Traceless Isoprenylation Traceless Isoprenylation of Aldehydes via N -Boc- N -(1,1-dimethyl-allyl)hydrazones

: A short isoprenylation protocol starting from non-conjugated N -Boc- N -(1,1-dimethylallyl)hydrazones was developed utilising Thomson's traceless bond construction. This type of [3,3]-sigmatropic rearrangement is catalysed by the Brønsted acid triflimide and liberates only gaseous by-products. The required N -Boc- N -allylhydrazine precursor is available in three


Introduction
The [3,3]-sigmatropic rearrangement is a common but impressive tool for the formation of new C-C-bonds in synthetic chemistry. [1] In 1973 Stevens showed that N-allylhydrazones undergo such a rearrangement under release of N 2 as well, but due to very harsh reaction conditions (300°C) and low yields, this reaction was limited in its applicability. [2] For several decades, synthetic chemists did not see any real benefit of this unique rearrangement, until 2010, when Thomson and co-workers published the traceless bond construction (TBC), an improved variant of Stevens' [3,3]-sigmatropic rearrangement, working with N-Boc-N-allylhydrazones (A, Scheme 1a) and catalytic amounts of the Brønsted superacid triflimide (HNTf 2 ). [3] It was now possible to lower the temperature of the rearrangement to 125°C and the yields of the products could be increased. This pioneering work of Thomson allowed the synthesis of various 1,2disubstituted olefins (B) and one 1,1-disubstituted olefin (Scheme 1a). Mono-substituted olefins could not be obtained by this way. Later our group extended the scope to the synthesis of 1,1-disubstituted olefins (D, Scheme 1b), bearing an isopropyl group in 1-position, which resulted in a methylene branched end, a motif which is found in the side chains of steroidal natural products, e.g. episterol. [4] In the same year we steps starting from a known diazene using biocatalytic aldol addition and Tebbe olefination as key steps. Allylhydrazones are prepared via condensation with appropriate aldehydes. Scope and limitations of the [3,3]-sigmatropic rearrangements are analysed.
reported the synthesis of terminal vinylsilanes (F, Scheme 1c) using TBC, which opened a new route to diversely substituted olefins. [5] Scheme 1. a) Original TBC by Thomson and co-workers. [3] b) Extension of the TBC to the synthesis of 1,1-disubstituted olefins bearing an isopropyl group. [4] c) TBC yielding terminal vinylsilanes. [5] d) Introduction of an isoprenyl group via TBC developed in this work.
In this work we present a protocol for the introduction of an isopentenyl (isoprenyl) residue to aldehydes (Scheme 1d). The isoprenyl function is a common structural element in terpenoid

Results and Discussion
The synthesis of the required N-Boc-N- (1,1-dimethylallyl)hydrazine building block 8a (Scheme 2; G in Scheme 1d) started with the two-step synthesis of known N-Troc-N-Boc-protected diazene 2. [21] Conversion into aldehyde 6a was performed on two different routes. Route A used commercially available silyl enol ether 3, which was activated by LiOTf and TBAF. The idea was to achieve a controlled O-Si-bond cleavage in 3 by slow addition of the fluoride source. Simultaneously, the presence of significant amounts of lithium ions should lead to an immediate formation of the lithium enolate. However, the addition of 3 to 2 did not proceed in a regioselective manner, and a 50:50 mixture of the isomeric aldehydes 6a and its regioisomer 6b was obtained. It is noteworthy, that the regioselectivity of this reaction could not be measured in this step, hence, it was determined retrospectively after conversion into 8a/8b after the last step. Both isomers showed identical chromatographic behaviour and no distinct signals enabling quantification of the ratio of regioisomers could be observed by NMR spectroscopy until reaching 8a/8b. Because of the lack of regioselectivity, an alternative approach to intermediate 6a utilising organocatalysis [22,23] was worked out (route B). For this Aldol-type reaction with isobutyraldehyde (4), three catalysts were explored: Lproline, [24] L-phenylalanine, [25] and Ley's (S)-5-(pyrrolidin-2-yl)-1H-tetrazole (5). [21,26] Tetrazole catalyst 5 gave the best result with 68 % yield and the isomeric ratio could be improved to 91:9 (determined retrospectively by 1 H NMR spectroscopy) of the desired aldehyde 6a and its regioisomer 6b. Methylenation of the aldehyde function of 6a/6b gave the olefins 7a and 7b. Different methods like Wittig, [27] Nysted-Takai [28] and Tebbe [29] olefination were tested, whereby the first two methods did not result in any product. Under Tebbe conditions the desired terminal olefin 7a and its regioisomer 7b were obtained in an acceptable yield of 48 % as an inseparable mixture. Scheme 2. Route A leading to an equimolar mixture of 8a/8b starting from silyl enolether 3. Route B provides 8a, contaminated with 9 % of isomer 6b starting from aldehyde 4. *The ratios of the isomers were determined retrospectively by NMR spectroscopy of the product 8a/8b. The X-ray crystal structure of the desired isomer 8a is shown on the left. Diazene 2 was synthesised according to literature. [21] Chemoselective reductive Troc cleavage with zinc powder gave a still inseparable mixture of the desired olefin 8a and its constitutional isomer 8b in excellent yield. However, at this stage NMR spectroscopy enabled determination of the ratio of main side product, and even right at the beginning of the reaction, the corresponding Boc-deprotected allylhydrazone was observed, a compound which does not undergo the rearrangement. This is in accordance with the observations of Thomson and could not be prevented. [3] This prompted us to further investigate an alternative carbamate residue, which might be less prone to premature acidic cleavage. We prepared the ethyl carbamate analogue S5a of 8a starting from ethyl carbazate on a route analogous to route B shown in Scheme 2 (for details, see Supporting Information). Two N-Boc-N-allylhydrazones S6g and S6i derived thereof were subjected to the previously determined best reaction conditions for rearrangement (HNTf 2 , diglyme, 125°C), but though the starting materials were fully consumed, none of the expected rearrangement products could be identified by GC/MS analysis. Consequently, the Boc group cannot be replaced in this protocol by the smaller ethoxycarbonyl group.
Scheme 4 shows the following rearrangement of substrates 9. The allylhydrazones 9a-c derived from n-alkanals underwent sigmatropic rearrangement providing the appropriate olefins 10a-c in 20-21 % isolated yields. The poor yields are in part due to the high volatility of the olefinic products, as demonstrated by an increased yield (25 %) of 10g on a larger scale (3 mmol). The rearrangement product 10d of isobutyraldehydederived N-allylhydrazone 9d could be detected by GC/MS, but could not be isolated due to its very high volatility (b.p. 135-136°C [32] ). Ester 9e did not undergo rearrangement to the corresponding olefin and only the Boc-deprotected allylhydrazone was found. N-Allylhydrazones derived from cycloalkane carboxaldehydes (9f, 9g) underwent rearrangement to olefins 10f and 10g with a yield of 20 % for both compounds (Scheme 4). In contrast, allylhydrazone 9h derived from an α, -unsaturated aldehyde did not undergo rearrangement and again only Boc-deprotected allylhydrazone was isolated. The attempted rearrangements of variously substituted arylidene hydrazones failed as well (9i-m). During the purification process of the attempted rearrangements of 9i and 9j crystalline solids were obtained, Obviously, acid-mediated removal of both the Boc and the dimethylallyl residue took place in these experiments. Next to those, once again Boc-deprotected allylhydrazones 11a/b were formed. Introduction of both electron-donating (methoxy compound 9l) and electron-withdrawing groups (nitro compound 9m) did not lead to successful rearrangements, and the same holds for hydrazones derived from heteroaromatic aldehydes (thiophene 9n and pyridine 9o). After these experiments it became evident which type of allylhydrazones would undergo the attempted acid-catalysed rearrangement. Non-conjugated allylhydrazones, like aliphatic systems 9a-d, 9f, and 9g form the corresponding olefins, in contrast to allylhydrazones conjugated with aryl or ester groups, which do not show any rearrangement. The following experiments supported this assumption: Non-conjugated N-allylhydrazone 9p derived from phenylpropanal showed a successful rearrangement with 19 % yield, whereas its cinnamaldehyde-derived congener 9q did not give the desired alkene 10q and only Boc-deprotected allylhydrazone was obtained. Thomson also reported on problems during the development of methods for hydrazone rearrangements, but with aliphatic systems, [3,33] which resulted in unidentified decomposition products. However, the rearrangement of aryl-substituted allylhydrazones worked well in his setup. Bocdeprotected allylhydrazones were observed in every reaction as by-products by GC/MS analysis, but no rearrangement takes place with these deprotected forms under our conditions. The deprotection reaction outcompetes the rearrangement and is a possible reason for the observed yields. This finding validates computational studies towards the mechanism of the triflimidecatalysed [3,3]-sigmatropic rearrangement by Gutierrez et al. indicating that conversion of deprotected allylhydrazones does not proceed well or not at all. [34] Conclusion In summary, we present a unique method for traceless isoprenylation of aliphatic aldehydes via triflimide-catalysed [3,3]sigmatropic rearrangement of N-Boc-N-allylhydrazones. The central N-Boc-N-allylhydrazine building block 8a is available in four steps utilising organocatalysis and Tebbe methylenation. This method opens a new route to isoprenyl compounds. This novel protocol is compromised by poor yields in the final step and its limitation to non-conjugated systems. Nevertheless, it broadens the scope of Stevens-type traceless bond constructions and represents the first example of a TBC for the introduction of an isoprenyl group into readily available aliphatic aldehydes. Therefore, this work extends the repertoire of methods for the total synthesis of isoprenoid natural products.

Experimental Section
General Information: All reactions were carried out in oven-dried Schlenk flasks equipped with a septum and a magnetic stirring bar which were evacuated and back filled with dry nitrogen. Solvents were dried according to standard methods by distillation over drying agents. Thin layer chromatography (TLC) was performed using polyester sheets polygram SIL G/UV254 covered with SiO 2 (layer thickness 0.2 mm, 40 × 80 mm) from Macherey-Nagel. Spots were visualized with a CAM (ceric ammonium molybdate) solution followed by heating. Flash column chromatography was performed using SiO 2 60 (0.040-0.063 mm, 230-400 mesh ASTM) from Merck. For chromatography distilled solvents were used. NMR spectra were recorded on JNM-Eclipse 400 (400 MHz), JNM-Eclipse 500 (500 MHz), Avance III HD 400 MHz Bruker Biospin (400 MHz) and Avance III HD 500 MHz Bruker Biospin (500 MHz) with CryoProbe™ Prodigy. Chemical shifts δ are reported as δ values in ppm relative to the deuterated solvent peak. The chemical shifts are reported in parts per million [ppm] and refer to the δ scala. Coupling constants J are indicated in Hertz [Hz]. For the characterization of the observed signal multiplicities the following abbreviations were applied: s (singlet), d (doublet), dd (doublet of doublet), dt (doublet of triplet), t (triplet), q (quartet), quint (quintet), m (multiplet), br (broad). Infrared spectra were recorded from 4000-650 cm -1 on a PERKIN ELMER Spectrum BX-59343 FT-IR instrument. For detection a Smiths Detection DuraSamp IR II Diamond ATR sensor was used. The absorption bands are reported in wave numbers (cm -1 ). High resolution mass spectra (HRMS) were recorded on a Jeol Mstation 700 (Fa. Jeol, Peabody, USA) or JMS GCmate II Jeol instrument for electron impact ionisation (EI) equipped with a quadrupole doublet based lens system. Thermo Finnigan LTQ FT (Fa. Thermo Electron Corporation, Bremen, Germany) was used for electrospray ionization (ESI) equipped with an ion trap. Melting points were measured with a Büchi apparatus B-540 (Büchi, Flawill, Switzerland) and are reported in°C and are not corrected. Gas chromatography (GC) was performed on a Varian 3800 gas chromatograph coupled to a Saturn 2200 ion trap from Varian (Darmstadt, Germany). The autosampler was from CTC Analytics (Zwingen, Switzerland) and the split/splitless injector was a Varian 1177 (Darmstadt, Germany). Instrument control and data analysis were carried out with Varian Workstation 6.9 SP1 software (Darmstadt, Germany). A Varian VF-5ms capillary column of 30 m length, 0.25 mm i.d. and 0.25 μm film thickness (Darmstadt, Germany) was used at a constant flow rate of 1.4 mL/min. Carrier gas was helium 99.999 % from Air Liquide (Düsseldorf, Germany). The inlet temperature was kept at 300°C and injection volume was 1 μL with splitless time 1.0 min. The initial column temperature was 50°C and was held for 1.0 min. Then the temperature was ramped up to 250°C with 50°C/min. Then the products were eluted at a rate of 5°C/min until 310°C (hold time 3 min). Total run time was 20 min. Transfer line temperature was 300°C and the ion trap temperature was 150°C. The ion trap was operated with electron ionization (EI) at 70 eV in scan mode (m/z 50-650) with a solvent delay of 6.3 min.
Crystallography: All X-ray intensity data were measured on a Bruker D8 Venture TXS system equipped with a multilayer mirror optics monochromator and a Mo K α rotating-anode X-ray tube (λ = 0.71073 Å). The data collections were performed at 103 K. The frames were integrated with the Bruker SAINT Software package. [35] Data were corrected for absorption effects using the Multi-Scan method (SADABS). [36] The structures were solved and refined using the Bruker SHELXTL Software Package. [37] All C-bound hydrogen atoms were calculated in positions having ideal geometry riding on their parent atoms.
Deposition Number(s) 1907495 (for 8a) contain(s) the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service www.ccdc.cam.ac.uk/structures.

General Procedure for the Synthesis of Olefins via [3,3]-Sigmatropic Rearrangement (GP2):
In an oven dried two-necked Schlenk flask HNTf 2 (10 mol-%) was dissolved in dry diglyme (1 mL). A solution of the appropriate N-Boc-N-allylhydrazone 9 (1.0 equiv.) in dry diglyme (2 mL + 1 mL rinse) was added at room temperature. The reaction mixture was fitted with a N 2 flashed reflux condenser and immediately heated to 125°C in a pre-heated oil bath. After completion of the rearrangement detected by TLC (75 min), the reaction was immediately cooled to room temperature via water bath and then quenched with a sat. aq. NaHCO 3 solution (4 mL). Pentane (10 mL) was added and the organic layer was washed with at least 100 mL water. The solvent was removed in vacuo (30°C, max. 700 mbar) and the crude product was purified by flash column chromatography.