Elucidating the structure and interactions of proteins in native environments is a fundamental goal of structural biology. Nuclear magnetic resonance (NMR) spectroscopy is well suited for this task but often suffers from low sensitivity, especially in complex biological settings. Here, we use a sensitivity-enhancement technique called dynamic nuclear polarization (DNP) to overcome this challenge. We apply DNP to capture the membrane interactions of the outer membrane protein Ail, a key component of the host invasion pathway of Yersinia pestis. We show that the DNP-enhanced NMR spectra of Ail in native bacterial cell envelopes are well resolved and enriched in correlations that are invisible in conventional solid-state NMR experiments. Furthermore, we demonstrate the ability of DNP to capture elusive interactions between the protein and the surrounding lipopolysaccharide layer. Our results support a model where the extracellular loop arginine residues remodel the membrane environment, a process that is crucial for host invasion and pathogenesis.

Characterizing the epoch of reionization (EoR) at z 6 via the redshifted 21 cm line of neutral Hydrogen (H i) is critical to modern astrophysics and cosmology, and thus a key science goal of many current and planned low-frequency radio telescopes. The primary challenge to detecting this signal is the overwhelmingly bright foreground emission at these frequencies, placing stringent requirements on the knowledge of the instruments and inaccuracies in analyses. Results from these experiments have largely been limited not by thermal sensitivity but by systematics, particularly caused by the inability to calibrate the instrument to high accuracy. The interferometric bispectrum phase is immune to antenna-based calibration and errors therein, and presents an independent alternative to detect the EoR H i fluctuations while largely avoiding calibration systematics. Here, we provide a demonstration of this technique on a subset of data from the Hydrogen Epoch of Reionization Array (HERA) to place approximate constraints on the brightness temperature of the intergalactic medium (IGM). From this limited data, at z=7.7 we infer "1σ"upper limits on the IGM brightness temperature to be ≤ 316 "pseudo"mK at κ =0.33 "pseudo"h Mpc-1 (data-limited) and ≤ 1000 "pseudo"mK at κ =0.875 "pseudo"h Mpc-1 (noise-limited). The "pseudo"units denote only an approximate and not an exact correspondence to the actual distance scales and brightness temperatures. By propagating models in parallel to the data analysis, we confirm that the dynamic range required to separate the cosmic H i signal from the foregrounds is similar to that in standard approaches, and the power spectrum of the bispectrum phase is still data-limited (at 106 dynamic range) indicating scope for further improvement in sensitivity as the array build-out continues.