Genetic Identification of Rickettsial Isolates from Fatal Cases of Brazilian Spotted Fever and Comparison with Rickettsia rickettsii Isolates from the American Continents

Abstract

Fifteen bacterial isolates from spotted fever group rickettsiosis in Brazil were genetically identified as Rickettsia rickettsii. In a phylogenetic analysis with other R. rickettsii isolates from GenBank, the Central/South American isolates showed low polymorphism and formed a clade distinct from two North American clades, with the North American clades having greater in-branch polymorphism.

TEXT

The tick-borne bacterium Rickettsia rickettsii is the etiologic agent of Rocky Mountain spotted fever (RMSF), the most severe rickettsiosis affecting humans in the Western Hemisphere (1). In Brazil, the disease has been referred to as Brazilian spotted fever (BSF). While RMSF fatality rates are usually 5% to 10% in the United States, general rates of 20% to 40% have been reported in Brazil (1).

In the state of São Paulo in southeastern Brazil, where BSF is a disease for which notification is compulsory, various spotted fever group (SFG) rickettsial isolates have been obtained from human clinical cases during the past few years (2, 3). Although all these isolates were confirmed to be SFG rickettsiae through indirect immunofluorescence assays using anti-R. rickettsii human sera (4), they were not genetically identified through molecular analysis. Herein, we performed genetic identification and molecular characterization of these rickettsial isolates and compared their genetic profiles with those of isolates from ticks in Brazil and with those of R. rickettsii isolates from other American countries.

Blood clot or skin lesion biopsy specimens from BSF-suspected cases were processed by the shell vial technique, as described previously (5). Once rickettsiae were visualized within Vero cells through an immunofluorescence assay using anti-R. rickettsii human polyclonal serum (4, 6), 1st-passage-infected cells were harvested, and one aliquot was used for DNA extraction through the PureLink genomic DNA kit (Invitrogen, Carlsbad, CA). Extracted DNA was assayed according to PCR protocols using primers targeting the rickettsial genes gltA, ompA, and ompB and the RR0155-rpmB, RR1240-tlc5^b, and cspA-ksgA intergenic regions, as shown in Table 1. In addition, DNA extracted from 3rd-passage-infected cells of three R. rickettsii isolates, previously isolated from Amblyomma sculptum (reported as Amblyomma cajennense), Amblyomma aureolatum, and Rhipicephalus sanguineus ticks in Brazil (12–14), was tested according to intergenic region PCR protocols. PCR products were sequenced in an ABI automated sequencer (model ABI 3500 Genetic Analyzer; Applied Biosystems/Thermo Fisher Scientific, Foster City, CA) with the same primers used for PCR. The generated sequences were compared with each other and submitted to BLAST analyses (www.ncbi.nlm.nih.gov/blast) to infer the closest similarities available in GenBank.

TABLE 1.

Primer pairs used for amplification of rickettsial genes (gltA, ompA, and ompB) or intergenic regions (RR0155-rpmB, RR1240-tlc5^b, and cspA-ksgA)

Primer pair	Target	Primer	Primer sequence (5′ to 3′)	Amplicon size (bp)	Reference
1	gltA	CS-239	GCTCTTCTCATCCTATGGCTATTAT	834	7
		CS-1069	CAGGGTCTTCGTGCATTTCTT		7
2	ompA	Rr190.70p	ATGGCGAATATTTCTCCAAAA	530	8
		Rr190.602n	AGTGCAGCATTCGCTCCCCCT		8
3	ompB	120-M59	CCGCAGGGTTGGTAACTGC	862	9
		120–807	CCTTTTAGATTACCGCCTAA		9
4	RR0155-rpmB	Forward	TTTCTAGCAGCGGTTGTTTTATCC	290	10
		Reverse	TTAGCCCATGTTGACAGGTTTACT		10
5	RR1240-tlc5b	Forward	CGGGATAACGCCGAGTAATA	357	11
		Reverse	ATGCCGCTCTGAATTTGTTT		11
6	cspA-ksgA	Forward	CATCACTGCTTCGCTTATTTT	405	10
		Reverse	ATTTCTTTTCTTCCTCTTCATCAA		10

Phylogenetic analyses were performed using PAUP version 4.0b10 (15) to maximum parsimony (MP); the confidence values for individual branches of the resulting tree were determined by bootstrap analysis with 1,000 replicates. Bayesian analysis (BA) was performed with MrBayes version 3.1.2 (16) software with 1,000,000 generations using the GTR+I+G substitution model. Partial DNA sequences obtained from the amplified PCR products (gltA, RR0155-rpmB, RR1240-tlc5^b, cspA-ksgA, and ompB) were concatenated and aligned with corresponding sequences of different strains of R. rickettsii available in GenBank using CLUSTAL X (17) and adjusted manually using GeneDoc (18). Partial sequences of the ompA gene were not included because the region of the gene that was amplified in the present study showed no polymorphism among the R. rickettsii isolates. Corresponding sequences of R. rickettsii strain Hlp#2 (CP003311) and Rickettsia philipii strain 364D (CP003308) were used as outgroups.

Fifteen rickettsial isolates (designated IAL 1 to 15) from BSF patients (13 fatal cases) in the state of São Paulo, southeastern Brazil, were identified as R. rickettsii, since their gltA (737 nt), ompA (491 nt), and ompB (787 nt) DNA sequences were 100% identical to each other and to corresponding sequences on the genome of the R. rickettsii strain Brazil (GenBank accession no. CP003305). While the ompA partial sequences were also 100% identical to corresponding sequences of R. rickettsii strains from North America (e.g., CP000848 and CP000766), the ompB partial sequences were 100% identical to that of the North American Sheila Smith strain (CP000848) and, at the same time, differed from those of other North American strains (e.g., CP000766 and CP003307) by one single nucleotide polymorphism. The gltA sequences of the 15 Brazilian human isolates differed by an extra codon (CGG) compared to those of several North American strains, such as Sheila Smith (CP000848) and Bitterroot (U59729). This extra codon was also present in the R. rickettsii tick isolates (Taiaçu, Itu, and Rs1) from Brazil (12–14). The sequences of three intergenic regions (249 nucleotides [nt] for RR0155-rpmB, 315 nt for RR1240-tlc5^b, and 362 nt for cspA-ksgA) were determined for the 15 human isolates and for the three tick isolates (Taiaçu, Itu, and Rs1) from Brazil. For each intergenic region, the sequences were 100% identical to each other (no polymorphisms were detected), and when subjected to BLAST analysis, they were 100% identical to the corresponding sequences of the R. rickettsii strain Brazil (CP003305).

For the concatenated phylogenetic analysis, which included a total of 2,392 nt, the sequences of the 15 human isolates (IAL 1 to 15) and 3 tick isolates (Taiaçu, Itu, and Rs1) from Brazil were aligned with the corresponding sequences from 10 other R. rickettsii isolates available in GenBank (6 from the United States, 1 from Costa Rica, 2 from Colombia, and 1 from Brazil) (Table 2). The 15 human and 3 tick isolates of R. rickettsii from Brazil formed a clade under high bootstrap support (99% to 100%), with all three South American isolates available in GenBank (1 from Brazil and 2 from Colombia) and with the Central American isolate from Costa Rica (Fig. 1). This Central/South American clade, ecologically associated with 4 different tick species (namely, A. aureolatum, Amblyomma patinoi, A. sculptum, and R. sanguineus) (Table 2), had a sister group formed by the North American Sheila Smith and Bitterroot strains that were ecologically associated with the tick vector Dermacentor andersoni. A more divergent clade was composed of North American isolates that have been ecologically associated with the ticks Dermacentor variabilis and R. sanguineus.

TABLE 2.

Isolates of Rickettsia rickettsii used in the phylogenetic analysis of the present study

Isolate	Isolation source	Clinical outcome	Source of DNA sequences for rickettsial genes (GenBank accession no. or reference no.) for:					Haplotypea	Tickb
			gltA	RR0155-rpmB	RR1240-tlc5b	cspA-ksgA	ompB
Iowa	Tick		CP000766	CP000766	CP000766	CP000766	CP000766	A	Dermacentor variabilis
Hino	Human	Fatal	CP003309	CP003309	CP003309	CP003309	CP003309	A	D. variabilis
Hauke	Human	Fatal	CP003318	CP003318	CP003318	CP003318	CP003318	A	D. variabilis
Arizona	Human	Fatal	CP003307	CP003307	CP003307	CP003307	CP003307	B	Rhipicephalus sanguineus sensu lato
Sheila Smith	Human	Fatal	CP000848	CP000848	CP000848	CP000848	CP000848	C	Dermacentor andersoni
Bitterroot	Tick		RRU59729	EF216032	EF215983	EF215860	X16353	C	D. andersoni
Costa Rica	Human	Fatal	19, 27	EF216038	EF215987	EF215872	27	D	?
I12B (Villeta)	Tick		KJ735644	KJ735647	KJ735648	KJ735649	KJ735646	D	Amblyomma patinoi
Rs1	Tick		13	This study	This study	This study	13	D	Amblyomma sculptum
Itu	Tick		KF742602	This study	This study	This study	KF742604	D	A. sculptum
Taiaçu	Tick		DQ115890	This study	This study	This study	12	D	Amblyomma aureolatum
IAL 1–2c	Human	Fatal	This study	This study	This study	This study	This study	D	A. aureolatum
IAL 4, 9d	Human	Cure	This study	This study	This study	This study	This study	D	A. sculptum
IAL 3, 5–8, 10–15e	Human	Fatal	This study	This study	This study	This study	This study	D	A. sculptum
Brazil	Human	Unknown	CP003305	CP003305	CP003305	CP003305	CP003305	D	A. sculptum
Colombia	Human	Fatal	CP003306	CP003306	CP003306	CP003306	CP003306	E	A. patinoi

^aRefers to the concatenated haplotypes shown in the phylogenetic tree (Fig. 1).

^bFor tick isolates, refers to the tick species from which the isolate was obtained; for human isolates, refers to the incriminated vector of R. rickettsii to humans in the area of origin of the isolate, according to Ogrzewalska et al. (32) for Amblyomma aureolatum; A. A. Faccini-Martínez, F. B. Costa, T. E. Hayama-Ueno, A. Ramírez-Hernández, J. Cortés-Vecino, M. B. Labruna, and M. Hidalgo (submitted for publication), Nava et al. (33), and Pacheco et al. (13) for Amblyomma patinoi and Amblyomma sculptum; and Karpathy et al. (10) for Dermacentor variabilis, Dermacentor andersoni, and Rhipicephalus sanguineus sensu lato. The question mark represents an unknown vector, according to Hun et al. (27).

^cGeographic origins (municipalities in the state of São Paulo, Brazil) of these isolates: São Paulo, IAL 1; São Bernardo do Campo, IAL 2.

^dGeographic origins (municipalities in the state of São Paulo) of these isolates: Piracicaba, IAL 4; Valinhos, IAL 9.

^eGeographic origins (municipalities in the state of São Paulo) of these isolates: Valinhos, IAL 3, 5, 8, 11, 12, and 15; Campinas, IAL 6 and 13; Jaguariúna, IAL 7; Piracicaba, IAL 10; Limeira, IAL 14.

FIG 1.

Molecular phylogenetic analysis of Rickettsia rickettsii isolates from the United States, Costa Rica, Colombia, and Brazil. A total of 2,392 unambiguously aligned nucleotide sites of two rickettsial genes (gltA and ompB) and three intergenic regions (RR0155-rpmB, RR1240-tlc5^b, and cspA-ksgA) were concatenated and subjected to analysis by maximum-parsimony and Bayesian methods. Corresponding sequences of Rickettsia philipii strain 364D and R. rickettsii strain Hlp#2 were used as an outgroup. Numbers at nodes are support values derived from bootstrap and posterior probability for MP and BA analyses (MP/BA). Sequence codes A to E, each with a different color, represent the five haplotypes generated from the 28 isolates of R. rickettsii described in Table 2. Gray braces or arrows at the clades indicate the tick species that has been ecologically associated with the R. rickettsii isolates. The geographical region of origin of the 28 isolates and their corresponding haplotypes (A to E) are indicated on the map of the American continents. (The map is reprinted from http://www.usgs.gov/.)

As reported in previous studies (10, 19–21), the North American isolates of R. rickettsii presented relatively high polymorphism compared to that of the Central/South American isolates. Our concatenated analysis showed the formation of 3 North American haplotypes (A, B, and C), each associated with a different tick species (Fig. 1, Table 2). Conversely, excluding the Colombia strain (haplotype E), there was a single haplotype (D) in Central/South America, although it was associated with 4 different tick species. Geographic distances cannot be inferred from this discrepancy because the distance between Costa Rica and southeastern Brazil is much higher than the distances between distinct North American isolates (Fig. 1). Interestingly, while low-, mild-, and high-virulent strains of R. rickettsii have been reported in the eastern and western parts of the United States (20, 22), only high-virulent strains, responsible for high fatality rates, have been reported in Central/South America, regardless of the tick vector (2, 23–29). While our results of no polymorphisms among Central/South American isolates was biased because most of these isolates were derived from fatal cases, a recent study reported the same clade distribution as shown in our Fig. 1 when analyzing the intergenic regions of R. rickettsii derived from fatal cases from the United States, Mexico, and Central/South America (21).

The relatively high level of polymorphism among North American isolates of R. rickettsii and the contrasting low level of polymorphism in Central/South America suggest that R. rickettsii radiated in North America and was introduced into South America during more recent periods. This scenario may also explain why there is a mixture of highly and less virulent R. rickettsii strains in North America (due to longer coevolving periods with vertebrates), while only highly virulent strains have been found in South America.

It has been suggested that the higher fatality rates of BSF, compared to those of RMSF in the United States, are related to delayed treatment, the use of less effective antibiotics, and more virulent R. rickettsii strains occurring in Brazil (2, 30, 31). The present study corroborates previous studies (10, 19–21) that provided genetic evidence for a very low level of polymorphism occurring among R. rickettsii isolates from South America. This fact should be a significant reason for the much higher fatality rates of BSF, although the others discussed above may also be contributing factors.

Nucleotide sequence accession numbers.

The GenBank nucleotide sequence accession numbers of the partial sequences of R. rickettsii generated in this study are KJ994337, KJ994338, and KJ994339 for the gltA, ompA, and ompB genes, respectively, and KJ994340, KJ994341, and KJ994342 for the RR0155-rpmB, RR1240-tlc5^b, and cspA-ksgA intergenic regions, respectively.