Energy-Conformational Studies of ,B-Endorphins:
Identification of Plausible Folded Conformers
GILDA LOEW, JACK COLLINS, PHILIP PAYNE, AMRIT K. JUDD, and
KEVIN H. WACKNOW
SRI International. 333 Ravenswood Avenue, Menlo Park, California 94025, USA
Abstract
{I-Endorphins are 31 amino acid endogenous opioid peptides with high receptor affinity and antinociceptive acvitity. Because of their importance as neurohormones and the significant experimental effort that has
been made to understand their struc1ure activiiy profiles, we have begun to develop procedures that could
be useful first to identify low-energy conformers of {I-endorphins and ul!imately their bioactive form. In
the initial studies reported here, we have identified plausible initial structures of the full peptide by calculating and comparing the conformational preference of all possible extended tetrapeptide fragments of
/3-endorphin starting from each of the first 28 residues. Comparisons of fragment energies suggested two
types of compact folded /3-endorphin conformers were plausible: a helix-rum-helix and an antiparallel
,8-sheet conformer. These structures, as well as an extended a-helical and /3-slrand conformer, were assembled and total geometry optimization performed using the empirical-energy-based program AMBER. The
resuhs yield an a-helical structure as the lowest energy form consistent with recently reported NMR studies of ,B-endorphin. The two more compact folded structures obtained, however, are reasonable staning
conformations for further planned molecular dynamics simulation studies and could yield competing lowenergy structures as candidates for the bioactive form of these peptides.
Introduction
{3-Endorphins, 3 1 amino acid fragments of a larger prohormone, are potent endogenous opioid peptides with high receptor affinity and antinociceptive activity [l].
That the sequence of {3-endorphins is remarkably conserved across a variety of species is an indication that more than just the amino tenninal met-enkaphalin-like portion of the peptides is important for activity. Since its discovery, about one hundred
different analogs of {3-endorphins have been synthesized in an attempt to detennine
the importance of individual residues and regions to the affinity and activity of the
peptide [1-3]. These extensive structure-activity studies include replacement, omission, and addition of residues and incorporation of disulfide bridges [l-3) . In another
approach, variations in residues in the 13-31 regions were made based on the hypothesis f4] that all that is required for opioid activity in {3-endorphins is that this region form an amphiphilic helical structure. While the resulting analogs have shown
binding affinity ranging from less than micromolar to nanomolar, and varying efficacies as analgesics, the fundamental stereoelectronic properties that determine these
variations have not yet been identified.
International Journal of Quantum Chemistry, Quantum Biology Symposium, 15, 055-066 ( 1988)
© 1988 by John Wiley & Sons, Inc.
CCC 0360-8832188/010055-12$04.00
1 of 12
Celltrion, Inc., Exhibit 1076
56
LOEW ET AL.
This lack is not surprising since the /3-endorphins belong to a class of intermediate
size bioactive peptides for which characterization of conformation profiles is most
difficult. Such peptides pose discouraging difficulties for each of the three new disciplines: x-ray crystal structure determination, NMR studies, and theoretical energyconformational studies, that in principle could be most useful for such studies. They
are, in general, difficult to crystallize, have conformational flexibilities at room temperature, and have many possible stable structures, i.e., many local conformational
minima of varying relative energies.
Recently, an NMR study of /3-endorphin has been published (5]. While conformational flexibility in water did not allow an.alysis of the spectra, dilution with methanol
did. The extensive analysis made of the NMR results, in water-methanol solution, including NOE spectra, were consistent with a predominantly a-helical structure. The
NMR data do not preclude, however, a possible tum near residues 10- 15. However,
these studies are only a first step in addressing the question of the bioactive conformers of ,8-endorphins, i.e., the form in which they bind to opioid receptors. There are
undoubtedly a number of lower-energy candidate structures and the environment at
the opioid receptor binding site is most likely devoid of bulk solvent.
The two more formidable obstacles to energy conformation studies of peptides of
the size of j3-endorphin, are the existence of large numbers of stable conformers, i.e.,
the multiminimum problem, and the practical difficulties of constructing confonners
which are good initial approximations to these minima. Human /3-endorphin, for example, has 184 nonhydrogen torsion angles which render impractical the search
strategies routinely used for smaller peptides such as nested rotations and "buildup"
from low energy conformers of single amino acids. Thus entirely different procedures
must be developed to search conformational space of these peptides for low-energy
conformers.
The most common approach used to rel.ate amino acid sequence to secondary structure in' large peptides is based on statistical analysis of x-ray structure data of
proteins (6). By contrast, in the work reported here, a novel search strategy, based on
comparisons of calculated optimized energies of sequential peptide fragments was developed to help identify plausible secondary structure regions for j3-endorphin. These
fragments were then used as guides to fold the peptide into a small number of qualitatively different conformations, i.e., into a set of tertiary structures which were then
subjected to complete geometry optimizations.
The results obtained thus far indicate that the search strategy developed could be
useful to construct initial conformers of bioactive peptides in general and to address
aspects of the protein folding problem.
Methods and Procedures
As a guide to construction of plausible conformers of 13-endorphin, 28 overlapping
elongated and derivatized tetrapeptide fragments were constructed of the form:
CH3CONH-ala-(res;-res 1+ 3)ala-CONHCH3 , i == 1- 28. Each of the 28 fragments
were constructed in 6 idealized backbone conformations corresponding to an a-helix,
a /3-strand, and four /3-turns; I, I' , II, and II'. The tetramers were extended to hexam-
2 of 12
Celltrion, Inc., Exhibit 1076
/HNDORPHlNS
57
ers by adding an alanine on each side to allow even-handed comparisons of the energies of these various secondary structures since a minimum of 6- , 5-, and
4-contiguous residues are required for favorable H-bonding in a-helical, {3-strand and
/3-tum fragments respectively. The N-terminal and C-terminal ends were appropriately derivatized by N-acetyl groups and carboxy N-methyl to more realistically
mimic the conformational behavior of the segment as part of the larger peptide chain.
Facile construction of these fragments was possible using the capabilities of an interactive structure generating program called MOLECULE, described elsewhere (7] ,
which has a library of single amino acid structures and the ability to automatically
generate peptides of a chosen backbone conformation with extended side chain torsion angles. For fragments 10, 11 , 12, and 13, which include the proline residue 13,
two optimized proline ring geometries called PROu and PROd were used. In fragment 11 , in which Prol3 is the second residue in the P-tum, only tum types 1 and ll'
are possible. For fragment 12, in which Pro 13 is the first residue in the turn, only
turn types II and II are possible. All of the initial structures generated for the
28 fragments were optimized in two steps, side chain angles only and then full torsion angle optimization using a quasi-Newton- Raphson energy minimization program caJled PEP that was developed in our laboratory. The five-term empirical
energy expression in the program called ECCEP (8) formed the basis of this optimization. It contains contributions from electrostatic, hydrogen-bonded, dispersion,
repulsion, and torsion angle potentials , and is described in detail elsewhere [8, 9).
In a buildup procedure similar to that used by Scheraga and co-workers [10), plausible folded structures of {3-endorphin were constructed by linking energy optimized
fragments corresponding to different types of secondary structures. This process was
not automatic, but involved extensive use of graphics capabilities, and distance optimization to obtain interfragment side-chain conformers which eliminated major steric
repulsions.
A set of 5 initial conformers generated for P -endorphin were energy optimized using the empirical energy expression contained in the program AMBER [l l) described
in detail elsewhere. This program allows total geometry optimization and contains a
7-term energy expression including bond angle and bond length variations in addition
to the torsion angle variation and four other types of terms similar to those in the
ECCEP potential. In these calculations, all atoms were explicitly included; with a
nonbonding atom distance cutoff of IO A. A distance-dependent dielectric, e = r ,
was used. In this program, the polar side chain residues are assumed to be ionized.
To achieve charge neutrality, salt bridges were formed between oppositely charged
nearby amino acid side chains and the remainder of the charged residues (Lys) were
neutralized by addition of negative counterions with Van Der Waal's radii of 3 A.
This procedure resulted in salt bridge formation in the antiparallel {3-sheet between
Lys29-Glu3 l and Lysl9-Glu8 and counterions at Lys9, 24, and 28. In the {3-strand,
salt bridges were formed between Glu8-Lys9; Lys29- Glu3 l , and counterions were
placed at Lys19, 24, and 28. In the a-helical structure, salt bridges were formed between Glu8- Ly9 , Lys24-Glu31 , and counterions placed at Lysl9, 28, and 25 . Finally, in the turn-helix-tum- helix-turn structure in the left-handed helix , salt
bridges were formed between Gly8-Ly9, Lys28-Glu3 1, and counterions were placed
3 of 12
Celltrion, Inc., Exhibit 1076
58
LOEW ET AL.
at Lysl9, 24, and 29; while in the right-handed helix salt bridges were fanned between Glu8- Lys9, Lys29-Glu31, and counterions placed at Lysl9, 24, and 28.
Results and Discussion
The optimized energies of the 28 extended tetrameric fragments of 13-endorphin in
the different backbone confonnations are summarized in Table I relative to that of the
alpha helical form. As shown in this table, a /3-tum conformation i~ preferred for the
enkephalin portion of the peptide. These results are consistent with our own [121 and
other previously reported energy-conformation studies of both tetra and pentapeptides
and NMR studies of metenkephalin [13, 14] in which evidence for both a gly-gly and
a gly-phe /3-bends have been reported. For fragments 10, 11, 12, and 13, which include the proline 13 residue, one proline ring geometry, Pro-D, definitely favored an
optimized alpha helical structure with some deviations from ideal backbone angles.
No turns were possible with this proline ring geometry. For the other proline ring
geometry however, /31-type turns , again deviating from ideal, were favored for
fragments 10, 11, and 12, identifying a crucial turn region in the middle of the
13-endorphin sequence which could allow a highly folded structure. A third turn region involving C-terminal residues 28-31 was also suggested by these results, though
an a-helix is somewhat favored .
In addition to identification of possible tum regions, the results suggest that the remaining contiguous region of the peptide, i.e., residues 5- 10, and residues 15-27 are
in modified alpha helical rather than /3-strand conformations. Thus the most plausible
folded conformer of ,8-endorphins, is predicted to be of the helix- tum-helix pattern
with possible additional turns at both the N-terminal and C-terminal end and an internal tum beginning at residue 11 or 12. Two highly folded structures of this type were
constructed, one with a right-handed and the other a left-handed helix, to explore the
effect of the sense of the helical portions of the conformation and energy. A totally
helical structure was also constructed as a possible variation of these compact folded
structures.
While /3-strands, with only a few exceptions, were not energetically favorable secondary structures for fragments, nevertheless, the possibility existed that a fully extended /3-strand could be a low-energy conformer of the full peptide since /j-strands
are likely to have less hindered interfragment interactions than a-helices. It is also
possible that a highly folded /3-sheet structure, which can be formed from antiparallel
/3-strands by interstrand H-bonding could be energetically favorable if the energy
gained from interstrand interactions outweighs the favorable a-helical versus /3-strand
contiguous domain energies. These possibilities were explored by including among
the initial conformers to be optimized an extended ,8-strand and an antiparallel
/3-sheet conformer with turns at residues l- 4, 11- 14, and 23-26.
The results of total geometry optimization of the five types of conformers of
/3-endorphins using the AMBER program are summarized in Table II. These optimized conformations are shown in Figure 1; their backbone conformations in Figure 2, and the corresponding backbone torsion angles are listed in Tables III-Vl. As
seen in Tables Ill and lV, fairly regular a-helical [Fig. l(a)), and /3-strand [Fig. l (b)]
structures are retained after optimization. This is not true for the more folded struc-
4 of 12
Celltrion, Inc., Exhibit 1076
59
,B·ENDORPHINS
TABJ..El. Optimized energies' of .Bh·endor·
phin fragments CH,CONH- ala[Res,Res<1+lJ)-ala-CONHCH 3 •
Fragment
I
2
3
4
5
6
7
8
9
10
II
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Tyr
Gly
Gly
Phe
Met
Thr
Ser
Glu
Lys
Seru•
SerD
GlnU<·•
GlnD
ThrUC,C
ThrD
Proue
ProD
Leu
AE11s
6..Br
3
- 4(1I')b
7
- l J(I)
17(11)
21
20
17
14
16
17
23
-I
2
-6
I
0
2
7
7
19(1)
I 1(1)
8(1')
5(1)
-
4(1)
25(Il)
- 9(1)
HIND1
- 2(1)
HIND
- 1(1)
H1ND
4(1)
21
HIND
23(I)b
Val
14
l 5(ll)
Thr
Leu
Phe
Lys
Asn
Ala
Ile
Ile
Lys
Asn
Ala
Tyr
Lys- 8
12
15
14
9(1)
13(1)
9(11)
9(1)
17(1)
16
15
16
16
14(!)
16(11)
14
12(1)
11(11)
9(11)
15
15
3
9(1)
2(ll)
14
14
7(1)
'tJ.E in kcal/mo! relative to energy of ahelix.
b( ) = optimized tum with lowest energy
of I, I', II , II'.
•Two proline ring geometries called U
and D were used in these fragments.
"only /3-tum types I and II' possible with
Prol3.
•only ,8-tum types I and II possible with
Prol3.
'Tums sterically hindered.
'The C-term.inal fragment is Lys-LysGly-Gln.
5 of 12
Celltrion, Inc., Exhibit 1076
60
LOEW ET AL.
TABLE
ll.
Optimized energies for four types /34 -endorphin structures.
Helix- tum-helix
Energy'
a.t· helix
A£ total
bond
nonbond
1-4 nb
angle
eel
1-4 eel
dihed
hbond
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
aL
J3-sheet
J3-strand
125.5
2.0
- 1.0
- 5.5
35.7
16.9
27.9
89.3
0.4
4.6
- 2.7
5.4
59.8
247.4
- 1.3
62.4
5.l
4.1
0.5
53.5
20.8
0.4
11.9
- 0.7
0.7
aR
41.7
0.4
- 16.8
1.2
9.4
4.9
24.9
17. l
0.4
- 3.9
65.J
'All energies relative to a-helix.
(a)
(b)
(d)
(e)
Figure I . Five optimized conformers of /3-endorphin showing both backbone and side
chains: (a) /3-strand, (b) a-helix , (c) antiparallel /3-shect, (d) /3r- arfJr-arJ3r. (d) f3r-aL -
f3r-a,_ -/Jr.
tures [Fig. l(c)-(e)] in which extensive interstrand side chain interaction causes large
deviations from ideal secondary structure. As seen in Table V [Fig. l(c)], in the
antiparallel ,B-sheet structure residues 1-4 form a distorted ,Bll' tum; 11 - 14, a dis-
6 of 12
Celltrion, Inc., Exhibit 1076
61
,8-ENDORPHINS
(a)
(b)
(c)
(d)
Figure 2.
f;
(e)
Five optimized backbone conformers of ,6-endorphin: (a) ,6-strand, (b) a -helix ,
(c) antiparallel /3-shect , (d) f3r-ar/3r-ap-/3r , (d) /3r-aL-/3r-0tL-.Br·
torted ,8lII tum, and residues 23-26 a distorted {31 tum. The remainder of the structure consists of antiparallel /3-strand distorted from ideal values to minimize steric
hindrance. Similar distortions from ideal secondary structure can be seen in the other
highly folded structures.
As seen in Table V, the left-handed helical structure, /3r-acf3r-aL.- f3r , has a ,8II' type turn in residues 1-4 and 11- 14, with the remainder of the structure in distorted
al helical form. The right-handed helical compact structure <f3r-arf3r-aR-/3r) has a
/3ll' tum at the N-terminal and C-terminal segments, a very broad tum around Prol3,
with the remaining segments distorted aR-helices.
As seen in Table II, the alpha helical structure is the most stable. Of the more
folded structures, our results thus far favor the f3r-aR-,8r-arf3r structure. This
structure is significantly stabilized over the left-handed helical folded structure and
the antiparallel ,B-sheet. While the interstrand interactions in the ,8-sheet has considerably lower energy relative to the {3-strand, the energy gained by interstrand
H-bonding was not sufficient to make it the more favored compact structure.
The finding that the two lowest energy conformers are predominantly a-helical is
in agreement with the recently published NMR studies of ,8-endorphin which also infers large helical regions [5]. Other features of the structures which are consistent
with those observed NMR spectra are a turn at the C-terminal end and a relatively
close distance of the Phe 18 side chain with the delta methyl group of isoleucine 23
7 of 12
Celltrion, Inc., Exhibit 1076
62
LOEW ET AL.
Ill. Backbone torsion angles for
a-Helical' 13.·endorphin structure.
TABLE
tli;
W;
-64
-54
- 42
- 43
-49
- 49
- 47
- 36
- 52
-52
-31
-65
- 56
-62
- 43
- 48
176
175
179
- 177
178
172
178
- 180
171
-171
-171
175
-175
-61
-45
-60
-59
-55
-49
-49
179
179
-177
176
178
-179
176
173
179
179
176
173
180
174
172
-176
<t>.
Tryl
Gly2
Gly3
Phe4
Met5
Thr6
Ser7
Glu8
Lys9
Ser IO
Glall
Thrl2
Prol3
Leul4
Vall5
Thrl6
Levl7
Phe18
Lysl9
Asn20
Alu21
lll22
11123
Lys24
Asn25
Ala26
Try22
Lys28
Lys29
Gly30
0
-49
-54
-64
-60
-67
-56
-55
- 64
-73
-50
-56
- so
-62
-51
- 59
-51
-44
- 47
-48
-51
-40
-41
-54
- 36
-27
- 60
-29
-87
-57
-53
-59
-60
- 57
-54
-60
-64
177
' Ideal torsion angle values for a-helix
are q,, = - 72°, t/11 = - 54°. Fairly regular
structure retained after optimization.
A) which is consistent with the upfield shift of this methyl group due to ring current perturbations. These results also support the growing experimental evidence that
small peptides can fold with regular secondary structure patterns most often favoring
those with predominantly a-helical regions [15]. While it is premature to select these
conformers as the bioactive form in which ~-endorphins bind to opioid receptors, the
result thus far also supports the hypothesis that structures with alpha helical segments
in the regions of residues 16-27 could be involved (4).
In conclusion, the results obtained thus far indicate that {3-endorphin can fold into
conformers with clearly de.fined structural motifs that can be inferred from energy op(5
8 of 12
Celltrion, Inc., Exhibit 1076
63
,8-EN DORPHINS
TABLE IV. Backbone torsion angles for optimized ,8-strand' confonner /3.-endorphin.
Try!
Gly2
Gly3
Phe4
Met.5
Thr6
Ser?
Glu8
Lys9
SertO
Glall
Thrl2
Prol3
Leul4
Vall5
Thrl6
Levi?
Phe18
Lys19
Asn20
Alu21
Ill22
lll23
Lys24
Asn25
Ala26
Try22
Lys28
Lys29
Gly30
<P1
l/J;
W;
0
-173
- 178
- 164
-157
- 155
- 174
- 157
-153
- 152
-169
-156
- 77
- 67
- 131
- 153
-166
- 152
-158
- 147
-149
- 160
-151
-142
-129
- 109
-169
- 163
- 160
- 170
174
173
171
159
154
175
156
158
139
159
155
141
177
180
178
175
168
177
172
- 170
179
-175
176
-176
173
172
172
-179
173
175
171
176
177
176
- 179
175
170
176
180
179
173
-179
65
126
154
172
153
152
143
145
148
153
150
129
126
169
152
153
158
169
'Ideal values for ,8-strand: <P1 "" - 140°,
=
tji1
135°. Fairly regular structure maintained except for a "kink" near Pro 13.
timized fragments and are not random conformations. While an a-helix appears to be
favored thus far, it is possible that additional side chain variations and some changes
in backbone structure will lead to lower energy forms of the more compact folded
structures. We are continuing to explore the conformational behavior of this large
peptide, by molecular dynamic studies using each of the optimized folded conformers
obtained here as starting configurations. The results should lead to additional variations and refinements of the lowest energy conformers among which the bioactive
form will ultimately be identified.
9 of 12
Celltrion, Inc., Exhibit 1076
64
LOEW ET AL.
Backbone torsion angles from optimized anti parallel /3-sheet' ,8. -endorphin
eon former.
T ABLE V .
Tryl
01)'2
Oly3
Phe4
Met5
Thr6
Ser7
Glu8
Lys9
Serio
Glall
Thrl2
Pro13
Leu14
Va115
Thrl6
Levi?
Phcl8
Lysl9
Asn20
Alu21
Ill22
lll23
Lys24
Asn25
Ala26
Try22
Lys28
Lys29
Oly30
c/>1
t/11
w,
0
70
- 47
- 134
- 145
- 134
- 179
- 75
- 71
- 49
- 165
- 44
36
- 73
- 54
148
164
146
53
67
86
129
96
- 46
- 39
107
135
157
114
75
67
145
JIO
179
177
156
174
173
179
172
178
158
-165
- 165
176
-179
172
173
-161
166
179
-168
172
168
-179
-175
173
168
-170
162
156
-155
-177
55
- 171
54
- 124
- 147
- 131
- 96
- 58
- 138
- 126
- 87
- 50
- 135
- 162
- 122
- 14·6
- 57
- 81
55
149
-28
40
107
109
126
98
45
' Residue 1-4 fonn a distort.e d /3fi' tum;
11-14 a distorted /3fil tum and 23-26 a distoned /31 tum. The remainder of the structures are antiparallel {3-strands distorted
from ideal values to minimize steric hindrance of side chains.
10 of 12
Celltrion, Inc., Exhibit 1076
65
P-ENDORPHINS
TABLE VI.
Backbone torsion angles for optimi1.ed Pr-a-fir-a-Pr conformers of PA ·
endorphin.
aR·helix
ai -helix
A
B
Tyrl
Gly2
Gly3
Phe4
Met5
0.00
62.12
- 139.84
-138.65
- 39.92
166.84
- 91.29
52.37
-168.99
- 48.49
- 170.91
172.59
177.69
-168.72
- 177.08
0.00
80.01
- 57.32
-126.26
66.72
14.55
- 72.17
- 46.40
137.07
107.31
175.07
177.48
170.44
- 175.35
- 159.85
Thr6
Ser7
Glu8
Lys9
SerlO
-
67.04
57.56
55.36
80.80
46.79
-
46.79
47.78
27.78
58.67
38.86
173.00
178.21
169.51
- 172.73
177.91
94.56
- 135.73
78.87
40.97
151.34
- 16.61
121.72
138.77
41.57
178.05
146.36
- 140.54
-156.89
-165.67
-170.12
Glnll
Thrl2
Prol3
Leul4
Va115
-
53.57
51.19
48.67
IOl.47
45.49
- 39.69
- 51.65
- 66.80
125.96
- 49.41
-160.56
-156.72
- 175.21
- 172.93
179.48
63.74
121.01
- 48.18
- 114.39
-1 16.06
63.04
- 53.29
- 42.25
147.14
130.43
173.64
165.57
166.03
166.78
- 138.00
Thrl6
Leul7
Phe18
Lysl9
Asn20
-
60.55
69.75
53.14
56.55
61.43
-
41.66
37.15
56.53
47.29
52.82
179.36
169.09
- 176.53
177.01
- 171.13
69.16
53.11
156.06
70.58
85.46
147.67
51.94
99.69
84.27
96.92
-156.63
-159.97
- 173.12
-145.84
-153.17
Ala2l
lle23
Ile23
Lys24
Asn25
-
64.62
63.07
66.89
51.91
49.34
-
41.16
37.08
44.58
53.07
44.32
177.68
174.04
172.39
172.41
176.36
74.52
62.85
64.95
81.09
59.48
104.92
115.34
87.58
134.17
60.47
-147.21
-167.45
-166.72
-175.08
-147.68
Ala26
Try27
Lys28
Lys29
Gly30
-
59.75
58.20
69.53
137 .88
63.34
-
45.72
44.27
39.51
21.99
43.50
273.47
- 174.93
-170.96
172.64
171.61
119.24
61.47
74.52
67.84
- 79.45
92.84
116.96
I 19.22
- 71.70
60.75
-136.60
- 162.93
-176.12
169.79
-166.89
11 of 12
Celltrion, Inc., Exhibit 1076
66
LOEW ET AL.
Acknowledgment
Support for this work from NIDA grant DA02622 is gratefully acknowledged. We
are also grateful the use of the San Diego Supercomputer Center Class VI computers
and the helpful guidance and support of the SDSC staff.
Bibliography
[I]
12)
(3)
[4]
(SJ
[6]
(7]
[8)
[9}
[!OJ
[ 11]
[12]
[13)
[ 14]
[151
C.H. Li, in Hormonal Proteins and Peptides. C . H. Li, Ed. (Academic, New York. 1980). p. 10.
P. Nicholas, R. G. Hammonds, Jr., and C . H. Li, Proc. Natl. Acad. Sci. U.S.A. 81, 3074 ( 1984).
P. Nicholas, R . G. Hammonds, Jr. , and C. H. Li, Proc. Natl. Acad. Sci. U.S.A. 79, 2191 ( 1982).
J. W. Taylor and E.T. Kaiser, Phannacol. Rev. 38, 291 (1986).
0. Lichtarge, 0. Jardetsky, and C.H. Li, Biochemistry 26, 5916 (1987).
P. Y. Chou and G.D. Fasman, J. Mol. Biol. 115, 135 (1977).
J. T. Egan, J. Hart, S. K. Burt, and R. D. MacElroy, Comput. Graphics 6, 1977 (1982).
F. A . Momany, R. F. McGuire, A. W. Burgess, and H . A . Scheraga, J . Phys. Chem. 79, 7381
(1975).
G. Nemethy, M. S. Pottle, and H. A. Scheraga, J. Phys. Chem. 87, 1883 (1983).
M. Vasquez and H. A. Scheraga, Biopolymers 24, 1437 (1985).
J. S. Weiner, P. A. Kollman , D. A. Case, U. C . Singh, C. Ghio, G. Alagona, S. Profeta, and P. J.
Weiner, J. Am. Chem. Soc. 106, 765 (1984).
G . Loew, H. Hashimoto, L. Williamson, S. Burt, and W. Ande.rson , Molec. Phann. 22, 467 (1982).
G. Gupta, FEBS Letters 198, 245 (1986).
Z . Li and H. A. Scheraga, Proc. Natl. Acad. Sci. U.S.A., 84, 6611 (1987).
S. Marqusee and R. L. Baldwin. "The Prot.ein Folding Problem," Abstracts, AAAS Meeting, Boston
MA , February 1988.
Received May 3, 1988.
12 of 12
Celltrion, Inc., Exhibit 1076