Mol Genet Genomics DOI 10.1007/s00438-014-0820-x

Original Paper

Lz‑0 × Berkeley: a new Arabidopsis recombinant inbred line population for the mapping of complex traits Arnaud Capron · Xue Feng Chang · Chun Shi · Rodger Beatson · Thomas Berleth 

Received: 1 October 2012 / Accepted: 25 January 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  This study describes the generation and test of a genetic resource suited to identify determinants of cell biological traits in plants. The use of quantitative trait loci (QTL) mapping for a better genetic understanding of cell biological traits is still at an early stage, even for biotechnologically important cell properties, such as the dimensions of fiber cells. A common strategy, the mapping of QTLs in recombinant inbred line (RIL) populations, is limited by the fact that the existing RIL populations exploit only a small fraction of the existing natural variation. Here, we report the mapping of QTLs impacting on the length of fiber cells in Arabidopsis inflorescence stems in a newly generated RIL population derived from a cross between the accessions Berkeley and the little known Lz-0. Through inbreeding of individual F2 plants, a total of 159 new F8 lines were

produced and genotyped with a set of 49 single nucleotide polymorphism markers. The population was successfully used not only for the mapping of three QTLs controlling fiber length, but also to map five QTL controlling flowering time under short and long-day conditions. Our study demonstrates the usefulness of this new genetic resource by mapping in it QTLs underlying a poorly explored cellular trait as well as an already better explored regulatory pathway. The new RIL population and an online platform for the continuous supplementation of genetic markers will be generally available to substantially broaden the genetic diversity through which loci with impact on plant quantitative traits can be identified. Keywords Arabidopsis · Recombinant inbred line · SNP · QTL mapping · Flowering time · Fiber

Communicated by A. Schnittger. Electronic supplementary material  The online version of this article (doi:10.1007/s00438-014-0820-x) contains supplementary material, which is available to authorized users. A. Capron · C. Shi · T. Berleth (*)  University of Toronto-CSB, 25 Willcocks St., Toronto, ON M5S 3B2, Canada e-mail: [email protected] Present Address: A. Capron  Department of Botany, University of British Columbia, 6270 University Blvd., Vancouver, BC V6T 1Z4, Canada X. F. Chang · R. Beatson  BCIT, 3700 Willingdon Ave., Burnaby, BC V5G 3H2, Canada Present Address: C. Shi  Agriculture and Agri-Food Canada, 2585 County Road 20, Harrow, ON N0R 1G0, Canada

Introduction The relatively recent evolutionary appearance of angiosperms accentuates the usefulness of genetic model systems such as Arabidopsis thaliana. Mutant screens in genetic model plants have identified key regulators in many cellular and organismal processes, which were then found to be recognizably conserved in many agronomically important plant species. For example, the gene families mediating auxin-responsive gene expression were found to be conserved from Arabidopsis to poplar (Kalluri et al. 2007) and rice (Wang et al. 2007). The mapping of loci controlling quantitative traits of interest in Arabidopsis accessions, significantly expands the power of Arabidopsis genetics, because of the inclusion of an ever-growing number of highly diverse local variants (accessions) and, because many important plant traits may not lend themselves to

13



identification through single knock-out mutants. Cell parameters that can be measured by semi-automated devices, such as cell dimensions, are particularly suited to be genetically addressed through quantitative trait loci (QTLs), but this line of research is still at an early stage. Genetic advances in understanding molecular mechanisms underlying plant traits have greatly benefited from novel strategies for mapping and eventually cloning of QTLs in Arabidopsis. The power of this approach is proportional to the amount of genetic diversity of accessions introduced into mapping populations, which makes the inclusion of more distantly related accessions in mapping populations and ongoing requirement. The use of recombinant inbred lines (RILs) for the mapping of QTLs has numerous advantages, including the unlimited availability of material for phenotyping and the possibility to compare results across many mapping experiments in the entire research community, which, in turn, continuously increases the density of the available genetic map for each RIL population as illustrated by the Col4  ×  Ler-0 set released by Lister and Dean with 67 markers in 1993 (Lister and Dean 1993) and expanded throughout the years up to 773 in 2012 (Sanyal and Randal Linder 2012). These advantages, as well as the relatively easily achievable high mapping resolution in RILs in combination with the compact genome of Arabidopsis has made it possible that in Arabidopsis, QTLs for important plant traits were not only mapped but eventually molecularly identified (for review, see Alonso-Blanco et al. 2009, also the V.A.S.T. website contains useful information on Arabidopsis RIL population http://www.inra.fr/vast/RILs.htm). The only possible disadvantage in the use of RILs, as opposed to individual mapping populations, lies in the risk of focusing too strongly on too few well-established RIL populations, but this can be compensated by the introduction of novel, outlier genotypes into new RIL populations. The physical and chemical properties of fibers in trees are of great significance to produce pulp for subsequent formation of paper products. Variations in pulp fiber morphological and physical properties, such as fiber length, width, strength and coarseness, have a profound influence on the properties of paper products (Dinwoodie 1965; Page and Seth 1980). Tree breeding has been used for fiber optimization in woody plants used for pulp production (Via et al. 2004). However, systematic studying of fiber property genetics in trees is obstructed by long generation times, fewer genetic and genomic resources and even the less detailed characterization of fiber ontogenies. In this study we have sought to combine: (a) a suitable, semi-automatically recordable cellular trait with (b) the power of Arabidopsis QTL mapping in RILs and (c) the introduction of an outlier accession into a newly established RIL population. The new Lz-0  ×  Berkeley

13

Mol Genet Genomics

population comprises 159 RILs, which were genotyped for an initial genetic map comprising 49 markers. The population and the genetic map were successfully used to map five QTLs controlling flowering time as well as three QTLs affecting fiber length in inflorescence stem. The population is generally available to search for expanded genetic diversity within the available RIL collections and for future mapping experiments.

Materials and methods Plant material The set of RILs was produced from a cross between Lz0 and Berkeley. The parents were crossed and 40 F1 were selected to start the collection and allowed to self and one F2 plant from each was selected and five F3 individuals from each F2 were kept, raising the number of plants in the collection to 200. Each individual was allowed to self and one plant was selected at each generation to produce the next one by selfing, up to the F8. 159 F8 were eventually recovered. In addition, F2 populations from the crosses Cibc-17 × Gr-3, Knox-10 × Gr-3, Knox-10 × Wei-0, Lz0 × En-1, Lz-0 × Hs-0 were produced. All parental lines were obtained from the ABRC from the following stock: CS22615 (Lz-0), CS8068 (Berkeley), CS22566 (Knox-10), CS1202 (Gr-3), CS1136 (En1), CS22622 (Wei-0), CS22655 (Hs-0) and CS22603 (Cibc-17). Plant DNA extraction and MassARRAY genotyping A single cauline leaf was harvested from each plant for genotyping (RIL and F2 populations). The tissue was frozen in liquid nitrogen and ground in a Mini BeadBeater96 (BioSpec Products, Inc., Bartlesville, OK, USA) in 2-mL screw-cap tubes containing around ten 1.0-mm zirconia/ silica beads (BioSpec Products, Bartlesville, OK, USA). The samples were shaken two times for 30 s. The powder was then processed using a DNAeasy 96 Plant Kit (Qiagen, Venlo, Netherlands) following the manufacturer’s instructions. DNA concentration was measured using a Nanodrop (Thermo Fisher Scientific, Waltham, MA, USA) and adjusted at 10 ng μL−1. Genotyping was performed using a previously described 149 markers assay (Platt et al. 2010) using the iPlex MassARRAY technology (Sequenom, San Diego, CA, USA). Flowering time experiments Seeds were surface sterilized, plated on ½ MS agar medium (MS salts 2.1 g L−1, sucrose 10 g L−1, agar 8 g L−1, pH 5.8)

Mol Genet Genomics

and vernalized for 48 h at 4 °C. Seedlings were transferred to Professional Pro-Mix ‘BX’/Mycorise Pro (Premier Horticulture Lté, Rivière du Loup, Canada) supplemented with 0.3 % (v/v) Nutricote 14-14-14 Type 100 fertilizer (ChissoAsahi Fertilizer C. Ltd., Tokyo, Japan) after 10 days of continuous light. For flowering time experiments, plants were grown under a 16/8 h day/night cycle at 21 °C day and 18 °C night temperature (long day) or under a 8/16 h day/night cycle at 21 °C day and 18 °C night temperature. Light was maintained at 200 μmol s−2 m−2. Flowering was assessed by counting the number of visible rosette leaves at the time of bolting, and five individuals were used for each genotype, whenever possible.

Fiber quality analyzer measurements Fiber length was determined using a fiber quality analyzer (FQA) (OpTest, Hawkesbury, ON, Canada) with a cytometric flow cell and image analysis system capable of rapidly and accurately measuring fiber curl, kink, and length distributions (Olson et al. 1995; Roberston et al. 1999), as described in Capron et al. (2013). Five milliliters of the fiber suspension was dispensed into a 600-mL plastic beaker for FQA analysis. This sample was then automatically diluted to exactly 600 mL by the FQA. The fiber input was adjusted to a targeted an EPS (events per second) measurement range of 25–40 fibers per second. Data analysis

Production of inflorescence stems for fiber length measurements Plants were grown in conditions identical to the long-day conditions described in the previous paragraph. Post-flowering inflorescence stems were collected for fiber quality analyzer (FQA) processing and two plants per lines were bulked for processing. Isolation of fibers from plant stems Fibers were isolated mainly as described in Capron et al. (2013). Samples (10 mg) of air-dried stem material taken from the bottom 5 cm of one stem or each of two stems were placed in a 20-mL test tube for the pulping reaction. The stem sample was compacted using a glass rod before adding 2 mL each of distilled water and acetic acid (glacial) and heating in a boiling water bath for 2 min. After adding 2 mL 30 % hydrogen peroxide, the samples were returned to the boiling water bath for 90 min. The resulting solution was then carefully decanted to retain the cooked stem tissue within the tube and the tissue was rinsed gently with distilled water three times to remove residual reagents. The delignified stem tissue was transferred to a screw-cap conical plastic centrifuge tube with 35 mL of distilled water and then agitated vigorously to disintegrate fiber bundles and form a homogenous fiber suspension. The suspension was subsequently filtered through a Britt Dynamic Drainage Jar (DDJ) using 3 L of water, stirred by an overhead stirrer at 200 rpm (TAPPI 1992) to collect fibers retained on the 200mesh screen (105 μm opening). The retained fibers were rinsed off the filter mesh, collected in a 50-mL centrifugal tube and diluted to a total volume of 50 mL with distilled water. The fiber suspensions were visually inspected and fiber bundles, if present, were removed manually using a fine needle. Triplicate measurements were performed on 10 mg samples taken from individual stems of the same line.

We used the genotyping data from five F2 populations and from the Lz-0xBerkeley RILs to generate a composite linkage map using JoinMap 3.0 (Van Ooijen and Voorrips 2001). Statistical analysis was carried out using the R statistical language (R Core Team 2009). Heritability was estimated using analysis of variance component using the lme4 package. Variance components for the lines and for the experiments were calculated with the model:

lmer(TRAIT ∼ ( 1|LINE) + ( 1|EXPERIMENT)) and used to calculate the heritability with the formula h2 = Va/[Va + Ve]:

 h2 = Var(lines) [Var(lines) + Var(Residual)] Composite interval mapping was performed with the R/qtl (Broman et al. 2003) analysis package. The imputation method was selected, with 256 draws, error 0.05 and 2.5 cM steps. The composite interval mapping was done with the imputation method, with three covariates and a 10 cM window: cim(cross, method = “imp”, window = 10, n.marcovar = 3)

The significant thresholds were determined by performing 1,000 permutations. The multiple QTL model was devised using the fitqtl function. Results obtained from the composite interval mapping analysis were used to construct the QTLs and their positions were used in a simple additive model. The resulting model was used to assess the effect of each QTL on the trait (percentage of explanation of the observed total variance). Bayesian interval mapping (BIM) was performed using the R/qtlbim package (Yandell et al. 2007). The data from both short- and long-day experiments were pooled and the photoperiod was used as fixed effect and an interactive covariate. G×E was allowed and epistasis disabled in the model.

13



Results RILs generation, genotyping and genetic map construction From a series of crosses between Arabidopsis accessions generated for a previous study (Beatson et al., in preparation), a series of F2 plants initially produced from the ecotypes Berkeley and Lz-0 was selected to generate a new RIL population. The choice was determined by the relative uniqueness of the Lz-0 accessions. In effect, this accession was found to form its own cluster in the milestone study from Platt et al. (2010). Furthermore, a simple phylogenetic tree generated from the data used in the same study places Lz-0 in a different clade than the usual workhorses of Arabidopsis genetics: Col, Ler, Ws… (Figure S1). A series of 40 F2 plants from this cross was expanded to 200 F3 and those plants were then taken to F8 by selfing. A total of 159 F8 plants were recovered from the initial lot of 200 and used for further experiments. F8 progeny from 159 Lz-0 × Berkeley RILs, along with the F2 progeny from the several crosses (Cibc-17  ×  Gr3, Knox-10  ×  Gr-3, Knox-10  ×  Wei-0, Lz-0  ×  En-1, Lz0 × Hs-0), was genotyped using a 149-marker iPlex MassARRAY assay (Platt et al. 2010). In the Lz-0xBerkeley population, 73 out of the 149 markers were found to show no polymorphism and 8 failed to produce any data. Residual heterozygosity in the 68 remaining markers was 3.07 %, larger than the expected 0.8 %, indicating a possible selection for heterozygosity. Individual genotyping failure rate for those 68 markers was 6.27 %. There were two segments on chromosomes 1 and 5 showing significant deviation from the expected 1:1 ratio: between markers AtMSQTSNP48 and ATMSQTSNP53 (p