- Email: [email protected]
Phylodynamics of Crimean Congo hemorrhagic fever virus in South Russia
Infection, Genetics and Evolution 59 (2018) 23–27
Contents lists available at ScienceDirect
Infection, Genetics and Evolution journal homepage: www.elsevier.com/locate/meegid
Opinion
Phylodynamics of Crimean Congo hemorrhagic fever virus in South Russia A.N. Lukashev a b c
a,c,⁎
, A.A. Deviatkin
T
b
Martsinovsky Institute of Medical Parasitology, Tropical and Vector Borne Diseases, Sechenov University, Moscow, Russia Institute of Molecular Medicine, Sechenov University, Moscow, Russia Chumakov Federal Scientific Center for Research and Development of Immune-and-Biological Preparations, Moscow, Russia
A R T I C L E I N F O
A B S T R A C T
Keywords: CCHFV Phylodynamics Reassortment
Phylodynamics of Crimean Congo Hemorrhagic fever virus (CCHFV) genotype V in South Russia was analyzed using 244 partial (452–571 nt) sequences in all three genomic segments and 38 complete genomic sequences. Despite increased number of sequences, the Russian lineage of the European genotype V (commonly termed GtVa) was distinct from GtV isolates from Turkey and the Balkan countries. No geographic pattern was observed in phylogenetic subgrouping of CCHFV within South Russia. Identical isolates could be found at distant locations spaced by hundreds of kilometers, while relatively divergent viruses circulated in the same region. Full genome analysis indicated that reassortment events within GtVa occurred every few decades (median half-life of a nonreassortant node 30–40 years) and involved M and S segments. Therefore, in South Russia CCHFV represents a highly dynamic population of frequently reassorting viruses.
1. Introduction Crimean Congo hemorrhagic fever (CCHF) is a severe tick-borne zoonosis. Fatality rates are usually around 1–5%, but can reach 30% upon human-to-human transmission (Bente et al., 2013). The virus is a significant health concern in Africa, Asia and Southern Europe. The CCHF virus (CCHFV) is a member of the Bunyavirales order, the Nairoviridae family and the Orthonairovirus genus. It is an enveloped virus with three RNA segments: S (small), around 1.7 Kb, which encodes nucleoprotein; M (medium), around 4.5 Kb, which encodes two glycoproteins and a mucin-like domain; and L (large), 12.1 Kb, which encodes polymerase. The S segment is predominantly used for phylogenetic studies. Analysis of CCHFV strains collected worldwide shows a clear geographic pattern of CCHFV genotypes (Deyde et al., 2006; Hewson et al., 2004a). Reassortment of CCHFV genome segments is well documented (Hewson et al., 2004b) and was estimated to occur approximately once every 100 years (Lukashev et al., 2016). As a result, the phylogenetic relations of viruses and the genotype systems may differ between segments. Virus lineage that is predominant in Southern Europe (South Russia, Turkey, Bulgaria, Albania) was termed genotype V (GtV) in all genome segments (Deyde et al., 2006). There is evidence of reassortment within GtV, but not between GtV and other genotypes (Lukashev et al., 2016). Rarely, divergent isolates of GtVI were also found in Europe (Sherifi et al., 2014). Phylogeographic studies of the “European” GtV indicate higher virus diversity in Turkey and South Russia than in Balkan countries, and
⁎
assume westward virus spread, which is also consistent with epidemiological data. Bayesian coalescent phylogenetic analysis of partial S segment sequences estimated that the most recent common ancestor of GtV existed 82 (95% high probability density interval: 52–112) years ago, while a study of full-genome sequences yielded more ancient dates, ranging from 248 (126–400) to 402(289–538) years ago in different genome segments (Lukashev et al., 2016) or 257–517 years ago (Carroll et al., 2010). The likely location of GtV emergence was inferred to be South Russia (Lukashev et al., 2016; Zehender et al., 2013). A further subdivision of GtV into subtypes Va (South Russia) and Vb-Ve (Turkey and the Balkans) was suggested (Zehender et al., 2013). Complete CCHFV genomes provide good resolution of phylogenetic methods (robust statistical support for most tree nodes), but only about 80 are available worldwide. Short genome fragments may lack resolution on a short timescale, because the virus accumulates mutations at a rate of approximately 1 × 10−4–3 × 10−4 substitutions/site/year (Carroll et al., 2010; Lukashev, 2005; Sherifi et al., 2014), and thus about one mutation per 100 nt in 30–100 years. In 2016, 21 complete genomes (KR814833-KR814893, KU161582KU161587) and partial sequences in all three genome segments of over 440 CCHFV samples from South Russia (S segment: KR814894KR815339, positions 115–652 according to AF428144; M segment: KU161576-KU161581, positions 4640–5074 according to KR814873 and L segment: positions 110–546 according to KX013468) were produced by the Stavropol Antiplague Scientific Research Institute of the State Sanitary Surveillance of Russia. Sequences KU161582–KU161587
Corresponding author at: Martsinovsky Institute of Medical Parasitology, Tropical and Vector Borne Diseases, Malaya Pirogovskaya 20-1, Moscow 119435, Russia. E-mail address: [email protected] (A.N. Lukashev).
https://doi.org/10.1016/j.meegid.2018.01.016 Received 1 November 2017; Received in revised form 19 January 2018; Accepted 21 January 2018 1567-1348/ © 2018 Elsevier B.V. All rights reserved.
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
Fig. 1. Incidence of CCHF in 2015 (2014) and the number of unique sequences available in Genbank from the endemic regions of South Russia. Total incidence was taken from annual reports on infectious disease incidence by the Russian Federal Service for Surveillance on Consumer Rights Protection and Human Wellbeing - Rospotrebnadzor - (Rospotrebnadzor, 2016). Dots correspond to branch colors on the phylogenetic tree (Fig. 2a). The map was obtained using ArcGIS software [Esri Inc., version 10.4.1, Redlands, California, USA].
coalescent analysis. Next, different clock assumptions and population models were compared by a Bayes factors test. The highest Bayes factor was observed in the combination of the uncorrelated lognormal relaxed clock and the exponential population model in all segments, although there were no significant differences between the estimates for other combinations of two clock (strict and relaxed clock) and two population models (constant and exponential). Trees were sampled every 10,000 (5000; 3000) generations for 100 (50; 30) million generations in total and annotated with a burn-in of 10 (5; 3) million generations in L, M and S segments, respectively. The convergence of parameter estimates was checked using Tracer (v1.5) and indicated by an effective sample size > 200.
were analysed previously (Kulichenko et al., 2016). These new sequences allowed for the investigation of CCHFV phylodynamics with an unprecedented resolution. CCHF is endemic in a large part of South Russia between the Black sea on the West and the Caspian sea on the East (about 700 km), and between Caucasian mountains on the South and Volgograd region on the North (about 1000 km). The annual number of reported cases varied between 80 and 162 in 2013-2016 (Fig. 1). Most cases are sporadic; however outbreaks involving up to 40 epidemiologically related cases have also occurred (Onishchenko et al., 2001).
2. Materials and methods 3. Results and discussion
As of December 2016, there were 487 partial S segment sequences from South Russia available in Genbank. To simplify analysis and data presentation, redundant entries (sequences of the same isolates or 100% identical sequences from the same region) were omitted. Importantly, sequences of several isolates were published independently by different laboratories and were identical, indicating high data quality. The sequences (genome positions 150–652 according to AF428144) were aligned with homologous Genbank reference sequences of GtV from other countries, also omitting identical sequences. The final data set contained 244 sequences (Online resource 1). For reassortment analysis, all available complete genome sequences of viruses isolated in Russia were used (as of December 2016). Samples without known collection dates were omitted. Strains Kelkit06 (Turkey) and Hoti (Kosovo) were added to the data set to root the trees. The multiple sequence alignment was performed using MAFFT server (Kuraku et al., 2013). Only protein-coding regions were used for the analysis. The final data set consisted of 38 sequences. A reversible jump-based substitution model (Bouckaert et al., 2014) was used to choose the substitution model and estimate the appropriate number of parameters, while sampling the tree was used for Bayesian
As expected, phylogenetic analysis of 244 relatively short and very similar sequences could not produce a robust bootstrap support for most tree nodes. Nevertheless, several important conclusions could be drawn. South Russian isolates were distinct from viruses found in Turkey and the Balkans. (Fig. 2). Despite a hugely increased dataset, there was no evidence of virus transfers between South Russia and Turkey-Balkan region; however, viruses from both these areas were occasionally introduced to Iran (Fig. 2). Isolates from distinct regions of South Russia (sometimes hundreds of kilometers apart, Fig. 1) were intermixed. There were a number of occasions when identical isolates could be obtained in distinct (and often non-adjacent) regions. The most vivid examples of such virus distribution, circled in Fig. 2, included five or six identical strains from the Astrakhan, Rostov, Stavropol, Krasnodar, Kalmykia and Volgograd regions. Similar conclusions could be drawn regarding M and L segment analysis; however, as fewer sequences of these regions were available from Turkey and Southern Europe, this data is not shown. 24
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
Fig. 2. Phylogenetic reconstruction for partial S segment sequence (positions 150–652 according to AF428144) of 217 unique CCHFV GtVa strains (all available unique sequences) and 27 reference sequences of GtV isolates. The tree was inferred using the Maximum Likelihood algorithm and the uncorrected nucleotide distance in MEGA 7 (Kumar et al., 2016). The final tree was split vertically to improve visualization. Label colors highlight the isolation area according to Fig. 1. Red circles indicate groups of more than two identical sequences isolated in different administrative regions of South Russia. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
European CCHFV strains (Anagnostou and Papa, 2009). This substitution rate inferred for South Russia isolates corresponds to roughly one substitution every 30–40 years within the analyzed 503 nt genome fragment. Of course, such estimates are very imprecise because the uncertainty of Bayesian analysis is combined with the stochastic nature of single substitutions. Given the range of rate values obtained in different studies (see above), one substitution in 503 nt could correspond to 10–70 years of circulation. Therefore, the virus was commonly carried over hundreds of kilometers within decades. Two mechanisms could provide such rapid virus spread: birds that can carry infected ticks and transportation of livestock carrying feeding adult ticks, which can indeed serve as a CCHFV spread vehicle (Chisholm et al., 2012). No virus transfers were observed between Russia and Turkey, although the transboundary migration of some birds was described (Elphick and Lovejoy, 2011); therefore the explanation for virus spread within a region involving livestock transportation associated with socioeconomic activity seems more plausible. As a result of such frequent long-distance transfers, there was no
Several studies have estimated the substitution rate of CCHFV to be between 1 × 10−4 and 3 × 10−4 substitutions/site/year (Carroll et al., 2010; Lukashev, 2005; Sherifi et al., 2014). According to our results (38 complete genome sequences from South Russia, Fig. 3), the substitution rates of S, M and L segments were approximately 0.60 × 10−4, 0.92 × 10−4 and 0.64 × 10−4 s/s/y, respectively. It is noteworthy that consistently higher substitution rates were observed for M segment compared to S and L segments here and in previous studies (Anagnostou and Papa, 2009; Carroll et al., 2010; Lukashev et al., 2016). The inferred rates were about two times lower than estimates from previous studies using the same method on the CCHFV sequences sampled globally (Carroll et al., 2010; Lukashev et al., 2016). This could be explained by small sample size in this study, which can lead to lower inferred substitution rates (Lukashev et al., 2016), or by the fact that the climate in South Russia limits virus transmission in winter. On the other hand, these rates were twice higher than the estimates by TreePuzzle software (0.28 × 10−4) that were obtained for S segment of the 25
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
Fig. 3. Phylogenetic trees of the complete L, M, S segments of GtV CCHFV. Only viruses with a complete sequence available for all three genome segments were used. Green circles indicate nodes that were supported in L segment (posterior probability > 0.95) and were not reliably reassortant (group supported or not reliably rejected in the S and M genome regions). Red squares indicate reliably supported reassortment events (conflicting grouping observed in M or S segments). All nodes above a reassortant node were also considered reassortant. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
is fixed on average once every 30–40 years in a 500 nt fragment (see above), which would not allow distinguishing more frequent reassortment events. In the whole S segment, one substitution is fixed on average every 7–13 years, which should provide better resolution of future reassortment analysis. Early analysis suggested that there was probably no reassortment between S and L segments (Hewson et al., 2004b). Later studies showed that reassortment involving any segment was possible (Deyde et al., 2006; Lukashev et al., 2016; Zhou et al., 2013). Among six nodes where the reassortment pattern could be established, four resulted from introduction of M segment to a preserved pair of L and S segments, while two reassortment events were a consequence of S segment introduction into a conserved pair of M and L segments. In a similar study of the South African CCHFV isolates, only reassortment of M segment relative to conserved S + L segment was observed (Goedhals et al., 2014), while a study that used a global sequence dataset found approximately equal reassortment of S and M segments and a less common reassortment of the L segment (Zhou et al., 2013). It is therefore doubtful that a suggestive fine adaptation of S and L segments involved in replication could explain relatively higher conservation of S + L segment pairing observed here and in other studies, especially given that many reassortment events in our data set involved viruses that differed by just a few third-position nucleotide positions. On the other hand, higher reassortment of M segment can be one of the factors that resulted in higher inferred substitution rates (see above).
clear geographic pattern in sequence grouping within the South Russian region, which supposedly emerged only few hundred years ago. The availability of novel sequences also allowed for a more precise reassortment frequency investigation. Classical analysis relies on a robust phylogenetic grouping support; therefore, only complete genomic sequences could be used. Recently, we reported that the median halflife of a non-reassortant node in full-genome CCHFV phylogenetic trees was about 100 years, and that this number could decrease with the increasing sample (Lukashev et al., 2016). The number of available GtV full-genome sequences almost doubled in 2016, which allowed for a more precise analysis. Phylogenetic trees were reconstructed for all 38 GtVa isolates, for which complete sequences were available in all three segments, using Bayesian coalescent analysis with integrated molecular clock (Fig. 3). Virus grouping was compared across segments. A tree node was considered reassortant if a reliably supported grouping (posterior probability > 0.95) in one segment was converted into a conflicting reliably supported grouping in another segment. In other cases (in which the node was conserved across all segments or converted into a conflicting but unsupported topology), the node was considered non-reassortant. Node height was measured from the most recent isolate within the node. Despite using the most conservative approach, the reassortment was very frequent within the South Russian GtV subgroup. As follows from the Bayesian phylogenetic analysis, the median height (from the most recent isolate) of a non-reassortant node was, respectively, 34 (range: 6–67), 30 (range: 7–55) and 40 (range: 3–58) years for S, M and L segments. Importantly, further refinement of this result would be possible only using long sequences. One substitution (which is generally not sufficient for a robust phylogenetic support) 26
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
4. Conclusion
2013. Crimean-Congo hemorrhagic fever: history, epidemiology, pathogenesis, clinical syndrome and genetic diversity. Antivir. Res. 100, 159–189. Bouckaert, R., Heled, J., Kuhnert, D., Vaughan, T., Wu, C.H., Xie, D., Suchard, M.A., Rambaut, A., Drummond, A.J., 2014. BEAST 2: a software platform for Bayesian evolutionary analysis. PLoS Comput. Biol. 10, e1003537. Carroll, S.A., Bird, B.H., Rollin, P.E., Nichol, S.T., 2010. Ancient common ancestry of Crimean-Congo hemorrhagic fever virus. Mol. Phylogenet. Evol. 55, 1103–1110. Chisholm, K., Dueger, E., Fahmy, N.T., Samaha, H.A., Zayed, A., Abdel-Dayem, M., Villinski, J.T., 2012. Crimean-Congo hemorrhagic fever virus in ticks from imported livestock, Egypt. Emerg. Infect. Dis. 18, 181–182. Deyde, V.M., Khristova, M.L., Rollin, P.E., Ksiazek, T.G., Nichol, S.T., 2006. CrimeanCongo hemorrhagic fever virus genomics and global diversity. J. Virol. 80, 8834–8842. Elphick, J., Lovejoy, T.E., 2011. The Atlas of Bird Migration: Tracing the Great Journeys of the World's Birds. Firefly Books, Richmond Hill, Canada. Estrada-Pena, A., Palomar, A.M., Santibanez, P., Sanchez, N., Habela, M.A., Portillo, A., Romero, L., Oteo, J.A., 2012. Crimean-Congo hemorrhagic fever virus in ticks, Southwestern Europe, 2010. Emerg. Infect. Dis. 18, 179–180. Goedhals, D., Bester, P.A., Paweska, J.T., Swanepoel, R., Burt, F.J., 2014. Next-generation sequencing of southern African Crimean-Congo haemorrhagic fever virus isolates reveals a high frequency of M segment reassortment. Epidemiol. Infect. 142, 1952–1962. Hewson, R., Chamberlain, J., Mioulet, V., Lloyd, G., Jamil, B., Hasan, R., Gmyl, A., Gmyl, L., Smirnova, S.E., Lukashev, A., Karganova, G., Clegg, C., 2004a. Crimean-Congo haemorrhagic fever virus: sequence analysis of the small RNA segments from a collection of viruses world wide. Virus Res. 102, 185–189. Hewson, R., Gmyl, A., Gmyl, L., Smirnova, S.E., Karganova, G., Jamil, B., Hasan, R., Chamberlain, J., Clegg, C., 2004b. Evidence of segment reassortment in CrimeanCongo haemorrhagic fever virus. J. Gen. Virol. 85, 3059–3070. Kulichenko, A.N., Volynkina, A.S., Kotenev, E.S., Pisarenko, S.V., Shaposhnikova, L.I., Lisitskaya, Y.V., Vasilenko, N.F., Tsygankova, O.I., Evchenko, Y.M., Tohov, Y.M., Savel'ev, V.N., Tihonov, S.N., Penkovskaya, N.A., 2016. A new genetic variant of the Crimean–Congo hemorrhagic fever virus isolated in Crimea. Mol. Genet. Microbiol. Virol. 31, 94–101. Kumar, S., Stecher, G., Tamura, K., 2016. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874. Kuraku, S., Zmasek, C.M., Nishimura, O., Katoh, K., 2013. aLeaves facilitates on-demand exploration of metazoan gene family trees on MAFFT sequence alignment server with enhanced interactivity. Nucleic Acids Res. 41, W22–28. Lukashev, A.N., 2005. Evidence for recombination in Crimean-Congo hemorrhagic fever virus. J. Gen. Virol. 86, 2333–2338. Lukashev, A.N., Klimentov, A.S., Smirnova, S.E., Dzagurova, T.K., Drexler, J.F., Gmyl, A.P., 2016. Phylogeography of Crimean Congo hemorrhagic fever virus. PLoS One 11, e0166744. Onishchenko, G.G., Markov, V.I., Merkulov, V.A., Vasil'ev, N.T., Berezhnoi, A.M., Androshchuk, I.A., Maksimov, V.A., 2001. [Isolation and Identification of CrimeanCongo Hemorrhagic Fever Virus in the Stavropol Territory]. Zhurnal mikrobiologii, epidemiologii, i immunobiologii. pp. 7–11. Rospotrebnadzor, 2016. Letter №01/500-16-32 “On Crimean Hemorrhagic Fever Epidemiological Situation in Southern, North Caucasian and Crimean Federal Districts of Russia Federation and Forecast for 2016”. Russian Federal Service for Surveillance on Consumer Rights Protection and Human Wellbeing (Rospotrebnadzor), Moscow. Sherifi, K., Cadar, D., Muji, S., Robaj, A., Ahmeti, S., Jakupi, X., Emmerich, P., Kruger, A., 2014. Crimean-Congo hemorrhagic fever virus clades V and VI (Europe 1 and 2) in ticks in Kosovo, 2012. PLoS Negl. Trop. Dis. 8, e3168. Zehender, G., Ebranati, E., Shkjezi, R., Papa, A., Luzzago, C., Gabanelli, E., Lo Presti, A., Lai, A., Rezza, G., Galli, M., Bino, S., Ciccozzi, M., 2013. Bayesian phylogeography of Crimean-Congo hemorrhagic fever virus in Europe. PLoS One 8, e79663. Zhou, Z., Deng, F., Han, N., Wang, H., Sun, S., Zhang, Y., Hu, Z., Rayner, S., 2013. Reassortment and migration analysis of Crimean-Congo haemorrhagic fever virus. J. Gen. Virol. 94, 2536–2548.
Analysis of recent GtVa CCHFV sequences from South Russia indicates that the virus population is highly dynamic in terms of longdistance virus transmission and reassortment. The European genotype V of CCHFV emerged just 100–500 years ago according to various estimates. Within the South Russia, genetic variants of GtVa from different areas were intermixed, and identical viruses could be isolated from distant locations. Therefore, we can conclude that numerous long-distance transfers occurred within South Russia over a few hundred years. On the other hand, there was no evidence of recent virus transfers between South Russia and Turkey, and only one transfer from Russia to Iran could be suggested. Apparently, ecological barriers (Caucasian mountains that split the habitat of steppe ticks) or socio-economic barriers (borders that limit livestock transport) are poorly penetrable to the virus. In the absence of such barriers, the high efficiency of virus spread observed in South Russia implies that a CCHFV focus, once established, can grow and occupy the whole suitable niche over several decades. This observation is especially important for predicting CCHFV spread in Southern Europe, where the virus is spreading West-ward from South Russia/Turkey (Zehender et al., 2013) and can be introduced by migratory birds (Estrada-Pena et al., 2012). Within a focus, CCHFV undergoes reassortment at least once every few decades. A detectable reassortment implies common co-infection by two phylogenetically distinct virus variants; thus, frequent reassortment further highlights a high frequency of virus intermixing within a region. Therefore, even on a timescale of decades, the virus population within the South Russia exists as a cloud of swiftly spreading and frequently reassorting genome segments. It is currently unknown if the co-infection required for reassortment occurs in a tick or in a mammal, and both hypotheses can be justified. High-throughput sequencing of virus populations can answer this question in the future. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.meegid.2018.01.016. Funding This study was supported by the Russian Science Foundation Grant 14-15-00619. Competing interests The authors declare that they have no competing interests. References Anagnostou, V., Papa, A., 2009. Evolution of Crimean-Congo hemorrhagic fever virus. Infect. Genet. Evol. 9, 948–954. Bente, D.A., Forrester, N.L., Watts, D.M., McAuley, A.J., Whitehouse, C.A., Bray, M.,
27
Contents lists available at ScienceDirect
Infection, Genetics and Evolution journal homepage: www.elsevier.com/locate/meegid
Opinion
Phylodynamics of Crimean Congo hemorrhagic fever virus in South Russia A.N. Lukashev a b c
a,c,⁎
, A.A. Deviatkin
T
b
Martsinovsky Institute of Medical Parasitology, Tropical and Vector Borne Diseases, Sechenov University, Moscow, Russia Institute of Molecular Medicine, Sechenov University, Moscow, Russia Chumakov Federal Scientific Center for Research and Development of Immune-and-Biological Preparations, Moscow, Russia
A R T I C L E I N F O
A B S T R A C T
Keywords: CCHFV Phylodynamics Reassortment
Phylodynamics of Crimean Congo Hemorrhagic fever virus (CCHFV) genotype V in South Russia was analyzed using 244 partial (452–571 nt) sequences in all three genomic segments and 38 complete genomic sequences. Despite increased number of sequences, the Russian lineage of the European genotype V (commonly termed GtVa) was distinct from GtV isolates from Turkey and the Balkan countries. No geographic pattern was observed in phylogenetic subgrouping of CCHFV within South Russia. Identical isolates could be found at distant locations spaced by hundreds of kilometers, while relatively divergent viruses circulated in the same region. Full genome analysis indicated that reassortment events within GtVa occurred every few decades (median half-life of a nonreassortant node 30–40 years) and involved M and S segments. Therefore, in South Russia CCHFV represents a highly dynamic population of frequently reassorting viruses.
1. Introduction Crimean Congo hemorrhagic fever (CCHF) is a severe tick-borne zoonosis. Fatality rates are usually around 1–5%, but can reach 30% upon human-to-human transmission (Bente et al., 2013). The virus is a significant health concern in Africa, Asia and Southern Europe. The CCHF virus (CCHFV) is a member of the Bunyavirales order, the Nairoviridae family and the Orthonairovirus genus. It is an enveloped virus with three RNA segments: S (small), around 1.7 Kb, which encodes nucleoprotein; M (medium), around 4.5 Kb, which encodes two glycoproteins and a mucin-like domain; and L (large), 12.1 Kb, which encodes polymerase. The S segment is predominantly used for phylogenetic studies. Analysis of CCHFV strains collected worldwide shows a clear geographic pattern of CCHFV genotypes (Deyde et al., 2006; Hewson et al., 2004a). Reassortment of CCHFV genome segments is well documented (Hewson et al., 2004b) and was estimated to occur approximately once every 100 years (Lukashev et al., 2016). As a result, the phylogenetic relations of viruses and the genotype systems may differ between segments. Virus lineage that is predominant in Southern Europe (South Russia, Turkey, Bulgaria, Albania) was termed genotype V (GtV) in all genome segments (Deyde et al., 2006). There is evidence of reassortment within GtV, but not between GtV and other genotypes (Lukashev et al., 2016). Rarely, divergent isolates of GtVI were also found in Europe (Sherifi et al., 2014). Phylogeographic studies of the “European” GtV indicate higher virus diversity in Turkey and South Russia than in Balkan countries, and
⁎
assume westward virus spread, which is also consistent with epidemiological data. Bayesian coalescent phylogenetic analysis of partial S segment sequences estimated that the most recent common ancestor of GtV existed 82 (95% high probability density interval: 52–112) years ago, while a study of full-genome sequences yielded more ancient dates, ranging from 248 (126–400) to 402(289–538) years ago in different genome segments (Lukashev et al., 2016) or 257–517 years ago (Carroll et al., 2010). The likely location of GtV emergence was inferred to be South Russia (Lukashev et al., 2016; Zehender et al., 2013). A further subdivision of GtV into subtypes Va (South Russia) and Vb-Ve (Turkey and the Balkans) was suggested (Zehender et al., 2013). Complete CCHFV genomes provide good resolution of phylogenetic methods (robust statistical support for most tree nodes), but only about 80 are available worldwide. Short genome fragments may lack resolution on a short timescale, because the virus accumulates mutations at a rate of approximately 1 × 10−4–3 × 10−4 substitutions/site/year (Carroll et al., 2010; Lukashev, 2005; Sherifi et al., 2014), and thus about one mutation per 100 nt in 30–100 years. In 2016, 21 complete genomes (KR814833-KR814893, KU161582KU161587) and partial sequences in all three genome segments of over 440 CCHFV samples from South Russia (S segment: KR814894KR815339, positions 115–652 according to AF428144; M segment: KU161576-KU161581, positions 4640–5074 according to KR814873 and L segment: positions 110–546 according to KX013468) were produced by the Stavropol Antiplague Scientific Research Institute of the State Sanitary Surveillance of Russia. Sequences KU161582–KU161587
Corresponding author at: Martsinovsky Institute of Medical Parasitology, Tropical and Vector Borne Diseases, Malaya Pirogovskaya 20-1, Moscow 119435, Russia. E-mail address: [email protected] (A.N. Lukashev).
https://doi.org/10.1016/j.meegid.2018.01.016 Received 1 November 2017; Received in revised form 19 January 2018; Accepted 21 January 2018 1567-1348/ © 2018 Elsevier B.V. All rights reserved.
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
Fig. 1. Incidence of CCHF in 2015 (2014) and the number of unique sequences available in Genbank from the endemic regions of South Russia. Total incidence was taken from annual reports on infectious disease incidence by the Russian Federal Service for Surveillance on Consumer Rights Protection and Human Wellbeing - Rospotrebnadzor - (Rospotrebnadzor, 2016). Dots correspond to branch colors on the phylogenetic tree (Fig. 2a). The map was obtained using ArcGIS software [Esri Inc., version 10.4.1, Redlands, California, USA].
coalescent analysis. Next, different clock assumptions and population models were compared by a Bayes factors test. The highest Bayes factor was observed in the combination of the uncorrelated lognormal relaxed clock and the exponential population model in all segments, although there were no significant differences between the estimates for other combinations of two clock (strict and relaxed clock) and two population models (constant and exponential). Trees were sampled every 10,000 (5000; 3000) generations for 100 (50; 30) million generations in total and annotated with a burn-in of 10 (5; 3) million generations in L, M and S segments, respectively. The convergence of parameter estimates was checked using Tracer (v1.5) and indicated by an effective sample size > 200.
were analysed previously (Kulichenko et al., 2016). These new sequences allowed for the investigation of CCHFV phylodynamics with an unprecedented resolution. CCHF is endemic in a large part of South Russia between the Black sea on the West and the Caspian sea on the East (about 700 km), and between Caucasian mountains on the South and Volgograd region on the North (about 1000 km). The annual number of reported cases varied between 80 and 162 in 2013-2016 (Fig. 1). Most cases are sporadic; however outbreaks involving up to 40 epidemiologically related cases have also occurred (Onishchenko et al., 2001).
2. Materials and methods 3. Results and discussion
As of December 2016, there were 487 partial S segment sequences from South Russia available in Genbank. To simplify analysis and data presentation, redundant entries (sequences of the same isolates or 100% identical sequences from the same region) were omitted. Importantly, sequences of several isolates were published independently by different laboratories and were identical, indicating high data quality. The sequences (genome positions 150–652 according to AF428144) were aligned with homologous Genbank reference sequences of GtV from other countries, also omitting identical sequences. The final data set contained 244 sequences (Online resource 1). For reassortment analysis, all available complete genome sequences of viruses isolated in Russia were used (as of December 2016). Samples without known collection dates were omitted. Strains Kelkit06 (Turkey) and Hoti (Kosovo) were added to the data set to root the trees. The multiple sequence alignment was performed using MAFFT server (Kuraku et al., 2013). Only protein-coding regions were used for the analysis. The final data set consisted of 38 sequences. A reversible jump-based substitution model (Bouckaert et al., 2014) was used to choose the substitution model and estimate the appropriate number of parameters, while sampling the tree was used for Bayesian
As expected, phylogenetic analysis of 244 relatively short and very similar sequences could not produce a robust bootstrap support for most tree nodes. Nevertheless, several important conclusions could be drawn. South Russian isolates were distinct from viruses found in Turkey and the Balkans. (Fig. 2). Despite a hugely increased dataset, there was no evidence of virus transfers between South Russia and Turkey-Balkan region; however, viruses from both these areas were occasionally introduced to Iran (Fig. 2). Isolates from distinct regions of South Russia (sometimes hundreds of kilometers apart, Fig. 1) were intermixed. There were a number of occasions when identical isolates could be obtained in distinct (and often non-adjacent) regions. The most vivid examples of such virus distribution, circled in Fig. 2, included five or six identical strains from the Astrakhan, Rostov, Stavropol, Krasnodar, Kalmykia and Volgograd regions. Similar conclusions could be drawn regarding M and L segment analysis; however, as fewer sequences of these regions were available from Turkey and Southern Europe, this data is not shown. 24
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
Fig. 2. Phylogenetic reconstruction for partial S segment sequence (positions 150–652 according to AF428144) of 217 unique CCHFV GtVa strains (all available unique sequences) and 27 reference sequences of GtV isolates. The tree was inferred using the Maximum Likelihood algorithm and the uncorrected nucleotide distance in MEGA 7 (Kumar et al., 2016). The final tree was split vertically to improve visualization. Label colors highlight the isolation area according to Fig. 1. Red circles indicate groups of more than two identical sequences isolated in different administrative regions of South Russia. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
European CCHFV strains (Anagnostou and Papa, 2009). This substitution rate inferred for South Russia isolates corresponds to roughly one substitution every 30–40 years within the analyzed 503 nt genome fragment. Of course, such estimates are very imprecise because the uncertainty of Bayesian analysis is combined with the stochastic nature of single substitutions. Given the range of rate values obtained in different studies (see above), one substitution in 503 nt could correspond to 10–70 years of circulation. Therefore, the virus was commonly carried over hundreds of kilometers within decades. Two mechanisms could provide such rapid virus spread: birds that can carry infected ticks and transportation of livestock carrying feeding adult ticks, which can indeed serve as a CCHFV spread vehicle (Chisholm et al., 2012). No virus transfers were observed between Russia and Turkey, although the transboundary migration of some birds was described (Elphick and Lovejoy, 2011); therefore the explanation for virus spread within a region involving livestock transportation associated with socioeconomic activity seems more plausible. As a result of such frequent long-distance transfers, there was no
Several studies have estimated the substitution rate of CCHFV to be between 1 × 10−4 and 3 × 10−4 substitutions/site/year (Carroll et al., 2010; Lukashev, 2005; Sherifi et al., 2014). According to our results (38 complete genome sequences from South Russia, Fig. 3), the substitution rates of S, M and L segments were approximately 0.60 × 10−4, 0.92 × 10−4 and 0.64 × 10−4 s/s/y, respectively. It is noteworthy that consistently higher substitution rates were observed for M segment compared to S and L segments here and in previous studies (Anagnostou and Papa, 2009; Carroll et al., 2010; Lukashev et al., 2016). The inferred rates were about two times lower than estimates from previous studies using the same method on the CCHFV sequences sampled globally (Carroll et al., 2010; Lukashev et al., 2016). This could be explained by small sample size in this study, which can lead to lower inferred substitution rates (Lukashev et al., 2016), or by the fact that the climate in South Russia limits virus transmission in winter. On the other hand, these rates were twice higher than the estimates by TreePuzzle software (0.28 × 10−4) that were obtained for S segment of the 25
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
Fig. 3. Phylogenetic trees of the complete L, M, S segments of GtV CCHFV. Only viruses with a complete sequence available for all three genome segments were used. Green circles indicate nodes that were supported in L segment (posterior probability > 0.95) and were not reliably reassortant (group supported or not reliably rejected in the S and M genome regions). Red squares indicate reliably supported reassortment events (conflicting grouping observed in M or S segments). All nodes above a reassortant node were also considered reassortant. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
is fixed on average once every 30–40 years in a 500 nt fragment (see above), which would not allow distinguishing more frequent reassortment events. In the whole S segment, one substitution is fixed on average every 7–13 years, which should provide better resolution of future reassortment analysis. Early analysis suggested that there was probably no reassortment between S and L segments (Hewson et al., 2004b). Later studies showed that reassortment involving any segment was possible (Deyde et al., 2006; Lukashev et al., 2016; Zhou et al., 2013). Among six nodes where the reassortment pattern could be established, four resulted from introduction of M segment to a preserved pair of L and S segments, while two reassortment events were a consequence of S segment introduction into a conserved pair of M and L segments. In a similar study of the South African CCHFV isolates, only reassortment of M segment relative to conserved S + L segment was observed (Goedhals et al., 2014), while a study that used a global sequence dataset found approximately equal reassortment of S and M segments and a less common reassortment of the L segment (Zhou et al., 2013). It is therefore doubtful that a suggestive fine adaptation of S and L segments involved in replication could explain relatively higher conservation of S + L segment pairing observed here and in other studies, especially given that many reassortment events in our data set involved viruses that differed by just a few third-position nucleotide positions. On the other hand, higher reassortment of M segment can be one of the factors that resulted in higher inferred substitution rates (see above).
clear geographic pattern in sequence grouping within the South Russian region, which supposedly emerged only few hundred years ago. The availability of novel sequences also allowed for a more precise reassortment frequency investigation. Classical analysis relies on a robust phylogenetic grouping support; therefore, only complete genomic sequences could be used. Recently, we reported that the median halflife of a non-reassortant node in full-genome CCHFV phylogenetic trees was about 100 years, and that this number could decrease with the increasing sample (Lukashev et al., 2016). The number of available GtV full-genome sequences almost doubled in 2016, which allowed for a more precise analysis. Phylogenetic trees were reconstructed for all 38 GtVa isolates, for which complete sequences were available in all three segments, using Bayesian coalescent analysis with integrated molecular clock (Fig. 3). Virus grouping was compared across segments. A tree node was considered reassortant if a reliably supported grouping (posterior probability > 0.95) in one segment was converted into a conflicting reliably supported grouping in another segment. In other cases (in which the node was conserved across all segments or converted into a conflicting but unsupported topology), the node was considered non-reassortant. Node height was measured from the most recent isolate within the node. Despite using the most conservative approach, the reassortment was very frequent within the South Russian GtV subgroup. As follows from the Bayesian phylogenetic analysis, the median height (from the most recent isolate) of a non-reassortant node was, respectively, 34 (range: 6–67), 30 (range: 7–55) and 40 (range: 3–58) years for S, M and L segments. Importantly, further refinement of this result would be possible only using long sequences. One substitution (which is generally not sufficient for a robust phylogenetic support) 26
Infection, Genetics and Evolution 59 (2018) 23–27
A.N. Lukashev, A.A. Deviatkin
4. Conclusion
2013. Crimean-Congo hemorrhagic fever: history, epidemiology, pathogenesis, clinical syndrome and genetic diversity. Antivir. Res. 100, 159–189. Bouckaert, R., Heled, J., Kuhnert, D., Vaughan, T., Wu, C.H., Xie, D., Suchard, M.A., Rambaut, A., Drummond, A.J., 2014. BEAST 2: a software platform for Bayesian evolutionary analysis. PLoS Comput. Biol. 10, e1003537. Carroll, S.A., Bird, B.H., Rollin, P.E., Nichol, S.T., 2010. Ancient common ancestry of Crimean-Congo hemorrhagic fever virus. Mol. Phylogenet. Evol. 55, 1103–1110. Chisholm, K., Dueger, E., Fahmy, N.T., Samaha, H.A., Zayed, A., Abdel-Dayem, M., Villinski, J.T., 2012. Crimean-Congo hemorrhagic fever virus in ticks from imported livestock, Egypt. Emerg. Infect. Dis. 18, 181–182. Deyde, V.M., Khristova, M.L., Rollin, P.E., Ksiazek, T.G., Nichol, S.T., 2006. CrimeanCongo hemorrhagic fever virus genomics and global diversity. J. Virol. 80, 8834–8842. Elphick, J., Lovejoy, T.E., 2011. The Atlas of Bird Migration: Tracing the Great Journeys of the World's Birds. Firefly Books, Richmond Hill, Canada. Estrada-Pena, A., Palomar, A.M., Santibanez, P., Sanchez, N., Habela, M.A., Portillo, A., Romero, L., Oteo, J.A., 2012. Crimean-Congo hemorrhagic fever virus in ticks, Southwestern Europe, 2010. Emerg. Infect. Dis. 18, 179–180. Goedhals, D., Bester, P.A., Paweska, J.T., Swanepoel, R., Burt, F.J., 2014. Next-generation sequencing of southern African Crimean-Congo haemorrhagic fever virus isolates reveals a high frequency of M segment reassortment. Epidemiol. Infect. 142, 1952–1962. Hewson, R., Chamberlain, J., Mioulet, V., Lloyd, G., Jamil, B., Hasan, R., Gmyl, A., Gmyl, L., Smirnova, S.E., Lukashev, A., Karganova, G., Clegg, C., 2004a. Crimean-Congo haemorrhagic fever virus: sequence analysis of the small RNA segments from a collection of viruses world wide. Virus Res. 102, 185–189. Hewson, R., Gmyl, A., Gmyl, L., Smirnova, S.E., Karganova, G., Jamil, B., Hasan, R., Chamberlain, J., Clegg, C., 2004b. Evidence of segment reassortment in CrimeanCongo haemorrhagic fever virus. J. Gen. Virol. 85, 3059–3070. Kulichenko, A.N., Volynkina, A.S., Kotenev, E.S., Pisarenko, S.V., Shaposhnikova, L.I., Lisitskaya, Y.V., Vasilenko, N.F., Tsygankova, O.I., Evchenko, Y.M., Tohov, Y.M., Savel'ev, V.N., Tihonov, S.N., Penkovskaya, N.A., 2016. A new genetic variant of the Crimean–Congo hemorrhagic fever virus isolated in Crimea. Mol. Genet. Microbiol. Virol. 31, 94–101. Kumar, S., Stecher, G., Tamura, K., 2016. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874. Kuraku, S., Zmasek, C.M., Nishimura, O., Katoh, K., 2013. aLeaves facilitates on-demand exploration of metazoan gene family trees on MAFFT sequence alignment server with enhanced interactivity. Nucleic Acids Res. 41, W22–28. Lukashev, A.N., 2005. Evidence for recombination in Crimean-Congo hemorrhagic fever virus. J. Gen. Virol. 86, 2333–2338. Lukashev, A.N., Klimentov, A.S., Smirnova, S.E., Dzagurova, T.K., Drexler, J.F., Gmyl, A.P., 2016. Phylogeography of Crimean Congo hemorrhagic fever virus. PLoS One 11, e0166744. Onishchenko, G.G., Markov, V.I., Merkulov, V.A., Vasil'ev, N.T., Berezhnoi, A.M., Androshchuk, I.A., Maksimov, V.A., 2001. [Isolation and Identification of CrimeanCongo Hemorrhagic Fever Virus in the Stavropol Territory]. Zhurnal mikrobiologii, epidemiologii, i immunobiologii. pp. 7–11. Rospotrebnadzor, 2016. Letter №01/500-16-32 “On Crimean Hemorrhagic Fever Epidemiological Situation in Southern, North Caucasian and Crimean Federal Districts of Russia Federation and Forecast for 2016”. Russian Federal Service for Surveillance on Consumer Rights Protection and Human Wellbeing (Rospotrebnadzor), Moscow. Sherifi, K., Cadar, D., Muji, S., Robaj, A., Ahmeti, S., Jakupi, X., Emmerich, P., Kruger, A., 2014. Crimean-Congo hemorrhagic fever virus clades V and VI (Europe 1 and 2) in ticks in Kosovo, 2012. PLoS Negl. Trop. Dis. 8, e3168. Zehender, G., Ebranati, E., Shkjezi, R., Papa, A., Luzzago, C., Gabanelli, E., Lo Presti, A., Lai, A., Rezza, G., Galli, M., Bino, S., Ciccozzi, M., 2013. Bayesian phylogeography of Crimean-Congo hemorrhagic fever virus in Europe. PLoS One 8, e79663. Zhou, Z., Deng, F., Han, N., Wang, H., Sun, S., Zhang, Y., Hu, Z., Rayner, S., 2013. Reassortment and migration analysis of Crimean-Congo haemorrhagic fever virus. J. Gen. Virol. 94, 2536–2548.
Analysis of recent GtVa CCHFV sequences from South Russia indicates that the virus population is highly dynamic in terms of longdistance virus transmission and reassortment. The European genotype V of CCHFV emerged just 100–500 years ago according to various estimates. Within the South Russia, genetic variants of GtVa from different areas were intermixed, and identical viruses could be isolated from distant locations. Therefore, we can conclude that numerous long-distance transfers occurred within South Russia over a few hundred years. On the other hand, there was no evidence of recent virus transfers between South Russia and Turkey, and only one transfer from Russia to Iran could be suggested. Apparently, ecological barriers (Caucasian mountains that split the habitat of steppe ticks) or socio-economic barriers (borders that limit livestock transport) are poorly penetrable to the virus. In the absence of such barriers, the high efficiency of virus spread observed in South Russia implies that a CCHFV focus, once established, can grow and occupy the whole suitable niche over several decades. This observation is especially important for predicting CCHFV spread in Southern Europe, where the virus is spreading West-ward from South Russia/Turkey (Zehender et al., 2013) and can be introduced by migratory birds (Estrada-Pena et al., 2012). Within a focus, CCHFV undergoes reassortment at least once every few decades. A detectable reassortment implies common co-infection by two phylogenetically distinct virus variants; thus, frequent reassortment further highlights a high frequency of virus intermixing within a region. Therefore, even on a timescale of decades, the virus population within the South Russia exists as a cloud of swiftly spreading and frequently reassorting genome segments. It is currently unknown if the co-infection required for reassortment occurs in a tick or in a mammal, and both hypotheses can be justified. High-throughput sequencing of virus populations can answer this question in the future. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.meegid.2018.01.016. Funding This study was supported by the Russian Science Foundation Grant 14-15-00619. Competing interests The authors declare that they have no competing interests. References Anagnostou, V., Papa, A., 2009. Evolution of Crimean-Congo hemorrhagic fever virus. Infect. Genet. Evol. 9, 948–954. Bente, D.A., Forrester, N.L., Watts, D.M., McAuley, A.J., Whitehouse, C.A., Bray, M.,
27