*Article* **Biochemical Characterization of the** *GBA2* **c.1780G>C Missense Mutation in Lymphoblastoid Cells from Patients with Spastic Ataxia**

**Anna Malekkou 1,2,†, Maura Samarani 3,†, Anthi Drousiotou 1,2,\*, Christina Votsi 2,4, Sandro Sonnino 3, Marios Pantzaris 2,5, Elena Chiricozzi 3, Eleni Zamba-Papanicolaou 2,6, Massimo Aureli 3,\*, Nicoletta Loberto 3,‡ and Kyproula Christodoulou 2,4,‡**


Received: 14 August 2018; Accepted: 8 October 2018; Published: 10 October 2018

**Abstract:** The *GBA2* gene encodes the non-lysosomal glucosylceramidase (NLGase), an enzyme that catalyzes the conversion of glucosylceramide (GlcCer) to ceramide and glucose. Mutations in *GBA2* have been associated with the development of neurological disorders such as autosomal recessive cerebellar ataxia, hereditary spastic paraplegia, and Marinesco-Sjogren-Like Syndrome. Our group has previously identified the *GBA2* c.1780G>C [p.Asp594His] missense mutation, in a Cypriot consanguineous family with spastic ataxia. In this study, we carried out a biochemical characterization of lymphoblastoid cell lines (LCLs) derived from three patients of this family. We found that the mutation strongly reduce NLGase activity both intracellularly and at the plasma membrane level. Additionally, we observed a two-fold increase of GlcCer content in LCLs derived from patients compared to controls, with the C16 lipid being the most abundant GlcCer species. Moreover, we showed that there is an apparent compensatory effect between NLGase and the lysosomal glucosylceramidase (GCase), since we found that the activity of GCase was three-fold higher in LCLs derived from patients compared to controls. We conclude that the c.1780G>C mutation results in NLGase loss of function with abolishment of the enzymatic activity and accumulation of GlcCer accompanied by a compensatory increase in GCase.

**Keywords:** GBA2; non-lysosomal β-glucosylceramidase; β-glucocerebrosidase; spastic ataxia; glucosylceramide; plasma membrane; lymphoblastoid cell lines

#### **1. Introduction**

Sphingolipids (SLs) are a class of lipids mainly associated with the external leaflet of the plasma membrane (PM) of all eukaryotic cells, playing an important role in the structural integrity of the PM and cellular signaling [1]. Glycosphingolipids (GSLs) are SLs with a head-group formed by a mono- or oligosaccharide moiety. Glucosylceramide (GlcCer) is the simplest member of GSLs and it is formed in the Golgi complex by the glycosylation of ceramide [2]. SLs play a fundamental role in cell physiology and this can be demonstrated by the numerous genetic diseases that arise from mutations in enzymes involved in SL metabolism and transport [3,4]. Cells can alter their lipid composition by the action of different hydrolases that are active at the lysosomes or at the PM, such as sphingomyelinase (SMase), β-hexosaminidase (β-Hex), β-galactosidase (β-gal), β-glucocerebrosidase (GCase), and the non-lysosomal β-glucosylceramidase (NLGase) [5,6].

GCase and NLGase are both involved in the catabolism of GlcCer to ceramide and glucose. GCase (EC 3.2.1.45) is encoded by the *GBA* gene (MIM\_606463), located on chromosome 1q21. GCase is a membrane glycoprotein of 497 amino acids with a β-barrel structure, ubiquitously expressed in all tissues and mainly localized in lysosomes. The catalytic site of GCase contains two highly conserved residues of glutamic acid, which are necessary for the two-step mechanism of action, the nucleophilic attack, and the subsequent protonation. Loss of function mutations in the *GBA* gene cause Gaucher disease, the most common lysosomal storage disorder [7,8].

NLGase (EC 3.2.1.45) is encoded by the *GBA2* gene localized on chromosome 9p13.3. The catalytic site of the protein is characterized by both nucleophile and acid/base residues, Glu-528 and Asp-678 respectively. These residues define NLGase as a retaining β-glucosidase belonging to the CAZy glycosyl hydrolase family 116 (GH116) [9]. As previously described, the retaining β-glucosidases utilize a double-displacement mechanism [10]. In the first step of the reaction, called glycosylation, the nucleophile residue attacks the glucose anomeric center to create the glycosyl-enzyme intermediate, whereas the acid/base residue protonates the glycosydic oxygen, leading to the release of a glycone [11]. In the second step, the group working as an acid in the first step acts as a base catalyst that together with incoming water determines the de-glycosylation of the nucleophile. This mechanism allows the retention of the configuration at the anomeric carbon of the released glucose molecule [12].

The intracellular localization of NLGase has been controversial. It was reported to be a single pass transmembrane protein [13] and later was identified as a cytoplasmic membrane-associated protein of the ER and Golgi complex [14]. The activity of NLGase at the PM is directly modulated by the efflux of protons through the proton pumps associated with the cell surface [15,16].

The pathological involvement of NLGase was initially studied by generating *GBA2*-knockout mice [17]. Knockout mice present impairment in liver regeneration [18], and male infertility due to GlcCer accumulation. This accumulation causes dysregulation of lipid homeostasis due to a more ordered lipid composition of the PM [19,20]. As a result, the cytoskeletal dynamics are altered and the formation of sperm-head shaping and acrosome is affected [19].

Despite the abnormal GlcCer accumulation in brain, *GBA2*-knockout mice do not display any neurological symptoms or defects [19]. On the contrary, *GBA2*-knockdown zebrafish show abnormal motor neuron development [21], and mutations in the human *GBA2* gene have been found to lead to neurological disorders like spastic ataxia (SA) [22,23], hereditary spastic paraplegia (HSP) [21,24,25], and more recently Marinesco-Sjogren-Like Syndrome [26]. The molecular mechanism(s) leading to the development of disease are not currently known. Only one mutation of *GBA2* has been so far functionally characterized in vivo (i.e., zebrafish model) [21]. A more detailed biochemical analysis of the different mutants of *GBA2* in patient derived cells is still missing.

We have previously identified a *GBA2* missense mutation [c.1780G>C (p.Asp594His)] in a Cypriot family with progressive spastic ataxia [22]. In vitro characterization of this mutation in COS7 and HeLa cells showed that it causes a reduction at both the protein and enzyme activity levels [25]. In the present study we have undertaken the biochemical characterization of the *GBA2* c.1780G>C missense mutation in lymphoblastoid cell lines (LCLs) derived from spastic ataxia patients homozygous for the mutation. Our results contribute to the understanding of the biochemical consequences of mutations in the *GBA2* gene.

#### **2. Results**

#### *2.1. The c.1780G>C Mutation Results in NLGase Loss of Activity and GlcCer Accumulation*

*GBA2* mRNA expression levels were measured in LCLs obtained from four healthy individuals and from three patients homozygous for the c.1780G>C mutation. As shown in Figure 1-panel A, the same level of *GBA2* transcript was found in control and patient LCLs, suggesting that the mutation does not affect the *GBA2* mRNA expression and stability.

**Figure 1.** *GBA2* mRNA expression and non-lysosomal glucosylceramidase (NLGase) activity. (**A**) The graph represents the *GBA2* mRNA levels of controls (light grey) and patients (dark grey) relative to the average value of controls (*n* = 4) after normalization with the endogenous β-actin gene (*ACTB*). Values represent the mean ± SEM of two independent triplicate experiments. (**B**) NLGase activity associated with the total cell lysates and plasma membrane (PM) of controls (light grey) and patients (dark grey) derived lymphoblastoid cell lines (LCLs). Enzymatic activity was expressed as pmoles/106 cells/h. Data are expressed as mean <sup>±</sup> SD of three independent triplicate experiments (\*\*\* *p* < 0.0001 vs. controls).

Subsequently, we measured the NLGase enzymatic activity on the total cell lysate of the same LCLs. We observed that NLGase activity was almost undetectable in patients' cells with respect to controls (Figure 1B), which are characterized by an average specific activity of 78 ± 19 pmoles/106 cells/h.

In several cell lines of different origin, NLGase was found to be associated with the external leaflet of the PM where it catalyzes the in situ hydrolysis of glucosylceramide (GlcCer) to ceramide [27–29]. For this reason, we measured the enzymatic activity of NLGase directly at the cell surface of control and patient derived lymphoblastoid living cells. In patients' cells NLGase activity was also strongly reduced at the cell surface, showing 2–4% of residual activity with respect to that found in controls (Figure 1B; controls 44 ± 8 pmoles/106 cells/h, patients 2.3 ± 0.5 pmoles/106 cells/h). We can exclude that the enzymatic activity measured at the PM was due to other β-glucocerebrosidases because, by adding AMP-DNM, a specific inhibitor of NLGase to the assay solution abolished the enzymatic activity at the cell surface. Our data demonstrate that the presence of the homozygous *GBA2* c.1780G>C mutation results in an important loss of NLGase activity.

Furthermore, we assessed the effect of the NLGase loss of function on the GlcCer content. To this purpose, total lipid extracts obtained from the same LCLs used for the evaluation of the enzymatic activity were subjected to SFC-MS/MS analysis. This method is capable of quantifying and also distinguishing GlcCer from galactosylceramide (GalCer).

As shown in Table 1, patient derived LCLs are characterized by a two-fold increase in the GlcCer content with respect to controls. No differences were found among controls and cells expressing the mutated protein in the level of GalCer. Moreover, no difference was found in cellular cholesterol content between patients and controls, suggesting that GBA2 loss-of-function does not affect cholesterol homeostasis. In Gaucher disease, where we have GCase loss-of-function, the accumulated GlcCer is converted to glucosphingosine (GlcShp) by the action of the acid ceramidase [30]. We therefore also analyzed the levels of GlcShp by SFC-MS/MS and these were found to be hardly detectable, without any significant difference between control and patient cells.

**Table 1.** Hexosylceramides of lymphoblastoid cell lines (LCLs) from controls and patients. Glucosylceramide (GlcCer) and galactosylceramide (GalCer) contents were evaluated by SFC-MS/MS in controls (WT, *n* = 4) and patients (c.1780G>C, *n* = 3) LCLs. Data are expressed as pmoles/mg of cell protein ± Error (*n* = 3).


Quantitative analysis of the different molecular species of GlcCer showed that C16 is the most abundant species in both control and patient derived LCLs. In addition, it emerged that all the molecular species are doubled in pathological cells with respect to controls (Table 2).

**Table 2.** SFC-MS/MS analysis of glucosylceramide (GlcCer) molecular species in control (WT, *n* = 4) and patient (c.1780G>C, *n* = 3) derived LCLs. Data are the mean of three independent experiments and are expressed as pmoles/mg of cell proteins ± SEM.


In order to show that the increase of GlcCer was the result of an impairment of its catabolism, we labeled the cell sphingolipids at steady state using radioactive sphingosine [1-3H]Sph. This experimental procedure is based on the recycling of the radioactive precursor [1-3H]Sph that is used by the anabolic pathways as the endogenous counterpart. When there is an impairment of a catabolic enzyme (i.e., NLGase), the [1-3H]Sph is not further recycled and no degraded radioactive sphingolipid accumulates. As shown in Figure 2, despite the inter-individual variability in the content of lactosylceramide (LacCer) and globotriaosylceramide (Gb3), no significant differences were found between controls and patients for the other sphingolipids characteristic of LCLs. The only exception was the GlcCer and the ganglioside GM3 content. All patient derived LCLs showed an increase of more than two-fold of radioactive GlcCer with respect to control LCLs as well as a decrease of the radioactive GM3. The reduction in patient LCLs of the ganglioside GM3, which is a lipid typically associated with the external leaflet of the cell PM, with the concomitant increase in GlcCer, supports the hypothesis that the impairment of the catabolic pathway occurs at the cell PM level.

**Figure 2.** Radioactive sphingolipid pattern of control and patient LCLs. Total lipid extracts were separated by thin layer chromatography using the solvent system Chloroform/Methanol/Water 110:40:6 (*v*:*v*:*v*). (**A**) Representative digital autoradiogram obtained by the Beta-Imager TRacer equipment (BioSpace Lab). Same quantities of radioactivity were applied per lane. Ceramide (Cer), glucosylceramide (GlcCer), phosphatidylethanolamine (PE), lactosylceramide (LacCer), globotriaosylceramide (Gb3), sphingomyelin (SM), and ganglioside GM3. (**B**) Distribution of the radioactive sphingolipids associated with the total lipid extract expressed as % of the total radioactivity. Data are expressed as mean ± SD of three independent triplicate experiments (\*\*\* *p* < 0.0003 vs. controls).
