Keywords

6.1 Introduction

Hikizimycin (1, also known as anthelmycin, Scheme 6.1), a highly oxygenated nucleoside natural product isolated from Streptomyces A-5 and Streptomyces longissimus [1, 2], displays potent anthelmintic activity against various common parasites and modest antibacterial properties. These biological activities originate from its inhibitory effects against pro- and eukaryotic ribosomal peptidyl transferase, which is essential for protein biosynthesis [3, 4]. Many nucleoside antibiotics are known to have a wide range of pharmacologically useful biological activities [5,6,7,8], suggesting that 1 may serve as a promising drug lead in the development of therapeutic agents [9,10,11,12].

Scheme 6.1
A series of chemical reactions leading to the formation of a compound labeled d galactose derivative has various skeletal structures and reactions leading to different intermediate compounds. Some processes involved are desymmetrization, chain extension, convergent assembly, and beta-elimination.

Structure of hikizimycin (1), Schreiber and Ikemoto’s total synthesis, and our synthetic plan

Full size image

The unique 4-amino-4-deoxyundecose sugar (hikosamine) of 1 contains one amino and ten hydroxy groups on its C1–C11 linear carbon chain [13]. Nucleobase (cytosine) and 3-amino-3-deoxyglucose (kanosamine) are appended to the hikosamine structure at the C1 and C6O positions, respectively, through glycosidic linkages [14, 15]. The densely functionalized C(sp3)-rich structure of 1 with its multiple stereocenters significantly heightens the synthetic challenge, which has been tackled by many research groups over the last half-century since its first isolation. Specifically, two major problems arise in the synthesis of 1: (1) stereoselective introduction of the ten contiguous stereocenters of the hikosamine structure and (2) site- and stereoselective attachment of the cytosine and kanosamine moieties to hikosamine. The first problem was solved by Secrist and Barnes [16], Danishefsky and Maring [17, 18], and Fürstner and Wuchrer [19], as well as by our group [20], culminating in four distinct syntheses of the hikosamine structure. However, none of these approaches addresses the second problem to complete the total synthesis of 1, thereby highlighting the associated difficulties. The second problem requires differentiation of the C6O group from the surrounding oxygen functionalities in order to attach kanosamine and activation of the C1 anomeric position in order to introduce cytosine. Schreiber and Ikemoto accomplished the first total synthesis of 1 by designing a differentially protected hikosamine intermediate [21, 22]. Remarkably, the team exploited the latent C2-symmetry embedded in the hikosamine structure, using pairwise functionalizations via iterative C–C and C–O bond formations to achieve total synthesis of 1 from l-diisopropyl tartrate (3) in 27 steps.

In 2020, we reported the convergent total synthesis of 1 from bis-TMS-cytosine (6) and three hexose derivatives (5, 11, and 12) [23]. We selected starting materials with preinstalled oxygen functional groups [24, 25] and increased the molecular complexity upon single coupling reaction [26,27,28], thereby minimizing the number of functional group manipulations and drastically shortening the synthetic route to 1 (longest linear sequence: 17 steps). To directly couple densely functionalized fragments derived from two hexoses 11 and 12, we devised a novel radical addition to an aldehyde [23, 29, 30]. We found that the specific protective group pattern of the synthetic intermediates facilitated key radical coupling in a stereoselective manner while also securing regio- and stereoselective introduction of the cytosine and kanosamine moieties. In this chapter, we detail the development of the radical-based and protective group strategies that enabled the total synthesis of 1. Interested readers can consult our review of total syntheses of hikosamine and 1 from our and other groups [31].

6.2 Synthetic Plan for Hikizimycin

We planned to assemble hikizimycin (1) from three components: bis-TMS-cytosine 6, kanosamine derivative 5, and differentially protected hikosamine 7-α (Scheme 6.1). The C1 acetal carbon center and C6 oxygen atom of 7-α were discriminated from other hetero functionalities to allow for selective appending of 6 and 5 by the two glycosylation reactions. In these reactions, neighboring participation from the proximal C2O and C13O benzoyl groups of 7-α and 5 would control the stereochemistry of the C1 and C12 positions. In principle, C6 alcohol 7-α would be directly coupled by an anionic reaction between anion A and aldehyde 8 or by a radical reaction between radical B and 8 [32]. The β-elimination propensity of the C4 nitrogen functional group from anion A and presence of electrophilic ester groups impeded us from adopting the anion-based approach [33]. Hence, we selected the radical-based approach [34].

Radical reactions serve as powerful tools for forging congested C(sp3)–C(sp3) bonds without affecting diverse oxygen and nitrogen functionalities. Therefore, they have been extensively utilized in the total synthesis of densely functionalized C(sp3)-rich natural products [28, 35,36,37,38,39,40]. We previously developed a decarbonylative radical reaction of α-alkoxyacyl telluride under Et3B/O2-mediated conditions and incorporated this powerful reaction into the synthesis of diverse densely oxygenated natural products [41, 42].

To date, intermolecular radical coupling reactions with aldehydes have been underexplored [43,44,45]. We recently realized Et3B/O2-mediated coupling between α-alkoxyacyl telluride 13 and aldehyde 14 [23] (Scheme 6.2). The reaction mechanism responsible for this unusual coupling was rationalized as follows. First, an Et radical was produced from Et3B and O2 [46], leading to C–Te bond cleavage, formation of acyl radical C, and decarbonylation to generate α-alkoxy radical D [47,48,49]. The polarity-matched intermolecular coupling between nucleophilic radical D and electrophilic aldehyde 14 was a fast but endergonic process [50, 51], generating oxyl radical E, which was higher in energy than D and generally underwent β-scission to readily reverse the process. However, the present reaction system overrode this energetically unfavorable step by capturing unstable E with Et3B to afford stable borinate F, the hydrolysis of which led to alcohol 16 in 40% yield. Consequently, Et3B played two important roles in this transformation: initiating and terminating the radical reaction. The unwanted compound 15 (47%) was also generated presumably through direct hydrogen atom abstraction by D [52]. The formation of D was suppressed by the addition of Lewis acid BF3·OEt2 to afford 16 (46%) and 15 (23%) [53].

Scheme 6.2
A series of chemical reactions leading to the formation of compounds labeled 15 and 16 has various skeletal structures and reactions leading to different intermediate compounds. Some compounds involved are labeled 13, C, D, E, and F.

Et3B/O2-mediated decarbonylative radical coupling reaction of α-alkoxyacyl telluride 13 to aldehyde 14

Full size image

This outcome prompted us to select the Et3B/O2 reagent system and α-alkoxyacyl telluride 10 as a precursor of radical B (Scheme 6.1). The C5 and C6 stereochemistries of 7-α needed to be controlled in the coupling reaction with 8. We envisioned to install the requisite stereocenters through strategic placement of the protective groups (PGs) of both 8 and 10 [20, 42]. To enable this challenging task, we decided to specify the appropriate protective group patterns by screening the substrates. Compounds 8 and 10 would be prepared from d-mannose (12) and d-galactose derivative 11, respectively, which together carried the six stereocenters of 1 (C2, C3, C7, C8, C9, and C10). Maximum use of the intrinsic chiralities of the two hexoses would streamline the route to 1 [24, 25]. Overall, our strategy was designed to maximize convergency and minimize the number of functional group transformations toward 1.

6.3 Protective Group Optimization of Aldehyde 8

We set out to optimize the protective groups of the right-half aldehyde 8 using readily available 13 [54] as a model acyl telluride (Table 6.1). In doing so, penta-benzoate 8a and bis-acetonide 8b were prepared and submitted to Et3B/O2-mediated radical reactions. The radical coupling reaction between 13 and 8a afforded the mixture of C6 alcohol 17a and C7 alcohol 18, which was generated via 1,2-benzoyl migration from 17a (entry 1). Despite the modest yield (17a: 36% and 18: 5.5%) and low C6 diastereoselectivity (α/β = 1:1.1 for 17a), a densely oxygenated carbon chain with nine consecutive stereogenic centers was notably generated by this single transformation. The yield and C6 stereoselectivity were improved by altering the substrate from 8a to 8b (entry 2). Submission of 13 and 8b to Et3B and air at −30 °C produced a 2.1:1 diastereomeric mixture of 17b in 66% yield. In contrast to coupling of the simple aldehyde 14 (Scheme 6.2), the addition of BF3·OEt2 to aldehyde 8b had a negative effect on the reaction efficiency and the yield of 17b was negligible (2.5%), presumably due to the acid-labile protective groups of 8 and 17b. Hence, neutral conditions omitting Lewis acids were to be applied for the total synthesis of 1.

Table 6.1 Intermolecular radical coupling of α-alkoxy carbon radicals with the right-half aldehyde 8
Full size table

The model radical reactions shown in Table 6.1 simultaneously installed C5 and C6 stereocenters. The stereochemical outcomes were attributable to the three-dimensional (3D) shapes of the radical donor (13) and acceptors (8a/8b) with the protective groups (Scheme 6.3). C–C bond formation proceeded from the convex face of the acetonide-protected 5/5-cis-fused ring system of α-alkoxy radical D, thereby establishing the C5 stereocenter [54]. Meanwhile, introduction of the C6 stereogenic center was attributed to the preferred conformations of acceptors 8a and 8b. The Felkin–Ahn transition states 8a-α and 8a-β were energetically similar, and both accepted the radical to lead to comparable amounts of C6 diastereomers 17a-α and 17a-β. In contrast, severe steric interaction occurred with transition state 8b-β between D and the methyl group of the 6/6-cis-fused ring system (highlighted in gray), resulting in higher energy than 8b-α, which led to the requisite C6α stereochemistry. Therefore, the unique 3D structure of 8b forced by the two acetonide groups was likely to induce the desired C6α selectivity of 17b-α.

Scheme 6.3
2 reactions for the formation of compounds 17 a beta and 17 b beta. Top. The compounds 8 a alpha and 8 a beta react reversibly with each other to form the compounds 17 b alpha and 17 b beta, respectively. Bottom. The compounds 8 beta alpha and 8 b beta react reversibly to form 17 b alpha and 17 b beta.

Rationale for the stereochemical outcomes of the coupling reactions between radical D and aldehydes 8a/8b

Full size image

6.4 Protective Group Optimization of Acyl Telluride 10

Next, we systematically altered the structure of the left-half acyl telluride 10 to attain the requisite C5 stereochemistry upon radical reaction. Namely, four substrates, 10a-α, 10b-α, 10b-β, and 10c-β, possessing distinct C4-substituted nitrogen functionalities (trifluoroacetamide, phthalimide, and azide) and different C1 stereochemistries were prepared as radical precursors (Scheme 6.4). To investigate C5 stereoselectivity, methyl vinyl ketone (19) was employed as an acceptor [54]. When C4 trifluoroacetamide 10a-α with C1α stereochemistry was utilized in the presence of Et3B in air, only the undesired C5β stereoisomer 20a-α was produced (Scheme 6.4a). Simply altering the C4 substituent from the trifluoroacetamide of 10a-α to the phthalimide of 10b-α exclusively generated the desired C5α stereoisomer 21b-α in high yield (76%, Scheme 6.4b) [20]. Intriguingly, the C1 stereochemistry affected the reaction efficiency and stereoselectivity. The reaction of C1 epimeric C4 phthalimide 10b-β provided the desired adduct 21b-β (25%) along with ketone 22b-β (38%), generated through direct addition of the acyl radical intermediate (Scheme 6.4c). The application of C4 azide 10c-β with the C1β stereocenter resulted in the formation of ketone 22c-β (30%) and radical elimination of the azide group to produce 3,4-dihydro-2H-pyran 23c-β (58%) [55, 56] (Scheme 6.4d). Accordingly, we selected 10b-α as the optimal radical donor for the total synthesis of 1.

Scheme 6.4
A four-part series of chemical reactions leading to the formation of compounds labeled 20 a alpha, 21 b alpha, 21 b beta and 22 b beta, 23 c beta, and 22 c beta has various skeletal structures and reactions leading to different intermediate compounds.

Decarbonylative radical coupling of the left-half glucosamine-derived α-alkoxyacyl telluride 10 with distinct C4N protective group and C1 stereochemistry

Full size image

These disparate radical reaction results can be rationalized by analyzing the 3D shapes of the radical intermediates (Scheme 6.5). First, Et3B/O2-induced homolysis of the weak C–Te bonds of 10a-α/10b-α/10b-β gave rise to acyl radicals with chair conformations Ga-αʹ/Gb-αʹ/Gb-βʹ and equilibrating boat conformations Ga-α/Gb-α/Gb-β. In the case of the C1α isomers, boat forms Ga-α and Gb-α had stabilizing secondary orbital interactions between the C5–CO σ*-orbital and the antiperiplanar oxygen lone pair and were thus more energetically preferred over Ga-αʹ and Gb-αʹ (Scheme 6.5a, b). The same orbital interaction between the C5–CO σ*-orbital and the oxygen lone pair also facilitated C5–CO scission via decarbonylation to generate α-alkoxy C5 radicals Ba-α and Bb-α [47,48,49]. The boat conformations of Ba-α and Bb-α were stabilized because the C5 radical interacted with the parallel oxygen lone pair and C4–N σ*-orbital [57, 58]. Enone 19 approached the upper face of Ba-α with the assistance of an intermolecular hydrogen bond between the C4 trifluoroacetamide of Ba-α and the carbonyl group of 19 to establish the unwanted C5β stereochemistry of 20a-α [59] (Scheme 6.5a). In contrast, the bulky C4 phthalimide of Bb-α shielded the top face of Bb-α, forcing coupling with 19 from the opposite face to establish the desired C5α stereocenter of 21b-α (Scheme 6.5b). Alternatively, acyl radical Gb-β with C1β stereochemistry was an unstable conformer due to steric repulsion between the C1 methoxy and C4 phthalimide substituents (Scheme 6.5c). The preferable chair conformer Gb-βʹ lacked the parallel relationship between the C5–CO σ*-bond and the oxygen lone pair. Due to slow C5–CO cleavage, decarbonylation toward C5 radical Bb-β competed with direct interception of Gb-βʹ with 19, providing a mixture of 21b-β and ketone 22b-β. Thus, the C1α stereochemistry of 10b-α contributed to accelerate α-alkoxy C5 radical formation and stabilize the boat conformation, while the C4 phthalimide group controlled the desired C5α stereoselectivity.

Scheme 6.5
A three-part series of chemical reactions leading to the formation of compounds labeled 22 a alpha, 20 a alpha, 22 b alpha, and 21 b alpha has various skeletal structures and reactions leading to different intermediate compounds.

Rationale for the coupling reaction outcomes between acyl radical G and methyl vinyl ketone (19)

Full size image

6.5 Synthesis of Protected Hikizimycin

Having clarified the significance of the C1α methoxy and C4 phthalimide groups of 10b-α and the bis-acetonide structure of 8b in stereoselective radical coupling, we turned our attention to the synthesis of protected hikizimycin 36. In addressing this task, the chemical structures of the radical donor and acceptor needed further modification. To enable C1 cytosine introduction in the last synthetic stage, radical donor 10d-α was designed to have a more activated C1α acetoxy group instead of the C1α methoxy group of 10b-α (Scheme 6.6). Moreover, the tert-butyldiphenylsilyl group of radical acceptor 8b was replaced with the benzoyl group of 8c, which would be removed together with other nucleophile-sensitive protective groups such as the C2O and C3O benzoyl and C4 phthalimide groups.

Scheme 6.6
A series of chemical reactions leading to the formation of a compound labeled 8 c has various skeletal structures and reactions leading to different intermediate compounds like 25, 26, 27, 29, 30, and 31.

Preparation of fragments 10d-α and 8c

Full size image

The left-hand fragment 10d-α was prepared from commercially available d-galactose derivative 11-β in nine steps (Scheme 6.6). First, benzoylation of triol 11-β protected the equatorial C2 and C3 alcohols and left the axial C4 alcohol untouched to afford bis-benzoate 24. Triflation of alcohol 24 and subsequent C4 stereo-inversion using KNPhth installed the C4 phthalimide group of 25. Acidic treatment of 25 with Ac2O exchanged the methyl and trityl groups with acetyl groups, producing a 1:8.3 mixture of 26-α and 26-β. The obtained C1 diastereomeric mixture was subjected to MeOCHCl2 and ZnCl2 [60] to introduce the β-oriented chloride of 27. Hg(OAc)2 in AcOH then promoted replacement of the C1β chloride of 27 with the equatorial C1α acetoxy group of 26-α [61]. Site-selective i-Bu2AlH reduction of the least-hindered C6 acetoxy group of 26-α and subsequent AZADOL-catalyzed oxidation of the liberated primary alcohol of 28 furnished carboxylic acid 29 [62]. The requisite radical donor 10d-α was derivatized from 29 in a one-pot procedure through activation of the ester, followed by addition of i-Bu2AlTePh derived in situ from (PhTe)2 and i-Bu2AlH [63].

The right-hand fragment 8c was prepared from d-mannose (12) in six steps. The known three-step sequence converted 12 to dithioacetal 30 [64]. The TBDPS group of 30 was detached using TBAF to generate the primary alcohol, which was benzoylated with BzCl and pyridine to form benzoate 31. The dithioacetal of 31 was in turn hydrolyzed by the action of mercury salts, giving rise to aldehyde 8c.

Next, we turned our attention to the unprecedented and challenging radical coupling of the two densely oxygenated fragments 10d-α and 8c (Scheme 6.7), which involved individual optimization of many parameters, such as reagent amounts, order of reagent addition, method of oxygen addition, concentration, and temperature. Ultimately, 7-α and the minor C6 epimer 7-β were obtained in 65% combined yield (7-α/7-β = 2.2:1) when air was slowly introduced into a mixture of α-alkoxyacyl telluride 10d-α (1.0 equiv), aldehyde 8c (3.0 equiv), and Et3B (5.0 equiv) in CH2Cl2 (0.1 M) at −30 °C. Hence, the desired 7-α with its ten contiguous stereocenters was constructed by forging the hindered C(sp3)–C(sp3) bond under simple and mild conditions with C5 and C6 stereoselectivities. The observed C5 and C6 stereochemical outcomes were in accordance with those of the aforementioned preliminary investigations (Sects. 6.3 and 6.4).

Scheme 6.7
A series of chemical reactions leading to the formation of a compound labeled 36 a and b has various skeletal structures and reactions leading to different intermediate compounds labeled 33, 34 and 35, and 36 b.

Synthesis of protected hikizimycin 36

Full size image

C6 alcohol 7-α was then elaborated into protected hikizimycin 36 through stepwise attachment of bis-TMS-cytosine 6 and kanosamine derivative 5. Prior to these two glycosylation reactions, the protective groups of 7-α were manipulated in the ensuing three steps. First, the C6 alcohol was capped as the benzyl ether by the action of N-phenyl-2,2,2-trifluoroacetimidate and catalytic TfOH, yielding 32 [65]. Second, the two acid-labile acetonide groups of 32 were detached using BF3·OEt2 and 1,3-propane dithiol without damaging the potentially reactive C1 acetal structure [66]. Third, the resultant tetraol was peracetylated to produce pentaacetate 33. C1α acetate 33 and bis-TMS-cytosine 5 were subjected to TMSOTf to produce C1α benzoyl cytosine 34 as a single isomer after in situ N-benzoylation. Complete C1α stereoselectivity was attributable to neighboring participation of the β-oriented C2O benzoyl group to form the stabilized intermediate H.

To prepare for the second glycosylation, the C6 hydroxy group of 35 was liberated in a site-selective manner by DDQ-induced oxidation of the C6O benzyl group of 34. Then, we investigated the reactivity of the differentially N-protected glycosyl donors 5a and 5b [67]. Treatment of C14 phthalimide 5a with TMSOTf in the presence of 35 merely led to decomposition. In contrast, TMSOTf-promoted glycosylation of 35 with C4 azide 5b smoothly afforded protected hikizimycin 36b with the C12β kanosamine moiety as the major isomer (36b/12-epi-36b = 1.8:1) [68]. The requisite C12β stereoselectivity would be controlled by neighboring participation of the α-oriented C13O benzoyl group, while the different reactivities of 5a and 5b were explained by the boat-like 3D structures of cationic intermediates Ia and Ib, respectively. Whereas the axially oriented bulky phthalimide of Ia impeded intermolecular attack of 35, the smaller azide group of Ib accepted hindered C6 secondary alcohol 35 upon glycosylation. Therefore, the requisite C1α–N and C12β–O glycosidic bonds were stereoselectively constructed via the strategically placed protective groups.

6.6 Total Synthesis of Hikizimycin

The last task of the total synthesis of hikizimycin (1) from 36b was removal of the seven benzoyl, four acetyl, and one phthaloyl groups and hydrogenolysis of the C14 azide moiety. Because of the nucleophilic and electrophilic sites of the highly complex structure of 36b, the reaction order and conditions needed to be carefully tuned. Our preliminary model experiment revealed sluggish hydrogenolysis of the sterically cumbersome C4 azide of des-cytosine derivative 37 (Scheme 6.8). Even when C4 amine was generated, the nucleophilic nitrogen atom of 38 attacked the C=O bond of the proximal benzoyl group to induce 1,2-benzoyl migration, generating N-benzoylated products 39 and 40. The resistance of the amide bonds of 39 and 40 to basic hydrolysis required suppression of ester–amide exchange. This reaction indicated that removal of all acyl protective groups must precede azide hydrogenolysis.

Scheme 6.8
A chemical reaction leads to the formation of a compound labeled 38, 39, and 40 with the reactant labeled 37. The reaction undergoes the reduction of C 14 N 3.

Azide hydrogenolysis of des-cytosine model compound 37

Full size image

Next, we investigated simultaneous hydrolysis of multiple acyl groups of 36b (Scheme 6.9a). When n-Bu4NOH was employed, the seven benzoyl and four acetyl groups were completely removed from 36b, but phthalimide hydrolysis only partially afforded phthalamic acid 41 in 96% yield after HPLC purification. To avoid unwanted hydrogenation of the cytosine ring, a Lindlar catalyst was utilized for the subsequent hydrogenolysis of azide 41, leading to amine 42. The remaining phthalamic acid of 42 was detached using a large excess amount of ethylenediamine in refluxing t-BuOH (boiling point: 82 °C) [69]. These harsh conditions of the last reaction caused partial decomposition of the electrophilic cytosine moiety, leading to pure hikizimycin (1) only being obtained in low yield from 41 (21% over two steps) after the crude mixture was purified with HPLC.

Scheme 6.9
2 reactions. a. The compound 36 b undergoes H P L C purification with n B u 4 N O H and M e O H to form compound 41, which further reacts with Lindlar catalyst to form compound 42, followed by hikizimycin. b. The 36 b undergoes H P L C purification to form compound 43, followed by 44 and hikizimycin.

a Three-pot protocol for converting protected hikizimycin 36b to hikizimycin (1). b One-pot protocol for converting protected hikizimycin 36b to hikizimycin (1)

Full size image

To develop a more efficient protocol, the phthalimide needed to be converted to the corresponding amine during the first hydrolysis step. Ultimately, we established a one-pot procedure to generate 1 (Scheme 6.9b). n-BuNH2 was expected to detach the phthaloyl group of 36b via the formation of phthalamic acid, following facile 5-exo cyclization of 43, with N-butyl phthalimide expelled [70]. Indeed, the addition of n-BuNH2 to refluxing MeOH (boiling point: 65 °C) removed all 12 protective groups to form primary amine 44. In the same pot, reduction of the C14 azide substituent of 44 was realized under an H2 atmosphere in H2O in the presence of the Lindlar catalyst. This optimized protocol improved reaction efficiency and delivered 1 from 36b in 50% yield after HPLC purification. Thus, the total synthesis of 1 from 11-β was completed in 17 steps.

6.7 Conclusion

In summary, we accomplished highly efficient total synthesis of hikizimycin (1) from 11-β (longest linear sequence: 17 steps). The exceptionally complex structure of 1 was constructed from three hexose derivatives and one cytosine structure (5b, 6, 11-β, and 12), without extra carbon extension or oxygen atom introduction. The radical-based and protective group strategies were specifically devised to enable the present novel convergent route to 1. First, highly oxygenated α-alkoxyacyl telluride 10d-α and aldehyde 8c were derivatized from hexose structures 11-β and 12, respectively, and then combined under newly developed Et3B/O2-mediated conditions for radical addition to the aldehyde. The mild, yet robust reaction linked two hindered trisubstituted carbons and installed the desired C5α and C6α stereocenters, thereby assembling the protected hikosamine structure 7-α with ten contiguous stereocenters. Notably, these stereochemical outcomes were controlled by the strategically placed C4 phthalimide and bis-acetonide groups. Subsequently, the cysteine moiety and kanosamine with the small C14 azide were attached by two TMSOTf-promoted reactions in a C1 and C12 stereoselective manner. Remarkably, the proximal benzoyl groups stereoselectively forged the C1α and C12β glycosidic bonds. Lastly, protected hikizimycin 36b was achieved in a one-pot by the removal of the 12 acyl protective groups and reduction of the one azide functionality, affording 1.

The present data corroborate the importance of radical coupling reactions for convergent assembly of two densely oxygenated fragments and the significance of strategically placed protective groups for precise control of reactivity and stereo- and site-selectivity. It is our hope that further application and improvement of the radical-based convergent synthetic approach will benefit advances in both chemical and pharmaceutical sciences.