key: cord-0807650-w1nxt8iy
authors: Drexler, Jan Felix; Corman, Victor Max; Wegner, Tom; Tateno, Adriana Fumie; Zerbinati, Rodrigo Melim; Gloza-Rausch, Florian; Seebens, Antje; Müller, Marcel A.; Drosten, Christian
title: Amplification of Emerging Viruses in a Bat Colony
date: 2011-03-03
journal: Emerg Infect Dis
DOI: 10.3201/eid1703.100526
sha: 4a25649bbc72d2ef0e57a19221b21c486271bcfd
doc_id: 807650
cord_uid: w1nxt8iy

Bats host noteworthy viral pathogens, including coronaviruses, astroviruses, and adenoviruses. Knowledge on the ecology of reservoir-borne viruses is critical for preventive approaches against zoonotic epidemics. We studied a maternity colony of Myotis myotis bats in the attic of a private house in a suburban neighborhood in Rhineland-Palatinate, Germany, during 2008, 2009, and 2010. One coronavirus, 6 astroviruses, and 1 novel adenovirus were identified and monitored quantitatively. Strong and specific amplification of RNA viruses, but not of DNA viruses, occurred during colony formation and after parturition. The breeding success of the colony was significantly better in 2010 than in 2008, in spite of stronger amplification of coronaviruses and astroviruses in 2010, suggesting that these viruses had little pathogenic influence on bats. However, the general correlation of virus and bat population dynamics suggests that bats control infections similar to other mammals and that they may well experience epidemics of viruses under certain circumstances.

Bats host noteworthy viral pathogens, including coronaviruses, astroviruses, and adenoviruses. Knowledge on the ecology of reservoir-borne viruses is critical for preventive approaches against zoonotic epidemics. We studied a maternity colony of Myotis myotis bats in the attic of a private house in a suburban neighborhood in Rhineland-Palatinate, Germany, during 2008, 2009, and 2010. One coronavirus, 6 astroviruses, and 1 novel adenovirus were identifi ed and monitored quantitatively. Strong and specifi c amplifi cation of RNA viruses, but not of DNA viruses, occurred during colony formation and after parturition. The breeding success of the colony was signifi cantly better in 2010 than in 2008, in spite of stronger amplifi cation of coronaviruses and astroviruses in 2010, suggesting that these viruses had little pathogenic infl uence on bats. However, the general correlation of virus and bat population dynamics suggests that bats control infections similar to other mammals and that they may well experience epidemics of viruses under certain circumstances.

B ats (Chiroptera) constitute ≈20% of living mammal species and are distributed on all continents except Antarctica (1) . Their ability to fl y and migrate, as well as the large sizes of social groups, predispose them for the acquisition and maintenance of viruses (2) . Although the ways of contact are unknown, bat-borne viruses can be passed to other mammals and cause epidemics (2, 3) . Several seminal studies have recently implicated bats as sources of important RNA viruses of humans and livestock, including lyssaviruses, coronaviruses (CoVs), fi loviruses, henipaviruses, and astroviruses (AstVs) (2, 4) . DNA viruses, including herpesviruses and adenoviruses (AdVs), have also been detected in bats, although with less clear implications regarding the role of bats as sources of infection for other mammals (5) (6) (7) (8) . While most of the above-mentioned viruses are carried by tropical fruit bats (Megachiroptera), the predominant hosts of mammalian CoVs, including those related to the agent of severe acute respiratory syndrome (SARS), are insectivorous bats (Microchiroptera) that are not restricted to tropical climates (1) . By demonstrating the presence of SARS-related CoV in Europe, we have recently shown that the geographic extent of its reservoir is much larger than that of other bat-borne viruses, including Ebola, Marburg, Nipah, and Hendra (9) .

In spite of the potential for serious consequences of virus epidemics emerging from bats, knowledge is currently lacking on the ecology of bat-borne viruses in bat reservoirs. We do not know how viruses with human pathogenic potential are maintained in bat populations, whether and how they are amplifi ed and controlled, and whether they cause effects on individual bats or on bat populations. The current lack of data is due to diffi culties in monitoring virus populations (rather than bat populations) in suffi cient density. Available studies have focused on lyssaviruses, hennipaviruses, and fi loviruses, which have extremely low detection frequencies, thus causing viruses to be encountered too rarely to enable the characterization of virus frequency and concentration over time (10) (11) (12) (13) (14) (15) . These studies have therefore relied on antibody testing, which provides higher detection rates by making indirect and cumulative assessments of virus contact during the lifetime of bats (10) (11) (12) (13) (14) (15) . However, results of antibody testing fail to correlate with the current presence of virus, preventing reliable analysis of a time component.

In a recent study, we obtained preliminary statistical hints that bats were more likely to carry CoV if they were young (16) . In adult bats, a signifi cant risk of carrying virus was identifi ed for lactating females (16) . Taking these clues together, we speculated that maternity roosts, inhabited predominantly by lactating females and newborns, with few adult males (17) , might serve as the compartment of CoV amplifi cation within the yearly life-cycle of bats in temperate climates. We therefore investigated the patterns of maintenance and amplifi cation of specifi c RNA-and DNA viruses by direct and quantitative virus detection in a maternity colony over 3 consecutive years. RNA-and DNA viruses were examined because of their different abilities to persist and to rapidly generate new variants. Viruses identifi ed included 1 CoV, 6 different AstVs, as well as a novel bat AdV. To assess the pathogenic infl uence of these viruses on bats, we quantifi ed the reproductive success of the colony over the same time period.

Permission for this work on protected bats was obtained from the environmental protection authority (Struktur-Und Genehmigungsbehörde Nord Koblenz) of the German federal state of Rhineland-Palatinate. Sampling took place over 3 consecutive years: 2008, 2009, and 2010. The sampling site was the attic of a private house in a suburban area in the state of Rhineland-Palatinate, western Germany ( Figure 1 ). The study did not involve any direct manipulations of bats and relied entirely on collection of fecal samples from the attic fl oor. Classifi cation of bats as Myotis myotis was confi rmed by mitochondrial DNA typing as described (9) . Adult female bats leaving the roost were counted by trained fi eld biologists before and after parturition. Pups were counted in the sampling site after the departure of adults. For each sampling date, plastic fi lm was spread in the evening on the ground of a 20-m 2 attic compartment, and fresh droppings were collected with clean disposable forks the following night. Each sample consisted of exactly 5 fecal pellets collected in proximity and added to RNAlater RNA preservative solution (QIAGEN, Hilden, Germany). The equivalent of ≈100 mg was purifi ed by the Viral RNA kit (QIAGEN) according to manufacturer's instructions.

Five microliters of RNA/DNA eluate were tested by broad range reverse transcription-PCR (RT-PCR) assays for the whole subfamily Coronavirinae (16) , the family Astroviridae (4), and the genus Mastadenovirus (18) . Specifi c real-time RT-PCR oligonucleotides were designed within the initial PCR fragments (those used are shown in Table 1 ). All 4 described real-time RT-PCR assays showed comparable lower limits of detection in the single copy range. Twenty-fi ve-microliter reactions used the SuperScript III PlatinumOne-Step qRT-PCR Kit (Invitrogen, Karlsruhe, Germany) for detecting CoVs and AstVs in M. myotis bats or the Platinum Taq DNA Polymerase Kit (Invitrogen) for For quantifi cation, PCR amplicons from the initial screening assay were TA cloned in a pCR 4.0 vector (Invitrogen). Plasmids were then purifi ed and reamplifi ed with vector-specifi c oligonucleotides, followed by in vitro transcription with a T7 promotor-based Megascript kit (Applied Biosystems, Darmstadt, Germany). The in vitro-transcribed RNAs or, in the case of AdVs, the photometrically quantifi ed plasmid alone, were used as calibration standards for virus quantifi cation in bat fecal samples, as described previously (19) .

Sanger sequencing of PCR products was done by using dye terminator chemistry (Applied Biosystems). Nucleic acid alignments with prototype virus sequences were done based on amino acid code by the BLOSUM algorithm in the MEGA4 software package (www.megasoftware. net). Neighbor-joining phylogenies used an amino-acid percentage distance substitution model and 1,000 bootstrap reiterations. All sequences were submitted to GenBank under accession nos. HM368166-HM368175. All analyses were performed with Epi Info 3.5.1 (www.cdc.gov/epiinfo) and with SPSS 17 (SPSS, Munich, Germany).

In a fi rst step, the M. myotis maternity colony was surveyed for bat-borne RNA viruses. Broad-range RT-PCR assays for CoVs and AstVs were employed on samples taken in 2008. Screening was extended to include AdVs described in microchiroptera and megachiroptera bats (6, 8, 20) . As shown in Figure 2 , a CoV, 6 different AstVs, and 1 novel AdV were found. The CoV (GenBank accession no. HM368166) was a member of the genus Alphacoronavirus and belonged to a tentative species defi ned by bat-CoV HKU6 (97.4% amino acid identity in RNA-dependent RNA polymerase [RdRp], typing criteria as defi ned in [9] ). The 6 different mamastroviruses (GenBank accession nos. HM368168-HM368175) clustered phylogenetically with bat-associated AstV, which has been described previously (4, 21) , showing 65.0%-86.0% amino acid identities with related bat-associated AstV from M. chinensis and M. ricketti bats from the People's Republic of China ( Figure  2 ). The AdV constituted a novel Mastadenovirus species (GenBank accession no. HM368167) that was clearly separated from a clade of AdV recently reported in a M. ricketti bat in China and a Pipistrellus pipistrellus bat in Germany (6,20) (A. Kurth, pers. comm.). The closest relatives were bovine AdV C10 (GenBank accession no. AF282774) and Tupaia AdV (GenBank accession no. NC_004453), with 90.0% and 91.0% identity on the amino acid level, respectively. Amino acid identity with the Chinese bat AdV TJM (GenBank accession no. GU226970) was 83.5%.

For all 3 viruses, strain-specifi c real-time RT-PCR assays, including cloned, in vitro-transcribed RNA or plasmid DNA quantifi cation standards, were generated (Table 1) . For AstV, 2 assays had to be designed to cover the high diversity of AstVs that was found. These assays were used to monitor virus abundance in the M. myotis bat maternity colony over time. CoV For AstV, no amplifi cation was associated with parturition in the same samples. Total detection rate of astroviruses was 51.2% before birth of the fi rst pup and 40.5% thereafter. However, prevalence and virus concentration signifi cantly increased in the second sampling than in the fi rst and fourth samplings, respectively (χ 2 7.4, p = 0.006); ANOVA, F = 4.4, p = 0.03). This pattern resembled the amplifi cation after formation of the colony as also observed in CoV. Figure 3 , panel B, shows AstV RNA concentrations over time.

Concentration and detection rates of AdV were determined next. As shown in Figure 3 , panel C, no marked variation in prevalence was seen. Detection rate was 46.4% before birth of the fi rst pup and 57.7% thereafter. Although statistically signifi cant variation in virus concentrations could be observed (ANOVA, F = 8.2, p<0.001), this was exclusively contributed by slightly lower virus concentrations in the fi rst sampling than in the succeeding samples ( Table 2) .

Because of the diverging pattern of amplifi cation of the RNA viruses (CoVs, AstVs) against the DNA virus (AdV), the investigation was repeated the next year (2009). All viruses were detected again ( Figure 3) . Unfortunately, the colony was found to be abandoned after the fi rst postparturition sampling, leaving an incomplete dataset for that year. Still, it could be seen and statistically confi rmed that the CoV was beginning to be amplifi ed after parturition (χ 2 7.85, p = 0.005), while no signifi cant variation in prevalence or virus concentration was visible for the other viruses (data not shown).

A repetition of the full sampling scheme was attempted again in 2010. All 5 sampling dates could be completed, yielding a sample of 187 pools in total, equivalent to 935 individual fecal pellets. As shown in Figure 3 , the CoV showed the same 2 amplifi cation peaks as observed in 2008, one after formation of the colony and one after parturition. Mean virus concentrations these samples were signifi cantly increased compared with the samples taken at other times (ANOVA, F = 22.0, p<0.001). The detection rate during the fi rst peak was 100.0%, followed by 2-fold and 5-fold decreases 3 and 6 weeks later (χ 2 52.0, p<0.001), and an augmentation to 97.5% after parturition (χ 2 77.7, p<0.001). The maximal CoV concentration in 2010 was higher than in 2008, at 50,495,886,830 RNA copies/g of feces. The amplifi cation pattern of AstV showed clearer similarities to that of CoV in 2010. An initial peak of detection rate was 97.5%, followed by a detection rate of 22.2%-22.4% in subsequent samples and 97.5%-100% after parturition (χ 2 56.2 and 92.2, respectively, p<0.001). Virus concentrations were signifi cantly increased in these amplifi cation peaks (ANOVA, F = 7.8, p<0.001). The amplifi cation was almost completely contributed by one of the AstV lineages (represented by BtAstV/N58-49), while the other lineages were constant (Figure 3, panel B) . Notably, the BtAstV/ N58-49 lineage had been present only sporadically in the years before (Figure 3, panel B) . Detection frequency for AdV was 58.6% before parturition and 40.3% thereafter without any signifi cant variation in virus concentrations between sampling dates (ANOVA, F = 0.5, p = 0.72).

CoV, AstV, and AdV are clearly pathogenic for other mammals. To determine whether the presence of these viruses had any infl uence on bats' health, the reproductive success of the maternity colony was evaluated in 2008 and 2010. The data are summarized in Figure 4 . In a census taken 2008 before parturition, the colony comprised 581 female adult bats. A second census after parturition yielded 394 adults and 220 newborns. The decline in adult females and the moderate number of pups contrasted with observations made in 2010, when 480 adult females were counted before parturition and 437 thereafter, along with 285 pups. The gain in total colony size was signifi cantly greater in 2010 than in 2008 (χ 2 18.3, p<0.001).

Viral host switching is probably determined by the chances of interspecies contact, as well as by the concentration and prevalence of virus in the donor species. To judge zoonotic risks associated with bats, when and where these 2 variables would favor transmission must be determined. In this study, we found that strong and specifi c amplifi cation of the RNA viruses, but not of the DNA virus, occurred upon colony formation and following parturition.

The viruses monitored in our study were selected because they are regularly encountered in bats and thus provide a certain chance of detection. Attempts to characterize virus dynamics in bat populations have been made earlier by using the examples of lyssaviruses (rabies virus and related species), fi loviruses (Ebola and Marburg viruses) and henipaviruses (Hendra and Nipah viruses). However, because these viruses are found rarely, only vague conclusions have so far been made. For instance, increased contact between bats and humans through bat migratory events or fruit harvesting periods have been temporally linked with individual human cases of Ebola and Nipah virus infection (22, 23) . One study has shown Empty columns indicate pools that tested negative. In panel B, light and dark gray bars identify results by 2 different real-time RT-PCRs that were used simultaneously to cover the large astrovirus diversity encountered.

that the success of Nipah virus isolation from Pteropus spp. bats depended on seasonal factors, which was interpreted as evidence for season-dependent variation of virus concentration or prevalence (24) . Furthermore, the reproductive cycle of bats has been tentatively connected with seasonality of henipavirus, fi lovirus and lyssavirus seropositivity in bats as well as with the temporal distribution of Nipah virus outbreaks in humans (10,12,,15, 25-27) . Our direct data on virus concentration and prevalence for CoV and AstV integrate many of these independent observations and provide a model that might be transferable to other viruses. The initial peak in annual CoV and AstV prevalence observed in our study was probably due to the formation of a contiguous population of suffi cient size and density, bringing together enough susceptible bats to establish a critical basic reproductive rate of infection (28, 29) . The second amplifi cation peak after parturition was most probably associated with the establishment of a susceptible subpopulation of newborn bats who had not yet mounted their own adaptive immunity. Sporadic vertical transmission from mothers to pups as observed in Pteropus spp. bats artifi cially infected with Hendra virus would probably initiate this second wave of infection (30) . The main driver of the second wave would then be a horizontal transmission between pups. The latency between parturition and the second wave of virus amplifi cation indicates a certain level of perinatal protection conferred by mothers during the fi rst weeks of life as demonstrated for other small mammals, and as indirectly suggested for bats (13, (31) (32) (33) . This protection may be differentially effective against different viruses, as indicated by the differential amplifi cation patterns between CoVs and AstVs. While CoVs were amplifi ed both in 2008 and 2010, AstVs underwent postparturition amplifi cation only in 2010 when a new virus lineage gained predominance in the population. This fi nding strongly indicates antigenspecifi c immune control of virus circulation. A common, but unproven, assumption is that bats are resistant to even highly pathogenic viruses (2, 3) . In this study, we have correlated direct measurements of virus burden with the reproductive success of a bat colony. The rate of successful reproduction is probably a sensitive indicator of the presence or absence of disease, given the tenuous conditions under which bats breed in temperate climates. Indeed, no effects of CoV and AstV on reproduction were initially apparent; although postparturition amplifi cation of both viruses was more effi cient in 2010 than in 2008, the overall breeding success was signifi cantly better in 2010. This result may merely have been a consequence of a positive correlation between virus amplifi cation and colony size, which was larger in 2010 due to better breeding success. On the other hand, the individual prenatal amplifi cation peaks of both CoV and AstV were higher in 2010, which may have enabled better perinatal protection and thus better survival of newborns. The grouping of large numbers of pregnant females before birth is a specifi c characteristic of bats that may contribute to their puzzling ability to maintain highly pathogenic viruses without experiencing die-offs. Our noninvasive approach did not allow any further analyses such as the testing of blood and colostrum samples for antibodies. Nevertheless, the general picture obtained in this study by correlating virus and bat population dynamics suggests that bats control infections in similar ways to other mammals, and that they may well experience virus epidemics.

Another intriguing fi nding of our study was the difference in the amplifi cation pattern of the RNA viruses and that of the DNA virus. We selected these viruses because, in humans, AdVs are typically capable of persisting in tissue (34) and thus do not depend so much on continuous transmission and consistent amplifi cation on the population level. Indeed, it appeared that AdV did not make use of periodic amplifi cation in our bat colony. Persistence on the level of individual bats is more common for DNA viruses than for RNA viruses. RNA viruses ensure that they are maintained on a population level by a much higher error rate of the enzymes they use for genome replication and consequent higher levels of antigenic variability, causing waves of epidemic spread as confi rmed for bat-borne RNA viruses in this study. This factor can explain why most emerging viruses, including those from bats, are indeed RNA viruses (2, 35) . For CoV, our study indicates clearly that virus amplifi cation takes place in maternity colonies, confi rming our earlier statistical implications from studies in a different region and on a different species (16) . High peak RNA concentrations in the range of 10 9 -10 10 copies/g were observed, which is tremendously higher than CoV concentrations observed in earlier studies outside the parturition period (19) . Similarly high RNA virus concentrations are observed in human diseases transmitted through the fecal-oral route, e.g., picornaviruses or noroviruses, which suggests that maternity roosts may involve an elevated risk of virus transmission to other hosts. It is interesting to reconsider the potential genesis of the SARS epidemic in this light. Although an origin of SARS-related CoV in bats is confi rmed (9, 36) , SARS-CoV precursors have existed in carnivores some time before the SARS epidemic and have been transmitted from carnivores to humans again at least one additional time after the end of the epidemic (37, 38) . these data provide an intriguing explanation of how the SARS agent may have left its original reservoir (39, 40) . The data also indicate a feasible and ecologically sensible means of prevention. Because carnivores are known to enter maternity roosts to feed on dead newborn bats, bat maternity roosts should be left undisturbed by humans and kept inaccessible to domestic cats and dogs. 

Mammal species of the world: a taxonomic and geographic reference

Bats: important reservoir hosts of emerging viruses

What links bats to emerging infectious diseases?

Genomic characterizations of bat coronaviruses (1A, 1B and HKU8) and evidence for co-infections in Miniopterus bats

Discovery of herpesviruses in bats

New adenovirus in bats

Partial molecular characterization of alphaherpesviruses isolated from tropical bats

Isolation of novel adenovirus from fruit bat (Pteropus dasymallus yayeyamae)

Genomic characterization of SARS-related coronavirus in European bats and classifi cation of Coronaviruses based on partial RNA-dependent RNA polymerase gene sequences

Spatial and temporal patterns of Zaire ebolavirus antibody prevalence in the possible reservoir bat species

Prevalence of rabies and LPM paramyxovirus antibody in non-hematophagous bats captured in the Central Pacifi c coast of Mexico

A longitudinal study of the prevalence of Nipah virus in Pteropus lylei bats in Thailand: evidence for seasonal preference in disease transmission

Henipavirus infection in fruit bats (Pteropus giganteus)

Henipavirus and Tioman virus antibodies in pteropodid bats

Host immunity to repeated rabies virus infection in big brown bats

Detection and prevalence patterns of group I coronaviruses in bats, northern Germany

Reproductive success, colony size and roost temperature in attic-dwelling bat Myotis myotis

Rapid typing of human adenoviruses by a general PCR combined with restriction endonuclease analysis

Distant relatives of severe acute respiratory syndrome coronavirus and close relatives of human coronavirus 229E in bats

Host range, prevalence, and genetic diversity of adenoviruses in bats

Detection of diverse astroviruses from bats in China

Human Ebola outbreak resulting from direct exposure to fruit bats in Luebo, Democratic Republic of Congo

Foodborne transmission of Nipah virus

Isolation of Nipah virus from Malaysian Island fl ying-foxes

Reproduction and nutritional stress are risk factors for Hendra virus infection in little red fl ying foxes (Pteropus scapulatus)

A longitudinal study of the prevalence of Nipah virus in Pteropus lylei bats in Thailand: evidence for seasonal preference in disease transmission. Vector Borne Zoonotic Dis

Recurrent zoonotic transmission of Nipah virus into humans

The pre-vaccination epidemiology of measles, mumps and rubella in Europe: implications for modelling studies

Estimation of effective reproduction numbers for infectious diseases using serological survey data

Experimental hendra virus infection in pregnant guinea-pigs and fruit Bats (Pteropus poliocephalus)

Protection of pregnant mice, fetuses and neonates from lethality of H5N1 infl uenza viruses by maternal vaccination

Role of milk-derived IgG in passive maternal protection of neonatal ferrets against infl uenza

Transfer of antibody via mother's milk

Adenovirus immunoregulatory genes and their cellular targets

Host range and emerging and reemerging pathogens

Bats are natural reservoirs of SARS-like coronaviruses

Recombination, reservoirs, and the modular spike: mechanisms of coronavirus cross-species transmission

Cross-host evolution of severe acute respiratory syndrome coronavirus in palm civet and human

Domestic cat predation on vampire bats (Desmodus rotundus) while foraging on goats, pigs, cows and human beings

Temporal patterns in the emergence behaviour of pipistrelle bats, Pipistrellus pipistrellus, from maternity colonies are consistent with an anti-predator respose

We are grateful to Monika Eschbach-Bludau and Sebastian Brünink for excellent technical assistance and to Manfred Braun for legal and expert advice. We thank Christian Nowak and the volunteers at the Bonn Consortium for Bat Conservation for fi eld assistance.This study was funded by the European Union