Structure confirmation, reactivity, bacterial mutagenicity and quantification of 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione in drinking water

Abstract

The presence of two new disinfectant by-product (DBP) groups in the UK was recently shown using non-target analysis, halogenated-hydroxycyclopentenediones and halogenated-methanesulfonic acids. In this work, we confirmed the structure of 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione (TBHCD), and quantified it together with dibromomethanesulfonic acid at 122 ± 34 and 326 ± 157 ng L^-1 on average in London's drinking water, respectively (n = 21). We found TBHCD to be photolabile and unstable in tap water and at alkaline pH. Furthermore, spectral and computational data for TBHCD and three other halogenated-hydroxycyclopentenediones indicated they could act as a source of radicals in water and in the body. Importantly, TBHCD was calculated to have a 14.5 kcal mol^-1 lower C-Br bond dissociation enthalpy than the N-Br bond of N-bromosuccinimide, a common radical substitution reagent used in organic synthesis. TBHCD was mutagenic in Salmonella/microsome assays using strains TA98, TA100 and TA102. This work reveals the unique features, activity and toxicity of trihalogenated hydroxycyclopent-4-ene-1,3-diones, prompting a need to more comprehensively assess their risks.

Fig. 1. Synthesis, NMR characterisation and structural assignment of 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione.

A Synthetic route, B ¹³C{¹H} NMR inverse-gated spectrum and C X-ray molecular structure of 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione (50% probability ellipsoids).

Fig. 2. Mean concentrations of 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione (TBHCD) and dibromomethanesulfonic acid (DBMSA) in tap water from seven London households.

Samples were taken in mid-August, end of August and mid-September 2023 (see also Tables S2 and S3). Error bars represent standard deviation over samples taken at three separate timepoints and not error uncertainty across repeated measurement of the same sample.

Fig. 3. Calculation of C-Br and C-Cl homolytic cleavage BDEs (in kcal mol^-1) for the bonds of N-bromosuccinimide.

A 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione (TBHCD) (B), 2,4-dibromo-2-chloro-5-hydroxycyclopent-4-ene-1,3-dione (D), 2-bromo-2,4-dichloro-5-hydroxycyclopent-4-ene-1,3-dione (E), and 2,2,4-trichloro-5-hydroxycyclopent-4-ene-1,3-dione (F). Spin density distribution plot of the radical anion generated with C-Br homolysis of TBHCD is represented in C, spin density in purple.

Fig. 4. Homolysis of 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione.

A Resonance and tautomeric structures for deprotonated 2,4-dibromo-5-hydroxycyclopent-4-ene-1,3-dione in drinking water (B); resonance structures for deprotonated 2,4-dibromo-5-hydroxycyclopent-4-ene-2-yl-1,3-dione (C). All structures are represented in anionic form reflecting ionization state at nearly-neutral pH.

Fig. 5. Results of a representative experiment reporting mutagenic activities 2,2,4-tribromo-5-hydroxycyclopent-4-ene-1,3-dione with Salmonella strains.

A TA98, B TA100, C TA102 without metabolic activation, and D strain 100 with metabolic activation. Columns indicate mean ± SD from three plates. * positive according to the two-fold rule.