VECTOR/PATHOGEN/HOST INTERACTION,TRANSMISSION Conservation of Transmission Phenotype of marginale (: ) Strains Among and Rhipicephalus Ticks (Acari: Ixodidae)

GLEN A. SCOLES,1 MASSARO W. UETI,2 SUSAN M. NOH, DONALD P. KNOWLES, 2 AND GUY H. PALMER

USDAÐARS, Animal Disease Research Unit, Washington State University, Pullman, WA 99164

J. Med. Entomol. 44(3): 484Ð491 (2007) ABSTRACT Before the eradication of Boophilus ticks from the United States, Rhipicephalus (Boophilus) microplus (Canestrini) and Rhipicephalus (Boophilus) annulatus (Say) were impor- tant biological vectors of the cattle pathogen Anaplasma marginale Theiler. In the absence of Boophilus ticks, A. marginale continues to be transmitted by Dermacentor ticks. However, a few U.S. strains are not transmissible by Stiles, (Say), or both, raising the question of how these strains evolved and how they are maintained. We hypothesize that the U.S. non-Dermacentor-transmissible strains of A. marginale were formerly Boophilus-transmitted strains that have been maintained by a combination of persistent infection and mechanical transmission since the eradication of their biological vector from the United States. To test this hypothesis, we attempted to transmit a well-documented non-Dermacentor- transmissible A. marginale strain (Florida), by using D. andersoni and the two Boophilus that formerly occurred in the United States. For comparison, we examined tick-borne transmission of a strain of A. marginale (Puerto Rico), which has previously been shown to be transmissible by both D. andersoni and B. microplus. All three species of tick transmitted the Puerto Rico strain, and immunohistochemical (IHC) analysis conÞrmed the presence of A. marginale colonies in their salivary glands. All three tick species failed to transmit the Florida strain. Although both D. andersoni and B. microplus acquired transient midgut and salivary gland infections after acquisition feeding, we were unable to detect colonies of the Florida strain in the salivary glands with IHC. This demonstrates that the transmission phenotype of A. marginale strains is conserved among tick species, and it suggests that the failure of the Florida strain to be transmitted by ticks is related to a general inability to efÞciently invade or replicate in tick cells, rather than to a failure to invade or replicate in cells of a speciÞc tick species.

KEY WORDS D. andersoni, Boophilus, tick-borne transmission, immunohistochemistry

Anaplasma marginale Theiler is a tick-borne rickettsial old for microscopic detection in blood smears (French pathogen of cattle that also can be transmitted me- et al. 1998). Persistently infected cattle are clinically chanically by biting ßies and fomites (Ewing 1981, healthy, and they serve as reservoirs of infection for Potgieter et al. 1981, Hawkins et al. 1982). After trans- competent tick vectors. mission to cattle, A. marginale undergoes sequential Strains of A. marginale that are not transmissible by cycles of invasion, replication, and release from eryth- Dermacentor sp. ticks have been reported previously rocytes. During acute infection, cell-associated rick- (Smith et al. 1986, Wickwire et al. 1987, de la Fuente ettsemia may reach 109 infected erythrocytes (IE) per et al. 2001b, Scoles et al. 2006). The origin of these milliliter of blood. This results in clinical strains and mechanisms for their maintenance have that is characterized by anemia, weight loss, and abor- not been well investigated; however, it seems that tion, and it may result in death. Animals that survive levels of infection in persistently infected cattle may acute infection maintain a life-long persistent infec- be too low for efÞcient mechanical transmission by tion characterized by repeated cycles of rickettsemia biting ßies (Scoles et al. 2005a). ranging from 102.5 to 107 IE per ml, below the thresh- Although current data on the prevalence of A. mar- ginale infection in U.S. cattle are available for only a limited number of areas, anaplasmosis has historically 1 Corresponding author, e-mail: [email protected]. been most prevalent in two broad regions of the coun- 2 Program in Vector-Borne Disease, Department of Veterinary Mi- crobiology and Pathology, Washington State University, Pullman, WA try, the southeastern and the western United States 99164-7040. (Saulmon 1962). In the west (including the northwest May 2007 SCOLES ET AL.: CONSERVATION OF A. marginale TRANSMISSION PHENOTYPE 485 and PaciÞc Coast states), Dermacentor andersoni Stiles Materials and Methods and Dermacentor occidentalis Marx have been consid- A. marginale Strains. The Florida strain of A. mar- ered the primary vectors. A nontick-transmissible ginale used in this study was originally isolated by D. A. strain of A. marginale from California has been re- Sanders in 1955 from naturally infected cattle in Flor- ported, although the tick species tested was not spec- ida (Ristic and Carson 1977). This strain has been iÞed (de la Fuente et al. 2003). California is the only shown to be nontransmissible by D. andersoni in sev- state in this region to have once had a population (now eral studies (for examples, see Friedhoff and Ristic eradicated) of Boophilus annulatus (Say) (Cooley 1966, Wickwire et al. 1987). 1946). In the southeastern United States, the primary The Puerto Rico strain of A. marginale used in this vectors of A. marginale were B. annulatus and to a study was originally received as a frozen stabilate from lesser extent, Boophilus microplus (Canestrini), until K. Kuttler in 1985 and had one calf passage in our these species were eliminated from the United States laboratory before use in this study. This strain has in the early 1940s (Bram et al. 2002, George et al. 2002). previously been shown to be transmissible by our Since the removal of Boophilus species from the laboratory colonies of D. andersoni and B. microplus United States, A. marginale has continued to be trans- (Futse et al. 2003). mitted in the southeast by Dermacentor variabilis Tick Colonies. The D. andersoni Reynolds Creek (Say) and presumably by ßy-borne mechanical trans- strain used in these studies was a laboratory colony mission. Several strains of A. marginale that are not established from ticks originally collected in Owyhee transmissible by Dermacentor sp. have been collected Co., in southwestern Idaho. These ticks have been Ͼ from this region. These strains include the well char- maintained in colony for 15 yr at the University of acterized Florida, Illinois, and Okeechobee strains Idaho Holm Research Center (HRC), and they are competent vectors of the Puerto Rico strain as well as (Smith et al. 1986, Wickwire et al. 1987, de la Fuente several other A. marginale strains (Futse et al. 2003). et al. 2001b). These non-Dermacentor tick-transmis- “ The B. microplus and B. annulatus used in these sible isolates have been tested only for transmissibil- ” studies originated from colonies maintained at the ity by Dermacentor sp.; the ability of these strains to be USDA cattle fever tick research laboratory in Mission, transmitted by Boophilus ticks has not been tested TX. The B. microplus La Minita strain came from ticks previously. collected in Starr Co., TX, in 1996. This strain has been Strains of A. marginale from regions where they routinely maintained at HRC since 1999 and at the would be naturally transmitted by B. microplus form a time of these studies was in approximately the 20th distinct, genetically deÞned clade based on the se- laboratory generation; ticks of this strain have been quence of the msp4 , whereas most temperate shown to be competent vectors of Puerto Rico (Futse region strains that would be biologically transmitted et al. 2003) as well as other strains of A. marginale. The by Dermacentor sp. or mechanically transmitted by B. annulatus El Indio strain came from ticks collected biting ßies fall into a discretely different clade (de la in Maverick Co., TX, in 2003; larvae were shipped to Fuente et al. 2002, Futse et al. 2003). With the excep- HRC in April 2004 and reared for one generation tion of the Illinois strain, non-Dermacentor tick-trans- before use in this study, at the time of this study they missible strains (Okeechobee, FL, and Mississippi) were in the fourth laboratory generation. The vector cluster together with the B. microplus transmitted competence of this B. annulatus El Indio strain for A. clade (de la Fuente et al. 2002). The msp4 sequence marginale has not previously been examined. of the Puerto Rico strain of A. marginale places it in this Cattle. Holstein calves 5Ð6 mo old were used in clade along with strains from regions where B. micro- these experiments. All cattle were cared for in facil- plus is found (Futse et al. 2003). ities located at the HRC following procedures ap- We hypothesize that the non-Dermacentor trans- proved by the University of Idaho Institutional Animal missible strains of A. marginale collected from the Care and Use Committee. All calves were PCR neg- southeastern United States are strains that were for- ative for A. marginale (see below for details of the test), and before use in the study, they were conÞrmed merly Boophilus-transmitted and have been main- to be seronegative for A. marginale by using compet- tained, since the elimination of Boophilus sp. from the itive enzyme-linked immunosorbent assay (cELISA) United States, by a combination of persistent infec- (VMRD, Pullman, WA). tion, iatrogenic transmission, and ßy-borne mechani- Both of the calves used for tick acquisition feeding cal transmission. To test our hypothesis, we have used had been infected by intravenous inoculation with B. microplus, B. annulatus, and D. andersoni and at- frozen A. marginale stabilates. Calf c1031bl was in- tempted to transmit the Florida strain, a well-docu- fected with the Puerto Rico strain on 31 April 2004, mented non-Dermacentor tick-transmissible strain of and it was in the persistent phase of infection, below A. marginale. For comparison, we have used the microscopic detection, at the time of tick acquisition Puerto Rico strain that has previously been shown to feeding; calf c1049bl was infected with the Florida be transmissible by D. andersoni and B. microplus and strain on 24 September, 2004 and was postacute but that falls within a clade of strains collected from re- not yet below microscopic detection at the time of gions where B. microplus is the primary vector of A. acquisition feeding. Infection levels of the cattle at the marginale (Futse et al. 2003). time of tick feeding were established using quantita- 486 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 44, no. 3

Table 1. Attempted transmission of the Puerto Rico and Florida strains of A. marginale with B. microplus, B. annulatus, and D. andersoni

Acquisition A. marginale Vector Transmission No. ticks c Prepatent Peak Transmission d e host strain species host Applieda Recoveredb period PPEs c1031b1f PRAM B. microplus c1060bl 120 80 Yes 18 7.6 (44) B. annulatus c1059b1 66 36 Yes 18 0.9 (46) D. andersoni c1057bl 59 53 Yes 18 4.9 (44) c1049b1g FAM B. microplus c1048bl 209 68 No B. annulatus c1051bl 42 12 No D. andersoni c1061bl 109 105 No

PRAM, Puerto Rico strain; FAM, Florida strain. a Number of acquisition fed ticks applied for transmission. b Number of ticks recovered after transmission feeding. c Did transmission occur? d Number of days before Þrst microscopic detection on stained blood smear. e Peak percentage of parasitized erythrocytes (days to peak). f Level of infection with the Puerto Rico strain at acquisition feeding 5.55 ϫ 106 A. marginale per ml. g Level of infection with the Florida strain at acquisition feeding 7.05 ϫ 107 A. marginale per ml. tive polymerase chain reaction (qPCR) (described level of infection to which the acquisition-fed ticks under DNA Preparation and nPCR and qPCR). were exposed. Male ticks of each of the three species Tick Rearing and Transmission Trial. Tick acqui- were acquisition fed in separate feeding patches on sition and transmission feedings were conducted using each of the A. marginale-infected calves (Puerto Rico the intrastadial transmission model: acquisition and and Florida). After7dofacquisition feeding, ticks transmission by adult male ticks. Intrastadial transmis- were removed from the host and held for 3 d. Then, sion by adult male D. andersoni is the most efÞcient each group of ticks was placed on a separate trans- mode of transmission, and B. microplus also has been mission host for 7 d (1Ð8 November). The numbers of shown to efÞciently transmit both the Puerto Rico and each species of ticks applied and recovered from the the St. Maries strains of A. marginale by this route transmission feedings is provided in Table 1. Trans- (Kocan et al. 1993, Futse et al. 2003). D. andersoni were mission fed calves were monitored for development of fed as nymphs on a rabbit; fully engorged nymphs A. marginale infection by microscopic examination of dropped off on 17 September 2004, and they were held Giemsa-stained blood smears, by nested PCR (nPCR) at 25ЊC and 92% RH for molting and until adults were and by cELISA (VMRD Inc.). used for the acquisition feeding. A cohort of B. microplus and D. andersoni were Adult male Boophilus sp. ticks can only survive off dissected 3 d after acquisition feeding and another the host for 4Ð6 d. Because of this, B. microplus and B. cohort 24 h after transmission feeding. The number of annulatus males were reared just before use in the ticks dissected is provided in Table 2. All B. annulatus transmission trial. Larvae of each species were placed ticks were used for transmission and for immunohis- on cattle on 4Ð6 October 2004 and allowed to feed to tochemistry (IHC). Ticks were dissected in HanksÕ fully engorged nymphs. Engorged nymphs were re- balanced salt solution (Sigma-Aldrich, St. Louis, MO). moved from the cattle 19Ð20 October (13Ð16 d) and Each tick was dissected on a fresh piece of dental wax maintained in an incubator at 25ЊC and 98% RH for with a new razor blade and clean forceps. Forceps molting to the adult stage. Adult males were sorted and were thoroughly cleaned between ticks by sonication used for acquisition within 2Ð4 d of molting. in 5% sodium dodecyl sulfate (SDS) for 5 min, fol-

Ticks were acquisition fed for 7 d (from 22 to 29 lowed by two rinses with sterile distilled H2O, and October 2004). Blood was collected from each calf at then dipping and ßaming in 95% ethanol. Midguts and the end of the feeding (29 October) to determine the salivary glands were dissected from each tick, rinsed

Table 2. Detection of the Florida and Puerto Rico strains of A. marginale in tick midgut and salivary glands

A. marginale Acquisition feed Transmission feed Tick species strain n Gutsa SGb n Gutsa SGb Puerto Rico B. microplus 20 20 (100) 20 (100) 67 67 (100) 67 (100) D. andersoni 20 18 (90) 19 (95) 37 36 (97) 36 (97) Florida B. microplus 40 4 (10) 23c (57) 58 1 (1.7) 1 (1.7) D. andersoni 39 7 (18) 10d (26) 80 0 (0) 1 (1.2)

The n value is number of ticks dissected. a Number with nPCR-positive guts (%). b Number with nPCR-positive salivary glands (SG) (%). c Only three ticks were also gut positive, total of 24 (60%) ticks positive in either the gut, salivary gland or both. d Only one tick was also gut positive, total of 16 (41%) ticks positive in either the gut, salivary gland, or both. May 2007 SCOLES ET AL.: CONSERVATION OF A. marginale TRANSMISSION PHENOTYPE 487 in a fresh drop of sterile HanksÕ balanced salt solution, samples predicted to be above 106 (i.e., most blood placed separately in tubes with 100 ␮l of proteinase K samples) were diluted 1:10 and 1:100 so that the results buffer with 2ϫ enzyme (10 mM Tris, pH 7.8, 5 mM would fall within the range of the standard curve. EDTA, 0.5% SDS, and 100 ␮g/ml proteinase K), and Immunohistochemistry. The numbers of ticks pre- frozen at Ϫ20ЊC before DNA isolation. pared and sectioned for IHC are provided in Table 1. Cohorts of B. microplus and D. andersoni (10 ticks Ticks were prepared as described previously (Futse et of each species) were prepared for IHC after acqui- al. 2003, Ueti et al. 2003). Brießy, after acquisition or sition feeding, and a cohort of each of the three species transmission feeding, the ticks were promptly Þxed in of ticks (10 ticks each for B. microplus and D. andersoni 10% formaldehyde for eight to18 h and embedded and 12 ticks for B. annulatus) were prepared for IHC in parafÞn. Repeated tick sections of 4 ␮m were after transmission feeding. deparaÞnized in Clear-Rite (Richard-Allan ScientiÞc, DNA Preparation and nPCR and qPCR. Dissected Kalamazoo, MI) and hydrated in an ethanol gradient. tick tissues (gut and salivary gland processed sepa- For antigen retrieval, the sections were treated with rately) were incubated in the proteinase K solution at Zymed citrate solution (Dako North America, Inc., 50ЊC for 1 h. IsoQuick (Orca Research, Bothel, WA) Carpinteria, CA) in a steam bath for 30 min, and then lysis solution (100 ␮l) was added, and the tissues were the sections were blocked with serum-free protein ground with a disposable plastic pestle (Bel-Art Prod- blocker (Dako North America, Inc.) for 25 min and ucts, Pequannock, NJ). After grinding, tubes were treated with 3% hydrogen peroxide in distilled water incubated for an additional hour at 50ЊC. DNA was for 15 min. The sections were stained for 25 min with extracted from the lysate by using the IsoQuick DNA 0.1 ␮g/ml of either speciÞc monoclonal antibody 115/ extraction kit (Orca Research) according to a modi- 152.20.19 against a conserved region of A. marginale Þed protocol (Schwartz et al. 1997) with the following MSP-3 or as a negative control with an isotype- additional modiÞcation: before the DNA precipitation matched monoclonal antibody against Trypanosoma step, 1.0 ␮l of a 20 mg/ml solution of glycogen (Roche brucei. The MSP-3 antibody was conÞrmed to work Diagnostics, Indianapolis, IN) was added to the tick equally well for both the Florida and the Puerto Rico salivary gland samples to improve the precipitation of strains of A. marginale. The secondary antibody, horse- small quantities of DNA. DNA was resuspended in 50 radish peroxidase-labeled anti-mouse (Dako North ␮ Њ l of double distilled H2O and stored at 4 C. America, Inc.), was applied to the sections and incu- Tick tissues were tested for A. marginale by using a bated for 25 min. Tick sections were then incubated nPCR that targets the single copy A. marginale msp5 for 5 min in 3-amino-9-ethylcarbazole containing hy- gene (Scoles et al. 2005b). Based on repeated nPCR drogen peroxide (Dako North America, Inc.). All sec- ampliÞcation of a serially diluted cloned msp5 frag- tions were counterstained with MayerÕs hematoxylin ment (Scoles et al. 2005b), we determined that the for 2 min and cover slipped by using an aqueous sensitivity threshold for this nPCR is Ͻ10 A. marginale mounting medium. Ticks fed on uninfected calves and organisms. As a control for the quality of the DNA processed in the same manner were examined as ad- preparation all samples that were nPCR negative for ditional negative controls. Organisms of both A. mar- A. marginale were tested by conventional PCR for the ginale strains were isolated from infected erythro- presence of tick DNA by using mitochondrial 16s cytes, embedded in 2% agarose gel, and Þxed in 10% rDNA primers 16sϩ1 and 16sϪ1 as described previ- formaldehyde as positive control for IHC. ously (Norris et al. 1996). Blood samples from the infected calves used for Results acquisition feeding, and a sample of nPCR-positive salivary glands and guts was tested by qPCR by using The Puerto Rico strain of A. marginale was trans- a TaqMan protocol described previously (Futse et al. mitted by all three of the tick species (D. andersoni, B. 2003), with modiÞcations to the primers and probes as microplus, and B. annulatus), whereas transmission of described below to adapt the assay to work with both the Florida strain failed for all three species (Table 1). the Florida and Puerto Rico strains. For the modiÞed The prepatent period (i.e., the number of days before TaqMan assay the forward primer was 5Ј-CTT CCG infected erythrocytes were Þrst detected microscop- AAG TTG TAA GTG AGG GCA-3Ј, and the reverse ically) for all three Puerto Rico strain transmission primer was 5Ј-CTT ATC GGC ATG GTC GCC TAG calves was 18 d; the peak percentage of parasitized TTT-3Ј; the TaqMan probe sequence was 5Ј-GCC erythrocytes (PPEs) ranged from 0.9 to 7.6% (Ϸ107.8Ð TCC GCG TCT TTC AAC AAT TTG GT-3Ј. The 108.7) and was reached in 44Ð46 d (Table 1). All three annealing temperature for this modiÞed assay was Florida strain transmission calves were conÞrmed to 55ЊC (as opposed to 50ЊC for the original assay). Sam- be free of A. marginale by repeated testing with nPCR ples were run in triplicate and expressed as the mean and serology until 100 d after the tick transmission of the three replicates. The detection threshold for feeding. As further conÞrmation of their infection free this qPCR assay is between 10 and 100 A. marginale status, the three calves were splenectomized, and organisms (i.e., less sensitive than nPCR but more there was no evidence of recrudescence of a subclin- sensitive than IHC). The standard curve was con- ical infection. All three calves were then challenged by structed using dilutions of a cloned msp5 fragment intravenous inoculation with stabilates of Florida from 106 to 103 copies; concentration of samples with strain A. marginale to conÞrm their susceptibility. Al- Ͻ103 copies per ml were determined by extrapolation; though none of these calves were infected with the 488 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 44, no. 3

Fig. 1. Colonization of the Puerto Rico strain but not the Florida strain of A. marginale in the salivary glands of three tick species: D. andersoni, Rhipicephalus (B.) microplus, and R.(B.) annulatus. Sections of whole ticks were probed by immu- nohistochemistry by using an anti-A. marginale MSP3 monoclonal antibody (mAb) (115/152.20.19), and, as a negative control, an isotype-matched control mAb speciÞc for T. brucei (Tryp1E1). Arrows indicate the presence of A. marginale colonies within the salivary gland acini.

Florida strain by tick feeding, all were susceptible to Puerto Rico strain-infected calf also had a high infec- infection with this strain, because they all developed tion rate (Ͼ90%) in both the gut and the salivary acute infection after intravenous challenge. glands, both after acquisition feeding and after trans- At the time of acquisition feeding the Florida strain mission feeding (Table 2). Similar to B. microplus, the infected calf had a bacteremia of 7.05 ϫ 107 A. mar- number of A. marginale per salivary gland pair in- ginale per ml, Ͼ12-fold higher than the level of infec- creased by nearly 1.5 logs in the D. andersoni salivary tion of the Puerto Rico strain-infected calf, which had glands between acquisition and transmission, from 5.55 ϫ 106 organisms per ml at the same time point. 8.02 ϫ 103 to 2.46 ϫ 105, again indicative of efÞcient Nested PCR results on dissected guts, and salivary replication of the Puerto Rico strain in the salivary glands of D. andersoni and B. microplus are presented glands. This was evident when sections were exam- in Table 2. There were insufÞcient numbers of B. ined by IHC (Fig. 1); large colonies of A. marginale annulatus for dissection and all B. annulatus ticks were were detected in the salivary glands of each of the used for transmission and IHC. All B. microplus that three vector species. were fed on the Puerto Rico strain-infected calf were In contrast to the Puerto Rico strain, a much lower nPCR positive in both the gut and the salivary glands proportion of ticks that were fed on the Florida strain- at 3 d after acquisition feeding and immediately after infected calf were nPCR positive in the gut and sali- transmission feeding. There were 8.01 ϫ 102 A. mar- vary glands 3 d after acquisition feeding. In total, 24/40 ginale organisms per salivary gland pair 3 d after ac- (60%) B. microplus and 16/39 (41%) D. andersoni had quisition feeding, and this number increased by more nPCR-positive guts and/or salivary glands 3 d after than 2 logs to 9.65 ϫ 104 organisms per salivary gland being removed from the Florida-infected acquisition pair after transmission feeding, indicating that the host. In addition, a larger proportion of ticks had Puerto Rico strain had invaded and was replicating in nPCR-positive salivary glands than guts, suggesting the salivary glands. D. andersoni that were fed on the that midgut infection is transient (Table 2). After May 2007 SCOLES ET AL.: CONSERVATION OF A. marginale TRANSMISSION PHENOTYPE 489 transmission feeding, the number of ticks with positive has previously been shown to be transmissible by both salivary glands was markedly decreased (Table 2). Not D. andersoni and by B. microplus (Futse et al. 2003). withstanding their failure to transmit, a very small Although both B. microplus and B. annulatus were proportion of Florida-exposed B. microplus (1/58) and present in the United States before eradication, B. D. andersoni (1/80) had nPCR-positive salivary glands annulatus was much more widespread than was B. after transmission feeding. All salivary glands of ticks microplus, having been collected from parts of 16 fed on the Florida-infected acquisition host were be- southern and western states, whereas B. microplus was low levels detectable by qPCR (i.e., Ͻ100 organisms). probably only present in the southernmost parts of Consistent with the low percentage of infected sali- Florida and Texas (Cooley 1946). vary glands and the lack of replication in the glands, The capacity of an A. marginale strain to be biolog- colonies of the Florida strain could not be detected by ically transmitted by ticks depends on its ability to IHC in salivary glands of any of the three tick species invade, survive, replicate, and escape Þrst from the (Fig. 1). midgut epithelium and then to invade and replicate in the salivary gland acini and be shed in the saliva; infection of the midgut epithelium is a prerequisite to Discussion infection of the salivary glands (Kocan 1986, Kocan et In this study, we tested the hypothesis that a non- al. 1993). Although we were unable to show the pres- Dermacentor tick-transmissible strain of A. marginale, ence of colonies of the Florida strain of A. marginale isolated from a region formerly endemic for Boophilus in the midgut epithelium by using qPCR or IHC, the ticks, would be transmissible by Boophilus ticks. Al- detection of the Florida strain by nPCR in the salivary though the failure of the Florida strain to be trans- glands of ticks 3 d after acquisition feeding provides de mitted by D. andersoni has been well documented facto evidence of infection of the midgut. Contami- (Friedhoff and Ristic 1966, Wickwire et al. 1987), the nation of the salivary gland with gut material is an transmissibility of this strain by Boophilus ticks has not unlikely explanation for the high level of nPCR-pos- been reported. Before their elimination from the itive salivary glands, because only a small proportion United States in 1943 (Bram et al. 2002, George et al. of the guts were nPCR positive, and there were many 2002), B. annulatus and B. microplus would have been ticks with nPCR-positive salivary glands and negative the primary natural vectors of A. marginale in Florida guts (Table 2). Furthermore, previous studies have and elsewhere in the southern United States. We have shown that DNA of the Florida strain is not detectable suggested that non-Dermacentor tick-transmissible in the gut lumen of exposed ticks beyond the second strains of A. marginale may be Boophilus-transmitted day postdetachment (Stiller et al. 1989). These data strains that have been maintained by mechanical suggest a transient infection of the gut in which the transmission in the absence of their natural vectors organism is able to invade and survive in the epithelial since the elimination of Boophilus from the United cells, but it is unable to replicate to high levels. Salivary States. The data we present here do not support this glands that are nPCR positive in the absence of trans- hypothesis. Rather, the tick transmissibility pheno- mission suggest that the organism is either adhering to types of the Florida and Puerto Rico strains of A. the exterior of the salivary gland basil lamina and marginale seem to be conserved across tick species. unable to invade or replicate, or if it is able to cross the We suggest that the failure of the Florida strain to be basil lamina, it is not able to invade and/or replicate in transmitted by ticks is related to a broadly based fail- salivary gland acini, as would be required for trans- ure of this strain to efÞciently invade, survive, or rep- mission. Detection of the Florida strain in the salivary licate in tick cells, rather than to a failure to invade or glands by nPCR (sensitivity Ͻ10 organisms) in the replicate in cells of speciÞc tick species. absence of detection by qPCR (sensitivity Ͼ100 or- In this experiment, the failure of the Florida strain ganisms) and IHC suggests that the organism is to invade or replicate in Boophilus ticks cannot be present in low levels, adhering to or invading the attributed to dose. The bacteremia level of the Florida salivary gland tissue, but not replicating. Thus, it seems strain-infected calf was more than a log higher than more likely that the defect in the Florida strain that the level of the Puerto Rico strain-infected calf at the prevents it from being transmitted by ticks is related time of acquisition feeding, a situation that should to an inability to replicate in tick cells rather than a skew the probability of transmission in favor if the failure to invade tick midgut epithelial cells as has Florida strain. Nonetheless, the Puerto Rico strain was been proposed previously (de la Fuente et al. 2001a, transmitted, whereas the Florida strain was not. 2001b, 2003). The Florida and Puerto Rico A. marginale strains An earlier preliminary study (S.M.N., unpublished were chosen for this study because both originated observations) used IHC to demonstrate colonies of from areas where Boophilus ticks have historically the Florida strain in the midgut epithelium of D. ander- been present. Both B. microplus and B. annulatus occur soni; however, we have not been able to replicate that in Puerto Rico, both historically and currently observation here. The sensitivity of IHC is consider- (Cooley 1946, Guglielmone et al. 2003). The Florida ably lower than nPCR. We suggest that low sensitivity, strain was originally isolated in southern Florida coupled with the low level of replication and the where both B. microplus and B. annulatus would have transient nature of the infection in the gut epithelial been found before the eradication program (Cooley cells, works against detection of nonproductive infec- 1946, Guglielmone et al. 2003). The Puerto Rico strain tions by using IHC. 490 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 44, no. 3

In this study, the Boophilus ticks had lower feeding proteins 1a and 1b of the ehrlichial cattle pathogen success than did D. andersoni (Table 1). Boophilus Anaplasma marginale to bovine erythrocytes and tick species are adapted to molting on the host; molting off cells. Int. J. Parasitol. 31: 145Ð153. of the host, as in our experimental design, causes some de la Fuente, J., J. C. Garcia-Garcia, E. F. Blouin, B. R. stress that may lead to higher levels of mortality. How- McEwen, D. Clawson, and K. M. Kocan. 2001b. Major ever, D. andersoni is a robust tick that is well adapted surface protein 1a effects tick infection and transmission to molting off the host. In spite of the lower feeding of Anaplasma marginale. Int. J. Parasitol. 31: 1705Ð1714. de la Fuente, J., R.A.V.D. Bussche, J. C. Garcia-Garcia, S. D. success, Boophilus species were still efÞcient vectors Rodrı´guez, M. A. Garcı´a, A. A. Guglielmone, A. J. of the Puerto Rico strain of A. marginale, and in a Mangold, L.M.F. Passos, M.F.B. Ribeiro, E. F. Blouin, et previous study, B. microplus handled in the same way al. 2002. Phylogeography of New World isolates of efÞciently transmitted both the Puerto Rico and St. Anaplasma marginale based on major surface protein se- Maries strains of A. marginale (Futse et al. 2003). Only quences. Vet. Microbiol. 88: 275Ð285. 12 of the B. annulatus that had been acquisition fed on de la Fuente, J., J. C. Garcia-Garcia, E. F. Blouin, and K. M. the Florida strain-infected calf survived the transmis- Kocan. 2003. Characterization of the functional domain sion feeding. These ticks failed to transmit, and ex- of major surface protein 1a involved in adhesion of the amination by IHC did not reveal Florida strain A. Anaplasma marginale to host cells. Vet. Micro- marginale colonies within the salivary glands. Twelve biol. 91: 265Ð283. ticks may be too few to be considered a deÞnitive Ewing, S. A. 1981. Transmission of Anaplasma marginale by demonstration of the inability of this species to trans- arthropods, pp. 425Ð434. In R. J. Hidalgo and W. E. Jones [eds.], Proceedings of the Seventh National Anaplasmo- mit the Florida strain; however, for an efÞcient vector sis Conference, 21Ð23 October 1981, Starkville, MS. Mis- very few ticks will be required for transmission. We sissippi State University, Mississippi State, MS. have previously reported the transmission of the St. French, D. M., T. F. McElwain, T. C. McGuire, and G. H. Maries strain of A. marginale by using only three D. Palmer. 1998. Expression of Anaplasma marginale major andersoni males (Scoles et al. 2005a). surface protein 2 variants during persistent cyclic rick- Although the data presented here do not support ettsemia. Infect. Immun. 66: 1200Ð1207. our hypothesis that non-Dermacentor-transmissible Friedhoff, K. T., and M. Ristic. 1966. Anaplasmosis. XIX. A strains of A. marginale are strains that were transmit- preliminary study of Anaplasma marginale in Dermacen- ted by Boophilus spp. before their elimination from the tor andersoni (Stiles) by ßuorescent antibody technique. United States, these data do show that the transmission Am. J. Vet. Res. 27: 643Ð646. phenotype of the Florida strain of A. marginale is Futse, J. E., M. W. Ueti, D. P. Knowles, Jr., and G. H. Palmer. conserved across three major tick vector species. In 2003. Transmission of Anaplasma marginale by Boophilus microplus: retention of vector competence in the absence addition, these data suggest that the defect in trans- of vector-pathogen interaction. J. Clin. Microbiol. 41: mission of the nontick-transmissible Florida strain 3829Ð3834. may be in its ability to replicate in tick cells rather than George, J. E., R. B. Davey, and J. M. Pound. 2002. Intro- in its ability to invade them. DeÞning the barriers to duced ticks and tick-borne diseases: the threat and ap- tick-borne transmission of nontick-transmissible proaches to eradication. Vet. Clin. North Am. Food Anim. strains of A. marginale, like the Florida strain, may lead Pract. 18: 401Ð416. to identiÞcation of transmission-blocking targets that Guglielmone, A. A., A. Estrada-Pena, J. E. Keirans, and R. G. could become important components of strategies for Robbins. 2003. Ticks (Acari: Ixodidae) of the Neotropi- interrupting tick-borne transmission of A. marginale. cal zoogeographic region. The International Consortium on Ticks and Tick-borne Disease, Houten, The Nether- lands. Acknowledgments Hawkins, J. A., J. N. Love, and R. J. Hidalgo. 1982. Mechan- ical transmission of anaplasmosis by tabanids (Diptera: We thank Sara Davis and Ralph Horn for superior tech- Tabanidae). Am. J. Vet. Res. 43: 732Ð734. nical assistance. James Allison assisted with animal handling. Kocan, K. M. 1986. Chapter 22: Development of Anaplasma We also thank Ron Davey for supplying larvae of B. microplus marginale Theiler in ixodid ticks: coordinated develop- and B. annulatus to start colonies. This work was supported ment of a rickettsial organism and its tick host, pp. 472Ð by USDAÐARS CRIS 5348-32000-016-00D. S.M.N. was sup- 505. In J. R. Sauer and J. A. Hair [eds.], Morphology, ported by National Institutes of Health grant K0Ð8AI052412, Physiology, and Behaviorial Biology of Ticks. Ellis Hor- and M.W.U. was supported by National Institutes of Health wood Limited, Chinchester, England. training grant T32 AI007025. Kocan, K. M., W. L. Goff, D. Stiller, W. Edwards, S. A. Ewing, P. L. Claypool, T. C. McGuire, J. A. Hair, and S. J. Barron. 1993. Development of Anaplasma marginale in salivary References Cited glands of male Dermacentor andersoni. Am. J. Vet. Res. 54: Bram, R. A., J. E. George, R. E. Reichard, and W. J. Tabach- 107Ð112. nick. 2002. Threat of foreign arthropod-borne patho- Norris, D. E., J. S. Klompen, J. E. Keirans, and W.C.T. Black. gens to livestock in the United States. J. Med. Entomol. 39: 1996. Population genetics of scapularis (Acari: Ix- 405Ð416. odidae) based on mitochondrial 16S and 12S . Cooley, R. A. 1946. The genera Boophilus, Rhipicephalus, J. Med. Entomol. 33: 78Ð89. and Haemaphysalis (Ixodidae) of the New World. Natl. Potgieter, F. T., B. Sutherland, and H. C. Biggs. 1981. At- Inst. Health Bull. No. 187: 1Ð54. tempts to transmit Anaplasma marginale with Hippobosca de la Fuente, J., J. C. Garcia-Garcia, E. F. Blouin, and K. M. rufipes and Stomoxys calcitrans. Onderstepoort J. Vet. Res. Kocan. 2001a. Differential adhesion of major surface 48: 119Ð122. May 2007 SCOLES ET AL.: CONSERVATION OF A. marginale TRANSMISSION PHENOTYPE 491

Ristic, M., and C. A. Carson. 1977. Methods of immunopro- of Anaplasma marginale (Rickettsiales: Anaplasmata- phylaxis against bovine anaplasmosis with emphasis on ceae) is not transmissible by Dermacentor andersoni use of the attenuated Anaplasma marginale vaccine, pp. (Acari: Ixodidae), whereas ticks from two Canadian D. 151Ð188. In L. H. Miller, J. A. Pino, and J. John J. McKelvey andersoni populations are competent vectors of a U.S. [eds.], Immunity to blood parasites of animals and man. strain. J. Med. Entomol. 43: 971Ð975. Plenum, New York. Smith, R. D., M. G. Levy, M. S. Kuhlenschmidt, J. H. Adams, Saulmon, E. E. 1962. Changes in the anaplasmosis map, pp. D. L. Rzechula, T. A. Hardt, and K. M. Kocan. 1986. 2Ð3. In Proceedings of the Fourth National Anaplasmosis Isolate of Anaplasma marginale not transmitted by ticks. Conference, 26Ð27 April 1962, Reno, NV. Am. J. Vet. Res. 47: 127Ð129. Schwartz, I., S. Varde, R. B. Nadelman, G. P. Wormser, and Stiller, D., W. L. Goff, S. Landry, L. W. Johnson, and J. D. Fish. 1997. Inhibition of efÞcient polymerase chain Gorham. 1989. Persistence of Anaplasma marginale in reaction ampliÞcation of DNA in ticks: considerations in determining the tick infection blood-fed ticks. Am. J. Trop. Med. Hyg. 56: 339Ð342. rate. pp. 177Ð181. In Proceedings of the Eighth National Scoles, G. A., A. B. Broce, T. J. Lysyk, and G. H. Palmer. Veterinary Hemoparasitic Disease Conference, 10Ð12 2005a. Relative efÞciency of biological transmission of April 1989, St. Louis, MO. Anaplasma marginale (Rickettsiales: Anaplasmataceae) Ueti, M. W., G. H. Palmer, L. S. Kappmeyer, G. A. Scoles, and by Dermacentor andersoni (Acari: Ixodidae) compared to D. P. Knowles. 2003. Expression of equi merozoite an- mechanical transmission by Stomoxys calcitrans (Diptera: tigen 2 during development of Babesia equi in the midgut Muscidae). J. Med. Entomol. 42: 668Ð675. and salivary gland of the vector tick Boophilus microplus. Scoles, G. A., M. W. Ueti, and G. H. Palmer. 2005b. Variation J. Clin. Microbiol. 41: 5803Ð5809. among geographically separated populations of Derma- Wickwire, K. B., K. M. Kocan, S. J. Barron, S. A. Ewing, R. D. centor andersoni (Acari: Ixodidae) in midgut susceptibil- Smith, and J. A. Hair. 1987. Infectivity of three Anaplasma ity to Anaplasma marginale (Rickettsiales: Anaplasmata- marginale isolates for Dermacentor andersoni. Am. J. Vet. ceae). J. Med. Entomol. 42: 153Ð162. Res. 48: 96Ð99. Scoles, G. A., T. F. McElwain, F. R. Rurangirwa, D. P. Knowles, and T. J. Lysyk. 2006. A Canadian bison isolate Received 24 October 2006; accepted 18 January 2007.