Molecular mechanism of polyamine stimulation of the synthesis of oligopeptide-binding protein.

Polyamine stimulation of the synthesis of oligopeptide-binding protein (OppA) was shown to occur mainly at the level of translation by measuring OppA synthesis and its mRNA level. Several artificial oppA genes were constructed by site-directed mutagenesis. These synthesize different kinds of OppA mRNAs: mRNAs differing in the size of 5′-untranslated region; mRNAs having the Shine-Dalgarno (SD) sequence in a different position; mRNAs having different secondary structure in the region of the SD sequence; and fusion mRNAs consisting of the 5′-untranslated region of OppA mRNA and the open reading frame of β-galactosidase. By measuring the synthesis of OppA or β-galactosidase from these mRNAs, we found that the 171-nucleotide 5′-untranslated region and 145 nucleotides of the ORF of OppA mRNA are involved in the polyamine stimulation of OppA synthesis. When the secondary structure of the above region of OppA mRNA was analyzed by optimal computer folding, it was shown that the degree of polyamine stimulation of OppA protein synthesis was dependent on the structure of the SD sequence in addition to its position. Loose base pairing of the SD sequence with other regions of the mRNA caused strong polyamine stimulation, while intense base pairing of the SD sequence with other regions of the mRNA resulted in insignificant or weak polyamine stimulation.

Polyamines, aliphatic cations present in almost all living organisms, are known to be necessary for normal cell growth (1). Their proliferative effects are probably caused by the stimulation of nucleic acid and protein synthesis. We previously reported that polyamines can stimulate some kinds of protein synthesis in both prokaryotic and eukaryotic cell-free systems (2,3) and in vivo (4,5). Furthermore, it has been reported that assembly of 30 S ribosomal subunits is stimulated by polyamines (6,7) and the fidelity of protein synthesis is increased by polyamines (8,9). We also found that most polyamines exist as a polyamine-RNA complex in cells, and that the amount of polyamines (spermidine plus spermine) bound to RNA in rat liver is about 2 mol/100 mol of phosphate of RNA (10). Under the condition of spermine-stimulated globin synthesis in a rabbit reticulocyte cell-free system (11), the amount of polyamine bound to RNA was very close to the estimated value in cells.
In Escherichia coli, synthesis of a protein (polyamine-induced protein) was strongly stimulated by the addition of putrescine to growing cells of a polyamine-requiring mutant MA261 (12). The protein was identified, by cloning the corresponding gene, as OppA 1 and is a periplasmic substrate-binding protein of the oligopeptide uptake system (13). In the present work, we have shown that stimulation of OppA synthesis by polyamines occurs mainly at the level of translation, and the position and secondary structure of the Shine-Dalgarno (SD) sequence (14) are probably involved in the stimulation of protein synthesis by polyamines.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Culture Conditions-E. coli MA261 (speB speC gly leu thr thi) was kindly provided by Dr. W. K. Maas, New York University School of Medicine. E. coli MA261 oppA::Km and lacZ::Em were prepared as described previously (15,16). These cells were grown at 37°C in medium A in either the presence (100 g/ml) or absence of putrescine (13). Methionine content in medium A was decreased from 100 to 3 g/ml in order to label proteins with [ 35 S]methionine; this modification did not influence their growth rate. Another polyaminerequiring mutant, HT283, was kindly provided by Dr. H. Tabor, National Institutes of Health, and was grown according to the method of Hafner et al. (17).
Plasmids-pACYCoppA, equivalent to pPI5.1, was prepared as described previously (13). The 4.3-kb HindIII-SalI fragment of pPI57 (13) was ligated into the same restriction sites of pUC119. Subsequently, the 3.2-kb oppA gene-containing SmaI fragment of the plasmid was ligated into the same restriction site of pMW119 purchased from Nippon Gene (pMW975). pMW211 was constructed by inserting the 3.4-kb NruI fragment of pACYCoppA into the SmaI site of pMW119. A PCR product was obtained using the HindIII-digested pMW975 as template, and 5Ј-CAAATAGGTTACCTGGT-3Ј and 5Ј-GGGGAATTCCAATCTCTATT-TGATTGA-3Ј as primers. pMW412 was constructed by inserting the 1.1-kb EcoRI-BstEII fragment of the PCR product into the same restriction sites of pMW211.
Site-directed mutagenesis by overlap extension using PCR (18) was performed to prepare pMWSD Ϫ and pMWSD*. The template used for first PCR was EcoRI digested pMW975. Primers used for first PCR were 5Ј-GGGGAATTCCGGGCAATAAGGGCGC-3Ј (P1, sequence for Ϫ958 to 942 of oppA gene), 5Ј-TTTTTGGACTCAATCATTATAATT-3Ј (complementary sequence for Ϫ27 to Ϫ4 of oppA gene except two underlined bases), 5Ј-AATTATAATGATTGAGTCCAAAAA-3Ј (sequence for Ϫ27 to Ϫ4 of the oppA gene except two underlined bases) and 5Ј-TGGCTC-CAGCCAAACCATTCT-3Ј (P2, complementary sequence for 1068 -1088 of the oppA gene) to construct pMWSD Ϫ , and were P1, 5Ј-TCATT-GTTTTTCCTACTCCCTCATT-3Ј (complementary sequence for Ϫ21 to 4 of oppA gene except three underlined bases), 5Ј-AATGAGGGAGTAG-GAAAAACAATGA-3Ј (sequence for Ϫ21 to 4 of oppA gene except three underlined bases) and P2 to construct pMWSD*, respectively. Then, a second PCR was performed using initial PCR products as templates and P1 and P2 as primers. Two PCR products thus obtained were digested with EcoRI and BstEII, and inserted into the same restriction sites of pMW211. The same method was applied to prepare pMWSD1 to pM-WSD5 with appropriate primers. * This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan, and by Research Aid from the Uehara Memorial Foundation, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) D83137.
To make pMW9-lacZ and pMW45-lacZ, PCR was performed using EcoRI digested pMW975 as template. Primers used were 5Ј-GGGG-GATCCGTCGCGAAAGAATAATGT-3Ј (P3) and 5Ј-AACGCCAGCTG-CAGCTAAACTTCTCT-3Ј to construct pMW9-lacZ, and P3 and 5Ј-ACT-GAACTTCTGCAGCATTGTTACGTA-3Ј to construct pMW45-lacZ. After the products were digested with BamHI and PstI, they were inserted into the same restriction sites of pUC119. The 3.1-kb PstI fragment containing the lacZ gene was obtained from pMC1871 fusion vector (19) and was then inserted into the same restriction site of the above plasmids (pUC9-lacZ and pUC45-lacZ). Then, the 3.4-and 3.5-kb KpnI-HindIII fragments of pUC9-lacZ and pUC45-lacZ were inserted into the same restriction sites of pMW119.
The strains and plasmids used in this study are listed in Table I. Determination of the Transcription Initiation Sites of OppA mRNA and Quantitative Measurement of the mRNA-The nucleotide sequence of the upstream region of the oppA gene was determined by the dideoxy method of Sanger et al. (20) using the M13 phage system. S1 nuclease mapping was performed according to the method of Ausubel et al. (21). A 5Ј-terminal 32 P-labeled 600-nucleotide fragment of single-stranded DNA (5 ϫ 10 4 cpm), which is complementary to the upstream region of OppA mRNA (from Ϫ579 to 21), was used as a probe. The probe was hybridized with 60 g of total RNA from E. coli MA261 prepared by the method of Emory and Belasco (22) and digested with S1 nuclease. The length of the remaining DNA fragment was determined by electrophoresis on a 5% polyacrylamide sequencing gel. OppA mRNA was measured by dot blot analysis (23) using a 32 P-labeled 319-nucleotide fragment of DNA (from 996 to 1314) as a probe. The DNA fragment was labeled with [␣-32 P]dCTP using TaKaRa BcaBEST labeling kit.
Purification of OppA and Preparation of Immunoglobulin for the Protein-OppA was purified from the 100,000 ϫ g supernatant of E. coli MA261 as described previously (12). Immunoglobulin for OppA was prepared as described previously (24) and partially purified from the antiserum by precipitation with 40% saturated (NH 4 ) 2 SO 4 .
Measurement of OppA Synthesis by an Immunoprecipitation Method-E. coli MA261 was grown in polyamine-deficient medium. When A 540 reached 0.15, the cells were divided into 5-ml aliquots and grown in the presence (100 g/ml) or absence of putrescine for 10 min. Then, [ 35 S]methionine (1 MBq) was added to each 5-ml aliquot, and the cells were allowed to grow for 20 min. They were harvested after the addition of methionine at a final concentration of 20 mM and disrupted by a French pressure cell at 20,000 p.s.i. containing 1 ml of buffer A (10 mM sodium phosphate, pH 7.4, 100 mM NaCl, 1% Triton X-100, and 0.1% SDS). The amount of OppA protein synthesized was determined using 1,000,000 cpm of [ 35 S]methionine-labeled protein according to the method of Philipson et al. (25) with some modifications (26). Radioactivity of labeled OppA protein was quantified using a Fujix Bas 2000II imaging analyzer.
Measurement of Polyamine Contents-Polyamine levels in E. coli were determined by high pressure liquid chromatography as described previously (27).
Prediction of the Secondary Structure of RNA-Optimal computer folding of the Ϫ65 to 65-nucleotide region of OppA mRNA or lacZ mRNA was performed according to the method of Zuker and Stiegler (28). Free energy (⌬G) for the formation of the secondary structure was calculated on the basis of the data of Turner et al. (29). Fig. 1, we confirmed polyamine stimulation of OppA synthesis in polyamine-requiring mutant MA261 cells by immunoprecipitation of [ 35 S]methionine-labeled OppA protein. If the oppA gene was disrupted, there was no OppA synthesis. When cells were transformed with pACYCoppA, a relatively high copy number plasmid, large amounts of OppA were synthesized, and polyamines did not stimulate OppA synthesis significantly. In contrast, when cells were transformed with pMW975, a low copy number plasmid, OppA synthesis was greatly stimulated by polyamines. These results indicate that polyamine stimulation of OppA synthesis is observed only when the copy number of the oppA gene is low. Under these conditions, putrescine and spermidine contents in cells grown in the presence and absence of 100 g/ml putrescine were 43.5 and Ͻ0.1 nmol/mg protein for putrescine and 5.45 and 0.45 nmol/mg protein for spermidine. Thus, it is clear that significant amounts of putrescine were converted into spermidine. The function of polyamines was not replaced by  Insertion of 975 nucleotides (nt) of 5Ј-upstream region (5Ј-UR) of the initiation codon ATG and ORF of oppA into pMW119 pMW412

Polyamine Stimulation of OppA Synthesis Is Dependent on Gene Copy Number-As shown in
Insertion of 412 nt of 5Ј-UR and ORF of oppA into pMW119 pMW211 Insertion of 211 nt of 5Ј-UR and ORF of oppA into pMW119 pMWSD Ϫ Nonexistence of SD sequence in pMW975 pMWSD* Shift of SD sequence from Ϫ17 ϳ Ϫ12 to Ϫ10 ϳ Ϫ7 in pMW975 pMWSD1 Site-directed mutagenesis of GGG(Ϫ55 ϳ Ϫ53) of stem I to AAA in pMW975 pMWSD2 Site-directed mutagenesis of GGG(Ϫ55 ϳ Ϫ53) of stem I to TTT in pMW975 pMWSD3 Site-directed mutagenesis of CCC(Ϫ63 ϳ Ϫ61) of stem II to AAA in pMW975 pMWSD4 Site-directed mutagenesis of CCC(Ϫ63 ϳ Ϫ61) of stem II to TTT in pMW975 pMWSD5 Site-directed mutagenesis of CCC(Ϫ63 ϳ Ϫ61) of stem II to TTT in pMWSD1 pMW 9-lacZ Fusion of oppA 5Ј-UR and ORF for 9 amino acids and lacZ ORF in pMW211 pMW 45-lacZ Fusion of oppA 5Ј-UR and ORF for 45 amino acids and lacZ ORF in pMW211 Mg 2ϩ or Ca 2ϩ (data not shown).
Determination of the Transcription Initiation Sites of the oppA Gene-First, we determined the nucleotide sequence upstream of the oppA gene. As shown in Fig. 2, insertion sequence 2 (IS2) (30) was found in the upstream region. Although IS2 was observed in another polyamine-requiring mutant HT283 (EWH319), it was not in E. coli W3110, the parental strain of MA261 and HT283. The upstream region of the gene also included the leucine responsive element observed in the upstream region of the oppA gene of Salmonella typhimurium (31). We confirmed leucine stimulation of OppA synthesis in E. coli HT283, but polyamine stimulation of OppA synthesis was not influenced by leucine (data not shown). The effect of leucine on OppA synthesis in E. coli MA261 could not be determined since leucine is necessary for cell growth in this strain.
Transcription initiation sites of the oppA gene were determined by S1 nuclease mapping (Fig. 3). In E. coli MA261, there were three initiation sites (P1, P2, and P3). However, initiation mainly occurred from P1, suggesting that IS2 has a strong promoter activity. When OppA mRNA was synthesized from pMW975, pMW412, and pMW211, the major initiation site of transcription was P2, P2, and P3, respectively. It remains to be clarified why P1 is not the initiation site in pMW975.
Polyamine Stimulation of OppA Synthesis at the Translational Level-To determine the level of polyamine stimulation of OppA synthesis, the amount of OppA mRNA and OppA synthesis were measured by dot-blotting of RNA and immunoprecipitation of [ 35 S]methionine-labeled OppA protein (Fig. 4). When OppA mRNA was transcribed from P1, polyamines significantly stimulated OppA mRNA synthesis (3.2-fold). When OppA mRNA was transcribed from P2 and P3, polyamines only slightly stimulated the OppA mRNA synthesis (1.2-1.4-fold). In the region upstream of P2 and P3, there were no typical Ϫ35 and Ϫ10 promoter regions, but transcriptional activity was stronger from P2 than from P3. OppA synthesis measured by immunoprecipitation of [ 35 S]methionine-labeled OppA protein was stimulated by polyamines with all versions of the OppA mRNAs (Fig. 4). The degree of polyamine stimulation relative to the transcription initiation site was in the order P1 Ͼ P2 Ն P3. This result indicates that polyamine stimulation of OppA synthesis occurs mainly at the level of translation, and that transcription from P1 was also stimulated by polyamines. It also suggests that the 171-nucleotide 5Ј-UTR is enough to cause polyamine stimulation of OppA synthesis.
Determination of RNA Structure Necessary for Polyamine Stimulation of OppA Synthesis at the Translational Level-We previously reported that polyamines stimulate protein synthesis mainly at the level of translational initiation (2). For initiation of protein synthesis, the most important elements in the mRNA are the initiation codon AUG and SD sequence. When the SD sequence was removed from the OppA mRNA, there was no significant OppA synthesis (Fig. 5). These results confirmed the importance of the SD sequence in protein synthesis (32)(33)(34). The SD sequence of the OppA mRNA was relatively distant (12 nucleotides) from the AUG compared with other mRNAs, in which the typical position of SD sequence was 7 nucleotides upstream from the AUG. Therefore a new SD sequence was inserted 7 nucleotides upstream from the AUG. When OppA was synthesized from the mRNA with the new SD sequence, polyamine stimulation was observed, but to a lesser degree (pMWSD*, 2.3-fold) (Fig. 5). However, OppA synthesis from the mRNA with the new SD sequence in the absence of putrescine was greater than that from the normal OppA mRNA. The results suggest that the position of the SD sequence may influence polyamine stimulation of OppA synthesis.
We next examined whether the 5Ј-UTR of the OppA mRNA is sufficient for polyamine stimulation of OppA synthesis. As shown in Fig. 6, synthesis of ␤-galactosidase from the chromosomal lacZ gene was not stimulated by polyamines. When the 5Ј-UTR and 27 nucleotides encoding first 9 amino acids for OppA protein was fused to ␤-galactosidase (9-lacZ mRNA), the degree of polyamine stimulation of ␤-galactosidase synthesis was 1.4-fold. The length of OppA mRNA used for the construction of a fused mRNA was then increased to include the 5Ј-UTR and 135 nucleotides encoding first 45 amino acids (45-lacZ mRNA). As a result, the degree of polyamine stimulation increased to 4.2-fold. Possible secondary structure of the initiation codon surrounding region (Ϫ65 to 65) of OppA mRNA and 9-lacZ mRNA was then compared (Fig. 7). Stability of the mRNAs was nearly equal (Ϫ46.6 and Ϫ47.8 kcal/mol). The SD sequence of the 9-lacZ mRNA was tightly base paired. However, the SD sequence of OppA mRNA, which is equivalent to  5. Effect of SD sequence on polyamine stimulation of OppA synthesis. pMWSD Ϫ which encodes OppA mRNA without the SD sequence, and pM-WSD* which encodes OppA mRNA having the SD sequence 7 nucleotides upstream from the AUG were prepared as described under "Experimental Procedures." OppA protein synthetic activities directed by these OppA mRNAs were compared with that directed by normal OppA mRNA .  FIG. 7. Possible secondary structure of OppA mRNA (A) and 9-lacZ mRNA (B). Optimal computer folding of the 130 nucleotides (-65 to 65) of mRNAs was performed by the method of Zucker and Stiegler (28). The structure of 45-lacZ mRNA at this region is the same as that of OppA mRNA (A). Initiation codon AUG is circled and the SD sequence is boxed. Continuous GC stems I and II are shaded.
FIG. 6. Effect of the 5-UTR of OppA mRNA on polyamine stimulation of ␤-galactosidase synthesis. Fused mRNA having oppA 5Ј-UTR and ORF for 9 amino acids and lacZ ORF or having oppA 5Ј-UTR and ORF for 45 amino acids and lacZ ORF was prepared as described under "Experimental Procedures." Synthesis of ␤-galactosidase was measured by the same method as that of OppA protein synthesis. Antiserum for ␤-galactosidase was obtained from Sigma. 45-lacZ mRNA, was loosely base paired. This suggests that the secondary structure of the SD sequence may play an important role in polyamine stimulation of OppA synthesis.
The secondary structure of the SD sequence was then changed by site-directed mutagenesis at positions of the oppA gene corresponding to the 5Ј-UTR of OppA mRNA (Table I).
Since polyamines bind to double-stranded RNA, especially GCrich RNA rather than to single-stranded RNA (10,35), the GC-rich double-stranded region of the 5Ј-UTR of OppA mRNA (stems I and II of Fig. 7A) was mutated. As shown in Table II, the change of the secondary structure of the SD sequence was observed by the mutation on stem I (pMWSD1 and 2) or stems I and II (pMWSD5), but not by the mutation on stem II (pM-WSD3 and 4). When the SD sequence was loosely base-paired as in mRNAs synthesized from pMW975, pMWSD4, pMWSD3, pMWSD5, and pMWSD1, polyamines significantly stimulated OppA synthesis (Table II). The degree of polyamine stimulation was 3.7-5.1-fold. The amounts of OppA mRNA from the cells cultured with putrescine were nearly equal to those from the cells cultured without putrescine. When the SD sequence was tightly base paired as in mRNAs synthesized from pMW9-lacZ and pMWSD2, polyamines did not influence protein synthesis significantly. The secondary structure of the AUG region was not correlated with the polyamine stimulation of OppA synthesis (Fig. 7). 2

DISCUSSION
Polyamines can stimulate synthesis of proteins such as OppA protein (1,2) and ribosomal proteins (26), which are important for cell growth. In this communication, the molecular mechanism of polyamine stimulation of OppA synthesis has been studied. It was found that polyamine stimulation occurs mainly at the translational level.
In E. coli cells, IS2 was inserted at some stage of evolution. IS2 enhanced transcriptional efficiency of oppA gene and the IS2-dependent transcription was stimulated by polyamines. This also contributed to the polyamine stimulation of OppA synthesis. In E. coli W3110 lacking IS2 at the upstream region of oppA gene, the transcription started mainly from P2 and the efficiency was low because only weak Ϫ35 and Ϫ10 regions exist upstream of P2 (Fig. 2). Thus, IS2 was inserted at some stage of evolution by chance so that OppA protein was synthesized more effectively in E. coli MA261 and HT283. A plasmid containing 461 nucleotides of the 3Ј-end of IS2 (pMW975) led to transcriptional initiation starting from P2 (Fig. 3). This was unexpected, and it remains to be clarified why P1 is not the initiation site in pMW975. The whole sequence of IS2 may be necessary as a signal for transcriptional initiation.
When transcription started from P2 or P3, polyamines did not influence transcriptional efficiency significantly. Thus, we could study the polyamine effect on OppA synthesis at the level of translation using E. coli MA261oppA::Km/pMW975 or pMW211. When pMW975 and pMW211 were used as the template for OppA mRNA synthesis, the size of the 5Ј-UTR was 266 and 171 nucleotides, respectively. Although the polyamine effect on OppA protein synthesis was slightly greater with pMW975 than with pMW211, essentially the same results were obtained with both plasmids. The results with pMW211 were only shown with fused mRNAs containing the 5Ј-UTR of OppA mRNA and the open reading frame of lacZ mRNA.
For initiation of protein synthesis in E. coli, the most important elements in the mRNA are the initiation codon AUG and SD sequence. The latter can base pair with the 3Ј-end of 16 S rRNA so that translational efficiency increases. Thus, translational efficiency decreases if the SD sequence undergoes intrastrand base pairing. The polyamine effect on OppA synthesis was examined by changing the secondary structure of the SD sequence and the AUG region through site-directed mutagenesis at positions of oppA gene corresponding to the 5Ј-UTR of OppA mRNA. Although the secondary structure of the AUG region did not influence the polyamine stimulation of OppA protein synthesis, that of SD sequence did. When the SD sequence was loosely base paired with another region of OppA mRNA, polyamines significantly stimulated protein synthesis. When the SD sequence was strongly base paired, polyamines did not influence protein synthesis. Since polyamines bind to double-stranded RNA more strongly than to single-stranded RNA (10,35), it is to be expected that a strongly base paired SD sequence would become further stabilized by spermidine binding. It is also noted that the disappearance of the GC-rich stem I in OppA mRNA synthesized from pMWSD2 caused no polyamine stimulation of OppA synthesis. Our results suggest that polyamines may contribute to unwinding weak secondary structure of the SD sequence through their binding to a region(s) such as stem I on the OppA mRNA close to SD sequence. However, an alternative explanation may also be possible. Experiments are in progress to determine how polyamines change the secondary structure of OppA mRNA.
In contrast, polyamines are known to inhibit some kinds of protein synthesis like ribosome modulation factor and OmpC protein (36). It is of interest to know the secondary structure of the SD sequence of these mRNAs. In case of eukaryotic protein synthesis, polyamines can regulate the initiation complex formation of Met-tRNA i •mRNA•40 S ribosomal subunits positively and initiation factor-dependent RNA helicase activity negatively (37). Thus, polyamine regulation of protein synthesis is dependent on the size and base composition of the 5Ј-UTR of the mRNA. There is a tendency that polyamines regulate  Fig. 7).
b Relative amount of OppA or ␤-galactosidase was shown as the amount compared with that of OppA synthesized in E. coli MA261 transformed with pMW975 in the absence of putrescine. protein synthesis directed by the mRNAs having long 5Ј-UTR in both prokaryotes and eukaryotes. We consider that these studies will help to establish the physiological importance of polyamines.