STRUCTURAL CHARACTERIZATION AND BIOCHEMICAL ANALYSIS OF ID2, AN INHIBITOR OF DNA-BINDING MARIE VIVIAN WONG TZU YENN (B.Sc.), University of Melbourne A THESIS SUBIMTTED FOR THE DEGREE OF PHILOSOPHY OF SCIENCE DEPARTMENT OF BIOLOGICAL SCIENCES NATIONAL UNIVERSITY OF SINGAPORE 2012 ACKNOWLEDGEMENTS I would like to thank my supervisor Dr Prasanna R Kolatkar for the opportunity to work in his lab and for the valuable insight given during the course of this project I would like to thank Dr Paaventhan Palasingham and Dr Jeremiah Joseph for their mentorship and help in the structural determination at various stages I am grateful to Dr Robert Robinson and Dr Howard Robinson for their assistance with X-ray beamtime and data collection I am thankful to my parents and sister who are always there when help is needed I am also grateful to my husband for his support Finally, I would like to acknowledge all the students and lab mates who made life in the lab a great experience ! i! ! TABLE OF CONTENTS TABLE OF CONTENTS .ii! SUMMARY v! LIST OF TABLES vi! LIST OF FIGURES vii! LIST OF SYMBOLS .xi! CHAPTER 1: INTRODUCTION 1! 1.1! Classes(of(basic(helix0loop0helix((bHLH)(proteins( (1! 1.2! bHLH(structures( (3! 1.3! IDs(are(Group(D(HLH(proteins( (7! 1.4! ID(proteins(in(development( (12! 1.5! ID(proteins(and(myogenesis( (14! 1.6! ID(proteins(and(neurogenesis( .(14! 1.7! IDs(in(cancer( (16! 1.8! Properties(and(roles(of(ID2( (17! 1.9! Aim(and(Scope(of(Project( (20! CHAPTER 2: MATERIALS and METHODS 22! 2.1! Cloning( (22! 2.2! Site(directed(mutagenesis( (24! 2.3! Protein(expression(optimization( (26! 2.4! Native(protein(expression( (26! 2.5! Seleno0Methionine((Se0Met)(substituted(protein(expression( (27! 2.6! Cell(Harvesting( (27! 2.7! Protein(Purification( (28! 2.8! Electrophoretic(mobility(shift(assay( (28! ! ii! 2.9! Crystallization( .(29! 2.10! X0ray(data(collection(and(processing( .(30! CHAPTER 3: RESULTS and DISCUSSION 31! (Expression to X-ray Data Collection) 31! 3.1! Cloning(and(Small0scale(Protein(Expression( (31! 3.2! Protein(Expression(and(Purification( (35! 3.3! Protein(Identification( .(39! 3.4! Crystallization( .(40! 3.5! Data(Collection( (42! CHAPTER 4: RESULTS and DISCUSSION 46! (Structure Solution and Insights) 46! 4.1! Structure(solution(and(Refinement( (46! 4.2! Overall(Structure( .(50! 4.3! Dimer(Interface( (53! 4.3.1! Hydrophobic(Core( (53! 4.3.2! Hydrogen(Bonds( .(53! 4.3.3! Comparison(of(ID3(homology(model(homodimer(interactions( (57! 4.3.4! Disulfide(bond(in(ID2(homodimer(formation( .(59! 4.4! Loop(region( (59! 4.5! N0terminal(Helix01(region( (61! CHAPTER 5: RESULTS and DISCUSSION 64! (Biochemical Studies) 64! 5.1! ID2(protein(activity( (64! 5.2! ID(heterodimer(binding(specificity(and(affinity( (66! 5.3! ID(helix01(residues(in(binding(specificity( (68! 5.4! Exploring(other(differences(in(ID(residues( (71! ! iii! 5.5! MASH1(and(the(ID(proteins( (76! CHAPTER 6: CONCLUSION and FUTURE DIRECTIONS 78! 6.1! Conclusions( (78! 6.2! Future(Directions( .(81! BIBLIOGRAPHY: 83! LIST OF PUBLICATIONS 97! Appendix 1: Protein Sequences (Human) 98! Appendix 2: Purified proteins used in EMSA studies 99! Appendix 3: E47 & MYOD1 cloning, expression and purification for EMSA studies 100! Appendix 4: Summary of expression and purification protocols for ID mutants 103! Appendix 5: ID1 & ID3 cloning, expression and purification 104! Appendix 6: ID2 as a dimer in solution Gel filtration profile 105! Appendix 7: ID2 coordinates 106! ! iv! SUMMARY ! ! The ID proteins, a class of transcription regulators, were named for their role as inhibitors of DNA-binding and differentiation They contained a helix-loop-helix (HLH) domain without a basic DNA-binding domain and worked by dimerizing with basic-HLH transcription factors to inactivate their DNA-binding abilities Although the HLH domain was highly conserved and shared similar topology, the IDs preferentially antagonized group A bHLHs such as E47 (TCF3) but not the group B MYC In general, group A bHLHs contained proteins that bound the enhancer-box (Ebox) motif CANNTG and the consequences of their transcriptional inactivation were implicated in cell cycle regulation, cell lineage determination, differentiation, myogenesis, neurogenesis and tumourigenesis ID2, a member of the ID family, was used to study this protein family Cloning strategies to overcome the instability of this protein family were explored in addition to the expression and purification approaches required to produce enough soluble protein for crystallization The crystal structure of ID2 was solved to 2.1 Å using a seleno-methionine template model in molecular replacement The structure showed for the first time, a loop ion that was previously unreported in HLH structures Residues involved in ioninteractions were investigated for their roles in the structure of ID2 Besides the hydrophobic core, an inspection of the ID2 structure showed that specific hydrogen bonds were required for dimerization Comparisons of the ID2 crystal structure with homology models, previous studies of specific residues, and the ID3 NMR structure were done to examine how these residues might play a role in the structure and function of ID2 Finally, mutations to key residues would be made and their activities tested in competitive EMSAs to gauge their importance in dimerization of the ID protein family ! v! LIST OF TABLES Table 1: Representative structures of bHLH-containing proteins from the PDB for each group 4! Table 2: ID2 constructs and their theoretical biochemical properties estimated by ProtParam (Wilkins, et al., 1999) Constructs described in detail (yellow highlight) 23! Table 3: Primer base for BP cloning (Invitrogen) to create the entry clone for Gateway LR reaction (Invitrogen) attB sites (italics), sequence transferred into pDonr vector during BP reaction (bold), protease sites (underlined) Final selected protease is highlighted in yellow 23! Table 4: Sequences for each construct were added to the primer base in Table to complete the primer sequences used for BP cloning 23! Table 5: Mutagenesis primers Mutation shown after first underscore and changed residue denoted by red bold letter Forward and reverse primers denoted by _F and _R respectively Changed nucleotide (s) denoted by grey highlight 24! Table 6: Domain prediction results for ID2 from Ensembl release 67 32! Table 7: LC/MS/MS mass spectrometry top hits for the purified proteins (Figure 8) Searches were done against all nr as well as human nr to show that the fragments captured always belonged to ID2 Note that the N-HLH-82-L contained the intact N-terminus (matched peptides in bold red) whereas the shorter form HLH24-82-L and the seleno-methionine version did not 39! Table 8: Crystallographic Data Collection Statistics 44! Table 9: Phasing statistics of Se-Met construct HLH24-82-L-Se-Met 47! Table 10: Refinement statistics for native ID2 N-HLH82-L construct 48! Table 11: Positions of residues thought to be important for heterodimerization with MYOD1 62! Table 12: Constructs created for use in protein expression for E47 and MYOD1 human proteins showing their theoretical biochemical properties estimated by ProtParam(Wilkins, et al., 1999) 100! Table 13: Changes to ID2 protocol for expression and purification of ID2 and ID3 mutants 103! Table 14: ID1 & ID2 constructs and their theoretical biochemical properties estimated by ProtParam(Wilkins, et al., 1999) 104! Table 15: Changes to ID2 protocol for expression and purification of ID1 and ID3 HLH domains 104! ! vi! LIST OF FIGURES Figure 1: Hydrophobic core packing of bHLH-containing proteins 5! Figure 2: Cartoon representation of ID3 (PDB: 2LFH) NMR structure Monomer shown as dark blue N-terminal residual tag, green unfolded N-terminus, pale red helix 1, green loop, red helix 6! Figure 3: T-coffee multiple alignment of full length ID proteins to show the highly conserved HLH region and the divergent N & C-termini with only a few small regions of similarity such as the D-box (destruction box) element 9! Figure 4: Reported binders and non-binders of ID proteins The general structure of binders had shorter helices unlike non-binders, such as MYC, which had the additional leucine zipper Overall, topology conformed to the same helical bundle 10! Figure 5: Cartoon representation of ID3 (PDB: 2LFH) aligned with E47 (Ellenberger private communication) to illustrate how the heterodimerization might take place ID3 in red, E47 in blue 11! Figure 6: Representative small-scale protein expression tests 33! Figure 7: Stability of HLH24-82-L containing polypeptide stabilizer over days at room temperature (25°C) SDS-PAGE 12% gel: marker (lane M), Day (lane 1), Day (lane 2), Day (lane 3), Day (lane 4) 35! Figure 8: ID2 proteins’ expression and purification 37! Figure 9: ID2 proteins’ purity check by SDS_PAGE: marker (lane M, kDa) N-HLH82L (gel A, lane 1), HLH24-82-L (gel B, lane 2), HLH24-82-L-Se-Met (gel C, lane 3) 38! Figure 10: HLH24-82-L crystals in 0.1 M MES pH 6.5, 2.5 M Lithium Acetate grown at 18°C 41! Figure 11: Crystals from manual hanging-drop optimization grown at 18°C 42! Figure 12: HKL view of reflections in the kl plane in reciprocal space for N-HLH82-L crystal at 2.1Å resolution 45! Figure 13: Ramachandran plot of ID2 N-HLH82-L by RAMPAGE (http://wwwcryst.bioc.cam.ac.uk/rampage/) (Lovell, et al., 2003) 49! Figure 14: Diagrammatic representation of ID2 HLH structure 51! Figure 15: Cartoon representation of the crystal structure of ID2 at 2.1Å resolution showing the positive loop ion and missing basic region 52! Figure 16: Ribbon representations of ID2 homodimer interactions ID2: chain A in purple, chain B in brown, loop in green and potassium ion in grey 55! Figure 17: Loop region mutants of ID2 and ID3 SDS-PAGE: marker (kDa, lane M), before induction (lane U), insoluble pellet fraction (lane P), soluble fraction (S) ! vii! Red boxes denote expected expression region Gel A and B expression vector was pDest-565 induced at 17°C Gel C expression vector was pDestHisMBP induced at 17°C 56! Figure 18: Predicted interactions based on ID3 homology model (Wibley, et al., 1996) were not found in either the ID2 crystal structure nor ID3 NMR structure 58! Figure 19: Structural alignment of the bHLH domain of ID proteins and their binding partners Alignments were done manually using Pymol’s align function as a guide 58! Figure 20: E47 homodimer showing the network of glutamines that were predicted to form hydrogen bonds but the distances were too far for most of them Perhaps E47 also had a positive ion in the loop coordinated by two of the glutamines that held it rigid? (grey sphere) 60! Figure 21: Ribbon representation of ID2 and ID3 opposing chains to illustrate residues thought to play an important role in heterodimerization with MYOD1 Residues from ID2 (Y37, D41) and ID3 (D42, H46) pointed away from the dimer interface ID2-K47 and potentially ID3-R52 had interactions with the loop ion that was necessary for homodimer formation of ID2 63! Figure 22: EMSA controls 64! Figure 23: EMSA 6% native gel showing that increasing concentration of ID2 inhibited E47 binding to DNA Lanes without ID2 (lanes and 2) denoted by “-“ Number of “+” denoted relative concentration of ID2 added All lanes contained μM E47 This showed that the purified ID2 used for crystallization was active 65! Figure 24: EMSA 6% native gel showing the different ID-HLH binding affinities to 0.05 µM human E47 Residues for each human ID protein given in parentheses “+” denoted presence of E47 All lanes contained 200nM DNA Concentrations of each ID protein provided in the table above the gel All ID proteins bound E47 to varying degrees 66! Figure 25: EMSA 6% native gel showing the different ID-HLH binding affinities to 0.2 µM human MYOD1 (tagged with His-MBP) Residues for each human ID protein given in parentheses “+” denoted presence of MYOD1 All lanes contained 100nM DNA Concentrations of each ID protein provided in the table above the gel ID1 and ID2 showed weak interactions with MYOD1 where a large fraction seemed to form an intermediate rather than complete inhibition ID3 did not bind MYOD1 67! Figure 26: EMSA 6% native gel showing the different ID-HLH binding affinities to 0.2 µM human MYOD1 (tagged with His-MBP) heterodimerized with E47 (0.05µM) Residues for each human ID protein given in parentheses “+” denotes presence of MYOD1 and/or E47 All lanes contained 200nM DNA Concentrations of each ID protein provided in the table above the gel MYOD1 had high propensity to bind E47 All IDs showed the same binding pattern as seen in Figures 24 and 25 67! ! viii! (Depinho, et al., 1986), Class IV contained Myc-dimerizing proteins such as Mad (Ayer, et al., 1993) and Max (Blackwood, et al., 1991) Class V contained the inhibitors of Class I and II proteins such as ID (Benezra, et al., 1990) and Class VI contained proteins like Hairy (Klambt, et al., 1989) which has a proline in its basic region Finally, Class VII contained proteins like Arnt (Crews, 1998) which has a bHLH-PAS domain Some years later, a different classification method based on phylogenetic profiling of 242 HLH-containing proteins using evolutionary relationships gave rise to four major groups A-D (Atchley, et al., 1997) Each group was based not only on their DNA-binding specificities but also on conservation of residues at specific positions Group A contained all proteins that bound to the specific E-box motif (CAGCTG) such as E12 and MyoD Group B contained those that bound to motif CACGTG such as Myc and HAIRY These two groups were further characterized based on specific amino acids at defined positions Group C contained Group B derived proteins but with no defined amino acid configuration such as Arnt and finally Group D proteins were those without the basic domain such as ID (Atchley, et al., 1997) More recently, with many newly sequenced genomes, these original phylogenetic groups were updated and extended to incorporate two new groups, E and F (Ledent, et al., 2002) Group A now contained CAGCTG or CACCTG binding proteins such as MyoD and E47 Group B contained CACGTG or CATGTTG binding proteins such as Myc, Mad and Max Group C contained the PAS domain containing proteins that bound to ACGTG or GCGTG such as Arnt Group D remained the same, containing proteins that lacked a basic domain such as ID The new Group E contained Hairy which bound preferentially to N-box motifs (CACGCG or CACGAG) and contained an additional orange domain And finally Group F contained proteins with an additional COE domain such as Coe (Vervoort, et al., 1999) ! 2! 1.2 bHLH structures ! The protein data bank (PDB) (Berman, et al., 2000) currently has over 60 bHLH transcription factors representing a cross section of five out of the seven different groups of bHLH-containing proteins as either homo- or heterodimers bound or unbound to DNA as shown in Table As the structures show, bHLHs exist as dimers conforming to a parallel 4-helix bundle with varying N and C termini The dimeric form is the functional form of bHLHs and monomers are unable to activate transcription (Murre, et al., 1989) ! 3! Table 1: Representative structures of bHLH-containing proteins from the PDB for each group Protein E47NeuroD1 heterodimer on DNA 2QL2 A (Longo, et al., 2008) c-Myc (red) –Max (blue heterodimer on DNA 1NKP B (Nair, et al., 2003) Arnt-HIF2A heterodimer 2A24 C (Card, et al., 2005) ID3 homodimer 2LFH D unpublished EBF3 (COE3) ! PDB ID 3N50 F (Siponen, et al., 2010) Structure Group Reference 4! It was found that dimerization required only the HLH domain (Murre, et al., 1989, Sun, et al., 1991, Voronova, et al., 1990) Hydrophobic packing between the adjoining helices of the two monomers created a core that stabilized the HLH into the recognized four-helix bundle (Figure 1) (Ellenberger, et al., 1994, Ma, et al., 1994) Besides these hydrophobic interactions, Group A proteins like E47 have been shown to contain a network of hydrogen bonds that stabilized the loop (Ellenberger, et al., 1994, Ma, et al., 1994) Figure 1: Hydrophobic core packing of bHLH-containing proteins ! ! 5! Group B proteins like Myc-Max contained an extra leucine zipper (Leucines at every 7th position) C-terminal of the HLH in addition to the basic residues N-terminal of the HLH (Murre, et al., 1989) The leucine zipper possibly aided in dimerization by adding to the hydrophobic core that brought the monomers together (Kajimoto, et al., 1994, Landschulz, et al., 1988) The very recently deposited ID3 NMR structure (Group D) had none of these added structures and contained only the HLH (Figure 2) with a random coil at the N-terminus The C-terminus from residue 84 onwards was not included in the deposition Figure 2: Cartoon representation of ID3 (PDB: 2LFH) NMR structure Monomer shown as dark blue N-terminal residual tag, green unfolded N-terminus, pale red helix 1, green loop, red helix ! 6! For structures with DNA bound, arginines at fixed positions within the basic domain conferred DNA-binding specificity to the E-box element (Murre, et al., 1989) Binding would take place as the basic residues of the dimer fixed in the major groove of the DNA and held together by hydrogen bonds between these residues and the phosphates in the E-box motif (Ellenberger, et al., 1994, Sun, et al., 1991) Deletion of this region or point mutations to conserved residues abolished DNA-binding but did not prevent dimerization (Voronova, et al., 1990) Some bHLH transcription factors were able to form homodimers on DNA and activate transcription of downstream genes such as the muscle creatine kinase (MCK) gene An example of this was the E47 homodimer which strongly activated the immunoglobulin genes (Church, et al., 1985) Others such as the tissue-specific MyoD, homodimerized weakly on DNA but were not efficient in transcriptional activation unless heterodimerized with a ubiquitous HLH such as E47 (Lassar, et al., 1991) 1.3 IDs are Group D, HLH-containing proteins ! The ID proteins fall into the D group of HLHs The ID genes were named for their roles as inhibitors of DNA-binding and differentiation They were unique in that they did not contain the customary basic domain and therefore had no propensity to bind DNA (Benezra, et al., 1990) Instead, they disrupted the DNA-binding ability of a variety of transcription factors that contained a basic-helix-loop-helix (bHLH) (Benezra, et al., 1990, O'Toole, et al., 2003) motif by heterodimerizing with them As such, their primary function was within the nucleus (Kurooka, et al., 2005, O'Toole, et al., 2003, Tu, et al., 2003) of many tissues where they exhibited distinctive expression patterns particularly during growth and development (Cooper, et al., 1997, Jen, et al., 1996, Jen, et al., 1997) In general, ID mRNAs were detected at high ! 7! levels during development but were reduced in mature, differentiated tissues (Israel, et al., 1999) Although the full complement of ID binding partners has yet to be established, several studies have shown that many of them are transcription factors that are specific to Group A As previously mentioned, this group contained the ubiquitously expressed E proteins (e.g E12/ELSPBP1, E47/TCF3) and the tissue specific myogenic and neurogenic proteins (e.g MYOD1, NEUROD1) (Massari, et al., 2000, Murre, et al., 1989) In binding to these transcription factors, ID proteins inactivated their transcriptional function and in so doing, regulated cell fate and differentiation not only in muscle tissue but in a variety of cell lineages (Benezra, et al., 1990, Jen, et al., 1992, Kee, 2009, Yokota, 2001) as well There was also evidence to suggest that IDs bound to non-bHLH-containing proteins (Hara, et al., 1996, Iavarone, et al., 1994, Lasorella, et al., 1996) such as retinoblastoma protein (pRb), a tumour suppressor This added promiscuity enhanced their functionality into the areas of cell cycle and tumourigenesis (Norton, 2000) besides their known roles in modulating myogenesis and neurogenesis There were four mammalian ID paralogs discovered by various groups (Benezra, et al., 1990, Biggs, et al., 1992, Christy, et al., 1991, Riechmann, et al., 1994, Sun, et al., 1991) The mammalian family consisted of four members, namely ID1, ID2, ID3 and ID4 (Norton, 2000) and the human forms were mapped to chromosomes 20q11, 2p25, 1p36 and 6p22 respectively (Norton, et al., 1998) Ensembl (Flicek, et al., 2012) genome browser reported that orthologous ID proteins had an overall identity of over 90% in primates and over 80% in higher vertebrates They even shared an average 25% identity to the emc gene in Drosophila (Riechmann, et al., 1994) whose function was similar to the dominant negative regulation of bHLH transcription factors in its mammalian counterparts (Campuzano, 2001) Multiple alignment by T-coffee ! 8! (Di Tommaso, et al., 2011) is given in Figure to show the peptide sequence conservation of the HLH region Figure 3: T-coffee multiple alignment of full length ID proteins to show the highly conserved HLH region and the divergent N & C-termini with only a few small regions of similarity such as the D-box (destruction box) element ! 9! Although the longest human isoforms varied in size - ID1 has 155 residues, ID2 has 134 residues, ID3 has 119 residues and ID4 has 161 residues - their gene structure and organization were highly similar at the intron-exon boundaries Thus, it had been suggested that they evolved via duplication events from a single ancestral gene (Deed, et al., 1994, Mantani, et al., 1998, Mathew, et al., 1995, Rigolet, et al., 1998) Protein sequence alignments of ID paralogs showed a high degree of conservation within the HLH domain but a divergence at the N and C-termini apart from small pockets of similarity (Pagliuca, et al., 1995) This HLH domain conservation was preserved beyond the IDs to other HLHs such as E47, MYOD, Max and Mad (Phillips, 1994) from multiple sequence alignment analysis Overall protein identity across the different IDs and other bHLH-containing proteins was fairly low, averaging around 35% Yet, crystal structures of E47, MYOD, Myc, Max and Mad showed a similar topology of the conserved HLH region consisting of a parallel, fourhelix bundle (Wibley, et al., 1996) (Figure 4) Figure 4: Reported binders and non-binders of ID proteins The general structure of binders had shorter helices unlike non-binders, such as MYC, which had the additional leucine zipper Overall, topology conformed to the same helical bundle ! 10! With such a high conservation of both the HLH domain and structure, Wibley et al used these structures as a template to create a 3D homology model of ID3 in order to predict how the ID homodimers could exist as well as how they functioned to disrupt DNA-binding in their heterodimerization with other bHLH transcription factors (Wibley, et al., 1996) They postulated that IDs were able to homodimerize without DNA for stability mainly because of better hydrophobic core packing and that they did not bind DNA due to a non-basic, coiled-coil structure in the corresponding basic region of other bHLH-containing proteins (Wibley, et al., 1996) Since then, the NMR structure of ID3 (2LFH) (Eletsky, et al., 2011) was deposited into the PDB and will be discussed in detail in a future chapter To illustrate a possible heterodimer of ID3 with E37, the structures were structurally aligned and a monomer from each removed to reveal the much shorter helix of ID3 at the N-terminal end where the basic domain of E47 would normally be located in (Figure 5) Figure 5: Cartoon representation of ID3 (PDB: 2LFH) aligned with E47 (Ellenberger private communication) to illustrate how the heterodimerization might take place ID3 in red, E47 in blue The next few sections serve to highlight the role of IDs in development, neurogenesis, myogenesis and tumourigenesis ! 11! 1.4 ID proteins in development ! Splice variants of the E2A gene, known as the E-proteins, were made up of E47, E12 and E2-5 (Murre, et al., 1989, Sun, et al., 1991) E47 and E12 proteins were ubiquitously expressed in various tissues such as the brain, heart, liver and skeletal muscle (Murre, et al., 1989, Watada, et al., 1995) In particular, E47 was able to homodimerize on the E-box motif (CANNTG) of enhancer regions of a number of genes to activate their transcription One such example was the enhancer region of the heavy and light chain of the immunoglobulin genes where two putative E47 binding sites could be found (Aronheim, et al., 1991, Ephrussi, et al., 1985) Transcription of immunoglobulin genes via the bound E-proteins to these E-boxes led to B cell differentiation (Henthorn, et al., 1990) Confirmation of E47’s role in B cell differentiation was captured in EMSA experiments where E47 was complexed with the two putative E-box sequences (BCF1 and BCF2) (Murre, et al., 1991) It was with BCF1 that saw the co-migration of the E47 homodimer in the EMSA gel (Murre, et al., 1991) It was also found that the presence of BCF1 changed with the transition of pre-B to mature B cells and consequently changed the expression of E47 at the different stages of B cell differentiation (Bain, et al., 1993) To study the regulatory role of ID proteins in development, several types of experiments were performed to look at gain-of-function and loss-of-function phenotypes in different organs Genetically modified mice were generated and studied for ID1, ID2 and ID3 knock-out phenotypes (Lyden, et al., 1999, Rivera, et al., 2000, Yan, et al., 1997, Yokota, et al., 1999) Many in vivo functional studies of ID protein overexpression were also carried out (Cai, et al., 2000, Kim, et al., 1999, Martinsen, et al., 1998, Morrow, et al., 1999, Sun, 1994, Wice, et al., 1998) E2A knock-out mice were created to target and disrupt the bHLH domain (Bain, et al., 1994) Given E2A gene products’ role in all stages of B cell differentiation, the ! 12! mouse phenotype did not produce B cells During initiation of B cell differentiation, ID1 and ID2 mRNA levels were found to be high and decreased as the cells matured (Sun, et al., 1991) Since the E2A proteins were found mostly only after the pro-B cell stage (Jacobs, et al., 1993), it was postulated that prior to this, the ID proteins bound the E-proteins and inactivated them till they were ready to progress to the next stage of differentiation (Sun, et al., 1991) Transgenic mice overexpressing ID1 showed severe impairment in B cell differentiation as was expected since ID1 would have sequestered the E2A proteins, thus disabling the transition of pro-B to pre-B cells (Sun, 1994) In another study, ID2 expression in the chick was found to play a role in cell fate determination of neural crest cells At different stages of chick embryogenesis, ID2 was expressed in different areas of the neural system with early expression in the dorsal cranial neural fold and at a later stage in the neural tube, as well as in some migrating cranial neural cells (Martinsen, et al., 1998) By overexpressing ID2 near the embryonic surface, the ectoderm covering the neural tube vanished and these cells became neural crest cells There was also an uncontrolled growth of the dorsal neural tube Overexpression of ID2 in the mesenchymal layer did not cause these types of effects, suggesting that there was a bHLH partner that ID2 bound to in the ectodermal layer whose regulation by ID2 determined cell fate In a separate study, knock-out of ID2 caused lactation defects in female mice only during pregnancy due to a lack of mature lactating cells, which was otherwise normal (Mori, et al., 2000) ! 13! 1.5 ID proteins and myogenesis ! Muscle cell fate was another process controlled by ID proteins MyoD (aka MYOD1), a master regulator of myogenesis is expressed only in skeletal muscle cells in vivo and myogenic cell lines in vitro (Davis, et al., 1987) Davis showed that transfected MyoD cDNA converted mouse fibroblasts to myoblasts As previously discussed, MyoD homodimerized poorly and only had strong activation of downstream gene transcription when heterodimerized with an E-protein such as E47 Similar to the E-proteins, myogenenic bHLH transcription factors also bound to the Ebox motif, an example being in the enhancer region of the muscle creatine kinase gene (Lassar, et al., 1989) Lassar’s experiments provided evidence that heterodimers of E47 and MyoD existed in muscle cell extracts and that the addition of anti-sense E2A to C3H 10T1/2 cells transfected with myoD were less terminally differentiated Hence, an E2A-MyoD complex was deduced to drive myogenesis ID1 and ID3 were shown to bind only weakly to MyoD as compared to E12 (Langlands, et al., 1997) In vivo experiments using fluorescence resonance energy transfer (FRET) analysis showed the ID1-E2A interaction to be stronger than the MyoD-E2A interaction (Lingbeck, et al., 2008) Taken together, these observations supported the hypothesis that myogenesis was regulated indirectly by ID proteins bound to E-proteins, thus taking away MyoD’s heterodimerization partners and hence, the ability of MyoD to initiate transcription 1.6 ID proteins and neurogenesis ! bHLH transcription factors involved in neurogenesis included the human MASH1 (Guillemot, et al., 1993) and NEUROD (Lee, et al., 1995) genes Similar to their group B counterparts, they heterodimerized with E-proteins on the E-box DNA motif (Murre, et al., 1989) to drive transcription that led to neuronal differentiation ! 14! The structure of mouse NeuroD1-E47 complexed on DNA (E-box element) was deposited in the PDB as 2QL2 (Longo, et al., 2008) The structure together with the EMSA experiments clearly showed the existence of the heterodimer bound to DNA MASH1 was shown to be present in increasing fashion during murine embryonic development via northern blot analysis (Guillemot, et al., 1993) Furthermore, it was found to be expressed in both the central (CNS) as well as peripheral (PNS) nervous systems (Guillemot, et al., 1993, Lo, et al., 1991) MASH1 null mice were lethal within 24 hours of birth and lacked a nerve supply to the digestive and respiratory tracts as well as underdeveloped olfactory neurons (Guillemot, et al., 1993) These results strongly suggested a role for MASH1 in neurogenesis All ID proteins were expressed in developing neuroblast cells at different stages Interestingly, ID2 and ID4 were also expressed in some adult specific neurons (Neuman, et al., 1993, Riechmann, et al., 1995, Tzeng, et al., 1998) Inactivation of a single ID gene did little in early mouse development Similarly, in neurogenesis, individual ID1 and ID3 null mice showed no detectable changes in phenotype until both genes were knocked out which led to embryonic lethality and smaller brains (Lyden, et al., 1999) However, in the ID2 knock-out, there was a 25% chance of lethality as well as significant changes to the olfactory and frontal cortex (Yokota, et al., 1999) ID4 knock-outs showed extreme malformation of the mouse brains and stem cells isolated from these embryos did not differentiate Taken together, the ID proteins had intricate roles that overlapped but at the same time were distinct from each ID or set of IDs based on timing of expression and locality (Yun, et al., 2004) Although MASH1 and the ID proteins clearly had roles in neurogenesis, EMSA studies described in the results section will clearly show that ID1, ID2 and ID3 not bind to human MASH1 Instead, it was hypothesized that IDs worked in a similar ! 15! fashion in neuronal cells as in myoblast cells by sequestering E-proteins that were required for the assembly of functional transcriptional units of MASH1 and NEUROD1 1.7 IDs in cancer ! ID gene expression had been reported to be deregulated in many tumour types ID1 had been found to be elevated in breast (Lin, et al., 2000), thyroid (Kebebew, et al., 2000), endometrial (Takai, et al., 2001), cervical (Schindl, et al., 2001), ovarian (Schindl, et al., 2003), esophageal (Hu, et al., 2001), melanoma (Polsky, et al., 2001) and prostate (Ouyang, et al., 2002) cancers ID1, ID2 and ID3 had all been found in head and neck (Langlands, et al., 2000), colorectal (Wilson, et al., 2001), astrocytic (Vandeputte, et al., 2002) and pancreatic (Maruyama, et al., 1999) cancers Finally, all four IDs had shown to be expressed in testicular cancers (Sablitzky, et al., 1998) In some cases, high levels of ID were associated with disease severity and poor prognosis Even though recent reports of a frequent structural variation in human locus 20q11.21 containing ID1 was found in different types of cancer (Beroukhim, et al., 2010), so far, there have been no reports of ID specific gene defects in tumours Hence, it was thought to be more likely that up-regulation of these genes had implications in cell immortality, thus generating an increased propensity for tumourigenesis Arrested cells showed low to no expression of ID genes Mitogen stimulated fibroblast cells, on the other hand, showed induced ID expression that was sustained throughout G1 phase of the cell cycle and increased as cells progressed to the S phase (Hara, et al., 1994, Norton, et al., 1998) By directly interacting with some nonbHLH-containing proteins, ID proteins were able to influence cell cycle progression In the above example, ID proteins antagonistically interacted with ETS domain proteins such as SAP1 and ELK1 to down-regulate c-fos and egr-1 resulting in an inhibition of the MAP kinase signaling pathway (Yates, et al., 1999) By sequestering ! 16! ... et al., 19 99) 10 0! Table 13 : Changes to ID2 protocol for expression and purification of ID2 and ID3 mutants 10 3! Table 14 : ID1 & ID2 constructs and their theoretical biochemical. .. al., 19 90, Biggs, et al., 19 92, Christy, et al., 19 91, Riechmann, et al., 19 94, Sun, et al., 19 91) The mammalian family consisted of four members, namely ID1, ID2, ID3 and ID4 (Norton, 2000) and. .. al., 19 99) 10 4! Table 15 : Changes to ID2 protocol for expression and purification of ID1 and ID3 HLH domains 10 4! ! vi! LIST OF FIGURES Figure 1: Hydrophobic core packing of