点击显示 收起
Laboratory of Human Bacterial Pathogenesis, Rocky Mountain Laboratories, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Hamilton, Montana
Population genetics and comparative genomics analyses of the pathogenic Yersinia species have indicated that arthropodborne transmission is an evolutionarily recent adaptation in Yersinia pestis, the agent of plague. We show that the infectivity of Y. pestis to its most proficient vector, the rat flea Xenopsylla cheopis, and subsequent transmission efficiency are both low. The poor vector competence of fleas likely imposed selective pressure that favored the emergence and continued maintenance of a hypervirulent Y. pestis clone. In particular, the rapidly fatal gram-negative sepsis that typifies plague is a consequence of the high threshold bacteremia level that must be attained to complete the transmission cycle. Epidemiological modeling predicts that, to compensate for a relatively short period of infectivity of the mammalian host for the arthropod vector, plague epizootics require a high flea burden per host, even when the susceptible host population density is high.
Yersinia pestis, the highly virulent agent of bubonic plague, is unique among the enteric group of gram-negative bacteria in having adopted an arthropodborne route of transmission. The results of population genetics analyses have indicated that Y. pestis strains worldwide constitute a highly uniform clone that diverged from the closely related Yersinia pseudotuberculosis within the last 150020,000 years, and comparative genomics analyses have confirmed a high degree of genetic identity [13]. Despite this recent common ancestry and genetic relatedness, the 2 sister species differ radically in natural history. Y. pseudotuberculosis is transmitted perorally in contaminated food and water and produces a relatively benign, self-limited enteric disease in humans and many animals. Y. pestis, by contrast, is transmitted primarily by fleas and is highly invasive and virulent. Presumably, these changes occurred during the evolutionarily short period since Y. pestis diverged from the Y. pseudotuberculosis progenitor.
To produce a transmissible infection, Y. pestis grows in the flea digestive tract as a biofilm that adheres to and fills the proventriculus, the valve that connects the esophagus to the midgut. The biofilm interferes with the normal valvular action and eventually blocks the flow of blood through the proventriculus. Transmission occurs when bacteria released from the biofilm are refluxed into the bite site when blocked fleas attempt to feed [46]. Within the first week after transmission from the flea, bacteria disseminate from the intradermal bite site to the regional lymph nodes to produce bubonic plague. From there, the bacteria rapidly invade the bloodstream and infect many internal organs. Infected tissues typically contain massive numbers of bacteria, and a high-density, usually fatal septicemia is a hallmark of plague [7]. Plague is characterized by an irregular pattern of periodic eruptive epizootics among susceptible rodents, which results in rapid amplification and geographic spread, followed by regression to enzootic focal areas [7].
Approximately 80 species of fleas, associated with some 200 species of wild rodents, have been found to be susceptible to Y. pestis infection [8]. Although the comparative vector efficiency of some of these fleas has been examined [814], their susceptibility to infection and transmission efficiency have not been directly measured. For the present study, we evaluated the vector competence of the rat flea Xenopsylla cheopis, the most efficient plague vector among 40 species that have been examined [814]. The data indicate that the transmission of Y. pestis by fleas is inefficient, and they suggest that one consequence of the development of a vectorborne route of transmission was the emergence and continued maintenance of a highly virulent Y. pestis clone. These and other direct measurements of parameters of X. cheopis vectorial capacity presented here permit epidemiologic modeling of plague transmission dynamics and predictions about the outcome of vector control.
MATERIALS AND METHODS
Infection of fleas with Y. pestis.
X. cheopis fleas were infected with Y. pestis strain KIM6+ by use of an artificial feeding system and mouse blood that contained known bacterial concentrations, prepared as described elsewhere [5, 15, 16]. To determine the infectious dose of Y. pestis for fleas, blood meals that contained 4.0 × 1064.2 × 108 Y. pestis/mL were used to infect X. cheopis. For each experiment, a sample of 20 female fleas was collected and stored at -80°C immediately after feeding; this sample was later used to determine the average dose of Y. pestis acquired in the blood meal. The remaining fleas were maintained and fed on healthy, uninfected mice on days 2 and 6 after infection, and a second sample of 20 female fleas was collected on day 7 to determine the infection rate and bacterial load. The volume of the blood meal was calculated in female fleas, collected immediately after their infectious blood meal, by dividing the average number of Y. pestis colony-forming units by the number of colony-forming units per milliliter of the blood they fed on. The results from 27 separate feeding experiments involving 469 fleas were used to determine the average and SD.
Other groups of fleas that ingested an infectious blood meal containing 5 × 108 Y. pestis/mL were fed twice weekly for 4 weeks on uninfected mice. After each of these subsequent feedings, the fleas were examined under a dissecting microscope for evidence of proventricular blockage, which was diagnosed by the presence of fresh blood only in the esophagus [16].
Flea survival and blockage rates.
Data were compiled from 930 X. cheopis infected with Y. pestis strains KIM6+ or 195/P in 10 separate experiments [16, 17]. All fleas fed on blood from either a membrane feeder or a mouse with a septicemia level of 5 × 108 Y. pestis/mL. Fleas were kept at 21°C and 75% relative humidity, fed twice weekly on uninfected mice, and monitored for mortality and proventricular blockage for 4 weeks after the infectious blood meal [16].
Fleaborne transmission.
Individual blocked fleas were allowed to feed for 1 h on 0.3 mL of sterile heparinized mouse blood in a miniature artificial feeding device. Alternatively, blocked fleas were placed on shaved areas of uninfected mice. Immediately after feeding attempts, mice were killed; the resulting skin-biopsy samples of bite sites were stored at -80°C. For both the mouse and artificial feeding experiments, all fleas were examined microscopically, to verify blockage and attempted feeding.
Quantitation of Y. pestis in fleas and tissues.
Individual fleas were surface sterilized and triturated, and dilutions of each triturate were plated on brain-heart infusion agar that contained 0.1% Irgasan (Ciba-Geigy), for colony-forming unit counts after a 72-h incubation at 28°C [5, 16]. To determine the number of Y. pestis colony-forming units transmitted by blocked fleas, blood was removed from the membrane feeder, the mouse skin membrane was washed 3 times with 0.1 mL of PBS, and blood and washings were spread on Yersinia selective agar (YSA) plates. The washed skin from the membrane feeder and the biopsies of bite sites from mice were digested with 5 L of a 10% (wt/vol) solution of collagenase A (Roche Biochemicals) in 500 L of 154 mmol/L NaCl, 5.6 mmol/L KCl, and 1.7 mmol/L CaCl2 for 1 h at 37°C. The digests were then manually triturated and plated on YSA. Control experiments, in which colony-forming units from replicate aliquots of the buffer containing Y. pestis and uninfected skin with or without added collagenase, verified that the enzyme treatment did not affect bacterial viability or growth.
Quantitative TaqMan polymerase chain reaction (PCR) was performed on total DNA extracted from skin-biopsy samples of bite sites by use of a primer and probe set for the Y. pestis pla gene. Numbers of Y. pestis in the bite samples were determined by extrapolation from a standard curve generated from TaqMan results of uninfected skin-biopsy samples to which known numbers of Y. pestis were added.
RESULTS AND DISCUSSION
Oral infectious dose of Y. pestis for X. cheopis. The ID50 was determined by allowing groups of fleas to feed on blood that contained different concentrations of Y. pestis. The ID50 for X. cheopis was 4.8 × 103 Y. pestis (95% confidence interval, 2.9 × 1031.4 × 104), determined by linear-regression analysis of the data in table 1. In comparison, the ID50 of Y. pestis for susceptible mammals is <10 bacteria [7]. The average volume of blood ingested by X. cheopis was calculated to be 0.12 ± 0.07 L; estimates of 0.030.5 L have been reported elsewhere [18]. Thus, the threshold septicemia level that would provide 1 ID50 (4800 Y. pestis) in the 0.12 L ingested by a feeding flea is 4.0 × 107 bacteria/mL peripheral blood. The high ID50 of Y. pestis for X. cheopis was not strain dependent or an artifact of the membrane feeder infection method; the rate and kinetics of infection of X. cheopis by the Y. pestis strain 195/P that fed on a mouse with a septicemia level of 4.3 × 108 Y. pestis/mL were the same as those observed for fleas artificially infected by use of blood with the same concentration of Y. pestis 195/P or KIM6+ [16]. Earlier studies of X. cheopis and other species consistently reported that fleas could not be reliably infected unless they fed on an animal with terminal septicemia (typically, >108 Y. pestis/mL peripheral blood) and suggested that the threshold of infection was high [8, 14, 18]. Our results confirm that hypothesis. The high oral infectious dose for fleas can be explained by the fact that Y. pestis does not adhere to or invade the midgut epithelium and is subject to rapid elimination through peristalsis and excretion in flea feces [4, 6].
Transmission efficiency of X. cheopis.
Individual blocked fleas were allowed to feed on mice or on a membrane feeder, and the number of Y. pestis colony-forming units transmitted per flea was determined (figure 1). Less than one-half (14/31 [45%]) of the samples were positive; these contained 24970 Y. pestis (average ± SD, 636 ± 1324 Y. pestis; median, 82 Y. pestis). These results are consistent with the low LD50 of Y. pestis in mammals [7] and with the results of previous reports that plague results from the bite of an individual blocked X. cheopis only 36%47% of the time [9, 1214]. Only 2 of 14 positive samples contained >1000 Y. pestis, even though the corresponding transmitting fleas contained an average of 4.8 × 105 Y. pestis (median, 4.0 × 105 Y. pestis; range, 2.1 × 1031.9 × 106 Y. pestis)a median transmission efficiency of <0.1%.
A likely explanation for the irregular distribution (figure 1) is that, because the transmitted bacteria derive from a biofilm in the flea proventriculus, the infectious units may consist not only of individual bacteria but also of various-sized clusters of bacteria that break free from the biofilm [6]. Quantitative PCR (which measured the number of Y. pestis genome equivalents transmitted) performed on an additional 25 skin biopsies verified that quantitation of the colony count did not underestimate the actual number of bacteria transmitted (data not shown). Our results differ from a previous indirect estimate of 11,00024,000 Y. pestis transmitted per flea, which was based on a single blocked X. cheopis flea [9].
Evolution of increased virulence.
The high infectious threshold for fleas and the low rate of transmission of small numbers of bacteria imply that only a strain able to produce high-density septicemia from a small intradermal infectious dose is able to complete the vector-host life cycle. Thus, the evolutionary change to fleaborne transmission would have imposed new selective pressures that strongly favored the coevolution of increased virulence. For the Y. pestis progenitor, the acquisition of 2 plasmids, one of which encodes a phospholipase D required for survival in the flea midgut and the other of which encodes a plasminogen activator that increases invasiveness in the mammalian host, were important steps in this evolutionary process [17, 1921]. Epidemiological models have hypothesized that pathogens evolve a level of virulence that maximizes their basic reproduction ratio (R0)the number of new infections that a single infected host gives rise to during its infectious life span [22]. R0 is directly proportional to the transmission rate and to the length of time during which a host remains infectious. A high transmission rate correlates with high parasite density in the host, but the associated virulence tends to shorten the host's life span and to reduce the infectious period. A low level of pathogen reproduction in the host has little effect on host longevity (decreased virulence) but results in a lower transmission rate. Therefore, evolutionarily stable strains with an optimal R0 are thought to have struck a balance between these 2 opposing tendencies that tempers virulence [2224].
Does Y. pestis fit this model The poor vector competence of fleas mandates that Y. pestis achieve high density in peripheral blood, and because gram-negative septicemias are rapidly fatal, the infectious period of the host to feeding fleas is brief. The transmission rate of an arthropodborne agent, however, depends only partly on vector competence or efficiency. Other factors, such as vector density, the number of bites experienced per day, the vector survival rate during the time it takes for a transmissible infection to develop, and the density of the susceptible host population also contribute to the overall vectorial capacity, which is a better measure of transmission rate [25]. Most rodents maintain a permanent flea population in their fur and in their nests or burrows, and fleas feed frequently. Thus, even though the time between threshold septicemia and death is short, if the flea burden is sufficiently high, there is good probability that 1 resident flea will ingest an infectious blood meal. Because fleas do not readily leave their host mammal, killing the host may actually be important to Y. pestis transmission. Death of the host compels the resident infected fleas to seek a new, healthy hosta necessary step to complete the transmission cycle [26]. The persistent feeding behavior of a blocked, starving flea also enhances vectorial capacity. These ecological factors weaken the linkage between extreme virulence and decreased opportunity for transmission, which is one of the key assumptions of most epidemiological models, and likely permitted the evolution of hypervirulence in Y. pestis.
For plague, models predicting a trade-off between virulence and transmission may better fit the insect host. Unlike most arthropodborne agents, Y. pestis is potentially fatal to its vector, because fleas with proventricular blockage rapidly starve to death [10, 11, 14]. However, only approximately one-half of chronically infected X. cheopis develop blockage, and this rate is lower for several other wild rodent flea species [5, 8, 9, 13, 16]. Thus, both the ID50 and the LD50 of Y. pestis are much higher for X. cheopis than for mammals, which may be important for the maintenance of transmission cycles. The evolution of virulence has been hypothesized to be more restricted in the arthropod than in the mammal, because an active vector is more important for transmission than a healthy, infected host [23].
Vector density and transmission dynamics.
The X. cheopis density per primary host that is necessary to support an epizootic was calculated by use of a modification of Macdonald's equation, a standard epidemiologic model of vectorborne disease dynamics [22, 25, 27]:
where m is the flea density per host; a is the biting rate per day of an uninfected flea; a is the biting rate per day of a blocked flea; b is the probability of blockage developing in a flea after it feeds on an infected host with septicemia at the threshold level or greater; p and q are the daily survival probability of unblocked and blocked fleas, respectively; n is the number of days required for proventricular blockage to develop after the infectious blood meal (the extrinsic incubation period) and pn is the probability of a flea surviving until blockage develops; and 1/r is the duration of infectivity (life expectancy) of the host after the threshold septicemia level is reached. In the modified equation, distinct terms were used for blocked (infective) fleas, because their daily biting rate and survival differ greatly from those of unblocked fleas. Experimental values for these parameters were determined from 930 infected fleas: 38% of the fleas developed proventricular blockage 728 days (median, 14 days) after the infectious blood meal (b = 0.38; n = 14), and mortality of unblocked fleas at 14 days was 15% (pn = 0.85) for both artificially infected and naturally infected fleas [15, 16] (unpublished data). Unblocked fleas fed every 3 days (a = 0.33); blocked fleas attempted to feed every day (a = 1.0), but they died from starvation within 25 days. In 2 separate infection kinetics studies, 14 Brown Norway rats survived only 2 days after their septicemia levels reached 1 × 106 Y. pestis/mL peripheral blood (r = 0.5; unpublished data).
The X. cheopis densities required to sustain plague enzootics and epizootics in a susceptible population were calculated by use of the experimentally derived terms for this equation (table 2). The calculated flea density per host required to maintain Y. pestis infection in a population (R0 = 1) was 4.7, and the flea density required to support an epizootic (R0 2) was 9.4 (figure 2). These are likely to be minimum estimates, because flea survival data were from optimal environmental conditions and saturation host density was assumed. During the last pandemic in India, where X. cheopis is the principal vector, the incidence of rodent and human plague correlated directly with seasonal changes in flea abundance [28, 29]. An X. cheopis index (the average number of fleas in the fur of live-trapped hosts) of 15 was cited as being critical for spread of an epidemic [8, 30]. However, the flea index underestimates the total flea density per host (m), because a majority of fleas may remain in the rodent nest or burrow [8, 14, 29, 31]. Field counts of X. cheopis per rat usually range from 1 to 12, but numbers fluctuate with climactic conditions, and flea indices of 2080 have been recorded for X. cheopis [30, 32]. Our results support models that have postulated that plague epizootics depend on a coincidence of high vector and host densities and that flea control is a sufficient preventative measure and is less likely to engender cases of human plague than is rodent control [32, 33].
Proventricular blockage occurs more frequently and rapidly after an infectious blood meal in X. cheopis than in other flea species, which accounts for the greater vector efficiency of X. cheopis that has been reported in previous comparative studies [814]. However, complete blockage of the proventriculus is not essential for transmission. Partial obstruction and interference with the valvular function of the proventriculus can also result in efficient transmission [34]. For vector species in which complete blockage is less common, such as the rock squirrel flea Diamanus (Oropsylla) montanus in the western United States [9, 13], transmission dynamics may differ in some respects from those for X. cheopis. For example, the life span of infected and infective (transmission-competent) fleas has been reported to be higher in other vector species than in X. cheopis [13, 14]. Because partially blocked fleas can ingest blood, their daily survival probability (q) would be expected to be higher than that of blocked fleas. Thus, the number of times that a partially blocked infective flea feeds during its life span (a/-ln q) would be correspondingly higher than that for a fully blocked infective flea. This could compensate, in part, for a lower proventricular infection rate (b) and a longer extrinsic incubation period (n), compared with those in X. cheopis. Although oral transmission due to cannibalizing of dead nest mates is thought to play a role in sustaining plague in some rodent populations [35], direct transmission of pneumonic plague does not, because infected rodents, even those with lung involvement, do not generate aerosolized respiratory droplets [36, 37].
The vector competence of fleas is low, compared with those characterized for mosquito, sandfly, and tick vectors of other pathogens [3842], which may be a reflection of the recent transition of Y. pestis to arthropodborne transmission. With a longer coevolutionary history, Y. pestis strains better able to colonize fleas and with a lower threshold of infection may evolve. Such strains would be predicted to have a selective advantage, which would be augmented if this advantage was coupled with reduced virulence that prolonged the infectious period of an infected host. The interaction of relatively few flea species with highly clonal Y. pestis strains that spread worldwide during plague pandemics during the past 700 years have been examined, and vector-host cycles with different transmission and disease dynamics may exist in the more ancient plague foci in central Asia [43].
Another example of a correlation between low vector competence and increased virulence is provided by the St. Louis encephalitis (SLE) virus. In the western United States, the primary vector is the most competent one, Culex tarsalis. This mosquito does not occur in the eastern United States, where the primary transmitter is Culex pipiens, a less competent vector with a higher threshold of infection for the virus. Eastern SLE strains produce higher levels of viremia in infected birds and cause more-severe disease in humans than do western strains, which suggests that reliance on the poorer vector has been compensated for by increased virulence [44]. The examples of X. cheopis and C. pipiens indicate that an inefficient vector is not necessarily a more benign one, and this has implications for disease-control strategies that are based on increasing vector resistance to infection.
Acknowledgments
We thank T. Schwan, for helpful discussions; R. Rebeil and D. Erickson, for critical reading of the manuscript; and C. Jarrett, for technical assistance.
References
1. Achtman M, Zurth K, Morelli G, Torrea G, Guiyoule A, Carniel E. Yersinia pestis, the cause of plague, is a recently emerged clone of Yersinia pseudotuberculosis. Proc Natl Acad Sci USA 1999; 96:140438. First citation in article
2. Hinchcliffe SJ, Isherwood KE, Stabler RA, et al. Application of DNA microarrays to study the evolutionary genomics of Yersinia pestis and Yersinia pseudotuberculosis. Genome Res 2003; 13:201829. First citation in article
3. Chain PSG, Carniel E, Larimer FW, et al. Insights into the evolution of Yersinia pestis through whole-genome comparison with Yersinia pseudotuberculosis. Proc Natl Acad Sci USA 2004; 101:1382631. First citation in article
4. Bacot AW, Martin CJ. Observations on the mechanism of the transmission of plague by fleas. J Hygiene 1914; 13(Plague Suppl 3):42339. First citation in article
5. Hinnebusch BJ, Fischer ER, Schwan TG. Evaluation of the role of the Yersinia pestis plasminogen activator and other plasmid-encoded factors in temperature-dependent blockage of the flea. J Infect Dis 1998; 178:140615. First citation in article
6. Jarrett CO, Deak E, Isherwood KE, et al. Transmission of Yersinia pestis from an infectious biofilm in the flea vector. J Infect Dis 2004; 190:78392. First citation in article
7. Perry RD, Fetherston JD. Yersinia pestisetiologic agent of plague. Clin Microbiol Rev 1997; 10:3566. First citation in article
8. Pollitzer R. Plague. Monograph series 22. Geneva: World Health Organization, 1954. First citation in article
9. Burroughs AL. Sylvatic plague studies: the vector efficiency of nine species of fleas compared with Xenopsylla cheopis. J Hyg (Lond) 1947; 45:37196. First citation in article
10. Kartman L, Prince FM. Studies on Pasteurella pestis in fleas. V. The experimental plague-vector efficiency of wild rodent fleas compared with Xenopsylla cheopis, together with observations on the influence of temperature. Am J Trop Med Hyg 1956; 5:105870. First citation in article
11. Kartman L, Prince FM, Quan SF. Studies on Pasteurella pestis in fleas. VII. The plague-vector efficiency of Hystrichopsylla linsdalei compared with Xenopsylla cheopis under experimental conditions. Am J Trop Med Hyg 1958; 7:31722. First citation in article
12. Kartman L, Prince FM, Quan SF. Studies on Pasteurella pestis in fleas: comparative plague-vector efficiency of Xenopsylla vexabilis hawaiiensis and Xenopsylla cheopis. Bull World Health Organ 1956; 14:681704. First citation in article
13. Wheeler CM, Douglas JR. Sylvatic plague studies. V. The determination of vector efficiency. J Infect Dis 1945; 77:112. First citation in article
14. Eskey CR, Haas VH. Plague in the western part of the United States. Public health bulletin 254. Washington, DC: US Public Health Service, 1940. First citation in article
15. Wade SE, Georgi JR. Survival and reproduction of artificially fed cat fleas, Ctenocephalides felis Bouché (Siphonaptera: Pulicidae). J Med Entomol 1988; 25:18690. First citation in article
16. Hinnebusch BJ, Perry RD, Schwan TG. Role of the Yersinia pestis hemin storage (hms) locus in the transmission of plague by fleas. Science 1996; 273:36770. First citation in article
17. Hinnebusch BJ, Rudolph AE, Cherepanov P, Dixon JE, Schwan TG, Forsberg . Role of Yersinia murine toxin in survival of Yersinia pestis in the midgut of the flea vector. Science 2002; 296:7335. First citation in article
18. Douglas JR, Wheeler CM. Sylvatic plague studies. II. The fate of Pasteurella pestis in the flea. J Infect Dis 1943; 72:1830. First citation in article
19. Carniel E. Evolution of pathogenic Yersinia, some lights in the dark. Adv Exp Med Biol 2003; 529:312. First citation in article
20. Brubaker RR, Beesley ED, Surgalla MJ. Pasteurella pestis: role of pesticin I and iron in experimental plague. Science 1965; 149:4224. First citation in article
21. Sodeinde OA, Subrahmanyam YV, Stark K, Quan T, Bao Y, Goguen JD. A surface protease and the invasive character of plague. Science 1992; 258:10047. First citation in article
22. Anderson RM, May RM. Population biology of infectious diseases. New York: Springer Verlag, 1982. First citation in article
23. Ewald PW. Host-parasite relations, vectors, and the evolution of disease severity. Annu Rev Ecol Syst 1983; 14:46585. First citation in article
24. Lenski RE, May RM. The evolution of virulence in parasites and pathogens: reconciliation between two competing hypotheses. J Theor Biol 1994; 169:25365. First citation in article
25. Fine PEM. Epidemiological principles of vector-mediated transmission. In: McKelvey JJ Jr, Eldridge BF, Maramorosch K, eds. Vectors of disease agents. New York: Praeger, 1981:7791. First citation in article
26. Brubaker RR. Yersinia pestis and bubonic plague. In: Dworkin M, Falkow S, Rosenberg E, Schleifer K-H, Stackelbrandt E, eds. The prokaryotes: an evolving electronic resource for the microbiological community. Available at: http://et.springer-ny.com:8080/prokPUB/index.htm. Accessed 27 April 2005. First citation in article
27. Macdonald G. Epidemiologic models in studies of vector-borne diseases. Public Health Rep 1961; 76:75364. First citation in article
28. Martin CJ. Insect porters of bacterial infections. BMJ 1913; 11:5968. First citation in article
29. Hirst LF. The conquest of plague. Oxford: Clarendon Press, 1953. First citation in article
30. Traub R. Notes on fleas and the ecology of plague. J Med Entomol 1972; 9:603. First citation in article
31. Schwan TG. Seasonal abundance of fleas (Siphonaptera) on grassland rodents in Lake Nakuru National Park, Kenya, and potential for plague transmission. Bull Entomol Res 1986; 76:63348. First citation in article
32. Keeling MJ, Gilligan CA. Metapopulation dynamics of bubonic plague. Nature 2000; 407:9036. First citation in article
33. Gratz NG. Problems and developments in the control of flea vectors of disease. In: Traub R, Starcke H, eds. Fleas. Rotterdam: AA Balkema, 1980:21740. First citation in article
34. Bacot AW. Further notes on the mechanism of the transmission of plague by fleas. J Hygiene 1915; 14(Plague Suppl 4):7746. First citation in article
35. Rust JH Jr, Harrison DN, Marshall JD Jr, Cavanaugh DC. Susceptibility of rodents to oral plague infection: a mechanism for the persistence of plague in inter-epidemic periods. J Wildl Dis 1972; 8:12733. First citation in article
36. Meyer KF. Pneumonic plague. Bacteriol Rev 1961; 25:24961. First citation in article
37. Wu L-T. A treatise on pneumonic plague. Geneva: League of Nations Health Organization, 1926. First citation in article
38. Chambers T, Monath TP. The flaviviruses: detection, diagnosis, and vaccine development. In: Maramorosch K, Murphy F, Shatkin AJ, eds. Advances in virus research. Vol 61. San Diego: Elsevier Academic Press, 2003. First citation in article
39. Hardy JL. Susceptibility and resistance of vector mosquitoes. In: Monath TP, ed. The arboviruses: epidemiology and ecology. Boca Raton, FL: CRC Press, 1988:87126. First citation in article
40. Collins WE, Jeffery GM. A retrospective examination of mosquito infection on humans infected with Plasmodium falciparum. Am J Trop Med Hyg 2003; 68:36671. First citation in article
41. Probst RJ, Wellde BT, Lawyer PG, Stiteler JS, Rowton ED. Rhesus monkey model for Leishmania major transmitted by Phlebotomus papatasi sandfly bites. Med Vet Entomol 2001; 15:1221. First citation in article
42. Mather TN, Ginsburg HS. Vector-host-pathogen relationships: transmission dynamics of tick-borne infections. In: Sonenshine DE, Mather TN, eds. Ecological dynamics of tick-borne zoonoses. New York: Oxford University Press, 1994. First citation in article
43. Anisimov AP, Lindler LE, Pier GB. Intraspecific diversity of Yersinia pestis. Clin Microbiol Rev 2004; 17:43464. First citation in article
44. Monath TP, Cropp CB, Bowen GS, Kemp GE, Mitchell CJ, Gardner JJ. Variation in virulence for mice and rhesus monkeys among St. Louis encephalitis virus strains of different origin. Am J Trop Med Hyg 1980; 29:94862. First citation in article