Expression of Arabidopsis WEE1 in tobacco induces unexpected morphological and developmental changes

WEE1 regulates the cell cycle by inactivating cyclin dependent protein kinases (CDKs) via phosphorylation. In yeast and animal cells, CDC25 phosphatase dephosphorylates the CDK releasing cells into mitosis, but in plants, its role is less clear. Expression of fission yeast CDC25 (Spcdc25) in tobacco results in small cell size, premature flowering and increased shoot morphogenetic capacity in culture. When Arath;WEE1 is over-expressed in Arabidopsis, root apical meristem cell size increases, and morphogenetic capacity of cultured hypocotyls is reduced. However expression of Arath;WEE1 in tobacco plants resulted in precocious flowering and increased shoot morphogenesis of stem explants, and in BY2 cultures cell size was reduced. This phenotype is similar to expression of Spcdc25 and is consistent with a dominant negative effect on WEE1 action. Consistent with this putative mechanism, WEE1 protein levels fell and CDKB levels rose prematurely, coinciding with early mitosis. The phenotype is not due to sense-mediated silencing of WEE1, as overall levels of WEE1 transcript were not reduced in BY2 lines expressing Arath;WEE1. However the pattern of native WEE1 transcript accumulation through the cell cycle was altered by Arath;WEE1 expression, suggesting feedback inhibition of native WEE1 transcription.


Results
Arath;WEE1 expression in tobacco plants results in premature flowering, altered root system growth and spontaneous shoot formation in culture. Constitutive Arath;WEE1 expression in tobacco ( Fig. S1) caused significant changes in plant development and led to premature flowering (Fig. 1a). WT plants grown in a growth chamber took approximately 150 days to flower (production of first visible bud) from day of sowing, whereas the Arath;WEE1 -expressing transgenic plants (NT-Arath;Wee1#8 and #2) flowered significantly earlier, after about 100 days (Fig. 1b). Moreover, WT plants flowered when they had produced more than 20 leaves longer than 10 cm, while transgenic plants expressing Arath;WEE1 formed only around seven leaves of this size before they started to flower (Fig. 1c).
Expression of Arath;WEE1 in tobacco plants also affected root development. NT-Arath;Wee1#8 plants had a significantly shorter primary root and when they were considered together, they had significantly fewer lateral roots + root primordia (Fig. 2a,b). However, when considered separately, there was no difference in the number of lateral roots between WT and Arath;WEE1-expressing plants, while there were fewer primordia in the transgenic line (Fig. 2c). This indicates that Arath;WEE1-expressing plants form less primordia with better capacity for outgrowth into fully grown lateral roots.
Further effects of expressing Arath;WEE1 in tobacco plants were seen in culture. When growing on standard cultivation media without any growth regulators, 1 cm long stem cuttings from Arath;WEE1-expressing plants formed on average 15 new shoots compared to WT cuttings that formed only callus (Fig. 3). NT-Arath;Wee1#8 tobacco stem segments cultivated on shoot induction medium also showed significantly greater capacity to form new shoots, producing 30, on average, from each stem cutting, while WT cuttings formed on average only 14. Arath;WEE1 expression in tobacco BY2 cells resulted in a reduction in mitotic cell size and a shortening of G2. When Arath;WEE1 was expressed constitutively in three independent BY2 cell lines (c-WEE1 lines 2, 10 and 12; Fig. S2a), significant reductions in mitotic cell size (P < 0.05) were detected compared with the empty vector line (EV) or with WT ( Fig. 4a,b). A similar result was obtained when BY2 cells were transformed with Arath;WEE1 using an inducible vector and expression induced by DEX in two independent lines (i-WEE-1 and i-WEE-6; Fig. S2b-d). A consistent effect on cell size was not seen when DEX was added to BY2 cells transformed with the empty pTA002 vector (Fig. S3). The effect of Arath;WEE1 on cell size correlated with effects on cell cycle progression. When expression of Arath;WEE1 was induced in line i-WEE-1 that was www.nature.com/scientificreports www.nature.com/scientificreports/ synchronised using aphidicolin, the mitotic index curve rose sooner (1-2 h) and peaked earlier (4-5 h) compared with the minus DEX control in which expression of Arath;WEE1 was not induced (Fig. 5b). These curves are consistent with a shortened G2 when Arath;WEE1 was expressed, as shown by histone H4 profiles used to measure the duration of S-phase, which was 5 to 6 h −DEX (Fig. 5a) and 4 to 5 h in the +DEX treatment (Fig. 5c). The interval between peaks (indicated by arrows in Fig. 5b) spans a cell cycle time of 13 and 12 h in the −DEX and +DEX treatments, respectively. Hence following induction of Arath;WEE1 expression, the major effect on the cell cycle was an 8-fold shortening of G2 compensated by a 3-fold lengthening of G1 (Fig. 5d).
Total WEE1 protein levels increased on expression of Arath;WEE1, and the pattern of WEE1 protein levels and activity were altered in synchronised cultures. To establish the mechanism of the cell cycle changes, effects on the timing of changes in WEE1 protein during the cell cycle were investigated. In synchronised BY2 cell lines transformed with an inducible Arath;WEE1 construct, without addition of DEX, total WEE1 protein levels increased through S/G2, however, as the majority of the cells entered mitosis (after 7 h), the WEE1 protein level decreased (Figs 6a,b; S6). When expression of Arath;WEE1 was induced by addition of DEX, WEE1 protein levels increased during S/G2 and again fell rapidly as cells entered mitosis (after 5 h). Hence, changes in WEE1 protein through the cell cycle followed the altered timing of mitosis in the induced cells.
A WEE1 kinase inhibition assay was used to investigate whether the WEE1 protein levels correlated with changes in the timing of WEE1 kinase activity. WEE1 activity was measured as the inhibitory action of immunoprecipitated WEE1 protein on CDK activity, using histone H1 as substrate (Fig. 6c). Sampling times were selected to coincide with early S phase and G2/M in both ±DEX. Without induction of Arath;WEE1 by addition of DEX, WEE1 kinase activity was maximal in early S phase and decreased by 31% in late G2 reaching a minimum during mitosis, consistent with the observed decrease in WEE1 protein level. In induced cultures, the WEE1 kinase activity was again maximal in early S phase and decreased by 29% in G2 and by a further 28% when the mitotic index peaked. Thus, WEE1 kinase activity also followed WEE1 protein levels and the altered timing of the mitotic peak. expression of Arath;WEE1 resulted in a premature increase in Nicta;CDKB1 activity. A logical hypothesis is that premature cell division would require early increases in CDK activity, which would drive cells into early mitoses. This hypothesis was tested by measuring kinase activity of both Nicta;CDKA;1 (referred to here, as CDKA) and Nicta;CDKB;1 (referred to here as CDKB) in the inducible Arath;WEE1 line 1 with and without DEX induction. CDKA activity was relatively constant regardless of the addition of DEX (Figs 7a; S8). However DEX-induction of Arath;WEE1 resulted in a significant increase in CDKB activity, compared to uninduced cells 1 h following release of the cells from aphidicolin when both induced and uninduced cells were in early S phase. In addition induced cells showed a significant reduction in CDKB activity at 5-7 h following aphidicolin release. At this point the +DEX treated cells were at G2/M, while the uninduced cells were only at S/G2 (Figs 7b; S8). Thus the induction of Arath;WEE1 resulted in an earlier peak in CDKB activity consistent with the earlier mitotic peak.
Arath;WEE1 perturbed the pattern of Nicta;WEE1 expression in synchronously dividing cells. Premature entry into mitosis at a reduced cell size could be regulated at the transcriptional, level.
Expression of Nicta;WEE1 in exponential phase BY2 cell cultures carrying the Arath;WEE1 inducible construct following DEX induction, was compared with Nicta;WEE1 expression in exponential phase WT BY2 cell cultures (Fig. S2e). There was a small decrease in Nicta;WEE1 expression when Arath;WEE1 was induced compared to www.nature.com/scientificreports www.nature.com/scientificreports/ WT expression, but this is unlikely to be sufficient to explain the cellular, protein and kinase changes seen in the Arath;WEE1-expressing cell lines. Similarly there was no significant change in the total WEE1 transcripts (Nicta;WEE1 + Arath;WEE1) when Arath;WEE1 expression was induced by DEX in BY2 cells carrying the inducible construct compared to exponential phase WT BY2 cell cultures (Fig. S2f).
However comparing Nicta;WEE1 expression ±DEX in synchronised cell lines, clear differences in the timing of Nicta;WEE1 expression were evident. In WT BY2 cells and in the uninduced Arath;WEE1 line, expression www.nature.com/scientificreports www.nature.com/scientificreports/ of Nicta;WEE1 peaked in mid S-phase ( Fig. 8a,b). However, when Arath;WEE1 was induced, the pattern of Nicta;WEE1 expression was perturbed so that the peak of its expression was shifted into mitosis/early G1 (Fig. 5c). Following induction, Arath;WEE1 was expressed more constantly through the cell cycle than Nicta;WEE1, as Arath;WEE1 expression was regulated by the 35S promoter, but significant peaks in expression were still seen in S phase and late G1 (Fig. 5d).

over-expression of Nicta;WEE1 in BY2 cells did not lead to a small mitotic size phenotype.
To test whether the effect of Arath;WEE1 expression in BY2 cells was a general effect of excess WEE1 expression, or whether it was specific to Arath;WEE1, Nicta;WEE1 was over-expressed in BY2 cells using the same DEX inducible system in two independent lines (Fig. 9a). However when the BY2 cells were synchronised with aphidicolin, and the Nicta;WEE1 expression was induced with DEX the mitotic peak was not anticipated as was found when  www.nature.com/scientificreports www.nature.com/scientificreports/ Arath;WEE1 expression was induced, in fact there was a very slight delay in mitosis (Fig. 9b). Mitotic cell area was also unaffected by over-expression of Nicta;WEE1 in BY2 cells (Fig. 9c).

Discussion
The flowering phenotype seen in the Arath;WEE1 tobacco plants shows strong similarities to the phenotype seen when Spcdc25 was expressed in tobacco 14,20,21 . The reduction in time to flowering (a 1.5 fold reduction) and number of leaves produced before flowering (a 2.8 fold reduction) was almost identical. However in contrast to tobacco plants expressing Spcdc25, expression of Arath;WEE1 in tobacco plants did not result in additional flowering from lateral branches. Based on grafting experiments 20 it was hypothesised that the anticipation of flowering in the Spcdc25 expressing plants may be result from an earlier competence of the shoot apical meristem to respond to the floral stimulus 15 . A similar mechanism may be operating in the tobacco plants expressing Arath;WEE1. It is also possible that Arath;WEE1 tobacco plants have similar perturbations in cytokinin signalling and carbohydrate status that were noted in Spcdc25 expressing tobacco plants 15,18,19 , although this would require further verification.
The reduction in primary root length and lateral root production in tobacco plants expressing Arath;WEE1 contrasts with the effect of Spcdc25 in increasing lateral root production noted by 26 . However, it is consistent with later reports of a restriction in root growth elicited by Spcdc25 expression in tobacco and attributed to a replacement of cytokinin effects in the roots 15 . Shorter primary roots were also found when Arath;WEE1 was over-expressed in Arabidopsis 24 and is consistent with a negative effect of increased WEE1 on root meristematic cell division. . The pairs of dark and light arrows mark the cell cycle times for each line/treatment: BY2 cells blocked in late G1 and S-phase by aphidicolin and then released following drug removal show an initial rise in the curve when cells trapped at the end of S-phase during the aphidicolin block, are the first to traverse G2 and enter mitosis following removal of the block. Since the first peak is when the bulk of synchronised cells enter mitosis, this point in time minus S-phase is an alternative measure of G2. Either way, G2 is less than 1 hour in the +DEX treatment, and 4 h −DEX (representative data from replicate experiments). Above and below the cell cycle plots, are mean expression profiles of histone H4 as percentages of maximum expression (±SD) without (a) and with (c) DEX used to calculate S-phase (4.5 h +DEX, 5.5 h −DEX). SEM was <3% throughout; n = 3. The duration of M-phase was calculated from the average mitotic index for each treatment (M) using formulae developed by Nachtwey and Cameron (1968) which account for exponential growth: dM = C/ln2 × ln (M + 1). G1 is calculated by difference. All phase durations are in hours. (2019) 9:8695 | https://doi.org/10.1038/s41598-019-45015-3 www.nature.com/scientificreports www.nature.com/scientificreports/ The spontaneous formation of shoots in the absence of added cytokinins was also seen both in tobacco expressing Spcdc25 and Arath;WEE1. However it contrasts with the phenotype seen in Arabidopsis plants over-expressing Arath;WEE1 where cultured hypocotyls from the Arath;WEE1 over-expressors produced fewer shoots than WT 24 . In fact the phenotype of the tobacco plants expressing Arath;WEE1 in this respect is more similar to the Arath;WEE1 knockout mutant lines, which produced more shoots from cultured hypocotyls than WT 24 .
Thus at a plant and organ level there are strong similarities between the effects of expressing Spcdc25 and Arath;WEE1 in tobacco. This is surprising given the opposing functions of the enzymes encoded. The difference between the expression of Arath;WEE1 in tobacco and Arabidopsis confirms that Arath;WEE1 does indeed induce the expected phenotype when expressed in its native environment. However the effects of its expression www.nature.com/scientificreports www.nature.com/scientificreports/ in tobacco are more consistent with a dominant negative effect, somehow repressing the action of the native Nicta;WEE1.
At a cellular level expression of Arath;WEE1 also had a positive effect on cell division, very similar to that seen with the expression of Spcdc25 17 . This effect was independent of the insertion location or the construct since multiple tobacco BY2 lines of both constitutively expressed and inducible Arath;WEE1 created multiple times in the lab all had the same phenotype. In most of the transgenic lines, the reduction in mitotic cell area in Arath;WEE1 expressing tobacco BY2 cells was not quite as severe as that seen when Spcdc25 was expressed, and indeed the Arath;WEE1 expression did not induce the formation of double files of cells as was seen in the Spcdc25 expressing cell lines 17 . However in one line, c-WEE1 line 10 where mitotic cell area was as low as seen in Spcdc25 expressing lines, double files of cells were also visible. This indicates a threshold effect for the production of double cell files. As previously suggested 15 the double cell files are reminiscent of the initial divisions in the pericycle that lead to the production of lateral root primordium. It is possible that the increase in lateral roots seen in some Spcdc25 expressing tobacco plants may be related to the severity of the effect on meristematic cell size. The reduction in root mass and in lateral roots in Spcdc25 and Arath;WEE1 expressing plants may therefore be consistent with a less severe cellular phenotype when the transgene is expressed constitutively as was the case here and in Bell et al. 14 as opposed to an inducible vector 26 .
Effects on cell cycle progression again were strikingly similar between BY2 cells expressing Arath;WEE1 or Spcdc25 with both showing a dramatic reduction in the length of the G2 phase and a lengthening of G1 + M phase 17 . In Spcdc25-expressing cells the anticipated mitotic peak was matched by an earlier increase in CDKB activity. Consistent with previous reports 4,5 , CDKB activity was also high at G2/M in uninduced cells. However, it peaked much earlier, in S phase in the cultures expressing Arath;WEE1. The anticipation of the mitotic peak when Arath;WEE1 expression was induced, was also accompanied by a premature fall in WEE1 protein and kinase activity, consistent with the changes in WEE1 seen in WT cells 23 . Thus at a cellular level the induced Arath;WEE1 expressing cell cultures are consistent with an early induction of mitosis after a short G2 resulting in a smaller mitotic cell size.
Expression of Arath;WEE1 in tobacco BY2 cells resulted in the opposite phenotype to that found with Solly;WEE1 expression in BY2 cells 25 and indeed over-expression of Nicta;WEE1 in the tobacco BY2 cells essentially had no effect. The results here also contrast with the effects on cell size seen when Arath;WEE1 was over-expressed in Arabidopsis plants 24 where root meristematic cells were larger than in WT. www.nature.com/scientificreports www.nature.com/scientificreports/ One hypothesis to explain these unexpected results was that the expression of Arath;WEE1 in the BY2 cells was causing an overall reduction of WEE1 protein perhaps due to a reduction of the native Nicta;WEE1 transcript. However overall WEE1 protein was higher, and neither Nicta;WEE1 or overall WEE1 transcript (Arath;WEE1 + Nicta;WEE1) changed dramatically on induction of Arath;WEE1 expression in exponentially growing BY2 cells. This indicates that the phenotypic effect is not due to a cell cycle-independent activation of the RNAi degradation pathway, which can be activated even with sense expression of transgenes 27 . A sense silencing mechanism is also less plausible given that in all three vector systems (BIN-HYG-TX 28 , pTA7002 29 and pKanII-SPYCE 30 ) used to express Arath;WEE1 in BY2 cells the orientation of the constructs is such that read through of antisense transcript from the selectable marker construct is not possible. This was shown to be a key factor in sense-mediated silencing 27 .
However the apparent shift in the expression of the native Nicta;WEE1 may form the underlying mechanism for the activation of a premature mitosis with the resulting phenotypic effects seen at a cellular, organ and whole plant level. In both WT and uninduced BY2 cells, Nicta;WEE1 transcripts are most abundant during S phase. This is consistent with the slightly later accumulation of WEE1 protein during S + G2 phase. However, when www.nature.com/scientificreports www.nature.com/scientificreports/ Arath;WEE1 is expressed, the peak of Nicta;WEE1 transcripts in S phase seems to be replaced by a later expression peaking in M/G1. Arath;WEE1 expression in these induced cultures is expressed more evenly through the cell cycle with a slight peak in S phase. This pattern is broadly consistent with reports on the expression of the 35S promoter during the cell cycle which show either a peak in S phase 28 or constant expression throughout all phases 29 . One possible mechanism is that Arath;WEE1 transcript production and translation into protein during S + G2 results in a feedback to Nicta;WEE1 transcription, delaying the accumulation of native WEE1 transcripts. www.nature.com/scientificreports www.nature.com/scientificreports/ This could be mediated through the large number of transcription factors that are thought to regulate WEE1 expression that include AtTCP15 30 , SOG1 31 and many others. An alternative mechanism may act at the protein level. The accumulation of Arath;WEE1 protein in S/G2 may activate the proteasome machinery prematurely due to differences in its sequence (Figs S3 and S4) and/ or conformation to trigger an early mitosis.
In conclusion the key finding is that expression of Arath;WEE1 in tobacco causes an anomalous phenotype consistent with a dominant negative effect and a phenotype that strongly resembles expression of the positive regulator of G2/M progression, Spcdc25. This can be used as a useful tool to explore effects of down-regulating WEE1 action on plant development and cellular function. Furthermore, a full understanding of the underlying mechanism may throw light on the interaction of WEE1 with cellular machinery at a transcriptional and/or protein level.

Materials and Methods
WEE1 constructs. For expression of Arath;WEE1 in BY2 cells the Arath;WEE1 open reading frame was PCR amplified using primers P35SX (5′-AGGCCCCGGCTCGAGATGTTCGAGAAGAACGG-3′) and P36SS (5′GCACACTAGTCGACTCAACCTCGAATCCTAT-3′) and cloned into the BIN HYG TX vector 32 under an attenuated form of the 35S promoter (as described in 24,33 ) for constitutive expression, or into the inducible vector pTA7002 34 using Xho I/Spe I. Individual clones were sequenced and a clone for each construct in which the amino acid sequence was intact was chosen for further work. For expression in whole tobacco plants, Arath;WEE1 was cloned into pkanII-SPYCE(M) 35 as described in Lentz Grønlund et al. 35,36 . Nicta;WEE1 was cloned into the pTA7002 vector as described in Cook et al. 23 .

Transformation of tobacco BY2 cells and induction of transgene in inducible lines. Stable
transformation of tobacco (Nicotiana tabacum) BY2 cells was achieved using a modified version of the method described by 37 with the addition of 20 µM acetosyringon (Sigma-Aldrich) during co-cultivation of the Agrobacterium (LBA4404) with the BY2 cells. Transformants were selected on solidified BY2 medium (0.8% agar) supplemented with 250 µg/ml Timentin and 80 µg/ml hygromycin. Calli were cultured in 50 ml BY2 medium, 250 µg/ml Timentin and 80 µg/ml hygromycin until stationary phase (1-3 weeks). Cultures were subjected to at least four rounds of sub culturing before being used in synchrony experiments.
Induction of WEE1 expression in BY2 cells carrying the pTA002 construct was achieved by addition of DEX (Sigma, UK) to a final concentration of between 1 µM and 100 µM. Induction of Arath;WEE1 was achieved using 1 µM DEX, while for the Nicta;WEE1 lines 30 and 100 µM were tested. DEX was added immediately following release from the aphidicolin block for synchronised cells, and three days after subculture for assays on exponentially growing cultures.

Tobacco and arabidopsis plant transformation. Young leaves from Nicotiana tabacum var Samsun
plants grown in soil were surface sterilised in 5% hypochlorite solution containing 100 ul/l Triton X-100 for 5 min with gentle agitation. Leaves were rinsed three times in sterile distilled water and cut into 1 cm 2 squares using a razor blade. Leaf squares were co-cultivated for 20-30 min in 100 ml of Rhizobium radiobacter (Agrobacterium tumefaciens) LBA4404 cell suspension (containing the WEE1 construct) at OD600 of 0.5 in 1 × MS medium in 140 mm diameter Petri dishes. Leaf squares were then transferred to shooting medium (1 × MS, 3% sucrose, 0.8% agar, NAA 0.1 ug/l, BAP 1ug/l). Following 48 h at 22 °C in the light, Leaf squares were then transferred to shooting medium including 50 ug/ml hygromycin and 200 ug/ml carbenicillin and incubation was continued for 4-6 weeks with a weekly subculture until calli and shoots were visible. Shoots were then excised and further cultured in rooting medium (1 × MS, 3% sucrose, 0.8% agar) to induce rooting. Plantlets were transferred to soil and grown to maturity. Expression of the transgene was analysed by PCR using primers AtWEE1fw (AGCTTGTCAGCTTTGCCT) and AtWEE1rv (TCAACCTCGAATCCTATCA). Two lines expressing the transgene (lines #2 and #8) were selected for further experiments.
Analysis of tobacco plants. Wild type and transgenic tobacco plants were grown from seed in a growth chamber at 22/18 °C day/night thermoperiod with 16 hrs illumination (irradiance 435 W m −2 ), and a relative humidity 50-75% as described in 38 . The leaves were numbered from the base (1 oldest) and when the first flower bud emerged, the length of leaves without the petiole was measured and leaves above 10 cm in length were counted. The age of the plants is given as days of growth after sowing.
Tobacco roots analysis. Sterilized tobacco seeds were sown on a square Petri dish containing MS medium (Murashige and Skoog Basal Salt Mixture, plant cell culture tested, Sigma-Aldrich, St. Louis, USA) containing 3% sucrose, 2 cm apart. After 21 days of cultivation at 25 °C with 16 h illumination with PFD (photon flux density) approximately 100 μmol m −2 s −1 (daylight fluorescent tubes; Osram, Wintherthur, Switzerland) as described in 38 . The length of the main root was measured and lateral roots counted semi-automatically with Smart Root software. For visualisation of root primordia the clearing method was used. The roots were fixed in acetone overnight and then fixed in phosphate buffer and mounted in 65% aqueous glycerol. They were observed with an Olympus BX51 microscope equipped with anvApogee U4000 digital camera.
For all semi quantitative RT-PCR experiments, cycle number was reduced and optimised rigorously as described previously 24,40 (Fig. S7) so that product amount was proportional to input amount of total RNA. This was verified with a dilution series of cDNA in each PCR experiment. Relative expression was normalised using primers to 18S rRNA as described previously 41 . A minimum of three replicate PCRs were performed for each primer set and products quantified from ethidium bromide stained agarose gels using the GeneGenius (Syngene, Cambridge, UK).

Protein extraction, Western blotting and histone kinase assays. Proteins were extracted from
Arabidopsis or tobacco leaves essentially as described in 42 . The WEE1 antibody and Western blotting were described in 35 . The antibody was used at a dilution of 1:1000 followed by α-rabbit IgG at 1:2500 (Sigma Dorset, UK). ECL reagents (Amersham Biosciences, Amersham, UK) were used to visualise the proteins.
For histone kinase assays proteins were extracted from 5 ml of synchronised cultures and assayed essentially as described in Cockcroft et al. 42 . Immunoprecipitations were carried out using antisera raised to Nicta;CDKA;1 and Nicta;CDKB1 as described in Sorrell et al. 4 . H1 protein kinase assays were as previously described 33,42 using 2 µl of antiserum. Incorporation was assayed by quantitation of autoradiographs using the GeneGenius (Syngene, Cambridge, UK).