Approaches to Styrenyl Building Blocks for the Synthesis of Polyene Xanthomonadin and its Analogues

: A number of aryl building blocks for the synthesis of two xanthomonadin natural product pigments, as well as a related analogue, were accessed using a divergent hydroboration/bromoboration approach from a key alkynyl intermediate. A new approach towards substitution patterns around the ring was adopted following the isolation of an unexpected regioisomer from the bromination reaction. Potential coupling reactions onto these building blocks were explored, with a successful Sonogashira coupling performed on the key alkynyl intermediate, and with the key debrominated styrenyl boronate ester intermediate functionalised both by preliminary Suzuki-Miyaura coupling and via iododeboronation/Heck-Mizoroki coupling. Coupling reactions onto brominated styrenyl intermediates proved much more challenging due to the instability of the intermediates to cross-coupling, but some studies have shown promise. total synthesis of methodology required to complete the synthesis.


Introduction
Polyene natural products are ubiquitous in nature, and a wide variety of synthetic methods for their construction have been developed. [1] Cross-coupling, and iterative cross-coupling in particular, represents an extremely powerful methodology in this respect, and has consequently seen widespread use in natural product total synthesis. One drawback of such methodology is that the conditions for cross-coupling are frequently more forcing than desirable for the synthesis of such intrinsically unstable products. We have shown that Heck-Mizoroki (HM) reactions can often perform well at lower temperatures than is common for Suzuki-Miyaura cross-couplings, making this reaction potentially better suited to the construction of complex polyenes. [2][3][4][5][6][7][8][9] With this in mind, we envisaged the total synthesis of xanthomonadin 1, and its derivatives, would be an ideal test for the development of mild, polyene-compatible methodology; more especially because we have frequently found electron deficient alkenyl coupling partners to be challenging, with extended chain lengths doing little to aid stability.
Xanthomonadin campestris (black rot of crucifers) and members of the genus Xanthomonas are the cause of a number of plant diseases. These bacteria form characteristic yellow colonies due to the yellow, membrane-bound pigments they produce. [10][11][12][13][14][15][16][17][18][19][20][21] Andrewes and Starr pioneered investigations into these yellow pigments, first proposing the combination of arylated, polyenic and halogenated structures in 1973. [22] Andrewes then reported an attempted total synthesis of one of the proposed structures later that year, although the characteristics of this compound did not match those of any previously isolated pigments. [23] The first of these pigments, isobutyl xanthomonadin 1a, was successfully isolated and characterised from Xanthomonas juglandis strain XJ103 in 1976, and the micro-and resonance Raman spectroscopic characteristics obtained by Sharma et al. in 2012. [24,25] Interestingly, it has been postulated that these bacteria produce such compounds as biological, photo-protective agents. However, despite the similarity of such polyene compounds to the carotenoids, no efforts have been reported to synthesise this general class of compounds in order to test this photoprotective hypothesis. It is also interesting to observe that the nature of the ester function seems to depend upon the alcohol used in the extraction procedure, suggesting either that the ester itself is highly labile or that the free carboxylic acid is in fact the natural product; this would make xanthamonodinic acid a potentially more appropriate name, and allow the compound to be better incorporated into biological membranes.
During their investigations, Andrewes and Starr identified a number of different pigments in addition to xanthomonadin 1a by mass spectrometry, the most common of these being the putative debrominated xanthomonadin pigment 2. [22,26] This raises questions about the purpose of the bromine in xanthomonadin 1 and its relatives, specifically whether it provides an improvement in the activity or of stability to the pigment. However, a lack of complete spectroscopic data is a challenge in the synthesis of these pigments, particularly the lack of detailed NMR data. Indeed, the extent of the NMR data available for xanthomonadin 1a is detailed in Figure 1, with only mass spectrometry and UV data available for isobutyl debrominated xanthomonadin 2a (UV data was obtained on a mixture of pigments containing 2a). Therefore, one aim of our work was also to corroborate the characterisation data for all the xanthomonadins and build a full spectroscopic profile of the pigments.
With our experience in highly stereoselective polyene construction, [6,27] we anticipated stability issues in the construction of these pigments and therefore envisaged that an alkynyl analogue such as 3 might be useful, not only in terms of an interesting derivative for assessing biological activity (note that such an alkynyl function is a useful addition in synthetic retinoids, such as EC23 and related structures, having previously shown that such analogues can have the desired beneficial effects whilst retaining biological activity [28][29][30] ), but also because of the potential for imparting increased stability to the structure as a whole and making it more likely that useful stable analogues could be accessed. [28][29][30] [24] Our original retrosynthesis involved the synthesis of tetraenyl building blocks, employing a Suzuki-Miyaura (SM) coupling to complete the synthesis. As a result, we would require one key polyenyl intermediate, which could then be coupled with the appropriate polyenyl aryl intermediates to furnish each of the three target pigments. With this in mind, we considered the synthesis of all-trans polyene unit 4 using our Heck-Mizoroki (HM)/iododeboronation (IDB) iterative cross-coupling (ICC) methodology (Scheme 1) which we had previously applied in the synthesis of other polyene natural products. [2][3][4][5][6][7][8][9]31] Alternatively, if an all-trans heptaene 11 could be accessed, then this could be coupled directly onto suitable styrenyl and alkynyl aryl building blocks 8-10.
Given the planned use of the HM/IDB methodology, the exact nature of the alkenyl iodide-boronic acid coupling partners in the construction reaction could remain flexible. We anticipated that arenyl building blocks 8-10 could be used to access fragments 5-7 via palladium-catalysed cross-coupling, and therefore, these represented the key first targets for our intended approach, allowing us to choose the most appropriate route to access the desired pigments once we understood more about the stability and reactivity of the various intermediates. Herein, we report our approaches to the synthesis of these key building blocks, and their application in cross-coupling protocols to access polyene natural products systems and their polyenyl analogues.

Results and Discussion
Access to the ideal aryl tetraenyl iodide intermediates 12 and 13 from styrenyl building blocks 18 and 19 was envisioned from either a bromoboration or hydroboration reaction of Sonogashira phenylacetylene analogue 10 (Scheme 2). Sonogashira coupling onto building block 10 could also furnish a key polyenyl intermediate such as 22, or provide direct access to desired analogue 3. In turn, access to phenylacetylene analogue 10 was envisaged to be readily achievable from meta-iodoanisole 21 (Scheme 2). The initially desired bromination of 3-iodoanisole 21 proved more challenging than expected, with elemental bromine giving a mixture of regioisomers and NBS proving unreactive. Fortunately, lowering of the reaction temperature was found to give adequate regiocontrol (95:5) for the reaction employing Br2 (Equation 1), however, the two isomers could not be readily separated. The identity of the major regioisomer could not be unambiguously determined at this stage. Hence, the mixture was carried forward through the next synthetic steps with the intention of determining the major regioisomer at a later stage. The subsequent Sonogashira coupling and alkyne deprotection sequence was found to be successful under standard conditions (Scheme 3), furnishing the desired building block 27 for the key stereoselective bromoboration reaction. Attempts to perform a direct conversion were unsuccessful; however, after screening several conditions, a two-step process involving initial formation of a boronic acid 28 was developed. This involved initial reaction with boron tribromide followed by hydrolysis and esterification to give pinacol ester 29 as the desired Z-alkene stereoisomer. Whilst the boronic acid 28 proved difficult to handle, as the corresponding pinacol ester 29 it was readily handled and had the advantage of providing crystalline material. Subsequent single crystal X-ray crystallography provided proof of both the regiochemical outcome of the bromination and stereocontrol in the bromoboration reactions (Scheme 3). However, as can be seen from Figure 2, although the bromoboration of 27 gave the desired stereochemical result, the outcome of the SEAr reaction to derive the starting bromide 24 was not that anticipated.
Indeed, X-ray crystallography showed the bromine atom was in fact located ortho to the alkene, i.e. forming structure 30 (see Figure 2) and showing that the major regioisomer formed in the original bromination of 3-iodoanisole was in fact 23. As a result of these results and having uncovered the actual regioisomeric control, the rest of the route towards a key arenyl intermediate was now established and hence, a new selective entry to the required building block 20 was required.  We next examined an alternative synthesis of the desired iodide 20 utilizing a Sandmeyer approach from aniline 31 which proceeded readily to give the iodide 20 in quantitative yield (Scheme 4). Iodide 20 was then taken through the previously developed series of reactions to give alkyne 10 in excellent overall yield (Scheme 4); the only deviation from the previous route being use of TBAF rather than NaOH to cleanly convert TMS-alkyne 32 to required alkyne 10. With alkyne building block 10 unambigously produced, we could develop the synthesis of the two required styrenyl precursors to the different xanthomonadins, i.e. 18 and 19 (Scheme 5). As noted previously, it was envisaged that debrominated styrenyl building block 19 could be synthesised by hydroboration of alkyne 10. This was achieved using a copper(I)-catalysed borylation with Xantphos, sodium tert-butoxide and B2Pin2 in THF/methanol, and with some optimisation the styrenyl pinacolate ester 19 was isolated in a 79% yield and an overall 68% yield from aniline 31 (Scheme 5). Turning to bromo-analogue 18, application of the bromoboration conditions previously discussed (vide supra) and subsequent pinacol ester formation gave key brominated styrenyl boronate ester 18 in a 75% yield over the two steps. This ester was also successfully recrystallized and the structure confirmed by X-ray crystallography ( Figure 3). This route gave key boronate ester 18 from aniline 31 in an overall 65% yield (Scheme 5). It was found that the bromoboration step converting 10 to 18 was highly dependent upon the quality of the BBr3 used. Alkene 34 was routinely isolated as a by-product, presumably due to HBr  Comparison of 1 H NMR data obtained for styrenyl units 18 and 30 with those obtained by Andrewes et al. for the corresponding section of iso-butyl xanthomonadin 1a, showed a good agreement between the correct regioisomer 18 and 1a, something that had not been observed during the synthesis of the incorrect regioisomer 30 (Table 1). With the desired building blocks in hand, attention turned to the investigation of cross-coupling methods to enable access the required natural product pigments and analogue 3. A Sonogashira coupling was attempted on alkyne 10 (Equation 3) as a model cross-coupling reaction allowing access to a simplified alkynyl analogue of polyeneyne 3. This proved successful, giving the resulting enyne 36 in 89% yield (Equation 3).
Attention was then turned to Suzuki-Miyaura cross-coupling of styrenyl boronate analogues 18 and 19. Initially, these crosscouplings (Equations 4 and 5) were unsuccessful, and although it was noted the boronate esters were stable during the course of the SM coupling reaction, the alkenyl iodide partner 35 decomposed. In order to prevent this, cross-coupling of pinacol esters 18 and 19 was attempted using silver(I) oxide as base, which resulted in the desired SM products being be identified in the crude products according to 1 H NMR and mass spectrometry. However, isolation of the products 37 and 38 respectively proved difficult to isolate due to their tendency to polymerise (Equation 4).
We considered that the temperature used in the cross-coupling reactions above could be a cause of the decomposition of iodide 35. Hence, optimisation of the cross-coupling conditions between styrenyl boronate 19 and iodoacrylate 35 was carried out in order to develop improved reaction efficiency and allow for lower temperatures, as shown in Table 2.
Examination of Table 2 shows that generally, NMR yields of the model diene product 38 were improved at lower temperatures, particularly perhaps due to its sensitivity to polymerisation. At 60 °C ( Table 2, entry 1), product formation was poor, improving significantly both with lower temperature and increased palladium loading ( Table 2, Entries 2 and 3). Further reduction in temperature ( Table 2, Entries 4 and 5) made little difference at 30 °C but reduced product formation at room temperature. Changing bases also had an impact with silver carbonate improving the room temperature reaction ( Table  2, Entry 6) and tert-butoxide causing significant by-product formation (Table 2, Entry 7); a by-product that was generally observed in all reactions. This generally minor side-product 39 was observed, with 1 H NMR signals at δ 6.12 (1H, d, J = 18.3 Hz), 6.55-6.60 (1H, m) and 7.90 (1H, dd, J = 12.4, 7.8 Hz) inter alia, but remained elusive to isolation due to its high susceptibility towards polymerisation and hence, was not fully structurally identified.
Given our original aim of developing a flexible route to the synthesis of polyenyl intermediates, styrenyl boronate ester 19 was also subjected to an IDB/HM cross-coupling sequence, as an alternative route to the construction of debromo xanthomonadin analogues such as 2. Indeed, this was successful (Scheme 6), giving pinacol boronate 19 diene 42 in a 51% yield over the two steps, and interestingly, with diene 42 showing good stability and especially compared with the more electron deficient system 38.  [b] Multiple side-products observed. [c] Major product was the minor side-product observed in all other reactions.
[d] Conversion after 14.5 h. The SM tolerated lower temperatures, but benefitted from an increase in catalyst loading to improve the reaction rate ( Attention was then turned to reaction of the brominated styrenyl boronate analogues to give an alkenyl iodide as an alternative building block. Our standard iododeboronation conditions (NaOMe, ICl at -78 °C) were found to have limited success when applied to boronate ester 18. A literature procedure was found which could affect the conversion of boronic acids to halides using N-halosuccinimides at room temperature in acetonitrile [32] which, when attempted using N-iodosuccinimide (NIS) on pinacol ester 18 gave only unreacted starting material and alkyne 10. These conditions were, however, successfully applied to convert styrenyl boronic acid 33 to styrenyl iodide 43 in a 78% yield (Equation 5). Indeed, iodide 43 proved to be quite stable, and perhaps surprisingly so given the dihalogenated alkene moiety; our previous observations of related compounds showed instability despite storage at -18 °C under argon in the dark, whereas 43 proved stable for several months in these conditions.
The HM cross-coupling potential of styrenyl iodide 43 was then explored with vinylboronate 41 (Equation 6), which unfortunately proved unsuccessful, with starting material, alkyne 34 and alkyne 10 observed. In light of this, other types of cross coupling were considered. Both Stille and Sonogashira couplings were especially appealing, as these could be performed at room temperature. A Stille coupling was therefore attempted on iodide 43 with vinyl stannane 45 (Equation 8), however, although the desired product 44 was obtained, it was produced in an inseparable mixture of products, including alkyne 10 and starting material. A range of conditions were explored (see ESI for details), which improved the conversion to an extent, but not sufficiently to allow for isolation of the pure product. A Sonogashira reaction of iodobromostyrene 43 with TMS acetylene 25 under standard conditions (Equation 9) was attempted, with similar results to the previous Stille coupling. The desired product 47 was again produced, but could not be isolated pure. We anticipate that this reoccuring theme of challenging purification of cross-coupling products may be something that improves when using longer polyene systems.
The difficulties associated with iodide 43 meant attention was then turned to SM coupling of the more stable 1,2-borobromo styrene system 18, to endeavour to access bromo diene 37. This approach proved more successful (see Table 3) and while enyne 36 proved to be the major product under all conditions, careful choice of catalyst and base did result in formation of the desired product 37. It was clear that further work was needed if conditions suitable for a total synthesis of the brominated xanthomonadins were to be developed. Table 3 Attempted Suzuki-Miyaura couplings onto brominated styrenyl pinacol ester 18.
[a] 1 H NMR yields calculated due to product diene 37's tendency to polymerise, using characteristic dd at δ 6.12 ppm for the diene (easily identifiable) versus the singlet appearing at δ 6.43 ppm or the doublet appearing at δ 6.32 ppm for pinacolate ester 18 and enyne 35, respectively.

Conclusions
A number of key styrenyl building blocks for the synthesis of brominated xanthomonadin 1, debrominated xanthomonadin 2 and desired alkynyl analogue 3 were successfully synthesised, and their reactivity in several model cross-couplings for the construction of the xanthomonadins and their analogues were examined. Regioselective hydroboration and stereoselective bromoboration proved to be robust and efficient routes to the desired styrenyl boronate esters 18 and 19, using desired alkynyl building block 10 as their key intermediate, representing an efficient way to access these systems. The reactivity of these building blocks proved to be as anticipated, with the Sonogashira onto alkynyl building block 10 proving extremely facile. The successful Suzuki-Miyaura cross-coupling onto debrominated styrenyl boronate ester 19 along with the demonstrated reactivity towards iterative cross-coupling does indeed provide the intended flexible route to debrominated xanthomonadin 2. Whilst the brominated styrenyl analogues proved to be as challenging to cross-couple as expected, the unexpected stability of iodide 43, combined with a number of promising results across a number of different cross-coupling reactions provides encouragement that suitable conditions for reacting onto these intermediates will be found, thus allowing access to brominated xanthomonadin 1. Should this prove not to be the case, the successful Sonogashira coupling of alkyne 10 also opens up the possibility of functionalising the alkyne at a later stage in the synthesis via a hydrobromination reaction. This body of work therefore represents considerable progress

Experimental section General experimental
Except where specified, all reagents were purchased from commercial sources and were used without further purification. All 1 H NMR were recorded on Bruker Avance-400, Varian VNMRS-600, Varian VNMRS-700 spectrometers. 13

Supporting information statement
All relevant 1 H, 13 C and 11 B spectra and crystal data are detailed in the supplementary information. The CIF files have been deposited at the Cambridge Crystallographic Data Centre as CCDC-1537383 (18), 1537382 (30) and 1819789 (35).