The Formation of Impossible Rings in Macrocyclooligomerizations for Cyclodepsipeptide Synthesis: The 18-from-12 Paradox

Abstract

A reinvestigation into the macrocyclooligomerization (MCO) of a tetradepsipeptide is reported, uncovering a paradox in which the MCO of depsipeptide monomers can produce “impossible” ring sizes: a 12-atom chain produced the expected 24-membered ring, alongside unexpected 18- and 30-membered cyclic oligomeric depsipeptides (CODs). We report an alternative preparation of authentic 18- and 36-membered macrocycles for this case using a stepwise synthesis that provides definitive analytical characterization for each ring size. Our investigation yields a recharacterization and reassignment of two macrocycles originally reported in this MCO series, along with updated yields and isothermal titration calorimetry data after implementation of new critical protocols for purification and subsequent analysis. Initial studies to probe this mechanistic conundrum are described.

Graphical Abstract

INTRODUCTION

We recently reported an approach to small collections of cyclic depsipeptides in a single reaction from simple monomers using a controlled oligomerization/macrocyclization process termed macrocyclooligomerization (MCO).^1,2 The salient features of this process were the use of didepsipeptides and tetradep-sipeptides to prepare (macro)cyclic oligomeric depsipeptides (CODs) with 18, 24, 36, and higher rings sizes. Two distinct advantages were highlighted: the formation of (typically) three CODs in significant yield with a remarkably small unused mass balance and rapid access to large ring sizes in as few as six steps from readily available starting materials. A Mitsunobu reaction was leveraged for the chain elongation and macrocyclization steps, and evidence for ion templation (sodium, potassium, and cesium) was documented for several cases.

These studies were originally driven by our interest in (−)-verticilide,³ an insect ryanodine receptor (RyR) antagonist (Figure 1A, 24-membered COD from ent-1).⁴ Our discovery that ent-verticilide, but not nat-verticilide, is a potent and selective inhibitor of mammalian RyR2⁵ led us to wonder whether the lower and higher oligomers of this structure might also exhibit activity. To accomplish this while developing a size-selective synthesis of the (unnatural) 36-membered oligomer (Figure 1, 36-membered COD), we explored a traditional stepwise synthesis strategy. We report here the successful realization of this initial goal and the finding that the data originally assigned to this oligomer is consistent instead with the 18-membered oligomer. Importantly, such a product is ostensibly precluded by the MCO of a tetradepsipeptide that should produce only ring sizes that are multiples of 12 (i.e., 12-, 24-, 36-, and 48-membered rings). Our in-depth study of this paradox has led to the conclusion that, in this case,⁶ for reasons not entirely clear, a competitive chain-cleaving step occurs to produce the unexpected 18- and 30-membered rings (Figure 1B). This study provides incontrovertible data for each COD ring in this tetradepsipeptide MCO case and tests several mechanistic hypotheses for its formation. Evidence suggests that a depsipeptide fragmentation contributes the needed monomer to access the paradoxical ring sizes.

Figure 1.

Macrocyclooligomerization (MCO) approach to cyclic oligomeric depsipeptides (CODs): (A) partial revision of original findings and (B) documenting the formation of paradoxical ring-size oligomers.

RESULTS AND DISCUSSION

Preparation and investigation of molecules with sizes larger than most “small molecule” drugs but smaller than typical “small peptide” drugs have drawn increasing attention. This size regime has adopted the moniker “beyond the rule of five”, alluding to Lipinski’s rules⁷ but highlighting the potential for discovery when considering this largely unexplored chemical space. This endeavor is not without limits, including the need for concise access to conformationally well-defined molecules that might also exhibit favorable drug-like properties. Our effort has targeted cyclic depsipeptides for these reasons, culminating in the finding that small depsipeptide monomers could be converted to cyclic oligomeric depsipeptides by a straightforward oligomerization/cyclization process using the Mitsunobu reaction. This approach was further refined by the addition of Lewis acids with the anticipation that the cyclization step might be templated and therefore accelerated relative to further chain elongation.

Our use of this approach complements substantial prior work by others who used an MCO strategy to prepare oligomeric esters and amides. Most notable is the use of activated seco-acids to form macrodiolides (and triolides)^8–14 and amidations of activated peptidic amino acids to construct the macrocyclic peptides valinomycin and westiellamide.^15–17 A common feature to these amidation reactions is a five-membered heterocycle, such as a thiazole or oxazole, installed in the peptide backbone.^18–27 This presumably enhances a conformation more favorable for cyclization.

A specific tetradepsipeptide (1, Figure 1) was noted to form unusually large cyclic depsipeptides that were 36- and 60-membered rings in size, alongside the 24-membered ring. Assignments to these ring sizes were made on the basis of fractionation in the workflow of purification, and the observation of the corresponding mass ions, albeit with relatively weak intensity. This case was reinvestigated to better understand the factors leading to larger ring-size production. A key factor was assumed to be the length of the depsipeptide monomer, with the tetradepsipeptide (1, contributing 12 backbone atoms) leading to allowed ring sizes that included 24-, 36-, 48-, and 60-membered rings. A route to the 36-membered ring using a stepwise chain-extension sequence and final macrocyclization was planned and executed to acquire larger amounts of this target. The first chemical synthesis of nat-verticilide employed a linear route in which the stereochemistry of the α-hydroxyheptanoic acid moiety was installed using a chiral auxiliary, requiring six steps overall to form the didepsipeptide.⁴ They recently developed a second-generation synthesis, which uses serine as a starting material to shorten the step-count to five steps.²⁸ Our synthesis of the hydroxyheptanoic acid leveraged enantioselective catalysis and Umpolung Amide Synthesis to arrive at the didepsipeptide in three steps.

Our approach relies on O-deprotected benzyl ester 4 and carboxylic acid 6 prepared using optimized procedures of those previously reported (Scheme 1).¹ These didepsipeptide units were coupled using a Mitsunobu protocol to form the tetradepsipeptide 7 (Scheme 2). This material was divided, with half that was O-deprotected to 9, and half that was hydrogenolyzed to 8 prior to segment coupling [4+4] using another Mitsunobu reaction. The MOM-protecting group of the resulting octadepsipeptide 10 was removed, and the resulting alcohol was reacted with tetradepsipeptide carboxylic acid 8 using Mitsunobu conditions. This furnished the dodecadepsipeptide 11 in 60% yield. This material was deprotected at both termini, subjected to a Mitsunobu macrocyclization protocol (1 mM concentration), and then isolated as a 36-membered COD (12) (70% for three steps).

Scheme 1.

Enantioselective Synthesis of Differentially Protected and Unprotected Didepsipeptide Monomers

Scheme 2.

Linear, Stepwise Synthesis of the 36-Membered Cyclic Dodecadepsipeptide.

Analysis of 12 revealed a ¹H NMR spectrum with the characteristic number and type of peaks in this COD series, but it was distinct from the ¹H NMR spectrum of the material originally assigned to the 36-membered ring (Figure 1A). More telling was the observation of an intense parent ion peak ([M + Na]⁺ 1217.7146) by HRMS consistent with a 36-membered COD. This parent ion was far more intense than that observed in the original material (Figure 2). A second distinguishing feature in the HRMS spectrum for the authentic 36-membered ring is the absence of a singly charged 620 Da ion [M¹⁸ + Na]⁺. Instead, the ion peak at 620.3518 corresponds to [M³⁶ + 2Na]²⁺.

Figure 2.

Comparison of HRMS data from originally assigned “‘36” COD (A) formed by MCO and authentic 36-membered COD (B) formed through stepwise linear synthesis. [exact mass calculated for [M³⁶ + Na]⁺ = 1217.7148; [M¹⁸ + Na]⁺ = 620.3523; [M³⁶ + 2Na]²⁺ = 620.3523].²⁹

The successful synthesis of 12 using a stepwise route provided compelling data to fingerprint the 36-membered COD 12. Given that all other aspects of the spectroscopic data were consistent with a highly symmetrical oligomeric depsipeptide, the assignment of this material was considered sound, and the identity of the material originally assigned as the 36-membered COD was reevaluated. We ultimately postulated that the original material could be an 18-membered COD, an “impossible” product of the 12-mer oligomerization (Figure 1B). Data consistent with this assignment included (a) a high degree of symmetry by ¹H NMR, indicating a product of the Mitsunobu reaction that produces homochiral α-oxy amide subunits, and (b) intense adducts of the parent ions including [M + H]⁺ at 598 and [M + Na]⁺ at 620 consistent with an 18-membered COD (Figure 2). Verification of this structure by independent synthesis was straightforward, leveraging a stepwise route analogous to the synthesis of the cyclo-dodecadepsipeptide 12. Hence, treatment of an O-deprotected tetradepsipeptide 9 with a benzyl deprotected didepsipeptide 6 under Mitsunobu conditions yielded the linear hexapeptide in 93% yield (Scheme 3). Final MOM deprotection, hydrogenolysis, and cyclization under dilute conditions gave the authentic 18-membered COD in a 75% yield over three steps. The ¹H NMR (A, Figure 3) and HRMS (Figure 2) data mirrored the data associated with the compound originally assigned as the 36-membered ring (D, Figure 3). It should be noted that broadening and slight shifting of α-oxy amide methine peaks in the proton NMR spectra are evident when any amount of trifluoroacetic acid (TFA) from HPLC purification is present (see discussion on residual TFA below and in the Supporting Information). Even small amounts of TFA appear to affect the chemical shift and coupling, as illustrated in Figure 3, where the originally characterized 18-membered ring (spectrum C) appears reasonably distinct from the other samples (spectra B and D). This difference contributed to the initial belief that two different compounds were forming in the MCO reactions. We have since confirmed that upon complete removal of TFA, by exhaustive base washes, the spectra overlap (see the Supporting Information). From this data, we are able to conclude that the 18-membered COD is the major product when using didepsipeptide monomer.

Scheme 3.

Linear, Stepwise Synthesis of the 18-Membered Cyclic Hexadepsipeptide.

Figure 3.

Full-scale ¹H NMR spectrum of (A) the authentic 36-membered COD (12, Scheme 2), (B) an authentic sample of the 18-membered COD (16, Scheme 3), (C) the originally assigned 18-membered ring formed in the MCO reaction,³⁰ and (D) the originally assigned 36-membered ring formed in the MCO reaction (Figure 1A). Inset: expansion from 5.2 to 4.4 ppm.³¹

Although it is a minor COD formed from the tetradepsipeptide MCO, the 18-membered COD is clearly formed. The 36-membered COD is not formed in either MCO reaction, and attempts to change the reaction conditions to favor its formation have been unsuccessful to date.

A stepwise synthesis of the 60-membered ring was considered but not pursued. However, a parallel investigation of the data led to the conclusion that this material is most likely another paradoxical product: the 30-membered COD. This reassignment is made based on an analysis of HRMS data exhibiting a more intense singly charged mass ion for the 30-membered ring relative to the mass ion corresponding to a 60-membered COD, as well as MS/MS data (see the Supporting Information). Figure 4 illustrates our final assignments as well as the general similarities and differences of each ring-size congener. The use of ¹H NMR analysis can clearly distinguish each COD, but the irregular movement of specific peak types undermines its utility as a predictor of ring size.

Figure 4.

Full-scale ¹H NMR spectra of authentic samples of all COD rings sizes.

High-Resolution Mass Spectrometry Investigation.

At the time of the original investigation, the HRMS data supported the conclusions but required the perceived mechanistic limitations of a 12-atom monomer leading to lengths of only 24/36/48/60 to be decisive. In every spectra of the 18-membered ring and “36”-membered ring formed in the MCO reactions, only singly charged ion peaks at both 620 and 1217 Da were observed. For an 18-membered ring, these ion peaks would correspond to [M¹⁸ + Na]⁺ and [2M¹⁸ + Na]⁺. For a 36-membered ring, these would correspond to [1/2M³⁶ + Na]⁺ and [M³⁶ + Na]⁺.³² However, in all cases, the ion at 620 Da was more intense. Only after the synthesis of the authentic samples of each ring size through a stepwise route was a clear difference apparent in the MS spectra (Figure 5). To further probe behaviors for these oligomers by HRMS, MS/MS experiments were applied to authentic samples of the 18- and 36-membered rings.

Figure 5.

Comparison of the HRMS spectra and corresponding MS/MS spectrum of 620 and 1217 ion peaks in each sample of authentic 18-membered and 36-membered rings (z = charge).

When the 620 Da ion is fragmented in each experiment, the results are remarkably different. From the 18-membered ring, the singly charged 620 Da ions gave no fragments above m/z of 620. However, when the doubly charged 620 Da ion for 12 [M³⁶+Na]²⁺ was fragmented, an intense 1217 ion was observed, corresponding to [M³⁶ − Na]⁺. Additionally, there were a handful of fragments from the ring itself. When fragmenting the 1217 Da ion in each sample, again the MS2 patterns were distinct. The MS-MS of the 1217 ion from 16 showed the largest ion at 620 Da. This would correspond to the dissociation of the noncovalent dimer [2M¹⁸ + Na]⁺ with loss of one macrocycle (M¹⁸). Perhaps most intriguing is the presence of fragments heavier than 620 Da. This would imply that the non-covalent interaction that forms the singly charged dimer ion is rather strong as some of these dimers are experiencing fragmentation of the covalent bonds in the backbone of the macrocycle first.

Rationalizing the 18-from-12 Paradox.

The formation of an 18-membered COD from a 12-atom chain could result from several general pathways: (1) the presence of shorter monomers as impurities that co-oligomerize (e.g., 12+6) during MCO, (2) tetradepsipeptide monomer (12-atom chain) decomposition during reaction but pre-MCO, generating 6-atom chains, and (3) decomposition of oligomers to 6-atom monomers that then react with tetradepsipeptide monomer (e.g., 6+12). In the synthesis of these macrocyclic depsipeptides, the key oligomerization step occurs through a Mitsunobu reaction in which the stereochemistry of the α-hydroxy amide is inverted. When probing the different mechanistic possibilities for simple tetradepsipeptide cleavage, almost all of the options result in two linear peptides, heterochiral at the terminal α-hydroxy amide carbon. If this occurred, the isolated macrocycles would be diastereomers heterochiral at one α-oxy amide, and the dissymmetry would likely be evident by ¹H NMR. Furthermore, these stereochemical analogues are separable by HPLC purification based on our previous work.¹ Throughout this reinvestigation, the combined yields of the three major MCOs were high, suggesting that any chain cleavage process was likely not random.

To rule out the impure starting material as the source of adventitious didepsipeptide, the tetradepsipeptide was purified via reversed-phase HPLC. There was no indication of didepsipeptide during this purification, and the material purified in this manner produced the same array of products from MCO (Scheme 4, eq 1). HPLC-purified tetradepsipeptide was subjected to the same purification again, and no HPLC-induced decomposition was noted. We next tested the hypothesis that residual water in the reaction may cleave the central ester bond in the tetradepsipeptide in the MCO reaction or, separately, affect the level of possible anhydride species formed during the reaction. The formation of anhydrides in the Mitsunobu reaction has been studied,³³ and this pathway would be expected to form non-symmetrical CODs that we have yet to detect. Reactions with the tetradepsipeptide under dry conditions (Scheme 4, eq 2) or in the presence of a variety of drying agents did not decrease the amount of 18-membered ring or 30-membered ring formed in the reaction. Moreover, when varying amounts of water were added to the reaction (ranging from 0.25 to 5 equivalents), it yielded only the starting material (no macrocycles) (Scheme 4, eq 3). To explore the possibility that DIAD or PPh₃ could cause cleavage of the tetradepsipeptide, we added each reagent to tetradepsipeptide in benzene for 12 h. Both of these experiments returned only unreacted tetradepsipeptide with no sign of decomposition (Scheme 4, eqs 4 and 5). Additionally, no decomposition was noted after stirring the starting material alone in benzene for several days. Hypothesizing that the rings could decompose after formation and then recyclize to make smaller macrocycles, the authentic 36-membered ring was re-subjected to MCO reaction conditions (Scheme 4, eq 6). Again, no sign of decomposition was observed, and the starting material was recovered quantitatively. Finally, an experiment was investigated in which partially protected MOMO–tetradepsipeptide–CO₂H (8) was subjected to MCO reactions to determine if it was self-cleaving from the carboxylic acid terminus concomitant with the MCO reaction (Scheme 5). While this experiment did result in mostly recovered starting material (65%), the protected didepsipeptide 6 was isolated by HPLC purification in 13% yield.³⁴

Scheme 4.

Summary of Experiments Probing the Correlation of Conditions to the Formation of Paradoxical Ring-Size CODs.

Scheme 5.

Evidence That MCO Conditions Can Promote Depsipeptide Self-Cleavage.

The observation of protected depsipeptide instability under the typical Mitsunobu reaction conditions employed in MCO provides a tangible basis on which to consider the 18-from-12 paradox. The formation of 6 from 8 does not readily explain the formation of the 18-membered ring (16) since an analogous cleavage of 1 would lead to 5 and epi-5. While didepsipeptide 5 would give the expected heterochiral CODs (l,d), didepsipeptide epi-5 would lead to either homochiral CODs (d,d) or an epimer of 12 or 16, both distinguishable by ¹H NMR. However, our best rationale for the formation of 6 from 8 (Scheme 6) involves carboxylate addition to the central ester, collapse of this tetrahedral intermediate (a hemiorthoester)³⁵ to the anhydride isomer 18, and then either anhydride hydrolysis, or similar transacylation to form an active ester. Despite our careful reexamination, this is our best evidence that small amounts of the linear depsipeptide chain are self-cleaving under the MCO reaction conditions.

Scheme 6.

Mechanistic Hypothesis for C-Terminal Didepsipeptide Self-Cleavage from the Tetradepsipeptide Prior to MCO.

This occurrence, however, raises the question of why evidence of diastereomer formation is not observed. While the protected depsipeptide 6 is observed, didepsipeptide epi-5, where the α-hydroxy chiral center has already been inverted in a Mitsunobu reaction, should accompany it. The α-hydroxy amide configuration, therefore, is a label to determine the source of an MCO product; the expected heterochiral CODs (l,d) result from single-inversion Mitsunobu reactions. While the exact mechanism of cleavage is not certain, we outline a general pathway to explain this. These mechanistic proposals all stem from the formation of an anhydride (e.g., 18). In one scenario, the pendant alcohol resulting from the anhydride formation could intramolecularly cleave the anhydride, resulting in diketomorpholine 19 and the dipeptide (5) with the correct stereochemistry to participate in the MCO reaction to generate the heterochiral CODs (l,d) observed. The 18-and 30-membered ring formation support the presence of didepsipeptide 5, which could react with the tetradepsipeptide and octadepsipeptide, respectively, to form the requisite COD linear precursors (pre-18, pre-30). Unfortunately, we have yet to isolate any diketomorpholine products from these reactions, and it is not clear if any are present in the crude reaction mixtures (¹H NMR) for these reactions, given their complexity.³⁶ We speculate that this occurs with the tetradepsipeptide, but it is plausible that an analogous cleavage process could occur on a variety of linear depsipeptides.

Alternatively, a C-terminal didepsipeptide cleavage could occur on any linear oligomer (12, 24, 36, etc.). If an anhydride were to form at the C-terminal end (similar intermediate to 17 in Scheme 6), one might imagine a disproportionation mechanism, in which a single didepsipeptide unit is transferred to another linear depsipeptide (Figure 6). This possibility provides an explanation why the diketomorpholine, homochiral macrocycles, and diastereomeric macrocycles are not observed under these reaction conditions. This type of mechanism could be related to a competitive transesterification process responsible for chain scission in the polymerization of hydroxy acids.^37,38

Figure 6.

Possible disproportionation mechanism to paradoxical ring sizes.

Isothermal Titration Calorimetry and Correlations to Ring-Size Influences during MCO by Salt Additives.

The preparation of authentic 36-membered COD 12 and the reassignment of mass initially attributed to 12, but reallocated to 18-membered COD 16, demanded a reevaluation of the isothermal titration calorimetry (ITC) measurements from this series. We noted that our reported ion binding measurements for these macrocycles exhibited very different patterns for the 18-membered ring and putative 36-membered ring.² With the knowledge that these compounds should be identical, we sought an understanding of the measurement difference with a prejudice toward possible contaminant(s). This careful analysis led to the finding that some of the original samples used for ITC contained (small) residual amounts of trifluoroacetic acid (TFA) from reversed-phase HPLC purification. Detection of this residual TFA was possible only via ¹⁹F NMR since the ¹H NMR of these samples can appear identical when there are only small amounts of TFA in the sample. Furthermore, this finding was unexpected since all of these samples had been subjected to a standard base wash in a work-up after the HPLC purification. Hence, these macrocycles exhibit a strong propensity to bind TFA.

Armed with this knowledge, the original macrocycle samples were extensively base-washed until TFA removal was confirmed by 19F NMR, prior to repeating the ITC experiments. Truly, TFA-free samples labeled 18- or “36”-membered ring from the MCO reactions then provided the same reproducible isotherms (see the Supporting Information).

A complete series of TFA-free samples of 18-, 24-, 30-, and 36-membered rings were next analyzed by ITC. This data is provided in Figure 7. Overall, while the remeasured data still delineates cases of strong, selective binding, the binding data for this particular depsipeptide does not support⁴⁰ a templating effect with sodium, potassium, or cesium. The benefit of adding cesium is clear, however, as the combined yield of three CODs is 86%, compared to 57% without cesium salt additive. This increase in yield results from the significant increase in production of the paradoxical 18-membered ring (46% vs 29%).

Figure 7.

Comparison of MCO reaction yields (di- and tetradepsipeptide monomers) and ITC data (corrected). Association constant (K_a), binding stoichiometry (N), and enthalpy of binding (ΔH either positive (+) or negative (−)) were determined by ITC using methanol solutions, whereas MCO was performed in benzene. See the Supporting Information for complete details. K salt is KSCN for ITC measurements and KCl for MCO reaction conditions, Na is NaPF₆ for ITC measurements, NaBF₄ for MCO reaction conditions, and Cs salt is CsCl for both ITC data and MCO reaction conditions.³⁹ A value of <1.00 × 10³ indicates that no binding was observed, and therefore fields denoted with N/A (not applicable) were not measured. ITC data for the 36-membered ring was collected using material prepared via the synthesis in Scheme 2 since this ring size is not formed in MCO reaction. Yields are averages of three or more replicates for each case, based on a theoretical yield of 18.3 mg (from didepsipeptide) and 19.1 mg (from tetradepsipeptide). The maximal observed yield variation was 12% (~2 mg) in the didepsipeptide series and 13.5% (2.6 mg) in the tetradepsipeptide series. ITC data for the 24-membered ring is the same as that previously reported;² these macrocycle samples contained no TFA and were replicated 10 times.

As we noted originally, the ITC measurements could only be comprehensively obtained by the use of methanol solvent. This contrasts the use of benzene for MCO reactions, although the salt additives are used at saturating concentrations in benzene. Attempts to prepare ITC runs in other organic solvents or solvent mixtures led to insufficient solubility of either the salts or the salt–macrocycle complexes. At the time of initial investigation with this series, we noted that the solvent can have a strong effect on the magnitude of binding affinity; however, the goal was to observe possible trends in relative binding interactions. The trends observed in methanol, with the corrected ring sizes, still hold true. While it is disappointing that a correlation of affinity and MCO selectivity cannot be concluded for this particular MCO reaction (Figure 1A),⁴⁰ the potential ion binding abilities of this series of CODs remain unchanged.

We have carefully reviewed other series of MCO reactions and have no reason to suspect TFA impurities in any other cases. First and foremost, there are no other series of macrocycles that display the same data that lead us to believe there was a mischaracterization. Each COD reported in other series have very distinct NMRs and significantly different HPLC retention times. For ITC data, one factor that indicated the presence of TFA in samples was isotherms that exhibited atypical S-shaped curves. While this is characteristic of aggregate formation at the beginning of titration^41,42 and our initial suspicion, we now know that TFA impurity was responsible. No other series of COD isotherms appear to display this behavior. Additionally, some samples used in original experiments were still on hand and confirmed to be TFA-free by ¹⁹F NMR. The potential application of ion binding in future chemistry remains promising, and this study highlights an important consideration for any depsipeptide MCO.

The ability of residual TFA, despite routine base wash of HPLC fractions, to adversely affect COD behavior is evident. While this was a problem specifically with depsipeptides here, routine ¹⁹F NMR analysis should be a consideration when using reversed-phase HPLC purification buffered with TFA. Several patterns of behavior have been documented here, which could be essential when ¹⁹F NMR is not accessible.

CONCLUSIONS

In summary, we have re-examined the single case in which we isolated a material from an MCO reaction, utilizing a tetradepsipeptide, with the original assignment as a 36-membered ring. Through synthesis of an authentic 36-membered ring using a stepwise route and analysis of its product, we have reassigned the “original 36” material to be the 18-membered ring, a size not possible by direct oligomerization of a tetradepsipeptide (12-atom chain). While its relative stereochemistry confirms that it is formed during an MCO, its size is paradoxical since a hexadepsipeptide cannot form directly from a tetradepsipeptide. Our reanalysis of this case illuminates the paradox and explores several possible mechanisms, narrowing to an elongation that succumbs to an unusually competitive self-cleavage step. Importantly, the additive effects that have been shown to template certain MCO reactions are re-evaluated in this case, highlighting the possible insensitivity⁴⁰ of this specific tetradepsipeptide conversion to added sodium or potassium salts.

EXPERIMENTAL SECTION

General Information.

Glassware was flame-dried under vacuum for all non-aqueous reactions. All reagents and solvents were commercial grade and purified prior to use when necessary. Benzene and dichloromethane (CH₂Cl₂) were dried by passage through a column of activated alumina as described by Grubbs.⁴³ Flash column chromatography was performed using Sorbent Technologies 230–400 mesh silica gel with solvent systems indicated. Analytical thin layer column chromatography was performed using Sorbent Technologies 250 μm glass backed UV254 silica gel plates and were visualized by fluorescence upon 250 nm radiation and/or the by use of ceric ammonium molybdate (CAM), phosphomolybdic acid (PMA), or potassium permanganate (KMnO₄). Solvent removal was affected by rotary evaporation under vacuum (~25–40 mm Hg). All extracts were dried with MgSO₄ or NaSO₄ unless otherwise noted. Preparative HPLC was performed on an Agilent 1260 system (column: Zorbax Eclipse XDB-C18; 21.2 mm × 150 mm, 5 μm, flow rate 8 mL/min) with a 210 nm monitoring wavelength and acetonitrile/water (+0.1% TFA) gradient as indicated. Nuclear magnetic resonance (NMR) spectra were acquired on a Bruker AV-400 (400 MHz), Bruker DRX500 (500 MHz), or Bruker AV II-600 (600 MHz) instrument. Chemical shifts are measured relative to residual solvent peaks as an internal standard set to δ 7.26 and δ 77.16 (CDCl₃), unless otherwise specified. Mass spectra were recorded on a high-resolution Thermo Electron Corporation MAT 95XP-Trap mass spectrometer by use of the ionization method noted by the Indiana University Mass Spectrometry Facility. IR spectra were recorded on a Nicolet Avatar 360 spectrophotometer and are reported in wave-numbers (cm⁻¹) as neat films on a NaCl plate (transmission). Melting points were measured using an OptiMelt automated melting point system (Stanford Research Systems) and are not corrected. Chiral HPLC analysis was conducted on an Agilent 1100 series Infinity instrument using the designated ChiralPak column. Optical rotations were measured on either a Jasco P-2000 polarimeter or an AUTOPOL III (Rudolph Research Analytical) polarimeter. ITC measurements were performed using a Microcal PEAQ-ITC instrument (MicroCal Inc. Northampton, MA). All experiments were performed at 25 °C in anhydrous methanol. Optimal settings for consistent results in methanol include reference power set to 5 μcal/mol, stir speed set to 1000 rpm, and instrument feedback set to low.

General Procedures.

Methoxymethylene Ether (MOM) Deprotection.

A flame-dried round-bottom flask was charged with the depsipeptide (1 equiv) dissolved in dry dichloromethane (0.05 M), pentamethyl benzene (3 equiv), and BF₃·Et₂O (3 equiv). The reaction was allowed to stir for 50 min at ambient temperature. The crude reaction mixture was quenched with satd aq NaHCO₃, washed with brine, dried, concentrated, and subjected to flash column chromatography to afford the alcohol.

Benzyl Deprotection.

A round-bottom flask was charged with the depsipeptide (1 equiv) dissolved in methanol (0.1 M) and treated with 10% Pd/C (10 mol %). The reaction flask was evacuated with a light vacuum (~40 Torr). Hydrogen (balloon) was added, and then the flask was cycled through a light vacuum three times. The reaction was stirred for 1.5 h. The crude reaction mixture was filtered through Celite and concentrated to afford the carboxylic acid.

Didepsipeptide MCO.

A flame-dried round-bottomed flask under inert atmosphere was charged with seco-acid (1 equiv), PPh₃ (6 equiv), and benzene (0.02 M). DIAD (5 equiv) was then added to the stirred solution in 15 aliquots over 120 min. The reaction was allowed to stir at ambient temperature for 24 h and then concentrated to afford a residue that was subjected to the MCO purification protocol.

Tetradepsipeptide MCO.

A flame-dried round-bottomed flask under inert atmosphere was charged with seco-acid (1 equiv), PPh₃ (3 equiv), and benzene (0.005 M). DIAD (2.5 equiv) was then added to the stirred solution in 5 aliquots over 40 min. The reaction was allowed to stir at ambient temperature for 24 h and then concentrated to afford a residue that was subjected to the MCO purification protocol.

General MCO Purification Information.

The crude residue was purified following the procedure reported previously with minor changes.¹ The crude residue was filtered through a silica gel plug (2 cm × 9 cm) to remove excess Mitsunobu reagents. A stepwise MeOH/DCM gradient was used, and the mixture was collected into two fractions (Fraction 1: 0.5–1% MeOH in DCM; Fraction 2: 20% MeOH in DCM). After analysis by ¹H NMR, fraction 1 was discarded as Mitsunobu byproducts only. Fraction 2 was subjected to HPLC purification. Preparative HPLC was performed on an Agilent 1260 system (column: Zorbax Eclipse XDB-C18; 21.2 mm × 150 mm, 5 μm, 8 mL/min flow rate) with 210, 254, and 198 nm monitoring wavelength. The gradient used is listed in the Supporting Information. HPLC fractions containing depsipeptides (macrocyclic or linear) were then subjected to extractive workup with ethyl acetate. These compounds are sensitive to cleavage under the high-heat conditions necessary to remove the acidic water–acetonitrile solvent system via rotary evaporation. Additionally, these compounds retain polar solvents and TFA, so the washes (water for acidic depsipeptides and satd aq NaHCO₃ for non-acidic depsipeptides) are necessary for complete removal.

Experimental and Characterization Data for Reported Compounds.

((S)-2-((((R)-2-Hydroxyheptanoyl)-d-alanyl)oxy)-heptanoyl)-d-alanine (1).

Following the general benzyl deprotection procedure, the benzyl protected tetradepsipeptide (405 mg, 799 μmol) afforded the acid (312 mg, 95%) as a pale-yellow oil. All spectral data are in agreement with literature values.¹

(R)-2-(Methoxymethoxy)-1-nitroheptane (2).

Following the Evans protocol,⁴⁴ IndaBOX ((3aR,3a′R,8a-S,8a′S)-2,2′-(propane-2,2-diyl)bis(3a,8a-dihydro-8H-indeno-[1,2-d]oxazole)) (321 mg, 897 μmol) and Cu(OAc)₂·H₂O (162 mg, 813 μmol) were stirred at ambient temperature in isopropanol (32.6 mL) for 1 h. The cerulean blue solution was then cooled to 0 °C, and hexanal (2.00 mL, 16.3 mmol) was added and allowed to stir for 10 m before nitromethane (9.95 mL, 163 mmol) addition. After stirring for 4 days at ambient temperature, the reaction was quenched dropwise at 0 °C with 1 N HCl and the aqueous layer was extracted with CH₂Cl₂. Following drying and concentration under reduced pressure, the crude alcohol was dissolved in CHCl₃ (81.6 mL), treated with P₂O₅ (23.1 g, 163 mmol) and dimethoxymethane (33.9 mL, 326 mmol), and stirred at ambient temperature overnight. The reaction mixture was diluted with DCM and decanted from the solid. The organic layer was then washed with NaHCO₃ and brine. The organic layers were dried and concentrated to afford an oil that was subjected to flash column chromatography (SiO₂, 3–6% diethyl ether in hexanes) to afford the title compound as a pale-yellow oil (2.54 g, 76%, 2 steps). All spectral data are in agreement with literature values.¹

Benzyl ((R)-2-(Methoxymethoxy)heptanoyl)-d-alaninate (3).

A round-bottom flask was charged with 2 (289 mg, 1.41 mmol), benzyl d-alaninate (515 mg, 2.87 mmol), DME (5.2 mL), and H₂O (127 μL, 7.04 mmol). The mixture was then treated with DBTCE (550 mg, 1.69 mmol) followed by NaI (21.1 mg, 141 μmol), K₂CO₃ (397 mg, 2.87 mmol), and O₂ (balloon). The heterogeneous solution was vigorously stirred for 2 days at ambient temperature and then treated at 0 °C with 1 N HCl and poured into CH₂Cl₂. The aqueous layer was extracted with CH₂Cl₂, and the combined organic layers were washed with satd aq Na₂S₂O₃, dried, and concentrated. The crude residue was subjected to flash column chromatography (SiO₂, 15% ethyl acetate in hexanes) to afford the amide as an orange solid (231 mg, 47%).⁴⁵ All spectral data are in agreement with literature values.¹

Benzyl ((R)-2-Hydroxyheptanoyl)-d-alaninate (4).

Following the general MOM deprotection procedure, 3 (1.07 g, 3.02 mmol) afforded a crude pale-yellow solid. Flash column chromatography (SiO₂, 15–40% ethyl acetate in hexanes) afforded the alcohol (805 mg, 87%) as a pale-yellow oil. All spectral data are in agreement with literature values.¹

((R)-2-Hydroxyheptanoyl)-d-alanine (5).

Following the general benzyl deprotection procedure, 4 (441 mg, 1.44 mmol) afforded the deprotected acid (310 mg, 99%) as a pale-yellow oil. All spectral data are in agreement with literature values.¹

((R)-2-(Methoxymethoxy)heptanoyl)-d-alanine (6).

Following the general benzyl deprotection procedure, 5 (776 mg, 2.21 mmol) afforded the deprotected acid (566 mg, 99%) as a pale-yellow oil. All spectral data are in agreement with literature values.¹

Benzyl ((S)-2-((((R)-2-(Methoxymethoxy)heptanoyl)-d-alanyl)-oxy)heptanoyl)-d-alaninate (7).

A flame-dried round-bottom flask was charged with PPh₃ (1.04 g, 3.94 mmol), DIAD (775 μL, 3.94 mmol), and benzene (39.4 mL). The reaction was allowed to stir for 30 min at ambient temperature. Compound 4 (604 mg, 1.97 mmol) was added followed by 6 (566 mg, 2.17 mmol), and the reaction was allowed to stir for 24 h. The crude reaction mixture was concentrated and subjected to flash column chromatography (SiO₂, 15–60% ethyl acetate in hexanes) to afford the tetradepsipeptide (857 mg, 80%) as a pale-yellow solid. All spectral data are in agreement with literature values.¹

((S)-2-((((R)-2-(Methoxymethoxy)heptanoyl)-d-alanyl)oxy)-heptanoyl)-d-alanine (8).

Following the general benzyl deprotection procedure, the protected tetradepsipeptide (200 mg, 363 μmol) afforded the acid (165 mg, 99%) as a pale-yellow oil. All spectral data are in agreement with literature values.¹

Benzyl ((S)-2-((((R)-2-Hydroxyheptanoyl)-d-alanyl)oxy)-heptanoyl)-d-alaninate (9).

Following the general MOM deprotection procedure, the protected tetradepsipeptide (500 mg, 908 μmol) afforded a crude pale-yellow solid. Flash column chromatography (SiO₂, 15–40% ethyl acetate in hexanes) afforded the alcohol (405 mg, 88%) as a pale-yellow oil. All spectral data are in agreement with literature values.⁵

Benzyl ((S)-2-((((S)-2-((((S)-2-((((R)-2-(Methoxymethoxy)-heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alaninate (10).

A flame-dried round-bottom flask was charged with PPh₃ (85.5 mg, 326 μmol), DIAD (64.2 μL, 326 μmol), and benzene (3.3 mL). The reaction was allowed to stir for 30 min at ambient temperature. Compound 9 (82.6 mg, 163 μmol) was added, followed by 8 (82.6 mg, 179 μmol), and the reaction was allowed to stir for 24 h. The crude reaction mixture was concentrated and subjected to flash column chromatography (SiO₂, 20–40% ethyl acetate in hexanes) to afford the octadepsipeptide (124 mg, 80%) as a colorless oil. All spectral data are in agreement with literature values.⁵

Benzyl ((S)-2-((((S)-2-((((S)-2-((((S)-2-((((S)-2-((((R)-2-(Methoxy-methoxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)-heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alaninate (11).

A flame-dried round-bottom flask was charged with PPh₃ (56.7 mg, 216 μmol), DIAD (43 μL, 216 μmol), and benzene (2.2 mL). The reaction was allowed to stir for 30 m at ambient temperature. The octadepsipeptide free alcohol (98.0 mg, 108 μmol) was added followed by 8 (54.8 mg, 119 μmol), and the reaction was allowed to stir for 24 h. The crude reaction mixture was concentrated and subjected to flash column chromatography (SiO₂, 20–50% ethyl acetate in hexanes) to afford the dodecadepsipeptide (108 mg, 72%) as a colorless oil. ${\left[\alpha \right]}_{\text{D}}^{20}-23\left(c0.86,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.37 (40% EtOAc/hexanes); IR (film) 3311, 2930, 2862, 1752, 1657, 1539, 1456, 1380, 1311, 1195, 1155 cm⁻¹; ¹H NMR (600 MHz, CDCl₃) δ 7.75 (d, J = 5.3 Hz, 2H), 7.71 (d, J = 4.1 Hz, 1H), 7.66 (d, J = 6.0 Hz, 1H), 7.39 (d, J = 7.4 Hz, 1H), 7.36–7.28 (m, 5H), 7.00 (d, J = 5.7 Hz, 1H), 5.21–5.11 (m, 5H), 5.19 (d, J = 12.9 Hz, 1H), 5.13 (d, J = 12.4 Hz, 1H), 4.70 (dd, J = 8.8, 6.7 Hz, 2H), 4.63 (dq, J = 7.3, 7.3 Hz, 1H), 4.33–4.17 (m, 5H), 4.04 (dd, J = 6.3, 4.8 Hz, 1H), 3.42 (s, 3H), 1.96–1.69 (series of m, 12H), 1.50 (d, J = 7.4 Hz, 3H), 1.498 (d, J = 7.0 Hz, 3H), 1.496 (d, J = 7.1 Hz, 3H), 1.491 (d, J = 7.4 Hz, 3H), 1.486 (d, J = 7.2 Hz, 3H), 1.45 (d, J = 7.4 Hz, 3H), 1.41–1.27 (m, 36H), 0.90–0.84 (m, 18H); 13C{¹H} NMR (150 MHz, CDCl₃): Due to extensive overlap of methylene peaks, line listing is not given. Refer to the image of the ¹³C spectrum;⁴⁶ HRMS (ESI) m/z: [M + H]⁺ calcd for C₆₉H₁₁₅N₆O₂₀ 1347.8161; Found 1347.8120.

(3R,6S,9R,12S,15R,18S,21R,24S,27R,30S,33R,36S)-3,9,15,21,27-,33-Hexamethyl-6,12,18,24,30,36-hexapentyl-1,7,13,19,25,31-hexaoxa-4,10,16,22,28,34-hexaazacyclohexatriacontane-2,5,8,11,14-,17,20,23,26,29,32,35-dodecaone (12).

A flame-dried round-bottom flask was charged with S1 (10.0 mg, 8.2 μmol) dissolved in dichloromethane (100 μL). The seco-acid was diluted with benzene (8.2 mL) and PPh₃ (12.9 mg, 41 μmol) was added. DIAD (8.1 μL, 50 μmol) was added over 40 min in 5 aliquots. The reaction was allowed to stir at ambient temperature for 12 h. The crude mixture was concentrated and passed through a silica plug (F1: 0.5% MeOH/DCM, 1% MeOH/DCM; F2: 20% MeOH/DCM). Fraction 2 was purified by preparative HPLC (45–95% aqueous acetonitrile, 210 nm, flow rate: 20 mL/min, R_t = 9.8 m) to afford the 36-membered macrocycle (4.2 mg, 42%) as a colorless oil. ${\left[\alpha \right]}_{\text{D}}^{23}-13\left(c0.40,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.60 (4% MeOH/DCM); IR (film) 3316, 2955, 2926, 2857, 1750, 1657, 1544, 1457, 1379, 1193, 1156, 1064 cm⁻¹; ¹H NMR (400 MHz, CDCl₃) δ 7.25 (d, J = 7.8 Hz, 1H), 5.16 (dd, J = 8.5, 4.1 Hz, 1H), 4.45 (dq, J = 7.0, 7.0 Hz, 1H), 1.93–1.70 (series of m, 2H), 1.47 (d, J = 7.3 Hz, 3H), 1.40–1.28 (m, 6H), 0.87 (t, J = 6.7 Hz, 3H); ¹³C{¹H} NMR (150 MHz, CDCl₃) ppm 171.6, 170.6, 74.4, 48.7, 31.5, 31.4, 24.7, 22.5, 17.2, 14.1; HRMS (ESI) m/z: [M + H]⁺ calcd for C₆₀H₁₀₃N₆O₁₈ 1195.7323; Found 1195.7355.

Benzyl ((S)-2-((((S)-2-((((R)-2-(Methoxymethoxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alaninate (13).

A flame-dried round-bottom flask was charged with PPh₃ (97.5 mg, 372 μmol), DIAD (73.2 μL, 372 μmol), and benzene (3.7 mL). The reaction was allowed to stir for 30 min at ambient temperature. Compound 9 (94.0 mg, 186 μmol) was added followed by 6 (53.3 mg, 204 μmol), and the reaction was allowed to stir for 24 h. The crude reaction mixture was concentrated and subjected to flash column chromatography (SiO₂, 20–30% ethyl acetate in hexanes) to afford the hexadepsipeptide (130 mg, 93%) as a colorless oil. ${\left[\alpha \right]}_{\text{D}}^{20}+2.3\left(c1.42,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.33 (40% EtOAc/hexanes); IR (film) 3295, 2952, 2956, 2857, 1749, 1655, 1541, 1457, 1378, 1155 cm⁻¹; ¹H NMR (600 MHz, CDCl₃) δ 7.51 (d, J = 4.8 Hz, 1H), 7.36–7.30 (m, 5H), 7.34 (d, J = 7.0 Hz, 1H), 6.99 (d, J = 5.6 Hz, 1H), 5.20 (d, J = 12.4 Hz, 1H), 5.20–5.18 (m, 2H), 5.13 (d, J = 12.4 Hz, 1H), 4.70 (d, J = 6.7 Hz, 1H), 4.68 (d, J = 6.7 Hz, 1H), 4.62 (dq, J = 7.3, 7.3 Hz, 1H), 4.32 (dq, J = 7.0, 5.6 Hz, 1H), 4.31 (dq, J = 7.1, 5.4 Hz, 1H), 4.04 (dd, J = 6.4, 4.8 Hz, 1H), 3.42 (s, 3H), 1.97–1.71 (series of m, 6H), 1.48 (d, J = 7.2 Hz, 3H), 1.48 (d, J = 7.2 Hz, 3H), 1.44 (d, J = 7.3 Hz, 3H), 1.40–1.21 (br m, 18H), 0.89–0.85 (m, 9H); ¹³C{¹H} NMR (150 MHz, CDCl₃) ppm 173.6, 172.6, 172.5, 172.4, 170.5, 169.8, 135.8, 128.6, 128.3, 128.2, 96.5, 77.7, 74.4, 74.3, 67.0, 56.4, 49.4, 49.0, 48.1, 32.7, 31.6, 31.5, 31.39, 31.36, 29.8, 24.7, 24.5, 24.4, 22.6, 22.52, 22.47, 17.5, 16.9, 16.4, 14.12, 14.09, 14.0; HRMS (ESI) m/z: [M + Na]⁺ calcd for C₃₉H₆₃N₃NaO₁₁ 772.4360; Found 772.4361.

Benzyl ((S)-2-((((S)-2-((((R)-2-Hydroxyheptanoyl)-d-alanyl)oxy)-heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alaninate (14).

Following the general MOM deprotection procedure, the hexadepsipeptide (130 mg, 173 μmol) afforded a crude pale-yellow solid. Flash column chromatography (SiO₂, 20–40% ethyl acetate in hexanes) afforded the alcohol (104.0 mg, 85%) as a colorless oil. ${\left[\alpha \right]}_{\text{D}}^{20}-4.1\left(c1.27,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.20 (40% EtOAc/hexanes); IR (film) 3312, 2955, 2927, 2859, 1749, 1655, 1538, 1457, 1381, 1155 cm⁻¹; ¹H NMR (600 MHz, CDCl₃) δ 7.54 (d, J = 5.1 Hz, 1H), 7.37–7.30 (m, 5H), 7.30 (d, J = 7.6 Hz, 1H), 7.04 (d, J = 5.6 Hz, 1H), 5.20 (d, J = 12.4 Hz, 1H), 5.19–5.16 (m, 2H), 5.13 (d, J = 12.4 Hz, 1H), 4.63 (dq, J = 7.3, 7.3 Hz, 1H), 4.32 (dq, J = 7.0, 5.6 Hz, 1H), 4.31 (dq, J = 7.1, 5.4 Hz, 1H), 4.11 (ddd, J = 8.2, 8.2, 4.3 Hz, 1H), 2.85 (d, J = 4.9 Hz, 1H), 1.97–1.59 (series of m, 6H), 1.47 (d, J = 7.2 Hz, 3H), 1.47 (d, J = 7.1 Hz, 3H), 1.44 (d, J = 7.3 Hz, 3H), 1.40–1.22 (br m, 18H), 0.90–0.85 (m, 9H); ¹³C{¹H} NMR (150 MHz, CDCl₃) ppm 174.9, 172.7, 172.6, 172.4, 170.5, 169.9, 135.7, 128.7, 128.4, 128.2, 74.42, 74.36, 72.2, 67.0, 49.3, 49.0, 48.1, 34.6, 31.7, 31.6, 31.5, 31.40, 31.37, 24.69, 24.67, 24.5, 22.6, 22.52, 22.48, 17.5, 16.8, 16.5, 14.11, 14.08, 14.0; HRMS (ESI) m/z: [M + Na]⁺ calcd for C₃₇H₅₉N₃NaO₁₀ 728.4098; Found 728.4083.

((S)-2-((((S)-2-((((R)-2-Hydroxyheptanoyl)-d-alanyl)oxy)-heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanine (15).

Following the general benzyl deprotection procedure, the hexadepsipeptide (140 mg, 198 μmol) afforded deprotected acid (119 mg, 98%) as a colorless oil. ${\left[\alpha \right]}_{\text{D}}^{20}-0.7\left(c0.98,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.17 (80% EtOAc/hexanes); IR (film) 3310, 3073, 2930, 2863, 1749, 1655, 1540, 1457, 1380, 1203, 1155, 1064 cm⁻¹; ¹H NMR (400 MHz, (CD₃)₂SO) δ 8.30 (d, J = 7.3 Hz, 1H), 8.24 (d, J = 7.0 Hz, 1H), 7.93 (d, J = 5.6 Hz, 1H), 4.94 (dd, J = 7.6, 4.7 Hz, 1H), 4.91 (dd, J = 7.5, 4.9 Hz, 1H), 4.35 (dq, J = 7.1, 4.5 Hz, 1H), 4.34 (dq, J = 7.0, 4.6 Hz, 1H), 4.12 (dq, J = 7.0, 7.0 Hz, 1H), 3.88 (dd, J = 7.7, 4.0 Hz, 1H), 1.75–1.39 (series of m, 6H), 1.34 (d, J = 7.2 Hz, 6H), 1.33–1.16 (br m, 18H), 1.25 (d, J = 7.2 Hz, 3H), 0.87–0.83 (m, 9H), [OH and CO₂H not observed]; ¹³C{¹H} NMR (150 MHz, (CD₃)₂SO) ppm 174.6, 173.7, 171.8, 171.5, 169.5, 168.4, 73.6, 73.4, 70.6, 47.8, 47.73, 47.67, 34.1, 31.3 (2C), 31.1, 30.79, 30.77, 24.2, 24.0, 23.8, 22.0, 21.88, 21.86, 17.4, 17.0, 16.7, 13.9, 13.8 (2C); HRMS (ESI) m/z: [M + Na]⁺ calcd for C₃₀H₅₃N₃NaO₁₀ 638.3629; Found 638.3624.

(3R,6S,9R,12S,15R,18S)-3,9,15-Trimethyl-6,12,18-tripentyl-1,7,13-trioxa-4,10,16-triazacyclooctadecane-2,5,8,11,14,17-hexaone (16).

A flame-dried round-bottom flask was charged with the fully deprotected hexadepsipeptide (10.0 mg, 16.2 μmol), PPh₃ (12.8 mg, 48.7 μmol), and benzene (1.62 mL). DIAD (8.00 μL, 40.5 μmol) was added in 5 aliquots over 40 min. The reaction was allowed to stir at ambient temperature for 24 h. The crude reaction mixture was concentrated and subjected to the general MCO purification protocol to afford the 18-membered macrocycle (6.0 mg, 62%, R_t = 23.8 min) as an amorphous white solid. ${\left[\alpha \right]}_{\text{D}}^{23}+25\left(c0.95,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.30 (6% MeOH/DCM); IR (film) 3208, 3046, 2922, 2852, 1756, 1669, 1556, 1456, 1379, 1260, 1211, 1171, 1105 cm⁻¹; ¹H NMR (400 MHz, CDCl₃) δ 6.98 (d, J = 7.8 Hz, 1H), 5.13 (dd, J = 7.6, 4.7 Hz, 1H), 4.58 (dq, J = 7.4, 7.3 Hz, 1H), 1.93–1.74 (m, 2H), 1.47 (d, J = 7.1 Hz, 3H), 1.40–1.22 (series of m, 6H), 0.88 (t, J = 6.8 Hz, 3H); ¹³C{¹H} NMR (100 MHz, CDCl₃ ) ppm 171.6, 169.9, 75.0, 48.7, 31.6, 31.5, 24.7, 22.6, 17.5, 14.1; HRMS (ESI) m/z: [M + Na]⁺ calcd for C₃₀H₅₁N₃NaO₉ 620.3523; Found 620.3518.

((2S)-2-((((2S)-2-((((2S)-2-(((2-((((S)-2-((((R)-2-Hydroxyheptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)-heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanyl)oxy)heptanoyl)-d-alanine (S1).

A flame-dried vial was charged with N-H depsipeptide (48.0 mg, 35.6 μmol), pentamethyl benzene (26.4 mg, 178 μmol), and dry DCM (712 μL). BF₃·Et₂O (22 μL, 180 μmol) was then added to the reaction mixture. The reaction was allowed to stir at ambient temperature for 45 m, and it was then quenched by the dropwise addition of satd aq NaHCO₃. The aqueous layer was extracted with DCM. The organic layer was dried and concentrated. The crude reaction mixture was concentrated and subjected to flash column chromatography (SiO₂, 40% ethyl acetate in hexanes) to afford the alcohol as a colorless oil (39.4 mg, 30.2 μmol). This material was loaded into a vial and then treated with 10% Pd/C (39.4 mg, l mass equiv.) and dissolved in a 1:1 MeOH/CH₂Cl₂ mixture (302 μL). The reaction vial was evacuated and backfilled with hydrogen (balloon), and this cycle was repeated. The reaction was allowed to stir for 35 min. After purging the flask with nitrogen, the crude reaction mixture was filtered through Celite and concentrated. The reaction mixture was then purified by preparative HPLC (45–95% aqueous acetonitrile, 210 nm, flow rate: 20 mL/min, R_t = 9.8 m) to afford the depsipeptide (35 mg, 95% 2-step) as a yellow oil. ${\left[\alpha \right]}_{\text{D}}^{20}+2.0\left(c0.34,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.26 (4% MeOH/DCM); IR (film) 3315, 2928, 2861, 1752, 1657, 1541, 1457, 1380, 1196, 1155, 1065 cm⁻¹; ¹H NMR (600 MHz, CDCl₃) δ 7.74 (d, J = 4.8 Hz, 2H), 7.66 (d, J = 5.5 Hz, 1H), 7.64 (d, J = 6.2 Hz, 1H), 7.31 (d, J = 6.9 Hz, 1H), 7.12 (d, J = 5.3 Hz, 1H), 5.23 (dd, J = 8.1, 4.3 Hz, 1H), 5.20 (dd, J = 5.9, 4.0 Hz, 1H), 5.18 (dd, J = 4.9, 3.9 Hz, 1H), 5.16 (dd, J = 7.9, 3.8 Hz, 1H), 5.12 (dd, J = 8.1, 3.7 Hz, 1H), 4.55 (dq, J = 7.2, 7.2 Hz, 1H), 4.45 (dq, J = 6.9, 6.9, 1H), 4.38 (dq, J = 7.2, 7.2 Hz, 1H), 4.31 (dq, J = 7.2, 7.2 Hz, 2H), 4.23 (dq, J = 7.1, 7.0 Hz, 1H), 4.12 (dd, J = 7.9, 3.8 Hz, 1H), 1.97–1.76 (series of m, 12H), 1.67–1.59 (m, 1H), 1.50 (d, J = 7.2 Hz, 6H), 1.49 (d, J = 7.0 Hz, 3H), 1.48 (d, J = 7.9 Hz, 3H), 1.47 (d, J = 7.3 Hz, 3H), 1.46 (d, J = 7.2 Hz, 3H), 1.45–1.27 (m, 36H), 0.89 (t, J = 7.0 Hz, 3H), 0.88 (t, J = 6.9 Hz, 6H), 0.87 (t, J = 6.7 Hz, 6H), 0.86 (t, J = 6.8 Hz, 3H), [CO₂H not observed]; 13C{¹H} NMR (150 MHz, CDCl₃) Due to extensive overlap of methylene peaks, line listing is not given. Refer to the image of the ¹³C spectrum;⁴⁶ HRMS (ESI) m/z: [M + H]⁺ calcd for C₆₀H₁₀₅N₆O₁₉ 1213.7429; Found 1213.7387.

(3R,6S,9R,12S,15R,18S,21R,24S)-3,9,15,21-Tetramethyl-6,12,18,24-tetrapentyl-1,7,13,19-tetraoxa-4,10,16,22-tetraazacy-clotetracosan-2,5,8,11,14,17,20,23-octaone (S2).

Following the tetradepsipeptide MCO general procedure, seco-acid 1 (20.0 mg, 48.0 μmol) was stirred for 24 h at ambient temperature to afford a pale-yellow oil. Preparative HPLC following the general MCO purification protocol afforded the 24-membered macrocycle (8.1 mg, 44%, R_t = 24.8 min) as a white solid. All spectral data are in agreement with literature values.¹

(3R,6S,9R,12S,15R,18S,21R,24S,27R,30S)-3,9,15,21,27-Penta-methyl-6,12,18,24,30-pentapentyl-1,7,13,19,25-pentaoxa-4,10,16-,22,28-pentaazacyclotriacontane-2,5,8,11,14,17,20,23,26,29-decaone (S3).

Following the tetradepsipeptide MCO general procedure, seco-acid 1 (20.0 mg, 48.0 μmol) was stirred for 24 h at ambient temperature to afford a pale-yellow oil. Preparative HPLC following the general MCO purification protocol afforded the 30-membered macrocycle (2.8 mg, 15%, R_t = 35.6 min) as a colorless oil. ${\left[\alpha \right]}_{\text{D}}^{23}-15\left(c0.65,{\text{CHCl}}_{\text{3}}\right)$; R_f = 0.35 (4% MeOH/DCM); IR (film) 3315, 2957, 2924, 2853, 1747, 1664, 1547, 1458, 1380, 1260, 1199, 1158, 1064 cm⁻¹; ¹H NMR (400 MHz, CDCl₃) δ 7.10 (d, J = 6.4 Hz, 1H), 5.10 (dd, J = 8.1, 4.0 Hz, 1H), 4.45 (dq, J = 6.9, 6.9 Hz, 1H), 1.93–1.83 (m, 1H), 1.83–1.68 (m, 1H), 1.45 (d, J = 7.2 Hz, 3H), 1.41–1.20 (m, 6H), 0.88 (t, J = 6.6 Hz, 3H); ¹³C{¹H} NMR (100 MHz, CDCl₃) ppm 172.1, 170.4, 74.4, 48.7, 31.5, 31.4, 24.8, 22.6, 17.1, 14.1; HRMS (ESI) m/z: [M + Na]⁺ calcd for C₅₀H₈₅ NaN₅O₁₅ 1018.5934; Found 1018.5936.