Letter Cite This: Org. Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/OrgLett
Chemoenzymatic Synthesis of C6-Modified Sugar Nucleotides To Probe the GDP‑D‑Mannose Dehydrogenase from Pseudomonas aeruginosa Sanaz Ahmadipour,† Giulia Pergolizzi,‡ Martin Rejzek,‡ Robert A. Field,‡ and Gavin J. Miller*,† †
Lennard-Jones Laboratory, School of Chemical and Physical Sciences, Keele University, Keele, Staffordshire ST5 5BG, United Kingdom ‡ Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Norwich, NR4 7UH, United Kingdom
Downloaded by BETHEL UNIV at 17:00:13:674 on May 30, 2019 from https://pubs.acs.org/doi/10.1021/acs.orglett.9b00967.
S Supporting Information *
ABSTRACT: The chemoenzymatic synthesis of a series of C6-modified GDP-D-Man sugar nucleotides is described. This provides the first structure−function tools for the GDP-D-ManA producing GDP-D-mannose dehydrogenase (GMD) from Pseudomonas aeruginosa. Using a common C6 aldehyde functionalization strategy, chemical synthesis introduces deuterium enrichment, alongside one-carbon homologation at C6 for a series of mannose 1-phosphates. These materials are shown to be substrates for the GDP-mannose pyrophosphorylase from Salmonella enterica, delivering the required toolbox of modified GDPD-Mans. C6-CH3 modified sugar-nucleotides are capable of reversibly preventing GDP-ManA production by GMD. The ketone product from oxidation of a C6-CH3 modified analogue is identified by high-resolution mass spectrometry.
A
lginate constitutes the major component of the exopolysaccharide coating within mucoidal Pseudomonas aeruginosa infections, which are particularly deleterious for cystic fibrosis patients,1 contributing to the biofilm environment and augmenting the antibiotic resistance profile of the bacterium. The molecular genetics and regulation of alginate biosynthesis2 have been well-characterized in this human pathogen.3 However, the biological chemistry of the underpinning biosynthetic polymer production and modification processes is still insufficiently understood. Pivotal to this biosynthetic pathway is GDP-D-mannose dehydrogenase (GMD), which is the enzyme required to form the alginate sugar nucleotide feedstock, GDP-D-ManA (Figure 1a).4,5 Since there is no corresponding enzyme in humans, specific inhibition of GMD could be envisaged as a tactic to stop alginate production and interfere with biofilm formation in chronic mucoid P. aeruginosa infections. GMD is a member of a small group of NAD+-dependent four-electron-transfer dehydrogenases, which includes UDP-glucose dehydrogenase (UGD),6,7 and converts GDP-D-Man 1 to GDP-D-ManA 2. The GMD-catalyzed oxidation step is proposed to have four discrete steps (Figure 1b), coordinated by an active site Cys268 © XXXX American Chemical Society
Figure 1. (a) Oxidation of GDP-D-Man 1 to GDP-D-ManA 2, which is the feedstock sugar nucleotide for alginate biosynthesis. R= H or Ac. (b) Proposed oxidative conversion of 1 to 2, using a key active site cysteine residue.
residue, with the oxidation of C6 aldehyde 3 to thioester 5 (intermediated by thiohemiacetal 4), followed by hydrolysis to give 2. Received: March 19, 2019
A
DOI: 10.1021/acs.orglett.9b00967 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters In order to investigate this mechanism of action for GMD and to lay a foundation for the development of potential inhibitors, we report herein the design, synthesis, and preliminary evaluation of a series of C6-modified GDP-DMan tools (Figure 2). Modified sugar nucleotide analogues
Scheme 1. Synthesis of Diastereomerically Pure C6-Methyl Mannose 1-Phosphates 9 and 10 from Common Aldehyde Precursor 6
Figure 2. Strategy to GDP-D-Man tools, modified at C6 from a common aldehyde intermediate.
separated into amounts of each required anomeric 1-phosphate at a later stage. Mixture 7 proved difficult to separate by silica gel chromatography, and, therefore, we opted to continue a single synthetic route toward 9 and 10, first completing benzylation of the C6-OH. Following this reaction, the C6 stereochemistry was assigned by comparison to 1H and 13C NMR data of known 6S-methyl-tetrabenzyl thioglycoside reported by Davis,14 who confirmed the stereochemistry at C6 using nuclear Overhauser effect (nOe) analysis of a 4,6 benzylidene derivative. Our data for the mixture of C6 diastereoisomers showed a strong match (C1 δ 85.6 vs 85.5 ppm reported; H1 δ 5.83 (d, J = 1.7 Hz) vs 5.81 ppm reported) and showed the 6S diastereoisomer to be the minor component in our mixture (data for the 6R diastereoisomer was distinct from 6S (C1 δ 85.8; H1 δ 5.62 (d, J = 1.9 Hz)). We next completed anomeric 1-phosphate installation (in protected form) using dibenzylphosphate (DBP) as the acceptor under thioglycoside activation conditions (NIS/ AgOTf) in satisfactory yield (63%, two steps). 1H and 31P NMR confirmed the presence of the anomeric phosphate with the characteristic doublet of doublets observed for H1 coupling to H2 and 31P (3JH1−H2 = 2.0 Hz, 3JH1−P = 6.1 Hz for 10). This approach then allowed a late-stage chromatographic resolution of the diastereomeric material 8S/8R before a final hydrogenolysis to deliver free anomeric phosphates 9 and 10. The final deprotection reaction required some optimization to align the solubility differences of the organic starting material and the aqueous soluble glycosyl 1-phosphate. We evaluated several alternative solvent systems, finally settling on a 6:2:1 mixture of EtOH:THF:5% aqueous NaHCO3 (to provide the sodium salt form of the phosphate). A catalyst loading of 0.2 mol equivalents of 1:1 w/w Pd(OH)2/C:Pd/C per OBn group was also established. We found that the use of pressures above that of a balloon of hydrogen (up to 10 bar in a Parr vessel) and increasing the temperature of the reaction (up to 55 °C from ambient) provided no immediate increase in the rate of completion, rather leading to mixtures of products and incomplete hydrogenolysis, which was then susceptible to anomeric degradation as the reaction continued. Our optimized conditions consistently delivered the deprotected glycosyl 1-phosphates in good yields (vide infra). For the synthesis of monodeuterated 6R and 6S mannose 1phosphate targets 16 and 17, we started from known uronate
have valuable potential as probes to study glycosyltransferases and other enzymes that use these activated glycosylating agents as substrates.8,9 Therefore, the synthesis (both enzymatic and chemical) of natural and non-natural sugar nucleotides is a topic of continuing interest and challenge.10,11 We selected targets with either deuterium or small alkyl group (Me) modifications at C6 of the parent pyranose ring to enable: (i) probing of the GMD active site space for non-native substrate binding, (ii) establishing evidence for a ketone or thiohemiketal intermediate in the oxidation mechanism (in place of 3 or 4), and (iii) capability to assess the diastereoselectivity of proton abstraction during the oxidation of 1 to 3. D-Mannose systems diastereoselectively deuterated at C6 have previously provided important chemical tools, illustrated by the synthesis of ADP-[6″-D]-D,D-Hep for elucidating the mechanism of ADP-L-glycero-D-manno-heptose 6-epimerase.12 Furthermore, pyranose C6 homologation using small alkyl groups has provided important mechanistic probes for the study of UDP-glucose dehydrogenase, which is the enzyme responsible for the oxidative conversion of UDP-Glc to UDP-GlcA.6 We hypothesized that, starting from D-mannose, a suitably protected C6 aldehyde thioglycoside donor would serve as the common material to access a series of C6-modified glycosyl 1phosphates and sugar nucleotide targets (Figure 2). Therefore, we began our synthesis from D-mannose, accessing the required C6 aldehyde 6, following established procedures on a multigram scale (see Figure S1 in the Supporting Information (SI)). Using MeMgBr, we first sought to evaluate the addition of small alkyl groups to 6 (Scheme 1), to enable one-carbon homologation. However, this reaction yielded significant amounts (up to 80% by crude 1H NMR) of a competing C4−C5 elimination product. In order to reduce the basicity of the organometallic reagent (reactions using reduced equivalents of Grignard did not prevent the elimination), transmetalation of MeMgBr with CeCl3 afforded the corresponding organocerium13 reagent in situ. Subsequent treatment of 6 with this material suppressed competing elimination and delivered the target secondary alcohol 7 on a multigram scale as a mixture of C6 diastereoisomers (diastereomeric ratio (dr) of 4:6) in good yield (86%). While the observed diastereoselectivity for this nucleophilic addition was low, it was, for our intended purposes, beneficial, because it allowed the material to be B
DOI: 10.1021/acs.orglett.9b00967 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters ester 1115 (Scheme 2), reasoning that a diastereoselective reduction of deuterated aldehyde 13 would be an effective
13
C−2H coupling of 23.1 Hz. Having completed their stereochemical assignments, both 14 and 15 were converted through to their respective 1-phosphates 16 and 17 using the previously established procedures (Scheme 2). From thioglycoside 12, we also synthesized a doubly C6-deuterated mannose 1-phosphate 18. These materials, along with 9 and 10, completed a series of regio-defined and stereo-defined mannose 1-phosphate chemical tools, which were next evaluated as substrates for a recombinant mannose pyrophosphorylase (from Salmonella enterica18) in converting them to their respective GDP conjugates. Glycosyl 1-phosphates 9 and 10 and 16−18 were individually incubated with GDP-mannose pyrophosphorylase at 37 °C. TLC monitoring of the reactions at 18 h indicated that the majority of the glycosyl 1-phosphates had been consumed and the reactions were stopped. The sugar nucleotide products 19−23 were then purified by strong anion exchange chromatography and C18 high-performance liquid chromatography (HPLC) to deliver the final, pure materials (see Table 1, as well as Figure S4 in the Supporting Information). The GDP-D-Man pyrophosphorylase showed good substrate promiscuity toward non-native phosphates 9 and 10 and 16− 18, rapidly converting all but 6R-Me derivative 9 and delivered, following purification, milligram quantities of the required sugar nucleotides in good yields (53%−64%). Glycosyl 1phosphate 9 showed a slower conversion by TLC and HRMS
Scheme 2. Synthesis of a C6-Deuterated Mannose 1Phosphates 16−18
strategy and negate the higher costs and availability associated with chiral, deuterated reducing agents. Reduction of 11 with NaBD4 to doubly C6-deuterated pyranose 12 proceeded in good yield (81%), furnishing material for a multigram scale oxidation, which was completed using the Parikh−Doering method16 to give deuteroaldehyde 13 in 92% crude yield. With 13 in hand, we completed independent diastereoselective reductions using either (R)(+)-Alpine-Borane or (S)-(−)-Alpine-Borane to access the respective 6R and 6S deuterated thioglycosides 14 and 15. To confirm the diastereomeric identity of these alcohols, we synthesized small amounts of their 1,6-anhydro derivatives, which enabled nOe experiments to establish their 6-position configuration (Figure 3).
Table 1. Enzymatic Synthesisa of C6-Modified GDP-D-Man Derivatives 19−23
Figure 3. Synthesis of 1,6-anhydro derivatives from 14 and 15 to facilitate stereochemical assignment at C6 (red/blue arrows represent the observed nOe; green arrow represents the direction of view for Newman projection).
For the H6 endo compound, (derived from 15), we observed an nOe from H6 to H4 (and H5) along with transfer from H4 to H 6 (see Figure S3 in the Supporting Information). Comparatively for H6 exo (derived from 14), we observed only transfer from H6 to H5. From the one-dimensional (1D) 1 H NMR data, a very small 3J H6−5 coupling of 0.8 Hz for the H6 endo inferred that the dihedral angle between these two protons was approaching 90°, further supporting our assignment and matching 1D NMR data reported previously by Meguro.17 2D-HSQC data showed the expected correlation between H6 and C6 for both diastereoisomers, with a 1J
a
See Figure S4 in the Supporting Information for the general enzymatic procedure. bSynthesized chemically.
C
DOI: 10.1021/acs.orglett.9b00967 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters
a reduced ability to produce NADH, compared to GDP-Dmannose 1. Although 19 and 20 can theoretically only produce 1 equiv of NADH (compared to 2 equiv for 1), they fall short of achieving that level of conversion. This is consistent with the expected keto-sugar product serving as an inhibitor of GMD activity. For the reaction of GMD with 20, we obtained direct evidence of the ketone oxidation product via high-resolution mass spectrometry (HRMS). While this finding is subject to further evaluation, it is evident that the ketone is not tightly bound: spiking 60 min incubations that contained GMD and 19 and 20 with GDP-Dmannose 1 led to renewed NADH production (Figure 4b). C6-deuterated analogues 21−23 were processed by the enzyme to a comparable level to 1, as expected: further kinetic evaluation of these materials is underway. In conclusion, we have developed a first series of sugar nucleotide probes to target the GDP-D-Man dehydrogenase from P. aeruginosa, using a chemoenzymatic approach. Strategic chemical modifications at C6 of the pyranose ring (deuteration and one-carbon homologation) followed by enzymatic pyrophosphorylation were effected. Access to these materials will enable a more-detailed mechanistic investigation of GMD and other important biochemical enzymes that utilize GDP-D-Man, including nucleotidylyltransferases, glycosyltransferases, and phosphorylases.
analyses (see Figures S4 and S6 in the Supporting Information), even after several small-scale repeats and a prolonged 38 h reaction time. This analogue was instead synthesized chemically, using GMP-morpholidate, to ensure that sufficient material could be isolated to characterize and complete the series. The substrate scope for the phosphorylase suggests a relaxed specificity at C6 of mannose, possibly inferring that it is not involved in binding to the enzyme. This is consistent with observations made by Lowary and coworkers,18 using the same pyrophosphorylase for the synthesis of 6-deoxy and 6-OMe GDP-D-Mans. Sugar nucleotides 19−23 were incubated with GMD from P. aeruginosa, and the colorimetric release of NADH was monitored at 460 nm. Figure 4a illustrates the relative activity of each of probes 19−23 over the first 5 min of incubation. As expected, the two C6-homologated species, 19 and 20, showed
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.9b00967. Experimental procedures, characterization data, and 1H, 13 C, and 31P NMR spectra for all new compounds, alongside enzymatic methods (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Gavin J. Miller: 0000-0001-6533-3306 Author Contributions
All authors have given approval to the final version of the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The Engineering and Physical Research Council (EPSRC) are thanked for project grant funding (No. EP/P000762/1) to G.J.M. at Keele. We also thank the EPSRC UK National Mass Spectrometry Facility (NMSF) at Swansea University, Dr. Richard Darton (Keele) for running nOe spectra, Prof. Peter Tipton (University of Missouri) for providing the GMD plasmid, and Professor Todd Lowary (University of Alberta) for providing the GDP-mannose pyrophosphorylase clone. Work at the JIC is supported by the Biotechnology and Biological Science Research Council (BBSRC) Institute Strategic Program on Molecules from NatureProducts and Pathways (No. BBS/E/J/000PR9790) and the InnovateUK IBCatalyst [Nos. BB/M02903411 and EP/N033167/10].
Figure 4. (a) Initial rates of NADH production over 5 min for sugar nucleotides 19−23. GMD (50 μg/mL), GDP sugars (50 μM), NAD+ (200 μM), the activity is a mean ± standard error of three repeats. A negative control experiment was run with no GMD. Data are not normalized for a maximum of 1 equiv NADH produced by 19 and 20. (b) GMD function with probes 19 and 20 over 120 min, 50 μM spiking with 1 at 60 min. D
DOI: 10.1021/acs.orglett.9b00967 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters
■
REFERENCES
(1) Franklin, M. J.; Nivens, D. E.; Weadge, J. T.; Howell, P. L. Front. Microbiol. 2011, 2, 167. (2) Tatnell, P. J.; Russell, N. J.; Gacesa, P. Microbiology 1994, 140, 1745. (3) Low, K. E.; Howell, P. L. Curr. Opin. Struct. Biol. 2018, 53, 32. (4) Roychoudhury, S.; Chakrabarty, K.; Ho, Y. K. J. Biol. Chem. 1992, 267, 990. (5) Snook, C. F.; Tipton, P. A.; Beamer, L. Biochemistry 2003, 42, 4658. (6) Campbell, R. E.; Tanner, M. E. J. Org. Chem. 1999, 64, 9487. (7) Campbell, R. E.; Tanner, M. E. Angew. Chem., Int. Ed. Engl. 1997, 36, 1520. (8) Errey, J. C.; Mann, M. C.; Fairhurst, S. A.; Hill, L.; McNeil, M. R.; Naismith, J. H.; Percy, J. M.; Whitfield, C.; Field, R. A. Org. Biomol. Chem. 2009, 7, 1009. (9) Poulin, M. B.; Zhou, R.; Lowary, T. L. Org. Biomol. Chem. 2012, 10, 4074. (10) Ahmadipour, S.; Miller, G. J. Carbohydr. Res. 2017, 451, 95. (11) Ahmadipour, S.; Beswick, L.; Miller, G. Carbohydr. Res. 2018, 469, 38. (12) Read, J. A.; Ahmed, R. A.; Tanner, M. E. Org. Lett. 2005, 7, 2457. (13) Liu, H.-J.; Shia, K.-S.; Shang, X.; Zhu, B.-Y. Tetrahedron 1999, 55, 3803. (14) Doores, K. J.; Fulton, Z.; Hong, V.; Patel, M. K.; Scanlan, C. N.; Wormald, M. R.; Finn, M. G.; Burton, D. R.; Wilson, I. A.; Davis, B. G. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 17107. (15) van den Bos, L. J.; Dinkelaar, J.; Overkleeft, H. S.; van der Marel, G. A. J. Am. Chem. Soc. 2006, 128, 13066. (16) Inuki, S.; Aiba, T.; Kawakami, S.; Akiyama, T.; Inoue, J.-I.; Fujimoto, Y. Org. Lett. 2017, 19, 3079. (17) Hori, H.; Nishida, Y.; Ohrui, H.; Meguro, H. J. J. Carbohydr. Chem. 1990, 9, 601. (18) Zou, L.; Zheng, R. B.; Lowary, T. L. Beilstein J. Org. Chem. 2012, 8, 1219.
E
DOI: 10.1021/acs.orglett.9b00967 Org. Lett. XXXX, XXX, XXX−XXX