Translational Enhancement by an Element Downstream of the Initiation Codon in Escherichia coli*

The translation initiation of Escherichia coli mRNAs is known to be facilitated by a ciselement upstream of the initiation codon, called the Shine-Dalgarno (SD) sequence. This sequence complementary to the 3′ end of 16 S rRNA enhances the formation of the translation initiation complex of the 30 S ribosomal subunit with mRNAs. It has been debated that acis element called the downstream box downstream of the initiation codon, in addition to the SD sequence, facilitates formation of the translation initiation complex; however, conclusive evidence remains elusive. Here, we show evidence that the downstream box plays a major role in the enhancement of translation initiation in concert with SD.

Sprengart and co-workers (1,2) have proposed that in genes 0.3 and 10 of bacteriophage T7, a specific region located downstream of the initiation codon serves as an independent translational signal. This region of the RNA sequence designated as the downstream box (DB) 1 is complementary to bases 1469 -1483 within the Escherichia coli 16 S rRNA (anti-DB sequence). It is speculated that formation of a duplex between the DB and anti-DB of 16 S rRNA is responsible for translational enhancement (2). The DB sequence has also been implicated in the translation of the c1 mRNA, an mRNA that lacks any untranslated region and the SD sequence (3). Interestingly, c1 translation was enhanced at 42°C in a temperature-sensitive strain in which the amount of ribosomal protein S2 decreased at 42°C. It was proposed that the anti-DB sequence in S2deficient ribosomes indirectly becomes more accessible to DB, resulting in enhancement of translation initiation of the c1 mRNA (3,4). However, the role of the DB in c1 translation initiation was disputed by Resch and co-workers (5). These authors constructed lacZ translational fusions with the c1 gene to test the DB function. Since a deletion of 6 bases encompassing a portion of the DB sequence did not reduce the formation of the translation initiation complex, they disputed the existence of DB. Despite these elusive roles of DB (6), we have shown that the presence of a DB sequence in cold-shock mRNAs plays an important role in translation efficiency, and we proposed that the DB is involved in the formation of a stable initiation complex at low temperature before the induction of cold ribosomal factors (7). Furthermore, we pointed out the SELEX enrichment of DB-like sequences in an mRNA when the 30 S ribosomal subunit was used as a ligand (8,9).
It was suggested that the results obtained by Resch et al. (5) could be explained by recreating a new DB as a result of the deletion of the original DB in the c1 mRNA (6). Indeed, a 6-base deletion eliminating 5 out of 8 matches in the original DB recreated a new 9-base matching DB sequence including the initiation codon. Therefore, the results by Resch and collaborators (5) could be explained by the newly created DB. In the present paper, we performed both biochemical and genetic experiments to reexamine the role of the DB in translation initiation, and we determined that DB plays a crucial role in regulation of gene expression in E. coli by enhancing the formation of the translation initiation complex.
␤-Galactosidase Activity-E. coli AR137 (pcnB Ϫ ) (11) or JM83 (pcnB ϩ ) harboring different plasmids were grown at 37°C to mid-log phase in 20 ml of LB medium containing 50 g/ml ampicillin in a 125-ml flask. The cultures were then transformed to a 15°C shaking water bath, or isopropyl-␤-D-thiogalactopyranoside (IPTG, 1 mM) was added to a final concentration of 1 mM. A 100-l culture was taken at each time point. ␤-Galactosidase activity was measured according to Miller's procedure (12).
Primer Extension-E. coli AR137 (pcnB Ϫ ) or JM83 (pcnB ϩ ) carrying different plasmids were grown under the same condition as used for the ␤-galactosidase assay described above. To estimate the mRNA amounts of the lacZ fusion constructs, 1.5 ml of culture was taken at each time point, and total RNA was extracted by the hot phenol method (13).
For the mRNA stability experiments at 15°C, rifampicin (0.2 mg/ml) was added 30 min after the temperature downshift from 37°C. The mRNA amounts from the cspA-lacZ fusion constructs were estimated by primer extension using the 32 P-labeled M13-47 antisense primer complementary to the region of lacZ between codons 14 and 22 as described previously (7). The reverse transcription reaction was carried out with AMV-RT according to the manufacturer's procedure (Boehringer Mannheim), and the cDNA products were resolved on a 6% Sequencing Gel and quantified by PhosphorImager (Bio-Rad).
Pulse Labeling-Cultures of E. coli AR137 (pcnB Ϫ ) cells carrying pINZ or pINZDB1 were grown at 37°C under the same conditions used for the ␤-galactosidase assay. IPTG (1 mM) was added at mid-log phase to each culture. At each time point, 1 ml of the culture was labeled for 5 min with 100 Ci of trans-[ 35 S]methionine (1, 175 Ci/mmol) (NEN Life Science Products) as described previously (14). Cell extracts from each time point were loaded on a 5% SDS-PAGE, and ␤-galactosidase synthesis was measured by PhosphorImager.
Ribosome Isolation-Cultures of E. coli JM83 (pcnB ϩ ) cells carrying pINZ or pINZDB1 were grown in 600 ml of LB medium in a 4-liter flask under the same condition as described above. At mid-log phase IPTG (1 mM) was added to each culture. Ribosomal particles were isolated by the procedure described by Dammel and Noller (15) with some modifications as follows: a 100-ml aliquot from the original culture was taken at each time point, and chloramphenicol was added to a final concentration of 0.1 mg/ml to stop cell growth. Cells were immediately collected by centrifugation (5,000 ϫ g for 10 min at 4°C), resuspended in buffer 1 (20 mM Tris-HCl (pH 7.6), 10 mM MgCl 2 , 100 mM NH 4 Cl, 6 mM ␤-mercaptoethanol and 1 mg/ml lysozyme), and frozen at Ϫ80°C for few hours. The cells were lysed by the freeze-thaw method (16). The cell extracts (0.5 ml) were then layered on top of a 5-40% (w/w) sucrose gradient (7.5 ml), and the polysomes and ribosomal subunits were separated by centrifugation at 151,000 ϫ g for 2.5 h at 4°C using a Beckman SW-41 rotor. The polysome profiles were detected by a fast protein liquid chromatography system, and a total of 15 fractions of 0.5 ml each were collected.
Detection of the lacZ-mRNA-From each polysome fraction (0.5 ml) 0.2 ml was spotted on a nitrocellulose membrane using the Minifold II Slot-Blot System (Schleicher & Schuell). The lacZ-mRNA was detected by hybridization using the 32 P-labeled M13-47 primer (17), and the amount of lacZ-mRNA was estimated by PhosphorImager.
In Vitro Translation-Using the E. coli S30 Extract System for Linear Templates Kit (Promega), the transcription-translation coupled reaction was carried out according to the manufacturer's protocol as follows. To 20 l of pre-mix containing all the amino acids except for methionine, 10 Ci of trans-[ 35 S]methionine (1,175 Ci/mmol; NEN Life Science Products) and the E. coli S30 extracts, 160 ng of pINZ or pINZDB1 (1 l), were added, and the mixture was incubated at 37°C for 45 min. The products were precipitated with acetone and analyzed by 15% SDS-PAGE.

RESULTS
The Effect of a Perfectly Matching DB-It has been previously shown that DB is essential for the production of CspA at low temperature (7). However, the wild-type DB of cspA has 10 matches out of 15 possible matches to the anti-DB in 16 S rRNA (7). Therefore, we added DBs of 12 (pJJG78DB1) or 15 (pJJG78DB2) bases that are complementary with the anti-DB of 16 S rRNA to the site after the 5th codon of lacZ under the cspA regulatory system in pJJG78 (see Fig. 1A) to examine if they enhance lacZ expression at 15°C. Mid-log phase cells (pcnB Ϫ ) grown at 37°C were shifted to 15°C, and ␤-galactosidase activity was measured at 1, 2, and 3 h after the shift. Fig. 1B shows that at 1 h at 15°C the ␤-galactosidase activity was 3-and 8-fold higher with pJJG78DB1 and pJJG78DB2, respectively, than with pJJG78. After 2 and 3 h at 15°C, the ␤-galactosidase activity was increased 3.5 and 10.5 times with pJJG78DB1 and pJJG78DB2, respectively, than with pJJG78 ( Fig. 1B). Moreover, the effect of the DB was observed at 37°C in which the ␤-galactosidase activity of pJJG78DB1 and pJJG78DB2 was 2-and 4-fold higher as compared with pJJG78. The amount of the lacZ mRNA ( Fig. 1C) at each time point as well as the mRNA stability ( Fig. 1D) did not vary significantly between these constructs. The lacZ mRNA halflife from pJJG78, pJJG78DB1, and pJJG78DB2 was calculated to be 27, 23, and 25 min, respectively. In addition, computer analysis (18) revealed no significant differences in the mRNA secondary structures among pJJG78, pJJG78DB1, and pJJG78DB2, suggesting that the insertion of the perfectly matching DB may not have a particular effect in the mRNA secondary structures that could account for the difference in their ␤-galactosidase expression. These results indicate that DB functions as a translational enhancer and that greater The cspA gene structure from its 5Ј end is shown at the top. pJJG78DB1 and pJJG78DB2 were constructed from pJJG78 as described under "Experimental Procedures." The DB sequences of pJJG78DB1 (12 matches) and pJJG78DB2 (15 matches) are shown at the bottom. B, ␤-Galactosidase activity of the cspA-lacZ fusion constructs after cold shock at 15°C. E. coli AR137 cells transformed with pJJG78, pJJG78DB1, or pJJG78DB2 were grown in LB medium, and at mid-log phase (A 600 ϭ 0.4) cultures were shifted from 37 to 15°C. ␤-Galactosidase activity was measured before (time 0) and 1, 2, and 3 h after the shift. C, detection of the cspA-lacZ mRNAs. Total RNA from E. coli AR137 cells carrying pJJG78, pJJG78DB1, or pJJG78DB2 was extracted at the same time points indicated above (B) and used as a template for primer extension as described under "Experimental Procedures." D, mRNA stability from the cspA-lacZ constructs. E. coli AR137 cells transformed with pJJG78, pJJG78DB1, and pJJG78DB2 were grown as described above. At mid-log phase the cultures were shifted to 15°C, and after 30 min rifampicin was added to a final concentration of 0.2 mg/ml (time 0). Total RNA was extracted at 5, 10, and 40 min after rifampicin addition. The cspA-lacZ mRNAs were detected by primer extension as described under "Experimental Procedures." complementarity to the anti-DB improves translational efficiency and/or that specific base pairings like the first three nucleotides of the DB from pJJG78DB2 may play an important role for the DB activity.
DB Functions at 37 o C-The experiments described above were carried out at 15°C. In order to examine whether DB also works at 37°C, the cspA cold-shock regulatory regions upstream of SD of pJJG78 and pJJG78DB2 were replaced with the constitutive lpp promoter and the lac promoter-operator fragment using a pINIII vector (19), yielding pINZ and pIN-ZDB1, respectively ( Fig. 2A). Cells (pcnB Ϫ ) transformed with pINZ or pINZDB1 showed very low ␤-galactosidase activity in the absence of IPTG, an inducer of the lac promoter (Fig. 2B,  time 0). Upon the addition of 1 mM IPTG, ␤-galactosidase activity was induced in both cells. After 3 h induction, ␤-galactosidase activity increased 18-and 37-fold for pINZ and pIN-ZDB1, respectively (Fig. 2B). However, the levels of ␤-galactosidase activity show a dramatic difference between the two; the activity with DB (pINZDB1) was 34 times higher than that without DB (pINZ), demonstrating that DB functions at 37°C as well. Specific activities of ␤-galactosidase produced from vector pINZ and pINZDB1 are almost identical (data not shown), and thus the addition of the five amino acid residues in the ␤-galactosidase sequence of pINZDB1 (due to DB) does not affect the enzymatic activity. Furthermore, the stabilities of ␤-galactosidase from pINZ and pINZDB1 are also identical with a half-life of approximately 3 h (data not shown).
SD Requirement for DB Function-Next, we examined whether DB functions independently from SD, the initial ribosome-binding site. For this purpose, the SD sequence of pIN-ZDB1, GAGG was changed to GCCC, yielding pINZDB2 (Fig. 2,  A and C). The ␤-galactosidase level of pINZDB2 induction was reduced to 1/7 of that of pINZDB1 but still 3-fold higher than that of pINZ at 3 h after IPTG induction (Fig. 2B). However, since pINZDB2 has the second AUG codon 6 codons down-stream as a result of DB insertion, it might serve as a secondary initiation codon for the lacZ gene. Indeed, the N-terminal sequence analysis showed that 100% of the ␤-galactosidase produced from pINZDB2 is initiated at the second AUG codon as compared with pINZDB1, which is more than 90% initiated at the first AUG codon (data not shown). Furthermore, the second AUG codon is presided by a potential but poor SD sequence (AAGG) at the region corresponding to the 2nd and 3th codons (underlined; Fig. 2C). Indeed, when this secondary SD was removed by deletion of the 15-base sequence (codons 1-5; pINZDB3 in Fig. 2A), ␤-galactosidase activity at all time points was reduced to the background level (Fig. 2B), indicating that the secondary SD played a crucial role in the translation of the pINZDB2 lacZ mRNA. When the SD sequence was recreated by 5-base substitution in pINZDB3 (pINZDB4; Fig. 2C), ␤-galactosidase activity of this construct was recovered to a comparable level to that of pINZDB1 (Fig. 2B). It is important to notice that the DB sequence starting from the first AUG codon was eliminated in pINZDB4 (Fig. 2C). Therefore, the high expression of ␤-galactosidase from pINZDB1 and pIN-ZDB4 is due to the perfectly matching DB sequence (Fig. 2B). These results indicate that (a) DB functions only in the presence of SD, and (b) the position of DB is flexible starting from either codon 1 or 6.
Enhancement of Protein Synthesis by DB-The ␤-galactosidase activity shown in Fig. 2B indicates that DB enhances the translation of pINZDB1. Therefore, in order to test the effect of the DB in translation efficiency, the rate of ␤-galactosidase synthesis from pINZ and pINZDB1 was analyzed. The rate of ␤-galactosidase synthesis was measured by pulse labeling cells for 5 min with [ 35 S]methionine after the addition of IPTG using cells harboring pINZ and pINZDB1. After SDS-PAGE, the amounts of radioactive ␤-galactosidase were estimated using a PhosphorImager (Fig. 3). Prior to the addition of IPTG, the rate of ␤-galactosidase synthesis from pINZ and pINZDB1 was FIG. 2. Perfectly matching DB enhances translation at 37°C. A, pIN-lacZ constructs. The XbaI-SalI fragment from pJJG78 or pJJG78DB2 was inserted into the XbaI-SalI sites of pIN-III to create pINZ and pINZDB1, respectively, which then were used to create pINZDB2, pINZDB3, and pINZDB4 as described under "Experimental Procedures." B, ␤-Galactosidase activity of the pINZ-lacZ constructs. Cultures of E. coli AR137 cells transformed with pINZ, pINZDB1, pINZDB2, pINZDB3, and pINZDB4 were grown at 37°C under the same conditions described in Fig. 1. IPTG (1 mM) was added at mid-log phase to each culture. ␤-Galactosidase activity was measured before (time 0) and at 0.5, 1, 2, and 3 h after IPTG addition. C, mRNA sequences of the pIN-lacZ constructs showing the position of SD, AUG, and DB. The lacZ in pJJG78 has a 10-match DB. The perfectly matching DB located after the 5th codon has 16 residues complementary with the anti-DB.
identical. However, upon IPTG induction the rates of ␤-galactosidase synthesis from pINZDB1 were continuously increasing at each time point, whereas the rate of ␤-galactosidase synthesis from pINZ was almost not affected. After 4 h of IPTG addition the rate of ␤-galactosidase synthesis from pINZDB1 was 6.5 times higher than that of pINZ. This result demonstrates that DB enhances the translation efficiency of pIN-ZDB1 as reflected by the increment in the synthesis of ␤-galactosidase.
Enhancement of Translation Initiation by DB-In order to examine whether DB enhances translation initiation, we next analyze the ability of lacZ mRNA from pINZ and pINZDB1 to form polysomes. For this experiment, pcnB ϩ cells were used to amplify the effect of DB. Interestingly, cells with pINZDB1 could not form colonies on LB plates in the presence of 1 mM IPTG, whereas cells with pINZ formed colonies. The lethal effect of IPTG on the cells with pINZDB1 is considered to be due to overexpression of ␤-galactosidase. After the addition of IPTG, cell growth was stopped by the addition of chloramphenicol (0.1 mg/ml) at 15, 30, and 60 min, and then polysome profiles were examined as shown in Fig. 4. From each gradient fraction (500 l), 200 l were spotted on a nitrocellulose membrane, and the amount of the lacZ mRNA analyzed using a 24-base antisense oligonucleotide (M13-47 oligonucleotide). The amounts of the lacZ mRNA were quantified by a Phosphor Imager and are displayed in Fig. 4. Although the polysome profiles are similar, there are significant differences in the distribution of the lacZ mRNA; at 15 min the lacZ mRNA mainly exists in the upper half of the gradient (fraction 8 -14, corresponding to 70 S to 30 S ribosomes) with pINZ, while with pINZDB1 a major peak (fraction 3 to 8) is formed in the lower half of the gradient. At 30 min, the lacZ mRNA from pINZ moved to the position of the 70 S ribosome, whereas the lacZ mRNA from pINZDB1 maintained a similar pattern as that at 15 min. At 60 min a major fraction of the lacZ mRNA from pINZ remained in the upper half of the gradient, whereas the lacZ mRNA from pINZDB1 was broadly distributed from higher order polysomes to 70 S ribosome fraction. Therefore, the reason why cells harboring pINZDB1 could not form colonies on LB plates containing 1 mM IPTG may be due to a decrease in the concentration of free ribosomes as a result of the massive expression of a highly translatable DB-containing mRNA (20). These results indicate that DB enhances the efficiency of polysome formation probably due to a translation initiation enhancement.
In order to estimate the exact effect of DB from the above experiment, the amount of the lacZ mRNA and the ␤-galacto-sidase activity were measured at the same time points taken in the polysome profiles (15,30, and 60 min after IPTG induction). As shown in Fig. 5A, the amounts of the lacZ mRNA reached almost the maximal level at 15 min for both pINZ and pIN-ZDB1. The PhosphorImager analysis of this result revealed that the amounts of the pINZDB1 mRNA are 1.5, 1.4, and 1.3 times higher than those of the pINZ mRNA at 15, 30, and 60 min, respectively. The higher mRNA levels for pINZDB1 are probably attributable to the highly efficient polysome formation of pINZDB1 that may stabilize the mRNA (21). The induction of ␤-galactosidase activity is shown in Fig. 5B. In the case of pINZDB1, the activity is very high even in the absence of IPTG, and upon the addition of IPTG, it increased from 18,500 to 64,400 units (3.5-fold) after 2.5 h incubation. In the case of pINZ, the background activity prior to IPTG induction was much lower, and it increased from 900 to 2,900 units (3-fold) at the 2.5-h time point. The increment of the ␤-galactosidase activity of pINZDB1 between 30 and 60 min is 35 times higher than that of pINZ, and therefore the efficiency of ␤-galactosidase production for pINZDB1 is calculated to be 26 times higher than that for pINZ on the basis of the amount of mRNA. Therefore, the higher levels of ␤-galactosidase production from pINZDB1 are due to a high efficiency of polysome formation.
Next, in order to demonstrate more directly the translationenhancement effect of DB, the ␤-galactosidase synthesis was examined in a cell-free system using pINZ and pINZDB1. The Fig. 6A) was 8-fold higher than that with pINZ (1st lane), whereas the ␤-lactamase (band L) production was almost identical in both lanes. Fig. 6B shows a time course of in vitro production of ␤-galactosidase from pINZ and pINZDB1 performed as described above. The same reaction was carried out with non-radioactive methionine, spotted on a nitrocellulose membrane, and hybridized with the M13-47 oligonucleotide as described under "Experimental Procedures." As shown in Fig. 6C, at each time point the amount of lacZ mRNA from pINZ and pINZDB1 was almost identical. This result supports the role of the DB as a translational enhancer from the in vivo data described above.
Further Enhancement of DB-dependent Translation by S2 ts -It has been proposed that in the absence of ribosomal protein S2, structural changes in 16 S rRNA result in the release of the anti-DB sequence from the penultimate stem making it more accessible to base pair with DB (3,4). We analyzed the ␤-galactosidase expression of pINZ, and pINZDB1 in E. coli CS239 that carries an S2 temperature-sensitive mutation (3). Fig. 7A shows that the ␤-galactosidase activity of pINZDB1 significantly increases upon shifting the temperature from 30 to 42°C in the S2 ts strain (CS239) (6.3-fold from 0 to 3.5 h), whereas the activity in the wild-type strain (CS240) slightly increased (1.1-fold from 0 to 3.5 h). If the initial ratio of the activity of CS239 to that of CS240 at time 0 is taken as 1, the ratio dramatically increased, reaching 5.8 at 3.5 h after temperature shift (Fig. 7B). In contrast, the lacZ gene without DB, pINZ, did not show any significant differences in its expression between CS240 and CS239, and the ratios of the activity of CS239 to that of CS240 remained also at the initial level throughout the incubation time (Fig. 7B). A similar experiment was carried out with pINZDB3 (SD Ϫ , DB ϩ ), and the ratio of the activity in CS239 to that in CS240 increased 3.4fold at 3.5 h after the temperature shift (data not shown). These results clearly demonstrate that the low levels of S2 protein at 42°C causes significant stimulation of the lacZ expression only if the lacZ gene contains DB, consistent to the proposal of Shean and Gottesman (3) .   FIG. 3. Rate of ␤-galactosidase synthesis of the pINZ-lacZ constructs. Cultures of E. coli AR137 cells carrying pINZ or pINZDB1 were grown at 37°C under the same conditions described above. IPTG (1 mM) was added at mid-log phase to each culture. Rate of ␤-galactosidase synthesis was measured before (time 0) and 0.5, 1, 2, 3, and 4 h after IPTG addition. Cells were pulse-labeled with trans-[ 35 S]methionine as described under "Experimental Procedures." Cell extracts from each time point were analyzed by 5% SDS-PAGE, and the ␤-galactosidase synthesis was measured by PhosphorImager. The rate of ␤-galactosidase synthesis from pINZ (filled circles) and pINZDB1 (open circles) is shown at each time point.

DISCUSSION
The present results clearly demonstrate that there is a cis element downstream of the initiation codon, which in concert with SD plays an important role in the translation efficiency of mRNAs in E. coli. Our results provide supporting evidence for the DB hypothesis (1, 2) that DB forms a complex with the anti-DB in the 16 S rRNA to enhance translation initiation of DB-containing mRNA.
The present data reveal the following five features of the DB function. (a) The ␤-galactosidase activity of the pJJG78DB2 is 2-3-fold higher than that of the pJJG78DB1, indicating that a better complementarity between DB and anti-DB yields better translation and/or that the first 3 residues (AUG) in the DB are required for the better activity of the DB. (b) The position of DB in the mRNA is quite flexible as DB can start from codon 1 (pINZDB1) or codon 6 (pINZDB4). (c) DB itself works very poorly in the absence of SD, suggesting that the formation of translation initiation complex could be first initiated by the SD-16 S rRNA interaction, which subsequently leads to the DB-anti-DB interaction. Alternatively, the SD-16 S rRNA interaction could be stabilized by DB-anti-DB base pairing. It has been shown that the 30 S ribosomal subunit can bind to mRNA in the absence of initiator tRNA to form an intermediate translation initiation complex (22). This intermediate complex could be stabilized by the interaction of DB with anti-DB. (d) DB enhances the translation initiation as judged by increased mRNA translational efficiency both in vivo and in vitro as observed by the increase of polysome formation and ␤-galactosidase production in a cell-free system. (e) In the absence of S2 protein, the DB function is enhanced, consistently with the notion that anti-DB becomes more accessible to DB (3,4). Since that the major conclusions for the behavior of the DB in translation have been made from artificially created DBs in overexpression systems as shown above, it would be important to analyze the effect of DB under more physiological conditions. In this regard, we have reported earlier that the DB is crucial for FIG. 4. Ribosomal fractionation of E. coli JM83 cells transformed with pINZ or pINZDB1. Ribosomal particles were isolated as described by Dammel and Noller (15). Cultures of E. coli JM83 cells carrying pINZ or pINZDB1 were grown at 37°C in LB medium containing 50 g/ml ampicillin. At mid-log phase (A 600 ϭ 0.4) 1 mM of IPTG was added to each culture. Chloramphenicol (0.1 mg/ml) was added at 15, 30, and 60 min after IPTG addition. The cell extracts prepared as described under "Experimental Procedures" were then layered on top of a 5-40% (w/w) sucrose gradient. The polysomes and ribosomal subunits were separated by centrifugation at 151,000 ϫ g for 2.5 h at 4°C. The polysome profiles were then detected by using a fast protein liquid chromatography system. 0.2 ml from each fraction (0.5 ml) were spotted on a nitrocellulose membrane using the Minifold II Slot-Blot System (Schleicher & Schuell). The lacZ mRNA was detected by hybridization using the 32 P-labeled M13-47 as described under "Experimental Procedures." PhosphorImager values from the hybridization are plotted at the right. The pINZ and pINZDB1 mRNAs are shown in closed and open squares, respectively.
FIG. 5. Translational enhancement by a perfectly matching DB at 37°C. A, estimation of pINZ and pINZDB1 mRNAs. Cultures of E. coli JM83 carrying pINZ or pINZDB1 were grown at 37°C under the same conditions described in Fig. 4. Total RNA extracted at 15, 30, and 60 min after IPTG (1 mM) addition was used as a template for primer extensions according to the procedure described previously (7). B, ␤-galactosidase activity of pINZ and pINZDB1 in multi-copy expression system. E. coli JM83 cells transformed with pINZ or pINZDB1 were grown at 37°C under the same condition described in Fig. 4. ␤-Galactosidase activity was measured before (time 0) and 0.5, 1, 1.5, 2, and 2.5 h after IPTG (1 mM) addition (open circles and squares). Closed circles and squares represent the activities in the absence of IPTG. the induction of the cold-shock protein A (CspA) at low temperature and, furthermore, that major cold-shock genes contain DB sequences (7). Moreover, we have recently postulated that the DB is essentially required for the induction of major cold-shock proteins under conditions completely blocking protein synthesis at low temperature (23). Future experiments testing artificially created DBs using single copies of DB-containing lacZ mRNAs would greatly support the conclusions about the role of the DB as a translational enhancer.
It should be noted that there are a number of mRNAs without the 5Ј-untranslated region such as c1 (3), mRNA from other phages (5), and Caulobacter crescentus (24). In these leaderless mRNAs, conclusive evidence indicates that DB plays an essential role in the formation of the translation initiation complex in the absence of SD (3). Such complexes in the absence of SD may be formed easier without any extra sequence upstream of the initiation codon (3,24).
Among a number of cis elements in E. coli mRNAs known to enhance translation (25), DB next to SD appears to be found most often (6, 26 -28). It is engaging to elucidate which of the two sequences (DB or SD) binds first to 30 S ribosomes for the formation of the initiation complex. In this regard, it should be noted that the SELEX method using 30 S ribosomes as a ligand resulted in enrichment of DB sequences accompanied with SD (8,9).