A Copper Cofactor for the Ethylene Receptor ETR1 from Arabidopsis

See allHide authors and affiliations

Science  12 Feb 1999:
Vol. 283, Issue 5404, pp. 996-998
DOI: 10.1126/science.283.5404.996


The ETR1 receptor from Arabidopsis binds the gaseous hormone ethylene. A copper ion associated with the ethylene-binding domain is required for high-affinity ethylene-binding activity. A missense mutation in the domain that renders the plant insensitive to ethylene eliminates both ethylene binding and the interaction of copper with the receptor. A sequence from the genome of the cyanobacteriumSynechocystis sp. strain 6803 that shows homology to the ethylene-binding domain of ETR1 encodes a functional ethylene-binding protein. On the basis of sequence conservation between theArabidopsis and the cyanobacterial ethylene-binding domains and on in vitro mutagenesis of ETR1, a structural model for this copper-based ethylene sensor domain is presented.

Small gaseous molecules act as signals for a variety of organisms. In many cases, signal perception involves the use of a transition metal cofactor that mediates the interaction between the signal and its proteinaceous receptor. For example, sensors for NO in animal cells and O2 in bacteria use a heme moiety to achieve high-affinity binding of the signal (1). The plant hormone ethylene is effective at nanomolar concentrations, reflecting the presence of high-affinity receptors (2). Theoretical considerations indicated Cu(I) as a possible receptor cofactor (3, 4). The opportunity to directly investigate the role of a metal cofactor in ethylene sensing has been provided by the cloning and characterization of the ETR1 gene fromArabidopsis (5). The ETR1 protein forms a membrane-associated disulfide-linked homodimer both in plant tissues and when heterologously expressed in yeast (6). Expression of truncated forms of ETR1 in yeast indicated that the first 165 amino acids of the protein contain the sequences that are necessary and sufficient to bind ethylene (7).

To further delineate the ethylene-binding domain of the ETR1 protein and to facilitate the purification of the binding activity, a chimeric gene consisting of the coding sequence for the first 128 amino acids of the ETR1 protein fused to the glutathione S-transferase (GST) coding sequence [ETR1(1-128)GST] was constructed (8). Yeast cells expressing the fusion protein showed ethylene-binding activity equivalent to that detected in cells expressing full-length ETR1 (Fig. 1A). Introduction of an amino acid substitution into the fusion protein, corresponding to the etr1-1 mutant allele [Cys65 → Tyr65 (C65Y)], resulted in the complete loss of binding activity by the expressed protein, as was previously observed for the full-length etr1-1 protein (7). Both the mutant and the wild-type fusion proteins formed membrane-associated disulfide-linked homodimers and were expressed at equivalent levels as determined by immunoblot analysis using antibodies to GST (9).

Figure 1

Requirements for ethylene binding to the ETR1 protein expressed in yeast. Saturable ethylene binding is indicated as the difference in 14C-ethylene between samples treated with 14C-ethylene (0.1 μl/liter) (white or hatched bars) and identical samples treated with14C-ethylene (0.1 μl/liter) plus 12C-ethylene (100 μl/liter) (overlapping black bars). DPM, disintegrations per minute. (A) Saturable ethylene-binding activity in yeast cells expressing different ETR1-derived constructs. Expressed proteins are depicted diagramatically, with the hydrophobic (small white squares), GAF (19) (white diamond), histidine kinase (black rectangle), receiver (black oval), and GST (white oval) domains indicated. (B) Effects of CuSO4addition to membranes extracted from yeast cells expressing the control vector (pYcDE) or the indicated ETR1-derived constructs. Ethylene-binding assays were performed with assay buffer alone (white bars) or with 300 μM CuSO4 (hatched bars) (11). (C) Effects of CuSO4 and other transition metals on ethylene-binding activity in yeast membranes expressing the ETR1 protein. Ethylene-binding assays were performed with assay buffer alone (white bar), or with 300 μM of the indicated metal salts (hatched bars).

We assessed binding in membrane extracts of yeast expressing both the full-length ETR1 protein and the ETR1(1-128)GST fusion protein (10, 11). Addition of 300 μM CuSO4to the isolated membranes led to a 10- to 20-fold increase in ethylene-binding activity (Fig. 1B); up to twice that observed in intact cells. The addition of copper had no effect on ethylene binding in native yeast membranes nor in membranes expressing the mutant etr1-1(1-128)GST fusion protein. Of other transition metals tested (Fig. 1C), only silver ions stimulated ethylene-binding activity in membranes containing the ETR1 protein.

We were able to solubilize and purify the ETR1(1-128)GST fusion protein from yeast membranes in an active form (Fig. 2). The copper content of the purified protein was determined by graphite furnace atomic absorption spectroscopy. ETR1(1-128)GST preparations contained sixfold higher concentrations of copper (Fig. 2D) than did either native yeast (pYcDE) or mutant etr1-1(1-128)GST preparations. The net amount of copper (2.7 nmol) associated with the ETR1(1-128)GST sample was stoichiometrically similar to the amount of protein dimer (2.8 nmol) calculated from the difference in protein concentration between the pYcDE control and the affinity-purified ETR1(1-128)GST preparation. However, the estimated number of ethylene-binding sites in the purified preparation of ETR1(1-128)GST was only 0.9 nmol, based on its binding activity (Fig. 2B), which indicates that not all the purified fusion protein was biochemically active. Samples of purified mutant etr1-1(1-128)GST dimer (yield, 4.0 nmol) did not copurify with copper and lacked saturable binding activity, which indicates that Cys65 is an essential residue for both copper association and ethylene binding to the receptor.

Figure 2

Copurification of copper with the ETR1 ethylene-binding domain. (A) Saturable ethylene-binding activity was determined with 50-μl aliquots of the indicated affinity-purified preparations (1 ml total volume) (20) in the presence of 14C-ethylene (white bars) or14C-ethylene plus 12C-ethylene (overlapping black bars) (11). (B) Protein immunoblots of affinity-purified samples immunodecorated with antibodies to GST (20). (C) Copper contents of the purified samples determined by graphite furnace atomic absorption; ppb, parts per billion.

In the development of a structural model for the ethylene-binding domain of ETR1, specific constraints are provided by amino acid residues that are conserved between functionally related proteins. These include the ERS1 protein from Arabidopsis and theNr (Never-ripe) gene product from tomato (12), both of which showed ethylene-binding activity when expressed in yeast (13). Database searches also revealed that an open reading frame, designated slr1212, from the cyanobacteriumSynechocystis strain 6803 (GenBank accession number D90905) showed sequence homology restricted to the ethylene sensor domain of ETR1. Ethylene-binding assays on cultured Synechocystis cell lines with wild-type and disrupted slr1212 (Fig. 3) showed that the wild-type gene encodes a functional ethylene-binding domain (14).

Figure 3

Ethylene-binding activity inSynechocystis lines containing the intact (slr1212+) or disrupted (slr1212) slr1212 ORF. The structural elements of the slr1212 coding sequence are indicated as follows: hydrophobic region related to ETR1 (small white rectangles), PAS (22) domains (black circles), and GAF motif (white diamond). The position of the insertion of theKan r gene is also shown. Ethylene-binding assays (14) were performed with 14C-ethylene (white bars) or 14C-ethylene plus 12C-ethylene (overlapping black bars).

Alignment of Synechocystis and plant sequences indicate that conserved residues in the second and third hydrophobic subdomain align along a single face when these regions are modeled as α helices (Fig. 4). Conservation in the first helix does not fall along a single face, which may mean that an α-helical structure is incorrect for this region or that this helix may contact other protein regions along more than one face.

Figure 4

(A) Helical net model of the ethylene-binding domain of the ETR1 protein. The model represents the first 128 amino acids of the ETR1 protein. The hydrophobic domain is modeled as α helices according to computer algorithms that predict membrane topology (23). Residues that are conserved between ETR1, ERS1, Nr, and slr1212 fromSynechocystis are are outlined in blue (24). Nonconserved residues are represented as solid black circles. In vitro mutagenesis of potential metal-ligand residues either disrupted (red circles) or did not disrupt (white letters) ethylene binding in the yeast-expressed protein (7). (B) Ethylene binding by yeast expressing wild-type and mutant forms of ETR1. Ethylene-binding assays were performed as in Fig. 1A. Relative levels of expressed protein were determined as previously described (7).

Evidence for a copper cofactor in ethylene binding to ETR1 places additional constraints on any structural model for the binding site. We hypothesized that particular amino acid residues, such as Cys and His, would serve as metal-coordinating ligands for the copper ion. As shown in Fig. 4, mutagenesis of Cys65 and His69residues to Ser and Ala, respectively, resulted in complete loss of binding activity in the yeast-expressed ETR1 (Fig. 4B). These two residues align along a single face when the second hydrophobic subdomain is modeled as an α helix (Fig. 4A). Mutagenesis of other Cys, His, and Met residues in the ethylene-binding domain did not result in the loss of binding activity (Fig. 4B). These results, coupled with the lack of copper copurification with the C65Y mutant etr1-1(1-128)GST, prompt us to suggest that Cys65 and His69 may serve as ligands for a Cu(I) ion in the ETR1 binding site for ethylene.

The results presented here provide a mechanistic basis for how high-affinity ethylene binding to the sensor domain of ETR1 is achieved. We propose that ethylene interacts with a Cu(I) cofactor in an electron-rich hydrophobic pocket formed by membrane-spanning helices of the ETR1 dimer. The binding site must confer some unusual chemistry on the copper ion, because the stability of this ethylene/receptor complex (half-life for dissociation = 11 hours) (7) is much different from that observed for artificial copper complexes (15). The ability of Ag(I) ions to interact with the receptor and bind ethylene is of interest because silver ions can inhibit ethylene responses when applied to plant tissues (4). Silver ions may occupy the binding site and interact with ethylene but fail to induce the changes in the receptor that are needed to elicit downstream signaling.

The discovery of a functional ethylene-binding domain in the slr1212 coding sequence from Synechocystis raises interesting questions about the evolutionary origin of the higher plant receptors. Synechocystis is thought to share a common ancestor with the cyanobacterial lineage that evolved into the modern chloroplast of higher plants (16). The presence of both the ethylene-sensor domain and histidine-kinase transmitter domains in the cyanobacterial genome may have provided the raw materials for the evolution of the higher plant form of the ethylene receptors. Sequence homology to the ethylene-binding domain has not been identified in any other bacterial genomes sequenced to date, which supports a single origin for this functional domain in the evolution of photosynthetic organisms.

  • * Present address: Department of Biochemistry and Molecular Biology, 46 College Road, University of New Hampshire, Durham, NH 03824, USA.

  • To whom correspondence should be addressed. E-mail: bleecker{at}


View Abstract

Navigate This Article