**Contents**


### *Editorial* **Metal Promoted Cyclocarbonylation Reactions in the Synthesis of Heterocycles**

**Laura Antonella Aronica**

Dipartimento di Chimica e Chimica Industriale, University of Pisa, Via G. Moruzzi 13, 56124 Pisa, Italy; laura.antonella.aronica@unipi.it

Oxygen and nitrogen heterocycle systems are found in a vast number of natural substrates and biologically active molecules such as antimycotics, antibiotics, antitumors and antioxidants, in addition to pigments and fluorophores. Therefore, several procedures dedicated to the building of such heterocycles have been developed. Many of them are based on the cyclization of suitable substrates [1–3], multi-component reactions [4,5] and ring expansion processes [6,7]. In this field, metal-catalysed cyclocarbonylative reactions represent atom-economical and efficient methods for the synthesis of several functionalized compounds. Indeed, when the cyclization reaction is performed under CO pressure, the potentiality of the process is enhanced, since the formation of the ring takes place with the contemporary introduction of the carbonyl functional group. An improvement in the field of carbonylation reactions is represented by the substitution of the CO gas by surrogates, molecules which are able to generate the carbon monoxide inside the reaction vessel or may act as CO synthons [8–10].

The present Special Issue collected two research articles and three reviews focused mainly on the preparation of different heterocyclic compounds such as lactones, furans, epoxides, chromanes, chromanones, chromenones, indolinones, tetrahydroquinolines, quinolinones, lactams, benzoimidazoles and benzooxazoles via cyclocarbonylation reactions.

The first research article [11] concerns the selective monocarbonylation of epoxides into the corresponding lactones. The reactions were carried out in the presence of a chromium (III)-phthalocyanine derivative connected to a porous organic polymer. The catalyst was highly effective in promoting the ring expansion reaction and preliminary tests indicated that the catalyst can be reused without losing its catalytic activity.

The preparation of bicyclic lactones has been investigated in the second research article [12] as a key step for the total synthesis of Jaspine B, which has shown promising biological activity as an antitumor against several types of cancer cells. In order to avoid the use of dangerous carbon monoxide gas, the authors have developed a new protocol for Pd-catalysed carbonylation reactions based on the use of iron pentacarbonyl Fe(CO)<sup>5</sup> as a CO surrogate. After the optimization of the reaction sequence under batch conditions, the carbonylation reactions were performed also in a flow rector that provided the desired bicyclic lactones in comparable yields to standard batch conditions.

Metal-mediated cyclizations are important transformations in a natural product total synthesis. In the first review [13] of this Special Issue, the Co, Rh, and Pd catalysed Pauson–Khand reactions (PKR) published in the last five years have been summarized. In particular, their application to the synthesis of cyclopentenone and lactone-containing structures have been highlighted. In many examples, the carbonyl moiety has been inserted, employing metal carbonyl compounds as a masked CO source, through the transition metal decarbonylation to in situ generated CO (Co2(CO)8, Mo(CO)3(DMF)3, Rh(CO)2Cl]2). The hetero-Pauson–Khand reaction has also been considered, since it represents a useful tool for the generation of bicyclic γ-butyrolactones and unsaturated lactams. The final part of the review is focused on the synthesis of natural macrocyclic compounds containing cyclopentenone motif.

**Citation:** Aronica, L.A. Metal Promoted Cyclocarbonylation Reactions in the Synthesis of Heterocycles. *Catalysts* **2022**, *12*, 353. https://doi.org/10.3390/ catal12040353

Received: 18 March 2022 Accepted: 18 March 2022 Published: 22 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

A different approach to the synthesis of carbonyl-containing oxygen and nitrogen heterocycles based on visible light photocatalytic radical carbonylation has been summarized in the second review [14] of this Special Issue. Acyl radicals serve as the key intermediates in these transformations and can be generated from the addition of alkyl or aryl radicals to carbon monoxide (CO), or from various acyl radical precursors such as aldehydes, carboxylic acids, anhydrides, acyl chlorides or α-keto acids. The discussion of the literature is organized based on the types of acyl radical precursors, and an exhaustive analysis of the transformations leading to the different heterocycles is reported, with particular attention to the mechanistic aspects.

The last review [15] is focused on the synthesis of heterocyclic rings of different sizes, nature and potentialities containing both a silyl and a carbonyl moiety. Intramolecular silylformylation and silylcarbocyclization reactions are the key step for the cyclocarbonylation to occur. The content of this review is divided into two sections: the first is dedicated to a detailed description of intramolecular silylformylation reactions with the corresponding synthesis of oxa- and aza-silacyclane, while the second is centred on the silylcarbocyclization of functionalized acetylenes. In each section, particular emphasis is given to the heterocycles which can be obtained, as well as a special look into used metal catalysts.

In summary, the contribution of the articles collected in this Special Issue will be stimulating for those authors working in the field of heterocycles synthesis and will provide a valuable guide to develop new innovative methodologies for cyclocarbonylative reactions performed under batch, flow and photoredox catalytic conditions, with particular attention to the use of CO surrogates or equivalent synthons.

Finally, I would like to thank all authors for their valuable contributions.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Cr-Phthalocyanine Porous Organic Polymer as an E**ffi**cient and Selective Catalyst for Mono Carbonylation of Epoxides to Lactones**

### **Vinothkumar Ganesan and Sungho Yoon \***

Department of Chemistry, Chung-Ang University, 84, Heukseok-ro, Dongjak-gu, Seoul 06974, Korea; vinothcau@cau.ac.kr

**\*** Correspondence: sunghoyoon@cau.ac.kr

Received: 19 July 2020; Accepted: 6 August 2020; Published: 8 August 2020

**Abstract:** A facile, one-pot design strategy to construct chromium(III)-phthalocyanine chlorides (Pc'CrCl) to form porous organic polymer (POP-Pc'CrCl) using solvent knitting Friedel-Crafts reaction (FCR) is described. The generated highly porous POP-Pc'CrCl is functionalized by post-synthetic exchange reaction with nucleophilic cobaltate ions to provide an heterogenized carbonylation catalyst (POP-Pc'CrCo(CO)4) with Lewis acid-base type bimetallic units. The produced porous polymeric catalyst is identical to that homogeneous counterpart in structure and coordination environments. The catalyst is very selective and effective for mono carbonylation of epoxide into corresponding lactone and the activities are comparable to those observed for a homogeneous Pc'CrCo(CO)<sup>4</sup> catalyst. The (POP-Pc'CrCo(CO)4) also displayed a good catalytic activities and recyclability upon successive recycles.

**Keywords:** Cr-phthalocyanine; porous organic polymer; Friedel–Crafts reaction; heterogeneous catalysis; carbonylation; β-lactones; catalyst recyclability

#### **1. Introduction**

β-Lactones are an important class of energetically favored four-membered heterocycles with prevalent utilities in the chemical industry, since they are crucial intermediates for the production of various derivatives of β-hydroxy acids, biodegradable poly(β-hydroxyalkanotes), succinic anhydrides, and acrylic acids [1–9]. Their inherent ring strain facilitates excellent reactivity, allowing them to undergo a range of transformations to provide products with a variety of applications ranging from polymer chemistry to natural product synthesis [4]. However, synthetic routes to β-lactones are limited. Recently, ring-expansion epoxide carbonylation utilizing inexpensive C1 sources have emerged as a convenient and direct method to produce β-lactones with good atom economy [10].

Although there have been numerous reports of ring-expansion carbonylation catalysts, well-defined Lewis acid–base ion pairing catalysts of the common type [Lewis acid] <sup>+</sup> [Co(CO)4] <sup>−</sup> have demonstrated high efficiency for these transformations [11–13]. Among the reported Lewis acid–base pair catalysts, the porphyrin-based [OEPCr(THF)2] <sup>+</sup> [Co(CO)4] <sup>−</sup> (OEP = Octaethylporphyrinato, THF = tetrahydrofuran) catalyst has demonstrated high reactivity and high selectivity toward mono carbonylation under homogeneous conditions [14]. However, tedious catalyst synthesis and product separation have limited the use of this catalytic system and motivated the search for viable alternatives, including heterogeneous systems [15–17]. In addition to the heterogenization of catalysts to improve recyclability, facile synthesis of such catalytic systems is being actively researched [18–20].

Phthalocyanine (Pc), which is a porphyrinoid analogue, is easily synthesized with excellent yields and could serve as an alternative for porphyrin systems; because of their planar tetradentate dianionic ligation, phthalocyanines are excellent structural analogs to porphyrins and are synthetically facile. Recently, we demonstrated that a catalyst generated in situ from commercial (AlPcCl) and Co2(CO)<sup>8</sup> displayed excellent activity for mono and double carbonylation [21]. But, the selectivity toward β-lactones was very poor, hence, Lewis acidic Al3<sup>+</sup> containing [Lewis acid]<sup>+</sup> [Co(CO)4] <sup>−</sup> type ion pairing catalysts are proven to be active also for double carbonylation and generally resulting in a mixture of β-lactones and anhydrides [3,21,22]. Therefore, Cr3<sup>+</sup> containing [PcCr(III)]<sup>+</sup> [Co(CO)4] − type catalyst could be a suitable Lewis acidic part for selective monocarbonylation of epoxides into β-lactones [12,13,23]. However, partial solubility of Pc metal complex due to intermolecular π–π stacking interactions leads to reduced collision between the Pc metal complex and substrate (epoxide) resulting in low catalytic activity. These enforced further structural tunings on the Pc ring to improve the catalyst solubility and activity by controlling such a π–π stacking interactions; in addition, the intensely colored Pc metal complexes were difficult to separate from the reaction mixture [21,24,25].

Immobilization of soluble Pc metal complexes on a flexible POP addresses both, solubility and separation issues, by providing a heterogeneous catalytic system. In this regard, we strategically designed and synthesized a new phthalocyanine chromium(III) chloride complex (Pc'Cr(III)Cl) and heterogenized using a simple, one-pot solvent-knitting Friedel–Crafts reaction (FCR); the resulting complex was functionalized with [Co(CO)4] <sup>−</sup> anion to generate a highly active and recyclable [POP-Pc'Cr(III)]<sup>+</sup> [Co(CO)4] <sup>−</sup> catalytic system.

#### **2. Results and Discussion**

#### *2.1. Synthesis and Characterization of POP-Pc'Cr(III)Cl*

A synthetic strategy of POP-Pc'Cr(III)Co(CO)<sup>4</sup> (**4**) is shown in Scheme 1. At first, the ligand **1** is synthesized by the mild base-catalyzed condensation [26]. The monomer Pc'Cr(III)Cl (**2**) was synthesized with excellent yield according to the modified literature procedure and characterized by FTIR and UV-Visible spectroscopic techniques, further confirmed by high-resolution mass spectrometry as shown in Figures S1–S4 [27,28]. The substituted 2-isopropylphenolic group not only improves the solubility of the monomer **2** but also acts as a knitting group through covalent linkages by AlCl3-catalyzed FCR using methylene dichloride both as a crosslinker and as a solvent. The resulting dark green color porous organic polymer POP-Pc'CrCl (**3**) is stable under open atmosphere conditions and not soluble in most commonly used organic solvents as a result of extensive cross-linking [18,19]. The compositional homogeneity and the surface topography of POP **3** was probed by a scanning electron microscope (SEM) and a transmission electron microscope (TEM) analysis. Figure 1a,b shows the SEM and TEM images of **3** respectively, as an aggregate of polydisperse spherical shape particles of 1 µm size average (Figures S5 and S6). Energy dispersive X-ray (EDS) analysis shows the relative abundance of constituent elements throughout the Pc' polymeric matrix indicating uniform distribution of elements after polymerization (Figure S5) [29]. A powder X-ray diffraction analysis of the resulting polymer showed that a wide peak at 2θ = 13.1◦ attributes to the construction of amorphous polymeric material (Figure S7a). Subsequently, the absence of distinctive sharp monomeric diffraction peaks (Figure S7b) suggests that the solvent knitting polymerization is thoroughly completed and the resulting POP is free from crystalline monomer residues [18]. The thermal endurance of the resulting polymeric material was characterized by thermogravimetric analysis (TGA) as shown in Figure S8, the polymeric material was stable up to 400 ◦C indicating possible outstanding thermal stability under harsh reaction conditions.

The porosity of POP-Pc'CrCl, **3** was investigated by N<sup>2</sup> sorption measurement carried out at 77 K. The Pc'-based polymer **3** exhibited characteristic type-I adsorption isotherms (as per IUPAC adsorption isotherms classification), as depicted in Figure 1c [30]. A steep N<sup>2</sup> uptake at a lower relative pressure (P/P<sup>0</sup> = 0–0.1) region is attributed to the microporous character of polymer **3**, whereas the hysteresis loop behavior throughout the range of relative pressure can be attributed to the existence of mesoporosity. The BET (Brunauer–Emmett–Teller) surface area of the polymeric material is 725 m<sup>2</sup> g −1 and the total pore volume is 0.388 cm<sup>3</sup> g −1 (Figure 1c and Figure S9a). These porosity results are comparable with other reported Pc-based porous polymers and indicate high surface area, which enables larger exposure of active sites per unit mass of the material; they also confirm the distribution of high grade porous structure that can accommodate bulky functional groups via the labile chloro ligand for specific catalytic applications [18,31–33].

**Scheme 1.** Synthesis of catalyst 4 by the Friedel-Crafts reaction (FCR).

**Figure 1.** (**a**) SEM image of POP-Pc'CrCl, (**b**) TEM image of **3**, (**c**) N<sup>2</sup> sorption isotherms of **3** and **4** at 77 K, and (**d**) FTIR spectra of **2**, **3**, and **4**.

<sup>−</sup> − − The chemical structure of the resulting polymer was analyzed using FTIR spectroscopy. Comparative FTIR spectral analysis of the monomeric complex and the formed porous polymer was performed to evaluate the structural integrity of the resulting polymers, as displayed in Figure 1d. IR peaks appearing in the range of 3068–2856 cm−<sup>1</sup> are attributed to −C=C−H stretching vibrations of the aromatic functional groups of the phthalocyanine complex and polymer **3**, as well as to the newly formed methylene crosslinking bridges [34]. The peaks at 1610, 1470, and 1390 cm−<sup>1</sup> can be assigned for C=C, C–N, and C=N stretching vibrations, respectively, of the Pc' ring, which contains benzene, aza, and pyrrole functional groups; these are present in the formed polymers as well as the parent monomers [31,35,36]. These FTIR spectroscopic analysis shows that the resulting POP retains most

<sup>−</sup> − −

of the feature peaks of its corresponding monomer and is consistent with the anticipated polymeric structure. Thus, these analysis results unambiguously verify the direct heterogenization of Pc'Cr(III)Cl complex by the one-pot FCR to produce very stable, and heterogeneous porous polymeric materials.

#### *2.2. Synthesis and Characterization of POP-Pc'Cr(III)Co(CO)<sup>4</sup>*

Heterogeneous phthalocyanine polymer matrix is a probable candidate for Lewis-acid-enabled catalytic transformations. Particularly, the Pc'Cr(III)Cl complex resembles TPPCr(III)Cl (TPP = Tetraphenylporphyrinato), the best candidate for the carbonylation of epoxides with an additional Lewis base incorporation [13,37]. Interestingly, polymeric material **3** can be incorporated with an appropriate base via its labile Cl<sup>−</sup> anions in order to form a Lewis acid–base ion pair catalyst [12–14]. Accordingly a metathesis reaction of Co(CO)<sup>4</sup> <sup>−</sup> anions can replace the labile Cl<sup>−</sup> ions to generate a heterogeneous bimetallic frustrated Lewis acid-base ion pair type catalyst ([Lewis acid]+[Co(CO)4] <sup>−</sup>)to promote the epoxide ring-expansion carbonylation [17–20]. As such, polymer **3** was treated with excess KCo(CO)<sup>4</sup> to generate the heterogeneous epoxide carbonylation catalyst [POP-Pc'Cr]<sup>+</sup> [Co(CO)4] − (**4**) [12–14]. At first the resulting catalyst **4** was characterized by FTIR spectroscopic technique. Compared to polymer **<sup>3</sup>** (contains Cl−), the Co(CO)<sup>4</sup> <sup>−</sup> anions exchanged catalyst **4** exhibits a strong new absorption peak at 1882 cm−<sup>1</sup> (Figure 1d). This peak is characteristic of typical ν(CO) from newly exchanged tetrahedral Co(CO)<sup>4</sup> <sup>−</sup> ions, consistent with that of previously reported well-defined homogeneous Cr-containing [Lewis acid]<sup>+</sup> [Co(CO)4] <sup>−</sup>-type catalysts [17,19]. SEM and TEM images of the catalyst indicate no morphology change after Co(CO)<sup>4</sup> <sup>−</sup> anion exchange. Subsequently, EDS analysis confirms the incorporation of Co into the polymeric frameworks along with other constituent elements, distributed uniformly all over the polymer(Figure 2a,b, Figures S10 and S11). Atomic absorption spectroscopy (AAS) and inductively coupled plasma–optical emission spectroscopy (ICP-OES) revealed that the Co and Cr contents were 1.78 and 3.63 wt%, respectively, against, 4.63 and 4.09 wt% calculated for Co and Cr, respectively, in well-defined homogeneous catalyst. The molar ratio of the Cr/Co content in catalyst **4** is 1.8 (determined by ICP-AAS), indicating partial exchange of Cl<sup>−</sup> ion and a part of Lewis-acidic Cr3<sup>+</sup> remains combined with Cl<sup>−</sup> ions; they could be buried inside the microporous channels and/or inaccessible for cobaltate exchange. Limited molecular exchange of cobaltate ion pairs is consistent with SEM-EDS analysis and also reported previously [17,19].

The coordination environment of the catalyst **4** metal species was characterized using X-ray photoelectron spectroscopy (XPS). The XPS peak for Cr *2p* shows a characteristic doublet at 577.21 and 586.70 eV as shown in Figure 2c, which matches well with the structural analogues, the POP-TPP-supported Cr(III) species, and analogous metal center on porous organic networks [17,19,20,38]. As shown in Figure 2d, the XPS peaks for Co species are detected at 796.90 eV and 781.55 eV along with the typical shoulder is for the Co *2p*1/<sup>2</sup> and Co *2p*3/<sup>2</sup> orbitals of the Co(CO)<sup>4</sup> <sup>−</sup> species, respectively. The observed Co XPS peaks values are also consistent with those of Co(CO)<sup>4</sup> <sup>−</sup>-exchanged similar TPPAl, CTF-Al(OTf), and TPPCr heterogeneous catalysts [16–20]. Finally, the porosity retention is evident from TEM images (Figure S11) and N<sup>2</sup> gas sorption measurements carried out at 77 K, which afford type-I isotherms and exhibiting hysteresis loop behavior, displaying a combination of micro and mesoporosity (Figure 1c). However, the BET surface area is reduced to 550 m<sup>2</sup> g −1 for catalyst **4,** and a decreased total pore volume of 0.28 cm<sup>3</sup> g −1 (related to the parent polymeric network) is observed (Figure 1c and Figure S9b). This suggests that the exchanged Co(CO)<sup>4</sup> <sup>−</sup> anions partly occupy the porous channels of polymeric network, thereby decreasing the total available pore volume as observed previously [17–19]. Nevertheless, the catalyst maintains a porous structure to allow the substrate epoxide and the product β-butyrolactone molecules to diffuse over the Lewis acid–base-ions paired porous channels.

−

−

− −

−

− −

− −

−

−

−

**Figure 2.** STEM-EDS mapping image of (**a**) Cr atoms and (**b**) Co atoms and X-ray photoelectron profiles of **4** for deconvoluted (**c**) Cr*2p* and (**d**) Cobalt *2p* core level.

#### *2.3. Carbonylation Activity of POP-Pc'Cr(III)Co(CO)<sup>4</sup>*

−

ν <sup>−</sup>

Catalyst **4** was tested for carbonylation catalytic activity in a 50 mL stainless steel custom-made one inch tubular reactor. Propylene oxide (PO) was used as an epoxide substrate with CO under 6 MPa pressure. Various solvents were tested, since carbonylation is affected by the type of solvent [14,22]. The crude reaction mass was analyzed by <sup>1</sup>H NMR spectral analysis using an internal standard naphthalene; the results are summarized in Table 1. Among the solvents screened, weakly coordinating DME is the most active solvent system for PO carbonylation to β-butyrolactone with >99% conversion and selectivity (entries 1–4), consistent with earlier reports for POP-TPPCrCo(CO)<sup>4</sup> and analogs [17,19,20]. Using similar reaction conditions, we evaluated the catalytic activity of homogenous well-defined [Pc'Cr][Co(CO)4]; we observed conversion of >99% and selectivity of 99% toward β-butyrolactone (entry 5). The activity was tested with a higher PO ratio, and the yield was reduced to a ratio of 200 (entry 6). Reactions were performed for 12 h and 1 h under same reaction conditions (entry 7 and 8, respectively) to get initial rates of conversion, and the carbonylation yields of 70 and 22% were observed, respectively, with a site time yield of 44 h−<sup>1</sup> . Finally, the activity was tested at room temperature; 40% yield was achieved with a substrate/catalyst ratio of 100 (entry 9). The activity of POP-Pc'Cr(III)Cl was also evaluated under the same reaction conditions; we observed only PO to polyether conversion, owing to the absence of a Lewis base for carbonyl group insertion (entry 10) [21].

β

−

−


**Table 1.** Carbonylation activity of catalyst 4.

− −

β

β

−

−

<sup>a</sup> Reactions performed in DME solution of epoxide (1.8 M) under 6 MPa CO pressure at respective temperature. The mixture was stirred in a preheated oil bath to maintain respective temperature. <sup>b</sup> Calculated based on ICP-AAS value for Co content. <sup>c</sup> Determined by <sup>1</sup>H-NMR spectra with an internal standard naphthalene. <sup>d</sup> Polyether was formed.

Before testing recyclability, the heterogeneous nature of catalyst **4** was examined using a hot filtration test; a suspension of catalyst **4** in DME solvent was stirred at 60 ◦C for 6 h, and the treated catalyst was separated by filtration [15,18]. The dried solid catalyst and filtrate were subjected to the standard carbonylation conditions (2 mol% of catalyst, 6 MPa of CO, 60 ◦C, and 24 h reaction time) separately. Only the separated solids promoted epoxide carbonylation: no significant epoxide conversion was observed in the presence of the filtrate under the same conditions. This confirms that catalyst **4** retains heterogeneity [18,19]. Finally, catalyst **4** was evaluated for recyclability. Epoxide carbonylation was carried out with catalyst **4** at 30 ◦C temperature for 24 h under 6 MPa CO pressure in DME solvent. After the reaction, the reaction mixture was filtered inside a glove box to isolate the solid catalyst, which was then washed with dry DME and dried under vacuum; the dried catalyst was used for successive cycles. <sup>1</sup>H NMR spectral analysis of the recovered filtrate was conducted to evaluate the recycling ability of the catalyst, as listed in Table 2. The activity was reduced from complete conversion to 98% in the second cycle. The activity decreased further to 85 ± 6% in the third cycle. After the third cycle, the catalyst was analyzed by SEM-EDS to understand the reason for the decreased activity. SEM-EDS analysis of catalyst **4** after third cycle shows no changes in the catalyst morphology, but did reveal an increase in the Cr/Co ratio due to reduced Co content in the catalyst. As shown in Figure S12, the ratio of Cr/Co increased from 1.8:1 to 3.5:1 after the third cycle [16–19]. This suggests that the decreased Co content in the catalyst causes reduced activity during recycling [18,19]. Notably, the spent catalyst (after three cycles) was subjected to treatment with KCo(CO)<sup>4</sup> for regeneration [17–19]. The regenerated catalyst revealed restoration of the catalytic activity upon testing. This further substantiates that leaching of Co causes catalyst deactivation during recycling and the catalyst activity can be restored by treating with cobaltate ions to replenish the activity. Thus, POP-Pc'CrCo(CO)4, prepared via the solvent-knitting FCR, is an efficient and recyclable heterogeneous catalyst.


**Table 2.** Recyclability of **4.**

Reaction conditions: catalyst 2 mol%, 6 MPa of CO pressure, 30 ◦C, DME solvent. The PO conversion was determined by <sup>1</sup>H NMR spectra measured with internal standard naphthalene. \* regenerated catalyst.

#### **3. Experimental Section**

#### *3.1. Materials and Methods*

All chemicals and reagents were procured from commercial dealers and used as received unless otherwise mentioned. Chemicals 4-nitrophthalonitrile, 2-isopropylphenol, anhydrous aluminum chloride (AlCl3), dicobaltoctacarbonyl (Co2(CO)8), tetrahydrofuran (THF), dimethanoxyethane (DME), 1,4-dioxane, toluene, and propylene oxide (PO) were purchased from Sigma-Aldrich (Seoul, Korea). The solvent THF, DME, 1,4-dioxane, and toluene were distilled over sodium/benzoquinone and PO was distilled over CaH<sup>2</sup> under argon atmosphere. Deuterated solvents were purchased from Cambridge Isotope Laboratories, Inc. (T&J Tech Inc, Seoul, Korea). Research grade carbon monoxide was purchased from Air Liquide Korea Co., Ltd. (Seoul, Korea) with 99.998% purity and used as received. The KCo(CO)<sup>4</sup> was synthesized according to the reported procedure [39,40]. All manipulations of air and moisture sensitive compounds were carried out inside the glove box under argon atmosphere. Attenuated total reflectance infrared (ATR-IR) measurements were carried out on a Nicolet iS 50 (Thermo Fisher Scientific, Waltham, MA, USA). Scanning electron microscopy (SEM) and energy-dispersive spectroscopy (EDS) measurements were performed using a JEM-7610F (JEOL Ltd., Tokyo, Japan) operated at an accelerating voltage of 20.0 kV. The morphology of the prepared catalysts was observed by a transmission electron microscope Tecnai G2 (FEI Company, Hillsboro, OR, USA). TEM-EDX elemental mapping was obtained with transmission emission microscopy Talos F200X (Thermo Fisher Scientific, Waltham, MA, USA). The X-ray photoelectron spectrum (XPS) was obtained using K-Alpha X-ray photoelectron spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). The binding energies were corrected by the C1s peak from carbon contamination to 284.6 eV. The metal content of the catalysts was analyzed by inductively coupled plasma optical emission spectroscopy (ICP-OES) (iCAP 6000 series, Thermo Fisher Scientific, Waltham, MA, USA) using a microwave-assisted acid digestion system (MARS6, CEM/USA). Samples (~20.0 mg) were digested in a mixture of conc. HCl (20.0 mL) and conc. H2SO<sup>4</sup> (10.0 mL) solution under microwave rays at 210 ◦C for 60 min (ramp rate <sup>=</sup> <sup>7</sup> ◦C/min). N<sup>2</sup> adsorption-desorption measurements were conducted in an automated gas sorption system (Belsorp II mini, MicrotracBEL, Osaka, Japan) at 77 K; the samples were degassed for 12 h at 80 ◦C before the measurements. The Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) methods were used for calculating the surface areas and pore size distributions, respectively. Powder X-ray diffraction (PXRD) was measured on a RIGAKU D/Max 2500 V using CuKα radiation. <sup>1</sup>H and <sup>13</sup>C NMR were measured on a 600 MHz Varian VNS NMR spectrometer (Varian, Inc., CA, USA) and 400 MHz NMR spectrometer Bruker Avance III 400 (Bruker Korea Co., Ltd., Seoul, Korea). Simultaneous DSC-TGA instrument (TA instruments, New Castle, DE, USA) was used for the thermogravimetric analysis (TGA) with a heating rate of 10 ◦C/min from 25 ◦C to 800 ◦C under nitrogen atmosphere. UHR-MS measurements were performed on Bruker compact mass spectrometer (Bruker Korea Co., Ltd., Seoul, Korea).

#### *3.2. Synthesis of Pc' Ligand*

4-Nitrophthalonitrile (5.01 g, 0.03 mol), 2-isopropylphenol (4.34 g, 0.03 mol), and K2CO<sup>3</sup> (6.00 g, 0.04 mol) were stirred in anhydrous *<sup>N</sup>*,*N*-dimethylformamide (20 mL) at 52 ◦C for 24 h under N<sup>2</sup> atmosphere. A dark brown solution was obtained and was poured into ice-cold water (200 mL). The resulting brown precipitate was filtered, washed with water, and dissolved in dichloromethane (200 mL); the organic phase was purified by water extraction (3 × 100 mL). The desired product was purified by flash column chromatography (silica gel; hexane/ethyl acetate: 10:1) and recrystallized in hot methanol to obtain a pale white crystalline solid in 90% yield. FTIR: (cm−<sup>1</sup> ) 3085, 2977, 2870, 2233, 1590, 1481, 1446, 1415. 1311, 1280, 1246, 1218, 1184, 1084, 952, 872, 852, 775, 752; <sup>1</sup>H NMR (600 MHz, CDCl3, ppm) δ 7.71 (d, *J* = 8.7 Hz, 1H), 7.42 (dd, *J* = 7.5, 1.8 Hz, 1H), 7.32–7.25 (m, 2H), 7.23 (d, *J* = 2.5 Hz, 1H), 7.18 (dd, *J* = 8.7, 2.6 Hz, 1H), 6.93 (dd, *J* = 7.8, 1.3 Hz, 1H), 3.03 (dt, *J* = 13.8, 6.9 Hz, 1H), 1.18 (d, *J* = 6.9 Hz, 6H).13C NMR (151 MHz, CDCl3) δ 162.33 (s), 150.63 (s), 140.91 (s), 135.56 (s), 128.14 (s), 127.90 (s), 127.05 (s), 121.08 (s), 121.03 (s), 120.85 (s), 117.84 (s), 115.55 (s), 115.13 (s), 108.65 (s), 27.34 (s), 23.15 (s).

#### *3.3. Synthesis of Pc'Cr(III)Cl*

In a glove box, CrCl3·3THF (0.36 g, 0.96 mmol) and 4-(2-isopropylphenoxy)phthalonitrile (1.00 g, 3.82 mmol) were added to a 20 mL ampoule, which was then sealed under high vacuum. The ampoule was heated at a rate of 60 ◦C per hour to 250 ◦C and maintained at the same temperature for 5 h. The ampoule was then cooled to room temperature to obtain a dark product. The product was subsequently removed from the ampoule and purified by Soxhlet extraction using dichloromethane for 48 h to obtain a very dark green crystalline product in 80% yield. FTIR: (cm−<sup>1</sup> ) 3174, 2962, 2865, 1612, 1470, 1396, 1334, 1276, 1222, 1072, 1045, 952, 872, 818, 790, 748; UV-Vis: (THF) λmax 281, 366, 491, 622, 690 nm; HRMS (ESI Q-TOF) m/z calculated [C68H56CrN8O4] <sup>+</sup> 1100.3830, found [M-Cl]<sup>+</sup> 1100.3832.

#### *3.4. Synthesis of POP-Pc'Cr(III)Cl*

Under Ar atmosphere, Pc'Cr(III)Cl (1.00 g, 0.87 mmol) was suspended in 40 mL dichloromethane, the reaction mixture was cooled to 0 ◦C, and fresh anhydrous AlCl<sup>3</sup> (1.87 g, 14.07 mmol) was added. The reaction mixture was then stirred at 0 ◦C for 4 h, 30 ◦C for 8 h, 40 ◦C for 12 h, 60 ◦C for 12 h, and 80 ◦C for 24 h to obtain a dark-colored polymerized solid suspension. The resulting solid suspension was quenched using 50 mL of a HCl-H2O mixture (*v*/*v* = 2:1), washed with water thrice and with ethanol twice, then with THF, methanol, water, acetone, pentane, and ether (100 mL each). It was further purified by Soxhlet extraction with 1:1 methanol/THF for 48 h, and then dried in a vacuum oven at 80 ◦C for 24 h to obtain a dark green solid. FTIR: (cm−<sup>1</sup> ) 2931, 2854, 1608, 1465, 1392, 1334, 1226, 1080, 1049, 879, 825, 748.

### *3.5. Synthesis of [POP-Pc'Cr(III)]][Co(CO)4]*

Inside the glove box, the heterogenized POP-Pc'Cr(III)Cl (1.00 g) was suspended in 10 mL dry THF and was added to a THF solution of KCo(CO)<sup>4</sup> (1.02 g). The solution was stirred at room temperature for 48 h, following which the reaction mixture was filtered to remove the dark precipitate, which was washed with THF (3 × 50 mL) and dried under high vacuum for 8 h to yield a dark green solid. FTIR: (cm−<sup>1</sup> ) 2965, 2870, 1882, 1608, 1458, 1396, 1334, 1218, 1080, 1053, 1049, 879, 825, 748.

#### *3.6. PO Carbonylation Procedure*

A stainless steel carbonylation reactor was dried overnight and placed inside the glove box. The reactor was charged with the POP-Pc'Cr(III)Co(CO)<sup>4</sup> catalyst (0.01 g, 12.16 µmol) and a dimethoxyethane solution of propylene oxide (1.8 M in 2.5 mL, PO/catalyst ratio = 200). The reactor was tightened completely and pressurized to 6 MPa of CO after the removal from the glove box and then placed in a preheated oil bath at 60 ◦C for 24 h. At 60 ◦C, the pressure was 6.2 MPa; after completion

of the reaction, the reactor was brought to room temperature (pressure ~6 MPa) and cooled in an ice bath, following which CO gas was vented slowly inside the fume hood. The filtrate of the reaction mixture was analyzed by <sup>1</sup>H NMR spectroscopy using internal standard naphthalene (*Caution*: carbon monoxide (CO) is a highly toxic gas, should be handled with extreme care inside the well-ventilated hood with a proper CO detector).

#### **4. Conclusions**

A new design strategy was presented for the facile synthesis of a chromium(III)phthalocyanine-based porous organic polymer (POP-Pc'CrCl) through a solvent-knitting Friedel–Crafts reaction. The constructed POP-Pc'CrCl has a high porosity with a BET specific surface area of 725 m<sup>2</sup> g −1 . When functionalized with cobaltate ([Co(CO)4] <sup>−</sup>) anions, the resulting heterogenized bimetallic Lewis acid–base ion pair catalyst exhibits epoxide ring-expansion carbonylation activity comparable to that of its homogeneous counterpart with slightly reduced activity during successive recycles which can be replenished upon catalyst regeneration. This new design strategy is useful for the synthesis of soluble metallophthalocyanines and one step construction of porous organic polymer for specific catalytic applications.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4344/10/8/905/s1. Figure S1: <sup>1</sup>H NMR spectrum of ligand **1** measured in CDCl<sup>3</sup> . Figure S2: <sup>13</sup>C NMR spectrum of ligand **1** measured in CDCl<sup>3</sup> . Figure S3: UV-Visible spectrum of Pc'Cr(III)Cl (**2**). Figure S4: HR-MS of Pc'Cr(III)Cl (**2**). Figure S5: SEM and EDS mapping images of **3**. Figure S6: TEM images of **3**. Figure S7: Powder X-ray diffraction pattern of **3**. Figure S8: TGA plots of **2** and **3**. Figure S9: BJH pore size distribution graph of **3** and **4**.. Figure S10: SEM and EDS mapping images of **4**. Figure S11: TEM and EDS mapping images of **4**. Figure S12: SEM-EDS images of catalyst **4** after cycle three.

**Author Contributions:** V.G. and S.Y. designed the experiments. V.G. conducted the experiments. V.G. and S.Y. wrote the original draft. Review and edited by S.Y. Supervision, project administration, and funding acquisition by S.Y. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the C1 Gas Refinery Program (No. 2018M3D3A1A01018006) and ERC program (No. 2020R1A5A1018052) through the National Research Foundation of Korea (NRF) grants funded by the Ministry of Science, ICT and Future Planning, Republic of Korea.

**Acknowledgments:** We acknowledge the financial support by the C1 Gas Refinery Program (No. 2018M3D3A1A01018006) and ERC program (No. 2020R1A5A1018052) through the National Research Foundation of Korea (NRF) grants funded by the Ministry of Science, ICT and Future Planning, Republic of Korea.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Flow Pd(II)-Catalysed Carbonylative Cyclisation in the Total Synthesis of Jaspine B č**

**Pavol Lopatka 1 , Michal Gavenda 1 , Martin Markoviˇc 1,2, \*, Peter Koóš 1,2, \* and Tibor Gracza 1**


**Abstract:** This work describes the total synthesis of jaspine B involving the highly diastereoselective Pd(II)-catalysed carbonylative cyclisation in the preparation of crucial intermediates. New conditions for this transformation were developed and involved the *p*BQ/LiCl as a reoxidation system and Fe(CO)<sup>5</sup> as an in situ source of stoichiometric amount of carbon monoxide (1.5 molar equivalent). In addition, we have demonstrated the use of a flow reactor adopting proposed conditions in the large-scale preparation of key lactones.

**Keywords:** Jaspine B; flow chemistry; palladium catalysis; cyclisation; carbonylation

#### **1. Introduction**

The examination of natural resources clearly remains the basis of the discovery of new bioactive substances. Since these newly discovered natural compounds become the inspiration for novel drug candidates, many groups in the scientific community create their research programs aiming at these novel structures. As a result, the newly developed transformations are then presented in terms of their applicability in the synthesis of such targets. However, in many cases such synthetic demonstrations do not provide accessibility to all derivatives and/or usable amounts of promising target molecule for further testing. In particular, progress in later stages of pharmaceutical/biomedical research might then be negatively affected. Although the synthetic optimisation is not so scientifically valued in the chemical community, it is still of great importance. This is especially so today when we are facing environmental and climate changes and need to be concerned about a sustainable future.

Since its discovery, jaspine B (pachastrissamine) **1** has drawn an immense attention from the scientific community (Figure 1).

**Figure 1.** Structure of pachastrissamine (jaspine B) (**1**), jaspine A (**2**) a D-ribophytosphingosine (**3**).

This natural sphingolipid **1** was independently isolated in 2002 and 2003 by T. Higa et al. and by the C. Debitus group from marine sponges, *Pachastrissa* sp. (family Calthropellidae) and *Jaspis* sp., respectively [1,2]. Due to its similar structure to other bioactive sphingolipids, jaspine B **1** has been also tested for its pharmacological properties and it has shown in vitro cytotoxic activity against several types of cancer cell (A-549, P-388, HT-29,

**Citation:** Lopatka, P.; Gavenda, M.; Markoviˇc, M.; Koóš, P.; Gracza, T. Flow Pd(II)-Catalysed Carbonylative Cyclisation in the Total Synthesis of Jaspine B. *Catalysts* **2021**, *11*, 1513. https://doi.org/10.3390/catal11121513 č

Academic Editor: Laura Antonella Aronica

Received: 26 November 2021 Accepted: 9 December 2021 Published: 12 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

MeL-28, MCF-7, KB, HCT-116, U2OS, MDA-231. HeLa, CNE, MGC-803, EC-9706, PC-3, A-375, WM-115, Caco-2, Jurkat, SNU-638 and Caki-1) in the micro and sub-micromolar range [1–15]. The group of Y. Salma described how the cytotoxic activity of jaspine B **1** is based on the inhibition of SMS (sphinghomyelin synthetase) enzyme activity, which is responsible for maintaining stable concentration levels of ceramides (Cer) in the cell. Thus, higher concentration of Cer induces the cell apoptosis by a caspase-dependent pathway.

To this date, an enormous effort has been made to prepare and study this natural compound 1. There are 35 known syntheses of jaspine B **1**, ten of which are based on an asymmetric step [12,16–24] and twenty-five are chiral pool approaches [4,8,9,25–45]. Moreover, the promising biological activity of this natural compound has resulted in the syntheses of various derivatives of this molecule for structure–activity relationship study purposes (Figure 2).

**Figure 2.** Synthetic derivatives of jaspine B **1**.

In summary, the biological activity of **1** has been found to be highly dependent on the stereo-configuration of the ring substituents and on the length of the aliphatic chain of the natural product [4,9]. The best bioactivity was observed while keeping the original configuration and the length of aliphatic chain [31]. However, the oxygen atom in the heterocycle of jaspine B **1** has not been found to be crucial for its cytotoxic properties [46].

#### **2. Results and Discussion**

In the course of our long-term research program directed towards CO gas-free carbonylative cyclisations, flow transformations and their synthetic applications, we have developed a new flow protocol for Pd-catalysed carbonylation reactions based on the use of iron pentacarbonyl [47]. To this date, only a few flow applications of this stereoselective reaction are reported in the literature. However, many total syntheses of natural products utilised this transformation as a batch process [48–50]. As a result, the reaction conditions have undergone many changes and the reoxidation system as one of the most modified parameters has been varied/adjusted from its original conditions to substrate specific requirements (*p*BQ/CuCl2/O2) [51,52].

Based on the previous results and our experience with the Pd-catalysed carbonylation in the total synthesis of natural products, we have proposed a total synthesis of jaspine B **1** utilising the flow carbonylative step as a key transformation.

#### *2.1. Total Synthesis of Jaspine B*

The synthesis of jaspine B **1** was designed to build the chiral centres via the stereoselective Pd-catalysed carbonylation. In addition, the synthesis was optimised to reach one of the key aspects—the compatibility of batch reaction conditions for the following application in the flow system. Thus, this key transformation would provide *N*-protected lactone **21** with correct configuration at chiral centres. The substrate for this cyclisation—unsaturated *N*-protected amino diol can be easily accessible from L-serine (Scheme 1).

**Scheme 1.** Retrosynthetic analysis of jaspine B **1**.

As depicted in the retrosynthetic analysis, further transformation (chain elongation) of lactone **21** functional group would lead to the natural jaspine B.

Accordingly, the synthesis of **1** started from a commercially available *N*-Boc (*N*-*t*butoxycarbonyl) protected Garner's aldehyde **26**-Boc. This aldehyde **26**-Boc can be easily prepared in few synthetic steps starting from L-serine. The whole sequence involving an esterification of L-serine, amino group protection of derivate **23**, the formation of oxazolidine **25** and the following reduction of the ester group is very well described in the literature [53] (Scheme 2).

**Scheme 2.** Preparation of *N*-Boc protected Garner's aldehyde **27**-Boc.

At the beginning, the *N*-Boc protected Garner's aldehyde **26**-Boc was transformed to unsaturated *N*-Boc protected aminodiol **22**-Boc by a two step sequence (Scheme 3a) [54]. In detail, the addition of vinylmagnesium bromide to the starting aldehyde **26**-Boc in THF (tetrahydrofuran) provided a mixture of diastereomeric alcohols **27**-Boc in 91% yield. The selectivity of this reaction varies from 90:10 to 60:40 depending on the reaction temperature [54–56]. The major isomer, 2*S*,3*R-*alcohol **27**-Boc leads to a final product **1** with correct stereo configuration. The following selective deprotection of the acid labile oxazolidine group of **27**-Boc using *p*TSA (*p*-toluenesulfonic acid) provided *N*-Boc protected aminodiols **22**-Boc in 76% yield as an inseparable mixture [57]. Such a mixture of diastereomeric alcohols was then submitted to Pd-catalysed carbonylative cyclisation. This reaction was performed using well-established conditions with Fe(CO)<sup>5</sup> as a CO surrogate and the desired products-lactones **21**-Boc were obtained in 75% combined yield. At this stage, the lactone **21**-Boc with all *syn* configuration was separated using MPLC as a major substance from the diastereomeric mixture. Next, the reduction of lactone functional group using DIBAL-H (diisobutylaluminum hydride) provided lactol **28**-Boc in 88% yield (Scheme 3b).

**Scheme 3.** Total synthesis of jaspine B **1**.

The final step sequence included a Wittig reaction, double bond reduction and oxazolidinone cleavage (Scheme 3c). These synthetic steps have already been described in the literature and the authors used them to furnish the final natural compound **1** by employing *N*-Cbz (*N*-benzyloxycarbonyl) protected lactone **21**-Cbz as a starting material [35]. However, based on the inspection of spectroscopic data, Davies et al. [58] later described the epimerization at C-2 carbon occurred during the Wittig reaction and the authors prepared 2-*epi*-jaspine B [35]. This discrepancy was accounted for by a retro-Michael/Michael epimerisation reaction pathway upon treatment of lactol **28** with excess Wittig reagent (Scheme 4b).

**Scheme 4.** Formation of oxazolidinone ring (**a**) and possible epimerisation at C-2 carbon in Wittig reaction (**b**).

The authors later described how this epimerisation occurred in the DIBAL-H reduction step using old reagent containing base via same reaction pathway [15].

In our case, by applying the described conditions we were able to prepare the chiral oxazolidinone **29**-Boc in 40% yield. The formation of oxazolidinone ring via intramolecular attack of the hydroxy anion to the carbamate group furnished aldehyde **31** (Scheme 4a). Following a reduction of the double bond using Pd/C and H<sup>2</sup> gave compound **30** in 97% yield. Finally, jaspine B **1** was provided by a cleavage of the oxazolidinone ring of **30** using aqueous KOH in 69% yield. The comparison of spectroscopic data of prepared jaspine B **1** to described data [34] revealed that the epimerisation of **31** during Wittig reaction or reduction has not occurred.

In summary, we have accomplished the total synthesis of jaspine B **1** in seven steps starting from *N*-Boc protected Garner's aldehyde. The stereoselective Pd(II)-catalysed carbonylative cyclisation was used in the preparation of key intermediate-lactone **21**-Boc. The yield of jaspine B was 10% over all synthetic steps.

In addition, the synthesis of jaspine B was also performed starting from the commercially available *N*-Cbz protected Garner's aldehyde **26**-Cbz. Following the same synthetic route, we were able to increase the overall yield of jaspine B up to 16% yield over seven reaction steps (Scheme 5).

**Scheme 5.** Synthesis of jaspine B **1** via *N*-Cbz protected lactone.

Similarly, the Wittig reaction of lactol **28**-Cbz proceeded cleanly to oxazolidinone **29** without unwanted epimerisation at the C-2 carbon centre as in the previous case and the desired product **29** was isolated in 83% yield.

After successful optimisation of the total synthesis of jaspine B in batch, we focused our attention on the application of flow chemistry for the preparation of key intermediatelactone **21**. Thus, the flow Pd(II)-catalysed carbonylative cyclisation of **22** was proposed.

#### *2.2. Flow Synthesis of the Key Intermediate 21 for the Preparation of Jaspine B 1*

Over the last few decades, the flow chemistry has shown many advantages in organic synthesis, and it has grown into a modern and enabling tool for new synthetic methods utilising dangerous and/or toxic chemicals [59,60]. Moreover, many total syntheses of various biorelevant compounds have utilised this technique as a fundamental instrument in the preparation of key intermediates [61–63] or as a multiple step telescoped system [64–66]. At present, the flow chemistry has become an integral part of scientific research and it is commonly used in synthetic laboratories at universities and in pharmaceutical companies.

The flow chemistry technique as a part of our research program has been used in the application of CO surrogates in the carbonylative transformations. We have previously demonstrated that Fe(CO)<sup>5</sup> can be utilised as a CO donor in Pd(II)-catalysed carbonylative cyclisation [47]. In the continuation of our research, we have focused on the development of new conditions for this flow transformation and on its application in the flow synthesis of bioactive compounds. Thus, we have proposed a new flow system for the preparation of the key intermediate **21** in the jaspine B **1** synthesis (Figure 3).

**Figure 3.** Proposed telescoped flow system for the synthesis of bicyclic lactone **21**.

The proposed telescoped flow system consists of three parts involving the Grignard addition, selective deprotection and crucial cyclisation. The aim of the design was to perform these steps in a one-telescoped system without executing any isolation procedure. Consequently, the flow system would provide the crude final lactone **21** starting from commercially available substrates.

Our investigation started with a series of batch experiments aimed to address the minimal requirements and compatibility of reaction conditions for flow procedure. At first, we examined various reaction conditions of the first two steps sequence of the total synthesis—addition of vinyl magnesium bromide to *N*-protected Garner's aldehyde **26** and selective cleavage of oxazolidine **27** without the isolation after the first step (Scheme 6).

**Scheme 6.** Optimisation of the RMgX addition and deprotection reaction sequence.

Compared to the previously described batch conditions (Scheme 3a), we tried and modified mainly the cleavage of 2,2-dimethyl oxazolidine **27** using different acidic conditions (Table 1). The best results were achieved using *p*TSA.H2O as a H<sup>+</sup> source in MeOH in the second step (Table 1, Entry 1 and 5). Following products **22**-Cbz and **22**-Boc were isolated in 87% and 75% yield over two steps, respectively. In general, altering the temperature of RMgX addition did not affect the yield of this sequence and only the difference in the diastereoselectivity outcome was observed. The addition step at 0 ◦C of this two-step procedures provided in all cases a 1:1 mixture of diastereomeric alcohols **22**. In addition, the SO3H polymer supported resin, Amberlyst 15 was also tried for the deprotection step as the implementation of polymer supported reagents in flow reactions is described very well [64,67]. In this case, the yield of product **22**-Boc was decreased due to the partial cleavage of an acid labile *N*-Boc protecting group of substrate **26**-Boc. The formation of completely deprotected aminodiol **22** was confirmed by LC-MS analysis and the full cleavage of protecting groups of **26**-Boc was also observed in the case of the reaction performed in AcOH (Table 1, entry 3).


**Table 1.** Optimisation of the RMgX addition and deprotection reaction sequence in batch.

<sup>a</sup> Yield based on the amount of isolated product **22**. <sup>b</sup> The capacity of Amberlyst 15 is 4.7 m equivalents per 1g by dry weight. <sup>c</sup> Partial cleavage of *N*-Boc protecting group was observed by LC-MS analysis.

> Based on the optimisation of reaction sequence under batch conditions, we performed a series of experiments using a flow system as depicted in Scheme 7. At first, only nucleophilic addition of RMgX was examined. In this case, the second stage of the flow setup was omitted, and the Grignard reaction was performed at 0 ◦C to ensure the homogeneity of the reaction stream (the reaction at −78 ◦C is not homogeneous). Thus, the flow reaction using a 9 mL reactor coil provided products **27** with full conversion of substrates **26** in acceptable yields with lower stereoselectivity (Table 2, entry 1 and 2). −

**Scheme 7.** Optimisation of the flow addition/deprotection reaction sequence.

The following flow experiments employing the second stage of the system were performed using a reaction coil or Diba column (optional) depending on the use of H<sup>+</sup> donor in the acetonide deprotection step. In the case of the reaction using an excess of Amberlyst 15 (1.7 g, 8 equivalents) in the Diba column, the reagent also secured the filtration of reaction stream and Mg(II) salts formed after mixing (quenching) the Grignard reaction stream with MeOH were caught on the polymer resin. However, the larger excess of Amberlyst 15 also caused a parallel carbamate cleavage and decreased the yield of products. Thus, the flow system using Amberlyst 15 provided products **22**-Boc and **22**-Cbz in 34 and 49% yield, respectively (Table 2, entry 3 and 4).

The best results were achieved using 1.1 M solution of *p*TSA in MeOH in the second step and the products were isolated in similar yields to batch (Table 2, entry 5 and 6). In this case, the concentration and flow rate of *para*-toluenesulfonic acid stream were adjusted to ensure the catalytic amount of H<sup>+</sup> necessary for the deprotection of acetonide group. Since, the 0.25 of 0.275 mmol/min amount of this acid was immediately quenched in the reactor by the excess of Mg(II) salts, the 0.275 mmol/min amount really represents only the catalytic amount of H<sup>+</sup> . Thus, the flow reactions using 8.65 mmol of substrates **27**-Cbz or **27**-Boc led to the formation of desired *N*-protected unsaturated aminodiols **22** in 82 and 62% yields, respectively (Table 2, entry 5 and 6). With optimised conditions for this addition/deprotection sequence in hand, we turned our attention to the continuous carbonylative cyclisation step.


**Table 2.** Optimisation of the flow addition/deprotection reaction sequence.

<sup>a</sup> 0.5 M solution of substrate in THF (tetrahydrofuran) was used. <sup>b</sup> Yield based on the amount of isolated product **22**. <sup>c</sup> Only addition step was performed in the flow system, the yield corresponds to the addition product **27**. <sup>d</sup> Deprotection of *N*-Boc group was also observed.

> In 2018, we reported on pros and cons of the flow and batch stereoselective Pdcatalysed carbonylative cyclisation of unsaturated polyols/aminoalcohols using known conditions [68] and Fe(CO)<sup>5</sup> as an in situ donor of carbon monoxide [47]. By adjusting the concentration of reaction streams and the amount of Cu2+/Li<sup>+</sup> inorganic salts required for the reoxidation Pd<sup>0</sup> , we were able to prepare a series of various cyclisation products in a flow system (Scheme 8).

**Scheme 8.** Previously described continuous Pd(II)-catalysed carbonylation of unsaturated polyols and aminoalcohols.

Even though aforementioned continuous flow system was successfully applied in the large-scale preparation of desired bicyclic lactones, there are still limitations of the system regarding the formation of insoluble copper salts in the reactor. Also, an active mixer and interchangeable filtration unit were necessary to perform the transformation over longer periods.

With the aim to improve the conditions for continuous flow Pd-catalysed oxy/ aminocarbonylation, we adopted the use of *p*BQ (*p*-benzoquinone) instead of Cu2+ as a reoxidant.

At first, we performed a batch reaction using 2 equivalents of *p*BQ, 0.6 equivalent of Fe(CO)<sup>5</sup> and 0.1 equivalent PdCl2(MeCN)<sup>2</sup> using diastereomeric mixture **22**-Boc (0.23 mmol) in acetic acid (0.9 mL, 0.25 M reaction). The reaction at 60 ◦C proceeded with full conversion of starting material after 1 h, however noticeable amounts of insoluble material were observed. The homogeneity of the reaction mixture over the whole course of reaction was achieved by the addition of 1 equivalent of LiCl. The reaction proceeded smoothly with full conversion of **22**-Boc in 1 h at 60 ◦C. A comparison of newly optimised and previously described conditions is shown in Figure 4.


**Figure 4.** Comparison of known and new conditions for Pd-catalysed carbonylative cyclisation.

The new reaction conditions were then tested in the preparation of key intermediatelactone **21** for the synthesis of jaspine B **1** in continuous flow mode (Scheme 9). The flow setup consisted of two reaction streams which were pumped using HPLC Azura pumps via injection coils to a preheated reactor coil. The composition of stock solutions were adjusted to avoid the decomposition of Fe(CO)<sup>5</sup> in the presence of oxidation reagents (*p*BQ). The following continuous reaction of **22** on 0.5 mmol scale provided the desired *N*-protected bicyclic lactones **21** in comparable yields to standard batch and flow conditions (Table 3).

**Scheme 9.** Continuous Pd(II)-catalysed carbonylations of unsaturated *N*-protected aminodiols **22**.


**Table 3.** Comparison of Pd(II)-catalysed carbonylation of **22** employing batch and flow conditions.

<sup>a</sup> Combined yield of both diastereomers based on the isolated amounts of products **21**. <sup>b</sup> Batch reaction of **22** (0.5 mmol) using PdCl2. 2CH3CN (0.048 mmol, 0.1 equivalent), CuCl<sup>2</sup> (1.9 mmol, 4 equivalents), LiOAc (1.9 mmol, 4 equivalents) and Fe(CO)<sup>5</sup> (0.15 mmol, 0.3 equivalent) in 1.9 mL of AcOH [68]. <sup>c</sup> Reaction was performed on 1.22 mmol scale (substrate **22**) using the conditions and flow system as described in the literature [47]. <sup>d</sup> Reaction was performed using flow system as depicted in Scheme 9.

In detail, flow transformation using new conditions as depicted in Scheme 9 provided both lactones **21** in comparable yields to reactions performed under typical conditions (Table 3, flow yield column). The new reoxidation system for Pd<sup>0</sup> -PdII cycle ensures homogeneity of the reaction stream thus enabling better scaling of this flow transformation. Compared to the batch reaction (Table 3, typical conditions, batch yield column), the designed flow transformation (under new conditions) has several advantages. The batch reaction using typical conditions [47] can be undertaken only on a small scale due to the excessive pressure in the glass reaction tube. Upscaling the batch reaction 20 mmol may cause a few problems. As Fe(CO)<sup>5</sup> immediately decomposes after contact with the reaction mixture, it releases 1 equivalent of CO resulting in foaming and problematic stirring of heterogenous reaction mixture. Also, the pressure in a 120 mL reaction tube can raise up to 95 psi after few minutes.

To prove the robustness of the described flow setup, a six-hour long experiment was performed. Based on our previous experience with this type of transformation, we lowered the reaction stream concentration (0.25 M down to 0.125 M) to avoid the gas-liquid segment formation in the reaction coil. Continuous flow transformation of diastereomeric mixture **22**-Boc (5.57 g) using such minimally modified conditions provided the desired lactones **21**-Boc in 71% (4.5 g) combined yield (using the same flow setup as depicted in Scheme 10). However, a formation of precipitate at the exit of the reactor after cooling the reaction stream was observed. To prevent the potential clogging of the tubing, THF was employed as co-solvent in the case of a large-scale continuous reaction using **22**-Cbz (Scheme 10). In this case, the prepared stock solutions of substrate **22**-Cbz and reagents were pumped directly through the HPLC pumps into the larger reactor (47 mL) therefore allowing us to use higher flow rates (0.785 mL/min) and transform a larger amount of starting material **22**-Cbz in shorter time. In detail, the diastereomeric mixture *N*-Cbz protected aminodiols **22**-Cbz (7.6 g) was easily transformed over 2.5 h into bicyclic lactones **21**-Cbz (d.r.: 2.6:1, 6.3 g) in 75% combined yield. The pure diastereomer **21**-Cbz with all *syn*-configuration was obtained after MPLC purification in 54% yield (4.5 g).

In conclusion, we have designed and optimised an enhanced synthesis of the key intermediate-lactones **21**-Cbz and **21**-Boc utilising the stereoselective Pd-catalysed cyclocarbonylation of corresponding unsaturated aminodiols **22**. The key lactones **21** were then successfully transformed into natural jaspine B **1** over a four-step sequence in batch. Also, we have demonstrated the applicability of the flow reactor in two steps preparation of *N*-protected aminodiols **22** in comparable yields to the batch process. Importantly, new conditions for Pd-catalysed cyclocarbonylation of unsaturated polyols/aminoacohols were developed involving *p*BQ/LiCl as a reoxidation system and Fe(CO)<sup>5</sup> as an in situ source of stochiometric amount of carbon monoxide (only 1.5 molar equivalents). Such conditions were easily applied to continuous flow mode allowing us to prepare gram quantities of intermediates **21** for jaspine B **1** synthesis. This flow setup has shown several advantages compared to previous versions of the flow reaction system and the homogeneity of the reaction stream facilitated the use of a common flow system without the implementation of any other special devices.

**Scheme 10.** Large-scale continuous synthesis of *N*-Cbz protected lactone **21**-Cbz.

#### **3. Experimental Section**

#### *3.1. Material and Methods*

Commercial materials which were obtained from Sigma-Aldrich, Acros Organics, Alfa Aesar or Fisher Scientific were used without further purification. Reactions were monitored using TLC on silica gel. Compound purification was undertaken by flash chromatography. All solvents were distilled before use. Hexanes refer to the fraction boiling at 60–65 ◦C. Flash column liquid chromatography (FLC) was performed on silica gel Kieselgel 60 (15–40 µm, 230–400 mesh) and analytical thin-layer chromatography (TLC) was performed on aluminium plates pre-coated with either 0.2 mm (DC-Alufolien, Merck) or 0.25 mm silica gel 60 F254 (ALUGRAM® SIL G/UV254, Macherey–Nagel). Analysed compounds were visualized by UV fluorescence and by dipping the plates in an aqueous H2SO<sup>4</sup> solution of cerium sulphate/ammonium molybdate followed by charring with a heat gun. Melting points were obtained using a Boecius apparatus and are uncorrected. <sup>1</sup>H and <sup>13</sup>C NMR spectra were recorded on either 300 (75) MHz MercuryPlus or 600 (151) MHz Unity Inova spectrometers from Varian (Supplementary Materials). Chemical shifts (δ) are quoted in ppm and are referenced to the tetramethylsilane (TMS), CDCl<sup>3</sup> or DMSO-d<sup>6</sup> as internal standard. FTIR spectra were obtained on a Nicolet 5700 spectrometer (Thermo Electron) equipped with a Smart Orbit (diamond crystal ATR) accessory, using the reflectance technique (400–4000 cm−<sup>1</sup> ). High-resolution mass spectra (HRMS) were recorded on an OrbitrapVelos mass spectrometer (Thermo Scientific, Waltham, MA, USA; Bremen, Germany) with a heated electrospray ionization (HESI) source. The mass spectrometer was operated with a full scan (50–2000 amu) in positive or negative FT mode (at a resolution of 100,000). The sample was dissolved in methanol and infused via syringe pump at a rate of 5 mL/min. The heated capillary was maintained at 275 ◦C with a source heater temperature of 50 ◦C and the sheath, auxiliary and sweep gases were at 10, 5 and 0 units, respectively. Source voltage was set to 3.5 kV.

#### *3.2. Representative Flow Procedures*

#### 3.2.1. Grignard Reaction

The flow setup consisted of three HPLC pumps (Knauer Azura 4.1S with 10 mL pump head). The pumps were used to introduce a solution of substrate **26** (1 mmol, 0.5 M) in anhydrous THF (Feed A), a commercial solution of vinylmagnesium bromide (1.0 M in THF, Sigma-Aldrich, Feed B), and MeOH (Feed C). Injection loops (PTFE, 0.8 mm i.d., 1.6 mm o.d.; internal volume: 2.0 mL, Feed A, and 2.3 mL, Feed B) were used to deliver the two starting feeds. At start of the experiment, the whole reactor was flushed with an anhydrous THF (Feed A and Feed B) and MeOH (Feed C). Both solutions were loaded into their corresponding injection loops. Feed A and feed B were pumped from the injection loops and mixed in a T-shaped connector (PEEK) in a cooling bath (0 ◦C). The combined mixture passed through a coil reactor (PTFE, 0.8 mm i.d., 1.6 mm o.d.; internal volume: 9.0 mL) at 0 ◦C before the mixture was combined with MeOH (Feed C) in a T-shaped

connector (PEEK) at the same temperature. Final reaction mixture left the system through Upchurch BPR (15 psi). The mixture was then collected in the flask, and evaporated in vacuo. The residue was purified by MPLC (mixture of hexanes and EtOAc) providing the desired alcohols **27**.

#### 3.2.2. Preparation of Unsaturated Aminodiols

The flow setup consisted of three HPLC pumps (Knauer Azura 4.1S with 10 mL pump head). The pumps were used to introduce a solution of substrate **26** (0.5 M) into anhydrous THF (Feed A), a commercial solution of vinylmagnesium bromide (1.0 M in THF, Sigma-Aldrich, Feed B), and quenching solvent/mixture (MeOH, AcOH, mixture AcOH/H2O or 1.1 M solution of *p*TSA in MeOH), Feed C). Injection loops (PTFE, 0.8 mm i.d., 1.6 mm o.d.; internal volume: 2.0 mL, Feed A, and 2.3 mL, Feed A) were used to deliver the starting two feeds. At the beginning of the experiment, the complete reactor setup was flushed with anhydrous THF (Feed A and Feed B) and corresponding solvent/mixture (according to the conditions in Table 2, Feed C). Both solutions were loaded into their corresponding injection loops. Feed A and feed B were pumped from the injection loops and mixed in a T-shaped connector (PEEK) in a cooling bath (0 ◦C). The combined mixture passed through a coil reactor (PTFE, 0.8 mm i.d., 1.6 mm o.d.; internal volume: 9.0 mL) at 0 ◦C before the mixture was combined with Feed C (corresponding solvent or mixture) in a T-shaped connector (PEEK) at the same temperature. The mixture was then pumped through a second coil reactor at 50 or 70 ◦C (PTFE, 1.5 mm i.d., 3.2 mm o.d.; internal volume: 4.0 or 10.0 mL) or glass Omnifit column at 40 ◦C (10 mm i.d. × 100 mm length) filled with corresponding amount of Amberlyst 15. At the end, the reaction mixture left the system through Upchurch BPR and it was collected in the flask. The solvent from collected crude material was concentrated in vacuo (if the *p*TSA was used, collected stream was at first quenched with saturated water solution of NaHCO<sup>3</sup> and extracted with EtOAc). The residue was purified by MPLC (mixture of hexanes and EtOAc) providing the desired alcohol **22**.

#### 3.2.3. Carbonylative Cyclisation Using pBQ/LiCl Reoxidation System

The flow setup consisted of two HPLC pumps (Knauer Azura 4.1S with 10 mL pump head). These pumps were used to introduce a solution of substrate **22** (0.25 M) and iron pentacarbonyl (0.3 equivalent) in glacial AcOH (Feed A), and solution of *p*BQ (2.5 equivalents), LiCl (1 equivalent) and PdCl2(MeCN)<sup>2</sup> (0.1 equivalent) in the solvent (glacial AcOH or THF/AcOH = 2:1, Feed B). Injection loops (PTFE, 0.8 mm i.d., 1.6 mm o.d.; internal volume: 2.0 mL, Feed A, and 2.3 mL, Feed B) were used to deliver the two feeds. At the beginning of the experiment, the complete reactor setup was flushed with glacial AcOH (Feed A) and corresponding solvent/mixture (Feed B). Both solutions were loaded into their corresponding injection loops. (Stock solutions were pumped directly via HPLC pumps in the case of long runs). Feed A and feed B were pumped from the injection loops and mixed in a T-shaped connector (PEEK). The combined mixture went through a reactor coil (PTFE, 0.8 mm i.d., 1.6 mm o.d.; internal volume: 17.1 or 47.1 mL) at 60 ◦C before the flow stream left the system through Upchurch BPR (100 psi). The whole reaction stream was collected in the flask, and evaporated in vacuo. The residue was purified by MPLC (mixture of hexanes and EtOAc) providing the appropriate bicyclic lactones **21**.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/catal11121513/s1. All experimental procedures for batch and flow transformations, copies of <sup>1</sup>H and <sup>13</sup>C NMR spectra for all prepared compounds are included.

**Author Contributions:** Experimental work, P.L., M.G. and M.M.; design of experiments M.M. and P.K.; writing—original draft preparation, writing—review and editing, P.K. and M.M.; supervision, M.M., P.K. and T.G.; project administration, P.K. and M.M.; funding acquisition, P.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research work was funded by SLOVAK GRANT AGENCIES APVV and VEGA, grant number APVV-20-0105 and VEGA No. 1/0552/18 and VEGA No. 1/0766/20.

**Data Availability Statement:** The datasets supporting the conclusions of this article are included within the article and Supplementary Materials.

**Acknowledgments:** We acknowledge the SLOVAK GRANT AGENCIES APVV, VEGA (APVV-20- 0105 and VEGA No. 1/0552/18 and VEGA No. 1/0766/20) and Georganics Ltd. for funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


*Review*
