Search

Real-time monitoring of biological activity can be achieved through the use of genetically encoded fluorescent indicators (GEFIs). GEFIs are protein-based sensing tools whose biophysical characteristics can be engineered to meet experimental needs. However, GEFIs are inherently complex proteins with multiple dynamic states, rendering optimization one of the most challenging problems in protein engineering. Most GEFIs are engineered through trial-and-error mutagenesis, which is time and resource-intensive and often relies on empirical knowledge for each GEFI. We applied an alternative approach using machine learning to efficiently predict the outcomes of sensor mutagenesis by analyzing established libraries that link sensor sequences to functions. Using the GCaMP calcium indicator as a scaffold, we developed an ensemble of three regression models trained on experimentally derived GCaMP mutation libraries. We used the trained ensemble to perform an in silico functional screen on a library of 1423 novel, untested GCaMP variants. The mutations were predicted to significantly alter the fluorescent response, and off-rate kinetics were advanced for verification in vitro. We found that the ensemble’s predictions of novel variants’ biophysical characteristics closely replicated what we observed of the variants in vitro. As a result, we identified the novel ensemble-derived GCaMP (eGCaMP) variants, eGCaMP and eGCaMP+, that achieve both faster kinetics and larger fluorescent responses upon stimulation than previously published fast variants. Furthermore, we identified a combinatorial mutation with extraordinary dynamic range, eGCaMP2+, that outperforms the tested 6th, 7th, and 8th generation GCaMPs. These findings demonstrate the value of machine learning as a tool to facilitate the efficient prescreening of mutants for functional characteristics. By leveraging the learning capabilities of our ensemble, we were able to accelerate the identification of promising mutations and reduce the experimental burden associated with screening an entire library. Machine learning tools such as this have the potential to complement emerging high-throughput screening methodologies that generate massive datasets, which can be tedious to analyze manually. Overall, these findings have significant implications for developing new GEFIs and other protein-based tools, demonstrating the power of machine learning as an asset in protein engineering.

Through the efforts of many groups, a wide range of fluorescent protein reporters and sensors based on green fluorescent protein and its relatives have been engineered in recent years. Here we explore the incorporation of sensing modalities into de novo designed fluorescence-activating proteins, called mini-fluorescence-activating proteins (mFAPs), that bind and stabilize the fluorescent cis-planar state of the fluorogenic compound DFHBI. We show through further design that the fluorescence intensity and specificity of mFAPs for different chromophores can be tuned, and the fluorescence made sensitive to pH and Ca2+ for real-time fluorescence reporting. Bipartite split mFAPs enable real-time monitoring of protein–protein association and (unlike widely used split GFP reporter systems) are fully reversible, allowing direct readout of association and dissociation events. The relative ease with which sensing modalities can be incorporated and advantages in smaller size and photostability make de novo designed fluorescence-activating proteins attractive candidates for optical sensor engineering., Fluorescent protein reporters based on GFP exist, but have intrinsic disadvantages. Here the authors incorporate pH, Ca2+ and protein–protein interaction sensing modalities into de novo designed mini-fluorescence-activating proteins (mFAPs), with increased photostability and smaller size, which bind a range of DFHBI chromophore variants.

INTRODUCTION: The Wnt proteins play key roles in the development, homeostasis and disease progression of many organs including the prostate. However, the spatiotemporal expression patterns of Wnt proteins in prostate cell lineages at different developmental stages and in prostate cancer remain inadequately characterized. METHODS: We isolated the epithelial and stromal cells in the developing and mature mouse prostate by flow cytometry and determined the expression levels of Wnt lignads. We used Visium spatial gene expression analysis to determine spatial distribution of Wnt ligands in the mouse prostatic glands. Using laser-capture microscopy in combination with gene expression analysis, we also determined the expression patterns of Wnt signaling components in stromal and cancer cells in advanced human prostate cancer specimens. To investigate how the stroma-derived Wnt ligands affect prostate development and homeostasis, we used a Col1a2-CreER(T2) mouse model to disrupt the Wnt transporter Wntless (Wls) specifically in prostate stromal cells. RESULTS: We showed that the prostate stromal cells are a major source of several Wnt ligands. Visium spatial gene expression analysis revealed a distinct spatial distribution of Wnt ligands in the prostatic glands. We also showed that Wnt signaling components are highly expressed in the stromal compartment of primary and advanced human prostate cancer. Blocking stromal Wnt secretion attenuated prostate epithelial proliferation and regeneration but did not affect cell survival and lineage maintenance. DISCUSSION: Our study demonstrates a critical role for stroma derived Wnt ligands in prostate development and homeostasis.