Structure Revision of Four Acylphloroglucinols Isolated from the Leaves of Syzygium polyanthum

The plant family Myrtaceae is a major source of phloroglucinol natural products with potent and varied biological activities [1]. Syzygium is a diverse genus within the family and contains over 1000 rainforest shrub and tree species that occur mainly in South East Asia and Australasia [2]. The importance of this genus is borne out by the many species that are used as foods, medicines, and timber, with for example, the spice trade from the 17th century being based upon the exploitation of spices such as cloves, the intensely aromatic flower buds of Syzygium aromaticum (L.) Merr. & L.M. Perry, harvested from plants growing on a select few islands of the Indonesian archipelago, to be sold in markets in Europe [3]. Surprisingly only a relatively few papers have reported on the non-volatile chemicals present in Syzygium species but acylphloroglucinol compounds predominate, with over 20 compounds from this class being reported [1,4–6]. It was therefore surprising when Saifudin et al. reported in 2012 that the leaves of Syzygium polyanthum (Wight) Walp. collected in Indonesia contain four compounds with an unusual 1-acyl-2,3,5-trihydroxy-4-methylphenyl substitution pattern [7]. Three of these compounds showed micromolar inhibition of the enzyme PTP1B that has been implicated with diabetes [7]. Further investigation of the literature revealed that another study on the same species, published at about the same time by Har et al., reported compounds containing the more typical acylphloroglucinol structure class and interestingly, these compounds had similar length acyl side chains to those reported by Saifudin et al. [5,7]. An additional paper also reported long chain acylphloroglucinol derivatives from Syzygium levinei (Merr.) Merr., although these compounds lacked an aromatic methyl group [6]. These conflicting reports suggested that one of the group of authors may have incorrectly assigned their structures, and since the compounds reported by Saifudin et al. do not fit the expected polyketide biogenetic path-