Purification and Uptake Studies of Recombinant Human N-α-D-Acetylglucosaminidase from Sf9 Insect Cells

(1)

from Sf9 Insect Cells by

Geoffrey Morris

B.Sc., University of Victoria, 2012

Grad. Cert. Learning and Teaching in Higher Education, University of Victoria, 2014 A Thesis Submitted in Partial Fulfillment

of the Requirements for the Degree of MASTER OF SCIENCE in the Department of Biology

(2)

ii

Supervisory Committee

Purification and Uptake Studies of Recombinant Human N-α-D-Acetylglucosaminidase from Sf9 Insect Cells

by Geoffrey Morris

B.Sc., University of Victoria, 2012

Grad. Cert. Learning and Teaching in Higher Education, University of Victoria, 2014

Supervisory Committee

Dr. Francis Choy, (Department of Biology) Supervisor

Dr. Robert Chow, (Department of Biology) Departmental Member

Dr. Raad Nashmi, (Department of Biology) Departmental Member

(3)

iii

Abstract

Supervisory Committee

Dr. Francis Choy, (Department of Biology) Supervisor

Dr. Robert Chow, (Department of Biology) Departmental Member

Dr. Raad Nashmi, (Department of Biology) Departmental Member

Human α-N-acetylglucosaminidase (Naglu) is a lysosomal enzyme implicated in the rare metabolic storage disorder Mucopolysaccharidosis III type B (MPS IIIB). A deficiency in Naglu results in a buildup of heparan sulfate in lysosomes, which is most detrimental in the central nervous system, causing mental retardation and a shortened lifespan. Enzyme replacement therapy is currently ineffective in treating the neurological symptoms of MPS IIIB due to the inability of Naglu to cross the blood-brain barrier. This laboratory uses a Spodoptera frugiperda insect cell system to express recombinant Naglu conjugated to a synthetic protein transduction domain with the intent to allow Naglu to cross the blood-brain barrier and treat the neurological symptoms.

In the present study, we aimed to purify a recombinant Naglu-PTD4 fusion protein in order to assess its capacity to cross cellular membranes. A three-step method involving multi-modal, hydrophobic interaction, and gel filtration

chromatography was optimized to achieve pure Naglu-PTD4, in good yield. Cellular uptake by human MPSIIIB fibroblasts of Naglu-PTD4 was not detectable. It is

hypothesized that additional amino acids, including a hexahistidine domain, following the PTD4 domain limited the fusion protein’s membrane transduction capacity. Future

studies will focus on removing the additional amino acids and adjusting the purification method as necessary. The ultimate goal of this research is to develop a large-scale recombinant Naglu production protocol for enzyme replacement therapy of MPS IIIB.

(4)

iv

List of Tables

Table 3.1. Yield of active Naglu through three-step analytical-scale purification ... 39 Table 3.2. Yield of active Naglu through three-step preparative-scale purification ... 44 Table 3.3. Fibrobloast uptake study of ΔS-Naglu-PTD4 with variable incubation times 49

(7)

vii

List of Figures

Figure 1.1. Stick-and-ribbon representation of the crystal structure of human Naglu ... 7

Figure 1.2 Schematic representation of Naglu construct ... 20

Figure 3.1. Protein elution profile of analytical-scale purification from A) MMC, B) HIC, and C) GFC columns ... 38

Figure 3.2. SDS-PAGE Gels visualizing results of analytical-scale purification by A) silver stain or B) Coomassie blue stain ... 40

Figure 3.3. Immunoblot of anti-Naglu from analytical scale purification ... 41

Figure 3.4. Protein elution profiles of preparative-scale purification from A) MMC, B) HIC, and C) GFC columns... 43

Figure 3.5. SDS-PAGE gel visualizing results of three-step preparative-scale purification by Coomassie blue stain ... 46

Figure 3.6. Western blot visualizing results of preparative-scale purification by probing against Naglu ... 47

Figure 3.7. Western blot visualizing purified ΔS-Naglu-PTD4 by probing against V5 epitope ... 48

Figure 3.8. Digestion of purified ΔS-Naglu-PTD4 using Carboxypeptidase A. ... 51

Figure 3.9. Digestion of purified ΔS-Naglu-PTD4 using Carboxypeptidase Y. ... 52

(8)

viii

List of Abbreviations

% percent °C degrees Celsius 4-mu-Naglu 4-methyl-umbelliferyl-N-acetyl-α-D-glucosaminide 6xHis hexahistidine

AAV adeno-associated virus AFU arbitrary fluorescence unit BBB blood-brain barrier

BCSFB blood-cerebrospinal fluid barrier BoCPA bovine carboxypeptidase A BSA bovine serum albumin

bp base pair

cDNA complimentary DNA CHO Chinese hamster ovary CNS central nervous system

CP carboxypeptidase

CPY carboxypeptidase Y

CV column volume

ΔSNaglu mutagenized Naglu

DMEM Dulbecco’s modified Eagle medium DNA deoxyribonucleic acid

(9)

ix EDTA ethylenediaminetetracaacetic acid

ER endoplasmic reticulum ERT enzyme replacement therapy FBS fetal bovine serum

FPLC fast protein liquid chromatography GAG glycosaminoglycan

GFC gel filtration chromatography GFP green fluorescent protein

HGMD Human Gene Mutation Database

HIC hydrophobic interaction chromatography HRP horseradish peroxidase

HS heparan sulfate

HSPG heparan sulfate proteoglycan IEX ion exchange chromatography

kb kilobase pair

kDa kilo Dalton kHz kilohertz

M molar

M6P mannose-6-phosphate

MCH064 normal skin fibroblast cell line

µg microgram

µL microliter

(10)

x mg milligram mM millimolar mL millilitre MMC multi-modal chromatography MPS Mucopolysaccharidosis

MPS III Mucopolysaccharidosis type III MPS IIIB Mucopolysaccharidosis type IIIB Naglu N-α-D-acetylglucosaminidase ND not detectable

NEAA non-essential amino acids

nm nanometre

nM nanomolar

nmol nanomole

NNSS native Naglu secretion signal PAGE polyacrylamide gel electrophoresis PBS phosphate buffered saline

PBST PBS with Tween 20

pM picomolar

PMSF phenylmethylsufonyl fluoride PTD protein transduction domain

PTD4 synthetically optimized protein transduction domain PVDF polyvinylidene fluoride

(11)

xi rpm revolutions per minute

SDS sodium dodecyl sulfate

Sf9 Spodoptera frugiperda 9 cells

TAT trans-activator of transcription UTR untranslated region

UV ultra-violet

v/v volume to volume ratio

V volts

WG0421 Sanfilippo Syndrome B skin fibroblast cell line w/v weight to volume ratio

(12)

xii

Acknowledgments

I am deeply thankful to my supervisor, Dr. Francis Choy, for his encouragement, support, and enthusiasm. Thank you to my labmates, including Kelly Turner, Rhea Ashmead, Anique Le Roux, Jen He, Kourtnee Hoitsema, Glynis Byrne, Olivia de Goede, and Alex Jack, for the time and effort they have spent discussing, teaching and also troubleshooting as I progressed through this project. Thank you to my committee members Dr. Nashmi and Dr. Chow, for their helpful insights and suggestions during committee meetings. I am grateful to Dr. Howard for allowing me the use of his FPLC machine.

(13)

xiii

Dedication

To my wife, Julia-Anne, for her strength, support, and encouragement. And for making endless meals and snacks.

(14)

Chapter 1 Introduction

1.1 Mucopolysaccharidosis III Type B

1.1.1 Overview

Mucopolysaccharidosis III, also known as Sanfilippo Syndrome, is a rare autosomal recessive metabolic disorder. There are four different confirmed subtypes of MPS III. The cumulative reported incidence of all subtypes of MPS III varies between populations, ranging from 0.28 to 4.1 per 100,000 lives births (Valstar et al. 2008). Each subtype is due to a deficiency in one of four lysosomal enzymes that participate in the degradation of heparan sulfate (HS): heparan N-sulfatase (type A),

N-α-D-acetylglucosaminidase (type B), heparanacetyl-CoA:α-glucosaminide N-acetyltransferase (type C), and N-acetylglucosamine-6-sulfatase (type D), respectively (Schmidtchen et al. 1998).

1.1.2 Clinical Manifestations and Prognosis

Mucopolysaccharidosis type IIIB (MPS IIIB, Sanfilippo Syndrome B) is an autosomal, recessive lysosomal storage disorder resulting from a deficiency of the acid hydrolase N-α-D-acetylglucosaminidase (Naglu, EC 3.2.1.50). If a cell does not express at least one functional copy of the NAGLU gene, a lack of active Naglu results in the build-up of partially degraded heparan sulfate within the lysosome that may eventually result in cell death. The partially degraded HS then accumulates within tissues and abnormally large amounts are excreted in urine.

Heparan sulfate is toxic to cells of the central nervous system (CNS), causing progressive cerebral atrophy (Weber et al. 1999). Patients with Sanfilippo Syndrome B are typically characterized by progressive, severe central CNS involvement and mild

(15)

2 somatic pathology. The neural degeneration results in aggressive behaviour,

developmental delays, hyperactivity, mental retardation, and a lifespan generally no greater than two decades (Hopwood 2007; Valstar et al. 2008).

The progression of Sanfilippo Syndrome B can be divided into three main stages. The concern of parents who observe an increase in behavioural disturbance, loss of language skills, or delayed development in their child usually begins diagnosis. At the time of clinical presentation, generally between two and four years of age, minimal somatic pathology is observed. In cases where somatic symptoms are present, they can include facial coarsening, hepatosplenomegaly, and skeletal changes (Schmidtchen et al. 1998; Barone et al. 2001).

The second stage of the disease is marked by noticeable, and therefore diagnostic, neurological degeneration. This stage occurs in most Sanfilippo Syndrome B patients by the age of six and is followed by a deterioration of behaviour and loss of learned skills. During the second stage, behaviour is characterized by frequent temper tantrums, hyperactivity, aggression, rapid loss in attention span, and severe sleep disturbance. Affected children are physically strong with good mobility, making the second stage of the illness the most difficult for caregivers to manage (Weber et al. 1999; Hopwood 2007).

General health and strength worsen in the final stage of the disorder. Most patients suffer from progressive dementia, loss of balance, increasing numbers of seizures, and feeding difficulties. Mobility is severely impaired by joint disease and increased spasticity. Death occurs in severely affected children by their mid to late teens,

(16)

3 commonly caused by a combination of respiratory infection and the debilitating nature of the disease (Cleary & Wraith 1993).

1.1.3 Sanfilippo Syndrome B at the DNA Level

The native human NAGLU gene was identified in 1994 (Zhao et al. 1994). Soon after, it was localized to chromosome 17q21.3 as a 8.5 kilobase (kb) segment containing six exons and five introns (Zhao et al. 1996). The sequence of human Naglu cDNA used for this project is 2229 base pairs in length, including 69 base pairs encoding the N-terminal native secretion-signaling peptide. The amino acid sequence of the secretion signal is MEAVAVAAAVGVLLLAGAGGAAG (Weber et al. 1996; Zhao et al. 1996). NAGLU is considered a housekeeping gene. As such, expression is expected at roughly uniform and ubiquitous levels in different tissues in the body. Northern blot analysis of multiple tissues shows Naglu mRNA at detectible levels in all tissues but thymus, with the highest levels seen in liver, ovary, and peripheral blood leukocytes (Weber et al. 1996).

At present, over 150 disease-causing mutations in the NAGLU gene are reported in the Human Gene Mutation Database (HGMD); most mutations are missense, but nonsense mutations, point mutations, insertions, deletions, and splice site mutations have also been reported. Allelic frequency of the different mutations is very low; most

mutations are unique to individual families. Nonetheless, some mutations are more common than others. For example R297X, a nonsense mutation, causes a severely affected phenotype and is found in 11.5% of patients worldwide and 23% of Dutch patients (Weber et al. 1999). In this mutation, as well as many others including R262X, R643H and R674H, the alteration occurs at CpG sites, areas known to be mutagenic hotspots (Zhao et al. 1996).

(17)

4 Though its incidence can vary widely, Sanfilippo Syndrome B is known to

manifest in all ethnic groups. The incidence was 0.36 per 100,000 per live births in Germany (Baehner et al. 2005), 0.42 in the Netherlands (Poorthuis et al. 1999), 0.28 in Taiwan (Lin et al. 2009), and 0.30 in Korea (Kim et al. 2013). MPSIIIB was also the most highly prevalent subtype reported for Portugal (Pinto et al. 2004), Taiwan (Lin et al. 2009), and Greece (Héron et al. 2010). Interestingly a high prevalence of mutations p.Y140C, p.H414R, and p.R626X in the NAGLU gene in Greek patients indicates possible founder effects (Beesley et al. 2005; Héron et al. 2010).

Sanfilippo Syndrome B has poor correlation between genotype and clinical phenotype (Weber et al. 1999). The allelic heterogeneity is considered a contributor to the wide spectrum of clinical phenotypes (van de Kamp et al. 1981; Weber et al. 1999; Yogalingam & Hopwood 2001; Valstar et al. 2008). Numerous polymorphisms that potentially modify disease severity complicate the prediction of genotype-phenotype relation in MPS IIIB (Yogalingam & Hopwood 2001). For example, the G79C mutation has been identified in a homozygous state in severely affected individuals (Bunge et al. 1999). Conversely, the F48L, G69S, S612G, and R643C mutations in NAGLU appear to confer a milder phenotype of the Sanfilippo B phenotype (Yogalingam & Hopwood 2001; Selmer et al. 2011). In compound heterozygous states with severe mutations, the S612G mutation appears to provide enough residual enzyme activity so as to mitigate disease progression and present as a more attenuated phenotype (Selmer et al. 2011).

1.1.4 Sanfilippo Syndrome B at the Biochemical Level

The Naglu enzyme was first associated with MPS-IIIB in 1972 by O'Brien. Since then, the enzyme has been purified and characterized from a variety of sources (Figura

(18)

5 1977a; b; Sasaki et al. 1991; Weber et al. 1996; Zhao & Neufeld 2000). Purified NAGLU has an isoelectric point of 5.1, pH optimum of 4.5 ± 0.3, and exists as tissue specific isozymes with sizes ranging from 73 kilo Daltons (kDa) to 86 kDa (Sasaki et al. 1991). The enzyme is inactivated easily by increases in heat and in pH (Figura 1977b).

As with all secreted or integral membrane proteins, the Naglu peptide undergoes co-translational insertion into the endoplasmic reticulum (ER). In the case of Naglu, the amino terminal signal sequence of 23 hydrophobic amino acids is eventually cleaved, resulting in a mature Naglu protein of 720 amino acids in length (Zhao et al. 1996). In the lumen of the ER, the Naglu enzyme is converted from its native linear form to the active functional form by appropriate folding into its three dimensional structure. This process is facilitated by chaperones, including those from the Hsp 70 family of proteins (Hebert & Molinari 2007).

The enzyme undergoes glycosylation catalyzed by oligosaccharial transferase, which is attached to the ER membrane. This occurs via N-glycosylation through addition of an acetyl-glycosyl group to the nitrogen of the asparagine residue in the consensus sequence asparagine-X-serine/threonine where X is any amino acid other than proline. Based on sequence analysis, six potential glycosylation sites are located at asparagine residues 261, 272, 435, 503, 526, and 532 (Weber et al. 2001). Upon completion of folding and glycosylation, the enzyme is transported out of the ER and into the Golgi for further modifications. In the cis-Golgi, Naglu is modified by the addition of mannose-6-phosphate (M6P) moieties at the terminal oligosaccharide residues (Lee et al. 2002). The resultant Naglu is released from the Golgi apparatus in vesicles that fuse with endosomes, which eventually develop to a mature lysosome (Luzio et al. 2003; Bagshaw et al. 2005).

(19)

6 The crystal structure of human Naglu was elucidated only recently (Meiyappan et

al. 2012). While an ortholog of human Naglu, CpGH89 from C. perfringens, was

crystallized several years earlier, it provided only limited insight: CpGH89 has about 30% overall amino acid sequence identity to Naglu and less than half of the CpGH89 amino acid residues were in the crystallized protein (Ficko-Blean et al. 2008). Crystalline Naglu has three domains (I, II, III). Amino acids 24-126 form a small α/β domain in Domain I. Containing the catalytic residues, amino acids 127-467 form a (α/β)8 domain in

Domain II. Amino acids 468-743 form an all α-helical domain in Domain III (Meiyappan

et al. 2012). The stick-and-ribbon representation of Naglu is shown in Figure 1.1.

Interestingly, while four mutations occur within the active site (Y140C, W268R, F410S, and W649C) there are no known mutations of the catalytic residues. It has been

hypothesized that such mutations may be so detrimental as to cause an embryonic lethal phenotype (Ficko-Blean et al. 2008).

(20)

7

Figure 1.1. Stick-and-ribbon representation of the crystal structure of human Naglu.

(21)

8 Heparan sulfate is the natural glycosaminoglycan (GAG) substrate for Naglu. It consists of a disaccharide repeat unit containing D-glucosamine linked to a hexauronic acid, and it is sulfated at various positions along its length. Heparan sulfate proteoglycans (HSPGs) are composed of a protein core to which heparan sulfate polysaccharides are attached. HSPGs are found ubiquitously in mammalian tissues, both extruding from cell surfaces and as a component of the extracellular matrix (Bernfield et al. 1999). Naglu is responsible for a step in the lysosomal catabolism of heparan sulfate. Specifically, Naglu hydrolyzes the α1-4 glycosidic linkage between N-acetylglucosamine and the

neighbouring uronic acid (Valstar et al. 2011). In the absence of functional Naglu, heparan sulfate builds up in the lysosomal vacuoles within cells. This storage leads to distortion and swelling of the cells (Li et al. 1999). Within the nervous system, the excess heparan sulfate may potentially interfere with neuroplasticity and neurogenesis (Li et al. 2002).

1.1.5 Neuropathological Effects

The mechanism of neuropathogenesis in MPS IIIB is not yet fully understood. Previously, a relatively straightforward method was thought to instigate the CNS disease: heparan sulfate accumulation within the lysosomes of neuronal cells damages several cellular functions and results in progressive CNS disease. Studies in both animal models and humans revealed that alterations in cellular homeostasis, resulting from incomplete breakdown of GAGs, are more complex and involve metabolism pathways not directly related to GAGs (Jolly & Walkley 1997; Clarke 2008). For example, accumulating GAGs may affect complex signalling pathways that occur between cells (Bishop et al. 2007),

(22)

9 while deficiency of one lysosomal hydrolase appears to lead to changes in the activity of other lysosomal hydrolases (Li et al. 1999).

Also of note is the accumulation of glycosphingolipids, which is surprising because their degradation does not require Naglu, nor any of the hydrolases involved in the MPSs (Constantopoulos et al. 1978; McGlynn et al. 2004). Recent studies in the MPS IIIB murine model have shown co-localized accumulation of free unesterified cholesterol as a potential response to the presence of excess ganglioside (McGlynn et al. 2004). As gangliosides and cholesterol are important in lipid rafts, which in turn are important in membrane signalling, their combined presence may represent possible new mechanisms to explain neuronal dysfunction in MPS IIIB (McGlynn et al. 2004).

Brain magnetic resonance imaging of patients and murine models with MPS has demonstrated mild to moderate cortical atrophy and ventricular enlargement (Barone et

al. 1999; Zafeiriou et al. 2001). Those neurons that remain in the cerebral cortex and the

anterior horn of the spinal cord have storage materials in the cytoplasm. Moreover, changes are seen in the cerebral white matter that suggest mucopolysaccharide-filled perivascular space. The molecular layer of the cerebellum shows prominent dendritic swelling of Purkinje cells in addition to gliosis in the cerebral white matter and thalamus (Tamagawa et al. 1985; Ferrer et al. 1988). Deterioration of the retinal pigmented epithelium and degeneration of photoreceptors have also been reported (Hamano et al. 2008).

1.1.6 Current and Proposed Therapies

At present, no disease modifying therapy exists for Sanfilippo Syndrome B; treatment focuses on supportive interventions. The characteristic aggression and

(23)

10 hyperactivity respond poorly to behavioural treatment strategies. Creating a ‘safe

environment’ within the home by way of soft furnishings, toughened glass, and wall padding can reduce damage to the home and child. In order to cope, many families find drug treatments, including antipsychotic agents and sedatives, aimed at controlling these symptoms are necessary. In the later stages of the disorder, when seizures become more common, anticonvulsant medication is also administered with relative success (Cleary & Wraith 1993). Treatment of sleep disturbances has been successful by administering melatonin (Guerrero et al. 2006).

Allogeneic bone transplantation has been used to benefit some patients. In one study, donor stem cells repopulated various tissues and delivered enzyme to correct storage in host cells; however, interpreting the effect of the therapy was difficult as neurological symptoms were present (Hoogerbrugge et al. 1995). A similar study on twins suffering from MPS IIIB found some improvement in cognitive function. Both twins still exhibited an overall decline in intellectual development and were considered to have significant developmental delays (Vellodi et al. 1992). Therefore, bone marrow transplantations are generally only effective for patients that have no evidence of neurological symptoms.

Though not yet approved for treatment in humans, gene therapy has received immense interest in recent years. A study using direct injection of adeno-associated virus (AAV) vectors encoding Naglu into 6-week old MPS IIIB mice showed widespread correction of biochemical and histological parameters in the brain at 38 weeks of age (Cressent 2004). Similarly, a lentiviral-Naglu vector injected intravenously into an 8-10 week old MPS IIIB mouse was able to maintain Naglu activity and reduced levels on

(24)

11 GAGs in liver lung and spleen (DiNatale 2005). In an effort to correct both the cerebral and systemic aspects of the disease, Heldermon et al. (2013) used a combination of both intracranial AAV-Naglu and intravenous lentiviral- Naglu injections. Compared to untreated MPS IIIB mice, the combination therapy resulted in significant improvements in motor function, hearing, and circadian rhythm. The combination therapy also nearly doubled the median life span: 322 days to 612 days (Heldermon et al. 2013).

Since the discovery that coculturing fibroblasts with distinct MPS subtypes resolved GAG storage by cross-correction of enzyme deficiency (Fratantoni et al. 1968), the potential of enzyme replacement therapy (ERT) as a cure for lysosomal storage disorders has been investigated. ERT was first shown to be successful in the treatment of Type 1 Gaucher Disease (Andersson et al. 2008), and has since been successful for other lysosomal storage disorders including Fabry disease, Pompe disease (Hopwood 2007), and for the treatment of somatic symptoms in MPS I, MPS II, and MPS VI (Wraith et al. 2004; Harmatz et al. 2006; Muenzer et al. 2006). However, enzymes injected

intravenously, including Naglu, do not appear able to cross the blood-brain barrier (BBB) on their own.

1.2 Treatment of Sanfilippo Syndrome B neurological symptoms

1.2.1 The Blood-Brain Barrier

The CNS is the control center for the body: processing sensory input, generating programs and strategies, controlling motor output, and coordinating many of the activities performed by the body’s tissues. To function properly, a highly regulated extracellular environment must be maintained, where the concentration of ions such as Na+, K+, and Ca2+_{remain within very narrow ranges. Additionally, the high level of metabolic activity}

(25)

12 within the CNS demands a considerable 20% of the oxygen consumed in humans (Rolfe & Brown 1997). The CNS is also very sensitive to a broad range of chemicals, including much of what is consumed within our diet. Many of the substances are readily

metabolized and excreted; if left to travel throughout the brain, however, significant neurological damage could result (Abbott & Friedman 2012).

It is therefore essential that the interface between the CNS and the surrounding circulatory system be selective in what can and cannot enter. The interface must maintain an appropriate ionic balance, allowing the entrance of nutrients and exit of waste

products, all while acting as a barrier to potentially harmful molecules and organisms. There are three main sites of CNS interface barriers: the endothelium of the brain microvessels (the BBB), and the epithelium of the choroid plexus and of the arachnoid matter (together the blood-cerebrospinal fluid barrier, BCSFB) (Abbott et al. 2010). At each of these sites, tight junctions between adjacent cells restrict the diffusion of polar solutes and transporter proteins within the membranes regulate any exchange. In all, the three interfaces act as physical, transport, enzymatic, and immunologic barriers. The barrier functions are able to respond to signals from both the blood and the brain side, and they can be significantly disturbed during pathology (Hawkins & Davis 2005).

Of particular interest is the BBB, as it is by far the largest interface for blood-brain exchange. For the average human adult, there is a total area of exchange in the blood-brain between 12 and 18 m2 (Strazielle & Ghersi-Egea 2013). The BBB is also considered to be the bottleneck in brain drug development and the limiting factor for future growth in neurotherapeutics (Pardridge 2010). Drugs that are intended to act in the CNS can be administered systemically if they have the ability to cross the BBB. Otherwise they must

(26)

13 be directly injected into the CNS via invasive methods. The relative impermeability of the BBB means only small gaseous molecules or those with appropriate hydrophobicity and size will diffuse from blood into the CNS.

However, the overwhelming majority (98-100%) of small, hydrophilic, and large molecules do not cross the BBB (Gabathuler 2010). Instead, these molecules require interaction with specific transporters and/or receptors present on the luminal side of the endothelial cells. Required substances such as glucose, insulin, and growth hormones are recognized by specific receptors/transporters and selectively transported into the brain. Delivery of pharmaceutical and therapeutic agents, therefore, is an ongoing challenge.

1.2.2 Potential Strategies for Neurological Drug Delivery

One approach to overcome the BBB roadblock is the administration of drugs directly into the brain. Direct delivery methods generally employ stereotactic procedures to locate target sites inside the brain in reference to anatomical landmarks, thus bypassing the BBB by injecting the treatment directly into the brain parenchyma (Deng 2010). The three most common approaches are: intracerebral implantation, intracerebroventricular infusion, and convection enhanced diffusion (Pardridge 2005).

Savas et al. (2004) explored the direct method of drug delivery for the treatment of MPS IIIA, an autosomal recessive disorder caused by a deficiency in the enzyme heparan N-sulfatase (sulfamidase). Using an MPS IIIA mouse model that was shown to have a disease progression similar to that of human patients, the authors delivered intracerebral injections of recombinant human sulfamidase directly into the murine dentate gyrus. This approach was found to delay neurodegenerative changes in young MPS IIIA mice (Savas et al. 2004). Further studies demonstrated that repeated doses of

(27)

14 sulfamidase (Hemsley et al. 2007), or removal of mannose-6-phosphate residues from sulfamidase (Rozaklis et al. 2011) improved the effectiveness of treatment.

The main advantage of the direct injection technique is the ability to bypass the BBB, improving delivery and requiring less molecular and chemical modification of neurotherapeutics. However, disadvantages include limited therapeutic efficacy due to inadequate drug diffusion from the injection sites, potential damage to untargeted brain tissue from needle penetration, and risk of complications such as infection from invasive procedures (Deng 2010).

Ultrasound, defined as waves at a frequency at or above 20 kHz, can be focused, reflected, and refracted through different mediums (Kodama et al. 2006). Due to the unique characteristics of ultrasound, it is not surprising that this technique has been explored as a potential method of inducing BBB disruption for drug delivery into the brain. Two distinct ultrasound-based approaches have received the majority of attention. The first method utilizes the thermal effects of focused ultrasound; the second method employs mechanical effects and acoustic cavitation. Regarding the thermal approach, studies have demonstrated that the BBB was disrupted near brain tissue that had received high intensity focused ultrasound. The second approach utilizes the mechanical effects of acoustic cavitation initiated by preformed microbubbles in the brain vasculature. Effects such as shear stress due to bubble oscillation and shock waves from bubbles collapsing have been identified as important factors involved in increasing the permeability of cellular membranes and endothelial barriers (Rosenthal et al. 2004).

Through a combination of stable and inertial cavitation, many studies have found biological effects both in vitro and in vivo (Miller et al. 1996; Deng et al. 2004). For

(28)

15 example, Konishita et al. (2006) delivered a Herceptin monoclonal antibody to a targeted region of the mouse brain. Herceptin was the first antibody drug to be successfully used in the treatment of metastatic breast cancer, however, it had been unsuccessful in patients with brain tumours due to its inability to cross the BBB (Slamon et al. 2001). Using MRI-guided focused ultrasound BBB disruption, Herceptin was able to cross the barrier and reach the target tissue with minimal histological changes (Kinoshita et al. 2006).

Regarding the thermal approach, a window for BBB disruption without irreversible tissue damage has yet to be determined. Therefore, this method has

considerable disadvantages compared to the mechanical approach. Conversely, the use of microbubbles offers a number of potential advantages: reduced ultrasound intensity resulting in less tissue damage, containment of disruptive cavitation events within the microvasculature, and less likelihood of irreversible neuronal damage (Deng 2010). However, the detailed mechanism for this second approach is yet to be elucidated. Therefore, further experiments are needed to determine its potential clinical success.

In the past several years, there have been a growing number of publications on nanotechnology and its use in drug delivery. Nanotechnologies are devices and materials that have a functional organization in at least one dimension on the nanometer scale (Silva 2008). It has been demonstrated that nanoparticles can be used to maintain drug levels in a therapeutically desirable range, and increase half-lives, solubility, stability and permeability of drugs. Nanoparticles can be widely adapted to optimize drug delivery and reduce side effects by targeted delivery (Yang 2010).

Afergan et al. (2008) described a novel use of liposome as a secondary delivery vehicle: monocytes phagocytised serotonin-encapsulated liposomes and permeated across

(29)

16 the BBB. As a result, the phagocytised serotonin was selectively delivered to the brain. In another example, the brain-targeting ligand lactoferrin was coupled to a dendrimer for directed delivery of DNAs. The lactoferrin-coupled dendrimer was reported to have high BBB-crossing efficiency compared to unmodified counterparts (Huang et al. 2009).

1.2.3 Protein Transduction Domains

With the discovery of cell penetrating protein transduction domains (PTDs), new opportunities to develop proteins fused to PTDs have surfaced. Instead of injecting a virus that contains a gene of interest, the cell-penetrating portion, derived from the virus, is attached to a protein directly, avoiding the issue of viral integration. PTDs are small peptides capable of transversing biological membranes, and bringing their fusion partners along with them. Several studies have shown the potential of PTDs in drug delivery (Jin

et al. 2001), including the transduction of proteins as large as 110 kDa (Dietz & Bähr

2004).

The most intensely studied, yet less understood peptide in protein transduction is the PTD of the HIV Tat. The functional 11 amino acid portion of trans-activator of transcription (Tat) PTD (YGRKKKRRQRRR) has the ability to transport fusion partners, including large proteins across the BBB (Schwarze et al. 1999). In the study by Schwarze

et al. (1999) recombinant protein containing the enzyme β-galactosidase fused to

Tat-PTD was found to enter the mouse brain in vivo following intraperitoneal injection without any notable toxicity or immunogenicity from Tat-PTD. The wildtype Tat-PTD is an amphipathic alpha-helix. An improved variant of Tat-PTD, named PTD4, was

developed by Ho et al. (2001). PTD4 contains optimized alpha helical structure and placement of basic arginine residues (amino acid sequence YARAAARQARA). The

(30)

17 result is a 33 times greater transduction in vitro and five times greater transduction in vivo over wildtype Tat-PTD (Ho et al. 2001).

Since the Schwarze et al. study (1999), a large number and variety of molecules have been shown to cross the BBB and positively effect the target site when fused to HIV Tat (Zhang et al. 2012). For example, Neuroglobin, an agent to protect against brain hypoxic-ischemic injury, was fused to Tat PTD (Cai et al. 2011). Normally unable to cross the BBB, mice treated with the Neuroglobin-PTD fusion protein showed

significantly less brain infarct volume and had better neurologic outcomes following middle cerebral artery occlusion when compared to a control group. Similar success has been shown with Bcl-xL (Doeppner et al. 2010) and heat shock protein fusion products (Doeppner et al. 2009).

Once again, this type of treatment has enormous potential for CNS-based disorders and diseases. Owing to its high cationic charge, and the anionic properties of the cell surface, electrostatic interactions are considered to be the most significant effect that associates Tat PTD to the cell surface. It has been shown that PTD-fusion partners are internalized by endocytosis (Richard et al. 2005; Tuennemann et al. 2006) by simultaneously using 3 endocytic pathways: macropinocytosis, clathrin-mediated endocytosis, and caveolae/lipid-raft-mediated endocytosis (Duchardt et al. 2007). More recently, it was postulated that the cell penetrating ability of Tat is a function of the number and spatial array of its guanidinium groups in the PTD of Tat (Stanzl et al. 2013).

(31)

18

1.3 Expression of Human Recombinant Naglu

1.3.1 Removal of Cryptic Splice Site

Our laboratory had consistently experienced markedly low expression levels of Naglu in the Sf9 insect cell system. Using RNA isolated from Sf9 cells expressing recombinant human Naglu cDNA construct, reverse transcription polymerase chain reaction (RT-PCR) analysis indicated the presence of a non-target band about 400 bp in length. Sequencing data on this fragment placed it at 426 bp in length and containing sequence from both the 5’ and 3’ ends of Naglu cDNA, with a large central portion (1735 bp) missing (Truelson et al. 2011). Most interesting was that the internally deleted

segment had a GU dinucleotide and an AG dinucleotide at its expected flanking positions. These are the most common splice donor and acceptor signals, respectively (Shapiro & Senapathy 1987), and lead to the hypothesis of cryptic splicing. A “cryptic” splice site is a sequence present in precursor mRNA that very closely resembles the consensus sequence for a splice site. Normally, cryptic splice sites are not recognized by the splicing machinery and therefore not acted upon during mRNA processing (Jantzen et

al. 2013). However, when authentic splice sites are not present, as in cDNA, cryptic

splice signals can activate.

Site directed mutagenesis was performed to modify the splice signals without altering the amino acid sequence. Fusion PCR through combination of three mutagenized Naglu fragments was able to create the full length 2.2 kb mutagenized Naglu cDNA (ΔSNaglu). RT-PCR analysis of Sf9 cells expressing the ΔSNaglu cDNA showed the presence of the expected 2.2 Kb band and the complete absence of the 426 bp putative splice band (Truelson et al. 2011). In order to test the yield of the ΔSNaglu, stably

(32)

19 selected Sf9 cell lines were generated using Naglu, Naglu-PTD4, and ΔSNaglu-PTD4. Data from suspension cultures showed the mutagenized ΔSNaglu-PTD4 had a 3.99 fold higher level of expression compared to both the Naglu and Naglu-PTD4 cultures (Jantzen

et al. 2013). This was a significant improvement and should assist in producing sufficient

enzyme for purification steps.

1.3.2 Expression System

Recombinant Naglu used for this project was expressed in an Sf9 cell system. This cell line is derived from the moth Spodoptera frugiperda which is widely used for the production of human glycoproteins (Altmann et al. 1999). The expression system was selected based on the capacity of Sf9 cells to perform post-translational modifications similar to those in mammalian cells (Pfeifer et al. 2001) and its ability to secrete large amounts of recombinant protein (Sinclair et al. 2006). Notably, however, insect cells produce N-glycans of a lower average molecular weight and are unable to phosphorylate mannose residues (Kost et al. 2005). Insect cells have several advantages over

mammalian cells, including their relative economy, and ability to be adapted for large-scale fermenters (Pfeifer 1998). In the past, insect cell lines were commonly used along side a baculovirus expression system. While this system was effective in creating transfected cell lines, it had several major drawbacks including sub-optimal

post-translational protein modification and eventual lysis of the host cells (Pfeifer et al. 2001). Following the creation of ΔSNaglu-PTD4 in a p2ZoptcxF plasmid (Truelson et al. 2011), the construct was transferred into a pIZT/V5-His plasmid vector (Jantzen 2011). Unlike the baculovirus system that requires reinfection of new cell lines due to eventual lysis and death of infected cells, plasmid-mediated transfection allows for the creation of

(33)

20 stable recombinant cell lines. The pIZT/V5-His vector contains a Zeocin™-resistance gene (Sh ble) for antibiotic selection, as well as a cycle 3 green fluorescent protein (GFP) gene for rapid detection of transfected cells. Within the multiple cloning site is a V5 epitope, to allow detection of recombinant protein, and a hexahistidine (6xHis) tag, to facilitate purification (Life Technologies 2012). The Naglu construct is represented in Figure 1.2. For the full pIZT/V5-His polylinker map, see Supplementary Figure 1.

Figure 1.2 Schematic representation of Naglu construct.

NNSS: N-terminal native secretion signal. ∆SNaglu: Naglu cDNA with cryptic splice sites removed. PTD4: Optimized protein transduction domain, derived from HIV Tat PTD. V5: V5 epitope binding domain. 6x His: Hexahistidine purification tag.

1.4 Purification of Naglu

1.4.1 Fast-Protein Liquid Chromatography

Fast Protein Liquid Chromatography (FPLC) is a technique developed for the purification of proteins from a complex mixture (Gonzalez-Llano et al. 1990; Choy & Woo 1991). FPLC can separate proteins based on size, charge, hydrophobicity,

isoelectric point, or specific ligand recognition (Madadlou et al. 2011). GE Healthcare’s ÄKTA FPLC is a recent system that allows fully automated liquid chromatography designed for research scale protein purification. This system has several benefits over its predecessors, including better protein purification conditions and reduced cost (Madadlou

et al. 2011). FPLC is optimal for the rapid purification of active enzymes as protocols can

(34)

21 the enzyme of interest to maintain its native conformation and minimize the amount of inactivation (Sheehan & O'Sullivan 2004).

Previous attempts by our lab to purify Naglu have been unsuccessful. A single step purification method was developed using a Naglu-cellulose binding domain fusion enzyme and microcrystalline cellulose as a purification medium. While this method resulted in Naglu purified to near homogeneity, final yields of active enzyme were well below 10% (Bandsmer 2004) Similarly, purification using a Naglu-polyhistidine tag fusion protein and nickel affinity chromatography did not result in acceptable yield values. In both cases, the conditions required to elute Naglu from the column were too stringent and laborious. Purification of Naglu using a combination of more traditional chromatography methods, including ion-exchange, hydrophobic interaction, and/or gel-filtration, has been more successful (Zhao & Neufeld 2000; Weber et al. 2001;

Meiyappan et al. 2012).

1.4.2 Ion-Exchange

Possibly the most popular method for the purification of proteins and other charged molecules, ion-exchange chromatography (IEX) relies on charge-charge

interactions between proteins and an immobilized resin. IEX can be subdivided into three categories: cation IEX involves positively charged sample binding to negatively charged resin; anion IEX, conversely, involves negatively charged sample binding to positively charged resin; multi-modal chromatography (MMC) combines hydrophobic and/or other types of interactions with IEX. Buffers of low ionic strength are used to equilibrate the column and promote binding of proteins. With increasing ionic strength or change in pH, proteins begin to elute from the column. Increasing salt concentrations allow salt ions in

(35)

22 the buffer to compete for binding with the charges on the resin, and the bound material is displaced. With changes in pH, bound proteins are titrated and eventually become

noncharged or of identical charge to the resin, leading to repulsion and elution (Jungbauer & Hahn 2009).

The popularity of IEX is based on its capacity to concentrate and simultaneously purify. MMC systems, in particular, have received increased utilization for a number of applications including small molecule (Hou & Cramer 2011), protein (Burton et al. 1997), and antibodies (Chen et al. 2010). MMC also has the capacity to bind proteins at higher salt levels than conventional IEX, meaning cell culture samples do not require dialysis prior to loading.

1.4.3 Hydrophobic Interaction

Hydrophobic interaction chromatography (HIC) exploits the difference in relative hydrophobicity of proteins to purify them from one another. Just as nonpolar molecules aggregate in polar solvents (such as water) to increase the overall entropy of the system, the hydrophobic regions present in macromolecules bind to the hydrophobic ligands present in the column (McCue 2009). Generally, a high salt buffer is used to induce binding by the greatest number of proteins, and the salt level is gradually dropped to pure water so that proteins with the lowest relative hydrophobicity are released first. Organic solvents can also be added in order to disrupt hydrophobic interactions and elute the strongly bound proteins.

The conditions used for HIC are relatively gentle, allowing for the biological activity of the sample to be maintained (O'Farrell 2004). Several studies have had good success in using HIC to help purify lysosomal enzymes (Chen et al. 2000; Sinclair et al.

(36)

23 2006; Meiyappan et al. 2012). As Naglu is a membrane-associated protein, its strong hydrophobicity characteristics are expected to lend well to HIC as a form of purification. Indeed, Jantzen (2011) was able to increase the purity of Naglu from cell culture media while maintaining approximately 60% of original activity levels.

1.4.4 Size Exclusion

Just as IEX and HIC exploit a particular characteristic of proteins to separate them from a mixture, size exclusion, or gel filtration chromatography (GFC), uses porous beads to separate proteins on the basis of size (Stellwagen 2009). In contrast with conventional forms of purification, the gel filtration column retains none of the proteins. Instead, molecules with small hydrodynamic diameter have their progress through the column slowed, as they are forced through the porous beads, while molecules too large for the pores pass over them. Thus, the order of proteins eluted from a GFC column is the inverse of their hydrodynamic diameter (Rhodes & Laue 2009) and, assuming all the proteins have similar shape, by their molecular weight (Stellwagen 2009).

The non-retaining feature of GFC makes it ideal for the purification of delicate proteins as no damage occurs through the binding and releasing from the

chromatographic support (Hagel 2011). This is also a limitation of GFC, however.

Stellwagen (2009) hypothesized that a maximum of 10 proteins can be resolved from one another by any GFC column due to the lack of column binding and non-ideal flow around the beads. Therefore, GFC is best used later in a purification procedure when few

proteins are left in the sample, and the protein of interest is of significantly different size from the others.

(37)

24

1.5 Project Overview

Now that strong expression of ΔSNaglu-PTD4 has been achieved (Jantzen et al. 2013), the next steps are to purify ΔSNaglu-PTD4 and conduct uptake studies. The Sf9 cell line expressing the recombinant fusion protein ΔSNaglu-PTD4 was constructed previously in our lab, and preliminary small-scale purification trials were conducted. I will use a range of analytical column chromatography methods, including ion exchange, hydrophobic interaction, and gel-filtration, to optimize the purification of Naglu-PTD4. The most appropriate purification protocol will be increased to a semi-preparative scale in order to generate sufficient quantities of Naglu-PTD4 for cellular uptake studies, using disease-state fibroblast cultures. The ultimate goal of this research is to develop an efficient system for economical, large-scale production of human recombinant Naglu that has the potential to be successfully used for enzyme replacement therapy of MPS IIIB.

(38)

25

Chapter 2 Materials and Methods

2.1 Materials

2.1.1 Chemicals and Reagents

Abcam, Toronto, ON: polyclonal rabbit anti-Naglu antibody (ab137685). ACP Chemicals, Montreal, QC: sodium acetate. BioRad, Hercules, CA: BioRad Protein Assay dye

reagent, Precision Plus Protein™ Dual Color Standards. Commercial Alcohols Inc.,

Brampton, ON: 95% ethanol. EMD Millipore, Billerca, MA:

4-methylumbelliferyl-N-acetyl-α-D-glucosaminide. Invitrogen, Burlington, ON: Monoclonal mouse anti-V5-HRP-conjugated antibody, pIZT-V5/His vector system. Life Technologies, Burlington, ON: High Glucose Dulbecco’s Modified Eagle Medium (DMEM), Fetal Bovine Serum (FBS), Minimum Essential Media Non-Essential Amino Acids (MEM NEAA) (100X),

Ethylenediaminetetracaacetic acid (EDTA). New England Biolabs, Beverly, MA: 100x bovine serum albumin (BSA), 3x sodium dodecyl sulphate (SDS) sample buffer, 1.25 M dithiothreitol (DTT). Pierce Biotechnology, Rockford, IL: Carboxypeptidase Y,

SuperSignal® West Dura chemiluminescent reagent. Sigma-Aldrich, Oakville, ON: Carboxypeptidase A, phenylmethylsulfonyl fluoride (PMSF), sodium phosphate dibasic (Na₂HPO₄-7H₂O), potassium phosphate monobasic (KH₂PO₄), sodium chloride (NaCl), potassium chloride (KCl). Sodium hydroxide (NaOH), Tween® 20. Thermo Fisher

Scientific, Waltham, MA: goat anti-rabbit horseradish peroxidase (HRP)-conjugated

(39)

26 2.1.2 Prepared Media and Solutions

Buffers: CPA digestion buffer: 25 mM Tris-HCl pH 7.5, 100 mM NaCl. CPY digestion

buffer: 100 mM sodium acetate pH 5.5. Glycine-NaOH buffer: 0.5 M glycine-NaOH

buffer, pH 10.5. Methanol transfer buffer: 10% (v/v) methanol, 25 mM tris base, 19.2 mM glycine. PBS: 137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4, 1.44 mM KH2PO4, pH

7.4. PBST: PBS, 0.2% (v/v) Tween® 20. Purification buffer A: 50 mM sodium phosphate, pH 5.8. Purification buffer B: 1M NaCl, 50 mM sodium phosphate, pH 7.0. Purification

buffer C: 2 M NaCl, 50 mM sodium phosphate, pH 5.8. Purification buffer D: 150 mM

NaCl, 50 mM sodium phosphate, pH 5.8. Purification buffer E: 500 mM NaCl, 50 mM sodium phosphate, pH 6.2. SDS Buffer Mix: 90% (v/v) 3x SDS sample buffer, 10% (v/v) 30x DTT. Tris-Glycine electrophoresis buffer: 0.1% (v/v) SDS, 200 mM glycine, 25 mM Tris base.

Culture Media: Fibroblast growth medium: (DMEM High Glucose, NEAA 1% (v/v), FBS 10% (v/v).

Other Prepared Solutions: 4-mu-Naglu substrate: 0.1M sodium acetate, 0.5 mg/ml BSA, 0.2 mM 4-methylumbelliferyl-N-acetyl-α-D-glucosamide, pH 4.3. Coomassie fixative

solution: 50% (v/v) ethanol, 2% (v/v) phosphoric acid. Coomassie staining solution:

(10% (v/v) phosphoric acid, 10% (w/v) ammonium sulphate, 20% (v/v) methanol, 0.12% (w/v) Coomassie blue G-250. Silver stain fixative solution: 50% (v/v) ethanol, 5% (v/v) acetic acid. Silver stain developer solution: 20 mg/mL sodium carbonate, 0.04% (v/v) formaldehyde. Silver stain pre-drying solution: 10% (v/v) ethanol, 5% (v/v) glycerol.

(40)

27 2.1.3 Equipment and Software

Barnstead International, Dubuque, IA: Sequoia-Turner digital fluorometer (model 450). BioRad, Hercules, CA: Immunoblot® polyvinylidene fluoride (PVDF), Mini

PROTEAN® Tetra Cell electrophoresis system, Mini Trans-Blot® Cell electrophoretic transfer system. Carl Zeiss Microimaging, Jena, Germany: AxioVert 100 inverted microscope. Eastman Kodak, Rochester, NY: Kodak® X-OMAT 2000A processor. Processor. EMD Millipore, Billerca, MA: Amicon® Ultra-4 30K centrifugal filter device, Amicon®_{Ultra-15 30K centrifugal filter device, Stericup}®_{-HV 0.45 µm Durapore PVDF}

Filter Units. Forma Scientific, Marietta, ON: Water-Jacketed Incubator Model 3332. GE

Healthcare Life Sciences, Mississauga, ON: Äktaprime™ chromatographic system,

Capto™ MMC media, HiLoad™ 16/600 Superdex™ 200 prep grade prepacked column, HiScreen™ Capto™ MMC prepacked column, HiScreen™ Butyl-S FF prepacked column, HiTrap™ Butyl-S FF prepacked column, PrimeView 5.0 software, XK 26/50 empty column. New Brunswick Scientific, Edison, NJ: C24 Incubator Shaker. Pall

Corporation, Mississauga, ON: 0.2 µm Supor®-200 filtration membranes. PerkinElmer,

Waltham, MA: Wallac 1420 Victor2™ Microplate Reader. Pierce Biotechnology,

Rockford, IL: CL-X Posure™ film Sarstedt, Newton, NC: PE Vented Cap T25, T75, and

T175 cell culture flasks, SUREGrip 60 mm tissue culture dish for adherent cells, rubber tipped 25 cm cell scrapers. Thermo Fisher Scientific, Waltham, MA: 250 mL and 1 L: Fisherbrand® shaker flasks.

(41)

28 2.1.4 Cell Lines

Invitrogen, Burlington, ON: Gibco™ Spodoptera frugiperda 9 cells. McGill University Health Centre Cell Repository, Montreal, QC: MCH064 normal skin fibroblasts,

WG0421 Sanfilippo Syndrome B skin fibroblasts.

2.2 Methods

2.2.1 Sf9 cell growth and expression

Sf9 stocks, expressing NNSS-ΔSNaglu-PTD4, stored at -80°C were thawed and cultured in T25 flasks with 5 mL SFX media at 26°C in a non-humidified water-jacketed incubator. SFX was changed every 3-4 days until cultures reached 90-100% confluency. To generate large-scale suspension cultures, Sf9 adherent cultures were scaled up in a stepwise manner until a T175 flask was covered to confluency; at this stage, the cells were resuspended and made up to 100 mL in a 250 mL unbaffled vented shaker flask. Cultures were continuously shaken at 120 rpm, 26-28°C. Once a density of 2 x 106 cells/mL was reached, 75 mL of the small-scale shaker culture was used to seed a 1 L unbaffled shaker flask to a total volume of 400 mL. The cells were allowed to recover for three days before increasing the total culture volume to 500 mL. Harvesting and replacing approximately 450 mL every 7-9 days, depending on the level of ΔS-Naglu-PTD4

activity indicated by the fluorogenic assay, maintained cultures. Harvested medium was clarified by centrifugation at 5000xg and stored at -20°C. Crude medium was thawed prior to protein purification.

(42)

29 2.2.2 Preparative-Scale ΔS-Naglu-PTD4 Purification

The preparative-scale protocol was derived from an analytical-scale protocol. For ease of understanding, the preparative-scale is described in full. The analytical-scale differences are described following the full preparative-scale description.

All steps were completed on an ÄktaPrime FPLC system (GE Healthcare Life Sciences). A continuous UV readout of each step was saved using PrimeView software. All solutions were filter-degassed through a 0.45 µm Durapore PVDF membrane (EMD Millipore) prior to loading onto columns.

MMC purification

An XK 26/20 column (GE Healthcare Life Sciences) was custom packed to a final volume of 50 mL with Capto™ MMC media (GE Healthcare Life Sciences). All steps were carried out at a flow rate of 7 mL/min. For each purification run, the column was first rinsed with 5 column volumes (CV) of distilled water, then equilibrated with 5 CV of Buffer A.

Crude, cell clarified, Sf9 media was removed from the -20°C freezer, thawed completely, and centrifuged at 6000xg for 30 minutes. The supernatant was adjusted to a pH of 5.8. Prepared media was loaded onto the equilibrated column.

Following loading, the column was washed with Buffer A until the UV trace returned to baseline. ΔS-Naglu-PTD4 elution was done using an increase in both pH and NaCl concentration through a 2 CV linear gradient from Buffer A to Buffer B followed by 5 CV of Buffer B. Fractions of 15mL were collected and stored at 4°C.

All fractions were analyzed for activity with an artificial substrate 4-methylumbelliferyl-N-acetyl-α-D-glucosaminide (4-mu-Naglu, EMD Millipore). Fractions that contained high levels of activity were pooled and adjusted to 2M NaCl,

(43)

30 50mM sodium phosphate, and pH 5.8. A small portion of the pooled and adjusted sample was saved for future analysis. The column was preserved with 4 CV of distilled water followed by 4 CV of 20% ethanol.

HIC purification

Two pre-packed HiScreen Butyl-S columns (GE Healthcare Life Sciences) were connected in series to produce a single 20cm, 9.4mL column. All steps were carried out at a flow rate of 1 mL/min. The column was equilibrated with 2 CV of Buffer A, followed by 4 CV of Buffer C. The MMC-purified sample was loaded onto the column.

The column was washed with Buffer C, until the UV trace returned to baseline. A 5 CV desalting gradient from Buffer C to Buffer A was run to elute ΔS-Naglu-PTD4 from the column. Fractions of 2.5 mL were collected and stored at 4°C.

All fractions were analyzed for activity with 4-mu-Naglu substrate. Fractions that contained high levels of activity were pooled, dialyzed with Buffer D and concentrated to approximately 1 mL using Amicon Ultra-15 centrifugal filter devices (EMD Millipore). A small portion of the final sample was saved for future analysis. The column was preserved with 4 CV of distilled water and 4 CV of 20% ethanol.

GFC purification

A pre-packed HiLoad Superdex 200 gel filtration column (GE Healthcare Life Sciences) was equilibrated with 2 CV of Buffer A at 1 mL/min, 3 CV of Buffer D at 1.6 mL/min and 1 CV at 0.5 mL/min. Prepared sample was injected onto the column through the use of a sample loop at 0.5 mL/min. Buffer D was run through the column while collecting fractions of 1 mL. Fractions were analyzed for activity with 4-mu-Naglu substrate.

(44)

31

Changes from analytical-scale purification

During the analytical-scale purification, a 4.7 mL HiScale Capto™ MMC column (GE Healthcare Life Sciences) was used for the MMC step. The column was run at 1 mL/min and eluted with a 2 CV gradient from Buffer A to Buffer E (500 mM NaCl, 50 mM sodium phosphate, pH 6.2), followed by 5 CV of Buffer E. Fractions of 2 mL were collected.

For the Butyl-S step, a 1 mL HiTrap Butyl-S column (GE Healthcare Life Sciences) was run at 0.5 mL/min, using identical buffers to the preparative-scale procedure. Fractions of 1 mL were collected. Those that contained high levels of ΔS-Naglu-PTD4 activity were pooled, dialyzed with Buffer D and concentrated to

approximately 1 mL using Amicon Ultra-4 centrifugal filter devices (EMD Millipore). For the GFC step, all conditions were identical.

2.2.3 Bradford Protein Concentration and Naglu Activity Assays

Total protein activity of each sample was measured using a Bradford assay as follows. Bio-Rad Protein Assay concentrated dye was diluted to 50% in water. In a 96-well plate, 80 µL of reagent was mixed with 120 µL of each sample, incubated shaking at RT for 10 minutes, and read at 595 nm using a Wallac 1420 Victor2™ Microplate Reader (PerkinElmer). Protein concentration was calculated using a BSA standard curve ranging from 0 to 50 µg/mL. In order to generate values that fell within the linear range of this curve, crude medium prior to purification was diluted to 10% in SFX, while purification samples were diluted to 5% in Buffer D.

An activity assay was run on each sample using the following procedure (Chow & Weissmann 1981; Zhao & Neufeld 2000): a combination of 25 µL sample and 25 µL 4-mu-Naglu substrate was incubated at 37°C for 60 min. After addition of 1.5 mL of

(45)

32 glycine-NaOH buffer, the fluorescence of the resulting sample was measured using a Novaspec® visible fluorometer (360 nm narrow band excitation filter, 415 nm sharp cut emission detection filer, span fully counter-clockwise). The arbitrary fluorescence unit (AFU) generated by this assay was converted to units of enzyme activity using a 4-methylumbelliferone standard curve.

2.2.4 Carboxypeptidase Digestion

Purified ΔS-Naglu-PTD4 was subjected to C-terminal digestion with Bovine Carboxypeptidase A (BoCPA) (Sigma-Aldrich) or Carboxypeptidase Y (CPY) (Thermo-Fisher).

For digestion with BoCPA, protein substrates were dialyzed in CPA digestion buffer. BoCPA was added to approximate an enzyme to substrate molar ratio of 1:20. The final volume of each digest was 20 µL. The digests were incubated at room temperature for 1 hour or overnight, and the reactions were quenched with 2 µL 0.5 M

ethylenediaminetetracaacetic acid (EDTA).

For digestion with CPY, 100 mM sodium acetate pH 5.5 was used to dialyze the proteins. CPY was added to the sample at a 1:200 or 1:20 ratio. In some cases, 100 mM SDS was added to denature ΔS-Naglu-PTD4 while maintaining CPY activity. The digests were incubated for 1 hour or overnight at 37°C. PMSF was added at 1 mM to inhibit further digestion. Digested samples were assayed for activity using the 4-mu-Naglu substrate.

2.2.5 SDS-PAGE Protein Analysis

Samples were combined with 3X SDS buffer mix (NEB) containing 42 mM dithiothreitol (DTT) and boiled for 5 min. Sample pH was adjusted with NaOH (1 M),

(46)

33 and centrifuged at 13,000 rpm for 5 min. The prepared samples were loaded on an SDS-PAGE gel consisting of a 10% (v/v) glycine resolving layer and a 4% (v/v) tris-glycine stacking layer and resolved on a Mini-Protean® Tetra Cell electrophoresis unit (BioRad) at 125V. Apparent molecular weights were calculated relative to a pre-stained protein ladder (Precision Plus Protein™ Dual Color Standards, BioRad) that was loaded beside the samples.

2.2.6 Silver Stain Analysis

Proteins from SDS-PAGE gels were visualized using silver stain as follows. Gels were placed in fixative and microwaved for 90 seconds. Gels were placed in wash solution and microwaved for 90 seconds, followed by sensitize solution (0.02% w/v sodium thiosulfate) and microwaved 90 seconds. Gels were rinsed in distilled water for 90 seconds at room temperature and stained with 0.2% (w/v) silver nitrate at room temperature, shaking for 20 minutes. Gels were rinsed for 90 seconds at room temperature with distilled water before being developed in a solution of 0.04% (v/v) formaldehyde and 2% (w/v) sodium carbonate for as long as necessary. Developing was stopped with 5% (v/v) glacial acetic acid at room temperature with agitation for 5 minutes.

2.2.7 Coomassie Blue Analysis

A modified Neuhoff “Blue Silver” Coomassie procedure (Candiano et al. 2004) was also used for visualization of protein bands on SDS-PAGE gels. Gels were placed in fixative (50% v/v ethanol, 2% v/v phosphoric acid) and microwaved for 90 seconds. Gels were washed twice with distilled water, shaking for 20 minutes. Gels were stained (10% v/v phosphoric acid, 10% w/v ammonium sulphate, 20% v/v methanol, 0.12% w/v

(47)

34 Coomassie blue G-250) overnight with shaking. To remove background staining, gels were rinsed in distilled water.

2.2.8 Western Blot Analysis

Following SDS-PAGE, proteins were transferred from the acrylamide gel to Immunoblot® polyvinylidene fluoride (PVDF) membrane (BioRad) in 10% v/v methanol transfer buffer (25 mM tris-HCl, 200 mM glycine) at 10 V overnight. Membranes were rinsed in PBST and blocked with 5% (w/v) skim milk in PBST for 1 hour at room

temperature, or overnight at 4°C. Membranes were washed with PBST before incubation with various antibodies.

When probing against Naglu, polyclonal rabbit anti-Naglu ab137685 (Abcam) was applied at 1:2000 in PBST and incubated for 2 hours at room temperature or

overnight at 4°C. Unbound primary antibody was washed off with 3 cycles of PBST for 5 minutes each. Membranes were incubated with secondary goat anti-rabbit HRP (Thermo Scientific) conjugated antibody applied at 1:2000 in blocking reagent for 1 hour, and washed as above. When probing against the V5 epitope, a monoclonal mouse anti-V5-HRP-conjugated antibody (Invitrogen) was applied at 1:500 in blocking reagent for 1 hour at room temperature, and washed as above.

Blotted membranes were reacted with SuperSignal® West Dura extended exposure substrate (Pierce) for 5 minutes. Membranes were exposed to CL-X Posure™ film (Pierce) for times varying from 10 seconds to 2 hours and visualized using a Kodak X-OMAT 2000A Processor.

(48)

35 2.2.9 Cellular Uptake Studies

Human MPS IIIB skin fibroblasts (WG0421, McGill University Health Centre Cell Repository) and normal skin fibroblasts (MCH064, McGill University Health Centre Cell Repository) were grown in 60 mm cell culture plates containing fibroblast growth media. When 80-100% confluency was reached, the medium was removed and MPS IIIB fibroblasts cells were incubated with 1 mL of DMEM containing 400 units/mL of

purified ΔS-Naglu-PTD4. 1 mL of DMEM containing no ΔS-Naglu-PTD4 was added to normal skin fibroblasts and MPS IIIB skin fibroblasts as a positive and negative control, respectively. Following incubation for 30 minutes to 24 hours at 37°C and 5% CO2, cells

were harvested by mechanical passaging. The cell pellets were washed and resuspended in PBS. Cells were lysed via sonication for 20 seconds and centrifuged at 10,000 xg for 15 minutes. A 25 µL aliquot of each supernatant was used to measure Naglu activity, as described above.

Studies with variation in incubation medium, digestion via CPY and CPA, ΔS-Naglu-PTD4 concentration, and resuspension buffer were also conducted using a similar method.

(49)

36

Chapter 3 Results

3.1 Analytical Scale Purification

Medium containing active ΔS-Naglu-PTD4 was harvested from continuous shaker cultures. Purification trials were conducted on an ÄKTAprime™ system with small-scale columns containing a variety of chromatography media, including Affinity, HIC, and IEX. After assessing a range of buffer conditions on each medium, a three-step purification protocol was found to be most effective at purifying ΔS-Naglu-PTD4.

MMC was chosen as the initial capture step. 300 mL of crude Sf9 culture medium was equilibrated to a pH of 5.8 before application to a 4.7 mL MMC column. Figure 3.1A shows the protein elution during the column run, as measured by in-line UV at 280 nm. 4-mu-Naglu activity analysis found that ΔS-Naglu-PTD4 eluted from the column during the continuous-gradient increase of salt concentration and pH. As depicted by the box in Figure 3.1A, elution occurred over a volume of 32 mL, with 16 fractions of 2 mL measured to have significant enzyme activity. The partially purified fractions were pooled and adjusted to Buffer C conditions: 2 M NaCl, 50 mM Na acetate, and pH 5.8.

Butyl-S, a weak HIC medium, was chosen as the intermediate purification step. 48 mL of adjusted MMC-purified sample was applied to a 1.0 mL column. Figure 3.1B shows the protein elution during the column run, as measured by in-line UV at 280 nm. ΔS-Naglu-PTD4 was seen to elute from the column early in the desalting gradient. Elution occurred over a volume of 22.5 mL, with 15 fractions of 1.5 mL measured to have significant enzyme activity. The partially purified fractions were pooled, dialyzed, and concentrated to 1 mL in 150 mM NaCl, 50 mM Na acetate, and pH 5.8.

(50)

37 GFC was chosen as a polishing step. Concentrated sample was applied to a 120 mL column. As seen in the UV trace, protein eluted from the column in multiple peaks, with the ΔS-Naglu-PTD4 peak centered at 60 mL (Figure 3.1C). Two protein peaks are visible near the Naglu peak; other peaks occurred significantly after Naglu elution (not shown).

Relative to the equilibrated crude medium that was applied to the column, 46% of the total activity was present following MMC purification and adjustment to pre-HIC conditions. 28% of the total activity was still present following HIC purification and adjustment to pre-GFC conditions. After the completion of the three-step purification protocol, the yield of activity Naglu was 15% (Table 3.1).

Purification and Uptake Studies of Recombinant Human N-α-D-Acetylglucosaminidase from Sf9 Insect Cells

Supervisory Committee

Abstract

Table of Contents

List of Tables

List of Figures

List of Abbreviations

Acknowledgments

Dedication

Chapter 1 Introduction

1.1 Mucopolysaccharidosis III Type B

1.2 Treatment of Sanfilippo Syndrome B neurological symptoms

1.3 Expression of Human Recombinant Naglu

1.4 Purification of Naglu

1.5 Project Overview

Chapter 2 Materials and Methods

2.1 Materials

2.2 Methods

Chapter 3 Results

3.1 Analytical Scale Purification