A photoflow reactor for the continuous photoredox-mediated synthesis of C-glycoamino acids and C-glycolipids

Article abstract of DOI:10.1002/anie.201200593

Go with the flow: A simple flow reactor has been developed to accommodate highly absorbing [RuL₃]²⁺ photosensitizers for light-starved photoredox reactions (see picture). The use of vessels having a thin diameter increases the efficiency of the reaction. This methodology has been applied to the divergent synthesis of C-glycoconjugates. Copyright

Full text of DOI:10.1002/anie.201200593

.
Angewandte
Communications
DOI: 10.1002/anie.201200593
Photochemistry
A Photoflow Reactor for the Continuous Photoredox-Mediated
Synthesis of C-Glycoamino Acids and C-Glycolipids**
R. Stephen Andrews, Jennifer J. Becker, and Michel R. Gagnꢀ*
Glycoconjugates are essential biological compounds having
various functions. For example, glycoproteins are involved in
intercellular recognition events, such as immune response,
and along with glycolipids they help to form mammalian cell
surfaces.[1] These insights combined with their potential in
vaccine therapeutics[2] continue to drive investigations to
characterize the role of glycoconjugates in biological pro-
cesses. A common feature of glycoconjugates is the metabolic
instability of the O-glycosidic linkage,[3] which has stimulated
the development of C-glycoside isosteres[4] to overcome this
hydrolytic instability and thus extend bioavailability. Despite
extensive efforts, improved methods of C-glycoside synthesis
are still in demand.[5] To this end we have reported mild
methods of C-glycoside synthesis through nickel-catalyzed[6,7]
and photoredox processes.[8] The latter approach has received
Scheme 1. Planned synthesis of C-glycoconjugates from common alde-
significant interest as a mild method of generating radicals
using photochemical energy.[9] We report herein a high-
yielding and scalable approach to C-glycopeptides and
C-glycolipids which utilizes a continuous (flow) photoredox
process.[10,11]
hyde intermediate.
vent the disadvantages of previous approaches.[16,17] Aldehyde
1 can also enable the synthesis of C-linked glycosyl lipids by
reduction and acylation. These goals therefore required
a one-step scalable synthesis, and we envisioned that our
photoredox methodology would provide the needed multi-
gram quantities of 1.
Our efforts focused on the light-mediated conjugate
addition of glycosyl radicals into acrolein (Scheme 2), which
is high yielding in the presence of 1 mol% of [Ru(dmb)3]2+
Previously reported syntheses of C-glycopeptides include
the cross-metathesis of chiral alkenes,[12] Ramberg–Bꢀcklund
olefination/hydrogenation seuqences,[13] addition of chiral
carbon nucleophiles to glycosyl electrophiles,[14] and organo-
catalytic amidation of aldehydes.[15] While effective, these
methods require multiple synthetic steps, use expensive or
toxic reagents, harsh reaction conditions (strong acid or base),
or the use of chiral starting materials (chiral auxiliaries or
amino-acid-derived). As a more efficient alternative, we
envisioned that aldehyde 1, accessible in one step from
commercial sources using our recently reported visible-light
photoredox-mediated methodology,[8a] could function as a key
intermediate for the divergent synthesis of C-glycoconjugate
mimics (Scheme 1).
Recent advances in organocatalytic modification of
aldehydes would enable the synthesis of a series of useful
amino acid derivatives from a single intermediate to circum-
Scheme 2. Conversion and TOF based on vessel size.
[*] R. S. Andrews, Prof. M. R. Gagnꢀ
Department of Chemistry
(dmb = 4,4’-dimethyl-2,2’-bipyridine).[6d] Although scale-up
experiments under standard reaction conditions in a 25 mL
Schlenk flask proceeded smoothly, reactions on this scale
required 24 hours of irradiation to reach 85% conversion, for
a net turnover frequency (TOF) of 3.5 hÀ1. A similar reaction
in a 5 mm NMR tube afforded 73% conversion after only one
hour of irradiation, thus corresponding to a TOF of 70 hÀ1.
Throughout our investigations of photoreductant-mediated
reactions we have noted that reaction vessels having a thinner
diameter generally result in faster reaction rates.
University of North Carolina at Chapel Hill
Chapel Hill, NC 27599-3290 (USA)
mgagne@ unc.edu
Dr. J. J. Becker
US Army Research Office, Research Triangle Park, NC 27709 (USA)
[**] J.J.B. thanks the Army Research Office for Staff Research Funding,
and M.R.G. thanks the Department of Energy (DE-FG02-
05ER15630) for their support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201200593.
4140
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 4140 –4143

Angewandte
Chemie
Since the molar extinction coefficient for [RuL3]2+ com-
plexes are high (17000mÀ1 cmÀ1 for [Ru(dmb)3]2+), we con-
sidered the possibility that the reaction was light-starved and
the light source thus fails to irradiate the entire volume of the
solution.[18] A simple analysis of the absorption profile of
a theoretical vessel using the Beer–Lambert law at relevant
concentrations of [Ru(dmb)3]2+ is shown in Figure 1. This
Figure 2. Diagram of the designed photoflow reactor.
modules to be connected in series to increase residence time
without decreasing the flow rate.
By running the reaction depicted in Scheme 2, the
efficiency of the photoflow reactor was examined for different
concentrations of [Ru(dmb)3]2+ using two different tubing
diameters (Figure 3). For tubing having a 1.6 mm interior
diameter and 1.1 mm [Ru(dmb)3]2+ a TOF of 30 hÀ1 was
Figure 1. The percent transmittance versus distance from the wall (d)
2+
*
as calculated from the Beer–Lambert law. 0.5 mm [Ru(dmb)3]
,
2+
2+
~
&
1 mm [Ru(dmb)3]
,
2 mm [Ru(dmb)3] .
analysis indicates that the vast majority of the reaction vessel
receives negligible amounts of light. At a catalyst concen-
tration of 1 mm, 98% of the incident light is absorbed within
1 mm of the vessel wall, while at 2 mm, this occurs within
0.5 mm. Since the light initiates the cascade of events that
leads to free radical generation, the “active volume” is
effectively determined by the surface area of the glass being
irradiated; the remaining volume is not effectively irradiated.
This analysis suggests that decreasing the diameter of the
reaction vessel should increase the flux of photons throughout
the vessel and thereby increase the effective concentration of
active catalyst and thus the rate of the reaction. The data in
Figure 1 suggest that achieving reasonable irradiation of the
entire reaction volume would require a vessel having a sub-
millimeter diameter. Our solution to the problem of obtaining
sufficiently thin diameters without sacrificing reactor volume
was a photoflow reactor, which allows the continuous
irradiation of a reaction mixture as it flows around a light
source.
The basic design principle was initially demonstrated by
Booker-Milburn,[11a] and several other examples of photoflow
reactors for UV-light irradiation have since been
reported.[10,11] As a result of the weak absorptivity of the
reactants in these examples, these reactors are limited only by
the flux of the light source and benefit from tubing with
a thicker diameter and multiple layers of wrapping around the
light source. In contrast, the above analysis suggested the
optimal design would utilize tubing with a thin diameter and
a single wrap around the light source.
Figure 3. TOF versus [Ru(dmb)3]2+ for FEP tubing of two different
inner diameters.
obtained with one module. At the same flow rate of
0.1 mLminÀ1 and an increased [Ru(dmb)3]2+ concentration
(2.2 mm) the rate of conversion was lower (TOF = 17 hÀ1),
whereas a lowered concentration increased the rate (TOF =
50 hÀ1 at 0.5 mm). Consistent with the analysis in Figure 1, the
reaction rate increased upon decreasing the inner diameter of
the FEP tubing, that is, at 1.1 mm [Ru(dmb)32+] and with
tubing having an inner diameter of 0.8 mm the TOF was
72 hÀ1. Again, an inverse relationship between catalyst
concentration and TOF was observed, with 0.5 mm [Ru-
(dmb)3]2+ giving the highest observed TOF (120 hÀ1). These
flow reactions demonstrated higher TOFs than either the
flask or batch reactor. More importantly, the TOFs increase
significantly with decreasing tubing diameter, thus supporting
the prediction provided from the data in Figure 1 of a photon-
starved reaction. Decreasing the catalyst concentrations also
led to an increase in the TOF, which is consistent with
a previously photon-starved reaction which now enjoys an
increase in photon flux throughout the diameter of the
vessel.[20,21]
As shown in Figure 2, our simple design utilizes clear
fluorinated ethylene propylene (FEP)[19] tubing coiled once
around a Liebigs condenser, three strips of blue LEDs placed
inside the condenser, cool water passing through the con-
denser jacket for temperature control, and a flow rate that is
controlled by the pump of a preparatory HPLC. The ends of
the tubing are fitted to Swagelok connectors, which enable the
For our final reactor design, we chose a FEP tubing with
a 1.6 mm inner diameter and a concentration of 1 mm
Angew. Chem. Int. Ed. 2012, 51, 4140 –4143
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4141

.
Angewandte
Communications
[Ru(dmb)3]2+ as these parameters provided the highest
yields.[22] On a 18.2 mmol scale, this flow reactor, using two
modules connected in series, was operated continuously
(24 h) to obtain 4.5 g of 1 in 70% yield (entry 1, Table 1).
The yield of 1 increased to 85% by increasing the concen-
tration of acrolein, that is, 5.5 g of 1 after 24 h of irradiation.
The pivaloate-protected substrate 4 was slower to react, thus
resulting in only 75% conversion using two modules. How-
ever, full conversion of 4 was obtained by attaching a third
module (entries 3 and 4, Table 1).
for the hydrolysis (65% H2SO4) or alcoholysis (HCl in
MeOH) of the nitrile failed to produce the desired product,
hydration with Parkinꢁs catalyst (9) provided the amide
intermediate with no observable epimerization.[23] Alcohol-
ysis of the primary amide provided protected C-serine
aminoester 12 in excellent yields,[24] although these reaction
conditions were problematic with acetate-protected sugars.
To further demonstrate the versatility of 1 as a launch pad
for glycoconjugate synthesis, we performed a series of
derivatizations of the aldehyde (Scheme 4). Selective a-
chlorination with l-prolinamide and NCS and subsequent
Table 1: Continuous flow reaction run for 24 hours for Ac- and Piv-
protected sugars.
Entry Substrate
3
No. of Product Conv. [%][a] Yield [%][a]
(equiv) modules
1
2
3
4
2
2
4
4
2
4
4
4
2
2
2
3
1
1
5
5
>97
>97
75
70 (65)
85 (77)
n.d.
>97
85 (46)
[a] Determined by 1H NMR spectroscopy using p-dimethoxybenzene as
internal standard. The yield of isolated product is given within the
parentheses.
With the development of an efficient route to multigram
quantities of 1, we then focused on the synthesis of a C-serine
glycoamino acid (Scheme 3). One-pot asymmetric Strecker
cyanation leads to the aminonitrile in high yields and
diastereoselectivity for both the acetate- (7) and pivaloate-
protected (8) glycosides. While traditional reaction conditions
Scheme 4. Synthesis of glycoconjugates from 1. Boc=tert-butoxycar-
bonyl, DMAP=4-dimethylaminopyridine, DMF=N,N-dimethylform-
amide, DMSO=dimethylsulfoxide, NCS=N-chlorosuccinimide,
PMP=para-methoxyphenyl, pyr=pyridine, THF=tetrahydrofuran,
Ts =4-toluenesulfonyl.
oxidation provided the chloroester 13 as a single diastereo-
mer.[25] Azide substitution led to the protected C-glucosyl
alanine derivative as the azidoester in short order with a 51%
overall yield from 2. Proline-catalyzed addition of 1 into the
iminoglyoxalate 15 and subsequent protection provided 16 as
a single diastereomer.[26] Reduction of 1 and acylation with
dodecanoyl chloride afforded the model C-glycolipid 17.
In summary, a simple flow reactor provides an efficient
solution to the problem of photon-starved large-scale photo-
redox reactions. Highly absorbing [RuL3]2+ catalysts create
strong concentration gradients that localize photoactivated
catalysts near the surface of the vessel, thus requiring thinner
reaction vessels to maximize efficiencies. This efficiency was
achieved through a flow apparatus using FEP tubing. The
light-starved nature of the photoredox reaction was con-
firmed through assessing the relationship between tubing size
and TOFs, which were considerably higher than those of
typical batch reactions and independent of scale. This design
Scheme 3. Synthesis of C-linked isostere of glucosyl-serine. Bn=ben-
zyl, M.S.=molecular sieves, TMS=trimethylsilyl.
4142
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 4140 –4143

Angewandte
Chemie
2011, 16, 7522 – 7550; c) J. Wegner, S. Ceylan, A. Kirschning,
Chem. Commun. 2011, 47, 4583 – 4592.
concept was applied to the synthesis of a key C-glycoside
intermediate, which was subsequently converted into a series
of C-linked glycoconujugates.
[11] For recent examples of photoflow reactors, see: a) B. D. A.
Hook, W. Dohle, P. R. Hirst, M. Pickworth, M. B. Berry, K. I.
Booker-Milburn, J. Org. Chem. 2005, 70, 7558 – 7564; b) Y. S. M.
Vaske, M. E. Mahoney, J. P. Konopelski, D. L. Rogow, W. J.
McDonald, J. Am. Chem. Soc. 2010, 132, 11379 – 11385; c) P.
Laurino, R. Kikkeri, N. Azzouz, P. H. Seeberger, Nano Lett.
2011, 11, 73 – 78; d) F. Lꢃvesque, P. H. Seeberger, Org. Lett. 2011,
13, 5008 – 5011; e) A. C. A. C. Gutierrez, T. F. Jamison, Org. Lett.
2011, 13, 6414 – 6417; f) R. A. Bourne, X. Han, M. Poliakoff,
M. W. George, Angew. Chem. 2009, 121, 5426 – 5429; Angew.
Chem. Int. Ed. 2009, 48, 5322 – 5325; g) F. Lꢃvesque, P. H.
Seeberger, Angew. Chem. 2012, 124, 1738 – 1741; Angew. Chem.
Int. Ed. 2012, 51, 1706 – 1709.
[12] a) F. W. Schmidtmann, T. E. Benedum, G. J. McGarvey, Tetra-
hedron Lett. 2005, 46, 4677 – 4681; b) A. Dondoni, P. P. Giovan-
nini, A. Marra, J. Chem. Soc. Perkin Trans. 1 2001, 2380 – 2388.
[13] a) R. W. Franck, M. Tsuji, Acc. Chem. Res. 2006, 39, 692 – 701;
b) G. Chen, M. Chien, M. Tsuji, R. W. Franck, ChemBioChem
2006, 7, 1017 – 1022; c) X. Zhu, Y. Jin, J. Wickham, J. Org. Chem.
2007, 72, 2670 – 2673.
[14] a) B. J. Dorgan, R. F. W. Jackson, Synlett 1996, 859 – 861; b) R. F.
Barghash, A. Massi, A. Dondoni, Org. Biomol. Chem. 2009, 7,
3319 – 3330; c) A. Dondoni, A. Marra, A. Massi, J. Org. Chem.
1999, 64, 933 – 944.
[15] A. Nuzzi, A. Massi, A. Dondoni, Org. Lett. 2008, 10, 4485 – 4488.
[16] Previous syntheses of 1 and similar compounds utilize tin-
mediated conditions. See: a) S. Liu, R. N. Ben, Org. Lett. 2005, 7,
2385 – 2388; b) K. Gotanda, M. Matsugi, M. Suemura, C. Ohira,
A. Sano, M. Oka, Y. Kita, Tetrahedron 1999, 55, 10315 – 10324.
[17] For a review of organocatalytic modifications of aldehydes to
amino acids, see M. J. Gaunt, C. C. C. Johansson, A. McNally,
N. T. Vo, Drug Discovery Today 2007, 12, 8 – 27.
Received: January 20, 2012
Published online: March 22, 2012
Keywords: C-glycosides · glycolipids · glycopeptides ·
.
photochemistry · redox chemistry
[1] a) R. Owens, J. Nettleship, Functional and Structural Proteomics
of Glycoproteins, Springer, Dordrecht, New York, 2011; b) A.
Varki, Glycobiology 1993, 3, 97 – 130; c) H. J. Allen, E. C.
Kisailus, Glycoconjugates: Composition, Structure, and Function,
Dekker, New York, 1992; d) R. A. Dwek, Chem. Rev. 1996, 96,
683 – 720.
[2] a) K. J. Doores, D. P. Gamblin, B. G. Davis, Chem. Eur. J. 2006,
12, 656 – 665; b) S. Danishefsky, J. R. Allen, Angew. Chem. 2000,
112, 882 – 912; Angew. Chem. Int. Ed. 2000, 39, 836 – 863; c) S.-i.
Hakomori, Y. Zhang, Chem. Biol. 1997, 4, 97 – 104; d) G.-J.
Boons, Z. Guo, Carbohydrate-Based Vaccines and Immuno-
therapies, Wiley, Hoboken, NJ, 2009; e) B. Kuberan, R. J.
Lindhardt, Curr. Org. Chem. 2000, 4, 653 – 677.
[3] a) P. Sears, C.-H. Wong, Angew. Chem. 1999, 111, 2446 – 2471;
Angew. Chem. Int. Ed. 1999, 38, 2300 – 2324; b) L. A. Marcaur-
elle, C. R. Bertozzi, Chem. Eur. J. 1999, 5, 1384 – 1390.
[4] a) G. Yang, J. Schmieg, M. Tsuji, R. W. Franck, Angew. Chem.
2004, 116, 3906 – 3910; Angew. Chem. Int. Ed. 2004, 43, 3818 –
3822; b) D. E. Levy, C. Tang, The Chemistry of C-Glycosides,
Pergamon, Tarrytown, 1995; c) C. H. Lin, H. C. Lin, W. B. Yang,
Curr. Top. Med. Chem. 2005, 5, 1431; d) W. Zou, Curr. Top. Med.
Chem. 2005, 5, 1363; e) P. G. Hultin, Curr. Top. Med. Chem.
2005, 5, 1299; f) F. Nicotra, Top. Curr. Chem. 1997, 187, 55.
[5] a) A. Dondoni, A. Marra, Chem. Rev. 2000, 100, 4395 – 4421;
b) M. H. D. Postema, C-Glycoside Synthesis, CRC, Boca Raton,
FL, 1995; c) P. Meo, H. M. I. Osborn, Best Synthetic Methods:
Carbohydrates, Academic Press, London, 2003; d) L. Somsꢂk,
Chem. Rev. 2001, 101, 81; e) Y. Du, R. J. Linhardt, I. R. Vlahov,
Tetrahedron 1998, 54, 9913.
[18] Molar absorptivity reported in acetonitrile. See: A. Juris, V.
Balzani, F. Barigelletti, S. Campagna, P. Belser, A. von Zelewsky,
Coord. Chem. Rev. 1988, 84, 85 – 277.
[19] Purchased from NewAge Industries (www.newageindustries.-
com). Cost per foot: 1.6 mm tubing (1/16’’ I.D.) = $0.998, 0.8 mm
tubing (1/32’’ I. D.) = $0.528.
[20] This increase in rate with decreasing catalyst concentration has
been previously observed. See Ref. [6e].
[21] During peerreview, a similar flow reactor design was published
by F. R. Bou-Hamdan, P. H. Seeberger, Chem. Sci. 2012, DOI:
10.1039/C2SC01016J.
[6] a) H. Gong, R. Sinisi, M. R. Gagnꢃ, J. Am. Chem. Soc. 2007, 129,
1908; b) H. Gong, M. R. Gagnꢃ, J. Am. Chem. Soc. 2008, 130,
12177; c) H. Gong, R. S. Andrews, J. L. Zuccarello, S. J. Lee,
M. R. Gagnꢃ, Org. Lett. 2009, 11, 879 – 882.
[22] 93% based on recovered starting material. See the Supporting
Information for full details.
[23] a) T. Ghaffar, A. W. Parkins, Tetrahedron Lett. 1995, 36, 8657 –
8660; b) T. Ghaffar, A. W. Parkins, J. Mol. Catal. A 2000, 160,
249 – 261.
[24] Stereochemistry assigned from crystal structure. See the Sup-
porting Information for details.
[25] Stereochemistry assigned from crystal structure of alcohol
derived from NaBH4 reduction of aldehyde intermediate. See
the Supporting Information for details.
[7] For a related nickel-catalyzed method, see: X. Yu, Y. Dai, T.
Yang, M. R. Gagnꢃ, H. Gong, Tetrahedron 2011, 67, 144 – 151.
[8] a) R. S. Andrews, J. J. Becker, M. R. Gagnꢃ, Angew. Chem. 2010,
122, 7432 – 7434; Angew. Chem. Int. Ed. 2010, 49, 7274 – 7276;
b) R. S. Andrews, J. J. Becker, M. R. Gagnꢃ, Org. Lett. 2011, 13,
2406 – 2409.
[9] For recent reviews of photoredox catalysis, see: a) K. Zeitler,
Angew. Chem. 2009, 121, 9969 – 9974; Angew. Chem. Int. Ed.
2009, 48, 9785 – 9789; b) T. P. Yoon, M. A. Ischay, J. Du, Nat.
Chem. 2010, 2, 527 – 532; c) J. M. R. Narayanam, C. R. J. Ste-
[26] Attempts to determine stereochemistry by single-crystal X-ray
diffraction were unsuccessful. Stereochemistry assigned by
analogy to previously reported method. See: A. Cꢄrdova, S.-i.
Watanabe, F. Tanaka, W. Notz, C. F. Barbas, J. Am. Chem. Soc.
2002, 124, 1866 – 1867.
´
phenson, Chem. Soc. Rev. 2011, 40, 102 – 113; d) F. Teply, Collect.
Czech. Chem. Commun. 2011, 76, 859 – 917.
[10] For perspectives and reviews on photoflow microreactors, see:
a) E. E. Coyle, M. Oelgemçller, Photochem. Photobiol. Sci.
2008, 7, 1313 – 1322; b) M. Oelgemçller, O. Shvydkiv, Molecules
Angew. Chem. Int. Ed. 2012, 51, 4140 –4143
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4143