Enhancement of +1 Frameshift by Polyamines during Translation of Polypeptide Release Factor 2 in Escherichia coli*

Polypeptide release factor 2 (RF2) in Escherichia coli is known to be synthesized by a +1 frameshift at the 26th UGA codon of RF2 mRNA. Polyamines were found to stimulate the +1 frameshift of RF2 synthesis, an effect that was reduced by excess RF2. Polyamine stimulation of +1 frameshift of RF2 synthesis was observed at the early logarithmic phase, which is the important phase in determination of the overall rate of cell growth. A Shine-Dalgarno-like sequence was necessary for an efficient +1 frameshift of RF2 synthesis, but not for polyamine stimulation. Spectinomycin, tetracycline, streptomycin, and neomycin reduced polyamine stimulation of the +1 frameshift of RF2 synthesis. The results suggest that a structural change of the A site on 30 S ribosomal subunits is important for polyamine stimulation of the +1 frameshift. The level of mRNAs of ribosomal proteins and elongation factors having UAA as termination codon was enhanced by polyamines, and OppA synthesis from OppA mRNA having UAA as termination codon was more enhanced by polyamines than that from OppA mRNA having a UGA termination codon. Furthermore, synthesis of ribosomal protein L20 and elongation factor G from the mRNAs having a UAA termination codon was enhanced by polyamines at the level of translation and transcription. The results suggest that some protein synthesis from mRNAs having a UAA termination codon is enhanced at the level of translation through polyamine stimulation of +1 frameshift of RF2 synthesis. It is concluded that prfB encoding RF2 is a new member of the polyamine modulon.

Polyamines (putrescine, spermidine, and spermine), aliphatic cations present in almost all living organisms, are necessary for normal cell growth (1)(2)(3). Because polyamines interact with nucleic acids and mostly exist as polyamine-RNA complexes in cells (4,5), their proliferative effects are presumed to be caused by stimulation of nucleic acid and protein synthesis. In this context, it has been reported that polyamines stimulate the synthesis of some proteins in vitro and in vivo (6 -9), increase the fidelity of protein synthesis (10 -12), and induce the in vivo assembly of 30 S ribosomal subunits (13)(14)(15), together suggesting that polyamines regulate protein synthesis at several different steps.
We have previously studied whether the translation of a defined set of proteins is enhanced by polyamines using Escherichia coli MA261, a polyamine-requiring mutant. The strain MA261 is unable to synthesize putrescine, and cell growth slows down in the absence of exogenous putrescine (16). When putrescine is added to the culture medium, it is taken up into cells, and spermidine is synthesized from putrescine, leading to the recovery of cell growth. By comparing the protein expression pattern in MA261 cultured with or without putrescine, we reported previously that polyamines increase the synthesis of several proteins at the level of translation (17)(18)(19)(20). These include OppA (17), a periplasmic substrate-binding protein of the oligopeptide uptake system; adenylate cyclase (18); RpoS ( 38 ) (19), one of the subunits of RNA polymerase; FecI factor ( 18 ) (20), a transcription factor for the iron transport operon; and Fis (20), a global regulator of transcription of some growthrelated genes including those for rRNA and some tRNAs. We proposed that the genes whose expression is enhanced by polyamines at the level of translation can be classified as a "polyamine modulon." In the case of RpoS ( 38 ) synthesis, UAG amber codon-dependent Gln-tRNA supE binding to ribosomes at the 33rd position of open reading frame (ORF) 3 of rpoS mRNA was enhanced by polyamines (19). The termination codon in the ORF of mRNA can be decoded not only by readthrough but also by ϩ1 frameshift. In mammalian cells, it has been reported that antizyme (AZ) mRNA has the termination codon UGA at the 69th position of its ORF, and the ϩ1 frameshift is strongly enhanced by polyamines (21,22). AZ is involved in the negative regulation of polyamine content in cells by stimulating the degradation of ornithine decarboxylase and by inhibiting the polyamine uptake (23)(24)(25)(26). In E. coli, polypeptide release factor 2 (RF2), which recognizes UGA and UAA as termination codons, is known to be synthesized by a ϩ1 frameshift at the 26th UGA termination codon of RF2 mRNA (27,28). In this study, we examined whether the ϩ1 frameshift of RF2 mRNA is enhanced by polyamines. mg of MgSO 4 ⅐7H 2 O, 13 mg of CaCl 2 , 2 mg of thiamine, 100 mg each of alanine, asparagine, aspartic acid, glutamic acid, glycine, methionine, proline, serine, and threonine, and 50 mg each of cysteine, histidine, isoleucine, leucine, phenylalanine, tyrosine, and valine, 5 mg of tryptophan, and 1 g of arginine/liter of water) was used. An ornithine deficiency was achieved by the addition of 1 mg of arginine/ml to the medium. For growth of MA261, medium B (4 g of glucose, 7 g of K 2 HPO 4 , 3 g of KH 2 PO 4 , 500 mg of sodium citrate, 1 g of (NH 4 ) 2 SO 4 , 100 mg of MgSO 4 ⅐7H 2 O, 2 mg of thiamine, 10 mg of biotin, 100 mg each of leucine, threonine, methionine, serine, glycine, and ornithine/liter of water) was used. The cells were cultured in the absence or presence of 100 g/ml putrescine dihydrochloride. Cell growth was monitored by measuring the absorbance at 540 nm.
The plasmid pMW-OppA(UGA) encoding oppA having UGA as termination codon was prepared from pMW-OppA(UAA) having UAA as termination codon (pMW975 in Ref. 14) with site-directed mutagenesis by overlap extension using PCR (32). The nucleotide sequence of the plasmids was confirmed by the Gene Rapid System (Amersham Biosciences).
Western Blot Analysis-Western blot analysis was performed by the method of Nielsen et al. (33) using ECL TM Western blotting reagents (Amersham Biosciences). Antibody against RF2 (34) was kindly sup-FIGURE 1. Effect of polyamines on ؉1 frameshift of RF2 synthesis in polyaminerequiring mutants. A, structure of RF2 mRNA and its product. B, MA261 and DR112 cells were cultured in the absence and presence of 100 g/ml putrescine and harvested at A 540 ϭ 0.1, 0.15, or 0.2. Western blotting of RF2 was performed using 10 g of protein of cell lysate. C, dot blot analysis of RF2 mRNA was performed as described under "Experimental Procedures." The experiments were repeated three times, and the S.E. was within Ϯ10% for each value. plied by Dr. Y. Nakamura (the University of Tokyo, Tokyo, Japan), and that against maltose-binding protein was from New England BioLabs. Antibodies against ribosomal protein L20 and elongation factor G (EF-G) were prepared as described previously (35,36). The level of protein was quantified by a LAS-1000 plus luminescent image analyzer (Fuji Film).
Dot Blot Analysis and DNA Microarray Analysis-Total RNA was prepared from E. coli MA261 and DR112 cells according to the method of Emory and Belasco (37). Dot blot analysis was performed using various amounts of total RNA and the 32 P-labeled probe (38). PCR of prfB gene was performed using primers P7 and P8, and PCR product was labeled with [␣-32 P]dCTP using a BcaBEST TM labeling kit (Takara Shuzo Co.) and used as probe. DNA microarray analysis was performed as described previously (20) with total RNA isolated from E. coli MA261 cultured in the presence and absence of putrescine and harvested at A 540 ϭ 0.15 using TaKaRa Intelli Gene E. coli CHIP, version 2.0 (Takara Shuzo Co.).
In Vitro Translation-Purified RF2 protein was kindly provided by Dr. Y. Nakamura (34). DNA-dependent transcription-translation reactions (39) were carried out using an E. coli T7 S30 Extract System for Circular DNA (Promega) following the supplier's manual with minor modifications. A reaction mixture (0.01 ml) contained 0.4 g of pBluescript II SKϩ derived DNA, 120 kBq of [ 35 S]methionine (27 TBq/ mmol), 0.1 mM each of 19 amino acids without methionine, 3 l of T7 S30 extract, 4 l of the attached reaction mixture, and magnesium acetate and spermidine at the specified concentration. If the reaction mixture (0.01 ml) was used as it is, the concentration of magnesium acetate and spermidine was 9 and 0 mM, respectively. To examine the effect of RF2 protein and antibiotics on spermidine stimulation of the ϩ1 frameshift of RF2 synthesis, the various levels of RF2 and antibiotics shown in the figures were added to the reaction mixture. After incubation at 37°C for 60 min, 2 l of 20 mM methionine and 1 ml of ice-cold 5% trichloroacetic acid were added to the reaction mixture. After removal of supernatant, the pellet was dissolved with 25 l of sample buffer for SDS-PAGE (40) and boiled for 2 min. A 20-l aliquot was used for 7.5% SDS-PAGE, and fluorography was performed according to the method of Laskey and Mills (41). Radioactivity of labeled proteins was quantified using a Fuji BAS-2000II imaging analyzer (Fuji Film).

Identification of Spermidine Cross-linked to 30 S Ribosomal Proteins
through Dimethyl Suberimidate-Preparation of salt (0.5 M NH 4 Cl)washed ribosomes from E. coli Q13 (rna pnp) and 30 S ribosomal subunits from ribosomes were carried out as described previously (42,43). The reaction mixture (0.1 ml) for spermidine binding to 30 S ribosomal subunits contained 20 mM Tris-HCl (pH 7.5), 10 mM magnesium acetate, 100 mM NH 4 Cl, 3 A 260 units of 30 S ribosomal subunits, 1 mM [ 14 C]spermidine (148 kBq), and antibiotics at the specified concentration. After incubation at 30°C for 30 min, an equal volume (0.1 ml) of cross-linking reagent containing 66 mM sodium borate, 10 mM magnesium acetate, 100 mM NH 4 Cl, and 45 mM dimethyl suberimidate, was added to the reaction mixture, the final pH being 8.2. Further incubation was performed at 30°C for 30 min. Ribosomal proteins were extracted with 66% acetic acid and 30 mM magnesium acetate (44). The supernatant, obtained by centrifugation at 15,000 rpm for 10 min, was then treated with 5 volumes of acetone. The precipitate thus obtained was dissolved in the sample loading buffer containing 6 M urea and 50 mM 2-mercaptoethanol. Ribosomal proteins were analyzed by two-dimensional gel electrophoresis; the first dimension gel electrophoresis was run in an acidic system, and the second dimension separation was carried out in the presence of SDS (45). Identification of the individual proteins was according to the nomenclature proposed by Kaltschmidt and Wittman (45). The radioactivity incorporated into each of the ribosomal protein was quantified using a Fuji BAS-2000II imaging analyzer (Fuji Film).

Effect of Polyamines on in Vivo ϩ1
Frameshift of RF2 Synthesis-It is well known that AZ synthesis involves a ϩ1 frameshift at the 69th termination codon of ORF of AZ mRNA in mammalian cells, and the efficiency of this frameshift is enhanced by polyamines (21,22). Because RF2 in E. coli is also synthesized by a ϩ1 frameshift at the 26th termination codon (Fig. 1A) (27,28), the effect of polyamines on the efficiency of this frameshift was examined using the polyamine-requiring mutants MA261 and DR112. When these two mutants were cultured in the presence of 100 g/ml putrescine, polyamines (putrescine and spermidine) in cells increased, and cell growth became faster (data not shown). The level of RF2 was increased ϳ1.7-fold by polyamines only at the early FIGURE 2. Effect of polyamines on ؉1 frameshift of RF2 synthesis in the translational termination/frameshift assay system. A, three products in the translation termination/ frameshift assay system. B, MA261 and DR112 cells transformed with pUC-malE-RF2 (UGA) were cultured with or without putrescine and harvested at A 540 ϭ 0.2. Three products were analyzed by Western blotting using antibody against maltose-binding protein, and 8 or 24 g of protein of cell lysate prepared from cells cultured with or without putrescine. The experiments were repeated three times, and the S.E. was within Ϯ10%.
logarithmic phase (Fig. 1B), which is the important phase with the regard to the overall rate of cell growth. The level of RF2 increased with the progression of cell growth and the effect of polyamines declined. The level of RF2 mRNA was nearly equal in cells cultured with or without putrescine (Fig. 1C). As a control, the level of 70 , one of the subunits of RNA polymerase, was measured, and it was not increased by polyamine (data not shown). The results suggest that the ϩ1 frameshift of RF2 synthesis is enhanced by polyamines and negatively regulated by RF2 (28).
The effects of polyamines were then examined using translational termination/frameshift assay system (46). In this system, three kinds of products can be identified by antibody against maltose-binding protein; those are: a ϩ1 frameshift product (55 kDa), a readthrough product (50 kDa), and a termination product (45 kDa) ( Fig. 2A). Efficiency of the ϩ1 frameshift was calculated as the percentage of the level of ϩ1 frameshift product divided by the total level of three products. As shown in Fig. 2B, the ϩ1 frameshift of RF2 was enhanced by polyamines 2.0 -2.2-fold. The polyamine effect was greater in this assay system, probably because of the existence of higher level of mRNA synthesized from the multicopy plasmid. The results confirm that the ϩ1 frameshift of RF2 synthesis is enhanced by polyamines.
Factors that influence polyamine enhancement of ϩ1 frameshift were then examined using E. coli DR112. The termination codon (UGA) at the 26th codon of ORF of RF2 mRNA was changed to other termination codons, UAG and UAA, or to UGG for Trp (Fig. 3A). However, polyamine enhancement of the ϩ1 frameshift was not influenced by changing this codon (Fig. 3B). The influence of a SD-like sequence at the 5Ј-end of the termination codon was also studied. If the SD-like sequence was present, polyamine enhancement of the ϩ1 frameshift was observed regardless of the nucleotide sequence (AGGGGG and AGGAGG). When the SD-like sequence was changed to a non-SD-like sequence (AGGGAG), the ϩ1 frameshift decreased greatly (Fig. 3, C  and D). Thus, it remains to be clarified whether polyamines are involved in the interaction between the SD-like sequence and 16 S rRNA.

Effect of Spermidine, RF2, and SDϪlike Sequence on in Vitro ϩ1
Frameshift of RF2 Synthesis-An in vitro DNA-dependent protein synthetic system (39) was used to study the factors influencing the ϩ1 frameshift of RF2 synthesis. As shown in Fig. 4A, RF2 synthesis was enhanced 1.8-fold by 1.5 mM spermidine. Confirming this, spermidine, but not Mg 2ϩ , enhanced the ϩ1 frameshift of RF2 synthesis in a translational termination/frameshift assay (Fig. 4, B and C). The effect of RF2 on the ϩ1 frameshift of RF2 synthesis was then examined. As shown in Fig. 4 (D and E), the addition of RF2 reduced ϩ1 frameshift of RF2 synthesis and reduced the polyamine enhancement of this frameshift.
Because it was not clear in an in vivo assay system whether polyamines function at the level of the interaction between the SD-like sequence and 16 S rRNA, this was examined in an in vitro system. When the SD-like sequence was absent, the efficiency of the ϩ1 frameshift greatly decreased. However, polyamine enhancement of the frameshift was still observed (Fig. 4F). The results suggest that polyamines may act on 30 S ribosomal subunits rather than on RF2 mRNA. We next tested whether AZ synthesis was enhanced by polyamines in an in vitro E. coli cell-free system. Although AZ mRNA does not have the SD-like sequence (22) and efficiency of ϩ1 frameshift was low, spermidine  clearly enhanced ϩ1 frameshift of synthesis (data not shown), suggesting again that polyamines may act on 30 S ribosomal subunits.

Effect of Various Antibiotics on Polyamine Stimulation of in Vitro ϩ1
Frameshift of RF2 Synthesis-This was examined using an in vitro translational termination/frameshift assay system. As shown in Fig. 5, polyamine stimulation of the ϩ1 frameshift of RF2 was decreased by the addition of streptomycin, neomycin, tetracycline, and spectinomycin, which act at the A site of 30 S ribosomal subunits (47)(48)(49), in parallel with the decrease in protein synthesis (Fig. 5, A-D). However, polyamine stimulation of frameshifting was not affected by edeine and pactamycin, which act at the P or E site of 30 S ribosomal subunits (49,50), although protein synthetic activity decreased greatly (Fig. 5, E and F). The results suggest that structural change at the A site of 30 S ribosomal subunits by polyamines is important for polyamine stimulation of the ϩ1 frameshift of RF2 synthesis.
Polyamine stimulation of the ϩ1 frameshift of RF2 was more strongly affected by spectinomycin and tetracycline than by streptomycin and neomycin (Fig. 5, A-D). It has been reported that spectinomycin and tetracycline bind close to the A site (Fig. 5G) (47)(48)(49). Ribosomal proteins S4 and S5 are also located close to the A site (Fig. 6E) (52,53). Thus, the level of spermidine bound to the A site of 30 S ribosomal subunits was measured in the presence and absence of spectinomycin through spermidine cross-linking to 30 S ribosomal proteins by dimethyl suberimidate (44). As shown in Fig. 6 (A-D), spermidine cross-linked to S4 was decreased in the presence of spectinomycin. However, we could not estimate the spermidine level cross-linked to S5 because spermidine cross-linked to S5 was very low. Similar results were obtained with 1 mM tetracycline (data not shown). Along these lines, the inhibition of growth of a polyamine requiring mutant DR112 by spectinomycin and tetracycline was reduced by polyamines (data not shown). These results indicate that the level of spermidine bound to the A site of 30 S ribosomal subunits decreased in the presence of spectinomycin and tetracycline and vice versa. Thus, it is deduced that the decrease in polyamine stimulation of the ϩ1 frameshift of RF2 synthesis by these antibiotics is due to a decrease in spermidine bound to the A site of 30 S ribosomal subunits. Analysis of mRNA Species Enhanced by Polyamines at the Early Logarithmic Phase-Experiments were carried out to study the physiological significance of the stimulation of ϩ1 frameshift of RF2 synthesis at the early logarithmic phase. In these studies, mRNAs that are up-regulated by polyamines were identified by DNA microarray analysis. It has been reported that recognition of the termination codon UAA by RF2 is weaker than that of UGA (51). At the early logarithmic phase, expression of 347 genes was enhanced by polyamines more than 2-fold among 2,714 detected genes at A 540 ϭ 0.15. In particular, expression of the genes in the categories of energy production and translation were enhanced by polyamines (Fig. 7A). Because expression of genes involved in translation such as ribosomal proteins and elongation factors of protein synthesis was not significantly enhanced by polyamines at the middle logarithmic phase (A 540 ϭ 0.5) (20), the usage frequency of UAA as a termination codon was investigated in the mRNAs involved in translation. As shown in Fig. 7B, the usage frequency of UAA as a termination codon among the mRNAs expressed at the early logarithmic phase was 65.1%, and that among the mRNAs up-regulated by polyamines was 69.5%. However, the usage frequency of UAA as a termination codon among the mRNAs involved in translation was 82.3%, and that in the mRNAs up-regulated by polyamines was 93.9%. Thus, it was tested whether protein synthesis from mRNA having UAA as termination codon at the early logarithmic phase would be enhanced by polyamines at the translational level because of the increase in RF2 using OppA mRNA whose synthesis is most strongly enhanced by polyamines (17). As shown in Fig. 8, synthesis of OppA from the mRNA having UAA as termination codon was more strongly enhanced by polyamines at the level of translation than that from the mRNA having UGA as termination codon. The level of both OppA mRNAs was enhanced by polyamines by 1.8 -1.9-fold (Fig. 8). These results suggest that the mRNAs having UAA as a termination codon were more efficiently translated through quick termination of the synthesis of these proteins. Synthesis of ribosomal protein L20 and EF-G from the mRNAs having UAA termination codon was actually enhanced by polyamines at the level of not only transcription but also translation, i.e. the level of these mRNAs was enhanced by polyamines 1.9-fold, and the level of the proteins was enhanced by polyamines 2.1-fold (Fig. 8B). As a control, synthesis of 24 subunit of RNA polymerase from the mRNA having a UGA termination codon was measured. The level of both 24 subunit protein and its mRNA was enhanced by polyamines 1.2-fold (data not shown), i.e. synthesis of the 24 subunit protein was not enhanced by polyamines at the level of translation.
Effect of Polyamines on Ϫ1 Frameshift of DnaX mRNA-In E. coli, one more protein is synthesized by the frameshift, that is, the ␥ subunit of DNA polymerase III holoenzyme is synthesized by the Ϫ1 frameshift of DnaX mRNA (54). The effect of polyamines on Ϫ1 frameshift of DnaX mRNA was examined using in vivo and in vitro systems. As shown in Fig. 9, efficiency of the Ϫ1 frameshift in vivo is very high even in the absence of polyamines (80 -90% in both MA261 and DR112 strains), and polyamines did not influence the Ϫ1 frameshift of DnaX mRNA in vivo and in vitro. FIGURE 7. Identification of genes whose expression was enhanced by polyamines at the level of transcription (A) and the usage frequency of termination codon (B). A, DNA microarray analysis was performed as described under "Experimental Procedures." The data were the average of three experiments. Classification of genes by functional category was carried out using the data base of GenoBase (ecoli.aist-nara.ac.jp/gb4/search/function/orffunc.php). Among the 15 categories, a category labeled with red letters means that the percentage of the category in genes up-regulated in the presence of putrescine increased Ͼ1.5-fold compared with the percentage in 2,714 detected genes. B, the usage frequency of termination codons. The percentage of three kinds of termination codon used in the mRNAs was calculated using the data base of GenoBase.

DISCUSSION
Our results indicate that the ϩ1 frameshift of RF2 synthesis in E. coli was enhanced by polyamines. However, this was only observed at the early logarithmic phase of cell growth, because synthesis of RF2 is autoregulated by RF2 protein (28), and the level of RF2 was increased with the progress of cell growth. It is also well known that polyamines enhance the ϩ1 frameshift of AZ synthesis in mammalian cells (22) and yeast (55). Polyamine enhancement of the ϩ1 frameshift in synthesis of RF2 and AZ seems rational. When the level of RF2 is low at the early logarithmic phase, polyamines stimulate RF2 synthesis to release the completed protein from ribosomes rapidly. When high levels of polyamines become potentially toxic to cells, AZ is induced to decrease the level of polyamines in eukaryotic cells. AZ is known to stimulate degradation of ornithine decarboxylase, a rate-limiting enzyme in polyamine biosynthesis (23,24), and to inhibit polyamine uptake (25,26). It should be noted, however, that the ϩ1 frameshift at CUU codon of gag-pol protein for transposition of Ty1 in yeast is decreased by polyamines (56). With regard to Ϫ1 frameshifts, the synthesis of ␥ subunit of DNA polymerase III holoenzyme (this work) and gag-pol fusion protein of the L-A double-stranded RNA virus in yeast (56) is not affected by polyamines. The results suggest that the mechanism of ϩ1 and Ϫ1 frameshifting on ribosomes is different.
For the ϩ1 frameshift in RF2 synthesis, an SD-like sequence was necessary to slow down the speed of translation. However, an SD-like sequence did not affect polyamine stimulation of the ϩ1 frameshift. Therefore, we thought that polyamines may act on ribosomes rather than mRNA to stimulate ϩ1 frameshifting. Indeed, spectinomycin and tetracycline, which bind to the A site of 30 S ribosomal subunits (47)(48)(49), reduced the polyamine enhancement of ϩ1 frameshifting. However, pactamycin and edeine, which bind closely to the interaction site with an SD-like sequence of RF2 mRNA on 30 S ribosomal subunits, did not influence the ϩ1 frameshifting. The results strongly suggest that spermidine binds to and causes a structural change in the A site to enhance the ϩ1 frameshift.
It has been reported that premature release of the E site tRNA from the ribosomes is coupled with high level frameshifting (57). Our results suggest that polyamines do not stimulate premature release of the E site tRNA, because edeine and pactamycin, which act at the P or E site of 30 S ribosomal subunits, did not influence polyamine stimulation of the ϩ1 frameshift of RF2 synthesis.
Polyamines preferentially bind to double-stranded RNA, especially GC-rich double-stranded RNA (4,58). We have also shown that the existence of a GC-rich double-stranded region close to SD sequence of OppA mRNA is important for spermidine stimulation of fMet-tRNA binding to ribosomes (59). It is also estimated that ϳ10 -15% of phosphate in RNA interacts with polyamines in cells (4,5), and the binding site of polyamines on 16 S rRNA has been determined (60). Several binding sites of polyamines and those of antibiotics acting at the A site of 30 S ribosomal subunits were close to each other, and the binding of spermidine at the A site was inhibited by these antibiotics judging from the decrease in cross-linked spermidine to S4 protein. Our results suggest that spermidine binding to the A site is important for spermidine stimulation of ϩ1 frameshift during RF2 synthesis. Because RF2 also binds close to the A site (61), the binding of RF2 probably inhibits the ϩ1 frameshift by competing with spermidine binding to the A site. These results, taken together, suggest that a structural change of the decoding site by polyamines is important for the ϩ1 frameshift of RF2 synthesis, and antibiotics acting on the A site and RF2 disturb the polyamine stimulation of the ϩ1 frameshift through their binding to the decoding site.
It is noted that expression of many mRNAs encoding ribosomal proteins and elongation factors was enhanced by polyamines at the early logarithmic phase. Because these mRNAs contain the UAA termination codon, it is quite possible that the efficiency of translation from these mRNAs is increased by polyamine enhancement of RF2 synthesis judging from the results of OppA synthesis and polyamine enhancement of the synthesis of ribosomal protein L20 and EF-G at the levels of both transcription and translation shown in Fig. 8. The mechanism of polyamine stimulation of the synthesis of these mRNAs is now under investigation.
It is known that polyamines stimulate cell growth through enhancement of the assembly of 30 S ribosomal subunits (13)(14)(15) and stimulation of the synthesis of specific kinds of proteins that are important for cell growth (20,62). We proposed that the genes whose expression is enhanced by polyamines at the level of translation are classified as polyamine modulon. Those are oppA, cya, rpoS, FIGURE 9. Effect of polyamines on ؊1 frameshift of Pol III␥ synthesis in in vivo and in vitro assay systems. A, structure of DnaX mRNA and its products. B, two products (0 frame and Ϫ1 frameshift) in the translation termination/frameshift assay system. C, MA261 and DR112 cells transformed with pUC-malE-DnaX were cultured in the absence and presence of 100 g/ml putrescine and harvested at A 540 ϭ 0.2. Western blotting was performed as described in the legend of Fig. 1. D, MalE-DnaXmRNAderivedproteinsynthesisinDNAdependent transcription and translation reactions was performed in the presence of various concentrations of spermidine shown in the figure, and the efficiency (%) of the Ϫ1 frameshift of Pol III␥ synthesis was shown. The experiments were repeated three times, and the S.E. was within Ϯ10%.
fecI, and fis (20). Among these genes, four genes encode transcription factors. Accordingly, expression of mRNAs regulated by the above transcription factors are enhanced by polyamines. We found that expression of ϳ300 mRNAs is enhanced by polyamines. (available at www.jbc.org). However, polyamine enhancement of expression of these mRNAs was not explained completely by the above transcription factors. Experiments are in progress to identify a new polyamine modulon from the genes encoding transcription factors including a factor for ribosomal protein mRNAs. Our present results indicate that prfB encoding RF2 is a new member of polyamine modulon, although RF2 is not a transcription factor. Fis is known to enhance the synthesis of ribosomal RNA (63). Enhancement of the synthesis of ribosomal proteins by polyamines, partly because of the enhancement of RF2 synthesis, at the early logarithmic phase is also important for efficient protein synthesis through active ribosome formation.