Extensive developmental and metabolic alterations in cybrids Nicotiana tabacum (+ Hyoscyamus niger) are caused by complex nucleo-cytoplasmic incompatibility

Authors

  • Mikhajlo K. Zubko,

    Corresponding author
    1. Institut für Pflanzengenetik und Kulturpflanzenforschung Gatersleben, IPK Corrensstrasse 3, 06466 Gatersleben, Germany,
    2. International Institute of Cell Biology, Zabolotnogo Str. 148, 252143 Kiev-143, Ukraine, and
    Search for more papers by this author
    • Present address: School of Biological Sciences, University of Manchester, 3.614 Stopford Building, Oxford Road, Manchester M13 9PT, UK

  • Elena I. Zubko,

    1. International Institute of Cell Biology, Zabolotnogo Str. 148, 252143 Kiev-143, Ukraine, and
    Search for more papers by this author
    • Present address: Leeds Institute for Plant Biotechnology and Agriculture (LIBA), The University of Leeds, Leeds LS2 9JT, UK.

  • Alexander V. Ruban,

    1. Department of Molecular Biology and Biotechnology, Robert Hill Institute for Photosynthesis, Western Bank, Firth Court, Sheffield S10 2TN, UK
    Search for more papers by this author
  • Klaus Adler,

    1. Institut für Pflanzengenetik und Kulturpflanzenforschung Gatersleben, IPK Corrensstrasse 3, 06466 Gatersleben, Germany,
    Search for more papers by this author
  • Hans-Peter Mock,

    1. Institut für Pflanzengenetik und Kulturpflanzenforschung Gatersleben, IPK Corrensstrasse 3, 06466 Gatersleben, Germany,
    Search for more papers by this author
  • Simon Misera,

    1. Institut für Pflanzengenetik und Kulturpflanzenforschung Gatersleben, IPK Corrensstrasse 3, 06466 Gatersleben, Germany,
    Search for more papers by this author
  • Yuri YU. Gleba,

    1. International Institute of Cell Biology, Zabolotnogo Str. 148, 252143 Kiev-143, Ukraine, and
    Search for more papers by this author
  • Bernhard Grimm

    1. Institut für Pflanzengenetik und Kulturpflanzenforschung Gatersleben, IPK Corrensstrasse 3, 06466 Gatersleben, Germany,
    Search for more papers by this author

For correspondence (fax +44 161 275 39 38; e-mail mikhajlo.zubko@man.ac.uk).

Summary

The genetic basis of multiple phenotypic alterations was studied in cell-engineered cybrids Nicotiana tabacum (+ Hyoscyamus niger) combining the nuclear genome of N. tabacum, plastome of H. niger and recombinant mitochondria. The plants possess a complex, maternally inheritable syndrome of nucleo-cytoplasmic incompatibility, severely affecting growth, metabolism and development. In vivo, the syndrome was manifested as: late germination of seeds; dramatic decrease of chlorophyll and carotenoids in cotyledons and leaves; altered morphology of cotyledons, leaves and flowers; and dwarfism. The leaf phenotype depended on light intensity. In ‘green flowers’ (an extreme phenotype), homeotic function B was downregulated. In vitro, the incompatibility syndrome was restricted to the pigment deficiency of cotyledons. Electron microscopy revealed perturbations in the differentiation of chloroplasts and palisade parenchyma cells in bleached leaves. The pigment deficiency accompanied by retarded growth is discussed as a result of plastome–genome incompatibility, whereas other features are likely to be due to nucleo-mitochondrial incompatibilities.

Introduction

The development of higher plants is genetically controlled by three interacting genomes: nuclear, plastidic and mitochondrial. Evolutionarily synchronized regulatory interactions between these cellular compartments match their biogenesis, development and functions (Kutzelnigg and Stubbe, 1974; Leon et al., 1998). A disruption of subcellular interactions by alloplasmic substitutions or by rearrangements in cytoplasmic genes results in nucleo-cytoplasmic incompatibility, or compartmental genetic incompatibility (Babiychuk et al., 1995; Hanson, 1991; Stubbe, 1989; Stubbe and Herrmann, 1982). Classic examples of nucleo-cytoplasmic incompatibility in interspecific sexual hybrids of Oenothera manifested different degrees of chlorophyll deficiency designated as ‘hybrid bleaching’ or ‘hybrid variegation’ (Kutzelnigg and Stubbe, 1974; Stubbe, 1989; Stubbe and Herrmann, 1982). Non-sexual cytoplasmic hybrids (cybrids) combining a nucleus of Atropa belladonna and a plastome of Nicotiana tabacum were generated by protoplast fusions, and were characterized by complete chlorophyll deficiency (Kushnir et al., 1991). Some nucleus–cytoplasm combinations could not be established in somatic hybrids because of a severe nucleo-cytoplasmic incompatibility (Kushnir et al., 1991; Thanh et al., 1988; Wolters et al., 1993). In one case the incompatibility between a tobacco nucleus and potato chloroplasts (Thanh et al., 1988; Wolters et al., 1993) was overcome by recombination of plastid DNA of these two species (Thanh and Medgyesy, 1989).

In the examples presented above, nucleo-cytoplasmic incompatibility was attributed to the transfer and inheritance of a plastome into a different nuclear background. A distinctive criterion of incompatibility between a nuclear genome and a plastome of incompatible species was the failure to form fully green chloroplasts (Stubbe, 1989). However, a taxonomically distant chloroplast transfer often results in functional green plants, as was shown for cybrids with a nuclear genome from tobacco and plastomes from Petunia (Dragoeva et al., 1999; Glimelius and Bonnett, 1986); Atropa (Kushnir et al., 1987); Salpiglossis (Thanh et al., 1988); Lycium, Nolana, Physochlaine (Babiychuk et al., 1995); and Scopolia (Babiychuk et al., 1995; Zubko et al., 1996). A more detailed analysis of some of the cybrids with green plastids revealed structural and quantitative alterations of LHCII polypeptides (Babiychuk et al., 1995; Kushnir et al., 1987) that cause slight changes in photosynthetic parameters (Peter et al., 1999). These observations were explained by a weak compartmental genetic incompatibility expressed at the molecular and physiological levels. Similarly, the AA/IV nucleo-plastome combination of Oenothera, previously known as compatible, actually has a suboptimal photosynthetic apparatus that reflects a subtle version of plastome–genome incompatibility (Glick and Sears, 1994).

It was suggested that alloplasmic forms of cytoplasmic male sterility also reflect nucleo-cytoplasmic incompatibility that is caused by the transfer and rearrangements of mitochondrial genomes (Hanson, 1991; Leaver and Gray, 1982). The dwarfism in hybrids of Epilobium has also been attributed to rearrangements of mitochondrial genomes (Schmitz and Michaelis, 1988).

Nucleo-cytoplasmic incompatibility is a suitable model for studies of the interactions between nucleus and cytoplasm, which are still poorly understood in higher plants (Babiychuk et al., 1995; Hanson, 1991; Kushnir et al., 1991; Leaver and Gray, 1982; Stubbe, 1989). However, the lethality of chlorophyll-deficient plants and biparental inheritance of cytoplasm limit genetic and physiological studies on compartmental incompatibility. On the other hand, recent multidisciplinary studies revealed co-ordinated signal transduction pathways involved in mutual interactions between nuclear and organellar genes (Brutnell and Langdale, 1998; Hedtke et al., 1999; Leon et al., 1998; Mackenzie and McIntosh, 1999; Poyton and McEwen, 1996). These studies highlighted the need to generate new model systems with distinct properties.

Here we report on a novel, complex nucleo-cytoplasmic incompatibility in the de novo-generated cybrids Nicotiana tabacum (+ Hyoscyamus niger) that manifested severe developmental and metabolic perturbations during vegetative and generative development. We present the initial systematic studies of developmental, genetic and physiological aspects of the incompatibility syndrome.

Results

Generation of cybrids by substitution of cytoplasm

Cybrids containing a N. tabacum nuclear genome and a H. niger plastome (Figure 1) were generated by a one-step cytoplasm transfer in protoplast-fusion experiments (Zubko et al., 1996), followed by subsequent backcrosses. The parental plants were represented by a recipient line of a plastome albino mutant of N. tabacum and donor wild-type plants of H. niger. Three independent tobacco-like green lines (Rhn1, Rhn2 and Rhn3) were recovered after protoplast fusions. In a separate experiment, the cytoplasm of the line Rhn1 was re-transferred into nuclear background of a plastome chlorophyll-deficient tobacco mutant DSR A15, a derivative of N. tabacum SR1 (Svab and Maliga, 1986). The resulting plants (repeated cybrids) were designated Drhn lines. Rhn and Drhn plants rooted in vitro were transferred to soil and placed in separate chambers in the greenhouse for further backcrosses with wild-type tobacco. All plants exhibited cytoplasmic male sterility (Zubko et al., 1996). Seeds resulting from backcrosses of cybrid lines were germinated either in soil or in vitro for further analysis.

Figure 1.

Vegetative phenotypic characteristics of wild-type tobacco and cybrids under different growth conditions.

(a) 2-week-old seedlings of tobacco Wisconsin 38 under HL. (b) 2-week-old seedlings of cybrid Rhn × W4 under HL. (c) 2-week-old seedlings of Wisconsin 38 under LL. (d) 2-week-old seedlings of Rhn × W4 under LL. (e) 3-week-old plantlets of Wisconsin 38 under HL. (f) 3-week-old plantlets of Rhn × W4 under HL. (g) 3-week-old plantlets of Wisconsin 38 under LL. (h) 3-week-old plantlets of Rhn × W4 under LL. (i) 6-week-old plants of Rhn × W4 under HL. (j) A 9-week-old plant of Wisconsin 38 under LL. (k) A 9-week-old plant of Rhn × W4 under LL. (l) Top leaves of a Wisconsin 38 plant grown for 9 weeks under LL and 6 days under HL. (m) Top leaves of a Rhn × W4 plant grown for 9 weeks under LL and 6 days under HL. (n) The Rhn × W4 plant grown for 9 weeks at LL and 2 weeks at HL. (o) The Rhn × W4 plant grown for 9 weeks at LL and 4 weeks at HL. (p) Wisconsin 38 plants (top) and Rhn × W4 plants (bottom) grown in vitro from seed for 6 weeks. (q) A Wisconsin 38 plant (left) and an Rhn × W4 plant (right) during flowering. W4 is an abbreviation for the fourth consequent backcross using tobacco Wisconsin 38.

Stable inheritance of H. niger plastome in cybrids and their progenies

The completeness of the H. niger plastome in all cybrids analysed during their vegetative propagation (>5 years) and after repeated backcrosses was confirmed by restriction analysis of cpDNA (data not shown) and by Southern blot hybridization analysis using an rbcL gene fragment as a specific probe for the chloroplast genome (Figure 2). The hybridization patterns of the parental species H. niger (lane 2) and the selected cybrids (lanes 3–7) were identical, and different from those of N. tabacum (lane 1). It is concluded that the chloroplast DNA of all initial Rhn cybrids, repeated Drhn cybrids, and their backcross-derived progenies are identical to the cpDNA of H. niger. No recombination with the N. tabacum cpDNA was found in HindIII digests of cpDNA from cybrids (data not shown).

Figure 2.

Stable inheritance of H. niger plastome in cybrids.

Total DNA was isolated from N. tabacum Wisconsin 38 (lane 1), H. niger (lane 2) and randomly selected cybrid lines: Rhn1 (lane 3), Rhn2 (lane 4), Rhn × W2S (lane 5), Drhn3 × W-2 (lane 6), Drhn9 × W-1 (lane 7). DNA was digested with PstI+SalI enzymes; fragments were separated on agarose gel, blotted and hybridized with an rbcL probe. Lanes contained 1–3 µg DNA. HindIII digest of lambda DNA was used as a marker.

Mitochondrial genomes are polymorphic within the cybrid lines

Some rearrangements of mitochondrial DNA of the original and repeated cybrids have been shown previously (Zubko et al., 1996). We extended the analyses to compare the parental and cybrid mitochondrial genomes in plants with diverse flower morphology. MtDNA within the primary Rhn cybrids and their backcross-derived progenies, as well as from sexual offspring of repeated cybrids Drhn3 and Drhn9, were probed with gene fragments of atpα and nad3/rps12. The results are shown in Figure 3, and summarized in Table 1.

Figure 3.

Recombination of mtDNA in cybrids.

(a) BamHI DNA digests (hybridization with atpα probe).

(b) BamHI DNA digests (hybridization with nad3/rps12 probe); lanes contained 3–4 µg of DNA. Numbering of plant lines corresponds to Table 1. HindIII digest of lambda DNA was used as a marker.

Table 1.  Southern blot hybridization patterns of mtDNA and flower types in cybrids
Lines analysedTotal
hybridization
pattern
Flower type (and
its representative
in Figure 4)
  1. Total DNAs of cybrids (1–8), H. niger (9) and two tobacco lines, N.t. R100a1 (10) and N.t. cv. Wisconsin 38 (11) were digested with BamHI and probed with mitochondrial probes: (1) atpα and (2) nad3/rps12. Total hybridization patterns were determined on the basis of their uniqueness recognized as a set of obvious major bands (below 9 kb) for both probes. Different patterns B1−2 (abbreviated from BamHI; probes 1 and 2) were designated by arabic numbers of the corresponding cybrid lines in which the patterns were seen first.

  2. Hn, H. niger; (Nt) N. tabacum;CMS, cytoplasmic male sterility; gf, corolla-less ‘green flower’; fc, funnel-like corolla; tc, tobacco-type corolla; hc, Hyoscyamus-type corolla.

1 (Drhn9 × W-1)B1−2-1gf, CMS (e)
2 (Drhn3 × W-1)B1−2-2fc, CMS (g)
3 (Drhn3 × W-2)B1−2-2fc, CMS (g)
4 (Rhn × W3-1)B1−2-4tc, CMS (c)
5 (Rhn × W3-2)B1−2-4tc, CMS (c)
6 (Rhn × W3-3)B1−2-4tc, CMS (c)
7 (Rhn-2)B1−2-7gf, CMS (i)
8 (Rhn-1)B1−2-8gf, CMS (d)
9 (H.n.)Hnhc, fertile (a)
10 (N.t. R100a1)Nttc, fertile (b)
11 (N.t. W-38)Nttc, fertile (b)

The hybridization patterns of mtDNA from N. tabacum and H. niger differ from each other. Only major hybridization bands smaller than 9 kb were taken into account for assessment of uniqueness patterns (Table 1). Most hybridizing bands were shared by cybrid and parental mtDNA. The cybrid mitochondrial genomes are apparently generated from different proportions of both parental mtDNAs. Lines 4, 5, 6 and 7 are obviously recombinant, and some major bands of H. niger are missing in their patterns. Line 1 is also likely to be a recombinant, as only its middle band is derived from the H. niger pattern (Figure 3a). Lines 2 and 3 appear to contain all detectable bands of both parents in atpα patterns (Figure 3a). Lines 1, 2, 3, 7 and 8 either do not contain upper tobacco bands, or contain them at lower stoichiometric proportions in nad3/rps12 patterns (Figure 3b). At least five different groups of mtDNA hybridization patterns could be suggested from these data (Table 1). The same total mtDNA patterns were not observed in plants with different flower types.

General phenotypic characteristics are severely altered in cybrid plants

Germination and early development

The developmental differences between cybrid plants and tobacco became apparent during germination. Air-dried seeds of cybrids had a reduced weight (on average 1.9 times less than tobacco seeds). The seeds were germinated in soil and in vitro, under high light (HL) and low light (LL) conditions (Table 2). Cybrid seeds manifested lower germination rates (Table 2). The lightest seeds did not germinate (data not shown). In general, cybrid seeds germinated 3–5 days later than tobacco seeds. The delay in germination was less pronounced for seeds sown in vitro under LL, and was not apparent under LL/soil conditions (Table 2). This can be explained by an increased photosensitivity in germinating cybrid seeds. The enhanced viability of cybrid seeds in vitro(Table 2) suggests that heterotrophic growth compensates for their metabolic perturbations in soil.

Table 2.  Dynamics of seed germination for tobacco Wisconsin 38 and cybrids Rhn × W4 under different light and growth conditions (%)
Germination
time
(days)
In vitro, HLSoil, HLIn vitro, LLSoil, LL
W38CybridW38CybridW38CybridW38Cybrid
cscscscscscscscs
  1. c, Cotyledonary stage of seedlings; s, seedlings with true leaves; nd, not determined after stationary stage of germination.

3372150
8 9468 97541008310065
13 94472 972335 1008310065
21 94 84 97 58 1008310065
24 94 84 97 58 100355010065
34 nd nd nd nd nd 85871365
35 nd nd nd nd nd 85673365
36 nd nd nd nd nd 85406065
39 nd nd nd nd nd nd13873431
44 nd nd nd nd nd nd 1002540
49 nd nd nd nd nd nd 100 65

Cotyledons of cybrid seedlings germinated under HL were yellow-green, much smaller (Figure 1b) and paler than green seedlings of tobacco (Figure 1a). These differences in the cotyledon phenotype were less pronounced under LL (Figure 1c,d), where the cybrid cotyledons remained yellow-green for only 2–3 days after germination. The ratio between dry and fresh weight (D/F,%) of soil-grown Rhn × W4 cotyledons was much higher for cybrids (25 in HL; 28 in LL) than for tobacco (4.6 in HL; 11 in LL). It suggests severe alterations in metabolic rates and water regimes of the cybrid plants. D/F values for the cotyledons grown in vitro (13.3 in HL and 25 in LL for cybrids; 6.5 in HL and 18.2 in LL for tobacco) were less different under LL than under HL. All cybrid cotyledons had more conspicuous lanceolate shape of the tips than tobacco cotyledons (Figure 1a,b).

Later development under high light

During further development under HL, cybrids grew much more slowly than tobacco plants (Figure 1e,f). Their first leaves appeared after 13–20 days, 5–12 days later then in tobacco seedlings (Table 2). The leaves were lanceolate with curly, upturned edges (Figure 1i). The extent of leaf elongation varied noticeably in different plants. Plants formed several short shoots that branched from the base and remained yellow-green for approximately 1 month. By 6–8 weeks after germination, leaves of Rhn plants gradually became greener, and before flowering had reached a green pigmentation comparable to that of control tobacco plants (Figure 1q).

Flowering

Different lines of cybrids displayed remarkable phenotypic variation of male sterile flowers (Figure 4). The flower phenotypes could be separated into two characteristic classes representing corolla-containing (Figure 4c,f,g) and corolla-less flowers or ‘green flowers’(Figure 4d,e,h,i). ‘Green flowers’ did not possess stamens. These two basic flower types were stable in cybrids backcrossed with wild-type tobacco (Figure 4b). They appeared to be less stable after parasexual transmission of cytoplasm from the cybrid Rhn1 (‘green flowers’) to tobacco nuclear background, as a range of phenotypic variations of flowers of both types have been generated (Figure 4e–i).

Figure 4.

Diversity of flower morphology in cybrids.

(a) H. niger;(b) tobacco Wisconsin 38; (c) Rhn3 × W3; (d) Rhn-1; (e) Drhn9 × W-1; (f) Drhn3; (g) Drhn3 × W-1; (h) Drhn9 × W-1H; (i) Drhn9 × W-2.

The accumulation of transcripts hybridizing to the GLO cDNA probe (derived from an Antirrhinum gene GLOBOSA specifying petal and stamen formation) was significantly reduced in cybrid ‘green flowers’(Figure 5a). Transcripts corresponding to the tobacco NAG1 gene (a homologue of Arabidopsis AGAMOUS) that specifies carpel formation were accumulated in amounts comparable to those in parental flowers containing corollas and stamens (Figure 5a). The transcript pattern of the nad3/rps12 mitochondrial gene cluster was altered in cybrid ‘green flowers’, whereas transcript profiles of two other randomly selected mitochondrial genes, atp9 and coxIII, were the same in tobacco flowers and ‘green flowers’(Figure 5b).

Figure 5.

Expression of nuclear homeotic genes and mitochondrial genes in flowers of tobacco (Nt), H. nigrum (Hn), and ‘green flowers’ of cybrid Drhn9 × W-2 (Dr9).

(a) Northern blot of RNA from mature flowers probed with fragments of the genes for floral functions B and C.

(b) Northern blot of RNA from young flowers (yf) and mature flowers (mf) probed with mitochondrial genes.

Phenotypic alterations are limited in plants vegetating in culture in vitro and under low light

Cybrid seeds germinated in vitro manifested chlorophyll deficiency only in cotyledonary leaves. During further growth in vitro, cybrid plants did not exhibit any differences from control plants in the rate of growth and colour of the true leaves (Figure 1p). Rooted in vitro and then transferred into soil, cybrids did not display those differences from tobacco characteristic for the vegetative stage of Rhn plants grown in soil from the seeds. Being backcrossed with tobacco, these plants produced seeds that developed either into tobacco-like plants when grown in vitro, or into chlorophyll-deficient dwarf plants when grown in soil under HL.

Cybrids germinated in soil under LL displayed bleached cotyledons that became greener during the development of true leaves (Figure 1d,h). Surprisingly, the cybrid true leaves under LL (Figure 1g,k) were identical in their pigmentation and morphology to wild-type tobacco grown under HL and LL (Figure 1e,j). A transfer of the LL-grown cybrids to HL caused rapid bleaching of their top leaves within several days (Figure 1m), while tobacco leaves remained green (Figure 1l). Longer exposure at HL (≈2–4 weeks) led to dramatic deviations in the morphology of bleached leaves, and then to their gradual, patchy re-greening (Figure 1n). During re-greening of bleached leaves, new leaves appeared completely green under HL (Figure 1n,o). These observations suggest a light-sensitive step during early leaf development that affects leaf morphology and plastid structure in the cybrids.

Developmentally regulated reduction of photosynthetic pigments in cybrids

Consistent with their phenotypes, cybrid cotyledons grown in vitro contained one-quarter the amount of chlorophyll and one-third the amount of carotenoids seen in control cotyledons (Figure 6). Consequently, the carotenoid/chlorophyll (Car/Chl) ratio and the Chl a/b ratio were also higher in cotyledons of cybrids than in tobacco cotyledons. True aseptic leaves of cybrids and control plants were more similar with regard to these parameters than the corresponding cotyledons; the decrease in chlorophyll and carotenoid concentration in cybrids was less pronounced, and the Chl a/b ratio was even higher in tobacco leaves than in cybrid leaves.

Figure 6.

Pigment composition of cotyledons (C) and leaves (L) of tobacco plants (closed bars) and cybrids Rhn × W3 (open bars).

(a) Chlorophyll concentration (µg g−1 FW tissue); (b) chlorophyll a to b ratio; (c) total carotenoid concentration (µg g−1 FW tissue); (d) carotenoid to chlorophyll ratio.

Chlorophyll and carotenoid contents of cybrid cotyledons grown in soil were reduced to a lesser extent than those of seedlings germinated in vitro. The Chl a/b ratio was slightly higher in cybrids than in tobacco plants. In true leaves of HL-grown cybrid plants in soil, the content of both chlorophylls and carotenoids was lower than those under LL. The Chl a/b ratio and the Car/Chl ratio were higher in cybrids than in control plants. Surprisingly, the pigment content of the cybrids grown under LL was noticeably higher than in control plants.

Carotenoid composition analysis showed an enormous increase of xanthophyll cycle pigments in cybrids due to accumulation of zeaxanthin and antheraxanthin (Figure 7). Relative amounts of neoxanthin and lutein, the main carotenoids of the light-harvesting complexes, were not affected in the cybrids. The relative content of carotenoids was significantly increased in leaves of cybrid plants grown in soil at HL and LL.

Figure 7.

Carotenoid composition of tobacco (closed bars) and cybrid (open bars) leaves. Neo, Vio, Ant, Lut, Zea and BC are neoXanthin, violaXanthin, antheraXanthin, lutein and zeaXanthin, respectively. XC is a percentage of all xanthophyll cycle carotenoids (violaXanthin + antheraXanthin + zeaXanthin). DES is a de-epoxydation state calculated as (Zea + 0.5Ant) / (Vio + Ant + Zea).

Chloroplast biogenesis and cell differentiation are impaired in bleached tissues

Cells of cotyledons from 3-week-old tobacco seedlings contained three to 10 chloroplasts per cell section. They were full of large starch grains, regular in shape, with well developed thylakoids and grana (Figure 8a). Cells of yellow cotyledons of cybrid Rhn × W2 contained one to five chloroplasts per cell section. They were smaller and less abundant in starch grains than those of tobacco. Some cybrid chloroplasts had normal morphology, but less-developed thylakoid membranes (Figure 8b). The other plastids were dramatically perturbed. They had developed neither lamella system nor grana, and only primary thylakoids were observed (Figure 8c). The biogenesis of these chloroplasts was apparently impaired in cybrids at a very early stage of cotyledon development. The reduced accumulation of starch indicates early metabolic differences between cybrids and control tobacco plants, probably due to low photosynthetic capacity.

Figure 8.

Ultrastructure of chloroplasts and topography of cells in tobacco and cybrids.

(a) Chloroplast with developed grana (g) from a cell of a green tobacco cotyledon. (b) Chloroplasts from a cell of a yellow cotyledon from the cybrid. (c) Highly perturbed chloroplast with malformed thylakoids (mth) from a yellow cotyledon of the cybrid. (d) Area of tobacco leaf cell with differentiated chloroplasts. (e) Cell area of a yellow cybrid leaf (chloroplasts are elongated). (f) Fragment of a cell from re-greened cybrid leaf (chloroplasts are of normal shape). Starch grains (sg) are clearly visible in chloroplasts from green cotyledons and leaves; bar = 1 µm. (g–i) Transverse sections were made through (g) tobacco leaf; (h) green cybrid leaf; (i) yellow cybrid leaf. pp, Palisade parenchyma cells; upp, undifferentiated palisade parenchyma cells; bar = 30 µm.

The number of chloroplasts in leaf cells of plants grown in the soil were always higher (five to 25 per cell section) than in cotyledons. In the yellow-green leaves of cybrid Rhn × W2, plastids were slightly more elongated, and contained fewer starch grains and a reduced membrane system (Figure 8e) in comparison to green leaves of tobacco (Figure 8d) and greening leaves of the cybrid (Figure 8f). These perturbations were less pronounced than in plastids of yellow cybrid cotyledons (Figure 8c). Cybrid leaves that restored normal pigmentation contained well differentiated chloroplasts (Figure 8f). This suggests that the normalization of chloroplast biogenesis in greening cybrid leaves occurs at a later stage of development.

Leaf cells of tobacco were regular in their oval shape, and palisade and spongy parenchyma cells were well differentiated and equally distributed (Figure 8g). The transverse sections from yellow-green cybrid leaves showed dramatic perturbations in cell morphology (Figure 8i). They had an irregular shape and more intercellular spaces. A palisade parenchyma layer was not clearly differentiated, despite the elongation of corresponding cells. In contrast, the majority of cells from greened leaves of the cybrid (Figure 8h) were similar to cells of tobacco.

Discussion

Maternally inheritable phenotypic alterations in cybrids are caused by nucleo-cytoplasmic incompatibility

We describe a maternally inherited complex of phenotypic traits derived de novo in cybrids combining a nucleus from N. tabacum, chloroplasts of H. niger and recombinant mitochondria. There are two distinctive components in this complex: (1) a constant component that characterizes all cybrids and consists of a transient chlorophyll deficiency in early development under HL and retarded growth; and (2) a variable component where different lines show multiple alterations of vegetative and floral development. Cytoplasmic inheritance of the phenotypic complex has been proven by three to six backcrosses with wild-type tobacco. In addition, the repeated cybrids (with Rhn1 cytoplasm re-transferred to tobacco) manifested the expected altered phenotypes. Therefore it is highly unlikely that the observed altered phenotypes in cybrids are due to the presence of nuclear genetic material from H. niger.

All functional cybrids previously generated in Solanaceae were identical in general phenotype to recipient plants (Babiychuk et al., 1995; Dragoeva et al., 1999; Glimelius and Bonnett, 1986; Kushnir et al., 1987; Peter et al., 1999; Thanh et al., 1988). Among some cybrids (Bonnema et al., 1995; Bonnett et al., 1993) rare, variegated mutants similar to mitochondrial nonchromosomal stripe mutants (Newton and Coe, 1986) were isolated. The phenotypes described here cannot be the result of these or any other cytoplasmic mutations (Börner and Sears, 1986; Hanson, 1991; Kutzelnigg and Stubbe, 1974) because all independently produced lines had similar phenotypes. Backcrosses and fusion experiments with three nuclear backgrounds of tobacco revealed that the alterations in cybrid phenotype are not cultivar-dependent. Therefore we conclude that the described phenotypic alterations could be interpreted in terms of nucleo-cytoplasmic incompatibility between the nucleus of N. tabacum and the cytoplasm derived from H. niger. In contrast to albino cybrids A. belladonna (+ N. tabacum) (Kushnir et al., 1991) demonstrating a severe sexual incongruity in Solanaceae, the plants N. tabacum (+ H. niger) are viable in soil and thus are suitable for genetic and developmental studies.

Plastome and chondriome could contribute differentially to the whole incompatibility syndrome

Pigment deficiency and retarded growth are permanent characteristics of all cybrids that stably inherited cpDNA from H. niger. Thus we suggest that this constant phenotypic property reflects the incompatibility between the nucleus of N. tabacum and the plastome of H. niger. The cybrid growth retardation is likely to be a secondary effect caused by a low photosynthetic capacity and a delay in germination. In Oenothera hybrids, the incompatibility syndrome is associated with biparental plastid inheritance. Different proportions of the parental plastomes in hybrid tissues cause a range of variegation patterns (Kutzelnigg and Stubbe, 1974; Stubbe, 1989). More homogeneous pigmentation of chlorophyll-deficient leaves in our cybrids N. tabacum (+ H. niger) might be explained by the exclusive presence of H. niger plastids. Their uniparental inheritance simplifies genetic studies of the Nicotiana (+ Hyoscyamus) incompatibility syndrome.

The variations in cybrid flower morphology (Figure 4) generally reflected the diversity of mtDNA patterns (Figure 3; Table 1). Each group of plants with certain morphological types of flowers tended to display its own pattern of mtDNA. These observations allow speculation about possible correlations between variants in flower composition and specific alterations in their mtDNA. The different flower types are inherited, although some plants with new flower morphologies occasionally arise after backcrosses (data not shown). Variations in flower phenotypes and corresponding mtDNA patterns are known for alloplasmic cytoplasmically male-sterile lines (Dragoeva et al., 1999; Kofer et al., 1991). Therefore we could attribute the cybrid flower alterations to nucleo-mitochondrial incompatibilities resulting from rearrangements in mtDNA.

The recombination of plant mtDNA (Hanson, 1991; Leaver and Gray, 1982) is especially pronounced in cytoplasmic hybrids (Belliard et al., 1979; Gleba and Sytnik, 1984). The recombinant Rhn-mtDNA might have a biological advantage as a mechanism affecting nucleo-cytoplasmic composition in somatic hybrids (Wolters et al., 1993). An incompatibility between mitochondria of N. tabacum and plastome of H. niger within a tobacco nuclear background also could be hypothesized as a mechanism selecting recombinant mitochondria.

We related constant and variable components of cybrid phenotypes to nucleo-plastome and nucleo-chondriome interactions on the basis of co-segregational analysis (Gleba and Sytnik, 1984) of organelle DNA patterns and phenotypic characteristics. However, some of the variable features (dwarfism, elongation/curling of leaves, branching of stems) could also result from altered biochemical (Mackenzie and McIntosh, 1999) and molecular (Hedtke et al., 1999) interactions between plastids and mitochondria.

‘Green flowers’ imply an impact of cytoplasm on floral homeotic functions

Corolla-less and stamen-less flowers of cybrids N. tabacum (+ H. niger) resemble in their characteristics nuclear homeotic mutants defective in function B that specifies petal and stamen structures (Weigel and Meyerowitz, 1994). The present study revealed that homeotic function B is downregulated transcriptionally in the cybrid ‘green flowers’, which have also been found in cybrids N. tabacum (+ Scopolia carniolica) (Zubko et al., 1996). Therefore ‘green flowers’ could represent a novel model of ‘cytoplasmic homeosis’, in which normal homeotic genes are affected in their expression by rearrangements in mitochondrial genomes. The GLOBOSA-like gene is a first nuclear candidate for the particular inter-compartmental interaction. It is tempting to consider the nad3/rps12 cluster as a possible mitochondrial component for such interaction. Interestingly, mtDNA patterns of lines 1, 7 and 8 (all with ‘green flowers’) contain only one tobacco-derived major band hybridizing to nad3/rps12, instead of two bands in plants with pink-coloured corollas (Figure 3b). This fact could be related to the reduced expression pattern of nad3/rps12 cluster in ‘green flowers’(Figure 5b). The extent to which an intact structure and tobacco-like expression pattern of nad3/rps12 are associated with a proper performance of the homeotic function B need to be studied in more detail.

Alterations of the photosynthetic apparatus are modulated developmentally and conditionally in cybrid plants

Transient chlorophyll deficiency is observed in the cybrids from the beginning of their development in HL. A distinctive characteristic of the cybrids is a gradual restoration to normal chlorophyll levels in nearly all leaves at later stages. Similar incompatibility effects characterize Oenothera hybrids, but their pigmentation changes occur in the order green–bleached–green (Glick and Sears, 1994; Kutzelnigg and Stubbe, 1974). An alteration in hormonal status was suggested to contribute to the bleaching of Oenothera leaves (Glick and Sears, 1994). It could play a similar role in the cybrid incompatibility syndrome, in particular in stem branching normally related to hormonal regulation (Brutnell and Langdale, 1998). The bleaching and greening might also reflect a delayed assembly and functioning of the photosynthetic apparatus in cells at certain developmental stages. In vitro and under LL in soil, chlorophyll deficiency was observed only in cybrid cotyledons. This implies (1) a different regulatory mechanism of chlorophyll synthesis in cotyledonary leaves than in true leaves; and (2) a dependence of the incompatibility syndrome expression on growth and light conditions.

Pigment accumulation is strictly correlated with co-ordinated expression and assembly of the pigment–protein complexes (Bassi et al., 1993). The major chlorophyll carriers are nuclear-encoded antenna complexes of photosystems I and II. It is assumed that import and translocation of these proteins in the foreign chloroplast thylakoid membranes are perturbed in the cybrids, and consequently the assembly of the photosynthetic apparatus is delayed. A decreased amount of chlorophyll and carotenoids in the cybrids is consistent with the preferential loss of LHCIIb. The full recovery of this parameter at later stages of cybrid development in vitro could reflect acclimation. Significant increases of the carotenoid-to-chlorophyll ratio, the xanthophyll cycle pool size, and the extent of violaxanthin de-epoxidation are indications of photosensitization and protective reactions of cybrids to prevent photoinhibitory damage of the photosynthetic apparatus (Ruban and Horton, 1995).

Electron microscope analysis revealed that perturbations in thylakoid formation occur very early in cybrids, during development of cotyledons that are most sensitive in their response to incompatibility. This response was not dependent on the light intensity and nutritional conditions, in contrast to the response of true leaves of cybrids growing in vitro and in soil.

Aberrations in palisade cell differentiation in bleached cybrid leaves are similar to those in mutants defective in plastid development: Arabidopsis pale cress (Reiter et al., 1994); sharpdragon dag (Chatterjee et al., 1996) and tomato dcl (Keddie et al., 1996). Enlarged intercellular spaces found in the cybrids were also observed in the pale cress mutant. Genes corresponding to the mutations were also suggested to control palisade cell differentiation, apart from chloroplast biogenesis (Chatterjee et al., 1996; Keddie et al., 1996). The homologous nuclear genes are unlikely to be mutated in the cybrids. Therefore the involvement of those genes in chloroplast development and palisade differentiation is not direct, and several pathways can mediate these processes. Moreover, it suggests that nucleo-cytoplasmic interactions are involved in palisade cell differentiation as the process could be interrupted by nucleo-cytoplasmic incompatibilities, resulting in phenotypes similar to the mutational. Thus a putative plastid signal (Brutnell and Langdale, 1998; Hess et al., 1993/94) might be involved in palisade cell differentiation. Consequently, the altered leaf morphology in cybrids under HL could be caused by impaired cell differentiation.

Conclusion

The cybrids Nicotiana (+ Hyoscyamus) revealed that cytoplasm transfer could generate global developmental and metabolic alterations. The new traits are maternally inherited and are caused by nucleo-cytoplasmic incompatibility. In contrast to uniform characters born from a nucleo-plastome incompatibility, variable alterations caused by nucleo-mitochondrial incompatibilities provide a range of phenotypes due to different mtDNA rearrangements. These findings support a novel concept of intercompartmental genetics that operates with a new source of cytoplasmically inherited variation, based on the altered status of interactions between nuclear and cytoplasmic genes in a certain distant nucleo-cytoplasmic combination. This concept implies the possibility of associating the discrete variable traits with particular groups of interacting nuclear and cytoplasmic genes.

Experimental procedures

Plant material and growth conditions

Cybrids Nicotiana tabacum (+ Hyoscyamus niger) containing nuclear genetic material of N. tabacum and cytoplasm of H. niger were produced by protoplast fusions (Zubko et al., 1996). Aseptic plants were propagated in vitro on basal Murashige and Skoog medium with 30 g l−1 sucrose at 25°C, illumination regime 10/14 h darkness/light, and light intensity 80–100 µmol m−2 sec−1 (lamps TSP FL40 SS W/37, Sanyo). For observations of growth and development, plants were grown from seeds in aseptic culture and soil under high (180–200 µmol × m−2 sec−1) and low (10–20 µmol × m−2 sec−1) light intensities (lamp Powerstar HQI-TS 250 W NDL, Osram Galogen) at 25°C and 60% humidity.

Backcrosses

Three cybrid lines, Rhn1, Rhn2 and Rhn3, which resulted from independent protoplast fusion events, were transferred to soil in the glasshouse. Flowering plants were pollinated by pollen from wild-type N. tabacum, mainly cv. Wisconsin 38 kindly provided by Hermine Gelin-Kowallis (Berlin, Germany). Seeds from these backcrosses were germinated either in vitro or in the soil to obtain progenies of cybrids for molecular, morphological and genetic analyses. In some experiments, backcrosses with cultivars Turkish Samsun, SR1 and Lechija were used for comparative morphological analyses.

Analyses of cpDNA and mtDNA

Total DNAs from aseptic leaves were isolated according to Dellaporta et al. (1983) and digested with restriction enzymes BamHI, and PstI +SalI. DNA digests were fractionated by electrophoresis (3–5 µg per lane) in 0.8% agarose gels, blotted on Hybond+ membrane, crosslinked by UV and hybridized with 32P-labelled probes at 65°C according to Koes et al. (1987). For analysis of cpDNA, a filter with PstI + SalI digests was probed with a 1.75 kb EcoRI fragment of rbsL plastid structural gene from spinach (Zurawski et al., 1981) kindly provided by Dr I.K. Komarnitsky (Kiev, Ukraine). For analysis of mtDNA, filters with BamHI digests were probed with a 1.0 kb BamHI fragment of mitochondrial atpα gene and a 2.2 kb EcoRI fragment of mitochondrial nad3/rps12 gene cluster from Arabidopsis (kindly provided by Dr W. Schuster, Berlin, Germany).

Northern blot hybridization

Total RNA from flowers was fractionated (5 µg per lane) on 1% (W/V) agarose gels with 6% (V/V) of formaldehyde, blotted on Hybond+ membrane, crosslinked by UV and probed at 65°C (Koes et al. (1987). The following probes were used: 0.848 kb EcoRI fragment of pcGLO containing GLOBOSA cDNA (kindly provided by Dr Z. Schwarz-Sommer, Köln, Germany); 0.75 kb HindIII/HincII fragment of pSK93 containing NAG1 cDNA (a gift from Dr M. Yanofsky, La Jolla, CA, USA); 1.85 kb BamHI fragment of atp9 gene, 1.1 kb EcoRI/PstI fragment of coxIII gene, and a 2.2 kb fragment of the nad3/rps12 gene cluster (kindly provided by Dr W. Schuster, Berlin, Germany).

Pigment measurements

Spectrophotometric pigment measurements (for chlorophyll and carotenoids) were performed according to Lichtenthaler (1987) in 100% acetone extracts. Pigments were extracted from cotyledons of 2-week-old seedlings and from the first leaves of 4-week-old plants, growing both in aseptic culture and in soil. In vitro plants grew at 25°C, 10/14 h of darkness/light, with a light intensity of 80–100 µmol m−2 sec−1 (lamps TSP FL40 SS W/37, Sanyo). In soil, plants grew under high (150–180 µmol m−2 sec−1) and low (15–20 µmol m−2 sec−1) light intensities (lamp TLD 50 W/83HF, Philips) at 25°C and 65% humidity.

HPLC analysis of carotenoids

Pigments were extracted from the first leaves of 4-week-old plants grown in soil at high light and analysed by HPLC as described by Thayer and Bjorkman (1990) and modified by Hartel et al. (1996).

Electron microscopy

Leaves and cotyledons were cut into small (≈1 × 1 mm) pieces, fixed for 2 h at 4°C in 3% glutaraldehyde + 0.6% KMnO4, and washed four times for 15 min in 25% (v/v) acetone. The material was transferred in a Lynx™ tissue processor (Leica, Bensheim, Germany) and subjected to a stepped acetone series, followed by a transfer to ethanol in four steps. From ethanol, the pieces were transferred over a stepped sequence (LR-W: 25, 50, 75 and 100% LR-W) into LR-White resin, and were polymerized in gelatine capsules in a drying oven at 60°C for 48 h. Thin sections were cut on a Reichert Ultracut-S microtome (Leica) and evaluated in a Zeiss CEM 902A transmission electron microscope (Zeiss-Leo, Oberkochen, Germany).

Acknowledgements

We thank Drs P.M. Debnam and M.E. Cannell for critical reading of the manuscript, and Mrs Heike Ernst for photography. This work was supported by the Koerber-Stiftung (Hamburg) and a visiting scientist grant (M.K.Z.) from IPK (Gatersleben).

Ancillary