Inhibition of Gas Hydrate Nucleation and Growth: Efficacy of anAntifreeze Protein from the Longhorn Beetle Rhagium mordaxChristine Malmos Perfeldt,† Pei Cheng Chua,‡ Nagu Daraboina,† Dennis Friis,§ Erlend Kristiansen,§

Hans Ramløv,§ John M. Woodley,# Malcolm A. Kelland,‡ and Nicolas von Solms*,†

†Center for Energy Resources Engineering, Department of Chemical and Biochemical Engineering, Technical University of Denmark,Søltoft Plads, Building 229, DK-2800 Kgs., Lyngby, Denmark#Center for Process Engineering and Technology (PROCESS), Department of Chemical and Biochemical Engineering, TechnicalUniversity of Denmark, Søltoft Plads, Building 229, DK-2800 Kgs., Lyngby, Denmark‡Department of Mathematics and Natural Science, Faculty of Science and Technology, University of Stavanger, N-4036 Stavanger,Norway§Department of Science, Systems and Models, Roskilde University, Universitetsvej 1, P.O. Box 260, DK-4000 Roskilde, Denmark

*S Supporting Information

ABSTRACT: Antifreeze proteins (AFPs) are characterized by their ability to protect organisms from subfreezing temperaturesby preventing tiny ice crystals in solution from growing as the solution is cooled below its freezing temperature. This inhibition ofice growth is called antifreeze activity, and in particular, certain insect AFPs show very high antifreeze activity. Recent studieshave shown AFPs to be promising candidates as green and environmentally benign inhibitors for gas hydrate formation. Here weshow that an insect antifreeze protein from the longhorn beetle, Rhagium mordax (RmAFP1), the most potent protein yet foundfor freezing inhibition, can inhibit methane hydrates as effectively as the synthetic polymeric inhibitor polyvinylpyrrolidone(PVP). In high pressure rocking cell experiments, onset hydrate nucleation temperatures and growth profiles showed repeatableresults. RmAFP1 clearly showed inhibition of hydrates compared to amino acids (L-valine and L-threonine) and the proteinbovine serum albumin (BSA). This indicates that proteins or amino acids do not generally inhibit hydrate formation. Thepromising performance of RmAFP1 as a new green kinetic hydrate inhibitor could further the development and increasedproduction of green hydrate inhibitors.

1. INTRODUCTION

Gas hydrates are crystalline, icelike solid compounds of smallgas molecules and water which form at low temperature andhigh pressure. In oil and gas production, formation of gashydrates can cause pipeline blockage with subsequent loss ofproduction and possible safety concerns.1,2 To prevent hydrateformation and ensure the flow of hydrocarbons (flowassurance), the energy industry generally uses so-calledthermodynamic inhibitors (THIs), such as methanol andglycols, in high dosage concentrations (20−50 wt %).2 Byadding THIs the hydrate phase boundary is shifted towardlower temperatures, meaning lower temperatures will berequired to form hydrates for a given operating pressure. Dueto the amounts of THIs needed, operational issues (e.g.,storage, transportation, and handling) and environmental issuesconnected to toxicity and flammability lead to high operating(OPEX) and capital expenditure (CAPEX).1,3

An alternative class of chemicals, kinetic hydrate inhibitors(KHIs), have been developed which can be applied in muchlower concentrations (0.1−1.0 wt %).1,4 Often KHIs are water-soluble polymeric compounds that prevent or delay hydrateformation. Despite the success of these compounds at very lowconcentrations, their poor biodegradability means that they arenot certified for use in all regions (for example the North Sea)which has prompted a search for new environmentally friendlyhydrate inhibitors.5,6 AFPs are a group of proteins that protect

organisms at subfreezing temperatures by preventing icecrystals from growing upon cooling below the freezing point.This prevention of ice growth is called antifreeze activity and isdefined as the difference between the melting temperature andthe freezing temperature of a solution. This thermal hysteresiscan readily be measured using a nanoliter osmometer. Theadsorption inhibition mechanism of antifreeze proteins is anoncolligative phenomenon where AFPs are irreversiblyadsorbed on the ice crystal.7 However, the exact mechanismis not well understood, and different explanations are discussedin the literature.8,9 Ice and gas hydrates have different althoughrelated structures, and it appears that the inhibition mechanismtoward ice and hydrate formation is different. In ice, AFPs allowthe formation of nuclei, to which they bind, after which growthis inhibited. In hydrates the nucleation itself is inhibited.10

AFPs are expressed in a variety of organisms, such as bacteria,fungi, plants, fish and insects, where the reported activities varyfrom about 0.3 °C in bacteria to 5−7 °C in insects.11,12 Recentstudies have shown that antifreeze proteins (AFPs) may havepotential as kinetic hydrate inhibitors while at the same timebeing environmentally benign.10,13−21 Zeng et al.17 found thatPleuronectes americanus (winter flounder) antifreeze protein

Received: February 7, 2014Revised: April 9, 2014

Article

pubs.acs.org/EF

© XXXX American Chemical Society A dx.doi.org/10.1021/ef500349w | Energy Fuels XXXX, XXX, XXX−XXX

increases the induction time of THF hydrate formation morethan PVP (available as a commercial inhibitor). Zeng et al.22

also found that Choristoneura fumiferana (an insect AFP) ismore effective than PVP in inhibiting THF hydrate formation.Gordienko et al.20 have observed that antifreeze proteinsmodify THF hydrate structure by adhering to the hydratesurface a way similar to PVP. Jensen et al.15 studied theinfluence of type III HPLC12 (originally identified in oceanpout) and ISP type III found in meal worm (Tenebrio molitor)on methane hydrate formation and reported that lowconcentrations of these proteins are more effective in reducinghydrate growth than PVP. Daraboina et al.6,13,14,21 used amultiscale approach to investigate the influence of two fishAFPs: type I AFP (AFP-I; average molecular weight, 3.3−4.5kDa) and type III AFP (AFP-III; Swiss Protein Databaseaccession number P19414; average molecular weight, 7 kDa)on methane/ethane/propane gas hydrate formation anddecomposition. They reported that these proteins showsignificant inhibition activity with less structural and composi-tional complexity during the dissociation process. Jensen etal.,16 also showed that type III AFP showed good hydrateinhibition capability compared to PVCap.From all these results it was evident that AFPs from different

species had the ability to inhibit gas hydrate formation.Furthermore, many studies were only indicative and qualitativeor comparative and were not performed at conditions close tothose encountered in natural gas production (such as studies onTHF hydrate). The search for active antifreeze proteinscontinues in order to develop strong environmentally benignalternatives to commercial kinetic inhibitors. Recently, thelonghorn beetle Rhagium mordax was shown to exhibit highantifreeze activity which during the cold season may be inexcess of 8 °C.23 In the AFPs of R. mordax (RmAFP1), severalrepeats of a motif rich in the amino acid threonine (TxTxTxTmotifs) is suspected to make up the ice-binding site.24

Antifreeze activity measured for mutations of the fourthreonine residues in fish AFP-I replaced with serine showedno antifreeze activity. However, replacement of the threoninehydroxyl-groups with methyl-groups (threonine to valine)reduced the antifreeze activity to 30−50% of the AFP-I.25−27

This indicates that the loss of the methyl group is moredetrimental to the antifreeze activity than the loss of thehydroxyl group. The ice crystal growth of valine mutations areobserved to cause slow crystal growth27 or bipyrimidal growth,whereas for serine mutations the crystals were growing freely.26

This indicates that the hydrophobic methyl group could beinvolved in water exclusion at the adsorption site of the protein.This means that the hydrophobic methyl group in threoninecould be important in the antifreeze inhibition mechanism ofice.9,25 However, the slow growth of valine mutations suggeststhat the hydroxyl groups of threonine residues in fish AFPs(AFP-I) could be essential to make the antifreeze protein−iceinteraction permanent.9 However, no data were available as tohow effective this hyperactive protein from longhorn beetle R.mordax is on the gas hydrate formation.Hydrophobic amino acids such as KHIs have also been

recently studied28 for CO2 gas hydrate inhibition. Glycineshowed KHI performance comparable to that of PVP in freshwater, and the increasing length of the amino acid alkyl sidechain showed decreasing inhibition. However, amino acids asKHIs for methane gas hydrate inhibition have not yet beeninvestigated.

Thermal stability of proteins is the equilibrium between thenative form and an ensemble of the nonnative form also calledthe unfolded form.29 For AFPs the thermal stability has beenobserved to vary for different organisms. Chao et al.30 reportedfor the type I antifreeze protein (AFP9) from winter flounder amelting temperature estimated to be 18 °C. Li et al.31 found forthe beetle, Dendroides canadensis, that the thermal stability wasapproximately 84 °C. The melting temperature of RmAFP1from R. mordax was recently determined by Friis et al.29 to be28.5 °C. However, thermal stability can be affected by thecomposition of the buffer.32 The thermal stability of RmAFP1might affect its performance as hydrate inhibitor. In order toavoid the influence of RmAFP1 unfolding during hydrateinhibition experiments the temperature range in this study waschosen to be well below the melting point of RmAFP1.Various techniques have been applied to investigate hydrate

nucleation in the presence of inhibitors. These include high-pressure differential scanning calorimetry,10,13,33−35 stirredtank,6 NMR,21,22 Raman14 and PXRD.36 In order to simulatemore realistic hydrate nucleation scenarios a so-called high-pressure rocking cell is increasingly being used by the oilindustry and gas hydrate research laboratories.37,38 The rockingof the cell and movement of the ball inside the cell creates shearforces and turbulence which better simulate conditions in thepipelines. In addition, hydrate nucleation is a stochastic processwhich leads to scattered results making evaluation of thestrength of these inhibitors difficult. To improve repeatability, asignificant number of experiments under the same experimentalconditions are preferred. Rocking cells with multiple units canmore rapidly provide statistically meaningful results forinhibition activity of hydrate formation37,38 and are thusincreasingly being adopted in the research community forevaluating kinetic inhibition activity. We used rocking cells withconstant cooling experiments, and these cells allowed us toreliably measure hydrate nucleation and growth data in thepresence of inhibitors. In this study we report repeatable resultsfor both the hydrate nucleation and growth showing that theAFP from R. mordax performs as effectively as PVP to inhibitmethane hydrate formation under realistic conditions.The goal of this study was to investigate how effective an

AFP from R. mordax is as a green KHI compared to acommercial synthetic inhibitor. Furthermore, the amino acidsessential to inhibit the ice growth and a nonantifreeze proteinwere studied in order to examine if the amino acids themselvesperform as inhibitors or if the structure of AFPs is unique forthis application.

2. EXPERIMENTAL SECTION2.1. Materials. In all the experiments methane gas and 50 mM

NaCl in Milli-Q water as aqueous phase was used, which will formstructure I hydrates. The methane gas was supplied from AGA (purity99.5%). The synthetic kinetic hydrate inhibitors were two poly-vinylpyrrolidones (PVP; MW 10,000 and MW 40,000) supplied fromSigma-Aldrich. The recombinant antifreeze protein (RmAFP1)originates from the bark beetle R. mordax. The genetic engineeringof the gene and the expression of the protein are described inKristiansen et al.24 Supernatant from cell lysate corresponding to 2 Loriginal cell culture (around 30 mL), was applied to a HisTrap FFcolumn with a column volume (CV) of 5 mL (GE Healthcare). Thecolumn was washed with 4 CV resuspension buffer (50 mM NaCl, 10mM imidazole, 25 mM Tris-HCl, pH 8), before eluting by applying agradient of imidazole 10−200 mM in 40 min at a flow rate of 2 mL/min. The fraction containing the RmAFP1 was dialyzed against 500volumes of deionized water for at least 24 h with change of water once

Energy & Fuels Article

dx.doi.org/10.1021/ef500349w | Energy Fuels XXXX, XXX, XXX−XXXB

or twice during this time. The dialyzed product was lyophilized andstored at −18 °C until use. Upon use, the lyophilized RmAFP1 wasdissolved in 50 mM NaCl in Milli-Q water. Concentration of theRmAFP1 was determined by amino acid analysis (AAA) (Departmentof Systems Biology, DTU, Denmark) to 2770 ppm. The RmAFP1 hasa molecular weight of 13.9 kDa. The two amino acids, L-threonine(Thr, purity ≥98%) and L-valine (Val, purity ≥98%) supplied fromSigma-Aldrich and a nonantifreeze protein bovine serum albumin(BSA; MW 68 kDa; purity ≥98%) supplied from Sigma-Aldrich, wereused as controls.2.2. Apparatus. The hydrate nucleation experiments were

conducted in a Rocking Cell produced by PSL Systemtechnik (seeFigure 1). The Rocking Cell has five test cells placed on a cell drive in

the cooling bath. Each cell has a working pressure up to 200 bar. Thetemperature is adjusted by the cooling bath in the range −20 °C to+60 °C. The detailed experimental setup and procedure is explainedelsewhere.37

Briefly, the apparatus consists of stainless steel test cells with avolume of 40 mL and has a stainless steel ball that freely rolls over theentire length of the cell when it is tilted. The cells can rock from anangle of −45° to +45°. When the cell is rocked, the ball inside the cellrolls along the length of the cell thereby mixing the contents. Themovement of the ball creates shear forces and turbulence in the cell,better reproducing conditions in the pipeline.2.3. Procedures. Measurements of the onset of hydrate nucleation

temperatures by constant cooling experiments were studied using arocking cell apparatus. The aim was to obtain the onset temperaturedifference in delay of hydrate nucleation for synthetic and biologicalKHIs compared to a noninhibited system in order to investigateeffectiveness of AFP as an inhibitor. The sample from the firstformation is then used for the second hydrate formation (memorysolution) in order to obtain more repeatable nucleation temperatures.The inhibitors were tested at a concentration of 2770 ppm and at 1385ppm.First the cell was loaded with 10 mL of test liquid. To eliminate air

in the cells, they were evacuated using a vacuum pump, pressurizingwith 2−3 bar of methane gas, rocking the cells, and finally using thevacuum pump again. The system without inhibitor was pressurized to95 bar at the initial experimental temperature of 20.5 °C, and theconstant cooling/heating cycle was from 20.5 to 2 °C to 20.5 °C. Thebaseline experiments without inhibitor were only performed with freshsolution. The KHI experiments were pressurized to 95 bar at the initialexperimental temperature of 13 °C and the constant cooling/heatingcycle was from 13 to 2 °C to 14 °C in order to maintain the stability ofthe protein while fully melting the hydrates in the cells. Theequilibrium temperature for the water−methane system at 95 bar wascalculated using CSMGem software to 12.5 °C. The equilibriumtemperature for the methane−water system at 95 bar was determinedby first forming hydrates and then heating very slowly (0.25 °C/h)from a temperature of 2 °C to a temperature about 2 °C higher thanthe expected equilibrium temperature. The intersection of the twocurves obtained from the cooling and heating steps gives thedissociation point. This is a standard laboratory technique.39 The

experiments conducted resulted in an equilibrium temperature of 12.9± 0.10 °C.

We used the WinRC software (PSL Systemtechnik) to program theexperiments and to log the temperature and pressure for each cell andthe temperature for the cooling bath throughout the experiment. Twocycles were set up for each experiment in order to perform fresh andmemory experiments. A typical pressure−temperature−time recordingfor one cell is shown in Figure 2. The cells were pressurized with

methane to 95 bar and set to rock at a rate of 20 rocks/min at an angleof 40° and to cool at a constant rate of 0.1 °C/min from 13 to 2 °C.During cooling the pressure decreases at a constant rate as thetemperature decreases until hydrate nucleation, where the pressuredrops faster due to hydrate formation. The temperature at whichhydrate nucleates is called the onset temperature, To (see Figure 2).

The cells were then heated from 2 to 14 °C at a rate of 0.5 °C/minin order to quickly dissociate the hydrates. By maintaining thetemperature at 14 °C for 30 min we could ensure full dissociation ofthe hydrates before repeating the memory cycle. Full dissociation ofhydrates was established, since the pressure reached the initial setpressure (before hydrate formation) and stabilized at a constanttemperature. Each inhibitor was tested at least 4 times.

3. RESULTS AND DISSCUSION3.1. Hydrate Nucleation. Previous studies showed that

AFPs do inhibit structure I hydrate nucleation and structure IIhydrate nucleation.6,10,13,14,16,21,22 The efficacy of AFPs indelaying hydrate nucleation has been shown to vary accordingto type, and no correlation between protein type and inhibitionefficacy has yet been found. Previous studies conducted at highpressure with methane gas are with AFPs sourced from plantsand fish. Hence, the performance of hyperactive insect AFPs onmethane hydrate nucleation at relevant pressures could lead tonew potential hydrate inhibitor candidates.The measured onset hydrate nucleation temperatures are

summarized in Table 1 and 2 for the fresh and memory tests.The relative standard deviation in the onset of hydratetemperature values were for fresh solution tests between 0.4−6.9% (SD 0.05−0.63 °C) and for the memory solution testsbetween 1.0−5.7% (SD 0.11−0.53 °C) indicating goodrepeatability of the tests. When using fresh solutions (Table1) of the amino acids L-valine and L-threonine, hydrate

Figure 1. Rocking cell (PSL Systemtechnik. (A) Rocking cell setupconsisting of the cooling bath with the five cells and the high pressurepanel. (B) One of the test cells.

Figure 2. Temperature and pressure recorded during the constantcooling test of PVP10 in 50 mM NaCl-Milli-Q. Two cycles arerecorded where the first one is the fresh solution test and the second isthe memory solution test. The red solid line indicates the pressuredecrease as the temperature decreases. The measured pressure willdeviate from the red solid line as hydrates nucleate. This is indicatedby a black arrow pointing at the onset temperature (To).

Energy & Fuels Article

dx.doi.org/10.1021/ef500349w | Energy Fuels XXXX, XXX, XXX−XXXC

nucleation temperatures were obtained at 11.5 and 11.3 °C,respectively. Comparison of these nucleation temperatures tothat of the noninhibited system (11.5 °C) shows that the aminoacids do not exhibit any KHI effect by themselves as they donot delay hydrate nucleation. Furthermore, addition of 2770ppm RmAFP1 in a fresh solution decreased the hydratenucleation temperature from 11.5 to 8.6 °C compared to thenoninhibited system by a depression in temperature ofapproximately 3 °C. This was more effective than theperformance of BSA (10.3 °C) and comparable to PVP10(9.0 °C) under similar experimental conditions. Whenrepeating the experiment using memory solution (Table 2),the average hydrate nucleation temperature of 2770 ppmRmAFP1 (9.6 °C) again showed better inhibition than BSA(10.4 °C). Experiments conducted at a lower concentration of1385 ppm RmAFP1 (10.5 °C) showed poorer inhibition thanPVP10 (9.2 °C) and PVP40 (9.5 °C). We suspect that aprotein concentration of 1385 ppm is too low to inhibithydrates as efficiently as PVP at the same concentrations.The main reason for the discrepancy in the number of tests

performed for the different substances was the limited amountof antifreeze protein at our disposal. Nevertheless, we believethe number of tests done for AFP with memory water at 2770ppm (8) were enough to conclude that it was comparable toPVP10 and better than the other substances tested. Similarly,the smaller number of amino acid tests performed werenevertheless enough to see that they were not good inhibitors.Since synthetic inhibitors such as PVP10 are the industrybenchmark, a larger number of tests were performed here.In Figure 3 the onset hydrate nucleation temperatures for

fresh and memory cycles are illustrated at 2770 ppm. On the

basis of these results BSA, PVP, and RmAFP1 perform betterthan the noninhibited system in delaying hydrate formation.The statistical differences between the KHIs are determined bycomparing the onset hydrate nucleation values usingindependent sample t tests with equal variances assumed. Forranking the performance of KHIs it is assumed that p valueslower than 0.05 confirm a statistically significant differencebetween two sets of onset hydrate nucleation values. On thebasis of these results, we find the ranking of the inhibitors to bethe same, regardless of the concentration (2770 or 1385 ppm)or the history of the water (memory or fresh). This indicatesthat both methodologies result in the same ranking. At the highconcentration (2770 ppm): RmAFP1 = PVP10 > BSA > L-threonine = L-valine = 50 mM NaCl in Milli-Q water > Milli-Qwater. It should be noted that only one RmAFP1 in freshsolution was performed at 2770 ppm due to limited amount ofRmAFP1. At the low concentration (1385 ppm) (SupportingInformation, Figure S1): PVP10 > PVP40 > RmAFP1 > 50mM NaCl in Milli-Q water > Milli-Q water.For the synthetic inhibitors and the RmAFP1 a shift to

higher values in the recorded onset hydrate nucleationtemperatures is observed when using memory rather thanfresh solutions. Using memory solutions clearly results inhigher temperatures which are expected as the memory effectmakes it easier to form hydrates. Similar observations have beenreported by others for AFPs:6,10 Daraboina et al.6,13,14,21

observed similar results for PVP and fish AFPs I and III forinhibition of natural gas hydrates. By using an isothermalnucleation procedure to obtain the nucleation time (time delayof hydrate formation) it was shown that PVP and AFP obtainedvery similar nucleation times using both a stirred tank, high-pressure microdifferential scanning calorimeter (HP μDSC),and magnetic resonance imaging (MRI).The RmAFP1 was also compared with two PVPs of different

average molecular weights at a lower concentration of 1385ppm (Tables 1 and 2). As expected, the lower-molecular weightPVP10 performs better than its higher-molecular weightcounterpart, PVP40. O’Reilly et al.40 explain the betternucleation inhibition of the lower-molecular weight polymerPVP by reasoning that this KHI has greater surface to volumeratio. Hence, the polymer perturbs the water structure morethan higher-molecular weight polymers which might in additionbe curled up or entangled through intramolecular hydrogen-

Table 1. Experimental Conditions and Average Results forFresh Solutions and Constant Cooling Tests for MethaneHydrate

solutionnumberof tests

concentration(ppm)

avg. onsettemp. (°C)

SD(°C)

RSD(%)

Milli-Q 4 − 11.8 0.19 1.650 mM NaCl−Milli-Q

4 − 11.5 0.05 0.4

L-valine 5 2770 11.5 0.05 0.5L-threonine 5 2770 11.3 0.31 2.3BSA 5 2770 10.3 0.08 0.8RmAFP1 1 2770 8.6a − −PVP10 14 1385 8.8 0.42 4.8

8 2770 9.0 0.63 6.9PVP40 6 1385 9.3 0.31 3.4aOnly one test was conducted due to limited amount of AFP.

Table 2. Experimental Conditions and Average Results forMemory Solutions and Constant Cooling Tests for MethaneHydrate

solutionnumber of

testsconcentration

(ppm)avg. onsettemp. (°C)

SD(°C)

RSD(%)

L-valine 5 2770 11.5 0.11 1.0L-threonine 5 2770 11.4 0.36 3.2BSA 21 2770 10.4 0.13 1.2RmAFP1 8 1385 10.5 0.23 2.2

7 2770 9.6 0.24 2.5PVP10 28 1385 9.2 0.26 2.8

20 2770 9.2 0.53 5.7PVP40 8 1385 9.5 0.30 3.1

Figure 3. Onset hydrate nucleation temperatures for the freshsolutions and memory solutions at a concentration of 2770 ppm (errorbars represent standard deviations (SD)).

Energy & Fuels Article

dx.doi.org/10.1021/ef500349w | Energy Fuels XXXX, XXX, XXX−XXXD

bonding, making it correspondingly less available to perturb thefree water structure. The molecular weight difference betweenthe proteins (RmAFP1 and BSA) might be thought to explainthe observed results for hydrate nucleation. BSA (having highmolecular weight) shows less inhibition than the low-molecularweight RmAFP1 as BSA might be entangled in the solution.However, for antifreeze proteins at equimolar concentrationsan increase in protein size decreases the solubility of the proteinin the water phase. This, in turn, will lead to increasedantifreeze activity.9 Most importantly though, AFPs have aunique structure consisting of a hydrophobic flat binding sitethat structurally matches the water molecules in ice41 whichother proteins such as BSA generally will not have.3.2. Hydrate Growth. Hydrate inhibition studies show that

AFPs inhibit hydrate growth when using plant, fish, and insectAFPs.10,13,15,21,42 It is reported that AFP may act as inhibitor orpromoter, depending on the temperature and pressureconditions during experiments.19 The efficacy of RmAFP1 onhydrate growth was studied and compared to that of the otherKHIs in order to evaluate its performance. The recordedhydrate growth profiles in the presence of RmAFP1 areillustrated in Figure 4. The figure shows that the recorded

profiles are very similar. Thus, not only the onset hydratenucleation temperature but also the hydrate growth can berepeated in the rocking cells.In order to compare the hydrate growth for the different

systems, hydrate growth profiles starting from the onset hydratenucleation temperature are normalized according to pressure.The profiles at a concentration of 2770 ppm for the freshsolutions are shown in Figure 5A and for the memory solutionsin Figure 5B. The measured pressure drops are indicative of theamount of hydrate formed during growth. In fresh solution thenoninhibited system and the control L-valine show a similargrowth profile and pressure drop of 10%, whereas L-threonine(11%) shows a slightly greater hydrate conversion.In fresh solution, RmAFP1 (11%) shows rapid initial growth

which is in contrast to the other KHIs. BSA (12%) showsslower hydrate growth rate but greater hydrate conversion.However, a larger pressure drop was observed in the presenceof PVP10 (14%), whereas the noninhibited system showed theleast growth (10%). This indicates that the KHIs promotehydrate growth although they significantly delay hydratenucleation. It is reported that once hydrates are formed inthe presence of inhibitor the hydrate growth can be promotedcompared to a noninhibited system.37,43 O’Reilly et al.40

observed that higher-molecular weight PVP prevents THFhydrate growth better than low-molecular weight PVP. This isexplained by the longer PVP chains more strongly adsorbingonto the hydrate crystal surfaces. In our study the growth ofhydrates using BSA is slower than AFP. This could be due tothe size of the BSA molecule. However, RmAFP1 shows slightlyless growth which indicates that the structure of the protein isimportant for growth inhibition.In the memory solutions the growth profiles of L-valine and

L-threonine appears to be similar. After 375 min of growth L-valine shows additional rapid growth resulting in a hydrateconversion of 11% for L-valine, whereas the hydrate conversionis 10% for L-threonine. The growth profile of RmAFP1 (8%)changes significantly by showing the least growth of hydrateswhich means RmAFP1 changes from promoting to inhibitingthe hydrate growth. Again, PVP10 (12%) shows significantgrowth, but BSA (13%) shows even greater growth although ata slower rate. Inhibition of methane hydrate growth was alsoreported16 when using similar methodology for the insect AFP,

Figure 4. Hydrate growth recorded for RmAFP1 during memorysolutions with cooling tests at a concentration of 2770 ppm. The graphshows the seven repeated experimental recordings.

Figure 5. Normalized hydrate growth profiles at a concentration of 2770 ppm for (A) fresh solutions and (B) memory solutions. The graph begins attime zero which is the onset of hydrate nucleation.

Energy & Fuels Article

dx.doi.org/10.1021/ef500349w | Energy Fuels XXXX, XXX, XXX−XXXE

Tenebrio molitor (TmISP), fish AFP (ISP type III HPLC12),and PVP. In that work, pressure was increased or decreased tomove into or out of the stable hydrate zone. Here we do so bychanging temperature. Different concentrations (0.01−0.1 wt%) were studied for the ISP type III HPLC12, and byincreasing concentration the amount of hydrate formeddecreased. We observed that the hydrate growth for RmAFP1at 1385 ppm initially was more rapid than at 2770 ppm.However, the pressure dropped by 8% at both concentrations,indicating a similar amount of hydrate growth (SupportingInformation Figure S2). This could be due to the highconcentration of RmAFP1, as Jensen et al.16 only observe a verysmall increase in inhibition of growth from the ISP type IIIHPLC12 at 0.05 to 0.1 wt %. Furthermore, they also observedthat 0.1 wt % PVP performed as hydrate growth inhibitor,whereas in our study we used the same molecular weight ofPVP and observed that 0.277 wt % PVP promoted the hydrategrowth. This again could be due to the higher concentration ofPVP. Kumar et al.44 reported that, for methane/propanehydrates in the presence of 1 wt % PVP, catastrophic hydrategrowth was observed compared to low concentration (0.1−0.5wt %) PVP. Conversely, Daraboina et al.45 observed that byincreasing concentration of PVP from 0.1 wt % to 1.0 wt %increasing hydrate growth inhibition was obtained.3.3. Stability of AFP. The onset hydrate nucleation

temperatures for RmAFP1 (2770 ppm) were observed tochange throughout the test period of 16 days (Figure 6). In

comparison BSA showed no change in hydrate nucleationtemperatures during the test time. The first data at day zero isfrom a constant cooling test using fresh solutions. The memoryRmAFP1 solution was then kept at 5 °C for a week beforefurther tests were conducted. The constant cooling resultsusing the memory solution indicate an increase in the averageonset temperature from 9.3 to 9.9 °C over 7.5 days. This couldsuggest an increased memory effect of the RmAFP1 or that theprotein may be decomposing with time. The BSA solution waskept at 8 °C between the tests and no change of performancewas observed.

■ CONCLUSIONThe efficacy of antifreeze protein from the longhorn beetle,Rhagium mordax (RmAFP1), for inhibiting methane hydratenucleation and growth was investigated using high-pressurerocking cells. The performance of the antifreeze protein wascompared to that of the synthetic inhibitor PVP, the proteinBSA, and the amino acids L-valine and L-threonine. Constantcooling experiments showed that the amino acids (L-Val and L-Thr) do not exhibit kinetic hydrate inhibition. However, usingfresh solutions showed that RmAFP1 (2770 ppm) performedas effectively as PVP for inhibition of hydrate nucleation. Inaddition the (non-antifreeze) protein BSA showed slightinhibition of hydrates. The efficacy of RmAFP1 in memoryexperiments was also comparable to PVP. The efficacy seems todecrease with time though, indicating possible decompositionof the protein.For hydrate growth the RmAFP1 changes from being

hydrate growth promoter in the fresh solution to becominghydrate growth inhibitor in the memory solution. In contrast,all other KHIs perform as hydrate promoters in both fresh andmemory solutions. By comparing onset nucleation temper-atures and growth profiles it is observed that there is no clearcorrelation between the efficacies of KHIs. However, RmAFP1in memory solution performs as an effective inhibitor forhydrate formation. These results confirm that AFPs such asRmAFP1 have a unique ability to inhibit hydrate formation andgrowth. Hyperactive insect AFPs such as RmAFP1 maytherefore represent a new class of green hydrates inhibitors.

■ ASSOCIATED CONTENT*S Supporting InformationThis material is available free of charge via the Internet athttp://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*Telephone: +45 4525 2867. Fax: +45 4588 2258. E-mail:nvs@kt.dtu.dk.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by Grant 10-082261 from TheDanish Council for Independent Research - Technology andProduction Sciences, the Danish National Advanced Technol-ogy Foundation, and the Technical University of Denmark

■ REFERENCES(1) Kelland, M. A. History of the Development of Low DosageHydrate Inhibitors. Energy Fuels 2006, 20 (3), 825−847.(2) Koh, C. A. Towards a Fundamental Understanding of NaturalGas Hydrates. Chem. Soc. Rev. 2002, 31 (3), 157−167.(3) Frostman, L.; Thieu, V.; Crosby, D.; Downs, H. Low-DosageHydrate Inhibitors (LDHIs): Reducing Costs in Existing Systems andDesigning for the Future. International Symposium on OilfieldChemistry; SPE, Houston, TX, 2003(4) Perrin, A.; Musa, O. M.; Steed, J. W. The Chemistry of LowDosage Clathrate Hydrate Inhibitors. Chem. Soc. Rev. 2013, 42 (5),1996−2015.(5) Del Villano, L.; Kommedal, R.; Kelland, M. A. Class of KineticHydrate Inhibitors with Good Biodegradability. Energy Fuels 2008, 22(5), 3143−3149.

Figure 6. Onset hydrate nucleation temperatures for RmAFP1 andBSA in fresh and memory solutions at a concentration of 2770 ppmover time.

Energy & Fuels Article

dx.doi.org/10.1021/ef500349w | Energy Fuels XXXX, XXX, XXX−XXXF

(6) Daraboina, N.; Linga, P.; Ripmeester, J.; Walker, V. K.; Englezos,P. Natural Gas Hydrate Formation and Decomposition in thePresence of Kinetic Inhibitors. 2. Stirred Reactor Experiments. EnergyFuels 2011, 25 (10), 4384−4391.(7) Raymond, J. A.; Devries, A. L. Adsorption Inhibition as aMechanism of Freezing Resistance in Polar Fishes. Proc. Natl. Acad.Sci. U.S.A. 1977, 74 (6), 2589−2593.(8) Wen, D. Y.; Laursen, R. A. A Model for Binding of an AntifreezePolypeptide to Ice. Biophys. J. 1992, 63 (6), 1659−1662.(9) Kristiansen, E.; Zachariassen, K. E. The Mechanism by WhichFish Antifreeze Proteins Cause Thermal Hysteresis. Cryobiology 2005,51 (3), 262−280.(10) Ohno, H.; Susilo, R.; Gordienko, R.; Ripmeester, J.; Walker, V.K. Interaction of Antifreeze Proteins with Hydrocarbon Hydrates.Chem.Eur. J. 2010, 16 (34), 10409−10417.(11) Duman, J. G.; Olsen, T. M. Thermal Hysteresis Protein Activityin Bacteria, Fungi, and Phylogenetically Diverse Plants. Cryobiology1993, 30 (3), 322−328.(12) Duman, J. G.; Bennett, V.; Sformo, T.; Hochstrasser, R.; Barnes,B. M. Antifreeze Proteins in Alaskan Insects and Spiders. J. InsectPhysiol 2004, 50 (4), 259−266.(13) Daraboina, N.; Ripmeester, J.; Walker, V. K.; Englezos, P.Natural Gas Hydrate Formation and Decomposition in the Presenceof Kinetic Inhibitors. 1. High Pressure Calorimetry. Energy Fuels 2011,25 (10), 4392−4397.(14) Daraboina, N.; Ripmeester, J.; Walker, V. K.; Englezos, P.Natural Gas Hydrate Formation and Decomposition in the Presenceof Kinetic Inhibitors. 3. Structural and Compositional Changes. EnergyFuels 2011, 25 (10), 4398−4404.(15) Jensen, L.; Ramlov, H.; Thomsen, K.; von Solms, N. Inhibitionof Methane Hydrate Formation by Ice-Structuring Proteins. Ind. Eng.Chem. Res. 2010, 49 (4), 1486−1492.(16) Jensen, L.; Thomsen, K.; von Solms, N. Inhibition of Structure Iand II Gas Hydrates using Synthetic and Biological Kinetic Inhibitors.Energy Fuels 2011, 25, 17−23.(17) Zeng, H.; Wilson, L. D.; Walker, V. K.; Ripmeester, J. A. TheInhibition of Tetrahydrofuran Clathrate-Hydrate Formation withAntifreeze Protein. Can. J. Phys. 2003, 81 (1−2), 17−24.(18) Zeng, H.; Wilson, L. D.; Walker, V. K.; Ripmeester, J. A. Effectof Antifreeze Proteins on the Nucleaton, Growth, and the MemoryEffect during Tetrahydrofuran Clathrate Hydrate Formation. J. Am.Chem. Soc. 2006, 128 (9), 2844−2850.(19) Al-Adel, S.; Dick, J. A. G.; El-Ghafari, R.; Servio, P. The Effect ofBiological and Polymeric Inhibitors on Methane Gas Hydrate GrowthKinetics. Fluid Phase Equilib. 2008, 267 (1), 92−98.(20) Gordienko, R.; Ohno, H.; Singh, V. K.; Jia, Z. C.; Ripmeester, J.A.; Walker, V. K., Towards a Green Hydrate Inhibitor: ImagingAntifreeze Proteins on Clathrates. PLoS One 2010, 5, (2), -.(21) Daraboina, N.; Moudrakovski, I.; Ripmeester, J. A.; Walker, V.K.; Englezos, P. Assesing the Performance of Commercial andBiological Gas Hydrate Inhibitors Using Nuclear Magnetic ResonanceMicroscopy and a Stirred Autoclave. Fuel 2013, 105, 630−635.(22) Zeng, H.; Moudrakovski, I. L.; Ripmeester, J. A.; Walker, V. K.Effect of Antifreeze Protein on Nucleation, Growth and Memory ofGas Hydrates. AIChE J. 2006, 52 (9), 3304−3309.(23) Wilkens, C.; Ramlov, H. Seasonal Variations in AntifreezeProtein Activity and Haemolymph Osmolality in Larvae of the BeetleRhagium mMordax (Coleoptera: Cerambycidae). Cryoletters 2008, 29(4), 293−300.(24) Kristiansen, E.; Wilkens, C.; Vincents, B.; Friis, D.; Lorentzen,A. B.; Jenssen, H.; Lobner-Olesen, A.; Ramlov, H. HyperactiveAntifreeze Proteins from Longhorn Beetles: Some Structural Insights.J. Insect Physiol 2012, 58 (11), 1502−1510.(25) Haymet, A. D. J.; Ward, L. G.; Harding, M. M. Winter Flounder“Antifreeze” Proteins: Synthesis and Ice Growth Inhibition ofAnalogues That Probe the Relative Importance of Hydrophobic andHydrogen-Bonding Interactions. J. Am. Chem. Soc. 1999, 121 (5),941−948.

(26) Haymet, A. D. J.; Ward, L. G.; Harding, M. M.; Knight, C. A.Valine Substituted Winter Flounder ‘Antifreeze’: Preservation of IceGrowth Hysteresis. Febs Letters 1998, 430 (3), 301−306.(27) Zhang, W.; Laursen, R. A. Structure-Function Relationships in aType I Antifreeze Polypeptide - The Role of Threonine Methyl andHydroxyl Groups in Antifreeze Activity. J. Biol. Chem. 1998, 273 (52),34806−34812.(28) Sa, J. H.; Kwak, G. H.; Lee, B. R.; Park, D. H.; Han, K.; Lee, K.H. Hydrophobic Amino Acids As a New Class of Kinetic Inhibitors forGas Hydrate Formation. Sci. Rep.-Uk. 2013, 3.(29) Friis, D. S.; Johnsen, J. L.; Kristiansen, E.; Westh, P.; Ramlov, H.Low Thermodynamic- but High Kinetic Stability of an AntifreezeProtein from Rhagium mordax. Protein Sci. 2014, DOI: 10.1002/pro.2459.(30) Chao, H.; Hodges, R. S.; Kay, C. M.; Gauthier, S. Y.; Davies, P.L. A Natural Variant of Type I Antifreeze Protein with Four Ice-Binding Repeats Is a Particularly Potent Antifreeze. Protein Sci. 1996, 5(6), 1150−1156.(31) Li, N.; Andorfer, C. A.; Duman, J. G. Enhancement of InsectAntifreeze Protein Activity by Solutes of Low Molecular Mass. J. Exp.Biol. 1998, 201 (15), 2243−2251.(32) Salvay, A. G.; Santos, J.; Howard, E. I. Electro-Optical PropertiesCharacterization of Fish Type III Antifreeze Protein. J. Biol. Phys.2007, 33 (5−6), 389−397.(33) Daraboina, N.; Malmos, C.; von Solms, N. Investigation ofKinetic Hydrate Inhibition Using a High Pressure Micro DifferentialScanning Calorimeter. Energy Fuels 2013, 27 (10), 5779−5786.(34) Sharifi, H.; Ripmeester, J.; Walker, V. K.; Englezos, P. KineticInhibition of Natural Gas Hydrates in Saline Solutions and Heptane.Fuel 2014, 117, 109−117.(35) Lachance, J. W.; Sloan, E. D.; Koh, C. A. Determining GasHydrate Kinetic Inhibitor Effectiveness Using Emulsions. Chem. Eng.Sci. 2009, 64 (1), 180−184.(36) Ohno, H.; Moudrakovski, I.; Gordienko, R.; Ripmeester, J.;Walker, V. K. Structures of Hydrocarbon Hydrates during Formationwith and without Inhibitors. J. Phys. Chem. A 2012, 116 (5), 1337−1343.(37) Daraboina, N.; Malmos, C.; von Solms, N. Synergistic KineticInhibition of Natural Gas Hydrate Formation. Fuel 2013, 108, 749−757.(38) Chua, P. C.; Kelland, M. A. Poly(N-vinyl azacyclooctanone): AMore Powerful Structure II Kinetic Hydrate Inhibitor than Poly(N-vinyl caprolactam). Energy Fuels 2012, 26 (7), 4481−4485.(39) Tohidi, B.; Burgass, R. W.; Danesh, A.; Ostergaard, K. K.; Todd,A. C. Improving the Accuracy of Gas Hydrate Dissociation PointMeasurements. Ann. NY Acad. Sci. 2000, 912, 924−931.(40) O’Reilly, R.; Ieong, N. S.; Chua, P. C.; Kelland, M. A. CrystalGrowth Inhibition of Tetrahydrofuran Hydrate with Poly(N-vinylpiperidone) and other Poly(N-vinyl lactam) Homopolymers. Chem.Eng. Sci. 2011, 66 (24), 6555−6560.(41) Kristiansen, E.; Ramlov, H.; Hojrup, P.; Pedersen, S. A.; Hagen,L.; Zachariassen, K. E. Structural Characteristics of a Novel AntifreezeProtein from the Longhorn Beetle Rhagium inquisitor. Insect Biochem.Mol. 2011, 41 (2), 109−117.(42) Bruusgaard, H.; Lessard, L. D.; Servio, P. Morphology Study ofStructure I Methane Hydrate Formation and Decomposition of WaterDroplets in the Presence of Biological and Polymeric KineticInhibitors. Cryst. Growth Des 2009, 9 (7), 3014−3023.(43) Lederhos, J. P.; Long, J. P.; Sum, A.; Christiansen, R. L.; Sloan,E. D. Effective Kinetic Inhibitors for Natural Gas Hydrates. Chem. Eng.Sci. 1996, 51 (8), 1221−1229.(44) Kumar, R.; Lee, J. D.; Song, M.; Englezos, P. Kinetic InhibitorEffects on Methane/Propane Clathrate Hydrate-Crystal Growth at theGas/Water and Water/n-Heptane Interfaces. J. Cryst. Growth 2008,310 (6), 1154−1166.(45) Daraboina, N.; Linga, P. Experimental Investigation of the Effectof Poly-N-vinyl pyrrolidone (PVP) on Methane/Propane ClathratesUsing a New Contact Mode. Chem. Eng. Sci. 2013, 93, 387−394.

Energy & Fuels Article

dx.doi.org/10.1021/ef500349w | Energy Fuels XXXX, XXX, XXX−XXXG