The Unusual 23S rRNA Gene of Coxiella burnetii: Two Self-Splicing Group I Introns Flank a 34-Base-Pair Exon, and One Element Lacks the Canonical ΩG

Abstract

We describe the presence and characteristics of two self-splicing group I introns in the sole 23S rRNA gene of Coxiella burnetii. The two group I introns, Cbu.L1917 and Cbu.L1951, are inserted at sites 1917 and 1951 (Escherichia coli numbering), respectively, in the 23S rRNA gene of C. burnetii. Both introns were found to be self-splicing in vivo and in vitro even though the terminal nucleotide of Cbu.L1917 is adenine and not the canonical conserved guanine, termed ΩG, found in Cbu.L1951 and all other group I introns described to date. Predicted secondary structures for both introns were constructed and revealed that Cbu.L1917 and Cbu.L1951 were group IB2 and group IA3 introns, respectively. We analyzed strains belonging to eight genomic groups of C. burnetii to determine sequence variation and the presence or absence of the elements and found both introns to be highly conserved (≥99%) among them. Although phylogenetic analysis did not identify the specific identities of donors, it indicates that the introns were likely acquired independently; Cbu.L1917 was acquired from other bacteria like Thermotoga subterranea and Cbu.L1951 from lower eukaryotes like Acanthamoeba castellanii. We also confirmed the fragmented nature of mature 23S rRNA in C. burnetii due to the presence of an intervening sequence. The presence of three selfish elements in C. burnetii's 23S rRNA gene is very unusual for an obligate intracellular bacterium and suggests a recent shift to its current lifestyle from a previous niche with greater opportunities for lateral gene transfer.

Group I introns are a distinct class of catalytic RNAs (ribozymes) that are considered a legacy of a primordial RNA world (40). They are able to self-splice by means of a two-step transesterification reaction using a guanosine molecule as a cofactor. The catalytic structure that is conserved among all group I introns consists of a specific arrangement of about 10 paired elements (P) that are capped by loops and connected by junctions (J) (38). The core of each intron is comprised of two separately folding helical domains made up of P4-P5-P6 and P3-P7-P9. The P4-P6 domain structurally supports the P3-P9 domain, which contains the active site (17). Splice sites are determined by helix P1, which pairs with the 5′ substrate strand (5′ exon), and by helix P10, which pairs with the 3′ substrate strand (3′ exon). A conserved G · U wobble pair contributes to the recognition of the 5′ splice site, whereas the 3′ splice site is recognized in part by a conserved terminal guanine, termed ΩG (23). Based on conserved secondary-structure characteristics, group I introns are further classified into 13 subgroups (23, 34). These selfish genetic elements are distributed widely in nature, albeit with a bias towards fungi, plants, and red or green algae (14). Although group I introns are abundant in mitochondria and chloroplasts of lower eukaryotes, they are relatively rare in bacteria and until recently were unknown in bacterial structural RNAs. Group I introns in 23S rRNA genes have been reported to occur in hyperthermophilic bacteria of the genus Thermotoga (26) and in Simkania negevensis, a Chlamydiales member (7). A putative group I intron was detected in the sole 23S rRNA gene of Coxiella burnetii (strain Nine Mile phase I RSA493) when the entire genome was sequenced (31); in addition, an intervening sequence (IVS) that produces a fragmented 23S rRNA had been reported earlier by Afseth et al. (1).

C. burnetii, the etiological agent of human Q fever, is a gram-negative obligate intracellular bacterium. This category B select agent is distributed almost worldwide and has a broad range of susceptible hosts, including arthropods, fish, birds, and wild and domestic mammals (6). Human infection is acquired mainly through inhalation of aerosols of animal origin and can cause a wide array of conditions, ranging from asymptomatic infection to an acute self-limiting febrile illness or severe, chronic diseases like endocarditis and hepatitis (21).

The aim of this study was to characterize the splicing properties of the intron reported by Seshadri et al. (31) in the 23S rRNA gene of C. burnetii, since both nonsplicing and self-splicing introns are known to occur in bacterial 23S rRNA (7, 26). During our investigation, we discovered that, in contrast to what was noted in the earlier report, two group I self-splicing introns flank a 34-bp exon and interrupt the 23S rRNA gene of C. burnetii. We also discovered that the terminal nucleotide of one of the introns was adenine and not the canonical ΩG. This is the first time that a group I intron without ΩG has been found in nature. This is significant because the ΩG is considered invariable and thought to play an essential role in RNA splicing (38, 40). This novel intron is also smaller than most group I introns, as it does not encode a homing endonuclease (HE), making it more difficult to identify in whole-genome sequences than relatively larger introns that code for an LAGLIDADG HE. Nevertheless, a nucleotide BLAST search against GenBank revealed a number of group I introns with conserved sequences. Secondary structures for both introns were also inferred by locating conserved sequences. In addition to sequencing the Nine Mile strain, we sequenced the introns of eight other strains of C. burnetii representing seven additional genomic groups (3, 15) and found them to be highly conserved (Table 1), implying a possible role in C. burnetii's biology.

TABLE 1.

C. burnetii 23S rRNA gene group I introns are highly conserved among eight genotypes

Genomic groupa	Strain	Source, yr of isolation	Sequence characteristic(s) ofb:
			Cbu.L1917	Cbu.L1951
I	Nine Mile phase I RSA493	Montana, tick, 1935	Reference sequence	Reference sequence
I	Nine Mile phase II RSA439 (clone 4)	Montana, tick, 1935	Ref	Ref
II	Henzerling RSA331	Italy, human blood, 1945	A247 deleted	Ref
III	Idaho Goat Q195	Idaho, goat, 1981	Ref	G393A,c G402Ac
IV	MSU Goat Q177	Montana, goat, 1980	Ref	Ref
IV	K Q154	Oregon, human heart valve, 1976	Ref	Ref
V	G Q212	Nova Scotia, human heart valve, 1981	Ref	C403Tc
VI	Dugway 7E9-12	Utah, rodents, 1958	Ref	Ref
VII	Q321	Russia, cow's milk	Ref	Ref
VIII	Le Bruges	France, unknown source	Ref	Ref

^aAs defined previously (3,15).

^bRef, same as the reference sequence.

^cThe mutation is in the HE-encoding gene, not the rRNA.

MATERIALS AND METHODS

C. burnetii strains and cultivation.

The primary strain of C. burnetii used in this study was C. burnetii Nine Mile phase II RSA439 (clone 4). Genomic DNA from strains belonging to seven other genomic groups used for determining the sequences of both introns are listed in Table 1. C. burnetii Nine Mile phase II and genomic DNA of the seven other genotypes were generous gifts from Robert A. Heinzen and Paul A. Beare at Rocky Mountain Laboratories, Hamilton, MT. C. burnetii Nine Mile phase II was propagated in African green monkey kidney (Vero) fibroblasts (CCL-81; American Type Culture Collection, Manassas, VA) grown in RPMI medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (HyClone, Logan, UT). Bacteria were purified from infected cells at 7 days postinfection by Renografin (Bracco Diagnostics, Princeton, NJ) gradient centrifugation as previously described (39).

Sequence analyses.

All sequence data were obtained from plasmid DNA or PCR products with an automated DNA sequencer (model ABI3130x1) and a BigDye Terminator cycle sequencing ready reaction kit (ABI, Foster City, CA). Sequence analysis was accomplished using the BLAST 2 Sequences tool at the NCBI website, http://www.ncbi.nlm.nih.gov/BLAST/bl2seq/wblast2.cgi (36).

Intron secondary structure.

Nucleotide BLAST searches (http://www.ncbi.nlm.nih.gov/BLAST/) were done to identify other introns with similar sequences (2). Secondary structures were inferred by manually locating conserved paired helices (P1-P10) using published intron secondary structures (20, 26) as a reference and were drawn using PowerPoint 2003 (Microsoft, Redmond, WA).

RNA and DNA preparations.

RNA was isolated using a Ribopure bacterium kit (Ambion, Austin, TX) and DNA was isolated using a High Pure PCR template preparation kit (Roche Diagnostics, Basel, Switzerland) according to the manufacturers' recommendations.

Reverse transcription-PCR (RT-PCR), cloning, and in vitro transcription.

cDNA was made from RNA using an iScript cDNA synthesis kit (Bio-Rad) per the instructions of the manufacturer. PCR primers (Sigma-Aldrich, St. Louis, MO) for analyzing introns (intron F, GTGGCTGCGACTGTTTAC, and intron R, ATTTCCGACCGTGCTGAG, and exon F, AACGGTCCTAAGGTAGCG, and exon R, TTCGCTACCTTAGGACCG) and IVS (IVS F, CGTGGTGGAAAGGGAAAC and IVS R, TGTCAGCATTCGCACTTC) were designed using Beacon Designer 6 software (Bio-Rad, Hercules, CA). Intron F and R primer sites flank the reported intron sequence (31) 266 bp upstream and 274 bp downstream, respectively, and the IVS F and R primers flank the reported IVS (1) 199 bp upstream and 78 bp downstream, respectively (Fig. 1). The exon F and R primers are specific to the 34-bp exon separating the two introns (see Results) and were used in combination with the intron F and R primers to clone each intron independently into the pCR2.1-TOPO vector as described below. PCR was done in a Mastercycler (Eppendorf AG, Hamburg, Germany). The reaction conditions were 94°C for 5 min, 30 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min, with a single cycle of 72°C for 5 min. PCR products were cloned into pCR2.1-TOPO using a TOPO TA cloning kit (Invitrogen) and selected for proper insertional orientation to utilize the T7 promoter on the plasmid. In vitro transcription was done using a MAXIscript in vitro transcription kit (Ambion), and the resulting RNA was purified using NucAway spin columns (Ambion) per the manufacturer's instructions.

FIG. 1.

C. burnetii 23S rRNA gene linkage map. Two group 1 introns, Cbu.L1917 and Cbu.L1951, flank an essential 34-bp 23S rRNA gene exon. Cbu.L1951 also encodes a putative HE (Cbu_0182). The position of the IVS is also shown, with its nested open reading frame (Cbu_2096). Primer sets used in the study are italicized, and their positions are indicated by small arrows.

In vitro protein synthesis.

The protein product from the gene coding for the HE in Cbu.L1951 was produced in vitro from its endogenous promoter using an Escherichia coli S30 extract system for circular DNA (Promega, Madison, WI) as per the manufacturer's instructions and analyzed using 0.1% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and autoradiography as previously described (24).

Phylogenetic analyses.

Group IB introns (including Cbu.L1917) and group IA3 introns (including Cbu.L1951) were preliminarily aligned by using CLUSTALW (37). Other group IB introns included in the alignment were Thermotoga subterranea (GenBank accession no. AJ556793), Synechococcus sp. strain C9 (DQ421380), Chlorosarcina brevispinosa (L49150), Chaetosphaeridium globosum (AF494279), Chlamydomonas zebra (L43356), Chlamydomonas humicola (L42989), Chlamydomonas monadina (L49149), Chlamydomonas frankii (L43352), Chlamydomonas komma (L43502), Mesostigma viride (AF353999), Oltmannsiellopsis viridis (DQ291132), Suillus luteus (L47586, intron 2), and Chara vulgaris (AY267353, introns 4, 5, and 6); the other group IA3 introns were Monomastix sp. strain M722 (L44124), Chlorella vulgaris (AY008338), Acanthamoeba castellanii (U12380), Chlorosarcina brevispinosa (L49150), Nephroselmis olivacea (AF110138), and Stigeoclonium helveticum (DQ630521). The alignments were next manually refined in BioEdit (11) to explicitly incorporate conserved secondary structures (P1 to P9) and subsequently removed variable regions that could not be reliably aligned, as described in Haugen and Bhattacharya (12). Alignments used for phylogeny reconstruction had 170 nucleotides (group IB) and 237 nucleotides (group IA3). Phylogenies were reconstructed for both data sets by distance, maximum parsimony, and maximum likelihood methods implemented in PAUP* version 4.0b10 (35). Phylogenies for group IB introns were outgroup rooted with the distantly related cox1 intron of Catalpa fargesii (AJ223411 [12]); group IA3 phylogenies were unrooted. The models of sequence evolution used in the distance and likelihood methods for both data sets were selected on the basis of the Akaike information criterion, as implemented in Modeltest (28). For the group IB data set, the model selected was the general time-reversible (GTR) model, with the rate of heterogeneity among nucleotide sites estimated by a discrete (four rate categories) approximation of a gamma (G) distribution (GTR-plus-G model); for the group IA3 data set, the model selected was the general time-reversible model, with the rate of heterogeneity among nucleotide sites estimated by the proportion of invariant (I) sites (GTR-plus-I model). Heuristic searches (for 10 replicate sequences with random sequence additions and with branch swapping by the tree bisection reconnection method) were performed for both the distance and likelihood analyses. Distance analyses were performed according to the least-squares optimality criterion. Unweighted, unordered parsimony reconstructions were inferred by branch and bound, which is guaranteed to find the most parsimonious tree(s). Phylogenies were bootstrap replicated with either 10,000 (distance and parsimony) or 100 (likelihood) replicates. To test whether alternative phylogenetic hypotheses explained the aligned sequence data equally well (i.e., whether their likelihood scores were statistically significantly different), we used the Shimodaira-Hasegawa test of topological congruence (32), as implemented in PAUP*.

Nucleotide sequence accession number.

The sequence data for the two introns designated Cbu.L1917 and Cbu.L1951 were deposited in GenBank and assigned the accession number EF632073.

RESULTS

C. burnetii 23S rRNA gene contains two introns that splice in vivo.

Using the published genome sequence as a reference (31), PCR primers (intron F and R [Fig. 1]) flanking the reported intron insertion site in 23S rRNA gene were synthesized. PCR was done using these primers and genomic DNA from C. burnetii Nine Mile phase II. PCR products were sequenced to confirm that the published 23S rRNA gene sequence of Nine Mile phase I was conserved in the phase II strain (data not shown). Since unspliced and self-splicing group I introns have been reported to occur in 23S rRNA of bacteria (7, 26), splicing characteristics of the intron in the 23S rRNA gene of C. burnetii were analyzed using RT-PCR and PCR (intron F and R primers) on total RNA and genomic DNA, respectively, from 7-day-old cultures (∼2.5 × 10¹⁰ cells). While the resulting PCR product from cDNA was only 551 bp in size, the PCR product from genomic DNA using the same primer pair was 1,559 bp (Fig. 2, lanes 1 and 2), indicating that the intron was spliced out of the mature rRNA. When the PCR product from cDNA was sequenced, it revealed to our surprise that, unlike in the previous report, two introns actually interrupted the 23S rRNA gene of C. burnetii, as shown in Fig. 1. The two introns flank a 34-bp exon corresponding to bases 1918 to 1951 in domain IV of E. coli 23S rRNA (41). Based on their positions, the two introns were designated Cbu.L1917 and Cbu.L1951, respectively, using standard nomenclature (18). Sequencing data also showed that the two introns are spliced in vivo to generate mature rRNA. Further, these data allowed us to determine the exact sequence and the 5′ and 3′ boundaries of both introns. We also analyzed the previously reported (1) IVS in the 23S rRNA gene of C. burnetii using RT-PCR and PCR (IVS F and R primers [Fig. 1]) as described above. While genomic DNA gave the expected 722-bp PCR product, no PCR product was observed with cDNA (Fig. 2, lanes 5 and 6), confirming that the mature 23S rRNA of C. burnetii is fragmented due to the excision of the IVS without subsequent ligation, as previously reported (1).

FIG. 2.

In vivo analysis. PCR was done using intron or IVS primer sets on genomic DNA (gDNA), cDNA, RNA isolated from C. burnetii (Nine Mile phase II), or a no-template control (NTC). An ethidium bromide-stained agarose gel (1% agarose [wt/vol]) is shown. Sizes of the amplicons were determined from standards and are given to the left in base pairs.

Both introns are self-splicing in vitro.

To determine whether the introns could self-splice in vitro, the PCR product generated from the primers intron F and intron R (Fig. 1) and C. burnetii genomic DNA were cloned using a TOPO TA cloning kit (Invitrogen). Utilizing the T7 promoter present in the pCR2.1-TOPO vector, RNA was transcribed and converted into cDNA to create a PCR template for the intron F and R primers. As observed in vivo (Fig. 2), the PCR product from cDNA was 551 bp in size (Fig. 3, lane 2), indicating that the introns could self-splice in vitro in the transcription buffers. Subsequent sequencing of the 551-bp PCR product confirmed that both introns self-splice in vitro without the aid of any proteins (data not shown). Separate constructs containing Cbu.L1917 or Cbu.L1951 were also generated (see Materials and Methods), and the introns were observed to be able to self-splice independently in vitro (data not shown). In contrast to what occurred with introns, when PCR was done using the IVS F and R primers (Fig. 1), both cDNA and genomic DNA gave similarly sized bands (Fig. 3, lanes 5 and 6), showing that the IVS does not self-splice in vitro.

FIG. 3.

In vitro analysis. PCR was done using intron or IVS primer sets on genomic DNA (gDNA), cDNA (from RNA generated by in vitro transcription), in vitro-transcribed RNA, or a no-template control (NTC). An ethidium bromide-stained agarose gel (1% agarose [wt/vol]) is shown. Sizes of the amplicons were determined from standards and are given to the left in base pairs.

Both Cbu.L1917 and Cbu.L1951 are group I introns.

Introns are classified into four major classes based on their splicing mechanisms: group I, group II/III, spliceosomal introns, and tRNA/archaeal introns (14). Hallmarks of group I introns include autocatalytic activity, conserved P helices, the wobble pair G · U at the 5′ splice site, and the conserved ΩG. Based on our sequencing data and previous reports (12, 31), it was easy to determine that intron Cbu.L1951 is a group I intron. Cbu.L1951 has all the typical features of the group I introns described above and also encodes an LAGLIDADG HE (Fig. 1 and 4B), which might play a role in its mobility (12). In fact, when the expression of the cloned HE from its endogenous promoter was checked using a prokaryotic in vitro transcription and translation kit (Promega), an ∼17-kDa product that was not present in vector reactions was observed on autoradiographs (data not shown). The presence of the P2 helix and of helices P7.1 and P7.2 in Cbu.L1951 are characteristics of subgroup IA3 introns (10, 20). The internal guide sequence that pairs with 5′ and 3′ exons and forms the substrate for the ribozyme was also recognizable (helices P1 and P10 in Fig. 4B) in Cbu.L1951. In contrast, it was difficult to classify the intron Cbu.L1917, as it does not code for an HE and does not have the canonical ΩG. A BLAST search using the nucleotide sequence of Cbu.L1917 identified other group I introns in the same 23S rRNA gene position (1917) with conserved sequences. Using the published secondary structure of the intron Tsu.bL1917 as a reference (26), the secondary structure of Cbu.L1917 was predicted. From the secondary structure (Fig. 4A), it is clear that this intron belongs to group I, even though it lacks ΩG. The lack of the P2 helix and an extensive P5 loop indicates that Cbu.L1917 belongs to subgroup IB2 (10). In addition, the internal guide sequence (P1 and P10 in Fig. 4A) is much better defined in Cbu.L1917 than in Cbu.L1951.

FIG. 4.

Predicted secondary structures for Cbu.L1917 (A) and Cbu.L1951 (B). Positions of conserved, paired helices common to group I introns are designated P1 to P10. The 5′- and 3′-terminal bases are encircled. The site of the HE-encoding open reading frame of Cbu.L1951 in P8 is indicated by “HE.” Different line types are used to facilitate easy visualization of intron secondary structures. Exon sequences are shown in lowercase letters, while introns are in uppercase letters.

Phylogenetic analyses.

Phylogeny reconstructions of the group IB intron data were generally similar, and the least-squares distance phylogeny is shown (Fig. 5A). The phylogenetic position of the C. burnetii intron Cbu.L1917 was not definitively resolved. Cbu.L1917 formed a clade with the Thermotoga subterranea intron in the least-squares phylogeny and the three equally most parsimonious trees, but with only modest bootstrap support. In the likelihood phylogeny, Cbu.L1917 was basal to Thermotoga, Synechococcus, and the Chara/Chlamydomonas clade. However, the likelihood score of an alternative tree constrained to the least-squares topology (Fig. 5A) was not significantly different from that of the maximum likelihood tree determined by a Shimodaira-Hasegawa test (−ΔL = 7.70463, where L is the likelihood score; P = 0.193), indicating that both trees explain the data equally well.

FIG. 5.

Phylogenetic analyses of Cbu.L1917 and Cbu.L1951. (A) Least-squares phylogeny for group IB introns, including Cbu.L1917. The tree was rooted with the cox1 intron of Catalpa fargesii. Bootstrap values for the least-squares/maximum likelihood/maximum parsimony analyses are indicated at the nodes, for which the bootstrap value was >50% in at least one analysis. (B) Unrooted least-squares phylogeny for group IA3 introns, including Cbu.L1951. Bootstrap values were as described for panel A. sub, substitution.

Phylogeny reconstructions of the group IA3 intron data were qualitatively similar (least-squares distance phylogeny [Fig. 5B]). L1951 introns (including Cbu.L1951) grouped together to the exclusion of introns with insertion sites at other locations (Nephroselmis intron L2593 and Stigeoclonium intron L730), as previously found (12). Although Cbu.L1951 formed a clade with the Chlorella and Acanthamoeba introns in all trees, its sister taxon was not definitively resolved (Acanthamoeba in the least-squares and parsimony trees and Chlorella in the likelihood tree).

Intron sequences are conserved in other strains of C. burnetii.

Genomic DNA from strains belonging to all eight genomic groups (3) of C. burnetii was analyzed for both introns to determine their presence, absence, or sequence variation. The data indicate that both introns were present in a common ancestor of all eight genomic groups of C. burnetii and are highly conserved among them (≥99% nucleic acid sequence identity, with most variation occurring in the HE gene) (Table 1).

DISCUSSION

In the current study, we identified two self-splicing group I introns in the 23S rRNA gene of C. burnetii, one of which, Cbu.L1917, is the first of its kind. The lack of ΩG in Cbu.L1917 is intriguing, since all other group I introns known to date have guanine as the 3′-terminal nucleotide (4). The conserved ΩG forms a Hoogsteen base-triple with a G-C pair in the P7 helix, sandwiched by three additional base-triples formed by residues in the P7 helix and the J6/7 region to form the guanosine-binding pocket (33, 40). Even though it is possible that the splicing efficiency of Cbu.L1917 is lower than that of Cbu.L1951, as shown by earlier studies in which ΩG was replaced with adenine, cytosine, or uracil, which greatly reduced the rates of 3′ cleavage and exon ligation in vitro (22, 29), we could never detect intermediate forms containing spliced Cbu.L1951 and unspliced Cbu.L1917 in C. burnetii RNA preparations. Moreover, both in vitro and in E. coli expressing the introns, we could detect intermediate forms with either intron spliced out in comparable amounts when internal primer sets were used, indicating that both introns splice out independently and with similar efficiencies (data not shown). It would be interesting to learn the splicing mechanism used by Cbu.L1917 with adenine as its terminal base. Possibly the ΩA uses the same guanosine-binding pocket used by other group I introns to facilitate 3′ cleavage, or it might use a novel mechanism to select the 3′ splice site. It is also very likely that accessory proteins are involved in intron splicing in vivo in spite of the introns' ability to self-splice in vitro, as previously described (19).

It is highly unusual that C. burnetii's lone 23S rRNA gene contains two group I introns, especially considering the rarity of even single elements in bacterial rRNA genes (26). In addition, the large ribosomal-subunit RNA gene of C. burnetii carries an IVS near its 5′ end (Fig. 1). Unlike with introns, the IVS is excised from the RNA without exon ligation, a cleavage process mediated by RNase III (8), resulting in a fragmented but functional rRNA. Moreover, unlike with other obligate intracellular bacteria, which have few or no insertion sequences—presumably due to limited opportunities for lateral gene transfer—C. burnetii's genome possesses a large number of insertion sequence elements (29 full-length and 3 degenerate). This unusually high number of selfish DNAs in the genome of C. burnetii, especially the acquisition of three “selfish genes” in its vital 23S rRNA gene, seems antithetical to the observation that obligate intracellular bacteria undergo reductive evolution. However, this is likely a relic from a past niche that had greater opportunities for lateral gene transfer, with the pathogen recently shifting to its present obligate intracellular lifestyle, a hypothesis supported by the higher percentage of coding genes (89.1%) in its genome than the percentages in other intracellular bacteria, like Rickettsia prowazekii and Mycobacterium leprae (both 76%). It is likely that given sufficient time, C. burnetii's genome will undergo further reductive evolution as suggested by Seshadri et al. (31).

Group I introns are genetic elements that can spread from an intron-positive strain to an intron-minus strain by two known mechanisms. The best-studied and well-understood mechanism is a process termed homing, where the intron moves to an intronless allele of the same gene using its encoded HE (5). The alternative to the HE-dependent mechanism is reverse splicing, where the intron recognizes and integrates into homologous or ectopic RNA sites coupled with RT and recombination (30). The Goddard-Burt model of intron evolution suggests that host populations go through cyclical intron-containing and intronless states, with horizontal transmission being necessary for the long-term persistence of introns in any population (9). According to this model, a group I intron with a full-length HE invades an intronless population by homing. The HE is lost once the intron becomes fixed in a population. Assuming no host benefit, the intron is also subsequently lost from the population. The intron reappears only when it regains access to the population via lateral gene transfer. This model seems appropriate for C. burnetii. Even though available sequence data were not sufficient to specifically identify the donors of both introns, it is apparent that the introns have invaded the 23S rRNA gene of C. burnetii from organelles of lower eukaryotes or other bacteria (Fig. 5). With the introns fixed in the population (Table 1), they appear to be in the process of losing HEs, as suggested by the lack of an HE in Cbu.L1917. However, in contrast to earlier studies that found introns in various stages of degeneration in multiple populations of the same organism (13, 25), a screening of eight genomic groups of C. burnetii revealed that both introns are highly conserved (>99%). This could be due either to insufficient evolutionary time for mutations to accumulate or to selective pressure acting on the introns to maintain them. If the latter is the case, it would be highly informative to understand what function(s) the introns serve, as they are generally considered to be selfish genes that make no contribution towards the reproductive success of the host. Preliminary work in our laboratory suggests a role for the introns in modulating the growth rate of C. burnetii, akin to the intron-mediated ribosomal inhibition observed when Tetrahymena group I intron was expressed in E. coli (27). Currently, we are investigating this growth modulation to understand its potential role in C. burnetii's life cycle. Another exciting avenue for further investigation is the potential for using self-splicing introns as antimicrobial targets to treat Q fever. For example, pentamidine, a drug used for the treatment and prophylaxis of Pneumocystis carinii (Pneumocystis jiroveci) pneumonia (16), is thought to act by inhibiting rRNA group I intron self-splicing (42). Although pentamidine has toxic side effects, it might be a potential therapeutic agent for treating patients with chronic Q fever endocarditis.