Louse borne typhus (also called epidemic typhus) was one of man's major scourges, and epidemics of the disease can be reignited when social, economic, or political systems are disrupted. The fear of a bioterrorist attack using the etiologic agent of typhus, Rickettsia prowazekii, was a reality. An attenuated typhus vaccine, R. prowazekii Madrid E strain, was observed to revert to virulence as demonstrated by isolation of the virulent revertant Evir strain from animals which were inoculated with Madrid E strain. The mechanism of the mutation in R. prowazekii that affects the virulence of the vaccine was not known. We sequenced the genome of the virulent revertant Evir strain and compared its genome sequence with the genome sequences of its parental strain, Madrid E. We found that only a single nucleotide in the entire genome was different between the vaccine strain Madrid E and its virulent revertant strain Evir. The mutation is a single nucleotide insertion in the methyltransferase gene (also known as PR028) in the vaccine strain that inactivated the gene. We also confirmed that the vaccine strain E did not cause fever in guinea pigs and the virulent revertant strain Evir caused fever in guinea pigs. We concluded that a single nucleotide insertion in the methyltransferase gene of R. prowazekii attenuated the R. prowazekii vaccine strain E. This suggested that an irreversible insertion or deletion mutation in the methyl transferase gene of R. prowazekii is required for Madrid E to be considered a safe vaccine.
Citation: Liu Y, Wu B, Weinstock G, Walker DH, Yu X-j (2014) Inactivation of SAM-Methyltransferase is the Mechanism of Attenuation of a Historic Louse Borne Typhus Vaccine Strain. PLoS ONE 9(11): e113285. https://doi.org/10.1371/journal.pone.0113285
Editor: Roman Ganta, Kansas State University, United States of America
Received: August 14, 2014; Accepted: October 21, 2014; Published: November 20, 2014
Copyright: © 2014 Liu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. The genome sequence of Rickettsia prowazekii was deposited in NCBI BioProject under ID PRJNA254815.
Funding: This study was supported by a grant from the Institute of National Health (U01AI71283). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The co-author Xue-jie Yu is a PLOS ONE Editorial Board member, but this does not alter the authors' adherence to PLOS ONE Editorial policies and criteria.
Louse borne typhus (also called epidemic typhus) was one of man's major scourges from the 1500 s through the mid-20th century and frequently played a decisive role in wars in Europe during this period. Thus, it has affected the course of European history . It has killed millions of people, and the threat of louse borne typhus is still real. Louse-borne typhus occurs in epidemics when social, economic, or political systems are disrupted exposing a large population such as refugees to louse infestation due to lack of hygiene and reactivation of latent infection associated with the harsh and stressful conditions. This situation has been observed in outbreaks of typhus in Burundi, Algeria, Peru, and Russia . In 1997, it was estimated that as many as 100,000 cases of typhus occurred in the refugee camps of Burundi . Rickettsia prowazekii, the causative agent of louse borne typhus, may be also used deliberately by terrorists in a biologic weapon attack. Rickettsia prowazekii is listed as a Select Agent with rigorous legal restrictions on its possession and use in research. The former Soviet Union developed R. prowazekii as a biologic weapon during the 1930s . During World War II, the Japanese performed human experiments with rickettsial agents for purposes of biologic weapon development during their occupation of China .
Epidemic typhus is transmitted by the human body louse, Pediculus humanus corporis. However, the louse is only a vector, but not a reservoir because the louse dies 5–7 days after infection with R. prowazekii. Rickettsia prowazekii multiplies in louse gut epithelial cells, which detach, rupture and release rickettsiae into the feces and louse hemocoel, killing the insect. The louse feces containing rickettsiae is scratched into the skin, rubbed into mucous membranes such as the conjunctiva, or inhaled. Humans can develop latent infection after acute louse-borne typhus and serve as reservoirs of R. prowazekii. Rickettsia prowazekii is subsequently reactivated causing recrudescent typhus fever (Brill-Zinsser disease) when latently infected persons experience waning immunity or stressful conditions . Rickettsia prowazekii is also maintained in a zoonotic cycle involving flying squirrels (Glaucomys volans) and their specific fleas and lice in the eastern United States. Sporadic epidemic typhus occurring in the United States is transmitted to humans by the flea of flying squirrels .
Lethality of epidemic typhus in the pre-antibiotic era was usually 15%, and as high as 60%. Even with the availability of effective rickettsiostatic antibiotic treatment, the mortality rate is approximately 4% because of late diagnosis and delay in starting appropriate therapy .
After infection, humans develop lifelong immunity to R. prowazekii infection or become latently infected. Louse borne typhus is preventable by vaccination using live or dead organisms. The attenuated Madrid E (E) strain of R. prowazekii, which had been demonstrated as an effective vaccine, is a spontaneous laboratory variant that was originally isolated from a typhus patient in Madrid in 1941 and passed in rapid succession in embryonated eggs 255 times. The E strain was limited in virulence for guinea pigs and humans. The E strain is protective and provided long term immunity against louse borne typhus when tested as a vaccine during the 1950s and 1960s. Ninety-four percent (170/181) of immunized persons were protected from natural infection by epidemic typhus compared to the unvaccinated controls in a 14-month period after vaccination . Ninety-six percent (27/28) of volunteers who were vaccinated with the E strain and subsequently challenged with Breinl strain at intervals from 2 months to 36 months remained healthy following challenge, and 83% (5/6) of the volunteers who were challenged after 48 to 66 months were protected . However, the E strain vaccine caused a late reaction in up to 14% of vaccinated persons 9–14 days after inoculation. The late reaction varied from simple malaise and mild headache to modified typhus characterized by fever, headache, malaise and occasionally a rash in a small proportion of subjects . The reason for the late reaction was not known at the time. In 1971 a virulent revertant strain (Evir) was isolated by passage of E strain in guinea pigs and/or mice. After the isolation of Evir strain, E and Evir strains have been compared for their biological properties and genetic difference. E strain grows poorly in both human and mouse macrophages, whereas virulent Evir strain and another virulent strain (Breinl) grow very well in these cells . A previous study demonstrated that the patterns of methylation of lysine in the surface protein antigens of E and Evir were different. The surface antigen, outer membrane protein B, of the virulent strains Breinl and Evir is extensively methylated, and that of the avirulent strain E is deficient in protein methylation –. Coincidently, we recently demonstrated that the SAM-methyltransferase gene (smt) was inactivated in E strain by a single nucleotide insertion, which was absent in the Evir strain . The important question is whether smt is the only gene difference between E strain and Evir strain that causes the reversion of avirulent E strain to virulent Evir strain. To identify the mechanism of attenuation of E strain, we sequenced the genome of Evir strain and compared it to the genome of E strain in the database.
Materials and Methods
Rickettsia prowazekii strain and genomic DNA preparation
Rickettsiae were cultivated in L929 cell monolayers in 150 cm2 flasks. For preparation of genomic DNA, when 100% of L929 cells were heavily infected, cells were disrupted by vortexing vigorously with 4-mm glass beads to release rickettsiae; the cell suspension was centrifuged at 200×g for 10 minutes to remove unbroken cells and nuclei. The supernatant was passed sequentially through a 5 µm-syringe filter and a 3 µm-syringe filter. The filtered solution was centrifuged at 12,100×g for 30 min to collect the rickettsiae. Rickettsial genomic DNA was extracted using the QIAamp DNA Mini Kit (Valencia, CA). In order to determine the TCID50, rickettsial stocks were diluted in 10-fold increments from 10−1 to 10−9. Each dilution was inoculated onto 4 wells of L929 cell monolayers in 6-well plates, and infectivity was determined by examining stained monolayers 12 days later. The TCID50 was calculated by the Reed-Muench formula.
All experiments were approved by the Institutional Animal Care and Use Committee of the University of Texas Medical Branch (Protocol #: 05-11-074) and conducted in strict compliance with the NIH Guidelines for the Care and Use of Laboratory Animals to minimize any pain or suffering of the animals.
We used guinea pigs to test the virulence of R. prowazekii vaccine E strain and virulent strain Evir because guinea pigs are susceptible to R. prowazekii infection and develop fever after infection. Other animals except for monkeys are not susceptible to R. prowazekii infection. Only animal experiments can determine the virulence of R. prowazekii strains, which vary from avirulence to virulence to humans. Therefore, guinea pigs were used in this study. Twenty one male Hartley strain guinea pigs weighing 250 to 300 grams were randomly divided into seven groups with three guinea pigs in each group. The first group was inoculated i.p. with 106 uninfected L929 cells (ATCC CCL-1) in which rickettsiae were cultivated as a negative control. The second, third and fourth groups were inoculated i.p. with Evir strain at doses of 104, 105, and 106 organisms, respectively. The fifth, sixth, and seventh groups were inoculated i.p. with E strain at doses of 104, 105, and 106 organisms, respectively. Body weight and rectal temperature of the guinea pigs were measured daily from the day of inoculation to day 13 after inoculation. On day 14 post infection all guinea pigs were anesthetized by inhaling isofluorane followed by cardiac puncture to collect blood. The guinea pigs were euthanized by using CO2 after bleeding.
Genomic DNA sequencing
The Evir strain was sequenced using the Illumina Genome Analyzer. A subset of reads from which DNA contaminants representing L929 cells and low quality reads had been removed, were de novo assembled into contigs using the Velvet assembler with optimized parameters. The nucmer algorithm (MUMmer package, v3.19) was used to align assembled contigs to the parent Madrid E strain genome and to generate single nucleotide polymorphisms (SNPs) and insertion or deletion (indel) calls. Illumina reads were also mapped directly to the parent genome using BWA, and SNPs were called with Samtools. Variants detected by both programs were further analyzed and validated by PCR for both strains, E and Evir. The Evir strain genome sequence was deposited in GenBank (Access No. PRJNA254815).
Alignment of the genomes of E strain and Evir strain with nucmer revealed that the two genomes were perfectly aligned. There was no large insertion or deletion and no inversion of DNA sequences.
Initial comparison of the genomes of the Evir strain and the E strain (DOW) in GenBank revealed 5 SNPs and 26 indels. We amplified and sequenced the DNA of these SNPs and indels from strain E (DHW) and Evir strain from our laboratory to confirm the SNPs and indels. We found that only one nucleotide was different between Evir strain and E strain (DHW), which was an insertion of a single A at 732 bp in the SAM methyltransferase gene of the strain E (DHW) genome and interrupted the SAM methyltransferase gene near the middle (Table 1). All SNPs and other indels were not present in the strain E (DHW), suggesting that either the genome sequence of strain E(DOW) in GenBank, which was one of the earliest bacterial genomes reported, has errors or the E strain used for genome sequencing was different from E strain (DHW) in our laboratory.
To determine whether the SNPs and indels in E strain (DOW) genomic sequence in GenBank were genuine or caused by sequence errors, we acquired a strain E (DOW) that was used for the original R. prowazekii genomic sequence from Dr. David Wood's laboratory at the University of South Alabama. We amplified and sequenced all SNPs and indels of strain E (DOW). In addition to the insertion of a T at 732 bp in the SAM methyltransferase gene, strain E (DOW) has two additional nucleotide insertion mutations, which interrupted genes Rp061 and recO, respectively (Table 1). An insertion of a T at position 64 occurred in Rp061, which introduced a premature stop codon, resulting in a peptide that lacks the first 19 amino acids. Rp061 is a conserved protein in Rickettsia and is not found in other bacteria. The Rickettsia recO encoded a DNA repair protein. The insertion of a T at position 326 of recO in R. prowazekii also interrupted the gene. Other SNPs which were different between Evir strain and E strain (DOW) in GenBank were not identified in strain E (DHW) and strain (DOW). It suggested that these differences were most likely caused by sequence errors in the GenBank sequence (Accession No: AJ235269.1) (Table 2). The genotype of strain E (DHW) was smt−, recO+, rp061+ and the genotype of strain E (DOW) was smt−, recO−, rp061−.
All guinea pigs inoculated with 105 and 106 organisms of Evir strain developed a fever. After inoculation with a high dose (106 organisms) of Evir strain, one guinea pig had fever from day 5 to day 10, one guinea pig had fever from day 6 to day 11, and one had fever on days 7 and 8. With the medium dose (105 organisms) of Evir strain, all 3 guinea pigs had fever from day 8 to day 11. With the low dose (104 organisms) of Evir strain, only 1 of 3 guinea pigs had a fever from day 8 to 10 after inoculation. The guinea pigs inoculated with the highest dose of Evir strain had a shorter incubation period and a longer period of fever (Figures 1 and 2). In contrast, regardless of the dose (from 104 to 106), no guinea pig inoculated with strain E (DHW) developed fever, and the growth rate of guinea pigs as determined by body weight was not affected by strain E (DHW) infection compared to L929 cell mock-infected guinea pigs. During the febrile period, the growth rate of the guinea pigs inoculated with 105 and 106 organisms of strain Evir was significantly slower than the guinea pigs inoculated with strain E (DHW) or L929 cells. This result confirmed that strain E (DHW) was avirulent and strain Evir was virulent for guinea pigs.
L929 cell suspension was used as control.
We sequenced the genome of R. prowazekii strain Evir and compared it with the genome sequence of strain E (DOW) in GenBank and identified 31 insertions, deletions and SNPs between Evir and E strain (DOW). We sequenced all insertion, deletion and SNPs in strain E (DHW), which we had in our laboratory. We found only one mutation was truly a mutation in strain E (DHW). The mutation interrupted smt into two open reading frames, which was designated as rp027 and rp028 previously . We were curious whether the remaining differences between Evir and E (DOW) strain were genuine or caused by sequence errors in the GenBank sequence. We amplified and sequenced all SNPs and indels in a prototype strain E (DOW), which was used for generating the genome sequence of R. prowazekii in GenBank (AJ235269.1). Sequence of strain E (DOW) demonstrated that three nucleotides were different between strains Evir and E (DOW), in addition to smt, two additional single nucleotide mutations interrupted recO, and rp061 in strain E (DOW). The recO and rp061 mutations have been found in E strain from other laboratory previously . recO and rp061 mutations were not identified in all strains of R. prowazekii sequenced.
E strain is avirulent for humans and animals, but passages in animals selected a more virulent strain, Evir. We do not know the origins of strain E (DHW) and strain E (DOW). It is most likely that strain E first acquired the mutation in smt and further passages led to accumulation of more mutations such as recO and rp061.
Mild typhus observed in 14% of vaccinated humans was most likely caused by a mixed population of virulent and avirulent strains. Previously we demonstrated that the methyltransferase gene is inactivated by a frame shift mutation in E strain, but the mutation does not exist in the virulent strain Evir . To determine whether mutation in smt determines the attenuation of E strain, we evaluated the virulence of E (DHW) strain in guinea pigs. We demonstrated that E strain (DHW) was avirulent for guinea pigs, but Evir strain is virulent for guinea pigs. After we found that strain E (DHW) and strain E (DOW) were genetically different, we further tested the virulence of E (DOW) in guinea pigs. As expected that E (DOW) also did not cause fever or body weight loss in guinea pigs (data not shown). Since the only three genes different between Evir and E (DOW) and one gene (smt) different between strain Evir and strain E (DHW), we reasoned that the reverse mutation in smt determined the virulence of R. prowazekii. As smt encodes a methyltransferase, these results clearly indicate that the methyltransferase gene is critical for the virulence of R. prowazekii and the attenuation of E strain might be caused by inactivation of SAM-methyltransferase. It is apparent that SAM-methyltransferase is a virulence factor and deficiency of this enzyme may affect the growth or survival of Rickettsia inside the animal host.
A previous study demonstrated that avirulent E strain and Erus had a 12 nucleotide deletion in rp827, but all virulent strains including Breinl, Rp22, and Evir had no deletion in rp827. In contrast, we found that a 12 nucleotide stretch was deleted from all three strains derived from E strain including Evir, strain E (DOW), and E strain (DHW). It seems highly improbable to have a reverse mutation that inserts 12 nucleotides in rp827 in Evir strain to restore the mutated gene.
We concluded that a single nucleotide insertion in the methyltransferase gene smt attenuates R. prowazekii vaccine strain E, and a reversion mutation makes the revertant strain Evir. It suggests that a safe and irreversible vaccine of R. prowazekii can be achieved by genetic deletion of nucleotides of the smt gene of Rickettsia prowazekii E strain.
Conceived and designed the experiments: XJY GW DHW. Performed the experiments: YL BW. Analyzed the data: XJY YL. Wrote the paper: XJY DHW.
- 1. Conlon JM (2014) The historical impact of epidemic typhus. Available: http://entomology.montana.edu/historybug/typhus-conlon.pdf.
- 2. Raoult D, Woodward T, Dumler JS (2004) The history of epidemic typhus. Infect Dis Clin North Am 18: 127–140.
- 3. Alibek K, Handelman S (1999) Biohazard the chilling true story of the largest covert biological weapons program in the world, told from the inside by the man who ran it. Random House, New York.
- 4. Harris S (1992) Japanese biological warfare research on humans: a case study of microbiology and ethics. Ann N Y Acad Sci 666: 21–52.
- 5. Zinsser H, Castaneda MR (1933) On the isolation from a case of Brill's disease of a typhus strain resembling the European type. New England Journal of Medicine 209: 815–819.
- 6. Duma RJ, Sonenshine DE, Bozeman FM, Veazey JM Jr, Elisberg BL, et al. (1981) Epidemic typhus in the United States associated with flying squirrels. JAMA 245(22): 2318–2323.
- 7. Wisseman CL Jr (1972) Concepts of louse-borne typhus control in developing countries: the use of the living attenuated E strain typhus vaccine in epidemic and endemic situations; in: Klingberg MA, (ed): Immunity in viral and rickettsial diseases. New York, Plenum Publishing Corporation, pp 97–130.
- 8. Fox JP, Jordan ME, Gelfand HE (1957) Immunization of man against epidemic typhus by infection with avirulent Rickettsia prowazekii strain E. J Immunol 79: 348–354.
- 9. Penkina GA, Ignatovich VF, Balayeva NM (1995) Interaction of Rickettsia prowazekii strains of different virulence with white rat macrophages. Acta Virol 39(4): 205–209.
- 10. Rodionov AV, Eremeeva ME, Balayeva NM (1991) Isolation and partial characterization of the M(r) 100 kD protein from Rickettsia prowazekii strains of different virulence. Acta Virol 35(6): 557–565.
- 11. Turco J, Winkler HH (1994) Cytokine sensitivity and methylation of lysine in Rickettsia prowazekii EVir and interferon-resistant R. prowazekii strains. Infect Immun 62(8): 3172–3177.
- 12. Ching WM, Carl M, Dasch GA (1992) Mapping of monoclonal antibody binding sites on CNBr fragments of the S-layer protein antigens of Rickettsia typhi and Rickettsia prowazekii. Mol Immunol 29: 95–105.
- 13. Ching WM, Wang H, Davis J, Dasch GA (1993) Amino acid analysis and multiple methylation of lysine residues in the surface protein antigen of Rickettsia prowazekii; in: Angeletti RH, (ed): Techniques in Protein Chemistry vol. IV. San Diego, Academic Press, pp 307–314.
- 14. Zhang JZ, Hao JF, Walker DH, Yu XJ (2006) A mutation inactivating the methyltransferase gene in avirulent Madrid E strain of Rickettsia prowazekii reverted to wild type in the virulent revertant strain Evir. Vaccine 24: 2317–2323.
- 15. Andersson SG, Zomorodipour A, Andersson JO, Sicheritz-Ponten T, Alsmark UC, et al. (1998) The genome sequence of Rickettsia prowazekii and the origin of mitochondria. Nature 396: 133–140.
- 16. Bechah Y, El KK, Mediannikov O, Leroy Q, Pelletier N, et al. (2010) Genomic, proteomic, and transcriptomic analysis of virulent and avirulent Rickettsia prowazekii reveals its adaptive mutation capabilities. Genome Res 20: 655–663.