Challenges in the Conversion of Manual Processes to Machine-Assisted Syntheses: Activation of Thioglycoside Donors with Aryl(trifluoroethyl)iodonium Triflimide

Article abstract of DOI:10.1021/acs.orglett.7b03940

The steps needed to adapt a stable iodonium promoter for use in automated fluorous-assisted solution-phase oligosaccharide synthesis are described. Direct adaptation of the originally reported batch procedure resulted in the formation of an orthoester or

Full text of DOI:10.1021/acs.orglett.7b03940

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
Challenges in the Conversion of Manual Processes to Machine-
Assisted Syntheses: Activation of Thioglycoside Donors with
Aryl(trifluoroethyl)iodonium Triflimide
Regis C. Saliba,† Zachary J. Wooke,† Gabriel A. Nieves,† An-Hsiang Adam Chu,‡,∥ Clay S. Bennett,*,‡
and Nicola L. B. Pohl*,†,§
†Department of Chemistry, Indiana University, 212 South Hawthorne Drive, Bloomington, Indiana 47405, United States
‡Department of Chemistry, Tufts University, 62 Talbot Avenue, Medford, Massachusetts 02155, United States
§Radcliffe Institute of Advanced Study, Harvard University, 8 Garden Street, Cambridge, Massachusetts 02318, United States
S
* Supporting Information
ABSTRACT: The steps needed to adapt a stable iodonium
promoter for use in automated fluorous-assisted solution-phase
oligosaccharide synthesis are described. Direct adaptation of the
originally reported batch procedure resulted in the formation of
an orthoester or protecting group transfer to the glycosyl
acceptor. Fortunately, the addition of inexpensive β-pinene as
an acid scavenger avoided both of these side reactions. The
utility of this newly developed protocol was applied to the
automated solution-phase synthesis of a β-glucan fragment.
ith increasing interest in reproducibility and the
automation of specialized chemical production comes
iodonium species7 to carry out a glycosylation reaction, and
we show its use in an iterative multistep automated synthesis of
a glucan fragment. The development of these processes
highlights general strategies for translating traditional manual
benchtop processes to machine-assisted procedures.
W
the need for reactions compatible with the liquid handling
platforms used to carry out machine-assisted solution-phase
and solid-phase syntheses, including flow chemistry.1 Unfortu-
nately, the vast majority of reported chemical reactions are not
homogeneous from start to finish, but rely on principles such as
precipitation of components to drive reactions forward or
gradual dissolution of components to limit solvent usage. These
issues are becoming especially problematic in the recent push
to automate oligosaccharide synthesis. Despite over a decade of
work to that endcomprising automated solid-supported
synthesis on a commercial synthesizer (Glyconeer 2.1),2
HPLC-assisted solid-supported automated synthesis,3 auto-
mated glycolipid-assisted solution-phase synthesis,4 automated
fluorous-assisted solution-phase synthesis,5 and automated
iterative electrochemical solution-phase synthesis6only a
small subset of glycosylation and deprotection procedures
have been translated from manual synthetic procedures to
machine-assisted processes to create robustly reproducible
protocols. Regardless of whether the reactions occur in solution
or upon polymeric supports, robotic handling platforms at
present are still most reliable when dispensing liquids and
solutions rather than solids and precipitates. Therefore, many
traditional glycosylation methods that rely on the formation of
solids such as silver salts can generally only be performed as
manual batch processes that are not readily adapted to iterative
machine-assisted protocols. Herein we report the first example
of a machine-assisted protocol that uses a hypervalent
For a protocol to easily translate to automated synthesis
platforms, all reagents and reaction mixtures need to be soluble
in a solvent amenable to the desired reaction in order to be
transferred via a liquid handling system. Although this
constraint is also true of continuous flow systems, all current
iterative automation platforms known for biopolymer synthesis
are based on batch reactors and therefore precipitates can form
during reaction as long as they can be solubilized later in an
appropriate solvent during the purification step, whether that
be rinsing of a solid-phase or loading of a fluorous solid-phase
extraction cartridge. Ideally, any machine-assisted glycosylation
would provide the desired product with at least a 2:1
stereoselectivity and a 70% yield to allow ease of product
identification and purification using recycling HPLC
(RHPLC)8 or other methods. In addition, reactions should
be conducted in the temperature range most readily accessed in
automation platforms (−45 to 250 °C). Finally, all reagents
should ideally be stable in solution at room temperature under
argon for the length of the automation cycle in order to
minimize required interventions. This latter requirement has
Received: December 18, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.7b03940
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
started turning our focus away from the more reactive
trichloroacetimidate donors5 toward thioglycosides.9 To this
end, hypervalent iodonium salts 1A7 and 1B,7b which are stable
to ambient air and moisture, look promising for carrying out
glycosylations at ambient temperatures (Figure 1).
Figure 1. Hypervalent iodonium promoters.
Given the general importance of reagent and reactant
solubility in machine-assisted protocols, data on the solubility
of aryl(trifluoroethyl)iodonium triflimide 1A in dichloro-
methane and dichloroethane, the most commonly used solvents
in glycosylation, were needed first. Saturation concentrations
were determined by saturating solvent with 1A, evaporating an
aliquot, reconstituting in acetonitrile, and measuring UV
absorbance against a calibration curve. UV absorbance was
measured using an Agilent 1100 HPLC by injecting the samples
directly from the autosampler into the diode array detector.10
As expected, this promoter was soluble in both solvents but at a
low concentration (<50 mM, Table 1, entries 1 and 2). Since
Table 1. Solubility of Aryl(trifluoroethyl)iodonium
Triflimide 1A in Different Solvents at 23 °C
a
a
entry
solvent
CH2Cl2
saturation (mM) saturation (mg·mL−1
)
1
2
3
4
44
46
2
2
6
1
25
26
1
1
3
1
ClCH2CH2Cl (DCE)
1:20, CH3CN/DCE
1:10, CH3CN/DCE
246
582
140
330
a
Errors bars represent duplicate measurements.
this promoter has also been used in nitrile solvents to promote
β-stereoselectivity in the glycosylation of donors possessing
benzyl ethers at the 2-position,7b the solubility of the promoter
in various 1,2-dichloroethane/CH3CN mixtures was also
investigated without the use of added heat. Addition of the
nitrile cosolvent proved to increase the solubility of the
promoter by 5- to 12-fold (Table 1, entries 3 and 4). From
these data, a 1:10 mixture of acetonitrile in 1,2-dichloroethane
(DCE) was chosen as the reaction solvent for machine-assisted
synthesis trials.
Acetonitrile is known to favor β-selectivity in glycosylation
via a nitrilium intermediate, potentially competing with the
attachment of the acceptor to the activated donor, thus
decreasing the rate of the reaction.11 To investigate the effect of
the acetonitrile solvent, a glycosylation reaction between
acceptor 2 and donor 3A was conducted on a solution-phase
automated synthesis platform (Figure 2). The synthesizer,
consisting of a robotic arm coupled to two syringe pumps,
transfers reagents in solution to an array of double-jacketed
reactors under argon that can be vortexed and either heated or
cooled (see Supporting Information). Small aliquots were
picked at different time points from the reactor, transferred to a
sample rack, and quenched in order to be further analyzed by
HPLC. This experiment proved that the small amount of
acetonitrile did not impact the apparent reaction rate
Figure 2. Optimization of the reaction conditions. (1) Base = 7 (4
equiv), [2] = 113 mM; (2) base = 7 (4 equiv), [2] = 50 mM; (3) base
= 7 (4 equiv), [2] = 15 mM; (4) base = 8 (4 equiv), [2] = 50 mM; (5)
base = 8 (4 equiv), [2] = 15 mM; (6) base = 8 (4 equiv), [2] = 5 mM;
(7) base = 9 (4 equiv), [2] = 50 mM; (8) base = 10 (4 equiv), [2] =
15 mM; (9) base = 10 (10 equiv), [2] = 15 mM; (10) base = 10 (20
equiv), [2] = 15 mM; (11) no base, [2] = 50 mM; (12) no base, [2] =
15 mM.
significantly. Total disappearance of the limiting acceptor was
observed by HPLC after only 30 min. However, to our surprise
two products of the same molecular weight but different
retention times were formed (see Supporting Information).
Further isolation identified the second more apolar product as
the orthoester 5 (Figure 2).
Because orthoester 5 was the major side product of the
reaction with the addition of the more polar solvent, other
reaction parameters were investigated (Figure 2). Switching the
tri-tert-butylpyrimidine 9 (TTBP) to a base with a higher pKa
such as 2,6-di-tert-butylpyridine 7 (DTBP) or 2,6-di-tert-butyl-
4-methyl-pyridine 8 (DTBMP)12 did slightly improve the
formation of the desired product. Diluting the reaction also
slightly improved the conversion to the desired product 4
(Figure 2, columns 1−7). However, elimination of the base
B
DOI: 10.1021/acs.orglett.7b03940
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
entirely only resulted in the transfer of an acetate protecting
group to the free hydroxyl of the acceptor (Figure 2, columns
11 and 12, compound 6). Given the evidence for these
hindered bases being potentially nucleophilic,13 strained olefins
as nonbasic proton scavengers14 were explored as an alternative
to amine bases. To this end, β-pinene 10, of comparable cost to
the solvent DCE, was tested for its ability to hinder formation
of the orthoester as well as the transfer of the protecting group
in the glycosylation reaction. Interestingly, 4 equiv of this
nonbasic proton scavenger did inhibit the formation of the
orthoester completely, but the acetate transfer only partially
(Figure 2, column 8). However, increasing the amount of β-
pinene proved to inhibit both side reactions (Figure 2, columns
9 and 10). Therefore, all subsequent reactions were run at 15
mM in the presence of 20 equiv of the β-pinene.
In theory at least, the reaction conditions should now be
amenable to robotic liquid handling systems required by solid-
phase- and solution-phase-based automated synthesis plat-
forms. To probe this glycosylation reaction in practice and to
test the compatibility of the new conditions with common
deprotection steps without the need for purification between
these two reaction types, several donors with different
protecting groups at their 6-positions that are commonly
used in carbohydrate chemistry were procured. Each iterative
cycle of oligosaccharide synthesis is commonly composed of a
glycosylation step, followed by the selective deprotection of the
next position to be glycosylated, and finally purification of the
newly formed acceptor intermediate by some means such as, in
our case, fluorous solid-phase extraction (FSPE, Table 2).
(Table 2). In each case, compatibility between the glycosylation
and the deprotection steps5b,15 was observed in the automation
cycles, giving 11 as a major product after FSPE purification as
seen by HPLC analysis. Purification by HPLC allowed us to
obtain the disaccharide 11 in 26% to 72% unoptimized yield
depending on the protecting group and the glycosylation
conditions used (Table 2). For most automated syntheses,
optimization of yields is less important than the ability of a
general set of conditions to provide sufficient material for
further study in the first attempt.
Having established that common deprotection conditions
could be directly carried out after glycosylation with the
hypervalent iodonium salt 1A and β-pinene 10 without the use
of any intermediate purification, we applied this new automated
strategy for the formation of a larger oligosaccharide. Starting
with acceptor 2, three iterative cycles were implemented on the
automation platforms using only 2 equiv of donor 3B in each
glycosylation step and the conditions for the levulinoyl cleavage
(Figure 3). After the last FSPE and further reversed-phase
purification, the desired β-1,6-glucan tetrasaccharide 12 was
obtained in 16% over 6 steps (74% average per reaction step).
Table 2. Results for the Glycosylation/Deprotection/FSPE
Cycle with Donors 3A−D
Figure 3. Automated solution-phase synthesis of β-1,6-glucan
tetramer.
In conclusion, we have demonstrated that the air- and water-
stable iodinated promoter 1A can be translated for use in an
automation protocol using robotic liquid handling to activate
bench stable thioglycoside donors. In addition, the new
conditions used here for glycosylation have been shown to be
compatible with common deprotection conditions without the
need for intermediate purification. The use of iterative
glycosylation/6-O-Lev-deprotection/FSPE cycles was applied
to achieve the formation of a β-1,6-glucan tetramer. This
strategy extends the possible activation chemistries that can be
implemented in the known automated oligosaccharide
protocols to a variety of thioglycoside donors and glycoside
acceptors to potentially access a wider range of structures. Key
to the successful adaption of the manual protocol to
automation was careful reaction monitoring of undesired
glycosylation side products and the use of a strained olefin
entry
donor
proton scavenger
deprotection condition
yield (%)
1
2
3
4
3B
3B
3C
3D
8
1
1
2
3
26
72
58
47
10
10
10
Purification between the glycosylation and the deprotection
steps is usually not required for successful removal of most
masking groups and therefore is avoided to save time. However,
these new glycosylation conditions produce a new set of
reagents and byproducts that could potentially inhibit the usual
deprotection programs.5b A levulinate ester (Lev), chloroace-
tate ester (ClAc), and a tert-butyldimethylsilyl ether (TBDMS)
were chosen for testing due to their ability to potentially be
removed in the presence of benzoyl esters and benzyl ethers
C
DOI: 10.1021/acs.orglett.7b03940
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(7) (a) Chu, A. H.; Minciunescu, A.; Montanari, V.; Kumar, K.;
Bennett, C. S. Org. Lett. 2014, 16, 1780−1782. (b) Chu, A. H.;
Minciunescu, A.; Bennett, C. S. Org. Lett. 2015, 17, 6262−6265.
(8) Nagy, G.; Peng, T.; Pohl, N. L. B. Anal. Methods 2017, 9, 3579−
3593.
(9) (a) Goswami, M.; Ellern, A.; Pohl, N. L. Angew. Chem., Int. Ed.
2013, 52, 8441−8445. (b) Goswami, M.; Ashley, D. C.; Baik, M. H.;
Pohl, N. L. J. Org. Chem. 2016, 81, 5949−5962. (c) Kabotso, D. E. K.;
Pohl, N. L. B. Org. Lett. 2017, 19, 4516−4519.
(10) Pan, L.; Ho, Q.; Tsutsui, K.; Takahashi, L. J. Pharm. Sci. 2001,
90, 521−529.
(11) Lahmann, M.; Oscarson, S. Org. Lett. 2000, 2, 3881−3882.
(12) Crich, D.; Smith, M.; Yao, Q. J.; Picione, J. Synthesis 2001, 2001,
323−326.
(13) Lichtenthaler, M. R.; Stahl, F.; Kratzert, D.; Benkmil, B.;
Wegner, H. A.; Krossing, I. Eur. J. Inorg. Chem. 2014, 2014, 4335−
4341.
(14) (a) Gu, X.; Chen, L.; Wang, X.; Liu, X.; You, Q.; Xi, W.; Gao, L.;
Chen, G.; Chen, Y. L.; Xiong, B.; Shen, J. J. Org. Chem. 2014, 79,
1100−1110. (b) Chen, G.; Yin, Q.; Yin, J.; Gu, X.; Liu, X.; You, Q.;
Chen, Y. L.; Xiong, B.; Shen, J. Org. Biomol. Chem. 2014, 12, 9781−
9785. (c) Lloyd, D.; Bylsma, M.; Bright, D. K.; Chen, X. Z.; Bennett,
C. S. J. Org. Chem. 2017, 82, 3926−3934.
(15) (a) Schworer, R.; Zubkova, O. V.; Turnbull, J. E.; Tyler, P. C.
Chem. - Eur. J. 2013, 19, 6817−6823. (b) Yang, W.; Yoshida, K.; Yang,
B.; Huang, X. Carbohydr. Res. 2016, 435, 180−194.
acid scavenger, a process that holds lessons for future attempts
to develop new chemistries for automated synthesis platforms.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.7b03940.
Automated solution-phase synthesis protocols, synthetic
procedures, and spectral data (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: npohl@indiana.edu.
*E-mail: clay.bennett@tufts.edu.
ORCID
Regis C. Saliba: 0000-0003-0090-204X
Clay S. Bennett: 0000-0001-8070-4988
Nicola L. B. Pohl: 0000-0001-7747-8983
Present Address
∥Pfizer, 1 Burtt Road, Andover, MA 01810, United States
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
N.P. and C.B. both acknowledge support from the U.S.
National Institutes of Health Common Fund Glycosciences
Program (1U01 GM116248 and 1U01GM120414 respectively)
to spur this collaboration. N.P. also thanks the Joan and Marvin
Carmack Fund at Indiana University and Radcliffe Institute for
Advanced Study for the Edward, Frances, and Shirley B.
Daniels Fellow position.
REFERENCES
■
(1) (a) Ley, S. V.; Fitzpatrick, D. E.; Myers, R. M.; Battilocchio, C.;
Ingham, R. J. Angew. Chem., Int. Ed. 2015, 54, 10122−10136.
(b) Selekman, J. A.; Qiu, J.; Tran, K.; Stevens, J.; Rosso, V.; Simmons,
E.; Xiao, Y.; Janey, J. Annu. Rev. Chem. Biomol. Eng. 2017, 8, 525−547.
(2) (a) Plante, O. J.; Palmacci, E. R.; Seeberger, P. H. Science 2001,
291, 1523−1527. (b) Plante, O. J.; Palmacci, E. R.; Seeberger, P. H.
Adv. Carbohydr. Chem. Biochem. 2003, 58, 35−54. (c) Hahm, H. S.;
Schlegel, M. K.; Hurevich, M.; Eller, S.; Schuhmacher, F.; Hofmann, J.;
Pagel, K.; Seeberger, P. H. Proc. Natl. Acad. Sci. U. S. A. 2017, 114,
E3385−E3389. (d) Geert Volbeda, A.; van Mechelen, J.;
Meeuwenoord, N.; Overkleeft, H. S.; van der Marel, G. A.; Codee,
J. D. C. J. Org. Chem. 2017, 82, 12992−13002. (e) Komba, S.;
Terauchi, T.; Machida, S. J. Appl. Glycosci. 2009, 56, 193−206.
(3) (a) Ganesh, N. V.; Fujikawa, K.; Tan, Y. H.; Stine, K. J.;
Demchenko, A. V. Org. Lett. 2012, 14, 3036−3039. (b) Pistorio, S. G.;
Nigudkar, S. S.; Stine, K. J.; Demchenko, A. V. J. Org. Chem. 2016, 81,
8796−8805.
(4) Bauer, J.; Brandenburg, K.; Zahringer, U.; Rademann, J. Chem. -
Eur. J. 2006, 12, 7116−7124.
(5) (a) Tang, S. L.; Linz, L. B.; Bonning, B. C.; Pohl, N. L. J. Org.
Chem. 2015, 80, 10482−10489. (b) Tang, S. L.; Pohl, N. L. Org. Lett.
2015, 17, 2642−2645. (c) Tang, S. L.; Pohl, N. L. B. Carbohydr. Res.
2016, 430, 8−15.
(6) (a) Nokami, T.; Hayashi, R.; Saigusa, Y.; Shimizu, A.; Liu, C. Y.;
Mong, K. K.; Yoshida, J. Org. Lett. 2013, 15, 4520−4523. (b) Nokami,
T.; Isoda, Y.; Sasaki, N.; Takaiso, A.; Hayase, S.; Itoh, T.; Hayashi, R.;
Shimizu, A.; Yoshida, J. Org. Lett. 2015, 17, 1525−1528.
D
DOI: 10.1021/acs.orglett.7b03940
Org. Lett. XXXX, XXX, XXX−XXX