With a production of four million tons worldwide (http://www.fao.org) common carp (Cyprinus carpio L.) is one of the most important fish for aquaculture. Moreover, its ornamental variety, the koi, is one of the most expensive ornamental fish in the world. A severe infection that leads to high mortalities of Cyprinus carpio stocks has been reported repeatedly since 1997. This disease is associated with skin lesions and gill necrosis (Bretzinger et al., 1999). The causative agent was first described as an aquatic herpesvirus in Hedrick et al. (2000) and was named koi herpes virus (KHV) after its host. Later, it was given the scientific name Cyprinid herpesvirus 3 (CyHV-3) and was classified as a member of the family Alloherpesviridae (Waltzek et al., 2005). With a genome size of 295 kbp, KHV is the largest known herpesvirus. The 11 complete genomes published so far show a very high sequence identity (>99.9%; Aoki et al., 2007; Li et al., 2015; Hammoumi et al., 2016).

To gain more knowledge on the diversity of this virus, a duplex PCR was established (Bigarre et al., 2009) to discriminate between the different virus introductions into Europe. This PCR uses two marker regions to determine the relationship between different KHV variants. Unfortunately, the temporal and spatial resolution of this assay is too low to permit phylogeographical analyses or molecular tracing for origin detection. Subsequently, eight DNA regions with variable number of tandem repeats (VNTRs) were selected and tested as a tool for KHV discrimination (Avarre et al., 2011). Using the fluctuating copy number combined with hierarchical clustering, it was possible to obtain information on the phylogeographical background of some investigated KHV specimens (Avarre et al., 2012).

The present study aimed at designing a new qPCR assay able to distinguish between different variants of the Asian and European lineages. A second goal was to improve the previously described MLVA using different primer pairs covering the eight complete VNTR sequences. Finally, these two methods were applied to trace the variability of four KHV isolates passaged in vitro multiple times. With this additional information, a more confident clustering of KHV isolates as well as molecular tracing now seem to be possible.

For KHV, 11 complete sequenced genomes are now available in the databases. While full length genomes are compulsory for phylogeographical investigations, the costs for next-generation sequencing, and the lack of handiness for the analysis of raw sequence data still hamper wide genome-based phylogeographical studies. To reduce these drawbacks, the duplex PCR was designed to distinguish between different types, variations or possible introductions of KHV to Europe (Bigarre et al., 2009). Unfortunately, this method was able to detect four different variations or types of KHV only. To overcome this possible gap, an MLVA-based method was established and included in this study. Based on eight different VNTR sequences, it was possible to distinguish more precisely the phylogeny of KHV. It was decided to use and sequence the VNTRs further on, even though they represent only 0.56% of the entire KHV genome. Additionally, by increasing the numbers of tested virus isolates and/or specimens by duplex PCR and VNTR-3 qPCR analyses, the informative value of the VNTRs will become strikingly more precise. In the case of the complete genome of the Chinese KHV isolate GZ11, an intermediate status appears in comparison to the European and Asian lineages. Complete genome analysis indicated that this Chinese virus had a phylogenetic position between European viruses like KHV-U or KHV-I and Asian types like KHV-J or KHV-T Considering the information obtained by duplex PCR, the isolate GZ 11 shares the same genotype as the attenuated AK-129. However, the phylogenies obtained from whole genome sequences and VNTR data indicate that GZ11 does not cluster together with AK-129. This shows that the markers used in the duplex PCR are not informative enough and may lead to incorrect conclusions. Detecting different variants of KHV in a single isolate or specimen is a major challenge for diagnostics. Combating the spread of KHV depends on these efforts. Moreover, different types of KHV may possess different pathogenic features which determine the viability of a population. Tracking Asian and European lineages by qPCR, the first cut was done for applied diagnostics. The importance is shown by the data (Figure 4) obtained with samples from KHVD outbreaks and with experimental samples. The most interesting samples are those which include both, the European and the Asian lineage, e.g., F02/05 from Poland (KHV-P), “atypically reacting” samples F34/13-2 and F11/16 from Germany and the KHV isolate from China (KHV-Ch, named also GZ for Guangzhou). These samples seem to have both features. This can either be the result of mixed infections with different KHV lineages in one isolate or of in vivo replication of the infecting agent that generates different variations or types, thereby increasing the likeliness of infection of a fish or a population at the same time. Mixed infections with differential KHV have already been reported, although further investigations are required to elucidate their ecological significance (Sunarto et al., 2011; Avarre et al., 2012; Li et al., 2015). However, a more important point is that this VNTR-3 qPCR neither detected CyHV-1, CyHV-2 nor HVA or CEV. Especially, discrimination of CEV from KHV is important because of the very similar symptoms in carp they cause and their cumulated emergence (Miyazaki et al., 2005; Jung-Schroers et al., 2015). The fact that CyHV-1 and CyHV-2 could only be detected with VNTR-2 primers indicates that the improved MLVA method does not recognize DNAs of other Cypriniviruses. This is especially important because of the high similarity of the three cyprinid viruses to each other as well as to the other species member HVA. By reducing the low specificity of the primer pair for VNTR-2 PCR, it should be possible to perform phylogenetic studies even with samples comprising more than one Cyprinivirus like KHV, carp pox virus, goldfish herpesvirus and/or HVA as a PAN-Cyprinivirus PCR.

The improved VNTR analysis was suitable to monitor the effects of KHV passages. During the 100 passages of KHV-T and KHV-E as well as KHV-Israel and KHV-G1 onto CCB cells at 20°C, it was possible to observe different changes in the VNTR sequences. Moreover, it was possible to detect a shift in lineage affiliation of KHV-E. After passage 78, the virus showed features of Asian KHV variants by both VNTR-3 qPCR and duplex PCR.

It is known from bacteria that VNTRs play an important role in adaptation, especially in relation to pathogenesis and virulence (van Belkum et al., 1997). Here, since the VNTRs are located in putative open reading frames, varying number of repeats may be a way to adapt to changing conditions by modifying the structural properties (i.e., transmembrane domains) of the corresponding proteins. Furthermore, it was found that changes in lineage affiliation of European KHV may occur. By both duplex PCR and VNTR-3 qPCR, shifts were visible. All three European-typed viruses shifted into the Asian lineage at different time points. Surprisingly, KHV-I was able to shift back, which was confirmed by duplex PCR. While with duplex PCR it was only possible to detect the lineage shift, only the VNTR-3 qPCR made it possible to see that in some cases, both Asian and European types were present. This indicated that these genetic variations were most likely due to the presence of several virus variations in the sample with one overcoming the other(s), of course the most virulent and fastest growing isolate. It is striking that only the European samples shifted into the other lineage but not the used Asian isolate KHV-T. In this case it must be clarified using other Asian type isolates that they are more stable in passaging. Nevertheless, most genetic variations were observed in European viruses. This may indicate that the Asian type is better adapted than the European types which may underline the fact that KHV was introduced to Europe from Asia. Since its introduction to Europe it needs to struggle with new ecological conditions like water temperature, seasonal cycles and changes, and different fish populations, which leads to an ongoing adaptation process. Moreover the phylogenic position of KHV GZ11 supports this possibility (Figure 2A). Because of its intermediate status, KHV GZ11 seems to be on the half way between the Asian cluster and the other European samples like KHV-U or KHV-I. But it still belongs to the European lineage.

Now it should be possible to detect different types of KHV more rapidly as well as to find possible subtypes or mixtures of KHV variants in a sample. However, this method is not yet an alternative for KHV diagnostics. It may however serve as an additional tool for deeper insights into KHV, its adaptation, its changes, and its distribution in these variations.

Beyond the phylogenetic analyses it might be possible to predict the virulence of a KHV with this method, if the appropriate marker genes are used. However, studies need to be performed to gain deeper insights into the virulence of KHV, especially which genes or modifications are important for high or low virulent virus variants.

Finally, this method does not only permit to monitor KHV in the wild and in carp or koi pond cultures but also to use it in cases of KHVD in these fish. Moreover, this technique has the potential to trace back introduced KHV inducing KHVD to their source. This might become useful, if Australia realizes its plan to eradicate or control the common carp population by using KHV (http://www.csiro.au).

SK and SB planned and performed experimental work. SK, QW, WZ, YW, YL, SZ, JK, PL, MM, and SB discussed the results.