Bacterial Stress Responses Lower mRNA-Protein Level Correlations

By integrating transcriptomics and proteomics data across three human bacterial pathogens under ten stress conditions, this study reveals that environmental stressors significantly weaken mRNA-protein level correlations, particularly for essential genes and those involved in osmotic stress, highlighting the critical role of post-transcriptional regulation in bacterial adaptation.

The Gist

Imagine a bacterial cell as a busy, high-tech factory. Inside this factory, there are two main departments working together to keep the business running: the Blueprint Department (which holds the DNA and makes mRNA copies of the instructions) and the Assembly Line (which reads those instructions to build actual proteins, the workers and machines that do the work).

Usually, these two departments are in sync. If the Blueprint Department sends out 100 new instructions, the Assembly Line builds 100 new machines. It's a predictable relationship: More blueprints = More machines.

However, this new study by Sena Gizem Suer and her team at Umeå University asks a fascinating question: What happens when the factory is under attack?

They looked at three different types of bacterial "factories" (Salmonella, Yersinia, and Staphylococcus) and subjected them to ten different kinds of "hostile environments," like extreme heat, starvation, or high salt (osmotic stress). They wanted to see if the Blueprint Department and the Assembly Line stayed in sync when things got tough.

Here is the simple breakdown of what they found:

1. The "Stress Disconnect"

Under normal, calm conditions, the factory runs smoothly. The number of blueprints (mRNA) matches the number of machines (proteins) quite well.

But when the bacteria face stress, the connection breaks.

• The Analogy: Imagine the Blueprint Department is frantically shouting out new instructions because the factory is on fire. They are printing thousands of new blueprints. But the Assembly Line is confused, overwhelmed, or broken. It can't keep up. So, you have a pile of blueprints on the floor, but no new machines are being built.
• The Result: The scientists found that under stress, the correlation between blueprints and machines drops significantly. You can no longer guess how many machines a bacterium has just by counting its blueprints.

2. The "Essential Workers" vs. The "New Hires"

The study also looked at two types of genes:

• Essential Genes: These are the "vital workers" (like the power plant operators or the security guards) that the bacteria must have to survive.
• Non-Essential Genes: These are the "nice-to-have" workers (like the cafeteria staff or the gardeners).

The Finding: The vital workers were surprisingly stable. Even when the factory was on fire, the Blueprint Department and Assembly Line for these essential genes stayed in sync. They kept producing the necessary machines reliably.

• The Metaphor: When the building is on fire, the fire alarm (essential gene) and the sprinkler system (protein) work perfectly together. But the "Decorations" (non-essential genes) get chaotic; the instructions for them might change wildly, but the actual decorations might not appear, or they might appear even if the instructions say they shouldn't.

3. The "Salt Shock" Surprise

The researchers tested many stressors, but one stood out: Osmotic Stress (like putting the bacteria in a super-salty solution).

• The Analogy: Imagine the factory is suddenly flooded with salt water. The Blueprint Department goes into overdrive, screaming out instructions to build "Salt Shields." But the Assembly Line is physically jammed. The salt has messed up the machinery's ability to start building.
• The Discovery: In this specific scenario, the bacteria had lots of blueprints for salt-shield proteins, but very few actual proteins were made. The factory was screaming instructions, but the workers couldn't hear them or couldn't start the engines.

4. Why Does This Matter?

For a long time, scientists thought: "If we know the blueprint, we know the machine." This study proves that this is not true during an infection.

When bacteria infect a human, they face a hostile environment (low iron, high salt, immune attacks). During these times, they rely on "hidden" regulatory mechanisms—like pausing the assembly line or recycling old machines—to survive.

The Takeaway:
If we want to develop new antibiotics or treatments, we can't just look at the bacterial "blueprints" (mRNA). We have to look at the actual "machines" (proteins) because, under stress, the two tell very different stories. The bacteria are masters of improvisation, and sometimes they stop building new things even while they are frantically writing new instructions.

In short: Bacteria are like a factory that, when under attack, stops listening to its own instructions to survive. To understand how they survive, we have to watch the workers, not just read the memos.

Technical Summary

1. Problem Statement

Bacterial pathogens employ complex regulatory mechanisms to adapt to diverse environmental stresses encountered during infection (e.g., nutrient limitation, pH shifts, hypoxia, osmotic stress). While transcriptomics (mRNA levels) is widely used to study these responses, there is a well-documented but poorly understood weak correlation between mRNA and protein levels. This discrepancy complicates the prediction of bacterial physiology and adaptation strategies based solely on transcriptomic data. Specifically, it is unclear how different infection-relevant stressors modulate the relationship between mRNA abundance and protein synthesis, and which specific regulatory mechanisms (post-transcriptional or post-translational) drive these decoupling events.

2. Methodology

The study employed an integrative multi-omics approach combining transcriptomics and proteomics across three distinct human bacterial pathogens:

• Organisms: Salmonella enterica serovar Typhimurium (Gram-negative), Yersinia pseudotuberculosis (Gram-negative), and Staphylococcus aureus (Gram-positive).
• Conditions: Ten infection-relevant stress conditions (including hypoxia, nutritional downshift, osmotic stress, low iron, etc.) plus an unexposed control.
• Data Sources:
  • Transcriptomics: Utilized existing RNA-seq data from the PATHOgenex RNA Atlas.
  • Proteomics: Generated new mass-spectrometry (LC-MS/MS) proteomic data using TMT11plex labeling and a modified sp3 digestion protocol.
• Experimental Validation:
  • Polysome Profiling: Performed on Y. pseudotuberculosis to assess global translation status under osmotic stress.
  • Puromycin Incorporation Assay: Used to measure active translation efficiency under osmotic stress, low iron, and control conditions.
• Computational Analysis:
  • Normalization: Variance Stabilizing Normalization (VSN) was selected for proteomic data to ensure tight replicate clustering.
  • Correlation Analysis: Pearson correlation coefficients were calculated between log2-transformed mean TPM (mRNA) and Relative Intensity (RI) (protein) values.
  • Clustering: Genes were classified into six clusters based on discordant regulation patterns (e.g., mRNA upregulated but protein unchanged).
  • MOBILE Framework: A machine learning-based pipeline (using Lasso regression) was employed to identify condition-specific mRNA-protein associations and construct Integrated Association Networks (IANs).
  • Functional Enrichment: Gene Ontology (GO) and Over-Representation Analysis (ORA) were used to identify enriched biological processes in discordant gene clusters.

3. Key Contributions

• Multi-Species Comparative Analysis: Provided the first comprehensive comparison of mRNA-protein correlations across three phylogenetically distinct pathogens under a unified set of ten stress conditions.
• Quantification of Stress-Induced Decoupling: Systematically demonstrated that stress conditions significantly reduce the correlation between mRNA and protein levels compared to unstressed controls.
• Identification of Essential Gene Stability: Highlighted that essential genes maintain higher expression levels, lower variance, and stronger mRNA-protein correlations across all conditions compared to non-essential genes.
• Mechanistic Insight into Osmotic Stress: Specifically linked the severe mRNA-protein decoupling observed under osmotic stress to impaired translation initiation, validated by polysome profiling and puromycin assays.
• Framework for Discordant Regulation: Established a classification system for genes exhibiting non-correlating expression patterns, revealing that translation machinery and metabolic enzymes are primary targets of post-transcriptional regulation during stress.

4. Key Results

• General Correlation Trends: Positive correlations between mRNA and protein levels were observed across all species (R values ranging from 0.46 to 0.72), but these correlations were significantly lower under stress conditions.
• Stress-Specific Effects:
  • Hypoxia and Nutritional Downshift: These conditions exhibited the lowest mRNA-protein correlations, suggesting strong post-transcriptional regulation.
  • Osmotic Stress: Showed the most pronounced decoupling. Genes within the osmotic stress-specific stimulon had markedly lower correlations than other genes.
• Essential vs. Non-Essential Genes: Essential genes consistently showed higher expression levels and stronger mRNA-protein correlations with lower variance, underscoring their critical role in survival and adaptability.
• Discordant Gene Clusters:
  • Cluster 1 & 6 (mRNA changed, Protein stable): Enriched for translation machinery components (rRNA, tRNA, ribosomal proteins), suggesting protein stability or translational buffering.
  • Cluster 5 (mRNA up, Protein down): Enriched for central metabolism (TCA cycle), indicating post-transcriptional repression or rapid protein degradation.
• Mechanism of Osmotic Stress Decoupling:
  • MOBILE analysis revealed that osmotic stress-specific associations were enriched for translation-related processes in Y. pseudotuberculosis.
  • Polysome Profiling: Showed no global change in polysome distribution, indicating global translation remained active.
  • Puromycin Assay: Revealed a significant reduction in active translation efficiency specifically under osmotic stress.
  • Conclusion: The decoupling is caused by a bottleneck where transcriptional induction of stress genes outpaces the capacity of the impaired translational machinery, preventing efficient protein synthesis despite high mRNA levels.

5. Significance

• Therapeutic Implications: The study identifies specific gene products and regulatory mechanisms (e.g., translation impairment under osmotic stress) that are critical for bacterial survival in hostile host environments. These could serve as novel targets for antimicrobial therapies.
• Methodological Caution: The findings emphasize that inferring protein abundance solely from transcriptomic data is unreliable during stress responses, particularly for osmotic and hypoxic conditions. Integrated omics are necessary for accurate physiological modeling.
• Evolutionary Insight: The conservation of non-correlating expression patterns across distantly related species suggests that post-transcriptional and post-translational regulatory mechanisms are evolutionarily conserved strategies for stress adaptation.
• Resource Creation: The dataset serves as a valuable resource for training predictive models of protein abundance from mRNA levels, specifically improving model accuracy under stress conditions where standard linear relationships fail.