Abstract
Herein, we report the development of continuous flow photoreactors for large scale ESIPT-mediated [3+2]-photocycloaddition of 2-(p-methoxyphenyl)-3-hydroxyflavone and cinnamate- derived dipolarophiles. These reactors can be efficiently numbered up to increase throughput two orders of magnitude greater than the corresponding batch reactions.
Keywords: photochemistry, flow chemistry, ESIPT, photocycloaddition, rocaglate, natural product, aglain
Graphical abstract

1. Introduction
Rocaglates are a class of structurally complex secondary metabolites isolated from plants of the Aglaia genus [1-3]. Over 100 rocaglates have been isolated or synthesized, all sharing the cyclopenta[b]benzofuran core structure (Figure 1). The naturally occurring family members rocaglamide (1), rocaglaol (2), silvestrol (3), and the synthetic analogue RHT (4) possess potent cytotoxicity for a number of cancer cell lines [4-8]. Much of the anticancer activity of these compounds is believed to be derived from translation suppression through inhibition of the RNA helicase eIF4A. However, there have been a number of new and potentially important biological activities reported for the natural products and analogues. Methyl rocaglate (4) was found to be a specific inhibitor of TNFα-R or PMA-induced NF-κB activity in T cell lines [4,5][9-13] and rocaglamide was recently reported to inhibit cancer cell migration through Rho GTPase inhibition [14]. In a collaboration with Kramnik and co-workers, we found that the synthetic rocaglate derivatives (3 and 4) synergized with low concentrations of IFNγ in primary macrophages to stimulate expression of some IFN-inducible genes, including a key regulatory factor, Irf1, which activates autophagy [15].
Figure 1. Biologically active rocaglate natural products and analogues.
It is clear that the rocaglates have a number of unique biological activities that could lead to novel therapeutics. However, the chemical synthesis of rocaglate natural products and analogues has proven highly challenging [16-19]. The common approach leverages an excited-state intramolecular proton transfer (ESIPT)-mediated [3+2]-photocycloaddition to generate aglains (7) from 3-hydroxyflavone derivatives (5) and dipolarophiles (6) which was developed by the Porco laboratory (Scheme 1) [19][20]. Subsequent α-ketol rearrangement and hydroxyl-directed reduction affords the rocaglate skeleton (8) [21-25].
Scheme 1. ESIPT-mediated [3+2] photocycloaddition.
While this synthetic approach has been widely utilized, there has remained significant challenges related to practicality and scalability of the photochemical cycloaddition. Thus, synthesis of sufficient material to support medicinal chemistry studies of rocaglate analogues in vitro and in vivo has been a significant hurdle for development of this class of molecule. Herein, we report a numbering up approach for developing a multigram scale continuous flow reactor for ESIPT [3+2]-photocycloaddition.
2. Results and Discussion
The ESIPT [3+2]-photocycloaddition of 3-hydroxyflavone (9) and methyl trans-cinnamate (10) to obtain aglain (11) is typically carried out in batch, but is difficult to scale due to long irradiation times (typically 12 hours), low temperature (0 °C), and high dilution required (30 mM). The typical isolated yield of aglain (11) is 40-50% affording on average <150 mg per reaction (Scheme 2a)[21]. In 2012, Tremblay and coworkers reported the use of a recirculating flow reactor for synthesis of methyl rocaglate and related derivatives (Scheme 2c). This three-step sequence (photocycloaddition, α-ketol rearrangement, and hydroxyl-directed reduction) was initiated with 3-hydroxyflavone derivative (11) and methyl trans-cinnamate and gave an overall 20 % yield of the methyl rocaglate derivative (13) on a 3 g scale [26]. The synthesis of aza-rocaglates via ESIPT [3+2]-photocycloaddition was also recently reported using a recirculating flow device [27] wherein the aza-aglain derivative (15) was obtained in 46 % over nine hours of irradiation on a 500 mg scale. However, these methods remain limited in terms of scale and throughput [28]. To enable synthesis and medicinal chemistry programs related to the rocaglate scaffolds and to address the challenges of this reaction, we considered a continuous-flow approach which has been demonstrated to be highly beneficial for large scale synthesis of active pharmaceutical ingredients, natural products, and polymers have been reported [29-32]. In particular, photochemical reactions performed in a continuous-flow can be significantly more efficient than the corresponding batch reactions due to better light penetration, resulting in reduced reaction times, cleaner reactions, and greater scalability [33-39].
Scheme 2.
Previous studies a) [3+2] photocycloaddition in batch, b) recirculating flow reactor, c) aza-aglain synthesis using a recirculating flow reactor.
2.1 Optimization of ESIPT [3+2]-photocycloaddition in flow
While there have been significant advances in developing flow reactors for large scale photochemical reactions [40-42], there are still few available options which could be applied to challenging reactions requiring low temperatures and are simple, affordable, and modular. We designed and built a novel flow photoreactor that would enable UV irradiation >300 nm while maintaining a low temperature for the photocycloaddition (Figure 2 A). The reactor utilizes a 150 W metal-halide lamp (>300 nm) which is placed in the center of a reactor containing wrapped FEP tubing. The inner chamber of the reactor is cooled by recirculation of glycol. We built the first flow reactor (Reactor A) using 1/16″ OD FEP tubing resulting in 3.5 mL reactor volume.
Figure 2.
Flow reactor A.
We began optimization of the reaction using 3-hydroxyflavone 9 [26] and methyl cinnamate 10. Initial reactions revealed that under typical conditions (30 mM) full conversion was observed with only 15-minute residence time, compared to 9-hour reaction times required for the corresponding batch reaction (Table 1, entries 1 and 2). Increasing the concentration to 60 mM did not affect the reaction and we obtained an increased isolated yield (Table 1, entry 3). In order to maximize the potential throughput and minimize solvent waste, we increased the concentration to 120 mM at which point a longer 30-minute residence time was required for full conversion (Table 1, entry 4). Although isolated yields for the flow reactions were slightly lower than the batch reaction these initial efforts resulted in nearly a 20-fold increase in throughput.
Table 1. Reaction optimization with flow reactor.
| ||||||
|---|---|---|---|---|---|---|
| Entry | [9] | [3+2] Adduct | Residence Time | Flow Rate | Yield | |
| 1 | 30 mM | 10 | 11 | Batch (9 hr) | -- | 55 % |
|
| ||||||
| 2 | 30 mM | 10 | 11 | 15 min. | 233 μL/min | 32 % |
| 3 | 60 mM | 10 | 11 | 15 min. | 233 μL/min | 41 % |
| 4 | 120 mM | 10 | 11 | 30 min. | 117 μL/min | 45 % |
|
| ||||||
| 5 | 30 mM | 16 | 17 | Batch (9 hr) | -- | 47 % |
|
| ||||||
| 6 | 120 mM | 16 | 17 | 30 min. | 117 μL/min | 36 % |
Targeting a key intermediate for medicinal chemistry studies, we evaluated the potential for large-scale synthesis of the Weinreb amide-bearing aglain 17 derived from ESIPT [3+2]-photocycloaddition with hydroxamate 16 [43]. Aglain 17 is the precursor to the rocaglate congener RHT, an important tool molecule for studying inhibition of eIF4A, and an intermediate in the synthesis of many analogues including furan 6. Photoreaction of 3-hydroxyflavone 9 and 16 afforded 47 % yield after 9 hours irradiation in a batch reactor (Table 1, entry 5) while the flow reaction at 120 mM required 30-minute residence time affording a 36 % isolated yield (Table 1, entry 6).
To increase the throughput of the reaction, we envisioned a numbering up approach [44] wherein we would daisy-chain reactors rather than splitting the reaction stream into separate reactors. In this way, we would effectively increase the total reactor volume allowing for greater flow rate, and subsequent throughput (Figure 2, B and C). We initially combined two reactors (reactor A) which effectively doubled the reactor volume to 7 mL. Reaction of hydroxyflavone 9 and dipolarophile 16 afforded a comparable isolated yield and the reaction throughput was increased from 0.12g/h to 0.22 g/h. Using this system, we synthesized 890 mg of aglain product 17 in only four hours (Table 2, entry 2). By redesigning the reactor to have a larger surface area, it was possible to increase the volume to 6 mL/reactor (reactor B). Using this design, both a single and double reactor afforded similar yield, significantly increasing the throughput to 0.6 g/h (Table 2, entries 3 & 4).
Table 2. Reaction carried out in the multi-reactor setup.
| |||||||
|---|---|---|---|---|---|---|---|
| Entry | [9] | Dipolarphile | Reactor Volume | Residence Time | Flow Rate | Yield | Throughput |
| 1 | 120 mM | 16a | 3.5 mL (Single A) | 30 min. | 117 μL/min | 36 % | 0.14 g/h |
| 2 | 120 mM | 16a | 7.0 mL (Double A) | 30 min. | 233 μL/min | 29 % | 0.22 g/h |
|
| |||||||
| 3 | 120 mM | 16b | 6.0 mL (Single B) | 30 min. | 200 μL/min | 28 % | 0.17 g/h |
| 4 | 120 mM | 16b | 12.0 mL (Double B) | 20 min. | 600 μL/min | 32 % | 0.61 g/h |
|
| |||||||
| 5 | 120 mM | 16b | 37.5 mL (Triple C) | 35 min. | 1.07 mL/min | 35 % | 1.16 g/h |
In order to further increase the reactor volume and throughput, we built a new reactor with 1/8″ OD FEP, resulting in a total volume of 12.5 mL for each reactor (reactor C). We found that the impact of the longer path length was minimal and we could obtain similar isolated yields with only a slightly longer residence time of 35 minutes. Numbering up this larger reactor to three resulted in a total volume of 37.5 mL, thereby maintaining a 35% yield to realize a throughput of 1.16 g/h (Table 2, entry 5).
A large scale [3+2] photocycloaddition was performed using this latest design. 21.3 g of 3-hydroxyflavone (9) and 124.3 g of dipolarophile (16) with a total solution volume of 670 mL took 10.5 hours with a flow rate of 1.07 mL/min to reach completion. A total of 12.1 grams of aglain 17 was obtained after purification.
2.2 Synthesis of rocaglate analogs in flow
To validate that this flow system would be generally useful for medicinal chemistry and the synthesis of an array of rocaglate analogues, we evaluated a number of diopolarophiles in the flow ESIPT [3+2]-photocycloaddition. As outlined in Table 43, multiple reaction partners were found to be successful using the system, generally providing moderate to good isolated yields and excellent throughput. With the triple reactor configuration, we observed an increase to 61% isolated yield for methyl cinnamate. Reaction with p-bromo ethyl cinnamate (19) afforded aglain 24 in 52% isolated yield and the corresponding Weinreb amide (20) was equally successful in the reaction (52%). Reaction with o-bromo and m-bromo Weinreb amides (21, 22) afforded lower isolated yields of the corresponding aglains (26: 32% and 27: 38% respectively) but maintained excellent throughput. The electron rich p-methoxy Weinreb amide (23) afforded a good yield (47%) and throughput of the expected aglain (28).
3. Conclusion
We have developed a series of simple continuous flow photoreactors for carrying out the ESIPT [3+2]-photocycloaddition, generating aglain derivatives on unprecedented scale. We have demonstrated the reactors can be effectively numbered up to increase the reaction throughput without significant change in the overall reaction outcome. Using a triple reactor, we successfully synthesized 12.1 grams of the critical aglain 17 in 10.5 hours. The ability to carry out these challenging reactions on a large scale has significantly improved the prospects for advancing these highly interesting molecules through in vitro and in vivo medicinal chemistry programs.
4. Experimental
4.1 General information
Photoreactors were constructed with fluorinated ethylene propylene (FEP) tubing (1/16″ [OD] – 1/32″ [ID], and 1/8″ [OD] – 1/16″ [ID]). The light source was metal halide lamp (MHL 150 G12) purchased from Ushio America (Cypress, CA) with metal halide ballast (V90D7130K) purchased from Venture Lighting International (Twinsburg, Ohio). Solutions were flowed through the TPFE photoreactors with Harvard Apparatus (Holliston, MA) PHD Ultra syringe pumps or KD Scientific (Holliston, MA) LEGATO 200 syringe pumps.
4.2 General procedure of [3+2]-photocycloaddition
The solvent (70:30 chloroform:2,2,2-trifluoroethanol) for photoreactions was degassed with argon for 10 minutes prior to dissolution of the reaction substrates. A flask containing 3-hydroxyflavone 9 and dipolarophile was evacuated under high vacuum for 30 minutes then charged with argon. The degassed solvent mixture was added to the flask to prepare the solution in the desired concentration. The solution subsequent reaction solution was transferred to appropriate syringes and placed into the flow system. The crude material was collected and after the reaction was complete the solution was concentrated dry loaded onto Celite followed by chromatography (SiO2, hexanes/ethyl acetate/Et3N). From the chromatography column was collected a mixture of the major product with regioisomers and diastereomers still present. Subsequent trituration with ethyl acetate/hexanes solution or recrystallization from acetonitrile afforded pure aglain product as a single diastereomer.
Supplementary Material
Figure 3.
Daisy-chained reactors.
Scheme 4. Large scale [3+2] –triple reactor C (37.5 mL).
Table 3.
Reaction with addition reaction partners using triple reactor C.
| |||||
|---|---|---|---|---|---|
| Entry | [9] | DipolarphileaDipolarophilea | [3+2] Adduct | Yield | Throughput |
| 1 | 120 mM | 10 | 11 | 61 % | 1.92 g/h |
| 2 | 120 mM | 19 | 24 | 52 % | 1.78 g/h |
| 3 | 120 mM | 20 | 25 | 52 % | 1.80 g/h |
| 4 | 120 mM | 21 | 26 | 32 % | 1.10 g/h |
| 5 | 120 mM | 22 | 27 | 38 % | 1.31 g/h |
| 6 | 120 mM | 23 | 28 | 47 % | 1.58 g/h |
10.0 equiv.
Acknowledgments
Financial support from Boston University and National Institutes of Health (ABB R33AI105944) is gratefully acknowledged. We thank Dr. Norman Lee (Boston University) for high-resolution mass spectrometry data. NMR (CHE-0619339) and MS (CHE-0443618 facilities at Boston University are supported by the NSF.
Footnotes
Supporting Information Available: Supplementary data is available in the online version at XXXXX
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Proksch P, Edrada R, Ebel R, Bohnenstengel FI, Nugroho BW. Curr Org Chem. 2001;5:923–938. [Google Scholar]
- 2.Kraus GA, Sy JO. J Org Chem. 1989;54:77–83. [Google Scholar]
- 3.Trost BM, Greenspan PD, Yang BV, Saulnier MG. J Am Chem Soc. 1990;112:9022–9024. [Google Scholar]
- 4.King ML, Chiang CC, Ling HC, Fugita E, Ochiai M, McPhail AT. J Chem Soc, Chem Commun. 1982:1150–1151. [Google Scholar]
- 5.Lee SK, Cui B, Mehta RR, Kinghorn AD, Pezzuto JM. Chem Biol Interact. 1998;115:215–228. doi: 10.1016/s0009-2797(98)00073-8. [DOI] [PubMed] [Google Scholar]
- 6.Hwang BY, Su BN, Chai H, Mi Q, Kardono LBS, Afriastini JJ, iswan S, Santarsiero BD, Mesecar AD, Wild R, Fairchild CR, Vite GD, Rose WC, Farnsworth NR, Cordell GA, Pezzuto JM, Swanson SM, Kinghorn AD. J Org Chem. 2004;69:3350–3358. doi: 10.1021/jo040120f. [DOI] [PubMed] [Google Scholar]
- 7.Chen W, Pan L, Kinghorn AD, Swanson SM, Burdette JE. BMC Cancer. 2016;16:17. doi: 10.1186/s12885-015-1988-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kogure T, Kinghorn AD, Yan I, Bolon B, Lucas DM, Grever MR, Patel T. Plos One. 2013;8:e76136. doi: 10.1371/journal.pone.0076136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Baumann B, Bohnenstengel F, Siegmund D, Wajant H, Weber C, Herr I, Debatin KM, Proksch P, Wirth T. J Biol Chem. 2002;277:44791–44800. doi: 10.1074/jbc.M208003200. [DOI] [PubMed] [Google Scholar]
- 10.Fahrig T, Gerlach I, Horváth E. Mol Pharmacol. 2005;67:1544–1555. doi: 10.1124/mol.104.008177. [DOI] [PubMed] [Google Scholar]
- 11.Rivero-Cruz JF, Chai HB, Kardono LBS, Setyowati FM, Afriatini JJ, Riswan S, Farnsworth NR, Cordell GA, Pezzuto JM, Swanson SM, Kinghorn AD. J Nat Prod. 2004;67:343–347. doi: 10.1021/np0304417. [DOI] [PubMed] [Google Scholar]
- 12.Lee SK, Cui B, Mehta RR, Kinghorn AD, Pezzuto JM. Chemico-Biol Interact. 1998;115:215–228. doi: 10.1016/s0009-2797(98)00073-8. [DOI] [PubMed] [Google Scholar]
- 13.Pan L, Acuña UM, Li J, Jena N, Ninh TN, Pannell CM, Chai H, Fuchs JR, Carcache de Blanco EJ, Soejarto DD, Kinghorn AD. J Nat Prod. 2013;76:394–404. doi: 10.1021/np3007588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Becker MS, Müller PM, Bajorat J, Schroeder A, Giaisi M, Amin E, Ahmadian MR, Rocks O, Köhler R, Krammer PH, Li-Weber M. Oncotarget. 2016;7:51908–51921. doi: 10.18632/oncotarget.10188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bhattacharya B, Chatterjee S, Devine WG, Kobzik L, Beeler AB, Porco JA, Jr, Kramnik I. Sci Rep. 6:24409. doi: 10.1038/srep24409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kraus GA, Sy JO. J Org Chem. 1989;54:77–83. [Google Scholar]
- 17.Davey AE, Taylor RJK. J Chem Soc, Chem Commun. 1987:25–27. [Google Scholar]
- 18.Davey AE, Schaeffer MJ, Taylor RJK. J Chem Soc, Chem Commun. 1991:1137–1139. [Google Scholar]
- 19.Dobler MR, Bruce I, Cederbaum F, Cooke NG, Diorazio LJ, Hall RG, Irving E. Tetrahedron Lett. 2001;42:8281–8284. [Google Scholar]
- 20.Gerard B, Jones G, II, Porco JA., Jr J Am Chem Soc. 2004;126:13620–13621. doi: 10.1021/ja044798o. [DOI] [PubMed] [Google Scholar]
- 21.Gerard B, Cencic R, Pelletier J, Porco JA., Jr Angew Chem Int Ed Engl. 2007;46:7831–7834. doi: 10.1002/anie.200702707. [DOI] [PubMed] [Google Scholar]
- 22.Roche SP, Cencic R, Pelletier J, Porco JA., Jr Angew Chem Int Ed. 2010;49:6533–6538. doi: 10.1002/anie.201003212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rodrigo CM, Cencic R, Roche SP, Pelletier J, Porco JA., Jr J Med Chem. 2011:558–562. doi: 10.1021/jm201263k. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lajkiewicz NJ, Cognetta AB, III, Niphakis MJ, Cravatt BF, Porco JA., Jr J Am Chem Soc. 2014;136:2659–2664. doi: 10.1021/ja412431g. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Gerard B, Sangji S, O'Leary DJ, Porco JA., Jr J Am Chem Soc. 2006;128:7754–7755. doi: 10.1021/ja062621j. [DOI] [PubMed] [Google Scholar]
- 26.Liu T, Nair SJ, Lescarbeau A, Belani J, Peluso S, Conley J, Tillotson B, O'Hearn P, Smith S, Slocum K, West K, Helble J, Douglas M, Bahadoor A, Ali J, McGovern K, Fritz C, Palombella VJ, Wylie A, Castro AC, Tremblay MR. J Med Chem. 2012;55:8859–8878. doi: 10.1021/jm3011542. [DOI] [PubMed] [Google Scholar]
- 27.Wang W, Cencic R, Whitesell L, Pelletier J, Porco JA., Jr Chem Eur J. 2016;22:12006–12010. doi: 10.1002/chem.201602953. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Efforts in the synthesis of gram quantities of silvestrol have also been pursued at National Cancer Institute, Develpmental Therapeutics Program. https://dtp.cancer.gov/organization/dscb/smChemistry/silvestrol.htm.
- 29.Wenger J, Ceylan S, Kirschning A. Adv Synth Catal. 2012;354:17–57. [Google Scholar]
- 30.Brzozowski M, O'Brien M, Ley SV, Polyzos A. Acc Chem Res. 2015;48:349–362. doi: 10.1021/ar500359m. [DOI] [PubMed] [Google Scholar]
- 31.Porta r, Benaglia M, Puglisi A. Org Process Res Dev. 2016;20:2–25. [Google Scholar]
- 32.Junkers T. Macromol Chem Phys. 2016;218:1600421. [Google Scholar]
- 33.Knowles JP, Elliott LD, Booker-Milburn KI. Beilstein J Org Chem. 2012;8:2025–2052. doi: 10.3762/bjoc.8.229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gilmore K, Seeberger PH. Chem Rec. 2014;14:410–418. doi: 10.1002/tcr.201402035. [DOI] [PubMed] [Google Scholar]
- 35.Atodiresei I, Vila C, Rueping M. ACS Catal. 2015;5:1972–1985. [Google Scholar]
- 36.Tucker JW, Zhang Y, Jamison TF, Stephenson CRJ. Angew Chem Int Ed. 2012;51:4144–4147. doi: 10.1002/anie.201200961. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Telmesani R, Park SH, Lynch-Colameta T, Beeler AB. Angew Chem Int Ed. 2015;54:11521–11525. doi: 10.1002/anie.201504454. [DOI] [PubMed] [Google Scholar]
- 38.Yueh H, Voevodin A, Beeler AB. J Flow Chem. 2015;5:155–159. [Google Scholar]
- 39.Cambié D, Bottecchia C, Straathof NJW, Hessel V, Noël T. Chem Rev. 2016;116:10276–10341. doi: 10.1021/acs.chemrev.5b00707. [DOI] [PubMed] [Google Scholar]
- 40.Elliott LD, Berry M, Harji B, Klauber D, Leonard J, Booker-Milburn KI. Org Process Res Dev. 2016;20:1806–1811. [Google Scholar]
- 41.Josland S, Mumtaz S, Oelgemöller M. Chem Eng Tech. 2016;39:81–87. [Google Scholar]
- 42.Gutierrez AC, Jamison TF. J Flow Chem. 2011;1:24–27. [Google Scholar]
- 43.Nahm S, Weinreb SM. Tetrahedron Lett. 1981;22:3815–3818. [Google Scholar]
- 44.Su Y, Kuijipers K, Hessel V, Noël T. React Chem Eng. 2016;1:73–81. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.






