Different RNA recognition by ProQ and FinO depends on the sequence surrounding intrinsic terminator hairpins

Abstract

Escherichia coli ProQ and FinO proteins both have RNA-binding FinO domains, which bind to intrinsic transcription terminators, but each protein recognizes distinct RNAs. To explore how ProQ and FinO discriminate between RNAs, we transplanted sequences surrounding terminator hairpins between RNAs specific for each protein, and compared their binding to ProQ, the isolated FinO domain of ProQ (ProQ^NTD), and FinO. The results showed that the binding specificity of chimeric RNAs toward ProQ, ProQ^NTD, or FinO was determined by the origin of the transplanted sequence. Further analysis showed that the sequence surrounding the terminator hairpin, including a purine-purine mismatch, in natural RNA ligands of FinO and in chimeric RNAs, weakened their binding by ProQ^NTD Overall, our studies suggest that RNA sequence elements surrounding the intrinsic terminator hairpin contribute to the discrimination between RNAs by ProQ and FinO.

Keywords: FinO; FinO domain; ProQ; bacterial regulatory RNA; sRNA.

FIGURE 1.

Comparison of malM-3′ and FinP RNA binding to ProQ, ProQ^NTD, and FinO. (A) Secondary structures of malM-3′ (black font) and FinP RNAs (red font), which were predicted using RNAstructure software (Reuter and Mathews 2010). The lowercase g denotes guanosine residue added on 5′ ends of RNA molecules to enable T7 RNA polymerase transcription. (B–D) The gelshift analysis of malM-3′ and FinP binding to full-length ProQ (B), ProQ^NTD (C), and FinO (D). Free ³²P-labeled RNA is marked as R and RNA-protein complexes as RP. (E) The plots of fraction-bound data versus protein concentration from B to D are shown. The fitting of the quadratic equation into malM-3′ binding data provided a K_d value of 18 nM for binding to ProQ, and 5.5 nM for binding to ProQ^NTD, while the K_d value for binding to FinO was estimated as higher than 200 nM. The fitting of the quadratic equation into FinP binding data provided a K_d value of 74 nM for binding to ProQ, 18 nM for binding to ProQ^NTD, and 109 nM for binding to FinO. The average equilibrium dissociation constant (K_d) values calculated from at least three independent experiments are shown in Table 1.

FIGURE 2.

Comparison of malM-FinP and FinP-malM chimeras binding to ProQ, ProQ^NTD, and FinO. (A) Secondary structures of malM-FinP and FinP-malM chimeras, which were predicted using RNAstructure software (Reuter and Mathews 2010). The sequences originating from malM-3′ are shown in black font and the sequences from FinP in red font. The lowercase g denotes guanosine residue added on 5′ ends of RNA molecules to enable T7 RNA polymerase transcription. (B) The respective binding data for ProQ, the ProQ^NTD, and FinO are shown on the graphs below each RNA. The fitting of the quadratic equation into malM-FinP data provided a K_d value of 85 nM for binding to ProQ and 83 nM for binding to FinO, while the K_d value for binding to ProQ^NTD was estimated as higher than 200 nM. The fitting of the quadratic equation into FinP-malM data provided a K_d value of 27 nM for binding to ProQ, 4.9 nM for binding to ProQ^NTD, and 327 nM for binding to FinO. Gels corresponding to the data in the plots are shown in Supplemental Figure S6. Average K_d values are shown in Table 1.

FIGURE 3.

Comparison of malM-3′, malM-3′-A-GU₆, and malM-3′-A-GAU₅ binding to ProQ^NTD, and to FinO. (A) Secondary structures of malM-3′, malM-3′-A-GU₆, and malM-3′-A-GAU₅, which were predicted using RNAstructure software (Reuter and Mathews 2010). The nucleotides from FinP which were substituted into malM-3′ are shown in red underlined font. The lowercase g denotes guanosine residue added on 5′ ends of RNA molecules to enable T7 RNA polymerase transcription. (B) The respective binding data for ProQ^NTD and FinO are shown on the graphs below each RNA. The fitting of the quadratic equation into malM-3′-A-GU₆ data provided a K_d value of 1.8 nM for binding to ProQ^NTD, and 151 nM for binding to FinO. The fitting of the quadratic equation into malM-3′-A-GAU₅ data provided a K_d value of 122 nM for binding to FinO, while the K_d value for binding to ProQ^NTD was estimated as higher than 200 nM. The data shown for malM-3′ are the same as in Figure 1. Gels corresponding to the data in the plots are shown in Supplemental Figure S7. Average K_d values are shown in Table 1.

FIGURE 4.

Comparison of FinP, FinP-U-UAU₄, and FinP-U-U₆ binding to ProQ^NTD, and to FinO. (A) Secondary structures of FinP, FinP-U-UAU₄, and FinP-U-U₆, which were predicted using RNAstructure software (Reuter and Mathews 2010). The nucleotides from malM-3′ which were substituted into FinP are shown in black underlined font. The lowercase g denotes guanosine residue added on 5′ ends of RNA molecules to enable T7 RNA polymerase transcription. (B) The respective binding data for ProQ^NTD and FinO are shown on the graphs below each RNA. The fitting of the quadratic equation into FinP-U-UAU₄ data provided a K_d value of 4.6 nM for binding to ProQ^NTD, and 184 nM for binding to FinO. The fitting of FinP-U-U₆ data using the quadratic equation provided a K_d value of 11 nM for binding to ProQ^NTD, and 160 nM for binding to FinO. The data shown for FinP are the same as in Figure 1. Gels corresponding to the data in the plots are shown in Supplemental Figure S10. Average K_d values are shown in Table 1.

FIGURE 5.

Comparison of cspE81-3′, cspE81-FinP chimera, and cspE81-FinP-stem chimera binding to ProQ^NTD, and to FinO. (A) Secondary structures of cspE81-3′, cspE81-FinP chimera, and cspE81-FinP-stem chimera, which were predicted using RNAstructure software (Reuter and Mathews 2010). The sequences originating from cspE81-3′ are shown in dark blue font, and the sequences from FinP in red font. The lowercase g denotes guanosine residue added on 5′ ends of RNA molecules to enable T7 RNA polymerase transcription. (B) The respective binding data for ProQ^NTD and FinO are shown on the graphs below each RNA. The fitting of the quadratic equation into cspE81-3′ data provided a K_d value of 2.7 nM for binding to ProQ^NTD, while the K_d value for binding to FinO was estimated as higher than 200 nM. The fitting of the quadratic equation into cspE81-FinP data provided a K_d value of 5.0 nM for binding to ProQ^NTD, and 99 nM for binding to FinO. The fitting of the quadratic equation into cspE81-FinP-stem data provided a K_d value of 55 nM for binding to FinO, while the K_d value for binding to ProQ^NTD was estimated as higher than 200 nM. Gels corresponding to the data in the plots are shown in Supplemental Figure S13. Average K_d values are shown in Table 1.

FIGURE 6.

Comparison of RepX, RepX-malM chimera, and malM-RepX chimera binding to ProQ^NTD, and to FinO. (A) Secondary structures of RepX, RepX-malM chimera, and malM-RepX chimera, which were predicted using RNAstructure software (Reuter and Mathews 2010). The sequences originating from RepX are shown in orange font, and the sequences from malM-3′ in black font. The lowercase g denotes guanosine residue added on 5′ ends of RNA molecules to enable T7 RNA polymerase transcription. (B) The respective binding data for ProQ^NTD and FinO are shown on the graphs below each RNA. The fitting of the quadratic equation into RepX data provided a K_d value of 79 nM for binding to FinO, while the K_d value for binding to ProQ^NTD was estimated as higher than 200 nM. The fitting of the quadratic equation into RepX-malM data provided a K_d value of 3.0 nM for binding to ProQ^NTD, and 148 nM for binding to FinO. The fitting of the quadratic equation into malM-RepX data provided a K_d value of 137 nM for binding to FinO, while the K_d value for binding to ProQ^NTD was estimated as higher than 200 nM. Gels corresponding to the data in the plots are shown in Supplemental Figure S15. Average K_d values are shown in Table 1.

FIGURE 7.

The modeling of RNA-binding surfaces of E. coli ProQ and F-like plasmid FinO. The figure shows the α-helix 3 and surrounding region from E. coli ProQ (left) and the corresponding α-helix 4 from F-like plasmid FinO (right) with those amino acid residues marked, which side chains are directed toward the modeled location of RNA helix. The modeling of interactions was done using Chimera X (Pettersen et al. 2021) by aligning the NMR structure of the FinO domain of E. coli ProQ (Gonzalez et al. 2017) and the X-ray structure of the F-like plasmid FinO protein (Ghetu et al. 2000) with the X-ray structure of the FinO domain of L. pneumophila RocC in complex with the terminator hairpin of RocR RNA (Kim et al. 2022). The side chains of amino acid residues located in the corresponding positions of both proteins are marked in color, with arginine, lysine, and histidine residues marked in red, serine and threonine in green, and tyrosine in orange. The descriptions of corresponding amino acids are located in corresponding places on the figure. Those amino acids which are different, but located in corresponding positions, are marked by underlining. The structure of the L. pneumophila RocR hairpin is shown in gray.

Maria D. Mamońska