The MADS-box gene DAL1 is a potential mediator of the juvenile-to-adult transition in Norway spruce (Picea abies)


(fax +46 18 559885; e-mail


Progression through the plant life cycle involves change in many essential features, most notably in the capacity to reproduce. The transition from a juvenile vegetative and non-reproductive to an adult reproductive phase is gradual and can take many years; in the conifer Norway spruce, Picea abies, typically 20–25 years. We present a detailed analysis of the activities of three regulatory genes with potential roles in this transition in Norway spruce: DAL1, a MADS-box gene related to the AGL6 group of genes from angiosperms, and the two LEAFY-related genes PaLFY and PaNLY. DAL1 activity is initiated in the shoots of juvenile trees at an age of 3–5 years, and then increases with age, whereas both LFY genes are active throughout the juvenile phase. The activity of DAL1 further shows a spatial pattern along the stem of the tree that parallels a similar gradient in physiological and morphological features associated with maturation to the adult phase. Constitutive expression of DAL1 in transgenic Arabidopsis plants caused a dramatic attenuation of both juvenile and adult growth phases; flowers forming immediately after the embryonic phase of development in severely affected plants. Taken together, our results support the notion that DAL1 may have a regulatory role in the juvenile-to-adult transition in Norway spruce.


Plants go through a period of vegetative growth before they reach maturity and start to reproduce. During this vegetative juvenile phase a gradual maturation occurs. Because of the modular growth of plants, a maturation gradient is formed from the basal, juvenile, parts formed by a young shoot apical meristem, to the apical parts formed more recently by an older meristem (Fortanier and Jonkers, 1976; Greenwood, 1995; Poethig, 1990). As many woody plants pass through several growth seasons before reproducing, the juvenile phase may be very long. In the conifer Norway spruce (Picea abies) it lasts 20–25 years or even longer (Chalupka and Cecich, 1997). During this time various traits change with age, and also with position along the stem of the tree (Greenwood, 1995).

In angiosperms, the flowering plants, three post-embryonic developmental phases have been recognized: a juvenile, an adult vegetative, and an adult reproductive (reviewed by Poethig, 1990). These different phases are accompanied by changes in morphological and physiological traits. The most important characteristic distinguishing the adult phase from the juvenile is the competence for reproduction, gradually gained over the adult vegetative phase, but manifested only in the adult reproductive phase. In Arabidopsis the molecular background of the floral transition has been well documented (reviewed by Araki, 2001; Battey and Tooke, 2002). Preceding flower development, environmentally and autonomously controlled pathways eventually lead to the activation of the transcription factor LEAFY (LFY), which is expressed only at low levels during the vegetative phase (Blázquez et al., 1997). When present at a certain threshold level LFY activates the MADS-box gene APETALA1 (AP1), a floral-meristem identity gene, and subsequently LFY together with AP1 activate the floral organ-identity genes (Parcy et al., 1998). Transition to flowering may be interconnected with the juvenile-to-adult phase change (Battey and Tooke, 2002). Interestingly, the long time of vegetative growth in the woody angiosperms citrus and poplar is reduced by constitutive expression of the Arabidopsis genes LFY and AP1 (Pena et al., 2001; Rottmann et al., 2000; Weigel and Nilsson, 1995).

Despite the long juvenile phase of conifers and woody angiosperms, and its importance as a major obstacle hindering efficient breeding, the mechanisms that control maturation are unknown (Martin-Trillo and Martinez-Zapater, 2002). This is in contrast to later stages of development: components of the reproductive development process of conifers and other gymnosperms have begun to be elucidated by the identification of genes phylogenetically and functionally related to key developmental regulators in angiosperms (Becker et al., 2000; Carlsbecker et al., 2003; Fukui et al., 2001; Mellerowicz et al., 1998; Mouradov et al., 1998a,b, 1999; Rutledge et al., 1998; Shindo et al., 1999, 2001; Sundström et al., 1999; Tandre et al., 1995; Winter et al., 1999, 2002; Zhang et al., 2004). These studies suggested an evolutionary conservation in regulatory processes between angiosperms and gymnosperms: the floral-organ identity genes of angiosperms have homologs that may have similar functions in conifer cone development. This is the case for both the C-class MADS-box genes, which in angiosperms control the identities of stamens and carpels, and the B-class genes that determine petal and stamen identity (Mouradov et al., 1999; Rutledge et al., 1998; Sundström and Engström, 2002; Sundström et al., 1999; Tandre et al., 1995, 1998). The identification of LFY-related genes in conifers suggested that reproductive initiation might also be controlled by similar mechanisms in angiosperms and conifers. However, unlike most angiosperms, conifers and other gymnosperms have two LFY-like genes: the ‘Leafy’ gene orthologous to the angiosperm LFY, and the ‘Needly’ gene, the ortholog of which is thought to have been lost in the evolution of angiosperms (Frohlich and Parker, 2000). Further, no ortholog of AP1 or the closely related SEPALLATA (SEP) genes has been identified in conifers or in other gymnosperms (Becker et al., 2000; Litt and Irish, 2003; Sundström et al., 1999; Tandre et al., 1995; Theissen et al., 2000). AP1 and SEP phylogenetically belong to a larger group of genes, which also includes AGL6 (Purugganan, 1997; Theissen et al., 2000). Two paralogous groups of AGL6 orthologs have been isolated from conifers, including the spruce DAL1 and DAL14 genes and the pine genes PrMADS2 and PrMADS3. DAL14 and PrMADS2 form a pair of orthologous genes active exclusively in reproductive tissues (Mouradov et al., 1998b; our unpublished results) and DAL1 and PrMADS3 form the second pair, both active in reproductive as well as in vegetative shoots (Mouradov et al., 1998b; Tandre et al., 1995).

In this study we examined the involvement of three regulatory genes, DAL1 and the two spruce LFY-like genes PaNLY and PaLFY, in the juvenile-to-mature transition process in Norway spruce.


Expression of DAL1 is initiated in mid-juvenile phase and increases with age

A previous study reported DAL1 expression in vegetative shoots as well as in developing cones in mature spruce trees, but not in seedlings (Tandre et al., 1995). To determine the age at which DAL1 expression is initiated we performed RNA blot analyses on vegetative shoots collected from juvenile and adult plants of different age; 2 month–4-year-old potted plants and field-grown 3 years and older plants. The experiment was performed with post-dormant field-grown plant material collected before and after bud-break, and with material collected during winter dormancy. In seedlings and young trees up to an age of 4 years expression of DAL1 was non-detectable. DAL1 expression was initiated in 5-year old trees and increased with age throughout the juvenile phase (Figure 1a,b). In cone-bearing trees, DAL1 expression was readily detected in male and female cones and in vegetative shoots. Similar results were obtained with samples collected before and after bud-break. No expression of DAL1 was detected in samples collected during winter dormancy (data not shown).

Figure 1.

Expression of DAL1, PaLFY, and PaNLY in reproductive, and in juvenile and adult vegetative shoots.
(a) RNA blot hybridization with a gene-specific DAL1 probe to RNA samples derived from vegetative shoots collected from pot-grown plants (lanes 2–4), field-grown plants sampled before bud-break (collecting date April 20; lanes 5–9), and after bud-break (May 18; lane 10), from developing pollen- and seed cones (May 10; lanes 11–12) and cambium (c.; lane 13). A histogram shows the DAL1 signal intensity normalized relative to the control UBI.
(b) Hybridization of a DAL1-specific probe to RNA from vegetative shoots from pot-grown plants (lanes 2–7), field-grown plants sampled after bud-break (May 18; lanes 8–13) and developing seed cones (May 10; lane 14). The histogram shows the DAL1 signal intensity normalized relative to the rRNA control. RNA from transgenic Arabidopsis plants expressing DAL1 was included as positive control [lanes 1 in (a) and (b)].
(c) RT-PCR using DAL1, PaLFY, or PaNLY-specific primers, and 10 ng total RNA as template. Samples were from seedlings (s.; lane 1), vegetative shoots collected before bud-break (April 18; lanes 2–6), after bud-break (May 18; lanes 7–8), and pre-dormant seed cones (Sept. 13; lane 9) and pollen cones (Oct. 23; lane 10). The histograms show the relative expression levels of the three genes normalized against UBI.
Age of the trees sampled is indicated in months (m) or years (y) in each lane.

The two spruce LFY-related genes are active throughout the juvenile phase

We cloned partial cDNA sequences of two spruce LFY homologs using RT-PCR and 3′-RACE with primers designed to match the Pinus radiata PrFLL and NEEDLY sequences. Resulting spruce LFY sequences were 93 and 94% identical over the nucleotide sequence to the corresponding pine sequence and are referred to as PaLFY and PaNLY respectively. As PaNLY could not be detected by RNA blot-hybridization techniques we compared the expression profile of DAL1 with that of PaLFY and PaNLY using RT-PCR. The temporal gradient of DAL1 expression over the juvenile phase detected by RNA blot experiments was confirmed by the RT-PCR experiments (Figure 1c). With RT-PCR we also detected small amounts of DAL1 mRNA in vegetative shoots of 3-year-old plants, not detected by RNA blot hybridization. In contrast to DAL1 PaLFY transcript was present in all samples tested (Figure 1c). In addition, PaNLY transcript was detected in all samples, but at an apparently low level in seedlings and vegetative shoots collected after bud burst from 30- and 50-year-old trees (Figure 1c). Thus, PaLFY and PaNLY expression is initiated in seedlings and both genes are expressed in the juvenile as well as in the adult phase.

DAL1 is expressed in an apical–basal gradient along the stem of juvenile trees

Because maturation in conifers is thought to occur in an apical–basal gradient along the stem of the tree (Borchert, 1976; Fortanier and Jonkers, 1976), we examined the level of DAL1, PaLFY and PaNLY expression in post-dormant, vegetative shoots collected after bud-break from apically and basally positioned whorls of branches of 4, 5, and 6-year old trees (Figure 2). Consistent with the results shown in Figure 1, expression of DAL1 was higher in 5 and 6-year-old trees than in 4-year-old trees. A clear difference in expression levels was also detected between shoots within the tree, collected from different positions along the stem. Apical shoots showed significantly higher level of DAL1 expression, and expression levels decreased towards the base of the tree in all trees examined (sign test; n = 8 P < 0.008; Figure 2b,c). This trend was most obvious in the 4-year-old trees. DAL1 expression levels also differed consistently between trees of similar age. One of the 4-year-old trees (III) showed no detectable DAL1 expression and the absolute level of transcript varied among the trees (Figure 2b). The expression levels of both PaNLY and PaLFY were relatively uniform between samples from trees of different ages, and between samples collected from branches in different positions along the stem of the trees (Figure 2b,c).

Figure 2.

Expression of DAL1, PaNLY, and PaLFY in vegetative shoots along the stem of juvenile trees.
(a) Schematic drawings of the growth pattern of 4-, 5- and 6-year-old spruce trees. The tree grows by the consecutive addition of one whorl of branches each year, as shown by numbers 1–6, which indicate the year of origin after germination of the stem section. Filled circles indicate positions of the shoots sampled for RNA.
(b) RT-PCR analysis of nine individual plants (I–IX), 4, 5, or 6 years old. The position in the tree of the sampled shoot is noted above each lane as indicated in (a). DAL1 RT-PCR reactions were performed with 10, 50 or 100 ng, and PaNLY or PaLFY reactions with 10 or 100 ng of DNase-treated total RNA.
(c) Relative expression levels of DAL1, PaLFY and PaNLY normalized against UBI is shown in the histograms. Results for DAL1 for the 4-year-old trees are shown from the RT-PCR with 100 ng and for the 5- and 6-year-old trees with 10 ng RNA as template. The expression level for sample A from each tree was set to 1 to facilitate comparisons of expression levels within individual trees. Results for sample I B are not shown (n.s.) due to artifactually low expression of the UBI control for this sample.

Temporal and spatial expression patterns of DAL1, PaNLY, and PaLFY

To further examine possible functions of DAL1 and the two LFY genes we performed mRNA in situ localization experiments on juvenile and mature vegetative shoots collected at primordium initiation in autumn and at the time of shoot elongation in spring. In addition, the expression pattern of DAL1 in young male and female cones was analyzed.

Expression of DAL1, in adult vegetative shoots in early autumn samples, was primarily detected in needle primordia and the cortex of the shoot (Figure 3a). Expression was low in the meristem and pith. This pattern of expression was similar in leading apical shoots and lateral shoots (data not shown). Juvenile shoots collected from 5-year or older trees showed an expression pattern similar to that of adult shoots (Figure 3b). In younger trees the transcript was not localized specifically to needles, but was detected at low levels (barely above background level) in the entire shoot (Figure 3c). In post-dormant adult shoots expression of DAL1 was detected at the flanks of the meristem, in bud-scale primordia, and vascular strands of elongating needles and stems (Figure 3d–g). Young pre-dormant female cones expressed DAL1 in the entire shoot, at high levels in ovuliferous scale primordia (Figure 3h). In later developmental stages expression of DAL1 was confined mainly to the developing ovuliferous scales (Figure 3i). No signal was detected in the meristem or in the apical parts of the bracts subtending the ovuliferous scales. After differentiation of ovules a hybridization signal was detected in the nucellus and the integuments (data not shown). As in female cones, DAL1 expression was detected in entire male cones at early stages of development; both in the microsporophylls and the shoot axis (Figure 3j). In later stages of pollen cone development DAL1 expression was maintained in all tissues examined (data not shown).

Figure 3.

In situ localization of DAL1, PaLFY, and PaNLY mRNA. The figure shows longitudinal sections of vegetative and reproductive shoots hybridized to a DAL1 probe (a–j), a PaNLY probe (k–m), or a PaLFY probe (n–q). Panel (a) shows an apical vegetative shoot from an adult tree (collection date Sept. 8), the insert shows a control hybridization to a DAL1 sense probe, (b) a vegetative shoot of a 5-year-old tree (Nov. 5), (c) a vegetative shoot from a juvenile tree less than 3 years old (Nov. 5), (d, e) a vegetative shoot from an adult tree (May 5), the insert shows a control hybridization to a DAL1 sense probe, (f, g) a needle, expression in vasculature indicated, (h) the apical part of a pre-dormant seed cone (Oct. 13), (i) the apical part of a post-dormant seed cone (April 19), (j) a developing pollen cone (Sept. 1), (k) a vegetative shoot, (l, m) a vegetative shoot with developing lateral bud (note signal in lateral bud indicated by an arrowhead), (n, o) a vegetative shoot (note signal in needle anlagen indicated by arrowheads), (p, q) a vegetative shoot, arrowheads mark signal in needles and in lateral buds. The pictures are dark-field micrographs, with exceptions of (e), (g), (m), (o), and (q) that are light-field micrographs. The signal appears as white grains in the dark-field micrographs. Aggregates of phenolic compounds appear as bright white areas. Note that (a, k, n) and (d, l, p) are consecutive sections of the same structures. Abbreviations: bract (b), bud scale (bs), crown region (cr), lateral bud (lb), meristem (m), microsporophyll (ms), needle (n), needle primordium (np), ovuliferous scale (os), sense (s). Bars represent 100 μm.

In adult vegetative shoots, in early autumn, expression of PaNLY was confined to the basal parts of the shoots; a zone just above the crown region, separating the new shoot from the previous year's shoot (Figure 3k). No PaNLY expression was detected in the meristem and apical needle primordia. In post-dormant developmental stages the signal was detected in lateral buds but not in needles, bud scales, or stem tissue (Figure 3l,m).

The PaLFY transcript was specifically localized to needle anlagen and young needle primordia in early autumn shoots (Figure 3n,o). In post-dormant shoots, expression of PaLFY was confined to bud-scale primordia, basal parts of differentiated needles, and lateral buds (Figure 3p,q).

No signal above background was detected with the sense probes corresponding to PaLFY, PaNLY or DAL1 in any tissue examined (Figure 3a,d, not shown).

Each of the three genes examined, thus, displayed a unique pattern of expression, implying distinct and different functions; the broad domains of DAL1 expression contrasted to the more specific expression patterns of PaNLY and PaLFY, in early as well as late developmental stages.

DAL1 causes early flowering in transgenic Arabidopsis plants

The spatial and temporal activity pattern of DAL1 suggests the gene to have a role relating to the maturation process in spruce. To further explore this hypothesis we examined the influence of DAL1 on the development of Arabidopsis thaliana, by constructing transgenic plants (ecotype Wassilevskija) expressing DAL1 under the control of the constitutive 35S promoter. Of 160 kanamycin-resistant primary transformant 35S::DAL1 plants, representing 20 independent lines, 65% (with representatives in most lines) displayed a markedly altered phenotype, including a short hypocotyl and small, dark green, cup-shaped, almost sessile cotyledons (class I transformant plants; Figure 4b). The remaining 35% showed no visible alteration in cotyledon morphology (class II plants). Ten of the 20 lines were studied in consecutive T2 and T3 generations. Phenotypic properties were inherited in successive generations. In several transformant lines seed germination was reduced and development of some seedlings was arrested at a very early stage (data not shown). Offspring of the primary transformant plants displayed a variation in the degree of phenotypic deviation from wild-type. The level of expression of DAL1, as determined by RNA blot experiments, was positively correlated with the severity of the phenotypic deviations in different lines (Figure 4a).

Figure 4.

Effects of DAL1 expression on transgenic Arabidopsis development.
(a) An RNA blot hybridized to a DAL1-specific probe. Samples are from wild-type seedlings (lane 1) and seven different lines of transgenic 35S::DAL1 seedlings, displaying different phenotypic aberrations, increasing in severity from lane 2 to 8. Lanes 2–4: class II plants ranging from wild-type like plants (2), to plants with curling of the rosette leaves (3 and 4). Lanes 5–8: class I plants; lanes 6, 7, and 8 representing class I plants with an extremely early flowering, like those shown in (f) and (g).
(b) A 35S::DAL1 seedling (right) and a wild-type seedling (left) of the same age. Note the difference in size and the sessile cotyledons and leaves of the transgenic plant.
(c) A 35S::DAL1 plant just after bolting. Note abaxial trichomes visible on the first true leaf and the development of a floral bud in the axil of this leaf.
(d) A class I 35S::DAL1 plant with a less strong phenotype, note the three rosette leaves showing abaxial trichomes.
(e) A 35S::DAL1 plant (right) and a wild-type plant (left) of equal age. The wild-type plant has produced a few flowers whereas the 35S: DAL1 plant has completed the life cycle.
(f, g) Examples of extreme phenotypic properties of 35S::DAL1 plants grown under long-day (f) or short-day (g) conditions. The plants have a few leaf or sepal-like organs, a pistil, and a few aberrantly shaped stamens.
(h) A 35S::DAL1 plant (right) and wild-type (left) of the same age, grown under short-day conditions. Note that the wild-type plant has not yet bolted.
Bars represent 1 mm.

Many of the class II plants were somewhat reduced in size compared with wild-type and had curled rosette and cauline leaves (data not shown). The class I plants were generally reduced in size; the inflorescence reaching a height of 2–15 cm, when compared with 30 cm or more for the wild-type (Figure 4e,h), and rarely produced lateral branches or secondary inflorescences. Occasionally, the normally indeterminate inflorescence meristem terminated by the production of floral organs in both class I and II plants.

Under inductive long-day light conditions wild-type plants bolted and entered the reproductive phase after the production of six to seven rosette leaves. In class I plants reproduction was initiated considerably earlier than in wild-type; in most lines it was after the formation of only two to four rosette leaves (Figure 4c,d). Leaves formed during the juvenile phase in Arabidopsis normally lack abaxial trichomes (Telfer et al., 1997). The 35S::DAL1 plants initiated abaxial trichomes early. Whereas wild-type developed adaxial trichomes after production of five or six leaves the class I plants bolting after the production of four to five leaves had abaxial trichomes from leaf 3 or 4, whereas those bolting after producing two to four leaves developed abaxial trichomes on their first leaf (Figure 4d).

Plants with very severe phenotypic aberrations were observed particularly in two independent transformant lines, both carrying two unlinked T-DNA inserts as judged by kanamycin resistance segregation (data not shown) and expressing DAL1 at high levels. These plants bolted without the previous development of rosette leaves and produced only one or a few sessile leaves on a very small inflorescence before the termination of the meristem in a floral structure (Figure 4f,g). Equally severe phenotypic alterations were observed in several independent primary transformant plants that could not be propagated and followed in consecutive generations. A minor proportion of the transformants displayed internode elongation between cotyledons and the first leaves (Figure 4d,f,g). This was more pronounced in plants grown under short-day conditions. Under this light regime the class I plants flowered ca. 3 weeks earlier than wild-type (Figure 4h), approximately at the same time and developmental stage as under long-day conditions.


The activity pattern of DAL1 suggests involvement in the juvenile-to-adult transition

In long-lived plants like trees the transition from a juvenile and non-reproductive to an adult growth phase may occur after many years of growth and, thus, requires the plant to be able to monitor life cycle progression over several growth seasons. The mechanisms by which the plant achieves this long-term monitoring of growth phase are unknown, but likely involve a gradual change in the transcriptional activities of regulatory genes (Battey and Tooke, 2002; Pena et al., 2001; Weigel and Nilsson, 1995). On the basis of data presented in this report we propose that the MADS-box gene DAL1 may be part of such a regulatory mechanism in the conifer Norway spruce.

DAL1 expression is activated in shoots of the juvenile tree at an age of 3–5 years and then increases with age. Expression is maintained at high levels in vegetative shoots, and later in reproductive shoots, the developing cones, of the adult tree. Activation of the gene, thus, precedes the entry of the tree into reproductive development, and the increased gene activity of DAL1 is correlated in time with the transition to the adult phase, the process which in conifers is referred to as the maturation of the tree. In addition to this temporal correlation, the activity of the gene also shows a spatial pattern along the axis of the tree that correlates with a similar pattern of maturation in several physiological and morphological characters in conifers. For example, the rooting capacity of cuttings decreases both with the age of the donor tree, and the distance from the base of the tree from which the cutting originates (Tousignant et al., 1995). The behavior of explants in tissue culture shows a similar dependence on the apical basal position within the tree (von Aderkas and Bonga, 2000; Greenwood, 1995). This pattern of maturation is most evident in the distribution of reproductive organs along the stem of the adult spruce tree, seed cones being born predominantly on branches close to the top of the tree, while pollen cones form on branches in more basal positions, and lower branches on the stem normally fail to develop reproductive organs.

The increase in DAL1 expression in a shoot with the distance between the shoot and the base of the tree is consistent with its expression being dependent on the ontogenetic age of the shoot. One possible mechanism for the generation of this type of pattern may be if DAL1 expression in the leader shoot meristem would increase with age, and lateral branch primordia derived from the meristem inherit the gene expression pattern of the leading shoot. Alternatively, positional information, e.g. the physical distance from the root, or metabolic signals that indirectly reflect the maturation status (Borchert, 1976), may be part of the regulation of the gene. The expression pattern of DAL1, together with the fact that other MADS-box genes in angiosperms, and likely in conifers, act as developmental regulators, suggests that this gene may have a direct role in the regulation of phase transition. As spruce trees, grown under normal conditions, set cones only after ca. 20–25 years the transcriptional activation of DAL1 alone appears insufficient to promote reproductive development. Instead DAL1 activity may be required for the shoot to attain reproductive competence, additional factors being required for reproduction to occur. Alternatively, DAL1 may activate reproduction directly, but only when above a certain threshold level of activity. If this hypothesis is correct, the onset of DAL1 expression in the leading shoot can be seen as the end point of the juvenile phase, as defined by Poethig (1990), and would define the entry into an adult vegetative phase.

DAL1 promotes phase change in Arabidopsis

Phenotypic alterations in Arabidopsis development caused by DAL1 expression are consistent with our hypothesis on DAL1 function in the juvenile-to-adult transition in spruce. The early production of leaves with predominantly abaxial trichomes indicates that DAL1 acts to induce flowering specifically by promoting a precocious development of adult features. Because high levels of expression of DAL1, in addition to shortening the juvenile phase, also cause a suppression of the adult phase, an alternative possibility is that DAL1 instead, or in addition, acts as a very strong inducer of flowering in Arabidopsis. Data derived from experiments of this kind are intrinsically difficult to interpret, in that the evolutionary distance between conifers and angiosperms is considerable, and the expression of a given gene in a heterologous host plant may cause artifactual effects, for example by unspecific interference with the host plant regulatory machinery. Early flowering is known to occur as a result of high level expression of a range of other MADS-box genes in transgenic plants, including the DAL2 and DAL10 genes from spruce (Carlsbecker et al., 2003; Tandre et al., 1998). The effects of DAL1 on flowering time, however, are quantitatively considerably more severe than those of most other MADS-box genes and the phenotypic effects of DAL1 on Arabidopsis development are highly specific. Furthermore, work on conifer and other gymnosperm B- and C-type MADS-box genes (Rutledge et al., 1998; Sundström and Engström, 2002; Tandre et al., 1998; Zhang et al., 2004) as well as on LFY homologs (Mouradov et al., 1998a; Shindo et al., 2001), and homologs to theKNOTTED-1 class of homeobox genes (Sundås-Larsson et al., 1998), have demonstrated functional conservation of transcription factors between gymnosperms and angiosperms. We, therefore, suggest that the phenotypic effects of DAL1 expression in Arabidopsis reflect a conservation of gene function between spruce and Arabidopsis.

DAL1 is a putative ortholog of the two Arabidopsis genes AGL6 and AGL13 that are both poorly characterized as regards function. AGL6 is preferentially expressed not only in flowers but also in stems (Ma et al., 1991) and AGL13 is expressed in the inflorescence and in siliques and at low levels in leaves and seedlings (Rounsley et al., 1995). Recently, an orchid AGL6 ortholog, OMADS1, was reported to cause early flowering by inducing the flowering time genes SOC1, FT, LFY, and AP1 when constitutively expressed in Arabidopsis (Hsu et al., 2003). The similarities in phenotype between 35S::DAL1 plants and 35S:OMADS1 suggest that DAL1 may act by similar mechanisms. AGL6-like genes are paralogous to the AP1- and SEP-like genes (Purugganan, 1997; Theissen et al., 2000), which have been implicated in flowering transition in a range of angiosperm species (Bonhomme et al., 1997; Chung et al., 1994; Elo et al., 2001; Immink et al., 2002; Yu and Goh, 2000) suggesting that this type of function might be common to AGL6, AP1 and SEP-type genes. We note that the phenotype reported for plants expressing a combination of SEP3 and AP1 under the constitutive 35S promoter (Pelaz et al., 2001) is very similar to that of the class I 35S::DAL1 plants.

DAL1 and the LFY-related genes in juvenile and adult development in spruce

In analogy with the proposed function of the AGL6-like genes in angiosperms, as activators of flowering time genes, one possible function for DAL1 in spruce would be to act as a transcriptional activator of the spruce LFY homologs, to promote reproductive development. Our expression data are not readily reconcilable with this hypothesis, as the two LFY-related genes are already active in young seedlings, and their transcription level does not appear to be influenced by the age of the tree and/or the level of DAL1 expression. In developing cones, however, DAL1 expression partly overlaps with the expression of both the LFY-related genes, allowing for an interaction between these genes in this stage of development.

The expression profiles of both the LFY genes and of DAL1 in spruce suggest the genes to function in juvenile vegetative as well as in adult reproductive development. A function for the LFY genes in juvenile development appears not to be unique to spruce. Both Pinus LFY genes are expressed in very early stages of development (Mellerowicz et al., 1998; Mouradov et al., 1998a), and in angiosperms LFY has, apart from determining flowering time (Blazquez et al., 1997), been implicated in the control of plant architecture in maize (Bomblies et al., 2003), leaf shape determination in tomato and pea (Hofer et al., 1997; Molinero-Rosales et al., 1999), and in shoot-apical meristem development in tobacco (Ahearn et al., 2001). Interestingly, also Am in Arabidopsis has a dual function, both in the vegetative-to-reproductive transition and in the development of the outer flower organs (Gustafson-Brown et al., 1994; Irish and Sussex, 1990; Mandel et al., 1992). Also the SEP genes are required for the development of the flower organs, as their gene products interact with the B-function gene products AP3 and PI, as well as with the C-gene product AG (Honma and Goto, 2001; Pelaz et al., 2000). It is, therefore, interesting to note that the expression of DAL1 overlaps in both pollen cones and seed cones with that of DAL2, the spruce C-class ortholog (Tandre et al., 1998) and in the pollen cones with DAL11, DAL12, and DAL13, spruce MADS-box genes related to the angiosperm B-genes (Sundström et al., 1999). DAL1 expression pattern also overlaps with DAL10, a MADS-box gene implicated in reproductive initiation in spruce (Carlsbecker et al., 2003). A possible function for DAL1 in spruce cones, thus, would be to act in concert with DAL10 and the B- and C-type MADS-domain proteins in the determination of micro- and megasporophyll organ identity in spruce.

In this study we have identified the MADS-box gene DAL1 as a potential regulatory component of the juvenile-to-adult phase transition process in Norway spruce. This may open interesting possibilities for exploring DAL1 as a tool to influence the length of the juvenile phase in conifers, thereby enhancing their breeding potential, and addressing the intriguing question of how plants monitor life cycle progression.

Experimental procedures

Plant material

Developing pollen cones, seed cones and vegetative shoots in different ontogenetic stages were collected throughout the growth season from several adult trees of Norway spruce [Picea abies (L.) Karst.] from locations outside Uppsala, Sweden. The approximate age of these adult cone-bearing trees was appreciated by counting whorls of branches. Vegetative samples were collected mainly from branches close to the apex of the tree, excluding the leading shoot, except for a 50-year-old tree from which samples were taken from lower pollen-cone carrying branches. Samples of vascular cambium, including differentiating phloem and xylem, were collected during the spring flush from the trunk of an adult tree, using a vegetable peeler. Juvenile material were from three sources, either plants grown from seeds in growth chambers under continuous light, obtained from nurseries, or from plants of known age in the field. Plants obtained from the nursery were kept in growth chambers or outdoors until budburst, when material was collected. For most experiments vegetative shoots from three to five individual plants were collected predominantly from the apical parts of the plants and pooled. Alternatively, vegetative shoots from individual plants were collected from easily distinguishable separate whorls of branches, avoiding subordinate branches, representing yearly growth. The seedling sample included all above ground tissue. All material for RNA preparations was collected and immediately frozen in liquid nitrogen and then kept at −70°C until use. Material for in situ hybridization was immediately fixed in fixation buffer (see below).

RNA blot analysis

Samples containing 30 μg of total RNA, prepared according to Chang et al. (1993), were size fractionated on 1% denaturating formaldehyde gels and blotted onto Hybond N+ membranes (Amersham Pharmacia Biotech, Sweden). The membrane were hybridized with a 32P-labeled 504 bp gene-specific PstI/ HindIII fragment of the DAL1 cDNA, spanning the K-box and the 3′-end of the gene. Hybridization and high stringency washing was performed according to membrane manufacturer's instructions. The membrane was rehybridized with a PCR-amplified labeled fragment specific for UBIQUITIN (UBI). The hybridization signal was detected using a BAS2000 Bio Analyzer (Fuji Photo Film, Tokyo, Japan) and the signal intensity was quantified using the Image Gauge v. 3.3 software, and normalized relative to the UBI or rRNA controls.


RT-PCR was performed using the Access RT-PCR kit (Promega, Madison, WI, USA) according to the manufacturer's recommendations. The amplification conditions were: 45 min at 48°C, 2 min at 94°C, followed by 35 cycles of 0.5 min at 94°C, 1 min at an annealing temperature determined by the primers used, and 2 min at 68°C, this was followed by a 7-min extension at 68°C. Ten, 50 or 100 ng of total RNA, DNase treated using SNAP total RNA Isolation Kit (Invitrogen, Leek, The Netherlands), was used as template in the RT-PCR reactions as indicated in the Results section. Selected PCR-positive products were sequenced. As control, parallel amplification reactions with primers for a UBI cDNA clone (Sundås et al., 1992) was performed. The UBI PCR conditions were as described but with 22 cycles, experimentally determined to represent late exponential phase (not shown). Densitometry of band intensities was performed using Image Gauge v. 3.3, and the intensities were normalized relative to the UBI control.

RNA in situ hybridization

Plant material for RNA in situ hybridization was dissected, fixed in a phosphate buffer containing 4% paraformaldehyde and 0.25% glutaraldehyde, and embedded in 100% Histowax (Histolab, Gothenburg, Sweden). In situ hybridization was performed essentially as described (Jackson, 1991). Sections of 7 μm were hybridized to 35S-labeled probes. For each slide 0.2 ng probe was applied, and after hybridization high stringency washing was performed. A 150-nucleotide fragment, covering parts of the 3′-transcribed region of DAL1, or 175 and 186 nt gene-specific fragments of PaNLY or PaLFY respectively, was used as probes. The probes were obtained by PCR, using a T7-promoter sequence attached to either forward or reverse primers, producing template for sense and antisense probes respectively. Primer sequences are available upon request. In vitro transcription was performed using the TransProbe T kit (Amersham Pharmacia Biotech) or the SP6/T7 Transcription Kit (Roche, Mannheim, Germany), with 8 μl purified PCR products, using High Pure PCR Product Purification Kit (Roche), as templates. The slides were coated with NBT2 emulsion (Eastman Kodak Co, Rochester, NY, USA) and exposed for 8 weeks. After development the sections were stained in 20% Gill's hematoxylin (BDH, Poole, UK), mounted and photographed in a Leica microscope. The pictures were processed using the Adobe Photoshop 7.0 software.

Cloning of spruce LFY genes

Poly(A)+ purified RNA, from pollen cones and seed cones, was isolated using PolyATract (Promega). cDNA was prepared with 5′/3′ RACE kit (Roche, Indianapolis, IN, USA), and used as template in PCR reactions with primers designed from the sequences of the two P. radiataLFY sequences (Mellerowicz et al., 1998; Mouradov et al., 1998a). Primer sequences are available upon request. The amplifications were performed with AmpliTaq Gold polymerase (Perkin Elmer Inc., Foster City, NJ, USA). Resulting PCR products were directly cloned using the TOPO-TA cloning kit (Invitrogen) and sequenced.

Transgenic plants

The DAL1 cDNA clone 3IB was corrected for an artifactual in-frame stop codon (Tandre et al., 1995). An ApoI fragment, covering the coding region, was inserted downstream of the 35S promoter in pHTT202 (Elomaa et al., 1993), and subsequently introduced into the Agrobacterium tumefaciens strain C58::pGV2260 by triparental mating. The resulting A. tumefaciens strain was used to transform Arabidopsis thaliana, ecotype Wassilevskija, by an infiltration protocol (Bechtold et al., 1993). Transformants were selected on growth media containing 50 μg ml−1 kanamycin. Phenotypes of the resulting plants were studied in T1, T2, and T3 generations grown under either long (16 h light and 8 h darkness) or short-day (9 h light and 15 h darkness) conditions.


Jens Sundström is thanked for fruitful discussions and critical reading of the manuscript. Agneta Ottosson and Gun-Britt Berglund are acknowledged for assistance with the plant transformation, and RNA preparations respectively. Lugnet's and Ed's nurseries are acknowledged for generous gifts of spruce plantlets, and Stefan Granath at Uppsala stift is thanked for information and for providing field grown juvenile plant material. This work was supported by grants from the Swedish Foundation for Strategic Research, the Swedish Science Research Council, and the Wallenberg Consortium North for functional genomics.

GenBank accession numbers:

AY701763 (PaLFY) and AY701762 (PaNLY).