1: \documentstyle[preprint,aps,prl,psfig]{revtex}
2:
3: \begin{document}
4: \topmargin -0.05cm
5: \title{Energetic Components of Cooperative Protein Folding }
6: \author{H\"useyin KAYA and Hue Sun CHAN}
7: \address{Department of Biochemistry, and \\
8: Department of Medical Genetics \& Microbiology \\
9: Faculty of Medicine, University of Toronto \\
10: Toronto, Ontario M5S 1A8, Canada}
11:
12:
13: \def\kB{{k_{\scriptscriptstyle B}}}
14: \def\gD{{g_{\scriptscriptstyle{\rm D}}}}
15: \def\gcD{{{\rm g}_{\scriptscriptstyle{\rm D}}}}
16: \def\HD{{H_{\rm D}}}
17: \def\sH{{\sigma_H}}
18: \def\eHDkT{{{\rm e}^{-H_{\rm D}/(\kB T)}}}
19: \def\eHkT{{{\rm e}^{-H/(\kB T)}}}
20: \def\DHv{{\Delta H_{\rm vH}}}
21: \def\DHc{{\Delta H_{\rm cal}}}
22: \def\DHvDHc{{\Delta H_{\rm vH}/\Delta H_{\rm cal}}}
23:
24: \maketitle
25:
26: \begin{abstract}
27: A new lattice protein model with a four-helix bundle ground
28: state is analyzed by a parameter-space Monte Carlo histogram
29: technique to evaluate the effects of an extensive variety of
30: model potentials on folding thermodynamics. Cooperative
31: helical formation and contact energies based on a 5-letter
32: alphabet are found to be insufficient to satisfy calorimetric
33: and other experimental criteria for two-state folding.
34: Such proteinlike behaviors are predicted, however, by models
35: with polypeptide-like local conformational restrictions and
36: environment-dependent hydrogen bonding-like interactions.
37:
38: \vskip 1cm
39: \noindent
40: {\underline {PACS Numbers}:} 87.15.Aa, 87.15.Cc, 87.15.He, 87.15.By
41: \end{abstract}
42: %===================================================================
43: \eject
44:
45: %\begin{multicols}{2}
46:
47:
48: Proteins are complex systems. Insight into their behaviors has
49: been gained by simplified models of generic proteins [1-6].
50: To serve as stepping stones towards an elucidation of the physics of
51: real proteins, however, these models must be subjected to rigorous
52: evaluations against experiments. Recently, we found that
53: a number of popular lattice protein models with 2, 3, and 20 residue
54: types and pairwise additive contact energies do not
55: satisfy the experimental criteria for two-state thermodynamic
56: cooperativity
57: which, among other conditions, requires a protein's van't Hoff to
58: calorimetric
59: enthalpy ratio $\Delta H_{\rm vH}/\Delta H_{\rm cal}\approx 1$ [7,8].
60: While certain G\=o models are cooperative,
61: they do not address the physical nature of protein interactions since
62: their potentials are {\it teleological} [8] in the sense that
63: the energetic favorability of the native conformation as a whole is
64: presupposed.
65:
66: It has since been proposed that a cooperative interplay between local
67: conformational preferences and nonlocal interactions could give rise to
68: proteinlike thermodynamics [7].
69: To evaluate the viability of this scenario, we introduce in this Letter a
70: 55-mer chain model with a four-helix bundle ground state (Fig.~1) that
71: shares common features with the corresponding protein motif [9].
72: The cubic-lattice helices are right-handed
73: (called type (i) in Ref.~[2]), as are most $\alpha$-helices. The model
74: has the following energetic components. (In the analysis below all
75: energies are dimensionless and temperature independent.)
76:
77: {\it Contact energies.}
78: A reduced 5-letter alphabet (Fig.~1) identical to the one
79: recently optimized [5] is used for nearest-neighbor interactions,
80: with energy parameters from Table~III of Kolinski et al. [10].
81: While these energies are proteinlike to some extent because
82: hydrophobic groups are placed in the native core (Fig.~1), they do not
83: represent the full interactions between amino acid residues [10].
84: Here they are adopted to capture
85: heterogeneous aspects of intraprotein interactions [11].
86:
87: {\it Unfavorable local conformations.}
88: Two types of local non-proteinlike bond geometries
89: are discouraged: the initiation of a left-handed (lh) helix (Fig.~2a),
90: and one end of a helix taking a sharp turn (st) to fold back
91: onto itself (Fig.~2b) are assigned unfavorable ($>0$) energies
92: to take into account that in real
93: proteins left-handed $\alpha$-helices are sterically disfavored, and that
94: polypeptides are stiffer than a fully flexible chain [12].
95:
96: {\it Environment-dependent hydrogen bonding.}
97: The favorable many-body interactions (${\cal E}_{\rm Hb}<0$) in Fig.~2c,d
98: are introduced to explore an idea, suggested by experiments [13], that the
99: collective strength of hydrogen bonds is stronger when they are buried in
100: the core of a protein than when they are exposed to water.
101: Hence we focus mainly on $b>1$ cases below. Analogous
102: interactions have been used before [7,14].
103:
104: {\it Cooperative helical propagation.}
105: An extra favorable energy is assigned to every two consecutive helical
106: turns (Fig.~2e) to encourage helix elongation. Such an effect
107: may arise from dipole-dipole interactions between amide groups
108: in real $\alpha$-helices [12].
109:
110: The total energy of a conformation from these contributions is
111: $$
112: E= E_{\rm contact} + \gamma_{\rm lh}N_{\rm lh}
113: + \gamma_{\rm st}N_{\rm st} + {\cal E}_{\rm Hb} N^{(6)}_{\rm Hb}
114: + b{\cal E}_{\rm Hb} N^{(8)}_{\rm Hb} + {\cal E}_{\rm Helix}N_{\rm Helix}
115: \; ,
116: \eqno(1)
117: $$
118: where $E_{\rm contact}$ is the sum of 5-letter
119: contact energies, $N_{\rm lh}$ and $N_{\rm st}$
120: are respectively the numbers of all incidences of Fig.~2a,b.
121: In the present analysis, $N_{\rm Helix}$ only counts those helices
122: (Fig.~2e) that are parts of the four helices in the ground-state
123: conformations (monomers 1--12, 15--26, 30--41, and 44--55; see Fig.~1);
124: and the hydrogen bonding pairs counted by $N^{(6)}_{\rm Hb}$
125: (Fig.~2c) and $N^{(8)}_{\rm Hb}$ (Fig.~2d) are the 36 $(i,i+3)$ or
126: $(i,i+5)$ contacting monomer pairs in the native helices.
127: A first-principle treatment would have assigned
128: hydrogen bonds and helical segments in a manner that do not require
129: knowledge of the native structure. However, progress can nonetheless
130: be made by the approach taken here,
131: which seeks, as a first step in the inquiry, to ascertain
132: the consequence of presupposing {\it local}
133: native preferences, a presupposition that is notably less dependent on
134: {\it
135: a priori} knowledge than the assumption of {\it global} native preference
136: in G\=o potentials.
137:
138: To efficiently determine how thermodynamic properties vary with the model
139: energetic parameters, we use a generalization [3,15]
140: of the standard Metropolis Monte Carlo histogram technique [8,16]
141: to eliminate the need to perform separate direct simulations
142: for every parameter set of interest.
143: Typical simulations are carried out with ${\cal E}_{\rm Hb}$ $=$
144: ${\cal E}_{\rm Helix}=0$ at a certain temperature $T^\prime$, during which
145: numbers $P$ of sampled conformations are binned into a
146: multiple-dimensional
147: array (histogram) according to
148: $(E^\prime,N^{(6)}_{\rm Hb},N^{(8)}_{\rm Hb},N_{\rm Helix})$, where
149: $E^\prime$ is the energy of the conformation. Thus the density
150: of states of the simulated system
151: $g(E^\prime,N^{(6)}_{\rm Hb},N^{(8)}_{\rm Hb},N_{\rm Helix})$
152: $=$ $P(E^\prime,N^{(6)}_{\rm Hb},N^{(8)}_{\rm Hb},N_{\rm Helix})
153: e^{E^\prime/\kB T^\prime}$, where $\kB$ is Boltzmann's constant.
154: It follows that the partition function for any ${\cal E}_{\rm Hb}$,
155: ${\cal E}_{\rm Helix}$ at any temperature $T$ is given by $\sum
156: g(E^\prime,N^{(6)}_{\rm Hb},N^{(8)}_{\rm Hb},N_{\rm Helix})e^{-E/\kB T}$,
157: where the summation is over
158: $E^\prime$, $N^{(6)}_{\rm Hb}$, $N^{(8)}_{\rm Hb}$, and $N_{\rm Helix}$,
159: and
160: $E=E^\prime + {\cal E}_{\rm Hb}(N^{(6)}_{\rm Hb}
161: + bN^{(8)}_{\rm Hb}) + {\cal E}_{\rm Helix}N_{\rm Helix}$.
162: A similar procedure is used to study the effects of
163: $\gamma_{\rm lh}$ and $\gamma_{\rm st}$.
164:
165: Each Monte Carlo run consists of $1.53\times 10^9$ attempted
166: moves, the first $3\times 10^7$ of which are
167: excluded from data acquisition. To generate a multiple-dimensional
168: histogram, a total of 20 runs at 10 different simulation $T^\prime$s
169: (around the transition region) are performed, with two different random
170: initial conformations for each $T^\prime$. Values of $\kappa_2$ estimated
171: from different $T^\prime$s agree well, with
172: standard deviation $\approx 4\%$. We have conducted extensive comparisons
173: with direct simulations to validate the method [17].
174: No energy lower than that of the structure in Fig.~1 has been encountered.
175:
176: As in experimental calorimetry [18], we characterize the
177: thermodynamic cooperativity of a model protein by its specific heat
178: capacity
179: $C_P$ and $\Delta H_{\rm vH}/\Delta H_{\rm cal}$.
180: When baseline subtraction is not applied to the $C_P$ function,
181: $\Delta H_{\rm vH}/\Delta H_{\rm cal}$ may be equated to
182: $\kappa_2\equiv$ $2T_{\rm max}\sqrt{\kB C_P(T_{\rm max})}/\Delta H_{\rm
183: cal}$,
184: where $\Delta H_{\rm cal}=\int_0^\infty dT C_P(T)$ is the calorimetric
185: enthalpy, and $C_P(T)$ is maximum at $T=T_{\rm max}$ [8].
186: Baseline subtractions amount to
187: defining a multi-conformation native state
188: and ignoring a part of the enthalpic variation in the denatured ensemble.
189: This effectively reduces both the calorimetric enthalpy and the maximum
190: heat
191: capacity value, resulting in a modified enthalpy ratio
192: $\kappa_2^{({\rm s})}$ ($>\kappa_2$) [8].
193:
194: Our main findings are summarized in Figs.~3 and 4.
195: The apparent $\Delta H_{\rm vH}/\Delta H_{\rm cal}$ $=\kappa_2^{({\rm
196: s})}$
197: after empirical baseline subtractions can often be close or equal to unity
198: even when $\kappa_2$ is low (Fig.~4).
199: However, as is the case for a 3-letter model, a large discrepancy
200: between $\kappa_2$ and $\kappa_2^{({\rm s})}$ is often symptomatic of
201: non-proteinlike significant post-denaturational chain expansion at
202: $T\gg T_{\rm max}$ [8]. Therefore, for model evaluation, proteinlike
203: thermodynamic cooperativity requires {\it both} a small
204: $\kappa_2^{({\rm s})}$ $-$ $\kappa_2$ {\it and}
205: $\kappa_2^{({\rm s})}\approx 1$.
206:
207: Fig. 3 compares three models by this criterion. The least cooperative
208: is a flexible chain model with only pairwise additive contact energies.
209: A second model with polypeptide-like local sterics has
210: slightly enhanced cooperativity because populations of nonnative
211: conformations with
212: non-proteinlike local geometries that are hitherto favorable in a fully
213: flexible chain model are reduced. However, both of these models
214: are not proteinlike because of their significant post-denaturational chain
215: expansions, as is evident from their thick denatured $C_P$ ``tails'' at
216: $T\gg T_{\rm max}$ (Fig.~3), which account for these models' relatively
217: large differences between
218: $\kappa_2^{({\rm s})}$ and $\kappa_2$ [8]. On the other hand, the
219: model that also incorporates environment-dependent hydrogen
220: bonding has more proteinlike thermodynamics:
221: Its native $C_P$ tail at $T\ll T_{\rm max}$
222: is thin, with $C_P$ values lower on average than that
223: of its denatured tail (Fig.~3).
224: This implies that its native conformational diversity
225: is limited, thus conforming better to NMR
226: data [19] than a previously considered 20-letter model [8].
227: As for real proteins [20], its average radius of gyration undergoes a
228: sharp change around $T_{\rm max}$, but has no appreciable
229: post-denaturational
230: increase (data not shown).
231:
232: Fig.~4 surveys a range of energetic parameters.
233: Remarkably, local helical cooperativity has only a small
234: impact on overall folding cooperativity, and proteinlike
235: thermodynamics is possible at ${\cal E}_{\rm Helix}=0$.
236: While ${\cal E}_{\rm Helix}<0$ stabilizes the native state, it also
237: stabilizes
238: denatured conformations with partially intact native helices.
239: Therefore, its effect on calorimetric cooperativity is not substantial
240: because it cannot widen the average enthalpy difference between native
241: and denatured states significantly [7,8]. In contrast, calorimetric
242: cooperativity increases sharply with more negative ${\cal E}_{\rm Hb}$.
243: For $0.5\le b\le 1.5$, this effect is not very sensitive to
244: $b$. For example, ${\cal E}_{\rm Hb}=-0.5$, ${\cal E}_{\rm Helix}=0$,
245: $b=0.5$ (and $\gamma_{\rm lh}=6.0$, $\gamma_{\rm st}=5.0$) lead to
246: $\kappa_2$, $\kappa_2^{({\rm s})}=$ $0.85$, $0.99$, which are only
247: slightly lower than the $\kappa_2$, $\kappa_2^{({\rm s})}$
248: values of $0.90$, $1.0$ for the same ${\cal E}_{\rm Hb}$ and
249: ${\cal E}_{\rm Helix}$ in Fig.~4 for $b=1.5$.
250:
251: In addition, we found that a lattice version of
252: helix capping [21] has a slight attenuating effect on cooperativity [17].
253: Consistent with experiments [22], in our model, tertiary interactions
254: are essential in stablizing helices in native structures.
255: When folded independently, the two 12-mer sequences
256: for the native helices at 1--12 and 15--26 are much less stable
257: ($T_{\rm max}\sim 0.3$), and their thermal transitions are not
258: calorimetrically cooperative.
259:
260: Figs.~3 and 4 suggest that a cooperative interplay between local
261: conformational preferences and nonlocal contact interactions, as
262: exemplified
263: by the environment-dependent hydrogen bonding in the model, is a viable
264: mechanism for proteinlike thermodynamics;
265: and that the required cooperative effect ($b{\cal E}_{\rm Hb}=-0.8$)
266: need not be exceedingly strong relative to the pairwise contact energies
267: (average magnitude $=0.66$). This observation is consistent with a
268: previous
269: high-coordination lattice model study [23], though the latter did not
270: address the calorimetric criterion.
271: The present approach did not consider non-native
272: hydrogen bonding. Cooperativity in the present model would be
273: reduced if such non-native conformations are favored.
274: Further investigations using continuum models with
275: polypeptide chain geometry are necessary to ascertain
276: whether real proteins can have substantial number of
277: non-native hydrogen bonds.
278: While the present model
279: should be regarded as tentative because it
280: relies on local native information, its proteinlike features
281: do not follow trivially from part of its interactions' native-centric
282: nature {\it per se}.
283: Important physical principles have emerged from our analysis
284: because not all native-centric interaction schemes can bring about
285: comparable enhancements in thermodynamic cooperativity: ({\it i}) The fact
286: that
287: ${\cal E}_{\rm Helix}<0$ is neither necessary nor sufficient
288: for proteinlike thermodynamics suggests that multi-body interactions that
289: favors local native conformation irrespective of tertiary packing cannot
290: account for calorimetric cooperativity.
291: ({\it ii}) A G\=o model for the ground-state conformations in Fig.~1,
292: which exploits both local and nonlocal native information but is based
293: exclusively on pairwise additive interactions, has $\kappa_2=0.73$ and is
294: thus less cooperative than the model with $\kappa_2=0.91$ in Fig.~3.
295: ({\it iii}) Proteinlike steric effects contribute
296: to cooperativity. If non-proteinlike local
297: conformations were not disfavored in the latter model
298: (i.e., if $\gamma_{\rm lh}=\gamma_{\rm st}=0$),
299: $\kappa_2$ would be reduced to $0.85$.
300:
301: We have thus mapped out a general investigative strategy and established
302: the
303: viability of a folding scenario. It should be emphasized, however,
304: that satisfying the
305: requirements for thermodynamic cooperativity is clearly necessary but not
306: sufficient for the validity of a scenario's underlying physical
307: mechanisms.
308: Whether hydrogen bonding is favorable to native stability remains
309: controversial [13,24]. The present choice of ${\cal E}_{\rm Hb}<0$
310: is motivated by experiments [13]. But there have been theoretical
311: suggestions
312: that hydrogen bonding disfavors the folded state of a protein [24].
313: In the present modeling framework, that would be detrimental to
314: calorimetric
315: cooperativity as it corresponds to ${\cal E}_{\rm Hb}>0$, leading to
316: $\kappa_2$s even smaller than the ${\cal E}_{\rm Hb}=0$ case (Fig.~3).
317: For instance, when ${\cal E}_{\rm Hb}=+0.1$ and ${\cal E}_{\rm Helix}=0$,
318: (and $\gamma_{\rm lh}=6.0$, $\gamma_{\rm st}=5.0$),
319: $\kappa_2$, $\kappa_2^{({\rm s})}=$ $0.51$, $0.80$ for $b=0.5$ and
320: $\kappa_2$, $\kappa_2^{({\rm s})}=$
321: $0.45$, $0.70$ for $b=1.5$, respectively. If that turns out to be the
322: case,
323: there would be added impetus to extend the present method
324: to ascertain the role of other mechanisms such as sidechain packing
325: [4,8,25] in
326: protein calorimetric two-state cooperativity.
327:
328: %$\null$
329:
330: %\noindent
331: This work was supported by Medical Research Council of Canada
332: grant MT-15323.
333:
334:
335: %===========================================================================
336:
337: \par\vfill\eject
338:
339: %\noindent
340: %{\large\bf References}
341:
342: %\kern -1.5cm
343:
344: %\footnotesize
345: \begin{thebibliography}{999}
346:
347: \bibitem{1}
348: J. D. Bryngelson {\it et al.},
349: Proteins: Struct. Funct. Genet. {\bf 21}, 167 (1995);
350: K. A. Dill {\it et al.}, Protein Sci. {\bf 4}, 561 (1995);
351: H. Li {\it et al.}, Science {\bf 273}, 666 (1996); V.S. Pande,
352: A. Yu. Grosberg, and T. Tanaka, Biophys. J. {\bf 73}, 3192 (1997);
353: E.I. Shakhnovich, Curr. Opin. Struct. Biol. {\bf 7}, 29 (1997);
354: H. S. Chan and K. A. Dill, Proteins: Struct. Funct. Genet. {\bf 30}, 2
355: (1998);
356: C. Micheletti {\it et al.}, Phys. Rev. Lett. {\bf 82}, 3372 (1999).
357:
358: \bibitem{2}
359: H. S. Chan and K. A. Dill, J. Chem. Phys. {\bf 92}, 3118 (1990); {\bf
360: 107},
361: 10353 (1997).
362:
363: \bibitem{3}
364: N.D. Socci and J.N. Onuchic, J. Chem. Phys. {\bf 103}, 4732 (1995).
365:
366: \bibitem{4}
367: D. K. Klimov, and D. Thirumalai, Fold. Des. {\bf 3}, 127 (1998).
368:
369: \bibitem{5}
370: J. Wang and W. Wang, Nature Struct. Biol. {\bf 6}, 1033 (1999).
371:
372: \bibitem{6}
373: H. S. Chan, Nature (London) {\bf 392}, 761 (1998).
374:
375: \bibitem{7}
376: H. S. Chan, Proteins: Struct. Funct. Genet. {\bf 40}, 543 (2000).
377:
378: \bibitem{8}
379: H. Kaya and H. S. Chan, Proteins: Struct. Funct. Genet. {\bf 40}, 637
380: (2000).
381:
382: \bibitem{9}
383: C. Branden and J. Tooze, {\it Introduction to Protein Structure}
384: (Garland, New York, 1991).
385:
386: \bibitem{10}
387: A. Kolinski, A. Godzik, and J. Skolnick, J. Chem. Phys. {\bf 98}, 7420
388: (1993).
389:
390: \bibitem{11}
391: H. S. Chan and K. A. Dill, Proteins: Struct. Funct. Genet. {\bf 24}, 335
392: (1996); P. G. Wolynes, Nature Struct. Biol. {\bf 4}, 871 (1997);
393: H. S. Chan, Nature Struct. Biol. {\bf 6}, 994 (1999).
394:
395: \bibitem{12}
396: C. R. Cantor and P. R. Schimmel,
397: {\it Biophysical Chemistry} (Freeman, San Francisco, 1980).
398:
399: \bibitem{13}
400: S. M. Habermann and K. P. Murphy, Protein Sci. {\bf 5}, 1229 (1996);
401: J. K. Myers and C. N. Pace, Biophys. J. {\bf 71}, 2033 (1996).
402:
403: \bibitem{14}
404: S. Takada, Z. Luthey-Schulten, and P. G. Wolynes,
405: J. Chem. Phys. {\bf 110}, 11616 (1999).
406:
407: \bibitem{15}
408: J.-E. Shea {\it et al.}, J. Chem. Phys. {\bf 109}, 2895 (1998).
409:
410: \bibitem{16}
411: A. M. Ferrenberg and R. H. Swendsen, Phys. Rev. Lett. {\bf 63}, 1195
412: (1989).
413:
414: \bibitem{17}
415: H. Kaya and H. S. Chan, in preparation.
416:
417: \bibitem{18}
418: P. L. Privalov and S. A. Potekhin, Meth. Enzymol. {\bf 131}, 4 (1986);
419: E. Freire, in {\it Protein Stability and Folding: Theory and Practice},
420: edited by B. A. Shirley, Methods in Molecular Biology, Vol. 40
421: (Humana, Totowa, New Jersey, 1995).
422:
423: \bibitem{19}
424: D. Yang {\it et al.}, J. Mol. Biol. {\bf 272}, 790 (1997).
425:
426: \bibitem{20}
427: T. R. Sosnick and J. Trewhella, Biochemistry {\bf 31}, 8329 (1992);
428: Y. Hagihara {\it et al.}, Fold. Des. {\bf 3}, 195 (1998).
429:
430: \bibitem{21}
431: L. G. Presta and G. D. Rose, Science {\bf 240}, 1632 (1988);
432: R. Aurora, R. Srinivasan, and G. D. Rose, Science {\bf 264}, 1126 (1994).
433:
434: \bibitem{22}
435: K. A. Dill, Biochemistry {\bf 29}, 7133 (1990).
436:
437: \bibitem{23}
438: A. Kolinski, W. Galazka, and J. Skolnick,
439: Proteins: Struct. Funct. Genet. {\bf 26}, 271 (1996).
440:
441: \bibitem{24}
442: B. Honig and A.-S. Yang, Adv. Protein Chem. {\bf 46}, 27 (1995);
443: N. Ben-Tal {\it et al.}, J. Phys. Chem. B {\bf 101}, 450 (1997).
444:
445: \bibitem{25}
446: E. I. Shakhnovich and A. V. Finkelstein, Biopolymers {\bf 28}, 1667
447: (1989);
448: S. Bromberg and K. A. Dill, Protein Sci. {\bf 3}, 997 (1994).
449:
450: \end{thebibliography}
451:
452: %\end{multicols}
453:
454:
455: %\end{document}
456: \par\vfill\eject
457:
458: %===================================================================
459:
460: \noindent
461: {\large\bf Figure Captions}\\
462:
463: {\bf Fig.~1}. Ground state conformations of the model.
464: Black beads denote nominally hydrophobic monomers.
465: The numbers label selected sequence positions.
466: Each of the 3 short loops
467: at positions (13, 14), (27, 28, 29), and (42, 43) has two
468: iso-energetic local conformations. Thus the ground state has 8
469: conformations, one of which is shown.
470: \\
471:
472: {\bf Fig.~2}. Energetic components of the model.
473: (a,b) Each contact marked by a double arrow is assigned an energy (as
474: shown).
475: The dotted lines in (b) depict an alternate path of the chain from
476: monomer $i-3$ to $i$ that has one instead of two unfavorable contacts.
477: Similarly, an energy is associated with each buried hydrogen bond (c,d)
478: and
479: each occurrence of two consecutive turns (layers) of a right-handed
480: lattice
481: helix (e). In (c,d), hydrogen bonds are represented as ladders linking
482: pairs
483: of encircled monomers. Their burial requires occupation of at least 6 of
484: their
485: neighbor sites. The energy of a completely buried bond with 8 occupied
486: neighbor sites (d) can be stronger ($b>1$) or weaker ($b<1$) than
487: a partially buried bond with 6 or 7 occupied neighbor sites. The
488: latter two cases have the same energy ${\cal E}_{\rm Hb}$
489: and are accounted for collectively by $N^{(6)}_{\rm Hb}$ in Eq.~(1).
490: \\
491:
492: {\bf Fig.~3}. Specific heat capacity functions.
493: In all three cases shown, ${\cal E}_{\rm Helix}=0$.
494: $T_{\rm max}$s are marked by vertical lines.
495: >From left to right, the first model has only the 5-letter pairwise
496: contact
497: energies ($\gamma_{\rm lh}=\gamma_{\rm st}={\cal E}_{\rm Hb}=0$).
498: In addition to these, the second model incorporates the
499: repulsive interactions in Fig.~2a,b, with $\gamma_{\rm lh}=6.0$ and
500: $\gamma_{\rm st}=5.0$. The third model further adds
501: the favorable hydrogen-bonding
502: energies in Fig.~2c,d, with ${\cal E}_{\rm Hb}=-0.53$,
503: $b{\cal E}_{\rm Hb}=-0.8$, and $b=1.5$.
504: $\kappa_2=0.55$, $0.62$, and $0.91$, and
505: after the plotted inclined baselines are subtracted,
506: $\kappa_2^{({\rm s})}=0.93$, $0.93$, and $1.01$, respectively.
507: \\
508:
509: {\bf Fig.~4}.
510: Calorimetric cooperativity as a function of local helical cooperativity
511: and environment-dependent hydrogen bonding strength, with $b=1.5$,
512: $\gamma_{\rm lh}=6.0$ and $\gamma_{\rm st}=5.0$.
513: $\kappa_2^{({\rm s})}$ and $\kappa_2$ are
514: given by the upper and lower surfaces, respectively.
515: The black dots connected by a vertical bar mark the
516: parameters for the most cooperative case in Fig.~3.
517: $\kappa_2^{({\rm s})}$ are calculated using empirical
518: baselines [8] constructed as tangents of the $C_P$ function at
519: $C_P^{\prime\prime}(T)/ C_P^{\prime\prime}(T_{\rm max})=-0.001$,
520: where $C_P^{\prime\prime}\equiv d^2 C_P/dT^2$ (see Fig.~3).
521: \\
522:
523: \end{document}
524:
525: