Cloning of Pinus sylvestris SCARECROW gene and its expression pattern in the pine root system, mycorrhiza and NPA-treated short roots


Author for correspondence: Marjatta Raudaskoski Tel: +358 23338092 Fax: +358 23338075 Email:


  • • The SCARECROW (SCR) gene is central to root radial patterning. Its expression has not been investigated in conifers with morphologically different root types. Additional interest in SCR functions in the Pinus sylvestris root system comes from the effect of ectomycorrhiza formation on the short root apical structure.
  • • Here, the P. sylvestris SCR gene (PsySCR) was cloned and its expression investigated by northern blot and in situ hybridization of primary, lateral and short roots and mycorrhiza. Short root dichotomization was induced by auxin transport inhibitor (N-1-naphthylphthalamic acid (NPA)).
  • • PsySCR has conserved GRAS family protein motifs at the C-terminus and a variable N-terminus. PsySCR expression occurred in young root tissue and mycorrhiza. In root sections the PsySCR signal runs through the tip in initials for stele and root cap column and becomes upwards-restricted to endodermis in all root types.
  • • The PsySCR expression pattern suggests for the first time a regulatory role for SCR in maintaining the endodermal characteristics and radial patterning of roots with open meristem organization. The specific PsySCR localization is also an excellent marker for investigation of the dichotomization process in short roots.


In Arabidopsis, several genes encoding putative trancription factors regulating the differentiation of the root radial patterning have been isolated. Among the most central ones are SCARECROW (SCR) and SHORT ROOT (SHR) genes of the GRAS family (Pysh et al., 1999). SCR is necessary for periclinal division of a cell close to the stem cells at the quiescent center (QC) of the root. After the periclinal division, one of the daughter cells gives rise to the endodermis and the other to the cortex (Di Laurenzio et al., 1996). SCR has also been shown to be necessary for QC functions (Sabatini et al., 1999, 2003; Heidstra et al., 2004) and to be involved in repression of cell divisions leading to formation of middle cortex (Paquette & Benfey, 2005). The SHR gene functions upstream of SCR and it is synthesized in vascular tissue, from where it moves into the cells with SCR expression (Helariutta et al., 2000; Nakajima et al., 2001).

In spite of the central role of SCR and SHR genes in the root radial pattern formation, their presence and expression patterns have not yet been very extensively investigated in plants other than Arabidopsis thaliana. The expression and the universal function of SCR have been verified in maize (Lim et al., 2000, 2005), rice (Kamiya et al., 2003) and pea (Sassa et al., 2001), while no reports exist in angiospermous trees or gymnosperms. In particular, conifers like Scots pine (Pinus sylvestris) have a complicated root system consisting of morphologically and anatomically different root types, in which the primary root or the main root has an undetermined capacity for continuous growth, and the lateral roots have a somewhat limited ability, and the short roots a very limited ability, to elongate (Wilcox, 1968). The establishment of ectomycorrhizal symbiosis by the growth of fungal mycelium in the intercellular space between cortical cells of short roots is also an essential feature of the P. sylvestris root system. Additional interest in the occurrence and function of the central transcription factors in the P. sylvestris root system comes from the effect of fungal growth in the root on the structure of short root apical meristem. After the fungus has invaded the root, the tip splits to form first a dichotomous root, which through further divisions may form a coralloid root structure with up to 30 root tips (Niini & Raudaskoski, 1998). The formation of dichotomous roots may occur without contact with a fungus, for instance in response to stress conditions, and thus it is an endogenous property of the short roots (Faye et al., 1980). Previously it has been considered that auxin produced by the fungus plays a central role in inducing the dichotomization of the meristem (Slankis, 1973; Rupp & Mudge, 1985). Recently it was demonstrated that auxin transport inhibitors also induce dichotomous short roots (Kaska et al., 1999).

In the present work, we cloned the PsySCR gene from P. sylvestris root and followed its expression in different root types by in situ hybridization also when short root morphology was modified by the treatment with auxin transport inhibitor N-1-naphthylphthalamic acid (NPA) and by ectomycorrhizal symbiosis. Visualization of the PsySCR transcripts in the endodermis and apical meristem of different root types contributed to our understanding of differences in growth patterns of the main root, and lateral and short roots. The work also illuminated the structural features of short roots that make them preferential hosts for the development of ectomycorrhizal symbiosis.

Materials and Methods

Plant and fungal cultures, production of mycorrhiza and NPA and indole acetic acid treatments

The plant and fungal cultures and production of mycorrhiza followed the methods developed for biochemical analysis of Pinus sylvestris L. roots and P. sylvestrisSuillus bovinus mycorrhizal cultures (Niini et al., 1996; Tarkka et al., 1998, 2001) with slight modifications.

After c. 1 month, when the lateral and first short roots had developed on P. sylvestris seedlings, the roots were either treated with 10 µm NPA or indole acetic acid (IAA) or inoculated with the ectomycorrhizal fungus S. bovinus (Pers.) Roussel. NPA (Ducheva, Harlem, the Netherlands) and IAA (Sigma-Aldrich, Schnelldorf, Germany) were first dissolved in DMSO or ethanol, respectively, to 25 mm solution. From the stock solution, NPA and IAA were diluted to 10 µm concentration in sterile water agar and poured to a thin layer in a Petri dish. After the water agar had solidified, slices approx. 1.5 cm wide and 8 cm long were cut and transferred to the top of the P. sylvestris root system growing in test tubes. The control roots were covered with slices of water agar in which DMSO was diluted to the same concentration as water agar with NPA. In IAA treatments, ethanol was removed by evaporation. After 8 wk, when dichotomous short roots had developed, the seedlings were fixed in situ, and the number and structure of lateral and short roots were documented.

DNA and RNA isolation and northern hybridization

Pinus sylvestris DNA was isolated from primary roots frozen in liquid nitrogen using the DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA). For RNA isolation P. sylvestris primary, lateral and short roots and mycorrhizas, as well as stem and needles, were quickly cut from the seedlings and immediately frozen in liquid nitrogen and stored at –80°C. Total RNA was isolated as described in Nehls et al. (1998). S. bovinus mycelia grown on dialysis membranes were also collected in liquid nitrogen and stored at –80°C. From mycelia, total RNA was isolated as described in Russo et al. (1992). Northern hybridization experiments were carried out as reported in Gorfer et al. (2001).

Cloning of PsySCR

The first fragment of PsySCR was isolated by PCR using P. sylvestris DNA and primers PA1 5′-GCAGACAACTTTGAAGAAGCC-3′ and PA2 5′-AGGGAGAGGAGAATACATTCC-3′ designed on the basis of a putative 240-bp-long Norway spruce (Picea abies) SCR gene fragment obtained from Prof. Yrjö Helariutta (University of Helsinki, Finland). With the aid of the amplified DNA sequence of the P. sylvestris PCR fragment, the SCR1 5′AACTCTCCACCCCCTATGGCAACTCGG-3′ and SCR2 5′-CAAAATAGGCAGCCACTCGTTGCACCG-3′ primers were constructed. These primers were used for 3′- and 5′-RACE of P. sylvestris SCR cDNA, respectively, with the SMART RACE cDNA Amplification Kit (CLONETECH Laboratories, Inc., Palo Alto, CA, USA) according to the manufacturer's instructions. The preparation of P. sylvestris 5′- and 3′-RACE Ready cDNA from primary root total RNA was conducted using the respective primers for 3′- and 5′-RACE provided by the kit under standard PCR conditions.

To ensure that the fragments obtained by 5′- and 3′-RACE-PCR represented parts of the same PsySCR cDNA, the encoding region was amplified using primer 5 5′-GACATGGCAGCTTGTATTGC-3′ from the N-terminus and primer 6 5′-TGATGAGCAGGTCTCCATGC-3′ from the C-terminus of the putative ORF and 5′-RACE-Ready cDNA of P. sylvestris. In addition, the PsySCR gene was PCR-amplified from P. sylvestris genomic DNA by using primers 5 and 6, but also as two fragments representing the beginning and end of the gene by primer combinations 5 and SCR2, and 6 and SCR1, respectively.

All the PCR products were cloned into the pCR2.1 TOPO vector (Invitrogen, Carlsbad, CA, USA) and sequenced at the DNA Sequencing Service, University of Turku or GATC Biotech AG (Konstanz, Germany).

Histochemical techniques, riboprobe preparation and in situ hybridization

Seedlings were fixed in situ by filling the test tubes with freshly made 4% paraformaldehyde in phosphate-buffered saline (PBS), pH 7.4, in vacuum for 1 h. After the preliminary fixation, the seedlings were moved from the test tubes into fresh fixative in which the number of different root types was recorded under a dissecting microscope. The roots were cut from the seedlings and fixed from 2 to 3 h more in a vacuum. After several rinses with PBS, the root material was arranged on slides and covered with a thin layer of 1% agarose solution prepared in RNase free sterile water. The agarose blocks with the roots tips were dehydrated, embedded in paraffin and sectioned. Serial microtome sections, 8–10 µm thick, were collected on objective slides treated with 3-aminopropyltriethoxysilane (Sigma) to increase the attachment of the sections to the slides. However, it was difficult to obtain serial sections of mycorrhizal short roots at the early stage of dichotomization and, therefore, for the in situ hybridization, short roots were treated with NPA.

From the cloned PsySCR ORF, a 379 bp PCR fragment was produced with primer 8 5′-GAGTATGTGGGTGCAACCAA-3′ and primer 9 5′-CATTGCTTCTGCAAAATAGG-3′. The fragment encoding amino acids from S-399 to M-524 from the center of the PsySCR ORF (Fig. 1b) was cloned in sense and antisense orientation in the pCR2.1 TOPO vector in order to produce sense and antisense riboprobes from BamHI linearized vectors. Sense and antisense complementary RNA probes for PsySCR mRNA were transcribed in vitro using the DIG RNA Labeling Kit (Boehringer-Mannheim, Roche, Mannheim, Germany), which contained a T7-RNA polymerase and DIG-RNA-ribonucleotide mixture with digoxigenin-11-UTP.

Figure 1.

Functional domains of Pinus sylvestris SCARECROW (PsySCR) and alignment of SCARECROW (SCR) amino acid sequences. (a) The functional domains in boxes are LHRI, LHRII (leucine heptad repeat), VHIID, PFYRE and SAW, named according to the conserved amino acids in their respective domains (Pysh et al., 1999). The putative two-partite nuclear localization signal at the beginning of the sequence is marked by two vertical bars. The two LRRs (leucine-rich regions) are unique to PsySCR. (b) Alignment of SCR amino acid sequences from Cucumis sativus (CAI30893), Arabidopsis thaliana (AAB06318), Pisum sativum (BAB39155), Pharbitis (Ipomoea) nil (BAE48702), Oryza sativa (BAD22576) and Zea mays (AAG13663) and PsySCR (ABH85406). The putative two-partite nuclear localization signal at the beginning of the sequence is shown by asterisks and the LRRs by bars with triangles. In the highly conserved C-terminus, the labeling and underlining of the motifs is in accordance with Pysh et al. (1999), except for the labeling of the LXXLL motif by asterisks at the beginning of LHRI.

The in situ hybridization and immunological detection were performed as previously described (Di Laurenzio et al., 1996; Lim et al., 2000; Sassa et al., 2001) with slight modifications. The best results were obtained when the in situ hybridization was started at 80°C and the temperature was allowed gradually to decrease to 42°C.

Starch staining

Paraffin was removed from the sections representing primary roots, short roots and mycorrhiza, rehydrated and stained with Lugol's solution (Merck, Darmstadt, Germany), washed and mounted in 10% glycerol in sterile water.


The samples were viewed with a DMBL research microscope (Leica Microsystems GmbH, Wetzlar, Germany) and images were captured electronically with a PCO CCD 12 BIT Cooled Imaging Video Camera (SensiCAM, Kelheim, Germany) and displayed with Image-Pro Plus software (Media Cybernetics, Silver Spring, MD, USA). Images were further processed using Adobe Photoshop (Adobe Systems, San Jose, CA, USA) or Corel PHOTOPAINT and CorelDRAW (Corel Corporation, Ottawa, Ontario, Canada).

Accession number of PsySCR and PsySCR sequences: DQ855418, ABH85406.


Cloning and characterization of PsySCR

The 157-bp-long PCR fragment obtained from P. sylvestris genomic DNA with homologies to AtSCR (AAB06318) was used to design the oligonucleotide primers SCR1 and SCR2 for 3′- and 5′-RACE-PCR in order to obtain a full-length PsySCR cDNA. The 5′- and 3′-RACE-PCR produced approx. 1700 and 1200 bp fragments, respectively, which were cloned. From both RACE PCR products, three clones were sequenced and each of the three sets had the same nucleotide sequence except for length variation in 5′- and 3′-UTRs. The longest 5′-fragment was 1678 bp and it encoded amino acid sequences 42 and 80% identical to amino acids 141–188 and 284–334 in A. thaliana SCR (AAB06318). The amino acid sequence of the first 1000 bp of the 3′-fragment was 75% identical to amino acids from 319 to 651 in the 653 amino-acid-long protein of A. thaliana SCR (Fig. 1). This suggested that we had cloned cDNA fragments related to the N- and C-terminal parts of A. thaliana SCR.

Pinus sylvestris 5′-RACE-Ready cDNA and genomic DNA were further amplified using a primer from the beginning and the end of the coding region of the putative PsySCR cDNA. Both amplifications produced a 2500 bp PCR fragment. The nucleotide sequences of the cDNA and genomic PCR fragments were 99% identical among themselves and with the joined 5′- and 3′-RACE and -genomic PCR products. This indicated that fragments obtained originally by 5′- and 3′-RACE-PCR represented parts of the same PsySCR cDNA. The same size of the PCR products from cDNA and genomic DNA indicated that the gene contains no intron. The gene was named PsySCR and its ORF consists of 2526 nucleotides encoding 842 amino acids (Fig. 1).

The PsySCR ORF was c. 40% identical to the SCR of Pharbitis (Ipomoea) nil, Cucumis sativus, Pisum sativum, A. thaliana, Oryza sativa and Zea mays, the highest similarity (44%) being with P. nil and the lowest (40%) with Z. mays. SCR of P. nil, C. sativus and P. sativum is c. 800 amino acids long, as is PsySCR, while in A. thaliana, O. sativa and Z. mays, SCR is shorter, around 650 amino acids (Fig. 1b). The size difference is the result of the long and highly divergent N-terminal region which ends at amino acid 469 in PsySCR. However, in the N-terminal region of PsySCR, the amino acid residues from 196 to 244 are similar to the other described SCR proteins, and the 3′-end of this region from amino acid 237–250 contains a leucine-rich region (LRR) which has some features of a leucine heptad repeat region (LHR). A similar sequence also occurs from amino acid 312–349 in a region unique to PsySCR (Fig. 1b). Both sequences consist of three heptad repeats, one of which is separated by an internal loop, containing 11 amino acids in the first LRR and eight in the second, from the other two heptads. Comparable regions are not detected in the N-terminus of the other SCRs. In addition, the basic amino acids RRKR and RR from 43 to 46 and 57 and 58 at the N-terminus of PsySCR could represent a two-partite nuclear localization signal (Fig. 1).

The C-terminal region of PsySCR protein from amino acid 470–835 is highly similar, containing over 75% identical amino acids with the other investigated SCR proteins. The motifs LHRI, VHIID, LHRII, PFYRE and SAW characteristic to the SCR and other GRAS gene family proteins (Pysh et al., 1999) are clearly recognized in PsySCR (Fig. 1). In PsySCR, a LLALL sequence, and in the other SCR proteins a LLTLL sequence, occurs at the beginning of the first LHRI region. In mammalian cells, the LXXLL sequence is a protein-recognition motif widely used in transcriptional regulation (Plevin et al., 2005). The phylogenetic analysis of the 294 amino acids from the C-terminal region of PsySCR and of 63 other GRAS family members from public databases divided the members into groups comparable to those reported by Bolle (2004) and placed PsySCR in the SCR branch (Fig. 2).

Figure 2.

A maximum-likelihood tree of Pinus sylvestris SCARECROW (PsySCR) and 63 related GRAS protein sequences. Different groups of GRAS family proteins have been delineated according to Bolle (2004). Note that some sequences fall outside these groups. Conifer proteins PsySCR (ABH85406) and PrSCL1 (ABG77970), as well as three Pinus taeda sequences (TC60068, TC60455 and TC67393), occur in SCR, SCL9, SHR and HAM groups. For more information on the origin of P. taeda TC sequences, see the text. Species abbreviation in the gene names are as follows: At, Arabidopsis thaliana; Bn, Brassica napus; Cs, Cucumis sativus; In, Ipomoae nil; Ll, Lilium longiflorum; Os, Oryza sativa; Ph, Petunia ¥ hybrida; Pr, Pinus radiata; Psy, Pinus sylvestris; Pt, Pinus taeda; Ps, Pisum sativum; Sl, Solanum lycopersicum; Ta, Triticum aestivum; Zm, Zea mays. xxx denotes unannotated proteins. The annotation of the majority of rice GRAS proteins is in accordance with Tian et al. (2004). The tree was generated using the MrBayes 3.0 program and a data set of 2949 characters selected from the conserved part of the multiple sequence alignment (generated and edited using the XCED program). For clarity, only posterior probabilities of branches deviating from 1.00 are shown at the respective locations. The tree was calculated using the amino acid substitution model of Jones and represents sampling of 3013 trees from the 10 001 calculated trees (the first 390 were rejected as unstable initial trees), while a 99% credible set was 2917 trees.

Identification of SCR and other GRAS family genes in the Pinus gene index

When the nucleotide sequence of PsySCR ORF was used in tblastn to screen the Pinus taeda (loblolly pine) EST and TC sequences (, several TC clones homologous with PsySCR were detected, representing a large range of GRAS family proteins. TC61519, consisting of the overlapping P. taeda EST clones DR095931, CO362002, CO361927, CX714326 and CX714276, was 98.7% identical at the nucleotide level with the terminal region of PsySCR ORF from nucleotide 1837–2526. The nucleotide sequence encodes the end of the VHIID motif and the complete LHRII, PFYRE and SAW motifs (Fig. 1a). Out of the nine nucleotide differences between P. sylvestris and P. taeda nucleotide sequences, only two caused a difference in the SCR amino acid sequences (Supplementary material, Fig. S1a). The high identity between the nucleotide sequences of TC61519 and PsySCR continued to 3′-UTR, so that the 3′-end of the TC61519 nucleotide sequence is 97% identical to the longest-3′-UTR sequence of PsySCR (Fig. S1b). All the EST sequences of P. taeda except for DR095931 represent SCR gene transcripts with longer 3′-UTRs than those in P. sylvestris. Unfortunately the search found no sequence homology to the N-terminal region of PsySCR. P. taeda SCR-like cDNAs have also been identified in microarray analyses of 2178 P. taeda cDNAs with transcripts from Picea glauca (Moench) Voss embryos at different developmental stages (Stasolla et al., 2003, 2004).

BlastX against public databases with P. taeda TC60068 nucleotide sequence, which also showed high similarity to PsySCR, revealed that TC60068 is 58% identical to the 748 amino-acid-long SCARECROW-like protein (LISCL) of Lilium longiflorum (Morohashi et al., 2003) at the C-terminal region of the protein from amino acid 308–747.

Comparison of the other sequences, TC60455, TC67393, TC62756 and TC73035 obtained from the TIGR Pinus gene index by tblastn with PsySCR revealed several other GRAS family genes. TC60455 encodes the C-terminal region of the P. taeda SHR gene with 54% identity at amino acid level to AtSHR, while TC67393 showed 44% identity to AtSCL6. In the dendrogram of GRAS family proteins, TC60068, TC60455 and TC67393 appear in their respective branches (Fig. 2). Unfortunately TC61519, as well as TC62756 and TC73035, were too short to be included in the tree calculation.

Characterization of the PsySCR expression pattern in pine root system

Three different EcoRI fragments, representing the N-terminal, middle and C-terminal parts encoding the ORF in PsySCR, were used in northern hybridization experiments. All three fragments gave the same hybridization pattern. A strong signal was obtained with total RNA isolated from primary and short roots and a less strong signal with ectomycorrhizal short roots. No signal was obtained with needle and stem RNA or with RNA from the ectomycorrhizal fungus S. bovinus (Fig. S2).

This result led to investigation of the expression of PsySCR in the P. sylvestris root system at cellular level by in situ hybridization using an antisense and a sense probe prepared from a 379 bp fragment from the center of PsySCR. The 5′ region of the PsySCR probe had low, and the 3′ region high, homology with other SCR genes in public databases. In the longitudinal sections of primary (Fig. 3a), lateral (Fig. 3b) and short roots (Fig. 3c), in situ hybridization localized PsySCR mRNA in a cell layer that was interpreted to represent the endodermis, as it separated the vascular tissue from cortical cell layers. P. sylvestris primary and lateral roots have an open root meristem (Esau, 1977) in which the apical initials for the vascular tissue appear to be continuous with the columella root cap cells (Fig. 3a). In such roots, the QC is not as easily identified as in A. thaliana and Z. mays roots with closed meristem. Median longitudinal sections of the primary, lateral and short roots suggested that the PsySCR signal extended, although sometimes with less intensity, through the very center of the tip in the initials for the vascular tissue and root cap columella cells. On both sides of columella initials, PsySCR expression appeared to be two to three cells wide (Fig. 3a) but became upwardly restricted mainly to one cell layer that represented the endodermis (Fig. 3b). In the endodermis, PsySCR mRNA was detected running along the meristem and elongation zone but no longer in the older parts with thick cell walls. The lack of the PsySCR signal in the differentiated endodermal cells was clearly observed in short roots in which the root cap, meristem and elongation zone are less developed (Fig. 3c) than in the primary and lateral roots. No signal was obtained with the PsySCR sense probe (Fig. 3d).

Figure 3.

Localization of Pinus sylvestris SCARECROW (PsySCR) transcripts in longitudinal sections of Pinus sylvestris roots. (a) PsySCR expression occurs in cells between stele (s) and cap (c) in primary roots. Note the two rows of columella cells (cc) in the cap. Arrowheads mark the PsySCR signal in the cells on both sides of columella initials. (b) PsySCR expression in the root apex and endodermis in a lateral root. (c) PsySCR expression in the endodermis of a short root. (d) No signal is seen in the short root hybridized with PsySCR sense probe. (e) Dichotomous mycorrhizal short root with a PsySCR expression in the right side tip. (f, g) A weak PsySCR signal (marked by arrowheads) occurs in the mycorrhizal short root tip between stele and cap comprising few cell layers. Note the mycelial sheath (ms) around the root tip. (h) A mycorrhizal root tip hybridized with PsySCR sense probe. No signal is seen. Bars, 20 µm (a, f–h); 100 µm (b–e).

The northern blot hybridization suggested that PsySCR is also expressed in the mycorrhizal short roots (Fig. S2), which was confirmed by in situ hybridization. The examination of the dichotomous mycorrhizal roots (Fig. 3e) at higher magnification revealed a signal at the tips of these roots. The cells with the PsySCR signal separated the poorly developed cap cells from the initial cells of the vascular tissue and extended into a few endodermal cells between the vascular tissue and the cortical cells (Fig. 3f,g). No signal was obtained in the mycorrhizal short roots with the PsySCR sense probe (Fig. 3h).

The structure of NPA-treated short roots

It has been reported that polar auxin transport inhibitor NPA induces in P. sylvestris growth of dichotomous short roots comparable to those in mycorrhiza (Kaska et al., 1999). Therefore 1-month-old seedlings in which the first short roots started to appear were treated with 10 µm NPA or IAA and the formation of dichotomous short roots was recorded after 2 months (Table 1; Figs 4, 5). No statistically significant difference in the number of lateral and short roots per seedling occurred between control, NPA- and IAA-treated samples except for the significantly higher (t-test P < 0.001) number of dichotomous short roots per seedling in NPA-treated material (Table 1). In a few samples the NPA-treated short root tips had split several times (Fig. 4b), which also happens in mycorrhizal short roots (Fig. 4c,d), although the distance between the first and second root tip splitting was much longer in NPA-treated than in mycorrhizal short roots. The occurrence of dichotomous short roots in control seedlings indicates that the process is an endogenous property in P. sylvestris that can be accelerated by inhibition of polar auxin transport by NPA if the treatment takes place at the right time. In the present experiment, no special effect of exogenous IAA on formation of dichotomous short root was found (Table 1).

Table 1.  Number of different root types in 3-month-old control, N-1-naphthylphthalamic acid (NPA)- and indole acetic acid (IAA)-treated Pinus sylvestris seedlings
Root typeControl10 µm NPA10 µm IAA
  1. Pairwise comparisons of root numbers per seedling from different treatments indicate a significant increase in dichotomous short root number in NPA-treated seedlings according to t-test (***, P < 0.001). Data are means ± SE.

Number of seedlings111714
Lateral roots per seedling34.3 ± 7.928.4 ± 3.134.7 ± 4.8
Short roots per seedling79.0 ± 13.272.4 ± 9.565.6 ± 9.5
Dichotomous short  roots per seedling 1.9 ± 1.0 7.5 ± 1.3*** 0.7 ± 0.3
Figure 4.

Control, NPA (N-1-naphthylphthalamic acid)-treated and mycorrhizal short roots of Pinus sylvestris. (a) Short roots extending from pieces of lateral roots. (b) Branched tips in NPA-treated short roots. (c) A short root surrounded by Suillus bovinus mycelium. (d) Dichotomous mycorrhizal short root tips. Bar, 1 mm.

Figure 5.

In situ hybridization analysis of Pinus sylvestris SCARECROW (PsySCR) in control and NPA (N-1-naphthylphthalamic acid)-treated Pinus sylvestris short roots. (a–c) PsySCR transcripts are seen in the short endodermis and at tip between stele and poorly developed root cap in three longitudinal serial sections of a control short root. (d) The short root tip in (b) at higher magnification. A strong PsySCR expression is seen in endodermal cells and a weak signal spread in columella cells (arrow) and in cells on both sides of columella that could represent initials for cortical and cap cells. No spreading of the signal is seen on stele side. (e–g) Three longitudinal serial sections of a NPA-treated short root. The expression of PsySCR appears to be similar to that in control short roots. The root cap is thinner than in control short roots. (h) The short root tip in (f) at higher magnification. The meristem of NPA-treated short roots appears to be less well organized than in control short roots. Note the three cell groups with weaker PsySCR signal (arrows), perhaps representing columella initials. (i) PsySCR expression in the tips of a dichotomous NPA-treated short root. (j) PsySCR expression in endodermis in a cross-section of a root tip from a NPA-treated short root branch. (k) A longitudinal section of a NPA-treated short root at an early stage of bifurcation. Note the cell with a big vacuole (arrowhead) in the center of the root and the weak PsySCR signal in each tip. (l) Higher magnification of part of the root tip on the left in (k). Note the elongated cells defined by arrows in vascular tissue differentiating to new endodermis left from the central vacuole (arrowhead). (m) Higher magnification of part of the root tip on the right in (k). Note the long cells defined by arrows in vascular tissue differentiating to new endodermis. The cells originate from PsySCR-expressing cells at the new root tip to the right from the central vacuole (arrowhead). Bars, 100 µm (a–c, e–g, i–k); 20 µm (d, h, l, m).

In situ hybridization of serial sections of control (Fig. 5a–d) and NPA-treated (Fig. 5e–h) short roots showed PsySCR transcripts in the root tip and in the endodermis, where the occurrence of the label was very short (Fig. 5a–c,e–g). The comparison of the structure of control and NPA-treated root tips at higher magnification (Fig. 5d,h) indicated that, in addition to a strong PsySCR signal in the short endodermis, a weak signal was spread in one or two cell layers at the root tip, the lower layer of these cells belonging to root cap tissue, including columella cells. In the control roots, the columella was formed of two to four cells in two rows (Fig. 5d), while in NPA-treated short roots the columella cells were smaller and less well organized (Fig. 5h). The endodermis and the root cap appeared to be shorter in NPA-treated than in control short roots, and this length difference was proved to be statistically significant (Fig. S3). No statistically significant difference occurred in the diameter of the short root or stele between control and NPA-treated short roots, although in longitudinal sections the diameter of NPA roots and stele usually appeared to be larger.

The PsySCR signal occurred also in dichotomous short roots from NPA-treated material (Fig. 5i,j). In a few cases it was possible to catch longitudinal sections of short roots at an early stage of dichotomization (Fig. 5k). A weak signal of PsySCR was seen in the two young tips isolated by a cell with a big vacuole. The new endodermis appeared as elongated cells extending into vascular tissue from the PsySCR-labeled cell layer at the new tips on both sides of the central vacuole (Fig. 5l,m).

Location of starch grains in primary, lateral and short roots

Staining with Luganol-reagent indicated that, in primary and lateral roots, starch grains accumulated at the bottom of the cap cells, especially in the columella area (Fig. 6a,b). In short roots, the location of starch grains was not associated with columella cells but they occurred in an outer cap cell layer, at the upper side of the cells (Fig. 6c). In NPA-treated short roots, the starch grains also appeared in the cap cell layer below the columella cells and were located at the upper but also at lateral sides of the cells (Fig. 6d). Similar distribution of starch grains also occurred in mycorrhizal short roots (Fig. 6e).

Figure 6.

Visualization of starch grains in the root cap cells of Pinus sylvestris. (a) Starch grains in the lower part of root cap cells in a longitudinal section of a lateral root. (b) A higher magnification of the lateral root in (a) shows clearly the basal location of starch grains in columella cells. (c) In a short root, the starch grains occur in a cap cell layer close to the root surface at the upper part of the cells. No location in columella cells could be detected which could explain the absence of gravity response from short roots. (d) In NPA (N-1-naphthylphthalamic acid)-treated short roots, the location of starch grains in cap cells is also in the upper part and sides of the cap cells. (e) A few starch grains are seen in the upper part of the cap cells in a mycorrhizal short root covered by a mycelial sheath (ms). Bars, 100 µm (a, c); 10 µm (b, d, e).


Phylogenetic analysis using the highly conserved C-terminal amino acids of 64 GRAS proteins showed that the conifereous and angiospermous sequences appear together in the same GRAS subfamily branches (Fig. 2). This indicates a close relationship between gymnosperms and angiosperms in the development and function of GRAS family proteins. GRAS genes of these two plant groups may play physiologically and developmentally similar roles, although the highly divergent N-terminus of the GRAvS proteins, as in PsySCR, could also give different specificity to the genes, leading to regulation of different processes in gymnosperms than in angiosperms.

Interestingly, no introns were detected in PsySCR. All the sequenced SCR genes from monocots, including Z. mays (Lim et al., 2000) and O. sativa (Kamiya et al., 2003), and from the dicots A. thaliana (Di Laurenzio et al., 1996), P. sativum (Sassa et al., 2001) and C. sativus (AJ870307) contain an intron at the same site in the conserved C-terminus of SCR genes. The lack of this intron from PsySCR may suggest that the intron has been introduced into the angiospermous genome after the divergence of angiosperms and gymnosperms. This assumption has to be verified when more SCR gene sequences will be available from gymnosperms. In conifer genes encoding enzymes for terpene synthases, introns are frequent and their location is, in many cases, similar to those in A. thaliana (Trapp & Croteau, 2001).

Motifs in the predicted PsySCR protein

The highly conserved LHRI-VHIID-LHRII region present in GRAS proteins including PsySCR has been suggested to function as a DNA-binding domain analogous to the bZIP protein–DNA interaction, with the LHRs mediating protein–protein interactions and the VHIID motif mediating protein–DNA interaction (Pysh et al., 1999; Bolle, 2004). In the LRR regions of the PsySCR N-terminus, an internal loop occurs between heptad repeats. A heptad separating loop is also present in the C-terminal LHRI and LHRII regions defined by Pysh et al. (1999), and a loop has been described recently in leucine zippers typical to plant HD-START family transcription factors (Schrick et al., 2004). However, the leucine heptads in PsySCR, as in all GRAS proteins, are very short in comparison to the conventional leucine zippers (Schumacher et al., 1999) and therefore may not function as true leucine zippers. As the regions are able to form coiled coils (Lupas, 1997), they could regulate intramolecular or intermolecular interactions of SCR proteins. LRR regions of PsySCR could replace the homopolymeric sequences assumed to regulate the protein structure in the N-terminal part of the other SCR proteins (Di Laurenzio et al., 1996; Lim et al., 2000; Sassa et al., 2001; Kamiya et al., 2003; Kitazawa et al., 2005).

In A. thaliana SCR, a sequence in front of the LHRI domain contains several basic amino acids and it has been suggested to function as a nuclear localization signal (NLS) (Di Laurenzio et al., 1996). No homology is detected between A. thaliana SCR and PsySCR in this region. Instead, at the beginning of the N-terminus of PsySCR there is a two-partite amino acid sequence RRKR X10 RR, which has similarity with the nucleoplasmin NLS (Dingwall & Laskey, 1991) and Opaque2 NLS B (Varagona & Raikhel, 1994) in the sense that it consists of two groups of basic amino acids separated by 10 amino acids. The sequence also has similarity with many NLS sequences known to be involved in binding to specific members of the importin superfamily (Poon & Jan, 2005). Whether the RRKR X10 RR sequence directs PsySCR to the nucleus remains to be tested.

PsySCR functions

In northern hybridization a PsySCR signal was obtained from young root tissues and mycorrhiza, but not in needles and stems, while in A. thaliana and P. sativum the expression of SCR has been detected in other plant organs, too (Di Laurenzio et al., 1996; Sassa et al., 2001). In P. sylvestris needles and stems, the PsySCR expression could be too low to be detected by northern hybridization. In P. sativum no signal was obtained in RNA blots from either older parts of the shoot or the root proper (Sassa et al., 2001).

In all conifers the primary, lateral and short roots have an open type of apical meristem in which the initials for stele, ground tissue, epidermis and root cap cells occur in one cell layer (Esau, 1977). No sharp boundary is seen between root apical meristem and root cap (Jiang & Feldman, 2005; Fig. 7a) as in Arabidopsis primary roots with a closed type of apical organization and with three different layers/tiers of initials, the innermost giving rise to the stele, the second to the ground tissue (endodermis and cortex) and the third to the root cap and epidermis (Esau, 1977). In the center of the Arabidopsis root apical meristem, a few mitotically inactive cells constitute the QC surrounded by the mitotically active stem cells giving rise to different tissues (Benfey & Scheres, 2000). In the conifer Calocedrus (Libocedrus) decurrens, the initial cell layer was less active in incorporating radioactive thymidine and was therefore suggested to represent the QC of the root (Wilcox, 1962). The stele and the root cap column were thought to develop from cells above and below the central initials and the ground tissue, and the peripheral region of the root cap was thought to have a common origin from the initial cells that surround the proximal end of the root cap column (Wilcox, 1962).

Figure 7.

A hypothetical model of auxin distribution in main (a) and short roots (b) in Pinus sylvestris. The description of auxin flow in main roots (black arrows) is based on the knowledge of auxin distribution and the location of auxin transporters in Arabidopsis (Friml, 2003; Leyser, 2005; Petrášek et al., 2006). The dichotomous short root pattern occurring as a result of NPA (N-1-naphthylphthalamic acid) treatment (c) and ectomycorrhiza formation (EMR) (d) are suggested to result from alterations in auxin distribution. The highly specific and restricted expression pattern of PsySCR (dashed line) makes it an excellent marker for the developmental stage of each root type. (a) Thick arrows in the center of the root show polar auxin transport towards the root tip which creates auxin maximum at the border of the meristem and columella. From the columella, auxin is distributed in all directions into the root cap, from where it is transported basipetally towards the root elongation zone. It is also suggested that auxin may return from the basipetal flow in the elongation zone to the polar auxin transport route in the center of the root. (b) In the short roots the auxin stream is shown by thinner and shorter arrows to describe the putative lower auxin concentration in short roots than in main roots. (c) Treatment of short roots with auxin transport inhibitor (NPA) disrupts the auxin maximum at the center of the root tip. NPA increases and spreads the auxin accumulation in the root tip to superoptimal concentrations (triangle), which leads to cessation of meristematic activity at the center of the root tip (cross). At the border of the meristem, the auxin accumulation remains optimal for meristematic activity and two new root tips with a few columella cells and their own auxin recycling are formed. (d) In the ectomycorrhizal short roots, the tight hyphal network in the apoplastic space between epidermal and cortical cells (Hartig net, marked by wavy lines) could interfere with the basipetal auxin transport and induce a superoptimal auxin concentration and cessation of meristematic activity at the center of the short root tip (cross) and induction of two new root tips with a few columella cells and their own auxin recycling at the edge of the meristem as in NPA treatment. The elongation of the dichotomous short roots in NPA treatment (Fig. 4b) but not in ectomycorrhiza (Figs 4c,d) suggests that some basipetal auxin flow (dotted arrows) could occur in the former roots. The growth of the new root tips in (c) and (d) is illustrated by rows of small squares representing columella cells.

The strong expression of PsySCR in differentiated endodermis suggests a regulatory role for SCR in roots with open meristem organization in the radial patterning of the root and in maintaining the endodermal characteristics comparable to that in angiospermous roots with closed meristem organization (Di Laurenzio et al., 1996; Lim et al., 2000, 2005; Sassa et al., 2001; Kamiya et al., 2003). In addition, PsySCR expression in P. sylvestris roots occurs in the cells between the provascular and root cap columella cells. These cells fit well with the description of initial cells with QC properties (Wilcox, 1962; Esau, 1977). In Arabidopsis the functional analysis based on AtSCR mutant alleles, root tissue specific markers and laser ablation of QC has shown that SCR is required for QC cell specification, which maintains the flanking stem cells in an undifferentiated stage (Sabatini et al., 2003; Heidstra et al., 2004; Xu et al., 2006). In the longitudinal sections of P. sylvestris roots, PsySCR is also detected in cells flanking the early root cap columella cells. These cells perhaps represent stem cells for the peripheral region of the root cap (Fig. 3a, 5d). The PsySCR expression in the cells between the flanking initials for root cap peripheral cells and the differentiated endodermis could be comparable to the AtSCR occurrence in ground tissue stem cells (Heidstra et al., 2004). In these cells SCR induces the periclinal division which causes the separation of two cells, one of which then differentiates to endodermis and the other to cortex. The anticlinal and periclinal divisions are not easy to recognize in P. sylvestris root apical meristem, where they might not be as strictly defined as in Arabidopsis root meristem.

In P. sylvestris the cortex is formed of several cell layers, as it is in maize (Lim et al., 2000), pea (Sassa et al., 2001) and rice (Kamiya et al., 2003). Recently, it has been shown that an extra cell layer, middle-cortex, is formed in A. thaliana roots in distance from the root apex (Baum et al., 2002; Paquette & Benfey, 2005). The middle-cortex has its origin in periclinal divisions of the mature endodermal cells, in which SCR appears to inhibit these divisions close to the root apex. It is possible that in P. sylvestris, as well as in rice and maize, the inhibitory role of SCR is not present, which allows the formation of multiple cortex layers from very young endodermal cells, a process visualized clearly in rice root (Kamiya et al., 2003).

The expression and function of SCR are dependent on interplay with another GRAS protein, the putative SHR transcription factor (Sabatini et al., 2003; Heidstra et al., 2004). SHR is transcribed in stele (Helariutta et al., 2000), from where it is transported as a protein (Nakajima et al., 2001) to those cells in which SCR is expressed. In QC and in stem cells, the SHR protein is necessary for induction of SCR transcription, while during the periclinal division and in determining the endodermal characteristics the interplay between SHR and SCR appears to work on a different basis (Heidstra et al., 2004; Sena et al., 2004). In P. sylvestris the expression and function of SCR could also be dependent on the interplay with a SHR-like transcription factor, since in the Pinus gene index a partial sequence of this gene is found.

Short root development

The localization of PsySCR transcripts in short roots indicated that the structure of root cap, meristem and elongation zone is poorly developed in comparison with those in the primary root (Fig. 7a,b). In addition, the starch grains are missing from the columella cells and those present in the other root cap cells are fewer in number and differently located than in the primary and lateral roots, which could explain the absence of gravity responses from short roots (Niini & Raudaskoski, 1998). In addition to the putative SCR and SHR transcription factors, the maintenance and function of the root apical meristem in A. thaliana is dependent on auxin (Sabatini et al., 1999; Leyser, 2005; Kepinski, 2006). Auxin promotes both the cell division and cell expansion and regulates the root gravity responses (Teale et al., 2006). The reduced growth and structure of the short roots could be the result of faulty auxin distribution in the P. sylvestris root system, leading to lower auxin concentration in short roots than in primary and lateral roots (Fig. 7a,b). This idea is further supported by the sensitivity of short root development to auxin transport inhibitors.

Dichotomization of NPA-treated and mycorrhizal short roots

The present experiments confirmed that NPA enhances the dichotomization of P. sylvestris short root tips. In addition to NPA, precursors of ethylene synthesis (Kaska et al., 1999) or ethylene itself (Rupp & Mudge, 1985) are known to induce dichotomization of short roots. The similar phenotype is suggested to result from the activation of ethylene biosynthesis by an increased auxin concentration (Kaska et al., 1999), caused by NPA treatment at the root tip (Sabatini et al., 1999; Friml et al., 2002). In the apical center of the short roots, ethylene could induce differentiation of the meristem cells to vacuolated parenchymatous cells seen here in NPA-treated (Fig. 5k–m), and reported previously in mycorrhizal dichotomous, short roots (Wilcox, 1968; Piche et al., 1982; Niini & Raudaskoski, 1998).

In the mycorrhizal short roots, a tight Hartig net is formed by fungal hyphae in the apoplastic/intercellular space between epidermal and cortical cells proximal to the meristem. The Hartig net could reduce auxin recycling in the cortex and epidermis (Friml, 2003; Leyser, 2005) and increase auxin accumulation in the central short root meristem, causing the establishment of two lateral apical meristems in the same way as the NPA treatment is suggested to effect (Fig. 7c,d).

During the development of dichotomous short roots, the cells on both sides of the central vacuolated tissue maintain their meristematic activity, probably because of more moderate accumulation of auxin at the sides than in the center of the apex (Fig. 7c,d). The new endodermis for the two new root tips is formed on both sides of the vacuolated central part of the short root from cells of the provascular tissue in stele, in the same manner as maize root regeneration to two new root tips happens after half of the root tip region has been removed (Lim et al., 2000). In maize during the regeneration of the root tips, the localization of ZmSCR expression facilitated the definition of different stages during the regeneration process. In contrast, in P. sylvestris, the expression of PsySCR was clear in the developing new root tips but weak in differentiating endodermal cells in the provascular tissue. The occurrence of a strong PsySCR signal in the new endodermis after bifurcation suggests, however, that SCR has an important role in establishing the correct radial patterning in the developing dichotomous short roots as well.


This work was supported by a grant from the Academy of Finland, the Finnish Cultural Foundation and the Niemi Foundation. We thank Prof. Y. Helariutta for support and advice at the start of the work. The help with image processing given by Dr Mika Keränen to MR is highly appreciated.