Insights into the Regioselectivity of Metal-Catalyzed Aryne Reactions

Abstract

The synthetic potential of unsymmetrically substituted aryne intermediates is significantly hindered by regioselectivity issues. Current methods for inducing regioselectivity all rely on substrate control and are focused on non-metallated arynes. Before our initial disclosure, there was no systematic study regarding the regioselectivity of metal-catalyzed aryne reactions. By exploiting ligand control, we have induced regioselectivity in a palladium-catalyzed aryne annulation to form phenanthridinones (up to 9:91 r.r.). Through this study we have investigated: ligand effects, influence of steric perturbation, and the impact of the aryne precursor.

1 Introduction

2 Inducing Regioselectivity via Ligand Control

3 A Comparison of o-Borylaryl Triflate Aryne Precursors to Kobayashi Aryne Precursors

4 Conclusion

Key words

catalysts - ligands - precursors - selectivity - aryne

1

Introduction

Courtney C. Roberts (right) grew up in Los Angeles, CA, and attended Pepperdine University for her undergraduate education. In 2011, she became the first graduate student in the lab of Prof. Simon Meek at UNC Chapel Hill, where she developed catalysts for enantioselective diene hydrofunctionalizations. After obtaining her PhD in 2016, she moved to the University of Michigan for postdoctoral work with Prof. Melanie Sanford, where she worked on high-valent Ni catalysis. Courtney began her independent career in 2019 at the University of Minnesota, where her group works in the fields of early-transition-metal catalysis, aryne chemistry, and photoredox catalysis. Erin E. Plasek (middle) grew up in Apple Valley, MN, and completed her undergraduate studies at St. Olaf College. In 2019, she joined the first class of graduate students in the lab of Prof. Courtney Roberts at UMN, where she now works on enhancing the regioselectivity of metal-catalyzed reactions of unsymmetrical arynes and enabling the synthesis of previously inaccessible five-membered heteroarynes. Brylon N. Denman (left) grew up in Farmington, MO, and attended St. Louis University for her undergraduate education. In 2020, she joined the lab of Prof. Courtney Roberts at UMN, where she now works on resolving issues with regioselectivity of metal-catalyzed aryne reactions while also working to establish a regioselective model for metalated arynes.

1.1

Arynes as Building Blocks

Arynes have been sought-after targets for over 120 years due to their highly reactive nature.[1] The excess strain of these compounds (ΔH_dehydro ≈ 100 kcal mol^–1)[2] can be leveraged to effectively synthesize difunctionalized arenes in a single step. This has allowed access to many value-added species such as ligands, natural products, and other conjugated compounds.[3] [4] [5] [6] [7] [8] [9] [10] [11] Although arynes have been used with much success, aryne methodology continues to be plagued with regioselectivity challenges.[8] ^, [12–15] Reactions of unsymmetrical arynes can (and often will) result in mixtures of regioisomeric products and lead to poor reaction efficiency. This is demonstrated in the total synthesis of natural product ellipticine, in which an unsymmetric 3,4-pyridyne intermediate results in the formation of regioisomers in near equimolar amounts (Figure [1]).[14] Efforts have been made to increase regioselectivity of free arynes;[16] however, the regioselectivity of metal-mediated aryne reactions remains underexplored.

Figure 1 Synthesis of ellipticine

1.2

Regioselectivity of Free Arynes

According to the Aryne Distortion Model, the observed regioselectivity of free arynes is dependent on the level of distortion within the triple bond.[2] [17] This level of distortion is indicative of regioselectivity, as nucleophilic attack is preferred at the most linear position of the triple bond. The most linear position is favored because less geometric distortion between the aryne and nucleophile is required to reach its transition state. Electronically activating functional groups ortho to the triple bond can increase distortion and lead to high levels of regioselectivity. This is illustrated in a report by Yoshida and co-workers for the synthesis of phenoxathiins, as only one regioisomer was observed when using the electronically influenced methoxy aryne (Figure [2]).[18] In contrast, a 54:46 ratio of regioisomers was observed when using a Kobayashi precursor to generate an o-methyl aryne. Overall, this illustrates that regioselectivity of free arynes is substrate-controlled, as regioselectivity is dependent on the level of aryne distortion.

Figure 2 Aryne distortion model

Electronically influencing groups have been leveraged to induce aryne distortion and regioselectivity for unsymmetrically substituted arynes with poor inherent regioselectivity. This was demonstrated by Garg and co-workers in a sophisticated study that involved the addition of functional groups ortho to the triple bond of 3,4-pyridyne. This allowed for both enhanced and reversed regioselectivity compared to the parent unsubstituted species (Figure [3]).[16] [19] Although this strategy was immensely effective for 3,4-pyridyne across a range of nucleophiles, it does not allow access to all moieties as pre-existing functional groups may occupy the positions ortho to the triple bond. Additionally, the removal or subsequent derivatization of the electronically influencing group requires further synthetic measures, which lowers reaction efficiency and atom economy.

Figure 3 Using the aryne distortion model to induce regioselectivity

1.3

Regioselectivity of Metal-Bound Arynes

Although many different models have been developed to explain regioselectivity trends for free arynes, to the best of our knowledge, no such models exist for metallated arynes. Furthermore, the aryne distortion effects described above have no bearing on the regioselectivity of metallated arynes. The most compelling example of this comes from Hosoya and co-workers, where a discrete nickel-bound o-methoxy aryne shows poor regioselectivity upon stoichiometric reaction with allyl bromide and trimethylsilyl cyanide (Figure [4]).[20] This is in direct contrast to the complete selectivity of the o-methoxy aryne when reacted in the absence of metals (Figure [2])[18] and supports the conclusion that the distortion model does not govern the regioselectivity of metal-mediated aryne reactions.

Figure 4 Regioselectivity of nickel-bound o-methoxybenzyne

The lack of a regioselectivity model for metal-bound arynes highlights a significant limitation for the synthetic potential of arynes. As some transformations are only accessible through metal-mediated processes, understanding the factors influencing regioselectivity is a worthwhile pursuit. Catalytic transformations of arynes demonstrated by Huang and co-workers illustrate that regioselectivity is currently not well understood. In their report, the same o-methyl aryne exhibits drastically different regioselectivities when coupled with an allyl-substituted iodocyclohexenone vs. iodofuranone (Figure [5]).[21] It is hypothesized that the steric clash between the methyl group and benzyl protons of the minor product is unfavorable and thus drives the regioselectivity of the furanone coupling. However, due to the location of the gem-dimethyl moiety, a greater clash may occur for the methyl group in the major product, thus this may not be the best explanation for regioselectivity. In the same report, the regioselectivity observed using an o-methoxy aryne coupled with furanone (15:85) is thought to be due to chelation effects of the oxygen. Evidently, there are a variety of steric and electronic factors that influence the regioselectivities of metal-mediated aryne reactions, but none have been extensively studied or used to the advantage of improving selectivity.[21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] Prior to our initial disclosure, ligands had never been utilized to alter selectivity of aryne transformations, despite innumerable reports of ligand-controlled selectivity in other areas of organometallic chemistry (i.e., allylic substitution, Heck reactions, olefin metathesis).[49–57]

Figure 5 Divergent regioselectivity in palladium-catalyzed aryne annulations

2

Inducing Regioselectivity via Ligand Control

2.1

Establishing a Baseline

Inspired by ligand-controlled selectivity in other metal-mediated transformations,[49] [50] [51] [52] [53] [54] [55] [56] [57] we postulated that ligands could be used to control the binding and subsequent functionalization of the aryne. We selected a palladium-catalyzed annulation developed by Larock and co-workers for the synthesis of phenanthridinones as our model reaction.[58] Phenanthridinones are attractive compounds to study as they have found use as PARP inhibitors.[59] Initially, we sought to target the o-methyl aryne substrate due to its small steric and electronic profile. We screened a variety of symmetrical ligand systems to establish a consistent baseline selectivity of the o-methyl aryne towards its phenanthridinone regioisomers (Figure [6]).[60]

Figure 6 Establishing baseline regioselectivity. Yield determined as the sum of both regioisomers through ¹H NMR spectroscopy with mesitylene as an internal standard. Each entry is an average of two experiments.

Interestingly, each of these systems resulted in slightly different regioisomeric ratios. While this was preliminary evidence that ligands could alter the regioselectivity of metal-mediated aryne reactions, it also posed a question: What baseline selectivity should we choose to compare to? Using 22 mol% of tricyclohexylphosphine (PCy₃) delivered the lowest levels of regioselectivity; however, due to its monodenticity, it was ambiguous if all palladium species would remain bisligated throughout the cycle. We ultimately decided that the reactions using no ligand and diphenylphosphinomethane (dppm) were the fairest to use as a baseline. The system with no ligand allows us to observe the regioisomeric ratio without ligand influences; however, it does not provide a direct comparison to other ligands. Therefore, we also utilized dppm, the symmetric bidentate phosphine ligand Larock and co-workers found optimal.

2.2

Inducing Regioselectivity via Ligand Control

Figure 7 Ligand effects on regioselectivity. Yield determined as the sum of both regioisomers through ¹H NMR spectroscopy with mesitylene as an internal standard. Each entry is an average of two experiments.

After screening multiple classes of ligands, we found monodentate phosphine ligands to be optimal with respect to both yield and regioselectivity. A series of monodentate phosphorus ligands of variable size and electron donation were screened (Figure [7]).[61] [62] [63] The use of 11 mol% tri-tert-butylphosphine (P^tBu₃) provided the greatest regioselectivity and yield. We initially attributed this to its large cone angle; however, attempts to further increase cone angle using tri(o-tolyl)phosphine (P(o-tol)₃) led to diminished regioselectivity. Within the series of ligands, we also noticed a trend of greater regioselectivity with increased electron donation of the ligand. Thus, we proposed that both ligand size and electron donation affect regioselectivity.

2.3

Steric Perturbation of o-Borylaryl Triflate Aryne Precursors

Figure 8 Charton analysis of o-borylaryl triflates precursors with P^tBu₃ and dppm. Yield determined as the sum of both regioisomers through ¹H NMR spectroscopy with mesitylene as an internal standard. Each entry is an average of 2–3 experiments.

To further investigate the factors influencing regioselectivity, we performed Charton analysis to probe the relationship between the steric encumbrance of the aryne intermediate with the effect of the ligand (Figure [8]). Charton analysis is a linear free energy relationship that uses Charton values, derived from Van der Waals radii, to gauge the steric influence of a key substituent on selectivity.[64] [65] [66] [67] [68] We began by screening a set of aryne precursors that differ in their steric profile (R= Me, Et, ⁱPr, ^tBu) under our optimized conditions with P^tBu₃. Here we saw a preference for regioisomer B as the steric encumbrance of the aryne increased. It should be noted that the regioisomeric products are separable via chromatography. By plotting the log of the regioisomeric ratio as a function of the Charton value, v, we observed a highly linear relationship, which indicates that regioselectivity is influenced by steric effects of the aryne substituent. We next performed the same analysis using our baseline ligand (dppm) to study the impact of the ligand on regioselectivity. As with our optimized conditions, we saw a linear correlation between the regioselectivity and size of the aryne substituent; however, here we observe reduced regioselectivity. Yields were also decreased using dppm, likely due to increased rigidity and lessened lability of this bidentate ligand. Overall, this demonstrates that our optimized conditions do outperform the symmetric ligand environment for each aryne precursor.

To further explore why our optimized ligand system delivers higher levels of regioselectivity, we looked to the slopes of the Charton plots, which correspond to the sensitivity values (ψ). The sensitivity values indicate the degree to which steric encumbrance affects regioselectivity. The sensitivity value for P^tBu₃ (ψ = 2.11) is significantly larger in magnitude than that of the symmetric dppm system (ψ = 0.73). We propose that the large cone-angle of P^tBu₃ allows for greater steric interaction with the aryne, which amplifies regioselectivity. This difference observed between ligand environments showcases ligand control and is strong evidence that ligands can be used to intentionally control regioselectivity within aryne reactions.

3

A Comparison of o-Borylaryl Triflate Aryne Precursors to Kobayashi Aryne Precursors

3.1

Supporting the Presence of an Aryne Intermediate

After establishing regioselectivity, we next wanted to support the presence of an aryne intermediate. Prior to our initial disclosure, there was only one example of o-borylaryl triflates being used in a catalytic aryne reaction. In a report by Greaney and co-workers, o-borylaryl triflates were subjected to palladium-catalyzed conditions to produce a trimeric product.[69] Although this reaction was shown to occur via an aryne intermediate, we questioned if these precursors could generate our desired phenanthridinone products via an alternative iterative cross-coupling pathway including a Suzuki–Miyaura aryl–aryl coupling and an aryl-N Buchwald–Hartwig amidation.

Figure 9 Support for aryne intermediate. Yield determined as the sum of both regioisomers through ¹H NMR spectroscopy with mesitylene as an internal standard. Each entry is an average of 2–3 experiments.

Given that o-borylaryl triflates are relatively underexplored in catalysis, we sought to support the presence of an aryne intermediate via regioselectivity studies (Figure [9]). Both regioisomers of the o-borylaryl triflates (1 and 2) and Kobayashi precursors (o-silylaryl triflates) (3 and 4) were synthesized and screened under optimized conditions. Kobayashi precursors were chosen for comparison because they are well-studied and widely accepted to generate aryne intermediates. When all precursors were subjected to optimized conditions with P^tBu₃, the same regioisomer was favored in a similar magnitude. Overall, through these regioselectivity studies, we have gathered support for the presence of an aryne intermediate.

3.2

Regioselectivity of Kobayashi Aryne Precursors

During our inquiry to support an aryne intermediate, we discovered differences between o-borylaryl triflate precursors and Kobayashi precursors. Throughout the same steric series described in our steric perturbation (Section 2.3), lower levels of regioselectivity were observed using Kobayashi precursors; however, the same regioisomers were always favored (Figure [10]). This further supports the conclusion that o-borylaryl triflates generate an aryne intermediate in this reaction.

Figure 10 A comparison of Kobayashi and o-borylaryl triflate aryne precursors. Yield determined as the sum of both regioisomers through ¹H NMR spectroscopy with mesitylene as an internal standard. Each entry is an average of three experiments.

Another discrepancy is that Kobayashi precursors consistently resulted in lower yields. One key difference between these two precursors is their method of aryne activation. Kobayashi aryne precursors are hypothesized to undergo an elimination pathway after fluoride activation of the silyl group; however, their exact method of metallation is poorly understood. Unlike Kobayashi precursors, o-borylaryl triflate precursors require the presence of a metal to generate an aryne, as oxidative addition and transmetallation are required to occur. We hypothesize that the decrease in yield when using Kobayashi precursors may be due to uncontrolled off-cycle aryne formation. We believe this is circumvented when using o-borylaryl triflate precursors as the aryne is only able to form as fast as catalytic turnover allows. The variations in regioselectivity and yield suggest differences in metallation of these two precursor types may impact the rate and regioselectivity determining steps.

4

Conclusion

We have begun to develop a model for regioselectivity of metal-catalyzed aryne reactions. Much of the selectivity seems to arise from steric effects of the aryne; however, we have shown this effect can be amplified by increasing ligand size and electron donation. Overall, we have demonstrated that ligand control can be an effective maneuver to induce regioselectivity in metal-catalyzed aryne reactions. Through these studies we have also highlighted the synthetic potential of o-borylaryl triflates as aryne precursors. Further work is underway to determine the rate and regioselectivity determining step of this transformation. Quantitative structure-selectivity studies of the ligand are underway to interrogate the ligand parameters impacting regioselectivity. Additionally, we are testing this effect in other aryne systems to investigate the generality of this approach. Overall, we hope to gain more insight into the factors that govern metallated aryne selectivity to create a model of regioselectivity and improve the potential of such sought-after intermediates for the broader synthetic community.

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgment

We thank the T. Hoye group for equipment to obtain IR measurements. Mass spectrometry analysis was performed at the UMN Department of Chemistry Mass Spectrometry Laboratory (MSL), supported by OVPR, CSE, and the Department of Chemistry at UMN, as well as the National Science Foundation (NSF) (CHE-1336940).

• References

• 1 Stoermer R, Kahlert B. Ber. Dtsch. Chem. Ges. 1902; 35: 1633
• 2 Goetz AE, Bronner SM, Cisneros JD, Melamed JM, Paton RS, Houk KN, Garg NK. Angew. Chem. Int. Ed. 2012; 51: 2758
• 3 Tadross PM, Stoltz BM. Chem. Rev. 2012; 112: 3550
• 4 Anthony SM, Wonilowicz LG, McVeigh MS, Garg NK. JACS Au 2021; 1: 897
• 5 Spence KA, Chari JV, Niro MD, Susick RB, Ukwitegetse N, Djurovich PI, Thompson ME, Garg NK. Chem. Sci. 2022; 13: 5884
• 6 Chari JV, Spence KA, Susick RB, Garg NK. Nat. Commun. 2021; 12: 3706
• 7 Shi J, Li L, Li Y. Chem. Rev. 2021; 121: 3892
• 8 García-López J.-A, Greaney MF. Chem. Soc. Rev. 2016; 45: 6766
• 9 Berthelot-Bréhier A, Panossian A, Colobert F, Leroux FR. Org. Chem. Front. 2015; 2: 634
• 10 Shi F, Waldo JP, Chen Y, Larock RC. Org. Lett. 2008; 10: 2409
• 11 Dubrovskiy AV, Markina NA, Larock RC. Org. Biomol. Chem. 2012; 11: 191
• 12 Bhattacharjee S, Guin A, Gaykar RN, Biju AT. Org. Lett. 2020; 22: 9097
• 13 Seo JH, Ko HM. Tetrahedron Lett. 2018; 59: 671
• 14 May C, Moody CJ. J. Chem. Soc., Chem. Commun. 1984; 926
• 15 Gribble GW, Saulnier MG, Sibi MP, Obaza-Nutaitis JA. J. Org. Chem. 1984; 49: 4518
• 16 Goetz AE, Garg NK. J. Org. Chem. 2014; 79: 846
• 17 Medina JM, Mackey JL, Garg NK, Houk KN. J. Am. Chem. Soc. 2014; 136: 15798
• 18 Kanemoto K, Sakata Y, Hosoya T, Yoshida S. Chem. Lett. 2020; 49: 593
• 19 Goetz AE, Garg NK. Nat. Chem. 2013; 5: 54
• 20 Sumida Y, Sumida T, Hashizume D, Hosoya T. Org. Lett. 2016; 18: 5600
• 21 Huang X, Sha F, Tong J. Adv. Synth. Catal. 2010; 352: 379
• 22 Jeganmohan M, Bhuvaneswari S, Cheng C.-H. Angew. Chem. Int. Ed. 2009; 48: 391
• 23 Henderson JL, Edwards AS, Greaney MF. Org. Lett. 2007; 9: 5589
• 24 Bhuvaneswari S, Jeganmohan M, Cheng C.-H. Org. Lett. 2006; 8: 5581
• 25 Chatani N, Kamitani A, Oshita M, Fukumoto Y, Murai S. J. Am. Chem. Soc. 2001; 123: 12686
• 26 Feng M, Tang B, Xu H.-X, Jiang X. Org. Lett. 2016; 18: 4352
• 27 Xie C, Liu L, Zhang Y, Xu P. Org. Lett. 2008; 10: 2393
• 28 Bhuvaneswari S, Jeganmohan M, Cheng C.-H. Chem. Commun. 2008; 5013
• 29 Peng X, Ma C, Tung C.-H, Xu Z. Org. Lett. 2016; 18: 4154
• 30 Zeng Y, Li G, Hu J. Angew. Chem. Int. Ed. 2015; 54: 10773
• 31 Reiner BR, Tonks IA. Inorg. Chem. 2019; 58: 10508
• 32 Jayanth TT, Jeganmohan M, Cheng C.-H. Org. Lett. 2005; 7: 2921
• 33 Jayanth TT, Cheng C.-H. Angew. Chem. Int. Ed. 2007; 46: 5921
• 34 Qiu Z, Xie Z. Angew. Chem. Int. Ed. 2009; 48: 5729
• 35 Lin Y, Wu L, Huang X. Eur. J. Org. Chem. 2011; 2993
• 36 Yang Y, Huang H, Wu L, Liang Y. Org. Biomol. Chem. 2014; 12: 5351
• 37 Li R.-J, Pi S.-F, Liang Y, Wang Z.-Q, Song R.-J, Chen G.-X, Li J.-H. Chem. Commun. 2009; 46: 8183
• 38 Pi S.-F, Yang X.-H, Huang X.-C, Liang Y, Yang G.-N, Zhang X.-H, Li J.-H. J. Org. Chem. 2010; 75: 3484
• 39 Pi S.-F, Tang B.-X, Li J.-H, Liu Y.-L, Liang Y. Org. Lett. 2009; 11: 2309
• 40 Yao T, He D. Org. Lett. 2017; 19: 842
• 41 Liu Z, Larock RC. Angew. Chem. Int. Ed. 2007; 46: 2535
• 42 Yoshida H, Honda Y, Shirakawa E, Hiyama T. Chem. Commun. 2001; 1880
• 43 Tang C.-Y, Wu X.-Y, Sha F, Zhang F, Li H. Tetrahedron Lett. 2014; 55: 1036
• 44 Garve LK. B, Werz DB. Org. Lett. 2015; 17: 596
• 45 Zeng Y, Hu J. Org. Lett. 2016; 18: 856
• 46 Zeng Y, Zhang L, Zhao Y, Ni C, Zhao J, Hu J. J. Am. Chem. Soc. 2013; 135: 2955
• 47 Jeganmohan M, Cheng C.-H. Synthesis 2005; 1693
• 48 Henderson JL, Edwards AS, Greaney MF. J. Am. Chem. Soc. 2006; 128: 7426
• 49 Helmchen G, Pfaltz A. Acc. Chem. Res. 2000; 33: 336
• 50 Behenna DC, Stoltz BM. J. Am. Chem. Soc. 2004; 126: 15044
• 51 Margalef J, Biosca M, de la Cruz Sánchez P, Faiges J, Pàmies O, Diéguez M. Coord. Chem. Rev. 2021; 446: 214120
• 52 Connon R, Roche B, Rokade BV, Guiry PJ. Chem. Rev. 2021; 121: 6373
• 53 Blackham EE, Booker-Milburn KI. Angew. Chem. Int. Ed. 2017; 56: 6613
• 54 Sun Z.-M, Zhang J, Zhao P. Org. Lett. 2010; 12: 992
• 55 Coeffard V, Guiry PJ. Curr. Org. Chem. 2010; 14: 212
• 56 Endo K, Grubbs RH. J. Am. Chem. Soc. 2011; 133: 8525
• 57 Paradiso V, Costabile C, Grisi F. Beilstein J. Org. Chem. 2018; 14: 3122
• 58 Lu C, Dubrovskiy AV, Larock RC. J. Org. Chem. 2012; 77: 8648
• 59 Aleti RR, Festa AA, Voskressensky LG, Van der Eycken EV. Molecules 2021; 26: 5560
• 60 Denman BN, Plasek EE, Roberts CC. Organometallics 2023; 42: 859
• 61 Jover J, Cirera J. Dalton Trans. 2019; 15036
• 62 Tolman CA. Chem. Rev. 1977; 77: 313
• 63 Coll DS, Vidal AB, Rodríguez JA, Ocando-Mavárez E, Añez R, Sierraalta A. Inorg. Chim. Acta 2015; 436: 163
• 64 Sigman MS, Miller JJ. J. Org. Chem. 2009; 74: 7633
• 65 Zahrt AF, Athavale SV, Denmark SE. Chem. Rev. 2020; 120: 1620
• 66 Guan Y, Buivydas TA, Lalisse RF, Attard JW, Ali R, Stern C, Hadad CM, Mattson AE. ACS Catal. 2021; 11: 6325
• 67 Saint-Denis TG, Lam NY. S, Chekshin N, Richardson PF, Chen JS, Elleraas J, Hesp KD, Schmitt DC, Lian Y, Huh CW, Yu J.-Q. ACS Catal. 2021; 11: 9738
• 68 Mantilli L, Gérard D, Torche S, Besnard C, Mazet C. Chem. Eur. J. 2010; 16: 12736
• 69 García-López J.-A, Greaney MF. Org. Lett. 2014; 16: 2338

Publikationsverlauf

Eingereicht: 01. Juli 2023

Angenommen nach Revision: 31. Juli 2023

Artikel online veröffentlicht:
18. September 2023