Frizzled receptors (FZDs) are class-F G-protein-coupled receptors (GPCRs) that function in Wnt signalling and are essential for developing and adult organisms. As central mediators in this complex signalling pathway, FZDs serve as gatekeeping proteins both for drug intervention and for the development of probes in basic and in therapeutic research. Here we present an atomic-resolution structure of the human Frizzled 4 receptor (FZD4) transmembrane domain in the absence of a bound ligand. The structure reveals an unusual transmembrane architecture in which helix VI is short and tightly packed, and is distinct from all other GPCR structures reported so far. Within this unique transmembrane fold is an extremely narrow and highly hydrophilic pocket that is not amenable to the binding of traditional GPCR ligands. We show that such a pocket is conserved across all FZDs, which may explain the long-standing difficulties in the development of ligands for these receptors. Molecular dynamics simulations on the microsecond timescale and mutational analysis uncovered two coupled, dynamic kinks located at helix VII that are involved in FZD4 activation. The stability of the structure in its ligand-free form, an unfavourable pocket for ligand binding and the two unusual kinks on helix VII suggest that FZDs may have evolved a novel ligand-recognition and activation mechanism that is distinct from that of other GPCRs.

Wnt/β-catenin signaling is activated when extracellular Wnt ligands bind Frizzled (FZD) receptors at the cell membrane. Wnts bind FZD cysteine-rich domains (CRDs) with high affinity through a palmitoylated N-terminal "thumb" and a disulfide-stabilized C-terminal "index finger," yet how these binding events trigger receptor activation and intracellular signaling remains unclear. Here we report the crystal structure of the Frizzled-4 (FZD) CRD in complex with palmitoleic acid, which reveals a CRD tetramer consisting of two cross-braced CRD dimers. Each dimer is stabilized by interactions of one hydrophobic palmitoleic acid tail with two CRD palmitoleoyl-binding grooves oriented end to end, suggesting that the Wnt palmitoleoyl group stimulates CRD-CRD interaction. Using bioluminescence resonance energy transfer (BRET) in live cells, we show that WNT5A stimulates dimerization of membrane-anchored FZD CRDs and oligomerization of full-length FZD, which requires the integrity of CRD palmitoleoyl-binding residues. These results suggest that FZD receptors may form signalosomes in response to Wnt binding through the CRDs and that the Wnt palmitoleoyl group is important in promoting these interactions. These results complement our understanding of lipoprotein receptor-related proteins 5 and 6 (LRP5/6), Dishevelled, and Axin signalosome assembly and provide a more complete model for Wnt signalosome assembly both intracellularly and at the membrane.

Wnt/β-catenin signalling is initiated by a ternary Wnt-Frizzled (FZD)-LDL receptor-related protein (LRP) 5/6 binding event. The resulting conformational changes in the FZD and LRP5/6 receptors promote the assembly of an intracellular signalosome driven by Dishevelled and Axin co-polymerization. Recent evidence suggests that the FZD receptor and LRP5/6 participate in the assembly of this signalosome by forming regulatory scaffolds for stabilizing Dishevelled and Axin adapters. In this review, we focus on the contributions of Wnts and their receptors in the assembly of the signalosome. We present an emerging model, which unifies Wnt receptor oligomerization with intracellular signalosome formation, and then discuss how FZD receptors might be targeted to either disrupt or enhance their capacity as a dynamic sensor of Wnt binding. LINKED ARTICLESThis article is part of a themed section on WNT Signalling: Mechanisms and Therapeutic Opportunities. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v174.24/issuetoc IntroductionWingless/int1 (Wnt) signal activation is tightly controlled by a dynamic signalosome consisting of Class Frizzled GPCRs (FZDs), LDL receptor-related protein (LRP) 5/6 coreceptors and Dishevelled and Axin adapters (Cong et al., 2004;Kaykas et al., 2004;Bilic et al., 2007;Gammons et al., 2016a). In this review, we discuss emerging evidence for a model of FZD and LRP5/6 co-oligomerization. In this model, oligomerization of the Wnt signalling complex is facilitated by domains with weak homo-oligomerization propensities such as the FZD cysteine-rich domain (CRD), the LRP5/6 β-propeller (βP) domains, and the Dishevelled EGL-10 and pleckstrin (DEP) and Dishevelled and Axin (DIX) domains (Dann et al., 2001;Gammons et al., 2016b). In the absence of Wnt, inactive complexes of FZD, LRP5/6 and DVL pose as static sensors for Wnt binding (Chen et al., 2014) ( Figure 1A). When Wnt binds FZD receptors and LRP5/6, the FZD receptor and LRP5/6 complexes are activated and oligomerize to create an extensive scaffold for Dishevelled stabilization and interaction with LRP5/6 (Bilic et al., 2007) ( Figure 1B). Dishevelled then co-polymerizes with Axin through shared DIX domains to form a trap which sequesters the β-catenin destruction complex (Schwarz-Romond et al., 2007;Fiedler et al., 2011).Consequentially, β-catenin translocates to the nucleus where it works in concert with Wnt transcription factors to turn on Wnt target genes (van de Wetering et al., 1991). Wnt family ligands are FZD-specific, highly conserved and often determine developmental patterning and cell fate (van de Wetering et al., 1997). Many of these physiological consequences are the result of Wnt/β-catenin signalling which involves the assembly of an intracellular Dishevelled/Axin signalosome (Cong et al., 2004; SchwarzRomond et al., 2007). Unlike norrin, an atypical FZD 4 / LRP5 agonist, all 19 human Wnts share a highly conserved two-domain structure which enables it to attach the FZD receptor CRD and bin...

Non-negative matrix factorization (NMF) is an intuitively appealing method to extract additive combinations of measurements from noisy or complex data. NMF is applied broadly to text and image processing, time-series analysis, and genomics, where recent technological advances permit sequencing experiments to measure the representation of tens of thousands of features in millions of single cells. In these experiments, a count of zero for a given feature in a given cell may indicate either the absence of that feature or an insufficient read coverage to detect that feature ("dropout"). Unlike spectral decompositions such as Singular Value Decomposition (SVD) or Principal Component Analysis (PCA), NMF is an ideal method for handling single-cell data with ambiguous zeros due to its strictly positive imputation of signal. While single-cell datasets contain many ambiguous zero counts, most analysis pipelines apply SVD or PCA on transformed counts because these implementations are fast and current NMF implementations are slow. We present an accessible NMF implementation that is much faster than PCA and rivals the runtimes of state-of-the-art SVD. NMF models learned with our implementation from raw count matrices yield intuitive summaries of complex biological processes, capturing coordinated gene activity and enrichment of sample metadata. Our NMF implementation, available in the RcppML (Rcpp Machine Learning library) R package, improves upon current NMF implementations by introducing a scaling diagonal to enable convex L1 regularization for feature engineering, reproducible factor scalings, and symmetric factorizations. RcppML NMF easily handles sparse datasets with millions of samples, making NMF an attractive replacement for PCA in the analysis of single-cell experiments.

Angiogenesis is essential for cancer metastasis, thus the discovery and characterization of molecules that inhibit this process is important. Thalidomide is a teratogenic drug which is known to inhibit angiogenesis and effectively inhibit cancer metastasis, yet the specific cellular targets for its effect are not well known. We discovered that CUL5 (previously identified as VACM-1), a scaffold protein in E3 ligase complexes, is involved in thalidomide-dependent inhibition of endothelial cell growth. Our results show that in human endothelial cells (HUVEC), thalidomide-dependent decrease in cell growth was associated with decreased nuclear localization of CUL5. In HUVEC transfected with anti-VACM-1 siRNA, thalidomide failed to decrease cell growth. Previously it was established that the antiproliferative effect of CUL5 is inhibited in rat endothelial cells (RAMEC) transfected with mutated CUL5 which is constitutively modified by NEDD8, a ubiquitin-like protein. In this study, the antiproliferative response to thalidomide was compromised in RAMEC expressing mutated CUL5. These results suggest that CUL5 protein is involved in the thalidomide-dependent regulation of cellular proliferation in vitro. Consequently, CUL5 may be an important part of the mechanism for thalidomide-dependent inhibition of cellular proliferation, as well as a novel biomarker for predicting a response to thalidomide for the treatment of disorders such as multiple myeloma and HIV infection.