•In seed plants, current knowledge concerning embryonic pattern formation by polar auxin transport (PAT) and WUSCHEL-related homeobox (WOX) gene activity is primarily derived from studies on angiosperms, while less is known about these processes in gymnosperms. In view of the differences in their embryogeny, and the fact that somatic embryogenesis is used for mass propagation of conifers, a better understanding of embryo development is vital.
•The expression patterns of PaWOX2 and PaWOX8/9 were followed with quantitative reverse transcription–polymerase chain reaction (qRT-PCR) and in situ hybridization (ISH) during seed and somatic embryo development in Norway spruce (Picea abies), and in somatic embryos treated with the PAT inhibitor N-1-naphthylphthalamic acid (NPA).
•Both PaWOX2 and PaWOX8/9 were highly expressed at the early growth stages of zygotic and somatic embryos, and shared a similar expression pattern over the entire embryo. At later embryo stages, high expression of PaWOX8/9 became restricted to cotyledon primordia, epidermis, procambium and root apical meristem (RAM), which became most evident in NPA-treated somatic embryos, while expression of PaWOX2 was much lower.
•Our results suggest an ancestral role of WOX in seed plant embryo development, and strengthen the proposed connection between PAT, PIN-FORMED (PIN) and WOX in the regulation of embryo patterning in seed plants.
The two major taxa of extant seed plants, gymnosperms and angiosperms, are believed to have diverged from a common ancestor > 300 Myr ago (Stuessy, 2004), although their evolutionary relationship is still a matter of controversy (Chaw et al., 2000; Burleigh & Mathews, 2004). Many aspects of growth and development differ between gymnosperms and angiosperms, including early embryogeny (Raghavan & Sharma, 1995). For example, in gymnosperms, the multicellular haploid female gametophyte both holds egg cells within several archegonia and serves as a nutritive tissue for the growing embryo, similar to the triploid endosperm of angiosperms. Unlike those of most angiosperms, conifer proembryos have a brief period of free nuclear divisions before cellularization of the embryo (Singh, 1978), and embryos do not contain a cell corresponding to the hypophysis, which in the model plant Arabidopsis thaliana has been shown to be important for radicle and root cap formation (Berleth & Chatfield, 2002).
There seems to be a connection between PAT and the spatial separation of WUSCHEL-related homeobox (WOX) transcription factors in the formation of the main body axis in A. thaliana embryos. Haecker et al. (2004) recently identified two members of the WOX gene family, WOX2 and WOX8, as being initially co-expressed in the A. thaliana egg cell and the zygote, and then specifically expressed in the apical and basal cell lineages, respectively, after zygotic division. STIMPY/WOX9, a close homologue of WOX8, is expressed after the first division of the zygote and is required to maintain cell division in the embryo and the suspensor (Wu et al., 2007), suggesting that WOX genes are intrinsic determinants in early asymmetric divisions during A. thaliana embryogenesis. WOX2 and WOX8 act redundantly with MONOPTEROS (MP), an auxin response factor (ARF) involved in the TRANSPORT INHIBITOR RESPONSE 1 (TIR1)-AUXIN (AUX)/INDOLE-3-ACETIC ACID (IAA)-ARF pathway, in PIN1 regulation (Breuninger et al., 2008), while the expression of STIMPY/WOX9 is altered in mp mutants (Haecker et al., 2004). Auxin is thought to modulate the transcription of multiple PIN genes through the TIR1-Aux/IAA-ARF pathway (Sauer et al., 2006; Wenzel et al., 2007). After the establishment of the apical–basal pattern, other members of the gene family, WUSCHEL (WUS) and WOX5, can be detected in the SAM and RAM, respectively, with both genes promoting stem cell identity (Mayer et al., 1998; Sarkar et al., 2007). Other angiosperms show similar WOX expression patterns, including the monocot Zea mays, indicating that the WOX gene family is ancient in the angiosperm lineage (Nardmann et al., 2007; reviewed e.g. by Dodsworth, 2009; Van der Graaff et al., 2009).
Previous WOX gene phylogenies for basal land plants, gymnosperms, and angiosperms support a distribution of genes into three evolutionary lineages with a modern clade (WUS/WOX1-7) restricted to seed plants (Deveaux et al., 2008; Palovaara & Hakman, 2008; Nardmann et al., 2009; Van der Graaff et al., 2009). Recently, Nardmann et al. (2009) identified a single WUS/WOX5 homologue in two nonconiferous gymnosperm genomes, suggesting that distinct WUS and WOX5 genes are restricted to angiosperms and that the genome of the last common gymnosperm/angiosperm ancestor contained a single WUS/WOX5 precursor. Given the importance and proposed connection of PAT and WOX transcription factors during embryo development in angiosperms, and the evolutionary considerations and developmental differences mentioned above, we set out to analyse the expression patterns of a WOX2 (PaWOX2) and a WOX8/9 (PaWOX8/9) homologue, genes that had earlier been identified as being expressed in conifer embryos (Palovaara & Hakman, 2008), during P. abies seed and somatic embryo development, and in somatic embryos treated with the PAT inhibitor N-1-napthylphthalamic acid (NPA).
The results of the present study suggest that PaWOX2 and PaWOX8/9 are involved in P. abies embryo formation and differentiation, and, together with PaPIN1 and PAT (Hakman et al., 2009; Palovaara et al., 2010), act in embryo patterning. Thus, despite changes during seed plant evolution, our results strengthen the proposed connection between PAT, PIN and WOX in regulating seed embryo patterning. Also, the comparison of expression in somatic and zygotic embryos reveals somatic embryogenesis as a useful model system for conifer embryo development.
Materials and Methods
Plant material and somatic embryo culture
One highly embryogenic cell line of Norway spruce (Picea abies (L.) Karst.), which contains a mixture of early staged embryos, cell aggregates and single cells of diverse form and size, recently described by Palovaara & Hakman (2008), was used in this study. ‘Calli’ were collected 1 wk after their subculture onto fresh medium, while somatic embryos of various stages were isolated with the aid of a dissection microscope after being stimulated to develop to maturity by the subculture of calli onto maturation medium as recently described (Palovaara & Hakman, 2008; Hakman et al., 2009). Tissues with a mixture of slightly different staged embryos were isolated 1, 1–2, 2–4, and 4–8 wk, respectively, after subculture. To impair PAT, which is needed for proper embryo development (Hakman et al., 2009; Palovaara et al., 2010), somatic embryos were treated with NPA as previously described (Hakman et al., 2009).
Ovules and developing seeds of P. abies were isolated from seed cones collected from a seed orchard on the island of Öland (55°), Sweden on 9, 23 and 30 June, 7, 14, 21 and 28 July, 4 August and 1 September 2009. Mature seeds were germinated at 25°C under a 12-h photoperiod in a growth chamber and seedlings were harvested after 2 wk and divided into roots, shoot apices, hypocotyls, and cotyledons.
Plant tissues were frozen separately in liquid nitrogen for later RNA isolation and also processed for RNA in situ hybridization (ISH).
Database analysis, isolation and characterization of PaWOX8/9
A potential P. abies WOX8/9 homologue (PaWOX8/9; accession no. GU944670) was isolated by screening the spruce (Release 3.0 (July 11, 2008)) expressed sequence tag (EST) database at The Gene Index (TGI) databases (http://compbio.dfci.harvard.edu/tgi/) using the TBLAST algorithm (Altschul et al., 1997) with a pine WOX8/9 homologue (accession no. DR692518), previously described to be expressed in embryos by Palovaara & Hakman (2008), as a query. Several ESTs from spruce assembled together with DR692518 and based on these results primers were designed from the spruce contig TC72223 and EST DR588627. Rapid ampification of the 3′-cDNA and 5′-cDNA ends was performed using the SMARTTM RACE cDNA amplification kit (Clonetech, Palo Alto, CA, USA) and cDNA from somatic embryos. Resulting PCR products were cloned using the TOPO-TA cloning kit (Invitrogen, Carlsbad, CA, USA) and sequenced by Eurofins MWG Operon in Ebersberg, Germany.
Total RNA was extracted from the plant samples as previously described (Palovaara & Hakman, 2008) and quantified using the NanoDrop 2000 (Thermo Scientific, Wilmington, DE, USA). For cDNA synthesis, 2 μg of RNA was reverse-transcribed with oligo (dT) and random hexamer primers using the iScriptTM cDNA synthesis kit (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s protocol.
qRT-PCRs were performed with the MiniOpticon Real-Time PCR Detection System (Bio-Rad) using gene-specific primers (Supporting Information Table S1). Reactions were performed in a 25-μl volume containing 12.5 μl of iQ SYBR Green Supermix (Bio-Rad), 0.3 μM of both primers and 2 μl of cDNA template. The thermal cycling conditions were as follows: 3 min at 95°C, followed by 40 cycles of 10 s at 95°C, 10 s at 58°C, and 10 s at 72°C. Expression levels were determined as the number of cycles needed for the amplification to reach a threshold fixed in the exponential phase of the PCR reaction (Ct) (Walker, 2002). Absolute quantification of gene expression was performed as described previously (Palovaara & Hakman, 2008). All samples and the standard curve were run in triplicate (n =3) and three assays were performed using three independent samples of each tissue collected. The statistical significance of differences was evaluated by one-way ANOVA, followed by post hoc Tukey’s HSD comparisons, and Student’s t-test using statistica v7.1 (StatSoft, Inc., Tulsa, OK, USA). Differences of P <0.05 were regarded as significant.
Full-length and homeodomain (HD) WUS/WOX sequences from other species were obtained by searching the Genbank Non-Redundant Database with TBLASTN, using the WUS/WOX proteins of A. thaliana as queries. Oryza sativa full-length sequences were derived from its established genome sequence (http://rice.plantbiology.msu.edu) as recently described by Nardmann et al. (2009), and were named, along with the full-length sequences of Populus trichocarpa, according to their homology with members of the A. thaliana WOX family. Accession numbers and genome loci are given in Table S2. Amino acid sequence alignments were generated using Clustal X 2.0 (Thompson et al., 1997), with minor manual adjustments. Phylogenetic reconstructions were performed using the neighbour-joining (NJ) method in paup 4.0 (Swofford, 2000) using default parameters. Estimates of confidence were obtained by bootstrapping (5000 replicates) and a consensus tree was computed using the 50%-majority rule.
RNA in situ hybridization
Chemical fixation and tissue processing for ISH were performed as described by Tahir et al. (2006). Specific primers containing T7 (sense) and SP6 (antisense) sites were used to amplify sequences from nonconserved regions of PaWOX2 and PaWOX8/9, respectively, as well as from the conserved homeobox region of PaWOX2 (Table S1, Fig. S1). The PCR products were then used for in vitro transcription using DIG-11-UTP, as described in the DIG RNA labelling kit (Roche Diagnostics GmbH, Mannheim, Germany). Tissue treatments, pre-hybridization, post-hybridization washes, antibody treatment, and detection of DIG-labelled probes were conducted as described previously (Palovaara et al., 2010). Micrographs were taken with a Leica DMRE and a Zeiss Axiovert 10 microscope equipped with a Leica DC 500 and DFC 320 camera, respectively, while pictures were further processed using Adobe® Photoshop® CS4.
PaWOX8/9 isolation, characterization and phylogenetic analysis
Recently, we isolated and characterized a P. abies WOX2 homologue, PaWOX2, and also investigated the evolutionary relationships between various conifer WOX transcription factor family members and the WUS/WOX sequences from both angiosperms and nonseed plants (Palovaara & Hakman, 2008). In this analysis, we established that the assembled conifer representatives cluster with members of the ancient WOX10/13/14 clade, the WOX8/9/11/12 clade and the modern WUS/WOX1-7 clade, although no WUS or WOX5 conifer representatives were found. We also established that several conifer sequences, all derived from embryo libraries, clustered with WOX8 and WOX9, thus indicating their involvement in embryo patterning as previously shown in angiosperms (see e.g. Haecker et al., 2004; Nardmann et al., 2007; Van der Graaff et al., 2009). Here we set out to isolate and characterize the expression pattern of a WOX8/9 homologue together with that of PaWOX2 during seed and somatic embryo development in P. abies.
A full-length P. abies cDNA sequence, here designated PaWOX8/9 (accession no. GU944670), was isolated. PaWOX8/9 encodes a putative protein of 274 amino acids and contains the regular WUS-homeodomain subtype with 65 amino acids (66 aa in WUS; Mayer et al., 1998; Haecker et al., 2004) (Fig. S2). However, the protein does not contain the WUS box (TLXLFPXX, where X can be any amino acid), a distinctive feature in members of the modern WUS/WOX1-7 clade that seems to act as a transcriptional repressor in plants and is necessary for the maintenance of initial cells in SAMs (Haecker et al., 2004; Ikeda et al., 2009). PaWOX8/9, like the other non-WUS WOX family members, shows variation at this position (Van der Graaff et al., 2009). Alignment of full-length PaWOX8/9 with WOX proteins from different plant species revealed an N-terminal and a C-terminal motif (Deveaux et al., 2008; Mukherjee et al., 2009) conserved in all the retrieved WOX8/9 and WOX8/9/11/12 sequences, respectively (highlighted in red and blue, respectively, in Fig. S2). The function of these motifs is unknown.
PaWOX8/9 clusters with both A. thaliana WOX8 and WOX9 as well as with WOX9 of other plant species in phylogenetic reconstructions of both the full-length WOX proteins and their homeodomains (HDs) (Figs 1, S3, S4). It should be noted that, of the plant species with fully sequenced genomes, only A. thaliana has both a WOX8 and WOX9 protein. The WOX proteins, both full-length and HD, divided into three different orthologous groups, the modern WUS/WOX1-7 clade, the WOX8/9/11/12 clade, and the ancient WOX10/13/14 clade, which is consistent with previous results (Deveaux et al., 2008; Palovaara & Hakman, 2008; Nardmann et al., 2009). Furthermore, the results confirm the previous findings of Nardmann et al. (2009) that Pinus sylvetris WUS, which does not contain the WUS-specific extra tyrosine (Y) between helixes 1 and 2 in the HD, clusters with WOX5 in the full-length phylogenetic reconstruction, while it groups to the WOX1/6 branch in the HD reconstruction (Figs S3, S4).
Quantitative expression profiles of PaWOX2 and PaWOX8/9 during somatic embryo development using qRT-PCR
Absolute qRT-PCR was used to quantify changes of both PaWOX2 and PaWOX8/9 expression in somatic embryos during development (Fig. 2a) (for verification of the method, see Table S3). The two genes showed similar expression levels in samples taken from calli, while, by contrast, the expression of PaWOX8/9 increased significantly in tissue samples taken 1 wk after transfer to the maturation medium, in which more somatic embryos with a better defined structure were present (‘early embryos’) (Hakman et al., 2009), whereas PaWOX2 decreased slightly. The expression profiles of the two genes thereafter showed a similar pattern of decrease in embryos as they matured further into the club-like precotyledonary stage and later into the cotyledonary stage, where cotyledons had just started to form on embryos, although the PaWOX8/9 transcript level was much higher compared with PaWOX2 at all stages. In mature embryos, the expression levels of PaWOX2 and PaWOX8/9 had decreased even more, showing a 9.2- and 6.5-fold decline, respectively, compared with the initial expression in calli. It must be pointed out here that despite the relatively uniform development of the cultured somatic embryos development is not totally synchronized, and that the different samples taken for RNA isolation, in particular from the earlier embryo stages, always contained a mixture of embryo stages with considerable morphological variability among them, as well as other cells besides the embryos. The expression patterns, however, were comparable to those seen for WOX2 and WOX9 genes in microspore-derived Brassica napus embryos (MDEs) (Malik et al., 2007). Also, as with WOX2 and WOX9 in B. napus, PaWOX2 and PaWOX8/9 were expressed at low levels in P. abies seedling tissues (Fig. S5).
The results for PaWOX2 were similar to those we reported previously (Palovaara & Hakman, 2008), although expression was higher than before in calli and in mature somatic embryos. This disparity could be explained by differences in the callus proliferation rate and in callus quality, which changes slightly over time, and/or by the way of sample collection.
Expression profile of PaWOX2 and PaWOX8/9 during somatic embryo development using ISH
ISH was employed to analyse the expression pattern of PaWOX2 and PaWOX8/9 transcripts to the tissue level in somatic embryos during development. Results showed that both PaWOX2 and PaWOX8/9 were transcribed in the aggregates of smaller cells within proliferating calli (Fig. 3a,e) and throughout the early somatic embryo, including its suspensor cells (Fig. 3b,f). As the embryo progressed to the club-like precotyledonary stage, staining of transcripts for PaWOX2 was not very clear but appeared to be slightly more prevalent in the basal part of the embryo proper (Fig. 3c), while transcripts of PaWOX8/9 appeared to localize particularly at the future RAM and cotyledon initiation sites (Fig. 3g), with no transcripts at all in the SAM area. The transcriptional activity of PaWOX2 ceased in the later stages and no hybridization signal could be detected in mature embryos (Fig. 3d). For PaWOX8/9, transcription in cotyledonary-stage embryos became confined to the procambial cells that demarcate the future site of vascular tissue differentiation of the embryo and was seen extending all the way out to the tip of the developing cotyledons (Fig. 3h). Also, PaWOX8/9 transcription activity could be observed in the epidermis and subepidermal cell layers of emerging cotyledons (Fig. 3i), in the epidermal cells extending down to the junction zone between the hypocotyl and the root cap, and in the RAM and the central area of the developing root cap (Fig. 3h,i). No transcripts, however, were seen in the shoot apex or in the epidermis that covers it. In more mature embryos, when the cytohistological zonation of the embryos became clearer, PaWOX8/9 transcripts were detected in the differentiating procambium, running all the way from the cotyledons down through the embryo axis to the RAM (Fig. 3j,k). Also here were PaWOX8/9 transcripts seen in the epidermis down to the junction zone at the root area (Fig. 3l), and to the zone that gives rise to the embryonic cortex.
Taken together, both the ISH results and the qRT-PCR data indicate a higher transcription level of PaWOX8/9 than PaWOX2 in developing somatic embryos, and that gene expression declines as the embryos mature. However, qRT-PCR showed much lower transcript abundance of PaWOX8/9 in mature embryos than the ISH results seem to indicate. Previously, we identified a number of WOX8/9-like sequences from conifers (Palovaara & Hakman, 2008) and it is possible that our PaWOX8/9 in situ probe detected other WOX8/9 homologous genes in the P. abies somatic embryos.
Expression level of both PaWOX2 and PaWOX8/9 increased in NPA-treated embryos
Recently, we showed that PaWOX2 was up-regulated in precotyledonary somatic embryos treated with the PAT inhibitor NPA (Palovaara & Hakman, 2009), and that NPA treatment caused both an expanded procambium and a broader RAM in embryos (Hakman et al., 2009). As WOX expression appears to be associated with differentiation of such cells and structures in conifer somatic embryos, we wanted to examine further the expression of both genes in precotyledonary embryos but also in mature embryos that had attained the cup- or pin-like morphology after NPA treatment as previously described (Hakman et al., 2009). Compared with untreated embryos, NPA treatment appeared to significantly up-regulate the gene activity of both WOX genes in precotyledonary embryos (3.0-fold and 1.6-fold for PaWOX2 and PaWOX8/9, respectively), as revealed by qRT-PCR (cf. Fig. 2a,b). While no significant difference compared with untreated embryos could be seen in mature embryos for PaWOX2, expression was significantly up-regulated for PaWOX8/9 (3.2-fold) (cf. Fig. 2a,b). Also, PaWOX8/9 expression was significantly higher than PaWOX2 expression in embryos of both developmental stages.
We next asked if there was any difference in PaWOX2 and PaWOX8/9 expression patterns in NPA-treated somatic embryos compared with the untreated embryos, or if the higher expression of the WOX genes seen using qRT-PCR only reflected the exaggerated proliferation of vascular tissue in embryos treated with NPA. ISH results showed that both PaWOX2 and PaWOX8/9 were transcribed throughout the embryo proper in the precotyledonary-stage embryos, perhaps with a slightly stronger signal for PaWOX8/9 (Fig. 4a,e). As the embryos developed further, the PaWOX8/9 signal was very strong in the emerging cotyledons and became restricted to the procambial cells of the embryo and to the RAM (Fig. 4f,g), while PaWOX2 gave only a very weak signal in the same tissues (Fig. 4b–d). In addition, PaWOX8/9 transcripts were detected in the upper lateral derivatives of the root cap (Fig. 4f,g), a pattern that became more evident in the more developed embryos, where PaWOX8/9 also accumulated in the vascular tissue of the malformed cotyledons of cup-like embryos and in the epidermis (Fig. 4h). In embryos that had grown into the pin-like phenotype, the PaWOX8/9 transcription pattern was similar, but with a much stronger staining reaction compared with that of untreated and morphologically normal-looking embryos (Fig. 4i); that is, the localization of PaWOX8/9 to the vasculature was clear and high expression of PaWOX8/9 was seen in the expanded root area.
Thus, PaWOX2 and PaWOX8/9 transcription patterns overlapped to some degree in NPA-treated somatic embryos, but the expression signal was much higher for PaWOX8/9 in all stages, similar to the qRT-PCR results. Also, the high transcription activity of PaWOX8/9 in NPA-treated embryos was localized to cotyledon primordia, epidermis, differentiating vascular tissue and root area, substantiating the results found in normally developing embryos.
Expression profile of PaWOX2 and PaWOX8/9 during seed development obtained using ISH
One interesting feature of the WOX genes in A. thaliana is that WOX2 and WOX8 are initially co-expressed in the zygote, but become restricted to the apical and basal domains of the embryo, respectively, during development (Haecker et al., 2004; Van der Graaff et al., 2009). Because of the major differences during early embryogeny in angiosperms and conifers, we wanted to examine the expression patterns of WOX2 and WOX8/9 during P. abies seed development, but also to compare the expression patterns in somatic and zygotic embryos.
As far as could be determined, fertilization had not yet occurred at the first collection date, but, because the time of fertilization varies from year to year in a given locality as well as between various trees, some egg cells may still have gone through syngamy. As there is no clear fusion of gametes in conifers, it is also difficult to distinguish the zygote from the egg cell (Allen, 1946; Runions & Owens, 1999), and the cell with its centrally located nucleus is therefore here referred to as the egg cell. The female gametophytes in early June contained multiple archegonia with jacket cells enclosing the large egg cell. Transcripts of both PaWOX2 and PaWOX8/9 appeared to be present in the egg cell but also in the surrounding archegonial jacket cells (Fig. 5a,f).
After fertilization, the zygote of conifers undergoes a sequence of divisions without cytokinesis to enter a free nuclear phase with the nuclei situated at the archegonial base opposite the micropyle and surrounded by a dense ‘neocytoplasm’ believed to be newly formed cytoplasm (Runions & Owens, 1999). The free nuclei are arranged into two planes with four nuclei in each which, after cellularization, become separated from each other, and, after internal divisions, a four-tiered proembryo is formed. In the present study, the proembryo included stages from the two-nucleate stage up to the stage in which the primary suspensor cells elongated and the embryo broke through the archegonial base. Both PaWOX2 and PaWOX8/9 gave intense staining reactions in the proembryo after ISH while the egg cell only stained weakly (Fig. 5b,g), which could be a consequence of egg cytoplasm degeneration which starts at proembryo development (Allen, 1946).
At the early embryo stage, high PaWOX2 and PaWOX8/9 transcript levels were detected both in the embryo proper and perhaps to a lesser extent in the suspensor (Fig. 5c,i), which might be a result of the highly vacuolated nature of the cells. Later, transcripts of PaWOX2 became confined to the middle part of the embryo proper (Fig. 5d) and later still, at the histodifferentiation of embryos, to the actively dividing procambial cells, where the hybridization signal was very weak (Fig. 5e). PaWOX2 transcriptional activity thereafter ceased in more developed embryos (not shown). The transcription pattern for PaWOX2 differed somewhat compared with somatic embryos, probably because of the clearer cytohistological zonation in zygotic embryos. For PaWOX8/9, the transcription pattern followed that seen in somatic embryos more closely, with an intense staining reaction localized particularly to the area of RAM formation and to the position of cotyledon initiation and procambium, which became even clearer in more mature embryos, while no signal was seen in SAM (Fig. 5i,j).
Localization of WOX gene expression in the shoot apex of P. abies seedlings using ISH
As PaWOX2 and PaWOX8/9 were both expressed at low levels in P. abies seedling tissues, we used ISH to localize their transcription patterns in the same tissues (root, hypocotyl, apical shoot and cotyledon). No transcripts were detected in any of the seedling tissues (not shown) when the probe designed against PaWOX2 (Table S1) was used. However, when we used the highly conserved WUS/WOX homeobox region from PaWOX2 as a probe, a hybridization signal could be detected in a diffuse pattern throughout the shoot apex, with the signal being somewhat stronger in the outer cell layers of the shoot apex and in needle primordia (Fig. 6a). For PaWOX8/9, transcription activity was somewhat higher and specifically detected in the outer cell layers of the shoot apex and in the needle primordia, where some cells were deeply stained, as seen in serial sections (Fig. 6b,c). No transcripts were observed in any other part of the seedling, the hypocotyl, the cotyledons, or the root after ISH with these probes (not shown). Alignment of the nucleotide sequences showed that the PaWOX2 homeobox probe sequence is similar to the homeobox sequences of WUS and WOX5 from other species, including gymnosperms (Fig. S1), and thus our results suggest the possibility of transcription also of other WOX genes than PaWOX8/9 in the shoot apex.
In angiosperms, the establishment of distinct transcriptional domains during early embryo development leads to correct apical–basal axis formation, which is essential for optimal post-embryonic growth. In gymnosperms, a better understanding of this process is essential, not only from an evolutionary perspective, but also because of the potential of somatic embryogenesis as a propagation method in forestry and as a model system for basic studies of embryo development (Hakman, 1985; Hakman et al., 1985). A fascinating aspect of somatic embryogenesis is the similarity between somatic and zygotic embryo developmental patterns, even though the physical and chemical environments differ. Conifer somatic embryos develop into polar structures with a well-defined embryo proper and suspensor, similar to zygotic embryos, without the influence of the surrounding tissue of the archegonium and the female gametophyte. The early stages of zygotic proembryo development, however, are not easily recognized in somatic embryo development. In this study, we show that similar expression patterns of PaWOX2 and PaWOX8/9 occur in P. abies embryos, regardless of their origin.
WOX genes are expressed in both zygotic and somatic embryos during early development
PaWOX2 and PaWOX8/9 were both expressed in the egg cell/zygote (Fig. 5a,f), similar to WOX2 and WOX8 in A. thaliana (Haecker et al., 2004) and WOX2A in Z. mays (Nardmann et al., 2007), and at high levels in proembryos (Fig. 5b,g). High transcript levels of both genes were also seen in the small rapidly dividing cells, present as single cells or in cell aggregates, in the calli (Fig. 3a,e) which presumably gave rise to the somatic embryos seen in the cultures (Hakman et al., 1987). At early embryo stage, which starts when primary suspensor cells elongate and ends at meristem differentiation, high PaWOX2 and PaWOX8/9 expression was seen in both the embryo proper but also in the suspensor cells of both somatic and zygotic embryos (Figs 3b,f, 5c,h), which may reflect the importance of both genes during early P. abies embryo development. This is further suggested by the up-regulation of PaWOX8/9 during early somatic embryo development (Fig. 2). The expression of PaWOX8/9 throughout the early embryo is similar to that seen for STIMPY/WOX9 in the proembryo of A. thaliana (Wu et al., 2007).
In A. thaliana, the wox2 mutation confers division defects in the apical cell lineage of the embryo, while embryo arrest is seen in wox8 stimpy/wox9 double mutants, with aberrant divisions in both embryo and suspensor cell lineages (Breuninger et al., 2008). As WOX2 restores apical cell fate in wox8 stimpy/wox9 mutant embryos, its expression is a downstream function of WOX8 and STIMPY/WOX9.
The gene activity of both PaWOX2 and PaWOX8/9 in P. abies suspensor cells (Fig. 3b,c,f,g) could perhaps reflect their higher mitotic activity compared with that of A. thaliana, in which the suspensor only consists of a few cells that rarely divide. Also, the archegonial jacket cells (Fig. 5a,f), which are implicated in the nutrition of the egg/proembryo (Hakman & Oliviusson, 2002), expressed both WOX genes. The jacket cells are derived from cells surrounding the archegonial initials (Singh, 1978) and mitotic divisions are common also here. Thus, high WOX expression could be linked to cell proliferation as well as to cell specification. Indeed, PaWOX2 is highly expressed in a somatic cell line with a very high proliferation rate, but with poor embryo maturation capacity, suggesting its involvement in regulating cell proliferation and/or differentiation (Palovaara & Hakman, 2008).
WOX2 and WOX8/9 may act together in conifer embryo patterning
In later stages of P. abies embryo development, the transcription pattern of PaWOX8/9 share many similarities with that of STIMPY/WOX9 in A. thaliana embryos (Figs 3g–l, 5i,j) (Wu et al., 2007). After being expressed throughout the A. thaliana proembryo, STIMPY/WOX9 becomes restricted to the cotyledon primordia and the outer cell layers, including the epidermis, in the basal half of the embryo, much like what was seen for PaWOX8/9 in P. abies embryos (Figs 3g–j, 5i). As STIMPY/WOX9 seems to promote, partially redundantly with WOX8, cell division during A. thaliana embryo development (Wu et al., 2007), expression patterns here indicate a similar role for PaWOX8/9 in P. abies embryos.
We recently showed that PaPIN1 is transcribed in the differentiating procambium of P. abies somatic embryos (Palovaara et al., 2010), and that correct auxin transport is needed for proper cotyledon and SAM and RAM initiation and development (Hakman et al., 2009). Here, high levels of PaWOX8/9 transcripts overlapped with a weaker PaWOX2 transcription pattern in the differentiating procambium of embryos, which was most clearly seen after NPA treatment of embryos (Fig. 4b–d,f). The up-regulation of both genes in NPA-treated precotyledonary embryos (Fig. 2) indicates that PaWOX8/9 and PaWOX2 may act together in procambium initiation and/or maintenance together with PaPIN1, which is also up-regulated in such embryos (Hakman et al., 2009; Palovaara et al., 2010).
Although WOX2, WOX8 and STIMPY/WOX9 are not expressed in the procambium of A. thaliana embryos (Haecker et al., 2004; Wu et al., 2007), it was recently shown that WOX2 and WOX8 act redundantly with MP to promote PIN1 expression in the cotyledonary vasculature, while also both being involved in cotyledon separation (Wu et al., 2007; Breuninger et al., 2008). Presumably, a PIN1-mediated auxin maximum is required to regulate cotyledon initiation and separation, partially through the MP-mediated control of CUP-SHAPED COTYLEDON (CUC) gene expression (Aida et al., 2002; Furutani et al., 2004). Furthermore, evidence suggests that auxin and WOX4 are both involved in procambium differentiation and/or maintenance in A. thaliana by controlling the HD-Zip class III gene HOMEOBOX GENE 8 (ATHB8), which specifies procambial cells, possibly through MP action (Donner et al., 2009; Ji et al., 2010). Together with the recent discovery of a CUC2-like and MP-like gene in EST libraries from Pinus taeda embryos (Cairney et al., 2006), our results thus imply that there are conserved embryo pattern events regulating procambium initiation/maintenance and, indirectly, cotyledon separation in gymnosperms and angiosperms.
WOX genes may contribute to shoot apex function in conifers
As in embryos, PaWOX8/9 and STIMPY/WOX9 share similar transcription patterns in the shoot apex of P. abies seedlings (Fig. 6b,c) and the SAM of A. thaliana, respectively (Wu et al., 2005). Recently, Skylar et al. (2010) showed that cytokinin signalling is involved in activation of STIMPY/WOX9 expression in meristematic tissues where it promotes undifferentiated growth by maintaining cell division and WUS expression (Wu et al., 2005). Loss of STIMPY/WOX9 leads, for example, to a failure in leaf-primordia initiation. High PaWOX8/9 expression in P. abies needle primordia could thus be linked to cytokinin signalling in needle formation, but also to cell cycle control, as some single cells of the needle primordia stained more intensely than others. However, whether these cells really are cycling and also express genes concerned with cell cycle control needs to be determined. When the PaWOX2 homeobox probe was used, a faint expression pattern was seen in the P. abies SAM (Fig. 6a), which might indicate transcription of WOX genes other than PaWOX2 and PaWOX8/9. Several WOX family members are expressed in the SAM of A. thaliana, O. sativa, Z. mays and Petunia × hybrida, with different roles in organ development, including leaf formation (Matsumoto & Okada, 2001; Nardmann et al., 2004, 2007; Park et al., 2005; Dai et al., 2007; Deveaux et al., 2008; Rebocho et al., 2008; Shimizu et al., 2009; Vandenbussche et al., 2009). Thus, our results indicate that PaWOX8/9 and possibly other WOX genes may contribute to shoot apex function, including needle initiation, also in P. abies.
STIMPY/WOX9 is expressed in the roots of A. thaliana seedlings and has a similar role as in the aerial part of the plant; that is, to maintain cell division and growth (Wu et al., 2005). As no expression of PaWOX8/9 could be found in P. abies seedling root after ISH, this suggests that WOX9 has acquired new functions in angiosperms or that other WOX8/9-like genes function in this tissue in conifers, which needs to be investigated further. Also, the PaWOX2 homeobox probe, which should have broader and less specific hybridization binding, did not detect any WOX transcripts in the seedling root or in the developing SAM of embryos. However, as genes homologous to CLAVATA1, 2 and 3, which are involved in WUS regulation in A. thaliana (Schoof et al., 2000), have been discovered in EST libraries from loblolly pine embryos (Cairney & Pullman, 2007), this implies that a WUS/WOX-CLAVATA3 (CLV3)/ENDOSPERM SURROUNDING REGION (ESR)-related (CLE) protein signalling regulation mechanism also functions in conifers, and so this needs to be investigated further.
Taken together, our results indicate that WOX2 and WOX8/9 are involved in conifer embryo formation and differentiation, and, together with PIN1 and PAT, in embryo patterning, similar to the situation in angiosperms. Thus, despite changes during seed plant evolution, a conserved mechanism, involving PAT and WOX transcription factors, may regulate seed embryo patterning. The comparison of expression in somatic and zygotic embryos also reveals somatic embryogenesis as a useful model system for conifer embryo development. Finally, similar mechanisms, including WOX gene activities, seem to function in both cotyledon and needle formation.
Support from the Linnaeus University (former Kalmar University) and the Royal Swedish Academy of Agriculture and Forestry is gratefully acknowledged.