Orthogonal recognition in dimeric coiled coils via buried polar-group modulation

Article abstract of DOI:10.1021/ja076265w

We describe the design and exploration of new buried polar groups to control coiled-coil dimerization. Employing our recently described method for on-resin guanidinylation, we have prepared coiled-coil peptides with a single core guanidine, spaced from the backbone by 1-3 methylene groups. Heterodimeric mixtures of these sequences with guanidine, amide, and carboxylic acid binding partners form a large number of reasonably stable coiled coils (T_m ≥ 60°C). A detailed stability trend examination reveals that asparagine/acid pairs are sharply sensitive to acid residue chain length (Asn/Asp much worse than Asn/Glu), while guanidine/acid pairs are largely insensitive. This has been exploited to create orthogonal recognition pairs which establish the capacity to form two distinct heterodimeric coiled coils by simple mixing of four different peptides. One dimer has buried core asparagines, while the other pairs aspartic acid with any of three guanidinylated side chains. Specificity of this behavior is underscored by failure of glutamic acid substituted sequences to perform accordingly. The successful alternate pairs are further characterized by various biophysical methods (circular dichroism, ultracentrifugation, thermal and chemical denaturation, affinity tags).

Full text of DOI:10.1021/ja076265w

Published on Web 01/03/2008
Orthogonal Recognition in Dimeric Coiled Coils via Buried
Polar-Group Modulation
Maria L. Diss and Alan J. Kennan*
Department of Chemistry, Colorado State UniVersity, Fort Collins, Colorado 80523
Received August 20, 2007; E-mail: kennan@lamar.colostate.edu
Abstract: We describe the design and exploration of new buried polar groups to control coiled-coil
dimerization. Employing our recently described method for on-resin guanidinylation, we have prepared
coiled-coil peptides with a single core guanidine, spaced from the backbone by 1-3 methylene groups.
Heterodimeric mixtures of these sequences with guanidine, amide, and carboxylic acid binding partners
form a large number of reasonably stable coiled coils (Tm g 60 °C). A detailed stability trend examination
reveals that asparagine/acid pairs are sharply sensitive to acid residue chain length (Asn/Asp much worse
than Asn/Glu), while guanidine/acid pairs are largely insensitive. This has been exploited to create orthogonal
recognition pairs which establish the capacity to form two distinct heterodimeric coiled coils by simple mixing
of four different peptides. One dimer has buried core asparagines, while the other pairs aspartic acid with
any of three guanidinylated side chains. Specificity of this behavior is underscored by failure of glutamic
acid substituted sequences to perform accordingly. The successful alternate pairs are further characterized
by various biophysical methods (circular dichroism, ultracentrifugation, thermal and chemical denaturation,
affinity tags).
Introduction
Precise control and understanding of macromolecular structure
underlies numerous active research areas, from nanostructure
synthesis to deciphering disease pathologies. Self-assembly
events dominate many such investigations, placing a premium
on development and discovery of molecular recognition motifs.
Numerous bio-inspired synthetic structures have been devel-
oped.1 Identification of new routes to specific recognition
expands design capabilities, consequently enhancing odds of
success in any given application. Of particular interest is the
ability to program simultaneous recognition events which occur
without leakage to undesired complexes. Such systems dramati-
cally increase the number of reasonable supramolecular targets.
We have recently been interested in self-assembling peptide
systems derived from R-helical coiled coils, whose ubiquitous
biological roles and substantial catalog of structure-function
studies make them logical candidates for further design work.2
These complexes, composed of two or more supercoiled helical
strands, exhibit a regular sequence repeat in seven residue blocks
(labeled abcdefg). Assembly is driven by formation of a
hydrophobic interface juxtaposing a/d position side chains,
augmented by more typically electrostatic contacts between e/g
residues (see Figure 1 for an example).
Figure 1. Peptides used. Helical wheel projection (above left) demonstrates
interactions in heterodimers. Buried polar residues are represented by X or
X′ (see position 14). Sequences of each peptide are given below, with
nonstandard amino acids represented by the following code (structures of
each given above right): Z ) guanidinylated diaminopropionic acid, Z* )
guanidinylated diaminobutyric acid. Underlined lysine side chains are capped
with acetamidobenzoyl groups as spectroscopic labels.
(1) (a) Goodman, C. M.; Choi, S.; Shandler, S.; DeGrado, W. F. Nat. Chem.
Biol. 2007, 3, 252-262. (b) Hill, D. J.; Mio, M. J.; Prince, R. B.; Hughes,
T. S.; Moore, J. S. Chem. ReV. 2001, 101, 3893-4011. (c) Gellman, S. H.
Acc. Chem. Res. 1998, 31, 173-180.
(2) Coiled-coil reviews: (a) Woolfson, D. N. AdV. Protein Chem. 2005, 70,
79-112. (b) Parry, D. A. D.; Squire, J. M. AdV. Protein Chem. 2005, 70,
1-10. (c) Lupas, A. N.; Gruber, M. AdV. Protein Chem. 2005, 70, 37-78.
(d) Cheng, R. P. Curr. Opin. Struct. Biol. 2004, 14, 512-520. (e) DeGrado,
W. F.; Summa, C. M.; Pavone, V.; Nastri, F.; Lombardi, A. Annu. ReV.
Biochem. 1999, 68, 779-819.
Efforts to control and manipulate coiled coil formation have
targeted oligomerization state, relative strand orientation, and
specificity for homo- or heteromeric structures.3 Design of e/g
interfaces has largely focused on maximizing potentially favor-
9
10.1021/ja076265w CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 1321-1327
1321

A R T I C L E S
Diss and Kennan
able glutamic acid/lysine (Glu/Lys) interactions, while avoiding
presumably repulsive Glu/Glu or Lys/Lys pairings. Strategies
for a/d control include alignment of sterically matched side
chains or buried polar groups. This array of fundamental design
work has facilitated applications in protein misfolding models,4
materials chemistry,5 and biotechnology.6
We have previously developed steric matching of hydropho-
bic core side chains as a route to specific trimer formation.7
Seeking to establish similar levels of control in dimeric systems,
we have begun to explore new buried polar interactions. The
most common use of buried polar contacts in dimeric coiled
coils is to include a single core asparagine (Asn), which has
been shown to impart structural uniqueness with respect to both
oligomerization state and strand orientation, although other polar
contacts have also been developed.8 Despite unquestionable
utility, the Asn strategy is necessarily limited to a single
recognition event. Considerably more self-assembly problems
could be addressed if two independent parallel dimers could be
generated by simply mixing four different peptides. This
demands new core recognition pairs, which can direct specific
and stable dimer formation while ignoring the presence of
alternative complexes.
In the course of other investigations, we have developed a
convenient synthetic methodology for preparing guanidinylated
amino acid residues during solid-phase peptide synthesis.9 Here
we report use of a slight modification to this method to prepare
several peptides with a single core guanidinylated side chain
of varying length. A detailed evaluation of heterodimeric
assemblies with guanidine/guanidine, guanidine/amide, or guani-
dine/carboxylic acid buried polar contacts reveals a large number
of reasonably stable coiled coil dimers (Tm g 60 °C). Most
significantly, three new combinations are demonstrated to
function in the presence of an Asn/Asn driven dimerization.
Results and Discussion
As a starting sequence for investigating core interactions we
focused on the Acid-p1/Base-p1 heterodimer designed by Kim
and co-workers.10 Each strand of our parent dimer thus contains
a leucine (Leu) core with a single Asn at position 14, and either
Glu (pAsnE) or Lys (pAsnK) at each e/g location (Figure 1).11
The other sequences differ in polar core residue identity,
featuring guanidine or carboxylic acid side chains of varying
length. In addition to arginine (Arg, three methylenes in the
side chain), we employed guanidinylated diaminopropionic acid
(Dap*, one methylene) and diaminobutyric acid (Dab*, two
methylenes). The acidic (pDap*E, pDab*E, pArgE) and basic
(pDap*K, pDab*K, pArgK) versions of each sequence differ only
in e/g substitution. We also investigated acidic peptides bearing
aspartic acid (Asp) or Glu at the polar core position (pAspE,
pGluE).
The sequences containing only natural side-chain structures
were prepared via standard solid-phase methods using com-
mercial amino acids. The requisite chain shortened Arg ana-
logues were prepared using a modification of our previously
reported on-resin guanidinylation method.12
With the peptides in hand, we began by investigating
heterodimers bearing identical guanidine side chains in place
of the parent asparagines (Figure 2). Although we anticipated
that burial of like charges might substantially destabilize
potential complexes, circular dichroism (CD) spectra of 1:1
pDab*K/pDab*E, and pArgK/pArgE mixtures exhibit dramatic
helicity increases compared to the average component signals
(Figure 2C/E).13 Thermal unfolding experiments reveal coopera-
tive transitions only for the mixtures, which also have reasonably
(3) (a) Tsurkan, M. V.; Ogawa, M. Y. Inorg. Chem. 2007, 46, 6849-6851.
(b) Steinmetz, M. O.; Jelesarov, I.; Matousek, W. M.; Honnappa, S.; Jahnke,
W.; Missimer, J. H.; Frank, S.; Alexandrescu, A. T.; Kammerer, R. A.
Proc. Nat. Acad. Sci. U.S.A. 2007, 104, 7062-7067. (c) Shang, J.; Geva,
E. J. Phys. Chem. B 2007, 111, 4178-4188. (d) Ryan, S. J.; Kennan, A.
J. J. Am. Chem. Soc. 2007, 129, 10255-10260. (e) Portwich, M.; Keller,
S.; Strauss, H. M.; Mahrenholz, C. C.; Kretzschmar, I.; Kramer, A.;
Volkmer, R. Angew. Chem., Int. Ed. 2007, 46, 1654-1657. (f) Nikolaev,
Y.; Pervushin, K. J. Am. Chem. Soc. 2007, 129, 6461-6469. (g) Yoder,
N. C.; Kumar, K. J. Am. Chem. Soc. 2006, 128, 188-191. (h) Son, S.;
Tanrikulu, C.; Tirrell, D. A. ChemBioChem 2006, 7, 1251-1257. (i) Meier,
M.; Burkhard, P. J. Struct. Biol. 2006, 155, 116-129. (j) Mason, J. M.;
Schmitz, M. A.; Mueller, K. M.; Arndt, K. M. Proc. Nat. Acad. Sci. U.S.A.
2006, 103, 8989-8994. (k) Liu, J.; Zheng, Q.; Deng, Y.; Cheng, C.-S.;
Kallenbach, N. R.; Lu, M. Proc. Nat. Acad. Sci. U.S.A. 2006, 103, 15457-
15462. (l) Lee, K.-H.; Cabello, C.; Hemmingsen, L.; Marsh, E. N. G.;
Pecoraro, V. L. Angew. Chem., Int. Ed. 2006, 45, 2864-2868. (m) Hadley,
E. B.; Gellman, S. H. J. Am. Chem. Soc. 2006, 128, 16444-16445. (n)
Dong, H.; Hartgerink, J. D. Biomacromolecules 2006, 7, 691-695. (o)
Bunagan, M. R.; Cristian, L.; DeGrado, W. F.; Gai, F. Biochemistry 2006,
45, 10981-10986. (p) Bjelic, S.; Karshikoff, A.; Jelesarov, I. Biochemistry
2006, 45, 8931-8939. (q) Balakrishnan, G.; Hu, Y.; Case, M. A.; Spiro,
T. G. J. Phys. Chem. B 2006, 110, 19877-19883. (r) Acharya, A.; Rishi,
V.; Vinson, C. Biochemistry 2006, 45, 11324-11332. (s) Yadav, M. K.;
Redman, J. E.; Leman, L. J.; Alvarez-Gutierrez, J. M.; Zhang, Y.; Stout,
C. D.; Ghadiri, M. R. Biochemistry 2005, 44, 9723-9732. (t) Sakurai, Y.;
Mizuno, T.; Hiroaki, H.; Gohda, K.; Oku, J.-i.; Tanaka, T. Angew. Chem.,
Int. Ed. 2005, 44, 6180-6183. (u) Plecs, J. J.; Harbury, P. B.; Kim, P. S.;
Alber, T. J. Mol. Biol. 2004, 342, 289-297. (v) Marti, D. N.; Bosshard,
H. R. Biochemistry 2004, 43, 12436-12447. (w) Kwok, S. C.; Hodges, R.
S. J. Biol. Chem. 2004, 279, 21576-21588. (x) Gurnon, D. G.; Whitaker,
J. A.; Oakley, M. G. J. Am. Chem. Soc. 2003, 125, 7518-7519. (y)
McClain, D. L.; Gurnon, D. G.; Oakley, M. G. J. Mol. Biol. 2002, 324,
257-270. (z) Vu, C.; Robblee, J.; Werner, K. M.; Fairman, R. Protein
Sci. 2001, 10, 631-637.
(4) (a) Dong, H.; Hartgerink, J. D. Biomacromolecules 2007, 8, 617-623. (b)
Yadav, M. K.; Leman, L. J.; Price, D. J.; Brooks, C. L., III; Stout, C. D.;
Ghadiri, M. R. Biochemistry 2006, 45, 4463-4473. (c) Pagel, K.; Wagner,
S. C.; Samedov, K.; Von Berlepsch, H.; Boettcher, C.; Koksch, B. J. Am.
Chem. Soc. 2006, 128,2196-2197. (d) Ambroggio, X. I.; Kuhlman, B. J.
Am. Chem. Soc. 2006, 128,1154-1161 (e) Cerasoli, E.; Sharpe, B. K.;
Woolfson, D. N. J. Am. Chem. Soc. 2005, 127, 15008-15009.
(5) (a) Papapostolou, D.; Smith, A. M.; Atkins, E. D. T.; Oliver, S. J.; Ryadnov,
M. G.; Serpell, L. C.; Woolfson, D. N. Proc. Nat. Acad. Sci. U.S.A. 2007,
104,10853-10858. (b) Zimenkov, Y.; Dublin, S. N.; Ni, R.; Tu, R. S.;
Breedveld, V.; Apkarian, R. P.; Conticello, V. P. J. Am. Chem. Soc. 2006,
128,6770-6771. (c) Shen, W.; Zhang, K.; Kornfield, J. A.; Tirrell, D. A.
Nat. Mater. 2006, 5,153-158. (d) Raman, S.; Machaidze, G.; Lustig, A.;
Aebi, U.; Burkhard, P. Nanomedicine 2006, 2,95-102. (e) Kovaric, B. C.;
Kokona, B.; Schwab, A. D.; Twomey, M. A.; De Paula, J. C.; Fairman, R.
J. Am. Chem. Soc. 2006, 128,4166-4167. (f) Farmer, R. S.; Argust, L.
M.; Sharp, J. D.; Kiick, K. L. Macromolecules 2006, 39, 162-170.
(6) (a) Mason, J. M.; Mueller, K. M.; Arndt, K. M. Biochemistry 2007, 46,
4804-4814. (b) Leman, L. J.; Weinberger, D. A.; Huang, Z.-Z.; Wilcoxen,
K. M.; Ghadiri, M. R. J. Am. Chem. Soc. 2007, 129, 2959-2966. (c)
Yuzawa, S.; Mizuno, T.; Tanaka, T. Chem. Eur. J. 2006, 12, 7345-7352.
(d) Woolley, G. A.; Jaikaran, A. S. I.; Berezovski, M.; Calarco, J. P.; Krylov,
S. N.; Smart, O. S.; Kumita, J. R. Biochemistry 2006, 45,6075-6084. (e)
Zhang, K.; Diehl, M. R.; Tirrell, D. A. J. Am. Chem. Soc. 2005, 127,-
10136-10137. (f) Magliery, T. J.; Wilson, C. G. M.; Pan, W.; Mishler,
D.; Ghosh, I.; Hamilton, A. D.; Regan, L. J. Am. Chem. Soc. 2005, 127,-
146-157. (g) Zhou, M.; Ghosh, I. Org. Lett. 2004, 6, 3561-3564.
(7) (a) Schnarr, N. A.; Kennan, A. J. Org. Lett. 2005, 7, 395-398. (b) Schnarr,
N. A.; Kennan, A. J. J. Am. Chem. Soc. 2004, 126, 14447-14451.
(8) (a) Straussman, R.; Ben-Ya’acov, A.; Woolfson, D. N.; Ravid, S. J. Mol.
Biol. 2007, 366, 1232-1242. (b) Lear, J. D.; Gratkowski, H.; Adamian,
L.; Liang, J.; DeGrado, W. F. Biochemistry 2003, 42, 6400-6407. (c)
Schneider, J. P.; Kretsinger, J.; J. Am. Chem. Soc. 2003, 125, 7907. (d)
Akey, D. L.; Malashkevich, V. N.; Kim, P. S. Biochemistry 2001, 40, 6352-
6360. (e) Oakley, M. G.; Kim, P. S. Biochemistry 1998, 37, 12603-12610.
(f) Zeng, X.; Herndon, A. M.; Hu, J. C. Proc. Natl. Acad. Sci. U.S.A. 1997,
94, 3673-3678. (g) Schneider, J. P.; Lear, J. D.; DeGrado, W. F. J. Am.
Chem. Soc. 1997, 119, 5742. (h) Gonzalez, L., Jr.; Woolfson, D. N.; Alber,
T. Nat. Struct. Biol. 1996, 3, 1011-1018. (i) Lumb, K. J.; Kim, P. S.
Biochemistry 1995, 34, 8642-8.
(9) Zhang, Y.; Kennan, A. J. Org. Lett. 2001, 3, 2341-2344. See Supporting
Information for details.
(10) O’Shea, E. K.; Lumb, K.; Kim, P. S. Curr. Biol. 1993, 3, 658-667.
(11) These two peptides are not technically the same as Acid-p1/Base-p1, only
because they differ in the chosen spectroscopic label (tryptophan versus
acetamidobenzoyl). To avoid confusion they are given names consistent
with the nomenclature scheme of the other sequences.
(12) See Supporting Information for details.
(13) See Supporting Information for similar plots involving all other heterodimers
examined in this work.
9
1322 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008

Orthogonal Recognition in Dimeric Coiled Coils
A R T I C L E S
Figure 3. Mixed guanidine-guanidine interactions. Wavelength (A) and
thermal denaturation (B) CD spectra for the indicated equimolar mixtures.17
Figure 2. Buried guanidine-guanidine interactions. Wavelength (A, C,
E) and thermal denaturation (B, D, F) CD spectra for pure solutions of
pXaaK (red) and pXaaE (blue), where Xaa ) Dap* (A, B), Dab* (C, D),
and Arg (E, F). In each case, traces for an equimolar mixture (green) and
the calculated weighted-average signal (open circles) are also given.17
high stability (Figure 2D/F, Tm ) 59 °C in each case), albeit
diminished from the parent pAsnE/pAsnK complex (Tm ) 77 °C).
Both observations point to a specific and significant interaction
between the component peptides, consistent with heterodimer
formation. The pDapK/pDapE mixture displays a more subtle
but similar effect, supporting an overall trend of increased
heterodimer stability with increases in side-chain length (Figure
2A/B).
The observation of an Arg/Arg contact that is only somewhat
destabilizing compared to an Asn/Asn interaction is consistent
with some contact potential estimates derived from statistical
analyses of protein structures.14 One such method predicts
contact energies of -1.68 and -1.55 kcal/mol for Asn/Asn and
Arg/Arg, respectively.14b Both are significantly less stabilizing
than Leu/Leu, Ile/Ile, or Val/Val contacts (-7.37, -6.54, -5.52
kcal/mol, respectively), consistent with increased Tm values for
coiled coils with purely hydrophobic cores (which nonetheless
typically form nonspecific complexes).
Having observed reasonable stability in complexes which pair
identical guanidine side chains, we deemed it worthwhile to
examine the analogous mixed-core systems. Spectra of equimo-
lar pDap*K/pDab*E, pDap*K/pArgE, and pDab*K/pArgE mixtures
also exhibit reasonable room-temperature helicities and coopera-
tive thermal unfolding profiles (Figure 3, Table 1). A subtle
but noticeable stability trend again favors those complexes with
more total methylenes in the buried polar side chains.
Figure 4. Core-swapped mixed guanidine-guanidine interactions. Wave-
length (A) and thermal denaturation (B) CD spectra for the indicated
equimolar mixtures.17
Table 1. Observed Tm Values for Guanidine Core Pairs
sample
Tm
(°
C)
sample
Tm (°C)
pDap*K/pDap*E
pDab*K/pDab*E
pArgK/pArgE
pDap*K/pDab*E
pDap*E/pDab*K
49
59
59
43
49
pDap*K/pArgE
pDab*K/pArgE
pDap*E/pArgK
pDab*E/pArgK
47
55
47
51
Mixed-core complexes introduce another possible complica-
tion, since exchanging the polar core residues on the acidic and
basic strands produces a distinct complex (e.g., pDap*K/pDab*E
is not the same as pDap*E/pDab*K). Although we expected any
differences to be minimal, we nonetheless examined the reverse
mixed-core mixtures pDap*E/pDab*K, pDap*E/pArgK, and
pDab*E/pArgK (Figure 4, Table 1). Somewhat surprisingly, a
modest compression in stability range was observed, such that
the two mixed pDap*E complexes are very similar to each other
and to the original pDap*K/pDap*E mixture. Even the pDab*E/
pArgK signal is closer to the others than its counterpart is.
Though observable, the differences were small enough that we
hesitate to apply detailed structural rationales.
Having discovered surprisingly stable complexes with elec-
trostatically repulsive guanidine/guanidine pairings, we next
examined the interaction of each guanidinylated side chain with
the parent neutral Asn residue, expecting a further increase in
(14) (a) Gromiha, M. M.; Selvaraj, S. Prog. Biophys. Mol. Biol. 2004, 86, 235-
277. (b) Miyazawa, S.; Jernigan, R. L. J. Mol. Biol. 1996, 256, 623-644.
(c) For a discussion of potential difficulties with such analyses see: Thomas,
P. D.; Dill, K. A. J. Mol. Biol. 1996, 257, 457-489.
9
J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008 1323

A R T I C L E S
Diss and Kennan
Figure 6. Guanidine/Asp core interactions. Wavelength (A) and thermal
denaturation (B) CD spectra for equimolar mixtures of pAspE with pDap*K
(blue), pDab*K (red), and pArgK (green).17
Figure 5. Guanidine/Asn core interactions. Wavelength (A) and thermal
denaturation (B) CD spectra for the indicated equimolar mixtures.17 Closed
and open circles of the same color indicate data from complexes that differ
by exchange of the indicated buried polar side chains (from the acidic or
basic peptide to the other).
Table 2. Observed Tm Values for Guanidine Amide/acid Core
Pairs
Figure 7. Guanidine/Glu core interactions. Wavelength (A) and thermal
denaturation (B) CD spectra for equimolar mixtures of pGluE with pDap*K
(red), pDab*K (blue), and pArgK (green).17
sample
Tm
(°
C)
sample
Tm (°C)
pDap*K/pAsnE
pDab*K/pAsnE
pArgK/pAsnE
pDap*E/pAsnK
pDab*E/pAsnK
pArgE/pAsnK
73
71
71
63
63
65
pDap*K/pAspE
pDab*K/pAspE
pArgK/pAspE
pDap*K/pGluE
pDab*K/pGluE
pArgK/pGluE
61
67
65
59
63
63
overall stability. Examination of wavelength and thermal
unfolding CD data support this expectation (Figure 5, Table
2). Equimolar mixtures of pDap*K, pDab*K, and pArgK with
pAsnE exhibited cooperative unfolding transitions and substantial
melting temperatures (Tm ) 71-73 °C). To determine the
impact of local strand interactions, the reverse complexes,
formed by mixing pDap*E, pDab*E, and pArgE with pAsnK,
were also investigated (Figure 5, Table 2). As with the mixed
guanidine complexes, a small but definite stability distinction
exists, favoring those with guanidinylated side chains on the
basic peptide (e.g., Tm ) 71 °C for pArgK/pAsnE, versus 65 °C
for pArgE/pAsnK). The difference is particularly pronounced in
the room-temperature helicity of the two Dap* complexes,
although the gap narrows at the melting transition. Overall, a
reduced stability variance is observed within each set, meaning
that length of the guanidine side chain seems largely irrelevant
when paired opposite Asn, as opposed to another guanidine.
The final set of interactions examined mix each guanidine
side chain with Asp or Glu residues, resulting in an overall
neutral complex which may benefit from core electrostatic
attraction. As with other the other core pairings, equimolar
mixtures of pDap*K, pDab*K, and pArgK with either pAspE or
pGluE exhibit good helicities and thermal stabilities. (Figure 6-7,
Table 2). The range of Tm values (59 to 67 °C) is comparable
to those of the guanidine/amide pairs, albeit slightly lower. The
guanidine/acid pairs also display a similar insensitivity to polar
residue chain length, with only a small change between the
pDap*K/pAspE and pArgK/pGluE complexes.
Figure 8. Melting curve comparisons. Thermal unfolding data for equimolar
solutions of pDap*K (A), pDab*K (B), and pArgK (C) with each of the
indicated acidic peptides replotted for comparison.17
guanidine side chain paired with each of its partners reveals
some interesting features (Figure 8). In each case, solutions with
pAsnE as the partner core residue exhibit the most stable profiles,
followed by the pAspE and pGluE complexes, and finally the
various guanidine/guanidine mixtures. The degree of separation
between these classes varies quite significantly, with sharp
distinctions drawn for interactions with pDap*K, and a gradual
With a now substantial array of feasible heterodimers
available, we began to explore prospects for new orthogonal
recognition pairs. A comparison of melting curves for each
9
1324 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008

Orthogonal Recognition in Dimeric Coiled Coils
A R T I C L E S
Figure 10. Ni-NTA analysis of Arg/Arg, Dap*/Dap* and Dab*/Dab*
complexes. Equimolar mixtures of pArgK/pArgE/pDap*K/pDap*E His (left)
and pArgK/pArgE/pDab*K/pDab*E His (right) give rise to supernatant and
elution fractions containing all reasonable complexes. Solutions are as in
Figure 9.
Figure 9. Ni-NTA analysis of pArgE/pArgK pDab*K Asp/Glu selective
dimer formation. Equimolar initial mixtures of pArgK/pArgE/pDab*K/
pAspE His (left) and pArgK/pArgE/pDab*K/pGluE His (right) give rise to
supernatant and elution fractions containing all reasonable complexes. Initial
solutions are 20 µM total peptide, pH 7.0, 150 mM NaCl, 10 mM phosphate.
blurring of those distinctions in moving to the longer side chains.
At the extreme, the trace for pArgK/pArgE is not that different
from those of the pAspE and pGluE mixtures (Figure 8C).
The reasonable stability of the Arg/Arg core pair led us to
examine it as an initial candidate for an interaction which could
be preserved in the presence of another heterodimer. Similarly,
the complexes with Dab*/Asp and Dab*/Glu pairs demonstrated
reasonable melting temperatures and were chosen as the alternate
partners. Thus it was hoped that equimolar mixtures of pArgK/
pArgE/pDab*K/pAspE and pArgK/pArgE/pDab*K/pGluE would
each afford two separate dimers, rather than a mixture of the
four energetically reasonable assemblies (discounting electro-
statically incompatible binding of two acidic or basic peptides).
To assay for dimer specificity, we employed a Ni-NTA
affinity method we have previously used extensively to measure
coiled-coil stoichiometries.7,12 An N-terminal GlyGly(His)6 tag
confers affinity for Ni-NTA agarose beads upon both the tagged
peptide and any species which binds to it. After exposure to
beads, removal of the supernatant, washing to eliminate low-
affinity peptides, and elution with imidazole buffer to recover
bound material, each fraction is analyzed by HPLC. In the
present case, if one of the four peptides is tagged and orthogonal
recognition is operative, one expects the tagged peptide and its
partner to dominate the elution fraction, while the two matched
untagged peptides reside in the supernatant.
For the initial two mixtures, tagged derivatives of the Asp
(pAspE His) and Glu (pGluE His) peptides were synthesized using
standard methods. Unfortunately, analysis of each mixture
demonstrates essentially nonspecific recognition (Figure 9). Both
elution fractions contain each of the basic peptides, suggesting
that pArgK/pAspE and pArgK/pGluE complexes form in addition
to the desired ones. The supernatants also contain both basic
species, consistent with formation of the unintended pDab*K/
pArgE complex in each case.
Following this lack of success using two new core-core
matchups, we next investigated whether guanidine/guanidine
peptides could recognize each other in the context of the known
Asn/Asn interaction, which would require identifying only one
new viable core-core interaction. Thus equimolar solutions of
pAsnK/pAsnE/pDap*K/pDap*E, pAsnK/pAsnE/pDab*K/pDab*E,
and pAsnK/pAsnE/pArgK/pArgE were analyzed by Ni-NTA
affinity (the basic Asn peptide was His tagged in each case).
Once again, essentially no specificity was observed, as both
supernatant and elution fractions contained significant quantities
of each acidic peptide in all three experiments.12
Figure 11. Asn/acid interactions. Wavelength (A) and thermal denaturation
(B) CD spectra for equimolar mixtures of pAsnK/pAsnE (green), pAsnK/
pGluE (blue), and pAsnK/pAspE (red).17
Table 3. Observed Tm Values for Asn/acid Core Pairs
sample
Tm
(°
C)
sample
Tm (°C)
pAsnK/pAspE
pAsnK/pGluE
51
65
pAsnK/pAsnE
77
predict likely candidates. By this measure, our initial efforts
seemed less promising. Using melting temperatures as a crude
measure of stability, we recognized that the reasonable pArgK/
pArgE interaction could be balanced by the comparable pDab*K/
pArgE (or pDab*K/pGluE) affinity. Similarly, the very stable
pAsnK/pAsnE contact (Tm ) 77 °C) is nearly balanced by the
uniformly high guanidine/amide Tm values (more than balanced
for the relatively weak pDab*K/pDab*E interaction).
With this view in mind, we next targeted all-guanidine
systems. Mixtures of pDap*K with pDap*E or pArgE exhibited
similar stabilities (Tm ) 49 and 47 °C, respectively), while
pArgK demonstrated a preference for pArgE over pDap*E (Tm
) 59 vs 47 °C). Similar arguments apply to pDab*K, with
slightly more favorable numbers (Tm values of 59 vs 51 °C and
59 vs 55 °C). Despite this determination, analysis of equimolar
pArgK/pArgE/pDap*K/pDap*E, and pArgK/pArgE/pDab*K/pDab*E
mixtures (using pArgK His) again failed to support high levels
of specificity (Figure 10). However, some relative preference
was observed (indicated by slightly unequal ratios of the acidic
peptides in supernatant and elution fractions). We therefore
sought to extend this approach to even more unbalanced pairs.
Using the reasoning above, selective recognition seems most
likely when at least one of the unintended complexes is
destabilized. Thus to exploit the stable Asn/Asn contact we
sought another pair in which at least one partner is mismatched
with Asn. Although the guanidine/Asn contacts were largely
favorable, we wondered if the same would apply to Asp or Glu
matched with Asn. Accordingly, we measured wavelength and
thermal unfolding CD profiles for equimolar solutions of pAsnK/
pAspE and pAsnK/pGluE, and compared them with the parent
pAsnK/pAsnE complex (Figure 11, Table 3).
In light of these results we began to consider that selection
of two reasonable complexes was not a sufficient criterion to
ensure specific recognition. In particular, we examined stabilities
of the alternative complexes as an additional consideration to
Encouragingly, Asp appears to be relatively mismatched with
Asn (Tm ) 51 °C). In contrast, the interaction with Glu seems
9
J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008 1325

A R T I C L E S
Diss and Kennan
Figure 13. Ni-NTA analysis of four component Asn/Asn and Gdn/Glu
mixtures. Equimolar ratios of (A) pAsnKHis/pAsnE/pDap*K/pGluE, (B)
pAsnK His/pAsnE/pDab*K/pGluE, and (C) pAsnK/pAsnE/pArgK/pGluEHis all
demonstrate nonspecific recognition. Elution fractions and supernatants are
consistent with formation of all possible dimers in significant quantities.
Solutions are as in Figure 9.
Figure 12. Ni-NTA analysis of four component Asn/Asn and Gdn/Asp
mixtures. Equimolar ratios of (A) pAsnK His/pAsnE/pDap*K/pAspE, (B) pAsn
His/pAsnE/pDab*K/pAspE, and (C) pAsnK His/pAsnE/pArgK/pAspE all demK-
onstrate specific recognition. Elution fractions are dominated by the Asn
peptides, indicating preferential binding of pAspE. Solutions are as in
Figure 9.
relatively strong (Tm ) 65 °C). Thus we predicted that
guanidine/Asp peptides should be able to specifically recognize
each other in the presence of the Asn/Asn contact, while the
corresponding guanidine/Glu complexes would be nonspecific.
To examine both of these possibilities, equimolar solutions of
pAsnK/pAsnE with pDap*K/pAspE, pDab*K/pAspE, pArgK/
pAspE, pDap*K/pGluE, pDab*K/pGluE, and pArgK/pGluE were
investigated by Ni-NTA methods (using pAsnK His in all but
the latter case, which used pGluE His).
Figure 14. Ni-NTA analysis of Gdn/Asp heterodimers. In all three cases,
one equivalent of the guanidinylated peptide is retained in the elution
fraction, starting from equimolar mixtures. Solutions are as in Figure 9,
except that initial concentrations are 10 µM total peptide.
Analysis of the Asp complexes reveals that indeed these pairs
are capable of orthogonal recognition in the presence of an Asn/
Asn interaction (Figure 12). In each case, the elution fraction
is dominated by pAsnK His and pAsnE, while the supernatant
contains largely the guanidine/Asp peptides. To our knowledge
this is the first example of simultaneous self-selection of
independent heterodimeric coiled coils driven only by a single-
residue buried polar-residue substitution.
Further supporting the exquisite specificity of this interaction,
the corresponding guanidine/Glu complexes, differing only by
the addition of a single methylene unit in the core, do not display
correspondingly specific binding (Figure 13). In each case, as
with the early experiments, significant quantities of each possible
complex is indicated by the presence of additional peptides in
both supernatant and elution fractions.
The apparent success of guanidine/Asp complexes in specific
orthogonal recognition led us to further characterize each
heterodimer. Support for dimer formation and stoichiometry was
obtained by Ni-NTA experiments. When equimolar mixtures
of pDap*K, pDab*K, and pArgK with a His-tagged pAspE
derivative are exposed to Ni-NTA agarose beads, the elution
fractions in each case contain one equivalent of each peptide
(Figure 14).
Figure 15. GdnHCl (A) and urea (B) denaturation data for equimolar
mixtures of pAspE with pDap*K (red), pDab*K (blue), and pArgK (green).
Solutions are as in Figure 9.
pDab*K/pAspE (Mr calcd ) 7214, Mr obs ) 7589), and pArg*K/
pAspE (Mr calcd ) 7228, Mr obs ) 7634) all gave apparent
molecular weights within six percent of the calculated ones.15
Finally, we subjected each guanidine/Asp complex to chemi-
cal denaturation experiments (Figure 15). Although use of
GdnHCl gave transitions with limited cooperativity, we specu-
lated that ionic interactions could be important for such highly
charged species.16 Accordingly, the experiments were repeated
using urea unfolding, and the expected cooperative curves were
observed. In addition to characterizing the Gdn/Asp dimers, we
Sedimentation equilibrium analytical ultracentrifugation re-
sults are also consistent with dimer formation. Equimolar
solutions of pDap*K/pAspE (Mr calcd ) 7200, Mr obs ) 7543),
(15) Mr obs values for pDap*K and pDab*K were obtained by assigning the
unnatural amino acid a partial specific volume equal to the average value
of all proteinogenic side chains. Substitution with either maximal or minimal
values had little effect. See Supporting Information for details.
9
1326 J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008

Orthogonal Recognition in Dimeric Coiled Coils
A R T I C L E S
conducted Ni-NTA, centrifugation, and unfolding experiments
on many of the other heterodimers reported here.12
Dap* residue exhibiting much more selective binding prefer-
ences than arginine in the same environment. Despite the
overall relative tolerance of varying core residue chain length
and charge states, only a select subset of these new heterodimers
were capable of recognition based only on polar core residue
identity. The outcomes of these experiments suggest that spec-
ific formation of two independent heterodimers from initial four-
peptide mixtures requires that at least one of the alternative
complexes must be mismatched. Thus poor stability of Asn/
Asp cores is capable of directing impressive recognition
levels. The lessons learned here provide a scaffold for
future design of even more complicated and selective assemblies,
which will expand further the already extensive appli-
cations of coiled coils in biotechnology and materials chem-
istry.
To further quantify buried core interactions we considered a
double-mutant cycle analysis, but selection of a usable reference
state proved problematic. The standard Ala-Ala contact is
unsuitable, as that substitution in the required core position
dramatically alters oligimerization preferences in similar systems
(a failing likely exhibited by other hydrophobic options).8i Polar
groups capable of specifying oligomerization state would in turn
presumably exhibit measurable new interactions in the single-
mutant complexes.
Conclusions
The work described above establishes a surprisingly broad
array of reasonably stable coiled-coil heterodimers with an
assortment of buried polar interactions. Although dimers with
neutral core pairings are more stable, the gap was smaller than
expected, particularly for cores with two buried guanidines.
Intriguing chain-length effects are observed, with the short-chain
Acknowledgment. This work was supported by an NSF
CAREER award to A.J.K. (Grant CHE-0239275).
Supporting Information Available: Detailed experimental
procedures, descriptions of guanidinylation chemistry, additional
CD, Ni-NTA, and analytical ultracentrifugation data. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(16) (a) Schellman, J. A. Quart. ReV. Biophys. 2005, 38,351-361. (b) Schellman,
J. A. Biophys. Chem. 2002, 96, 91-101. (c) Monera, O. D.; Kay, C. M.;
Hodges, R. S. Protein Sci. 1994, 3, 1984-91. (d) Tanford, C. AdV. Protein
Chem. 1970, 24, 1-95. (e) Robinson, D. R.; Jencks, W. P. J. Am. Chem.
Soc. 1965, 87, 2462-70.
(17) All wavelength spectra were recorded at 25°C, and solutions for all CD
experiments contained 10 µM total peptide in PBS buffer (150 mM NaCl,
10 mM phosphate, pH 7.0).
JA076265W
9
J. AM. CHEM. SOC. VOL. 130, NO. 4, 2008 1327