Results:

1
LIN12/NOTCH REPEATS (LNRs) IMPARTING TUNABLE
REDOX RESPONSIVENESS IN LESS THAN 35 AMINO ACIDS
THROUGH AN INTERPLAY BETWEEN Ca2 BINDING AND
DISULFIDE BOND FORMATION Janelle L. Jackson,
Angie Seo, Didem Vardar Ulu, Wellesley College,
Chemistry Department, Wellesley, MA, USA.

Introduction Many multi-domain proteins contain
small protein modules whose global folds are
stabilized by metal binding or disulfide bonds
rather than an extensive hydrophobic core or
secondary structures. Lin-12/Notch Repeat (LNR)
is such a module, first identified in Notch
receptors and more recently, within functionally
unrelated multi-domain proteins, such as
pregnancy associated plasma proteins and stealth
proteins. Prosite database defines LNR
(PDOC50258) as a 35 amino acid module with three
conserved Asp/Asn residues and six Cys residues
engaged in a particular disulfide pattern favored
by the presence of Ca2. However, homology
searches reveal naturally occurring LNRs with
only four of the conserved Cys residues, as well
as deviations in the proposed Ca2 binding
residues (Figure 1A). A.
B. In this work we
investigated the impact of free Ca2 and the
total number of disulfides on the reduction of
LNRs under various redox potentials through a
comparative study of multiple LNRs. Our results
indicate that while bound Ca2 provides
significant protection and prolonged chemical
stability against reduction under even strong
reducing conditions for the canonical LNRs, this
Ca2 based tunable stability is eliminated for
LNRs missing the first pair of disulfide bonds,
despite the presence of all the conserved Ca2
binding residues. Taken together with our earlier
findings we propose that LNRs are small protein
modules that have evolved to provide varying
amounts of redox sensitivity to the multi-domain
protein they are incorporated in through a
protein specific arrangement of multiple LNR
modules with subtle, yet critical sequence
variations.
Results
Below the fully saturating Ca2 concentration,
the rate of hN1LNRA reduction is redox potential
dependent
Ca2 protects the disulfide bonds against
reduction and prolongs the chemical stability of
folded hN1LNRA
Figure 1. (A) Sequence alignments of the first
LNRs from three different proteins Human Notch 1
(hN1LNRA), N-acetylglucosamine-1-phosphate
transferase (GS LNRA), and Human Notch 4
(hN4LNRA). Cysteine residues are in orange, and
the Ca2 coordinating D/N residues in yellow. The
characteristic disulfide bonding-pattern is
indicated above the sequence.
97.5 LNRA-Ca2 complex
99.8 LNRA-Ca2 complex
(B) NMR structure of hN1LNRA with disulfide bonds
shown as orange sticks and Ca2 coordinating
residues represented as yellow sticks.1
Figure 4. Percent of reduced hN1LNRA, complexed
with Ca2 to varying degrees, after 60 min. of
reaction with 250 µM (E-3386 mV) and 2.5 mM DTT
(E-3611 mV). At Ca2 concentrations below what
is needed to form essentially 100 LNRA-Ca2
complex, the amount of reduction at a given time
was directly proportional to the reducing power
of the environment. When hN1LNRA was fully
saturated with Ca2, however, it was fully
protected against reduction at both redox
potentials.
Figure 3. Reduction of folded hN1LNRA as a
function of time in varying Ca2 and 2.5 mM
DTT. Percent values of reduced hN1LNRA in 0 µM,
25 µM, 100 µM, 1 mM and 10 mM free Ca2 were
plotted as a function of time in 15 minute
intervals up to 1 hr.
Conclusions
2.5 mM DTT hN1LNRA Reduced hN1LNRA Reduced hN1LNRA Reduced hN1LNRA Reduced hN1LNRA Reduced
Ca2 0 min. 15 min. 30 min. 45 min. 60 min.
0 µM Ca2 (500 µM EDTA) 0 LNRA-Ca2 complex 3.16 85.33 92.15 91.57 88.5
25 µM Ca2 42.7 LNRA-Ca2 complex 0.58 12.71 32.71 38.73 49.64
100 µM Ca2 77.9 LNRA-Ca2 complex 0 4.93 13.46 14.99 19.9
1 mM Ca2 97.5 LNRA-Ca2 complex 0 0 0.43 1.55 1.87
10 mM Ca2 99.8 LNRA-Ca2 complex 0 0 0 0 0
Redox Potential (mV) in EDTA -373 -365 -364 -364 -367
Mean Redox Potential ( SD) (mV) in Ca2 -346 7 -341 8 -343 6 -341 7 -343 7

• Ca2 binding LNRs, like hN1LNRA and GS LNRA,
  are protected against reduction by free Ca2 in
  the environment. The rate of reduction is
  inversely proportional to the free Ca2 .
• The rate of reduction of an LNR is dependent on
  both the free Ca2 and the redox potential of
  the environment, unless the free Ca2 is fully
  saturating. At this point, there is complete
  protection against reduction for the range of
  tested redox potentials.
• Ca2 does not does not have an impact on the
  reduction of hN4 LNRA because hN4 LNRA does not
  bind Ca2 . However, even in the absence of any
  Ca2, hN4LNRA, with one fewer disulfide bond is
  still more susceptible to reduction at a given
  redox potential than its three disulfide-bonded
  homologs.

Materials and Methods Sample Preparation
Bacterially expressed hN1LNRA and synthesized GS
and hN4LNRA were folded in a refolding buffer
that allowed the formation of native disulfide
bonds and were purified via reverse phase High
Performance Liquid Chromatography (RP-HPLC).
Folded proteins were then dialyzed against 20 mM
HEPES, pH 8.0, 150 mM NaCl,  containing
predetermined concentrations of free Ca2 as
calculated by Visual Minteq.2 Reduction
Reactions Dialyzed proteins were aliquoted into
reaction tubes to have a final concentration of
15 µM and were placed in an AtmosBag
(Sigma-Aldrich) purged with N2. Dialysis buffers
supplemented with 250 µM or 2.5 mM DTT were
added under these anaerobic conditions to start
the reduction reaction. At predetermined time
points, reactions were quenched via acidification
and assayed via analytical RP-HPLC, using a C18
column running a 0.1 / min. acetonitrile
gradient. Data Analysis The areas under each
corresponding peak on the 280 nm chromatograms
were integrated to quantitate the amounts of
oxidized, reduced, and misfolded proteins (Figure
2). Percent oxidized, reduced, and misfolded
proteins were calculated by dividing the
integrated peak area of the corresponding peaks
by the total area of all protein peaks from
individual chromatograms (Table
1). Calculation of protein-Ca2
complex percentage The percent of protein
complexed with Ca2 at each free Ca2 was
calculated using the formula3 ML LoMoKD)
((LoMoKD)2 4MoLo)1/2 / 2, describing
reversible binding between a receptor and a
ligand (ML bound complex, Mo initial
protein 15 µM, Lo initial Ca2 of the
experiments included in the tables, and Kd
dissociation constant, previously determined to
be 25 µM). Calculation of Redox Potentials To
determine the amount of reduced and oxidized DTT
present at each experimental time point, a set of
experiments with known concentrations of oxidized
and total DTT were performed in the absence of
protein. Peaks on the chromatograms corresponding
to the oxidized (only at 280 nm) and reduced (at
229 and 280 nms) were integrated to determine the
corresponding absorbances. These values were
substituted into Beers Law (A elc) to
calculate the extinction coefficients of the
oxidized and reduced forms of DTT. These
extinction coefficients were used to determine
the actual concentration values at the assayed
time points of the reduction experiments. The
redox potentials for each experiment and time
point were calculated using the Nernst equation
Eh (in mV) Eo (RT/nF) ln(DTTred2/DTToxid)
Eo -323 mV at pH 7.0 with an adjustment of
-5.9 mM / 0.1 increase in pH.4
Discussion

• Ca2 is an integral part of the LNR structure
  possessing three disulfide bonds1. Fluorescence
  experiments have shown that binding of Ca2 to
  the LNR alters the surface exposure of residues
  around the binding pocket. Hence the Ca2
  dependent protection against reduction observed
  for these LNRs can be attributed to the increased
  chemical stability of the LNR-Ca2 complex
  compared to its apo form due to a decrease in the
  solvent accessibility of the disulfide bonds upon
  Ca2 binding.
• In comparison to the cytoplasm, which has a
  resting redox potential of approximately -230 mV
  that can be significantly altered by cellular
  status (-240, -200, and -170 mV during cell
  proliferation, differentiation, and apoptosis,
  respectively), the ER lumen provides a relatively
  constant and more oxidative environment (-180
  mV)4 critical for the proper folding of
  extracellular proteins. Unlike redox potential,
  though, the free Ca2 in the ER can change
  significantly based on cellular demands.6
• LNRs that are the focus of this study are found
  as repeated units within different multi-domain
  proteins targeted to the cell membrane. During
  folding in the ER, their disulfide bonds are
  formed and broken until the correct bonding
  pattern is achieved. The ability of free Ca2 in
  the environment to selectively fine tune the
  chemical stability of the LNRs by altering their
  redox sensitivity offers a novel mechanism for
  the cells to regulate this process and preserve
  any correctly folded regions of the protein over
  misfolded regions, which can continue to shuffle
  their disulfide bonds in search of the most
  stable conformation.

Table 2. Quantification of hN1LNRA reduction over
time in 20 mM HEPES, pH 8.0, 150 mM NaCl, 2.5 mM
DTT, and varying free Ca2. The percent of
reduced hN1LNRA and the redox potentials for each
experiment were determined as described in the
Materials and Methods section. Since the presence
of metal ions impacts the stability of DTT5, the
redox potentials calculated in the presence of
different Ca2 were averaged for each time
point, excluding the EDTA conditions, which are
reported separately. Based on the calculated
redox potentials, oxidized hN1LNRA experienced
very similar reducing environments (redox
potential -3436) in the presence of any amount
of Ca2 during the course of the experiment.
Under these redox conditions, essentially full
protection against reduction was achieved when
there was enough free Ca2 to ensure gt97.5
LNRA-Ca2 complex.
Figure 2. Overlay of chromatograms recorded at
280 nm at 0, 15, 30, 60, 180 min. for the series
of hN1LNRA reduction experiments. The four sets
of quantified peaks are annotated on the figure.
Folded hN1LNRA and GS LNRA are more protected
against reduction by Ca2 than hN4LNRA
Table 1. Integrated peak areas (IPA) of the
corresponding peaks on the chromatogram, in
addition to the calculated oxidized and reduced
hN1LNRA percentages and redox potentials for 250
µM DTT in 0 µM Ca2 (50 µM EDTA) at 0, 15, 30,
60, 180 min.

• References
• Vardar, D., North, C.L., Sanchez-Irizarry, C.,
  Aster, J. C., Blacklow, S. C. (2003) Nuclear
  Magnetic Resonance Structure of a Prototype
  Lin12-Notch Repeate Module from Human Notch1.
  Biochemistry, 42, 7061-7067.
• Visual MINTEQ http//www2.lwr.kth.se/English/OurS
  oftware/vminteq/
• Jakubowski. Chapter 5 Binding. A Reversible
  Binding 1 Equations and Curves. (2010)
  http//employees.csbsju.edu/hjakubowski/classes/ch
  331/bind/olbindderveq.html.
• Shafer, F.Q., Buettner, G.R. (2001) Redox
  environment of the cell as viewed through the
  redox state of the glutathione disulfide/glutathio
  ne couple. Free Radical Biology Medicine,
   30(11), 1191-1212.
• Burmeister Getz, E., Xiao, M., Chakrabarty, T.,
  Cooke, R., Slevin, P.R. (1999) A comparison
  between the sulfhydryl reductants
  tris(2-carboxyethyl)phosphine and dithiothreitol
  for use in protein biochemistry, Analytical
  Biochemistry, 273, 73-80.
• Bygrave, F.L, Benedetti, A. (1996) What is the
  concentration of calcium ions in the endoplasmic
  reticulum? Cell Calcium, 19 (6), 547-551.

250 µM DTT Reduced Reduced Reduced
Ca2 Time (min.) hN1LNRA GS LNRA hN4LNRA Redox Potential (mV)
  0 0 0.74 0 -333
0 µM Ca2 (50 µM EDTA) 15 5.16 2.63 9.11 -320
0 LNRA-Ca2 complex 30 13.19 15.99 25.61 -313
60 34.18 43.18 54.05 -305
  0 0 0 1.95 -324
10 mM Ca2 15 0 0 9.92 -304
99.8 LNRA-Ca2 complex 30 0.65 0 28.41 -309
60 1.02 0 54.72 -305
Time (min.) Oxidized DTT IPA Oxidized hN1 LNRA IPA Reduced hN1 LNRA IPA Mis- Folded PA Oxidized hN1LNRA Reduced hN1LNRA Calculated redox potential (mV)
0 N/A 409529 0 8333 98.01 0 N/A
15 17934 318971 18902 28092 87.16 5.16 -325
30 27861 250626 44043 39291 75.05 13.19 -319
60 48648 151864 96588 34173 53.73 34.18 -311
180 82116 24604 216413 68202 7.96 69.99 -303
Table 3. Quantification of the three homologous
LNRs over time in Ca2-free and Ca2-saturated
environments. The percent of reduced hN1LNRA, GS
LNRA, and hN4LNRA in 0 µM and 10 mM free Ca2 and
250 µM DTT at 0, 15, 30, and 60 minutes, as well
as the redox potentials calculated as described
in the Materials and Methods section. hN4LNRA,
with one fewer disulfide bond, has a faster rate
of reduction in comparison to hN1LNRA and GS
LNRA. The varying concentrations of free Ca2, in
fact, had no effect in the reduction of hN4LNRA,
while the saturation of Ca2 significantly
protected and prolonged the chemical stability of
hN1LNRA and GS LNRA against reduction.

• Funding
• Camille and Henry Dreyfus Faculty Start-up Award
  (DVU)
• Wellesley College Sophomore Early Research
  Program (JJ, AS)
• Protein Science Young Investigator Travel Grant
  (JJ)
• Wellesley College Science Center Travel Award
  (JJ, AS)