Gammaherpesvirus infection triggers the formation of tRNA fragments from premature tRNAs

Transfer RNAs (tRNAs) are fundamental for both cellular and viral gene expression during viral infection. In addition, mounting evidence supports biological function for tRNA cleavage products, including in the control of gene expression during conditions of stress and infection. We previously reported that infection with the model murine gammaherpesvirus, MHV68, leads to enhanced tRNA transcription. However, whether this has any influence on tRNA transcript processing, viral replication, or the host response is not known. Here, we combined two new approaches, sequencing library preparation by Ordered Two Template Relay (OTTR) and tRNA bioinformatic analysis by tRAX, to quantitatively profile full-length tRNAs and tRNA fragment (tRF) identities during MHV68 infection. We find that MHV68 infection triggers both pre-tRNA and mature tRNA cleavage, resulting in the accumulation of specific tRFs. OTTR-tRAX revealed not only host tRNAome changes, but also the expression patterns of virally-encoded tRNAs (virtRNAs) and virtRFs made from the MHV68 genome, including their base modification signatures. Because the transcript ends of several host tRFs matched tRNA splice junctions, we tested and confirmed the role of tRNA splicing factors TSEN2 and CLP1 in MHV68-induced tRF biogenesis. Further, we show that CLP1 kinase, and by extension tRNA splicing, is required for productive MHV68 infection. Our findings provide new insight into how gammaherpesvirus infection both impacts and relies on tRNA transcription and processing.

Importance

Diverse conditions of infection and cellular stress incite the cleavage of transfer RNAs, leading to the formation of tRNA fragments which can directly regulate gene expression. In our study of gammaherpesviruses, such as the murine herpesvirus 68 and human oncogenic Kaposi Sarcoma associated Herpesvirus, we discovered that transfer RNA regulation and cleavage is a key component of gene reprogramming during infection. We present the first in-depth profile of tRNA fragment generation in response to DNA virus infection, using state-of-the-art sequencing techniques that overcome several challenges with tRNA sequencing. We present several lines of evidence that tRNA fragments are made from newly-transcribed premature tRNAs and propose that this may be a defining characteristic of tRNA cleavage in some contexts. Finally, we show that tRNA splicing machinery is involved with the formation of some MHV68-induced tRNA fragments, with a key regulator of splicing, CLP1, required for maximal viral titer. Together, we posit that tRNA processing may be integral to the elegant shift in gene expression that occurs during viral take-over of the host cell.

Gammaherpesvirus infection alters transfer RNA splicing and triggers tRNA cleavage

Transfer RNAs (tRNAs) are fundamental for both cellular and viral gene expression during viral infection. Moreover, mounting evidence supports a noncanonical role for tRNA cleavage products in the control of gene expression during diverse conditions of stress and infection. We previously reported that infection with the model murine gammaherpesvirus, MHV68, leads to altered tRNA transcription, suggesting that tRNA regulation may play an important role in mediating viral replication or the host response. To better understand how viral infection alters tRNA expression, we combined Ordered Two Template Relay (OTTR) with tRNA-specific bioinformatic software called tRAX to profile full-length tRNAs and fragmented tRNA-derived RNAs (tDRs) during infection with MHV68. We find that OTTR-tRAX is a powerful sequencing strategy for combined tRNA/tDR profiling and reveals that MHV68 infection triggers pre-tRNA and mature tRNA cleavage, resulting in the accumulation of specific tDRs. Fragments of virally-encoded tRNAs (virtRNAs), as well as virtRNA base modification signatures are also detectable during infection. We present evidence that tRNA splicing factors are involved in the biogenesis of MHV68-induced cleavage products from pre-tRNAs and, in the case of CLP1 kinase, impact infectious virus production. Our data offers new insights into the importance of tRNA processing during gammaherpesvirus infection.

tRNATyr has an unusually short half-life in Trypanosoma brucei

Following transcription, tRNAs undergo a series of processing and modification events to become functional adaptors in protein synthesis. Eukaryotes have also evolved intracellular transport systems whereby nucleus-encoded tRNAs may travel out and into the nucleus. In trypanosomes, nearly all tRNAs are also imported from the cytoplasm into the mitochondrion, which lacks tRNA genes. Differential subcellular localization of the cytoplasmic splicing machinery and a nuclear enzyme responsible for queuosine modification at the anticodon "wobble" position appear to be important quality control mechanisms for tRNATyr, the only intron-containing tRNA in T. brucei. Since tRNA-guanine transglycosylase (TGT), the enzyme responsible for Q formation, cannot act on an intron-containing tRNA, retrograde nuclear transport is an essential step in maturation. Unlike maturation/processing pathways, the general mechanisms of tRNA stabilization and degradation in T. brucei are poorly understood. Using a combination of cellular and molecular approaches, we show that tRNATyr has an unusually short half-life. tRNATyr, and in addition tRNAAsp, also show the presence of slow-migrating bands during electrophoresis, we term these conformers: alt-tRNATyr and alt-tRNAAsp respectively. Although we do not know the chemical or structural nature of these conformers, alt-tRNATyr has a short half-life resembling that of tRNATyr; the same is not true for alt-tRNAAsp. We also show that RRP44, which is usually an exosome subunit in other organisms, is involved in tRNA degradation of the only intron-containing tRNA in T. brucei and partly responsible for its unusually short half-life.

RNA binding by ADAR3 inhibits adenosine-to-inosine editing and promotes expression of immune response protein MAVS

Members of the ADAR family of double-stranded RNA–binding proteins regulate one of the most abundant RNA modifications in humans, the deamination of adenosine to inosine. Several transcriptome-wide studies have been carried out to identify RNA targets of the active deaminases ADAR1 and ADAR2. However, our understanding of ADAR3, the brain-specific deaminase-deficient ADAR family member, is limited to a few transcripts. In this study, we identified over 3300 transcripts bound by ADAR3 and observed that binding of ADAR3 correlated with reduced editing of over 400 sites in the glioblastoma transcriptome. We further investigated the impact of ADAR3 on gene regulation of the transcript that encodes MAVS, an essential protein in the innate immune response pathway. We observed reduced editing in the MAVS 3′ UTR in cells expressing increased ADAR3 or reduced ADAR1 suggesting ADAR3 acts as a negative regulator of ADAR1-mediated editing. While neither ADAR1 knockdown or ADAR3 overexpression affected MAVS mRNA expression, we demonstrate increased ADAR3 expression resulted in upregulation of MAVS protein expression. In addition, we created a novel genetic mutant of ADAR3 that exhibited enhanced RNA binding and MAVS upregulation compared with wildtype ADAR3. Interestingly, this ADAR3 mutant no longer repressed RNA editing, suggesting ADAR3 has a unique regulatory role beyond altering editing levels. Altogether, this study provides the first global view of ADAR3-bound RNAs in glioblastoma cells and identifies both a role for ADAR3 in repressing ADAR1-mediated editing and an RNA-binding dependent function of ADAR3 in regulating MAVS expression.

Distinct Stress-Dependent Signatures of Cellular and Extracellular tRNA-Derived Small RNAs

The cellular response to stress is an important determinant of disease pathogenesis. Uncovering the molecular fingerprints of distinct stress responses may identify novel biomarkers and key signaling pathways for different diseases. Emerging evidence shows that transfer RNA-derived small RNAs (tDRs) play pivotal roles in stress responses. However, RNA modifications present on tDRs are barriers to accurately quantifying tDRs using traditional small RNA sequencing. Here, AlkB-facilitated methylation sequencing is used to generate a comprehensive landscape of cellular and extracellular tDR abundances in various cell types during different stress responses. Extracellular tDRs are found to have distinct fragmentation signatures from intracellular tDRs and these tDR signatures are better indicators of different stress responses than miRNAs. These distinct extracellular tDR fragmentation patterns and signatures are also observed in plasma from patients on cardiopulmonary bypass. It is additionally demonstrated that angiogenin and RNASE1 are themselves regulated by stressors and contribute to the stress-modulated abundance of sub-populations of cellular and extracellular tDRs. Finally, a sub-population of extracellular tDRs is identified for which AGO2 appears to be required for their expression. Together, these findings provide a detailed profile of stress-responsive tDRs and provide insight about tDR biogenesis and stability in response to cellular stressors.

Disruption in A-to-I Editing Levels Affects C. elegans Development More Than a Complete Lack of Editing

A-to-I RNA editing, catalyzed by ADAR proteins, is widespread in eukaryotic transcriptomes. Studies showed that, in C. elegans, ADR-2 can actively deaminate dsRNA, whereas ADR-1 cannot. Therefore, we set out to study the effect of each of the ADAR genes on the RNA editing process. We performed comprehensive phenotypic, transcriptomics, proteomics, and RNA binding screens on worms mutated in a single ADAR gene. We found that ADR-1 mutants exhibit more-severe phenotypes than ADR-2, and some of them are a result of non-editing functions of ADR-1. We also show that ADR-1 significantly binds edited genes and regulates mRNA expression, whereas the effect on protein levels is minor. In addition, ADR-1 primarily promotes editing by ADR-2 at the L4 stage of development. Our results suggest that ADR-1 has a significant role in the RNA editing process and in altering editing levels that affect RNA expression; loss of ADR-1 results in severe phenotypes.

A protein-protein interaction underlies the molecular basis for substrate recognition by an adenosine-to-inosine RNA-editing enzyme

Adenosine deaminases that act on RNA (ADARs) convert adenosine to inosine within double-stranded regions of RNA, resulting in increased transcriptomic diversity, as well as protection of cellular double-stranded RNA (dsRNA) from silencing and improper immune activation. The presence of dsRNA-binding domains (dsRBDs) in all ADARs suggests these domains are important for substrate recognition; however, the role of dsRBDs in vivo remains largely unknown. Herein, our studies indicate the Caenorhabditis elegans ADAR enzyme, ADR-2, has low affinity for dsRNA, but interacts with ADR-1, an editing-deficient member of the ADAR family, which has a 100-fold higher affinity for dsRNA. ADR-1 uses one dsRBD to physically interact with ADR-2 and a second dsRBD to bind to dsRNAs, thereby tethering ADR-2 to substrates. ADR-2 interacts with >1200 transcripts in vivo, and ADR-1 is required for 80% of these interactions. Our results identify a novel mode of substrate recognition for ADAR enzymes and indicate that protein-protein interactions can guide substrate recognition for RNA editors.

The C. elegans neural editome reveals an ADAR target mRNA required for proper chemotaxis

ADAR proteins alter gene expression both by catalyzing adenosine (A) to inosine (I) RNA editing and binding to regulatory elements in target RNAs. Loss of ADARs affects neuronal function in all animals studied to date. Caenorhabditis elegans lacking ADARs exhibit reduced chemotaxis, but the targets responsible for this phenotype remain unknown. To identify critical neural ADAR targets in C. elegans, we performed an unbiased assessment of the effects of ADR-2, the only A-to-I editing enzyme in C. elegans, on the neural transcriptome. Development and implementation of publicly available software, SAILOR, identified 7361 A-to-I editing events across the neural transcriptome. Intersecting the neural editome with adr-2 associated gene expression changes, revealed an edited mRNA, clec-41, whose neural expression is dependent on deamination. Restoring clec-41 expression in adr-2 deficient neural cells rescued the chemotaxis defect, providing the first evidence that neuronal phenotypes of ADAR mutants can be caused by altered gene expression.

DNA is the blueprint that tells each cell in an organism how it should operate. It encodes the instructions to make proteins and other molecules. To make a protein, a section of DNA known as a gene is used as a template to make molecules known as messenger ribonucleic acids (or mRNAs for short). The message in RNA consists of a series of individual letters, known as nucleotides, that tell the cell how much of a protein should be produced (referred to as gene expression) as well as the specific activities of each protein.

The letters in mRNAs can be changed in specific cells and at certain points in development through a process known as RNA editing. This process is essential for animals to grow and develop normally and for the brain to work properly. Errors in RNA editing are found in patients suffering from a variety of neuropathological diseases, including Alzheimer’s disease, depression and brain tumors. Humans have millions of editing sites that are predicted to affect gene expression. However, many studies of RNA editing have only focused on the changes that alter protein activity.

The ADAR proteins carry out a specific type of RNA editing in animals. In a microscopic worm known as Caenorhabditis elegans the loss of an ADAR protein called ADR-2 reduces the ability of the worm to move in response to chemicals, a process known as chemotaxis. Deffit et al. found that loss of ADR-2 affected the expression of over 150 genes in the nervous system of the worm. To identify which letters in the mRNAs were edited in the nervous system, Deffit et al. developed a new publically available software program called SAILOR (software for accurately identifying locations of RNA editing). This program can be used to detect RNA editing in any cell, tissue or organism.

By combining the experimental and computational approaches, Deffit et al. were able to identify a gene that was edited in normal worms and expressed at lower levels in the mutant worms. Increasing the expression of just this one of gene in the mutant worms restored the worms’ ability to move towards a chemical “scent”.

Together, these findings suggest that when studying human neuropathological diseases we should consider the effect of RNA editing on the amount of gene expression as well as protein activity. Future work should investigate the importance of RNA editing in controlling gene expression in other diseases including cancers.

Genetic transformation of chickens using irradiated male gametes

Results have been obtained which corroborate those of Pandey and Patchell (Molec. Gen. Genet., 186, 305, 1982) in demonstrating that genetic material from irradiated semen is incorporated into the embryo and expressed, albeit at rather a low rate, and is subsequently transmitted to progeny of the transfected birds. The method provides a technically straightforward means of transferring genetic material where rapid and reliable means of detecting the transferred gene exist. An advantage of the method is that regulatory regions are likely to be carried with the transferred gene but there is equally a disadvantage in the simultaneous transfer of unwanted material.

Stress of transportation for broilers

Transporting birds for two to four hours over distances of up to 224 km (140 miles) induced hyperlipacidaemia and hypoglycaemia while the concentration of plasma cholesterol decreased and then increased significantly. These responses were the same in summer and winter. Body temperature was not affected by the experimental conditions. There was a consistent increase in plasma corticosterone which was greater in winter than in summer.

Failure to induce stress reactions following vaccination against Marek's disease or Newcastle disease

Chickens aged one or 21 days were given single injections of a vaccine against Marek's disease or Newcastle disease respectively, and monitored over a three week period for any signs of a stressor response. No consistent evidence of such a response was found. Some data on normal ontogenetic changes in various adrenal and plasma variables are also presented.

Re-establishment of the stress response in Gallus domesticus after hatching

Cold caused a hypercorticosteronaemia and hypothermia in the 2-day-old chick. Heat did not affect plasma corticosterone concentrations, despite the birds becoming hyperthermic. The 2-day-old chick has a lower resting concentration of corticosterone than the 1-day-old. The minimum effective dose of corticotrophin in stimulating a hypercorticosteronaemia was 0.5 IU/kg in the former, 1.0 IU/kg in the latter. It is concluded that stress non-responsiveness persists for about 48 hr after hatching and that it is due to a temporary inhibition of hypothalamic function.

Dietary ascorbic acid or procaine penicillin and the response of the immature fowl to stressors

1. The effectiveness of dietary ascorbic acid and procaine penicillin in ameliorating the response to the stressors glucagon, corticotrophin, withdrawal of food and water and extremes of environmental temperature, both short and long-term, has been assessed. 2. No consistent changes in the responses were found, but it is suggested that the rates of inclusion of the substances may be important.

An evaluation of adrenal mass and adrenal cholesterol as measures of adrenal activity

1. The effects of twice daily injections of corticotrophin (1 IU/kg body weight) or restriction of food intake to 75% of normal on body mass, adrenal mass and adrenal cholesterol were determined on chicks from 1 to 21 d of age. 2. Only the birds subjected to restricted feeding showed a reduced growth rate. 3. There was no adrenal hypertrophy in birds receiving corticotrophin but in the restricted group there was transient hypertrophy at 2 weeks. 4. Depletion of adrenal cholesterol was noted only in the birds receiving corticotrophin. 5. It is concluded that neither depletion of cholesterol nor hypertrophy is an inevitable consequence of enhanced adrenal cortical activity.

Some responses of the immature fowl to the withdrawal of a stressor

1. Light Sussex chicks were treated with corticotrophin (30 IU/kg) five times a week from hatching to 3 weeks of age. Their responses to the cessation of treatment were determined over the subsequent 2 weeks. 2. Growth rate was impaired by the treatment but had returned to normal after 7 d. 3. Adrenal mass was not affected by treatment but hypertrophy was evident after 14 d of recovery. Relative adrenal weight was greater in the treated group throughout the recovery period. 4. Adrenal cholesterol stores were depleted at the end of treatment, were repleted after 7 d but significantly depleted again at 14 d. 5. Plasma glucose concentration was not affected by treatment but birds were hyperglycaemic 7 and 14 d after treatment had ceased. 6. Plasma free fatty acid (FFA) concentrations were not affected at any time. 7. Plasma corticosterone concentration was not affected by the treatment but 7 d later it was greater in the treated birds. 8. Significant correlations between plasma glucose, FFA and corticosterone were found.

Short-term stressor effects of propranolol

1. The effects of propranolol (10 mg/kg) on plasma cholesterol, glucose, corticosterone and free fatty acids, on adrenal mass and cholesterol content and on body temperature were determined over a 24-h period in 3-week-old Light Sussex chicks. 2. Changes in plasma cholesterol were equivocal, though there was perhaps a trend towards hypercholesteraemia. 3. Birds became hypoglycaemic within an hour but were hyperglycaemic between 2 and 4 h. At 24 h they were normoglycaemic. 4. Plasma corticosterone was increased markedly at 1 h but was decreased compared with the control at 2 h ( P < 0.01). Thereafter the concentration was in the normal range. 5. There was an enhanced hyperlipacidaemia in the treated birds after 1 h. 6. Adrenal cholesterol stores were decreased by 10% at 4 h but were unaffected at all other times. 7. Colonic temperature decreased by 1x6 degrees C after 2 h but was normal by 4 h. 8. It is concluded that propranolol has mild stressor activity which is lost within 24 h.

The effects of repeated injections of adrenaline on the response of the fowl to further alarm stimulation

Chicks, aged two weeks, were injected with either adrenaline (300 microgram/kg) or saline daily for seven days. One day after the seventh injection it was found that the birds treated with adrenaline had become hypolipacidaemic and hypocholesteraemic and that there had been adrenal hypertrophy and an increase in the adrenal store of cholesterol. Plasma concentrations of glucose and corticosterone were within the normal range. The two groups (adrenaline- or saline-treated) were further subdivided each into two subgroups and were now injected with either adrenaline or saline, and their responses measured over a 120 min period. A significantly shorter period of hyperglycaemia was found in the birds pretreated with adrenaline and given a further injection of the hormone. These birds also showed an enhanced lipacidaemic response but the corticosterone response was not altered.

Stressor effects of handling on the immature fowl

Chicks were handled for 5 min twice daily for five consecutive days per week over a three-week period. The control birds were not handled once they had been allotted to their brooders. The growth rate of the treated birds was significantly decreased. There was no adrenal hypertrophy; the concentration, but not the content, of adrenal cholesterol was significantly increased. Plasma glucose, FFA and cholesterol concentrations were similar in the groups at the end of the experiment. The lipolytic response of the handled birds to a further stimulus of handling was significantly greater than that of unhandled birds experiencing novel handling stimulus. There was also a significant decrease in the concentration of corticosterone in the plasma of the handled birds but not in the unhandled group. There was thus no evidence of habituation.

Short-term stressor effects of reserpine

The effects of reserpine on plasma glucose, FFA and cholesterol and on adrenal cholesterol were determined in 1-and 21-d-old chicks over a 24-h period. 2. Irrespective of age, reserpinised chicks became hyperglycaemic and hypocholesteraemic, their stores of adrenal cholesterol were depleted and the mobilisation of the lipid was impaired. The response of the older bird was quantitatively larger than that of the younger bird. 3. Propranolol prevented the increase in plasma glucose concentration seen 1 h after reserpine but it did not affect subsequent changes. 4. It is concluded that reserpine acts as a stressor and it is shown that its effects have not been entirely overcome after 24 h.

Stressor effects of adrenaline on the immature fowl

Chicks were injected with L-adrenaline (500 microgram/kg) thrice weekly from hatching for one, two or three weeks. The growth rate of treated chicks was depressed, particularly in the first week. When four-week-old chicks were given nine injections of adrenaline over a three-week period, growth rate was also depressed but less severely. The greater sensitivity of the younger chicks is thought to result from an immature blood-brain barrier. Relative adrenal mass (mg/kg) was generally increased but adrenal cholesterol depletion occurred only in three-week-old birds treated with adrenaline throughout. Plasma glucose and cholesterole concentrations were within the normal range after three weeks of treatment but there was a significant hypolipacidaemia in the younger birds treated for two or three weeks.

Responses of the immature fowl to a single injection of adrenocorticotrophic hormone

Chicks, aged 1 or 21 d, were injected with long-acting ACTH (6, 12, 30 or 60 IU/kg) and changes in plasma glucose and cholesterol and in adrenal weight and cholesterol measured. 2. No consistent response was observed in the 1-d-old chicks. 3. Hyperglycaemia and adrenal cholesterol depletion could be demonstrated in the 3-week-old bird at all doses. 4. The duration of the hyperglycaemic response was dose-dependent but there was only limited evidence that adrenal cholesterol depletion was similarly dependent. 5. Starving the 3-week-old bird overnight led to a significant hypercholesteraemia which was reduced after ACTH. In contrast no change in plasma cholesterol concentration was noted in fully fed birds.

Failure of procaine penicillin and zinc bacitracin to modify the response of the fowl to stressors

1. The effect of either procaine penicillin or zinc bacitracin (25 mg active base/kg food) on the response of young birds to three different stressors - withdrawal of food for 18 h, exposure to cold for 24 h or treatment with ACTH- has been examined. 2. Neither antibiotic had any consistent effect on the stress responses. 3. It is concluded that neither penicillin nor bacitracin has any stress-ameliorating activity at least under the conditions of the test.

Dietary aureomycin and the response of the fowl to stressors

1. Conventional or gnotobiotic chicks, when injected from 1 d to 3 weeks of age with adrenocorticotrophic hormone (120IU/kg, three times weekly), showed a depressed growth rate, adrenal hypertrophy and depletion of cholesterol form the adrenal glands. 2. Feeding a diet supplemented with aureomycin (10 mg/kg) did not have any consistent ameliorating effect on the response of the stressed bird as judged by the above parameters. 3. It was found that treating germfree chicks with five daily injections of sterile milk on days 3 to 7 did not depress growth rate at any time, nor could differences in adrenal size or cholesterol stores be detected at the end of the 21-d experimental period. The responses were not modified by feeding an aureomycin-supplemented diet.

The response of the immature fowl to multiple injections of adrenaline

Chicks were injected, from one to 21 days of age, with either L-adrenaline (500 mug/kg) or the solvent three times a week. Birds treated with adrenaline showed an imparied growth rate. Relative (mg/kg0.75), but not absolute (mg) adrenal weight was significantly increased. No changes in adrenal cholesterol content or concentration were noted.

The response of the immature fowl to multiple injections of adrenocorticotrophic hormone

1. Treating chicks, from 1 d or 5 weeks of age with ACTH three times weekly for 3 weeks depresses growth and causes adrenal hypertrophy at dose rates of 30 IU/kg or more, and depletion of adrenal cholesterol (greater than 10 IU/kg). 2. Treating chicks five times weekly at a dose rate of 30 IU/kg was as effective as a dose of 120 IU/kg three times weekly. 3. Plasma glucose and FFA concentrations of chicks treated thrice weekly with 120 IU ACTH/kg for 3 weeks were within the normal range. 4. Rhode Island Reds were more sensitive to ACTH than Light Sussex. 5. Adrenal cholesterol stores in normal chicks show significant variations with season.