Professional Documents
Culture Documents
Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech
h i g h l i g h t s
1
COD.
) did not affect nitrogen removal.
Anaerobic effluent changed bacterial community structure and diversity.
Anammox and denitrifying bacteria were both found in the reactor.
Anammox can be applied to remove nitrogen from pre-treated municipal wastewater.
a r t i c l e
i n f o
Article history:
Received 20 January 2016
Received in revised form 18 March 2016
Accepted 19 March 2016
Available online 22 March 2016
Keywords:
Anammox
COD/N ratio
Anaerobic effluent
DGGE
Nitrogen removal
a b s t r a c t
Long-term effects of COD/N ratios on the nitrogen removal performance and bacterial community of an
anammox reactor were evaluated by adding a synthetic medium (with glucose) and real anaerobic effluent to a SBR. At a COD/N ratio of 2.8 (COD, 390 mgL 1) ammonium removal efficiency was 66%, while
nitrite removal remained high (99%). However, at a COD/N ratio of 5.0 (COD, 300 mgL 1), ammonium
and nitrite removal efficiencies were high (84% and 99%, respectively). High COD, nitrite, and ammonium
removal efficiencies (80%, 90% and 95%, respectively) were obtained on adding anaerobically pre-treated
municipal wastewater (with nitrite) to the reactor. DGGE revealed that the addition of anaerobic effluent
changed the bacterial community structure and selected for DNA sequences related to Brocadia sinica and
Chloroflexi. Adding glucose and anaerobic effluent increased denitrifiers concentration threefold. Thus,
the possibility of using the anammox process to remove nitrogen from anaerobically pre-treated municipal wastewater was demonstrated.
2016 Elsevier Ltd. All rights reserved.
1. Introduction
Anaerobic ammonium oxidation (anammox) is a microbial process that can be used, in principle, for the treatment of ammoniumrich wastewaters with a low C:N ratio. Anammox bacteria oxidize
258
Table 1
Characterization of the anaerobic effluent from a UASB reactor that treats municipal
wastewater.
Compound
Unit
Median value
pH
Conductivity
DO
Temperature
Total BOD
Total COD
Soluble COD
TSS
VSS
Total nitrogen
TNK
NH+4
NO2
NO3
PO34
SO24
7.2 (0.1)a
744 (64)
0.3 (0.1)
25.7 (2.3)
72 (22)
150 (92)
80 (35)
71 (43)
51 (30)
46 (8)
45 (8)
27 (4)
0.0 (0.0)
0.1 (0.1)
6.9 (4.0)
14.2 (4.5)
lscm
mgL
C
mgL
mgL
mgL
mgL
mgL
mgL
mgL
mgL
mgL
mgL
mgL
mgL
1
1
1
1
1
1
1
1
1
1
1
1
1
1
259
260
(a)
(b)
(c)
Fig. 1. (a) Performance of the SBR during operation with autotrophic medium (Phase I, anammox cultivation), increasing influent COD/N ratios (Phase II, glucose addition),
and the addition of real anaerobic effluent (Phase III). Box plot of ammonium (b) and nitrite (c) removal efficiencies of the SBR during operation with autotrophic medium
(Phase I), increasing COD/N ratios (Phase II), and the addition of real anaerobic effluent (Phase III).
The
stoichiometric
coefficient
(consumption
of
NNO2 /consumption of N-NH+4) determined during Phase I was 1.46
(as shown in Table 2), which is close to that (1.32) reported for
an anammox reaction by Strous et al. (1998). Quan et al. (2008)
reported a coefficient of 1.46 in an UASB reactor operated in anammox conditions. The average value found for the coefficient of NNO3 production/N-NH+4 consumption was 0.34, slightly higher than
the one reported by Strous et al. (1998), of 0.26. Date et al. (2009)
reported a coefficient of 0.33 (production of N-NO3 /consumption
of N-NH+4) in an upflow reactor with biomass from a wastewater
treatment tank from a pig production facility. A similar result
(0.36) was reported by Pereira et al. (2014).
The results obtained in Phase I demonstrated that the anammox
process was established in the SBR, thus confirming that the two
261
NH+4-N removal
(mg N/L)a
Anammox
Denitrification
Production
Removal
Final
116.0
33.6
169.4
0.0
39.6
0.0
II
0.7:1
1.4:1
2.8:1
3.5:1
R**
5.0:1
49.1
46.4
43.2
11.4
35.0
31.5
5.5
9.0
5.3
0.9
16.9
8.0
71.7
67.8
63.1
16.6
51.2
37.7
10.9
1.0
16.9
61.6
6.1
1.4
16.7
15.8
14.7
3.9
11.9
9.0
5.3
6.8
11.0
1.1
5.3
3.3f
Phase
SD
COD influent
(mg/L)
COD removal
(mg/L)
Stoichiometrye
39.6
0.0
0.0
1:1.46:0.34
11.5
9.0
3.7
2.8
6.6
12.3
97.5
195.0
390.0
487.5
0.0
300.0
67.8
162.2
261.0
89.2
0.0
174.4
1:1.30:0.31
III A
3.0:1
31.7
15.8
34.8
0.3
6.3
0.9
5.4
238.2
204.4
1:1.10:0.20
III B
3.5:1
48.5
10.1
82.4
32.1
9.7
0.9
8.8
172.2
133.1
1:1.70:0.20
these values and those obtained during Phase I, followed by multiple comparisons of medians (a = 5%; p = 0.1531). However, for
ammonium removal efficiency, statistical differences between
the average values obtained from Phase II and Phase I were
detected by KruskalWallis test (a = 5%; p = 0.01356). In addition,
for multiple comparisons of median values, the KruskalWallis test
(a = 5%) was performed and statistical differences were observed
between periods with COD/N ratios of 0.7 and 1.4 (p = 0.0412),
0.7 and 2.8 (p = 0.0453), 0.7 and 3.5 (p = 0.0167), 0.7 and the recovery phase (p = 0.0102) and 0.7 and 5.0 (p = 0.0201). Thus, the addition of glucose (at concentrations of 195 mgL 1, 390 mgL 1,
487.5 mgL 1 and 300 mgL 1, corresponding to COD/N ratios of
1.4, 2.8, 3.5 and 5.0, respectively) affected ammonium removal efficiency and consequently the anammox process. Moreover, these
results indicate that the COD concentration of 487.5 mgL 1 was
the one that severely affected and inhibited the anammox process
since ammonium removal was significantly reduced (as shown in
Fig. 1b) and the color of the biomass changed from brownishyellow to a darker color (as shown in Supplementary Fig. S1).
The nitrogen mass balance was calculated for all operational
phases to evaluate the participation of different processes (Table 2).
The results indicated that the addition of glucose at different concentrations (during Phase II) gradually reduced the participation of
anammox process in the nitrite removal (from 87% at COD/N ratio
of 0.723% at COD/N of 3.5) while favored the denitrification process, being the highest consumption of nitrite by denitrification
was observed at the COD/N ratio of 3.5 (Table 2). Nitrate removal
during Phase II was also observed, being the highest nitrate consumption via denitrification was observed at COD/N ratio of 2.8
(Table 2).
The COD was monitored during Phase II to determine whether
there was glucose consumption (COD removal) by other bacterial
groups or whether COD was accumulating inside the reactor. The
average COD removal efficiencies in each of the steps (shown in
Fig. 2) were 71% (COD of 97.50 mgL 1, COD/N = 0.7), 82% (COD of
195 mgL 1, COD/N = 1.4), 62% (COD of 390.0 mgL 1, COD/
N = 2.8), 16% (COD of 487.50 mgL 1, COD/N = 3.5), and 55% (COD
of 300.00 mgL 1, COD/N = 5.0), indicating that COD was being consumed inside the reactor probably by denitrifying bacteria (confirming the mass balance results).
Phase II may be viewed as the most critical phase of the whole
study, as it was the one in which nitrogen removal performance by
262
Fig. 2. Box plot of COD removal efficiencies at different COD/N ratios during Phase II (glucose addition) and Phase III (addition of anaerobic effluent amended with nitrite).
efficiencies were higher (85%) than in Phase IIIA (51%) (Fig. 1b).
Statistical differences were observed between these values when
they were compared by MannWhitney test (a = 5%;
p = 0.00016). By the end of Phase IIIB, an increase in ammonium
removal efficiencies to higher than 95% was observed.
Previous studies have shown that anammox bacteria enriched
in reactors fed with organic compounds such as propionate and
acetate are capable of competing with denitrifying heterotrophic
bacteria for the electron acceptor (nitrite) because of their affinity
for the substrate (Kartal et al., 2012). Anammox bacteria have very
low growth rate and cellular yield (coefficient Y = 0.066 0.01)
when compared to denitrifying heterotrophic bacteria
(Y = 0.27 0.3), according to Strous et al., 1998. Therefore, denitrifying bacteria may grow at a higher rate when organic compounds
are combined with ammonium and nitrite in the wastewater being
treated.
Nitrate production (via anammox process) was monitored during Phase III, and the average values obtained were 6.3 and
9.7 mgL 1 in steps A and B, respectively (Table 2). These values
were lower than those obtained in Phases I and II. This might be
explained by an increase in denitrifying activity following the addition of anaerobic effluent to the SBR. In fact, the mass balance
results (Table 2) confirmed this observation and showed that 28%
of the nitrite added to the reactor was consumed by denitrification
process (in Phase IIIB).
A COD consumption assessment was performed during Phase
III, and the results varied depending on the composition of the
anaerobic effluent used to feed the SBR. The average COD value
determined for the anaerobic effluent was 208.44 54 mgL 1.
During Phase III, COD removal efficiency was nearly 85% in step
A and 74% in step B (Fig. 2). Therefore, along with the anammox
process, heterotrophic denitrification was also taking place inside
the reactor (as shown in Table 2), as a large part of COD (from
the real anaerobic effluent) was consumed, and that COD removal
explained the high nitrite consumption, and, consequently, the
higher observed stoichiometric coefficient (N-NO2 /N-NH+4) of 1.7
(Table 2). In Phase IIIB, nitrite consumption via denitrification
was 28% while via anammox was 72%, indicating that denitrification occurred in the reactor but anammox process prevailed.
The results from Phase III suggest that the anammox process
can be applied to the post-treatment of anaerobic effluents with
high COD and low ammonium concentrations in order to remove
nitrogen. An important aspect of the present study was the application of the anammox process to the treatment of real anaerobic
263
Table 3
Identification of DNA band sequences obtained from DGGE.
3.2. Bacterial community structure in the SBR before and after the
addition of glucose and real anaerobic effluent
DGGE profiles and band patterns from biomass developed in the
SBR and sampled at day 349 (after Phase I), day 482 (after Phase II,
with glucose addition), day 448 (recovery phase-R, without glucose
addition), and day 558 (Phase III, after the addition of anaerobic
effluent) were evaluated (Fig. 3). The bacterial community structure changed after the addition of glucose (Phase II) and anaerobic
effluent (Phase III) to the reactor (Fig. 3a).
Representative bands were excised, and the DNA was
sequenced to identify microorganisms present in each phase of
reactors operation (Table 3). Most of the DGGE bands yielded
sequences related to those of bacteria within the phyla Proteobacteria, Chloroflexi, and Planctomycetes.
DGGE results showed that bands A and H, with sequences closely related to the anammox bacteria Candidatus Brocadia caroliniensis and Candidatus Brocadia sinica, respectively, were found in
every phase of the study (Fig. 3a and Table 3). These species are
commonly found in wastewater treatment systems (Hu et al.,
2010). Ca. Brocadia sinica has been reported in the biomass of
anammox reactors from China, Japan, and Germany, indicating
the worldwide distribution of these bacteria (Oshiki et al., 2011).
Physiologic characteristics of the species Ca. Brocadia sinica
were compared to the characteristics of Ca. Brocadia anammoxidans
and Ca. Kuenenia stuttgartiensis by Oshiki et al. (2011). They found
that Ca. Brocadia sinica had a higher growth rate with less affinity
Band
RDP classifier
Blast
Similaritya
(%)
Acc. No.
KF810110.1
97
KF810114.1
C
D
Myxococcales
Chloroflexi
Ca. Brocadia
caroliniensis
Denitratisoma
oestradiolicum
Myxococcales
Anaerolineaceae
98
B, J
Ca.
Brocadiaceae
Rhodocyclaceae
92
97
FJ552616.1
Rhodospirillales
Acetobacteraceae
91
EU193085.1
Chloroflexi
Caldilinea sp.
84
HQ043268.1
Acetobacteraceae
Acetobacteraceae
93
EU921207.1
99
I
K
Chlorobi
Actinomycetales
Chlorobi
Propioniferax sp.
88
90
AB565477.1
KJ941776.1
Anaerolineaceae
Chloroflexi
97
JF703582.1
Caldilinea sp
Chloroflexi
94
AY921707.1
Planctomycetes
Planctomycetales
80
FJ517124.1
Thermoflexus sp.
Chloroflexi
95
AY921707.1
Chloroflexi
Chloroflexi
96
AY921913.1
HE648186.1
KP419698.1
Percentages indicate the similarity between the DGGE band sequences and the
closest matched sequences in GenBank. Words in bold indicate: DNA bands A and H
related to anammox bacteria (Brocadia); bands B, J and K are related to bacteria able
to degrade glucose and might be involved in COD removal.
Fig. 3. Bacterial community analysis by denaturating gradient gel electrophoresis (DGGE). (a) DGGE profile of the bacterial community in biomass sampled at day 349 (I)
(after Phase I, anammox cultivation), day 482 (II) (after Phase II, glucose addition), day 448 (R) (after 35 days without glucose-recovery phase), and day 558 (III) (Phase III,
after the addition of anaerobic effluent). (b) Dendrogram based on the DGGE profiles.
264
(a)
(b)
Fig. 4. (a) Nitrogen removal rate and abundance of anammox, denitrifying (inferred from the nosZ gene), and total bacterial populations in the biomass of the SBR in Phases I
(anammox cultivation), II (after glucose addition), and III (after anaerobic effluent addition). (b) Relative abundance (in percentage) of anammox bacteria and denitrifiers in
relation to total bacterial populations.
were 2.49, 2.50 and 2.82, respectively. The addition of anaerobically pre-treated municipal wastewater to the SBR did not select
against anammox bacteria, on the contrary, it allowed for the coexistence of anammox, denitrifying bacteria and Chloroflexi.
The presence of denitrifying bacteria in the SBR was also confirmed by qPCR (Fig. 4a and b). Higher concentrations of denitrifying bacteria (as measured by nosZ gene abundance) were observed
in the SBR biomass with increasing COD/N ratios (Phase II, with
glucose addition) and after the addition of real anaerobic effluent
(Phase III). These results indicate that the addition of glucose and
the anaerobic effluent favored the enrichment of denitrifiers and
increased their concentration threefold (from 2.57 109 to
7.57 109 copies/g of sludge) (Fig. 4a). However, the concentration
of anammox bacteria (as measured by 16S rRNA gene abundance)
did not change with an increase in COD or application of anaerobic
effluent, indicating that long-term operation of the SBR enriched
and sustained the anammox population (ranging from 2.08 109
to 1.90 109 copies/g of sludge). Ni et al. (2010), investigating an
anammox UASB reactor, observed an anammox concentration of
4.6 108 copies/g of sludge when the reactor reach a nitrogen
removal efficiency of 94%.
The relative abundances of anammox bacteria and denitrifiers
in relation to total bacteria are presented in Fig. 4b. The results also
showed that the addition of glucose and the real anaerobic effluent
(Phases II and III, respectively) changed the bacterial community
by increasing the abundance of denitrifiers more than fourfold in
relation to the anammox population.
4. Conclusions
High concentrations of COD (above 487 mgL 1) inhibited the
anammox process. However, a COD/N ratio of 5.0 (with COD concentrations up to 300 mgL 1) did not inhibit the process and
allowed for the coexistence of anammox and denitrifying bacteria.
High COD, nitrite, and ammonium removal efficiencies (80%, 90%,
and 95%, respectively) were obtained with the addition of real
anaerobic effluent to the SBR. This changed the bacterial community structure and selected for DNA sequences related to Brocadia
sinica and Chloroflexi. Thus, the feasibility of applying the anammox
process to nitrogen removal from anaerobically pre-treated municipal wastewater was demonstrated.
Conflict of interest
All authors declare that they have no conflict of interest.
Acknowledgements
We are thankful to the following brazilian agencies: Coordenao de Aperfeioamento de Pessoal de Nvel Superior (CAPES),
to Fundao de Amparo a Pesquisa do Estado de Minas Gerais
(FAPEMIG) and to Conselho Nacional de Desenvolvimento Cientfico e Tecnolgico do Brasil (CNPq).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.biortech.2016.03.
107.
References
Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990. Basic local
alignment search tool. J. Mol. Biol. 2015, 403410.
APHA, 2012. Standard Methods for the Examination of Water and Wastewater,
22nd ed. American Public Health Association, Washington, DC, USA.
265
266
Tang, C.J., Zheng, P., Wang, C.H., Mahmood, Q., 2010. Suppression of anaerobic
ammonium oxidizers under high organic content in high-rate Anammox UASB
reactor. Bioresour. Technol. 101, 17621768.
Van de Graaf, A.A., Bruijn, P., Robertson, L.A., Jetten, M.S.M., Kuenen, J.G., 1996.
Autotrophic growth anaerobic ammonium-oxidation micro-organisms in a
fluidized bed reactor. Microbiology 142, 21872196.
Yokota, A., Tamura, T., Takeuchi, M., Weiss, N., Stackebrandt, E., 1994. Transfer of
Propionibacterium innocuum Pitcher and Collins 1991 to Propioniferax gen. nov.
as Propioniferax innocua comb. nov. Int. J. Syst. Bacteriol. 44, 579582.
Zhang, L., Zheng, P., Tang, C., Jin, R., 2008. Anaerobic ammonium oxidation for
treatment of ammonium-rich wastewaters. J. Zhejiang Univ. Sci. B 9, 16426.