• atmospheric pollution;
  • biological gas treatment;
  • extremophiles;
  • microbial community dynamics;
  • mycoremediation


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The treatment of air contaminated with toluene, ethylbenzene, and p-xylene was assayed in three laboratory-scale biofilters, each consisting of two modules connected in series, packed with a pelletized organic fertilizer and inoculated with a toluene-degrading liquid enrichment culture. Biofilters were operated in parallel for 185 days in which the volumetric organic loading rate was progressively increased. The operation regime was subjected to drying out, so that packing humidity generally remained below 40%. Significant process failure occurred with ethylbenzene and p-xylene, but the toluene biofilter comparatively sustained a significant elimination capacity. Microbial community characterization by quantitative PCR and denaturing gradient gel electrophoresis showed substantial fungal enrichment in the toluene biofilter. Ribotypes identical to the well-known toluene-degrading black yeast Exophiala oligosperma (Chaetotyriales) were found among the dominant species. The microbial community structure was similar in the biofilters loaded with toluene and ethylbenzene but with p-xylene was quite specific and encompassed other chaetothyrialean fungi. Several species of Actinomycetales were found in the packing while the inoculum was dominated by representatives of the Burkholderiales and Xanthomonadales. One single fungal ribotype homologous to Acremonium kiliense was detected in the inoculum. The implications of xerophilic biofilter operation on process biosafety and efficiency are discussed.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

An air biofilter is the simplest configuration of a vapor-phase bioreactor used for the treatment of air polluted with volatile organic compounds. The contaminated gas stream is forced through a column packed with a porous support material, where the volatile contaminants are ad/absorbed and metabolized by a microbial biofilm that develops on the packing. A conventional biofilter is characterized by the absence of a free liquid phase, and nutrients needed for microbial growth are supplied either by the packing itself, when organic materials such as compost or peat are used, or are supplemented through discontinuous irrigation of the bed in case of inert or synthetic materials. Humidification of the biofilter bed is therefore a critical control parameter to keep the availability of water for microbial growth at an acceptable level (Bohn & Bohn, 1999). Yet, excess of water has been reported as detrimental for the elimination of relatively insoluble substrates because of the mass transfer limitation phenomena and packing deterioration (Kennes & Veiga, 2004).

Air biofiltration was initially developed for the removal of odorous compounds, but its use has progressively been extended to the biodegradation of hazardous volatile pollutants such as benzene, toluene, ethylbenzene, and xylene (Kennes et al., 2009) – monoaromatic hydrocarbons known collectively as BTEX. Significant atmospheric BTEX pollution arises from petrochemical and chemical industries, particularly where containment and/or gas treatment is inadequate (Swoboda-Colberg, 1995). The biological treatment of gas streams containing relatively low concentrations of BTEX has been found to be more economical and environmentally friendly than other physicochemical alternatives such as adsorption, condensation, and incineration, which require more input (i.e. materials, reagents, and energy) and produce toxic wastes (van Groenestijn & Hesselink, 1993). As reviewed recently (Kennes & Veiga, 2004), an increasing number of studies have pointed to the fact that the presence of fungi in the biofiltration of specific pollutants offers some advantages with respect to the stability and biodegradation activity. Fungi are generally better suited than bacteria to grow on a solid support, to deal with water and nutrient scarcity and to tolerate low pH conditions. Hence, fungi have primarily been used in combination with inert packing media (e.g. perlite, polyurethane foam, vermiculite, etc.) for an extended biofilter operation, and sustained elimination capacity (EC) values ranging from 20 to 100 g m−3 h−1 have generally been obtained (Cox et al., 1996; García-Peña et al., 2001; Kennes & Veiga, 2004; Prenafeta-Boldú et al., 2008).

Research on the microbial ecology of biofilters related to the biofiltration of BTEX-polluted gases has fundamentally been focused on population dynamics and composition of heterotrophic bacteria, particularly when biofilters were operated under optimal or suboptimal conditions (Cabrol et al., 2012). Extensive reviews have also been published on the aerobic metabolism of aromatic hydrocarbons by bacteria (Gibson & S Harwood, 2002; O'Leary et al., 2002), but studies on the fungal counterparts are comparatively scarce (Prenafeta-Boldú et al., 2006). Interactions between fungi and bacteria in gas biofilters, and the consequences on the overall bioreactor performance, remain a controversial issue. Fungal dominance has been claimed both as beneficial and detrimental to the feasibility of biofiltration. Besides the already mentioned advantages of fungi to withstand growth-limiting conditions, the development of fungal biomass in biofilters has also been related to the presence of potential pathogens (Prenafeta-Boldú et al., 2006), the frequent incidence of clogging due to the filamentous biomass (Aizpuru et al., 2005), and the generally lower metabolic rates when compared to most of aerobic bacteria (Prenafeta-Boldú et al., 2001). While the presence of an adapted microbial community is essential for the stable operation of a biofilter, relatively little research has been carried out in characterizing the structure and dynamics of the underlying microbial populations. A biofilter is an ecosystem in which diverse organisms are subjected to several environmental stresses, including changes in water content, pH, temperature, and exposure to toxic chemicals. Moreover, the system is likely to be heterogeneous, and important environmental gradients might occur inside the bioreactor in time and/or space. Altogether, biofilter operation and performance may be correlated directly with the microbial community dynamics. A biofilter is analogous to a soil bed, and molecular techniques aimed to microbial community analysis in soil would seem directly applicable to biofiltration research.

In this study, the microbial communities of bacteria and fungi in samples taken from different laboratory-scale biofilters were studied by culture-independent molecular methods to gain new insights into the relationship between microbial community structure and biofilter operational parameters. The biofilters were used for the treatment of gas streams containing toluene, ethylbenzene, and p-xylene. Operation was performed at relatively low water content to study the effect of bed drying out on biofilter performance and on the microbial populations from the bioreactor bed. Microbial community shifts were characterized by comparing the dominance and composition of fungal and bacterial genotypes enriched in the biofilters, in relation to those initially present in the original inoculum and organic packing.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Biofilter set-up and operation

Gas biofiltration experiments were performed in three identical columns, fed respectively with toluene, ethylbenzene, and p-xylene. Each biofilter consisted of two interchangeable cylindrical modules mounted on top of each other, named M1 and M2, respectively, accordingly to their initial vertical position. Each module was made of PVC (10 cm diameter × 33 cm length) and had a hollowed bottom plate to sustain the packing. Biofilter modules were packed with a pelletized mixture of animal manure and sawdust, commercialized as a plant fertilizer under the name of AbonlirTM (SLIR S. L., Carcastillo, Spain). Pellets were previously sieved to homogenize the particle size and characterized in physicochemical terms (Table 1). The packing was then soaked with a toluene-degrading liquid enrichment culture prepared as described elsewhere (Elías et al., 2010), and about 1.6 dm3 of the material was loaded into each biofilter module. Subsequent gas feeding in biofiltration experiments was applied in a down-flow mode, and the polluted air was generated by mixing two compressed air streams that were regulated with rotameter assemblies. The main stream consisted of air bubbled through a distilled water column for humidification, while a secondary air stream was saturated with the selected volatile substrate in a gas washing bottle. The relative humidity of the contaminated air at the biofilter inlet remained always higher than 98%. Biofiltration experiments lasted for 185 days during which step changes were applied to the volumetric organic loading rate (OLR), either by increasing the substrate concentrations in the influent air, the fed gas flow rate, or by removing one of the biofilter modules. The effect of drought was studied by applying sporadic irrigations and interchanging the position of the two biofilter modules. The water content in each packed module was monitored gravimetrically upon drying a sample of support material (2–3 g) at 105 °C. An equivalent mass of the withdrawn filter media was replaced with new packing material. All biofiltration experiments were performed at room temperature (approximately 25 °C).

Table 1. Physicochemical characteristics of the bed packing material used in biofiltration experiments
  1. a

    Average and standard deviation values.

  2. b

    Value provided by the manufacturer.

Physical parameter
Particle shapeCylindrical
Bed apparent density (g mL−1)0.98 ± 0.07
Pellet apparent density (g mL−1)1.29 ± 0.08
Material real density (g mL−1)2.72 ± 0.21
BET superficial area (m2 g−1)12.06 ± 0.09
Langmuir superficial area (m2 g−1)17.41 ± 0.74
Chemical composition
Total organic matter (%)b72
Labile organic matter (%)40
Total carbon content (%)32
Total nitrogen content (%)2
Total hydrogen content (%)3
Total phosphorous content (%)0.2
Total sulfur content (%)3
Elemental sulfur content (%)< 0.1
Sulfate content< 0.1
Water content (%)23.2

Analytical and microscopy methods

In relation to the physicochemical characterization of the packing material, the pellets were previously sieved so that only those with a diameter between 6.3 and 8 mm were used as packing material. ASTM sieves were used for the particle size distribution. The BET surface area and external surface area were estimated by means of the nitrogen adsorption technique with an ASAP 2010 V5.02H analyzer (Micromeritics, Georgia). The original moisture content was determined in a Leco TGA 500 thermobalance (LECO Corporation, MI). The pH value was measured in a 1 : 9 dry material weight to water volume ratio. Total C, H, and N contents were measured in an EAI Exeter Analytical analyzer (Exeter Analytical Inc., MA), and the S total content was determined in a Leco SC 132 analyzer (LECO Corporation). The content of elemental sulfur was determined by dissolving the elemental sulfur component of the sample into carbon sulfide.

The EC and efficiency in each biofilter were determined by measuring inlet and outlet concentrations of the applied volatile organic compounds in a gas chromatograph (microGC CP 4900) equipped with a CP-Sil 5 CB (6m × 0.15mm × 2μm) column and a TCD detector. The column and injector temperatures were set at 80 °C and 100 °C, respectively. For scanning electron microscopy (SEM) imaging, samples of the packing were fixed in 2% glutaraldehyde in 0.1 M cacodylate buffer (pH = 7.4), washed in iso-osmolar cacodylate/sucrose buffer, and postfixed in 1% osmium tetroxide in cacodylate buffer. Samples were then dehydrated through an ethanol series and washed in hexamethyldisilazane prior to air-drying. Finally, samples were mounted onto stubs and gold-coated using a JEOL fine-coat ion sputter JFC-1100. Samples were visualized and micrographed using a scanning electron microscope (Hitachi S-4800) at 15 kV accelerating voltage.

Denaturing gradient gel electrophoresis (DGGE) molecular profiling

Bed samples were withdrawn at the end of experiments from each biofilter bottom module (M2). Samples were also taken from the carrier material. Total DNA was extracted from approximately 0.25 g of each sample with the PowerSoilTM DNA Isolation kit (MoBio Laboratories, Inc., Carlsbad, CA), according to the instructions of the manufacturer. Two primer sets were used to selectively amplify bacterial and fungal rDNA fragments. Universal eubacterial forward F341GC and reverse R907 primers were used to amplify the hypervariable V3-V5 region from the 16S rRNA gene, as previously reported (Yu & Morrison, 2004). The fungal first internal transcriber spacer (ITS1) from the ribosomal DNA was amplified with the primer pair ITS5 and ITS2 (White et al., 1990). The forward primer ITS5 and F341 contained the GC clamp 5′-CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGG-3′. All PCRs were performed with a Mastercycler (Eppendorf, Hamburg, Germany), and each reaction mix (25 μL mix/reaction) contained 1.25 U of ExTaq DNA polymerase (Takara Bio, Otsu, Shiga, Japan), 12.5 mM dNTPs, 0.25 μM of each primer and 100 ng of DNA.

The obtained PCR amplicons were loaded in two 8% (w/v) polyacrylamide gels with a chemical denaturing gradient ranging from 30% to 70% [100% denaturant contained 7 M urea and 40% formamide (w/v)] and electrophoretically resolved in a DGGE-4001 equipment (CBS Scientific Company, Del Mar, CA). Electrophoresis was carried out at 60 °C and at 100 V for 16 h in a 1× TAE buffer solution (40 mM Tris, 20 mM sodium acetate, 1 mM EDTA, pH 7.4). The DGGE gels were stained for 45 min in 1× TAE buffer solution containing SybrGold (Molecular Probes, Inc., Eugene, OR) and then scanned under blue light by means of a blue converter plate (UV Products Ltd, Cambridge, UK). Predominant DGGE bands were excised with a sterile filter tip, resuspended in 50 μL of sterilized Milli-Q water, and stored at 4 °C overnight. A 1 : 50 dilution of the supernatants was subsequently reamplified by PCR as described previously and sequenced using R907 and ITS2 primers, for eubacterial and fungal sequences, respectively.

Sequencing was accomplished using the ABI Prism Big Dye Terminator Cycle-Sequencing Reaction kit v. 3.1 and an ABI 3700 DNA sequencer (both Perkin–Elmer Applied Biosystems, Waltham, MA), according to the manufacturer's instructions. Sequences were edited using the bioedit software package v. 7.0.9 (Ibis Biosciences, Carlsbad, CA) and aligned with the NCBI genomic database using the blast search alignment tool. The bacterial 16S rRNA and fungal ITS1 rRNA gene nucleotide sequences determined in this study were deposited in the GenBank database under accession numbers JN982532JN982549 and JN982550JN982558, respectively.

Quantitative PCR assay

Gene copy numbers of eubacterial 16S rRNA and fungal ITS1 rRNA fragments were quantified with the quantitative real-time PCR (qPCR). Each sample was analyzed in triplicate by means of three independent DNA extracts. The analysis was carried out using Brilliant II SYBR® Green qPCR Master Mix (Stratagene, La Jolla, CA) in a Real Time PCR System MX3000-P (Stratagene) operated with the following protocol: 10 min at 95 °C, followed by 40 cycles of denaturation at 95 °C for 30 s, annealing for 30 s at 50 and 55 °C (for 16S rRNA and ITS1 rRNA, respectively), extension at 72 °C for 45 s, and fluorescence measurement at 80 °C. The specificity of PCR amplification was determined by observations on a melting curve and gel electrophoresis profile. Melting curve analysis to detect the presence of primer dimers was performed after the final extension by increasing the temperature from 55 to 95 °C with a heating rate of 0.05 °C per cycle. Each reaction was performed in a 25-μL volume containing 2 μL of DNA template (approximately 100 ng of DNA), 200 nM of each primer, 12.5 μL of the ready reaction mix, and 30 nM of ROX reference dye. The primer set for eubacterial population was 519F and 907R (Lane, 1992; Muyzer et al. 1995) and for fungal population was ITS5 and ITS2 (both couple of primers were purified by HPLC). The standard curves were performed with the following reference genes: 16S rRNA gene from Desulfovibrio vulgaris ssp. vulgaris ATCC 29579, inserted in a TOPO TA vector (Invitrogen, Belgium); and an ITS1 gene fragment obtained from a single DGGE band (GenBank accession no. JN982550) cloned onto the PGEM plasmid vector using PGEM-T Easy Vector System I (Promega, Madison, WI). All reference genes were quantified by Quant-iT PicoGreen® dsDNA Reagent using MX3000P (Stratagene) as a detector system. Tenfold serial dilutions of known copy numbers of the plasmid DNA in the range from 101 to 108 copies were subjected to a qPCR assay in duplicate to generate the standard curves. The qPCR efficiencies of amplification were greater than 96%; the Pearson correlation coefficients (R2) of the standard curves were between 0.997 and 0.994; and the slopes were between −3.353 and −3.416 for 16S rRNA and ITS rRNA, respectively. All the results were processed by means of MxPro QPCR Software.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Operation of biofilters

High removal efficiencies (RE > 90%) were immediately achieved in the start-up operation at low OLRs [OLR < 10 g m−3 h−1, empty bed residence time (EBRT) > 2 min], conditions that were maintained constant for about 30 days (Fig. 1). Thereafter, the OLR was progressively increased by applying higher substrate concentrations in the influent air (Ci), until a significant drop in the RE was observed. An adaptation period of variable length, depending on the substrate, was generally required prior to the recovery of the previously recorded RE values. The maximum EC that was achieved during this first period was 40 g m−3 h−1 for toluene (RE = 97%), 34 g m−3 h−1 for ethylbenzene (RE = 97%), and 29 g m−3 h−1 for p-xylene (RE = 92%). Neither irrigation nor nutrient amendments were applied during this initial stage, and the water content remained close to that of the original packing material (23%) in all modules that composed the biofilters treating toluene and ethylbenzene. Conversely, the p-xylene biofilter displayed a marked humidity gradient (Fig. 2); while water content in the top module (M1) stabilized at about 20%, it remained close to 40% in the bottom module (M2). To check whether such behavior was caused by heterogeneous packing, the position of the two modules was reversed after 85 days of operation in all biofilters prior to bed watering, so that the bottom module (M2) was replaced at the top position and vice versa. In just a few days of operation after module reversal (as measured in day 116), the water content of modules M1 now displaced to the bottom position raised in all biofilters up to 27–36%, while it remained fairly constant in the now top modules M2 in relation to the previously measured values.


Figure 1. OLR (○) and substrate RE (●) of three identical biofilters operated in parallel and fed respectively with toluene (a), ethylbenzene (b) and p-xylene (c). Different operation phases in relation to the EBRT and influent substrate concentration (Ci) are separated by vertical lines.

Download figure to PowerPoint


Figure 2. Evolution of the moisture content of the packing material from the biofilter modules M1 (a) and M2 (b) in the biofilter degrading toluene (■), ethylbenzene (●), and p-xylene (♦) The following operations are also indicated: watering plus module exchange (dotted line), watering plus bed mixing (dashed line), and watering plus removal of upper module (solid line) only for toluene and ethylbenzene.

Download figure to PowerPoint

In a second operational stage, the EBRT was reduced to about 90 s by increasing the gas flow after 111 days of operation of the biofilters with toluene and ethylbenzene, and after 122 days for the p-xylene biofilter. Yet, despite packing moistening and the application of a lower OLR at the beginning of this second period, the performance of the biofilters tended to deteriorate and the previously recorded EC values could not be maintained for a long time, particularly for the biofilter with ethylbenzene. The highest EC achieved in this period were 39 g m−3 h−1 for toluene (RE = 85%), 33 g m−3 h−1 for ethylbenzene (RE = 84%), and 20 g m−3 h−1 for p-xylene (RE = 90%). During this phase, the packing of all bioreactor modules was irrigated and mixed up (on day 124). Upon resuming biofilter operation, all modules tended to dry out but dewatering was particularly strong (from 35% to 21%) in the p-xylene M2 module, now placed at the biofilter top. Humidity levels in the toluene and ethylbenzene M2 modules were further reduced down to 16–18%.

A third and final operational stage consisting on a second EBRT reduction down to 44 and 47 s was applied only to the toluene and ethylbenzene biofilters, respectively, by removing the module M1, while the previous operational conditions were maintained for the biofilter fed with p-xylene. These conditions were kept for more than 30 days, until the end of the experiments. During this time, the EC and RE in the toluene biofilter progressively dropped from 60 to 40 g h−1 m−3 and from 80% to 50%, respectively. In the case of ethylbenzene, however, the biofilter performance experienced a significant failure and, despite the return to low OLR values, the RE was at the end of the experiments as low as 12% for inlet substrate concentrations of only 29–167 ppm. Despite the fact that the EBRT in the p-xylene biofilter was kept at 94 s, its performance tended to deteriorate as well, a process that was temporarily reversed by the application of relatively low OLR (RE = 100% for OLR of 53–67 g h−1 m−3, for 10 days). However, the EC and RE at the end of experiments were below 24 g m−3 h−1 and 64%, respectively.

Microbial community characterization

The fungal/bacterial biomass ratio in the liquid culture (inoculum), the original packing, and bed samples taken at the end of the biofiltration experiments was estimated from the number of fungal and bacterial ribosomal gene copies per gram of fresh weight (Fig. 3). Bacterial gene copy numbers remained within the same magnitude order in all packing samples, ranging from 5 × 1011 to 5 × 1012 gene copies g−1. Instead, the number of fungal ITS rRNA displayed a larger variability, from 5 × 109 gene copies g−1 in the p-xylene biofilter to 3 × 1011 gene copies g−1 in the toluene biofilter. Microbial gene counts in the liquid enrichment culture were significantly lower than in the biofilter samples (6 × 107 and 2 × 106 copies g−1 for bacteria and fungi, respectively). In relation to the fungal/bacterial ratio of the original packing, a significant increase in this value was only manifested in the biofilter with toluene. The proliferation of fungi was also visualized macroscopically (results not shown) and microscopically by SEM images on bed samples at the beginning and after prolonged biofilter operation (Fig. 4). Abundant fungal biomass was primarily observed in the toluene biofilter, depicted as a profuse network of filaments with an approximate width of 5–10 μm and therefore compatible with fungal hyphae and mycelial cords.


Figure 3. Average copy number of bacterial 16S rRNA (dark gray) and fungal ITS1 (light gray) partial ribosomal genes from three independent DNA extracts in different biomass samples (left-side log scale) per gram of sample fresh weight (FW); standard deviations are represented as error bars. The fungal/bacterial ratio is represented by a solid line on the right-site axis. Samples correspond to the inoculum (I) and original packing (P), and from biofilters treating toluene (T), ethylbenzene (E), and p-xylene (X) after 185 days of operation.

Download figure to PowerPoint


Figure 4. SEM pictures from two packing samples of the toluene-fed biofilter, taken at the beginning of the experimental run (a), and after 185 days of operation (b).

Download figure to PowerPoint

The microbial community structure of three similar biofilters used respectively for the biodegradation of toluene, ethylbenzene, and p-xylene was characterized by DGGE molecular profiling of bacterial 16S and fungal ITS1 rRNA genes (Fig. 5). All three biofilters were inoculated with the same toluene-degrading enrichment culture obtained from activated sludge. Four predominant bacterial ribotypes were depicted in the inoculum as DGGE bands and were successfully excised and sequenced (Table 2). The sequence from band 1 was somewhat related to different members in the Burkholderia cepacea species complex, while that of band 2 was identical to the type strain of Pandoraea pnomenusa, a species closely related to, and commonly misidentified as, B. cepacea. Bands 3 and 4 were very similar (99% sequence homology) to several uncultured ribotypes belonging to the Xanthomonadaceae family that have previously been observed in activated sludge from municipal wastewater treatment plants (Table 2). In contrast, the fungal biodiversity from the liquid culture used as inoculum was limited to one single ribotype (band 25), with an ITS1 rRNA sequence that was identical to that of the species type strain of the hypocrealean ascomycete Acremonium kiliense (Table 3).


Figure 5. DGGE profiles for bacterial 16S (a) and fungal ITS1 (b) rRNA genes from the initial inoculum (I) and packing material (P) used in the biofiltration experiments, and from three biofilters used respectively for the treatment of toluene (T), ethylbenzene (E), and p-xylene (X) after 185 days of operation (Fig. 1). Numbered DGGE bands were successfully excised and sequenced.

Download figure to PowerPoint

Table 2. The most closely related sequences found in the GenBank database (NCBI, USA) for the DGGE bands from bacterial 16S rRNA genes obtained from the DGGE profiles from Fig. 5
BandSample presenceAccessionOrderReference species, strain or uncultivated microorganism (environmental source)AccessionH (%)
  1. The most homologous sequence and the closest phylogenetically relevant match are shown (preferably type strainsT).

1IJN982532BurkholderialesBurkholderia cepacia ATCC25416THQ84907896
B. cenocepacia J2315TAM747720H96
B. vietnamiensis LMG10929TQ84910796
2IJN982533BurkholderialesPandoraea oxalativorans TA25TAB469785100
P. pnomenusa CCUG38742TAY268170100
3IJN982534XanthomonadalesUncultured (activated sludge)FJ53687499
Frateuria aurantia IFO3245TAB09119493
4IJN982535XanthomonadalesUncultured (activated sludge)HQ44007899
Gynumella flava YC6842TGQ36912294
5, 11T, EJN982536BurkholderialesAlcaligenes sp. C4M17DQ08974999
Pusillimonas noertemannii T7-7TDQ41760698
6TJN982537ActinomycetalesMicrobacterium xylanilyticum S3-ETAJ85390899
M. hydrocarbonoxydans DSM16089TAJ69872699
M. testaceum DSM20166TNR_02616399
7TJN982538ActinomycetalesStreptomyces baliensis NBRC104276TAB44171897
S. griseoplanus NBRC12779TAB18413897
S. radiopugnans R97T91293097
8TJN982539ActinomycetalesStreptomyces chungwhensis AA-98TAY38229298
S. ferralitis strain SFOp68TNR_02908798
S. paucisporeus strain 1413TNR_0432498
9, 13, 23T, E, X, PJN982542ActinomycetalesStreptomyces halotolerans YIM 90017AY37616699
10T, EJN982540SphingobacterialesUncultured (domesticated horse feces)EU463479100
Pedobacter bauzanensisBZ42TGQ16199096
12, 22E, PJN982541ActinomycetalesUncultured (cattle feedlot)FJ67156199
Aeromicrobium ginsengisoli GBS39TAB24539496
14XJN982543SphingobacterialesUncultured (activated sludge)FN59778497
Lewinella nigricans ATCC23147TAM29525586
15, 19X, PJN982544CytophagalesUncultured (algal mat)HM35704792
Flexibacter aggregans IFO15974AB07803890
16, 24T, E, X, PJN982545ActinomycetalesRhodococcus coprophilus DSM43347TX8062699
17XJN982546ActinomycetalesActinomadura madurae XMU324HM36864196
A. nitritigenes NBRC15918AB36459596
18XJN982547ThermoleophilalesUncultured (contaminated soil)AM93569497
Thermoleophilum minutum YS-4TNR_03693282
T. album HS-5TNR_02554382
20PJN982548RhizobialesPseudaminobacter sp. G210 (beach sand)GU19900398
P. salicylatoxidans BN12TNR_02871097
21X, PJN982549ActinomycetalesRuania albidiflava AS4.3142TDQ34315396
Table 3. The most closely related sequences found in the GenBank database (NCBI, USA) for the DGGE bands from fungal ITS1 rRNA genes obtained from the DGGE profiles from Fig. 5
BandSample presenceAccessionOrderReference species, strain or uncultivated microorganism (environmental source)AccessionH (%)
  1. The most homologous sequence and the closest phylogenetically relevant match are shown (preferably type strainsT).

25, 33I, XJN982553HypocrealesAcremonium killiense MUCL9724TFN691446100
26, 28, 39T, E, PJN982550ChaetotyrialesExophiala oligosperma CBS113408AY857531100
Exophiala oligosperma CBS725.88TAY16355199
27, 31, 40T, E, PJN982558EurotialesAspergillus sydowii CBS593.65TAY37386999
29EJN982551HypocrealesUncultured (watermelon rhizosphere)GQ86619099
Acremonium chrysogenum ATCC14615TACU5767294
32, 37T, E, P, XJN982552EurotialesUncultured (airfilter sample)GQ99931899
Aspergillus versicolor UOA/HCPF8640FJ87862598
34XJN982554ChaetotyrialesCladophialophora saturnica CBS114326AY857507100
30, 35E, X, PJN982555ChaetotyrialesFonsecaea sp. CBS102252JN999999100
36XJN982556SordarialesUnidentified IBL03178 (coffee seedlings)DQ68260198
38E, PJN982557HypocrealesCylindrocarpon destructans CBS185.36AM41906299

The organic packing used in the biofiltration experiments already contained a diverse population of both bacteria and fungi. In relation to the bacterial domain, the sequence from band 19 was distantly related to several, mainly uncultured, ribotypes belonging to the Cytophagales class and might, thus, belong to a yet undescribed species. Band 20 was similar in sequence (98% homology) to an undefined Pseudoactinobacter sp. as the closest phylogenetically defined match (Table 2). The remaining bacterial bands were associated with different species in the Actinomycetales: sequences from bands 21 and 22 were distantly related (96% sequence homology) to the type strain of Ruania albidiflava and Aeromicrobium ginsengisoli, respectively, while bands 23 and 24 were highly homologous (98%) to Streptomyces halotolerans and Rhodococcus coprophilus. The phylogenetic identity of some important DGGE bands from the original packing, the sequences of which could not be directly obtained, was assigned on the basis of an identical migration pattern toward sequenced bands from other related biofilter packing samples. In relation to this, the presence in the original packing of an unknown species in the Sphingobacteriales with a 16S rRNA partial sequence identical to an uncultured microorganism from horse feces was established by comparisons of the toluene sample DGGE band pattern (band 10). Several fungi were also detected in the same DNA extract from the original biofilter packing, although fungal diversity depicted by DGGE appears to be comparatively low. Most of the detected species belonged to the Chaetotyriales, Eurotiales, and Hypocreales, with the three predominant bands (3840) assigned to Cylindrocarpon destructans, Exophiala oligosperma, and Aspergillus versicolor on grounds of a high sequence homology (< 99%) toward reference type strains.

The predominant bacterial populations described at the end of the biofiltration experiments contained indigenous representatives from the original packing, such as S. halotolerans or R. coprophilus, which were present in all packing samples, but some enriched ribotypes were also found. The toluene and ethylbenzene biofilters presented a relatively low biodiversity and a similar microbial composition profile. Specific ribotypes were closely related to some species encompassed in the genera Alcaligenes and Microbacterium (bands 56; sequence homology > 99%), as well as to different Streptomyces species (bands 78; sequence homology > 97%). The biofilter exposed to p-xylene displayed a more complex microbial community, from the phylogenetic perspective, that was also more similar to the indigenous microbial population of the packing. Specific ribotypes included the very distinct bands 14 and 18, which had poor sequence homology toward any known microorganism but were relatively similar (97%) to uncultured ribotypes found previously in activated sludge and polluted soil, respectively (Table 2). In relation to the fungi, most of the predominant species found in the biofilters fed with toluene and ethylbenzene were already detected in the original packing material. Those included the previously mentioned E. oligosperma (bands 26 and 28) and A. versicolor (bands 27 and 31). Ribotypes related to A. versicolor and an uncultured Aspergillus species (bands 30 and 32; sequence homology > 98%) were also found in the ethylbenzene biofilter. The presence of C. destructans, an indigenous species in the original packing, was also found in the ethylbenzene biofilter by means of band position matching. Additional chaethothyrialean species were found in the ethylbenzene and p-xylene biofilters, which included Cladophialophora saturnica and an as yet apparently undescribed new species (bands 34 and 35; 100% sequence homology). Interestingly, the A. kiliense detected in the inoculated liquid culture was also found in the p-xylene biofilter (band 33).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In this study, three laboratory-scale biofilters treating vapors of toluene, ethylbenzene, and p-xylene were started up and operated for an initial period of 85 days without bed irrigation. The water content in the two modules that composed each biofilter dropped from an initial 32% and became stabilized to approximately 20% in the units run with toluene and ethylbenzene (Fig. 2). These humidity levels were significantly lower than the 40–60% range that has previously been reported as optimal for the biofiltration of alkylbenzenes (Cox et al., 1996; Sun et al., 2002). Yet, a relatively stable biofilter operation and performance was achieved for toluene, with acceptable EC and RE values (Fig. 1). As reviewed by Kennes & Veiga (2004), significantly higher EC values have previously been reported in biofilters fed with alkylbenzenes but, in general, those elimination rates were sustained for relatively short periods, only, and/or with higher irrigation frequencies. Under identical operational conditions, biofilter performance tended to fail when fed with ethylbenzene indicating that, despite a similar chemical structure, the substrate characteristics had a strong influence on the overall bioreactor performance. In fact, toluene has generally been found to be more easily biodegradable than ethylbenzene and p-xylene, by both fungi and bacteria (Mallakin & Ward, 1996; Prenafeta-Boldú et al., 2002). In relation to the biofilter fed with p-xylene though, a strong vertical gradient of humidity was observed between the top and bottom modules as the module from the lower position consistently had a water content of 30–40% despite bed homogenization and module reversal (Fig. 2). Water accumulation in the deeper sections of the biofilter bed is not uncommon, particularly under down-flow operation mode, because of the effect of gravity and the drying effect of the incoming air (Sakuma et al., 2009). Bed homogenization is known to improve operating conditions by reducing bed drying (Znad et al., 2007) but, despite further attempts to re-water the biofilter bed, humidity continued to drop, particularly in the biofilters fed with toluene and ethylbenzene. Such substrate-dependent spatial and temporal humidity gradients point to the fact that the water balance is somehow related to microbial population dynamics which, in its turn, are selected by the organic volatile substrate.

It is now widely accepted that in gas biofilters, microbial community structure and dynamics are strongly influenced by environmental conditions (i.e. bioreactor operational parameters). But we do not yet fully understand to what extent microbial interactions drive the macroscopic process functioning in terms of biodegradation performance and system stability (Cabrol & Malhautier, 2011). Earlier studies on the biofiltration of BTEX compounds have pointed out that operation under relatively dry and/or acidic conditions favors the development of fungi rather than bacteria, and that this effect was generally beneficial for the process (Kennes & Veiga, 2004). However, quantitative evidence supporting this claim has commonly been based on visual macroscopic and microscopic observations (Weber et al., 1995; Prenafeta-Boldú et al., 2001), as well as on culture-dependent microbial counts (García-Peña et al., 2001; Sun et al., 2002), which are inherently biased toward high propagule-producing and fast-growing microorganisms on laboratory media (Cabrol & Malhautier, 2011). The quantification of fungal-specific biomarkers in toluene biofilters demonstrated significant growth of fungi, but bacteria were then overlooked (Prenafeta-Boldú et al., 2008). Here, the culture-independent quantification of specific fungal and bacterial genes has shown a significant increase in the F/B ratio only in the case of toluene (Fig. 3). Such an increment was primarily due to the fungal biomass fraction; the number of ribosomal gene copies in fungi increased by one order of magnitude, in relation to the content of the original packing, while the bacterial numbers remained fairly constant. Contradicting the common belief that microbial enrichment in liquid cultures tends to select for bacteria while filamentous fungi are prone to the colonization of solid state-like fermentation systems (Prenafeta-Boldú et al., 2001), the F/B ratio from the biofilter inoculum was significantly higher than that of the original packing; though, the overall microbial gene content in the former was substantially lower.

Certain fungi and bacteria are known to assimilate alkylbenzenes as the sole source of carbon and energy, but the metabolism of these substrates by bacteria is better known, from both enzymatic and genomic perspectives. Several members of the Pseudomonadales, Burkholderiales, and Xanthomonadales are known to assimilate toluene and related substrates (Timmis et al., 2010) and it is therefore not surprising that the predominant bacterial ribotypes found in the inoculum (activated sludge enriched upon toluene additions) were related to these taxa. Burkholderia cepacia is in fact a complex of at least nine closely related species, including B. cenocepacia and B. vietnamiensis, commonly isolated from soil and plant roots with a known ability to degrade several organic pollutants. They are also opportunistic pathogens capable of causing life-threatening respiratory tract infections in predisposed patients (Mahenthiralingam et al., 2005). One bacterial ribotype from the inoculum was closely related to different Pandoraea spp. (Burkholderiales), commonly associated with activated sludge, but also with lung infection (Coenye et al., 2000). A distinct representative of the Burkholderiales closely related to Pusillimonas noertemannii was also observed in packing samples from the toluene and ethylbenzene biofilters. This species is able to mineralize substituted salicylates and aromatic acids (Stolz et al., 2005), analogues of some intermediates of the ethylbenzene biodegradation pathway (Gunsch et al., 2005).

Nevertheless, the bacterial ribotypes identified in the biofilter bed samples belonged predominantly to the order Actinomycetales. In several aspects, actinobacteria can be considered as the bacterial counterparts of common fungi. Just like fungal hyphae, many actinobacteria form filamentous multicellular structures, the most suitable microbial morphology for the colonization of solid substrates. Consequently, in several cases at least, filamentous growth is related to the biological resistance toward low water activity stress. Because of the need to exploit and protect a spatially defined resource, like many saprotrophic fungi, actinobacteria have developed a complex array of secondary metabolites (antibiotics, volatile compounds, etc.), as well as extracellular enzymes for the hydrolysis of polymeric substrates (McCarthy & Williams, 1992). Moreover, dispersal of such organisms into the environment primarily relies on air-born propagules, which might therefore be more easily be encountered in air biofilters. A few actinobacteria are also known to assimilate aromatic hydrocarbons, like Microbacterium hydrocarbonoxidans and related species (Schippers et al., 2005), detected here in the toluene biofilter. Other Streptomyces spp. were also present in the DGGE patterns though with lower intensity in the toluene biofilter, in contrast to the very xerophilic S. halotolerans, which was found to predominate in all packing samples. It is interesting to mention that different sequences detected in the original packing could be related to microorganisms that have been related to the ruminant digestive system; this is to be expected considering that animal dung is one of the chief packing components. These included a few ribotypes that could not be phylogenetically assigned (bands 10, 12, and 22), as well as the actinobacteria R. coprophilus (Table 2). However, the dominance of these presumably enteric microorganisms, on the basis of DGGE relative band intensity compared to that in the original packing of the biofilter, tended to decrease in most of cases, indicating that they played a minor role in the biodegradation processes.

Interestingly, one single fungus was detected in the toluene-enriched liquid culture used as inoculum for the biofiltration experiments, the ascomycete A. kiliense (Bionectriaceae), which apparently is responsible for the relatively high F/B ratio of this particular sample. This fungus is a very cosmopolitan fungus that has commonly been isolated from soil, but it has also been claimed that it can biodegrade aromatic hydrocarbons (April et al., 2000). At the end of the experiments, A. kiliense was detected in the p-xylene biofilter only, indicating that it played a minor role in the biodegradation of toluene and ethylbenzene under biofilter conditions. A fungus related to an unidentified strain in the Sordariales isolated from coffee seedlings was strongly enriched in the p-xylene biofilter. Several volatile aromatic compounds, including the xylene isomers, have been found to be emitted by green coffee (Holscher et al., 1995) and might, thus, explain the occurrence of this particular strain. Instead, the DGGE profiles indicate that the species E. oligosperma was strongly enriched in the biofilter treating toluene and might therefore contribute significantly to the high F/B biomass ratio seen in samples from this biofilter. This species was also detected in the ethylbenzene biofilter though with a lower intensity, which is consistent with the lower F/B ratio observed in this case. Exophiala oligosperma appears to be among the most common species isolated from biofilters treating volatile alkylbenzenes (Kennes & Veiga, 2004; Prenafeta-Boldú et al., 2006). Interestingly, DGGE profiles also indicated that E. oligosperma was already present in the original packing material and might thus be related to its constitutive materials (i.e. sawdust and manure). This fungus is one of the so-called black yeasts, a functional group of fungi that owe its name to their strongly melanized thallus and by an ability to grow either as filaments, budding cells, or by forming meristematic structures. Such physiological flexibility and melanin pigmentation enables members of this group to colonize a wide range of hostile and sometimes very unusual environments, so that many species are in fact considered as extremophilic eukaryotic microorganisms (de Hoog, 1999). In recent years, it has become apparent that black yeast members of the Chaetothyriales are consistently isolated from environments that are polluted with aromatic hydrocarbons, and the assimilation of toxic aromatics such as toluene and styrene as sole carbon and energy sources has been demonstrated for an increasing number of species (Prenafeta-Boldú et al., 2006). Besides melanization, these fungi are also characterized by an extremely hydrophobic biomass, which has been of advantage for the selective isolation of these organisms by extraction on mineral oil (Satow et al., 2008). This hydrophobicity could have contributed to the poor bed watering that was observed in the toluene and ethylbenzene biofilters. The p-xylene biofilter displayed a more distinct and diverse microbial profile, and the F/B ratio reached the lowest value measured (Fig. 3). Also, in contrast to the toluene and ethylbenzene biofilters, E. oligosperma was absent from the dominant fungal population. However, two other chaetothyrialean black yeasts were detected instead, though with minor intensity: the well-known toluene-growing C. saturnica (Badali et al., 2009) and a new Fonsecaea species that will soon be described (unpublished data). The latter species was also distinguished in the original packing and in the ethylbenzene biofilter. In contrast to the simpler alkylbenzenes, such as toluene and ethylbenzene, the utilization of the xylene isomers by black yeasts as the sole carbon and energy source remains inconclusive, because of the most commonly reported biodegradation by co-metabolism (Prenafeta-Boldú et al., 2001; Prenafeta Boldú et al., 2002). The comparatively lower fungal biomass and the more complex biodegradation of p-xylene might thus explain the poor performance of the biofilter fed with this substrate.

It is interesting to contemplate how certain ecological traits (xerotolerance, assimilation of aromatic hydrocarbons, and opportunistic pathogenicity) interact among fungi and bacteria in air biofilters. Among the fungal component, we are mainly concerned with members of the genera Exophiala and Cladophialophora (Herpotrichiellaceae, Chaetothyriales). In our study, E. oligosperma appears to be one of the dominant fungi from the original packing and its predominance clearly increased upon exposure to toluene. Besides being a common biofilter species, this fungus has also been found in woody materials and as an agent of opportunistic infections (de Hoog et al., 2003). The same pattern of opportunism and assimilation of alkylbenzenes has been observed with related species, such as Exophiala xenobiotica, Exophiala lecanii-corni, Phialophora sessilis, etc. (Prenafeta-Boldú et al., 2006). There are some presumably highly specialized, and also extremely virulent, human pathogens among these genera (de Hoog & Guarro, 2000), which include agents of deep skin lesions (Cladophialophora carrionii and Exophiala spinifera) as well as infections of the brain (Cladophialophora bantiana and Exophiala dermatitidis). To our knowledge, though, not a single report has been made on the occurrence of such dangerous pathogens in air biofilters. Recent evidence suggests that a process of speciation might be going on among these fungi, manifested by the occurrence of highly similar sibling species evolving, respectively, toward virulence or to saprotrophy in extreme environments (Badali et al., 2011).

The identified bacterial genera also harbor, or are related to, an important number of opportunistic pathogens which display infection patterns that are similar to those seen among the black yeasts. Several actinobacteria species are associated with subcutaneous lesions (Fahal & Hassan, 1992) and are even connected with dissemination with cerebral involvement (Hobson et al., 1995). These parallels suggest the possibility that phylogenetically very diverse groups of hydrocarbon-degrading organisms may share some common factors predisposing them to particular patterns of human pathogenicity, for example, lipophily, extremotolerance, metabolism of aromatic compounds, etc. Although tentative, the connection appears to be worth exploring as more genomic information becomes available about the groups in question. This information is fundamental for evaluating the biosafety of biofilters treating monoaromatic hydrocarbons and for the development of biotechnological processes with minimal biohazard.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The research described in this article was partly funded by the Spanish Ministry of Science and Education (CTM2006-07976). Gorka Gallastegui was supported by the Government of the Basque Country (PIFB001/2007/02) and by SGIker (UPV/EHU). We kindly acknowledge Walter Gams for the critical reading of the manuscript.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Aizpuru A, Dunat B, Christen P, Auria R, García-Peña I & Revah S (2005) Fungal biofiltration of toluene on ceramic rings. J Environ Eng 131: 396402.
  • April TM, Foght JM & Currah RS (2000) Hydrocarbon-degrading filamentous fungi isolated from flare pit soils in northern and western Canada. Can J Microbiol 46: 3849.
  • Badali H, Carvalho VO, Vicente VA, Attili-Angelis D, Kwiatkowski IB, Van den Ende AHG & de Hoog GS (2009) Cladophialophora saturnica sp. nov., a new opportunistic species of Chaetothyriales revealed using molecular data. Med Mycol 47: 5162.
  • Badali H, Prenafeta-Boldú FX, Guarro J, Klaassen C, Meis JF & de Hoog GS (2011) Cladophialophora psammophila, a novel species of Chaetothyriales with a potential use in the bioremediation of volatile aromatic hydrocarbons. Fungal Biol 115: 10191029.
  • Bohn HL & Bohn KH (1999) Moisture in biofilters. Environ Prog 18: 156161.
  • Cabrol L & Malhautier L (2011) Integrating microbial ecology in bioprocess understanding: the case of gas biofiltration. Appl Microbiol Biotechnol 90: 837849.
  • Cabrol L, Malhautier L, Poly F, Lepeuple A-S & Fanlo J-L (2012) Bacterial dynamics in steady-state biofilters: beyond functional stability. FEMS Microbiol Ecol 79: 260271.
  • Coenye T, Falsen E, Hoste B, Ohlén M, Goris J, Govan JR, Gillis M & Vandamme P et al. (2000) Description of Pandoraea gen. nov. with Pandoraea apista sp. nov., Pandoraea pulmonicola sp. nov., Pandoraea pnomenusa sp. nov., Pandoraea sputorum sp. nov. and Pandoraea norimbergensis comb. nov. Int J Syst Evol Microbiol 50: 887899.
  • Cox HHJ, Magielsen FJ, Doddema HJ & Harder W (1996) Influence of the water content and water activity on styrene degradation by Exophiala jeanselmei in biofilters. Appl Microbiol Biotechnol 45: 851856.
  • de Hoog GS (1999) Ecology and Evolution of Black Yeasts and Their Relatives, Vol. 43. CBS, Baarn/Delft, pp. 208.
  • de Hoog GS & Guarro J (2000) Atlas of Clinical Fungi. Centraalbureau voor Schimmelcultures, Baarn/Reus.
  • de Hoog GS, Vicente VA, Caligiorne RB, Kantarcioglu S, Tintelnot K, Gerrits van den Ende AHG & Haase G (2003) Species diversity and polymorphism in the Exophiala spinifera clade containing opportunistic black yeast-like fungi. J Clin Microbiol 41: 47674778.
  • Elías A, Barona A, Gallastegi G, Rojo N, Gurtubay L & Ibarra-Berastegi G (2010) Preliminary acclimation strategies for successful startup in conventional biofilters. J Air Waste Manag Assoc 60: 959967.
  • Fahal AH & Hassan MA (1992) Mycetoma. Br J Surg 79: 11381141.
  • García-Peña EI, Hernández S, Favela Torres E, Auria R & Revah S (2001) Toluene biofiltration by the fungus Scedosporium apiospermum TB1. Biotechnol Bioeng 76: 6169.
  • Gibson J & S Harwood C (2002) Metabolic diversity in aromatic compound utilization by anaerobic microbes. Annu Rev Microbiol 56: 345369.
  • Gunsch CK, Cheng Q, Kinney KA, Szaniszlo PJ & Whitman CP (2005) Identification of a homogentisate-1,2-dioxygenase gene in the fungus Exophiala lecanii-corni: analysis and implications. Appl Microbiol Biotechnol 68: 405.
  • Hobson R, Gould I & Govan J (1995) Burkholderia (Pseudomonas) cepacia; as a cause of brain abscesses secondary to chronic suppurative otitis media. Eur J Clin Microbiol Infect Dis 14: 908911.
  • Holscher W, Steinhart H & George C (1995) Aroma compounds in green coffee. Dev Food Sci 37: 785803.
  • Kennes C & Veiga MC (2004) Fungal biocatalysts in the biofiltration of VOC-polluted air. J Biotechnol 113: 305319.
  • Kennes C, Rene ER & Veiga MC (2009) Bioprocesses for air pollution control. J Chem Technol Biotechnol 84: 14191436.
  • Lane DJ (1991) 16S/23S rRNA Sequencing in Nucleic Acid Techniques in Bacterial Systematics (Wiley, New York), pp 205248.
  • Mahenthiralingam E, Urban TA & Goldberg JB (2005) The multifarious, multireplicon Burkholderia cepacia complex. Nat Rev Microbiol 3: 144156.
  • Mallakin A & Ward OP (1996) Degradation of BTEX compounds in liquid media and in peat biofilters. J Ind Microbiol 16: 309318.
  • McCarthy AJ & Williams ST (1992) Actinomycetes as agents of biodegradation in the environment – a review. Gene 115: 189192.
  • Muyzer G, Teske A, Wirsen CO & Jannasch HW (1995) Phylogenetic relationships of Thiomicrospira species and their identification in deep-sea hydrothermal vent samples by denaturing gradient gel electrophoresis of 16S rDNA fragments. Arch Microbiol 164: 165172.
  • O'Leary ND, O'Connor KE & Dobson ADW (2002) Biochemistry, genetics and physiology of microbial styrene degradation. FEMS Microbiol Rev 26: 403417.
  • Prenafeta-Boldú FX, Vervoort J, Grotenhuis JTC & van Groenestijn JW (2002) Substrate interactions during the biodegradation of benzene, toluene, ethylbenzene, and xylene (BTEX) hydrocarbons by the fungus Cladophialophora sp strain T1. Appl Environ Microbiol 68: 26602665.
  • Prenafeta-Boldú FX, Kuhn A, Luykx D, Anke H, van Groenestijn JW & de Bont JAM (2001) Isolation and characterisation of fungi growing on volatile aromatic hydrocarbons as their sole carbon and energy source. Mycol Res 105: 477484.
  • Prenafeta-Boldú FX, Summerbell R & de Hoog GS (2006) Fungi growing on aromatic hydrocarbons: biotechnology's unexpected encounter with biohazard? FEMS Microbiol Rev 30: 109130.
  • Prenafeta-Boldú FX, Illa J, van Groenestijn JW & Flotats X (2008) Influence of synthetic packing materials on the gas dispersion and biodegradation kinetics in fungal air biofilters. Appl Microbiol Biotechnol 79: 319327.
  • Sakuma T, Hattori T & Deshusses MA (2009) The effects of a lower irrigation system on pollutant removal and on the microflora of a biofilter. Environ Technol 30: 621627.
  • Satow MM, Attili-Angelis D, de Hoog GS, Angelis DF & Vicente VA (2008) Selective factors involved in oil flotation isolation of black yeasts from the environment. Stud Mycol 61: 157163.
  • Schippers A, Bosecker K, Spröer C & Schumann P (2005) Microbacterium oleivorans sp. nov. and Microbacterium hydrocarbonoxydans sp. nov., novel crude-oil-degrading Gram-positive bacteria. Int J Syst Evol Microbiol 55: 655660.
  • Stolz A, Bürger S, Kuhm A, Kämpfer P & Busse H-J (2005) Pusillimonas noertemannii gen. nov., sp. nov., a new member of the family Alcaligenaceae that degrades substituted salicylates. Int J Syst Evol Microbiol 55: 10771081.
  • Sun YM, Quan X, Chen JW, Yang FL, Xue DM, Liu YH & Yang ZH (2002) Toluene vapour degradation and microbial community in biofilter at various moisture content. Process Biochem 38: 109113.
  • Swoboda-Colberg NG (1995) Chemical contamination of the environment: sources, types, and fate of synthetic organic chemicals. Microbial Transformations and Degradation of Toxic Organic Chemicals (Young LY & Cerniglia CE, eds), pp. 2774. Wiley-Liss. Inc., New York.
  • Timmis KN, McGenity T, van der Meer JR & de Lorenzo V (2010) Handbook of Hydrocarbon and Lipid Microbiology. Springer-Verlag, Berlin, pp. 4699.
  • van Groenestijn JW & Hesselink PGM (1993) Biotechniques for air pollution control. Biodegradation 4: 283301.
  • Weber FJ, Hage KC & de Bont JAM (1995) Growth of the fungus Cladosporium sphaerospermum with toluene as the sole carbon and energy source. Appl Environ Microbiol 61: 35623566.
  • White TJ, Bruns TD, Lee S & Taylor J (1990) Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. PCR Protocols: A Guide to Methods and Applications (Innis MA, Gelfand DH, Sninsky JJ & White TJ, eds), pp. 315322. Academic Press, New York.
  • Yu Z & Morrison M (2004) Comparisons of different hypervariable regions of RSS genes for use in fingerprinting of microbial communities by PCR-denaturing gradient gel electrophoresis. Appl Environ Microbiol 70: 48004806.
  • Znad HT, Katoh K & Kawase Y (2007) High loading toluene treatment in a compost based biofilter using up-flow and down-flow swing operation. J Hazard Mater 141: 745752.