Translation initiation region dependency of translation initiation in Escherichia coli by IF1 and kasugamycin Serhiy Surkov 1 , Hanna Nilsson 1 , Louise C. V. Rasmussen 2 , Hans U. Sperling-Petersen 2 and Leif A. Isaksson 1 1 Department of Genetics, Microbiology and Toxicology, Stockholm University, Sweden 2 Department of Molecular Biology, Aarhus University, Denmark Introduction Translation initiation factor 1 (IF1), encoded by infA, is a small protein consisting of 71 amino acids in Escherichia coli. It is essential for cell viability [1], even though the exact reason for this remains obscure [2,3]. IF1 is highly conserved, and homologous proteins are present in all three domains of life (IF1 in bacteria, aIF1A in archeaons, and eIF1A in eukaryotes) [4,5]. IF1 is the smallest of the three initiation factors in E. coli. During initiation of translation, IF1 binds to the 30S ribosomal subunit in the A-site region [6,7], presumably through electrostatic interactions [8]. It stimulates the action of the other two factors, espe- cially translation initiation factor 2 (IF2) [9,10]. A crystal structure analysis of IF1 bound to the 30S ribosomal subunit shows that the factor is located in a cleft formed between helix 44, the 530 loop of 16S RNA, and ribosomal protein S12. Besides direct con- formational changes in helix 44 and the neighboring region, this binding induces small but significant changes in overall 30S subunit conformation, tilting the head of the subunit towards the A-site. It is pos- sible that these conformational changes could be even bigger in the absence of crystal lattice constraints [11]. Keywords bipA; cspA; infA(IF1); kasugamycin; yggJ Correspondence L. A. Isaksson, Department of Genetics, Microbiology and Toxicology, Stockholm University, S-10691 Stockholm, Sweden Fax: +46 8 164315 Tel: +46 8 164197 E-mail: leif.isaksson@gmt.su.se (Received 26 October 2009, revised 17 February 2010, accepted 17 March 2010) doi:10.1111/j.1742-4658.2010.07657.x Translation initiation factor 1 (IF1) is an essential protein in prokaryotes. The nature of IF1 interactions with the mRNA during translation initiation on the ribosome remains unclear, even though the factor has several known functions, one of them being RNA chaperone activity. In this study, we analyzed translational gene expression in vivo in two cold-sensitive chromo- somal mutant variants of IF1 with amino acid substitutions, R40D and R69L, using two different reporter gene systems. The strains with the mutant IF1 gave higher reporter gene expression than the control strain. The extent of this effect was dependent on the composition of the transla- tion initiation region. The Shine–Dalgarno (SD) sequence, AU-rich ele- ments upstream of the SD sequence and the region between the SD sequence and the initiation codon are important for the magnitude of this effect. The data suggest that the wild-type form of IF1 has a translation initiation region-dependent inhibitory effect on translation initiation. Kasu- gamycin is an antibiotic that blocks translation initiation. Addition of kasugamycin to growing wild-type cells increases reporter gene expression in a very similar way to the altered IF1, suggesting that the infA mutations and kasugamycin affect some related step in translation initiation. Genetic knockout of three proteins (YggJ, BipA, and CspA) that are known to interact with RNA causes partial suppression of the IF1-dependent cold sensitivity. Abbreviations IF1, translation initiation factor 1; IF2, translation initiation factor 2; SD, Shine–Dalgarno; TIR, translation initiation region. 2428 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS Several functions are attributed to IF1. It facilitates IF2-dependent fMet-tRNA binding to the P-site [10,12,13], probably by stabilizing IF2 binding to the 30S subunit [14], and stimulates the GTPase activity of IF2 [15]. It also increases binding of mRNA to the ini- tiation complex in the presence of IF2 [16]. Together with IF2, it stimulates drop-off of peptidyl-tRNAs with short polypeptides from 70S ribosomes [17]. In a recent study, it was shown that IF1 together with IF2 recognizes the formylmethionine moiety of initiator aminoacyl-tRNA and discriminates against unformy- lated and deacylated tRNA f Met [3]. IF1 is necessary for IF2 recycling after subunit joining and GTP hydrolysis [13,18]. IF1 directly contacts domains III–V of IF2 during initiation of translation [19,20]. Additionally, IF1 stim- ulates both ribosomal dissociation and subunit associa- tion without affecting the equilibrium point [21]. Even though IF1 (together with fMet-tRNA) has many effects on translation initiation, it is not crucial in the in vitro translation system based on purified compo- nents, in contrast to the other translation initiation factors IF2 and translation initiation factor 3 [22]. IF1 contains an oligomer-binding motif with high homology to the RNA-binding domains of ribosomal protein S1 and polynucleotide phosphorylase [23,24], and the factor binds to different synthetic polynucleo- tides in solution [25]. It has an RNA chaperone activ- ity both in vivo and in vitro [26]. IF1 can act as a transcriptional antiterminator in vivo, but this function of the factor is not essential for cell growth [27]. Cold shock stimulates expression of IF1 at the levels of both transcription and translation [28,29]. Other studies sug- gest that heterologous expression of E. coli IF1 in Bacillus subtilis can complement the double deletion of the cold shock-inducible genes cspB and cspC [30]. Kasugamycin is an aminoglycoside antibiotic that selectively inhibits initiation of translation in prokary- otes [31]. The antibiotic impairs binding of fMet-tRNA to the P-site on the 30S subunit and on 70S ribosomes [31]. Recent X-ray analyses have located a kasugamy- cin-binding site on the 30S subunit or on the 70S ribo- some in the region of the mRNA-binding tunnel in the E-site and P-site [32,33]. As no effect of kasugamycin on mRNA binding to the 30S subunit was shown [34], it was proposed that the antibiotic effectively distorts the mRNA structure near the P-site codon, thus pre- venting efficient fMet-tRNA binding [32,33,35]. Trans- lation of different mRNAs is affected by kasugamycin [36], depending on the nature of the nucleotides in mRNA corresponding to the E-site [32]. Translation of leaderless mRNA starting directly from AUG is insen- sitive to kasugamycin action [37]. Expression of a reporter gene is increased in E. coli strains that carry mutations in the chromosomal infA (IF1) gene [38]. These mutant strains grow consider- ably slower than the parental MG1655 strain, and some of them are cold sensitive for growth. The repor- ter gene is significantly overexpressed in two cold-sen- sitive chromosomal IF1 mutant strains or through addition of kasugamycin to the corresponding wild- type strain. In this study, we have made an extensive in vivo analysis of the mRNA sequence composition that causes such overexpression. We demonstrate simi- lar effects on TIR-dependent gene expression of the IF1 mutations and the antibiotic kasugamycin. The IF1-dependent cold sensitivity is partly suppressed by elimination by genetic knockout of some other pro- teins involved in RNA recognition (YggJ, BipA, and CspA). Results Increased gene expression by a mutant form of IF1 IF1 is essential for cell viability, although the reason for this is obscure. A set of mutants with mutations in infA, giving an altered IF1, has been isolated [38]. Some of these mutants are cold sensitive for growth, and have increased gene expression in an in vivo repor- ter system. We wanted to characterize the determinants of this effect, as they could reveal interactions between the factor and the rest of the translation initiation machinery in the growing cell. Comparison of several mutants with different chromosomal infA mutations motivated a closer study of two mutants, with an R40D alteration (strain CVR40D) or an R69L alter- ation (strain CVR69L), in IF1. Both of these mutants are cold sensitive, with CVR40D being more sensitive than CVR69L, and both can survive with the mutated IF1 gene in a single chromosome. Furthermore, both mutants have moderately increased expression of both the b-galactosidase and the A¢ reporter gene systems, indicating an effect of the altered IF1 [2]. The A¢ reporter gene system is based on a plasmid with a test gene (3A¢) with a varied sequence and an internal standard gene (2A¢) (see below). In this sys- tem, the 3A¢⁄2A¢ ratio is dependent on sequence changes introduced in 3A¢ as long as the control gene 2A¢ is not altered. As IF1 is involved in expression of both the 3A¢ test gene and the 2A¢ internal control gene, it was not clear whether an observed increase in the 3A¢⁄2A¢ ratio in the infA mutant bacteria was the result of an increase in 3A¢ expression or a decrease in 2A¢ expression, or both. S. Surkov et al. TIR dependence of translation initiation by IF1 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2429 To address this question, we performed a radioactive double-labeling experiment using the wild-type strain and the two IF1 cold-sensitive mutants CVR40D and CVR69L [2]. The pSS101 vector with the 3A¢ and 2A¢ genes was used to study protein A¢ expression in the strains by gel scanning. Proteins in CVR40D and CVR69 were labeled with [ 3 H]lysine. The MG1655 parental strain was labeled during cultivation with [ 14 C]lysine. Cells were cultivated separately, but each mutant was pooled with MG1655 when harvested. The 3A¢-encoded and 2A¢-encoded double-labeled proteins in the mixture were purified and separated on gels. The 3 H ⁄ 14 C isotope ratios were determined for the 2A¢ and 3A¢ gel bands, and compared with the isotope ratios of the total cellular protein, from the sample taken before the purification step. As can be seen in Fig. 1, values for 2A¢ reference gene expression in the IF1 mutants were quite similar to the values for total proteins. In contrast, 3A¢ expression was increased by the IF1 mutations. The increased values for the 3A¢⁄2A¢ ratio agree with previous determinations obtained by scanning of Comassie-stained gel bands. The results suggest that the increased 3A¢⁄2A¢ expres- sion ratio in the IF1 mutants studied here is mainly the result of increased 3A¢ expression, whereas 2A¢ is essentially not affected by the IF1 mutation. This implication is supported by preliminary 2D gel analysis of the CVR40D proteome showing that most cellular proteins are not affected by the R40D mutation, even though some are increased and a few are decreased in expression (not shown). Taken together, the total pro- tein mixture can be used as a reliable standard refer- ence, suggesting that the two infA (IF1) cold-sensitive mutant strains, CVR40D and CVR69L, both show sig- nificant 3A¢ reporter gene overexpression, as compared with the 2A¢ reference gene and total cellular proteins. Plasmid copy numbers in MG1655 and CVR40D were evaluated by spectrophotometric and electropho- retic analysis. The results indicated that changes in plasmid copy number do not play any role in the observed increase in reporter gene expression in the infA mutant strains. By use of a northern blotting technique, it was found that the mRNA levels corre- sponding to the 3A¢ and 2A¢ genes were not altered by the infA mutations (not shown). Expression of IF1 can be increased under different physiological conditions [28,29], even though infA is not under auto-control [46]. By using IF1-specific monoclonal antibodies [40], we have found that the levels of IF1 in CVR40D and CVR69L, using EF-Tu as a reference, are slightly higher (1.50- and 1.33-fold, respectively) than in the wild-type strain. However, no increase was seen in the 2A¢⁄total protein ratio as a result of the IF1 mutations, whereas the 3A¢⁄total pro- tein ratio was increased about two-fold for both of them (Fig. 1). Preliminary data suggest that overpro- duction of wild-type IF1 from a multicopy plasmid does not cause any increase in 3A¢ expression. The data suggest that the increased 3A¢⁄2A¢ ratio is mostly the result of changed functionality of the mutant IF1 and not of an altered IF1 level. Because the 3A¢ and the 2A¢ genes in the pSS101 plasmid are different in their translation initiation region (TIR) composition, the data suggest that there are altered functional inter- actions between the mutated IF1 factor and some sequence signals in the TIR region. We decided to ana- lyze these signals. Effect of the downstream region composition The influence of different sequences downstream of AUG on gene expression at the translational level has been well characterized [41]. Even though IF1 binds to the A-site of the 30S subunit, different +2 codons do not cause significantly changed levels of protein expres- sion in different IF1 mutant strains [38], indicating a 0.0 0.5 1.0 1.5 2.0 2.5 3A ′ 2A ′ Total protein 3 H/ 14 C ratio CVR40D/MG1655 CVR69/MG1655 MG1655 + ksg/MG1655 pSS101 3A′ TIR cuagcuaauaaauuaAGGAGGauuuaaauAUGaaaccucuagagucgacu 2A′ TIR cggauaacaauuucacacAGGAaacagaccAUGgaauugcaacacgauaag Fig. 1. Protein expression from the pSS101 plasmid as measured by 3 H ⁄ 14 C double labeling. Proteins in the parental MG1655 strain were labeled with [ 14 C]lysine. Proteins in CVR40D and CVR69L or in MG1655 in the presence of 175 lgÆmL )1 kasugamycin were labeled with [ 3 H]lysine. Cultures were pooled together, and the iso- tope ratios for the total cellular proteins, as well as the 3A¢ and 2A¢ protein bands in PAGE gels, were calculated. The isotope ratios for the 3A¢ reporter gene and the 2A¢ reference gene are shown in relation to the isotope ratio for the total cellular protein, which is taken as unity. TIRs of the 3A¢ and 2A¢ genes in pSS101 are indi- cated by the SDs in capital letters. The AUG initiation codon is in underlined capital letters. TIR dependence of translation initiation by IF1 S. Surkov et al. 2430 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS lack of codon specificity. Using CVR40D, we used the 3A¢⁄2A¢ test system to analyze other plasmids with dif- ferent sequences downstream of the initiation codon in 3A¢ (downstream regions DR-A, DR-B, DR-C, and DR-D) (Fig. 2) [42] but with a constant upstream sequence. In CVR40D, the expression levels of these 3A¢ variants with the Shine–Dalgarno (SD) + sequence were elevated, giving an approximately two-fold increase (Fig. 2), which is similar to the observed value for the original construct pSS101. This increased gene expression in CVR40D relative to the parental strain MG1655 is independent of the composition of the downstream region. Influence of the SD sequence and its upstream sequence IF1 inhibits the joining of 50S to the preinitiation complex if the SD sequence in the mRNA is extended. This effect is not seen for a four base SD [43]. For this reason, we studied the influence on gene expression of the length of the SD sequence in CVR40D and CVR69L and their parental strain MG1655. A set of constructs with different lengths of the SD sequence (4–10 bases) was used. As shown in Fig. 3A, the gene variants with a long SD sequence (6–10 bases) gave increased 3A¢⁄2A¢ values in the mutants, whereas a four base SD sequence gave very similar expression values in the wild-type and mutant strains. This obser- vation is in line with the fact that the 2A¢ reference gene, with an SD + sequence that is four bases short (Fig. 1), is expressed at an unchanged level in the infA mutant strains (Figs 1 and 3A). However, the total removal of the SD sequence from the test gene does not abolish the elevated expression level in the infA mutant strains, which remains 1.8-fold higher than in MG1655 (Fig. 3B). To further analyze the influence of the TIR sequence on IF1-dependent gene expression, several reporter gene variants were used. As can be seen in Fig. 3B for CVR40D, all of the analyzed TIR sequences gave higher 3A¢⁄2A¢ ratios than for the control strain MG1655. This was particularly true for pSS201 with its S1-binding site. The sequence dependency on reporter gene expres- sion was similar but less pronounced in CVR69L. The exception was pSS201. However, at 30 °C (test temperature for CVR69L), expression from this plasmid is toxic for MG1655, which makes correct evaluation of the expression levels difficult. The reason for this toxic effect is not clear, and requires further investigation. Increased expression was also obtained by the addi- tion of kasugamycin to MG1655, as discussed below. The effect of the length of the spacer between a canon- ical SD + sequence and the initiation codon was also analyzed. As shown in Fig. 3C, the longer spacers, especially with a 12 base spacer, gave higher gene expression in the infA mutants. Comparison of two different reporter gene systems Relevant reporter gene sequences were also analyzed by using the b-galactosidase assay system in MG1655 and CVR40D. For this purpose, the initiation region of the b-galactosidase gene from the pCMS71 plasmid [44] was replaced with some of the corresponding DR-A cuagcuaauaaauuaAGGAGGauuuaaauAUGAAAGCAAUUUUCGUAc DR-B cuagcuaauaaauuaAGGAGGauuuaaauAUGAGUGAAUCACAAGCCc DR-C cuagcuaauaaauuaAGGAGGauuuaaauAUGAAAAAGGAGUCGACUc DR-D cuagcuaauaaauuaAGGAGGauuuaaauAUGACCGAGGGUGUUUCCc 3A′ 2A′ DR-A 3A′/2A′ 0.16 0.08 0.370.170.390.210.660.31 CVR40DMG1655CVR40DMG1655CVR40DMG1655CVR40DMG1655 DR-B DR-C DR-D Fig. 2. PAGE analysis of reporter gene expression in MG1655 and CVR40D. Protein bands corresponding to the 3A¢ reporter gene and 2A¢ reference gene products are indicated. Sequences of TIRs of the reporter gene are shown with the different down- stream regions, DR-A, DR-B, DR-C, and DR-D, in bold letters. The expression ratios (3A¢⁄2A¢) for the two strains are given. S. Surkov et al. TIR dependence of translation initiation by IF1 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2431 initiation regions of 3A¢ gene variants. The results obtained by using the b-galactosidase or the 3A¢ repor- ter gene systems are compared in Fig. 4. It can be seen that the cold-sensitive CVR40D shows similarly increased reporter gene expression, as compared with the parental strain MG1655, for corresponding initia- tion region sequences for both assay systems. Increased sensitivity of IF1 mutants to kasugamycin The mRNA sequence-specific inhibition of translation initiation by the antibiotic kasugamycin is well documented [32,36]. It binds to the mRNA upstream of the initiation codon, and X-ray crystallography has AB C 3A′/2A′ ratio 3A′/2A′ ratio 3A′/2A′ ratio Fig. 3. (A) Influence of the SD sequence length on reporter gene expression. Strains and the numbers of bases in SD are indicated. (B) TIR- dependent reporter gene expression in MG1655, CVR40D and CVR69L or in MG1655 in the presence of 175 lgÆmL )1 kasugamycin (ksg). Sequences with different TIRs are shown. The SD region is in bold capital letters, and the initiation codon is in underlined capital letters. pSS301 and pSS101 represent SD ) and SD + versions of the parental 3A¢ plasmid. pSS201 carries an extension by the ribosome-binding sequence of ribosomal protein S1 (capital letters). The six bases upstream of SD + (capital letters) in pSS103 are from the 2A¢ gene in pSS101. pSS144 carries an extension of four bases in the spacer downstream of the SD + , as compared with pSS103. pSS133 carries three altered bases in the spacer as compared with pSS101. (C) Influence of the distance (Dis) between the SD sequence and the AUG initiation codon on reporter gene expression in MG1655, CVR40D and CVR69L. The number of bases forming the distance are indicated. The SD region is in bold capital letters, and the initiation codon is in underlined capital letters. TIR dependence of translation initiation by IF1 S. Surkov et al. 2432 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS located the antibiotic to the ribosomal tunnel region. As the effects of the R40D and the R69L mutations are dependent on the sequence upstream of the initia- tion codon, we wanted to analyze the effects of these mutations on growth and reporter gene expression in comparison with the action of kasugamycin. Growth of CVR40D and that of its parental strain MG1655 were compared in the presence of kasugamy- cin (Fig. 5A). Addition of kasugamycin reduced the growth rate of MG1655. During the growth curve acquisition for CVR40D cells in broth medium, it became apparent that addition of kasugamycin to a concentration of  70 lgÆmL )1 or higher had a delayed bacteriostatic effect, stopping growth of CVR40D at a D 590 nm of 0.8–1.0. In comparison, MG1655 did not show a bacteriostatic response unless as much as  140 lgÆmL )1 kasugamycin was used (Fig. 5A). Minimal inhibitory concentrations were determined for the three strains in minimal medium. These values were 40, 50 and 75 lgÆmL )1 for CVR40D, CVR69L, and MG1655, respectively. In summary, both CVR40D and CVR69L are more sensi- tive to kasugamycin during growth than is MG1655. Influence of kasugamycin on gene expression The effect of kasugamycin on expression of the 3A¢ reporter gene in plasmid pSS101 in the parental strain MG1655 was analyzed. The 3A¢⁄2A¢ ratio was increased approximately two-fold at a kasugamycin concentration of 175 lgÆmL )1 , as determined by gel scanning (Fig. 5B). The other bacteriostatic antibiotics tested (chloramphenicol and tetracycline) markedly decreased the growth rate of MG1655 cells but did not influence the 3A¢⁄2A¢ ratio associated with pSS101 (Figs 1 and 5B). Thus, the increased expression caused by kasugamycin was specific for this antibiotic and was not caused by the other two antibiotics, which also inhibit translation. Analysis of reporter gene expression using plasmid pSS101 showed that the pres- ence of 50 lgÆmL )1 kasugamycin in LB medium had almost no effect on the 3A¢⁄2A¢ ratio in MG1655, but caused an increased 3A¢⁄2A¢ ratio in CVR40D and CVR69L (Figs 1 and 5C). At this antibiotic concentra- tion (50 lgÆmL )1 ), no growth rate reduction was visi- ble for either MG1655 or the mutant strains. The 3A¢⁄2A¢ ratios were also measured for the different strains during growth in the presence of 70 or 100 lgÆmL )1 kasugamycin. At these concentrations, cells were collected at a D 590 nm of 0.6, before the bac- teriostatic effect of the antibiotic is seen. As can be seen in Fig. 5C, the increased 3A¢⁄2A¢ ratios reveal that reporter gene expression is more sensitive to kasu- gamycin for both IF1 mutants than for MG1655. In MG1655, kasugamycin at 175 lgÆmL )1 caused an increase in the 3A¢⁄2A¢ ratio in a TIR-dependent man- ner (Fig. 3B). This was particularly true in the case of pSS201, which contains the extended S1-binding site, and pSS144, which has a 12 base spacer between the SD sequence and the initiation codon. The same sequences gave increased gene expression in CVR40D and CVR69L in the absence of kasugamycin. Thus, addition of kasugamycin to MG1655 has very similar effects on reporter gene expression as the R40D and the R69L mutations (Figs 1, 3B and 5B). Suppression of cold sensitivity of an IF1 mutant by inactivation of other genes We investigated whether the inactivation of any other genes could suppress the cold-sensitive phenotype of CVR40D. The less cold-sensitive mutant CVR69L was not analyzed. The KEIO collection of E. coli gene knockout strains was used to introduce nonessential 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 A B CR40D/MG1655 ratio pSS201, S1 binding site pSS101, SD + pSS301, SD – pSS103, mutated sequence upstream of SD pSS133, mutated spacer pSS144, spacer extended by 4 bases Fig. 4. Comparison of gene expression measurements using two different reporter gene systems. The ratios of gene expression val- ues (3A¢⁄2A¢) in CVR40D and MG1655 as measured by the b-galac- tosidase (A) and protein A¢ (B) assays are shown. The plasmids used for the comparison are indicated, and their TIR sequences are given in Fig. 3. Spacer refers to the sequence between the initia- tion codon AUG and the SD region. S. Surkov et al. TIR dependence of translation initiation by IF1 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2433 gene knockouts by P1 transduction into CVR40D and kanamycin selection [45]. For screening, we chose 69 nonessential genes that are known to be associated with ribosome function or maturation as well as cold shock response. The double mutant strains were tested for growth on LB plates at 18 °C. It was found that inactivation of bipA, yggJ or cspA partly suppressed the cold-sensitive phenotype of the IF1 mutant CVR40D (Fig. 6). Discussion IF1 stimulates translation initiation in an in vitro sys- tem by promoting formation of the preinitiation com- plex [9,22]. However, it was found by Croitoru et al. [38] that cold-sensitive chromosomal infA mutants with the mutations R40D and R69L had increased reporter gene expression in vivo as compared with the wild-type strain. This depends on the TIR in the mRNA [38]. The overexpression found in the mutant as compared with the wild-type strain, using two different reporter gene systems, depends on the composition of the TIR. Such TIRs include the downstream region following the initiation codon, the length of the SD + sequence, provided that it is longer than four bases, and the dis- tance between the initiation codon and the SD + sequence, provided that it is shorter than 15 bases but longer than six bases. IF1 is known to have antitermination activity during transcription [27], and the intracellular level of IF1 can be increased by physiological treatments [28,29], but infA itself is not subject to any feedback control [46]. We found that the reporter mRNA levels in the mutants are not altered by the infA mutations. Analy- sis with IF1-specific monoclonal antibodies suggests that the level of IF1 is slightly increased in CVR40L and CVR40D. No corresponding increase is seen for the 2A¢⁄total protein ratio as compared with the Growth in the presence of kasugamycin 0 1 2 3 4 5 6 A 0 200 400 600 Time (min) D 590 MG1655 MG1655, ksg 70 µg·mL –1 MG1655, ksg 140 µg·mL –1 CVR40D CVR40D, ksg 70 µg·mL –1 CVR40D, ksg 140 µg·mL –1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Tet 0.5 µg·mL –1 Cam 2 µg·mL –1 Cam 1 µg·mL –1 Ksg 175 µg·mL –1 Ksg 100 µg·mL –1 0 3A′/2A′ ratio Kasu g amycin (µ g ·mL –1 ) 0 0.5 1 1.5 2 2.5 0 50 100 MG1655 CVR40D CVR69L 3A′/2A′ ratio B C Fig. 5. (A) Growth in LB medium in the presence of kasugamycin (ksg). Filled symbols represent MG1655 and open symbols repre- sent the mutant CVR40D. (B) Reporter gene expression in MG1655 in the presence of kasugamycin (ksg) or other antibiotics. The plas- mid was pSS101, as described in Fig. 1. Cam, chloramphenicol; Tet, tetracycline. (C) Reporter gene expression (pSS101) in MG1655 or in CVR40D and CVR69L in the presence of different kasugamycin concentrations. 1 3 2 4 Fig. 6. Growth of CVR40D and derivatives. The indicated strains were incubated on an LB plate at 18 °C for 4 days. 1, CVR40D; 2, CVR40D DcspA; 3, CVR40D DbipA; 4, CVR40D DyggJ. TIR dependence of translation initiation by IF1 S. Surkov et al. 2434 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS wild-type strain, whereas the 3A¢⁄total protein ratio is increased about two-fold. Taken together, the data suggest that most of the increase in the 3A¢⁄2A¢ ratio in the mutants is the result of changed functionality of IF1 and not overproduction of mutant IF1. Normally, one would expect a mutationally altered protein to show lowered, not increased, activity. Our analysis of reporter gene expression in the two cold- sensitive chromosomal IF1 mutants suggests that IF1 plays a role as a TIR-dependent repressor of transla- tion initiation, and that this negative effect is less pro- nounced in the mutants. It has been shown by Milon et al. [43] that IF1 acts as an inhibitor of formation of the initiation complex in vitro. In that study, inhibition by IF1 was found for an mRNA with a six base SD sequence, but not if this sequence was only four bases long. Our in vivo results described here for the R40D IF1 mutant are in line with their results, as the mutant IF1 showed increased expression in the case of a six base, but not a four base, SD sequence. The results reported here were obtained using two different reporter gene systems, two independent IF1 mutations and addition of kasugamycin to growing wild-type or mutant bacteria. The question can be asked as to what extent the results are applicable to normal genes. Preliminary proteomic analysis suggests that most normal genes are unaffected by the IF1 alter- ations, thus resembling the 2A¢ gene. Some genes, how- ever, show increased expression, being similar to 3A¢, and others show decreased expression. Kasugamycin increases the 3A¢⁄2A¢ ratio in our reporter gene system. The antibiotic is an inhibitor of translation initiation. This suggests that the apparently increased expression of 3A¢ relative to 2A¢ or to total protein caused by kas- ugamycin reflects a lower sensitivity of 3A¢ than of 2A¢ to kasugamycin. Upon addition of kasugamycin to growing bacteria, expression of some natural genes is either increased or decreased, also suggesting a differ- ent sensitivity to the antibiotic (not shown). IF1 binds to different synthetic polynucleotides in solution, and it contains an oligomer-binding motif with high homology to the RNA-binding domains of ribosomal protein S1 and polynucleotide phosphory- lase [23–25]. The crystal structure of IF1 in complex with the 30S ribosomal subunit suggests that IF1 could directly contact mRNA nucleotides in the ribosomal A-site [11]. IF1 affects the conformation of 16S rRNA, causing a movement of helix 44 and a global confor- mational change in the 30S subunit. This is visible as the movement of the head of the subunit towards the body and flipping of bases A1492 and A1493. This flipping has been shown to constitute an important part of the quality control signaling during tRNA or RF factor recognition of the A-site [11]. The R40D mutant described here is altered in its binding pocket for the base A1493 [2]. This suggests a direct interac- tion effect of the IF1 mutation. As IF1 influences the splicing of a group I intron in vivo and in vitro, and influences RNA annealing in vitro, the factor has an RNA chaperone activity [26]. It is conceivable that the mutant forms of IF1 are less capable of setting the intricate balance between favoring and disfavoring higher RNA structures, in the rRNA, mRNA or both, that are necessary for translational initiation. As a result, the initiation machinery could be biased such that the initiation conformation in the mutants is too high, giving the observed decreased mutant growth rate and cold sensitivity. The very similar expression responses to a number of different TIR sequences that were seen for CVR40D and CVR69L as compared with the addition of kasu- gamycin to the wild-type strain are compelling. The antibiotic binds in the mRNA channel, just upstream of AUG between bases G926 and A774 in 16S rRNA, according to a current X-ray structure model. It dis- torts the P-site, thereby disturbing the start codon position and preventing fMet-tRNA binding during the stage of 50S subunit joining [32,33]. Several aspects of kasugamycin action are still unknown. However, it is known that the resistance mutation (ksgA) affecting modification of the 16S rRNA gives an effect at the level of subunit joining [32]. In parallel with that, it has been shown that the wild-type form of IF1 reduces subunit joining depending on the mRNA TIR in vitro [43]. We have shown the apparent resemblance in the TIR dependence of IF1 and kasugamycin action as well as a synergistic effect on gene expression levels between an infA mutation and addition of kasugamy- cin. The results suggest that the actions of kasugamy- cin and IF1 are dependent on a closely related target or step during translation initiation. The correlation suggests that those mechanisms underlying the actions of the antibiotic and IF1 are similar, possibly at the level of subunit joining. The effects of kasugamycin and mutationally altered IF1 appear to be synergistic. Using the Keio strain collection, we found that the cold sensitivity associated with the R40D mutation is partly suppressed by inactivation of yggJ, bipA,or cspA. YggJ is a methylase that specifically modifies uri- dine 1498 of the 16S rRNA. This base is located in the mRNA channel upstream of AUG. The residue directly contacts the kasugamycin molecule in the X-ray structure [32]. It appears likely that it influences the TIR selection or AUG adjustment specificity of IF1. BipA is a protein that disrupts SD–antiSD interactions in some mRNAs during the first steps of S. Surkov et al. TIR dependence of translation initiation by IF1 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2435 translation [47]. Both of these two gene products are connected to the TIR-specific response of the IF1 mutants, suggesting that the imbalance in the mRNA translation is the primary reason for the cold sensitiv- ity of the infA mutants studied here. It is conceivable that elimination of yggJ or bipA could partially com- pensate for the enhanced translation initiation of mRNAs that is observed in the IF1 mutant strains. CspA facilitates translation initiation at low temper- atures by melting the mRNA secondary structure [39]. Cold shock stimulates expression of IF1 at the levels of both transcription and translation [28,29]. Thus, cold shock constitutes a common denominator for cspA and infA. The compensation of the cold sensitiv- ity of IF1 mutant strains by the inactivation of cspA is another functional link to IF1. In view of the finding that IF1 has RNA chaperone activity [26], the effects of the elimination of BipA, YggJ and CspA are conceivable, as all of them influ- ence the structure of RNA. The results are all in line with a model in which IF1, together with some other proteins, recognizes and shapes the TIR region in mRNA on the 30S ribosomal subunit. Experimental procedures Chemicals All chemicals used were of the highest purity commercially available. Enzymes were from New England Biolabs (Ipswich, MA, USA), Invitrogen (Carlsbad, CA, USA), Promega (Madison, WI, USA), and Fermentas Life Sciences (Vilnius, Lithuania). DNA extraction kits were from Qiagen (Hilden, Germany). Plasmids were prepared using kits from Qiagen, Fermentas Life Sciences, and GE Healthcare (Waukesha, WI, USA). Radioactive lysine was from GE Healthcare. Media and antibiotics Strains were grown in LB broth with tryptone and yeast extract or in M9 defined minimal medium [48] supplemented with all amino acids except for lysine. Ampicillin was used at a concentration of 100 lgÆmL )1 . Other antibiotics were used as indicated. Kasugamycin (Sigma, St Louis, MO, USA) was dissolved in water at 12.5 mgÆmL )1 , and the pH was adjusted with NaOH. Protein labeling Overnight cultures of MG1655, CVR40D or CVR69L cells, carrying the pSS101 plasmid with the 3A ¢ reporter system, were grown in M9 minimal medium supplemented with ampicillin and all amino acids except lysine. 50 lL of the overnight cell culture were inoculated into 5 mL of the same medium, and they were grown with intensive aeration to a D 590 nm of 0.2. At this point, 75 lL of a lysine solution labeled with 14 C (50 lCiÆmL )1 ) was added to MG1655 cells, and 35 lL of lysine solution labeled with 3 H (1 mCiÆmL )1 ) was added to CVR40D and CVR69L cells. The cells were grown to a D 590 nm of 0.6. Then, cold lysine was added to the final concentration of 0.25 mm, and the cultures were grown for an additional 20 min at 37 °C for CVR40D and at 30 °C for CVR69L. The difference in the growth temperature is due to different cold sensitivities of the strains. CVR40D shows a decreased growth rate, down to 50%, at 37 °C. CVR69L is grown at 30 °C to obtain a similar decrease in growth rate. The cultures were cooled, and MG1655 cells were combined with CVR40D or CRV69L cells. The com- bined cells were washed, harvested, and processed for pro- tein A¢ analysis by gel electrophoresis [49]. The double- labeled protein bands were excised and kept overnight for extraction in 300 lL of a solution containing 30% H 2 O 2 and 1% NH 4 OH at 37 °C. The resulting solution was placed in a scintillation vial containing 2 mL of Ultima Gold XR cock- tail (Perkin Elmer, Norwalk, CT, USA), and shaken for 1 h at room temperature. Radioactivity was counted in a 1219 Rackbeta scintillation counter (LKB, Bromma, Sweden). The amounts of radioactivity corresponding to 3 H and 14 Cin each sample were corrected according to the energy spectra of the pure elements, giving the ratio of counts originally derived from 3 H-labeled and 14 C-labeled lysine. The radioactivity of total protein samples was measured by placing 300 lL of the pooled 3 H-labeled and 14 C-labeled cell cultures into 10% trichloroacetic acid with 1% casamino acids. The precipitates were washed with 5% trichloroacetic acid containing 0.1% casamino acids, whereafter the filters were dried and radioactivity was measured in a scintillation counter. Alternatively, the pooled 3 H-labeled and 14 C-labeled cell lysate was loaded onto a polyacrylamide gel, and all of the resulting protein bands in one lane were excised together and counted as described above. Both methods gave similar 3 H ⁄ 14 C ratios. Similar PAGE experiments were performed with MG1655 cells expressing 3A¢ and 2A¢ proteins encoded by the pSS101 plasmid in the presence or absence of 175 lgÆmL )1 kasugamycin. Protein A¢ assay The protein A¢ reporter system has been extensively described [49]. Briefly, a plasmid carries two genes under the control of identical trc promoters. Both proteins are composed of identical A¢ building blocks derived from the IgG-binding domain (also known as the Z domain) of Staphylococcus aureus protein A. One gene, encoding three A¢ domains (3A¢, 21 kDa) is a reporter gene that can be modified and used to study the influence of different mRNA sequences on gene expression. The second gene in the plasmid, encoding two A¢ domains (2A¢, 14 kDa) is an TIR dependence of translation initiation by IF1 S. Surkov et al. 2436 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS internal control gene. The protein products of both genes are purified by affinity chromatography using IgG Sepha- rose, and the relative expression ratio 3A¢⁄2A¢ is estimated. Protein A¢ is not toxic, and this system does not have any transcriptional polarity effects. Expression of the 3A¢ and 2A¢ genes was analyzed by using 15% SDS ⁄ PAGE electrophoresis. Gels were stained with Comassie Brilliant Blue R (Sigma) and scanned using a LAS1000 plus (FujiFilm) camera. Bands corresponding to 3A¢ and 2A¢ proteins were quantified using image gauge v. 4.0 (FujiFilm). All experiments were repeated at least four times. b-Galactosidase assay Wild-type MG1655 and the IF1 mutants CVR40D and CVR69L were grown overnight at 37 °C in M9 medium supplemented with all amino acids at the recommended concentrations and 100 lgÆmL )1 ampicillin. These cultures were used for inoculatation into the same medium at 37 °C. Exponentially growing cells (D 590 nm of 0.4–0.5) were har- vested without isopropyl thio-b-d-galactoside induction, as the trc promoter is leaky, giving significant expression even in the absence of induction. b-Galactosidase acivity of the lysed uninduced cells was then determined as described pre- viously [50]. Plasmids and strains All plasmids used are based on the pHN109 vector [44], which carries the 2A¢ internal control gene and the 3A¢ test gene. The different initiation regions in the 3A¢ test gene in the plasmids used are shown in the corresponding figures. P1 transduction was performed according to Miller [51]. The MG1655 strain (F ) , ilvG, rfb-50, rph) was used as a wild-type reference strain. Its derivatives CVR40D and CVR69L have the same genotype except for the R40D or R69L mutations, respectively, in the infA gene on the chromosome [2]. Growth curves Thirty microliter volumes of overnight cultures were inocu- lated into 3 mL of fresh LB. Cells were grown with intensive shaking to a D 590 nm of 0.6, and then diluted in LB to obtain a D 590 nm of 0.05 in the presence or absence of kasugamycin; this was followed by measurements of bacterial growth. Cold sensitivity complementation test The gene deletions tested were transferred from the KEIO strain collection [45] by P1 transduction into CVR40D on LB plates, with selection for kanamycin resistance. Plates were incubated for 4 days at 18 °C. Immunoblotting Cells from overnight cultures were lysed by sonication, and cell lysates were loaded into a 15% SDS ⁄ PAGE gel along with IF1 and EF-Tu standards. Proteins were transferred to a nitrocellulose membrane by electroblotting. Buffer [0.9% NaCl, 50 mm Tris ⁄ HCl (pH 7.5), 1% (v ⁄ v) Tween] was used for the wash and incubation steps. Dry milk was used for blocking, three types of mouse monoclonal anti- bodies against IF1 (1BD3, 3AE12 and 2EF10 from [40]) and rabbit polyclonal antibodies against EF-Tu were used as primary antibodies, and horseradish peroxidase-conju- gated swine anti-(rabbit IgG) and horseradish peroxidase- conjugated rabbit anti-(mouse IgG) were used as secondary antibodies. The blot was developed using a GE Healthcare ECL kit and exposed to X-ray film. The X-ray film was scanned using a GS-800 calibrated densitometer (Bio-Rad, Richmond, CA, USA), and analyzed with BioRad quantity one software. Acknowledgements We thank V. Croitoru for the CVR strains, as well as help and advice. This work was supported by grants from the Swedish Science foundation (L. C. V. Rasmussen) and from the Swedish Institute (Visby Program). We thank Genkaku for strains from the Keio collection. References 1 Cummings HS & Hershey JW (1994) Translation initia- tion factor IF1 is essential for cell viability in Escheri- chia coli. J Bacteriol 176, 198–205. 2 Croitoru V, Bucheli-Witschel M, Hagg P, Abdulkarim F & Isaksson LA (2004) Generation and characteriza- tion of functional mutants in the translation initiation factor IF1 of Escherichia coli. Eur J Biochem ⁄ FEBS 271, 534–544. 3 Antoun A, Pavlov MY, Lovmar M & Ehrenberg M (2006a) How initiation factors maximize the accuracy of tRNA selection in initiation of bacterial protein synthe- sis. Mol Cell 23, 183–193. 4 Kyrpides NC & Woese CR (1998) Universally con- served translation initiation factors. Proc Natl Acad Sci USA 95, 224–228. 5 Sorensen HP, Hedegaard J, Sperling-Petersen HU & Mortensen KK (2001) Remarkable conservation of translation initiation factors: IF1 ⁄ eIF1A and IF2 ⁄ eIF5B are universally distributed phylogenetic markers. IUBMB Life 51, 321–327. 6 Moazed D, Samaha RR, Gualerzi C & Noller HF (1995) Specific protection of 16 S rRNA by S. Surkov et al. TIR dependence of translation initiation by IF1 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2437 [...]... usage of infA in response to cold shock in Escherichia coli FEBS Lett 580, 539–544 Giangrossi M, Brandi A, Giuliodori AM, Gualerzi CO & Pon CL (2007) Cold-shock-induced de novo transcription and translation of infA and role of IF1 during cold adaptation Mol Microbiol 64, 807–821 Weber MH, Beckering CL & Marahiel MA (2001) Complementation of cold shock proteins by translation initiation factor IF1 in vivo... Binding of formylmethionyl-tRNA to 30S ribosomal sub-units Nature 217, 1013–1016 Kay AC & Grunberg-Manago M (1972) The mechanism of action of initiation factor IF1: non-analogy with elongation factor EFTs Biochim Biophys Acta 277, 225–230 Stringer EA, Sarkar P & Maitra U (1977) Function of initiation factor 1 in the binding and release of initiation factor 2 from ribosomal initiation complexes in Escherichia. ..TIR dependence of translation initiation by IF1 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 S Surkov et al translational initiation factors J Mol Biol 248, 207– 210 Dahlquist KD & Puglisi JD (2000) Interaction of translation initiation factor IF1 with the E coli ribosomal A site J Mol Biol 299, 1–15 Celano B, Pawlik RT & Gualerzi CO (1988) Interaction of Escherichia coli translation- initiation factor... structure of the translational initiation factor IF1 from E coli contains an oligomer-binding motif EMBO J 16, 1436–1443 Gribskov M (1992) Translational initiation factors IF-1 and eIF-2 alpha share an RNA-binding motif with prokaryotic ribosomal protein S1 and polynucleotide phosphorylase Gene 119, 107–111 Schleich T, Verwolf GL & Twombly K (1980) A circular dichroism study of Escherichia coli initiation. .. Differential inhibition of coliphage MS2 protein synthesis by ribosome-directed antibiotics J Mol Biol 70, 41–55 37 Moll I & Blasi U (2002) Differential inhibition of 30S and 70S translation initiation complexes on leaderless mRNA by kasugamycin Biochem Biophys Res Commun 297, 1021–1026 38 Croitoru VV, Bucheli-Witschel M & Isaksson LA (2005) In vivo involvement of mutated initiation factor IF1 in gene expression... Cooperative effects by the initiation codon and its flanking regions on translation initiation Gene 273, 259–265 42 Jin H, Zhao Q, Gonzalez de Valdivia EI, Ardell DH, Stenstrom M & Isaksson LA (2006) In uences on gene expression in vivo by a Shine–Dalgarno sequence Mol Microbiol 60, 480–492 43 Milon P, Konevega AL, Gualerzi CO & Rodnina MV (2008) Kinetic checkpoint at a late step in translation initiation Mol... CM, Jin H, Major LL, Tate WP & Isaksson LA (2001) Codon bias at the 3¢-side of the initiation TIR dependence of translation initiation by IF1 45 46 47 48 49 50 51 codon is correlated with translation initiation efficiency in Escherichia coli Gene 263, 273–284 Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL & Mori H (2006) Construction of Escherichia coli K-12 inframe,... IF1 and IF2 synergistically remove peptidyl-tRNAs with short polypeptides from the P-site of translating Escherichia coli ribosomes J Mol Biol 281, 241–252 Chae YB, Mazumder R & Ochoa S (1969) Polypeptide chain initiation in E coli: isolation of homogeneous initiation factor E2 and its relation to ribosomal proteins Proc Natl Acad Sci USA 62, 1181–1188 Brock S, Szkaradkiewicz K & Sprinzl M (1998) Initiation. .. & SperlingPetersen HU (2005) Initiation of protein synthesis in bacteria Microbiol Mol Biol Rev 69, 101–123 Antoun A, Pavlov MY, Lovmar M & Ehrenberg M (2006b) How initiation factors tune the rate of initiation of protein synthesis in bacteria EMBO J 25, 2539– 2550 Carter AP, Clemons WM Jr, Brodersen DE, MorganWarren RJ, Hartsch T, Wimberly BT & Ramakrishnan V (2001) Crystal structure of an initiation. .. Initiation factors of protein biosynthesis in bacteria and their structural relationship to elongation and termination factors Mol Microbiol 29, 409–417 Allen GS, Zavialov A, Gursky R, Ehrenberg M & Frank J (2005) The cryo-EM structure of a translation initiation complex from Escherichia coli Cell 121, 703–712 van der Hofstad GA, Buitenhek A, van den Elsen PJ, Voorma HO & Bosch L (1978) Binding of labeled . Translation initiation region dependency of translation initiation in Escherichia coli by IF1 and kasugamycin Serhiy Surkov 1 , Hanna Nilsson 1 , Louise C. V. Rasmussen 2 , Hans U. Sperling-Petersen 2 and. TIRs of the 3A¢ and 2A¢ genes in pSS101 are indi- cated by the SDs in capital letters. The AUG initiation codon is in underlined capital letters. TIR dependence of translation initiation by IF1. of bases forming the distance are indicated. The SD region is in bold capital letters, and the initiation codon is in underlined capital letters. TIR dependence of translation initiation by IF1