Structural characterization and biochemical analysis of ID2, an inhibitor of DNA binding 9

Figure 22: EMSA controls A ID proteins did not bind e-box DNA, only E47, a bHLH transcription factor bound to DNA B bHLH transcription factors E47, Mash1, MYOD1 did not bind to mutant

CHAPTER 5: RESULTS and DISCUSSION

(Biochemical Studies)

!

!

5.1 ID2 protein activity

To ensure that the purified ID2 protein was functionally active, competitive

electromobility shift assays (EMSAs) were performed to look for loss of binding of a

Group A bHLH transcription factor to its cy5-labeled DNA probe with an increasing

concentration of ID2

E47, a Group A bHLH transcription factor and known binding partner of ID2 was

cloned, expressed and purified (Appendix 3) for use in these studies E47 bound an

E-box element found in the DNA sequence of the MCK promoter containing the motif

(CANNTG) The sequence GGATCCCCCCAACACCTGCTGCCTGA was ordered as

cy5-labeled forward and reverse probes and annealed in a thermal cycler Negative

controls were used to ensure that ID proteins did not bind DNA (Figure 22A) and a

mutant e-box element showed that E47 only bound to the wild-type e-box element

(Figure 22B)

Figure 22: EMSA controls

(A) ID proteins did not bind e-box DNA, only E47, a bHLH transcription factor bound to DNA (B) bHLH transcription factors (E47, Mash1, MYOD1) did not bind to mutant e-box DNA The ID proteins did not bind the mutant e-box probe

!

! 65!

ID2 at different concentrations was incubated with E47 for 10 mins at room

temperature in EMSA binding buffer Cy5-labeled DNA was added and allowed to

compete with ID2 for E47 binding for another 15 mins at room temperature to a final

reaction volume of 20 µl The mixtures were loaded on a native gel and scanned for

cy5 intensity Results showed that the ID2 N-HLH82-L protein used in the

crystallization experiments was successful in binding E47 and was therefore

functionally active (Figure 23)

Figure 23: EMSA 6% native gel showing that increasing concentration of ID2 inhibited E47 binding to DNA Lanes without ID2 (lanes 1 and 2) denoted by “-“ Number of “+” denoted relative concentration of ID2 added All lanes contained 2 μM E47 This showed that the purified ID2 used for crystallization was active

!

!

5.2 ID heterodimer binding specificity and affinity

Although the ID2 and ID3 structures revealed a very similar fold to other classes

of bHLH-containing proteins such as E47, MYOD1, NeuroD1 and MAX, previous

studies showed that the ID proteins did not bind to any class of bHLH-containing

proteins Instead, they preferentially bound to Group A bHLH-containing proteins

Previous reports showed that ID1 and ID2 bound to E47 and MYOD1 (Group A)

Using EMSAs, it was found that ID1, ID2 and ID3 all bound to E47 (Figure 24) ID1

and ID2 but not ID3 bound weakly to MYOD1 (Figure 25) This confirmed the

reported observations by Langlands (Langlands, et al., 1997) The E47-MYOD1

heterodimer EMSA experiment (Figure 26) showed the same binding pattern as with

E47 and MYOD1 homodimers This affirmed that IDs preferentially bound E47 and

likely inhibited MYOD1 indirectly by binding to MYOD1’s functional dimeric partner It

was clear from the EMSA that MYOD1 preferentially bound E47 than itself and

confirmed MYOD1’s weak homodimerization capablities

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

!

!

!

! 67!

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

25

5.3 ID helix-1 residues in binding specificity

Since IDs shared an average 80% identity across the HLH domain, differences

between ID2 and ID3 had to account for the differential binding to MYOD1

Langlands had proposed that three residues in helix-1 were important for binding

specificity using a mammalian-2-hybrid system (Langlands, et al., 1997) In the ID2

structure, these three residues pointed away from the dimer interface (Figure 21)

The only residue that could potentially be affected was the lysine at position 47

because of its interaction with the positive ion in the loop To find out if these

residues alone played a role in binding specificity, equivalent human positions from

the Langlands paper were mutated in ID2 as follows Y37D (Y in ID1), D41H (G in

ID1), K47R (K in ID1) as well as a double mutant Y37D-D41H The expectation was

that binding to E47 would not change but binding to MYOD1 (Appendix 3) would be

affected Using EMSAs and comparing binding of these mutants to E47 (Figure 27)

and MYOD1 (Figure 28), it was found that there were no significant differences and

they all bound like wild-type ID2 As with the previous EMSA for wild-type ID proteins,

the mutants showed similar binding patterns for the MYOD1-E47 heterodimer as with

MYOD1 and E47 homodimers (Figure 29) Hence, these residues were unlikely to

confer binding specificity of the different ID proteins as was hypothesized from the

ID2 and ID3 structures

! 69!

Figure 27: EMSA 6% native gel showing ID2 helix-1 mutants binding affinities to 0.2 µM human E47 “+” denotes presence of E47 All lanes contained 100nM DNA Concentrations of each ID protein provided in the table above the gel All mutants bound to E47

!

Figure 29: EMSA 6% native gel showing ID2 helix-1 mutants binding affinities to 0.2 µM human MYOD1 (HisMBP tagged) heterodimerized with 0.2 µM E47 “+” denotes presence of MYOD1 and/or E47 All lanes contained 100nM DNA Concentrations of each ID protein provided in the table above the gel IDs bound with similar affinities as with the E47 and MYOD1 homodimers

! 71!

5.4 Exploring other differences in ID residues

Other residues where ID1 and 2 differed from ID3 were at ID2 positions Q55, N56,

K58, K61 and H67 Of these, Q55 and K61 had associated interactions in the

structure; Q55 to the ion and K61 as an intra-chain hydrogen bond to N40 The latter

bond was also found in the E47 homodimer at N353-K375 and was not lost in the

heterodimeric structure, E47-NeuroD1, at E47.N559-E47.K581 Mutations were

made to these residues to change their size as well as mirror the opposing ID protein

ID2 mutations were Q55A, Q55R, K61A, K61Q and corresponding ID3 mutations

were R60A, R60Q, Q66A and Q66K Of these mutations, ID2.K61A, ID3.Q66K and

ID3.Q66A did not produce soluble protein (Figure 17), suggesting a requirement for

this bond to form a stable monomer ID2.K61Q had expression similar to wild type

and bound both E47 (Figure 30, top gel) and MYOD1 (Figure 30, bottom gel) that

was similar to wt-ID2

Finally, a deletion at ID1.Q78 (ID2.Q55) caused partial binding loss to MYOD1

but not to E47 (Pesce, et al., 1993) It was found that ID2.Q55A and ID2.Q55R

showed partial binding loss to both MYOD1 (Figure 31, bottom gel) and E47 (Figure

31, top gel) We postulate that the alanine causes a loss of one of the interactions to

the positive ion and arginine would add a repelling force at that position The reverse

mutation in ID3 was striking ID3.R60A and ID3.R60Q had higher protein solubility

than wildtype, suggesting added stability All previous experiments were done with

the solubility tag, His-MBP attached to ID3 However, these 2 mutants were soluble

without the tag and EMSAs were performed with these untagged ID3 mutants It was

found that they bound better to E47 than wt-ID3, both with (Figure 32) and without

(Figure 33) the solubility tag R60A made no difference to MYOD1 binding But R60Q

showed partial binding to MYOD1

Pesce and Benezra had shown that the loop of ID1 was not as flexible and was

important for specificity (Pesce, et al., 1993) It is believed that this is also true for ID2

These results show that an ionic interaction in the loop makes for a more rigid loop

that when disrupted, causes a conformational change that is not conducive to

heterodimerization The hypothesis is that ID1 and 2 preferentially bind other

bHLH-containing proteins with a similar loop formation Further, the arginine at position 60

in the loop of ID3 could have implications for its preference to different binding

partners than ID1 and ID2

Figure 30: EMSA 6% native gels showing ID2 loop region mutants wt = wild-type ID2, E47

concentration=100nM, DNA concentration=100nM, MYOD1 concentration=200nM

Concentrations of ID2 are labeled Top gel shows binding to E47, bottom gel to MYOD1 Apart from the double mutant, the other ID2 mutants bound to E47 and MYOD1 as well as wild-type ID2

! 73!

Figure 31: EMSA 6% native gel showing ID2 loop mutants wt = wild-type ID2, E47

concentration=100nM, DNA concentration=100nM, MYOD1 concentration=200nM

Concentrations of ID2 are labeled Top gel shows binding to E47, bottom gel to MYOD1 Both mutants showed partial binding loss compared to wild-type ID2

Figure 32: EMSA 6% native gels showing ID3 loop region mutants wt = wild-type ID3 (His-MBP tag), E47 concentration=100nM, DNA concentration=100nM, MYOD1 concentration=200nM Concentrations of ID3 are labeled Top gel shows binding to E47, bottom gel to MYOD1 R60Q and R60A were both tagged with His-MBP R60Q appeared to bind better than wild-type ID3

! 75!

Figure 33: EMSA 6% native gels showing ID3 loop region mutants wt = wild-type ID3 (His-MBP tag), E47 concentration=100nM, DNA concentration=100nM, MYOD1 concentration=200nM Concentrations of ID3 are labeled Top gel shows binding to E47, bottom gel to MYOD1 R60Q and R60A were both untagged Tagged (Figure 32) or untagged, R60Q showed better binding than wild-type ID3

5.5 MASH1 and the ID proteins

As mentioned in the introduction, MASH1 transcription factor played a role in

neurogenesis It was hypothesized that IDs indirectly inhibited the function of MASH1

by sequestering E47, thus removing MASH1’s functional binding partner However,

there had been conflicting reports on whether IDs bound MASH1 directly An initial

report showed that all 4 ID proteins dimerized with E47 but not with human MASH1

(Jogi, et al., 2002) Later co-immunoprecipitation studies identified ID1 as a direct

binding partner of MASH1 (Obayashi, et al., 2009) As with the previous studies,

EMSAs were performed with MASH1 homodimer (Figure 34), MASH1-E47

heterodimer (Figure 34) and MASH1-MYOD1 heterodimer (Figure 35) None of them

showed ID1, ID2 or ID3 bound directly to MASH1 Based on all the EMSAs, it was

found that IDs bound strongly to E47, weakly to MYOD1 but not to MASH1

Figure 34: EMSA 6% native gel showing ID proteins bound to MASH1 (left gel) and MASH1-E47 heterodimer (right gel) E47 concentration=50nM, MASH1 concentration = 0.5µM, DNA concentration=100nM Concentrations of ID2 are labeled on top of the gels IDs did not bind MASH1, only E47

! 77!

Figure 35: EMSA 6% native gel showing ID proteins bound to MASH1-MYOD1 heterodimer MYOD1 concentration=0.2µM, MASH1 concentration = 0.5µM, DNA concentration=100nM Concentrations of ID2 are labeled on top of the gel IDs bound weakly to MYOD1 but did not bind

to MASH1

CHAPTER 6: CONCLUSION and FUTURE DIRECTIONS

!

!

6.1 Conclusions

Constructs were created to enhance solubility and stability by choosing to omit

the C-terminal region of ID2 that contained the D-box motif that would signal the

protein for degradation via APC and Cdh1 However, even with the D-box region

removed and the addition of surrounding residues to stabilize the HLH domain, the

protein was insoluble without the help of a solubility tag like His6-GST A cloning

artifact that ended up adding a short C-terminal polypeptide completely stabilized the

protein The polypeptide was initially used only for the main HLH domain (residues

24-82) However, it was found that even adding the full N-terminus did not stabilize

ID2 so the polypeptide was added to terminate ID2 (residues 1-82) and it was then

found to be stable and soluble at room temperature

With the short and long form of ID2, crystallization trials yielded crystals in

conditions of high salt The crystals were optimized for growth and tested for

diffraction quality before being shipped to a synchrotron light source for data

collection The short form of ID2 had very different looking crystals to the long form of

ID2 and ID2-Se-Met short form It was noted that the long form and the Se-Met form

had similar looking crystals despite being grown in completely different mother

liquors

The short form of ID2-Se-Met was solved as a SAD dataset using the Peak

energy The structure had poor statistics in the highest resolution shells and was

used as a template for molecular replacement of the long form The structure of the

long form of ID2 was solved at 2.1Å The ID2 short form without the Se-Met could not

be solved

Structural comparisons were made between the ID2 crystal structure and

published results about residues that were predicted to play an important role in

! 79!

dimerization There were 4 main regions within the protein that were deemed crucial

for dimerization

The first region was Helix-1 This region contained 3 residues that were

interchanged between ID1 and ID3 This residue exchange was proposed to confer

increased binding affinities of ID3 for MYOD1, which it did not otherwise bind

(Langlands, et al., 1997) The second region was a cysteine residue that was shown

to aid in homodimerization (Liu, et al., 2000) The third was the loop region where

mutations in certain residues abolished dimerization (Pesce, et al., 1993) And lastly,

a set of proposed bonds based on an ID3 homology model (Wibley, et al., 1996)

The helix-1 residues that were mentioned in the publication all pointed

outwards in the ID2 model This was confirmed in the ID3 NMR model The only

residue that could affect the dimer would be the lysine at position 47 in ID2, which

had a backbone oxygen that interacted with a loop ion believed to stabilize the

helices To test these hypotheses, mutants at these positions were made in ID2 to

mirror both ID1 and ID3 As expected, none of the mutants changed the binding

affinities of ID2 to either E47 or MYOD1 in competitive EMSA studies Even the

mutation at K47R did not change the binding much So it was concluded that these

residues were not important for dimer formation

The cysteine residue mentioned was proposed to create disulfide bonds to aid in

dimerization But the ID2 model showed that this residue pointed away from the

dimer interface and also pointed outwards in the two monomers The atomic distance

was too large for a disulfide bond However, it could not be discounted that such a

bond could transiently bring two monomers in close proximity together to form the

homodimer

The ID3 homolog model predicted 3 bond types, a hydrogen bond, a repelling

force and a salt bridge between 2 helices to form the homodimer However, an

inspection of both ID2 and ID3 showed that these bonds were unlikely due to the

sheer distances between the side chains

Finally, the loop region, at the first turn after helix-1, contained density in

Fo-Fc maps that was too large to be water A potassium ion was modeled in the density

based on the diameter of the density, the coordination of the surrounding residues

and the fact that the crystals were grown in a high concentration of potassium The

ion was found in the loops of both monomers and was coordinated by four residues,

one from helix-1 and three from the loop with distances within 3Å The closest

interaction was with I53 and the farthest was with Q55 A study of ID1’s equivalent

positions showed that mutations at these positions caused complete loss of

expression and a partial loss respectively (Pesce, et al., 1993) It is possible then that

ID1 shared this loop phenomenon especially since it was reported to have a rigid

loop It is highly likely that the ion has a part to play in the loop’s rigidity; probably to

aid in a stable monomer by holding the 2 helices together Looking at residues that

differed between ID2 and ID3 near the loop region, mutants were made to exchange

the ID2 for the ID3 residue and vice versa to see if there would be differences in

binding At residue position 55 in ID2, mutation of Q to an R/A caused binding to be

slightly worse than wild-type Conversely, mutation of the ID3 residue at the

equivalent position 61 from a K to and R or A conferred better binding than wild-type

It was possible that since there was some interaction of this residue with the loop’s

ion, that the positively charged arginine repelled the positive ion enough to cause a

change that did not favour the ion’s placement to hold the loop rigidly in place If

rigidity was required for sucessful heterodimer formation, it would explain the

differences in binding affinity

Other interactions that are conserved between the ID2 crystal structure and

the ID3 NMR structure were the hydrogen bonds and the hydrophobic core Wibley

! 81!

had noted that the hydrophobic core was all that was needed to hold the ID dimer

together but mutation studies to residues conferring interhelical hydrogen bonds such

as those of Y71 (chain A) and Q76 (chain B) completely abolished soluble protein

production and was required for proper folding of the monomer and hence, the

stability of the dimer

In conclusion, the loop region was shown to be critical to dimer formation as well

as overall stability of the monomers Apart from the loop, interhelical hydrogen bonds

also played a pivotal role in the protein’s stability

!

6.2 Future Directions

A detailed study of the loop region of ID2 would be prudent Loop swaps between

ID2 and the other IDs as well as with the binding partners to look for differences in

binding or perhaps even in vivo to allow transcription of a downstream gene

Residue by residue mutations at each of the positions near the loop till the end of

the loop to study if the predicted residues affecting binding to the ion would either

abolish soluble protein production or binding to E47

Only one hydrogen bond mutant was done so far The others such as the Y43

that were conserved in ID as well as other bHLH-containing proteins should be tested

Many proteins in the PDB are bound to divalent ions Experiments such as using

crown ether to chelate the potassium ion, as well as EDTA and EGTA to chelate

divalent ions would be a way to check that the ion really was a potassium, especially

since the crystallization solution had such a high potassium salt content and could

have competed off the biological ion In addition, the other ID proteins should also

have the same assay done to assess whether they also contained the loop ion and if

chelating the ion would result in misfolded protein

Crystal structures of the heterodimers of ID, such as ID2-E47 and/or ID3-E47

would provide a definitive answer as to the specific residues involved in these

interactions

A high-resolution crystal structure of E47 or biochemical assays would be

worthwhile to test for the presence of a loop ion as predicted from the ID2-E47 model

comparisons Mutations of residues predicted to hold the E47 loop in place could

also be tested

Mutant forms of ID2 that caused misfolded proteins could be tested in vivo in cell

culture systems to look for phenotypic changes in cancer cells as a way to validate

the hypothesis that the ion interactions and the specific hydrogen bonds were

necessary for ID function

! 83!

1 Adams, P D., Grosse-Kunstleve, R W., Hung, L W., Ioerger, T R., McCoy, A J., Moriarty, N W., Read, R J., Sacchettini, J C., Sauter, N K and Terwilliger, T C (2002) "PHENIX: building new software for automated crystallographic structure

determination", Acta Crystallogr D Biol Crystallogr, 58, 1948-54

2 Alani, R M., Hasskarl, J., Grace, M., Hernandez, M C., Israel, M A and Munger,

K (1999) "Immortalization of primary human keratinocytes by the helix-loop-helix

protein, Id-1", Proc Natl Acad Sci U S A, 96, 9637-41

3 Aronheim, A., Ohlsson, H., Park, C W., Edlund, T and Walker, M D (1991)

"Distribution and characterization of helix-loop-helix enhancer-binding proteins from

pancreatic beta cells and lymphocytes", Nucleic Acids Res, 19, 3893-9

4 Atchley, W R and Fitch, W M (1997) "A natural classification of the basic loop-helix class of transcription factors", Proceedings of the National Academy of

helix-Sciences of the United States of America, 94, 5172-6

5 Ayer, D E., Kretzner, L and Eisenman, R N (1993) "Mad: a heterodimeric

partner for Max that antagonizes Myc transcriptional activity", Cell, 72, 211-22

6 Bain, G., Gruenwald, S and Murre, C (1993) "E2A and E2-2 are subunits of

B-cell-specific E2-box DNA-binding proteins", Mol Cell Biol, 13, 3522-9

7 Bain, G., Maandag, E C., Izon, D J., Amsen, D., Kruisbeek, A M., Weintraub, B C., Krop, I., Schlissel, M S., Feeney, A J., van Roon, M and et al (1994) "E2A proteins are required for proper B cell development and initiation of immunoglobulin

gene rearrangements", Cell, 79, 885-92

8 Benezra, R., Davis, R L., Lassar, A., Tapscott, S., Thayer, M., Lockshon, D and Weintraub, H (1990) "Id: a negative regulator of helix-loop-helix DNA binding

proteins Control of terminal myogenic differentiation", Ann N Y Acad Sci, 599, 1-11

9 Benezra, R., Davis, R L., Lockshon, D., Turner, D L and Weintraub, H (1990)

"The protein Id: a negative regulator of helix-loop-helix DNA binding proteins", Cell,

61, 49-59

10 Berman, H M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T N., Weissig, H., Shindyalov, I N and Bourne, P E (2000) "The Protein Data Bank", Nucleic Acids

Res, 28, 235-42

11 Beroukhim, R., Mermel, C H., Porter, D., Wei, G., Raychaudhuri, S., Donovan, J., Barretina, J., Boehm, J S., Dobson, J., Urashima, M., Mc Henry, K T., Pinchback, R M., Ligon, A H., Cho, Y J., Haery, L., Greulich, H., Reich, M., Winckler, W., Lawrence, M S., Weir, B A., Tanaka, K E., Chiang, D Y., Bass, A J., Loo, A., Hoffman, C., Prensner, J., Liefeld, T., Gao, Q., Yecies, D., Signoretti, S., Maher, E., Kaye, F J., Sasaki, H., Tepper, J E., Fletcher, J A., Tabernero, J., Baselga, J., Tsao,

M S., Demichelis, F., Rubin, M A., Janne, P A., Daly, M J., Nucera, C., Levine, R L., Ebert, B L., Gabriel, S., Rustgi, A K., Antonescu, C R., Ladanyi, M., Letai, A., Garraway, L A., Loda, M., Beer, D G., True, L D., Okamoto, A., Pomeroy, S L., Singer, S., Golub, T R., Lander, E S., Getz, G., Sellers, W R and Meyerson, M (2010) "The landscape of somatic copy-number alteration across human cancers",

Nature, 463, 899-905

12 Biggs, J., Murphy, E V and Israel, M A (1992) "A human Id-like helix-loop-helix

protein expressed during early development", Proc Natl Acad Sci U S A, 89, 1512-6

13 Blackwood, E M and Eisenman, R N (1991) "Max: a helix-loop-helix zipper

protein that forms a sequence-specific DNA-binding complex with Myc", Science, 251,

1211-7

14 Brunger, A T., Adams, P D., Clore, G M., DeLano, W L., Gros, P., Kunstleve, R W., Jiang, J S., Kuszewski, J., Nilges, M., Pannu, N S., Read, R J., Rice, L M., Simonson, T and Warren, G L (1998) "Crystallography & NMR system:

Grosse-A new software suite for macromolecular structure determination", Grosse-Acta Crystallogr D

Biol Crystallogr, 54, 905-21

15 Cai, L., Morrow, E M and Cepko, C L (2000) "Misexpression of basic loop-helix genes in the murine cerebral cortex affects cell fate choices and neuronal

helix-survival", Development, 127, 3021-30

16 Campuzano, S (2001) "Emc, a negative HLH regulator with multiple functions in

Drosophila development", Oncogene, 20, 8299-307

17 Card, P B., Erbel, P J and Gardner, K H (2005) "Structural basis of ARNT PAS-B dimerization: use of a common beta-sheet interface for hetero- and

homodimerization", J Mol Biol, 353, 664-77

18 Christy, B A., Sanders, L K., Lau, L F., Copeland, N G., Jenkins, N A and Nathans, D (1991) "An Id-related helix-loop-helix protein encoded by a growth factor-

inducible gene", Proc Natl Acad Sci U S A, 88, 1815-9

19 Church, G M., Ephrussi, A., Gilbert, W and Tonegawa, S (1985)

"Cell-type-specific contacts to immunoglobulin enhancers in nuclei", Nature, 313, 798-801