FEMS Microbiology Ecology 48 (2004) 369–378

www.fems-microbiology.org

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

Reactor performance and microbial community dynamics
during anaerobic biological treatment of wastewaters at 16–37 °C
Sharon McHugh *, Micheal Carton, Gavin Collins, Vincent O’Flaherty
Microbial Ecology Laboratory, Department of Microbiology, National University of Ireland, Galway, Ireland
Received 16 April 2003; received in revised form 14 October 2003; accepted 20 February 2004

First published online 9 April 2004

Abstract

The anaerobic biological treatment of volatile fatty acid (VFA) – and sucrose – based wastewaters was investigated in two an-
aerobic bioreactors, R1 and R2, over a 300-day trial period. During the trial, the operating temperature of both reactors was lowered,
in a stepwise fashion, from 37 to 16 °C. The VFA-fed reactor maintained an excellent level of performance, regardless of operating
temperature, reaching COD removal efficiencies of 95% at 18 °C, and a biogas methane content in excess of 70% at 16 °C, at an
imposed OLR of 20 kg COD m
3 d
1 . However, an increase in the applied liquid upflow velocity to the bottom chamber of the reactor
from 5 to 7.5 m h
1 on day 236 resulted in a considerable decline in reactor performance. COD removal efficiencies in excess of 80%
were achieved by the sucrose-fed reactor at 18 °C, at an imposed OLR of 20 kg COD m
3 d
1 . An increase in the liquid upflow velocity
applied to the sucrose-fed reactor resulted in enhanced reactor performance and stability, with respect to decreasing temperature. The
different responses of both reactors to increased upflow velocity was associated with variations in the microbial population structure
of the sludges, as determined by culture-independant molecular approaches, specifically the presence of high levels of d-Proteobacteria
and hydrogenotrophic methanogens in the VFA-fed biomass. High levels of Methanomicrobiales sp., in particular Methanocor-
pusculum parvum sp., were observed in both R1 and R2 during the trial. There was a distinct shift from acetoclastic methanogenic
dominance to hydrogenotrophic dominance in both reactors in response to a decrease in the operating temperature.
Ó 2004 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.

Keywords: Psychrophilic anaerobic digestion; 16S rRNA genes; Microbial community dynamics; T-RFLP; Hydrogenotrophic methanogens

1. Introduction further bioengineering innovations, which will allow for

the implementation of psychrophilic anaerobic digestion
Psychrophilic anaerobic digestion is an attractive on a more global basis for the treatment of a wide va-
option for the treatment of wastewaters discharged at riety of wastewaters.
low or ambient temperatures, offering technical and In particular, advanced research into the structure,
economical benefits over more conventional treatment function and biological properties of the microbial
processes [1,2]. The potential drawbacks previously as- communities involved in psychrophilic anaerobic diges-
sociated with low temperature reactor operation, such as tion is essential, in order to fully understand the mi-
low biogas production rates and poor stability, have crobial ecology of the process and to allow future
been largely overcome by the introduction of novel re- improvement and optimisation of the technology. The
actor designs, such as the expanded granular sludge bed success, or otherwise of an anaerobic treatment system is
(EGSB) reactor [3], and improved modes of operation. entirely dependant on the actions and interactions of
However, much scope still exists for the application of complex microbial communities. However, many of the
current bioreactor designs and operating routines have
*
Corresponding author. Tel.: +353-91-524411x3783; fax: +353-91- very little, if any, support from a microbiological point
525700. of view [4]. Furthermore, at present very little is known
E-mail address: sharon.mchugh@nuigalway.ie (S. McHugh). about low temperature methanogenesis or psychrophilic

0168-6496/$22.00 Ó 2004 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.femsec.2004.02.012
370 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378

methanogenic communities. Anaerobic microbial deg- in the Netherlands (Paques B.V.), to a final sludge
radation of organic matter has been reported at tem- concentration of 13.9 g volatile suspended solids (VSS)
peratures as low as 2 °C [5], however conflicting reports l
1 . Prior to the commencement of this study, both R1
have been obtained regarding the pathways of degra- and R2 were operated at 37 °C for 200 days. The op-
dation under low temperature conditions and the mi- erational data for this period are described in detail in a
crobial trophic groups involved [6,7]. The application of separate paper by the same authors (unpublished data;
nucleic acid-based molecular techniques to the study of manuscript submitted). Throughout this initial 200-day

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

the microbial populations involved in psychrophilic period, R1 was fed the VFA-based wastewater for 177
anaerobic digestion should provide a greater insight into days and the sucrose-based wastewater for the final 23
the process, aiding the successful implementation of this days, with the opposite feeding regime employed for R2.
treatment technology.
In the present study, the feasibility of psychrophilic 2.2. Determination of specific methanogenic activity
anaerobic digestion was investigated by monitoring the
effect of temperature decrease on reactor performance Maximum specific methanogenic activity (SMA)
and sludge biomass activity in two anaerobic bioreactors profiles of sludge samples taken from R1 and R2 on
over a 300-day period. In addition, the microbial pop- days 165 and 300, and also from the biomass removed
ulation structure and dynamics of both reactors during from the fixed-biofilm section of each reactor on day
the trial were investigated using molecular microbial 300, were determined using the pressure transducer
ecology approaches. technique [1,10]. Briefly, the test procedure involved the
measurement of the biogas pressure increase developing
in sealed vials fed with the non-gaseous substrates,
2. Materials and methods ethanol (30 mM), propionate (30 mM), butyrate (15
mM) and acetate (30 mM), or of the pressure decrease in
2.1. Reactor design and experimental setup vials pressurized with H2 /CO2 (80:20) to 1 atm. Tests
were carried out, in triplicate, at 37, 22 and 15 °C.
Two 3.8 l laboratory-scale upflow reactors, R1 and
R2, were used in this study. An effluent recirculation 2.3. DNA extraction
facility was applied to the lower chamber of each reactor
to give a liquid upflow velocity of 3.5 m h
1 in this Sludge samples were initially crushed with a pestle
section. This was increased for R1 and R2 to 5.0 and 7.5 and DNA extracted from the crushed sludge, in tripli-
m h
1 on days 118 and 236 of the study, respectively. cate, using the MoBio Soil DNA extraction kit (Cam-
The top chamber of each reactor was designed as a bio, Cambridge, UK), according to the manufacturer’s
fixed-film section, comprising randomly packed poly- instructions. Cell lysis efficiency was determined using a
ethylene ring-shaped matrix pieces. The initial temper- modified version of the method of Bitton et al. [11], as
ature of both reactors was 37 °C and this was decreased described previously [1].
during the trial to 30, 25, 20, 18 and 16 °C on days 45,
98, 167, 184 and 198, respectively. R1 was fed a sucrose- 2.4. Generation of 16S rDNA clone libraries, amplified
based wastewater, prepared by dissolving table sugar rDNA restriction analysis (ARDRA) and phylogenetic
(>99.9% sucrose; Irish Sugar Ltd., Co. Carlow, Ireland) classification
in tap water to a final concentration of 10 g chemical
oxygen demand (COD) l
1 . R2 was fed a synthetic Archaeal and bacterial clone libraries were generated
wastewater, consisting of acetate, ethanol, butyrate and by PCR amplification of 16S rRNA genes using the
propionate, at a COD ratio of 1:1:1:1 to a total of 10 archaeal primers 21F (50 -TTCCGGTTGATCCYG
g COD l
1 . The wastewaters were buffered with CCGGA-30 ) [12] and 958R (50 -YCCGGCGTTGAM
NaHCO3 (10 g l
1 ) and supplemented with macro- (10 TCCAATT-30 ) [13] and the bacterial primers 27F (50 -
ml l
1 ) and micro- (1 ml l
1 ) nutrients, as recommended GAGTTTGATCCTGGCTCAG-30 ) [13] and 1392R (50 -
by Shelton and Tiedje [8]. Throughout the trial, both R1 ACGGGCGGTGTGTRC-30 ) [14]. Reaction mixtures
and R2 were operated at a hydraulic retention time (50 ll) contained NH4 buffer (16 mM (NH4 )2 SO4 , 67
(HRT) of 12 h, corresponding to an organic loading rate mM Tris–HCl (pH 8.8 at 25 °C), 0.01% Tween-20), 200
(OLR) of 20 kg COD m
3 d
1 . Samples of reactor ef- lM dNTP (dATP, dCTP, dGTP, dTTP), 100 ng of each
fluent and biogas were routinely taken for volatile fatty primer, 100 ng template DNA, 0.5 U Taq DNA poly-
acids (VFAs), ethanol, COD and CH4 determination, as merase and 3.0 mM MgCl2 (archaeal PCR) or 1.5 mM
described previously [9]. MgCl2 (bacterial PCR). The PCR conditions were initial
Both reactors were initially seeded with a 10-year-old denaturation at 95 °C for 5 min followed by 30 cycles of
unfed sludge, consisting of mesophilic granular sludges denaturing (95 °C, 1 min), annealing (55 °C – archa-
obtained from full-scale industrial wastewater digesters eal; 52 °C – bacterial, 1 min) and primer extension
 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378 371

(72 °C, 1 min) with a final extension of 72 °C for 10 min. TRF lengths were determined for each sequence ob-
PCR products were electropheresed on 0.8% agarose tained through clone library analysis by locating the
and bands of the correct size were cut out and eluted primer and enzyme restriction site within the retrieved
using Snap minicolumns (Invitrogen, Groningen, The sequences and counting the number of base pairs in the
Netherlands). terminal fragment. In addition, predicted TRF lengths
The purified DNA amplicons were ligated into pCRÒ of known archaeal and bacterial species were obtained
2.1-TOPO (Invitrogen) plasmids and the ligation prod- using the TAP-TRFLP tool of the RDP database [18].

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

ucts transformed into E. coli TOP 10 competent cells
using kanamycin selection. Plasmid inserts were ampli-
fied by PCR using the M13 PCR set [1]. Insert-con- 3. Results
taining clones were restricted using HaeIII, grouped into
operational taxonomic units (OTUs) on the basis of 3.1. Operational performance of R1 and R2
their restriction profiles and representatives from se-
lected OTUs were chosen for gene sequencing. Template During the initial 45 days of this study, a steady in-
DNA was prepared from overnight cultures of selected crease in COD removal efficiencies, from 60% to 90%,
clones using an alkaline miniprep kit (Qiagen, Heidel- was observed for R1, the sucrose-fed reactor (Fig. 1(a)).
berg, Germany). Sequencing was performed on a Licor During this period, the methane content of the biogas
gel sequencer using vector specific primers (MWG Bio- remained relatively stable at approximately 50% and
tech, Milton Keynes, UK). The resultant sequence data effluent VFA concentrations remained low (<500
were compared to nucleotide databases using basic local mg l
1 ). Following the initial decrease in operating
alignment search tool (BLASTn) as described previously temperature, from 37 to 30 °C on day 45, a drop in both
[15]. The presence of chimeric amplification products COD removal efficiencies (to 50–60%) and in the
was screened for using the Ribosomal Database Project methane content of the biogas was observed. This was
(RDP) Chimera_Check software package [16]. None associated with an immediate rise in effluent propionate
were present in the data generated from this study. Se- levels, which remained high until day 70, after which
quence data for the retrieved sequences was manually time a decrease in propionate concentration and a par-
aligned to sequences obtained from the RDP. The allel increase in acetate and butyrate levels was noted.
phylogenetic inference package Paup* 4.0b8 was used An increase in COD removal efficiency was also ob-
for all phylogenetic analyses [17]. The partial 16S rRNA served from day 70 for R1, with the system reaching
gene sequences determined in this study were deposited COD removal efficiencies in excess of 75% by day 75. A
in the Genbank database under accession numbers further temperature decrease (to 25 °C) on day 98 also
AY231301 – AY231364. resulted in a reduction in R1 reactor performance, with
a decrease in COD removal efficiency to 40% (Fig. 1(a))
2.5. Terminal restriction fragment length polymorphism and a reduction in the methane content of the biogas to
(T-RFLP) 20%, the lowest values obtained during the trial. The
decrease in reactor performance was associated with
PCR was carried out, as above, but with fluorescently elevated effluent concentrations of VFA, with acetate,
labelled forward and reverse primers. The forward propionate and butyrate concentrations of 1433, 1065
primers, 21F and 27F, were 50 -hexachloroflourescein and 286 mg l
1 , respectively, recorded on day 100. An
(HEX) labelled and the reverse primers, 958R and applied increase in the liquid upflow velocity of the
1392R, were 50 – [6] – carboxyflourescein (FAM) la- system to 5 m h
1 on day 118 resulted in a considerable
belled. All oligonucleotide primers were obtained from improvement in reactor performance with an increase in
Oswel, Southampton, UK. The PCR reactions were R1 COD removal efficiency to approximately 75%
carried out in triplicate for each sample and resulting (Fig. 1(a)). Lowering the operational temperature of R1
products grouped together, ethanol precipitated and to 20 °C, and then to 18 °C, on days 167 and 184, re-
resuspended in 20 ll of sterile water prior to restriction. spectively, did not appear to have a detrimental effect on
Restrictions were carried out separately with the tetra- reactor performance with COD removal efficiencies in
meric endonucleases HhaI and AluI for 6 h at 37 °C excess of 75% consistently achieved under these tem-
using 10 ll of DNA and 10 units of enzyme to produce a perature conditions (Fig. 1(a)). Moreover, the biogas
mixture of variable length end-labelled 16S rRNA gene methane content and effluent VFA concentrations re-
fragments. The lengths of the terminal restriction frag- mained stable in response to these temperature de-
ments (>50 bp) were determined by comparison with creases. After 28 HRTs at 18 °C, the operating
internal standards using an automated ABI PRISMTM temperature of R1 was lowered by a further 2–16 °C.
Genetic Analyzer (Oswel Genetic Analysis, UK) and This resulted in a considerable decline in R1 COD re-
GeneScanÒ 3.1 analysis software (Applied Biosystems, moval efficiency to 50–60% and an increase in effluent
Foster City, CA), as described previously [1]. Predicted VFA concentrations, particularly acetate. However,
372 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

Fig. 1. CODsol removal efficiencies achieved by R1 (a) and R2 (b) during the 300-day study. A – temperature decrease to 30 °C; B – temperature
decrease to 25 °C; C – liquid upflow velocity increase from 3.5 to 5.0 m h
1 ; D – temperature decrease to 20 °C; E – temperature decrease to 18 °C; F –
temperature decrease to 16 °C; G – liquid upflow velocity increase from 5.0 to 7.5 m h
1 .

following this perturbation, stable reactor performance to >70%, and it remained stable at this level for the
was achieved with a slow, steady increase recorded in remainder of the study, with the highest values obtained
COD removal efficiency, to levels of approximately 70% during operation at 16 °C (average – 74%). The increase
after a 30-day period (Fig. 1(a)). In an attempt to further in applied liquid upflow velocity on day 236 had a sig-
improve the treatment capacity of the reactor, the ap- nificant effect on R2 reactor performance, resulting in
plied liquid upflow velocity in the bottom chamber of considerable increases in effluent propionate, butyrate
R1 was increased to 7.5 m h
1 on day 236, resulting in and in particular, acetate levels, and a subsequent de-
an increase in both COD removal efficiency (Fig. 1(a)) cline in R2 COD removal efficiency to 40% (Fig. 1(b)).
and in the biogas methane content, with stable and ef- By day 480, however, the performance of R2 had re-
ficient performance levels achieved for the final 20 days covered with effluent VFA concentrations >1000 mg l
1
of the study. and stable COD removal efficiency of 80% (Fig. 1(b)).
The COD removal efficiency of R2, the VFA-fed re-
actor, remained very stable (<80%) for the initial 240 3.2. Specific methanogenic activity profiles of R1 and R2
days of the study, regardless of operating temperature, biomass samples
with the highest removal efficiencies achieved at 18 °C
(average – 94.5%; Fig. 1(b)). R2 effluent VFA concen- The seed sludge originally used to inoculate R1 and
trations remained low for the first 240 days of R2 op- R2 displayed low SMA values towards all substrates
eration, with the most prominent VFA, acetate, tested, with no propionate activity detected at 22 °C
generally detected at levels <800 mg l
1 , with propionate (Table 1), although considerable methanogenic activity
(<400 mg l
1 ) and butyrate (<100 mg l
1 ) also detected. developed within the biomass in both reactors over the
Furthermore, the methane content of the biogas pro- 200-day trial period prior to commencement of this
duced by R2 increased over the initial 60 days of the trial study. In the present study, R1 sludge bed biomass ex-
 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378 373

Table 1
Average SMA values for R1 and R2 biomass (ml CH4 g VSS
1 d
1 (STP); n ¼ 3)
Substrate/test Seed sludgea Day 165 Day 300 Biofilmb
temperature (°C) R1 R2 R1 R2 R1 R2
37 22 37 22 37 22 37 22 15 37 22 15 37 37
Acetate 7.9 0.9 104.3 31.0 255.6 170.8 32.2 21.2 14.7 151.7 59.7 26.4 16.8 15.5

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

Ethanol 10.9 1.1 197.1 133.4 1113.5 350.1 230.9 130.6 86.2 494.5 82.9 21.3 29.4 23.1
Butyrate 10.7 2.1 71.9 37.9 358.6 187.3 38.5 14.3 11.7 241.2 73.6 23.8 14.0 2.7
Propionate 6.4 ND* 13.2 ND* 135.8 72.3 11.1 10.6 ND* 83.7 57.1 11.5 13.2 6.5
H2 /CO2 27.8 1.5 212.5 109.9 174.3 89.6 298.5 171.3 101.8 234.7 168.6 78.7 35.4 27.9
ND*: no activity detectable.
a
Original seed sludge used to inoculate R1 and R2.
b
Biomass removed from the fixed-biofilm section of R1 and R2 on day 300.

hibited a higher methanogenic activity than R2 biomass tained at 37 °C was obtained at 22 °C, indicating the
with H2 /CO2 at all temperatures tested. R2 biomass, adaptation of the propionate-degrading bacteria within
conversely, exhibited higher activities for all soluble the sludge to the lower temperatures during the course
substrates, with the exception of ethanol at 22 and 15 °C of the study.
on day 300 (Table 1), demonstrating the influence of
feed composition on the methanogenic activity of an- 3.3. Archaeal population dynamics within R1 and R2
aerobic sludge. The biofilm sample removed from the biomass
upper chamber of R1 on day 300 exhibited a higher
activity than R2 on all substrates tested. The values for Analysis of the T-RFLP profiles of biomass samples
the SMA profiles of the fixed biofilm samples recorded removed from R1 and R2 on days 45 (operating tem-
at 37 °C were generally lower than those obtained from perature 37 °C), 142 (operating temperature 25 °C), 197
the sludge bed biomass (Table 1). A temperature opti- (operating temperature 18 °C) and 300 (operating tem-
mum of 37 °C, with respect to SMA, was recorded for perature 16 °C) revealed a comparable shift in the ar-
all R1 and R2 sludge samples tested, with lower activi- chaeal community structure within both bioreactors. In
ties recorded as the test temperature was decreased, R1, the initial predominant archaeal species were Met-
implying that a true psychrophilic biomass had not de- hanosarcina sp. (555; 137 – Table 2), although these were
veloped during the trial (Table 1). However, the devel- subsequently replaced as the major archaeal group by
opment of a psychrotolerant biomass was observed in Methanosaeta sp. (187; 582 – Table 2) by day 142.
both reactors, with activity at 15 °C detected in R1 and However, by day 197, and for the remainder of the
R2 biomass on day 300 of the trial. In addition, attempts study, members of the hydrogenotrophic Methanomi-
to set up an anaerobic bioreactor at 18–20 °C using the crobiales group (516; 122 – Table 2) were the predomi-
original seed sludge from this study as inoculum were nant methanogens detected in R1 biomass samples.
unsuccessful, with no significant COD removal or bio- Similarly for R2, the initially dominant Methanosarcina
gas production achieved over a six month trial period sp. were replaced by Methanomicrobiales sp. during the
(data not shown). The absence of propionate activity trial, with the population shift occurring by day 142,
from the sucrose-fed R1 biomass at 22 °C on day 165 which was earlier than that recorded for R1. The pre-
correlates with the results obtained during the previous dominance of Methanomicrobiales sp. in the final bio-
200-day trial period (unpublished data). However, by mass samples from R1 and R2 was also demonstrated
day 300 of the trial, a comparable activity to that ob- using clone library analysis, with clones showing a

Table 2
Length (bp) of the predominant peaks in the archaeal T-RFLP profiles generated from R1 and R2 biomass
Days R1 R2
Alu21F Alu958R Hha21F Hha958R Alu21F Alu958R Hha21F Hha958R
45 555 211 137 328 555 211 137 328
142 187 211 582 195 516 211 122 328
197 516 211 122 328 516 211 122 328
300 516 211 122 328 516 211 122 328
374 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378

strong similarity to Methanocorpusculum parvum ac- 3.4. Bacterial population dynamics within R1 and R2
counting for 84.2% and 90.6% of total clones analysed biomass
for R1 and R2 biomass, respectively. In total, 107 clones
were analysed and 10 OTUs were observed within each The bacterial populations present within R1 and R2,
library. In addition to M. parvum – like clones, se- as determined by T-RFLP analysis, appeared to be
quences closely related to Methanosaeta sp. (R1 – 12.7% considerably more diverse and dynamic than the ar-
of total clones analysed; R2 – 4.4% of total clones chaeal populations, even during periods of stable reactor

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

analysed) were also detected in both sludges. Clones performance, as reported in several previous studies
correlating to Methanosarcina sp. (3.2%) were detected [1,19]. Variations in bacterial population structure be-
in R1 sludge biomass only, while clones affiliated to tween reactors and sampling times were observed during
Methanospirillum hungatei (4.5%) were only detected in the trial, although no clear pattern could be discerned.
R2 biomass. Phylogenetic classification of the sequences Many recurring peaks were present, including 292 and
obtained with known sequences from the RDP is illus- 301 (Hha reverse), 339 and 349 (Alu reverse) and 240
trated in Fig. 2. (Alu forward) (Table 3). However, many of the peaks

Fig. 2. Phylogenetic classification of archaeal sequences obtained from R1 and R2 biomass, sampled on day 300. The tree was constructed with
evolutionary distances calculated based on the Kimura-2 model and the neighbour joining method of Saitou and Nei [26]. Three bacterial sequences
were defined as outgroups during phylogenetic reconstruction. Numbers at nodes represent bootstrap values (100 replicates).
 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378 375

Table 3
Length (bp) of the predominant peaks in the bacterial T-RFLP profiles generated from R1 and R2 biomass
Days R1 R2
Alu27F Alu1392R Hha27F Hha1392R Alu27F Alu1392R Hha27F Hha1392R
45 142 272 362 292 252 339 94 301
142 200 349 73 292 252 349 58 301
197 252 339 1093 301 609 339 357 301

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

300 240 339 377 301 240 339 94 301

within the bacterial TRF profiles had poor phylogenetic des sp. (30%), Spirochaeata sp. (6%), TMF phylum (8%)
resolution, with several distinct unrelated bacterial and 5% were closely related only to uncultured clones
groups generating TRFs of equal size and, therefore, the for which phylogenetic information is unknown, al-
sequences obtained from clone library analysis were though phylogenetic classification suggested that these
used to obtain more phylogenetically informative re- clones (R1_4; R1_20; R1_21; R1_24; Fig. 3) were affil-
sults. Of the clones analysed from the R1 library, 10 of iated with the Gram positive bacteria. Similarly, R2
the OTUs (51% of total clones analysed) were closely biomass samples contained sequences affiliated to Bac-
affiliated to the Gram positive bacteria, in particular the teroides sp. (33.4%), Gram positive bacteria (7.2%) and
Bacillus–Lactobacillus–Streptococcus group (Fig. 3). The Spirochaeta sp. (1.2%), as well as a large number of
remaining clones showed strong similarity to Bacteroi- unaligned clones (29.6%). However, additional clones

Fig. 3. Phylogenetic classification of bacterial sequences obtained from R1 biomass sampled on day 300. The tree was constructed with evolutionary
distances calculated based on the Kimura-2 model and the neighbour joining method of Saitou and Nei [26]. Three archaeal sequences were defined
as outgroups during phylogenetic reconstruction. Numbers at nodes represent bootstrap values (100 replicates).
376 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378

bacterial TRFs [22], subtle population changes within

bacterial groups and/or the presence of low numbers
of psychrophilic microorganisms may have been
overlooked.
A distinct shift in archaeal community structure was
observed in both R1 and R2 during the trial with a
proliferation of Methanomicrobiales sp. and a parallel

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

decrease in Methanosarcina sp./Methanosaeta sp. ob-
served. The reason for the shift in population structure
from acetate-utilising methanogens to hydrogen-utilis-
ing methanogens is unclear. However, it has been sug-
gested that in stressed systems, a syntrophic relationship
between an acetate-utilising organism and a hydrogen-
utilising methanogen serves as the major route of
methane production from acetate [23]. The combination
of high loading rate and low operating temperature may
have resulted in a stressed state within the reactors,
encouraging the propagation of hydrogenotrophic
methanogens. In a similar study by the same authors
[24], a similar proliferation of M. parvum-like organ-
isms, coupled with a decrease in the relative abundance
of acetoclastic methanogens, was observed in an EGSB
bioreactor treating whey-based wastewater at 12–20 °C,
following an increase in OLR. This further suggests the
contribution of low temperature and high loading rate
to the shift in archaeal population structure. The posi-
tive effect of low temperature on hydrogenotrophic
methanogens, due to the decrease in H2 threshold con-
Fig. 4. Phylogenetic classification of bacterial sequences obtained from centrations [6], may also have contributed to the emer-
R2 biomass sampled on day 300. The tree was constructed with evo- gence of Methanomicrobiales sp. during the trial, as may
lutionary distances calculated based on the Kimura-2 model and the
neighbour joining method of Saitou and Nei [26]. Three archaeal se-
have changes within the microbial composition of the
quences were defined as outgroups during phylogenetic reconstruction. higher trophic groups. Regardless of the reason to why
Numbers at nodes represent bootstrap values (100 replicates). the proliferation occurred, the results do show the im-
portance of hydrogenotrophic methanogens in the low
related to d-Proteobacteria (25.9%), green sulphur bac- temperature anaerobic degradation of organic matter.
teria (1.2%) and green non-sulfur bacteria (1.2%) were The increase in upflow velocity to R1 resulted in
also present (Fig. 4). improved performance, presumably due to enhanced
substrate-biomass contact within the reactor, and also
appeared to confer on the system a better ability to deal
4. Discussion with decreased temperature. However, the increase in
applied liquid upflow velocity (from 5.0 to 7.5 m h
1 ) to
Current economic, social and environmental trends R2 had an adverse effect on reactor performance, with
have necessitated the development of high-rate, low-cost a significant decrease in COD removal efficiencies and
waste treatment systems. In this study, psychrophilic an increase in effluent VFA concentrations observed.
anaerobic digestion was demonstrated to provide the The reason for the differing responses to this parameter
potential for such a system. Treatment efficiencies in change by R1 and R2 is likely due to variations within
excess of 95% and 80% were achieved at 18 °C and the microbial community structure of both reactors. A
imposed organic loading rates of 20 kg COD m
3 d
1 for considerable number of d-Proteobacteria sequences
VFA- and sucrose-based wastewaters respectively, re- were detected in the R2 bacterial clone library only,
sults which are comparable and, in some cases, superior with peaks corresponding to the reverse TRF lengths of
to those obtained at mesophilic temperatures [20,21]. No this group (338 – Alu, 301 – Hha) predominant in the
specific shift towards psychrophilic microorganisms was TRF profiles generated from R2 samples taken sub-
detected throughout the trial, with R1 and R2 biomass sequent to the upflow velocity increase. Members of
on day 300 displaying mesophilic temperature optima this group, such as Syntrophobacter wolinii, are known
against all substrates tested by SMA analysis. How- to be important in propionate degradation and con-
ever, due to the poor phylogenetic resolution of many comitant acetate production in anaerobic digesters. The
 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378 377

increased upflow velocity in R2 may have allowed for rate treatment systems and ultimately, the increased ef-
the proliferation of these acetogenic bacteria, due to ficiency and exploitation of anaerobic technologies.
increased substrate–biomass contact within the reactor.
This would account for the low levels of propionate
(<200 mg l
1 ) and butyrate (below the detection References
threshold) detected in R2 effluent following the applied
upflow velocity increase, and the subsequent rise in [1] Collins, G., Woods, A., McHugh, S., Carton, M.W. and O’Flah-

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

effluent acetate concentrations. The low density of erty, V. (2003) Microbial community structure and methanogenic
activity during start-up of psychrophilic anaerobic digesters
acetoclastic methanogens present within R2 biomass treating synthetic industrial wastewaters. FEMS Microbiol. Ecol.
was presumably inadequate for sufficient degradation 1578, 1–12.
of the high levels of generated acetate, resulting in the [2] Lettinga, G., Rebac, S. and Zeeman, G. (2001) Challenge of
accumulation of this VFA in R2 effluent and the sub- psychrophilic anaerobic wastewater treatment. Trends Biotechnol.
sequent decline in reactor performance. Previous stud- 19 (9), 363–370.
[3] Zoutberg, G.R. and de Been, P. (1997) The BiobedÒ EGSB
ies have suggested that high densities of both (Expanded Granular Sludge Bed) system covers shortcomings of
acetoclastic and hydrogenotrophic methanogens are the upflow anaerobic sludge blanket reactor in the chemical
necessary for balanced and successful anaerobic bio- industry. Water Sci. Technol. 35 (10), 183–188.
logical treatment, particularly during perturbations to [4] Wilderer, P.A., Bungartz, H.J., Lemmer, H., Wagner, M., Keller,
the system [24,25]. The results of this trial suggest that J. and Wuertz, S. (2002) Modern scientific methods and their
potential in wastewater science and technology. Water Res. 36 (2),
although methanogenic populations comprised pre- 370–393.
dominantly of hydrogenotrophs are capable of achiev- [5] Nozhevnikova, A.N., Holliger, C., Amann, A. and Zehnder,
ing high COD removal efficiencies within anaerobic A.J.B. (1997) Methanogenesis in sediments from deep lakes at
bioreactors (<95% COD removal in R2), the absence of different temperatures (2–70 °C) Wat. Water Sci. Technol. 36 (6–
acetoclastic methanogens within an anaerobic reactor 7), 57–64.
[6] Conrad, R. and Wetter, B. (1990) Influence of temperature on
infers an increased susceptibility to process disturbance energetics of hydrogen metabolism in homoacetogenic, methano-
on the system. genic, and other anaerobic bacteria. Arch. Microbiol. 155 (1), 94–
Previously, anaerobic bioreactor trials were moni- 98.
tored exclusively by material balances, such as pH, COD [7] Lay, J.J., Miyahara, T. and Noike, T. (1996) Methane release rate
removal efficiency and biogas methane content. Al- and methanogenic bacterial populations in lake sediments. Water
Res. 30, 901–908.
though all of these parameters are essential elements in [8] Shelton, D.R. and Tiedje, J.M. (1984) General method for
the successful operation of anaerobic biological waste- determining anaerobic biodegradation potential. Appl. Environ.
water treatment systems, this study illustrates that Microbiol. 47, 850–857.
analysis of the microbial composition of anaerobic bi- [9] Pender, S. (2000). Mesophilic and thermophilic anaerobic treat-
oreactors during reactor operation can provide a greater ment of molasses-based wastewater. PhD Thesis, National Uni-
versity of Ireland, Galway, Ireland.
understanding of the process, potentially explaining re- [10] Colleran, E., Concannon, F., Golden, T., Geoghegan, F., Crum-
actor performance and system disturbance. In the lish, B., Killilea, E., Henry, M. and Coates, J. (1992) Use of
present study, it was demonstrated that bacterial com- methanogenic activity tests to characterize anaerobic sludges,
munities display high levels of complexity and dyna- screen for anaerobic biodegradability and determine toxicity
mism, even during periods of functional stability. As a thresholds against individual anaerobic trophic groups and
species. Water Sci. Technol. 25 (7), 31–40.
consequence, it would appear that biomonitoring of the [11] Bitton, G., Koopman, B., Jung, K.H., Voiland, G. and Kotob, M.
archaeal community dynamics within anaerobic biore- (1993) Modification of the standard epiflourescence microscopic
actors is more effective in assessment of the response of method for total bacterial counts in environmental samples. Water
the reactor biomass to environmental change. However, Res. 27, 1109–1112.
with respect to bacterial populations, more specific [12] Stackebrandt, E. and Goodfellow, M. (1991) Nucleic acid
techniques in bacterial systematics. Wiley, England.
primer-enzyme combinations should be used to monitor [13] DeLong, E.F. (1992) Archaea in coastal marine sediments. Proc.
specific bacterial groups or species within sludge bio- Natl. Acad. Sci USA 89, 5685–5689.
mass, allowing for the generation of more informative [14] Lane, D., Pace, J.B., Olsen, G.J., Stahl, D.A., Sogin, M.L. and
results than those obtained in the present study using Pace, N.R. (1985) Rapid determination of 16S ribosomal RNA
universal primers. Molecular techniques are now an in- sequences for phylogenetic analysis. Proc. Natl. Acad. Sci. USA
82, 6955–6959.
valuable tool for the characterisation of complex mi- [15] Altschul, S., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z.,
crobial communities and, when used in conjunction with Miller, W. and Lipman, D.J. (1997) Gapped BLAST and PSI-
process engineering and physiological measurements, BLAST: a new generation of protein database search programs.
they will allow a more extensive knowledge and under- Nucl. Acids Res. 25, 3389–3402.
standing of the physiology and biochemistry of the mi- [16] Maidak, B.L., Cole, J.R., Lilburn, T.G., Parker Jr., C.T., Saxman,
P.R., Stredwick, J.M., Garrity, G.M., Li, B., Olsen, G.J.,
crobial populations involved in anaerobic digestion. Pramanik, S., Schmidt, T.M. and Tiedje, J.M. (1999) The RDP
This will allow for the identification of potential bot- (Ribosomal Database Project) continues. Nucl. Acids Res. 28,
tlenecks and problems, the development of novel high- 173–174.
378 S. McHugh et al. / FEMS Microbiology Ecology 48 (2004) 369–378

[17] Swofford, D.L. (2001) PAUP*. Phylogenetic Analysis Using [22] Osborn, A.M., Moore, E.R.B. and Timmins, K.N. (2000) An
Parsimony (*and Other Methods). Version 4. Sinauer Associates, evaluation of terminal restriction fragment length polymorphism
Sunderland, MA. (TRFLP) analysis for the study of microbial community structure
[18] Marsh, T.L., Saxman, P., Cole, J. and Tiedje, J. (2000) Terminal and dynamics. Environ. Microbiol. 2, 39–50.
Restriction Fragment Length Polymorphism Analysis program, a [23] Schnurer, A., Zellner, G. and Svensson, B.H. (1999) Mesophilic
web-based research tool for microbial community analysis. Appl. syntrophic acetate oxidation during methane formation in biogas
Environ. Microbiol. 66 (8), 3616–3620. reactors. FEMS Microbiol. Ecol. 29 (3), 249–261.
[19] Fernandez, A., Huang, S., Seston, S., Xing, J., Hickey, R., Criddle, [24] McHugh, S. (2003). Low temperature anaerobic treatment of food

Downloaded from https://academic.oup.com/femsec/article/48/3/369/583722 by Universidad Nacional de Colombia user on 01 November 2023

C. and Tiedje, J. (1999) How stable is stable. Function versus processing wastewaters – microbiology and process technology.
community composition. Appl. Environ. Microbiol. 65, 3697–3704. PhD Thesis, National University of Ireland, Galway,
[20] Philpott, U. (2001). Mesophilic and thermophilic anaerobic Ireland.
treatment of sulphate containing wastewater. PhD Thesis, Na- [25] Nozhevnikova, A.N., Rebac, S., Kotsyurbenko, O.R., Parshina,
tional University of Ireland, Galway, Ireland. S.N., Holliger, C. and Lettinga, G. (2000) Anaerobic production
[21] Gonzalez, J.S., Rivera, A., Borja, R. and Sanchez, E. (1998) and degradation of volatile fatty acids in low temperature
Influence of organic volumetric loading rate, nutrient balance and environments. Water Sci. Technol. 41 (12), 39–46.
alkalinity: COD ratio on the anaerobic sludge granulation of an [26] Saitou, N. and Nei, M. (1987) The neighbour-joining method: a
UASB reactor treating sugar cane molasses. Int. Biodet. Biode- new method of reconstructing phylogenetic trees. Mol. Biol. Evol.
grad. 41 (2), 127–131. 4, 406–425.