Cultivation of pure cell cultures
Pure cultures of Escherichia coli (DSM 1116), Pseudomonas fluorescens (DSM 50090), and Clostridium thermocellum (DSM 1237) as well as a mixed culture of Methanosarcina barkeri (DSM800) and Propionibacterium acne (DSM1897) were cultivated under defined conditions as control samples. Therefore, cryo beads of E. coli and P. fluorescens were pre-cultivated over night at 37°C (E. coli) or 30°C (P. fluorescens) in filtrated Nutrient Broth (NB) medium. For this pre-culture, approx. 106 cells per ml were used to inoculate 100 ml fresh and NB medium. These cultures were incubated for 10 h at the respective optimal growth temperature to obtain the working culture. C. thermocellum cells were cultivated in GS2 nutrient solution  at anaerobic conditions at 55°C for 30 h. M. barkeri and P. acne were cultivated in mixed culture in DSM medium 120 (Leibniz Institute DSMZ, German Collection of Microorganisms and Cell Cultures, Germany) at anaerobic conditions at 37°C for 48 h. All culture media were sterilized by autoclaving process before use.
Operation and sampling of the biogas reactor
The design and operation of the upflow anaerobic solid state (UASS) reactor connected with a downstream anaerobic filter (AF) reactor was described in detail by Pohl et al. (2012) . For this study, chopped wheat straw was used as substrate at an organic loading rate (OLR) of 2.5 g volatile substances (VS) per liter and day. The UASS reactor was operated at mesophilic temperatures (37°C). Two liquid samples were taken from the effluent of the UASS reactor at various times (hereafter referred to as UASS-1 und UASS-2). Samples were processed immediately after sampling for further analyses.
Sample fixation was carried out immediately after sampling according to a protocol after Kepner and Pratt (1994) . Therefore, 10 ml of pure cultures or liquid samples from the UASS reactor, respectively, were fixed with 30 ml of a 3.7% formaldehyde solution (diluted in 1× PBS pH 7.4) for 4 h at 4°C. After fixation, the samples were centrifuged at 8,000 × g for 20 min at room temperature (RT). The supernatant was discarded and the pellet was washed twice in 1× PBS using same centrifugation conditions as before. The 1× PBS was prepared of 140 mM NaCl, 10 mM Na2HPO4, 2.7 mM KCl, and 1.8 mM KH2PO4. The pH was adjusted to 7.4 with HCl (all reagents were provided by Carl Roth GmbH & Co. KG, Germany). After washing the pellet was re-suspended in 5 ml 1× PBS, mixed with 5 ml 96% ethanol p.a. and stored until further use at −20°C.
Alternatively, a fixation with 50% ethanol (diluted in 1× PBS pH 7.4) was performed for Gram-positive prokaryotes. In this case, the samples were centrifuged at 8,000 × g for 20 min. The pellets were re-suspended in 5 ml 1× PBS, mixed with 5 ml 96% ethanol p.a. and stored until further use at −20°C.
Sample pre-treatment for Flow-FISH analyses
Six different pre-treatment techniques for sample purification taken from the recent literature (in the following denominated as procedure 1 to procedure 6) were tested on both, pure cultures and UASS biogas reactor samples. An overview about all pre-treatment procedures and their modifications is given in Table 1.
The following modifications of these procedures were conducted: (a) varying concentrations of detergents (Table 1, index C), (b) ultrasound treatment at varying intensities (Table 1, index S), (c) without and with homogenization at varying intensities (Table 1, index H), and (d) with or without filtration (Table 1, index F).
For procedure 1, 10 ml of fixed sample was centrifuged at 8,000 × g for 20 min at room temperature. For procedures 2–6, a similar volume was centrifuged at 15,000 × g for 5 min at room temperature. Afterwards, all preparations were washed once with 1× PBS (pH 7.4) to remove ethanol. The solid residues were re-suspended according to the respective literature. All applications were carried out in triplicates. In the following, purification procedure 1 is described in detail because this procedure is the optimized pre-treatment method for Flow-FISH, while the other pre-treatment techniques were carried out as published previously (Table 1). All applied modifications are described in Table 1.
Procedure 1 modified after Singh-Verma  and Bakken [24, 26]: The cell pellet was washed with sterile 1× PBS (pH 7.4). After centrifugation at 8,000 × g for 20 min the cell pellet was re-suspended in 10 ml sterile 0.5% sodium hexametaphosphate (pH 8.5, Sigma-Aldrich, Germany). After 10 min of incubation the sample was sonicated at 65 W for 1 min (Sonoplus GW2070, Bandelin, Berlin, Germany). A centrifugation step at 650 × g for 2 min was conducted to separate microorganisms from organic or inorganic particles in the sample. The supernatant containing free cells was transferred in a sterile tube for further application. The residual cell pellet was re-suspended in 10 ml sterile 0.5% sodium hexametaphosphate (pH 8.5) and incubated for 10 min followed by a further ultrasonic treatment and centrifugation step. The sodium hexametaphosphate incubation step, the ultrasound step, and the centrifugation step were repeated up to five times depending on sample consistence. After five repetitions, the remaining pellet should consist mainly of organic and inorganic material and a negligible quantity of free microbial cells. The supernatants containing free microbial cells were pooled in a sterile tube. The cells were collected by centrifugations at 8,000 × g for 20 min. The supernatant was discarded and the pelleted cells were re-suspended in 10 ml 1× PBS (pH 7.4). Afterwards, a vacuum filtration of the sample using a sterile filter with 12–15 μm pore size was conducted. The filter was washed once with 40 ml 1× PBS (pH 7.4). Subsequently, the filtrate was centrifuged at 8,000 × g for 20 min. The supernatant was discarded, and the pellet was re-suspended in 10 ml of 1× PBS (pH 7.4) and used for the Flow-FISH analysis. In addition, the residues on the filter were collected described as following: to re-suspend particles and cells the filter was transferred into a 50 ml tube and incubated in 9 ml 1× PBS (pH 7.4) at room temperature for 20 min with slow rotation. After incubation, the filter was carefully rinsed off with 1 ml 1× PBS (pH 7.4). The residues on the filter were subsequently used for the microscopic verification of purification success.
All samples purified by the six procedures were stored at 4°C no longer than 12 h until further processing.
Verification of purification procedures
One important criterion for a purification method is a minimized loss of cells. Unfortunately, cell densities of untreated biogas reactor samples could not be calculated by particle counting due to interfering particles and cell aggregates. Hence, pure cultures of E. coli were used for determination of cell losses during the purification procedures. Cell counts were determined in triplicates by Coulter Counter (Multisizer™ 3 Coulter Counter®, Beckman Coulter, Germany). Each triplicate was measured three times and the standard deviation of the nine measurements was calculated. Measurements were carried out with a 50 μm capillary, and the measurement volume was 50 μl. To determine the particle number and size within the electrolyte solution (‘background control’), the electrolyte was measured without addition of any microorganisms.
For the verification of the purification success in terms of cell aggregates disbandment and detachment of microorganisms from particles, the washed pellets, the supernatants, and the residues on the filter were visually evaluated by fluorescence microscopy. For microscopic analyses 10 μl of residues on the filter, pellet samples, and supernatants each diluted 1:500 in sterile water were coated on separate wells of a 10-well-slide in triplicates. After drying the samples at 40°C the antifading reagent Citifluor A1 (PLANO GmbH, Wetzlar, Germany) was added to coat each well and 0.2 μl of a 20 μg ml-1 stock solution of 4’,6-diamidino-2-phenylindole (DAPI) were carefully injected into the Citifluor A1 drop. The size of cell aggregates was determined by microscopic field analyses using an ocular micrometer at 630× magnification. Five randomly chosen microscopic fields from each sample were analyzed in terms of the sizes of cell aggregates, the presence of organic and inorganic particles, and their microbiological growth. One microscopic field comprised the total area of 144 μm2 and was divided into 10 × 10 sub-fields of 5.76 μm2 each. All microscopic analyses were conducted with a Nikon Optiphot-2 microscope (Nikon, Duesseldorf, Germany) fitted with a DAPI AMCA filter tube or with an Olympus BX51 fluorescence microscope (Olympus GmbH, Hamburg, Germany) fitted with a U-MWU2 filter module.
Fluorescence in situ hybridization (FISH)
FISH was carried out with domain specific probes EUB338 (5′-GCTGCCTCCCGTAGGAGT-3′)  and ARCH915 (5′-GTGCTCCCCCGCCAATTCCT-3′)  for the detection of bacteria and archaea, respectively. For the detection of undesired cross hybridization with non-target microorganisms the nonsense probe NonEUB338 (5′-ACTCCTACGGGAGGCAGC-3′)  was used. Furthermore, negative controls without the addition of probes were performed to determine autofluorescence effects. All FISH probes were labeled with fluorescent dye Alexa488 and were manufactured by Eurofins MWG GmbH (Ebersberg, Germany). Flow-FISH was carried out in triplicates which were each analyzed three times by flow cytometry. Based on these in total nine measurements an average with a standard deviation was calculated.
The modified protocol for Flow-FISH of biogas reactor samples established in this study consists of following steps: 250 μl fixed sample was centrifuged at 8,000 × g for 20 min. All centrifugation steps were conducted at room temperature. The supernatant was discarded, and the pellet was re-suspended in 221 μl of 46°C preheated hybridization buffer (0.9 M NaCl, 20 mM Tris/HCl (pH 7.2), 0.1% SDS and 50% formamide) and 21 μl of the FISH probe (50 ng μl-1). During incubation at 46°C for 2 h, the sample was repeatedly inverted. A centrifugation step at 8,000 × g for 20 min ensured the pelleting of microbial cells. The cell pellet was washed twice with 500 μl 0.05 M PBS pH 7.0 using the same centrifugation conditions as before. The phosphate buffered saline (PBS) was prepared of 137 mM NaCl, 2.7 mM KCl, 40.6 mM Na2HPO4, and 7.1 mM KH2PO4. The pH was adjusted to 7.0 with HCl and the buffer was finally filtered with a 0.2 μm membrane filter.
For comparison, the following conventional FISH protocol according to Amann et al. (1990) , Wallner et al. (1993) , and Grzonka (2008)  was also performed: 1 ml fixed sample was centrifuged at 8,000 × g for 20 min. The pellet was dehydrated stepwise in 1 ml 50%, 80% and 96% ethanol for 3 min each. After each ethanolic treatment a centrifugation at 8,000 × g for 20 min was conducted. After completed dehydration the pellet was re-suspended in 46°C preheated hybridization buffer (0.9 M NaCl, 20 mM Tris/HCl (pH 7.2), 0.1% SDS, and 50% formamide) containing FISH probe with an end concentration of 5 ng per μl. The hybridization was carried out in the dark for 2 h at 46°C in a water bath with occasional inverting. To remove hybridization buffer and non-bound probes the samples were centrifuged at 8,000 × g for 20 min and washed with 0.05 M PBS (pH 7.0). After further centrifugation at 8,000 × g for 20 min, the pellet was re-suspended in 0.05 M PBS (pH 7.0) to obtain a cell concentration of approximately 106 cells per ml suited for subsequent flow cytometric analysis.
For flow cytometry, a Cytomics FC500 (Beckman Coulter, Deutschland) or a CyFlow ML (Partec, Deutschland) platform were used. In case of the Cytomics FC500, the field stop was set on 1 - 19°, and the discriminator to reduce background noise was set on the side scatter (SS = 2). For all platforms, the fluorescence of the probes was excited with a laser at a wavelength of 488 nm and the emission was measured using a photomultiplier and a band pass filter of 525 ± 25 nm (Cytomics FC500) or 536 ± 40 nm (CyFlow ML). Samples without probes were measured to adjust the voltage of the photomultiplier to avoid the measurement of autofluorescent cells. A total of 10,000 (Cytomics FC500) or 100,000 (CyFlowML) events were collected in all runs.
Determination of the microbial metabolic activity
The low hybridization rate for bacteria in the UASS biogas reactor samples indicated that not all bacteria possessed the high metabolic activity essential for a strong fluorescence signal. Hence, the metabolic activity of the microbial cells needed to be evaluated. Therefore, the dehydrogenase activity was determined by incubation with 5-cyano-2,3-ditolyl tetrazolium chloride (CTC) according to the protocol of Preuss and Hupfer (1998)  based on a modified protocol of Rodriguez and co-workers (1992) .
This assay was tested with growth series of pure cultures of E. coli and C. thermocellum as well as with a time series of UASS reactor samples. Samples of the E. coli and C. thermocellum culture were taken every 3 h between 3 and 36 h of growth. Samples from UASS biogas reactor were taken 1, 3, 5, 7, 9, 20, and 22 h after last feeding.
From each sample, triplicates of 1 ml were inoculated with 100 μl of a 0.16% CTC solution and incubated at 37°C for 60 min with constant shaking at 450 rpm (Thermomixer comfort, Eppendorf, Germany) and at dark conditions. As negative controls, 1 ml triplicates of each sample were inactivated for 20 min at 95°C with constant shaking at 700 rpm (Thermomixer comfort, Eppendorf, Germany) and treated as described above. The CTC reaction was stopped by adding 10 μl 37% formaldehyde. From each sample, a dilution series (100-, 500- and 1000-fold) was performed with sterile water.
For microscopic quantification of active and inactive cells 10-well-slides were coated with an aqueous solution of 0.1% gelatin and 0.01% CrK (SO4). 10 μl of each sample dilution was added to the wells and dried by air at room temperature. Subsequently, 5 μl antifading reagent Citifluor A1 (PLANO GmbH, Wetzlar, Germany) was added to coat each well, and 0.2 μl of a 5 μM stock solution of SYTO60 were carefully injected into this drop. After 20 min incubation the samples were ready to use for microscopic analysis by confocal laser scanning microscopy (TCS SP5 II, Leica Microsystems, Germany) using LAS AF Leica software. Following system settings were used: scan mode xyz - pinhole 1.50 airy, Acusto-Optical Tunable Filter (AOTF) 514 nm (10%), AOTF 633 nm (10%); sequential scan settings for SYTO60 - 633 nm, photo multiplier tubes (PMT) 650–770 nm; sequential scan settings for CTC - AOTF 514 nm, PMT 570–640 nm. The settings for picture size, gain, and offset were varied during the experiment to reach best image resolution and fluorescence signal strength.
In addition, samples were analyzed by flow cytometry. The Cytomics FC500 platform was used with following settings: excitation of CTC fluorescence at 488 nm, photomultiplier wavelength 615–620 nm. All further details were as given above.