A-type lamins and nuclear architecture
Lamins are nuclear intermediate filament proteins with essential roles in nuclear shape, architecture and function. Mammalian cells encode two types of lamins, termed A- and B-type. More than ten diseases ('laminopathies') map to mutations in LMNA, the gene encoding A-type lamins. To explain these diseases, A-type lamins were proposed to be important for either the mechanical integrity of the nucleus, or tissue-specific gene regulation, or both. My thesis work tested and supported both models. The mechanical model was supported by my discovery that lamins bind directly to nuclear titin in vivo and in vitro, and that A-type lamins also specifically bind and bundle F-actin in vitro. Titin and actin were each identified as proteins that bind the Ig-fold domain of A-type lamins in a two-hybrid screen of a skeletal muscle cDNA library. Titin was recovered twice and actin forty-seven times out of ∼200 positives. Biochemical analysis showed a small polypeptide near the C-terminus of titin was sufficient to bind A- and B-type lamins. When expressed in mammalian cells, this lamin-binding fragment competed for binding to lamins and profoundly disrupted nuclear architecture, demonstrating the interaction is biologically relevant. Furthermore the nuclear envelope localization of titin depends on lamin, as shown by RNA-interference-mediated downregulation studies in Caenorhabditis elegans. This work established nuclear titin as a novel architectural partner for A-type lamins in vivo. G-actin was previously shown by others to bind residues 563-646 in the tail of lamin A. I discovered a second actin-binding region comprising residues 461-536 (the Ig-fold domain), and proposed that lamins might therefore bundle or crosslink F-actin. Through biochemical analysis I found the tail domain of lamin A, but not lamin B1, binds F-actin with an affinity of 3 nM. This affinity was reduced four-fold by the 50-residue deletion that causes Hutchinson-Gilford Progeria Syndrome (HGPS); this deletion removes part of one actin-binding region but does not affect the Ig-fold domain. Interestingly both wildtype and HGPS-mutant lamin A tails were able to bundle F-actin, suggesting the existence of a third actin-binding site. The "gene expression" model for A-type lamins was supported by my microarray and Real Time PCR analysis of mRNA levels in the hearts of lmna-null mice versus wildtype siblings. Six affected genes were identified, of which three showed increased expression (TRAP220, TNF-α induced protein 8, and Riken cDNA NM_023557) and three showed decreased expression (Special AT rich Binding protein 1, Hypoxia induced gene 1, Splicing factor Arg/Ser rich 1) in lmna-null hearts. This pilot study demonstrated that the molecular mechanism of cardiac phenotypes in EDMD and other laminopathies is likely due to perturbation of multiple activities including transcriptional regulation, mRNA splicing, and cell signaling. In summary, my work suggests that A-type lamins have multiple roles in both nuclear architecture and gene expression.