Distribution of Extracellular Enzyme Producing Bacteria in the Digestive Tracts of Four Brackish water Fish Species

Abstract

Abstract: Occurrence and distribution of enzyme-producing bacteria in the proximal (PI), middle (MI), and distal (DI) segments of
the gastrointestinal tracts of 4 brackish water teleosts (Scatophagus argus, Terapon jarbua, Mystus gulio, and Etroplus suratensis) have
been investigated. Data were presented as log viable counts g–1
intestine (LVC). Te heterotrophic bacterial population had the highest
occurrence in the DI regions of all fsh species studied except M. gulio. Proteolytic and amylolytic bacteria had the highest occurrence
in the DI of M. gulio (LVC = 5.50 and 5.93, respectively), while cellulolytic and lipolytic populations exhibited highest occurrences in
the DI regions of T. jarbua (LVC = 6.33) and S. argus (LVC = 5.78), respectively. Out of the 81 bacterial isolates, the most promising 3
isolates were determined through quantitative enzyme assay and studied through 16S rRNA gene sequence analysis for identifcation.
Both the strains SA2.2 isolated from S. argus and TJ2.3 isolated from T. jarbua showed high similarity to diferent strains of Brevibacillus
parabrevis, while another strain, MG4.2, isolated from M. gulio, was similar to Bacillus licheniformis. Te NCBI GenBank accession
numbers of the 16S rRNA gene sequences for isolates SA2.2, TJ2.3, and MG4.2 were KF377322, KF377324, and KF377323, respectively.
Te present study might ofer scope for further research to evaluate prospects for application of the gut-associated extracellular enzyme-
producing bacteria in brackish water aquaculture.
Key words: Brackish water fsh, gut bacteria, enzyme, Brevibacillus parabrevis

Full text

79
http://journals.tubitak.gov.tr/zoology/
Turkish Journal of
Zoology
Turk J Zool
(2014) 38: 79-88
© TÜBİTAK
doi:10.3906/zoo-1205-3
Distribution of extracellular enzyme-producing bacteria in the digestive tracts of 4
brackish water sh species
Paramita DAS
1
, Sudipta MANDAL
1
, Argha KHAN
1
, Sanjib Kumar MANNA
2
, Koushik GHOSH
1,
*
1
Aquaculture Laboratory, Department of Zoology, e University of Burdwan, Golapbag, Burdwan, West Bengal, India
2
Central Inland Fisheries Research Institute, Barrackpore, West Bengal, India
* Correspondence: kghosh@zoo.buruniv.ac.in
1. Introduction
e microora within the gastrointestinal (GI) tract of
marine and freshwater sh species has been widely inves-
tigated (Austin, 2002; Ghosh et al., 2010; Ray et al., 2010;
Askarian et al., 2011, 2012) e nutrient-rich GI tract of
sh is a favorable growth environment for these bacteria
(Kar et al., 2008). During the last decade, there has been an
improved understanding of the importance of commen-
sal intestinal microora in sh (Bairagi et al., 2002; Ghosh
et al., 2002a, 2010; Ringø et al., 2010; Khan and Ghosh,
2012; Mandal and Ghosh, 2013). e gut microora may
be categorized as either autochthonous (indigenous) or
allochthonous (transient) depending upon its ability to
colonize and adhere to the mucus layer in the digestive
tract (Ringø and Birkbeck, 1999; Ringø et al., 2003). e
bacterial ora within the GI tract of sh shows very broad
and variable enzymatic potential, and these enzymatic
masses may interfere positively in the digestive process of
sh (Ray et al., 2010). Fish gut bacterial isolates have been
demonstrated to break down chitin (Danulat and Kausch,
1984; MacDonald et al., 1986; Itoi et al., 2006), p-nitro-
phenyl-b-N-acetylglucosamine and protein (MacDonald
et al., 1986; Belchior and Vacca, 2006), cellulose (Saha and
Ray, 1998; Bairagi et al., 2002; Ghosh et al., 2002a, 2010;
Saha et al., 2006; Mondal et al., 2008), starch (Sugita et al.,
1997; Ghosh et al., 2002a, 2010), phytate (Li X et al., 2008,
Li XY et al., 2008; Roy et al., 2009; Khan et al., 2011; Khan
and Ghosh, 2012, 2013), and tannin (Mandal and Ghosh,
2013). Previous studies conducted in the carps advocated
the benecial aspects of gut-associated microbiota in the
host sh with regard to nutrition (Ghosh et al., 2002a,
2002b, 2003; Ray et al., 2010). Meanwhile, information on
the enzyme-producing gut bacteria in the brackish water
sh species is scanty (Rani et al., 2004; Sivasubramanian
et al., 2012). In this context, the search for extracellular
enzyme-producing benecial gut bacteria to be used as
probiotics for the culturable brackish water sh species
may be of interest.
Abstract: Occurrence and distribution of enzyme-producing bacteria in the proximal (PI), middle (MI), and distal (DI) segments of
the gastrointestinal tracts of 4 brackish water teleosts (Scatophagus argus, Terapon jarbua, Mystus gulio, and Etroplus suratensis) have
been investigated. Data were presented as log viable counts g
–1
intestine (LVC). e heterotrophic bacterial population had the highest
occurrence in the DI regions of all sh species studied except M. gulio. Proteolytic and amylolytic bacteria had the highest occurrence
in the DI of M. gulio (LVC = 5.50 and 5.93, respectively), while cellulolytic and lipolytic populations exhibited highest occurrences in
the DI regions of T. jarbua (LVC = 6.33) and S. argus (LVC = 5.78), respectively. Out of the 81 bacterial isolates, the most promising 3
isolates were determined through quantitative enzyme assay and studied through 16S rRNA gene sequence analysis for identication.
Both the strains SA2.2 isolated from S. argus and TJ2.3 isolated from T. jarbua showed high similarity to dierent strains of Brevibacillus
parabrevis, while another strain, MG4.2, isolated from M. gulio, was similar to Bacillus licheniformis. e NCBI GenBank accession
numbers of the 16S rRNA gene sequences for isolates SA2.2, TJ2.3, and MG4.2 were KF377322, KF377324, and KF377323, respectively.
e present study might oer scope for further research to evaluate prospects for application of the gut-associated extracellular enzyme-
producing bacteria in brackish water aquaculture.
Key words: Brackish water sh, gut bacteria, enzyme, Brevibacillus parabrevis
Received: 03.05.2012
Accepted: 21.08.2013 Published Online: 01.01.2014 Printed: 15.01.2014
Research Article

80
DAS et al. / Turk J Zool
erefore, the primary objective of the present study
was to detect the autochthonous extracellular enzyme-
producing bacteria from the proximal (PI), middle (MI),
and distal (DI) segments of the GI tracts of 4 culturable
brackish water teleosts. Furthermore, the study was
intended to evaluate the gut bacteria’s ability for protease,
amylase, cellulase, and lipase production, and to identify
the most promising bacterial strains by 16S rRNA gene
sequence analysis.
2. Materials and methods
2.1. Fish species examined
Four brackish water sh species, Scatophagus argus,
Mystus gulio, Terapon jarbua, and Etroplus suratensis,
were collected from 3 brackish water culture ponds of the
Fish Technological Station, Junput, West Bengal, India
(21°43.232′N, 87°48.884′E) and brought to the laboratory
in oxygen-packed plastic bags. Physicochemical
parameters of the collection ponds ranged between pH
8.1 and 8.4, with water temperature of 29.8–31.4 °C and
salinity of 16–18 ppt. Feeding habits, average live weight,
average sh length, and average gut weight of the shes
examined are presented in Table 1.
2.2. Processing of specimens
Nine specimens of each species collected from 3 ponds
(3 from each pond) were evaluated for the present study.
e shes were starved for 48 h to empty the GI tracts
(Ray et al., 2010). Aer starvation, sh were anesthetized
with 0.03% tricaine methane-sulfonate (MS222), and the
ventral surface of each sh was thoroughly scrubbed with
1% iodine solution for surface decontamination (Trust
and Sparrow, 1974). e shes were dissected aseptically
within a laminar airow and their alimentary tracts
were removed. Gut samples were processed for isolation
of adherent (autochthonous) bacteria as described by
Ringø (1993), with minor modication. e GI tracts
were divided into proximal (PI), middle (MI), and distal
(DI) segments, cut into pieces, and ushed carefully 3
times with 0.9% sterile saline solution using an injection
syringe in order to remove nonadherent (allochthonous)
microora, according to Ghosh et al. (2010). Gut segments
from the 3 specimens of a species collected from the same
pond were pooled together region-wise for each replicate,
and thus there were 3 replicates for each gut segment from
each sh species. e gut segments were homogenized
with 10 parts of sterilized, prechilled 0.9% NaCl solution,
as described elsewhere (Das and Tripathi, 1991). Pooled
samples were used to avoid erroneous conclusions due
to individual variations in gut microora, as described
elsewhere (Ghosh et al., 2010).
2.3. Microbial culture
Homogenized samples of each gut segment were used
separately aer appropriate serial (1:10) dilutions (Beveridge
et al., 1991). Diluted samples (0.1 mL) were poured
aseptically within a laminar airow onto sterilized tryptone
soy agar (TSA, HiMedia, India) plates to determine the
culturable heterotrophic autochthonous aerobic/facultative
anaerobic microbial population. For determination of
protease-, cellulase-, amylase-, and lipase-producing
bacterial populations, diluted samples (0.1 mL) were poured
onto peptone gelatin agar (PG), carboxymethylcellulose
agar (CMC), starch agar (SA), and tributyrin agar (TA)
plates, respectively. e culture plates were incubated at
30 °C for 48 h. e colony-forming units (CFUs) per unit
sample volume of gut homogenate were determined by
multiplying the number of colonies formed on each plate by
the reciprocal dilution (Rahmatullah and Beveridge, 1993),
and the data were presented as log viable counts g
–1
intestine
(LVC). Colonies with apparently dierent morphological
appearances (such as color, conguration, surface, margin,
and opacity) from a single plate were streaked separately on
respective plates to obtain pure cultures.
Table 1. Food habits, average live weight, average sh length (standard length), and average gut weight of the shes examined.
Fish species Feeding habit
Average live
weight (g)
Average sh
length (cm)
Average gut
weight (g)
Scatophagus argus
Omnivorous; feeds on shes, insects, algae,
and sand-dwelling invertebrates
53.87 ± 4.92 13.76 ± 0.56 3.02 ± 0.2
Terapon jarbua
mnivorous; feeds on worms, crustaceans,
insects, and plant matter
46.5 ± 3.56 14.2 ± 0.81 1.7 ± 0.14
Mystus gulio Carnivorous 24.2 ± 1.94 11.92 ± 0.45 1.22 ± 0.17
Etroplus suratensis
Omnivorous; feeds on lamentous algae, plant
material, and insects
33.46 ± 1.4 11.99 ± 0.28 0.53 ± 0.03
Results are mean ± SD of the 3 observations.

81
DAS et al. / Turk J Zool
2.4. Screening of isolates by qualitative assay for
exoenzyme production
Out of the 81 extracellular enzyme-producing isolates
from the sh species examined, 21 isolates were primarily
selected (on the basis of growth potential at 30 °C) for
qualitative enzyme assay. For extracellular amylase
production, isolates were inoculated on SA plates and
incubated at 30 °C for 48 h. e culture plates were ooded
with 1% Lugol’s iodine solution to identify amylase activity
by the formation of a transparent zone (halo) surrounding
the colony (Jacob and Gerstein, 1960). Similarly, for
extracellular protease, the isolates inoculated on PG plates
were incubated at 30 °C for 48 h; the appearance of a
halo aer ooding the plates with 15% HgCl
2
indicated
the presence of proteolytic activity (Jacob and Gerstein,
1960). For determination of cellulase production, isolates
grown on CMC plates at 30 °C for 48 h were ooded
with Congo red dye prepared with 0.7% agarose (Teather
and Wood, 1982). Congo red selectively binds with
unhydrolyzed CMC. Appearance of a halo due to the
presence of hydrolyzed CMC surrounding the bacterial
colony indicated cellulase production in the medium.
Lipase producers showed a halo surrounding their colony
in 1% tributyrin plates (Sangiliyandi and Gunasekaran,
1996). ere were 3 replicates for each experimental set.
Qualitative extracellular enzyme activity was assessed
based on the measurement of the halo zone (diameter in
mm) around the colony and presented as scores, as follows:
0 (0–5 mm), 1 (low, 6–10 mm), 2 (moderate, 11–15 mm), 3
(good, 16–20 mm), 4 (high, 21–25 mm), and 5 (very high,
>25 mm).
2.5. Quantitative enzyme assay
On the basis of the qualitative assay, 10 extracellular
enzyme-producing isolates were selected for quantitative
assay using broth culture to screen the promising isolates.
Quantitative assay for the production of amylase, cellulase,
protease, and lipase were performed following the methods
described by Bernfeld (1955), Denison and Koehn (1977),
Walter (1984), and Bier (1955), respectively. A detailed
description for measurement of extracellular enzyme
production and quantitative enzyme assay has been
mentioned elsewhere (Bairagi et al., 2002). Quantitative
enzyme activities were expressed as units (U).
2.6. Identication of isolates by 16S rRNA gene sequence
analysis
e most promising 3 of the extracellular enzyme-
producing isolates were investigated by means of their
quantitative enzyme assays (amylase, cellulase, protease,
and lipase) in addition to their 16S rRNA partial gene
sequence analysis for identication. e gene encoding
16S rRNA was amplied from the isolates by polymerase
chain reaction (PCR) using universal primers 27f
(5´-AGAGTTTGATCCTGGCTCAG-3´) and 1492r
(5´-GGTTACCTTGTTACGACTT-3´). e PCR
reactions were performed using PCR mix containing 200
µM of deoxynucleotides (dNTPs), 0.2 µM of each primer,
2.5 mM MgCl
2
, 1X PCR buer, and 0.2 U of Taq DNA
polymerase (Invitrogen). e template DNA was obtained
by extracting genomic DNA using the GenElute Bacterial
Genomic DNA Kit (Sigma-Aldrich) from a fresh colony
grown on a nutrient agar slant. e following cycle was
used for PCR reaction: initial denaturation at 95 °C for 3
min, followed by 35 cycles at 95 °C for 1 min, annealing at
55 °C for 1 min, and extension at 72 °C for 2 min, and a nal
extension at 72 °C for 3 min (Lane, 1991). PCR products
were sent to a commercial house for Sanger sequencing
using an automated DNA sequencer (Applied Biosystems
Ltd.). Sequenced data were aligned and analyzed for nding
the closest homolog of the microbes using a combination
of NCBI GenBank and RDP databases. e phylogenetic
tree was constructed incorporating 16S rRNA partial
gene sequences of isolates SA2.2, TJ2.3, and MG4.2 and
their phylogenetically closest type strains using MEGA
5.2 soware following the minimum evolution method.
Partial sequences of 16S rRNA from the 3 selected isolates
were deposited in the NCBI GenBank database to obtain
accession numbers.
2.7. Media composition
TSA medium contained (g L
–1
): pancreatic digest of casein,
15; papaic digest of soybean meal, 5; NaCl, 5; agar, 15; pH of
7.5. PG medium contained (g L
–1
): beef extract, 3; peptone,
5; gelatin, 4; agar, 20; pH of 7.5. CMC medium contained
(g L
–1
): beef extract, 5; peptone, 5; NaCl, 5; carboxymeth-
ylcellulose, 2; agar, 20; pH of 7.5. SA medium contained (g
L
–1
): beef extract, 5; peptone, 5; NaCl, 5; starch (soluble), 2;
agar, 20; pH of 7. TA medium contained (g L
–1
): tributyrin-
agar, 10; peptone, 5; agar, 15; pH of 7.5.
2.8. Statistical analysis
Data pertaining to specic extracellular enzyme production
by the selected isolates were subjected to analysis of
variance (ANOVA) followed by Tukey’s test following Zar
(1974) using SPSS 10 (Kinear and Gray, 2000).
3. Results
Heterotrophic as well as protease-, cellulase-, amylase-,
and lipase-producing bacterial populations present in
the PI, MI, and DI segments of the GI tracts of all the
sh species examined are presented in Table 2. Analysis
of the bacterial populations in the GI tracts of 4 brackish
water sh revealed that the heterotrophic population
on TSA plates diered among dierent sh species as
well as among dierent regions of the gut. e bacterial
population on TSA plate was highest in the DI region of
M. gulio (LVC = 7.49 g
–1
intestine), followed by T. jarbua
(LVC = 7.48 g
–1
intestine), while it was lowest in the PI

82
DAS et al. / Turk J Zool
region of M. gulio (LVC = 6.38 g
–1
intestine). e highest
amylolytic bacterial population was detected in the DI
region of M. gulio (LVC = 5.93 g
–1
intestine), followed by
S. argus (LVC = 5.73 g
–1
intestine); it was lowest in the PI
region of E. suratensis (LVC = 4.87 g
–1
intestine). Cellulase-
producing bacteria showed the highest concentration
in the DI region of T. jarbua (LVC = 6.33 g
–1
intestine),
followed by S. argus (LVC = 6.06 g
–1
intestine), while it
was lowest in the PI region of E. suratensis (LVC = 4.83
g
–1
intestine). Proteolytic bacterial population showed the
highest concentration in the DI region of M. gulio (LVC =
5.50 g
–1
intestine), followed by T. jarbua (LVC = 5.41 g
–1
intestine); the lowest concentration was in the PI region
of S. argus (LVC = 4.69 g
–1
intestine). Lipolytic bacteria
rates were highest in the DI region of S. argus (LVC =
5.78 g
–1
intestine), followed by the MI region of the same
species (LVC = 5.76 g
–1
intestine). Altogether, 21 enzyme-
producing bacterial isolates were primarily selected from
dierent sh species, and extracellular enzyme production
by the bacterial isolates was assayed qualitatively.
Qualitative extracellular enzyme activities were presented
as scores (Table 3), maximum and minimum scores being
18 and 2, respectively. Based on the qualitative assay, 10
bacterial isolates were selected for the quantitative enzyme
assay. Results of the quantitative enzyme assay revealed
signicant dierences in the enzyme activities among
dierent bacterial isolates (Table 4). Maximum amylase
and cellulase activities were recorded in SA2.2 isolated
from the DI of S. argus (44.03 ± 0.43 U and 13.12 ± 0.23 U,
respectively). Protease activity was highest in TJ2.3 isolated
from the DI of T. jarbua (26.89 ± 0.28 U), while the best
lipase activity was noticed in MG4.2 isolated from the DI
of M. gulio (11.1 ± 0.31 U). Considering all 4 enzymatic
activities, isolates SA2.2, TJ2.3, and MG4.2 were found
to have the most potential among the 10 selected isolates.
Based on the nucleotide homology and phylogenetic
analysis of the 16S rRNA gene sequences, isolates SA2.2
and TJ2.3 were both identied as Brevibacillus parabrevis
(GenBank Accession Nos. KF377322 and KF377324). e
isolate SA2.2 showed 100% similarity with B. parabrevis
HDYM-18 (EF428244), while isolate TJ2.3 showed 99%
similarity with B. parabrevis M3 (AB215101). us,
SA2.2 and TJ2.3 were 2 dierent strains of B. parabrevis.
e isolate MG4.2 was identied as Bacillus licheniformis
Table 2. Log viable counts (LVC) of autochthonous adherent bacteria isolated from the proximal (PI), middle (MI), and distal (DI)
segments of the GI tracts of the sh species examined.
LVC g
–1
intestine
Fish species
Bacterial count in
TSA plate
Amylolytic
bacteria
Cellulolytic
bacteria
Proteolytic
bacteria
Lipolytic
bacteria
Scatophagus argus
PI 6.93 4.92 5.06 4.69 5.58
MI 7.11 5.61 5.84 4.99 5.76
DI 7.29 5.73 6.06 5.12 5.78
Terapon jarbua
PI 6.98 5.33 5.18 4.85 5.09
MI 7.28 5.50 5.92 5.20 5.43
DI 7.48 5.62 6.33 5.41 5.55
Mystus gulio
PI 6.38 5.61 5.69 5.05 5.31
MI 7.09 5.66 5.29 5.27 5.35
DI 7.49 5.93 4.93 5.50 5.50
Etroplus suratensis
PI 6.78 4.87 4.83 4.74 4.96
MI 6.92 4.93 4.85 4.89 5.08
DI 7.25 4.94 4.99 5.08 5.23

83
DAS et al. / Turk J Zool
(GenBank Accession No. KF377323). It showed 100%
similarity with B. licheniformis GLU 113 (FN678352). It
appeared from the phylogenetic tree that strains SA2.2
and TJ2.3 were closest to the B. parabrevis type strain
(AB112714) and were grouped together, whereas MG4.2
was distantly placed with B. licheniformis (Figure).
4. Discussion
Diverse microbial communities in the GI tracts of
freshwater or marine carnivorous, herbivorous, and
omnivorous sh species have been reported abundantly
(for review, see Ray et al., 2012). However, the
endosymbiotic community among the brackish water
sh species has remained poorly investigated (Esakkiraj
et al., 2009). Digestive tracts of endotherms are colonized
mainly by obligate anaerobes (Finegold et al., 1983),
while the predominant bacterial genera isolated from
most sh guts have been aerobes or facultative anaerobes
(Trust and Sparrow, 1974; Horsley, 1977; Sakata, 1990;
Bairagi et al., 2002; Ghosh et al., 2002a). In the present
investigation, aerobic/facultative anaerobic extracellular
enzyme-producing bacterial symbionts were detected in
the GI tracts of 4 brackish water sh species. As the sh
were starved for 48 h and their GI tracts were thoroughly
Table 3. Qualitative extracellular enzyme activity of some bacterial strains isolated from the GI tracts of the sh species examined.
Enzyme activities were presented as scores as described in the text.
Fish species
Bacterial
strains
Enzyme activity (scores)*
Total score
Isolated from Amylase
1
Cellulase
2
Protease
3
Lipase
4
Terapon jarbua
TJ3.1 PI 0 2 0 0 02
TJ1.2 MI 2 2 3 2 09
TJ1.1 DI 4 3 3 2 12
TJ2.3 DI 5 4 5 4 18
TJ4.1 DI 4 3 4 4 15
Scatophagus argus
SA3.1 PI 0 4 3 2 09
SA1.1 MI 4 3 3 4 14
SA3.2 MI 0 2 3 3 08
SA2.2 DI 5 4 5 3 17
SA1.2 DI 5 3 3 2 13
SA2.1 DI 3 2 3 2 10
SA4.3 DI 0 4 3 2 09
Mystus
gulio
MG3.1 PI 0 2 0 4 06
MG4.1 MI 4 2 4 3 13
MG1.1 MI 0 3 4 2 09
MG4.2 DI 5 3 5 5 18
MG1.2 DI 4 2 3 2 11
Etroplus suratensis
ES3.1 PI 0 2 4 2 08
ES1.2 MI 0 2 4 0 06
ES2.1 DI 3 3 4 0 10
ES4.3 DI 2 2 4 0 08
*: With pure culture of bacterial isolates.
1
On starch (SA) plate;
2
on carboxymethylcellulose (CMC) plate;
3
on gelatin-peptone (GP) plate;
4
on tributyrin-agar (TA) plate.

84
DAS et al. / Turk J Zool
Table 4. Prole of specic enzyme activities (mean ± SE) in the selected isolates from the GI tracts of the sh species
examined.
Bacterial
strain
Enzyme activity (U)
Amylase* Cellulase
$
Protease
#
Lipase
@
MG4.1 26.82 ± 0.34
c
7.42 ± 0.24
a
20.31 ± 0.28
e
8.20 ± 0.27
c
SA2.2 44.03 ± 0.43
e
13.12 ± 0.23
d
20.95 ± 0.20
e
8.45 ± 0.51
c
TJ2.3 42.25 ± 0.70
e
11.76 ± 0.16
c
26.89 ± 0.28
g
8.79 ± 0.38
d
TJ4.1 34.19 ± 0.51
d
12.59 ± 0.16
d
22.87 ± 0.20
f
7.56 ± 0.52
b
MG4.2 36.98 ± 0.72
d
10.03 ± 0.13
b
11.78 ± 0.12
a
11.1 ± 0.31
e
TJ1.1 30.22 ± 0.69
c
10.80 ± 0.38
c
12.31 ± 0.32
b
5.66 ± 0.60
a
SA1.1 22.13 ± 0.08
b
9.95 ± 0.32
b
12.43 ± 0.27
b
7.78 ± 0.25
b
SA1.2 31.98 ± 0.16
c
9.1 ± 0.13
a
13.43 ± 0.38
c
5.82 ± 0.27
a
MG1.2 22.43 ± 0.12
b
9.61 ± 0.62
b
18.17 ± 0.43
d
8.92 ± 0.22
d
ES2.1 17.43 ± 0.27
a
8.41± 0.54
a
19.91 ± 0.33
d
Nil
Data are means ± SE of 3 determinations. Values with the same superscripts in the same vertical column are not
signicantly dierent (P < 0.05).
*:
1 unit (U) = 1 µg maltose liberated mL
–1
of enzyme-extract min
–1
#
: 1 unit (U) = 1 µg tyrosine liberated mL
–1
of enzyme-extract min
–1
$
: 1 unit (U) = 1 µg glucose liberated mL
–1
of enzyme-extract min
–1
@
: 1 unit (U) = 1 µmol fatty acid liberated mL
–1
of enzyme-extract min
–1
Bacillus subtilis JQ424889
Bacillus mojavensis AB021191
Bacillus amyloliquefaciens AB325583
Bacillus atrophaeus AB363731
Bacillus licheniformis MG4.2 KF377323
Bacillus licheniformis
NR074923.1
Bacillus aerius AJ831843
Brevibacillus brevis AB101593
Brevibacillus choshinensis AB112713
Brevibacillus agri AB112716
Brevibacillus parabrevis AB112714
Brevibacillus parabrevis SA2.2 KF377322
Brevibacillus parabrevis TJ2.3 KF377324
Alicyclobacillus pohliae AJ564766
9
9
78
68
51
10
0
88
10
0
10
0
9
8
43
10
0
0.02
Figure. e phylogenetic tree showing the relationship among Brevibacillus parabrevis
strains SA2.2 and TJ2.3, Bacillus licheniformis MG4.2, and their phylogenetically
closest type strains. e GenBank accession numbers of the type strains and studied
strains are shown following species names. Distance matrix was calculated by Kimura’s
2-parameter model. e scale bar indicates 0.02 substitutions per nucleotide position.
Alicyclobacillus pohliae AJ564766 served as an out-group.

85
DAS et al. / Turk J Zool
washed with sterile, chilled 0.9% saline prior to isolation
of microora, it may be suggested that the microorganisms
isolated in the present study belong to the autochthonous
adherent microora, as suggested elsewhere (Ghosh et
al., 2010). e rate of microora present within the GI
tract of sh is much higher than that of the surrounding
water, indicating that the GI tract of sh provides favorable
ecological niches for these microorganisms (Mondal et
al., 2008). However, isolation and identication alone
might not give a realistic depiction of the gut microora
in dierent regions of the GI tract with an appraisal of
their likely function (Khan and Ghosh, 2012). erefore, it
was considered legitimate in the present study to quantify
heterotrophic bacteria along with specic extracellular
enzyme-producing bacteria in dierent regions of
the GI tracts in the sh species studied, as the major
endeavor in the present study was to gather information
on extracellular enzyme-producing gut bacteria in some
brackish water shes.
In the present study, gut bacteria were isolated by
conventional culture-based methods. It is generally argued
that culture-dependent techniques are time-consuming,
lack accuracy (Ase et al., 2003), and do not represent a
correct picture of the bacterial diversity in the sh gut,
even if several dierent media are used (Ray et al., 2010).
However, the use of a culture-based technique employing
a specic substrate containing selective media is justiable,
as the major aim of the present study was to detect dierent
extracellular enzyme-producing gut bacteria. Besides, in
the present study, conventional methods in combination
with 16S rRNA analysis have been employed to identify
the potent enzyme-producing gut isolates, as suggested
elsewhere (Ghosh et al., 2010; Mondal et al., 2010; Ray et
al., 2010).
Proper information regarding the relative importance
of exogenous enzymes produced by the endosymbionts
of the GI tract and digestive enzymes produced by the
host is essential for understanding the contribution
of endosymbionts in digestion (Clements, 1997). In
the present investigation, all the sh species examined
exhibited considerable amylolytic, proteolytic, cellulolytic,
and lipolytic bacterial populations (Table 2). is can be
correlated with their feeding habits. Being omnivore sh
species, the occurrence of protease-, amylase-, cellulase-,
and lipase-producing bacterial populations in the digestive
tracts of S. argus, T. jarbua, and E. suratensis is justied.
e occurrence of proteolytic, cellulolytic, and amylolytic
bacteria in the gut has been suggested as an omnivorous
feeding aptitude of the sh by Creach (1963) and Ghosh
et al. (2002a, 2010). Previously, Bairagi et al. (2002)
failed to detect cellulolytic bacteria in the GI tracts of
carnivorous catsh and murrels; however, the results of the
present investigation showed the presence of cellulolytic
bacteria in carnivorous M. gulio. Stickney and Shumway
(1974) opined that omnivores and carnivores might
pick up cellulolytic ora from invertebrates that harbor
the bacteria, which might explain the presence of the
cellulolytic bacteria within the GI tract of M. gulio in the
present study. e present study indicated that cellulolytic
bacteria exist in the GI tracts of all the brackish water
sh species studied, which supports the hypothesis that
bacteria might contribute to the degradation of cellulose
in sh (Ray et al., 2010). e presence of a huge population
of cellulolytic bacteria and their vital role in extracellular
cellulase production in sh has been documented in
several investigations (Das and Tripathi, 1991; Saha and
Ray, 1998; Bairagi et al., 2002; Saha et al., 2006; Mondal
et al., 2008; Mondal et al., 2010). In their previous study
with carp, Shcherbina and Kazlawlene (1971) suggested
that cellulose absorption takes place in the DI, which may
indicate the presence of microbial cellulase in this region.
Our observation is in accordance with this hypothesis as
most cellulase-producing bacteria were recorded in the DI
of all sh species studied except in M. gulio, which was
supposed to be a carnivorous sh. Furthermore, it may be
mentioned that except for cellulolytic bacteria in M. gulio,
the heterotrophic microbial population was observed to be
highest in the DI regions of all the sh species studied when
compared to the PI and MI regions, which is in harmony
with previous reports (Mondal et al., 2008; Ghosh et al.,
2010; Ray et al., 2010).
e assay of extracellular enzyme production showed
the highest values for amylase and cellulase production in
SA2.2 isolated from the DI of S. argus. However, protease
and lipase productions were highest in TJ2.3 isolated
from the DI of T. jarbua and MG4.2 isolated from DI
of M. gulio, respectively. Qualitative and quantitative
determination of extracellular enzyme production
exhibited poor performance by the isolates from E.
suratensis when compared with the isolates from the other
shes studied. In the present study, the 2 ecient enzyme-
producing strains (SA2.2 and TJ2.3) were established
through quantitative enzyme assay and identied as
Brevibacillus parabrevis based on 16S rRNA sequence
analysis. Although both promising isolates belonged to B.
parabrevis, the strain SA2.2 isolated from DI of S. argus
showed the most similarity to B. parabrevis HDYM-18
(EF428244) (Ping and Ge, 2007; unpublished data), while
strain TJ2.3 isolated from DI of T. jarbua showed closeness
to B. parabrevis M3 (AB215101) (Suzuki et al., 2005;
unpublished data). Another strain, MG4.2, isolated from
the DI of M. gulio, showed 16S rRNA sequence similarity
to B. licheniformis GLU 113 (FN678352) (Shariati, 2010;
unpublished data). Diverse strains of extracellular enzyme-
producing bacteria have been identied from the GI tracts
of freshwater and marine shes (for review, see Ray et

86
DAS et al. / Turk J Zool
al., 2012). e occurrence of B. licheniformis within the
gut of freshwater shes has been reported previously by
several authors (Mondal et al., 2010; Dan and Ray, 2013).
However, to the authors’ knowledge, extracellular enzyme-
producing Brevibacillus sp. has not been reported from
sh gut previously. In addition, reports on gut-inhabiting
extracellular enzyme-producing bacteria from brackish
water sh species are scanty (De et al., 2012).
An extensive range of enzymes produced by GI bacteria
could be a contributing source of digestive enzymes in
sh (Ray et al., 2012). Characterization of the microbial
populations in the intestinal microenvironment of sh, and
understanding of the physiological interactions between the
indigenous microora and the host, might have important
implications (Silva et al., 2005). Enzymes produced by
intestinal sh microora might have a signicant role in
digestion, mainly for substrates such as cellulose, which
few animals can digest, and also for other substrates
(Smith, 1989). Luczkovich and Stellwag (1993) opined that
the GI microora of pinsh (Lagodon rhomboides) might
contribute to the breakdown of plant material. Kar et al.
(2008) indicated that the enzyme-producing gut bacteria
are able to utilize carbohydrates such as mannose, xylose,
ranose, cellobiose, and cellulose. ese substances are
mainly found in plant foodstus. erefore, cellulase and
amylase activities by the gut bacteria might indicate their
ability to aid in the digestion of plant foodstus. e use
of benecial bacteria as probiotics has a long tradition in
animal husbandry (Stavric and Kornegay, 1995). Benecial
bacteria could be introduced in commercial aquaculture
by incorporating them into formulated sh diets, or in the
form of bacteria biolm to achieve colonization in the GI
tract to a higher degree (Bairagi et al., 2002, 2004; Ghosh
et al., 2002b, 2003, 2004a, 2004b; Ramachandran et al.,
2005; Ramachandran and Ray, 2007; Askarian et al., 2011;
Saha and Ray, 2011). It has been suggested that benecial
gut bacteria are continuously competing with pathogens
through competitive exclusion (Ray et al., 2012). ese
topics could be addressed in upcoming studies. Whether
the gut bacteria can contribute to the host’s nutrition has
not been elucidated in the present study. Assessment of
the role of the enzyme-producing gut bacteria in brackish
water sh culture should therefore be given high priority
in future studies.
Acknowledgments
e authors are grateful to the Head of the Department
of Zoology, e University of Burdwan, West Bengal,
India; the Department of Science and Technology (FIST
Programme), New Delhi, India; and the University Grants
Commission (Special Assistance Programme), New
Delhi, India, for providing research facilities and nancial
support. e authors are obliged to Nilanjan Maitra for
rendering help in the analyses of sequenced data.
References
Ase, M., Yoshijima, T. and Sugita, H. 2003. Characterization of the
goldsh fecal microora by the uorescent in situ hybridization
method. Fisheries Sci. 69: 21–26.
Askarian, F., Kousha, A., Salma, W. and Ringø, E. 2011. e eect
of lactic acid bacteria administration on growth, digestive
enzymes activity and gut microbiota in Persian sturgeon
(Acipenser persicus) and beluga (Huso huso) fry. Aquacult.
Nutr. 17(5): 488–497.
Askarian, F., Zhou, Z., Olsen, R.E., Sperstad, S. and Ringø, E.
2012. Culturable autochthonous gut bacteria in Atlantic
salmon (Salmo salar L.) fed diets with or without chitin.
Characterization by 16S rRNA gene sequencing, ability to
produce enzymes, and in vitro growth inhibition of four sh
pathogens. Aquaculture 326–329: 1–8.
Austin, B. 2002. e bacterial microora of sh. Sci. World J. 2:
558–572.
Bairagi, A., Ghosh, K.S., Sen, S.K. and Ray, A.K. 2002. Enzyme
producing bacterial ora isolated from sh digestive tracts.
Aquacult. Int. 10: 109–121.
Bairagi, A., Sarkar Ghosh, K., Sen, S.K. and Ray, A.K. 2004.
Evaluation of nutritive value of Leucaena leucocephala leaf
meal inoculated with sh intestinal bacteria Bacillus subtilis
and Bacillus circulans in formulated diets for rohu, Labeo rohita
(Hamilton) ngerlings. Aquacult. Res. 35: 436–446.
Belchior, S.G. and Vacca, G. 2006. Fish protein hydrolysis by a
psychrotrophic marine bacterium isolated from the gut of hake
(Merluccius hubbsi). Can. J. Microbiol. 52: 1266–1271.
Bernfeld, P. 1955. Amylase. In: Methods in Enzymology (Eds. S.P.
Colowick and N.O. Kaplan). Academic Press, New York, NY,
USA, pp. 149–158.
Beveridge, M.C.M., Sikdar, P.K., Frerichs, G.N. and Millar, S. 1991.
e ingestion of bacteria in suspension by the common carp
Cyprinus carpio L. J. Fish Biol. 39: 825–831.
Bier, M. 1955. Methods in Enzymology. Vol. I. Academic Press, New
York, NY, USA, pp. 627–642.
Clements, K.D. 1997. Fermentation and gastrointestinal
microorganisms in shes. In: Gastrointestinal Microbiology.
Vol I. Gastrointestinal Ecosystems and Fermentations (Eds.
R.I. Mackie and B.A. White). Chapman and Hall, New York,
NY, USA, pp. 156–198.
Creach, P.V. 1963. e proteolytic enzymes of poisons. In: Nutrition
in Poecilothermes (Ed. CNRS). CNRS, Paris, France, pp. 375–
471.
Dan, S.K. and Ray, A.K. 2013. Characterization and identication of
phytase-producing bacteria isolated from the gastrointestinal
tract of four freshwater teleosts. Ann. Microbiol. DOI: 10.1007/
s13213-013-0664-3.

87
DAS et al. / Turk J Zool
Danulat, E. and Kausch, H. 1984. Chitinase activity in the digestive
tract of cod, Gadus morhua (L.). J. Fish Biol. 24: 125–133.
Das, K.M. and Tripathi, S.D. 1991. Studies on the digestive enzymes
of grass carp, Ctenopharyngodon idella (V). Aquaculture 92:
21–32.
De, D., Ghoshal, T.K. and Ananda Raja, R. 2012. Characterization
of enzyme-producing bacteria isolated from the gut of Asian
seabass, Lates calcarifer and milksh, Chanos chanos and
their application for nutrient enrichment of feed ingredients.
Aquacult. Res. DOI:10.1111/are.12099.
Denison, D.A. and Koehn, R.D. 1977. Cellulase activity of Poronia
oedipus. Mycologia 69: 592–601.
Esakkiraj, P., Immanuel, G., Sowmya, S.M., Iyapparaj, P. and
Palavesam, A. 2009. Evaluation of protease-producing ability
of sh gut isolate Bacillus cereus for aqua feed. Food Bioprocess
Technol. 2(4): 383–390.
Finegold, S.M., Sutter, V.L. and Mathisen, G.E. 1983. Normal
indigenous intestinal ora. In: Human Intestinal Microora in
Health and Disease (Ed. D.J. Hentgens). Academic Press, New
York, NY, USA, pp. 3–31.
Ghosh, K., Roy, M., Kar, N. and Ringø, E. 2010. Gastrointestinal
bacteria in rohu, Labeo rohita (Actinopterygii: Cypriniformes:
Cyprinidae): scanning electron microscopy and bacteriological
study. Acta Ichthyol. Piscatoria 40(2): 129–135.
Ghosh, K., Sen, S.K. and Ray, A.K. 2002a. Characterization of bacilli
isolated from the gut of rohu, Labeo rohita, ngerlings and its
signicance in digestion. J. Appl. Aquacult. 12(3): 33–42.
Ghosh K., Sen, S.K. and Ray, A.K. 2002b. Growth and survival of
rohu, Labeo rohita (Hamilton) spawn fed diets supplemented
with sh intestinal microora. Acta Ichthyol. Piscatoria 32(1):
83–92.
Ghosh, K., Sen, S.K. and Ray, A.K. 2003. Supplementation of an iso-
lated sh gut bacterium, Bacillus circulans, in formulated di-
ets for rohu, Labeo rohita, ngerlings. Israeli J. Aquacult. 55:
13–21.
Ghosh, K., Sen, S.K. and Ray, A.K. 2004a. Growth and survival of
rohu, Labeo rohita (Hamilton, 1822) spawn fed diets fermented
with intestinal bacterium, Bacillus circulans. Acta Ichthyol.
Piscatoria 34(2): 155–165.
Ghosh, K., Sen, S.K. and Ray, A.K. 2004b. Supplementation of
Lactobacillus acidophilus in compound diets for Labeo rohita
ngerlings. Indian J. Fish. 51(4): 521–526.
Horsley, R.W. 1977. A review of the bacterial ora of teleosts and
elasmobranches, including methods for its analysis. J. Fish Biol.
10: 529–553.
Itoi, S., Okamura, T., Koyama, Y. and Sugita, H. 2006. Chitinolytic
bacteria in the intestinal tract of Japanese coastal shes. Can. J.
Microbiol. 52: 1158–1163.
Jacob, M.B. and Gerstein, M.J. 1960. Handbook of Microbiology. D.
Van Nostrand Co. Inc., Princeton, NJ, USA.
Kar, N., Roy, R.N., Sen, S.K. and Ghosh, K. 2008. Isolation and
characterization of extracellular enzyme-producing bacilli
in the digestive tracts of rohu, Labeo rohita (Hamilton) and
murrel, Channa punctatus (Bloch). Asian Fish. Sci. 21(4):
421–434.
Khan, A. and Ghosh, K. 2012. Characterization and identication of
gut-associated phytase-producing bacteria in some freshwater
sh cultured in ponds. Acta Ichthyol. Piscatoria 42(1): 37–45.
Khan, A. and Ghosh, K. 2013. Evaluation of phytase production by
sh gut bacterium, Bacillus subtilis for processing of Ipomea
aquatica leaves as probable aquafeed ingredient. J. Aquat. Food
Prod. Technol. 22(5): 508–519.
Khan, A., Mandal, S., Samanta, D., Chatterjee, S. and Ghosh, K. 2011.
Phytase-producing Rhodococcus sp. (MTCC 9508) from sh
gut: a preliminary study. Proc. Zool. Soc. 64: 29–34.
Kinnear P.R. and Gray, C.D. 2000. SPSS for Windows Made Simple.
Release 10. Psychology Press, Sussex, UK.
Lane, D.J. 1991. 16S/23S rRNA sequencing. In: Nucleic Acid
Techniques in Bacterial Systematics (Eds. E. Stackebrandt and
M. Goodfellow). Wiley, Chichester, UK, pp. 130–141.
Li, X., Chi, Z., Liu, Z., Yan, K. and Li, H. 2008. Phytase production
by a marine yeast Kodamea ohmeri BG3. Applied Biochem.
Biotechnol. 149: 183–193.
Li, X.Y., Liu, Z.Q. and Chi, Z.M. 2008. Production of phytase by a
marine yeast Kodamaea ohmeri BG3 in an oats medium:
optimization by response surface methodology. Bioresour.
Technol. 99: 6386–6390.
Luczkovich, J.J. and Stellwag, E.J. 1993. Isolation of cellulolytic
microbes from the intestinal tract of pinfish, Lagodon
rhomboids: size-related changes in diet and microbial
abundance. Mar. Biol. 116: 381–388.
MacDonald, N.L., Stark, J.R. and Austin, B. 1986. Bacterial microora
in the gastrointestinal tract of Dover sole (Solea solea L.), with
emphasis on the possible role of bacteria in the nutrition of the
host. FEMS Microbiol. Lett. 35: 107–111.
Mandal, S. and Ghosh, K. 2013. Isolation of tannase-producing
microbiota from the gastrointestinal tracts of some freshwater
sh. J. Appl. Ichthyol. 29: 145–153.
Mondal, S., Roy, T. and Ray, A.K. 2010. Characterization and
identication of enzyme-producing bacteria isolated from the
digestive tract of bata, Labeo bata. J. World Aquacult. Soc. 41:
369–377.
Mondal, S., Roy, T., Sen, S.K. and Ray, A.K. 2008. Distribution of
enzyme-producing bacteria in the digestive tracts of some
freshwater sh. Acta Ichthyol. Piscatoria 38: 1–8.
Rahmatullah, S.M. and Beveridge, M.C.M. 1993. Ingestion of bacteria
in suspension by Indian major carps (Catla catla, Labeo rohita)
and Chinese carps (Hypophthalmichthys molitrix, Aristichthys
nobilis). Hydrobiologia 264: 79–84.
Ramachandran, S. and Ray, A.K. 2007. Nutritional evaluation
of fermented black gram (Phaseolus mungo) seed meal in
compound diets for rohu, Labeo rohita (Hamilton), ngerlings.
J. Appl. Ichthyol. 23: 74–79.

88
DAS et al. / Turk J Zool
Ramachandran, S., Bairagi, A. and Ray, A.K. 2005. Improvement of
nutritive value of grass pea (Lathyrus sativus) seed meal in the
formulated diets for rohu, Labeo rohita (Hamilton) ngerlings
aer fermentation with a sh gut bacterium. Bioresour.
Technol. 96: 1465–1472.
Rani, S., Garg, S.K., Sabhlok, V.P. and Bhatnagar, A. 2004. Intestinal
enzyme activity and enzyme-producing microbial ora in
relation to feeding behaviour in some brackish water teleosts.
J. Aquacult. 12: 55–68.
Ray, A.K., Ghosh, K. and Ringø, E. 2012. Enzyme-producing bacteria
isolated from sh gut: a review. Aquacult. Nutr. 18: 465–492.
Ray, A.K., Roy, T., Mondal, S. and Ringø, E. 2010. Identication of
gut-associated amylase, cellulase, and protease-producing
bacteria in three species of Indian major carps. Aquacult. Res.
41:1462–1469.
Ringø, E. 1993. e eect of chromic oxide (Cr
2
O
3
) on aerobic
bacterial populations associated with the epithelial mucosa
of Arctic charr (Salvelinus alpinus L.). Can. J. Microbiol. 39:
1169–1173.
Ringø, E. and Birkbeck, T.H. 1999. Intestinal microora of sh larvae
and fry. Aquacult. Res. 30: 73–93.
Ringø, E., Olsen, R.E., Gifstad, T.O., Dalmo, R.A., Amlund, H.,
Hemre, G.I. and Bakke, A.M. 2010. Prebiotics in aquaculture: a
review. Aquacult. Nutr. 16: 117–136.
Ringø, E., Olsen, R.E., Mayhew, T.M. and Myklebust, R. 2003.
Electron microscopy of the intestinal microora of sh.
Aquaculture 227: 395–415.
Roy, T., Mondal, S. and Ray, A.K. 2009. Phytase-producing bacteria
in the digestive tracts of some freshwater sh. Aquacult. Res.
40: 344–353.
Saha, A. and Ray, A.K. 1998. Cellulase activity in rohu ngerlings.
Aquacult. Int. 6: 281–291.
Saha, S. and Ray, A.K. 2011. Evaluation of nutritive value of water
hyacinth (Eichhornia crassipes) leaf meal in compound diets
for rohu, Labeo rohita (Hamilton, 1822) ngerlings aer
fermentation with two bacterial strains isolated from sh gut.
Turkish J. Fish. Aqua. Sci. 11: 199–209.
Saha, S., Roy, R.N., Sen, S.K. and Ray, A.K. 2006. Characterization of
cellulase-producing bacteria from the digestive tract of tilapia,
Oreochromis mossambica (P) and grass carp, Ctenopharyngodon
idella (V). Aquacult. Res. 37: 380–388.
Sakata, T. 1990. Microora in the digestive tract of sh and shellsh.
In: Microbiology in Poecilotherms (Ed. R. Lesel). Elsevier,
Amsterdam, the Netherlands, pp. 171–176.
Sangiliyandi, G. and Gunasekaran, P. 1996. Extracellular lipase-
producing Bacillus licheniformis from an oil mill renery
euent. Indian J. Microbiol. 36: 109–110.
Shcherbina, M.A. and Kazlawlene, O.P. 1971. e reaction of the
medium and the rate of absorption of nutrients in the intestine
of carp. J. Ichthyol. 11: 81–85.
Silva, F.C.P., Brito, M.F.G., Brito, L.M. and Nicoli, J.R. 2005.
Composition and antagonistic activity of the indigenous
intestinal microbiota of Prochilodus argenteus Agassiz. J. Fish
Biol. 67: 1686–1698.
Sivasubramanian, K., Ravichandran, S. and Kavitha, R. 2012. Isola-
tion and characterization of gut microbiota from some estua-
rine shes. Marine Sci. 2(2): 1–6.
Smith, L.S. 1989. Digestive functions in teleost shes. In: Fish
Nutrition (Ed. J.E. Halver). Academic Press, San Diego, CA,
USA, pp. 331–42.
Stavric, S. and Kornegay, E.T. 1995. Microbial probiotics for pigs
and poultry. In: Biotechnology in Animal Feeds and Animal
Feeding (Eds. R.J. Wallace and A. Chuson). Wiley, New York,
NY, USA, pp. 205–231.
Stickney, R.R. and Shumway, S.E. 1974. Occurrence of cellulase
activity in the stomachs of shes. J. Fish Biol. 6: 779–790.
Sugita, H., Shibuya, K., Harada, H. and Deguchi, Y. 1997.
Antibacterial abilities of intestinal microora of the river sh.
Fisheries Sci. 63: 378–383.
Teather, R.M. and Wood, P.J. 1982. Use of Congo red-polysaccharide
interactions in enumeration and characterization of cellulolytic
bacteria from the bovine rumen. Appl. Environ. Microbiol. 43:
777–780.
Trust, T.J. and Sparrow, R.A. 1974. e bacterial ora in the
alimentary tract of fresh water salmonid shes. Can. J.
Microbiol. 20: 1219–1228.
Walter, H.E. 1984. Methods of Enzymatic Analysis. Verlag Chemie,
Weinheim, Germany.
Zar, J.E. 1974. Biostatistical Analysis. Prentice Hall, NJ, USA.