* To whom correspondence should be addressed. Email: sstott@mgh

Subscriber access provided by Karolinska Institutet, University Library

Article

Molecular dynamics at the interface between ice and poly(vinyl alcohol) and ice recrystallization inhibition Lindong Weng, Shannon L Stott, and Mehmet Toner Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b03243 • Publication Date (Web): 04 Dec 2017 Downloaded from http://pubs.acs.org on December 5, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Molecular dynamics at the interface between ice and poly(vinyl alcohol) and ice recrystallization inhibition Lindong Wenga,b, Shannon L. Stotta,c,d,* and Mehmet Tonera,b,e,* a

Center for Engineering in Medicine and BioMEMS Resource Center, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02129, United States b Department of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114, United States c Massachusetts General Hospital Cancer Center, Harvard Medical School, Charlestown, MA 02129, United States d Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114, United States e Shriners Hospital for Children, Boston, MA 02114, United States ABSTRACT Ice formation is a ubiquitous process that poses serious challenges for many areas. Nature has evolved a variety of different mechanisms to regulate ice formation. For example, many cold-adapted species produce antifreeze proteins (AFPs) and/or antifreeze glycoproteins (AFGPs) to inhibit ice recrystallization. Although several synthetic substitutes for AF(G)Ps have been developed, the fundamental principles of designing AF(G)P mimics are still missing. In this study, we explored the molecular dynamics of ice recrystallization inhibition (IRI) by poly(vinyl alcohol) (PVA), a well-recognized ice recrystallization inhibitor, to shed light on the otherwise hidden ice-binding mechanisms of chain polymers. Our molecular dynamics simulations revealed a stereoscopic, geometrical match between the hydroxyl groups of PVA and the water molecules of ice, and provided microscopic evidence of the adsorption of PVA to both the basal and prism faces of ice and the incorporation of short-chain PVA into the ice lattice. The length of PVA, i.e., the number of hydroxyl groups, seems to be a key factor dictating the performance of IRI, as the PVA molecule must be large enough to prevent the joining together of adjacent curvatures in the ice front. The findings in this study will help pave the path for addressing a pressing challenge in designing synthetic ice recrystallization inhibitors rationally, by enriching our mechanistic understanding of IRI process by macromolecules.

* To whom correspondence should be addressed. Email: [email protected] (SLS) and [email protected] (MT) ACS Paragon Plus Environment

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

INTRODUCTION Ice formation is a ubiquitous phenomenon, posing serious challenges for many areas, such as frost damage in agriculture,1 icing hazard for aviation transportation and wind power industries,2-3 and freezing injury in cryobiology.4-6 In nature, many species of fish, insects and plants that inhabit cold environments produce antifreeze proteins (AFPs) and/or antifreeze glycoproteins (AFGPs) as specialized adaptation to modulate the effect of ice on their lives,7-10 including thermal hysteresis, dynamic ice shaping and ice recrystallization inhibition. Ice recrystallization is a process in which large ice crystals grow at the cost of small ones, effectively reducing the interfacial energy of the system (i.e, Ostwald ripening).11 AF(G)Ps have been known to adsorb to multiple faces of ice crystals and inhibit their growth.12-15 Knight et al.14 found that antifreeze proteins from winter flounder and Alaskan plaice adsorbed onto the pyramidal plane of ice whereas the sculpin AFP adsorbs on the secondary prism plane. Using molecular dynamics (MD) simulation, Nada and Furukawa16 demonstrated that a mutant of winter flounder AFP could bind to the pyramidal plane of the ice lattice stably with its hydrophobic residues, yielding a dramatic decrease in the ice growth rate near the ice-binding site. The lowtemperature (250 K) computational study of the solvated AFP from the spruce budworm Choristoneura fumiferana also found that the water structure was ordered and the dynamics slowed down around the ice-binding face of the AFP, suggesting that the preconfigured solvation shell around the ice-binding face was involved in the initial recognition and binding of the AFP to ice by lowering the barrier for binding and consolidation of the AFP-ice interaction surface.17 Very recently, Kuiper et al.18 proposed an icebinding mechanism for the spruce budworm Choristoneura fumiferana AFP. They found that AFP bound indirectly to the prism face of ice crystals through a linear array of ordered water molecules which were structurally distinct from ice.18 However, the high cost, low availability and instability of AF(G)Ps impede their large-scale production and use in industrial applications.19 Therefore, a number of synthetic substitutes have been discovered or developed, such as poly(vinyl alcohol) PVA,11, 19-21 double hydrophilic block copolymers,22 zirconium acetate,23 D-glucose derivatives bearing β-linked para-methoxyphenyl,24 graphene oxide,25 and selfassembled, amphipathic metallohelicies.26 For example, graphene oxide (5 mg/ml) was found to greatly suppress the growth and recrystallization of ice crystals in an 8 mg/ml sodium chloride solution.25 The accompanying MD simulations revealed that oxidized groups on the basal plane of graphene oxide could form more hydrogen bonds with ice than with liquid water due to the honeycomb hexagonal scaffold of graphene, elucidating the molecular mechanism of the ice recrystallization inhibition by graphene oxide.25 Among the abovementioned AFP substitutes, however, PVA has been recognized as an ice recrystallization inhibitor of unusually high activity.11, 26 Inada and Lu19 investigated the recrystallization of ice grains after annealing between −2.3 and −2.0 °C in a calcium chloride solution with the addition of 5 mg/ml PVA ( =7200-98000 g/mol), poly(ethylene glycol) (PEG), poly(acrylic acid) (PAA), vinyl triethoxy silane or Tween. They found that PVA was the most effective in inhibiting ice recrystallization whereas the other four additives had no or insignificant effect of ice recrystallization inhibition (IRI). Budke and Koop20 determined the effect of PVA ( =27000 or 145000 g/mol) on ice recrystallization in an aqueous sucrose solution, and proposed that PVA might adsorb to both the primary and secondary prism faces of hexagonal ice. Although it has been proposed that the adsorption of AF(G)P mimics onto the ice crystal faces plays an essential role in IRI, the ice-binding mechanism, especially at the molecular level, is not fully understood.20 Such lack of understanding limits the rational design of IRI molecules, resulting in a trialand-error process. Towards addressing this challenge, in this study, we explored the molecular dynamics of ice recrystallization inhibition by PVA. We found that there is a stereoscopic, geometrical match between the -OH groups of PVA and the water molecules in the ice lattice, and provided microscopic

2

ACS Paragon Plus Environment

Page 2 of 15

Page 3 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

evidence of the adsorption of PVA to multiple ice faces and the incorporation of short-chain PVA into the ice lattice. The number of hydroxyl groups is a key factor dictating the IRI performance of PVA, as the PVA molecule must be large enough to prevent the joining together of adjacent curvatures in the ice front that are otherwise separated by the PVA molecule. COMPUTATIONAL METHOD Simulation Details Three lengths of PVA were investigated in this study, namely, =5, 10 and 20, as shown in Scheme 1. In each simulation box, a single layer of hexagonal ice,27 containing 1280 TIP4P/200528 molecules, was fixed in the x-z plane, serving as the seed ice (Scheme 2). On each side of the seed ice, there were abundant water (TIP4P/2005) molecules that had been pre-equilibrated at 230 K and 101.325 kPa. There were a total of 16266-16431 water molecules in the simulation box, generating an initial size of 81.14 Å x 80 Å x 80.96 Å, regardless of the length of PVA molecules. One PVA molecule was inserted into each water layer with a center-to-center distance of 15 Å (for PVA5 and PVA10) or 17.5 Å (for PVA20) from the seed ice. The starting conformation of PVA was taken from a 10-ns equilibration of the macromolecule solvated in water at 300 K and 101.325 kPa. Each simulation was run for 100 ns with an isothermalisobaric (NpT) ensemble in which the temperature and pressure were fixed at 230 K and 101.325 kPa, respectively. The simulation temperature is about 22 K below the melting point of TIP4P/2005 (i.e., 252.1 K).28 Such supercooling extent was found to facilitate the ice crystal growth.29 The force field parameters for PVA were generated through MATCH.30 We analyzed the dynamics and molecule arrangement over the simulation period of 0-70 ns. The ice layers began to meet its periodic counterparts along the y axis after 70 ns, which may jeopardize the accuracy of the analysis as an unfrozen water cluster may be trapped within the ice phase and interrupt the ice lattice that has been formed previously.

Scheme 1. Chemical structure of poly(vinyl alcohol) (top right) and the backbone chain of PVA5 (=5), PVA10 (=10) and PVA20 (=20) (bottom). Open circles indicate the carbon atoms. The oxygen atoms are color coded and represented by close circles, which will serve as the color legend for Figure 4 and 5. The hydrogen atoms are not shown.

3


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 15

Scheme 2. The PVA-ice-water simulation system viewed in the x-y plane. A single layer of hexagonal ice slab is placed in the center, surrounded by over 16000 water molecules. There is a PVA molecule on each side of the seed ice. The ice face that can be seen in this figure is the basal plane, while the prism face laying in the x-z plane is exposed to the liquid phase. The simulations in this study were conducted by using the MD simulation package NAMD.31 The distance beyond which electrostatic and van der Waals interactions are truncated is 10 Å. When the distance is beyond 8 Å, the switching functions begin to take effect to smoothly reduce electrostatic and van der Waals interactions to zero. The time step is 2 fs. The SHAKE algorithm32 is used to fix the vibrations of the fastest atoms. The temperature is fixed using Langevin dynamics33 with a damping coefficient of 1 ps-1. The pressure is fixed by a modified Nosé-Hoover method, which is a combination of the constant pressure algorithm proposed by Martyna et al.34 with piston fluctuation control implemented using Langevin dynamics35. The atoms of the seed ice were held fixed in the simulation by not adjusting the coordinates of the fixed atoms when rescaling the unit cell to maintain a constant pressure. The NpT barostat is anisotropic in which the x, y and z lengths of the simulation box fluctuated independently. Periodic boundary conditions were applied and the Particle Mesh Ewald (PME) method was employed,36 such that the configuration of the seed ice was periodically replicated in the x-z plane. Averaged Local Bond Order Parameter We used the averaged local bond order parameters ( ) developed by Lechner and Dellago37 to distinguish solid-like and liquid-like molecules in the (PVA-)ice-water systems. holds the structural information of both the first and second shells around particle . It is defined as ∑ | |

(1)

∑

(2).

where

4


Page 5 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

runs over all neighbors of particle plus the particle itself. Here, the sum from =0 to The complex vector of particle is defined as38 ∑"

!"

(3)

where is the number of nearest neighbors of particle , # is a free integer parameter, and $ is an integer that runs from $=−# to $=+#. The function !" is the spherical harmonics and !" is the vector from particle to particle '. The cutoff distance to identify neighboring oxygen atoms for the calculation of for a given oxygen atom is 3.5 Å, indicating the position of the first minimum of the radial distribution function of the oxygen-oxygen pair in liquid water. It has been demonstrated that for TIP4P/2005 the parameter ( is sufficient to distinguish between the solid-like (either cubic or hexagonal ice) and liquid-like molecules. If a TIP4P/2005 oxygen atom possesses ( >0.358, the molecule it belongs to will be identified as ice.39 The calculated number of ice molecules near time zero is 1184-1291, very close to the number of molecules comprising the seed ice, which demonstrates the validity of the above criteria. Hydrogen Bond The hydrogen bonds in the (PVA-)ice-water systems were determined via the geometrical criteria.40 A certain aggregate between two oxygen atoms is identified as a hydrogen bond if the O···O distance does not exceed 3.5 Å and the angle O–H···O is greater than 145°. RESULTS AND DISCUSSION During the simulation, the ice crystal grew layer by layer. Figure 1 displays representative snapshots of the ice-water systems in the absence or presence of PVA at )=10, 30 and 60 ns, respectively. At )=10 ns, ice crystals had grown two additional layers from each side of the seed ice. Meanwhile, PVA, regardless of the chain length, began to adsorb onto the frontier ice layers. Pronounced interruption against the frontier ice layers by PVA was observed at )=30 ns, showing evident structural defects. As the simulations were run for another 30 ns, however, the ice layers engulfed PVA5 and PVA10 into the ice crystal lattice. In contrast, the structural defect attributed to PVA20 persisted and expanded during the 60-ns period. As a result, the ice layers only accumulated on the sides in the PVA20-ice-water system. In the simulation setup, the PVA molecule below the seed ice (y0 Å). Therefore, the two PVA molecules had different but complementary configurations in the x-z plane facing the ice front. Given that curvatures were generated on both sides of the seed ice due to the PVA20-ice binding (Figure 1), the starting conformation of PVA20 facing the ice front did not affect the ice-binding activity significantly. As seen in Scheme 2, it is the primary prism face that is exposed to the liquid phase in our simulations. We performed an additional PVA20-ice-water simulation with the secondary prism plane of the seed ice facing the liquid phase. We observed that the PVA20 molecules still bound to the growing ice front and pushed the ice layers to accumulate only on their sides (See Figure S1 in Supporting Information). The similar phenomena shown in Figure 1 and S1 indicate that the ice-binding activity of PVA is independent of the ice face that is exposed to the liquid phase, due to the stereoscopic nature of the ice-binding by PVA as will be discussed later. In this study, the PVA molecules that were initially placed in the water box for pre-equilibration at 300 K were isotactic. As the PVA samples that were studied in the experiments usually have random stereochemistry, we conducted another PVA20-ice-water simulation in which the very initial structure of PVA20 is atactic (See Figure S2 A in SI). The result presented in Figure S2 B demonstrates that the ice-

5


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

binding activity of PVA20 is also independent of the polymer tacticity as curvatures had emerged upon the adsorption of PVA20 onto the ice front by )=40 ns.

Figure 1. Snapshots of the (PVA-)ice-water systems viewed in the x-y plane at )=10, 30 and 60 ns as the ice crystal grew at 230 K. The kinetics of ice growth in the absence or presence of PVA has been quantified based on the averaged local bond order parameter ( (Figure 2). As seen in Figure 2, the PVA20-ice-water system constantly had fewer ice molecules than the other three systems (i.e., ice-water, PVA5-ice-water and PVA10-ice-water). For example, at )=70 ns, the ice phase in the PVA20-ice-water system contained only 11648 molecules, whereas the systems of PVA5-ice-water and PVA10-ice-water had 15194 and 14358 solid-like molecules, 30.4% and 23.3% more, respectively.

6


Page 6 of 15

Page 7 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 2. Ice growth kinetics in the absence or presence of PVA. The number of ice-like (either cubic or hexagonal) molecules (*+ ) is plotted as a function of time ()). During the propagation of ice, noticeable curvatures emerged as PVA20 adsorbed onto the ice front and excluded ice layers to grow on the sides. To better demonstrate this phenomenon, Figure 3 illustrates the adsorption of PVA20 to the ice front, showing two adjacent periodic images along the x axis. It is evident that there is a well-defined curvature (guided by the solid lines) on each side of the PVA20 molecule. Similar curvatures were also found in the z-y plane (See Figure S2 in SI). These curvatures can lead to the depression of the local equilibrium freezing temperature, thereby retarding ice growth, known as the Gibbs-Thomson effect.20, 41 The adsorption of PVA5 and PVA10 to the ice front was also found to generate similar curvatures. But these curvatures are short-lived and joined together easily due to the small size of PVA5 or PVA10 as the spacer (data not shown). Therefore, the synthetic ice recrystallization inhibitors must possess a large ice-binding area in the plane orthogonal to the direction of ice propagation to prevent the adjacent curvatures from merging, in order to maintain the GibbsThomson effect. These atomic-scale mechanisms are supported by the experimental findings in the literature. Congdon et al.11 observed that a minimum number of repeat units between 10 and 20 should be necessary for significant ice recrystallization inhibition by PVA. Similarly, it was found that the IRI performance of PVA increased with its molecular weight, i.e., the chain length.19

7


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. Illustration of the Gibbs-Thomson effect due to the adsorption of PVA20 to the ice front (x-y plane, time point: 60 ns). For an explicit demonstration, two adjacent periodic images along the x axis were joined together. When PVA5 molecules were incorporated into the ice lattice after 50 ns, they became significantly less mobile. In contrast, at least part of the PVA20 molecules maintain high mobility. Figure 4 compares the trajectories of the hydroxyl oxygen atoms of PVA5 and PVA20 in the x-y plane during the simulation period of 50-70 ns. The corresponding trajectories in the z-y plane can be found in Figure S3 in SI. As seen in Figure 4 A, all the hydroxyl oxygen atoms of PVA5 exhibit the same scale of movement as the ice oxygen atoms (shown in the circle in Figure 4 B). However, there are 6-10 hydroxyl oxygen atoms of PVA20 still showing high or moderate mobility (Figure 4 B). A closer look at the identities of these mobile oxygen atoms of PVA20 revealed that they are all exposed to the liquid phase (data not shown).

8


Page 8 of 15

Page 9 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 4. The x-y plane trajectories of the oxygen atoms of PVA5 (A) and PVA20 (B) over the period of 5070 ns. Two PVA5 or PVA20 molecules (either above or below the seed ice) are shown in each graph. The trajectories (blue) of two representative ice oxygen atoms are circled in B. Refer to Scheme 1 for the color code. To uncover the ice-binding mechanism of PVA, we superimposed the positions of the hydroxyl oxygen atoms (,-. ) of PVA5 onto the ice lattice at 60 ns (Figure 5). It has been shown in Figure 1 that the PVA5 molecules have been completely incorporated into the ice lattice by 60 ns. Also, a total of five oxygen atoms of PVA5 will not compromise the visual clarity of the superimposition. Interestingly, all the hydroxyl oxygen atoms of the top PVA5 molecule overlap with the oxygen atoms (,*+ ) constituting the ice lattice in both the basal (Figure 5 A) and secondary prism (Figure 5 B) planes. All hydroxyl oxygen atoms except for ,/ of the bottom PVA5 geometrically match the ice lattice by replacing the ice oxygen atoms to fit into the ice lattice. Such geometrical match has also been identified for ice-binding AF(G)Ps.19 Figure 5 C shows the radial distribution function of the atomic pair between ,-. of the bottom PVA5 and ,*+ , averaged over the simulation period of 50-70 ns. The blue line represents the 0 of the pair ,*+ -,*+ , showing the first primary peak around 2.75 Å, the first minimum around 3.5 Å, and the second primary peak around 4.5 Å. This agrees well with the experimental data42 (i.e., the grey line) as shown in the inset of Figure 5 C. It is reassuring that almost all the hydroxyl oxygen atoms of the bottom PVA5 reproduce the interactions with ,*+ largely the same way as the ice oxygen atoms do between themselves. The only exception is ,/ , showing neither minimum around 3.5 Å nor maximum around 4.5 Å, which agrees with the previous observation that ,/ is unable to fit into the ice lattice (Figure 5 A). Generally, it is clear from Figure 5 that the geometrical match between the hydroxyl groups of PVA and the ice lattice should be stereoscopic.

Figure 5. The geometrical match between the -OH groups of PVA5 and the ice lattice (A: x-y plane and B: z-y plane). The radial distribution function of the ,-, pair between the bottom PVA5 and its surrounding ice-like molecules (C). The blue line represents the 0 of the ,*+ -,*+ pair, the experimental counterpart of which is also given in the inset of C (the grey line). Refer to Scheme 1 for the color code To clarify the effect of oligomer tacticity on the ability of PVA to inhibit ice recrystallization, we conducted another simulation in which an atactic PVA20 molecule (Figure 6 A) was placed on each side of

9


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the seed ice. As shown in Figure 6 B, the atactic PVA20 molecules dock to the ice front and only allow the ice layers to grow on the sides, generating curvatures that lead to the Gibbs-Thomson effect. It is also found that the oxygen atoms of some of the -OH groups of PVA20, as indicated by the black arrows in Figure 6 C, overlap with the oxygen atoms of the ice front. For longer chains like PVA20, the stereoscopic ice-binding is facilitated by the intrinsic flexibility of the polymer backbone, in which the backbone adjusts its conformation to help the -OH groups fit the ice lattice. We speculate that such intrinsic flexibility could be critical for other relatively large IRI molecules that need to alter conformations to match the ice lattice for binding. Overall, the results shown in Figure 6 demonstrate that the oligomer tacticity of PVA would not significantly affect the ability to inhibit ice recrystallization.

Figure 6. The effect of oligomer tacticity on the ability of PVA to inhibit ice recrystallization. (A) An atactic PVA20 chain with a random arrangement of the -OH groups. (B) Illustration of the Gibbs-Thomson effect due to the PVA20-ice binding (x-y plane, time point: 40 ns) with two adjacent periodic images along the x axis joined together. (C) The geometrical match between some of the -OH groups of PVA20 (indicated by the black arrows) and the lattice of ice front. The hydrogen-bonding characterization is shown in Figure 7. As seen in Figure 7 A-C, each hydroxyl oxygen atom of a PVA molecule is able to donate its bonded hydrogen atom to a surrounding water/ice molecule to form one hydrogen bond, while as a 1-acceptor, it can form 1 or 2 hydrogen bond(s) with its surrounding water/ice molecule(s). In a stable ice lattice, each ice molecule typically forms two hydrogen bonds as 1-donor and two hydrogen bonds as 1-acceptor (Figure 7 D). Moreover, we suggest that the hydroxyl oxygen atom of PVA will form about three hydrogen bonds with the surrounding ice molecules (i.e., .2345 ≈3) when it fits into the ice lattice perfectly. Figure 7 A shows that both ,6 and , of PVA5 have about three hydrogen bonds with the surrounding ice molecules. Their corresponding 0 s show a primary peak around 2.75 Å, a minimum around 3.5 Å and a second primary peak around 4.5 Å (Figure 5 C), demonstrating the essential features of an ice lattice. However, a moderate match between ,-. and ,*+ (e.g., ,7 and ,8 ) or a complete mismatch (e.g., ,/ ) will contribute to .2345 ≤2. The validity of .2345 being an indicator for the fit of ,-. into the ice lattice was also confirmed for PVA20 on ,7 and ,/ which were found to perfectly fit into the ice lattice, and ,7/ which was highly mobile and exposed to the liquid phase (See Figure 7 C and Figure S4 in SI).

10


Page 10 of 15

Page 11 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 7. Hydrogen-bonding characterization. The number of hydrogen bonds formed between PVA (the one below the seed ice) and TIP4P/2005 (either water or ice). (A: PVA5; B: PVA10; C: PVA20) The number of hydrogen bonds formed between ice molecules (D). Grey bars: -OH group of PVA as the 1 acceptor; black bars: -OH group of PVA as the 1 donor. CONCLUSIONS Molecular dynamics investigations have been conducted to reproduce the ice-binding phenomenon of AF(G)Ps18 and graphene oxide25. AF(G)Ps have stable secondary and tertiary structures, yielding a rigid ice-binding plane, while graphene oxide has a structure of monomolecular sheet, serving as a natural plane for ice-binding. However, chain polymers like PVA are highly flexible, changing configurations frequently, especially in aqueous environments. Such intrinsic flexibility of chain polymers may invalidate any plain extrapolation of the molecular mechanisms learned about AF(G)Ps or graphene oxide. Previous speculation about the geometrical match between PVA and ice lattice was ideally based on a two-dimensional, straight-line configuration.20 In this study, by conducting MD simulations at the ice-PVA interfaces, we provided direct evidence of the adsorption of PVA to both the basal and prismatic ice faces and the engulfment of PVA into the ice lattice, showing the explicit representation of the geometrical match between the hydroxyl groups of PVA and the ice lattice. The PVA molecule must be large enough to prevent the merge of adjacent curvatures in the ice front that are otherwise separated, providing a hypothesis for the effect of the PVA chain length on the activity of ice recrystallization inhibition. As the current work is clarifying the molecular mechanisms of a well-established ice recrystallization inhibitor, our findings will help pave the path for addressing the pressing challenge in designing synthetic ice recrystallization inhibitors rationally, by enriching our mechanistic understanding of IRI process by macromolecules. For future studies, modifications to the PVA structure such as substituting acetate or methyl groups for hydroxyl would be useful for understanding the role of hydrophobicity in ice recrystallization inhibition. The simulations about the emerging AFP mimics,

11


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

especially small molecules, are also suggested, such as the carbohydrate and amino acid derivatives developed by Ben and his colleagues.43 SUPPORTING INFORMATION Supporting Information Available: Figures of the ice-binding phenomenon observed in an additional PVA20-ice-water system, the illustration of the Gibbs-Thomson effect, the z-y plane trajectories of the oxygen atoms of PVA5 and PVA20, and the radial distribution function of the ,-, pair between PVA20 and its surrounding water/ice molecules. This material is available free of charge via the Internet at http://pubs.acs.org. ACKNOWLEDGMENTS The authors would like to thank Dr. Jorge R. Espinosa and Dr. Carlos Vega from Universidad Complutense de Madrid for valuable suggestions on MD simulation of ice crystallization. The authors would also like to acknowledge the Enterprise Research Infrastructure & Services at Partners Healthcare for the provision of the ERISone cluster environment. Portions of this research were conducted on the Orchestra High Performance Compute Cluster at Harvard Medical School. This study was financially supported by NIH P41 EB002503 (MT). REFERENCES (1) Pearce, R. S., Plant freezing and damage. Ann. Bot 2001, 87, 417-424. (2) Kreder, M. J.; Alvarenga, J.; Kim, P.; Aizenberg, J., Design of anti-icing surfaces: smooth, textured or slippery? Nat. Rev. Mater 2016, 1, 15003. (3) Parent, O.; Ilinca, A., Anti-icing and de-icing techniques for wind turbines: Critical review. Cold Reg. Sci. Technol 2011, 65, 88-96. (4) Mazur, P., Cryobiology: the freezing of biological systems. Science 1970, 168, 939-949. (5) Fahy, G. M.; Wowk, B.; Wu, J.; Phan, J.; Rasch, C.; Chang, A.; Zendejas, E., Cryopreservation of organs by vitrification: perspectives and recent advances. Cryobiology 2004, 48, 157-178. (6) Fowler, A.; Toner, M., Cryo-injury and biopreservation. Ann. N. Y. Acad. Sci. 2006, 1066, 119-135. (7) Griffith, M.; Ala, P.; Yang, D. S.; Hon, W.-C.; Moffatt, B. A., Antifreeze protein produced endogenously in winter rye leaves. Plant Physiol. 1992, 100, 593-596. (8) Duman, J. G., Antifreeze and ice nucleator proteins in terrestrial arthropods. Annu. Rev. Physiol 2001, 63, 327-357. (9) Fletcher, G. L.; Hew, C. L.; Davies, P. L., Antifreeze proteins of teleost fishes. Annu. Rev. Physiol 2001, 63, 359-390. (10) Daley, M. E.; Spyracopoulos, L.; Jia, Z.; Davies, P. L.; Sykes, B. D., Structure and dynamics of a βhelical antifreeze protein. Biochemistry 2002, 41, 5515-5525. (11) Congdon, T.; Notman, R.; Gibson, M. I., Antifreeze (glyco) protein mimetic behavior of poly (vinyl alcohol): detailed structure ice recrystallization inhibition activity study. Biomacromolecules 2013, 14, 1578-1586. (12) Olijve, L. L.; Meister, K.; DeVries, A. L.; Duman, J. G.; Guo, S.; Bakker, H. J.; Voets, I. K., Blocking rapid ice crystal growth through nonbasal plane adsorption of antifreeze proteins. Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 3740-3745. (13) Raymond, J. A.; Wilson, P.; DeVries, A. L., Inhibition of growth of nonbasal planes in ice by fish antifreezes. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 881-885. (14) Knight, C.; Cheng, C.; DeVries, A., Adsorption of alpha-helical antifreeze peptides on specific ice crystal surface planes. Biophys. J. 1991, 59, 409-418.

12


Page 12 of 15

Page 13 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(15) Celik, Y.; Drori, R.; Pertaya-Braun, N.; Altan, A.; Barton, T.; Bar-Dolev, M.; Groisman, A.; Davies, P. L.; Braslavsky, I., Microfluidic experiments reveal that antifreeze proteins bound to ice crystals suffice to prevent their growth. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 1309-1314. (16) Nada, H.; Furukawa, Y., Growth inhibition mechanism of an ice–water interface by a mutant of winter flounder antifreeze protein: a molecular dynamics study. J. Phys. Chem. B 2008, 112, 7111-7119. (17) Nutt, D. R.; Smith, J. C., Dual function of the hydration layer around an antifreeze protein revealed by atomistic molecular dynamics simulations. J. Am. Chem. Soc 2008, 130, 13066-13073. (18) Kuiper, M. J.; Morton, C. J.; Abraham, S. E.; Gray-Weale, A., The biological function of an insect antifreeze protein simulated by molecular dynamics. Elife 2015, 4, e05142. (19) Inada, T.; Lu, S.-S., Inhibition of recrystallization of ice grains by adsorption of poly (vinyl alcohol) onto ice surfaces. Cryst. Growth Des. 2003, 3, 747-752. (20) Budke, C.; Koop, T., Ice recrystallization inhibition and molecular recognition of ice faces by poly (vinyl alcohol). ChemPhysChem 2006, 7, 2601-2606. (21) Deller, R. C.; Vatish, M.; Mitchell, D. A.; Gibson, M. I., Synthetic polymers enable non-vitreous cellular cryopreservation by reducing ice crystal growth during thawing. Nat. Commun. 2014, 5, 3244. (22) Mastai, Y.; Rudloff, J.; Cölfen, H.; Antonietti, M., Control over the structure of ice and water by block copolymer additives. ChemPhysChem 2002, 3, 119-123. (23) Mizrahy, O.; Bar-Dolev, M.; Guy, S.; Braslavsky, I., Inhibition of ice growth and recrystallization by zirconium acetate and zirconium acetate hydroxide. PLOS ONE 2013, 8, e59540. (24) Capicciotti, C. J.; Kurach, J. D.; Turner, T. R.; Mancini, R. S.; Acker, J. P.; Ben, R. N., Small molecule ice recrystallization inhibitors enable freezing of human red blood cells with reduced glycerol concentrations. Sci. Rep. 2015, 5, 9692. (25) Geng, H.; Liu, X.; Shi, G.; Bai, G.; Ma, J.; Chen, J.; Wu, Z.; Song, Y.; Fang, H.; Wang, J., Graphene Oxide Restricts Growth and Recrystallization of Ice Crystals. Angew. Chem. Int. Ed. 2017, 56, 997-1001. (26) Mitchell, D. E.; Clarkson, G.; Fox, D. J.; Vipond, R. A.; Scott, P.; Gibson, M. I., Antifreeze protein mimetic metallohelices with potent ice recrystallization inhibition activity. J. Am. Chem. Soc 2017, 139, 9835-9838. (27) Hayward, J.; Reimers, J., Unit cells for the simulation of hexagonal ice. J. Chem. Phys. 1997, 106, 1518-1529. (28) Abascal, J. L.; Vega, C., A general purpose model for the condensed phases of water: TIP4P/2005. J. Chem. Phys. 2005, 123, 234505. (29) Weiss, V. C.; Rullich, M.; Köhler, C.; Frauenheim, T., Kinetic aspects of the thermostatted growth of ice from supercooled water in simulations. J. Chem. Phys. 2011, 135, 034701. (30) Yesselman, J. D.; Price, D. J.; Knight, J. L.; Brooks, C. L., MATCH: An atom-typing toolset for molecular mechanics force fields. J. Comput. Chem. 2012, 33, 189-202. (31) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K., Scalable molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781-1802. (32) Ryckaert, J.-P., Special geometrical constraints in the molecular dynamics of chain molecules. Mol. Phys 1985, 55, 549-556. (33) Brünger, A. T., X-PLOR: version 3.1: a system for x-ray crystallography and NMR. Yale University Press: 1992. (34) Martyna, G. J.; Tobias, D. J.; Klein, M. L., Constant pressure molecular dynamics algorithms. J. Chem. Phys. 1994, 101, 4177-4189. (35) Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks, B. R., Constant pressure molecular dynamics simulation: the Langevin piston method. J. Chem. Phys. 1995, 103, 4613-4621. (36) Darden, T.; York, D.; Pedersen, L., Particle mesh Ewald: An N⋅log (N) method for Ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089-10092.

13


Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(37) Lechner, W.; Dellago, C., Accurate determination of crystal structures based on averaged local bond order parameters. J. Chem. Phys. 2008, 129, 114707. (38) Steinhardt, P. J.; Nelson, D. R.; Ronchetti, M., Bond-orientational order in liquids and glasses. Phys. Rev. B 1983, 28, 784. (39) Sanz, E.; Vega, C.; Espinosa, J.; Caballero-Bernal, R.; Abascal, J.; Valeriani, C., Homogeneous ice nucleation at moderate supercooling from molecular simulation. J. Am. Chem. Soc 2013, 135, 1500815017. (40) Weng, L.; Chen, C.; Zuo, J.; Li, W., Molecular dynamics study of effects of temperature and concentration on hydrogen-bond abilities of ethylene glycol and glycerol: implications for cryopreservation. J. Phys. Chem. A 2011, 115, 4729-4737. (41) Voets, I. K., From ice-binding proteins to bio-inspired antifreeze materials. Soft Matter 2017, 13, 4808-4823. (42) Soper, A., The radial distribution functions of water and ice from 220 to 673 K and at pressures up to 400 MPa. Chem. Phys. 2000, 258, 121-137. (43) Capicciotti, C. J.; Leclere, M.; Perras, F. A.; Bryce, D. L.; Paulin, H.; Harden, J.; Liu, Y.; Ben, R. N., Potent inhibition of ice recrystallization by low molecular weight carbohydrate-based surfactants and hydrogelators. Chem. Sci 2012, 3, 1408-1416.

14


Page 14 of 15

Page 15 of 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Table of Contents Graphic

15


* To whom correspondence should be addressed. Email: sstott@mgh

Recommend Documents