**1. Introduction**

Lignans are a large natural product class of structurally and functionally diverse phenylpropanoids, isolated from over 70 known families of plants worldwide [1–4]. The lignan framework is derived from the oxidative dimerisation of two phenylpropane (C-6–C-3) moieties, which produces a linkage between each monomer's propyl side chains at the respective C-8 carbons [1,5].

Arctigenin is a natural product belonging to the dibenzylbutyrolactone subclass of lignans (Figure 1). These structures comprise a γ-lactone core, with dibenzyl substitution at the C-8 and C-8' positions in an anti-relationship [6,7]. Arctigenin possesses a range of biological activities and consequently has been well-studied to determine structure-activity relationships of its derivatives [6–12]. Results from previous studies have confirmed the lignan's biological activities, with arctigenin analogues having cytotoxic, anti-tumour and hypoglycaemic activities, amongst others [10,12–14]. Many of these derivatives have explored modifications of the aromatic rings, and to a slightly lesser extent the benzylic positions, [9,11,13,15,16], but there is a large underrepresentation of lactone ring modifications.

To date, only one synthetic derivative of arctigenin with C-9 modification has been reported—a compound containing a methylenehydroxy group at C-5 in the lactone ring (C-9 according to lignan nomenclature, Figure 1 right) [17,18]. This compound showed the induction of apoptosis in Jurkat T cells with only 2% necrosis. This derivative was accessed using an acyl-Claisen rearrangement as the key step to establish the necessary *trans* relationship between C-8 and C-8' groups. The prolific activities of arctigenin and lack of SAR information at the C-9 position inspired this work to synthesise additional C-9 analogues. Herein, we report the synthesis of 15 arctigenin derivatives with different C-9substitution, and their anti-proliferative activities.

**Citation:** Paulin, E.K.; Leung, E.; Pilkington, L.I.; Barker, D. Synthesis and Anti-Proliferative Evaluation of Arctigenin Analogues with C-9- Derivatisation. *Int. J. Mol. Sci.* **2023**, *24*, 1167. https://doi.org/10.3390/ ijms24021167

Academic Editors: Barbara De Filippis, Marialuigia Fantacuzzi and Alessandra Ammazzalorso

Received: 7 December 2022 Revised: 22 December 2022 Accepted: 4 January 2023 Published: 6 January 2023

**Copyright:** © 2023 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/).

**Figure 1.** Arctigenin (**left**) and racemic C-9methylenehydroxy analogue (**right**).

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

#### *2.1. Retrosynthetic Analysis of C-9*-*Arctigenin Analogues*

The proposed pathway to access the targeted arctigenin analogues also exploited an acyl-Claisen rearrangement [17] to introduce the correct relative stereochemistry between benzyl groups in the morpholine pentenamide **1** and converged two parallel pathways (Scheme 1). Cyclisation of the rearrangement product was envisaged to establish the core DBL lactone framework and included the C-9 substitution of a methylenehydroxy group, from which derivatisations could be prepared. The acyl-Claisen precursors, an acid chloride **2** and allylic morpholine **3**, could be prepared from vanillin **4** and 4-allyl-1,2 dimethoxybenzene **5**, respectively, through two separate routes.

**Scheme 1.** Retrosynthetic approach to racemic C-9analogues.

#### *2.2. Synthesis of Acid Chloride* **2**

Acid chloride **2** was prepared when required from the more stable carboxylic acid **6**. Synthesis of **6** began from vanillin **4**, which was subjected to a Wittig olefination with (carbethoxymethylene)triphenylphosphorane, following literature methods [17] to give α,βunsaturated ethyl ester **7** in 80% yield (*E*:*Z*, 92:8). The newly installed alkene was reduced by catalytic hydrogenation to give saturated ester **8** in quantitative yield (Scheme 2). With the saturated ester **8** in hand, the phenol substituent was protected as the benzyl ether to give **9** in 93% yield. Hydrolysis of the ester **9** gave carboxylic acid **6** in 92% yield and

the resulting acid chloride was prepared in situ at the time of the successive step due to its instability.

**Scheme 2.** Synthesis of acid **6**. Reagents and conditions: (**i**) Ph3PCHCO2Et (1.1 equiv.), CH2Cl2, rt, 21 h, **7** 80%; (**ii**) H2, Pd/C (10% *w*/*w*), EtOAc, rt, 20 h, **8** quant.; (**iii**) BnBr (3 equiv.), K2CO3 (3 equiv.), MeCN, 80 ◦C, 46 h, **9** 93%; (**iv**) NaOH (4.5 equiv.), MeOH, rt, 2.5 h, **6** 92%.

#### *2.3. Synthesis of Allylic Morpholine* **3**

To obtain allylic morpholine **3**, 4-allyl-1,2-dimethoxybenzene **5** was first dehydroxylated, under Upjohn conditions [19], to give diol **10** in 94% yield (Scheme 3). The newly formed diol moiety then underwent oxidative cleavage using NaIO4 [17], to afford aldehyde **11** in 97% yield, which was used immediately, due to its tendency to degrade even stored at low temperatures. Thus, **11** underwent a Horner-Wadsworth-Emmons (HWE) reaction, forming the respective (*E*)-α,β-unsaturated ethyl ester **12** in 73% yield [20–22]. During repeated syntheses of α,β-unsaturated ester **12**, the *E*-selectivity of the Horner-Wadsworth-Emmons reaction was found to vary with the formation of both the *Z*-isomer and a third regioisomer, which was determined to be a β,γ-unsaturated ester **13** (Scheme 3). Separation of the desired *E*-isomer **12** on AgNO3-treated silica was possible but poor, leading to diminished yields.

**Scheme 3.** Synthesis of allylic ester **12**. Reagents and conditions: (**i**) OsO4 (2.5 mol-%), NMO (3 equiv.), *t*-BuOH/H2O (1:1), rt, 4 days, **10** 94%; (**ii**) NaIO4 (1.2 equiv.), MeOH/H2O (3:1), rt, 3 days, **11** 97%; (**iii**) (EtO)2POCH2CO2Et (1.5 equiv.), NaH (2 equiv.), THF, 0 ◦C to rt, 18 h, *E*-**12**, 39%, *Z*-**12** 2%, **13** 7%.

#### *2.4. Prevention of Isomeric Esters*

Only *E*-isomer *E*-**12** was required for further steps, therefore, in order to prevent double bond migration to form **13**, different conditions were trialled for the formation of allylic ester **12**. While the mechanism of rearrangement was not confirmed, it was proposed to be base-mediated, through abstraction of the γ-proton after formation of the initial α,β-unsaturated product (Scheme 3), with the resulting β,γ-unsaturated product **13** stabilised by increased conjugation. As a result, different bases were screened. The migrated isomer was observed to a lesser extent under kinetic control or with the use of hindered bases, such as DBU under Masamune-Roush conditions [23], but unfortunately, these reactions still had poor *E*/*Z* stereocontrol. After reports of MeMgBr use to suppress isomerisation in PhCH2CHO aldehydes, this was applied as a base in the HWE reaction between triethylphosphonoacetate and aldehyde **11** [24]. As a result, neither the migrated species **13** nor *Z*-isomer *Z*-**12** were observed. On a larger scale, good selectivity was maintained (9:1 *E*:*Z*), but unfortunately, the yield was poor (10%), so the exploration of other methods was resumed.

#### *2.5. Grubbs Cross Metathesis Pathway; Revised Route to* **3**

An alternative route which did not involve a HWE reaction was then developed by implementing a cross metathesis [25] approach between 4-allyl-1,2-dimethoxybenzene **5** and ethyl acrylate. Using Grubb's second-generation catalyst at a loading of 5 mol-% and three equivalents of ethyl acrylate, full conversion to the *E* product *E*-**12** took place in 91% yield. No migrated product **13** was observed, allowing large scale synthesis of ester *E*-**12**.

Ester *E*-**12** was then fully reduced to primary allylic alcohol **14** using DIBAL-H (Scheme 4). The final step towards allylic morpholine **3** involved substitution of the hydroxyl group in **14** for a morpholine moiety. The reaction was attempted using various strategies, including via mesylation, tosylation and bromination with all giving the desired product **3**, but in poor yields.

**Scheme 4.** Synthetic pathway to allylic morpholine **3** using cross metathesis approach. Reagents and conditions: (**i**) Ethyl acrylate (3 equiv.), Grubbs II (1.9 mol-%), CH2Cl2, rt, 24 h, **12** 91%; (**ii**) DIBAL-H (2.9 equiv.), PhMe, −10 ◦C to rt, 20 h, **14** quant.; (**iii**) Ac2O (2.2 equiv.), Et3N (3 equiv.), DMAP (10 mol-%), CH2Cl2, 0 ◦C to rt, 25 h, **15**; (**iv**) morpholine (1.9 equiv.), Pd(PPh3)4 (5 mol-%), THF, reflux, 5 days, **3** 83% (two steps).

In an alternate approach, acetate **15** was then synthesised from alcohol **14**, then subjected to Tsuji-Trost allylation conditions, using palladium tetrakis Pd(PPh3)4 and morpholine. With the possibility of two regioisomers of the allylic amine product, thermodynamic control was implemented to ensure the desired linear isomer was obtained over the possible kinetic branched product [26]. Over two steps from the allylic alcohol **14**, the desired allylic morpholine **3** was achieved in 83% yield as solely the *E*-isomer, linear product (Scheme 4).
