Molecular Investigations of the CMT4D Gene N-myc Downstream-Regulated Gene 1(NDRG1)
Michael Hunter
This thesis is presented for the degree of Doctor of Philosophy at The University of Western Australia
Western Australian Institute for Medical Research And Centre for Medical Research School of Medicine and Pharmacology 2006
i ABSTRACT
Hereditary Motor and Sensory Neuropathy Lom (HMSNL) is a severe autosomal recessive peripheral neuropathy, the most common form of demyelinating Charcot- Marie-Tooth (CMT) disease in the Roma (Gypsy) population. The mutated gene, N-myc downstream-regulated gene 1 (NDRG1) on chromosome 8q24, is widely expressed and has been implicated in a wide range of processes and pathways. In this study we have aimed to assess the overall contribution of this gene to the pathogenesis of peripheral neuropathies, in cases where the most common causes of CMT disease have been excluded, as well as to gain clues about its function through the identification of its interactions with other proteins. Sequence analysis of NDRG1 in 104 patients with CMT disease and of diverse ethnicity identified one novel disease-causing mutation, IVS8- 1G>A (g.2290787G>A), which affects the splice-acceptor site of IVS8 and results in the skipping of exon 9. In addition, we have detected homozygosity for the known Roma R148X mutation in two affected individuals. Mutations in NDRG1 thus accounted for 2.88% of our overall group of patients, and for 4.68% of cases with demyelinating neuropathies. To gain an insight into NDRG1 function yeast two-hybrid screening of mouse sciatic nerve, human fetal brain and murine pluripotent cell line libraries were performed and identified interacting proteins whose known functions suggest NDRG1 involvement in cellular trafficking. Further analyses, focusing on apolipoproteins A-I (APOA1) and A-II (APOA2), confirmed their interaction with NDRG1 in mammalian cells. The results suggest a defect in Schwann cell lipid trafficking as a major pathogenetic mechanism in CMT4D. At the same time, database searches showed that the chromosomal location of NDRG1 coincides with a reported High-Density Lipoprotein-Cholesterol Quantitive Trait Locus (HDL-C QTL) in humans and in mice. A putative role of NDRG1 in the general mechanisms of HDL-mediated cholesterol transport was supported by biochemical studies of blood lipids, which revealed an association between the Gypsy founder mutation, R148X, and decreased HDL-C levels. These findings suggest that while peripheral neuropathy is the drastic result of NDRG1 deficiency, the primary role of the protein may be related to general mechanisms of lipid transport/metabolism.
ii TABLE OF CONTENTS
ABSTRACT ...... ii TABLE OF CONTENTS...... iii LIST OF TABLES ...... iv LIST OF FIGURES...... v ACKNOWLEDGEMENTS...... vii DECLARATION OF CONTENT...... ix PUBLICATIONS ARISING FROM THIS THESIS ...... x CHAPTER 1 – INTRODUCTION ...... 1 1.1 GENERAL OVERVIEW OF CHARCOT-MARIE-TOOTH (CMT) DISEASE ...... 1 1.2 HEREDITARY MOTOR AND SENSORY NEUROPATHY-LOM (HMSNL) ...... 3 1.3 PROPOSED ROLES OF NDRG1 DERIVED FROM NON-NEURAL TISSUE AND CELL-TYPES .. 10 1.4 CURRENT KNOWLEDGE OF NDRG1 EXPRESSION AND POSSIBLE FUNCTION IN PERIPHERAL NERVE ...... 15 1.5 AIMS OF THIS PHD PROJECT ...... 16 CHAPTER 2 – THE IDENTIFICATION OF NDRG1 MUTATIONS IN PATIENTS WITH CHARCOT-MARIE-TOOTH DISEASE ...... 18 2.1 BACKGROUND...... 18 2.2 METHODS...... 20 2.3 RESULTS ...... 26 2.4 DISCUSSION ...... 35 CHAPTER 3 – THE IDENTIFICATION OF NDRG1 PROTEIN-PARTNERS BY YEAST TWO-HYBRID LIBRARY SCREENING ...... 39 3.1 BACKGROUND...... 39 3.2 METHODS...... 44 3.3 RESULTS ...... 59 3.4 DISCUSSION ...... 76 CHAPTER 4 - CONFIRMING NDRG1-PROTEIN INTERACTIONS ...... 83 4.1 BACKGROUND...... 83 4.2 METHODS...... 87 4.3 RESULTS ...... 96 4.4 DISCUSSION ...... 115 CHAPTER 5 – EFFECT OF NDRG1 DEFICIENCY ON EXTRACELLULAR CHOLESTEROL TRANSPORT ...... 124 5.1 BACKGROUND...... 124 5.2 METHODS...... 127 5.3 RESULTS ...... 129 5.4 DISCUSSION ...... 133 CHAPTER 6 - MOUSE MODELS OF NDRG1 DEFICIENCY...... 137 6.1 BACKGROUND...... 137 6.2 METHODS...... 140 6.3 RESULTS ...... 143 6.4 DISCUSSION ...... 150 CHAPTER 7 – SUMMARY AND CONCLUDING DISCUSSION...... 152 REFERENCES...... 156
iii LIST OF TABLES
Table 1.1 Genes and loci implicated in different forms of CMT disease and their modes of inheritance ...... 2 Table 1.2. Conditions and agents regulating NDRG1 expression...... 11 Table 2.1. Incidence of mutations in common CMT genes determined from large cohort screening of CMT patients...... 18 Table 2.2. Primers used in the cDNA analysis of sequence variants potentially affecting RNA splicing ...... 24 Table 2.3. Single nucleotide polymorphisms in NDRG1 in 104 unrelated Caucasian individuals with CMT disease ...... 29 Table 2.4. Nucleotide diversity in NDRG1...... 31 Table 3.1. NDRG1-LexBD interactors detected by yeast two-hybrid screening of a murine pluripotent cell line (EML C.1) ...... 41 Table 3.2 False positive interactors identified from yeast two-hybrid screening of human fetal brain and mouse sciatic nerve libraries...... 63 Table 3.3. Candidate NDRG1 interactors identified from yeast two-hybrid screening in nerve tissue...... 64 Table 3.4. PRA1-protein partners identified from yeast two-hybrid screens ...... 79 Table 5.1. Blood lipid parameters and genotype ...... 132 Table 7.1. CMT diseases caused by mutations in genes implicated in trafficking...... 154
iv LIST OF FIGURES
Figure 1.1. Demyelination, axonal loss and onion bulbs in HMSNL nerve...... 6 Figure 1.2. Cytoplasmic inclusions in Schwann cells in HMSNL nerve...... 6 Figure 1.3. Curvilinear axonal inclusions in HMSNL nerve ...... 7 Figure 1.4. Myelin sheaths fail to compact in HMSNL nerve...... 7 Figure 1.5 Genomic and cDNA organisation of the NDRG1 gene and the R148X mutation ...... 9 Figure 1.6 Predicted domains of the NDRG1 protein point to a role in lipid metabolism ... 13 Figure 2.1. The IVS8-1G>A mutation results in exon 9 skipping...... 27 Figure 2.2. Prediction of the secondary structure of A) wild-type NDRG1 protein and B) mutant NDRG1 protein missing exon 9 as a result of the splicing IVS8-1G>A mutation...... 32 Figure 2.3. Surface accessibility predictions of residues of A) wildtype NDRG1 protein and B) mutant NDRG1 protein missing exon 9...... 33 Figure 3.1. Principles of the yeast two-hybrid method...... 40 Figure 3.2. Schematic overview of the yeast two-hybrid systems utilized in the screening for NDRG1-interacting proteins ...... 43 Figure 3.3. NDRG1 yeast two-hybrid constructs, domains, primer sequences and PCR conditions...... 45 Figure 3.4. Mouse sciatic nerve yeast two-hybrid library production ...... 50 Figure 3.5. Prenylated Rab Acceptor 1 yeast two-hybrid constructs, domains, primer sequences and PCR conditions...... 55 Figure 3.6. Apolipoprotein A-I yeast two-hybrid constructs, primer sequences and PCR conditions...... 56 Figure 3.7. Apolipoprotein A-II yeast two-hybrid constructs, primer sequences and PCR conditions...... 57 Figure 3.8 Mouse apolipoprotein family yeast-two hybrid constructs, primer sequences and PCR conditions...... 58 Figure 3.9. PCR amplification of library inserts from HIS and LacZ +ve colonies identified from screening a human fetal brain cDNA library with NDRG1-LexABD.. 60 Figure 3.10. PCR amplification of library inserts from ADE, HIS and LacZ +ve colonies identified from screening a mouse peripheral nerve cDNA library with NDRG1- Gal4BD...... 61 Figure 3.11. Reticulon-1C interacts with NDRG1...... 65 Figure 3.12. Mapping minimal NDRG1-LexABD sites of interaction with the full-length human Reticulon-1C AD fusion ...... 66 Figure 3.13. Full-length mouse and human PRA1 Gal4AD fusion proteins interact with NDRG1-Gal4BD but not with LaminC-Gal4BD nor NDRG1/IVS8-1G>A-Gal4BD. . 68 Figure 3.14. Mapping minimal NDRG1-Gal4BD sites of interaction with the full-length mouse Pra1-Gal4AD fusion...... 69 Figure 3.15. Mapping minimal Pra1-Gal4AD sites of interaction with the full-length NDRG1–Gal4BD fusion...... 70 Figure 3.16. Full-length mouse and human ApoA-I and ApoA-II Gal4AD fusion proteins interact with NDRG1-Gal4BD but not with LaminC-BD nor NDRG1/IVS8-1G>A- BD...... 72
v Figure 3.17. Mapping minimal NDRG1-Gal4BD sites of interaction with the full-length mouse Apoa1 and Apoa2-Gal4AD fusion proteins...... 73 Figure 3.18. Schematic of mouse Apoa1 -Gal4AD proteins and analysis of interaction with the full-length NDRG1-Gal4BD protein...... 74 Figure 3.19. NDRG1 binds with mouse Apoa1 and Apoa2 but not with other apolipoprotein family members...... 75 Figure 4.1. Schematic representation of the principle of detecting the co-localisation of two proteins using immunochemistry...... 83 Figure 4.2. A schematic overview of the confirmatory protein-protein interaction assays utilized to confirm the NDRG1 and apolipoprotein interaction...... 85 Figure 4.3. Co-localisation of NDRG1 and APOA1 in transfected HepG2 cells...... 97 Figure 4.4. Co-localisation of endogenous NDRG1 and APOA1 in HepG2 cells...... 100 Figure 4.5. Confocal microscopy of human sural nerve sections stained for NDRG1 and APOAl...... 103 Figure 4.6. Western blot of the pulldown binding assay between NDRG1/GST recombinant protein (expressed and purified from E. coli BL21) and HA-Tagged Apoa1 (expressed and purified from COS7 cells) reveals no interaction...... 105 Figure 4.7. Western blot of the pulldown assay between recombinant NDRG1-GST fusion protein (expressed and purified from Sf9 insect cells) and purified human APOA1 (Sigma) fails to detect the NDRG1 and APOA1 interaction...... 106 Figure 4.8. Western blot of the Pulldown assay between NDRG1/GST recombinant protein (expressed and purified from Sf9 insect cells) and mouse sciatic nerve and kidney lysates...... 107 Figure 4.9. Western blots of co-immunoprecipitation experiments of NDRG1 and apolipoprotein A-I from 1) mouse sciatic nerve 2) mouse kidney and 3) HepG2 cell lysates fails to show an interaction...... 109 Figure 4.10. NDRG1 interacts with mouse Apoa1 in living mammalian cells only when Rluc is tagged to Apoa1 and not NDRG1...... 112 Figure 4.11. NDRG1 interacts with human APOA1 and APOA2 in living mammalian cells : Bioluminescence resonance energy transfer (BRET) assays...... 113 Figure 4.12. NDRG1 interacts with the mouse Prenylated rab acceptor protein 1 (Pra1) in living mammalian cells : Bioluminescence resonance energy transfer (BRET) assay.114 Figure 4.13 Protein-protein interaction network links NDRG1 to membrane trafficking and lipid transport pathways...... 118 Figure 5.1. The reverse cholesterol transport pathway...... 125 Figure 5.3. NDRG1 is a positional candidate for the human HDL-QTL on chromosome 8q24...... 130 Figure 6.1. Northern blotting of kidney RNA...... 144 Figure 6.2 RT-PCR and cDNA sequence of the stretcher (str) mutation...... 145 Figure 6.3. Western blotting of kidney lysates ...... 147 Figure 6.4. Light micrographs of sciatic nerve in wild-type and str mutant mice...... 149
vi ACKNOWLEDGEMENTS
I would like to extend thanks to a number of people who have made this research possible. Firstly to my principal supervisor Luba, whose motivation, drive and knowledge has, and always will be an inspiration to me. I thank you for helping me realize my potential and providing me with the guidance and opportunity to work on such a unique and stimulating project. I would also like to thank my co-supervisor Dora, who has shared the joys and frustrations of our first venture into the wonderful world of protein research. It has been fantastic sharing this journey with you Dora, thank you for your advice and support, we have learnt so much together. A big thank you also to my friends at WAIMR; Rebecca, Janina, Dave, Bharti, Tom, Lorna and Matt for all the fun times and advice you have all provided throughout the project. Also to the guys at ECU, where
I first started this work, Liz, Andrea, Ian, Cheryl and Salvatore.
I would also like to thank Evan Ingley, Jim Tiao and Amerigo Carrello for their regular advice and support of all things proteomic. Also to Rosalind King for her hospitality in the UK and teaching me the art of cryosectioning and immunostaining; and
Marcel Kwa at the AMC for inter-contential Western blotting when antibodies were not available here in WA. Also gratitude to Dr Kokame and Dr Miyata for recently supplying us with the excellent NDRG1 antibody that has made much of the work in this thesis possible. I would also like to thank the ongoing support of patients and family members who have participated in this project over the years and our Bulgarian colleagues for collection and analyses of blood samples. Thanks also to Gerald Watts and Dick Chan at the Lipoprotein Research unit WAIMR for the blood lipid and statistical analyses.
vii I would also like to thank my family; my Mum and Max, my Dad and Hayley, and my sister Michele for their ongoing support and encouragement. Also to my mates;
Harro, Crawley, Kingy, Yak, Shag and Ryno for regularly dragging me out of bed to go surfing at the crack of dawn for some much needed saltwater therapy. A special mention also to my long-time friend and companion of the furry, four-legged variety, Candy, who sadly passed away before the completion of this thesis at the grand age of 17.
Finally and most importantly I would like thank my partner Zoë, whose love, support, patience and understanding has made the completion of this thesis possible.
viii DECLARATION OF CONTENT
To my best knowledge and belief, I certify that this thesis does not incorporate any material previously submitted for a degree or diploma in any other institution of higher education. For any work in this thesis that has been co-published with other authors, I have the permission of all co-authors to include this work in my thesis. The thesis complies with the guidelines of the University of Western Australia and the Faculty of Medicine and Dentistry.
Michael Hunter (PhD candidate)
Prof Luba Kalaydjieva (Principal supervisor)
Dr Dora Angelicheva (Co-supervisor)
ix PUBLICATIONS ARISING FROM THIS THESIS
Journal Publications
Hunter M, Bernard R, Freitas E, Boyer A, Morar B, Martins IJ, Tournev I, Jordanova A, Guergelcheva V, Ishpekova B, Kremensky I, Nicholson G, Schlotter B, Lochmuller H, Voit T, Colomer J, Thomas PK, Levy N, Kalaydjieva L. Mutation Screening in the N-myc Downstream-Regulated Gene 1 (NDRG1) in patients with Charcot-Marie Tooth disease. Human Mutation. 2003. 22(2):129-35 (Chapter 2).
Hunter M, Angelicheva D, Tournev I, Ingley E, Chan D, Watts GF, Kremensky I, Kalaydjieva L. The CMT4D gene N-myc Downstream-Regulated Gene 1 (NDRG1) interacts with apolipoproteins A-I and A-II and is a functional candidate for the HDL-C QTL on chromosome 8q. Biochem. Biophys. Res. Comm. 2005. 15; 332(4): 982-92 . (Chapters 3, 4 & 5).
Abstracts
Hunter M, Angelicheva D, Tournev I, Chan D, Watts G, Kalaydjieva L. Peripheral neuropathy gene is a functional candidate for the HDL-C locus on Chr8q24 . 4th Australasian Human Gene Mapping Meeting 2004, Fremantle, Western Australia (Awarded first prize for best student paper). (Chapters 3, 4 & 5).
Hunter M., Angelicheva D., Ingley E., Kalaydjieva L. Identifying NDRG1 protein-partners - clues to the neuropathology of Charcot-Marie-Tooth disease type 4D. Combio 2004, Burswood, Western Australia. (Chapters 3, 4 & 5).
x CHAPTER 1 – INTRODUCTION
1.1 General overview of Charcot-Marie-Tooth (CMT) Disease
Charcot-Marie-Tooth (CMT) disease, or Hereditary Motor and Sensory Neuropathy
(HMSN), is a heterogeneous group of inherited peripheral neuropathies, which generally share a phenotype characterised by weakness and wasting of the distal limb muscles that is frequently associated with distal sensory loss and skeletal deformities. Two main sub groups have been defined by histopathological and electrophysiological features, and include demyelinating forms (CMT1) (where nerve conduction velocities (NCVs) are <38 m/s), and axonal forms (CMT2) (where NCVs > 38 m/s) (1-4). Recently an intermediate subgroup of
CMT diseases have been described with NCVs that overlap CMT1 and CMT2 and present with a neuropathology displaying characteristics of both demyelination and axonal loss
(5,6).
At least 20 genes and several loci, (where the genetic aberration is yet to be identified), are implicated in the pathogenesis of CMT disease (Table 1.1) and include genes encoding proteins with obvious roles in nerve, such as myelin and axonal structural proteins.
In addition mutations in many widely expressed genes involved in fundamental processes such as transcription, translation, signalling and membrane trafficking have been implicated in the molecular pathogenesis of some CMTs and point towards important roles which are essential to normal peripheral nerve function and morphology (2,7). The large genetic heterogeniety and varied functional categories to which CMT genes belong makes molecular diagnosis a difficult task. This is a problem which is compounded by the fact that different mutations in the same gene can be inherited as autosomal dominant or autosomal recessive traits, and also can result in markedly different phenotypes.
1 Table 1.1 Genes and loci implicated in different forms of CMT disease and their modes of inheritance
CMT1 CMT2 CMT-I Demyelinating Axonal Intermediate
Dominant Dominant Dominant Gene Locus Gene Locus Gene Locus CMT1A PMP-22 17p11 CMT2A1 KIF1B 1p36 CMTDIA unknown 10q24 CMT1B P0 1q22 CMT2A2 MFN2 1p36 CMTDIB DNM2 19p12 CMT1C LITAF 16p13 CMT2B RAB7 3q13-q22 CMTDIC unknown 1p34 CMT1D EGR2 10q21 CMT2C unknown 12q23-q24 CMTDI3 P0 1q22 CMT1E P0 1q22 CMT2D GARS 7p14 CMT 2E NF-L 8p21 CMT1F NF-L 8p21 CMT2E NF-L 8p21 HNPP PMP-22 17p11 CMT2F HSPB1 7q11-q21 HMSN3 PMP-22, P0 8q23 CMT2G unknown 12q12 SOX10 22q13 CMT2I P0 1q22 Cxn-31 1p35 CMT2J P0 1q22 ARHGEF10 8p23 CMT2L HSBP8 12q24
Recessive Recessive Recessive Gene Locus Gene Locus Gene Locus CMT4A GDAP1 8q21.1 AR-CMT2A LaminA/C 1q21 CMTRIA GDAP1 8q21.1 CMT4B MTMR2 11q23 AR-CMT2B unknown 19q13.3 CMT4B2 SBF2 11p15 CMT 2K GDAP1 8q21.1 CMT4C SH3TC2 5q23-q33 CMT 2H unknown 8q21.3 CMT4D NDRG1 8q24 CMT4E EGR2 10q21 CMT4F PXN 19q13 HMSNR HK1 ? 10q23 CMT4H unknown 12q12 CCFDN CTDP1 18qter
X-linked X-linked Gene Locus Gene Locus CMTX Cx32 Xq13.1 CMTX unknown Xq24-q26
Adapted from the Neuromuscular Disease Center Database, Washington University, USA (available from www.neuro.wustl.edu/neuromuscular/time/hmsn.html)
2 Relative to autosomal dominant CMT disease, autosomal recessive forms are rare, however they are clinically more severe (8) and as they are less likely to result from mutations in structural myelin proteins, understanding their molecular basis may provide information into unknown mechanisms of axon-glial interaction and PNS development. Our laboratory has been involved in the mapping of genes and identification of founder mutations underlying a number of hypo/demyelinating autosomal recessive neuropathies occuring in Gypsies: hereditary motor and sensory neuropathy-Lom (HMSNL) (MIM
601455) (9,10), hereditary motor and sensory neuropathy-Russe (HMSNR) (MIM 605285)
(11,12), congenital cataracts facial dysmorphism neuropathy (CCFDN) syndrome (MIM
604168) (13,14) and CMT4C (MIM 601596) (15). The identification of the underlying mutation in these disorders has implicated a number of seemingly unlikely genes in the molecular pathogenesis and include; the Carboxyl Terminal Domain Phosphatase 1
(CTDP1) of mRNA Polymerase II, mutated in CCFDN (13,14), Hexokinase 1 (HK1) as a possible cause of HMSNR (11,12) and the N-myc Downstream Regulated Gene 1 (NDRG1) where a truncating mutation is responsible for the demyelinating CMT4D disease, HMSNL
(9,10). These findings point towards hitherto unknown roles in the PNS and add to the functional diversity of genes implicated in CMT disease. The focus of this PhD project is on
NDRG1, a gene whose function is yet to be resolved and where knowledge about its contribution to peripheral neuropathy is incomplete.
1.2 Hereditary motor and sensory neuropathy-Lom (HMSNL)
Hereditary motor and sensory neuropathy-Lom (HMSNL) (CMT4D) occurs in
Romani (Gypsy) groups descended from a small founder population known as the Vlax
Roma. The neuropathy was first described in 14 affected individuals from the endogamous
Gypsy community of Lom, a small town on the Danube river in the North-West of Bulgaria
3 and 6 affected individuals belonging to the Wallachian Gypsy group (9). Since that time, cases have been detected and described in Gypsy patients throughout Europe (16-19) and the identification of closely related haplotypes at the HMSNL locus among geographically dispersed Gypsy groups has supported the common origin of the mutation (10,20). The carrier frequency of the HMSNL mutation ranges from 2% overall, up to 16% for specific high-risk groups, making it the most common and widespread cause of neuropathy in the
Gypsy community (20,21).
Clinical phenotype of HMSNL
HMSNL is a severe, early onset neuropathy with a clinical phenotype that includes distal muscle wasting and atrophy, skeletal and foot deformities, tendon areflexia, sensory loss affecting all modalities and severe reduction in nerve conduction velocities (NCVs), indicative of demyelination (10,17,18,22,23). Affected patients present with gait disorder during the first decade of life followed by upper limb weakness in the second decade. Many patients progress to severe disability by the fifth decade. Deafness is an invariant feature of the disorder and develops during the second or third decade of life, with abnormalities in the brain-stem auditory-evoked potentials suggesting involvement of the entire tract, including the central auditory pathways (18,24). Other sensory deficits include loss of sensation of vibration, pin-prick and impaired spatial positioning of the lower extremities (10,18,23-25).
Like other autosomal recessive forms of CMT disease the onset is insidious, with a rapid progression to severe disability. The early onset of clinical features suggests impairment during the early development of the PNS.
4 HMSNL neuropathology indicates a defect in Schwann cell-axonal interactions
Neuropathological studies of peripheral nerve biopsies from HSMNL patients and more recently in Ndrg1-knockout mice, which present with a neuropathy resembling the human phenotype (26), reveal a severe depletion in the myelinated nerve fibre population and a reduction in the size of preserved myelinated fibres (Figure 1.1) (23,27).
Ultrastructural changes suggest Schwann cell dysfunction, characterised by hypomyelination and demyelination/remyelination, and poor hypertrophic response (onion-bulb formation) to the demyelination process. Onion bulb formations are indicative of repeated demyelination and remyelination and are often seen in peripheral neuropathies in which the myelination process is adversely affected (28). In biopsies from young HMSNL patients, basal lamina onion bulbs, characterised by a central unmyelinated fibre or demyelinated axons are present. In the older patients, these hypertrophic changes are less frequent. Instead, short circumferential Schwann cell processes without a central axon are noted (29). These differences can be attributed to initial demyelinating and remyelinating episodes resulting in onion bulb formation, but it would seem that failure of sustained communication between myelin and the axon results in axonal death, subsequently resulting in the regression of the onion bulbs over time. The widespread axonal loss and impaired regenerative activity point towards a primary defect in Schwann cell-axonal interactions crucial to axonal survival.
Aberrations to Schwann cell morphology include the presence of myelin debris inside Schwann cells surrounding demyelinated axons and cytoplasmic inclusions (Figure
1.2). These inclusions consist of amorphous granular material enclosed in irregular membrane-bound components. Axonal inclusions are also evident (Figure 1.3) and consist of multiple-curvilinear bodies (23,27) resembling those seen in experimental vitamin E deficiency (30). These features, coupled with evidence of hypomyelination, perhaps suggest an aberration in trafficking as a possible pathogenic mechanism.
5 Figure 1.1. Demyelination, axonal loss and onion bulbs in HMSNL nerve Sural nerve biopsy showing evidence of demyelination and axonal loss (indicated by solid arrows) and onion bulb formations (indicated by open arrows). (image courtesy RHM King).
Figure 1.2. Cytoplasmic inclusions in Schwann cells in HMSNL nerve Pleomorphic inclusions (indicated by asterisk) are found in the Schwann cell cytoplasm. (m - myelin; a - axon) (image courtesy RHM King).
6 Figure 1.3. Curvilinear axonal inclusions in HMSNL nerve Curvilinear inclusions (inset) in the axon resemble those found in experimental vitamin E deficiency (image courtesy RHM King).
Figure 1.4. Myelin sheaths fail to compact in HMSNL nerve Failure of compaction of the innermost myelin lamellae (asterisks) ensheathing the axon (a) (image courtesy RHM King).
7 Other disruptions to normal myelination in HMSNL include failure of compaction of the innermost myelin lamellae (Figure 1.4) and abnormalities of the myelin sheath presenting as oblique Schmidt-Lanterman incisures. These incisures are thought to function as a pathway between the inner and outer compartments of the Schwann cell cytoplasm, providing a route by which metabolites can pass through towards the axon (28). Variation in myelin thickness is evident in adjacent internodes and presents as an abrupt step like- reduction in thickness, resulting in the formation of pseudonodes. Complete demyelination of scattered internodes is also noted.
Other histological observations in HMSNL include non-specific morphological changes to other tissues besides Schwann cells and axons. These include minor changes to blood vessels, such as the reduplication of the basal lamina around epithelial cells and the pericytes of the endoneural microvessels (17). Prominent endoneural fibrosis (27) and the accumulation of extracellular collagenous material (18) has also been noted in some patients.
The truncating R148X mutation in NDRG1 is a cause of HMSNL
The HMSNL disease gene interval was mapped using linkage analysis to the chromosomal location 8q24 (9). Refined mapping to a 206kb region and subsequent large- scale genomic sequencing identified a single nonsense mutation resulting in a premature stop codon (R148X) in the gene NDRG1 (10) (Figure 1.5) predicted to result in a truncated protein.
NDRG1 encodes an evolutionary conserved protein with homologues identified across species from sunflower (31) to primates, and belongs to a family of genes whose members include; NDRG2 on chromosome 14q11.2, NDRG3 on chromosome 20q11.21- q11.23 and NDRG4 on chromosome 16q21–q22.3 (10,32,33). A conserved family of genes
8 A GC Box
TATA Box 5 kb TATA-less CpG island Enhancer 58 119 bp
B 3014 bp
1 2 3 4 5 6 7 8 9 10 11 12-15 16
Untranslated region C WT
C
T
R148X
Figure 1.5 Genomic and cDNA organisation of the NDRG1 gene and the R148X mutation A) Genomic organisation as described by Kalaydjieva et al, (10). NDRG1 consists of 16 exons spanning 60kb of genomic sequence. A CpG island overlaps the first exon and the 5” end of the intron 1. A potential promoter with a TATA box, a GC box and a TATA-less enhancer are located upstream of the first exon. The 5'UTR is split between the first two exons. B) cDNA structure. Translation starts from nucleotide position 123. C) HMSNL mutation. Chromatogram of the the C to T change at base 564 which results in a premature stop signal at codon 148 in exon 7 (R148X). Nucleotide positions are given relative to the published sequence (Genbank accession number D87953).
9 has also been identified in mice (34). The gene has been identified previously in a number of unrelated studies investigating various pathologies and physiological states, resulting in a varied nomenclature: Reducing agent and Tunicamycin-responsive Protein, RTP (35);
Differentiation-Related Gene, DRG1 (36); Calcium Activated Protein CAP43 (37); Reduced in Tumor, 42kDa rit42 (38) and Protein Regulated by OXYgen1 PROXY-1 (39).
1.3 Proposed roles of NDRG1 derived from non-neural tissue and cell-types
At the outset of this PhD project, existing concepts of NDRG1 function were largely derived from in vitro studies in different cell-types far removed from peripheral nerve. It was first identified as gene upregulated in HUVEC cells exposed to homocysteine (35,40) and also by lysophosphatidylcholine (41) suggestive of involvement in the response to oxidative stress and the pathogenesis of atherosclerosis. Other studies investigating its expression have linked NDRG1 to a multitude of pathways and processes including stress responses to hypoxia (39,42) and heavy metal exposure (37), differentiation and growth arrest (36,38,43), carcinogenesis (44,45) and apoptosis (46). A role as a developmental gene has been documented for Ndr1 which, in the mouse embryo, is repressed by N-myc and upregulated in cells committed to terminal differentiation (34). Numerous inducing agents and conditions have been shown to influence its expression (Table 1.2) making assigning a definitive biological pathway or function difficult.
10 Table 1.2. Conditions and agents regulating NDRG1 expression
Up-regulating conditions and agents References Cell differentiation (36,47) Activation of p53 (38) Hypoxia (39,48) Decreased glucose concentration (49) Homocysteine (35) Mercaptoethanol (35) Lysophophatidinylcholine (41) Tunicamycin (35) Okadaic acid (37) Calcium ionophore (37,48) DNA-damaging agents (38) 1,25-(OH)2 Vitamin D3 (43) Synthetic retinoids (43) Phorbol myristate acetate (37,43) Forskolin (37)
Down-regulating conditions and agents N-myc/c-myc pathways (34) Cell proliferation (43) Androgens (50-52) Malignant transformation (38) Histone deacetylase (53)
Localisation studies suggest NDRG1 is a translocating molecule
Investigations in various tissues and cell-types have shown that NDRG1 protein is widely expressed (54), with high levels detected in colonic epithelium (36,45), ovary and prostate tissue (47), epithelia cells lining the kidney tubules (54) and placental membranes
(55,56). Differences in its subcellular distribution among varying tissue types and cell lines, and also at various stages of cell differentiation have been noted. NDRG1 has been shown to reside in both the cytoplasm and nucleus in prostate epithelium and placental chorion
(36,38). A diffuse cytoplasmic pattern of distribution is noted in poorly differentiated colon epithelial cells, with a shift to the plasma membrane upon differentiation (36). Punctate 11 distribution suggestive of association with secretory vesicles has been noted in mast cells
(57). The multiple subcellular localisation patterns suggest that NDRG1 may be an actively translocating enzyme and/or signaling molecule, functions which are perhaps supported by evidence that the protein can undergo phosphorylation (56,58).
Protein domains suggest a role for NDRG1 in lipid metabolism
Homology modeling and functional domain predictions have identified a number of features of the NDRG1 protein that suggest a role in lipid metabolism and may be important in revealing a function in peripheral nerve (59) (Figure 1.6). The protein has been shown to be an atypical member of the alpha/beta hydrolase (abhydrolase) super-family of proteins
(59) that include a large number of enzymes involved in lipid and cholesterol metabolism and biosynthesis (60). Although their substrate specifity is diverse, three-dimensional modeling and crystallography analysis of abhydrolase-type enzymes, has revealed a topology common to all members (61) that consists of at least five parallel beta-strands, nucleophilic elbow and a conserved three-residue triad that imparts catalytic activity (62).
While the abhydrolase fold is conserved in the structure of NDRG1, the necessary residues that form the catalytic triad are absent making it unclear whether the protein can function as an abhydrolase (59). However other regions of the protein suggest it may have an enzymatic function. Bioinformatic analyses consistently identify a putative phosphopantetheine- attachment site in the N-terminal first third of the protein.
Phosphopantetheine is an essential cofactor in the biosynthesis of fatty acids (63), forming the reactive unit of acyl carrier protein and coenzyme A (CoA). The moeity can act as a “swinging arm” that is thought to bring substrates to proximity to thioesterase active sites. Interestingly, a putative esterase/lipase/thioesterase active site is predicted to span the phosphopantethiene-binding domain of NDRG1 suggesting these two domains may function
12 αβ-hydrolase fold (aa81 – 309)
1 394
Esterase/lipase/thioesterase active site Phospopantetheine binding domain Tri-Deca repeat motif
Figure 1.6 Predicted domains of the NDRG1 protein point to a role in lipid metabolism The NDRG1 protein has a molecular weight of 43 kDa and consists of 394 amino acids. In silico protein domain predictions have revealed that NDRG1 is a member of the αβ- hydrolase superfamily (59) however it lacks the functional catalytic triad common to other hydrolases. A predicted esterase/lipase/thioesterase active site and a phosphopantethiene attachment site point towards involvement in lipid metabolism. A unique motif repeated 3 times at the C-terminal end of the protein has been shown to bind metal species in vitro (64) however its biological substrates have yet to be identified. The 3D structure of the protein is unknown.
13 in tandem in the binding and possible activation of phosphopantetheine-containing molecules. Phosphopantetheine and its derivatives play an important role in peripheral nervous tissue as evidenced by neuropathy caused by mutations to the peroxisomal phosphopantetheine-derived enzymes, phytanoyl-CoA hydroxylase and alpha-methylacyl-
CoA rasemase (65), and dietary deficiency of pantothenic acid, a precursor of phosphopantetheine has also been shown to result in peripheral nerve degeneration (66) and neuropathy (67).
A unique characteristic of NDRG1 is the presence of a three tandem, 10 amino acid repeat at its C-terminus (GTRSRSHTSE)3. To date this motif shows no homology to any known domain or protein in the public databases. An in vitro study has shown that the repeat sequence of NDRG1 can participate in the binding of the metal ions, Ni2+ and Cu2+
(64). Interestingly, previous investigations have shown that NDRG1 expression is induced by a number of reactive metal species including nickel compounds (48). The biological significance of these findings and whether they are related to this motif is not known.
The variation in cell types and tissues investigated, coupled with the multitude of agents and conditions known to influence NDRG1 expression have made it difficult to assign a definitive function. Moreover, the existing proposed roles in protecting cells from chemical and physical stress, and preventing or counteracting malignant transformation, are not reflected in the phenotype that results from deficiency of the protein. This suggests that the
NDRG1 interacts in a pathway or process that is critical for normal nerve function.
14 1.4 Current knowledge of NDRG1 expression and possible function in peripheral nerve
Prior to the identification of involvement in the pathogenesis of CMT4D disease, information about the expression of NDRG1 in peripheral nerve was unknown. Subsequent information obtained from S.A.G.E library screening (10), and more recently through quantitative analysis of its mRNA transcript (68), have shown abundant expression in peripheral nerve, significantly in excess of other tissues. Immunohistochemical analyses reveal that the protein is localised in myelinating Schwann cells with no evidence of axonal expression (10,26,68,69) indicating that the axonal loss in CMT4D is due to Schwann cell dysfunction. The exact role NDRG1 plays in the Schwann cell is unclear. The protein is found within multiple subcellular locations including the abaxonal cytoplasm and perinuclear regions, and has been shown to colocalise with the Myelin-associated glycoprotein (MAG) in paranodal regions, incisures, and in the inner mesaxon of myelin sheaths (68).
A role in myelination is suggested through expression studies during early development of the PNS which show NDRG1 levels rise sharply after birth and remain high throughout the major myelination period, paralleling the expression of other major myelin genes such as PMP22 (68). This is also evident in nerve crush and transection studies where high NDRG1 levels are detected in Schwann cells during active remyelination (68,69) pointing towards an important role in the myelination process.
However, NDRG1 does not appear to be essential for the initial formation of the myelin membrane, as ultrastructural studies in the knockout mice show myelin sheaths develop normally (26). It is only after myelination is established in which there is a rapid degradation and associated axonal loss. Whether this is through a failure in the direct delivery of myelin components necessary for maintenance, and/or due to the effects of an
15 aberration in Schwann cell-axonal communication that results in a reversion to an undifferentiated state and subsequent axonal death is unclear.
The mechanisms underlying SC-axonal interactions and the process of myelination are among the most complex in biology (70-72). Recent gene array studies have revealed a multitude of pathways activated during Schwann cell differentiation and development of the
PNS (73,74). Understanding the role of NDRG1 and the molecular pathogenesis of CMT4D may provide some insight into at least one pathway essential to the normal SC-axonal interactions. The expression of the protein in myelinating Schwann cells together with the neuropathological observations in NDRG1 deficiency point towards a breakdown in SC- axonal communication that is critical in maintaining the myelin membrane and survival of axons. As axonal loss, rather than demyelination, is the primary cause of disability in demyelinating neuropathies, understanding the molecular basis of CMT4D may also provide clues to the general pathogenetic mechanisms of peripheral neuropathies.
1.5 Aims of this PhD project
The fact that the pathology of CMT4D is restricted to the peripheral nervous system, strongly suggests that NDRG1 has a specific function in nervous tissue. It is likely that the protein is involved in a pathway that is essential in maintaining the myelin membrane either through the delivery of components or Schwann cell-axonal interactions. The abundant expression of the NDRG1 protein in peripheral nerve during stages of active myelination and throughout adulthood, coupled with the protein domain predictions, could point towards involvement in the lipid turnover and biosynthetic pathways operating in myelinating
Schwann cells. The following chapters describe the experiments that were undertaken in order to provide information regarding the contribution of NDRG1 to the molecular pathogenesis of CMT disease and its role in normal nerve.
16 The specific aims of the current investigation were:
To identify and assess the contribution of NDRG1 mutations to the pathogenesis of
CMT disease among a large and ethnically diverse sample of patients with various
forms of CMT disease.
. To gain an insight into the physiological role of NDRG1 and pathogenesis of
CMT4D through the identification of its interacting proteins.
17 CHAPTER 2 – THE IDENTIFICATION OF NDRG1 MUTATIONS IN PATIENTS WITH CHARCOT-MARIE-TOOTH DISEASE
2.1 Background
Mutations in genes encoding the common myelin proteins such as peripheral myelin protein 22 (PMP22), myelin protein zero (MPZ) and connexin 32 (Cx32) account for a large percentage of cases of CMT disease (75-80), however a significant number of cases exist where the underlying mutation(s) have yet to be identified (Table 2.1). Identifying these mutations is difficult given the large number of genes involved in the pathogenesis of CMT disease, and also the fact that different mutations in the same gene can be inherited as autosomal dominant or autosomal recessive traits, and can result in markedly different phenotypes. Often cases of CMT disease are sporadic and the absence of extensive family history and availability of DNA samples precludes genetic linkage studies, thus contributing to the difficulty in choosing candidate genes for mutation identification.
Table 2.1. Incidence of mutations in common CMT genes determined from large cohort screening of CMT patients
Sample PMP22 Cx32/GJB1 MPZ Unknown Reference number 819 579 - - 240 (75) (70.7%) (30.3%) 174 61 12 6 95 (76) (35.1%) (6.8%) (3.5%) (54.6%) 170 100 12 4 54 (77) (58.8%) (7.1%) (2.3%) (31.8%) 153 84 11 5 53 (78) (54.9%) (7.2%) (3.2%) (34.7%) 127 46 14 12 55 (79) (35.9%) (10.9%) (9.4%) (43.8%) 63 43 - - 20 (80) (68.3%) (31.7%)
18 Prior to this investigation the private Gypsy R148X mutation (10) had remained the only evidence pointing to a unique role of NDRG1 in the peripheral nervous system and placing it on the list of genes involved in CMT disease. The overall contribution of other
NDRG1 mutations to the molecular basis of CMT disease was unknown. The aim of this part of the study was to identify and assess the contribution of mutations in NDRG1 to the molecular pathogenesis of peripheral neuropathies in patients where preliminary analyses had excluded the most common genetic causes of CMT disease.
By identifying and determining the incidence of NDRG1 mutations it would indicate the future feasibility of including this gene in mutation screening in patients in which the common causes of CMT disease have been excluded. At the same time, identifying novel
NDRG1 mutations could also contribute to our understanding of the functionally important domains of the protein and how they contribute to the normal maintenance of peripheral nerve.
19 2.2 Methods
2.2.1 Subjects
The study included 104 unrelated individuals affected by different forms of peripheral neuropathy: 51 male and 53 female, aged between 1 and 88 years. DNA samples were provided by a number of centres including; the Laboratoire de Génétique
Moléculaire, Département de Génétique Médicale, Hôpital d’Enfants de la Timone,
Marseille, France, the Department of Neurology and Laboratory of Molecular Pathology,
Medical University, Sofia, Bulgaria; The Neurobiology Department, ANZAC Research
Institute, University of Sydney, Australia; The Friedrich-Baur Institute, Ludwig-
Maximilians University, Germany; the Department of Pediatrics, University Hospital
Essen, Germany, Hospital Sant Joan de Déu, Barcelona, Spain; Institute of Neurology and
THE National Hospital for Neurology and Neurosurgery, London, UK.
Based on neurophysiological investigations (nerve conduction velocity measurements) 64 patients had been classified as having demyelinating forms of CMT disease, 15 as axonal, 9 mixed, and 16 were unspecified. According to mode of inheritance, 62 cases were sporadic (with evidence of consanguinity in 6 of the families in this group), 10 were documented autosomal recessive, 15 were autosomal dominant and 17 unknown. The ethnic origin of the patients was as follows: 32 French, 24 Bulgarian, 16
Italian and Corsican, 11 German, 5 British, 3 Lebanese, 3 Romani, 1 Austrian, 1
Hungarian, 1 Croatian, 1 Turkish, 1 Greek, 1 Persian, 1 Indian, 1 Spanish, 1 Moroccan and
1 Argentinian.
Common causes of peripheral neuropathies, such as 17p11.2 duplications and mutations in Peripheral Myelin Protein 22 (PMP22), Myelin Protein Zero (MPZ),
Connexin 32 (Cx32) and Early Growth Response 2 (EGR2), had been excluded in these
20 patients. Informed consent had been obtained from all participating individuals or, in the case of minors, from their parents.
2.2.2 Sequencing NDRG1
Sequencing of the NDRG1 gene was a collaborative effort carried out by memebrs of our laboratory including Michael Hunter, Ian Martins and Elizabeth Freitas, and the
Laboratoire de Génétique Moléculaire, Département de Génétique Médicale, Hôpital d’enfants de la Timone, Marseille, France. Screening for sequence variation in NDRG1 was conducted by PCR amplification and direct sequencing of the PCR products of all coding regions and at least 100 bp of the flanking introns, including part of the 5' and 3'
UTR. Amplification was performed in 20µL reactions using 30ng of genomic DNA, 1X
PCR buffer (Qiagen Pty Ltd, Clifton Hill, Victoria, Australia), 20ng of each primer, 1U
Taq or 2.5U Hot Star™ DNA polymerase (Qiagen, Pty Ltd, Clifton Hill, Victoria,
Australia), and 25µM dNTPs. The PCR primer sequences and cycling conditions for each exon were as described (10). Sequencing was done using the PCR primers and the ABI
PRISM Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer -
Applied Biosystems Incorporated, Forster City, CA). The sequencing reactions were run on an ABI377 DNA Analyzer. The chromatograms were edited and aligned using “MT
Sequence Navigator” (Perkin Elmer - Applied Biosystems Incorporated) and "Sequencher"
(Genome Corporation) and compared with the published NDRG1 mRNA sequence
(Accession number XM_005243) and the genomic sequence in contig NT_008150 available from Genbank, (http://www.ncbi.nlm.nih.gov/Genbank/).
Sequence variants were named according to established recommendations (81), with positive IVS (intervening sequence) numbers indicating intronic SNPs located downstream from the G of the invariant GT donor-site, and negative IVS numbers
21 indicating those located upstream from the G of the invariant AG acceptor-site. Nucleotide positions in the chromosome 8 contig NT_008150 correspond to sequence gi 22049640.
The possible role of sequence variants as disease-causing mutations was assessed on the basis of predicted effect on the protein, family segregation, and information on the known sequence variation in the NDRG1 gene published in the SNP databases at NCBI (available at http://www.ncbi.nlm.nih.gov/SNP/) and The Human Genome Variation Base database
(HGVbase) (available at http://hgvbase.cgb.ki.se/) .
2.2.3 Determining the effects on mRNA splicing - RNA isolation and cDNA synthesis
Possible effects on RNA splicing were investigated for sequence variants located within or in close proximity to splice sites, namely IVS8-1G>A, IVS11-5C>T, IVS2-
6T>C, and IVS11+10T>C. The analysis was done on RNA obtained from total blood.
Blood samples were taken in the storage buffer provided in the PAXgene Blood RNA Kit
(Qiagen) and RNA extracted according to the manufacturer’s protocol. Reverse transcription was done with the Superscript Preamplification System for First Strand cDNA Synthesis Kit (Life Technologies, Rockville, MD). The primers used for the subsequent PCR reactions are shown in Table 2.2.
The reactions were performed in a 50µL volume, using 2µL of a 1:3 dilution of the cDNA, 20 ng each of the PCR primers, 1x Qiagen PCR buffer, 1.5mM MgCl2, 25µM dNTPs and 2.5U Taq DNA polymerase (Qiagen). The PCR was performed on a Perkin
Elmer 2400 thermal cycler (Perkin Elmer - Applied Biosystems Incorporated) under the following conditions: 95°C for 5 mins, followed by 35 cycles at 95°C for 15 sec, 60°C for
1 min and 72°C for 1 min, and final extension at 72°C for 10 mins. The PCR products
22 were sequenced in both directions using the PCR primers. Sequencing and sequence analysis were as described above.
23 Table 2.2. Primers used in the cDNA analysis of sequence variants potentially affecting RNA splicing
Sequence Fully Partially Sequence F Sequence R Product Splice variant variant covered covered size (bp) detected exons exons
IVS2-6 T>C E2-E3 E1 , E4 ggcaccgtcctttccttt acagcgtgacgtgaacagag 195 None
IVS8-1G>A E6-E14 E5, E15 gaggacatgcaggagatcac ccagaggctgtgcgggacc 754 Exon 9 skipped
IVS11-5 C>T, E11-E14 E10, E15 ctgcacctgttcatcaatgc ccctgcacgaagtacttgaa 263 None IVS11+10 T>C
24 2.2.4 Calculating nucleotide diversity in the NDRG1 gene
Nucleotide diversity was calculated for the NDRG1 coding sequence (1182bp), and the flanking non-coding intronic sequences (3429bp) and part of the 3' UTR
(147bp). All sequence variants detected in this study were included in the analysis.
Calculations were performed using the software package DnaSP (Ver.3.52) (available at http://www.ub.es/dnasp/ ) (82).
2.2.5 In silico analyses of the mutant NDRG1 protein
The full-length peptide sequence of both wild-type NDRG1 (Genbank:
NP_006087) and mutant NDRG1 missing the 19 residues encoded by exon 9, were submitted for secondary-structural analysis using the PSIpred program (available at http://bioinf.cs.ucl.ac.uk/psipred/psiform.html) (83) and PHDsec (available at http://cubic.bioc.columbia.edu/predictprotein/) (84). Surface accessibility predictions were performed using the Scansite program (available at http://scansite.mit.edu/motifscan_seq.phtml) (85).
Protein motif and domain searches were carried out using the Prosite database
(available at http://au.expasy.org/prosite/) (86). The Pfam- Protein Families Database of
Alignments (available at http://www.sanger.ac.uk/Software/Pfam/index.shtml) (87) was used for abhydrolase family information [accession number PF00561].
25 2.3 RESULTS
2.3.1 The identification of disease-causing mutations in NDRG1
Analysis of the coding sequence of NDRG1 in 104 unrelated individuals affected by hereditary motor and sensory neuropathies identified two disease-causing mutations, R148X and IVS8-1G>A (g.2290787G>A).
R148X
The truncating mutation described previously as the founder mutation causing
HMSNL, was also detected in this study. Two female patients, one residing in Bulgaria and the other in France, were found to be homozygous for R148X. Both suffered from severe demyelinating CMT disease with early onset and associated deafness. Upon re- examination of the family history, it was established that the patients belonged to
Romani families with multiple cases of HMSNL.
IVS8-1 G>A (g.2290787G>A)
A G>A transition at the canonical AG splice-acceptor site in IVS8 was found to occur in the homozygous state (Figure 2.1) in a patient affected by severe demyelinating neuropathy. This finding was confirmed in two independent assays, one of which was performed on DNA extracted from a newly obtained blood sample. The same nucleotide substitution was detected in the heterozygous state in both parents of the affected individual. The parents were unaware of any relatedness, however both originated from the same small village in the southern part of Bulgaria. The patient is a
27-year old male, born from an uneventful pregnancy. Onset of the disease was at age 7 years, with gait disturbances, distal weakness and numbness in the lower limbs.
Orthopaedic correction of both feet was undertaken at age 9 years. Involvement of the upper limbs became apparent at age 10 years, and hearing loss – at 20 years. Myopia was present since early childhood. Bowel dysfunction developed in the last 12 months.
Recent neurological examination revealed areflexia in the upper and lower limbs,
26 Figure 2.1. The IVS8-1G>A mutation results in exon 9 skipping The G >A substitution at the invariant AG acceptor splice site in IVS8 of NDRG1 results in an mRNA species that lacks exon 9. A) Genomic DNA sequence of part of IVS8 and exon 9: the mutant IVS8-1 G >A allele was present in the homozygous state in the patient, and both parents were found to be carriers of this rare novel mutation. B) Sequencing analysis of the aberrant transcript resulting from the IVS8-1G >A (g.2290787G>A) mutation, showing the skipping of exon 9.
27 markedly reduced distal muscle strength (distal quadriplegia) and reduced proximal muscle strength in the lower limbs. Sensation was reduced in the limbs: all modalities were affected, with particularly severe disturbances of the sense of joint position in the lower limbs. Intentional tremor and peripheral ataxia were present.
The hearing loss was profound. Severe deformities of the spine (scoliosis) and of the feet and hands were observed. Electrophysiological studies, performed at age 9 years, showed severely reduced nerve conduction: motor nerve conduction velocity was
10.3 m/sec in the ulnar nerve, and unmeasurable in the fibular nerve.
The effect of the mutation on the splicing of the primary NDRG1 transcript was tested using RNA extracted from white blood cells. Reverse transcription and PCR amplification were performed to generate a cDNA product spanning the coding sequence from codon 79 in exon 5 to codon 330 in exon 15. Sequencing analysis of the cDNA from the affected individual and from a control revealed that exon 9 was missing in the patient’s sample (Figure 2.1B). The IVS8-1G>A (g.2290787G>A) mutation thus results in exon 9 skipping, with a preservation of the reading frame in the resultant transcript.
2.3.2 Single nucleotide polymorphisms (SNPs) in NDRG1
The NCBI SNP database lists six variants in the NDRG1 coding sequence, three missense (H41R, M67V and M111L), and three synonymous (G102, A169 and P293).
Neither these, nor any other variation in the coding sequence (apart from the pathogenic mutation R148X) were detected in this study. The analysis of the flanking intronic sequences of NDRG1 identified a total of 12 single nucleotide polymorphisms (Table
2.3) database searches indicated that 5 of the SNPs detected in this study are novel.
These have been submitted to the HGVbase database. The diversity of our sample, and the small number of individuals from some of the ethnic groups made it impossible to
28 estimate allele frequencies in different populations. Therefore the frequency of each
variant (Table 2.3) refers to the overall Caucasian sample.
Table 2.3. Single nucleotide polymorphisms in NDRG1 in 104 unrelated Caucasian individuals with CMT disease
SNP Genotype frequencies in % Sequence variationa reference #
IVS1+48C>T (g.2333426C>T) b SNP001745694 C/C = 96.2 C/T = 3.8 T/T = 0
IVS2-6T>C (g.2316464T>C) rs2272653 T/T = 20.2 T/C = 48.1 C/C = 31.7
IVS5+64G>C (g.2298174G>C) rs2233324 G/G = 2.2 G/C = 2.2 C/C = 95.6
IVS10+77G>A (g.2286554G>A) b SNP001745695 G/G = 98.1 G/A = 1.9 A/A = 0
IVS10+84A>G (g.2286547A>G) b SNP001745696 A/A = 83.7 A/G = 14.4 G/G = 1.9
IVS10-51A>C (g.2285013A>C) rs2233335 A/A = 72.1 A/C = 25.0 C/C = 2.9
IVS11+10T>C (g.2284896T>C) rs2233336 T/T = 88.5 T/C = 11.5 C/C = 0
IVS11-5C>T (g.2284122C>T) rs2227262 C/C = 81.7 C/T = 12.5 T/T = 5.8
IVS13+148T>C (g.2282659T>C) rs2233339 T/T = 75.0 T/C = 18.3 C/C = 6.7
IVS14+39C>T (g.2280507C>T) rs2233341 C/C = 91.3 C/T = 7.7 T/T = 1.0
IVS15-90G>A (g.2275400G>A) b SNP001745697 G/G = 99.0 G/A = 0 A/A = 1.0
IVS15-91C>T (g.2275401C>T) b SNP001745698 C/C = 99.0 C/T = 1.0 T/T = 0
a Nucleotide numbering according to Contig NT_008150.10 (gi number 22049640). b New SNP detected in this study with the designated HGVbase reference number
Three SNPs, IVS2-6T>C (rs2272653), IVS11-5C>T (rs2227262), and
IVS11+10T>C (rs2233336) were in close proximity to splicing sites. A role as disease-
causing mutations appeared unlikely based on the fact that all three have been registered
in the public SNP databases and found to occur at high frequency (Table 2.3).
29 Predictably, sequencing analysis of the cDNA fragments, obtained from blood cell
RNA, failed to produce evidence of aberrant splicing.
One variant, IVS15-90G>A (g.2275400G>A), was identified in the homozygous state in a Spanish patient and not found in any other individual. RNA from that patient was not available for analysis and thus no conclusive evidence against its role in CMT disease could be established. A pathogenic effect appears improbable as the nucleotide substitution is located deep in the intron and the variant sequence, GCACATAAA, bears no resemblance to a cryptic splicing signal.
2.3.3 Nucleotide diversity in the NDRG1 gene
The sequence information derived from the analysis of NDRG1 was used to calculate nucleotide diversity π (Table 2.4). The lack of variation in the coding region of NDRG1, with only one sequence variant detected, resulted in a SNP frequency of 1 per 1185 coding nucleotides and a low value for this region (0.00006).
In the non-coding regions of NDRG1, nucleotide diversity was markedly higher, with a SNP frequency of 1 per 275 non-coding nucleotides and a π value of 0.00066.
The majority of non-coding SNPs (75%) were located deep in introns. Transitions accounted for 86% of all SNPs detected, with a 2-fold bias towards C>T and G>A mutation events. Transversions were rarer, with only one G>C and one A>C change identified.
30 Table 2.4. Nucleotide diversity in NDRG1
Genomic No. of Variants Frequency Nucleotide Std. region nucleotides detected (SNP/bp) diversity +/- ()
Coding 1182 1 1/1182 0.00006 0.00003
Non-coding 3576 13 1/275 0.00066 0.00003
Total 4758 14 1/339 - -
2.3.4 In silico analysis – effects of the IVS8-1G>A mutation on protein structure
To determine the effect of the deletion of the 19 amino acid residues encoded by exon 9 upon the predicted properties of the NDRG1 protein, in silico analyses including secondary-structure predictions, as well as motif and domain identification and comparisons with the wild-type NDRG1 protein were performed. A consistent finding from the secondary-structure predictions (PHDsec and PSIpred) is that the deletion of residues 181-198 (encoded by exon 9) results in the shortening of a helix normally present in the wild-type protein (Figure 2.2A). An extended coil structure is added to this shortened helix in the mutant NDRG1 protein (Figure 2.2B).
31 A Wild-type NDRG1
exon 8 exon 9 exon 10 B Mutant NDRG1
exon 8 exon 10
L
H - helix C - coil Confidence of prediction
Figure 2.2. Prediction of the secondary structure of A) wild-type NDRG1 protein and B) mutant NDRG1 protein missing exon 9 as a result of the splicing IVS8- 1G>A mutation. Skipping of exon 9 results in the loss of 19 amino acid residues which, in the wild-type NDRG1 protein, are predicted to form a helix. Deletion of these residues results in an extended coil (highlighted in red box) encoded by the last residue of exon 8 and the first 5 residues of exon 10. Analysis of the full-length peptide sequences of both wildtype NDRG1 and mutant NDRG1 was performed by PSIpred software (available at http://bioinf.cs.ucl.ac.uk/psipred/psiform.html).
32 The shortening of the helix and the addition of a coil structure may lead to significant changes in protein folding resulting in changes in accessibility to functional domains. This is supported by surface accessibility predictions performed by the
Scansite program which shows that the helix encoded by exon 9 may be exposed at the outer surface of the protein (Figure 2.3). This predicted change may have functional significance, particularly if exon 9 encodes a domain whose surface accessibility is necessary for protein interaction or post-translational modification.
A Wild-type NDRG1
B Mutant NDRG1
Figure 2.3. Surface accessibility predictions of residues of A) wildtype NDRG1 protein and B) mutant NDRG1 protein missing exon 9. Scores above 1.0 indicate residues have a high probability of facing the outer surface of the protein. Arrow denotes region predicted to be exposed on the outside of the wild- type NDRG1 protein does not occur when exon 9 is deleted. Predictions were performed by Scansite (available at http://scansite.mit.edu/motifscan_seq.phtml) .
33 To determine whether exon 9 encodes any functionally important regions, in silico domain and motif searches using the full-length wild-type NDRG1 were performed. Database searches revealed that exon 9 falls within the large region of
NDRG1 (residues 89-309) identified by Pfam as an alpha/beta hydrolase (abhydrolase) fold. When submitting the mutant NDRG1 sequence (missing residues encoded by exon
9) for analysis, the mutant protein was still classified as an abhydrolase. Sequence alignment showed that exon 9 does not fall within any of the 5 motifs that confer membership to the abhydrolase superfamily (54,59).
34 2.4 Discussion
Mutations in NDRG1 are a rare cause of peripheral neuropathy
NDRG1 is one of an ever-growing list of genes implicated in the molecular pathogenesis of CMT disease and related disorders. The problem of genetic heterogeneity is compounded by the fact that different mutations in the same gene can be inherited as autosomal dominant or autosomal recessive traits, and also can result in different phenotypes. For example, mutations in the MPZ and EGR2 genes can cause autosomal dominant, as well as recessive, forms of CMT disease (88-91), and mutations in Gangliosidase-Induced Differentiation-Associated Protein 1 (GDAP1) and MPZ have been found to occur in both demyelinating and axonal neuropathies (92-94). These previously reported findings, together with the phenotypic features of HMSNL involving both Schwann cells and axons, led us to adopt an inclusive approach to the selection of affected individuals for this study and examine both demyelinating and axonal forms of CMT disease of autosomal recessive, as well as dominant, inheritance.
The majority of affected subjects were sporadic cases where the common genetic causes of peripheral neuropathy had been excluded, an important category which presents diagnostic difficulties and which is also likely to contain a high proportion of autosomal recessive forms.
The analysis of the overall sample of 104 CMT patients identified three subjects with mutations in NDRG1 (2.88%), two of whom were unrecognised cases of HMSNL.
In the group of demyelinating CMT disease, NDRG1 mutations accounted for 4.68% of patients. The contribution of NDRG1 to the molecular etiology of CMT disease thus appears to be modest, similar to that reported for other genes, such as Early Growth
Response Factor 2 (EGR2), Periaxin (PRX), and Myotubularin-Related protein 2
(MTMR2) (78,95-97). This makes the molecular diagnosis of CMT disease a difficult task in about one third of the patients, where the common causes, such as PMP22
35 duplications, and PMP22, Cx32 and MPZ mutations (78,98) have been excluded. At the same time, one should emphasise that R148X is the single most common cause of
CMT disease in individuals of Romani descent. Carrier rates for this mutation can be as high as 16% in some Romani groups (10) and testing for its presence may be advisable in Caucasian patients of unknown ethnicity. The IVS8-1G>A (g.2290787G>A) is the second disease–causing NDRG1 mutation reported to-date. The two known mutations are very similar in terms of mode of inheritance, as well as resulting clinical phenotype
– demyelinating neuropathy with early onset, rapid progression, severe early reduction in nerve conduction velocity, and associated deafness, suggesting that NDRG1 mutation screening may be most appropriate in patients with the above features.
Lack of variation in the NDRG1 coding region
The high degree of evolutionary conservation of NDRG1, from sunflower to primates, points to the functional importance of the protein. This is reflected in the low diversity of the coding region observed in this study. The SNPs in the coding sequence of NDRG1, registered in the public databases, are synonymous or result in conservative amino acid substitutions, in support of selective pressure contributing to the conservation of NDRG1. It should be noted that these are not validated SNPs and the fact that none were found in our analysis of more than 200 chromosomes suggests that they are very rare variants, and some may be artifacts. The observed nucleotide diversity and SNP frequency in the intronic regions of NDRG1 is consistent with published data for non-coding genomic regions (99). So too, is the bias in favour of transitions observed over large parts of the genome (100).
36 Characterisation of the novel NDRG1 mutation IVS8-1G>A
The mechanism whereby exon 9 skipping results in NDRG1 loss of function is difficult to explain given the limited knowledge of the wild-type protein. Exon 9 falls within the large region of NDRG1, between amino acids 89 to 309, classified by Pfam as an alpha/beta hydrolase (abhydrolase) fold. The inclusion of NDRG1 in the superfamily of abhydrolases, enzymes with similarity in the tertiary structure but with diverse substrate specificity, should be regarded as putative. At present, there is no functional information suggesting that NDRG1 acts as an enzyme as it lacks the conserved serine and histidine residues at the characteristic positions forming the active abhydrolase triad (59). The missing exon does not overlap with any of the five motifs shared by members of this protein family, and the mutant protein is still classified as an abhydrolase by Pfam analysis.
It is highly likely that the IVS8-1G>A mutation and resulting deletion of exon 9 in the protein leads to misfolding. This is supported by the secondary structure predictions (84) which suggest that amino acids encoded by exon 9 form a helical structure, therefore the absence of this region could have profound effects on tertiary folding and subsequently the orientation of domains necessary for protein interactions or post-translational modifications. This may perhaps be reflected in the surface accessibility predictions, which show that this helix lies on the outer surface of the wild- type NDRG1 protein. The misfolding of NDRG1 could lead to mislocalisation or aggregation and degradation of the mutant protein via classical protein degradation pathways or to the disruption of protein interactions.
Although both NDRG1 mutations, R148X and IVS8-1G>A, result in a similar clinical phenotype the underlying cellular mechanisms of dysfunction may be very different according to the effect each mutation has on the encoded NDRG1 protein. In the case of R148X the dysfunction is assumed to result from the effects of a severely
37 truncated protein or more likely through its complete absence, arising from the nonsense-mediated decay of the mRNA transcript. The IVS8-1G>A mutation on the other hand, would result in a protein that is only 19 residues smaller than the wild-type
NDRG1 and may have different functional properties and localisation fates compared to its wild-type counterpart. Ultrastructural comparisons of nerve biopsies from patients with the R148X mutation and from the patient with the IVS8-1G>A mutation could perhaps provide important information as to whether the similarities in clinical phenotype caused by both mutations are shared at the molecular level, for instance does
IVS8-1G>A result in similar Schwann cell and axonal pathology as that noted in
HMSNL? (27,101). The feasibility and practicalities of investigations involving a private mutation like IVS8-1G>A however, makes such studies difficult. Nonetheless the IVS8-1G>A mutation and resulting protein provide an interesting model for future in vitro functional investigations and comparisons with the wild-type NDRG1 protein.
Summary and future directions in determining NDRG1 function and its role in CMT disease
Understanding the molecular pathogenesis of peripheral neuropathies caused by mutations in NDRG1 depends on knowledge of the physiological role of the protein in the Schwann cell and in Schwann cell-axonal interactions. The existing hypotheses on
NDRG1 function in tumour suppression and response to cellular stress are derived from studies of non-neural tissues. The fact that the only known clinical phenotypes resulting from NDRG1 mutations are restricted to the peripheral nervous system points to a crucial and unique role in this tissue. The characterisation of NDRG1-interacting protein partners may provide an insight into its function in peripheral nerve and facilitate the understanding of its protein domains and their roles. It is the search for these interacting protein partners that is presented in the following chapter.
38 CHAPTER 3 – THE IDENTIFICATION OF NDRG1 PROTEIN-PARTNERS BY YEAST TWO-HYBRID LIBRARY SCREENING
3.1 Background
The yeast two-hybrid method (102) is a powerful technique that is used to identify protein interactions. It is particularly useful for assigning proteins with unknown function, such as NDRG1, to specific pathways through “guilt-by-association” with proteins whose function is known. In this method, one cDNA construct, the "bait" molecule, is fused to a
DNA-binding domain (BD) (e.g. E. coli LexA protein or yeast GAL4 protein) and is used to screen a library of "prey" cDNAs, fused to an activation domain (AD). The constructs are expressed in yeast and if the proteins interact, a bipartite transcription factor is reconstituted and can activate reporter genes. Interaction is assayed by nutritional selection
(for ADE2 and HIS3) and colorimetric assays detecting α-galactosidase or β-galactosidase activity encoded by the MEL1 or LacZ genes respectively. (Figure 3.1).
Isolation and sequencing of the plasmid encoding the interacting protein can then be performed to establish identity. If the identified protein partner(s) belong to known or well- characterised pathways, clues to the biological relevance of the interaction can be inferred and hypotheses formed. Interesting candidates can be subjected to complementary confirmation experiments, such as in-vitro binding assays, co-immunoprecipitation, bioluminescence resonance energy transfer (BRET) assays, and confocal co-localisation tests to ensure the interaction is biologically relevant and not an artifact of the yeast two- hybrid method. Ultimately, assays can be developed to show the functional significance of the interaction.
39 A Prey fused to AD AD
Bait fused to BD Reporter Genes B
Promoter HIS3 ADE2 LacZ MEL1
C AD BD Promoter HIS3 ADE2 LacZ MEL1
Figure 3.1. Principles of the yeast two-hybrid method A) The prey protein is fused to the activation domain (AD) and B) the bait protein is fused to the binding domain (BD) and binds to the yeast promoter. C) If the two proteins interact, transcription of reporter genes under the control of the promoter proceeds. Interaction is indicated by growth on selective media for HIS3 and ADE2, and by colorimetric assays for LacZ and Mel1. (Model is based on yeast strain AH109 and yeast two-hybrid plasmids pGBKT7-BD and pGADT7-AD). For clarity, a single promoter for all reporter genes is depicted.
40 In recent years, yeast two-hybrid technology and the diversity of library availability and construction methods have improved. This has made it possible to screen three cDNA libraries throughout the duration of this PhD project in the search for NDRG1 protein- partners. The screening of a murine pluripotent cell line (EML C.1) library (provided by
Dr. Evan Ingley, WAIMR and carried out by Dr Dora Angelicheva, WAIMR) using a full- length human NDRG1-LexA/BD construct as “bait” identified a number of NDRG1- interacting proteins listed in Table 3.1.
Table 3.1. NDRG1-LexBD interactors detected by yeast two-hybrid screening of a murine pluripotent cell line (EML C.1)
Protein Function
BCL2/Adenovirus E1B 19kda Mitochondrial protein, induces apoptosis, may protein-interacting protein 3 function as a tumor suppressor.
ARL-6 interacting protein-1 Localised to intracytoplasmic membranes and involved in protein transport and membrane trafficking. Belongs to Ras family.
Herpud1 Ubiquitin-like membrane protein induced by ER stress and involved in the retro-translocation of misfolded proteins.
Scotin Localised to ER and nuclear membrane. Induces apoptosis.
The functional categories to which these candidates belong are consistent with previous findings suggesting NDRG1 involvement in apoptosis (36,46) and the response to cellular stress (35,40,52). At the same time the identification of proteins like ARL-6 interacting protein-1 and Herpud1 implicated NDRG1 in trafficking. Originally identified as a gene upregulated alongside NDRG1 in response to homocysteine (35), Herpud1 has recently shown to be part of a protein complex proposed to be involved in the 41 ubiquitylation and retro-translocation of misfolded proteins from the ER to the cytoplasm
(103) and ARL-6 interacting protein-1 may function in protein transport and membrane trafficking (104).
In order to obtain information about the role of NDRG1 in nervous tissue and gain clues to the neuropathology underlying CMT4D we searched for NDRG1-interacting proteins in nerve tissue cDNA libraries. At the time, the only available yeast two-hybrid library representative of this tissue source was derived from human fetal brain. While CNS pathology is not a feature of CMT4D and the expression of the NDRG1 protein in human brain is yet to be characterised, this library was screened on the basis that similarities in pathways and cellular processes between neurons, myelin-producing and supporting cells in the CNS and PNS may result in the identification of functionally relevant interactors.
Closely following this screen, there was a major commercial development by
Clontech which facilitated yeast two-hybrid library production, allowing us to produce and screen an adult mouse peripheral nerve cDNA library in the search for NDRG1 protein- partners. Clearly, this library source was more relevant to the CMT4D phenotype and could reveal tissue-specific NDRG1 interactors. An overview of the two library screens utilized in these investigations is presented in Figure 3.2 and the results of the search for NDRG1 interactors identified from both library sources is presented in the following sections.
42 A B
Human fetal brain Adult mouse sciatic nerve library screen library screen
LexA-based Y2H system Gal4-based Y2H system Reporter genes Reporter genes HIS3/LacZ HIS3/ADE2/LacZ/MEL1
Bait Bait Prey
Library double-stranded NDRG1 in NDRG1 in cDNA (dscDNA) pBTM116-BD pGBKT7-BD
Linearised pGADT7-rec
Co-transform into Prey yeast strain AH109 Transform bait into yeast strain L40
Transform library Library in into yeast strain pACT2-AD Library dscDNA homologously recombined with L40/pBTM116-NDRG1 pGADT7-rec, simultaneous screening with NDRG1-Gal4BD
Plate on media lacking Tryptophan, Plate on TDO media lacking Plate on QDO media lacking Leucine, Histidine to select for HIS3 Tryptophan, Leucine, Tryptophan, Leucine, reporter gene activation Histidine Histidine & Adenine (MEDIUM STRINGENCY) (HIGH STRINGENCY)
Filter lift β -galactosidase assay to test Filter lift β-galactosidase assay to test for for LacZ reporter gene activation LacZ reporter gene activation and/or Liquid α-galactosidase assay to test for MEL1 reporter gene activation
Isolate Library Plasmids → Sequence, Re-transform favourite candidates into yeast and re-test for interaction. Clone full-length constructs and re-test for interaction
Confirm interaction by alternative assays (Co-immunoprecipitation, Pulldown assays, BRET, In-vivo co-localisation)
Figure 3.2. Schematic overview of the yeast two-hybrid systems utilized in the screening for NDRG1-interacting proteins A) a human fetal brain library and B) a mouse sciatic nerve library. Binding domain plasmids pBTM116 and pGBKT7 carry the TRP selection marker. Activation domain plasmids pACT2 and pGADT7-rec carry the LEU selection marker. Transformants carrying both plasmids are selected for on media lacking Trp and Leu with positive interaction indicated by colony growth in the absence of histidine and/or adenine. Colorimetric assays can detect the activity of an additional reporter gene product, β-galactosidase encoded by LacZ in L40 and AH109 and secreted α- galactosidase encoded by the the MEL1 gene in AH109. 43 3.2 Methods
3.2.1 Yeast two-hybrid “bait” constructs
The complete human NDRG1 reading frame (1182nt / 394aa) and truncation mutants encompassing the predicted functional domains of the protein, were PCR amplified from IMAGE clone #3181396 (The I.M.A.G.E Consortium) using sequence-specific primers appended with EcoRI (5') and BamHI (3') restriction sites according to conditions detailed in Figure 3.3. PCR products were electrophoresed on agarose gels containing ethidium bromide and visualised using UV illumination. Images were captured using a
Kodak Digital Camera and KDC software. PCR products were double-digested with EcoRI and BamHI restriction enzymes (New England Biolabs). Each digest reaction contained 1 x
EcoRI restriction buffer, 1% BSA and 1 unit of each enzyme in a total volume of 50µL.
The reactions were incubated at 37°C for 3-6 hours and gel-purified using QG buffer
(Qiagen) and Qiagen PCR Purification columns.
The yeast two-hybrid Binding Domain (BD) vectors, pBTM116-BD for screening of the human fetal brain library (a gift from the Laboratory for Cancer Medicine, WAIMR), and pGBKT7-BD (Clontech) for screening of the mouse peripheral nerve library, were prepared for ligation in a similar manner as the PCR inserts, except that each was dephosphorylated after restriction digestion using 1 unit of shrimp alkaline phosphatase
(SAP) and incubated at 37°C for 1 hour, followed by heat inactivation of the SAP at 65°C for 20 minutes. Digested products were ligated into cloning sites of the SAP-treated yeast two-hybrid vectors using the T4 Rapid DNA Ligation kit (Roche) or with T4 ligase
(Promega) according to manufacturers instructions.
44 A B NDRG1-BD Constructs PCR Forward Primer Reverse Primer Conditions
NDRG1 (aa 1-394)-BD 1 394 ggaattcatgtctcgggagatgcagg cgggatcccctagcaggagacctcca TD68-61ºC
NDRG1 IVS8-1G>A-BD 1 ggaattcatgtctcgggagatgcagg cgggatcccctagcaggagacctcca TD68-61ºC
ggaattcatgtctcgggagatgcagg cggatccgcatgtatcccatgccctgc TD68-61ºC NDRG1 (aa 1-316)-BD 1 316 ggaattcatgtctcgggagatgcagg cggatccaggcgcctgctcctgttcc TD68-61ºC NDRG1 (aa 1-146)-BD 1 146 ggaattccctgtcatcctcacc cggatccaggcgcctgctcctgttcc 55ºC NDRG1 (aa 57-146)-BD 57 146 ggaattcatgcttcctggagtc cggatccaggcgcctgctcctgttcc 55ºC NDRG1 (aa 121-146)-BD 121 146
NDRG1 (aa 146-394)-BD 146 394 ggaattctacaccctaactcgatttgc cgggatcccctagcaggagacctcca TD68-61ºC
NDRG1 (aa146-316)-BD 146 316 ggaattctacaccctaactcgatttgc cggatccgcatgtatcccatgccctgc 55ºC
NDRG1 (aa 316-394)-BD 316 394 ggaattctcggctagcatgacccgcctg cgggatcccctagcaggagacctcca 55ºC Legend Esterase/lipase/thioesterase active site Phospopantetheine binding domain Tri-Deca repeat motif
Figure 3.3. NDRG1 yeast two-hybrid constructs, domains, primer sequences and PCR conditions. A) Schematic of NDRG1-BD constructs. Full-length NDRG1 construct – NDRG1(aa1-394)-BD, was used for library screening. All other constructs were used in domain mapping experiments. Numbers indicate amino acid codons. Shaded regions denote predicted NDRG1 protein domains as indicated. B) Primer sequences and PCR conditions used to amplify inserts for cloning into the pGBKT7-BD or pBTM116-BD vectors.
45 3.2.2 Preparation of competent bacterial cells
Ligation reactions were transformed into chemically competent Eschericia coli
DH5-α cells. The cells were prepared by inoculating 100µL of stock cell culture into 10mL of Luria Broth (LB) and incubating at 37°C with shaking for approximately 2 hours.
Optical density (OD) was measured on a Beckmann DU3 spectrophotometer and cells were harvested upon reaching an OD660 of 0.6-1.0. Cells were cooled on ice and pelleted for 5 minutes at 2000xg at 4°C. The supernatant was discarded and the pellet was resuspended in
5mL of ice-cold 0.1M calcium chloride (CaCl2) and left on ice for 40 minutes. Cell pellets were collected by centrifugation, resuspended in 1mL 0.1M CaCl2 / 20% Glycerol, and aliquoted in separate 1.5mL pre-chilled Eppendorf tubes to a final volume of 50µL.
Competent cells were used immediately for transformation and unused aliquots stored at –
70°C.
3.2.3 Transformation of competent bacterial cells
Typically, 5µL of ligation mix was added to 50µL of competent cells, mixed gently and left on ice for 30 minutes. Transformation mixes were heat-shocked at 42°C in a water bath for 90 seconds and chilled on ice for 5 minutes. 450µL of Luria Broth or SOC medium were added and cells were incubated at 37°C with vigorous shaking for 1 hour. Cells were collected by centrifugation at 2500xg for 20 seconds, the supernatant was removed and the pellet was resuspended in 70-100µL of sterile Tris-EDTA (T.E) buffer. Cells were plated out on MaConkey agar (Difco) or Luria Broth agar plates, containing an appropriate antibiotic (100µg/mL ampicillin or 50µg/mL kanamycin) and incubated at 37°C for 16-24 hours.
46 3.2.4 Bacterial colony screening and plasmid isolation
Bacterial colonies were checked for the presence of the correct plasmid insert by
PCR screening. Typically, up to ten different colonies were screened at a time. Half a colony was picked with a sterile pipette tip and resuspended into a PCR reaction mixture containing 1x Qiagen PCR Buffer, 20 µM dNTPS, 2.5mM MgCl2, 20 ng primers, 1 unit of
Taq polymerase and dH2O to a final volume of 12µL. PCR cycling conditions included initial denaturation at 94°C for 5 minutes followed by 35 cycles of 94°C for 30 seconds, annealing at 55°C for 30 seconds and extension at 72°C for 60 seconds, with a final extension step at 72°C for 5 minutes. PCR products were run on agarose gels containing ethidium bromide, and visualised by UV illumination.
Colonies containing plasmids with correct inserts were picked and seeded into 2mL of Luria Broth with selective antibiotic (100µg/mL ampicillin or 50µg/mL kanamycin) and cultures were grown with shaking at 37°C for 14-16 hours. Plasmid DNA was isolated using the Qiagen Mini-Prep Purification kit or the Eppendorf Fast-Plasmid Isolation kit according to manufacturers instructions. Where large concentrations of plasmid were needed, culture volumes were increased and the Qiagen Midi-Prep Plasmid Isolation kits were used. Plasmids were checked by restriction digest to ensure each contained the correct insert as described in section 3.2.1.
To ensure that the insert was in the correct reading frame, plasmids were sequenced using ABI PRISM Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Perkin
Elmer - Applied Biosystems Incorporated, Forster City, CA). The sequencing reactions were run on an ABI377 DNA Analyzer and sequences analysed using Sequence
NavigatorTM software.
47 3.2.5 Yeast two-hybrid library amplification – human fetal brain library
A 20µL aliquot of a human fetal brain yeast two-hybrid library in pACT2-AD pre- transformed in Eschericia coli DH5-α, was seeded into 25mL of Luria Broth with ampicillin and grown with vigorous shaking at 37°C for 8 hours. 250µL aliquots of the culture were spread onto each of 100 large Luria Broth agar plates with ampicillin and incubated overnight at 37°C. Plates were removed from the incubator and refrigerated at
4°C for 30 minutes. 1mL of Luria Broth with ampicillin was added to the plates and colonies were scraped with a sterile glass rod into a large sterile beaker. Isolation of library plasmid DNA was performed using the Qiagen Maxi-Prep kit according to the manufacturer’s instructions.
3.2.6 Yeast two-hybrid library construction – mouse peripheral nerve library
Homologous recombination-mediated library construction was performed in
Saccharomyces cerevisiae strain AH109 with the genotype; [MATa, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4Δ, gal80 Δ, LYS2::GAL1UAS-GAL2TATA–ADE2, URA3::MEL1UAS-
MELTATA-lacZ] using the Matchmaker III Library Construction and Screening Kit
(Clontech). The cDNA library was made from sciatic nerves removed from 4 adult BalbC mice. The nerves were washed in sterile PBS, frozen in liquid nitrogen and pestle-ground in a chilled RNAse-free mortar. The tissue was suspended in 1mL of Triazol reagent (Gibco) and homogenised by passing approximately 20 times through a 22-gauge needle and syringe. RNA was isolated according to the Triazol protocol, and DNAse (Qiagen) treated on RNeasy purification columns (Qiagen). Single stranded cDNA was synthesized from
1µg of RNA using Oligo d(T) and random primers (Clontech) in separate reactions, according to manufacturer’s instructions. The reaction products were combined and a 2µL aliquot was used as a template for long-range PCR to produce double-stranded cDNA 48 (dscDNA). PCR reagents included 1x Advantage 2 PCR Buffer (Clontech), 2mM dNTPs,
20 ng of each linker primer (5'-TTCCACCCAAGCAGTGGTATCAACGCAGAGTGG-3' and 5'-GTATCGATGCCCACCCTCTAGAGGCCGAGGCGGCCGACA-3'), 1x GC-Melt solution (Clontech), 1 unit of Advantage 2 Polymerase Mix (Clontech) and dH2O in a total volume of 100µL. Thermal cycling conditions included denaturation at 95°C for 30 seconds followed by 22 cycles of 95°C for 10 seconds denaturation and a combined extension and annealing step at 68°C for 6 minutes in the first cycle. An additional 5 seconds was added to each successive extension and annealing cycle, followed by a final 5 minute incubation step of 68°C. The reaction products were purified with Chroma-
Spin+TE-400 columns (Clontech) to eliminate DNA molecules shorter than 200 base pairs.
The quality and length of library inserts were checked by running 5µL on 1.5% agarose gels alongside a positive control human fetal brain RNA sample (Clontech) treated in the same manner as outlined (Figure 3.4).
49 1 2 3 4 5 6 7
5.0kb
1.6kb
1.0kb
0.2kb
Figure 3.4. Mouse sciatic nerve yeast two-hybrid library production dscDNA products were run on a 1.5% agarose gel prior to AH109-mediated homologous recombination with the pGADT7-Rec AD yeast two-hybrid library vector. From a starting amount of 1µg of total adult mouse sciatic nerve RNA, two pools of random-hexamer and oligo-d(T) primed cDNA were produced. Subsequent long-range PCR conversion to double-stranded cDNA (dscDNA) linked with primer sequences homologous to regions of the pGADT7-Rec AD vector produced library inserts ranging from 200 base pairs to 1.5kb in length. These inserts were pooled and transformed along with the pGADT7-Rec AD vector into yeast strain AH109 to produce a library consisting of 2.9 x 106 clones. Lane 1) GeneRuler 1kb ladder. Lane 2) Control oligo d(T) dscDNA derived from 1µg of total human fetal brain RNA (Clontech). Lanes 3 & 4) mouse sciatic nerve oligo d(T) dscDNA after 22 PCR cycles. Lane 5 & 6) mouse sciatic nerve random hexamer dscDNA after 22 PCR cycles. Lane 7) Control random hexamer dscDNA derived from 1µg of total human fetal brain RNA (Clontech).
50 3.2.7 Testing of the NDRG1-Gal4BD fusion protein in yeast strain AH109
Prior to library screening, the complete human NDRG1-Gal4BD fusion protein was tested to ensure it did not have any toxic effects when transformed into yeast, display any auto-transcriptional activity of reporter genes, nor interact with the activation domain of the pGADT7-AD fusion protein. This was achieved using a small-scale co-transformation of pGBKT7/NDRG1-BD and the pGADT7-AD vector into LiAc-treated competent AH109 cells. Comparison of the growth-rate of transformation cultures containing the NDRG1-
Gal4BD protein with cultures containing the pGBKT7-BD and pGADT7-AD vectors alone, showed no differences in cell density as measured by optical density. This indicated that the NDRG1-Gal4BD protein did not have a toxic effect when expressed in yeast.
NDRG1-Gal4BD and pGADT7IAD transformants were plated on yeast media lacking the nutritional selectors tryptophan and leucine (-Trp/-Leu) and on yeast TDO (-
Trp/-Leu/-His) and QDO (-Trp/-Leu/-His/-Ade) media. Absence of induction of β- galactosidase activity from transformants grown on (-Trp/-Leu) plates and lack of growth on selective plates (-Trp/-Leu/-His) and (-Trp/-Leu/-His/-Ade) confirmed that the full- length NDRG1-Gal4BD fusion was suitable to use as a “bait” in the yeast two-hybrid assay. Similar tests had previously been performed in our laboratory for the NDRG1-
LexA/BD fusion protein in the L40 yeast strain (not shown), leading to the same conclusions.
3.2.8 Yeast two-hybrid library screening– human fetal brain library
Saccharomyces cerevisiae strain L40 with the genotype [MATa, his3 Δ200, trp1-
901, leu2-3,112ade2, LYS:: (lexAop)4-HIS3, URA3:: (lexAop)8-lacZ, GAL4, gal80 ] was transformed with the pBTM116/NDRG1-BD vector construct using a small-scale transformation protocol developed by Stan Hollenberg of the Fred Hutchinson Cancer 51 Research Center, Seattle USA. For large-scale library screening, a 10mL overnight culture of L40/pBTM116/NDRG1-BD was seeded into 100mL of the yeast selective medium (-
Trp, -Ura) and grown for 16 hours. This culture was used to inoculate 1L of YPAD media pre-warmed to 30°C and grown with vigorous shaking (280rpm) for approximately 3 hours until an optical density of 0.5-0.6 was reached. The cells were pelleted at 2500rpm for 5 minutes and washed in 500mL of sterile TE buffer. The cells were resuspended in 20mL of
100mM LiAc/0.5xT.E and incubated at room temperature for 10 minutes. To these cells were added 1.0mL of 10mg/mL denatured salmon sperm DNA and 500µg of fetal brain library plasmid DNA followed by 140mL of 100mM LiAc/ 40% PEG-3350 (Sigma)/1x TE and incubated at 30°C for 30 minutes. 17.6mL of DMSO was added and the cells were heat-shocked at 42°C for 6 minutes. The cells were rapidly cooled by immediately adding
400mL of YPA and placing the flask in a water bath set at 23°C. Cells were pelleted, washed with 500mL of YPA then resuspended in 1L of YPAD and incubated at 30°C for 1 hour. Cells were pelleted, washed in 500mL of SD (-Ura/-Trp/-Leu), repelleted again and grown with vigorous shaking for 6 hours at 30°C in 1L of SD-Ura/-Trp/-Leu. Cells were recovered, washed twice in SD-Trp/-His/-Ura/-Leu/-Lys and resuspended in a final volume of 10mL. 500ul of the transformed cells were spread on each of 20 large plates containing
SD (-Trp/-His/-Ura/-Leu/-Lys) media and incubated at 30°C. Colonies were picked over a period of 5-14 days.
3.2.9 Yeast two-hybrid library screening – adult mouse sciatic nerve
The AH109 strain was co-transformed with pGBKT7/NDRG1-BD, linearized pGADT7-Rec AD and mouse sciatic nerve dscDNA using the LiAc method as outlined in the Clontech MM-III handbook. The transformants were plated on 20 large plates containing yeast triple-dropout medium (TDO) (-Trp/-Leu/-His) to select for medium 52 stringency interactions, and on 20 large plates containing yeast quadruple-dropout medium
(QDO) (-Trp/-Leu/-His/-Ade), to select for high stringency interactions. Plates were incubated at 30°C and colonies were picked over a period of 5-9 days.
3.2.10 α-galactosidase (MEL1) and β-galactosidase (LacZ) assays
L40 and AH109 yeast transformants were tested for activation of the colorimetric reporter gene LacZ in the presence of the β-galactosidase substrate, X-gal, using a filter-lift assay. Colonies were transferred onto sterile nitrocellulose filters and cells were ruptured by brief immersion in LiN2. Filters were left to thaw then placed in a Petri dish lined with a sheet of Whatman paper immersed in Z-buffer (60mM Na2HPO4, 40mM NaH2PO4, 10mM
KCl, 1mM MgSO4, 100µM β-mercaptoethanol, pH7.0) containing X-gal (Sigma). The filters were incubated at 30°C and colour development was periodically checked over a 24 hour period.
Liquid α-galactosidase assays (105) to detect MEL1 activation in AH109 transformants were performed according to instructions detailed in the Yeast Protocols handbook (Clontech PT3024-1).
3.2.11 Recovery of yeast plasmids
Library plasmids were recovered from yeast colonies as detailed (106) and propagated in Eschericia coli DH5-α as described in section 3.2.3. Bacterial plasmid DNA was isolated using Qiagen Miniprep columns. In cases where yeast plasmid isolation yield was insufficient for shuttling to DH5-α, up to 5µL of the yeast plasmid preparation was used as a template for PCR using 20ng each of the library screening primers 5'-
ATGAAGATACCCCACCAAACC-3' and 5'-CGGGGTTTTTCAGTATCTACG-3'
(Clontech), 1x PCR buffer (Qiagen), 1.5mM MgCl2, 2mM dNTPs, 1 unit of Taq 53 polymerase and dH20 in a total volume of 50µL. Typical PCR conditions included initial denaturation at 95°C followed by 40 cycles of denaturation at 95°C for 30 seconds, annealing at 55°C for 30 seconds and extension at 72°C for 45 seconds with a final extension step at 72°C for 7 minutes. In some cases, PCR amplification from the yeast plasmid preparations revealed multiple inserts and, where possible, these were cut from the gel, purified and sequenced. Interesting candidates were re-ligated into the activation domain vectors, transformed back into yeast and re-tested for interaction with the NDRG1-
BD fusion protein. In some cases attempts to isolate colonies containing single interacting library plasmids only were performed by streaking out colonies on selective media for 3 or more generations.
3.2.12 Establishing clone identity
All clones were sequenced as detailed in section 3.2.4 and identity was determined using the nBLAST and pBLAST programs at NCBI (available at http://www.ncbi.nlm.nih.gov). Protein domains were analysed using Pfam- Protein
Families Database of Alignments (available at http://www.sanger.ac.uk/Software/Pfam/index.html).
3.2.13 Cloning full-length constructs
Human homologues and full-length cDNAs of interacting clones were amplified from human pooled cDNA (Clontech) or mouse sciatic nerve cDNA. The full length and truncation constructs used for the domain-mapping, primer sequences and PCR conditions are shown in figures 3.5 to 3.8. All PCR products were cloned into yeast-two hybrid vectors as detailed in sections 3.2.1 to 3.2.4.
54 A B
Human PRA1 Forward Primer Reverse Primer PCR Conditions 1 185 ggaattcatggcagcgcagaag ggatcctcacacgggttccatc 55ºC
Mouse Pra1
1 185 ggaattcatggcggcccagaag ggatccttacacaggttccatc 55ºC
1 165 ggaattcatggcggcccagaag ggatccttaggagcctatgagtac 55ºC
1 112 ggaattcatggcggcccagaag ggatccttacagatagagaatg 55ºC
79 185 ggaattcgtgttcgtgtttctc ggatccttacacaggttccatc 55ºC
132 185 ggaattcgccctggctggtgg ggatccttacacaggttccatc 55ºC Legend Cytoplasmic domains Transmembrane domains Linker domain
Figure 3.5. Prenylated Rab Acceptor 1 yeast two-hybrid constructs, domains, primer sequences and PCR conditions. A) Schematic of full-length human PRA1 and mouse Pra1 constructs. Domain organization as proposed in (107). B) Primers and PCR conditions.
55 A B Human APOA1 Forward Primer Reverse Primer PCR H1 H2 H3 H4 H5 H6 H7 H8 H9 H10 Conditions
APOA1 (aa 1-243)-AD 1 ggaattcgatgaacccccccag cgggatcctcactgggtggc 55ºC 243 Mouse Apoa1 cgcggatccatgatgaaccccag cggatcctcactgggcagtcag TD68-61ºC Apoa1(aa 1-240)-AD 1 240 Apoa1 (aa 1-148)-AD1 148 cgcggatccatgatgaaccccag ccgctcgaggcgaaattcctca 55ºC
Apoa1 (aa 83-148)-AD 83 148 cgggatccatcaggagatgaac ccgctcgaggcgaaattcctca 55ºC
Apoa1 (aa 1-160)-AD 1 160 cgcggatccatgatgaaccccag cgggatcctgtgcgcagagagtc 55ºC
Apoa1 (aa 83-240)-AD 83 cgggatccatcaggagatgaac ccgctcgagtcactgggcagtcagagtc TD68-61ºC 240
Apoa1 (aa 83-210)-AD 83 210 cgggatccatcaggagatgaac ccgctcgagcctccagcgcagg TD68-61ºC
Apoa1 (aa 161-240)-AD 161 cgaattccagctagcgccccac ccgctcgagtcactgggcagtcagagtc 55ºC 240
Apoa1 (aa 138-190)-AD 138 190 cgggatccatagactgtcccctgtg ccgctcgagcgtggtactcgttcaag 55ºC
Apoa1 (aa190-210)-AD 190 210 cgggatccatcacaccagggcc ccgctcgagcctccagcgcagg 55ºC
Legend Lipid binding regions
LCAT activation regions H1 Helix
Figure 3.6. Apolipoprotein A-I yeast two-hybrid constructs, primer sequences and PCR conditions. A) Schematic of the full-length human APOA1–Gal4AD and mouse Apoa1 constructs used for domain mapping experiments. Functional regions and helices are indicated as proposed by Brouillette et al (2001) (108). B) Primer sequences and PCR conditions used for the amplification of inserts.
56 A B Forward Primer Reverse Primer PCR Conditions Human APOA2 1 79 ggaattcgttcggagacag cgggatcctcactgggtggc 55ºC
Mouse Apoa2 1 81 ggaattcggggttaagagacag cgggatcctcacttagccgcagg 55ºC
Figure 3.7. Apolipoprotein A-II yeast two-hybrid constructs, primer sequences and PCR conditions. A) Schematic of human and mouse full-length mature apolipoprotein A-II constructs. B) Primer sequences and PCR conditions.
57 A B Mouse apolipoprotein family constructs Forward Primer Reverse Primer PCR Conditions Apoa1 1 240 cgcggatccatatgaaagctgtggtg ccgctcgagtcagctctccagagg TD68-61ºC
Apoa2 1 81 ggaattcggggttaagagacag cgggatcctcacttagccgcagg 55ºC ggaattcatgaggctcttcatcgc ccgctcgagctacacgttttttattg 55ºC Apoc1 1 88 ggaattcatggggtctcggttct ccgctcggagctactctcccctcag 55ºC Apoc2 1 97 ggaattcatgcagccccggacgc ccgctcgagtcacgactcaatagc 55ºC Apoc3 1 99 ggaattcatggtgaccatgctg ccgctcgagttacaggaagtccgg TD63-55ºC Apod 1 189 ggaattcatgaaggctctgtgg cgggatccattctcctgggccac TD63-55ºC Apoe 1 311
Figure 3.8 Mouse apolipoprotein family yeast-two hybrid constructs, primer sequences and PCR conditions. A) Schematic of full-length apolipoprotein family members amplified from nerve cDNA. Numbers indicate amino-acid lengths of the apolipoproteins B) Primers and PCR conditions used to amplify inserts for cloning into pGADT7-AD vector.
58 3.3 Results
3.3.1 The identification of colonies activating reporter genes
To identify proteins that interact with NDRG1 in CNS tissue, a LexA-based yeast two-hybrid screen of a human fetal brain library consisting of 1.2 x 106 clones was performed. From a co-transformation efficiency of 1.0 x 106 clones and after 14 days of incubation, 115 colonies grew on selective-dropout media lacking histidine, indicating activation of the HIS3 reporter gene. From these, 21 colonies were able to activate the additional second reporter gene LacZ (Figure 3.9).
To identify proteins that interact with NDRG1 in peripheral nervous tissue, a Gal4- based yeast two-hybrid screen of a mouse sciatic nerve library consisting of 2.9 x 106 clones was performed. From a co-transformation efficiency of 3.62 x 105 clones and after 9 days of incubation, 71 colonies grew on high stringency QDO media (-Trp/-Leu/-His/-Ade) indicating activation of both ADE2 and HIS3 reporter genes. Of these colonies, 14 activated the additional colorimetric reporter gene LacZ. From 293 colonies that grew on medium stringency TDO media (-Trp/-Leu/-His), 27 were able to activate all three reporter genes
HIS3, ADE2 and LacZ when transferred to QDO media. A total of 41 colonies were identified which activated all 3 reporter genes assayed in this screen (Figure 3.10).
59 ) 3 B E R 5 C 2 ( y n l i i e t 1 m a o a r f e e r e p r o m m t m e c g i y y c y I r n z z I a z r i n n a F n n d V ) ) e i e e n C n e e i C C g t g s g A A A o - - e i n B o n t a n 3 3 3 t i P P i r i t u t d e a t l p S i S y y y a a a v g l l l o x r i g i i i g N N g n s e e e e e o S ( ( u u 9 a 9 t t t t u 9 o j m m m j n g l j l c a a a a 2 7 2 2 a a a n n l l l l o n a C C n r f f f E i i L L L o e L o p o o o o e 1 1 o f r c l l l s e e e s s s s - - c l n c d d i i i i m a c a a e n n n a n n o d n n n o i n n i o o o o o o o o i o R a t r i m m m i t o t t m t t t t t l t l t i a i h z s s s o o o i L h o l u u P i i i u d d d d s s s c u s c u s c c F b e e e e q o o o o i i q H H H
A o o q n l l l l M i t i t t t i i i i i Z b b b a K b i b i i y i e b 3 3 e 3 i a a a a b A r L 1 c U R R C F F F F R R U H H R R H M T U P l l o o r r t t n n o o c c l l a a g g - - X X e e v v i i t t a i s g o e P B N
Figure 3.9. PCR amplification of library inserts from HIS and LacZ +ve colonies identified from screening a human fetal brain cDNA library with NDRG1-LexABD A) pACT2 Fetal Brain library screening primers were used to amplify library inserts isolated from yeast plasmid preparations. PCR products were sequenced and candidate interactors were chosen for re-transformation into L40 to retest for interaction. In lanes with multiple products, each band was separated and excised from the gel and sequenced or alternatively yeast plasmid preps from which the products were derived were used to transform DH5α and bacterial plasmids were isolated and sequenced as described. B) Filter lift β-galactosidase assay of yeast colonies grown on selective media (-Trp/ -Leu/ -His) from which library inserts in panel A were isolated and amplified. Blue colour indicates activation of LacZ, signifying a positive interaction.
60 ) 1 a r P ( 1 r I o I I I - I t I I - t 3 3 p - i - t i 3 3 e n R n L L u c A A u T l l b c 8 8 b a a u U 1 1 n n u s A ’ i 0 0 m m i s 3 e 8 8 o o e e e s b s s 1 1 e t s t a 5 5 5 5 s o a o 5 0 0 a d a o b b E E E E o 0 i 0 E d i r i s s s R i s s s 0 0 x r r r r e e e r r r A t t t r r t t t d l l l l t t t l l l l x r r r r r 7 7 r r r o l o o l o o o y a a a a a a a a a a a t t t d t r o e p e e t o o p e e e a 1 1 a l l l r h t t i s s s c c i c c o s s s c e e m m m m o o o m e m e m m r t r o a a n e e a a a a o n n n a t n n n B l s s s o o o o i i i f f o a k f f o d o o d a i i i f n v v n i i i m m s a s s s r a s s i i p s s p v n n n o o t t o l n n n n c a i o o o o i o n i o o o o l o e e i e l o o e e e r i o t m o r c c t t t t l l l l o o o o l l l b b b b c l y o c b i b b b c h i i h i i n h i a a o i i i h i i i M p p i i p p p p t t t t t - - u d d d a s i c i i c o c o n i c i i o n n t t t R a a R a a R R - R t t t a R e e e R R o o q o o + o l l l i i o i i b e o e l l l e i l l l t t i t t n t t t t t p t s t r s s s i i i p s b s b s s u u u k i i i k i i i u u u i r u u b i y o a i y i a a a 8 8 8 8 8 2 k 8 8 n n n n n 1 A I F 1 A F A F R M P C 1 M U I 2 M I I M M 1 C R A 1 1 I M N C M R M 1 1
B
Figure 3.10. PCR amplification of library inserts from ADE, HIS and LacZ +ve colonies identified from screening a mouse peripheral nerve cDNA library with NDRG1-Gal4BD. A) Library screening primers were used to amplify library inserts isolated from yeast plasmid preparations. PCR products were sequenced and candidate interactors were chosen for re-transformation into AH109 to retest for interaction. In lanes with multiple products, each band was separated and excised from the gel and sequenced or alternatively yeast plasmid preps from which the products were derived were used to transform DH5-α and bacterial plasmids were isolated and sequenced as described. B) Filter lift β-galactosidase assay of yeast colonies grown on QDO (-Trp/ -Leu/ -Ade/ -His) media from which library inserts in panel A were isolated and amplified. Blue colour indicates activation of LacZ signifying a positive interaction.
61 3.3.2 Eliminating false-positive clones
A commonly encountered problem in any yeast-two hybrid screen is the large number of false-positive clones that may be identified. Sorting and eliminating these clones can be a laborious process and involves co-transforming the library plasmid into yeast with an unrelated bait (e.g. LaminC-BD) from that used in the initial screen. Additional tests include co-transformation with the BD fusion construct alone, as well as transformation of the library plasmid only to test for reporter gene auto-activation in the absence of the BD fusion. This process can be expedited by sequencing clones identified in the library screening, to determine which clones are worthy of further analysis and which can be discarded.
A number of clones from both screens were identified as “true” false-positive and
“probable” false-positive interactors and were eliminated from further analysis. The criteria for omission were as follows: Those assigned “probable” false-positives had been previously identified as common false-positives in screens using different “baits” and appear on lists published by laboratories proficient in the yeast two-hybrid technique
(http://www.fccc.edu/research/labs/golemis/InteractionTrapInWork.html). The majority of these common false-positives are cytoskeletal, ribosomal and mitochondrial proteins, as well as elongation and transcription factors (109,110).
Those deemed “true” false-positives displayed auto-transcriptional activity when tested against a LaminC-BD fusion or the BD fusion alone and were discarded. Other clones had multiple stop codons resulting in short reading frames that showed no homology to any proteins in the NCBI database, suggesting they are artifact fusion proteins capable of reporter gene activation in the yeast two-hybrid system. A list of these identified from both screens is presented in Table 3.2.
62 Table 3.2 False positive interactors identified from yeast two-hybrid screening of human fetal brain and mouse sciatic nerve libraries
Human fetal brain screen
ORF Number Criteria for omission DNA homology Protein homology length of clones a.a Ribosomal protein L29 Ribosomal protein L29 Various 3 Common false-positive 1 Ribosomal protein L7a Ribosomal protein L7a 22 1 Common false-positive 1 H3 Histone family 3A H3 Histone family 3A 82 3 Common false-positive 1 Ubiquitin-conjugating Ubiquitin-conjugating 117 2 Common false-positive 1 enzyme enzyme Cytochrome C Oxidase Cytochrome C Oxidase 74 1 Common false-positive 1 VIIc VIIc Translation elongation Translation elongation 51 1 Common false-positive 1 factor A1 factor A1 Mitochondrial solute Mitochondrial solute carrier 131 1 Common false-positive 1 carrier family 25 family 25 cAMP responsive binding cAMP responsive binding 24 1 Transcription factor protein-3 (CREB3) protein-3 (CREB3) PLZF zinc finger protein PLZF zinc finger protein 120 1 Auto-activator of LacZ
Mouse sciatic nerve screen
Initiation Factor E5 Initiation Factor E5 Various 7 Transcription factor Ribosomal 18s, 28s, 12s Ribosomal Various 8 Common false-positives 1 proteins Progesterone Receptor None 17 1 Artifact protein and auto- Membrane Component activator of LacZ Cytochrome oxidase Cytochrome oxidase subunit 16 2 Common false-positive 1 subunit -II II Ubiquitin B Ubiquitin B 36 1 Common false-positive 1 Mitochondrial ribosomal Mitochondrial ribosomal 9 & 22 2 Common false-positive 1 protein L33 protein L33 Riken clone 1700018018 None 23 2 Artifact short protein gi20844292 Carbonic Anhydrase III None 41 1 3' UTR adjacent to (Car3) gi 20341103 polyadenylation signal - Artifact protein 1 Identified as common false positive (http://www.fccc.edu/research/labs/golemis/InteractionTrapInWork.html)
63 3.3.3 Choosing candidate interactors for further analysis
Four candidate NDRG1-interactors were chosen for further yeast two-hybrid analysis. These are listed in Table 3.3. Various tests included cloning the full-length constructs and human homologues (for those identified from the mouse sciatic nerve library screen) into AD vectors and assaying for reporter gene activation when co-transformed with the NDRG1-BD construct. Further analyses included testing for non-specific binding with the negative control LaminC-BD fusion protein as well as domain-mapping experiments to delineate important binding regions. Testing for interaction with the IVS8-
1G>A mutant NDRG1-Gal4BD construct cloned from a patient with CMT4D (Chapter 2) was also performed for clones identified in the mouse sciatic nerve screen. The results of these experiments are presented in the following sections.
Table 3.3. Candidate NDRG1 interactors identified from yeast two-hybrid screening in nerve tissue
Length of Length of Number DNA/Protein homology clone wild-type of clones ORF (a.a) ORF (a.a)
Human fetal brain screen
Reticulon-1C (RTN-1C) 1-208 208 2
Mouse sciatic nerve screen
Apolipoprotein A-I (Apoa1) 83-240 240 1 Apolipoprotein A-II (Apoa2) 1- 81 81 1 Prenylated Rab Acceptor 1 1- 185 185 1 (Pra1)
64 3.3.4 Reticulon-1C interacts with the NDRG1-LexA/BD fusion
Yeast two-hybrid screening of the human fetal brain library revealed two clones encoding the full-length Reticulon-1C protein which activated the reporter genes, HIS3 and
LacZ when retransformed with the NDRG1-LexA/BD construct in the yeast strain L40,
(Figure 3.11). Interaction was not evident when the same clones were co-transformed with a LaminC-LexA/BD construct suggesting that binding with NDRG1 is specific. The interaction was maintained in a “switch” experiment where Reticulon-1C was subcloned into the pBTM116-BD vector and tested with a NDRG1-AD fusion protein produced by subcloning into the pVP16-AD vector. This test showed that the orientation of the
LexA/BD and LexA/AD tags have no bearing on the binding capabilities of the full-length fusion proteins.
A B C
-Trp/-Leu -Trp/-Leu LacZ -His
NDRG1-BD + Reticulon-1C-AD
Reticulon-1C-BD + NDRG1-AD
LaminC-BD + Reticulon-1C-AD
LaminC-BD + NDRG1-AD
Figure 3.11. Reticulon-1C interacts with NDRG1. A) L40 colonies bearing indicated constructs grow on media lacking tryptophan and leucine indicating successful co-transformation B) Colonies bearing Reticulon-1C and NDRG1 constructs grow on histidine-deficient media indicating activation of the HIS3 reporter gene and signalling an interaction. Reticulon-1C-AD and NDRG1-AD fusion proteins do not interact with the negative control LaminC-LexABD C) Filter-lift β- galactosidase assay. Blue colour indicates activation of the second reporter gene LacZ indicating Reticulon-1C interacts with NDRG1.
65 Minimal NDRG1 residues necessary for maintaining the interaction with the full- length Reticulon-1C-AD fusion protein were mapped. A NDRG1-LexA/BD construct spanning amino-acid residues 146 - 316 interacted with Reticulon-1C as indicated by growth on media lacking histidine and activation of the LacZ gene (Figure 3.12). Other longer NDRG1-LexA/BD constructs encompassing the 146 - 316 region did not interact with Reticulon-1C, suggesting folding and binding-site accessibility differs between N’ terminal and C’ terminal truncated constructs.
Reticulon-1C NDRG1-BD Construct Schematic -Trp/-Leu -Trp/-Leu LacZ -HIS
NDRG1 (aa 1-394)-BD 1 394
NDRG1 (aa146-316)-BD 146 316 - NDRG1 (aa 1-316)-BD 1 316 -
NDRG1 (aa 146-394)-BD 146 394
NDRG1 (aa 316-394)-BD 316
-
Esterase/lipase/thioesterase active site Phospopantetheine binding domain Tri-Deca repeat motif
Figure 3.12. Mapping minimal NDRG1-LexABD sites of interaction with the full- length human Reticulon-1C AD fusion Both the full-length NDRG1-LexABD construct and a construct spanning residues 146-316 interact with Reticulon-1C-AD fusion as indicated by activation of the reporter genes HIS3 and LacZ. Red box indicates minimal interaction region. Other NDRG1-LexABD constructs containing residues 146-316 did not interact with Reticulon-1C.
66 3.3.5 The Prenylated Rab Acceptor 1 (Pra1) interacts with the NDRG1-Gal4BD fusion
A clone encoding the full-length mouse Prenylated Rab Acceptor 1 (Pra1) was identified in the mouse sciatic nerve screen as a NDRG1 interactor as indicated by growth on TDO (-Trp/-Leu/-His) and QDO (-Trp/-Leu/-His/-Ade) media and activation of the
LacZ reporter gene (Figure 3.13). Interaction was also confirmed for human PRA1 which shares 94% homology with the murine protein. Both the human and mouse Pra1-AD proteins did not interact with the negative control LaminC-Gal4BD fusion protein indicating interaction is specific to NDRG1 and not the Gal4BD protein. Switching the tags abolished interaction, suggesting that N-terminal tagging with Gal-4BD interferes with folding and accessibility of binding sites and has previously been noted in other interaction studies using the Pra1-Gal4BD fusion (111).
Interaction was abolished when the mouse Pra1-AD construct was co-transformed with the IVS8-1G>A mutant NDRG1-Gal4BD construct (Figure 3.14). It is possible that changes in protein folding caused by the exon skipping mutation (as predicted by secondary structural analysis in Chapter 2, Section 2.6) interrupts the accessibility of binding sites necessary for the interaction.
67 A B C D
-Trp/-Leu -Trp/-Leu -Trp/-Leu LacZ -His -His/-Ade
NDRG1-BD + mouse Pra1-AD
LaminC-BD + mouse Pra1-AD
NDRG1-BD + human PRA1-AD
LaminC-BD + human PRA1-AD
NDRG1/IVS8-1G>A -BD + PRA1-AD
Figure 3.13. Full-length mouse and human PRA1 Gal4AD fusion proteins interact with NDRG1-Gal4BD but not with LaminC-Gal4BD nor NDRG1/IVS8-1G>A- Gal4BD. A) AH109 colonies bearing indicated constructs grow on media lacking tryptophan and leucine indicating successful co-transformation B) Colonies with interacting constructs grow on histidine-deficient media indicating activation of HIS3 reporter gene C) and on histidine and adenine-deficient media indicating activation of HIS3 and ADE2 reporter genes D) Filter-lift β-galactosidase assay. Blue colour indicates activation of the LacZ reporter gene.
68 Domain mapping experiments failed to identify any deletion NDRG1 constructs that interacted with Pra1-Gal4AD, indicating that the full-length NDRG1 is needed for interaction (Figure 3.14). Similar experiments were performed to map the minimal Pra1-
Gal4AD binding sites necessary for interaction with the full-length NDRG1-Gal4BD
(Figure 3.15). Under medium stringency conditions (-His) a smaller construct encompassing the N-terminal 165 amino acids of Pra1 was found to interact with NDRG1.
Pra1 -Trp/-Leu -Trp/-Leu LacZ NDRG1-BD Construct Schematic -HIS
NDRG1 (aa 1-394)-BD1 394 -
NDRG1 (aa 1-316)-BD 1 316 -
NDRG1 (aa 57-146)-BD 57 146 NDRG1 (aa 121-146)-BD 121 146
NDRG1 (aa 146-394)-BD 146 394 - NDRG1 (aa146-316)-BD 316
NDRG1 (aa 316-394)-BD 316
Esterase/lipase/thioesterase active site Phospopantetheine binding domain Tri-Deca repeat motif
Figure 3.14. Mapping minimal NDRG1-Gal4BD sites of interaction with the full- length mouse Pra1-Gal4AD fusion. The full-length NDRG1-Gal4BD fusion was the only construct that interacted with the full- length NDRG1-Gal4BD protein.
69 NDRG1-Gal4BD
Pra1-Gal4AD -Trp/-Leu -Trp/-Leu -Trp/-Leu LacZ A -HIS -HIS-ADE
1 185 -
1 165
1 112
79
132
Cytoplasmic domains Transmembrane domains
B 1
185 Cytoplasm 79 112 132 165 Membrane
Lumen
Figure 3.15. Mapping minimal Pra1-Gal4AD sites of interaction with the full-length NDRG1–Gal4BD fusion. A) The full-length mouse Pra1-Gal4AD fusion protein and a construct spanning residues 1- 165 interact with the full-length NDRG1-Gal4BD fusion as indicated by activation of the reporter genes HIS3 and LacZ. The interaction with the shorter construct (aa1-165) occurs under medium stringency (-His) but not under high stringency (-His/-Ade) conditions. Minimal interaction region indicated by red box. B) Schematic cellular organization of Pra1 domains as proposed by Lin et al. (107).
70 3.3.6 Apolipoproteins A-I and A-II interact with the NDRG1-Gal4BD fusion
Yeast two-hybrid screening in the mouse sciatic nerve library identified one clone encoding the C-terminus (amino acids 83-240) of apolipoprotein A-I (Apoa1) and one clone encoding the mature apolipoprotein A-II (Apoa2). Interaction with the NDRG1-
Gal4BD fusion protein was subsequently confirmed for the full-length constructs of both mouse and human apolipoproteins as indicated by growth on selective media and the activation of the colorimetric reporter gene LacZ in the presence of X-gal (Figure 3.16).
Specificity of the interaction was confirmed by experiments which showed that all clones interacted with the NDRG1-Gal4BD fusion and not with the LaminC-Gal4BD fusion protein. Interaction was not observed when the construct encoding the mutant
NDRG1/GAL4-BD fusion protein missing exon 9 (as a result of the IVS8-1G>A mutation) was co-transformed with the human APOA1 and APOA2 full-length AD constructs (Figure
3.16).
Domain-mapping experiments show that the N-terminal region of NDRG1 containing both the phosphopantetheine-binding site and part of a region identified as an esterase/lipase/thioesterase active site, is necessary for interaction with mouse Apoa1 and
Apoa2 (Figure 3.17).
Apolipoprotein A-I regions important for NDRG1 binding were mapped using a similar strategy (Figure 3.18). The mouse Apoa1 C-terminus (amino acid residues 161-
210, which encode helices 7 and 8), was necessary for interaction with the NDRG1-GalBD fusion protein as indicated by activation of the HIS3, LacZ and MEL1 reporter genes.
71 A B C D
ApoA-I -Trp/-Leu -Trp/-Leu -Trp/-Leu LacZ -His -His/-Ade
NDRG1-BD + mouse Apoa1-AD
LaminC-BD + mouse Apoa1-AD
NDRG1-BD + human APOA1-AD
LaminC-BD + human APOA1-AD
NDRG1/IVS8-1G>A -BD+ human APOA1-AD
ApoA-II
NDRG1-BD + mouse Apoa2-AD
LaminC-BD + mouse Apoa2-AD
NDRG1-BD + human APOA2-AD
LaminC-BD + human APOA2-AD
NDRG1/IVS8-1G>A-BD + human APOA2-AD
Figure 3.16. Full-length mouse and human ApoA-I and ApoA-II Gal4AD fusion proteins interact with NDRG1-Gal4BD but not with LaminC-BD nor NDRG1/IVS8- 1G>A-BD. A) AH109 colonies co-transformed with indicated constructs grow on media lacking tryptophan and leucine indicating successful transformation B) Colonies with interacting constructs grow on histidine-deficient media indicating activation of HIS3 reporter gene and on C) histidine and adenine-deficient media indicating activation of HIS3 and ADE2 reporter gene D) Filter-lift β-galactosidase assay. Blue colour indicates activation of the LacZ reporter gene signalling interaction.
72 A Apoa1 -Trp/-Leu -Trp/-Leu LacZ NDRG1-Gal4BD Construct Schematic -HIS
NDRG1 (aa 1-394)-BD 1 394 -
NDRG1 (aa 57-146)-BD 57 146
NDRG1 (aa 1-316)-BD 1 316 -
NDRG1 (aa 121-146)-BD 121 146
NDRG1 (aa 146-394)-BD 146 394 -
NDRG1 (aa146-316)-BD 316 - NDRG1 (aa 316-394)-BD 316 -
B Apoa2 -Trp/-Leu -Trp/-Leu LacZ -HIS NDRG1-Gal4BD Construct Schematic
NDRG1 (aa 1-394)-BD 1 394 -
NDRG1 (aa 1-316)-BD 1 316 - -
NDRG1 (aa 57-146)-BD 57 146
NDRG1 (aa 121-146)-BD 121 146
NDRG1 (aa 146-394)-BD 146 394 -
NDRG1 (aa146-316)-BD 316 - NDRG1 (aa 316-394)-BD 316 -
Esterase/lipase/thioesterase active site Phospopantetheine binding domain Tri-Deca repeat motif
Figure 3.17. Mapping minimal NDRG1-Gal4BD sites of interaction with the full- length mouse Apoa1 and Apoa2-Gal4AD fusion proteins. A) NDRG1 residues 57-146 interact with the Apoa1-Gal4AD fusion protein as indicated by activation of the reporter genes HIS3 and LacZ. Minimal interaction region indicated by red box. B) NDRG1 residues 1-316 interact with the Apoa2-Gal4AD fusion protein.
73 NDRG1-Gal4BD -Trp/-Leu Mouse Apoa1 - Schematic Apoa1-Gal4AD Constructs -His LacZ MEL1 H1 H2 H3 H4 H5 H6 H7 H8 H9 H10
Apoa1 (aa 1-240)-AD 1 240
Apoa1 (aa 1-148)-AD 1 148
83 148
Apoa1 (aa 1-160)-AD 1 160
Apoa1 (aa 83-240)-AD* 83
Apoa1 (aa 83-210)-AD 83 210
Apoa1 (aa 161-240)-AD 161
Apoa1 (aa 138-190)-AD 138 190
Apoa1 (aa190-210)-AD 210
Lipid binding regions 20 40 60 80 100 LCAT activation regions α-galactosidase milliunits (ml/cells) H1 Helices
Figure 3.18. Schematic of mouse Apoa1 -Gal4AD proteins and analysis of interaction with the full-length NDRG1-Gal4BD protein. A) Bars represent truncated Apoa1-GAL4AD protein with amino acid positions and known functional domains as indicated in legend. * denotes clone identified from initial yeast two-hybrid screen. Red box indicates minimal interaction region delineated by helices 7 and 8 required for binding with the NDRG1-Gal4BD protein. B) Colony growth on media lacking the nutritional selector histidine (HIS) indicating activation of the HIS3 reporter gene. C) Colormetric ß-galactosidase filter lift assays. D) Quantitative evaluation of interactions were tested for activation of MEL1 using liquid α-galactosidase assays. Standard deviations are shown (n=3). 74 In order to investigate whether NDRG1 interacts specifically with Apoa1 and
Apoa2, and not with other members of the mouse apolipoprotein family that share similar structure or biochemical activity (112), the apolipoproteins Apoc1, Apoc2, Apoc3, Apod and Apoe were cloned into the pGADT7-AD vector and tested for interaction with
NDRG1-Gal4BD by yeast two-hybrid analysis. No interactions with any of these apolipoproteins were detected (Figure 3.19), suggesting the association between NDRG1 and Apoa1 and Apoa2 is specific and not due to similar domains shared across the mouse apolipoprotein family.
A B C D
-Trp/-Leu -Trp/-Leu -Trp/-Leu LacZ -His -His/-Ade
NDRG1-BD + Apoa1-AD
NDRG1-BD + Apoa2-AD
NDRG1-BD + Apoc1-AD
NDRG1-BD + Apoc2-AD
NDRG1-BD + Apoc3-AD
NDRG1-BD + Apod-AD
NDRG1-BD + Apoe-AD
Figure 3.19. NDRG1 binds with mouse Apoa1 and Apoa2 but not with other apolipoprotein family members. A) Apolipoproteins C-I, C-II, C-III, D and E do not interact with the NDRG1-Gal4BD fusion. Co-transformants containing NDRG1-Gal4BD construct and apolipoproteins A-I and A-II Gal4AD constructs (highlighted in red) grow on histidine (–His) and adenine (- Ade)-deficient media and activate LacZ signifying a positive interaction. Co-transformants of other apolipoproteins with the NDRG1-Gal4BD construct grow on media lacking tryptophan (-Trp) and leucine (-Leu) indicative of successful co-transformations but do not grow on histidine (–HIS) or adenine (-Ade) -deficient media, nor activate LacZ signifying no interaction with NDRG1.
75 3.4 Discussion
Yeast two-hybrid screening in nervous tissue libraries identified four interesting
NDRG1 interactors which activated all reporter genes assayed, indicating a positive interaction. The specificity of interaction with the NDRG1-Gal4BD protein was supported by the fact that none of the candidates interacted with the negative control LamC-BD protein. Three of the candidates tested (Pra1, Apoa1 and Apoa2) did not interact with a mutant NDRG1 protein (resulting from the IVS8-1G>A mutation) previously shown to be a cause of CMT4D (Ch. 2). (RTN-1C was not tested).
Domain mapping experiments identified possible important functional regions of the NDRG1 proten. The N-terminal region of NDRG1 encompassing the predicted phosphopantetheine-binding domain and the esterase/lipase/thioesterase active site was important for interaction with the apolipoproteins A-I and A-II identified in the mouse peripheral nerve library screen. The central region of NDRG1, in which no specific protein domains are predicted, was important for interaction with RTN-1C identified from the fetal brain screen. The fact that different protein partners and regions of interaction were detected in the NDRG1 protein is consistent with its multiple roles.
Three out of the four candidates identified are widely expressed proteins (Pra1,
Apoa1 and Apoa2) and were not detected across library screens. The fact that there was no overlap of interacting clones may perhaps be explained by the low reproducibility in identifying protein-partners across yeast two-hybrid screens or may reflect library differences in cDNA quality or tissue expression levels.
76 Determining likely functional candidates for further analysis
A general outline of the known functions and expression patterns of the four
NDRG1- interactors and why they may be important functional candidates is presented in the following sections.
Reticulon-1C (RTN-1C)
RTN-1C belongs to the Reticulon family (formerly designated Neuroendocrine-
Specific Proteins) which are mainly expressed, but not limited to, tissues of neural and neuroendocrine origin (113,114). The Reticulon-1 gene (14q21-q22) encodes three overlapping proteins, RTN-1A (135kDa), RTN-1B (45kDa) and RTN-1C (23kDa) which share a common carboxyl terminus that is anchored to membranes of the endoplasmic reticulum (ER). It is through the association with the ER that reticulons may play a role in the regulation of intracellular calcium levels (115). The exact role of RTN-1C is unknown, however it has been linked to growth cone formation, neuronal differentiation and is a marker of neuroendocrine malignancies (114,116). A proapoptotic role for RTN-1C has also been suggested through its interaction with Bcl-xL (117) and also glucosylceramide synthase in neuroepithelioma cells (118). Recently RTN-1C has been linked to vesicle transport pathways via its interaction with snare proteins such as VAMP2 and syntaxin 13
(119).
To date, no disease-causing RTN-1C mutations have been reported, however the expression of RTN-1C has been found to be down-regulated in CNS pathologies including
Down syndrome and in Alzheimer’s disease (120) where it is thought to reflect neurodegenerative changes thus pointing towards an important role in this tissue. In the
CNS, RTN-1C is expressed in neurons of the cerebellar cortex and the cerebral neocortex, as well as in axons of the pituitary neurohypophysis. While NDRG1 RNA transcripts have been detected in the brain (32), the expression of the human NDRG1 protein in this tissue 77 has yet to be determined (54). There is evidence that the Rat NDRG1 homologue is expressed for a short period of time in hippocampal pyramidal neurons before being strongly expressed by glial cells (astrocytes) into adulthood (121). The RTN-1C protein is not expressed in glial cells in the CNS, (114), thus making an in vivo physical interaction with NDRG1 in this tissue difficult to envisage.
Expression of RTN-1C in the PNS has been noted in nerves of the stomach, small intestine and lung, however it is not known if the protein is expressed in peripheral glial cells such as Schwann cells, or if expression is limited to neurons and axons as is the case in the CNS (114). High RTN-1C expression has been detected in hNT2 cells that have been induced to differentiate to a neuronal phenotype (116). NDRG1 is not expressed in differentiated hNT2 cells, nor in axons (10,26,68,69). The fact that both proteins seem to show differences in cell-type expression makes a functional link between the two proteins hard to justify. Clearly more data need to be gathered concerning the expression of both proteins in nervous tissue to determine if the interaction detected by the yeast two-hybrid method is a biologically relevant one. The recent commercial availability of a RTN-1C antibody paves the way for such investigations to be undertaken in the future.
Prenylated rab-acceptor protein 1 (PRA1)
The Prenylated rab-acceptor protein (PRA1) is a Golgi-associated protein involved in regulating the membrane targeting and intracellular trafficking of small GTPases including
Ras and Rab proteins (122,123). Ras proteins have been shown to regulate various cellular functions such as proliferation, differentiation, cell death and survival (124), while Rab proteins are involved in biosynthetic/secretory and endocytic pathways (125). Multiple binding partners of PRA1 have been identified and include not only a large number of Ras
78 and Rab-related proteins but also viral proteins (Table 3.4). This large number of interactors and their varied roles suggest that PRA1 mediates a number of important cellular processes.
Table 3.4. PRA1-protein partners identified from yeast two-hybrid screens
PRA1 protein-partners Function Reference
Rab proteins – Rab3a, Biosynthetic/secretory and endocytic (122) 4a, 4b 5a, 5c, 7, 17, 22 pathways Ras proteins – Ha-Ras, Proliferation, differentiation, cell death and (123) RhoA, TC21, Rap1a survival v-Snare Vamp2 Vesicular trafficking (126) BHRF1 Homologue of anti-apoptic Bcl-2 (127) Viral proteins – Rotovirus spike protein involved in cell (111,128) SIVgp41 CD, VP4 spike attachment, penetration and protein hemagglutination
No disease-causing mutations have been identified in the PRA1 gene, however a PRA1 protein-partner, Rab7 (122), has recently been shown to be mutated in CMT2B disease
(129). The mutation results in a mild to severe axonal neuropathy with sensory and motor deficits, including marked distal muscle weakness and wasting. The mechanism as to how a mutation in Rab7 results in the neuropathology is unclear. However over-expressing Rab7 in Niemann-Pick type C disease fibroblasts can correct lipid trafficking and reduce the accumulation of intracellular cholesterol (130). As neuropathy is also a feature of
Niemann-Pick type C disease (MIM#257220), this finding may point towards an important role of Rab7 in lipid trafficking pathways in the nervous system. Whether PRA1, through its association with Rab7 or indeed NDRG1, is involved in this or a similar process is not known.
79 The expression of the PRA1 protein in peripheral nerve is unknown. The lack of a suitable PRA1 antibody explains why investigations into its tissue and cell-type distribution have been limited to analysis of its RNA transcript (122). The PRA1 transcript is expressed in a wide variety of tissues including the brain, and is similar to the tissue distribution of the NDRG1 transcript (32,54). However, no data exists concerning PRA1 expression in peripheral nerve, a fact that needs to be investigated if the interaction with NDRG1 detected by this yeast two-hybrid screen, is biologically relevant. In transfected cells, PRA1 constructs have been shown to move from the ER to the Golgi and are anchored in the membrane with parts of the protein having access to the cytoplasm (107). The subcellular expression pattern of NDRG1 has revealed both ER and cytoplasmic localisation (36) therefore an in vivo interaction between the two proteins would be spatially possible.
The fact that mutation to a PRA1 interactor results in a neuropathy, coupled with the proposed vesicular trafficking role of PRA1, suggests that PRA1 may have an important role in the nervous tissue. Therefore this interactor is a worthy candidate, that should be considered for future confirmatory studies.
Apolipoproteins A-I and A-II
The apolipoproteins A-I and A-II are the principal protein components of high- density lipoprotein (HDL). While the exact role of APOA2 is currently unknown, APOA1 has been shown to be involved in the efflux of intracellular cholesterol in a process known as reverse cholesterol transport (RCT) (131-133). The removal of excess cholesterol is not only crucial for regulating cellular cholesterol homeostasis, but can also regulate receptor endocytosis and recycling (134), and cell signaling (135).
APOA1 is an interesting functional candidate as it has been associated with neuropathies via either direct mutation to the gene (136), or through disruption to linked 80 components of the cholesterol transport pathway. Peripheral neuropathy is a consistent clinical finding in Tangier disease (MIM#205400), caused by mutations in the ATP-
Binding Cassette transporter (ABCA1), which encodes a putative APOA1 docking site essential for the removal of intracellular cholesterol (137). Neuropathy is also a feature of
Niemann-Pick disease type C (MIM#257220) caused by compromised intracellular cholesterol transport from late endosomes-lysosomes to other cellular compartments (138).
The interaction between NDRG1 and APOA1 detected by yeast two-hybrid screening in peripheral nerve provides the first direct evidence linking NDRG1 to a role in a lipid/cholesterol pathway and perhaps an important role for APOA1 in the peripheral nervous system.
A comparison of sequence homology of apolipoprotein A-I across species shows that more than 60% of the protein sequence is conserved (108,112). While similarities in primary sequence may seem relatively low, it is the secondary structure conferred by the conservation of tandemly repeated 22-residue amphipathic alpha-helices that is functionally important. These helices allow hydrophobic residues to face inwards facilitating a conformation so as to encapsulate around the surface of lipid particles. This feature is thought to confer lipid-binding properties to all the apolipoprotein family members (112).
The fact that the NDRG1-BD fusion did not bind to apolipoproteins C-I, C-II, C-III, D nor
E suggests that the interaction with the apolipoproteins A-I and A-II is specific. A previous yeast two-hybrid screen using bait unrelated to NDRG1 has also identified clones encoding both APOA1 and APOA2 (139) emphasising their structural and functional similarities.
Apoa1 and Apoa2 interacted with different NDRG1-Gal4BD constructs encompassing the esterase/lipase/thioesterase active site and phosphopantetheine-binding region perhaps pointing towards unique and specific binding sites within these domains. Delineating the exact residues necessary for the interactions may provide clues to the structural and 81 functional properties of the proteins. The minimal region of interaction for the Apoa1 protein was mapped to the C-terminus, a region that has been shown to be important for lipid efflux and HDL biogenesis (108,140).
Future directions
Yeast two-hybrid library screening often involves the identification of many candidate interactors. Confirming interaction by alternative methods and delineating the biological significance of each of these interactors is a large task. Given the time and budget constraints imposed on a PhD project, not all interactors can be pursued. Thus, choosing a candidate for further confirmatory interaction and functional assays is an important decision that was made on the basis of a number of factors, including the known functional relevance, links to nerve pathology and practical considerations such as the availability of antibodies for endogenous in vivo analysis. APOA1 was deemed a primary candidate for further investigations for these reasons. Its choice as a functional candidate was also strengthened by evidence linking NDRG1 to lipid pathways such as the predicted phosphopantetheine and esterase/lipase/thioesterase domains, membership to the abhydrolase family (59) and the upregulation of expression by agents invoking cardiovascular disease (141,142).
The following chapter describes experiments undertaken to confirm the NDRG1 protein interactions outside of the yeast two-hybrid system and investigations into the expression of NDRG1 and APOA1 in peripheral nerve. In vitro confirmatory experiments undertaken to provide evidence for NDRG1 interaction with APOA2 and PRA1 are also presented.
82 CHAPTER 4 - CONFIRMING NDRG1-PROTEIN INTERACTIONS
4.1 Background
The identification of NDRG1 interactors using the yeast two-hybrid technique is an important step towards elucidating a role for NDRG1. In order to determine whether these are biologically relevant interactions and not artifacts of the experimental system in which they were discovered, additional confirmatory experiments need to be performed. While the yeast two-hybrid method is a powerful tool in identifying protein-partners, the technique does have limitations. One justified criticism of this system is that the interaction is measured in the nucleus of the cell. This is a cellular compartment that may be different from the normal subcellular localisation of the proteins in vivo. One way to provide support that the proteins may associate in the normal biological state is to see whether the proteins co-localise. This may be achieved by immunofluorescent detection of the proteins in tissue sections or cultured cells (Figure 4.1). In the absence of specific antibodies to the target proteins, tagged contructs encoding recombinant proteins may be transfected into cells.
A B
Figure 4.1. Schematic representation of the principle of detecting the co-localisation of two proteins using immunochemistry. Protein A is labelled with antibodies conjugated to the CY3 (Red) fluorophore, Protein B is labelled with antibodies conjugated to the FITC (Green) fluorophore. If the proteins are in close proximity, the spectral overlap of the red and green fluorophores results in fluorescence in the yellow channel.
83 In both methods, localisation is detected using confocal microscopy. While co-localisation of the proteins does not provide conclusive proof of interaction, it does show that the proteins share the same cellular compartment making a physical association spatially possible.
In addition to co-localisation experiments, a number of different confirmatory protein binding assays (Figure 4.2) can be performed to provide additional support that an interaction detected by the yeast two-hybrid method is a real one. Some of the assays include pulldown experiments, where the bait protein is expressed as a recombinant
Glutathione-S-transferase (GST) fusion protein in bacterial or mammalian cells and is immobilized on glutathione beads. A cell lysate containing the prey protein, or the purified prey protein alone is then passed over the GST-bait fusion and the protein complex is separated by SDS-PAGE, transferred to membrane by Western blotting and the membrane probed using antibodies to detect the proteins.
Isolation of interacting proteins can also be achieved by co-immunoprecipitation. In this method, an antibody specific to a bait or prey protein is added to a tissue lysate or a lysate from cells overexpressing the proteins of interest. The antibodies bind to epitopes on the target protein and then the immuno-bound protein is captured on sepharose beads which bind to the IgG portion of the bound antibody. Any interacting proteins or protein complexes that may be associated with the immuno-bound protein can then be detected by separation using SDS-PAGE, Western transfer to a membrane and probing with antibodies.
An advantage of this technique over the pulldown assay is that if antibodies to the target proteins are available, the proteins are in the native form as opposed to relying upon appending artifical tags which may affect the structural properties of the protein.
84 A – GST Pulldown A B
Bait Prey
Bait (e.g NDRG1) is Bait/prey complex is captured Bait/prey complex is boiled and separated on expressed as recombinant by glutathionine beads which SDS-PAGE gel → Western transfer and the blot GST fusion protein and bind to the GST moeity and the is probed with antibodies to bait and prey: incubated with prey protein complex is washed to remove A) Expected band pattern if interaction (eg. tagged construct, cell non-specific binding proteins. B) Expected band pattern if no interaction. lysate or purified protein).
B – Co-immunoprecipitation
A B
IgG Heavy chain Bait Prey IgG Light chain
Bait or prey-specific Bait/prey immuno-complex is Bait/prey immuno-complex is boiled and antibodies are added to a captured by sepharose beads separated on SDS-PAGE gel → Western transfer lysate containing Bait (e.g and washed to remove non- and the blot is probed with antibodies to bait and native or tagged NDRG1) and specific binding proteins. prey: Prey protein (eg. tagged A) Expected band pattern if interaction constructs or tissue lysate). B) Expected band pattern if no interaction.
C – Bioluminescence Resonance Energy Transfer (BRET)
Donor Acceptor Donor Acceptor Donor Acceptor
Light emission EYFP Rluc EYFP Rluc (530nm) Rluc EYFP Coelentrazine
Light BRET Coelentramide (480nm)
Mammalian cells are The Rluc substrate If the Rluc and EYFP moities transfected with Bait tagged coelentrazine is added and are within less than 100Ǻ the with Renilla luciferase (Rluc) oxidation results in Rluc light acceptor molecule EYFP is and Prey tagged with emission. stimulated and a BRET signal enhanced yellow fluroscence is generated indicating protein (EYFP). interaction.
Figure 4.2. A schematic overview of the confirmatory protein-protein interaction assays utilized to confirm the NDRG1 and apolipoprotein interaction.
85 The pulldown assay and co-immunoprecipitation technique are limited to detecting stable and strong protein-protein interactions that can survive the numerous washing steps used to eliminate non-specific interactions and background noise. To circumvent these shortcomings, a highly sensitive technology known as bioluminescence resonance energy transfer (BRET) has been developed and shown to be useful for detecting transient or weak protein-protein interactions in intact living mammalian cells (143). In the BRET methodology, the “bait” construct is fused to a donor molecule (Renilla luciferase or Rluc) and the “prey” construct is fused to an acceptor molecule (Enhanced yellow fluorescent protein or EYFP) and are co-expressed in mammalian cells. If the two proteins do not interact, only one signal, emitted by the luciferase, can be detected after the addition of the
Rluc substrate coelentrazine. If the proteins interact, resonance energy transfer occurs between the luciferase and the EYFP and an additional signal, emitted by the EYFP, can be detected. Similar to co-localisation experiments, the BRET assay is reliant upon the close proximity of the proteins, however it is far superior in providing proof of interaction as resonance energy transfer can only occur if the proteins are within 10-100 Angstrom (Å), a distance that is indicative of a molecular interaction.
This chapter describes the experiments that were undertaken to provide additional support for the NDRG1 interaction with APOA1, APOA2 and Pra1. The specific aims of these investigations were to utilize the aforementioned technologies to :
Visualise the subcellular co-localisation of NDRG1 and APOA1 in transfected
mammalian cells and in the endogenous state in different cells and tissues.
Provide additional support for the NDRG1-APOA1 interaction by using the GST
pull down assay, and co-immunoprecipitation technique.
Confirm the NDRG1 interaction with APOA1, APOA2 and Pra1 in mammalian
cells using BRET. 86 4.2 Methods
4.2.1 Cell lines and Transfections
COS7 and HepG2 cells (American Type Tissue Culture, Manassas, VA, USA) were grown in DMEM (Gibco, BRL) supplemented with 2mM L-glutamine, 100U/mL penicillin/streptomycin and 5% (v/v) fetal calf serum (FCS) at 37°C in an atmosphere of
TM 5% CO2. Transfections were performed in a 6-well format with LipofectAMINE -2000 reagent (Invitrogen) or GenejuiceTM (Novagen) according to manufacturers instructions.
4.2.2 Immunocytochemistry and immunohistochemistry experiments
For the immunocytochemical detection of the NDRG1 and APOA1 co- localisation, human hepatoma HepG2 cells were grown on coverslips, transfected with 1µg each of pcDNA3- NDRG1/HIS and pcDNA3- APOA1/ EYFP expression constructs and fixed at 48 hrs post-transfection with 4% paraformaldehyde. The cells were blocked with
3% BSA/TBST and the NDRG1/HIS fusion protein was probed using an antibody to the
HIS epitope tag (1:200) (Santa Cruz, USA), and an anti-IgG Rhodamine-conjugated secondary antibody (1:250) (Rockland, USA). EYFP fused to the C-terminal of APOA1 was detected by direct fluorescence.
For immunodetection of endogenous NDRG1 and APOA1, HepG2 cells were fixed and blocked in the same manner and incubated overnight at 4 ºC with affinity-purified polyclonal rabbit anti-NDRG1 antibody (1:250) (A gift from K. Kokame & T. Miyata ,
National Cardiovascular Center Research Institute, Japan) and mouse anti-human APOA1 antibody (1:100) (Southwestern Medical School, Center for Biomedical Inventions,
University of Texas, USA). Secondary antibodies included FITC-conjugated anti-rabbit
IgG (1:250) (Rockland, USA) and Cy3-conjugated anti-mouse IgG (1:250) (Sigma).
87 For immunohistochemistry, a section of human sural nerve (taken from a region
2cm proximal to a neuroblastoma biopsy) was mounted in Cryomatrix embedding medium
(Thermo Shandon, PA, USA) and frozen in LiN2-cooled methyl-butane for 20 seconds and stored in LiN2 until sectioning. 8µm thick sections were cut at –19ºC using a Reichert-Jung
2900 Frigocut E cryotome and thaw mounted on 0.1% poly-L-lysine coated slides. Tissue sections were fixed in acetone for 10 minutes, rehydrated with PBS for twenty minutes and blocked with 3%BSA/0.1%PVP/0.05%Triton X-100 in PBS for 30 minutes. The sections were incubated with primary antibodies or blocking diluent only (for negative controls) overnight at 4ºC. The sections were washed in PBS and incubated with secondary antibodies (FITC-conjugated anti-rabbit IgG (1:250) (Rockland, USA) and Cy3-conjugated anti-mouse IgG (1:250) (Sigma)) for 1 hour at room temperature. All slides in the immunohistochemistry and immunocytochemistry experiments were washed 3 times in
PBS and mounted in Vectasheild with DAPI (Vector Labs Inc, CA, USA). Fluorescence was visualised using a Bio-Rad MRC 1024 UV Laser Scanning Confocal Microscope (Bio-
Rad, Hercules, CA). Images were merged using the Confocal Assistant software (version
2.1).
4.2.3 Pull-down assays
4.2.3.1 Bait production - pGEX4T-1 NDRG1-GST construct
To produce the NDRG1/GST fusion construct for use in the pull-down assays, the complete human NDRG1 reading frame (1182nt / 394aa) was PCR amplified from IMAGE clone #3181396 (The I.M.A.G.E Consortium) using primers (5'-
GGAATTCATGTCTCGGGAGATGCAGG-3' & 5'-
CCGCTCGAGCGGCTAGCAGGAGACCTCCA-3') appended with EcoRI (5') and XhoI
(3') restriction sites, digested and cloned into the pGEX4T-1 vector (Amersham 88 Biosciences). The PCR reaction consisted of 1x PCR buffer (Invitrogen), 1.5mM MgSO4,
25µM dNTPs, 20ng of each primer, 1 unit of Platinum Pfx Taq polymerase (Invitrogen),
1µL of template DNA and dH2O to a final volume of 50µL. PCR conditions included denaturation at 95°C for 5 minutes, followed by 40 cycles of 94°C for 30 seconds, with a touchdown regime of 68°C to 61°C for the first 15 cycles and 61°C for the remaining cycles. Extension was at 72°C for 40 seconds, with a final extension step at 72°C for 7 minutes. The PCR product was double-digested with EcoRI and XhoI restriction enzymes
(New England Biolabs). The digest reaction contained 1 x EcoRI restriction buffer, 1%
BSA and 3 units of each enzyme in a total volume of 50µL. The reaction was incubated at
37°C for 6 hours and the reaction product was electrophoresised on 1.5% agarose gel, excised and purified using QG buffer (Qiagen) and Qiagen PCR Purification columns.
The pGEX4T-1 vector was prepared for ligation in a similar manner, except for an additional dephosphorylation reaction after restriction digestion, using 1 unit of shrimp alkaline phosphatase (SAP), 1 hour incubation at 37°C, followed by heat inactivation of the
SAP at 65°C for 20 minutes. The digested NDRG1 PCR product was ligated into EcoRI and XhoI sub-cloning sites in pGEX4T-1 downstream of the GST open-reading frame using the T4 Rapid DNA Ligation kit (Roche) according to manufacturers instructions.
Transformation of the ligation reaction into competent bacteria, propagation, plasmid DNA isolation and sequencing were performed as detailed in Chapter 3, Section
3.2.2 with the exception that the protease-deficient Eschericia coli BL21 strain was used.
4.2.3.2 Recombinant NDRG1-GST protein production and purification
A single bacterial colony containing the pGEX4T-1/NDRG1/GST construct was seeded into 10mL of LB medium with 100µg/mL ampicillin and grown overnight at 37°C
89 with vigorous shaking. This culture was used to inoculate 100mL of LB/ampicillin and was grown for approximately 3 hours until an OD600 of 1.0 was reached. Protein expression was induced by adding 2mM isopropyl ß-D-thiogalactopyranoside (IPTG). The culture was left to grow for a further 5 hours and the cells were pelleted by centrifuging at 5000rpm for 5 minutes. The pellet was resuspended in 1mL of TNTG lysis buffer (40mM Tris pH 8.0,
100mM NaCl, 5mM EDTA, 1% Triton X-100, 25mM NaF, 1mM Benzamidine (Sigma),
0.1mM PMSF (Boehringer), 10µg/mL, Aprotinin (Sigma), 10 µg/mL Leupeptin (Sigma)) and was subjected to 3 freeze-thaw cycles to lyse the cells. The homogenate was then sonicated 3 times, (ten seconds each) and the cell debris were cleared by centrifuging at 13
000rpm for 20 minutes at 4°C. The supernatant was removed to a fresh tube and 200µL of glutathione-agarose was added. Binding of the NDRG1-GST fusion protein to the glutathione-agarose was achieved by overnight incubation at 4°C with gentle rotation. The
NDRG1-GST/glutathione-agarose complexes were washed 5 times with 1mL of TNTG buffer and were used immediately for binding assays. During the course of this PhD project a recombinant NDRG1/GST protein expressed in baculovirus-infected Sf9 insect cells was kindly produced and provided as a cell pellet by Dr A.Akkari, (ANRI). Recombinant
NDRG1/GST protein from a stock of this source maintained at –70°C was isolated and purified according to the above protocol. This protein was utilized in subsequent experiments on the assumption that eukaryotic post-translational modifications inherent to
Sf9 insect cells may produce a more functionally active NDRG1-GST recombinant protein than the prokaryotic E.coli produced protein.
90 4.2.3.3 Expression of recombinant Apoa1/HA protein in COS7 cells for use in the pulldown assay
The complete ORF of Apoa1 was amplified from mouse sciatic nerve cDNA with primers (5'-CGGGATCCAAAGCTGTGGTGCTGGCCG-3' & 5'-
CCGCTCGAGTCACTGGGCAGTCAGAGTC-3') and ligated into BamHI and XhoI restriction sites in the pcDNA3-HA vector. The construct was propagated, isolated and sequenced according to the protocol detailed in Chapter 3, Section 3.2.2.
Up to 2µg of pcDNA3-HA/Apoa1 vector DNA was transfected into a T75 tissue culture flask of COS7 cells using LipofectAMINETM-2000 reagent (Invitrogen) according to manufacturers instructions. The cells were incubated for 48 hours and were detached with
500µL of 1x trypsin at 37°C for 5 minutes. The cells were transferred to an Eppendorf tube and spun at 3000rpm for 2 minutes. The supernatant was removed and the cell pellet was washed with ice-cold PBS. The cells were resuspended in 500µL of lysis buffer, incubated on ice for 30 minutes and spun at 14 000rpm for 30 minutes at 4°C. Up to 200µL of the supernatant fraction were incubated with the NDRG1-GST fusion protein with gentle rocking at 4°C for 4 hours.
4.2.3.4 Preparation of purified human APOA1 protein for use in the pulldown assay
Purified lyophylised human APOA1 was purchased from Sigma (cat# A0722) and resuspended in sterile phosphate buffered saline (PBS). Different quantities (2µg and 5µg) of the reconstituted protein were incubated with the NDRG1-GST fusion protein at either
4°C or 37°C for four hours in 200µL of binding buffer (40mM Tris pH7.5, 50mM NaCl,
PBS).
91 4.2.3.5 Preparation of tissue lysates for use in the pulldown assay
Whole tissue homogenates used in the binding assays were prepared as follows.
Mouse sciatic nerve and kidneys were removed from adult BalbC mice and washed in sterile PBS, frozen in LiN2 and stored at –70°C until processing. Protein from up to 8 sciatic nerves or 1 kidney were processed at each time. The tissues were defrosted in 1mL of TNTG lysis buffer, homogenised by cutting with scissors and ground with an automated pestle tip in a 1.5mL Eppendorf tube. The lysate was passed through a 19 gauge needle up to 20 times and sonicated for 30 seconds. Tissue debris were collected by centrifuging at 13
000rpm for 20 minutes at 4°C. Up to 200µL of the supernatant was then added to 25µL of
NDRG1-GST fusion protein cross-linked to glutathione-agarose and the tubes were incubated at 4°C overnight with gentle rocking to facilitate binding.
4.2.4 Co-immunoprecipitation
Protein homogenates were prepared from human hepatoma (HepG2) cells, mouse kidney and mouse sciatic nerve as described above. Endogenous immunoglobulins were cleared by incubating the supernatant with 50µL of Protein G beads (Amersham
Biosciences) with gentle rocking at 4°C for 2 hours. The supernatant was transferred to fresh tubes and diluted 1:2 with TNTG lysis buffer to a total volume of 200µL. Two antibodies including polyclonal rabbit anti-Cap43 (NDRG1) (Zymed, USA) and polyclonal rabbit anti-NDRG1 (A gift from K. Kokame & T. Miyata, National Cardiovascular Center
Research Institute, Japan) were used to alternatively immunoprecipitate NDRG1. Human
APOA1 and mouse Apoa1 were immunoprecipitated using anti-human APOA1
(Southwestern Medical School, Center for Biomedical Inventions, University of Texas,
USA) or anti-mouse Apoa1 (Rockland, USA) antibodies respectively. In all cases, antibodies were added to seperate tubes of diluted lysate in a final ratio of 1:200 and were 92 rocked gently for 2 hours at 4°C. The lysates were spun briefly at 200rpm and 50µL of protein G beads in PBS slurry were added. The tubes were left to rock for a further 2 hours at 4°C. Beads were collected by spinning tubes for 10 seconds at 400rpm. The supernatant was discarded and the beads were resuspended and washed three times in 500µL of wash buffer. 50µL of protein loading buffer (50mM Tris-HCl pH6.8
2% SDS, 10% Glycerol, 1% -Mercaptoethanol, 12.5mM EDTA, 0.02 % Bromophenol
Blue) was added and the samples were boiled for 5 minutes. 20µL of each sample were loaded onto a 4-20% gradient Tris-Glycine SDS-PAGE gel (Novex) and electrophoresed at
125V for 2 hours.
4.2.5 Western blotting
Proteins were transferred to Hybond-P™ membranes by Western transfer (30V overnight at 4°C or 200V for 2 hours at 4°C) using a Biorad transfer tank. The membranes were blocked with 5% milk in TBST (25mM Tris pH 7.5, 150mM NaCl, 0.1% Tween®-20
(ICI America LTD)) and incubated with anti-NDRG1 (1:1000), anti-GST (1:1000), anti- human APOA1 (1:1000) or anti-mouse Apoa1 (1:3000) primary antibodies for one hour with gentle agitation and washed 3 times in 50 mL of TBST for 15 minutes. The membranes were probed with secondary antibodies anti-mouse IgG (1:16 000) (Cell- signalling USA), anti-rabbit IgG (1:15 000) (Sigma) or anti-goat IgG (1: 15 000) (Sigma) for one hour and were washed three times in TBST for 15 minutes. Immuno-labelled protein bands were visualised using the ECL+ Chemiluminescence kit (Amersham
Biosciences) and exposure to Hyperfilm™ ECL Chemiluminescence film (Amersham
Biosciences) for periods of between 20 seconds to 20 minutes. The film was developed using an AGFA CP100 film processer.
93 4.2.6 Bioluminescence Resonance Energy Transfer Assays
4.2.6.1 Production of BRET Expression constructs
Full-length human cDNA for NDRG1 was amplified with sense and antisense primers harboring unique EcoRI and XhoI sites following the PCR conditions in section
4.2.3.1. The fragment was then sub-cloned to be in-frame with either Rluc or EYFP in the
Renilla luciferase expressing vector pcDNA3.1-Rluc or the enhanced yellow variant of
GFP, pcDNA3.1-EYFP (both vectors gifts from K. Eidne, 7TM laboratory, WAIMR).
Constructs with either N-terminal and C-terminal expression of Rluc or EYFP were produced. Full-length expression constructs of human APOA1, APOA2 and mouse Apoa1 were produced as detailed in the preceeding chapter (section 3.2.1 to 3.2.4).
4.2.6.2 BRET Assay
For the BRET assay, a total of 1µg of plasmid DNA (in a 1:3 or 1:4 ratio of Rluc construct to EYFP construct) was co-transfected into COS7 cells grown in 6-well plates.
After 48 hours of incubation the cells were detached with 1x trypsin/PBS and washed twice in PBS. Approximately 50 000 cells per well were distributed in a 96-well microtiter plate and incubated with 5µM coelenterazine (h form) (Molecular Probes). Repeated readings were taken over a 15 minute period using a custom-designed BRET instrument (Berthold,
Australia) that allows sequential integration of the signals detected in the 440-500 nm and
510-590 nm windows. BRET ratios were normalised for expression of EYFP to take into account any differences in the efficiency of expression between EYFP-only and EYFP- fusion constructs. The BRET ratio is defined as [(emission at 510-590) (emission at 440-
500) × cf]/(emission at 440-500), where cf corresponds to (emission at 510-590/emission at 94 440-500) for the Rluc construct expressed alone in the same experiment. Flow cytometry was used to monitor expression of EYFP-tagged constructs in a Becton Dickinson
FACSCaliburTM (San Jose, CA). Cells were gated to exclude dead cells and debris, and analysis was typically carried out on 10,000 events. The data were analysed using
CellQuestTM.
95 4.3 Results
4.3.1 NDRG1 and APOA1 co-localise in transfected mammalian cells
To determine whether NDRG1 and APOA1 share the same cellular compartment, co-localisation experiments were performed in human hepatoma HepG2 cells transiently co-expressing NDRG1/HIS and APOA1/EYFP recombinant proteins. The use of tagged constructs was necessary at this stage of the investigation as suitable antibodies capable of detecting the endogenous NDRG1 and APOA1 proteins were not available. At 48 hours post-transfection, confocal analysis showed that the NDRG1/HIS fusion protein was localised in the nucleus, perinuclear regions and in a punctate pattern throughout the cytoplasm (Figure 4.3A). NDRG1/HIS was also detected on the cell membrane of some cells. This pattern of staining and localisation are similar to reports of NDRG1 translocation between the nucleus, cytoplasm and cell membrane (36,54,57).
APOA1/EYFP was localised to perinuclear regions and was also visible in a punctate pattern in the cytoplasm (Figure 4.3B) and is consistent with its association with the trans-
Golgi network and secretory granules (139,144). Co-localisation of NDRG1 and APOA1 was observed throughout the cytoplasm and perinuclear regions (Figure 4.3D) but not on the cell membrane.
96 A B
NDRG1-HIS APOA1-EYFP
C D
DAPI Merged
Figure 4.3. Co-localisation of NDRG1 and APOA1 in transfected HepG2 cells. HepG2 cells were co-transfected with (A) HIS-tagged NDRG1 and (B) EYFP-tagged APOA1 and fixed and stained 48 hours post-transfection. (D) Merged overlay of confocal images (A) and (B) show NDRG1-HIS and APOA1-EYFP co-localise in a punctate pattern throughout the cytoplasm as indicated by yellow colour. (C) DAPI staining of DNA was used to locate the nucleus.
97 4.3.1 Endogenous NDRG1 and APOA1 co-localisation in mammalian cells
As introducing and over-expressing artificial constructs into cells may not reflect the true in vivo localisation of endogenous proteins, it was necessary to determine if the native NDRG1 and APOA1 proteins co-localise in cultured cells and tissue. A number of antibodies targeted against NDRG1 were available throughout the course of this project and included two commercially-produced antibodies; one raised against a peptide encoding the unique C-terminal repeat of the protein and the other targeted to the N-terminal region.
Subsequent testing of these antibodies on Western blots by our laboratory and colleagues at the Neurogenetics Laboratory, Academic Medical Center, Amsterdam, showed that both antibodies cross-reacted with other proteins, specifically with the closely related NDRG3 protein and would not be suitable for immuno-staining experiments. Attempts made in our laboratory to raise and purify an antibody targeted to the C-terminal half of NDRG1 in white New Zealand rabbits were unsuccessful. Recently, an affinity-purified NDRG1- specific antibody suitable for use on tissue sections was produced and kindly provided by
K. Kokame and T. Miyata from the National Cardiovascular Center Research Institute,
Japan and was utilized for co-localisation experiments in conjunction with an anti-APOA1 antibody obtained from the Southwestern Medical School, Center for Biomedical
Inventions, University of Texas, USA.
These two antibodies were used to investigate whether the proteins co-localise in the human hepatoma-derived HepG2 cell-line. This cell-type was chosen because of the known high expression levels of both APOA1 and NDRG1 in the liver (54) and the possible role of the interaction in the mechanism of cholesterol transport. The specificity of the antibodies in HepG2 cell lysates was determined by Western blotting, which revealed single bands corresponding to NDRG1 (43kDa) and APOA1 (28kDa) respectively (Figure 98 4.4F). Immunocytochemistry and confocal microscopy revealed extensive co-localisation of NDRG1 and APOA1 in perinuclear regions and throughout the cytoplasm in a punctate pattern (Figure 4.4C). This pattern of colocalisation was similar to that seen in the transfected cell experiments detailed above. Strong NDRG1 staining was also detected in nuclei (Figure 4.4A).
99 A B C
NDRG1 APOA1 Merged
D E F Western blot
HepG2
NDRG1 (43kDa)
APOA1 (28kDa) Negative-FITC Negative-CY3
Figure 4.4. Co-localisation of endogenous NDRG1 and APOA1 in HepG2 cells. The expression of endogenous (A) NDRG1 and (B) APOA1 in HepG2 cells was detected using anti-NDRG1 and anti-APOA1 sera. (C) Merged overlay of panels (A) and (B) showed NDRG1 co-localises with APOA1 in a punctate pattern in the cytoplasm but was also localised without APOA1 in the nucleus. (D & E) Negative controls were incubated with secondary antibodies only. (F) Specificity of the NDRG1 and APOA1 antibodies were tested by Western blotting of HepG2 cell lysates where single bands corresponding to NDRG1 (43kDA) and APOA1 (28kDA) were detected.
100 4.3.2 Co-localisation of NDRG1 and APOA1 in human nerve
Having established co-localisation of NDRG1 and APOA1 in cultured cells we then sought to determine whether the proteins co-localise in vivo. Because the primary tissue affected in CMT4D is peripheral nervous tissue, an adult human sural nerve biopsy was used to determine the localisation of NDRG1 and APOA1, and to investigate if the proteins overlap, thus providing evidence of a possible interaction. Immunohistochemistry and confocal microscopy showed NDRG1 in multiple locations throughout Schwann cells, including strong expression in the cytoplasm, sometimes in a punctate pattern (Figure 4.5
A), and in other cases in a reticular pattern (Figure 4.5 B), as well as on the cell membrane
(Figure 4.5 C). No evidence of axonal localisation of NDRG1 was detected.
APOA1 was generally localised in a punctate pattern mainly in myelin (Figure 4.5,
D & E) and is consistent with its previous co-purification with the myelin fraction in the chick nerve (145). In rare instances APOA1 was found in the Schwann cell cytoplasm
(Figure 4.5 E, F & H) and it was in this cellular region in which co-localisation with
NDRG1 was detected (Figure 4.5 I).
101 NDRG1 is localised in the Schwann cell cytoplasm and the cell membrane A B C
NDRG1 (60x) NDRG1 (60x) NDRG1 (40x)
APOA1 is localised in the Schwann cell cytoplasm and in myelin sheaths
D E F
APOA1 (40x) APOA1 (60x) APOA1 (60x)
NDRG1 and APOA1 co-localise in the Schwann cell cytoplasm in human peripheral nerve G H II
NDRG1 (60x) APOA1 (60x) Merged (60x)
Negative controls H&E Stain L J K L
Negative - FITC Negative – CY3 H&E (20x) (Figure legend overleaf) 102 Figure 4.5. Confocal microscopy of human sural nerve sections stained for NDRG1 and APOAl. Immunostaining for NDRG1 shows that the protein is localised to multiple cellular compartments in Schwann cells including (A) punctate (arrow) or strong (arrow head) expression in the cytoplasm; (B) in a reticular pattern (arrows); and (C) on the cell membrane (arrow heads). (D & E) APOA1 was localised mainly to myelin sheaths (arrows) and (E & F, H) in the Schwann cell cytoplasm in a punctate pattern (arrow heads). (I) Merging of images (G) and (H) shows that NDRG1 and APOA1 co-localise in the Schwann cell cytoplasm as indicated by yellow colour. (J & K) Negative controls. (L) Hematoxilin and Eosin (H&E) stain of the sural nerve biopsy section used for the immunostaining.
4.3.3 Pulldown assays
As the co-localisation experiments revealed evidence of overlapping expression of
NDRG1 and APOA1, further assays were undertaken in an attempt to provide additional support that the proteins interact. In the first pulldown assay, recombinant NDRG1/GST protein, expressed and isolated from E.coli BL21 cells, was incubated with a lysate of
COS7 cells expressing a recombinant APOA1/HA protein (Figure 4.6). No binding between these proteins was detected under the conditions used in this assay. In a second pulldown experiment, a recombinant NDRG1/GST protein isolated and produced in SF9 insect cells, was incubated with a purified preparation of APOA1 (Sigma) under different incubation temperatures (4°C and 37°C) and concentrations (2µg and 5µg). Again, no binding between NDRG1/GST and APOA1 was detected (Figure 4.7). Another strategy used was to determine whether the native form of apolipoprotein A-I was necessary for binding to the NDRG1/GST recombinant protein. Tissue homogenates prepared from mouse sciatic nerve and kidney were used as a source of apolipoprotein A-I and were incubated with the NDRG1/GST recombinant protein. Although apolipoprotein A-I was detected in both the sciatic nerve and kidney lysates, the results of the pulldown assay were inconclusive (Figure 4.8). A negative bleached band migrating slightly higher than the
103 expected size of Apoa1 was detected in the pulldown lanes. The bleaching phenomenon is often indicative of excess of target protein, however neither subsequent stripping and reprobing of the membrane, nor reloading a diluted amount of lysate onto a new gel and transfer to a membrane showed that these bands correspond to apolipoprotein A-I.
104 1 2 3 A 75kDa NDRG1/GST 50kDa
B 35kDa
30kDa Apoa1/HA 28kDa
Figure 4.6. Western blot of the pulldown binding assay between NDRG1/GST recombinant protein (expressed and purified from E. coli BL21) and HA-Tagged Apoa1 (expressed and purified from COS7 cells) reveals no interaction A) upper panel probed with anti-GST antibody. B) lower panel probed with anti-HA antibody. Lane 1) Negative binding control: Lysate from COS7 cells transfected with pcDNA3/Apoa1-HA was incubated with glutathione beads only without the NDRG1/GST recombinant protein. Lane 2) Pulldown assay: Purified NDRG1-GST recombinant protein was incubated with COS7 cell lysate expressing Apoa1/HA. A band corresponding to Apoa1-HA recombinant protein is not detected in this lane signifying no binding with the NDRG1-GST fusion protein under these conditions Lane 3) Lysate of COS7 cells transfected with pcDNA3/Apoa1-HA that was used for the binding assay shows that a band corresponding to the correct molecular size of Apoa1/HA is expressed.
105 1 2 3 4 5 6 7 8 9 10 A 78kDa NDRG1-GST (66kDa) 53kDa
B 34kDa APOA1 (28kDa) 21kDa
Figure 4.7. Western blot of the pulldown assay between recombinant NDRG1-GST fusion protein (expressed and purified from Sf9 insect cells) and purified human APOA1 (Sigma) fails to detect the NDRG1 and APOA1 interaction. A) Upper panel probed with anti-NDRG1 antibody raised against the N-terminus of the NDRG1 protein. B) lower panel probed with anti-human APOA1 antibody. Lane 1) An aliquot of purified APOA1 was loaded as a probe control, where probing with the anti-APOA1 antibody revealed a band at 28kDa corresponding to the molecular weight of APOA1. Lane2) recombinant NDRG1-GST protein was incubated for 4 hours with 5µg of APOA1 at 4 ºC or Lane 3) 2µg of APOA1 at 4 ºC. or Lane 4) 5µg of APOA1 at 37ºC or Lane 5) 2µg of APOA1 at 37ºC. No apoliprotein A-I is detected in any of these lanes signifying that no binding with the NDRG1-GST fusion protein occurs under conditions used in this assay. Lanes 6 & 7) Negative binding control: 5µg and 2µg of APOA1 were incubated with the glutathione beads only and the beads were boiled and loaded on the gel. No non-specific binding of APOA1 to the beads was detected. Lanes 9 & 10) APOA1 is eluted in the pooled wash fractions from the incubation/pulldown experiments evident of lack of binding with the NDRG1-GST protein under the conditions used in this assay.
106 1 2 3 4
NDRG1/GST 60kDa probed with anti-GST antibody
B 34kDa
Apoa1 28kDa 23kDa probed with anti-Apoa1 antibody
Figure 4.8. Western blot of the Pulldown assay between NDRG1/GST recombinant protein (expressed and purified from Sf9 insect cells) and mouse sciatic nerve and kidney lysates. Lane 1 ) NDRG1/GST fusion protein (upper panel lane 1) incubated with mouse nerve lysate. Lane 2) Probe control: nerve lysate only (5µg loaded). Lane 3) NDRG1/GST fusion protein (upper panel lane 3) incubated with mouse kidney lysate. Lane 4) Probe control: Kidney lysate only (5µg loaded). Although bands corresponding to Apoa1 were detected in the nerve and kidney lysates (Lanes 2 and 4) no definitive bands corresponding to the expected molecular size of Apoa1 were detected in the lanes where the lysates were incubated with the NDRG1/GST recombinant protein (Lanes 1 & 3). Subsequent stripping and reprobing of panel B with anti-Apoa1 antibody in different concentrations failed to cross-react with the slower migrating white bands.
107 4.3.4 Co-immunoprecipitation
As the pulldown assays were reliant upon a recombinant form of NDRG1 as bait, and to circumvent the possibility that the GST moiety may cause misfolding, potentially interfering with binding sites necessary for the apolipoprotein interaction, co- immunoprecipitations of endogenous NDRG1 and apolipoprotein A-I from different cell types and tissues were attempted. Protein lysates were prepared from human hepatoma
(HepG2) cells, mouse sciatic nerve and kidney. Figure 4.9 shows that while both NDRG1 and apolipoprotein A-I are expressed in these lysates, as revealed by immunoblotting using the respective antibodies, NDRG1 could not be successfully precipitated using the polyclonal affinity-purified antibody raised against the full-length NDRG1 protein.
Attempts to immunoprecipitate NDRG1 were also made using a commercially produced antibody (Zymed, USA) raised against a peptide consisting of the repeat C-terminal sequence of NDRG1 but were not successful (data not shown). In contrast, the human and mouse apolipoproteins A-I were succesfully precipitated from these lysates (Figure 4.9,
Lanes 1C, 2C & 3C) using anti-human APOA1 (Southwestern Medical School, Center for
Biomedical Inventions, University of Texas, USA) or anti-mouse Apoa1 (Rockland, USA) antibodies respectively.
Probing for the presence of NDRG1 in the lanes in which the apolipoproteins were successfully precipitated failed to detect any band corresponding to the expected molecular size of the NDRG1 protein. Thus it was concluded that NDRG1 does not co- immunoprecipitate with apolipoprotein A-I under the conditions used in these experiments, possibly due to antibodies blocking binding sites or the fact that the interaction is not stable enough to survive the handling steps of this technique.
108 Mouse sciatic nerve 1 A B C D IP IP Lysate Lysate NDRG1 Apoa1 Beads
IB NDRG1 (43kDa) NDRG1 IB Apoa1 (28kDa) Apoa1
Mouse kidney
2 A B C D IP IP Lysate Lysate NDRG1 Apoa1 Beads
IB NDRG1 (43kDa) NDRG1
Apoa1 (28kDa) IB Apoa1
Human HepG2 cells A B C D 3 IP IP Lysate Lysate NDRG1 Apoa1 Beads
IB NDRG1 (43kDa) NDRG1
IB Apoa1 (28kDa) Apoa1
Figure 4.9. Western blots of co-immunoprecipitation experiments of NDRG1 and apolipoprotein A-I from 1) mouse sciatic nerve 2) mouse kidney and 3) HepG2 cell lysates fails to show an interaction. Lanes 1A, 2A & 3A) Bands corresponding to the correct molecular sizes of NDRG1 (43kDa) and apolipoprotein A-I (28kDa) are detected in lystates prepared from mouse sciatic nerve, mouse kidney and human hepatoma (HepG2) cells (~5µg of protein lysate loaded). Lanes 1B, 2B & 3B) Immunoprecipitation (IP) using the anti-NDRG1 antibody, fails to precipitate NDRG1 from the lysates as determined by the absence of a 43kDa band after immunoblotting (IB) with the anti-NDRG1 antibody. Lanes 1C, 2C & 3C) Immunoprecipitation (IP) using the anti-mouse Apoa1 antibody (for mouse nerve and kidney lysates) or the anti-human APOA1 antibody (for HepG2 cells) reveals that apolipoprotein A-I is successfully precipitated from the lysates, however immunoblotting (IB) of the upper panel of these lanes with anti-NDRG1 antibody reveals no band corresponding to the molecular size of NDRG1, signifying that NDRG1 is not co- immunoprecipated with apolipoprotein A-I under these conditions. Lanes 1D, 2D & 3D) As a negative control the lysates were incubated with the protein G beads without the addition of antibody. For clarity, the heavy and light chains of the immunoprecipitating antibodies which migrate at 56 kDa and 26 kDa respectively, have been cropped from the images.
109 4.3.5 Bioluminescence resonance energy transfer (BRET) assays confirm the NDRG1 and apolipoprotein interaction
Because neither pulldown nor co-immunoprecipitation assays are suitable for detecting weak or transient interactions, the sensitive BRET technology was utilized to provide additional support that NDRG1 interacts with the apolipoproteins. An advantage of this assay is that the technique does not involve disrupting or lysing the cells, instead the interaction is measured in intact living cells.
Constructs consisting of the full-length open-reading frames of mouse Apoa1 and human NDRG1 tagged at either the N- or C-terminus with Rluc or EYFP were produced and co-expressed in COS7 cells for 48 hours (Figure 4.10). The BRET signal was measured immediately following the addition of coelenterazine. No signal was observed in control experiments with cells co-expressing the Rluc tagged constructs with pcDNA3.
A low BRET signal was generated when the Rluc/Apoa1 or Rluc /NDRG1 constructs were co-expressed with pcDNA3-EYFP, indicative of random proximity of the Rluc and EYFP moieties. Co-transfection with NDRG1/EYFP and Rluc/Apoa1 (Figure 4.10 A) produced a
BRET signal over and above that for EYFP alone, indicating interaction between the two protein moieties. The orientation of tagging was shown to be critical for detecting the interaction as no specific BRET signal was generated when the Rluc moeity was added to the N-terminal of NDRG1 and co-expressed with an Apoa1 construct C-terminally tagged with EYFP (Figure 4.10 B). These results suggest that tagging NDRG1 at the N-terminus may result in misfolding and cause inaccessibility of regions that were previously shown in
Chapter 3 as being important for the interaction with the apolipoproteins.
To investigate whether NDRG1 interacts with the human Apolipoproteins A-I and
A-II in the BRET assay, constructs encoding full-length human APOA1 and APOA2 tagged with Rluc at the N-terminus were produced and co-expressed in COS7 cells with
110 NDRG1 tagged C-terminally with EYFP for 48 hours. Again, as with the previous experiments, no BRET signal was observed in control experiments with cells co- expressing the Rluc/APOA constructs with pcDNA3. A low BRET signal was generated when one of the Rluc/APOA constructs was co-expressed with pcDNA3-EYFP, indicative of random proximity of the Rluc and EYFP moieties. Co-transfection with NDRG1/ EYFP and Rluc/APOA1 (Figure 4.11 A) or NDRG1/ EYFP and Rluc/APOA2 (Figure 4.11 B) produced a BRET signal over and above that for EYFP alone, indicating interaction between the two protein moieties. The results were reproducable in three independent wells transfected with a 1:3 or 1:4 ratio of Rluc APOA1 or APOA2 to NDRG1/EYFP plasmid DNA.
The BRET assay was also used to confirm the NDRG1 interaction with the prenylated rab acceptor protein 1 (Pra1) identified in the yeast two-hybrid library screen of the mouse peripheral nerve library (Chapter 3). A construct encoding the full-length mouse
Pra1, which shares 94% protein homology to the human PRA1 protein, was tagged at the
C-terminus with EYFP and co-expressed with full-length Rluc/NDRG1 in COS7 cells.
Positive BRET ratios over and above the background BRET signal were obtained indicating interaction (Figure 4.12). Unlike the apolipoprotein BRET experiments, appending the Rluc moiety to the N-terminus of NDRG1 did not interfere with binding to
Pra1 suggesting that the N-terminal region of NDRG1 is not involved in the Pra1 interaction. This is consistent with the yeast two-hybrid domain-mapping experiments where N-terminal NDRG1 constructs failed to interact with Pra1 (Chapter 3).
111 BRET : NDRG1 interacts with mouse Apoa1 in COS7 cells - tagging orientation is crucial to detecting the interaction
0.12
0.1
0.08 o i t a r 0.06 T E
R 0.04 B
0.02
0 Rluc Rluc Rluc Rluc Rluc Rluc Apoa1 Apoa1 Apoa1 NDRG1 NDRG1 NDRG1 + + + + + + pcDNA3 EYFP NDRG1 pcDNA3 EYFP Apoa1 EYFP EYFP
A B
Figure 4.10. NDRG1 interacts with mouse Apoa1 in living mammalian cells only when Rluc is tagged to Apoa1 and not NDRG1. A) BRET was measured in COS7 cells co-transfected with full-length mouse Rluc/Apoa1 fusion constructs with either pcDNA3 vector, or pcDNA3- EYFP as negative controls, or with pcDNA3-NDRG1/EYFP. Readings were taken 48 hours post-transfection. No BRET signal is generated when Rluc/Apoa1 is co-expressed with pcDNA3 alone. A small BRET signal is generated when Rluc /Apoa1 is co-expressed with EYFP alone indicative of random proximity of the Rluc and EYFP moeities. When the same amount of NDRG1/EYFP construct is co-transfected with Rluc/Apoa1 the BRET signal over and above that of EYFP alone (dashed line) indicates that NDRG1 interacts with Apoa1 in this assay. B) When the same experiments were performed when switching the Rluc moiety to NDRG1 and the EYFP moiety to Apoa1 no BRET signal above the background was generated indicating that tagging the proteins in this combination induces conformational changes to the proteins that abolish binding.
112 BRET : NDRG1 interacts with human APOA1 in COS7 cells
0.50 0.45 0.40 0.35 o i t
a 0.30 r
T 0.25
E 0.20 R
B 0.15 0.10 0.05 0.00 Rluc Rluc Rluc Rluc Rluc APOA1 APOA1 APOA1 APOA1 APOA1 + + + + + pcDNA3 EYFP EYFP NDRG1 NDRG1 (1:3) (1:3) (1:4) EYFP EYFP (1:3) (1:4)
BRET : NDRG1 interacts with human APOA2 in COS7 cells
0.50 0.45 0.40 0.35 o i t
a 0.30 r
T 0.25 E
R 0.20 B 0.15 0.10 0.05
Rluc Rluc Rluc Rluc Rluc APOA2 APOA2 APOA2 APOA2 APOA2 + + + + + pcDNA3 EYFP EYFP NDRG1 NDRG1 (1:3) (1:3) (1:4) EYFP EYFP (1:3) (1:4)
Figure 4.11. NDRG1 interacts with human APOA1 and APOA2 in living mammalian cells : Bioluminescence resonance energy transfer (BRET) assays. BRET was measured in COS7 cells co-transfected with full-length human A) APOA1 or B) APOA2 -Rluc fusion constructs with either pcDNA3 vector, or pcDNA3- EYFP as negative controls, or with pcDNA3-NDRG1/EYFP in ratios of either 1:3 or 1:4 (total DNA conc.1µg). Readings were taken 48 hours post-transfection. No BRET signal is generated when Rluc APOA1 or Rluc APOA2 are co-expressed with pcDNA3 alone. A small BRET signal is generated when Rluc APOA1 or Rluc APOA2 are co-expressed with EYFP alone indicative of random proximity of the Rluc and EYFP moeities (this background signal is indicated by dashed line). When the same amount of NDRG1 EYFP construct is cotransfected with Rluc APOA1 or Rluc APOA2 a positive BRET signal over and above that of EYFP alone indicates that NDRG1 interacts with APOA1 and APOA2 in this assay.
113 BRET : NDRG1 interacts with Pra1 in COS7 cells
0.50 0.45 0.40 0.35 o i t
a 0.30 r
T 0.25 E
R 0.20 B 0.15 0.10 0.05
Rluc Rluc Rluc NDRG1 NDRG1 NDRG1 + + + pcDNA3 EYFP Pra1 (1:3) EYFP (1:3)
Figure 4.12. NDRG1 interacts with the mouse Prenylated rab acceptor protein 1 (Pra1) in living mammalian cells : Bioluminescence resonance energy transfer (BRET) assay. BRET was measured in COS7 cells co-transfected with a full-length human Rluc/NDRG1 construct with either pcDNA3 vector, or pcDNA3-EYFP as negative controls, or with pcDNA3-Pra1/EYFP. Readings were taken 48 hours post-transfection. No BRET signal is generated when Rluc/NDRG1 is co-expressed with pcDNA3 alone. A small BRET signal is generated when Rluc/NDRG1 is co-expressed with EYFP alone indicative of random proximity of the Rluc and EYFP moeities (this background signal is indicated by dashed line). When the same amount of Pra1/EYFP construct is cotransfected with Rluc/NDRG1, a positive BRET signal over and above that of EYFP alone, indicated that NDRG1 interacts with Pra1 in this assay.
114 4.4 Discussion
Alternate methods confirm NDRG1 interaction with APOA1, APOA2 & Pra1
In this chapter, various approaches were used to provide evidence that NDRG1 interacts with the trafficking proteins identified in the yeast two-hybrid screen of the mouse peripheral nerve library. The BRET technology, which measures interaction in an intact and living cellular environment and is far more sensitive than the pulldown and co- immunoprecipitation techniques, confirmed that NDRG1 interacts with the lipid transporters Apolipoproteins A-I and A-II, as well as with the Golgi-associated trafficking protein Prenylated rab acceptor 1. The fact that the less sensitive methods did not detect the
NDRG1 and APOA1 interaction suggests that the binding is weak and not stable enough to survive the harsh washing steps of the pulldown and co-immunoprecipitation assays, or that the interaction is dependent on the membrane fraction, a subcellular compartment that was not isolated using these techniques. While the BRET assays provided unequivocal proof that the recombinant forms of these proteins can associate with NDRG1 in transfected living mammalian cells, additional support for an in vivo interaction between
NDRG1 and APOA1 was provided by the co-localisation of the proteins in cultured cells and in peripheral nerve.
NDRG1-APOA1 localisation patterns point to a role in trafficking
The immunohistochemical analysis has shown that NDRG1 is localised to many different cellular regions. In HepG2 cells, the protein was detected in the nucleus and perinuclear regions, as well as in a punctate distribution throughout the cytoplasm. In peripheral nerve, localisation was confined to the Schwann cell, consistent with previous findings in nerve (10,26,68). Strong expression was detected in the cytoplasm, in some cases in a reticular-like pattern or on the cell membrane. The multiple sites of intracellular
115 localisation are consistent with previous studies in various cell-types and tissues (36,38) and are thought to be indicative of a role as translocating enzyme or signalling molecule.
Our localisation and interaction data build upon these concepts by linking the protein to a role in trafficking pathways. This is supported by the localisation of the protein to perinuclear regions of the cell that house trafficking organelles such as the Golgi and in a punctate pattern that is consistent with association with secretory or vesicle structures. It is in these regions that colocalisation with APOA1 was found.
NDRG1 interacts with trafficking proteins
Prior to this investigation, knowledge about the role of NDRG1 was derived from in vitro expression studies in response to different reagents and states in cell-lines very different from peripheral nerve (36-39,43,45,47,48,146-150). It is through these studies that NDRG1 has been linked to a number of different pathways and cellular processes resulting in proposed functions that are hard to reconcile with the clinical phenotype of
CMT4D. A common theme emerging from our interaction data is that NDRG1 may be play a role in the inter-related pathways of lipid transport and cellular trafficking.
Intracellular trafficking of proteins and lipids is a complex process and relies upon a number of regulated and coordinated steps including packaging and sorting of specific vesicle populations, the directional movement of the vesicles through the cell, and the fusion of the vesicles with specific cellular compartments.
A vast array of proteins have been implicated in trafficking pathways, some of which include membrane receptors and fusion proteins such as SNARE (Soluble N- ethylmaleimide-sensitive factor attachment protein receptor), NSF (N-ethylmaleimide- sensitive fusion protein) and SNAP (Soluble NSF-Attachment Protein); vesicle and coat proteins like VAMP (Vesicle-Associated Membrane Protein) and Clathrin; regulatory
116 GTPases such as Ras, Rab and Dynamin; as well as cytoskeletal proteins that provide the scaffold on which vesicular traffic moves. Adding to this complex and dynamic environment is the regulatory effect of cholesterol upon vesicle transport and protein sorting, signal transduction and membrane composition (151,152).
Our interaction data and the results of other researchers have allowed us to construct a protein-protein interaction network (Figure 4.13) that shows that a number of
NDRG1-interacting proteins converge in common trafficking pathways. These interactors include the Heat Shock Cognate 70kDa protein (Hsc70) (153) which plays a role in membrane fusion through its ATPase activity leading to uncoating of clathrin coats from vesicles (154), and PICK-1 and p47 (153) which also regulate vesicular and plasma membrane fusion by interacting with NSF-SNAP and functionally similar complexes
(155). The recently identified NDRG1-interaction with the non-structural viral protein
NS5A (156), which enters cells via endocytotic mechanisms and also interacts with the apolipoproteins A-I and A-II (139), provides further evidence for NDRG1-association with vesicular trafficking machinary.
Prenylated Rab Acceptor 1 (PRA1), identified and confirmed as a NDRG1- interacting protein in our study, is required for transport vesicle formation from the Golgi complex and is involved in the recruitment of proteins necessary for cargo sequestration and subsequent vesicle docking and fusion (122,123,157). PRA1 interacts with a number of Rab and SNARE proteins, including Rab7 and VAMP2. Rab7, a member of the family of small GTPases which regulate transport between cell organelles (158), plays a crucial role in late endosomal traffic (159). Reticulon 1-C, another NDRG1-interacting protein, suggested by our Y2H screening, belongs to a family of endoplasmic reticulum membrane proteins, related to axonal regeneration and apoptosis (160). RTN1C has been shown to
117 Hsc70 Herpud1 CSP p47 NDRG1 NS5A Pick1 APOA2
Rab7 APOA1 Pra1 RTN1C
SYN13 BACE1 PS1 ABCA1 VAMP2 APP Key
indicates direct interactions detected in this study Membrane trafficking and regulatory proteins indicates direct interactions detected in other studies Lipid transport proteins indicates association as a protein complex Amyloid beta forming proteins ER stress response protein Viral protein Protein Notes Refs ABCA1 - ATP binding cassette 1 Lipid efflux (161,162) APOA1- Apolipoprotein A1 Lipid transporter (139,162,163) APOA2 - Apolipoprotein A2 Lipid transporter (139) APP - Amyloid precursor protein Cleaved to form amyloid beta (163,164) BACE1 - β-secretase APP-cleaving enzyme 1 Formation of amyloid beta (164-166) PS1 - Presenilin 1 Formation of amyloid beta (164,165) Herpud1- Homocysteine endoplasmic reticulum Upregulated with NDRG1 by homocysteine, responsive ubiquitin-like domain 1 part of retro-translocation transport complex (103,141,167) PRA1 - Prenylated rab acceptor protein 1 Membrane trafficking of Rab/Ras proteins (122,123) Rab7 - Ras associated protein 7 Mutated in CMT2B (122,129) Hsc70 - Heat shock cognate protein 70 Membrane trafficking/chaperone protein (153,168) RTN1C - Reticulon 1C Interacts with membrane fusion proteins (119,166) Syn13 - Syntaxin 13 Vesicle-membrane fusion protein (119,161) VAMP2 - Vesicle associated membrane protein 2 Vesicle-membrane fusion protein (119,156) CSP - Cysteine string protein Vesicle-membrane fusion protein (168) p47 – protein of MW 47kDa (ATPase) ATPase involved in membrane fusion (153) Pick1 - Protein Interacting with C Kinase 1 Regulator of membrane fusion (153) NS5A - Hepatitis C non structural viral protein Enters cell via endocytosis (139,156)
Figure 4.13 Protein-protein interaction network links NDRG1 to membrane trafficking and lipid transport pathways. Interaction data was obtained from the yeast two-hybrid library screens performed in this investigation and searches of the NCBI PubMed database, the Biomolecular Interaction Network Database (BIND) (available at http://www.bind.ca/) (169) and the Database of Interacting Proteins (DIP) (available at http://dip.doe-mbi.ucla.edu/) (170).
118 interact directly with SNARE proteins (119), including the PRA1 interactor VAMP2, and syntaxin 13. In turn, syntaxin 13 is an interacting partner of ABCA1, the ATP-binding cassette transporter, whose function in the efflux of cholesterol from cells to APOA1/HDL particles is closely associated with vesicular transport (161). APOA1 also interacts with
Presenilin-1 (PS1) which along with the endosomal and lipid raft-resident β-secretase
(BACE1) cleaves Amyloid Precursor Protein (APP) to form β-amyloid peptides, implicated in the neuropathogenesis of Alzheimer’s disease (AD)(171). APP interacts with the NDRG1–interactor Herpud1 whose upregulation results from the activation of the unfolded protein response (UPR) pathway and subsequent dysregulation of genes involved in triglyceride and cholesterol biosynthesis (172) and has recently been shown to be part of a complex involved in the ubiquitylation and transport of misfolded proteins (103).
Dysregulation of cholesterol synthesis and/or transport has been proposed as one possible mechanism underlying the pathogenesis of AD (173,174). The converging network of
NDRG1 interactions and their relationship to cell trafficking and lipid transport, involvement in other disorders affecting nerve, and their possible regulation by cholesterol, provides an important basis for our understanding of the biological role of NDRG1 and the neuropathogenesis of CMT4D disease.
Trafficking and lipid transport defects are a cause of CMT disease and other neuropathies
The inter-relationship between cellular trafficking and lipid transport is crucial for normal nerve function, as illustrated by the presence of peripheral neuropathy in genetic disorders of cholesterol transport, such as APOA1 deficiency (136,175), Tangier disease
(137,176,177), and Niemann-Pick disease type C (138). Aberrant trafficking is emerging as a major pathogenetic mechanism in some forms of CMT disease. Examples include
CMT2B caused by mutations in the PRA1-interacting protein Rab7 (129) where the
119 underlying defect may be related to disrupted lipid transport given the role of Rab7 in targeting sphingolipids to the Golgi, and its ability, when overexpressed, to correct the lipid trafficking defects and cholesterol accumulation in Niemann-Pick C cells (130). A defect in vesicular transport of lipids and/or proteins, resulting from altered phosphoinositol metabolism, has been proposed to play a major role in CMT4B1 pathogenesis (178) caused by mutations in Myotubularin-related protein-2 (MTMR2)
(178,179). Aberrant membrane trafficking may also underly CMT1C, resulting from mutation in Small Integral Membrane Protein of Lysosome/late Endosome (SIMPLE) which encodes an adapter protein that retains ubiquinated clathrin-enriched proteins in the endosome (180). A dominant intermediate form of CMTB is caused by mutation in the
Dynamin 2 (DYN2) gene (181) which encodes an ubiquitously expressed GTPase protein that has been implicated in many aspects of cell trafficking, including clathrin-mediated endocytosis, internalization of caveolae, synaptic vesicle recycling, and vesicular trafficking to and from the Golgi complex (182).
Mutations in genes encoding motor and cytoskeletal proteins upon which vesicular traffic moves have also been implicated in the pathogenesis of CMT disease. These include
CMT2A caused by mutation in the Kinesin family member 1Bβ (KIF1B) gene encoding a microtubule motor protein that functions in the anterograde transport of synaptic vesicle precursors (183), and the GTPase Mitofusin 2 (MFN2) which regulates mitochondrial membrane fusion and architecture (184). CMT2L and CMT2F are caused by mutations in
HSPB8 (185) and HSPB1 (186) which encode the Heat Shock Proteins 22kDa (HSP22) and 27kDa (HSP27) respectively, and may play a role in organizing the neurofilament network in nerve which is important for maintaining the axonal cytoskeleton and transport.
Why aberrations to these widely expressed proteins involved in various aspects of cell trafficking result in phenotypes that are restricted to peripheral nerve is not clear but
120 are likely to be related to the specialised trafficking needs of Schwann cells and axons.
Myelinating Schwann cells have a high demand for the mobilisation of protein and lipid components for producing and maintaining the myelin membrane, and an intact axonal transport mechanism is necessary for the delivery of components between the nerve cell body and distal regions of the axon. The survival of either cell type is dependent upon the correct functioning and communication with the other and this relationship may be particularly vunerable. Cholesterol homeostasis in this system is extremely important as it is not only a major constituent of the myelin membrane (187), but is also a component of intracellular and plasma membranes where its concentration may influence a number of cellular pathways in nervous tissue including cell signalling, adhesion, axonal guidance and vesicular trafficking (188). The fact that NDRG1 interacts with the apolipoproteins, as well as trafficking proteins that function in pathways tightly regulated by cholesterol, implies that aberrant cholesterol homeostasis may play a part in the pathogenesis of
CMT4D disease. This may possibly affect Schwann cell trafficking, resulting in abnormal targeting of lipids/proteins to the myelin membrane and/or impaired SC-axonal communication. The neuropathological features of HMSNL support this: intracytoplasmic
Schwann cell and axonal inclusions point towards a defect in trafficking, and repeated failed demyelination/remyelination events preceding axonal loss, indicate a breakdown in communication between Schwann cells and axons. Another feature of the disorder is hypomyelination and failure of compaction which may be the direct result of in a defect in the delivery of components essential for the formation and maintenance of the myelin membrane.
Given the lack of a suitable in vitro model of myelination, the functional significance of the NDRG1 and apolipoprotein interaction in peripheral nerve and whether it is related to the pathology of CMT4D disease can only be inferred based upon the
121 current knowledge of the roles and expression patterns of the proteins in this tissue type.
Neuropathological observations in NDRG1 knockout mice ((26) & described in Chapter 6) which present with a demyelinating neuropathy resembling HMSNL, suggest that NDRG1 is involved in the maintenance of the myelin membrane, once initial myelination is established. This is a role which is also supported in nerve crush studies (68,69) in normal mice, which show that NDRG1 is highly expressed in terminally differentiated Schwann cells at a time when remyelination is occuring. A role in myelination has been proposed for apolipoprotein A-I where the protein is detected in nerve following experimental crush injury (189), most likely functioning in the transport of lipids/cholesterol needed for myelin biogenesis and repair. It is also expressed in astrocytes in the human CNS (190), in neurons of rat spinal chord (191), and also during development of the chick PNS during active myelination, where it has been shown to co-purify with myelin and cytoplasmic subfractions (145). According to extensive literature searches, our study is the first to investigate APOA1 expression in human peripheral nerve, where we have found localisation to the myelin membrane and in rare instances in the Schwann cell cytoplasm.
It is in the cytoplasmic region where colocalisation with NDRG1 was detected. Given that the proposed role of the apolipoproteins in nerve are related to the transport of myelin components, a defect in intracellular trafficking crucial for normal myelin maintenance and axonal survival may be a pathogenetic mechanism underlying CMT4D disease.
NDRG1 interaction with the apolipoproteins points towards a role in the general mechanisms of lipid transport/metabolism.
While a severe Mendelian disorder of the PNS is the immediate drastic consequence of NDRG1 deficiency, the interaction with the apolipoproteins also points towards involvement in the more general processes of HDL-mediated cholesterol
122 metabolism. Current knowledge about the function of APOA1 focuses upon its role as an extracellular acceptor and transporter of cholesterol and lipids. Its intracellular function has not been extensively studied, however it is likely that it also plays role in the movement of lipids in the intracellular environment. In HDL-producing cells, previous investigations have found that APOA1 is associated with perinuclear regions, co-localising with the
Golgi-complex (139,144) and secretory granules (144), where it can undergo lipidation with cholesterol and phospholipids (192,193). As our co-localisation data in HepG2 cells supported an interaction with APOA1, a role for NDRG1 in the general processes of cholesterol transport and HDL formation was considered and is explored in the following chapter.
123 CHAPTER 5 – EFFECT OF NDRG1 DEFICIENCY ON EXTRACELLULAR CHOLESTEROL TRANSPORT
5.1 Background
Apolipoprotein A-I, as a component of the HDL particle, is a key player in maintaining cholesterol homeostasis through a mechanism known as Reverse Cholesterol
Transport (RCT) (reviewed in (131-133)). The pathway involves the efflux of cholesterol from peripheral cells, integration into and remodelling of HDL particles, and delivery and uptake at the liver (Figure 5.1). The role of APOA2 in this process is still largely unknown
(194) but it may have a modulating effect by changing the conformation of HDL particles and their ability to invoke efflux (195). The mechanisms of APOA1-mediated efflux include interactions with specific lipid domains in the cell membrane (196-198) as well as with the Scavenger Receptor B1 (SR-B1) (199,200) and with the ATP-binding cassette A1
(ABCA1) (201).
Aberrations to components of the RCT pathway and HDL metabolism can contribute to variation in plasma HDL levels which is a known risk factor in the development of atherosclerosis (202-206). HDL levels are thought to be the end product of the interaction between various genes and environmental effects, with genetic factors estimated to contribute about 70% of the variance (207). As with other complex traits, the identification of the genes that contribute to variation in HDL levels is far from complete.
Rare mutations in genes, encoding the known major proteins involved in reverse cholesterol transport (RCT) and HDL metabolism can cause Mendelian disorders with drastic reductions in HDL levels and increased risk for cardiovascular disease (CVD), for example familial APOA1 deficiency (208,209); Tangier disease, caused by mutations in
ABCA1 (137,176,177); and lecithin cholesterol acyltransferase (LCAT) deficiency (210). A recent study has demonstrated that rare mutations in some of these genes can also
124 FC CE FC
PERIPHERAL CELL LDL-R CE ABCA1 APOA1 UPTAKE EFFLUX FC KIDNEY ApoB/VLDL APOA1 APOA1 FC Preβ1-HDL CUBILIN
FC Preβ2-HDL LCAT α3-HDL
CE
α 2-HDL
CE
α1-HDL Key
ABCA1 - ATP binding cassette transporter A1 APOA1 - Apolipoprotein AI CE - cholesterol ester ApoB/VLDL CETP - Cholesterol ester transfer protein FC - Free cholesterol CETP HDL- High density lipoprotein UPTAKE PLPT CE CE HL - hepatic lipase HL LCAT - Lecithin:cholesterol acyltransferase LDL-R SR-B1 LDL-R - low density lipoprotein receptor LIVER Pl - phospholipid Hepatocyte PLTP - phospholipid transfer protein SRB1 - Scavenger receptor type B1 lipid-free CE CE VLDL – very low density lipoprotein APOA1 CE lipidated Pl APOA1 GOLGI
Figure 5.1. The reverse cholesterol transport pathway The RCT pathway involves the efflux of cholesterol by a lipid poor APOA1/HDL species known as preβ1-HDL which results in the formation of preβ2-HDL. The cholesterol in preβ2-HDL is then esterified by the enzyme lecithin-cholesterol acyltransferase (LCAT) to form α3-HDL. The α3-HDL species acquire more cholesterol from preβ-HDL that transforms these particles into larger α2-HDL and α1-HDL. The next step involves the transfer of cholesterol esters (CE) to apoB-containing lipoproteins via the action of the plasma cholesterol ester transfer protein (CETP) or the selective uptake of CE via SRB1 receptors. The particles are then remodelled into smaller α3-HDL and lipid- free APOA1. APOA1 can then be re-lipidated by cellular phospholipid and cholesterol to reform preβ1-HDL particles, or can be taken up by the kidney for metabolism and excretion. Hepatocytes contribute to the HDL pool by the secretion of a lipid poor APOA1 and also a smaller fraction of newly formed APOA1 that is lipidated with cholesterol and phospholipid at the Golgi to form nascent or preβ1-HDL particles. The role of APOA2 in this pathway is not clear but may modulate the shape of HDL particles affecting their capacity to invoke efflux (195). (Figure adapted from (133) and information from references (192,193,211-213)). 125 contribute significantly to quantitative variation in HDL-C in the general population (214).
However, these are not sufficient to explain the variation, and genome scans and quantitative trait loci (QTL) mapping have suggested the involvement of a large number of additional, unidentified genes. QTL mapping of HDL-C levels has identified 27 candidate regions in mice and 22 in humans (207) and, as with other complex traits, identifying the underlying genes has been difficult.
The identification of NDRG1 interaction with the two major protein components of
HDL and its colocalisation with APOA1 in the HDL-producing cell-line HepG2 suggested that the gene may be a possible functional candidate contributing to variation in HDL-C levels. To test these hypotheses, the aims of our follow-up investigations were as follows:
To determine whether the chromosomal locations of the human NDRG1 and the
mouse Ndrg1 loci coincide with known HDL-QTLs.
To utilize HMSNL as a human knockout model to investigate whether the
R148X genotype is associated with differences in blood lipid levels.
126 5.2 Methods
5.2.1 Literature and database searches
Searches of the NCBI PubMed database (available at http://www.ncbi.nlm.nih.gov/) and Mouse Genome Informatics database (available at http://www.informatics.jax.org/) (215) were carried out to identify HDL-QTLs falling within proximity of the chromosomal locations of human NDRG1 (Chr 8) and mouse
Ndrg1 (Chr 15). The positions of the markers underlying these HDL-QTLs were compared to the physical positions of the human NDRG1 and mouse Ndrg1 genes according to the current NCBI sequence map (build 35, version 1). Genetic distances were determined according to the Marshfield map for human NDRG1 and the MGI map for mouse Ndrg1.
5.2.2 Biochemical, clinical & statistical methods
The collection and analysis of data described in this section was performed by our
European collaborators and the Lipoprotein Research Unit, WAIMR, Royal Perth Hospital.
Venous blood samples were collected for analysis from individuals residing mostly in the Roma neighbourhood in Lom, Bulgaria, as part of the R148X carrier testing programme. During sampling, information was collected on weight, height, waist circumference, smoking habits and alcohol consumption. BMI was calculated as weight
(kg) divided by the square of the height (m). Testing for the R148X mutation was as described (10). The samples were analysed for serum glucose, creatinine, triglycerides, insulin, total cholesterol, HDL-cholesterol, and apolipoprotein concentrations. To minimise confounding effects such as shared genetic background and pathological states, first degree relatives and individuals with evidence of diabetes (fasting glucose >7.1 mmol/L) or renal failure (serum creatinine >120µmmol/L) were excluded from the study.
Plasma triglyceride and cholesterol concentrations were determined by standard enzymatic
127 methods. HDL-cholesterol was measured by an enzymatic colorimetric method using a commercial kit (Boehringer Mannheim, Mannheim, Germany). LDL-cholesterol was calculated using the Friedewald equation. APOA1, APOA2 and APOB were determined by immuno-nephelometry. HDL particle size was estimated as the HDL cholesterol/protein ratio, where protein = APOA1 + APOA2 concentrations. Statistical analyses were carried out using SPSS 10.1 (SPSS, Inc., Chicago, IL, USA). The results were expressed as mean+standard deviation (SD). Data were compared among the three genotype groups using ANOVA with Bonferroni adjustment. Statistical significance was defined at the 5% level using a two-tailed test.
128 5.3 Results
5.3.1 NDRG1 is a positional candidate for the HDL-QTL on chromosome 8q
The unexpected finding of NDRG1 interaction with the major HDL apolipoproteins prompted us to investigate whether the NDRG1 locus coincides with a HDL-QTL. Our literature searches identified two studies conducted in different human populations providing evidence for a HDL-QTL on the long arm of chromosome 8. Analysis of Finnish families with low HDL-C and history of familial combined hyperlipidemia resulted in a 2- point lod score of 4.7 for marker D8S1132 on chromosome 8q24 (216). Linkage of HDL-
C levels to 8q has been found also in a study of Mexican American families, with maximum multipoint LOD scores above 4 obtained for markers D8S1128 and D8S1990
(217). NDRG1 is located within the 144-145 cM genetic reference interval on 8q24.3,
(Figure 5.3) and is bracketed by the markers giving the highest lod scores in the Mexican
American study (217). The QTL interval contains 19 genes, none of which have an obvious functional relationship to HDL-C metabolism. A HDL-C QTL has also been identified in the conserved syntenic region on mouse chromosome 15 (218). Mouse Ndrg1, which encodes a protein with 93% identity to human NDRG1, resides within 5Mb of the
D15mit17 marker giving the highest LOD score in that study.
129 Figure 5.3. NDRG1 is a positional candidate for the human HDL-QTL on chromosome 8q24 NDRG1 is located between the markers D8S1128 and D8S1909 giving the highest LOD scores in Mexican-American families with low HDL-C levels (LOD peak adapted from Almasy et al., 1999). D8S1132, the marker giving the highest LOD score for low HDL-C levels in a Finnish sample with familial combined hyperlipidemia (Soro et al.,2002), lies centromeric to NDRG1 and D8S1128. Physical positions in base pairs are given according to the current NCBI sequence map (build 35, version 1) and genetic distances are according to the Marshfield map.
130 5.3.2 The effect of the R148X genotype on serum lipid, lipoprotein and apolipoprotein concentrations
Based on the detected interaction with apolipoproteins A-I and A-II and the localisation of the NDRG1 gene within an HDL-C QTL, we reasoned that in addition to its unique role in the PNS, NDRG1 may be involved in the general mechanisms of RCT. To test this possibility, we investigated the effect of the truncating R148X mutation on blood
HDL-C levels. Blood samples were collected from subjects of Roma ethnicity, belonging to communities with a high frequency of the founder mutation. After testing for the presence of the R148X mutation, the subjects were classified into the three genotype groups: wild-type homozygotes, R148X carriers and R148X homozygotes. We observed a statistically significant reduction in serum HDL-C concentrations in male R148X homozygotes compared with same sex wild-type controls (0.73 ± 0.05 vs 1.18 ± 0.27 mmol/L, P <0.01) (Table 5.1). A trend towards lower HDL-C levels in R148X carriers compared to wild-type homozygotes was also noted. There were no group differences in
APOA1 and APOA2 concentrations. However, HDL particle size, estimated as the HDL cholesterol/protein ratio, and the ratio of total cholesterol/HDL-cholesterol were significantly reduced in homozygous mutant individuals (P<0.05) compared with wild- type homozygotes. In females, the mean HDL-C levels were lower in R148X homozygotes (1.23± 0.18 mmol/L) compared to R148X heterozygotes (1.35± 0.15 mmol/L) and matched controls (1.31± 0.23 mmol/L) but the differences were not significant. The statistically significant group differences in lipid-related variables in the men in Table 5.1 were not related to variations in BMI, waist circumference, insulin levels or alcohol intake, which were similar in the wide-type carriers, R148X heterozygotes and
R148X homozygotes.
131 Table 5.1. Blood lipid parameters and genotype
WT/WT WT/R148X R148X/R148X ANOVA (n=27) (n=13) (n=11) (P-value)
Men Women Men Women Men Women Men Women (n=14) (n=13) (n=5) (n=8) (n=5) (n=6)
HDL–cholesterol (mmol/L) 1.18 ± 0.27 1.31 ± 0.23 1.04 ± 0.18 1.35 ± 0.15 0.73 ± 0.05** 1.23 ± 0.18 0.005 0.806 LDL-cholesterol (mmol/L) 2.33 ± 0.63 2.9 ± 0.77 3.02 ± 1.02 3.22 ± 0.52 2.59 ± 0.52 3.5 ± 0.72 0.192 0.562 Total cholesterol (mmol/L) 4.19 ± 0.65 4.81 ± 1.01 4.99 ± 1.14 5.40 ± 0.67 4.25 ± 0.25 5.34 ± 0.94 0.117 0.292 Triglycerides (mmol/L) 1.47 ± 0.48 1.22 ± 0.42 2.01 ± 0.79 1.79 ± 0.63 2.01 ± 0.93 1.32 ± 0.47 0.158 0.057 ApoA-I (g/L) 1.25 ± 0.21 1.39 ± 0.20 1.24 ± 0.16 1.52 ± 0.17 1.18 ± 0.33 1.35 ± 0.13 0.911 0.184 ApoA-II (g/L) 0.30 ± 0.05 0.31 ± 0.04 0.32 ± 0.02 0.32 ± 0.03 0.28 ± 0.09 0.30 ± 0.05 0.594 0.617
ApoB (g/L) 0.77 ± 0.18 0.89 ± 0.2 1.08 ± 0.33* 1.00 ± 0.22 0.95 ± 0.13 0.98 ± 0.25 0.042 0.510 ApoB/A-I 0.64 ± 0.20 0.64 ± 0.11 0.91 ± 0.36 0.66 ± 0.11 0.86 ± 0.35 0.72 ± 0.17 0.125 0.403 HDL-cholesterol/ protein 0.76 ± 0.10 0.78 ± 0.09 0.67 ± 0.09 0.74 ± 0.08 0.55 ± 0.15* 0.76 ± 0.15 0.030 0.714 Total cholesterol/ HDL- cholesterol 3.71 ± 1.07 3.68 ± 0.75 4.97 ± 1.70 4.02 ± 0.67 5.78 ± 0.22** 4.37 ± 0.79 0.005 0.180
Sample includes individuals with BMI < 30 kg/m2, TG < 3.5 mmol/L and tobacco consumption <200 mg/day, means ± SD; Compared with WT/WT by ANOVA with Bonferroni adjustments: *P <0.05; ** P<0.01
132 5.4 Discussion
The causal link between plasma HDL-C levels and the risk of developing atherosclerosis and cardiovascular disease is well established (202-206), however knowledge about the genes that contribute to the variation in these levels is far from complete. While quantitative-trait loci (QTL) mapping is a useful strategy to identify candidate regions underlying complex traits such as HDL-C levels, the low resolution of the technique makes pinpointing the causative genes a difficult task. This problem is compounded not only by our limited understanding of the factors influencing HDL-C regulation but also often by the lack of functional information about the genes residing within QTL regions.
Support for a putative role of NDRG1 as one of the factors contributing to variation in HDL-C concentrations came from the detection of its interaction with the two major apolipoproteins of HDL, the location of the NDRG1 gene within an HDL-QTL, and the observed significant decrease in blood HDL-C levels in male individuals homozygous for the truncating R148X mutation. Why this association is only appearant among males homozygous for R148X may be related to the effects of sex hormones on HDL-C levels: sex-specific HDL-QTLs have been identified in mice (218), and testosterone has been shown to have an effect on human HDL-C concentrations (219) as well as on NDRG1 expression (47). Whether a common hormonal mechanism between NDRG1 expression and HDL-C regulation can explain these findings is unknown.
The mechanism for the markedly lower serum HDL-cholesterol in the R148X homozygotes may involve any of a number of steps of the cholesterol transport or HDL metabolic pathway and include reduction in cellular cholesterol efflux, increased transfer of cholesterol ester from HDLs to apoB-containing lipoproteins and/or increased hepatic uptake of cholesterol esters from HDL (220,221). In the absence of significantly higher 133 plasma triglycerides and insulin resistance, the first mechanism appears the most likely.
Whether this also represents decreased concentration or functional activity of pre-β HDL, the first acceptor step in cellular cholesterol efflux remains to be elucidated. A recent investigation has shown that lipidation of apolipoprotein A-I, which results in the formation of nascent or pre-β HDL particles, also occurs intracellularly in hepatocytes and contributes to the total circulating HDL pool (193). The acquisition of cholesterol and phospholipids by a fraction of newly formed apolipoprotein A-I occurs in the Golgi-complex
(192,211,212,222,223), before its secretion out of the cell. Based upon our colocalisation data, the NDRG1 and APOA1 interaction in the hepatoma cell-line seems to occur in perinuclear regions rather than on the cell membrane (Chapter 4, Figures 4.3 & 4.4), suggesting that the basis for the low HDL levels in the R148X homozygotes may possibly be related to the formation of nascent or pre-β HDL particles. While our colocalisation data points towards an intracellular interaction between NDRG1 and APOA1, our findings of
NDRG1 localisation to the cell membrane suggests that an interaction with extracellular
APOA1 is also plausable. The current experiments were not designed to investigate this possibility and could be undertaken in the future. While the exact cellular mechanism of the
NDRG1 and APOA1 interaction and its functional significance are yet to be elucidated, it is perhaps significant that the NDRG1 protein is highly expressed in other HDL-producing and recycling tissues including the intestine and kidney (54). In a recent study, the
Hepatocyte Nuclear Factor transcription factor (HNF-4) was shown to bind to the NDRG1 promoter, implying an important role for NDRG1 in the liver (224).
The low HDL-cholesterol together with the high total cholesterol/HDL-cholesterol ratio and the small HDL particle size signify that males homozygous for R148X and possibly heterozygous carriers, would be at substantially increased risk of cardiovascular disease (225). Interestingly, NDRG1 has previously been linked to the pathogenesis of 134 atherosclerosis: it was originally identified as being upregulated by homocysteine (35,40), and also by lysophosphatidylcholine (41), a component of oxidatively modified lipoproteins. Both reagents induce in vitro changes analogous to the in vivo mechanisms of atherosclerosis, possibly through the activation of the unfolded protein response (UPR) pathway and the dysregulation of genes involved in triglyceride and cholesterol biosynthesis (172). The functional significance of this upregulation and its relationship to the findings of our study remain to be determined.
Whether partial deficiency of NDRG1 conferred by the heterozygous R148X genotype, is a predisposing risk factor in the development of early atherosclerosis among the Roma remains to be established and would rely upon the recruitment of larger numbers of heterozygotes in order to perform meaningful genotype/phenotype correlations. Current information about the major health indicators in the Roma is limited (226), but higher rates of cardiovascular morbidity and an increased risk of cardiovascular disease, due to lifestyle factors, marginization from the health-care system and genetic predisposition have been suggested (227,228). Whether the R148X allele is an additional genetic risk factor for cardiovascular disease in this group is yet to be determined.
Our biochemical data are based on the study of a rare mutation in an isolated founder population. The association of other sequence variants in NDRG1 with HDL-C variation in the general population merit investigation. Currently the NCBI SNP database identifies 3 SNPs in the N-terminal coding region of NDRG1 that are predicted to result in non-synonymous changes (H41R, M67V and M111L). While these variants are yet to be validated, if real, they occur close to, or within regions necessary for interaction with the apolipoproteins detected in this study (Chapter 3, Figure 3.17) and may perhaps affect binding with the NDRG1 protein. The search for these and other NDRG1 alleles that may be associated with atherosclerotic risk indicators such as low HDL and carotid intima 135 media thickness (CIMT), could form the basis for association studies in large cohorts, where information on blood lipids and CIMT have already been collected (229). This type of investigation could also be complemented by a study into the susceptability to the development of atherosclerosis in mouse models of Ndrg1-deficiency where factors such as diet and genetic background can be controlled. The molecular characterization of such a model is presented in the following chapter.
136 CHAPTER 6 - MOUSE MODELS OF NDRG1 DEFICIENCY
6.1 Background
Determining the functional significance and cellular mechanisms of the interactions identified in this study is a complex task and will rely upon suitable experimental models that replicate the lack of NDRG1. Recently, two Ndrg1 deficient mouse-models, both presenting with peripheral neuropathy resembling the human CMT4D disease phenotype, have become available and colonies of each type are currently being established in Western
Australia for use in future investigations.
The first mouse model was produced by Hisato Kondoh (Laboratory of
Developmental Biology) and Toshiyuki Miyata (National Cardiovascular Centre Research
Institute) in Osaka, Japan (26). These mice present with symptoms of a progressive demyelinating peripheral neuropathy by 3 months of age. Histological analyses have revealed that Schwann cell proliferation and initial myelination appear normal after birth, but by 5 weeks of age sporadic myelin degeneration becomes evident. As the disease progresses, ultrastructural observations include a large number of demyelinated or thinly myelinated axons, onion bulb pathology with SC processes, thin myelin sheaths, endoneurial collagenisation, and macrophage infiltration. The neuropathogenesis thus closely mimics the human CMT4D disease phenotype (23-25,101). No published information about cardiovascular pathology or blood lipid profiles of these mice currently exists.
Although Ndrg1 expression in these mice has been eliminated by deleting the promoter and exon 1 of the gene, Northern blotting has shown that limited leaky transcription occurs in this mouse (26) as a result of the preservation of the gene downstream of exon 2, including the initiation codon. This leaky expression may be a
137 confounding factor in the use of this model for the study of molecular events downstream of Ndrg1 where the effects of even a small amount of Ndrg1 protein are unknown.
A second mouse model was first identified by the Morahan laboratory at WEHI,
Australia and has resulted from a spontaneous deletion of a large fragment of the Ndrg1 gene arising in the diabetic NOD.Slc9a1b congenic mouse strain. The mice show a characteristic stretching of the rear limbs, especially when handled for examination. This feature inspired the mutation to be named stretcher (str), a nomenclature that will be used in subsequent descriptions of this model in this chapter. To preserve the mutation it has been bred onto the nondiabetic strain BALB/c by 10 generations of backcrossing females to
BALB/c males and maintained further by sib mating.
In order to map the str locus, the group at WEHI performed genome scans in affected mice, where a single-point LOD score of 23.9 was obtained on chromosome 15 for the marker D15Mit63. High resolution genotyping mapped the str locus to an interval of approximately 2cM between the markers D15Mit233 and D15Mit144. Using additional markers developed at WEHI, the region was further refined to a 3Mb interval containing 11 known gene transcripts, including Ndrg1. As the phenotypic features of the str mice suggested a neural deficit, Ndrg1 was an excellent candidate given its involvement in the neuropathogenesis of CMT4D disease (10,230). Sequencing of the Ndrg1 gene in the str mice revealed no polymorphisms, however in the course of this work, the WEHI group were unable to amplify exons 10, 11, 12, 13 or 14. This pointed towards a likely deletion of a large genomic fragment spanning these regions. Primers flanking exon 9 and 15 were designed where the exact break points were defined. It was revealed that over 5kb of genomic DNA encompassing exons 10 to 14 had been deleted.
To investigate the molecular basis of the deficiency and the suitability of the str mouse as a model of CMT4D, mice were sent to our laboratory for RNA and protein 138 analyses and also to Dr Rosalind King at the Department of Neurology, Royal Free
Hospital, UK for neuropathological investigations. The results of these experiments are presented in this chapter.
139 EXPRESSION STUDIES OF Ndrg1 IN THE STR MUTANT MOUSE
6.2 Methods
6.2.2 RNA isolation, cDNA synthesis and sequencing
To investigate the effect of the genomic deletion upon the Ndrg1 mRNA transcript, kidney tissue which is an abundant source of Ndrg1 RNA and protein (54,121,231), was used for the expression experiments. For RNA extraction whole kidneys from wild-type
BALB/c mice or BALB/str mutant mice were homogenised in 500µL of Triazol (Gibco) reagent according to manufacturers instructions. For cDNA synthesis 2µg of the RNA was reverse transcribed using 1 x Reverse transcription buffer (Promega), 1U of RNase inhibitor (Invitrogen), 2mM of dNTPs, 50ng/µL of random hexamers (Promega), and 8U of
MMLV-reverse transciptase (Promega) in a total volume of 20µL. Reaction mixes were incubated at 42°C for 60 minutes and the reaction stopped by heat inactivation at 95°C for
10 minutes. The cDNA was used as a template for the amplification of a PCR product spanning exon 6 to 15 of the Ndrg1 gene. The reaction consisted of 1x PCR buffer, 2.5mM
MgCl2, 0.5mM dNTPs, 1.5U Taq, 20ng of each primer (5'-
GAGGACATGCAGGAGATCAC-3' & 5'-CAGAGGCTGTGCGGGACC-3') and dH2O in a total volume of 50µL. PCR cycling conditions consisted of initial denaturation at 95°C followed by 40 cycles of denaturation at 95°C for 30 seconds, annealing at 55°C for 30 seconds and extension at 72°C for 45 seconds with a final extension step at 72°C for 7 minutes. The products were cleaned with PCR purification columns and sequenced using
BigDye Terminator chemistry.
6.2.3 Northern Blotting
For Northern blotting, 5µg of RNA was electrophoresised on a 1.2% agarose/formamide gel for 2 hours in MOPS buffer. The RNA was transferred to a
140 nitrocellulose membrane via capillary-wick blotting in SSCE buffer for 3 hours and the membrane was dried in a oven set at 80ºC for ten minutes. The RNA was fixed onto the membrane by a 4 minute exposure under a UV lamp (312nm). A DNA probe was constructed from a 233bp PCR product spanning exons 2 to 4 of Ndrg1 amplified from mouse kidney cDNA using primers 5'-GACCTCGCTGAGGTGAAGCC-3' & 5'-
GTGATCTCCTGCATGTCCTC-3'. The PCR product was labelled with 32P-dCTP
(Amersham) using a Random Primed Labelling kit (Roche) according to manufacturers instructions. The membrane was incubated in 3mL of Ultrahyb® hybridization solution
(Ambion) for 30 minutes at 42ºC and replaced with 5mL of fresh solution containing the denatured labelled probe (activity of 6.0 x 105 cpm/mL). Hybridization was carried out in glass tubes rotated at 42ºC for 24 hours. The membrane was then washed twice in 2x SCC,
0.1% SDS buffer pre-warmed to 42ºC for 10 minutes and twice with 0.1x SCC, 0.1% SDS buffer for 15 minutes. The membrane was wrapped in cling film and exposed to Medical
X-Ray film for 16 hours at –80ºC. The film was developed using an AGFA CP100 processer.
6.2.4 Western Blotting
For the isolation of total protein, whole kidneys were homogenised in RIPA buffer
(1% Nonidet P-40, 0.1%SDS, 0.5% Deoxycholate, 150mM NaCl, 50mM Tris pH 8.0,
10µg/mL aprotinin, 1mM PMSF, 1mM benzamidine, 0.1mM Na3VO4) and centrifuged at
13,000rpm for 20 minutes at 4ºC. The supernatant were transferred to a new tube and quantitated using a BCA Protein Assay kit (Pierce). 10µg of protein was loaded into wells of a 12% SDS-PAGE stacking gel and electrophoresed at 200V for 1.5 hours. The proteins were transferred to PVDF membranes by Western blotting at 30V overnight at 4°C. The membranes were probed first with the affinity-purified polyclonal antibody raised against 141 the full-length NDRG1 protein (26). After exposure and subsequent stripping, the membrane was then re-probed with an antibody directed against the N-terminus of the
NDRG1 protein. Immuno-labelled protein bands were visualised using the ECL+
Chemiluminescence kit (Amersham Biosciences) and exposure to Hyperfilm™ ECL
Chemiluminescence film (Amersham Biosciences).
142 6.3 Results
6.3.1 Analysis of the Ndrg1 RNA transcript in the stretcher (str) mouse
To determine whether the genomic deletion results in a lack of Ndrg1 mRNA,
Northern blotting using a Ndrg1-specific probe targeted to exons 2 to 4 was performed on kidney RNA isolated from str and wild-type mice. It was found that the str mice produce a shorter Ndrg1 mRNA species compared to the wild-type Ndrg1 transcript (Figure 6.1A).
The fact that a comparable quantity of Ndrg1 mRNA was detected suggested that the deficiency in the str mutant was not due to the decay of the mutant transcript but pointed towards a truncated Ndrg1 protein.
To investigate this possibility, RT-PCR was used to determine what effect the 5kb genomic deletion has upon the reading frame of the mRNA transcript. A RT-PCR product spanning exons 6-15 of Ndrg1 was amplified from kidney RNA isolated from str and wild- type mice. A significant size discrepancy between the RT-PCR product amplified from wild-type mice and the product amplified from str mice was detected after gel electrophoresis (Figure 6.2A). Sequencing of the RT-PCR products and in silico translation using the Sequence Navigator™ software program showed that exons 10 to 14
(297bp) of the Ndrg1 mRNA from the str mice were absent. The open reading frame was maintained with exon 9 spliced in-frame with exon 15 (Figure 6.2B).
143 Northern Blot A B
WT str WT str
28s Wild-type Ndrg1 Mutant mRNA Ndrg1 mRNA
18s
Probed with a RNA loading 32P-CTP-labelled 233bp control N-terminal fragment of Ndrg1 EtBr-stained DNA
Figure 6.1. Northern blotting of kidney RNA (A) Northern blotting of 5µg of kidney RNA from (WT) the wild-type and (str) the str mutant mouse using a 233bp probe spanning exons 2-4 of Ndrg1 shows the str mutant mouse produces a mRNA species shorter than the wild-type Ndrg1 mRNA. (B) Ethidium bromide-staining of the agarose gel prior to Northern transfer shows comparative intensities of ribosomal 28s and 18s in both lanes indicating equal amounts of RNA were loaded.
144 A RT-PCR – Ndrg1 WT str
750bp
500bp
250bp
B Chromatogram
WT Exon 9 Exon 10
str Exon 9 Exon 15
Figure 6.2 RT-PCR and cDNA sequence of the stretcher (str) mutation (A) Using Ndrg1 primers located in exon 6 and exon 15, RT-PCR was performed on cDNA prepared from kidney RNA from (WT) wild-type BALB/c and (str) the str mutant mouse. A 629bp product was detected in the wild-type. A significantly shorter product was amplified from the str cDNA. (B) chromatograms and translated protein sequences of Ndrg1 cDNA prepared from kidney RNA from (WT) wild-type and (str) the str mutant mouse show that the deletion results in the skipping of exons 10-14. Exon 15 is spliced in- frame with exon 9.
145 6.3.2 Analysis of the Ndrg1 protein in the stretcher (str) mouse
As the RNA expression experiments in the str mice detected a significant amount of the mutant Ndrg1 mRNA and also revealed that the deletion results in the preservation of the open-reading frame, it was feasable that a mutant form of the Ndrg1 protein could be present. According to in silico molecular weight (M.W) calculation (available at http://www.changbioscience.com/genetics/mw.html), the loss of exons 10-14 would result in a mutant Ndrg1 protein with a predicted molecular weight of 32 kDa. This is approxiamately 10-11 kDa smaller than the wildtype Ndrg1 protein (42-43kDa). To determine whether this smaller Ndrg1 protein is present in kidney tissue from the str mice,
Western blotting using two antibodies raised against different parts of the human NDRG1 protein, and that cross-react with mouse Ndrg1, were performed. Probing with a polyclonal antibody raised against the full-length protein (26) detected a band at the expected molecular weight for Ndrg1 (43kDa) in lysates prepared from the wild-type mice, but failed to detect any full-length Ndrg1 protein, nor any mutant Ndrg1 protein (predicted M.W = 32 kDa) in the kidney lysate prepared from the str mice (Figure 6.3A).
To investigate the possibility that the lack of detection of the mutant Ndrg1 protein was due to missing epitopes encoded by exons 10 to14, probing with an antibody directed against the N-terminal part of the protein was performed. This antibody cross-reacts with the closely related NDRG family member, NDRG3 (41 kDa), however literature searches indicated tha only a small amount of NDRG3 transcript is present in kidney (232).
Therefore this antibody would be suitable for specifically detecting Ndrg1 in this tissue lysate. Hybridisation of the Western blot using this antibody detected a band corresponding to the correct molecular size of Ndrg1 (43 kDa) in the lanes where the kidney lysates from wild-type mice were loaded (Figure 6.3B), but did not detect any full-length Ndrg1 protein, nor any truncated Ndrg1 protein in the str mice. 146 A B C
WT str WT str WT str
53kDa 53kDa
Ndrg1 Ndrg1 (43kDa) (43kDa)
34kDa 34kDa
23kDa 23kDa
Probed with Probed with antibody raised Protein loading control antibody raised against against N-terminal of Coomassie-stained full-length NDRG1 NDRG1
Figure 6.3. Western blotting of kidney lysates Western blotting and probing using (A) an anti-NDRG1 antibody raised against the full- length protein, or (B) an antibody raised against the N-terminal region, reveals that the Ndrg1 protein is detected in kidney lysates prepared from (WT) wild-type BALB/c but is absent in the lysate from (str) the str mutant mouse. No bands at the predicted molecular weight of the mutant protein (32kDa) were detected. The faint bands migrating at approximately 25kDa in both lanes in (B) are non-specific artefacts. (C) Coomassie-blue staining of the gel that was used for Western transfer verifies that equal amounts of protein were loaded.
147 Some faint bands migrating at approxiamately 25 kDa detected in both the wild-type and str lanes were deemed to be non-specific artefacts. The failure of two different antibodies, targeted to different parts of Ndrg1, to detect the full-length protein or the predicted shortened protein suggested that the basis for the deficiency in the str mouse is due to a complete lack of Ndrg1 protein. The lack of Ndrg1 may be due to inhibition of translation of the mutant RNA transcript or the destruction of the misfolded protein by the ER- associated protein degradation pathway.
6.3.3 Neuropathology of the stretcher (str) mice
Ultrastructural characterisation of nerves in the str mice during early developmental stages of the PNS including time periods of Schwann cell differentiation and myelination are currently being performed by Dr Rosalind King, at the Department of Neurology, Royal
Free Hospital, U.K. Her preliminary findings indicate that there are few abnormalities in the sciatic nerves at 2 weeks of age, but by 5 weeks moderately extensive pathological changes are evident despite the very mild clinical symptoms at this time-point. The ventral roots in particular showed extensive demyelination and many axons with abnormally thin myelin. Occasional fibres undergoing axonal degeneration could be identified. There were no signs of axonal regeneration even in the oldest mice examined (10 weeks of age)
(Figure 6.4B).
Electron microscopy showed pleomorphic adaxonal Schwann cell inclusions similar to those seen in HMSNL (23). In contrast to the Japanese Ndrg1-knockout mouse, endoneurial collagenization was not prominent and there was no evidence of onion bulb changes but the axonal changes and adaxonal Schwann cells are an additional feature.
Electrophysiological studies suggest that demyelination starts between weeks 4 and 5 consistent with the ultrastructural changes evident at this time-point. 148 A - Wt
B - str
40x
40x
Figure 6.4. Light micrographs of sciatic nerve in wild-type and str mutant mice A) Cross section of sciatic nerve of the wild-type (WT) mouse B) Sciatic nerve from a 10 week old str mouse shows degeneration (arrows) and thin myelin (arrowheads). (Images courtesy of RHM King).
149 6.4 Discussion
The availability of animal models that replicate the CMT4D phenotype is a vital resource for use in determining the neuropathological progression of the disease and the functional significance and cellular mechanisms of the interactions identified in this study.
Extensive characterisation of the neurophysiological and neuropathological features of the two models will provide important information about the progression of the disability and may also deliver some insight into the mechanisms of SC-axonal interactions necessary for normal PNS development and function.
A number of neuropathological hallmarks of HMSNL are replicated in the Ndrg1- deficent mice suggesting that they are an appropriate model to investigate the early manifestations and evolution of the disorder, something which is not possible in humans.
However thorough phenotypic characterisations need to be performed to determine their suitability for downstream investigations. A small percentage of str mice develop cataracts, a feature that has not been noted in patients with CMT4D disease (10,18,23-25), nor the
Japanese knockout mouse (26). Whether this is a direct result of Ndrg1 deficiency in this strain or due to another unidentified polymorphism in another gene is not known. Leaky expression in the Japanese Ndrg1-knockout may also be a confounding factor in the use of this model in some experiments where even the effects of a residual amount of Ndrg1 protein is unknown.
The mouse models will also be valuable in the testing of the current hypothesis, generated by in-vitro proteomics experiments of Ndrg1 involvement in cell trafficking, including the transport of lipids. The mice will provide a source of primary cell-lines that may be harvested and cultured for assays to measure the effect of lack of Ndrg1 upon lipid transport. For instance, cell-types traditionally used in cholesterol efflux assay such as
150 peritoneal macrophages, may provide a readily accessible model to investigate Ndrg1 involvement in cholesterol efflux. The proposed link to susceptibility to the development of atherosclerosis could also be investigated in the Ndrg1-deficient mice and may include a comparison of the characteristics of the two models with the view of further research into modifying factors and treatment strategies. This would include an analysis and comparison of the extent of aortic lesions (if any) after feeding with a high fat diet and the measurement of blood lipid profiles, including the various HDL subfractions. This may aid in pin- pointing which specific step of HDL formation is affected.
The mice may also be used to verify in vitro data and further investigations of the effect of Ndrg1 deficiency on the gene expression profile of peripheral nerve. They will provide a tissue source to confirm the expression of a number of up-regulated and down- regulated genes that were identified in Serial Analysis Gene Expression (SAGE) and macroarray experiments of cultured SCs from a HMSNL nerve biopsy previously performed by Dr David Chandler at the Amsterdam Medical Centre, Netherlands. Some of these genes include Caveolin-1, which encodes an integral membrane protein that acts as a scaffold for a number of signalling molecules including Ras-related GTPases and may also play a critical role in cholesterol trafficking in Schwann cells (233,234), and Semaphorin
3C encoding a protein involved in axonal guidance and growth cone collapse (235).
Utilizing peripheral nerve as a source of RNA and protein from the mice will determine whether the in vivo expression of these genes are indeed influenced by the presence or absence of Ndrg1 without the confounding effects inherent to cell culture systems.
151 CHAPTER 7 – SUMMARY AND CONCLUDING DISCUSSION
Since the identification of the founder R148X mutation as a cause of HMSNL (10) the contribution of other NDRG1 mutations to the pathogenesis of CMT4D has remained unknown. So too has a function of the protein which can adequately explain its role in the peripheral nervous system, and the pathology that results from its absence. This investigation has sought to address both these issues by first identifying and assessing the contribution of mutations in NDRG1 as a cause of peripheral neuropathy in a large sample of patients where the common CMT genes had been excluded. And secondly, to use an in vitro proteomic approach to identify NDRG1-interacting proteins to gain clues to its function with particular emphasis upon possible roles in peripheral nervous tissue.
The sequence analysis of NDRG1 among a large sample of patients presenting with various forms of peripheral neuropathy identified a novel exon skipping mutation, IVS8-
1G>A, which is only the second known NDRG1 disease-causing mutation. The large degree of genetic heterogeneity underlying CMT disease, coupled with the fact that different mutations in the same gene can result in different phenotypes, and that the neuropathological features of HMSNL pointed towards both SC and axonal involvement, lead us to adopt an inclusive approach in selecting the sample of patients for the sequence analysis. Thus patients presenting with either demyelinating or axonal forms and different modes of inheritance were included. Like R148X, the IVS8-1G>A mutation is recessively inherited and results in an early onset demyelinating neuropathy marked by clinical features of severe reduction in nerve conduction velocity and deafness, similar to HMSNL.
Our results indicate that aside from the founder R148X mutation, which was detected in two patients of Romani ancestry in our sample, other mutations in NDRG1 are a
152 rare cause of neuropathy, similar to other genes such as Early Growth Response Factor 2
(EGR2), Periaxin (PRX), and Myotubularin-Related protein 2 (MTMR2) (78,95-97). The gene should be considered for mutation analysis in patients presenting with clinical features of HMSNL and where the common CMT disease genes have been excluded. Mutation screening for R148X should especially by undertaken in cases where Romani ancestry is reported or suspected.
Delineating the molecular pathogenesis of neuropathies caused by NDRG1 mutations relies upon knowledge about the function of the protein. Prior to this investigation, knowledge concerning the role of NDRG1 was largely derived from studies of its expression in various cell-lines and tissues unrelated to peripheral nerve. This has resulted in diverse proposed roles that are hard to reconcile with the clinical features of
CMT4D disease, largely restricted to the PNS and replicated in the mouse models. This investigation, along with two recently published studies (153,156), are the first to identify
NDRG1–interacting proteins and point towards involvement in the complex and inter- related mechanisms of membrane trafficking and lipid transport. These are roles that are also supported by the in silico predictions of its protein domains linking NDRG1 to lipid metabolism and are perhaps more relevant to our understanding of its function in nerve.
Peripheral neuropathy is often part of the clinical phenotype of lipid transport disorders (136-138), and as discussed in Chapter 4 and summarised in Table 7.1 below, is
153 Table 7.1. CMT diseases caused by mutations in genes implicated in trafficking
Disease Gene Role in trafficking Reference Mutated CMTDIB DYN2 GTPase regulator of clathrin-mediated endocytosis, (181) internalisation of caveolae, vesicle recycling and trafficking to and from the Golgi CMT1C SIMPLE Ligand protein regulating transport of ubiquinated clathrin- (180) enriched proteins in the endosome CMT2A KIF1B Microtubule protein involved in anterograde vesicle (183) trafficking CMT2A MFN2 Regulator of mitochondrial membrane trafficking (184) CMT2B RAB7 Regulator of vesicle and sphingolipid transport (129) CMT2F HSPB1 Organization of neurofilament network thus maintaining (186) axonal transport CMT2L HSPB8 Organization of neurofilament network thus maintaining (185) axonal transport CMT4B1 MTMR2 Implicated in aberrant vesicle transport of lipids and proteins (178,179) due to altered phosphoinositol metabolism CMT4D ? NDRG1 Interacts with trafficking proteins - ApoA1/A2, PRA1, (153) and RTN1C, ARL6-IP, PICK1, HSC70 and p47 Chapter 3
the only obvious clinical phenotype in the growing number of CMT disease subtypes whose molecular basis involves genetic defects in cell trafficking. Our data suggest that
CMT4D disease may be a disorder of Schwann cell trafficking, resulting in abnormal targeting of lipids/proteins to the myelin membrane and impaired SC-axonal communication. Some of the ultrastructural findings in HMSNL peripheral nerve
(18,23,24,27) and the mouse models of NDRG1-deficiency ((26) & Chapter 6) support this.
These include the presence of irregular membrane-bound inclusion bodies in the Schwann cell cytoplasm and axon, as well as evidence of hypomyelination, defects in myelin morphology and severe axonal loss. These are features that are perhaps indicative of a failure in the transport of components necessary for normal myelin production and maintenance and communication between the SC and axon.
While peripheral neuropathy is the drastic clinical outcome of complete NDRG1 deficit, the results of this investigation suggest that the essential physiological role of the
154 protein is most likely related to the general mechanisms of lipid metabolism and trafficking. The unexpected finding of NDRG1 interaction with the major protein components of HDL, suggested a role in cholesterol transport/metabolism, this was supported by blood lipid analysis where NDRG1 deficit conferred by the homozygous
R148X genotype is associated with a significant reduction in plasma HDL-C levels in males, and a trend towards a dosage effect was evident in heterozygotes. This unique role for NDRG1 was also supported by the chromosomal location of the gene within a HDL
QTL in both humans and mice (216,217) and previous in vitro studies linking NDRG1 to atherosclerosis (41,141). Based on our co-localisation data in hepatoma cells this reduction may be related to the formation of nascent HDL particles and/or a decreased efficiency in reverse cholesterol transport, hypotheses which should be tested in future HDL sub- fractionation studies and efflux experiments utilizing Ndrg1-deficient cell-lines.
The neuropathology and rapid progression to severe disability in CMT4D disease, combined with the pattern of expression of the NDRG1 protein, make it an important model in understanding the mechanisms of axonal loss, which is the cause of the neurological deficit in demyelinating CMT disease. Based upon our findings it is perhaps plausable that a defect in lipid metabolism and/or trafficking and its effect upon Schwann cell-axonal communication and survival, is the basis for the neuropathy in CMT4D disease.
At the same time HMSNL, like other neurological disorders of lipid trafficking and metabolism such as Tangier disease in which part of the clinical features include neuropathy and abnormal blood lipid profiles, may possibly serve as a model in our understanding of the general mechanisms of lipid metabolism and trafficking pathways.
Determining the cellular mechanisms of the interactions detected in this study and their functional significance is the next challenge in assigning a definitive role for this protein. 155 REFERENCES
1. Pareyson, D. (1999) Charcot-marie-tooth disease and related neuropathies: molecular basis for distinction and diagnosis Muscle Nerve 22, 1498-1509 2. Bennett, C. L., and Chance, P. F. (2001) Molecular pathogenesis of hereditary motor, sensory and autonomic neuropathies Curr Opin Neurol 14, 621-627 3. Hanemann, C. O. (2001) Hereditary demyelinating neuropathies: from gene to disease Neurogenetics 3, 53-57 4. Reilly, M. M. (2000) Classification of the hereditary motor and sensory neuropathies Curr Opin Neurol 13, 561-564 5. Villanova, M., Timmerman, V., De Jonghe, P., Malandrini, A., Rizzuto, N., Van Broeckhoven, C., Guazzi, G., and Rossi, A. (1998) Charcot-Marie-Tooth disease: an intermediate form Neuromuscul Disord 8, 392-393 6. Kennerson, M. L., Zhu, D., Gardner, R. J., Storey, E., Merory, J., Robertson, S. P., and Nicholson, G. A. (2001) Dominant intermediate Charcot-Marie-Tooth neuropathy maps to chromosome 19p12-p13.2 Am J Hum Genet 69, 883-888 7. Meuleman, J., Timmerman, V., Nelis, E., and De Jonghe, P. (2000) Molecular genetics of inherited peripheral neuropathies: who are the actors? Acta Neurol Belg 100, 171-180 8. Harding, A. E., and Thomas, P. K. (1980) Autosomal recessive forms of hereditary motor and sensory neuropathy J Neurol Neurosurg Psychiatry 43, 669-678 9. Kalaydjieva, L., Hallmayer, J., Chandler, D., Savov, A., Nikolova, A., Angelicheva, D., King, R. H., Ishpekova, B., Honeyman, K., Calafell, F., Shmarov, A., Petrova, J., Turnev, I., Hristova, A., Moskov, M., Stancheva, S., Petkova, I., Bittles, A. H., Georgieva, V., Middleton, L., and Thomas, P. K. (1996) Gene mapping in Gypsies identifies a novel demyelinating neuropathy on chromosome 8q24 Nat Genet 14, 214-217. 10. Kalaydjieva, L., Gresham, D., Gooding, R., Heather, L., Baas, F., de Jonge, R., Blechschmidt, K., Angelicheva, D., Chandler, D., Worsley, P., Rosenthal, A., King, R. H., and Thomas, P. K. (2000) N-myc downstream-regulated gene 1 is mutated in hereditary motor and sensory neuropathy-Lom Am J Hum Genet 67, 47-58 11. Rogers, T., Chandler, D., Angelicheva, D., Thomas, P. K., Youl, B., Tournev, I., Gergelcheva, V., and Kalaydjieva, L. (2000) A novel locus for autosomal recessive peripheral neuropathy in the EGR2 region on 10q23 Am J Hum Genet 67, 664-671 12. Hantke, J., Rogers, T., French, L., Tournev, I., Guergueltcheva, V., Urtizberea, J. A., Colomer, J., Corches, A., Lupu, C., Merlini, L., Thomas, P. K., and Kalaydjieva, L. (2003) Refined mapping of the HMSNR critical gene region--construction of a high-density integrated genetic and physical map Neuromuscul Disord 13, 729-736 13. Varon, R., Gooding, R., Steglich, C., Marns, L., Tang, H., Angelicheva, D., Yong, K. K., Ambrugger, P., Reinhold, A., Morar, B., Baas, F., Kwa, M., Tournev, I., Guerguelcheva, V., Kremensky, I., Lochmuller, H., Mullner-Eidenbock, A., Merlini, L., Neumann, L., Burger, J., Walter, M., Swoboda, K., Thomas, P. K., von Moers, A., Risch, N., and Kalaydjieva, L. (2003) Partial deficiency of the C- terminal-domain phosphatase of RNA polymerase II is associated with congenital cataracts facial dysmorphism neuropathy syndrome Nat Genet 35, 185-189 14. Angelicheva, D., Turnev, I., Dye, D., Chandler, D., Thomas, P. K., and Kalaydjieva, L. (1999) Congenital cataracts facial dysmorphism neuropathy (CCFDN) syndrome: a novel developmental disorder in Gypsies maps to 18qter Eur J Hum Genet 7, 560-566 156 15. Gooding, R., Colomer, J., King, R., Angelicheva, D., Marns, L., Parman, Y., Chandler, D., Bertranpetit, J., and Kalaydjieva, L. (2005) A novel Gypsy founder mutation, p.Arg1109X in the CMT4C gene, causes variable peripheral neuropathy phenotypes J Med Genet 42, e69 16. Colomer, J., Iturriaga, C., Kalaydjieva, L., Angelicheva, D., King, R. H., and Thomas, P. K. (2000) Hereditary motor and sensory neuropathy-Lom (HMSNL) in a Spanish family: clinical, electrophysiological, pathological and genetic studies Neuromuscul Disord 10, 578-583. 17. Merlini, L., Villanova, M., Sabatelli, P., Trogu, A., Malandrini, A., Yanakiev, P., Maraldi, N. M., and Kalaydjieva, L. (1998) Hereditary motor and sensory neuropathy Lom type in an Italian Gypsy family Neuromuscul Disord 8, 182-185. 18. Butinar, D., Zidar, J., Leonardis, L., Popovic, M., Kalaydjieva, L., Angelicheva, D., Sininger, Y., Keats, B., and Starr, A. (1999) Hereditary auditory, vestibular, motor, and sensory neuropathy in a Slovenian Roma (Gypsy) kindred Ann Neurol 46, 36- 44. 19. Baethmann, M., Gohlich-Ratmann, G., Schroder, J. M., Kalaydjieva, L., and Voit, T. (1998) HMSNL in a 13-year-old Bulgarian girl Neuromuscul Disord 8, 90-94. 20. Morar, B., Gresham, D., Angelicheva, D., Tournev, I., Gooding, R., Guergueltcheva, V., Schmidt, C., Abicht, A., Lochmuller, H., Tordai, A., Kalmar, L., Nagy, M., Karcagi, V., Jeanpierre, M., Herczegfalvi, A., Beeson, D., Venkataraman, V., Warwick Carter, K., Reeve, J., de Pablo, R., Kucinskas, V., and Kalaydjieva, L. (2004) Mutation history of the roma/gypsies Am J Hum Genet 75, 596-609 21. Kalaydjieva, L., Gresham, D., and Calafell, F. (2001) Genetic studies of the Roma (Gypsies): a review BMC Med Genet 2, 5 22. Leonardis, L., Zidar, J., Popovic, M., Timmerman, V., Lofgren, A., Van Broeckhoven, C., and Butinar, D. (2000) Hereditary motor and sensory neuropathy associated with auditory neuropathy in a Gypsy family Pflugers Arch 439, R208- 210 23. Baethmann, M., Gohlich-Ratmann, G., Schroder, J. M., Kalaydjieva, L., and Voit, T. (1998) HMSNL in a 13-year-old Bulgarian girl Neuromuscul Disord 8, 90-94 24. Kalaydjieva, L., Nikolova, A., Turnev, I., Petrova, J., Hristova, A., Ishpekova, B., Petkova, I., Shmarov, A., Stancheva, S., Middleton, L., Merlini, L., Trogu, A., Muddle, J. R., King, R. H. M., and Thomas, P. K. (1998) Hereditary motor and sensory neuropathy-Lom, a novel demyelinating neuropathy associated with deafness in gypsies: Clinical, electrophysiological and nerve biopsy findings Brain 121, 399-408 25. Kalaydjieva, L., King, R. H., Gresham, D., Molnar, M., Tournev, I., Angelicheva, D., Butinar, D., Colomer, J., Corches, A., Lupu, C., Merlini, L., Zanetti, M., Bergonzoni, C., and Thomas, P. K. (2001) Hereditary motor and sensory neuropathy Lom Acta Myologica 20, 192-202 26. Okuda, T., Higashi, Y., Kokame, K., Tanaka, C., Kondoh, H., and Miyata, T. (2004) Ndrg1-deficient mice exhibit a progressive demyelinating disorder of peripheral nerves Mol Cell Biol 24, 3949-3956 27. King, R. H., Tournev, I., Colomer, J., Merlini, L., Kalaydjieva, L., and Thomas, P. K. (1999) Ultrastructural changes in peripheral nerve in hereditary motor and sensory neuropathy-Lom Neuropathol Appl Neurobiol 25, 306-312 28. Thomas, P. K., Landon, D. N., and King, R. H. M. (1997) in Greenfield's Neuropathology (L., L. P., ed) Vol. 2, pp. 367-448, 2 vols., Arnold, London 157 29. King, R. H., Tournev, I., Colomer, J., Merlini, L., Kalaydjieva, L., and Thomas, P. K. (1999) Ultrastructural changes in peripheral nerve in hereditary motor and sensory neuropathy-Lom Neuropathol Appl Neurobiol 25, 306-312. 30. Lampert, P., Blumberg, J. M., and Pentschew, A. (1964) An electron microscopic study of dystrophic axons in the gracile and cuneate nuclei of vitamin E-deficient rats: axonal dystrophy in vitamen E deficiency Journal of Neuropatholgy and Experimental Neurology 23, 60-77 31. Krauter-Canham, R., Bronner, R., Evrard, J. L., Hahne, G., Friedt, W., and Steinmetz, A. (1997) A transmitting tissue- and pollen-expressed protein from sunflower with sequence similarity to the human RTP protein Plant Science 129, 191-202 32. Zhou, R. H., Kokame, K., Tsukamoto, Y., Yutani, C., Kato, H., and Miyata, T. (2001) Characterization of the human NDRG gene family: a newly identified member, NDRG4, is specifically expressed in brain and heart Genomics 73, 86-97 33. Zhao, W., Tang, R., Huang, Y., Wang, W., Zhou, Z., Gu, S., Dai, J., Ying, K., Xie, Y., and Mao, Y. (2001) Cloning and expression pattern of the human NDRG3 gene Biochim Biophys Acta 1519, 134-138 34. Shimono, A., Okuda, T., and Kondoh, H. (1999) N-myc-dependent repression of ndr1, a gene identified by direct subtraction of whole mouse embryo cDNAs between wild type and N-myc mutant Mech Dev 83, 39-52 35. Kokame, K., Kato, H., and Miyata, T. (1996) Homocysteine-respondent genes in vascular endothelial cells identified by differential display analysis. GRP78/BiP and novel genes J Biol Chem 271, 29659-29665 36. Van Belzen, N., Dinjens, W. N., Diesveld, M. P., Groen, N. A., van der Made, A. C., Nozawa, Y., Vlietstra, R., Trapman, J., and Bosman, F. T. (1997) A novel gene which is up-regulated during colon epithelial cell differentiation and down- regulated in colorectal neoplasms Lab Invest 77, 85-92 37. Zhou, D., Salnikow, K., and Costa, M. (1998) Cap43, a novel gene specifically induced by Ni2+ compounds Cancer Res 58, 2182-2189 38. Kurdistani, S. K., Arizti, P., Reimer, C. L., Sugrue, M. M., Aaronson, S. A., and Lee, S. W. (1998) Inhibition of tumor cell growth by RTP/rit42 and its responsiveness to p53 and DNA damage Cancer Res 58, 4439-4444 39. Park, H., Adams, M. A., Lachat, P., Bosman, F., Pang, S. C., and Graham, C. H. (2000) Hypoxia induces the expression of a 43-kDa protein (PROXY-1) in normal and malignant cells Biochem Biophys Res Commun 276, 321-328 40. Kokame, K., Kato, H., and Miyata, T. (1998) Nonradioactive differential display cloning of genes induced by homocysteine in vascular endothelial cells Methods 16, 434-443 41. Sato, N., Kokame, K., Shimokado, K., Kato, H., and Miyata, T. (1998) Changes of gene expression by lysophosphatidylcholine in vascular endothelial cells: 12 up- regulated distinct genes including 5 cell growth-related, 3 thrombosis-related, and 4 others J Biochem (Tokyo) 123, 1119-1126 42. Chen, B., Nelson, D. M., and Sadovsky, Y. (2005) N-Myc downregulated gene 1 (Ndrg1) modulates the response of term human trophoblasts to hypoxic injury J Biol Chem 43. Piquemal, D., Joulia, D., Balaguer, P., Basset, A., Marti, J., and Commes, T. (1999) Differential expression of the RTP/Drg1/Ndr1 gene product in proliferating and growth arrested cells Biochim Biophys Acta 1450, 364-373
158 44. Chang, J. T., Wang, H. M., Chang, K. W., Chen, W. H., Wen, M. C., Hsu, Y. M., Yung, B. Y., Chen, I. H., Liao, C. T., Hsieh, L. L., and Cheng, A. J. (2005) Identification of differentially expressed genes in oral squamous cell carcinoma (OSCC): overexpression of NPM, CDK1 and NDRG1 and underexpression of CHES1 Int J Cancer 114, 942-949 45. Nimmrich, I., Erdmann, S., Melchers, U., Finke, U., Hentsch, S., Moyer, M. P., Hoffmann, I., and Muller, O. (2000) Seven genes that are differentially transcribed in colorectal tumor cell lines Cancer Lett 160, 37-43 46. Rutherford, M. N., Bayly, G. R., Matthews, B. P., Okuda, T., Dinjens, W. M., Kondoh, H., and LeBrun, D. P. (2001) The leukemogenic transcription factor E2a- Pbx1 induces expression of the putative N-myc and p53 target gene NDRG1 in Ba/F3 cells Leukemia 15, 362-370 47. Ulrix, W., Swinnen, J. V., Heyns, W., and Verhoeven, G. (1999) The differentiation-related gene 1, Drg1, is markedly upregulated by androgens in LNCaP prostatic adenocarcinoma cells FEBS Lett 455, 23-26 48. Salnikow, K., Kluz, T., and Costa, M. (1999) Role of Ca(2+) in the regulation of nickel-inducible Cap43 gene expression Toxicology & Applied Pharmacology 160, 127-132 49. van Belzen, N., Dinjens, W. N., Diesveld, M. P., Groen, N. A., van der Made, A. C., Nozawa, Y., Vlietstra, R., Trapman, J., and Bosman, F. T. (1997) A novel gene which is up-regulated during colon epithelial cell differentiation and down- regulated in colorectal neoplasms Laboratory Investigation 77, 85-92 50. Lin, T. M., and Chang, C. (1997) Cloning and characterization of TDD5, an androgen target gene that is differentially repressed by testosterone and dihydrotestosterone Proc Natl Acad Sci U S A 94, 4988-4993. 51. Piquemal, D., Joulia, D., Balaguer, P., Basset, A., Marti, J., and Commes, T. (1999) Differential expression of the RTP/Drg1/Ndr1 gene product in proliferating and growth arrested cells Biochimica et Biophysica Acta 1450, 364-373 52. Segawa, T., Nau, M. E., Xu, L. L., Chilukuri, R. N., Makarem, M., Zhang, W., Petrovics, G., Sesterhenn, I. A., McLeod, D. G., Moul, J. W., Vahey, M., and Srivastava, S. (2002) Androgen-induced expression of endoplasmic reticulum (ER) stress response genes in prostate cancer cells Oncogene 21, 8749-8758 53. Shimono, A., Okuda, T., and Kondoh, H. (1999) N-myc-dependent repression of Ndr1, a gene identified by direct subtraction of whole mouse embryo cDNAs between wild type and N-myc mutant Mechanisms of Development 83, 39-52 54. Lachat, P., Shaw, P., Gebhard, S., van Belzen, N., Chaubert, P., and Bosman, F. T. (2002) Expression of NDRG1, a differentiation-related gene, in human tissues Histochem Cell Biol 118, 399-408 55. Xu, B., Lin, L., and Rote, N. S. (1999) Identification of a stress-induced protein during human trophoblast differentiation by differential display analysis Biol Reprod 61, 681-686 56. Agarwala, K. L., Kokame, K., Kato, H., and Miyata, T. (2000) Phosphorylation of RTP, an ER stress-responsive cytoplasmic protein Biochem Biophys Res Commun 272, 641-647 57. Taketomi, Y., Sugiki, T., Saito, T., Ishii, S., Hisada, M., Suzuki-Nishimura, T., Uchida, M. K., Moon, T. C., Chang, H. W., Natori, Y., Miyazawa, S., Kikuchi- Yanoshita, R., Murakami, M., and Kudo, I. (2003) Identification of NDRG1 as an early inducible gene during in vitro maturation of cultured mast cells Biochem Biophys Res Commun 306, 339-346 159 58. Sugiki, T., Murakami, M., Taketomi, Y., Kikuchi-Yanoshita, R., and Kudo, I. (2004) N-myc downregulated gene 1 is a phosphorylated protein in mast cells Biol Pharm Bull 27, 624-627 59. Shaw, E., McCue, L. A., Lawrence, C. E., and Dordick, J. S. (2002) Identification of a novel class in the alpha/beta hydrolase fold superfamily: the N-myc differentiation-related proteins Proteins 47, 163-168 60. Heikinheimo, P., Goldman, A., Jeffries, C., and Ollis, D. L. (1999) Of barn owls and bankers: a lush variety of alpha/beta hydrolases Structure Fold Des 7, R141- 146 61. Hotelier, T., Renault, L., Cousin, X., Negre, V., Marchot, P., and Chatonnet, A. (2004) ESTHER, the database of the alpha/beta-hydrolase fold superfamily of proteins Nucleic Acids Res 32, D145-147 62. Nardini, M., and Dijkstra, B. W. (1999) Alpha/beta hydrolase fold enzymes: the family keeps growing Curr Opin Struct Biol 9, 732-737 63. Kleinkauf, H. (2000) The role of 4'-phosphopantetheine in t' biosynthesis of fatty acids, polyketides and peptides Biofactors 11, 91-92 64. Zoroddu, M. A., Kowalik-Jankowska, T., Kozlowski, H., Salnikow, K., and Costa, M. (2001) Ni(II) and Cu(II) binding with a 14-aminoacid sequence of Cap43 protein, TRSRSHTSEGTRSR J Inorg Biochem 84, 47-54 65. Ferdinandusse, S., Denis, S., Clayton, P. T., Graham, A., Rees, J. E., Allen, J. T., McLean, B. N., Brown, A. Y., Vreken, P., Waterham, H. R., and Wanders, R. J. (2000) Mutations in the gene encoding peroxisomal alpha-methylacyl-CoA racemase cause adult-onset sensory motor neuropathy Nat Genet 24, 188-191 66. Follis, R. H., and Wintrobe, M. M. (1945) A comparison of the effects of pyridoxine and pantothenic acid deficiency in swine Journal of Experimental Medicine 81, 539-552 67. Bean, W. B., Hodges, R. E., and Daum, K. E. (1955) Pantothenic acid deficiency induced in human subjects Journal of Clinical Investigation 34, 1073-1084 68. Berger, P., Sirkowski, E. E., Scherer, S. S., and Suter, U. (2004) Expression analysis of the N-Myc downstream-regulated gene 1 indicates that myelinating Schwann cells are the primary disease target in hereditary motor and sensory neuropathy-Lom Neurobiol Dis 17, 290-299 69. Hirata, K., Masuda, K., Morikawa, W., He, J. W., Kuraoka, A., Kuwano, M., and Kawabuchi, M. (2004) N-myc downstream-regulated gene 1 expression in injured sciatic nerves Glia 47, 325-334 70. Dong, Z., Sinanan, A., Parkinson, D., Parmantier, E., Mirsky, R., and Jessen, K. R. (1999) Schwann cell development in embryonic mouse nerves J Neurosci Res 56, 334-348. 71. Jessen, K. R., and Mirsky, R. (1999) Developmental regulation in the Schwann cell lineage Adv Exp Med Biol 468, 3-12 72. Jessen, K. R., and Mirsky, R. (1998) Origin and early development of Schwann cells Microscopy and Research Technique 41, 393-402 73. Verheijen, M., Chrast, R., Burrola, P., and Lemke, G. (2003) Local regulation of fat metabolism in peripheral nerves Genes Dev 17, 2450-2464 74. Nagarajan, R., Le, N., Mahoney, H., Araki, T., and Milbrandt, J. (2002) Deciphering peripheral nerve myelination by using Schwann cell expression profiling Proc Natl Acad Sci U S A 99, 8998-9003 75. Nelis, E., Van Broeckhoven, C., De Jonghe, P., Lofgren, A., Vandenberghe, A., Latour, P., Le Guern, E., Brice, A., Mostacciuolo, M. L., Schiavon, F., Palau, F., 160 Bort, S., Upadhyaya, M., Rocchi, M., Archidiacono, N., Mandich, P., Bellone, E., Silander, K., Savontaus, M. L., Navon, R., Goldberg-Stern, H., Estivill, X., Volpini, V., Friedl, W., Gal, A., and et al. (1996) Estimation of the mutation frequencies in Charcot-Marie-Tooth disease type 1 and hereditary neuropathy with liability to pressure palsies: a European collaborative study Eur J Hum Genet 4, 25-33 76. Mersiyanova, I. V., Ismailov, S. M., Polyakov, A. V., Dadali, E. L., Fedotov, V. P., Nelis, E., Lofgren, A., Timmerman, V., van Broeckhoven, C., and Evgrafov, O. V. (2000) Screening for mutations in the peripheral myelin genes PMP22, MPZ and Cx32 (GJB1) in Russian Charcot-Marie-Tooth neuropathy patients Hum Mutat 15, 340-347 77. Mostacciuolo, M. L., Righetti, E., Zortea, M., Bosello, V., Schiavon, F., Vallo, L., Merlini, L., Siciliano, G., Fabrizi, G. M., Rizzuto, N., Milani, M., Baratta, S., and Taroni, F. (2001) Charcot-Marie-Tooth disease type I and related demyelinating neuropathies: Mutation analysis in a large cohort of Italian families Hum Mutat 18, 32-41 78. Boerkoel, C. F., Takashima, H., Bacino, C. A., Daentl, D., and Lupski, J. R. (2001) EGR2 mutation R359W causes a spectrum of Dejerine-Sottas neuropathy Neurogenetics 3, 153-157 79. Numakura, C., Lin, C., Ikegami, T., Guldberg, P., and Hayasaka, K. (2002) Molecular analysis in Japanese patients with Charcot-Marie-Tooth disease: DGGE analysis for PMP22, MPZ, and Cx32/GJB1 mutations Hum Mutat 20, 392-398 80. Wise, C. A., Garcia, C. A., Davis, S. N., Heju, Z., Pentao, L., Patel, P. I., and Lupski, J. R. (1993) Molecular analyses of unrelated Charcot-Marie-Tooth (CMT) disease patients suggest a high frequency of the CMTIA duplication Am J Hum Genet 53, 853-863 81. Antonarakis. (1998) Recommendations for a nomenclature system for human gene mutations. Nomenclature working group Human Mutation 11, 1-3 82. Rozas, J., and Rozas, R. (1999) DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis Bioinformatics 15, 174-175 83. McGuffin, L. J., Bryson, K., and Jones, D. T. (2000) The PSIPRED protein structure prediction server Bioinformatics 16, 404-405 84. Rost, B. (1996) Predicting one-dimensional protein structure by profile based neural networks Meth Enzm 266, 525-539 85. Obenauer, J. C., Cantley, L. C., and Yaffe, M. B. (2003) Scansite 2.0: Proteome- wide prediction of cell signaling interactions using short sequence motifs Nucleic Acids Res 31, 3635-3641 86. Hulo, N., Sigrist, C. J., Le Saux, V., Langendijk-Genevaux, P. S., Bordoli, L., Gattiker, A., De Castro, E., Bucher, P., and Bairoch, A. (2004) Recent improvements to the PROSITE database Nucleic Acids Res 32 Database issue, D134-137 87. Bateman, A., Coin, L., Durbin, R., Finn, R. D., Hollich, V., Griffiths-Jones, S., Khanna, A., Marshall, M., Moxon, S., Sonnhammer, E. L., Studholme, D. J., Yeats, C., and Eddy, S. R. (2004) The Pfam protein families database Nucleic Acids Res 32 Database issue, D138-141 88. Hayasaka, K., Himoro, M., Sato, W., Takada, G., Uyemura, K., Shimizu, N., Bird, T. D., Conneally, P. M., and Chance, P. F. (1993) Charcot-Marie-Tooth neuropathy type 1B is associated with mutations of the myelin P0 gene Nat Genet 5, 31-34
161 89. Warner, L. E., Hilz, M. J., Appel, S. H., Killian, J. M., Kolodry, E. H., Karpati, G., Carpenter, S., Watters, G. V., Wheeler, C., Witt, D., Bodell, A., Nelis, E., Van Broeckhoven, C., and Lupski, J. R. (1996) Clinical phenotypes of different MPZ (P0) mutations may include Charcot-Marie-Tooth type 1B, Dejerine-Sottas, and congenital hypomyelination Neuron 17, 451-460 90. Warner, L. E., Mancias, P., Butler, I. J., McDonald, C. M., Keppen, L., Koob, K. G., and Lupski, J. R. (1998) Mutations in the early growth response 2 (EGR2) gene are associated with hereditary myelinopathies Nat Genet 18, 382-384 91. Warner, L. E., Svaren, J., Milbrandt, J., and Lupski, J. R. (1999) Functional consequences of mutations in the early growth response 2 gene (EGR2) correlate with severity of human myelinopathies Hum Mol Genet 8, 1245-1251 92. Baxter, R. V., Ben Othmane, K., Rochelle, J. M., Stajich, J. E., Hulette, C., Dew- Knight, S., Hentati, F., Ben Hamida, M., Bel, S., Stenger, J. E., Gilbert, J. R., Pericak-Vance, M. A., and Vance, J. M. (2002) Ganglioside-induced differentiation-associated protein-1 is mutant in Charcot-Marie-Tooth disease type 4A/8q21 Nat Genet 30, 21-22 93. Cuesta, A., Pedrola, L., Sevilla, T., Garcia-Planells, J., Chumillas, M. J., Mayordomo, F., LeGuern, E., Marin, I., Vilchez, J. J., and Palau, F. (2002) The gene encoding ganglioside-induced differentiation-associated protein 1 is mutated in axonal Charcot-Marie-Tooth type 4A disease Nat Genet 30, 22-25 94. Vance, J. M. (2000) The many faces of Charcot-Marie-Tooth disease Arch Neurol 57, 638-640 95. Timmerman, V., De Jonghe, P., Ceuterick, C., De Vriendt, E., Lofgren, A., Nelis, E., Warner, L. E., Lupski, J. R., Martin, J. J., and Van Broeckhoven, C. (1999) Novel missense mutation in the early growth response 2 gene associated with Dejerine-Sottas syndrome phenotype Neurology 52, 1827-1832 96. Bolino, A., Lonie, L. J., Zimmer, M., Boerkoel, C. F., Takashima, H., Monaco, A. P., and Lupski, J. R. (2001) Denaturing high-performance liquid chromatography of the myotubularin-related 2 gene (MTMR2) in unrelated patients with Charcot- Marie-Tooth disease suggests a low frequency of mutation in inherited neuropathy Neurogenetics 3, 107-109 97. Takashima, H., Boerkoel, C. F., De Jonghe, P., Ceuterick, C., Martin, J. J., Voit, T., Schroder, J. M., Williams, A., Brophy, P. J., Timmerman, V., and Lupski, J. R. (2002) Periaxin mutations cause a broad spectrum of demyelinating neuropathies Ann Neurol 51, 709-715 98. Hayasaka, K., Ohnishi, A., Takada, G., Fukushima, Y., and Murai, Y. (1993) Mutation of the myelin P0 gene in Charcot-Marie-tooth neuropathy type 1 Biochem Biophys Res Commun 194, 1317-1322 99. Cargill, M., Altshuler, D., Ireland, J., Sklar, P., Ardlie, K., Patil, N., Shaw, N., Lane, C. R., Lim, E. P., Kalyanaraman, N., Nemesh, J., Ziaugra, L., Friedland, L., Rolfe, A., Warrington, J., Lipshutz, R., Daley, G. Q., and Lander, E. S. (1999) Characterization of single-nucleotide polymorphisms in coding regions of human genes Nat Genet 22, 231-238 100. Schneider, J. A., Pungliya, M. S., Choi, J. Y., Jiang, R., Sun, X. J., Salisbury, B. A., and Stephens, J. C. (2003) DNA variability of human genes Mech Ageing Dev 124, 17-25 101. King, R. H., Tournev, I., Colomer, J., Merlini, L., Kalaydjieva, L., and Thomas, P. K. (1999) Ultrastructural changes in the peripheral nerve in hereditary motor and sensory neuropathy-Lom Neuropathology Applied Neurobiology 4, 306-312 162 102. Fields, S., and Song, O. (1989) A novel genetic system to detect protein-protein interactions Nature 340, 245-246 103. Schulze, A., Standera, S., Buerger, E., Kikkert, M., van Voorden, S., Wiertz, E., Koning, F., Kloetzel, P., and Seeger, M. (2005) The ubiquitin-domain protein HERP forms a complex with components of the endoplasmic reticulum associated degradation pathway J Mol Biol 354, 1021-1027 104. Ingley, E., Williams, J. H., Walker, C. E., Tsai, S., Colley, S., Sayer, M. S., Tilbrook, P. A., Sarna, M., Beaumont, J. G., and Klinken, S. P. (1999) A novel ADP-ribosylation like factor (ARL-6), interacts with the protein-conducting channel SEC61beta subunit FEBS Lett 459, 69-74. 105. Aho, S., Arffman, A., Pummi, T., and Uitto, J. (1997) A novel reporter gene MEL1 for the yeast two-hybrid system Anal Biochem 253, 270-272 106. Singh, M., and Weil, P. (2002) A method for plasmid purification directly from yeast Analytical Biochemistry 307, 13-17 107. Lin, J., Liang, Z., Zhang, Z., and Li, G. (2001) Membrane topography and topogenesis of prenylated Rab acceptor (PRA1) J Biol Chem 276, 41733-41741 108. Brouillette, C. G., Anantharamaiah, G. M., Engler, J. A., and Borhani, D. W. (2001) Structural models of human apolipoprotein A-I: a critical analysis and review Biochim Biophys Acta 1531, 4-46 109. Serebriiskii, I., Estojak, J., Berman, M., and Golemis, E. A. (2000) Approaches to detecting false positives in yeast two-hybrid systems Biotechniques 28, 328-330, 332-326 110. Bartel, P., Chien, C. T., Sternglanz, R., and Fields, S. (1993) Elimination of false positives that arise in using the two-hybrid system Biotechniques 14, 920-924 111. Enouf, V., Chwetzoff, S., Trugnan, G., and Cohen, J. (2003) Interactions of rotavirus VP4 spike protein with the endosomal protein Rab5 and the prenylated Rab acceptor PRA1 J Virol 77, 7041-7047 112. Weinberg, R. B. (1994) Identification of functional domains in the plasma apolipoproteins by analysis of inter-species sequence variability J Lipid Res 35, 2212-2222 113. Roebroek, A. J., van de Velde, H. J., Van Bokhoven, A., Broers, J. L., Ramaekers, F. C., and Van de Ven, W. J. (1993) Cloning and expression of alternative transcripts of a novel neuroendocrine-specific gene and identification of its 135-kDa translational product J Biol Chem 268, 13439-13447 114. Senden, N. H., Timmer, E. D., Boers, J. E., van de Velde, H. J., Roebroek, A. J., Van de Ven, W. J., Broers, J. L., and Ramaekers, F. C. (1996) Neuroendocrine- specific protein C (NSP-C): subcellular localization and differential expression in relation to NSP-A Eur J Cell Biol 69, 197-213 115. van de Velde, H. J., Roebroek, A. J., Senden, N. H., Ramaekers, F. C., and Van de Ven, W. J. (1994) NSP-encoded reticulons, neuroendocrine proteins of a novel gene family associated with membranes of the endoplasmic reticulum J Cell Sci 107 ( Pt 9), 2403-2416 116. Hens, J., Nuydens, R., Geerts, H., Senden, N. H., Van de Ven, W. J., Roebroek, A. J., van de Velde, H. J., Ramaekers, F. C., and Broers, J. L. (1998) Neuronal differentiation is accompanied by NSP-C expression Cell Tissue Res 292, 229-237 117. Tagami, S., Eguchi, Y., Kinoshita, M., Takeda, M., and Tsujimoto, Y. (2000) A novel protein, RTN-XS, interacts with both Bcl-XL and Bcl-2 on endoplasmic reticulum and reduces their anti-apoptotic activity Oncogene 19, 5736-5746
163 118. Di Sano, F., Fazi, B., Citro, G., Lovat, P. E., Cesareni, G., and Piacentini, M. (2003) Glucosylceramide synthase and its functional interaction with RTN-1C regulate chemotherapeutic-induced apoptosis in neuroepithelioma cells Cancer Res 63, 3860-3865 119. Steiner, P., Kulangara, K., Sarria, J. C., Glauser, L., Regazzi, R., and Hirling, H. (2004) Reticulon 1-C/neuroendocrine-specific protein-C interacts with SNARE proteins J Neurochem 89, 569-580 120. Kim, S. H., Yoo, B. C., Broers, J. L., Cairns, N., and Lubec, G. (2000) Neuroendocrine-specific protein C, a marker of neuronal differentiation, is reduced in brain of patients with Down syndrome and Alzheimer's disease Biochem Biophys Res Commun 276, 329-334 121. Wakisaka, Y., Furuta, A., Masuda, K., Morikawa, W., Kuwano, M., and Iwaki, T. (2003) Cellular distribution of NDRG1 protein in the rat kidney and brain during normal postnatal development J Histochem Cytochem 51, 1515-1525 122. Bucci, C., Chiariello, M., Lattero, D., Maiorano, M., and Bruni, C. B. (1999) Interaction cloning and characterization of the cDNA encoding the human prenylated rab acceptor (PRA1) Biochem Biophys Res Commun 258, 657-662 123. Figueroa, C., Taylor, J., and Vojtek, A. B. (2001) Prenylated Rab acceptor protein is a receptor for prenylated small GTPases J Biol Chem 276, 28219-28225 124. Vojtek, A. B., and Der, C. J. (1998) Increasing complexity of the Ras signaling pathway J Biol Chem 273, 19925-19928 125. Martinez, O., and Goud, B. (1998) Rab proteins Biochim Biophys Acta 1404, 101- 112 126. Hutt, D. M., Da-Silva, L. F., Chang, L. H., Prosser, D. C., and Ngsee, J. K. (2000) PRA1 inhibits the extraction of membrane-bound rab GTPase by GDI1 J Biol Chem 275, 18511-18519 127. Li, L. Y., Shih, H. M., Liu, M. Y., and Chen, J. Y. (2001) The cellular protein PRA1 modulates the anti-apoptotic activity of Epstein-Barr virus BHRF1, a homologue of Bcl-2, through direct interaction J Biol Chem 276, 27354-27362 128. Evans, D. T., Tillman, K. C., and Desrosiers, R. C. (2002) Envelope glycoprotein cytoplasmic domains from diverse lentiviruses interact with the prenylated Rab acceptor J Virol 76, 327-337 129. Verhoeven, K., De Jonghe, P., Coen, K., Verpoorten, N., Auer-Grumbach, M., Kwon, J. M., FitzPatrick, D., Schmedding, E., De Vriendt, E., Jacobs, A., Van Gerwen, V., Wagner, K., Hartung, H. P., and Timmerman, V. (2003) Mutations in the small GTP-ase late endosomal protein RAB7 cause Charcot-Marie-Tooth type 2B neuropathy Am J Hum Genet 72, 722-727 130. Choudhury, A., Dominguez, M., Puri, V., Sharma, D. K., Narita, K., Wheatley, C. L., Marks, D. L., and Pagano, R. E. (2002) Rab proteins mediate Golgi transport of caveola-internalized glycosphingolipids and correct lipid trafficking in Niemann- Pick C cells J Clin Invest 109, 1541-1550 131. Groen, A. K., Oude Elferink, R. P., Verkade, H. J., and Kuipers, F. (2004) The ins and outs of reverse cholesterol transport Ann Med 36, 135-145 132. Fielding, C. J., and Fielding, P. E. (1995) Molecular physiology of reverse cholesterol transport J Lipid Res 36, 211-228 133. Sviridov, D., and Nestel, P. (2002) Dynamics of reverse cholesterol transport: protection against atherosclerosis Atherosclerosis 161, 245-254
164 134. Gagescu, R., Demaurex, N., Parton, R. G., Hunziker, W., Huber, L. A., and Gruenberg, J. (2000) The recycling endosome of Madin-Darby canine kidney cells is a mildly acidic compartment rich in raft components Mol Biol Cell 11, 2775-2791 135. Fielding, C. J., and Fielding, P. E. (2004) Membrane cholesterol and the regulation of signal transduction Biochem Soc Trans 32, 65-69 136. Ng, D. S., O'Connor, P. W., Mortimer, C. B., Leiter, L. A., Connelly, P. W., and Hegele, R. A. (1996) Case report: retinopathy and neuropathy associated with complete apolipoprotein A-I deficiency Am J Med Sci 312, 30-33 137. Brooks-Wilson, A., Marcil, M., Clee, S. M., Zhang, L. H., Roomp, K., van Dam, M., Yu, L., Brewer, C., Collins, J. A., Molhuizen, H. O., Loubser, O., Ouelette, B. F., Fichter, K., Ashbourne-Excoffon, K. J., Sensen, C. W., Scherer, S., Mott, S., Denis, M., Martindale, D., Frohlich, J., Morgan, K., Koop, B., Pimstone, S., Kastelein, J. J., Hayden, M. R., and et al. (1999) Mutations in ABC1 in Tangier disease and familial high-density lipoprotein deficiency Nat Genet 22, 336-345 138. Liscum, L., Ruggiero, R. M., and Faust, J. R. (1989) The intracellular transport of low density lipoprotein-derived cholesterol is defective in Niemann-Pick type C fibroblasts J Cell Biol 108, 1625-1636 139. Shi, S. T., Polyak, S. J., Tu, H., Taylor, D. R., Gretch, D. R., and Lai, M. M. (2002) Hepatitis C virus NS5A colocalizes with the core protein on lipid droplets and interacts with apolipoproteins Virology 292, 198-210 140. Chroni, A., Liu, T., Fitzgerald, M. L., Freeman, M. W., and Zannis, V. I. (2004) Cross-linking and lipid efflux properties of apoA-I mutants suggest direct association between apoA-I helices and ABCA1 Biochemistry 43, 2126-2139 141. Kokame, K., Kato, H., and Miyata, T. (1996) Homocysteine-respondent genes in vascular endothelial cells identified by differential display analysis The Journal of Biological Chemistry 271, 29659-29665 142. Sato, T., Hanada, M., Bodrug, S., Irie, S., Iwama, N., Boise, L. H., Thompson, C. B., Golemis, E., Fong, L., Wang, H. G., and et al. (1994) Interactions among members of the Bcl-2 protein family analyzed with a yeast two-hybrid system Proc Natl Acad Sci U S A 91, 9238-9242. 143. Xu, Y., Piston, D. W., and Johnson, C. H. (1999) A bioluminescence resonance energy transfer (BRET) system: application to interacting circadian clock proteins Proc Natl Acad Sci U S A 96, 151-156 144. Banerjee, D., Rodriguez, M., and Rajasekaran, A. K. (1997) Rapid movement of newly synthesized chicken apolipoprotein AI to trans-Golgi network and its secretion in Madin-Darby canine kidney cells Exp Cell Res 235, 334-344 145. LeBlanc, A. C., Foldvari, M., Spencer, D. F., Breckenridge, W. C., Fenwick, R. G., Williams, D. L., and Mezei, C. (1989) The apolipoprotein A-I gene is actively expressed in the rapidly myelinating avian peripheral nerve J Cell Biol 109, 1245- 1256 146. Mackay, A., Jones, C., Dexter, T., Silva, R. L., Bulmer, K., Jones, A., Simpson, P., Harris, R. A., Jat, P. S., Neville, A. M., Reis, L. F., Lakhani, S. R., and O'Hare, M. J. (2003) cDNA microarray analysis of genes associated with ERBB2 (HER2/neu) overexpression in human mammary luminal epithelial cells Oncogene 22, 2680- 2688 147. Guan, R. J., Ford, H. L., Fu, Y., Li, Y., Shaw, L. M., and Pardee, A. B. (2000) Drg- 1 as a differentiation-related, putative metastatic suppressor gene in human colon cancer Cancer Res 60, 749-755
165 148. Gomez-Casero, E., Navarro, M., Rodriguez-Puebla, M. L., Larcher, F., Paramio, J. M., Conti, C. J., and Jorcano, J. L. (2001) Regulation of the differentiation-related gene Drg-1 during mouse skin carcinogenesis Mol Carcinog 32, 100-109 149. Bandyopadhyay, S., Pai, S. K., Gross, S. C., Hirota, S., Hosobe, S., Miura, K., Saito, K., Commes, T., Hayashi, S., Watabe, M., and Watabe, K. (2003) The Drg-1 gene suppresses tumor metastasis in prostate cancer Cancer Res 63, 1731-1736 150. Bandyopadhyay, S., Pai, S. K., Hirota, S., Hosobe, S., Tsukada, T., Miura, K., Takano, Y., Saito, K., Commes, T., Piquemal, D., Watabe, M., Gross, S., Wang, Y., Huggenvik, J., and Watabe, K. (2004) PTEN up-regulates the tumor metastasis suppressor gene Drg-1 in prostate and breast cancer Cancer Res 64, 7655-7660 151. Gruenberg, J. (2001) The endocytic pathway: a mosaic of domains Nat Rev Mol Cell Biol 2, 721-730 152. Salaun, C., James, D. J., and Chamberlain, L. H. (2004) Lipid rafts and the regulation of exocytosis Traffic 5, 255-264 153. Sugiki, T., Taketomi, Y., Kikuchi-Yanoshita, R., Murakami, M., and Kudo, I. (2004) Association of N-myc downregulated gene 1 with heat-shock cognate protein 70 in mast cells Biol Pharm Bull 27, 628-633 154. Ungewickell, E. (1985) The 70-kd mammalian heat shock proteins are structurally and functionally related to the uncoating protein that releases clathrin triskelia from coated vesicles Embo J 4, 3385-3391 155. Otter-Nilsson, M., Hendriks, R., Pecheur-Huet, E. I., Hoekstra, D., and Nilsson, T. (1999) Cytosolic ATPases, p97 and NSF, are sufficient to mediate rapid membrane fusion Embo J 18, 2074-2083 156. Ahn, J., Chung, K. S., Kim, D. U., Won, M., Kim, L., Kim, K. S., Nam, M., Choi, S. J., Kim, H. C., Yoon, M., Chae, S. K., and Hoe, K. L. (2004) Systematic identification of hepatocellular proteins interacting with NS5A of the hepatitis C virus J Biochem Mol Biol 37, 741-748 157. Gougeon, P. Y., Prosser, D. C., Da-Silva, L. F., and Ngsee, J. K. (2002) Disruption of Golgi morphology and trafficking in cells expressing mutant prenylated rab acceptor-1 J Biol Chem 277, 36408-36414 158. Zerial, M., and McBride, H. (2001) Rab proteins as membrane organizers Nat Rev Mol Cell Biol 2, 107-117 159. Bucci, C., Thomsen, P., Nicoziani, P., McCarthy, J., and van Deurs, B. (2000) Rab7: a key to lysosome biogenesis Mol Biol Cell 11, 467-480 160. Oertle, T., and Schwab, M. E. (2003) Nogo and its paRTNers Trends Cell Biol 13, 187-194 161. Bared, S. M., Buechler, C., Boettcher, A., Dayoub, R., Sigruener, A., Grandl, M., Rudolph, C., Dada, A., and Schmitz, G. (2004) Association of ABCA1 with syntaxin 13 and flotillin-1 and enhanced phagocytosis in tangier cells Mol Biol Cell 15, 5399-5407 162. Tsujita, M., Wu, C. A., Abe-Dohmae, S., Usui, S., Okazaki, M., and Yokoyama, S. (2005) On the hepatic mechanism of HDL assembly by the ABCA1/apoA-I pathway J Lipid Res 46, 154-162 163. Koldamova, R. P., Lefterov, I. M., Lefterova, M. I., and Lazo, J. S. (2001) Apolipoprotein A-I directly interacts with amyloid precursor protein and inhibits A beta aggregation and toxicity Biochemistry 40, 3553-3560 164. Verdile, G., Martins, R. N., Duthie, M., Holmes, E., St George-Hyslop, P. H., and Fraser, P. E. (2000) Inhibiting amyloid precursor protein C-terminal cleavage promotes an interaction with presenilin 1 J Biol Chem 275, 20794-20798 166 165. Hebert, S. S., Bourdages, V., Godin, C., Ferland, M., Carreau, M., and Levesque, G. (2003) Presenilin-1 interacts directly with the beta-site amyloid protein precursor cleaving enzyme (BACE1) Neurobiol Dis 13, 238-245 166. He, W., Lu, Y., Qahwash, I., Hu, X. Y., Chang, A., and Yan, R. (2004) Reticulon family members modulate BACE1 activity and amyloid-beta peptide generation Nat Med 10, 959-965 167. Sai, X., Kawamura, Y., Kokame, K., Yamaguchi, H., Shiraishi, H., Suzuki, R., Suzuki, T., Kawaichi, M., Miyata, T., Kitamura, T., De Strooper, B., Yanagisawa, K., and Komano, H. (2002) Endoplasmic reticulum stress-inducible protein, Herp, enhances presenilin-mediated generation of amyloid beta-protein J Biol Chem 277, 12915-12920 168. Braun, J. E., Wilbanks, S. M., and Scheller, R. H. (1996) The cysteine string secretory vesicle protein activates Hsc70 ATPase J Biol Chem 271, 25989-25993 169. Bader, G. D., Betel, D., and Hogue, C. W. (2003) BIND: the Biomolecular Interaction Network Database Nucleic Acids Res 31, 248-250 170. Salwinski, L., Miller, C. S., Smith, A. J., Pettit, F. K., Bowie, J. U., and Eisenberg, D. (2004) The Database of Interacting Proteins: 2004 update Nucleic Acids Res 32, D449-451 171. Glabe, C. (2001) Intracellular mechanisms of amyloid accumulation and pathogenesis in Alzheimer's disease J Mol Neurosci 17, 137-145 172. Werstuck, G. H., Lentz, S. R., Dayal, S., Hossain, G. S., Sood, S. K., Shi, Y. Y., Zhou, J., Maeda, N., Krisans, S. K., Malinow, M. R., and Austin, R. C. (2001) Homocysteine-induced endoplasmic reticulum stress causes dysregulation of the cholesterol and triglyceride biosynthetic pathways J Clin Invest 107, 1263-1273 173. Wellington, C. L. (2004) Cholesterol at the crossroads: Alzheimer's disease and lipid metabolism Clin Genet 66, 1-16 174. Ehehalt, R., Keller, P., Haass, C., Thiele, C., and Simons, K. (2003) Amyloidogenic processing of the Alzheimer beta-amyloid precursor protein depends on lipid rafts J Cell Biol 160, 113-123 175. de Sousa, M. M., Vital, C., Ostler, D., Fernandes, R., Pouget-Abadie, J., Carles, D., and Saraiva, M. J. (2000) Apolipoprotein AI and transthyretin as components of amyloid fibrils in a kindred with apoAI Leu178His amyloidosis Am J Pathol 156, 1911-1917 176. Rust, S., Rosier, M., Funke, H., Real, J., Amoura, Z., Piette, J. C., Deleuze, J. F., Brewer, H. B., Duverger, N., Denefle, P., and Assmann, G. (1999) Tangier disease is caused by mutations in the gene encoding ATP-binding cassette transporter 1 Nat Genet 22, 352-355 177. Bodzioch, M., Orso, E., Klucken, J., Langmann, T., Bottcher, A., Diederich, W., Drobnik, W., Barlage, S., Buchler, C., Porsch-Ozcurumez, M., Kaminski, W. E., Hahmann, H. W., Oette, K., Rothe, G., Aslanidis, C., Lackner, K. J., and Schmitz, G. (1999) The gene encoding ATP-binding cassette transporter 1 is mutated in Tangier disease Nat Genet 22, 347-351 178. Berger, P., Bonneick, S., Willi, S., Wymann, M., and Suter, U. (2002) Loss of phosphatase activity in myotubularin-related protein 2 is associated with Charcot- Marie-Tooth disease type 4B1 Hum Mol Genet 11, 1569-1579 179. Bolino, A., Muglia, M., Conforti, F. L., LeGuern, E., Salih, M. A., Georgiou, D. M., Christodoulou, K., Hausmanowa-Petrusewicz, I., Mandich, P., Schenone, A., Gambardella, A., Bono, F., Quattrone, A., Devoto, M., and Monaco, A. P. (2000)
167 Charcot-Marie-Tooth type 4B is caused by mutations in the gene encoding myotubularin-related protein-2 Nat Genet 25, 17-19 180. Street, V. A., Bennett, C. L., Goldy, J. D., Shirk, A. J., Kleopa, K. A., Tempel, B. L., Lipe, H. P., Scherer, S. S., Bird, T. D., and Chance, P. F. (2003) Mutation of a putative protein degradation gene LITAF/SIMPLE in Charcot-Marie-Tooth disease 1C Neurology 60, 22-26 181. Zuchner, S., Noureddine, M., Kennerson, M., Verhoeven, K., Claeys, K., De Jonghe, P., Merory, J., Oliveira, S. A., Speer, M. C., Stenger, J. E., Walizada, G., Zhu, D., Pericak-Vance, M. A., Nicholson, G., Timmerman, V., and Vance, J. M. (2005) Mutations in the pleckstrin homology domain of dynamin 2 cause dominant intermediate Charcot-Marie-Tooth disease Nat Genet 37, 289-294 182. Hinshaw, J. E. (2000) Dynamin and its role in membrane fission Annu Rev Cell Dev Biol 16, 483-519 183. Zhao, C., Takita, J., Tanaka, Y., Setou, M., Nakagawa, T., Takeda, S., Yang, H. W., Terada, S., Nakata, T., Takei, Y., Saito, M., Tsuji, S., Hayashi, Y., and Hirokawa, N. (2001) Charcot-Marie-Tooth disease type 2A caused by mutation in a microtubule motor KIF1Bbeta Cell 105, 587-597 184. Zuchner, S., Mersiyanova, I. V., Muglia, M., Bissar-Tadmouri, N., Rochelle, J., Dadali, E. L., Zappia, M., Nelis, E., Patitucci, A., Senderek, J., Parman, Y., Evgrafov, O., Jonghe, P. D., Takahashi, Y., Tsuji, S., Pericak-Vance, M. A., Quattrone, A., Battaloglu, E., Polyakov, A. V., Timmerman, V., Schroder, J. M., and Vance, J. M. (2004) Mutations in the mitochondrial GTPase mitofusin 2 cause Charcot-Marie-Tooth neuropathy type 2A Nat Genet 36, 449-451 185. Tang, B. S., Zhao, G. H., Luo, W., Xia, K., Cai, F., Pan, Q., Zhang, R. X., Zhang, F. F., Liu, X. M., Chen, B., Zhang, C., Shen, L., Jiang, H., Long, Z. G., and Dai, H. P. (2005) Small heat-shock protein 22 mutated in autosomal dominant Charcot-Marie- Tooth disease type 2L Hum Genet 116, 222-224 186. Evgrafov, O. V., Mersiyanova, I., Irobi, J., Van Den Bosch, L., Dierick, I., Leung, C. L., Schagina, O., Verpoorten, N., Van Impe, K., Fedotov, V., Dadali, E., Auer- Grumbach, M., Windpassinger, C., Wagner, K., Mitrovic, Z., Hilton-Jones, D., Talbot, K., Martin, J. J., Vasserman, N., Tverskaya, S., Polyakov, A., Liem, R. K., Gettemans, J., Robberecht, W., De Jonghe, P., and Timmerman, V. (2004) Mutant small heat-shock protein 27 causes axonal Charcot-Marie-Tooth disease and distal hereditary motor neuropathy Nat Genet 36, 602-606 187. Saher, G., Brugger, B., Lappe-Siefke, C., Mobius, W., Tozawa, R., Wehr, M. C., Wieland, F., Ishibashi, S., and Nave, K. A. (2005) High cholesterol level is essential for myelin membrane growth Nat Neurosci 8, 468-475 188. Tsui-Pierchala, B. A., Encinas, M., Milbrandt, J., and Johnson, E. M., Jr. (2002) Lipid rafts in neuronal signaling and function Trends Neurosci 25, 412-417 189. Boyles, J. K., Notterpek, L. M., and Anderson, L. J. (1990) Accumulation of apolipoproteins in the regenerating and remyelinating mammalian peripheral nerve. Identification of apolipoprotein D, apolipoprotein A-IV, apolipoprotein E, and apolipoprotein A-I J Biol Chem 265, 17805-17815 190. Harr, S. D., Uint, L., Hollister, R., Hyman, B. T., and Mendez, A. J. (1996) Brain expression of apolipoproteins E, J, and A-I in Alzheimer's disease J Neurochem 66, 2429-2435 191. Fujii, H., Saito, K., Hamakawa, H., Maekawa, N., Fujigaki, S., Wada, H., Shimokawa, K., Seishima, M., and Takami, T. (2002) Immunohistochemical
168 localization and mRNA expression of apolipoprotein A-I in rat spinal cord J Atheroscler Thromb 9, 93-98 192. Chisholm, J. W., Burleson, E. R., Shelness, G. S., and Parks, J. S. (2002) ApoA-I secretion from HepG2 cells: evidence for the secretion of both lipid-poor apoA-I and intracellularly assembled nascent HDL J Lipid Res 43, 36-44 193. Zheng, H., Kiss, R. S., Franklin, V., Wang, M. D., Haidar, B., and Marcel, Y. L. (2005) ApoA-I lipidation in primary mouse hepatocytes: Separate controls for phospholipid and cholesterol transfers J Biol Chem 194. Hedrick, C. C., and Lusis, A. J. (1994) Apolipoprotein A-II: a protein in search of a function Can J Cardiol 10, 453-459 195. Tailleux, A., Duriez, P., Fruchart, J. C., and Clavey, V. (2002) Apolipoprotein A-II, HDL metabolism and atherosclerosis Atherosclerosis 164, 1-13 196. Oram, J. F., and Yokoyama, S. (1996) Apolipoprotein-mediated removal of cellular cholesterol and phospholipids J Lipid Res 37, 2473-2491 197. Gillotte, K. L., Zaiou, M., Lund-Katz, S., Anantharamaiah, G. M., Holvoet, P., Dhoest, A., Palgunachari, M. N., Segrest, J. P., Weisgraber, K. H., Rothblat, G. H., and Phillips, M. C. (1999) Apolipoprotein-mediated plasma membrane microsolubilization. Role of lipid affinity and membrane penetration in the efflux of cellular cholesterol and phospholipid J Biol Chem 274, 2021-2028 198. Gillotte, K. L., Davidson, W. S., Lund-Katz, S., Rothblat, G. H., and Phillips, M. C. (1998) Removal of cellular cholesterol by pre-beta-HDL involves plasma membrane microsolubilization J Lipid Res 39, 1918-1928 199. Williams, D. L., Connelly, M. A., Temel, R. E., Swarnakar, S., Phillips, M. C., de la Llera-Moya, M., and Rothblat, G. H. (1999) Scavenger receptor BI and cholesterol trafficking Curr Opin Lipidol 10, 329-339 200. Liadaki, K. N., Liu, T., Xu, S., Ishida, B. Y., Duchateaux, P. N., Krieger, J. P., Kane, J., Krieger, M., and Zannis, V. I. (2000) Binding of high density lipoprotein (HDL) and discoidal reconstituted HDL to the HDL receptor scavenger receptor class B type I. Effect of lipid association and APOA-I mutations on receptor binding J Biol Chem 275, 21262-21271 201. Attie, A. D., Kastelein, J. P., and Hayden, M. R. (2001) Pivotal role of ABCA1 in reverse cholesterol transport influencing HDL levels and susceptibility to atherosclerosis J Lipid Res 42, 1717-1726 202. Gordon, T., Castelli, W. P., Hjortland, M. C., Kannel, W. B., and Dawber, T. R. (1977) High density lipoprotein as a protective factor against coronary heart disease. The Framingham Study Am J Med 62, 707-714 203. Gordon, D. J., Probstfield, J. L., Garrison, R. J., Neaton, J. D., Castelli, W. P., Knoke, J. D., Jacobs, D. R., Jr., Bangdiwala, S., and Tyroler, H. A. (1989) High- density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies Circulation 79, 8-15 204. Manninen, V., Elo, M. O., Frick, M. H., Haapa, K., Heinonen, O. P., Heinsalmi, P., Helo, P., Huttunen, J. K., Kaitaniemi, P., Koskinen, P., and et al. (1988) Lipid alterations and decline in the incidence of coronary heart disease in the Helsinki Heart Study Jama 260, 641-651 205. Miller, G. J., and Miller, N. E. (1975) Plasma-high-density-lipoprotein concentration and development of ischaemic heart-disease Lancet 1, 16-19 206. Rhoads, G. G., Gulbrandsen, C. L., and Kagan, A. (1976) Serum lipoproteins and coronary heart disease in a population study of Hawaii Japanese men N Engl J Med 294, 293-298 169 207. Wang, X., and Paigen, B. (2002) Quantitative trait loci and candidate genes regulating HDL cholesterol: a murine chromosome map Arterioscler Thromb Vasc Biol 22, 1390-1401 208. Schaefer, E. J., Heaton, W. H., Wetzel, M. G., and Brewer, H. B., Jr. (1982) Plasma apolipoprotein A-1 absence associated with a marked reduction of high density lipoproteins and premature coronary artery disease Arteriosclerosis 2, 16-26 209. Norum, R. A., Lakier, J. B., Goldstein, S., Angel, A., Goldberg, R. B., Block, W. D., Noffze, D. K., Dolphin, P. J., Edelglass, J., Bogorad, D. D., and Alaupovic, P. (1982) Familial deficiency of apolipoproteins A-I and C-III and precocious coronary-artery disease N Engl J Med 306, 1513-1519 210. Kuivenhoven, J. A., Pritchard, H., Hill, J., Frohlich, J., Assmann, G., and Kastelein, J. (1997) The molecular pathology of lecithin:cholesterol acyltransferase (LCAT) deficiency syndromes J Lipid Res 38, 191-205 211. Banerjee, D., Mukherjee, T. K., and Redman, C. M. (1985) Biosynthesis of high density lipoprotein by chicken liver: intracellular transport and proteolytic processing of nascent apolipoprotein A-1 J Cell Biol 101, 1219-1226 212. Howell, K. E., and Palade, G. E. (1982) Heterogeneity of lipoprotein particles in hepatic Golgi fractions J Cell Biol 92, 833-845 213. Moestrup, S. K., and Kozyraki, R. (2000) Cubilin, a high-density lipoprotein receptor Curr Opin Lipidol 11, 133-140 214. Cohen, J. C., Kiss, R. S., Pertsemlidis, A., Marcel, Y. L., McPherson, R., and Hobbs, H. H. (2004) Multiple rare alleles contribute to low plasma levels of HDL cholesterol Science 305, 869-872 215. Eppig, J. T., Bult, C. J., Kadin, J. A., Richardson, J. E., Blake, J. A., Anagnostopoulos, A., Baldarelli, R. M., Baya, M., Beal, J. S., Bello, S. M., Boddy, W. J., Bradt, D. W., Burkart, D. L., Butler, N. E., Campbell, J., Cassell, M. A., Corbani, L. E., Cousins, S. L., Dahmen, D. J., Dene, H., Diehl, A. D., Drabkin, H. J., Frazer, K. S., Frost, P., Glass, L. H., Goldsmith, C. W., Grant, P. L., Lennon- Pierce, M., Lewis, J., Lu, I., Maltais, L. J., McAndrews-Hill, M., McClellan, L., Miers, D. B., Miller, L. A., Ni, L., Ormsby, J. E., Qi, D., Reddy, T. B., Reed, D. J., Richards-Smith, B., Shaw, D. R., Sinclair, R., Smith, C. L., Szauter, P., Walker, M. B., Walton, D. O., Washburn, L. L., Witham, I. T., and Zhu, Y. (2005) The Mouse Genome Database (MGD): from genes to mice--a community resource for mouse biology Nucleic Acids Res 33, D471-475 216. Soro, A., Pajukanta, P., Lilja, H. E., Ylitalo, K., Hiekkalinna, T., Perola, M., Cantor, R. M., Viikari, J. S., Taskinen, M. R., and Peltonen, L. (2002) Genome scans provide evidence for low-HDL-C loci on chromosomes 8q23, 16q24.1-24.2, and 20q13.11 in Finnish families Am J Hum Genet 70, 1333-1340 217. Almasy, L., Hixson, J. E., Rainwater, D. L., Cole, S., Williams, J. T., Mahaney, M. C., VandeBerg, J. L., Stern, M. P., MacCluer, J. W., and Blangero, J. (1999) Human pedigree-based quantitative-trait-locus mapping: localization of two genes influencing HDL-cholesterol metabolism Am J Hum Genet 64, 1686-1693 218. Korstanje, R., Li, R., Howard, T., Kelmenson, P., Marshall, J., Paigen, B., and Churchill, G. (2004) Influence of sex and diet on quantitative trait loci for HDL cholesterol levels in an SM/J by NZB/BlNJ intercross population J Lipid Res 45, 881-888 219. Semmens, J., Rouse, I., Beilin, L. J., and Masarei, J. R. (1983) Relationship of plasma HDL-cholesterol to testosterone, estradiol, and sex-hormone-binding globulin levels in men and women Metabolism 32, 428-432 170 220. Barter, P., Kastelein, J., Nunn, A., and Hobbs, R. (2003) High density lipoproteins (HDLs) and atherosclerosis; the unanswered questions Atherosclerosis 168, 195- 211 221. von Eckardstein, A., Nofer, J. R., and Assmann, G. (2001) High density lipoproteins and arteriosclerosis. Role of cholesterol efflux and reverse cholesterol transport Arterioscler Thromb Vasc Biol 21, 13-27 222. Banerjee, D., and Redman, C. M. (1984) Biosynthesis of high density lipoprotein by chicken liver: conjugation of nascent lipids with apoprotein A1 J Cell Biol 99, 1917-1926 223. Banerjee, D., and Redman, C. M. (1983) Biosynthesis of high density lipoprotein by chicken liver: nature of nascent intracellular high density lipoprotein J Cell Biol 96, 651-660 224. Odom, D. T., Zizlsperger, N., Gordon, D. B., Bell, G. W., Rinaldi, N. J., Murray, H. L., Volkert, T. L., Schreiber, J., Rolfe, P. A., Gifford, D. K., Fraenkel, E., Bell, G. I., and Young, R. A. (2004) Control of pancreas and liver gene expression by HNF transcription factors Science 303, 1378-1381 225. Sacks, F. M. (2002) The role of high-density lipoprotein (HDL) cholesterol in the prevention and treatment of coronary heart disease: expert group recommendations Am J Cardiol 90, 139-143 226. Ginter, E., Krajcovicova-Kudlackova, M., Kacala, O., Kovacic, V., and Valachovicova, M. (2001) Health status of Romanies (Gypsies) in the Slovak Republic and in the neighbouring countries Bratisl Lek Listy 102, 479-484 227. Krajcovicova-Kudlackova, M., Blazicek, P., Ginter, E., and Spustova, V. (2002) Insulin levels in Gipsy minority Bratisl Lek Listy 103, 459-461 228. Krajcovicova-Kudlackova, M., Blazicek, P., Spustova, V., Valachovicova, M., and Ginter, E. (2004) Cardiovascular risk factors in young Gypsy population Bratisl Lek Listy 105, 256-259 229. Foundation, N. (1990) in National Heart Foundation of Australia and Australian Institute of Health 230. Hunter, M., Bernard, R., Freitas, E., Boyer, A., Morar, B., Martins, I. J., Tournev, I., Jordanova, A., Guergelcheva, V., Ishpekova, B., Kremensky, I., Nicholson, G., Schlotter, B., Lochmuller, H., Voit, T., Colomer, J., Thomas, P. K., Levy, N., and Kalaydjieva, L. (2003) Mutation screening of the N-myc downstream-regulated gene 1 (NDRG1) in patients with Charcot-Marie-Tooth Disease Hum Mutat 22, 129-135 231. He, J., and Zhou, R. (2003) [Correlation of NDRG1 gene with liver tissue differentiation and hepatocarcinogenesis] Beijing Da Xue Xue Bao 35, 471-475 232. Okuda, T., and Kondoh, H. (1999) Identification of new genes Ndr2 and Ndr3 which are related to Ndr1/RTP/Drg1 but show distinct tissue specificity and response to N-myc Biochemical & Biophysical Research Communications 266, 208-215 233. Mikol, D. D., Scherer, S. S., Duckett, S. J., Hong, H. L., and Feldman, E. L. (2002) Schwann cell caveolin-1 expression increases during myelination and decreases after axotomy Glia 38, 191-199 234. Mikol, D. D., Hong, H. L., Cheng, H. L., and Feldman, E. L. (1999) Caveolin-1 expression in Schwann cells Glia 27, 39-52 235. Fiore, R., and Puschel, A. W. (2003) The function of semaphorins during nervous system development Front Biosci 8, s484-499
171 172