Quantitative analyses of interactions between SpoVG and RNA/DNA

Abstract

The Borrelia burgdorferi SpoVG protein has previously been found to be a DNA- and RNA-binding protein. To aid in the elucidation of ligand motifs, affinities for numerous RNAs, ssDNAs, and dsDNAs were measured and compared. The loci used in the study were spoVG, glpFKD, erpAB, bb0242, flaB, and ospAB, with particular focus on the untranslated 5′ portion of the mRNAs. Performing binding and competition assays yielded that the 5′ end of spoVG mRNA had the highest affinity while the lowest observed affinity was to the 5’ end of flaB mRNA. Mutagenesis studies of spoVG RNA and ssDNA sequences suggested that the formation of SpoVG-nucleic acid complexes are not entirely dependent on either sequence or structure. Additionally, exchanging uracil for thymine in ssDNAs did not affect protein-nucleic acid complex formation.

1. Introduction

SpoVG proteins are conserved across diverse groups of bacteria, including Firmicutes, Spirochaetota, and Myxococcota (Delta-proteobacteria) [1]. The protein’s nomenclature derives from early studies of Bacillus spp., where it was observed that mutations within the spoVG locus impaired sporulation at stage V [2]. The SpoVG homologues of Borrelia burgdorferi and other non-sporulating bacterial species affect expression of a broad range of genes, and spoVG mutants exhibit physiological defects [3–8]. Our group has studied SpoVG proteins from B. burgdorferi, Staphylococcus aureus, and Listeria monocytogenes, and found that they exhibit affinities for specific DNA sequences [1]. Subsequent studies by Burke and Portnoy found that L. monocytogenes SpoVG exhibits a substantially greater affinity for single-stranded RNAs than for cognate double-stranded DNAs (dsDNAs), and hypothesized that SpoVG functions primarily as an RNA-binding protein [3]. Evaluation of the B. burgdorferi homologue found a similarly greater affinity for RNAs [4].

The Lyme disease spirochete, B. burgdorferi, is a particularly useful model for studies of gene and protein regulation in a vector-borne pathogen [9–11]. In addition to being the cause of a significant human disease, B. burgdorferi infects both vertebrates and ticks [12]. The Lyme spirochete must, therefore, possess mechanisms to accurately determine which host it is in, and produce proteins and other essential factors for each stage of its infectious cycle. B. burgdorferi must also recognize when a tick is feeding, in order to undergo the physiological changes that are necessary for transmission into the vertebrate host. Our group’s previous data indicate that B. burgdorferi SpoVG affects expression levels of numerous proteins and affects bacterial physiology, apparently through its activities as an RNA- and/or DNA-binding protein [4]. To gain further insights into SpoVG function, we identified high- and low-affinity ligands of B. burgdorferi SpoVG, which were then employed to gain insights on the ability of SpoVG to bind single-stranded and double-stranded nucleic acids.

2. Materials and methods

2.1. Purification of recombinant proteins

Polyhistidine-tagged B. burgdorferi SpoVG was purified essentially as described previously [1,4]. Briefly, Escherichia coli Rosetta II (Invitrogen, MA) was transformed with pBLJ132, which consists of spoVG cloned into pET101 [1]. E. coli were then grown to an OD₆₀₀ of at least 1.0 in Super Broth (32 g Tryptone, 20 g Yeast Extract, and 5 g NaCl per liter), and recombinant SpoVG expression was induced for 1h by adding isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. Bacteria were harvested by centrifugation at 5400×g for 30 min and frozen at −80 °C until needed. Resuspended cells were lysed by sonication with the addition of B-PER bacterial protein extraction reagent to 2% v/v (Thermo-Fisher, MA). Recombinant proteins were purified using MagneHis nickel particles (Promega, WI), then dialyzed against EMSA buffer (50 mM Tris-HCl, 25mMKCl, 10% glycerol (v/v), 0.01% Tween 20, 100 nM dithiothreitol (DTT), and 1 mM phenylmethanesulfonyl fluoride (PMSF)). Proteins were concentrated using 10 kDa Amicon centrifugal units (MilliporeSigma, MA) and aliquots were stored at −80 °C until needed. Protein purity and concentration were assessed by SDS-PAGE, Quick Start Bradford protein assay (Bio-Rad), and bicinchoninic acid assay (BCA) (Thermo-Fisher, MA).

2.2. Electromobility shift assay (EMSA)

Fluorescently tagged and untagged DNA and RNA oligonucleotides were synthesized by Integrated DNA Technologies (IDT, IA). The sequences of oligonucleotides used in this study are listed in Table 1 and Supplemental Table 1. DNA and RNA probes were tagged on the 5’ end with an IRDye 800 fluorescent tag (LI-Cor, NE) or an Alexa Fluor 488, respectively. Double-stranded DNA (dsDNA) probes and competitors were produced from pairs of separately-synthesized complimentary oligonucleotides by mixing equal molar concentrations, heating to 95 °C, and slowly cooling to room temperature.

Table 1.

5′-3′ sequences of dsDNA, ssDNA and RNA used throughout. Red highlighted nucleotides in the ssDNA and RNA sections deviate from the spoVG probe. ss22 and ss23 have a deoxyribose backbone with uracil bases instead thymine.

Name	Sequence
dsDNA
SpoVG DNA	AGTTATGTACTTTTTGCGGGAGGCTTATAA
erpAB DNA	CTTATGGAGAAATTTATGAATAAGAAAATGAAA
glpFKD DNA	ATTAAATATAATTTTAATAAGGCTTTTATTAGAAAAATTAAT
bb0242 DNA	GTATTCAAAAAATAAACTGTCTAAACCTTTTGAAAAGG
OSpAB DNA	GTATTAAGTTATATTAATATAAAAGGAGAATATATT
flaB DNA	AGGCAAAAGGATTTGCCAAAGTCAGAAATT
ssDNA
spoVG CD	TATGTACTTTTTGCGGGAGGCTTATAA
SpoVG NC	TTATAAGCCTCCCGCAAAAAGTACATA
ss01	AATGTACTTTTTGCGGGAGGCTTATAA
ss02	GATGTACTTTTTGCGGGAGGCTTATAA
ss03	TATGTACAAAAAGCGGGAGGCTTATAA
ss04	TATGTACCCCCCGCGGGAGGCTTATAA
ss05	TATGTACGGGGGGCGGGAGGCTTATAA
ss06	TATGTACTTTTTGCGGGTGGCTTATAA
ss07	TATGTACTTTTTGCGGGTCCGAAATAA
ss08	TATGTACTTTTTGCGGGGAATCCATAA
ss09	GTACTTTTTGCGGGAGGCTTATAA
ss10	CTTTTTGCGGGAGGCTTATAA
ss11	TATGTACATTTTGCGGGAGGCTTATAA
ss12	TATGTACTTATTGCGGGAGGCTTATAA
ss13	TATGTACTTTTAGCGGGAGGCTTATAA
ss14	TATGTACAATTTGCGGGAGGCTTATAA
ss15	TATGTACTAATTGCGGGAGGCTTATAA
ss16	TATGTACTTAATGCGGGAGGCTTATAA
ss17	TATGTACTTTAAGCGGGAGGCTTATAA
ss18	TATGTACAAATTGCGGGAGGCTTATAA
ss19	TATGTACTAAATGCGGGAGGCTTATAA
ss20	TATGTACTTAAAGCGGGAGGCTTATAA
ss21	TATGTACAAAAAGCGGGTCCGAAATAA
ss22	TATGTACUUUUUGCGGGAGGCTTATAA
ss23	UAUGUACUUUUUGCGGGAGGCUUAUAA
RNA
SpoVG RNA	UAUGUACUUUUUGCGGGAGGCUUAUAA
erpAB RNA	AUGGAGAAAUUUAUGAAUAAGAAAAUGAAA
glpFKD RNA	AUAAUUUUAAUAAGGCUUUUAUUAGAAAAAUUAAU
bb0242 RNA	GUAUUCAAAAAAUAAACUGUCUAAACCUUUUGAAAAGG
ospAB RNA	GUAUUAAGUUAUAUUAAUAUAAAAGGAGAAUAUAUU
flaB RNA	AGGCAAAAGGAUUUGCCAAAGUCAGAAAUU
RNA-03	UAUGUACAAAAAGCGGGAGGCUUAUAA
RNA-17	UAUGUACUUUAAGCGGGAGGCUUAUAA
RNA-20	UAUGUACUUAAAGCGGGAGGCUUAUAA

EMSAs were performed as previously described [1,4]. Unless stated otherwise, purified proteins were added to 100 nM of labeled nucleic acid probes and incubated at room temperature for 5–15 min. Unlabeled nucleic acid competitors were added prior to the addition of protein. EGTA was added to the indicated reactions to inhibit RNAse activity in protein aliquots. Poly-dI-dC was added to indicated gels as a non-specific competitor. One-sixth volume of EMSA loading dye (15 mg/mL Ficol 400, 0.8 mg/mL Orange G) was added. Electrophoresis was performed using pre-run 6% TBE gels (Invitrogen, MA) at 100-V in 0.5x TBE buffer. Gels were imaged and analyzed via densitometry with a ChemiDoc MP and Image-Lab software, respectively (Bio-Rad, CA). Lanes and bands were added manually according to the strongest free probe and the strongest shifted band. Equivalent bands were then added across the entire EMSA.’ Background noise was calculated and accounted for by the Image-Lab software. Percent shifted was graphed using GraphPad Prism 9 (Dotmatics, MA).

2.3. Calculation of apparent K_D and IC50 values

For direct binding assays, the fraction of bound complex was fit to a one-site saturation ligand binding model (Fraction Bound = Bmax SpoVG/(Kd + SpoVG)) where Bmax is the maximum fraction bound and K_D is the apparent dissociation equilibrium constant. [13]. It should be noted that this estimate of the binding constant is limited by the high concentration of nucleic acid in the binding assays which likely puts acids in the “Titration Regime” but can still be used to determine relative affinities of one sequence to another [13]. In competition assays, data were fit globally to determine the concentration of inhibitor resulting in 50% of the complex being disrupted (IC50) to the equation fraction bound to labeled probe = m – (m-b)/(1+IC50/[C]) where m is the maximum signal, b is the minimum (background) signal, [C] is the concentration of the unlabeled nucleic acid competitor, and IC50 is the concentration of the competitor that results in 50% of the complex being disrupted.

3. Results

3.1. SpoVG is a site-specific RNA- and DNA-binding protein

We previously reported that B. burgdorferi SpoVG can bind to the 5′ ends of the borrelial spoVG and glpFKD mRNAs, and then cognate dsDNAs [4]. It was also found that SpoVG binds within the transcript of a small gene of unknown function, bb0242, that is located between the glpK and glpD genes [4]. The present studies delved further into the interactions between SpoVG and other nucleic acid sequences to assess relative affinities and mechanisms of binding. Noting that spoVG and the glpFKD operon are highly expressed during tick colonization but not during mammalian infection, we investigated the ability of SpoVG to bind other borrelial sequences; specifically the 5’ ends of erpAB (which is repressed during tick colonization but highly expressed during mammalian infection), ospAB (which is highly expressed during tick colonization but repressed in mammals), and flaB (which is constitutively expressed throughout the borrelial tick-mammal cycle) [4,12,14–18]. Both RNA and dsDNA were examined for each target.

Of the tested probes and competitors, SpoVG exhibited the greatest affinity for the spoVG mRNA 5′ end, henceforth referred to as spoVG_RNA (Fig. 1 and Fig. 2). This high affinity binding site was further tested by competition EMSAs using an unlabeled spoVG_RNA as a competitor (Fig. 1A and Supp. Fig. 1). No intermediate shift was observed indicating that there was no formation of a dsRNA probe. The protein-RNA complex shift stayed above 50% until 1 μM (10x) of competitor was added. Addition of up to 100 ng/μL poly-dI-dC did not result in an appreciable difference of SpoVG binding to the spoVG_RNA (Fig. 1B), further indicating that binding of SpoVG to spoVG_RNA is specific. SpoVG also bound to the glpFKD_RNA, bb0242_RNA, and erpAB_RNA probes, but EMSAs did not approach saturation in the tested ranges of SpoVG protein concentrations, indicating that the affinity of SpoVG for those three RNAs is weaker than spoVG_RNA (Fig. 2).. (Insert Figs. 1 and 2)

Fig. 1. spoVG RNA is a SpoVG high affinity binding site.

A. Competition assay using 100 nM labeled spoVG_RNA. Free probe is indicated, and the shifted probe is indicated by an asterisk. SpoVG concentration was 500 nM in lanes 2–9. Unlabeled spoVG_RNA was added at 10 nM, 25 nM, 100 nM, 250 nM, 1 μM, 2.5 μM, and 10.0 μM in lanes 3–9 respectively. EGTA was added to 5 mM. This assay was done in triplicate (Sup. Fig. 1). B. Competition assay using labeled spoVG_DNA and spoVG_RNA probe against poly-dI-dC. Lanes 1–7 included 100 nM spoVG_DNA probe. Lanes 8–14 included 100 nM spoVG_RNA probe. SpoVG was added to lanes 2–7 at 10 μM and lanes 9–14 at 100 nM. Poly-dI-dC was added in at 1,5,10,50, and 100 ng/μL in lanes 3–7 and 10–14 respectively.

Fig. 2. SpoVG has a higher binding affinity for RNA than DNA.

A. EMSA using labeled spoVG_RNA, erpAB_RNA, glpFKD_RNA, and bb0242_RNA at 100 nM from lanes 1–5, 6–10, and 11–15, 16–20 respectively. Free probe is indicated, and the shifted probe is indicated by an asterisk. SpoVG was added to lanes 2–5, 7–10, 12–15, and 17–20 at 50 nM, 100 nM, 200 nM, and 500 nM. B. EMSAs using labeled spoVG_DNA, glpFKD_DNA, erpAB_DNA, and bb0242_DNA at 100 nM from 1 to 5, 6–10, 11–15, 16–20 respectively. SpoVG was added to lanes 2–5, 7–10, 12–15, and 17–20 at 500 nM, 1 μM, 5 μM, and 10 μM respectively. C. Graph representing the percentage of 100 nM probe shifted due to SpoVG binding from the EMSAs in A and B.

Our initial studies of B. burgdorferi and S. aureus SpoVG found that both proteins preferentially bound to certain dsDNA sequences [1]. In contrast, a study of the L. monocytogenes orthologue with a variety of DNA probes led those researchers to conclude that DNA binding appeared to be non-specific [3]. To build upon those observations, we assessed the affinities of B. burgdorferi SpoVG for the above-described borrelial sequences as dsDNA. Affinities for all tested dsDNAs were significantly lower than their cognate RNAs (Fig. 2). The relative affinity of SpoVG to the spoVG_DNA probe was over 100x weaker than for spoVG_RNA, with addition of 10 μM SpoVG protein yielding approximately 30% shifted probe (Fig. 2B). This difference signifies that a substantially higher concentration of SpoVG is required to form a complex with SpoVG_DNA dsDNA compared to SpoVG_RNA.

These results were highlighted by performing EMSAs of the spoVG_DNA, spoVG_RNA, glpFKD_DNA, and glpFKD_RNA probes (Supp. Fig. 2). EMSAs with spoVG RNA achieved close to 100% shift at 200 nM compared to the unshifted dsDNA at 200 nM concentration (Supp. Fig. 3A). This was also observed with the glpFKD_RNA. Saturation of the glpFKD_RNA probe occurred between 1 μM and 10 μM SpoVG, whereas the glpFKD_DNA probe showed almost no detectable shift at that concentration. While SpoVG had an inherently weaker interaction with the DNAs tested than the RNAs, there were still substantial differences when comparing affinities for the DNA probes out of the four tested. SpoVG had the lowest relative affinity for the erpAB_DNA probe (Fig. 2B and C).

The relative affinities of SpoVG for the ospAB_RNA and flaB_RNA were determined by use of unlabeled RNAs as competitors against labeled spoVG_RNA or glpFKD_RNA probes. When using the spoVG_RNA probe, addition of 100x excess of the unlabeled ospAB_RNA reduced the shifted protein-RNA complex by approximately 18% (IC50 = 5.9±4.6 × 10^–5 M), while 100x excess of the unlabeled flaB_RNA did not detectably affect the SpoVG-spoVG_RNA complex (Fig. 3A and B). When labeled glpFKD_RNA was used as a probe, addition of 25x excess of unlabeled flaB_RNA led to a 20% reduction in the SpoVG-glpFKD_RNA complex (Fig. 3C). In contrast, 25x excess of unlabeled glpFKD_RNA, erpAB_RNA, or ospAB_RNA reduced binding to the probe by approximately 80%, and unlabeled spoVG_RNA virtually eliminated the SpoVG-glpFKD_RNA complex (Fig. 3C). The data above indicate that SpoVG has a substantially greater affinity for RNA compared to DNA, but that this affinity differs based on the sequence of the nucleic acid.

Fig. 3. SpoVG is a specific RNA binding protein.

A. Competition assay using 100 nM labeled spoVG_RNA. Free probe is indicated, and the shifted probe is indicated by an asterisk. The SpoVG concentration was 500 nM in lanes 2–9. Unlabeled ospAB_RNA was added at 10 nM, 25 nM, 100 nM, 250 nM, 1 μM, 2.5 μM, and 10.0 μM in lanes 3–9 respectively. EGTA was added to 5 mM. B. Competition assay using unlabeled flaB_RNA and 100 nM labeled spoVG_RNA. Lane values match (A), except unlabeled flaB_RNA was used instead of ospAB_RNA. C. Competition assay using 100 nM labeled glpFKD_RNA (lanes 8–14). SpoVG concentration was 500 nM. Competitors were added at a concentration of 2.5 μM in the following sequence glpFKD_RNA, spoVG_RNA, erpAB_RNA, bb0242_RNA, and flaB_RNA. The gamma of entire image was increased to more accurately visualize differences in competition.

3.2. Effects of nucleic acid sequence on SpoVG-binding

As the SpoVG protein appears to have the strongest affinity for the spoVG_RNA probe, we made a series of sequence variants to determine whether the sequence or structure of the RNA was important in recognition and affinity of the protein to the nucleic acid. EMSAs were undertaken using the labeled spoVG_RNA probe and numerous unlabeled variant nucleic acids as competitors (Table 1). As a cost saving first approach, ssDNA competitors were used for the initial analyses. Sequence variations in competitors included changing the 5′ nucleotide, removing 3 or 6 nucleotides from the 5’ end, altering the run of five consecutive thymines, and altering an AGGCT sequence near the 3′ end (Table 1).

Of the 21 tested unlabeled ssDNAs, only a subset of the oligonucleotides that disrupted the five-thymine run had appreciable impacts on competition for SpoVG (Fig. 4A and B). Competition of the spoVG_RNA with nucleic acid fragments that replace the sequence TTTTT in the unlabeled probe with CCCCC (ss04) or GGGGG (ss05) did not appreciably affect their ability to compete with the wild type spoVG sequence (Fig. 4A). However, replacing the run of thymines in the sequence with adenines in some positions: (AAAAA) (ss03), TTTAA (ss17), or AAATT (SS18), or with Us (ss21) resulted in reduced competition. Lastly, when combining the changes of the TTTTT sequence to AAAAA with deletion of the AGGCT sequence (ss21), the competition was reduced but was similar to that of the AAAAA (ss03) change only.

Fig. 4. Mutagenesis of spoVG ssDNA and RNA.

A. EMSA using 100 nM spoVG_RNA probe and 100 nM of SpoVG in lanes 2–13. ssDNA competitors were added to lanes 3–13 at 5 μM. Free probe is indicated, and the shifted probe is indicated by an asterisk. Competitors used were spoVG + through ss10 (Table 1 and Supp. Table 1). B. EMSA using 100 nM spoVG_RNA probe and 100 nM of SpoVG in lanes 2–15. Competitors at 5 μM were added to lanes 3–15. Competitors were spoVG_CD, ss03, followed by ss11 through ss21 (Table 1). C. EMSA using 100 nM spoVG_RNA probe with 100 nM SpoVG added to lanes 2–6. Unlabeled spoVG_RNA, RNA-03, RNA-17, and RNA-20 were added at 1.5 μM to lanes 3–6 respectively. D. EMSA using 100 nM of spoVG_RNA probe and 100 nM of SpoVG added to lanes 2–8. ssDNA competitors consisted of spoVG_CD, ss22, and ss23 at 500 nM in lanes 3–5 and 2 μM in lanes 6–8.

Based on above results, three unlabeled RNAs were synthesized for use as EMSA competitors: RNA-03 changed the run of five uracils (UUUUU) to adenines (AAAAA), RNA-17 changed the run to UUUAA, and RNA-20 changed it to UUAAA (Table 1). When added to EMSAs at 15-fold excess, each of these mutant competitors was slightly less effective at competing for SpoVG than the spoVG_RNA sequence (Fig. 4C). RNA-03 saw a 29% decrease in competition, whereas RNA-17 and RNA-20 saw a 7% decrease in competition compared to the wild type sequence. All RNAs effectively competed for SpoVG when added at 50x excess whereas the ssDNA counter-parts, ss03 and ss17, were less effective than spoVG_CD (Table 1 and Fig. Supp. 3A). In contrast, 50x excess of unlabeled glpFKD_RNA competitor did not compete nearly as well as any tested spoVG mutant RNA sequence. (Fig. Supp. 3B). Altogether, this suggests that SpoVG differentially interacts with additional aspects of the spoVG_RNA that are not present in the glpFKD_RNA.

Due to the observed differential affinities above between the RNA probes derived from the start of transcription and their cognate dsDNAs, we performed competition assays comparing the coding (spoVG_CD) and non-coding (spoVG_NC) strands of DNA. EMSAs were performed in which the two ssDNAs competed against labeled erpAB_RNA, which was used to inhibit possible base pairing. This competition EMSA indicated that spoVG_NC better competed against erpAB_RNA than spoVG_CD (Supp. Fig. 4A and B). A substantial difference was observed when competitors were added at 100 nM concentration: spoVG_NC was able to compete away 62% of the SpoVG-erpAB_RNA complex compared to spoVG_CD, which only competed away 12%.

3.3. Differences of uracil vs thymine alone can not explain the differences in RNA and ssDNA binding

The differential competition of the SpoVG derived ssDNA and RNA sequences lead us to further investigate whether the presence of uracils in RNA and thymines in ssDNA contributed to differences in SpoVG affinity. To that end, we produced two ssDNAs based on the spoVG_CD sequence: competitor ss22 replaced only the TTTTT with UUUUU, whereas competitor ss23 replaced every thymine in the sequence with uracil (Table 1). Competition analysis using the altered ssDNA sequences did not result in more effective competition than the original SpoVG ssDNA sequence at 5x and 15x competitor concentrations (Fig. 4D).

4. Discussion

B. burgdorferi SpoVG is a sequence-specific RNA- and DNA-binding protein, exhibiting binding constants of 5.9±2.1 × 10^–8 M from direct binding assays and with an apparent IC50 of 1.033 × 10^–6 M ± 2.7 × 10^–7 for spoVG_RNA. Binding affinities for tested dsDNAs appear to be over 100-times weaker than for cognate RNAs, yet selectivity and differential affinity for certain dsDNA sequences was observed.

Analyses with the high-affinity RNA sequence derived from the spoVG start of transcription provided insights on the nature of those interactions. This RNA contains a run of five uracils, UUUUU, which, when changed to adenines, AAAAA, substantially affected SpoVG binding. Two additional base replacements, UUUUU to UUUAA or UUAAA, also led to decreases in SpoVG-RNA complex formation. Note that two other RNAs that SpoVG bound with high affinity also contains runs of uracils: the glpFKD_RNA probe contains two poly-U tracts, UUUUAAU and UUUUAUU, and the erpAB_RNA probe contains one poly-U tract, UUUAU. However, ospAB_RNA contains runs of only two sequential uracils, UU, and the flaB_RNA contains a single uracil tract of UUU. These results suggest that a poly-U tract may play a role in SpoVG binding. Noting that unlabeled glpFKD_RNA was less effective than any of the mutated spoVG RNAs, it is evident that additional features have important impacts on complex formation.

Having observed that spoVG ssDNA and RNA sequence variants showed a wide variation in their abilities to bind SpoVG, we used RNA structure v6.4 to predict the secondary structures of these sequences and the other natural substrates that were observed (Supp. Table 1) [19]. A majority of the constructs that were the least effective competitors e.g. (SS-03 and RNA-03, SS-17 and RNA-17, SS-20 and RNA-20) are predicted to share a secondary structure where the nucleic acid has a stem of 4 or 5 base pairs and a loop with a sequence GGAG. Lacking a 2′ hydroxyl group, the RNA sequences may fold into a more A-form helix compared to the same sequence of DNA, providing a possible rationale for why SpoVG prefers RNA over DNA. This seems particularly likely in that replacing thymines with uracils (ss22 and ss23) was insufficient to increase affinity of the SpoVG protein for these sequences. Yet, sequence ss21, which is predicted to have a different structure, with variations in both the stem and the loop, was still an efficient competitor, suggesting that there are likely other determinants to binding.

The greater affinity of SpoVG for RNA over cognate ssDNA containing uracil implies that SpoVG is more likely to form complexes with RNA rather than DNA either through direct interactions with the backbone or because of RNA’s preferred tertiary folds compared to those of DNA. Previous site-directed mutagenesis of SpoVG revealed that two domains are involved with nucleic acid-binding [1]. Residues in the sole alpha helix of the protein confer nucleic acid sequence preference. Two positively charged residues (argi-nine or lysine) in a separate loop are conserved across all bacterial species and changing either residue to an alanine eliminated nucleic acid binding. Taken together, these results suggest that the charged residues of the loop may bind to the sugar-phosphate backbone and may be better positioned to interact with ribose rather than deoxyribose.

We acknowledge that the sequences used in this study were short fragments of the 5’ end of mRNAs and lack the complexity of full-length mRNAs of coding genes. We are presently examining the B. burgdorferi transcriptome by RNA immunoprecipitation -sequencing (RIP-Seq) to identify additional high-affinity SpoVG binding sites, which can then be compared to the binding sites herein, to provide an enhanced rational approach to characterizing RNA features that are involved with SpoVG-binding.

In conclusion, the results of these studies indicate that there is definitive preferential affinity of SpoVG for certain RNAs and DNAs within the B. burgdorferi genome. Furthermore, SpoVG can bind to numerous sites throughout the transcriptome and genome. B. burgdorferi controls levels of SpoVG throughout the infection cycle [4]. Thus, differential expression of SpoVG will result in occupancy of greater or fewer binding sites, facilitating a variety of different phenotypes.