Plant Physiol. (1998) 117: 1433-1443

Premature Polyadenylation at Multiple Sites within a Bacillus thuringiensis Toxin Gene-Coding Region¹

Scott H. Diehn², Wan-Ling Chiu³, E. Jay De Rocher, and Pamela J. Green^*

Michigan State University-Department of Energy Plant Research Laboratory (S.H.D., W.-L.C., E.J.D.R., P.J.G.), Department of Botany and Plant Pathology (S.H.D.), and Department of Biochemistry (P.J.G.), Michigan State University, East Lansing, Michigan 48824

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

Some foreign genes introduced into plants are poorly expressed, even when transcription is controlled by a strong promoter. Perhaps the best examples of this problem are the cry genes of Bacillus thuringiensis (B.t.), which encode the insecticidal proteins commonly referred to as B.t. toxins. As a step toward overcoming such problems most effectively, we sought to elucidate the mechanisms limiting the expression of a typical B.t.-toxin gene, cryIA(c), which accumulates very little mRNA in tobacco (Nicotiana tabacum) cells. Most cell lines transformed with the cryIA(c) B.t.-toxin gene accumulate short, polyadenylated transcripts. The abundance of these transcripts can be increased by treating the cells with cycloheximide, a translation inhibitor that can stabilize many unstable transcripts. Using a series of hybridizations, reverse-transcriptase polymerase chain reactions, and RNase-H-digestion experiments, poly(A⁺) addition sites were identified in the B.t.-toxin-coding region corresponding to the short transcripts. A fourth polyadenylation site was identified using a chimeric gene. These results demonstrate for the first time to our knowledge that premature polyadenylation can limit the expression of a foreign gene in plants. Moreover, this work emphasizes that further study of the fundamental principles governing polyadenylation in plants will have basic as well as applied significance.

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

The ability to express foreign genes in plants has been an invaluable tool in understanding normal plant growth and development. Many molecular and biochemical questions concerning plant metabolism, physiology, development, and responses to environmental cues have been addressed in this fashion. With respect to plant biotechnology, the introduction of foreign genes has led to a variety of significant advancements in crop improvement. As a result, the literature contains numerous reports demonstrating the successful expression of foreign genes in a variety of plants. However, this is not always the case. There is a growing list of foreign genes that are poorly expressed in plants.

The gene encoding the GFP from Aequorea victoria, which has been used as a reporter gene in many different biological systems, is not expressed well in some plant species. Little or no GFP-related fluorescence can be detected in Arabidopsis, tobacco (Nicotiana tabacum), or barley protoplasts, or in Arabidopsis and tobacco plants transformed with the GFP gene, even when transcription is directed by the CaMV 35S promoter (Haseloff and Amos, 1995; Reichel et al., 1996; Haseloff et al., 1997; Rouwendal et al., 1997). Similarly, the genes encoding T4 lysozyme, Klebsiella pneumoniae cyclodextrin glycosyltransferase and bacterial mercuric ion reductase are expressed at very low levels in plants despite the use of strong or tissue-specific promoters to direct their transcription. Potato plants expressing the T4 lysozyme or K. pneumoniae cyclodextrin glycosyltransferase genes accumulate very low levels of the corresponding mRNA and protein (Oakes et al., 1991; Düring et al., 1993). Tobacco plants expressing the T4 lysozyme gene under the control of the mannopine synthase promoter instead of the CaMV 35S promoter used in the transgenic potato plants also accumulate very low levels of lysozyme protein (Düring, 1988). The full-length transcript of the bacterial mercuric ion reductase gene is not detectable in transgenic petunia plants (Thompson, 1990); instead, two short transcripts of approximately 800 nucleotides accumulate.

The genes best known for their low expression in plants are the Bacillus thuringiensis (B.t.)-toxin genes (for review, see Diehn et al., 1996). This family of genes from the gram-positive soil bacterium B.t. encodes potent insecticidal proteins that target specific orders of insects (Höfte and Whiteley, 1989; Aronson, 1993). Initial efforts to express B.t.-toxin genes in plants using standard approaches yielded transgenic plants that produced little or no B.t.-toxin mRNA or protein (Barton et al., 1987; Vaeck et al., 1987). The few plants generated that did express B.t. toxin were only resistant to those insect species that were the most susceptible to the toxin (Fischhoff et al., 1987; Delanney et al., 1989). The problem appeared to be at the level of mRNA accumulation, because the plants that accumulated detectable B.t.-toxin mRNA were also the most insect resistant (Barton et al., 1987; Vaeck et al., 1987; Cheng et al., 1992; Dandekar et al., 1994).

Because of the potential agronomic importance of B.t.-toxin genes, considerable effort has been made to increase the expression of these genes in plants. These efforts include expressing only the region of the gene encoding the insecticidal domain, modifying the 5 and 3 UTRs, generating protein fusions, and using a variety of strong promoters (for review, see Diehn et al., 1996). The mRNA and protein levels were eventually increased by resynthesizing the genes to be more "plant like." In most cases, this included changing the codon usage to a plant-preferred codon bias (for review, see Diehn et al., 1996), which also has the effect of raising the G/C content of the wild-type gene. Many plant RNA-processing signals, in particular those for polyadenylation, mRNA decay, and splicing, are A/T rich. Wild-type B.t.-toxin genes have an A/T content of approximately 65%. Therefore, increasing the G/C content of the genes may eliminate potential RNA-processing signals.

Why the transcripts of some poorly expressed foreign genes fail to accumulate in plants remains unclear in most cases. The problem can occur at one or more steps in gene expression. mRNA accumulation may be limited at the level of transcription by sequences within the coding region that adversely affect transcription initiation or elongation (Adang et al., 1987; Oakes et al., 1991). Alternatively, the problem may occur posttranscriptionally (Fischhoff et al., 1987; Vaeck et al., 1987; Oakes et al., 1991) as a result of aberrant splicing and/or degradation of the transcript. Recently, the transcripts of GFP and a cryIA(b) B.t.-toxin gene were found to contain one and three cryptic introns, respectively (Haseloff and Amos, 1995; Van Aarssen et al., 1995; Haseloff et al., 1997). The splicing of these transcripts was shown to be partly responsible for the low expression of the GFP and cryIA(b) genes in plants.

It has been proposed that B.t.-toxin-coding regions contain plant-polyadenylation signals (Adang et al., 1987; Perlak et al., 1991). However, no reports documenting polyadenylation in the coding region of B.t.-toxin transcripts have been published. In this report we provide direct evidence demonstrating that the transcript of a cryIA(c) B.t.-toxin gene is polyadenylated in the coding region. Multiple sequence elements in the B.t.-toxin-coding region are recognized by plant cells as polyadenylation signals. Recognition of these signals appears to be one factor contributing to the low accumulation of the full-length B.t.-toxin transcript in plant cells. Another limitation of cryIA(c) B.t.-toxin transcript accumulation in plant cells, posed at the level of mRNA stability, is described in the accompanying paper (De Rocher et al., 1998). Elucidating the mechanisms responsible for the low accumulation of B.t.-toxin mRNA in plants may make it easier in the future to engineer novel foreign genes and other B.t.-toxin genes for high expression. It may also provide insight into natural gene-expression mechanisms in plants.

	MATERIALS AND METHODS

Top Abstract Introduction Methods Results Discussion References

Plant Materials and Treatments

1

Plasmid Construction

The resulting plasmid, p995, was digested with BglII and BamHI to release the B.t.-toxin-coding region with its modified 5 and 3 ends (lacking the pSP64-poly[A⁺] multiple cloning site and the poly[A⁺] sequence). This B.t.-toxin-coding region fragment was then used to replace the -globin-coding region of p1185, which originally contained an expression cassette consisting of a 2X35S-globin-coding region and the 3 UTR from the pea Rubisco small subunit rbcS-E9 gene to generate p1204. The construction of p1185 is described below. The resulting 2X35S-B.t. toxin-E9 gene cassette from p1204 was then inserted into the HindIII site of the binary vector pBI121 (accession no. X77672) to make p1205, which was used for integration into the genome of BY-2 cells

p1185 is a pUC 8 plasmid containing a 2X35S, a globin-coding region, and the pea rbcS-E9 3 UTR that was generated from pMF6 (Goff et al., 1991), a plasmid kindly provided by Michael Fromm at U.S. Department of Agriculture/University of California, Berkeley. To create the 2X35S a copy of the CaMV 35S enhancer contained on a HincII-EcoRV fragment of pMF6 was inserted into the EcoRV site of a second pMF6 CaMV 35S promoter. The nopaline-synthase polyadenylation sequence of the modified pMF6 plasmid, now called p1079, was removed by digestion with KpnI and ScaI. After creating blunt ends at the KpnI site of p1079, a BglII-ScaI fragment from a second p1079 plasmid, also treated to create blunt ends, was ligated to the vector to generate the plasmid p1138.

The alcohol dehydrogenase 1 (ADH1) intron 1 from the original pMF6 plasmid was removed from p1138 by digestion with EcoRV and BamHI. To replace the region of the 2X35S that was excised with the intron, the EcoRV-BamHI fragment from p1163 was ligated into p1138 to form the plasmid p1166. The p1163 EcoRV-BamHI fragment consisted of the region downstream of 90 (EcoRV) of the CaMV 35S promoter through the 5 UTR (BglII) of the CaMV 35S-globin-E9 cassette described by Newman et al. (1993), plus irrelevant polylinker sequences between BglII and BamHI that were excised in the next cloning step. Finally, the human -globin-coding region and E9 3 UTR on a BglII-ClaI fragment from p977 (described by De Rocher et al., 1998) was inserted into BglII-ClaI-digested p1166 behind the 2X35S to generate p1185. The 2X35S-globin-E9 gene cassette was removed from p1185 by HindIII digestion for insertion into the HindIII site of the binary vector pBI121 to make p1528.

To construct the globin-B.t. chimeric gene, the AccI-BamHI fragment (segment 4) was excised from the cryIA(c) B.t.-toxin-coding region in p995. After creating blunt ends with T4 DNA polymerase, the fragment was introduced into the unique EcoRV site of p948, a Bluescript II SK() vector in which the region of the polylinker between the SacI and ClaI sites was replaced with a polylinker (GAGCTCAGATCTTCTAGAGATATCGGATTCATCGAT) containing BglII, XbaI, EcoRV, and BamHI restriction sites to generate p1167. After digestion with BglII and BamHI, the segment 4 DNA fragment was inserted into the unique BamHI site between the globin-coding region and the E9 3 UTR of p1160, which contained an expression cassette consisting of 2X35S-ADH1 intron 1-globin-coding region-rbcS-E9 3 UTR (De Rocher et al., 1998), to create p1171. A DNA fragment carrying B.t.-toxin segment 4 and the E9 3 UTR was excised from p1171 with BamHI and ScaI and substituted for the region of p1185 between BamHI and ScaI containing the E9 3 UTR to create the expression cassette 2X35S-globin-B.t.-toxin segment 4-E9 in p1188. The expression cassette was then excised with HindIII and introduced into the HindIII site of the binary vector pBI121 to make plasmid p1194 for standard A. tumefaciens-mediated transformation of BY-2 cells.

RNA Methods

1

View larger version (30K): [in this window] [in a new window]   Figure 3. Characterization of the short B.t.-toxin transcripts by hybridization with different segments of the coding region. A, The cryIA(c) B.t.-toxin-coding region was divided into four segments using convenient restriction sites. Segment 1 consists of an SphI-XbaI restriction fragment from positions 28 to 295; segment 2 is a restriction fragment extending from positions 295 to 670 and is defined by XbaI sites at each end; segment 3 is an XbaI-AccI restriction fragment from positions 670 to 1201; and segment 4 is the remainder of the coding region from the AccI site at position 1201 to the BamHI site at position 1851. B, An RNA gel blot consisting of alternating lanes of poly(A+) RNA from stably transformed tobacco cells (left lane) or untransformed cells (right lane) was sectioned and hybridized with a probe for the entire coding region or with the individual segments of the B.t.-toxin-coding region (designated below each panel). The arrowheads indicate the full-length transcript, and the dark circles indicate the 900- and 600-nucleotide transcripts. The box diagrams indicate the segments of the B.t.-toxin-coding region found in each transcript.

RNA probes corresponding to the entire E9 3 UTR or the AccI/BamHI segment of the B.t.-toxin-coding region (segment 4) were in vitro transcribed using a riboprobe system (Promega) from the linearized Bluescript II SK() plasmids p1425 and p1522, which contain only the E9 and segment 4 sequences, respectively. [-³²P]UTP was used in the labeling reaction and the probes were purified with push columns (Stratagene). Prehybridization of the blots was performed as described above; however, hybridization with the riboprobe was done overnight at 65°C in hybridization solution (described by De Rocher et al., 1998), which was modified by increasing the formaldehyde concentration to 50% (v/v) and decreasing the SSC concentration to 1×. Blots were washed as described above but with an additional wash at 0.2× SSC, 0.1% (w/v) SDS.

Transcripts synthesized in vitro for RNase-H-cleavage experiments were generated from approximately 0.5 µg of p995 that was linearized with AccI. The reaction was assembled at room temperature and consisted of 40 mM Tris-HCl, pH 8.0, 19 mM MgCl₂, 5 mM DTT, 2 mM spermidine, 0.01% Triton X-100, 4 mM each ATP, CTP, UTP, and GTP, 16 units of RNAsin (Promega), 1 unit of yeast pyrophosphatase, and 200 units of T7 RNA polymerase (GIBCO-BRL) in a 50-mL volume. After 6 h of incubation at 37°C, 1 µL of RNase-free DNase I (GIBCO-BRL) was added, and the reaction was incubated for another 15 min at 37°C. Sterile distilled water at a volume of 130 µL was added followed by 20 µL of 5 M NH₂OAc/100 mM EDTA. The RNA was then phenol/chloroform extracted and precipitated. The reaction resulted in a 1201-nucleotide in vitro transcript containing segments 1, 2, and 3. Fifty picograms of the transcript was added to 20 µg of total RNA from untransformed BY-2 cells before electrophoresis through an RNA gel.

RT-PCR Analysis

1

1

1

The reverse primer, PG-192 (5-GGCCACGCGTCGACTAGTAC-3), which anneals to the adapter region of the adaptor primer used to generate the cDNA, was also added to a final concentration of 200 nM. The PCRs of samples without prior cDNA synthesis, or the size of the amplified products in combination with the presence of a poly(A⁺) tail at the 3 end of the cDNA clone, was used to verify that genomic DNA was not amplified. The amplification protocol was for 5 min at 94°C followed by 30 cycles of 2 min at 94°C, 2 min at 55°C (for PG-177/PG-192), or 2 min at 65°C (for PG-170/PG-192) and 3 min at 72°C. A 15-min incubation at 72°C completed the amplification. The B.t.-toxin PCR products were digested with XbaI and SalI, whereas the globin-B.t. PCR products were digested with BamHI and SalI. All of the PCR products were cloned into p948, the Bluescript II SK() vector described above. Four positive clones were identified by gel electrophoresis or by Southern-blot analysis using a DNA probe consisting of the cryIA(c) B.t.-toxin-coding region. Cycle sequence analysis of the clones identified the poly(A⁺) addition sites (DNA Sequencing Facility, Michigan State University).

Oligo-Directed RNase-H-Cleavage Analysis

1218

1218

RNase-H-cleavage experiments were also performed using 3.5 µg of poly(A⁺) RNA from untreated tobacco cells annealed to 0.3 µg of either PG-229 or PG-234. The hybrids were incubated in 20 mM Tris-HCl, pH 7.5, 10 mM MgCl₂, 100 mM KCl, and 5% (w/v) Suc, with 3 units of RNase-H for 60 min at 37°C. All samples were alcohol precipitated and resuspended in formaldehyde loading buffer before electrophoresis through a 2% (v/v) formaldehyde/1.7% (w/v) agarose gel in 1× Mops buffer (described above). The gel was blotted and the membrane probed with the full-length B.t.-toxin-coding region as described above.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Low Accumulation of B.t.-toxin Transcripts and the Detection of Short, Polyadenylated Transcripts in Tobacco Cells

View larger version (11K): [in this window] [in a new window]   Figure 1. Structure of the genes stably introduced into tobacco cells. A, Portion of the wild-type B.t. subsp. kurstaki cryIA(c) gene encoding the insecticidal domain (amino acids 9-613; Schnepf and Whiteley 1985) used in this study. Transcription of the chimeric gene was controlled by the CaMV 35S promoter modified by duplicating the upstream enhancer region (2X35S). The E9 3 UTR provides the elements necessary for polyadenylation. B, Chimeric globin-B.t.-toxin gene used to identify the polyadenylation site in segment 4 of the cryIA(c) B.t.-toxin-coding region. A 650-bp AccI-BamHI restriction fragment of the B.t.-toxin-coding region (segment 4) was inserted between a -globin reporter gene under the control of the 2X35S and the E9 3 UTR. Use of any poly(A+) addition sites within the segment 4 insert will result in polyadenylated transcripts that lack the E9 3 UTR sequences.

Tobacco cells stably transformed with this B.t.-toxin derivative were analyzed for expression on RNA gel blots. However, as shown in the example in Figure 2A, lane 1, little or no accumulation of the full-length B.t.-toxin transcript (i.e. transcripts approximately 2300 nucleotides in length terminating in the E9 3 UTR) could be observed in total RNA samples. The full-length transcript could not be visualized in most of the transformed cell lines that were examined, presumably because the transcript was below the level of detection (data not shown). This is despite the fact that transcription is directed by the 2X35S. Run-on transcription experiments have shown that the gene is efficiently transcribed in nuclei isolated from stably transformed cells, arguing against the possibility that the B.t.-toxin-coding region contains a repressor sequence capable of inhibiting transcription initiation or elongation (De Rocher et al., 1998). The discrepancy between the RNA-gel-blot and the run-on-transcription results suggests that the mechanisms responsible for the low abundance of the B.t.-toxin transcript in plant cells are posttranscriptional.

View larger version (40K): [in this window] [in a new window]   Figure 2. Poor accumulation of B.t.-toxin mRNA and the detection of short, polyadenylated transcripts in tobacco cells. A, Total RNA (T) from stably transformed tobacco cells was isolated from a pool of kanamycin-resistant calli growing in liquid culture and electrophoresed next to poly(A+) RNA (A) isolated from the same cells. Cell cultures were either treated (+) or not treated () with CHX. The RNA gel blot was probed with a 730-bp DNA fragment corresponding to the 5 portion of the B.t.-toxin-coding region. The autoradiograph was overexposed to show the B.t.-toxin transcripts in tobacco cells not treated with CHX. B, Detection of poly(A+) tails on the B.t.-toxin transcripts. Total RNA was isolated from two different CHX-treated tobacco cell lines (A and B) expressing the B.t.-toxin gene. Both cell lines accumulate the 900- and 600-nucleotide (nt) B.t.-toxin transcripts, but only one cell line (A) accumulates the full-length B.t.-toxin transcript. RNA from each line was incubated with RNase-H, oligo(dT)12, or both, and electrophoresed next to a 1227-nucleotide in vitro-transcribed B.t.-toxin transcript that was incubated in the presence or absence of RNase-H. Hybridization of oligo(dT)12 to the poly(A+) tail of a transcript results in cleavage of the poly(A+) tail by RNase-H. As a consequence, deadenylated transcripts will have an increased mobility in an RNA gel. The RNA gel blot was hybridized with a probe for the B.t.-toxin-coding region. Lane 9 contains total RNA from untransformed BY-2 cells that was used as a carrier for the in vitro-transcribed transcript.

Figure 2A, lane 2, shows the poly(A⁺) RNA fraction from the same cell line shown in Figure 2A, lane 1. The full-length B.t.-toxin transcript can be easily detected in this fraction, indicating that it is polyadenylated, as expected. However, two short transcripts of 900 and 600 nucleotides can also be detected. The abundance of these two short transcripts were significantly increased by treating the cell line with the translation inhibitor CHX (compare Figure 2A, lanes 1 and 3 and lanes 2 and 4). CHX can increase the abundance of unstable transcripts because many of these transcripts are degraded in a translation-dependent manner. Nearly all of the stably transformed cell lines analyzed accumulated the two short transcripts in the presence of CHX, even when the full-length B.t.-toxin transcript was below the limit of detection (Fig. 2B, lanes 4-6). This is consistent with the rapid degradation of the 900- and 600-nucleotide transcripts in plant cells. The instability of cryIA(c) B.t.-toxin transcripts is the focus of the accompanying paper (De Rocher et al., 1998).

Characterization of the Short B.t.-Toxin Transcripts

The existence of a poly(A⁺) tail on the short transcripts was verified using oligo-directed RNase-H cleavage, as shown in Figure 2B. RNase-H cleaved the RNA strand of an RNA/DNA hybrid. Therefore, annealing oligo(dT) to the poly(A⁺) tail will result in cleavage of the tail by RNase-H. The removal of the poly(A⁺) tail can be detected by an increased mobility of the transcripts on an RNA gel blot. As shown in Figure 2B, lanes 2 and 5, the 900- and 600-nucleotide transcripts from two different cell lines decreased in size upon treatment with oligo(dT) and RNase-H. This is consistent with the removal of the poly(A⁺) tail. The migration of these transcripts was unaltered when RNase-H or oligo(dT) was not present in the reactions (Fig. 2B, lanes 1 and 3 and lanes 4 and 6). The full-length B.t.-toxin transcript also had an increased mobility in the presence of RNase-H and oligo(dT), confirming that it is polyadenylated (Fig. 2B, lane 2). An in vitro-synthesized cryIA(c) B.t.-toxin transcript corresponding to the first 1201 nucleotides (segments 1, 2, and 3; see below) was also incubated with oligo(dT) in the presence or absence of RNase-H (Fig. 2B, lanes 7 and 8). The migration of this transcript was unaltered in the presence of RNase-H, indicating that oligo(dT) did not anneal nonspecifically to the 900- and 600-nucleotide transcripts or to other sequences within the first 1201 nucleotides of the full-length transcript.

The RNA gel blot in Figure 2A was probed with the first 730 nucleotides of the 1850-nucleotide B.t.-toxin-coding region. Detection of the 900- and 600-nucleotide transcripts with this probe and the fact that these transcripts are polyadenylated argues against their resulting from degradation of the full-length transcript. Instead, the data support splicing or polyadenylation within the coding region as the mechanism responsible for the formation of the short transcripts.

To further delineate which sequences were present in the 900- and 600-nucleotide transcripts, the B.t.-toxin-coding region was divided into four segments using convenient restriction sites, as shown in Figure 3A. Each segment was then used to generate probes to hybridize against poly(A⁺) RNA isolated from CHX-treated cells producing the full-length transcript and both short transcripts (Fig. 3B). All of the probes hybridized to the full-length transcript, as expected. However, the 600-nucleotide transcript hybridized to only the segment 1 and 2 probes and the 900-nucleotide transcript hybridized to the segment 1, 2, and 3 probes. Hybridization was detected in the 900-nucleotide region with the segment 4 probe. However, this was also evident in the untransformed control and, therefore, does not correspond to the 900-nucleotide B.t.-toxin transcript. Neither the 600- nor the 900-nucleotide transcript hybridized with the probe spanning the E9 3 UTR (Fig. 3B, last panel), indicating that the poly(A⁺) tail is attached directly to the B.t.-toxin-coding region. The simplest explanation of these data is that sequences within the cryIA(c) B.t.-toxin-coding region are recognized as polyadenylation signals. The 600-nucleotide transcript is consistent with polyadenylation within segment 2 of the coding region, and the 900-nucleotide transcript is consistent with polyadenylation in segment 3. Other minor transcripts can be observed hybridizing to some of the B.t.-toxin probes (Fig. 2A); however, unlike the 900- and 600-nucleotide transcripts, these transcripts were not consistently found in the stably transformed cell lines. Thus, these transcripts were not pursued further.

Identification of Polyadenylation Sites within the B.t.-Toxin-Coding Region

View larger version (20K): [in this window] [in a new window]   Figure 4. Identification of three poly(A+) addition sites within the B.t.-toxin-coding region. cDNA was synthesized from total RNA isolated from stably transformed tobacco cells and amplified by PCR. The resulting PCR products were cloned and sequenced to determine the polyadenylation sites that are indicated.

The difference in size between the 900- and 600-nucleotide transcripts and that expected based on the mapped poly(A⁺) addition sites is likely due to the length of the poly(A⁺) tail. To test this possibility and to confirm that the segment 3 polyadenylation site identified by RT-PCR corresponds to the same polyadenylation site used to generate the 900-nucleotide in vivo transcript, oligo-directed RNase-H-cleavage analysis was performed. A DNA oligonucleotide hybridizing to an mRNA upstream of the poly(A⁺) site should direct cleavage of that RNA by RNase-H, whereas an oligonucleotide hybridizing downstream should not. As shown in Figure 5A, an oligonucleotide that hybridizes starting 65 bases upstream of the segment 3 polyadenylation site directs cleavage of the full-length and 900-nucleotide transcripts in the presence of RNase-H. The mobility of the 600-nucleotide transcript, which lacks sequences complementary to the oligonucleotide, was unaltered in the presence of RNase-H, as expected.

View larger version (47K): [in this window] [in a new window]   Figure 5. The mapped poly(A+) addition site in segment 3 corresponds to the polyadenylation site of the 900-nucleotide (nt) in vivo transcript. Total RNA from CHX-treated tobacco cells expressing the B.t.-toxin gene was incubated with oligonucleotides that hybridize either upstream (A) or downstream (B) of the segment 3 polyadenylation site identified by RT-PCR. Poly(A+) RNA from transgenic tobacco cells not treated with CHX was also incubated with the upstream and downstream oligonucleotides (C). Incubations were performed in the presence (+) or absence () of RNase-H. If sequences complementary to the oligonucleotides are present in the 900-nucleotide transcript, RNase-H will cleave the RNA strand of the RNA/DNA duplex, resulting in a band shift on the RNA gel blots. The bracketed bands, which may correspond to the 3 ends of the full-length transcript, were detected only when poly(A+) RNA was digested with RNase-H.

The 180-base decrease in the size of the 900-nucleotide transcript to approximately 720 nucleotides in the presence of RNase-H (Fig. 5A) indicates that the poly(A⁺) tail is about 115 bases long. Sixty-five bases of the 180-base difference are accounted for by the sequences to which the oligonucleotide anneals, as well as the distance between the oligonucleotide and the poly(A⁺) addition site. A poly(A⁺) tail length of 115 bases corresponds well with the 110- to 120-base poly(A⁺) tail predicted from the discrepancy between the position of the segment 3 polyadenylation site and the in vivo size of the transcript.

Similar RNase-H-cleavage experiments were carried out using an oligonucleotide that hybridizes starting approximately 10 bases downstream of the polyadenylation site mapped in segment 3 (Fig. 5B). This oligonucleotide should not anneal to the 900-nucleotide transcript unless it actually extends beyond the mapped poly(A⁺) site. Figure 5B shows that the size of the 900-nucleotide transcript was not altered after incubation with the oligonucleotide and RNase-H, demonstrating that sequences immediately downstream of the segment 3 polyadenylation site are not present in the 900-nucleotide transcript.

Cleavage experiments using poly(A⁺) RNA from untreated tobacco cells instead of total RNA from CHX-treated cells show the same results with both the upstream and downstream oligonucleotides (see Fig. 5C). These data indicate that the segment 3 polyadenylation site identified by RT-PCR is the same site used to generate the 900-nucleotide transcript. More importantly, they also indicate that the same polyadenylation site is used in both CHX-treated and untreated transformed tobacco cells. RNase-H-cleavage experiments were also performed to confirm the segment 2 polyadenylation sites. However, the diffuse nature of the 600-nucleotide transcript made it difficult to assess precisely the shift in bands (data not shown). Yet, an oligonucleotide complementary to a region upstream of the cleavage site was able to shift the 600-nucleotide transcript to a smaller size, whereas an oligonucleotide hybridizing downstream of the cleavage sites did not appear to decrease the size of the transcript (data not shown).

Sequences Typical of Plant Polyadenylation Signals Are Present Upstream of the Identified Poly(A⁺) Sites

View larger version (34K): [in this window] [in a new window]   Figure 6. The B.t.-toxin-coding region contains elements characteristic of plant-polyadenylation signals. A, Schematic representation comparing the structure of typical plant and mammalian polyadenylation signals. Sequence motifs characteristic of the plant FUE and NUE, as well as the mammalian downstream element (DSE), are indicated. The poly(A+) addition sites are represented by arrows. Plants can use multiple poly(A+) addition sites downstream of specific NUEs within a transcript. The cleavage of a plant transcript usually occurs at a Py/A dinucleotide (YA). Mammalian transcripts are usually cleaved at a single site corresponding to a C/A dinucleotide (CA). B, Identification of elements characteristic of plant polyadenylation signals upstream of the poly(A+) addition sites in the B.t.-toxin-coding region. The putative plant polyadenylation signals in segments 2, 3, and 4 of the B.t.-toxin-coding region were compared with the most commonly used polyadenylation sites in the rbcS-E9 and octopine synthase (OCS) genes. The positions of the FUE and NUE relative to the cleavage site (CS) are indicated.

Located approximately 40 to 150 bases upstream of the cleavage site, the FUE is required for efficient 3 end formation. The most common motif among known FUEs is the presence of multiple U/G-rich regions (for review, see Hunt, 1994; Wu et al., 1995; Rothnie, 1996). Several U/G-rich stretches can be found in the B.t.-toxin-coding region upstream of the segment 2 and segment 3 poly(A⁺) addition sites in positions that correspond to a putative FUE (Fig. 6B).

The NUE is an A/U-rich element typically found 10 to 30 bases upstream of the cleavage site. These elements are essential for polyadenylation and control poly(A⁺) addition at specific cleavage sites. Therefore, a plant transcript with multiple polyadenylation sites will have a NUE corresponding to each site. NUEs can contain the mammalian canonical AAUAAA sequence, as in the CaMV polyadenylation signal (Sanfacon et al., 1991; Rothnie et al., 1994), but more often an AAUAAA-like sequence is present in which one or two of the bases do not match (for review, see Hunt, 1994; Wu et al., 1995; Rothnie, 1996). Upstream of the three poly(A⁺) addition sites in the B.t.-toxin-coding region, an AAUAAA-like sequence typical of plant polyadenylation signals can be identified. A comparison of these sequences with other known NUEs revealed sequence similarity (Fig. 6B).

Cleavage of the B.t.-toxin transcript in both segments 2 and 3 occurred at Py/A dinucleotides (Fig. 6B). This is consistent with other known plant poly(A⁺) addition sites. Again, no strict consensus sequence is known that defines the cleavage site in plants. However, cleavage typically occurs at a Py/A dinucleotide in plant transcripts (for review, see Hunt, 1994; Wu et al., 1995; Rothnie, 1996).

Further sequence analysis of the B.t.-toxin gene revealed the presence of other possible plant polyadenylation signals in the coding region. In particular, segment 4 (Fig. 3A) contains sequences that resemble plant-polyadenylation signals. Transcripts terminating in this segment were not detected on RNA gel blots or by RT-PCR using total RNA from CHX-treated cells as the template. However, these transcripts may be very unstable or may be produced in very small amounts if a substantial portion of the transcripts is polyadenylated in segments 2 and 3. To determine whether segment 4 of the B.t.-toxin-coding region contains a functional plant polyadenylation signal, a chimeric gene was constructed consisting of the segment inserted between a 2X35S-driven -globin reporter gene and the E9 3 UTR (Fig. 1B). A polyadenylated transcript terminating in segment 4 would result if a polyadenylation signal exists in the segment; otherwise, the transcript would be polyadenylated at the E9 poly(A⁺) sites. The RNA gel blot in Figure 7A shows that stably transformed tobacco cells expressing the globin-B.t. gene accumulated only a small amount of transcript at a position consistent with termination in the E9 region. Most of the globin-B.t. transcripts in these cells accumulated as discrete bands at sizes more consistent with termination in segment 4. Hybridization with the E9 3 UTR showed that these abundant transcripts lack the E9 region (Fig. 7A). Similar transcript patterns were reproducibly observed in transgenic tobacco plants and in protoplasts transiently expressing the gene (data not shown).

View larger version (48K): [in this window] [in a new window]   Figure 7. Identification of a fourth polyadenylation site in the B.t.-toxin-coding region. A, Truncated transcripts are produced in tobacco cells expressing a chimeric globin-B.t.-toxin gene containing nucleotides 1201 to 1851 of the B.t.-toxin-coding region. An RNA gel blot of total RNA extracted from two pools of 100 tobacco calli stably transformed with the globin-segment 4 chimeric gene (4) (p1194) described in ``Materials and Methods'' was probed with either the globin-coding region (left panel) or the E9 3 UTR (right panel). Total RNA from tobacco calli expressing a gene containing the globin-coding region and rbcS-E9 3 UTR (p1528) described in "Materials and Methods," but lacking the segment 4 insert, was used as a control (0). The arrows indicate the positions of the transcripts terminating in the E9 3 UTR for the control and globin-B.t.-toxin genes. The bracket marks the position of the globin-B.t.-toxin transcripts polyadenylated in the segment 4 insert. The autoradiographs were overexposed to show the globin-B.t.-toxin transcripts terminating in the E9 3 UTR. B, The short globin-B.t.-toxin chimeric transcripts are polyadenylated. Total RNA isolated from transgenic tobacco cells expressing the control gene (0) or the globin-segment 4 chimeric gene (4) was incubated with oligo(dT)12-18 in the presence (+) or absence () of RNase-H. After electrophoresis, the RNA was blotted and probed with the globin-coding region. The arrows show the positions of the transcripts terminating in the E9 3 UTR for the control and globin-B.t.-toxin transcripts. The bracket indicates the position of the globin-B.t.-toxin transcripts polyadenylated in the segment 4 insert.

The presence of a poly(A⁺) tail on the short transcripts was determined using RNase-H and oligo(dT). As shown in Figure 7B, the short transcripts decreased in size in the presence of RNase-H, indicating that they are polyadenylated. These data are supported by RT-PCR analysis, which identified a poly(A⁺) addition site 201 bases into segment 4. This site is located 13 bases downstream of a putative NUE in segment 4 (Fig. 6B). Taken together, these data show that segment 4 of the B.t.-toxin-coding region does contain sequences that function as polyadenylation signals in plants. How much polyadenylation at this site contributes to the poor accumulation of the full-length B.t.-toxin mRNA is not known. We cannot rule out the possibility that our globin-B.t. construct activates a cryptic polyadenylation site within segment 4 (Luehrsen and Walbot, 1994). However, when a chimeric B.t.-toxin gene was created by fusing synthetic plant-like versions of segments 1, 2, and 3 to a wild-type version of segment 4, most of the resulting mRNA was polyadenylated in segment 4, rather than at the E9 poly(A⁺) site located downstream (S.H. Diehn, W.-L. Chiu, E.J. De Rocher, and P.J. Green, unpublished results). This suggests that the segment 4 poly(A⁺) site does function within a B.t.-toxin gene context.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The goal of this study and the one that follows (De Rocher et al., 1998) was to determine what processes play a role in limiting the accumulation of the cryIA(c) B.t.-toxin transcript in plants. Elucidating the mechanisms responsible for the low accumulation of this transcript may make it easier to resynthesize novel B.t.-toxin genes, but more importantly, it may provide an understanding of why the transcripts of some foreign genes fail to accumulate in plants. In this report we have demonstrated that the cryIA(c) B.t.-toxin-coding region contains multiple sequence elements that are recognized by plant cells as polyadenylation signals. We suggest that use of these polyadenylation signals appears to be at least partially responsible for the low accumulation of the cryIA(c) B.t.-toxin transcript in plants.

To the best of our knowledge, this study is the first to show that sequences within the coding region of a foreign gene can be recognized as polyadenylation signals by plants. It had been observed previously that tobacco plants expressing a cryIA(c) protoxin gene or a 3-truncated version produced a polyadenylated 1.7-kb transcript. The size of this transcript was shorter than expected for either gene, prompting the suggestion that the B.t.-toxin-coding region contains plant polyadenylation signals (Adang et al., 1987). However, the transcript disappeared as the plants matured (Murray et al., 1991); therefore, the mechanism responsible for the production of the 1.7-kb transcript could not be determined, and the disappearance of the transcript during development was not explained. In another study, transcripts of 1.6 and 0.9 kb were detected in the poly(A⁺) RNA fractions of plants expressing a cryIA(b) B.t.-toxin gene. However, these transcripts were believed to be degradation intermediates (Murray et al., 1991).

In this study nearly every tobacco cell line stably transformed with the cryIA(c) B.t.-toxin gene accumulated polyadenylated transcripts about 900 and 600 nucleotides in length. Hybridization, RT-PCR, and RNase-H-mapping experiments all confirmed that these short transcripts were the result of polyadenylation at two nearby sites in segment 2 and at another site in segment 3. Although the short transcripts could not be detected in total RNA preparations, the abundance of these transcripts could be increased by treating the cells with CHX to a point at which they were the most abundant B.t.-toxin transcripts in both the total and poly(A⁺) RNA fractions. In addition, these transcripts accumulated to detectable levels in the poly(A⁺) RNA fractions of untreated tobacco cells. As discussed in the accompanying paper (De Rocher et al., 1998), the full-length and short B.t.-toxin transcripts are rapidly degraded in tobacco. Although they are unstable, we suggest that the short B.t.-toxin transcripts are produced in significant amounts and thereby contribute to the low accumulation of the full-length mRNA in plants.

The high A/T content of B.t.-toxin genes raises the possibility that regions of the cryIA(c) B.t.-toxin mRNA are recognized as introns in plant cells. The presence of spliced B.t.-toxin transcripts was recently reported in tobacco cells expressing a cryIA(b) gene (Van Aarssen et al., 1995). Our results demonstrate that the 900- and 600-nucleotide transcripts are not a result of splicing. The hybridization data, the RT-PCR analysis, and the RNase-H experiments were consistent with the conclusion that these transcripts are a result of polyadenylation in the B.t.-toxin-coding region. This is not to suggest that splicing of the cryIA(c) B.t.-toxin transcript does not occur in tobacco cells. Splicing could account for some of the minor transcripts hybridizing to the various B.t.-toxin probes observed on our RNA gel blots. However, most of these transcripts were not reproducible in our transformed tobacco cell lines and, therefore, were not pursued further.

The presence of plant poly(A⁺) addition sites within the cryIA(c) B.t.-toxin-coding region raises the question of whether other closely related B.t.-toxin genes might contain plant polyadenylation signals within their coding regions. The cryIA(c) gene belongs to the cryI class, one of six classes of B.t.-toxin genes (see Höfte and Whiteley, 1989; Feitelson et al., 1992, for a detailed description of the different classes). These genes share significant nucleotide sequence identity with each other and encode insecticidal proteins that are active against the insect order Lepidoptera. A subclass of the cryI genes, the cryIA genes, contains members designated cryIA(a), cryIA(b), cryIA(c), and cryIA(d), which are more than 80% identical at the nucleotide level. Alignment of the cryIA(c)- and cryIA(b)-coding regions shows that a region extending 230 bp upstream from the second segment 2 polyadenylation site (position 509) is identical between the two genes (data not shown). This 230-bp region should be of sufficient length to contain the elements necessary for polyadenylation in plants. A similar alignment with the cryIA(a) gene shows that the same segment 2 poly(A⁺) signals are probably common to this gene as well (data not shown).

The cryIA(d)-coding region has 95% nucleotide identity over a region spanning the putative segment 2 polyadenylation signals (data not shown). Approximately the same sequence identity over the region containing the segment 3 polyadenylation signal can be observed for the cryIA(a), cryIA(b), and cryIA(d) genes. One would expect, therefore, that the same poly(A⁺) sites are used in the cryIA(a-d)-coding regions, although this remains to be proven. Other B.t.-toxin genes, such as cryIE(a), cryIF, cryID, and Prt A, share lower levels of nucleotide identity in the region of the poly(A⁺) signals (e.g. 57%-87% for the segment 2 poly[A⁺] signals), which could affect poly(A⁺) signal recognition, so it is difficult to predict if the same poly(A⁺) addition sites are used in these genes.

B.t.-toxin transcripts are not likely to be the only foreign transcripts that are prematurely polyadenylated in plants. Although there are no other documented cases yet to our knowledge, other examples are expected to arise as the expression of other problematic foreign genes is investigated. This contention is supported by the finding that even a transcript normally produced in one plant species can be differentially polyadenylated when it is transcribed in another. Specifically, the maize activator (Ac) transposase transcript is polyadenylated at four sites within a 200-bp region of exon 2 when it is expressed in Arabidopsis plants (Jarvis et al., 1997; Martin et al., 1997). Recognition of these poly(A⁺) addition sites has been suggested to contribute to the low abundance of correctly processed transposase transcripts and hence the low frequency of transposition in this plant species. The low accumulation of the T4 lysozyme and Klebsiella pneumoniae cyclodextrin glycosyltransferase transcripts in potato plants also may be a result of premature poly(A⁺) addition sites. Like B.t.-toxin genes, these genes have a high A/U bias. Currently, it is not possible to predict which putative polyadenylation signals will be recognized in plants strictly on the basis of sequence analysis. Nevertheless, as more poly(A⁺) signals are scrutinized and the mechanisms by which they are recognized are elucidated, designing an algorithm to achieve this goal may indeed be feasible.

	FOOTNOTES

   Received November 11, 1997; accepted May 9, 1998.

	ABBREVIATIONS

Abbreviations: 2X35S, doubly enhanced cauliflower mosaic virus 35S promoter. BY-2, Bright Yellow 2. CaMV, cauliflower mosaic virus. CHX, cycloheximide. FUE, far-upstream element. GFP, green fluorescent protein. NUE, near-upstream element. Py, pyrimidine. RT, reverse transcriptase. UTR, untranslated region.

	ACKNOWLEDGMENTS

We thank Dr. Ambro van Hoof and Dr. Dan Vernon for their comments on the manuscript. We are also grateful to Dr. Pedro Gil for the construction of plasmid p995, Dr. Christie Howard for the construction of p1425, Marlene Cameron for computer graphics, and Kurt Stepnitz for photographic services.

	LITERATURE  CITED

Top Abstract Introduction Methods Results Discussion References

Adang MJ, Firoozabady E, Klein J, DeBoer D, Sekar V, Kemp JD, Murray E, Rocheleau TA, Rashka K, Staffeld G, and others (1987) Expression of a Bacillus thuringiensis insecticidal crystal protein gene in tobacco plants. In CJ Arntzen, C Ryan, eds, Molecular Strategies for Crop Protection, Vol 46. Alan R. Liss, New York, pp 345-353

Adang MJ, Staver MJ, Rocheleau TA, Leighton J, Barker RF, Thompson DV (1985) Characterized full-length and truncated plasmid clones of the crystal protein of Bacillus thuringiensis subsp. kurstaki HD-73 and their toxicity to Manduca sexta. Gene 36: 289-300 [CrossRef][ISI][Medline]

An G (1985) High-efficiency transformation of cultured tobacco cells. Plant Physiol 79: 568-570 [Abstract/Free Full Text]

Aronson AI (1993) The two faces of Bacillus thuringiensis: insecticidal proteins and post-exponential survival. Mol Microbiol 7: 489-496 [CrossRef][Medline]

Barton KA, Whiteley HR, Yang N-S (1987) Bacillus thuringiensis -endotoxin expressed in transgenic Nicotiana tabacum provides resistance to lepidopteran insects. Plant Physiol 85: 1103-1109 [Abstract/Free Full Text]

Bigelow CC, Channon M (1982) In GD Fasman, ed. CRC Handbook of Biochemistry and Molecular Biology, Ed 3: Proteins, Vol 1. CRC Press, Boca Raton, FL, pp 209-224

Cheng J, Bolyard MG, Saxena RC, Sticklen MB (1992) Production of insect resistant potato by genetic transformation with a -endotoxin gene from Bacillus thuringiensis var. kurstaki. Plant Sci 81: 83-91 [CrossRef]

Dandekar AM, McGranahan GH, Vail PV, Uratsu SL, Leslie C, Tebbets JS (1994) Low levels of expression of wild type Bacillus thuringiensis var. kurstaki cryIA(c) sequences in transgenic walnut somatic embryos. Plant Sci 96: 151-162 [CrossRef]

Delannay X, LaVallee BJ, Proksch RK, Fuchs RL, Sims SR, Greenplate JT, Marrone PG, Dodson RB, Augustine JJ, Layton JG, and others (1989) Field performance of transgenic tomato plants expressing the Bacillus thuringiensis var. kurstaki insect control protein. Bio/Technology 7: 1265-1269

De Rocher EJ, Vargo-Gogola TC, Diehn SH, Green PJ (1998) Direct evidence for rapid degradation of Bacillus thuringiensis toxin mRNA as a cause of poor expression in plants. Plant Physiol 117: 1445-1461 [Abstract/Free Full Text]

Diehn SH, De Rocher EJ, Green PJ (1996) Problems that can limit the expression of foreign genes in plants: lessons to be learned from B.t.-toxin genes. In JK Setlow, ed, Genetic Engineering: Principles and Methods, Vol 18. Plenum Press, New York, pp 83-99

Düring K (1988) Wundinduzierbare Expression und Sekretion von T4 Lysozym und monoklonalen Antikörpern in Nicotiana tabacum. PhD thesis. University of Cologne, Cologne, Germany

Düring K, Porsch P, Fladung M, Lörz H (1993) Transgenic potato plants resistant to the phytopathogenic bacterium Erwinia carotovora. Plant J 3: 587-598

Feinberg AP, Vogelstein B (1983) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity (addendum). Anal Biochem 137: 266

Feitelson JS, Payne J, Kim L (1992) Bacillus thuringiensis: insects and beyond. Bio/Technology 10: 271-275 [CrossRef]

Fischhoff DA, Bowdish KS, Perlak FJ, Marrone PG, McCormick SM, Niedermeyer JG, Dean DA, Kusano-Kretzmer K, Mayer EJ, Rochester DE, and others (1987) Insect tolerant transgenic tomato plants. Bio/Technology 5: 807-813 [CrossRef]

Goff SA, Cone KC, Fromm ME (1991) Identification of functional domains in the maize transcriptional activator C1: comparison of wild-type and dominant inhibitor proteins. Genes Dev 5: 298-309 [Abstract/Free Full Text]

Haseloff J, Amos B (1995) GFP in plants. Trends Genet 11: 328-329 [CrossRef][ISI][Medline]

Haseloff J, Siemering KR, Prasher DC, Hodge S (1997) Removal of a cryptic intron and subcellular localization of green fluorescent protein are required to mark transgenic Arabidopsis plants brightly. Proc Natl Acad Sci USA 94: 2122-2127 [Abstract/Free Full Text]

Höfte H, Whiteley HR (1989) Insecticidal crystal proteins of Bacillus thuringiensis. Microbiol Rev 53: 242-255 [Abstract/Free Full Text]

Hunt AG (1994) Messenger RNA 3 end formation in plants. Annu Rev Plant Physiol Plant Mol Biol 45: 47-60