Positive fluorescent selection permits precise, rapid, and in-depth overexpression analysis in plant protoplasts.

Transient genetic modification of plant protoplasts is a straightforward and rapid technique for the study of numerous aspects of plant biology. Recent studies in metazoan systems have utilized cell-based assays to interrogate signal transduction pathways using high-throughput methods. Plant biologists could benefit from new tools that expand the use of cell culture for large-scale analysis of gene function. We have developed a system that employs fluorescent positive selection in combination with flow cytometric analysis and fluorescence-activated cell sorting to isolate responses in the transformed protoplasts exclusively. The system overcomes the drawback that transfected protoplast suspensions are often a heterogeneous mix of cells that have and have not been successfully transformed. This Gateway-compatible system enables high-throughput screening of genetic circuitry using overexpression. The incorporation of a red fluorescent protein selection marker enables combined utilization with widely available green fluorescent protein (GFP) tools. For instance, such a dual labeling approach allows cytometric analysis of GFP reporter gene activation expressly in the transformed cells or fluorescence-activated cell sorting-mediated isolation and downstream examination of overexpression effects in a specific GFP-marked cell population. Here, as an example, novel uses of this system are applied to the study of auxin signaling, exploiting the red fluorescent protein/GFP dual labeling capability. In response to manipulation of the auxin response network through overexpression of dominant negative auxin signaling components, we quantify effects on auxin-responsive DR5::GFP reporter gene activation as well as profile genome-wide transcriptional changes specifically in cells expressing a root epidermal marker.


INTRODUCTION
It has been demonstrated that flow cytometric analysis and fluorescence activated cell sorting (FACS) of plant protoplasts is practicable, moreover, this technique has yielded valuable results in a number of different fields of research (Harkins and Galbraith, 1984;Galbraith et al., 1995;Sheen et al., 1995). For instance, FACS of protoplasts from Arabidopsis plants expressing tissue-specific fluorescent protein markers has been used to examine both basal and environmentally stimulated transcriptional profiles in particular cell types (Birnbaum et al., 2003;Brady et al., 2007;Gifford et al., 2008;Dinneny et al., 2008) and flow cytometry has been employed to analyze reactive oxygen species production and programmed cell death tobacco protoplasts (Nicotiana tabacum;Lin et al., 2006). A broad selection of fluorescent tools is available to study a plethora of physiological parameters in plants, e.g. cis-regulatory elements fused to fluorescent proteins (Haseloff and Siemering, 2006), genetically-encoded molecular sensors (Looger et al., 2005) or dye-based sensors (Haugland, 2002) can be used in combination with cytometry to measure diverse biological processes.
Here, we document the development of a protoplast transfection system that employs cytometry and a transient transformation vector harboring a fluorescent positive selection marker (pBeaconRFP; Fig. 1). The notable advantage of this system is that it allows for the exclusive analysis of the transformed cells and facilitates high-throughput dual color analysis.
The new vector for use in this system is designed in such a way that it not only expresses a gene-of-interest but also expresses monomeric red fluorescent protein (mRFP). Furthermore, it is compatible with the Gateway recombinase-mediated cloning system, permitting fast and easy cloning. Because of its red emission spectrum, the mRFP marker can easily be used in combination with the commonly utilized green fluorescent protein (GFP). We present two examples of this system's use in the analysis of an important signal transduction cascade involved in many aspects of plant development, namely the auxin perception pathway (Fig. 2;Guilfoyle and Hagen, 2007). Promising alternative uses of the system are further discussed. 6 agent (usually Agrobacterium tumefaciens), transfection of protoplasts can be achieved in just one day and entails only raw DNA and either a chemical-or electroporation-based transfection method. Additionally, transient transformation analyses can overcome problems encountered with stable over-expression such as pleiotropic developmental effects or nonviability, when a cell-based assay is appropriate. However, due to the fact that protoplast transformation efficiency is never 100%, results can be convoluted by the non-transformed cells.
Transformation efficiencies are often low and variable (e.g. Cummins et al., 2007;<10%) and depend on the employed method as well as properties of the protoplasts and DNA used. We usually get efficiencies ranging from 5 to 20% using Arabidopsis root protoplasts.
Others in the field, however, have reported efficiencies of up to 90% using Arabidopsis mesophyll protoplasts (Sheen, 2001). Nonetheless, even a relatively small contamination with non-transformed cells can obscure effects and lead to a misinterpretation of results. For example, supposing one wanted to measure the ability of a dominant negative signaling component to inhibit the activation of downstream targets and one still sees a level of activation after transfection of the protoplasts; albeit significantly reduced as compared to a control, is the remaining activity due to a partial inhibition or is it only present in the nontransformed cells? If a way were to be found to select for successfully transformed cells, on the other hand, a much more precise measurement of the parameter of interest could be obtained.
As a first example we used the pBeaconRFP transient transformation system for the rapid analysis of a regulatory circuit by means of reporter gene readout. We over-express dominant negative mutant isoforms of the Aux/IAA transcription factors (IAAnmII; Fig. 2; Tiwari et al., 2001) in protoplasts derived from the roots of Arabidopsis seedlings stably transformed with the auxin sensitive reporter DR5:: GFP (Fig. 3A;Ottenschläger et al., 2003; the pBeaconRFP system is also usable in mesophyll cells as mRFP is readily distinguishable from chlorophyll autofluorescence cytometrically [data not shown]). In this first experiment, we validate the system using the elegant experiments pioneered in the Guilfoyle lab (Ulmasov 7 cell-type-specific GFP marker line (P WER ::GFP, the WEREWOLF promoter fused to GFP), which expresses primarily in atrichoblasts (Lee and Schiefelbein, 1999), transfected with two different IAAnmII isoforms. Subsequent genome-wide transcriptional profiling of auxintreated and mock-treated IAAnmII expressing cells makes it possible to distinguish distinctive patterns of gene expression regulated by the different mutant Aux/IAA isoforms.
The use of high-throughput cell-based screening methods in the study of regulatory networks has become a conventional and effective approach in animal systems (e.g. Müller et al., 2005;Palmer et al., 2006). Cytometric and FACS-based analyses have also been much more widespread and prolific in animal or microbiology research than in plant research. The combination of a selectable protoplast transformation system along with the use of cytometry now allows us to take these powerful techniques to a new level in plant research.

Modification of reporter gene activation by transient over-expression
In order to demonstrate the use of the pBeaconRFP system to study signal transduction, we took advantage of the auxin-sensitive DR5::GFP reporter gene. DR5 is a highly active synthetic auxin response element created by Ulmasov and coworkers (1997) and derived from the soybean (Glycine max) GH3 indole-3-acetic acid amido synthetase promoter. Upon treatment of seedlings or protoplasts harboring a DR5 reporter gene with auxin, the reporter is activated throughout the plant or protoplast suspension (Ulmasov et al., 1997).
Measurement of auxin-induced GUS activity relative to luciferase activity showed a reduced induction of GUS activity in protoplast suspensions transfected with the Aux/IAA effector plasmid as compared to those transfected with a control vector, indicative of the repressive effect on auxin responses of this family of transcription factors. A drawback of this initial system is that GUS activity induced in protoplasts that have been transformed with the reporter and not the effector will also be measured. The relative amount of protoplasts transformed with less than all three of the applied vectors will vary from experiment to experiment and among different effector plasmids. An improved version of this system, in which mesophyll protoplasts from a stably transformed DR5::GUS Arabidopsis line were utilized, avoided the need for the co-transfection with the reporter and allowed for an analysis of the reporter in a more natural chromatin environment but did not address the issue of measuring response only in transformed cells (Tiwari et al., 2006).
It has been demonstrated previously that stabilizing mutations in Domain II of Aux/IAA proteins lead to a repression of auxin-responsive reporter gene activation (Fig. 2;Tiwari et al., 2001). These authors used the carrot protoplast system described above and presented results indicating that over-expression of these dominant negative mutant isoforms caused a marked reduction in reporter gene activation, although it appeared that the inhibition was incomplete.
Here, we have constructed an mRFP positive marker containing Gateway-compatible transient transformation vector, pBeaconRFP ( Fig. 1), and have cloned the dominant negative Aux/IAA isoforms IAA7mII and IAA19mII, provided by the Guilfoyle lab, into this vector.
pMON999_mRFP was utilized as a control vector, expressing only mRFP. These vectors were used to transfect protoplasts derived from the roots of 1-week-old DR5::GFP Arabidopsis seedlings (Fig. 3A). After an overnight incubation, giving the transformed protoplasts the opportunity to start expressing the IAAnmIIs and mRFP, protoplast suspensions were treated with 5 µM indole-3-acetic acid (IAA) or mock-treated with solvent and monitored cytometrically. Figure  Interestingly, the quantification also showed that the GFP signal in mock-treated IAA7mIIand IAA19mII-expressing protoplasts was already less intense than in the protoplasts transformed with the control vector, a 1.7-and 2.6-fold repression, respectively (Fig. 3D). An independent experiment is presented, showing a time course analysis of GFP induction (Fig.   3E), reiterating the previous result and allowing examination of the kinetics of reporter gene activation.
These results corroborate previous results (Tiwari et al., 2001) and validate the pBeaconRFP system. Furthermore, they demonstrate that we were able to measure reporter gene activation specifically in the transformed cells and indicate that both IAA7mII and IAA19mII effectively repress auxin-induced DR5::GFP expression. This system permits a highly quantitative live analysis and has the potential for large scale screening of candidate genes for effects on reporter gene activation.

Transcriptional analysis of cell-type-specific transient over-expression
In order to demonstrate an entirely novel use for the system, we used pBeaconRFP in combination with a cell-type-specific GFP marker to isolate dually labeled cells by FACS and analyze the effects of over-expression in a particular cell population.  (Gifford et al., 2008;Dinneny et al., 2008). Auxin responses are also expected to diverge between different cell types, this can be deduced from, among other evidence, the cell-typespecific expression of the different isoforms of the ARF-Aux/IAA auxin perception pathway (Weijers and Jürgens, 2004).
We have used pBeaconRFP to transiently express IAAnmIIs in protoplasts derived from the roots of P WER ::GFP Arabidopsis seedlings. Following overnight incubation and a 3 hour treatment with IAA or solvent alone, dually labeled protoplasts were separated using FACS and transcriptionally profiled by means of microarray analysis (Fig. 4A). Protoplast suspensions were transfected with the pMON999_mRFP control vector, pBeaconRFP_IAA7mII or pBeaconRFP_IAA19mII. Figure 4B shows microscopic images of a labeled protoplasts would be sorted (Fig. 4C). The experiment was performed in triplicate; 9 separate transfections, 18 treatments, sorts and microarrays. In corroboration with known auxin responses and our own data, the expression of Arabidopsis GH3.5, as measured by microarray, resembles the DR5::GFP expression measured in the previous experiment, displaying a drastically reduced auxin-induced increase in expression level and a basal repression of expression in protoplasts transformed with the IAAnmIIs (Fig. 3, D and E; Fig.   4D). Furthermore, genes displaying a response to auxin in the protoplasts transformed with the control vector generally exhibited a dampened response in the protoplasts expressing IAAnmIIs ( Table I). Analysis of the data as a whole showed that the protoplasts transformed with the IAAnmIIs were fundamentally already very different compared to the protoplasts transformed with the control vector. Interestingly, although they were more similar to each other than to the control, there was also a substantial amount of statistically significant gene expression differences between protoplasts expressing IAA7mII and IAA19mII ( Fig. 4D; Table I).
These results provide a proof of concept for the feasibility of transcriptional profiling after transient protoplast transformation. This is now possible due to the fact that the system eliminates any contaminating effects of non-transformed cells. Furthermore, the dual-color cell sorting approach makes it possible to analyze the effect of over-expression in a specific population of cells. In this case, the system allowed us to compare the outcome of expression of two highly homologous signal transduction cascade components and the results indicate that IAA7 and IAA19 have both overlapping and unique downstream consequences in protoplasts derived from the Arabidopsis root epidermis. These results can be pursued to investigate mechanisms that lead to the specificity of auxin signal transduction. This demonstrates how the pBeaconRFP system can be used as a tool for rapid and highthroughput as well as in-depth analysis of genetic circuitry. is Gateway-compatible, making it quick and easy to clone genes-of-interest and amenable to high-throughput approaches (e.g. Sutter et al., 2005).

Making use of the pBeaconRFP vector and FACS-based collection of cells permits
analysis not only of effects on GFP-reporter gene activation or transcriptional profiles, as demonstrated here, but also of any other measurable parameters, such as enzymatic activities or metabolite levels. In combination with cell-identity markers, this system now also makes it possible to quickly analyze over-expression effects in a cell-type-specific manner.
Additionally, measuring effects in a particular cell population, as opposed to a heterogeneous mix of protoplasts, allows for a more defined and specific analysis. Moreover, there is the potential of measuring multiple parameters at once, for instance, one could measure the effect of manipulation of upstream signal transduction elements both on mitogen activated protein kinase activation and its ultimate downstream transcriptional responses. Of course this system does not have to be used exclusively with flow cytometry or FACS; for example, it could also be used to select transformed protoplasts for individual analyses such as patch clamping or subcellular protein localization studies. Alternatively, a use in combination with more basic fluorometric analyses could be envisioned, such as microscopic analyses or assays performed with plate readers. Lastly, the system is conceptually well suited for high-throughput screening purposes (e.g. looking for genes that activate or inhibit activation of a favorite reporter gene or complementation screens in mutant backgrounds). In conclusion, the technique described here opens up a wide field of possibilities not previously feasible in plant research.
Further development and enhancement of this system is ongoing. A transient silencing vector containing a positive selection marker will allow for RNAi manipulations.
Enhancement with glucocorticoid receptor protein fusion or a transcriptionally inducible system will make it possible to time the activation or over-expression of one's gene-ofinterest (Moore et al., 2006). A Gateway-compatible multicolor protein tagging set will give the possibility of high-throughput protein localization studies as well as protein interaction screens. Additionally, vectors with alternate positive selection markers, such as GFP or other fluorescent proteins, will permit analysis of protoplasts transformed with multiple effectors.
Lastly, the development of low-stress-eliciting protoplast transfection procedures will allow examination of protoplasts that more closely resemble their natural state. The pBeaconRFP vector will be made available through the Flanders Institute of Biotechnology (VIB, http://www.psb.ugent.be/gateway/), where the backbone originated. The microarray data has been deposited in the Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) database under accession GSE13783. agaaagctgggttcactcgtctactcctctag and subsequently re-amplified with primers AttB1 ggggacaagtttgtacaaaaaagcaggct and AttB2 ggggaccactttgtacaagaaagctgggt using Phusion polymerase. The PCR products were recombined into pDONR221 using BP clonase and subsequently shuttled into pBeaconRFP with LR clonase (Invitrogen).

Protoplast isolation and transfection
Protoplast isolation and polyethylene glycol (PEG) -mediated transfection was performed basically as described by the Sheen lab (http://genetics.mgh.harvard.edu/sheenweb/). Roots of 1-week-old seedlings were harvested with a scalpel and placed into a gently shaking flask with 100 ml protoplasting solution for 3 h. Protoplasting solution was prepared as follows: Protoplast suspensions were incubated overnight in 24-well plates in the dark.

Flow cytometry and FACS
that accounts for multiple testing (Significance Analysis of Microarrays, two-class unpaired test, Wilcoxon statistic; q<10% False Discovery Rate). In order to assess the effects of IAAnmII expression on auxin responses, the transcripts that showed a significant difference between mock-treated control vector and IAA-treated control vector (basal auxin response, n=809) were then tested for their fold-change response in experiments in which protoplasts were transiently transformed with pBeaconRFP_IAAnmII and mock-treated or treated with auxin. Increases and decreases in average expression were converted to an absolute fold change to measure the overall effect of the over-expression on the basal auxin response. IAA7 and IAA19 were removed from analysis in their respective over-expressor samples.    . Transcriptional analysis of cell-type-specific IAA7mII and IAA19mII expression. A, A schematic representation of the experiment. Protoplasts derived from the roots of 1-week-old P WER ::GFP seedlings were transfected with either pMON999_mRFP, expressing only mRFP, or pBeaconRFP expressing IAA7mII or IAA19mII. After an overnight incubation, protoplast suspensions were treated for 3 hours with 5 µM IAA or solvent alone. Dually labeled protoplasts were isolated by FACS and use for microarray analysis. B, Microscopic examination of protoplasts derived from the roots of 1-week-old P WER ::GFP seedlings that were transfected with pMON999_mRFP. Scale bar represents 50 µm. C, Fluorescence activated cell sorting of the transfected protoplast suspensions. Dot plots are shown depicting the controls used to set up the gates: an untransfected protoplast suspension derived from wild-type roots (blank), an untransfected protoplast suspension derived from P WER ::GFP roots and a protoplast suspension derived from wild-type roots transfected with pMON999_mRFP. In addition, a dot plot depicting a protoplast suspension derived from P WER ::GFP roots transfected with pMON999_mRFP is shown; protoplasts falling within the gate marked "Double" were sorted and used for microarray analysis. 100,000 events are displayed in each dot plot. D, Transcriptional analysis of sorted protoplasts. A log-scale heatmap and a histogram quantifying the differences in gene expression between the various collected protoplasts are shown. The heatmap displays all the genes that exhibit any significant difference between either mock-and auxin-treatment, between transformation with the different vectors or on an interaction level (see MATERIALS AND METHODS) as measured by microarray analysis, rows represent genes and columns represent treatment and transformation vectors (n=3). The histogram presents the difference in GH3.5 expression (as measured by microarray) between the various collected protoplasts +/-s.d. (n=3).

Table I
Transcriptional changes induced by auxin-treatment and IAAnmII over-expression. The average fold-change in expression of the 809 genes responsive to auxin treatment in the pMON999_mRFP control vector is presented for protoplasts transformed with the control vector as well as protoplasts transformed with the IAA7mII and IAA19mII over-expressors. The number of genes with statistically significant differences (see MATERIALS AND METHODS) in expression between all vectors and treatments are given.