Final Report Summary - PBS (Dissecting the new 'platform-binding site' on the bacterial ribosome. Role of particular ribosomal proteins in regulating translation initiation in bacteria)
I. Background
X-ray crystallographic and cryo-electron microscopy data {Marzi et al., 2007, Cell, 130, 1019-1031} indicate that structured mRNAs bind to the bacterial ribosome at a newly characterised ribosomal site located at the 'platform' of the small ribosomal subunit (the 'platform binding' site or PBS) at an early step of translation initiation. In a second step, the mRNAs are supposed to unfold and to migrate to the P-site to interact with the initiator tRNA. Ribosomal protein S1, a protein essential for translation initiation and involved in many other biological functions, could be responsible for the mRNA unfolding during the second step.
During these two years in Paris, Alexey Korepanov (AK) studied several functions of ribosomal protein S1 and the role of other ribosomal proteins (r-proteins) localised at the PBS and potentially contacting the structured 5'-terminal parts of mRNAs. An important part of the work was done in collaboration with Stefano Marzi in the team of Pascale Romby at the IBMC in Strasbourg.
II. Ribosomal protein S1
Protein S1 belongs to the small ribosomal subunit (30S) and is the largest r-protein {Subramanian, 1983, Prog Nucleic Acid Res Mol Biol, 28, 101-42}. In addition to its essential role in translation initiation, S1 is known to be one of the subunits of the phage Qß replicase {Blumenthal et al., 1976, J Biol Chem, 251, 2740-3}. In vitro experiments indicate that S1 is also involved in RNA unfolding {Rajkowitsch and Schroeder, 2007, RNA, 13, 2053-60}, transcriptional cycling {Sukhodolets et al., 2006, RNA, 12, 1505-13}, binding tmRNA {Wower et al., 2002, Biochemistry, 41, 8826-36}, forming a complex with a phage protein involved in general recombination {Venkatesh and Radding, 1993, J Bacteriol, 175, 1844-6}, and in stimulating phage T4 RegB endonuclease {Durand et al., 2006, Nucleic Acids Res, 34, 6549-60}. Most of these properties have not yet been studied in vivo.
The role of ribosomal protein S1 in adaptation
The existence of an mRNA primary binding site on the 30S platform (PBS) implies existence of an activity that allows the structured mRNAs to unfold and 'adapt' to the mRNA channel. This adaptation allows the initiation codon to occupy the P-site, allowing the correct positioning of the initiator tRNA. Our hypothesis was that S1 is responsible for this adaptation. On the contrary, mRNA with an unfolded translation initiation region could bind to the P site without the S1 unfolding activity.
In collaboration with S. Marzi in the team of Pascale Romby, we have provided evidence that a structured mRNA binds to the PBS and, in a second step, needs S1 to accommodate into the mRNA channel to give a functional translation initiation complex. The situation is different with an unstructured mRNA, which seems to accommodate directly in the mRNA channel, probably because this mRNA does not need to be unfolded.
The domains of the S1 r-protein
S1 protein is made out of six so-called 'S1 motifs' found in many proteins interacting with nucleic acids {Salah et al., 2009, Nucleic Acids Res, 37, 5578-88}. In vitro experiments of Stefano Marzi (from the Strasbourg team) showed that most of the in vitro activities of S1 during initiation are mainly attributed to the first three N-terminal domains of S1. Alexey Korepanov constructed a set of chromosomal deletions that eliminate five out the six domains one after the other starting from the 3' end of the gene. Deletions of the two C-terminal domains are viable but confer a cold-sensitive phenotype. Bacteria expressing S1 deleted of the three C-terminal domains is viable but very sick and those expressing S1 with larger deletions are not viable. This data correlated well with the in vitro results.
The effect of C-terminal deletions of S1 on rRNA levels
During his studies, Alexey Korepanov had noticed that the strains carrying deletions of the two C-terminal domains contained lower 23S rRNA levels than the wild-type strain. To find out at what level the defect could be (transcription initiation, anti-termination, elongation, and/or termination). Alexey Korepanov made a set of constructs that allow the cloning on the chromosome of promoters, anti-terminators, and terminators or different combinations of these transcriptional signals upstream of a gene encoding a mutated but stable version of a tRNAArg. The tRNA level was measured with northern blots using a fluorescent probe. In this construct, the gene for beta-galactosidase was placed downstream of the tRNA precursor gene. The corresponding transcript is matured, because of the tRNA precursor, to give the same lacZ mRNA whatever is placed upstream. We have tested our system and shown that it quantitatively detects promoters, anti-terminators and terminators in more valuable way than previous systems {Albrechtsen et al., 1990, J Mol Biol, 213, 123-34}. Using this system, Alexey Korepanov obtained evidence that the deletion of the C-terminal domain of S1 affects anti-termination efficiency of rRNA operon.
A study of transcription termination
Alexey Korepanov showed that the constructs carrying a promoter followed by either a Rho-dependent or Rho-independent terminator upstream of the tRNA precursor gene and lacZ, give a Lac- phenotype. This allowed us to select for Lac+ mutants. Interestingly, these Lac+ mutants were of two kinds. The first kind was found with Rho-independent terminators and all the corresponding mutations were located in the terminator, thus acting in cis. The second kind was found with the Rho-dependent terminators. A majority of the mutants were located in the gene for transcription termination factor Rho. The selection is very easy and yielded many new rho alleles that have been sequenced but remain to be studied biochemically. The selection also yields mutations in genes different from rho, potentially in the genes for Nus factors or even uncharacterised genes. The corresponding mutations are being mapped.
The effect of C-terminal deletions of S1 on the cellular physiology
Comparative proteomics have been carried out by S. Marzi and O. Fuchsbauer in the laboratory of Pascale Romby revealed that the deletion of the two last domains (five and six) of S1, altered the level of several proteins. Their results show that, in the strains with C-terminally deleted versions of S1, the level of flagellin was strongly decreased, and the level of CTP synthetase, cysteine synthetase A, and pyruvate formate lyase I were reproducibly decreased. On the contrary, the level of iron superoxide dismutase and trigger factor was increased. Alexey Korepanov measured the expression of all these genes using translational fusions to lacZ expressed from the endogenous and/or exogenous promoters. These experiments indicate that the expression of fliC, the gene for flagellin is decreased in a strain carrying a deletion for the last domain of S1. However, this S1-dependent activation might be indirect since it occurs at the transcriptional level and not at the translational level. The expression of fliC is regulated at the level of transcription by fliA (the gene for a flagella-specific sigma factor) and flhD (the gene for a regulator of flagella synthesis). We are investigating whether the translation of fliA and/or flhD is S1-dependent. Fusions of lacZ to sodB, the gene encoding iron superoxide dismutase, show that the expression of sodB is enhanced in the strain expressing the S1 protein lacking the two last domains showing that the regulation takes place at the translational level. However, the Strasbourg team has shown that the translation of the sodB mRNA is S1-independent in vitro. We believe that the S1-dependent enhancement we see in vivo is the consequence of the inability of the defective S1 to translate the majority of S1-dependent mRNAs, allowing the enhancement of the translation of the minority of S1-independent mRNAs. Interestingly enough, the sodB mRNA is predicted to be unstructured, which fits with our hypothesis that such an mRNA is directly accommodated in the mRNA channel with the initiation codon at the P site, without the help of S1.
III. Role of the proteins of the ribosomal platform in translation initiation and regulation
Based on the way the expression of the rpsO gene, encoding ribosomal protein S15, is auto-regulated at the level of translation initiation, the work of {Marzi et al., 2007, Cell, 130, 1019-1031} proposed that a group of four r-proteins (S2, S7, S11 and S18) are involved in the first steps of translation initiation and recognition of structured mRNAs (not only for rpsO), and constitute the PBS.
Our general approach was to introduce mutant patches (corresponding to changes in one to several amino acids) into these genes and analyse their effects on the expression and the regulation of rpsO mRNA. The mutant patches were chosen on the basis of the work of {Marzi et al., 2007, Cell, 130, 1019-1031} and the structure of the ribosome. The amino acid patches correspond to conserved residues that should be in contact with the rpsO mRNA when it binds to the platform of the 30S subunit during the first steps of translation initiation.
For the moment, we have isolated multiple mutant patches in S2, S7 and S18. Only a limited number of patches affect the expression/regulation of the rpsO gene. The corresponding mutants have been handed over to the Strasbourg team who is analysing the effect of these patches on the whole E. coli proteome.
IV. Conclusions and further directions
Our work indicates that r-protein S1 has many properties required to unfold structured mRNAs during translation initiation. In addition, we show that S1 activity is mainly dependent of domains one to three but domains four, five, and six increase the efficiency of the process. The work done with mutants of S2, S7, and S18 belonging to the PBS requires to be continued to find out if the constructed patches affect the first steps of translation initiation. However, the genetic system constructed by Alexey Korepanov is powerful enough to select for such mutants instead of constructing them.
In addition, Alexey Korepanov has constructed a manageable and powerful system that allows a precise measurement of the efficiency of transcription initiation, termination and anti-termination. This system has been used to select for a set of cis and trans-acting mutants modified in transcription termination. The system has allowed the selection of a many new mutants of termination factor Rho and in other genes that remain to be characterised. In addition, the system has given indications that S1 acts at transcription anti-termination. Whether this effect is direct or indirect remains to be characterised.
X-ray crystallographic and cryo-electron microscopy data {Marzi et al., 2007, Cell, 130, 1019-1031} indicate that structured mRNAs bind to the bacterial ribosome at a newly characterised ribosomal site located at the 'platform' of the small ribosomal subunit (the 'platform binding' site or PBS) at an early step of translation initiation. In a second step, the mRNAs are supposed to unfold and to migrate to the P-site to interact with the initiator tRNA. Ribosomal protein S1, a protein essential for translation initiation and involved in many other biological functions, could be responsible for the mRNA unfolding during the second step.
During these two years in Paris, Alexey Korepanov (AK) studied several functions of ribosomal protein S1 and the role of other ribosomal proteins (r-proteins) localised at the PBS and potentially contacting the structured 5'-terminal parts of mRNAs. An important part of the work was done in collaboration with Stefano Marzi in the team of Pascale Romby at the IBMC in Strasbourg.
II. Ribosomal protein S1
Protein S1 belongs to the small ribosomal subunit (30S) and is the largest r-protein {Subramanian, 1983, Prog Nucleic Acid Res Mol Biol, 28, 101-42}. In addition to its essential role in translation initiation, S1 is known to be one of the subunits of the phage Qß replicase {Blumenthal et al., 1976, J Biol Chem, 251, 2740-3}. In vitro experiments indicate that S1 is also involved in RNA unfolding {Rajkowitsch and Schroeder, 2007, RNA, 13, 2053-60}, transcriptional cycling {Sukhodolets et al., 2006, RNA, 12, 1505-13}, binding tmRNA {Wower et al., 2002, Biochemistry, 41, 8826-36}, forming a complex with a phage protein involved in general recombination {Venkatesh and Radding, 1993, J Bacteriol, 175, 1844-6}, and in stimulating phage T4 RegB endonuclease {Durand et al., 2006, Nucleic Acids Res, 34, 6549-60}. Most of these properties have not yet been studied in vivo.
The role of ribosomal protein S1 in adaptation
The existence of an mRNA primary binding site on the 30S platform (PBS) implies existence of an activity that allows the structured mRNAs to unfold and 'adapt' to the mRNA channel. This adaptation allows the initiation codon to occupy the P-site, allowing the correct positioning of the initiator tRNA. Our hypothesis was that S1 is responsible for this adaptation. On the contrary, mRNA with an unfolded translation initiation region could bind to the P site without the S1 unfolding activity.
In collaboration with S. Marzi in the team of Pascale Romby, we have provided evidence that a structured mRNA binds to the PBS and, in a second step, needs S1 to accommodate into the mRNA channel to give a functional translation initiation complex. The situation is different with an unstructured mRNA, which seems to accommodate directly in the mRNA channel, probably because this mRNA does not need to be unfolded.
The domains of the S1 r-protein
S1 protein is made out of six so-called 'S1 motifs' found in many proteins interacting with nucleic acids {Salah et al., 2009, Nucleic Acids Res, 37, 5578-88}. In vitro experiments of Stefano Marzi (from the Strasbourg team) showed that most of the in vitro activities of S1 during initiation are mainly attributed to the first three N-terminal domains of S1. Alexey Korepanov constructed a set of chromosomal deletions that eliminate five out the six domains one after the other starting from the 3' end of the gene. Deletions of the two C-terminal domains are viable but confer a cold-sensitive phenotype. Bacteria expressing S1 deleted of the three C-terminal domains is viable but very sick and those expressing S1 with larger deletions are not viable. This data correlated well with the in vitro results.
The effect of C-terminal deletions of S1 on rRNA levels
During his studies, Alexey Korepanov had noticed that the strains carrying deletions of the two C-terminal domains contained lower 23S rRNA levels than the wild-type strain. To find out at what level the defect could be (transcription initiation, anti-termination, elongation, and/or termination). Alexey Korepanov made a set of constructs that allow the cloning on the chromosome of promoters, anti-terminators, and terminators or different combinations of these transcriptional signals upstream of a gene encoding a mutated but stable version of a tRNAArg. The tRNA level was measured with northern blots using a fluorescent probe. In this construct, the gene for beta-galactosidase was placed downstream of the tRNA precursor gene. The corresponding transcript is matured, because of the tRNA precursor, to give the same lacZ mRNA whatever is placed upstream. We have tested our system and shown that it quantitatively detects promoters, anti-terminators and terminators in more valuable way than previous systems {Albrechtsen et al., 1990, J Mol Biol, 213, 123-34}. Using this system, Alexey Korepanov obtained evidence that the deletion of the C-terminal domain of S1 affects anti-termination efficiency of rRNA operon.
A study of transcription termination
Alexey Korepanov showed that the constructs carrying a promoter followed by either a Rho-dependent or Rho-independent terminator upstream of the tRNA precursor gene and lacZ, give a Lac- phenotype. This allowed us to select for Lac+ mutants. Interestingly, these Lac+ mutants were of two kinds. The first kind was found with Rho-independent terminators and all the corresponding mutations were located in the terminator, thus acting in cis. The second kind was found with the Rho-dependent terminators. A majority of the mutants were located in the gene for transcription termination factor Rho. The selection is very easy and yielded many new rho alleles that have been sequenced but remain to be studied biochemically. The selection also yields mutations in genes different from rho, potentially in the genes for Nus factors or even uncharacterised genes. The corresponding mutations are being mapped.
The effect of C-terminal deletions of S1 on the cellular physiology
Comparative proteomics have been carried out by S. Marzi and O. Fuchsbauer in the laboratory of Pascale Romby revealed that the deletion of the two last domains (five and six) of S1, altered the level of several proteins. Their results show that, in the strains with C-terminally deleted versions of S1, the level of flagellin was strongly decreased, and the level of CTP synthetase, cysteine synthetase A, and pyruvate formate lyase I were reproducibly decreased. On the contrary, the level of iron superoxide dismutase and trigger factor was increased. Alexey Korepanov measured the expression of all these genes using translational fusions to lacZ expressed from the endogenous and/or exogenous promoters. These experiments indicate that the expression of fliC, the gene for flagellin is decreased in a strain carrying a deletion for the last domain of S1. However, this S1-dependent activation might be indirect since it occurs at the transcriptional level and not at the translational level. The expression of fliC is regulated at the level of transcription by fliA (the gene for a flagella-specific sigma factor) and flhD (the gene for a regulator of flagella synthesis). We are investigating whether the translation of fliA and/or flhD is S1-dependent. Fusions of lacZ to sodB, the gene encoding iron superoxide dismutase, show that the expression of sodB is enhanced in the strain expressing the S1 protein lacking the two last domains showing that the regulation takes place at the translational level. However, the Strasbourg team has shown that the translation of the sodB mRNA is S1-independent in vitro. We believe that the S1-dependent enhancement we see in vivo is the consequence of the inability of the defective S1 to translate the majority of S1-dependent mRNAs, allowing the enhancement of the translation of the minority of S1-independent mRNAs. Interestingly enough, the sodB mRNA is predicted to be unstructured, which fits with our hypothesis that such an mRNA is directly accommodated in the mRNA channel with the initiation codon at the P site, without the help of S1.
III. Role of the proteins of the ribosomal platform in translation initiation and regulation
Based on the way the expression of the rpsO gene, encoding ribosomal protein S15, is auto-regulated at the level of translation initiation, the work of {Marzi et al., 2007, Cell, 130, 1019-1031} proposed that a group of four r-proteins (S2, S7, S11 and S18) are involved in the first steps of translation initiation and recognition of structured mRNAs (not only for rpsO), and constitute the PBS.
Our general approach was to introduce mutant patches (corresponding to changes in one to several amino acids) into these genes and analyse their effects on the expression and the regulation of rpsO mRNA. The mutant patches were chosen on the basis of the work of {Marzi et al., 2007, Cell, 130, 1019-1031} and the structure of the ribosome. The amino acid patches correspond to conserved residues that should be in contact with the rpsO mRNA when it binds to the platform of the 30S subunit during the first steps of translation initiation.
For the moment, we have isolated multiple mutant patches in S2, S7 and S18. Only a limited number of patches affect the expression/regulation of the rpsO gene. The corresponding mutants have been handed over to the Strasbourg team who is analysing the effect of these patches on the whole E. coli proteome.
IV. Conclusions and further directions
Our work indicates that r-protein S1 has many properties required to unfold structured mRNAs during translation initiation. In addition, we show that S1 activity is mainly dependent of domains one to three but domains four, five, and six increase the efficiency of the process. The work done with mutants of S2, S7, and S18 belonging to the PBS requires to be continued to find out if the constructed patches affect the first steps of translation initiation. However, the genetic system constructed by Alexey Korepanov is powerful enough to select for such mutants instead of constructing them.
In addition, Alexey Korepanov has constructed a manageable and powerful system that allows a precise measurement of the efficiency of transcription initiation, termination and anti-termination. This system has been used to select for a set of cis and trans-acting mutants modified in transcription termination. The system has allowed the selection of a many new mutants of termination factor Rho and in other genes that remain to be characterised. In addition, the system has given indications that S1 acts at transcription anti-termination. Whether this effect is direct or indirect remains to be characterised.