Data collection statistics are from HKL3000 (Otwinowski and Minor, 1997) or XDS (I- experiment) (Kabsch, 2010). Rsym = Σ | Ii- < Ii > | / Σ Ii, where < Ii > is the average intensity of symmetry-equivalent reflections. CC1/2, CCwork and CCfree are defined in (Karplus and Diederichs, 2012). Phasing power = RMS (|F|/ε), where |F| is the heavy-atom structure factor amplitude and ε is the residual lack of closure error. Rcullis is the mean residual lack of closure error divided by the dispersive or anomalous difference. Rwork = Σ | Fobs – Fcalc | / Σ | Fobs |, where Fobs and Fcalc are the observed and calculated structure factors, respectively. Rfree is calculated using a subset (~10%) of reflection data chosen randomly and omitted throughout refinement. Figure of merit is indicated after density modification and phase extension starting from 9.0 Å in DM. R.m.s.d: root mean square deviations from ideal geometry. Numbers in parentheses indicate the highest resolution shells and their statistics.