Characterization of protein folding determinants for LIN-12/Notch-Repeats (LNRs) using Human Notch1 LNR-B as a model system

1
Characterization of protein folding determinants
for LIN-12/Notch-Repeats (LNRs) using Human
Notch1 LNR-B as a model system Sharline
Madera Advisor Dr. Didem Vardar-Ulu Wellesley
College, Massachusetts
NSF REU Chemistry Leadership Group Travel
Award Recipient NSF Award CHE-0754512

• Abstract

Metal Specificity
5 Constructs
LNRB_orig L N F N D P W K N C T Q S L Q C W K Y F S D G H C D S Q C N S A G C L F D G F D C Q R A E G Q Mut 1,5C?A LNRB_orig L N F N D P W K N A T Q S L Q C W K Y F S D G H C D S Q C N S A G A L F D G F D C Q R A E G Q
LNRB_short K N C T Q S L Q C W K Y F S D G H C D S Q C N S A G C L F D G F D C Q LNRB_delAB K N C T Q S L Q C W K Y F S D G H C D S Q C N S A G C L F D G F D C Q R A E G Q LNRB_delBC L N F N D P W K N C T Q S L Q C W K Y F S D G H C D S Q C N S A G C L F D G F D C Q Mut 1,5C?A LNRB_delBC L N F N D P W K N A T Q S L Q C W K Y F S D G H C D S Q C N S A G A L F D G F D C Q
LNRB_int D P W K N C T Q S L Q C W K Y F S D G H C D S Q C N S A G C L F D G F D C Q Mut 1,5C?A LNRB_int D P W K N A T Q S L Q C W K Y F S D G H C D S Q C N S A G A L F D G F D C Q
Human Notch1 is a member of a conserved family of
heterodimeric type 1 transmembrane receptors that
control differentiation in multicellular animals.
Notch proteins possess three contiguous
LIN-12/Notch-Repeats (LNRs), LNRA, LNRB and LNRC,
in their extracellular domain that maintain the
receptor in its resting conformation in the
absence of ligand. These conserved LNRs display a
characteristic disulfide bonding pattern and
require Ca2 for folding. In the receptor, they
are separated by two linkers, linker_AB and
linker_BC, 10 and 6 amino acids long,
respectively. Previously, we had shown that LNRB
folding was affected by residues 6-10 of
linker_AB and not of linker_BC (1, Figure 5). In
this study we investigated the dependence of LNRB
folding on the identity of the metal ion as well
as the number of conserved disulfide bonds. Our
results indicate that LNRB folding is selective
for Ca2 and the linker_AB length requirement and
total number of disulfide sulfide bonds needed
for effective folding are interdependent.
Together these studies represents the initial
steps toward defining the minimum requirements
for a correctly folding LNR module using LNRB
from human Notch1 as a model system.
Figure 7. Chromatograms of LNRB_orig and LNRB_int
folded under varying metal ion conditions. Panel
A Superposition of chromatograms of LNRB_orig
folded in the presence of 10mM CaCl2-red and 10mM
MgCl2-black. Panel B Superposition of
chromatograms of LNRB_orig folded in the presence
of 1mM CaCl2-green and 1mM ZnCl2-blue. These
chromatograms demonstrate the selectivity of
LNRB_orig folding for Ca2 which result in a
single folded peak, compared to other potential
divalent cations Mg2 and Zn2, which yield an
array of multiple peaks indicating the lack of
one predominant native fold.
Figure 4. Sequence Alignment for LNRB Constructs
used for the study Red Residues Ca 2
coordinating residues Blue Residues Mutated
cysteines to alanines Orange Residues
Disulfide bonded cysteines
linker_BC
linker_AB
C C
C C
C C
SLNFNDPWKN
QRAEGQ
LNRA
LNRC
LNRB
Table 1. Summary of HPLC Mass Spectrometry
Results

• Results

Figure 1. Human Notch 1 LNRs and linkers.
Construct Buffer B Elution Buffer B Elution 100mM DTT Correctly Folded Calculated MW (Da) Mass Spec MW (Da)
LNRB_orig 28 30 Yes 5368.8 5365.65
Mut LNRB_orig 28.2-30.2 --- No 5304.7 ---
LNRB_short 21-25 26 No 3940.3 3937.54
LNRB_int 25 27 Yes 4338.7 4291.93
Mut LNRB_int 22.7-32.3 --- No 4274.6 4227.36
LNRB_delAB 23-25 25 No 4481.8 4480.47
LNRB_delBC 27 30 Yes 4827.2 4823.22
Mut LNRB_delBC 26.4 --- Yes 4763.1 4760.86

• Introduction

Notch Proteins are large Ca2 binding,
transmembrane receptors that control
differentiation in multicellular animals. In
mammals, there are four Notch homologs Notch1-4.
These proteins function via a highly conserved
mechanism referred to as the Notch signaling
pathway, which is important for cell-cell
communication, involving gene regulation
mechanisms that control multiple cell
differentiation processes during embryonic and
adult life. Deregulation of normal Notch
activation has been noted in certain human
leukemias, (2) Alagille (3, 4) and CADASIL (5)
syndromes, indicating that perturbations of Notch
signaling underlie several forms of human
diseases (6). Notch proteins exhibit a highly
conserved modular architecture (Figure 2), in
which distinct repeated structural units are
associated with different functional roles in the
intact receptor (7). Ligand binding to the
N-terminal EGF-repeats activates these proteins
by facilitating a proteolytic cleavage by a
metalloprotease at site S2, and triggers the
gamma-secretase cleavage that permits the
translocation of intracellular Notch (ICN) into
the nucleus to activate the transcription of
target genes (8, 9, 10, 11). The Negative
Regulatory Region (NRR) of all Notch receptors
has three tandem, independently folding
LIN-12/Notch Repeats (LNRs) that wrap around the
HD domain containing the regulatory cleavage site
S2, and mask the S2 site in the resting receptor
(Figure 3) (12-14). Hence the interactions
between the LNRs and the HD are critical in
stabilizing the NRR and preventing activation
prior to ligand binding. Each of the LNRs
contains six cysteines with a unique disulfide
bonding pattern and coordinate a single Ca2
(Figure 3), however the minimum requirements that
would ensure an LNR to fold independently are not
known. This work utilizes different LNRB
constructs that all contain the 32 amino acid
stretch from cysteine 1 to cysteine 6 in the
second LNR of hN1and various numbers of the
residues that flank these residues (Table 1), to
define the minimum length requirement for hN1
LNRB and to investigate the impact of metal ions
and number of disulfide bonds on its autonomous
folding.
B
Figure 5. Chromatograms of LNRB constructs
LNRB_orig- green, LNRB_int- purple, LNRB_delBC-
blue. LNRB_delAB- brown, LNRB_short- orange Panel
A) Superposition of the chromatograms of the
three LNRB constructs that folded into a single
predominant peak indicating a single
thermodynamically favored correctly folded
species. Small neighboring peaks are indicative
of misfolded species. Top left inset
Representative chromatogram detailing the elution
gradient used and the pressure during the run.
Panel B) Superposition of the chromatograms of
the two unfolded constructs after dialysis 3.
Note no predominant peak is obtained
demonstrating no preference for correctly folded
species for these two constructs. Top right
insets for both panels Superposition of
chromatograms for the same constructs after DTT
treatment showing that all peaks of the folded
chromatograms collapse to a single peak when
reduced and elute later in the gradient (see
Table 1).

• Conclusions

In this study we investigated the relative
importance of the number of potential disulfide
bonds and the divalent ion identity for the
independent folding of LNRB in relation to the
presence or absence of the linker regions
flanking LNRB. Previous folding studies
identified constructs LNRB_delBC, LNRB_orig and
LNRB_int as autonomously folding constructs. In
this study we show two of the modified LNRB
constructs where cysteines 1 and 5 are mutated to
alanines to eliminate the potential of the first
disulfide bond, Mut 1,5C?A LNRB_int and Mut
1,5C?A LNRB_orig, are unable to fold into a
single thermodynamically favored folded state. In
contrast, Mut 1,5C?A LNRB_delBC retains the
ability to fold autonomously. These data show
that for the formation of a single
thermodynamically favored LNRB species upon in
vitro refolding there is a minimum requirement
for the total number of stabilizing interactions
that participate in the folding process. The
relative impact of each of these interactions,
which involve linker_AB, the disulfide bonds, and
the Ca2 ion, in ensuring autonomous folding are
highly interdependent. Furthermore the
attainment of one thermodynamically favored
species in the folding of LNRB is dependent and
selective for Ca2 compared to the two other
divalent cations abundant in cells, Mg2 and
Zn2.

• Future Directions

The findings of this study are the initial steps
in defining the minimum requirements for an
autonomously folding LNR module using LNR_B as a
model from human Notch1. They will be followed
by studies that include further investigations on
the effects of altering the cysteine arrangement
of other LNR modules with varying lengths and
direct correlation of these findings to hN4LNRA,
a wild type LNR module within the human Notch4
receptor and contains only two disulfide bonds.
Figure 2. Domain organization of the Notch
Receptor.
Figure 3. Crystal structure of Human Notch2 NRR
(15).
Figure 6. Superposition of chromatograms for
LNRB_delBC-black and Mutant LNRB_delBC-red.
Similar to LNRB_delBC, mutant LNRB_delBC displays
a single predominant peak. This finding
demonstrates that the mutant LNRB_delBC that
contains only two disulfide bonds is capable of
folding into a single thermodynamically favored
state whose stability is comparable to that of
wild type LNRB_delBC.

• Experimental

• References

• Wildtype hN1LNRB was expressed as inclusion
  bodies in BL21(DE3)PlysS E.coli. cell line as a
  fusion protein with a modified form of the trpLE
  sequence using the pMML vector (kind gift of S.
  Blacklow, BWH).
• The plasmids for the mutant hN1LNRBs were
  obtained from the corresponding wildtype pMML
  vector using the QuikChange Site-Directed
  Mutagenesis protocol (Stratagene) where cysteines
  1 and 5 were changed into alanines eliminating
  the potential for the first disulfide bond.
• LNRB was cleaved from the hydrophobic leader
  sequence by cyanogen bromide cleavage in 70
  formic acid and was separated from the leader
  sequence through precipitation of the leader
  sequence upon pH increase.
• Protein identity for each construct was
  confirmed by Mass Spectrometry.

• Soluble LNRB constructs (175 ?M) were folded
  for two days in a refolding buffer with daily
  buffer changes.
• 100mM NaCl
• 20mM Tris pH 8
• 10mM CaCl2/ 10mM MgCl2/ 1mM ZnCl2
• 2.5mM cysteine
• 0.5mM cystine

• Madera, S. and Vardar-Ulu D. Characterization of
  protein folding determinants for
  LIN-12/Notch-Repeats (LNRs) using Human Notch1
  LNR-B as a model system Poster Presentation,
  21st Annual Symposium of the American Protein
  Society, Protein Society, Boston July 2007.
• 2. Ellisen, L. W. et al. (1991) Cell.
  66649661.
• 3. Li, L., et al. (1997) Nat. Genet. 16
  243251.
• 4. Oda, T. et al. (1997) Nat. Genet. 16
  235242.
• 5. Joutel, A. et al. (1996) Nature. 383
  707710.
• 6. Rand, M. et al. (2000) Molec. and Cell. Biol.
  20 1825-1835.
• 7. Vardar, D. et al. (2003) Biochemistry. 42
  7061-7067.
• 8. Sanchez-Irizarry, C. et al. (2004) Molec. and
  Cell. Biol. 24 9265-9273.
• 9. Logeat, F. et al. (1998) Proc. Natl. Acad.
  Sci. USA. 95 8108-8112.
• 10. Brou, C. et al. (2000) Mol. Cell. 2
  207-216.
• 11. Lawrence, N. (2000) Development. 127
  3185-3195.
• 12. Aster, J. et al. (1999) Biochemistry. 38
  4736-4742.
• 13. Weng, A.P. et al. (2004) Science. 306
  269271.
• 14. Kopan, R. et al. (2000) Genes Dev. 14
  2799-2806.
• 15. Gordon, W. R. et al. (2007) Nature Structural
  and Molec. Biol. 14 295-300.

• On day 3 the constructs were moved into a
  dialysis buffer that did not contain any redox
  reagent (cysteine/cystine).
• All constructs from day 3 of dialysis were
  analyzed on a reverse phase HPLC using a C18
  column and 0.25/min gradient elution under the
  following experimental conditions
• Buffer A 10 Acetonitrile, 90 H2O, 0.1
  TFA
• Buffer B 90 Acetonitrile, 10 H2O, 0.1
  TFA
• A sample of each folded construct was also
  incubated with 100 mM DTT at room temp for 2 hrs
  and analyzed by RP-HPLC.

Figure 7. Panel A) Superposition of the
chromatograms of LNRB_orig-black and Mutant
LNRB_orig-red. Panel B) Superposition of the
chromatograms of LNRB_int-black and Mutant
LNRB_int-red. Unlike the wild type constructs
LNRB_orig and LNRB_int, mutant LNRB_orig and
mutant LNRB_int do not display a single peak that
corresponds to one thermodynamically stable
folded state. Instead the chromatograms of the
mutant constructs show various peaks indicating
the presence of several thermodynamically similar
species and not one favored folded state.