The ATP-Binding Cassette Transporter ABCB19 Regulates Postembryonic Organ Separation in Arabidopsis

The phytohormone auxin plays a critical role in plant development, including embryogenesis, organogenesis, tropism, apical dominance and in cell growth, division, and expansion. In these processes, the concentration gradient of auxin, which is established by polar auxin transport mediated by PIN-FORMED (PIN) proteins and several ATP-binding cassette/multi-drug resistance/P-glycoprotein (ABCB/MDR/PGP) transporters, is a crucial signal. Here, we characterized the function of ABCB19 in the control of Arabidopsis organ boundary development. We identified a new abcb19 allele, abcb19-5, which showed stem-cauline leaf and stem-pedicel fusion defects. By virtue of the DII-VENUS marker, the auxin level was found to be increased at the organ boundary region in the inflorescence apex. The expression of CUP-SHAPED COTYLEDON2 (CUC2) was decreased, while no obvious change in the expression of CUC3 was observed, in abcb19. In addition, the fusion defects were greatly enhanced in cuc3 abcb19-5, which was reminiscent of cuc2 cuc3. We also found that some other organ boundary genes, such as LOF1/2 were down-regulated in abcb19. Together, these results reveal a new aspect of auxin transporter ABCB19 function, which is largely dependent on the positive regulation of organ boundary genes CUC2 and LOFs at the postembryonic organ boundary.


Introduction
Throughout the lifespan of most higher plants, new organs are initiated continuously from pluripotent cells in the shoot apical meristem. This essential process is associated with the establishment of boundaries separating the newly formed organs from adjacent tissues [1]. Such boundaries are composed of a specialized group of saddle-shaped cells that are morphologically different from the adjacent cells [2]. The unique shape of these cells is attributed to elongation along the organ boundary, contraction along the axis perpendicular to the boundary, and cell division leading to a new cell wall parallel to the boundary [2][3][4]. These boundaries emerge at the early stage of primordia initiation, and their positions are determined by signals from the central region of the meristem [1,2,5]. The boundaries act as a barrier to separate and maintain different cell types [1], and, when localized at the base of leaves, they have the potential to produce axillary meristems, which contribute greatly to the overall architecture of plants [6].
Several lines of evidence show that auxin plays a significant role in organ patterning and boundary establishment by controlling CUC gene expression [18][19][20][21]. Mutations in the putative auxin efflux carrier PIN1 produce naked inflorescence stems resulting from the ectopic expression of CUC2 at a ring-like domain characterized by primordia-specific gene expression [18]. PINOID (PID) and ENHANCER OF PINOID (ENP) regulate PIN1 localization and function to promote cotyledon initiation bilaterally by preventing CUC1, CUC2, and STM from expanding to the primordia during embryonic development [20,21]. MONO-PTEROS (MP)/AUXIN RESPONSE FACTOR5 (ARF5), a transcriptional activator of auxin signaling, participates with PIN1 in cotyledon separation through the partial regulation of CUC2 [19]. Together, these results suggest a relationship among auxin, auxin transporters, and the regulation of CUC expression in the process of organ patterning and boundary establishment [2].
The auxin concentration gradient is a crucial signal during plant development that is established by polar auxin transport [22,23]. Two protein families, PIN-FORMED (PINs) efflux carriers and ATP-binding cassette/multi-drug resistance/P-glycoprotein (ABCB/MDR/PGP) transporters, are involved in auxin efflux [23][24][25][26]. PIN encodes a 67-kilodalton protein with similarity to bacterial and eukaryotic carrier proteins [23]. There are eight members of the PIN family in the Arabidopsis genome [22]. As described above, pin-formed1 (pin1) was first characterized by needle-like inflorescence stems [23]. pin1 also exhibits defects in vascular patterning, organogenesis, and phyllotaxis [23,27,28]. Physiological studies performed to date in planta and/or heterologous systems have demonstrated that at least five PINs act as a rate-limiting step in cellular auxin efflux. And consistent with their role as auxin polar transporters, some of the PIN proteins display polar localization, especially in embryonic development and organogenesis, although some are distributed without prominent polarity in certain tissues (for a review, see [22]).
Here, we identified a new allele of abcb19, named abcb19-5, which shows organ fusion defects in addition to the phenotypes already described for abcb19 [24]. CUC2 was down-regulated in abcb19 and cuc3 greatly enhanced the organ fusion phenotype of abcb19-5, reminiscent of the cuc2 cuc3. Further more, some other organ boundary genes were also down-regulated in abcb19. Our results reveal a new function for the auxin transporter ABCB19.

Results
ABCB19 is necessary for organ separation at stem-cauline leaf and -pedicel junctions in Arabidopsis To characterize novel components in flowering time control, we screened a T-DNA insertion mutant library and identified a mutant with a delay in the transition to flowering ( Figure 1A). In addition, the mutant exhibited epinastic cotyledons and wavy roots and hypocotyls at the seedling stage ( Figure 1B-E). Furthermore, organ fusion defects occurred at both stem-cauline leaf junctions (the abnormal growth of the proximal part of the cauline leaf fused with the stem) ( Figure 1F and G) and stem-pedicel junctions ( Figure 1H and I). This fusion, which was seen on the primary and secondary branches, was most obvious on rosette branches. The stem-cauline leaf fusions caused bending of the stem ( Figure 1G). Stem-pedicel fusions were more obvious for the first several siliques, and, as a result, the angle between the stem and pedicel was significantly reduced for the first eight siliques ( Figure 1J).
TAIL-PCR was performed to obtain the flanking sequence of the T-DNA left border. Sequencing of the PCR products showed that the adjacent gene was the auxin transporter ABCB19/PGP19/ MDR1, and that the insertion was in the last exon ( Figure 2A). No full-length transcript was detected for this new allele, which was named abcb19-5; however, a partial transcript was present ( Figure 2A).
To confirm that the mutation of abcb19 was responsible for these developmental defects, the T-DNA insertion allele abcb19-3 (mdr1-3) was obtained [24]. We found that abcb19-3 behaved very similar to abcb19-5; abcb19-3 showed the same fusion phenotype ( Figure 2B, C, and D) in addition to the other phenotypes described above (data not shown). F1 plants produced by crossing abcb19-5 with abcb19-3 also behaved like both of the parents ( Figure 2B, C, and D). Moreover, these defects, including epinastic cotyledons, rosette leaf shape, and stem-cauline leaf and -pedicel fusion defects, were successfully rescued by the transformation of ABCB19 into abcb19-5 ( Figure 2E-H). These results demonstrate that abcb19-5 is a new allele of abcb19. Given that there is still no study about ABCB19 in organ separation, we focused on the organ fusion phenotype of the mutant.

Auxin distribution is altered by ABCB19 mutation
It was reported that ABCB19 is required for the basipetal auxin transport out of the shoot apex of seedling and inflorescence [24], and that loss of ABCB19 function increased auxin retention in the apical tissues of seedling by quantification of endogenous IAA levels and radiotracer studies [42]. Due to the organ separation defects of abcb19, we are curious about the endogenous auxin level at the organ boundary region in abcb19. However, as a result of the auxin distribution gradient, with levels being highest in the primordia and lowest in the organ boundaries [1], it is difficult to analyze the alteration of auxin levels at the site of organ fusion using the auxin-responsive marker DR5::GUS/GFP. Fortunately, the DII-VENUS (termed domain II fusion with fast maturating variant of YFP, VENUS) marker is more sensitive than DR5::GUS/GFP, images of which are like a photographic negative of auxin levels [43][44][45]. And there are strong signals at the organ boundary at the inflorescence apical region [43].
Consequently, the DII-VENUS marker was introduced into abcb19-5. We found that the overall fluorescence signal was dropped in abcb19 compared with wild type plants at the inflorescence apex including the inflorescence meristem (IM) and organ boundary region ( Figure 3). As a negative indicator of auxin, the reduction of DII-VENUS indicates that the auxin level is increased at the inflorescence apex in abcb19, consistent with the abnormal basipetal auxin transport activity in abcb19. Thus, by means of DII-VENUS, we show that the endogenous auxin level is increased both in the organ boundary region and in inflorescence meristem ( Figure 3).

CUC2 and CUC3 expression is differentially regulated in abcb19
Among the genes that function in postembryonic organ boundary separation, CUC2 and CUC3 of the NAC family are two well-known, important regulators [8]. And it was indicated that the expression of CUC2 are inhibited by auxin [18,19]. To examine the expression of these genes in wild-type and abcb19 plants, CUC2::GUS and CUC3::GUS [46] were crossed into abcb19-5, respectively.
The expression patterns of CUC2::GUS and CUC3::GUS were analyzed at different developmental stages in wild-type and abcb19 plants. In wild-type, CUC2::GUS and CUC3::GUS activity was detected at the organ boundary in cotyledons, stem-cauline leaf junctions, and at the boundary of stem-pedicel junctions, consistent with previous in situ results [8]. In abcb19-5, the CUC2::GUS was down-regulated after the plants undergoing the floral transition and after bolting ( Figure 4A-B). Furthermore, CUC2::GUS activity was reduced by about 37% in both the stemcauline leaf junctions and inflorescences of abcb19 plants according to our b-glucuronidase assay results ( Figure 4C-D). The CUC2::-GUS activity showed similar down-regulation pattern in another allele, abcb19-3 ( Figure 4E-F). However, under the same conditions, the expression of CUC3 as shown by histological staining and a b-glucuronidase assay was not obviously changed in our experiment ( Figure 4H, I, and J). The decrease in the CUC2 expression in abcb19 was further confirmed in both abcb19 alleles by determining the level of CUC2 mRNA using the real-time quantitative-PCR (q-PCR) ( Figure 4G). Thus, mutations in ABCB19 may specifically affect CUC2 expression by increased auxin level in the organ boundary region, with no or little effect on CUC3 expression, during postembryonic growth.
The genetic relationship between ABCB19 and CUC2 or CUC3 in Arabidopsis CUC2, but not CUC3, expression was obviously reduced in abcb19-5. Given this, we hypothesized that the organ separation defects in abcb19 cuc3 would be enhanced compared to those in abcb19, while the elimination of cuc2 would not be as efficient as the elimination of cuc3 in terms of phenotype enhancement.
To test this hypothesis, we generated abcb19-5 cuc2-3 and abcb19 cuc3-105 plants. Consistent with our expectations, cuc3-105 enhanced the fusion defects seen in abcb19 dramatically, while cuc2-3 contributed to the observed defects to a lesser extent ( Figure 5A-C). The extent of fusion was greatly enhanced at stemcauline leaf junctions and inflorescence stem-pedicel junctions in abcb19-5 cuc3 compared with abcb19-5 ( Figure 5A and B) and fusion of the axillary shoot to the main stem was observed in abcb19 cuc3,showing the phenotype equivalence between abcb19 cuc3 and cuc2 cuc3 ( Figure 5A, shown by an white arrow) [8]; however, the degree of fusion was still less than that seen in cuc2 cuc3, due to the residual expression of CUC2 in abcb19-5. cuc2 enhanced the fusion defects in abcb19-5 slightly and less effectively than cuc3 (Figure 5A and B). We next determined the frequency (%) of fusion defects at stem-cauline leaf junctions. In primary stem-cauline leaf junctions, the number of fusion events in abcb19-5 cuc3-105 was significantly increased compared with abcb19-5 ( Figure 5C); in comparison, the number of fusion events in abcb19-5 cuc2 was not significantly different from abcb19-5 ( Figure 5C). The rate of fusion in abcb19 cuc3 was even higher than that in cuc2 cuc3, however, the difference was not significant shown by the t-test ( Figure 5C).
In general, the lesion of cuc3 significantly reinforced the fusion defects in abcb19 in terms of the degree and frequency of fusions, while cuc2 contributed less. This is largely consistent with the reduced expression of CUC2 (but not of CUC3) in abcb19 (Figure 4).

Other organ boundary-specific genes besides CUC2 may be involved in ABCB19-mediated organ separation
As CUC2 and CUC3 participate redundantly in postembryonic organ separation, each single mutant shows no obvious fusion defect [8]; thus, only reduction in CUC2 in abcb19 does not account for the organ fusion phenotype observed. Since ABCB19 acts as an auxin transporter, the auxin distribution pattern in abcb19 is altered obviously (Figure 3). Auxin is such an important regulator of plant development that a number of factors may be changed to varying degrees at the organ boundaries in abcb19. Variations in these factors together with the down-regulation of CUC2 may contribute to the fusion defect observed in abcb19.
We tested a number of organ boundary-specific factors in abcb19 by semi-quantitative RT-PCR, and observed that BOP was elevated in abcb19 (the elevated BOP expression is similar to the situation in lof1 [6]); LOF1 was reduced slightly and LOF2 was down-regulated obviously; LAS and RAX1 were not distinguishable from the wild type plants ( Figure 6). Since it has been shown that the lof1 knock-out considerably enhances the cuc2 phenotype [6], the down-regulation of the two LOFs in abcb19 might at least to some extent explain why the cuc2 phenotype does not match the abcb19 phenotype.
Therefore, these results demonstrate that ABCB19, as an auxin transporter, control a variety of organ boundary genes to guarantee the establishment of the organ boundary.

ETT may function in postembryonic organ separation
Auxin functions mainly through AUXIN RESPONSE FACTORs (ARFs). ETTIN (ETT)/ARF3 are reportedly involved in flower development [47], adaxial-abaxial patterning during leaf development [48], and in the vegetative phase change as the target of trans-acting (ta) siRNA-ARFs (tasiR-ARF) [49]. We observed that ett-3 showed moderate cauline stem-cauline leaf fusion defects ( Figure 7A). When we combined ett-3 with abcb19-5, the extent of fusion was dramatically enhanced ( Figure 7A). The rate of  Figure 7B). This suggests that ABCB19 participates in a pathway parallel with ETT to control postembryonic organ separation.

Discussion
ABCB19 participates in postembryonic organ separation in Arabidopsis ABCB19, as an auxin transporter [24,29,31,32,39], has been implicated in a multitude of biological processes, including normal growth and development in multiple tissues [24,39], photomorphogenesis [32,40], and gravitropic responses [29,41]. In this study, we generated several lines of evidence showing the novel function of ABCB19 in postembryonic organ separation based on a mutant identified from our genetic screen. The similar organ separation defects in two alleles of abcb19 and the appearance of the same defect in F1 plants from a cross between abcb19-3/mdr1-3 and abcb19-5, as well as transgenic complementation (Figure 1 and Figure 2), all demonstrate the role of ABCB19 in organ separation control.
When ABCB19 is knocked out, the auxin concentration is increased in the boundary region, as is shown by the newly developed DII-VENUS marker (Figure 3). This may result in abnormal cell growth and then the organ fusion defects. We also found that AUXIN RESPONSE FACTOR-ARF3/ETT is involved in postembryonic organ separation (Figure 7), and that ABCB19 may participate in a pathway parallel with ETT to control postembryonic organ boundary formation.
ABCB19 plays a role in organ separation by partially regulating CUC2 and some other organ boundary genes Previous studies have indicated that auxin plays a critical role in organ boundary establishment by controlling CUC gene expression [18][19][20][21]. CUC2 and CUC3 play redundant roles in postembryonic organ separation [8]. CUC2 has been frequently reported to be repressed by high auxin concentrations [18,19]. Notably, we found that the expression of CUC2 was obviously down-regulated at the postembryonic boundary in abcb19 compared with wild-type ( Figure 4A-G). In contrast, CUC3 expression was not obviously changed ( Figure 4H-J), indicating the differential regulation of these homologs at the transcriptional level by ABCB19 through the control of auxin distribution. Consistently, it was cuc3 rather than cuc2 that enhanced the fusion defects in abcb19 significantly ( Figure 5). Besides CUC2, we also found that the some other organ boundary genes, such as LOF1, LOF2, and BOP, were also shown altered expression in abcb19 ( Figure 6). Together, our gene expression and genetic results indicate that ABCB19 may promote postembryonic organ separation via the regulation of CUC and other organ boundary genes, probably through the depletion of auxin at the boundary.  In summary, we demonstrated that the auxin efflux carrier ABCB19 participates in postembryonic organ boundary specification by partially regulating the NAC family transcription factor CUC2 and some other organ boundary genes.

Plant Materials and Growth Conditions
The Arabidopsis thaliana plants used in this work were all in the Columbia-0 (Col-0) background. abcb19-3 (mdr1-3) was kindly provided by Dr. Edgar P. Spalding; abcb19-5, which carries a T-DNA insertion [50], was cloned by TAIL-PCR. abcb19-5 was crossed with CUC::GUSs to produce abcb19-5 CUC::GUSs. In the F2 generation, plants homologous for abcb19-5 that carried CUC::GUS were identified by PCR. In the next generation, thirty seedlings of several different lines were analyzed by GUS staining to identify lines homologous for CUC::GUS. abcb19-5 cuc2-3, which exhibited an abcb19-specific leaf shape and smooth leaf margin (cuc2-3 phenotype), was first identified by leaf appearance and then by PCR analysis. abcb19-5 cuc3-105 was characterized by PCR analysis. abcb19-5 ett-3 was identified by abnormal carpel development (ett-3) and the PCR analysis of abcb19-5.
Seeds were sterilized in 75% ethanol for 1 min, washed three times with sterile water, kept at 4uC for 2 days to promote germination, and then grown on Murashige and Skoog medium.

Plasmid Construction and Plant Transformation
The full-length CDS of ABCB19 was amplified from Arabidopsis cDNA reverse-transcribed from total seedling RNA using the following primers: ABCB19-c-F (59-CGGGATCCATGTCG-GAAACTAACACAACC- 39) and ABCB19-c-R (59-GGGGTACCTCAAATCCTATGTGTTTGAAGC-39). After sequencing, the ABCB19 CDS was cleaved with BamHI and KpnI and ligated to the pCAMBIA1300 binary vector under the control of the CaMV 35S promoter. The construct was then transformed into GV3101 cells and introduced to abcb19-5 by Agrobacterium tumefaciens-mediated floral infiltration as described previously [51].

GUS Staining
GUS staining and subsequent Paraplast Plus sectioning were performed as described previously [52]. A b-glucuronidase assay was performed according to the protocol of Jefferson [53].

Confocal Microscopy
Immediately after the plants were bolting, the inflorescences were cut and placed on a slide. Almost all visible buds were cut off and left only the tiny region including the inflorescence meristem. The fluorescent pictures were taken at 406lens at the excitation of 514 nm on an inverted Zeiss 510 microscope.