1: \documentclass[12pt]{article}
2: \usepackage{graphics}
3: \usepackage{cite}
4: \usepackage{epsfig}
5: \newcommand {\etal} {{\it et~al.}}
6: %UGLY \renewcommand{\baselinestretch}{2.0}
7: \textwidth 432pt
8: %\textwidth 26pc
9: \flushbottom
10: \textheight 594pt
11: \topmargin 72pt
12: \headheight 0pt
13: \headsep 0pt
14: \footskip 54pt
15: \oddsidemargin 0pt
16: \parindent 0in
17: \parskip 3ex
18:
19: \renewcommand{\textfraction}{0.1}
20: \renewcommand{\topfraction}{0.9}
21: \renewcommand{\bottomfraction}{0.9}
22: \renewcommand{\floatpagefraction}{1.0}
23: \renewcommand{\thefootnote}{\fnsymbol{footnote}}
24: %\renewcommand\citepunct{; }
25: %\renewcommand\citeleft{(} % parentheses around list
26: %\renewcommand\citeright{)} % parentheses around list
27: %\makeatletter \renewcommand\@biblabel[1]{} \makeatother
28: % JOURNALS
29: \newcommand {\APC} {{\it Adv. Protein Chem.\ }}
30: \newcommand {\Bioch} {{\it Biochemistry\ }}
31: \newcommand {\Biomet} {{\it Biometrika\ }}
32: \newcommand {\Biopol} {{\it Biopolymers\ }}
33: \newcommand {\BC} {{\it Biophys.\ Chem.\ }}
34: \newcommand {\BJ} {{\it Biophys.\ J.\ }}
35: \newcommand {\CPL} {{\it Chem. Phys. Lett.\ }}
36: \newcommand {\COSB} {{\it Curr.\ Opin.\ Struct.\ Biol.\ }}
37: \newcommand {\EL} {{\it Europhys.\ Lett.\ }}
38: \newcommand {\JCC} {{\it J.\ Comput.\ Chem.\ \ }}
39: \newcommand {\JCoP} {{\it J.\ Comput.\ Phys.\ }}
40: \newcommand {\JCP} {{\it J.\ Chem.\ Phys.\ }}
41: \newcommand {\JMB} {{\it J.\ Mol.\ Biol.\ }}
42: \newcommand {\JP} {{\it J.\ Phys.\ }}
43: \newcommand {\JPC} {{\it J.\ Phys.\ Chem.\ }}
44: \newcommand {\JPSJ} {{\it J. Phys. Soc. (Jap)\ }}
45: \newcommand {\JSP} {{\it J.\ Stat.\ Phys.\ }}
46: \newcommand {\JTB} {{\it J.\ Theor.\ Biol.\ }}
47: \newcommand {\Mac} {{\it Macromolecules\ }}
48: \newcommand {\MC} {{\it Makromol.\ Chem.,\ Theory Simul.\ }}
49: \newcommand {\MP} {{\it Molec.\ Phys.\ }}
50: \newcommand {\NAR} {{\it Nucleic\ Acids\ Res.\ }}
51: \newcommand {\Nat} {{\it Nature}}
52: \newcommand {\NSB} {{\it Nat.\ Struct.\ Biol.\ }}
53: \newcommand {\NP} {{\it Nucl.\ Phys.\ }}
54: \newcommand {\Phy} {{\it Physica\ }}
55: \newcommand {\Pro} {{\it Proteins\ Struct.\ Funct.\ Genet.\ }}
56: \newcommand {\ProEng} {{\it Protein\ Eng.\ }}
57: \newcommand {\ProSci} {{\it Protein\ Sci.\ }}
58: \newcommand {\Pa} {{\it Physica}}
59: \newcommand {\PL} {{\it Phys.\ Lett.\ }}
60: \newcommand {\PNAS} {{\it Proc.\ Natl.\ Acad.\ Sci.\ USA\ }}
61: \newcommand {\PR} {{\it Phys.\ Rev.\ }}
62: \newcommand {\PRL} {{\it Phys.\ Rev.\ Lett.\ }}
63: \newcommand {\PRS} {{\it Proc.\ Roy.\ Soc.\ }}
64: \newcommand {\PS} {{\it Protein\ Sci.\ }}
65: \newcommand {\RMP} {{\it Rev.\ Mod.\ Phys.\ }}
66: \newcommand {\Sci} {{\it Science\ }}
67: \newcommand {\SFD} {{\it Structure with Folding \& Design\ }}
68: \newcommand {\Str} {{\it Structure\ }}
69: \newcommand {\TBS} {{\it Trends Biochem. Sci.\ }}
70: \newcommand {\ZP} {{\it Z.\ Physik}}
71: %
72: \newcommand{\beq}{\begin{equation}}
73: \newcommand{\eeq}{\end{equation}}
74: \newcommand{\beqa}{\begin{eqnarray}}
75: \newcommand{\eeqa}{\end{eqnarray}}
76: \newcommand{\bea}{\begin{eqnarray}}
77: \newcommand{\eea}{\end{eqnarray}}
78: %
79: \newcommand {\ev}[1] {\langle #1\rangle}
80: \newcommand {\Ca} {C${}_{\alpha}$}
81: \newcommand {\Cb} {C${}_{\beta}$}
82: \newcommand {\Cp} {C${}^{\prime}$}
83: \newcommand {\Eloc} {E_{\mbox{{\scriptsize loc}}}}
84: \newcommand {\Cv} {C_{\mbox{{\scriptsize v}}}}
85: \newcommand {\Esa} {E_{\mbox{{\scriptsize sa}}}}
86: \newcommand {\Ehb} {E_{\mbox{{\scriptsize hb}}}}
87: \newcommand {\Eaa} {E_{\mbox{{\scriptsize AA}}}}
88: \newcommand {\ephi} {\epsilon_\phi}
89: \newcommand {\epsi} {\epsilon_\psi}
90: \newcommand {\esa} {\epsilon_{\mbox{{\scriptsize sa}}}}
91: \newcommand {\ehb} {\epsilon_{\mbox{{\scriptsize hb}}}}
92: \newcommand {\eaa} {\epsilon_{\mbox{{\scriptsize AA}}}}
93: \newcommand {\shb} {\sigma_{\mbox{{\scriptsize hb}}}}
94: \newcommand {\saa} {\sigma_{\mbox{{\scriptsize AA}}}}
95: \newcommand {\QFU} {Q_{\mbox{{\scriptsize FU}}}}
96: \newcommand {\QBU} {Q_{\mbox{{\scriptsize BU}}}}
97: \newcommand {\dBU} {\delta_{\mbox{{\scriptsize BU}}}}
98: \newcommand {\dFU} {\delta_{\mbox{{\scriptsize FU}}}}
99: \newcommand {\Rg} {R_{\mbox{{\scriptsize g}}}}
100:
101: \input{psfig}
102:
103: \begin{document}
104:
105: \begin{flushright}
106: LU TP 00-22\\
107: Revised version\\
108: October 29, 2000
109: \end{flushright}
110:
111: \vspace{0.4in}
112:
113: \begin{center}
114: {\LARGE \bf Three-helix-bundle Protein}
115:
116: {\LARGE \bf in a Ramachandran Model}
117:
118: \vspace{.3in}
119:
120: \large
121: Anders Irb\"ack, Fredrik Sjunnesson and
122: Stefan Wallin\footnote{E-mail: irback,\,fredriks,\,stefan@thep.lu.se}\\
123: \vspace{0.10in}
124: Complex Systems Division, Department of Theoretical Physics\\
125: Lund University, S\"olvegatan 14A, S-223 62 Lund, Sweden \\
126: {\tt http://www.thep.lu.se/tf2/complex/}\\
127:
128: \vspace{0.3in}
129:
130: Submitted to \PNAS
131:
132: \end{center}
133: \vspace{0.3in}
134: \normalsize
135: Abstract:\\
136: We study the thermodynamic behavior of a model protein with
137: 54 amino acids that forms a three-helix bundle in its native state.
138: The model contains three types of amino acids and five to six
139: atoms per amino acid and has the Ramachandran torsional angles
140: $\phi_i$, $\psi_i$ as its degrees of freedom. The force field
141: is based on hydrogen bonds and effective hydrophobicity forces.
142: For a suitable choice of the relative strength of these interactions,
143: we find that the three-helix-bundle protein undergoes an abrupt folding
144: transition from an expanded state to the native state. Also shown is
145: that the corresponding one- and two-helix segments are less stable
146: than the three-helix sequence.
147:
148: \newpage
149:
150: \section{Introduction}
151:
152: It is not yet possible to simulate the formation of proteins' native
153: structures on the computer in a controlled way. This goal has been achieved
154: in the context of simple lattice and off-lattice models, where typically
155: each amino acid is represented by a single interaction site corresponding
156: to the \Ca\ atom, and such studies have provided valuable insights into
157: the physical principles of protein
158: folding~\cite{Sali:94,Bryngelson:95,Dill:97,Klimov:98,Nymeyer:98}
159: and the statistical properties of functional protein
160: sequences~\cite{Pande:94,Irback:96}.
161: However, these models have their obvious limitations.
162: Therefore, the search for computationally feasible models
163: with a more realistic chain geometry remains a highly relevant task.
164:
165: In this paper, we discuss a model based on the well-known fact that the
166: main degrees of freedom of the protein backbone are the Ramachandran
167: torsional angles $\phi_i,\psi_i$~\cite{Ramachandran:68}. Each amino acid is
168: represented by five or six atoms, which makes this model computationally
169: slightly more demanding than \Ca\ models. On the other hand, it also
170: makes interactions such as hydrogen bonds easier to define. The
171: formation of native structure is, in this model, driven by hydrogen-bond
172: formation and effective hydrophobicity forces; hydrophobicity is
173: widely held as the most important stability factor in
174: proteins~\cite{Dill:90,Privalov:92}, and hydrogen bonds are essential
175: to properly model the formation of secondary structure.
176:
177: In this model, we study in particular a three-helix-bundle protein with
178: 54 amino acids, which represents a truncated and simplified version of the
179: four-helix-bundle protein {\it de novo} designed by Regan and
180: DeGrado~\cite{Regan:88}.
181: This example was chosen partly because there have been earlier studies
182: of similar-sized helical proteins using models at comparable levels of
183: resolution~\cite{Rey:93,Guo:96,Zhou:97,Koretke:98,Shea:99,Hardin:99,Takada:99}.
184: The behavior of small fast-folding proteins is a current topic in both
185: theoretical and experimental research, and a three-helix-bundle protein
186: that has been extensively studied both
187: experimentally~\cite{Bottomley:94,Bai:97} and
188: theoretically~\cite{Zhou:97,Shea:99,Guo:97,Kolinski:98}
189: is fragment B of staphylococcal protein A.
190:
191: In addition to the three-helix protein, to study size dependence,
192: we also look at the behavior of the corresponding one- and two-helix
193: segments. By using the method of simulated
194: tempering~\cite{Lyubartsev:92,Marinari:92,Irback:95}, a careful
195: study of the thermodynamic properties of these different chains
196: is performed.
197:
198: Not unexpectedly, it turns out that the behavior of the model
199: depends strongly on the relative strength of the hydrogen-bond
200: and hydrophobicity terms. In fact, the situation is somewhat
201: reminiscent of what has been found for homopolymers with
202: stiffness~\cite{Kolinski:86,Doniach:96,Bastolla:97,Doye:98},
203: with hydrogen bonds playing the role of the stiffness term.
204: Throughout this paper, we focus on one specific empirical choice of
205: these parameters.
206:
207: For this choice of parameters, we find
208: that the three-helix-bundle protein has the
209: following three properties. First, it does form a stable
210: three-helix bundle (except for a 2-fold topological degeneracy).
211: Second, its folding transition is abrupt, from an expanded
212: state to the native three-helix-bundle state. Third, compared to
213: the one- and two-helix segments, it forms a more stable secondary
214: structure. It should be stressed that these properties are found
215: without resorting to the popular G\=o approximation~\cite{Go:78},
216: in which interactions that do not favor the desired structure
217: are ignored.
218:
219: \section{The Model}
220:
221: The model we study is a reduced off-lattice model. The chain
222: representation is illustrated in Fig.~\ref{fig:1}. As mentioned in the
223: introduction, each amino acid is represented by five or six atoms.
224: The three backbone atoms N, \Ca\ and \Cp\ are all included.
225: Also included are the H and O atoms shown in Fig.~\ref{fig:1}, which we
226: use to define hydrogen bonds. Finally, the side chain is represented by
227: a single atom, \Cb, which can be hydrophobic, polar or absent.
228: This gives us the following three types of amino acids: A with
229: hydrophobic \Cb, B with polar \Cb, and G (glycine) without \Cb.
230:
231: \begin{figure}
232: \vspace{0mm}
233: \hspace{40mm}
234: \rotatebox{270}{\epsfig{figure=fig1.eps,width=5cm,height=6.2cm}}
235: \vspace{5mm}
236: \caption{Schematic figure showing the representation of one amino acid.}
237: \label{fig:1}
238: \end{figure}
239:
240: The H, O and \Cb\ atoms are all attached to the backbone in a rigid
241: way. Furthermore, in the backbone, all bond lengths, bond angles and
242: peptide torsional angles ($180^\circ$) are held fixed.
243: This leaves us with two degrees of freedom per amino acid, the
244: Ramachandran torsional angles $\phi_i$ and $\psi_i$ (see Fig.~\ref{fig:1}).
245: The parameters held fixed can be found in Table~\ref{tab:1}.
246:
247: \begin{table}[t]
248: \begin{center}
249: \begin{tabular}{llll}
250: \multicolumn{2}{l}{Bond lengths (\AA)}&
251: \multicolumn{2}{l}{Bond angles (${}^\circ$)}\\
252: \hline
253: N\Ca & 1.46 & \Cp N\Ca & 121.7 \\
254: \Ca\Cp & 1.52 & N\Ca\Cp & 111.0 \\
255: \Cp N & 1.33 & \Ca\Cp N & 116.6 \\
256: NH & 1.03 & N\Ca\Cb & 110.0 \\
257: \Ca\Cb & 1.53 & \Cp\Ca\Cb & 110.0 \\
258: \Cp O & 1.23 & & \\
259: \end{tabular}
260: \caption{Geometry parameters.}
261: \label{tab:1}
262: \end{center}
263: \end{table}
264:
265: Our energy function
266: \beq
267: E=\Eloc+\Esa+\Ehb+\Eaa
268: \label{e}\eeq
269: is composed of four terms.
270: The local potential $\Eloc$ has a standard form with 3-fold symmetry,
271: \beq
272: \Eloc=\frac{\ephi}{2}\sum_i(1 + \cos3\phi_i)
273: + \frac{\epsi}{2}\sum_i(1 + \cos3\psi_i)\,.
274: \eeq
275: The self-avoidance term $\Esa$ is given by a hard-sphere potential
276: of the form
277: \beq
278: \Esa=\esa\mathop{{\sum}'}_{i<j}
279: \bigg(\frac{\sigma_{ij}}{r_{ij}}\bigg)^{12}\,,
280: \label{sa}\eeq
281: where the sum runs over all possible atom pairs except those
282: consisting of two hydrophobic \Cb. The hydrogen-bond term $\Ehb$
283: is given by
284: \beq
285: \Ehb= \ehb \sum_{ij}u(r_{ij})v(\alpha_{ij},\beta_{ij})\,,
286: \label{hb}\eeq
287: where
288: \begin{eqnarray}
289: u(r_{ij})&=& 5\bigg(\frac{\shb}{r_{ij}}\bigg)^{12} -
290: 6\bigg(\frac{\shb}{r_{ij}}\bigg)^{10}\\
291: v(\alpha_{ij},\beta_{ij})&=&\left\{
292: \begin{array}{ll}
293: \cos^2\alpha_{ij}\cos^2\beta_{ij} & \ \alpha_{ij},\beta_{ij}>90^{\circ}\\
294: 0 & \ \mbox{otherwise}
295: \end{array} \right.
296: \end{eqnarray}
297: In Eq.~\ref{hb} $i$ and $j$ represent H and O atoms, respectively,
298: and $r_{ij}$ denotes the HO distance, $\alpha_{ij}$ the NHO
299: angle, and $\beta_{ij}$ the HO\Cp\ angle. Any HO pair can form a
300: hydrogen bond. The last term in Eq.~\ref{e}, the hydrophobicity
301: term $\Eaa$, has the form
302: \beq
303: \Eaa=\eaa\sum_{i<j}\bigg[
304: \bigg(\frac{\saa}{r_{ij}}\bigg)^{12}
305: -2\bigg(\frac{\saa}{r_{ij}}\bigg)^6\,\bigg]\,,
306: \eeq
307: where both $i$ and $j$ represent hydrophobic \Cb. To speed up the simulations,
308: a cutoff radius $r_c$ is used,\footnote{The cutoff procedure is
309: $f(r)\mapsto\tilde f(r)$ where $\tilde f(r)=f(r)-f(r_c)-(r-r_c)f^\prime(r_c)$
310: if $r<r_c$ and $\tilde f(r)=0$ otherwise.} which is 4.5\AA\ for $\Esa$
311: and $\Ehb$, and 8\AA\ for $\Eaa$.
312:
313: In this energy function, roughly speaking, the first two terms,
314: $\Eloc$ and $\Esa$, enforce steric constraints,
315: whereas the last two terms, $\Ehb$ and $\Eaa$, are the
316: ones responsible for stability.
317: Force fields similar in spirit, emphasizing hydrogen bonding
318: and hydrophobicity, have been used with some success to predict
319: structures of peptides~\cite{Ishikawa:99} and small helical
320: proteins~\cite{Koretke:98}.
321:
322: The parameters of our energy function were determined largely by
323: trial and error. The final parameters are listed in
324: Table~\ref{tab:2}. The parameters $\sigma_{ij}$ of Eq.~\ref{sa}
325: are given by
326: \[
327: \sigma_{ij}=\sigma_{i}+\sigma_j+\Delta\sigma_{ij}\,,
328: \]
329: where $\sigma_i,\sigma_j$ can be found in Table~\ref{tab:2}, and
330: $\Delta\sigma_{ij}$ is zero except for \Cb\Cp, \Cb N and \Cb O pairs
331: that are connected by three covalent bonds. In these three cases, we put
332: $\Delta\sigma_{ij}=0.625$\AA . This could equivalently be described
333: as a change of the local $\phi_i$ and $\psi_i$ potentials. In
334: Fig.~\ref{fig:2}, we show $\phi_i,\psi_i$ scatter plots for nonglycine (A
335: and B) and glycine for our final parameters, which are in
336: good qualitative agreement with the $\phi_i,\psi_i$ distributions
337: of real proteins~\cite{Ramachandran:68,Zimmerman:77}.
338:
339: Finally, we determined the strengths of the hydrogen-bond and
340: hydrophobicity terms on the basis of the resulting overall thermodynamic
341: behavior of the three-helix sequence. For this purpose, we performed a
342: set of trial runs for fixed values of the other parameters. An alternative
343: would have been to use the method of Shea~\etal~\cite{Shea:98}.
344: The result of our empirical determination of $\ehb$ and $\eaa$ does
345: not seem unreasonable; at the folding temperature of the three-helix
346: sequence (see below), we get $\ehb/kT\approx4.3$ and $\eaa/kT\approx3.4$.
347:
348: \begin{table}[t]
349: \begin{center}
350: \begin{tabular}{ccccccccccccc}
351: \multicolumn{5}{c}{}&\multicolumn{6}{c}{$\sigma_i$(\AA)}\\
352: \cline{6-11}
353: $\ephi$ & $\epsi$ & $\esa$ & $\ehb$ & $\eaa$ & N & \Ca & \Cp & H & \Cb & O &
354: $\shb$ (\AA) & $\saa$ (\AA) \\
355: \hline
356: 1 & 1 & 0.0034 & 2.8 & 2.2 & 1.65 & 1.85 & 1.85 & 1.0 & 2.5 & 1.65 & 2.0 & 5.0
357: \end{tabular}
358: \caption{Parameters of the energy function. Energies are
359: in dimensionless units, in which the folding transition occurs
360: at $kT\approx0.65$ for the three-helix-bundle protein (see below).}
361: \label{tab:2}
362: \end{center}
363: \end{table}
364:
365: \begin{figure}
366: \vspace{-40mm}
367: \mbox{
368: \hspace{-30mm}
369: \psfig{figure=fig2a.ps,width=10cm,height=14.1cm}
370: \hspace{-30mm}
371: \psfig{figure=fig2b.ps,width=10cm,height=14.1cm}
372: }
373: \vspace{-40mm}
374: \caption{$\phi_i,\psi_i$ scatter plots for nonglycine and glycine, as obtained
375: by simulations of the chains GXG for X=A/B and X=G, respectively, at $kT=0.625$
376: (shown is $\phi_i,\psi_i$ for X).}
377: \label{fig:2}
378: \end{figure}
379:
380: In this model, we study the three sequences shown in Table~\ref{tab:3},
381: which contain 16, 35 and 54 amino acids, respectively. Following the
382: strategy of Regan and DeGrado~\cite{Regan:88}, the A and B amino acids
383: are distributed along the sequence 1H in such a way that this segment
384: can form a helix with all hydrophobic amino acids on the same side.
385: The sequence 3H, consisting of three such stretches of As and Bs
386: plus two GGG segments, is meant to form a three-helix bundle.
387: This particular sequence was recently studied by
388: Takada~\etal~\cite{Takada:99}, who used a more elaborate model
389: with nonadditive forces.
390:
391: \begin{table}
392: \begin{center}
393: \begin{tabular}{ll}
394: 1H:& BBABBAABBABBAABB\\
395: 2H:& 1H--GGG--1H\\
396: 3H:& 1H--GGG--1H--GGG--1H
397: \end{tabular}
398: \caption{The sequences studied.}
399: \label{tab:3}
400: \end{center}
401: \end{table}
402:
403: \section{Results}
404:
405: To study the thermodynamic behavior of the chains described in the
406: previous section, we use the method of simulated tempering.
407: This means that we first select a set of allowed temperatures and
408: then perform simulations in which the temperature is a dynamical
409: variable. This is done to speed up low-temperature simulations. In addition,
410: it provides a convenient method for calculating free energies.
411:
412: An example of a simulated-tempering run is given in Fig.~\ref{fig:3},
413: which shows the Monte Carlo evolution of the energy $E$ and
414: radius of gyration $\Rg$ (calculated over all backbone atoms) in a
415: simulation of the three-helix sequence.
416: Also shown, bottom panel, is how the system jumps between the
417: different temperatures. Two distinct types of behavior can be
418: seen. In one case, $E$ is high, fluctuations in size are large,
419: and the temperatures visited are high. In the other case, $E$ is low,
420: the size is small and almost frozen, and the temperatures visited
421: are low. Interesting to note is that there is one temperature, the
422: next-lowest one, which is visited in both cases. Apparently, both
423: types of behavior are possible at this temperature.
424:
425: \begin{figure}
426: \vspace{-33mm}
427: \hspace{20mm}
428: \psfig{figure=fig3.ps,width=10cm,height=14.1cm}
429: \vspace{-17mm}
430: \caption{Monte Carlo evolution of the energy and radius of gyration
431: in a typical simulation of the three-helix sequence. The bottom panel shows
432: how the system jumps between the allowed temperatures $T_j$, which
433: are given by
434: $T_j=T_{\min}(T_{\max}/T_{\min})^{(j-1)/(J-1)}$~\protect\cite{Hansmann:97}
435: with $kT_{\min}=0.625$, $kT_{\max}=0.9$ and $J=8$. The temperature $T_{\min}$
436: is chosen to lie just below the collapse transition, whereas $T_{\max}$ is well
437: into the coil phase (see Fig.~\protect\ref{fig:4}).}
438: \label{fig:3}
439: \end{figure}
440:
441: In Fig.~\ref{fig:4}a we show the specific heat as a function of
442: temperature for the one-, two- and three-helix sequences.
443: A pronounced peak can be seen that gets stronger with
444: increasing chain length. In fact, the increase in
445: height is not inconsistent with a linear dependence on chain
446: length, which is what one would have expected if it had been a
447: conventional first-order phase transition with a latent heat.
448:
449: \begin{figure}
450: \vspace{-42mm}
451: \mbox{
452: \hspace{-51mm}
453: \psfig{figure=fig4a.ps,width=10cm,height=14.1cm}
454: \hspace{-51mm}
455: \psfig{figure=fig4b.ps,width=10cm,height=14.1cm}
456: \hspace{-51mm}
457: \psfig{figure=fig4c.ps,width=10cm,height=14.1cm}
458: }
459: \vspace{-47mm}
460: \caption{Thermodynamic functions against temperature for the
461: sequences 1H ($\diamond$), 2H ($\times$) and 3H ($+$) in
462: Table~\ref{tab:3}. (a) Specific heat $\Cv=(\ev{E^2}-\ev{E}^2)/NkT^2$,
463: $N$ being the number of amino acids.
464: (b) Hydrogen-bond energy per amino acid, $\Ehb/N$.
465: (c) Chain entropy per amino acid, $\delta S/N=[S-S(kT=0.9)]/N$.
466: The full lines in (a) represent single-histogram
467: extrapolations~\protect\cite{Ferrenberg:88}. Dotted lines are
468: drawn to guide the eye.}
469: \label{fig:4}
470: \end{figure}
471:
472: Our results for the radius of gyration (not displayed) show that
473: the specific heat maximum can be viewed as the collapse temperature.
474: The specific heat maximum is also where hydrogen-bond formation
475: occurs, as can be seen from Fig.~\ref{fig:4}b. Important to note
476: in this figure is that the decrease in hydrogen-bond energy
477: {\it per amino acid} with decreasing temperature is most rapid for
478: the three-helix sequence, which implies that, compared to the shorter ones,
479: this sequence forms more stable secondary structure. The results
480: for the chain entropy shown in Fig.~\ref{fig:4}c provide further
481: support for this; the entropy loss per amino acid with decreasing
482: temperature is largest for the three-helix sequence.
483:
484: It should be stressed that the character of the collapse transition
485: depends strongly on the relative strength of the hydrogen-bond
486: and hydrophobicity terms. Figure~\ref{fig:4} shows that the transition is
487: very abrupt or ``first-order-like'' for our choice $(\ehb,\eaa)=(2.8,2.2)$.
488: A fairly small decrease of $\ehb/\eaa$ is sufficient to get a very
489: different behavior with, for example, a much weaker peak in the specific
490: heat. In this case, the chain collapses to a molten globule without specific
491: structure rather than to a three-helix bundle. A substantially weakened
492: transition was observed for $\ehb=\eaa=2.5$.
493: If, on the other hand, $\ehb/\eaa$ is too large,
494: then it is evident that the chain will form one long helix instead
495: of a helical bundle.
496:
497: We now turn to the three-dimensional structure of the three-helix
498: sequence in the collapsed phase. It turns out that it does form a
499: three-helix bundle. This bundle can have two distinct topologies: if we
500: let the first two helices form a U, then the third helix can be either
501: in front of or behind that U. The model is, not unexpectedly,
502: unable to discriminate between these two possibilities. To characterize
503: low-temperature conformations, we therefore determined two representative
504: structures, one for each topology, which, following~\cite{Takada:99},
505: are referred to as FU and BU, respectively. These structures
506: are shown in Fig.~\ref{fig:5}. They were generated by
507: quenching a large number of low-$T$ structures to
508: zero temperature, and we feel convinced that they provide good
509: approximations of the energy minima for the respective topologies.
510: Given an arbitrary conformation, we then measure the root-mean-square
511: distances $\delta_i$ ($i=$\,FU,\,BU) to these two structures (calculated
512: over all backbone atoms). These distances are
513: converted into similarity parameters $Q_i$ by using
514: \beq
515: Q_i=\exp(-\delta_i^2/100\mbox{\AA}^2)\,.
516: \label{sim}\eeq
517:
518: \begin{figure}
519: \vspace{-00mm}
520: \mbox{
521: \hspace{40mm}
522: \epsfig{figure=fig5a.eps,width=1.53cm,height=2.97cm}
523: \hspace{30mm}
524: \epsfig{figure=fig5b.eps,width=1.53cm,height=2.97cm}
525: }
526: \caption{Representative low-temperature structures, FU and BU, respectively.
527: Drawn with RasMol~\protect\cite{Sayle:95}.}
528: \label{fig:5}
529: \end{figure}
530:
531: At temperatures above the specific heat maximum, both $Q_i$ tend
532: to be small. At temperatures below this point, the system is
533: found to spend most of its time close to one or the other of the
534: representative structures; either $\QFU$ or $\QBU$ is close to 1.
535: Finally, at the peak, all three of these regions in the $\QFU,\QBU$
536: plane are populated, as can be seen from Fig.~\ref{fig:6}a. In particular,
537: this implies that the folding transition coincides with the
538: specific heat maximum.
539:
540: \begin{figure}
541: \vspace{-40mm}
542: \mbox{
543: \hspace{-30mm}
544: \psfig{figure=fig6a.ps,width=10cm,height=14.1cm}
545: \hspace{-30mm}
546: \psfig{figure=fig6b.ps,width=10cm,height=14.1cm}
547: }
548: \vspace{-40mm}
549: \caption{(a) $\QFU,\QBU$ (see Eq.~\protect\ref{sim}) scatter plot
550: at the specific heat maximum ($kT=0.658$).
551: (b) Free energy $F(Q)$ as a function of $Q=\max(\QFU,\QBU)$ at
552: the same temperature.}
553: \label{fig:6}
554: \end{figure}
555:
556: The folding transition can be described in terms of a single
557: ``order parameter'' by taking $Q=\max(\QFU,\QBU)$ as a
558: measure of nativeness. Correspondingly, we put
559: $\delta=\min(\dFU,\dBU)$. In Fig.~\ref{fig:6}b,
560: we show the free-energy profile $F(Q)$ at the folding temperature.
561: The free energy has a relatively sharp minimum at $Q\approx0.9$,
562: corresponding to $\delta\approx3$\AA. This is followed by a weak
563: barrier around $Q=0.7$, corresponding to $\delta\approx6$\AA.
564: Finally, there is a broad minimum at small $Q$, where $Q=0.2$ corresponds
565: to $\delta\approx13$\AA.
566:
567: What does the nonnative population at the folding temperature
568: correspond to in terms of $\Rg$ and $\Ehb$? This can be seen from
569: the $Q,\Rg$ and $Q,\Ehb$ scatter plots in Fig.~\ref{fig:7}.
570: These plots show that the low-$Q$ minimum of $F(Q)$ corresponds to
571: expanded structures with a varying but not high
572: secondary-structure content. Although a detailed kinetic study
573: is beyond the scope of this paper, we furthermore note that the
574: free-energy surfaces corresponding to the distributions in
575: Fig.~\ref{fig:7} are relatively smooth. Consistent with that, we
576: found that standard fixed-temperature Monte Carlo simulations were able to
577: reach the native state, starting from random coils.
578:
579: \begin{figure}
580: \vspace{-40mm}
581: \mbox{
582: \hspace{-30mm}
583: \psfig{figure=fig7a.ps,width=10cm,height=14.1cm}
584: \hspace{-30mm}
585: \psfig{figure=fig7b.ps,width=10cm,height=14.1cm}
586: }
587: \vspace{-40mm}
588: \caption{(a) $Q,\Rg$ and (b) $Q,\Ehb$ scatter plots at the
589: folding temperature ($kT=0.658$).}
590: \label{fig:7}
591: \end{figure}
592:
593: Let us finally mention that we also performed simulations of some
594: random sequences with the same length and composition as the
595: three-helix sequence. The random sequences did not form stable
596: structures and collapsed more slowly with decreasing temperature
597: than the designed three-helix sequence.
598:
599: \section{Summary and Outlook}
600:
601: We have studied a reduced protein model where the formation
602: of native structure is driven by a competition between hydrogen
603: bonds and effective hydrophobicity forces. Using this force field,
604: we find that the three-helix-bundle protein studied
605: has the following properties:
606: \begin{itemize}
607: \item It does form a stable three-helix-bundle state, except
608: for a 2-fold topological degeneracy.
609: \item It undergoes an abrupt folding transition from an expanded
610: state to the native state.
611: \item It forms more stable secondary structure than the
612: corresponding one- and two-helix segments.
613: \end{itemize}
614: An obvious question that remains to be addressed is what is
615: needed to lift the topological degeneracy. Not obvious, however,
616: is whether this question should be addressed at the present
617: level of modeling, before including full side chains.
618:
619: A first-order-like
620: folding transition that takes the system directly from the
621: unfolded state to the native one is what one expects
622: for small fast-folding proteins. For the model to
623: show this behavior, careful tuning of the relative
624: strength of the hydrogen-bond and hydrophobicity terms, $\ehb/\eaa$,
625: is required. This $\ehb/\eaa$ dependence may at first glance seem unwanted
626: but is not physically unreasonable; $\ehb$ can be thought of partly as a
627: stiffness parameter, and chain stiffness has important implications
628: for the phase structure, as shown by recent work on
629: homopolymers~\cite{Kolinski:86,Doniach:96,Bastolla:97,Doye:98}.
630: Note also that incorporation of full side chains makes the chains
631: intrinsically stiffer, which might lead to a weaker
632: $\ehb/\eaa$ dependence.
633:
634: Our three-helix sequence has previously been studied by
635: Takada~\etal~\cite{Takada:99}, who used a more elaborate force
636: field. It was suggested that it is essential to use
637: context-dependent hydrogen bonds for the three-helix-bundle
638: protein to make more stable secondary structure than its one-helix fragments.
639: Our model shows this behavior, although its hydrogen bonds are
640: context-independent.
641:
642: Let us finally stress that we find a first-order-like
643: folding transition without using the G\=o approximation.
644: Evidence for first-order-like folding transitions has been found
645: for proteins with similar lengths in some \Ca\
646: models~\cite{Zhou:97,Nymeyer:98,Shea:99,Shea:98}, but these
647: studies use this approximation.
648:
649: \section*{Acknowledgements}
650:
651: This work was in part supported by the Swedish Foundation for Strategic
652: Research.
653:
654: \newpage
655:
656: %\bibliography{/home/people/irback/tex/bib/refs}
657: %\bibliographystyle{unsrt}
658:
659: \begin{thebibliography}{}
660:
661: \bibitem{Sali:94}
662: S\u{a}li,~A., Shakhnovich,~E. \& Karplus,~M. (1994)
663: ``Kinetics of Protein Folding: A Lattice Model Study
664: of the Requirements for Folding to the Native State'',
665: \JMB {\bf 235}, 1614--1636.
666:
667: \bibitem{Bryngelson:95}
668: Bryngelson,~J.D., Onuchic,~J.N., Socci,~N.D. \& Wolynes,~P.G. (1995)
669: ``Funnels, Pathways, and the Energy Landscape of Protein
670: Folding: A Synthesis'',
671: \Pro {\bf 21}, 167--195.
672:
673: \bibitem{Dill:97}
674: Dill,~K.A. \& Chan,~H.S. (1997)
675: ``From Levinthal to Pathways to Funnels'',
676: \NSB {\bf 4}, 10--19.
677:
678: \bibitem{Klimov:98}
679: Klimov,~D.K. \& Thirumalai,~D. (1998)
680: ``Linking Rates of Folding in Lattice Models of Proteins
681: with Underlying Thermodynamic Characteristics'',
682: \JCP {\bf 109}, 4119--4125.
683:
684: \bibitem{Nymeyer:98}
685: Nymeyer,~H., Garc\'\i a,~A.E. \& Onuchic,~J.N. (1998)
686: ``Folding Funnels and Frustration in Off-lattice Minimalist
687: Protein Landscapes'',
688: \PNAS {\bf 95}, 5921--5928.
689:
690: \bibitem{Pande:94}
691: Pande,~V.S., Grosberg,~A.Y. \& Tanaka,~T. (1994)
692: ``Thermodynamic Procedure to Synthesize Heteropolymers that
693: Can Renature to Recognize a Given Target Molecule'',
694: \PNAS {\bf 91}, 12976--12979.
695:
696: \bibitem{Irback:96}
697: Irb\"ack,~A., Peterson,~C. and Potthast,~F. (1996)
698: ``Evidence for Nonrandom Hydrophobicity Structures in Protein Chains'',
699: \PNAS {\bf 93}, 9533--9538.
700:
701: \bibitem{Ramachandran:68}
702: Ramachandran,~G.N. \& Sasisekharan,~V. (1968)
703: ``Conformation of Polypeptides and Proteins'',
704: \APC {\bf 23}, 283--437.
705:
706: \bibitem{Dill:90}
707: Dill,~K.A. (1990)
708: ``Dominant Forces in Protein Folding'',
709: \Bioch {\bf 29}, 7133--7155.
710:
711: \bibitem{Privalov:92}
712: Privalov,~P.L. (1992)
713: ``Physical Basis of the Stability of the Folded Conformations of Proteins'',
714: in {\it Protein Folding}, ed. Creighton,~T.E. (Freeman, New York), pp. 83--126.
715:
716: \bibitem{Regan:88}
717: Regan,~L. \& DeGrado,~W.F. (1988)
718: ``Characterization of a Helical Protein Designed from First Principles'',
719: \Sci {\bf 241}, 976--978.
720:
721: \bibitem{Rey:93}
722: Rey,~A. \& Skolnick,~J. (1993)
723: ``Computer Modeling and Folding of Four-helix Bundles'',
724: \Pro {\bf 16}, 8--28.
725:
726: \bibitem{Guo:96}
727: Guo,~Z. \& Thirumalai,~D. (1996)
728: ``Kinetics and Thermodynamics of Folding of a {\it de Novo}
729: Designed Four-helix Bundle Protein'',
730: \JMB {\bf 263}, 323--343.
731:
732: \bibitem{Zhou:97}
733: Zhou,~Z. \& Karplus,~M. (1997)
734: ``Folding Thermodynamics of a Model Three-helix-bundle Protein'',
735: \PNAS {\bf 94}, 14429--14432.
736:
737: \bibitem{Koretke:98}
738: Koretke,~K.K., Luthey-Schulten,~Z. \& Wolynes,~P.G. (1998)
739: ``Self-consistently Optimized Energy Functions for Protein Structure
740: Prediction by Molecular Dynamics'',
741: \PNAS {\bf 95}, 2932--2937.
742:
743: \bibitem{Hardin:99}
744: Hardin,~C., Luthey-Schulten,~Z. \& Wolynes,~P.G. (1999)
745: ``Backbone Dynamics, Fast Folding, and Secondary Structure Formation
746: in Helical Proteins and Peptides'',
747: \Pro {\bf 34}, 281--294.
748:
749: \bibitem{Shea:99}
750: Shea,~J.-E., Onuchic,~J.N. \& Brooks,~C.L.,~III (1999)
751: ``Exploring the Origins of Topological Frustration:
752: Design of a Minimally Frustrated Model of Fragment B of Protein A'',
753: \PNAS {\bf 96}, 12512--12517.
754:
755: \bibitem{Takada:99}
756: Takada,~S., Luthey-Schulten,~Z. \& Wolynes,~P.G. (1999)
757: ``Folding Dynamics with Nonadditive Forces: A Simulation
758: Study of a Designed Helical Protein and a Random Heteropolymer'',
759: \JCP {\bf 110}, 11616--11629.
760:
761: \bibitem{Bottomley:94}
762: Bottomley,~S.P., Popplewell,~A.G., Scawen,~M., Wan,~T., Sutton,~B.J.
763: \& Gore,~M.G. (1994)
764: ``The Stability and Unfolding of an IgG Binding Protein Based upon
765: the B Domain of Protein A from {\it Staphylococcus Aureus} Probed
766: by Tryptophan Substitution and Fluorescence Spectroscopy'',
767: \ProEng {\bf 7}, 1463--1470.
768:
769: \bibitem{Bai:97}
770: Bai,~Y., Karimi,~A., Dyson,~H.J. \& Wright,~P.E. (1997)
771: ``Absence of a Stable Intermediate on the Folding Pathway of Protein A''
772: \ProSci {\bf 6}, 1449--1457.
773:
774: \bibitem{Guo:97}
775: Guo,~Z., Brooks,~C.L.,~III \& Boczko,~E.M. (1997)
776: ``Exploring the Folding Free Energy Surface of a Three-helix
777: Bundle Protein'',
778: \PNAS {\bf 94}, 10161-10166.
779:
780: \bibitem{Kolinski:98}
781: Kolinski,~A., Galazka,~W. \& Skolnick,~J. (1998)
782: ``Monte Carlo Studies of the Thermodynamics and Kinetics of Reduced
783: Protein Models: Application to Small helical, $\beta$ and $\alpha/\beta$
784: Proteins'',
785: \JCP {\bf 108}, 2608--2617.
786:
787: \bibitem{Lyubartsev:92}
788: Lyubartsev,~A.P., Martsinovski,~A.A., Shevkunov,~S.V. \&
789: Vorontsov-Velyaminov,~P.V. (1992)
790: ``New Approach to Monte Carlo Calculation of the Free Energy:
791: Method of Expanded Ensembles'',
792: \JCP {\bf 96}, 1776--1783.
793:
794: \bibitem{Marinari:92}
795: Marinari,~E. \& Parisi,~G. (1992)
796: ``Simulated Tempering: A New Monte Carlo Scheme'',
797: \EL {\bf 19}, 451--458.
798:
799: \bibitem{Irback:95}
800: Irb\"ack,~A. \& Potthast,~F. (1995)
801: ``Studies of an Off-lattice Model for Protein Folding: Sequence
802: Dependence and Improved Sampling at Finite Temperature'',
803: \JCP {\bf 103}, 10298--10305.
804:
805: \bibitem{Kolinski:86}
806: Kolinski, A., Skolnick, J. \& Yaris, R. (1986)
807: ``Monte Carlo Simulations on an Equilibrium Globular Protein Folding Model'',
808: \PNAS {\bf 83}, 7267--7271.
809:
810: \bibitem{Doniach:96}
811: Doniach,~S., Garel,~T. \& Orland, H. (1996)
812: ``Phase Diagram of a Semiflexible Polymer Chain in a $\theta$ Solvent:
813: Application to Protein Folding'',
814: \JCP {\bf 105}, 1601--1608.
815:
816: \bibitem{Bastolla:97}
817: Bastolla,~U. \& Grassberger, P. (1997)
818: ``Phase Transitions of Single Semi-stiff Polymer Chains'',
819: \JSP {\bf 89}, 1061--1078.
820:
821: \bibitem{Doye:98}
822: Doye,~J.P.K., Sear,~R.P. \& Frenkel,~D. (1998)
823: ``The Effect of Chain Stiffness on the Phase Behaviour of Isolated
824: Homopolymers'',
825: \JCP {\bf 108}, 2134--2142.
826:
827: \bibitem{Go:78}
828: G\=o,~N. \& Taketomi,~H. (1978)
829: ``Respective Roles of Short- and Long-range Interactions
830: in Protein Folding'',
831: \PNAS {\bf 75}, 559--563.
832:
833: \bibitem{Ishikawa:99}
834: Ishikawa,~K., Yue,~K. \& Dill,~K.A. (1999)
835: ``Predicting the Structures of 18 Peptides Using Geocore'',
836: \ProSci {\bf 8}, 716--721.
837:
838: \bibitem{Zimmerman:77}
839: Zimmerman,~S.S., Pottle,~M.S., N\'emethy,~G. \& Scheraga,~H.A. (1977)
840: ``Conformational Analysis of the 20 Naturally Occurring Amino
841: Acid Residues Using ECEPP'',
842: \Mac {\bf 10}, 1--9.
843:
844: \bibitem{Shea:98}
845: Shea,~J.-E., Nochomovitz,~Y.D., Guo,~Z. \& Brooks,~C.L.,~III (1998)
846: ``Exploring the Space of Protein Folding Hamiltonians:
847: The Balance of Forces in a Minimalist $\beta$-barrel Model'',
848: \JCP {\bf 109}, 2895--2903.
849:
850: \bibitem{Hansmann:97}
851: Hansmann,~U.H.E. \& Okamoto,~Y. (1997)
852: ``Numerical Comparisons of Three Recently Proposed Algorithms
853: in the Protein Folding Problem'',
854: \JCC {\bf 18}, 920--933.
855:
856: \bibitem{Ferrenberg:88}
857: Ferrenberg,~A.M. \& Swendsen,~R.H. (1988)
858: ``New Monte Carlo for Studying Phase Transitions''
859: \PRL {\bf 61}, 2635--2638, and erratum (1989) {\bf 63}, 1658,
860: and references given in the erratum.
861:
862: \bibitem{Sayle:95}
863: Sayle,~R. \& Milner-White,~E.J. (1995)
864: ``RasMol: Biomolecular Graphics for All'',
865: \TBS {\bf 20}, 374--376.
866:
867: \end{thebibliography}
868:
869: \end{document}
870:
871:
872:
873:
874:
875: