Transfer of Plastid DNA to the Nucleus Is Elevated during Male Gametogenesis in Tobacco

The Frequency of Plastid-to-Nucleus DNA Transposition Is Lower in the Female Germline Than in the Male Germline

In a previous study (Huang et al., 2003a), the frequency of ptDNA transfer to the tobacco nucleus was determined by analysis of progeny derived from fertilization of wild-type plants with pollen from a transplastomic line (tp7). Contained within the tp7 plastid genome was a 35SneoSTLS2 transgene tailored specifically for nuclear expression (Fig. 1, A and C ). This transgene consisted of a region encoding neomycin phosphotransferase under the regulatory control of the cauliflower mosaic virus 35S promoter and terminator (Benfey and Chua, 1990), which allowed selection of nuclear integration events by screening for kanamycin resistance during early seedling growth. A nuclear intron (STLS2), designed to prevent translation of a functional protein even if unexpected transcription occurred in the plastid, was inserted into the neo reading frame (Fig. 1A). When tp7 was used as a male parent in a cross to wild-type tobacco, one in approximately 16,000 progeny were resistant to kanamycin, indicative of multiple independent transfers of this plastid transgene to the nucleus and subsequent nuclear expression (Huang et al., 2003a). The selected kanamycin-resistant plants each carried a copy, or sometimes several copies, of 35SneoSTLS2 integrated into chromosomal DNA that, with some exceptions (Huang et al., 2003a), behaved as single Mendelian loci in genetic studies.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Generation of tp7 and tpGUS transplastomic lines. A and B, Structure of the tp7 (A) and tpGUS (B) transplastomic cassettes. neoSTLS2 and gusSTLS2 are under the control of the 35S promoter (arrows) and 35S terminator sequences (gray boxes). The STLS2 intron is indicated by a black box. The aadA spectinomycin resistance gene is under the control of plastid regulatory sequences (Huang et al., 2003a; Zubko et al., 2004). C and D, Location of neoSTLS2 (C) and gusSTLS2 (D) in the tobacco plastid genomes of tp7 and tpGUS, respectively. In line tp7, neoSTLS2 is integrated into the inverted repeats (shown in gray) and is therefore present as two copies per plastome. In tpGUS, gusSTLS2 is integrated into the large single-copy region of the plastid genome.

To determine whether the frequency of maternal plastid-to-nucleus DNA transfer was similar to that in the male germline, crosses were performed using the tp7 transplastomic as the female parent (tp7 ♀ × NtBAR/GUS ♂), i.e. the reverse direction of that used by Huang et al. (2003a). The male parent used in this reciprocal cross was homozygous for a BAR/GUS nuclear transgene, which acted as a nuclear marker to confirm that any emerging kanamycin-resistant progeny were derived from the cross. A complication in screening progeny from the cross in which tp7 was used as the female parent was that 35SneoSTLS2 in the tp7 transplastome conferred partial kanamycin resistance (Huang et al., 2003a; Stegemann et al., 2003), and this was maternally inherited by all the progeny, thus precluding an identical screen to that used previously (Huang et al., 2003a). To overcome this background resistance, the concentration of kanamycin used in screening the reciprocal cross progeny was increased from 150 μg mL⁻¹ to 300 μg mL⁻¹ (Stegemann et al., 2003). To validate effective selection at this antibiotic concentration, emasculated tp7 flowers were fertilized with pollen from plants hemizygous for newly transposed 35SneoSTLS2 in their nuclear genomes (kr18 described previously [Huang et al., 2003a], and kr2.2, kr2.3, kr2.7, kr2.9, and kr2.10 described below; kr, kanamycin resistant). In each cross, all progeny contained the maternal tp7 transplastome, while one-half of the progeny were expected to also contain a nuclear 35SneoSTLS2 copy inherited from the hemizygous paternal kanamycin-resistant parent. For each of these crosses, 13 seeds were plated at defined positions among a high density of tp7 seeds on medium containing 300 μg mL⁻¹ kanamycin. After 3 months, progeny containing both plastid and nuclear copies of 35SneoSTLS2 from all genotypes except kr2.3 were clearly distinguished at this antibiotic concentration from seedlings exhibiting the background resistance conferred by the transplastome alone (Fig. 2A ). While it is possible that the kr2.3 nuclear genotype is not detectable above the background resistance of the transplastome, a more likely explanation is that none of the progeny contained a nuclear copy of 35SneoSTLS2, because this genotype is known to be highly unstable (A.E. Sheppard and J.N. Timmis, unpublished data). Nevertheless, it can be concluded that this screen is very close in sensitivity to that used previously for crosses where the transplastomic parent was male.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Detection and analysis of ptDNA transfer to the nucleus. A, Four kanamycin-resistant plants from the segregating progeny of tp7 ♀ × kr2.7 ♂ among 2,000 progeny of self-fertilized tp7 grown for 3 months on medium containing 300 μg mL⁻¹ kanamycin. B, One kanamycin-resistant plant (kr2.4) among 2,000 progeny of NtBAR/GUS ♀ × tp7 ♂ grown for 11 weeks on medium containing 150 μg mL⁻¹ kanamycin. C, Progeny of wild-type ♀ × kr3.1 ♂ grown for 6 weeks on medium containing 150 μg mL⁻¹ kanamycin. D, Progeny of wild-type ♀ × kr4.1 ♂ grown for 3 weeks on medium containing 150 μg mL⁻¹ kanamycin. E to I, Leaf (E–G), anther (H), and root (I) tissues from tpGUS plants that have been histochemically stained for GUS expression. The arrow in F indicates a stained trichome. J, A 3.5-cm leaf of NtBAR/GUS that has been stained for GUS expression. K, A 6-cm leaf of gs1.1 that has been stained for GUS expression. L, An 8-cm leaf of gs1.2 that has been stained for GUS expression. Scale bars = 20 mm (A and B), 5 mm (C and D), 2 mm (E), 0.3 mm (F and G), and 0.5 mm (H–L).

Having established the ability of the selection regime to recognize transposition events efficiently, approximately 273,000 seeds from tp7 ♀ × NtBAR/GUS ♂ crosses were screened under these conditions. After 3 months, no kanamycin-resistant seedlings similar to the positive controls were observed. However, in contrast to the progeny of the reverse cross (Huang et al., 2003a), some of the plants were not killed by the antibiotic, though they showed severe growth restriction (Fig. 2A). After a further 6 to 9 months, a number of seedlings were still alive, but these lacked the clear phenotype associated with a nuclear 35SneoSTLS2 in the transplastomic background (Fig. 2A). The most promising candidates for true kanamycin resistance (green seedlings that were clearly larger than other seedlings on the same plate) were tested by backcrossing to female wild type to remove plastid-localized 35SneoSTLS2 genes. Of these, only a single plant (kr3.1) produced a proportion of kanamycin-resistant progeny, and these appeared to be chimeric for kanamycin resistance, containing both resistant and susceptible sectors on their cotyledons (Fig. 2C), suggesting unstable expression of neo in the nucleus. Consequently, it was not possible to determine unequivocally the proportion of kanamycin-resistant progeny, because seedlings with only very small resistant sectors could not be clearly distinguished from fully sensitive seedlings. Positive GUS expression in progeny of kr3.1 confirmed that it had resulted from a cross with the homozygous NtBAR/GUS parent (data not shown). PCR analysis confirmed the presence of neo in kanamycin-resistant plants (data not shown), indicating that kr3.1 arose from transposition of the plastid-encoded neo to the nucleus. Nevertheless, the phenotype of the progeny of this plant was atypical compared with those reported earlier (Huang et al., 2003a), and the antibiotic selection time required for its identification was much greater than that necessary for the positive controls (Fig. 2A).

Given the remarkably low frequency of transplastomic 35SneoSTLS2 transposition in the female germline (one atypically resistant plant in 273,000), screening of the original cross undertaken by Huang et al. (2003a) was repeated using the parental lines described above to confirm elevated transfer rates in pollen. That is, pollen from the tp7 transplastomic line was used to fertilize NtBAR/GUS plants, and seedlings derived from this cross were screened for kanamycin resistance using 150 μg mL⁻¹ kanamycin. Ten kanamycin-resistant seedlings (kr2.1–kr2.10) were isolated from 126,000 seedlings (Fig. 2B), and all except kr2.8 survived to maturity after transplanting to soil. To ensure that the elevated levels of kanamycin (300 μg mL⁻¹) used for screening progeny from the reciprocal crosses did not affect the assay, a further 20,000 seedlings from progeny of the NtBAR/GUS ♀ × tp7 ♂ crosses were screened at the higher kanamycin concentration, resulting in the isolation of three additional kanamycin resistant plants.

Positive GUS expression in kr2.1 to kr2.10 (except kr2.8, which did not survive to maturity and was therefore not tested) confirmed that each had resulted from a cross with the homozygous NtBAR/GUS parent (data not shown). The independent origin of these kanamycin-resistant plants was demonstrated by DNA blotting of total cellular DNA restricted with XbaI (Fig. 3 ). DNA from each line showed a unique combination of restriction patterns when hybridized with aadA- and neo-specific probes. The NtBAR/GUS transgenotype DNA showed weak cross hybridization at high molecular size to the aadA probe. DNA of tp7 showed the expected hybridization at 10.9 kb to the neo probe and at 11.4 kb and 18.5 kb to the aadA probe (Huang et al., 2003a). An unexpected band was also seen at approximately 2.5 kb with the aadA probe, which is most likely due to a rearrangement of the transplastome arising from recombination between the native psbA gene and the plastid-derived psbA terminator sequence that regulates aadA.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

DNA-blot analysis of XbaI-digested DNA from nine kanamycin-resistant plants hybridized with neo (left) or aadA (right) probes. Each lane contained 10 μg of DNA except for tp7, which contained 50 ng of DNA. Size markers and approximate sizes for tp7 bands are shown in kb. The arrow indicates a faint band for kr2.5.

An overall frequency of one kanamycin resistance event in approximately 11,000 pollen grains was obtained from the two experiments that used seedlings derived from NtBAR/GUS ♀ × tp7 ♂ crosses. This is consistent with the previously reported transposition frequency of one event in 16,000 pollen grains (Huang et al., 2003a). In marked contrast, only a single atypical transposition event could be detected among 273,000 progeny of the reciprocal cross. These data demonstrate an approximately 25-fold greater frequency of transposition and stable integration of the plastid-encoded 35SneoSTLS2 gene into the plant nuclear genome in the male germline compared with the female germline. Even if we include only those kanamycin-resistant lines where transposition was confirmed at the molecular level, which gives a more conservative estimate of nine events out of 126,000 progeny from the cross where the transplastomic parent was male, the reciprocal difference described here is highly significant (P < 0.001; Fisher's Exact Test). As a final confirmation of the reliability of the selection regime used for screening tp7 ♀ × NtBAR/GUS ♂ crosses, 40,000 seedlings (sufficient to give >97% chance of recovering an event assuming a male germline transposition frequency of one in 11,000) of self-fertilized tp7 were screened at 300 μg mL⁻¹ kanamycin. Similarly to the tp7 ♀ × NtBAR/GUS ♂ cross described above, no resistant seedlings could be observed after 3 months, but some were still alive after a further 3.5 months. The most promising candidates for true kanamycin resistance were tested by backcrossing to female wild type, and a single plant (kr4.1) was identified that produced Mendelian ratios of kanamycin-resistant progeny (Fig. 2D).

Monitoring Plastid-to-Nucleus DNA Transposition Using a gus Reporter Gene

To investigate the timing and frequency of ptDNA transposition in somatic cells, a second transplastomic line (tpGUS) analogous to tp7 but containing a gus reporter gene in place of neoSTLS2 was generated (Fig. 1B). The tpGUS line is homoplastomic (data not shown) for a single 35SgusSTLS2 and aadA cassette inserted into the plastid genome near rbcL (which encodes the large subunit of ribulose bisphosphate carboxylase-oxygenase; Fig. 1D). Due to the presence of a nuclear promoter (35S) and intron (STLS2), the gus gene was expected to be expressed only upon transfer to the nucleus.

Histochemical staining of tpGUS plants identified sectors of GUS-positive tissue in leaves, cotyledons, roots, vasculature, anther walls, and trichomes. Staining appeared as small, discrete foci of expression in these tissues surrounded by areas where staining was not detectable (Fig. 2, E–I). To verify that these sectors were the result of transfer of 35SgusSTLS2 to the nucleus rather than activation of the gene in the plastid genome, the cellular localization of the GUS protein was examined. After transfer of 35SgusSTLS2 from the plastid to the nucleus, the resulting GUS enzyme, which does not contain any organelle targeting signals, would be expected to accumulate in the cytosol where the gus mRNA is translated. Nonlocalized and uniform staining of cells is diagnostic of GUS located in the cytosol, and this was observed in all the blue sectors examined in tpGUS plants. Figure 4 shows examples of sectors composed of one (Fig. 4, A, C and E), two (Fig. 4, B and D ), or more (Fig. 4, F and G) blue cells in tpGUS plants. Leakage of cytosolic contents results in patchy staining of adjacent cells and intercellular spaces (Fig. 4, C–G). These results are clearly distinguishable from expression in the chloroplasts of a control transplastomic line containing the gus gene driven by a plastid promoter (pUM79; Kode et al., 2006) in which GUS activity is confined to chloroplasts (Fig. 4H).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Microscopic analysis of GUS expression in tpGUS transplastomic plants. A and B, Single (A) and double (B) guard cells (labeled GC) expressing GUS. C, Transverse section showing a pallisade cell expressing GUS. Labels indicate epidermis (Ep), pallisade mesophyll (Pm), and spongy mesophyll (Sm). D, Two mesophyll cells expressing GUS. E, Epidermal cell expressing GUS. F and G, Multiple cells expressing GUS. H, Mesophyll cells in a previously isolated pUM79 transplastomic plant (see Kode et al., 2006) showing GUS localized to chloroplasts (cp). Scale bars = 20 or 100 microns.

In a single 18-cm leaf, 228 GUS-stained sectors were observed. Using the same estimation of cell number (Hannam, 1968) as Stegemann et al. (2003), this implies transfer and expression of 35SgusSTLS2 in at least one in 200,000 somatic cells, if each sector is assumed to result from a single transfer event. This frequency is approximately 25 times higher than that measured by Stegemann et al. (2003) where the observed transposition events required stable integration and subsequent cell division and plant regeneration.

Due to the significant discrepancy between transfer frequencies, we sought to provide a more detailed analysis of the DNA transfer frequency. Seeds were germinated in vitro and GUS sectors scored when cotyledons and the first true leaves had reached a length of 3 to 4 mm (Fig. 5 ). Large variations in sector numbers were found in different cotyledons and leaves, which is reflected in large sds. No sectors were found in the cotyledons and leaves of a transplastomic line lacking the gus gene (negative control). All four tpGUS transplastomic lines (tpGUS5.3, 5.6, 8.4, and 9.4), which were derived from independent transformation events, gave a similar range of sector numbers in cotyledons and leaves. This consistency between transplastomic lines supports plastid-to-nucleus transfer of 35SgusSTLS2 as an explanation for the origin of sectors. The possibility that the sectors arise from activation of a silenced gus gene, inadvertently introduced into the nucleus during transformation, is unlikely given the comparable sector frequencies in the four transplastomic lines. An average of between five and six sectors per cotyledon or leaf was found when the results from all four transplastomic tpGUS lines were combined (111 cotyledons and 89 leaves). The total number of cells per cotyledon or leaf was estimated to be 100,000 (see “Materials and Methods”), which corresponds to a DNA transfer frequency of approximately one event per 18,000 cells, if each sector is assumed to result from an independent transfer event. The correspondence in sector frequencies between cotyledons and leaves probably reflects the similar numbers of cells present in the organs when sectors were scored and a similar frequency of transfer.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Analysis of sector frequency in cotyledons (3–4 mm) and first true leaves (3–4 mm) in tpGUS lines and a control transplastomic line lacking the gus gene. sds are shown. n = 24 cotyledons for tpGUS lines and 16 cotyledons for the control line. n = 15 to 21 leaves for tpGUS lines and 12 leaves for the control line.

GUS-expressing sectors were found in all types of leaf cells, and their appearance appeared to be random. The variations in sector sizes probably reflect the timing of plastid-to-nucleus DNA transfer during leaf development. Among 225 sectors examined, we found 121 single-cell sectors, 60 two-cell sectors, 23 three-cell sectors, and 12 four-cell sectors. The remaining nine sectors contained five to 10 cells. Guard cells that are derived from a common guard mother cell (Pillitteri et al., 2007) provide a clear example of DNA transfer in progenitor and terminally differentiated cells. DNA transfer in a terminally differentiated cell gives rise to a single GUS-positive guard cell (Fig. 4A), whereas two GUS-positive sister guard cells probably result from DNA transfer in their shared mother cell (Fig. 4B).

The Effect of Gene Copy Number and Location in the Transplastome on Transposition Frequency

Potentially, both copy number and location of a reporter gene within the plastid genome could affect the frequency of its transposition to the nucleus. In tp7, the 35SneoSTLS2 transgene was inserted into the inverted repeat region of the plastid genome between the 16S rRNA and rps7/12 genes and is therefore present as two gene copies per plastid genome (Fig. 1C). In contrast, the 35SgusSTLS2 reporter gene is located adjacent to rbcL in the large single-copy region of the plastid genome (Fig. 1D). Therefore, these different transplastomic types may show different frequencies of plastid-to-nucleus DNA transfer because of the copy number difference of the reporter gene in the transplastome or different transposition frequencies existing among plastid sequences.

This hypothesis was tested by performing histochemical GUS staining to identify progeny with widespread GUS-positive staining from self-fertilized tpGUS plants. Two GUS-positive plants, gs1.1 and gs1.2, were identified in a nondestructive screen of 98,000 seedlings, giving a frequency of one recoverable transposition event in approximately 49,000 progeny. The histochemical staining phenotypes of these two lines were similar to nuclear 35Sgus-positive control plants (Fig. 2, J–L). Spliced gus mRNA, lacking the STLS2 intron, was amplified from gs1.1 and gs1.2, verifying that these two lines were the result of nuclear transposition of 35SgusSTLS2 (Fig. 6 ). In contrast, the plastid-localized 35SgusSTLS2 gene remained unspliced. Furthermore, after backcrossing gs1.1 to female wild type, GUS staining was present in approximately one-half of the progeny (data not shown), consistent with the expected segregation pattern for a nuclear gene. Therefore, prima facie, the frequency of plastid-to-nucleus DNA transfer observed in tpGUS appears to be approximately 4-fold lower than that observed for 35SneoSTLS2 in tp7. This relatively small difference in transposition frequencies could be due to sampling error arising from the rarity of the events or could be due to kanamycin selection being more efficient than vital GUS staining in identifying seedlings resulting from transfer events in pollen. However, from these experiments, it may be concluded that the insertion of transgenes in these two very different transplastomic locations with two quite different reporter genes does not appear to have a large effect on transposition frequency.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

RT-PCR analysis. RT-PCR was performed using gus primers spanning the STLS2 intron. The higher band (450 bp) represents unspliced transcript, and the lower band (250 bp) represents spliced transcript. In the NtBAR/GUS control, the gus gene does not contain an intron. Control RT-PCRs with ribosomal protein L25 primers are also shown, giving a single 370-bp band. Lanes marked − indicate no RT.

Plant Growth Conditions

Tobacco (Nicotiana tabacum) plants grown in soil were kept in a controlled environment chamber with a 14-h-light/10-h-dark and 25°C-day/18°C-night growth regime.

Analysis of Kanamycin Resistance

Kanamycin selection was performed using 0.5× Murashige and Skoog salt medium (Murashige and Skoog, 1962) containing 150 or 300 μg mL⁻¹ kanamycin for plants with wild-type plastids or 300 μg mL⁻¹ kanamycin for plants with tp7 plastids. Screens were performed by plating surface-sterilized seeds on 150-mm plates containing 80 mL of the above medium at a density of 2,000 seeds/plate. Progeny testing of kanamycin-resistant plants was performed by plating surface-sterilized seeds on 90-mm plates containing 20 mL of the above medium. All plates were incubated at 25°C with 16 h light/8 h dark.

Analysis of GUS Activity

For histochemical GUS assays, tissues were fixed by vacuum infiltration in 100 mm sodium phosphate buffer, pH 7.0, 0.12% formaldehyde, 0.1% β-mercaptoethanol, 0.1% Triton X-100 for 10 min, washed three times with 100 mm sodium phosphate buffer, pH 7.0, and stained in 45 mm sodium phosphate buffer, pH 7.0, 0.45 mm potassium ferricyanide, 0.45 mm potassium ferrocyanide, 0.1% Triton X-100, 0.05% chloramphenicol, 0.1% β-mercaptoethanol, 10% dimethyl sulfoxide, 0.1% X-Gluc overnight at 37°C. After staining, tissues were cleared in 70% ethanol.

Viable GUS staining was performed in tissue culture as described (Martin et al., 1992).

For the analysis of GUS sectors in seedlings, seeds were germinated on Murashige and Skoog medium as described (Kode et al., 2006). Whole seedlings were incubated in buffer containing X-Gluc at 37°C, then chlorophyll was removed with ethanol before mounting on slides to examine sectors using a Leica S8APO Stereo Zoom microscope. For sections, leaf pieces were fixed (0.3% formaldehyde, 30 min) and stained in X-Gluc buffer as described (Klösgen and Weil, 1991). Leaf pieces were embedded in wax (Thermoshandon Histocentre 3), sectioned into 5-μm slices (Microm HM330), cleared (Histo-Clear; National Diagnostics), and mounted on slides for microscopy (Leica DMR microscope). The average number of cells per leaf or cotyledon was estimated to be 100,000 cells based on counting the average number of cells in a 100-μm × 100-μm square (20 cells), the average cotyledon area (9.2 mm²), or leaf area (9.6 mm²) and the observation of six cell layers in the cotyledons and leaves examined (e.g. see Fig. 4C).

Southern Blotting

DNA blot analyses were carried out as described (Ayliffe and Timmis, 1992; Huang et al., 2003b).

Generation of tpGUS

A gus gene containing the second intron of the potato (Solanum tuberosum) STLS-1 gene inserted into the open reading frame, with 35S promoter and terminator sequences, was amplified from p35S GUS INT (Vancanneyt et al., 1990) using the following primers (NotI and SacII sites underlined): 5′-ATCGTAGCGGCCGCAACATGGTGGAGCACGACACTCTCGTCTAC-3′ and 5′-TGACTACCGCGGCATGCCTGCAGGTCACTGGATTTTGGTTTTAGG-3′.

The resulting PCR product was cloned into pGEM-T Easy (Promega). The SacII/NotI fragment of this vector containing gus and the ApaI/SacII fragment of pUM35 containing aadA flanked by the Brassica napus 16S rrn promoter and psbC terminator regions (Zubko et al., 2004) were then ligated into ApaI/NotI-digested pATB27-link (Zubko et al., 2004) to generate the transformation vector. Transplastomic plants were isolated by selecting for dark-green spectinomycin-resistant shoots following particle bombardment of pale-green ΔrbcL leaves as described (Kode et al., 2006).

Reverse Transcription-PCR Analysis

RNA extraction was performed using an RNeasy Plant Mini kit (Qiagen) and genomic DNA contamination removed using a TURBO DNA-free kit (Ambion). Reverse transcription (RT) was then performed using an Advantage RT-for-PCR kit (CLONTECH) with oligo(dT) primer. All kits were used in accordance with the manufacturers' instructions. PCR amplification was performed using Taq DNA Polymerase (New England Biolabs) according to standard protocols. Primers used were 5′-TCATTACGGCAAAGTGTGGGTC-3′ and 5′-GTAGAGCATTACGCTGCGATGG-3′ for gus PCRs and 5′-AAAATCTGACCCCAAGGCAC-3′ and 5′-GCTTTCTTCGTCCCATCAGG-3′ for L25 PCRs.