Letter Cite This: Org. Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/OrgLett
Late-Stage Diversification of Phosphinic Dehydroalanine Pseudopeptides Based on a Giese-Type Radical C‑Alkylation Strategy Kostas Voreakos,† Laurent Devel,‡ and Dimitris Georgiadis*,† †
Downloaded via IDAHO STATE UNIV on April 1, 2019 at 18:45:18 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Department of Chemistry, Laboratory of Organic Chemistry, National and Kapodistrian University of Athens, Panepistimiopolis, Zografou, 15771 Athens, Greece ‡ CEA, Institut des Sciences du Vivant Frédéric Joliot, Service d’Ingénierie Moléculaire des Protéines (SIMOPRO), Université Paris-Saclay, Gif-sur-Yvette 91190, France S Supporting Information *
ABSTRACT: A straightforward, late-stage diversification strategy for the installation of side chains on readily accessible unsaturated phosphinopeptidic scaffolds based on a Giese-type addition of alkyl radicals has been investigated. Among different alternatives, the preferred methodology is operationally simple as it can be carried out in an open flask with no need for protection of acidic moieties. Direct application to the synthesis of SPPS-compatible building blocks or to longer peptides is also reported.
P
hosphinic peptides constitute a privileged class of Znmetalloprotease inhibitors, distinguished by their ability to exhibit enhanced potency and improved selectivity profiles by acting as transition-state analogues.1 Even though their value as efficient inhibitors of Zn proteases has been widely recognized,1−3 cumbersome synthetic routes and lack of attractive “late-stage” diversification strategies are crucial limiting factors for the wide screening of diverse structures during drug discovery.1a,4 A landmark achievement in the field of phosphinic peptides is the development of a practical protocol for the incorporation of phosphinic pseudodipeptide isosters within peptide sequences on solid phase,5 a strategy that has allowed the identification of many important inhibitors through combinatorial chemistry techniques (Figure 1).6 However, screening possibilities are restricted by the limited structural diversity of P1 and P1′ positions of available phosphinic building blocks which generally determines the affinity and specificity of a successful inhibitor. Postmodification of suitable phosphinic precursors may seem like a convenient alternative; nevertheless, only a few relevant examples have been reported in the literature.7 Most of them have found limited use, with the sole exception of a 1,3-dipolar cycloaddition protocol toward P1′ isoxazole-substituted phosphinic peptides (Figure 1),7d a method that has successfully delivered potent and selective inhibitors of several Zn-metalloproteases.3a,6b,8 The need to expand structural diversity beyond the isoxazole motif has prompted us to explore other postmodification possibilities that would combine mild conditions, simple manipulations, wide scope, and high efficiency. In this regard, we present herein a Giese-type free-radical C-alkylation © XXXX American Chemical Society
Figure 1. Generic structure of a phosphinic peptide.
strategy for the facile installation of alkyl side chains onto suitable dehydroalanine pseudopeptidic precursors. Recently, we reported on the chemistry of α-alkyl(aryl)idene-β-phosphinyl carboxylic derivatives and the possible synthetic applications rising from their electrophilic double bond.7a Indeed, in an early application of such scaffolds on a Received: March 8, 2019
A
DOI: 10.1021/acs.orglett.9b00857 Org. Lett. XXXX, XXX, XXX−XXX
Letter
Organic Letters derivatization synthetic scheme, α-methylidene-β-phosphinyl carboxylic derivatives (dehydroalanine phosphinic isosters, DPIs) were transformed to pseudocysteine analogues via conjugate addition of thiolates.7c,f However, even though heteroatom-centered nucleophiles seem to perform well, carbanions react sluggishly, raising important chemoselectivity issues.7f As an alternative to traditional two-electron synthetic approaches, the conjugate addition of carbon-centered radicals (Giese-type addition) to DPI precursors could allow for the development of a robust and practical late-stage diversification protocol. Although we were unable to trace any literature precedent concerning phosphorus-containing pseudopeptides, a similar rationale has been applied to amino acids,9,10 to oligopeptides,11 and very recently even to whole proteins, opening an avenue of seemingly unlimited opportunities for discoveries in chemical biology.12 Inspired by these developments, we embarked to explore the possibilities of a radical Calkylation protocol for the late-stage diversification of DPIs. Optimization of reaction conditions was performed using phosphinic derivatives of type 1,7a as model electrophilic radical traps. Table 1 summarizes the results of selected optimization experiments (for a complete list, see Table S1). Initially, aiming to an operationally simple protocol that would be efficient in room or lower temperature and therefore
compatible with thermally sensitive substrates, we explored Et3B/O2 as an initiator. Sibi and co-workers have successfully used tristrimethylsilylsilane (TTMSS) as an efficient H-donor/ chain carrier for Et3Β-initiated Giese-type additions, in combination with a Lewis acid and a large excess of Et3Β (5 equiv).13 Application of such forcing conditions, falling in the so-called “high-oxygen regime” as proposed recently by Curran and McFadden,14 was not effective in our model system, leading to a complex mixture (entry 1). However, replacement of TTMSS with tri-n-butyltin hydride (TBTH) afforded 45% of 2aa, as determined by 31P NMR (entry 2). MS and NMR analysis showed formation of oxidized byproducts (presumably due to addition of alkoxy radicals) that was completely suppressed after removing Lewis acid, raising the conversion for 2aa to 66% (entry 3). Further adjustment of conditions revealed a positive effect of TBTH excess to the overall efficiency (entries 4 and 5). This requirement was associated with the partial consumption of TBTH due to the formation of P−O−Sn adducts.15 Indeed, control NMR experiments showed broadening of 31P signal, typical of P−Sn coupling, as well as the disappearance of an acidic proton in 1H NMR (Scheme S2). 16 Consequently, phosphinic acid 1a is temporarily masked during the reaction by the bulky nBu3Sn “protecting” group. High isolated yields were also obtained for diacid 1c (entry 7), whereas for ester 1b similar yields were achieved with 1 equiv less of TBTH since phosphinic functionality is already protected (entry 9).17 Initiation by AIBN was also explored, and satisfactory conversions for both 1a and 1b were obtained only by using a large excess of TBTH/RX; however, starting material was never completely consumed (entries 8−10). On the contrary, only 2 equiv of TTMSS as a H-donor was enough to achieve high conversions to 2aa (entry 11), thus offering access to a tin-free alternative. This observation reflects the lower tendency of TTMSS to reduce alkyl radicals, as compared to TBTH.18 However, this property of TTMSS was accompanied by a lower propagating ability when bromides were used as a source of stabilized alkyl radicals (entry 12), instead of iodides (entry 13). Furthermore, TTMSS caused extensive cleavage of phosphinate ester 1b (entry 14), presumably due to in situ formation of silyl iodide species, which are well-known dealkylating reagents of phosphinates.19 NaBH3CN, another efficient H-donor in Giese-type additions,20 proved to be incompatible with 1a (entry 15). This may result from deprotonation of P−OH by NaBH3CN which repels the nucleophilic cyanoborane radical anion during H atom transfer.20a In line with this hypothesis, 1b reacted smoothly, leading to 93% conversion under optimized conditions (entry 17). Interestingly, t-BuI was incapable of efficiently generating 2bb (entry 18), which renders this protocol unsuitable for stabilized alkyl radicals. Having established the reactivity profile of type 1 phosphinic acceptors toward three different Giese-type addition protocols, we proceeded to the investigation of substrate scope with respect to the halide (Scheme 1). In most cases, the Et3B/air/ TBTH protocol (Method A) performed well, and tin byproducts were easily removed after saponification and aqueous workup. Method A was compatible with bromides, affording 2bc−2ec in good isolated yields. As expected, iodides offered higher flexibility, leading to high conversions to products of type 2, not only with Method A but also with Methods B (AIBN/TTMSS) and C (AIBN/NaBH3CN) (see Table S2 for a detailed comparison of methods). Except from
Table 1. Optimization of Giese-Type Addition to DPIsa
entry
1
initiator
RX (equiv)
1c,d 2c,e 3e 4 5f 6 7f 8 9 10 11 12 13 14 15
1a 1a 1a 1a 1a 1b 1c 1a 1a 1b 1a 1a 1a 1b 1a
Et3Be Et3B Et3B Et3B Et3B Et3B Et3B AIBN AIBN AIBN AIBN AIBN AIBN AIBN AIBN
i-PrI (10) i-PrI (3) i-PrI (3) i-PrI (3) i-PrI (4) i-PrI (4) i-PrI (4) i-PrI (3) i-PrI (6) i-PrI (6) i-PrI (3) t-BuBr (3) t-BuI (3) i-PrI (3) i-PrI (3)
16
1b
AIBN
i-PrI (3)
17
1b
AIBN
i-PrI (3)
18
1b
AIBN
t-BuI (3)
H-donor (equiv) TTMSS (5) TBTH (2) TBTH (2) TBTH (3) TBTH (3) TBTH (3) TBTH (3) TBTH (2) TBTH (5) TBTH (5) TTMSS (2) TTMSS (2) TTMSS (2) TTMSS (2) NaBH3CN (3) NaBH3CN (3) NaBH3CN (5) NaBH3CN (5)
2
NMR yieldb (%)
2aa