• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
PLoS One. 2011; 6(9): e24867.
Published online Sep 19, 2011. doi:  10.1371/journal.pone.0024867
PMCID: PMC3176232

Molecular Population Genetics of Elicitor-Induced Resistance Genes in European Aspen (Populus tremula L., Salicaceae)

Daniel Ortiz-Barrientos, Editor

Abstract

Owing to their long life span and ecological dominance in many communities, forest trees are subject to attack from a diverse array of herbivores throughout their range, and have therefore developed a large number of both constitutive and inducible defenses. We used molecular population genetics methods to examine the evolution of eight genes in European aspen, Populus tremula, that are all associated with defensive responses against pests and/or pathogens, and have earlier been shown to become strongly up-regulated in poplars as a response to wounding and insect herbivory. Our results show that the majority of these defense genes show patterns of intraspecific polymorphism and site-frequency spectra that are consistent with a neutral model of evolution. However, two of the genes, both belonging to a small gene family of polyphenol oxidases, show multiple deviations from the neutral model. The gene PPO1 has a 600 bp region with a highly elevated KA/KS ratio and reduced synonymous diversity. PPO1 also shows a skew toward intermediate frequency variants in the SFS, and a pronounced fixation of non-synonymous mutations, all pointing to the fact that PPO1 has been subjected to recurrent selective sweeps. The gene PPO2 shows a marked excess of high frequency, derived variants and shows many of the same trends as PPO1 does, even though the pattern is less pronounced, suggesting that PPO2 might have been the target of a recent selective sweep. Our results supports data from both Populus and other species which have found that the the majority of defense-associated genes show few signs of selection but that a number of genes involved in mediating defense against herbivores show signs of adaptive evolution.

Introduction

Plants have evolved numerous defensive mechanisms to protect themselves against biotic and abiotic stress [1], [2]. Herbivores and pathogens, in turn, respond to plant defenses by evolving counter-adaptations, that render current defenses less effective or even useless [3], [4]. This can lead to an “arms race” between plants and their enemies [3], [4]. Alternatively natural selection can favor the maintenance of multiple varieties of a defensive trait within a population, either because of negative frequency-dependent selection favoring rare alleles, or due to spatially variable selection favoring different alleles in different environments [4], [5], a model known as the “balanced polymorphism” model or the “trench-warfare model” [3].

Earlier molecular population genetic surveys of plant defense genes have shown that specialist defenses (i.e. defenses active against a small number of natural enemies) appear to be under stronger selection than generalist defenses (i.e. defenses active against a broad number of enemies) [4], [6] and that selection is more likely to maintain variation within specialist defense genes. This suggest that the type of selection on genes mediating herbivore and/or pathogen responses are experiencing, is largely dependent on whether the trait in question is effective against a broad or narrow number of natural enemies: specialist defenses are more likely to evolve in a manner consistent with the balanced-polymorphism model, while generalist defenses often show pattern of intraspecific diversity that are consistent with either a neutral model or, in rare cases, show evidence for recent positive selection, which would be consistent with an arms-race model [6].

In addition to the distinction between specialist and generalist defenses, defense responses can be classified as either constitutive or inducible, with constitutive defense being always present in a plant, whereas induced defenses are only expressed after an enemy attack has occurred [7]. Induced defenses are primed through molecular signaling pathways that allow the plants to tune their response to the type of attacker they are facing. Plant hormones such as salicylic acid (SA), jasmonic acid (JA) and ethylene (ET) are known to play important roles in the regulation of the defense response, and the effectiveness of the different pathways depends on the type of attacker the plant is facing [1]. The SA-dependent defense response is in general more active against pathogens that require a living host to complete their lifecycle (biotrophic pathogens), and SA induction is associated with the accumulation of reactive oxygen species (ROS) and diverse defense related genes, including several pathogenesis-related (PR) proteins. One of the most visible manifestation of SA is the activation of a hypersensitive response (HR) at the local site of infection that ends in programmed cell death (PCD) of surrounding cells [8]. JA is synthesized through the octadecanoid biosynthesis pathway and is only one of several bioactive compounds within the class of octadecanoids which regulates a broad spectrum of plant responses [9] and is induced by, among other things, mechanical wounding and herbivory. Induction of JA triggers a suit of defense-associated biochemical responses, such as the production of protease inhibitors and polyphenol oxidase that serve to deter the herbivore, either by slowing down its development, decreasing its reproductive success, or even repelling it from the plant [10].

Owing to their long life span and ecological dominance in many communities, forest trees are subject to the attack of a diverse array of herbivorous insects and mammals throughout their range. Forest trees have developed a large number of both constitutive and inducible defenses to deal with this suite of natural enemies [11], [12]. By virtue of the large suite of genomics tools available, members of the genus Populus (Salicaceae) have become the de facto model species for studying the effects of pathogens and herbivores, both in greenhouse and under field conditions [13][16]. In this paper we have used molecular population genetics methods to examine the evolution of eight wound-induced genes in European aspen, Populus tremula. These eight genes are all associated with defensive responses against pests and/or pathogens and have earlier been shown to be strongly up-regulated in poplars as a response to insect herbivory [17][20]. We analyze patterns of nucleotide diversity and divergence to detect departures from neutral expectations, with a special emphasis on evaluating the likelihood of either recurrent selective sweeps or balancing selection maintaining amino acid polymorphisms.

Materials and Methods

Plant material and DNA sequencing

Samples of P. tremula were collected from the Swedish Aspen collection (SwAsp collection), which includes 116 individuals sampled from twelve different sites throughout Sweden. The SwAsp collection has been described in detail elsewhere and for more information please refer to Luquez et al. [21]. We selected the following eight genes for study: Allene oxide synthase (AOS5) and Phenylalanine-ammonia lyase (PAL1) encode proteins that are involved in the early steps of the octadecanoid- and phenylpropanoid biosynthesis pathways, respectively [9]. NFXL1 encodes a putative transcription factor that functions as a repressor of elicitor-induced defense responses and is involved in the crosstalk between different defensive responses associated with the SA and abscisic acid (ABA) pathways [22], [23]. Endoxylanase is a family 10 glycoside hydrolase (GHf10) that catalyzes the initial breakdown of xylan in cell walls and has been shown to be associated with cell wall degradation [24]. Its role in plant defense is not well understood, but other GHs, such as chitinases and glucanases, have been identified as pathogen responsive (PR) proteins [25]. CathL encodes a cysteine proteinase of the papain family which is induced through a hypersensitive response and takes part in the programmed cell death (PCD) of cells that surround the entry site of the pathogen infection [26]. The remaining three genes we studied are all from the polyphenol oxidase gene family (PPO1, PPO2, and PPO3), which codes for a group of copper-containing enzymes induced by JA that catalyses the oxidation of diphenolic compounds into quinones that cross-link or alkylate dietary proteins and thereby decrease the nutritional value for herbivorous insects or pathogens [27][29].

Between 11 and 27 individuals were selected to be equally distributed among the 12 sampled populations, to capture as much variation as possible present in the SwAsp collection. These individuals were sequenced for the eight genes using primers that were designed from the corresponding gene model in P. trichocarpa v2.0, publicly available in Phytozome (http://www.phytozome.net/poplar). The gene models used were: POPTR_0009s11280.1 (AOS5), POPTR_0005s18520.1 (CathL), POPTR_0004s17900.1 (GHf10), POPTR_0012s04070.1 (NFXL1), POPTR_0006s12870.1 (PAL1), POPTR_0011s10950.1 (PPO1), POPTR_0001s39920.1 (PPO2) and POPTR_0011s04710.1 (PPO3), respectively). Five of the genes (AOS5, NFXL1, PPO1, PPO2 and PPO3) are intronless, whereas PAL1 has two exons and CathL and GHf10 have four exons each. An effort was made to sequence the entire coding region of all genes, but this was not achieved for GHf10, PPO2, and PPO3. For GHf10 this was due to a region of 500 bp of missing data in the middle of the gene, which was amplified in two, non-overlapping pieces. For PPO2 and PPO3, low quality sequence data at the 3′ and 5′ ends prevented us from obtaining the entire coding region of these genes. Sequencing of AOS5, CathL, NFXL1, PAL1 and PPO1 yielded the entire coding region and short sequences of the 3′ and 5′ UTRs. The length of the sequenced regions ranged from 1035 bp for GHf10 to 3454 bp for NFXL1 (Table 1) and the length of the coding regions ranged from 594 for GHf10 to 3336 bp for NFXL1 (Table 1).

Table 1
Summary statistics of eight inducible defense genes and a genome-wide average calculated over 76 reference genes.

Gene fragments were amplified from diploid genomic DNA using polymerase chain reaction (PCR) and directly sequenced on a Beckman Coulter CEQ 2000XL capillary sequencer at Umeå Plant Science Center. Sequences were verified manually and contigs were assembled using the computer program Sequencher v4.0. Regions with missing or low-quality data were trimmed from all sequences. Multiple sequence alignments were made using Clustal W [30] and adjusted manually using BioEdit (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). Alignments were annotated on the basis of the corresponding gene from the P. trichocarpa genome sequence (available at http://www.phytozome.net/poplar.php).

Digital northerns

The full-length coding regions for all eight genes were obtained from the annotated P. trichocarpa genome (version 2.0, http://www.phytozome.net/poplar). We then obtained all ESTs from the experiments of Sterky et al. [31] and and Ralph et al. [18] and an unpublished study from the Arborea project (http://www.arborea.ulaval.ca/) from GenBank or from PopulusDB (http://www.populus.db.umu.se/) (ESTs downloaded on 2010-10-05). We used these EST resources to obtain “digital northern” analysis of all genes. These digital northern analyses were performed to verify that we had selected genes from the Populus genome that do show wound induction.

Publicly available EST data sets can confound such digital expression analyses in a number of ways and several measures were taken to minimize these effects. First, ESTs are usually sequenced from only one end of the cDNA clone but in some cases EST clones are sequenced from both ends. To identify such cases, all clone IDs were retrieved and duplicate IDs were excluded from the analysis. Second, a relevant comparison of EST abundance is best made if ESTs originate from non-normalized cDNA libraries. Normalization will introduce a certain level of noise in the data set and will lead to an underestimation of mRNA abundance for highly expressed genes. However, it is unlikely that it should change the rank order of genes with respect to gene expression, and data from normalized data still provide semi-quantitative data on transcript abundance. We included data from normalized libraries but these have been labeled in subsequent analyses to aid in interpreting the data. The EST data sets used were further derived from a number of different species of Populus, but the low sequence divergence seen in coding regions between different species in the genus suggest that this will not influence the estimation of expression profiles from the EST libraries [32].

Full-length coding sequences were filtered with the XBLAST program to mask out repetitive elements. Expression profiles were then obtained for all genes BLASTN searches against each EST data set using stringent matching criteria. Alignments were required to show at least a 90% identity across 100 bp to be recorded as a match, to avoid over-estimating transcript abundance of genes belonging to small gene families. The number of BLASTN hits was used as a proxy for expression levels and normalized by the total number of ESTs in each data set, to enable comparisons between different treatments and/or tissues. Because wounding has been shown to affect the expression of most genes involved in both primary and secondary metabolism [33], we have not included a housekeeping gene as reference control which is normally done in comparative studies of gene expression.

Population genetic analyses

Estimates of nucleotide polymorphism (segregating sites (S), nucleotide diversity (π)), between-species divergences, for synonymous and non-synonymous sites separately, and statistical tests of neutrality (Tajima's D, Fay & Wu's H and the McDonald-Kreitman test), were obtained using the computer program DnaSP v.4.50.2 (http://www.ub.es/dnasp/). For analyses requiring an outgroup, a single sequence of the homologous region in P. trichcarpa was used. To provide an estimate of the distribution of genome-wide levels of polymorphism we used data from 77 loci corresponding to 76 unique genes evenly distributed throughout the genome of Populus. These reference genes were chosen without prior knowledge of their function [see 34 for more details] and we refer to them as collectively as “reference genes” hereafter.

In order to evaluate significance statistics of Fay and Wu's H and Tajima's D, we performed coalescent simulations using Richard Hudson's ms program (available at http://home.uchicago.edu/~rhudson1/source/mksamples.html). For each summary statistic the null distribution under neutrality was obtained from 104 simulations, conditioned on the original sample configuration and sequence length. The demographic history of P. tremula is characterized by a moderately strong bottleneck that occurred roughly 770 kya and to account for this, coalescent simulations were performed based on the bottleneck model described in Ingvarsson [34]. Briefly, for each simulation replicate, parameters for the bottleneck model were drawn from the posterior distribution of these parameters as described in Ingvarsson [34]. The corresponding summary statistic calculated from the observed data was then compared to the full distribution of the simulated data sets to determine where they fell in the range of simulated values.

Since LD extends only a few hundred bps in P. tremula [34], it is possible that any signals of natural selection only affect a small region around the target site of selection. We therefore also performed sliding window analyses for synonymous diversity, Tajima's D, Fay and Wu's H and the KA/KS ratio to study how patterns of polymorphism, divergence and site frequency spectra (SFS) vary across genes. The sliding window analysis of diversity at synonymous sites was performed using a 75 bp window that was moved in 8 bp increments across the gene, while the rest of the analyses were performed using a 250 bp window moved in 25 bp increments. Smaller window sizes yielded too little polymorphism in many windows, while larger windows increase the risk of averaging over too many sites with independent evolutionary histories, thereby reducing the power to detect regions with enhanced genetic differentiation. We used coalescent simulation to determine whether specific window segments showed significant deviations from neutral expectations. We simulated the same number of alleles that were present in our sample and simulations were run conditional on the number of segregating sites observed in each window. We performed simulations without recombination as we are simulating only short gene-segments. Furthermore, ignoring recombination is conservative with respect to identifying outlier-values for Tajima's D and Fay and Wu's H. For the KA/KS ratio we adopted a slightly different method to assess outlier windows. We calculated the 95% percentile from the distribution of KA/KS from the 76 control loci and used this value as a cutoff limit to assess which window-regions that show elevated KA/KS ratios. We regard the sliding window analyses as exploratory and as a way to identify regions worth studying in greater detail and we therefore do not correct for multiple testing in the analysis.

Results

Digital northerns

We compared gene expression across different tissues using digital northerns based on available EST data sets (Figure 1). Although there is substantial variation across different tissues and wounding treatments, 7 of 8 genes studied here show a consistently higher induction in wounded or infected tissues (paired t-test, t = 3.07, p = 0.017, Figure 1). The only exception is AOS5 which showed very low expression across all data sets we analyzed. This confirms that the majority of the genes we studied are indeed wound induced, showing higher gene expression following wounding.

Figure 1
Heat map of digital northern analyses of expression data for eight defense-associated genes in Populus.

Intraspecific patterns of nucleotide polymorphism

Between 22 and 54 alleles were sequenced for AOS5, CathL, GHf10, NFXL1, PAL1, PPO1, PPO2 and PPO3 (Table 1). The complete coding region together with some flanking regions were sequenced for five of the genes (AOS5, CathL, NFXL1, PAL1 and PPO1) while only a part of the coding regions was obtained for GHf10, PPO2 and PPO3. Seven of the genes show substantial levels of polymorphism, with the number of segregating sites ranging from 38 to 87 (Table 1). NFXL1, by contrast, contained only 12 segregating sites despite having the longest contiguous stretch of sequence. Overall, the three PPO genes show higher diversity at both synonymous and non-synonymous sites, compared to the genome-wide average in P. tremula, while AOS5, CathL, and GHf10 are showing slightly to moderately reduced levels of sequence diversity at synonymous sites but elevated levels of diversity at non-synonymous. NFXL1 and PAL1 show relatively low levels of sequence diversity at both synonymous and non-synonymous sites (Table 1).

The sliding window analysis of the synonymous diversity shows a striking dip in a region in the middle of PPO1 with a three-fold reduction in polymorphism (Figure 2F). Even though PPO2 shows a similar pattern, with a reduction of synonymous polymorphisms in the center of the gene, the reduction it is not as pronounced as in PPO1. The remaining six genes show a more homogenous distribution of polymorphism (Figure 2).

Figure 2
Sliding window of intraspecific diversity at synonymous sites (πsyn).

If any of the genes are under the influence of strong natural selection it should be possible to determine this from shifts in the site frequency spectrum (SFS). An excess of intermediate frequency variants in the SFS, as indicated by a positive Tajima's D, suggest that the gene might be under balancing selection whereas an excess of high frequency derived variants, indicated by a negative value of Fay and Wu's H, is a sign of directional selection, as high frequency derived variants are very rare under a neutral model. Two of the defense genes show significant shifts in the site frequency spectrum compared to neutral expectations (Table 1), with PPO1 having a significantly positive Tajima's D (D = 1.09, p = 0.02) and PPO2 have a significantly negative Fay and Wu's H (H = −2.63, p = 0.006). The other six genes (AOS5, CathL, GHf10, NFXL1, PAL1 and PPO3) do not show any deviations from neutral expectations (Table 1).

The sliding window analyses show that in PPO1, a region at the 5′ end of the coding region harbors an excess of both intermediate- and high frequency variants (positive Tajima's D and negative Fay and Wu's H, Figure 3F). Furthermore, PAL1 and PPO2 show regions with very distinct negative values of Fay and Wu's H, for PAL1 at the 5′ end of exon 2 and for PPO2 most clearly at the 5′ end of the gene. For the remaining genes there are a few regions showing weak deviations from neutrality but none of the genes show any systematic patterns. The simulated confidence intervals for each window across the eight genes are presented in Table S1.

Figure 3
Sliding window of Tajima's D (red line) and Fay and Wu's normalized H (black line).

Polymorphism and Divergence

Three of the defense genes (CathL, GHf10 and NFXL1) have lower or equal divergence at synonymous sites compared to the genome-wide average, whereas the other five defense genes are showing a higher degree of divergence. At non-synonymous sites only PAL1 and NFXL1 exhibit a lower or equal degree of divergence compared to the genome-wide average (Table 1).

None of the eight defense genes are showing KA/KS ratios that exceed unity. However, sliding window analyses identifies a 600 bp region in the middle of the PPO1 gene that show a greatly elevated KA/KS ratio (KA/KS>15, Figure 4). CathL, GHf10, PPO2, and to lesser extent AOS5 and NFXL1, are also harboring small regions with elevated KA/KS ratios, but these are less distinct than the very clear peak of elevated KA/KS seen in PPO1. These results suggest that adaptive evolution has occurred in the PPO1 gene, as evidenced by the preferential fixation of non-synonymous mutations compared to synonymous mutations. In line with this, we find that a McDonald-Kreitman (MK) test, which compares the rate of divergence to levels of standing nucleotide polymorphism within species, also identifies PPO1 as a gene whichs show a significant deviation from neutral expectations with a -log10 neutrality index (NI) of 0.721 (Fisher's exact test, p = 3×10−5), suggesting again preferential fixation of non-synonymous mutations at PPO1. None of the other seven genes show any deviations from expectations in the MK-tests (Table 2).

Figure 4
Sliding window of nonsynonymous-to-synonymous ratio of diversity (red line) and divergence (black line).
Table 2
Number of fixed and polymorphic mutations at synonymous and replacement sites in the coding regions of Populus tremula inducible defense genes.

Discussion

We have examined patterns of nucleotide variation in eight wound-induced genes from European aspen (P. tremula). These genes are representative of the two main pathways that control wound-induced defense responses, the octadecanoid and the salicylate pathway [35]. The digital northern analyses confirm that all but one gene show increased gene transcription following wounding, confirming that these genes (with the exception of AOS5) are wound induced (Figure 1), in line with a number of recent studies that have investigated transcriptional changes following wounding and herbivory in Populus [17][19], [36], [37].

With a few notable exceptions (see below), patterns of nucleotide polymorphism in the eight wound induced genes conform to neutral expectations, and most genes also show low KA/KS ratios, indicating that purifying selection acted on most non-synonymous polymorphisms in these genes.

While the majority of genes we have studied show few sign of selection apart from a purifying selection on non-synonymous polymorphisms, two genes from the polyphenol oxidase gene family, PPO1 and PPO2, show several signs that are indicative of the action of natural selection at these two genes. Both PPO1 and PPO2 show significant skew in their site frequency spectra, at least for parts of their coding regions. PPO1 show an excess of intermediate frequency variants across the entire gene (D = 1.09, p = 0.02), and the sliding window analysis of the SFS shows that this patterns is most pronounced in the regions close to the 5′ and 3′ ends of the gene. In addition, the 5′ end of the gene shows a clear excess of high frequency derived variants (negative Fay and Wu's H, Figure 3). Furthermore, even though the KA/KS ratio does not exceed unity for the entire PPO1 gene, a 600 bp region in the center of the gene display a greatly elevated KA/KS ratio, exceeding unity by up to twentyfold (Figure 4) and an MK-test also indicates that PPO1 harbors an excess of non-synonymous fixations. In addition, the same region showing an elevated KA/KS ratio also shows markedly reduced levels of synonymous polymorphism compared to surrounding regions in the PPO1 gene (Figure 2). Taken together, this suggests that PPO1 has experienced recurrent adaptive evolution, with a preferential fixation of non-synonymous mutations. The deviations we observe in the SFS, with an excess of both high frequency derived and intermediate frequency variants hints at a strong haplotype structure in the 5′ region of PPO1 (Figure 3), that could be a signature of a recently completed sweep at PPO1 (eg. [38][40]).

When searching for conserved domains in PPO1 (www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi), we found that the 600 bp region with elevated KA/KS ratio and reduced synonymous diversity is located close to the end of the tyrosinase superfamily domain, the domain that contains the active site for all PPOs. This domain consists of a di-nuclear copper center, with two copper ions, each coordinated to three histidine side chains [41]. Both the peak in the KA/KS ratio and the deepest dip of the intraspecific diversity lies within the second histidine residue of the tyrosinase domain, hinting at the possibility that the possible target of selection might affect the enzymatic catalytic properties of PPO1. Further biochemical characterizations of the different allelic variants of PPO1 are required to test this hypothesis.

In contrast to PPO1, PPO2 does not show an elevated KA/KS ratio, with the exception of a short region in the 3′ end of the gene (Figure 4) in a domain that is generally thought to be involved in activation of the enzyme by proteolytic cleavage [41]. Furthermore, PPO2 shows a significant excess of high frequency derived variants across the entire gene (Fay and Wu's H = −2.63, p = 0.006). The sliding window analysis also identified a region in the middle of the gene that shows no deviation from neutrality, but that is flanked on both sides by regions showing an excess of high frequency derived variants (Figure 3)- a sign that could indicate that also PPO2 went through a recent selective sweep [38], [39]. The MK-test only indicated a weak trend towards an excess of non-synonymous fixations at PPO2 (p = 0.107), but the MK-test only detects significant deviations from neutrality when the gene has been under long-term adaptive evolution. Interestingly, judging by the location of possible flanking regions showing significant deviations in the SFS, a putative target of selection in PPO2 would also be located in the tyrosinase domain of the protein, just as in PPO1.

The two genes showing signs of either long-term or more recent natural selection are members of the polyphenol oxidase gene family, a small gene family consisting of up to 10–12 different members in the Populus genome [29]. PPO enzymes are generally categorized as generalist defenses when it comes to substrate specificity [29], as the function of PPOs in the defensive response is to catalyze the oxidation of diphenolic compounds into quinones, which in turn cross-link or alkylate essential amino acids and sulfhydryl groups in dietary proteins [29]. This leads to a general reduction in the nutritional value of the tissue, and ultimately to lower growth rates or even starving herbivores [29]. Transgenic plants with altered expression of PPOs have also been shown to have detrimental effects on insect performance [42][47], although results vary between different herbivore guilds and with the age of herbivores [48] In Populus both PPO1 and PPO2 have been shown to be wound-inducibe, with PPO1 being exclusively expressed in damaged leaves while PPO2 is primarily expressed in stems, petioles and roots [28].

Genes like PPO, involved in generalized defense, often show few signs of natural selection compared to genes involved in specialized defenses [5], [6] and it may therefore seem rather surprising that two of the three PPO genes included in this study show signs of natural selection. In fact, the two PPO genes in this study show patterns of selection that are reminiscent of the patterns found from trypsin inhibitor (TI) genes in Populus [20], [49], [50]. The TI genes are generally thought to represent a much more specialized defense mechanism than that provided by the PPO genes due to the highly selective nature of TI genes when targeting specific protases [11], [51]. However, recent results have shown that PPO proteins are not only resistant to proteolysis in insect guts, but that limited proteolysis actually is required to activate the protein from the latent form in which it is stored in plant leaves [27], [41], [45]. In this respect it is interesting to note that the region showing strong signs of adaptive evolution in PPO1, and to a lesser extent also in PPO2, is located close to the active site of the PPO protein and also close to a region that has been speculated to be involved in controlling conditions under which the PPO protein becomes activated in the insect gut [29], [41]. One possibility is therefore that the signs of selection we observe in PPO1 and PPO2 are linked to natural selection in herbivores, preventing or reducing the activation of the PPO protein, and to counter-selection in plants for maintaining this function. Such a co-evolutionary scenario is clearly a speculation at this point and more data on both the functional consequences of mutations at sites targeted by natural selection and possible counter-selection in herbivores are needed before this question can be answered conclusively.

To conclude, we have shown that patterns of polymorphism and divergence at most of the eight wound-induced genes we studied showed little deviation from neutrality. The exceptions a two polyphenol oxidase genes which both appear to have been the target of recent selective sweeps, and in the case of PPO1, also long-term adaptive evolution. Our results support the notion that the majority of wound-induced genes are not under strong selection in Populus. However, the results show that some biochemical herbivore defenses, that are strongly induced following mechanical damage or insect attack, show signs of adaptive evolution in Populus [49], [50]. These results supports data from other species, such as Arabidopsis and maize, which have found that the the majority of defense-associated genes show few signs of selection and that genes showing signs of natural selection are usually involved in specialist defenses [5], [6], [52], [53].

Supporting Information

Table S1

Number of segregating sites, nucleotide diversity, observed values of Tajima's D and Fay and Wu's H and their simulated 95% confidence intervals for the sliding window analysis.

(PDF)

Acknowledgments

We are grateful to F. Bokma and D. Hall for comments on earlier versions that notably improved the manuscript.

Footnotes

Competing Interests: The authors have declared that no competing interests exist.

Funding: This work was funded by grants from the Swedish Research Council and C.F. Lundström's stiftelse to PKI. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1. Kessler A, Baldwin IT. Plant responses to insect herbivory: The Emerging Molecular Analysis. Annu Rev Plant Biol. 2002;53:299–328. doi: 10.1146/annurev.arplant.53.100301.135207. [PubMed]
2. Howe GA, Jander G. Plant Immunity to Insect Herbivores. Annu Rev Plant Biol. 2008;59:41–66. doi: 10.1146/annurev.arplant.59.032607.092825. [PubMed]
3. Bergelson J, Dwyer G, Emerson JJ. Models and data on plant-enemy coevolution. Annu Rev Genet. 2001;35:469–499. doi: 10.1146/annurev.genet.35.102401.090954. [PubMed]
4. de Meaux J, Mitchell-Olds T. Evolution of plant resistance at the molecular level: ecological context of species interactions. Heredity. 2003;91:345–352. [PubMed]
5. Tiffin P, Moeller DA. Molecular evolution of plant immune system genes. Trends in Genetics. 2006;22:662–670. doi: 10.1016/j.tig.2006.09.011. [PubMed]
6. Tiffin P, Hacker R, Gaut BS. Population Genetic Evidence for Rapid Changes in Intraspecific Diversity and Allelic Cycling of a Specialist Defense Gene in Zea. Genetics. 2004;168:425–434. doi: 10.1534/genetics.103.023028. [PMC free article] [PubMed]
7. Karban R, Baldwin IT. Induced responses to herbivory. Chicago, IL: University of Chicago Press; 1997.
8. Vlot AC, Dempsey DA, Klessig DF. Salicylic Acid, a Multifaceted Hormone to Combat Disease. Annu Rev Phytopathol. 2009;47:177–206. doi: 10.1146/annurev.phyto.050908.135202. [PubMed]
9. Schaller F. Enzymes of the biosynthesis of octadecanoid-derived signalling molecules. J Exp Bot. 2001;52:11–23. doi: 10.1093/jexbot/52.354.11. [PubMed]
10. Kant MR, Baldwin IT. The ecogenetics and ecogenomics of plant-herbivore interactions: rapid progress on a slippery road. Current Opinion in Genetics & Development. 2007;17:519–524. doi: 10.1016/j.gde.2007.09.002. [PubMed]
11. Constabel CP. A survey of herbivore-induced defense proteins and phytochemicals. In: Agrawal AA, Tuzun S, Bent E, editors. Induced plant defenses against pathogens and herbivores: biochemistry, ecology and agriculture. St. Paul, MN: APS Press; 1999. pp. 137–166.
12. Eyles A, Bonello P, Ganley R, Mohammed C. Induced resistance to pests and pathogens in trees. New Phytologist. 2010;185:893–908. doi: 10.1111/j.1469-8137.2009.03127.x. [PubMed]
13. Whitham TG, Young WP, Martinsen GD, Gehring CA, Schweitzer JA, et al. Community and ecosystem genetics: a consequence of the extended phenotype. Ecology. 2003;84:559–573. doi: 10.1890/0012-9658(2003)084[0559:CAEGAC]2.0.CO;2.
14. Constabel CP, Major IT. Molecular biology and biochemistry of induced insect defense in Populus. Recent Advances in Phytochemistry. 2005;39:119–143.
15. Osier TL, Lindroth RL. Genotype and environment determine allocation to and costs of resistance in quaking aspen. Oecologia. 2006;148:293–303. doi: 10.1007/s00442-006-0373-8. [PubMed]
16. Philippe RN, Bohlmann J. Poplar defense against insect herbivores. Bot. 2007;85:1111–1126. doi: 10.1139/B07-109.
17. Christopher ME, Miranda M, Major IT, Constabel CP. Gene expression profiling of systemically wound-induced defenses in hybrid poplar. Planta. 2004;219:936–947. doi: 10.1007/s00425-004-1297-3. [PubMed]
18. Ralph S, Oddy C, Cooper D, Yueh H, Jancsik S, et al. Genomics of hybrid poplar (Populus trichocarpa deltoides) interacting with forest tent caterpillars (Malacosoma disstria): normalized and full-length cDNA libraries, expressed sequence tags, and a cDNA microarray for the study of insect-induced defences. Molecular Ecology. 2006;15:1275–1297. doi: 10.1111/j.1365-294X.2006.02824.x. [PubMed]
19. Ralph S, Chun HJ, Cooper D, Kirkpatrick R, Kolosova N, et al. Analysis of 4,664 high-quality sequence-finished poplar full-length cDNA clones and their utility for the discovery of genes responding to insect feeding. BMC Genomics. 2008;9:57. doi: 10.1186/1471-2164-9-57. [PMC free article] [PubMed]
20. Philippe RN, Ralph SG, Külheim C, Jancsik SI, Bohlmann J. Poplar defense against insects: genome analysis, full-length cDNA cloning, and transcriptome and protein analysis of the poplar Kunitz-type protease inhibitor family. New Phytologist. 2009;184:865–884. doi: 10.1111/j.1469-8137.2009.03028.x. [PubMed]
21. Luquez V, Hall D, Albrectsen B, Karlsson J, Ingvarsson P, et al. Natural phenological variation in aspen (Populus tremula): the SwAsp collection. Tree Genetics & Genomes. 2008;4:279–292. doi: 10.1007/s11295-007-0108-y.
22. Asano T, Masuda D, Yasuda M, Nakashita H, Kudo T, et al. AtNFXL1, an Arabidopsis homologue of the human transcription factor NF-X1, functions as a negative regulator of the trichothecene phytotoxin-induced defense response. The Plant Journal. 2008;53:450–464. doi: 10.1111/j.1365-313X.2007.03353.x. [PubMed]
23. Mussig C, Schroder F, Usadel B, Lisso J. Structure and putative function of NFX1-like proteins in plants. Plant Biology. 2010;12:381–394. doi: 10.1111/j.1438-8677.2009.00303.x. [PubMed]
24. Caspers MPM, Lok F, Sinjorgo KMC, Van Zeijl MJ, Nielsen KA, et al. Synthesis, processing and export of cytoplasmic endo-β-1,4-xylanase from barley aleurone during germination. The Plant Journal. 2001;26:191–204. doi: 10.1046/j.0960-7412.2001.01019.x. [PubMed]
25. Minic Z. Physiological roles of plant glycoside hydrolases. Planta. 2007;227:723–740. doi: 10.1007/s00425-007-0668-y. [PubMed]
26. Xia Y. Proteases in pathogenesis and plant defence. Cell Microbiol. 2004;6:905–913. doi: 10.1111/j.1462-5822.2004.00438.x. [PubMed]
27. Constabel CP, Yip L, Patton JJ, Christopher ME. Polyphenol Oxidase from Hybrid Poplar. Cloning and Expression in Response to Wounding and Herbivory. Plant Physiol. 2000;124:285–296. doi: 10.1104/pp.124.1.285. [PMC free article] [PubMed]
28. Wang J, Constabel CP. Three polyphenol oxidases from hybrid poplar are differentially expressed during development and after wounding and elicitor treatment. Physiologia Plantarum. 2004;122:344–353. doi: 10.1111/j.1399-3054.2004.00403.x.
29. Constabel CP, Barbehenn R. Induced plant resistance to herbivores. New York, NY: Springer; 2008. Defensive roles of polyphenol oxidases in plants. pp. 253–269.
30. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994;22:4673–4680. [PMC free article] [PubMed]
31. Sterky F, Bhalerao RR, Unneberg P, Segerman B, Nilsson P, et al. A Populus EST resource for plant functional genomics. Proceedings of the National Academy of Sciences of the United States of America. 2004;101:13951–13956. doi: 10.1073/pnas.0401641101. [PMC free article] [PubMed]
32. Ingvarsson PK. Gene Expression and Protein Length Influence Codon Usage and Rates of Sequence Evolution in Populus tremula. Mol Biol Evol. 2007;24:836–844. doi: 10.1093/molbev/msl212. [PubMed]
33. Rehrig EM, Appel HM, Schultz JC. Measuring “normalcy” in plant gene expression after herbivore attack. Molecular Ecology Resources. 2011;11:294–304. doi: 10.1111/j.1755-0998.2010.02929.x. [PubMed]
34. Ingvarsson PK. Multilocus Patterns of Nucleotide Polymorphism and the Demographic History of Populus tremula. Genetics. 2008;180:329–340. doi: 10.1534/genetics.108.090431. [PMC free article] [PubMed]
35. Reymond P, Farmer EE. Jasmonate and salicylate as global signals for defense gene expression. Current Opinion in Plant Biology. 1998;1:404–411. doi: 10.1016/S1369-5266(98)80264-1. [PubMed]
36. Major IT, Constabel CP. Molecular analysis of poplar defense against herbivory: comparison of wound- and insect elicitor-induced gene expression. New Phytologist. 2006;172:617–635. doi: 10.1111/j.1469-8137.2006.01877.x. [PubMed]
37. Miranda M, Ralph SG, Mellway R, White R, Heath MC, et al. The Transcriptional Response of Hybrid Poplar (Populus trichocarpa x P. deltoides to Infection by Melampsora medusae Leaf Rust Involves Induction of Flavonoid Pathway Genes Leading to the Accumulation of Proanthocyanidins. MPMI. 2007;20:816–831. doi: 10.1094/MPMI-20-7-0816. [PubMed]
38. Kim Y, Stephan W. Detecting a Local Signature of Genetic Hitchhiking Along a Recombining Chromosome. Genetics. 2002;160:765–777. [PMC free article] [PubMed]
39. Przeworski M. The Signature of Positive Selection at Randomly Chosen Loci. Genetics. 2002;160:1179–1189. [PMC free article] [PubMed]
40. Meiklejohn CD, Kim Y, Hartl DL, Parsch J. Identification of a Locus Under Complex Positive Selection in Drosophila simulans by Haplotype Mapping and Composite-Likelihood Estimation. Genetics. 2004;168:265–279. doi: 10.1534/genetics.103.025494. [PMC free article] [PubMed]
41. Marusek CM, Trobaugh NM, Flurkey WH, Inlow JK. Comparative analysis of polyphenol oxidase from plant and fungal species. Journal of Inorganic Biochemistry. 2006;100:108–123. doi: 10.1016/j.jinorgbio.2005.10.008. [PubMed]
42. Li L, Steffens J. Overexpression of polyphenol oxidase in transgenic tomato plants results in enhanced bacterial disease resistance. Planta. 2002;215:239–247. doi: 10.1007/s00425-002-0750-4. [PubMed]
43. Thaler JS, Farag MA, Paré PW, Dicke M. Jasmonate-deficient plants have reduced direct and indirect defences against herbivores. Ecology Letters. 2002;5:764–774. doi: 10.1046/j.1461-0248.2002.00388.x.
44. Thipyapong P, Hunt MD, Steffens JC. Antisense downregulation of polyphenol oxidase results in enhanced disease susceptibility. Planta. 2004;220:105–117. doi: 10.1007/s00425-004-1330-6. [PubMed]
45. Wang J, Constabel CP. Polyphenol oxidase overexpression in transgenic Populus enhances resistance to herbivory by forest tent caterpillar (Malacosoma disstria). Planta. 2004;220:87–96. doi: 10.1007/s00425-004-1327-1. [PubMed]
46. Mahanil S, Attajarusit J, Stout MJ, Thipyapong P. Overexpression of tomato polyphenol oxidase increases resistance to common cutworm. Plant Science. 2008;174:456–466. doi: 10.1016/j.plantsci.2008.01.006.
47. Bhonwong A, Stout MJ, Attajarusit J, Tantasawat P. Defensive Role of Tomato Polyphenol Oxidases against Cotton Bollworm (Helicoverpa armigera) and Beet Armyworm (Spodoptera exigua). J Chem Ecol. 2008;35:28–38. doi: 10.1007/s10886-008-9571-7. [PubMed]
48. Barbehenn R, Jones C, Yip L, Tran L, Constabel C. Limited impact of elevated levels of polyphenol oxidase on tree-feeding caterpillars: assessing individual plant defenses with transgenic poplar. Oecologia. 2007;154:129–140. doi: 10.1007/s00442-007-0822-z. [PubMed]
49. Ingvarsson PK. Molecular Population Genetics of Herbivore-induced Protease Inhibitor Genes in European Aspen (Populus tremula L., Salicaceae). Mol Biol Evol. 2005;22:1802–1812. doi: 10.1093/molbev/msi171. [PubMed]
50. Talyzina N, Ingvarsson P. Molecular Evolution of a Small Gene Family of Wound Inducible Kunitz Trypsin Inhibitors in Populus. Journal of Molecular Evolution. 2006;63:108–119. doi: 10.1007/s00239-005-0249-2. [PubMed]
51. Major IT, Constabel CP. Functional Analysis of the Kunitz Trypsin Inhibitor Family in Poplar Reveals Biochemical Diversity and Multiplicity in Defense against Herbivores. Plant Physiol. 2008;146:888–903. doi: 10.1104/pp.107.106229. [PMC free article] [PubMed]
52. Moeller DA, Tiffin P. Genetic Diversity and the Evolutionary History of Plant Immunity Genes in Two Species of Zea. Molecular Biology and Evolution. 2005;22:2480–2490. doi: 10.1093/molbev/msi247. [PubMed]
53. Bakker EG, Traw MB, Toomajian C, Kreitman M, Bergelson J. Low Levels of Polymorphism in Genes That Control the Activation of Defense Response in Arabidopsis thaliana. Genetics. 2008;178:2031–2043. doi: 10.1534/genetics.107.083279. [PMC free article] [PubMed]

Articles from PLoS ONE are provided here courtesy of Public Library of Science
PubReader format: click here to try

Formats: