RESEARCH ARTICLE

Variation in the hindgut microbial communities of the Florida

manatee, Trichechus manatus latirostris over winter in Crystal
River, Florida
Samuel D. Merson1, Diane Ouwerkerk2,3, Lisa-Maree Gulino2,3, Athol Klieve3,4, Robert K. Bonde5,
Elizabeth A. Burgess1 & Janet M. Lanyon1
1
Marine Vertebrate Ecology Research Group, School of Biological Sciences, The University of Queensland, Brisbane, Qld, Australia; 2Rumen

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

Ecology Unit, Agri-Science Queensland, Department of Agriculture, Fisheries and Forestry, Ecosciences Precinct, Brisbane, Qld, Australia; 3Centre
for Animal Science, Queensland Alliance for Agriculture and Food Innovation, The University of Queensland, Brisbane, Qld, Australia; 4School of
Agricultural and Food Sciences, The University of Queensland, Brisbane, Qld, Australia; and 5Sirenia Project, U.S. Geological Survey, Southeast
Ecological Science Centre, Gainesville, FL, USA

Correspondence: Janet M. Lanyon, Marine Abstract

Vertebrate Ecology Research Group, School
of Biological Sciences, The University of The Florida manatee, Trichechus manatus latirostris, is a hindgut-fermenting
Queensland, Brisbane, St Lucia QLD 4072, herbivore. In winter, manatees migrate to warm water overwintering sites
Australia. where they undergo dietary shifts and may suffer from cold-induced stress.
Tel.: +617 336 54416; Given these seasonally induced changes in diet, the present study aimed to
fax: +617 336 51655;
examine variation in the hindgut bacterial communities of wild manatees over-
e-mail: j.lanyon@uq.edu.au
wintering at Crystal River, west Florida. Faeces were sampled from 36 manatees
Received 13 May 2013; revised 30 October
of known sex and body size in early winter when manatees were newly arrived
2013; accepted 5 November 2013. Final and then in mid-winter and late winter when diet had probably changed and
version published online 9 December 2013. environmental stress may have increased. Concentrations of faecal cortisol
MICROBIOLOGY ECOLOGY

metabolite, an indicator of a stress response, were measured by enzyme immu-

DOI: 10.1111/1574-6941.12248 noassay. Using 454-pyrosequencing, 2027 bacterial operational taxonomic units
were identified in manatee faeces following amplicon pyrosequencing of the
Editor: Julian Marchesi 16S rRNA gene V3/V4 region. Classified sequences were assigned to eight pre-
viously described bacterial phyla; only 0.36% of sequences could not be classi-
Keywords
fied to phylum level. Five core phyla were identified in all samples. The
bacteria; faeces; colon; 16S rRNA gene;
454-pyrosequencing. majority (96.8%) of sequences were classified as Firmicutes (77.3  11.1% of
total sequences) or Bacteroidetes (19.5  10.6%). Alpha-diversity measures
trended towards higher diversity of hindgut microbiota in manatees in mid-
winter compared to early and late winter. Beta-diversity measures, analysed
through PERMANOVA, also indicated significant differences in bacterial commu-
nities based on the season.

manatee movement and foraging patterns (Deutsch et al.,

Introduction
2003). During spring and summer months, manatees feed
The Florida manatee (Trichechus manatus latirostrus), a on a wide variety of high-nutrient plants. However, aqua-
subspecies of the West Indian manatee, is a generalist tic plants undergo seasonal fluctuations in standing crop,
herbivore, consuming up to 60 species of freshwater energy content, epiphyte load and proximate constituents
plants and seagrasses (Campbell & Irvine, 1977; Best, (e.g. protein) (Deutsch et al., 2003), so that in winter, less
1981; Reep & Bonde, 2006; Alves-Stanley et al., 2010). and poorer quality forage is available (Gilbert & Clark,
Diet composition appears to depend on forage availability 1981; Deutsch et al., 2003).
(Campbell & Irvine, 1977), and regional variation in In response to decreasing temperatures, Florida mana-
quantity and quality of aquatic vegetation is thought to tees undertake annual migrations to warm water sources
be one of the major factors determining intraseasonal (Langtimm et al., 1998; Deutsch et al., 2003, 2006; Laist

FEMS Microbiol Ecol 87 (2014) 601–615 ª 2013 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
602 S.D. Merson et al.

& Reynolds, 2005). These seasonal migrations are usually other herbivores are generally highly responsive to factors
accompanied by pronounced dietary shifts, including such as diet, age, health, geographic location, seasons
from marine to freshwater plants, which reflect availabil- and feeding regime (Yu & Forster, 2005), so that we
ity of local forage along the migration path and within might expect to see microbial variation in response to
commuting distance of overwintering sites (Deutsch dietary change in manatees. A study using denaturing
et al., 2003). Manatees show strong annual site fidelity to gradient gel electrophoresis (DGGE) to examine the bacte-
winter refugia and then spend on average fifty percentage rial community of the related dugong, a more specialised
of their time at the site throughout the winter period grazer, suggests that its hindgut microbial community is
(Deutsch et al., 2006). However, due to lack of sufficient responsive to shifts in diet, including those that occur
local vegetation at these refugia, manatees often commute during ontogenetic development (Eigeland et al., 2012).

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

to feeding grounds up to 5 km away (Deutsch et al., A preliminary study by Mao et al. (2011) investigated
1998). If winter temperatures are particularly cold, dura- the microbial communities of the skin and hindgut of
tion and commuting distance of feeding forays may be wild West Indian manatees (Trichechus manatus) from
shortened; prolonged cold periods have resulted in mana- eastern Florida (Atlantic coast subpopulation) and Belize,
tees fasting for periods of a week or more (Deutsch et al., using 454-amplicon pyrosequencing, DGGE and terminal
2006), presumably suffering some degree of nutritional restriction fragment length polymorphism (T-RFLP), and
stress. found Bacteroidetes and Firmicutes to be the dominant
Crystal River (CR) on the central-west coast of Florida bacterial phyla. Using 454-pyrosequencing, this current
is a winter refuge, accessed via waterways from the Gulf study profiled the hindgut bacterial communities of live
of Mexico, with a series of warm (22 °C) freshwater wild manatees from a different subpopulation overwinter-
springs that support more than 600 manatees from ing in CR, central-west Florida, and investigated the rela-
November through February (Hartman, 1979; J. Kleen, tionships amongst the gut microbiota and each of
USFWS, pers. commun.). As the manatee population ontogeny (size class), assumed dietary change (different
utilising the warm spring waters increases during winter, winter periods) and chronic stress (including thermal
benthic plants are overgrazed locally so that manatees and/or nutritional).
must move greater distances to forage, which may result
in a shift of diet (J.P. Reid, USGS, unpublished data).
Materials and methods
The effects of such dietary shifts on the digestive and
stress physiology of the manatee are unknown. Manatees
Faecal sampling
have evolved effective digestive systems for extracting
nutrients from plants that are typically high in fibre and Freshly voided faeces were collected from live free-ranging
low in protein and caloric value (Burn, 1986; Burn & Florida manatees, Trichechus manatus latirostris, at Three
Odell, 1987; Reep & Bonde, 2006). Their well-developed Sisters Spring in CR, Citrus County, central-west Florida,
molariform dentition aids in physical breakdown of USA, one of the most important winter refuge sites for
fibrous and/or abrasive plants (Domning, 1983) with this species (Nabor & Patton, 1989; Reep & Bonde,
subsequent fermentation of plant fibre by symbiotic mic- 2006). Faecal samples were collected on three separate
robiota in the hindgut (Burn, 1986). Fibre fermentation occasions during winter 2010: ‘early winter’ (November–
is a slow process but is optimised in the manatee by the mid-December) when manatees had newly arrived from
capacious hindgut (20–30 m long) that weighs up to summer foraging grounds, ‘mid-winter’ (mid-December–
70% of total gut mass (Reynolds, 1980; Reynolds & January) when manatees had become established within
Rommel, 1996) with slow digesta passage rates of their refuge and ‘late winter’ (February–March) when the
4–10 days compared to 2 days in the horse (Lomolino & effects of any cold-related stress would presumably be
Ewel, 1984; Larkin et al., 2007). Apparent digestibility of most apparent. The objective was to collect faecal samples
fibre is significantly higher than in most mammalian her- from at least three manatees of each sex and body size
bivores at 83–91% (Lomolino, 1977; Burn & Odell, 1987; class (calves < 245 cm body length, subadults 246–
Reynolds & Rommel, 1996). The composition and func- 265 cm, adults > 265 cm: Tripp et al., 2011) at each of
tion of microbial communities have not been well stud- the three times; however, the final sample size was depen-
ied in the Florida manatee. It is likely that these gut dent on which animals visited the spring on the sampling
microorganisms are important contributing factors in the days. A total of 36 manatees were sampled including 14
manatee’s ability to efficiently digest fibre and assimilate manatees from early winter, 10 from mid-winter and 12
nutrients and in their capacity to consume a broad diet from late winter (Table 1). Faeces were collected from
or undertake frequent dietary shifts without serious each manatee by approaching quietly in water and duck-
consequence. Gastrointestinal microbial communities in diving to remove, by hand, the faecal plug extruding from

ª 2013 Federation of European Microbiological Societies. FEMS Microbiol Ecol 87 (2014) 601–615
Published by John Wiley & Sons Ltd. All rights reserved
Microbial communities in the manatee hindgut 603

Table 1. Biological data for faecal samples collected from 36 live wild Florida manatees from CR, western Florida, including winter period in
which sample was collected (EW – early winter, MW – mid-winter, LW – late winter), sample collection number and sequenced sample identifier
code (identifier), sex, age class based on body size (A – adult, SA – subadult, C – calf), faecal cortisol level (value ng g 1 and relative level/
category), number of microbial sequences obtained and analysed, and number of microbial OTUs identified in each sample
1
Winter period Sample no. Identifier Sex Age class Cortisol ng g Cortisol category No. of sequences No. of OTUs
EW 10-322-F1 AMEW1 Male A 18.33 Low 5970 155
EW 10-322-F2 AFEW2 Female A 20.77 Medium 7636 397
EW 10-336-F3 AMEW3 Male A 14.23 Low 7828 364
EW 10-336-F5 AFEW4 Female A 11.18 Low 5017 258
EW CCR-10-11 AFEW5 Female A 19.79 Low 3795 314

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

EW 10-341-F1 AFEW6 Female A 28.11 Medium 3967 364
EW 10-336-F1 SAFEW9 Female SA 8.34 Low 11 804 509
EW 10-336-F2 SAMEW10 Male SA 10.56 Low 5365 346
EW 10-336-F4 SAFEW11 Male SA 7.27 Low 4792 347
EW Liz 1 SAUknEW12 Unknown SA 8.5 Low 4243 261
EW 10-336-F6 CFEW7 Female C 26.35 Medium 4749 303
EW 10-336-F7 CFEW8 Female C 12.06 Low 4840 276
MW 10-355-F2 AFMW13 Female A 12.51 Low 9866 581
MW 10-363-F2 AMMW14 Male A 8.31 Low 7514 427
MW 10-363-F3 AFMW15 Female A 32.51 Medium 6506 555
MW 11-007-F2 AMMW16 Male A 47.22 High 4954 489
MW 11-007-F6 AMMW17 Male A 24.46 Medium 4712 327
MW 11-007-F8 AFMW18 Female A 62.76 High 4552 446
MW 10-355-F1 SAFMW21 Female SA 23.86 Medium 9264 638
MW 11-029-F3 SAMMW22 Male SA 46.66 High 4489 315
MW 10-363-F4 CMMW19 Male C 27.26 Medium 5152 379
MW 11-032-F1 CMMW20 Male C 25.01 Medium 3292 302
LW MF0055 AMLW23 Male A 15.64 Low 6943 369
LW MF0056 AMLW24 Male A 15.93 Low 4493 316
LW MF0058 AFLW25 Female A 16.93 Low 4637 349
LW MF0017 AFLW26 Female A 24.74 Medium 5052 237
LW MF0018 AMLW27 Male A 15.99 Low 9073 376
LW MF0021 AFLW28 Female A 62.77 High 6616 300
LW MF0057 SAFLW32 Female SA 16.76 Low 6436 222
LW MF0020 SAMLW33 Male SA 17.53 Low 6593 541
LW MF0023 SAMLW34 Male SA 23.42 Medium 7168 255
LW MF0024 SAMLW35 Male SA 25.09 Medium 7599 395
LW MF0053 SAFLW36 Female SA 27.96 Medium 5846 366
LW MF0059 CMLW29 Male C 12.54 Low 5917 246
LW MF0060 CMLW30 Male C 12.86 Low 7878 279
LW MF0022 CMLW31 Male C 50.01 High 6665 456

the anus. Samples were subsequently removed from the DNA was quantified using a Nanodrop 8000 Spectro-
core of the faeces to avoid environmental contamination photometer (Thermo Scientific, Wilmington, DE). Sam-
and frozen at 80 °C until analysis. Manatees were sexed ples were diluted to a final concentration of 10 ng lL 1
by examining the anal-genital region (Lanyon et al., to ensure sample standardisation for subsequent PCR
2009). Body size class was estimated by experienced assays. A 4 lL aliquot of extracted DNA was mixed with
researchers, and on occasion, it was measured in water to 1 lL of 6X loading buffer (Fermentas, Hanover, MD)
ensure precision. and loaded onto a 1% agarose gel in 1X TBE buffer
containing the DNA stain GelRed (Jomar Biosciences,
Kensington, SA, Australia). A 2 lL aliquot of size mar-
DNA extraction
ker, 1 Kb DNA ladder (Fermentas), was loaded and gel
Genomic DNA was extracted from 0.25 g of each faecal electrophoresed at 120 V for 35 min. The DNA was vis-
sample using the PowerSoil DNA kit (MO BIO Labora- ualised under UV light, and an image obtained using the
tories, Carlsbad, CA) following the manufacturer’s Bio-Rad GelDoc system (Bio-Rad Laboratories, Hercules,
instructions. The concentration of recovered genomic CA).

FEMS Microbiol Ecol 87 (2014) 601–615 ª 2013 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
604 S.D. Merson et al.

and 470 bp in length were retained. Operational taxo-

454-amplicon pyrosequencing barcoded PCR
nomic units (OTUs) were defined at the 97% sequence
Extracted gDNA from each of the manatee faecal samples similarity level using Uclust, and a sequence representa-
was used as a template to prepare samples for pyrose- tive of the OTU was selected based on the most abun-
quencing as described below. The V3/V4 region of the dant sequence found in each OTU. Representative
16S rRNA gene from each sample was PCR amplified OTUs were classified using the Ribosomal Database
using the primers, 341F (5′CCATCTCATCCCTGCGTGT Project (RDP) classifier (Cole et al., 2009) and the 2011
CTCCGAC -3′) and 787R (5′-CCTATCCCCTGTGTGCTT Greengenes 16S rRNA gene database (http://www.
GGCAGTC-3′) (Lane et al., 1985) modified to incorporate secondgenome.com/go/2011-greengenes-taxonomy/). The
a 10 bp error-correcting DNA barcode (Hamady et al., sequences were aligned using the python nearest align-

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

2008). PCRs were prepared to a final volume of 50 lL ment space termination (PYNAST) tool (Caporaso et al.,
with the following reagents and final concentrations: 1 2010b) against the 16S rRNA Greengenes core set (De
Unit of Phusion polymerase (Finnzymes Oy, Espoo, Santis et al., 2006). The relative abundance of sequences
Finland); 250 mM of (each) forward and reverse primers; at the phylum level was calculated from the total num-
200 nM dNTP mix (Roche Diagnostics, Castle Hill, ber of classified sequences for each sample. A phyloge-
NSW, Australia); HF Phusion Buffer and 20 ng of tem- netic tree of the representative OTUs was constructed
plate DNA. Thermocycling was performed in a Bio-Rad using the tree-building program FASTTREE and used for
C100 (Bio-Rad Laboratories, Richmond, CA) using the downstream processing (e.g. calculating beta-diversity -
following conditions: lid preheated to 101 °C; initial Price et al., 2010). A readable matrix of OTU by sample
denaturation at 98 °C for 30 s; followed by 30 cycles of: was created, and all OTUs and core OTUs (i.e. OTUs
98 °C, 10 s; 65 °C, 20 s; 72 °C, 15 s; followed by a final that occur in all samples) were visualised as a heatmap
extension step at 72 °C, 10 min and a ‘cooling’ step at generated in RStudio: integrated development environ-
12 °C, 5 min. The PCR product (50 lL) was mixed with ment for R version 0.96.122 (R Development Core Team,
10 lL of loading buffer (Fermentas) and split between 2008) (http://www.rstudio.org), using the gplots and
two wells of a 2% agarose gel. A 5 lL aliquot of 100 bp heatmap2 functions. Rarefaction plots were created for
DNA marker (Fermentas) was loaded into wells 1 and 10, individual samples to assess sampling coverage and levels
and the gel electrophoresed (100 V, 30 min) followed by of diversity. Boxplots were generated for the top five
poststaining with 3X GelRed. PCR amplicons of the cor- most abundant groups at the class and genus level, using
rect size were excised in a dark room under UV light R version 0.96.122 (R Development Core Team, 2008)
using GelX 6.5 gel excision tips (Cleaver Scientific, Rugby, (http://www.rstudio.org). The microbial diversity (alpha-
UK). The excised bands were purified using a QIAquick diversity) within a sample was determined by calculating:
Gel Extraction Kit (Qiagen, Valencia, CA) and quantified Chao1, phylogenetic distance, observed species (no. of
using the Qubit fluorometer (Invitrogen, Carlsbad, CA) OTUs), Shannon–Wiener index, Simpson index and the
as per manufacturer’s instructions. C. 300 ng of each number of singletons; at a depth of 3300 sequences (i.e.
purified PCR amplicon was sent to the Australian Gen- minimum rounded number of sequences in a sample).
ome Research Facility (AGRF, Brisbane, Australia) for The computation of diversity between samples (beta-
pyrosequencing on a 454 Genomic Sequencer (Roche diversity) was then compared through the generation of
Diagnostics, Indianapolis, IN). 2D principal coordinates analysis (PCoA) plots using
QIIME. PERMANOVA was conducted in QIIME, using 9999
permutations and the categorical variables of season and
454-amplicon pyrosequencing data analysis
age class.
Data were returned as unaltered standard flowgram files Denoised sequences and metadata were submitted to
(sample.sff). Mothur version 1.1.8.0 (Schloss et al., 2009) MG-RAST (http:/metagenomics.anl.gov) and are stored
was used to convert .sff files into .sff.txt; .fasta and .qual under the project name ‘Manatee gut microbiology’ –
files. Data were denoised using ACACIA (Bragg et al., record numbers for each amplicon library are as follows:
2012), and the denoised data processed using the Quan- 4521911.3, 4521901.3, 4521912.3, 4521902.3, 4521903.3,
titative Insights Into Microbial Ecology (QIIME version 4521904.3, 4521919.3, 4521920.3, 4521927.3, 4521931.3,
1.5.0) software pipeline (Caporaso et al., 2010a). Specifi- 4521926.3, 4521936.3, 4521908.3, 4521916.3, 4521909.3,
cally, sequences were deconvoluted into individual sam- 4521917.3, 4521918.3, 4521910.3, 4521924.3, 4521925.3,
ples based on their barcodes and simultaneously filtered 4521930.3, 4521935.3, 4521913.3, 4521914.3, 4521905.3,
to remove poor/low-quality sequences. Sequences with 4521906.3, 4521915.3, 4521907.3, 4521921.3, 4521922.3,
an average quality value of < 25 were discarded, across a 4521923.3, 4521928.3, 4521932.3, 4521933.3, 4521934.3,
sliding window of 50 bp, and sequences between 430 4521929.3.

ª 2013 Federation of European Microbiological Societies. FEMS Microbiol Ecol 87 (2014) 601–615
Published by John Wiley & Sons Ltd. All rights reserved
Microbial communities in the manatee hindgut 605

number of sequences ranged from 3292 to 11804, and the

Hormone extraction and cortisol analysis
number of OTUs from 155 to 638 for individual mana-
Manatee faecal samples were freeze-dried in a lyophiliser tees (Table 1). Rarefaction analysis suggested that pyrose-
and pulverised until homogeneous. Faecal steroid quencing depth was insufficient to adequately uncover
hormones were extracted by adding 4 mL of 80% methanol the full diversity within the samples (Fig. 1), and this was
to 0.2  0.01 g aliquot of dried faecal sample in a glass supported by estimates of Good’s coverage (0.45–0.64).
scintillation vial (Wasser et al., 1994). Samples were sealed, Alpha-diversity measures indicated that bacterial diversity
vortexed (30 s) until homogenised and then put on a rotat-
ing mixer overnight (minimum of 12 h) to solubilise the
steroid hormones from the faecal mass. Samples were (a)

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

Key Early winter (3% dissimilarity)
centrifuged for 15 min at 1790 g and the methanol super-
Adult
natant was decanted for storage at 20 °C.

Number of OTUs
Concentrations of cortisol metabolites in faecal extracts Sub-adult
were measured using enzyme immunoassay following Calf
Burgess et al. (2013). Microtitre plates were coated with
50 lL of antibody (1 : 8500 R4866; C. Munro, University
of California, Davis, CA) per well. Plates were sealed with
an acetate sealer, incubated overnight at 4 °C, then
washed five times with 250 lL of wash solution and
drained to remove unbound antiserum. Manatee samples
(diluted 1 : 2 faecal extract), standards (3.9–1000 pg/ Number of sequences
50 lL) and controls, made up in standard assay buffer,
(b)
were loaded in 50 lL volumes to wells and then 50 lL of Mid winter (3% dissimilarity)
antigen conjugated to horseradish peroxidase (1 : 15,000;
C. Munro, University of California) was added to each
Number of OTUs

well. Following 1 h incubation at room temperature, the

plates were washed and 100 lL of the colour substrate
solution was added to each well. Plates were sealed, sha-
ken for 5 min and left at room temperature until colour
developed and the optical density of the zero wells (assay
buffer only) reached 1.0. Plates were read using a micro-
plate reader (absorbance wavelength of 405 nm) with
Revelation software (Dynex MRX II, Q-Lab, Chantilly,
Number of sequences
VA). All samples, controls and standards were assayed in
duplicate, with the resulting coefficient of variation (c)
Late winter (3% dissimilarity)
between all duplicates required to be < 10% for accep-
tance. Variation in faecal cortisol levels was examined
Number of OTUs

against winter period (early, mid, late), and size class of

manatee (subadult, adult) using 2-way ANOVA on log10-
transformed data to meet assumptions of normality and
equal variance.

Results

Bacterial diversity in wild manatee faeces

Number of sequences
The total number of bacterial DNA sequences obtained
(postdenoising and filtering) from manatee faecal samples Fig. 1. Rarefaction curves of bacterial communities from faecal
samples of 36 wild Florida manatees sampled in CR, Florida, USA.
was 243 933, with a mean length of 425 bp. Overall, a
Number of sequences is plotted against the number of bacterial OTUs
total of 2027 bacterial OTUs was identified (at 97% (3% dissimilarity level) and is presented for each of the periods in
sequence similarity, postdenoising, filtering and chimera which the manatees were sampled (a) early winter, (b) mid-winter, (c)
removal), indicating a diverse microbiome present in the late winter. The different age classes based on body size are adult
hindgut of the 36 wild manatees sampled from CR. The (red), subadult (blue), calf (grey).

FEMS Microbiol Ecol 87 (2014) 601–615 ª 2013 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
606 S.D. Merson et al.

varied considerably between samples (Table 2). The high- for the level of phylum (Table 2) and the top five most
est bacterial diversity was detected in manatees sampled abundant groups at the order level (Fig. 2a) and genus
in mid-winter according to each of the diversity indices level (Fig. 2b). Eight bacterial phyla were identified in
(i.e. Chao1, Phylogenetic Diversity, Number of OTUs, manatee faecal samples with the two dominant phyla
Shannon–Wiener, Singletons) with the exception of the being Firmicutes and Bacteroidetes. There was a small
Simpson index (Table 2). number of sequences (0.36%) that could not be classified
The number and diversity of bacteria varied amongst to the phylum level, and these were assigned to the group
individual manatees at each level, and this is displayed Unclassified Bacteria (Tables 3 and 4). At deeper

Table 2. Mean alpha-diversity measures calculated for the bacterial communities of faeces from 36 wild Florida manatees sampled at CR,

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

Florida, USA, at a depth of 3300 sequences, grouped into categories based on the season the sample was collected; early winter, mid-winter and
late winter

Season Chao1 Phylogenetic diversity Number of OTUs Shannon Simpson Singletons

Early winter 489.94a 21.55a 256a 5.10a 0.92a 126.83a
Mid- winter 607.89b 26.51b 328b 5.87b 0.96a 166.4b
Late winter 438.65a 20.64a 244a 4.85a 0.91a 118.50a
Chao1, Phylogenetic diversity, the number of OTUs, Shannon–Wiener index, Simpson index and the number of singletons are presented. In gen-
eral, the higher the value, the greater was the diversity. Means with the same superscripts in a column are not significantly different at the
P < 0.01 level.

(a)
Early winter Adult
0.8

Mid winter
0.8
Sub-adult
Percent abundance (%)

Percent abundance (%)

Late winter Calf

0.6

0.6
0.4

0.4
0.2

0.2
0.0
0.0

Clostridiales Bacteroidales Coriobacteriales Burkholderiales Unclassified Clostridiales Bacteroidales Coriobacteriales Burkholderiales Unclassified
Bacteria Bacteria

ORDER ORDER

(b)
Early winter Adult
0.4

0.4

Mid winter Sub-adult

Percent abundance (%)

Percent abundance (%)

Late winter Calf

0.3

0.3
0.2
0.2

0.1
0.1

0.0
0.0

Clostridium Unclassified Unclassified Unclassified Epulopsicium Clostridium Unclassified Unclassified Unclassified Epulopsicium
Bacteroidales Clostridiaceae Clostridiales Bacteroidales Clostridiaceae Clostridiales

GENUS GENUS

Fig. 2. The top five most abundant groups of the bacterial community of faeces sampled from 36 wild Florida manatees with bacterial 16S
rRNA gene sequences taxonomically classified at the level of, (a) order, (b) genus for different age classes (adult, subadult, calf) and sampling
period (early winter, mid-winter, late winter).

ª 2013 Federation of European Microbiological Societies. FEMS Microbiol Ecol 87 (2014) 601–615
Published by John Wiley & Sons Ltd. All rights reserved
Microbial communities in the manatee hindgut 607

Table 3. Mean relative abundance of the OTUs classified to eight known phyla and OTUs unable to be classified to known phyla (Unclassified
Bacteria) across age classes (calves, subadults, adults) or sampling period (early winter, mid-winter, late winter)

Calves Subadults Adults Early winter Mid-winter Late winter

Taxon Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD

Firmicutes 0.7096 0.0819 0.7856 0.1401 0.7893 0.0979 0.8249 0.0944 0.7156 0.0921 0.7688 0.1222
Bacteroidetes 0.2578 0.0704 0.1932 0.1313 0.1726 0.0948 0.1338 0.0805 0.2464 0.0807 0.2119 0.1197
Actinobacteria 0.0156 0.0045 0.0111 0.0051 0.0271 0.0532 0.0305 0.0646 0.0185 0.0153 0.0120 0.0044
Proteobacteria 0.0074 0.0108 0.0026 0.0023 0.0023 0.0023 0.0024 0.0019 0.0072 0.0089 0.0014 0.0011
Fibrobacteres 0.0039 0.0079 0.0014 0.0034 0.0016 0.0022 0.0027 0.0061 0.0035 0.0036 0.0002 0.0003
Unclassified Bacteria 0.0031 0.0022 0.0038 0.0022 0.0037 0.0027 0.0027 0.0014 0.0049 0.0026 0.0035 0.0027

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

Tenericutes 0.0011 0.0008 0.0008 0.0004 0.0017 0.0014 0.0013 0.0014 0.0017 0.0012 0.0010 0.0007
Spirochaetes 0.0008 0.0008 0.0008 0.0012 0.0010 0.0013 0.0009 0.0011 0.0016 0.0015 0.0004 0.0007
Fusobacteria 0.0008 0.0018 0.0006 0.0009 0.0007 0.0013 0.0007 0.0014 0.0006 0.0006 0.0007 0.0016

Table 4. Mean relative abundance of the bacteria classified to eight 10 OTUs also represented the most abundant OTUs pres-
known phyla and bacteria unable to be classified to known phyla
ent in the manatee hindgut, according to average abun-
(Unclassified Bacteria) present in faecal samples of 36 live wild Florida
dance (Table 5). Six of the 10 most abundant OTUs had
manatees sampled at CR
99% sequence identity to OTUs present in a barcoded
% of total sequences* amplicon library data set obtained from wild dugong fae-
Phylum Mean  SD
cal samples (Eigeland et al., 2012).
Firmicutes 77.27  11.14
Bacteroidetes 19.54  10.58
Actinobacteria 2.00  3.80 Variation in microbial diversity of wild
Unclassified Bacteria 0.36  0.24 manatees
Proteobacteria 0.34  0.53
Fibrobacteres 0.20  0.41 PCoA was performed to determine whether the variables
Tenericutes 0.13  0.11 of body size/age, cortisol levels, sex and/or winter sam-
Spirochaetes 0.09  0.12 pling period could account for the variation displayed
Fusobacteria 0.07  0.13 within the microbial community. No clustering was evi-
Bacterial phyla are listed in order of descending abundance. dent in the PCoA plots (Fig. 4), with the variance
*The percentage of total sequences was calculated on the mean total explained by PC1 and PC2 being 37.77% and 16.65%,
number of sequences from the partial 16S rRNA gene amplicon
respectively, with no single variable accounting for the
libraries constructed from faeces samples obtained from 36 wild man-
differences in microbial diversity, and PERMANOVA supports
atees, taxonomically classified using the RDP classifier tool.
this for the category age class (F = 0.927, P = 0.49), but
demonstrates that there are significant differences between
taxonomic levels, 18 classes, 26 orders, 47 families, 91 seasons (F = 4.08, P = 0.0003).
genera and 2022 OTUs were identified. Bacteroidales and
Clostridiales were the predominant orders (Fig. 2a) in all
Faecal cortisol levels in wild overwintering
manatees except one adult male sampled in early winter
manatees
(AMEW1), with Clostridiales and Coriobacteriales as dom-
inant. Clostridium was the predominant classifiable genus There were significant differences in faecal cortisol meta-
throughout all sampled manatees (Fig. 2b). bolite (ng g 1) levels between manatees sampled at differ-
Twenty-three core OTUs, that is OTUs present in ent winter periods. Manatees sampled in early winter had
≥ 95% of sampled manatees, were identified, and a heat- significantly lower mean cortisol levels (mean  SD:
map displaying their abundance is shown in Fig. 3. The 15.5  7.1 ng g 1) than those sampled during mid-win-
closest sequence similarity matches using the BLAST algo- ter (mean  SD: 31.1  16.7 ng g 1) and late winter
rithm against the NCBI GenBank and Greengene databas- periods (24.2  14.7 ng g 1): 2-way ANOVA: winter period
es and wild dugong amplicon sequence data set (Eigeland x size class interaction F4,27 = 1.24, P = 0.32; winter per-
et al., 2012) are contained in the Supporting Information, iod F2,27 = 4.08, P = 0.03; size class F2,27 = 0.28,
Table S1. Core OTUs that could be classified to the level P = 0.76, but there were no differences between mid-win-
of genus included Adlercreutzia, Bacteriodes, Butyrivibrio, ter and late winter levels (Fig. 5).
Caloramator, Cellulosilyticum, Clostridium, Eubacterium, There were no differences in mean cortisol levels of
Ruminococcus and Sedimentibacter. Of these core OTUs, subadults and adults.

FEMS Microbiol Ecol 87 (2014) 601–615 ª 2013 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
608 S.D. Merson et al.

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

Caloramotor sp.(32)

Bacteroides sp.(4005)
Eubacterium sp. (6540)
Unclassified Clostridiales (5299)
Clostridium sp. (4540)
Unclassified Sphingobacteriales (5507)

Unclassified Ruminococcaceae (4377)

Unclassified Flavobacteriales (3288)

Aldecreutzia sp. (91)

Bacteroides sp. (2043)
Unclassified Bacteroidales (4614)
Sedimentibacter sp. (4431)
Eubacterium sp. (2004)

Clostridium sp. (3515)

Clostridium sp. (2156)

Cellulosilyticum sp. (4773)

Unclassified Ruminococcaceae (3739)
Unclassified Clostridiales (4153)
Butyrivibrio sp. (6224)
Unclassified Bacteroidales (1747)

Clostridium sp. (3961)

Clostridium sp. (5326)
Clostridium sp. (29)
CFEW7
AMEW1
AFMW15
AFMW13
SAMMW22
CMMW19
SAFMW21
AFMW18
AMEW3
AFEW4
SAMEW10

SAFEW9
SAFEW11

AFEW6
CMMW20
AMMW14

AMMW16
AMMW17

AFEW2
AFEW5
CFEW8
SAUknEW12
SAFLW32
AMLW27
AFLW28
CMLW29
SAMLW35
CMLW30

AMLW24
AFLW26
SAMLW34
AMLW23
AFLW25
SAFLW36

SAMLW33

CMLW31

Fig. 3. Heatmap of the 23 core bacterial OTUs found in the faecal samples of 36 wild Florida manatees from CR, Florida, USA. Individual
manatee identifier codes with the following key: age class (A – adult, SA – subadult, C – calf) Sex (M – male, F – female) Sample period (EW –
early winter, MW – mid-winter, LW – late winter); Sample numbers run along the x-axis.

Fusobacteria were also present but at lower abundance.

Discussion OTU sequences that could not be classified to any known
The hindgut microbiota was diverse in overwintering Flor- phyla comprised 0.36% of the total. These findings con-
ida manatees in CR, central-west Florida, and included trast with those of Mao et al. (2011) who found that Bac-
representatives of several bacterial phyla. Firmicutes were teroidetes comprised a significantly larger proportion of
by far the most dominant of the hindgut bacteria the bacterial community (63.5%) than Firmicutes (32.3%)
(77.3%), with Bacteroidetes (19.5%), and Actinobacteria in the hindgut of manatees from the Blue Springs, located
(2%) the next most abundant phyla. Members of c. 160 km due east of CR within the range of the Atlantic
Proteobacteria, Fibrobacteria, Tenericutes, Spirochaetes and subpopulation. It is possible that these phylogenetic

ª 2013 Federation of European Microbiological Societies. FEMS Microbiol Ecol 87 (2014) 601–615
Published by John Wiley & Sons Ltd. All rights reserved
 Table 5. The 10 most abundant bacterial OTUs identified in faecal samples collected from 36 wild Florida manatees in CR, Florida, USA, listed in descending order of abundance

Average% of total
OTU ID sequences  SD NCBI/ BLAST Host Reference Greengenes/ BLAST Host Reference
29 13.99  5.7 Clone Sass_faecal-45 Kakapo (Strigops habroptilus) Waite et al. (2012) Clostridium glycolicum CM- Cattle (Bos taurus) – Bagge et al. (2010)
99% (418/419) – faeces C52 99% (417/419) faeces
3961 7.07  4.44 Clostridium butyricum Human (Homo sapiens) – V.B. Lanjekar, Clostridium butyricum str. Cattle (Bos taurus) – Bagge et al. (2010)
str NMPYG-4 100% faeces N.P. Marathe, CM-C97 100% (420/420) faeces
(420/420) Y.S. Shouche and
D.R. Ranade
(unpublished)

FEMS Microbiol Ecol 87 (2014) 601–615

5326 6.43  6.6 Clone dgA-143 99% Captive Dugong (Dugong Tsukinowa Clostridium chartatabidum Sheep (Ovis aries) – Rainey &
(419/420) dugong) – faeces et al. (2008) str. DSM5482 99% rumen contents Stackebrandt (1993)
(417/420)
Microbial communities in the manatee hindgut

1747 6.05  8.52 Clone African elephant faeces Ley et al. (2008) Barnesiella intestinihominis Human (Homo sapiens) – Sakamoto & Ohkuma
AFEL3_aao13a08 (Loxodonta africana) str. YIT 11860 88% faeces (2010)
94% (412/439) (390/440)
3739 4.89  3.38 Clone Domesticated horse (Equus Ley et al. (2008) Ethanoligenens harbinense Anaerobic sludge Ren et al. (2007)
horsem_aai93f07 equus) – faeces str. CGMCC 1152 89%
98% (415/423) (376/423)
4153 4.76  3.65 Clone WA_aaa03a08 Somali wild ass (Equus Ley et al. (2008) Garciella sp. str. GK3 86% Anaerobic digester – M. Kanno, T. Katayama,
99% (420/423) asinus) – faeces (367/423) butanol-tolerant bacteria H. Tamaki, Y. Mitani,
X.Y. Meng, T. Hori,
T. Narihiro, N. Morita,
T. Hoshino, I. Yumoto,
N. Kimura, S. Hanada
and Y. Kamagata
(unpublished)
6224 3.95  3.84 Clone CK66 97% Holstein calves (Bos taurus) – J.Q. Wang, P. Yu, Butyrivibrio hungatei str. Su6 Cattle (Bos taurus) – van de Vossenberg &
(423/436) rumen D.P. Bu, K.L. Liu, 93% (406/436) rumen Joblin (2003)
D. Li and S.G. Zhao
(unpublished)
5507 3.32  5.15 Clone KO2_aai20b11 Red Kangaroo (Macropus Ley et al. (2008) Hevizibacter sp. str. P2K-21 Red water-lily (Nymphaea A.K. Borsodi, J. Makk,
90% (400/443) rufus) – faeces 84% (368/438) rubra) – rhizoplane V. Vagany, A. Rusznyak,
P. Schumann,
K. Marialigeti and
E.M. Toth
(unpublished)
2156 2.88  1.6 Clone dgB-84 99% Captive Dugong (Dugong Tsukinowa Clostridium sp. str Kt11 98% Hot spring – soil Takahashi et al.
(419/420) dugon) – faeces et al. (2008) 411/420 (2010)
32 1.95  2.42 Clone B21.68 99% 55 day anaerobic digestion Xing et al. (2011) Caloramator proteoclasticus Mesophilic granular Tarlera et al. (1997)
(415/420) of Microcystis blooms str. Uruguayensis DAM methanogenic sludge
10124 95% (401/420)
The BLAST algorithm was used to compare the 10 OTUs to 16S rRNA gene sequences within a data set obtained from faeces of wild dugongs, the NCBI GenBank and Greengene databases. The
best match for each OTU (Highest% sequence similarity) is listed for each search, along with the source of the sample (host).

Published by John Wiley & Sons Ltd. All rights reserved

ª 2013 Federation of European Microbiological Societies.
609

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

610 S.D. Merson et al.

Sub-adult High

Adult Medium
PC2 Percent variation explained 16.65%

PC2 Percent variation explained 16.65%

Calf Low

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

PC1 Percent variation explained 37.77% PC1 Percent variation explained 37.77%
Age Cortisol

Late winter
Male
Early winter
PC2 Percent variation explained 16.65%

PC2 Percent variation explained 16.65%

Unknown
Mid winter
Female

PC1 Percent variation explained 37.77% PC1 Percent variation explained 37.77%
Season Sex

Fig. 4. Weighted PCoA of the faecal bacterial communities of 36 wild Florida manatees sampled in CR, Florida, USA. PCoA was performed
examining the following variables: sex, age class (as a function of body size), winter period and chronic stress (faecal cortisol) levels.

differences in bacterial composition reflect geographic et al., 2012) with Firmicutes (76% of flora) and Bacteroi-
and/or dietary differences between these west and east detes (20%) being dominant, and Actinobacteria and Pro-
coast subpopulations. teobacteria being the next most prevalent phyla. Further,
Bacterial phyla found in the manatees of this study from six bacterial phyla identified in a single captive
were similar to those that dominate the hindgut flora of dugong held in Japan (Tsukinowa et al., 2008), a similar
the related wild dugong in eastern Australia (Eigeland distribution was observed: Firmicutes 83.1%, Bacteroidetes

ª 2013 Federation of European Microbiological Societies. FEMS Microbiol Ecol 87 (2014) 601–615
Published by John Wiley & Sons Ltd. All rights reserved
Microbial communities in the manatee hindgut 611

Clostridiales (OTU 4153), were most similar to each of

four bacteria found within gut contents of the foregut-
fermenting red kangaroo (Macropus rufus), and the hind-
gut-fermenting African elephant (Loxodonta africana),
domestic horse (Equus ferus caballus) and Somali wild ass
(Equus africanus somalicus), respectively (Ley et al., 2008).
Interestingly, Ley et al. (2008) found distinct microbial
clustering between foregut- and hindgut-fermenting her-
bivores. This corresponds with six of the manatees’ domi-
nant OTUs being most closely related to those in other

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

hindgut-fermenting herbivores. Two OTUs (OTU 5326,
2156) were congruous with uncultured Clostridium
species clones found within the faecal microbiota of a
Fig. 5. Faecal cortisol levels in wild Florida manatees of three size
classes (adult, subadult, calf) over winter. x-Axis displays three winter
captive dugong (Tsukinowa et al., 2008). Furthermore,
periods (early, mid, late). y-Axis displays faecal cortisol levels in units comparison of the manatee’s 23 core OTUs with a bar-
of nanogram per gram (ng g 1). coded 16S rRNA gene amplicon database from wild
dugong faecal samples (Eigeland et al., 2012) identified
11 OTUs, including six of the dominant OTUs, with high
15%, with the remaining phyla (Verrucomicrobia, Actino- sequence homology to dugong OTUs (using the BLAST
bacteria, Lentisphaerae and Proteobacteria) at < 1%. algorithm). Given the sympatric evolution of sea cow
Although there has been considerable taxonomic diversity communities (Velez-Juarbe et al., 2012), this similarity in
in microorganisms at the phylum level in all of the sire- hindgut bacterial flora might be expected.
nians surveyed to date, variation across the order appears The hindgut bacterial communities of wild manatees
to be conservative. Contributing factors to similarities sampled in CR were similar across all age/size classes,
presumably relate to phylogenetic, geographic and/or die- despite assumed dietary differences between immature
tary factors (after Velez-Juarbe et al., 2012), but these and mature animals. Manatee calves remain dependent
have not been distinguished. Furthermore, many of the on their mothers’ milk for 1–2 years (Reid et al., 1995)
bacterial OTUs that have been identified in both mana- and then wean slowly over a 1- to 2-year period (Koelsch,
tees and dugongs appear to represent novel and/or previ- 2001), and all calves sampled here were still nursing or at
ously unidentified bacteria and appear to constitute a least supplementing their plant diet with milk. It is possi-
unique microbial assemblage within this group of herbiv- ble that calves and adult Florida manatees eat similar
orous mammals. plant forage, as has been found with Antillean manatees
Ten of the 23 core bacterial OTUs from the hindgut of in Puerto Rico and Mexico (Mignucci-Giannoni & Beck,
wild CR manatees were from the order Clostridiales, and 1998; Castelblanco-Martinez et al., 2009) and that plant
five of these were amongst the ten dominant OTUs. OTU dietary composition is informative of microbial composi-
29 represented 14  5.7% of the total sequences tion. In contrast, distinct bacterial communities have been
obtained. Interestingly, its closest match using BLAST found between size classes of wild dugongs within a single
against the Greengenes database was to the bacterium population, and these differences have been interpreted as
Clostridium glycolicum, a known reductive acetogen in responses to ontogenetic dietary shifts from milk (calves),
eastern grey and red kangaroos (Ouwerkerk et al., 2009). to milk and seagrass (weaning juveniles) and then to sea-
Reductive acetogenesis combines hydrogen and carbon, grass alone (subadults, adults) (Eigeland et al., 2012).
forming acetate as a byproduct, which is utilised by the Ontogenetic dietary shifts have also been implicated as
host animal as a form of energy (Ouwerkerk et al., 2009). determining factors for variation in gastrointestinal
As this species is known to be a facultative reductive ace- microbial community structure in herbivorous marine
togen, it is possible that as a dominant species in the fishes (Rimmer, 1986; Clements & Choat, 1993; Moran
Florida manatee, it may reduce emissions of methane et al., 2005). The ontogenetic stability in hindgut micro-
from hindgut fermentation. The next most abundant bial communities in CR manatees may be anomalous
OTU 3961, an unclassified Clostridiaceae, represented compared to some other herbivorous vertebrates (e.g.
7  4.4% of the total sequences obtained and had 100% dugongs), but may reflect maintenance of a diverse and
identity to Clostridium butyricum, previously isolated generalist microbial biota capable of digesting a broad
from human and cattle faeces. Four of the next dominant diet, throughout all life stages.
OTUs, that is two unclassified Bacteroidales (OTU 5507, Interestingly, and perhaps contrary to expectations,
1747), Ruminococcaceae (OTU 3739) and an unclassified hindgut microbial diversity in Florida manatees remained

FEMS Microbiol Ecol 87 (2014) 601–615 ª 2013 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
612 S.D. Merson et al.

high throughout the winter period even as dietary shifts could be attributable to CSS in Citrus County, which
were likely to be occurring. When manatees migrate to includes CR (FWC, 2012). Manatee survival in the face of
warm water sources, including CR, dietary changes are surrounding freezing temperatures and presumed forage
usual, as available forage alters spatially and nutritionally shortage might suggest the high quality of the CR over-
(Gilbert & Clark, 1981; Deutsch et al., 2003). For some wintering habitat and/or the physiological resil-
manatees, these dietary shifts may be profound, for exam- ience (including digestive health) of this species to harsh
ple, from marine to freshwater vegetation (Etheridge conditions.
et al., 1985). Changes in forage availability continue to In summary, the high diversity and stability of the
occur as winter progresses, as vegetation becomes over- hindgut microbial system of the Florida manatee probably
grazed and/or seasonally deficient (R.K. Bonde, USGS, reflect its generalist herbivore strategy. Furthermore, the

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

unpublished data). It was therefore expected that individ- apparent resilience of the diversity of this microbial com-
uals sampled during early winter might display a more munity in the face of seasonal dietary change and under-
diverse or at least different bacterial community, concom- lying physiological stress is probably an effective
itant with higher prewinter forage availability (Etheridge adaptation for a migratory species. It is also possible that
et al., 1985), compared to those sampled later in winter CR is a particularly good overwintering site (i.e. reliable
when forage deficiencies were more likely to result in a warm waters, adjacent foraging grounds) and that its
lower nutrient plane or fasting (Deutsch et al., 2006). peculiar traits may have resulted in the relative stability
However, according to the alpha-diversity index, mana- of microbial communities in its residents compared to
tees sampled in mid-winter exhibited higher microbial manatees at overwintering sites where more pronounced
diversity than early and late winter manatees. Beta-diver- nutritional and/or thermal stress may be experienced.
sity measures, analysed through PERMANOVA, indicated sig- This possibility cannot be discounted without compara-
nificant differences in bacterial communities based on the tive studies amongst and within the four subpopulations
season in which the samples were taken; however, this of this species.
was not reflected in the PCoA plots. Interpreting these Florida manatees are of considerable evolutionary sig-
two measures simultaneously suggests that whilst mana- nificance because sirenians are the only herbivorous mar-
tees during mid-winter exhibited a bacterial community ine mammals. Our findings suggest that Florida
of higher diversity, these bacteria were not sufficiently manatees, similar to dugongs, have a novel microbial
distinct based upon the entire community. community in the hindgut. It is also interesting to con-
Temporal variation in microbial biota, directly related sider the evolution of the sirenians with respect to their
to available forage and forage quality, has been recorded gut microbiota. Of 23 core OTUs identified across all 36
in other mammalian herbivores including red deer, cattle manatees, 44% were also common to wild dugongs
and horses (Hobson et al., 1975; Dierenfeld, 1997; (Eigeland et al., 2012). This suggests that a considerable
Kocherginskaya et al., 2001; Kobayashi et al., 2006). The proportion of the core bacterial community in sirenians
maintenance of a highly diverse microbial biota in Florida predates the evolutionary divergence of manatees and
manatees throughout a period of potential nutritional dugongs. As far as we are aware, no research currently
and thermal stress was surprising. It is likely that environ- exists on the microbial communities of the manatees’
mental factors including changed food availability and/or extant relatives, the other species of manatees, nor to the
temperature may have caused the cortisol stress response more distantly related extant terrestrial mammals, the ele-
detected in manatees as winter progressed. Although nor- phant and rock hyrax. Future research into the microbial
mal faecal cortisol metabolite levels have not been estab- communities of related animals would provide informa-
lished for Florida manatees, circulating cortisol has been tion about their evolution and insight into the transition
shown to be the predominant glucocorticoid and diag- from terrestrial to marine herbivory.
nostic of stress in this species (Tripp et al., 2011) and in
the related dugong (Burgess et al., 2013). Such short-term
Acknowledgements
glucocorticoid release is considered beneficial in helping
mammals survive stressful conditions, whilst prolonged Manatee faeces were collected under U.S. Fish and Wild-
stress, reflected in chronically elevated glucocorticoid con- life Service federal research permit MA791721 issued to
centrations, can inhibit immune and inflammatory pro- Sirenia Project, United States Geological Survey, and
cesses (Sapolsky et al., 2000). imported into Australia under CITES permit 2010-AU-
The winter of 2010 was one of the coldest on record in 608083 and AQIS IP10018161. Thanks to Helen Sneath
Florida with a record 282 manatee mortalities directly (University of Queensland) for assistance in the field, and
attributable to cold stress syndrome (CSS) (FWC, 2012; Damien Finn and Emilio Martinez (Rumen Ecology Unit,
Stith et al., 2012). However, only two manatee mortalities DAFF) for assistance in the laboratory. This project was

ª 2013 Federation of European Microbiological Societies. FEMS Microbiol Ecol 87 (2014) 601–615
Published by John Wiley & Sons Ltd. All rights reserved
Microbial communities in the manatee hindgut 613

funded by the Winifred Violet Scott Foundation and Sea De Santis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL,
World Australia. A Queensland-Smithsonian fellowship Keller K, Huber T, Dalevi D, Hu P & Andersen GL (2006)
awarded to EAB funded the faecal cortisol assays. Any use Greengenes, a chimera-checked 16S rRNA gene database
of trade, product or firm names is for descriptive pur- and workbench compatible with ARB. Appl Environ
poses only and does not imply endorsement by the U.S. Microbiol 72: 5069–5072.
Government. Order of authorship follows the ‘first-last- Deutsch CJ, Bonde RK & Reid JP (1998) Radio-tracking
manatees from land and space: tag design, implementation,
then others-author-emphasis’ approach.
and lessons learned from long-term study. Mar Technol Soc
J 32: 18–29.
References Deutsch CJ, Reid JP, Bonde RK, Easton DE, Kochman HI &

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

O’Shea TJ (2003) Seasonal movements, migratory behavior,
Alves-Stanley CD, Worthy GA & Bonde RK (2010) Feeding and site fidelity of West Indian manatees along the Atlantic
preferences of West Indian manatees in Florida, Belize, and coast of the United States. Wildl Monogr 151: 1–77.
Puerto Rico as indicated by stable isotope analysis. Mar Ecol Deutsch CJ, Edwards HH & Barlas ME (2006) Florida manatee
Prog Ser 402: 255–267. foraging behavior around a winter warm-water refuge. Final
Bagge E, Persson M & Johansson KE (2010) Diversity of Report to the Wildlife Foundation of Florida. Grant No.
spore-forming bacteria in cattle manure, slaughterhouse CWT0405-12. FWC/FWRI file FCWT0405-12-F. pp. 42.
waste and samples from biogas plants. J Appl Microbiol 109: Dierenfeld ES (1997) Captive wild animal nutrition: a
1549–1565. historical perspective. Proc Nutr Soc 5: 989–999.
Best RC (1981) Foods and feeding habits of wild and captive Domning DP (1983) Marching teeth of the manatee. Nat Hist
sirenia. Mamm Rev 11: 3–29. 92: 8–11.
Bragg L, Stone G, Imelfort M, Hugenholtz P & Tyson G Eigeland KA, Lanyon JM, Trott DJ, Ouwerkerk D, Blanshard
(2012) Fast, accurate error-correction of amplicon W, Milinovich GJ, Guilino L-M, Martinez E, Merson S &
pyrosequencings using Acacia. Nat Methods 9: 425–426. Klieve AV (2012) Bacterial community structure in the
Burgess EA, Brown JL & Lanyon JM (2013) Sex, scarring, and hindgut of wild and captive dugongs, Dugong dugon. Aquat
stress: understanding seasonal costs in a cryptic marine Mamm 38: 402–411.
mammal. Conserv Physiol 1: Article ID cot014. DOI: 10. Etheridge K, Rathbun GB, Powell JA & Kochman HI (1985)
1093/conphys/cot014. Consumption of aquatic plants by the West Indian manatee.
Burn DM (1986) The digestive strategy and efficiency of the J Aquat Plant Manag 23: 21–25.
West Indian manatee, Trichechus manatus. Comp Biochem Florida Fish and Wildlife Conservation Commission FWC
Physiol 85: 139–142. (2012) Annual Manatee Death Count Shows Cold Weather an
Burn DM & Odell DK (1987) Volatile fatty acid concentrations Unusually Big Factor Again. Available at: http://myfwc.com/
in the digestive tract of the West Indian manatee, Trichechus news/news-releases/2012/January/04/manatee-deaths-2011/.
manatus. Comp Biochem Physiol 88: 47–49. Gilbert S & Clark K (1981) Seasonal variation in standing crop
Campbell HW & Irvine AB (1977) Feeding ecology of the of the seagrass Syringodium filiforme and associated
West Indian manatee Trichechus manatus Linnaeus. macrophytes in the northern Indian River, Florida. Estuaries
Aquaculture 12: 249–251. 4: 223–225.
Caporaso GJ, Kuczynski J, Stombaugh J et al. (2010a) QIIME Hamady M, Walker JJ, Harris JK, Gold NJ & Knight R (2008)
allows analysis of high-throughput community sequencing Error-correcting barcoded primers for pyrosequencing
data. Nat Methods 7: 335–336. hundreds of samples in multiplex. Nat Methods 5: 235–237.
Caporaso GJ, Bittinger K, Bushman FD, DeSantis TZ, Hartman DS (1979) Ecology and behaviour of the manatee
Andersen GL & Knight R (2010b) PYNAST: a flexible tool for (Trichechus manatus) in Florida. Amer Soc Mammal 5: 153.
aligning sequences to a template alignment. Bioinformatics Hobson PN, Mann SO, Summers R & Staines BW (1975)
26: 266–267. Rumen function in red deer, hill sheep and reindeer in the
Castelblanco-Martinez DN, Morales-Vela B, Hernandez-Arana Scottish Highlands. Proc R Soc Edinb Nat Environ 75:
HA & Padilla-Saldivar J (2009) Diet of the manatees 181–198.
(Trichechus manatus manatus) in Chetumal Bay, Mexico. Kobayashi Y, Koike S, Miyaji M, Hata H & Tanaka K (2006)
Lat Am J Aquat Mamm 7: 39–46. Hindgut microbes, fermentation and their seasonal
Clements KD & Choat JH (1993) Influence of season, variations in Hokkaido native horses compared to light
ontogeny and tide on the diet of the temperate marine horses. Ecol Res 21: 285–291.
herbivorous fish Odax pullus (Odacidae). Mar Biol 117: Kocherginskaya SA, Aminov RI & White BA (2001) Analysis
213–220. of the rumen bacterial diversity under two different diet
Cole JR, Wang Q, Cardenas E et al. (2009) The Ribosomal conditions using denaturing gradient gel electrophoresis,
Database project: improved alignments and new tools for random sequencing, and statistical ecology approaches.
rRNA analysis. Nucleic Acids Res 37: 141–145. Anaerobe 7: 119–134.

FEMS Microbiol Ecol 87 (2014) 601–615 ª 2013 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved
614 S.D. Merson et al.

Koelsch J (2001) Reproduction in female manatees observed in Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,
Sarasota Bay, Florida. Mar Mamm Sci 17: 331–342. URL http://www.R-project.org.
Laist DW & Reynolds JE III (2005) Florida manatees, Rainey FA & Stackebrandt E (1993) 16S rDNA analysis reveals
warm-water refuges, and an uncertain future. Coast Manage phylogenetic diversity among the polysaccharolytic
33: 279–295. Clostridia. FEMS Microbiol Lett 113: 125–128.
Lane D, Pace B, Olsen GJ, Stahl DA, Sogin ML & Pace NR Reep RL & Bonde RK (2006) The Florida Manatee Biology and
(1985) Rapid determination of 16S ribosomal RNA Conservation. pp. 189. University Press of Florida,
sequences for phylogenetic analysis. P Nat Acad Sci USA 82: Gainesville, FL.
6955–6959. Reid JP, Bonde RK & O’Shea TJ (1995) Reproduction and
Langtimm CA, O’Shea TJ, Pradel R & Beck CA (1998) Estimates mortality of radio-tagged and recognizable manatees on the

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

of annual survival probabilities for adult Florida manatees Atlantic coast of Florida. Population Biology of the Florida
(Trichechus manatus latirostris). Ecology 79: 981–987. Manatee. Information and Technology Report 1. (O’Shea TJ,
Lanyon JM, Sneath HL, Ovenden JR, Broderick D & Bonde Ackerman BB & Percival HF, eds), pp. 171–191. National
RK (2009) Sexing sirenians: validation of visual and Biological Service, Washington, D.C.
molecular sex determination in both wild dugongs (Dugong Ren N, Xing D, Rittmann BE, Zhao L, Xie T & Zhao X (2007)
dugon) and Florida manatees (Trichechus manatus Microbial community structure of ethanol type
latirostris). Aquat Mamm 35: 187–192. fermentation in bio-hydrogen production. Environ Microbiol
Larkin LV, Fowler VF & Reep RL (2007) Digesta passage rates 9: 1112–1125.
in the Florida manatee (Trichechus manatus latirostris). Zoo Reynolds JE III (1980) Aspects of the structural and functional
Biol 26: 503–515. anatomy of the gastrointestinal tract of the West Indian
Ley RE, Hamady M, Lozupone C et al. (2008) Evolution of manatee, Trichechus manatus. PhD Thesis, University of
mammals and their gut microbes. Science 320: 1647–1651. Miami, Miami, FA, USA.
Lomolino MV (1977) The ecological role of the Florida Reynolds JE III & Rommel SA (1996) Structure and
manatee Trichechus manatus latirostris in water hyacinth function of the gastrointestinal tract of the Florida
dominated ecosystems. MS Thesis, University of Florida, manatee, Trichechus manatus latirostris. Anat Rec 245:
Gainesville, FA. 539–558.
Lomolino MV & Ewel KC (1984) Digestive efficiencies of the Rimmer DW (1986) Changes in diet and the development of
West Indian manatee (Trichechus manatus). Florida Sci 47: microbial digestion in juvenile buffalo bream Kyphosus
176–179. cornelii. Mar Biol 92: 443–448.
Mao Y, Yannarell AC, Hong P & Mackie RI (2011) Different Sakamoto M & Ohkuma M (2010) Usefulness of the hsp60 gene
groups of novel Bacteroidetes dominate the gut microbial for the identification and classification of Gram-negative
community of wild manatees (Trichechus manatus) from anaerobic rods. J Med Microbiol 59: 1293–1302.
Belize and Florida. Congress on Gastrointestinal Function, Sapolsky RM, Romero LM & Munck AU (2000) How do
University of Chicago, Chicago, IL. 18–20 April 2011. glucocorticoids influence stress-responses? Integrating
Mignucci-Giannoni AA & Beck CA (1998) The diet of the permissive, suppressive, stimulatory, and adaptive actions.
manatee (Trichechus manatus) in Puerto Rico. Mar Mamm Endocr Rev 21: 55–89.
Sci 14: 394–397. Schloss PD, Westcott SL, Ryabin T et al. (2009) Introducing
Moran D, Turner SJ & Clements KD (2005) Ontogenetic mothur: open source, platform-independent,
development of the gastrointestinal microbiota in the community-supported software for describing and
marine herbivorous fish Kyphosus sydneyanus. Microb Ecol comparing microbial communities. Appl Environ Microbiol
49: 590–597. 75: 7537–7541.
Nabor P & Patton GW (1989) Aerial studies of the West Stith BM, Slone DH, de Wit M, Edwards HH, Langtimm CA
Indian manatee (Trichechus manatus) from Anna Maria & Swain E (2012) Passive thermal refugia provided warm
Florida to Northern Charlotte Harbor including the Myakka water for Florida manatees during the severe winter of
River: recommended habitat protection and manatee 2009–2010. Mar Ecol Prog Ser 462: 287–301.
management strategies. Mote Marine Laboratory Technical Takahashi Y, Suto K & Inoue C (2010) Polysulfide reduction
Report no 134: pp. 94. County of Sarasota, Florida. by Clostridium relatives isolated from sulfate-reducing
Ouwerkerk D, Maguire AJ, McMillen L & Klieve AV (2009) enrichment cultures. J Biosci Bioeng 109: 372–380.
Hydrogen bacteria from the forestomach of eastern grey Tarlera S, Muxi L, Soubes M & Stams AJ (1997) Caloramator
(Macropus giganteus) and red (Macropus rufus) kangaroos. proteoclasticus sp. nov., a new moderately thermophilic
Anim Prod Sci 49: 1043–1051. anaerobic proteolytic bacterium. Int J Syst Bacteriol 47:
Price MN, Dehal PS & Arkin AP (2010) FASTTREE 2 – 651–656.
approximately maximum-likelihood trees for large Tripp KM, Verstegen JP, Deutsch CJ, Bonde RK, Wit MD,
alignments. PLoS One 5: e9490. Manire CA, Gaspard J & Harr KE (2011) Evaluation of
R Development Core Team (2008) R: A Language and adrenocortical function in Florida manatees (Trichechus
Environment for Statistical Computing. R Foundation for manatus latirostris). Zoo Biol 30: 17–31.

ª 2013 Federation of European Microbiological Societies. FEMS Microbiol Ecol 87 (2014) 601–615
Published by John Wiley & Sons Ltd. All rights reserved
Microbial communities in the manatee hindgut 615

Tsukinowa E, Karita S, Asano S, Wakai Y, Oka Y, Furuta M & Xing P, Guo L, Tian W & Wu QL (2011) Novel Clostridium
Goto M (2008) Fecal microbiota of a dugong (Dugong populations involved in the anaerobic degradation of
dugon) in captivity at Toba Aquarium. J Gen Appl Microbiol Microcystis blooms. ISME J 5: 792–800.
54: 25–38. Yu Z & Forster RJ (2005) Nucleic acid extraction,
van de Vossenberg JL & Joblin KN (2003) Biohydrogenation oligonucleotide probes and PCR methods. Methods in Gut
of C18 unsaturated fatty acids to stearic acid by a strain of Microbial Ecology for Ruminants (Makkar HPS, &
Butyrivibrio hungatei from the bovine rumen. Lett Appl McSweeney CS, eds), pp 81–104. Springer, the
Microbiol 37: 424–428. Netherlands.
Velez-Juarbe J, Domning DP & Pyenson ND (2012) Iterative
evolution of sympatric seacow (Dugongidae, Sirenia)
Supporting Information

Downloaded from https://academic.oup.com/femsec/article/87/3/601/576328 by MarÃ-a Otero user on 21 October 2024

assemblages during the past 26 million years. PLoS One 7:
e31294.
Waite DW, Deines P & Taylor MW (2012) Gut microbiome of Additional Supporting Information may be found in the
the critically endangered New Zealand parrot, the Kakapo online version of this article:
(Strigops habroptilus). PLoS One 7: e35803.
Table S1. Core OTUs identified in faecal samples col-
Wasser SK, Monfort SL, Southers J & Wildt DE (1994)
lected from 36 wild Florida manatees in Crystal River,
Excretion rates and metabolites of oestradiol and
progesterone in baboon (Papio cynocephalus) faeces.
Florida, USA.
Reproduction 101: 213–220.

FEMS Microbiol Ecol 87 (2014) 601–615 ª 2013 Federation of European Microbiological Societies.
Published by John Wiley & Sons Ltd. All rights reserved