Anaerobe xxx (2014) 1e7
Contents lists available at ScienceDirect
Anaerobe journal homepage: www.elsevier.com/locate/anaerobe
Molecular biology, genetics and biotechnology
Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters Lucia Blasco a, *, Minna Kahala a, Elina Tampio a, Satu Ervasti a, Teija Paavola a,1, Jukka Rintala a, b, Vesa Joutsjoki a a b
MTT Agrifood Research, Finland Tampere University of Technology, Finland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 May 2013 Received in revised form 17 April 2014 Accepted 18 April 2014 Available online xxx
This study describes the microbial community richness and dynamics of two semi-continuously stirred biogas reactors during a time-course study of 120 days. The reactors were fed with untreated and autoclaved (160 C, 6.2 bar) food waste. The microbial community was analysed using a bacteria- and archaea-targeting 16S rRNA gene-based Terminal-Restriction Fragment Length Polymorphism (T-RFLP) approach. Compared with the archaeal community, the structures and functions of the bacterial community were found to be more complex and diverse. With the principal coordinates analysis it was possible to separate both microbial communities with 75 and 50% difference for bacteria and archaea, respectively, in the two reactors fed with the same waste but with different pretreatment. Despite the use of the same feeding material, anaerobic reactors showed a distinct community profile which could explain the differences in methane yield (2e17%). The community composition was highly dynamic for bacteria and archaea during the entire studied period. This study illustrates that microbial communities are dependent on feeding material and that correlations among specific bacterial and archaeal T-RFs can be established. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Biogas Anaerobic digestion Microbial communities T-RFLP
1. Introduction Anaerobic digestion (AD) is a sustainable waste management approach for a wide range of organic waste types in agriculture and industry [1]. The AD of organic waste is an environmentally useful technology. This treatment is advantageous because it produces valuable energy in the form of biogas. During the process, organic matter is converted to methane, carbon dioxide, water, inorganic nutrients and humus through sequential metabolic processes: hydrolysis, acidogenesis, acetogenesis mainly carried out by bacteria, and methanogenesis performed exclusively by archaeal species [2,3]. Consequently, anaerobic digesters are characterized by complex microbial communities [4]. Regardless of the growth of AD use, the relationship between these microbial communities and process efficiency is not thoroughly understood. Currently, the anaerobic digestion process can only be empirically optimized by adjusting
* Corresponding author. Tel.: þ358 295317133. E-mail address: lucia.blasco@mtt.fi (L. Blasco). 1 Present address: Biovakka Suomi Ltd, Finland.
digester conditions such as: temperature, pH, redox potential, organic loading rate (OLR) and solid and hydraulic retention time (SRT, HRT). Usually, parameters such as pH, volatile fatty acid (VFA) concentration, the total alkalinity in the digester and biogas composition are monitored and used for operational control [5]. Domestic food waste is an energy-rich substrate for anaerobic digestion. With pre-treatment such as autoclaving, the material hydrolyses. This process is followed by easier degradation during digestion. However, the treatment temperature and pressure as well as the characteristics of the treated material affect the anaerobic digestion process and biogas production. Autoclaving acts as a hygienisation step prior to digestion, but at temperatures 130e 180 C methane production has been reported to decrease [6e8] due to formation of hardly biodegradable Maillard compounds [9]. Generic tools and approaches to relate microbial community structure and dynamics to reactor performance include fingerprinting techniques such as: amplified ribosomal DNA restriction analysis (ARDRA), single strand conformation polymorphism (SSCP), denaturing gradient gel electrophoresis (DGGE), temperature gradient gel electrophoresis (TGGE), terminal restriction fragment length polymorphism (T-RFLP) and ribosomal intergenic spacer analysis (RISA) [10]. Furthermore, microbial identification
http://dx.doi.org/10.1016/j.anaerobe.2014.04.011 1075-9964/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
2
L. Blasco et al. / Anaerobe xxx (2014) 1e7
can be done using fluorescence in situ hybridization (FISH), DNA microarrays or sequencing [10]. T-RFLP is a molecular tool for monitoring microbial community compositions and their relative abundances [11]. It has been successfully applied in studying different ecosystems, including methanogenic reactors [12e14]. In previous studies, T-RFLP technique results were found to reflect the change in the microbial populations and agree with other techniques based on the 16S rDNA gene but showing more accurate information [15e20]. T-RFLP is currently recognized as one of the most powerful fingerprinting methods in microbial ecology, because it allows high sample throughput and precise fragment length determination. Therefore, although T-RFLP gives information about the diversity, structure and dynamic of the complex microbial community in anaerobic reactors, its main disadvantages are PCR bias and low resolution. On the other hand, it is fast, cheap and semi-quantitative. The present study monitored microbial community structure and dynamics from two laboratory-scale methanogenic reactors, based on their bacterial and archaeal 16S rRNA gene-targeted TRFLP profiles. The study evaluated the impact of pretreatment (autoclaving) and OLR on the dynamics of the microbial community in an STR. This study connects microbial community structure and dynamics with operational conditions in different anaerobic digesters. 2. Materials and methods 2.1. Feed and STR operation The stirred tank reactors (STRs) used in the study were fed with source-segregated domestic food waste (FW) collected from the South Shropshire Biowaste digestion plant in Ludlow, UK. One portion was pre-treated with a novel double-auger autoclave (AeroThermal Group Ltd, UK) at 160 C and 6.2 bars (autoclaved FW); the other portion was left untreated (untreated FW). Both portions were then passed through a macerating grinder (S52/010 Waste Disposer, IMC Limited, UK), frozen, shipped to Finland and thawed prior to use. The STRs (Metener Ltd, Finland) were 11-l stainless-steel reactors with semi-continuous stirring. The operating temperature was mesophilic (37 C). Reactors were fed five times a week and samples for analyses were collected through the feeding inlet. The biogas produced was measured with a volume-calibrated cylindrical gas collector based on water displacement, after which the gas was collected in gas bags. The gas composition (CH4, H2S) from the bags was analysed Combimass GA-m gas analyser (Binder Engineering, Germany). At the beginning of the experimental run, both reactors (R1M untreated and R3A autoclaved) were inoculated with sewage sludge digestate from the Biovakka Suomi Ltd Turku plant. The reactors started off with an organic loading rate (OLR) of 2 kg VS/ m3 day and a hydraulic retention time (HRT) of 117 and 94 days for untreated R1M and autoclaved R3A respectively. On day 151 the OLR was raised to3 kg VS/m3 day and on day 256, to 4 kg VS/m3 day shortening HRTs from 78 d to 58 d in R1M and from 63 d to 47 d in R3A. Microbial sampling was done in OLRs of 3 and 4 kg VS/m3. The reactors were supplemented with trace element solutions containing the cationic elements Al (0.1 mg/l), B (0.1 mg/l), Co (1.0 mg/ l), Cu (0.1 mg/l), Fe (5.0 mg/l), Mn (1.0 mg/l), Ni (1.0 mg/l), Zn (0.2 mg/l) and oxyanions Mo (0.2 mg/l), Se (0.2 mg/l) and W (0.2 mg/l) [21]. More detailed feed handling and STR operation are described in Ref. [8]. Digestate samples for chemical analyses and microbial samples during the sampling period (days 231e328) were collected routinely once a week. The characteristics of the feed and biogas reactors are provided in Tables 1 and 2.
Table 1 Feed material characteristics with standard deviations.
pH TS (%) VS (%) VS/TS (%) TKN (g/kg) NH4eN (g/kg)
Untreated FW
Autoclaved FW
5.02 0.13 24.65 0.48 22.90 0.44 92.90 7.40 0.34 0.35 0.13
5.01 0.13 20.51 0.83 18.91 0.72 92.18 6.78 0.28 0.43 0.12
2.2. Analytics The total solids (TS) and Volatile solids (VS) were determined according to SFS 3008 [22] (Finnish Standard Association, 1990). The total Kjeldahl nitrogen (TKN) was analysed with a standard method [23] using a Foss Kjeltec 2400 Analyser Unit (Foss Tecator AB, Höganäs, Sweden) with Cu as a catalyst and ammonium nitrogen (NH4eN) according to Ref. [24], pH was determined using a VWR pH100 pH-analyser (VWR International). 2.3. Microbial samples and total genomic DNA extraction Samples of 5 ml were obtained from each STR every week and aliquots of 2 ml were prepared (mixing with glycerol 1:1) and stored at 80 C until DNA extraction was performed. A sample of the untreated feeding material was also analysed. The total community DNA was extracted from selected samples based on CH4, pH and NH4eN changes. Approximately 0.25 g of each sample was used for extraction using FastDNAÒ SPIN Kit for Soil (MP Biomedicals, US) according to the manufacturer protocol. DNA extractions were visualized by ethidium bromide staining following gel electrophoresis in 1% (w/v) agarose and 1 TBE buffer. Concentration measurement of genomic DNA was performed using a NanoDrop ND1000 (NanoDrop Technologies, Wilmington, DE, USA). 2.3.1. T-RFLP analysis For T-RFLP analyses, 16S rRNA genes were amplified in duplicate for each sample using the bacterial primers pA and pH [25] and the archaeal primers 2AF and 915R [26,27]. Both forward primers were labelled at the 50 -end with the phosphoramidite dye 6-FAM and the reverse primers with VICÒ (Applied Biosystems). 1 ml of DNA extract was applied in the PCR mix for bacteria and 2 ml for archaea. The cycle profiles used were: an initial denaturation at 95 C for 3 min followed by denaturation at 95 C for 1 min, annealing at 55 C (archaea) or 52 C (bacteria) for 1 min, extension at 72 C for 3 min; the number of cycles was 35, and a final extension of 20 min was performed at 72 C. The amplicons were purified using a PCR purification kit (Qiagen, Venlo, Netherlands) and quantified using the
Table 2 Reactor characteristics. Days (d) OLR HRT (d) CH4 (m3/kg VS)a H2S (ppm) pH (kg VS/m3 day) R1M R3A R1M R3A R1M R3A R1M R3A 230 251 265 286 294 307 321 328
3 3 4 4 4 4 4 4
78 78 58 58 58 58 58 58
63 63 47 47 47 47 47 47
0.511 0.505 0.490 0.508 0.552 0.512 0.489 N/A
0.457 0.452 0.453 0.392 0.515 0.469 0.480 N/A
25 60 110 200 255 370 N/A N/A
N/A 35 50 30 0 20 N/A N/A
7.8 7.8 7.8 7.7 7.8 7.8 7.7 7.8
7.5 7.4 7.5 7.5 7.5 7.5 7.4 7.5
N/A, not available. a CH4 concentration calculated as a weekly average.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
L. Blasco et al. / Anaerobe xxx (2014) 1e7
3
Fig. 1. Relative abundance of bacterial 16S rRNA gene fragments retrieved from the anaerobic reactors R1M and R3A during the sampling period based on T-RFLP analyses with HhaI enzyme forward and reverse labelled primers for bacterial (a) and archaeal (b) populations. The length of T-RFs in base pairs (bp) was indicated.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
4
L. Blasco et al. / Anaerobe xxx (2014) 1e7
NanoDrop ND-1000 (NanoDrop Technologies, Wilmington, DE, USA). The purified PCR product (100 ng) was digested with the HhaI and HpyAV restriction enzymes for bacteria and HhaI and TaqI for archaea (Fermentas, St. Leon-Rot, Germany). DNA fragments were precipitated using 95% ethanol and washed with 70% ethanol, then vacuum-dried and resuspended in 15 ml of distilled water. 1 ml of this suspension was mixed with 9 ml of formamide containing GeneScan 1200 LIZ Size Standard (Applied Biosystems, Halle, Belgium) and separated on a 3500l Genetic Analyser (Applied Biosystems, Halle, Belgium). 2.3.2. Statistical analysis of the T-RFLP data The T-RFLP electropherograms were analysed using Peak Scanner Software v. 1.0 (Applied Biosystems, Halle, Belgium). The relative abundances of T-RFs were determined by calculating the ratio between the height of each peak and the total peak height of all peaks within one sample. The cut off point for fragment sizes included in further analysis was >20 bp. Terminal restriction fragment length polymorphism (T-RFLP) analysis was done in triplicate for each sample. The restriction endonucleases selected for further analyses were those producing the highest numbers and best size distribution of T-RFs in silico using the online program MiCA ISPaR [28]. After testing these enzymes on the extracted DNA one enzyme seemed the most suitable. Thus HhaI using reverse and forward primers was chosen for further analyses of archaea and bacteria, respectively. MiCA APLAUS [28] was used to infer the plausible community structure data. Only peaks with more than 2% abundance were considered. The range-weighted richness was determined as the number of peaks in each electropherogram. The T-RFLP profile was interpreted using principal coordinates analysis (PCoA) with the Bray Curtis similarity index, using the software PAST v.2.15 (PAleontological STatistic [29] and Qiime [30]). Shannon’s diversity index (H) was calculated on T-RFLP data using PAST software as: H ¼ Sri ln ri [31]. A one-way analysis of variance (ANOVA) was performed (SASÒ software package, Version 9.2.) to statistically evaluate whether microbial assemblages varied between bioreactor treatments, and Spearman’s rank correlation test was used to assess possible correlations between abundance and presence of archaeal and bacterial T-RFs. 3. Results Microbial community structure was investigated during the sampling period (days 230e328) in both reactors by performing a PCR amplification of 16S rDNA. Using the same amount of DNA solution prepared from 250 mg of each sample, T-RFLP analysis of the samples was carried out under the same conditions. The T-RFLP fingerprints of bacterial 16S rRNA gene fragments revealed a total of 33 terminal restriction fragments (T-RFs) on the
Fig. 2. TKN and NH4eN during the studied period on both reactors.
forward fragment and 19 T-RFs on the reverse fragment of the enzyme HhaI for the bacterial population (Fig. 1). For archaea, 17 TRFs on the forward fragment and 18 T-RFs on the reverse fragment were generated with the same enzyme. This result implies a lower diversity of the archaeal community than of the bacterial population in both reactors (the tested enzymes HpyAV for bacteria and TaqI for archaea showed a lower number of T-RFs). Some T-RFs were detected in high abundance; thereby, on bacteria the T-RFs 377, 555, 560 and 1084 accounted for an average of 70% of the total population and the reverse fragments 431, 435 and 436 an average of 61%. At the same time the archaea HhaI forward fragments T-RFs 321 and 326 accounted for a range of abundance from 34 to 94% (Fig. 1). It was encountered that for fragments 182 and 376, TaqI enzyme accounted for 91% of total abundance of population (data not shown). The change of OLR did not significantly influence the appearance or disappearance of specific T-RFs but it did reveal a clear change of the abundance of some T-RFs on bacterial and archaeal populations. The presence of T-RFs within STRs was very similar during the sampling period, primarily displaying difference in abundances between the reactors. The methane production was notably higher (2e17%) in the untreated reactor (R1M) compared with the autoclaved (R3A) (Table 2). When comparing Shannone Wiener diversity indexes, the reactor R1M showed a higher diversity for bacterial communities than the R3A during the course of the process, whereas the correlation of indexes on archaeal populations appeared to be more or less random (Fig. 3). PCoA with the Bray Curtis similarity index was used to visualize the relationships among bacterial and archaeal communities. PCoA of 16S rDNA TRFLP data revealed clustering related to the type of reactor (for bacterial communities, variance between reactors was as high as 74%) but not so clearly related to date (13%) although there was a segregation based on the change of OLR on day 259 (see Fig. 4a). Associations between the community structure and pretreatments or OLR change were observed but not as evidently on the archaeal population compared to that of bacterial population (Fig. 4b). When archaeal T-RFLP data was analysed using the APLAUS application, we found out that the genus Methanosarcina was present during the entire period in both reactors. Meanwhile the genera Methanocalculus and Methanoculleus (both belonging to the order Methanomicrobiales) and the genus Methanococcus (family Methanococcaceae) were present in the R1M during the entire period but were undetectable in R3A after day 251. Moreover, a large fraction of 16S-rDNA T-RFs could only be assigned to uncultured archaeon, demonstrating that numerous microorganisms are still unclassified or unknown. When the bacterial population was analysed it was found that the genus Clostridia as well as Bacillus was found in both reactors, allocated in several sizes of T-RFs.
Fig. 3. ShannoneWiener diversity indexes of archaea and bacteria.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
L. Blasco et al. / Anaerobe xxx (2014) 1e7
5
microbial population composition in the untreated feeding material differed totally from that observed in the reactors (data not showed). Fig. 5 marks the common T-RFs and T-RFs which showed significant differences between reactors and illustrates the Spearman’s rank correlation coefficient between different T-RFs of archaeal and bacterial population for the enzyme HhaI. TKN, NH4eN and pH levels remained relatively stable during the period of study; however, in R3M the levels were lower (Fig. 2, Table 2). H2S content increased from
Contents lists available at ScienceDirect
Anaerobe journal homepage: www.elsevier.com/locate/anaerobe
Molecular biology, genetics and biotechnology
Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters Lucia Blasco a, *, Minna Kahala a, Elina Tampio a, Satu Ervasti a, Teija Paavola a,1, Jukka Rintala a, b, Vesa Joutsjoki a a b
MTT Agrifood Research, Finland Tampere University of Technology, Finland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 May 2013 Received in revised form 17 April 2014 Accepted 18 April 2014 Available online xxx
This study describes the microbial community richness and dynamics of two semi-continuously stirred biogas reactors during a time-course study of 120 days. The reactors were fed with untreated and autoclaved (160 C, 6.2 bar) food waste. The microbial community was analysed using a bacteria- and archaea-targeting 16S rRNA gene-based Terminal-Restriction Fragment Length Polymorphism (T-RFLP) approach. Compared with the archaeal community, the structures and functions of the bacterial community were found to be more complex and diverse. With the principal coordinates analysis it was possible to separate both microbial communities with 75 and 50% difference for bacteria and archaea, respectively, in the two reactors fed with the same waste but with different pretreatment. Despite the use of the same feeding material, anaerobic reactors showed a distinct community profile which could explain the differences in methane yield (2e17%). The community composition was highly dynamic for bacteria and archaea during the entire studied period. This study illustrates that microbial communities are dependent on feeding material and that correlations among specific bacterial and archaeal T-RFs can be established. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Biogas Anaerobic digestion Microbial communities T-RFLP
1. Introduction Anaerobic digestion (AD) is a sustainable waste management approach for a wide range of organic waste types in agriculture and industry [1]. The AD of organic waste is an environmentally useful technology. This treatment is advantageous because it produces valuable energy in the form of biogas. During the process, organic matter is converted to methane, carbon dioxide, water, inorganic nutrients and humus through sequential metabolic processes: hydrolysis, acidogenesis, acetogenesis mainly carried out by bacteria, and methanogenesis performed exclusively by archaeal species [2,3]. Consequently, anaerobic digesters are characterized by complex microbial communities [4]. Regardless of the growth of AD use, the relationship between these microbial communities and process efficiency is not thoroughly understood. Currently, the anaerobic digestion process can only be empirically optimized by adjusting
* Corresponding author. Tel.: þ358 295317133. E-mail address: lucia.blasco@mtt.fi (L. Blasco). 1 Present address: Biovakka Suomi Ltd, Finland.
digester conditions such as: temperature, pH, redox potential, organic loading rate (OLR) and solid and hydraulic retention time (SRT, HRT). Usually, parameters such as pH, volatile fatty acid (VFA) concentration, the total alkalinity in the digester and biogas composition are monitored and used for operational control [5]. Domestic food waste is an energy-rich substrate for anaerobic digestion. With pre-treatment such as autoclaving, the material hydrolyses. This process is followed by easier degradation during digestion. However, the treatment temperature and pressure as well as the characteristics of the treated material affect the anaerobic digestion process and biogas production. Autoclaving acts as a hygienisation step prior to digestion, but at temperatures 130e 180 C methane production has been reported to decrease [6e8] due to formation of hardly biodegradable Maillard compounds [9]. Generic tools and approaches to relate microbial community structure and dynamics to reactor performance include fingerprinting techniques such as: amplified ribosomal DNA restriction analysis (ARDRA), single strand conformation polymorphism (SSCP), denaturing gradient gel electrophoresis (DGGE), temperature gradient gel electrophoresis (TGGE), terminal restriction fragment length polymorphism (T-RFLP) and ribosomal intergenic spacer analysis (RISA) [10]. Furthermore, microbial identification
http://dx.doi.org/10.1016/j.anaerobe.2014.04.011 1075-9964/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
2
L. Blasco et al. / Anaerobe xxx (2014) 1e7
can be done using fluorescence in situ hybridization (FISH), DNA microarrays or sequencing [10]. T-RFLP is a molecular tool for monitoring microbial community compositions and their relative abundances [11]. It has been successfully applied in studying different ecosystems, including methanogenic reactors [12e14]. In previous studies, T-RFLP technique results were found to reflect the change in the microbial populations and agree with other techniques based on the 16S rDNA gene but showing more accurate information [15e20]. T-RFLP is currently recognized as one of the most powerful fingerprinting methods in microbial ecology, because it allows high sample throughput and precise fragment length determination. Therefore, although T-RFLP gives information about the diversity, structure and dynamic of the complex microbial community in anaerobic reactors, its main disadvantages are PCR bias and low resolution. On the other hand, it is fast, cheap and semi-quantitative. The present study monitored microbial community structure and dynamics from two laboratory-scale methanogenic reactors, based on their bacterial and archaeal 16S rRNA gene-targeted TRFLP profiles. The study evaluated the impact of pretreatment (autoclaving) and OLR on the dynamics of the microbial community in an STR. This study connects microbial community structure and dynamics with operational conditions in different anaerobic digesters. 2. Materials and methods 2.1. Feed and STR operation The stirred tank reactors (STRs) used in the study were fed with source-segregated domestic food waste (FW) collected from the South Shropshire Biowaste digestion plant in Ludlow, UK. One portion was pre-treated with a novel double-auger autoclave (AeroThermal Group Ltd, UK) at 160 C and 6.2 bars (autoclaved FW); the other portion was left untreated (untreated FW). Both portions were then passed through a macerating grinder (S52/010 Waste Disposer, IMC Limited, UK), frozen, shipped to Finland and thawed prior to use. The STRs (Metener Ltd, Finland) were 11-l stainless-steel reactors with semi-continuous stirring. The operating temperature was mesophilic (37 C). Reactors were fed five times a week and samples for analyses were collected through the feeding inlet. The biogas produced was measured with a volume-calibrated cylindrical gas collector based on water displacement, after which the gas was collected in gas bags. The gas composition (CH4, H2S) from the bags was analysed Combimass GA-m gas analyser (Binder Engineering, Germany). At the beginning of the experimental run, both reactors (R1M untreated and R3A autoclaved) were inoculated with sewage sludge digestate from the Biovakka Suomi Ltd Turku plant. The reactors started off with an organic loading rate (OLR) of 2 kg VS/ m3 day and a hydraulic retention time (HRT) of 117 and 94 days for untreated R1M and autoclaved R3A respectively. On day 151 the OLR was raised to3 kg VS/m3 day and on day 256, to 4 kg VS/m3 day shortening HRTs from 78 d to 58 d in R1M and from 63 d to 47 d in R3A. Microbial sampling was done in OLRs of 3 and 4 kg VS/m3. The reactors were supplemented with trace element solutions containing the cationic elements Al (0.1 mg/l), B (0.1 mg/l), Co (1.0 mg/ l), Cu (0.1 mg/l), Fe (5.0 mg/l), Mn (1.0 mg/l), Ni (1.0 mg/l), Zn (0.2 mg/l) and oxyanions Mo (0.2 mg/l), Se (0.2 mg/l) and W (0.2 mg/l) [21]. More detailed feed handling and STR operation are described in Ref. [8]. Digestate samples for chemical analyses and microbial samples during the sampling period (days 231e328) were collected routinely once a week. The characteristics of the feed and biogas reactors are provided in Tables 1 and 2.
Table 1 Feed material characteristics with standard deviations.
pH TS (%) VS (%) VS/TS (%) TKN (g/kg) NH4eN (g/kg)
Untreated FW
Autoclaved FW
5.02 0.13 24.65 0.48 22.90 0.44 92.90 7.40 0.34 0.35 0.13
5.01 0.13 20.51 0.83 18.91 0.72 92.18 6.78 0.28 0.43 0.12
2.2. Analytics The total solids (TS) and Volatile solids (VS) were determined according to SFS 3008 [22] (Finnish Standard Association, 1990). The total Kjeldahl nitrogen (TKN) was analysed with a standard method [23] using a Foss Kjeltec 2400 Analyser Unit (Foss Tecator AB, Höganäs, Sweden) with Cu as a catalyst and ammonium nitrogen (NH4eN) according to Ref. [24], pH was determined using a VWR pH100 pH-analyser (VWR International). 2.3. Microbial samples and total genomic DNA extraction Samples of 5 ml were obtained from each STR every week and aliquots of 2 ml were prepared (mixing with glycerol 1:1) and stored at 80 C until DNA extraction was performed. A sample of the untreated feeding material was also analysed. The total community DNA was extracted from selected samples based on CH4, pH and NH4eN changes. Approximately 0.25 g of each sample was used for extraction using FastDNAÒ SPIN Kit for Soil (MP Biomedicals, US) according to the manufacturer protocol. DNA extractions were visualized by ethidium bromide staining following gel electrophoresis in 1% (w/v) agarose and 1 TBE buffer. Concentration measurement of genomic DNA was performed using a NanoDrop ND1000 (NanoDrop Technologies, Wilmington, DE, USA). 2.3.1. T-RFLP analysis For T-RFLP analyses, 16S rRNA genes were amplified in duplicate for each sample using the bacterial primers pA and pH [25] and the archaeal primers 2AF and 915R [26,27]. Both forward primers were labelled at the 50 -end with the phosphoramidite dye 6-FAM and the reverse primers with VICÒ (Applied Biosystems). 1 ml of DNA extract was applied in the PCR mix for bacteria and 2 ml for archaea. The cycle profiles used were: an initial denaturation at 95 C for 3 min followed by denaturation at 95 C for 1 min, annealing at 55 C (archaea) or 52 C (bacteria) for 1 min, extension at 72 C for 3 min; the number of cycles was 35, and a final extension of 20 min was performed at 72 C. The amplicons were purified using a PCR purification kit (Qiagen, Venlo, Netherlands) and quantified using the
Table 2 Reactor characteristics. Days (d) OLR HRT (d) CH4 (m3/kg VS)a H2S (ppm) pH (kg VS/m3 day) R1M R3A R1M R3A R1M R3A R1M R3A 230 251 265 286 294 307 321 328
3 3 4 4 4 4 4 4
78 78 58 58 58 58 58 58
63 63 47 47 47 47 47 47
0.511 0.505 0.490 0.508 0.552 0.512 0.489 N/A
0.457 0.452 0.453 0.392 0.515 0.469 0.480 N/A
25 60 110 200 255 370 N/A N/A
N/A 35 50 30 0 20 N/A N/A
7.8 7.8 7.8 7.7 7.8 7.8 7.7 7.8
7.5 7.4 7.5 7.5 7.5 7.5 7.4 7.5
N/A, not available. a CH4 concentration calculated as a weekly average.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
L. Blasco et al. / Anaerobe xxx (2014) 1e7
3
Fig. 1. Relative abundance of bacterial 16S rRNA gene fragments retrieved from the anaerobic reactors R1M and R3A during the sampling period based on T-RFLP analyses with HhaI enzyme forward and reverse labelled primers for bacterial (a) and archaeal (b) populations. The length of T-RFs in base pairs (bp) was indicated.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
4
L. Blasco et al. / Anaerobe xxx (2014) 1e7
NanoDrop ND-1000 (NanoDrop Technologies, Wilmington, DE, USA). The purified PCR product (100 ng) was digested with the HhaI and HpyAV restriction enzymes for bacteria and HhaI and TaqI for archaea (Fermentas, St. Leon-Rot, Germany). DNA fragments were precipitated using 95% ethanol and washed with 70% ethanol, then vacuum-dried and resuspended in 15 ml of distilled water. 1 ml of this suspension was mixed with 9 ml of formamide containing GeneScan 1200 LIZ Size Standard (Applied Biosystems, Halle, Belgium) and separated on a 3500l Genetic Analyser (Applied Biosystems, Halle, Belgium). 2.3.2. Statistical analysis of the T-RFLP data The T-RFLP electropherograms were analysed using Peak Scanner Software v. 1.0 (Applied Biosystems, Halle, Belgium). The relative abundances of T-RFs were determined by calculating the ratio between the height of each peak and the total peak height of all peaks within one sample. The cut off point for fragment sizes included in further analysis was >20 bp. Terminal restriction fragment length polymorphism (T-RFLP) analysis was done in triplicate for each sample. The restriction endonucleases selected for further analyses were those producing the highest numbers and best size distribution of T-RFs in silico using the online program MiCA ISPaR [28]. After testing these enzymes on the extracted DNA one enzyme seemed the most suitable. Thus HhaI using reverse and forward primers was chosen for further analyses of archaea and bacteria, respectively. MiCA APLAUS [28] was used to infer the plausible community structure data. Only peaks with more than 2% abundance were considered. The range-weighted richness was determined as the number of peaks in each electropherogram. The T-RFLP profile was interpreted using principal coordinates analysis (PCoA) with the Bray Curtis similarity index, using the software PAST v.2.15 (PAleontological STatistic [29] and Qiime [30]). Shannon’s diversity index (H) was calculated on T-RFLP data using PAST software as: H ¼ Sri ln ri [31]. A one-way analysis of variance (ANOVA) was performed (SASÒ software package, Version 9.2.) to statistically evaluate whether microbial assemblages varied between bioreactor treatments, and Spearman’s rank correlation test was used to assess possible correlations between abundance and presence of archaeal and bacterial T-RFs. 3. Results Microbial community structure was investigated during the sampling period (days 230e328) in both reactors by performing a PCR amplification of 16S rDNA. Using the same amount of DNA solution prepared from 250 mg of each sample, T-RFLP analysis of the samples was carried out under the same conditions. The T-RFLP fingerprints of bacterial 16S rRNA gene fragments revealed a total of 33 terminal restriction fragments (T-RFs) on the
Fig. 2. TKN and NH4eN during the studied period on both reactors.
forward fragment and 19 T-RFs on the reverse fragment of the enzyme HhaI for the bacterial population (Fig. 1). For archaea, 17 TRFs on the forward fragment and 18 T-RFs on the reverse fragment were generated with the same enzyme. This result implies a lower diversity of the archaeal community than of the bacterial population in both reactors (the tested enzymes HpyAV for bacteria and TaqI for archaea showed a lower number of T-RFs). Some T-RFs were detected in high abundance; thereby, on bacteria the T-RFs 377, 555, 560 and 1084 accounted for an average of 70% of the total population and the reverse fragments 431, 435 and 436 an average of 61%. At the same time the archaea HhaI forward fragments T-RFs 321 and 326 accounted for a range of abundance from 34 to 94% (Fig. 1). It was encountered that for fragments 182 and 376, TaqI enzyme accounted for 91% of total abundance of population (data not shown). The change of OLR did not significantly influence the appearance or disappearance of specific T-RFs but it did reveal a clear change of the abundance of some T-RFs on bacterial and archaeal populations. The presence of T-RFs within STRs was very similar during the sampling period, primarily displaying difference in abundances between the reactors. The methane production was notably higher (2e17%) in the untreated reactor (R1M) compared with the autoclaved (R3A) (Table 2). When comparing Shannone Wiener diversity indexes, the reactor R1M showed a higher diversity for bacterial communities than the R3A during the course of the process, whereas the correlation of indexes on archaeal populations appeared to be more or less random (Fig. 3). PCoA with the Bray Curtis similarity index was used to visualize the relationships among bacterial and archaeal communities. PCoA of 16S rDNA TRFLP data revealed clustering related to the type of reactor (for bacterial communities, variance between reactors was as high as 74%) but not so clearly related to date (13%) although there was a segregation based on the change of OLR on day 259 (see Fig. 4a). Associations between the community structure and pretreatments or OLR change were observed but not as evidently on the archaeal population compared to that of bacterial population (Fig. 4b). When archaeal T-RFLP data was analysed using the APLAUS application, we found out that the genus Methanosarcina was present during the entire period in both reactors. Meanwhile the genera Methanocalculus and Methanoculleus (both belonging to the order Methanomicrobiales) and the genus Methanococcus (family Methanococcaceae) were present in the R1M during the entire period but were undetectable in R3A after day 251. Moreover, a large fraction of 16S-rDNA T-RFs could only be assigned to uncultured archaeon, demonstrating that numerous microorganisms are still unclassified or unknown. When the bacterial population was analysed it was found that the genus Clostridia as well as Bacillus was found in both reactors, allocated in several sizes of T-RFs.
Fig. 3. ShannoneWiener diversity indexes of archaea and bacteria.
Please cite this article in press as: Blasco L, et al., Dynamics of microbial communities in untreated and autoclaved food waste anaerobic digesters, Anaerobe (2014), http://dx.doi.org/10.1016/j.anaerobe.2014.04.011
L. Blasco et al. / Anaerobe xxx (2014) 1e7
5
microbial population composition in the untreated feeding material differed totally from that observed in the reactors (data not showed). Fig. 5 marks the common T-RFs and T-RFs which showed significant differences between reactors and illustrates the Spearman’s rank correlation coefficient between different T-RFs of archaeal and bacterial population for the enzyme HhaI. TKN, NH4eN and pH levels remained relatively stable during the period of study; however, in R3M the levels were lower (Fig. 2, Table 2). H2S content increased from
