Heavy metal-associated isoprenylated plant protein (HIPP): characterization of a family of proteins exclusive to plants

Authors

  • João Braga de Abreu-Neto,

    1. Programa de Pós-Graduação em Biologia Celular e Molecular, Universidade Federal do Rio Grande do Sul, Brazil
    Search for more papers by this author
  • Andreia C. Turchetto-Zolet,

    1. Programa de Pós-Graduação em Genética e Biologia Molecular, Universidade Federal do Rio Grande do Sul, Brazil
    Search for more papers by this author
  • Luiz Felipe Valter de Oliveira,

    1. Programa de Pós-Graduação em Genética e Biologia Molecular, Universidade Federal do Rio Grande do Sul, Brazil
    Search for more papers by this author
  • Maria Helena Bodanese Zanettini,

    1. Programa de Pós-Graduação em Genética e Biologia Molecular, Universidade Federal do Rio Grande do Sul, Brazil
    Search for more papers by this author
  • Marcia Margis-Pinheiro

    Corresponding author
    1. Programa de Pós-Graduação em Genética e Biologia Molecular, Universidade Federal do Rio Grande do Sul, Brazil
    • Programa de Pós-Graduação em Biologia Celular e Molecular, Universidade Federal do Rio Grande do Sul, Brazil
    Search for more papers by this author

Correspondence

M. Margis-Pinheiro, Departamento de Genética, Universidade Federal do Rio Grande do Sul, Brazil

Fax: +55 51 3308 7311

Tel: +55 51 3308 9814

E–mail: marcia.margis@ufrgs.br

Abstract

Metallochaperones are key proteins for the safe transport of metallic ions inside the cell. HIPPs (heavy metal-associated isoprenylated plant proteins) are metallochaperones that contain a metal binding domain (HMA) and a C–terminal isoprenylation motif. In this study, we provide evidence that proteins of this family are found only in vascular plants and may be separated into five distinct clusters. HIPPs may be involved in (a) heavy metal homeostasis and detoxification mechanisms, especially those involved in cadmium tolerance, (b) transcriptional responses to cold and drought, and (c) plant–pathogen interactions. In particular, our results show that the rice (Oryza sativa) HIPP OsHIPP41 gene is highly expressed in response to cold and drought stresses, and its product is localized in the cytosol and the nucleus. The results suggest that HIPPs play an important role in the development of vascular plants and in plant responses to environmental changes.

Abbreviations
ATX1

antioxidant protein 1

CCC2

yeast copper-transporting ATPase

CCH

copper chaperone

CCS

copper chaperone for Cu–Zn superoxide dismutase

COX17

copper chaperone for cytochrome c oxidase

HIPP

heavy metal-associated isoprenylated plant protein

HMA domain

heavy metal-associated domain

HPP

heavy metal-associated plant protein

Introduction

It is estimated that one-third to half of all proteins require metallic ions, either as a structural component or as a catalytic factor [1-3]. These ions possess chemical and physical properties that are essential for diverse metabolic reactions such as hydrolysis, electron transfer and oxidation of molecules. However, these same reactive characteristics make the metals potentially toxic. Thus, all organisms have developed fine-tuned mechanisms that allow them to obtain and maintain the necessary concentration of each of these trace elements without accumulating excess amounts [2, 4].

In eukaryotes, metallic ions that enter the cell are immediately chelated by proteins and small molecule ligands. For example, copper (Cu) binds to glutathione and various thiol-rich proteins, which prevents release of free Cu+ into the cytosol and protects the cell against Cu toxicity. However, to act as structural components or as enzymatic co-factors of many proteins, the metallic ions need to reach specific cellular sites. Transport of these essential ions is accomplished by proteins called metallochaperones [1, 5, 6].

Metallochaperones are generally soluble intracellular proteins that bind metallic ions tightly enough to prevent deleterious reactions with other cellular elements and transport the ions safely through the cell. Additionally, these proteins are labile enough to permit ion exchange, delivering the metal directly to a specific partner via protein–protein interactions [6]. The concept of metallochaperones is derived from study of the bacterial CopZ protein and the yeast antioxidant protein 1 (ATX1). These similar small proteins (69 and 73 residues, respectively) contain the Cys-XX-Cys metal-binding motif within the first loop of a ferredoxin-like structural fold (βαββαβ), which is usually referred to as a heavy metal-associated domain (HMA; pfam00403.6) [6]. These Cu metallochaperones are able to lower the kinetic barrier for Cu transfer in the cytosol, and catalyze Cu delivery through interaction between analogous HMA domains of the chaperone and a partner protein, which may be a transporter or the metal ‘target protein’ [4, 7].

In yeast, ATX1 was first shown to interact with CCC2, which is a Cu-transporting P–type ATPase localized in the membrane of the Golgi, where Cu is assembled into new metalloproteins [1, 4]. This appears to be one of the most basic and ancestral type of pairing between Cu proteins. For example, in organisms that do not need Cu to survive, such as Archaeoglobus fulgidus, Thermoplasma acidophilum and Thermotoga maritime, this pair of proteins forms a basic efflux system comprising a carrier (ATX1-like) that captures Cu and delivers it to a pump (CCC2-like) that exports it across the membrane [8].

A similar process is observed with two other well-conserved eukaryotic chaperones: copper chaperone for Cu–Zn superoxide dismutase (CCS) and copper chaperone for cytochrome c oxidase (COX17). CCS delivers Cu to a Cu/Zn superoxide dismutase, while COX17 mediates the transfer of Cu to cytochrome c oxidase, which is essential in the respiratory chain. In addition to an HMA domain similar to that of ATX1, both proteins contain other functional domains that are involved in recognition of their specific targets [1, 9, 10].

In addition to the three previously mentioned Cu chaperones (ATX1, CCS and COX17), another ATX1-like protein has been described in plants. The Arabidopsis thaliana copper chaperone CCH differs from other plant ATX1 homologs due to the presence of 46 additional residues at its C–terminus. Both AtATX1 and AtCCH are able to restore a normal Cu tolerance phenotype to the yeast ΔATX1 null mutant [11, 12]. However, the C–terminal extension of CCH is a plant-exclusive domain that acts independently of the N–terminal HMA domain and plays an important role in translocation of the protein to neighboring cells in plant vascular tissues, which is a characteristic specific of this metallochaperone [12, 13].

In other organisms, e.g. algae, fungi and animals, there are a small number of metallochaperone-like proteins. However, in plants, this group has diversified into a large family comprising two types of proteins: heavy metal-associated plant proteins (HPPs) and heavy metal-associated isoprenylated plant proteins (HIPPs) [14, 15].

HIPPs are characterized by the presence of one or two HMA domains and an isoprenylation motif. In addition, most HIPPs present a flexible glycine-rich region between these elements [14, 16]. Isoprenylation, also known as farnesylation, is a post-translational protein modification that involves addition of a C–terminal hydrophobic anchor that is important for interaction of the protein with membranes or other proteins. This occurs via covalent thioether binding of a 15–carbon farnesyl or 20–carbon geranylgeranyl group to the cysteine residue of a C–terminal CaaX motif (where ‘C’ is cysteine, ‘a’ is an aliphatic amino acid, and ‘X’ is any amino acid) [15, 17].

Although the HMA domain and the isoprenylation process are common in many organisms from bacteria to humans, the presence of both in the same protein was observed only in plants [18]. Only a few HIPPs have been studied so far, and the more detailed studies focused on the model plant Arabidopsis. In this species, 45 HIPPs and 22 HPPs have been identified, including AtATX1, AtCCH and AtCCS as HPPs [14, 15]. HIPPs were described as acting in two different ways: (a) in heavy metal homeostasis and detoxification mechanisms (especially cadmium tolerance) [15, 16, 19, 20], and (b) as regulatory elements in the transcriptional response to cold and drought [14, 21].

However, virtually nothing is known about the evolution of the HIPP gene family or even whether they are involved in other biological processes. In the present study, we present phylogenetic analyses for HIPP genes based on newly available data from recent genome sequencing projects of numerous organisms in order to help understand their evolutionary history in plants. Additionally, to suggest processes in which these genes may be involved, we performed and expression analysis of HIPP genes in the model plants Arabidopsis and rice. The diversity of expression patterns for HIPP genes suggests that HIPP may be involved in various roles in plant development and defense responses.

Results

Phylogenetic analysis of the HIPP gene family

Through a search of the available genomic databanks, putative HIPP sequences were identified in all vascular plants but not in any other organisms, including the moss Physcomitrella patens. In the lycophyte Selaginella moellendorffii, which belongs to the oldest groups of live vascular plants, five putative HIPP genes were found, and 59 HIPP genes were identified in rice. This illustrates the vast diversification of this gene family in more derived plants (Table S1). While only ATX1 genes are found in animals, fungi and algae, a larger number of other ATX1-like metallochaperone genes (HIPPs and HPPs) were identified in plants (Fig. 1).

Figure 1.

Phylogenetic tree of ATX1-like proteins inferred by Bayesian analysis using the HMA domain from ATX1, CCH, HPP and HIPP sequences of Saccharomyces cerevisiae (Sce), Schizosaccharomyces pombe (Spo), Homo sapiens (Hsa), Mus musculus (Mmu), Danio rerio (Dre), Aedes aegypti (Aae), Drosophila melanogaster (Dme), Selaginella moellendorffii (Smo), A. thaliana (Ath), O. sativa (Osa), Populus trichocarpa (Ptr), Setaria italica (Sit), Physcomitrella patens (Ppa), Chlamydomonas reinhardtii (Cre), Volvox carteri (Vca) and Ostreococcus lucimarinus (Olu).

A great number of duplications and diversification of HIPP genes were observed in this gene family. Phylogenetic analyses show genes that diversified at various times through plant evolution, some recently as observed for clusters exclusive to one species (e.g. PtrHIPP23–29 and PtrHIPP62) or plant family (e.g. OsaHIPP1–5 and SitHIPP11–12) (Fig. 2). In P. patens, 12 putative HPP genes were identified, suggesting a similar diversification in this group of plants (Fig. 1).

Figure 2.

Phylogenetic analysis of HIPPs inferred by the Bayesian analysis using HIPP peptide sequences of five plant species: Selaginella moellendorffii (black), A. thaliana (green), O. sativa (red), Populus trichocarpa (blue) and Setaria italica (orange).

The phylogenetic analysis of all the HIPPs found in five distinct plant genomes (five genes in S. moellendorffii, 45 in A. thaliana, 74 in Populus trichocarpa, 59 in O. sativa and 51 in Setaria italica) showed five distinct clusters (Fig. 2). Except for S. moellendorffii, all the plants that were analyzed possess genes in each of these clusters, indicating an early diversification of angiosperm HIPP genes in these five groups. Two S. moellendorffii proteins were not included in any of the five clusters (SmoHIPP02 and SmoHIPP04), but SmoHIPP01, SmoHIPP05 and SmoHIPP06 were grouped into cluster II (Fig. 2).

A relevant characteristic of the HIPPs is their size, which ranges from 113 to 584 amino acids. This is several times larger than AtATX1, which has 66 amino acids. Despite their great variation in size, most HIPPs have a conserved structure, which includes a globular HMA domain similar to that of ATX1, a C–terminal isoprenylation CaaX motif, and a flexible segment located between these two domains with predicted protein–protein binding elements (e.g. glycine-rich repeats and proline-rich motifs). A closer analysis of the architecture of HIPPs in rice and Arabidopsis shows conserved features within the five predicted clusters. For example, proteins from cluster I are distinguished by the presence of a second HMA domain, while larger proteins are found in cluster III, with a great quantity of glycine repetition (GGGGG) (Fig. 3).

Figure 3.

The architecture of HIPP proteins identified in A. thaliana (A) and the rice (Oryza sativa) genome (B). Dendogram of predicted HIPPs using the neighbor-joining method with 2000 bootstrap replications and the amino acid sequences of the HMA domains of putative HIPPs (MEGA 5.1 [40]). Protein architecture was determined, and conserved domains are indicated based on MEME 5.05 results. Red dot, HMA domain; blue box, isoprenylation motif; pink box, glycine-rich repetitions; yellow box, proline-rich motif.

Expression patterns of rice and Arabidopsis HIPP genes

The HIPP genes demonstrated a broad spectrum of expression profiles. Analysis of available microarray expression data from A. thaliana and O. sativa revealed the absence of a conserved pattern shared by all members of the HIPP family. Rather, some genes exhibited ubiquitous expression, but expression of others was more restricted to a specific organ and/or stage of plant development.

AtHIP34, AtHIPP42, OsHIPP31 and OsHIPP59 are examples of genes that are expressed throughout plant development and in diverse tissues, while other genes show a more specific expression pattern (Figs S1, S2 and S3). In rice, OsHIPP21 and OsHIPP28 were specifically and highly expressed in the leaf, while OsHIPP13 and OsHIPP48 were expressed mostly in the roots. Some genes, such as OsHIPP10, are expressed only in the inflorescence and reproductive tissues (Fig. S4).

Response to heavy metal exposure

Previous studies showed the involvement of HIPPs in cadmium (Cd) homeostasis. The expression of four HIPP genes was analyzed by quantitative RT–PCR in shoots of rice seedlings treated with 60 μm CdCl2. Three genes showed a small but statistically significant induction after 24 h treatment, OsHIPP28, OsHIPP41 and OsHIPP21 transcripts increased approximately two-, three- and fivefold, respectively, in seedlings exposed to excess Cd compared with control conditions (Fig. 4). However, this was not observed after 6 or 12 h treatment. Expression of OsHIPP50 did not show any significantly variation in the analyzed samples (Fig. S5).

Figure 4.

Response of OsHIPP21, OsHIPP28 and OsHIPP41 on rice seedlings subjected to diverse treatments: cadmium (exposed to 60 μm CdCl2 for 24 h; black bars), drought (15 days; gray bars) and cold (4 °C for 24 h; whitebars). Quantitative RT–PCR results showing the relative gene expression of each gene normalized against the expression of three control genes. Asterisks indicate statistically significant differences between control and treated plants (*< 0.05; **< 0.01).

Using the Genevestigator tool [22] , it was possible to identify other Cd-responsive HIPP genes based on the available microarray data. This analysis identified additional up-regulated genes (e.g. AtHIPP05, AtHIPP13, AtHIPP14, AtHIPP39, AtHIPP43, OsHIPP14 and OsHIPP44) and down-regulated genes (e.g. AtHIPP32, OsHIPP18, OsHIPP23 and OsHIPP38) in the roots of plants treated with Cd (Fig. S6).

Expression of OsHIPP41 is induced by cold and drought

OsHIPP21 and OsHIPP28, which are highly expressed in leaves under normal conditions and induced by excess Cd, were also tested by quantitative RT–PCR in cold- and drought-treated seedlings. The relative expression of OsHIPP41 exhibits a tenfold increase in response to drought and a more than 50-fold increase in response to low temperatures, while OsHIPP28 shows no significant changes, and OsHIPP21 is down-regulated in response to these conditions (Fig. 4).

Analysis of microarray data from rice seedlings exposed to drought, salt and cold stress has shown that most HIPP transcript levels did not change under these conditions. However, some HIPP genes were responsive under the tested conditions. The expression of OsHIPP09 and OsHIPP15 was down-regulated in response to drought and salt stress, respectively, while expression of OsHIPP23 and OsHIPP40 was down-regulated under both conditions. However, OsHIPP11 and OsHIPP45 were down-regulated only in response to cold stress (Fig. S5). OsHIPP41 was weakly expressed under normal conditions (Figs S3 and S4) but was highly expressed in response to cold and drought conditions (Fig. S7). In order to confirm this result, we performed quantitative RT–PCR for the OsHIPP41 gene in rice seedlings exposed to cold and drought treatment and found the same pattern.

Subcellular localization

Knowledge regarding the localization of a protein inside the cell is very important for elucidation of its function. To verify where the HIPP proteins are localized in vivo, RFP-tagged fusion proteins were expressed and observed in rice protoplasts. OsHIPP21 and OsHIPP41 were chosen for this analysis. Both proteins are present in the protoplast organ source (2-week-old seedling shoots) but showed contrasting expression patterns (Fig. 4). An N-terminal fusion protein construct was selected because isoprenylation has been shown to influence distribution of the HIPP protein inside cellular compartments [14, 16, 21].

In transformed protoplasts, fusion proteins were observed in the cytosol and in the cell nucleus, showing them to be soluble proteins that are not bound to cellular membranes by hydrophobic anchors. In a bioinformatics approach, three distinct prediction programs were used to infer the subcellular localization of all rice and Arabidopsis HIPP proteins. However, for most proteins, a consensus among programs was not obtained (Table S1). The result observed in vivo was in agreement with the prediction given by the CELLO program [23].

The nuclear dye 4′,6-diamidino-2-phenylindole (DAPI) was used to confirm the presence of recombinant proteins in the nucleus in a complementary experiment (Fig. 5G–J). A variation in the distribution of the tagged protein was found. In some protoplasts, the reporter protein is abundant in the nucleus and almost absent in other parts of the cell (Fig. 5C–E); however, in other protoplasts, such a distribution was not observed (Fig. 5F,G). This variation may be attributed to the existence of different cell types in the protoplast preparation.

Figure 5.

Subcellular localization in rice protoplasts. Green, false-color chlorophyll image; red, reporter protein. (A, B) Individual protoplasts expressing only RFP (positive control for transformation). (C, D) Localization of the RFP–OsHIPP21 fusion protein. (E–J) Localization of the RFP–OsHIPP41 fusion protein. 4′,6-diamidino-2-phenylindole (DAPI) nucleic acid stain (blue) was utilized to highlight the nucleus. Images (A–F) show individual cells from independent transformations; images (H–J) show the same protoplast under different light conditions. Imaging was performed using a confocal laser scanning microscope.

Discussion

The HMA domain and the isoprenylation motif are found in almost all species, but the presence of both in the same protein occurs only in vascular plants. While other types of metallochaperones are encoded by no more than a few genes in each organism, the HIPPs have diverged to form large families of more than 40 members (Fig. 1 and Table S1). All plant metallochaperone genes, including ATX, HPP and HIPP, started diversification early during plant evolution, and the phylogenetic analysis shows that they share a common ancestor in plants and algae species. This high diversification is in agreement with the important role of this gene family in all plant species. The phylogenetic relationships between metallochaperones of plant and metallochaperones of other eukaryotic organisms confirmed that the diversification occurred in the ancestral lineage of plant species, mainly within vascular plants (Fig. 1). This analysis also shows that HIPP and HPP genes have duplicated and diversified in several independent episodes during evolution of the land plants.

Phylogenetic analysis using only HIPP sequences from five distinct plant species shows that these proteins are grouped into five clusters. All species except S. moellendorffii have HIPP proteins in all clusters (Fig. 2), suggesting that these duplication events occurred early during angiosperm evolution.

Most HIPP proteins studied so far have been shown to be directly involved in heavy metal homeostasis. In 1999, while describing these proteins for the first time, Dykema et al. indicated their capacity for binding Cu2+, Ni2+ and Zn2+ [18]. Further studies in vitro have demonstrated that HIPPs can also bind to Cd2+, Hg2+ and Pb2+ (via the CXXC core motif of the HMA domain) but not to Ca2+, Mn2+ or Co2+ [16, 19].

The role of these proteins in heavy metal homeostasis and/or detoxification has been demonstrated by several independent studies. Over-expression of AtHIPP06 and AtHIPP26 conferred Cd tolerance to transgenic plants, and the triple knockout mutant AtHIPP20/21/22 was more sensitive to Cd than wild-type Arabidopsis plants were [15, 16, 19]. This capacity was also tested using heterologous expression methods; AtHIPP06 conferred improved Cd tolerance to wild-type yeast [24], while AtHIPP20, AtHIPP22, AtHIPP26 and AtHIPP27 conferred Cd tolerance to the Cd-hypersensitive yeast mutant ycf1. Remarkably, similar results were not observed for Cu2+ or Zn2+ tolerance, suggesting that these proteins may act specifically in Cd homeostasis [15].

The main hypothesis for the mechanism of cadmium homeostasis is that HIPPs protect the plant by trapping free Cd2+ in the cytosol, thus preventing the ion from binding to a more essential protein [15, 16]. Unlike other metallic ions, cadmium is not an element that is needed for the growth of most plants and animals, although it has similar properties to zinc. Indeed, the toxicity of Cd is a direct consequence of its chemical similarity to other divalent ions such as Zn2+ and Ca2+ [25]. It is likely that such a similarity allows Cd ions to replace essential ions and therefore interfere with Ca- or Zn-dependent processes, including protection against oxidative stress [25, 26]. In yeast, Cd2+ toxicity occurs when Cd2+ replaces Cu+ in the HMA domain of ATX1 and is transported to the Golgi, where it binds to Cu-binding sites in other proteins [27].

The transcription of many HIPPs is altered in response to heavy metal stresses, which indicates that HIPPs may be involved in homeostasis of these elements. For example, AtHIPP06 transcription was induced by Cd, Hg, Fe and Cu, while AtHIPP26 transcription was induced by Cd and Zn but not Fe or Cu [16, 19]. Because Cd induces very different HIPPs, we searched for other Cd-responsive genes using the Genevestigator tool. We found new genes that were up-regulated in the roots of Cd-treated plants (AtHIPP05, AtHIPP13, AtHIPP14, AtHIPP39, AtHIPP43, OsHIPP14 and OsHIPP44) that may also be involved in Cd homeostasis (Fig. S6).

However, genes that are not induced by metal excess may also be directly involved in heavy metal homeostasis, especially if their function is to help cells to cope with a metal deficit rather than a metal excess, as in the case of AtCCH and its paired Cu transporting ATPase responsive-to-antagonist1 (AtRAN1) [28]. AtHIPP32, OsHIPP18, OsHIPP23 and OsHIPP38 were down-regulated by Cd treatment (Fig. S6). The genes that respond to Cd belong to all five clusters of HIPPs, indicating that this may be a recurrent function within this group.

In addition to the simple function of chelating metals in the cytosol, HIPPs may act in a more elaborate mechanism. In 2009, Barth et al. identified nuclear HIPPs that function in the transcriptional response to cold [14]. They demonstrated a direct interaction between the HMA domain of AtHIPP26 and the zinc finger homeodomain of the drought stress-related transcription factor AtHB29 using a yeast two-hybrid approach and GST pull-down assays. Proteins from the same cluster (AtHIPP20, 21, 22, 23, 24, 27 and 30) were also able to interact with this same transcription factor (AtHB29), illustrating that additional HIPPs may play a role in transcriptional regulation [14].

A similar mechanism has already been described in the bacteria Enterococcus hirae, in which Cu+ of the metallochaperone CopZ displaces Zn2+ from the zinc finger of the repressor CopY. Zn2+ is essential for binding of CopY to the DNA, and, in the absence of this ion, CopY releases the promoter region of the cop operon, allowing transcription of this operon to proceed [14, 29].

To identify genes that may play related roles in rice, we searched for genes that were induced by cold, salt or drought, which were the same treatments that induced AtHIPP26 expression [14]. By analysis of microarray expression data, we identified OsHIPP41, a gene that possesses 61.6% sequence identity with the coding sequence of AtHIPP26. Quantitative RT–PCR analysis confirmed that OsHIPP41 is highly expressed in response to drought and cold (Fig. 4).

Although most metallochaperones are predicted to be cytosolic proteins that are involved in the storage and delivery of heavy metals, some have roles that require their presence in specific cell compartments. For example, Cox17, the metallochaperone responsible for delivering Cu to the respiratory chain, is found in the cytosol and the mitochondria [10]. Similarly, CCS delivers Cu for three plant Cu/Zn superoxide dismutase isoforms, each present in a different cell compartment: CSD1 in the cytosol, CSD2 in chloroplasts, and CSD3 in peroxisomes [30]. Using heterologous systems to express fusion proteins with fluorescent reporters (GFP or RFP), HIPPs were previously detected in the cytosol, associated with membranes, and in the cell nucleus [14, 16, 19, 21]. In the present study, we used a homologous system (rice protoplasts) to express a rice protein tagged with RFP. Protoplasts were obtained from tissues in which the natural proteins are expressed and where possible associated proteins are present. The protein OsHIPP41, which has a composition and expression pattern similar to the exclusively nuclear protein AtHIPP26, was visualized in the cell nucleus and in the cytosol. The same distribution was observed for OsHIPP21.

For many years, copper has been used as an important component in many chemical pesticides and antibiotics, with known antimicrobial activity [31]. It is also an essential element in the battle between plants and pathogens. An elegant example of this battle and co-evolution is found in the interaction between rice and Xanthomonas oryzae pv oryzae (Xoo), the agent of bacterial blight disease. Xoo is sensitive to Cu; however, it secretes an effector that is able to activate transcription of the rice susceptibility gene Xa13. This protein (Xa13), together with the copper transporters COPT1 and COPT5, promotes removal of Cu from xylem vessels of the infected plant, facilitating the spread of Xoo through the plant vessels. Plants with recessive alleles of Xa13, that encode a mutated variation of this protein, are more resistant to Xoo infection [32]. HIPPs may also be involved in a similar mechanism in plant–pathogen interactions. In previous studies, OsHIPP05 was described as a susceptibility factor for fungal infections that slows plant defense responses to a specific pathogen [33, 34]. It is know that deletion of the proline-rich motifs of OsHIPP05 (also designated Pi21; shown in Fig. 6), confers durable resistance to rice blast disease [34]. A recent study indicated that the wild-type form of this gene enhanced the movement of M. oryzae infection hyphae in host mesophyll cells. Nakao et al. demonstrated that M. oryzae hyphae were able to grow in lines of the non-host plant Arabidopsis when this plant over-expresses OsHIPP05, clearly indicating that this protein (with its proline-rich motifs intact) is essential for the fungal infection [33].

Figure 6.

Plant metallochaperones. Structural representation of the proteins showing their conserved domains based on MEME 5.05 results. Red dot, HMA domain; blue box, isoprenylation motif; pink box, glycine-rich repetitions; yellow box, proline-rich motif.

Protein domains that bind proline-rich motifs are frequently involved in signaling events. The proline-rich regions, which are present in most HIPPs (Figs 3 and 6), tend to adopt a secondary structure nominated polyproline helix. When located in the solvent-exposed N– or C–terminus of a globular protein, the helix forms a structure described as ‘sticky arms’ that is responsible for interaction with other proteins [35, 36].

Analysis of microarray data (Figs S8 and S9) indicated that many HIPP genes are up- or down-regulated in response to various micro-organisms, and thus may also be involved in the interaction between pathogens and plants.

Metallic ions are necessary for many biological functions, from basal metabolism to defenses against environmental stresses. The known metallochaperones play roles in intracellular transport of these elements; however, this family of proteins may also have developed a more fine-tuned function in plants. The HIPPs possess a conserved HMA domain for metal transport but vary in their ‘carboxylic arm’, which is a flexible structure with different protein-binding components that may allow precise interactions (Fig. 6). Thus, each HIPP may participate in protein complexes in specific tissues and situations.

The diversity of expression patterns presented by HIPP genes suggests that HIPPs may be involved in various roles in plant development and defense responses. However, the way in which these proteins act and their protein partners have yet to be identified.

In the present work, we have used phylogenetic analysis to ascertain the evolutionary history of HIPPs, which are involved in many different processes in plants, including the response to environmental stressors such as heavy metal excess, cold and drought, and also in plant–pathogen interactions. This broad functionality suggests that HIPPs play important roles in the responses of plants to the ever-changing environment.

Experimental procedures

Search for and identification of HIPP genes

We searched for putative HIPP genes in the National Center for Biotechnology Information (NCBI, https://www.ncbi.nlm.nih.gov/), PHYTOZOME (http://www.phytozome.net/) and GRAMENE (http://www.gramene.org/) databases using BLAST methodology [36a] and previously identified HIPP sequences as queries. From the obtained sequences, only those that contained a heavy metal-associated domain (HMA, pfam00403.6) and a C–terminal isoprenylation motif (CaaX) were identified as putative HIPPs. The presence of the isoprenylation motif, which was not characterized in the databases, was confirmed using PrePS [37].

Using the MEME suite (Multiple EM for Motif Elicitation, version 5.05) [38], the architecture and the most highly conserved motifs were identified in the full-length amino acid sequences of all HIPP proteins of rice and Arabidopsis.

Phylogenetic analysis

Phylogenetic analysis was performed after protein sequence alignment using both the neighbor-joining and Bayesian methods. The alignments were performed using MUSCLE [39] and implemented in MEGA 5.1 [40]. We used only the HMA domain to perform the phylogenetic analyses, as all metallochaperones possess this domain. The neighbor-joining analyses were performed in MEGA 5.1 [40] using a p–distance parameter with 2000 bootstrap replications. The Bayesian analyses were performed using beast1.7 software [41]. The model of protein evolution used in this analysis was the JTT model for protein matrix substitution [41a]. The Yule process was selected as a tree prior to Bayesian analysis, and 10 000 000 generations were performed using Markov chain Monte Carlo (MCMC) algorithms. The trees were visualized and edited using figtree version 1.3.1 software (http://tree.bio.ed.ac.uk/software/figtree/).

In silico expression analysis

The expression analysis of rice and Arabidopsis HIPP genes was performed in silico using RiceXPro [42] and Genevestigator [22], which allowed us to infer the expression patterns of genes during plant development and in response to diverse conditions.

An alternative approach was used to identify possible abiotic stress-responsive HIPPs. The raw microarray data from the GSE6901 dataset was obtained from GEO (http://www.ncbi.nlm.nih.gov/geo/). LIMMA was used to normalize raw data, through the robust multichip average (RMA) method [43], and identify genes that were significantly up- or down- regulated [44]. The significance was determined using the LIMMA moderated t test with  0.05.

Biological assays

For the expression analysis of rice HIPP genes in response to abiotic stress, both cold and drought conditions were tested. Plants were cultivated in a growth chamber with supplemental lighting (8 h dark/16 h light; 150 μmol·m−2·s−1) at 28 °C.

For the cold stress experiment, 2-week-old seedlings were maintained at 4 °C for 24 h, and the same number of plants was kept at 28 °C as control. After that period, plants from both conditions were immediately harvested and stored in liquid nitrogen for RNA extraction. The drought experiment was performed using seedlings at the four-leaf stage that were split into two treatments: drought stress and control. Control seedlings continued to receive a normal supply of water, whereas water was withheld from stressed seedlings for 15 days. After this period, plant material was collected from both treatments and stored in liquid nitrogen.

For the Cd experiment, 2-week-old seedlings cultivated in a low ionic-strength hydroponic medium [45] were divided into control and experimental groups. For the aluminum treatment, plants were cultivated with 450 μm AlCl3 for 8 h. For the cadmium treatment, plants were cultivated in medium supplemented with 60 μm CdCl2 for 6, 12 and 24 h.

Quantitative RT–PCR analysis

Real-time PCR experiments were performed using cDNA synthesized from total RNA purified using Trizol (Invitrogen, Carlsbad, CA, USA) and treated with RQ1 DNase-free RNase (Promega, Madison, WI, USA). The cDNAs were obtained using the M–MLV reverse transcriptase system and a poly T primer (Promega). After synthesis, the cDNAs were diluted 10–100-fold in sterile water for use in PCR reactions.

In total, eight genes were analyzed: four HIPPs (OsHIPP21, OsHIPP28, OsHIPP41 and OsHIPP50), which were chosen based their inferred expression pattern on the basis of in silico analyses, and four control genes (OsFDH, OsNSL, OseF1α1 and OsGAPDH), which were used to normalize the amount of mRNA present in each sample as described previously [46]. The following primer pairs were designed to produce DNA fragments ranging from 100 to 250 bp: OsHIPP21_RT (forward 5′-CCGGTTGATGAGAAGAAGGA-3′, reverse 5′-ATCTGGCAGCATGATGGACT-3′); OsHIPP28_RT (forward 5′-TCGGTCGAGACTGATCTCCT-3′, reverse 5′-GTCGGCCTTATCCTCATCTG-3′); OsHIPP41_RT (forward 5′-AGCTTGCCTCCGCCTTCAGC-3, reverse 5′-ATGGCATGGCCGCATGGGTG-3′); OsFDH_RT (forward 5′-TTCCAATGCATTCAAAGCTG-3′, reverse 5′-CAAAATCAGCTGGTGCTTCTC-3′); OsHIPP50_RT (forward 5′-TTAGCCCTTACACCGCCTAC-3′, reverse 5′-CATGCTTCATCACCAACCTC-3′); OseF1α1_RT (forward 5′-5′TTTCACTCTTGGTGTGAAGCAGAT-3′, reverse 5′-GACTTCCTTCACGATTTCATCGTAA-3′); OsNSF_RT (forward 5′-CGCTAAACCAGGCTGTTGAT-3′, reverse 5′-GCCCCTCACTGTCAAAAAGA-3′) and OsGAPDH_RT (forward 5′-GGGCTGCTAGCTTCAACATC-3′, reverse 5′-TTGATTGCAGCCTTGATCTG-3).

We performed four experimental repetitions plus four technical replicates using an Applied Biosystems (Foster City, CA, USA) StepOnePlus real-time PCR system with SYBR Green fluorescence detection (Molecular Probes, Eugene, OR, USA). The expression data analyses were performed after comparative quantification of the amplified products using the math formula method [47, 48]. Statistical analysis was performed by Student's t test to compare pairwise differences in expression. Values were considered to be significantly different for < 0.05.

Subcellular localization

Both in silico and in vivo approaches were utilized to predict the subcellular localization of HIPP proteins. For in silico analysis, three programs with different algorithms for localization prediction were used: CELLO version 2.5 [23], PredSL [49] and TargetP [50]. The in vivo visualization was performed in rice protoplasts using OsHIPP21 and OsHIPP41 proteins tagged with RFP at their N–termini. The reporter proteins were obtained by introducing the amplified cDNA sequences (cloned into the pENTR vector) into the p2RGW7 plasmid (Invitrogen, Carlsbad, CA, USA) using Gateway technology [51]. OsHIPP21 cDNA cloning was performed using the HIPP21F (5′-CACCTTTGGAGAAAACATGG-3′) and HIPP21R (5′- CTACATCTGGCAGCATGATG-3′) primers, and the cloning of OsHIPP41 was performed using the HIPP41F (5′-CACCATGGGCGTGGACACAT-3′) and HIPP41R (5′-CTACATGACGGCGCAGGA-3′) primers. The resulting vectors were used for protoplast transformation.

Protoplast isolation and transfection was performed as described by Tao et al. [52] with modifications. Briefly, transformed protoplasts were incubated in the dark for approximately 16 h at 28 °C before imaging. Fluorescence microscopy was performed using an Olympus (Tokyo, Japan) FluoView 1000 confocal laser scanning microscope. In complementary analyses, the cell nucleus was visualized using 4′,6-diamidino-2-phenylindole (DAPI) nucleic acid stain (0.1 mg·mL−1).

Acknowledgements

This study was supported by the Brazilian National Council of Technological and Scientific Development (CNPq).

Ancillary