Ribonucleoprotein particles of bacterial small non-coding RNA IsrA (IS61 or McaS) and its interaction with RNA polymerase core may link transcription to mRNA fate

Coupled transcription and translation in bacteria are tightly regulated. Some small RNAs (sRNAs) control aspects of this coupling by modifying ribosome access or inducing degradation of the message. Here, we show that sRNA IsrA (IS61 or McaS) specifically associates with core enzyme of RNAP in vivo and in vitro, independently of σ factor and away from the main nucleic-acids-binding channel of RNAP. We also show that, in the cells, IsrA exists as ribonucleoprotein particles (sRNPs), which involve a defined set of proteins including Hfq, S1, CsrA, ProQ and PNPase. Our findings suggest that IsrA might be directly involved in transcription or can participate in regulation of gene expression by delivering proteins associated with it to target mRNAs through its interactions with transcribing RNAP and through regions of sequence-complementarity with the target. In this eukaryotic-like model only in the context of a complex with its target, IsrA and its associated proteins become active. In this manner, in the form of sRNPs, bacterial sRNAs could regulate a number of targets with various outcomes, depending on the set of associated proteins.


IsrA expression
IsrA was fully expressed during the mid-logarithmic growth phase (OD600 of 1.5-1.9) in RL based strains (Supplementary Table), with levels being much lower at OD600 > 2.4, which is different from MG1655, where IsrA expression peaks at an OD600 of 2.5 during the transition to stationary growth (4). The RL rpoCHIS strain routinely used at the time in our lab for RNAP purification, and its parent, RL, have a different genetic background than the commonly used MG1655 (Supplementary Table). We observed a marked difference in the expression levels of IsrA between these strains or their derivatives (cf. Fig. 1E, lanes 3 and 2, 5, 6) during mid-log phase. Also a strain constructed from the original JC7623, the parental strain of RL324 obtained from a stock center (Supplementary Table) expressed very low levels of IsrA, undetectable by Northern blotting in total lysates from cells harvested at an OD600 of 1.5-1.9 (Fig. 1E, lane 4). A reason for this difference in expression, as found by sequencing of the isrA-locus, could be a single mutation in the -10 region that created an optimal promoter for RNAP: TATAAT in the RL324-based strains instead of TATAAC in MG1655 and JC7623 ( Supplementary Fig. S1F). Still, although minor amounts were detected, IsrA co-purifed with RNAP from MG1655 and JC7623 cells (Fig. 1E, lanes 9, 10), demonstrating that the association of this sRNA with RNAP is strain-independent.

Supplementary Figure S1. Schematic overview of recombinant strains and plasmids.
(A) Peptide sequence of the C-terminal region of RpoC with attached the C-terminal region of BCCP (yellow, encoded by nucleotides 211-468 of accB) that will be biotinylated on the boxed lysine. RNAP can be released from streptavidin sepharose by HRV3C cleavage of LEVLFQ/GP (red).
(B) Corresponding nucleotide sequence (top) and schematic overview of the integration cassette.
(D) Schematic of modifying the hfq locus by insertion of a stop-codon linked to the cat marker.
(E) Overview of the isrA gene disruption and (F) of IsrA mutants constructed in pGemT Easy within their genomic context (from -205 to +42).
The -10 region of isrA in RL rpoCHIS and RL contains a T which is a C in MG1655 and JC7623 (red arrow). IsrA (mutants) were detected by Northern hybridization with a 32 P-radioactively end-labeled oligo complementary to nt 4-26 (maroon arrow). Integration cassettes were amplified with primers indicated by purple arrows, and their insertion verified by colony PCR with primers shown as black arrows. Relevant nucleotide positions are given with respect to the start of the coding regions as available via http://ecocyc.org. The numbering of downstream positions is with respect to the end of the coding regions and include the stop codons (asterisk). The genes providing resistance to kanamycin (kan, cloned length is within parentheses), or chloramphenicol (cat), and the c1 primer are described in (1). Figure S2. Phylogenetic analysis of genes encoding IsrA homologs.

Supplementary
(A) Alignment of promoter regions. The -35 and -10 elements in E. coli are indicated, as well as the identified CRP element (4). The nucleotide in the -10 box that is mutated in RL-based strains (to a T), is indicated (red arrow).
(B) Alignment of IsrA genes with the resulting sequence logo at the top. Brackets indicate paired nucleotides according to mfold or according to phylogenetic comparison. Protein binding sites (CsrA) or regions implied in translational regulation of csgD or flhDC mRNAs are indicated and mostly well-conserved. Streptotag of IsrA does not affect its association with RNAP in vivo. Northern blot analysis of RNAs that co-purified with RNAP from strain RL rpoCBCCP ΔisrA in which isrA3-tag was expressed. Figure S4. Curli synthesis is not affected by IsrA overexpression after truncation of Hfq.

Supplementary
Congo-red plate-assay with isogenic strains (expressing RNAP with a biotinylated tag, RL rpoCBCCP ) used in the co-purification experiment (Figs. 1 D, E). Due to a promoter mutation, the parental strain (wt) overexpresses IsrA, which affects curli synthesis as indicated by reduced red-pigment formation (4,5). In the absence of IsrA (ΔisrA) or when a truncated version of Hfq (containing the first 65 amino acids and lacking the C-terminal tail; Supplementary Fig. S1C) is expressed (hfq65), curli-fibers are stainable. The right panels show a serial dilution series (and thereby cells at various growth phases), the left panels independently obtained duplicates of the indicated strains, uniformly spread as large spots. The bottom panel shows the scans before cleanup of surrounding background and digital enhancement, the results of which are shown at the top to better visualize the differences between the strains. Figure S5. Competitor RNAs.

Supplementary
Unlabeled competitor RNAs, separated on 10% PAGE and stained with methylene blue (left) or ethidium bromide (right). Figure S6. Controls to pull-down experiment of Figure 4D.

Supplementary
(A) Secondary structure model of IsrA with altered regions indicated. In IsrA-mutant isrA i2loop (i2loop) the loop closing stem 2 has been replaced with GAAG, whereas in mutants isrA ΔB (ΔB) and isrA iΔ2 (iΔ2) the regions in blue and red, respectively, were deleted (see Supplementary Fig. S1F).
(B) Plasmids from which the mutant IsrA RNAs were expressed, along with an empty vector (pGemT), were transformed into strain RL rpoCBCCP ΔisrA with a disrupted IsrA gene and expressing RNAP with a biotinylated tag that can be removed by HRV3C (3C) protease (Supplementary Table and   Alignment of 5' UTRs of pgaABCD (A), flhDC (B) and csgD (C) with indicated the transcription start site (gray arrow), ribosome binding site (SD), start codon (ATG). Secondary structure elements that are supported by phylogenetic evidence (i.e. compensatory nucleotide changes) are marked by brackets (with in red a stem-loop interfering with the model proposed by (6)). Regions of base pairing interactions with sRNAs IsrA, ArcZ, OxyS, OmrA/OmrB, GcvB and RprA that have been tested by compensatory base pair changes (apart from GcvB) are boxed (as reviewed in (6)). The names of downregulating sRNAs are in red. Asterisks preceding the names for strains on the left indicate that IsrA was not found in the genome of these strains (cf. Supplementary Figure S2D).      global sRNA chaperone HF-I; host factor for RNA phage Q beta replication .