Bioremediation of creosote contaminated soil in both laboratory and field scale: Investigating the ability of methyl-β-cyclodextrin to enhance biostimulation

International Biodeterioration & Biodegradation 106 (2016) 117e126

Contents lists available at ScienceDirect

International Biodeterioration & Biodegradation journal homepage: www.elsevier.com/locate/ibiod

Bioremediation of creosote contaminated soil in both laboratory and field scale: Investigating the ability of methyl-b-cyclodextrin to enhance biostimulation €kela € a, Juha Mikola a, Hannu Silvennoinen b, Suvi Simpanen a, *, Riikka Ma a, c Martin Romantschuk a b c

University of Helsinki, Department of Environmental Sciences, Niemenkatu 73, 15140, Lahti, Finland Nordic Envicon Ltd, Huopalahdentie 24, 00350, Helsinki, Finland Kazan Federal University, Institute of Environmental Sciences, 420008, Kazan, Russia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 June 2015 Received in revised form 19 October 2015 Accepted 19 October 2015 Available online xxx

We investigated the bioremediation of 16 polycyclic aromatic hydrocarbons (PAH) in historically creosote contaminated soil using both laboratory and field scale experiments. We found that nutrient amendments and circulation of methyl-b-cyclodextrin (CD) solution enhanced soil microbial degradation capacity. In the laboratory experiment, the degradation of lower molecular weight, 2e3 ringed PAHs was achieved already by circulating nutrient solution and the use of CD mainly increased the desorption and removal of larger, 4e5 aromatic ringed PAH compounds. The 1% CD concentration was most feasible for bioremediation as most of the extracted PAH compounds were degraded. In the 5% CD treatment, the PAH desorption from soil was too fast compared to the degradation capacity and 25% of the total PAH amount remained in the circulated solution. Similar lab-scale results have been generated earlier, but very little has been done in full field scale, and none in freezing conditions. Although freezing stopped circulation and degradation completely during the winter, PAH degradation returned during the warm period in the field test. Circulation effectiveness was lower than in the laboratory but the improved nutrient and moisture content seemed to be the main reason for decreasing soil PAH concentrations. It also appeared that PAH extraction yield in chemical analysis was increased by the CD treatment in field conditions and the results of CD-treated and non-treated soil may therefore not be directly comparable. Therefore, a positive effect of CD on PAH degradation velocity could not be statistically confirmed in the field test. Based on our results, we recommend initiating the bioremediation of PAH contaminated soil by enhancing the microbial degradation with nutrient amendments. The CD seems to be useful only at the later stage when it increases the solubilisation of strongly absorbed contaminants to some extent. More investigation is also needed of the CD effect on the PAH yield in the chemical analysis. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Soil bioremediation Cyclodextrin Surfactant Biodegradation Biostimulation Polycyclic aromatic hydrocarbon

1. Introduction Creosote has been used for decades in wood impregnation processes to preserve and waterproof wooden structures like railway sleepers, telephone poles and bridge and pier deckings. The annual creosote production has been up to 100,000 tons in both Europe and in the USA, the main use being in the wood preservation industry (Melber et al., 2004), where the spills in wood

Abbreviations: PAH, polycyclic aromatic hydrocarbon; CD, cyclodextrin. * Corresponding author. E-mail address: suvi.simpanen@helsinki.fi (S. Simpanen). http://dx.doi.org/10.1016/j.ibiod.2015.10.013 0964-8305/© 2015 Elsevier Ltd. All rights reserved.

preserving plants or releases from treated wood products have contaminated the soil. The European Soil Data Centre estimated that wood and paper industry contributes to nearly 4% of soil contamination in Europe and that polycyclic aromatic hydrocarbons (PAHs) have a role in nearly 11% of all soil contamination (Panagos et al., 2013). In Finland, approximately 6% of all contaminated land sites are contaminated by creosote and most of these are in the premises of old sawmills (Suni et al., 2007). Coal tar creosote is a dark, oily liquid formed by fractional distillation of crude coal tars in 200e400
C. It consists of a complex mixture of several hundred chemicals, of which only 20% are present in amounts greater than 1%. The composition of creosote varies

118

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

depending on the origin of the coal and the distillation process, but six major compound classes can be sorted out: i.e. aromatic hydrocarbons, tar acids/phenolics, nitrogen-, sulphur- or oxygen containing heterocycles and aromatic amines (Melber et al., 2004). The main compound class of creosote is the PAHs, which can make up approximately 85% of creosote composition (Mueller et al., 1989). PAHs are hydrophobic organic compounds with a low aqueous solubility, high melting and boiling points and low vapour pressure (Zhang et al., 2006). They are composed of fused, aromatic rings, whose biochemical persistence arises from the dense clouds of pelectrons on both sides of the ring structures, making PAHs resistant to nucleophilic attack (Zhang et al., 2006). These properties promote PAH accumulation in the solid phases of the terrestrial environment and make them highly persistent in the soil (Zhang et al., 2006). PAH compounds are also resistant to the utilization by bacteria, which can degrade chemicals only when these are dissolved in water (Harms and Bosma, 1997; Johnsen et al., 2005). Despite this and although PAHs may also undergo adsorption, volatilization, photolysis and chemical degradation, microbial degradation is the major elimination process of PAHs in the soil (Koivula et al., 2004; Haritash and Kaushik, 2009). Many bacterial (Pseudomonas aeruginosa, P. fluorescens, Mycobacterium spp., Haemophilus spp.) and fungal (Pleurotus ostreatus) species can degrade PAHs in contaminated soils or sediments (Haritash and Kaushik, 2009), but the degradation rate in natural environmental conditions is extremely slow, if not completely lacking, due to the high recalcitrance of PAHs (Breedveld and Sparrevik, 2000). This is often due to the slow rate of contaminant desorption from soil particles, i.e. poor bioavailability, rather than the inability of degradation by microorganisms (Breedveld and Karlsen, 2000). In many European countries, specific guideline values have been set for the most relevant soil and groundwater contaminants. The values are based on the ecological or health risk, but these values vary in different countries (Carlon, 2007). According to the decree on the Assessment of Soil Contamination and Remediation Needs, published by the Finnish Government in 2007 (214/2007) the threshold value (i.e. assessment is needed when the value is exceeded) for the sum concentration of 16 PAHs in soil is 15 mg kg1 d.w., the lower guideline value (residential use of soil) 30 mg kg1 d.w. and the higher guideline value (industrial use of soil) 100 mg kg1 d.w. Guideline values have also been given for seven individual PAHs (anthracene, benzo(a)anthracene, benzo(a) pyrene, benzo(k)fluoranthene, phenanthrene, fluoranthene and naphthalene): for these, the threshold, lower, and higher guideline values are 1, 5 and 15 mg kg1 d.w., respectively, except for the high cancer risk benzo(a)pyrene, for which the threshold value is 0.2 and the lower guideline value 2 mg kg1 d.w. As the water solubility of contaminants is the rate-limiting factor for microbial degradation, additives have been used to enhance contaminant availability. Surfactants are surface tension reducing compounds, which increase the solubility and bioavailability of hydrophobic organic compounds by accumulating at the interface of two phases that have distinct polarity (Singh et al., 2007). In the bioremediation of contaminated soils, both synthetic and biosurfactants have been used. One group of the biosurfactants are cyclodextrins (CDs), which are oligosaccharides formed in the enzymatic degradation of starch by bacteria (Wang and Brusseau, 1995; Semple et al., 2007). The three most common cyclodextrins are a-, b-, and g-CD, which contain 6, 7, and 8 monomeric glucopyranose units, respectively (Ravelet et al., 2002). These compounds are torus-shaped molecules with a hydrophilic exterior, which provides them with high aqueous solubility and a hydrophobic cavity, into which organic contaminants can be encapsulated (Ravelet et al., 2002; Del Valle, 2004; Papadopoulos

et al., 2007). Other advantages of using CDs in bioremediation are their biodegradability, non-toxicity, and relatively stable physiochemical properties in a range of solution chemistry conditions (McCray and Brusseau, 1998; Ko et al., 1999). Of the three CD homologues, beCDs or their derivatives are the most studied in soil remediation because they are the cheapest option (Wang and Brusseau, 1993; Badr et al., 2004; Viglianti et al., 2006), they are suitable in size for the encapsulation of aromatic hydrocarbons, and their water solubility can be chemically enhanced (Petitgirard et al., 2009). However, the evidence of the remediation effect of b-CDs in PAH contaminated soils is mainly based on laboratory studies (Viglianti et al., 2006; Petitgirard et al., 2009; Zhang et al., 2012; Sun et al., 2013) and the implementation of the procedures in pilot and full scale experiments is lacking although these are essential to understand how remediation treatments can be up-scaled and how they function in realistic field conditions. Here we present results from a study, where we investigated whether biodegradation of PAHs in contaminated soil can be enhanced using methyl-b-cyclodextrin. Based on the literature we hypothesized that CD would have a postitive effect on PAH removal in lab conditions, but the aim was to test whether this effect could be replicated in field scale in cold conditions. We first tested the cleaning efficiency of varying CD-concentrations (0%, 1% and 5%) in creosote contaminated soil in the laboratory, and then based on the results of the laboratory test, we chose the most effective CD concentration (1%) for a field-scale bioremediation test of the same soil. In both tests, soil remediation was based on the degradation capacity of the indigenous microbial community, stimulated using nutrient amendments. 2. Materials and methods 2.1. Experimental soil The soil used in this study was collected from a former impregnation plant area in Mikkeli, Eastern Finland, where railway sleepers had been treated with creosote oil from the beginning of the 20th century until 1982 (Suni et al., 2007). The restoration was started in 2002 using mainly soil excavation and off-site disposal. The excavated soil was stored at the waste disposal site of €sairila Ltd in Mikkeli, where the field experiment of this study Metsa was implemented. The texture of the soil was sand with a total C content of 7.6 g kg1 soil (d.w.) and total N content less than 1 g kg1 soil (d.w.) (analysed in Ramboll Analytics, Lahti). The total PAH concentration was around 1000 mg kg1 soil (d.w.) with highest concentrations composed of 3-ringed fluoranthene and 4ringed pyrene (Table 1). 2.2. Chemicals used in the remediation treatments In the laboratory and field experiment, a 50% (w/v) aqueous stock solution of methyl-b-cyclodextrin (Cawasol W7 M TL, Wacker Chemie AG, Germany) was used as a surfactant. In the field test, 75% calcium peroxide (CaO2) (IXPER®75C, Solvay S.A, Belgium), consisting of 17% (by weight) molecular O2, was used as an oxygen source. CaO2 is a powdery chemical, which decomposes and releases oxygen in the presence of water €nen et al., 2012). The N (2CaO2 þ 2H2O  2Ca(OH)2 þ O2) (Nyka sources used in the field experiment were methylene urea (Yara Suomi Oy, Finland) and saltpetre (YaraBela Suomensalpietari Seþ, Yara Suomi Oy, Finland) which contains 14.5% NH4eN and 12.5% NO3eN. Using methylene urea in soil bioremediation is recommendable because it releases N slowly and the overdose of ammonia and excessively high pH values can be avoided (Peltola et al., 2006). Ammonium nitrate (NH4NO3) solution was used as a

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

119

Table 1 The initial concentrations of PAH compounds and their reduction in soil (calculated as a percentage loss of the initial concentration) in the different treatments of the laboratory and field experiment. In the field experiment, the P-value gives the statistical significance for the treatment effect (from MANOVA with the multivariate Pillai's trace P ¼ 0.232). Aro-matic rings

PAH compound

Laboratory experiment Initial concentration (mg kg1 d.w.)

2

Naphthalene Acenaphthylene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benz(a)anthracene Chrysene Benzo(b)&(k)fluoranthene Benzo(a)pyrene Indeno(1,2,3-cd)pyrene and Dibenz(a,h)anthracene Benzo(ghi)perylene P 16 PAH

3

4

5þ

0.7 6 102 50 20 36 440 239 68 57 48 14 0 0 1081

Field experiment Reduction % 1% CD

Nutrient þ water

33 70 99 98 94 90 88 76 88 72 48 31

31 69 99 97 91 84 84 76 75 46 26 15

83

78

N source in the laboratory experiment due to the easiness of supplying soil with solution.

2.3. Laboratory experiment In the laboratory experiment, polyacrylic plastic tubes (diameter 4.5 cm, length 30 cm) were first filled with 300 g of creosote contaminated soil (Fig. 1). The lower parts of the tubes were perforated to enable circulation of treatment solutions through the soil. Circulation was used to intensify the surfactant and nutrient treatment and to increase the oxygen concentration in the soil. Treatment solutions were made using a phosphate buffer (10 mM K2HPO4 þ 10 mM NaH2PO4) with NH4NO3 concentration of Silicone tubing

Cap with hole

Peristaltic multichannel pump

Soil (~300 g)

Silicone tubing Sampling hole

Initial concentration (mg kg1 d.w.) 0.9 4 17 14 8 39 455 283 73 77 97 33 18 7 1125

Reduction % 1% CD

Water

Control

P

46 8 95 89 75 71 75 66 75 53 39 37 32

27 9 95 52 16 6 79 62 72 45 39 38 33

23 14 93 74 67 46 29 17 40 20 40 41 36

0.775 0.895 0.015 0.205 0.220 0.221 < 0.001 < 0.001 < 0.001 < 0.001 0.996 0.919 0.932

30 66

30 62

34 30

0.902 <0.001

0.7 g N l1 and CD concentrations of 0, 1 and 5%. Each CD concentration had four replicates. After filling the tubes, the soils were rinsed with 500 ml of treatment solution to compact the material in the column. After this the outflow was prevented and the soil was saturated with 50 ml of fresh solution. The columns were left to settle for one day, rinsed and drained again with 500 ml of fresh solution and the soils and the treatment solutions were sampled. Drained soils were then left to stand without treatment for four weeks. During this preliminary treatment surfactant molecules were able to loosen the bindings between the contaminants and soil particles. After this period, 300 ml of solution was continuously circulated through the soil using a multichannel peristaltic pump (Ismatec ISM 404, Germany) with a flow rate of 20e60 ml min1. Treatment solutions were replaced first after two days and then weekly to keep the nutrient and surfactant concentrations stable during the experiment. For technical reasons, the continuous circulation was replaced by semi-continuous circulation after four weeks: here the treatment solutions were poured to tubes manually on average 12 times per week for two and half weeks and the solutions were replaced weekly. The semi-continuous circulation had the advantage of more closely resembling the field scale activities presented below. Circulated solutions were collected before each solution replacement and stored at 20
C. Solutions were combined for the PAH analysis so that samples from the periods 30e32 days, 40e52 days and 53e74 days were analyzed as three separate samples. Soil samples (8 g) were taken from the depth of 20 cm at time points of 1 d, 29 d (before the beginning of continuous circulation), 39 d, 52 d and 74 d (counted from the start of the surfactant treatment) and stored at 20
C until analyzed. The experiment was conducted at 18
C.

Mesh fabric

2.4. Field experiment

Cap with hole and stopper Circulated solution

Fig. 1. Graphic presentation of the column set-up in the surfactant concentration test.

Of the treatment solutions tested in the laboratory test, 1% CD solution was chosen for the field test. The experiment was set up on the premises of Mets€ asairila Ltd, the municipal waste disposal corporation of the Mikkeli region in eastern Finland. The experiment started in September 2009 and lasted for 13 months. Three soil piles were built on the management field of the area (Fig. 2): nutrients and liquids were added to the soil of piles 1 (170 m3) and 2 (60 m3), while the pile 3 (15 m3) was used as a control without

120

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

Fig. 2. Implementation of the field scale experiment.

any treatments. To have maximum benefit of the treatments the piles that were expected to have the most efficient biodegradation were the largest, and to reduce the costs of the study, the pile 2 (treated with CD and nutrients) was smaller than pile 1 (treated with nutrients only). The nutrient amendments used in powder form consisted of calcium peroxide (1 g kg1), methylene urea (0.28 g kg1) and salpetre (0.18 g kg1), which were mixed with the soil using a front loader. The piles 1 and 2 were built on plastic pools and perforated pipes were installed under the soil to enable solution circulation. Two to four m3 water (pile 1) and 1% CD solution (pile 2) was added to the piles once or twice each month during the unfrozen period. After the start in September, solutions were added only once before the freezing of piles in the autumn, but additions continued again in May in the following year. Effective circulation of solutions was not urgently needed because oxygen was supplied via the CaO2 amendment. Soil samples (200 g) were collected using an auger (diameter 5 cm) from each side of the pile (n ¼ 3e6) at the depth of 20 to 90 cm three times: in the beginning of the experiment in September 2009, after 7 months in April 2010 when the soil had thawed, and after 13 months in October 2010. To achieve comparable samples between the treatments, samples were always collected from the same depth under the surface and at the same height from the ground in proportion to the pile height in all piles. Water samples (0.5e1 L, n ¼ 3) were taken by pumping water out from beneath the piles from the perforated tubes. Before sampling the water was run for a while to get a representative sample. 2.5. Soil physical and chemical analysis The dry mass of soil samples was determined after drying 2e5 g subsamples at 105
C overnight and the organic matter content was

measured as loss on ignition (4 h in 550
C) (Kauppi et al., 2011). For pH measurements, the soil was shaken with 0.01 M CaCl2 for 30 min (250 rpm) and incubated at room temperature for 12 h. In the laboratory experiment, the quantity of suspended solids in solutions was determined by weighing 3 ml of solution before and after evaporation in room temperature. 2.6. Bacterial analysis Soil bacterial abundances were determined by plating according to Hernesmaa et al. (2005). In brief, soil samples of 0.5 g, representing a thoroughly mixed original sample, and solution samples of 1 ml were mixed with 0.9% NaCl solution and serial dilutions were plated on 1/5 tryptone-glucose-yeast extract agar plates (1/5 TGY) containing cycloheximide 50 mg ml1 (Laine et al., 1997). Plates were incubated for 3 d at 18
C. Longer incubation did not significantly increase the number of colonies. 2.7. PAH analysis Characterization of the creosote was performed by analyzing the concentration of 16 priority polycyclic aromatic hydrocarbons (PAHs) listed by EPA (United States Environmental Protection Agency): i.e. naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benzo(a) anthracene, chrysene, benzo(b)fluoranthene, benzo(k)- fluoranthene, benzo(a)pyrene, benzo(ghi)perylene, ideno(1,2,3-cd) pyrene and dibenzo(a,h)anthracene. The PAHs were extracted from the soil as described by Karstensen (1996) with a few exceptions. Briefly, 2 g of soil taken from the mixed primary sample was incubated with internal standard (PAH-Mix 31, Dr. Ehrenstorfer, Germany) at 4
C for 24 h. PAHs were then extracted using

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

acetonitrile-water mixture (1:1), 0.05 M Na4O7P2*10H2O and toluene. The samples were sonicated in an ultrasound bath for 30 min and shaken for 16 h at 300 rpm in an orbital shaker. The toluene fraction was collected and the extraction repeated using pure toluene and brief hand shaking. Toluene fractions containing the extracted PAHs were dried using anhydrous Na2SO4 columns. For PAH extraction, the whole samples of solutions were filtered through two glass microfiber filters, pore size 1.6 mm (grade GF/A, Whatman®, England) and pore size 0.7 mm (grade GF/F, Whatman®, England) to remove particles. Methanol containing internal standard (PAH-Mix 31, Dr. Ehrenstorfer, Germany) was added to the sample volume of 100e200 ml and PAHs were extracted using solid phase extraction disks (ENVI™ 18 DSK 47 mm, Supelco, USA) and dichloromethane (DCM) as the elution solvent. Samples were dried with anhydrous Na2SO4 columns. Deuterated anthracene D10 (Dr. Ehrenstorfer, Germany) was used as a recovery standard. The extracts were analyzed using a gas chromatograph (Shimadzu GC-17A) e mass spectrophotometer (Shimadzu GCMSQP5000) (GC-MS) and a ZB-5MS capillary GC column (length 30 m, internal diameter 0.25 mm and film thickness 0.25 mm). Sample injection volume was 1 ml and the injector temperature 280
C. Helium was used as the carrier gas with a flow rate of 22.7 ml min1. The oven program consisted of an isothermal step at 80
C for 1 min, followed by increases to 250
C at the rate of 10
C min1, to 280
C at the rate of 7
C min-1, and finally to 320
C at the rate of 20
C min1. The final temperature lasted for 10 min and the GC interface temperature was set to 280
C. A single ion monitoring (SIM) method was chosen in order to be able to detect the 16 PAH compounds of interest. 2.8. Statistical analysis Data analyses were performed using the SPSS 15.0 statistical package (SPSS Inc., Cary, NC, USA). In the laboratory experiment, a multivariate analysis of variance (MANOVA) was used to test whether the CD treatment had an effect on the final proportion of individual PAH compounds that remained in the soil, were extracted into soil solution or were biodegraded (each portion tested separately). The Pillai's Trace test was used to examine the general, multivariate treatment effect and the post-hoc comparisons of treatment levels were performed using the StudentNewman-Keuls test. The homogeneity of error variances was tested using Levene's test and if necessary, log or square root transformations were used to meet the requirements of the analysis (the homogeneity of variances could not be reached in the case of solution fluorene and total PAH concentrations and biodegraded acenaphthene and fluorene concentrations). The analysis of the field data (i.e. the comparison of the three piles with different treatments) followed the same procedure except that Tukey's HSD was used for the post-hoc tests. In the field data, the homogeneity of variances could not be reached for fluorene, phenanthrene and anthracene concentrations. The effect of the pile treatment on total PAH reduction rate was tested using analysis of covariance (ANCOVA). 3. Results 3.1. Laboratory experiment 3.1.1. PAH-analyses The aim of the laboratory experiment was to test how much the methyl-b-cyclodextrin (CD) can improve cleaning of creosote contaminated soil and to find the most feasible CD concentration. During the 74-day experiment, soil concentrations of PAH compounds decreased in all treatments, including the control (i.e. 0% CD

121

treatment), but the reduction was more pronounced in the treatments with CD (Fig. 3). At the end of the experiment, the sum concentration of all 16 PAHs had decreased by 78%, 83% and 87% in the control and 1% and 5% CD treatments, respectively, and the differences between the treatments in the proportion of compounds remaining in the soil were statistically significant (Table 2). The removal efficiency differed among the PAH-compounds, however in general, the proportion that was lost of the initial soil concentration was higher for smaller, 2e3 ringed PAHs than larger, 4e5 ringed PAHs (with the exception of 2-ringed naphthalene and acenaphthylene, whose initial concentrations were too low to be reliably analyzed) (Table 1). The loss was statistically significantly increased by adding CD except for the 2-ringed acenaphthene and the 4-ringed pyrene (Table 2). For most compounds, both CD concentrations accelerated the loss, but for 3-ringed fluoranthene and 5-ringed benzo(a)pyrene only the 5% CD showed a statistically significant effect (Table 2). When CD was added, part of the PAHs were extracted and circulated in the solution until the solution was replaced, but without CD this effect was negligible (Fig. 4). The extraction into the solution was greatest after the pretreatment in all treatments and decreased during the experiment (Fig. 3). For all PAHs, the extraction was highest in the 5% CD treatment and lowest in the control (Table 2 and Fig. 4). Biodegradation of lower molecular weight, 2e3 ringed PAHs was higher in the control and 1% CD treatment than in the 5% CD treatment (Table 2), whereas for higher molecular weight PAHs, the 4-ringed chrysene, benzo(b and k)fluoranthene and the 5-ringed benzo(a)pyrene, biodegradation was 10e25% higher in the CD treatments than in the control. When comparing the endpoint result of the treatments, the 1% CD appeared to be the most efficient for removing the PAHs as the combined amount of PAHs that remained in the soil and in the circulated solution was lowest for this treatment. 3.1.2. Bacterial analyses Measurements of soil bacterial abundances were started shortly before the beginning of solution circulation (day 29) and were then repeated every second week. The abundances varied during the experiment and were lowest in the first measurement (Fig. 5). At this time, the abundance in soil was eight times higher in the 5% CD treatment than in the control and 1% CD treatment indicating that bacteria had utilized the desorbed organic compounds during the preliminary treatment. The highest bacterial abundances were always found in the CD treated soils except at day 60, when the circulation had recently been changed to semi-continuous (Fig. 5). The circulating solution was changed weekly and the abundance of bacteria in solutions was measured at the end of three circulation periods: i.e. after 46e52 days, 60e67 days and 67e74 days. Highest bacterial abundances were always observed in the 5% CD treatment and this effect was clearest during the early, continuous circulation period (Fig. 5). The control and the 1% CD treatment did not differ from each other. 3.2. Field experiment The potential for large-scale biostimulation intensification by methyl-b-cyclodextrin was tested in creosote contaminated soil piles. The experiment was conducted in natural conditions and the piles were exposed to seasonal variation. As the addition and circulation of treatment solutions was relatively infrequent and technically challenging, nutrients were mixed with the soil before the beginning of circulation. The same solutions were recirculated throughout the experiment and water was added only when the initial volume reduced due to absorption to dry soil and evaporation. Based on the earlier results of the laboratory experiment,

122

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

Fig. 3. Total PAH concentration (mean ± SD) in a) the soil and b) the circulated solution of the laboratory experiment. In b) the grey columns represent the PAH concentrations in the solution after pretreatment (the continuous circulation was started on day 30 and the semi continuous on day 57).

Table 2 MANOVA tests (consisting of the multivariate Pillai's trace and the univariate tests) of the effect of the CD treatment (0% vs. 1% vs. 5% CD) on the proportion of PAH compounds remaining in the soil, extracted into the solution or biodegraded during the test. Treatment comparisons show the statistically significant differences among the means of the treatments. Aromatic rings

2

3

4

5

Pillai's trace Naphthalene Acenaphthylene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benz(a)anthracene Chrysene Benzo(b)&(k)fluoranthene Benzo(a)pyrene P 16 PAH

Remained in soil

Extracted into solution

Biodegraded

P-value

P-value

Treatment comparison

P-value

Treatment comparison

1, 5 > 0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0 5>1>0

0.011 0.676 0.004 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.005 0.013 < 0.001

0, 1 > 5 0>1>5 0>1>5 0, 1 > 5 0, 1 > 5 0, 1 > 5 0, 1 > 5 1>0>5 1>5>0 1, 5 > 0 5 > 0, 1 1>0>5

0.014 0.957 0.360 0.062 0.001 0.032 < 0.001 < 0.001 0.291 < 0.001 < 0.001 0.001 0.002 < 0.001

Treatment comparison

0 > 1, 5 0 > 1, 5 0>1>5 0, 1 > 5 0 > 1, 5 0 > 1, 5 0 > 1, 5 0, 1 > 5 0>1>5

< 0.001 0.012 < 0.001 < 0.001 < 0.001 < 0.001 <0.001 < 0.001 <0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

123

Fig. 4. The percentages of initial masses of different PAH compounds that remained in the soil (black bars), were extracted into the treatment solutions (white bars) or were biodegraded (grey bars) in the laboratory experiment. The biodegraded portion was estimated using the measured values of remaining and extracted portions.

a)

9

0 % CD

1 % CD

5 % CD

log CFU g-1dw

8

7

6

5

4 29

46

60

74

days

b)

9

0 % CD

1 % CD

5 % CD

log CFU ml-1

8

7

6

5

4 46 - 52

60-67

67-74

days Fig. 5. The colony forming units of bacteria (mean ± SD) in a) soil (log CFU g1 d.w.) and in b) circulated solutions (log CFU ml1) of the laboratory experiment (measurements were started in the beginning of continuous circulation).

124

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

water and 1% CD were used as treatments in the field application. The experiment was set up in the autumn 2009 and because of the early winter there was no decrease in soil PAH concentrations during the first seven months (Fig. 6). Instead, a 23% increase in analysis result was observed in the CD treated pile (Fig. 6). During the warm season from April to October 2010, the total PAH concentration decreased by 72%, 60% and 31% in the CD treated, water treated and control pile, respectively, using April 2010 values as a baseline. The April to October reduction rate differed statistically significantly between the control and the CD treated pile, but not between the water treated and the other two piles [the reduction rate slopes were 60 (95% CI from 103 to 17), -105 (95% CI from 147 to 63) and 166 (95% CI from 208 to 124) for the control, water treated and CD treated piles, respectively]. In the multivariate test, the PAH concentrations were not generally affected by the pile treatment (P ¼ 0.232 for Pillai's trace in MANOVA), but the univariate tests for single compounds suggested that the concentrations of 2-ringed acenaphthene, 3-ringed fluoranthene, and 4-ringed pyrene, benz(a)anthracene and chrysene decreased significantly more in the CD and water piles than in the control pile (Table 1). Because the circulated solutions were not changed during the experiment there was no need for regular solution analyses. After 8.5 months, the total PAH concentration was 0.3 ± 0.09 (mean ± SE, n ¼ 2e3) and 1.2 ± 0.4 mg L1 in the water and CD solution, respectively. In both treatments, the most common PAH-compound was the 3-ringed fluoranthene, which covered 35% of the total PAH concentration in the solutions. In the beginning of the experiment, soil pH was 6 but after the start of the remediation treatments, pH increased in both water and CD treatments to 7 and stayed at this value until the end of the experiment. The pH of the circulated solutions was 7. 4. Discussion A number of studies have shown the effect of CD for both prediction of PAH's biodegradability (Juhasz et al., 2005; Semple et al., 2007; Mahmoudi et al., 2013) and for improvement of bioavailability as an enhancer of bioremediation in laboratory scale (Wang et al., 1998; Zhang et al., 2012). The study presented here represents, however, the first test where laboratory simulations have been compared to a full field scale application in environmental conditions, including sub-zero winter conditions. We show that also in these conditions CD may be used to improve the degradation of particularly larger PAH compounds present in creosote contaminated soil.

Fig. 6. The total PAH concentration (mean ± SE) in the control, water treated and CD treated piles in the field experiment. Reduction rates were calculated for the warm period using the April and October values.

Balancing the C:N ratio of soil microbial resources using nutrient amendments enhances the microbial purification of soils contaminated with organic compounds (Guerin, 1999; Kauppi et al., 2011). We found this effect in our laboratory experiment, where soil PAH concentrations decreased significantly in all treatments, also without CD additions. It is also likely that the fast circulation of the solutions in our study washed the soil mechanically, provided better oxygen conditions for microbes, and enabled microbial movement in the soil matrix. It thus appears that the biostimulation as such, without additional solubility enhancement by cyclodextrin, was sufficient to remove a certain portion of PAH compounds. Across all treatments, the decrease in concentration was deepest for acenaphthene, fluorene and phenanthrene, implying that these compounds are especially efficiently desorbed from the soil and easily degraded. Estimation of the proportion of biodegraded PAHs was based on the assumption that the proportion of the initial mass, which was absent from the soil and solution, was biologically degraded. The experimental soil had been subjected to natural attenuation for over 20 years before our experiments and it is likely that the volatile, water soluble and easily decomposed PAH compounds were already absent. Substantial loss of PAHs during the experiment due to volatility, photolysis or chemical degradation was, therefore, unlikely. In our treatments without the CD, the loss of PAHs from the soil was almost completely due to microbial degradation, whereas in CD containing treatments part of the PAHs were extracted and remained in the circulated solution. This suggests that the circulation of CD solution loosened PAH compounds from soil particles faster than microbes were able to degrade them and the PAHs accumulated into the solution, especially in the 5% CD treatment. The efficiency of PAH extraction from soil has earlier been found to be proportional to the CD concentration (Viglianti et al., 2006), but when the soil cleaning is based on biodegradation, the removal of PAHs from soil is ultimately proportional to degradation capacity. In the present study, the solutions in the lab scale experiment were changed weekly to maintain constant CD concentration and optimal nutrient content for biodegradation. If the solutions had been changed less frequently, it is likely that the proportion of extracted compounds would have been lower because of longer exposure to microbial degradation. As we monitored bacterial abundances using a conventional enumeration method based on plating on general growth media, the results can only give a relative estimation of the differences between the treatments. Nevertheless, the higher bacterial abundances in the CD treated soil, and especially in 5% CD treatment, indicate that the cultivable bacteria had benefitted of the CD treatment, probably because of increased carbon source. On the other hand, due to their biodegradable nature, biosurfactants can also themselves serve as a microbial carbon source (Ławniczak et al., 2013), and although microbial growth has not been detected when CD is used as a sole carbon source (Wang et al., 1998; Bardi et al., 2000), microbial utilization of CD cannot be completely excluded in this study. It is likely that the structure of the PAH compound has an effect on how efficiently it can be degraded or removed from the soil by surfactants. The molecular stability of PAH compounds affects their degradation and depends on the alignment of the fused rings; the linear being the most unstable and the angular the most stable (Guerin, 1999). PAH compounds of higher molecular weight and many aromatic rings are more recalcitrant, hydrophobic and thus absorb more efficiently to soil particles (Gao et al., 2009). On the other hand, the structure of the surfactant has also an impact on the nchez-Trujillo et al. (2013) release of PAHs from soil particles. Sa compared the extraction efficiency of various CDs and found that the most effective PAH extraction from aged, contaminated soil was

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

achieved using chemically modified 2-hydroxypropyl-b-cyclodextrin (HP-b-CD) and randomly methylated-b-cyclodextrin (RAMEB), the latter being of the type of CD that we used in our work. It further appeared that especially 3-ringed PAHs were extracted efficiently because their structure had a good fit for the hydrophobic cavity of RAMEB and that RAMEB was most efficient in extracting higher molecular weight 4- and 5-ringed PAHs, except nchez-Trujillo et al., 2013). This is in agreement with for pyrene (Sa our laboratory test, where the proportion of soil remaining 4- and 5-ringed PAHs was lowest in the CD treated soil, except for pyrene, which was released from the soil to the same extent in all treatments. When investigating the use of biosurfactants in bioremediation, a general problem is that laboratory studies are not performed in a way that can be exactly reproduced in real clean-up cases or that studies are carried out using microbial monocultures (Ławniczak et al., 2013). In laboratory studies, conditions are also often optimal and the implementation of equal remediation in the field is challenging or impossible. In our study, we tested the chosen CD concentration in field scale with nutrient amendment and circulation of solutions implemented in a way that was possible in this scale. Although we did not find clear differences in soil concentrations between the water and CD treated piles at the end of the experiment, the differences in reduction rates during the warm period (April 2010eOctober 2010) indicate that the concentrations of PAH compounds decreased faster in the CD treated pile. This was partly because CD had released PAH compounds from the soil during the winter, when microbial activity was low, and these were then degraded by microbes during the warm period. The increased PAH extractability caused by CD adds a factor of uncertainty to the result of the last time point sampling e the real difference in PAH concentrations between the water and the CD treatment at the end of the experiment may be larger than the observed 10%, but this cannot be verified with the available data. In their investigation, Gao et al. (2009) sorted PAHs into three fractions - desorbing, nondesorbing and bound residue fractions - according to how easily they were desorbed from the soil in a sequential extraction mass balance approach. They found that CD extraction (a mild extraction technique) removed the most bioavailable, desorbing fraction, but not all of their PAH compounds were extractable even with solvent extraction in ultrasonic bath (Gao et al., 2009). Consequently, the extraction efficiency has an effect on the PAH yield from soil. Although the optimal nutrient level for biodegradation is the most important factor in bioremediation processes, we found that the use of CD enhanced the soil purification in both laboratory and field test especially improving the removal of PAH compounds with complex structures. We also proved that remediation procedures performed in a laboratory scale can be implemented as a field application, and thereby, that laboratory simulations may be used in predicting bioremediation performance in field conditions. In this case slowly releasing nutrients and infrequent recirculation of treatment solution are workable for enhancing the bioremediation of PAH-contaminated soil. Although the threshold values were not reached in the field test, it is apparent that considerable time savings can be achieved with biostimulation, and the positive effect can be further enhanced using CD as a biosurfactant. 5. Conclusions Our results show that in aged, contaminated soils, the indigenous microbes can degrade the most bioavailable and easily degradable PAH compounds also without added surfactants while the solubilisation effect of cyclodextrin mostly benefits the removal of larger PAH compounds. In the laboratory test, we found that the 1% CD solution was the most feasible for bioremediation as it

125

produced a good end-point concentration in soil and most of the extracted PAH compounds were not detected as solutes, but were apparently degraded. These findings suggest that the most costefficient and practical bioremediation can be achieved by optimising the nutrient and moisture content in the contaminated soil and the use of CD is reasonable only at later stages when the soilabsorbed, more recalcitrant contaminants remain in the soil. By implementing the nutrient-CD based treatment as a field application, we proved that this treatment combination is usable also in real clean-up cases as on site eremediation or even as in situ eapplication if the leakage of treatment solution can be avoided. Acknowledgements The study was funded by Tekes e Finnish Funding Acency for Technology and Innovation, by the University of Helsinki, the €ij€ Finnish Culture Foundation's Pa at-H€ ame Regional Fund, the Onni and Hilja Tuovinen Foundation and the Alfred Kordelin Foundation. Acknowledgements to Juha Valonen for his assistance in laboratory analysis in this study. References Badr, T., Hanna, K., de Brauer, C., 2004. Enhanced solubilization and removal of naphthalene and phenanthrene by cyclodextrins from two contaminated soils. J. Hazard. Mater. B112, 215e223. http://dx.doi.org/10.1016/ j.jhazmat.2004.04.017. Bardi, L., Mattei, A., Steffan, S., Marzona, M., 2000. Hydrocarbon degradation by a soil microbial population with b-cyclodextrin as surfactant to enhance bioavailability. Enzyme Microb. Tech. 27, 709e713. http://dx.doi.org/10.1016/ S0141-0229(00)00275-1. Breedveld, G.D., Karlsen, D.A., 2000. Estimating the availability of polycyclic aromatic hydrocarbons for bioremediation of creosote contaminated soils. Appl. Microbiol. Biotechnol. 54, 255e261. http://dx.doi.org/10.1007/s002530000362. Breedveld, G.D., Sparrevik, M., 2000. Nutrient-limited biodegradation of PAH in various soil strata at a creosote contaminated site. Biodegradation 11, 391e399. http://dx.doi.org/10.1023/A:1011695023196. Carlon, C. (Ed.), 2007. Derivation Methods of Soil Screening Values in Europe. A Review and Evaluation of National Procedures towards Harmonization. European Commission, Joint Research Centre, Ispra, p. 306. EUR 22805-EN. Del Valle, E.M.M., 2004. Cyclodextrins and their uses: a review. Process Biochem. 39, 1033e1046. http://dx.doi.org/10.1016/S0032-9592(03)00258-9. Gao, Y., Zeng, Y., Shen, Q., Ling, W., Han, J., 2009. Fractionation of polycyclic aromatic hydrocarbon residues in soils. J. Hazard. Mater. 172, 897e903. http://dx.doi.org/ 10.1016/j.jhazmat.2009.07.084. Guerin, T.F., 1999. Bioremediation of phenols and polycyclic aromatic hydrocarbons in creosote contaminated soil using ex-situ landtreatment. J. Hazard. Mater. 65, 305e315. http://dx.doi.org/10.1016/S0304-3894(99)00002-3. Haritash, A.K., Kaushik, C.P., 2009. Biodegradation aspects of polycyclic aromatic hydrocarbons (PAHs): a review. J. Hazard. Mater. 169, 1e15. http://dx.doi.org/ 10.1016/j.jhazmat.2009.03.137. Harms, H., Bosma, T.N.P., 1997. Mass transfer limitation of microbial growth and pollutant degradation. J. Ind. Microbiol. Biotechnol. 18, 97e105. http:// dx.doi.org/10.1038/sj.jim.2900259. €rklo € f, K., Kiikkila €, O., Fritze, H., Haahtela, K., Romantschuk, M., Hernesmaa, A., Bjo 2005. Structure and function of microbial communities in the rhizosphere of Scots pine after tree-felling. Soil Biol. Biochem. 37, 777e785. http://dx.doi.org/ 10.1016/j.soilbio.2004.10.010. Johnsen, A.R., Wick, L.Y., Harms, H., 2005. Principles of microbial PAH degradation in soil. Environ. Pollut. 133, 71e84. http://dx.doi.org/10.1016/j.envpol.2004.04.015. Juhasz, A.L., Waller, N., Stewart, R., 2005. Predicting the efficacy of polycyclic aromatic hydrocarbon bioremediation in creosote-contaminated soil using bioavailability assays. Bioremediat. J. 9, 99e114. http://dx.doi.org/10.1080/ 10889860500276524. Karstensen, K.H., 1996. Nordic Guidelines for Chemical Analysis of Contaminated Soil Samples. Nordtest Report NT Techn Report 329, Espoo, pp. 126e130. Kauppi, S., Sinkkonen, A., Romantschuk, M., 2011. Enhancing bioremediation of diesel-fuel-contaminated soil in a boreal climate: comparison of biostimulation and bioaugmentation. Int. Biodeter. Biodegr. 65, 359e368. http://dx.doi.org/ 10.1016/j.ibiod.2010.10.011. Ko, S.-O., Schlautman, M.A., Carraway, E.R., 1999. Partitioning of hydrophobic organic compounds to hydroxypropyl-b-cyclodextrin: experimental studies and model predictions for surfactant-enhanced remediation applications. Environ. Sci. Technol. 33, 2765e2770. http://dx.doi.org/10.1021/es9813360. Koivula, T.T., Salkinoja-Salonen, M., Peltola, R., Romantschuk, M., 2004. Pyrene degradation in forest humus microcosms with or without pine and its mycorrhizal fungus. J. Environ. Qual. 33, 45e53. http://dx.doi.org/10.2134/ jeq2004.4500.

126

S. Simpanen et al. / International Biodeterioration & Biodegradation 106 (2016) 117e126

Laine, M.M., Haario, H., Jørgensen, K.S., 1997. Microbial functional activity during composting of chlorophenol-contaminated sawmill soil. J. Microbiol. Meth. 30, 21e32. http://dx.doi.org/10.1016/S0167-7012(97)00040-7. Mahmoudi, N., Slater, G.F., Juhasz, A.L., 2013. Assessing limitations for PAH biodegradation in long-term contaminated soils using bioaccessibility assays. Water Air Soil Pollut. 224, 1411. http://dx.doi.org/10.1007/s11270-012-1411-2. McCray, J.E., Brusseau, M.L., 1998. Cyclodextrin-enhanced in situ flushing of multiple-component immiscible organic liquid contamination at the field scale: mass removal effectiveness. Environ. Sci. Technol. 32, 1285e1293. http:// dx.doi.org/10.1021/es970579þ. Melber, C., Kielhorn, J., Mangelsdorf, I., 2004. Coal Tar Creosote. Concise International Chemical Assessment Document 62. World Health Organization, Geneva. Mueller, J.G., Chapman, P.J., Pritchard, P.H., 1989. Creosote-contaminated sites. Their potential for bioremediation. Environ. Sci. Technol. 23, 1197e1201. http:// dx.doi.org/10.1021/es00068a003. €nen, A., Kontio, H., Klutas, O., Penttinen, O.-P., Kostia, S., Mikola, J., Nyka Romantschuk, M., 2012. Increasing lake water and sediment oxygen levels using slow release peroxide. Sci. Total Environ. 429, 317e324. http://dx.doi.org/ 10.1016/j.scitotenv.2012.04.044. Panagos, P., Van Liedekerke, M., Yigini, Y., Montanarella, L., 2013. Contaminated sites in Europe: review of the current situation based on data collected through a European network. Article ID 158764 J. Environ. Public Health 11. http://dx.doi. org/10.1155/2013/158764. Papadopoulos, A., Paton, G.I., Reid, B.J., Semple, K.T., 2007. Prediction of PAH biodegradation in field contaminated soils using a cyclodextrin extraction technique. J. Environ. Monit. 9, 516e522. http://dx.doi.org/10.1039/B700720E. Peltola, R., Salkinoja-Salonen, M., Pulkkinen, J., Koivunen, M., Turpeinen, A.-R., Aarnio, T., Romantschuk, M., 2006. Nitrification in polluted soil fertilized with fast- and slow-releasing nitrogen: a case study at a refinery landfarming site. Environ. Pollut. 143, 247e253. http://dx.doi.org/10.1016/j.envpol.2005.11.029. Petitgirard, A., Djehiche, M., Persello, J., Fievet, P., Fatin-Rouge, N., 2009. PAH contaminated soil remediation by reusing an aqueous solution of cyclodextrins. Chemosphere 75, 714e718. http://dx.doi.org/10.1016/j.chemosphere. 2009.01.072. Ravelet, C., Ravel, A., Grosset, C., Villet, A., Geze, A., Wouessidjewe, D., Peyrin, E., 2002. Stoichiometry and formation constants of six PAHs with g-cyclodextrin determined by HPLC using a cyano stationary phase. J. Liq. Chromatogr. R T 25, 421e432. http://dx.doi.org/10.1081/JLC-120008757. S anchez-Trujillo, M.A., Morillo, E., Villaverde, J., Lacorte, S., 2013. Comparative effects of several cyclodextrins on the extraction of PAHs from an aged

contaminated soil. Environ. Pollut. 178, 52e58. http://dx.doi.org/10.1016/ j.envpol.2013.02.029. Semple, K.T., Doick, K.J., Wick, L.Y., Harms, H., 2007. Microbial interactions with organic contaminants in soil: definitions, processes and measurement. Environ. Pollut. 150, 166e176. http://dx.doi.org/10.1016/j.envpol.2007.07.023. Singh, A., Van Hamme, J.D., Ward, O.P., 2007. Surfactants in microbiology and biotechnology: part 2. Application aspects. Biotechnol. Adv. 25, 99e121. http:// dx.doi.org/10.1016/j.biotechadv.2006.10.004. Sun, M., Luo, Y., Teng, Y., Jia, Z., Li, Z., Deng, S., 2013. Remediation of polycyclic aromatic hydrocarbon and metal-contaminated soil by successive methyl-bcyclodextrin-enhanced soil washing-microbial augmentation: a laboratory evaluation. Environ. Sci. Pollut. R 20, 976e986. http://dx.doi.org/10.1007/ s11356-012-1064-0. Suni, S., Malinen, E., Kosonen, J., Silvennoinen, H., Romantschuk, M., 2007. Electrokinetically enhanced bioremediation of creosote-contaminated soil: laboratory and field studies. J. Environ. Sci. Heal. A 42, 277e287. http://dx.doi.org/ 10.1080/10934520601134213. Viglianti, C., Hanna, K., de Brauer, C., Germain, P., 2006. Removal of polycyclic aromatic hydrocarbons from aged-contaminated soil using cyclodextrins: experimental study. Environ. Pollut. 140, 427e435. http://dx.doi.org/10.1016/ j.envpol.2005.08.002. Wang, X., Brusseau, M.L., 1993. Solubilization of some low-polarity organic compounds by hydroxypropyl-b-cyclodextrin. Environ. Sci. Technol. 27, 2821e2825. http://dx.doi.org/10.1021/es00049a023. Wang, X., Brusseau, M.L., 1995. Cyclopentanol-enhanced solubilization of polycyclic aromatic hydrocarbons by cyclodextrins. Environ. Sci. Technol. 29, 2346e2351. http://dx.doi.org/10.1021/es00009a029. Wang, J.-M., Marlowe, E.M., Miller-Maier, R.M., Brusseau, M.L., 1998. Cyclodextrinenhanced biodegradation of phenanthrene. Environ. Sci. Technol. 32, 1907e1912. http://dx.doi.org/10.1021/es980011g. Zhang, X.-X., Cheng, S.-P., Zhu, C.-J., Sun, S.-L., 2006. Microbial PAH degradation in soil. Degradation pathways and contributing factors. Pedosphere 16, 555e565. http://dx.doi.org/10.1016/S1002-0160(06)60088-X. Zhang, Z.-X., Zhu, Y.-X., Li, C.-M., Zhang, Y., 2012. Investigation into the causes for the changed biodegradation process of dissolved pyrene after addition of hyroxypropyl-b-cyclodextrin (HPCD). J. Hazard. Mater. 243, 139e145. http:// dx.doi.org/10.1016/j.jhazmat.2012.10.011. Ławniczak, Ł., Marecik, R., Chrzanowski, L., 2013. Contributions of biosurfactants to natural or induced bioremediation. Appl. Microbiol. Biot. 97, 2327e2339. http:// dx.doi.org/10.1007/s00253-013-4740-1.