Retention of the cis proline conformation in tripeptide fragments of bovine pancreatic ribonuclease A containing a non-natural proline analogue, 5,5-dimethylproline

Article abstract of DOI:10.1021/ja9930317

Attention is focused on L-5,5-dimethylproline (dmP) as a substitute to lock L-proline (Pro) in a cis conformation in peptides and proteins, to prevent cis/trans isomerization when a protein with cis X-Pro peptide groups unfolds. Procedures have been devel

Full text of DOI:10.1021/ja9930317

11558
J. Am. Chem. Soc. 1999, 121, 11558-11566
Retention of the Cis Proline Conformation in Tripeptide Fragments of
Bovine Pancreatic Ribonuclease A Containing a Non-natural Proline
Analogue, 5,5-Dimethylproline
Seong Soo A. An,† Cathy C. Lester,† Jin-Lin Peng,† Yue-Jin Li,† David M. Rothwarf,†
Ervin Welker,† Theodore W. Thannhauser,† L. S. Zhang,‡ James P. Tam,‡ and
Harold A. Scheraga*,†
Contribution from the Baker Laboratory of Chemistry and Chemical Biology, Cornell UniVersity,
Ithaca, New York 14853-1301, and Department of Microbiology and Immunology, Vanderbilt UniVersity,
NashVille, Tennessee 37232
ReceiVed August 20, 1999. ReVised Manuscript ReceiVed October 27, 1999
Abstract: Attention is focused on L-5,5-dimethylproline (dmP) as a substitute to lock L-proline (Pro) in a cis
conformation in peptides and proteins, to prevent cis/trans isomerization when a protein with cis X-Pro peptide
groups unfolds. Procedures have been developed to obtain optically pure L-dmP and to incorporate this sterically
hindered residue as the central one in tripeptides that are suitable for fragment coupling to prepare synthetic
proteins. Based on the sequences of residues 92-94 (Tyr-Pro-Asn:YPN) and 113-115 (Asn-Pro-Tyr: NPY)
in bovine pancreatic ribonuclease A (RNase A), in which the X-Pro peptide groups are in the cis conformation,
the tripeptides Ac-Tyr-dmP-Asn (YdmPN) and Ac-Asn-dmP-Tyr (NdmPY) were synthesized, and their structures
1
were determined by 2D H nuclear magnetic resonance (NMR) spectroscopy. YdmPN was found to exist
solely in the cis conformation between 6 and 60 °C, whereas NdmPY was found to have some trans component
that increased from about 10% to about 21% as the temperature increased over the range between 6 and 80
°C. Both YdmPN and cis-NdmPY adopt a type VI reverse turn, as does proline. The NMR structures of
YdmPN and cis-NdmPY are comparable with the X-ray structures of the corresponding portions of RNase A,
and the NMR structure of trans-NdmPY is compatible with the X-ray structure of the isolated tripeptide,
Ac-NPY. These results demonstrate that L-dmP is a promising substitute for proline in a variety of protein
problems to constrain the X-Pro peptide group to the cis conformation.
Introduction
have been hampered by the heterogeneity of the X-Pro peptide
group.8-10 Formation of cis peptide groups at non-proline
residues is also biologically significant. A cis Gly-Gly peptide
bond has been suggested to play a role in amyloid formation,11
but investigations have been somewhat limited by the inability
to generate peptides containing a high percentage of the cis
peptide group. The ability to restrict a peptide group exclusively
to the cis conformation, in addition to facilitating progress in
the areas discussed above, would also provide an invaluable
tool for use in developing peptides for drug use and for the
generation of peptide antibodies targeted against the native
isomeric state of a protein. The formation of a cis peptide group
results in a chain reversal,3,12 a structural feature that may be
of great importance in dictating the formation of early folding
intermediates3,13 and thereby directing productive protein folding
pathways.3,13-18 Indeed, it is the role of proline in influencing
Proline is the only naturally occurring cyclic amino acid.
Consequently, unlike other peptide groups, which predominantly
adopt the trans form, the imidic X-Pro peptide group readily
exists in the cis as well as the trans form, even in the absence
of structural constraints. In the X-Pro peptide group, the cis
and trans states are of similar energies, with about a 20 kcal/
mol barrier for their interconversion.1 In unfolded proteins and
short peptides, the X-Pro peptide group exists as a mixture of
the two isomers;2 in native proteins, it almost always adopts
only one of these single isomeric states, dictated by other
interactions within the protein.
Because of its unique conformational properties, proline plays
a critical role in many biologically important processes, includ-
ing protein folding3,4 and signal transduction,5,6 and in biologi-
cally active and pharmaceutically relevant peptides.7 Peptide
and protein studies of the role of proline in structure formation
(8) Baldwin, R. L. Annu. ReV. Biochem. 1975, 44, 453.
(9) Escribano, J.; Grubb, A.; Calero, M.; Mendez, E. J. Biol. Chem. 1991,
266, 15758.
(10) Truckses, D. M.; Somoza, J. R.; Prehoda, K. E.; Miller, S. C.;
Markley, J. L. Protein Sci. 1996, 5, 1907.
* To whom correspondence should be addressed.
† Cornell University.
‡ Vanderbilt University.
(1) Zimmerman, S. S.; Scheraga, H. A. Macromolecules 1976, 9, 408.
(2) Grathwohl, C.; Wu¨thrich, K. Biopolymers 1981, 20, 2623.
(3) Lewis, P. N.; Momany, F. A.; Scheraga, H. A. Proc. Natl. Acad.
Sci. U.S.A. 1971, 68, 2293
(4) Veeraraghavan, S.; Nall, B. T. Biochemistry 1994, 33, 687.
(5) Cohen, G. B.; Ren, R.; Baltimore, D. Cell 1995, 80, 237.
(6) Soman, K. V.; Hanks, B. A.; Tien, H.; Chari, M. V.; O’Neal, K. D.;
Morrisett, J. D. Protein Sci. 1997, 6, 999.
(11) Weinreb, P. H.; Jarrett, J. T.; Lansbury, P. T., Jr. J. Am. Chem.
Soc. 1994, 116, 10835.
(12) Venkatachalam, C. M. Biopolymers 1968, 6, 1425.
(13) Matheson, R. R., Jr.; Scheraga, H. A. Macromolecules 1978, 11,
819.
(14) Jackson, S. E.; Fersht, A. R. Biochemistry 1991, 30, 10436.
(15) Houry, W. A.; Rothwarf, D. M.; Scheraga, H. A. Nat. Struct. Biol.
1995, 2, 495.
(7) Yaron, A.; Naider, F. Crit. ReV. Biol. Mol. Biol. 1993, 28, 31.
(16) Jager, M.; Pluckthun, A. FEBS. Lett. 1997, 418, 106.
10.1021/ja9930317 CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/04/1999

RNase A Fragments with Dimethylproline
J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11559
protein folding pathways that arguably may be its most
important role.
predominantly in the cis conformation in model peptides.23,30,31
Unfortunately, oxazolidine and thiazolidine compounds are acid
labile and, therefore, are not easily incorporated into pep-
tides.26,32,33 The proline analogues constrain the X-Pro peptide
group by the addition of bulky functional groups. In addition,
the presence of the heteroatom alters the ring puckering (the
dihedral angles, φ and ø) and hydrophobicity of the residue as
compared to proline, which may result in significant alterations
in the conformational properties of peptides containing oxazo-
lidine and thiazolidine proline analogues.31,34
In protein folding, numerous examples exist in which the rate-
determining step of in vitro folding is the isomerization of an
X-Pro peptide group.15,19-21 The most important purpose of
folding studies is not the characterization of the proline
isomerization processes but rather the determination of how the
sequence directs the folding process. Unfortunately, the het-
erogeneity of the X-Pro peptide groups in the unfolded state
obscures the conformational folding process and prevents
meaningful interpretation. In some cases, it has been possible
to eliminate the isomerization problem through experiments in
which the protein is rapidly unfolded, and refolding is initiated
before the X-Pro peptide groups have had time to isomerize
from their native state. This procedure has been used success-
fully to study the disulfide-intact refolding of bovine pancreatic
ribonuclease A (RNase A),22 but it severely limits the experi-
mental conditions and observation methods that can be used
and requires specialized equipment. In the case of oxidative
protein folding (beginning from the fully denatured and disulfide
reduced state), refolding rates are on a time scale similar to
that of proline isomerization, and double-jump stopped-flow
techniques are ineffective. In that situation, it has been impos-
sible to eliminate complications arising from the cis/trans
isomerization of proline peptide groups.
Clearly, the ability to constrain a peptide group to a single
unique confirmation is of considerable utility. A number of
attempts have been made to constrain the X-Pro peptide group
to a single conformation.23-31 Proline analogues such as
methanoproline24,25 and oxazolidine-4-carboxylic acid31 are very
effective at yielding high percentages of the trans conformer
while still retaining the imidic peptide bond. The most common
method of retaining a trans peptide bond, however, is simply
to replace proline with one of the other naturally occurring
amino acids, typically alanine. No such routine methods exist
for the incorporation of cis peptide groups. Among the more
successful approaches that have been used to constrain peptide
groups to the cis conformation is the use of substituted
thiazolidines and oxazolidines.23,30,31 One such analogue, 2,2-
dimethylthiazoladine-4-carboxylic acid, has been shown to exist
The pyrrolidine compound most similar to 2,2-dimethylthi-
azolidine-4-carboxylic acid is 5,5-dimethylproline (dmP).26,35
A dmP-containing dipeptide, Boc-Phe-dmP-OMe, has been
reported, and nuclear magnetic resonance (NMR) data indicated
that the peptide group was predominantly in the cis conformation
in methanol.26 Unfortunately, the results of that experiment were
not quantitative because efforts to separate the enantiomers of
dmP were unsuccessful. To ascertain the suitability of dmP for
use in peptides and proteins, we have investigated the properties
of optically pure L-dmP in aqueous solution, and we report here
the efficient separation of L-dmP from D-dmP. Since one of
our major purposes is to incorporate dmP into synthetic RNase
A and peptides derived from that protein, we have synthesized
two tripeptides, Ac-Tyr-dmP-Asn (acetylated tyrosine-5,5-
dimethylproline-asparagine, YdmPN) and Ac-Asn-dmP-Tyr
(acetylated asparagine-5,5-dimethylproline-tyrosine, NdmPY),
which correspond to residues 92-94 and 113-115, respectively,
of RNase A. In native RNase A, both of these peptides contain
cis X-Pro peptide groups.36 We also report here the solution
structure of YdmPN, cis-NdmPY, and trans-NdmpY, determined
1
by H NMR spectroscopy.
Experimental Section
Synthesis of 5,5-Dimethylproline (dmP). D,L-dmP was synthesized
at the Cornell University Biotechnology Resource Center by the method
described by Magaard et al.26
Resolution of L-dmP from D-isomer: Determination of Optical
Purity. As far as we are aware, optically pure L-dmP has not been
prepared previously. A rapid and efficient method for separating the
optical isomers of D,L-dmP was developed here, based on a method
that proved suitable to resolve the enantiomers of D,L-Pro.37 This method
is based on the differential solubility of the complexes formed between
the optical isomers of dmP and D-tartaric acid. The D,L-dmP (1.0 g/7.0
mmol, equivalent to 3.5 mmol of the L form) was dissolved in 2.0 mL
of water containing 1.05 g/7.0 mmol of D-tartaric acid at room
temperature. Then, 70 mL of absolute ethanol was added slowly with
mechanical stirring. After 24 h, a crystalline precipitate was collected
by filtration, washed with ethanol, and air-dried; yield of the L-dmP:
D-tartaric acid complex, 0.95 g (91% yield, 3.2 mmol). The dmP was
separated from the D-tartaric acid by anion-exchange chromatography
(Sigma IRA-410, Amberlite) and converted to the methyl ester using
SOCl2/MeOH.38
(17) Aronsson, G.; Brorsson, A. C.; Sahlman, L.; Jonsson, B. H. FEBS.
Lett. 1997, 411, 359.
(18) Rousseau, F.; Schymkowitz, J. W.; Sanchez, del Pino M.; Itzhaki,
L. S. J. Mol. Biol. 1998, 284, 503.
(19) Brandts, J. F.; Halvorson, H. R.; Brennan, M. Biochemistry 1975,
14, 4953.
(20) Texter, F. L.; Spencer, D. B.; Rosenstein, R.; Matthews, C. R.
Biochemistry 1992, 31, 5687.
(21) Dutta, S.; Maity, N. R.; Bhattacharyya, D. Biochim. Biophys. Acta
1997, 1343, 251.
(22) Houry, W. A.; Rothwarf, D. M.; Scheraga, H. A. Biochemistry 1994,
33, 2516.
(23) Sˇavrda, J. In Peptides 1976, Proceedings of the 14th European
Symposium, Brussels; Loffet, A., Ed.; 1976; p 653.
(24) Montelione, G. T.; Arnold, E.; Meinwald, Y. C.; Stimson, E. R.;
Denton, J. B.; Huang, S.-G.; Clardy, J.; Scheraga, H. A. J. Am. Chem. Soc.
1984, 106, 7946.
(25) Montelione, G. T.; Hughes, P.; Clardy, J.; Scheraga, H. A. J. Am.
Chem. Soc. 1986, 108, 6765.
The absolute configuration of the isomer of dmP forming the
insoluble crystalline complex with d-tartaric acid was identified as L
1
by H NMR spectroscopy of the methyl ester of the dmP isomer in
(R)-(-)-1-phenyl-2,2,2-trifluoroethanol.39 The purity of the L-dmP
(26) Magaard, V. W.; Sanchez, R. M.; Bean, J. W.; Moore, M. L.
Tetrahedron Lett. 1993, 34, 381.
(27) Zerkout, S.; Dupont, V.; Aubry, A.; Vidal, J.; Collet, A.; Vicherat,
A.; Marraud, M. Int. J. Pept. Protein Res. 1994, 44, 378.
(28) Mikhailov, D.; Daragan, V. A.; Mayo, K. H. Biophys. J. 1995, 68,
1540.
(29) Chalmers, D. K.; Marshall, G. R. J. Am. Chem. Soc. 1995, 117,
5927.
(30) Dumy, P.; Keller, M.; Ryan, D. E.; Rohwedder, B.; Wo¨hr, T.;
Mutter, M. J. Am. Chem. Soc. 1997, 119, 918.
(31) Keller, M.; Sager, C. S.; Dumy, P.; Schutkowski, M.; Fisher, G.
S.; Mutter, M. J. Am. Chem. Soc. 1998, 120, 2714.
(32) Sheehan, J. C.; Yang, D.-D. H. J. Am. Chem. Soc. 1958, 80, 1158.
(33) Wo¨hr, T.; Wahl, F.; Nefzi, A.; Rohwedder, B.; Sato, T.; Sun, X.;
Mutter, M. J. Am. Chem. Soc. 1996, 118, 9218.
(34) Frau, J.; Donoso, J.; Mun˜oz, F.; Blanco, F. G. HelV. Chim. Acta
1994, 77, 1557.
(35) Bonnett, R.; Clark, V. M.; Giddey, A.; Todd, S. A. J. Chem. Soc.
1959, 2087.
(36) Wlodawer, A.; Svensson, L. A.; Sjo¨lin, L.; Gilliland, G. L.
Biochemistry 1988, 27, 2705.
(37) Yamada, S.; Hongo, C.; Chibata, I. Agric. Biol. Chem. 1977, 41,
2413.
(38) Brenner, M.; Huber, W. HelV. Chim. Acta 1953, 36, 1109.

11560 J. Am. Chem. Soc., Vol. 121, No. 49, 1999
An et al.
(>98%; mp 193.5 °C) was determined by using 2,3,4,6-tetra-O-benzyl-
â-D-glucopyranosyl isothiocyanate (BGIT), which can form diastere-
omeric thioureas with D,L-dmP without racemization.40 The diastere-
omers can be separated on a nonchiral reversed-phase (RP) C-18 column
(Vydac C-18, 4.6 × 100) by high-pressure liquid chromatography
(HPLC), using as a mobile phase 10 mM phosphate, pH 7, 60-70%
MeOH gradient. The specific rotation of the optically pure L-dmP ([R]D
) -51.2°) was determined in water using a Perkin-Elmer model 241
polarimeter. Concentrations were determined by weight using desiccated
material.
obtained for each of 432 t1 increments. Chemical shifts were measured
and referenced relative to tetramethylsilane (TMS) by assigning the
1
HOD resonance a value of 4.80 ppm. 2D H NMR experiments were
carried out in the phase-sensitive mode using time-proportional phase
incrementation.43 Suppression of the water signal was accomplished
by presaturation during the initial delay between transients. Data were
processed with the VNMR data processing program or with the Felix
(Molecular Simulations Inc.) program. The data matrix of 432 t1
experiments was zero-filled to 2K points along both dimensions; prior
to Fourier transformation, it was multiplied by a 0°-shifted sine-bell
curve along both t1 and t2 for the double-quantum-filtered shift-
correlated spectroscopy (DQF-COSY)44 experiment and a 75°-shifted
squared sine-bell curve along both t1 and t2 for the nuclear Overhauser
effect spectroscopy (NOESY),45 rotating frame nuclear Overhauser
effect spectroscopy (ROESY),46 and total correlated spectroscopy
(TOCSY)47,48 (τmix ) 75 ms) experiments. NOESY spectra were
acquired with mixing times of 150, 250, and 450 ms at 25 °C and with
a mixing time of 250 ms at 10 and 15 °C. ROESY experiments were
carried out with mixing times of 250 ms at 20 and 25 °C.
Determination of Isomerization Rate Constants. As will be shown
in the Results and Discussion section, YdmPN is solely in the cis
conformation, whereas NdmPY is an equilibrium mixture of cis and
trans forms. Assuming that cis/trans interconversion is a two-state
process, the rate constants of isomerization were estimated by following
the temperature-jump procedures reported by Grathwohl and Wu¨thrich.2,49
In these experiments, the kinetics of the isomerization process were
followed by measuring the relative intensities of the corresponding 1H
resonances of the cis/trans conformers as a function of time after
initiating a jump in temperature. For these experiments, the initial
temperature was set to 80 °C by heating the NdmPY peptide sample
before insertion into the magnet. The final temperatures in the magnet
were set to 10, 15, 20, 25, 30, and 35 °C. After an initial delay of 20
s, spectra were recorded at 6.8-s intervals until thermal equilibrium at
the final temperature was attained. The cis/trans isomerization rate
constants (kcft and ktfc; Figure 3A, below) were then calculated from
the ratio of the peak areas for the cis and trans species.2,50,51 For
obtaining kcft, the following equation was used,
Synthesis of Tripeptides (Fragments of RNase A). The tripeptides
Ac-YdmPN and Ac-NdmPY were synthesized by a combination of
solid- and solution-phase methods at the Cornell University BioRe-
source Center by the following methods.
Ac-Y(OBut)-dmP-OH and Ac-N(Trt)-dmP-OH (Trt ) 1,1,1-
Triphenylmethyl). The bulky methyl groups on the δ carbon of dmP
hampered formation of the X-dmP peptide bond. In one procedure,
Fmoc-Y(OBut)-dmP-OMe (Fmoc ) (fluorenylmethoxy)carbonyl) was
prepared by coupling Fmoc-Y(OBut)-OH (138 mg, 0.3 mmol) and dmP-
OMe (47 mg, 0.3 mmol) using N-dicyclohexylcarbodiimide (DCC, 62
mg, 0.3 mmol) in dry DMF. The Fmoc and the methyl ester were
removed by treatment with a minimal excess of 1 M NaOH/methanol
and then acetylated with acetic anhydride in a one-pot reaction. The
product, Ac-Y(OBut)dmP-OH (46 mg, 0.114 mmol, 38% yield), was
purified by RP-HPLC on a Nova-Pak C18 (8 × 25 cm) radial
compression cartridge employing a Waters model 600E HPLC. Ac-
N(Trt)-dmP-OH was prepared in a similar fashion with 32% yield (38
mg, 0.0704 mmol) from Fmoc-N(Trt)-OH (131 mg, 0.22 mmol) and
dmP-OMe (35 mg, 0.22 mmol) using DCC (46 mg, 0.22 mmol) in dry
DMF. In a second procedure, Boc and Fmoc protecting groups and a
variety of coupling reagents were tried, to form the Tyr-dmP peptide
bond.
Ac-Y-dmP-N-OH and Ac-N-dmP-Y-OH. Ac-Y(OBut)dmP-OH
(13.5 mg, 0.033 mmol) was activated with 2-(1H-benzotriazol-1-yl)-
1,1,3,3-tetramethyluronium tetrafluoroborate (TBTU)/N-hydroxyben-
zotriazole (HOBt) and coupled to an equimolar amount of NH2-N(Trt)-
Novasyn resin (500 mg, 0.1 mM equivalent NH2, Nova Biochem). The
resulting tripeptide resin was cleaved and deprotected with TFA/EDT/
thioanisol (95/2.5/2.5, v/v) (EDT ) 1,2-ethanedithiol). The Ac-Y-dmP-
N-OH was then purified by RP-HPLC as described above with a yield
of 55% (8.5 mg, 0.018 mmol). Ac-N-dmP-Y-OH was prepared by
coupling Ac-N(Trt)-dmP-OH (12.4 mg, 0.023 mmol) to an NH2-
Y(OBut)-Novasyn resin in a similar manner. The cleavage of the
protecting groups and purification of Ac-N-dmP-Y-OH were ac-
complished as described above with a yield of 58% (6 mg, 0.013 mmol).
The two tripeptides were characterized by analytical HPLC, capillary
electrophoresis (CE), and matrix-assisted laser desorption/chemical
ionization mass spectrometry (MALDI MS). Each peptide was found
to be >98% pure. The (M + H)+ ions of these peptides were 464.8
((0.1%) and 464.5 ((0.1%) Da, respectively (expected 464.4 Da).
In the second procedure for synthesizing YdmPN, higher yields
(75%) were obtained by using a Boc protecting group and the bis(2-
oxo-3-oxazolidinyl)phosphinic chloride (BOP-Cl) coupling reagent,41
or an Fmoc protecting group and the tetramethylfluoroformamidinium
hexafluorophosphate coupling reagent42 (TFFH, generously provided
by PerSeptive Inc.).
[cis]t - [cis]∞
[cis]0 - [cis]∞
-kcft
) exp
t
(1)
(
)
[trans]∞
where [cis]0, [cis]t, [cis]∞, and [trans]∞ are the concentrations of cis or
trans at times 0, t, and equilibrium, respectively. A corresponding
equation was used to obtain ktfc. Cis and trans resonances of the methyl
protons on the δ carbon of dmP were selected for the quantitative
measurement because they were well separated from each other and
from other resonances. The activation energy, Eq, for the cis f trans
interconversion was computed from the Arrhenius equation,
ln kcft ) ln A - Eq/RT
(2)
To obtain the steady-state ratio between the cis and trans populations
of the two peptides at a given temperature, the 1D spectra of YdmPN
and NdmPY were recorded at 6 °C, in 5-deg steps from 10 to 60 °C,
and at 80 °C. Steady-state equilibrium constants were calculated from
Kcft ) [trans]/[cis], and a van’t Hoff plot was used to obtain the
thermodynamic parameters (∆G°, ∆H°, and ∆S°; Figure 3B, below).
At 25 °C, for example, K ) 0.143, or the % trans is 13% for NdmPY.
Since only one set of resonances was observed for YdmPN throughout
the temperature range in this study, it was concluded that no
isomerization of the YdmPN group had occurred.
NMR Measurements. For the NMR experiments, 2-3 mg of
peptide was added to 240 µL of 90%/10% H2O/D2O solution in Shigemi
NMR tubes (BMS-005V). D2O (99.9%) from Cambridge Isotope was
used. The pH was adjusted with aliquots of 0.5 M NaOH and checked
with a Beckmann Futura Plus electrode without correction for the
presence of D2O, which was added for the deuterium lock. The pH for
all experiments was 5.3 ( 0.1 at 20 °C.
(43) Marion, D.; Wu¨thrich, K. Biochem. Biophys. Res. Commun. 1983,
113, 967.
(44) Rance, M.; Sørensen, O. W.; Bodenhausen, G.; Wagner, G.; Ernst,
R. R.; Wu¨thrich, K. Biochem. Biophys. Res. Commun. 1983, 117, 479.
(45) Bodenhausen, G.; Kogler, H.; Ernst, R. R. J. Magn. Reson. 1984,
58, 370.
1
All H NMR experiments were carried out on a Varian Unity 500-
MHz spectrometer equipped with a triple-resonance probe and pulsed-
field z-gradients. The spectral width was 4199.7 Hz. For each 2D NMR
experiment, 32 transients of 4096 complex points per transient were
(46) Bothner-By, A. A.; Stephens, R. L.; Lee, J.-M.; Warren, C. D.;
Jeanlos, R. W. J. Am. Chem. Soc. 1984, 106, 811.
(47) Braunschweiler, L.; Ernst, R. R. J. Magn. Reson. 1983, 53, 521.
(48) Bax, A.; Davis, D. G. J. Magn. Reson. 1985, 65, 355.
(49) Akasaka, K.; Naito, A.; Nakatani, H. J. Biomol. NMR 1991, 1, 65.
(50) Gutowsky, H. S.; Holm, C. H. J. Chem. Phys. 1956, 25, 1228.
(51) Neuman, R. C., Jr; Jonas, V. J. Am. Chem. Soc. 1968, 90, 1970.
(39) Pirkle, W. H.; Beare, S. D. J. Am. Chem. Soc. 1969, 91, 5150.
(40) Lobell, M.; Schneider, M. P. J. Chromatogr. 1993, 633, 287.
(41) Tung, R. D.; Rich, D. H. J. Am. Chem. Soc. 1985, 107, 4342.
(42) Carpino, L. A.; El-Faham, A. J. Am. Chem. Soc. 1995, 117, 5401.

RNase A Fragments with Dimethylproline
J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11561
Table 1. J Coupling Constants in Hertz (A) and Cis/Trans Population Ratios at Various Temperatures in H2O (B)
(A) J Coupling Constants
JHN-HR
JHR-Hâ1
JHR-Hâ2
JHR-Hâ1,2
JHN-HR
JHR-Hâ1
JHR-Hâ2
8.7
dmP
6.8
YdmPN
Tyr
10.4
Asn
8.9
dmP
8.4, 2.0
dmP
Asn
4.7
Tyr
7.3
Tyr
7.0
4.3
6.4
8.3
5.4
4.4
7.0
8.2
8.72
7.3
7.0
80
cis-NdmPY
9.2, 1.9
dmP
7.25
7.36
trans-NdmPY
Asn
7.4
9.3, 2.0
(B) Cis/Trans Population Ratios
temp (°C)
% cis-NdmPY
6
90.3
10
89.6
15
88.9
20
88.5
25
30
86.8
35
86.2
40
85.1
45
84.3
50
84.1
55
82.5
60
80.7
87.5
78.9
Dihedral Angle and NOE Constraints. Scalar spin-spin coupling
constants (J; Table 1) were measured from the 1H-1H spin-spin
splittings in the 1D 1H spectra. To minimize the overlap of peaks, the
amplitude of the free induction decay (FID) was multiplied by a shifted
Gaussian window function before the Fourier transformation.
From the analysis of the NOEs from both NOESY and ROESY
spectra, similar relative NOE connectivity information was obtained.
The intensities of the NOESY cross-peaks from a 250-ms experiment
at 10 and 25 °C were measured. Distances were estimated (within upper
and lower bounds of (0.5 Å) by using the isolated spin-pair
approximation, with reference to the geminal proton cross-peak of Asn
corresponding to an interproton distance of 1.8 Å.52 This was justified
in view of the small size of the two peptides, which should involve
negligible spin diffusion.
angle ω was allowed to vary in the calculation, but only within the
neighborhood of 0° and 180° for the cis and trans conformations,
respectively.
The conformations of the three peptides were then determined by
incorporating the NOE distance and dihedral angle constraints into a
Powell energy minimization (EM) and simulated annealing (SA)
protocol.56-59 The SA method was based on Verlet restrained molecular
dynamics. The EM/SA protocol can be divided into five stages. First,
500 steps of EM were applied with a weight for the van der Waals
repulsion force constant set at 0.002 and a slope for the asymptotic
constant for the NOEs set at 0.1. Under these conditions, the atoms
were allowed to move through each other, providing greater freedom
to accommodate the initial NOE constraints. Next, conformational
sampling of structures was carried out by applying 10 ps of molecular
dynamics at 1000 K, again with the weight for the repulsion force
constant set at 0.002 and the slope for the asymptotic constant for the
NOEs set at 0.1. Distance penalties were evaluated by using a soft-
well square NOE potential. In the third stage, the structures were
subjected to 10 ps of molecular dynamics at 1000 K with the weight
for the repulsion force constant increased to 0.1 and the slope for the
asymptotic constant for the NOE increased to 1.0. Under these
conditions, the peptides were given less freedom to sample the
conformational space. For the fourth stage, the asymptotic constant for
the NOEs remained at 1.0, and the weight for the repulsion force
constant was increased to 1.0. The temperature was gradually lowered
from 1000 to 100 K in 1000 steps (1 ps/step) of molecular dynamics.
In the final stage, the square-well potential was replaced with a soft-
well square potential and 1000 steps of EM, and another 1000 steps
with inclusion of electrostatic interactions were applied. From the 100
initial random structures of each of the three peptides, 57, 48, and 24
structures of YdmPN, cNdmPY, and tNdmPY, respectively, were
selected on the basis of no violations of NOE or dihedral angle
constraints greater than 0.1 Å and 5°, respectively. An average structure
was then calculated for each peptide, and the 10 structures with the
smallest root-mean-square deviations (RMSDs) from the average
structure were finally chosen.
Stereospecific assignments of the Hâ resonances of Asn and Tyr in
the cis or trans conformers and their dihedral angles, ø1, were
3
determined from the JHRHâ coupling constants together with the
intraresidue NOE pattern53 from HR to Hâ1,2 and HN to Hâ1,2. For the
3
3
dihedral angle ø1 of 180°, the JHRHâ1 and JHRHâ2 coupling constants
are >10 and <5 Hz, respectively. The HR-â2 and HN-Hâ1 NOEs are
stronger than their corresponding HR-â1 and HN-Hâ2 NOEs, respec-
3
3
tively. For the dihedral angle ø1 of - 60°, the JHRHâ1 and JHRHâ2
coupling constants are <5 and >10 Hz, respectively. The HR-â1 and
HN-Hâ2 NOEs are stronger than their corresponding HR-â2 and HN-
Hâ1 NOEs, respectively.
Calculation of the Structures of YdmPN, cNdmPY, and tNdmPY.
The X-PLOR 3.1 program was used, with the IBM SP2 supercomputer
of the Cornell Theory Center, to calculate the structures of the
peptides.54 The resulting structures were viewed with the molecular
graphics program InsightII (Molecular Simulations Inc.).
Since the dmP residue was a new amino acid, its coordinates, partial
charges, and nonbonded parameters were incorporated into the X-PLOR
topology and parameter files, respectively, for calculating the structures
of YdmPN, cNdmPY, and tNdmPY. The nonbonded parameters, bond
lengths, and partial charges for the two methyl groups attached to the
δ carbon were taken from the methyl groups of the valine residue in
the X-PLOR protein topology and parameter files. The bond angles
and dihedral angles of dmP were taken as those of the Pro residue in
the down-puckering conformation, the down-puckering having been
observed in this NMR investigation.
Results and Discussion
NMR analyses of dmP, YdmPN, and NdmPY. The chemi-
cal shift assignments of L-dmP and the two tripeptides were
determined by 1D, DQ-COSY, NOESY, ROESY, and TOCSY
experiments using the methods of Wu¨thrich.52 The chemical
shifts for the free L-dmP residue and the two tripeptides are
listed in Table 2. For the free L-dmP, the Hâ resonances were
resolved and were stereospecifically assigned according to the
intensities of their NOEs with the HR proton. Both Hγ1 and Hγ2
resonances, as well as those of the two methyl groups, were
The default force constants for bond lengths, bond angles, and
dihedral angles of all amino acid residues were the same as reported
previously.55 The NOE penalty function was set to 50 kcal mol-1 Å-2
.
Initial sets of 100 random structures each were generated for the
YdmPN, cNdmPY, and tNdmPY peptides starting from random values
for the dihedral angles φ, ψ, and ø, and a value of 180° for the dihedral
angle ω, except for dmP in the cis conformation (ω ) 0°). The dihedral
(52) Wu¨thrich, K. NMR of Proteins and Nucleic Acids; John Wiley and
Sons: New York, 1986.
(56) Powell, M. J. D. Math. Programming 1977, 12, 241.
(53) Gu¨ntert, P.; Braun, W.; Billeter, M.; Wu¨thrich, K. J. Am. Chem.
Soc. 1989, 111, 3997.
(54) Brunger, A. T. X-PLOR Version 3.0 User Manual; Yale University
Press: New Haven, CT, 1991.
(57) Nilges, M.; Clore, G. M.; Gronenborn, A. M. FEBS Lett. 1988, 229,
317.
(58) Nilges, M.; Clore, G. M.; Gronenborn, A. M. FEBS Lett. 1988, 239,
129.
(55) Maurer, M. C.; Peng, J.-L.; An, S. S. A.; Trosset, J. Y.; Henschen-
Edman, A.; Scheraga, H. A. Biochemistry 1998, 37, 5888.
(59) Nilges, M.; Clore, G. M.; Gronenborn, A. M. Biopolymers 1990,
29, 813.

11562 J. Am. Chem. Soc., Vol. 121, No. 49, 1999
An et al.
Table 2. 1H Resonance Assignmentsa for L-dmP (A), Ac-YdmPN (B), and cis-Ac-NdmPYb (C) in Aqueous Solution
residue
HNc
H(Me)d
HR
Hâ1,2
otherse
(A) L-dmP
4.18
dmP
8.73, 7.56
2.46, 2.17
Hγ1,2, 1.92; Hꢀ1,2, 1.44
(B) Ac-YdmPN
Ac
1.94
Tyr
dmP
Asn
8.01
-
8.26
4.44
3.56
4.36
2.84, 2.94
1.66, 1.84
2.81, 2.66
H2,6, 7.17; H3,5, 6.87
Hγ1,2, 1.62; Hꢀ1, 1.27; Hꢀ2, 1.47
Hδ, 7.53, 6.78
(C) cis-Ac-NdmPY
Ac
1.96
(1.96)
Asn
dmP
Tyr
8.29
(8.17)
-
4.32
(5.14)
4.88
(4.45)
4.33
(4.37)
2.40, 2.46
Hδ, 7.35, 6.72
(2.71, 2.60)
2.28, 1.93
(2.16, 1.82)
3.11, 2.96
(2.96)
(Hδ, 7.54, 6.81)
Hγ1,2, 1.69, 1.50; Hꢀ1, 1.26; Hꢀ2, 1.36
(Hγ1,2, 1.73; Hꢀ1, 1.31; Hꢀ2, 1.45)
H2,6, 7.08; H3,5, 6.78
7.66
(7.15)
(H2,6, 7.07; H3,5, 6.80)
a Chemical shift (ppm) at pH 5.3. b Values in parentheses were observed for the trans-Ac-NdmPY. c There are two values for dmP because the
free N-terminal group is protonated. d Me is the CH3 of the acetyl groups. e Hꢀ corresponds to the protons of two methyl groups on the δ-carbon
of dmP. Hδ of Asn corresponds to the side-chain amide protons. H2,6 and H3,5 of Tyr correspond to the phenolic ring protons.
Figure 1. Regions of 2D NOESY spectrum of YdmPN: (A) NH-HR,Hâ and (B) HR-HR,Hâ regions. Experimental conditions for both spectra: 8
mM YdmPN in 90/10% H2O/D2O, pH 5.3, 500 MHz, 32 transients, delay time of 1.0 s, 432 FIDs, τmix ) 250 ms, 25 °C. All spectra were referenced
against HOD, 4.80 ppm. The asterisk in (B) indicates the connectivity between the R protons of Tyr and dmP.
degenerate. In the tripeptides, on the other hand, all resonances
of dmP were resolved, except the Hγ resonances of YdmPN
and tNdmPY.
pH 5.3. Therefore, the single conformation adopted by the Tyr-
dmP peptide group is cis in aqueous solutions.
Two distinct sets of resonances were observed in the one-
dimensional 1H spectrum of NdmPY with relative populations
of 90% and 10% at 10 °C, indicating that two average
conformations of the Asn-dmP peptide group are present in
solution. For the abundant species, a strong NOE was observed
between the Asn HR and the dmP HR, characterizing the peptide
group conformation for this species as cis. For the less abundant
The NOESY experiments with the tripeptides were used for
determining the conformation of the peptide group preceding
the proline residues. For an X-Pro peptide group, the NOE
connectivity between the HR of the X residue and the HR of the
Pro is indicative of a cis conformation, and a strong NOE
connectivity between the Hδ of Pro (or the Hꢀ3 of dmP) and the
HR of the X residue is indicative of a trans conformation.52 For
ꢀ
species, NOEs between the Asn HR and the CH3 protons of
1
YdmPN, only one set of resonances was observed in the H
dmP were observed (Figure 2), indicating that the Asn-dmP
peptide group adopts the trans conformation. The existence of
these two species and interconversion between them were also
confirmed by the presence of chemical exchange cross-peaks
between the cis and trans species for all resonances of NdmPY
in the NOESY experiments collected at 25 °C (Figure 2).
spectrum, demonstrating that a single average conformation
exists for this tripeptide. A strong NOE connecting the Tyr HR
to the dmP HR (Figure 1B) was observed while no significant
NOE connectivity between the Tyr HR and the resonances of
the two methyl groups (ꢀCH3) of dmP was observed at 25 °C,

RNase A Fragments with Dimethylproline
J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11563
Figure 2. Regions of 2D NOESY spectrum of NdmPY: cis-/trans-
NdmPY, HR-HR,δ region. Experimental conditions for the spectrum
were the same as in Figure 1. The asterisk indicates the connectivity
between the R protons of Asn and dmP in the cis conformation.
Figure 3. Plot of ln k (A) and ln Kcft (B) for NdmPY. The kinetic
data were obtained from the temperature jump experiments of NdmPY
in H2O, pH 5.3.
Chemical exchange cross-peaks were identified on the basis of
their phases, which are in phase with the diagonal peaks but
out of phase from the NOE cross-peaks.
the relative populations of cis and trans species were determined
from the integrated areas of the peaks corresponding to the
resonances of the methyl groups on the δ carbon of dmP. The
percentage of trans conformer increased by ∼11% over the
temperature range of 6-80 °C. Assuming that cis/trans isomer-
ization is a two-state process, the equilibrium constant was
expressed as [cis]/[trans] at each temperature, from which the
thermodynamic parameters (∆G°, ∆H°, and ∆S°) and the time
constant for isomerization were determined (Figure 3).2 The
values of ∆G°, ∆H°, and ∆S° are ∼1.15 ( 0.01 kcal/mol at 25
The observed coupling constants, together with NOE analyses,
indicate that the side-chain rotamers of Tyr in YdmPN and Asn
in cNdmPY have a dihedral angle ø1of ∼180° ( 30°, and the
Asn of YdmPN has a dihedral angle ø1 of -60° ( 30°. These
dihedral angles are relatively unusual, since most peptides of
this size do not have a specific conformation for the dihedral
3
angle ø1. JHRHâ coupling constants were not informative for
Asn and Tyr of tNdmPY and for Tyr of cYdmPN because their
values range between 6 and 8 Hz. In these cases, we assigned
Hâ stereospecifically on the basis of the relative intensity of
the NOESY cross-peak.
°C, ∼2.40 ( 0.01 kcal/mol,and ∼4 ( 0.1 cal mol-1 deg-1
,
respectively. The transition from the cis to the trans conforma-
tion is unfavored for NdmPY. The equilibrium constants
increased with increasing temperature. The free energy differ-
ence between cis and trans isomers of dmP (cis f trans) in
this peptide was different from the corresponding value for the
proline-containing peptide Asn-Pro-Tyr (∼ -0.93 kcal/mol at
25 °C, unpublished results).
Determination of Thermodynamic Parameters. In folded
states, X-Pro peptide groups usually adopt a single conformation
with some exceptions as in the case of Staphylococcal nuclease,
which displays multiple conformations because of cis/trans
isomerization of the peptide group.60 In unfolded proteins or
peptides, the population of the cis form varies from 0% to
70%2,61 and depends primarily on the preceding residue type,
as well as on the solvent, pH, temperature, and neighboring
residues. For YdmPN, only one set of resonances was observed
throughout the entire temperature range, indicating that, within
the detection limit of 1%, the Y-dmP peptide group adopts only
the cis conformation.
From the temperature dependence of the rate constant kcft
of eq 1, we obtained a barrier, ∆Eq, of about 13 ( 1.0 kcal/mol
between the cis and trans forms of Asn-dmP in NdmPY, which
was smaller than that which is usually found for the X-Pro group
(∼20 kcal/mol).2
NOE Connectivity for a Turn. In small peptides, even those
with a tendency to adopt a unique structure, NOEs are usually
very small or not seen, and their dihedral angles (i.e., J
couplings) do not provide any useful information, because of
the presence of an ensemble of conformations. Even though
the tripeptides investigated here are small, they give rise to
NOEs and coupling constants that indicate the existence of
populations with folded conformations. More long-range NOEs
were found for YdmPN than for NdmPY.
The relative populations of the cis and trans conformers of
NdmPY peptide are listed in Table 1 for the range of temper-
atures from 6 to 80 °C. From the two distinct sets of resonances,
(60) Evans, P. A.; Dobson, C. M.; Kautz, R. A.; Hatfull, G.; Fox, R. O.
Nature 1987, 329, 266.
(61) Yao, J.; Feher, V. A.; Espejo, B. F.; Reymond, M. T.; Wright, P.
E.; Dyson, H. J. J. Mol. Biol. 1994, 243, 736.

11564 J. Am. Chem. Soc., Vol. 121, No. 49, 1999
An et al.
Figure 4. Comparison of the i-i+2 NOEs of YdmPN, YPN, NdmPY,
and NPY (10 °C). NOEs were normalized with respect to the
corresponding NOEs between HR of the preceding amino acid and HR
of Pro/dmP in cis conformations.
Figure 5. Ensemble of 10 structures of YdmPN (A) and superposition
of that NMR structure which was closest to the average on the 92-94
fragment of the X-ray structure36 of RNase A (B). The N-terminus of
YdmPN is CH3CO, and the C-terminus is COO-. White and black
models in (B) represent the 92-94 YPN stretch of the X-ray structure36
and the NMR structure, respectively. N and C refer to the N- and
C-termini of the fragments.
The most convincing evidence for a turn comes from an NMR
criterion suggested by Yao et al.61 for indicating the presence
of a type VI turn (with a cis X-Pro peptide group) in a sequence
which contains aromatic residues preceding and/or following
proline, viz., a cross-turn NOE between HR of the residue
preceding and HN of the residue following Pro (i-i+2 NOE).
Observation of an i-i+2 NOE for the peptides in both the trans
and the cis conformations indicates the presence of type I and
VI turn conformations,52,62 respectively. NOEs between HR of
Tyr and HN of Asn, and between HR of Asn and HN of Tyr,
were observed in the spectra of YdmPN and cNdmPY,
respectively, indicating that they have a turn conformation. The
relative intensities of the i-i+2 NOEs of YdmPN and cNdmPY
were normalized with respect to the NOE between HR of the
residue preceding dmP and HR of dmP. The corresponding
i-i+2 NOEs from cis-YPN and cis-NPY were also normalized
with respect to their NOEs between HR of the residue preceding
Pro and HR of Pro, respectively. The intensities of the NOEs
from YdmPN and NdmPY were larger than those from YPN
and NPY, respectively (Figure 4). These results indicate that
the tripeptides with dmP have a higher percentage of a type VI
turn conformation than tripeptides with Pro. It seems that dmP
increases not only the cis propensity but also the turn propensity
in the cis population of the peptides and perhaps in proteins as
well.
Temperature Coefficient of HN Resonances. The temper-
ature dependence of the amide proton chemical shifts can be
used to determine whether these protons are protected from the
solvent. The temperature coefficients with an absolute value
lower than 5 ppb/deg have been used for identifying amide
protons which might be involved in hydrogen bonding in a close
family of peptides.61 We examined whether hydrogen bonding
contributes to the stabilization of the folded population of these
peptides. The 1D spectra of YdmPN and NdmPY, recorded from
6 to 60 °C in 5-deg steps, were analyzed. The temperature
coefficients of the YdmPN amide protons indicate that the
backbone HNs of Tyr and Asn may not be involved in hydrogen
bonds (∆δ/∆T ≈ - 9 ppb/K). However, the NOE between HR
of Tyr and HN of Asn in YdmpN suggests that HN of Asn is
positioned close to the carbonyl oxygen of Ac. For cNdmPY,
on the other hand, the temperature coefficients of Asn HN and
Tyr HN were ∼ - 7.9 and -6.4 ppb/K, respectively, suggesting
some protection for Tyr HN from the solvent. Interestingly, in
tNdmPY, the temperature coefficients of HN of Asn and Tyr
were ∼ - 5.5 and - 5.7 ppb/K, respectively, suggesting the
existence of some interaction that involves these amide protons.
These results are similar to previous results from Pro-containing
peptides that also did not show a strong hydrogen bond even
though their structures contained a type VI turn without a
hydrogen bond.62,63
These data suggest that there is a turn without a strong
intramolecular hydrogen bond involving the amide protons; i.e.,
there is no substantial stabilization of the turn conformation by
hydrogen bonding.
Solution Structures of YdmPN, cNdmPY, and tNdmPY.
For calculating the structures of YdmPN, cNdmPY, and
tNdmPY, 75, 42, and 18 NOEs, respectively, were used with
the X-plor program. Since the percentage of NdmPY molecules
containing a trans-NdmP group is low, the number of NOEs is
small and their intensities are weak. Three and two dihedral
angle constraints were used for the structures of YdmPN and
cNdmPY, respectively.
All structures are consistent with the constraints, with the
distance or dihedral angle violations being less than 0.1 Å or
5°, respectively. Superpositions of the 10 best structures, which
converged close to one conformation, are shown in Figures 5A,
6A, and 7A. The backbone RMSDs of each of the 10 structures
of YdmPN, cNdmPY, and tNdmPY, with respect to their own
average structure, are 0.022 ( 0.001, 0.131 ( 0.014, and 0.361
( 0.012 Å, respectively, suggesting that good convergence was
obtained in the calculation of these structures. The corresponding
backbone and side-chain heavy-atom RMSDs are 0.460 ( 0.020,
0.848 ( 0.072, and 0.954 ( 0.141 Å for YdmPN, cNdmPY,
and tNdmPY, respectively. The corresponding all-atom RMSDs
are 0.633 ( 0.017, 0.907 ( 0.187, and 1.033 ( 0.133 Å for
YdmPN, cNdmPY, and tNdmPY, respectively. The tNdmPY
structures are not as well characterized as are the other two cis
forms, because fewer NOEs were observed and used in the
calculation.
From the best structures in Figures 5A, 6A, and 7A, a close
interaction between the dmP side chain and the Tyr rings is
(62) Dyson, H. J.; Rance, M.; Houghton, R. A.; Lerner, R. A.; Wright,
P. E. J. Mol. Biol. 1988, 201, 161.
(63) Oka, M.; Montelione, G. T.; Scheraga, H. A. J. Am. Chem. Soc.
1984, 106, 7959.

RNase A Fragments with Dimethylproline
J. Am. Chem. Soc., Vol. 121, No. 49, 1999 11565
-23.3°, respectively, in the structures shown in Figures 5A and
6A. Even though the temperature coefficients were not signifi-
cantly low, suggesting no hydrogen bonding, these values
suggest that there is some type VIa turn character in these
peptides. Yao at al. found a similar kind of turn in a longer
peptide with a proline in the cis form.61
Finally, to examine the dependence of the structures on the
NMR constraints, 500 steps of EM without any NOE and
dihedral angle constraints were applied to the 10 best structures.
The minimized structures without any experimental constraints,
especially the backbone heavy atoms, did not deviate much from
the best starting structures, indicating that the best structures
with NOE constraints were, indeed, close to their low-energy
state. The backbone RMSDs of each of the newly minimized
10 structures of YdmPN, cNdmPY, and tNdmPY, with respect
to the previous average structure, are 0.159 ( 0.011, 0.196 (
0.100, and 0.385 ( 0.048 Å, respectively. The ring stacking
between Tyr and dmP was maintained, but the ring of Tyr
moved out to a greater extent from dmP than in the initial best
structures of YdmPN, cNdmPY, and tNdmPY. The type VIa
turn was maintained in both the YdmPN and cNdmPY peptides.
Comparison between Calculated Structures and Corre-
sponding Stretches from the X-ray Structure of the Bovine
RNase A. We superimposed the best structures of YdmPN and
the cNdmPY from the calculations onto the corresponding
stretches, 92-94 and 113-115, respectively, in the RNase A
X-ray crystal structure. The RMSDs of the backbone atoms are
0.162 and 0.652 for YdmPN and cNdmPY, respectively. In
YdmPN, the side chains of all residues preceding and following
dmP have similar interactions and positions with respect to the
side chains of the residue preceding and following Pro in wild-
type RNase A, respectively (Figure 5B). In cNdmPY, the side
chains of the Asn residue have a conformation similar to the
corresponding X-ray conformation, but the side chains of the
Tyr residues point in opposite directions. In the NMR structure
of the cNdmPY tripeptide, the side chain of Tyr is close to that
of dmP, whereas the Tyr115 side chain of RNase A is positioned
far away from Pro114 and close to the side chain of Tyr73 in
the X-ray structure. This is most likely due to the presence of
additional interactions of the side chain of Tyr in NPY with
neighboring residues in the whole protein which are absent in
the tripeptides.
Figure 6. Ensemble of 10 structures of cNdmPY (A) and superposition
of that NMR structure which was closest to the average on the 113-
115 NPY fragment of the X-ray structure36 of RNase A (B). The
N-terminus of cNdmPY is CH3CO, and the C-terminus is COO-. White
and black models in (B) represent the 113-115 stretch of the X-ray
structure36 and the NMR structure, respectively. N and C refer to the
N- and C-termini of the fragments.
Figure 7. Ensemble of 10 structures of tNdmPY (A) and superposition
of that NMR structure which was closest to the average on the X-ray
structure25 of the isolated NPY peptide (B). The N-terminus of tNdmPY
is CH3CO, and the C-terminus is COO-. White and black models in
(B) represent the X-ray fragment and the NMR structure, respectively.
N and C refer to the N- and C-termini of the fragments.
Conclusion
For the reasons stated in the Introduction, it is desirable to
have a proline derivative for which the X-Pro peptide group is
locked in the cis conformation. The experiments presented here
demonstrate that L-dmP satisfies this requirement completely
for YdmPN and largely (but not completely) for NdmPY.
Therefore, YdmPN, but not NdmPY, is suitable for the
incorporation into RNase A.
This investigation was facilitated by the development of
procedures to separate L-dmP from its D-isomer, and to couple
the sterically hindered L-dmP residue as the central residue
between two naturally occurring amino acids in good yields.
This will enable these tripeptides to be easily incorporated into
RNase A by fragment coupling, using a synthetic procedure64
that was developed for this purpose.
evident in all three peptides with the same degree of ring
stacking. The structures of YdmPN are consistent with the
widely different JHRHâ coupling constants (10.4 and 5.4 Hz),
which indicate that there is not much flexibility of the aromatic
rings and the side chain is quite rigid, since most of the JHRHâ
coupling constants from similar size peptides fall between 6
and 8 Hz. The Hꢀ1 of dmP is close to the Tyr ring preceding
dmP in YdmPN, while the Hꢀ2 of dmP is close to Tyr following
dmP in both cNdmPY and tNdmPY, which are consistent with
the NOEs and with the stereospecific assignments of the dmP
ring protons.
Type VI turns are common in proteins with a cis proline
residue. There are two sets of type VI turns in proteins, type
VIa (hydrogen-bonded) and VIb (non-hydrogen-bonded) with
dihedral angles ψ close to ∼0 and 150°, respectively.44 The
values of ψ of dmP in YdmPN and cNdmPY are ∼ 24.9° and
The solution structures of YdmPN and NdmPY have been
determined by 2D H NMR measurements. The ensembles of
calculated 3D structures satisfy the available distance and
dihedral angle constraints. YdmPN is observed to adopt a
1
(64) Welker, E.; Scheraga, H. A. Biochem. Biophys. Res. Commun. 1999,
254, 147.

11566 J. Am. Chem. Soc., Vol. 121, No. 49, 1999
An et al.
completely cis conformation over the temperature range from
6 to 60 °C, whereas NdmPY is mostly cis, with the trans
conformation increasing from about 10% to about 21% as the
temperature increased from 6 to 80 °C. Based on dRN(i,i+2)
NOEs, both tripeptides have a high population of a type VIa
reverse-turn conformation, similar to that in native RNase A;
i.e., the dmP residue adopts the same backbone structure as
proline and, therefore, can serve as a substitute for proline. Turn
formation is thought to be an initial event in protein folding,3
e.g., in chain-folding initiation sites.13
Compared to these dmP-related tripeptides, corresponding
peptides containing proline have a lower percentage of cis
conformation, e.g., YPN (29%),65 SYPNDV (40%),61 NPY
(13%),65 and SNPYDV (7.3%).61 Presumably, the high prefer-
ence for the cis conformation in the dmP peptides arises from
steric interaction of the two methyl groups on the δ-carbon atom
of the dmP residue with the preceding residue.
Finally, the NMR structure of YdmPN is compatible with
that of YPN in RNase A, the NMR structure of cNdmPY is
only partially compatible with that of NPY in RNase A
(presumably because of interactions with other residues in the
proteins), and the NMR structure of tNdmPY is compatible with
the trans-NPY fragment.
Acknowledgment. We thank Dr. T. L. Bogard of the
American Cyanamid Co. for the initial dmP sample. This work
was supported by a research grant from the National Institutes
of Health (Grant GM-24893). Support was also received from
the Cornell Biotechnology Resource Center. A preliminary
report was presented at the 15th American Peptide Symposium
(Nashville, TN, June 4, 1997; published in Peptides: Frontiers
of Peptide Science; Tam, J. P., Kaumaya, P. T. P., Eds.; Kluwer
Academic Publ.: Dordrecht, 1999; p 422).
(65) Stimson, E. R.; Montelione, G. T.; Meinwald, Y. C.; Rudolph, R.
K. E.; Scheraga, H. A. Biochemistry 1982, 21, 5252.
JA9930317