RNase III Processing of rRNA in the Lyme Disease Spirochete Borrelia burgdorferi

ABSTRACT

The rRNA genes of Borrelia (Borreliella) burgdorferi are unusually organized; the spirochete has a single 16S rRNA gene that is more than 3 kb from a tandem pair of 23S-5S rRNA operons. We generated an rnc null mutant in B. burgdorferi that exhibits a pleiotropic phenotype, including decreased growth rate and increased cell length. Here, we demonstrate that endoribonuclease III (RNase III) is, as expected, involved in processing the 23S rRNA in B. burgdorferi. The 5′ and 3′ ends of the three rRNAs were determined in the wild type and rnc_Bb mutants; the results suggest that RNase III in B. burgdorferi is required for the full maturation of the 23S rRNA but not for the 5S rRNA nor, curiously, for the 16S rRNA.

IMPORTANCE Lyme disease, the most common tick-borne zoonosis in the Northern Hemisphere, is caused by the bacterium Borrelia (Borreliella) burgdorferi, a member of the deeply branching spirochete phylum. B. burgdorferi carries a limited suite of ribonucleases, enzymes that cleave RNA during processing and degradation. Several ribonucleases, including RNase III, are involved in the production of ribosomes, which catalyze translation and are a major target of antibiotics. This is the first study to dissect the role of an RNase in any spirochete. We demonstrate that an RNase III mutant is viable but has altered processing of rRNA.

INTRODUCTION

Borrelia (Borreliella) burgdorferi is a spirochete that causes Lyme disease (1–3). The bacterium possesses a unique, segmented genome composed of a 910-kb linear chromosome with hairpin ends, as well as a couple dozen or so linear and circular plasmids (4–6). The chromosome harbors most of the housekeeping genes, while the plasmids, which have more genetic variation, predominantly carry genes encoding factors that mediate interactions with the vertebrate host and tick vector (5–11). B. burgdorferi is maintained in nature in an enzootic cycle, where it passes between a tick and a vertebrate (8, 10, 12). The B. burgdorferi genome has lost genes over its evolutionary history as the spirochete has become restricted to an obligate parasitic lifestyle (7, 9, 13–16). Likely as a consequence of this genome reduction, the rRNA genes have been repositioned into an idiosyncratic arrangement (5, 17–22).

In most bacteria, the three rRNAs (16S, 23S, and 5S) are produced as a single polycistronic transcript. Many microorganisms carry multiple copies of the rRNA operon (seven in Escherichia coli and 10 in Bacillus subtilis). Each operon almost always contains the three rRNA genes in the canonical 16S-23S-5S order, with minimal intergenic spacing. Various tRNA genes may be inserted between the rRNA genes. The polycistronic transcript is cotranscriptionally processed by endoribonucleases to separate each individual rRNA (23–27). Bacterial species that are host or niche restricted, with a reduced genome size and low GC content, generally possess fewer rRNA copies. Presumably, some operons are lost through recombination events during genome reduction; this can result in unusual rRNA gene organization, regulation, and chromosomal distribution (13, 28). B. burgdorferi and closely related Lyme disease species possess only a single copy of the 16S rRNA gene, and it is separated by more than 3 kb from the tandemly duplicated 23S-5S rRNA operons (5, 17–22, 29). The relapsing fever Borrelia species also have a split ribosomal operon, with additional genes inserted, but have only one copy of each rRNA gene (30). Although split ribosomal operons and uneven numbers of rRNA genes have been described (13, 28, 31–33), to our knowledge, the specific rRNA gene organization of B. burgdorferi has not been observed in any other genome.

Bacteria typically encode a suite of ribonucleases (RNases) that perform a variety of functions, including ribosome processing. RNase III, an endoribonuclease encoded by the rnc gene, acts as a homodimer that specifically recognizes and cleaves double-stranded RNA (34–36). The enzyme catalyzes the initial processing step of the 16S and 23S rRNAs, separating them from the polycistronic primary transcript prior to ribosome assembly (37–39). During transcription, stems form between the complementary 5′ and 3′ flanking regions of each rRNA and serve as double-stranded substrates for RNase III (40). RNase III cleaves both strands simultaneously, releasing the individual rRNAs. Exonucleases, and a few endonucleases, further process the remaining stem to generate the mature rRNAs (27, 35, 36). RNase III also processes mRNA to regulate transcript levels (35, 41–44). B. burgdorferi encodes an RNase III homolog with 44% identity to the B. subtilis enzyme and 35% identity to the E. coli enzyme. The 5S rRNA is processed by either RNase E in E. coli or RNase M5 in B. subtilis, following its liberation from the 30S transcript by RNase III cleavage (45, 46); B. burgdorferi possesses a homolog of RNase M5 (BB0626).

RNase III is not essential in many bacteria, including E. coli, because alternative RNA processing pathways can be utilized, although these rnc mutants have a growth phenotype (25, 47–50). The phenotype is likely due to aberrant ribosomes; the precursor 30S and pre-23S rRNA species accumulate in rnc null mutants (37, 39, 47, 49–52). mRNA and noncoding RNA (ncRNA) processing defects might also contribute to the phenotype, due to the role of RNase III in regulating levels of these RNA species (35, 41, 42, 44). However, RNase III is essential in B. subtilis (39) because it cleaves several prophage-encoded double-stranded toxin-antitoxin duplexes, preventing these mRNAs from being translated (53).

RNase III has not been previously investigated in spirochetes. We now demonstrate that RNase III is not essential in B. burgdorferi, but an rnc_Bb mutant has a growth and morphological phenotype. The 23S rRNA is processed, as expected, by RNase III. However, RNase III is not required to generate wild-type 16S rRNA ends, which may be related to the distinct rRNA arrangement in Lyme disease Borrelia species.

RESULTS

Generation of an RNase III (rnc_Bb) null mutant.

We hypothesized that the rRNA transcripts in B. burgdorferi, despite their unique organization, are processed as in other bacteria, including cleavage of the 16S and 23S rRNA by the endoribonuclease RNase III (37, 38, 39). The chromosomal rnc_Bb gene was replaced through homologous recombination with the gentamicin resistance cassette flgBp-aacC1 (Fig. 1A). Three transformants were isolated in two B. burgdorferi strains (297 and B31-A3). The clones were examined for the presence of the flgBp-aacC1 insertion (Fig. 1B) by PCR, using primers specific for the flanking regions of the rnc_Bb gene (Table 1; Fig. 1A). A second primer set specific for the flgBp-aacC1 cassette and flanking downstream rnc_Bb region was also used to verify insertion of the resistance marker into the chromosome (data not shown). In addition, the absence of rnc_Bb transcript in mutants was confirmed by reverse transcription-PCR (RT-PCR) (data not shown). Extensive attempts to complement the rnc_Bb null mutants, either in trans with rnc_Bb carried on a shuttle vector or in cis by genetic reconstitution, were unsuccessful; we speculate that our inability to transform the mutant was a consequence of an altered response to the competence preparation.

FIG 1.

A B. burgdorferi RNase III (rnc_Bb) null mutant exhibits growth and morphological phenotypes. (A) Schematic of the genetic approach taken to construct the rnc_Bb null mutant in B. burgdorferi. A gentamicin resistance cassette (flgBp-aacC1) replaced most of the chromosomal rnc_Bb gene via allelic exchange by homologous recombination. Primers rnc 1F+NdeI and rnc 738R (Table 1), which were used to obtain PCR products shown in panel B, are indicated by the small black arrows above the rnc_Bb gene and flgBp-aacC1 cassette. (B) Confirmation of an rnc_Bb null mutant (in a B31 background). DNA from a B. burgdorferi transformant and controls were amplified by PCR, using primers flanking the insertion site, and the products were electrophoresed in a 1% agarose gel. Lane 1, cloned recombination substrate; lane 2, genomic DNA from wild-type B. burgdorferi; lane 3, genomic DNA from B. burgdorferi rnc null mutant. (C) B. burgdorferi wild type and the rnc_Bb null mutant were inoculated in BSK II liquid medium at a cell density of 10⁴ cell ml⁻¹ and grown at 34°C until stationary phase. Cells were enumerated every 24 h, using a Petroff-Hausser counting chamber. (D) Microscopy images showing the cell morphology of the rnc_Bb null mutant (lower panel) compared with that of wild-type B. burgdorferi (upper panel). Live B. burgdorferi cells were stained with a wheat germ agglutinin (WGA)-Alexa Fluor 594 conjugate and assayed by fluorescence microscopy. The length of 50 cells for each strain was measured as described by Lybecker et al. (54) and sorted into bins of 5-μm increments.

TABLE 1.

Oligonucleotides used in this study

Purpose and name	Sequence (5′→3′)
rnc null mutant construction
bb705 U1260F	TTTAAAGGTTGAAAATGAAG
bb705 92R+AatII+AgeI	ACCGGTCAAGACGTCTAAAGTCAATGCTCAAATT
bb705 617F+AatII	GACGTCTTTTTTGTGTGGAACTTTAT
bb705 D1945R+AgeI	ACCGGTATGAATCTAGGGAAAAACA
rnc null mutant screening
rnc 1F+NdeI	CATATGAAAAAAAAATCTTCTGA
rnc 738R	TTAAAGGTTAATGTTTTCCA
pflgB5′+AgeI	ACCGGTACCCGAGCTTCAAGGAA
5′ and 3′ RACE
rrs 136R	CCCATCTCATAGGTAGATCATCCACGC
rrl 76R	GCTTTTCGCAGCTTACCACGACCTTC
rrl 198R	TTAGATGGTTCACTTCCCCTGGTATCGC
rrf 88R	CGAACTCGCAGTACCATCAGCGAATAAG
rrf 112R	CCCTGGCAATAACCTACTCTCCCGC
rrs 1365F	TGAATACGTTCTCGGGCCTTGTACACA
rrl 2770F	ACGTTCGGAAAGGATAACCGCTGAAAG
rrf 60F	CTTATTCGCTGATGGTACTGCGAGTTCG
rrf 49R	TGTGTTCGGAATGATAACAGGTGTTTCCTC
rrf 51R	TCTGTGTTCGGAATGATAACAGGTGTTTCC
rrf 8F	GGTTAAAGAAAAGAGGAAAC
rrl 76R	GCTTTTCGCAGCTTACCACGACCTTC
Northern blot probe: rnc 158R	CTAGATTTTTGATCCAACTCATTAGAATACGACGAATGACACAATG
Recombinant RNase III
rnc 1F+NdeI	CATATGAAAAAAAAATCTTCTGA
rnc 735R+SapI	GCTCTTCCGCAAAGGTTAATGTTTTCCATAG
Artificial 23S rRNA substrate
T7 prom	TAATACGACTCACTATAG
23S stem+T7	AGGAAGACAAAAATATGGCCAAAGTTGCCTTTGACCATATTTTTATCTTCCATCCTATAGTGAGTCGTATTA

Characterization of the rnc_Bb null mutant phenotype.

The rnc_Bb null mutants from both parental strains exhibited growth (Fig. 1C) and morphological (Fig. 1D) phenotypes. Wild-type B. burgdorferi cells reached mid-log-phase cell density within 1 day and stationary phase by 4 days. The rnc_Bb mutant did not reach mid-log-phase cell density until about 5 days and the log phase was extended until about 11 days. Cell density reached by the rnc_Bb null mutant at stationary phase was approximately a half log lower than that of the wild type. In addition to the growth phenotype, the rnc_Bb null mutant cells are elongated compared to wild-type cells (Fig. 1D), which is a phenotype previously found in B. burgdorferi hfq mutants (54). A subpopulation of mutant cells has a wild-type length, resulting in a bimodal distribution (Fig. 1D).

Analyses of rRNA in the rnc_Bb null mutant.

The ends of the three rRNAs in the rnc_Bb mutant were determined by 5′ and 3′ rapid amplification of cDNA ends (RACE) to assay the function of RNase III in rRNA processing in B. burgdorferi. Total RNA was isolated from both B. burgdorferi rnc null mutants, as well as from the wild type. The 5′ and 3′ RACE PCR products for each rRNA were cloned and sequenced. The primary nucleotide sequence of the rRNA is shown with black dots above (wild-type sequences) and below (rnc_Bb mutant sequences), indicating the experimentally determined location of the rRNA ends (Fig. 2). The ends of the 16S rRNA are the same in both the wild type and rnc_Bb mutant, and map to the annotated ends, although the 3′ end is more variable (Fig. 2B). Therefore, RNase III is not essential for 16S rRNA processing in B. burgdorferi. However, the RACE data show that RNase III is required for full processing of the 23S rRNA (Fig. 2C). The mature 23S rRNA transcript is longer at both ends in an rnc_Bb null mutant. The 5′ end of the 23S rRNA in the rnc_Bb mutant is 34 nucleotides downstream from the predicted promoter. There is no difference in the 5S rRNA ends generated in the rnc_Bb null mutant compared to those of the wild type, demonstrating that RNase III is not required for processing 5S rRNA (Fig. 2A), as expected (25–27, 45, 46).

FIG 2.

B. burgdorferi RNase III processes the 23S rRNA but not the 5S rRNA or the 16S rRNA. The 5′ and 3′ ends of the three rRNAs were examined by 5′ and 3′ RACE, respectively. Primary sequences for each rRNA 5′ end and 3′ end are shown, with the annotated rRNA sequence underlined. Individual sequencing events from multiple RACE PCR products for each rRNA are represented by black dots above the sequences for wild type (WT) and below the sequences for the rnc_Bb mutant in a B31 background (rnc null). (A) 5S rRNA. (B) 16S rRNA. (C) 23S rRNA.

rnc_Bb operon structure.

In B. burgdorferi, rnc is carried on the chromosome, in a region dense with genes related to translation. Northern blotting with an oligonucleotide probe to rnc_Bb detected two polycistronic transcripts (Fig. 3A). The shorter mRNA, which may be either a primary or a processed transcript, begins with rpmF (encoding ribosomal protein L32), and the longer mRNA begins in trmD and includes rplS (encoding ribosomal protein L19). Examination of transcriptome data from wild-type B. burgdorferi (Fig. 3B) confirms the Northern blotting data. Junctional RT-PCR detected transcription throughout a larger 16-gene region (data not shown), beginning with bb0690 (bicA, dps, or napA), which may be due to either a larger polycistronic mRNA or pervasive transcription (55).

FIG 3.

rnc_Bb is in an operon. (A) Northern blot analysis of 15 μg of total RNA fractionated on a formaldehyde-agarose gel, blotted to a nylon membrane, and hybridized with an oligonucleotide probe (gray bar in panel B). (B) Deep-sequencing results displayed as a coverage map. The chromosomal location and genomic context of the region around rnc_Bb (bb0705) are depicted above and below the coverage map, respectively. The wavy arrows represent the predicted polycistronic mRNAs, and the gray bar illustrates the location of the probe used in the Northern blot.

Cleavage of an artificial 23S rRNA substrate.

We biochemically assayed the activity of RNase III from B. burgdorferi using an in vitro cleavage reaction, with an artificial 23S rRNA stem-loop that serves as an RNase III substrate (56). The sequence is composed of the double-stranded stem portion of the B. burgdorferi 23S rRNA transcript with a loop of four unmatched nucleotides (Fig. 4A). RNase III from E. coli recognizes double-stranded stems and creates a staggered break with a 3′ two-nucleotide overhang (35, 36), although cleavage sites predicted from our RACE data (Fig. 2) result in a 5′ three-nucleotide overhang (Fig. 4A). The artificial 23S rRNA substrate was generated by in vitro transcription and radiolabeled with [³²P]UTP. A molar excess of the substrate was incubated with recombinant B. burgdorferi RNase III protein as previously described (56, 57). Cleavage products were produced with increasing concentrations of RNase III, indicating that the stem-loop serves as a substrate (Fig. 4B).

FIG 4.

RNase III cleaves an artificial 23S rRNA substrate in vitro. (A) The double-stranded secondary structure of the artificial B. burgdorferi 23S rRNA substrate, predicted using Mfold (84). Substrate design was as described by Amarasinghe et al. (56). Arrows indicate cleavage sites predicted from RACE data (Fig. 2). (B) Phosphorimage showing the increase in [³²P]UTP-labeled artificial 23S substrate cleavage products produced at 37°C after 5 min in the presence of increasing concentrations of recombinant RNase III (0 nM to 25 nM). The predicted species designations (right side) for each cleavage product are based on the work of Meng et al. (57) but have not been experimentally confirmed.

DISCUSSION

RNase III is a highly conserved endoribonuclease (35, 36, 42, 43, 58). This study is the first to characterize an RNase III homolog from a spirochete. Our results reveal that B. burgdorferi RNase III processes 23S rRNA, transcribed from the unusual tandem 23S-5S rRNA operons, as expected. However, the enzyme is not required to generate the mature 16S rRNA (Fig. 2), which is transcribed from a single gene separated from the 23S-5S rRNA operons. The experimentally determined 16S rRNA ends are not in the predicted double-stranded stem (Fig. 5), which may explain why these ends are not affected by the absence of RNase III.

FIG 5.

A model for rRNA processing in B. burgdorferi. Structures of the rRNA double-stranded stem regions are shown; the 16S rRNA structure (left side) was predicted using MacVector and Mfold (84), and the 23S and 5S rRNA structures (right side) are based on the work of Schwartz et al. (20). Experimentally determined 5′ and 3′ ends for the rRNAs from the RACE data (Fig. 2) are represented by black arrows (wild type) and open arrows (rnc_Bb null mutant); 5′ and 3′ ends for 16S and 5S rRNAs are the same in the wild type and the rnc_Bb mutant.

B. burgdorferi RNase III is more similar to that of the B. subtilis homolog than that of the E. coli homolog, and the palette of endoribonucleases in B. burgdorferi is more similar to the B. subtilis set. B. burgdorferi possesses a homolog of the 5S rRNA-processing enzyme RNase M5 from B. subtilis instead of RNase E from E. coli, in addition to an unannotated homolog of RNase Y (BB0504), which is also found in B. subtilis but not E. coli (M. L. Anacker, unpublished data) (59).

RNase III is essential in B. subtilis (39, 53), and rnc mutants of E. coli as well as those of S. aureus grow at a reduced rate, utilizing alternative RNA processing pathways (60, 61). The B. burgdorferi rnc mutant grows more slowly and reaches a lower cell density at stationary phase than the wild type (Fig. 1C). The growth phenotype may be due to defective ribosomes that are a consequence of incomplete 23S rRNA processing, or it may be due to defective processing of either a regulatory (44) or a housekeeping RNA. Abnormal 30S precursor ribosomal subunits containing partially processed 16S-23S-5S rRNAs are observed in rnc mutants of E. coli and B. subtilis (37, 39, 47, 50–52), and aberrantly long 23S rRNA can interfere with ribosome assembly (62). We were unable to complement the B. burgdorferi rnc mutants despite nearly exhaustive efforts, which may be another consequence of the growth defect; however, this is not uncommon, as complementation of mutants of other nucleic acid-interacting proteins has foundered in the spirochete (63–65).

RNase III also processes mRNA in order to globally regulate transcript levels (35, 41, 42, 66–68). An additional role for RNase III in S. aureus was found in tRNA and ncRNA, as well as mRNA regulation (68), and RNase III in Corynebacterium glutamicum regulates cell division via degradation of mraZ mRNA (44). Loss of mRNA and ncRNA regulation may also contribute to the growth and morphological phenotype of the B. burgdorferi rnc mutant. We hypothesize that the slow-growth and long-cell phenotype is due to delayed or improper cell division, caused by either inefficient translation or production of defective cell division machinery. Intriguingly, the rnc_Bb mutant population also includes cells with a wild-type length, yielding a bimodal distribution (Fig. 1D).

RNase III has a fundamental role in processing of the 16S and 23S rRNAs during ribosome genesis, so we hypothesized that this would be conserved in B. burgdorferi. We assayed processing by determining the 5′ and 3′ ends of the mature rRNAs in the B. burgdorferi rnc null mutant. In other bacteria, endonucleolytic processing occurs cotranscriptionally, first releasing the 16S rRNA and then separating the 23S and 5S rRNAs from the polycistronic transcript (23).

Comparison of wild-type B. burgdorferi and the rnc_Bb null mutant by 5′ and 3′ RACE revealed that the 23S and 5S rRNA transcripts exist in their canonical mature form, in spite of the tandem gene duplication (Fig. 2 and 5). The 23S rRNA is 15 nucleotides longer at the 5′ end and about 18 nucleotides longer at the 3′ end in the rnc_Bb null mutant, which supports a role for RNase III in 23S rRNA processing. The 23S rRNA 5′ end in the rnc_Bb null mutant maps to 34 nucleotides downstream of the predicted promoter, implying that this end is generated by minimal nucleolytic processing by an unknown RNase (Fig. 5). The 23S rRNA 3′ end in the rnc_Bb null mutant maps to the 5′ end of the 5S rRNA unprocessed stem (Fig. 5), suggesting that an RNase, possibly RNase Y, is responsible for processing this end, following RNase M5 cleavage of the 5S rRNA. RT-PCR data showed an increase in the amount of RNA across the junction between the tandem 5S-23S rRNA operons in the rnc_Bb null mutant (M. L. Anacker, A. F. Brinkworth, and D. S. Samuels, unpublished data), implying the presence of a large precursor transcript and a decrease in endonucleolytic processing, which is not without precedent, as a 30S unprocessed rRNA species is observed in both B. subtilis rncS and E. coli rnc null mutants (37, 39, 47, 50–52). On the other hand, there is no difference in the 5′ and 3′ ends of the 5S rRNA transcript between the wild type and rnc_Bb null mutant, which is consistent with processing of 5S rRNA carried out by RNase E in E. coli and RNase M5 in B. subtilis (45, 46).

The 16S rRNA gene is spatially separated from the tandem 23S-5S rRNA operons on the chromosome in B. burgdorferi (5, 17–22). A truncated open reading frame (ORF) (bb0425) is present upstream, and a tRNA^Ala is downstream of the marooned 16S rRNA gene. RT-PCR data suggest that the 16S rRNA is cotranscribed with the two upstream genes (Anacker et al., unpublished), and Bugrysheva et al. (69) showed that the tRNA^Ala was cotranscribed with the 16S rRNA gene, yielding a large polycistronic transcript. Data from RACE analyses (Fig. 2) indicate that the 16S rRNA ends are the same in the rnc_Bb null mutant as in the wild type, demonstrating that RNase III is not required for full processing of this rRNA. The experimentally determined ends of the 16S rRNA are not in the predicted double-stranded stem structure (Fig. 5) that potentially could be a cleavage site for RNase III. We propose the following model for generating the observed 16S rRNA 3′ end in the rnc_Bb null mutant. RNase P processing of the tRNA^Ala 5′ end downstream of the 16S transcript would release a region of single-stranded RNA that undergoes subsequent processing by PNPase, a single-stranded exonuclease, and YbeY, a single-stranded endonuclease, as observed in E. coli (47, 70, 71). The mechanism of 16S rRNA 5′ end maturation is unknown in B. burgdorferi. Our data demonstrate that RNase III in B. burgdorferi processes the 23S rRNA, as expected; other endonucleases, possibly RNase Y and RNase M5, process the 5S rRNA; and a unique and currently undefined mechanism processes the 16S rRNA.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

Low-passage-number B. burgdorferi strains B31-A3 (72) and 297 (BbAH130) (73) and the rnc_Bb mutants, were cultivated at 34°C in Barbour-Stoenner-Kelly II (BSK II) liquid medium containing 6% rabbit serum (74) without gelatin (75). Cell density was assayed as previously described, by enumeration using a Petroff-Hausser counting chamber (76). E. coli Rosetta (DE3) (Novagen), used to overexpress recombinant B. burgdorferi RNase III, was grown at 37°C in LB medium containing 50 μg ml⁻¹ ampicillin and 68 μg ml⁻¹ chloramphenicol. E. coli TOP10F′ and DH5α cells, used for cloning, were grown at 37°C in LB medium containing 50 μg ml⁻¹ kanamycin.

Generation of an rnc_Bb null mutant in B. burgdorferi.

The rnc_Bb gene (bb0705) was disrupted by replacement with flgBp-aacC1, which confers gentamicin resistance (72), as previously described (77, 78). Genomic regions encompassing 1.3 kb upstream and 1.3 kb downstream of rnc_Bb were amplified by PCR using Taq polymerase (Sigma-Aldrich), cloned into pCR2.1-TOPO, confirmed by DNA sequencing, and ligated together. The gentamicin resistance cassette was ligated into a synthetic AatII site between the two flanking sequences. The plasmid was linearized with ScaI and electroporated into B. burgdorferi, as previously described (75, 79). Transformants were cloned in liquid BSK II medium in 96-well plates (80) containing 40 μg ml⁻¹ gentamicin at 34°C in a 1.5% CO₂ atmosphere and screened by PCR. At 34°C, mutant clones appeared in approximately 90 days, which is considerably longer than the 10 to 14 days typically required. Positive clones were confirmed by reverse transcription-PCR (RT-PCR). Total RNA was isolated using TRIzol (Ambion), and rnc_Bb transcript was measured using a RETROscript kit (Ambion), as previously described (54, 77). The presence of endogenous plasmids was confirmed by PCR (81, 82). Two independent rnc_Bb null mutants from B. burgdorferi strains B31-A3 and 297 were isolated.

Fluorescence microscopy.

Live B. burgdorferi cells were stained using a wheat germ agglutinin (WGA)-Alexa Fluor 594 conjugate (Invitrogen) (78). Briefly, a small volume of cells was washed once in prewarmed Dulbecco's phosphate-buffered saline (dPBS; 138 mM NaCl, 2.7 mM KCl, 8.1 mM Na₂HPO₄, and 1.5 mM KH₂PO₄) plus 5 mM MgCl₂ and resuspended in dPBS containing WGA-Alexa Fluor 594 at 1:200 dilution. The cells were incubated for 5 min at 37°C before pelleting by centrifugation. Cells were resuspended in a small volume of prewarmed dPBS and placed on precleaned slides with coverslips. Images were captured using an Olympus BX51 microscope through a 100× objective equipped with a DP72 digital camera, controlled by DP2-BSW software. Fluorescence images were processed and the length of 50 spirochetes for each strain was measured using ImageJ (National Institutes of Health; http://rsbweb.nih.gov/ij/) (54).

Identification of 5′ and 3′ ends of rRNAs.

The 5′ and 3′ ends of each of the rRNAs (16S, 23S, and 5S) were determined by 5′ and 3′ rapid amplification of cDNA ends (RACE). B. burgdorferi cells were grown to mid-log-phase cell density and lysed with TRIzol; RNA was extracted with chloroform, precipitated with isopropanol, and washed with 70% ethanol, before being resuspended in RNase-free water. Prior to 3′ RACE, total RNA was polyadenylated with a poly(A) tailing kit (Ambion). For the RACE protocol, nested primers and reverse transcriptase from the BD Smart RACE cDNA Amplification kit (Clontech) were utilized to generate cDNA from the total RNA. The cDNA was amplified by PCR, using a universal primer set and gene-specific primers for a sequence either downstream of the 5′ end or upstream of the 3′ end for each rRNA gene (Table 1). PCR products were analyzed on a 2% agarose gel, purified with a QIAquick gel extraction kit (Qiagen), and cloned into pCR2.1-TOPO. Clones were sequenced at the Murdock DNA Sequencing Facility (University of Montana) with an Applied Biosystems genetic analyzer (GeneScan), and sequence data were analyzed using MacVector.

RNA-seq and Northern blotting of rnc_Bb operon.

Directional (strand-specific) RNA-seq cDNA libraries were constructed with a ligation-based protocol and analyzed as described previously (78). Operon predictions were made with DOOR² (http://csbl.bmb.uga.edu/DOOR/) (83). For Northern blotting, total RNA was fractionated on a formaldehyde-agarose gel, blotted to a nylon membrane, and hybridized with oligonucleotide rnc 158R (Table 1), as previously described (54).

RNase III cleavage assay.

B. burgdorferi RNase III protein was purified using the IMPACT (intein-mediated purification with an affinity chitin-binding tag) kit (New England BioLabs). Briefly, cultures of E. coli carrying pTXB1 with the B. burgdorferi rnc gene were induced with 0.4 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 4 h at 37°C and lysed by sonication. The lysate was centrifuged and loaded on chitin beads and washed with column buffer (20 mM Tris-HCl [pH 8.5], 500 mM NaCl, and 1 mM EDTA). Dithiothreitol (DTT) was added to the column to initiate self-cleavage of the intein tag. The unbound recombinant protein was then eluted from the column using column buffer.

An artificial 23S rRNA stem-loop that serves as an RNase III substrate was generated, as previously described (56); the sequence is composed of the 23S rRNA 5′ and 3′ complementary ends that form the stem portion of the rRNA transcript plus an intervening loop of four unmatched nucleotides. Oligonucleotides incorporating a T7 promoter (Table 1) were used to generate a template for transcribing a ³²P-labeled RNA molecule. Oligonucleotides were annealed in water by heating for 5 min at 65°C and then quick cooling on ice. A MEGAscript (Ambion) in vitro transcription reaction with [α-³²P]UTP (Perkin-Elmer) was incubated overnight at 37°C. Following subsequent DNase treatment (Turbo DNase), the substrate was resolved in a 15% Tris-borate-EDTA (TBE)–urea–polyacrylamide gel, extracted overnight using a gel elution buffer (0.5 M ammonium acetate, 1 mM EDTA, and 0.2% sodium dodecyl sulfate), and precipitated with ethanol. Detection of the substrate in the gel prior to extraction was performed using a phosphorimager (FLA-3000; Fujifilm). Following precipitation, RNA was resuspended in Tris-EDTA (TE) buffer (10 mM Tris-HCl, pH 8.0, and 1 mM EDTA).

The RNase III cleavage assay was performed as previously described (56, 57). The artificial 23S rRNA substrate was briefly heated at 100°C and rapidly cooled on ice to fold into a stem-loop structure. Substrate RNA (300 nM) and various levels of recombinant RNase III (0 to 25 nM) were combined in cleavage buffer (30 mM Tris-HCl [pH 8.0], 160 mM NaCl, 0.1 mM EDTA, 0.1 mM DTT, and 0.1 mM tRNA) and incubated for 5 min at 37°C to bind enzyme and RNA. Addition of 10 mM MgCl₂ initiated the cleavage reaction. Reactions were stopped after 5 min by adding gel loading buffer containing EDTA. Samples were resolved on a 15% (wt/vol) polyacrylamide gel containing 7 M urea and visualized on a phosphorimager.