Retrospective analysis of dissemination of the 2.MED1 phylogenetic branch of Yersinia pestis in the Caucasus

The 2.MED1 phylogenetic branch of Yersinia pestis of the medieval biovar became widespread in the Caspian Sea region, the Caucasus, and the Northern Aral Sea region in the 20th century, causing outbreaks and epizootics of plague there. Some of the formed natural foci of 2.MED1 still show epizootic activity and retain their epidemic potential. In this work, we carried out a phylogenetic analysis of 46 Y. pestis strains of the medieval biovar isolated in the Caucasus, the Caspian Sea, and the Northern Aral Sea regions during epidemic outbreaks and epizootics from 1922–2014. The obtained phylogenetic data, together with epidemiological and epizootological data accumulated over a period of about a hundred years, indicate the presence of two waves of penetration of the 2.MED1 branch into the Caucasus. The first occurred, apparently, in the first half of the 20th century as a result of the penetration of 2.MED1 from the foci of the Northern and North-Western Caspian Sea. The second wave was caused by the spread of 2.MED1 from the Northern Aral to the foci of the North-Western, Northern and Eastern Caspian Sea regions at the beginning of the second half of the 20th century, followed by introduction into the Pre-Caucasus and Transcaucasia. The rapid spread of 2.MED1 could be associated with the transfer of the pathogen by land and sea transport in the process of economic activity of the population.


Introduction
The Caucasus is located between the Black and Caspian Seas and is divided into the Pre-Caucasus and Transcaucasia, separated by the Main Caucasian mountain range. There are seven natural foci of plague, geographically located in the Russian Federation, Azerbaijan, Armenia and Georgia. In five of these foci, a highly virulent medieval biovar of the main subspecies of Yersinia pestis is common [1]. Three of the five foci-Dagestan plain-piedmont, Terek-Sunzha low-mountain, Central Caucasian high-mountain-belong to the foci of the souslik type and are located in the Pre-Caucasus and the Central Caucasus (S1 the gerbil type foci and are located in Transcaucasia-Transcaucasian plain-piedmont (a group of autonomous plain-piedmont foci: Bozchel, Kobystan, Mil-Karabakh, Jeyranchel and others) and Araks low-mountain foci [2]. In two other foci of the Caucasus, the East Caucasian highmountain and the Transcaucasian high-mountain (a group of autonomous mountain foci: Gyumri, Sevan and Zangezur-Karabakh), only the aucasian subspecies-Y. pestis subspecies caucasica (phylogenetic lineage 0.PE2) is currently found, which is virulent mainly for small mouse-like rodents, but can cause isolated cases of plague in humans.
In accordance with genetic nomenclature of branches, the medieval biovar corresponds to the phylogenetic lineage 2.MED [3][4][5][6]. The 2.MED lineage is divided into several phylogenetic branches: 2.MED0 (Russia, Central Caucasian high-mountain focus), 2.MED2 and 2.MED3 (China), and 2.MED1 (Eurasian foci). The 2.MED0 strains are endemic to the territory of the Central Caucasian high-mountain focus and are not found in other plague-enzootic regions of the world. The most widespread is the most recent branch 2.MED1, which occupies vast areas of natural plague foci of Eastern Europe and Central Asia. The place of origin of the medieval biovar (lineage 2.MED) is unknown. It is believed that it could have been near the Caspian Sea [7]. In the first half of the 20th century, multiple outbreaks of plague with a high mortality rate occurred in the North-Western and Northern Caspian Sea region with a peak incidence in the 1920s-1930s. In the same period, epizootics among rodents also occurred in this territory. All of these events were preceded by an outbreak of bubonic plague (444 cases, 363 deaths) in the village of Vetlyanka in the Astrakhan province in 1876-1879 [2]. Many dangerous diseases from other countries of the Caspian region were brought into the Russian Empire through the Astrakhan province, located at the confluence of the Volga River into the Caspian Sea. The phylogenetic affiliation of the Y. pestis strains that caused the outbreak of plague in Vetlyanka is unknown.
A study of Y. pestis strains obtained in the North-Western and Northern Caspian Sea regions in 1912-1945 showed that the etiological agent of those outbreaks were 2.MED1 strains, as well as strains of another newly identified 2.MED4 branch of the medieval biovar [8]. The 2.MED4 branch existed in the North-Western (Caspian North-Western steppe focus) and Northern (Volga-Ural steppe, Volga-Ural sandy foci) Caspian Sea regions in the first half of the 20th century. In the second half of the 20th century, 2.MED4 strains were no longer isolated. Only one strain of the 2.MED4 branch isolated in the Caucasus is known. It was obtained in 1931 in the Transcaucasian high-mountain focus (Zangezur-Karabakh autonomous focus) from a man who died of pneumonic plague.
In contrast to the North-Western and Northern Caspian Sea regions, there were no large plague outbreaks in the natural foci of the Caucasus during the 20th century, although some local outbreaks or isolated cases of plague did occur. The last major plague outbreaks in the Caucasian natural foci occurred in 1838-1843, and by the end of the 19th century, plague cases in the Caucasian foci were a rare phenomenon. In the 20th century, outbreaks of the plague occurred in the Transcaucasian plain-piedmont focus: in 1914 in the Baku district (30 cases. 30 deaths) and Baku region (50 cases, 50 deaths), in the Nagorno-Karabakh region in 1929-1931 (8 cases, 8 deaths), and in the Araks low-mountain focus in 1948 (15 cases, 12 deaths).
Based on phylogenetic analysis of Y. pestis strains in combination with epidemiological and epizootiological data, we studied the spatiotemporal patterns of 2.MED1 circulation in the foci of the North-Western and Northern Caspian regions in the 20th and early 21st centuries [8]. We assumed that changes in the level of the Caspian Sea affected the activity of these plague foci. Sharp fluctuations in the level of the Caspian Sea, which had a significant impact on the ecosystem of the Caspian and Turan lowlands, repeatedly occurred in the 20th century under the influence of cyclical climatic changes [9][10][11]. In our current work, based on a complex of phylogenetic, epidemiological and epizootiological data, we identified the most probable directions of dissemination of the 2.MED1 branch of the medieval biovar in the Caucasus in the 20th-early 21st centuries. This clarification is important for understanding the functioning of natural plague foci and predicting focus activation in the future.

Materials and methods
Yersinia pestis strains Y. pestis strains isolated from natural plague foci of the Caucasus, Northern Caspian region and other foci of Eastern Europe and Central Asia from carriers and vectors of plague, as well as from humans were studied (S1 Table) Y. pestis was grown in LB broth and LB agar for 24-48 hours. Traditional methods of laboratory diagnostics were used for the assessment of biochemical properties of the strains [12]. Fermentation of glycerol, rhamnose and arabinose was determined on Hiss media (1% peptone water, 0.5% sodium chloride, 1% Andrade indicator, pH 7.2) with the addition of 1% of the corresponding substrates, which were inoculated with 10 8 CFU of Y. pestis, and cultivated for 48 hours at 28˚C. The reaction was considered positive if the medium became crimson in color. To determine the denitrifying activity, 5 ml of LB broth with 0.1% potassium nitrate were inoculated with 10 8 CFU of Y. pestis, cultivated for 72 hours, and then the Griss reagent was added. Crimson staining of the medium indicated the presence of denitrifying activity.
(https://github.com/rrwick/Unicycler). Alignment of the obtained reads to the reference genome of Y. pestis CO92 (GenBank accession number NC_003143.1) and identification of plasmid sequences was performed using the DNASTAR Lasergene v15.3 software package [14]. Default parameters were used for all software. The sequence reads were assembled into contigs with average coverage per genome being 98.47% (45,9-fold depth) and an average genome assembly size of 4,55 Mb, The average GC% content ranged from 47.39% to 48.06%. Finally, we obtained from 102 to 449 contigs > 1 kb for each genome (S2 Table). The final assemblies were annotated using the NCBI Prokaryotic Genome Annotation Pipeline (PGAP) v6.1 [15]. Each genome contained from 2,445 to 3,978 coding sequences. Core SNPs were identified by aligning contigs of Y. pestis strains to CO92 genome through Snippy 4.6. software program, then, 28 homoplastic SNPs were excluded. Core SNP calls for all Y. pestis strains used are shown in S3 Table. The resulting set of SNPs included only the core region of the genome. Using the Snippy software package (https://github.com/tseemann/snippy) provides a set of core SNPs that can be used for constructing high-resolution phylogenetic trees, since it excludes sites of possible recombination. The jModelTest2 program [16] was used to select a model of nucleotide substitutions. Based on the adjusted Akaike criterion (AICc), the GTR model was selected. The Maximum Likelihood tree was constructed using software: SeaView 5.0.4.
According to the whole genome SNP analysis, based on 1324 core SNPs, a dendrogram of phylogenetic relations of Y. pestis strains from the plague foci of the Caucasus was constructed (Fig 1).
In total, whole genome sequences of 25 Y. pestis strains of the medieval biovar from the foci of the Caucasus and Transcaucasia, 24 strains of the medieval biovar from the Northern Caspian and Northern Aral Sea regions, and 3 strains of this biovar from Iran and China were used to construct the dendrogram. On the dendrogram, Y. pestis strains of the phylogenetic lineage 2.MED are divided into sequentially diverging branches: 2.MED0, 2. MED3, 2.MED2, 2.MED4, and 2.MED1. The 2.MED1 branch is preceded by the 2.MED4 branch with two strains from the Volga-Ural sandy (1924) and Volga-Ural steppe (1950) foci, and one strain from the Zangezur-Karabakh mountain autonomous focus (1931) included in it. 2.MED4 is followed by 2.MED1, the base trunk of which is represented by a polytomy (MN1 -Medieval Node 1, 12 SNPs in S4 Table), from which sub-branches and clusters outstretch, including Y. pestis strains from the North-Western, Northern, and Eastern Caspian Sea regions, as well as from the Caucasus and Transcaucasia. This large polytomy indicates a rapid expansion of Y. pestis across the Caspian Sea region, Caucasus, and Transcaucasia.
The earliest strains on the polytomy are represented by two sub-branches No. 2 (MN2, 1 SNP) from the Northern, North-Western Caspian and No. 3 (MN3, 1 SNP) from the Northern Caspian Sea regions of the early 20th century (1922)(1923)(1924). All of these strains were isolated from humans and were the etiological agents of plague outbreaks with high mortality rates that occurred in the Caspian North-Western steppe, Volga-Ural steppe and Volga-Ural sandy foci at the beginning of the past century. These early strains departing practically from the trunk of the polytomy (1 SNP), together with a large number of plague cases in this territory, suggest that the first manifestations of the spread of 2.MED1 in the Caspian Sea region and the Caucasus in the early 20th century began in the North-Western and Northern Caspian regions. This assumption is also supported by the data of the Maximum Parsimony dendrogram of phylogenetic relations of Y. pestis strains of the 2.   Table. https://doi.org/10.1371/journal.pone.0283670.g001 of strains isolated in the Caucasus and Transcaucasia from 1950-1986. Earlier strains of this sub-branch are absent, and therefore the exact place of its origin is unclear. The strains of subbranch No. 4 were obtained from carriers and vectors of plague, which indicates the persistence of this sub-branch in the ecosystem of Caucasian foci in the second half of the 20th century. Sub-branch No. 4 is divided on the dendrogram into two clusters, the first of which includes strains from the Mil-Karabakh, Bozchel, and Kobystan autonomous foci of the Transcaucasian plain-piedmont focus and the KIM10 strain from Kurdistan, Iran. The second cluster consists mainly of strains of the Central Caucasus, the Pre-Caucasus and the North-Western Caspian region.
The remaining strains, isolated from the Caucasus in 1953-2003, did not belong to these Caucasian sub-branches (extending directly from the stem of the 2.MED1 polytomy), but were introduced into the Caucasus from the foci of the Caspian and Northern Aral Sea regions. For example, sub-branch No.5 (MN5, 4 SNP) is composed of three strains, the first of which was isolated in the Volga-Ural sandy focus in 1945 from a human, and two strains of the cluster were obtained in 1953 from fleas in the Kobystan plain-piedmont autonomous focus. Apparently, those strains were introduced into the Kobystan focus from the Northern Caspian region (Fig 2).
Sub-branch No. 6 (MN6, 1 SNP) offshoots from the trunk of the polytomy. It is divided into two sub-branches-the Eastern Caspian (No. 7, MN7, 3 SNPs) and the Northern Aral (No. 8, MN8, 1 SNPs), which demonstrates the further expansion of the 2.MED1 branch to the east. Sub-branch No. 7 of the Eastern Caspian Sea region includes strains that date back to 1962- 1978 from the Ustyurt desert, Mangyshlak desert and Kobystan plain-piedmont autonomous foci. This population is the likely source for the introduction of 2.MED1 strains through the Caspian Sea into the Kobystan natural plague focus in 1965-1968. Sub-branch No. 7 includes two separate strains from the Ustyurt desert focus and two additional clusters, one of which consists of strains from the Kobystan focus (1965,1968), and the second of which consists of strains from the Mangyshlak (1978) and Kobystan (1976) foci. This suggests a path of spread for sub-branch No. 7 from the Ustyurt focus to the Mangyshlak focus and further through the Aktau port of the Mangyshlak Peninsula across the Caspian Sea to the Kobystan autonomous focus (Port Baku, Apsheron Peninsula) in Transcaucasia (Fig 2, S5 Fig).  (Fig 2).
In general, the phylogenetic analysis of Y. pestis strains obtained over a period of about a hundred years (1922-2014), together with epidemiological and epizootic data and climatic observations, testifies to intensive processes of dissemination of the 2.MED1 phylogenetic branch in the foci of the North-Western, Northern and Eastern Caspian Sea regions, as well as the Pre-Caucasus and Transcaucasia in the 20th century.

Discussion
The medieval biovar of the main subspecies of Y. pestis, phylogenetic lineage 2.MED is widely distributed in Eurasia in steppe, semi-desert, desert, plain-piedmont, low-mountain and highmountain landscapes with different composition of plague carriers and vectors. The exact place of origin of the 2.MED lineage is not known. The earliest of the branches of the medieval biovar 2.MED0 has been preserved in the Central Caucasian high-mountain focus in Russia.
Another, later branch, 2.MED4, existed in the Northern and North-Western Caspian regions at the beginning of the 20th century and possibly even earlier. From the same region, the last branch, 2.MED1, of the medieval biovar began to spread around the Caspian Sea southward to the Caucasus and eastward to the Eastern Caspian and Northern Aral Sea regions. The climate in the 19th century and the beginning of the 20th century in the Caspian lowland was cool with a large amount of precipitation, and the level of the Caspian Sea during this period was consistently high [9][10][11], http://www.geogr.msu.ru/casp/. This had a positive effect on the ecosystem of natural plague foci of the Caspian region. Against this background, strains of the medieval biovar caused multiple outbreaks of plague in the North-Western and Northern Caspian Sea regions, as well as outbreaks and isolated cases of plague in Central and East Asia in the 20th and 21st centuries [18,19].
In the Caucasus, infection with plague has been known for a long time. According to historical chronicles, outbreaks and epidemics of plague occurred there throughout the new era and in prehistoric times [20]. The phylogenetic affiliation of Y. pestis which caused outbreaks of plague in the Caucasus in the 19th century is unknown. In the 20th century, plague infections in the Caucasus occurred in the Transcaucasian plain-piedmont (1910,1914,1929), and then in the Araks low-mountain (1948,1967) and Dagestan plain-piedmont (1951) foci. As a rule, those were small outbreaks or isolated cases of plague. Two outbreaks happened in the Kobystan plain-piedmont autonomous focus in 1914. During one of them, 30 people got sick and died, and during the other 50 people got sick and died of pneumonic plague. Separate cases of plague also took place in the autonomous foci of the Caucasian high-mountain focus: in Zangezur-Karabakh in 1931 and 1975, and in Gyumri in 1958 and 1969 [18]. Two of those cases in 1958 and 1975 were caused by the Caucasian subspecies-Y. pestis ssp. caucasica, but such cases are rare [21].
This study reveals that two cases of plague in the Zangezur-Karabakh mountain autonomous focus in 1931 were caused by 2.MED4 and 2.MED1 strains. It can be assumed that the incidence of plague in the Caucasus at the very beginning of the 20th century (Transcaucasian plain-piedmont focus-1910, 1914, 1929) could also have been caused by 2.MED1 (possibly together with 2.MED4), as we showed for the foci of the North-Western and Northern Caspian regions. Directly from the stem of the 2.MED1 polytomy of the first half of the 20th century, sub-branch No. 4 evolved, which was preserved in the Caucasus through the second half of the century (Fig 1). At the same time, during the middle and in the second half of the 20th century, intensive processes of introduction of 2.MED1 into Transcaucasia and the Pre-Caucasus from the Northern and North-Western Caspian, Eastern Caspian and Northern Aral Sea regions took place. The rate of spread of 2.MED1 may have exceeded even the rate of spread of the phylogenetic lineage 1.ORI, which caused the third plague pandemic [22]. We have indicated that the second wave of 2.MED1 spread likely began in the late 1940s-early 1950s, when the 2.MED1 population preserved in the Northern Aral Sea during the inter-epizootic period gave rise to a new polytomy that swept the North-Western, Northern and Eastern Caspian regions and the Caucasus (Figs 1, 2; S3 and S4 Figs).
Interestingly, strains that came from the Northern Caspian Sea region (sub-branch No. 5), and descendants of strains from the Northern Aral Sea region (sub-branch No. 9) simultaneously circulated in the Kobystan autonomous focus in Transcaucasia in 1953 (Fig 1). In 1955-1966, strains of the local Caucasian sub-branch No. 4 were isolated here, and in 1965-1968 strains that came from the Eastern Caspian Sea region (sub-branch No. 7) were isolated. In a previous publication, we showed that in the period of 1962-1978 in the Eastern Caspian (Ustyurt Plateau) there was a population of the first wave of distribution of 2.MED1 from the Northern Caspian region of the early 20th century [23]. The Kobystan autonomous focus is located on the Apsheron Peninsula alongside the capital of Azerbaijan-Baku. Frequent introductions of plague to this region could have been due to this region's active economic ties to other regions of the Caspian Sea. Expansion of sub-branch No. 7 in the 1960s and subbranch No. 9 in the 1950s into the Kobystan focus from the Eastern Caspian regions most likely occurred through the Mangyshlak (port Aktau) and Karakum (port Turkmenbashi) desert foci across the Caspian Sea with commercial and economic cargo (Fig 2, S5 Fig). Previous studies of 580 strains isolated on the Apsheron Peninsula in the middle of the last century showed that they also belong to the medieval biovar of the main subspecies of the plague microbe [24].
The role of the South Caspian Sea region in the spread of 2.MED1 remains unclear. We assume that 2.MED1 was brought to the North-Western and Northern Caspian in the late 19th-early 20th century from Iran, since there was no plague in other nearby regions. Outbreaks of plague in Iran occurred in 1871-1877, 1899, then in 1906-1924 [25]. They may have been caused by 2.MED1. Their introduction across the Caspian Sea into the North-Western Caspian region in the Astrakhan province in 1876 (Vetlyanka station) or later served as the beginning of the spread of 2.MED1 in the foci of the Caspian, Caucasus and Northern Aral in the 20th century. During the period of 1899-1916, 1,758 cases of plague were registered among the populations on the territory of the Right-bank villages of the Volga in the Astrakhan region, with 1,582 of them ending in death [18,20]. It also cannot be excluded, that outbreaks in Baku and in the Baku region in 1914 were caused by the importation of the plague directly from Iran. Azerbaijan and Iran are connected by close cultural and economic ties, which could be the reason for historical plague penetrations into the Transcaucasia from the south.
The data presented here indicate that in the plain, plain-piedmont and low-mountain natural foci located in the arid landscapes of semi-deserts and deserts of the Caspian and Turan lowlands, under the influence of unfavorable climatic changes, the disappearance of local populations of Y. pestis may occur. A new activation of the epizootic process may happen due to the introduction of Y. pestis from other epizootically active areas and its next rooting in historical focal areas. In this work, we confirmed the introduction of Y. pestis strains of the medieval biovar of the 2.MED1 phylogenetic branch from the Caspian Sea and the Northern Aral Sea region to the Eastern Transcaucasia and Pre-Caucasus in the second half of the last century. The intensity and high rate of the spread of 2.MED1 were apparently caused by active economic and trade activities in the regions of the Caspian Sea and the Caucasus. The import of Y. pestis as a result of human economic activity have often led to its introduction into the ecosystems of natural focal areas in the presence of favorable climatic conditions. Another possible mechanism of plague dissemination, such as the natural migratory activity of carriersrodents (small souslik; midday gerbil, crested gerbil, great gerbil and fleas parasitizing on them) could hardly provide such a high rate of spread of the 2.MED1 population of medieval biovar over the vast expanses of Eastern Europe and Central Asia in the 20th and early 21st centuries.
Supporting information S1