Phylogeography of the second plague pandemic revealed through analysis of historical Yersinia pestis genomes

The second plague pandemic, caused by Yersinia pestis, devastated Europe and the nearby regions between the 14th and 18th centuries AD. Here we analyse human remains from ten European archaeological sites spanning this period and reconstruct 34 ancient Y. pestis genomes. Our data support an initial entry of the bacterium through eastern Europe, the absence of genetic diversity during the Black Death, and low within-outbreak diversity thereafter. Analysis of post-Black Death genomes shows the diversification of a Y. pestis lineage into multiple genetically distinct clades that may have given rise to more than one disease reservoir in, or close to, Europe. In addition, we show the loss of a genomic region that includes virulence-related genes in strains associated with late stages of the pandemic. The deletion was also identified in genomes connected with the first plague pandemic (541–750 AD), suggesting a comparable evolutionary trajectory of Y. pestis during both events.

This work provided an opportunity to explore health and disease of the city's inhabitants during a period of considerable population growth fed by migration to the expanding metropolis.
The discovery of a mass pit in the central area of the southern half of the burial ground, containing at least 42 individuals, provided an opportunity to investigate the first archaeologically excavated 17 th -century plague burial in London. The pit contained stacks of coffined and un-coffined burials, up to eight individuals deep. It was filled in a single event. The head ends of the coffins were alternated to allow the maximum use of space within the pit. A single perpendicular line of burials filled a gap at one end.
The archaeological dating, although imprecise and complicated by some intrusive finds introduced by later grave cuts, is consistent with a late 16 th -or early 17 th -century date. The east-west alignment of the pit was typical of burials from the earlier period of the burial ground's use, 1569 -1670. The fill contained a small pottery assemblage dated between 1550 and 1610, and coffins within the pit were of a type that appeared in the last quarter of the 16 th century and was ubiquitous from 1650 onwards. While radiocarbon dating was not sufficiently precise to distinguish between specific plague events, the dates clearly indicated that the Great Plague of 1665 was too late to have been responsible for the mass burial and the outbreaks in 1603, 1625 and 1636 were the most likely.

Augustinian Friary, Cambridge, England
Documentary evidence indicates that the Augustinian Friary in Cambridge was founded between 1277 and 1289, as a mendicant institution with a strong focus on learning and continued until the Dissolution of the Monasteries in 1538. The friary probably acquired burial rights in 1290 if not earlier and in 1302 these were extended to include individuals who were not members of the Augustinian order. Excavations in 2016-17 by the Cambridge Archaeological Unit as part of the redevelopment of the New Museums site by the University of Cambridge revealed two phases of friary related activity, both with associated burials 32 .
The earlier phase mainly comprised a cemetery located south of the friary church, where 32 burials that appear to be a mixture of friars and lay individuals were excavated. General archaeological evidence, including belt buckles associated with some of the burials, suggests that the cemetery spans the late 13 th to mid-14 th century, with radiocarbon dating of three burials in a stratigraphic sequence indicating that burial continued until at least the 1330s (all radiocarbon dates quoted to 95% probability). Ancient DNA analysis of one of the skeletons from the cemetery F.355 (NMS003) yielded a whole Y. pestis genome that likely has an identical genotype to other European genomes attributed to the Black Death episode (see Supplementary data 3), indicating that this burial likely dates to 1348/49. Stratigraphically F.355 (NMS003) predates two radiocarbon-dated burials F.332 and F.328. F.332 dates to 1286-1397 cal AD, but when factors including the marine reservoir effect, bone turnover/remodeling and stratigraphic sequence are considered it is dated to 1323-1426.
The second phase related to the friary is marked by the construction of the claustral ranges.
Typologically the cloister arcade is dated to between c. 1330 and c. 1390. This construction is unlikely to have predated the late 1330s, as it was not until then that the friary managed to acquire enough properties to embark on this major project. Additionally, the legs of F.355 (NMS003), which was likely buried in 1348/49, were removed when the foundation trench was dug for the wall of the cloisters. As this process did not disturb the rest of the bones it is probable that some time had passed since the interment, conservatively suggesting a date of no earlier than c. 1360 for the wall footing.
Six individuals were buried in the Chapter House in the eastern claustral range of the friary c.
1360/90-1538. As with the earlier cemetery the burials appear to be a mixture of friars and lay individuals. Three of the burials from the Chapter House tested positive for Yersinia pestis; F.230 (NMS002) yielded a whole-genome, whereas F.190 (NMS001) and F.310 (NMS005) remain to be confirmed. There is no stratigraphic evidence to indicate the relative sequence of these three burials and the Y. pestis DNA could relate to either a single plague outbreak or multiple outbreaks. F.230 (NMS002) has a copper alloy symmetrical double oval frame belt buckle, dating to between c. 1350 and the 16th century and the grave fill contained some 15 th -to 16 th -century lead window came.
F.230 has been radiocarbon dated to 1437-1619 cal AD, but when various factors are allowed−including the marine reservoir effect, the bone turnover/remodeling, the stratigraphic sequence and the fact that the burials date to 1538 at the latest−the probable date range is 1475-1536. There are several documented plague outbreaks in England during the late 15 th and early 16 th centuries, with major outbreaks in 1471 and 1479-80 and smaller scale recurrences in 1499-1500, 1509-21, 1526-32 and 1535-36, though it is uncertain which of these outbreaks affected Cambridge. The only specific dating evidence associated with F.190 (NMS001) is a small amount of pottery that indicates a 16 th century date, and the location of the grave suggests that it may be contemporary with F.230 (NMS002). F.310 is also accompanied by a copper alloy symmetrical double oval frame belt buckle, dating to between 1350 and the 16 th century.

"Sankt Johans Freidhof", Nabburg, Germany,
The "Sankt Johans Freidhof" churchyard (sic, see Hensch 2014) 33 in the city of Nabburg in Southern Germany was excavated between 2008 and 2012 next to the hospital church "St. Maria", consecrated in 1420. However, the earliest graves are attributed to the older parish church "St.
Johannes der Täufer" from the end of the 13 th century. The ~200 excavated graves were found in a maximum of eight layers and could be classified in two groups: The older graves from the late 13 th to late 14 th centuries are dug in a strict and regular layout, the younger graves from the 15 th century to the closure in 1597 are more irregular, presumably due to the lack of space. Only in four cases, the use of a coffin could be attested by the finds of iron nails and wooden remains.
Towards the western border of the cemetery, in total nine multiple burials were found with between two and four individuals. The four individuals that tested positive for Y. pestis originate from three multiple burials: NAB002 (arch. ID 451, early juvenile male) and NAB003 (arch. ID 452, early adult female) were found in a triple burial with an additional young woman, all piled up in a narrow grave pit in supine position. NAB004 (arch. ID 457, presumably female of 6-12 years, supine position) was buried in a simultaneous double burial together with an adult female on top in the opposite orientation. NAB005 (arch. ID 471, early mature female) was also buried in a simultaneous double burial, here with an adult male individual beneath in prone position.
A connection to the second plague pandemic was suggested by the archaeologists given the grave characteristics, despite the fact that there are no historical records for the Black Death in Nabburg.
However, epidemic records exist for the nearby towns of Amberg, Sulzbach, Burglengenfeld and Regenstauf from 1349 onwards.

St. Leonhardi, Manching-Pichl, Germany
During the renovation works of 1984/85 at St. Leonhardi, a mass grave was found under the sacristy. The mass grave revealed a minimum number of 75 individuals in four layers. The construction of the mass grave suggests that it was not dug into the ground but the individuals were placed on ground level without coffins and afterwards covered with earth 28 . However, the removal of the ground floor within the nave also revealed the remains of six additional individuals with some of them considered to be the church donors. The high amount on disarticulated skeletal elements within the mass grave hints towards a more intensive use of the site as a burial ground.
The construction of the sacristy can be dated to the second half of the 15 th century, which would give a terminus ante quem, assuming that the sacristy was built after the deposition. Radiocarbon dates of earlier studies gave two contradicting intervals, one spanning roughly the 12 th century and another spanning the 14 th century [25][26][27] , which could be explained by erroneously assigned scattered remains of earlier burials in this area.
The mass grave was repeatedly subject to aDNA studies that investigated the presence of Y. pestis DNA with PCR and qPCR assays [25][26][27][28]34 . In this study, individuals MAN008 (arch.  35 . The excavation revealed 365 burials, estimated to represent only one third of the cemetery and dating mostly between the 8 th /9 th and the 18 th centuries. However, five stone plate graves date between the middle of the 7 th and the early 9 th centuries. This indicates, that the cemetery was originally founded as a private burial ground by an aristocratic family, probably living nearby in a manor, in the mid-7 th century.
The first church can be dated to the middle of the 7 th century, too, and is actually one of the oldest known churches in Bavaria.
After the church was donated to Benediktbeuren Abbey in the 8 th or 9 th century, it was turned into a parish church for the surrounding village of Achheim. The Merovingian church was replaced by a larger one in the late 10 th or early 11 th century. This medieval church was once again completely rebuilt in the first half of the 15 th century.
Achheim stayed a small village at the shore of Lake Starnberg until the mid-19 th century, when it became part of the growing modern town of Starnberg. Starnberg Castle (first mentioned in 1245), which surmounts the former village of Achheim, was a temporary summer residence of the dukes of The medieval graves are characterized by the complete absence of grave goods and can only be dated by stratigraphy, radiocarbon analysis and their arm posture. In the years around 1300 burial rites changed and few grave goods can be found again, especially rosaries as well as occasional coins, buckles or finger rings.
One triple burial -situated just a few meters southeast of the sacristy and belonging to the late medieval period -was the only multiple burial found at this site. Due to this, it was examined here for the presence of Y. pestis. Individual STA001 (arch. ID 207, infans II to juvenile, aged 12-14, presumably male) tested positive for the bacterium through qPCR and high throughput sequencing (Supplementary table 2). In its hands the individual held a rosary, as shown by few remains of blue glass beads, which cannot be dated precisely. The juvenile was simultaneously buried together with two other children aged 3-5 (arch. ID 208, infans I) and 12 (arch. ID 214, infans II), in regular supine position. The rosary found with STA003 (arch. ID 214) can be dated between the late 14 th and the early 16 th centuries. This special long form of a rosary was composed of approximately 150 ring-shaped bone beads (15 mm in diameter), represented by 122 intact beads and 38 fragments.
Taking into account the 1-sigma interval of the radiocarbon dating, the date of the triple burial can be narrowed down to 1433-1494 (59,6 %), respectively to 1420-1523 (74,5 %) when considering the 2-sigma interval. Thus, these three subadults died during a historically undocumented plague outbreak in the 15 th or early 16 th century.

Kirchhof St. Johannis, Landsberg am Lech, Germany
After Landsberg was granted the town privilege in 1315 the small trading post on the crossroads of the salt route soon started to prosper 36  Imprints in the quicklime show that the dead were wrapped in shrouds and were then piled up in the burial pits. In contrast to the regular burials, which contained the typical furnishing (rosaries and multiple types of devotional objects) of catholic funerals of that period, just a few non-datable finds came from these epidemic burials.
Three individuals tested qPCR positive for Y. pestis. These were LBG002 (arch. ID 460, presumably male juvenile) was buried in a 11-fold burial, LBG005 (arch. ID 572, presumably female 6-12 years) from a single burial with quicklime, and LBG007 (arch. ID 598, presumably a female juvenile) buried in a quintuple burial with quicklime.

Nägeligasse, Stans, Switzerland
The partial excavation of the churchyard of "St. Peter und Paul" took place between 2015 and 2016.

Domlinden 12, Brandenburg an der Havel, Germany
During a survey in 2011, the burial of three individuals was found on the "Dominsel", the centre of the old town of Brandenburg, situated between two streams of the Havel. It is noteworthy that no context of a contemporary cemetery was found 41 . Instead, it appears that the individuals were buried in the backyard of a bourgeois house. One of the individuals was buried with a clay pipe bowl with the initials of the Dutch manufacturer Samuel Collier, setting the terminus post quem between 1630 and 1640. All three individuals were morphologically identified as adult males.
Moreover, one of them showed a healed impression fracture of the skull, indicative of interpersonal violence. Due to the poor health status and the hasty burial that did not follow the Christian burial practices, the individuals have been characterized as foreigners with low social status. Indeed, the oxygen and strontium isotope data show a non-local signature consistent with Scandinavia or the Baltic region. Therefore, it was hypothesized that the individuals were foreign soldiers housed with civilians during the occupation of the city by Swedish troops between 1626 and 1631 as part of the Supplementary figure 12 -Maximum likelihood phylogeny 2 (97% partial deletion) of all Y. pestis genomes used in this study. A total of 6,058 SNP positions were considered for the phylogeny. The tree is comprised of 233 modern isolates 3-9 , 35 second pandemic isolates, one first pandemic isolate 10 , one 2 nd -to 3 rd -century isolate 11 , three Bronze Age isolates 12,13 , and a Y. pseudotuberculosis isolate (IP32953) 14 that was used as outgroup for rooting the tree. Bootstrap values of 95 or higher are shown. All early 14 th -century genomes are shown in red, late 14 th -century genomes are shown in orange, 15 th -17 th century genomes are shown in green and 17 th -to18 th -century genomes are shown in blue.    Genomes are arranged from top to bottom according to their inferred divergence ages (from youngest to oldest), and all second pandemic strains are shown in red, orange, green or blue based on their age (see Figure 1). A lack of coverage across the plasmids in the Marseille OBS isolates is apparent since those regions were not used as part of the probe set for enrichment of these loci and are marked here with "X" 23 . Instead, missing regions in all other isolates represent real gene absences.
Supplementary figure 21 -Analysis of coverage across genes identified as absent in 17 th -18 th century second pandemic genomes BED and OBS. The presence or absence of genes located on the Y. pestis chromosome in representative ancient and modern genomes is shown. Specifically, second pandemic isolates from all studied archaeological sites (one representative chosen when more than one identical genome was recovered from a site), one published first pandemic genome from Germany (Altenerding 2148), one 2 nd -to 3 rd -century genome from the Tian Shan region (DA101) and three Bronze Age genomes from Russia (RT5, RISE505, RISE509) are shown for comparison. In addition, modern representatives from all phylogenetic clades are shown. The heatmap was constructed using the ggplot2 22 package in R version 3.4.1 15 . Genomes are arranged from top to bottom according to their inferred divergence ages (from youngest to oldest), and all second pandemic strains are shown in red, orange, green or blue based on their age (see Figure 1). A lack of coverage across certain Y. pestis-specific regions in Altenerding 2148, Barcelona 3031, and Bolgar 2370 are marked here with "X" since they were not used as part of the probe set for their enrichment 10,24 . Instead, missing regions in all other isolates represent real gene absences.