Preparation of the β2‐Homoselenocysteine Derivatives Fmoc‐(S)‐β2hSec(PMB)‐OH and Boc‐(S)‐β2hSec(PMB)‐OH for Solution and Solid‐Phase Peptide Synthesis

Fmoc‐β²hSer(^tBu)‐OH was converted to Fmoc‐β²hSec(PMB)‐OH in five steps. To avoid elimination of HSeR, the selenyl group was introduced in the second last step (Fmoc‐β²hSer(Ts)‐OAll→Fmoc‐β²hSec(PMB)‐OAll). In a similar way, the N‐Boc‐protected compound was prepared. With the β²hSe‐derivatives, 21 β²‐amino‐acid building blocks with proteinogenic side chains are now available for peptide synthesis.     

Preparation of Protected β2‐ and β3‐Homocysteine, β2‐ and β3‐Homohistidine,and β2‐Homoserine for Solid‐Phase Syntheses

The Ser, Cys, and His side chains play decisive roles in the syntheses, structures, and functions of proteins and enzymes. For our structural and biomedical investigations of β‐peptides consisting of amino acids with proteinogenic side chains, we needed to have reliable preparative access to the title compounds. The two β³‐homoamino acid derivatives were obtained by Arndt–Eistert methodology from Boc‐His(Ts)‐OH and Fmoc‐Cys(PMB)‐OH (Schemes 2–4), with the side‐chain functional groups' reactivities requiring special precautions. The β²‐homoamino acids were prepared with the help of the chiral oxazolidinone auxiliary DIOZ by diastereoselective aldol additions of suitable Ti‐enolates to formaldehyde (generated in situ from trioxane) and subsequent functional‐group manipulations. These include OH→O^tBu etherification (for β²hSer; Schemes 5 and 6), OH→STrt replacement (for β²hCys; Scheme 7), and CH₂OH→CH₂N₃→CH₂NH₂ transformations (for β²hHis; Schemes 9–11). Including protection/deprotection/re‐protection reactions, it takes up to ten steps to obtain the enantiomerically pure target compounds from commercial precursors. Unsuccessful approaches, pitfalls, and optimization procedures are also discussed. The final products and the intermediate compounds are fully characterized by retention times (t_R), melting points, optical rotations, HPLC on chiral columns, IR, ¹H‐ and ¹³C‐NMR spectroscopy, mass spectrometry, elemental analyses, and (in some cases) by X‐ray crystal‐structure analysis.     

β‐Peptide Conjugates: Syntheses and CD and NMR Investigations of β/α‐Chimeric Peptides,of a DPA‐β‐Decapeptide,and of a PEGylated β‐Heptapeptide

β³‐Peptides consisting of six, seven, and ten homologated proteinogenic amino acid residues have been attached to an α‐heptapeptide (all d‐ amino acid residues; 4 ), to a hexaethylene glycol chain (PEGylation; 5c ), and to dipicolinic acid (DPA derivative 6 ), respectively. The conjugation of the β‐peptides with the second component was carried out through the N‐termini in all three cases. According to NMR analysis (CD₃OH solutions), the (M)‐3₁₄‐helical structure of the β‐peptidic segments was unscathed in all three chimeric compounds (Figs. 2, 4, and 5). The α‐peptidic section of the α/β‐peptide was unstructured, and so was the oligoethylene glycol chain in the PEGylated compound. Thus, neither does the appendage influence the β‐peptidic secondary structure, nor does the latter cause any order in the attached oligomers to be observed by this method of analysis. A similar conclusion may be drawn from CD spectra (Figs. 1, 3, and 5). These results bode well for the development of delivery systems involving β‐peptides.     

Preparation of (S,S)‐Fmoc‐β2hIle‐OH, (S)‐Fmoc‐β2hMet‐OH,and (S)‐Fmoc‐β2hTyr(tBu)‐OH for Solid‐Phase Syntheses of β2‐ and β2/β3‐Peptides

The preparation of three new N‐Fmoc‐protected (Fmoc=[(9H‐fluoren‐9‐yl)methoxy]carbonyl) β²‐homoamino acids with proteinogenic side chains (from Ile, Tyr, and Met) is described, the key step being a diastereoselective amidomethylation of the corresponding Ti‐enolates of 3‐acyl‐4‐isopropyl‐5,5‐diphenyloxazolidin‐2‐ones with CbzNHCH₂OMe/TiCl₄ (Cbz=(benzyloxy)carbonyl) in yields of 60–70% and with diastereoselectivities of >90%. Removal of the chiral auxiliary with LiOH or NaOH gives the N‐Cbz‐protected β‐amino acids, which were subjected to an N‐Cbz/N‐Fmoc (Fmoc=[(9H‐fluoren‐9‐yl)methoxy]carbonyl) protective‐group exchange. The method is suitable for large‐scale preparation of Fmoc‐β²hXaa‐OH for solid‐phase syntheses of β‐peptides. The Fmoc‐amino acids and all compounds leading to them have been fully characterized by melting points, optical rotations, IR, ¹H‐ and ¹³C‐NMR, and mass spectra, as well as by elemental analyses.     

cyclo(β‐Asp‐β3‐hVal‐β3‐hLys) – Solid‐Phase Synthesis and Solution Structure of a Water Soluble β‐Tripeptide. Preliminary Communication

The H₂O‐soluble cyclic β³‐tripeptide cyclo(β‐Asp‐β³‐hVal‐β³‐hLys) ( 4 ) was obtained by on‐resin cyclization of the side‐chain‐anchored β‐peptide 3 (Scheme). In aqueous solution, 4 adopts a structure with uniformly oriented amide bonds and all side chains in lateral positions (Fig. 3).     

Synthesis of Halogenated α,β‐Unsaturated γ‐Butyrolactone Derivatives by Triphenylphosphine‐Catalyzed Cyclization of α‐Halogeno Ketones with Dialkyl Acetylenedicarboxylates (=Dialkyl But‐2‐ynedioates)

A Ph₃P‐catalyzed cyclization of α‐halogeno ketones 2 with dialkyl acetylenedicarboxylates (=dialkyl but‐2‐ynedioates) 3 produced halogenated α,β‐unsaturated γ‐butyrolactone derivatives 4 in good yields (Scheme 1, Table). The presence of electron‐withdrawing groups such as halogen atoms at the α‐position of the ketones was necessary in this reaction. Cyclization of α‐chloro ketones resulted in higher yields than that of the corresponding α‐bromo ketones. Dihalogeno ketones similarly afforded the expected γ‐butyrolactone derivatives in high yields.     

New Open‐Chain and Cyclic Tetrapeptides,Consisting of α‐, β2‐, and β3‐Amino‐Acid Residues,as Somatostatin Mimics – A Survey

Cyclo‐β‐tetrapeptides are known to adopt a conformation with an intramolecular transannular hydrogen bond in solution. Analysis of this structure reveals that incorporation of a β²‐amino‐acid residue should lead to mimics of ‘α‐peptidic β‐turns’ (cf. A, B, C ). It is also known that short‐chain mixed β/α‐peptides with appropriate side chains can be used to mimic interactions between α‐peptidic hairpin turns and G protein‐coupled receptors. Based on these facts, we have now prepared a number of cyclic and open‐chain tetrapeptides, 7 – 20 , consisting of α‐, β²‐, and β³‐amino‐acid residues, which bear the side chains of Trp and Lys, and possess backbone configurations such that they should be capable of mimicking somatostatin in its affinity for the human SRIF receptors (hsst_1–5). All peptides were prepared by solid‐phase coupling by the Fmoc strategy. For the cyclic peptides, the three‐dimensional orthogonal methodology (Scheme 3) was employed with best success. The new compounds were characterized by high‐resolution mass spectrometry, NMR and CD spectroscopy, and, in five cases, by a full NMR‐solution‐structure determination (in MeOH or H₂O; Fig. 4). The affinities of the new compounds for the receptors hsst_1–5 were determined by competition with [¹²⁵I]LTT‐SRIF₂₈ or [¹²⁵I] [Tyr¹⁰]‐CST₁₄. In Table 1, the data are listed, together with corresponding values of all β‐ and γ‐peptidic somatostatin/Sandostatin^® mimics measured previously by our groups. Submicromolar affinities have been achieved for most of the human SRIF receptors hsst_1–5. Especially high, specific binding affinities for receptor hsst₄ (which is highly expressed in lung and brain tissue, although still of unknown function!) was observed with some of the β‐peptidic mimics. In view of the fact that numerous peptide‐activated G protein‐coupled receptors (GPCRs) recognize ligands with turn structure (Table 2), the results reported herein are relevant far beyond the realm of somatostatin: many other peptide GPCRs should be ‘reached’ with β‐ and γ‐peptidic mimics as well, and these compounds are proteolytically and metabolically stable, and do not need to be cell‐penetrating for this purpose (Fig. 5).     

Synthesis of β3‐Peptides and Mixed α/β3‐Peptides by Thioligation

Five β‐peptide thioesters ( 1 – 5 , containing 3, 4, 10 residues) were prepared by manual solid‐phase synthesis and purified by reverse‐phase preparative HPLC. A β‐undecapeptide ( 6 ) and an α‐undecapeptide ( 7 ) with N‐terminal β³‐HCys and Cys residues were prepared by manual and machine synthesis, respectively. Coupling of the thioesters with the cysteine derivatives in the presence of PhSH (Scheme and Fig. 1) in aqueous solution occurred smoothly and quantitatively. Pentadeca‐ and heneicosapeptides ( 8 – 10 ) were isolated, after preparative RP‐HPLC purification, in yields of up to 60%. Thus, the so‐called native chemical ligation works well with β‐peptides, producing larger β³‐ and α/β³‐mixed peptides. Compounds 1 – 10 were characterized by high‐resolution mass spectrometry (HR‐MS) and by CD spectroscopy, including temperature and concentration dependence. β‐Peptide 9 with 21 residues shows an intense negative Cotton effect near 210 nm but no zero‐crossing above 190 nm, (Figs. 2–4), which is characteristic of β‐peptidic 3₁₄‐helical structures. Comparison of the CD spectra of the mixed α/β‐pentadecapeptide ( 10 ) and a helical α‐peptide (Fig. 5) indicate the presence of an α‐peptidic 3.6₁₃ helix.     

Preparation of β2‐Homotryptophan Derivatives for β‐Peptide Synthesis

In view of the prominent role of the 1H‐indol‐3‐yl side chain of tryptophan in peptides and proteins, it is important to have the appropriately protected homologs H‐β² HTrp OH and H‐β³ HTrp OH (Fig.) available for incorporation in β‐peptides. The β²‐HTrp building block is especially important, because β²‐amino acid residues cause β‐peptide chains to fold to the unusual 12/10 helix or to a hairpin turn. The preparation of Fmoc and Z β²‐HTrp(Boc) OH by Curtius degradation (Scheme 1) of a succinic acid derivative is described (Schemes 2–4). To this end, the (S)‐4‐isopropyl‐3‐[(N‐Boc‐indol‐3‐yl)propionyl]‐1,3‐oxazolidin‐2‐one enolate is alkylated with Br CH₂CO₂Bn (Scheme 3). Subsequent hydrogenolysis, Curtius degradation, and removal of the Evans auxiliary group gives the desired derivatives of (R)‐H β²‐HTrp OH (Scheme 4). Since the (R)‐form of the auxiliary is also available, access to (S)‐β²‐HTrp‐containing β‐peptides is provided as well.     

Oxidative Fragmentations of the 5α‐ and 5β‐Hydroxy‐B‐norcholestan‐ 3β‐yl Acetates

Oxidations of 5α‐hydroxy‐B‐norcholestan‐3β‐yl acetate ( 8 ) with Pb(OAc)₄ under thermal or photolytic conditions or in the presence of iodine afforded only complex mixtures of compounds. However, the HgO/I₂ version of the hypoiodite reaction gave as the primary products the stereoisomeric (Z)‐ and (E)‐1(10)‐unsaturated 5,10‐seco B‐nor‐derivatives 10 and 11 , and the stereoisomeric (5R,10R)‐ and (5S,10S)‐acetals 14 and 15 (Scheme 4). Further reaction of these compounds under conditions of their formation afforded, in addition, the A‐nor 1,5‐cyclization products 13 and 16 (from 10 ) and 12 (from 11 ) (see also Scheme 6) and the 6‐iodo‐5,6‐secolactones 17 and 19 (from 14 and 15 , resp.) and 4‐iodo‐4,5‐secolactone 18 (from 15 ) (see also Scheme 7). Oxidations of 5β‐hydroxy‐B‐norcholestan‐3β‐yl acetate ( 9 ) with both hypoiodite‐forming reagents (Pb(OAc)₄/I₂ and HgO/I₂) proceeded similarly to the HgO/I₂ reaction of the corresponding 5α‐hydroxy analogue 8 . Photolytic Pb(OAc)₄ oxidation of 9 afforded, in addition to the (Z)‐ and (E)‐5,10‐seco 1(10)‐unsaturated ketones 10 and 11 , their isomeric 5,10‐seco 10(19)‐unsaturated ketone 22 , the acetal 5‐acetate 21 , and 5β,19‐epoxy derivative 23 (Scheme 9). Exceptionally, in the thermal Pb(OAc)₄ oxidation of 9 , the 5,10‐seco ketones 10, 11 , and 22 were not formed, the only reaction being the stereoselective formation of the 5,10‐ethers with the β‐oriented epoxy bridge, i.e. the (10R)‐enol ether 20 and (5S,10R)‐acetal 5‐acetate 21 (Scheme 8). Possible mechanistic interpretations of the above transformations are discussed.     

NMR‐Solution Structures in Methanol of an α‐Heptapeptide,of a β3/β2‐Nonapeptide,and of an all‐β3‐Icosapeptide Carrying the 20 Proteinogenic Side Chains

The NMR‐solution structure of an α‐heptapeptide with a central Aib residue was investigated in order to verify that, in contrast to β‐peptides, short α‐peptides do not form a helical structures in MeOH. Although the central Aib residue was found to induce a bend in the experimentally determined structure, no secondary structure typical for longer α‐peptides or proteins was found. A β²/β³‐nonapeptide with polar, positively charged side chains was subjected to NMR analysis in MeOH and H₂O. Whereas, in MeOH, it folds into a 10/12‐helix very similar to the structure determined for a corresponding β²/β³‐nonapeptide with only aliphatic side chains, no dominant conformation could be determined in H₂O. Finally, the NMR analysis of a β³‐icosapeptide containing the side chains of all 20 proteinogenic amino acids in MeOH is described. It revealed that this 20mer folds into a 3₁₄‐helix over its whole length forming six full turns, the longest 3₁₄‐helix found so far. Together, our findings confirm that, in contrast to α‐peptides, β‐peptides not only form helices with just six residues, but also form helices that are longer than helical sections usually observed in proteins or natural peptides. The higher helix‐forming propensity of long β‐peptides is attributed to the conformation‐stabilizing effect of the staggered ethane sections in β‐peptides which outweighs the detrimental effect of the increasing macrodipole.     

Stereoselective Preparation of 3‐Amino‐2‐fluoro Carboxylic Acid Derivatives,and Their Incorporation in Tetrahydropyrimidin‐4(1H)‐ones,and in Open‐Chain and Cyclic β‐Peptides

The preparation of (2S,3S)‐ and (2R,3S)‐2‐fluoro and of (3S)‐2,2‐difluoro‐3‐amino carboxylic acid derivatives, 1 – 3 , from alanine, valine, leucine, threonine, and β³h‐alanine (Schemes 1 and 2, Table) is described. The stereochemical course of (diethylamino)sulfur trifluoride (DAST) reactions with N,N‐dibenzyl‐2‐amino‐3‐hydroxy and 3‐amino‐2‐hydroxy carboxylic acid esters is discussed (Fig. 1). The fluoro‐β‐amino acid residues have been incorporated into pyrimidinones ( 11 – 13 ; Fig. 2) and into cyclic β‐tri‐ and β‐tetrapeptides 17 – 19 and 21 – 23 (Scheme 3) with rigid skeletons, so that reliable structural data (bond lengths, bond angles, and Karplus parameters) can be obtained. β‐Hexapeptides Boc[(2S)‐β³hXaa(αF)]₆OBn and Boc[β³hXaa(α,αF₂)]₆‐OBn, 24 – 26 , with the side chains of Ala, Val, and Leu, have been synthesized (Scheme 4), and their CD spectra (Fig. 3) are discussed. Most compounds and many intermediates are fully characterized by IR‐ and ¹H‐, ¹³C‐ and ¹⁹F‐NMR spectroscopy, by MS spectrometry, and by elemental analyses, [α]_D and melting‐point values.     

Mixed β2/β3‐Hexapeptides and β2/β3‐Nonapeptides Folding to (P)‐Helices with Alternating Twelve‐ and Ten‐Membered Hydrogen‐Bonded Rings

The structural properties of four mixed β‐peptides with alternating β²/β³‐ or β³/β²‐sequences have been analyzed by two‐dimensional homonuclear ¹H‐NMR‐ and CD spectroscopic measurements. All four β‐peptides fold into (P)‐helices with twelve‐ and ten‐membered H‐bonded rings (Figs. 3–6). CD Spectra (Fig. 2) of the mixed β³/β²‐hexapeptide 4a and β³/β²‐nonapeptide 5a , indicating that peptides of this type also adopt the 12/10‐helical conformation, were confirmed by NMR structural analysis. For the deprotected β³/β²‐nonapeptide 5d , NOEs not consistent with the 10/12 helix have been observed, showing that the stability of the helix decreases upon N‐terminal deprotection. From the NMR structures obtained, an idealized helical‐wheel representation was generated (Fig. 7), which will be used for the design of further 12/10 or 10/12 helices.     

Synthesis of 5′‐O‐(2‐Azido‐2‐deoxy‐α‐D‐glycosyl)nucleosides and Their Antitumor Activities

The 1,3,4,6‐tetra‐O‐acetyl‐2‐azido‐2‐deoxy‐β‐D ‐mannopyranose ( 4 ) or the mixture of 1,3,6‐tri‐O‐acetyl‐2‐azido‐2‐deoxy‐4‐O‐(2,3,4,6‐tetra‐O‐acetyl‐β‐D ‐galactopyranosyl)‐β‐D ‐mannopyranose ( 10 ) and the corresponding α‐D ‐glucopyranose‐type glycosyl donor 9 / 10 reacted at room temperature with protected nucleosides 12 – 15 in CH₂Cl₂ solution in the presence of BF₃?OEt₂ as promoter to give 5′‐O‐(2‐azido‐2‐deoxy‐α‐D ‐glycosyl)nucleosides in reasonable yields (Schemes 2 and 3). Only the 5′‐O‐(α‐D ‐mannopyranosyl)nucleosides were obtained. Compounds 21, 28, 30 , and 31 showed growth inhibition of HeLa cells and hepatoma Bel‐7402 cells at a concentration of 10 μM in vitro.     

Microwave‐Assisted Synthesis of β‐Lactams and Cyclo‐β‐dipeptides

Different cyclo‐β‐dipeptides were prepared from corresponding N‐substituted β‐alanine derivatives under mild conditions using PhPOCl₂ as activating agent in benzene and Et₃N as base. To evaluate β³‐substituent influence, the amino acids 7 – 26 were synthesized, and a β‐lactam formation reaction was carried out instead of cyclo‐β‐dipeptide formation. The crystal structures of three derivatives of cyclo‐β‐peptides and one β‐lactam are presented.     

Synthesis,CD Spectra,and Enzymatic Stability of β2‐Oligoazapeptides Prepared from (S)‐2‐Hydrazino Carboxylic Acids Carrying the Side Chains of Val,Ala, and Leu

β‐Peptides offer the unique possibility to incorporate additional heteroatoms into the peptidic backbone (Figs. 1 and 2). We report here the synthesis and spectroscopic investigations of β²‐peptide analogs consisting of (S)‐3‐aza‐β‐amino acids carrying the side chains of Val, Ala, and Leu. The hydrazino carboxylic acids were prepared by a known method: Boc amidation of the corresponding N‐benzyl‐L ‐α‐amino acids with an oxaziridine (Scheme 1). Couplings and fragment coupling of the 3‐benzylaza‐β²‐amino acids and a corresponding tripeptide (N‐Boc/C‐OMe strategy) with common peptide‐coupling reagents in solution led to β²‐di, β²‐tri‐, and β²‐hexaazapeptide derivatives, which could be N‐debenzylated ( 4 – 9 ; Schemes 2–4). The new compounds were identified by optical rotation, and IR, ¹H‐ and ¹³C‐NMR, and CD spectroscopy (Figs. 4 and 5) and high‐resolution mass spectrometry, and, in one case, by X‐ray crystallography (Fig. 3). In spite of extensive measurements under various conditions (temperatures, solvents), it was not possible to determine the secondary structure of the β²‐azapeptides by NMR spectroscopy (overlapping and broad signals, fast exchange between the two types of NH protons!). The CD spectra of the N‐Boc and C‐OMe terminally protected hexapeptide analog 9 in MeOH and in H₂O (at different pH) might arise from a (P)‐3₁₄‐helical structure. The N‐Boc‐β²‐tri and N‐Boc‐β²‐hexaazapeptide esters, 7 and 9 , were shown to be stable for 48 h against the following peptidases: pronase, proteinase K, chymotrypsin, trypsin, carboxypeptidase A, and 20S proteasome.     

Synthesis of Three‐, Five‐, and Six‐Membered Heterocycles Derived from New β‐Amino‐α‐(trifluoromethyl) Alcohols

Nucleophilic trifluoromethylation of α‐imino ketones 2 , derived from arylglyoxal, with Ruppert–Prakash reagent (CF₃SiMe₃) offers a convenient access to the corresponding O‐silylated β‐imino‐α‐(trifluoromethyl) alcohols. In a ‘one‐pot’ procedure, by treatment with NaBH₄, these products smoothly undergo reduction and desilylation yielding the expected β‐amino‐α‐(trifluoromethyl) alcohols 4 . The latter were used as starting materials for the synthesis of diverse trifluoromethylated heterocycles, including aziridines 5 , 1,3‐oxazolidines 8 , 1,3‐oxazolidin‐2‐ones 9 , 1,3,2‐oxazaphospholidine 2‐oxides 10 , 1,2,3‐oxathiazolidine 2‐oxides 11 , and morpholine‐2,3‐diones 12 . An optically active 5‐(trifluoromethyl)‐substituted 1,3‐oxazolidin‐2‐one 9g was also obtained.     

Coα‐(1H‐Imidazolyl)‐Coβ‐methylcob(III)amide: Model for Protein‐Bound Corrinoid Cofactors

Coα‐(1H‐Imidazol‐1‐yl)‐Coβ‐methylcob(III)amide ( 4 ) was synthesized by methylation with methyl iodide of (1H‐imidazol‐1‐yl)cob(I)amide, obtained by electrochemical reduction of Coα‐(1H‐imidazol‐1‐yl)‐Coβ‐cyanocob(III)amide ( 5 ). The spectroscopic data and a single‐crystal X‐ray structure analysis indicated 4 to exhibit a base‐on constitution in solution and in the crystal. The crucial lengths of the axial Co−N and Co−CH₃ bonds also emerged from the crystallographic data and were found to be smaller by 0.1 and 0.02 Å, respectively, than those in methylcob(III)alamin ( 2 ). The data of 4 support the view, that the `long' axial Co−N bonds as determined by X‐ray crystallography for the B<sub>12</sub>‐dependent methionine synthase, for methylmalonyl‐CoA mutase, and for glutamate mutase represent stretched Co−N bonds. The thermodynamic effect (the `trans influence') of the 1H‐imidazole base in 4 on the organometallic reactivity of this model for protein‐bound organometallic B₁₂ cofactors was examined by studying Me‐group‐transfer equilibria in aqueous solution and using (5′,6′‐dimethyl‐1H‐benzimidazol‐1‐yl)cobamides (cobalamins) as reaction partners (Schemes 2 – 5, Table). In comparison with methylcob(III)alamin ( 2 ), 4 was found to be destabilized for an abstraction of the Co‐bound Me group by a Co^III electrophile. In contrast, the abstraction of the Co‐bound Me group by a radical(oid) Co^II species was not significantly influenced thermodynamically by the exchange of the nucleotide base. Likewise, exploratory Me‐group‐transfer experiments with Me−Co^III and nucleophilic Co^I corrinoids at pH 6.8 provided an apparent equilibrium constant near unity. However, this finding also was consistent with partial protonation of the imidazolylcob(I)amide at pH 6.8, suggesting an interesting pH dependence of the Megroup‐transfer equilibrium near neutral pH. Therefore, the replacement of the 5′,6′‐dimethyl‐1H‐benzimidazole base by an 1H‐imidazole moiety, as observed in methyl transferases and in C‐skeleton mutases, does not by itself strongly alter the inherent reactivity of the B₁₂ cofactors in the crucial homolytic and nucleophilic‐heterolytic reactions involving the organometallic bond, but may help to enhance the control of the organometallic reactivity by protonation/deprotonation of the axial base.     

Enzyme‐Mediated Preparation of (+)‐ and (−)‐β‐Irone and (+)‐ and (−)‐cis‐γ‐Irone from Irone alpha®

The (−)‐ and (+)‐β‐irones ((−)‐ and (+)‐ 2 , resp.), contaminated with ca. 7 – 9% of the (+)‐ and (−)‐trans‐α‐isomer, respectively, were obtained from racemic α‐irone via the 2,6‐trans‐epoxide (±)‐ 4 (Scheme 2). Relevant steps in the sequence were the LiAlH₄ reduction of the latter, to provide the diastereoisomeric‐4,5‐dihydro‐5‐hydroxy‐trans‐α‐irols (±)‐ 6 and (±)‐ 7 , resolved into the enantiomers by lipase‐PS‐mediated acetylation with vinyl acetate. The enantiomerically pure allylic acetate esters (+)‐ and (−)‐ 8 and (+)‐ and (−)‐ 9 , upon treatment with POCl₃/pyridine, were converted to the β‐irol acetate derivatives (+)‐ and (−)‐ 10 , and (+)‐ and (−)‐ 11 , respectively, eventually providing the desired ketones (+)‐ and (−)‐ 2 by base hydrolysis and MnO₂ oxidation. The 2,6‐cis‐epoxide (±)‐ 5 provided the 4,5‐dihydro‐4‐hydroxy‐cis‐α‐irols (±)‐ 13 and (±)‐ 14 in a 3 : 1 mixture with the isomeric 5‐hydroxy derivatives (±)‐ 15 and (±)‐ 16 on hydride treatment (Scheme 1). The POCl₃/pyridine treatment of the enantiomerically pure allylic acetate esters, obtained by enzymic resolution of (±)‐ 13 and (±)‐ 14 , provided enantiomerically pure cis‐α‐irol acetate esters, from which ketones (+)‐ and (−)‐ 22 were prepared (Scheme 4). The same materials were obtained from the (9S) alcohols (+)‐ 13 and (−)‐ 14 , treated first with MnO₂, then with POCl₃/pyridine (Scheme 4). Conversely, the dehydration with POCl₃/pyridine of the enantiomerically pure 2,6‐cis‐5‐hydroxy derivatives obtained from (±)‐ 15 and (±)‐ 16 gave rise to a mixture in which the γ‐irol acetates 25a and 25b and 26a and 26b prevailed over the α‐ and β‐isomers (Scheme 5). The (+)‐ and (−)‐cis‐γ‐irones ((+)‐ and (−)‐ 3 , resp.) were obtained from the latter mixture by a sequence involving as the key step the photochemical isomerization of the α‐double bond to the γ‐double bond. External panel olfactory evaluation assigned to (+)‐β‐irone ((+)‐ 2 ) and to (−)‐cis‐γ‐irone ((−)‐ 3 ) the strongest character and the possibility to be used as dry‐down note.     

(Carbonyl)Iron‐Selenolates: New Synthetic Methods and Reactions with Electrophiles. Crystal Structures of [(CO)6Fe2(μ‐SeR)2]; R = Me3SiCH2 and p‐O2NC6H4CH2, {(CO)6Fe2[μ‐η2‐Se,Se′‐o‐(SeCH2C6H4CH2Se)]}, and [(CO)6Fe2(μ‐SeFe(CO)2cp)2]

The reactions of Fe(CO)₅ or Fe₃(CO)₁₂ with NaBEt₃H or KB[CH(CH₃)C₂H₅]₃H, respectively and treatment of the resulting carbonylates M₂Fe(CO)₄, M = Na, K with elemental selenium in appropriate ratios lead to the formation of M₂[Fe₂(CO)₆(μ‐Se)₂]. Subsequent reactions with organo halides or the complex fragment cpFe(CO)₂⁺, cp = η⁵‐C₅H₅ afforded the selenolato complexes [Fe₂(CO)₆(μ‐SeR)₂], R = CH₂SiMe₃ ( 1 ), CH₂Ph ( 2 ), p‐CH₂C₆H₄NO₂ ( 3 ), o‐CH₂C₆H₄CH₂ ( 4 ) and cpFe(CO)₂⁺ ( 5 ) in moderate to good yields. A similar reaction employing Ru₃(CO)₁₂, Se and p‐O₂NC₆H₄CH₂Br leads to the formation of the corresponding organic diselenide. The X‐ray structures of 1 , 3 , 4 and 5 were determined and revealed butterfly structures of the Fe₂Se₂ cores. The substituents in 1 , 3  and 5 adopt different conformations depending on their steric demand. In 4 , the conformation is fixed because of the chelate effect of the ligand. The Fe–Se bond lengths lie in the range 235 to 240 pm, with corresponding Fe–Fe bond lengths of 254 to 256 pm. The ⁷⁷Se NMR data of the new complexes are discussed and compared with the corresponding data of related complexes.