Restriction endonucleases and Pfu polymerase were purchased from New England Biolabs and Stratagene, respectively. Sp6 and T7 RNA polymerases, RNasin, RNase-free DNase I, Luciferase Reporter Assay Systems, pGEM-T easy and pSP64 poly(A) vectors were from Promega. RNeasy kit and Oligotex mRNA Mini kit were from Qiagen. MAXIscript in vitro transcription kit and RPA III Ribonuclease Protection Assay Kit were from Ambion. m⁷GpppG cap analog and [α-³²P]dCTP were from Amersham/Pharmacia. The Sephadex G-25 Quick Spin Columns (TE) for radio-labeled DNA and RNA purification were from Roche Diagnostics. PCR Cloning kit, cell culture media and sera, Lipofectamine Plus and Lipofectin transfection reagents were all from Invitrogen. All other reagents were of molecular biology grade from Fisher or Sigma.

The full-length sequence of 5′-UTR of human p27 is 575 bases (9). The full-length and the two deletion mutants of 5′-UTR sequences of human p27 were amplified from human genomic DNA using a common reverse primer 5′-GGCCATGGCTTTCTCCCGGGTCTGCACGA-3′ and the following individual forward primers: 5′-GGACTAGTCCACCTTAAGGCCGCGCTCGC-3′ for full-length (−575 to −1) 5′-UTR; 5′-GGACTAGTAGCCTCCCTTCCACCGCCATA-3′ for deletion construct (−461 to −1); and 5′-GGACTAGTGGCCGTGGCTCGTCGGGGTCT-3′ for deletion construct (−150 to −1). These PCR products were then cloned into pGEM-T Easy vector.

To engineer dicistronic constructs containing different 5′-UTRs of human p27, the dicistronic vector pRF that contains Renilla and firefly luciferase genes (18) was used. The full-length and the two deletion constructs of human p27 5′-UTR in pGEM-T Easy were released by digestion with SpeI and NcoI and subsequently cloned into pRF vector to obtain pR-p27-F(−575), pR-p27-F(−461) and pR-p27-F(−150). The promoterless dicistronic constructs containing these 5′-UTRs were engineered by removing the vector SV40 promoter together with the chimeric intron as described previously (19). The hairpin structure in the dicistronic constructs were engineered by cloning the 5′-UTR sequences into the dicistronic constructs containing a hairpin prepared in a previous study (20). To generate constructs that can be used for in vitro production of transcripts with poly(A) tail, the 5′-UTR sequences in the dicistronic constructs were released by digestion with SpeI and NcoI and subsequently cloned into the pSP-R-HRV-F_A30 plasmid (19) by replacing the HRV IRES sequence digested with SpeI and NcoI.

In vitro transcription was performed as described previously (21). Briefly, pRFA₃₀-based plasmids were linearized by EcoRI and the capped transcripts were synthesized in vitro in the presence of 1 mM m⁷GpppG. The DNA template was then removed by digestion using RNase-free DNase I, and the RNA transcripts were purified using a Qiagen RNeasy Mini kit.

His-tagged PTB was expressed in Escherichia coli BL21(DE3) cells using the pET28aPTB construct (22) and purified using Ni-NTA Sepharose (Qiagen) according to the supplier's protocol. The purified His-PTB was extensively dialyzed against H100 buffer (10 mM Tris–HCl, pH 7.4, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT, 0.1 mM phenylmethylsulfonyl fluoride and 10 μg/ml leupeptin) prior to use in the in vitro translation system.

In vitro translation was performed as described previously (19,20,23). Briefly, 100 ng of capped dicistronic RNA transcripts were used to program cell-free translation in rabbit reticulocyte lysate (RRL) in a final volume of 10 μl containing 6.5 μl RRL. The translation mixture contained either 1 μl of H100 buffer (control) or 1 μl of purified His-PTB at 25 μg/ml.

Human large cell lung cancer cell line H1299 and human embryonic kidney 293T cells were maintained in RPMI 1640 and DMEM media, respectively, supplemented with 10% fetal bovine serum in humidified incubators at 37°C with 5% CO₂.

Transient transfection with DNA constructs were performed with Lipofectamine plus reagents as described previously (19,20,23). Briefly, ∼6 × 10⁴ cells/well in a 24-well plate were co-transfected with 0.4 μg of test DNA and 50 ng of β-galactosidase control DNA. Cells were then harvested 24 h post-transfection followed by analysis of luciferase and galactosidase activities.

Transient transfection with RNA transcripts were also performed using cationic liposomes as described previously (19,20,23). Briefly, ∼3 × 10⁵ cells/well were seeded onto 6-well plates on the day before transfection. Cells were washed once with Opti-MEM I reduced serum medium (Invitrogen) and left in the incubator with some medium during preparation of the liposome-polynucleotide complexes. One milliliter of Opti-MEM I medium in a 12 × 75 mm polystyrene snap-cap tube was mixed with 12.5 μg of Lipofectin reagent and 5 μg of RNA transcripts. The liposome–RNA–medium mixture was immediately added to cells. Eight hours following transfection, cells were harvested and processed for luciferase assay.

Cells co-transfected with dicistronic or monocistronic reporter plasmids and β-galactosidase control plasmid were lysed in the Passive Lysis Buffer (Promega). Both the firefly (Fluc) and Renilla luciferase (Rluc) activities were determined using the Dual-Luciferase Reporter Assay kit as described previously (19). The firefly luciferase activity from monocistronic constructs was determined using the Mono-Luciferase Reporter Assay kit. The activity of β-galactosidase was measured as described previously (24) and was used to normalize the differences of transfection efficiency.

Approximately 2 × 10⁶ cells in 10 cm plates were transfected with DNA (4 μg/plate) using Lipofectamine Plus reagent. At 48 h post-transfection, cells were harvested and the total RNAs were extracted using an RNeasy Mini Kit. Potential contaminating plasmid DNA in the total RNAs was digested with RNase-free DNase.

Ribonuclease Protection Assay (RPA) was performed using the RPA III kit according to the manufacturer's instruction. To prepare RNA probe, the region between −277 and −1 of human p27 5′-UTR sequence and +1 to +133 in the ORF of the firefly luciferase gene was amplified using PCR and cloned into pGEM-T Easy vector. The resulting plasmid was linearized with SpeI and labeled in vitro using T7 RNA polymerase in the presence of 0.5 mM each of ATP, UTP, GTP and 0.01 mM CTP supplemented with 50 μCi [α-³²P]CTP. The [α-³²P]labeled probe was digested with DNase I and purified using a Sephadex G-25 Quick Spin Column. About 5 × 10⁵ c.p.m. of probe was hybridized to 15 μg total RNA at 45°C overnight followed by digestion with RNase T1/A for 35 min at 37°C. The reaction was then stopped and the protected RNA samples were separated by electrophoresis on a 6% acrylamide/8 M urea gel. The gel was then dried for autoradiography.

Dicistronic DNA test has been used in the previous studies to show the existence of IRES activities in the 5′-UTR sequence of both human and murine p27, although it was thought that the length of the IRES element may vary between the two species (8,9). To further characterize the 5′-UTR sequence, we engineered three dicistronic constructs containing various lengths of the 5′-UTR sequence of human p27 (Figure 1B) flanked by Renilla and firefly luciferase genes (Figure 1A). These constructs were then transfected into H1299 cells for the analysis of potential IRES activity using luciferase assay. As shown in Figure 1C and Table 1, the expression of the second cistron firefly luciferase was greatly enhanced by the presence of the p27 5′-UTRs although the enhancement decreased gradually with the truncation of sequence from the 5′ end. The expression of Renilla luciferase was not changed by the insertion of the 5′-UTR sequence of p27. The increased ratio of firefly to Renilla luciferase activities by the 5′-UTR sequence of p27 compared with vector control suggests that there may be an IRES element in these sequences (Figure 1D). Insertion of a hairpin in front of the first cistron did not affect the expression of the second cistron firefly luciferase (data not shown). These observations are consistent with the previous findings and with the claim of the existence of IRES activity in the 5′-UTR sequence of p27 (8–10).

The above studies suggest that the 5′-UTR of p27 may contain an IRES element. However, the enhanced expression of the second cistron could also be due to the existence of promoters in the 5′-UTR or due to alternative splicing of the dicistronic transcripts as suggested previously (15–17,19). To further determine whether the 5′-UTR of p27 contains IRES activities, we performed a more stringent dicistronic mRNA test, which allows a direct analysis whether the 5′-UTR sequence of p27 in the intergenic region can increase the translation of the second cistron firefly luciferase without the transcriptional and splicing interference. For this purpose, we engineered additional constructs for making in vitro dicistronic transcripts containing 5′-caps and 3′-poly(A) tails (Figure 2A). In addition, a well-known viral HRV IRES sequence was also cloned into the intergenic region to serve as a positive control. These in vitro transcripts were then transfected into H1299 cells, which were then harvested for luciferase assay. As shown in Figure 2B and C, the expression of the second cistron firefly luciferase was enhanced only by the HRV IRES as expected. None of the 5′-UTR sequences of p27 enhanced the expression of the second cistron firefly luciferase. Thus, it is likely that the 5′-UTR of p27 does not contain a functional IRES element as previously thought.

The above results of dicistronic mRNA assay prompted us to investigate whether the 5′-UTR sequence of p27 contains a promoter that could be responsible for the observed stimulation of expression of the second cistron firefly luciferase in the dicistronic DNA assays (Figure 1). For this purpose, we first engineered promoterless dicistronic constructs as described previously (19,20) by simply removing the unique SV40 promoter together with the intron sequence from the pRF-based dicistronic constructs (Figure 3A). These promoterless dicistronic constructs will not generate dicistronic mRNAs after transfection into cells due to the lack of the SV40 promoter. Thus, the activity of any potential IRES element in the constructs should not be detected. Any expression of the second cistron firefly luciferase from these promoterless dicistronic DNAs will be due to the existence of a promoter in the 5′-UTR sequence of p27. As shown in Figure 3B and C, the enhanced expression profile of the second cistron firefly luciferase in the promoterless construct is similar to that generated in the presence of the SV40 promoter (compare with Figure 1). The actual firefly luciferase activities are similar to that generated in the presence of SV40 promoter, whereas the actual Renilla luciferase activity is drastically decreased due to the removal of the vector promoter (see Table 1). Thus, the 5′-UTR sequence of p27 likely contains promoters.

To ensure that the above observations are not specific to the cell line used, we performed similar studies using 293T cells, which has been used in a previous study of p27 5′-UTR (25). First, dicistronic mRNAs were transfected into 293T cells and luciferase activities were determined as described above. As shown in Figure 4A and B, the expression of the second cistron firefly luciferase was enhanced only by the HRV IRES as expected. None of the 5′-UTR sequences of p27 enhanced the expression of the second cistron firefly luciferase in 293T cells. These observations are consistent with the data shown using H1299 cells (compare with Figure 2). We next transfected the promoterless dicistronic constructs into 293T cells and determined luciferase activities. As shown in Figure 4C, the firefly luciferase expression was enhanced by the presence of the 5′-UTR sequence of p27 in the intergenic region, suggesting the presence of promoter activities (see also Table 1). This observation is also similar to that generated in H1299 cells (compare with Figure 3). In fact, the promoter activity in the longest form of 5′-UTR of p27 as indicated by the firefly luciferase activity is much more pronounced in 293T cells compared with H1299 cells, although the shorter 5′-UTR segments had similar activities in the two cell lines (see Table 1). Similar observations have also been made with HeLa cells that were used in two other previous studies (9,10). Taken together, we conclude that the previous findings of the putative ‘IRES’ activities in the 5′-UTR sequence of p27 was not due to the specific cell lines used and that most likely they are all misled from the promoter activities.

Recently, it was suggested that PTB enhanced the putative IRES activity of p27 5′-UTR in cell-free expression system and in vivo (25). We performed similar studies by translating dicistronic RNA transcripts in RRL in the absence or presence of purified PTB with HRV IRES as a positive control. Previously, we and others have observed that the HRV IRES is inactive in RRL unless HeLa cell extracts or PTB is added (19,26,27) (Figure 4D and Table 2). Contrary to the previous finding by Cho et al. (25), we did not observe any enhanced expression of the second cistron firefly luciferase by the 5′-UTR of p27 in the absence or presence of PTB (Figure 4D). In fact, the firefly luciferase activity generated is very low, >500-fold less than the Renilla luciferase activity (Table 2).

To confirm the above conclusion that the 5′-UTR of p27 contains promoters, we further tested if the transcripts with only the second cistron firefly luciferase exist in the cells transfected with dicistronic DNA constructs. RNase protection assay was performed on total RNAs isolated from H1299 cells following transfection with the traditional and promoterless dicistronic constructs using a RNA probe covering the region between −277 and −1 of human p27 5′-UTR sequence and the region between +1 and +133 of firefly luciferase ORF (Figure 5A). As shown in Figure 5B, a short fragment with an estimated size of 124 bases was produced from the traditional dicistronic vector (lane 4), which represents the protected luciferase gene from the dicistronic RNA transcripts. This product is barely detectable with the RNA isolated from cells transfected with the promoterless dicistronic vector (lane 8), suggesting that the functional SV40 promoter has been successfully removed. A fragment with an estimated size of 415 bases was produced from the traditional dicistronic constructs containing the 575 and 461 bases of p27 5′-UTRs (lanes 5 and 6). This product is likely generated from the dicistronic RNA produced using the vector SV40 promoter. However, a similar product, albeit with less intensity, was also observed with the promoterless dicistronic construct (lane 9), suggesting that a transcription start site may exist upstream of the probe used (Figure 5A, the marker X). Four fragments of 272, 258, 184 and 166 bases were also produced with both the traditional and the promoterless dicistronic constructs containing 575 and 461 bases of p27 5′-UTRs (lanes 5, 6, 9 and 10), suggesting that four transcription start sites likely exist at around −139, −125, −51 and −33. Interestingly, for the traditional dicistronic construct containing only 150 bases of the p27 5′-UTR (lane 7), only the product of 184 bases was observed, suggesting that except the transcription start site −51, all others are silent in this construct. The product of 290 bases (lane 7) was likely generated from the protection of the dicistronic RNA transcripts containing 150 bases of p27 5′-UTR. It appears that the shortest 5′-UTR segment (−150) in the promoterless construct apparently has promoter activity (Figure 3) but it is too weak for direct RNA analysis to detect the transcript (Figure 5, lane 11). It is possible that the transcription start site at −51 requires the vector SV40 promoter and/or the upstream p27 promoters, which may function as regulatory promoters (28). It is also possible that the production of the 184 bases fragment from the shortest traditional dicistronic construct was due to potential alternative splicing, which is absent with the removal of sequences together with the SV40 promoter in the shortest promoterless dicistronic plasmid (see below).

To further confirm the promoter activities in the 5′-UTR sequence of p27, we cloned these sequences into the standard promoter-testing monocistronic vector pGL3, which does not have any promoter element. These constructs containing the different sizes of p27 5′-UTR were then transfected into H1299 cells for the determination of luciferase activity. As shown in Figure 6B, similar firefly luciferase activities were observed with these constructs compared with that generated from the traditional and promoterless dicistronic constructs (compare with Figures 1 and 3). RNase protection analysis of the RNAs isolated from cells transfected with these monocistronic constructs also showed the production of various transcription products similar to that shown with the dicistronic constructs (compare Figure 6C with Figure 5) although, again, the shortest 5′-UTR segment had promoter activity which is too weak to be detected by direct RNA analysis.

Based on the above observations with promoterless dicistronic and the standard promoter-testing monocistronic constructs, we conclude that the 5′-UTR sequence of human p27 contains promoter activities that can initiate transcription at several sites. Transcription at these sites in the dicistronic DNA assay generates monocistronic transcripts containing only the firefly luciferase gene with short 5′-UTRs that can be translated easily using the scanning mechanism (12,13). These shorter monocistronic RNA transcripts are likely responsible for the enhanced expression of firefly luciferase in the dicistronic DNA test and they may represent the previously reported IRES activity.