Lafora disease (LD) is an autosomal recessive, neurodegenerative disorder resulting in myoclonus, epilepsy, dementia, and death [1–3]. Affected individuals experience an initial seizure during adolescence, followed by severe neurological decline until the patient’s death approximately ten years after the first seizure [1, 4]. Characteristic of the disease is the cytoplasmic accumulation of hyperphosphorylated glycogen-like particles called Lafora bodies (LBs) in various tissues including brain, muscle and liver [1, 5].
Approximately 50% of Lafora disease cases are caused by mutations in the EPM2A (epilepsy of progressive myoclonus type 2 gene A) gene that encodes the protein laforin [4–6]. EPM2A is conserved in all vertebrate genomes, but it is absent from the genome of most non-vertebrate organisms including standard model organisms such as Saccharomyces cerevisiae, Caenorhabditis elegans, and Drosophila melanogaster[7, 8]. An exception to this rule is a small subgroup of protists that synthesize floridean starch, an insoluble carbohydrate similar to LBs. Five protozoan laforin orthologs have been identified; however, sequence identity between these proteins and human laforin is <37% and the genes have major insertions and deletions [7, 8]. Thus, these proteins are not optimal orthologs to utilize for modeling human laforin.
Laforin is a bimodular protein with a carbohydrate-binding module (CBM) at its amino-terminus and a dual-specificity phosphatase (DSP) domain at its carboxy-terminus [9–11]. CBMs are most commonly found in glycosyl hydrolases and glucosyl transferases from bacteria, fungi or plants, and there are over 39 families of CBMs that bind a variety of carbohydrate substrates. Laforin belongs to the CBM20 family according to the Carbohydrate-Active Enzymes (CAZy; http://www.cazy.org) database . CBM20s are closely related to CBM48s, and both are classified as starch-binding domains with similar folds and binding sites [13–15]. Typical of DSPs, laforin is capable of hydrolyzing phosphotyrosine and phosphoserine/phosphothreonine substrates; however, laforin is unique among phosphatases in that it is the only phosphatase in humans containing a CBM, which targets laforin to glycogen [8, 9]. Laforin has been shown to bind and dephosphorylate glycogen and other glucans in vitro and in vivo[8, 9, 16–19].
Glycogen is an energy storage molecule synthesized by bacterial, fungal and animal species consisting of α-1,4 and α-1,6 linked residues of glucose, with 12-14 residues per branch . Glycogen has been shown to contain small amounts of phosphate, but the regulation and effects of this phosphorylation event are currently under debate [19, 21–25]. While the source of phosphorylation is disputable, data from multiple labs has clearly established that loss of laforin activity results in hyperphosphorylation and poorly branched glycogen, resulting in insoluble LBs [17–19, 23, 26].
Although the substrate and function of laforin have recently been elucidated, the structural basis for the unique glucan phosphatase activity of laforin remains unknown. Ourselves and others have experienced difficulty purifying laforin in sufficient quantities and of sufficient quality for crystallographic studies . One group recently demonstrated that recombinant human laforin expressed in E. coli is largely insoluble and must be purified from inclusion bodies . This procedure requires denaturation and refolding steps, involves harsh chemical treatments, and often yields low amounts of correctly folded protein. A subsequent report demonstrated that only the laforin CBM was soluble when expressed in E. coli.
Our lab has purified enough recombinant laforin from the soluble portion of bacterial cell lysates to perform in vitro assays [8, 16, 29–31]. However, the protein often aggregates and precipitates after the multistep purification procedure. In this study, we found that the addition of sugars to the lysis and purification buffers increases the yield of soluble laforin from lysates and improves stability. However, such additives interfere with methods such as isothermal titration calorimetry that directly measure protein-ligand interactions. Also, we have been unable to crystallize laforin purified in the presence of sugars (unpublished data). Our group recently determined the structures of two glucan phosphatases from Arabidopsis that are functionally similar to laforin, and the structures of other DSP domains and CBMs are available [32, 33]. However, these structures provide little information about the function of laforin due to low similarity between these domains and the domains of laforin. We then sought a laforin ortholog that is highly similar to human laforin (Hs-laforin) and, when expressed in bacteria, is less prone to aggregation and precipitation. We cloned and purified multiple laforin orthologs and optimized the purification of recombinant Gallus gallus laforin (Gg-laforin). Previously, the CBM of Gg-laforin was fused to a glutathione S-transferase (GST) tag and shown to bind glycogen . In this study, we purified SUMO-tagged full-length Gg-laforin and confirmed that Gg-laforin functions as a monomer, contrary to prior claims that laforin dimerization is necessary for phosphatase activity [27, 35]. Phosphatase and glucan binding assays indicate that the catalytic and binding ability of Gg-laforin is comparable to that of Hs-laforin [8, 30, 36]. Therefore, Gg-laforin is an excellent model for Hs-laforin and a better alternative for crystallization and other biophysical studies.