|
RNA
Expression Traits: Application of Microarrays
to
Questions in Evolutionary Ecology
Background
-
Microarrays
in Evolutionary Ecology
-
Recent advances in microarray technology for genomic
assessment of mRNA variation has rendered this approach feasible for
studies in population biology. Here we describe a novel
application of microarray technology that will identify mRNA expression
traits that are diagnostic for taxa and genotypes in the Piriqueta caroliniana
complex. This analysis is feasible with the Piriqueta system because
-
1. Replication of genotypes via vegetative
propagation within and among environments provides robust experimental
designs that allow the delineation of environment-specific mRNA
expression patterns and the identification of diagnostic,
taxon-specific transcript expression traits.
-
2.
The larger sample sizes required for these experiments are feasible
because of efficient and cost-effective procedures that have been
developed for microarray hybridization, which will allow us to conduct
mRNA expression assays at a fraction of the cost of traditional
protocols.
-
We will
conduct a series of genotype-replicated experiments that will evaluate
environment-dependent expression responses for large numbers of mRNA
transcripts and allow us to identify a subset of “expression
traits." Note that this approach does not require the development
of oligonucleotide microarrays- we will be treating mRNA expression
patterns in response to native environments as genetic traits. This
approach is reasonable as it is becoming increasing obvious from
genomic analyses with model systems that a large portion of the
phenotypic differences among taxa is due to regulation of mRNA
expression. Furthermore, the total expressed complement of mRNA
in organisms (the ‘transcriptome’) constitutes a critical link between
the genotype and the phenotype.
-
The Piriqueta caroliniana complex: a
Non-Traditional Model System
-
Attributes of the system:
-
Herbaceous perennial
native to southeastern North America, the Caribbean, and South America.
-
Short generation
time- 8 to 12 weeks from seed to flowering
-
Flowers year-round
in the greenhouse, large flowers (2 - 3 cm in diameter) last a single
day.
-
Distylous with self-
and intramorph incompatibility.
-
Easily propagated
froms seeds or vegetative cuttings.
-
Occurs in a broad
range of
habitats that represent the terrestrial environment extremes for
moisture regimes- from extreme drought to standing water.
-
Hybridization in central
Florida.
-
Results from
previous studies. In this section we describe results from
studies on the strength of interspecific mating barriers, historical
patterns of migration and dispersal, patterns of introgression, and
patterns of hybrid breakdown between the caroliniana (C) and viridis (V) morphotypes. The
C morphotype occurs in dry, sandhill/turkey oak scrub habitats in north
Florida, and the V morphotype occurs in mesic and periodacally
flooded. The "oldest" hybrid populations in the center of the
broad hybrid zone are desgnated as 'H' parental genotypes.
-
Interspecific mating barriers.
Only weak prezygotic
reproductive isolating barriers separating caroliniana and
viridis. The floral morphologies of these plants appear to be
nearly identical, so it is unlikely that pollinators distinguish
between them. Given the differences in habitat preference,
contact between these morphotypes must have occurred as a result of
long-distance pollen dispersal. To mimic these conditions, we
examined the frequency of intra- and interspecific fertilization after
pollination with mixtures of pollen that contained different
proportions of intra- and interspecific pollen (Wang and Cruzan 1998).
In general, the proportion of parental and hybrid genotypes in the
seeds produced reflected the proportion of each pollen type in the
pollen load, indicating a weak mating barrier between these morphotypes
(Gavrilets and Cruzan 1998). These
analyses indicate that the prezygotic mating barriers between these
morphotypes are extremely weak.
-
Phylogeography of Piriqueta in North
America. We
examined 10 individuals from each of 27 allopatric populations and 29
from the
hybrid zone for
sequence variation in hypervariable regions of
the chloroplast genome (Taberlet et al. 1991; Demesure et al. 1995)
using RFLP and heteroduplex (Wood and Bidwell 1996)(FMC BioProducts)
analyses (Maskas and Cruzan 2000). The
distributions of
the 17
haplotypes detected were analyzed to infer the effects of historical
processes on contemporary distributions using the methods of Templeton
et al. (1995; Templeton 1998). The results of these analyses indicate
that North American taxa of Piriqueta have been derived from three
independent emigrations from the northern Bahamas, the only islands on
which plants in the P. caroliniana complex occur north of Dominica
(Correll and Correll 1982; Arbo 1995). Populations of viridis were most
likely derived from an immigration event that occurred within the last
7000 years. Prior to this time, climates were relatively cool and
dry in south Florida (Webb 1990). The
combined evidence of analyses of cpDNA variation, clines for diagnostic
markers, and morphological characters indicates that this broad hybrid
zone was initiated in the last 5000 years in southern Florida and has
expanded northwards.
-
Genetic and morphological variation
across the hybrid zone.
Morphological analyses across hybrid populations of Piriqueta indicated
wide ranges in phenotypic variation that span and exceed parental
phenotypes (Martin and Cruzan 1999). Hybrid populations occurring
in xeric sites typically had high trichome densities and relatively
broad leaves (more similar to C), while populations in more mesic sites
had narrower leaves and were often completely glaborous (similar to
V). This pattern of
phenotypic variation and ecological associations among hybrid
populations suggests that selection has produced populations adapted to
local environmental conditions.
-
Patterns of Hybrid Breakdown in the
F1 Generation. We used
controlled crosses and common garden experiments to estimate levels of
heterosis and epistasis among morphotypes (caroliniana = C; viridis =
V) and their hybrid derivative populations (H) of Piriqueta. First
generation hybrids were compared with inbred and outbred parental lines
to delineate the confounding effects of heterosis (i.e., due to
outcrossing inbred populations) and dominant epistasis
(Dobzhansky-Muller incompatibilities) for hybrid fitness. Hybrid
derivative genotypes were from the region in the center of the hybrid
zone, where hybrid plants were morphologically uniform within and among
populations and there was no evidence of recent introgression (i.e.,
disequilibria among diagnostic markers was near zero: Barton and Gale
1993). Crosses were made within and among C, H, and V genotypes from at
least five maternal families from each of five to seven populations to
produce inbred (within population crosses), outbred (between
populations of the same parental type), and F1 offspring.
First generation hybrids were compared to their inbred and outbred
parental genotypes for measures of seed germination, vegetative
fitness, and sexual reproduction over two field seasons in transplant
gardens at Archbold Research Station in south-central Florida.
Transplant gardens were located in xeric sandhill scrub communities in
sites with sparse vegetation that were representative of native sites
for the C and H parental types in central and northern Florida.
Germination rates and vegetative fitness of hybrids were intermediate
between outbred (higher fitness) and inbred (lower fitness) parental
lines. More F1 than inbred or outbred individuals
reproduced, and F1 hybrids produced more flowers and fruits than either
parental group, indicating that levels of heterosis in the F1
generation exceeded levels seen for the outbred parental genotypes. Vegetative fitness in some first
generation hybrid genotypes was reduced by dominant epistasis, and in
all cases the expression of hybrid breakdown was dependent on the
cytoplasmic genotype.
-
Hybrid breakdown in the Backcross
Generation. First generation
hybrids of both cytotypes were reciprocally backcrossed to their two
parental genotypes (C, H, or V) and transplanted to field common
gardens in sandhill sites at Archbold Biological Station as described
above. Outbred parental crosses and F1 crosses were replicated in this
experiment for robust comparisons of growth and reproduction for
backcross genotypes under field and greenhouse conditions. Hybrid
breakdown in the backcross hybrids was more severe than in the F1
generation for the CxH and CxV crosses, indicating fitness effects of
recombination for many of these genotypes. The degree of
fitness loss in these second generation hybrids exceeded the
expectation for loss of the heterotic effects of outbreeding due to
segregation. While the majority of backcross hybrids had very low
fitness, the variance was very large such that some genotypes had
fitness values similar to or exceeding the parental genotypes.
The observation of increased hybrid breakdown in the backcross
generation compared to F1 hybrids is consistent with the hypothesis
that these genotypes are suffering from the breakup of coadapted gene
complexes (i.e., recessive/positive epistasis). For the HxV crosses, on
the other hand, the F1 genotypes were lower fitness than the
backcrosses, suggesting that Dobzhansky-Muller incompatibilities
(dominant epistasis) may be responsible for a larger proportion of
fitness loss. The contrasting
patterns of hybrid breakdown in these crosses suggests that the
relative roles of dominant intergenomic epistasis and recessive
intragenomic epistasis for hybrid breakdown differ depending on the
particular cross combination.
-
Patterns of Environmentally-Dependent
Hybrid Breakdown.
Measurements of growth were made for the backcross genotypes described
above under greenhouse conditions for the first three months following
germination. Even though initial resource levels were greater for
plants in the field compared to the greenhouse experiments (i.e., stem
cuttings vs. seeds), average levels of growth were higher in the
greenhouse (P < 0.0001), indicating that the field sites represented
more severe environmental conditions. Overall performance under
greenhouse conditions was fairly predictive of relative growth rate in
the field (R-square = 0.43, P < 0.001 for regression of standard
deviates
of vegetative fitness in each environment). However, the
difference in relative performance of plants was skewed such that a
substantially larger number of recombinant genotypes exhibited greater
than expected fitness reductions when growing under xeric field
conditions compared to the greenhouse (i.e., there are more points well
below the predicted line). These
results indicate that a portion
of hybrid breakdown under field conditions is due to maladaptive trait
combinations that only have significant fitness effects under
particular environmental conditions.
Microarray Development
-
-
Collection
of leaf samples for RNA expression library.
-
Collection of leaf samples from parental and hybrid
genotypes under a variety of environmental conditions. The goal
is to create a collection of RNA samples that represents a broad range
of environments and genotypes.
Environment
|
caoliniana morphotype
|
viridis
morphotype
|
Advanced-
Generation
Hybrids
|
F1
Hybrids
|
Backcross
Hybrids
|
Greenhouse/day
|
11
|
25
|
6
|
10
|
10 |
Greenhouse/dusk (at lights
off)
|
3
|
3
|
2
|
10
|
|
Greenhouse/night (at 3:00
AM)
|
3
|
3
|
2
|
10
|
|
Greenhouse/high calcium
|
2
|
2
|
2
|
|
|
Greenhouse/night/high
calcium
|
2
|
2
|
2
|
|
|
Lab/cold (4 degrees C)
|
3
|
3
|
2
|
|
|
Lab/wilt
|
3
|
3
|
2
|
|
|
| Lab/Heat shock (55 degrees
C) |
3
|
3
|
2
|
|
|
Field/drought (2003)
|
6
|
4
|
2
|
|
|
Field/wet (2003)
|
10
|
3
|
6
|
|
|
Field/day/dry (2004)
|
12
|
8
|
26
|
13
|
|
Field/night (2004)
|
14
|
8
|
28
|
13
|
|
Total genotypes
|
46
|
45
|
44
|
26
|
10
|
-
Total number of genotypes to date = 171
-
Total dumber of environments to date = 12
-
Total RNA samples to date = 289
|
