Genetic studies have identified the membrane protein EIN2 (ethylene insensitive 2) as a central component of ethylene signalling in Arabidopsis. In addition, EIN2 might take part in multiple hormone signalling pathways and in response to pathogens as demonstrated by recent genetic and biochemical studies. Here we show, by an integrated approach using in vivo and in vitro fluorescence techniques, that EIN2 is localized at the ER (endoplasmic reticulum) membrane where it shows specific interaction with the ethylene receptor protein ETR1.
The gaseous hormone ethylene regulates a diverse and complex range of developmental processes in plants but is also involved in mediating responses to various biotic and abiotic stresses [1,2]. Based on reverse genetics, several components involved in ethylene signalling have been identified in Arabidopsis thaliana. According to these studies, ethylene is perceived by a family of integral membrane receptors [ETR1 (ethylene resistant 1), ETR2, ERS1 (ethylene response sensor 1), ERS2 and EIN4 (ethylene insensitive 4)] that are similar to bacterial two-component histidine kinases [3–6]. However, in contrast to these receptors, which are found in the bacterial plasma membrane, the ethylene receptors are located at the ER (endoplasmic reticulum) [7,8]. The ethylene receptors activate CTR1 (constitutive triple response 1), a soluble protein kinase which co-localizes with the receptors at the ER membrane  and which negatively regulates downstream components in the ethylene signalling pathway . EIN2, an integral membrane protein acting downstream of CTR1, seems to play a central role in ethylene signalling as EIN2 is the only gene of all components involved in the ethylene pathway whose loss-of-function mutations lead to complete ethylene insensitivity . Although the central role of EIN2 for ethylene signalling is broadly recognized, current knowledge of this part of the ethylene signalling pathway is still limited. The localization of EIN2 in plant cells is not known yet. Furthermore, the molecular mechanism of signalling to and from EIN2 is largely unknown. A recent study by Qiao et al.  showing that degradation of EIN2 is triggered by the ethylene-controlled F-box proteins ETP1 and ETP2 provided the first important clue to unravelling the complex role of EIN2 in plant hormone signalling.
In the present study, we have probed the intracellular localization of EIN2 by fluorescence spectroscopy and have identified an unexpected interaction of this central protein of the ethylene signalling cascade. This interaction has not been found in previous genetic studies. The integrated approach of in planta and in vitro fluorescence measurements presented in the present study provides a detailed picture on protein–protein interactions in signalling networks, first by unravelling the intracellular localization of the complex and second by providing a precise thermodynamic parameter on the complex.
Cloning and transient expression of EIN2 and ETR1 in
The cDNA fragments encoding the genes for EIN2 or ETR1 were transferred via the Gateway BP-reaction into donor vector pDONR201 (Invitrogen). Plant expression vectors encoding ETR1 or EIN2 were generated via Gateway LR-reaction (Invitrogen) using C-terminal GFP (green fluorescent protein)- or mCherry-tagged destination vectors pABindGFP and pABindmCherry that are based on vector pMDC7 . Expression vectors (pABindGFP–0EIN2, pABindGFP–ETR1 and pABindmCherry–EIN2) as well as vector pER-Rb encoding an ER-marker protein  were transfected into Agrobacterium tumefaciens strain GV3101 pMP90 . Bacterial cells were resuspended in AS-Medium [5% (w/v) sucrose, 2 mM magnesium sulfate, 2mM glucose, 0.01% Silwet L77 and 0.225 mM Acetosyringon]. For transformation, A. tumefaciens cells containing GFP- and mCherry-tagged expression vectors and cells containing p19-plasmid coding for the P19 protein of tomato bushy stunt virus  to suppress gene silencing were mixed at 1:1:1 ratio (all to D600=0.4) and infiltrated in 5 week old N. benthamiana leafs. Infiltrated leafs were treated 36 h after transformation with 25 μM of proteasome inhibitor MG-132 and protein expression was induced by adding 20 μM β-oestradiol in 0.1% Tween 20.
Cloning, expression and purification of EIN2 and ETR1 in
The DNA fragment encoding the membrane-extrinsic C-terminal domain of EIN2 (EIN2479–1294) was amplified from a cDNA library obtained from etiolated seedlings of A. thaliana and cloned into the expression vector pET21a (Novagen, Darmstadt, Germany) containing a modified multiple cloning site derived from pET28a (Novagen). The resulting plasmid pET21a_MCSpET28a_EIN2479–1294, which encodes for the C-terminal part of EIN2 carrying N-terminal and C-terminal hexahistidine-tags was transformed into the E. coli strain BL21(DE3) and cells were grown in 2YT [1.6% (w/v) tryptone, 1% (w/v) yeast extract and 0.5% (w/v) NaCl] medium at 30°C. Expression of EIN2479–1294 was induced by addition of 0.5 mM IPTG (isopropyl β-D-thiogalactoside) at an attenuance of 0.8. Cells were harvested 6 h after induction by centrifugation and stored at −20°C. For purification of EIN2479–1294, cells were resuspended in 50 mM potassium phosphate, pH 7.6, 5% (w/v) glycerol, 300 mM NaCl, 0.002% (w/v) PMSF and passed through a pre-cooled French pressure cell at 12000 psi (1 psi=6.9 kPa). After centrifugation of the cell lysate at 100000 g, the supernatant was applied on a Ni-IDA (nickel-iminodiacetic acid) affinity column (10 mm diameter×200 mm length) previously equilibrated with 50 mM potassium phosphate buffer, pH 7.6, 5% (w/v) glycerol, 300 mM NaCl and 0.002% (w/v) PMSF. The column was washed with three bed volumes of the same buffer containing 0 mM, 75 mM and 100 mM imidazole respectively, in order to remove any proteins that were unspecifically attached to the matrix. Bound EIN2479–1294 was eluted from the column by applying a linear gradient of 100–250 mM imidazole. Elution fractions were analysed by SDS/PAGE. Fractions showing the highest purity were pooled, rebuffered in 50 mM Tris/HCl, pH 7.6, and 100 mM NaCl, concentrated and stored at 4°C or −70°C. ETR1 was expressed and purified as described in [17,18].
Preparation of tryptophan-less ETR1 and EIN2479–1294
Plasmids pET16b-ETR1  and plasmid pET21a_MCSpET28a_ EIN2479–1294 encoding the ethylene receptor protein ETR1 or the C-terminal part of EIN2 respectively, were used for the construction of tryptophan-less mutant proteins. Endogenous tryptophan residues at positions 11, 53, 74, 182, 265, 288 and 563 in ETR1 and at positions 622, 682, 812, 948, 1012, 1075, 1129, 1140 and 1158 in EIN2 were substituted either by phenylalanine or leucine by site-directed mutagenesis using sequential PCR steps as described by Cormack . Amplification reactions were carried out using Pfu polymerase, plasmids pET16b-ETR1 or pET21a_MCSpET28a_ EIN2479–1294 as template and mutagenic oligonucleotides as listed in Supplementary Figure S1 (at http://www.BiochemJ.org/bj/424/bj4240001add.htm). MCS (multiple cloning site) primers also listed in Supplementary Figure S1 were used to amplify the complete fragment encoding ETR1 or EIN2479–1294. The amplified fragments were agarose gel purified, digested with NdeI and BamHI for ETR1 or NheI and EagI for EIN2 and ligated into the equivalent sites of pET16b or pET21a_MCSpET28a. The inserted fragments were sequenced to confirm tryptophan substitutions and absence of additional mutations. Complete substitution of all endogenous tryptophan residues in ETR1 and EIN2 was obtained by the repetitive application of this protocol using the mutagenic primer listed in Supplementary Figure S1. The final plasmids containing all tryptophan substitutions were named pET16b-ETR1(ΔW) and pET21a_MCSpET28a_ EIN2479–1294(ΔW).
Steady-state fluorescence measurements on recombinant proteins were carried out in a Perkin-Elmer LS-55 spectrophotofluorometer at room temperature (20°C) using an excitation wavelength of 295 nm. Quenching of tryptophan fluorescence of purified wild-type ETR1 or wild-type EIN2479–1294 was measured at 345 nm. For the titration, 50 μl with increasing concentrations of tryptophan-less protein (0–1 μM EIN2479–1294 or 0–5 μM ETR1) was added to 450 μl of the wild-type protein (0.1 μM ETR1 or 0.5 μM EIN2479–1294). Tryptophan-less EIN2479–1294 was dissolved in a buffer containing 50 mM Tris/HCl, pH 7.6, and 50 mM NaCl; tryptophan-less ETR1 was dissolved in 50 mM Tris/HCl, pH 7.6, 100 mM potassium chloride, 50 mM NaCl and 0.1% (w/v) β-dodecylmaltoside. Wild-type EIN2 was dissolved in 50 mM Tris/HCl, pH 7.6, and 50 mM NaCl; wild-type ETR1 was dissolved in 50 mM Tris/HCl, pH 7.6, 100 mM potassium chloride, 50 mM NaCl and 0.1% (w/v) β-dodecylmaltoside. The dissociation constant of the ETR1–EIN2 complex was determined from the fluorescence data using the program GraFit (Erithacus Software) by a fit of the experimental data to a model assuming a single binding site in the interacting partners.
Confocal fluorescence microscopy
Confocal fluorescence microscopy was performed with a Zeiss LSM 510 Meta Laser Scanning Microscope (Carl Zeiss GmbH) using a Plan-Neoflur 40×/1.3 oil immersion objective. The following settings were used throughout the experiments: excitation 488 nm for GFP, 561 nm for mCherry, detection wavelength 497–550 nm for GFP and 583–636 nm for mCherry. Images were recorded and processed using Zeiss LSM 510 Software (Carl Zeiss GmbH). For FRET (fluorescence resonance energy transfer) interaction studies, GFP- and mCherry-tagged expression vectors encoding for ETR1 and EIN2 were co-infiltrated in N. benthamina. FRET was measured with the acceptor bleaching method. The measurements were performed as follows: Images (256×256 pixel, 2.55 μs scan speed/pixel) in the GFP and the mCherry channel were acquired before and after bleaching. The acceptor mCherry was bleached 100 times with 100% laser power of the 561 nm diode laser line in a ROI (region of interest). FRET was quantified by measuring GFP fluorescence intensity in the bleached area before (5 frames) and after (15 frames) bleaching. FRET efficiency (FRETeff), expressed as the percentage fluorescence increase, was calculated according to the following expression:
CD spectra of the purified C-terminal domain of EIN2 were recorded in a cylindrical quartz cuvette (path length 1 mm) at 20°C using a Jasco 715 Spectrophotopolarimeter (Jasco). Signals were monitored in a buffer containing 50 mM potassium phosphate, pH 7.6, and 50 mM NaCl at a protein concentration of 0.1 mg/ml. Secondary structure of the purified EIN2 protein was deduced from the spectra using Selcon3  and ContinLL .
Protein quantification and PAGE
Protein concentrations were determined by the bicinchoninic acid assay (Pierce Chemicals) using BSA as a standard. Purification of ETR1 and EIN2497–1294 was examined by SDS/PAGE as described by Laemmli  and Schägger and van Jagow . Proteins were separated on 10% polyacrylamide gels and visualized by silver staining as described by Heukeshoven and Dernick .
RESULTS AND DISCUSSION
Intracellular localization of the membrane protein EIN2
To examine the intracellular localization of EIN2, a full-length form of EIN2 fused to GFP at the C-terminal end was constructed and transiently expressed in N. benthamiana. Expression of the fluorescent EIN2 fusion protein was analysed by confocal microscopy. As shown in Figure 1(A), the characteristic net-like pattern of the ER membrane system was observed in the tobacco epidermal leaf cells. The identity of the stained pattern and the intracellular ER network was demonstrated by co-expression of the mCherry-labelled ER marker protein ER-Rb (Figure 1B) and the EIN2–GFP fusion protein. The perfect overlay of both expression patterns shown in Figure 1(C) demonstrated co-localization of both proteins and verified localization of EIN2 at the ER membrane. Staining of the nucleus by the fluorescent probe DAPI (4′,6-diamidino-2-phenylindole dihydrochloride) was used to analyse whether EIN2 is also targeted to the nuclear membrane. Analysis of the DAPI and GFP fluorescence signals shown in Figures 1(D) and 1(E) suggests that it was unlikely that EIN2–GFP also localizes to the nucleus, which was exclusively stained by DAPI but showed no GFP fluorescence. In summary, our localization study provides the first experimental evidence that a central component of the ethylene signalling pathway – the membrane protein EIN2 – localizes to the ER endomembrane system in plant cells. Other components of the ethylene signalling cascade such as the ethylene receptor family and the soluble Raf-like kinase CTR1 have been also localized at the ER in previous studies [7–9]. Hence signalling of the plant hormone might employ a ER-borne ternary super-complex (e.g. ETR1–CTR1–EIN2) or a binary complex formed by the receptor proteins and EIN2 (e.g. ETR1–EIN2).
Subcellular localization of EIN2 transiently expressed in
N. benthamiana epidermal leaf cells
Detection of EIN2–ETR1 interactions via FRET microscopy
Interaction of EIN2 with the ethylene receptor ETR1 was probed by FRET using the fluorophores GFP as donor and mCherry as acceptor. The acceptor fluorophor was fused to the C-terminus of EIN2, whereas the donor was attached to the C-terminus of ETR1. N. benthamiana epidermal leaf cells were co-transformed with both vector constructs and expression of both fusion proteins was induced by the addition of β-oestradiol. Upon simultaneous excitation of doubly labelled cells at 488 nm and 561 nm, both GFP and mCherry showed strong fluorescence. Energy transfer between both fluorophors and thus in planta interaction between ETR1 and EIN2 was demonstrated by the acceptor bleaching method  and is shown in Figures 2(A) and 2(B). From the increase in GFP fluorescence upon acceptor bleaching, an overall FRET efficiency of 18.3±2.3% is estimated. Similar transfer efficiencies were obtained in experiments studying the interaction of human tumour-necrosis-factor-receptor-associated-factors by FRET spectroscopy . Control experiments using tandem-labelled ETR1 show a transfer efficiency of 16.2±2.5%. Background fluctuations measured with ETR1–GFP are 3.5±1.2%. The specificity of the EIN2–ETR1 interaction was addressed in controls expressing ETR1–GFP and mCherry-labelled ER marker protein. A transfer efficiency of only 2.9±0.6% was observed in this control, which is considerably lower than the FRET efficiency found in cells expressing ETR1–GFP and EIN2–mCherry. Successful expression of all fusion proteins used in this study in the tobacco leaf cells was verified by monitoring fluorescence of the respective marker (GFP/mCherry) at the appropriate wavelength (see Supplementary Figure S2 at http://www.BiochemJ.org/bj/424/bj4240001add.htm).
in planta interaction of EIN2 and ETR1
The FRET efficiency observed for the ETR1–GFP–EIN2–mCherry donor–acceptor pair in our experiments (18.3±2.3%) was used to estimate the separation distance of the related protein–protein complex. Using a Förster distance of 5.1 nm for the GFP–mCherry donor–acceptor pair  a separation distance of 6.5 nm, with a range between 6.4 nm and 6.7 nm, was calculated from these numbers assuming random orientation of dipole moments. This value is clearly in the range that can be expected for an ETR1–EIN2 complex assuming an average diameter of 0.465 nm per transmembrane helix  and a sequence based number of 12 TMHs (transmembrane helices) for EIN2 and of three TMHs for ETR1, i.e. 15 TMHs or 7 nm in total. However, the distance fundamentally depends on the arrangement of the TMHs in the membrane as well as on the orientation of the C-termini of EIN2 and ETR1 carrying the acceptor and donor fluorophor relative to their TM domain. Furthermore, environment, orientation or movement of the fluorophors are not known for the plant system. The FRET efficiency of 18% obtained for the ETR1–GFP–EIN2–mCherry donor–acceptor pair is in clear contrast with background fluctuations and negative controls that are in the range of about 3%. Furthermore, even in a controlled in vitro system, only a maximum FRET efficiency of 35% was detected for the GFP–mCherry pair . Hence, we conclude that the FRET observed for the co-expressed ETR1–GFP and EIN2–mCherry fusions is mediated by a specific interaction of EIN2 and ETR1 rather than by non-specific binding of the overexpressed fluorescent probes.
In order to quantify the interaction of EIN2 and ETR1 revealed by FRET in the in planta studies, we have applied fluorescence titration studies using recombinant proteins. In these studies, we employed a tryptophan-less mutant of ETR1 for titration of wild-type EIN2 and vice versa, a tryptophan-less mutant of EIN2 for titration of wild-type ETR1.
To determine whether the substitution of the endogenous tryptophan residues in ETR1 led to significant changes in the protein structure, folding and secondary structure of wild-type and tryptophan-less mutant protein were checked by CD spectroscopy. Wild-type and mutant protein showed similar spectra. Autokinase activity of the recombinant proteins measured by the incorporation of radioactive labelled [γ-32P]ATP also clearly demonstrated that the tryptophan-less ETR1 protein retained wild-type characteristics (results not shown).
Cloning, expression and purification of the C-terminal domain of EIN2 in
The C-terminal part of EIN2 representing the putative membrane-extrinsic region of the receptor was amplified from a cDNA library obtained from etiolated seedlings of A. thaliana and cloned into a modified pET21a vector containing N-terminal and C-terminal hexahistidine-tags. Induction of E. coli BL21(DE3) pET21a_MCSpET28a_ EIN2479–1294 by addition of IPTG resulted in production of recombinant EIN2 protein in soluble form (Figure 3A). Expression of EIN2 was confirmed by Western blotting using anti-His antibodies (Figure 3B). Over-expressed recombinant EIN2 was purified to homogeneity from E. coli cell extracts by metal-chelate affinity chromatography using a Ni-IDA Superflow column. A single band on silver-stained SDS/PAGE confirmed purity and homogeneity of the protein (Figure 3C). By this procedure, approx. 1 mg of purified protein was routinely obtained from 1 litre of cell culture. Similar amounts were obtained with the tryptophan-less EIN2 mutant protein.
Expression and purification of EIN2479–1294
CD analysis of recombinant EIN2
The folding and secondary structure composition of recombinant wild-type and tryptophan-less EIN2479–1294 was estimated using CD spectroscopy. Wild-type and mutant proteins showed similar spectra. The observed minima at 206 nm and 220 nm (see Supplementary Figure S3 at http://www.BiochemJ.org/bj/424/bj4240001add.htm) indicate a predominately α-helical structure of the EIN2 protein. Secondary structure calculations using Selcon3 and ContinLL, two different algorithms for estimation of secondary structure composition of an unknown CD spectrum, yielded an α-helix content of 39% and a β-sheet percentage of 11–13% for the recombinant protein. Both results agree with the numbers predicted from the primary structure of the C-terminal of EIN2 using the program PROF [28,29] which are 46% α-helix and 10% β-sheet. In summary, the CD analysis demonstrates that the recombinant EIN2 protein adopts a well folded structure that is likely to reflect the natural structure of the protein. Furthermore, the agreement of CD results obtained with wild-type and tryptophan-less EIN2 indicate that folding and structure of the tryptophan-less mutant protein were not perturbed.
Fluorescence titration studies of ETR1 and EIN2
The intrinsic fluorescence of proteins is dominated by the emissions from tryptophan residues and tryptophan fluorescence provides a sensitive intrinsic probe to study changes in the local environment of a protein, e.g. caused due to the interaction of the protein with any other small or large molecule. In order to use tryptophan fluorescence as a sensitive tool to study protein–protein interactions, all endogenous tryptophan residues have to be removed from one of the proteins involved in the formation of the binary complex. Quenching of tryptophan fluorescence upon addition of the tryptophan-less binding partner can be used to evaluate the stability of the interaction and to determine the apparent dissociation constant of the complex.
When wild-type EIN2 was titrated with the tryptophan-less ETR1 mutant, a Kd of 360 nM was obtained (see Figure 4A), whereas titration of ETR1 by a tryptophan-less EIN2 mutant resulted in a slightly higher Kd of 440 nM as shown in Figure 4(B). Taken together, both studies underline that ETR1 and EIN2 form a highly specific interaction.
In vitro interaction studies of EIN2 and ETR1
The fluorescence titration measurements performed in the present study using recombinant wild-type and tryptophan-less proteins provides for the first time a quantitative understanding on a protein–protein interaction in the ethylene signalling cascade. The measurements require only small amounts of the recombinant proteins (10–6 M range) and allow an exact control of the external parameter. They have the potential to narrow down the interacting domains in both proteins and to address the stability of the complex in response to ethylene or ethylene antagonists in future experiments.
In summary, the localization and interaction studies presented in the present study define a new interaction within the ethylene signalling cascade. Together with studies from other labs [7–9], the results suggest that ethylene signalling might employ an ER-borne signalling complex consisting of ethylene receptor proteins and EIN2 or an even more complex structure containing ETR1, CTR1 and EIN2.
We thank Rüdiger Simon for stimulating discussion and helpful advice on the EIN2 localization studies. We also thank Elisa Buchen for assistance with the CD measurements and Daniel Schlieper for critical reading of the manuscript.
constitutive triple response 1
ethylene response sensor
fluorescence resonance energy transfer
green fluorescent protein
Melanie Bisson carried out the experiments, analysed data and helped in manuscript preparation. Andrea Bleckmann provided the vectors pABindGFP and pABindmCherry, and assisted in transient expression, FRET studies and data interpretation. Silke Allekotte established expression and purification of EIN2479–1294 and performed titration studies with recombinant EIN2 and tryptophan-less ETR1. Georg Groth designed the project, analysed the data and wrote the manuscript.
This work was supported by the Deutsche Forschungsgemeinschaft within the SFB 590 ‘Inhärente und adaptive Differenzierungsprozesse’ at the Heinrich-Heine-Universität Düsseldorf.