Hantaviruses represent a unique genus Hantavirus within the Bunyaviridae family. They are ”emerging viruses” which cause two human zoonoses, hemorrhagic fever with renal syndrome (HFRS) and hantavirus cardiopulmonary syndrome (HCPS) (Krüger et al., 2001; Mertz et al., 1997; Plyusnin et al., 2001a; Schmaljohn and Hjelle, 1997; Schmaljohn and Nichol, 2001). Prominent examples of hantaviruses that cause human disease are the Hantaan virus (HTNV), Seoul virus (SEOV), Puumala virus (PUUV) and Dobrava virus (DOBV) causing HFRS in Eurasia, and Sin Nombre virus (SNV) and Andes virus (ANDV) causing HCPS in the Americas. In contrast to other genera of the Bunyaviridaefamily, hantaviruses are not transmitted by arthropods. They are spread by aerosolised rodent excreta and produce chronic infection with no apparent harm in their natural hosts, rodents of the family Muridae

1.1 Short historical overview

Human disease now known as HFRS was probably for the first time described in Russian clinical records from Far Eastern Siberia in 1913 (Casals et al., 1970). However, Chinese writings from the 10th century already described a disease resembling HFRS (Lee et al., 1982). The first steps in elucidation of the etiology, epidemiology and ecology of HFRS were mostly done by military researches as the first epidemics in 20th century often occurred during military conflicts. Military physicians encountered the disease during World War I (“War nephritis”), invasion of Japan to Manchuria, World War II, as well as during the Korean Conflict in 1951 (Korean hemorrhagic fever). Despite a massive effort to isolate the etiologic agent during and after the Korean War, it was not until 1976 that the virus was discovered; the first hantavirus described, named Hantaan virus after the nearby Hantaan river in Korea, was isolated from the striped field mouse, Apodemus agrarius, by Ho-Wang Lee and co-workers (Lee et al., 1978). Soon, other HFRS-associated viruses such as PUUV from bank voles (Brummer-Korvenkontio et al., 1980), and SEOV from urban rats (Lee et al., 1982) were isolated using a similar approach. Antibodies to hantaviruses were found in rodents and humans all over the world and new related viruses were later isolated, e.g. Prospect Hill (PHV) in the USA from meadow voles, Microtus pennsylvanicus, or DOBV from A. flavicollis in Slovenia. Interestingly, the Thottapalyam virus, isolated from the shrew, an insectivore, in 1971 in India (Carey et al., 1971) was actually the first hantavirus ever isolated, but its relation to HFRS-causing viruses was detected many years later (Zeller et al., 1989; Xiao et al., 1994).

A new chapter in hantavirus research was opened in May 1993 when a sudden outbreak of a mysterious influenza-like prodrome of fever and myalgia that evolved into a relentless and often fatal syndrome of shock and edema occurred in the Four Corners region of USA. SNV was discovered as a causative agent of this disease, today known as HCPS (Nichol et al., 1993). This virus was shown to be carried by the common deer mouse (Peromyscus maniculatus) and other related, so called “New World hantaviruses” were soon discovered in other American rodents.

1.2 Hantaviruses within the Bunyaviridae family


Hantaviruses form a separate genus in the Bunyaviridae family. The family Bunyaviridae contains five genera: Bunyavirus, Phlebovirus, Nairovirus, HantavirusandTospovirus(Table 1). It representsone of the largest viral families with over 300 viruses. The features that encompass these diverse viruses within a single family are a common morphology, the tripartite genome, the absence of matrix in virions, and budding into intracytoplasmic vesicles from the internal membranes of the Golgi apparatus during replication. Ranging from 80 to 120 nm in size the viral particles are spherical or pleomorphic. They are enveloped viruses with a single stranded RNA genome of mostly negative polarity that is divided into three segments (Pringle, 1991).

Molecular genomic features are used to define the Bunyaviridae genera and antigenic data are used to separate the viruses within each genus. The most important human pathogens in the genus Bunyavirus include mosquito-borne California serogroup causing acute febrile illnesses and CNS infections (La Crosse, Tahyna). The Bunyamwera serogroup contains mostly tropical arboviruses. Crimean-Congo hemorrhagic fever virus is a representative of the Nairovirus genus, distinguished by the fact that its members are all transmitted by ticks. The Phlebovirus genus is divided into two groups: the Phlebovirus group with Sandfly fever and Toscana viruses transmitted by Phlebotomidae mites and causing a febrile illness with rash, and the Rift Valley fever virus, transmitted by mosquitoes, and causing veterinary and human outbreaks. Tomato spotted wilt virus is the only one member of the genus Tospovirus and is transmitted between plants by thrips, plant-feeding arthropods (Beaty and Calisher, 1991).

Hantaviruses (Table 2) are exceptional in that they have no arthropod vector. The reservoir hosts are specific rodents/insectivores. These features make it of special interest to study the genetic evolution of hantaviruses; the close phylogenetic correspondence between hantaviruses and their reservoir hosts has been interpreted as evidence for coevolution (cospeciation) of virus and host. Hantaviruses can cause severe, life threatening diseases in humans, however, they cause no detectable cytopathology in cell cultures and produce persistent non-pathogenic infections in rodents (Plyusnin et al., 1996; Plyusnin and Morzunov, 2001).


Table 1: Family Bunyaviridae. Taxonomical classification according to 7th report of The International Committee on Taxonomy of Viruses (ICTV) (Elliott et al., 1999) and selected significant pathogens


Type Species


Selected significant human/veterinary pathogens


Bunyamwera virus


La Crosse virus

Tahyna virus

Akabane virus

Oropouche virus


Hantaan virus


Hantaan virus

Sin Nombre virus


Dugbe virus


Crimean-Congo hemorrhagic fever virus

Nairobi sheep disease virus


Rift Valley fever virus


Rift Valley fever virus

Sandly fever-Sicilian virus


Tomato spotted wilt virus


Tomato spotted wilt virus

Table 2: Species in the genus Hantavirus, according to their rodent host.


rodent host

assigned abbreviation

Murinae -associated

Dobrava-Belgrade virus

Apodemus flavicollis


Hantaan virus

Apodemus agrarius


Seoul virus

Rattus norvegicus


Thailand virus

Bandicota indica


Arvicolinae -associated

Isla Vista virus

Microtus califonicus


Khabarovsk virus

Microtus fortis


Prospect Hill virus

Microtus pennsylvanicus


Puumala virus

Clethrionomys glareolus


C. rufocanus

Topografov virus

Lemmus sibiricus


Tula virus

Microtus arvalis



M. rossiaemeridionalis

Sigmodontinae -associated

Andes virus

Oligoryzomys longicaudatus


Bayou virus

Oryzomys palustris


Black Creek Canal virus

Sigmodon hispidus


Cano Delgadito virus

Sigmodon alstoni


El Moro Canyon virus

Reithrodontomys megalotis


Laguna Negra virus

Calomys laucha


Muleshoe virus

Sigmodon hispidus


New York virus

Peromyscus leucopus


Rio Mamore virus

Oligoyomys microtis


Rio Segundo virus

Reithrodontomys mexicanus


Sin Nombre virus

Peromyscus maniculatus


Insectivores -associated

Thottapalayam virus

Suncus murinus


* Only official virus species are shown. Tentative virus species, strains, or serotypes according to 7th report of ICTV (Elliott et al., 1999) are not shown.

1.3 Genome structure and replication

The virus genome consists of three segments of negative-stranded RNA; the large (L) segment encodes the viral RNA polymerase, the medium (M) segment the glycoproteins precursor (GPC), and the small (S) segment the nucleocapsid (N) protein (Figures 1, 2). The viral RNA-dependent RNA polymerase (L protein) acts as a replicase, transcriptase, endonuclease and possibly, RNA helicase. The GPC is cotranslationally cleaved into G1 and G2 proteins, which are thought to form a heterodimer. An open reading frame (ORF) of putative non-structural protein NSs have been found in many hantaviruses associated with Arvicolinae (e.g. PUUV, PHV, TULV) or Sigmodontinae rodents (e.g. SNV, BCCV, ELMCV) but not in Murinae-associated hantaviruses (HTNV, SEOV, DOBV). Despite the presence of nonstructural proteins in other bunyaviruses, NSs has not been found in hantavirus-infected cells. The functionality of these ORFs as well as their absence in Murinae-associated hantaviruses remains to be explained.


Sequence comparisons of distinct hantavirus species showed 60-70% identity at the nucleotide level for all three RNA segments. The corresponding values for deduced proteins are: 70-90% for L, 60-85% for N and 50-80% for G1, G2 proteins. The length of S segment varies significantly, mostly in its 3’ non-coding region (3’ NCR). However, within the hantavirus species, the length and sequence of 3’ NCR are conserved suggesting a functional role. Supposing the participation of S segment 3’ NCR in nucleic acid packaging, this step could be host-specific. In addition, the secondary structure of 3’ NCR might be crucial for this step of viral reproduction.

The 5’ and 3’-termini of all three genome segments are genus specific, highly conserved and complementary to each other. This enables to form panhandle structures of the RNA segments, which are typical also for other bunyaviruses. They are thought to play a role in viral transcription and in the proposed prime-and-realign mechanism of replication. Panhandles in hantaviruses are at least 17 nt long and the complementarity of the termini is incomplete.

Shortly after virion entry and uncoating in the cytoplasm of host cells, primary transcription by L protein occurs. The signal for switching from primary transcription to replication is most likely the accumulation of N protein as it is well documented for other negative-strand viruses. Transcription initiation requires suitable primers, which are acquired by the L protein from capped cellular mRNAs in the cytoplasm. The “prime-and-realign” model suggests that the terminal G residue of the host-derived primer aligns with the third nucleotide of the hantaviral RNA template (C residue) to initiate the transcription. After synthesis of a few nucleotides, the nascent RNA realigns by slipping backward three nucleotides on the repeated terminal sequences (AUCAUCAUC) of the L, M, and S RNA segments (Plyusnin et al., 1996; Schmaljohn and Nichol, 2001).


Figure 1: Scheme of a hantavirus particle (own drawing).

Figure 2: Hantavirus genome structure (adapted from Plyusnin et al., 1996).

1.4 Pathogenesis

1.4.1 Infection in natural host vs. humans

In both rodents and humans, replication occurs predominantly in pulmonary endothelial cells and macrophages. However, viral antigen is present in many organs and prominent in the spleen, kidney and lung. Hantavirus infection of rodents is asymptomatic. Although neutralising antibodies appear in rodents soon after infection, hantaviruses are not cleared from their rodent hosts. Animals are persistently infected and capable of transmitting the virus. Patients similarly develop neutralising antibody response but are able to clear the virus, suggesting that hantaviruses modulate rodent cellular host responses to effect viral persistence (Mackow and Gavrilovskaya, 2001). Recently, it was shown that HTNV can productively infect dendritic cells, upregulates costimulatory major histocompatibility complex (MHC) and adhesion molecules and induces the release of proinflammatory cytokines. This supports the hypothesis that in humans, hantaviruses elicit a strong immune response which could be an essential part of the virus-associated pathogenesis (Raftery et al., 2002).

1.4.2 Human diseases


Hantaviruses are known to cause two human diseases, HFRS and HCPS. They are transmitted to humans from rodents through inhalation of aerosolised excreted virus. Single report from South America suggests also person-to-person transmission of HPS-associated Andes virus (ANDV) but the relevance of this single event remains unclear (Padula et al., 1998).

Many clinical symptoms of both HFRS and HCPS are due to increased capillary permeability, which explains the hemorrhagic tendency and abdominal pain due to retroperitoneal edema in HFRS and extravasation of fluid to alveolar space and pulmonary edema occurring in HCPS. The specific feature that enables a certain hantavirus species to cause preferentially renal vs. pulmonary symptoms, or subclinical/mild vs. lethal manifestations remains to be explained (Krüger et al., 2001).

1.4.3 Hemorrhagic fever with renal syndrome

The clinical course can be usually divided into five distinct phases. After an incubation period of 2-4 weeks, there is an abrupt onset of disease with fever, chills, general malaise, headache and other influenza-like symptoms, nausea, back and abdominal pain, gastrointestinal symptoms. This febrile phase usually lasts for 3-7 days. Towards the end of this phase, conjunctival hemorrhages and fine petechiae occur at the body surface. The hypotensive phase can last from several hours to two days. The characteristic drop of the blood platelet begins. In severe cases, a clinical shock state occurs and one third of HFRS deaths are associated with irreversible shock at this stage. In the oliguric phase (3-7 days) due to renal failure, a massive proteinuria occurs. One half of fatalities occur during this phase. Typical findings are elevated concentrations of serum creatinine and urea. Blood pressure becomes normalised or even changes to hypertension. The onset of the diuretic phase is a positive prognostic sign for the patient. Diuresis of 3-6 litres is usually observed. The convalescent phase is characterised by recovery of the clinical and biochemical markers (Krüger et al., 2001).


The most severe forms of HFRS are caused by HTNV in Eastern Russia, China and Korea. Severe HFRS cases occur also in Europe, mostly in Balkan region, caused by DOBV. Most clinical cases due to SEOV infection exhibit a milder course than the HTNV infections. PUUV usually causes a rather mild form of HFRS, called Nephropathia epidemica(Krüger et al., 2001).

1.4.4 Hantavirus cardiopulmonary syndrome

In contrast to the five phases classically known for HFRS, clinical features of HCPS were originally divided into four phases (febrile, cardiopulmonary, diuretic, and convalescent). Febrile phase, typically lasting 3-5 days is characterised by fever, myalgia, and malaise. Other symptoms such as headache, dizziness, anorexia, nausea, vomiting, diarrhoea and abdominal pain may be present. Early recognition of HPS in this period is difficult, and the disease is indistinguishable from other viral prodromes. At the end, announcing the onset of pulmonary edema, non-productive cough and tachypnea occur. Cardiopulmonary phase is characterised by the presentation of shock and pulmonary edema. Hypotension and oliguria can be accompanied by shock. Tachypnea, exertional dyspnea, and non-productive cough are the clinical expression of pulmonary edema. Once pulmonary edema is present, the disease proceeds fast. Patients can decease within 24-48h; hypoxia, circulatory compromise, or both being the immediate cause of death. In diuretic phase, rapid clearance of pulmonary edema as well as resolution of fever and shock occur. Spontaneous diuresis is an early sign of this process. Convalescent phase may last up to two months, with patients recovering apparently completely. However, continued follow up is necessary to determine the long-term persistence of pulmonary dysfunction.

1.5 Virus ecology

Hantaviruses are exclusively maintained in the populations of their specific rodent hosts. Unlike other bunyaviruses, they are not transmitted by arthropod vectors. The order Rodentia contains the most species of mammalian orders and has members that are keystone species in most ecosystems of the world. Much of the ecological complexity can be found within the Muridae family. This family contains species that serve as reservoirs for all but one (Thottapalyam virus) known hantavirus species and also all but one species of arenavirus. Interestingly, such murid species are subject to large population changes, either as a result of human land-use practices, climate change, or normal population cycles, that pose the most serious threats to human health (Hjelle and Yates, 2001).


The mechanisms for maintenance of hantavirus infections among reservoir rodent populations are not well understood. Infection levels change rapidly as population densities in rodent populations fluctuate. Field studies have demonstrated that hantavirus seroprevalence is highly local, with rodent communities exhibiting 0-1% and as much as 50% prevalence. Seroprevalence at a particular site can vary greatly over time. For most hantaviruses, larger male animals are much more commonly infected than smaller males or females (Hjelle and Yates, 2001).

In a 4-year study of bank voles populations in southern Belgium, PUUV infection in adults was found to be associated with wounds at the end of breeding season, but not in spring. Sexually active animals were significantly more often wounded and positive for infection. Hantavirus infection was associated with higher mobility in juvenile and subadult males. Together with these behavioural and physiologic factors, the habitat also constitutes a crucial element influencing the hantaviral enzootic cycle by determining the distribution of the rodents (Escutenaire et al., 2002).

Olsson et al. (2002) showed that localised absence of PUUV coincided with the absence of overwintering specimens at several sites during population decline. Long-lived bank voles appear critical to the success of PUUV circulation and persistence within host populations. The chance of being seropositive is population density dependent.

1.6 Evolution of hantaviruses

1.6.1 Methods employed in phylogenetic analysis of viral sequences


The aim of phylogenetic analysis is to arrive at the best possible estimate of the true evolutionary history of the organisms, i.e. their phylogeny. Evolution is the accumulation of change in the genetic makeup of populations over time. As the genetic information is encoded as discrete states (occurrence of A, G, C or T/U in DNA/RNA, or amino acids in the encoded proteins) organised along linear molecules, it is ideally suited for computer analysis. The results of phylogenetic analyses are often presented as “trees”, a set of lines (branches) connecting the sequences (tips) via branching points (nodes) (Hungnes et al., 2000).

The methods for constructing phylogenetic trees from molecular data can be grouped according to whether the methods uses discrete character states or a distance matrix of pairwise dissimilarities. Distance matrix methods start by calculating a pairwise distance matrix from analysed sequences and then estimate the phylogenetic relationship. In character-state methods, the sequence information is not reduced to the set of pairwise distances but all characters are analysed separately and usually independently from each other (Vandamme, 2003).

In principle, distance methods try to fit a tree to a matrix of pairwise genetic distances. The distance is a single value based on the fraction of positions in which the two sequences differ, defined as p-distance. The p-distance is an underestimation of the true genetic distance because some of the aligned nucleotides are the results of multiple events. Therefore, in distance based methods, one tries to estimate the number of substitutions that have actually occurred by applying a specific evolutionary model that makes assumptions about the nature of evolutionary changes. Correct estimation of genetic distance is crucial and, in most cases, more important than the choice of method to infer the tree topology. Today the most commonly used method to construct distance trees is Neighbor-joining (NJ) method. The NJ method has been proven to be quite efficient in finding the “true” topologies or those that are close. It has the advantage of being very fast, which allows the construction of large trees including hundreds of sequences (Van de Peer, 2003).


The most commonly employed discrete-character methods are parsimony and maximum-likelihood methods. The basic idea of parsimony analysis is simple: one seeks the tree that minimises the amount of evolutionary change required to explain the data. Parsimony methods are most effective when rates of evolution are slow, in other words, the expected amount of change is low (Swofford and Sullivan, 2003). Because this is very often not the case for viral sequences, parsimony methods are in virology not as widely used as NJ or very recently maximum likelihood methods.

Maximum likelihood (ML) analysis is a widely used statistical method that is being applied to a broad range of analyses. In phylogenetic analysis it has been most extensively developed for nucleotide sequence data. With this method, the optimal tree is one that gives the highest probability of observing the actual sequences, given a particular model of evolution. The models allow the specification of probabilities for different kinds of mutations, for example transitions and transversions. ML is a very powerful method when the probabilistic model is realistic (Hungnes et al., 2000). Moreover, the ML method intrinsically estimates the standard error on the branch length and therefore gives some statistical support for each branch length and for the entire tree (Vandamme, 2003).

Compared to reality, even the most intricate models of evolution are necessarily simplistic. Thus, complex models are expected to perform better than simpler models, provided that the parameters are realistic. However, as parameters inferred from the small amount of data can be very misleading, in some cases a simple model will perform better than a complex one (Hungnes et al., 2000). The determination of an appropriate model can be accomplished using likelihood ratio tests, the Akaike information criterion, Bayesian information criterion or more subjective methods. Particularly the simpler models, such as Jukes and Cantor model (JC69), Kimura 2-parameters model (K80) or even Hasegawa, Kishino, and Yano model (HKY85) should never be used uncritically (Swofford and Sullivan, 2003).


Most of the tree constructing methods are implemented in two most widely used program packages, PHYLIP (Felsenstein, 1993) and PAUP* (Swofford, 2002). In addition, TREE-PUZZLE program for ML analysis (Schmidt et al ., 2002) is currently gaining the popularity because of its computationally fast quartet puzzling algorithm and additional features as Likelihood mapping analysis and parallel computing.

Recombination is increasingly seen as an important means of shaping genetic diversity in RNA viruses. Given the drastic effect that recombination can have on phylogenetic studies, it is highly desirable that the recombination analysis should be carried out on whole genomes wherever possible, so that putative recombinants can be identified and the possibility of misclassification of strains reduced (Twiddy and Holmes, 2003)

Typical approaches to identify recombination start with splitting the full alignment into a set of smaller overlapping alignments. Phylogenetic analyses are subsequently performed on each of the subalignments in order to see if different regions support different evolutionary histories. Similarity- or dissimilarity-based methods are among the fastest and theoretically least complicated of these methods. The similarity of a sequence to a given set of reference sequences is computed and plotted along the sequence using a sliding window fashion. If recombination results in the crossover between two sufficiently separated evolutionary lineages, it will be reflected in a plot of the query sequence by a gradual switch of the highest similarity from one reference sequence to another. In a similar sliding-window-based method, bootscaning, a combination of phylogenetic analysis and bootstrap values associated to specific clusters of sequences are used to map recombination breakpoints. Both similarity plots and bootscaning are implemented in the Simplot software (Lole et al ., 1999) , one of the most versatile of the currently available recombination exploration applications (Salminen, 2003)


In addition to these and other graphical applications (e.g. Split Decomposition analysis, TOPAL, PhylPro), more sophisticated procedures exist for locating crossover points (e.g. Informative Sites Analysis, LARD or Homoplasy test) (Worobey and Holmes, 1999).

1.6.2 Phylogenetic analysis of hantaviruses

Hantavirus species are strongly associated with one (or a few closely related) specific rodent species as their natural hosts. Absence of arthropod vector in the virus lifestyle seems to be a crucial in the evolution of hantaviruses and predetermined observed patterns of tight association of these viruses with their specific rodent hosts (Plyusnin and Morzunov, 2001). Phylogenetic analyses of hantaviruses reveal three well-differentiated clades corresponding to viruses circulating in three subfamilies (Murinae, Arvicolinae, and Sigmodontinae) of the rodent family Muridae (Figure 3). Moreover, in trees of M and L genes, the viruses of Arvicolinae and Sigmodontinae form a sister group and the viruses of Murinae rodents form outgroup to them. This phylogeny corresponds with a phylogeny of the murid subfamilies based on mitochondrial cytochrome b sequences, supporting the hypothesis that hantaviruses have coevolved with their mammalian hosts at least since the common ancestor of these three subfamilies, which probably occurred about 50 million years ago (MYA) (Hughes and Friedman, 2000).

Similar pattern of branching, reflecting cospeciation, can be observed from the basal nodes to the very fine terminal branches of hantavirus and rodent phylogenetic trees. However, clear pattern of cospeciation could be occasionally disrupted by host-switching events. A typical example could be the association of Khabarovsk virus (KHAV) with Microtus fortis and Topografov virus (TOPV) with Siberian lemmings (Lemmus sibiricus); while TOPV and KHAV are monophyletic, the respective rodent host species are only distantly related (Plyusnin and Morzunov, 2001; Vapalahti et al., 1999).


Evolutionary rate in hantaviruses was estimated using the rate of synonymous substitutions between the S genes from the Old World and New World clusters of viruses which are hosted by members of the genus Microtus. It is believed that the first separation of Old and New World Microtus occurred in the early Pleistocene, 1.8 – 2.0 MYA (Hoffmann and Koeppl, 1985). Assuming that the viruses have coevolved with their hosts, the corresponding substitution rate would be 2.41 – 2.68 x 10-7 substitutions per site per year. Although considerably lower than synonymous substitution rates of such viruses as influenza or HIV-1, this rate is about two orders of magnitude faster than typical mammalian mutation rate (Hughes and Friedman, 2000).

Although it is currently taken almost as a dogma, Holmes (2003) has recently questioned the virus-host coevolution and cospeciation concept for RNA viruses including hantaviruses. By using the best estimates for rates of evolutionary change (nucleotide substitution) and assuming an approximate molecular clock, it can be inferred that the families of RNA viruses circulating today could only have appeared very recently, probably not more than about 50,000 years ago. Besides the explanations that the molecular clock is not constant in RNA viruses and that the methods currently used to estimate evolutionary distances are flawed in some way, leading to a substantial underestimation of divergence times, the third explanation is that RNA viruses really have a recent origin. The match between virus and host phylogenies that has been taken as evidence for cospeciation over millions of years has to be explained by other mechanisms. One of them might be that the ability to jump species boundaries may be inversely dependent on the phylogenetic distance between hosts, so that it is easier to establish a new infection in a closely related host species than in a more distantly related one.

To determine the extent of homologous recombination in negative-sense RNA viruses, phylogenetic analyses of 35 negative-sense RNA viruses (a total of 2154 sequences) were carried out (Chare et al., 2003). Powerful evidence for recombination was found in only five sequences including hantavirus HTNV. In addition, more tentative evidence was found also in PUUV. Evidence for recombination has been documented also for TULV (Sibold et al., 1999a; Plyusnin et et al., 2002) and some indication was found also for DOBV (chapter 3.2.4; Klempa et al., 2003b). However, overall the rates of homologous recombination in including hantaviruses are very much lower than those of mutation. Consequently, recombination does not seem to be a main driving force in the evolution of hantaviruses or other negative-sense RNA viruses (Chare et al., 2003).


Figure 3: Phylogenetic relationship between the main hantavirus representatives corresponding to the three subfamilies of Murinae rodents.

Maximum likelihood phylogenetic tree (JTT evolutionary model) based on complete N protein amino acid sequence, calculated with TREE-PUZZLE, is shown.

1.7 Dobrava hantavirus

DOBV is intensively studied because of many of its unique properties. DOBV seems to be the most pathogenic European hantavirus. The severity of DOBV- associated HFRS can reach the death rate of up to 12%, as reported in Balkans (Antoniadis et al., 1996; Avsic-Zupanc et al., 1999; Lundkvist et al., 1997c; Papa et al., 1998). Most interestingly DOBV is hosted in Europe by at least two different rodent species, A. flavicollis (yellow necked mouse) and A. agrarius (striped field mouse) (Figure 4). Moreover, there is a hypothesis that the DOBV strains originating from different rodent hosts exhibit different pathogenicity towards humans. These interesting facts led recently to intensive discussions whether the strains originating from A. agrarius do represent a unique hantavirus species or only a lineage of original DOBV (Plyusnin and Morzunov, 2001; Nemirov et al., 2002; Brus Sjolander et al., 2002; Plyusnin, 2002; Plyusnin et al., 2003; Klempa et al., 2003b, c).

DOBV prototype strain (called here Dobrava/Slovenia, Slo/Af for short) was isolated from lungs of A. flavicollis captured in a natural focus of HFRS in Dobrava village, Slovenia. Immunofluorescent antibody assays using convalescent human sera and monoclonal antibodies (MAbs) indicated that DOBV differed from all other recognised hantaviruses but was most closely related to HTNV. Moreover, the part of viral RNA was amplified by RT PCR and in the restriction fragment length polymorphism analysis of the amplified product it was shown to be unique (Avsic-Zupanc et al., 1992). Later, the phylogenetic analysis of M and S segments confirmed that DOBV is similar to, but clearly distinct from HTNV and SEOV. Cloning and sequence analysis revealed the M segment to consist of 3644 nt, with a coding capacity of 1134 aa in the virus complementary sense RNA. Seven potential asparagine-linked glycosylation sites were identified in the M segment gene product, one in the G2 and six in the G1 coding regions. The S segment is 1667 nucleotides long, and has a single ORF encoding a protein of 429 aa (Xiao et al., 1993; Avsic-Zupanc et al., 1995).


Soon, reports about detection of DOBV in other European countries started to appear. DOBV nucleic acid was detected by PCR and sequencing in Greek and Albanian HFRS patients (Antoniadis et al., 1996). By Focus reduction neutralisation test (FRNT), DOBV neutralising antibodies were found in patient sera from Bosnia-Herzegovina (Lundkvist et al., 1997c), in sera of retrospectively studied HFRS outbreak in Russia (1991-92 in Tula-Ryazan region, Lundkvist et al., 1997a), in HFRS patient serum from Germany (Meisel et al., 1998), in human sera in Estonia (Lundkvist et al., 1998) and Slovakia (Sibold et al., 1999b).

Figure 4: Apodemus flavicollis (yellow necked mouse) -above- and A. agrarius (striped field mouse) -below-, the natural hosts of DOBV (own pictures).

Nineteen HFRS patients due to infection with DOBV from eastern and northern Germany and from Slovakia were investigated. They had shown typical signs of HFRS as fever and other influenza-like symptoms, acute renal failure, trombocytopenia and raise of serum creatinine. DOBV infection was verified by the FRNT where in all cases the sera did neutralise DOBV virus to a significantly higher extent than other virus species including the related HTNV and SEOV. The average age of the patients at the time of disease was 33.8 years; the youngest patient was 15 years old at onset of disease. Interestingly, only 10.5% (n=2) were females but 89.5% (n=17) were males. Two patients (10.5%) were reported to have developed pulmonary symptoms. No visible haemorrhages occurred. There were no deaths and the clinical course of the HFRS was mild or moderate. Only 18.8% of the patients underwent hemodialysis because of renal failure (Sibold et al., 2001).


Surprisingly, DOBV was detected in A. agrarius trapped in Estonian islands Saaremaa and Vormsi. Analysis of the partial S segment sequences revealed that they belong to the DOBV genotype and share an ancient ancestor with the Slovenian prototype strain but represent a distinct sublineage. These findings raised the questions whether DOBV can be maintained in both Apodemus species and whether the virus exists throughout Europe (Plyusnin et al., 1997b).

Subsequently, the virus detected in A. agrariustrapped on Saaremaa island in Estonia, was isolated on Vero E6 cells (strain Saaremaa/160V, Saa/160V for short) and its genetic and antigenic characteristics were analysed (Nemirov et al., 1999). A cross-neutralisation comparison with the Slovenian prototype strain isolated from A. flavicollisrevealed 2- to 4-fold differences in the end-point titers of some rabbit and human antisera. When studied with a panel of 25 MAbs, the Estonian (Saa/160V) and Slovenian (Slo/Af) isolates showed similar antigenic patterns that could be distinguished by two MAbs, PUUV N-specific 4E5 (Lundkvist et al., 1991) and the anti-HTNV-G1 MAb 8B6 (Arikawa et al., 1989). The complete sequences of the S and M segments and partial L segment sequence of the isolate were determined. The phylogenetic analysis confirmed that Estonian strains from A. agrarius (Saa/160V and Saa/90Aa) form a well-supported sublineage within the DOBV clade (Nemirov et al., 1999). Recently, a new DOBV cell culture isolate was established. DOBV/Ano Poroia (AP/Af) was isolated from an A. flavicollis mouse trapped in the North-Eastern Greece. It represents the third DOBV isolate in the world, second obtained from A. flavicollis(Papa et al., 2001).

DOBV strains in A. agrarius were detected also in Slovakia (Sibold et al., 1999b), Hungary (Scharninghausen et al., 1999) and Russia (Plyusnin et al., 1999b), where no casualties have been reported. This led to the idea that A. flavicollis-associated strains (Balkans) and A. agrarius-associated strains (North-East and Central Europe) exhibit different pathogenicity toward humans (Plyusnin et al., 1999b).


The existence of two distinct, rodent host-determined DOBV genetic lineages was definitively confirmed when strains of both lineages were found to occur sympatrically in Slovenia (Avsic-Zupanc et al., 2000) and Slovakia (Sibold et al., 2001). In Slovenia, partial S and M segment sequences were recovered by RT-PCR from nine A. flavicollis and rather exceptionally also from one A. agrarius. Sequence comparison and phylogenetic analysis revealed close relatedness and geographical clustering of all A. flavicollis-derived virus sequences. In contrast, the single strain harboured by A. agrarius clustered on phylogenetic trees with other DOBV strains derived from A. agrarius(Avsic-Zupanc et al., 2000). A “complementary” situation was found in the Eastern part of Slovakia, where screening for infected rodents revealed that A. agrarius represents the main reservoir for DOBV in this region but a single strain from A. flavicollis was also detected. Phylogenetic analysis placed this strain on a A. flavicollis-derived lineage (DOBV-Af) whereas other Slovakian strains belong to A. agrarius-derived lineage (DOBV-Aa) (Sibold et al., 2001).

These interesting findings raised several important questions. On one hand, closely related DOBV lineages are present in two distinct rodent species, on the other hand, two subspecies of A. agrarius are harbouring distantly related hantaviruses DOBV (A. agrarius agrarius) and HTNV (A. agrarius matchuricus). Such a discrepancy might be explained by a host-switch event of (pre)DOBV from A. flavicollis to A. agrarius(Plyusnin and Morzunov, 2001; Nemirov et al., 2002)Moreover, the occurrence of two significantly different lineages steadily linked to the two different host led to the question whether these two DOBV variants represent distinct subtypes or even distinct hantavirus species. This question still remains to be answered and is thoroughly discussed in this work.

In addition, DOBV was recently found in A. sylvaticus in South-European Russia and this virus strain could be associated with a severe HFRS case, which occurred in that geographical area (Tkachenko et al., 2001). More sequence data of DOBV strains from A. sylvaticus might reveal the definition of third, DOBV-As lineage within DOBV species.


Altogether, DOBV have been detected as an etiologic agent of HFRS in many European countries. Interestingly, most of the reports were only from three regions, Balkans, Central Europe and Russia/Estonia and are summarised in Table 3.

Table 3: The current state of the detection of DOBV. The summary of epidemiological and epizootiological data on DOBV from three regions of Europe. The data which were extended by this work are in bold.




Central Europe

Neutralising antibodies (FRNT)

In seroprevalence studies


+ 1

+ 2

In patients

+ 3

+ 4

+ 5

Clinically characterised patients

Dozens 6

1 7

26 (3) 5



Af, Aa 8

Aa, As 9

Aa, Af 10


+ 11

+ 12

+ 13

Virus isolation


Slo/Af, AP/Af 14

Saa/160V 15

SK/Aa 16





1 Golovljova et al., 2002; Lundkvist et al., 1998
2 Sibold et al., 1999b
3 Avsic-Zupanc et al., 1999; Lundkvist et al., 1997c
4 Lundkvist et al., 1997a; Golovljova et al., 2000
5 Meisel et al., 1998; Mentel et al., 1999; Schütt et al., 2001; Sibold et al., 2001; Klempa et al., 2004;
chapters 3.5, 3.6
6 Lundkvist et al., 1997c; Avsic-Zupanc et al., 1999; Markotic et al., 2002
7Golovljova et al., 2000
8 Avsic-Zupanc et al., 1992, 2000; Papa et al., 2000, 2001
9 Plyusnin et al., 1997b, 1999; Tkachenko et al., 2001
10Scharninghausen et al., 1999; Sibold et al., 2001; Klempa et al., 2003b; chapter 3.1
11 Antoniadis et al., 1996; Papa et al., 1998; Markotic et al., 2002
12 Tkachenko et al., 2001
13 Klempa et al., 2004; chapter 3.5
14Avsic-Zupanc et al., 1992, Papa et al., 2001, respectively
15Nemirov et al., 1999
16 Klempa et al., submitted; chapter 3.4
* Af stands for Apodemus flavicollis, Aa for A. agrarius and As for A. sylvaticus

1.8 Tula hantavirus

In the nineties, a new, PUUV-related hantavirus has been found in European common voles (Microtus sp.) and was called Tula hantavirus, TULV (Plyusnin et al., 1994; Sibold et al., 1995; Vapalahti et al., 1996). In the meantime, TULV has been detected in voles from several European regions (Bowen et al., 1997; Heyman et al., 2002; Plyusnin et al., 1995; Sibold et al., 1999a). For a long time, TULV was considered to be non-pathogenic towards humans. Only very recently, few data about TULV infections of humans are coming up. However the pathogenic potential of TULV is still not well determined. Moreover, TULV has been attractive to “hantavirologists” and molecular biologists because of the homologous recombination which was detected in TULV strains from East Slovakia (Sibold et al., 1999a).


The detection of TULV was for the first time described by Plyusnin et al. (1994). Six specimens containing hantavirus antigen related to PUUV and/or PHV, five from M. arvalis (Figure 5) and one M. rossiaemeridionalistrapped in Tula region, Russia, were found positive in RT-PCR. The sequence analysis of five cloned and sequenced complete S segments showed that the virus is genetically related to but distinct from known members of the Hantavirus genus. Immunochemical data obtained for TULV antigen derived from lung tissues of infected rodents and recombinant GST-TUL-N fusion protein confirmed the sequence data. Analysis of TULV with the panel of MAbs revealed that TULV is more related to PUUV and PHV than to HTNV and SEOV but is antigenically distinct from PUUV.

Independently and simultaneously, the TULV sequences were detected also in M. arvalis captured in western Slovakia (near Malacky town). Sequence analysis of a major part of the S segment showed this strain, called Malacky, to represent a new subtype within the newly characterised genotype (Sibold et al., 1995).

TULV was then detected also in five common voles from Moravia (Czech Republic). In addition to the full length S segment sequences, the proximal part of the M segment (about 1 kb) of two Moravian as well as two Russian strains was sequenced. Phylogenetic analysis suggested a similar evolutionary history for S and M genes of TULV. Comparison of the deduced N protein sequences showed that genetic drift in TULV can occur not only by accumulation of point mutations but also by the deletion of a nucleotide triplet (Ser252). Analysis of naturally expressed TULV N-antigen derived from lung tissue of infected voles with MAbs indicated antigenic heterogeneity among TULV strains (Plyusnin et al., 1995).


Subsequently, TULV was isolated from one out of five RT-PCR positive common voles from Moravia, Czech Republic. A Vero E6 cell culture isolate was established after initial passaging of lung samples in laboratory-colonised M. arvalis. TULV was defined as a classical serotype by a cross-FRNT. Moreover, serological evidence for a previous TULV infection was obtained from the serum of a blood donor from Moravia, Czech Republic, showing at least a 16-fold higher titer to TULV as compared to PUUV and other hantaviruses (Vapalahti et al., 1996).

Phylogenetic analysis of TULV genetic variants from Slovakia revealed interesting results. In complete S segment phylogenetic tree, strains from Eastern Slovakia clustered with Russian strains and strains from western Slovakia were situated closer to those from Czech Republic. However, phylogenetic analysis of the S segment 3’ NCR placed the Eastern Slovakian strains on branch together with Western Slovakian and Czech strains. A bootscan search revealed at least two recombination points in the S sequences of Eastern Slovakian strains which agreed well with the pattern of amino acid substitutions in the N protein and deletions/insertions in the 3’ NCR. Altogether, these data suggested that homologous recombination events could play some role in evolution of hantaviruses (Sibold et al., 1999a). These findings were confirmed by transfection-mediated generation of functionally competent TULV with recombinant S RNA segment. Independent attempts yielded S RNA molecules of similar structure, carrying a break point located close to one of the break points suggested for natural recombinants (Plyusnin et al., 2002).

Figure 5: Microtus arvalis (common vole), the natural host of TULV (© Rollin Verlinde – www.natuurbeleving.be)

1.9 Aims of the study


Hantaviruses in Europe are known for many years but the knowledge about their distribution, molecular evolution as well as pathogenic relevance is restricted only to some geographical regions, mostly Fennoscandia and Balkans. The aims of this study were to extend this knowledge to an important area of Central Europe, where DOBV, PUUV as well as TULV have been recently detected.

Currently, Dobrava virus (DOBV) is intensively studied because of its unique properties; different virus lineages exist in different regions of Europe which are harboured by different host reservoirs and which are probably of different virulence towards humans. Although a large overlap between the geographical distribution of A. agrarius and A. flavicollis exists in Europe, so far the two host-specific subtypes have been detected sympatrically only in the border region between Central and South-East Europe. This detection of DOBV strains fromA. agrarius andA. flavicollis occurring sympatrically in Eastern Slovakia (Sibold et al., 2001) offered an opportunity to remove the effects of geographic isolation on the supposedly host-specific genetic determinants that distinguish DOBV-Af from DOBV-Aa and to examine whether the two virus types are subject to genetic interactions with one another that could influence their evolutionary trajectories.Therefore, the first aim of this work was to extend the work of Sibold et al. (2001); to determine complete S and M segment nucleotide sequences of these strains, to analyse their evolutionary history and to study the role of reassortment and recombination processes in hantavirus evolution. In addition, we were intended to identify the candidate host-specific amino acid adaptations in the N proteins and glycoproteins.

By rodent screening, we wanted to identify additional DOBV strains and to infer their phylogenetic relationship. One of the main aims was to establish new DOBV isolates. Particularly an isolate from A. agrarius was needed which would better represent the Central European DOBV-Aa lineage than the only isolate from Estonia. For future comparative investigations on experimental virus evolution, virus life cycle, pathogenicity and virus-host interactions it will be crucial to use a viable DOBV-Aa prototype virus.


Because of recent speculations about different pathogenicity of DOBV-Aa and DOBV-Af lineages, we wanted to describe some interesting clinical cases which could contribute to characterisation of DOBV pathogenic potential. Moreover the aim was to obtain viral sequences directly from patient material. The genetic characterisation of viral gene sequences from HFRS patients, enabling molecular classification of the virus strain and providing direct evidence whether DOBV causes HFRS in Central Europe, was still missing. The obtained nucleotide sequence would add some epidemiological relevance to previous phylogenetic analyses based only on sequences obtained from rodents.

Moreover, our interests were not restricted to DOBV. We wanted to determine not only what hantaviruses are present in rodents populations of Germany and Slovakia but also which of them cause hemorrhagic fever with renal syndrome.

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated: