|

|

Received: 10 August 2016    Revised: 17 October 2016    Accepted: 19 October 2016 DOI: 10.1002/ece3.2639

ORIGINAL RESEARCH

Distinct patterns of natural selection in Na+/H+ antiporter genes in Populus euphratica and Populus pruinosa Yuxia Wu | Kuibin Meng | Xiaohui Liang State Key Laboratory of Grassland Agro-Ecosystem, School of Life Sciences, Lanzhou University, Lanzhou, Gansu, China Correspondence Yuxia Wu, State Key Laboratory of Grassland Agro-Ecosystem, School of Life Sciences, Lanzhou University, Lanzhou, Gansu, China. Email: [email protected] Funding information NSFC, Grant/Award Number: 31370396 and 30800865; Program for New Century Excellent Talents in the Ministry of Education in China, Grant/Award Number: NCET-090446

Abstract Salt tolerance genes constitute an important class of loci in plant genomes. Little is known about the extent to which natural selection in saline environments has acted upon these loci, and what types of nucleotide diversity such selection has given rise to. Here, we surveyed genetic diversity in three types of Na+/H+ antiporter gene (SOS, NhaD, and NHX, belonging to the cation/proton antiporter 1 family), which have wellcharacterized essential roles in plant salt tolerance. Ten Na+/H+ antiporter genes and 16 neutral loci randomly selected as controls were sequenced from 17 accessions of two closely related members of the genus Populus, Populus euphratica and Populus pruinosa, section Turanga, which are native to northwest China. The results show that salt tolerance genes are common targets of natural selection in P. euphratica and P. pruinosa. Moreover, the patterns of nucleotide variation across the three types of Na+/H+ antiporter gene are distinctly different in these two closely related Populus species, and gene flow from P. pruinosa to P. euphratica is highly restricted. Our results suggest that natural selection played an important role in shaping the current distinct patterns of Na+/H+ antiporter genes, resulting in adaptive evolution in P. euphratica and P. pruinosa. KEYWORDS

Na+/H+ antiporter, natural selection, nucleotide diversity, Populus euphratica, Populus pruinosa

1 | INTRODUCTION

salt tolerance-­related genes in plants. Positive selection will decrease nucleotide diversity in proximity to loci under selective pressure and

The ability of an organism to undergo biological adaptation to the en-

increase the prevalence of low-­frequency SNPs, whereas balancing

vironment is a product of long-­term evolution (Darwin, 1859). A pop-

selection or local adaptation increases both nucleotide diversity and

ulation is able to evolve when it contains individuals with heritable

the prevalence of medium-­frequency SNPs (Biswas & Akey, 2006; Fu

variation in traits facilitating functional diversification and adaptation.

& Li, 1993; Morrell, Lundy, & Clegg, 2003; Tajima, 1989). Although the

Adaptive evolution has been discussed in detail with respect to genes

effects of demographic history on nucleotide diversity and site fre-

involved in human diseases and/or pathogen response pathways

quency distributions can mimic the effects of selection, demographic

(Manjurano et al., 2015; Tishkoff et al., 2001; Vigano et al., 2012). In

influences affect the whole genome while selection targets specific

plant genomes, genes related to salt adaptation constitute an import-

loci. Thus, demographic effects can be estimated using a genomewide

ant group of loci. It is still largely unknown whether differences in mu-

set of reference genes (Akey et al., 2004; Glinka et al., 2003; Wright &

tation rates across the genome, consistent with natural selection in

Gaut, 2005). Multilocus scans that detect outliers from neutral expec-

response to environmental stresses, can account for the evolution of

tations are more apt to successfully identify plant genes influenced

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2016 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd. 82  |   www.ecolevol.org

Ecology and Evolution 2017; 7: 82–91

|

      83

WU et al.

by natural selection, as, for example, was carried out in studies on

salt adaptation of Populus and the other plants. For this study, we se-

drought and/or salt tolerance-­related genes in sunflower, wild tomato,

lected three types of Na+/H+ antiporter gene which were identified

and Pinus pinaster (Eveno et al., 2008; Fischer et al., 2011; Kane &

in previous studies as being affected by salt stress in P. euphratica (Hu

Rieseberg, 2007), “phenology genes” in balsam poplar (Populus balsam-

& Wu, 2014; Ottow et al., 2005; Wu et al., 2007; Ye et al., 2009), and

ifera, Keller et al., 2011, 2012), and candidate genes for cold hardiness

sequenced almost the complete gene coding regions for ten Na+/H+

in coastal Douglas fir (Pseudotsuga menziesii var. menziesii, Eckert et al.,

antiporter genes in these three classes, and 16 neutral loci randomly

2009).

selected as controls, in order to analyze nucleotide diversity in these

Recent advances in genetic analysis and the advent of the ge-

two Populus species.

nomic era have permitted the isolation and identification of genes responsible for salt tolerance pathways. Salt resistance is a quantitative character controlled by multiple genes in plants, with a given plant species typically containing hundreds or thousands of salt-­responsive genes (Brosche et al., 2005; Gong et al., 2005; Ma

2 | MATERIALS AND METHODS 2.1 | Plant materials

et al., 2013; Qiu et al., 2011; Sottosanto, Gelli, & Blumwald, 2004;

Leaf tissues from 17 different populations, six of P. pruinosa and

Yang et al., 2013). Na+/H+ antiporters provide one mechanism for

eleven of P. euphratica, were collected from sites covering the full ge-

the removal of sodium from the cytoplasm in order to maintain low

ographic range of each species in northwest China (Figure 2). For both

cytoplasmic sodium concentrations in plant cells (Bassil, Ohto, et al.,

species, ten to fifteen individuals were sampled from each population;

2011; Bassil, Tajima, et al., 2011; Fukuda et al., 2004; Shi et al., 2003;

the sampled individuals were separated by at least 100 m to minimize

Yokoi et al., 2002). Plant Na+/H+ antiporters belong to the monova-

the potential for sampling the same clonal individual. Four to eight

lent cation/proton antiporter 1 (CPA1) superfamily (Brett, Donowitz,

individuals per population were selected randomly for this study. DNA

& Rao, 2005). At least three types of Na+/H+ antiporter, NHX, NhaD,

was extracted from silica gel-­dried leaves by a modified CTAB method

and SOS, which differ in subcellular location, have been found in

(Doyle & Doyle, 1987).

plants (Barrero-­Gil, Rodriguez-­Navarro, & Benito, 2007; Blumwald & Poole, 1985; Shi et al., 2000), and it has been suggested that they play important roles in coping with increased Na+ influx and in

2.2 | Amplification, cloning, and sequencing

compartmentalizing Na+ within subcellular compartments under salt

For this work, we selected three types of Na+/H+ antiporter gene

stress (Bassil, Ohto, et al., 2011; Bassil, Tajima, et al., 2011; Fukuda

on the basis of previous studies that have identified genes affected

et al., 2004; Gaxiola et al., 1999; Shi et al., 2003; Yamaguchi et al.,

by salt stress in P. euphratica (Hu & Wu, 2014; Ottow et al., 2005;

2003). Plant salt tolerance can be enhanced by over-­expression of

Wu et al., 2007; Ye et al., 2009). SOS1 and SOS1B encode plasma

+

+

Na /H antiporters (Apse et al., 1999; Shi et al., 2003), and the ex-

membrane-­localized transporters, the six NHX genes belong to the

pression profiles of these genes differ between closely related salt-­

vacuolar type Na+/H+ antiporter group, and the two NhaD genes are

sensitive and salt-­tolerant species (Kant et al., 2006; Zahrana et al.,

chloroplast-­localized (Barrero-­Gil et al., 2007; Blumwald & Poole,

2007).

1985; Shi et al., 2000). We constructed primers (Table S1) for these

The sister species Populus euphratica Olivier and Populus pruinosa

Na/H antiporter genes using data from the poplar genome database

Schrenk are members of the Populus section Turanga Bunge; they grow

(http://genome.jgi-psf.org/Poptr1/Poptr1.home.html), making the

in semi-­arid regions of China and are known for their high levels of

assumption that the gene family consisted of the ten members previ-

stress tolerance (Figure 1). Both species play important roles in the

ously described (Barrero-­Gil et al., 2007; Blumwald & Poole, 1985;

arid ecosystems of northwest China (Li et al., 2003, 2006). Surveys of

Shi et al., 2000).

DNA sequence variability in salt tolerance genes from closely related

The sequenced regions of the Na+/H+ antiporter loci ranged from

Populus species distributed in similar high-­salinity environments may

3,890 to 8,858 bases in length and contained complete coding re-

contribute to a comprehensive understanding of the genetic basis of

gions (excluding some poor-­quality sequences in the case of SOS1B)

F I G U R E   1   Populus euphratica Olivier (left) and Populus pruinosa Schrenk (right), distributed in desert area in northwest China. Photographs by Yuxia Wu

|

WU et al.

84      

F I G U R E   2   Locations of populations sampled and accessions of Populus euphratica and Populus pruinosa used in this study. The geographic distribution ranges of P. euphratica and P. pruinosa in China are shown in green

for each gene (Figure S1), including a total of 17,812 bp of coding re-

3.1 (Excoffier, Laval, & Schneider, 2005) was used to calculate popu-

gion and 32,463 bp of noncoding region, resulting in a total length of

lation structure statistics (FST). The statistical significance of FST for

50,325 bp and 49,518 bp in P. euphratica and P. pruinosa, respectively,

each locus was calculated by comparing the observed values with the

across all genes. The total numbers of sequenced nonsynonymous, synonymous, and noncoding sites for all the Na+/H+ antiporter genes

distribution of FST calculated from 10,000 permutations of sequences among populations.

are given in Table S2. We directly sequenced most of the genes from

The influence of selection on individual salt tolerance genes was

PCR products after treatment with ExoSAP-­IT (USB Corp., Cleveland,

tested using multiple methods. For each species, we separately gen-

OH, USA). Several portions of the genes contained long fragments

erated 105 simulated data sets under the neutral model of Hudson,

with segregating indels that prevented direct sequencing. In these

Kreitman, and Aguade (1987) to simulate a genomewide pattern of

cases, PCR products were cloned into the pGEM19-­T vector (Takara,

diversity using diversity estimates from our 16 reference loci. These

China) after preparing the DNA using the recommended protocol for

data sets were used to assess the probability that diversity estimates

the AxyPrep DNA Gel Extraction kit (AXYGEN, China), and three to

for the salt tolerance genes were consistent with neutral expecta-

five clones were sequenced. The 16 reference loci (Table S3) were ran-

tions. The simulated data sets consisted of 80 and 50 chromosomes in

domly selected from a set used to develop estimates of nucleotide

P. euphratica and P. pruinosa, respectively, numbers which reflected

diversity in P. balsamifera (Olson et al., 2010). We PCR-­amplified and

­average sample sizes across loci for each species (http://home.uchi

directly sequenced these reference loci from all 76 sampled individu-

cago.edu/rhudson1/source/mksamples.html). Absolute nucleotide dif-

als (Table S4). The 16 reference loci contained 162–639 bp of coding

ferentiation for each salt tolerance gene was calculated as πT-S = πT -­ πS,

sequence and ranged in length from 415 to 687 bp per gene, with

where πT is the total nucleotide diversity using all the samples from

a total length of 9,761 bp. Aligned sequences were edited manually

the two species and πS is the average nucleotide diversity within each

using Aligner v.5.1.0 (Codon Code Corporation, Dedham, MA), with all

of the species (Charlesworth, 1998; Keller et al., 2012). Statistical sig-

posterior probabilities >.8 and all polymorphic and heterozygous sites

nificance was determined by comparing the observed πT -­πS to the

visually confirmed. For the ten Na+/H+ antiporter genes and 16 refer-

95% confidence interval calculated from the variation among 104 neu-

ence loci, exon and intron boundaries were determined from cDNA se-

tral coalescent simulations for each salt tolerance gene using DnaSP

quences obtained from the GenBank database or genomic sequences

v.5.10 (Librado & Rozas, 2009), and these estimates were based on

(Ma et al., 2013; Tuskan et al., 2006).

the number of segregating sites and assumed no recombination. The HKA test (Hudson et al., 1987) was applied to evaluate the ratio of si-

2.3 | Data analysis

lent polymorphisms within species to the divergence between species across multiple loci. We conducted an HKA test (Hudson et al., 1987)

To resolve haplotypes from sequences obtained directly from PCR

using the multilocus HKA program available from Jody Hey (https://

products, phase v. 2.1 (Stephens & Scheet, 2005; Stephens, Smith,

bio.cst.temple.edu/~hey/software/software.htm) to assess whether

& Donnelly, 2001) was used. After excluding insertions/deletions (in-

differences could be identified across all loci with 104 simulations. We

dels), genetic parameters were estimated using DnaSP v.5.10 (Librado

then used maximum-­likelihood HKA (mlHKA) to conduct statistical

& Rozas, 2009). We implemented the IMa2 (Hey, 2006, 2010; Hey &

tests comparing each salt tolerance locus between species (Wright

Nielsen, 2004) program using a Markov Chain Monte Carlo (MCMC)

& Charlesworth, 2004; http://wright.eeb.utoronto.ca/programs/). To

approach to analyze gene flow between the two species. Posterior

compare the model in which selection is hypothesized and the neutral

probability densities (proportional to likelihoods) of the model pa-

model at specific loci, the divergence time parameter (T) was set at

rameters were used to assess significance (Guo et al., 2013). Arlequin

15 (Ma et al., 2013; Tuskan et al., 2006) in the mlHKA program with

|

      85

WU et al.

105 MCMC cycles. Comparisons were performed by carrying out a

Most salt tolerance genes and all reference loci had higher nucle-

likelihood-­ratio test using chi-­squared values to determine statistical

otide diversity in P. pruinosa than in P. euphratica (Tables S2 and S4).

significance. The population selection parameters (γ = 2NeS, where

The average nucleotide diversity across all salt tolerance-­related genes

Ne is the effective population size and S is the selection parameter)

was higher in P. pruinosa (θw = 0.0030 ± 0.00097) than in P. euphratica

were determined using the mkprf software described in Bustamante

(θw = 0.0025 ± 0.00076; Table S2). The mean synonymous nucleotide

et al. (2002). We defined sites in the salt tolerance genes and the 16

diversity (πsil) across the 16 reference loci was 0.0034 for P. euphratica

reference loci as replacement or silent sites, depending on whether

and 0.0053 for P. pruinosa (Figure 3 and Table S4).

or not each polymorphism altered the amino acid at a given site in

The 16 reference loci exhibited differentiation among P. euphrat-

P. euphratica and/or P. pruinosa relative to the amino acid present at

ica populations (mean FST = 0.257, range = 0–0.424) but significantly

the same site in Populus trichocarpa. mkprf was based on the posterior

less differentiation among P. pruinosa populations (mean FST = 0.149,

distribution of the sample parameters using the Monte Carlo Markov

range = 0.024–0.381, Table S4). The FST values for salt tolerance

Chain (MCMC) approach. We ran the simulation for 107 cycles and,

genes among populations within species varied between 0.253 and

after discarding the first 104 as “burn-­in,” we sampled every 10th itera-

0.545 in P. euphratica and between 0.124 and 0.463 in P. pruinosa

tion. We summarized the selection parameters for each salt tolerance

(Table S2).

gene and reference locus using the mean and 95% distribution confidence intervals.

3.2 | Gene flow and introgression Gene flow between P. euphratica and P. pruinosa was examined using

3 | RESULTS

the IMa2 model, and the marginal posterior density distribution for

3.1 | Nucleotide diversity and population differentiation

migration rates is shown in Figure 4. Estimates of migration parameters were nonzero for both species at reference loci, with m1 = 0.219 (from P. euphratica to P. pruinosa) and m2 = 0.087 (from P. pruinosa to

Ten Na+/H+ antiporter genes were sequenced from a total of 76 in-

P. euphratica); the probabilities of the migration rate for both direction

dividuals: 52 individuals from eleven P. euphratica populations and 24

were strongly statistically significant (p