The Ectopic Expression of Btr2 in Aegilops tauschii Switches the Disarticulation Layer From Above to Below the Rachis Node

Seed dispersal among wild species belonging to the tribe Triticeae is typically achieved by the formation of a brittle rachis. The trait relies on the development of a disarticulation layer, most frequently above the rachis node (resulting in wedge type dispersal units), but in some species below the rachis node (resulting in barrel type dispersal units). The genes responsible for the former type are the complementary pair Btr1 and Btr2, while the genetic basis of the latter type has yet to be determined. Aegilops tauschii forms barrel type dispersal units and previous study showed this species lacked an intact copy of Btr1. Here it has been demonstrated that Ae. tauschii carries two of Btr2; and that Btr2 transcript is present in a region below the rachis node where the abscission zone forms. The implication is that in this species, the Btr2 product is involved in the formation of barrel type dispersal units.


INTRODUCTION
Many wild grass species ensure the dispersal of their seed by their inflorescence becoming brittle at physiological maturity; the trait requires the formation of one or multiple abscission zones (Cousens et al., 2008). The anatomy and positioning of the abscission zone varies from species to species, although not in a way which necessarily correlates with phylogeny (Yu et al., 2020b). Among species belonging to the Triticeae tribe, the zone forms either in the rachis or the rachilla (Sakuma et al., 2011). Rachis disarticulation most frequently occurs above the node, but in a few cases (as in, for example, Aegilops tauschii) it can occur below the node (Zohary et al., 2012).
Disarticulation above the node results in the formation of wedge type dispersal units, each consisting of a single spikelet attached to a short proximal section of the rachis. The presence on chromosome arm 3HS of the dominant, complementary and tightly linked genes Btr1 and Btr2 is necessary for the expression of the trait in wild forms of Hordeum vulgare (barley; Pourkheirandish et al., 2015). In Triticum and Aegilops species, Btr1 and Btr2 map to the short arm of the group 3 homeolog (Avni et al., 2017;Zeng et al., 2020b). The products of this pair of genes are presumed to determine the unique mode of disarticulation exhibited by the Aegilops longissima spike, in which only one or two of the central nodes are brittle (Zeng et al., 2020a). An alteration in the Btr1 coding region sequence is thought to be responsible for the non-brittle nature of the spikes formed by both einkorn (Triticum monococcum; Pourkheirandish et al., 2018;Zhao et al., 2019) and emmer (Triticum dicoccum; Avni et al., 2017) wheat. The transcription of Btr2 above the rachis node has been suggested as a critical determinant of the formation of a disarticulation zone in both barley (Pourkheirandish et al., 2015) and Ae. longissima (Zeng et al., 2020a).
The barrel type dispersal unit consists of a single spikelet attached to a short distal rachis segment; among Aegilops species, this form is only found in species harboring the D genome (van Slageren, 1994). According to Chen (2001), F 1 hybrids formed by crossing a weedy form of Triticum aestivum forming wedge type dispersal units with Triticum spelta (barrel type units) disarticulate both above and below the rachis node, suggesting that different gene(s) underlie the formation of the two forms of dispersal unit. The genetic basis of the barrel type unit was found to be a locus mapping to chromosome arm 3DL (Li and Gill, 2006;Zhang et al., 2015). The locus harbored a homolog of rice gene qSH1, TaqSH1-D (Katkout et al., 2015), which encodes a BELL1type homeobox protein implicated in the shattering phenotype (Konishi et al., 2006). An insertion present in the TaqSH1-D 3'-UTR has been proposed to be the genetic basis of rachis non-brittleness, although this hypothesis awaits verification. As yet, the genetic basis of the barrel type dispersal unit has yet to be identified.
The geographical range of the diploid species Ae. tauschii, the donor of the D genome harbored by bread wheat (T. aestivum), stretches across western and central Asia into south eastern Europe (van Slageren, 1994). Two phylogenetic lineages (L1 and L2) of the species have been recognized (Wang et al., 2013): the former includes accessions of ssp. tauschii, while the latter includes both ssp. tauschii and ssp. strangulata accessions. As it has been suggested that the L2 type is the one more closely related to the bread wheat D genome (Wang et al., 2013), and L2 representative has been selected for whole genome sequencing (Jia et al., 2013;Luo et al., 2017).
In both rice and sorghum, the products of multiple genes are known to determine seed shattering (Yu and Kellogg, 2018;Di Vittori et al., 2019). A comparison of the Btr1 and Btr2 sequences harbored by a range of grass species has shown that while Ae. tauschii lacks an intact copy of Btr1, two copies of Btr2 are present, both of which map along chromosome arm 3DS (Zeng et al., 2020b). Since Btr1 and Btr2 act as a complementary gene pair in most Triticeae species, Ae. tauschii represents an anomaly, raising the question as to whether the products of the two Btr2 genes underly the formation of the species' barrel-type disarticulation units. The present experiments sought to establish the involvement, if any, of the Btr2 product in the formation of barrel-type units by Ae. tauschii.

Plant Materials
Accessions of each of Ae. tauschii ssp. tauschii (AE 956) and Ae. longissima ssp. longissima (AE 417) were obtained from the IPK Genebank. 1 Plants were grown in a glasshouse at the National Institute of Crop Science (Tsukuba, Japan).

Histological and Anatomical Analysis
Spikelets were sectioned longitudinally by hand and stained by immersion for 3 h in 0.01% w/v acridine orange in distilled water, then rinsed in 10x PBS buffer (pH7; Acridine orange stains lignin green and anionic polysaccharides red, see Li and Reeve (2005) and Houtman et al. (2016). The sections were illuminated with a 488 nm diode laser and the output was captured using a confocal laser scanning microscope (LSM 700, Carl Zeiss, Tokyo, Japan) equipped with ZEN 2009 Light Edition CLSM software, scanning at 505-530 nm (green) and >600 nm (red). The surface of the rachis separation layer in mature dispersal units was characterized using a TM3000 scanning electron microscope (SEM; Hitachi, Tokyo, Japan), set to deliver an accelerating voltage of 1 kV.

DNA Sequence Analysis
Genomic DNA was extracted from fresh leaves following Komatsuda et al. (1998). The subsequent PCR amplification and DNA sequencing procedures followed those given by Pourkheirandish et al. (2015). CLC Sequences Viewer v8 software was used to align sequences. 2

Quantitative RT-PCR
Three replicates, each consisting of three 0.5-1.5 cm long spikes at the white anther stage, were sampled (Kirby and Appleyard, 1981), and RNA was extracted from these using the TRIzol reagent (Invitrogen, Carlsbad, CA, United States). After treatment with RNase-free DNase (Takara Bio, Kusatsu, Japan) and quantification using a NanoDrop 1,000 device (Thermo Scientific, Tokyo, Japan), a 1 μg aliquot was reverse-transcribed using SuperScript III (Invitrogen), and the resulting ss cDNA used as template for quantitative PCRs (qPCRs) based on the Thunderbird SYBR qPCR Mix kit (TOYOBO, Tokyo, Japan) and a CFX96 real-time PCR system (Bio-Rad, Tokyo, Japan). Each sample was represented by a minimum of three technical replicates.

RNA in situ Hybridization
Both an anti-sense and sense version of a segment of the Btr2 sequences were generated by PCR from Ae. longissima accession AE 417 (Zeng et al., 2020a). The PCR amplicons were purified using a QIAquick PCR purification kit (Qiagen, Germantown, MD, United States). One of the probes for Btr2-Lo-1 (identical to Btr2-D-1) covered 150 bp of the Btr2 coding region and 200 bp of its 3'-UTR, while the other one for Btr2-Lo-2 (identical to Btr2-D-2) covered 150 bp of the Btr2 coding region and 150 bp of its 3'-UTR. Adaptor sequences including promoter of T7 RNA polymerases (CGCGCGTAATACGACTCACTATAGGG) were added to the 5' end of the primers. The amplicons were validated by DNA sequencing before use. Hybridization probes 1 www.ipk-gatersleben.de/en/genebank/ 2 www.clcbio.com/ Frontiers in Plant Science | www.frontiersin.org were prepared using T7 RNA polymerase. The RNA in situ hybridization procedure, based on digoxigenin-labeled RNA probes, was performed on samples harvested at the white anther stage, following methods given by Komatsuda et al. (2007).

Disarticulation in the Mature Spikes of Aegilops tauschii and Aegilops longissima
The mature Ae. tauschii spike was brittle (Figure 1A), disarticulating below each rachis node, thereby generating barrel type dispersal units ( Figure 1B). In contrast, disarticulation in the Ae. longissima spike occurred above two nodes in the central portion of the rachis, resulting in the production of wedge type units (Figures 1C,D). The surface of the separation layer in the Ae. tauschii spike was smooth, as a result of all but the non-vascular cells having been capped by an intact cell wall (Figures 2A,C). In Ae. longissima, by way of contrast, although the surface of separation layer was also smooth, few cells were capped by a cell wall (Figures 2B,D). A typical LSM image of a stained longitudinal section of Ae. tauschii is shown in Figure 3A. No autofluorescence was observed in either the green or the red channel (data not shown), but following staining with acridine orange, there was evidence of lignification in cells lying below the rachis node ( Figure 3A). In particular, lignin accumulated in about four layers of small cells within the abscission zone (Supplementary Figure S1A), while the accumulation of anionic polysaccharides in this zone was relatively limited (Supplementary Figure S1C). The lignin was preferentially deposited in the secondary wall ( Figure 3B). No such cells were observed in the Ae. longissima disarticulation layer (Figures 3C,D), even in a higher magnification (Supplementary Figure S2), implying the absence of a distinct abscission zone.

The Profile of Btr2 Transcription in Aegilops tauschii
The primer sequences employed for qPCR were based on those targeting the Btr2 locus present (Supplementary Figure S3)    and barley Actin (GenBank accession AK362208) was used as the reference sequence. According to a qPCR assay, transcript of both Btr2-D-1 and -D-2 was present in spikes sampled at the white anther stage (Figure 5). When RNA in situ hybridization was employed using an antisense version of the Btr2-D-1 sequence as probe, signal was observed below the rachis node and spikelets, and also just above the rachis node ( Figure 6A); equivalent signals were not detected when the sense probe was employed ( Figure 6B). The same experiment using the Btr2-D-2 sequence as probe produced the same overall result (Figures 6C,D).

The Abscission Zone in Aegilops tauschii Forms Below the Rachis Node
Aegilops tauschii produces barrel type dispersal units by disarticulating below the rachis node. The point of breakage is ensured by the formation of an abscission zone in which an array of lignified cells is formed. A similar structure has been described in the spikes of both Elymus sibiricus (Zhao et al., 2017) and Brachypodium distachyon (Yu et al., 2020a), whereas the abscission zone cells in the panicles of both wild rice (Oryza nivara;  and wild sorghum (Sorghum virgatum; Lin et al., 2012) are not lignified. In wild Hordeum spp., by contrast, which form wedge type units, disarticulation is achieved by the formation of a layer of thin-walled cells in the separation zone (Pourkheirandish et al., 2015). As suggested by (Zhao et al., 2019), the mechanistic basis of rachis brittleness in the genus Triticum lies in repressing the synthesis of cell walls. The disarticulation surface in these species is characterized by cells not capped by a cell wall. In Ae. tauschii, the separation surface was similarly smooth, but, with the exception of vascular elements, the cells at the surface were capped by an undamaged cell wall. This phenotype resembles that of the separation layer seen in the rachilla of Elymus sibiricus (Zhao et al., 2017) and at the base of the pedicel in Oryza spp. (Konishi et al., 2006).  The co-expression of Btr1 and Btr2 is required for disarticulation to occur above the rachis node (Pourkheirandish et al., 2015;Zeng et al., 2020a). The expression of Btr2 above the rachis node determines where the disarticulation layer forms. Btr2 is not expressed below the rachis node in species which form wedge type disarticulation units (Pourkheirandish et al., 2015;Zeng et al., 2020a). The genome sequence of Ae. tauschii ssp. strangulata (Luo et al., 2017) lacks a copy of Btr1, and the assumption is that this is similarly the case in Ae. tauschii ssp. tauschii. The reason why Ae. tauschii fails to form a disarticulation layer above the rachis node is presumed to be the absence of an intact copy of Btr1, while its formation of a disarticulation layer below the rachis node reflects the operation of a cell-cell separation mechanism, as demonstrated here. The RNA in situ hybridization experiment confirmed that there was an abundance of Btr2 transcript in the abscission zone below the rachis node, which implies that the ectopic expression of Btr2 is required for the formation of this structure.
The genome of Ae. longissima (which does not form barrel type disarticulation units) harbors two copies of Btr2, neither of which is transcribed below the rachis node (Zeng et al., 2020a). The implication is that the unusual site of Btr2 transcription shown by Ae. tauschii is a recently acquired trait. An intriguing question is how this product can operate, given that its function is to ensure that disarticulation occurs below the rachis node. A further possibility is the formation of the abscission zone requires not just the product of Btr2 but also that of a gene(s) such as the ortholog of the rice locus qSH1 to achieve cell-cell separation in the abscission zone (Konishi et al., 2006;Katkout et al., 2015). Abscission zone is formed above glumes in Brachypodium, above the rudimentary glumes and below the empty glumes (sterile lemmas) in rice (Yoshida and Nagato, 2011;Zhou et al., 2012), and below glumes in Setaria (Yu et al., 2020a). Though they share distinct abscission zone, qSH1 is transcribed in each site corresponded to the abscission zone (Yu et al., 2020a). Thus, qSH1 appears to be a strong candidate to collaborate with Btr2 to form the abscission zone in Ae. tauschii.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
XZ and TK planned and designed the research and drafted the manuscript. XZ performed the morphological analysis, the DNA sequencing and the qPCRs, while AT and XZ performed the RNA in situ hybridization experiments. HS, SK, and TK monitored the acquisition of data. Each of the authors has read and agreed to the submitted version of the manuscript.

FUNDING
This research was financially supported in part by a grant-in-aid from the Japan Society for the Promotion of Science (JSPS) to TK (18H02184). XZ is grateful for the award of a Monbukagakusho scholarship given by from Government of Japan.

ACKNOWLEDGMENTS
We thank F. Blattner (IPK, Gatersleben, Germany) for the gift of Ae. tauschii and Ae. longissima germplasm; T. Koba, T. Oikawa, K. Mishina, and M. Ohta for advice on the conduct of the experiments; and T. Kawakatsu for the use of SEM facilities.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpls.2020.582622/ full#supplementary-material Signal was detected between the spikelet and the rachis node (black arrows) and above the rachis node (white arrows) in (A) and (C), but not in (B) or (D). sp, spikelet; gl, glume; ra, rachis.
Frontiers in Plant Science | www.frontiersin.org