Letter pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
A Solution to Chemical Pseudaminylation via a Bimodal Glycosyl Donor for Highly Stereocontrolled α- and β‑Glycosylation Ruohan Wei,†,∥ Han Liu,†,§,∥ Arthur H. Tang,‡ Richard J. Payne,‡ and Xuechen Li*,† †
Department of Chemistry, State Key Laboratory of Synthetic Chemistry, The University of Hong Kong, Hong Kong, P. R. China School of Chemistry, The University of Sydney, Sydney, Australia
‡
Org. Lett. Downloaded from pubs.acs.org by BETHEL UNIV on 05/02/19. For personal use only.
S Supporting Information *
ABSTRACT: A robust methodology for the stereocontrolled chemical glycosylation of pseudaminic acid has been developed to afford both α- (axial) and β- (equatorial) glycosides reliably with complete stereoselectivity, using a common glycosyl donor (7NCbz/5N-azido Pse thioglycoside) simply by changing the reaction conditions. In the CH2Cl2/MeCN cosolvent, highly β-selective pseudaminylation was observed, while addition of 5.0 equiv DMF in CH2Cl2 gave the α-pseudaminosides.
seudaminic acid (Pse), first discovered by Knirel et al. in 1984,1 has been identified in a variety of pathogenic bacteria as important surface glycan components.2 Belonging to the nonulosonic acid carbohydrate family, Pse is structurally related to its well-known congener, sialic acid. In native glycoconjugates, Pse with variable substitution patterns at N5 and N7 forms both axial (α)3 and equatorial (β)4 glycosidic linkages to different pyranoses/furanoses and the hydroxyl side chain of Ser/Thr residues5 (Figure 1). Such structural diversity makes the structure−function relationship of Pse-containing glycans both intriguing and challenging. Due to the lack of synthetic access to structurally defined pseudaminylated glycoconjuates, the biological and evolutional significances of bacterial pseudaminic acid and its glycoconjugates remain largely unexplored. To date, five synthetic routes toward Pse have been reported, and challenges in the preparation of Pse itself have been resolved to a great extent.6−10 However, the synthesis of Pse-containing oligosaccharides remains to be explored. Recently we reported the de novo synthesis of a Pse thioglycoside donor that was subsequently employed in the total synthesis of Pseudomonas aeruginosa 1244 pilin trisaccharide (α-5NβOHC47NFmPse-(2 → 4)-β-Xyl-(1 → 3)-FucNAc), in which highly α-selective glycosylation was obtained.9 In order to expand the synthetic application of the Pse donor, it is necessary to investigate the generality of the stereoselectivity initially observed through the glycosylation of diverse acceptors. More importantly, a robust method to access both α- and β-selectivity should also be investigated through either modification of the structure of the Pse donor or through variation of the glycosylation conditions such that both native anomeric configurations can be accessed. Contemporaneous with our studies, Crich et al. reported the β-selective glycosylation of the 5N/7N-bisazido Pse thioglycoside donor with limited acceptors tested.10 This Pse donor was synthesized from thiosialoside in 20 steps. Herein, we report
P
© XXXX American Chemical Society
Figure 1. Examples of α- and β-glycosides of pseudaminic acid discovered in bacterial glycoconjugates.
the development of a robust methodology for completely stereocontrolled pseudaminylation with a bimodal glycosyl donor to generate a range of pseudaminylated constructs, Received: March 20, 2019
A
DOI: 10.1021/acs.orglett.9b00990 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters
activating reagents (TolSCl/AgOTf) erodes this outcome, which is consistent with the observation in sialylation. Slightly improved β-selectivity was observed in DCM/MeCN, but with much inferior yield accompanied by glycal formation. To evaluate the generality of the observed selectivity, donors 2 and 3 were further tested. Acceptors 6−10, with varied reactivities, were selected based on the reported structures of Pse-containing oligosaccharides and glycopeptides, while acceptor 11 was chosen for its potential use in glycoconjugate preparation (Figure 3). As summarized in Table 1, good yields
which covers almost all the reported pseudaminylated glycosidic linkages identified in bacteria so far. In addition, our Pse donor, readily prepared in gram scale from de novo synthesis, carries orthogonal protecting groups at 5N/7N positions, making it ready to introduce different acyl groups, respectively. In 2017, our group reported the synthesis of 5N-Troc/7NCbz Pse donor 1 from L-allo-threonine in 16 steps and found the C-4 hydroxyl group of xylosides could be glycosylated with α-selectivity.9 However, the selectivity was found to be acceptor-dependent, and complete β-anomer was obtained in glycosylation with reactive acceptors such as benzyl alcohol. To develop both α- and β-selective glycosylation of the 5NTroc/7N-Cbz donor with a broader range of acceptors, we started by modifying the thiol aglycone as the most accessible position for tuning. Based on the widely accepted mechanism of chemical sialylation,11 the activation rate of the SR group could be tuned by varying the activation temperature, causing the change of absolute/relative concentration of reactive intermediates, thus the nucleophilic attack pathway would be varied to affect the selectivity of glycosylation.12,13 To test this hypothesis, two new 5N-Troc donors 2 (SAd) and 3 (SEt) were synthesized and submitted to the glycosylation with model acceptor 4. The representative results are summarized in Figure 2.
Figure 3. Acceptors selected for the pseudaminylation study.
Table 1. Pseudaminylation of 5N-Troc Donors with Different Acceptorsa
Figure 2. Effect of SR group on the selectivity of pseudaminylation.
entry
acceptor
donor
product
yield (%)b
α/βc
1
6
12a/b
2
7
3
8
4
9
5
10
6
11
2 3 2 3 2 3 2 3 2 3 2 3
95 93 88 96 97 92 83 85 75 88 66 89
20:1 1:6.0 2.9:1
13a/b 14a/b 15a/b 16a/b 17a/b
a
The reactions were conducted on 0.05 mmol scale in 1.0 mL of DCM using 1.0 equiv of donor and 2.0 equiv of acceptor at −78 °C (donor 2) or −40 °C (donor 3). bIsolated yields. cDetermined by 1H NMR of the mixture of anomers.
In the case of donor 1, when the glycosylation was conducted in DCM at −40 °C using NIS/TfOH as the activator, both the α and β anomers 5a/b were obtained in a 1.5:1 ratio, while the activation failed at −78 °C. Using more reactive TolSCl/AgOTf activator resulted in improved αselectivity at both temperatures. Addition of MeCN led to increased formation of the β-anomer as expected, but was not effective in reversing the α-selectivity. In the case of donor 2, similar α-selectivity was observed in DCM at −40 °C. The relatively armed donor 2 was readily activated by NIS/TfOH at −78 °C14 and the β-anomer was formed dominantly, while this β-selectivity was reversed when TolSCl/AgOTf activator was used. Interestingly, although not as reactive as donor 2, reactions with donor 3 afforded excellent α-selectivity at −40 °C even in the presence of MeCN. When activated by TolSCl/ AgOTf, the similar α-selectivity was achieved. Based on current observations, it is clear that the low reaction temperature facilitated by the highly reactive SAd donor contributes to the β-selectivity, while accelerating the activation process by using more reactive donor (2 or 3) together with more powerful
were obtained in most glycosylations. In general, SEt donor 3 gave moderate to good α-selectivity, while SAd donor 2 favored the formation of β-anomers. In the case of acceptor 9, the disaccharides 15a/b were obtained in a 1:1 ratio when donor 2 was used. To our surprise, both donors 2 and 3 favored formation of the α-anomer in the glycosylation of serine derived acceptor 10. These results clearly show the difference in the anomeric preference of donors 2 and 3, but with selectivity still strongly dependent on the nature of the acceptors. To further improve the stereocontrolled pseudaminylation, we shifted our attention to modify the 5N substituents. We hypothesized that our failure to achieve high βselectivity using 5N-Troc donor 2 could be attributed to the intramolecular remote participation of the carbamate group B
DOI: 10.1021/acs.orglett.9b00990 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters from the β-face of the oxocarbenium to induce the αglycosylation.15 In addition, we noted that Crich reported the stereoselective glycosylation of 5-amino-non-2-ulosonic acid species, in which the 5N-azido group contributed to the high equatorial selectivity.14b,c Thus, we modified the 5N-Troc substituent of donor 2 into the nonparticipating azide group (Scheme 1). Specifically, formamide 18, obtained from L-allo-
5.0 equiv of DMF, donor 22 gave complete α-selectivity in the glycosylation of acceptor 4. The use of DMF was first introduced by Mong et al. in 2011 as a strongly axial-selective glycosylation additive18 and recently applied by Codée et al. in the synthesis of a complex α-glucan,19 while the strategy has not been applied in ulosonic acid type glycosylation. As summarized in Table 3, glycosylation of 22 with acceptors 6−
Scheme 1. Synthesis of the 5N-Azido Pse Donora
Table 3. α-Selective Pseudaminylation of 5N-Troc and 5NAzido Donors Using DMF Additivea
a (a) 3% HCl (aq) in MeOH, 0 °C to rt, 8 h; (b) TfN3, 0.5 M Na2CO3, MeCN, 2 h; (c) O3, DCM, −78 °C, 0.5 h; then Me2S, 45% over 3 steps; (d) Ac2O, Py, DMAP, 0 °C to rt, 95%; (e) AdSH, BF3· OEt2, DCM, 16 h, 67%.
entry
acceptor
donor
product
yield (%)b
α/βc
1
4
2
6
3
7
4
8
5
9
6
10
7
11
2 22 2 22 2 22 2 22 2 22 2 22 2 22
5a 23a 12a 24a 13a 25a 14a 26a 15a 27a 16a 28a 17a 29a
75 77 80 76 79 75 75 73 56d,e 55d 75d 77d 78 80
>20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1 >20:1
9
Thr, was converted to azido imtermediate 19, which was directly submitted to ozonolysis to give 20 (45% over 3 steps) to avoid the intramolecular 1,3-dipolar cycloaddition.16 After acetylation and thioglycosylation, donor 22 was obtained as a mixture of α/β-anomers (1:4−1:5 favoring the β-anomer). The results from the glycosylation of donor 22 with various acceptors are summarized in Table 2. To our delight, high βselectivity was achieved in most cases regardless of the reactivity of the acceptor.
a
The reactions were conducted on 0.05 mmol scale in 1.0 mL DCM at −40 °C in the presence of 5.0 equiv of DMF, using 1.0 equiv of donor and 2.0 equiv of acceptor. bIsolated yields. cDetermined by 1H NMR of the mixture of anomers. dPreactivation procedure was used. e The α-selective glycosylation at the 4-OH of the acceptor was isolated as the side product in 12% yield.
Table 2. β-Selective Pseudaminylation of 5N-Azido Donor with Different Acceptorsa
entry
acceptor
product
yield (%)b
α/βc
1 2 3 4 5 6 7
4 6 7 8 9 10 11
23a/b 24b 25b 26b 27b 28b 29b
78 72 82 88 75 80 77
1:10
