Adding mRNA to the list of spatially organized components in bacteria

Abstract

Using LNA in situ hybridization, selected mRNAs have been shown to be spatially constrained to their chromosomal loci in two distantly related bacterial organisms. Translating ribosomes are diffusion limited by mRNA association.

This article refers to:

Bacteria are characterized by a lack of sub-cellular membrane enclosed compartments. Are biomolecules then free-floating within the cytoplasm of bacterial cells? No, evidence is accumulating that more and more components are actually spatially and temporally organized, and bacterial chromosomes is a prime example of an elaborate architecture in which specific loci occupy certain subcellular locations.1 Recently, the Jacobs-Wagner laboratory added mRNAs to the list of spatially organized cellular components in the freshwater bacterium Caulobacter cresentus and in Escherichia coli.2 Interestingly, all mRNAs investigated were born, lived, and died at their respective genomic loci contrary to contemporary models of gene transcription.2 These mRNAs,^oø encoding functionally disparate gene products, were confined to one or two sub-cellular foci depending on replication and segregation status, and could not be further resolved due to the limits of microscopy resolution. In the cases examined, mRNA foci overlapped with the position of the corresponding chromosomal locus. These observations are incompatible with free diffusion of mRNAs within the bacterial cytoplasm and instead support a model in which mRNAs, via unknown mechanisms, are anchored in the general vicinity of where they are made.

One key challenge with such measurements is a lack of sufficient sensitivity pertinent to low abundance mRNAs. In initial experiments, an all-recombinant live-cell assay was pursued using C. cresentus. RNA-binding protein from bacteriophage MS2 defective in coat assembly (MS2(dlFG)) was genetically fused to the green fluorescent protein (GFP) for detection. mcherry mRNA equipped with 48 tandem MS2 binding sites in the 3′ untranslated region (mcherry-bs48) to amplify the signal was used as the target. In C. cresentus cells expressing these constructs from chromosomal locations, green fluorescence distributed into a diffuse pattern, and, in some cells, formed fluorescent foci. In cells devoid of mcherry-bs48 mRNA, only diffuse green fluorescence was observed. This argued that mcherry-bs48 mRNA was localized, at least in part, to a specific subcellular site. Detection of the endogenous and naturally abundant groESL mRNA (encoding heat shock proteins GroES and GroEL) was demonstrated in fixed C. cresentus cells using a single Cy3 labeled locked nucleic acid (LNA) probe for RNA fluorescence in situ hybridization (FISH). These experiments also revealed fluorescent foci compatible with restrained dispersion of the groESL mRNA. However, neither method was deemed sufficiently sensitive for quantitative investigations of cellular mRNA distributions.

To enhance sensitivity a FISH strategy using a series of Cy3 labeled DNA oligonucleotides complementary to targets along the mRNA in question was used. This enabled detection of C. cresentus creS mRNA (encoding intermediate filament-like cell shape determinant creS), and of the E. coli lacZ transcript (encoding β-galactosidase). Both transcripts were detected as fluorescent foci. This approach, however, calls for the use of numerous fluorescently labeled custom oligonucleotides thereby increasing the risk of inadvertent binding to non-target mRNAs. As an alternative approach, a series of 120 tandem copies of a lac operator sequence was engineered into the 3′ untranslated region at the chromosomal position of the gene of interest hence reducing the required probe complexity while maintaining sensitivity. This allowed visualization of a number of endogenous chromosomal (although still recombinant) transcripts using a single antisense lacO-Cy3 LNA. All transcripts investigated in this manner (groESL-lacO₁₂₀, creS-lacO₁₂₀, divJ-lacO₁₂₀, ompA-lacO₁₂₀, fljK-lacO₁₂₀, and mcherry-lacO₁₂₀) were detected as distinct foci supporting that chromosomally expressed mRNAs are spatially confined occupying distinct sub-cellular sites. The approach increased the signal-to-noise ratio, reduced the background significantly relative to the multioligo-Cy3 approach, and, strikingly, did not appear to alter mRNA sub-cellular localization or turnover, at least in the case of groESL-lacO120.

These data suggested that translating ribosomes would also be spatially confined through mRNA association. Indeed, fluorescence recovery after photo bleaching experiments showed that ribosomal protein L1-GFP fusions (functional in translation), were dispersion limited in C. cresentus, whereas they were rapidly scattered when the cells were depleted for mRNA by treatment with transcription initiation inhibitor rifampicin prior to photo bleaching.

One might ask whether the above mRNA localization holds true for the entire transcriptome counting more than four thousand genes in E. coli. We do not know at the moment but increased mRNA dispersion is certainly a possibility as observed for groESL-lacO120 in C. cresentus under heat shock conditions increasing the steady state abundance of this transcript to a level where the localization mechanism(s) could be overloaded.2

RNaseE, a component of the RNA degradation machinery, was found distributed along the nucleoid framework further suggesting spatially ordered mRNA decay.2 The relationship between mRNAs destined for degradation and RNaseE warrant further exploration. The spatial dynamics of protein synthesis relative to mRNA transcript also calls for investigation. Perhaps it will soon be possible to explore these questions in live cells. Numerous studies have already established the use of peptide nucleic acid (PNA) oligomers for FISH analyses in fixed bacteria (example in Peleg, et al.6). Equipped with bacterial cell penetrating peptides,7 FISH PNAs could prove useful for spatio-temporal mRNA analyses in live bacteria if background fluorescence can be controlled.

A picture emerges of functionally compartmentalized bacteria with RNA localization and possibly decay organized according to the nucloid structure. Such architecture could facilitate interaction of gene products through proximity as pointed out by Montero, et al.2 Perhaps this provides a complement to the bacterial operon gene architecture enabling co-regulation of functionally interrelated genes. Irrespectively, the data emphasize that bacteria are architecturally highly sophisticated.

Commentary to: