Translation Start Sequences Affect the Efficiency of ... - Plant Physiology

Translation Start Sequences Affect the Efficiency of Silencing of Agrobacterium tumefaciens T-DNA Oncogenes1 Hyewon Lee, Jodi L. Humann, Jennifer S. Pitrak, Josh T. Cuperus, T. Dawn Parks, Cheryl A. Whistler, Machteld C. Mok, and L. Walt Ream* Departments of Microbiology and Horticulture, Oregon State University, Corvallis, Oregon 97331

Agrobacterium tumefaciens oncogenes cause transformed plant cells to overproduce auxin and cytokinin. Two oncogenes encode enzymes that convert tryptophan to indole-3-acetic acid (auxin): iaaM (tryptophan mono-oxygenase) and iaaH (indole-3-acetamide hydrolase). A third oncogene (ipt) encodes AMP isopentenyl transferase, which produces cytokinin (isopentenyl-AMP). Inactivation of ipt and iaaM (or iaaH) abolishes tumorigenesis. Because adequate means do not exist to control crown gall, we created resistant plants by introducing transgenes designed to elicit posttranscriptional gene silencing (PTGS) of iaaM and ipt. Transgenes that elicit silencing trigger sequence-specific destruction of the inducing RNA and messenger RNAs with related sequences. Although PTGS has proven effective against a variety of target genes, we found that a much higher percentage of transgenic lines silenced iaaM than ipt, suggesting that transgene sequences influenced the effectiveness of PTGS. Sequences required for oncogene silencing included a translation start site. A transgene encoding a translatable sense-strand RNA from the 5⬘ end of iaaM silenced the iaaM oncogene, but deletion of the translation start site abolished the ability of the transgene to silence iaaM. Silencing A. tumefaciens T-DNA oncogenes is a new and effective method to produce plants resistant to crown gall disease.

Crown gall tumors result from overproduction of auxin and cytokinin in plant cells transformed by Agrobacterium tumefaciens (Winans, 1992). These abnormally high phytohormone levels result from expression of three genes transferred stably into the plant genome from the A. tumefaciens tumor-inducing (Ti) plasmid: iaaM (Trp mono-oxygenase), iaaH (indole-3-acetamide hydrolase), and ipt (AMP isopentenyl transferase; Garfinkel et al., 1981; Ream et al., 1983). IaaM converts Trp into indole-3-acetamide, which IaaH converts into indole-3-acetic acid (auxin; Inze´ et al., 1984; Schroeder et al., 1984; Thomashow et al., 1985, 1986; Van Onckelen et al., 1986). Loss of either enzyme prevents auxin production. Ipt converts AMP into isopentenyl-AMP, a cytokinin (Akiyoshi et al., 1984; Barry et al., 1984; Buchmann et al., 1985). Inactivation of ipt and either one of the two auxin biosynthesis genes abolishes crown gall formation (Ream et al., 1983). Adequate means do not exist to control crown gall disease on grapes, fruit and nut trees, cane berries, chrysanthemum, rose, and other nursery crops. Inoculation of plants with Agrobacterium radiobacter strain K84 affords some protection against specific strains of A. tumefaciens (Moore, 1988); however, crown gall 1

This work was supported by the U.S. Department of Agriculture (grant no. 2001– 02760) and by the Oregon State University Agricultural Research Foundation. * Corresponding author; e-mail [email protected]; fax 541– 737– 0496. Article, publication date, and citation information can be found at www.plantphysiol.org/cgi/doi/10.1104/pp.103.026534. 966

disease remains a significant problem. Arabidopsis plants resistant to crown gall disease have been produced by traditional genetic approaches (Nam et al., 1997, 1999), but this strategy is not applicable to plants in which crown gall is a problem. Our strategy to produce crown-gall-resistant plants was based on the phenomenon called posttranscriptional gene silencing (PTGS) or RNA interference (RNAi; Napoli et al., 1990; van der Krol et al., 1990; Lindbo and Dougherty, 1992a, 1992b; Baulcombe, 2000; Tijsterman et al., 2002). Transgenes that elicit PTGS trigger sequence-specific destruction of transgene-encoded mRNA and other mRNAs that have sufficient sequence identity (Jorgensen et al., 1998). PTGS has effectively silenced a number of genes in plants, including endogenous genes, transgenes, and viral RNAs (Angell and Baulcombe, 1997; Conner et al., 1997; Marano and Baulcombe, 1998; Ruiz et al., 1998; Escobar et al., 2001). Transgenes that elicit PTGS need not reside in all cells of the plant to induce systemic silencing of target genes. Leaves infiltrated with a PTGS-eliciting transgene can trigger silencing in untreated leaves (Voinnet and Baulcombe, 1997; Voinnet et al., 1998), and transgenic rootstock can silence homologous genes in grafted non-transgenic tissue (Palauqui et al., 1997). A number of transgene constructions can trigger PTGS. Early observations implicated high levels of sense-strand transcription across an untranslatable (or wild-type) transgene as an elicitor of PTGS; silencing often correlated with multiple-copy transgenes (Napoli et al., 1990; van der Krol et al., 1990;

Plant Physiology, November 2003, Vol. 133, pp. 966–977, www.plantphysiol.org © 2003 American Society of Plant Biologists

Silencing Agrobacterium tumefaciens T-DNA Oncogenes

Dougherty et al., 1994). Transgenes in an inverted repeat configuration also elicit PTGS (Cluster et al., 1996; Stam et al., 1997). Recent work in plants (Angell and Baulcombe, 1997; Waterhouse et al., 1998; Hamilton and Baulcombe, 1999; Chuang and Meyerowitz, 2000), nematodes (Fire et al., 1998; Montgomery et al., 1998; Montgomery and Fire, 1998; Timmons and Fire, 1998), neurospora (Cogoni and Macino, 1999a, 1999b), insects (Kennerdell and Carthew, 1998; Tuschl et al., 1999), and trypanosomes (Ngo et al., 1998) showed that double-stranded RNA (dsRNA) molecules induce systemic, sequence-specific PTGS (Bass, 2000). Plants undergoing PTGS accumulate small RNAs complementary to silenced mRNAs (Hamilton and Baulcombe, 1999; Zamore et al., 2000). These small RNAs appear to determine the sequence specificity of PTGS (Baulcombe, 2002; Llave et al., 2002). Our goal was to create crown-gall-resistant plants by introducing transgenes that elicit posttranscriptional silencing of two A. tumefaciens oncogenes, iaaM and ipt. Although silencing has proven effective against a variety of target genes, we found that iaaM was silenced in a much higher percentage of transgenic lines than ipt, suggesting that target gene sequences influenced the effectiveness of gene silencing. Roots respond only to the auxin-producing oncogenes (iaaM and iaaH). The presence or absence of the cytokinin biosynthesis oncogene (ipt) does not affect crown gall development on roots of any plant species we have tested, including Arabidopsis, apple (Malus domestica), carrot (Daucus carota), potato (Solanum tuberosum), and Jerusalem artichoke (Helianthus tuberosus; Ream et al., 1983; Viss et al., 2003). In agricultural settings, crown gall usually occurs on roots. Therefore, crown-gall-resistant rootstocks only need to silence iaaM. This study and recent reports by Escobar et al. (2001, 2002) and Viss et al. (2003) indicate that transgenic plants capable of silencing A. tumefaciens oncogenes are resistant to crown gall.

RESULTS Construction of Genes Encoding Sense-Strand RNAs with Premature Stop Codons

Two A. tumefaciens oncogenes were silenced in this study: iaaM and ipt. Constructions were based on the observation that mRNAs containing premature stop codons can destabilize homologous translatable RNA molecules (Dougherty et al., 1994; Smith et al., 1994; Dougherty and Parks, 1995). To apply this phenomenon to crown gall disease, we attempted to silence two T-DNA oncogenes: iaaM and ipt. For each oncogene, we changed the third codon to a stop codon and introduced a frameshift mutation downstream, thereby creating additional stop codons in the reading frame (Fig. 1). The full-length ipt coding sequence was used, but the iaaM transgene contained only the first 1,797 bp of the coding sequence, which is 2,268 bp long (see accession NC_002377). In addition, we constructed a hybrid transgene designed to elicit PTGS of both ipt and iaaM. It contained iaaM sequences inserted into ipt (Fig. 1). We verified the DNA sequence of each transgene and inserted the sequences into a plant transformation vector (pPEV6; Fig. 1; Lindbo and Dougherty, 1992a) oriented such that the sense strand of each transgene was expressed from the cauliflower mosaic virus (CaMV) 35S promoter.

Gall-Resistant Plant Phenotypes

We introduced these transgenes (separately) into “haploid” tobacco (Nicotiana tabacum cvs Burley and Kentucky) and then doubled the chromosomes of kanamycin-resistant transformants to produce plants homozygous for each transgene. These plants were the subjects of our initial tests, which we repeated on the progeny of self pollinations. Figure 1. Transformation vector and transgene constructs. A, The T-DNA portion of the plant transformation vector pPEV6. Boxes labeled LB and RB indicate left- and right-hand border sequences; the right border includes an overdrive sequence. The box labeled CaMV represents the CaMV 35S promoter, and the lines labeled 5⬘UTR and 3⬘-UTR indicate the 5⬘- and 3⬘-UTRs from CaMV. Boxes labeled NOS and Term denote the nopaline synthase promoter and 3⬘UTR/polyadenylation signal, which facilitate expression of the neomycin phosphotransferase gene, labeled nptII. Arrowheads show the BamHI restriction site used to insert transgenes downstream of the CaMV 35S promoter and the EcoRI site used to estimate T-DNA copy number. B, The iaaM-stop (solid line) and ipt-stop (shaded line) transgenes. Sequences indicate start codons (ATG) in bold italic letters and inframe or native stop codons in bold letters.

Plant Physiol. Vol. 133, 2003

967

Lee et al.

Putative iaaM-silencing plants (TDP1 lines) were challenged with an ipt mutant (strain 338; Garfinkel et al., 1981). Of 63 lines tested, 35 did not respond to the wild-type iaaM gene on the incoming A. tumefaciens T-DNA (Fig. 2A; TDP1-B7). When inoculated with wild-type A. tumefaciens (strain A348), these lines produced “shooty” galls (Fig. 2A) typical of cytokinin-rich tumors (Garfinkel et al., 1981; Ream et al., 1983). Wild-type plants inoculated with an iaaM mutant also produce shooty galls (Garfinkel et al., 1981; Ream et al., 1983). Thus, these transgenic lines exhibited the phenotype expected of plants capable of silencing iaaM. For comparison, Figure 2B shows an iaaM-stop line (TDP1-B31) that did not silence iaaM; wild-type A. tumefaciens induced large unorganized galls, and an ipt mutant induced small necrotic galls. Similar responses were observed on transgenic plants (PEV6 lines) that contained only vector sequences (data not shown). Experiments with grafted plants showed that our iaaM-stop construct did not elicit systemic silencing of the iaaM oncogene. We grafted: (a) vector-only scions (PEV6-B14) to three iaaM-silencing rootstocks (TDP1-B7) and (b) iaaM-silencing scions to four vector-only rootstocks. Wild-type (A348) and iptmutant A. tumefaciens were inoculated onto vectoronly and iaaM-stop regions of each stem. All seven plants gave identical results: Tissue containing only vector sequences did not silence iaaM, regardless of its position in the plant (rootstock or scion). Vectoronly stems produced large unorganized galls when inoculated with wild-type A. tumefaciens, and they were susceptible to an ipt mutant. In contrast, all stems that contained the iaaM-stop transgene silenced iaaM. These stems did not respond to infection with an ipt mutant, and they produced shooty galls when inoculated with wild-type A. tumefaciens. Thus, the iaaM-stop transgene was highly effective in cells containing the construct, but it did not produce a

signal that could move across a graft (downward or upward). Escobar et al. (2003) reported similar observations in grafted transgenic tomato (Lycopersicon esculentum) plants that silence A. tumefaciens oncogenes. Of 45 putative ipt-silencing plants (CW1 lines), one (CW1-K27) did not respond to the wild-type ipt gene on the incoming A. tumefaciens T-DNA. This line did not respond when challenged with an iaaM mutant (strain 328; Fig. 2C). It produced small necrotic galls (Fig. 2C) typical of auxin-rich tumors (Garfinkel et al., 1981; Ream et al., 1983) when inoculated with wild-type A. tumefaciens. An ipt-stop line that failed to silence ipt (CW1-K52) produced shooty galls when inoculated with an iaaM mutant (data not shown). We tested 29 lines (CW4) that contained the fused ipt::iaaM-stop transgene. One line (CW4-B30) reduced tumorigenesis significantly, showing little or no response to inoculation with wild-type A. tumefaciens (Fig. 2D). Although this line reduced gall size markedly and prevented gall formation at some inoculation sites, it did not abolish tumorigenesis. Nine other CW4 lines produced shooty galls in response to infection by wild-type A. tumefaciens, indicating that these tumor cells expressed ipt but not iaaM. Two additional CW4 lines did not respond to infection with an ipt mutant, indicating that these lines also silenced iaaM. Thus, 41% (12 of 29) of the ipt::iaaM-stop plants silenced iaaM, but only one of these lines (3%) also silenced ipt. As controls, we constructed 11 kanamycin-resistant plant lines (designated PEV6) that contained the transformation vector T-DNA without a transgene downstream of the CaMV promoter. These lines developed unorganized tumors upon infection with wild-type A. tumefaciens (data not shown). An iaaM mutant induced shooty tumors, whereas an ipt mutant caused small necrotic galls, as expected (data not shown). Thus, PEV6 lines responded to A. tumefaciens in the same way as the parental non-transgenic to-

Figure 2. Silencing of the iaaM and ipt oncogenes in transgenic tobacco. A, The responses of iaaM-silencing tobacco line TDP1-B7 to infection with Ti-plasmidless (A136), wild-type (A348), and ipt-mutant (338) A. tumefaciens. B, The responses of non-silencing line TDP1-B31 to infection with the same strains. C, The responses of ipt-silencing tobacco line CW1-K27 to infection with Ti-plasmidless, wild-type, and iaaM-mutant (328) A. tumefaciens. D, The responses of ipt::iaaM-silencing line CW4-B30 to infection with these A. tumefaciens strains.

968



bacco cultivars, indicating that vector sequences did not block T-DNA oncogene expression. Transgene Structure and Copy Number

We performed Southern-blot analyses on representative plant lines of each type. Figure 3A shows the T-DNA structures from lines containing either the iaaM-stop or the ipt::iaaM-stop transgene. Each DNA sample was digested with BamHI, which will produce an 1,800-bp fragment from intact copies of either transgene (Fig. 1). The blot was probed with radiolabeled iaaM sequences, and all lines (except for PEV6 lines, which contained only vector sequences) exhibited an 1,800-bp BamHI fragment indicating the presence of an intact transgene (Fig. 3A, lanes 1–8). All lines that silenced incoming wild-type A. tumefa-

ciens oncogenes (TDP1 lines B7, B17, B27, and CW4B30) also contained at least one rearranged transgene copy (Fig. 3A, lanes 4–7). The T-DNAs each contained a single EcoRI site located in the 3⬘-untranslated region (UTR) downstream of the transgene (Fig. 1). Each integrated copy of the T-DNA, when digested with EcoRI, produced a restriction fragment extending from the EcoRI site in the 3⬘-UTR to the nearest EcoRI site upstream in the plant genome, resulting in fragments containing unknown lengths of plant DNA. Thus, the number of EcoRI fragments indicated the minimum transgene copy number. All plant lines examined contained at least two transgene copies (Fig. 3B), which proved sufficient to silence an incoming iaaM oncogene upon infection of lines TDP1-B17 and TDP1-B27. Transgenes present in more than two copies silenced target

Figure 3. Southern-blot analysis of transgene structure and copy number in transgenic tobacco. A, Transgene structure iaaM-stop (TDP1) and ipt::iaaM-stop (CW4) lines. Genomic DNA was digested with BamHI (lanes 2–6 and 10 [5 ␮g lane⫺1]; lanes 1, 7, and 8 [10 ␮g lane⫺1]; lane 9 [4 ␮g]); blots were probed with labeled iaaM-stop DNA. Lanes 1 through 4 contained DNA isolated from CW4 lines B1 (lane 1), B11 (lane 2), B22 (lane 3), and B30 (lane 4). Lanes 5 through 8 contained DNA from TDP1 lines B7 (lane 5), B17 (lane 6), B27 (lane 7), and B31 (lane 8). Lanes 9 and 10 contained DNA from PEV6 (vector-only) lines B2 (lane 9) and B14 (lane 10). B, Transgene copy number in TDP1 and CW4 tobacco lines. Genomic DNA was digested with EcoRI (5 ␮g lane⫺1); blots were probed with labeled iaaM-stop DNA. Lanes 1, 2, 7, and 8 contained DNA from TDP1 (iaaM-stop) lines B7, B27, B17, and B31, respectively. Lanes 3 through 6 contained DNA from CW4 (ipt::iaaM-stop) lines B1, B11, B22, and B30, respectively. Lanes 9 and 10 contained DNA from PEV6 (vector-only) lines B2 and B14. Lane M contained phage ␭-DNA digested with HindIII; numbers beside this lane indicated fragment size in base pairs. C, Transgene copy number and structure in ipt-stop (CW1) tobacco lines. Genomic DNA (10 ␮g lane⫺1) from lines CW1-K27 (lanes 1 and 4) and CW1-K52 (lanes 2 and 5) was digested with either Hinf1 (lanes 1 and 2) or EcoRI (lanes 4 and 5). Lane 3 contained 100 ng of pUC119-ipt plasmid DNA (equivalent to 5 copies genome⫺1) digested with Hinf1. Numbers on the left side of the gel indicate the size (in base pairs) and location of ␭-HindIII restriction fragments. Plant Physiol. Vol. 133, 2003

969

Lee et al.

oncogenes in some instances (e.g. lines TDP1-B7 and CW4-B30; Fig. 3B, lanes 1 and 6) but not in others (e.g. lines CW4-B1 and TDP1-B31; Fig. 3B, lanes 3 and 8). Thus, transgene copy number did not correlate with the ability to silence incoming A. tumefaciens oncogenes. Intact copies of the ipt-stop transgene in CW1 plant lines released a 682-bp restriction fragment upon digestion with Hinf1; both lines examined (CW1-K27 and CW1-K52) contained one intact copy (Fig. 3C). Two rearranged copies of ipt-stop were evident in both Hinf1 and EcoRI digests of the ipt-silencing line (CW1-K27), but the non-silencing line (CW1-K52) did not contain rearranged copies (Fig. 3C).

Transgene mRNA Accumulation

PTGS drastically reduces cytoplasmic accumulation of mRNA encoded by the eliciting transgene and by other genes that contain sufficient sequence identity (Dougherty et al., 1994). However, PTGS does not diminish transcription of target genes; therefore, the levels of target RNA in nuclei are not affected by PTGS (Dougherty et al., 1994). We used northern-blot analysis, which measures mRNA accumulation (but not transcription rate), to measure transgene mRNA levels in selected plant lines. Transgene RNA levels in isolated nuclei were measured by quantitative real-time reverse transcriptase-PCR (RT-PCR). As anticipated, all iaaM-silencing lines examined (TDP1-B7, -B17, and -B27) lacked sense-strand transgene mRNA, which should migrate at 2.08 kb (Fig. 4A, lanes 1–3). Three non-silencing lines also lacked sense-strand iaaM-stop mRNA; data are shown for one line (TDP1-B31; Fig. 4A, lane 4). Because these blots were exposed to film for an extended period, we observed signals from ribosomal RNAs (rRNA) in all lanes, including those with RNA from vector-only plants (PEV6-B2 and PEV6-B14; Fig. 4A, lanes 9 and 10), which shares no significant homology with the probe. Large antisense iaaM transcripts accumulated to high levels in iaaM-silencing lines (TDP1-B7 and -B17; Fig. 5A, lanes 1 and 2). These antisense RNAs included molecules that formed a distinct band at approximately 3.8 kb and smaller species of heterogeneous length, which formed a smear (Fig. 5A, lanes 1–3). The length of these antisense transcripts suggests that they may have originated at the nopaline synthase promoter of the transformation vector (Fig. 1). Ribosomal RNAs (Fig. 5B) and ubiquitin mRNA (Fig. 4C) in these preparations formed discrete bands indicating that they were not degraded. A plant line that reduced iaaM expression (TDP1-B27) contained a much smaller amount of iaaM-stop antisense RNA (Fig. 5A, lane 3) than lines that abolished iaaM expression (TDP1-B7 and -B17; Fig. 5A, lanes 1 and 2). A line that did not silence iaaM (TDP1-B31) contained just a trace of antisense RNA (Fig. 5A, lane 4). Thus, 970

Figure 4. Northern-blot analysis of sense-strand RNA accumulation in iaaM-stop (TDP1) and ipt::iaaM-stop (CW4) tobacco lines. Each lane contains 10 ␮g of total RNA isolated from lines TDP1-B7, -B17, -B27, and -B31 in lanes 1–4, and lines CW4-B1, -B11, -B22, and -B30 in lanes 5–8. Lanes 9 and 10 contain RNA from the vector-only lines PEV6-B2 and B14. A, Northern blot probed with radiolabeled iaaM antisense RNA. B, The same filter stained with methylene blue to reveal ribosomal RNAs. C, The same northern blot probed with radiolabeled ubiquitin DNA.

high levels of iaaM-stop antisense RNA correlated with silencing of this oncogene. Nuclear transgene RNA levels were also highest in iaaM-silencing lines. The real-time RT-PCR procedure was designed to detect both sense and antisense RNAs. Nuclei isolated from iaaM-silencing lines contained more transgene RNA (antisense and/or sense) than non-silencing lines. For example, nuclei from iaaM-silencing line TDP1-B17 contained three to 19 times more transgene RNA than the non-silencing lines (TDP1-B13, -B31, and -B62; data not shown). These results correlated well with the northern blots. Taken together, they indicate that iaaM-silencing lines contained high levels of iaaM-stop antisense RNA, but lines that failed to silence iaaM showed little transcription through the transgene in either direction. All three non-silencing lines examined expressed only traces of iaaM-stop RNA. In contrast, any line that contained a high level of this RNA also silenced the iaaM oncogene. Thus, the iaaM-stop transgene was highly effective at silencing the iaaM oncogene, except in cases where chromosomal position prevented adequate transcription of the transgene. Plants that contained the ipt::iaaM-stop transgene accumulated varying quantities of transgeneencoded sense-strand RNA of the predicted size, 2.8 kb (Fig. 4A), as expected from our observation that Plant Physiol. Vol. 133, 2003


In the successful ipt-silencing line (CW1-K27), we did not detect transgene-encoded mRNA in preparations of total cellular or nuclear RNA (Fig. 6, A, lane 1, and C). A line (CW1-K52) that did not silence ipt accumulated a large amount of transgene mRNA of the predicted size: 1 kb (Fig. 6A, lane 2). Thus, high levels of ipt-stop RNA did not trigger silencing of the ipt oncogene. We did not detect long antisense RNAs homologous to ipt in either plant line (Fig. 6A). The absence or presence of transgene mRNA in these lines correlated with their ability or inability to silence the ipt oncogene. The absence of transgene RNA in nuclei from plant line CW1-K27 (Fig. 6C) suggests that transcriptional gene silencing inhibited ipt oncogene expression. Genes Encoding Sense and Antisense RNA

Figure 5. Northern-blot analysis of antisense RNA accumulation in iaaM-stop (TDP1) tobacco lines. Each lane contains 10 ␮g of total RNA isolated from iaaM-stop lines TDP1-B7, -B17, -B27, and -B31 in lanes 1 through 4, respectively. Lanes 5 and 6 contain total RNA from vector-only lines PEV6-B2 and B14. A, A northern blot probed with radiolabeled iaaM-stop DNA (both strands). B, Ribosomal RNA on the same filter stained with methylene blue.

most of these lines did not silence the target oncogenes. Two non-silencing lines (CW4-B1 and -B11) accumulated large quantities of transgene RNA that hybridized to an iaaM antisense RNA probe (Fig. 4A, lanes 5 and 6) whereas another (CW4-B22) contained a small amount (Fig. 4A, lane 7). Transgene RNA accumulated to an intermediate level in line CW4B30 (Fig. 4A, lane 8), which silenced iaaM and severely limited the effect of the ipt oncogene upon infection with A. tumefaciens (Fig. 2D). Hybridization of these RNAs to radiolabeled ipt DNA produced similar results: Line CW4-B30 contained more 2.8-kb transgene RNA than line CW4-B22 (data not shown). Transgene RNA levels in nuclei from these plant lines corresponded to those in RNA preparations from whole cells: Line CW4-B11 contained two to four times more transgene RNA than CW4-B30 or CW4-B22, which had the least transgene RNA (data not shown). Plant Physiol. Vol. 133, 2003

dsRNAs are potent inducers of gene silencing (Angell and Baulcombe, 1997; Fire et al., 1998; Montgomery et al., 1998; Montgomery and Fire, 1998; Timmons and Fire, 1998; Waterhouse et al., 1998; Chuang and Meyerowitz, 2000; Wesley et al., 2001). Transcription of sense and antisense RNAs from a target gene can trigger PTGS in plants (Waterhouse et al., 1998; Chuang and Meyerowitz, 2000). In an attempt to silence both ipt and iaaM, we designed a transgene to express both sense and antisense RNA from fused iaaM and ipt sequences. We joined iaaM-stop and ipt-stop in tandem and inserted them in a plant transformation vector between opposing CaMV and figwort mosaic virus (FMV) promoters (Fig. 7). We verified the sequence of this construction (pJP17) and tested its ability to silence the ipt and iaaM oncogenes. Gene Silencing during Mixed Infections

We developed two mixed-infection assays to evaluate the efficacy of oncogene silencing transgenes. In these assays, Kalanchoe daigremontiana stems or potato tuber discs were infected with a mixture of two bacterial strains: (a) wild-type A. tumefaciens (A348) and (b) disarmed A. tumefaciens (EHA101) containing the oncogene-silencing plasmid (pJP17). Disarmed strains lack oncogenes, but they retain virulence (vir) genes required to move transgenes from a binary plasmid vector (e.g. pJP17) into plants. Mixed infection assays obviated the need to create transgenic plants. This assay involves multiple transformation events, thereby masking chromosomal position effects that can influence the effectiveness of a particular integrated transgene. K. daigremontiana stems develop pronounced tumors in response to A. tumefaciens. These tumors have distinct morphologies corresponding to overproduction of cytokinin, auxin, or both (Garfinkel et al., 1981; Ream et al., 1983; Miranda et al., 1992). If auxin overproduction is blocked, tumors are smooth and lack 971

Lee et al.

Figure 6. Northern-blot and real-time RT-PCR analyses of sense-strand RNA accumulation in ipt-stop (CW1) tobacco lines. A, Northern blot probed with radiolabeled ipt DNA; B, total RNA stained with methylene blue. Each lane contained 10 ␮g of total RNA isolated from ipt-stop lines CW1-K27 (lane 1) and CW1-K52 (lane 2). C, The results of real-time RT-PCR performed on RNA isolated from nuclei of an ipt-silencing line (CW1-K27), a non-silencing line (CW1-K52), and a vector-only control (PEV6-B14). RNA levels were normalized to the amount of 5.8S rRNA detected in each sample; pg RNA indicates picograms of ipt RNA per picogram of 5.8S rRNA. No ipt RNA (⬍0.07 fg) was detected in nuclei from ipt-silencing line CW1-K27 or vector-only line PEV6-B14, which lacked ipt sequences.

adventitious roots due to high cytokinin levels, whereas wild-type tumors are rough and produce adventitious roots. Co-inoculation of K. daigremontiana with wild-type A. tumefaciens (A348) and EHA101 carrying pJP17 (Fig. 7) resulted in cytokinin-rich tumors (Fig. 8A), indicating that iaaM, but not ipt, was silenced. Co-inoculation with wild-type A. tumefaciens and EHA101 containing the empty vector (pJP21) did not disrupt normal tumorigenesis by A348 (Fig. 8B). To quantify the degree of iaaM silencing, we coinoculated potato discs with the same A. tumefaciens mixtures used in the assays on K. daigremontiana. Potato tubers respond to auxin but not to cytokinin (Ream et al., 1983). Individual tumor foci develop on potato discs when they are inoculated with wild-type (or ipt-mutant) A. tumefaciens, but potatoes do not respond when inoculated with an iaaM mutant (Ream et al., 1983). This system has been used to compare quantitatively the virulence of A. tumefaciens strains (Pueppke and Benny, 1981; Shurvinton and Ream, 1991; Shurvinton et al., 1992; Sundberg et al., 1996; Dombek and Ream, 1997). The iaaM-silencing plasmid (pJP17) significantly reduced tumorigenesis on potato (Fig. 8C) compared with the empty vector (pJP21; Fig. 8D). In multiple tests, 2- to 5-fold fewer tumor foci formed on discs co-inoculated with wildtype and iaaM-silencing strains, compared with discs co-inoculated with wild-type and vector-only strains. 972

The decreases in tumor formation caused by the silencing strain were statistically significant: P ⬍ 0.0001 in a t test (n ⫽ 432). In this assay, the iaaM-silencing transgene did not prevent tumorigenesis completely. Perhaps some potato cells received only an oncogenic T-DNA. Alternatively, the iaaM-silencing transgene may have inserted into transcriptionally silent chromatin in some potato cells. Nevertheless, this assay is a rapid and quantitative means to evaluate the efficacy of iaaM-silencing transgenes; constructs that reduced tumorigenesis significantly (⬎ 2-fold) in this assay prevented tumorigenesis on roots of transgenic plants containing these transgenes (Viss et al., 2003). Translatable mRNA Elicits Gene Silencing Better Than Untranslatable Sense RNA

Sense-strand RNAs that contain premature stop codons elicit gene silencing if they can initiate translation from a start codon downstream of the premature stop (D.C. Baulcombe, personal communication; Que et al., 1997; Tanzer et al., 1997). In such cases, low Mr transgene RNA from silenced plant tissue is bound to polyribosomes in the cytoplasm, where transgene RNA degradation appears to occur (Tanzer et al., 1997). We searched iaaM- and ipt-stop for start codons downstream of the nonsense mutations at codon 3. The ipt transgene lacked an alternative Plant Physiol. Vol. 133, 2003


Figure 7. Transgenes encoding sense and antisense RNA. RB and LB indicate right and left T-DNA borders, respectively. Arrows labeled peCaMV and peFMV indicate the location and orientation of enhanced CaMV 35S and FMV promoters. The arrowhead labeled HSP70L represents the 5⬘-untranslated leader sequence from the petunia (Petunia hybrida) heat shock protein 70 gene. Arrows labeled iaaM-stop and ipt-stop show the location, orientation, and length of oncogene-silencing transgenes containing iaaM and ipt sequences. Numbers in parentheses list the portions of the iaaM-stop transgene contained in pJP17 and pJC14. The arrow labeled NptII shows the position and orientation of the neomycin phosphotransferase gene, which is flanked by the nopaline synthase promoter and 3⬘-UTR and polyadenylation signal.

translation start, but iaaM contained an obvious inframe start at codon 18. This codon is surrounded by excellent sequence context for efficient translation: CCA ACC AAA AUG G (nucleotides that match the consensus sequence are underlined). The “ideal” context is: GCC GCC A/GCC AUG G; the start codon and critical nucleotides are bold. A purine (usually A) at -3 is extremely important (Kozak, 1989, 1991). Although codon 1 is the first AUG in the open reading frame, it lies in a less favorable context for translation: UUU CUA ACA AUG U. Predicted IaaM pro-

Figure 8. Gene silencing during mixed infections. K. daigremontiana stems (A and B) and potato discs (C and D) co-inoculated with wild-type A. tumefaciens (A348) and disarmed strain EHA101 containing oncogene-silencing plasmid pJP17 (A and C) or an empty vector pJP21 (B and D). Plant Physiol. Vol. 133, 2003

tein sequences from A. tumefaciens, Agrobacterium rhizogenes, and Agrobacterium vitis are 58% to 64% identical, except for the first 17 amino acids, which are not conserved; compare accession numbers NP_536129, AAD30493, AAC77909, and Q09109. Both wild-type and iaaM-stop mRNAs may initiate translation at codon 18. To examine sequence requirements for silencing iaaM, we deleted 189 bp from the 5⬘ end of iaaM-stop to make an untranslatable construct (pJC14; Fig. 7). K. daigremontiana developed wild-type tumors when coinoculated with A348 and EHA101 containing this untranslatable iaaM transgene (data not shown), indicating that the region containing translation initiation sequences was necessary for silencing iaaM. Likewise, potato discs co-inoculated with A348 and EHA101(pJC14) developed as many tumor foci as those co-inoculated with A348 and a vector-only control strain [EHA101(pJP21)]. In side-by-side comparisons, the translatable iaaM construct (pJP17) reduced tumor foci 3-fold relative to a vector-only control (P ⫽ 0.0005; n ⫽ 39), whereas the untranslatable derivative (pJC14) had no significant effect: Tumor foci were 1.3-fold higher than the vector-only control (P ⫽ 0.3; n ⫽ 38). Although the untranslatable transgene contained 90% of the iaaM sequences present in the successful iaaM-silencing transgene, the truncated construct did not silence iaaM. DISCUSSION

We repeatedly observed marked differences in the ability of iaaM and ipt transgenes (with premature 973

Lee et al.

stop codons) to elicit destruction of transgene RNAs and silence wild-type oncogenes. These differences occurred even though both transgenes were expressed from the same promoter, flanked by identical 5⬘- and 3⬘-UTRs, and located in the same context relative to other T-DNA genes, such as the nopaline synthase promoter-neomycin phosphotransferase (Pnos-nptII) gene fusion. In the genome of transgenic plants, T-DNA location, copy number, and structure varies from one plant line to another, and each of these properties can affect the quantity and direction of transcription (sense, antisense, or both) through a transgene. Nevertheless, the pronounced difference between the performance of the iaaM- and ipt-stop transgenes occurred in a large number of independent lines, which suggests that properties of the specific target sequences, rather than other factors, were primarily responsible for the large difference in success rate. Translation start sequences in the iaaM-stop transgene may explain its ability to silence the incoming T-DNA oncogene and to trigger destruction of its own RNA. The iaaM-stop transgene contained an in-frame AUG (codon 18) flanked by sequences that promote initiation of translation. A small (189-bp) deletion that removed these sequences abolished the ability of this transgene to silence iaaM. In contrast, the ipt-stop transgene lacked alternative translation start sites. Thus, transgenes encoding untranslatable sense-strand RNAs failed to elicit gene silencing, whereas translatable iaaM-stop constructs were highly effective at silencing iaaM. Similar observations were made for other transgenes. Gene silencing elicited by sense-strand chalcone synthase transgenes is reduced by premature stop codons in the coding sequence (Que et al., 1997). In another example, tobacco etch virus coat protein RNA with a stop codon near the 5⬘ end triggers gene silencing. However, this RNA contains an alternative translation start within the coding sequence, and degraded transgene RNA is found in polysomes in silenced plant tissue (Tanzer et al., 1997). Levels of transgene RNA in oncogene-silencing and non-silencing tobacco support our conclusion that translatable iaaM-stop sense-strand RNA elicited gene silencing very effectively whereas untranslatable ipt-stop (and truncated iaaM-stop) RNA did not. If a sense-strand RNA was a potent elicitor of gene silencing, we expected it to silence the target gene in any transgenic line that strongly expressed this RNA. In addition, we predicted weak (or no) transgene expression in lines that failed to silence the target gene. This is exactly what we observed with the translatable iaaM-stop construct: All non-silencing lines contained very little transgene RNA, but all iaaM-silencing lines contained significant amounts of degraded transgene RNA. Conversely, we reasoned that transgene RNAs unable to trigger gene silencing (such as the untranslatable sense-strand ipt-stop 974

RNA) might be present at high levels in many nonsilencing lines. This was the case in the non-silencing ipt-stop line we examined. The sole ipt-silencing line exhibited transcriptional gene silencing, not PTGS. This further supports our conclusion that the untranslatable ipt-stop transgene was unable to trigger PTGS, in contrast to the translatable iaaM-stop transgene, which was highly effective in the majority of transgenic lines. In this study, we have shown that the iaaM oncogene is susceptible to gene silencing and that the T-DNA-encoded auxin biosynthetic pathway can be blocked by this strategy. Rootstocks are resistant to crown gall when they are engineered to contain transgenes that silence iaaM. We used this approach to create apple rootstocks resistant to crown gall (Viss et al., 2003). Escobar et al. (2001, 2002) also created gall-resistant plants using constructs that produce double-stranded hairpin RNAs, which are more potent inducers of gene silencing than sense/ antisense RNAs (Chuang and Meyerowitz, 2000; Smith et al., 2000; Wesley et al., 2001). Together these studies show that silencing ipt requires stronger PTGS inducers than are needed to silence iaaM, and they prove that silencing the iaaM oncogene is an effective means to produce rootstocks resistant to crown gall. MATERIALS AND METHODS Transgene Construction Three transgenes containing premature stop codons were constructed: iaaM-stop, ipt-stop, and ipt::iaaM-stop (Fig. 1). The third codon of each gene was converted to a stop codon by PCR using primers containing the desired mutations. In addition to the nonsense codon, a frameshift mutation and a new restriction site were introduced into each transgene. A plasmid containing EcoRI fragment 7 (pRK290-pBR325-Eco 7; Garfinkel et al., 1981) from the octopine-type Ti plasmid pTiA6NC served as template for PCR amplification of ipt and iaaM. Reaction mixtures (50 ␮L) contained 1.5 mm MgCl2, 50 mm KCl, 10 mm Tris-HCl, pH 8.3, 0.001% (w/v) gelatin, deoxynucleoside triphosphates at 0.25 mm each, 2 units of Tfl DNA polymerase (Epicenter, Madison, WI), primer oligonucleotides at 1 ␮m each, and 400 ng of template DNA. Reactions were incubated in a thermal cycler (480, PerkinElmer Life Sciences, Boston) at 94°C for 1 min followed by 30 cycles of 94°C for 1 min, 48°C for 1 min, and 72°C for 2 min. Upon conclusion of thermal cycling, reactions were cooled to 15°C. The upstream primer used to produce iaaM-stop (shown 5⬘ to 3⬘; CGGGATCCATGTCATGAACCTCTCCTTGATAAC) contained a BamHI site (GGATCC) upstream of the first iaaM codon (bold italicized ATG). This primer contained a stop codon (bold TGA) in place of the third codon (GCT) of iaaM, thereby creating a BspHI site (TCATGA); it also lacked the first two bases (TC) of the fourth codon. The downstream primer used to amplify iaaM-stop (CGGGATCCTGCGACTCATAGT) included the BamHI site within the iaaM gene (Barker et al., 1983), yielding an 1,807-bp product containing the first 1,797 bp of the iaaM-stop gene. The upstream primer used to amplify ipt-stop (GAAGATCTGATCATGGACTGAATCTAATTTTCGGTCC) contained BglII (AGATCT) and BclI (TGATCA) sites upstream of the ipt start codon (italicized). In this primer, the first base (C) of the third codon was deleted, and the first base (C) of the fourth codon was changed to A. These mutations created a TGA nonsense codon (bold type), a frameshift mutation, and a HinfI site (GANTC). The downstream primer (GAAGATCTGATCACTAATACATTCCGAACGG) contained the complement of the ipt termination codon (bold CTA) and BclI and BglII sites at the 5⬘ end. PCR with these primers produced a 747-bp amplicon containing the entire ipt-stop coding sequence, which contains 722 bp.



PCR amplicons digested with BamHI (iaaM-stop) or BglII (ipt-stop) were joined to BamHI-digested pUC119 (Yanisch-Perron et al., 1985) using T4 DNA ligase and transformed into Escherichia coli strain TG1 (Maniatis et al., 1989). We selected ampicillin-resistant Lac⫺ transformants on L agar containing 50 ␮g mL⫺1 ampicillin, 0.004% (w/v) 5-bromo-4-chloro-3-indolyl␤-d-galactoside, and 4 ␮m isopropylthio-␤-galactoside. The presence of the mutations in iaaM-stop and ipt-stop were confirmed by restriction with BspHI and HinfI, respectively, and by dye-terminator DNA sequencing using an Applied Biosystems (Foster City, CA) model 373A DNA sequencer. Restriction endonucleases and T4 DNA ligase were purchased from New England Biolabs (Beverly, MA) or Invitrogen (Carlsbad, CA) and used as directed. Mutant oncogenes were introduced into the BamHI site of plant transformation vector pPEV6 (Lindbo and Dougherty, 1992a) such that transcription from the vector’s CaMV 35S promoter produced sense-strand RNA from the oncogene (Fig. 1). The iaaM-stop oncogene was excised from pUC119 as a BamHI fragment and inserted into pPEV6, creating plasmid pTDP1 (Fig. 1). The pUC119::ipt-stop plasmid was transformed into E. coli strain GM119 (Marinus, 1984), an adenine methylase (dam-) mutant, to obtain unmethylated plasmid DNA susceptible to cleavage by BclI. We excised ipt-stop from pUC119 as a BclI fragment and inserted it into the BamHI site of pPEV6. The resulting plasmid, pCW1, contained a single BamHI site, which was located in ipt (Fig. 1). We inserted iaaM-stop into this BamHI site to create the hybrid transgene in pCW4 (Fig. 1). These pPEV6 derivatives, which encode kanamycin resistance, were mobilized from E. coli strain TG1 into Agrobacterium tumefaciens by tri-parental mating with E. coli strain MM294(pRK2013). The A. tumefaciens recipient harbored disarmed Ti plasmid pCIB542, which encodes resistance to spectinomycin and streptomycin (Lindbo and Dougherty, 1992a). We performed matings as described (Lindbo and Dougherty, 1992a) and selected transconjugants on AB minimal agar (Garfinkel et al., 1981) containing 250 ␮g mL⫺1 streptomycin, 25 ␮g mL⫺1 spectinomycin, and 25 ␮g mL⫺1 kanamycin. To create a transgene designed to produce both sense and antisense RNAs from iaaM and ipt, we fused the iaaM- and ipt-stop transgenes in tandem and situated them between opposing promoters (Fig. 7). First, we constructed a plasmid (pJP8) that contained opposing CaMV and FMV promoters. We amplified the enhanced FMV promoter from pCGN8063 (McBride and Summerfelt, 1990) using: GCCGAGATCTAGCTTCTGCAGGTCCTGCTC (forward primer) and AATAATTCTAGAAGGCCTGAATTCGAGCTCGGTACCGGATCCGTGTCCTTCAAATGGGAATG (reverse primer). This PCR product was inserted into pCR2.1-TOPO (Invitrogen), excised with BglII and XbaI, and ligated to pCGN8059 (McBride and Summerfelt, 1990) digested with BamHI and XbaI to yield pJP8. The ipt-stop transgene was amplified from pUC119::ipt-stop using: GAAGATCTAGAATGGACTGAATCTAATTTTCGGTCC (forward primer) and GAAGATCTGATCACTAATACATTCCGAACGG (reverse primer). The PCR amplicon was inserted into pCR2.1-TOPO. Next, ipt-stop was excised with XbaI and BclI and ligated to pJP8 digested with BamHI and XbaI to yield pJP11. The iaaM-stop transgene was amplified from pUC119::iaaM-stop using: CGGGATCCATGTCATGAACCTCTCCTTGATAAC (forward primer) and GGAATTCTAGATCCTGCGACTCATAGTCCAGGC (reverse primer). The PCR product was inserted into pCR2.1-TOPO, excised with BamHI and XbaI, and ligated to pJP11 digested with BglII and XbaI to yield pJP15. A NotI fragment containing the iaaM-ipt-stop transgene flanked by the CaMV and FMV promoters was excised from pJP15 and inserted into the NotI site of pCGN5927 (McBride and Summerfelt, 1990) to form pJP17 (Fig. 7). We constructed pJC14 in the same manner as pJP17, except that the iaaM-stop forward primer was: TTGCTGATCACTGGCCGATGGTCGCTTCCCC. We constructed the vector-only control plasmid (pJP21) by inserting a NotI fragment that contains the CaMV and FMV promoters from pJP8 into the NotI site of pCGN5927.

Distribution of Materials Upon request, all novel materials described in this publication will be made available in a timely manner for noncommercial research purposes, subject to the requisite permission from any third-party owners of all or parts of the material. Obtaining any permissions will be the responsibility of the requestor.


Plant Transformation Leaf discs of haploid tobacco (Nicotiana tabacum cvs Burley 49 and Kentucky 149) were transformed (Horsch et al., 1985; Lindbo and Dougherty, 1992a) with A. tumefaciens A136(pCIB542) derivatives harboring pTDP1, pCW1, pCW4, or pPEV6. Shoots were obtained on semisolid Murashige and Skoog (1962) medium containing 1 mg L⫺1 N6-benzyladenine, 0.1 mg L⫺1 ␣-naphthalene acetic acid, 400 mg L⫺1 timentin, and 200 mg L⫺1 kanamycin. Kanamycin-resistant shoots were transferred to Murashige and Skoog medium with 400 mg L⫺1 timentin and 400 mg L⫺1 kanamycin for rooting, and resulting plantlets were transplanted to soil. The chromosome numbers were doubled as described (Smith et al., 1994). Midveins from fully expanded leaves of greenhouse-grown plants were soaked in 70% (v/v) ethanol for 1 min and 10% (v/v) bleach for 10 min, cut in 2-cm sections, and placed on Murashige and Skoog medium with 2.5 mg L⫺1 kinetin, 4 mg L⫺1 ␤-indoleacetic acid, and 50 mg L⫺1 kanamycin. Shoots that formed on the cultured veins were transferred to Murashige and Skoog medium with 50 mg L⫺1 kanamycin for rooting. Plants were grown in the greenhouse, and seed was obtained from fertile plants.

Virulence Tests A. tumefaciens strains used to challenge potential oncogene-silencing transgenic plants contained either wild-type pTiA6NC (A348) or derivatives with Tn5 insertions in iaaM (strain 328) or ipt (strain 338) in the A136 genetic background (Garfinkel et al., 1981). Bacteria were cultured on AB-Glc minimal or yeast extract-peptone agar (Garfinkel et al., 1981) at 28°C for 3 d and then used to inoculate plants. We used sterile toothpicks to create 1- to 2-cm stem wounds, which we filled with A. tumefaciens and covered with laboratory film (Parafilm, American National Can, Greenwich, CT). Inoculated plants were grown in the greenhouse or a growth room (at 24°C with 16-h days) and scored for tumors after 4, 6, and 10 weeks.

Mixed Infection Assays A. tumefaciens cultures were grown overnight in yeast extract-peptone broth (Garfinkel et al., 1981) at 28°C with aeration. K. daigremontiana stems were wounded with a sterile toothpick, and each wound was inoculated with 5 ␮L of broth culture containing EHA101(pJP17; iaaM-silencing strain) and A348 (wild-type A. tumefaciens) mixed in a 4:1 (v/v) ratio. Plants were maintained in a greenhouse or growth room and were scored 6 to 12 weeks after inoculation. Potato discs were inoculated in the same manner, except cultures were washed once in phosphate-buffered saline buffer and suspended in 1 volume of phosphate-buffered saline before inoculation (Pueppke and Benny, 1981). Potato (cv Red Norland or cv Pontiac) discs 3 mm thick were cut from 7-mm-diameter cores, placed on water agar, and inoculated with 5 ␮L of mixed culture within 10 min of cutting. Discs were incubated at room temperature and were scored after 3 weeks.

RNA and DNA Isolation RNA and DNA were extracted from leaves of transgenic tobacco using SDS/phenol followed by LiCl precipitation as described previously (Ream and Field, 1999). Leaves (2.5 g) were frozen and ground to a fine powder in liquid nitrogen. Nucleic acids were extracted by mixing the ground tissue vigorously with 2.5 mL of extraction buffer (100 mm LiCl, 1% [w/v] SDS, 10 mm EDTA, and 100 mm Tris, pH 9) and 2.5 mL of phenol:chloroform (24:1, v/v). After centrifugation, the aqueous phase was recovered, mixed with an equal volume of 4 m LiCl, and incubated overnight at ⫺20°C to precipitate total RNA, which we collected by centrifugation. The supernatant, which contained genomic DNA, was mixed with 2 volumes of ethanol and incubated at 4°C to precipitate the DNA. The RNA and DNA pellets were dissolved in 200 ␮L of water or DNA buffer (10 mm Tris, pH 7.5, and 0.1 mm EDTA), precipitated again with 3 volumes of ethanol and 0.5 volume of ammonium acetate, and washed with 70% (v/v) ethanol. The pellets were dissolved in 25 ␮L of water or 50 ␮L of DNA buffer, and the A260 was measured.

Southern- and Northern-Blot Analysis Samples containing 5 to 10 ␮g of total RNA or restriction endonucleasedigested genomic DNA were subjected to agarose gel electrophoresis and

975

Lee et al.

were transferred to Gene Screen Plus nylon membranes (New England Nuclear, Boston). Procedures for capillary blotting and hybridization were as described (Ream and Field, 1999). DNA probes were generated using a nick translation kit (Invitrogen) to incorporate [␣-32P]dCTP (3,000 Ci mmol⫺1; New England Nuclear). Strand-specific RNA probes were created by inserting an iaaM-stop PCR amplicon into pCR-TOPO topoisomerasebound plasmid DNA (Invitrogen) as directed by the manufacturer. An in vitro transcription system (Promega, Madison, WI) was used to produce iaaM-stop RNA containing [␣-32P]UTP (3,000 Ci mmol⫺1). Specific activity of the probes varied from 106 to 109 cpm ␮g⫺1. Signals were detected on Kodak XAR film (Eastman Kodak, Rochester, NY).

Nuclear RNA Extraction and Real-Time RT-PCR Nuclei were isolated from 25 g of tobacco leaf tissue from plants incubated in the dark 2 d before harvest. Nuclear RNA isolation was carried out as described (Cox and Goldberg, 1988) with the following modification: Honda buffer added to ground tissue was supplemented with 0.5 m Suc. Real-time RT-PCR was carried out using an Applied Biosystems Prism 7700 Sequence Detector with version 1.7 software and the Quantitect SYBR Green RT-PCR kit (Qiagen USA, Valencia, CA). Three targets were examined: ipt, iaaM, and 5.8S rRNA. PCR primers amplified products of 100, 150, and 165 bp, respectively. The primers for iaaM and ipt amplified a region from the 3⬘ end of each transgene; primers were: iaaM sense strand, GAACCAAGCGGTTGATAACAGCC; iaaM antisense strand, CTGCGACTCATAGTCCAGGAATAC; ipt sense strand, CCATGCGCGCCAACAGGA; and ipt antisense strand, GAAGATCTGATCACTAATACATTCCGAAC. The entire 5.8s rRNA gene from tobacco was amplified using: GACTCTCGGCAACGGATATCTC (5⬘ primer) and ACGCGATGCGTGACGCCC (3⬘ primer). Regular RT-PCR was performed using the Qiagen One-Step RT-PCR kit to test primer efficiency and to produce standards for real-time RT-PCR. A standard curve was constructed for every primer set using RNA transcribed in vitro (Riboprobe kit, Promega). RNA stocks were diluted in 2-fold increments from 10 to 1,280 pg. Template plasmids for in vitro transcription contained full-length iaaM-stop (pHL183) and ipt-stop (pJH3) transgenes in pCR2.1 (Invitrogen, Carlsbad, CA). Plasmids were linearized with SacI before transcription. The 5.8s rRNA template (pJH13) contained an apple 5.8S rRNA gene obtained by RT-PCR and inserted into pCR4-TOPO (Invitrogen). This plasmid was linearized with PstI before in vitro transcription. Standard curves were obtained as outlined in the User Bulletin No. 2 for the ABI Prism 7700 Sequence Detection system. The iaaM and ipt RNA levels were normalized between samples to the 5.8s rRNA level. Real-time RT-PCR products were separated on 2% (w/v) agarose gels to check product size.

ACKNOWLEDGMENTS We thank Drs. Bill Dougherty, Bill Proebsting, and Stanton B. Gelvin for advice and encouragement, and we thank Dr. David C. Baulcombe for an especially productive discussion in which he suggested that we examine iaaM for alternate translation initiation sites. Received May 6, 2003; returned for revision May 30, 2003; accepted June 6, 2003.

LITERATURE CITED Akiyoshi DE, Klee H, Amasino RM, Nester EW, Gordon MP (1984) T-DNA of Agrobacterium tumefaciens encodes an enzyme of cytokinin biosynthesis. Proc Natl Acad Sci USA 81: 5994–5998 Angell SM, Baulcombe DC (1997) Consistent gene silencing in transgenic plants expressing a replicating potato virus X RNA. EMBO J 16: 3675–3684 Barker R, Idler K, Thompson D, Kemp J (1983) Nucleotide sequence of the T-DNA region of the Agrobacterium tumefaciens octopine Ti plasmid pTi15955. Plant Mol Biol 2: 335–350 Barry GF, Rogers SG, Fraley RT, Brand L (1984) Identification of a cloned cytokinin biosynthetic gene. Proc Natl Acad Sci USA 81: 4776–4780 Bass BL (2000) Double-stranded RNA as a template for gene silencing. Cell 101: 235–238 Baulcombe D (2002) An RNA microcosm. Science 297: 2002–2003

976

Baulcombe DC (2000) Unwinding RNA silencing. Science 290: 1108–1109 Buchmann I, Marner FJ, Schroder G, Waffenschmidt S, Schroder J (1985) Tumor genes in plants: T-DNA encoded cytokinin biosynthesis. EMBO J 4: 853–859 Chuang CF, Meyerowitz EM (2000) Specific and heritable genetic interference by double-stranded RNA in Arabidopsis thaliana. Proc Natl Acad Sci USA 97: 4985–4990 Cluster PD, O’Dell M, Metzlaff M, Flavell RB (1996) Details of T-DNA structural organization from a transgenic Petunia population exhibiting co-suppression. Plant Mol Biol 32: 1197–1203 Cogoni C, Macino G (1999a) Gene silencing in Neurospora crassa requires a protein homologous to RNA-dependent RNA polymerase. Nature 399: 166–169 Cogoni C, Macino G (1999b) Posttranscriptional gene silencing in neurospora by a RecQ DNA helicase. Science 286: 2342–2344 Conner JA, Tantikanjana T, Stein JC, Kandasamy MK, Nasrallah JB, Nasrallah ME (1997) Transgene-induced silencing of S-locus genes and related genes in Brassica. Plant J 11: 809–823 Cox KH, Goldberg RB (1988) Analysis of plant gene expression. In CH Shaw, ed, Plant Molecular Biology: A Practical Approach, Ed 1. IRL Press, Oxford, pp 1–35 Dombek P, Ream W (1997) Functional domains of Agrobacterium tumefaciens single-stranded DNA-binding protein VirE2. J Bacteriol 179: 1165–1173 Dougherty WG, Lindbo JA, Smith HA, Parks TD, Swaney S, Proebsting WM (1994) RNA-mediated virus resistance in transgenic plants: exploitation of a cellular pathway possibly involved in RNA degradation. Mol Plant-Microbe Interact 7: 544–552 Dougherty WG, Parks TD (1995) Transgenes and gene suppression: telling us something new? Curr Opin Cell Biol 7: 399–405 Escobar MA, Civerolo EL, Polito VS, Pinney KA, Dandekar AM (2003) Characterization of oncogene-silenced transgenic plants: implications for Agrobacterium biology and post-transcriptional gene silencing. Mol Plant Pathol 4: 57–65 Escobar MA, Civerolo EL, Summerfelt KR, Dandekar AM (2001) RNAimediated oncogene silencing confers resistance to crown gall tumorigenesis. Proc Natl Acad Sci USA 98: 13437–13442 Escobar MA, Leslie CA, McGranahan GH, Dandekar AM (2002) Silencing crown gall disease in walnut (Juglans regia L.). Plant Sci 163: 591–597 Fire A, Xu SQ, Montgomery MK, Kostas SA, Driver SE, Mello CC (1998) Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391: 806–811 Garfinkel DJ, Simpson RB, Ream LW, White FF, Gordon MP, Nester EW (1981) Genetic analysis of crown gall: fine structure map of the T-DNA by site-directed mutagenesis. Cell 27: 143–153 Hamilton AJ, Baulcombe DC (1999) A species of small antisense RNA in posttranscriptional gene silencing in plants. Science 286: 950–952 Horsch RB, Fry JE, Hoffmann NL, Eichholtz D, Rogers SG (1985) A simple and general method of transferring genes into plants. Science 227: 1229–1231 Inze´ D, Follin A, Van Lijsebettens M, Simoens C, Genetello C, Van Montagu M, Schell J (1984) Genetic analysis of the individual T-DNA genes of Agrobacterium tumefaciens: further evidence that two genes are involved in indole-3-acetic acid synthesis. Mol Gen Genet 194: 265–274 Jorgensen RA, Atkinson RG, Forster RLS, Lucas WJ (1998) An RNA-based information superhighway in plants. Science 279: 1486–1487 Kennerdell JR, Carthew RW (1998) Use of dsRNA-mediated genetic interference to demonstrate that frizzled and frizzled 2 act in the wingless pathway. Cell 95: 1017–1026 Kozak M (1989) The scanning model for translation: an update. J Cell Biol 108: 229–241 Kozak M (1991) Structural features in eukaryotic mRNAs that modulate the initiation of translation. J Biol Chem 266: 19867–19870 Lindbo JA, Dougherty WG (1992a) Pathogen-derived resistance to a potyvirus: immune and resistant phenotypes in transgenic tobacco expressing altered forms of a potyvirus coat protein nucleotide sequence. Mol PlantMicrobe Interact 5: 144–153 Lindbo JA, Dougherty WG (1992b) Untranslatable transcripts of the tobacco etch virus coat protein gene sequence can interfere with tobacco etch virus replication in transgenic plants and protoplasts. Virology 189: 725–733 Llave C, Xie Z, Kasschau KD, Carrington JC (2002) Cleavage of Scarecrowlike mRNA targets directed by a class of Arabidopsis miRNA. Science 297: 2053–2060



Maniatis T, Fritsch EF, Sambrook J (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Marano MR, Baulcombe D (1998) Pathogen-derived resistance targeted against the negative-strand RNA of tobacco mosaic virus: RNA strandspecific gene silencing? Plant J 13: 537–546 Marinus MG (1984) Methylation of procaryotic DNA. In A Razin, M Cedar, AD Riggs, eds, DNA Methylation, Ed 1. Springer, New York, pp 81–109 McBride KE, Summerfelt K (1990) Improved binary vectors for Agrobacterium-mediated plant transformation. Plant Mol Biol 14: 269–276 Miranda A, Janssen G, Hodges L, Peralta EG, Ream W (1992) Agrobacterium tumefaciens transfers extremely long T-DNAs by a unidirectional mechanism. J Bacteriol 174: 2288–2297 Montgomery MK, Fire A (1998) Double-stranded RNA as a mediator in sequence-specific genetic silencing and co-suppression. Trends Genet 14: 255–258 Montgomery MK, Xu SQ, Fire A (1998) RNA as a target of double-stranded RNA-mediated genetic interference in Caenorhabditis elegans. Proc Natl Acad Sci USA 95: 15502–15507 Moore LW (1988) The use of Agrobacterium radiobacter in agricultural ecosystems. Microbiol Sci 5: 92–95 Murashige T, Skoog F (1962) A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol Plant 15: 473–497 Nam J, Matthysse AG, Gelvin SB (1997) Differences in susceptibility of Arabidopsis ecotypes to crown gall disease may result from a deficiency in T-DNA integration. Plant Cell 9: 317–333 Nam J, Mysore KS, Zheng C, Knue MK, Matthysse AG, Gelvin SB (1999) Identification of T-DNA tagged Arabidopsis mutants that are resistant to transformation by Agrobacterium. Mol Gen Genet 261: 429–438 Napoli C, Lemieux C, Jorgensen RA (1990) Introduction of a chimeric chalcone synthase gene into Petunia results in reversible co-suppression of homologous genes in trans. Plant Cell 2: 279–289 Ngo H, Tschudi C, Gull K, Ullu E (1998) Double-stranded RNA induces mRNA degradation in Trypanosoma brucei. Proc Natl Acad Sci USA 95: 14687–14692 Palauqui JC, Elmayan T, Pollien JM, Vaucheret H (1997) Systemic acquired silencing: transgene-specific post-transcriptional silencing is transmitted by grafting from silenced stocks to non-silenced scions. EMBO J 16: 4738–4745 Pueppke SG, Benny UK (1981) Induction of tumors on Solanum tuberosum L. by Agrobacterium: quantitative analysis, inhibition by carbohydrates, and virulence of selected strains. Physiol Plant Pathol 18: 169–179 Que Q, Wang HY, English JJ, Jorgensen RA (1997) The frequency and degree of cosuppression by sense chalcone synthase transgenes are dependent on transgene promoter strength and are reduced by premature nonsense codons in the transgene coding sequence. Plant Cell 9: 1357–1368 Ream LW, Gordon MP, Nester EW (1983) Multiple mutations in the T-region of the Agrobacterium tumefaciens tumor-inducing plasmid. Proc Natl Acad Sci USA 80: 1660–1664 Ream W, Field KG (1999) Molecular biology techniques: an intensive laboratory course. Academic Press, San Diego Ruiz MT, Voinnet O, Baulcombe DC (1998) Initiation and maintenance of virus-induced gene silencing. Plant Cell 10: 937–946 Schroeder G, Waffenschmidt S, Weiler EW, Schroeder J (1984) The T-region of Ti plasmids codes for an enzyme synthesizing indole-3-acetic acid. Eur J Biochem 138: 387–391 Shurvinton CE, Hodges L, Ream W (1992) A nuclear localization signal and the C-terminal omega sequence in the Agrobacterium tumefaciens VirD2 endonuclease are important for tumor formation. Proc Natl Acad Sci USA 89: 11837–11841 Shurvinton CE, Ream W (1991) Stimulation of Agrobacterium tumefaciens T-DNA transfer by overdrive depends on a flanking sequence but not on


helical position with respect to the border repeat. J Bacteriol 173: 5558–5563 Smith HA, Swaney SL, Parks TD, Wernsman EA, Dougherty WG (1994) Transgenic plant virus resistance mediated by untranslatable sense RNAs: expression, regulation, and fate of nonessential RNAs. Plant Cell 6: 1441–1453 Smith NA, Singh SP, Wang MB, Stoutjesdijk PA, Green AG, Waterhouse PM (2000) Total silencing by intron-spliced hairpin RNAs. Nature 407: 319–320 Stam M, de Bruin R, Kenter S, van der Hoorn RAL, van Blokland R, Mol JNM, Kooter JM (1997) Post-transcriptional silencing of chalcone synthase in Petunia by inverted transgene repeats. Plant J 12: 63–82 Sundberg C, Meek L, Carroll K, Das A, Ream W (1996) VirE1 protein mediates export of the single-stranded DNA-binding protein VirE2 from Agrobacterium tumefaciens into plant cells. J Bacteriol 178: 1207–1212 Tanzer MM, Thompson WF, Law MD, Wernsman EA, Uknes S (1997) Characterization of post-transcriptionally suppressed transgene expression that confers resistance to tobacco etch virus infection in tobacco. Plant Cell 9: 1411–1423 Thomashow LS, Reeves S, Thomashow MF (1985) Crown gall oncogenesis: evidence that a T-DNA gene from the Agrobacterium Ti plasmid pTiA6 encodes an enzyme that catalyzes synthesis of indoleacetic acid. Proc Natl Acad Sci USA 81: 5071–5075 Thomashow MF, Hugly S, Buchholz WG, Thomashow LS (1986) Molecular basis for the auxin-independent phenotype of crown gall tumor tissues. Science 231: 616–618 Tijsterman M, Ketting RF, Plasterk RHA (2002) The genetics of RNA silencing. Annu Rev Genet 36: 489–519 Timmons L, Fire A (1998) Specific interference by ingested dsRNA. Nature 395: 854 Tuschl T, Zamore PD, Lehmann R, Bartel DP, Sharp PA (1999) Targeted mRNA degradation by double-stranded RNA in vitro. Genes Dev 13: 3191–3197 van der Krol AR, Mur LA, Beld M, Mol JN, Stuitje AR (1990) Flavonid genes in Petunia: addition of a limited number of gene copies may lead to a suppression of gene expression. Plant Cell 2: 291–299 Van Onckelen H, Prinsen E, Inze´ D, Rudelsheim P, Van Lijsebettens M, Follin A, Schell J, Van Montagu M, De Greef J (1986) Agrobacterium T-DNA gene codes for tryptophan 2-monooxygenase activity in tobacco crown gall cells. FEBS Lett 198: 357–360 Viss W, Humann JL, Cook M, Driver J, Ream W (2003) Crown-gallresistant transgenic apple trees that silence Agrobacterium tumefaciens T-DNA oncogenes. Molec Breeding (in press) Voinnet O, Baulcombe DC (1997) Systemic signalling in gene silencing. Nature 389: 553 Voinnet O, Vain P, Angell S, Baulcombe DC (1998) Systemic spread of sequence-specific transgene RNA degradation in plants is initiated by localized introduction of ectopic promoterless DNA. Cell 95: 177–187 Waterhouse PM, Graham MW, Wang MB (1998) Virus resistance and gene silencing in plants can be induced by simultaneous expression of sense and antisense RNA. Proc Natl Acad Sci USA 95: 13959–13964 Wesley SV, Helliwell CA, Smith NA, Wang MB, Rouse DT, Liu Q, Gooding PS, Singh SP, Abbott D, Stoutjesdijk PA et al. (2001) Construct design for efficient, effective and high-throughput gene silencing in plants. Plant J 27: 581–590 Winans SC (1992) Two-way chemical signaling in Agrobacterium-plant interactions. Micro Rev 56: 12–31 Yanisch-Perron C, Vieira J, Messing J (1985) Improved M13 phage cloning vectors and host strains: nucleotide sequence of the M13 mp18 and pUC19 vectors. Gene 33: 103–119 Zamore PD, Tuschl T, Sharp PA, Bartel DP (2000) RNAi: double-stranded RNA directs the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell 101: 25–33

977