京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第21654号 / 医博第4460号 / 新制||医||1035(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 朝長 啓造, 教授 妹尾 浩, 教授 萩原 正敏 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Cellular response to various stress conditions involves regulation of gene expression by different mechanisms. Translation is the final step in the flow of genetic information and regulation at this level allows an early response to changes in physiological conditions. Initiation of translation is the rate-limiting step of protein synthesis and hence is tightly regulated. Translation initiation in mammalian cells is mainly by “cap dependent pathway” wherein the 5’methyl guanosine “cap” structure is recognized by certain canonical initiation factors along with 40S ribosomal subunit and the complex scans the 5’UTR till it recognizes initiator AUG. This leads to the joining of the 60S ribosomal subunit and the initiation of translation. In an alternate mode of translation initiation called as the Internal ribosome entry site mediated translation (IRES), the ribosomes are recruited closer to the initiator AUG in a 5’ cap independent manner. Efficient translation by IRES mode requires some canonical initiation factors like eIF2 and eIF3 and other non-canonical IRES-trans-acting factors (ITAFs), which include human La antigen, polypyrimidine-tract binding protein (PTB),Upstream of N-Ras (Unr), Poly (rC) binding protein (PCBP2) etc. Various types of stress conditions, such as starvation of growth factors, heat shock, hypoxia, viral infection lead to down regulation of protein synthesis. However, translation of a subset of mRNAs continues or is up-regulated. Many of these mRNA may be translated by an IRES mode. It is believed that cellular IRESs become active during such conditions that abrogate the cap-dependent mode of translation so that the pool of vital proteins is maintained in the cell. In this thesis, presence of ‘IRES’ element has been investigated in the 5’UTR of Interferon regulatory factor -2 (IRF2) mRNA and the possible physiological significance has been studied. Further, it has been shown that polypyrimidine tract binding protein or PTB is important for the IRES activity. The probable mechanism of action of PTB has also been investigated which suggests that PTB interaction alters the IRF2 IRES conformation thus facilitating translation initiation. In the first part of the thesis, mRNAs that continue to be translated under heat-shocked condition, which is known to abrogate cap-dependent translation initiation, has been investigated by cDNA micro-array hybridization analysis of the ribosome bound RNA. The global protein synthesis was severely impaired under heat shock; however a number of mRNAs continued translation under this condition. Some of these mRNAs encode proteins that are likely to be involved in the heat shock response. Few of these genes are also reported to contain IRES element. Since the micro-array was performed from the RNA extracted from ribosome bound mRNA fraction in a condition when cap-dependent translation is impaired, it was hypothesized that some of the genes, which are up regulated under such condition, might operate via cap-independent mode of translation initiation. Based on this study, one candidate gene, the ‘interferon regulatory factor 2 (IRF2)’ was selected from the pool of up regulated genes and presence of an IRES element was investigated. Interferon regulatory factors are DNA-binding proteins that control interferon (IFN) gene expression. IRF2 has been shown to function as repressor of IFN and IFN-inducible genes. Real–Time and semi-quantitative RT-PCR assays were performed which validated the micro-array data. In the second part of the thesis, the presence of IRES element in the 5’UTR of IRF2 was investigated. Bicistronic assay showed comparable IRES activity with a known representative IRES, BiP, thus suggesting the presence of an IRES element in the IRF2 5’UTR. Stringent assays were then performed to rule out cryptic promoter activity, re-initiation/scanning or alternative splicing in the 5’UTR of the IRF2. RNA transfections using in vitro synthesized bicistronic RNAs further validated the presence of the IRES element. To understand the physiological significance of an IRES element in IRF2 mRNA, the cells were subjected to various stress conditions and IRES activity was studied. It seems IRF2 IRES function might not be sensitive to eIF4G cleavage, since its activity was only marginally affected in presence of Coxsackievirus 2A protease, which is known to cleave eIF 4G and thus inhibit the cap-dependent translation. Incidentally, Hepatitis A virus IRES was affected under such condition. Additionally, it was observed that compared to HCV or Bip IRES, the effect of Interferon α treatment was not so pronounced on the IRF2 IRES. This was further evidenced by its unchanged protein level post-treatment with interferon α. Furthermore, in cells treated with tunicamycin (a known agent causing ER stress), the IRF2 IRES activity and the protein levels were unaffected, although the cap dependent translation was severely impaired. The observations so far suggested that the IRF2 protein level is practically unchanged under conditions of ER stress and interferon treatment. Metabolic labeling followed by immunoprecipitation of IRF2 in cells treated with either tunicamycin or interferon suggested that de novo synthesis of the protein is continued under the above conditions thus validating our earlier data. In the third part of the thesis, the role of an IRES trans acting factor, PTB, in modulating the IRF2 IRES activity has been investigated. Analysis of the cellular protein binding with the IRF2 IRES suggested that certain cellular factors might influence its function under stress conditions. The IRF2 IRES was found to interact with a known trans-acting factor or PTB. To study the possible role of this trans acting factor, the PTB gene was partially silenced by PTB specific siRNA. This led to a decrease in the IRF2 IRES activity, suggesting that PTB is probably essential for the IRES activity. Interestingly, when Hela cells (with partially silenced PTB) were treated with tunicamycin (inducer of ER stress) the level of IRF2 protein was also found to be less thus pointing to an important role of PTB in IRF2 protein synthesis under such conditions. Western blot analysis and immunofluoroscence assay suggested that there was no significant nuclear-cytoplasmic relocalization of PTB under the condition studied. Primer extension inhibition assay or Toe-printing analysis was performed to detect the contact points of PTB on the IRF2 5’UTR. Many toe-prints were found on the 3’ end of the 5’UTR RNA. A 3’ deletion mutant was generated that showed reduced PTB binding. Incidentally the IRES activity of the mutant was also found to be less than the wt IRF2 RNA. Subsequently, structural analysis of the RNA was performed using enzymatic (CV1, RNase T1) and chemical modification (DMS) agents. Footprinting assay in presence of PTB suggested that there is change in the structure when PTB interacts with the RNA. To investigate this further, CD spectrum analysis of the IRF2 RNA in the presence of PTB was performed which indicated that there was a conformational change under such condition thus validating our earlier observation. The thesis reveals a novel cellular IRES element in the 5’UTR of IRF2 mRNA. The characterization of the IRES and possible role played by PTB protein in modulating its activity suggests that the regulated expression of IRF2 protein by its IRES element under various stress conditions would have major implications on the cellular response. Incidentally, this study constitutes the first report on translational control of interferon regulatory factors by internal initiation. The results might have far reaching implications on the possible role of IRF2 in controlling the intricate balance of cellular gene expression under stress conditions in general.

CoxsackievirusB3 (CVB3), a member of the Picornaviridae family is the causative agent of Virus-induced Myocarditis and Dilated Cardiomyopathy. The 5’UTR contains an Internal Ribosome Entry Site or IRES element that recruits ribosomes in a cap-independent manner. The ribosomes are recruited upstream of the AUG triplet at 591 (AUG591), also called as the cryptic AUG, after which they scan downstream for about 150 nucleotide, before initiating at the initiator AUG or AUG741. The 3’UTR of CVB3 is 99 nts long, highly structured RNA containing conserved domains, and is followed by a poly (A) tail of variable lengths. We have investigated possible involvement of host proteins which may interact with CVB3 IRES and influence its activity. We have demonstrated the role of Poly-pyrimidine tract binding protein (PTB) and established PTB as a bona-fide ITAF for CVB3, by characterizing the effect of partial silencing of PTB ex-vivo in HeLa cells. The IRES activity in BSC-1 cells, reported to have very low level of endogenous PTB, is found to be significantly low compared to that in HeLa cells. PTB is observed to interact with both the 5’ and 3’ UTR of CVB3, although with different affinities. Finer mapping of the interaction between PTB and the UTRs showed that the protein interacts with multiple regions of both UTRs. We have also shown the cis-acting effect of the CVB3-3’UTR on IRES mediated translation. The PTB contact points on the 3’UTRwas found to map to conserved regions, the deletion of which abrogates the 3’UTR mediated enhancement of the IRES activity. The possible role played by PTB in enhancing IRES activity by CVB3 3’UTR suggests that PTB protein might help in circularization of the CVB3 RNA by bridging the ends necessary for efficient translation of the viral RNA. In the second part, we have investigated possible role of some of the cis-acting element present in the CVB3 5’UTR RNA particularly the cryptic AUG. We have shown that mutation in cryptic AUG reduces the efficiency of translation mediated by the CVB3 IRES. Mutation in cryptic AUG moiety also reduces the interaction of mutant RNA with La protein. We have demonstrated that binding of 48S ribosomal complex with mutant IRES RNA was weaker compared to wt IRES RNA. We have investigated the possible alteration in secondary structure in the mutant RNA by chemical and enzymatic modification, which suggests that there is marginal alteration in the local structure due to mutation. It appears that integrity of cryptic AUG is important for efficient translation initiation by the CVB3 IRES. Results suggest that cryptic AUG plays a significant role in mediating internal initiation of translation of CVB3 RNA by mediating precise La binding and correct positioning of the 48S ribosomal complex. Finally, we have investigated the importance of a conserved hexa-nucleotide stretch in the apical loop within stem loop C (SLC, nt104-180), upstream of the ribosome landing site, on CVB3 IRES function. It has been already shown from our laboratory that the deletion at this apical loop resulted in significant decrease in IRES activity. This deletion mutant was shown to alter the secondary structure of the CVB3 5’UTR RNA. Here we have investigated the effect of point mutation in the apical loop SLC/c on CVB3 IRES activity by generating substitution mutation in the apical loop SLC/c in order to avoid possible alteration in secondary structure. Both the deletion or substitution mutation at this apical loop resulted in significant decrease in IRES activity. Both the mutant IRES RNAs (deletion and substitution mutant) failed to interact with certain trans-acting factors. Furthermore, expression of CVB3 2A protease significantly enhanced IRES activity of the wild type, but the effect was not so pronounced on the mutant IRESs. It is possible that the mutant RNAs were unable to interact with some trans-acting factors critical for enhanced IRES function. We have short-listed three proteins of approximate molecular mass of 56, 64 and 90 kDa, which showed reduced binding with mutant IRESs. By using RNA affinity column with biotinylated UTP labeled RNA we have purified couple of proteins and identified p64 as Cyto Keratin 1 protein by performing in-gel trypsin digestion followed by MALDI analysis. Overall, the results characterize the CVB3 IRES structurally and functionally, which could be useful in targeting critical RNA-protein interactions to develop candidate antiviral agent against Coxsackievirus infection.

Le virus de l’immunodéficience humaine de type 1 (VIH-1) est responsable du syndrome de l’immunodéficience acquise (SIDA). Il faut identifier de nouvelles cibles pour le développement d’agents anti-VIH-1, car ce virus développe une résistance aux agents présentement utilisés. Notre but est d’approfondir la caractérisation de l’étape du changement de cadre de lecture ribosomique en -1 (déphasage -1) nécessaire à la production du précurseur des enzymes du VIH-1. Ce déphasage est programmé et effectué par une minorité de ribosomes lorsqu’ils traduisent la séquence dite glissante à un endroit spécifique de l’ARN messager (ARNm) pleine-longueur du VIH-1. L’efficacité de déphasage est contrôlée par le signal stimulateur de déphasage (SSF), une tige-boucle irrégulière située en aval de la séquence glissante. La structure du SSF est déroulée lors du passage d’un ribosome, mais elle peut se reformer ensuite. Nous avons montré que des variations de l’initiation de la traduction affectent l’efficacité de déphasage. Nous avons utilisé, dans des cellules Jurkat-T et HEK 293T, un rapporteur bicistronique où les gènes codant pour les luciférases de la Renilla (Rluc) et de la luciole (Fluc) sont séparés par la région de déphasage du VIH-1. La Rluc est produite par tous les ribosomes traduisant l’ARNm rapporteur alors que la Fluc est produite uniquement par les ribosomes effectuant un déphasage. L’initiation de ce rapporteur est coiffe-dépendante, comme pour la majorité des ARNm cellulaires. Nous avons examiné l’effet de trois inhibiteurs de l’initiation et montré que leur présence augmente l’efficacité de déphasage. Nous avons ensuite étudié l’effet de la tige-boucle TAR, qui est présente à l’extrémité 5’ de tous les ARNm du VIH-1. TAR empêche la liaison de la petite sous-unité du ribosome (40S) à l’ARNm et module aussi l’activité de la protéine kinase dépendante de l’ARN double-brin (PKR). L’activation de PKR inhibe l’initiation en phosphorylant le facteur d’initiation eucaryote 2 (eIF2) alors que l’inhibition de PKR a l’effet inverse. Nous avons étudié l’effet de TAR sur la traduction et le déphasage via son effet sur PKR en utilisant TAR en trans ou en cis, mais à une certaine distance de l’extrémité 5’ afin d’éviter l’interférence avec la liaison de la 40S. Nous avons observé qu’une faible concentration de TAR, qui active PKR, augmente l’efficacité de déphasage alors qu’une concentration élevée de TAR, qui inhibe PKR, diminue cette efficacité. Nous avons proposé un modèle où des variations de l’initiation affectent l’efficacité de déphasage en modifiant la distance entre les ribosomes parcourant l’ARNm et, donc, la probabilité qu’ils rencontrent un SSF structuré. Par la suite, nous avons déterminé l’effet de la région 5’ non traduite (UTR) de l’ARNm pleine-longueur du VIH-1 sur l’efficacité de déphasage. Cette 5’UTR contient plusieurs régions structurées, dont TAR à l’extrémité 5’, qui peut interférer avec l’initiation. Cet ARNm a une coiffe permettant une initiation coiffe-dépendante ainsi qu’un site d’entrée interne des ribosomes (IRES), permettant une initiation IRES-dépendante. Nous avons introduit cette 5’UTR, complète ou en partie, comme 5’UTR de notre ARNm rapporteur bicistronique. Nos résultats démontrent que cette 5’UTR complète inhibe l’initiation coiffe dépendante et augmente l’efficacité de déphasage et que ces effets sont dus à la présence de TAR suivie de la tige-boucle Poly(A). Nous avons aussi construit un rapporteur tricistronique où les ribosomes exprimant les luciférases utilisent obligatoirement l’IRES. Nous avons observé que cette initiation par l’IRES est faible et que l’efficacité de déphasage correspondante est également faible. Nous avons formulé une hypothèse pour expliquer cette situation. Nous avons également observé que lorsque les deux modes d’initiation sont disponibles, l’initiation coiffe dépendante est prédominante. Finalement, nous avons étudié l’effet de la protéine virale Tat sur l’initiation de la traduction et sur l’efficacité de déphasage. Nous avons montré qu’elle augmente l’initiation de la traduction et que son effet est plus prononcé lorsque TAR est située à l’extrémité 5’ des ARNm. Nous proposons un modèle expliquant les effets de Tat sur l’initiation de la traduction par l’inhibition de PKR ainsi que par des changements de l’expression de protéines cellulaires déroulant TAR. Ces résultats permettent de mieux comprendre les mécanismes régissant le déphasage du VIH-1, ce qui est essentiel pour le développement d’agents anti-déphasage. / The human immunodeficiency virus type 1 (HIV-1) is responsible for the acquired immune deficiency syndrome (AIDS). HIV-1 develops a resistance towards the inhibitors used to treat infected patients. It is thus important to identify new targets for the development of novel antiretroviral agents. The aim of our work was to better characterize the programmed -1 ribosomal frameshift which generates the precursor of HIV-1 enzymes. The frameshift occurs at a specific sequence of HIV-1 full-length messenger RNA (mRNA), the slippery sequence, and is performed by a minority of the ribosomes translating this mRNA. The frameshift efficiency is controlled by the frameshift stimulatory signal (FSS), an irregular stem-loop located downstream of the slippery sequence. FSS structure is unfolded by every ribosome translating this region and can refold afterwards. We showed that HIV-1 frameshift efficiency is affected by changes in the rate of translation initiation. We transfected Jurkat-T and HEK 293T cells with a bicistronic reporter that contains the frameshift region of HIV-1 between the Renilla luciferase (Rluc) and the firefly luciferase (Fluc) genes. Rluc is produced by all ribosomes translating this reporter whereas only ribosomes that make a –1 frameshift produce Fluc. The translation of the reporter is initiated via a cap-dependant mode, like the majority of cellular mRNAs. We first determined the effect of three inhibitors of translation initiation. We showed that their presence increases the frameshift efficiency. We next determined the impact of the TAR stem loop, which is located at the 5’end of every HIV-1 mRNA. TAR is known to impair the binding of the small subunit of the ribosome (40S) to the mRNA. TAR also modulates the activity of the double-stranded RNA-dependent protein kinase (PKR). When PKR is activated, it phosphorylates the eukaryotic initiation factor 2 (eIF2), inhibiting translation initiation. The inhibition of PKR has the opposite effect. We studied the effect of TAR on PKR by positioning TAR at a distance of the 5’ end where it cannot interfere with the binding of the 40S. Our results showed that a small amount of TAR, which activates PKR, increases the frameshift efficiency whereas a large amount of TAR, which inhibits PKR, decreases it. A model is presented where the variations of translation initiation modulate HIV-1 frameshift efficiency by altering the distance between the elongating ribosomes. This influences the probability that these ribosomes encounter or not a folded FSS. We next observed the effect of the 5’ untranslated region (UTR) of HIV-1 full length mRNA on its frameshift efficiency. This 5’UTR contains several structured parts, including TAR at the 5’end, which can inhibit translation initiation. This mRNA has a cap and an internal ribosome entry site (IRES) and could then use a cap dependent and an IRES-dependent mode of translation initiation. We replaced the 5’UTR of our bicistronic reporter mRNA by the complete 5’UTR of HIV-1 full-length mRNA or a part of it. Our results showed that the presence of the complete 5’UTR inhibits cap-dependent initiation of translation and increases the frameshift efficiency. Those effects are mostly due to the presence of TAR followed by a Poly(A) stem-loop. We also constructed a tricistronic reporter where the ribosomes translating the luciferases have to use an IRES-dependent initiation mode. The rate of this initiation was low and the frameshift efficiency obtained was also low. We proposed a hypothesis accounting for this situation. We also observed that when both initiation modes are available, the cap-dependent mode seems to be highly favored. Finally, we studied the impact of the Tat viral protein on translation initiation and frameshift efficiency. We showed that the presence of Tat increases translation initiation and decreases the frameshift efficiency. Those effects are more important when TAR is present at the 5’end of mRNA. We propose a model explaining the effects of Tat on translation initiation by the inhibition of PKR and by changes in the expression of cellular proteins that are able to unfold TAR. Our results allow us to better understand the mechanisms controlling HIV-1 frameshift, which will help in the development of drugs targeting the HIV-1 frameshift.

Le virus de l’immunodéficience humaine de type 1 (VIH-1) est responsable du syndrome de l’immunodéficience acquise (SIDA). Il faut identifier de nouvelles cibles pour le développement d’agents anti-VIH-1, car ce virus développe une résistance aux agents présentement utilisés. Notre but est d’approfondir la caractérisation de l’étape du changement de cadre de lecture ribosomique en -1 (déphasage -1) nécessaire à la production du précurseur des enzymes du VIH-1. Ce déphasage est programmé et effectué par une minorité de ribosomes lorsqu’ils traduisent la séquence dite glissante à un endroit spécifique de l’ARN messager (ARNm) pleine-longueur du VIH-1. L’efficacité de déphasage est contrôlée par le signal stimulateur de déphasage (SSF), une tige-boucle irrégulière située en aval de la séquence glissante. La structure du SSF est déroulée lors du passage d’un ribosome, mais elle peut se reformer ensuite. Nous avons montré que des variations de l’initiation de la traduction affectent l’efficacité de déphasage. Nous avons utilisé, dans des cellules Jurkat-T et HEK 293T, un rapporteur bicistronique où les gènes codant pour les luciférases de la Renilla (Rluc) et de la luciole (Fluc) sont séparés par la région de déphasage du VIH-1. La Rluc est produite par tous les ribosomes traduisant l’ARNm rapporteur alors que la Fluc est produite uniquement par les ribosomes effectuant un déphasage. L’initiation de ce rapporteur est coiffe-dépendante, comme pour la majorité des ARNm cellulaires. Nous avons examiné l’effet de trois inhibiteurs de l’initiation et montré que leur présence augmente l’efficacité de déphasage. Nous avons ensuite étudié l’effet de la tige-boucle TAR, qui est présente à l’extrémité 5’ de tous les ARNm du VIH-1. TAR empêche la liaison de la petite sous-unité du ribosome (40S) à l’ARNm et module aussi l’activité de la protéine kinase dépendante de l’ARN double-brin (PKR). L’activation de PKR inhibe l’initiation en phosphorylant le facteur d’initiation eucaryote 2 (eIF2) alors que l’inhibition de PKR a l’effet inverse. Nous avons étudié l’effet de TAR sur la traduction et le déphasage via son effet sur PKR en utilisant TAR en trans ou en cis, mais à une certaine distance de l’extrémité 5’ afin d’éviter l’interférence avec la liaison de la 40S. Nous avons observé qu’une faible concentration de TAR, qui active PKR, augmente l’efficacité de déphasage alors qu’une concentration élevée de TAR, qui inhibe PKR, diminue cette efficacité. Nous avons proposé un modèle où des variations de l’initiation affectent l’efficacité de déphasage en modifiant la distance entre les ribosomes parcourant l’ARNm et, donc, la probabilité qu’ils rencontrent un SSF structuré. Par la suite, nous avons déterminé l’effet de la région 5’ non traduite (UTR) de l’ARNm pleine-longueur du VIH-1 sur l’efficacité de déphasage. Cette 5’UTR contient plusieurs régions structurées, dont TAR à l’extrémité 5’, qui peut interférer avec l’initiation. Cet ARNm a une coiffe permettant une initiation coiffe-dépendante ainsi qu’un site d’entrée interne des ribosomes (IRES), permettant une initiation IRES-dépendante. Nous avons introduit cette 5’UTR, complète ou en partie, comme 5’UTR de notre ARNm rapporteur bicistronique. Nos résultats démontrent que cette 5’UTR complète inhibe l’initiation coiffe dépendante et augmente l’efficacité de déphasage et que ces effets sont dus à la présence de TAR suivie de la tige-boucle Poly(A). Nous avons aussi construit un rapporteur tricistronique où les ribosomes exprimant les luciférases utilisent obligatoirement l’IRES. Nous avons observé que cette initiation par l’IRES est faible et que l’efficacité de déphasage correspondante est également faible. Nous avons formulé une hypothèse pour expliquer cette situation. Nous avons également observé que lorsque les deux modes d’initiation sont disponibles, l’initiation coiffe dépendante est prédominante. Finalement, nous avons étudié l’effet de la protéine virale Tat sur l’initiation de la traduction et sur l’efficacité de déphasage. Nous avons montré qu’elle augmente l’initiation de la traduction et que son effet est plus prononcé lorsque TAR est située à l’extrémité 5’ des ARNm. Nous proposons un modèle expliquant les effets de Tat sur l’initiation de la traduction par l’inhibition de PKR ainsi que par des changements de l’expression de protéines cellulaires déroulant TAR. Ces résultats permettent de mieux comprendre les mécanismes régissant le déphasage du VIH-1, ce qui est essentiel pour le développement d’agents anti-déphasage. / The human immunodeficiency virus type 1 (HIV-1) is responsible for the acquired immune deficiency syndrome (AIDS). HIV-1 develops a resistance towards the inhibitors used to treat infected patients. It is thus important to identify new targets for the development of novel antiretroviral agents. The aim of our work was to better characterize the programmed -1 ribosomal frameshift which generates the precursor of HIV-1 enzymes. The frameshift occurs at a specific sequence of HIV-1 full-length messenger RNA (mRNA), the slippery sequence, and is performed by a minority of the ribosomes translating this mRNA. The frameshift efficiency is controlled by the frameshift stimulatory signal (FSS), an irregular stem-loop located downstream of the slippery sequence. FSS structure is unfolded by every ribosome translating this region and can refold afterwards. We showed that HIV-1 frameshift efficiency is affected by changes in the rate of translation initiation. We transfected Jurkat-T and HEK 293T cells with a bicistronic reporter that contains the frameshift region of HIV-1 between the Renilla luciferase (Rluc) and the firefly luciferase (Fluc) genes. Rluc is produced by all ribosomes translating this reporter whereas only ribosomes that make a –1 frameshift produce Fluc. The translation of the reporter is initiated via a cap-dependant mode, like the majority of cellular mRNAs. We first determined the effect of three inhibitors of translation initiation. We showed that their presence increases the frameshift efficiency. We next determined the impact of the TAR stem loop, which is located at the 5’end of every HIV-1 mRNA. TAR is known to impair the binding of the small subunit of the ribosome (40S) to the mRNA. TAR also modulates the activity of the double-stranded RNA-dependent protein kinase (PKR). When PKR is activated, it phosphorylates the eukaryotic initiation factor 2 (eIF2), inhibiting translation initiation. The inhibition of PKR has the opposite effect. We studied the effect of TAR on PKR by positioning TAR at a distance of the 5’ end where it cannot interfere with the binding of the 40S. Our results showed that a small amount of TAR, which activates PKR, increases the frameshift efficiency whereas a large amount of TAR, which inhibits PKR, decreases it. A model is presented where the variations of translation initiation modulate HIV-1 frameshift efficiency by altering the distance between the elongating ribosomes. This influences the probability that these ribosomes encounter or not a folded FSS. We next observed the effect of the 5’ untranslated region (UTR) of HIV-1 full length mRNA on its frameshift efficiency. This 5’UTR contains several structured parts, including TAR at the 5’end, which can inhibit translation initiation. This mRNA has a cap and an internal ribosome entry site (IRES) and could then use a cap dependent and an IRES-dependent mode of translation initiation. We replaced the 5’UTR of our bicistronic reporter mRNA by the complete 5’UTR of HIV-1 full-length mRNA or a part of it. Our results showed that the presence of the complete 5’UTR inhibits cap-dependent initiation of translation and increases the frameshift efficiency. Those effects are mostly due to the presence of TAR followed by a Poly(A) stem-loop. We also constructed a tricistronic reporter where the ribosomes translating the luciferases have to use an IRES-dependent initiation mode. The rate of this initiation was low and the frameshift efficiency obtained was also low. We proposed a hypothesis accounting for this situation. We also observed that when both initiation modes are available, the cap-dependent mode seems to be highly favored. Finally, we studied the impact of the Tat viral protein on translation initiation and frameshift efficiency. We showed that the presence of Tat increases translation initiation and decreases the frameshift efficiency. Those effects are more important when TAR is present at the 5’end of mRNA. We propose a model explaining the effects of Tat on translation initiation by the inhibition of PKR and by changes in the expression of cellular proteins that are able to unfold TAR. Our results allow us to better understand the mechanisms controlling HIV-1 frameshift, which will help in the development of drugs targeting the HIV-1 frameshift.