TETRAHYDROFURAN HYDRATE INHIBITORS:
ICE-ASSOCIATING BACTERIA & PROTEINS
by
Emily Irene Huva
A thesis submitted to the Department of Biology
In conformity with the requirements for
the degree of Master of Science
Queen’s University
Kingston, Ontario, Canada
(May, 2008)
Copyright ©Emily Irene Huva, 2008
ii
ABSTRACT
Ice-associating proteins (IAPs) are proteins that interact directly with ice crystals, either by
offering a site for nucleation, i.e. ice nucleating proteins (INPs), or by binding to nascent
crystals to prevent addition of more water molecules, i.e. antifreeze proteins (AFPs). AFPs
have been found to inhibit the formation of clathrate-hydrates, ice-like crystalline solids
composed of water-encaged guest molecules. Study of AFP-hydrate interaction is leading to
a greater understanding of AFP adsorption and of the mechanism behind the “memory
effect” in hydrates, wherein previously frozen crystals reform more quickly after a brief melt.
AFP is currently the only known memory inhibitor. Such a low-dosage hydrate inhibitor
(LDHI) is of great interest to the oil and gas industry, as hydrate formation and reformation
in the field is a huge problem. Bacterial AFPs, though largely uncharacterized, may be the
best candidates for large-scale production of hydrate inhibitors, given the difficulties in
obtaining AFP from other sources.
The popular kinetic inhibitors (KIs) polyvinylpyrrolidone (PVP) and
polyvinylcaprolactam (PVCap) were used for points of comparison in experiments exploring
the hydrate-inhibition activity of several ice-associating bacteria and proteins. The addition
of the soil microbe, Chryseobacterium, increased the average lag-time to tetrahydrofuran (THF)
hydrate formation by 14-fold, comparable to PVP or PVCap. Samples containing
Pseudomonas putida, a bacterium having both ice-nucleation protein (INP) and AFP activity,
had lag-times double that of the control. Solutions with P. putida and Chryseobacterium
sometimes formed hydrate slurries of stunted crystal nuclei instead of solid crystals. No
inhibition of memory or nucleation was noted in bacterial assays, however bacteria with INP
iii
activity was linked to unusually rapid memory reformation. Quartz crystal microbalance
experiments with dissipation (QCM-D) showed that a tight adsorption to SiO2 and resistance
to rinsing are correlated with a molecule’s inhibition of hydrate formation and reformation.
These results support a heterogeneous nucleation model of the memory effect, and point to
the affinity of AFP for heterogeneous nucleating particles as an important component of
memory inhibition.
iv
CO-AUTHORSHIP
Several researchers made large contributions to the work chronicled in this thesis and are
gratefully accredited co-authorship.
Chapter 2
Dr. H. Zeng wrote the paper and performed and guided the QCM-D analyses and carried
out the induction time experiments with PVP. Dr. Hailong Lu also did some of the QCM-D
experiments. Drs. V.K. Walker and J.A. Ripmeester were co-supervisors.
Chapter 3
I wrote the paper. Raimond V. Gordienko performed the THF hydrate affinity purification
experiments with recombinant proteins, including purification of all recombinant proteins
used. Drs. H. Zeng assembled the apparatus and contributed advice on the induction time
experiments. Drs. V.K. Walker and J.A. Ripmeester were co-supervisors.
Chapter 4
I wrote the paper. Dr. H. Zeng assembled the apparatus and carried out the PVP induction
time experiments, as well as giving counsel regarding the experiments. Drs. V.K. Walker and
J.A. Ripmeester were co-supervisors.
Appendix A
Virginia K. Walker wrote the paper. R.V. Gordienko obtained the IRI and INP data, Dr. H.
Zeng did the QCM-D analyses, Dr. M. J. Kuiper did initial experiments on the interactions
of PVP with ice (not shown). Drs. V.K. Walker and J.A. Ripmeester were co-supervisors.
v
ACKNOWLEDGEMENTS
I would first like to extend my sincere gratitude to my supervisor Dr. Virginia Walker for not
only her exceptional leadership and guidance, but also for her friendship and inspiration as
an incredibly creative and capable individual.
I also express my great appreciation of the help provided by my co-supervisor, Dr.
John Ripmeester, and for the excellent opportunity to engage in the fertile intellectual
environment that exists at the Steacie Institute of Molecular Sciences, National Research
Council (NRC). Dr. Ripmeester’s broad research expertise constituted an invaluable resource
that complemented that from Queen’s Biology Department, while his candid observations
about both science and life never ceased to generate a smile.
Thanks go as well to my committee members, Drs. R.S. Brown and D.B. Layzell, for
offering their time and varied expertise to provide sound feedback and guidance. For
indoctrinating me into hydrate research during my first NRC “summer camp,” and his
continued assistance and friendship from afar, I submit great thanks to Dr. H. Zeng.
The members of the Walker laboratory were a continual source of helpfulness and
friendliness; in particular, several experiments would not have been as organised or as fun
without the assistance and good humour of Mr. Raimond Gordienko. I would also like to
acknowledge Dr. G.R. Palmer for the subcooling apparatus and merry conversations about
physics, Mrs. Suzi Wu for her introduction into the world of microbiology, and Dr. M.J.
Kuiper for his guidance with crystal morphology experiments and inspiring outlook on
science.
Appreciation is also due to my many friends and colleagues at the National Research
vi
Council. Special gratitude is due to Dr. Stephen Lang, for answering innumerable
experimental questions, Mr. Sergey Mitlin and Dr. Chris Ratcliffe for helpful discussions,
and Mr. Adebola Adeyamo for his laboratory assistance as well as office humour. Mr.
Conrad Rock, Mr. Benoît Charpentier, and the employees of the Sussex ITS help desk are
thanked for their logistic support. I am also grateful to Dr. Susan Logan of the NRC
Institute for Biological Sciences for generously allowing me the use of her microbiology
facilities and equipment while at the NRC, and the members of her laboratory for their
friendly assistance.
Finally, I could not have achieved this academic success without the continued
support of my family and friends, in Kingston, Ottawa, and beyond.
vii
TABLE OF CONTENTS
Abstract .............................................................................................................. ii Co-Authorship ..............................................................................................................iv Acknowledgements...........................................................................................................v Table of Contents ...........................................................................................................vii List of Tables ...............................................................................................................x List of Figures ............................................................................................................. xi List of Abbreviations .................................................................................................... xiii Chapter 1 Introduction 1
1.1. Ice-Associating Proteins .........................................................................................................1 1.1.1. Ice crystallisation 1 1.1.2. Freeze avoidance and freeze tolerance 2 1.1.3. Antifreeze proteins 3 1.1.4. Ice nucleating proteins 4 1.1.5. Applications 5
1.2. Clathrate-Hydrates...................................................................................................................5 1.2.1. Hydrate crystallisation 5 1.2.2. Applications 7 1.2.3. Problematic formation 8
1.3. Control of Hydrate Formation by Ice-Associating Proteins .............................................8 1.3.1. Hydrate promotion 8 1.3.2. Hydrate inhibition 9
1.4. Research Objectives ..............................................................................................................10 1.5. Literature Cited ......................................................................................................................12
Chapter 2 Differences in Nucleator Adsorption May Explain Distinct Inhibition Activities of Two Gas Hydrate Kinetic Inhibitors. ..........................................................................20
2.1. Abstract ...................................................................................................................................20 2.2. Introduction............................................................................................................................21 2.3. Experimental ..........................................................................................................................22 2.4. Results and Discussion .........................................................................................................23 2.5. Conclusion ..............................................................................................................................25
Chapter 3 The Search For “Green Inhibitors:” Perturbing Hydrate Growth with Bugs............. 31
3.1. Abstract ...................................................................................................................................31 3.2. Introduction............................................................................................................................32 3.3. Methods...................................................................................................................................34
3.3.1. Hydrate formation and reformation in the presence of ice-associating bacteria 34
viii
3.3.2. The development of a clathrate hydrate affinity purification (CHAP) technique35 3.4. Results & Discussion.............................................................................................................36
3.4.1. THF hydrate formation and reformation 36 3.4.2. Developing clathrate hydrate affinity purification 38
3.5. Conclusion ..............................................................................................................................39 3.6. Acknowledgements ...............................................................................................................39 3.7. Literature Cited ......................................................................................................................40
Chapter 4 Tetrahydrofuran Hydrate Crystallisation and Memory with Biological and Synthetic Inhibitior-Promoter Combinations..............................................................................46
4.1. Abstract ...................................................................................................................................46 4.2. Introduction............................................................................................................................47 4.3. Materials and Methods..........................................................................................................48
4.3.1. Sample Preparation 48 4.3.2. Data Analysis 49
4.4. Results .....................................................................................................................................50 4.5. Discussion...............................................................................................................................52 4.6. Literature Cited ......................................................................................................................59
Chapter 5 General Discussion ...................................................................................................... 61
5.1. Hydrate Inhibition by antifreeze Bacteria & Proteins ......................................................61 5.2. Memory Effect Inhibition by Antifreeze Bacteria & Proteins........................................62 5.3. Identification of Novel Hydrate-Associating Molecules and Bacteria...........................63 5.4. Hydrate Inhibitors: Future Directions................................................................................64 5.5. Literature Cited ......................................................................................................................66
Summary .............................................................................................................69 Appendix A The Mysteries of Memory Effect and its Elimination with Antifreeze Proteins ........70
A.1 Abstract ....................................................................................................................................70 A.2 Introduction.............................................................................................................................71 A.3 Methods....................................................................................................................................73
A.3.1 Preparation and characterization of potential inhibitors 73 A.1.1 Quartz crystal microbalance assessments 74
A.2 Results ......................................................................................................................................75 A.2.1 Characterization of ice-associating properties 75 A.2.2 Quartz crystal microbalance assessments 75
A.3 Discussion................................................................................................................................76 A.4 Acknowledgements ................................................................................................................79 A.5 Literature CIted.......................................................................................................................80
Appendix B
ix
Induction Time Experiments with the Hydrate Promoter, Snomax ..........................87 B.1 Introduction and Methods.....................................................................................................87 B.2 Results and Discussion...........................................................................................................87
Appendix C Induction Time Experiments with the Water Viscosifier, Xanthan Gum ..................89
C.1 Introduction and Methods.....................................................................................................89 C.2 Results and Discussion...........................................................................................................89 C.3 Literature Cited........................................................................................................................91
Appendix D Power-Law Model of Tetrahydrofuran Hydrate Formation........................................92
D.1 Parameter calculations ...........................................................................................................92 D.2 Conversion of Parameters into “Lag-Time” ......................................................................92 D.3 Non-Isothermal Power-Law Model ....................................................................................93 D.4 Power-Law Fits for Multi-Phasic Freezing.........................................................................93 5.6. Error Calculations for k-Value Comparisons....................................................................94
Appendix E Assessment of Tetrahydrofuran Hydrate Promotion from Supercooling Measurements .............................................................................................................99
E.1 Introduction.............................................................................................................................99 E.2 Methods....................................................................................................................................99
E.2.1 Sample preparation 99 E.2.2 Supercooling of samples in bulk and in microcapillaries 100 E.2.3 Differential Scanning Calorimetry 100
E.3 Results and Discussion........................................................................................................ 101 E.3.1 Supercooling of samples in bulk and in microcapillaries 101 E.3.2 Differential Scanning Calorimetry 103
E.4 Literature Cited .................................................................................................................... 109 Appendix F Morphology of Tetrahydrofuran Crystals in the Presence of Biological Inhibitors of Ice ............................................................................................................110
F.1 Introduction .......................................................................................................................... 110 F.2 Methods ................................................................................................................................. 110 F.3 Results .................................................................................................................................... 110 F.4 Discussion ............................................................................................................................. 111 F.5 Literature Cited..................................................................................................................... 113
x
LIST OF TABLES
Table 3.1. Comparison of isothermal THF hydrate formation at 0oC for various bacterial cultures...........................................................................................................................42
Table 5.1. Association of hydrate and ice inhibitors with various crystalline structures and one non-crystalline surface (polystyrene)..................................................................67
Table A.1. Kinetic parameters for adsorption of antifreeze proteins (wild type AFP and A17L mutant) as well as kinetic inhibitors on SiO2 and polystyrene, a control hydrophobic surface, as assessed by QCM-D..........................................................82
Table D.1. Parameters for power-law fits to THF hydrate induction time data. ....................95 Table D.2 Power-law parameters for a multi-phase model of THF hydrate formation in the
presence of ice-associating bacteria. ..........................................................................96 Table E.1 Bacterial concentrations at the time of experiments (CFU = colony forming unit
per mL) ....................................................................................................................... 104 Table E.2. Measurements of Heterogeneous Nucleation by Differential Scanning
Calorimetry................................................................................................................. 105
xi
LIST OF FIGURES
Figure 1.1. In van der Waals-mediated adsorption by antifreeze proteins to ice, surface-surface complementarity is a primary requirement. (Davies et al., 2002).............17
Figure 1.2. A schematic of the adsorption-inhibition model of antifreeze protein (AFP) adsorption to a growing ice crystal............................................................................18
Figure 1.3. Different-sized cages make up each type of hydrate structure; structure h not shown. (Adapted from: Sloan, 2003) ........................................................................19
Figure 2.1. (a) Effects of PVCap and PVP on the induction time of THF hydrate and the memory effect, and (b) the adsorption masses of PVCap and PVP on silica surface............................................................................................................................28
Figure 2.2. The relationships between dissipation factor (D) and adsorption mass (m) of PVCap (a) and PVP (b), as well as the relationships between the final R values, R2, and the concentrations of the same two inhibitors (c). ...................................29
Figure 2.3. (a) A diagram of the effect of rinses on the adsorption mass, m, as determined by QCM-D and (b) the relationship between percentage of adsorption mass remaining (mf/m) and concentration of PVCap and PVP. ....................................30
Figure 3.1. Crystallized fraction, nc , of (a) freshly made and (b) pre-frozen samples that had formed THF hydrate after time t at 0oC...................................................................43
Figure 3.2. Hydrate and aqueous fractions of dilute (a) cyclopentane and (b) THF solutions containing bromophenol blue, after being nucleated and left unstirred at 3oC for several days. ..................................................................................................................44
Figure 3.3. Typical images of polycrystalline THF hydrate formed in the presence of (a) fish and (b) plant antifreeze proteins tagged with green-florescent protein (GFP), as visualized under UV light. ..........................................................................................45
Figure 4.1. Induction time t to THF hydrate formation and reformation in the presence of (a) biological ice-inhibitors and (b) synthetic hydrate inhibitors...........................56
Figure 4.2. Average lag-times, τ, to THF formation and reformation in the presence of biological ice-inhibitors and synthetic hydrate inhibitors. .....................................57
Figure 4.3. Differential activity of cultures during THF hydrate formation...........................58 Figure 5.1. Proposed high-pressure apparatus for gas hydrate affinity purification (J.A.
Ripmeester, personal communication) .....................................................................68 Figure A.1. Assessment of ice association of Type I AFP, A17L, and PVP. .........................83 Figure A.2. Representative graphs showing (a) the adsorption of AFP and PVP as well as
(b) AFP and A17L on the SiO2 surface at 299 K. Frequency shift (f) vs. time was assessed using QCM-D. ......................................................................................84
Figure A.3. Representative graph showing the relationship between dissipation factor (D) and the frequency (f) of PVP (left arrow-head), AFP (triangle) and A17L (octagon) all at 12.5 µM on the SiO2 surface at 299 K...........................................85
Figure A.4. A cartoon depicting the reformation of hydrates in the presence or absence of certain AFPs. ................................................................................................................86
xii
Figure B.1. THF hydrate induction times in the presence of Snomax and PVP...................88 Figure C.1. THF hydrate induction times with and without xanthan gum..............................90 Figure D.1. Power-law fit of induction times in the presence of ice-associating bacteria
showing contrast between formation (a, b) and reformation (c, d) of THF hydrate. ..........................................................................................................................97
Figure D.2. Power-law fit of induction times in the presence of hydrate-associating polymers for the formation and reformation of THF hydrate. ............................98
Figure E.1. Subcoolings of samples cooled at -0.2oC/min in (a) 1 mL bulk aliquots and (b) 10 µL microcapillaries. ............................................................................................. 106
Figure E.2. Differential freezing (a) and melting (b) of a set of 1 mL samples during a subcooling experiment, compared with (c) freezing during an isothermal induction-time experiment. ..................................................................................... 107
Figure E.3. Subcoolings of samples containing metal wires.................................................. 108 Figure F.1. Morphology of THF crystals in the presence of 109 CFU cultures of (a) E. coli,
(b) P. putida, (c) Chryseobacterium, and (d-h) 0.05 mM type III AFP. .................. 112
xiii
LIST OF ABBREVIATIONS
AFP antifreeze protein
CFU colony forming units per mL
CHAP clathrate-hydrate affinity purification
DSC differential scanning calorimeter
GFP green fluorescent protein
IAP ice-associating protein
INP ice-nucleating protein
IRI ice recrystallisation inhibition
KI kinetic inhibitor
LDHI low dosage hydrate inhibitor
ME memory effect
NMR nuclear magnetic resonance
PVCap polyvinylcaprolactam
PVP polyvinylpyrrolidone
QCM-D quartz crystal microbalance with dissipation
SDS sodium dodecyl sulphate
TH thermal hysteresis
THF tetrahydrofuran
Chapter 1
INTRODUCTION
1.1. ICE-ASSOCIATING PROTEINS
1.1.1. Ice crystallisation
The water ice we see most commonly, ice 1h, has a crystalline structure in the form of a
hexagonal cylinder. Zero degrees Celsius is its “melting point”, or, the temperature at which
a warmed single crystal disappears. The freezing point is defined as the temperature at which
a stable single ice crystal suddenly undergoes rapid growth (Zachariassen and Kristiansen,
2000). Ice crystals held at just below the melting point, however, will slowly grow larger and
encompass smaller crystals, a phenomenon referred to as “ice recrystallisation.”
The occurrence of a “nucleation” event happens on a case-by-case basis; it is the
point at which agglomerated water molecules first reach a stable (microscopic) critical radius
that allows the nascent crystal to grow. While nucleation is possible at the freezing point, it
normally requires the presence of nucleating particles to act as a scaffold for a new crystal to
build upon; these might take the form of dust, already formed ice, or even ice-nucleating
proteins (INPs). This type of crystal induction is called “heterogeneous nucleation”, whereas
“homogeneous nucleation,” relies on, in theory, the random agglomerations of pure water
molecules at very low temperatures (typically at -40oC in “pure” water). The gap between the
melting point and the heterogeneous nucleation temperature is the “supercooling” (also
called “subcooling”), which can vary over repeated experiments since heterogeneous
nucleation is a stochastic process.
1
2
Water’s melting point (and hence its highest possible freezing point) may be
colligatively depressed by the addition of solutes. Alternatively, antifreeze proteins (AFPs)
can kinetically depress the freezing point while leaving the melting point virtually unchanged;
the separation between the two is termed “thermal hysteresis” (TH) and is an inherent
characteristic of the system being tested (DeVries et al., 1970; Duman et al., 1993; Ewart et al.,
1999).
1.1.2. Freeze avoidance and freeze tolerance
The formation of ice inside a living organism is often damaging if not fatal. Hence, in order
to exploit cooler climates, many organisms have developed strategies for freeze avoidance
and/or freeze tolerance.
Freeze avoidance is achieved when an organism is able to lower the heterogeneous
nucleation point of its fluids, thereby increasing its ability to supercool and remain unfrozen
at subzero temperatures. Freeze avoidance is achieved through AFP-mediated thermal
hysteresis, the accumulation of solutes (sugars and “cryoprotectants” like glycerol and
ethylene glycol), and avoidance of nucleators like ice and food particles, as with insects’ use
of a waxy external cuticle and emptying of the gut (Storey and Storey, 1998; Worland and
Block, 2003). Metabolic activity of bacteria has been observed in Antarctic sea ice at
temperatures as low as -20oC (Junge et al., 2004).
Despite the success of freeze avoidance, freezing is sometimes inevitable. In this
case, organisms attempt to tolerate freezing and still avoid cellular damage. Methods include:
the induction of extracellular ice formation by INPs, the changes in cell membrane fluidity,
metabolic depression, and the suppression of ice recrystallisation by AFPs (Storey and
Storey, 1998). Some psychrophilic bacteria also produce cold-shock proteins and make
3
amino acid substitutions in others (Deming, 2002; Beja et al., 2002).
1.1.3. Antifreeze proteins
AFP activity has been found in fish, arthropods, plants, bacteria, and fungi (Devries et al.,
1970, Zachariassen and Husby, 1982; Duman and Olsen, 1993). Activity of AFP is diverse.
In general, the presence of physiological concentrations of AFP in fish results in TH of ~1-
1.5oC while in some insects AFP normally results in a TH of ~5oC. Plant AFP has lower TH
(<0.5oC) but has good ice recrystallisation inhibition, which may be more useful for freeze-
thaw survival (Kuiper et al., 2001). Bacterial TH is often low, however a purified AFP
isolated from an arctic bacterium, Marinomonas primoryensis, has TH of as much as ~2oC
(Gilbert et al., 2005). Antifreeze activity is typically assessed by three tests: ice recrystallisation
inhibition (IRI), ice shaping, and TH activity (Wilson et al., 2006). Although such activity has
been identified in over a dozen bacteria, few bacterial AFPs have been characterized (e.g. Sun
et al., 1995; Ku et al., 1998; Yamashita et al., 2002; Gilbert et al., 2005).
The structural qualities of AFPs that allow ice-binding appear to be varied (Ewart et
al., 1999). Some proteins contain threonine repeats thought to correspond to the regularity
of certain inter-atom distances found in ice crystals (e.g., type I fish AFP, fish AFGP), while
there is evidence that others originated from proteins recognizing other molecules, including
carbohydrates (e.g., the globular type II fish AFP) and chitin (e.g., a winter rye AFP). Type III
AFP, also globular, and type IV AFP, a four-helix bundle, are also found in fish (Chao et al.,
1993; Deng et al., 1997). Type I (isoform HPLC6) from winter flounder (Pleuronectes
americanus) is the most studied AFP, and is an alpha helix consisting of three alanine-rich
repeats.
The interaction between AFPs and the crystalline surface of ice seems to be based on
4
surface-surface complementarity and mediated primarily by van der Waals forces and
hydrophobicity (Figure 1.1; Davies et al., 2002). Inhibition of ice growth is thought to follow
an adsorption-inhibition model, wherein the addition of water molecules is restricted to a
curved growth front between bound AFPs which has a maximal radius (Figure 1.2;
Raymond and DeVries, 1977). Some researchers have suggested that AFPs may also bind to
heterogeneous nucleators (Zeng et al., 2006a) and inhibit INP activity (Parody-Morreale et al.,
1988; Zátmecník and Jánacek, 1992; Defalco, 2007).
1.1.4. Ice nucleating proteins
INPs were discovered after atmospheric nucleation of snow was linked to decomposing leaf
litter, which contained ice nucleating bacteria (Schnell and Vali, 1972). Since then, they have
been reported in several organisms, sometimes in the same species as AFPs (Zachariassen
and Kristiansen, 2000), but proof of their existence by expression of cloned sequences has
only been demonstrated in a few bacteria (e.g. Green et al., 1985; Arai et al., 1989; Zhao et al.,
1990; Michigami et al., 1994). Like their AFPs, insect INPs are typically active at lower
temperatures than are fish INPs.
The similarities between INPs and AFPs may in part extend to function. Some
researchers have posited that AFPs sometimes act as nucleators (Duman et al., 2004; Wang,
2000) and INP ice-binding ability may allow them to act as antifreezes (Duman et al., 1993;
Kobashigawa et al., 2005; Holt, 2003). As mentioned, INPs may be inhibited by AFPs
(Chapter 1.1.3), while synthetic polymers such as polyglycerol and polyvinyl alcohol have
also been shown to impede INP activity (Wowk and Fahy, 2002).
INPs are repetitive polypeptides typically larger than AFPs, and increase nucleation
probability by forming membrane-bound aggregates (Duman, 2001; Mueller et al., 1990;
5
Wolber et al., 1986). These aggregates seem to be limited in size which in turn limits INP
activity (Yeung et al., 1991). As such, each protein produces a characteristic maximal
subcooling: type I INP initiates freezing between -2 and -4oC; type II, between -5 to -7 oC;
and type III, between -8 to -10 oC (Yankofsky et al., 1981).
1.1.5. Applications
Study of AFPs and INPs, or ice-associating proteins (IAPs) should lead to a better
understanding of cold tolerance mechanisms, however they may also have roles in the
natural world beyond that of freeze tolerance. It is hypothesized that some plant pathogens
use INP to cause damaging ice formation on host plants, allowing entrance of the bacteria
(Lindow, 1983), while they regularly seed snowfall and hence play an important role in the
global water cycle (Schnell and Vali, 1972; Christner et al., 2008; Morris et al., 2008).
Practical applications of IAPs are numerous. INP from Pseudomonas syringae is used in
artificial snow-making (Mlot, 1984) and ice- mutant strains have been used to protect crops
from frost damage (Lindow and Panapoulos, 1988). AFPs have commercial potential in the
frozen foods industry, especially in ice recrystallisation inhibition in ice cream (Griffith and
Ewart, 1995; Regand and Goff, 2006), as well as the cryopreservation of organs (Amir et al.,
2005).
1.2. CLATHRATE-HYDRATES
1.2.1. Hydrate crystallisation
Clathrate-hydrates are crystals composed of water encaging unbound guest molecule(s)
(Figure 1.3). Literally hundreds of small molecular species may become hydrate guests under
appropriate conditions, including common gasses like N2, O2 and CO2, hydrocarbons such
6
as methane, propane, cyclopentane, and mixed hydrates formed from natural gas
(Gudmundsson et al., 1994). The available guests and local conditions dictate the hydrate
structure formed: structure I and structure II are cubic structures consisting of two
differently-sized cages, while structure H is hexagonal and has cages of three different sizes
(Ripmeester et al., 1987). Methane, for instance, usually forms structure I while propane
forms structure II, as does the popular model hydrate former tetrahydrofuran (THF).
Hydrate crystallisation normally entails low, above-zero temperatures and high pressures, as
well as interface between guest molecules and water, however THF is both miscible in water
and capable of forming octahedral hydrate crystals just below its freezing point of -4.4oC at
atmospheric pressure. THF can also be used to help induce gas hydrates by occupying large
cages and increasing stability.
Like ice crystallization, hydrate formation is subject to the purity of the forming
materials, usually water or ice and pressurized gas. Analogous to “subcooling” tests for
biological nucleators, the stochastic event of hydrate formation in isothermal experiments is
measured as “induction time,” the time between initial isothermal equilibrium of the bulk
sample (t0) and the point at which rapid, sudden crystallisation occurs as measured by a
sudden temperature or (for gas hydrates) pressure change (Zeng et al., 2006a, 2006b). Note
that induction time is not a direct measure of nucleation in the bulk sample, as nucleation is
a microscopic event.
“Memory effect” is a phenomenon unique to hydrates wherein previously frozen
samples recrystallise much more rapidly. The probability of reformation has been linked to
sample thermal history (Takeya et al., 2000; Ohmura et al., 2003), however the mechanism
remains unknown. Melting is complete and no residual structure among water molecules has
7
been observed (Buchanan et al., 2005); it is hypothesized that the memory effect occurs when
heterogeneous nucleators are temporarily “imprinted” by hydrate crystals and become more
conducive to re-nucleation (Zeng et al., 2006a). This theory is supported by propane hydrate
reformation experiments in which induction time but not growth was accelerated, and by
DSC experiments in which the homogeneous nucleation temperature of THF hydrate
remained unchanged when samples had been pre-frozen (Zeng et al., 2006a, 2006b).
1.2.2. Applications
Massive deposits of natural gas hydrate exist in offshore sediments and under permafrost,
and are regarded as very promising potential energy resources (Kvenvolden, 1993; Collett
and Kuuskraa, 1998). At the same time, the long-term stability of these hydrates is of
concern, as rising global temperatures threaten to free the greenhouse gasses they contain
(Kvenvolden, 1993).
Synthetic hydrates have applications as efficient transporters of natural gas
(Kvenvolden, 1993; Collett and Kuuskraa, 1998; Kanda et al., 2005; Sun et al., 2003). As well,
they might one day be used for long-term sequestration of CO2 in the deep sea (Baes et al.,
1980; Holder et al., 1995). Others have suggested using hydrates for H2 storage for fuel cells
and exploitation of their thermodynamic properties in refrigeration (Lee et al., 2005; Bi et al.,
2004).
Hydrates have also been employed in industrial separation processes. Although the
desalination of seawater via solute exclusion by growing hydrate crystals has so far proven
impractical (Chatti et al., 2005; Max, 2006), hydrate-based filtration of CO2 from industrial
flue gasses remains promising (Kang and Lee, 2000).
8
1.2.3. Problematic formation
Clathrate-hydrates were first identified by the oil and gas industry, where their formation in
deep-sea pipelines is still an expensive and dangerous problem. Blockage of pipelines
requires shutdowns of hours or days, resulting in huge financial losses. Both offshore drilling
and onshore drilling in the Artic encounter problematic hydrate formation, which can not
only cause pipeline shutdowns but also poses a hazard to human lives (Cullen, 1990). The
heat and pressure created by drilling has a complex interplay with hydrate-forming
conditions (Goodman and Franklin, 1981). Although future exploitation of natural gas
hydrate deposits for energy is enticing, the difficulty in controlling their decomposition as
well as unscheduled formation has made industry wary of harnessing their power. Evidently,
more knowledge of hydrate formation and inhibition is needed before these problems can be
righted.
1.3. CONTROL OF HYDRATE FORMATION BY ICE-ASSOCIATING
PROTEINS
1.3.1. Hydrate promotion
Methods to accelerate hydrate formation are needed for many commercial applications
(Chapter 1.2.2) as formation is relatively slow. As well, the use of reliable hydrate promoters
in inhibitor tests may better reveal inhibitory mechanisms; i.e. whether heterogeneous
nucleation or growth is prevented. Stirring allows mixing of the gas and liquid phases in gas
hydrate formation and causes agitation, which speeds hydrate formation (Hussain et al.,
2006). Researchers have employed surfactants, such as sodium dodecyl sulfate (SDS), as
promoters (Sun et al., 2003; Gayet et al., 2005), as well as rust and AgI (Wilson et al., 2005).
9
The association of bacteria with natural gas hydrate deposits prompted researchers to
investigate bacterially-derived surface-active polymers (Rogers et al., 1999). Commercial
biosurfactants from Bacillus subtilis, Corynebacterium lepus; Pseudomonas aeruginosa, and
Acinetobacter calcoaceticus, and Snomax, the commercial INP of P. syringae (Mlot, 1984), have
since been shown to accelerate the formation of CO2 and natural gas hydrates (Rogers et al.,
1999; Morgan et al., 1993; Rogers et al., 2003).
1.3.2. Hydrate inhibition
Hydrate formation in pipelines may be reduced through the insulation of pipelines, to avoid
hydrate formation conditions, and the use of thermodynamic inhibitors such as methanol
and glycols, to increase the temperature of formation. However, both of these methods are
very expensive, especially considering the large requirements of methanol (>50 wt%) and the
need for methanol recovery (Sloan, 2005).
Consequently, researchers have been turning to alternatives: namely, so-called low-
dosage hydrate inhibitors (LDHIs) in the form of anti-agglomerants or kinetic inhibitors
(Sloan, 2005; Paez et al., 2001). Kinetic inhibitors are commonly large synthetic polymers like
poly(N-vinylpyrrolidone) (PVP) and poly(N-vinylcaprolactam) (PVCap) which interfere
directly with hydrate formation. These two remain among the best (Wua, 2007).
The similarities between ice and hydrate, which are different versions of crystalline
water, led researchers to test the inhibition of hydrates by AFPs. Using induction time
measurements, a fish (type I winter flounder) and insect (Choristoneura fumiferama) AFP have
both been shown to inhibit THF hydrate formation at levels comparable to the commercial
inhibitor PVP (Zeng et al., 2003). AFPs have since been shown to inhibit gas hydrate
formation (Zeng et al., 2006b). For example, type I AFP increased induction time to
10
formation and “memory” reformation of propane hydrate in a pressure cell, and slowed
methane hydrate formation in a nuclear magnetic resonance (NMR) experiment (Zeng et al.,
2006b). Inhibition by type I and insect AFP was also demonstrated during THF reformation
experiments (Zeng et al., 2006a).
The mechanism of hydrate inhibition likely includes growth inhibition similar to that
for ice, as AFPs modified the morphology of nascent THF crystals in some experiments
(Zeng et al., 2003). Hydrate inhibition activity is not always correlated with ice inhibition
activity, however. A mutant type I AFP having leucine substituted for alanine at position 17
(A17L), which is on the ice-binding face, had no TH activity but inhibited THF hydrate
formation (Zeng et al., unpublished).
Such AFP studies may also help to elucidate the mechanism of “memory,” and its
inhibition. Although the memory effect significantly decreased propane induction times, the
hydrate growth rate remained the same, lending support to the theory that memory effect is
a heterogeneous nucleation phenomenon (Zeng et al., 2006b). The addition of AFP increased
induction times to formation and reduced formation rates for both fresh and previously
frozen samples, and prevented the acceleration of induction time due to memory. This was
interpreted as AFP inhibition of both heterogeneous nucleation and growth, especially in
light of the earlier THF experiments showing memory inhibition (Zeng et al., 2006a).
1.4. RESEARCH OBJECTIVES
The primary aims of this thesis are to further elucidate: the mechanism of hydrate inhibition
by antifreeze proteins; the phenomenon of memory effect inhibition by antifreeze proteins;
and to examine, for the first time, the potential of bacterial antifreezes to inhibit hydrate
formation.
11
Chapter 2 serves both to illuminate the mechanisms of hydrate formation and
inhibition, and to offer a point of comparison for biologically-derived inhibitors. It describes
the hydrate inhibition activity and adsorption capabilities of the commercial polymers PVP
and PVCap, as well as their ability to influence memory reformation of hydrate. Inhibition
activity is assayed by induction time experiments, while adsorption is determined using a
quartz-crystal microbalance; both techniques are used in the study of hydrate inhibition by
antifreeze proteins (Chapter 3, Appendix A, Zeng et al., 2006).
In Chapter 3, a new arena of microbial hydrate inhibition is opened. This chapter
presents the effects of several ice-associating bacteria species on THF hydrate induction
times and memory recrystallisation, offering further insight into the mechanism of hydrate
inhibition by AFPs. Additionally, a new method for isolating novel hydrate-associating
particles, proteins, and micro-organisms is explored.
Chapter 4 pertains to the effects of nucleator/inhibitor combinations, both bacterial
and synthetic, on THF hydrate formation and “memory” reformation. Quantitative
comparisons are facilitated by a numerical model of hydrate induction times, which helps to
clarify the subtle differences among data sets and hence the mechanisms behind them.
The Appendices catalogue experiments complementary to the primary goals of the
thesis, including further investigations of the memory effect (Appendix A), induction-time
experiments with potential promoters (Appendices B and C), subcooling assays designed to
detect hydrate promotion (Appendix E), and preliminary observations of THF hydrate
morphology in the presence of ice-associating bacteria and proteins (Appendix F). In
addition, Appendix D provides details of the model used for induction time analysis in
Chapter 4.
12
1.5. LITERATURE CITED
Amir, G., Rubinsky, B., Basheer, S.Y., Horowitz, L., Jonathan, L., Feinberg, M.S., Smolinsky, A.K., Lavee, J.. 2005. Improved Viability and Reduced Apoptosis in Sub-Zero 21-Hour Preservation of Transplanted Rat Hearts Using Anti-Freeze Proteins. The Journal of Heart and Lung Transplantation 24(11): 1915-1929
Arai, S., Abe, K., Watabe, S., Emori, Y., Watanabe, M. 1989. Molecular cloning of an ice nucleation gene from Erwinia ananas and its expression in Escherichia coli. FEMS Microbiology Letters 61(1-2) 53-56
Baes, C.F.Jr., Beall., S.E., Lee, D.W.. 1980. The collection, disposal, and storage of carbon dioxide. In: W. Bach, J. Pankrath and J. Williams, Editor, Interactions of Energy and Climate, D. Reidel Dordrecht, Holland: pp. 495–519.
Beja, O., Koonin, E.V., Aravind, L., Taylor, L.T., Seitz, H., Stein, J.L., Bensen, D.C., Feldman, R.A., Swanson, R.V., DeLong, E.F. 2002. Comparative genomic analysis of archaeal genotypic variants in a single population and in two different oceanic provinces. Applied and Environmental Microbiology 68: 335–345
Bi, Y., T. Guo, T. Zhu, S. Fan, D. Liang, L. Zhang 2004. Influence of volumetric-flow rate in the crystallizer on the gas-hydrate cool-storage process in a new gas-hydrate cool-storage system. Applied Energy 78:111-121
Brewer, P.G.. 2000. Gas hydrates and global climate change. Annals of the New York Academy of Science 912: 195-199
Buchanan, P., Soper, A.K., Thompson, H., Westacott, R.E.. 2005. Search for memory effects in methane hydrate: Structure of water before hydrate formation and after hydrate decomposition. Journal of Chemical Physics 123: 164507
Chaeo, H., Davies, P.L., Sykes, B.D., Sönnichsen, F.D.. 1993. Use of proline mutants to help solve the NMR solution structure of type III antifreeze protein. Protein Science 2(9): 1411-1428
Chatti, I., Delahaye, A., Fournaison, L., Petitet, J.P.. 2005. Benefits and drawbacks of clathrate hydrates: a review of their areas of interest. Energy Conversion and Management 46(9-10): 1333-1343.
Christner, B.C., Morris, C.E., Foreman, C.M., Cai, R., Sands, D.C.. 2008. Ubiquity of Biological Ice Nucleators in Snowfall. Science 319: (5867), 1214
Collett, T. S., Kuuskraa, V.A. 1998. Hydrates contain vast store of world gas resources. Oil and Gas Journal 96(19): 90-95
Cullen, W. D.. 1990. The public inquiry into the Piper Alpha disaster. HMSO, London.
Davies P.L, Baardsnes, J., Kuiper, M.J., Walker, V.K.. 2002. Structure and function of antifreeze proteins. Philosophical Transactions of the Royal Society B 357(1423): 927-935
Defalco, T.. 2007. Undergraduate Thesis. Department of Biology, Queen’s University, Kingston, Canada.
13
Deming, J.W.. 2002. Psychrophiles and polar regions. Current Opinion in Microbiology 5(3): 301-309.
Deng, G., D. W. Andrews, and R. A. Laursen. 1997. Amino acid sequence of a new type of antifreeze protein, from the longhorn sculpin Myoxocephalus octodecimspinosis. FEBS Letters 402: 17-20
DeVries, A. L., Komatsu, S. K., Feeney, R. E. (1970) Chemical and physical properties of freezing point-depressing glycoproteins from Antarctic fishes. Journal of Biological Chemistry 245: 2901-2908
Duman, J.G. and Olsen, T.M. 1993. Thermal hysteresis protein activity in bacteria, fungi, and phylogenetically diverse plants. Cryobiology 30: 322-328.
Duman, J.G., 2001. Antifreeze and ice nucleator proteins in terrestrial arthropods. Annual Review of Physiology 63, pp. 327–357.
Duman, J.G., Bennett, V., Sformo, T., Hochstrasser, R., Barnes, B.M.. 2004. Antifreeze proteins in Alaskan insects and spiders. Journal of Insect Physilogy 50(4) 259-266.
Duman, J.G., Wu, D.W., Olsen, T.M., Urrutia, M. and Tursman, D. 1993. Thermal-hysteresis proteins. In: P.L. Steponkus, Editor, Advances in Low-Temperature Biology, Volume 2, Elsevier Science, JAI Press, London: 131-182.
Ewart K.V., Lin, Q., Hew, C.L.. 1999 Review: Structure, function and evolution of antifreeze proteins. Cellular and Molecular Life Sciences 55: 271–283
Gayet, P., Dicharry, C., Marion, G., Graciaa, A., Lachaise, J., Nesterov, A.. 2005. Experimental determination of methane hydrate dissociation curve up to 55 MPa by using a small amount of surfactant as hydrate promoter. Chemical Engineering Science 60: 5751-5758.
Goodman, M. A., Franklin, L. Thermal Model of a New Concept for Hydrate Control During Drilling. In: Proceedings of the Fourth Canadian Permafrost Conference, March 2-6, 1981, Calgary, Canada
Green, R.L., Warren, G.J.. 1985. Physical and functional repetition in a bacterial ice nucleation gene. Nature 317: 645 - 648
Griffith, M., Ewart, K.V., 1995. Antifreeze proteins and their potential use in frozen foods, Biotechnology Advances 13(3): 375-402.
Gudmundsson, J.S., Khokhar, A.A., Parlaktuna, M.. 1994. Storing natural gas as frozen hydrate. Society of Petroleum Engineers Production and Facilities 9(1): 69-73
Holder, G.D., Cugini, A.V., Warzinski, R.P.. 1995. Modeling clathrate hydrate formation during carbon dioxide injection into the ocean. Environmental Science Technology 29:276-278
Holt, C.B.. 2003. Substances which inhibit ice nucleation: A review. Cryoletters, 24(5): 269-274
Hussain, S.M.T., Kumar, A., Laik, S., Mandal, A., Ahmad., I.. 2006. Study of the kinetics and morphology of gas hydrate formation. Chemical Engineering and Technology 29(8): 937-943
Jack A. Gilbert, Peter L. Davies, Johanna Laybourn-Parry. 2005 A hyperactive, Ca2+-
14
dependent antifreeze protein in an Antarctic bacterium. FEMS Microbiology Letters 245:1 67
Junge, K., Eicken, H., Deming, J.W.. 2004. Bacterial Activity at −2 to −20°C in Arctic Wintertime Sea Ice. Applied and Environmental Microbiology. 70(1): 550–557.
Kanda H., Uchida K., Nakamura K., Suzuki T., Economics and Energy Requirements on Natural Gas Ocean Transportation in Form of Natural Gas Hydrate (NGH) pellets. In: Proceedings of the 5th International Conference on Gas Hydrates, Trondheim, Norway, 2005.
Kang, S.P., Lee, H.. 2000. Recovery of CO2 from flue gas using gas hydrate: thermodynamic verification through phase equilibrium measurements. Environmental Science Technology 34: 4397-4400
Kobashigawa, Y., Nishimiya, Y., Miura, K., Ohgiya, S., Miura, A., Tsuda, S.. 2005. A part of ice nucleation protein exhibits the ice-binding ability. FEBS Letters 579(6): 1493 - 1497
Kuiper, M. J., Davies, P.L., Walker, V.K.. 2001. A Theoretical Model of a Plant Antifreeze Protein from Lolium perenne. Biophysics Journal 81: 3560-3565
Kvenvolden, K.A.. 1993. Gas hydrates as a potential energy resource - a review of their methane content. In D. G. Howell, Editor, The Future of Energy Gases - U.S. Geological Survey Professional Paper 1570, United States Government Printing Office, Washington: pp. 555-561
Lee, H., Lee, J., Kim, D.Y., Park, J., Seo, Y.T., Zeng, H., Moudrakovski, I.L., Ratcliffe, C.I., Ripmeester, J.A.. 2005. Tuning clathrate hydrates for hydrogen storage. Nature 434:743-746
Lindow, S.E., Panapoulos, N.J.. 1988. Field test of recombinant Ice- Pseudomonas syringae for biological frost control in potato. In: Sussman, M., Collins, C.H., Skinner, F.A., Stewart-Tull, D.E. Eds. The Release of Genetically-Engineered Micro-Organisms, Academic, London: pp. 121-38.
Lindow, S.E.. 1983. The role of bacterial ice nucleation in frost injury to plants. Annual Review of Phytopathology. 21:363-84
Max, M.D.. 2006. Hydrate desalination for water purification. U.S. Patent #6991722
Michigami, Y., Watabe, S., Abe, K., Obata, H., Arai, S.. 1994. Cloning and sequencing of an ice nucleation active gene of Erwinia uredovora. Bioscience, Biotechnology, and Biochemistry 58(4):762-764
Mlot, C.. 1984. Making Snow the Microbial Way. Science News 126(17): 263
Morgan, J.J., Blackwell, V.R., Johnson, D.E., Spencer, D.F., North, W.J.. 1999. Hydrate formation from gaseous CO2 and water. Environmental Science Technology 33: 1448-1452
Morris, C.E., Sands, D.C., Vinatzer, B.A., Glaux, C., Guilbaud, C., Buffière, A., Yan, S., Dominguez, H., Thompson, B.M.. 2008. The life history of the plant pathogen Pseudomonas syringae is linked to the water cycle. The ISME Journal 2: 321–334
Mueller, G.M., Wolber, P.K., Warren, G.J.. 1990. Clustering of ice nucleation protein
15
correlates with ice nucleation activity. Cryobiology 27: 416–422
Muryoi, N., Sato, M., Kaneko, S., Kawahara, H., Obata, H., Yaish, M.W.F., Griffith, M., Glick, B.R.. 2004.Cloning and expression of afpA, a gene encoding an antifreeze protein from the arctic plant growth-promoting rhizobacterium Pseudomonas putida GR12-2. Journal of Bacteriology 186(17): 5661-5671
Ohmura, R., Ogawa, M., Yasuoka, K., Mori, Y.H.. 2003 Statistical study of clathrate-hydrate nucleation in a water/hydrochlorofluorocarbon system: Search for the nature of the "memory effect." Journal of Physical Chemistry B 107(22): 5289-5293
Paez, J. E., Blok, R., Vaziri, H., Islam, M. R.. 2001. Practical Guidelines for Field Remediation, Journal of Petroleum Technology 53(10): 29
Parody-Morreale, A., Murphy, K.P., Di Cera, E., Fall, R., DeVries, A.L. and Gill, S.J. 1988. Inhibition of bacterial ice nucleators by fish antifreeze glycoproteins. Nature 333: 782-783.
Raymond, J.A. and DeVries, A.L. 1977. Adsoption inhibition as a mechanism of freezing resistance in polar fishes. Proceedings of the National Academy of Sciences USA 74: 2589-2593.
Regand, A., Goff, H. D. 2006. Ice Recrystallisation Inhibition in Ice Cream as Affected by Ice Structuring Proteins from Winter Wheat Grass. Journal of Dairy Science 89: 49-57
Ripmeester, J.A., Tse, J.S., Ratcliffe, C.I., Powell, B.M.. 1987. A new clathrate hydrate structure. Nature 325, 135 – 136
Rogers R.E., Zhang G, Dearman J, Woods C. 2007. Investigations into surfactant/gas hydrate relationship. Journal of Petroleum Science and Engineering 56: 82-88
Rogers, R.E., Kothapalli, C., Lee, M.S., Woolsey, R.J.. 2003. Catalysis of gas hydrates by biosurfactants in seawater-saturated sand/clay. Canadian Journal of Chemical Engineering 81: 1-8
Schnell, R.C., Vali, G.. 1972. Atmospheric Ice Nuclei from Decomposing Vegetation. Nature 236: 163 - 165
Sloan, E.D.. 2003. Fundamental principles and applications of natural gas hydrates. Nature 426: 353-359
Sloan, E.D. 2005. A changing hydrate paradigm--from apprehension to avoidance to risk management. Fluid Phase Equilibria 228-229 (PPEPPD 2004 Proceedings): 67-74
Storey, K.B., Storey, J. M.. 1988. Freeze tolerance in animals. Physiological Reviews 68: 27-84
Sun, Z., Wang, R., Ma, R., Guo, K., Fan, S.. 2003. Natural gas storage in hydrates with the presence of promoters. Energy Conversion and Management 44: 2733-2742
Takeya, S., Hori, A., Hondoh, T., Uchida, T.. 2000. Freezing-Memory Effect of Water on Nucleation of CO2 Hydrate Crystals. Journal of Physical Chemistry B 104(17) 4164-4168
Walker, V.K., Palmer, G.R., Voordouw, G.. 2006. Freeze-Thaw Tolerance and Clues to the Winter Survival of a Soil Community. Applied and Environmental Microbiology 72: 1784-1792
Wang, J.H.. 2000. A Comprehensive Evaluation of the Effects and Mechanisms of Antifreeze Proteins during Low-Temperature Preservation. Cryobiology 41(1) 1-9.
16
Wilson, P.W., Lester, D.J., Haymet, A.D.J.. 2005. Heterogeneous nucleation of clathrates from supercooled tetrahydrofuran (THF)/water mixtures, and the effect of an added catalyst. Chemical Engineering Science 60: 2937-2941.
Wilson, S.L., Kelley, D.L., Walker, V.K.. 2006. Ice-active characteristics of soil bacteria selected by ice-affinity. Environmental Microbiology 8(10): 1816–1824
Wolber, P.K., Deininger, C.A., Southworth, M.W., Vandekerckhove, J., van Montagu, M., Warren, G.J. 1986. Identification and Purification of a Bacterial Ice-Nucleation Protein. Proceedings of the National Academy of Sciences USA 83(19): 7256-7260
Worland, M.R., Block, W.. 2003. Desiccation stress at sub-zero temperatures in polar terrestrial arthropods. Journal of Insect Physiology 49(3): 193-203.
Wowk, B., Fahy, G.M.. 2002. Inhibition of bacterial ice nucleation by polyglycerol polymers Cryobiology 44(1): 14-23
Wua, M., Wangb, S., Liuc, H.. 2007. A Study on Inhibitors for the Prevention of Hydrate Formation in Gas Transmission Pipeline. Journal of Natural Gas Chemistry 16(1): 81-85
Yankofsky, S. A., Levin, Z., Bertold, T., Sandlerman, N. 1981. Some basic characteristics of bacterial freezing nuclei. Journal of Applied Meteorology 20: 1013–1019
Yeung, K.L., Wolf, E.E., Duman, J.G.A.. 1991. Scanning tunneling microscopy study of an insect lipoprotein ice nucleator. Journal of Vacuum Science and Technology 9: 1197–1201
Zachariassen, K. E., Husby, J. A.. 1982. Antifreeze effect of thermal hysteresis agent protects highly supercooled insects. Nature 298:865–867.
Zátmecník, J., Jánacek, J.. 1992. Interaction of antifreeze proteins from cold hardened cereal seedlings with ice nucleation active bacteria. Cryobiology 29:718-719.
Zeng, H., Moudrakovski, I., Walker, V.K., Ripmeester, J. A., 2006b. Inhibition activity of an antifreeze protein on hydrocarbon hydrate formation. American Institute of Chemical Engineers Journal. 52: 3304-3309.
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J. A., 2006a. Effect of antifreeze proteins on the formation and reformation of tetrahydrofuran clathrate hydrate. Journal of the American Chemical Society 128: 2844-2850
Zhao, J.I., Orser, C.S.. 1990. Conserved repetition in the ice nucleation gene inaX from Xanthomonas campestris pv. translucens. Molecular and General Genetics 223(1): 163-166
17
Figure 1.1. In van der Waals-mediated adsorption of antifreeze proteins to ice, surface-surface complementarity is a requirement. (Davies et al., 2002)
18
Ice Growth
AFPSolution
Figure 1.2. A schematic of the adsorption-inhibition model, showing two antifreeze proteins (AFPs) binding to a growing ice crystal.
19
Figure 1.3. Different-sized cages make up each type of hydrate structure; structure h not shown. (Adapted from: Sloan, 2003)
20
Chapter 2
DIFFERENCES IN NUCLEATOR ADSORPTION MAY
EXPLAIN DISTINCT INHIBITION ACTIVITIES OF TWO
GAS HYDRATE KINETIC INHIBITORS.
Huang Zeng1, Hailong Lu1, Emily Huva2, Virginia K. Walker2,3 and John A. Ripmeester1,2
1 National Research Council of Canada, Ottawa, Ontario, K1A 0R6, Canada; 2 Department
of Biology, 3Department of Microbiology & Immunology, Queen’s University, Kingston,
Ontario, K7L 3N6, Canada
In: Chemical Engineering Science (in press: accepted April 2008, 9 pp)
2.1. ABSTRACT
Pipeline blockage by gas hydrate is a serious problem in the petroleum industry. Recently
low-dosage inhibitors have been developed. In particular, poly(N-vinylcaprolactam) (PVCap)
is a stronger inhibitor than poly(N-vinylpyrrolidone) (PVP). In this study, PVCap was also
found to have stronger inhibition activity compared to PVP, but it was less effective during
reformation of hydrate. To understand the mechanism, the adsorption of PVCap and PVP
on silica, a common nucleating agent, was examined using a quartz crystal microbalance with
the dissipation factor observation function (QCM-D). The results reveal that PVP forms a
21
loose film on silica whereas PVCap forms a relatively more rigid and compact film.
However, most of the PVCap film could be rinsed off. These results help explain the
different inhibition activities of PVCap and PVP.
2.2. INTRODUCTION
Gas hydrates are ice-like solids with guest molecules trapped in the frameworks of hydrogen-
bonded water molecules, and the formation of gas hydrates can cause serious clogging
problems during oil and gas production and transportation (Sloan, 1998). A common
method to prevent hydrate formation is to use large quantities of thermodynamic inhibitors
such as methanol (Koh et al., 2002). Recently, two groups of low-dosage hydrate inhibitors
(LDHIs) active at concentrations below 1 wt% have been developed: anti-agglomerants
(AAs) keep small hydrate particles dispersed (Kelland, 2006), and kinetic inhibitors (KIs)
retard hydrate formation (Sloan, 1998; Kelland, 2006). Some antifreeze proteins have been
also shown to be effective LDHIs (Zeng et al., 2006a, b). The mechanism of the LDHIs is
not well understood. Some studies related the inhibition activities to the effects of the
LDHIs on the water structures during homogeneous nucleation which prevent the
formation of the critical nuclei (Kelland, 2007; Moon et al., 2007). Some studies suggested
that the nucleation and/or crystal growth inhibition is achieved via adsorption on hydrate
surface (Larsen et al., 1998, Hutter et al., 2000; Moon, et al., 2007). Among the KIs tested,
poly(N-vinylcaprolactam)(PVCap) has been recognized to be more effective than poly(N-
vinylpyrrolidone) (PVP) (Lederhos et al., 1996). However, knowledge about their effects on
the unavoidable heterogeneous nucleation of gas hydrate is limited (Colles et al., 1999).
The diffusion barrier due to hydrate formation on the gas-liquid interface makes testing gas
hydrates very time-consuming. Thus, tetrahydrofuran (THF) hydrate has been employed as a
22
model hydrate for the inhibitor testing (Makogon et al., 1997). THF hydrate growth can be
inhibited by the same KIs known to be effective against gas hydrates and it also shows the
same memory effect, where recrystallisation occurs rapidly after a brief melting period (Zeng
et al., 2006a). Hydrate reformation in this case is due to heterogeneous, not homogenous
nucleation. Silica is a known ice-nucleating material and this nucleation can occur through
self-structuring around active centers at the silica surface (Klier et al., 1973). Considering the
similarity between ice and hydrate surface, it is likely that gas hydrate can form on ubiquitous
silica surfaces. In this study, the effects of PVP and PVCap on heterogeneous nucleation and
reformation of THF hydrate was examined. The adsorption of these two KIs on silica was
subsequently compared using a quartz crystal microbalance capable of monitoring the
adlayer status. The results provide important insights into the inhibition of heterogeneous
nucleation of gas hydrate, providing useful suggestions for future design of LDHIs.
2.3. EXPERIMENTAL
PVP (dry powder, mw ~40,000) was kindly provided by Dr. E. D. Sloan (Colorado School
of Mines). PVCap (dry powder, mw ~110,000) was kindly provided by Dr. L. Talley (Exxon
Mobil). The method to measure the induction time (ti) of THF hydrate was described
elsewhere (Zeng et al., 2006a). Investigations were carried out with THF solutions containing
PVCap or PVP at 10.0 mg/ml, a concentration previously proved effective (Larsen et al.,
1998; Makogon et al., 1997).
Surface adsorption was determined using a quartz crystal microbalance capable of
determining the energy loss, or dissipation factor (D)(QCM-D) (Rodahl, et al., 1995, 1997).
PVCap/PVP solutions were prepared with ultrapure water (18.2 mΩ·cm at 298 K) at 0.25,
23
0.5, 1.0, 2.0, 5.0 and 10.0 mg/ml. All measurements were conducted with a QCM-D(Q-
Sense D300, Q-Sense AB) with 5-MHz AT-cut quartz crystal coated with SiO2 at 299.0±0.02
K (details of the procedure were given in Zeng et al., 2007) . After the adsorption, the crystal
was rinsed with 0.5 ml ultrapure water for three times. Three individual samples were
measured for each concentration point.
2.4. RESULTS AND DISCUSSION
Induction times (ti) of THF hydrate formation in THF solutions containing PVP or PVCap
(10mg/ml) at 273 K are shown in Fig. 1A. Nt is the number of samples remaining un-
crystallized at time t, and N0 is the total sample number (~50). Therefore, Nt/N0 vs. time
represents the inhibition activity of the additive. The slower the curve declines, the stronger
the inhibition is. PVCap showed stronger inhibition activity for THF hydrate formation than
PVP. After 24 h, 41% of the PVCap-containing samples were hydrate-free compared to 32%
of the PVP-containing samples (Figure 2.1a).
The “memory effect” for hydrate formation is demonstrated when shorter induction
times are required in samples that have previously been crystallized and subsequently melted
for a certain period (Sloan, 1998). In this experiment, THF hydrate formed in a KI-
containing solution was melted at 279 K and kept at this temperature for 1h before cooling
to 273 K. Induction times for the recrystallisation were directly compared to the times for
the original solutions. Hydrate reformation in the presence of the two KIs was distinct. All
of the PVCap-containing samples recrystallized after 1h whereas 91.5% of the PVP-
containing samples formed hydrates at that point (Figure 2.1a). Remarkably, about 4% of the
PVP samples remained hydrate-free after 10h (Figure 2.1a). Thus, although neither KI
eliminates memory effect, PVP appears to be a better inhibitor than PVCap during the
24
reformation of THF hydrate.
It is well-known that a “sympathetic” surface can induce heterogeneous nucleation.
Thus, it is reasonable to assume that the probability of hydrate formation will be reduced if
the nucleating surface (e.g. hydrated oxides of Si or Fe) is covered by an inhibitor. Therefore,
an inhibitor’s adsorption properties on a nucleator should be correlated with its inhibition
activity. As expected, as the KI concentrations increased, QCM-D indicated more molecules
adsorbed on silica (Figure 2.1b). However, at any tested concentration, the adsorption
masses for PVP were higher than for PVCap although PVCap is a stronger inhibitor than
PVP. To understand this, one can analyze the viscoelastic properties of the adsorbed
molecules revealed by QCM-D. The ratio of the change of dissipation factor and adsorption
mass, R, calculated as R = ΔD/Δm, is a measure of the adlayer status. A large absolute R
value indicates a porous, flexible adlayer with considerable trapped liquid (Rodahl, et al.,
1995, 1997). It is notable that PVCap and PVP showed two distinct steps of adsorption,
indicating a rearrangement of the adsorbed layer as adsorption progressed, and R2 represents
the final status of the adlayer
For hydrate crystallisation to occur, sufficient water and THF molecules have to
reach nucleating sites. Based on our studies on antifreeze proteins, we proposed three
important factors for a LDHI to inhibit the gas hydrate: larger adsorption mass and rigid
adlayer contributes to higher inhibition activity, and stronger affinity with nucleating surface
contributes to the resistance to the reformation (Zeng et al., 2007). In the presence of
PVCap, the nucleating surface is covered with a compact PVCap film, and water and THF
molecules cannot so easily reach the nucleus, making heterogeneous nucleation less
probable. In contrast, because PVP forms a looser layer with more trapped solution,
25
crystallisation will more readily occur although the adsorption mass is greater. When the KI-
coated surface was rinsed (similar to the situation when gas hydrate is decomposed at
modest conditions and the surface of silica trapped inside gas hydrate is “rinsed” due to the
movement of the melted solution), however, the adsorption masses significantly changed.
More than 80% of PVCap was rinsed away compared to only about 17% of PVP (Figure
2.3). This helps to explain the relatively higher inhibition activity of PVP compared to
PVCap during recrystallisation (Figure 2.1b), indicating its higher affinity for the SiO2 surface
than PVCap. Similarly PVP may have a higher affinity for the hydrate residual clusters that
may be present after melting. It is possible too, that PVP associates with these clusters and
interferes with their re-assemblage when hydrate-favoring conditions are re-established. This
could potentially result in a change in the recrystallisation kinetics, and could resemble a type
of inhibition of “memory effect”. Indeed, the presence of PVP seemed to slow the
recrystallisation of THF hydrate (Figure 2.3), although it did not eliminate memory effect, as
did selected antifreeze proteins (Zeng et al., 2006a, 2006b, 2007).
These results suggest the importance of LDHIs on the heterogeneous nucleation of gas
hydrate. Silica is used as a model for the nucleators, other nucleating surfaces can also induce
the heterogeneous nucleation for hydrate formation. The system is under modification for
test on other nucleators.
2.5. CONCLUSION
The properties of adsorbed KI layers can be monitored effectively by QCM-D and the
results have provided useful information about the inhibition mechanism of the
heterogeneous nucleation of clathrate hydrate. It also offers an option for efficient large-
scale screening of potential LDHIs so that structure-function relationships can be
26
established to design better inhibitors.
27
2.6. LITERATURE CITED
Colle, K.S., Oelfke, R.H., Kelland, M.A., US Patent 5874660, February 23, 1999.
Hutter, J. L., King, H. E., Lin, M. Y., 2000. Polymeric Hydrate-Inhibitor Adsorption Measured by Neutron Scattering. Macromolecules 33: 2670-2679.
Kelland, M. A., 2006, History of the development of low dosage hydrate inhibitors. Energy and Fuels 20: 825-847.
Klier, K., Shen J. H., Zettlemoyer, A. C., 1973. Water on silica and silicate surfaces. I. Partially hydrophobic silicas. Journal of Physical Chemistry 77: 1458-1465.
Koh, C.A., Westacott, R.E., Zhang, W., Hirachand, K., Creek, J.L., Soper, A.K., 2002. Mechanisms of gas hydrate formation and inhibition. Fluid Phase Equilibria 194:143-151.
Larsen, R., Knight, C.A., Sloan, E.D. Jr., 1998. Clathrate hydrate growth and inhibition. Fluid Phase Equilibria 150-151: 353-360.
Lederhos, J.P., Long, J.P., Sum, A., Christiansen, R.L., Sloan, E.D. Jr., 1996. Effective kinetic inhibition for natural gas hydrates. Chemical Engineering Science 51: 1221-1229.
Makogon, T. Y., Knight, C. A., Sloan, E. D. Jr., 1997. Melt growth of tetrahydrofuran clathrate hydrate and its inhibition: Method and first results. Journal of Crystal Growth 179: 258-262.
Moon, C., Hawtin, R.W., Rodger, P. M., 2007. Nucleation and control of hydrates: insights from simulation. Faraday Discussions 136: 367-382
Sloan, E. D. Jr., Subramanian, S., Matthews, P. N., Lederhos, J. P., Khokhar, A. A., 1998. Quantifying hydrate formation and kinetic inhibition. Industrial and Engineering Chemical Research 37: 3124-3132
Sloan, E. D. Jr., 1998. Clathrate hydrates of natural gases, 2nd Ed. Marcel Dekker, N.Y.
Rodahl, M., Hook, F., Krozer, A., Brzezinski, P., Kasemo, B., 1995. Quartz crystal microbalance setup for frequency and Q-factor measurements in gaseous and liquid environments. Review of Scientific Instruments 66: 3924-3930.
Rodahl, M., Hook, F., Fredriksson, C., Keller, C., Krozer, A., Brzezinski, P., Voinova, M., Kasemo, B., 1997. Simultaneous frequency and dissipation factor QCM measurements of biomolecular adsorption and cell adhesion. Faraday Discuss. 107: 229-246.
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J. A., 2006a. Effect of antifreeze proteins on the formation and reformation of tetrahydrofuran clathrate hydrate. Journal of the American Chemical Society 128: 2844-2850
Zeng, H., Moudrakovski, I., Walker, V.K., Ripmeester, J. A., 2006b. Inhibition activity of an antifreeze protein on hydrocarbon hydrate formation. AIChE J. 52: 3304-3309.
Zeng, H., Walker, V. K., Rimpeester, J. A., 2007. Approaches to the design of better low-dosage gas hydrate inhibitors. Angewandte Chemie 119: 5498-5500
28
0 500 1000 1500
0.0
0.2
0.4
0.6
0.8
1.0N
t / N0
time / min.
0
100
200
300
400
500
600
700
A0 2 4 6 8 10 12
m (n
g/cm
2 )
conc.(mg/ml)
a B b
Figure 2.1. (a) Effects of PVCap and PVP on the induction time of THF hydrate and the memory effect. THF-hydrate was formed in the absence (filled star) or in the presence of 10 mg/ml PVCap (filled square) or PVP (filled triangle), or formed and subsequently melted and recrystallized in the absence (open star) or in the presence of 10 mg/ml PVCap (open square) or PVP (open triangle). (b) The adsorption masses of PVCap (squares) and PVP (triangles) on silica surface. Standard deviation error bars are indicated.
29
0 100 200 300
0
2
4
6
8
10
12
D (1
0-6 )
m (ng/cm2)
A
R1R2
0 200 400 600
0
2
4
6
8
10
12
B
R1
R2
0 2 4 6 8 10
a b
0
2
4
6
8
10
12
C c
R2 (1
0-9 n
g cm
2 )
conc. (mg/ml)
Figure 2.2. The relationships between dissipation factor (D) and adsorption mass (m) of PVCap (open square, a) and PVP (open triangle, b), as well as the relationships between their final R values, R2, and the concentrations of the same two inhibitors (c).
30
0 2 4 6 8 10 120
20
40
60
80
100
mf /
m (%
)
conc. (mg/ml)
B
0 10 20 30 40 50 60
0
50
100
150 r1
r2
mfm (n
g/cm
2 )
time (min.)
m
adsr3
A a
b
Figure 2.3. (a) A diagram of the effect of rinses on the adsorption mass, m, as determined by QCM-D and (b) the relationship between percentage of adsorption mass remaining (mf/m) and concentration of PVCap and PVP. mf: adsorption mass after three rinses; m: total adsorption mass before rinsing. Increase of m at point ads is due to introduction of PVP/PVCap into the measurement chamber. Decrease of m at point r1, r2 and r3 is due to the rinsing of the crystal.
31
Chapter 3
THE SEARCH FOR “GREEN INHIBITORS:” PERTURBING
HYDRATE GROWTH WITH BUGS
Emily I. Huva1, Raimond V. Gordienko1, John A. Ripmeester1,3, Huang Zeng1,3, Virginia K.
Walker1,2*
1Department of Biology, 2Department of Microbiology & Immunology, Queen’s University,
Kingston, Ontario, K7L 3N6, Canada; 3The Steacie Institute for Molecular Sciences,
National Research Council of Canada, 100 Sussex Drive, Rm 111, Ottawa, Ontario, K1A
OR6, Canada
In: Proceedings of the 6th International Conference on Gas Hydrates, Vancouver, 2008.
3.1. ABSTRACT
Certain organisms, including some bugs (both insects and microbes) are able to survive low
temperatures by the production of either ice nucleating proteins (INPs) or antifreeze
proteins (AFPs). INPs direct crystal growth by inducing rapid ice formation whereas AFPs
adsorb to ice embryos and decrease the temperature at which the ice grows. We have also
shown that certain AFPs can inhibit the crystallisation of clathrate hydrates and eliminate
more rapid recrystallisation or “memory effect”. Here we examine several bacterial species
32
with ice-associating properties for their effect on tetrahydrofuran (THF) hydrate
crystallisation. The bacteria Chryseobacterium sp. C14, which shares the ice recrystallisation
inhibition ability of AFPs, increased induction time to THF hydrate crystallisation in
isothermal experiments. In an effort to understand the association between AFPs and THF
hydrate we have produced bacterially-expressed AFPs as probes for hydrate binding.
Although the structure of hydrates is clearly distinct from ice, the apparent potential for
these products to perturb clathrate hydrate growth compels us to explore new techniques to
uncover “green inhibitors” for hydrate binding.
3.2. INTRODUCTION
Gas hydrates are formed when gas molecules are encaged by water molecules under
conditions of modest pressures and low (but not necessarily subzero) temperatures.
Although deposits of natural gas hydrates are regarded as a very promising potential energy
resource (Collett, 2002), unscheduled hydrate formation during hydrocarbon recovery and
transport can be costly, dangerous, and harmful to the environment (Cranswick, 2001; Zeng
et al., 2006).
Alternatives to thermodynamic inhibitors like methanol, which are required in large
amounts to be effective, are the low-dosage hydrate inhibitors, either anti-agglomerants or
kinetic inhibitors. Recently, a third group of low-dosage hydrate inhibitors have been
reported. Antifreeze proteins (AFPs), which confer cold tolerance to a variety of organisms
by adsorbing to ice crystals and inhibiting their growth, can also slow the growth of hydrate
crystals (Zeng et al., 2006; Zeng et al., 2003). These studies additionally present evidence that
some AFPs can eliminate the faster recrystallisation of hydrate after a brief melt, the so
called “memory effect.” Experiments designed to explore this phenomenon are presented
33
elsewhere in this volume (Appendix A).
To date, only AFPs from animals have been reported with hydrate inhibition activity.
However, they are costly to harvest from natural sources and it is difficult to obtain high
yields of active protein from recombinant Escherichia coli. Our hypothesis was that bacterial
AFPs would be more practical candidates as low dosage hydrate inhibitors for large-scale
production. We further believe that industry would be open to this novel approach to
hydrate inhibition because of the precedent set by the uses of bacterial proteins associated
with hydrate storage and transport. Commercial biosurfactants from Bacillus subtilis,
Pseudomonas aeruginosa, P. syringae, and Acinetobacter calcoaceticus have been used to accelerate the
formation of CO2 and natural gas hydrates (Morgan et al., 1999; Rogers et al., 2003; Rogers et
al., 2007).
In this paper we present the results of initial tests concerning the THF hydrate
affinity of several ice-associating bacteria. Chryseobacterium sp. C14 was initially isolated in our
lab as a highly freeze-thaw resistant microbe and was shown to exhibit ice recrystallisation
inhibition (Walker et al., 2006). Pseudomonas putida has AFP activity and ice nucleating activity
(Muryoi et al., 2004; B.R. Glick, personal communication; our unpublished observations). P.
borealis was isolated in our lab by ice affinity and shows ice nucleation activity (Wilson et al.,
2006) as do strains of P. syringae. We also examined the potential of ice affinity selection,
developed for microbial isolation (Wilson et al., 2006), to be adapted for hydrate affinity
selection.
34
3.3. METHODS
3.3.1. Hydrate formation and reformation in the presence of ice-associating bacteria
Bacteria were cultured in 10% Bacto™ Tryptic Soy Broth (Becton, Dickinson and Co.,
Franklin Lakes, NJ), inoculated with a single colony and grown to stationary phase (~109
CFU/mL) over 48 h at room temperature. Cultures (40- 100 mL) were then transferred to
3oC for 48-62 h. These cultures were mixed with OmniSolv® tetrahydrofuran (>99.9%
purity, unstabilized; EMD Chemicals Inc., Gibbstown, NJ) at a 3.34:1 culture:THF volume
ratio (15:1 molar ratio). Aliquots (3 mL) were dispensed into a series of 16 mm x 125 mm
Pyrex® culture tubes. Samples were immediately immersed in a 0.0 ± 0.2oC bath and stirred
individually at 300 rpm while temperature was monitored using thermocouples, as previously
described (Zeng et al., 2003). The temperature was chosen so that ice could not form (Gough
et al., 1971). For each solution tested, 3-4 independent experiments (a total of 33-56 vials)
were conducted. All vials, thermocouples and stir bars were well washed between
experiments: first with Sparkleen™ 1 (Fisher Scientific Co., Pittsburg, PA), then with tap
water, 75% acetone, and finally Milli-Q®-filtered H2O.
The temperature-time output of the thermocouples was used to measure the
induction time, t, between the time of sample equilibrium at 0oC and the onset of
crystallisation, indicated by a sudden increase in temperature. Note that the induction time to
crystallisation is distinct from the “time to nucleation,” as nucleation events did not
necessarily result in bulk crystallisation recorded by the thermocouples (see Results).
To examine THF hydrate recrystallisation, the same procedure was used, except that
freshly prepared samples were first frozen on dry ice (10 min), left at room temperature for 5
min and then thawed for 1 hr in a 6.4oC circulating bath (Model 9712; Polyscience, Niles,
35
IL).
Data from all experiments were pooled to examine trends graphically, while means
and standard deviations were calculated across experiments.
3.3.2. The development of a clathrate hydrate affinity purification (CHAP) technique
Initially cyclopentane and THF were each mixed with Milli-Q® water at volume ratios of
1:3.34 and 1:6.68. Sufficient bromophenol blue (Fisher Scientific) was added so as to be
visible. Samples were held at 3oC for days, both stirred and unstirred, after being nucleated
by a cooled copper wire. The solid and liquid fractions were then separated and the solid
fractions were rinsed with cold water before being visually compared.
Next, THF hydrate affinity purification was explored using several different
recombinant proteins marked with jelly fish green fluorescent protein (GFP). E. coli strains
were transformed with pET24a(+) plasmids encoding a His-tagged Type III AFP from the
fish Macrozoarces americanus, and a His-tagged AFP-GFP from the plant Lolium perenne.
Controls were produced from E. coli transformed with a pET20b(+) plasmid bearing a His-
tagged GFP.
After induction of the recombinant bacteria and collection of supernatants from cell
lysates, the proteins were purified (Sulkowski, et al., 1985) by immobilized metal affinity
chromatography with a cobalt-based resin (Clontech, Mountain View, CA). Purified proteins
were dialyzed overnight against a 0.1 M Tris HCl buffer (pH = 8) at 4°C. Protein
concentration was determined using a dye-binding assay (Smith at al., 1985).
THF-hydrate crystals were grown using the ice finger apparatus (Kuiper et al., 2003)
with the following modifications. The crystal was seeded at a bath temperature of 2.3°C and
36
later lowered into a pre-chilled 100 mL beaker containing one of the various proteins at 2
µM. The bath temperature was dropped approximately 1°C/h until the crystal diameter
approached half the original volume of the beaker. The hydrate crystal was then removed
from the copper finger and washed with 30 mL distilled water (<4°C) and observed under
UV light of wavelength 302 nm (Chromato-Vue transilluminator TM-36, UVP Incorp., San
Gabriel, CA)
3.4. RESULTS & DISCUSSION
3.4.1. THF hydrate formation and reformation
The presence of microbial cells in THF induction experiments varied depending upon the
type of bacteria examined. P. borealis and P. syringae appear to act as hydrate nucleators;
supercooling of THF solutions in the presence of these cells was reduced by ~ 2ºC
compared to controls (Appendix E). However, the addition of P. syringae cultures to THF did
not result in a significantly higher fraction of crystallized samples in our experiments (Figure
3.1a; Table 3.1). Similarly, E. coli cells did not decrease THF hydrate induction time
compared to H2O-THF controls (Figure 3.1a; Table 1). Crystallisation of both P. putida and
P. borealis solutions was reduced by up to 10% initially, but after 5 h was indistinguishable
from E. coli. It should be noted that THF is highly toxic to bacteria. Cells treated with THF
at the concentrations used here or even at 70% of this concentration (1:5.2 vol/vol) were
unculturable, forming no CFU (Huva, 2006). Therefore, it is possible that any putative
hydrate-inhibiting molecules associated with the cells could have been denatured or
disaggregated by the THF and therefore have been ineffective in this assay. Despite this
challenge, it is all the more remarkable that samples with the ice-associating bacteria
37
Chryseobacterium C14 took consistently longer to crystallize (Figure 3.1A). At the conclusion
of the experiment only 33% of the Chryseobacterium-containing samples had crystallized,
compared to 58% of the E. coli controls (Table 3.1).
Curiously, some Chryseobacterium and P. putida samples formed hydrate slurries instead
of the solid hydrate blocks normally seen after crystallisation. As many as 20% of these vials
contained suspensions of small macroscopic crystals, identified as millimeter-wide octahedra
in later morphology experiments (not shown). Such freezing was not seen on the
thermocouple readout, indicating that the multiple nucleation and growth events in these
vials were spread out over time; slurries were only recorded at the conclusion of the
experiments (t = 23 h).
Thus, the presence of Chryseobacterium sp. C14 appears to inhibit THF hydrate, at
least partially, by keeping crystals small for a certain time, superficially like the anti-
agglomerant low dosage hydrate inhibitors. Although THF is toxic to the cells, it is possible
then that at least some of the bacteria’s ice-associating molecules survived and were
responsible for mediating this effect. The proportion of Chryseobacterium vials containing
slurries was almost equivalent to that bacteria’s nc reduction compared to controls. This
suggests that the lower fraction of crystallized vials was due to growth inhibition and not to
nucleation inhibition. Nucleation inhibition did not occur, and indeed for P. putida, inclusion
of an estimated slurry fraction of 0.18 would raise nc to well above that of the control by 23
h. It may not be a coincidence that this increased THF hydrate nucleation combined with
the observed inhibited growth (Figure 1) mirrors P. putida’s known dual ice nucleating and
antifreeze activities.
All cultures recrystallized more rapidly after melting for 1 h at 6.4ºC, demonstrating
38
that the presence of cells did not interfere with memory effect. Despite some variability at 5
h, after 23 h nc was between 0.2 and 0.4 higher for every sample tested (Table 3.1, Figure
3.1b). This overall increase in proportion of samples frozen is consistent with previous
experiments on memory effect that show characteristic increases in freezing probability
(Takeya et al., 2000; Ohmura et al., 2003). Significantly, no slurries were seen with P. putida or
Chryseobacterium, or in any of the other samples, suggesting that for these microbes the
formation of nuclei inevitably resulted in solid freezing by the end of the experiment.
Vials containing Chryseobacterium again showed the lowest fraction frozen, ~0.12
below E. coli, over most of the experiment (Figure 3.1b). This was significantly slower than
the THF-water controls and for P. putida (Table 3.1). Certain AFPs eliminate the memory
effect (Zeng et al., 2006; Appendix A), but in this case since recrystallisation in the presence
of Chryseobacterium was still faster than initial crystallisation, there was no effect on memory.
3.4.2. Developing clathrate hydrate affinity purification
When ice or hydrates grow they exclude solutes, hence our initial experiments to develop
CHAP used a dye so that the exclusion could be easily visualized. Unstirred cyclopentane
solutions formed globular masses of hydrate that incorporated the bromophenol blue dye
(Figure 3.2a). Solutions that were stirred formed slurries of small crystals that made the
assessment of dye incorporation difficult. However, unstirred experiments using THF were
more promising: THF hydrate excluded bromophenol blue (Figure 3.2b).
This successful demonstration with polycrystalline hydrate encouraged us to examine
hydrate affinity for AFP. For ease of purification, the proteins were marked with a poly(His)
tag, and in order to visualize them in the hydrate a GFP tag was also incorporated into the
design of the plasmid constructs. When GFP alone was added to the THF-water solution,
39
there was only minimal incorporation into the polycrystalline solid that was formed (Figure
3). In contrast, both AFP-GFP constructs, one with a fish AFP and the other with a plant
AFP, appeared to be uniformly incorporated into the THF hydrate. These results show
much promise for the use of CHAP, or a “hydrate finger,” for the isolation of novel hydrate-
associating molecules and proteins.
3.5. CONCLUSION
THF hydrate formation was inhibited in the presence of cultures containing the ice-
associating bacterium Chryseobacterium C14. There was a 40% reduction in crystallisation,
showing potential as a commercial “green inhibitor” for hydrates. There was no elimination
of memory effect, however, and little compelling evidence was seen for nucleation inhibition
by any of the bacteria. Cyclopentane hydrates proved to be impractical for a new hydrate-
affinity purification method to isolate molecules interacting with hydrate, while THF’s
toxicity to bacteria may be limiting. Nonetheless, the use of such a hydrate-affinity based
technique to screen for potential “green” hydrate inhibitors is promising.
3.6. ACKNOWLEDGEMENTS
Dr. P.L. Davies, C. Garnham, A. Middleton and S. Gauthier are acknowledged for the
pET24a constructs and S. Wu is thanked for her help in making the GFP control. Dr. G.R.
Palmer developed the apparatus for the supercooling experiments. Partial support for
E.I. Huva was provided through the National Research Council’s guest worker program and
by a Queen’s University graduate award. NSERC (Canada) is also acknowledged for financial
support.
40
3.7. LITERATURE CITED
Collett, T.S.. 2002. Energy Resource Potential of Natural Gas Hydrates. American Association of Petroleum Geologists Bulletin 86: 1971-1992
Cranswick, D.. 2001. Brief overview of Gulf of Mexico OCS Oil and Gas Pipelines: Installation, Potential Impacts, and Mitigation Measures. In: OCS Report MMS 2001-067. U.S. Department of the Interior Minerals Management Service
Gough, S.R., Davidson, D.W.. 1971. Composition of Tetrahydrofuran Hydrate and the Effect of Pressure on the Decomposition. Canadian Journal of Chemistry 49: 2691-2699
Huva, E.I.. 2006. Microbial Nucleation and Inhibition of Tetrahydrofuran Hydrate. Undergraduate Thesis. Department of Biology, Queen’s University, Kingston, Ontario.
Kuiper, M.J., Lankin, C., Gauthier, S.Y., Walker, V.K., Davies, P.L.. 2003. Purification of antifreeze proteins by adsorption to ice. Biochemical and Biophysical Research Communications. 300: 645-648.
Morgan, J.J., Blackwell, V.R., Johnson, D.E., Spencer, D.F., North, W.J.. 1999. Hydrate formation from gaseous CO2 and water. Environmental Science Technology 33: 1448-1452
Muryoi, N., Sato, M., Kaneko, S., Kawahara, H., Obata, H., Yaish, M.W.F., Griffith, M., Glick, B.R.. 2004.Cloning and expression of afpA, a gene encoding an antifreeze protein from the arctic plant growth-promoting rhizobacterium Pseudomonas putida GR12-2. Journal of Bacteriology 186(17): 5661-5671
Ohmura, R., Ogawa, M., Yasuoka, K., Mori, Y.H.. 2003 Statistical study of clathrate-hydrate nucleation in a water/hydrochlorofluorocarbon system: Search for the nature of the "memory effect." Journal of Physical Chemistry B 107(22): 5289-5293
Rogers, R., Zhang, G., Dearman, J., Woods, C.. 2007. Investigations into surfactant/gas hydrate relationship. Journal of Petroleum Science and Engineering 56: 82-88
Rogers, R.E., Kothapalli, C., Lee, M.S., Woolsey, R.J.. 2003. Catalysis of gas hydrates by biosurfactants in seawater-saturated sand/clay. Canadian Journal of Chemical Engineering 81: 1-8
Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano, M.D., Fujimoto, E.K., Goeke, N.M., Olson, B.J., Klenk, D.C.. 1985. Measurement of protein using bicinchoninic acid. Analytical Biochemistry 150(1): 76-85.
Sulkowski, E.. 1985. Purification of proteins by IMAC. Trends in biotechnology 3: 1-7
Takeya, S., Hori, A., Hondoh, T., Uchida, T.. 2000. Freezing-Memory Effect of Water on Nucleation of CO2 Hydrate Crystals. Journal of Physical Chemistry B 104(17) 4164-4168
Walker, V.K., Palmer, G.R., Voordouw, G.. 2006. Freeze-Thaw Tolerance and Clues to the Winter Survival of a Soil Community. Applied and Environmental Microbiology 72: 1784-1792
Walker, V.K., Zeng, H., Gordienko, R.V., Kuiper, M., Huva, E.I., Ripmeester, J.A.. The
41
mysteries of memory effect and its elimination with antifreeze proteins. In: Proceedings of the 6th International Conference on Gas Hydrates, Vancouver, 2008.
Wilson, S.L., Kelley, D.L., Walker, V.K.. 2006. Ice-active characteristics of soil bacteria selected by ice-affinity. Environmental Microbiology 8(10): 1816–1824
Zeng, H., Moudrakovski, I.L., Ripmeester, J.A.. 2006. Effect of Antifreeze Protein on Nucleation, Growth, and Memory of Gas Hydrates. American Institute of Chemical Engineers 52(9): 3304-3309
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J.A.. 2003. The inhibition of tetrahydrofuran clathrate-hydrate formation with antifreeze protein. Canadian Journal of Physics 2003;81: 17-24
42
Table 3.1. Comparison of isothermal THF hydrate formation at 0oC for various bacterial cultures.
Known
activity1
Mean % Samples Crystallized
at: No. of:
Culture in solution with THF
AFP INP t = 5 h t = 23 h Exps Vials
None (H2O-THF) 36 ± 13 56 ± 13 3 36
E. coli 46 ± 22 58 ± 14 4 44
Chryseobacterium ● 19 ± 13 33 ± 14 3 36
P. putida ● ● 36 ± 7 64 ± 13 4 44
P. borealis ● ● 40 ± 22 51 ± 15 3 36
P. syringae ● 54 ± 25 64 ± 16 3 37
Recrystallisation2 Experiments
None (H2O-THF) 82 A ± 9 88 ± 5 3 33
E. coli 67AB ± 8 86 ± 5 3 36
Chryseobacterium 53 B ± 10 75 ± 22 3 36
P. putida 82 A ± 9 97 ± 5 3 33
1 observed antifreeze (AF) or ice nucleating (IN) activity in ice, published and unpublished 2 solutions were previously frozen and then melted (1 h at 6.4oC) A,B groupings of samples that do not differ significantly (Tukey-Kramer HSD; q*=3.28 fresh
samples, 3.20 pre-frozen; α=0.05).
43
0
0.5
1
0 5 10 15 20 25t (hours)
n C
a
0
0.5
1
0 5 10 15 20 25t (hours)
n C
b
a b
Figure 3.1. Crystallized fraction, nc , of (a) freshly made and (b) pre-frozen samples that had formed THF hydrate after time t at 0oC. Solutions consisted of THF in a 1:3.34 (vol) ratio with H2O (×) or 109 CFU cultures of: E. coli (□), Chryseobacterium (■), P. putida (●), P. borealis (▲), and P. syringae (∆).Pre-frozen samples had been melted for 1h at 6.4oC prior to the experiment. Data points at t = 0 are not displayed.
44
a b
Figure 3.2. Hydrate (left) and aqueous (right) fractions of dilute (a) cyclopentane and (b) THF solutions containing bromophenol blue, after being nucleated and left unstirred at 3oC for several days.
45
a b c
3 cm
Figure 3.3. Typical images of polycrystalline THF hydrate formed in the presence of (a) fish and (b) plant antifreeze proteins tagged with green-florescent protein (GFP), as visualized under UV light. Image (c) shows the result of an experiment containing a GFP not linked to AFP. The experiment was repeated three times.
46
Chapter 4
TETRAHYDROFURAN HYDRATE CRYSTALLISATION AND
MEMORY WITH BIOLOGICAL AND SYNTHETIC
INHIBITIOR-PROMOTER COMBINATIONS
Emily I. Huva1, Huang Zeng1,3, John A. Ripmeester1,3, Virginia K. Walker1,2
1Department of Biology, 2Department of Microbiology & Immunology, Queen’s University,
Kingston, Ontario, K7L 3N6, Canada; 3The Steacie Institute for Molecular Sciences,
National Research Council of Canada, 100 Sussex Drive, Rm 111, Ottawa, Ontario, K1A
OR6, Canada
In preparation for submission
4.1. ABSTRACT
Antifreeze proteins (AFPs), which inhibit ice formation, have recently been reported to
inhibit clathrate hydrate as well, suggesting that they may be useful as kinetic inhibitors. A
combination of AFPs and ice-nucleating proteins (INPs) is used by some cold-adapted
organisms to prevent the formation of large, potentially damaging ice crystals through the
generation of multiple small ice crystals. The use of inhibitors and promoter/inhibitor
combinations to encourage anti-agglomeration of a model hydrate, tetrahydrofuran (THF)
hydrate, was explored through a comparison of induction-times in the presence of bacteria
47
with dual antifreeze/ice nucleating properties, as well as with commercial polymers.
Pseudomonas putida and Pseudomonas borealis, which have INPs, prevented the formation of
solid hydrate only temporarily. Both commercial polymers increased induction times as
previously reported, but the addition of a surfactant, sodium dodecyl sulphate (SDS),
enhanced the activity of polyvinylcaprolactam (PVCap) modestly. Reformation of THF
hydrate after a brief melt was faster irrespective of the additive. These studies further
demonstrate the value of ice association as a predictor of hydrate inhibition.
4.2. INTRODUCTION
Clathrate-hydrates are problematic during oil and natural gas extraction and transport (Sloan,
2005). Thus the development of low-dosage kinetic inhibitors is of interest to industry in
order to address economic and environmental concerns. Two kinetic inhibitors,
polyvinylpyrrolidone (PVP) and polyvinylcaprolactam (PVCap), are viewed as promising
alternatives to current thermodynamic inhibitors, principally the toxic alcohols methanol and
ethylene glycol (Kelland, 2006).
Ironically, although gas hydrate formation is a re-occurring problem in the field,
clathrate hydrates often take a long time to form in a laboratory setting even given
appropriate temperature and pressure conditions, prompting the use of hydrate
“promoters”. Similarly, water also typically supercools, where small volumes in the
laboratory may not freeze until temperatures approache -40oC. In nature, freeze promotion
of water can be accomplished by ice-nucleating proteins (INPs) from bacteria, including
Pseudomonas syringae. With hydrates, the anionic surfactant, sodium dodecyl sulphate (SDS),
has been used to promote crystallisation (Zhong and Rogers, 2000; Lin et al., 2004; Ribeiro
and Lage, 2008), and biosurfactants like P. syringae INP have also been used (Rogers et al.,
48
2003).
Once hydrates form but then are melted for a short time by increasing temperatures
or reducing pressures, they subsequently reform at a much faster rate, a phenomenon termed
the “memory effect” reformation (Takeya et al., 2000; Ohmura et al., 2003). Recently,
antifreeze proteins (AFPs) from fish and insect have been observed to impede both the
formation and reformation of structure I and structure II hydrates (Zeng et al., 2006a,
2006b). These proteins are found in certain cold-resistant organisms and work to lower the
temperature at which ice grows. Other organisms have both INPs and AFPs presumably to
encourage the formation of multiple small nuclei, in effect avoiding large crystals and
concentrating cell solutes to depress further freezing (Storey and Storey, 1988). Applying this
concept to hydrate inhibition, the presence of both inhibitor molecules and promoting
agents together may generate an anti-agglomeration effect. We have examined this possibility
using the model structure II hydrate, tetrahydrofuran (THF) hydrate. P. borealis has INP and
ice-shaping activity (Wilson et al., 2006) and P. putida is reported to have INP and AFP
activities (Muryoi et al., 2004; B. Glick, personal communication). The synthetic inhibitors
PVP and PVCap are compared to the combination of SDS and PVCap.
4.3. MATERIALS AND METHODS
4.3.1. Sample Preparation
Solutions were composed of >99.9% purity tetrahydrofuran (unstabilized, OmniSolv®,
EMD Chemicals Inc., Gibbstown, NJ) in a 1:3.34 ratio (vol; 1:15 mol. ratio) with undiluted
bacterial culture as previously described (Chapter 3) or mixtures of milliQ H2O and synthetic
polymers. Single cell isolates of Pseudomonas putida, Pseudomonas borealis, and E.coli. TG2 were
49
cultured in 10% tryptic soy broth until they reached stationary phase at ~108 colony forming
units per mL as previously described (Walker et al., 2006). They were then stored at 3oC for
2-3 days. Aqueous solutions of commercial inhibitors were made from MilliQ water and
either PVP (10 mg/mL ), PVCap (10 mg/mL), or PVCap (10 mg/mL) and SDS (5 mg/mL).
These were compared with pure THF/water solutions to give a baseline for synthetic
inhibition. PVP in powder form (MW ~40,000) was kindly provided by Dr. E. D. Sloan
(Colorado School of Mines), and powdered PVCap (MW ~110,000) was generously donated
by Dr. L. Talley (Exxon Mobil).
Samples (3 mL) were crystallized isothermally as previously described (Zeng et al.,
2003). Induction time (t) was used to compare the effect of the various additives on the
period between thermal equilibrium (0.0 ± 0.2oC) and the onset of crystallization, indicated
by a rapid temperature increase.
4.3.2. Data Analysis
Data were fitted to a power-law relationship for analysis. Such models are useful descriptions
of time-dependent nucleation phenomena (Allnatt and Jacobs, 1968; Ding and Spruiell,
1998; Heneghan et al., 2001), even though inherent complexities prevent a detailed kinetic
analysis (e.g. Kashchieva and Firoozabadib, 2003; Lehtinena et al., 2007; Herhold et al., 1999).
Induction time was taken as a macroscopic physical property and hence only secondary
observations of kinetic phenomena were assessed so that nucleation was bundled with early
growth events. Nonetheless, a quantitative measure of the apparent crystallisation rate, k, is
useful. To obtain this, we have modeled the time taken for a discrete fraction of samples, nC,
to reached the onset of sudden, total crystallisation as:
50
( ) mC tktn +⋅= 1 (1)
where m represents the crystallisation index, which dictates the time-dependence of the
instantaneous crystallisation rate, dtdnC ; m=-1 indicates zero time-dependence; higher
values in the range -1<m<0 denote greater time-dependence; 0≤m<∞ gives increasing time-
dependency (Ding and Spruiell, 1998). This is seen most clearly from the instantaneous rate
of crystallisation:
mC tmkdt
dn)1( +⋅= (2)
The “average lag-time” to freezing will be defined as τ, as the time at which half the
samples will have crystallised:
( ) 5.0=τCn (3)
Hence,
( ) 11
2 +−= mkτ (4)
Model parameters were calculated in Microsoft Excel via log-log regression, i.e.:
( ) tmktnC log)1(loglog ++= (5)
Samples which crystallize at the equilibrium temperature (t=0) could not be included in these
calculations, although they were included in calculations of fraction frozen, nC. Since
crystallization was prompt (t=0) in many experiments, the data were pooled for model
fitting. Calculations of average fractions crystallized were done across separate experiments
(n=3-5; P. borealis reformation, n=1).
4.4. RESULTS
The dual INP/AFP bacteria, the addition of P. borealis or P. putida reduced crystallization
51
only during the first few hours (Figure 4.1a), showing rate constants reduced overall by 33 ±
17 % for P. borealis and negligibly for P. putida, compared to the control, E. coli.
Corresponding lag-times (Eq. 4) increased from 5±5 h (control) to 15 ± 17 h and 9 ± 8 h,
respectively. On the other hand, the ice-nucleating bacterium, P. syringae, was
indistinguishable from the control (Figure 4.1a; Appendix D). The contrast between strains
with INP, INP/AFP, and AFP is clear in a differential plot (Figure 4.3) which uses E. coli nC
data generated from power-law model parameters to “subtract” control behaviour,
effectively isolating hydrate activity. The same experiments conducted with the bacteria
Chryseobacterium (Chapter 3) exemplify AFP-only activity; Chryseobacterium samples reach an
equilibrium fraction frozen well below that of the control. While a similar effect occurred for
P. borealis samples, it was not significant (Chapter 3). Strikingly, up to 20% of P. putida
samples were identified as loose hydrate slurries at the conclusion of incubation yet had not
released sufficient heat of fusion to register as crystallized. This was not seen for solutions
containing synthetic additives or during reformation (memory effect) experiments.
Of the synthetic inhibitors, PVCap and PVCap/SDS samples had the lowest
crystallisation rates, 0.19 ± 0.01 and 0.15 ± 0.01, respectively, 67 ± 4 % and 74 ± 5 % lower
than the control (Figure 4.1b). Lag-times, at 11 ± 2 h and 17 ± 3 h, were similar to those of
the INP/AFP bacteria. PVP also reduced k, by 23 ± 2 %, corresponding to a lag-time of 2.2
± 0.4 hours. Although the instantaneous crystallisation rate, dtdnC , rapidly decayed (m≈-
0.84) for most samples, freezing was more sustained in PVCap (-0.59 ± 0.01) and
PVCap/SDS (-0.57 ± 0.02) samples.
Reformation was generally more rapid. By the end of the experiment, nC was
consistently 0.3-0.4 higher than for freshly prepared samples (Figure 4.1). During
52
reformation, samples containing both INP and AFP reformed substantially faster, having
formation rates 75 ± 26 % faster than before. In contrast, the rate for control solutions did
not significantly increase. P. syringae reformation was not tested. For mixtures containing
commercial polymers, k became very high, and lag-times dropped to effectively zero. The
sample most different from the control was PVCap/SDS, whose rate constant was
decreased by 13 ± 2 %. Reformation events decayed more quickly with time than previously:
m was approximately -0.85 for bacterial samples, and -0.98 for non-biological samples.
It should be noted that, overall, the power-law model fit both crystallisation and
recrystallisation data well (Figure 4.1 and 4.2), with each model having an R2 of between 0.81
and 0.99 (median: 0.96; see Appendix D).
4.5. DISCUSSION
Crystallisation of fresh samples of INP/AFP bacteria demonstrated inhibition activity that
was limited to the first few hours (Figure 4.3). However, the formation of loose slurries in P.
putida samples in addition to those frozen solid indicates that P. putida samples underwent
additional nucleation and subsequently growth inhibition, in order to preserve the slurry
state until the end of the experiment. This is the “agglomeration” effect anticipated for
INP/AFP bacteria. Because of its ability to inhibit THF hydrate growth, P. putida may be
useful as a hydrate inhibitor. Neither INP/AFP bacteria showed significant overall inhibition
compared to E. coli, possibly due to interference of AFPs by INPs. It may be that the surface
area of ice nuclei generated by INP reaches a point at which the binding sites of available
AFP are saturated, overall resulting in a delayed onset of crystallisation. Inhibition of INPs
by AFP, whether directly or not, is not a new idea (Parody-Morreale et al., 1988; Zátmecník
53
and Jánacek, 1992; Defalco, 2007). Such a saturation mechanism could be confirmed by
changing the relative concentrations of AFP and INP. The P. putida AFP is well-
characterized (Muryoi, 2004), so either this mechanism or hydrate inhibition assays could be
done with the purified protein.
Interestingly, samples with INPs reformed hydrate much quicker than the control in
memory experiments, despite the inhibition activity demonstrated earlier. This suggests that
the memory effect might be enhanced by INPs, consistent with current descriptions of the
memory effect as a heterogeneous nucleation phenomenon (Appendix A, Zeng et al., 2006a).
When more heterogeneous nucleators, such as INPs, are introduced into the system, there
are more nucleating points susceptible to the mechanism of memory crystallisation, whatever
that mechanism may be. As INPs are already conducive to crystal formation, perhaps no
imprinting occurs and instead, residual water structures surrounding the INPs are
responsible for the appearance of memory. This may explain the dependence of memory on
thermal history (Takeya et al., 2000; Ohmura et al., 2003) even though no residual water
structures have been found in melts of pure samples devoid of heterogeneous nucleators
(Buchanan et al., 2005).
Aside from the INP bacteria, each solution appeared to have the same additional
fraction freeze in memory experiments despite vast differences in composition. The
implication is that this particular sample size and thermal history result in a ~35% freezing
probability (Takeya et al., 2000; Ohmura et al., 2003), which was not influenced by biological
or synthetic additives. Hence, inhibition of memory effect was not seen, although other
researchers have noted memory inhibition by antifreeze proteins (Zeng et al., 2006).
The presence of SDS and PVCap together resulted in a formation rate, k, of -74 ± 5
54
%, even lower than that for PVCap alone (67 ± 4 %). During reformation, SDS/PVCap
samples still had a k marginally below the control (-13 ± 2 %), although the practical
implication of this difference is small as the lag-time was effectively zero. The mechanism
for this improvement was unlikely to be an anti-agglomerant-like effect, as no slurry-state
was observed. However, high concentrations of SDS (>1 mg/mL) have been found to
inhibit formation of methane hydrate without PVCap (Sun et al., 2007). In biochemistry,
SDS is used to break the non-covalent bonds of proteins, causing them to lose their tertiary
structure, in preparation for polyacrylamide gel electrophoresis (SDS-PAGE); in our case, it
may be that such a molecule could linearize a large polymer like PVCap, thereby increasing
the size of hydrate-binding sites and enhancing inhibition.
PVCap was previously noted to be a more efficient crystallisation inhibitor of THF
hydrate than PVP, but was poorer at reformation inhibition (Chapter 2). Analysis of this data
using the power-law model shows this first difference very clearly: the rate constant for PVP
samples is twice that of PVCap samples. However, upon reformation the difference between
PVP and PVCap was small (k differs by 7%) and the lag-time to 50% crystallisation was still
negligible. This suggests that the difference between PVP and PVCap during reformation
experiments is actually negligible. (No N23 data available; see Chapter 3).
The power law data fit resulted in a relatively high nucleation index, m (Ding and
Spruiell, 1998), for PVCap and PVCap/SDS samples. This is manifested in Figure 4.1b as a
straighter crystallisation curve upwards to the right, while curves for other polymer or even
bacterial samples exhibit a steeper cliff initially, followed by a plateau. In this latter case,
nucleation is quite time-dependent, meaning that more heterogeneous nucleators are initially
present. As this population of nucleators gets “used up,” the rate of crystallisation plateaus,
55
as seen for several samples. In the case where inhibitors are present, however, it is possible
that these are interfering with the natural population of impurities, such as dust, and so
samples appear to take their time in freezing.
The power law for time-dependent nucleation was found to fit very well to data,
despite the inherent complexity of the solutions being tested, and offers a valuable view of
the underlying time dependence of the system. As well, it is useful for error estimates and
extrapolation of data, especially since experiments with living organisms cannot be looped
multiple times for the sake of good statistics. Extrapolation of a “lag-time” (Figure 4.2)
allows a practical gauge of inhibition activity, while point-by-point comparison of data can
be made without the need for binning data (Figure 4.3). Preliminary fitting of the model to
supercooling data was also promising (Appendix D, Appendix E); it may merit further
examination.
In summary: dual INP/AFP activity temporarily delayed THF hydrate crystallisation.
Purification of the AFP from either P. borealis or P. putida would likely be necessary to
separate its effects from those of the INP, but the activity of the secreted P. putida protein is
promising as it caused hydrate slurry formation instead of solid freezing in some cases. The
addition of P. syringae did not appear to promote nucleation beyond that caused by the
presence of bacterial cells. The addition of the popular hydrate promoter, SDS, modestly
enhanced the activity of the inhibitor, PVCap. Given that PVCap is already a powerful
kinetic inhibitor, this small advantage may have large benefits.
56
0
0.5
1
0 5 10 15 20 25t (hours)
n C
a
0
0.5
1
0 5 10 15 20 25t (hours)
n C
b
Figure 4.1. Induction times to formation (solid symbols) and reformation (hollow symbols) of THF hydrate in the presence of (a) biological ice-inhibitors and (b) synthetic hydrate inhibitors. Samples consisted of 1:3.34 (vol) THF:solution, where solutions were (a) undiluted cultures of: E. coli (■/□), P. syringae (×), P. putida (●/o) and P. borealis (▲/∆); or (b) aqueous mixtures of: purified water (■/□), PVP (10 mg/mL; ♦/◊ ), PVCap (10 mg/mL; ●/o), and PVCap (10 mg/mL) + SDS (5 mg/mL; ▲/∆). Data are fitted with a power-law model (Eq. 1)
57
0.1
1
10
100
1000
E. coli
P. puti
da
P. bore
alis
P. syri
ngae
water
PVCap
PVCap/SDS
PVP
τ / h
*¤
Figure 4.2. Average lag-time, τ, to THF formation (grey) and reformation (white) in the presence of biological ice-inhibitors and synthetic hydrate inhibitors. Mixtures were 1:3.34 (vol) THF:culture or solution. Error bars were propagated in quadrature (Appendix D). ¤ Only one P. borealis experiment (11 samples) was performed, and all samples had induction times of zero. * P. syringae reformation experiments were not done.
58
-0.5
0
2 4 6 8 10 12 14 16 18 20 22 24
t (hours)
ni-n
cont
rol
control
promotion
inhibition
Figure 4.3. Differential activity of cultures during THF hydrate formation represented by the fraction of samples crystallized by time t (ni) minus that for an E. coli power-law model at time t. Solutions consisted of 1:3.34 (vol) THF:culture of: Chryseobacterium (■), P. putida (●), P. borealis (▲), and P. syringae (o).
59
4.6. LITERATURE CITED
Allnatt, A.R., Jacobs, P.W.M.. 1968. Theory of nucleation in solid state reactions. Canadian Journal of Chemistry 46: 111-116
Buchanan, P., Soper, A.K., Thompson, H., Westacott, R.E.. 2005. Search for memory effects in methane hydrate: Structure of water before hydrate formation and after hydrate decomposition. Journal of Chemical Physics 123: 164507
Defalco, T.. 2007. Undergraduate Thesis. Department of Biology, Queen’s University, Kingston, Canada.
Ding, Z., Spruiell, J.E.. 1998. Interpretation of the nonisothermal crystallization kinetics of polypropylene using a power law nucleation rate function. Journal of Polymer Science Part B: Polymer Physics 35(7): 1077-1093
Heneghan AF, Wilson PW, Wang G, Haymet ADJ. 2001. Liquid-to-crystal nucleation: Automated lag-time apparatus to study supercooled liquids. Journal of Chemical Physics. 2001; 115(16): 7599-7608
Herold, A.B., Ertas, D., Levine, A.J., King, H.E.Jr. 1999. Impurity mediated nucleation in hexadecane-in-water emulsions. Physical Review E 59(6): 6946-6955
Kashchieva, D., Firoozabadib, A.. 2003. Induction time in crystallization of gas hydrates. Journal of Crystal Growth 250: 499–515
Kelland, M.A.. 2006. History of the development of low dosage hydrate inhibitors. Energy and Fuels 20: 825-847
Lehtinena, K.E.J., Masob, M.D., Kulmalab, M., Kerminen, V.-M.. 2007. Estimating nucleation rates from apparent particle formation rates and vice versa: Revised formulation of the Kerminen–Kulmala equation. Journal of Aerosol Science 38(9): 988-994
Lin, W., Chen, G.J., Sun, C.Y., Guo, X.Q., Wu, Z.K., Liang, M.Y., Chen, L.T., Yang, L.Y.. 2004. Effect of surfactant on the formation and dissociation kinetic behavior of methane hydrate. Chemical Engineering Science 59: 4449.
Muryoi, N., Sato, M., Kaneko, S., Kawahara, H., Obata, H., Yaish, M.W.F., Griffith, M., Glick, B.R.. 2004.Cloning and expression of afpA, a gene encoding an antifreeze protein from the arctic plant growth-promoting rhizobacterium Pseudomonas putida GR12-2. Journal of Bacteriology 186(17): 5661-5671
Ohmura, R., Ogawa, M., Yasuoka, K., Mori, Y.H.. 2003 Statistical study of clathrate-hydrate nucleation in a water/hydrochlorofluorocarbon system: Search for the nature of the "memory effect." Journal of Physical Chemistry B 107(22): 5289-5293
Ohmura, R., Ogawa, M., Yasuoka, K., Mori, Y.H.. 2003 Statistical study of clathrate-hydrate nucleation in a water/hydrochlorofluorocarbon system: Search for the nature of the "memory effect." Journal of Physical Chemistry B 107(22): 5289-5293
Parody-Morreale, A., Murphy, K.P., Di Cera, E., Fall, R., DeVries, A.L. and Gill, S.J. 1988. Inhibition of bacterial ice nucleators by fish antifreeze glycoproteins. Nature 333: 782-783.
60
Ribeiro, C.P. Jr., Lage, P.L.C.. 2008. Modelling of hydrate formation kinetics: State-of-the-art and future directions, Chemical Engineering Science 63(8): 2007-2034.
Rogers, R.E., Kothapalli, C., Lee, M.S., Woolsey, R.J.. 2003. Catalysis of gas hydrates by biosurfactants in seawater-saturated sand/clay. Canadian Journal of Chemical Engineering 81: 1-8
Sloan, D.E.. 2005. A changing hydrate paradigm--from apprehension to avoidance to risk management. Fluid Phase Equilibria 228-229 (PPEPPD 2004 Proceedings): 67-74
Storey, K.B., Storey, J. M.. 1988. Freeze tolerance in animals. Physiological Reviews 68: 27-84
Sun, C.Y., Chen, G.J., Ma, C.F., Huang, Q., Luo, H., Li, Q.P.. 2007. The growth kinetics of hydrate film on the surface of gas bubble suspended in water or aqueous surfactant solution. Journal of Crystal Growth 306(2): 491-499.
Takeya, S., Hori, A., Hondoh, T., Uchida, T.. 2000. Freezing-Memory Effect of Water on Nucleation of CO2 Hydrate Crystals. Journal of Physical Chemistry B 104(17): 4164-4168
Takeya, S., Hori, A., Hondoh, T., Uchida, T.. 2000. Freezing-Memory Effect of Water on Nucleation of CO2 Hydrate Crystals. Journal of Physical Chemistry B 104(17) 4164-4168
Walker, V.K., Palmer, G.R., Voordouw, G.. 2006. Freeze-Thaw Tolerance and Clues to the Winter Survival of a Soil Community. Applied and Environmental Microbiology 72: 1784-1792
Wilson, S.L., Kelley, D.L., Walker, V.K.. 2006. Ice-active characteristics of soil bacteria selected by ice-affinity. Environmental Microbiology 8(10): 1816–1824
Zátmecník, J., Jánacek, J.. 1992. Interaction of antifreeze proteins from cold hardened cereal seedlings with ice nucleation active bacteria. Cryobiology 29:718-719.
Zeng, H., Moudrakovski, I., Walker, V.K., Ripmeester, J. A., 2006b. Inhibition activity of an antifreeze protein on hydrocarbon hydrate formation. AIChE J. 52: 3304-3309.
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J.A., 2006a. Effect of antifreeze proteins on the formation and reformation of tetrahydrofuran clathrate hydrate. Journal of the American Chemical Society 128: 2844-2850
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J.A.. 2003. The inhibition of tetrahydrofuran clathrate-hydrate formation with antifreeze protein. Canadian Journal of Physics 81: 17-24
Zhong, Y., Rogers, R.E.. 2000. Surfactant effects on gas hydrate formation. Chemical Engineering Science 55 (19): 4175-4187.
61
Chapter 5
GENERAL DISCUSSION
5.1. HYDRATE INHIBITION BY ANTIFREEZE BACTERIA &
PROTEINS
Work described in this thesis shows that THF hydrate formation can be inhibited in the
presence of certain bacteria with ice affinity characteristics. The addition of Chryseobacterium,
P. putida, and P. borealis to THF showed increased induction times to hydrate formation in
the first 5 h, but after 23 only Chryseobacterium samples had significantly longer induction
times. Hydrate slurries were observed in some Chryseobacterium and P. putida samples instead
of solid freezing (Chapter 3), which is suggestive of a growth-inhibition mechanism
analogous to that of freeze tolerant insects, mediated by AFPs and INPs. Further indirect
evidence of an interaction between bacteria, or AFP, and THF hydrate was seen when two
different GFP-tagged AFPs were selectively incorporated into a slowly grown multi-
crystalline mass (Chapter 3). In addition, single THF crystals appeared cloudy when grown in
the presence of bacteria with AFP activity (Appendix F). Changes to the shape of single
THF hydrate crystals in the presence of AFPs have been noted elsewhere (Zeng et al., 2003;
Zeng et al., 2006a), possibly arising from insufficient coverage of the crystal by AFPs,
however such morphological changes were not observed here. Moreover, because THF
hydrate is a cubic structure with eight identical faces, techniques used in ice-shaping assays
(Wilson et al., 2006), that might have been helpful to confirm or refute this observation, do
not apply here. Instead, visually monitoring crystal growth rate under controlled conditions
62
may provide a better assay of growth inhibition.
There was no direct evidence of nucleation inhibition by ice-associating bacteria
having antifreeze protein activity, but this has not been satisfactorily explored even for AFP
and water (Chapter 1.1.3; see Appendix A for AFP binding to a heterogeneous nucleator,
SiO2). Although the presence of hydrate slurries with Chryseobacterium and P. putida may
demonstrate growth inhibition, the slurries also indicate that nucleation had occurred. If the
slurry fraction (<20%) is added to the induction time curves, overall the same fraction of
samples containing Chryseobacterium underwent nucleation as in the controls. If analyzed in
this way, even more P. putida samples were nucleated, perhaps due to the presence of ice-
nucleating activity.
5.2. MEMORY EFFECT INHIBITION BY ANTIFREEZE BACTERIA &
PROTEINS
THF hydrate, formed in the presence of bacterial cell cultures, reformed at a slower rate
compared to samples containing commercial inhibitors (Chapter 3 and 4). However, none
appeared to inhibit the memory effect (Chapter 1.2.1). Others have demonstrated the
elimination of the memory effect by type I AFP and an insect AFP, by showing that
induction time curves remained the same whether fresh samples or previously frozen
samples were used (Zeng et al., 2006a). This was not the case here; in fact, “displacements”
of induction time curves were seen for all memory samples (Chapters 3 and 4; nC=+ 0.25 to
+0.4, or to nC=1). This makes the past findings of memory elimination by AFP (Zeng et al.,
2006a) even more exceptional. Moreover, pre-frozen solutions containing ice-nucleating
bacteria (i.e., P. borealis and P. putida) reformed at an even more accelerated rate than did the
control (Chapter 4), suggesting that the memory effect may arise from residual water
63
structures surrounding heterogeneous nucleators. In this case, there may even be a reduction
of the memory effect by stirring, although memory was seen in the experiments catalogued
in this thesis despite stirring of all samples placed at 0oC.
Since Chryseobacterium culture did not eliminate the memory effect, this may offer
additional, indirect evidence that the bacterium did not interfere with nucleation. However,
bacterial cells themselves seem to be able to initiate nucleation in some circumstances, so
this may have obscured any effect on memory inhibition even if it was there. A second
possibility is that Chryseobacterium does not optimally bind to the heterogeneous nucleators
associated with the memory effect. Due to the varied structures of AFPs (Chapter 1.1.3), it is
possible that they may show differences in their ability to interact with heterogeneous
nucleators. In addition, the QCM-D experiments demonstrate that there are inherent
differences among ice, hydrate, and memory hydrate inhibitors (Table 5.1). Since the spacing
of atoms in each type of crystal is different, it does not come as a surprise that some
molecules appear to adsorb more tightly than others (Appendix A).
5.3. IDENTIFICATION OF NOVEL HYDRATE-ASSOCIATING
MOLECULES AND BACTERIA
The success of clathrate-hydrate affinity purification (CHAP) of particles and proteins using
THF hydrate was demonstrated in Chapter 3. This technique has potential for high-
throughput searches for new hydrate-interacting molecules. However, THF is toxic to
bacteria (Author’s Undergraduate Thesis), precluding the isolation of culturable bacteria
using this system for CHAP. Nevertheless, the potential to identify novel hydrate-binding
bacteria is important, and so methods of identifying bacteria apart from culturing might be
explored, as well as a CHAP method using gas hydrates.
64
A proposed CHAP apparatus is displayed in Figure 5.1 (J.A. Ripmeester, personal
communication); partial formation of gas hydrate (e.g. propane hydrate) from a mixed
aqueous solution could be carried out in a quartz vessel, able to withstand moderate
pressurization. This would be followed by a separation of hydrate and liquid phases using a
fine filter. The hydrate particles could then be partially melted via a temperature change to
wash away any remaining solution adhering to the crystals. Hydrate particles will likely be
small (agitation will be required) however brief melts may be able to preserve sufficient
amounts of crystal and that repeated hydrate formation/melt/rinse cycles could allow
efficient separation of binding and non-binding particles. Another difficulty will be to
sufficiently dissolve the hydrate guest in the aqueous solution, as seen by the problems
associated with cyclopentane hydrate formation in preliminary experiments (Chapter 3).
5.4. HYDRATE INHIBITORS: FUTURE DIRECTIONS
Chryseobacterium samples formed THF hydrate at a rate constant 67 ± 21 % lower than the
control rate (Chapter 3, Appendix D). This was the same rate reduction shown for 10
mg/mL PVCap samples relative to the THF-water control, while PVP rates were less
divergent. The PVCap/SDS combination reduced the formation rate by even more (-74 ± 5
%). Given that PVCap is currently one of the most potent kinetic inhibitor of hydrates
available (Wua, 2007), these are very encouraging results. All three of these additives still
showed faster reformation from the melt, however pre-frozen samples containing
Chryseobacterium had a reformation rate equivalent to the that of fresh control samples
(k=0.39 ± 0.02; Appendix D). This too is promising, especially in light of the fact that the
1:15 (mol.) ratio of THF:culture can be harmful bacteria, rendering some isolates incapable
of colony formation after exposure (Author’s Undergraduate Thesis). The putative AFP
65
agent of hydrate inhibition activity associated with Chryseobacterium must be characterized and
isolated. Our working hypothesis is that this activity is conferred by its ice recrystallisation
inhibition activity, possibly and AFP. Once purified protein is obtained it may be seen to be
as potent a memory inhibitor as type I AFP (Zeng et al., 2006a, 2006b). The effect on gas
hydrate formation must also be assayed once pure protein is obtained. The same applies for
P. putida, whose samples had initially slower induction times and showed growth inhibition
through the appearance of hydrate slurries (Chapters 3, 4). These two species may have
future potential as practical and efficient hydrate inhibitors. In the meantime, it will now be
possible to undertake high-throughput searches of hydrate-associating polymers in the
search for more effective, “greener” inhibitors to be used in oil and gas exploration.
66
5.5. LITERATURE CITED
Sloan, D.E.. 2005. A changing hydrate paradigm--from apprehension to avoidance to risk management. Fluid Phase Equilibria 228-229 (PPEPPD 2004 Proceedings): 67-74
Wilson, S.L., Kelley, D.L., Walker, V.K.. 2006. Ice-active characteristics of soil bacteria selected by ice-affinity. Environmental Microbiology 8(10): 1816–1824
Wua, M., Wangb, S., Liuc, H.. 2007. A Study on Inhibitors for the Prevention of Hydrate Formation in Gas Transmission Pipeline. Journal of Natural Gas Chemistry 16(1): 81-85
Zeng, H., Moudrakovski, I., Walker, V.K., Ripmeester, J. A., 2006b. Inhibition activity of an antifreeze protein on hydrocarbon hydrate formation. AIChE J. 52: 3304-3309.
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J. A., 2006a. Effect of antifreeze proteins on the formation and reformation of tetrahydrofuran clathrate hydrate. Journal of the American Chemical Society 128: 2844-2850
67
Table 5.1. Association of hydrate and ice inhibitors with various crystalline structures and one non-crystalline surface (polystyrene).
Ice THF Hydrate Propane Hydrate SiO2 1
Poly-
styrene 1
F F M F M F Rinse F
type I AFP + 1 + 3 + 3 + 7 + 7 + + = A17L - 1 + 3 - 3 = - = PVP - 1 + 4 - 4 + 8 - 8 - - = PVCap + 5,6 - 5,6 + 8 - 8 = - Chrys. + 2 + 6 -? 6
+ binding capability or inhibition activity F First freeze/assay M memory 1 Appendix A 2 Wilson et al., 2006 3 Zeng et al., 2006a 4 Zeng et al., 2003 5 Chapter 2 6 Chapter 3, 4 7 Zeng et al., 2006b 8 Reviewed in Sloan, 2005
68
Figure 5.1. Proposed high-pressure apparatus for gas hydrate affinity purification (J.A. Ripmeester, personal communication)
69
SUMMARY
1. Regarding the manner of THF hydrate inhibition by bacteria with or proteins:
a) The strength and manner of binding to silica crystal was correlated to THF hydrate
inhibition activity.
b) The bacteria Chryseobacterium and P. putida were able to inhibit the growth of small
THF hydrate nuclei, although only Chryseobacterium was capable of significantly
slowing THF hydrate induction times.
c) AFPs appeared to be selectively incorporated into THF hydrate while non-ice-
binding proteins and dyes were expelled.
d) The presence of bacteria with both INP and AFP activity resulted in a short initial
period of hydrate inhibition.
e) No convincing evidence of nucleation inhibition was observed for the bacteria tested.
2. Regarding the nature of the memory effect in the presence of ice-associating
bacteria or proteins:
a) The ability of molecules to remain attached to silica crystal during rinsing was
correlated to inhibition of the memory effect in THF hydrate.
b) By freezing and melting the samples according to the procedures used here, the
probability of reformation was ~30% higher than for the first freeze. This is evidence
that the “memory effect” did occur.
c) The memory effect was not eliminated upon addition of any of the bacterial cultures.
d) The presence of bacterial INP enhanced the memory effect, suggesting that memory
is related to heterogeneous nucleation bodies.
70
A p p e n d i x A
THE MYSTERIES OF MEMORY EFFECT AND ITS
ELIMINATION WITH ANTIFREEZE PROTEINS
Virginia K. Walker1,2, Huang Zeng1, Raimond V. Gordienko1, Michael J. Kuiper1,2, Emily I.
Huva1, and John A. Ripmeester3.
1Department of Biology, 2Department of Microbiology & Immunology, Queen’s University,
Kingston, Ontario, K7L 3N6, Canada; 3The Steacie Institute for Molecular Sciences,
National Research Council of Canada, 100 Sussex Drive, Rm 111, Ottawa, Ontario, K1A
OR6, Canada
In: Proceedings of the 6th International Conference on Gas Hydrates, Vancouver, 2008.
A.1 ABSTRACT
Crystallisation of water or water-encaged gas molecules occurs when nuclei reach a critical
size. Certain antifreeze proteins (AFPs) can inhibit the growth of both of these, with most
representations conceiving of an embryonic crystal with AFPs adsorbing to a preferred face,
resulting in a higher kinetic barrier for molecule addition. We have examined AFP-mediated
inhibition of ice and clathrate hydrate crystallisation, and these observations can be both
explained and modeled using this mechanism for AFP action. However, the remarkable
ability of AFPs to eliminate ‘memory effect’ (ME) or the faster reformation of clathrate
hydrates after melting, prompted us to examine heterogeneous nucleation. The ubiquitous
71
impurity, silica, served as a model nucleator hydrophilic surface. Quartz crystal
microbalance-dissipation (QCM-D) experiments indicated that an active AFP was tightly
adsorbed to the silica surface. In contrast, polyvinylpyrrolidone (PVP) and
polyvinylcaprolactam (PVCap), two commercial hydrate kinetic inhibitors that do not
eliminate ME, were not so tightly adsorbed. Significantly, a mutant AFP (with no activity
toward ice) inhibited THF hydrate growth, but not ME. QCM-D analysis showed that
adsorption of the mutant AFP was more similar to PVCap than the active AFP. Thus,
although there is no evidence for “memory” in ice reformation, and the structures of ice and
clathrate hydrate are distinct, the crystallisation of ice and hydrates, and the elimination of
the more rapid recrystallisation of hydrates, can be mediated by the same proteins.
A.2 INTRODUCTION
Easily accessible traditional hydrocarbon supplies will be supplemented in the coming
decades with new stores situated in deeper off shore waters, in the permafrost, or sheathed
in crystalline water as gas hydrates. Prospecting, recovery and transport of this energy will be
not without its challenges. When pressure and temperature conditions are favorable,
hydrocarbon gases can form hydrate plugs that can lead to shutdown and financial losses. As
well, there is the potential for environmental damage due to unexpected hydrate formation.
For example, methane hydrate accidents in the Sea of Azov, Ukraine, resulted in mass fish
mortality even at sites located distant to the accident (Patin, 1999).
Traditionally, methanol is used to inhibit hydrates, but it is expensive due to the large
amounts required (10 to 50% of the water phase), and there are also environmental concerns
associated with its use in the Arctic and some European sectors (Kelland, 2006). Both
methanol and ethylene glycol, which has been used in the Gulf of Mexico, act by lowering
72
the crystallisation point of the hydrate. Thus cost considerations as well as environmental
concerns have motivated the search for more inexpensive, low dosage polymers (Kelland,
2006; Sloan et al., 1998). One such alternative to these traditional polyol inhibitors are newer
low dosage hydrate inhibitors (LDHIs), which can be classified as kinetic inhibitors (KIs)
and anti-agglomerates. At relatively low concentrations (<1% by weight), LDHIs appear to
function by slowing down hydrate growth or by reducing the likelihood of hydrate particle
agglomeration so that there is a reduced probability of blockage. Results have been
encouraging; when a KI was introduced into a flow line in the West Pembina (Alberta) field,
“downtimes” were reduced (Pickering et al., 2001). Thus the search for such new inhibitors is
to be encouraged.
The rapid addition of water molecules to embryonic crystals is an analogous problem
faced by certain organisms that live at temperatures lower than their theoretical freezing
points. Antifreeze proteins (AFPs) inhibit freezing in a non-colligative manner by adsorbing
to ice (DeVries, 1971). Since melting is effected in a colligative manner, the association
between AFPs and the ice surface results in a separation of the freezing point and the
melting point, a phenomenon termed thermal hysteresis (TH). The inhibition is thought to
derive from local ice surface curvature effects due to the presence of AFPs at the
ice/solution interface (the adsorption-inhibition hypothesis; Raymond and DeVries, 1977).
Because of this action, small quantities of AFPs can have large effects on ice crystal growth.
We hypothesized that gas hydrates, with their regular crystal lattice might act as an
alternative substrate for AFPs. Our experiments both with model gas hydrates
(tetrahydrofuran clathrate; THF) and methane and propane hydrates have shown that certain
AFPs inhibit gas hydrate crystallisation as well as eliminate faster recrystallisation (Zeng et al.,
73
A.3.1
2003, 2006a, 2006b). Despite these promising results, we know little of the mechanism of
this inhibition. The faster reformation or ‘memory effect’ (ME) is problematic in the field
when downtimes are expensive (Kelland, 2006; Sloan, 1988; Ohmura et al., 2003; Rodger,
2000; Kumar et al., 2008). It should be noted, however, that others have reported that ME is
an experimental artifact (e.g. Wilson et al., 2005). Nevertheless, here we have explored the use
of a quartz crystal microbalance to predict hydrate inhibition as a first step to explore the
observed differences in hydrate recrystallisation and eventually toward a more sophisticated
understanding of how these unique proteins interact with hydrates.
A.3 METHODS
Preparation and characterization of potential inhibitors
Polyvinylpyrrolidone (PVP K30, ~40,000 da; kindly provided by Dr. E.D. Sloan) as well as
wild type Type I AFP from winter flounder (kindly provided by Dr. G. Fletcher) solutions
were at a final concentration of 2.5, 12.5 and 25 μM. P polyvinylcaprolactam (PVCap;
~110,000 da; kindly provided by Dr. L. Talley) solutions were at 2 and 10 μM. All solutions
were prepared with Milli-Q® ultrapure water (Millipore, Bellerica, USA). A mutant Type I
AFP with a Leu for Ala substitution at position 17 (A17L) was chemically synthesized using
solid-phase peptide synthesis at the Queen’s University Protein Function Discovery Facility.
Peptide concentrations were routinely determined by amino acid analysis and A17L was used
in solutions at the final concentrations of 2.5, 12.5 and 25 μM.
TH was assessed using a nanoliter osmometer (Clifton Technical Physics, Hartford,
NY, USA) as previously described (Chakrabartty and Hew, 1991) and the ice crystal
morphology was noted. Inhibition of ice recrystallisation was assayed by rapidly freezing
74
A.1.1
samples in microcapillary tubes (10 μl) and allowing them to incubate at 267 K overnight
(Walker et al., 2008). After viewing through crossed polarizing filters those samples with no
recrystallisation inhibition activity were recognized by the presence of large ice crystals. Ice
nucleation assays were done as reported (Walker et al., 2008; Vali, 1971). Briefly, small
samples (10 μl) were loaded onto a polarizing filter and placed in a chamber where the
temperature was lowered from 272 K to 258 K. Digital photographs were captured every 60
sec though a crossed polarizing filter and automatically analyzed and transferred to a spread
sheet. The temperature at which 90% of the samples froze was taken as the ice nucleation
temperature.
Quartz crystal microbalance assessments
Surface adsorption was determined using a quartz crystal microbalance (QCM) equipped to
determine the energy loss or dissipation factor (D). This QCM-D (Q-Sense D300, Q-Sense
AB, Gothenburg, Sweden) with 5-MHz AT-cut crystals had a sensor crystal was coated with
SiO2 (QSX-303) on one side of the gold electrode. This was cleaned and placed in a 250 μl
measurement chamber with ultrapure water equilibrated at 299 K. Once a stable baseline was
achieved, aliquots (0.5 ml) of the solutions (1.5 ml) to be tested were introduced into the
measurement chamber, replacing the water. Frequency shift (f) and D were sampled at a rate
of ~1 Hz with a sensitivity of <0.5 Hz and 1×10-7, respectively. After the initial adsorption
measurements, the sensor was rinsed three times with 0.5 ml ultrapure water at 299 K.
Assessments of f and D were obtained after each rinse. Control experiments were done with
a hydrophobic substrate, a polystyrene surface (QSX-305), on the sensor crystal.
75
A.2.1
A.2.2
A.2 RESULTS
Characterization of ice-associating properties
Type I AFP showed TH activity consistent with previously published values. At a
concentration of 1 mg/ml it showed a TH of approximately 0.3ºK. The A17L mutation
showed no TH activity at this or higher concentrations. Ice crystals grown in the presence of
Type I AFP showed the diamond-like morphology typical of this AFP (Figure A.1a). In
contrast, the mutant A17L protein resulted in no change to the flat circular crystals seen in
the absence of AFPs. Since ice recrystallisation inhibition is a property that can be assessed
at very low concentrations of AFPs, this assay was also used to assess the activity of the two
AFPs. Type I AFP showed complete inhibition of ice recrystallisation even at the lowest
concentrations used, whereas the mutant A17L showed large crystals and the complete
absence of ice recrystallisation at every concentration tested (Figure A.1b; not shown).
Neither AFP nor the mutant A17L showed any ice nucleation activity (Figure A.1c; not
shown). PVP showed no ice association activity using ice recrystallisation inhibition activity
or ice nucleation assays (Figure A.1).
Quartz crystal microbalance assessments
When AFP, A17L, PVP and PVCap were analyzed using QCM-D all showed an initial
decrease in f before levelling out at a value consistent with the mass of the adsorbent on the
surface (Figure A.2; not shown). This adsorption was also concentration dependent; more
molecules were adsorbed to the SiO2 surface as the concentration of the potential inhibitors
was increased (Table A.1 and not shown). Adsorption masses, as assessed by the final values
of f, were higher for the PVP solutions than for PVCap and these masses were less for the
76
two proteins than those of the two polymers. For example, at 10-12 μM, the value of f for
PVP was -19.1 Hz, -0.59 Hz for PVCap, -2.1 Hz for AFP and -3.7 for A17L.
QCM-D analysis not only reveals the adsorption mass, but the dissipation factor
represents the viscoelastic properties of the adsorbed molecules (Rodahl et al., 1995; Rodal et
al., 1997). The D value for PVP, PVCap and A17L increased throughout the assay period,
but that for AFP remained almost constant (Figure A.3; not shown). The slope (R), of these
curves was calculated as R = ΔD/Δm. Thus R indicated the status of the adlayer. PVP,
PVCap and A17L showed two distinct steps of adsorption, R1 and R2, and for A17L at high
concentrations (> 12.5 μM), there was some evidence of a third step (Table A.1). The final
status of the adlayer is reflected in the last R values. PVP showed the highest final│R│values
(R2) of the four molecules tested, whereas the lowest final│R│values (0 Hz) were seen with
the wild type AFP (Table A.1).
When adsorption on the hydrophobic polystyrene surface was tested with the
potential inhibitors (PVP, AFP and A17L; PVCap was not assayed) the kinetic parameters
for adsorption showed final│R│values that were the same for all the tested molecules (-0.12
to -0.13 x 10-6 Hz-1; Table A.1).
A.3 DISCUSSION
Evolutionary pressures have designed AFPs to adsorb to the surface of seed ice crystals
thereby increasing the energy barrier for ice growth due to the Kelvin effect (Knight, 2000).
This inhibition mechanism occurs subsequent to heterogeneous nucleation. AFPs show a
remarkable variation in structure: some fish AFPs, including the Type I AFP used here are α-
helices, but with other fish and insect AFPs showing β-helices, β-rolls or more globular
77
structures. Therefore, it appears that the structure per se does not dictate ice association
properties, but a more general complementarity with particular residues exerting van der
Waals attractions and possibly a combination of hydrogen bonding and other hydrophobic
interactions (Haymet et al., 1998) facilitating the snug, irreversible fit to ice. Indeed, the 17th
residue in Type I AFP is so important for this interaction that a Leu substitution at this site
results in a loss of all ice association activity. We report here that there is no affinity for ice
by the mutant as assayed by ice crystal morphology and ice recrystallisation inhibition activity
(Figure A.1). The complete loss of TH activity in A17L was reported previously and was
used to support a new proposed “ice binding” face of the Type I AFP α-helix (Baardsnes et
al., 2001). PVP also did not show ice association as revealed by assays to detect inhibition of
ice recrystallisation or ice nucleation (Figure A.1).
AFPs have been shown to inhibit the growth of model and gas hydrates (Zeng et al.,
2003, 2006a, 2006b), but the mechanism for this inhibition is unknown. Significantly, the
A17L mutant showed similar inhibition of THF hydrate growth as the active AFP (Zeng et
al., unpublished), demonstrating that adsorption to hydrate and ice crystals is not mediated
by identical residues. Curiously, however, although the active AFP eliminated the more rapid
recrystallisation of hydrate after a brief melt (Zeng et al., 2006a), the A17L mutant did not
eliminate ME. Significantly, ME was also not eliminated by either of the tested KIs, PVP or
PVCap (Zeng et al., 2008). We reasoned that active AFP may be unique in its ability to
eliminate ME due to inhibition of heterogeneous nucleation. However, since A17L was
inactive against ice, it was a formal possibility that ice crystals could act as heterogeneous
hydrate nucleators, which grew to a critical size only in the absence of AFP activity. By
ensuring that the temperature of THF hydrate was kept above 272 K, this risk was
78
eliminated (Zeng et al., unpublished).
A second possibility was that active AFP could inhibit hydrate recrystallisation by
inhibiting heterogeneous nucleation on the surface of silica, a ubiquitous stable ice nucleator
(Klier et al., 1973). QCM-D analysis showed that the four tested molecules could be grouped
into three different classes based on their interaction with silica. PVP loosely associated with
silica; the adsorption mass was high, and this KI had the highest final │R│values of any of
the tested molecules, consistent with a porous adlayer with tapped water molecules. PVCap
and A17L formed a somewhat more compact film, with final│R│values that were 3-4 fold
less than the PVP values, indicating that less water was trapped on the surface. Remarkably
however, the lowest final│R│values of the tested molecules were found for wild type AFP,
consistent with a highly dense, compact adlayer, with very little water. These differences in
silica adsorption reflect the different effect the proteins and polymers had on ME, and
further suggest that such film differences could have a distinct influence on heterogeneous
nucleation.
We propose that hydroxylated silicon or hydrated silica nano or micro particles
present during the initial hydrate crystallisation become ‘imprinted’ and thus become more
effective nucleators for ME recrystallisation. This work shows that the hydroscopic polymers
PVP and PVCap as well as A17L can form films on SiO2 but since the aggregation is either
relatively loose or can easily wash off (not shown), there is no elimination of ME. In
contrast, active AFP adsorbed to silica and retained a tight compact surface with little
trapped liquid. As a consequence, water and guest molecules cannot so easily reach the
nucleating sites making heterogeneous nucleation on such a surface much less probable
(Figure A.4). As well, it appears that the ice binding site and the surface that tightly binds
79
silica coincide, given the differences observed with QCM-D in the adlayers for the active
AFP compared to the A17L mutant. The ice adsorption face of Type I AFP is relatively
hydrophobic (Baardsnes et al., 2001, Baardnes et al., 1999), and therefore the more
hydrophilic side of the active AFP would consistently face away from the silica surface.
In conclusion, these studies demonstrate that the properties of adsorbed layers can
be monitored effectively by QCM-D. The results have provided useful information about
the inhibition mechanism of heterogeneous nucleation of clathrate hydrate. This technique
offers opportunities to screen potential LDHIs, and to examine residues in AFPs that are
involved in silica adsorption, and by extension the inhibition of heterogeneous nucleation.
However, the most important practical ramification may be that these studies offer a
procedure to further investigate the mysterious memory effect, a phenomenon that has been
hotly debated by lab scientists who have denied its existence and those in the field who see
its destructive capability first hand.
A.4 ACKNOWLEDGEMENTS
We thank Mr. A. Brown for examining TH and ice morphology of the A17L mutant and Ms.
Z. Wu for the P. borealis ice nucleation data. The authors acknowledge Drs. G.R. Palmer and
P.L. Davies for assistance with equipment and their encouragement.
80
A.5 LITERATURE CITED
Baardnes, J., Kondejewski, L.H., Hodges, R.S., Chao, H., Kay, C., Davies, P.L.. 1999. New ice-binding face for type I antifreeze protein. FEBS Letts. 463: 87-91.
Baardsnes, J., Jelokhani-Niaraki, M., Kondejewski, L.H., Kuiper, M.J., Kay, C. M., Hodges, R.S., Davies, P.L.. 2001. Antifreeze protein from shorthorn sculpin: identification of the ice-binding surface. Protein Science 10: 2566-2576.
Chakrabartty, A., Hew, C.L.. 1991. The effect of enhanced α-helicity on the activity of a winter flounder antifreeze polypeptide European Journal of Biochemistry. 202: 1057-1063.
DeVries, A.L.. 1971. Glycoproteins as biological antifreeze agents in antarctic fishes. Science172: 1153-1155.
Haymet, A.D.J., Ward, L.G., Harding, M.M., Knight, C.A.. 1998. Valine substituted winter flounder “antifreeze”: preservation of ice growth hysteresis. FEBS Lett.430: 301-306.
Kelland, M.A.. 2006. History of the development of low dosage hydrate inhibitors. Energy and Fuels 20: 825-847.
Klier, K., Shen, J.H., Zettlemoyer, A.C.. 1973. Water on silica and silicate surfaces. I. Partially hydrophobic silicas. Journal of Physical Chemistry. 77: 1458-1459.
Knight, C.A.. 2000. Structural biology: adding to the antifreeze agenda. Nature 406: 249-251.
Kumar, R., Lee, J.D., Song, M., Englezos, P.. 2008. Kinetic inhibitor effects on methanepropane clathrate hydrate-crystal growth at the gas-water and watern-heptane interfaces. Journal of Crystal Growth 310: 1154-1166.
Ohmura, R., Ogawa, M., Yasuoka, K., Mori, Y.H.. 2003. Statistical study of clathrate-hydrate nucleation in a water/hydrochlorofluorocarbon system: Search for the nature of the “memory effect”. Journal of Physical Chemistry. B. 107: 5289-5293.
Patin, S.. 1999. Environmental impact of the offshore oil and gas industry. New York: EcoMonitor Publishing..
Pickering, P.F., Edmonds, B., Moorwood, R.A.S., Szczepanski, R., Watson, M.J.. 2001. Evaluating new chemicals and alternatives for mitigating hydrates in oil and gas production. In: IIR Conference, Aberdeen, Scotland..
Raymond, J.A., DeVries, A.L.. 1977. Adsorption inhibition as a mechanism of freezing resistance in polar fishes. Proceedings of the National Academy of Science USA 74: 2581-2593.
Rodahl, M., Hook, F., Fredricksson, C., Keller, C., Krozer, A., Brzezinski, P., Voinova, M., Kasemo, B.. 1997. Simultaneous frequency and dissipation factor QCM measurements of biomolecular adsorption and cell adhesion. Faraday Discussions.107: 229-246.
Rodahl, M., Hook, F., Krozer, A., Brzezinski, P., Kasemo, B.. 1995. Quartz crystal microbalance setup for frequency and Q-factor measurement in gaseous and liquid environments. Review of Scientific Instruments. 66: 3924-3930.
Rodger, P.M.. 2000. Methane hydrate melting and memory. Annals of the New York Academy of
81
Science 912: 474-482.
Sloan, E.D., Subramanian, S., Matthews, P.N., Lederhos, J.P., Khokhar, A.A.. 1998. Quantifying Hydrate Formation and Kinetic Inhibition. Industrial and Engineering Chemistry Research 37: 3124-3132.
Sloan, E.D. Jr.. 1988. Clathrate hydrates of natural gases, 2nd Ed. New York: Marcel Dekker.
Vali, G.. 1971. Quantitative evaluation of experimental results on the heterogeneous freezing nucleation of supercooled liquid. Journal of the Atmospheric Sciences 28: 402-409.
Walker, V.K., Wilson, S.L., Wu, Z., Miao, D., Zeng, H., Ripmeester, J.A., Palmer, G.R.. 2008. Selection for ice-associating molecules and their potential application as “green inhibitors” for gas hydrates In: V. Shah, Editor, Emerging Environmental Technologies, Springer Publications, New York: in press.
Wilson, P.W., Lester, D.J., Haymet, A.D.J.. 2005. Heterogeneous nucleation of clathrates from supercooled tetrahydrofuran (THF)/water mixtures, and the effect of an added catalyst. Chemical Engineering Science 60: 2937-2941.
Zeng, H., Moudrakovski, I.L., Ripmeester, J.A., Walker, V.K.. 2006b. Effect of Antifreeze Proteins on Nucleation, Growth and Memory in Hydrocarbon Hydrate Formation. AIChE Journal 52: 3304-3309.
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J.A.. 2006a. Effect of antifreeze proteins on the formation and reformation of tetrahydrofuran clathrate hydrate. Journal of the American Chemical Society 128: 2844-2850.
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J.A.. 2003. The inhibition of tetrahydrofuran clathrate hydrate formation with Antifreeze protein. Canadian Journal of Physics 81: 17-24.
Zeng, H., Lu, H., Huva, E.I., Walker, V.K., Ripmeester, J.A.. 2008. Differences in nucleator adsorption may explain distinct inhibition activities of two gas hydrate kinetic inhibitors. Chemical Engineering Science, in press (Accepted April 2008, 9 pp.)
82
Table A.1. Kinetic parameters for adsorption of antifreeze proteins (wild type AFP and A17L mutant) as well as kinetic inhibitors on SiO2 and polystyrene, a control hydrophobic surface, as assessed by QCM-D.
SiO2 polystyrene
Sample AFP A17L PVP PVCap AFP A17L PVP
~ Conc.
(μM)
25 12 25 12 25 12 10 25 25 25
R1
(x 10-6 Hz-1)
-0.04 -0.03 -0.04 -0.06 -0.14 -0.12 -0.36 -0.02 -0.03
R2
(x 10-6 Hz-1)
0.00 0.00 -0.01 -0.22 -0.19 -0.18 -0.06 -0.12 -0.12
-0.13
R3
(x 10-6 Hz-1)
_ _ -0.07 -0.05 _ _ _ _ _ _
f total
(Hz)
-11.1 -7.6 -19.9 -18.03 -29.5 -27.33 -1.6 -12.6 -17.1 -34.1
83
(a) (b)
H2O AFP A17L PVP
0
20
40
60
80
100
255260265270275
Tem perature (K)
% fr
ozen
(c)
Figure A.1. Assessment of ice association. (a) Microscopic ice crystals grown in the presence of Type I AFP at 1 mg/ml (left) buffer controls or the A17L mutant, which were indistinguishable. (b) Ice recrystallisation inhibition assays showing duplicate microcapillaries containing (from left to right) water controls, Type I AFP (2.5 μM), A17L (12.5 μM) and PVP (12.5 μM). The first row shows a portion of the capillary immediately after freezing and the second row shows the capillary after 16 h at temperatures close to the melt, with larger crystals in all samples except AFP. (c) Ice nucleation assays with 25 μM AFP (grey line), A17L (long dashed line) and PVP (short dashed line). P. borealis (solid line) has known ice nucleation activity.
84
(a)
5
f (H
z)
0 10 20 30 40-30
-20
-10
0
(b)
f (H
z)
0 6
- 1 0 0 1 0 2 0 3 0 4 0
- 4
- 2
0 A A w
time (min)
Figure A.2. Representative graphs showing (a) the adsorption of AFP (25 μM, solid line) and PVP (25 μM, dotted line) as well as (b) AFP (2.5 μM, solid line) and A17L (2.5 μM, dotted line) on the SiO2 surface at 299 K. Frequency shift (f) vs. time was assessed using QCM-D.
85
-30 -25 -20 -15 -10 -5 0-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Adsorption of AFPs and PVP (12.5uM) on Silica (299 K)
wfAFP A17L PVP
BD
(10
-6)
f (Hz)
Figure A.3. Representative graph showing the relationship between dissipation factor (D) and the frequency (f) of PVP (left arrow-head), AFP (triangle) and A17L (octagon) all at 12.5 µM on the SiO2 surface at 299 K. Assessments of PVCap compared to PVP were done at a different time and are therefore not included in this figure. Experiments were conducted using QCM-D.
86
vs.
= contaminating particle = H2O/gas = active
pipe wall
Figure A.4. A cartoon depicting the reformation of hydrates in the presence or absence of certain AFPs. The heterogeneous nucleator (light hexagon) is depicted as a large particle but hydrated silica nanoparticles or hydroxylated silicon are included. Hydrate (dark circles) assembles on the surface of these nucleators in the absence of active AFPs shown as α-helices (left scheme), but not in their presence (right scheme). It should be noted that AFPs do not need to entirely cover the silica surface, but sufficient numbers of molecules will make the addition of water molecules on the SiO2 surface unfavorable.
87
A p p e n d i x B
INDUCTION TIME EXPERIMENTS WITH THE HYDRATE
PROMOTER, SNOMAX
B.1 INTRODUCTION AND METHODS
The commercial ice-nucleator Snomax has been used to promote gas hydrate formation
(Chapter 1), and served to increase nucleation in subcooling experiments (Appendix E).
Induction time experiments (described in Chapter 2 and Chapter 3) were used to assess its
ability to promote THF hydrate.
B.2 RESULTS AND DISCUSSION
From Figure B.1 it is seen that Snomax activity is in fact closer to that of the hydrate
inhibitor, PVP, than to the control. The Snomax product contains freeze-dried media, i.e.
salts and protein, which makes up ~50% of its mass (pers. communication, Snomax
representatives). At the concentration used (0.05 mg/mL) the THF solution appeared quite
cloudly. It is possible that the colligative effects of these additives was sufficient to
substantially reduce the potential for hydrate formation. Future experiments might employ
laboratory-purified INP instead, or compare Snomax solutions of equivalent media or salt
content with solutions of bacterial cultures.
88
0
0.5
1
0 5 10 15 20 25t (hours)
n C
Figure B.1. THF hydrate induction times in the presence of no additives (■), 0.5 mg/mL Snomax (×), and 10 mg/mL PVP (♦). Curve represents best-fit power law (Appendix D)
89
A p p e n d i x C
INDUCTION TIME EXPERIMENTS WITH THE WATER
VISCOSIFIER, XANTHAN GUM
C.1 INTRODUCTION AND METHODS
A bioemulsifier from the bacteria Acinetobacter calcoaceticus was found to promote gas hydrate
formation in packed sand/clay at a rate comparable to that of Snomax (Rogers et al., 2003).
The commercial product, Emulsan, is a polysaccharide-lipid complex (Rosenberg, 1993).
Xanthan gum is an emulsifier produced by the bacteria Xanthomonas campestrins; some isolates
can also produce INPS (Kim et al., 1987). A two-pronged five-sugar chain, it is able to form
stiff rod-shaped dimers (Milas and Rinaudo, 1983) which enable it to act as a water
viscosifier often used in salad dressings, and to emulsify the fats in ice-creams to maintain
creaminess. Induction time experiments (described in Chapter 2 and Chapter 3) were used to
assess its ability to promote THF hydrate.
C.2 RESULTS AND DISCUSSION
Xanthan gum did not appear to significantly change the induction time of THF hydrate
crystallisation compared to experiments without additives
(t (hours)
Figure C.1):at the completion of the experiment, 57% of xanthan samples and 56% of non-
xanthan samples had crystallised. This finding supports the results of subcooling
experiments, in which xanthan gum did not elicit any increase in freezing temperature
(Appendix E).
90
0
0.5
1
0 5 10 15 20 25
n C
t (hours)
Figure C.1. THF hydrate induction times with (Δ) and without (■) 0.5 mg/mL xanthan gum. Curve represents best-fit power law (N=48; equation in Appendix D).
91
C.3 LITERATURE CITED
Kim, H.K., Orser, C., Lindow, S.E., Sands, D.C.. 1987. Xanthomonas campestris pv. translucens strains active in ice nucleation. Plant Disease 71(11): 994-997
Milas, M., Rinaudo, M.. 1983. Properties of the concentrated xanthan gum solutions. Polymer Bulletin 10(5): 271-273
Rogers, R.E., Kothapalli, C., Lee, M.S., Woolsey, R.J.. 2003. Catalysis of gas hydrates by biosurfactants in seawater-saturated sand/clay. Canadian Journal of Chemical Engineering 81: 1-8
Rosenberg, E.. 1993. Microbial Diversity as a Source of Useful Biopolymers. Journal of Industrial Microbiology 11: 131–137
92
A p p e n d i x D
POWER-LAW MODEL OF TETRAHYDROFURAN HYDRATE
FORMATION
D.1 PARAMETER CALCULATIONS
Parameters for fitting induction-time data to a power-law model:
1+⋅= mC tkn (1)
were calculated by log-log regression in Microsoft Excel (α = 0.05) and are listed in Table
D.1. (Note: log-log regression is the linear regression of log(nC) as a function of log(t)). The
fits are presented visually in Figure D.1 and Figure D.2.
D.2 CONVERSION OF PARAMETERS INTO “LAG-TIME”
For a lag-time (Chapter 4) of:
( ) 11
2 +−= mkτ (2)
the random errors of k and m are propagated in quadrature, that is:
( ) ...22
2 +⎟⎠⎞
⎜⎝⎛ Δ∂∂
+⎟⎠⎞
⎜⎝⎛ Δ∂∂
=Δ bbya
ayy (3)
giving the uncertainty on τ as:
( )( )
( )
2
2
2
12ln
1 ⎟⎟⎠
⎞⎜⎜⎝
⎛Δ⋅
+⋅
+⎟⎟⎠
⎞⎜⎜⎝
⎛Δ⋅
⋅+=Δ m
mkk
kmτττ (4)
Although lag-times are a useful quantity to use for comparing data sets, there are large
uncertainties associated with lag-time calculations for a power-law model (Table D.1) and it
93
is unable to distinguish between the effects of time-dependent and time-independent
nucleation.
D.3 NON-ISOTHERMAL POWER-LAW MODEL
For supercooling experiments,
α,
,ratecooling
Tsubcoolingt Δ= . (5)
Substitution of (4) into (1) yields the fitting equation:
1+
⎟⎠⎞
⎜⎝⎛ Δ⋅=
m
CTknα
. (6)
Note that values obtained for k and m vary with the chosen cooling rate.
D.4 POWER-LAW FITS FOR MULTI-PHASIC FREEZING
All bacterial samples save Chryseobacterium appeared to exhibit multiple phases of hydrate
formation, in contrast to either recrystallisation behaviour or trends for synthetic samples; a
log-log plot of data (Figure D.1b) shows sequences of ~3 straight lines, instead of the one
expected for a single power-law relationship and seen in other data sets (Figure D.1d, Figure
D.2b). Multiple freezing phases of sustained activity compared to control was also seen in
Figure 4.3. A breakdown of power-law parameters into “phases” (Table D.2), chosen at the
authors’ discretion, lends numerical support for this observation. Most samples at some
point shared control parameters, having k and m near 0.40 and -0.88, respectively. However,
extended periods of slower and steadier crystallisation were seen for type I AFP and
Chryseobacterium, which for the former was followed (t = ~21.5 h) by a period of rapid
formation analogous to that seen for both DIA bacteria at t = 4-5 h. Freezing rate for the
94
DIA bacteria later returned to normal (control).
5.6. ERROR CALCULATIONS FOR K-VALUE COMPARISONS
For samples with rate A compared to the control, B, the rate reduction is:
C = BBA )( −
(7)
and the propagation of error in quadrature (Eqn. 3) is:
22
⎟⎠⎞
⎜⎝⎛ Δ+⎟
⎠⎞
⎜⎝⎛ Δ⋅=Δ
BB
DDCC
(8)
where
( ) ( )22 BAD
BAD
Δ+Δ=Δ
−=
(9)
95
Table D.1. Parameters for power-law fits to THF hydrate induction time data.
k N τ Sample [fraction/h-m+1]
m R2
model total [hours]
E. coli 0.39 ± 0.05 -0.84 ± 0.07 0.85 18 44 4.7 ± 5.0E. coli+ wfAFP 1 0.12 ± 0.01 -0.42 ± 0.03 0.97 18 25 11.7 ± 2.2Chryseobacterium 0.13 ± 0.02 -0.68 ± 0.06 0.97 10 36 67.3 ± 62.2P. putida 0.36 ± 0.04 -0.85 ± 0.04 0.92 22 44 8.9 ± 8.4P. borealis 0.26 ± 0.04 -0.76 ± 0.08 0.93 11 37 15.3 ± 17.0P. syringae 0.40 ± 0.03 -0.86 ± 0.04 0.94 15 36 4.9 ± 3.5water 0.57 ± 0.00 -0.83 ± 0.00 0.98 34 48 0.5 ± 0.0Snomax 0.42 ± 0.01 -0.84 ± 0.01 0.97 16 38 3.0 ± 0.5PVCap 0.19 ± 0.01 -0.59 ± 0.01 0.96 31 55 10.6 ± 1.5PVCap + SDS 0.15 ± 0.01 -0.57 ± 0.02 0.96 20 37 16.4 ± 3.3PVP 0.44 ± 0.01 -0.84 ± 0.02 0.91 11 52 2.2 ± 0.4Recrystallisation E. coli 0.42 ± 0.04 -0.73 ± 0.05 0.97 27 36 1.9 ± 0.7E. coli+ wfAFP 1.01 ± 0.30 -0.98 ± 0.01 0.96 5 25 0.0 ± 0.0Chryseobacterium 0.39 ± 0.02 -0.81 ± 0.02 0.99 21 36 3.7 ± 1.1P. putida 0.63 ± 0.05 -0.81 ± 0.04 0.96 26 33 0.3 ± 0.1P. borealis 2 n/a n/a n/a n/a 11 0 ± 0water 0.96 ± 0.01 -0.98 ± 0.00 0.89 10 54 0.0 ± 0.0PVCap 0.99 ± 0.00 -0.97 ± 0.00 0.98 8 67 0.0 ± 0.0PVCap + SDS 0.84 ± 0.01 -0.99 ± 0.00 0.81 4 52 0.0 ± 0.0PVP 0.92 ± 0.00 -0.99 ± 0.00 0.91 9 47 0.0 ± 0.0
1 first data point excluded (included, R2 = 0.73 due to weighting in log-log regression) 2 insufficient data for model (nC = 1 at t = 0)
96
Table D.2 Power-law parameters for a multi-phase model of THF hydrate formation in the
presence of ice-associating bacteria.
I II III IV
Sample IAA3 k m Δt
[h]k m Δt
[h]k m Δt
[h] k m Δt
[h]E. coli 1 none 0.71 -0.34 0.4 0.44 -0.90 23.6
E. coli +AFP 1,2 AFP 0.19 -0.95 2.3 0.11 -0.36 10.6 0.39 -0.86 8.5 rapid formation* 1.6
Chryseobacterium 1 AFP 0.13 -0.69 23
P. putida 1 both 0.32 -0.90 4.7 rapid formation* 0.6 0.34 -0.80 17.7
P. borealis 1 both 0.25 -0.76 4.0 rapid formation* 1.1 0.35 -0.88 17.9
P. syringae 1 INP 0.36 -0.92 1.5 0.30 -0.66 1.6 0.44 -0.89 19.9 1 median R2 = 0.96 (4-10 data points per fit) 2 For I and III, R2 = n/a (2 data points per fit); AFP = 0.05 mM type I fish AFP (winter flounder) 3 Known ice-associating activity
* data did not fit model
97
0
0.5
1
0 5 10 15 20 25
n C
a0.1
1
0.01 0.1 1 10 100n C
b
t (hours) t (hours)
0
0.5
1
0 5 10 15 20 25
n C
c
0.1
1
0.01 0.1 1 10 100
n C
d
t (hours) t (hours)
Figure D.1. Power-law fit of induction times in the presence of ice-associating bacteria showing contrast between formation (a, b) and reformation (c, d) of THF hydrate. Crystallized sample fractions nC are shown in linear scale (a, c) and log-log scale (b, d). Solutions consisted of 1:3.34 (vol) THF:culture of: E. coli (□),E. coli+0.05 mM type I AFP (o), Chryseobacterium (■), P. putida (●), P. borealis (∆), and P. syringae (▲).
98
0
0.5
1
0 5 10 15 20 25
n C
a
0.1
1
0.01 0.1 1 10 100n C
b
t (hours) t (hours)
Figure D.2. Power-law fit of induction times in the presence of hydrate-associating polymers for the formation (solid shapes) and reformation (hollow shapes) of THF hydrate, on a linear (a) and log-log (b) scale. Solutions consisted of 1:3.34 (vol) THF:culture of: purified water (■/□); PVP (10 mg/mL; ♦/◊ ); PVCap (10 mg/mL; ●/ o); PVCap (10 mg/mL) + SDS (5 mg/mL; ▲/∆); and Snomax (0.5 mg/mL; ×)
99
E.2.1
A p p e n d i x E
ASSESSMENT OF TETRAHYDROFURAN HYDRATE
PROMOTION FROM SUPERCOOLING MEASUREMENTS
E.1 INTRODUCTION
For biologists, the promotion of water ice by ice-nucleators is normally gauged by the
temperature at which a cooled sample will freeze (Chapter 1). Although measurement of
isothermal induction time is perhaps the most common way of assessing hydrate promotion
or inhibition activity, a few researchers have also employed this method (Young et al., 1994;
Heneghan et al., 2002). Several kinds of subcooling experiments are explored as a means to
test the nucleation of tetrahydrofuran hydrate by biological ice-nucleators. Several putative
hydrate promoters were examined. Cultures of the ice-nucleator Pseudomonas syringae and ice-
shaper P. borealis (Wilson et al., 2006) were tested, along with commercial freeze-dried ice-
nucleator (Snomax) from P. syringae, and a commercial surfactant (xanthan gum) from ice-
nucleator Xanthomonas campestrins (Kim et al., 1987). Additionally, the potential for metal
thermocouples to incite hydrate formation was explored.
E.2 METHODS
Sample preparation
All bacteria were grown in 10% Tryptic Soy Broth (TSB) for 24h at RT (~25oC), shaken, and
then for 24h in a 4oC refrigerator. Bacteria were then kept in the fridge and age was counted
from this point. Cultures aged 1-3 days were used in experiments. The average bacterial
100
E.2.2
E.2.3
concentration after 2 days at 4oC was assessed to be ~1x109 by by plating in triplicate1 100
µL of a 10-6 dilution on 10% TSB agar plates (Table E1)
THF was mixed at a 1:3.34 volume ratio (1:15 molar ratio) with either milliQ purified
water (DSC experiments), de-ionized water (bulk and microcapillary experiments), or
bacterial culture on day two or three of incubation at 4oC.
Supercooling of samples in bulk and in microcapillaries
Aliquots of 1 mL were loaded into seven 2 mL capped vials. The vials were immersed in a
temperature-control bath while thermistors were used to monitor temperature; an eighth vial
contained 1 mL bath coolant (automobile ethylene-glycol mixture).
Thirty glass 10 µL microcapillaries were loaded with solution, plugged with
plasticine, and mounted into a frame that was immersed in stirred secondary bath in a pyrex
bowl around which coolant from the temperature-control bath flowed. A light shining
through the bottom of the bowl illuminated the samples against a strip of black paper
background. A digital camera automatically recorded images of the samples every minute
while the temperature of the secondary bath was digitally recorded.
The primary bath was cooled at a rate of -0.2 oC/min while the temperature of the
secondary bath followed after a lag time. Results were pooled for generation of figures.
Differential Scanning Calorimetry
A Differential Scanning Calorimeter (DSC) can be programmed for precise cooling and
heating of one small (~9 µL) samples. Its output of heat vs. temperature reveals the freezing
1 P. syringae was only plated in duplicate at this dilution
101
E.3.1
and melting temperatures of samples, while information regarding formation enthalpies
could distinguish between ice and hydrate formation at subzero temperatures.
Samples of 8.0 mg ± 0.6 were tightly sealed into DSC crucibles. Equilibrium at 10 oC
was followed by ramping at 2 oC/min to -42 oC, then ramping at 2 oC/min to +10 oC. The
device was first calibrated using MilliQ H2O and mercury.
The homogeneous melting temperature of THF hydrate was established as a control
using the following method: 0.5 mL THF was added to 1.5 mL silicon oil, then shaken by
hand; ~0.03 g Span60 surfactant was added and dissolved by 5 min sonication; finally, 0.1
mL THF-H2O 1:15 mol. ratio solution was mixed in by 30 min sonication in an ice-bath.
Stability of the emulsion was confirmed if no separation had occurred 10 min later.
Heterogeneous nucleation was assessed as a sudden spike in heatflow at temperatures above
the homogeneous nucleation temperature of THF.
E.3 RESULTS AND DISCUSSION
Supercooling of samples in bulk and in microcapillaries
Previously frozen bulk samples began crystallisation several degrees above the initial freeze,
with no distinction between Chryseobacterium and the control, whereas no “memory effect”
was observed in microcapillary experiments (Figure E.1a,b i) Ice nucleators P. syringae, P.
borealis, and Snomax marginally increased the temperature of hydrate formation in bulk (by
~2oC, ~10% of total subcooling), however the difference is less clear in microcapillary
experiments (Figure E.1a,b ii). Xanthan gum showed no increase in bulk freezing while it
slightly depressed freezing in microcapillaries (Figure E.1b,c iii). The addition of metal wires
(Figure E.3) did not significantly shift the crystallisation curve.
102
These results confirm the trends seen elsewhere in this work, however data was
difficult to interpret when the largest change seen was approximately 10%. Statistical
comparisons were done of the temperatures at which 10, 50, and 90 % of samples had
frozen, accross N=3 experiments (xanthan gum and Snomax and N<3 and were not
assayed). The only significant difference was between the 90% frozen temperature of P.
syringae or P. borealis and E. coli (Tukey HSD, df=5, p<0.05). Perhaps the driving force at
such large subcoolings in the presence of bacterial cells, which can themselves be nucleators,
scaled down the effects of additives.
Encouragingly, a fit of data to a power-law model (not shown; Appendix D) was
generally successful (R2≈0.92) except in memory freezing. Accurate fitting would however
require a better fit method than log-log regression method, which has weighting problems,
and further work to develop a “memory” model.
It was thought that the freezing and melting T vs. t data might offer insight into the
nature of the crystals formed, i.e. multicrystallinity, and whether ice or hydrate is formed. The
bath temperature was subtracted from sample temperatures and the differential freezing and
melting curves reviewed (Figure E.2a, b). Aside from blatant freezing of one sample in two
sections or non-freezing, generally the curves are indistinguishable. On a side note, analysis
of induction time freezing curves can be informative: a startling difference was revealed
between the speeds of growth of PVCap as compared to most other samples (Figure E.2c).
In conclusion, although supercooling experiments generally supported the
conclusions of induction time experiments, they were insufficient on their own to give clear
insight into promoters or inhibitors.
103
E.3.2 Differential Scanning Calorimetry
The homogeneous nucleation point of THF hydrate was found to be -37oC, and the melting
point, 4.14oC, in agreement with the literature. The heterogeneous nucleation temperature
with no additives present was, in contrast, ~18oC (Table E.2). Within the uncertainties of
this experiment, Pseudomonas syringae was the only culture showed significantly reduced
subcooling; Chryseobacterium slightly increased subcooling though not significantly.
Upon heating, melting curves for both hydrate and water ice were seen. At the
concentration used, the area of the melting curve (Heat flow x time) was sometimes twice
that for hydrate. A volume ratio of 1:2.17 was required to reliably remove any sign of an ice
melt, however this likely resulted in the coexistence of liquid THF with the hydrate phase
during the experiment.
Ice and hydrate clearly co-exist during the experiment, however even with kinetic
data available it is difficult to conclude that the heterogeneous freezing of THF was not
instigated by an ice nucleus. More importantly, the limit of DSC experiments to one sample
makes any statistical assessment of supercooling impractical. If the experiment was pursued,
E. coli culture would be a useful additional control. It may be possible to make a better
assessment of promotion by assessing nucleation with emulsions having increasing
concentration of ice-nucleators; at some critical concentration, each micelle would include a
nucleator and the freezing temperature of the micelle population would be indicative of the
promotion activity of the substance. However, this would have to be done with purified
proteins in buffer; the heterogeneity of whole cell cultures, as well as the risk of cell lysis
upon sonication, might pose problems.
104
Table E.1 Bacterial concentrations at the time of experiments (CFU = colony forming unit per mL)
Species CFU
E. coli 3.4 ± 0.1 x 108
P. syringae 1.3 ± 0.4 x 109
P. borealis 1.07 ± 0.05 x 109
Chryseobacterium 2.3 ± 0.2 x 109
105
Table E.2. Measurements of Heterogeneous Nucleation by Differential Scanning Calorimetry
Sample # Exp.
Temperature of Observed
Heterogeneous Nucleation Events
[oC]
Average Subcooling [oC]
H2O (emulsion) 3 -21, -29 25 ± 6
H2O 4 -18.98, -19.82, -15.81, -19, -14, -15.67, -19.95
18 ± 2
P. syringae 3 -8.81, -10.45, -11.99 10 ± 2
P. borealis 3 -16.60, -16.92, -19.29 18 ± 1
Chryseobacterium 2 -19.99, -26.52 23 ± 5
Tryptic Soy Broth 2 -25.31, -26.10 26 ± 1
106
0.0
0.2
0.4
0.6
0.8
1.0
8 10 12 14 16 18 20 22
1 - n
Ci
0.0
0.2
0.4
0.6
0.8
1.0
12 14 16 18 20
i
0.0
0.2
0.4
0.6
0.8
1.0
8 10 12 14 16 18 20 22
1 - n
C
ii
0.0
0.2
0.4
0.6
0.8
1.0
12 14 16 18 20
ii
0.0
0.2
0.4
0.6
0.8
1.0
8 10 12 14 16 18 20 22
Δ T / oC
1 - n
C
iii
a
0.0
0.2
0.4
0.6
0.8
1.0
12 14 16 18 20Δ T / oC
iii
b
Figure E.1. Subcoolings of samples cooled at -0.2oC/min in (a) 1 mL bulk aliquots and (b) 10 µL microcapillaries. (i) Crystallisation (solid) and recrystallisation (hollow) with antifreeze culture, comparing samples with E. coli (■/□), and Chryseobacterium (▲/Δ). (ii) Cultures of ice nucleators, comparing: E. coli (■), P. borealis (▲), and P. syringae (Δ). (iii) Synthetic or processed additives, including: 0.5 mg/mL Snomax (■), distilled water (□), and 0.25 mM xanthan gum (▲).
107
02468
101214161820
3700 4200 4700 5200 5700 6200 6700 7200
time / s
tem
pera
ture
/ o C
THF HydrateMelting Point
* *
a
-3-2.5
-2-1.5
-1-0.5
00.5
11.5
2
7500 8000 8500 9000 9500 10000 10500 11000
time / s
tem
pera
ture
/ o C
*
*
b
-1
0
1
2
3
4
5
100 300 500 700 900 1100 1300time / min
tem
pera
ture
/ 'C
c
Figure E.2. Differential freezing (a) and melting (b) of a representative set of 1 mL samples during a subcooling experiment (see text). Shown are: P. syringae (solid line), E. coli (dotted line) and Chryseobacterium (grey line). Stars indicate a sample that froze and subsequently melted in two sections. (c) Samples are water (solid line) and 10 mg/mL PVCap (dashed line).
108
0
0.2
0.4
0.6
0.8
1
8 10 12 14 16 18 20 22
ΔT / oC
1 - n
C
Figure E.3. Subcoolings of THF/H2O solutions containing: no additive (■), a copper wire (Δ), a tungsten wire (●), and both types of wire (□). 1 mL samples were cooled at -0.2oC/min.
109
E.4 LITERATURE CITED
Heneghan, A. F., Wilson, P. W., Haymet, A. D. J.. 2002. Heterogeneous nucleation of supercooled water, and the effect of an added catalyst. Proceedings of the National Academy of Sciences 99: 9631-9634
Kim, H.K., Orser, C., Lindow, S.E., Sands, D.C.. 1987. Xanthomonas campestris pv. translucens strains active in ice nucleation. Plant Disease 71(11): 994-997
Wilson, S.L., Kelley, D.L., Walker, V.K.. 2006. Ice-active characteristics of soil bacteria selected by ice-affinity. Environmental Microbiology 8(10): 1816–1824
Young, D.W.. 1994. How to characterize the Effectiveness of Kinetic Hydrate Inhibitors. Annals of the New York Academy of Sciences 715: 341-343
110
A p p e n d i x F
MORPHOLOGY OF TETRAHYDROFURAN CRYSTALS IN
THE PRESENCE OF BIOLOGICAL INHIBITORS OF ICE
F.1 INTRODUCTION
Zeng et al. (2003, 2006) showed that antifreezes could elicit plate-like growth of normally
octahedral single THF hydrate crystals grown at 2.5oC. This was demonstrated for: 0.05 and
0.25 mM winter flounder AFP, 0.05 mM insect AFP (Choristoneura fumiferama,), and 0.05 mM
PVP.
F.2 METHODS
Experiments were carried out as previously described (Zeng et al., 2006) with the exception
that crystal nucleation occurred in the solution of interest, not a THF-H2O solution, due to
the difficulty of transferring nascent crystals. Solutions including E. coli, P. putida and
Chryseobacterium were prepared as for induction time experiments (Chapter 3) or with cultures
first diluted 1/10 in milliQ H2O. Also, 0.05 and 0.1 mM of recombinant green fluorescent
protein (GFP) tagged fish type III AFP (Macrozoarces americanus; Chapter 3) was tested.
Several hundred crystals were grown over the course of ~8 experiments.
F.3 RESULTS
Every sample tested exhibited octahedral crystal morphology where single or double crystals
were successfully grown (Figure F.1). Undiluted Chryseobacterium and P. putida samples
resulted in cloudy octahedra (Figure F.1b, c), implying that some bacteria or bacteria product
111
was incorporated into the growing crystal. UV pictures of a single crystal grow with GFP-
tagged AFP (Figure F.1h; see also Chapter 3) also indicates incorporation. Little evidence
was found of plate-like growth, however, nucleation often resulted in multicrystalline growth
which did sometimes resemble stacked plates (Figure F.1b, e, f).
F.4 DISCUSSION
Incorporation of material into octahedra suggests that molecules capable of binding
to hydrate were present. The lack of convincing evidence for plate-like growth in the
presence of hydrate-associating particles, shown to inhibit hydrate in other experiments
throughout this work, is interesting. It is possible that observation of such growth more
characteristic of experimental method than of hydrate-binding: plate-like growth of THF
hydrate has also been observed in repeat-freezing experiments without inhibitors (Knight
and Rider, 2002; Tomoyuki et al., 2001). A second possibility is that plate-like growth results
when AFP coverage of a crystal is incomplete; due to the high concentration of bacteria in
solutions tested here, this effect may have been suppressed. In any case, more experiments
must be completed to fully document the behaviour of these additives on crystal growth.
Additionally, closer visual analysis and molecular genetics techniques on crystal melts may
yield a better description of what substances may actually be incorporated.
112
a b c
d e f
g h
Figure F.1. Morphology of THF crystals in the presence of 109 CFU cultures of (a) E. coli, (b) P. putida, (c) Chryseobacterium, and (d-h) 0.05 mM type III AFP. Figures (d), (e-f), and (g-h) are of three different crystals; (e) is a side view of (f) while the crystal in (d) is next viewed under UV light (e). Note that multiple crystals are featured in (a), (b), (e/f). The pipette tip is ~0.5 mm wide.
113
F.5 LITERATURE CITED
Knight, C. A., Rider, K. 2002. Free-growth forms of tetrahydrofuran clathrate hydrate crystals from the melt: Plates and needles from a fast-growing vicinal cubic crystal. Philosophical Magazine A 82(8): 1609 -1632
Tomoyuki, I., Hideaki, M., Takaaki, M., Yasuhiko, H. M.. 2001. Formation and dissociation of clathrate hydrate in stoichiometric tetrahydrofuran-water mixture subjected to one-dimensional cooling or heating. Chemical Engineering Science 56(16): 4747-4758.
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J. A., 2006. Effect of antifreeze proteins on the formation and reformation of tetrahydrofuran clathrate hydrate. Journal of the American Chemical Society 128: 2844-2850
Zeng, H., Wilson, L.D., Walker, V.K., Ripmeester, J.A.. 2003. The inhibition of tetrahydrofuran clathrate hydrate formation with Antifreeze protein. Canadian Journal of Physics 81: 17-24.
