Recombinational micro-evolution of functionally different metallothionein promoter alleles from Orchesella cincta

Abstract

Background: Metallothionein (mt) transcription is elevated in heavy metal tolerant field populations of Orchesella cincta (Collembola). This suggests that natural selection acts on transcriptional regulation of mt in springtails at sites where cadmium (Cd) levels in soil reach toxic values This study investigates the nature and the evolutionary origin of polymorphisms in the metallothionein promoter (pmt) and their functional significance for mt expression.

Results: We sequenced approximately 1600 bp upstream the mt coding region by genome walking. Nine pmt alleles were discovered in NW-European populations. They differ in the number of some indels, consensus transcription factor binding sites and core promoter elements. Extensive recombination events between some of the alleles can be inferred from the alignment. A deviation from neutral expectations was detected in a cadmium tolerant population, pointing towards balancing selection on some promoter stretches. Luciferase constructs were made from the most abundant alleles, and responses to Cd, paraquat (oxidative stress inducer) and moulting hormone were studied in cell lines. By using paraquat we were able to dissect the effect of oxidative stress from the Cd specific effect, and extensive differences in mt induction levels between these two stressors were observed.

Conclusion: The pmt alleles evolved by a number of recombination events, and exhibited differential inducibilities by Cd, paraquat and molting hormone. In a tolerant population from a metal contaminated site, promoter allele frequencies differed significantly from a reference site and nucleotide polymorphisms in some promoter stretches deviated from neutral expectations, revealing a signature of balancing selection. Our results suggest that the structural differences in the Orchesella cincta metallothionein promoter alleles contribute to the metallothionein -over-expresser phenotype in cadmium tolerant populations.

Figure 1

Architecture of the nine respective metallothionein promoter alleles (pmt). The respective putative transcription factor binding sites are represented as in the legend. Indels are indicated with black triangles. MRE, metal responsive element; ARE, anti-oxidant responsive element; DRE, DNA replication-related element; HERE, 20-hydroxyecdysone responsive element; Inr, initiator; DPE, downstream promoter element; C/EBP, CCAAT enhancer binding protein. The full sequence alignment is given in Additional File , 10.

Figure 2

95% confidence reticulate network of the eight described Orchesella cinca pmt alleles with the Orchesella villosa pmt clone as an out-group, following 1000 replicate bootstraps in Splitstree v4 (uncorrected p for nucleotide substitution, NeighborNet to calculate the distance and Reticulate to calculate the splits). The colored parallel edges refer to the colors in the recombination analysis.

Figure 3

Recombination analysis of Orchesella cincta metallothionein promoter alleles. Bootscanning analysis representing the percentage of permutated trees (left axis) that did coincide between the respective pmt alleles in a sliding window approach (200 bp width, 20 bp step size, Kimura 2-parameter for nucleotide substitution) relative to the sequence position. Only the relationships which trespass the 70% threshold of the permutated trees are presented. On the right axis the p-values of the respective breakpoints are indicated. The colours refer to the parallel edges in the reticulate network (Fig. 2).

Figure 4

Boxplot representing the log transformed basal expression RLU values. One Way ANOVA; F = 71.337 and p = 0.000. A Tukey post-hoc test revealed significance groups, represented by letters.

Figure 5

Dose response relationships of the six luciferase constructs. On the Y-axis the β-galactosidase normalized relative luciferase units (RLU) are presented. On the X-axis the exposure concentrations of cadmium (Cd) are indicated.

Figure 6

Dose response relationships of the six luciferase constructs. On the Y-axis the β-galactosidase normalized relative luciferase units (RLU) are presented. On the X-axis the exposure concentrations of paraquat are indicated.

Figure 7

Dose response relationships of the six luciferase constructs. On the Y-axis the β-galactosidase normalized relative luciferase units (RLU) are presented. On the X-axis the exposure concentrations of 20-hydroxyecdysone (20-E) are indicated.

Figure 8

A: RLU_max estimates from the cadmium (Cd) exposure data, B: Slope estimates from the Cd exposure data, C: EC₅₀ estimates from the Cd exposure data D: RLU_min from the 20-E exposure data.

Figure 9

Sliding window analysis of the Fu and Li's D statistic on the reconstituted dataset of the Plombières population. A step size of 10 bp and a window length of 100 bp were applied. Significant deviations from neutral expectations are represented with ● and ● for p-values < 0.05 and 0.02 respectively. A cartoon of the general architecture of the pmt locus is provided with the putative transcription factor binding sites.