Differentiation of Rickettsia felis and Rickettsia felis -Like Organisms via Restriction Fragment Length Polymorphism Analysis

Introduction

R ickettsia typhi and Rickettsia felis are obligately intracellular Gram-negative flea-borne pathogens, which cause murine typhus and flea-borne spotted fever, respectively. These rickettsiae are distributed in fleas throughout the world (Azad et al. 1997, Brown and Macaluso 2016). Recently, two other flea-borne rickettsiae have been discovered, Rickettsia asembonensis and Candidatus Rickettsia senegalensis. These two organisms are closely related to R. felis (they have been referred to as Rickettsia felis-like organisms), but they have enough genetic heterogeneity to be proposed as new species (Jiang et al. 2013, Mediannikov et al. 2015). However, these organisms have only been described recently, and little is known of their role in human disease.

As more information is sought to understand the significance of these organisms, such as their distribution within fleas and their ability to interfere with the acquisition and transmission of other pathogenic flea-borne rickettsiae, techniques to efficiently identify these genetically similar rickettsiae are sought. Although probe-based real-time PCR assays have been developed (Maina et al. 2016), they do not lend themselves to use in resource-limited settings. We herein describe an easy and inexpensive method, restriction fragment length polymorphism (RFLP) analysis, for the differentiation of R. felis and R. felis-like organisms.

Methods

Analysis of the nucleotide sequences was performed using the programs within the DNASTAR Lasergene 12 Core Suite (Madison, WI). Genomic sequences for the outer membrane protein B (sca5) of R. felis, R. asembonensis, Candidatus R. senegalensis, and R. typhi were acquired from the National Center for Biotechnology Information database. Sequences were trimmed using DNASTAR Editseq to represent an amplified segment corresponding to sca5 primers 120.2788 and 120.3599 as described elsewhere (Roux and Raoult 2000). These DNA sequences were then analyzed in DNASTAR SeqBuilder, for identification of restriction enzyme digestion sites.

Candidate restriction enzymes were chosen based on their predicted ability to create unique fragment patterns within the selected portion of sca5 for each rickettsial species. The rickettsial sca5 sequences and candidate enzymes were then transferred to DNASTAR GeneQuest, linearized, and run on a simulated agarose gel. The simulated gels produced for each sequence and enzyme pair were compared to determine which banding patterns were easily differentiated. Ultimately, NlaIV (New England BioLabs, Ipswitch, MA) was chosen due to its creation of specific individual patterns that were easily identifiable for each rickettsial species.

To confirm the predicted performance of NIaIV digestion, DNA from sequence confirmed R. felis-, R. asembonensis-, Candidatus R. senegalensis-, and R. typhi-infected Ctenocephalides felis fleas, collected from cats and opossums within Galveston, Texas, were tested (sequences as described elsewhere) (Blanton et al. 2016, Blanton et al. 2019). DNA extracts from these fleas were PCR amplified using the sca5 primer set 120.2788 and 120.3599 (Roux and Raoult 2000). Enzymatic digestion was performed on amplified products of sca5 by incubating 5 μL of PCR product with 3 μL of molecular grade water, 1 μL of NlaIV, and 1 μL of manufacturer-selected enzyme buffer for each sample. Incubation occurred overnight in a 37°C water bath.

The digested products were then electrophoresed on a 4–12% NOVEX TBE polyacrylamide gel (Thermo Fisher Scientific, Waltham, MA) at 200 V for 30 min, stained with ethidium bromide per manufacturer's instructions, and visualized with a UV transilluminator. Restriction profiles were then compared with those previously generated through the agarose gel simulation.

Results

Analysis identified digestion sites for 281 restriction enzymes. Of these, two produced fragment lengths differing for each of the rickettsiae, NlaIV and Hpy81. NlaIV was chosen due to the overall appearance of the unique digestion pattern with fewer bands of larger size when compared with digestion with Hpy81—the latter digested each species' PCR product into three to five bands, some of relatively small size, and with more narrow differences in spacing. The predicted digestion by NlaIV produced the following fragment lengths: R. asembonensis (2 bands of 490 and 323 base pairs [bp]), Candidatus R. senegalensis (2 bands of 683 and 130 bp), R. felis (1 band of 813 bp), and R. typhi (2 bands of 760 and 53 bp) (Fig. 1A).

OPEN IN VIEWER

FIG. 1.

(A) Simulated gel of predicted NlaIV restriction enzyme digestion patterns for the selected portion of sca5 for various flea-borne rickettsiae. Lane 1, molecular weight standards (117, 224, 415, 651, and 702 bp) as generated by DNASTAR GeneQuest software; lane 2, Rickettsia asembonensis; lane 3, Candidatus Rickettsia senegalensis; lane 4, Rickettsia felis; and lane 5, Rickettsia typhi. (B) Polyacrylamide gel of NlaIV digested sca5 amplicons from fleas containing various flea-borne rickettsiae. Lane 1, molecular weight standards (50 bp DNA ladder; New England BioLabs); lane 2, R. asembonensis; lane 3, Candidatus R. senegalensis; lane 4, R. felis; and lane 5, R. typhi.

When amplification and digestion procedures were performed on DNA from sequence confirmed R. felis-, R. asembonensis-, and Candidatus R. senegalensis-infected C. felis fleas, the restriction enzyme digestion patterns were easily discernable and consistent with the predicted digestion patterns (Fig. 1B). When R. typhi-infected fleas were tested, the small 53 bp band was present but faint (Fig. 1B).

Discussion

Flea-borne rickettsioses are important causes of febrile illness throughout the world (Blanton and Walker 2017). R. typhi, the agent of murine typhus, is a well-established pathogen and transmitted by the rat flea (Xenopsylla cheopis) and the cat flea (C. felis) (Azad et al. 1997). R. felis has a worldwide distribution in the cat flea and is increasingly reported as a cause of febrile illness, but its considerable prevalence within cat fleas is not commensurate with the number of cases supported by compelling evidence (Billeter and Metzger 2017, Blanton and Walker 2017). In addition to R. felis, phylogenetically similar organisms, R. asembonensis and Candidatus R. senegalensis, originally isolated from C. felis in Africa, have been increasingly reported in other areas of the world (Jiang et al. 2013, Mediannikov et al. 2015).

Currently, little is known regarding the pathogenicity of these R. felis-like organisms, although R. asembonensis has recently been found within the blood of those with acute undifferentiated febrile illness in Peru (Palacios-Salvatierra et al. 2018). As investigators conduct epidemiological and ecological studies to describe the prevalence of these rickettsiae, an easy and efficient method to identify infecting species within fleas is sought. The use of RFLP to distinguish among these bacteria offers these advantages. Although the use of probe-based real-time PCR assays and sequencing have been used (Maina et al. 2016), they do not lend themselves to use in resource-limited settings.

There are established techniques using other primer pairs and restriction enzymes to distinguish R. typhi from other flea-borne rickettsiae (Schriefer et al. 1994, Zavala-Velazquez et al. 2002), but they are unable to differentiate R. felis from R. felis-like organisms. Owing to a consistently faint 53 bp band (a likely result of limited ethidium bromide-DNA intercalation associated with a small PCR product), we do not consider this assay to be optimal for the differentiation of R. typhi from other flea-borne rickettsiae. Rather, it can be used in conjunction with other RFLP techniques. For example, digestion of a 17-kD antigen gene product with XbaI clearly distinguishes R. typhi from other flea-borne species (Schriefer et al. 1994), whereas the assay described here can further delineate R. felis, R. asembonensis, and Candidatus R. senegalensis.

Conclusions

This RFLP assay was developed to be an efficient and cost-effective method to screen fleas for R. felis and R. felis-like species. We believe this assay can aid in the epidemiological and ecological studies of flea-borne rickettsiae.