Characterization of the microbial communities in jet-loop (JACTO) reactors during aerobic olive oil wastewater treatment

ARTICLE IN PRESS

International Biodeterioration & Biodegradation 59 (2007) 226–233 www.elsevier.com/locate/ibiod

Characterization of the microbial communities in jet-loop (JACTO) reactors during aerobic olive oil wastewater treatment A. Euse´bioa, M. Mateusa, L. Baeta-Halla, M.C. Sa`a´guaa, R. Tenreirob, E. Almeida-Varaa, J.C. Duartea, a INETI, Departamento de Biotecnologia, Unidade de Monitorizac- a˜o e Ecotoxicidade, Estrada do Pac- o do Lumiar, 22, 1649-038 Lisboa, Portugal Universidade de Lisboa, Faculdade de Cieˆncias, Centro de Gene´tica e Biologia Molecular e Instituto de Cieˆncia Aplicada e Tecnologia, Edifı´cio ICAT, Campus FCUL, Campo Grande, 1749-016 Lisboa, Portugal

b

Received 18 March 2005; received in revised form 30 November 2005; accepted 21 November 2006 Available online 1 February 2007

Abstract The olive oil industry is one of the most typical and economically important Portuguese agro-industries, 29,900 tons of olive oil having been produced in 2002/2003. This industry generates large amounts of olive oil wastewaters (OOWW), which are difficult to degrade and thus cause a negative environmental impact. Jet-loop reactors (JACTO) developed and scaled-up by our group have been successfully used for biological treatment of winery and OOWW. This study aimed to determine the interactions of reactor hydrodynamics with microflora profiles during bio-treatment of OOWW. Bio-treatment was performed using a 20-dm3 JACTO bioreactor achieving a chemical oxygen demand (COD) and phenolic compounds removal rate of 70% at a hydraulic retention time of 12 days. Bio-treatment was scaled-up to 200-dm3 JACTO bioreactor, reaching 87% COD removal and 80% phenolic compounds removal. Microflora present on OOWW were identified on samples taken before, during, and at the end of bio-treatment. Identification of isolates was carried out at genus and/or species level. Samples from the bio-treatments did not show any fungi; most of the isolates belonged to the Bacillus genus (with predominance of Bacillus megaterium, Bacillus sphaericus, and Brevibacillus brevis). The good COD and phenolic compounds removal rate indicates that the microbial community selected during the treatment is well adapted to the stress conditions imposed by this special type of bioreactor. r 2006 Elsevier Ltd. All rights reserved. Keywords: Olive oil wastewater; Microbial community; Aerobic treatment; Jet-loop reactors

1. Introduction The olive oil industry is one of the most typical and economically important Portuguese agro-industries, 29,900 tons of olive oil having been produced in 2002/ 2003. This industry generates large volumes of dark wastewaters with high organic load and containing significant amounts of polyphenolic compounds, which are difficult to degrade. As a polluting source, olive oil wastewaters (OOWW) have existed for thousands of years. Their effects on the environment are more noticeable at present due to the remarkable increase of olive oil production, and also because the sensitivity of the public Corresponding author. Tel.: +351 21 092 4724; fax: +351 21 716 6966.

E-mail address: [email protected] (J.C. Duarte). 0964-8305/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.ibiod.2006.11.008

to environmental problems is much higher than in the past (Cabrera et al., 1996; Rozzi and Malpei, 1996). The olive oil sector was traditionally made up of a large number of small mills that used to discharge their effluents directly on the land and/or into receiving waters. However, due to environmental policies that are being implemented all over the world, industries are compelled to treat their effluents prior to discharge either into receiving waters or into municipal collectors. The average chemical composition of OOWW is water (83–92%), organic matter (4–16%) and minerals (1–2%) (Pozo et al., 2002; Tsioulpas et al., 2002). Due to their high organic load and content of phenolic compounds (10% of the total organic matter), OOWW have high chemical oxygen demand (COD) and high biochemical oxygen demand (BOD): These reach values of 80–200 and

ARTICLE IN PRESS A. Euse´bio et al. / International Biodeterioration & Biodegradation 59 (2007) 226–233

50–100 g l1, respectively (Balice et al., 1990; Rosario et al., 1999; Ergu¨der et al., 2000; Paixa˜o and Anselmo, 2002; Pozo et al., 2002). These values are around 200–400 times higher than those of a typical municipal sewage (Cossu et al., 1993) and, due to the seasonal nature of this industry, all the effluents are produced during a short period of time (November–March, in Portugal). These two characteristics prevent the disposal of OOWW into urban sewage treatment plants. The size of typical Portuguese olive oil mills (small and medium) and also their locations (they are very spread out geographically and in rural areas) are factors that impede the formation of associations to build joint small wastewater treatment plants (WWTP). Several physicochemical processes, including simple evaporation, reverse osmosis, and ultra-filtration, have been employed to detoxify these effluents, but these methods require costly investment and maintenance (D’Annibale et al., 1998) and are not compatible with the seasonal character of this agro-industry. Anaerobic bioreactors have been used to convert OOWW organic content into biogas that can be used for energy production. However, current methods for the anaerobic treatment are only applicable to highly diluted OOWW (Marques, 2001), and methanogenic bacteria are often inhibited by the high phenolic content of this type of effluent (D’Annibale et al., 1998). Several investigations have been carried out with fungi or bacteria, using either suspended or immobilized cultures (e.g., D’Annibale et al., 1998; Fadil et al., 2003; Fountoulakis et al., 2002; Kissi et al., 2001; Piperidou et al., 2000; Pozo et al., 2002). However, this type of process would be difficult for the olive oil producers due to the specialized knowledge needed to maintain the cultures and the inherent problems arising from the scale-up of the process. Feasible solutions to this environmental problem include aerobic treatments based on bioreactors that use the native effluent microbial consortia to degrade the polluting effluent charge. The use of jet aeration systems in the biological treatment of wastewaters, as a means of combining efficient oxygen transfer with high turbulent mixing, is a very promising technology. Jet-loop type reactors (JACTO) developed and scaled-up at INETI have been successfully used for biological treatment of winery and OOWW (Petruccioli et al., 2002; Duarte et al., 2004). This type of reactor has good oxygen transfer rates with low energy costs and, due to the reduced reactor volumes needed for treatment, small areas are required, resulting in significant savings in installation and maintenance costs. However, and in spite of the great expertise developed in bio-treatment systems, an adequate understanding of effluent microbial communities composition and of their synergies with reactor dynamics has not yet been achieved. Our previous results indicate that bioreactor performance could be linked to the microorganisms present and that a complete characterization of the microbial communities involved in the treatment would clarify its relationship with reactor efficiency (Euse´bio et al., 2004).

227

The study presented here aimed to characterize OOWW microflora present during the bio-treatments and to infer the interactions with reactor performance. 2. Materials and methods 2.1. Olive oil wastewater OOWW used to perform the bio-treatments were collected from one olive oil mill located at Bobadela, Oliveira do Hospital (north of Portugal) that works with a continuous olive oil extraction process. Effluent samples to be used for determination of crude OOWW physical–chemical parameters were kept in plastic bottles, frozen at 20 1C, until analysis. Samples for COD determination were acidified to pHo2 before storage. OOWW aliquots for microbial characterization were refrigerated at 4oC until use. Remaining OOWW was kept at room temperature (RT) and used to feed the bioreactors.

2.2. Determination of effluent parameters and analytical methods Samples were de-frozen at RT before analysis. Levels of COD, volatile suspended solids (VSS), and total suspended solids (TSS) were determined following Standard Methods for the Examination of Water and Wastewater (1998). Total phenolic compounds content was determined from 20-dm3 JACTO-treated samples using the Folin Index method as follows: 2 ml 1:50 diluted OOWW sample, 5 ml Folin-Ciocalteu reagent (10%), and 8 ml sodium carbonate (7.5%, w/v) were mixed in a 100 ml volumetric flask. The volume was completed with distilled water and left for 30 min. Distilled water was used as blank. Absorbance was determined at 700 nm against blank: Folin Index ¼ A700nm  1000. When samples of caffeic acid were used for calibration: Folin Index ¼ A700nm  20. For more relevant ‘‘mass balances’’ on the 200-dm3 JACTO reactor, the Singleton and Rossi method (Singleton and Rossi, 1965), which gives directly g l1 units, was preferred to determine total phenolic compounds. A linear correlation between these two methods was established using caffeic acid as a standard phenolic compound for calibration: Folin Index 100 corresponds to 3.1 g l1 of caffeic acid.

2.3. Bioreactor equipment OOWW bio-treatments were performed at lab-scale using the 20-dm3 jet-loop bioreactor prototype described in Euse´bio et al. (2004) and then scaled-up to a pilot bioreactor with a working capacity of 200 dm3 (JACTO). A schematic drawing of this reactor is shown in Fig. 1. The reactor had a cylindrical configuration, a plane bottom, and a cover to minimize the effect of foaming. The degassing tank was placed on the top of the cylindrical column. Airflow at the Venturi was 2 m3 h1, corresponding to 0.15 vvm (for a volume of 200 dm3). The aerated effluent passed from the column reactor to the degassing tank, from where it was recycled, passing again through the nozzle of the Venturi ejector into the column reactor. Probes for measurement of temperature, pH, and dissolved oxygen (%pO2) were placed on the top and side of the reactor, and a glass window was built in the cylindrical column, to allow visualization of biomass inside the reactor. The reactor was fed through a peristaltic pump from a reservoir containing the wastewater to be treated.

2.4. Reactor operation and control Bio-treatments were carried out using the 20-dm3 JACTO reactor and then scaled-up to the 200-dm3 JACTO reactor. Experiments with the 20dm3 JACTO reactor started under batch conditions for 20 days. The reactor regimen was then changed to continuous feeding, starting with a low flow rate of 132 ml h1 equivalent to a hydraulic retention time (HRT) of 6 days. The reactor worked continuously for 200 days under three different HRTs: 6 days at an effluent COD loading charge of 40 g l1, 3

ARTICLE IN PRESS 228

A. Euse´bio et al. / International Biodeterioration & Biodegradation 59 (2007) 226–233

drawn air

4

effluent inlet A

7

B

2 liquid recirculation

3

1 Effluent outlet

sludge recycle 5 6

Fig. 1. Schematic drawing of the pilot 200-dm3 JACTO bioreactor: 1, cylindrical column reactor; 2, degassing tank; 3, gravity lamellar decanter; 4, Venturi ejector; 5 and 6, centrifugal pump; 7, anti-foaming pump; A, temperature and pH electrodes; B, oxygen electrode.

days with an average loading charge of 40 g l1, and 12 days with an initial loading charge of 35.1 g l1. During the first two phases of the biotreatment (batch and HRT 6 day), the pH inside the reactor was adjusted to 7.0–7.5 by adding nitric acid, since in its absence pH values rose to 9.0 or more. Nitric acid was chosen for increasing nitrogen supply. The aeration rate was 0.7 vvm. Scale-up of OOWW bio-treatment was performed inoculating the 200dm3 JACTO bioreactor with the biomass grown during the lab scale treatment. OOWW was filtered through an ultrafiltration membrane (100 kDalton; 1.5 l vol) before reactor feeding. Bio-treatment was performed throughout a 220-day period using a working volume of 120 l and three different feeding flow rates: 5 l day1 (on 22nd day of bio-treatment), 10 l day1 (on 110th day of bio-treatment) and 20 l day1 (on 134th day of bio-treatment), corresponding to HRT of 20, 10, and 5 days, respectively. As in the 20-dm3 JACTO reactor, the pH inside the 200-dm3 JACTO reactor was adjusted to 7–7.5 by addition of nitric acid during the bio-treatment and until its end. Temperature, pH, and dissolved oxygen O2 (%pO2) were monitored on-line and registered daily during both treatments, using probes: the Mod 7F Modular Control Unit (SGI—INCELTECH) for temperature, Ingold electrode (Mod 7F Modular Control Unit from SGI—INCELTECH) for dissolved O2 (%pO2), and Ingold electrode (Mod 7) for pH.

2.5. Microbial counts and identification Samples from crude and treated OOWW were collected under aseptic conditions and used to perform microbial counts, isolation and identification of the microflora present. Microbial colonies were counted as colony forming units (CFU) using the spread plate method, as described in Standard Methods for the Examination of Water and Wastewater (1998). Bacterial and fungal isolation was carried out on either Plate Count Agar (PCA, Difco) or Cook Rose Bengal Agar (CRBA, Difco), incubated at 30 1C. The characterization and identification of the isolated bacterial

strains were based on colony morphology, Gram staining, spore formation, catalase and cytochrome oxidase enzyme activities, and the API system (API 20 NE, API 20 E and API 50 CHB; BioMe´rieux vitek, Inc. France). Fungi were identified to genus level using identification keys (Barnett and Barry, 1972).

3. Results and discussion 3.1. Characterization of crude and treated OOWW Physicochemical characteristics were determined on crude OOWW: pH 4.6; COD 80.1 g l–1; total phenolic compounds content 75 (Folin Index). OOWW native microflora were characterized. Table 1 shows the mean colony counts of all microorganisms isolated and identified in the crude OOWW sample. Although COD content of this particular crude OOWW was 80 g l1, when the continuous feeding bio-treatment began, the COD content had dropped to 40 g l1 due to the natural chemical and microbiological degradation occurring during storage (in a 800-dm3 container kept outdoors). This fact must be considered normal from a technological point of view, since storage of the effluent must be carried out in a reservoir to be used for continuous feeding of the treatment reactor. However, this decrease of COD levels is related to the elimination of organic from easily degradable compounds (sugars, some lipids, etc.). It must also be noted that most modern continuous mills produce effluents with COD levels in the order of 25–50 g l1.

ARTICLE IN PRESS A. Euse´bio et al. / International Biodeterioration & Biodegradation 59 (2007) 226–233

3.2. OOWW bio-treatment

229

PCA—Plate Count Agar medium; CRBA—Cook Rose Bengal medium; CFU—colony forming units.

The reactor worked through a period of 200 days without interruptions under three different HRTs: 6 days at an effluent COD loading charge of 40 g l1, 3 days with an average loading charge of 40 g l1, and 12 days with an initial loading charge of 35 g l1. Temperature, pH, and dissolved oxygen O2 (%pO2) were measured on-line and registered daily (Fig. 2). Temperature was maintained at a medium value of 35 1C using a simple refrigeration system consisting of an external refrigeration circuit of tap water. Jet-loop aeration was enough to keep dissolved oxygen values within a range of 70–80% of saturation, showing the good aeration transference ability of the JACTO bioreactor. COD and phenolic compound profiles, determined under the three different HRTs during the bio-treatment, are shown in Fig. 3. During the first 10 days, CODout increased due to the change from batch to continuous operation. After this stabilisation period, CODout started to decrease, attaining values of 17.4 g l1. An efficiency of about 60% was reached for COD removal under the 6-day HRT. The change of HRT to 3 days led to a significant

40

100

35

90

As described in the Section 2, OOWW bio-treatments were started on a 20-dm3 JACTO bioreactor using the native microflora as microbial inoculum. After the start-up period, under batch conditions, the reactor regimen was then changed to continuous feeding starting with a low flow rate of 132 ml h1.

Table 1 Characterization of the microbial population isolated and identified in crude OOWW Culture medium

CFU ml1

Identification

PCA

0.95  105 2.04  105 1.13  105 0.89  105

Bacillus sphaericus Burkholderia cepacia Stenotrophomonas maltophilia Aspergillus sp.

80

30

70

25

60

20

50

15

40 30

10

Oxygen (%pO2)

pH;Temperature (°C)

CRBA

20

5

10

0

0 0

20

40

60

80 pH

100 120 Time (days)

Temperature

140

160

180

200

%pO2

Fig. 2. Data acquisition during olive oil wastewater bio-treatment, using the lab-scale 20-dm3 JACTO bioreactor prototype: temperature, pH, and dissolved oxygen.

HRT=6d

50

HRT=3d

HRT=12d

90

COD (g/l)

70 60

30

50 40

20

30 20

10

Phenols (Folin index)

80 40

10 0

0 0

20

40

60

COD in

80

100 120 Time (days)

COD out

Phenols in

140

160

180

200

Phenols out

Fig. 3. Time course of olive oil wastewater treatment, using the lab-scale 20-dm3 JACTO bioreactor prototype: COD and total phenol content measured at the inlet and outlet.

ARTICLE IN PRESS 230

A. Euse´bio et al. / International Biodeterioration & Biodegradation 59 (2007) 226–233

COD and phenolic compound profiles determined under the three different HRTs tested during bio-treatment are shown in Fig. 5. The OOWW detoxifying capacity of the JACTO system was tested with a feeding flow rate of 20 l day1 with an initial COD loading charge of 22.9 g l1. A range of COD values, varying between 3.98 and 2.4 g l1, was achieved during bio-treatment, corresponding to a conversion of 87%. Phenolic compound removal efficiency was 70–80%.

40

100

35

90

3.3. Monitoring of the microbial consortia during biotreatments Isolates from the bio-treatment system were identified at genus or species level using conventional microbiology identification methods. Results are shown in Table 2. A higher diversity of bacterial species was observed compared with native OOWW microflora consortium; no fungal isolates (either filamentous fungi or yeasts) were detected. At the start-up of the bio-treatment, using the 20-dm3 JACTO and under batch conditions, five different

80

30

70

25

60

20

50

15

40 30

10

Oxygen (%pO2)

pH;Temperature (°C)

decrease of COD removal efficiency (30%). After 126 days of bio-treatment a third HRT was tested (12 days) with an initial COD loading charge of 35.1 g l1. A range of COD values varying between 12 and 5 g l1 was achieved during bio-treatment with a conversion of about 70%. Phenolic compounds removal reached an efficiency of 60–70% at HRTs of 6 and 12 days, decreasing to 10% at HRT 3 days. Bio-treatment scale-up was performed inoculating the 200-dm3 JACTO bioreactor with the biomass coming from the 20-dm3 JACTO reactor. Bio-treatment was performed throughout 207 days using a work volume of 120 l and starting under batch conditions (25 days). Reactor running conditions were then changed to continuous feeding and three different feeding flow rates were tested: 5 l day1 (at 25th day of bio-treatment), 10 l day1 (at 110th day of biotreatment), and 20 l day1 (at 134th day of bio-treatment), corresponding to HRTs of 20, 10, and 5 days, respectively. Temperature was self-maintained between 20 and 35 1C, pH was buffered at 7–8 with nitric acid, and dissolved oxygen values varied between 60% and 90% of air saturation (Fig. 4).

20

5

10

0

0 0

30

60

90 120 Time (days) pH

Temperature

150

180

210

%pO2

Fig. 4. Data acquisition during olive oil wastewater bio-treatment, using the pilot-scale 200-dm3 JACTO bioreactor prototype: temperature, pH, and dissolved oxygen.

Batch

30

HRT = 20 d

HRT = 10 d

HRT = 5 d

2.5 2

20 1.5 15 1 10

Total Phenol (g/L)

COD (g/L)

25

0.5

5 0

0 0

30

60 COD in

90 120 Time (days) COD out

Phenol in

150

180

210

Phenol out

Fig. 5. Time course of olive oil wastewater treatment, using the lab-scale 200-dm3 JACTO bioreactor prototype: COD and total phenol content measured at the inlet and outlet effluent.

ARTICLE IN PRESS A. Euse´bio et al. / International Biodeterioration & Biodegradation 59 (2007) 226–233

231

Table 2 Characterization and heterotrophic plate counts of microbial consortia during olive oil wastewater bio-treatments using the lab- and the pilot-scale JACTO bioreactors, under several hydraulic conditions Batch/HRT identification (CFU ml1  103)

Aer. salm. salmonicida Alcaligenes xylosoxidans Bacillus megaterium Bacillus megaterium 2 Bacillus sphaericus Bacillus stearothermophilus Bacillus subtilis Brevibacillus brevis Chryseobacterium sp. Com. testo./Ps. alcaligens Ewingella americana Pseudomonas fluorescens Sphingobacterium multivorum Stenotrophomonas maltophilia

20 dm3 JACTO

200 dm3 JACTO

Batch

3 days

12 days

20 days

10 days

5 days

— — 5.0 — 0.3 — 4.5 48.0 — — 17.5 — — —

12.5 — — 0.5 — — — — — — — — — 0.2

— — 2.0 1.5 0.2 0.5 — — — — 1.0 — — —

— — — 40.2 0.3 — — 22.9 38.7 — 6.0 45.8 13.6 7.7

— — — 0.3 0.1 — — 0.9 12.2 — — — — —

— 52.6 — 22.4 16.0 — — — — 22.9 — — — —

bacteria were isolated, revealing a predominance of grampositive species, mainly Brevibacillus brevis (63.7%). A significant shift in microbial community structure from gram-positive bacteria to gram-negative was observed when reactor conditions were changed to continuous feeding flow rate. In the sample collected during the 3 days HRT a predominance of Aeromonas salmonicida salmonicida (88%) was observed, indicating a dominance of gram-negative bacteria (89%) over gram-positive bacteria (11%). Bacillus megaterium 2 and Stenotrophomonas maltophilia were also observed. The presence of S. maltophilia can be associated with polyphenol biodegradation, as verified by Lee et al. (2002), who described phenol degradation by S. maltophilia at a broad pH range (5–8), which is consistent with the pH of the JACTO bioreactor bio-treatment conditions. During the 12 days HRT a very different microbial community was isolated, showing predominance of gram-positive bacteria (81%): Bacillus megaterium, Bacillus megaterium 2, Bacillus sphaericus, Bacillus stearothermophilus, and Ewingella americana. The overall result of microbial characterisation during treatment (using a 20-dm3 JACTO bioreactor) indicates the prevalence of gram-positive bacteria (67%), with the predominance of Brevibacillus brevis (51%), as opposed to gram-negative bacteria, represented by Ewingella americana (20%) and Aeromonas salmonicida salmonicida (13%). Characterisation of the microbial population isolated after the scale-up to a 200-dm3 JACTO bioreactor revealed a very different microbial consortium in terms of genus diversity, with the appearance of five different species that were not detected during the lab-scale treatment. At an HRT of 20 days, Pseudomonas fluorescens predominated at the beginning of the experiment; this later shifted to a predominance of Bacillus megaterium 2 and Chryseobacterium sp., but gram-negative bacteria (64%) prevailed

over gram-positive (36%). When the HRT changed to 10 days, CFU ml1 of Bacillus megaterium 2 and Chryseobacterium sp. decreased, but the predominance of gramnegative bacteria (87%) over gram-positive bacteria (13%) was maintained. With an HRT of 5 days, Chryseobacterium sp. was no longer found. Bacillus megaterium 2, Bacillus sphaericus, and Comamonas testosteroni/Ps. alcaligenes were present throughout this HRT. At the end of the bio-treatment (day 143), Alcaligenes xylosoxidans was isolated at a significant concentration, although it was not found in early samples. The predominance of gramnegative bacteria (67%) over gram-positive bacteria (33%) was maintained until the end of the bio-treatment. According to the results, dominance of gram-positive bacteria in the 20-dm3 JACTO bioreactor (67%) was changed to dominance of gram-negative bacteria in the 200-dm3 JACTO bioreactor (66%). Partially anoxic conditions, favouring denitrification, may occur in zones inside the reactor, such as in the layers of the biomass attached to the inner wall of the reactor or inside the suspended flocs. Under these conditions, some heterotrophic denitrifying bacteria, such as Alcaligenes, Bacillus, and Pseudomonas, can use nitrate as an alternative electron acceptor (Metcalf and Eddy, 2003; Knowles, 1982). These nitrate reducing and denitrifying bacteria may find advantageous conditions in such partially anoxic niches, from the additional source of nitrate provided by supplementation with nitric acid used to adjust the pH. This is in line with the observation (data not shown) of a strong decrease of N–NO3 during the last regimen (dilution rate) used when higher feeding rates were tested. From these results, and taking into account the problems encountered to screen (either quantitatively or qualitatively) all microorganisms present in these communities by conventional microbiological methods, it is difficult to establish a clear correlation between the microbial

ARTICLE IN PRESS 232

A. Euse´bio et al. / International Biodeterioration & Biodegradation 59 (2007) 226–233

consortium composition and the OOWW treatment. Nevertheless, it could be seen that five sub-populations (Bacillus megaterium 2, Bacillus sphaericus, Brevibacillus brevis, Ewingella Americana, and S. maltophilia) prevail in the microflora consortium during the 200 days of OOWW bio-treatment, using the 20-dm3 JACTO bioreactor, and that these could also be detected during the scale-up treatment (200-dm3). The changes in hydraulic conditions during both treatments did not affect the predominance of these sub-populations. Although degradation of phenolic compounds can give rise to other compounds that have antibacterial activity (Aggelis et al., 2003), our results cannot confirm this fact; no significant decrease was observed in bacterial counts, either during the OOWW bio-treatment carried out using the 20-dm3 JACTO bioreactor or in the 200-dm3 JACTO bioreactor. In fact, these strains must be highly resistant to any antibacterial properties of the polyphenol-derived compounds, since they are isolated from quasi-steady-state regimens in the reactors. However, we cannot exclude the possibility that some of these compounds are responsible for exerting some selective pressure on the isolated populations. It seems that, besides the mechanical stress imposed by the JACTO reactor aeration system that may be responsible for the elimination of fungi and filamentous bacteria, no other selection criteria can be inferred from these results to explain the actual diversity obtained. 4. Conclusions The microbial population developed during these biotreatments shows high adaptation to the JACTO bioreactor conditions, as well as its efficiency with respect to both COD and phenolic compound degradation and removal. For the 20-dm3 JACTO bioreactor, a HRT of 12 days was found to be the best condition tested with an initial loading charge of 35.1 g l1 day1 and a COD reduction yield of about 70%. This result was in fact much improved for the 200-dm3 JACTO bioreactor, where the short retention time tested (an HRT of 5 days) was found to be the best, with an initial loading charge of 22.9 g l1 day1. The CODout values varied between 3.98 and 2.4 g l1 during the aerobic bio-treatment, with a COD reduction of about 87% and efficiency between 72% and 84% for phenolic compound removal. Native OOWW microflora were used in these studies as an inoculum for bioreactor treatment. It was expected that selection pressure for some bacterial species would be imposed by the conditions in the bioreactor, resulting in differences in community composition. However, for both bio-treatments with the microbial population found in crude OOWW, the bacterial isolates belonged mostly to the genus Bacillus and no fungi were detected. The only persistent strain was Bacillus sphaericus that was present in crude OOWW and could be isolated during both lab-scale and pilot treatments. The predominance of this microorganism suggests that this sub-population is the best

suited to resist the shear forces generated at the nozzle of the bioreactor. Four other microorganisms (Bacillus megaterium 2, Brevibacillus brevis, Ewingella Americana, and S. maltophilia) prevailed in the microflora consortium, showing a high potential for OOWW bio-treatment with COD and phenolic compound removals of 87% and 84%, respectively. A similar selection of sub-populations was observed during our previous studies on treatment of winery wastewaters with a JACTO bioreactor (Euse´bio et al., 2004). In this case also the surviving populations detected at the end of the long-treatment belonged predominantly to the genera Bacillus and Pseudomonas. Isolation and identification of the cultivable microorganisms present throughout the JACTO bio-treatments provided a partial knowledge of the microbial consortia composition. However, the well-stated inability to cultivate the major fraction of the organisms, found by direct cell counts, and in addition non-cultivable microorganisms, hampers complete characterization. A complete analysis of the microbial communities present during the biotreatments, their structural shifts, and their individual contributions to the overall treatment will allow a more complete control of reactor performance and thus the increase of treatment efficiency. Acknowledgements This work was funded through the European INCO-MED Programme (Project ICA3-CT-1999–00010: Mediterranean usage of biotechnological treated effluent water—Medusa Water; 2001–2004) and the Portuguese FCT-SAPIENS Programme (Project: The Jet-Loop Reactor: A New System for the Treatment of AgroIndustrial Effluents; 2003–2005). References Aggelis, G., Iconomou, D., Christou, M., Bokas, D., Kotzailias, S., Christou, G., Tsagou, V., Papanikolaou, S., 2003. Phenolic removal in a model olive oil mill wastewater using Pleurotus ostreatus in bioreactor cultures and biological evaluation of the process. Water Research 37, 3897–3904. D’Annibale, A., Crestini, C., Vinciguerra, V., Sermanniet, G.G., 1998. The biodegradation of recalcitrant effluents from an olive mill by a white-rot fungus. Journal of Biotechnology 61, 209–218. Balice, V., Carrieri, C., Cera, O., 1990. Caratteristiche analitiche delle acque divegetazione. Rivista Italiana Sostanze Grasse 67, 9–16. Barnett, H.L., Barry, B.H. (Eds.), 1972. Ilustrated Genera of Imperfect Fungi, third ed. Burgess Publishing Company, USA. Cabrera, F., Lo´pez, R., Martinez-Bordiu´, A., Dupuy de Lome, A., Murillo, J.M., 1996. Land treatment of olive oil mill wastewater. International Biodeterioration & Biodegradation 38, 215–225. Cossu, R., Blakey, N., Cannas, P., 1993. Influence of co-disposal of municipal solid waste and olive vegetation water on the anaerobic digestion of a sanitary landfill. Water Science and Technology 27, 261–271. Duarte, J.C., Mateus, M., Almeida-Vara, E., Euse´bio, A., Sena-Martins, G., Seabra, S., Ferreira, A., 2004. Compact aerobic bioreactor for treatment of typical mediterranean agro-industries effluents. Journal of Catalytic Materials and Environment IV, 77–81.

ARTICLE IN PRESS A. Euse´bio et al. / International Biodeterioration & Biodegradation 59 (2007) 226–233 Ergu¨der, T.H., Gu¨ven, E., Demirer, G.N., 2000. Anaerobic treatment of olive mill wastes in batch reactors. Process Biochemistry 36, 243–248. Euse´bio, A., Petruccioli, M., Lageiro, M., Federici, F., Duarte, J.C., 2004. Microbial characterisation of activated sludge in jet-loop bioreactors treating winery wastewaters. Journal of Industrial Microbiology and Biotechnology 31, 29–34. Fadil, K., Chahlaoui, A., Ouahbi, A., Zaid, A., Borja, R., 2003. Aerobic biodegradation and detoxification of wastewaters from the olive oil industry. International Biodeterioration & Biodegradation 51, 37–41. Fountoulakis, M.S., Dokianakis, S.N., Kornaros, M.E., Aggelis, G.G., Lyberatos, G., 2002. Removal of phenolics in olive mill wastewaters using the white-rot fungus Pleurotus ostreatus. Water Research 36, 4735–4744. Kissi, M., Mountadar, M., Assobhei, O., Gargiulo, E., Palmieri, G., Giardina, P., Sannia, G., 2001. Roles of two white-rot basidiomycete fungi in decolorisation and detoxification of olive mill wastewater. Applied Microbiology and Biotechnology 57, 221–226. Knowles, R., 1982. Denitrification. Microbiological Reviews 46, 43–70. Lee, E.Y., Jun, Y.S., Cho, K.-S., Ryu, H.W., 2002. Degradation characteristics of toluene, benzene, ethylbenzene, and xylene by Stenotrophomonas maltophilia T3-c. Journal of the Air & Waste Management Association 52, 400–406. Marques, I.P., 2001. Anaerobic digestion treatment of olive mill wastewater for effluent re-use in irrigation. Desalination 137, 233–239. Metcalf and Eddy, Inc., 2003. Wastewater Engineering: Treatment and Reuse, fourth ed. McGraw-Hill, NY, USA. Paixa˜o, S.M., Anselmo, A.M., 2002. Effect of olive mill wastewaters on the oxygen consumption by activated sludge microorganisms: an acute toxicity test method. Journal of Applied Toxicology 22, 173–176.

233

Petruccioli, M., Duarte, J.C., Euse´bio, A., Federici, F., 2002. Aerobic treatment of winery wastewater using a jet-loop activated sludge reactor. Process Biochemistry 37, 821–829. Piperidou, C.I., Chaidou, C.I., Stalikas, C.D., Soulti, K., Pilidis, G.A., Balis, C., 2000. Bioremediation of olive oil mill wastewater: chemical alterations induced by Azotobacter vinelandii. Journal of Agricultural and Food Chemistry 48, 1941–1948. Pozo, C., Martı´ nez-Toledo, M.V., Rodelas, B., Gonza´lez-Lo´pez, J., 2002. Effects of culture conditions on the production of polyhydroxyalkanoates by Azotobacter chroococcum H23 in media containing a high concentration of alpechı´ n (wastewater from olive oil mills) as primary carbon source. Journal of Biotechnology 97, 125–131. Rosario, L., Ana, R., Bernardo, M., Gerardo, A.C., Antonio, G., 1999. Biodegradation of olive mill wastewaters (OMW) by strains of Scopulariopsis spp. isolated from OMW-contaminated habitats. Toxicology and Environmental Chemistry 72, 127–134. Rozzi, A., Malpei, F., 1996. Treatment and disposal of olive mill effluents. International Biodeterioration & Biodegradation 38, 135–144. Singleton, V.L., Rossi Jr., J.A., 1965. Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents. American Journal of Enology and Viticulture 16, 144–158. Standard Methods for the Examination of Water and Wastewater, 1998. American Public Health Association/American Water Works Association/Water Environment Federation (Eds.), 20th ed. Washington, DC, USA. Tsioulpas, A., Dimou, D., Iconomou, D., Aggelis, G., 2002. Phenolic removal in olive oil mill wastewater by strains of Pleorotus spp. in respect to their phenol oxidase (laccase) activity. Bioresource Technology 84, 251–257.