Department of Biological Chemistry, UCLA, David Geffen School of Medicine

Human lens fiber membrane intrinsic protein MP20 is the second most abundant membrane protein of the human eye lens. Despite decades of effort its structure and function remained elusive. Here, we determined the MicroED structure of full-length human MP20 in lipidic-cubic phase to a resolution of 3.5 Å. MP20 forms tetramers each of which contain 4 transmembrane α-helices that are packed against one another forming a helical bundle. We find that each MP20 tetramer formed adhesive interactions with an opposing tetramer in a head-to-head fashion. Investigation of MP20 localization in human lenses indicate that in young fiber cells MP20 is initially localized to the cytoplasm in differentiating fiber cells but upon fiber cell maturation is inserted into the plasma membrane, correlating with the restriction of the diffusion of extracellular tracers into the lens. Together these results suggest that MP20 forms lens thin junctions in vivo, confirming its role as a structural protein in the human eye lens essential for its optical transparency.

High-resolution information is important for accurate structure modeling but is challenging to attain in macromolecular crystallography due to the rapid fading of diffracted intensities at increasing resolution. While direct electron detection essentially eliminates the read-out noise during MicroED data collection, other sources of noise remain and limit the measurement of faint high-resolution reflections. Inelastic scattering significantly contributes to noise, raising background levels and broadening diffraction peaks. We demonstrate a substantial improvement in signal-to-noise ratio by using energy filtering to remove inelastically scattered electrons. This strategy results in sub-atomic resolution MicroED data from proteinase K crystals, enabling the visualization of detailed structural features. Interestingly, reducing the noise further reveals diffuse scattering that may hold additional structural information. Our findings suggest that combining energy filtering and direct detection provides more accurate measurements at higher resolution, facilitating precise model refinement and improved insights into protein structure and function.

The detailed understanding of the conformational pathway of fluticasone, a widely prescribed medicine for allergic rhinitis, asthma, and chronic obstructive pulmonary disease (COPD), from formulation to its protein-bound state, has been limited due to a lack of access to its high-resolution structures. The three-dimensional (3D) structure of fluticasone furoate 1 remains unpublished, and the deposited structure of fluticasone propionate 2 could be further refined due to refinement against new data. We applied microcrystal electron diffraction (MicroED) to determine the 3D structures of 1 and 2 in their solid states. The preferred geometries in solution were predicted by using density functional theory (DFT) calculations. A comparative analysis of the structures of 1 and 2 across three states (in solid state, in solution, and protein-bound conformation) revealed the course of the conformational changes during the entire transition. Potential energy plots were calculated for the most dynamic bonds, uncovering their rotational barriers. This study underscores the combined use of MicroED and DFT calculations to provide a comprehensive understanding of conformational and energy changes during drug administration. The quantitative comparison also highlights the subtle structural differences that may lead to significant changes in the pharmaceutical properties.

Mimicking metabolic pathways on electrodes enables in vivo metabolite monitoring for decoding metabolism. Conventional in vivo sensors cannot accommodate underlying complex reactions involving multiple enzymes and cofactors, addressing only a fraction of enzymatic reactions for few metabolites. We devised a single-wall-carbon-nanotube-electrode architecture supporting tandem metabolic pathway-like reactions linkable to oxidoreductase-based electrochemical analysis, making a vast majority of metabolites detectable in vivo. This architecture robustly integrates cofactors, self-mediates reactions at maximum enzyme capacity, and facilitates metabolite intermediation/detection and interference inactivation through multifunctional enzymatic use. Accordingly, we developed sensors targeting 12 metabolites, with 100-fold-enhanced signal-to-noise ratio and days-long stability. Leveraging these sensors, we monitored trace endogenous metabolites in sweat/saliva for noninvasive health monitoring, and a bacterial metabolite in the brain, marking a key milestone for unraveling gut microbiota-brain axis dynamics.

How early sensory experience during "critical periods" of postnatal life affects the organization of the mammalian neocortex at the resolution of neuronal cell types is poorly understood. We previously reported that the functional and molecular profiles of layer 2/3 (L2/3) cell types in the primary visual cortex (V1) are vision-dependent [S. Cheng et al., Cell 185, 311-327.e24 (2022)]. Here, we characterize the spatial organization of L2/3 cell types with and without visual experience. Spatial transcriptomic profiling based on 500 genes recapitulates the zonation of L2/3 cell types along the pial-ventricular axis in V1. By applying multitasking theory, we suggest that the spatial zonation of L2/3 cell types is linked to the continuous nature of their gene expression profiles, which can be represented as a 2D manifold bounded by three archetypal cell types. By comparing normally reared and dark reared L2/3 cells, we show that visual deprivation-induced transcriptomic changes comprise two independent gene programs. The first, induced specifically in the visual cortex, includes immediate-early genes and genes associated with metabolic processes. It manifests as a change in cell state that is orthogonal to cell-type-specific gene expression programs. By contrast, the second program impacts L2/3 cell-type identity, regulating a subset of cell-type-specific genes and shifting the distribution of cells within the L2/3 cell-type manifold. Through an integrated analysis of spatial transcriptomics with single-nucleus RNA-seq data, we describe how vision patterns cortical L2/3 cell types during the critical period.

We recently reported that microinjection of Xenopus nodal-related (xnr) mRNAs into β-catenin-depleted Xenopus embryos rescued a complete dorsal axis. Xnrs mediate the signal of the Nieuwkoop center that induces the Spemann-Mangold organizer in the overlying mesoderm, a process inhibited by the Nodal antagonist Cerberus-short (CerS). However, β-catenin also induces a second signaling center in the dorsal prospective ectoderm, designated the Blastula Chordin and Noggin Expression (BCNE) center, in which the homeobox gene siamois (sia) plays a major role. In this study, we asked whether the Xnrs and Sia depend on each other or function on parallel pathways. Expression of both genes induced β-catenin-depleted embryos to form complete axes with heads and eyes via the activation of similar sets of downstream organizer-specific genes. Xnrs did not activate siamois, and, conversely, Sia did not activate xnrs, although both were induced by β-catenin stabilization. Depletion with morpholinos revealed a robust role for the downstream target Chordin. Remarkably, Chordin depletion prevented all ectopic effects resulting from microinjection of the mRNA encoding the maternal cytoplasmic determinant Huluwa, including the radial expansion of brain tissue and the ectopic expression of the ventral gene sizzled. The main conclusion was that the BCNE and Nieuwkoop centers provide a double assurance mechanism for axial formation by independently activating similar downstream transcriptional target gene repertoires. We suggest that Siamois likely evolved from an ancestral Mix-type homeodomain protein called Sebox as a Xenopus-specific adaptation for the rapid differentiation of the anterior neural plate in the ectoderm.

Patterning of DNA methylation in eukaryotic genomes is controlled by de novo methylation, maintenance mechanisms and demethylation pathways. In Arabidopsis thaliana, DNA demethylation enzymes are clearly important for shaping methylation patterns, but how they are regulated is poorly understood. Here we show that the targeting of histone H3 lysine four trimethylation (H3K4me3) with the catalytic domain of the SDG2 histone methyltransferase potently erased DNA methylation and gene silencing at FWA and also erased CG DNA methylation in many other regions of the Arabidopsis genome. This methylation erasure was completely blocked in the ros1 dml2 dml3 triple mutant lacking DNA demethylation enzymes, showing that H3K4me3 promotes the active removal of DNA methylation. Conversely, we found that the targeted removal of H3K4me3 increased the efficiency of targeted DNA methylation. These results highlight H3K4me3 as a potent anti-DNA methylation mark and also pave the way for development of more powerful epigenome engineering tools.

Caspases are a family of highly homologous cysteine proteases that play critical roles in inflammation and apoptosis. Small molecule inhibitors are useful tools for studying caspase biology, complementary to genetic approaches. However, achieving inhibitor selectivity for individual members of this highly homologous enzyme family remains a major challenge in developing such tool compounds. Prior studies have revealed that one strategy to tackle this selectivity gap is to target the precursor or zymogen forms of individual caspases, which share reduced structural homology when compared to active proteases. To establish a screening assay that favors the discovery of zymogen-directed caspase-10 selective inhibitors, we engineered a low-background and high-activity tobacco etch virus (TEV)-activated caspase-10 protein. We then subjected this turn-on protease to a high-throughput screen of approximately 100 000 compounds, with an average Z value of 0.58 across all plates analyzed. Counter screening, including against TEV protease, delineated bona fide procaspase-10 inhibitors. Confirmatory studies identified a class of thiadiazine-containing compounds that undergo isomerization and oxidation to generate cysteine-reactive compounds with caspase-10 inhibitory activity. In parallel, mode-of-action studies revealed that pifithrin-μ (PFTμ), a reported TP53 inhibitor, also functions as a promiscuous caspase inhibitor. Both inhibitor classes showed preferential zymogen inhibition. Given the generalized utility of activation assays, we expect our screening platform to have widespread applications in identifying state-specific protease inhibitors.

Natural products are ligands and in vitro inhibitors of Alzheimers disease (AD) tau. Dihydromyricetin (DHM) bears chemical similarity to known natural product tau inhibitors. Despite having signature polyphenolic character, DHM is ostensibly hydrophobic owing to intermolecular hydrogen bonds that shield hydrophilic phenols. Our research shows DHM becomes ionized at near-neutral pH, allowing the formulation of salts with transformed solubility. The MicroED co-crystal structure with trolamine reveals DHM salts as metastable co-crystalline solids with unlocked hydrogen bonding and a thermodynamic bent to solubilize in water. All co-crystal formulations show better inhibitory activity against AD tau than the nonsalt form, with efficacies correlating to enhanced solubilities. In vitro and in vivo pharmacokinetic measures demonstrate that DHM co-crystals display enhanced absorption and distribution with altered rates of elimination, suggesting that co-crystal formulations could be strategically used to fine-tune delivery properties. These results underscore the role of structural chemistry in guiding the selection of solubilizing agents for chemical formulation. We propose DHM co-crystals are appropriate formulations for research as dietary supplements to promote healthy aging by combating protein misfolding, although central nervous system (CNS) delivery remains a major limitation. DHM may be a suitable backbone for medicinal chemistry and possible development of pharmaceuticals with enhanced CNS exposure.

Archaea share genomic similarities with Eukarya and cellular architectural similarities with Bacteria, though archaeal and bacterial surface layers (S-layers) differ. Using cellular cryo-electron tomography, we visualized the S-layer lattice surrounding Methanospirillum hungatei, a methanogenic archaeon. Though more compact than known structures, M. hungatei's S-layer is a flexible hexagonal lattice of dome-shaped tiles, uniformly spaced from both the overlying cell sheath and the underlying cell membrane. Subtomogram averaging resolved the S-layer hexamer tile at 6.4-angstrom resolution. By fitting an AlphaFold model into hexamer tiles in flat and curved conformations, we uncover intra- and intertile interactions that contribute to the S-layer's cylindrical and flexible architecture, along with a spacer extension for cell membrane attachment. M. hungatei cell's end plug structure, likely composed of S-layer isoforms, further highlights the uniqueness of this archaeal cell. These structural features offer advantages for methane release and reflect divergent evolutionary adaptations to environmental pressures during early microbial emergence.