Abstract
d-Glycero-β-d-manno-heptose 1,7-biphosphate (β-HBP) is a novel microbial-associated molecular pattern that triggers inflammation and thus has the potential to act as an immune modulator in many therapeutic contexts. To better understand the structure–activity relationship of this molecule, we chemically synthesized analogs of β-HBP and tested their ability to induce canonical TIFA-dependent inflammation in human embryonic kidney cells (HEK 293T) and colonic epithelial cells (HCT 116). Of the analogs tested, only d-glycero-β-d-manno-heptose 1-phosphate (β-HMP) induced TIFA-dependent NF-κB activation and cytokine production in a manner similar to β-HBP. This finding expands the spectrum of metabolites from the Gram-negative ADP–heptose biosynthesis pathway that can function as innate immune agonists and provides a more readily available agonist of the TIFA-dependent inflammatory pathway that can be easily produced by synthetic methods.
Introduction
Heptose is a seven-carbon sugar produced solely by bacteria (1). A phosphorylated heptose metabolite that is generated as an intermediate in the biosynthesis of the LPS carbohydrate heptose was recently recognized as a novel microbial-associated molecular pattern (MAMP) with immunostimulatory potential. This discovery arose from the finding that Neisseria gonorrhoeae released a small soluble factor that induced HIV production in latently infected CD4+ T cell lines (2). Subsequent genetic and mass spectrometry (MS)-based analyses led to the tentative conclusion that d-glycero-d-manno-heptose 7-phosphate (HMP-7), which was found to be a component of neisserial culture supernatants, elicited this response. This hypothesis was later rejected when a series of “deep-rough” mutants were generated via mutagenesis of enzymes within the ADP–heptose biosynthetic pathway. These mutants contained a truncated LPS deficient in heptose residues. Supernatants from strains lacking the HMP-7 kinase, HldA, were not inflammatory. In contrast, supernatants from an isogenic mutant defective in GmhB, the next enzyme in this pathway, did elicit an inflammatory response. This suggested that d-glycero-β-d-manno-heptose 1,7-biphosphate (β-HBP) was an active factor, whereas HMP-7 was not. To validate these findings using an independent approach, in vitro reactions with purified enzymes from the ADP–heptose biosynthetic pathway, starting from sedoheptulose 7-phosphate (S7P) (pathway alone shown in Fig. 1), were performed. Consistent with the bacterial supernatant-derived results, enzymatically synthesized β-HBP activated cellular NF-κB–dependent transcription, whereas HMP-7 did not (3). Notably, addition of the neisserial β-HBP phosphatase GmhB to these β-HBP preparations somewhat reduced the activity, suggesting that the next metabolite in the pathway, d-glycero-β-d-manno-heptose 1-phosphate (β-HMP), was also inactive (3).
Upon entry into the host cytoplasm via endocytosis (3) or direct delivery by bacteria that access the cytosol (3–8), β-HBP induces the threonine phosphorylation–dependent oligomerization of TRAF-interacting protein with forkhead-associated domain (TIFA). This results in the formation of large protein aggregates termed “TIFAsomes,” evident by fluorescence microscopy and native protein gels, which recruit TRAF6 to promote NF-κB nuclear translocation (2–8). The prolonged innate response of cells exposed to β-HBP, presumably stemming from the intracellular accumulation of this agonist (2), implies that β-HBP may be used to modulate immune responses in a wide variety of therapeutic applications, such as to promote HIV exit from latency in CD4+ lymphocytes (2) and to act as an adjuvant for both conventional vaccination and cancer immunotherapy. However, these translational studies have been hampered by the cost and technical difficulty in generating sufficient amounts of β-HBP from the established enzymatic or chemical synthesis approaches (3, 9–12). Although the synthesis of α-anomeric phosphates is straightforward (9–12), the chemical synthesis of β-anomeric heptose phosphates is particularly difficult and inefficient. This prompted us to synthesize and test a range of analogs, with the aim of determining the structure-activity relationship of this response and in the hopes that easier-to-synthesize targets with similar inflammatory potential to β-HBP could be revealed.
In this report, we chemically synthesized seven different heptose analogs (Fig. 2): HMP-7, β-HMP, d-glycero-α-d-manno-heptose 1-phosphate (α-HMP), β-d-mannose phosphate (β-MP), β-HBP, d-glycero-α-d-manno-heptose 1,7-biphosphate (α-HBP) and d-glycero-d-manno-heptose (H). These analogs were tested for their ability to activate NF-κB in human embryonic kidney–derived HEK 293T cells and to induce IL-8 production in colonic epithelial HCT 116 cells. Of the compounds tested, only β-HBP and β-HMP induced TIFA-dependent inflammation, indicating that both the β conformation and the phosphate in the C1 position are necessary for immune activation. This result was unexpected because we had previously observed that treatment of β-HBP with the β-HBP phosphatase GmhB, which should convert β-HBP to β-HMP, appeared to reduce its activity (albeit modestly) relative to untreated β-HBP (3). Thus, to confirm our new findings with the synthetic β-HMP, we conducted extensive enzymatic time course and chemical analyses using purified GmhB and chemically synthesized β-HBP and β-HMP. We confirmed that GmhB acts on β-HBP with high efficiency to generate β-HMP and that enzymatic and synthetic β-HMP have similar activities to β-HBP. Together, these data clearly indicate that the β-HMP is a previously unrecognized MAMP.
Materials and Methods
General
1H, 500 MHz] or Bruker [1H, 600 MHz]) reported with the solvent residual signal (4.79 ppm for 1H).
Chemical structures
Synthesis.
All compounds were synthesized using standard carbohydrate synthetic methods. HMP-7 and H (forgo phosphorylation) was synthesized using the method described by Güzlek et al. (13), and β-HMP and α-HMP were synthesized as described by Zamyatina et al. (14). β-MP was synthesized from acetylated d-mannose using the methods described in Zamyatina et al. (14). β-HBP and α-HBP were synthesized as described by Sauvageau et al. (9). Compound purity and identity were measured using 1H NMR, MS, and a high-performance anion exchange chromatography (HPAEC) system. HPAEC analyses showed only one signal apart for HMP-7, in which two signals were observed (Tables I, II).
HPLC
Phosphorylated derivatives were dissolved at a concentration of 50 μg/ml in deionized water and analyzed using an HPAEC. The derivatives (20 μl) were injected on a CarboPac PA1 column and detected with a pulsed amperometric detector from Dionex. The samples were eluted during a 30 min linear gradient (0–100% solvent B) and 5 min of isocratic elution (100% B). The solvents were as follows: solvent A, NaOH, 0.1 M; solvent B, AcONa, 1 M and NaOH, 0.05 M.
Enzymatic reactions
GmhA, HldA, and GmhB were purified as previously described (1). β-HBP (1 mg/ml, 2.2 mM) and β-HMP (1 mg/ml, 3.0 mM) were subjected to GmhA (4 or 8 μg) and HldA (5 μg), and/or only GmhB (4 μg), with or without ATP (5 mM), overnight at 30°C in 50 mM HEPES (pH 8), 20 mM KCl, and 10 mM MgCl2. In an independent experiment, S7P (2 mM) was incubated with GmhA (2 μg), and HldA (3 μg) was incubated with ATP (4 mM) overnight at 30°C in 20 mM HEPES (pH 8), 20 mM KCl, and 10 mM MgCl2. The reaction was stopped by incubating at 95°C for 10 min and then passed through a 0.22-μm filter. The filtrate was then incubated with or without 2 μg GmhB. At timepoints indicated, an aliquot of the reaction was removed and incubated at 95°C for 10 min to inactivate the enzyme, diluted accordingly to a final concentration of 50 μg/ml, and passed through a MilliporeSigma centrifugal 10 K filter (for the β-HBP and β-HMP reactions) or a 0.22 μm filter (for the S7P reaction). The fractions were then freeze-dried, analyzed with 1H NMR, diluted and analyzed via HPAEC, and assessed for their ability to activate NF-κB.
Biological assays
Synthetic compounds were resuspended in endotoxin-free water (Sigma-Aldrich, Oakville, ON, Canada). HCT 116 cells were maintained in McCoy’s 5A medium. HEK 293T cells were maintained in DMEM. All media were supplemented with 10% FBS and 1% GlutaMAX.
Reversible digitonin assays using wild-type and TIFA-knockout HEK 293T and HCT 116 cells were performed as previously described (6). Briefly, HEK 293T cells were cotransfected with an NF-κB–driven firefly luciferase reporter plasmid and a Renilla luciferase plasmid (Promega), the latter being used as a constitutively expressed internal control. After 24 h, cells were stimulated for 20 min with synthetic compounds at various concentrations in permeabilization buffer (5 μg/ml digitonin), washed once, and then incubated for 6 h in complete DMEM. Luciferase activity was determined using the Dual-Glo Luciferase Assay System (Promega) as previously described (3), and Renilla luminescence was measured using a luminometer (Cytation 5; BioTek). Results are expressed as fold increase relative to transfected, water-treated cells and normalized to Renilla luminescence. HCT 116 cells were stimulated for 20 min with the NOD1 agonist C12-iE-DAP (20 μg/ml; InvivoGen) or with synthetic compounds at various concentrations in permeabilization buffer (5 μg/ml digitonin), washed twice, and then incubated for 6 h in complete media. Quantitative measurements of IL-8 levels in the culture supernatant was then performed using an ELISA kit from BD Biosciences.
Results
To test the specificity of the TIFA-dependent response for β-HBP related to the biosynthetic pathway (Fig. 1), we synthesized the different analogs of β-HBP illustrated in Fig. 2 using methods described in previous reports (9, 13, 14) and then confirmed their purity and identity using NMR, MS spectrometry, and HPAEC (Tables I, II). Fig. 3, Table I, and Table II show the 1H NMR spectra, data, and MS spectrometry validating the identity of each heptose analog that we synthesized. To test for their induction of an inflammatory response, we have taken advantage of two well-established models: human embryonic kidney–derived HEK 293T cells transfected with an NF-κB–driven luciferase reporter and IL-8 expression by human colon–derived HCT 116 cells, in each of which we have also generated TIFA-deficient derivatives to confirm the specificity of the response. We confirmed that HMP-7 and heptose had no inflammatory activity because they did not stimulate NF-κB reporter or IL-8 expression (Figs. 4A, 5A). Similarly, neither α-HMP nor α-HBP showed any activity in either assay. Moreover, β-MP also showed no activity. Unexpectedly, β-HMP displayed substantial activity, inducing nearly a 100-fold increase in luciferase activity over background in HEK 293T cells and close to 1500 pg/ml IL-8 from HCT 116 cells. This activity was attributable to the canonical HBP-activated pathway because cell lines deficient in TIFA did not respond. When provided at equal mass/volume concentrations, β-HMP was more active than β-HBP by 2- to 5-fold (Figs. 4A, 5A). To test whether β-HMP was a more potent immune agonist than β-HBP, we compared equal molarities of β-HBP and β-HBP in HEK 293T cells and HCT 116 cells (Figs. 4B, 5B). Consistent with β-HBP (457.93 g/mol) having a higher m.w. compared with β-HMP (334.00 g/mol), their activities were comparable when provided at equal molarities (Figs. 4B, 5B). There remained modest differences between the responses to β-HMP and β-HBP, with the former consistently eliciting marginally more NF-κB activity (Fig. 4B) and IL-8 response (Fig. 4C) in HEK 293T, whereas the latter was more active in HCT 116 (Fig. 5B).
ADP–heptose biosynthetic pathway in neisserial species. S7P is transformed into HMP-7 with GmhA and then into β-HBP by HldA. Subsequent dephosphorylation via GmhB should lead to the production of β-HMP, which is transformed into ADP–heptose (H-ADP) by HldC. The enzymes used by Escherichia coli and other Gram-negative bacteria, specifically the bifunctional nature of E. coli HldE, are indicated in brackets.
Structures of chemically synthesized analogs. β-MP, α-HMP, α-HBP, H, HMP-7, β-HBP, and β-HMP.
1H NMR spectra of synthetic compounds. 1H NMR spectra in D2O measured on a Varian 500 MHz or on a Bruker Advance 600 MHz of chemically synthesized targets (A) h, (B) HMP-7, (C) β-MP, (D) β-HMP, (E) β-HBP, (F) α-HMP, and (G) α-HBP. h (A) is in the hemiacetal form and thus an α/β mixture as represented in Fig. 2 at the anomeric position. The chemical shifts of the α-anomeric proton (5.17 ppm) and the β-anomeric proton (4.87 ppm) are similar to those of HMP-7 (B), as expected. The main differences between these two spectra (A and B) is H6 (4.17 ppm) and H7 (4.09 ppm) being deshielded due to the presence of an O-phosphate ester at C7 for HMP-7 (B). These two spectra differ from β-MP (C), as at first, mannose is a pyranose and only an anomer is present (β). Clearly, the chemical shift of the anomeric proton is consistent with a β-phosphate (5.13 ppm), and the large coupling constant between this proton with phosphorous (9 Hz) confirms the presence of a phosphate ester at the anomeric position. Another very characteristic signal for a β-mannose is the shielded chemical shift of H5 (3.45 ppm). This pattern is repeated for β-HMP H5 [3.50 ppm, (D)] and β-HBP H5 [3.62 ppm, (E)]. A main difference between β-MP and β-HMP 1H spectra is that H6 is more deshielded for β-HMP (4.04 ppm) because it is a CH, whereas it is a CH2 for β-MP (3.70 ppm). These two spectra are both slightly different from the β-HBP spectrum. The same deshielding effect observed for H6 and H7 seen in HMP-7 is repeated for β-HBP with H6 (4.30 ppm) and H7 (4.20 ppm). The α anomers, α-HMP (F) and α-HBP (G) anomeric protons, both also show a large coupling constant indicator of a phosphate ester at the anomeric position. Differently to the β-anomers, the chemical shift of the α-anomers are deshielded (α-HMP, 5.37 ppm and α-HBP, 5.41 ppm) as well as H5 (α-HMP, 3.92 ppm and α-HBP, 3.93 ppm). The NMR data were in accordance with previously published data (9, 13, 14).
Inflammatory response of human embryonic kidney cells (HEK 293T) exposed to chemically synthesized heptose phosphate analogs. Wild-type (WT) or TIFA knockout (TIFA KO) cells were treated with compounds for 20 min in the presence of digitonin. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (A) Cells were treated with 10 μg/ml of the indicated compound or positive control. Data are means ± SE mean of technical triplicates. The results were analyzed using two-way ANOVA, Dunnett multiple comparisons test, comparing all treatment groups within each genotype to b-HMP treatment. (B) Cells were treated with two different molar concentrations of compound (30 or 150 μM). Two batches of β-HBP and β-HMP synthesized independently of one another (batch no. 1 and batch no. 2) were tested to confirm reproducibility of synthesis additionally to a NOD1 agonist. Data are means ± SE mean of technical triplicates and are representative of five independent experiments. The data were analyzed using two-way ANOVA, Sidak multiple comparisons test, comparing all treatment groups within each genotype to one another. (C) IL-8 production from wild-type (WT) or TIFA knockout (TIFA KO) HEK 293T cells was treated with the compounds in the presence of digitonin for 20 min and then removed by washing, and the IL-8 response was then measured by ELISA at 6 h posttreatment to complement the NF-κB luciferase data. The data were analyzed using two-way ANOVA, Dunnett multiple comparisons test, comparing all treatment groups within each genotype to β-HMP treatment.
Inflammatory response of human colonic epithelial cells (HCT 116) exposed to chemically synthesized analogs. IL-8 production from wild-type (WT) or TIFA knockout (TIFA KO) cells treated with compounds in the presence of digitonin for 20 min and then removed by washing, and the IL-8 response was then measured by ELISA at 6 h posttreatment. #None detected. (A) Cells were treated with 10 μg/ml of the indicated compound. Data are means ± SE mean of technical triplicates. The results were analyzed using two-way ANOVA, Dunnett’ multiple comparisons test, comparing all treatment groups within each genotype to β-HMP treatment. ****p < 0.0001. (B) Cells were treated with the NOD1 agonist C12-iE-DAP (20 μg/ml) or with two different molar concentrations of the compounds (30 or 150 μM). Two batches of β-HBP and β-HMP synthesized independently of one another (batch no. 1 and batch no. 2) were tested to confirm the reproducibility of synthesis. Data are means ± SE mean of two (150 μM treatment) or three (all other treatments) independent experiments. All treatments were performed in triplicate. The data were analyzed using two-way ANOVA, Sidak multiple comparisons test, comparing all treatment groups within each genotype to one another.
To establish the specificity of the enzymatic activity of GmhB and that it efficiently converts β-HBP to β-HMP, we incubated purified GmhB with chemically synthesized heptose derivatives. When GmhB was incubated with β-HMP, we did not observe any significant changes via 1H NMR or HPAEC or in NF-κB activation, confirming that GmhB does not act upon β-HMP (Fig. 6). In contrast, our 1H NMR and HPAEC data confirm that β-HBP is dephosphorylated into β-HMP, with the reaction reaching completion within 1 h. No changes were observed in the 1H NMR spectra time course or HPAEC data once β-HBP was converted to β-HMP, and no significant decrease was observed between the time points in NF-κB activation (Fig. 7), indicating that this reaction does not yield alternate products. To reproduce the conditions used in the previous report and determine whether there was any accumulation of ADP–heptose biosynthesis pathway intermediate products that alter the activity of the TIFA response, a three-step “one-pot” reaction containing S7P, ATP, GmhA, HldA, and GmhB was performed. This demonstrated complete conversion of S7P to β-HMP, and the reaction product induced similar levels of NF-κB activation as chemically synthesized β-HMP and β-HBP (Fig. 8). Notably, incubation of β-HMP with ATP and HldA yielded β-HMP and not ADP–heptose (Fig. 9), confirming that HldA is not a bifunctional enzyme.
GmhB does not modify β-HMP. Inflammatory activity and chemical analyses of synthetic β-HMP samples before and after treatment with GmhB for up to 48 h is displayed. (A) Effect of β-HMP treated with GmhB with or without ATP at various timepoints on human embryonic kidney cells (HEK 293T) encoding an NF-κB–driven luciferase reporter gene. Wild-type (WT) or TIFA knockout (TIFA KO) cells were treated with 30 μM of each compound for 20 min in the presence of digitonin, washed once, and then incubated in complete media for 6 h. NF-κB luciferase activity was measured 6 h after treatment. Data are means ± SE mean of one independent experiment performed in triplicate. (B) HPAEC analyses of β-HMP treated with GmhB with or without ATP show no significant differences compared with untreated β-HMP at all timepoints. For β-HMP treated with GmhB, the signals at a retention time of 0–5 min include glycerol from the enzyme solution and an unidentified compound from either the buffer/enzyme solution or the HEPES buffer. The signals at a retention time of 9.5 min is a baseline artifact. β-HMP is at a retention time of 9.9 min. ATP is at a retention time of 28 min. (C) 1H NMR (D2O, 600 MHz) of ATP, β-HMP, β-HMP + GmhB, and β-HMP + GmhB + ATP.
GmhB conversion of β-HBP to β-HMP does not reduce the TIFA-dependent inflammatory response. Data on time course on synthetic β-HBP treated with GmhB. (A) Effect of β-HBP treated with GmhB after 1, 4, 8, 12, 24, and 48 h on human embryonic kidney cells (HEK 293T) encoding an NF-κB–driven luciferase reporter gene. NF-κB luciferase activity in wild-type (WT) or TIFA knockout (TIFA KO) cells treated with compounds for 20 min in the presence of digitonin and then removed by washing was measured 6 h after treatment. Cells were treated with 10 μg/ml of the indicated compound. Data are means ± SE mean of technical triplicates. (B) HPAEC analyses of β-HBP treated with GmhB showing a clear conversion to β-HMP after 1 h, with no degradation observed after 48 h. The signals occurring at a retention time of 0–5 min are glycerol from the enzyme solution, an unidentified compound deriving from the buffer/enzyme solution and the HEPES buffer. At a retention time of 9.5 min, this is a baseline artifact, at a retention time of 9.9 min (β-HMP). (C) 1H NMR (D2O, 600 MHz) of β-HBP, β-HBP, and β-HBP + GmhB; t = 0, 1, and 48 h.
(A) Data on S7P + GmhA + HldA + gmhB at two different concentration of GmhA (4 μg, Reaction B, and 8 μg, Reaction A) and β-HMP treated with HldA and ATP (Reaction C) on human embryonic kidney cells (HEK 293T) encoding an NF-κB–driven luciferase reporter gene. NF-κB luciferase activity in wild-type (WT) or TIFA knockout (TIFA KO) cells treated with compounds for 20 min in the presence of digitonin and then removed by washing was measured 6 h after treatment. Cells were treated with 30 μM of the indicated compound. Data are mean ± SE mean of technical triplicates. (B) HPAEC analyses of S7P + GmhA (4 and 8 μg) + HldA + GmhB showing a clear conversion to β-HMP over 12 h. The signals at a retention time of 0–5 min are glycerol from the enzyme solution, an unknown related to the buffer/enzyme solution and the HEPES buffer. The signal at 9.5 min is a baseline artifact. β-HMP signal is at 9.9 min. β-HMP treated with HldA and ATP showed no conversion to H-ADP.
Rapid and complete conversion of enzymatically synthesized β-HBP treated by GmhB. (A) Effect of enzymatically synthesized HBP treated with GmhB or mock (water) after 0 h (boiled immediately after enzyme was added), 4, 8, or 12 h on human embryonic kidney cells (HEK 293T) encoding an NF-κB–driven luciferase reporter gene. NF-κB luciferase activity in wild-type (WT) cells treated with compounds for 20 min in the presence of digitonin and then removed by washing was measured 6 h after treatment. Data are means ± SE mean of technical duplicates. The data were analyzed using two-way ANOVA, Sidak multiple comparison test, comparing all groups within each genotype to one another. Reactions A, B, and C are not statistically significantly (ns) different from each other nor from HBP treatment. (B) HPAEC analyses of enzymatically synthetized β-HBP compared with enzymatically synthetized β-HBP treated with GmhB after 0, 4, 8, or 12 h showing that β-HMP is formed as soon as GmhB is mixed with β-HBP even at t = 0 (treatment of enzyme followed by heat degradation).
Discussion
We have previously used both enzymatic and chemical synthetic processes to demonstrate that β-HBP activates NF-κB via a TIFA-dependent signaling cascade (3, 9). Given that the process by which HBP is sensed within the cell remains to be elucidated, we aimed to generate a series of structurally related analogs to understand the requirements for eliciting the response. For this purpose, we chemically synthesized β-HBP, β-HMP, α-HBP, α-HMP, β-MP, H, and HMP-7, which allowed us to dissect the importance of phosphate presence, location, and configuration for signaling.
The synthesis and testing of the different isomers provided clear insights regarding the type of structures that elicit inflammation. First, the synthetic approach confirms that HMP-7 is not active, consistent with the enzyme synthesis-based studies (3), nor is simple heptose. Although previous work suggested that β-HMP may not be active, our synthetic, enzymatic, and structural studies described in this article make it clear that β-HMP elicits a similar response to β-HBP when compared on a molar basis. Critically, our observation that β-MP is not active indicates a clear specificity for heptose-based sugars as β-MP and β-HMP are indistinguishable except for the additional exocyclic carbon in the latter. To stimulate a response, the heptose must be phosphorylated at the reducing-end of the sugar, and this phosphate must exist in the β conformation, because α-anomers (α-HMP and α-HBP) do not show activity, whereas both β-anomers (β-HMP and β-HBP) do.
Although the activity of β-HMP indicates that the phosphate at the seven position is not essential, it is curious that it does influence the response in a cell type–specific manner. Specifically, we consistently observed that β-HMP elicited more activity than β-HBP with respect to both NF-κB reporter expression and IL-8 production in HEK 293 cells, whereas both compounds were similarly active in HCT 116 cells. It remains unclear whether this is due to differential expression of as yet unidentified signaling components or differential uptake of the agonists by these two cell types.
Although the modest effect of GmhB on enzymatically synthesized β-HBP was previously interpreted to suggest that β-HMP was not active (3), the analytical and in vitro data performed on both enzymatically produced and chemically synthesized analogs presented in this study clearly indicates that β-HMP and β-HBP are both immunologically active MAMPs that induce an NF-κB–dependent inflammatory response in a TIFA-dependent manner. This is highly impactful as we have identified a second heptose-derived metabolite within the Gram-negative bacterial LPS biosynthesis cascade that may contribute to inflammation during infection. Moreover, from a practical point of view, β-HMP is more readily available because it is significantly easier and cheaper to produce than β-HBP. This will facilitate future research to understand the TIFA-dependent innate immune cascade and to establish the use of β-HMP for immunomodulatory applications.
Disclosures
C.X.G., S.D.G.-O., J.S., and A.D.C. have filed patents concerning the therapeutic potential of β-HBP and β-HMP. The other author has no financial conflicts of interest.
Acknowledgments
We are grateful to Ken Chan and The-Minh Tu for measuring MS on the synthesized compounds. We thank Evgueni Vinogradov, Wei Zou, Dean Williams, Ryan Gaudet, Furkan Guvënç, and Amit Weiner for valuable discussions during the course of this work.
Footnotes
This work was supported by Canadian Institutes of Health Research Operating Grant HOP-137697 to S.D.G.-O.
Abbreviations used in this article:
- H
- d-glycero-d-manno-heptose
- α-HBP
- d-glycero-α-d-manno-heptose 1,7-biphosphate
- β-HBP
- d-glycero-β-d-manno-heptose 1,7-biphosphate
- α-HMP
- d-glycero-α-d-manno-heptose 1-phosphate
- β-HMP
- d-glycero-β-d-manno-heptose 1-phosphate
- HMP-7
- d-glycero-d-manno-heptose 7-phosphate
- HPAEC
- high-performance anion exchange chromatography
- MAMP
- microbial-associated molecular pattern
- β-MP
- β-d-mannose phosphate
- MS
- mass spectrometry
- NMR
- nuclear magnetic resonance
- S7P
- sedoheptulose 7-phosphate
- TIFA
- TRAF-interacting protein with forkhead-associated domain.
- Received July 20, 2018.
- Accepted August 17, 2018.
- Copyright © 2018 by The American Association of Immunologists, Inc.
In this issue
