Basic HTML Version
International Journal of Marine Science 2014, Vol.4, No.50, 1-22
http://ijms.biopublisher.ca
15
priori
as samples in which threshold cycle (C
t
) values
<30 were calculated.
3.8 VTE: host coral genotyping
It was hypothesized that the dramatic oceanographic
differences between Houbihu (UWS) and Houwan
(NUWS; Mayfield et al., 2012a), which are only ~15
km apart (Figure 1), could have driven a degree of
genetic divergence between coral populations of the
two sites of origin. To test this, one of the two
pseudo-replicated DNAs from each of the 12 tanks of
the VTE (
n
=6 for each site) was genotyped. Seven
species-specific microsatellite loci (Sh4-001,
Sh2-002, Sh3-004, Sh2-006, Sh2.15, Sh4.24, Sh4.28;
Tables 2 and A4) for
S. hystrix
developed in prior
works (Maier et al., 2001, 2005; 2009; Underwood et
al., 2006) were utilized to determine the extent of the
genetic divergence, if any, between these two
populations. Each microsatellite fragment was
amplified by PCR (15 µl) in all 12 DNA samples, and
reaction mixes contained 10-40 ng DNA, 1x PCR
buffer (MDbio), 0.2 mM dNTPs, 1.5 mM MgCl
2
, 200
nM each primer (Table A4), and 0.1 U
Taq
polymerase
(MDbio). Thermocycling was conducted at 95
℃
for
3 min followed by 35 cycles of 94
℃
for 40 s, 40 s at
the annealing temperatures listed in Table A4, and 72
℃
for 40 s. A final extension at 72
℃
for 1.5 min was
then conducted. An Applied Biosystems Genetic
Analyzer (3130, Life Technologies) was used to
separate fragments, and unique alleles were
determined with Peak Scanner™ (ver. 1.0, Life
Technologies).
Arlequin (ver. 3.1, Schneider et al., 2000) was used to
estimate the N
A
, the number of private alleles (N
PA
),
and the H
o
and H
e
, as well as conduct Hardy-Weinberg
equilibrium tests. F
IS
values were estimated using
Fstat ver. 2.9.3.2 (Goudet, 2001). Micro-Checker ver.
2.2.3 (van Oosterhout et al., 2004) was used to detect
possible null alleles. Genetic differentiation between
populations was estimated by calculating F
ST
values
with Arlequin 3.1. The significance of F
ST
values for
population comparisons was based on 1,000
permutations.
3.9 Statistical analyses
All statistical analyses were conduced with JMP™
(ver. 5, SAS Institute). Shapiro-Wilk W and Levene’s
tests were used to determine whether datasets were
normally distributed and characterized by homogeneous
variance, respectively. In certain cases, log, square
root, or rank transformations were required to generate
data suitable for parametric analyses. To assess
differences in both
Symbiodinium
and host gene
expression between treatments and over time in the
ETE, a repeated-measures ANOVA was conducted
with tank nested within treatment after confirming the
sphericity of the data with Mauchley’s sphericity tests
(
p
>0.05). To assess the effects of site of origin,
temperature treatment, and their interaction on both
host and
Symbiodinium
gene expression in samples of
the VTE, a 2-way ANOVA was conducted with tank
nested within site x temperature. In both studies, the
tank term was dropped when found to be
non-significant (
p
>0.05), and Tukey’s honestly
significant difference (HSD) tests were used to verify
individual mean differences when significant
differences were detected in the model (
p
<0.05).
Correlation analyses were conducted in order to
determine the degree of correlation between
expression levels of
Symbiodinium
and host coral
hsp70
in samples of the VTE. Slopes were considered
significant if
p
<0.01.
In order to further explore the effect of site of origin
on gene expression, the average expression of each
gene across the 24 samples from the ETE (all of which
were from the UWS) and the 12 samples of the VTE
originating from the UWS was compared against the
average expression of the 12 samples from NUWS, all
of which were from the VTE. This comparison is
referred to as the “global site of origin” comparison to
distinguish it from the site of origin comparisons
within the VTE described above. For the majority of
the global site of origin comparisons, Wilcoxon
rank-sum tests were utilized, as the variances tended
to differ significantly (Levene’s test,
p
<0.01) due to
the use of unequal sample sizes (
n
=36 and 12 for the
UWS and NUWS, respectively).
To compare the proportion of the 14 genes in which a
significant difference was detected between
experiments (ETE vs. VTE) and site of origin (UWS
vs. NUWS),
X
2
proportion tests were conducted
(
Quinn and Keough, 2002)
. Such tests were performed
in order to know whether variable temperature