponding glycerol phosphoinositols. Separation of these products by high performance chromatography (HPLC) confirmed the identity of PI and revealed that the minor lipid, observed in lanes 1 and 4 that co-migrates with PIP standard, is lyso-phosphatidylinositol as the deacylated product was determined to be glycerol 1-phosphoinositol. We hypothesize, but do not prove, that the production of lyso-PI from the PI product of Pis1 may be a result of endogenous bacterial phospholipase A activities.

To further characterize and confirm the identities of the lipids produced in our bacterial system, we utilized two additional approaches: 1) a borate TLC system (Walsh, Caldwell and Majerus, 1991) capable of differentiating phosphorylation at the D-3 and D-4 ring positions; and 2) deacylation of the glycerol-based bacterial lipids using methylamine, which enables high resolution HPLC separation of the stereomers of the water soluble glycero-phosphoinositol products. Examination of radiolabeled lipid products using borate TLC indicates clear resolution of the Pis1 and Pik1 produced PI(4)P (Figure 2B, lane 2) and Pis1 Vps34 produced PI(3)P (Figure 2B, lane 3). Additionally, separation using the borate system was an improvement as the lyso-PI observed in Pis1 and Pis1 Mss4 samples, no longer co-migrated with PIP thereby simplifying quantification of lipid products (Figure 2B, lanes 1 and 4).

Equally important in the identification of lipid products is the use of deacylation and separation of the resulting glycerol-inositol phosphates utilizing HPLC. Glycerol-inositol phosphate standards were prepared from osmotically shocked 3H-inositol radiolabeled yeast (Figure 3A) or radiolabeled mouse embryonic fibroblasts (MEFs) (Figure 3B). Total lipid deacylation of bacterial extracts confirmed that the Pis1 bacteria produced a single species coeluting with glycerol-inositol 1-phosphate (Figure 3C). Addition of either Pik1 or Vps34 along with Pis1 resulted in the appearance of glycerol-inositol 1-phosphate and an additional species co-eluting with glycerol-inositol 1,4-phosphate or glycerol-inositol 1,3-phosphate, respectively (Figures 3D and 3E). This method also confirmed that Mss4 is unable to phosphorylate PI as a single glycerol-inositol 1-phosphate species was observed in Pis1 Mss4 bacteria (Figure 3F).

Figure 3. HPLC analysis of Reconstituted Lipid Phosphoinositide phosphate synthesis.

Figure 3.

Open in a new tab

Representative HPLC traces from analysis of 3H-inositol labeled glycerol-inositol phosphates from deacylated PIPs. (A) Osmotically shocked wild-type yeast showing separation of gro-PI(3)P and gro-PI(4)P, and of gro-PI(3,5)P2 and gro-PI(4,5)P2. (B) Mouse embryonic fibroblasts (MEFs) showing relative elution positions of gro-PI(4)P, gro-PI(4,5)P2, and gro-PI(3,4,5)P3. (C) Pis1-expressing bacteria produce of gro-PI (D) Pis1-Pik1 bacteria produce of gro-PI and gro-PI(4)P. (E) Pis1-Vps34 bacteria have depletion of gro-PI and robust production of gro-PI(3)P. (F) Pis1-Mss4 bacteria showing inability of Mss4 to phosphorylate PI. (G) Pis1-Vps34-Mss4 bacteria convert gro-PI to gro-PIP3 with gro-PI(3)P and gro-PI(3,X)P2 intermediates. Despite its traditional role as a PI(4)P 5-kinase, Mss4 has activity against PI(3)P, and the resulting PI(3,X)P2 is converted very efficiently to PIP3, either by Mss4, or possibly by Vps34. (H) Pis1-Pik1-Mss4 bacteria produce PI(4,5)P2 thus demonstrating the canonical role of Mss4, but surprisingly also produce PI(3,4,5)P3 without expression of a separate PI(4,5)P2 3-Kinase. (I) Pis1-Pik1-Mss4-p110 (PIK3CA) bacteria produce PI(4,5)P2 and PI(3,4,5)P3.

While Mss4 is classically considered to be a PI(4)P specific 5-Kinase, moonlighting activities have also been reported, for example as a PI(3)P 4-kinase (Desrivières et al., 1998) and a PI(3,4)P2 5-kinase (Zhang et al., 1997). As a presumed negative control, we expressed Mss4 along with Vps34 and Pis1 and analyzed the radiolabeled cell products by TLC (Figure 2C) and by deacylation/HPLC (Figure 3E). Unexpectedly, we observed the appearance of two new species and concomitant reduction of PI(3)P (Figure 2C, lane 1). Based on standards, it appears PI(3)P is first converted to a PI(3,X)P2 species, likely PI(3,4)P2 based on previous reports (Zhang et al., 1997; Desrivières et al., 1998; Mitra et al., 2004). Secondly, we observed the appearance of a substantial amount of a species co-migrating with PIP3 (Figure 2C lane 1). These data were supported by analysis of the bulk deacylated Bligh-Dyer extracted radiolabeled lipids in which we observed the appearance of four (4) species, co-eluting with glycerol-inositol 1-phosphate, glycerol-inositol 1,3-bisphosphate, glycerol-inositol 1,3,x-trisphosphate and glycerol-inositol 1,3,4,5-tetrakisphosphate, respectively (Figure 3G).

We next co-expressed Mss4 along with both Pis1 and Pik1 and observed the expected depletion of PI4P and appearance of a species comigrating with PI(4,5)P2, consistent with its canonical role as a PI4P 5-Kinase (Figure 2C lane 2). In addition, we report that Mss4 expressed in a bacterial system enables the unexpected formation of PIP3, presumably a relatively low level of PIP2 3-kinase moonlighting activity, which has not previously been an observed for Mss4 (Figure 2C lane 2). Additional analysis of these bacterial lipids by deacylation confirmed these activities by the observation of four (4) species, co-eluting with glycerol-inositol 1-phosphate, glycerol-inositol 1,4-bisphosphate, glycerol-inositol 1,4,5-trisphosphate and glycerol-inositol 1,3,4,5-tetrakisphosphate, respectively (Figure 3H). As a control for PIP3 production, we also co-expressed a class III PIP2 3-kinase catalytic subunit, p110, and observed additional conversion of PIP2 to PIP3 by TLC (Figure 2C lane 3) and deacylated products (Figure 3I). Overall, our data demonstrate in cells that Mss4 is capable of acting as a promiscuous PIP multikinase in both a PI(3)P-and PI(4)P-dependent manner to produce PIP2 and PIP3 species.

In addition to expressing pairs of genes using the Duet Vector system, we also designed a synthetic operon for the expression of the genes required for the production of PI(4,5)P2, Pis1, Pik1, and Mss4 (Figure 4A). We confirmed utilizing TLC that bacteria expressing these three genes in a synthetic operon were able to produce PI, PI4P and PI(4,5)P2 (Figure 4B). We also noted that relative to expression with the Duet system, bacteria expressing the synthetic operon had more radioactivity in the aqueous phase. When this aqueous phase was resolved with HPLC we observed several peaks, despite not expressing the PI(4,5)P2 lipase PLC (Figure 4C). One difference between the operon and the Duet system was the plasmid backbone used to express the genes, importantly the antibiotic selection markers. The pET24 operon utilized Kanamycin for selection, and the Duet vectors used to express this same combination of genes, pET duet and pACYC duet, utilize Ampicillin and Chloramphenicol respectively. Importantly, Chloramphenicol has been previously identified as an inhibitor of bacterial esterases, including lipases (SMITH, WORREL and SWANSON, 1949), so we hypothesized that the peaks observed in the aqueous phase of the bacteria expressing the PI(4,5)2 operon may be the result of activity of bacterial lipases acting on PI(4,5)P2. In order to test this, we transformed the bacterial PI(4,5)P2 operon into BL21 Rosetta cells. These commercially available cells contain a plasmid encoding tRNAs that are rare in E. coli, and that importantly also encodes a Chloramphenicol resistance gene. When the bacterial operon was expressed in this BL21 Rosetta background with both Kanamycin and Chloramphenicol, the anomalous peaks in the aqueous phase were greatly reduced (Figure 4D). We found that utilization of a synthetic operon to express the genes responsible for PI(4,5)P2 production is a useful alternative to utilization of the Duet vector system, however co-expression with a Chloramphenicol selectable plasmid may be preferred to inhibit any activity of endogenous bacterial lipases that may be acting on inositides.

Reconstitution of Soluble Inositol Phosphates in E. coli

As a means to study the lipid-derived water-soluble inositol phosphate pathways, we used our heterologous bacterial system to examine to the activities of phosphoinositide selective phospholipase C and inositol phosphate kinase (IPK) enzymes. To this end, we utilized Plc1, a δ-class enzyme capable of cleaving PI(4,5)P2 to produce DAG and inositol 1,4,5-trisphosphate, I(1,4,5)P3. HPLC separation of 3H-inositol labeled water-soluble extracts from bacterial cells were compared to IPs produced in S. Cerevisiae overexpressing Plc (Figure 5A). In order to characterize the ability of Plc to cleave lipids in the bacterial system, we expressed Plc along with bacterial strains expressing the set of genes required to produce each PIP species. When Pis1 was co expressed with Plc, we observed no peaks corresponding with the IP standards (Figure 5B). However, when Plc was co expressed with Pis1 and Pik1 we observed a peak that co elutes with I(1,4)P2, indicating that Plc is cleaving PI(4)P (Figure 5C).

Figure 5. Reconstitution of Soluble Inositol Phosphate Synthesis.

Figure 5.

Open in a new tab

Representative traces from HPLC analysis of 3H inositol labeled soluble IPs. (A) S. cerevisiae overexpressing Plc1 produce I(1,4)P2, I(1,4,5)P3, I(1,4,5,6)P4, I(1,3,4,5,6)P5 and IP6. (B) Pis1-Plc bacteria do not produce IPs due to a lack of substrate lipids for Plc. (C) Pis1-Pik1-Plc bacteria produce I(4,5)P2, consistent with production of PI(4)P and relative lack of endogenous I(1,4)P2 1-phosphatase activity by CysQ. (D) Pis1-Pik1-Mss4 bacteria do not produce IPs. In the absence of Plc, inositide lipids and are not converted to soluble IPs. (E) Pis1-Pik1-Mss4-Plc bacteria produce I(1,4)P2 and I(1,4,5)P3. When Plc is added to bacteria producing PI(4)P and PI(4,5)P2, these lipids are cleaved to produce their corresponding soluble IPs. (F) Pis1-Pik1 Mss4-Plc-Ipk2 bacteria produce I(1,3,4,5,6)P5. When the IP multikinase Ipk2 is introduced into bacteria, it converts IP3 to IP5 consistent with its documented role across eukaryotes. (G) Pis1-Pik1-Mss4-Plc-Ipk2-Ipk1 bacteria produce I(1,3,4,5,6)P5(IP5A), IP6, as well as IP5B. Ipk1 phosphorylates IP5A to make IP6, however the presence of IP5B may be due to the activity of bacterial phytases. (H) Pis1-Pik1-Mss4-Ipk2-Ipk1 bacteria do not produce IPs, demonstrating that the soluble IP production is dependent on the action of Plc on select lipid inositide species.

In the absence of Plc, bacteria expressing Pis1-Pik1-Mss4 produce PI(4,5)P2 (Figure 3H), but do not produce any soluble IPs (Figure 5D). When Plc is expressed along with Pis1-Pik1-Mss4 species that comigrate with I(1,4,5)P3 and I(1,4)P2 are produced, indicating that the PI(4)P and PI(4,5)P2 are being cleaved by Plc (Figure 5E). The IP3 peak present in these bacteria is wider than the IP3 peak present in S. cerevisiae produced standards, and this may be due to the activities of bacterial phospholipases producing a glycerol-inositol phosphate.

Next, to test the ability of the IP multikinase, Ipk2 to phosphorylate IP3, and sequentially IP4 to form IP5, we expressed Ipk2 from Arabidopsis thaliana, along with Pis1-Pik1-Mss4-Plc, and a species coeluting with IP5 was observed (Figure 5F). No peak comigrating with IP4 is observed in this strain, indicating that the two-step Ipk2 catalyzed conversion of IP3 to IP5 proceeds in this bacterial system without accumulation of an IP4 intermediate. Of note, expression of the lipid inositide pathway, Plc, and the soluble IP kinases required the utilization of multiple plasmids, multiple antibiotics, and combined with the decreased growth rate in inositol-free M9 minimal media, the incorporation of 3H-inositol label was dramatically lower in this strain, potentially limiting our ability to detect minor species of IPs, such as, an IP4 intermediate. Finally, to form IP6, we expressed yeast Ipk1, which phosphorylates I(1,3,4,5,6)P5 at the 2-position. When Ipk1 is expressed along with Pis1-Pik1-Mss4-Plc1-Ipk2, a peak coeluting with IP6 is observed, along with peaks corresponding with two different IP5 species (Figure 5G). We hypothesize, but do not prove, that the fist IP5 peak to elute (IP5A) is I(1,3,4,5,6)P5, the product of Ipk2, and the substrate of Ipk1, and the second (IP5B) is the result of a bacterial phytase present in E. coli (Greiner, Konietzny and Jany, 1993). To demonstrate that this inositide pathway is lipid-dependent, we expressed Pis1-Pik1-Mss4-Ipk2-Ipk1, but without Plc, and expectedly do not observe production of any IP species (Figure 5H). Overall, these data indicate that in addition to reconstituting lipid inositide signaling, E. coli are capable of producing the soluble inositol phosphates as well.

Discussion

The ability to reconstitute eukaryotic metabolic signaling pathways in bacterial has functioned to both clarify as well as discover new activities associated with inositide enzyme regulation. Due to the important roles that inositides play in many biological processes, better understanding of the enzymatic activities controlling their production and breakdown is essential to delineating their roles in normal cellular function and in disease states. Eukaryotic systems are exceedingly complex, and there are many complex factors that regulate the inositide kinases and their localization and association with specific membrane compartments within the cell (Venditti et al., 2016). Because of this, utilization of a simpler cell-based system can provide a unique tool to elucidate precise enzymatic functions that may be otherwise difficult to ascertain. While in vitro analysis is useful for confirmatory experiments, a cell-based system enables reductionist-based approaches in a biological context. To this end, we provide evidence that inositide signaling can be successfully reconstituted in E. coli, offering a cell-based platform devoid of endogenous inositide signaling to enable characterization of inositide kinases, phosphatases, and regulatory proteins. During preparation of this manuscript, a report was published presenting a similar system for the expression of a portion of the inositide signaling pathway (Botero, Chiaroni-Clarke and Simon, 2019).

While the synthetic bacterial system presented here was able to reconstitute the canonical inositide synthesis pathway, we also observed a heretofore unappreciated route to PIP3 formation with expression of only Pis1, Pik1, and Mss4 (Figure 2C and Figure 3H). Mss4 has been previously annotated as a canonical PI(4)P 5-Kinase; however, to our knowledge, Mss4 has not been described as a PI(4)P 3-kinase nor as a PI(4,5)P2 3-kinase as our data suggests. An alternative explanation of these data is the presence of an endogenous heretofore undescribed bacterial lipid kinase and to date no putative bacterial inositide kinase has been identified experimentally or based on sequence homology. Description of a previously unrecognized activities of enzymes, such as the PIP multikinase activity of Mss4, are facilitated by our in vivo bacterial system and represent an example of the power of such an approach.

Although the primary pathway to produce PIP3 in mammalian cells is through PI(4)P and PI(4,5)P2, there is also evidence that a PI(3)P dependent pathway to PIP3 is present in both mammalian cells (Halstead et al., 2001) and fission yeast (Mitra et al., 2004) through the S. Pombe Vps34 and Mss4 homologues. Here we demonstrate this alternative pathway to PIP3. When bacteria were grown expressing a truncated form of Vps34 along with Pis1, we were able to observe the production of PI(3)P (Figure 2B). Further, with the addition of Mss4 we observed both PI(3,X)P2, and PIP3 (Figure 2C and Figure 3G). These data recapitulate this proposed alternate pathway to produce PIP3 with robust conversion of PI to PIP3 via PI(3)P and PI(3,X)P2 intermediates. While this PI(3)P-based pathway to PIP3 in this E. coli system was very robust, this is a secondary pathway in mammalian cells, a minor pathway in fission yeast only detectible after deletion of phosphatases, and interestingly, has yet to be observed in budding yeast. We believe this demonstrates the utility of our system to expose the activity of the enzymes free from alterations in cell compartmentalization, phosphatases, or other regulatory factors.

Reconstitution of the inositide pathway in E. coli revealed the formation of PIP3, an inositide not observed in S. cerevisiae, despite utilizing only S. cerevisiae enzymes (Figure 2C lane 1, Figure 3G). One possibility for this observation is the difference in spatiotemporal distribution of enzymes and their lipid substrates. A prokaryotic system enables characterization of inositide enzyme activities in a largely “noise-less” experimental context, where the kinases, phosphatases and regulatory proteins may have access to lipids without consideration for subcellular localization. Additionally, in eukaryotes the inositide pathway is highly regulated by phosphatases, while in this bacterial system there are no endogenous inositide phosphatases. In S. cerevisiae, a pathway to generate PIP3 may exist, but is suppressed by phosphatases that act on either PIP3 or the PI(3,X)P2 intermediate.

Our data also reconstitute and clarify the lipid-dependent and soluble inositide kinase pathways required to produce higher ordered IPs such as IP4, IP5 and IP6. The ability to recapitulate production of IP6 with six gene products is the first such proof of its kind. While the core of the inositide signaling pathway is conserved throughout eukaryotes, there are certain variations between single celled eukaryotes and metazoans that can help reveal important aspects of biology (Saiardi et al., 2018). The modular nature of our bacterial expression system lends itself well to studying these variations in the inositide pathway. Our system will enable future studies aimed at elucidating new IP kinases and in some cases their promiscuity, sometimes called “moonlighting” activities, of such enzymes.

In addition to studying the process of making PIPs and IPs, this synthetic system could be used to study the proteins that interact with PIPs and IPs. Many proteins have been identified as PIP-binding partners (Blind et al., 2014; Sablin et al., 2015; Sun and MacKinnon, 2017; Delgado-Ramírez et al., 2018), and this system could be used to co-express and study these proteins along with their endogenous binding lipids without the complex environment of the eukaryotic cell. In the case of IPs, a number of proteins have been shown to rely on IP6 as a structural cofactor (Macbeth et al., 2005; Sheard et al., 2010; Ouyang et al., 2016). For these proteins, expression of protein for structural studies was not possible using E. coli since they lack IP6, a key factor for their folding. This synthetic system may allow for structural studies of proteins that utilize IPs as cofactors to be conducted in E. coli rather than in more cumbersome expression systems.

While E. coli do not contain endogenous inositide signaling, there are still some potential caveats worth discussing. First, E. coli do encode a gene with homology to an inositol phosphatase, CysQ, that has homology to the mammalian I(1,4)P2 1-phosphatase inositol polyphosphate 1-phosphatase (INPP1). In vitro analysis indicates that CysQ’s basal activity for I(1,4)P2 is low (Spiegelberg et al., 1999), and its in vivo activity in this bacterial system is minimal, as we still observe appreciable signal where I(1,4)P2 elutes on HPLC (Figure 5C and 5E). Second, E. Coli contain bacterial phospholipases, and we observe the formation of small amounts of lyso-PI in some of our bacterial extracts examined by TLC (Figure 2A). However, combining TLC analysis of inositide lipids with HPLC analysis of deacylated glycerol-inositide phosphates, we are able to resolve lyso-PI from other species (Figure 3). We also observe that in bacteria expressing Pis1, Pik1 and Mss4, but not Plc, the bacteria convert some of the lipid signal into a soluble signal presumably through the activity of endogenous bacterial lipases (Figure 4C). We also find that this bacterial lipase activity can be inhibited by chloramphenicol (Figure 4D and Figure 5D), consistent with previously published work (SMITH, WORREL and SWANSON, 1949). Finally, the presence of bacterial phytase may catalyze the formation of additional IP species, although its overall effects appear to be small as the levels of IP6 produced are much greater than the levels of possible phytase products (Figure 5G). However, utilizing this system as a discovery tool for IP kinase/phosphatase activities may necessitate the inhibition of this phytase either genetically or pharmacologically.

While here we demonstrate an operon-based system for expression of the lipid portionof the inositide synthesis pathway, future iterations of this system will be able to stably express the entire pathway in a synthetic operon-based system, and even integration of this pathway into the E. coli genome. Additionally, we observed relatively inefficient radiolabeling with inositol, frequently having greater than 95% of the 3H-inositol in the unspent growth medium. This suggests that import of inositol into the cytosol of bacteria is relatively poor, as compared to eukaryotic cells. Thus, addition of an inositol transporter may be of benefit to improving the yield. Nonetheless, our data is a proof of concept that E. coli can be modified to express both the PIP and IP pathways, and can provide a useful tool for discovery and characterization of key inositide regulatory processes.

Acknowledgements:

We thank members of the York lab for helpful discussions and comments. This work was supported by funds from the Howard Hughes Medical Institute and from the National Institutes of Health grants R01 HL055672 and GM124404 (all to J.D.Y). J.D.Y. is an alumni investigator of the Howard Hughes Medical Institute.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Alcázar-Román AR and Wente SR (2008) ‘Inositol polyphosphates: A new frontier for regulating gene expression’, Chromosoma. doi: 10.1007/s00412-007-0126-4. [DOI] [PubMed] [Google Scholar]
  2. Ando H et al. (2018) ‘Remodeling of Ca2+ signaling in cancer: Regulation of inositol 1,4,5-trisphosphate receptors through oncogenes and tumor suppressors’, Advances in Biological Regulation. doi: 10.1016/j.jbior.2017.12.001. [DOI] [PubMed] [Google Scholar]
  3. Balla T (2013) ‘Phosphoinositides: Tiny Lipids With Giant Impact on Cell Regulation’, Physiological Reviews. doi: 10.1152/physrev.00028.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Berridge MJ et al. (1983) ‘Changes in the levels of inositol phosphates after agonist dependent hydrolysis of membrane phosphoinositides.’, The Biochemical journal. doi: 10.1042/bj2120473. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Blind RD et al. (2014) ‘The signaling phospholipid PIP 3 creates a new interaction surface on the nuclear receptor SF-1’, Proceedings of the National Academy of Sciences. doi: 10.1073/pnas.1416740111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bonangelino CJ et al. (2002) ‘Osmotic stress-induced increase of phosphatidylinositol 3,5-bisphosphate requires Vac14p, an activator of the lipid kinase Fab1p’, Journal of Cell Biology. doi: 10.1083/jcb.200201002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Botero S, Chiaroni-Clarke R and Simon SM (2019) ‘Escherichia coli as a platform for the study of phosphoinositide biology’, Science advances. doi: 10.1126/sciadv.aat4872. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cantley LC (2002) ‘The phosphoinositide 3-kinase pathway’, Science. doi: 10.1126/science.296.5573.1655. [DOI] [PubMed] [Google Scholar]
  9. Delgado-Ramírez M et al. (2018) ‘Regulation of Kv2.1 channel inactivation by phosphatidylinositol 4,5-bisphosphate’, Scientific Reports. doi: 10.1038/s41598-018-20280-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Desrivières S et al. (1998) ‘MSS4, a phosphatidylinositol-4-phosphate 5-kinase required for organization of the actin cytoskeleton in Saccharomyces cerevisiae’, Journal of Biological Chemistry. doi: 10.1074/jbc.273.25.15787. [DOI] [PubMed] [Google Scholar]
  11. Frederick JP et al. (2005) ‘An essential role for an inositol polyphosphate multikinase, Ipk2, in mouse embryogenesis and second messenger production’, Proceedings of the National Academy of Sciences of the United States of America. doi: 10.1073/pnas.0503706102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fruman D. a, Meyers RE and Cantley LC (1998) ‘Phosphoinositide kinases.’, Annual review of biochemistry. doi: 10.1146/annurev.biochem.67.1.481. [DOI] [PubMed] [Google Scholar]
  13. Greiner R, Konietzny U and Jany KD (1993) ‘Purification and characterization of two phytases from Escherichia coli’, Archives of biochemistry and biophysics. doi: 10.1006/abbi.1993.1261. [DOI] [PubMed] [Google Scholar]
  14. Halstead JR et al. (2001) ‘A novel pathway of cellular phosphatidylinositol(3,4,5) trisphosphate synthesis is regulated by oxidative stress’, Current Biology. doi: 10.1016/S0960-9822(01)00121-X. [DOI] [PubMed] [Google Scholar]
  15. Hatch AJ, Odom AR and York JD (2017) ‘Inositol phosphate multikinase dependent transcriptional control’, Advances in Biological Regulation. doi: 10.1016/j.jbior.2017.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hatch AJ and York JD (2010) ‘SnapShot: Inositol phosphates’, Cell. doi: 10.1016/j.cell.2010.11.045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Irvine RF (2016) ‘A short history of inositol lipids’, Journal of Lipid Research. doi: 10.1194/jlr.R071712. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Macbeth MR et al. (2005) ‘Structural biology: Inositol hexakisphosphate is bound in the ADAR2 core and required for RNA editing’, Science. doi: 10.1126/science.1113150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Majerus PW et al. (1988) ‘Inositol phosphates: Synthesis and degradation’, Journal of Biological Chemistry. [PubMed] [Google Scholar]
  20. Majerus PW and York JD (2009) ‘Phosphoinositide phosphatases and disease’, Journal of Lipid Research. doi: 10.1194/jlr.R800072-JLR200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Miller S et al. (2010) ‘Shaping development of autophagy inhibitors with the structure of the lipid kinase Vps34’, Science. doi: 10.1126/science.1184429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Mitra P et al. (2004) ‘A novel phosphateidylinositol(3,4,5)P3pathway in fission yeast’, Journal of Cell Biology. doi: 10.1083/jcb.200404150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Monserrate JP and York JD (2010) ‘Inositol phosphate synthesis and the nuclear processes they affect’, Current Opinion in Cell Biology. doi: 10.1016/j.ceb.2010.03.006. [DOI] [PubMed] [Google Scholar]
  24. Mulugu S et al. (2007) ‘A conserved family of enzymes that phosphorylate inositol hexakisphosphate’, Science. doi: 10.1126/science.1139099. [DOI] [PubMed] [Google Scholar]
  25. Nikawa J, Kodaki T and Yamashita S (1988) ‘Expression of the Saccharomyces cerevisiae PIS gene and synthesis of phosphatidylinositol in Escherichia coli.’, Journal of bacteriology. doi: 10.1128/jb.170.10.4727-4731.1988. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Otto JC et al. (2007) ‘Alterations in an inositol phosphate code through synergistic activation of a G protein and inositol phosphate kinases.’, Proceedings of the National Academy of Sciences USA. doi: 10.1073/pnas.0705729104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Otto JC and York JD (2010) ‘Molecular manipulation and analysis of inositol phosphate and pyrophosphate levels in Mammalian cells.’, Methods in molecular biology (Clifton, N.J.). doi: 10.1007/978-1-60327-175-2_3. [DOI] [PubMed] [Google Scholar]
  28. Ouyang Z et al. (2016) ‘Structural Basis and IP6Requirement for Pds5-Dependent Cohesin Dynamics’, Molecular Cell. doi: 10.1016/j.molcel.2016.02.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Raucher D et al. (2000) ‘Phosphatidylinositol 4,5-bisphosphate functions as a second messenger that regulates cytoskeleton-plasma membrane adhesion’, Cell. doi: 10.1016/S0092-8674(00)81560-3. [DOI] [PubMed] [Google Scholar]
  30. Sablin EP et al. (2015) ‘Structure of Liver Receptor Homolog-1 (NR5A2) with PIP3hormone bound in the ligand binding pocket’, Journal of Structural Biology. doi: 10.1016/j.jsb.2015.09.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Saiardi A et al. (2018) ‘Microbial inositol polyphosphate metabolic pathway as drug development target’, Advances in Biological Regulation. doi: 10.1016/j.jbior.2017.09.007. [DOI] [PubMed] [Google Scholar]
  32. Sheard LB et al. (2010) ‘Jasmonate perception by inositol-phosphate-potentiated COI1-JAZ co-receptor’, Nature. doi: 10.1038/nature09430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Shears SB et al. (2017) ‘The significance of the 1-kinase/1-phosphatase activities of the PPIP5K family’, Advances in Biological Regulation. doi: 10.1016/j.jbior.2016.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Shears SB and Wang H (2019) ‘Inositol phosphate kinases: Expanding the biological significance of the universal core of the protein kinase fold’, Advances in Biological Regulation. doi: 10.1016/j.jbior.2018.10.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. SMITH GN, WORREL CS and SWANSON AL (1949) ‘Inhibition of bacterial esterases by chloramphenicol.’, Journal of bacteriology. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Spiegelberg BD et al. (1999) ‘Cloning and characterization of a mammalian lithium-sensitive bisphosphate 3’-nucleotidase inhibited by inositol 1,4-bisphosphate’, Journal of Biological Chemistry. doi: 10.1074/jbc.274.19.13619. [DOI] [PubMed] [Google Scholar]
  37. Stack JH et al. (1993) ‘A membrane-associated complex containing the Vps15 protein kinase and the Vps34 PI 3-kinase is essential for protein sorting to the yeast lysosome-like vacuole.’, The EMBO journal. doi: 10.1111/hisn.12004_8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stolz LE et al. (1998) ‘INP51, a yeast inositol polyphosphate 5-phosphatase required for phosphatidylinositol 4,5-bisphosphate homeostasis and whose absence confers a cold-resistant phenotype’, Journal of Biological Chemistry. doi: 10.1074/jbc.273.19.11852. [DOI] [PubMed] [Google Scholar]
  39. Streb H et al. (1983) ‘Release of Ca2+ from a nonmitochondrial intracellular store in pancreatic acinar cells by inositol-1,4,5-trisphosphate’, Nature. doi: 10.1038/306067a0. [DOI] [PubMed] [Google Scholar]
  40. Sun J and MacKinnon R (2017) ‘Cryo-EM Structure of a KCNQ1/CaM Complex Reveals Insights into Congenital Long QT Syndrome’, Cell. doi: 10.1016/j.cell.2017.05.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Tolia NH and Joshua-Tor L (2006) ‘Strategies for protein coexpression in Escherichia coli’, Nature Methods. doi: 10.1038/nmeth0106-55. [DOI] [PubMed] [Google Scholar]
  42. Venditti R et al. (2016) ‘PI(4)P homeostasis: Who controls the controllers?’, Advances in Biological Regulation. doi: 10.1016/j.jbior.2015.09.007. [DOI] [PubMed] [Google Scholar]
  43. Walsh JP, Caldwell KK and Majerus PW (1991) ‘Formation of phosphatidylinositol 3- phosphate by isomerization from phosphatidylinositol 4-phosphate.’, Proceedings of the National Academy of Sciences of the United States of America. doi: 10.1073/pnas.88.20.9184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wilson MSC, Livermore TM and Saiardi A (2013) ‘Inositol pyrophosphates: between signalling and metabolism’, Biochemical Journal. doi: 10.1042/BJ20130118. [DOI] [PubMed] [Google Scholar]
  45. Zhang X et al. (1997) ‘Phosphatidylinositol-4-phosphate 5-kinase isozymes catalyze the synthesis of 3-phosphate-containing phosphatidylinositol signaling molecules’, Journal of Biological Chemistry. doi: 10.1074/jbc.272.28.17756. [DOI] [PubMed] [Google Scholar]

RESOURCES