cond-mat0211029/a.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %                                                                     %
3: %  A LaTeX file with some definitions at the top.                     %
4: %                                                                     %
5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6: %
7: %                                                       MTU-PHY-HA-02/4
8: %                                                       Oct, 25    2002
9: 
10: %% \documentstyle[aps,prb,preprint,floats]{revtex}
11: %\documentstyle[aps,prl,twocolumn,graphicx,floats]{revtex}
12: %\options [aps,prl,pre,preprint,twocolumn,floats,graphicx]
13: 
14: \documentstyle[aps,prb,preprint,graphicx,floats]{revtex}
15: 
16: 
17: \tightenlines
18: 
19: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
20: %    ONLY FOR ALVES PRINTER  %
21: %\voffset=1.6cm 
22: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23: 
24: 
25: \begin{document}
26: 
27: % \draft command makes pacs numbers print
28: \draft
29: % repeat the \author\address pair as needed
30: %{\wideabs{ 
31: \title{Solution effects and the order of the helix-coil transition 
32:          in polyalanine} 
33: \author{Yong Peng  and Ulrich H.E. Hansmann } 
34: \address{Department of Physics, Michigan Technological University,
35:          Houghton, MI 49931-1291, USA}
36: \author{ Nelson A. Alves}
37: \address{Departamento de F\'{\i}sica e Matem\'atica, FFCLRP
38:      Universidade de S\~ao Paulo. Av. Bandeirantes 3900. 
39:      CEP 014040-901 \, Ribeir\~ao Preto, SP, Brazil}
40: \date{\today}
41: \maketitle
42: \begin{abstract}
43: We  study helix-coil transitions  in an all-atom model of
44: polyalanine. Molecules of up to length 30 residues are investigated 
45: by multicanonical simulations. Results from two  implicit solvent 
46: models are compared with each other and with that from simulations in  
47: gas phase. While  the helix-coil transition  is in all three models 
48: a true thermodynamic phase transition,  we find that its  strength 
49: is  reduced by the protein-solvent interaction term. The order of
50: the helix-coil transition depends on the details of the solvation 
51: term.   
52: \end{abstract}
53: %\pacs{}
54: %}}
55: 
56: 
57: %%\begin{multicols}{2}
58: 
59: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
60: \section{Introduction}
61: A key step in the folding of a protein  is the formation of 
62: secondary structure elements   such as $\alpha$-helices or
63:  $\beta$-sheets. In the case of $\alpha$-helices this process
64: resembles crystalization or melting and has been extensively
65: studied \cite{Poland}. In Refs.~\onlinecite{HO98c,AH99b,AH00b} evidence
66: was presented that this helix-coil transition is for 
67: polyalanine a true thermodynamic phase transition.
68: The latter result was obtained from gas-phase simulations  
69: where interactions between all atoms in the
70: molecule were taken into account. While  available data from 
71: gas-phase experiments \cite{Jarrold} support the numerical
72: results of these simulations, their relevance for
73: the biologically more important case of solvated molecules
74: needs to be established. First attempts in this direction can be found  
75: in Ref.~\onlinecite{MO1},  and in more detail in Ref.~\onlinecite{PH01h}
76: where  various solvent models are compared. 
77: These previous investigations are restricted to homopolymers of
78: chain lengths 10.  However, a detailed study on
79: how the thermodynamic  characteristics of the helix-coil 
80: transition change with the added solvent term  requires to go 
81: to larger chains and to probe the size dependence of that transition.
82: For this purpose, we have
83: performed multicanonical simulations of polyalanine chains
84: of length up to 30 residues. The protein-water interaction
85: is included by either a term that is proportional to the 
86: solvent accessible surface area  \cite{oons} or by a
87: distance-dependent permittivity  \cite{lavarty}. 
88:  Our results are compared with that of gas-phase simulations. We
89: find that the order of the helix-coil transition depends 
90: strongly on the details of the solvation term.
91:  
92: 
93: 
94: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
95: 
96: \section{Methods}
97: \noindent
98: Our investigation of the helix-coil transition for polyalanine is
99: based on a detailed, all-atom representation of that homopolymer. 
100: The interactions between the atoms are 
101: described by a standard force field, ECEPP/2  ({\bf E}mpirical 
102: {\bf C}onformational {\bf E}nergy {\bf P}rogram for {\bf P}eptides,
103: version 2) \cite{EC} as implemented in the program package SMMP 
104: ({\bf S}imple {\bf M}olecular {\bf M}echanics for {\bf P}roteins)
105:  \cite{SMMP}:
106: \begin{eqnarray}
107: E_{ECEPP/2} & = & E_{C} + E_{LJ} + E_{HB} + E_{tor},\\
108: E_{C}  & = & \sum_{(i,j)} \frac{332q_i q_j}{\varepsilon r_{ij}},\\
109: E_{LJ} & = & \sum_{(i,j)} \left( \frac{A_{ij}}{r^{12}_{ij}}
110:                                 - \frac{B_{ij}}{r^6_{ij}} \right),\\
111: E_{HB}  & = & \sum_{(i,j)} \left( \frac{C_{ij}}{r^{12}_{ij}}
112:                                 - \frac{D_{ij}}{r^{10}_{ij}} \right),\\
113: E_{tor}& = & \sum_l U_l \left( 1 \pm \cos (n_l \chi_l ) \right).
114: \label{ECEPP/2}
115: \end{eqnarray}
116: Here, $r_{ij}$ (in \AA) is the distance between the atoms $i$ and $j$, and
117: $\chi_l$ is the $l$-th torsion angle.
118: 
119: The interactions between our homo-oligomer and water are approximated by
120: means of two implicit water models. In the first model 
121: one assumes that the free energy difference between  atomic groups 
122: immersed in the protein interior and groups exposed to  water   
123: is proportional to the solvent accessible surface area:
124: \begin{equation}
125: E_{solv}=\sum_i\sigma_iA_i,
126: \end{equation}
127: where $E_{solv}$ is the solvation energy, $A_i$ is the 
128: solvent accessible area (which depends on the configuration)
129:  of the $i$-th atom,  
130: and $\sigma_i$ is the  solvation parameter for the atom $i$.
131: For the present study we have chosen the  parameter set of 
132: Ref.~\onlinecite{oons}
133: that is often used in conjunction with the ECEPP force field. 
134: In the following, we will refer to this model as ASA (solvent
135: {{\bf A}ccessible {\bf S}urface {\bf A}rea).
136: Protein-water interactions are approximated differently 
137: in our second implicit water model to which we will refer as
138: DDE ({\bf D}istance {\bf D}ependent {\bf E}psilon). Here,  
139: a  distance  dependent  electrostatic permittivity  \cite{hingerty} 
140: \begin{equation}
141: \label{eps}
142: \varepsilon(r) = D-\frac{D-2}{2}[(sr)^2+2sr+2]e^{-sr} \, ,
143: \end{equation}
144: is introduced to model electrostatic interaction between the 
145: protein atoms in the presence of water.  The parameters $D$ and $s$  
146: are chosen such  that for large distances the permittivity takes 
147: the value $\varepsilon \approx 80$ of  bulk water, and 
148:  $\varepsilon=2$ for short distances (protein interior space). 
149: 
150: In detailed models of biological macromolecules the various
151: competing interactions lead to an energy landscape with a 
152: multitude of local minima separated by high energy barriers. 
153: Canonical Monte Carlo or molecular dynamics
154: simulations likely will  get trapped in one of these
155: minima and not thermalize within the available CPU time. Only recently,
156: with the introduction of new and sophisticated algorithms such as
157: {\it generalized-ensemble} techniques \cite{Review} was it
158: possible to alleviate this problem in  protein simulations \cite{HO}.
159: For this reason, we rely  in our project  on one of these 
160: sophisticated techniques, multicanonical sampling \cite{MU}, whose
161: performance for polyalanine is extensively discussed in 
162: Ref.~\onlinecite{OH95b}.
163: 
164: The multicanonical algorithm \cite{MU}
165: assigns a weight $w_{mu} (E)\propto 1/n(E)$ to conformations with energy $E$.
166: Here, $n(E)$ is the density of states.
167: A  simulation with this weight
168: will  lead to a uniform distribution of energy:
169: \begin{equation}
170:   P_{mu}(E) \,  \propto \,  n(E)~w_{mu}(E) = {\rm const}~.
171: \label{eqmu}
172: \end{equation}
173: Thus, the simulation generates a 1D random walk in the
174: energy space,
175: allowing itself to escape from any  local minimum.
176: Since a large range of energies is sampled, one can
177: use the reweighting techniques \cite{FS} to  calculate thermodynamic
178: quantities over a wide range of temperatures $T$ by
179: \begin{equation}
180: <{\cal{A}}>_T ~=~ \frac{{\int dx~{\cal{A}}(x)~w^{-1}(E(x))~
181:                  e^{-\beta E(x)}}}
182:               {{\int dx~w^{-1}(E(x))~e^{-\beta E(x)}}}~,
183: \label{eqrw}
184: \end{equation}
185: where $x$ stands for configurations 
186: and $\beta$ for the inverse temperature, $\beta = 1/k_B T$.
187: Note that the weights are
188: not {\it a priori} known in multicanonical simulations  and
189:  estimators  need to be determined. This is often done 
190: by an iterative procedure described in detail in 
191: Refs.~\onlinecite{Berg,PH01h}, with which we needed  between 
192: 100,000 ($N=10$) and 800,000 ($N=30$)
193: sweeps for the weight factor calculations. All thermodynamic quantities 
194: are  estimated then from one production run of $N_{sw}$ Monte Carlo sweeps 
195: starting from a random initial conformation. 
196:  Our  emphasis is on the ASA solvent for which we have 
197: chosen  $N_{sw}= 6,000,000$, while for the DDE solvent approximation  
198: our simulations rely on $N_{sw}= 2,000,000$ sweeps. In all runs, we 
199: store every 10th sweep for further analysis. Our error bars are obtained 
200: by the jackknife method, with 12 bins for ASA, and 8 bins for the 
201: DDE model. The results of these implicit solvent
202: simulations  are compared with that of polyalanine
203: in gas phase \cite{HO98c,AH99b,AFH01d} that rely on
204:  $N_{sw}$=500,000, 500,000, 1,000,000, and 3,000,000 sweeps
205: for $N=10$, 15, 20, and 30, respectively.
206: 
207: Thermodynamic quantities that we  calculate from these 
208: multicanonical simulations for our three models (gas phase,
209: ASA and DDE) include  the average energy, 
210: specific heat, helicity and susceptibility. We further
211: evaluate the complex partition function zeros whose analysis 
212: was introduced by us recently as a tool in studies of structural 
213: transitions in biomolecules \cite{AH99b,AH00b,AFH01d}.
214: Such an analysis is possible because the multicanonical algorithm 
215: allows one to calculate reliable estimates for the spectral density:
216: \begin{equation}
217:   n(E) = P_{mu} (E) w^{-1}_{mu} (E)~.
218: \end{equation}
219: We can therefore construct
220: the partition function in the complex variable
221: $u$ from these estimates,
222: \begin{equation}
223:      Z(\beta) = \sum_{E} n(E) u^{E} ,                         \label{eq:r1}
224: \end{equation}
225: where $u=e^{-\beta}$.
226: The complex solutions of the partition function (the so-called
227: Fisher zeros \cite{fisher,itzykson}) correspond to the complex
228: extension of the temperature variable and determine the critical 
229: behavior of the model.  Partition function zeros analysis is used here
230: again to study for polyalanine the effects of the two implicit
231: solvent models on the characteristics of the helix-coil transition.
232: 
233: 
234: 
235: 
236: 
237: 
238: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
239: 
240: \section{Results and Discussion}
241: We could show in previous work that polyalanine exhibits a pronounced 
242: helix-coil transition in gas phase \cite{OH95b,HO98c,AH99b,AH00b}.
243: In this paper, we investigate now how the characteristics of that
244: transition change in the presence of an implicit solvent.
245: A natural order parameter 
246: for the helix-coil transition is $q_H =<n_H(T)>/(N-2) $, i.e. the
247: average number of helical residues divided by the number of residues
248: that can be part of an $\alpha$-helix.  A residue is considered 
249: as ``helical'' if its backbone dihedral angles $(\phi,\psi)$ take 
250: values in the range $(-70^{\circ}\pm 30^{\circ},-37^{\circ}\pm30^{\circ})$ 
251:  \cite{OH95b}. The normalization factor $N -2$ is chosen instead
252: of $N$, the number of residues,  because the terminal residues are 
253: flexible and are usually not part of an $\alpha$-helix.  We  start our 
254: analysis by displaying in Fig.~1a and 1b over a larger temperature range
255: this order parameter  as calculated from ASA (Fig.~1a) and DDE (Fig.~1b)
256: simulations. We notice in both figures a clear separation 
257: between a high-temperature phase with few helical residues and 
258: a low-temperature phase that is characterized by a 
259: single $\alpha$-helix. The free energy difference per residue  
260:  $\Delta g_{hc}(T) = (G_{helix} (T) - G_{coil} (T))/(N-2)$  
261: between helix and coil configurations is plotted as a function 
262: of temperature in the inlets. At high temperatures, $\Delta g_{hc}$ 
263: is positive and coil configurations are favored. On the other hand, below
264: a transition temperature $T_c$ (that depends on the chain length) helical
265: states are favored and the  free energy difference is consequently negative. 
266: 
267: 
268: 
269: The  reason for the stability of the helical state at low temperature
270: is the large difference in intramolecular energy to  the coil structures
271: (that have a much larger entropy and dominate in the high temperature
272: phase).  For instance, at $T= 275$ K, well into the low-temperature
273: phase, we find in ASA simulations for polyalanine chains of 
274: length $N=30$ an average potential energy difference
275: \mbox{$<\Delta E_{hc}(ASA)>$} 
276:  $= <E_{tot}({helix}) - E_{tot} ({coil})> = -43.5(2.9)$ kcal/mol. 
277: This energy difference is the sum of 
278: two competing terms. Helices are favored over coil configurations 
279: by an ECEPP/2 energy difference of 
280: $\Delta E_{ECEPP/2} = -55.2(3.7)$ kcal/mol, 
281: however, the ASA solvation term
282: favors coil configurations and decreases that term 
283: by $\Delta E_{ASA} = 9.2(1.9)$  kcal/mol. In DDE simulations,
284: we have only one energy term, and here we find 
285: \mbox{$<\Delta E_{hc}(DDE)>$} $ = -73.0(1.2)$ kcal/mol. 
286:  For comparison, we have found in previous gas-phase simulations 
287: of polyalanine chains of the same length at this temperature
288: a value of \mbox{$<\Delta E_{hc}>$} $=-75.9(1.8)$ kcal/mol.
289: Hence, in both solvent models, helices are energetically less favored
290: than in gas-phase. Consequentely, the free energy difference per residue
291: between helix and coil configurations is,  with
292: $\Delta g_{hc}({ASA})= -0.60$ kcal/mol and 
293: $\Delta g_{hc}({DDE}) = -1.07$ kcal/mol,
294: smaller than in gas phase simulations where we have
295:  $\Delta g_{hc}({gas}) = -1.17$  kcal/mol.
296: This  result is in contradiction with  work 
297: by Mortenson {\it et al.} \cite{Wales} who claim that 
298: solvent effects enhance helix formation, but in agreement
299: with other recent studies by Vila {\it et al.}\cite{VRS00} and
300: Levy {\it et al.}\cite{LJB01}.
301: 
302: In  simulations with both solvent terms, the free energy 
303: difference increases with the size of the molecule  making 
304: the transition between both states sharper as the system size 
305: increases (see the inlets of Fig.~1a and 1b). This indicates 
306: that we observe  in both both solvent models
307: a phase transition for polyalanine.  Such a phase transition 
308: requires long range order in the infinite system. We test
309: our model for this criterion by calculating
310: the helix propagation parameter $s$ and the nucleation 
311: parameter $\sigma$  of the Zimm-Bragg model \cite{ZB}. These
312: two quantities are related to the average number
313: of helical residues $<n>$ and the average length $<\ell>$ of a helical
314: segment. For large number $N$ of residues, we have
315: \begin{eqnarray}
316: {{<n>} \over N}~ &=& ~{1 \over 2} - {{1-s} 
317:        \over {2\sqrt{(1-s)^2 + 4s \sigma}}}~, \\
318: <\ell>~~ &=& ~1 + {2s \over {1-s+\sqrt{(1-s)^2 +4s \sigma}}}~.
319: \end{eqnarray}
320: 
321: Table 1 lists the  values  of $s$ and $\sigma$ as calculated by
322: this equation for all chain lengths and both models. For comparison,
323: we have added also the corresponding values for polyalanine in gas
324: phase as listed in Ref.~\onlinecite{HO98c}.
325: We are  especially interested in the nucleation parameter $\sigma$
326: that characterizes the probability of a helix segment to break
327:   apart into two pieces. In Fig.~2, we draw a log-log plot of this
328: quantity as a function of chain length for $T=275$ K, deep in the low
329: temperature region. For both  implicit solvent models the data points 
330: form a straight line. Hence, $\sigma$ can be described by a power-law: 
331: \begin{equation}
332: <\sigma(N)> = \sigma_0 N^{-c}~,
333: \end{equation}
334: with $\sigma^{ASA}_0 = 0.6(4) $,  $c^{OONS} = 1.2(1)$ and 
335:  $\sigma^{DDE}_0 = 0.8(2)$, $c^{DDE}= 1.2(1)$.
336: It follows that $<\sigma (\infty)> = 0$ for both  solvent models, 
337: i.e.  the probability for a helical segment to break into pieces approaches
338: in both models to zero for infinite chain length. The same result was found
339: in Ref.~\onlinecite{HO98c} for polyalanine in gas phase, and the corresponding
340: values of $\sigma$ are  shown in Fig.~2 for comparison. 
341: It follows that polyalanine exhibits  long-range order in the 
342: low-temperature phase (and therefore allows for a phase transition) 
343: independently on whether the polymer is
344: simulated in gas phase or with ASA or DDE  implicit solvent.
345: We remark that our $s$-parameter values  (see table 1) are in 
346: agreement with the experimental results of Ref.~\onlinecite{CB} where they 
347: list values of $s$(Ala) between $1.5$ and $2.19$.
348: 
349: 
350: In order to analyze the helix-coil transition in more detail, and to
351: compare our results with that of previous gas phase simulations, we
352: determine first the transition temperatures from the corresponding 
353: plots in the specific heat $C_N(T)$ that are shown in Fig.~3a (ASA) and 
354: 3b (DDE). The so obtained values are listed in table 2 together with
355:  the corresponding gas-phase values of Refs.~\onlinecite{HO98c,AH99b}.
356: As already observed in Ref.~\onlinecite{PH01h}, the  transition 
357: temperatures are for ASA simulations lower than in gas phase, 
358: but higher for DDE simulations. However, the differences decrease 
359: with chain length.  If we  extrapolate the listed temperatures 
360: to the  infinite chain limit by
361: $T_c (L) = T_c(\infty) - a\, e^{-bN}$, 
362: we find for ASA simulations as the critical temperature
363: $T_c(\infty) = 480$ K,  which is only 30 K
364: % a=299 b=0.09
365: lower than the corresponding value for gas-phase simulations:
366:  $T_c (\infty) = 514$ K. For DDE simulations  we find $T_c(\infty) = 525$ K,
367: i.e. the difference to gas-phase  is only $\approx 10$ K.
368: We note that the transition temperatures are outside of the
369: range of physiologically relevant temperatures indicating  limitations
370: of our energy functions in protein folding studies. 
371: 
372: 
373: Further information on the helix-coil transition can be obtained from 
374: the finite-size scaling analysis of the peak of the specific heat. 
375: We show in Fig.~4 a log-log plot of the maximum
376: value $C_N^{\rm max}$ of the specific heat as calculated from gas-phase,
377: ASA and DDE simulations for all four chain lengths. One expects that
378: for sufficiently large chains  $C_N^{\rm max}$  can be described by the
379: scaling relation:
380: \begin{equation}
381: C_N^{\rm max} \propto  N^{\alpha/d\nu}~.
382: \label{alpha}
383: \end{equation}
384: Only the gas-phase values fall in the log-log plot  on a straight line. 
385: This allowed us  to calculate in Ref.~\onlinecite{HO98c,AH99b} the specific
386: heat exponent $\alpha$ of polyalanine in gas phase: $\alpha = 0.86(10)$.  
387: However, in  simulations with one of solvent approximations,  $C_N^{\rm max}$  
388: can not be described by a straight line in a log-log plot. They rather 
389: seem to converge towards a constant value. Such a behavior is expected 
390: if the system has a specific heat exponent $\alpha = 0$, in which case 
391: logarithmic corrections need to be taken into account when describing 
392: the scaling of the specific heat maximum.  A similar picture (Fig.~5) 
393: appears for the scaling of the peak in the susceptibility,
394: \begin{equation}
395: \chi(T) = (<n_H^2> - <n_H>^2)/(N-2) \, ,
396: \end{equation}
397: which is expected to follow the scaling law,
398: \begin{equation}
399: \chi_N^{\rm max} \propto  N^{\gamma/d\nu}~.
400: \label{gamma}
401: \end{equation}
402: Fig.~6 displays the corresponding log-log plot of this quantity for
403: all three models and the four chain lengths.  Again, the gas-phase values
404: lay on a straight line and allow us to calculate an estimate for the 
405: susceptibility exponent $\gamma = 1.06(14)$ \cite{HO98c}.  However, 
406: for ASA and DDE simulations, the peak values  converge again towards
407: a constant value indicating $\gamma = 0$ and logarithmic corrections to 
408: scaling.
409: 
410: 
411: Calculation of these correction terms is in general
412: difficult and requires very large chain lengths and 
413: statistics (see, for instance, Ref.~\onlinecite{ADH97} where
414: the problem was tackled for the $2D$-Ising model).
415: For polyalanine, the available molecule sizes and statistics 
416: do not allow  for a reliable fit where the logarithmic 
417: correction terms are taken into account.  Instead, we pursue  
418: a different way.  We have shown in recent articles 
419:   \cite{AH99b,AH00b,AFH01d,AHP01e} 
420: that considerable information on structural transitions in 
421: biological molecules can be gained from an analysis of the
422: partition function zeros of the system.  We display 
423: in tables III, IV and V, respectively, the first  complex 
424: zeros for polyalanine chains in the gas phase, and 
425: for both  solvent models. Methods and the evaluation
426: of these zeros are discussed in Ref. \onlinecite{AFH01d} 
427: for the gas-phase model. 
428:  We note that  constructing the partition function is more difficult
429: for the solvated molecules. Even while the larger number of sweeps
430: is larger than in gas phase we can obtain only the first three zeros 
431: and not  four zeros as was possible for the gas phase model. 
432: 
433: One way of extracting  information on the strength of
434: transitions in molecules from
435: the partition function zeros is the approach by Janke and Kenna \cite{JK} 
436: who assume that the zeros condense for large systems
437: on a single line,
438: \begin{equation}
439: u_j = u_c + r_j\,{\rm exp} (i \varphi)~,                   \label{a1}
440: \end{equation}
441: with the cumulative density of zeros given by the
442: average
443: \begin{equation}
444:   G_N(r_j) = \frac{2j-1}{2N}  \, .                      \label{poly1}
445: \end{equation}
446: Here, $j$ labels the complex zeros in order of increasing imaginary part,
447:  $u_j = {\rm exp}(-\beta_j), j=1,2, \cdots $. Janke and Kenna postulate
448:  that this density scales for sufficiently large chains as \cite{JK}:
449: \begin{equation}
450: \frac{2j -1 }{2N} = a_1 ({\rm Im}\, u_j(L))^{a_2} + a_3 \,.\label{a5}
451: \end{equation}
452: A necessary condition for the existence of a phase transition
453: %characterized by the collected zeros
454: is that $a_3$ is compatible with zero, else it  indicates
455: that the system is in a well-defined phase.  Indeed, we obtain from
456: tables IV and V values $a_3$ that are compatible with zero for
457: all chain lengths. The values of
458: the constants $a_1$ and $a_2$ then  characterize the phase transition.
459: For instance, for first order transitions
460: the constant $a_2$  should take values $a_2 \sim 1$,
461:  and in this case the slope of this equation is
462: related to the latent heat\cite{JK}.
463: On the other hand, a value of $a_2$ larger than $1$
464: indicates a second order transition whose specific heat
465: exponent is given by $\alpha = 2 -a_2$.
466: 
467: 
468: Table VI lists the parameter $a_2(N)$ of Eq.~(\ref{a5}).
469: The estimates are less precise for the implicit solvent models
470: than for polyalanine in gas phase. However, the values clearly 
471: exclude the possibility of a first order transition for the 
472: solvated molecule while  for the polymer in gas phase we obtained 
473: as our best value for $a_2 = 1.16(1)$ but could not exclude
474: a value $a_2 = 1$ when we fit all chain lenghts \cite{AFH01d,AHP01e}
475:  (i.e. the possibility of a
476: first order phase transition). On the other hand, 
477: a multiple fit\cite{JK} in the parameters $j$ and $N$   leads 
478: in the case of  the ASA  solvent  to $a_2 = 1.79(8)$ (all zeros considered)
479: and  $a_2 = 1.90(9)$  if only the lowest eight zeros are considered.
480: Similarly, one finds for DDE solvent $a_2 = 1.74(11)$ (all zeros 
481: considered) and $a_2 = 1.81(24)$  if one  considers only the lowest 8 
482: zeros. The corresponding estimates for the specific heat 
483: exponent, $\alpha = 0.10(9)$ for ASA solvent and $\alpha = 0.19(24)$ 
484: for DDE solvent, are  small and within the errorbars compatible
485: with zero. Hence, they support our conjecture that 
486: $\alpha = 0$ for polyalanine with an implicit solvent. Note, that
487: this calculations demonstrates the advantages of partition function zero
488: analysis over other approaches:  our data do not allow us to
489: obtain a reliable estimate for $\alpha$ for either of the two solvent models 
490: from the  finite-size scaling of the peak in the specific heat 
491: (see Fig.~4) while this is possible with partition function zero analysis.
492: 
493: 
494: Our analysis of the maxima of the specific heats and the partition 
495: function zeros indicate that the helix-coil transition in 
496: polyalanine with a DDE or ASA solvent  is second order with 
497: exponents that are consistent with $\alpha =  \gamma =0$, and
498: by means of the hyperscaling relation $\alpha = 2 - d\nu$ with
499:  $d\nu = 2$.
500: Hence, the order of the
501: transition in the solvated molecule is fundamentally different from the
502: one in gas phase whose exponents $\alpha = 0.86(14), \gamma = 1.06(10)$
503: and $d\nu = 0.93(7)$ \cite{HO98c,AH99b} are consistent with a 
504: (weak) first order transition or a strong second order transition.
505: 
506: 
507: 
508: 
509: 
510: The weakening of the helix-coil transition in models that
511: try to account for protein-water interactions is not unexpected.
512: A large part of the energy gain through helix formation comes 
513: from the formation of hydrogen bonds between a residue and the 
514: forth following one that characterizes an $\alpha$-helix. 
515: Within water  this process competes with the entropically more 
516: favorable formation of hydrogen bonds with the surrounding water. 
517: Another factor are hydrophobic forces that penalize the extended 
518: structure of an $\alpha$-helix for the hydrophobic alanine. The
519: corresponding smaller gain in potential energy through helix 
520: formation in water when compared to gas phase is  observed by
521: us for both implicit solvent models. Hence, one can expect that 
522: helix-formation is more favored
523: in gas-phase than in solvent. This is consistent with our observation
524: of a weaker transition in the solvated molecule than observed in
525: gas phase. 
526: 
527: 
528: 
529: 
530: 
531: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
532:  
533: 
534: \section{conclusion}
535: We have studied the helix-coil transition in two models of polyalanine
536: that attempt to approximate the interaction of the molecule with the
537: surrounding water by an implicit solvent. We find that the helix-coil 
538: transition in these models is a true thermodynamic phase transition. 
539: However, while we have for polyalanine in gas phase
540: critical exponents that are consistent with  a (weak) first order phase 
541: transition, our  results 
542: rather indicate for ASA and DDE solvent models
543:  a (weak) second order transition with critical exponents
544: $\alpha$ and $\gamma$ close to zero  and $d\nu$ close to two. This change 
545: in the strength of the helix-coil
546: transition is consistent with our understanding of the protein-water
547: interactions. While our results  show that the two implicit solvation
548: models  describe  qualitatively correct the physics of protein-water
549: interaction, the un-physiologically high transition temperatures
550: demonstrate also the limitations of these models.
551: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
552: \\
553:  
554: \noindent
555: {\bf Acknowledgements}: \\
556:  N.A. Alves gratefully acknowledges support by CNPq (Brazil), and
557:  U.H. Hansmann support by a research grant
558: from the National Science Foundation (CHE-9981874).
559:  Part of this work was done while
560: U.H. visited the Ribeir\~ao Preto campus of the University of S\~ao Paulo.
561: He thanks FAPESP for a generous travel grant and the Department of
562: Physics and Mathematics for kind hospitality.
563:  The authors are also grateful to DFMA (IFUSP) for the computer facilities
564: made available to them.
565: 
566: 
567: \clearpage
568: %%\end{multicols}
569: 
570: 
571: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
572:  
573: %\begin{thebibliography}{99}
574: \begin{references}
575: \bibitem{Poland} D.~Poland and H.A.~Scheraga, {\it Theory of Helix-Coil
576:          Transitions in Biopolymers} (Academic Press, New York, 1970).
577: \bibitem{HO98c} U.H.E.~Hansmann and Y.~Okamoto, {J. Chem.~Phys.} {\bf 110},
578:                 1267 (1999); {\bf 111} 1339(E) (1999).
579: \bibitem{AH99b} N.A. Alves  and U.H.E. Hansmann,\
580:                 {Phys. Rev. Lett.}  {\bf 84}, 1836 (2000).
581: \bibitem{AH00b} N.A. Alves and U.H.E.~Hansmann, {Physica A}
582:                 {\bf 292}, 509 (2001).
583: \bibitem{Jarrold} R.R. Hudgins, M.A. Ratner and M.F. Jarrold,
584:                   {J. Am. Chem. Soc.} {\bf 120}, 12974 (1998).
585: \bibitem{MO1}  A. Mitsutake and Y. Okamoto, {J. Chem. Phys.} {\bf 112},
586:                10638 (2000).
587: \bibitem{PH01h} Y.~Peng and U.H.E.~Hansmann, {Biophys. J.}
588:                 {\bf 82}, 3269 (2002). 
589: \bibitem{oons} T. Ooi, M. Obatake, G. Nemethy, H.A. Scheraga,
590:                {Proc. Natl. Acad. Sci.  USA} {\bf 8}, 3086 (1987).
591: \bibitem{lavarty} R. Lavery, H. Sklenar, K. Zakrzewska, B. Pullman,
592:                   {J. Biomol. Struct. \& Dynamics} {\bf  3}, 989 (1986).
593: \bibitem{EC} M.J. Sippl, G. N{\'e}methy, and H.A. Scheraga,
594:              {J. Phys. Chem.} {\bf 88}, 6231~(1984), 
595:              and references therein.
596: \bibitem{SMMP} F.~Eisenmenger, U.H.E.~Hansmann, Sh.~Hayryan, C.-K.~Hu,
597:                {Comp.~Phys.~Comm.} {\bf 138}, 192 (2001).
598: \bibitem{hingerty} B. Hingerty, R.H. Richie, T.L. Ferrel, J. Turner,
599:                   {Biopolymers} {\bf 24}, 427 (1985).
600: \bibitem{Review} U.H.E.~Hansmann and Y.~Okamoto, 
601:          in: Stauffer, D. (ed.) ``{\it Annual Reviews in Computational 
602:          Physics VI}'',(Singapore: World Scientific), p.129.  (1998).
603: \bibitem{HO} U.H.E. Hansmann and Y. Okamoto,  J.~Comp.~Chem.
604:              {\bf 14}, 1333 (1993).
605: \bibitem{MU} B.A. Berg  and T. Neuhaus,
606:              {Phys. Lett.} B {\bf 267}, 249 (1991).
607: \bibitem{OH95b} Y.~Okamoto and U.H.E.~Hansmann,  J.~Phys.~Chem.
608:                 {\bf 99}, 11276 (1995).
609: \bibitem{FS} A.M. Ferrenberg and R.H. Swendsen, {Phys.\ Rev.\ Lett.}
610:              {\bf  61}, 2635 (1988); {Phys. Rev. Lett.} {\bf 63},
611:              1658(E) (1989), and
612:               references given in the erratum.
613: \bibitem{Berg}  B.A. Berg, {J. Stat. Phys.} {\bf 82}, 323 (1996).
614: \bibitem{AFH01d} N.A. Alves, J.P.N. Ferrite and U.H.E. Hansmann,
615:                 {Phys. Rev. E} {\bf 65}, 036110 (2002).
616: \bibitem{fisher} M.E. Fisher, in {\it Lectures in Theoretical Physics},
617:                  vol. 7c p.1 (University of Colorado Press, Boulder, 1965).
618: \bibitem{itzykson} C. Itzykson, R.B. Pearson, and J.B. Zuber,
619:                    Nucl. Phys. B {\bf 220} [FS8], 415 (1983).
620: \bibitem{Wales} P.N. Mortenson, D.A. Evans and D.J. Wales, 
621:                 {J. Chem. Phys.} {\bf 117}, 1363 (2002). 
622: \bibitem{VRS00} J.A. Vila, D.R. Ripoll and H.A. Scheraga, 
623:                 Proc. Natl. Acad. Sci. USA {\bf 97}, 13075 (2000).
624: \bibitem{LJB01} Y. Levy, J. Jortner and O.M. Becker, 
625:                 Proc. Natl. Acad. Sci. USA {\bf 98}, 2188 (2001).
626: \bibitem{ZB} B.H. Zimm and J.K. Bragg,  J. Chem. Phys. {\bf 31}, 526 (1959).
627: \bibitem{CB} Chakrabartty, A.; R.L. Baldwin, R.L. In
628:     {\it Protein Folding: In Vivo and In Vitro}; Cleland J.;
629:     King, J. eds.; ACS Press: Washington, D.C., 1993; pp. 166--177.
630: \bibitem{ADH97} N.A. Alves, J.R.~Drugowich de Felicio and U.H.E.~Hansmann,
631:                 {Int.~J.~Mod.~Phys.~C} {\bf 8}, 1063 (1997).
632: \bibitem{AHP01e} N.A. Alves, U.H.E. Hansmann and Y. Peng,
633:                  {Int. J. Mol. Sci.} {\bf 3}, 17 (2002).
634: \bibitem{JK}   W. Janke and R. Kenna, J. Stat. Phys. {\bf 102}, 1211 (2001); 
635:                Nucl. Phys. Proc. Supl. {\bf B106}, 905 (2002); 
636:                {\it http://arXiv.org/abs/cond-mat/0103333};
637:                {\it http://arXiv.org/abs/cond-mat/0208014}.
638: \end{references}
639: %\end{thebibliography}
640: \vfil
641: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
642: 
643: \clearpage 
644: \newpage
645: 
646: 
647: %TABLE 1
648: \begin{table}[ht]
649: \renewcommand{\tablename}{Table}
650: \caption{Nucleation parameter $\sigma$ and helix-propagation parameter
651:  $s$ of the Zimm-Bragg model, as calculated from multicanonical simulations
652:  for polyalanine chains of length $N=10,15,20$ and $30$ in DDE or ASA
653:  solvent. The gas phase results (GP) are taken from Ref.~2}
654: %      are taken from Ref.~\onlinecite{HO98c}.}
655: \begin{center} 
656: \begin{tabular}{lccccccccc} \\
657: \\[-0.3cm]
658:  &&\multicolumn{2}{c}{\rm GP} && \multicolumn{2}{c}{\rm ASA}
659:  &&\multicolumn{2}{c}{\rm DDE}\\
660:  $N$&& $\sigma$ & $s$       & & $\sigma$ &  $s$     & & $\sigma$ &$s$ \\
661: \\[-0.35cm]
662: \hline
663: \\[-0.3cm]
664:  10 && 0.126(4) & 1.561(28) & & 0.122(1) & 1.345(29)& & 0.130(1) & 1.767(6)\\  
665:  15 && 0.074(2) & 1.679(10) & & 0.064(1) & 1.501(13)& & 0.073(1) & 1.860(11)\\
666:  20 && 0.056(1) & 1.780(13) & & 0.045(1) & 1.561(12)& & 0.052(1) & 1.890(10)\\
667:  30 && 0.036(1) & 1.845(25) & & 0.032(1) & 1.501(96)& & 0.034(1) & 1.934(11)\\
668: \end{tabular}
669: \end{center} 
670: \end{table}
671: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
672: % TABLE 2
673: \begin{table}[ht]
674: \renewcommand{\tablename}{Table}
675: \caption{\baselineskip=0.8cm Numerical results for various polyalanine 
676:        chain lenghts $N$: critical temperature $T_c$ 
677:        defined by the maximum of specific heat $C^{\rm max}$ and the
678:        maximum of susceptibility $\chi^{\rm max}$. The gas phase results (GP)
679:        are taken from Ref.~2.}
680: %      are taken from Ref.~\onlinecite{HO98c}.}
681: \begin{center}
682: \begin{tabular}{lcrclcrclcrcl}\\
683: \\[-0.3cm]
684:      & &    \multicolumn{3}{c}{\rm~~GP} & &
685:             \multicolumn{3}{c}{\rm ASA} & &
686:             \multicolumn{3}{c}{\rm ~DDE} \\
687: $~N$ & &$T_c$~~ &~$C^{\rm max}$ &$\chi^{\rm max}$ & &
688:         $T_c$~~ &~$C^{\rm max}$ &$\chi^{\rm max}$ & &
689:         $T_c$~~ &~$C^{\rm max}$ &$\chi^{\rm max}$ \\
690: \\[-0.35cm]
691: \hline
692: \\[-0.3cm]
693: ~10  & & 427(7)   &~~8.9(3)        & 0.49(2)          & &
694:          333(2)   & 10.2(1)        & 0.613(7)         & &
695:          482(1)   &~~9.9(1)        & 0.605(5)  \\
696: ~15  & & 492(5)   & 12.3(4)        & 0.72(3)          & &
697:          403(3)   & 13.1(2)        & 0.81(1)          & &
698:          517(3)   & 12.3(2)        & 0.71(1)  \\
699: ~20  & & 508(5)   & 16.0(8)        & 1.08(3)          & &
700:          430(2)   & 14.5(4)        & 0.91(3)          & &
701:          518(4)   & 14.0(4)        & 0.79(2)    \\
702: ~30  & & 518(7)   & 22.8(1.2)      & 1.50(8)          & &
703:          461(3)   & 15.5(1.1)      & 1.00(8)          & &
704:          523(7)   & 13.9(6)        & 0.81(4)     \\
705: \end{tabular}
706: \end{center}
707: \end{table}
708: 
709: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
710: 
711: % TABLE 3
712: \begin{table}[ht]
713: \renewcommand{\tablename}{Table}
714: \caption{\baselineskip=0.8cm Partition function zeros for polyalanine
715:           in the gas phase. The data are taken from 
716:           Ref.~19.}
717: %         Ref.~\onlinecite{AFH01d}.}
718: \begin{center}
719: \begin{tabular}{lcccccccc}\\
720: \\[-0.3cm]
721: $~N$~~&Re$(u_1)$ &Im$(u_1)$    & Re$(u_2)$ & Im$(u_2)$ &Re$(u_3)$ &Im$(u_3)$ 
722:       &Re$(u_4)$ &Im$(u_4)$  \\
723: \hline
724: \\[-0.3cm]
725: ~10 &0.30530(12) &0.07720(14)  &0.2823(13) &0.13820(61)&0.2459(72)&0.1851(63)
726:     &0.172(11)   &0.2200(71)    \\
727: ~15 &0.356863(61)&~0.053346(39)&0.34167(60)&0.10440(59)&0.3331(48)&0.1454(28)
728:     &~0.3067(81) &0.1689(32)    \\
729: ~20 &0.374016(41)&~0.042331(45)&0.36161(27)&0.08109(24)&0.3569(27)&0.1154(13)
730:     &~0.3336(56) &0.1470(27)    \\
731: ~30 &0.378189(19)&~0.027167(32)&~0.37399(14)&~0.05420(27)&0.3693(11)
732:     &0.0804(13)  &~0.35854(63) &0.1022(43)    
733: \end{tabular}
734: \end{center}
735: \end{table}
736: 
737: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
738: 
739: 
740: \clearpage 
741: \newpage
742: 
743: 
744: 
745: % TABLE 4
746: \begin{table}[ht]
747: \renewcommand{\tablename}{Table}
748: \caption{\baselineskip=0.8cm Partition function zeros for polyalanine
749:  chain lenghts $N$ with ASA solvent.}
750: \begin{center}
751: \begin{tabular}{lcccccc}\\
752: \\[-0.3cm]
753: $~N$~~&Re$(u_1)$ &Im$(u_1)$  & Re$(u_2)$ & Im$(u_2)$ &Re$(u_3)$  &Im$(u_3)$ \\
754: \\[-0.35cm]
755: \hline
756: \\[-0.3cm]
757: ~10 &0.2191(20) &0.06511(65) &0.1777(12) &0.1182(15) &0.1242(30) &0.1543(27) \\
758: ~15 &0.2870(24) &0.05092(83) &0.2695(21) &0.0981(14) &0.2463(27) &0.1265(38) \\
759: ~20 &0.3116(14) &0.04458(99) &0.2850(54) &0.0813(42) &0.2769(52) &0.1281(54) \\
760: ~30 &0.3378(23) &0.0385(28)  &0.3255(98) &0.0611(74) &0.304(15)  &0.0829(97) \\
761: \end{tabular}
762: \end{center}
763: \end{table}
764: 
765: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
766: 
767: % TABLE 5
768: \begin{table}[ht]
769: \renewcommand{\tablename}{Table}
770: \caption{\baselineskip=0.8cm Partition function zeros for polyalanine
771:  chain lenghts $N$ with DDE solvent.}
772: \begin{center}
773: \begin{tabular}{lcccccc}\\
774: \\[-0.3cm]
775: $~N$~~&Re$(u_1)$ &Im$(u_1)$  & Re$(u_2)$ & Im$(u_2)$ &Re$(u_3)$  &Im$(u_3)$ \\
776: \\[-0.35cm]
777: \hline
778: \\[-0.3cm]
779: ~10 &0.35256(86)&0.07316(63) &0.3329(12) &0.1424(12) &0.2959(32) &0.1907(16)\\
780: ~15 &0.3795(19) &0.05653(89) &0.3588(21) &0.0972(15) &0.3612(87) &0.1475(71)\\
781: ~20 &0.3794(30) &0.0487(22)  &0.3710(40) &0.0788(24) &0.3784(62) &0.1202(54)\\
782: ~30 &0.3883(36) &0.0449(31)  &0.3976(37) &0.0774(37) &0.378(20)  &0.1143(71)\\
783: \end{tabular}
784: \end{center}
785: \end{table}
786: 
787: 
788: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
789: 
790: % TABLE 6
791: \begin{table}[ht]
792: \renewcommand{\tablename}{Table}
793: \caption{\baselineskip=0.8cm  Estimates of the parameter $a_2$  
794:          for polyalanine. The gas phase results (GP) are from 
795:           Ref.~19. }
796: %         Ref.~\onlinecite{AFH01d}. }
797: \begin{center}
798: \begin{tabular}{lccc}\\
799: \\[-0.3cm]
800: $~N$ & GP   &  ASA   &  DDE \\
801: \\[-0.35cm]
802: \hline
803: \\[-0.3cm]
804: ~10 & 1.862(46) & 1.861(18) & 1.675(23) \\
805: ~15 & 1.664(16) & 1.750(73) & 1.70(23)  \\
806: ~20 & 1.558(24) & 1.541(20) & 1.79(31)  \\
807: ~30 & 1.473(30) & 2.12(19)  & 1.74(20)  
808: \end{tabular}
809: \end{center}
810: \end{table}
811: 
812: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
813: 
814: %\newpage
815: \clearpage
816: 
817: 
818: %FIGURE 1a
819: \begin{figure}[b]
820: %\begin{figure}[!ht]
821: \begin{center}
822: \begin{minipage}[t]{0.95\textwidth}
823: \centering
824: \includegraphics[angle=-90,width=0.72\textwidth]{Fig1a.eps}
825: \renewcommand{\figurename}{(Fig.1a)}
826: \caption{
827:   Temperature dependence of the  helicity order 
828:                parameter $q_{H} = <n_H>/(N-2)$  as obtained
829:                from simulations of polyalanine of chain length
830:                $N=10,15,20,30$ with ASA
831:                solvent representation. The 
832:                free energy difference per residue $g_{hc}$  
833:                between helical and coil states is shown in the
834:                corresponding inlets.}
835: \label{Fig. 1a}
836: \end{minipage}
837: \end{center}
838: \end{figure}
839: 
840: 
841: %FIGURE 1b
842: \begin{figure}[b]
843: %\begin{figure}[!ht]
844: \begin{center}
845: \begin{minipage}[t]{0.95\textwidth}
846: \centering
847: \includegraphics[angle=-90,width=0.72\textwidth]{Fig1b.eps}
848: \renewcommand{\figurename}{(Fig.1b)}
849: \caption{
850:   Temperature dependence of the  helicity order 
851:                parameter $q_{H} = <n_H>/(N-2)$  as obtained
852:                from simulations of polyalanine of chain length
853:                $N=10,15,20,30$ with
854:                 DDE solvent. The 
855:                free energy difference per residue $g_{hc}$  
856:                between helical and coil states is shown in the
857:                corresponding inlets.}
858: \label{Fig. 1b}
859: \end{minipage}
860: \end{center}
861: \end{figure}
862: 
863: 
864: \newpage
865: 
866: %FIGURE 2
867: \begin{figure}[b]
868: %\begin{figure}[!ht]
869: \begin{center}
870: \begin{minipage}[t]{0.95\textwidth}
871: \centering
872: \includegraphics[angle=-90,width=0.72\textwidth]{Fig2.eps}
873: \renewcommand{\figurename}{(Fig.2)}
874: \caption{
875:   Log-log plot of the nucleation parameter $\sigma$ as a
876:                function of the number of residues at $T=275$ K. The 
877:                gas-phase values (GP) are taken from Ref. 2.}
878: \label{Fig. 2}
879: \end{minipage}
880: \end{center}
881: \end{figure}
882: 
883: 
884: %FIGURE 3a
885: \begin{figure}[b]
886: %\begin{figure}[!ht]
887: \begin{center}
888: \begin{minipage}[t]{0.95\textwidth}
889: \centering
890: \includegraphics[angle=-90,width=0.72\textwidth]{Fig3a.eps}
891: \renewcommand{\figurename}{(Fig.3a)}
892: \caption{
893:   Specific heat $C(T)$ as a function of temperature $T$ for
894:                polyalanine molecules of chain length $N=10, 15, 20,$ 
895:                and $30$ with ASA
896:                solvent representation.}
897: \label{Fig. 3a}
898: \end{minipage}
899: \end{center}
900: \end{figure}
901: 
902: \newpage
903: 
904: 
905: %FIGURE 3b
906: \begin{figure}[b]
907: %\begin{figure}[!ht]
908: \begin{center}
909: \begin{minipage}[t]{0.95\textwidth}
910: \centering
911: \includegraphics[angle=-90,width=0.72\textwidth]{Fig3b.eps}
912: \renewcommand{\figurename}{(Fig.3b)}
913: \caption{
914:   Specific heat $C(T)$ as a function of temperature $T$ for
915:                polyalanine molecules of chain length $N=10, 15, 20,$ 
916:                and $30$ with DDE
917:                solvent representation.}
918: \label{Fig. 3b}
919: \end{minipage}
920: \end{center}
921: \end{figure}
922: 
923: 
924: %FIGURE 4
925: \begin{figure}[b]
926: %\begin{figure}[!ht]
927: \begin{center}
928: \begin{minipage}[t]{0.95\textwidth}
929: \centering
930: \includegraphics[angle=-90,width=0.72\textwidth]{Fig4.eps}
931: \renewcommand{\figurename}{(Fig.4)}
932: \caption{
933:  Log-log plot of the peak value $C^{\rm max}(N)$ of the specific
934:                heat as a function of chain length $N$ for 
935:                ASA and DDE simulations. For comparison, we show also
936:                the corresponding gas-phase values (GP) of 
937:                Refs. 2,3 and the straight-line
938:                fit through them.}
939: \label{Fig. 4}
940: \end{minipage}
941: \end{center}
942: \end{figure}
943: 
944: 
945: \newpage
946: 
947: %FIGURE 5a
948: \begin{figure}[b]
949: %\begin{figure}[!ht]
950: \begin{center}
951: \begin{minipage}[t]{0.95\textwidth}
952: \centering
953: \includegraphics[angle=-90,width=0.72\textwidth]{Fig5a.eps}
954: \renewcommand{\figurename}{(Fig.5a)}
955: \caption{
956:   Susceptibility $\chi(T)$ as a function of temperature $T$ for
957:                polyalanine molecules of chain length $N=10, 15, 20,$ 
958:                and $30$ with ASA
959:                solvent representation.}
960: \label{Fig. 5a}
961: \end{minipage}
962: \end{center}
963: \end{figure}
964: 
965: 
966: %FIGURE 5b
967: \begin{figure}[b]
968: %\begin{figure}[!ht]
969: \begin{center}
970: \begin{minipage}[t]{0.95\textwidth}
971: \centering
972: \includegraphics[angle=-90,width=0.72\textwidth]{Fig5b.eps}
973: \renewcommand{\figurename}{(Fig.5b)}
974: \caption{
975:   Susceptibility $\chi(T)$ as a function of temperature $T$ for
976:                polyalanine molecules of chain length $N=10, 15, 20,$ 
977:                and $30$ with DDE
978:                solvent representation.}
979: \label{Fig. 5b}
980: \end{minipage}
981: \end{center}
982: \end{figure}
983: 
984: \newpage
985: 
986: 
987: %FIGURE 6
988: \begin{figure}[b]
989: %\begin{figure}[!ht]
990: \begin{center}
991: \begin{minipage}[t]{0.95\textwidth}
992: \centering
993: \includegraphics[angle=-90,width=0.72\textwidth]{Fig6.eps}
994: \renewcommand{\figurename}{(Fig.6)}
995: \caption{
996:   Log-log plot of the peak value $\chi^{\rm max}(N)$ of the 
997:                susceptibility as a function of chain length $N$ for 
998:                ASA  and DDE simulations. The corresponding gas-phase 
999:                values (GP) of Ref. 2 and the straight-line
1000:                fit through them are shown for comparison.}
1001: \label{Fig. 6}
1002: \end{minipage}
1003: \end{center}
1004: \end{figure}
1005: 
1006: 
1007: %FIGURE 7
1008: \begin{figure}[b]
1009: %\begin{figure}[!ht]
1010: \begin{center}
1011: \begin{minipage}[t]{0.95\textwidth}
1012: \centering
1013: \includegraphics[angle=-90,width=0.72\textwidth]{Fig7.eps}
1014: \renewcommand{\figurename}{(Fig.7)}
1015: \caption{
1016:  Distribution of zeros for the polyalanine model with
1017:                ASA solvent. }
1018: \label{Fig. 7}
1019: \end{minipage}
1020: \end{center}
1021: \end{figure}
1022: 
1023: 
1024: \newpage
1025: 
1026: %FIGURE 8
1027: \begin{figure}[b]
1028: %\begin{figure}[!ht]
1029: \begin{center}
1030: \begin{minipage}[t]{0.95\textwidth}
1031: \centering
1032: \includegraphics[angle=-90,width=0.72\textwidth]{Fig8.eps}
1033: \renewcommand{\figurename}{(Fig.8)}
1034: \caption{
1035:  Distribution of zeros for the polyalanine model with
1036:                DDE solvent. }
1037: \label{Fig. 8}
1038: \end{minipage}
1039: \end{center}
1040: \end{figure}
1041: 
1042: 
1043: \end{document}
1044: 
1045: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1046: 
1047: 
1048: 
1049: 
1050: %%%\end{document}
1051: \newpage
1052: \cleardoublepage
1053: 
1054: 
1055: 
1056: %FIGURE 1a
1057: \begin{figure}[t]
1058: \begin{center}
1059: \begin{minipage}[t]{0.95\textwidth}
1060: \centering
1061: \includegraphics[angle=-90,width=0.72\textwidth]{Fig1a.eps}
1062: \renewcommand{\figurename}{Fig. }
1063: \caption{Temperature dependence of the  helicity order 
1064:                parameter $q_{H} = <n_H>/(N-2)$  as obtained
1065:                from simulations of polyalanine of chain length
1066:                $N=10,15,20,30$ with a) ASA
1067:                solvent representation and b) DDE solvent.}
1068: \label{fig1a}
1069: \end{minipage}
1070: \end{center}
1071: \end{figure}
1072: 
1073: 
1074: %FIGURE 1b
1075: \begin{figure}[t]
1076: \begin{center}
1077: \begin{minipage}[t]{0.95\textwidth}
1078: \centering
1079: \includegraphics[angle=-90,width=0.72\textwidth]{Fig1b.eps}
1080: %%%\renewcommand{\figurename}{Fig. 1b}
1081: \begin{quote}{    b) DDE solvent.}
1082: \end{quote}
1083: \label{fig1b}
1084: \end{minipage}
1085: \end{center}
1086: \end{figure}
1087: 
1088: 
1089: 
1090: \newpage
1091: \cleardoublepage
1092: 
1093: 
1094: %FIGURE 2
1095: \begin{figure}[!ht]
1096: \begin{center}
1097: \begin{minipage}[t]{0.95\textwidth}
1098: \centering
1099: \includegraphics[angle=-90,width=0.72\textwidth]{Fig2.eps}
1100: \renewcommand{\figurename}{Fig.}
1101: \caption{Log-log plot of the nucleation parameter $\sigma$ as a
1102:          function of the number of residues at $T=275$ K.}
1103: \label{fig2}
1104: \end{minipage}
1105: \end{center}
1106: \end{figure}
1107: 
1108: 
1109: 
1110: \newpage
1111: \cleardoublepage
1112: 
1113: 
1114: 
1115: %FIGURE 3a
1116: \begin{figure}[!ht]
1117: \begin{center}
1118: \begin{minipage}[t]{0.95\textwidth}
1119: \centering
1120: \includegraphics[angle=-90,width=0.72\textwidth]{Fig3a.eps}
1121: \renewcommand{\figurename}{Fig.}
1122: \caption{Specific heat $C(T)$ as a function of temperature $T$ for
1123:          polyalanine molecules of chain length $N=10, 15, 20,$ 
1124:          and $30$ with a) ASA
1125:          solvent representation and b) DDE solvent.}
1126: \label{fig3a}
1127: \end{minipage}
1128: \end{center}
1129: \end{figure}
1130: 
1131: 
1132: %FIGURE 3b
1133: \begin{figure}[!ht]
1134: \begin{center}
1135: \begin{minipage}[t]{0.95\textwidth}
1136: \centering
1137: \includegraphics[angle=-90,width=0.72\textwidth]{Fig3b.eps}
1138: \renewcommand{\figurename}{Fig.3b}
1139: \begin{quote}{  b) DDE solvent.}
1140: \end{quote}
1141: \label{fig3b}
1142: \end{minipage}
1143: \end{center}
1144: \end{figure}
1145: 
1146: 
1147: \newpage
1148: \cleardoublepage
1149: 
1150: 
1151: 
1152: %FIGURE 4
1153: \begin{figure}[!ht]
1154: \begin{center}
1155: \begin{minipage}[t]{0.95\textwidth}
1156: \centering
1157: \includegraphics[angle=-90,width=0.72\textwidth]{Fig4.eps}
1158: \renewcommand{\figurename}{Fig.}
1159: \caption{ Log-log plot of the peak value $C^{\rm max}(N)$ of the specific
1160:                heat as a function of chain length $N$ for gas-phase,
1161:                DDE and ASA simulations.}
1162: \label{fig4}
1163: \end{minipage}
1164: \end{center}
1165: \end{figure}
1166: 
1167: 
1168: \newpage
1169: \cleardoublepage
1170: 
1171: 
1172: %FIGURE 5a
1173: \begin{figure}[!ht]
1174: \begin{center}
1175: \begin{minipage}[t]{0.95\textwidth}
1176: \centering
1177: \includegraphics[angle=-90,width=0.72\textwidth]{Fig5a.eps}
1178: \renewcommand{\figurename}{Fig.}
1179: \caption{Susceptibility $\chi(T)$ as a function of temperature $T$ for
1180:                polyalanine molecules of chain length $N=10, 15, 20,$ 
1181:                and $30$ with a) ASA
1182:                solvent representation and b) DDE solvent.}
1183: \label{fig5a}
1184: \end{minipage}
1185: \end{center}
1186: \end{figure}
1187: 
1188: 
1189: %FIGURE 5b
1190: \begin{figure}[!ht]
1191: \begin{center}
1192: \begin{minipage}[t]{0.95\textwidth}
1193: \centering
1194: \includegraphics[angle=-90,width=0.72\textwidth]{Fig5b.eps}
1195: \renewcommand{\figurename}{Fig. 5b}
1196: \begin{quote}{ b) DDE solvent.}
1197: \end{quote}
1198: \label{fig5b}
1199: \end{minipage}
1200: \end{center}
1201: \end{figure}
1202: 
1203: 
1204: 
1205: \newpage
1206: \cleardoublepage
1207: 
1208: 
1209: %FIGURE 6
1210: \begin{figure}[!ht]
1211: \begin{center}
1212: \begin{minipage}[t]{0.95\textwidth}
1213: \centering
1214: \includegraphics[angle=-90,width=0.72\textwidth]{Fig6.eps}
1215: \renewcommand{\figurename}{Fig.}
1216: \caption{Log-log plot of the peak value $\chi^{\rm max}(N)$ of the 
1217:              susceptibility as a function of chain length $N$ for gas-phase,
1218:              DDE and ASA simulations.}
1219: \label{fig6}
1220: \end{minipage}
1221: \end{center}
1222: \end{figure}
1223: 
1224: 
1225: 
1226: \newpage
1227: \cleardoublepage
1228: 
1229: 
1230: %FIGURE 7
1231: \begin{figure}[!ht]
1232: \begin{center}
1233: \begin{minipage}[t]{0.95\textwidth}
1234: \centering
1235: \includegraphics[angle=-90,width=0.72\textwidth]{Fig7.eps}
1236: \renewcommand{\figurename}{Fig.}
1237: \caption{Distribution of zeros for the polyalanine model with
1238:                ASA solvent.}
1239: \label{fig7}
1240: \end{minipage}
1241: \end{center}
1242: \end{figure}
1243: 
1244: 
1245: %FIGURE 8
1246: \begin{figure}[!ht]
1247: \begin{center}
1248: \begin{minipage}[t]{0.95\textwidth}
1249: \centering
1250: \includegraphics[angle=-90,width=0.72\textwidth]{Fig8.eps}
1251: \renewcommand{\figurename}{Fig.}
1252: \caption{Distribution of zeros for the polyalanine model with
1253:          DDE solvent.} 
1254: \label{fig8}
1255: \end{minipage}
1256: \end{center}
1257: \end{figure}
1258: 
1259: 
1260: 
1261: 
1262: \newpage
1263: \cleardoublepage
1264: 
1265: 
1266: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1267: \bibitem{SKY} K.R.~Shoemaker, P.S.~Kim, E.J.~York, J.M.~Stewart, 
1268:               R.L.~Baldwin, {\it Nature} {\bf 326}, 563 (1987).
1269: \bibitem{curr_op}  U.H.E.~Hansmann and Y.~Okamoto,
1270:         {\it Curr.~Opin.~Struc.~Biol.} {\bf 9}, 177 (1999).
1271: \bibitem{GR1} S. Grossmann and W. Rosenhauer, Z. Physik 207 (1967)138.
1272: \bibitem{GR2} S. Grossmann and W. Rosenhauer, Z. Physik 218 (1969)437;
1273:               S. Grossmann and W. Rosenhauer, Z. Physik 218 (1969)449.
1274: \bibitem{BMH} P. Borrmann, O. M\"ulken and J. Harting,
1275:               Phys. Rev. Lett. 84 (2000)3511;
1276:               O. M\"ulken, P. Borrmann, J. Harting and H. Stamerjohanns,
1277:               Phys. Rev. A64 (2001) 013611;
1278:               O. M\"ulken and P. Borrmann, Phys. Rev. C63 (2001) 024306.
1279: