Single-Stranded DNA-Binding Protein Whirly1 in Barley Leaves Is Located in Plastids and the Nucleus of the Same Cell  

This article concerns the intriguing protein Whirly1 (Why1) that belongs to a small family of single-stranded DNA-binding proteins and has been described to have functions in the nucleus (Desveaux et al., 2002; Yoo et al., 2007). In contrast, in vitro import assays with isolated organelles and transient expression of a fusion construct with the gfp gene revealed that the protein is translocated into plastids (Krause et al., 2005). In this article, specific antibodies directed toward the Why1 protein of barley (Hordeum vulgare) were used to analyze the subcellular location of the native protein. The single-stranded DNA-binding factor Why1 belongs to a small protein family found mainly in land plants. Although most plant species have two Why proteins, Arabidopsis (Arabidopsis thaliana) has three of them (Desveaux et al., 2005; Krause et al., 2005). The three proteins share the putative DNA-binding domain KGKAAL, which is highly conserved in all Why proteins identified so far (Desveaux et al., 2005). The first member of the Why family to be identified was p24, which was later renamed StWhy1. StWhy1 was described as the DNA-binding component of the transcriptional activator PBF-2, which mediates elicitor-induced gene expression of the pathogenesis-related gene PR-10a of potato (Solanum tuberosum; Desveaux et al., 2000). Unlike other transcriptional activators that have been identified, StWhy1 was shown to bind to single-stranded DNA. Electrophoretic mobility shift assays indicated that StWhy1 binds to the inverted repeat sequence of the elicitor response element (ERE) of the PR10a gene of potato (Desveaux et al., 2000). Crystallographic analyses revealed that StWhy1 forms homotetrameres that have high preference for single-stranded DNA. It has been proposed accordingly that the protein may bind to melted promoter regions and may thus modulate transcription (Desveaux et al., 2002). Recent results obtained with Arabidopsis showed that Why1 also binds to single-stranded telomeric DNA and appears to modulate telomere length homeostasis by inhibiting the action of telomerase (Yoo et al., 2007). Though the Why1 protein fulfills different functions in the nucleus, computer-based analyses predicted its targeting to plastids (Desveaux et al., 2005; Krause et al., 2005; Schwacke et al., 2007). Indeed, AtWhy1 is imported into plastids, as demonstrated by in vitro import assays with isolated organelles and by transient expression of a fusion construct of the AtWhy1 gene and the gfp gene (Krause et al., 2005). Although AtWhy3 was shown to be targeted to plastids, too, AtWhy2 was shown to be targeted to mitochondria (Krause et al., 2005). Recently, it has been shown that AtWhy2 is associated with mitochondrial DNA and causes the development of dysfunctional mitochondria when it is overexpressed (Marechal et al., 2008). The amino acid sequence of Why1. The HvWhy1 sequence is compared to the sequences of StWhy1 and AtWhy1. An alignment of the mature protein sequence of Why1 from barley with the complete protein sequences of Why1 from potato and Arabidopsis was generated using the T-Coffee program (www.bioinformatics.nl/tools/t_coffee.html). Hv, Barley; At, Arabidopsis; St, potato; Cons, consensus sequence *; PTP, chloroP-predicted target peptide; PTD, putative transactivation domain (underlined); pNLS, potential nuclear localization signal; PAD, putative autoregulatory domain. The arrowheads indicate the position of the chloroP-predicted target peptide cleavage sites. Why domain, conserved ssDNA-binding domain is shaded in gray. Oligopeptides for antibodies in barley are boxed. Immunological detection of HvWhy1 in protein fractions derived from plastids and nuclei, respectively. A, Scheme depicting the segments I, II, and III excised from barley primary foliage leaves collected at 5 or 7 d after sowing. B, Plastidal and nuclear proteins were prepared from the same three leaf segments (I, II, and III). The blot was immunodecorated with the α-HvWhy1-P2 antibody (Fig. 1), followed by immunodecoration with an antibody specific for the cytochrome b 559 apoprotein A (9.5 kD) and an antibody specific for histone H2B (14 kD), respectively. C, Immunoblot analysis of plastid, stroma, and thylakoid membrane proteins. To show equal loading, a part of the Coomassie Blue-stained gel is shown. The specificity of the immunoreactions was further tested by comparison of the immunodetection pictures before and after preincubation of the antibody with either oligopeptide2 or bovine serum albumin as specific or unspecific competitor, respectively. In both cases total foliar protein extracts as well as protein extracts from purified chloroplasts and nuclei were electrophoretically separated and blotted. When the oligopeptide was added, the signal of the 25-kD HvWhy1 protein was not detectable, whereas it was still detectable when the same amount of bovine serum albumin was used instead of the oligopeptide (Supplemental Fig. S1). To gain insight into the distribution of Why1 within the plastid, stroma and membrane fractions were prepared from chloroplasts. By immunoblot analysis the major part of Why1 was detected in the stroma while a minor part was found in the membrane fraction (Fig. 2C). Immunohistochemical detection of Why1 in plastids and nuclei. Sections from segments of barley flag leaves (A–C) and from leaf segment II (D and E) were immunodecorated with the α-HvWhy1-P1 antibody. Labeling was obtained in nuclei (n) and plastids (p). A, Semithin section after immunogold labeling and silver enhancement as seen by light microscopy; scale bar is 10 μm. B, Fluorescence micrograph after DAPI staining of the same specimen as shown in A; scale bar is 10 μm. C, Electron micrograph depicting a section of a chloroplast after immunogold labeling silver enhancement; scale bar is 1 μm. D and E, Electron micrographs depicting sections of a nucleus (n) and a chloroplast (p), respectively, after immunogold labeling; scale bars are 1 μm (D) and 0.1 μm (E). Previously it has been shown that the AtWhy1-GFP fusion protein was exclusively targeted to the plastids (Krause et al., 2005). Unexpectedly, the transient transformation assays did not give any evidence for localization in the nucleus. This could be due to the high molecular mass of the fusion protein (68 kD). When the sequence encoding the plastid target peptide was removed from the construct, indeed GFP fluorescence was detected in the cytoplasm and in the nucleus after biolistic transformation of onion (Allium cepa) epidermal cells (data not shown). HvWhy1 forms homooligomers in the nucleus. Onion epidermal cells were transiently transformed with constructs expressing HvWhy1 and AtWhy1 without the PTP as well as with the full-length AtWhy1. Constructs were fused to either c-myc-YFPn173 or HA-YFPc155 and vice versa. The AtWhy1 fused to full-length GFP and the empty vectors were used as controls. All constructs were under the control of the 35S promoter. Fluorescence images are shown above the bright field images that are shown on the bottom row. A, AtWhy1-YFPn173 + AtWhy1-YFPc155. B, ΔPTP-AtWhy1-YFPn173 + ΔPTP-AtWhy1-YFPc155. C, ΔPTP-HvWhy1-YFPn173 + ΔPTP-HvWhy1-YFPc155. D, AtWhy1-GFP. E, AtWhy1-YFPn173 + empty vector. Scale bars are 200 μm (A and E) and 100 μm (B–D). Because the constructs used for these experiments lacked the PTP, no fluorescence was detected in plastids. When the full-length AtWhy1 was, however, fused to full-length GFP, fluorescence was detected in plastids as described previously (Krause et al., 2005; Fig. 4D). When the full-length sequence encoding AtWhy1 was fused instead with the N-terminal part and the C-terminal part of YFP, respectively, no complementation of the YFP fluorescence was achieved (Fig. 4A). When the full-length AtWhy1 was fused with full-length GFP and CFP, both fluorescence emissions were detected in plastids, respectively, but no overlay of fluorescence signals was observed (data not shown). This also suggests that Why1 does not form homooligomers when it is located in plastids. To investigate whether under conditions not showing a fluorescence complementation (Fig. 4, A and E) the constructs have been expressed in the cells, western-blot analysis with a GFP-specific antibody were performed. Immunoreactions confirmed that the N-terminal as well as the C-terminal construct with HvWhy1 (Supplemental Fig. S3) were both expressed in the cells. Homooligomerization in the nucleus, as here shown by bimolecular fluorescence complementation, is in accordance with the results of the structural analysis (Desveaux et al., 2002) and the proposed function of Why1 as single-stranded DNA-binding factor involved in regulation of transcription (Desveaux et al., 2004, 2005). Nevertheless, Why1 has been detected in the proteome of the transcriptionally active chromosome fraction prepared from chloroplasts of Arabidopsis (Pfalz et al., 2005). This finding is in accordance with the immunological detection of HvWhy1 in the membrane fraction (Fig. 2C) and by the association of gold particles with thylakoid membranes (Fig. 3C; Supplemental Fig. S2). The transcriptionally active chromosome fraction contains proteins bound to the ptDNA organized in nucleoids being associated to the thylakoid membrane of chloroplasts (Sato, 2001). Its protein composition is rather complex (Krause and Krupinska, 2000; Pfalz et al., 2005) and in addition to the subunits of the RNA polymerases contains proteins involved in posttranscriptional processes of plastid gene expression (Krause, 1999). Why1 in plastids could be involved in these processes. To achieve a coordination between the nucleus and the organelles and vice versa, anterograde and retrograde control mechanisms have developed (Beck, 2005; Nott et al., 2006). Intermediates of the tetrapyrrole biosynthesis, reactive oxygen species, plastid gene expression products, and changes in the redox state of the photosynthetic electron transport chain have been proposed as plastid signals. Putative signal-transducing components involved in plastid-to-nucleus signaling pathways such as the GUN proteins (Susek et al., 1993) and the Executer1 protein (Wagner et al., 2004) were identified by genetic studies. With regard to its dual localization in the nucleus and the plastids, Why1 is an excellent candidate for transducing signals between the plastids and the nucleus. Why1 as a DNA-binding protein in the nucleus as well as in plastids might contribute to the coordination between plastid gene expression and transcription in the nucleus by a still-unknown mechanism. The mechanism of Why1 distribution in the cell with regard to its putative signal-transducing function remains to be determined. Insight into the biological significance of its dual location is expected from investigations on transgenic plants with altered levels of Why1 in the nucleus and in plastids, respectively. Sequence data from this article can be found in the GenBank/EMBL data libraries under accession number BF6136. The following materials are available in the online version of this article. Supplemental Figure S1. Test of specificity of the α-HvWhy1-P2 antibody used for immunoblot analysis shown in Figure 1. Supplemental Figure S2. Overview electron micrograph showing the immunogold labeling of HvWhy1 in chloroplasts and in the nucleus. Supplemental Figure S3. Immunoblot analysis of extracts from onion tissue used for transient transformation assays described in Figure 3. Supplemental Materials and Methods S1. A supplemental “Materials and Methods” section. We thank Anke Schäfer and Marita Beese for technical assistance. We acknowledge Kirsten Krause (University of Tromsö, Norway) for constructive comments on the manuscript.