1: \documentclass[11pt,letterpaper]{article}
2: \usepackage{latexsym}
3: \usepackage{epsfig}
4: \usepackage{graphicx}
5: \usepackage{amssymb}
6: %%\DeclareGraphicsRule{.jpg}{eps}{*}{jpeg2ps}
7: \setlength{\oddsidemargin}{0.2in}
8: \setlength{\textwidth}{6.25in}
9: \setlength{\textheight}{8.75in}
10: \setlength{\topmargin}{-0.8in}
11:
12: \begin{document}
13:
14: \newcommand {\eqn}[1] {eqn.~(\ref{#1})}
15: \newcommand {\fig}[1] {fig.~\ref{#1}}
16:
17: \begin{center}
18: {\LARGE Force steps during viral DNA packaging ?} \\
19: \vspace{0.5cm}
20: \textsc{Prashant K. Purohit${}^*$, Jan\'{e} Kondev${}^\dag$,
21: and Rob Phillips${}^*$ } \\
22: \vspace{0.25cm}
23: \emph{${}^*$Division of Engineering and Applied Science, California
24: Institute of Technology, Pasadena, CA 91125. \\
25: ${}^\dag$Physics Department, Brandeis University, Waltham, MA 02454.} \\
26: \vspace{0.7cm}
27: \end{center}
28:
29:
30: \begin{abstract}
31: Biophysicists and structural biologists increasingly acknowledge the role
32: played by the mechanical properties of macromolecules as a critical
33: element in many biological processes. This change has been brought about,
34: in part, by the advent of single molecule biophysics techniques that have made it
35: possible to exert piconewton forces on key macromolecules and observe
36: their deformations at nanometer length scales, as well as to observe the
37: mechanical action of macromolecules such as molecular motors. This has
38: opened up immense
39: possibilities for a new generation of mechanical investigations that
40: will respond to such measurements in an attempt to develop
41: a coherent theory for the mechanical behavior of macromolecules
42: under conditions where thermal and chemical effects are on an equal
43: footing with deterministic forces. This paper presents an application
44: of the principles of mechanics to the problem of DNA packaging, one of the
45: key events in the life cycle of bacterial viruses with special reference
46: to the nature of the internal forces that are built up during the DNA
47: packaging process.
48: \end{abstract}
49:
50: \section{Introduction}
51: Mechanics has a long and rich tradition of reaching out to other fields.
52: For example, the mechanical analysis of thin films has revealed insights into
53: applications ranging from microelectronics to MEMS (see Freund and Suresh (2003)).
54: Mechanics arguments have demonstrated that dislocations and surface
55: instabilities play a key role in determining the mechanical and
56: electrical properties of such films. While the study of thin films
57: is a relatively new area to
58: have been influenced by mechanics there are others, like geology and
59: metallurgy, that have drawn on mechanics for decades. The most recent
60: developments in intersonic crack propagation (Rosakis (2002), Rice (2001))
61: have direct
62: analogues in the study of motion along geological fault lines. Geophysicists
63: have relied heavily on data emerging from shock compression of solids to
64: model the thermomechanical behavior of planets (see Meyers (1994)). In addition, the impact of dislocation
65: mechanics, phase transformations and fracture mechanics on the science
66: of metallurgy can scarcely be overestimated.
67:
68: Another field that promises to be visited by mechanicians with increasing
69: regularity is biology, though
70: the use of mechanical principles for both fluids and solids in understanding
71: biological processes is not at all new. For example, slender body theory from fluid
72: mechanics has been fruitfully applied to the study of motion in bacteria
73: and other microorganisms (see Bray (2001)). In addition, variational principles
74: for determining
75: optimal shapes have been used for the study of red-blood cells (see Boal (2002)).
76: Evans and Skalak (1980) show how lipid bilayer vesicles can be modeled as
77: shells and membranes with no shear resistance. The flagella of bacteria have been
78: idealized as rods in an effort to study how their rotation can give rise
79: to propulsion (Goldstein {\it et al.} (1998)).
80: The cytoskeletal framework of the cell is studied as a three dimensional
81: network of elements to obtain estimates of the response of cells to external
82: forces (see Boal (2002)). In this paper we restrict our attention to problems
83: in biological nanomechanics with the observation that there are a range of
84: fascinating connections between mechanics and biology at larger scales as well.
85: One of the most intriguiging features of problems in biological nanomechanics
86: is that they involve a rich interplay of statistical and deterministic forces
87: as will be seen below.
88:
89: It is only recently, however, that structural biologists have begun to understand
90: the significance of mechanical properties for the macromolecular processes that
91: sustain life. The role of bending stiffness of DNA in gene regulation
92: (Widom (2001)) and the stiffness of substrate in cell migration
93: (Lo {\it et al.} (2000)) are two examples. Such
94: insights have been garnered with the aid of sophisticated experiments that
95: have made it possible to apply piconewton forces on nanometer size objects
96: and measure the resulting deformations. The data from these experiments point
97: to a rich interplay of thermal and chemical forces with electrical and
98: mechanical forces. In this paper we show how experimental insights can be
99: combined with mechanical principles to construct a simple and quantitative
100: mechanical theory of DNA packaging in bacterial viruses that suggests
101: a new round of experiments.
102:
103:
104:
105: \section{Mechanics and the Viral Life Cycle}
106: Viruses have been studied extensively (see Alberts {\it et al.} (1997))
107: in the last three decades not just
108: with the goal of understanding their pathogenic nature but also for gaining
109: insights into structural biology of proteins and nucleic acids and for
110: studying gene regulation. Even though viruses are perhaps the simplest entities from
111: a biological perspective, they are nanotechnological marvels that embody a
112: wealth of physics at the nanometer scale. The principles operating at these scales
113: are being rapidly unravelled through ingenious experiments, such as that
114: by Smith {\it et al.} (2001) on DNA packaging in the \(\phi\)29 virus, which
115: serves as the primary motivation for the theoretical analysis presented below.
116:
117: Before going into the details of the experiment it is useful to get a
118: glimpse of the life-cycle of a virus (see fig.~\ref{fig:lcycle}). We are
119: especially interested in the class of
120: viruses known as bacteriophage and which infect bacteria
121: such as the well-known {\it E.coli}. In simplest terms, a bacteriophage
122: is nothing but a protein coat, the {\it capsid}, filled with nucleic acid
123: such as DNA or RNA. The capsids of bacteriophage attach to the surface of
124: the bacterial
125: cell which is under attack. The genetic material of the virus is {\it ejected} into
126: the bacterium leaving the capsid behind. Though the process
127: of DNA ejection is not completely
128: understood, it is argued that a mature bacteriophage capsid is highly
129: pressurised (see Smith {\it et al.} (2001)) and the genome is forcefully released
130: into the host cell whose contents are effectively at a much lower pressure.
131:
132: Once the viral DNA is inside the bacterium it hijacks the protein production
133: machinery of the bacterium to synthesize its own proteins such as those that
134: will ultimately make up the capsid. Copies of the viral DNA are also made.
135: Here again, experiments have shed light on the mechanics and kinetics of the
136: processes of {\it transcription} of the genome and its {\it translation}
137: into proteins. The process of transcription is mediated by a large protein
138: called RNA polymerase. This protein attaches to the DNA and then moves
139: along its length transcribing the DNA into
140: a molecule known as mRNA. As it moves along, the protein exerts forces
141: on the DNA molecule. These forces have been measured experimentally
142: (Wang {\it et al.} (1998)) and it
143: has been found that the protein can be stalled by exerting a force of roughly
144: 25pN. In fact, the rate of transcription is heavily dependent on the force
145: exerted. The type of data emerging from such experiments unmistakably points
146: to a plethora of mechanics problems still to be resolved at the level of
147: macromolecular assemblies.
148:
149: After a sufficient quantity of the viral proteins have been synthesized they
150: begin to {\it self-assemble} into hollow capsids. In the case of the \(\phi\)29
151: virus, the capsids have a portal at one end with an attached protein motor.
152: The motor
153: proteins identify the appropriate end of the as-yet unpackaged viral
154: DNA and push it into
155: the capsid. As more of the DNA gets pushed into the capsid the motor
156: has to perform work against an increasing resistive force due to confinement of
157: the packaged DNA. This causes the motor to
158: slow down as more DNA fills up, but ultimately all the DNA is packed
159: and the remaining proteins attach themselves to the capsid thus making it
160: ready to infect another bacterium. Once the capsids are fully assembled an
161: enzyme is released that breaks up the bacterial cell membrane and releases
162: the mature viruses so that they can repeat their evil action elsewhere (see Ptashne (1992)).
163:
164: \begin{figure}[h]
165: \centering
166: \includegraphics[scale=0.5]{vircycle.ps}
167: %% \includegraphics[scale=0.5]{vircycle}
168: %% \includegraphics[scale=0.5]{vircycle.jpg}
169: \caption{The life cycle of a virus. The critical
170: step which is the focus of this paper is the packaging of DNA into the
171: assembled capsid. The packaging process involves large forces on the
172: order of several tens of pico-newtons and it proceeds at high rates,
173: packing the full length of the genome in only a few minutes.}
174: \label{fig:lcycle}
175: \end{figure}
176: As noted above, a critical step in the life cycle (see fig.~\ref{fig:lcycle})
177: of the virus is
178: the packaging of DNA into the capsid by the portal motor. This step was the
179: focus of the experiment by Smith {\it et al.} (2001) in which they used optical
180: tweezers to measure the force exerted by the portal motor of the \(\phi\)29
181: bacteriophage on the DNA as a function of the amount of DNA already packed in
182: the capsid. One end of
183: the DNA was attached to a silica bead which was held in an optical trap while the
184: motor tugged on the other end. The capsid itself was attached to
185: a second silica bead that was held by a glass pipette thus immobilizing the
186: capsid during the packaging process. Through a series of careful
187: experiments, the average force exerted by the portal motor was measured and
188: plotted against the fraction of packed DNA. The rate at which the DNA is
189: packed was also measured as a function of the fraction packed. It was
190: found that during the early stages of packing the motor packs at a rate of
191: roughly 100 base pairs per second which correponds to \(34\)nm per second.
192: However, the
193: rate drops considerably after about \(70\)\% of the DNA has been packed and
194: it slows to a crawl as the packaging nears completion. We note
195: that the experiment was conducted {\it in vitro} in a solution which contained
196: \(50\)mM NaCl, \(5\)mM MgCl\(_{2}\) and tris-HCl buffer at pH 7.8. These
197: are not the conditions prevailing in a bacterium and one of the
198: interesting outcomes of the type of model being described here is
199: the recognition that the ionic strength is a key control parameter in
200: governing the physics of DNA packaging.
201: There are a variety of interesting mechanics issues to be confronted
202: relating to those stages of the viral life-cycle involving
203: both DNA packaging and ejection. One such issue involves determining what
204: is the maximum sustainable
205: pressure in a viral capsid (Purohit {\it et al.} (2003)). Another concerns
206: determining the amount
207: of DNA ejected by a bacteriophage as a function of the osmotic pressure in the
208: exterior solution (Evilevitch {\it et al.} (2003)).
209:
210: This paper puts forth a simple mechanical model of the packaging process in an
211: endeavour to understand the nature of the forces being exerted on the DNA.
212: Previously, we described results of an analysis based on a continuum
213: approximation to the mechanical model of the packaging force (Purohit {\it et al.}
214: (2003)). Here we analyze a discrete version of that same model which
215: reveals interesting new effects which could be accessible experimentally.
216:
217:
218:
219: \section{The Free Energy Function for Packed DNA}
220: In order to estimate the energetics of viral DNA packing,
221: we take a structural hint from the cryo-electron microscopy images of capsids
222: (see, for example, Cerritelli {\it et al.} (1997)) in various stages of the
223: packing process. It has been observed for several different
224: viruses that the packed DNA is arranged in a series of circular hoops (or
225: a spool). Each hoop finds itself at the center of a hexagon formed
226: by nearly parallel segments of the same overall strand. The exception are
227: the hoops
228: hugging the surface of the capsid, or those at the inner surfaces of the
229: spool, which are
230: surrounded on average by three nearest neighbors. We note that the geometry is
231: more subtle than the hoop packing adopted here since the packing involves a
232: helical pitch and hence the adjacent strands are probably not perfectly
233: parallel. Following earlier
234: work (see Riemer and Bloomfield (1978), Odijk (1998) and Kindt {\it et al.}
235: (2001)), the basic idea is to write a free energy function which characterizes
236: the energetics of structures like those described above as a sum of elastic
237: terms and interaction terms. It is well known that DNA
238: has a considerable bending stiffness at length scales of a few tens of
239: nanometers. This is evident from the fact that the persistence length of DNA is
240: about \(50\)nm. The persistence length is defined as the distance over which the
241: tangents to a fiber are correlated. The more flexible a fiber is the smaller its
242: persistence length. Polymer physicists use the persistence length as a measure
243: of the stiffness of a chain (see Doi and Edwards (1988)). Since the
244:
245: persistence length of DNA is of the same order
246: as the size of viral capsids
247: we expect the {\it bending} energy to play an important part in the packaging
248: process.
249: Such a description of the elasticity of DNA is now quite firmly established in
250: literature and is found to be useful in understanding key biological
251: processes related to genetic regulation and DNA condensation in chromosomes
252: of eucaryotic cells.
253:
254: Beyond its explicit mechanical properties another relevant characteristic of DNA is
255: its strong acidity. In particular, in an aqueous solution the backbone
256: of the molecule is highly charged, with two units of negative charge per
257: base pair. As a result there
258: are electrostatic penalties involved in bringing two strands in close
259: proximity. In the salty solutions found in cells the picture is more complicated
260: because charged ions screen these interactions. Another complication arises
261: from the
262: fact that close packing involves the ejection of water molecules from the
263: vicinity of the strands, which make sizable contributions to the free energy.
264: All of these complex interactions of the DNA with the ambient solution and with
265: itself result in an effective {\it interaction} energy. The
266: importance of these effects is immediately evident when we look at the physics
267: of DNA condensation in a solutions containing trivalent and tetravalent cations.
268: In particular, DNA is known to form hexagonally packed toroidal structures when
269: immersed in a solution containing these ions (Raspaud {\it et al.} (1998)).
270: The aim of the calculations to follow is to model viral packing, taking into account
271: both elastic and interaction energies.
272:
273: \subsection{Elastic Contribution to the Free Energy}
274:
275: We begin by idealizing DNA as an elastic rod
276: capable of sustaining bending deformations. The centerline of the
277: rod is parametrized by the reference coordinate \(x\). The bending energy stored
278: in length \(l\) of this rod is given by
279: \begin{equation}
280: e(l) = \frac{\kappa}{2}\int_{0}^{l}\frac{dx}{R(x)^{2}},
281: \end{equation}
282: where \(R(x)\) is the radius of curvature at reference position \(x\) along
283: the rod and \(\kappa\) is a bending modulus which we assume to be independent
284: of the reference position \(x\) though we note that the
285:
286: sequence dependence of the elasticity of DNA is one of its most
287:
288: intriguing features. For a rod bent into a hoop, the
289: radius of curvature \(R(x)\) is just the radius of the hoop \(R\),
290: and the length is \(l=2\pi R\), and so the integral reduces to
291: \begin{equation}
292: \label{eq:hoopel}
293: e(2 \pi R) = \frac{\kappa}{2R^{2}}\int_{0}^{2\pi R}dx = \frac{\pi\kappa}{R}.
294: \end{equation}
295: The modulus \(\kappa\) is usually taken to be \(EI\) where \(E\) is the
296: Young's modulus of the material and \(I\) is the moment of inertia of the
297: cross-section. However, it is convenient to express the elastic properties
298: of molecules differently since thermal oscillations play a dominant
299: role in their deformation. In particular, the deformation at a
300: given point on a molecule may be uncorrelated with the deformation at
301: another location because of the randomness caused by these thermal
302: motions. As noted above, the length at which the importance of bending energy
303: is comparable to that due to thermal vibrations is called the `bend persistence
304: length'
305: and is denoted by \(\xi_{p}\). More precisely, \(\xi_{p}\) is determined
306: by balancing the bending energy and the thermal energy and
307: results in \(\kappa = \xi_{p}k_{B}T\),
308: where \(\xi_{p}\) is the
309: persistence length, \(T\) is the temperature and \(k_{B}\) is Boltzmann's
310: constant; \(\xi_{p} = 50\)nm for DNA and \(k_{B}T = 4.1\)pNnm at \(T=300\)K.
311: Using this description of the elastic properties of DNA and adding up
312: the contributions due to all of the hoops in the spool, the
313: total bending energy is given by
314: \begin{equation}
315: E_{bend} = \pi \xi_{p}k_{B}T\sum_{i}\frac{N(R_{i})}{R_{i}},
316: \end{equation}
317: where \(N(R_{i})\) is the number of hoops in a column of radius \(R_{i}\).
318: Note that we neglect the fact that the DNA is packed in a helical arrangement
319: with an associated contribution to the curvature due to the helical pitch.
320: As noted in our earlier paper (Purohit {\it et al.} (2003)), this effect is
321: negligible.
322: For analytical simplicity the discrete expression given above can be converted
323: into an integral and written as
324: \begin{equation} \label{eq:engint}
325: E_{bend} = \frac{2\pi \xi_{p}k_{B}T}{\sqrt{3}d_{s}}\int_{R}^{R_{out}}
326: \frac{N(r)}{r}dr
327: \end{equation}
328: where \(d_{s}\) is the spacing between adjacent hoops, \(R_{out}\) is the
329: radius of the capsid and \(R\) is the radius of the innermost set of hoops.
330: The factor of \(\frac{\sqrt{3}}{2}\) appears since the spacing between two
331: adjacent columns of hoops is \(\frac{\sqrt{3}}{2}d_{s}\) in a hexagonal
332: array. The integral approximation has the virtue that it leads to
333: closed-form analytical expressions for the energies and forces in the viral
334: packing problem, at least for simple geometries like the cylinder and the sphere.
335: An interesting set of conclusions to be discussed below
336: are the differences between the discrete and continuum descriptions of this
337: problem, and their possible experimental consequences.
338:
339: \subsection{Free Energy of Interaction}
340: In addition to the role played by elastic bending we must also consider the
341: effect of the interaction between adjacent strands of DNA. For this we appeal
342: to the experiments of Parsegian {\it et al.} (1986) and Rau {\it et al.} (1984).
343: In these experiments, osmotic pressure was applied on hexagonally packed
344: DNA and the interstrand separation was measured as a function of the
345: pressure for a variety of different solvent conditions. For conditions
346: comparable to those of Smith {\it et al.} (2001), it was found that the
347: osmotic pressure could be related to the interstrand spacing as
348: \begin{equation}
349: p(d_{s}) = F_{0}\exp(-\frac{d_{s}}{c}),
350: \end{equation}
351: where \(F_{0}\) is a constant whose magnitude depends on the type and strength
352: of the ionic solution, and \(c= 0.27\)nm is a decay length which is roughly constant
353: over a range of ionic conditions. For a solution containing
354: \(500\)mM NaCl at 298K, measurements reveal
355: \(F_{0}=55000\)pN/nm\(^{2}\). To this piece of empirical
356: evidence we add the assumption that parallel strands interact through a pair
357: potential \(v(d_{s})\) per unit length and that interactions are limited only
358: to nearest neighbors. Given this description of the total energy
359: and the measured pressure vs interstrand spacing, we can deduce the functional
360: form of \(v(d_{s})\).
361:
362: In particular, consider \(N\) parallel strands of length \(l\) each
363: packed in a hexagonal array with a spacing \(d_{s}\). The total interaction
364: energy of this arrangement is
365: \begin{equation} \label{eq:bass}
366: E = 3N l v(d_{s}),
367: \end{equation}
368: where the factor of \(3\) appears because each strand interacts with \(6\) nearest
369: neighbours (ignoring surface effects) and we multiply by \(1/2\) to account for
370: double counting. The total volume of the assemblage is obtained by adding
371: together the (prismatic) volume occupied by each strand, thus
372: \begin{equation}
373: V = N \frac{\sqrt{3}}{2} d_{s}^{2}l.
374: \end{equation}
375: We now use the thermodynamic identity \(p=-\frac{\partial E}{\partial V}\) to
376: relate the expression
377: for pressure obtained from experiments to our simple model. We note that
378: \(dE = 3Nl \frac{\partial v}{\partial d_{s}} dd_{s}\) and
379: \(dV = \sqrt{3} d_{s} Nl dd_{s}\), so that
380: \begin{equation}
381: f(d_{s})= -\frac{\partial v(d_{s})}{\partial d_{s}}
382: = \frac{1}{\sqrt{3}} p(d_{s}) d_{s},
383: \end{equation}
384: where \(f(d_{s})\) is the force per unit length on the strands. We now
385: substitute the experimental result, namely
386: \(p(d_{s}) = F_{0}\exp(-\frac{d_{s}}{c})\) and solve the
387: differential equation above with the boundary condition \(v(\infty)=0\). This
388: results in the potential
389: \begin{equation}
390: v(d_{s}) = \frac{1}{\sqrt{3}}F_{0}(c^{2}+cd_{s}) \exp(-\frac{d_{s}}{c}).
391: \end{equation}
392: If this result is now exploited in the context of eqn. (\ref{eq:bass}) the
393: interaction energy can be written as
394: \begin{equation}
395: E_{int} = \sqrt{3}F_{0}(c^{2}+cd_{s})L \exp(-\frac{d_{s}}{c}),
396: \end{equation}
397: where \(L=Nl\) is the total length of the strands.
398:
399: Now we are in a position to write the total energy of the packaged DNA
400: as a sum of the bending energy and the interaction energy as follows
401: \begin{equation} \label{eq:mona}
402: E(L,d_{s}) = \sqrt{3}F_{0}L(c^{2}+cd_{s})\exp(-\frac{d_{s}}{c})
403: + \pi\xi_{p}k_{B}T\sum_{i}\frac{N(R_{i})}{R_{i}}.
404: \end{equation}
405: In the continuum approximation this takes the form
406: \begin{equation} \label{eq:totint}
407: E(L,d_{s}) = \sqrt{3}F_{0}L(c^{2}+cd_{s})\exp(-\frac{d_{s}}{c})
408: +\frac{2\pi \xi_{p}k_{B}T}{\sqrt{3}d_{s}}\int_{R}^{R_{out}}
409: \frac{N(r)}{r}dr.
410: \end{equation}
411: We still need to express the length \(L\) in terms of the radii of the hoops.
412: To do so, we simply add up the accumulated length of all the hoops as,
413: \begin{equation} \label{eq:discrlen}
414: L = \sum_{i} 2\pi R_{i} N(R_{i}).
415: \end{equation}
416: In the continuum approximation this can be rewritten as
417: \begin{equation} \label{eq:lenint}
418: L = \frac{4\pi}{\sqrt{3}d_{s}}\int_{R}^{R_{out}}rN(r)dr.
419: \end{equation}
420: With both the elastic and interaction contribution to the total
421: energy of packed DNA in hand, we turn now to a concrete investigation of the
422: implications of this model.
423:
424:
425:
426: \section{Geometry and Energetics of Packed DNA}
427: The calculation of the free energy associated with the DNA packed in a partially
428: filled capsid is predicated on key insights gained from experiments. We
429: assume throughout that the DNA adopts an inverse-spool geometry like
430: that described above, with the proviso that the class
431: of minimum free energy structures we consider is constrained to
432: the spool-type and that there are perhaps more complex
433: structures with lower free energy. We also assume that the strands have
434: local six-fold coordination.
435: As a result, the only variable that remains to be determined is the spacing
436: \(d_{s}\) between the strands as a function of the amount of DNA packed.
437: In the regime where the interaction between the DNA strands is repulsive
438: the interaction term dictates that the strands be as far apart as possible.
439: On the other hand, since the outer
440: dimensions of the capsid are fixed, larger spacings imply smaller radii of
441: curvature which results in a steep rise in the bending energy cost.
442: The competition between these two contributions to the free energy determines
443: the geometry of the packed DNA (Reimer and Bloomfield (1978), Odijk (1998),
444: Kindt {\it et al.} (2001)). We illustrate this through two
445: examples and also make a comparison with the experimental data of
446: Smith {\it et al.} (2001).
447:
448: For simplicity we begin by considering a cylindrical capsid which is a
449: first approximation
450: to the geometry of the \(\phi\)29 virus considered in the experiments of Smith
451: {\it et al.} (2001). In reality, the \(\phi\)29 virus is shaped like a hollow
452: oblong spheroid (Tao {\it et al.} (1998) and Wikoff and Johnson (1999))
453: which has an outer radius of
454: \(21\)nm and a height of about \(54\)nm. Its protein coat is about \(1.6\)nm
455: thick (on average) so that the radius of the capsid volume available to the DNA
456: is \(R_{out}=19.4\)nm while its height is roughly \(51.0\)nm. The capsid has
457: other interesting geometrical features, but for the purposes of the
458: present calculation we idealize it as a cylinder. The dimensions of this
459: idealized cylinder are given in fig.~\ref{fig:phi29} where the key point is
460: selecting the height of the cylinder so as to gaurantee that the volume
461: available for packing is identical to that of the real virus. The figure
462: shows a second geometry that will be used later as a more refined model
463: of the \(\phi\)29 capsid.
464: \begin{figure}[h]
465: \centering
466: \includegraphics[scale=0.6]{capcyjmp.ps}
467: %% \includegraphics{capcyjmp}
468: %% \includegraphics{capcyjmp.jpg}
469: \caption{Model representations of viral capsids. (a) Capsid modeled as
470: a cylinder with hemispherical caps shown in perspective. The curved rods
471: inside represent the packed DNA. (b) The cylindrical geometry shown in
472: cross-section with height adjusted to correspond to the correct internal
473: volume of the capsid. The gray circles represent the hexagonally packed
474: DNA hoops. (c) Idealization of \(\phi\)29 as a cylinder with
475: hemispherical caps shown in cross-section.}
476: \label{fig:phi29}
477: \end{figure}
478:
479: The \(\phi\)29 capsid encloses about \(6.6\mu\)m of double-stranded DNA
480: corresponding to a genome of \(19.3\) kilobase pairs. As noted above for
481: our cylindrical model we take the height of the cylinder as \(z=37.87\)nm
482: to ensure that the internal volume is the same as that of the actual capsid.
483: For such a capsid, the number of hoops at radius \(R_{i}\) is given by
484: \(N(R_{i}) = h/d_{s}\) where \(d_{s}\) is the
485: separation between the strands and \(h\) is the height of the cylinder as shown
486: in fig.~\ref{fig:phi29}.
487: One of the key themes of the present paper is the comparison of the
488: discrete and continuous treatments of the structure and energetics
489: of viral DNA packing. To that end, it is of interest to
490: examine the length of packaged DNA in the discrete setting.
491: Assuming that there are \(n\) columns of stacked hoops (see
492: fig.~\ref{fig:phi29} where we show two completed columns of hoops, i.e.~\(n=2\)),
493: the length of the packed DNA is
494: \begin{eqnarray}
495: L & =& 2\pi \frac{h}{d_{s}} \Big(R_{out} + R_{out}-\frac{\sqrt{3}}{2}d_{s}
496: +R_{out} -2\frac{\sqrt{3}}{2}d_{s} + ...
497: +R_{out}-(n-1)\frac{\sqrt{3}}{2}d_{s}\Big) \nonumber \\
498: & =& 2\pi n\frac{h}{d_{s}}\Big(R_{out}
499: -\frac{n-1}{2}\frac{\sqrt{3}}{2}d_{s}\Big). \label{eq:dislen}
500: \end{eqnarray}
501: This equation results from adding up the accumulated length of the hoops at
502: each radius as demanded by eqn.~(\ref{eq:discrlen}).
503: The hexagonal coordination of the strands appears once again through the
504: distance between adjacent columns of hoops which is \(\frac{\sqrt{3}}{2}d_{s}\).
505: As a concrete example, we use the numbers for the \(\phi\)29 phage and
506: compute how many
507: columns of stacked hoops it contains when fully packed.
508: X-ray measurements on packed \(\phi\)29 indicate that the spacing between the
509: strands in a fully packed \(\phi\)29 capsid is about \(2.8\)nm (Earnshaw
510: and Casjens (1980)). The length of the genome is \(L=6584\) nm, and
511: eqn. (\ref{eq:dislen}) implies \(n=5.61\) corresponding to \(5\) columns of
512: completely stacked
513: hoops and one innermost column of hoops that is only about \(60\%\) full.
514: There is a hollow cylindrical region inside whose radius is \(R_{in}
515: = 7.28\)nm. This calculation reveals that
516: there are not a large number of columns of hoops even in a completely
517: filled capsid and as a result, caution should be exercised in making the continuum
518: approximation in converting the discrete sums in equations (\ref{eq:engint})
519: and (\ref{eq:lenint}) to integrals.
520:
521: We now turn to the more detailed aspects of the geometry and see how by
522: minimizing the free energy we can obtain experimentally falsifiable insights
523: into the packing process.
524:
525: \subsection{Interstrand Spacing in Packed DNA Capsids: Discrete Model}
526: We mentioned earlier that the spacing between the strands is determined
527: by the competition between bending and interaction energies. We
528: demonstrate this minimization procedure with a very simple calculation. The model
529: as set forth above has one free parameter, namely, $F_0$, the parameter
530: characterizing the strength of the repulsion between adjacent DNA strands. We fix
531: \(F_{0} = 225500\)pN/nm\(^{2}\) since, as shown below, this leads to the best
532: fit to the experiment of Smith {\it et al.} (2001). We note that the ion
533: concentration in their experiment is small compared to the values reported
534: by Rau {\it et al.} (1984), which implies that the repulsive forces
535: will be stronger than those measured by Rau {\it et al.} (1984)
536: resulting, in turn, in a larger value of \(F_{0}\). We continue with
537: the cylindrical capsid geometry with the same internal volume as the
538: \(\phi\)29 phage.
539: The goal is to determine the spacing \(d_{s}\) between the strands for a given
540: length \(L\) of genome packed. To that end, we minimize the free energy with
541: respect to \(d_{s}\) at fixed length \(L\). In particular, we set
542: \(d_{s}\) to a fixed value in eqn. (\ref{eq:dislen}) and then solve for
543: the number of columns of hoops this implies while holding the length
544: \(L\) fixed. We then have
545: all the information needed to calculate the energy using eqn. (\ref{eq:mona}).
546: We repeat these steps for a range of different values of \(d_{s}\) to get a plot
547: of energy
548: vs \(d_{s}\). Finally, we use this plot to determine the \(d_{s}\)
549: corresponding to minimum energy at a given length of DNA. One such plot is
550: shown in fig.~\ref{engyeg}.
551: The interaction energy decreases as the spacing increases. This is expected since
552: the interaction energy is a decaying exponential and increasing $d_s$ implies
553: that the interactions come from deeper in the tail of the
554: pair potential. The bending energy
555: rises as the spacing increases since the radii of the innermost strands become
556: successively smaller. The total energy is dominated by
557: the interaction part at smaller spacings and by the bending energy at higher
558: spacings. For the case shown in fig.~\ref{engyeg} we can clearly identify
559: a minimum corresponding to \(d_{s}=3.81\)nm
560: in the total energy.
561: \begin{figure}[h]
562: \centering
563: \includegraphics[scale=0.6]{enerdis.ps}
564: %% \includegraphics{enerdis}
565: %% \includegraphics{enerdis.jpg}
566: \caption{Energy vs interstrand spacing for \(L=4.0\mu\)m
567: in a cylindrical capsid with \(R_{out}=19.4\)nm and \(z=37.87\)nm
568: and with repulsive parameters
569: \(F_{0} = 225500\)pN/nm\(^{2}\) and \(c=0.27\)nm. A minimum is
570: evident at \(d_{s} = 3.81\)nm. The sharp turn in the bending energy at
571: \(d_{s} = 3.81\)nm is the point where a new column of hoops with smaller radius
572: begins to form.}
573: \label{engyeg}
574: \end{figure}
575: \begin{figure}[h]
576: \centering
577: \includegraphics[scale=0.6]{disdscon.ps}
578: %% \includegraphics{disdscon}
579: %% \includegraphics{disdscon.jpg}
580: \caption{Spacing between the DNA strands as a function of the amount of DNA
581: packed for a model cylindrical capsid. The spacing between strands in
582: the early stages is larger when
583: the magnitude of the interaction energy is larger. In the later stages
584: of packing the spacing is dictated by geometrical constraints. The points
585: correspond to the discrete calculation for different ionic conditions and
586: the full curve corresponds to the continuum calculation.}
587: \label{dsdisc}
588: \end{figure}
589: More generally to find the trends for all values of the length packed,
590: this same procedure is repeated and then we
591: plot the optimal \(d_{s}\) (at minimum energy) against the fraction of DNA packed
592: for the \(\phi\)29 phage as shown in fig.~\ref{dsdisc} for
593: three different values of \(F_{0}\). We note that in the later stages of
594: packing all the curves collapse onto a single curve, though we also note that
595: the discrete and the continuous curves have different values in the large
596: packing limit. As shown in our earlier work (Purohit {\it et al.} (2003)),
597: \(d_{s}\) in the large packing
598: limit is dictated by packing the DNA in such a way as to maximize the distance
599: between adjacent strands and should be seen as a geometric limit.
600:
601: On the basis of the simple model
602: presented above, fig.~\ref{dsdisc} provides several distinct predictions.
603: First, on a gross level, it gives the history of the variation
604: of interstrand spacing during the packing process. This is a verifiable prediction
605: since it is possible to perform the viral packaging reaction for different ionic
606: concentrations of the ambient solution and for different values of the
607: genome length. Such experiments have been carried out for the T7 phage for
608: three different
609: genome lengths by Cerritelli {\it et al.} (1997). Similar experiments have
610: also been performed for the \(\lambda\)-phage (see Earnshaw and Harrison (1977)).
611: The second more subtle outcome of the model is
612: the possibility that there are actually discrete effects in the packing
613: process due to the packing of the DNA at a finite set of discrete radii.
614: It would be of interest to determine whether these effects in the packing spacing
615: (and a related force signature to be described below) are present
616: in experiments.
617:
618:
619:
620: \subsection{Interstrand Spacing in Packed Capsids: Continuum Model}
621: One of the interesting features of the calculations presented here is the
622: ability to contrast discrete and continuous descriptions of the energetics
623: of DNA packing. To that end, we now repeat the energy minimization argument
624: that was used to determine the interstrand separation \(d_{s}\) in the
625: discrete setting in the continuum approximation. In particular, we minimize
626: the free energy given in eqn. (\ref{eq:totint}) with respect to \(d_{s}\),
627: but subject to the constraint that \(L\) is constant. In concrete terms
628: this amounts to
629: \begin{equation}
630: \frac{dE}{dd_{s}} = -d_{s}F_{0}L\exp(-\frac{d_{s}}{c})
631: -\frac{4\pi\xi_{p}k_{B}T}{\sqrt{3}d_{s}^{3}}\int_{R}^{R_{out}}
632: \frac{z(R')}{R'}dR' + \frac{4\pi\xi_{p}k_{B}T}{\sqrt{3}R^{2}d_{s}^{3}}
633: \int_{R}^{R_{out}}R'z(R')dR'.
634: \end{equation}
635: where we have used the constraint \(\frac{dL}{dd_{s}} = 0\) which takes the
636: form
637: \begin{equation}
638: \frac{dR}{dd_{s}} = -\frac{2}{d_{s}}\frac{\int_{R}^{R_{out}}R'z(R')dR'}{Rz(R)}.
639: \end{equation}
640: We substitute for \(L\) using \eqn{eq:lenint} and then set
641: \(\frac{dE}{dd_{s}}=0\) to get the following general equation
642: \begin{equation} \label{eq:gencont}
643: \sqrt{3}F_{0}\exp(-\frac{d_{s}}{c}) = \frac{\xi_{p}k_{B}T}{R^{2}d_{s}^{2}}
644: -\frac{\xi_{p}k_{B}T}{d_{s}^{2}}\frac{\int_{R}^{R_{out}}z(R')/R'dR'}
645: {\int_{R}^{R_{out}}R'z(R')dR'}.
646: \end{equation}
647: This equation represents a competition between the interaction terms (on the
648: left) and the bending terms (on the right).
649: Note that the effect of the capsid geometry appears through the second term
650: on the right hand side of \eqn{eq:gencont}.
651: For a cylinder, \(z(R') = h\), a constant, and \eqn{eq:gencont} takes the form
652: \begin{equation} \label{eq:cylinnew}
653: \sqrt{3}F_{0}\exp(-\frac{d_{s}}{c}) = \frac{\xi_{p}k_{B}T}{R^{2}d_{s}^{2}}
654: -\frac{2\xi_{p}k_{B}T}{d_{s}^{2}}\frac{\log\big(\frac{R_{out}}{R}\big)}
655: {R_{out}^{2}-R^{2}}
656: \end{equation}
657: %%\marginpar{\it RP: fix eqn. 19}
658: Eqn.~(\ref{eq:lenint}) is similarly specialized for a cylinder to yield
659: \begin{equation}
660: \label{eq:cylingeo}
661: L = \frac{2\pi h}{\sqrt{3}d_{s}^{2}}(R_{out}^{2}-R^{2}).
662: \end{equation}
663: Eqns.~(\ref{eq:cylinnew}) and (\ref{eq:cylingeo}) feature
664: two unknowns, \(R\) and \(d_{s}\), and we solve for them using the Newton-Raphson
665: method. The result of this computation is a history of the interstrand spacing
666: \(d_{s}\) as a function of the length \(L\) of DNA packed. This has been
667: plotted as a thick line in fig.~\ref{dsdisc} for
668: \(F_{0} = 137500\)pN/nm\(^{2}\). It is evident that while the continuum
669: version captures the trend quite well, it underestimates the value of
670: \(d_{s}\) for the cylindrical geometry in comparison with the more exact
671: discrete version. Also, details captured by the discrete version, such as
672: the series of steps and jumps, are smoothed out in the continuum
673: approximation.
674:
675: \section{Forces during DNA packaging}
676: One of the key outcomes of the model presented here is the internal force that
677: builds up
678: during DNA packing. The force vs percent packed curves should be seen as a
679: second set of experimental implications for this mechanical model which
680: complements predictions of the spacing \(d_{s}\) vs percent packed.
681: \begin{figure}[h]
682: \centering
683: \includegraphics[scale=0.6]{disencon.ps}
684: %% \includegraphics{disencon}
685: %% \includegraphics{disencon.jpg}
686: \caption{Variation in the stored energy as a function of the
687: length of DNA packed for a model cylindrical capsid. The total energy rises
688: as a function of the amount of
689: DNA packed and with increasing value of \(F_{0}\). Also shown as a thick
690: line is the curve obtained from the continuum theory. Because the continuum
691: version of $d_s$ is smaller than that obtained from the discrete model, there
692: is a corresponding increase in the interaction energy which causes the
693: continuum energy to be larger than its discrete counterpart.}
694: \label{fig:endrtcon}
695: \end{figure}
696: Indeed, once we have the spacing \(d_{s}\) as a function of the amount of
697: DNA packed we can determine the energy using \eqn{eq:mona} or (\ref{eq:totint})
698: depending on whether we have adopted the discrete or continuum version of the
699: free energy. The variation
700: of the free energy has been plotted for the three \(F_{0}\) values in
701: fig.~\ref{fig:endrtcon}. Recall that the choice of \(F_{0}\) reflects the
702: ionic conditions present during the packaging process.
703: It is evident that the total energy increases at any given value of the
704: length packed as the strength of the repulsive interactions (the value
705: of \(F_{0}\)) is increased. Figure~\ref{fig:endrtcon} also shows the
706: energy obtained using the continuum approximation which overestimates the
707: energy but correctly captures the trend.
708: The force required to pack DNA is simply the derivative of the energy
709: with respect to the length packed. We have calculated this derivative
710: numerically and the results are plotted in fig.~\ref{fig:forcstep}.
711:
712: The behavior of the force curves is very much as expected, with higher forces
713: for larger values of \(F_{0}\). We note that the strong dependence of
714: the maximum packing force on $F_0$ is amenable to experimental
715: observation and is worth further attention. In particular, we note that the
716: maximum packing force can change by as much as a factor of two depending
717: upon the ionic conditions (and the related value of \(F_{0}\)). Perhaps
718: the most striking feature of the force curves is the appearance of steps.
719: The same steps are seen in the plots of \(d_{s}\) as a function of fraction
720: packed. They occur when a new column (stack) of hoops starts to
721: form on the inside of the spool. Curiously enough, the experimental force curve
722: reproduced in fig.~\ref{coolfigr} also seems to show discontinuities of
723: the slope at
724: 4-5 different points, which might correspond to the five columns of completely
725: stacked hoops needed to fill the $\phi$29 capsid with its DNA.
726: However, to establish that this actually occurs
727: during the packaging of a virus would require a more detailed analysis of the
728: kinetics of packing and more careful experiments. The presence
729: of such steps in the model results from implicitly assuming that the dynamics
730: of packing follows the sort of organized column-by-column hoop packing
731: described above. Therefore, the observation of such steps in experiment would
732: be not only a confirmation of the model, but a provocative hint about the
733: dynamics of packing.
734: \begin{figure}[h]
735: \centering
736: \includegraphics[scale=0.6]{dsfoinse.ps}
737: %% \includegraphics{dsfoinse}
738: %% \includegraphics{dsfoinse.jpg}
739: \caption{Internal force as a function of the amount of DNA packed for a model
740: cylindrical capsid. The plot
741: bears the signature of both the total energy (\fig{fig:endrtcon}) and the
742: \(d_{s}\) (\fig{dsdisc}) curve. Note that the maximum force (at 100\% packing)
743: is roughly proportional to \(F_{0}\) as shown in the inset.}
744: \label{fig:forcstep}
745: \end{figure}
746:
747: \section{Application to the \(\phi\)29 virus}
748: As a final calculation we go beyond the analytic simplicity offered by the
749: cylindrical capsid and consider the second approximate geometry shown in
750: fig.~\ref{fig:phi29} namely, the capped cylinder.
751: Though the cylindrical geometry offers considerable insight into the
752: mechanics of viral packaging it clearly represents an oversimplification
753: of the \(\phi\)29 capsid. The \(\phi\)29 capsid
754: is shaped like an oblong spheroid and we approximate it as a cylinder with
755: hemispherical caps in order to make comparisons with the experimental results
756: of Smith {\it et al.} (2001). The radius of the cylinder (and the hemispheres)
757: is taken to be \(R_{out}=19.4\)nm (see fig.~\ref{fig:phi29}). The height
758: of the cylindrical portion (also called the `waist') is \(z=12.0\)nm.
759: It is useful to study this geometry because one can obtain expressions for
760: spherical capsids (which are very good approximations to icosahedral viruses)
761: merely by setting the height of the waist to zero.
762: The expression for \(N(R_{i})\) is now a bit more complicated and is
763: given by
764: \begin{equation}
765: N(R_{i}) = \frac{z + 2 \sqrt{R_{out}^{2} - R_{i}^{2}}}{d_{s}}.
766: \end{equation}
767: Using this formula in \eqn{eq:discrlen} results in the following expression for
768: the length of DNA packed:
769: \begin{eqnarray}
770: L & =& 2\pi\Big[\frac{z}{d_{s}}R_{out}
771: +\frac{z + 2 \sqrt{d_{s}(\sqrt{3}R_{out}
772: -\frac{3}{4}d_{s})}}{d_{s}}(R_{out} - \frac{\sqrt{3}}{2}d_{s}) + ...
773: \nonumber \\
774: & & +\frac{z + 2 \sqrt{(n-1)d_{s}(\sqrt{3}R_{out}
775: -\frac{3}{4}(n-1)d_{s})}}{d_{s}}(R_{out} - (n-1)\frac{\sqrt{3}}{2}d_{s})\Big].
776: \end{eqnarray}
777: Unlike the cylindrical geometry, this expression is not analytically tractable.
778: However, the same procedure can be used for determining the interstrand
779: spacing, and the energy and force as a function of percent packed.
780: The results of numerical calculations for this geometry, both in the discrete
781: and the continuous setting, are plotted in fig.~\ref{coolfigr}.
782: We find that \(F_{0} = 225500\)pN/nm\(^{2}\) results in a good fit to the
783: experimental data. Moreover, the continuum and discrete versions of the theory
784: are in better agreement for this geometry than they were for cylindrical capsids.
785: \begin{figure}[h]
786: \centering
787: \includegraphics[scale=0.6]{coolgraf.ps}
788: %% \includegraphics{coolgraf}
789: %% \includegraphics{coolgraf.jpg}
790: \caption{Force and interstrand spacing as functions of the amount of DNA
791: packed in a capsid idealized as a cylinder with hemispherical caps. The
792: hexagons correspond to the experimental data of Smith {\it et al.} (2001).
793: The thick lines are results of the continuum model and the circles connected
794: with the thin line are obtained from the discrete model. A good fit to
795: their data is obtained for \(F_{0} = 225500\) pN/nm\(^{2}\).}
796: \label{coolfigr}
797: \end{figure}
798:
799: It is of interest to examine the effect of ionic concentration on the forces
800: exerted by the portal motor on the DNA. The ionic concentration affects the
801: value of \(F_{0}\) and we have already seen that larger values of \(F_{0}\)
802: imply larger values of the force. In the inset to fig.~\ref{fig:forcstep}, the
803: maximum force as a function of the parameter $F_0$ is shown.
804: This information can be potentially useful in the following way. It is known
805: (see Smith {\it et al.} (2001)) that the portal motor of the \(\phi\)29 virus
806: stalls at a force of \(57\)pN. In other words, if one were to conduct the
807: packaging reaction in a solution with ionic conditions such that
808: \(F_{0}>250,000\)pN/nm\(^{2}\) then
809: the genome would not be completely packed since the motor would stall prior to
810: complete packing. We are hopeful
811: that experiments carried out with different ionic concentrations would
812: permit an investigation of such predictions.
813: We have made similar calculations to those presented here for viruses other than
814: \(\phi\)29 and find interesting variations in the maximum packing force from
815: one virus to the next which should be similarly accessible experimentally.
816:
817:
818: \section{Conclusion}
819: Recent advances in the development of tools for single molecule manipulation
820: have permitted
821: the direct mechanical investigation of a variety of biological processes.
822: One intriguing recent example involves the direct measurement of mechanical
823: forces during DNA packaging in viruses. We have developed
824: a simple, analytically tractable model which responds to these experiments.
825: The model emphasizes the role of elastic and electrostatic effects in
826: the packaging process and is in good quantitative agreement with experimental data.
827: More importantly, the model makes a series of predictions that we hope will inspire
828: and guide future experiments. In particular, we predict the appearance of
829: steps in the force versus percent of DNA packed curve. There may be some
830: evidence of steps in the existing experimental data, but further experimentation
831: is needed.
832: The model also predicts a growth of the maximum force
833: required for packaging as the ion concentration is decreased; the underlying
834: cause being the stronger electrostatic repulsion between the DNA strands inside the
835: capsid. This points to
836: a new series of experiments that would look for incomplete packaging of
837: the viral DNA at lower salt concentrations than used previously. An alternative
838: approach would be to work at the same ion concentration but with longer
839: DNA constructs.
840:
841: The biggest shortcoming of the model as presently stated is that it does not
842: address dynamical issues associated with the packaging process. The fundamental
843: experimental observation is that as the packaging process proceeds the packaging
844: rate falls from the initial 100 base pairs per second to zero. We note that
845: the appearance of force steps described in this paper implicitly assumes a
846: precise dynamical pathway resulting in a helical packing of the DNA in a series
847: of helices of ever decreasing radius and are hopeful that the future work will
848: shed further light on the dynamics of viral packing.
849:
850:
851:
852:
853:
854: \section{Acknowledgements}
855: We are grateful to Ron Hockersmith, Paul Grayson, Dwight Anderson and Carlos
856: Bustamante for useful suggestions. We have benefitted from a long series
857: of stimulating discussions with Bill Gelbart, Alex Evilevitch and Chuck Knobler.
858: RP and PP acknowledge support of the NSF through grant number CMS-0301657,
859: the NSF supported CIMMS center and the support of the Keck Foundation. JK
860: is supported by the NSF under grant number DMR-9984471, and is a Cottrell
861: Scholar of Research Corporation.
862:
863:
864: \begin{thebibliography}{99}
865:
866: \bibitem{alberts1} Alberts B., Bray D., Johnson A., Lewis J., Raff M., Roberts K.
867: and Walter P., (1997) {\it Essential Cell Biology}, Garland Publishing, Inc.,
868: New York.
869:
870: \bibitem{boal} Boal D., (2002) {\it Mechanics of the cell}, Cambridge University
871: press, Cambridge.
872:
873: \bibitem{bray} Bray D. (2001) {\it Cell movements}, Garland Publishing, Inc.,
874: New York.
875:
876: \bibitem{busta} Bustamante C., Macosko J.~C. and Wuite G.~J. (2000), Grabbing
877: the cat by the tail: Manipulating molecules one by one, {\it Nat. Rev. Mol.
878: Cell Biol.} {\bf 1}, 130-136.
879:
880: \bibitem{cerritelli} Cerritelli, M.~E., Cheng, N., Rosenberg A.~H., McPherson C.~E.,
881: Booy, F.~P., and Alasdair C. Steven, A.~C., (1997), Encapsidated conformation
882: of bacteriophage T7 DNA, {\it Cell}, {\bf 91}, 271-280.
883:
884: \bibitem{doi} Doi M., and Edwards S.~F., (1988), {\it The theory of polymer
885: dynamics}, Oxford University Press, Oxford.
886:
887: \bibitem{evanska} Evans E. and Skalak R., (1980), Mechanics and
888: thermodynamics of biomembranes, {\it CRC reviews in Bioeng.}, CRC Press,
889: Boca Raton, Florida, 1-254.
890:
891: \bibitem{earnshaw} Earnshaw W.~C. and Casjens S.~R. (1980), DNA packaging
892: by double-stranded DNA bacteriophages, {\it Cell}, {\bf 21}, 319-331.
893:
894:
895:
896: \bibitem{earnshaw2} Earnshaw W.~C. and Harrison S.~C. (1977), DNA
897:
898: arrangement in isometric phage heads, {\it Nature}, {\bf 268}, 598-602.
899:
900: \bibitem{alex} Evilevitch A., Lavelle L., Knobler C.~M., Raspaud E. and
901: Gelbart W.~G., (2003), Osmotic pressure inhibition of DNA ejection from
902: phage, {\it Proc. Natl. Acad. Sci.}, in press.
903:
904: \bibitem{freusure} Freund L.~B. and Suresh S. (2003), {\it Thin film materials:
905: Stress, defect formation and surface evolution}, Cambridge University Press,
906: Cambridge.
907:
908: \bibitem{goldpow} Goldstein R.~E., Powers T.~R. and Wiggins C.~H., (1998)
909: Viscous non-linear dynamics of twist and writhe, {\it Phys. Rev. Lett.},
910: {\bf 80}, 5232-5235.
911:
912: \bibitem{kanamaru} Kanamaru S., Leiman P.~G., Kostyuchenko V.~A., Chipman P.~R.,
913: Mesyanzhinov V.~M., Arisaka F. and Rossmann M.~G., (2002), Structure of the
914: cell-puncturing device of bacteriophage T4, {\it Nature}, {\bf 415}, 553.
915:
916: \bibitem{kindt} Kindt, J.T., Tzlil, S., Ben-Shaul, A., and Gelbart, W. (2001),
917: DNA packaging and ejection forces in bacteriophage, {\it Proc. Nat. Acad. Sci.},
918: {\bf 98}, 13671-13674.
919:
920: \bibitem{lodembo} Lo C.~M., Wang H.~B., Dembo M., Wang Y.~L. (2000),
921: Cell movement is guided by the rigidity of the substrate, {\it Biophys. J.}
922: {\bf 79}, 144-152.
923:
924: \bibitem{marcm} Meyers M.~A., (1994), {\it Dynamic Behavior of Materials},
925: John Wiley \& Sons.
926:
927: \bibitem{odijk} Odijk, T., (1998), Hexagonally packed DNA within bacteriophage
928: T7 stabilised by curvature stress, {\it Biophys. J.} {\bf 75}, 1223-1227 .
929:
930: \bibitem{parseg} Parsegian V.~A., Rand R.~P., Fuller N.~L., and Rau D.~C.,
931: (1986), Osmotic stress for the direct measurement of intermolecular forces,
932: {\it Methods in Enzymology}, {\bf 127}, 400-416.
933:
934: \bibitem{ptashne} Ptashne, M., (1992), {\it A genetic switch}, Blackwell
935: Science (UK).
936:
937: \bibitem{purohit} Purohit P.~K., Kondev J., and Phillips R., (2003), Mechanics
938: of DNA packaging in viruses, {\it Proc. Nat. Acad. Sci.}, {\bf 100}(6),
939: 3173-3178.
940:
941: \bibitem{raspau} Raspaud E., de la Cruz M.~O., Sikorav J.~L., Livolant F., (1998),
942: Precipitation of DNA by polyamines: A polyelectrolyte behavior, {\it Biophys. J.},
943: {\bf 74}(1), 381-393.
944:
945: \bibitem{raup} Rau, D.~C., Lee, B., and Parsegian, V.~A., (1984), Measurement
946: of the repulsive force between the poly-electrolyte molecules in ionic solution
947: - hydration forces between parallel DNA helices, {\it Proc. Natl. Acad. Sci.},
948: {\bf 81}, 2621-2625.
949:
950: \bibitem{parsegian} Rau, D.~C. and Parsegian, V.~A.,(1992), Direct measurement
951: of the intermolecular forces between counterion-condensed DNA helices - evidence
952: of long-range attractive hydration forces, {\it Biophys. J.} {\bf 61 }, 246-259.
953:
954: \bibitem{riemer} Riemer, S.~C. and Bloomfield, V.~A., (1978), Packaging
955: of DNA in bacteriophage heads: Some considerations on energetics, {\it Biopolymers}
956: {\bf 17}, 785-794.
957:
958: \bibitem{rice} Rice J.~R., Lapusta N. and Ranjith K., (2001), Rate and
959: state dependent friction and the stability of sliding between elastically
960: deformable solids, {\it J. Mech. Phys. Sol.}, {\bf 49}(9), 1865-1898.
961:
962: \bibitem{rosakis} Rosakis A.~J. (2002) Intersonic shear cracks and fault
963: ruptures, {\it Adv. Phys.}, {\bf 51}(4), 1189-1257.
964:
965: \bibitem{smith} Smith D.~E., Tans S.~J., Smith S.~B., Grimes S., Anderson D.~L.
966: and Bustamante C., (2001), The bacteriophage $\phi$29 portal motor
967: can package DNA against a large internal force, {\it Nature}, {\bf 413}, 748.
968:
969: \bibitem{simpson} Simpson A.~A., Tao Y., Leiman P.~G., Badasso M.~O., He Y.,
970: Jardine P.~J., Olson N.~H., Morais M.~C., Grimes S., Anderson D.~L., Baker T.~S.
971: and Rossmann M.~G., (2000), Structure of the bacteriophage $\phi$29 DNA packaging
972: motor, {\it Nature}, {\bf 408}, 745.
973:
974: \bibitem{seifert} Seifert U., (1997), Configurations of fluid membranes and
975: vesicles, {\it Adv. Phys.}, {\bf 46}, 13.
976:
977: \bibitem{tao} Tao, Y., Olson N.~H., Xu, W., Anderson, D.~L., Rossman, M.~G.,
978: and Baker, T.~S., (1998), Assembly of tailed bacterial virus and its genome
979: release studied in three dimensions, {\it Cell}, {\bf 95}, 431-437.
980:
981: \bibitem{vogel} Vogel S., (1988), {\it Life's Devices: The physical world of
982: plants and animals}, Princeton University press, Princeton, New Jersey.
983:
984: \bibitem{wang} Wang M.~D., Schnitzer S.~J, Yin H., Landick R., Gelles J. and
985: Block S.~M. (1998), Force and velocity measured for single molecules of
986: RNA polymerase, {\it Science}, {\bf 282}, 902-907.
987:
988: \bibitem{widom} Widom J., (2001), Role of DNA sequence in nucleosome stability
989: and dynamics, {\it Q. Rev. of Biophys.}, {\bf 34}, 269-324.
990:
991: \bibitem{wikoff} Wikoff W.~R. and Johnson J.~E., (1999), Imaging a molecular
992: machine, {\it Curr. Bio.}, {\bf 9}(8), R296-R300.
993:
994: \end{thebibliography}
995:
996:
997: \end{document}
998:
999: