StGA2ox1 is induced prior to stolon swelling and controlsGA levels during potato tuber development

Bjorn Kloosterman1,*, Christina Navarro2, Gerard Bijsterbosch1, Theo Lange3, Salome Prat2, Richard G. F. Visser1 and

Christian W. B. Bachem1

1Graduate School Experimental Plant Sciences, Laboratory of Plant Breeding, Department of Plant Sciences,

Wageningen University and Research Centre, PO Box 386, 6700 AJ Wageningen, The Netherlands,2Centro Nacional de Biotecnologıa, Consejo Superior de Investigaciones Cientıficas, Campus Universidad Autonoma de

Madrid, c/Darwin 3, 28049 Madrid, Spain, and3Institute of Plant Biology, Technical University of Braunschweig, D-38106 Braunschweig, Germany

Received 21 March 2007; revised 9 June 2007; accepted 25 June 2007.*For correspondence (fax +31 317 483457; e-mail bjorn.kloosterman@wur.nl).

Summary

The formation and growth of a potato (Solanum tuberosum) tuber is a complex process regulated by different

environmental signals and plant hormones. In particular, the action of gibberellins (GAs) has been implicated

in different aspects of potato tuber formation. Here we report on the isolation and functional analysis of a

potato GA 2-oxidase gene (StGA2ox1) and its role in tuber formation. StGA2ox1 is upregulated during the early

stages of potato tuber development prior to visible swelling and is predominantly expressed in the subapical

region of the stolon and growing tuber. 35S-over-expression transformants exhibit a dwarf phenotype,

reduced stolon growth and earlier in vitro tuberization. Transgenic plants with reduced expression levels of

StGA2ox1 showed normal plant growth, an altered stolon swelling phenotype and delayed in vitro

tuberization. Tubers of the StGA2ox1 suppression clones contain increased levels of GA20, indicating altered

GA metabolism. We propose a role for StGA2ox1 in early tuber initiation by modifying GA levels in the

subapical stolon region at the onset of tuberization, thereby facilitating normal tuber development and

growth.

Keywords: GA 2-oxidase, potato tuber development, gibberellin metabolism.

Introduction

Potato (Solanum tuberosum) plants have the capability of

producing underground storage organs called tubers. In

general, tuberization is promoted by short days with the

degree of response largely dependent on the genotype and

physiological age of the plant (Ewing and Struik, 1992).

Plants induced to tuberize produce a graft transmissible

signal in the leaves that is transported basipetally to the

growing stolon tip where it promotes tuberization (Gregory,

1956; Jackson et al., 1998). The exact nature of this trans-

missible signal is still unknown but it is most likely based on

a mixture of both inducing and inhibiting types of signals

(Ewing and Struik, 1992; Jackson, 1999; Martinez-Garcia

et al., 2001). Interestingly, there are several lines of evidence

suggesting a similar regulatory pathway for both flowering

and tuberization photoperiodic responses in plants

(reviewed in Rodriguez-Falcon et al., 2006).

Gibberellins (GAs) have long been implicated to have a

regulatory role in potato stolon growth and tuber initiation

(Booth, 1963; Ewing, 1987; Kumar and Wareing, 1974).

Applications of biologically active GAs or inhibitors of GA

biosynthesis have been shown to either delay or promote

tuber formation under tuber-inducing conditions (Jackson

and Prat, 1996; Vreugdenhil and Struik, 1989). Furthermore,

several studies have reported the existence of cross-talk

between the perception of day length and GA metabolism

(Amador et al., 2001; Chen et al., 2003; Jackson et al., 2000).

In the last decade, significant progress has been made in

identifying the genes involved in GA metabolism (for review

see Hedden and Phillips, 2000; Hedden and Proebsting,

1999; Lange, 1998). In potato, Northern blot analysis of three

isolated GA20-oxidase genes showed expression of all three

genes in both stolon and tuber tissue (Carrera et al., 1999),

362 ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd

The Plant Journal (2007) 52, 362–373 doi: 10.1111/j.1365-313X.2007.03245.x

yet little is known about the corresponding enzyme activities

in these tissues. Over-expression of StGA20ox1 results in

plants that show delayed tuber formation, whilst suppres-

sion clones are semi-dwarfed and tuberize earlier under

short-day conditions (Carrera et al., 2000). Whether biolog-

ically active GAs or the precursors themselves are part of the

transmissible signal or play a role in the production,

sensitivity or transport of the tuberization-promoting or

-inhibiting signals within the plant remains to be resolved.

The identification of novel genes regulating GA levels

through transcriptional control (StBEL5, POTH1) or altering

GA sensitivity (PHOR1) in potato plants, which were shown

to affect tuberization onset (Amador et al., 2001; Chen et al.,

2003; Rosin et al., 2003), indicates the existence of a complex

mechanism for controlling GA levels during the formation of

potato tubers.

Another important line of evidence linking GAs to the

regulation of tuber development is the observation of a

decrease in the levels of GA1 within the subapical region of

in vitro grown microtubers prior to visible swelling (Xu et al.,

1998a). When potato plants are induced to tuberize, stolon

growth ceases and the reduction in the levels of GAs is

thought to result in the longitudinal reorientation of the cell

microtubules, allowing lateral cell expansion and division

(Fujino et al., 1995; Shibaoka, 1993; Xu et al., 1998b). The

regulatory mechanism by which levels of bioactive GAs are

reduced prior to stolon swelling is, however, still unknown.

In many plant species, bioactive GAs are inactivated

through 2b-hydroxylation to produce inactive forms of

GAs. This step is catalysed by a member of the 2-oxoglut-

arate-dependent dioxygenase family, GA 2-oxidase (Thomas

et al., 1999). Gibberellin 2-oxidase has also been shown to be

able to react with precursors of active GAs, i.e. GA20 (Martin

et al., 1999). Similar to the GA 20-oxidase and GA 3-oxidase

gene families, the GA 2-oxidase gene family consists of

multiple members with often overlapping functions. In

recent years, a large number of GA 2-oxidases have been

identified in various plant species and are considered to be

important regulators of GA metabolism (Hedden, 2001).

In a previous experiments we found strong upregulation

of a potato GA 2-oxidase family member after the switch

from long- to short-day conditions within the stolon tip

(Kloosterman et al., 2005). Here we present data on the

analysis of the expression pattern of the isolated potato GA

2-oxidase (StGA2ox1) and its role in the process of tuber

formation.

Results

StGA2ox1 is upregulated prior to tuber development

and is located within the subapical stolon region

Large-scale gene expression analysis of potato tuber

development has shown differential expression of many

genes during the early stages of tuber formation (Kloos-

terman et al., 2005). One striking expression profile that was

observed in this study was the upregulation of a potato

expressed sequence tag (EST) with homology to a GA 2-

oxidase. Sequence analysis of the corresponding cDNA

clone revealed a complete open reading frame of 340 amino

acids, coding for a protein with a predicted molecular weight

of 38 kDa. The deduced amino acid sequence showed high

levels of homology to other GA 2-oxidase genes isolated

[Nicotiana tabacum, NtGA2ox5 (gbjABO70986) and Nerium

oleander, NoGA2ox3 (gbjAAT72916)], sharing respectively

80.5% and 73.5% of sequence similarity at the amino acid

level. The isolated potato gene described here was named

StGA2ox1. Quantitative real-time polymerase chain reaction

(RT-PCR) analysis of StGA2ox1 expression during tuber

formation (Figure 1a,b) confirms a large increase in tran-

script levels at tuber organogenesis (Stage 3, 72 � 10-fold)

and further tuber growth (Stage 7, 128 � 4-fold) in com-

parison with the non-induced stolon at day 0 (Stage 1).

Expression levels in the tuber gradually decrease when it

reaches its final size (Figure 1a,b; Stage 8). Analysis of

StGA2ox1 tissue specificity shows that the corresponding

gene is predominantly expressed in the tuber while minor

expression was found in root, stem and stolon (Figure 1c).

To determine the precise location of StGA2ox1 transcrip-

tion during tuber development, in situ hybridization was

performed on longitudinal sections of swelling stolons and

very young tubers (<0.8 cm). StGA2ox1 transcripts were

detected within the subapical region of the stolon but not in

the apical meristem (Figure 2). In the swelling stolon, signal

is observed surrounding the vascular bundle in the region

where cell division and expansion are thought to initiate.

During stolon swelling, lower levels of transcript were also

observed in the central pith and possibly the cortex

(Figure 2a). In the longitudinal cross section of a very young

tuber, the region of StGA2ox1 expression is extended to

regions of high mitotic activity associated with tuber growth

and initiation of the perimedullary region (Figure 2b). This is

in good agreement with the upregulated expression pattern

that we observed at tuber initiation and the absence of any

significant expression in dormant tubers, where no further

tuber growth occurs. Thus, StGA2ox1 primary expression in

the subapical stolon region at the early stages of tuber

formation suggests a primary role for this gene in regulating

GA levels during tuber development.

Over-expression and suppression of StGA2ox1 expression

results in altered growth characteristics and tuber

morphology

To further elucidate the function of StGA2ox1 we produced

35S-CaMV over-expression and suppression (post-tran-

scriptional gene silencing, PTGS) clones of potato var.

Karnico. After initial screening, the transgenic clones

Role for a GA 2-oxidase in potato tuber development 363

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

showing the severest phenotypes were selected. These were

confirmed for the presence of at least one construct insertion

and grown in 10 replicates for further analysis. Plants were

kept under long-day conditions (16-h light) for a period of

6 weeks after which they were switched to short days (8-h

light) to promote tuber formation. Tuber induction in the

Karnico variety is not under strict photoperiod control and

some tuber formation was observed in both transgenic and

control plants under long-day conditions. Individual repli-

cates of each clone did not tuberize synchronously and

therefore we were not able to statistically determine a dif-

ference in the time point of tuberization between the trans-

genic or control plants. However, several clear phenotypic

differences were observed.

Transgenic plants over-expressing StGA2ox1 are semi-

dwarfed and have smaller leaves (Figure 3a,b). These

plants exhibited reduced stolon growth, with on average

fewer and shorter stolons, and formed small tuber initia-

tions directly on the underground stem or developing from

lateral stolon buds (Figure 3e). Formation of these small

(a)

(b)

(c)

Figure 1. Quantitative real-time polymerase chain reaction measuring

StGA2ox1 gene expression during potato tuber development and tissue

specificity.

(a) Representation of the tuber development range consisting of eight

developmental stages harvested over a period of 25 days after a switch from

a 16-h to a 8-h light period (from Kloosterman et al., 2005; ªBlackwell

Publishing, reprinted with permission).

(b) Relative gene expression of StGA2ox1 during the potato tuber develop-

mental stages. Expression levels (DDCt) are represented relative to potato ubi3

gene and the first harvest (Stage 1).

(c) Comparison of the relative expression levels of StGA2ox1 in the different

tissues. StGA2ox1 expression levels (fold change) were calculated relative to

the tissues in which expression levels was highest (Tuber). Tissues studied

include 3-week-old tuber, root, stem, stolon, non-tuberizing stolon tip from LD

plants (NTS), shoot apex, petiole, leaf and dormant tuber (stored for 4 weeks

at 4�C). Error bars indicate the SD between three replicated measurements.

(a) (b)

(c) (d)

Figure 2. In situ localization of StGA2ox1 expression in longitudinal sections

of swelling stolons (a, c) and young tubers (b, d).

Sections were hybridized with StGA2ox1 antisense (a, b) and sense probes (c,

d). The presence of StGA2ox1 mRNA is indicated by the dark stain under

bright-field microscopy. Bars correspond to 1 mm.

364 Bjorn Kloosterman et al.

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

tuber incipients coincides with stolon initiation, 4 weeks

after planting, but these did not continue to increase in

size. As in control plants, tuber development and further

tuber growth was observed after 6 weeks of growth under

long-day conditions. On average, a significant (P < 0.05)

reduction in total tuber yield was observed in the over-

expression clones (O5, 96 � 8 g; O8, 117 � 16 g; O6,

146 � 3 g) in comparison with the control plants

(226 � 12 g), these clones having either fewer or smaller

tubers (Figure 3j,k).

The suppression clones did not show marked changes in

the aerial part of the plant (Figure 3c). However, they

produced ‘elongated’ tuber swellings or had a higher degree

of stolon branching in the underground parts (Figure 3f–i).

The elongated swellings were most clearly visible during the

early stages of tuber development (Figure 3g). This pheno-

type becomes less distinct during secondary growth (Fig-

ure 3g–i), as tubers produced by these clones show a more

normal morphology after full maturation (Figure 3j–l). Total

tuber yield was on average reduced in the two severest

StGA2ox1 suppression clones (S9, 151 � 24.6 g; S10,

148 � 21.3 g), although this effect was less severe than that

observed for the over-expression clones. A significant

reduction in tuber yield could not be observed for the third

suppression clone (S15), and thus no conclusions about a

negative influence of StGA2ox1 suppression on tuber yield

can be made. The number of tubers did not change

significantly in these clones.

Tubers from both sets of transgenic plants do not differ in

the duration of their dormancy period in comparison to the

control tubers when stored at 18�C. However, sprout growth

(i.e. internode length) was reduced in the StGA2ox1 over-

expression clones, which is in agreement with dwarf

phenotype of these clones (Figure 3m,n).

For confirmation of altered StGA2ox1 expression in the

transgenic clones, transcript levels of StGA2ox1 were

determined in active sink tubers using quantitative (q)RT-

PCR and they are plotted relative to the expression levels

found in control tubers (Figure 4). The transcript levels of

StGA2ox1 in the over-expression clones O5 and O8 were

(a) (c)(b)

(d) (e) (f)

(g) (h) (i)

(j) (k) (l) (m) (n)

Figure 3. Observed phenotypes for StGA2ox1

over-expression clones O5, O6, O8 and suppres-

sion clones S9, S10, S15 and untransformed

control.

(a) Photograph of 4-week-old plants of control

and three independent over-expression clones

(O5, O6 and O8) grown under controlled envi-

ronmental conditions. (b) Reduced leaf size

(4 weeks old) of StGA2ox1 over-expression

clones (O5 and O8) relative to the untransformed

control. (c) Comparison between plant growth of

StGA2ox1 suppression clone S15 and untrans-

formed control plant. (d)–(f) Stolon growth and

tuber set in the untransformed control (d), over-

expression clone O5 (e) and StGA2ox1 suppres-

sion clone S9 (f). The presence of small tuber

incipients in the over-expression clone is indi-

cated with arrows (e). (g)–(i) Different growth

stages of the elongated tuber swelling pheno-

type observed in StGA2ox1 suppression clones.

(j)–(l) Tuber yield of control and transgenic plants

harvested after the plants had completely died

off. Sprouting of tubers stored in the dark from

untransformed control plants (m) and StGA2ox1

over-expression clone O8 (n).

Figure 4. Fold-change in the expression level of StGA2ox1 in tubers of the

over-expression (O5, O8) and suppression clones (S10, S15) relative to the

untransformed control.

Error bars indicate the SD between three replicated measurements.

Role for a GA 2-oxidase in potato tuber development 365

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

increased respectively, by 13.2 � 0.7- and 9.9 � 1.3-fold,

whilst in the suppression clones S10 and S15 transcript

levels were reduced by 3.9 � 0.3- and 3.6 � 0.3-fold. Hence,

although the expression levels in the suppression clones are

reduced, low levels of StGA2ox1 transcripts were still

detected in these plants and therefore low levels of

StGA2ox1 activity would remain in these clones throughout

tuber growth.

Gibberellin levels in tubers of StGA2ox1 over-expression

and PTGS clones

Gibberellin levels were measured in active sink tubers of

both over-expression and suppression clones and untrans-

formed control plants in order to correlate the observed

phenotypes with altered levels of bioactive GAs. An increase

of GA20 together with a smaller increase in GA1 and GA8 was

observed in the tubers of both suppression clones, S10 and

S15 (Table 1). Measurements of GA levels in tubers at a later

stage of tuber growth also showed elevated levels of GA20 in

both suppression clones S10 (22 ng g)1 FW) and S15

(87 ng g)1 FW) as compared with control plants (0.01 ng g)1

FW), thus further confirming these results.

No clear differences were found for GA metabolites of the

non-13-hydroxylation pathway (Table 1; GA12, GA9, GA51,

GA4 and GA34) and therefore it is assumed that StGA2ox1 is

primarily active in the early 13-hydroxylation pathway

(Table 1; GA53, GA20, GA29, GA1 and GA8).

Interestingly, no obvious difference in GA content could

be observed in the over-expression clones (O5 and O8) in

comparison with control plants (Table 1). A small and

consistent reduction in the levels of GA29 was observed

in both over-expression clones, though its significance

remains unclear. The levels of GA5, GA3 and GA29 catabolite

were not determined in the transgenic plants and therefore

we were unable to determine whether this part of the GA

pathway had changed (Figure 5).

Expression levels of GA metabolism genes are altered

during development and in StGA2ox1 transgenic plants

In order to further investigate GA metabolism during tuber

development we looked at the expression levels of known

potato GA 20-oxidase and GA 3-oxidase genes (Figure 6).

We found a similar expression profile for StGA20ox1 as was

observed for StGA2ox1 during the eight developmental

stages, with the strongest increase of expression levels

occurring in the swelling stage (15.6 � 2.0-fold change rel-

ative to day 0), after which expression levels remain rela-

tively constant (Figure 6a). StGA20ox3 transcript levels are

also increased at tuber organogenesis but the expression

profile is characterized by a more gradual increase, reaching

peak levels at stage 6 (Figure 6b; 5.7 � 0.9-fold increase). In

potato, two GA 3b-hydroxylase genes have been isolated

from sprouts, StGA3ox1 (gbjAAK91507) and StGA3ox2

(gbjAAK91506), but little is known about their expression

levels or activity within the plant. Using qRT-PCR we were

able to detect expression of one of the GA 3b-hydroxylase

genes (StGA3ox2) in the non-tuberizing stolon tips and early

stages of tuber development (Figure 6c). After the switch

from long-day to short-day conditions (Figure 1a; Stages 1

and 2), transcript levels of StGA3ox2 decrease until no

transcripts could be detected from tuber Stage 5 onwards.

Like StGA2ox1 and StGA20ox1, expression levels of

StGA3ox2 have already changed after the switch from long-

to short-day conditions prior to visible swelling and with the

strongest change in expression levels occurring at stolon

Table 1. Endogenous gibberellin (GA) levels in potato tubers fromuntransformed control plants and StGA2ox1 over-expression andsuppression clones

GA levels (ng g)1 FW)

Control Clone O5 Clone O8 Clone S10 Clone S15

Non-13-hydroxylation pathwayGA12 0.03 n.d. 0.05 0.02 n.d.GA9 0.01 n.d. 0.02 0.01 n.d.GA51 0.04 0.01 0.03 0.03 n.d.GA4 0.06 0.16 0.05 0.04 0.05GA34 0.01 0.01 0.01 0.01 0.02

13-Hydroxylation pathwayGA53 0.04 0.01 0.01 0.01 0.01GA20 0.01 n.d. n.d. 0.49a 0.33a

GA29 0.27 0.13 0.15 0.23 0.24GA1 0.01 0.02 0.04 0.09a 0.10a

GA8 0.03 0.01 0.02 0.23a 0.21a

FW, fresh weight; n.d., no dilution of the internal standard.aStatistically significant differences between control plants andStGA2ox1 suppression lines, Student’s t-test (P < 0.05).

Figure 5. Schematic overview of the major genes involved in the 13-hydrox-

ylation pathway in plants.

Steps catalysed by the different gibberellins (GA)-oxidase genes are

indicated; GA20-oxidase (20ox), GA3-oxidase (3ox), GA 2-oxidase (2ox).

Biologically active GAs are boxed.

366 Bjorn Kloosterman et al.

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

swelling (days 7–8). The strong induction in gene expression

of StGA2ox1, StGA20ox1, StGA20ox3 and the downregula-

tion of StGA3ox2 during tuber formation indicates the

existence of a complex regulatory mechanism of GA

metabolism at the transcriptional level. However, it is

important to realize that genes that are only expressed in

specific cell types, for instance meristematic cells, may

appear to be downregulated during tuber growth because of

a changing ratio in cell number between such specialized

cells and the newly formed parenchymatic cells that form

the bulk of the mature tuber.

To screen for changes in expression of the GA metabolism

genes as a result of altered StGA2ox1 transcript levels, we

also measured their expression in tubers of the different

transgenic clones. Interestingly, we found a strong reduction

of StGA20ox1 and StGA20ox3 expression levels in both

suppression clones (Figure 7). In particular StGA20ox3

shows a strong decrease in expression for suppression

clones S10 and S15, 22.9 � 0.9- and 22.6 � 1.7-fold respec-

tively, compared to the control plant. This downregulation

most likely reflects a negative feedback control because of

the elevated levels of either GA1 or GA20 found within the

suppression clones (Table 1). We did not find a significant

difference in the expression levels of the GA 20-oxidase

genes in both over-expression clones, consistent with the

seemingly unaltered GA status of the tuber. We were not

able to detect any transcripts of StGA3ox1 and StGA3ox2

within tubers of any of the transgenic plants analysed.

In vitro tuberization of StGA2ox1 transgenic plants

As the phenotype produced by the suppression clones was

partially lost because of secondary growth of the tuber, we

decided to study early tuberization events using an in vitro

tuberization assay (see Experimental procedures) in which

secondary growth is absent (Xu et al., 1998b). Interestingly,

in our in vitro experiment a clear increase in the induction of

tuber formation was found in the over-expression clones

and a delay in tuber formation in the suppression clones

(Figure 8). Already after 1 week’s incubation in the dark,

95.1% and 84.6% of the nodal stem cuttings from over-

expression clones O8 and O5, respectively, had formed

microtubers, whilst in suppression clone S10, tuber forma-

tion was still absent. Only after 3 weeks was a strong

increase in the amount of tuber formation in both suppres-

sion clones detected. In our soil-grown plants we were

not able to detect a difference in the time point of tuber

Figure 6. Quantitative real-time polymerase chain reaction of a number of

gibberellin (GA) metabolism genes during potato tuber development.

(a)–(c) RNA derived from eight tuber developmental stages harvested over a

period of 25 days after a switch from a 16-h to an 8-h light period (Figure 1a):

(a) StGA20ox1 (AJ291453); (b) StGA20ox3 (AJ291455); (c) StGA3ox2

(AY039110). Expression levels of GA metabolism of the eight developmental

stages were calculated relative to the potato ubi3 gene and the first harvest

(Stage 1). Error bars show the SE of the mean of three measurements.

Figure 7. Impact of altered transcript levels of StGA2ox1 on the expression of

GA20-oxidase genes, StGA20ox1 and StGA20ox3. Expression levels of

StGA20ox1 and StGA20ox3 were measured in sink tubers of over-expression

clones (O5 and O8) and the suppression clones (S10 and S15). Expression

levels in the transgenic plants were calculated relative to the expression levels

of the potato ubi3 gene and the untransformed control. Error bars show the

SE of the mean of three replicated measurements.

Role for a GA 2-oxidase in potato tuber development 367

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

initiation, although early small tuber incipients were obs-

erved in the over-expression clones. Moreover, morpho-

logically significant differences between the individual

transgenic clones and control plants could be observed

(Figure 9). In the over-expression clones, a large percentage

of the formed tubers (78.9%) are sessile where stolon growth

is largely inhibited. The suppression clones however, have a

much larger percentage of lateral tuber formation (43.1%)

and produce a small number of sessile tubers (7.6%) in

comparison to the control plants. Furthermore, while tubers

of the control plantlets generally have a round shape in

which the first and second nodes are incorporated in the

radial swelling, the suppression clones frequently have an

elongated phenotype between the first and second nodes of

the in vitro tuber similar to the phenotype observed in the

in vivo plants (Figure 3g). In addition, lateral tubers were

sometimes formed directly on the tuber eyes.

In an additional experiment, we applied different concen-

trations of GA3 and the GA biosynthesis inhibitor, ancym-

idol, to the tuber-inducing medium (Figure 9). The control

plants supplemented with 0.1 lM GA3 produced fewer

sessile tubers (17.0%) in comparison with the normal

tuber-inducing medium (47.0%) and had an increased

number of lateral tubers, indicating that lateral tuber forma-

tion is at least in part regulated by GA levels. Reduction in

the number of sessile tubers is further enhanced in medium

supplemented with 1.0 lM GA3 where the total number of

tubers formed decreases. Supplementing the over-expres-

sion clone O8 with 0.1 lM GA3 does not seem to have a large

effect on the number of sessile tubers formed, 78.9% and

72.1% respectively. However, when adding 1.0 lM GA3 the

percentage of sessile tubers formed decreased significantly

(42.5%) together with an increase in longitudinal stolon

growth. These findings are consistent with the inhibiting

effect of GA3 on tuber formation in potato. Because of the

expected increase of StGA2ox1 activity in the over-expres-

sion clones, higher levels of GA3 supplementation may be

required to cause a significant effect on tuber formation. In

suppression clone S15, a reduction in the percentage of

apical tubers was observed relative to the amount of GA3

supplied, but this did not lead to an increase in lateral tuber

formation as was found for the control. In conjunction with

these observations, when supplementing the tuber-inducing

medium with 0.5 lM ancymidol, a GA synthesis inhibitor,

stolon growth is largely absent in the over-expression clones

and control, while in the suppression clones the effect is

much smaller and longitudinal stolon growth persists. If GA

synthesis in the stolon is absent because of the presence of

an inhibitor, GA levels present within in the nodal stem

cuttings at the time point of harvest appear to be longer lived

in the suppression clones compared with the over-expres-

sion and control plantlets, possibly because of a reduction in

StGA2ox1 activity.

Taken together these findings suggest that GA levels are

reduced in the over-expression clones, resulting in inhibition

of stolon growth and formation of sessile tubers, whilst in

Figure 8. In vitro tuberization of single node

cuttings from untransformed control plants

(C, ), StGA2ox1 over-expression clones (O8,

and O5, ) and suppression clones (S10, and

S15, ). A large number of single node cuttings

(n) were collected from 4-week-old in vitro grown

plantlets and incubated on tuber-inducing med-

ium in the dark at 18�C. The relative percentage

of tuber formation was scored on a weekly basis.

Figure 9. Distribution of the type of tubers

scored in an in vitro tuberization experiment of

untransformed control plants (C), StGA2ox1

over-expression clone (O8) and StGA2ox1 sup-

pression clone (S15). Single node cuttings were

harvested from 4-week-old in vitro grown plant-

lets and incubated in the dark (18�C) on tuber-

inducing medium (TI) or tuber-inducing medium

supplemented with either 0.1 lM GA3, 1.0 lM

GA3 or 0.5 lM ancymidol and scored after

6 weeks. Images on the right panel illustrate the

type of tubers that were scored.

368 Bjorn Kloosterman et al.

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

the suppression clones GA levels remain high within the

stolon tip delaying tuber formation and promoting lateral

stolon growth and subsequent lateral tuber formation.

Discussion

StGA2ox1 is expressed during tuber formation in the

subapical region affecting tuber growth

StGA2ox1 is strongly upregulated at stolon swelling and

tuber formation, indicating a role in regulating GA levels at

tuber organogenesis. It has been well documented that GA

levels promote longitudinal cell division and elongation

within plants by affecting the orientation of microtubules

(Fujino et al., 1995; Sanz et al., 1996; Shibaoka, 1993). Xu

et al. (1998b) found that prior to in vitro tuber formation a

strong reduction in GA1 levels was associated with a reori-

entation of the plane of cell division and expansion, result-

ing in radial swelling of the subapical part of the stolon. In

our study, transcription of StGA2ox1 is strongly upregulated

at tuber formation and shown to be restricted to the sub-

apical region of the stolon tip in regions where active cell

divisions take place (Figure 2).

When reducing the number of StGA2ox1 transcripts in the

suppression clones we observe an elongated tuber pheno-

type. In these plants, the required reduction of GA levels at

tuber organogenesis may be insufficient, resulting in only a

partial reorientation of the plane of cell division and expan-

sion and thereby causing the elongated swelling morpho-

logy (Figure 3g–i). During secondary tuber growth, cell

divisions are randomly orientated in the perimedullary

region, often masking the initial elongated swelling pheno-

type. Together, these findings clearly indicate an important

role for StGA2ox1 in regulating the plane of cell division by

controlling GA levels within the subapical stolon region in

the early stages of tuber organogenesis. StGA2ox1 is

considered to be a primary response gene (i.e. a tuber

identity gene) as its expression specifically increases in the

subapical region only in cells undergoing transition to tuber

fate after tuber formation is induced. This is in agreement

with the absence of a significant difference in the time point

of tuberization between the control plants and the StGA2ox1

over-expression and suppression clones in the in vivo

experiment.

In sharp contrast, a clear difference in the time point of

tuberization was observed in the in vitro tuberization exper-

iment (Figure 8) in which either a clear induction or delay in

the rate of in vitro tuber formation was observed for the

over-expression and suppression clones, respectively, con-

sistent with altered GA content (Jackson and Prat, 1996;

Vreugdenhil and Struik, 1989; Xu et al., 1998a). Under in

vitro conditions, high sucrose levels are thought to be the

major signalling component in tuber formation, as sucrose

concentrations correlate negatively with GA levels in the

stolon tip (Xu et al., 1998a). Hence, under in vitro conditions

the time point of tuber formation appears to rely primarily

on sucrose/GA interactions within the stolon tip (Xu et al.,

1998a) whilst under in vivo conditions additional regulatory

factors are required to induce tuber formation. Interestingly,

the ratio of lateral in vitro tuber formation was significantly

increased in the suppression clones (Figure 9), indicating

the presence of a gradient of biologically active GAs within

the stolon. It has been suggested that concentrations of GA1

vary throughout the stolon, with the highest concentration

located in the stolon tip (Jackson, 1999).

Similarly, we did not observe a difference in the duration

of the dormancy period for both StGA2ox1 over-expression

and suppression clones in comparison with tubers of the

control plants. However, we did find that early sprout growth

(i.e. internode length) is reduced in the StGA2ox1 over-

expression clones consistent with the dwarfed plant pheno-

type (Figure 3a,m,n). We therefore conclude that StGA2ox1

does not play a role in regulating the length of the dormancy

period within the tuber.

Reduced transcript levels of StGA2ox1 result in increased

levels of GA20, GA1 and GA8

Within the suppression clones an apparently large increase

of GA20 and a less obvious increase of GA1 and GA8 were

observed (Table 1). The increased substrate availability of

GA20 most likely results in an increased flux towards the

synthesis of both GA1 and GA8. As the levels of the biolo-

gically inactive GA8 are elevated in the StGA2ox1 suppres-

sion clones and not in the over-expression clones, it seems

unlikely that StGA2ox1 acts directly on GA1. Surprisingly, we

did not find a significant increase or reduction of bioactive

GAs in the studied tubers of the over-expression clones.

However, the phenotypes of these plants, semi-dwarfed,

reduced stolon growth and earlier in vitro tuberization, are

consistent with a reduction in overall bioactive GA content,

indicating that very small changes in GA homeostasis may

already be sufficient to alter plant morphology.

The oxidation of GA29 to produce GA29 catabolite has

been shown to be catalysed by a GA 2-oxidase in pea shoots

(Lester et al., 1999; Martin et al., 1999). We did find a small

reduction in the levels of GA29 in the over-expression clones

suggesting a role for StGA2ox1 in this conversion, but no

supporting evidence was found with data obtained for the

suppression clones in which no difference were observed.

In maize shoots it was shown that GA3 can be synthesized

from GA20 via GA5 (Fujioka et al., 1990; Spray et al., 1996).

Abdala et al. (2002) measured significant levels of GA3 in

potato foliage, root, stolon and tuber. However, there are no

detailed data available on changing GA3 levels during the

transition of a stolon into a tuber as was reported for GA1

levels (Xu et al., 1998a) and as a result the role of GA3 in

tuber formation remains unclear. The GA 2-oxidase gene

Role for a GA 2-oxidase in potato tuber development 369

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

family in general consists of multiple members which are

often functionally redundant (Hedden and Proebsting, 1999;

Thomas et al., 1999) and therefore we cannot exclude

complementation of StGA2ox1 activity by a different family

member in the transgenic plants.

Transcriptional control of GA metabolism genes during

potato tuber development

A long-standing issue has been the understanding of how

endogenous GA levels are controlled during the process of

tuber formation. Based on the observed phenotype in the

StGA2ox1 suppression clones and its subapical localization,

inactivation of bioactive GAs or GA intermediates through

the activity of StGA2ox1 is an important regulatory step in

controlling endogenous GA levels in the subapical stolon

region. However, no StGA2ox1 transcript signal was

detected in the stolon apex, indicating that another control

mechanism is required for reducing GA content in the stolon

apex as reported by Xu et al. (1998a). We found that the

transcript levels of an enzyme catalysing the final step in the

biosynthesis of bioactive GAs, StGA3ox2, was strongly

downregulated at tuber onset (Figure 6c). Both the upreg-

ulation of a GA inactivation gene (StGA2ox1) and the

downregulation of a GA biosynthesis gene (StGA3ox2)

would allow a rapid reduction in GA content within the

swelling stolon required for normal tuber formation. Inter-

estingly, two genes involved in the synthesis of GA-precur-

sor GA20 (StGA20ox1 and StGA20ox3), show an upregulated

expression profile during the early stages of tuber devel-

opment, although less strong in comparison to StGA2ox1

(Figures 1b and 6a,b).

As both StGA20ox1 and StGA20ox3 transcript levels

have been shown to be under negative feedback control of

GA levels (Carrera et al., 1999), the increased expression of

both genes at tuber organogenesis may be the result of

feedback regulation because of a reduction in the concen-

tration of specific GAs. Similarly, the observed decrease in

the transcription levels of the same genes in the transgenic

tubers of StGA2ox1 suppression clones, having increased

levels of GA20, GA1 and GA8 (Table 1), can be explained by

the same feedback mechanism (Figure 7). Conversely,

several studies have shown that changes in StGA20ox1

transcript levels in the stolon can be directly altered

through over-expression of homeobox genes POTH1 and

StBEL5 that are able to bind to the StGA20ox promoter,

reducing its activity (Chen et al., 2003, 2004). As both

StBEL5 and StGA20ox1 are upregulated at tuber formation,

these findings bring up an interesting paradox as to the

primary transcriptional control mechanism and relative

importance of StGA20ox1 during tuber development within

the stolon and developing tuber. StGA20ox1 is highly

expressed in the leaves and a primary role for this gene in

controlling tuber induction in the leaves is further sup-

ported by the finding that over-expression of StGA20ox1

under control of a leaf specific promoter (STLS1) delays

tuber formation in a similar fashion as constitutive (CaMV

35S) over-expression lines (Carrera et al., 2000).

The relatively strong increase and decrease in transcript

levels of StGA2ox1 and StGA3ox2 at tuber onset raises the

question whether there is a common mechanism controlling

their expression. When decapitating pea shoots, the reduc-

tion of auxin levels throughout the plant leads to a decrease

of PsGA3ox1 expression and an increase of PsGA2ox1

expression levels (Ross et al., 2000). The effects can be

opposed by the application of indole-3-acetic acid (IAA)

indicating a regulatory role in GA metabolism. These

findings may support a similar role for IAA in regulating

StGA3ox2 and StGA2ox1 expression levels during potato

tuber development.

The changes in gene expression already observed after

the switch to short-day conditions within the stolon tip prior

to visible swelling implies that transcription of StGA2ox1,

StGA20ox1 and StGA3ox2 is closely associated with the

tuber induction signal. However, the mechanism by which a

tuber induction signal produced in the leaves translates to

GA metabolism or sensitivity within the stolon tip is still

under investigation, although much progress has been

made in recent years.

As mentioned, StBEL5 and its protein partner POTH1 can

alter tuber formation by mediating hormone levels in the

stolon tip. It was recently shown that transcripts of StBEL5

are upregulated in response to short-day conditions and are

present in phloem cells, capable of moving across a graft

union towards the stolon tip revealing a long-distance

signalling pathway (Banerjee et al., 2006). Different KNOX

proteins have been implicated in regulation of GA catabolic

enzymes to confine GA activity to the differentiating leaf

primordia (Jasinski et al., 2005). In another study, a MADS-

box transcription factor, AGAMOUS-like 15, has been shown

to bind to the regulatory region of AtGA2ox6, controlling its

expression level (Wang et al., 2004). Whether the expression

level of StGA2ox1, and thereby its activity, is also controlled

through similar types of regulatory protein needs to be

investigated. In addition, post-transcriptional control of

gene expression, protein modification and degradation,

provides the plant with additional tools for fine-tuning

of the GA metabolism pathway during potato tuber

development.

The data presented in this paper on the expression of

StGA2ox1 and other GA metabolism genes during potato

tuber development provide novel insights into GA-regulated

potato tuber formation at the organ level. From our expres-

sion studies in transgenic plants, we conclude that

StGA2ox1 fulfils a central role in the transition from longi-

tudinal stolon growth to tuber initiation by regulating GA

levels in the subapical stolon region facilitating radial

growth.

370 Bjorn Kloosterman et al.

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

Experimental procedures

Cloning of StGA2ox1 and sequence analysis

The sequence of the potato GA 2-oxidase was obtained bysequencing cDNA clone BI176613 deriving from an in vitro grownmicrotuber EST library (cv. Bintje). Sequence analysis revealed thatthe cDNA was full length and coded for a protein of 340 amino acids insize and was named StGA2ox1 (gb|EU003995). The complete codingsequence of StGA2ox1 was amplified by PCR (forward primer5¢-CACCTATGGTTGTTTTGTCTCA-3¢ and reverse primer 5¢-TGTCA-TGATTGAGCATTCT-3¢). The forward primer contains a 5¢-CACCpartial overhang to facilitate cloning of the PCR product inpENTR/SD/D-TOPO vector (Invitrogen, http://www.invitrogen.com/).StGA2ox1 was further subcloned by a LR recombination reactioninto Gateway plant transformation destination vector pk7WG2 andpk7GWIWG2 (Plant Systems Biology, University of Ghent, Belgium)(Karimi et al., 2002). The two resulting expression constructs werenamed pk7GA2ox1_O and pk7GAox1_S. pk7GA2ox_O expressionresults in heterologous expression of StGA2ox1 gene under controlof the 35S-CaMV promoter. The Pk7GA2ox_S expression constructconsists of a full-length StGA2ox1 inverted repeat harbouring anintron behind the 35S-CaMV promoter that when expressedproduces a hairpin structure resulting in PTGS.

Plant transformation and regeneration

pk7GA2ox1_O and pk7GAox1_S were transformed into Agrobacte-rium tumefaciens strain Agl0 using electroporation. In vitro shootsof the S. tuberosum cv. Karnico were used for A. tumefaciens-mediated transformation (Visser et al., 1991). After regeneration ofin vitro shoots on selective kanamycin Murashige and Skoog (MS)medium (100 mg L)1) (Murashige and Skoog, 1962), 45 indepen-dent suppression and 9 over-expression clones were transferred tosoil and grown in the greenhouse to check for altered expression ofStGA2ox1. Integration of at least one copy of either construct in theselected transgenic plants was confirmed by Southern blottingusing part of the 35S-CaMV promoter as probe. From these trans-genic plants, three strong over-expression and suppression cloneswere selected for further analysis. Plants regenerated fromuntransformed in vitro shoots were used as wild-type controls.

Plant material growth conditions and tuber harvest

Ten replicates of the in vitro control plants and of the three selectedover-expression and suppression clones were transferred to soiland grown in a climate chamber (16-h light 20�C and 8-h dark at18�C) for a period of 6 weeks after which plants were transferred toshort-day conditions (8-h light and 16-h dark period). Tubers fromtwo repeats of each plant were harvested 6 and 12 weeks after theswitch to short days and immediately frozen in liquid nitrogen.Frozen tubers were ground into a fine powder and used forexpression studies and GA measurements. Remaining plants wereleft to senescence and die after which the number of tubers and totaltuber weight was determined from at least three repeats from eachtransgenic clone and control plants. Significance for difference intotal tuber weight between the different transgenic clones andcontrol plant was determined using Student’s t-test (P < 0.05).

RNA isolation and quantitative RT-PCR

For expression studies within the different transgenic clones andcontrol plants, total RNA was isolated from tubers as described by

Bachem et al. (1996). For expression studies of the GA-oxidasegenes during the potato tuber developmental time series and in thedifferent potato tissues, total RNA was used from a previous study(Kloosterman et al., 2005). One additional tissue sample wasincluded, namely the shoot apex, and was harvested from 8-week-old greenhouse-grown plants (S. tuberosum cv. Karnico).

Relative expression levels of genes were determined by real-timeqRT-PCR on a Perkin Elmer Abi Prism 7700 sequence detector(Perkin Elmer, http://www.perkinelmer.com/) following the protocoldescribed in Kloosterman et al. (2005). Potato ubiquitin primers(ubi3) were used as a control. Relative quantification of the targetRNA expression level and SD was calculated from three replicatedmeasurements of the same sample using the comparative Ctmethod according to User Bulletin no. 2 (ABI Prism 7700 SequenceDetection System, December 1997, Applied Biosystems, http://www.appliedbiosystems.com/). The primer sequences for thegenes studied are as follows: StGA20ox1 (gb|CAC13036) forwardprimer 5¢-CGGCCCAACAAGCATCTAAG-3¢, reverse primer 5¢-AAGC-CATGACTCCGACACG-3¢; StGA20ox3 (gb|CAC13038) forwardprimer 5¢-GCAATGCCATGAGCACCC-3¢, reverse primer 5¢-GGCTC-AATCCCAAAAGTTCCA-3¢. StGA3ox2 (gb|AAK91506) forwardprimer 5¢-AGCTCATGTGGTCCGAAGGA-3¢; reverse primer 5¢-CGG-ACAAGCCGGGTAAGAAT-3¢; StGA2ox1 (gb|EU003995) forwardprimer 5¢-AGGCACAGAGTGATCGCAGAT-3¢, reverse primer5¢-TGGTGGCCCTCCAAAGTAAA-3¢; ubi3 (gb|L22576) forward pri-mer, 5¢-TCCGACACCATCGACAATGT-3¢, reverse primer 5¢-CGACCA-TCCTCAAGCTGCTT-3¢.

In situ hybridizations of StGA2ox1

Tissue preparation and in situ hybridization experiments were car-ried out according to Jackson et al. (1991), with small modificationsas described in Bradley et al. (1993) and Coen et al. (1990). Probesused to detect the StGA2ox1 transcript were prepared from a pGEM-T-Easy vector (Promega, http://www.promega.com/) that containeda 786-bp fragment of the StGA2ox1 cDNA inserted in sense and inantisense orientations. This insert was obtained by PCR amplifica-tion using the primers 5¢-TTTTCCAACAAACAACTATGGTTG-3¢ and5¢-CTGCAATGAGTCACCAACATTG-3¢. After T7 transcription, theprobe was hydrolysed for 40 min at 60�C in carbonate buffer, toobtain fragments of approximately 180 bp.

The GA measurements

For quantitative determination of endogenous GAs in tubers of thetransgenic clones (O5, O8, S10 and S15) and untransformed con-trols (C), tubers from three independent replicates were harvestedfrom 12-week-old plants, mixed and immediately frozen in liquidnitrogen before being ground into a fine powder. Five grams of FWtubers of each sample was spiked with 16, 17-d2-GA standards (5 ngeach; from Prof. L. Mander, Canberra, Australia). Samples wareextracted, purified, derivatized and analysed by gas chromatogra-phy mass spectrometry using selected ion monitoring as describedelsewhere (Lange et al., 2005). Significance for difference in GAcontent between the transgenic clones and control plant wasdetermined using Student’s t-test (P < 0.05).

In vitro tuberization assay

A modified version of in the in vitro tuberization method describedby Hendriks et al. (1991) was used for microtuber production.Instead of using in vivo-grown plants grown for single node harvest,in vitro plantlets were used. The upper three nodal stem sections

Role for a GA 2-oxidase in potato tuber development 371

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

containing a single axillary bud from 4-week-old in vitro-grownplantlets were harvested from two independent over-expressionand suppression clones and control plants. For each clone, morethan 100 stem sections (n) were put directly on the tuber-inducingmedium and incubated in the dark (18�C). Tuber-inducing mediumconsists of modified MS medium (Murashige and Skoog, 1962),containing 1/10 part of the standard amount of KNO3 and NH4NO3,8% w/v sucrose and 0.8% w/v agar with a final pH of 5.8. Tuberformation was scored weekly over a period of 6 weeks. In addition,around 50 stem sections of the control and different transgenicclones were incubated on tuber-inducing medium supplementedwith either 0.1 lM GA3, 1.0 lM GA3 or 1.0 lM ancymidol and scoredfor tuberization. As a non-tuberizing control, plantlets were incu-bated on tuber-inducing medium containing 1% w/v of sucrose. Theproportion of tuber formation and types of tuber (p) for each treat-ment was calculated relative to the number (n) of incubated stemsections with variance Var = p(1 – p)/n.

Acknowledgements

The authors acknowledge support from EU-SOL, EPS (GraduateSchool of Experimental Plant Sciences), Wageningen Universityand Plant Research International through its twinning projectprogramme. We thank Anja Liebrandt for performing the GAmeasurements and Niek Appeldoorn for critical reading of themanuscript.

References

Abdala, G., Castro, G., Miersch, O. and Pearce, D. (2002) Changes injasmonates and gibberellins during development of potato plants(Solanum tuberosum L.). Plant Growth Regul. 36, 121–126.

Amador, V., Monte, E., Garcia-Martinez, J.L. and Prat, S. (2001)Gibberellins signal nuclear import of PHOR1, a photoperiod-responsive protein with homology to Drosophila armadillo. Cell,106, 343–354.

Bachem, C.W., van der Hoeven, R.S., de Bruijn, S.M., Vreugdenhil,

D., Zabeau, M. and Visser, R.G. (1996) Visualization of differentialgene expression using a novel method of RNA fingerprintingbased on AFLP: analysis of gene expression during potato tuberdevelopment. Plant J. 9, 745–753.

Banerjee, A.K., Chatterjee, M., Yu, Y., Suh, S.G., Miller, W.A. and

Hannapel, D.J. (2006) Dynamics of a mobile RNA of potato in-volved in a long-distance signaling pathway. Plant Cell, 18, 3443–3457.

Booth, A. (1963) The role of growth substances in the developmentof stolons. In The Growth of the Potato (Ivins, J.D. and Milthorpe,F.L., eds). London: Butterworths, pp. 99–112.

Bradley, D., Carpenter, R., Sommer, H., Hartley, N. and Coen, E.

(1993) Complementary floral homeotic phenotypes result fromopposite orientations of a transposon at the plena locus ofAntirrhinum. Cell, 72, 85–95.

Carrera, E., Jackson, S.D. and Prat, S. (1999) Feedback control anddiurnal regulation of gibberellin 20-oxidase transcript levels inpotato. Plant Physiol. 119, 765–774.

Carrera, E., Bou, J., Garcia-Martinez, J.L. and Prat, S. (2000) Chan-ges in GA 20-oxidase gene expression strongly affect stemlength, tuber induction and tuber yield of potato plants. Plant J.22, 247–256.

Chen, H., Rosin, F.M., Prat, S. and Hannapel, D.J. (2003) Interactingtranscription factors from the three-amino acid loop extensionsuperclass regulate tuber formation. Plant Physiol. 132, 1391–1404.

Chen, H., Banerjee, A.K. and Hannapel, D.J. (2004) The tandemcomplex of BEL and KNOX partners is required for transcriptionalrepression of ga20ox1. Plant J. 38, 276–284.

Coen, E.S., Romero, J.M., Doyle, S., Elliott, R., Murphy, G. and

Carpenter, R. (1990) Floricaula: a homeotic gene required forflower development in antirrhinum majus. Cell, 63, 1311–1322.

Ewing, E.E. (1987) The role of hormones in potato (Solanumtuberosum L.) tuberization. In Plant Hormones and Their Role inPlant Growth and Development (Davies, P.J., ed). Boston: Marti-nus Nijhoff Publishers, pp. 515–538.

Ewing, E.E. and Struik, P.C. (1992) Tuber formation in potato:induction, initiation, and growth. Hortic. Rev. 14, 89–198.

Fujino, K., Koda, Y. and Kikuta, Y. (1995) Reorientation of corticalmicrotubules in the sub-apical region during tuberization insingle note stem segments of potato in culture. Plant Cell Physiol.36, 348–353.

Fujioka, S., Yamane, H., Spray, C.R., Phinney, B.O., Gaskin, P.,

MacMillian, J. and Takahashi, N. (1990) Gibberellin A3 is bio-synthesized from gibberellin A20 via gibberellin A5 in shoots ofZea Mays L. Plant Physiol. 94, 127–131.

Gregory, L.E. (1956) Some factors for tuberization in the potato. AnnBot. 41, 281–288.

Hedden, P. (2001) Gibberellin metabolism and its regulation. J. PlantGrowth Regul. 20, 317–318.

Hedden, P. and Phillips, A.L. (2000) Gibberellin metabolism: newinsights revealed by the genes. Trends Plant Sci. 5, 523–530.

Hedden, P. and Proebsting, W.M. (1999) Genetic analysis of gib-berellin biosynthesis. Plant Physiol. 119, 365–370.

Hendriks, T., Vreugdenhil, D. and Stiekema, W.J. (1991) Patatinand four serine proteinase inhibitor genes are differentiallyexpressed during potato tuber development. Plant Mol. Biol.17, 385–394.

Jackson, S.D. (1999) Multiple signaling pathways control tuberinduction in potato. Plant Physiol. 119, 1–8.

Jackson, S.D. and Prat, S. (1996) Control of tuberization in potato bygibberellins and phytochrome B. Physiol. Plant, 98, 407–412.

Jackson, D., Culianez-Macia, F., Prescott, A.G., Roberts, K. and

Martin, C. (1991) Expression patterns of myb genes from Antir-rhinum flowers. Plant Cell, 3, 115–125.

Jackson, S.D., James, P., Prat, S. and Thomas, B. (1998) Phyto-chrome B affects the levels of a graft-transmissible signalinvolved in tuberization. Plant Physiol. 117, 29–32.

Jackson, S.D., James, P.E., Carrera, E., Prat, S. and Thomas, B.

(2000) Regulation of transcript levels of a potato gibberellin 20-oxidase gene by light and phytochrome B. Plant Physiol. 124,423–430.

Jasinski, S., Piazza, P., Craft, J., Hay, A., Woolley, L., Rieu, I., Phillips,

A., Hedden, P. and Tsiantis, M. (2005) KNOX action in Arabidopsisis mediated by coordinate regulation of cytokinin and gibberellinactivities. Curr. Biol. 15, 1560–1565.

Karimi, M., Inze, D. and Depicker, A. (2002) GATEWAY vectors forAgrobacterium-mediated plant transformation. Trends Plant Sci.7, 193–195.

Kloosterman, B., Vorst, O., Hall, R.D., Visser, R.G.F. and Bachem,

C.W. (2005) Tuber on a chip: differential gene expression duringpotato tuber development. Plant Biotechnol. J. 3, 505–519.

Kumar, D. and Wareing, P.F. (1974) Studies on tuberization ofSolanum andigena. II. Growth hormones and tuberization. NewPhytol. 73, 833–840.

Lange, T. (1998) Molecular biology of gibberellin synthesis. Planta,204, 409–419.

Lange, T., Kappler, J., Fischer, A., Frisse, A., Padeffke, T., Schmidtke,

S. and Lange, M.J. (2005) Gibberellin biosynthesis in developingpumpkin seedlings. Plant Physiol. 139, 213–223.

372 Bjorn Kloosterman et al.

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373

Lester, D.R., Ross, J.J., Smith, J.J., Elliott, R.C. and Reid, J.B. (1999)Gibberellin 2-oxidation and the SLN gene of Pisum sativum. PlantJ. 19, 65–73.

Martin, D.N., Proebsting, W.M. and Hedden, P. (1999) The SLENDERgene of pea encodes a gibberellin 2-oxidase. Plant Physiol. 121,775–781.

Martinez-Garcia, J.F., Garcia-Martinez, J.L., Bou, J. and Prat, S.

(2001) The interaction of gibberellins and photoperiod in thecontrol of potato tuberization. J. Plant Growth Regul. 20, 377–386.

Murashige, T. and Skoog, F. (1962) A revised medium for rapidgrowth and bio-assays with tobacco tissue cultures. Physiol.Plant, 15, 473–497.

Rodriguez-Falcon, M., Bou, J. and Prat, S. (2006) Seasonal control oftuberization in potato: conserved elements with the floweringresponse. Annu. Rev. Plant Biol. 57, 151–180.

Rosin, F.M., Hart, J.K., Horner, H.T., Davies, P.J. and Hannapel, D.J.

(2003) Overexpression of a knotted-like homeobox gene of potatoalters vegetative development by decreasing gibberellin accu-mulation. Plant Physiol. 132, 106–117.

Ross, J.J., O’Neill, D.P., Smith, J.J., Kerckhoffs, L.H. and Elliott, R.C.

(2000) Evidence that auxin promotes gibberellin A1 biosynthesisin pea. Plant J. 21, 547–552.

Sanz, M.J., Mingo-Castel, A., Lammeren, A.A.M.V. and

Vreugdenhil, D. (1996) Changes in the microtubular cytoskele-ton precede in vitro tuber formation in potato. Protoplasma,191, 46–54.

Shibaoka, H. (1993) Regulation by gibberellins of the orientation ofcortical microtubules in plant cells. Aust. J. Plant Physiol. 20, 461–470.

Spray, C.R., Kobayashi, M., Suzuki, Y., Phinney, B.O., Gaskin, P. and

MacMillan, J. (1996) The dwarf-1 (dt) Mutant of Zea mays blocksthree steps in the gibberellin-biosynthetic pathway. Proc. NatlAcad. Sci. USA, 93, 10515–10518.

Thomas, S.G., Phillips, A.L. and Hedden, P. (1999) Molecular cloningand functional expression of gibberellin 2-oxidases, multifunc-tional enzymes involved in gibberellin deactivation. Proc. NatlAcad. Sci. USA, 96, 4698–4703.

Visser, R.G., Stolte, A. and Jacobsen, E. (1991) Expression of achimaeric granule-bound starch synthase-GUS gene in trans-genic potato plants. Plant Mol. Biol. 17, 691–699.

Vreugdenhil, D. and Struik, P.C. (1989) An integrated view ofthe hormonal regulation of tuber formation in potato (Solanumtubersosum). Physiol. Plant, 75, 525–531.

Wang, H., Caruso, L.V., Downie, A.B. and Perry, S.E. (2004) Theembryo MADS domain protein AGAMOUS-Like 15 directly reg-ulates expression of a gene encoding an enzyme involved ingibberellin metabolism. Plant Cell, 16, 1206–1219.

Xu, X., van Lammeren, A.A., Vermeer, E. and Vreugdenhil, D. (1998a)The role of gibberellin, abscisic acid, and sucrose in the regulationof potato tuber formation in vitro. Plant Physiol. 117, 575–584.

Xu, X., Vreugdenhil, D. and van Lammeren, A.A.M. (1998b) Celldivision and cell enlargement during potato tuber formation.J. Exp. Bot. 49, 573–582.

Role for a GA 2-oxidase in potato tuber development 373

ª 2007 The AuthorsJournal compilation ª 2007 Blackwell Publishing Ltd, The Plant Journal, (2007), 52, 362–373