Effect of 3-hydroxyproline residues on collagen stability

Article abstract of DOI:10.1021/ja034015j

Collagen is an integral part of many types of connective tissue in animals, especially skin, bones, cartilage, and basement membranes. A fibrous protein, collagen has a triple-helical structure, which is comprised of strands with a repeating Xaa-Yaa-Gly sequence. L-Proline (Pro) and 4(R)-hydroxy-L-proline (4-Hyp) residues occur most often in the Xaa and Yaa positions. The 4-Hyp residue is known to increase markedly the conformational stability of a collagen triple helix. In natural collagen, a 3(S)-hydroxy-L-proline (3-Hyp) residue occurs in the sequence: 3-Hyp-4-Hyp-Gly. Its effect on collagen stability is unknown. Here, two host-guest peptides containing 3-Hyp are synthesized: (Pro-4-Hyp-Gly)₃-3-Hyp-4-Hyp-Gly-(Pro-4-Hyp-Gly)₃ (peptide 1) and (Pro-4-Hyp-Gly)₃-Pro-3-Hyp-Gly-(Pro-4-Hyp-Gly)₃ (peptide 2). The 3-Hyp residues in these two peptides diminish triple-helical stability in comparison to Pro. This destabilization is small when 3-Hyp is in the natural Xaa position (peptide 1). There, the inductive effect of its 3-hydroxyl group diminishes slightly the strength of the interstrand 3-HypC=O...H-NGly hydrogen bond. The destabilization is large when 3-Hyp is in the nonnatural Yaa position (peptide 2). There, its pyrrolidine ring pucker leads to inappropriate mainchain dihedral angles and interstrand steric clashes. Thus, the natural regioisomeric residues 3-Hyp and 4-Hyp have distinct effects on the conformational stability of the collagen triple helix.

Full text of DOI:10.1021/ja034015j

Published on Web 04/30/2003
Effect of 3-Hydroxyproline Residues on Collagen Stability
†
‡,§
|
,†,‡
Cara L. Jenkins, Lynn E. Bretscher, Ilia A. Guzei, and Ronald T. Raines*
Contribution from the Department of Chemistry and Department of Biochemistry,
UniVersity of Wisconsin-Madison, Madison, Wisconsin 53706
Received January 2, 2003; E-mail: raines@biochem.wisc.edu
Abstract: Collagen is an integral part of many types of connective tissue in animals, especially skin, bones,
cartilage, and basement membranes. A fibrous protein, collagen has a triple-helical structure, which is
comprised of strands with a repeating Xaa-Yaa-Gly sequence. L-Proline (Pro) and 4(R)-hydroxy-L-proline
(
4-Hyp) residues occur most often in the Xaa and Yaa positions. The 4-Hyp residue is known to increase
markedly the conformational stability of a collagen triple helix. In natural collagen, a 3(S)-hydroxy-L-proline
3-Hyp) residue occurs in the sequence: 3-Hyp-4-Hyp-Gly. Its effect on collagen stability is unknown. Here,
two host-guest peptides containing 3-Hyp are synthesized: (Pro-4-Hyp-Gly) -3-Hyp-4-Hyp-Gly-(Pro-4-
Hyp-Gly) (peptide 1) and (Pro-4-Hyp-Gly) -Pro-3-Hyp-Gly-(Pro-4-Hyp-Gly) (peptide 2). The 3-Hyp residues
(
3
3
3
3
in these two peptides diminish triple-helical stability in comparison to Pro. This destabilization is small
when 3-Hyp is in the natural Xaa position (peptide 1). There, the inductive effect of its 3-hydroxyl group
diminishes slightly the strength of the interstrand 3-HypCdO‚‚‚HsNGly hydrogen bond. The destabilization
is large when 3-Hyp is in the nonnatural Yaa position (peptide 2). There, its pyrrolidine ring pucker leads
to inappropriate mainchain dihedral angles and interstrand steric clashes. Thus, the natural regioisomeric
residues 3-Hyp and 4-Hyp have distinct effects on the conformational stability of the collagen triple helix.
the collagen triple helix.11 This additional stability arises from
Introduction
stereoelectronic effects that preorganize a 4-Hyp residue in a
Collagen is the most abundant protein in animals, comprising
approximately one-third of the total protein by weight.
Collagen has a unique tertiary structure, which consists of three
12-15
1
,2
conformation that befits a triple helix.
In contrast to 4-Hyp,
no information is available on the effect of 3-Hyp residues on
collagen stability.
parallel left-handed polyproline-II-type strands wound tightly
around a common axis.3-6 Maintenance of this triple helix is
Here, we synthesize collagen-like peptides that contain a
3-Hyp residue. We use these peptides to assess the contribution
of 3-Hyp to the conformational stability of the collagen triple
helix. We find that the presence of a 3-Hyp residue, in surprising
contrast to a 4-Hyp residue, destabilizes a collagen triple helix.
Like the stability endowed by 4-Hyp residues, the instablity
imposed by 3-Hyp residues appears to arise largely from
inductive effects of its pendant hydroxyl group.
essential for many biological functions.7
Each strand of collagen consists of the repeating sequence
Xaa-Yaa-Gly, where Xaa and Yaa are often proline (Pro)
residues. The post-translational hydroxylation of some of these
proline residues generates 4(R)-hydroxy-L-proline (4-Hyp)
residues and, to a lesser extent, 3(S)-hydroxy-L-proline (3-Hyp)
residues. 3-Hyp, which is found in the natural triplet (3-Hyp-
8
4
-Hyp-Gly), arises from the action of the enzyme prolyl-3-
Experimental Section
9
hydroxylase rather than by regiochemical ambiguity by the
enzyme prolyl-4-hydroxylase.10
General. Reagents were obtained from Aldrich Chemical (Milwau-
kee, WI) or Fisher Scientific (Hanover Park, IL) and used without
further purification. Amino acids and their derivatives were obtained
from Fisher Scientific, Bachem Bioscience (King of Prussia, PA), or
Novabiochem (San Diego, CA). Dichloromethane was distilled over
In seminal work, Prockop and co-workers demonstrated that
-Hyp residues enhance greatly the conformational stability of
4
†
Department of Chemistry, University of Wisconsin-Madison.
Department of Biochemistry, University of Wisconsin-Madison.
Present address: Department of Biochemistry, Medical College of
‡
§
(
9) The crystalline structure of a bacterial proline 3-hydroxylase has been
determined by X-ray diffraction analysis (Clifton, I. J.; Hsueh, L.-C.;
Baldwin, J. E.; Harlos, K.; Schofield, C. J. Eur. J. Biochem. 2001, 268,
Wisconsin, Milwaukee, WI 53226.
|
To whom correspondence should be addressed regarding X-ray dif-
fraction analysis. Molecular Structure Laboratory, Department of Chemistry,
University of Wisconsin-Madison, 1101 University Avenue, Madison, WI
3706-1396.
6
625-6636). The product of its catalysis is the amino acid 3(R)-hydroxy-
L-proline, which has not been found in natural collagen.
5
(
(10) Tryggvason, K.; Risteli, J.; Kivirikko, K. Biochem. Biophys. Res. Commun.
1977, 76, 275-281.
1) Prockop, D. J.; Kivirikko, K. I. Annu. ReV. Biochem. 1995, 64, 403-434.
2) Myllyharju, J.; Kivirikko, K. I. Ann. Med. (Helsinki) 2001, 33, 7-21.
3) Ramachandran, G. N.; Kartha, G. Nature 1954, 269-270.
4) Ramachandran, G. N.; Kartha, G. Nature 1955, 176, 593-595.
5) Rich, A.; Crick, F. H. C. Nature 1955, 176, 915-916.
(
(
(
(
(
(
(
(11) Berg, R. A.; Prockop, D. J. Biochem. Biophys. Res. Commun. 1973, 52,
115-120.
(12) Holmgren, S. K.; Taylor, K. M.; Bretscher, L. E.; Raines, R. T. Nature
1998, 392, 666-667.
6) Rich, A.; Crick, F. H. C. J. Mol. Biol. 1961, 3, 483-506.
7) Fields, G. B. Connect. Tissue Res. 1995, 31, 235-243.
(13) Engel, J.; Prockop, D. J. Matrix Biol. 1998, 17, 679-680.
(14) Bretscher, L. E.; Jenkins, C. L.; Taylor, K. M.; DeRider, M. L.; Raines, R.
T. J. Am. Chem. Soc. 2001, 123, 777-778.
8) Gryder, R. M.; Lamon, M.; Adams, E. J. Biol. Chem. 1975, 250, 2470-
2
474.
(15) Jenkins, C. L.; Raines, R. T. Nat. Prod. Rep. 2002, 19, 49-59.
6422
9
J. AM. CHEM. SOC. 2003, 125, 6422-6427
10.1021/ja034015j CCC: $25.00 © 2003 American Chemical Society

Effect of 3-Hydroxyproline Residues on Collagen Stability
A R T I C L E S
Scheme 1
Scheme 2
2
CaH (s) or drawn from a Baker Cycletainer. Thin-layer chromatography
was performed by using aluminum-backed plates coated with silica
gel containing F254 phosphor and visualized by UV illumination or
R-(trifluoromethyl)phenylacetyl chloride (0.010 mL, 0.053 mmol), was
added, and the resulting mixture was stirred at room temperature for
15 min. EtOAc (1 mL) was then added to the reaction mixture, followed
staining with I
NMR spectra were obtained with Bruker AC-300 and Varian UNITY-
00 spectrometers. Mass spectra were obtained with a Micromass LCT
ESI or Perkin-Elmer Voyager MALDI-TOF instrument.
(S)-tert-Butoxy-L-proline tert-Butyl Ester Tosylate (5). N-(9H-
2
, p-anisaldehyde stain, or phosphomolybdic acid stain.
by concentrated HCl (2 drops) and H O (1 mL). The layers were mixed
2
thoroughly and then separated. The aqueous layer was extracted twice
5
more with EtOAc, and the combined organic extracts were dried over
19
MgSO
4
(s), filtered, and concentrated to a colorless oil. The F NMR
3
spectrum of the oil contained only a single peak at -72 ppm, indicating
Fluoren-9-ylmethoxycarbonyl)-3-tert-butoxy-L-proline (6) was synthe-
sized by the route shown in Scheme 1. 2-Methyl-propene (approxi-
mately 10 mL) was condensed into a pear-shaped flask at -78 °C and
then added to a suspension of 3-hydroxyproline (0.50 g, 3.8 mmol)
and p-toluenesulfonic acid hydrate (3.00 g, 15.8 mmol) in dichlo-
romethane (25 mL) at -78 °C. The reaction was stirred for 3 d, allowing
the mixture to come to room temperature. The reaction mixture was
then cooled to 0 °C, vented carefully, and then poured into a separatory
that compound 6 had maintained its stereochemical integrity.
13
N- C
2
-Acetyl-3(S)-tert-butoxy-L-proline tert-Butyl Ester (7).
[13
3
CH ]Ac-3-Hyp-OMe (8) was synthesized by the route shown in
Scheme 2. The C in the acetyl group facilitated the measurement of
amide bond trans/cis ratios because both the acetyl and ester methyl
groups overlap with other resonances in the H spectrum. Compound
(771 mg, 1.86 mmol) was dissolved in 25 mL of dry dichloromethane
and cooled to 0 °C. CH COCl (200 µL, 2.78 mmol) was added,
13
1
5
13
3
3
funnel and washed twice with a saturated aqueous solution of NaHCO .
The combined aqueous layers were extracted once with dichlo-
romethane, and the combined organic layers were then washed once
followed by triethylamine (800 µL, 5.74 mmol). The reaction was stirred
at 0 °C for 1 h and then at room temperature overnight. The reaction
mixture was transferred to a separatory funnel and washed twice with
4
with water, dried over MgSO (s), filtered, and concentrated under
water. The organic layer was dried over MgSO
concentrated under reduced pressure to a brown oil. The product was
purified by silica gel chromatography, eluting with CH Cl containing
MeOH (5% v/v) to yield compound 7 as a clear, colorless oil (355
4
(s), filtered, and
reduced pressure to yield compound 5 as a yellow oil (1.142 g, 72%).
1
H NMR (300 MHz, CDCl
3
) δ (ppm) 7.78 (d, J ) 8.1 Hz, 2H), 7.14
2
2
(d, J ) 8.1 Hz, 2H), 4.29 (m, 1H), 3.99 (d, J ) 2.9 Hz, 1H) 3.50-
3
1
.43 (m, 1H), 3.37-3.27 (m, 1H), 2.33 (s, 3H), 2.00-1.92 (m, 1H),
.83-1.76 (m, 1H), 1.46 (s, 9H), 1.18 (s, 9H).
1
mg, 66.8%). H NMR (300 MHz, CDCl ) δ (ppm) major isomer 4.28
(
3
bs, 1H), 4.17 (m, 1H), 3.65 (m, 2H), 2.20-1.80 (m, 2H) 2.09 (d, 1JCH
N-(9H-Fluoren-9-ylmethoxycarbonyl)-3(S)-tert-butoxy-L-pro-
) 132 Hz, 3H), 1.46 (s, 9H), 1.22 (s, 9H); minor isomer δ (ppm) 4.30
1
line (6). Compound 5 (412 mg, 0.99 mmol) was suspended in 1.0 M
NaOH (10 mL, 10 mmol) and stirred until the reaction mixture became
homogeneous, that is, for about 4 h. The reaction mixture was then
(m, 1H), 4.07 (s, 1H), 3.65 (m, 2H), 2.20-1.80 (m, 2H), 1.96 (d, J
CH
13
) 132 Hz, 3H), 1.48 (s, 9H), 1.24 (s, 9H). C NMR (75 mHz, CDCl3,
13
8 scans) δ (ppm) 22.10, 22.25 ( C-labeled carbon).
13
cooled to 0 °C and acidified with concentrated HCl to pH 0. Na
2
CO
3
N-( C -Acetyl)-3(S)-hydroxy-L-proline Methyl Ester (8). Com-
2
(
1.1 g) was added immediately, followed by dioxane (25 mL). Fmoc-
pound 7 (355 mg, 1.24 mmol) was dissolved in MeOH. Trifluoroacetic
acid (0.10 mL) was added, followed by a dropwise addition of SOCl2
(1.00 mL, 13.7 mmol). The reaction mixture was heated at reflux for
1 h and then cooled to room temperature and concentrated under
reduced pressure. The resulting oil was purified by chromatography
on silica gel, eluting with CH Cl containing MeOH (5% v/v) to yield
Cl (292 mg, 1.13 mmol) was added in small portions, and the resulting
mixture was stirred at room temperature for 4 h. The reaction mixture
was then poured into a separatory funnel with 50 mL of H O and
2
extracted with diethyl ether (3 × 25 mL). The aqueous layer was then
combined with 50 mL of ethyl acetate in a beaker, cooled to 0 °C, and
acidified with concentrated aqueous HCl to pH 0. The layers were
separated, and the aqueous layer was extracted again with EtOAc (25
mL). The combined ethyl acetate layers were dried over MgSO
filtered, and the solvent was removed under reduced pressure to yield
compound 6 as a white solid (0.148 g, 36.6% over two steps). H NMR
(
2
2
1
amide 8 as a clear, colorless oil (183 mg, 79%). H NMR (500 MHz,
CDCl , two rotamers) δ (ppm) major isomer 4.43 (s, 1H), 4.38 (d, 4
3
1
4
(s) and
Hz, 1H), 3.71-3.61 (m, 2H), 3.68 (s, 3H), 2.05 (d, J ) 128.5 Hz,
CH
3H), 2.14-2.06 (m, 1H), 2.01-1.94 (m, 1H); minor isomer δ (ppm)
4.50 (d, 4 Hz, 1H), 4.33 (s, 1H), 3.73 (s, 3H), 3.71-3.61 (m, 2H),
1
1
13
two rotamers, 500 MHz, CDCl
3
) δ (ppm) 7.75 and 7.70 (d, J ) 7 and
.5 Hz, 2H), 7.60 and 7.55 (dd, J ) 8, 8 Hz and 7.5, 10 Hz, 2H),
.41-7.27 (m, 4H), 4.68-4.12 (m, 5H), 3.76-3.61 (m, 2H), 2.12-
1.94 (d, J ) 128.5 Hz, 3H). C NMR (125 MHz, CDCl , two
CH
3
7
7
2
rotamers) δ (ppm) 174.44, 174.03, 172.98, 75.08, 73.66, 69.46, 67.68,
54.12, 53.83, 46.72, 45.15, 32.64, 30.93, 21.84, 21.60.
1
3
.01 (m, 1H), 1.89-1.80 (m, 1H), 1.22 (s, 9H). C NMR (125 MHz,
, two rotamers) δ (ppm) 175.52, 174.12, 155.92, 154.62, 144.00,
43.84, 143.74, 141.26, 127.71, 127.57, 127.05, 126.96, 125.15, 125.06,
pK Determinations. The pK values of the carboxylic acid groups
a
a
CDCl
3
of 3-HypOH and 4-HypOH were determined by H NMR spectroscopy
1
1
1
6
performed in aqueous solutions of different pH. D O stock solutions
2
24.96, 119.95, 119.86, 75.26, 75.10, 73.67, 67.94, 67.62, 67.34, 66.05,
consisted of 210 mM 3-HypOH (27.6 mg in 1.0 mL), 209 mM
4-HypOH (27.4 mg in 1.0 mL), 104 mM sodium 2,2-dimethyl-2-
silapentane-5-sulfonate (DSS, 40.8 mg in 1.0 mL), 100 mM D PO
5.93, 61.15, 61.13, 47.12, 45.46, 45.30, 33.09, 32.23, 28.08. ESIMS
+
(
m/z) [M - H] calculated, 408.1811; found, 408.1805.
3
4
To determine whether epimerization of the R-carbon had occurred
3 4
(68.5 µL of 85% w/v D PO in 10.0 mL), 100 mM DCl (156 µL of
upon saponification of the tert-butyl ester 5, the hydroxyl group was
deprotected by acidolysis of its tert-butyl protecting group and the
resulting hydroxyl group was converted into its Mosher ester. Briefly,
Fmoc-3-HypOH (17 mg, 0.048 mmol) was dissolved in 1 mL of
20% DCl in 10.0 mL), and 100 mM NaOD (67.6 µL of fresh 40% w/v
NaOD in 10.0 mL). Each sample contained 49 µL of the 3-Hyp or
4-Hyp stock solution, 1 µL of the DSS stock solution, and 950 µL of
buffer, according to Tables S1 and S2 in the Supporting Information.
The pH of each solution was measured by using a standard pH electrode
dioxane. K
2 3
CO (20 mg, 0.14 mmol) was added to this solution,
1
2
followed by H O (2 drops). A Mosher acid chloride, R-(-)-R-methoxy-
calibrated in nondeuterated buffers. H NMR spectra were obtained
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6423

A R T I C L E S
Jenkins et al.
(
32 scans, 24 °C), and the chemical shift of the R-proton was determined
A was H
2
O containing TFA (0.1% v/v) and solvent B was CH
3
CN
+
for each sample. The resulting pH and chemical shift data were fitted
containing TFA (0.1% v/v). MALDI MS (m/z) [M + H] calculated,
to eq 1 to yield the pK
a
values.
1905; found, 1906. Peptide 2, which was (Pro-4-Hyp-Gly)
3
-Pro-3-Hyp-
Gly-(Pro-4-Hyp-Gly)
+ H] calculated, 1889; found, 1890.
3
, was purified likewise. MALDI MS (m/z): [M
+
(pH-pKa)
δlow + δhigh10
δ )
(1)
m
Thermal Denaturation Experiments. Values for T for each triple
(pH-pKa)
1
+ 10
helix were determined in triplicate by thermal denaturation experiments
monitored by CD spectroscopy on an Aviv 202 SF instrument equipped
with an automated temperature controller. A 0.2 mM solution of peptide
1 or 2 in 50 mM HOAc was incubated at 4 °C for >24 h. Aliquots of
300 µL were placed in 0.1-cm path length quartz cuvettes that had
been equilibrated at 5 °C. Wavelength scans were performed from 200
to 260 nm at 5 and 50 °C, with a slit width of 1 nm and an averaging
time of 3 s. Thermal denaturation experiments were performed by
raising the temperature from 5 to 50 °C in 3 °C steps, equilibrating for
5 min at each temperature, and monitoring at 225 nm with a 20-s
The acidity of an aqueous solution of a molecule having a titratable
group with a fraction factor near unity can be measured in D O by
using a standard pH meter equilibrated against H O buffers and adding
a correction factor of 0.4 units to the reading. The difference in pK ’s
of amino acids in H O and D
O is approximately 0.5 units.17 Thus, a
correction of about 0.1 units can be applied to a pK measurement
2
2
16
a
2
2
a
obtained by measuring apparent pH values with a standard glass
electrode. To determine a more precise correction factor, we measured
the pK
a
of 3-HypOH and 4-HypOH in D
2
O in the same manner and
m
averaging time. Values of T , which is the temperature at the midpoint
found that the 4-HypOH value was within 0.02 units of the value
of the thermal transition, were determined by fitting the data to a two-
state model using the software package NLREG v. 4.0 (Philip Sherrod).
Molecular Modeling. Molecular models of triple-helical 1 and 2
were created by modifying the structure of crystalline collagen (PDB
1
8
reported previously. Thus, we did not apply any correction factor
herein to the 3-Hyp pK value.
a
13
2
Crystallization of N-( C -Acetyl)-3(S)-hydroxy-L-proline Methyl
20
Ester (8). Equal portions of a solution of amide 8 (67 mg) in 3.0 mL
of dichloromethane were placed in six half-dram vials. Approximately
entry 1CAG ). The atomic coordinates were imported into the program
SYBYL (Tripos, St. Louis, MO). The side-chain methyl group of the
three alanine residues were replaced with a hydrogen atom, and a section
of the (Pro-Hyp-Gly) triple helix was excised from the C-terminus.
This section was subjected to energy minimization with the Tripos force
15 drops (Pasteur pipet) of a different cosolvent (hexanes, ethyl acetate,
diethyl ether, tetrahydrofuran, tert-butyl methyl ether) or no cosolvent
was added to each vial. The vials were then capped loosely and allowed
to sit at room temperature undisturbed for several days. The vial with
ethyl acetate as a cosolvent contained the crystals most suitable for
X-ray diffraction analysis.
â
field. A hydroxyl group was added to C of the proline residue in the
Xaa position of the central triplet in each strand (to mimic peptide 1)
γ
â
or moved from C to C in the Yaa position of the central triplet in
each strand (to mimic peptide 2). The dihedral angles of the 3-Hyp
residues were then altered to match those of the structure of crystalline
amide 8. The resulting structures were not subjected to further
minimization.
X-ray Diffraction Analysis. A colorless crystal with approximate
3
dimensions 0.50 × 0.50 × 0.41 mm was selected under oil under
ambient conditions and attached to the tip of a glass capillary. The
crystal was mounted in a stream of N (g) at (173 ( 2) K and centered
2
in the X-ray beam by using a video camera. Crystal evaluation and
data collection were performed on a Bruker CCD-1000 diffractometer
with Mo KR (λ ) 0.710 73 Å) radiation and a diffractometer-to-crystal
distance of 4.9 cm.
Results and Discussion
Synthetic collagen mimics have been effective in dissecting
the basis for the conformational stability of the collagen triple
1
5,21
Initial cell constants were obtained from three series of ω scans at
different starting angles. Each series consisted of 20 frames collected
at intervals of 0.3° in a 6° range about ω with an exposure time of 10
s per frame. A total of 102 reflections were obtained. The reflections
were indexed successfully by an automated indexing routine built into
the SMART program. The final cell constants were calculated from a
set of 1508 strong reflections from the actual data collection.
Data were collected by using the hemisphere data collection routine.
The reciprocal space was surveyed to the extent of a full sphere to a
resolution of 0.80 Å. A total of 2911 data sets were harvested by
collecting one set of frames with 0.3° scans in ω with an exposure
time of 10 s per frame. This redundant data set was corrected for Lorentz
and polarization effects. The absorption correction was based on fitting
a function to the empirical transmission surface as sampled by multiple
helix.
In particular, host-guest studies have revealed
important insights on the contribution of individual amino acid
residues, both natural and nonnatural.
22-27
To reveal the effect
of 3-Hyp residues on triple-helical stability, we synthesized two
host-guest peptides in which the central triplet of (Pro-4-Hyp-
Gly)7 was replaced with one containing 3-Hyp. These peptides
were (Pro-4-Hyp-Gly)3-3-Hyp-4-Hyp-Gly-(Pro-4-Hyp-Gly)3 (1)
and (Pro-4-Hyp-Gly)3-Pro-3-Hyp-Gly-(Pro-4-Hyp-Gly)3 (2).
Peptide 1 was designed to elucidate the role of the 3-Hyp-4-
Hyp-Gly triplet as is found most often in natural collagen.
Peptide 2 was designed to reveal the effect of the regiochemistry
(3-Hyp versus 4-Hyp) of the pendant hydroxyl group. These
peptides were synthesized by standard Fmoc/tBu coupling
strategies using compound 6 (Scheme 1) to introduce 3-Hyp
residues.
1
9
equivalent measurements. All software and sources of the scattering
factors were contained in the SHELXTL (version 5.1) program library
(G. Sheldrick, Bruker Analytical X-ray Systems, Madison, WI).
Conformational Stability. The unfolding of triple helices
of peptides 1 and 2 was monitored by CD spectroscopy. The
Synthesis of Peptides 1 and 2. Peptide synthesis was performed
on an Applied Biosystems Synergy synthesizer using standard Fmoc/
tBu methodology and HBTU as the coupling reagent. 3-Hyp residues
were incorporated by using compound 6.
(20) Bella, J.; Eaton, M.; Brodsky, B.; Berman, H. M. Science 1994, 266, 75-
8
1.
(
21) Fields, G. B.; Prockop, D. J. Biopolymers 1996, 40, 345-357.
22) Shah, N. K.; Ramshaw, J. A. M.; Kirkpatrick, A.; Shah, C.; Brodsky, B.
Biochemistry 1996, 35, 10262-10268.
Peptide 1, which was (Pro-4-Hyp-Gly)
Hyp-Gly) , was purified by reversed-phase HPLC on a C-18 column,
eluting with a linear gradient (5-60% v/v) of solvent B, where solvent
3
-3-Hyp-4-Hyp-Gly-(Pro-4-
(
3
(
23) Ramshaw, J. A. M.; Shah, N. K.; Brodsky, B. J. Struct. Biol. 1998, 122,
86-91.
(
24) Kwak, J.; Jefferson, E. A.; Bhumralkar, M.; Goodman, M. Bioorg. Med.
Chem. 1999, 7, 153-160.
(
16) Glasoe, P. K.; Long, F. A. J. Phys. Chem. 1960, 64, 188-190.
(
17) Hyman, H. H.; Kaganove, A.; Katz, J. J. J. Phys. Chem. 1960, 64, 1653-
(25) Persikov, A. V.; Ramshaw, J. A. M.; Brodsky, B. Biopolymers 2000, 55,
436-450.
1
655.
(18) Fasman, G. D., Ed. Practical Handbook of Biochemistry and Molecular
Biology; CRC Press: Boca Raton, FL, 1989.
(26) Persikov, A. V.; Ramshaw, J. A. M.; Kirkpatrick, A.; Brodsky, B.
Biochemistry 2000, 39, 14960-14967.
(19) Blessing, R. H. Acta Crystallogr. 1995, A51, 33-38.
(27) Kersteen, E. A.; Raines, R. T. Biopolymers 2001, 59, 24-28.
6424 J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003

Effect of 3-Hydroxyproline Residues on Collagen Stability
A R T I C L E S
Table 1. Values of Tm for Synthetic
(Pro-Hyp-Gly)3-(Xaa-Yaa-Gly)-(Pro-Hyp-Gly)3 Triple Helices
peptide
Xaa-Yaa-Gly
T
m
(°C)a
3
1
4
2
Pro-4-Hyp-Gly
3-Hyp-4-Hyp-Gly
Pro-Pro-Gly
36 ( 2b
32.7 ( 0.9b
30.5 ( 2.2
21.0 ( 2.2
Pro-3-Hyp-Gly
a
Values of Tm were determined by CD spectroscopy for peptides (0.2
mM) in 50 mM acetic acid and are the average ((SE) of at least three
b
determinations. From ref 27.
resulting Tm values, along with those of triple helices of (Pro-
4-Hyp-Gly)7 (3) and (Pro-4-Hyp-Gly)3-Pro-Pro-Gly-(Pro-4-Hyp-
Gly)3 (4), are listed in Table 1. The values confirm that, in its
natural (i.e., Yaa) position, 4-Hyp provides more conformational
stability to a triple helix than does Pro. In contrast, 3-Hyp
provides less conformational stability than does Pro in either
its natural (i.e., Xaa) or a nonnatural (i.e., Yaa) position. The
additional stability endowed by 4-Hyp is known to arise from
stereoelectronic effects.1 What is the origin of the detrimental
effect of 3-Hyp on collagen stability?
4,15
Peptide Bond Isomerization. All of the peptide bonds in
triple-helical collagen are in the trans (i.e., Z) conformation.
The 4-hydroxyl group of a 4-Hyp residue increases the trans/
cis ratio of an Xaa-4-Hyp peptide bond by approximately 50%
compared to that of a Xaa-Pro bond.28 We used NMR
spectroscopy to determine the effect of the 3-hydroxyl group
of a 3-Hyp residue on its trans/cis ratio. We found that the trans/
cis ratios of amide 8 and Ac-Pro-OMe in D2O were 4.9 and
13
Figure 1. (A) Ortep diagram of crystalline N-( C2-acetyl)-3(S)-hydroxy-
L-proline methyl ester (8) drawn with 30% probability ellipsoids. (B)
Pyrrolidine ring pucker in crystalline amide 8 (left) and crystalline N-acetyl-
4(R)-hydroxy-L-proline methyl ester (right).33
4.6, respectively. These similar values indicate that peptide bond
isomerization makes a negligible contribution to the instability
of triple-helical 1 and 2.
Structure of a 3-Hyp Residue. We used X-ray diffraction
analysis to determine the structure of crystalline amide 8 (Figure
â
â
flap atom is Ci . In the twisted conformation, atoms Ni, Ci ,
29,30
δ
â
1A).
The pyrrolidine ring of crystalline amide 8 is puckered
and Ci form the basal plane. Atom Ci resides 0.456 ( 0.004
Å above that plane, and atom Ci resides 0.153 ( 0.005 Å below
R
â
γ
δ
such that its N-C -C -O bond has a dihedral angle of
-
81.46 ( 0.16°, as would be expected from the manifestation
that plane. The phase angle φ2 ) 9.6 ( 0.3°, and the puckering
amplitude is q2 ) 0.378 ( 0.002 Å.
14,31
of a gauche effect.
OMe (8) differs significantly from the ring pucker in crystalline
Ac-4-Hyp-OMe (Figure 1B). The conformation of the pyrroli-
The ring pucker in crystalline Ac-3-Hyp-
The structure of crystalline Ac-4-Hyp-OMe contains two
33
symmetry-independent molecules. (Thus, there are two num-
bers herein for each parameter.) In both molecules, the pyrro-
1
dine ring in amide 8 is intermediate between a E envelope and
1
32
5
C â
i
E2 twisted conformation. In the envelope conformation, the
lidine ring is in the T (that is,
TCi
with a characteristic phase angle of φ2 ) 129.8 ( 0.7° (127.8
0.7°) and puckering amplitude of q2 ) 0.354 ( 0.004 Å
γ
) twisted conformation
4
(
28) Eberhardt, E. S.; Panasik, N., Jr.; Raines, R. T. J. Am. Chem. Soc. 1996,
18, 12261-12266.
29) Systematic absences in the diffraction data were uniquely consistent with
the space group P2 for a crystal of amide 8‚H O, which yielded
(
1
3
2
(
(0.389 ( 0.006 Å). The basal plane is determined by atoms
1
2
1
2
1
2
δ
R
â
Ni, Ci , and Ci . Atom Ci resides 0.28 ( 0.01 Å (0.22 ( 0.01
chemically reasonable and computationally stable refinement results. A
successful solution of the structure by the direct methods provided most
of the non-hydrogen atoms from the E-map. The remaining non-hydrogen
atoms were located in an alternating series of least-squares cycles and dif-
ference Fourier maps. All non-hydrogen atoms were refined with anisotropic
displacement coefficients. All hydrogen atoms were included in the structure
factor calculation at idealized positions and were allowed to ride on the
neighboring atoms with relative isotropic displacement coefficients. The
asymmetric unit contains one molecule of amide 8 and one molecule of
water, which forms three hydrogen bonds. The final least-squares refinement
of 138 parameters against 1963 data points resulted in residuals R (based
γ
Å) above that plane, and atom Ci resides 0.35 ( 0.01 Å (0.36
(
0.01 Å) below the plane.
The ring puckers of 3-Hyp and 4-Hyp (Figure 1B) serve to
preorganize the φ (C′i-1-Ni-Ci-C′i) and ψ (Ni-Ci-C′i-Ni+1)
dihedral angles in a conformation that is appropriate for the
3
4-36
Yaa position of a collagen triple helix.
Likewise, a 3-Hyp
2
2
on F for I g 2σ) and wR (based on F for all data) of 0.0369 and 0.0965,
respectively. The final difference Fourier map was featureless. The absolute
configuration was assigned from the synthetic route.
(32) Giacovazzo, C.; Monaco, H. L.; Artioli, G.; Viterbo, D.; Ferraris, G.; Gilli,
G.; Zanotti, G.; Catti, M. Fundamentals of Crystallography, 2nd ed.; Oxford
University Press: Oxford, UK, 2002.
(
30) The conformation of crystalline amide 8 could be influenced by a hydrogen
bond with water (Figure 1A). To search for such an influence, the geometry
of amide 8 was optimized fully at the B3LYP/6-31+G* level of theory.
The structures obtained by experiment and theory are in good agreement
(33) Panasik, N., Jr.; Eberhardt, E. S.; Edison, A. S.; Powell, D. R.; Raines, R.
T. Int. J. Pept. Protein Res. 1994, 44, 262-269.
(34) Vitagliano, L.; Berisio, R.; Mazzarella, L.; Zagari, A. Biopolymers 2001,
58, 459-464.
(
data not shown). Some bond distances vary beyond the estimated standard
deviations but fall within the expected ranges. The pyrrolidine rings have
an essentially identical conformation with the hydroxyl oxygens overlapping
exactly.
(35) Berisio, R.; Vitagliano, L.; Mazzarella, L.; Zagari, A. Protein Sci. 2002,
11, 262-270.
(36) DeRider, M. L.; Wilkens, S. J.; Waddell, M. J.; Bretscher, L. E.; Weinhold,
F.; Raines, R. T.; Markley, J. L. J. Am. Chem. Soc. 2002, 124, 2497-
2505.
(31) O’Hagan, D.; Bilton, C.; Howard, J. A. K.; Knight, L.; Tozer, D. J. J.
Chem. Soc., Perkin Trans. 2 2000, 605-607.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6425

A R T I C L E S
Jenkins et al.
Table 2. Main-Chain Angles of Proline and Hydroxyproline Residues in Crystalline Amides and Triple Helices
amide
triple helix
Ac-4-Hyp-OMea
13
Pro (Xaa)b
Pro (Yaa)b
Pro (Xaa)c
Hyp (Yaa)c
angle
3
[ CH ]Ac-3-Hyp-OMe
φ
ψ
-50.9
145.2
-79.5
163.7
-75
164
-60
152
-69.8
162
-57.4
149.8
a
From ref 33. b From ref 35. c From ref 41.
residue would be stabilizing in the Xaa position relative to Pro
because its main-chain dihedral angles are favorable for that
position, whereas a 3-Hyp residue would be destabilizing in
the Yaa position because its dihedral angles differ significantly
from the optimal (Table 2). Although 3-Hyp is destabilizing in
both positions (Table 1), the preorganization that arises from
the gauche effect does favor a triple helix with 3-Hyp in the
Xaa rather than the Yaa position. Unlike 3-Hyp and 4-Hyp
residues, a Pro residue is not constrained by a gauche effect to
adopt a particular pucker and can, therefore, accommodate the
different φ and ψ angles of the Xaa and Yaa positions in a
triple helix (Table 2).
To discern whether hydroxyproline residues have similar
conformations in small molecule and collagen-related peptides
contexts, we compared the dihedral angles of Ac-4-Hyp-OMe
with those of a 4-Hyp residue in crystalline collagen (PDB entry
20
37
1
CAG ) by pseudorotational analysis. We found that the two
rings were essentially superimposable (data not shown), indicat-
ing that Hyp ring puckers are not dependent on context and
providing validity to our assumption that 3-Hyp would have
the same pucker in a peptide context as in amide 8.
Steric Effect on Collagen Stability. Could steric effects
contribute to the instability imparted by 3-Hyp? In a collagen
triple helix, residues in the Xaa position are more solvent-
exposed than are those in the Yaa position38 and are, therefore,
less likely to introduce unfavorable steric interactions. Indeed,
residues with bulky side chains are found only rarely in the
Yaa position of natural collagen.23 A steric clash between 3-Hyp
in the Yaa position with residues in neighboring strands (Figure
Figure 2. Model of a segment of the triple helix formed by peptide 1
top) and peptide 2 (bottom). Each 3-Hyp residue (green), the oxygen of
its hydroxyl group (red, labeled), and the steric clash in triple-helical peptide
(yellow) are indicated.
(
2
2
, bottom) is likely to diminish the stability of triple-helical 2.
This steric clash is absent in triple-helical 1 (Figure 2, top).
Effect of Hydrogen Bonds on Collagen Stability. In a
collagen triple helix, the CdO of the residue in the Xaa position
accepts a hydrogen bond from the glycine NsH of another
strand (Figure 3). The carboxyl pKa of XaaOH is a measure of
39
the ability of the Xaa residue to accept such a hydrogen bond.
The pKa values of the carboxyl groups of 3-HypOH and
-HypOH were determined by monitoring the effect of pH on
4
1
the H NMR chemical shift of their R-protons. The carboxyl
pKa of 4-HypOH was found to be 1.80, which is close to the
value of 1.82 reported previously,18 and that of ProOH is 1.95.18
Figure 3. Putative interstrand hydrogen bonds in triple-helical (3-Hyp-4-
Hyp-Gly)n.
In contrast, the pKa of 3-HypOH is 1.62, making the 3-Hyp
residue a weaker hydrogen bond acceptor than Pro. This attribute
is likely to contribute to the instability conferred by a 3-Hyp
residue in the Xaa position.
Biological Implications. 3-Hyp only decreases the Tm value
of triple-helical 2 by 3 °C compared to g5 °C for other residues
in the Xaa position of (Pro-Hyp-Gly)-based host-guest triple
helices.26 Thus, the insertion of 3-Hyp could serve to modulate
the local stability of triple helices. This idea is supported by
the 10-fold increase in 3-Hyp residues in basement-membrane
8,40
collagens as compared to fibrillar collagens. Collagen triple
helices in basement membranes interact with each other, as well
as with other biomolecules, in a more varied and complex
manner than do those in fibrils. This network of interactions
could require regions of finely tuned conformational stability,
and 3-Hyp residues could provide that tuning.
(
37) Chacko, K. K.; Swaminathan, S.; Veena, K. R. Curr. Sci. 1983, 52, 660-
6
63.
(
38) Jones, E. Y.; Miller, A. J. Mol. Biol. 1991, 218, 209-219.
39) Holmgren, S. K.; Bretscher, L. E.; Taylor, K. M.; Raines, R. T. Chem.
Biol. 1999, 6, 63-70.
(40) Kefalides, N. A. Int. ReV. Connect. Tissue Res. 1973, 6, 63-104.
(41) Berisio, R.; Vitagliano, L.; Mazzarella, L.; Zagari, A. Biopolymers 2001,
56, 8-13.
(
6426 J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003

Effect of 3-Hydroxyproline Residues on Collagen Stability
A R T I C L E S
Acknowledgment. We are grateful to J. A. Hodges, R.
Berisio, and L. Vitagliano for contributive discussions. This
work was supported by Grant AR44276 (NIH). C.L.J. was
supported by Chemistry-Biology Interface Training Grant
GM08506 (NIH). CD spectra were obtained at the University
of Wisconsin-Madison Biophysical Instrumentation Facility,
which was supported by Grants BIR-9512577 (NSF) and
RR13790 (NIH). NMR spectra were obtained at the Magnetic
Resonance Facility in the Department of Chemistry, which was
supported by Grants CHE-8813550, CHE-9629688, and CHE-
9208463 (NSF) and S10 RR08389-01 and S10 RR04981-01
(NIH).
Supporting Information Available: Data for pH titration of
3
-HypOH and 4-HypOH, thermal denaturation of triple-helical
peptides 1 and 2, and X-ray diffraction analysis of amide 8.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA034015J
J. AM. CHEM. SOC.
9
VOL. 125, NO. 21, 2003 6427