Osteogenesis imperfecta mutations in plastin 3 lead to impaired calcium regulation of actin bundling

Bone Research ARTICLE www.nature.com/boneres OPEN Osteogenesis imperfecta mutations in plastin 3 lead to impaired calcium regulation of actin bundling 1234567890();,: Christopher L. Schwebach 1,2, Elena Kudryashova 1, Weili Zheng Edward H. Egelman 3 and Dmitri S. Kudryashov 1,2,4 3 , Matthew Orchard1, Harper Smith1,4, Lucas A. Runyan1, Mutations in actin-bundling protein plastin 3 (PLS3) emerged as a cause of congenital osteoporosis, but neither the role of PLS3 in bone development nor the mechanisms underlying PLS3-dependent osteoporosis are understood. Of the over 20 identified osteoporosis-linked PLS3 mutations, we investigated all five that are expected to produce full-length protein. One of the mutations distorted an actin-binding loop in the second actin-binding domain of PLS3 and abolished F-actin bundling as revealed by cryo-EM reconstruction and protein interaction assays. Surprisingly, the remaining four mutants fully retained F-actin bundling ability. However, they displayed defects in Ca2+ sensitivity: two of the mutants lost the ability to be inhibited by Ca2+, while the other two became hypersensitive to Ca2+. Each group of the mutants with similar biochemical properties showed highly characteristic cellular behavior. Wild-type PLS3 was distributed between lamellipodia and focal adhesions. In striking contrast, the Ca2+-hyposensitive mutants were not found at the leading edge but localized exclusively at focal adhesions/stress fibers, which displayed reinforced morphology. Consistently, the Ca2+-hypersensitive PLS3 mutants were restricted to lamellipodia, while chelation of Ca2+ caused their redistribution to focal adhesions. Finally, the bundling-deficient mutant failed to co-localize with any F-actin structures in cells despite a preserved F-actin binding through a non-mutation-bearing actin-binding domain. Our findings revealed that severe osteoporosis can be caused by a mutational disruption of the Ca2+-controlled PLS3’s cycling between adhesion complexes and the leading edge. Integration of the structural, biochemical, and cell biology insights enabled us to propose a molecular mechanism of plastin activity regulation by Ca2+. Bone Research (2020)8:21 ; https://doi.org/10.1038/s41413-020-0095-2 INTRODUCTION Osteoporosis is a disease defined by low bone density and disruption of the bone architecture resulting in fragility and fractures.1 Hereditary forms of bone fragility called osteogenesis imperfecta (OI) or “brittle bone disease” are mostly linked to dysregulation of Type I collagen.2 Approximately 90% of OI cases stem from mutations in collagen I genes,3,4 while most of the remaining forms affect collagen-processing enzymes involved in collagen folding, posttranslational modifications, intracellular transport, or matrix incorporation.4 Recently, several cases of OI with classical clinical manifestations in hemizygous men and a variable phenotype in heterozygous women, but without an obvious link to collagen, were attributed to mutations in an Xchromosome gene coding an actin-bundling protein plastin 3 (PLS3).5–17 Among three vertebrate tissue-specific plastin isoforms,18 PLS3 (also known as T-plastin) is ubiquitously expressed in solid tissues19 and involved in cell migration,20 endocytosis,21 DNA repair,22 and membrane trafficking.23 In agreement with the essential role of PLS3 in bone and connective tissue development in vertebrates, a pls3 knockdown in zebrafish results in craniofacial dysplasia and malformations of body axis and tail,13 whereas PLS3 knockout mouse models showed impaired cortical bone acquisition with decreased osteoblast mineralization capacity24 and defects in the development of the epidermal basal membrane.25 In humans, PLS3 mutations were also associated with a diaphragmatic hernia.26 However, a detailed understanding of PLS3’s contribution to any of the above-mentioned cellular processes or to osteogenesis is missing. The domain structure of PLS3 (Fig. 1a) encompasses the Nterminal Ca2+-binding regulatory domain (RD) and a core consisting of two actin-binding domains (ABD1 and ABD2). RD contains two EF-hands and a calmodulin-binding motif (CBM), whereas each ABD is assembled from two tandem calponinhomology (t-CH) domains. Binding of Ca2+ ions by EF-hands potently inhibits F-actin bundling, but has only a marginal effect on F-actin binding by all human isoforms,27 suggesting that only one of the ABD’s binding to actin is inhibited. RD is connected to the ABD core via a linker (Fig. 1a), whose length and likely flexibility precluded, thus far, mapping of the RD’s place in the tertiary structure of plastins; hence, the mechanism of the Ca2+-dependent regulation remains unknown. Of over 20 OI-linked mutations of PLS3 identified to date in osteoporosis patients,5–17 five are insertions or missense 1 Department of Chemistry and Biochemistry, The Ohio State University, Columbus, OH 43210, USA; 2Molecular Cellular and Developmental Biology graduate program, The Ohio State University, Columbus, OH 43210, USA; 3Department of Biochemistry and Molecular Genetics, University of Virginia, Charlottesville, VA 22908, USA and 4Biophysics graduate program, The Ohio State University, Columbus, OH 43210, USA Correspondence: Dmitri S. Kudryashov () These authors contributed equally: Christopher L. Schwebach, Elena Kudryashova Received: 2 September 2019 Revised: 6 February 2020 Accepted: 23 March 2020 © The Author(s) 2020 Mutation in PLS3 and osteoporosis CL Schwebach et al. 2 a E249_A250insI-L_ _A253_L254insN N446S _A368D EF1 EF2 10 CBM 86 Linker 100 CH1 CH2 ABD1 123 CH3 CH4 ABD2 388 396 Regulatory domain (RD) b A589QfsX21 T578NfsX4 _L478P 630 Core c A253_L254insN E249_A250insI-L 0.12 N446S WT CH3 0.09 dF/dT L478P CH2 A368D L478P N446S A253_ L254insN A368D E249_ A250insI-L 0.06 0.03 0.00 -0.03 10 CH1 d 0.2 0.0 0.8 0.6 0.4 0.2 0.0 0 10 20 30 40 -1 [PLS3]/µmol·L 50 70 85 1.0 Fraction actin bundled 0.4 Fraction actin bundled Fraction actin bound 0.6 55 f 0.1 WT L478P N446S A253_ L254insN A368D E249_ A250insI-L 40 Temperature/°C e 1.0 0.8 25 CH4 0.8 0.6 0.4 0.2 0.0 0 5 10 15 -1 [PLS3]/µmol·L 8 7 6 5 4 pCa Fig. 1 PLS3 domain structure and effects of OI-linked PLS3 mutations on PLS3 properties (see also Supplementary Fig. S1). a A schematic diagram of plastin domain structure: EF EF-hands motifs, CBM calmodulin-binding motif, RD N-terminal regulatory domain, CH calponinhomology domains, ABD actin-binding domains, Core actin-binding core domain, Linker a flexible linker separating the CBM and ABD1. PLS3 amino acid residue numbers and the OI-causative PLS3 mutations are shown below and above the diagram, respectively. b A homologybased model of the PLS3 actin-binding core (color scheme as in a) generated by Phyre2.86 c Melting profiles of PLS3 osteoporosis mutants were recorded by DSF in three independent repetitions; averaged data were plotted as the negative first derivatives (...truncated)