Massively parallel identification of sequence motifs triggering ribosome-associated mRNA quality control.

Decay of mRNAs can be triggered by ribosome slowdown at stretches of rare codons or positively charged amino acids. However, the full diversity of sequences that trigger co-translational mRNA decay is poorly understood. To comprehensively identify sequence motifs that trigger mRNA decay, we use a massively parallel reporter assay to measure the effect of all possible combinations of codon pairs on mRNA levels in S. cerevisiae. In addition to known mRNA-destabilizing sequences, we identify several dipeptide repeats whose translation reduces mRNA levels. These include combinations of positively charged and bulky residues, as well as proline-glycine and proline-aspartate dipeptide repeats. Genetic deletion of the ribosome collision sensor Hel2 rescues the mRNA effects of these motifs, suggesting that they trigger ribosome slowdown and activate the ribosome-associated quality control (RQC) pathway. Deep mutational scanning of an mRNA-destabilizing dipeptide repeat reveals a complex interplay between the charge, bulkiness, and location of amino acid residues in conferring mRNA instability. Finally, we show that the mRNA effects of codon pairs are predictive of the effects of endogenous sequences. Our work highlights the complexity of sequence motifs driving co-translational mRNA decay in eukaryotes, and presents a high throughput approach to dissect their requirements at the codon level.

Supporting data driven translational patient-centered care using network analysis to visualize symptom distress in children with serious illness.

PURPOSE: There are an increasing number of techniques and tools to improve the capacity for children to relay their perceptions of their symptom experience while undergoing blood and marrow transplant (BMT). Network analysis (NA) is a tool that can illustrate associations between symptoms and the distress they cause. We aimed to develop a biopsychosocial assessment clinical analytic tool to examine symptom relationships for children undergoing BMT to find actionable relationships for intervention to improve clinical outcomes including mood. DESIGN AND METHODS: This pilot study used an analytical mobile application tool to support a wide scope of 15 biopsychosocial symptom distress levels and five mood assessments. Children recorded their symptom distress and mood using the app. NA was used to explore relationships between symptom distress and mood. RESULTS: Four children, 11-14 years old, undergoing BMT used the app daily during hospitalization. We found a strong presence of emotional distress and its associations symptom distress and mood. Multiple symptom associations were identified including associations between the set of symptoms difficulty breathing and fever (0.557), sad and worried (0.429). Of note, pain distress had a strong capacity to bridge other symptoms and was connected directly to many symptoms. PRACTICE IMPLICATIONS: We found the significance of patient struggles with emotional and symptom distress and the importance of this relationship to other clinical outcomes. This provides valuable insights and an improved understanding of the child's symptoms. Our findings support early assessment, intervention, and improved symptom communication to enhance sense of well-being and the child's care experience.

Use of consumer wearables to monitor and predict pain in patients with sickle cell disease.

Background: In sickle cell disease (SCD), unpredictable episodes of acute severe pain, known as vaso-occlusive crises (VOC), disrupt school, work activities and family life and ultimately lead to multiple hospitalizations. The ability to predict VOCs would allow a timely and adequate intervention. The first step towards this ultimate goal is to use patient-friendly and accessible technology to collect relevant data that helps infer a patient's pain experience during VOC. This study aims to: (1) determine the feasibility of remotely monitoring with a consumer wearable during hospitalization for VOC and up to 30 days after discharge, and (2) evaluate the accuracy of pain prediction using machine learning models based on physiological parameters measured by a consumer wearable. Methods: Patients with SCD (≥18 years) who were admitted for a vaso-occlusive crisis were enrolled at a single academic center. Participants were instructed to report daily pain scores (0-10) in a mobile app (Nanbar) and to continuously wear an Apple Watch up to 30 days after discharge. Data included heart rate (in rest, average and variability) and step count. Demographics, SCD genotype, and details of hospitalization including pain scores reported to nurses, were extracted from electronic medical records. Physiological data from the wearable were associated with pain scores to fit 3 different machine learning classification models. The performance of the machine learning models was evaluated using: accuracy, F1, root-mean-square error and area under the receiver-operating curve. Results: Between April and June 2022, 19 patients (74% HbSS genotype) were included in this study and followed for a median time of 28 days [IQR 22-34], yielding a dataset of 2,395 pain data points. Ten participants were enrolled while hospitalized for VOC. The metrics of the best performing model, the random forest model, were micro-averaged accuracy of 92%, micro-averaged F1-score of 0.63, root-mean-square error of 1.1, and area under the receiving operating characteristic curve of 0.9. Conclusion: Our random forest model accurately predicts high pain scores during admission for VOC and after discharge. The Apple Watch was a feasible method to collect physiologic data and provided accuracy in prediction of pain scores.

Quantification of elongation stalls and impact on gene expression in yeast.

Ribosomal pauses are a critical part of cotranslational events including protein folding and localization. However, extended ribosome pauses can lead to ribosome collisions, resulting in the activation of ribosome rescue pathways and turnover of protein and mRNA. While this relationship has been known, there has been little exploration of how ribosomal stalls impact translation duration at a quantitative level. We have taken a method used to measure elongation time and adapted it for use in Saccharomyces cerevisiae to quantify the impact of elongation stalls. We find, in transcripts containing Arg CGA codon repeat-induced stalls, a Hel2-mediated dose-dependent decrease in protein expression and mRNA level and an elongation delay on the order of minutes. In transcripts that contain synonymous substitutions to nonoptimal Leu codons, there is a decrease in protein and mRNA levels, as well as similar elongation delay, but this occurs through a non-Hel2-mediated mechanism. Finally, we find that Dhh1 selectively increases protein expression, mRNA level, and elongation rate. This indicates that distinct poorly translated mRNAs will activate different rescue pathways despite similar elongation stall durations. Taken together, these results provide new quantitative mechanistic insight into the surveillance of translation and the roles of Hel2 and Dhh1 in mediating ribosome pausing events.

Massively parallel identification of sequence motifs triggering ribosome-associated mRNA quality control.

Decay of mRNAs can be triggered by ribosome slowdown at stretches of rare codons or positively charged amino acids. However, the full diversity of sequences that trigger co-translational mRNA decay is poorly understood. To comprehensively identify sequence motifs that trigger mRNA decay, we use a massively parallel reporter assay to measure the effect of all possible combinations of codon pairs on mRNA levels in S. cerevisiae. In addition to known mRNA-destabilizing sequences, we identify several dipeptide repeats whose translation reduces mRNA levels. These include combinations of positively charged and bulky residues, as well as proline-glycine and proline-aspartate dipeptide repeats. Genetic deletion of the ribosome collision sensor Hel2 rescues the mRNA effects of these motifs, suggesting that they trigger ribosome slowdown and activate the ribosome-associated quality control (RQC) pathway. Deep mutational scanning of an mRNA-destabilizing dipeptide repeat reveals a complex interplay between the charge, bulkiness, and location of amino acid residues in conferring mRNA instability. Finally, we show that the mRNA effects of codon pairs are predictive of the effects of endogenous sequences. Our work highlights the complexity of sequence motifs driving co-translational mRNA decay in eukaryotes, and presents a high throughput approach to dissect their requirements at the codon level.

Transcriptional-translational conflict is a barrier to cellular transformation and cancer progression.

We uncover a tumor-suppressive process in urothelium called transcriptional-translational conflict caused by deregulation of the central chromatin remodeling component ARID1A. Loss of Arid1a triggers an increase in a nexus of pro-proliferation transcripts, but a simultaneous inhibition of the eukaryotic elongation factor 2 (eEF2), which results in tumor suppression. Resolution of this conflict through enhancing translation elongation speed enables the efficient and precise synthesis of a network of poised mRNAs resulting in uncontrolled proliferation, clonogenic growth, and bladder cancer progression. We observe a similar phenomenon in patients with ARID1A-low tumors, which also exhibit increased translation elongation activity through eEF2. These findings have important clinical implications because ARID1A-deficient, but not ARID1A-proficient, tumors are sensitive to pharmacologic inhibition of protein synthesis. These discoveries reveal an oncogenic stress created by transcriptional-translational conflict and provide a unified gene expression model that unveils the importance of the crosstalk between transcription and translation in promoting cancer.

Cognitive-motor dual-task interference in adults with sickle cell disease.

BACKGROUND: Sickle cell disease (SCD) is a genetic disorder that causes physical and cognitive impairment due to hemolysis, painful vaso-occlusion episodes, joint avascular necrosis, and strokes. As individuals with SCD age and develop conditions impacting their physical and cognitive function, their ability to multitask successfully and safely may decline. Cognitive-motor dual-task interference occurs when there is deterioration in one or both tasks while dual-tasking relative to single-tasking. Dual-task assessment (DTA) is a valuable measure of physical and cognitive function; however, there is limited data on DTA in adults with SCD. RESEARCH QUESTION: Is DTA a feasible and safe method of measuring physical and cognitive function in adults with SCD? What patterns of cognitive-motor interference occur in adults with SCD? METHODS: We enrolled 40 adults with SCD (mean age 44 years, range 20-71) in a single-center prospective cohort study. We used usual gait speed as the measure of motor performance and verbal fluency (F, A, and S) as the measure of cognitive performance. We measured feasibility as the proportion of consented participants able to complete the DTA. We calculated the relative dual-task effect (DTE %) for each task and identified patterns of dual-task interference. RESULTS: Most consented participants completed the DTA (91%, 40/44) and there were no adverse events. There were 3 main dual-task interference patterns for the first trial using letter 'A': Motor Interference (53%, n = 21), Mutual Interference (23%, n = 9), and Cognitive-Priority Tradeoff (15%, n = 6). For the second trial using letter 'S', there were two main dual-task interference patterns: Cognitive-Priority Tradeoff (53%, n = 21) and Motor Interference (25%, n = 10). STATEMENT OF SIGNIFICANCE: DTA was feasible and safe in adults with SCD. We identified specific patterns of cognitive-motor interference. This study supports further evaluation of DTA as a potentially useful tool to measure physical and cognitive function in ambulatory adults with SCD.

Quantification of elongation stalls and impact on gene expression in yeast.

Ribosomal pauses are a critical part of co-translational events including protein folding and localization. However, extended ribosome pauses can lead to ribosome collisions, resulting in the activation of ribosome rescue pathways and turnover of protein and mRNA. While this relationship has been known, the specific threshold between permissible pausing versus activation of rescue pathways has not been quantified. We have taken a method used to measure elongation time and adapted it for use in S. cerevisiae to quantify the impact of elongation stalls. We find, in transcripts containing Arg CGA codon repeat-induced stalls, a Hel2-mediated dose-dependent decrease in protein expression and mRNA level and an elongation delay on the order of minutes. In transcripts that contain synonymous substitutions to non-optimal Leu codons, there is a decrease in protein and mRNA levels, as well as similar elongation delay, but this occurs through a non-Hel2-mediated mechanism. Finally, we find that Dhh1 selectively increases protein expression, mRNA level, and elongation rate. This indicates that distinct poorly translated codons in an mRNA will activate different rescue pathways despite similar elongation stall durations. Taken together, these results provide new quantitative mechanistic insight into the surveillance of translation and the roles of Hel2 and Dhh1 in mediating ribosome pausing events.

Predicting Pain in People With Sickle Cell Disease in the Day Hospital Using the Commercial Wearable Apple Watch: Feasibility Study.

BACKGROUND: Sickle cell disease (SCD) is a genetic red blood cell disorder associated with severe complications including chronic anemia, stroke, and vaso-occlusive crises (VOCs). VOCs are unpredictable, difficult to treat, and the leading cause of hospitalization. Recent efforts have focused on the use of mobile health technology to develop algorithms to predict pain in people with sickle cell disease. Combining the data collection abilities of a consumer wearable, such as the Apple Watch, and machine learning techniques may help us better understand the pain experience and find trends to predict pain from VOCs. OBJECTIVE: The aim of this study is to (1) determine the feasibility of using the Apple Watch to predict the pain scores in people with sickle cell disease admitted to the Duke University SCD Day Hospital, referred to as the Day Hospital, and (2) build and evaluate machine learning algorithms to predict the pain scores of VOCs with the Apple Watch. METHODS: Following approval of the institutional review board, patients with sickle cell disease, older than 18 years, and admitted to Day Hospital for a VOC between July 2021 and September 2021 were approached to participate in the study. Participants were provided with an Apple Watch Series 3, which is to be worn for the duration of their visit. Data collected from the Apple Watch included heart rate, heart rate variability (calculated), and calories. Pain scores and vital signs were collected from the electronic medical record. Data were analyzed using 3 different machine learning models: multinomial logistic regression, gradient boosting, and random forest, and 2 null models, to assess the accuracy of pain scores. The evaluation metrics considered were accuracy (F1-score), area under the receiving operating characteristic curve, and root-mean-square error (RMSE). RESULTS: We enrolled 20 patients with sickle cell disease, all of whom identified as Black or African American and consisted of 12 (60%) females and 8 (40%) males. There were 14 individuals diagnosed with hemoglobin type SS (70%). The median age of the population was 35.5 (IQR 30-41) years. The median time each individual spent wearing the Apple Watch was 2 hours and 17 minutes and a total of 15,683 data points were collected across the population. All models outperformed the null models, and the best-performing model was the random forest model, which was able to predict the pain scores with an accuracy of 84.5%, and a RMSE of 0.84. CONCLUSIONS: The strong performance of the model in all metrics validates feasibility and the ability to use data collected from a noninvasive device, the Apple Watch, to predict the pain scores during VOCs. It is a novel and feasible approach and presents a low-cost method that could benefit clinicians and individuals with sickle cell disease in the treatment of VOCs.

A nascent peptide code for translational control of mRNA stability in human cells.

Stability of eukaryotic mRNAs is associated with their codon, amino acid, and GC content. Yet, coding sequence motifs that predictably alter mRNA stability in human cells remain poorly defined. Here, we develop a massively parallel assay to measure mRNA effects of thousands of synthetic and endogenous coding sequence motifs in human cells. We identify several families of simple dipeptide repeats whose translation triggers mRNA destabilization. Rather than individual amino acids, specific combinations of bulky and positively charged amino acids are critical for the destabilizing effects of dipeptide repeats. Remarkably, dipeptide sequences that form extended ß strands in silico and in vitro slowdown ribosomes and reduce mRNA levels in vivo. The resulting nascent peptide code underlies the mRNA effects of hundreds of endogenous peptide sequences in the human proteome. Our work suggests an intrinsic role for the ribosome as a selectivity filter against the synthesis of bulky and aggregation-prone peptides.

Translational buffering by ribosome stalling in upstream open reading frames.

Upstream open reading frames (uORFs) are present in over half of all human mRNAs. uORFs can potently regulate the translation of downstream open reading frames through several mechanisms: siphoning away scanning ribosomes, regulating re-initiation, and allowing interactions between scanning and elongating ribosomes. However, the consequences of these different mechanisms for the regulation of protein expression remain incompletely understood. Here, we performed systematic measurements on the uORF-containing 5' UTR of the cytomegaloviral UL4 mRNA to test alternative models of uORF-mediated regulation in human cells. We find that a terminal diproline-dependent elongating ribosome stall in the UL4 uORF prevents decreases in main ORF protein expression when ribosome loading onto the mRNA is reduced. This uORF-mediated buffering is insensitive to the location of the ribosome stall along the uORF. Computational kinetic modeling based on our measurements suggests that scanning ribosomes dissociate rather than queue when they collide with stalled elongating ribosomes within the UL4 uORF. We identify several human uORFs that repress main ORF protein expression via a similar terminal diproline motif. We propose that ribosome stalls in uORFs provide a general mechanism for buffering against reductions in main ORF translation during stress and developmental transitions.

Careful considerations for mHealth app development: lessons learned from QuestExplore.

Background: Mobile health (mHealth) applications (app) have proven to be useful in gathering symptom data for a variety of populations living with chronic and serious illnesses. These mHealth tools have been built for a variety of populations but can quickly lose their novelty over time due to the lack of changes and engagement between the mHealth tool and the user. High costs, constantly changing timelines, and difficulties in building compliant data storage systems are some of the reasons why mHealth development and implementation can be a challenge. Methods: Our team's tool, QuestExplore (QE), was built in collaboration with healthcare providers, child-life specialists, a music therapist, mobile app developers, data specialists, cyber security specialists, researchers, and children living with chronic illnesses alongside their families. Through this process, our team learned various ways to reduce costs, streamline the app development process, and build compliant data storage systems. In addition, our frequent interactions with stakeholders provided us with the ability to continuously make improvements, to build an engaging mHealth app. Results: Based upon our findings, our team needed to include prompting, condensing, gamification, data visualizations, and an engaging user design in the remodel of QE. Through a three-stage process of redesigning our previous symptom monitoring apps, QE was developed to better communicate between our users and their providers, with the overall hope of improving symptom management of these children. Conclusions: In the paper, we aim to explain how our team developed QE with feedback from our stakeholders, while also continuously improving our development process through the lessons we gained through the app's development. QE is now being used in both Duke University and the University of North Carolina at Chapel Hill and will soon be implemented in Amsterdam University Medical Center.

The amino acid sensor GCN2 suppresses terminal oligopyrimidine (TOP) mRNA translation via La-related protein 1 (LARP1).

La-related protein 1 (LARP1) has been identified as a key translational inhibitor of terminal oligopyrimidine (TOP) mRNAs downstream of the nutrient sensing protein kinase complex, mTORC1. LARP1 exerts this inhibitory effect on TOP mRNA translation by binding to the mRNA cap and the adjacent 5'TOP motif, resulting in the displacement of the cap-binding protein eIF4E from TOP mRNAs. However, the involvement of additional signaling pathway in regulating LARP1-mediated inhibition of TOP mRNA translation is largely unexplored. In the present study, we identify a second nutrient sensing kinase GCN2 that converges on LARP1 to control TOP mRNA translation. Using chromatin-immunoprecipitation followed by massive parallel sequencing (ChIP-seq) analysis of activating transcription factor 4 (ATF4), an effector of GCN2 in nutrient stress conditions, in WT and GCN2 KO mouse embryonic fibroblasts, we determined that LARP1 is a GCN2-dependent transcriptional target of ATF4. Moreover, we identified GCN1, a GCN2 activator, participates in a complex with LARP1 on stalled ribosomes, suggesting a role for GCN1 in LARP1-mediated translation inhibition in response to ribosome stalling. Therefore, our data suggest that the GCN2 pathway controls LARP1 activity via two mechanisms: ATF4-dependent transcriptional induction of LARP1 mRNA and GCN1-mediated recruitment of LARP1 to stalled ribosomes.

Coordinated missplicing of TMEM14C and ABCB7 causes ring sideroblast formation in SF3B1-mutant myelodysplastic syndrome.

SF3B1 splicing factor mutations are near-universally found in myelodysplastic syndromes (MDS) with ring sideroblasts (RS), a clonal hematopoietic disorder characterized by abnormal erythroid cells with iron-loaded mitochondria. Despite this remarkably strong genotype-to-phenotype correlation, the mechanism by which mutant SF3B1 dysregulates iron metabolism to cause RS remains unclear due to an absence of physiological models of RS formation. Here, we report an induced pluripotent stem cell model of SF3B1-mutant MDS that for the first time recapitulates robust RS formation during in vitro erythroid differentiation. Mutant SF3B1 induces missplicing of ∼100 genes throughout erythroid differentiation, including proposed RS driver genes TMEM14C, PPOX, and ABCB7. All 3 missplicing events reduce protein expression, notably occurring via 5' UTR alteration, and reduced translation efficiency for TMEM14C. Functional rescue of TMEM14C and ABCB7, but not the non-rate-limiting enzyme PPOX, markedly decreased RS, and their combined rescue nearly abolished RS formation. Our study demonstrates that coordinated missplicing of mitochondrial transporters TMEM14C and ABCB7 by mutant SF3B1 sequesters iron in mitochondria, causing RS formation.

Pain Intensity Assessment in Sickle Cell Disease Patients Using Vital Signs During Hospital Visits.

Pain in sickle cell disease (SCD) is often associated with increased morbidity, mortality, and high healthcare costs. The standard method for predicting the absence, presence, and intensity of pain has long been self-report. However, medical providers struggle to manage patients based on subjective pain reports correctly and pain medications often lead to further difficulties in patient communication as they may cause sedation and sleepiness. Recent studies have shown that objective physiological measures can predict subjective self-reported pain scores for inpatient visits using machine learning (ML) techniques. In this study, we evaluate the generalizability of ML techniques to data collected from 50 patients over an extended period across three types of hospital visits (i.e., inpatient, outpatient and outpatient evaluation). We compare five classification algorithms for various pain intensity levels at both intra-individual (within each patient) and inter-individual (between patients) level. While all the tested classifiers perform much better than chance, a Decision Tree (DT) model performs best at predicting pain on an 11-point severity scale (from 0-10) with an accuracy of 0.728 at an inter-individual level and 0.653 at an intra-individual level. The accuracy of DT significantly improves to 0.941 on a 2-point rating scale (i.e., no/mild pain: 0-5, severe pain: 6-10) at an inter-individual level. Our experimental results demonstrate that ML techniques can provide an objective and quantitative evaluation of pain intensity levels for all three types of hospital visits.

Seriously ill pediatric patient, parent, and clinician perspectives on visualizing symptom data.

OBJECTIVE: This study examined the perspectives on the use of data visualizations and identified key features seriously ill children, their parents, and clinicians prefer to see when visualizing symptom data obtained from mobile health technologies (an Apple Watch and smartphone symptom app). MATERIALS AND METHODS: Children with serious illness and their parents were enrolled into a symptom monitoring study then a subset was interviewed for this study. A study team member created symptom data visualizations using the pediatric participant's mobile technology data. Semi-structured interviews were conducted with a convenience sample of participants (n = 14 children; n = 14 parents). In addition, a convenience sample of clinicians (n = 30) completed surveys. Pediatric and parent participants shared their preferences and perspectives on the symptom visualizations. RESULTS: We identified 3 themes from the pediatric and parent participant interviews: increased symptom awareness, communication, and interpretability of the symptom visualizations. Clinicians preferred pie charts and simple bar charts for their ease of interpretation and ability to be used as communication tools. Most clinicians would prefer to see symptom visualizations in the electronic health record. DISCUSSION: Mobile health tools offer a unique opportunity to obtain patient-generated health data. Effective, concise symptom visualizations can be used to synthesize key clinical information to inform clinical decisions and promote patient-clinician communication to enhance symptom management. CONCLUSIONS: Effectively visualizing complex mobile health data can enhance understanding of symptom dynamics and promote patient-clinician communication, leading to tailored personalized symptom management strategies.

Metabolic excretion associated with nutrient-growth dysregulation promotes the rapid evolution of an overt metabolic defect.

In eukaryotes, conserved mechanisms ensure that cell growth is coordinated with nutrient availability. Overactive growth during nutrient limitation ("nutrient-growth dysregulation") can lead to rapid cell death. Here, we demonstrate that cells can adapt to nutrient-growth dysregulation by evolving major metabolic defects. Specifically, when yeast lysine-auxotrophic mutant lys- encountered lysine limitation, an evolutionarily novel stress, cells suffered nutrient-growth dysregulation. A subpopulation repeatedly evolved to lose the ability to synthesize organosulfurs (lys-orgS-). Organosulfurs, mainly reduced glutathione (GSH) and GSH conjugates, were released by lys- cells during lysine limitation when growth was dysregulated, but not during glucose limitation when growth was regulated. Limiting organosulfurs conferred a frequency-dependent fitness advantage to lys-orgS- by eliciting a proper slow growth program, including autophagy. Thus, nutrient-growth dysregulation is associated with rapid organosulfur release, which enables the selection of organosulfur auxotrophy to better tune cell growth to the metabolic environment. We speculate that evolutionarily novel stresses can trigger atypical release of certain metabolites, setting the stage for the evolution of new ecological interactions.

The serine transporter SdaC prevents cell lysis upon glucose depletion in Escherichia coli.

The amino acid serine plays diverse metabolic roles, yet bacteria actively degrade exogenously provided serine via deamination to pyruvate. Serine deamination is thought to be a detoxification mechanism due to the ability of serine to inhibit several biosynthetic reactions, but this pathway remains highly active even in nutrient-replete conditions. While investigating the physiological roles of serine deamination in different growth conditions, we discovered that Escherichia coli cells lacking the sdaCB operon, which encodes the serine transporter SdaC and the serine deaminase SdaB, lyse upon glucose depletion in a medium containing no exogenous serine but all other amino acids and nucleobases. Unexpectedly, this lysis phenotype can be recapitulated by deleting sdaC alone and can be rescued by heterologous expression of SdaC. Lysis of ΔsdaC cells can be prevented by omitting glycine from the medium, inhibiting the glycine cleavage system, or by increasing alanine availability. Together, our results reveal that the serine transporter SdaC plays a critical role in maintaining amino acid homeostasis during shifts in nutrient availability in E. coli.

Inverted translational control of eukaryotic gene expression by ribosome collisions.

The canonical model of eukaryotic translation posits that efficient translation initiation increases protein expression and mRNA stability. Contrary to this model, we find that increasing initiation rate can decrease both protein expression and stability of certain mRNAs in the budding yeast Saccharomyces cerevisiae. These mRNAs encode a stretch of polybasic residues that cause ribosome stalling. Our computational modeling predicts that the observed decrease in gene expression at high initiation rates occurs when ribosome collisions at stalls stimulate abortive termination of the leading ribosome or cause endonucleolytic mRNA cleavage. Consistent with this prediction, the collision-associated quality-control factors Asc1 and Hel2 (orthologs of human RACK1 and ZNF598, respectively) decrease gene expression from stall-containing mRNAs only at high initiation rates. Remarkably, hundreds of S. cerevisiae mRNAs that contain ribosome stall sequences also exhibit lower translation efficiency. We propose that inefficient translation initiation allows these stall-containing endogenous mRNAs to escape collision-stimulated reduction in gene expression.