Assessing the origins of the European Plagues following the Black Death: A synthesis of genomic, historical, and ecological information – pnas.org

Assessment of the Two Hypotheses.

For the purpose of understanding the evolution of the plague bacteria, more than 100 ancient Y. pestis genomes have been published to date. The last 17 were recently reported during a short period by four distinct research groups (79, 12). Using most of the ancient genomes (criteria for exclusion are described in Methods), along with 499 modern ones, we present here the most updated phylogeny (Fig. 1).

Phylogeny and archaeological site locations of ancient genomes. (A) A maximum likelihood phylogeny was obtained with 574 genomes of Y. pestis (including 75 ancient genomes) involved based on 12,608 SNPs. The numbers at each node indicate the bootstrap values of 1,000 replicates. Branches highlighted in blue correspond to the second pandemic, which is subdivided in three groups: the 14th to 15th century group, which also includes the Black Death and the Pestis secunda (1,357 to 1,366) strains; the 15th to 17th century group; and the 18th century group (which also includes the BED genomes for homogeneity). Branches in purple correspond to the first Pandemic, and branches in green correspond to the prehistoric plague. The ratio between the depth of pla and that of the entire pPCP1 plasmid for all ancient genomes is shown in the rightmost heatmap, with a color scale ranging from 0 (dark blue) to 130+ (dark red). (B) Geographic distribution of the three waves during the second pandemic.

The updated phylogeny confirms the almost clonal nature of the Black Death strains in comparison to all other lineages of the second plague pandemic, including the strains from the Pestis secunda [Ber37 and Ber 45, The Netherlands (6), and BolgarCity2370, Russia (3)], which are placed on Branch 1 [see also London-Ind6330, United Kingdom (3)], as well as to all other strains, which are placed on the postBlack Death branch. There is general agreement that the postBlack Death branch was hosted in a novel wild rodent reservoireither in Europe or outside Europe (38, 12, 14). The original hypothesis (Hypothesis 1) claims that such a plague reservoir existed in Western Europe (15), perhaps in the Alps (16). However, a newer hypothesis (Hypothesis 2) claims that the plague reservoir was in Asia, possibly close to Eastern Europe (6, 7, 9, 11, 13).

In order to more easily view the phylogeny from the second plague pandemic and to better contrast the evidence for the two hypotheses, we generated two schematic figures (Fig. 2) and a table (Table 1).

Schematic comparison between the two main hypotheses for the interpretation of the Y. pestis phylogeny of the second plague pandemic. Historic and evolutionary information is included in the schematic figures. In addition to the symbols explained in the figure, we outlined in red the strains showing the 49-kb deletion. Pla decay (meaning both, full, or partial absence of the pla gene) is indicated by the names in bold.

Main differences between the two competing hypotheses proposed to explain the phylogeny of Y. pestis of the second plague pandemic; genomic and evolutionary, historical and archaeological, and ecological arguments are considered

Hypothesis 1 is supported by a phylogenetic analysis based on the currently available ancient genomes, which infers high posterior probability for a Western European source of the transmissions on the postBlack Death branch (SI Appendix, Fig. S1). However, as the dataset includes 41 ancient genomes from Western Europe against only 8 strains from Eastern Europe (including Gdansk and Riga), the proposed origins from Western Europe are likely to be biased toward a European reservoir due to a size-effect bias. Notably, the most basal genome LAI009 (4) (the Black Deaths lineage), Bolgar (at the root of Branch 1), and the most recent genome [CHE1 (7)] all originated from Western Russia, implying that they might have been closer to a putative Asian or Eastern European reservoir. This continuity does represent strong evidence in support of Hypothesis 2.

Using only genomic data, Hypothesis 1 might be seen as the most parsimonious hypothesis since it proposes an internal source for all western Eurasian outbreaks. However, for two locations (Pestbacken, Sweden 1710 [PEB10] and Marseille, France 1722 [OBS]), an origin from the Ottoman Empire is historically and archaeologically well supported (7). Thus, Hypothesis 1 needs to account for a back and forth spread, which reintroduced plague on two occasions to the Ottoman Empire and back again to Western Europe. Notably, none of the strains from the 18th century appear to have originated in Western Europe according to historical sources (7, 9).

Hypothesis 1 assumes the existence of a wild rodent plague reservoir in the Alps, which is not supported by ecological evidence (13). Instead, a study of more than 7,000 historical plague outbreaks and 15 tree-ring datasets (four of which from the Alps) found climatic signals in support of frequent reimportations of plague from Asia into Eastern and Western European harbors (13).

Intriguingly, only a few genotyped strains are nodes on the backbone of the postBlack Death branch: the strains of the Black Death itself, the strain from Gdansk 1425 to 1469, and the strains from London (BED, 16th to 17th century). While the strains of the Black Death were notoriously imported into Western Europe from the Mongol Empire via Caffa in Crimea (10), both Gdansk and London were very active harbors also in historical times and were very often hit by plague. Interestingly, Y. pestis was also recovered from a rat found in Gdansk. Although the genome is partial due to the different single-nucleotide polymorphism (SNP) profile, it is clear that the strain from the rat could not have infected the victim (Gdansk8) (9). Being a port, Gdansk may indeed have hosted diverse importations of infected rats in the period from 1425 to 1469, as it happened in European harbors during the third pandemic (17).

Hypothesis 2 is consistent with the ecological as well as with the historical evidence (Fig. 2 and Table 1). The only Western European subcluster, the Alpine cluster formed by LBG (Landsberg, Germany), STN (Stans, Switzerland), BRA (Brandenburg, Germany), LAR (Lariey, French Alps), and SPN (San Procolo a Naturno, Italian Alps), may naturally be explained by the circulation of soldiers and troops in Europe during the Thirty Years War (1618 to 1648), which made up human chains of transmission with historically documented epidemic events (12, 18, 19). For three strains (SPN from the Italian Alps, LAR from the French Alps, and BRA from Northern Germany), the relationship with the time of the Thirty Years War is historically and archaeologically documented (4, 7, 12). Human chains of transmission, which do not require the presence of rats to start and sustain an epidemic, might explain the circulation of the plague within Europe over long periods of time. They might be due to interpersonal contacts, crowding, infected parasites in clothes or goods (14, 20, 21), or contact with infected pets or fur (6). Several chains of human transmission within Europe could be reconstructed for cases of the last century (17, 22) as well as for the second pandemic (23, 24).

To better understand the evolution of Y. pestis, we examined two more mutations, which were recently discovered in ancient strains. In the most recent subclade of the second pandemic, starting with BED, there is a 49-kb deletion with unknown function. This deletion was also present in the last lineage of the first pandemic and, in both cases, might have accounted for the decline of the pandemic (4, 7, 25). We found the same mutation in the Rostov 2033 strain in the 18th century clade (Figs. 1 and 2). By contrast, a second strain found in the same cemetery in Rostov (Rostov 2039) has a different SNP pattern and lacks the chromosomal deletion.

Another mutation, the depletion of the pla gene on the plasmid pPCP1, has recently been proposed as the cause of the disappearance of the second plague pandemic in the 18th century (8) given that the pla gene is an important virulence factor of Y. pestis. We checked for the presence of the pla+/pla plasmids in all published ancient strains. The ratio in coverage depth between pla and the whole pPCP1 plasmid indicates the status of pla loss in an organism (Fig. 3). If the depth of pla is significantly lower than that of pPCP1, it might properly be concluded that the pla gene was lost in some pPCP1 plasmids. Our analyses show that the ratio of pla in the Black Death and postBlack Death genomes appears to be different when compared with the prehistoric and the first pandemic lineages (Fig. 1). We have also checked randomly selected modern Y. pestis genomes in different lineages, and their depth of pla and pPCP1 are quite consistent, indicating no other pla loss in modern plagues. By contrast, the generalized depletion of pla extensively observed during the postBlack Death era and at the end of the first pandemic (Fig. 1) seems to be consistent. Given that the sequencing data were generated by several different research groups, a systemic error during sequencing is unlikely.

The decay of the pla gene. (A) Depth plot of the pPCP1 plasmid in strain CHE1 using Integrative Genomics Viewer. The annotated genes of the pPCP1 plasmid are marked with blue bars. The average sequencing depth of whole pPCP1 plasmid is 195.65x, while the average sequencing depth of the pla region is 96.04x. (B) Group-wise comparison of the ratio between the depth of pla and that of whole pPCP1 plasmid among three waves of the second pandemic. Boxplots depict the upper, median, and lower quartiles of the ratios; individual dots indicate outliers that lie outside of 1.5 times the interquartile range; and vertical lines indicate the range of all ratios except for outliers. The P values of group-wise comparison using the Wilcoxon test are labeled on the top, two of which are statistically significant (P < 0.05). Data of B are provided in Dataset S4.

It seems that full pla strains were slightly depleted at the end of the second pandemic (8), with the same phenomenon at the end of the first pandemic. Notably, however, Rostov2033, one of the most recent genomes of the second pandemic, shows full read pPCP1 plasmids, whereas CHE, the most recent historical strain, shows very slight pla decay (Fig. 3). This observation is not fully in agreement with the proposed hypothesis that pla depletion contributed to the end of the pandemic. An alternative explanation for this phenomenon (8) is that the differences observed in the full pla plasmids might be due to different forms of plague. In particular, bubonic plague and pneumonic plague need the pla gene to develop, whereas primary septicemic plague does not (8). It seems that plague existed in all three forms, at least from the time of the first pandemic; however, this does not add any specific evolutionary information to the observed variability.

We propose an evolutionary hypothesis for the presence of lineages with pla decay. One of the optimized survival strategies for an emerging pathogen is to balance its virulence to the main host with its transmission strategy. This trade-off hypothesis was previously demonstrated for Y. pestis (26, 27). This mechanism would allow the bacterium to reduce virulence and enhance the time of survival of the host and, consequently, of the pathogen (28). After experiencing the Black Death and successive waves, the pla decay strains might have attempted to acquire a fitness advantage, reducing their virulence by increasing the time to death. Indeed, we observe among the victims only pla+/pla mixed strains, whereas pla lineages might have survived longer in the host population, providing a milder form of the illness. The Eastern European/Asia clade of the 18th century (including CHE1) further lost the 49-kb region, which can be the result of an extension of a virulence attenuated pattern. Such events of attenuated virulence might have occurred multiple times in the Y. pestis evolutionary history and left out host-adapted lineages, such as for 0.PE2 and 0.PE4 (29). Therefore, the possible virulence reduction caused by pla decay and loss of the 49-kb region is not necessarily the reason for the extinction of plague at the end of the first and second pandemics but might be the result of a form of adaptation to a new host, which may be the wild rodent in the putative Western European reservoir (Hypothesis 1), a new host in the Asian reservoir, or the human host (Hypothesis 2) as well as their vectors. We observed that the newly published strains from Lariey [French Alps (12)] do not show pla decay in contrast to other Alpine lineages (SPN). This evidence might exclude the hypothesis of an adaptation to a host in a Western European reservoir. Thus, we tentatively propose that this mechanism of pla decay would support the presence of human-to-human transmission chains mediated by human ectoparasites (fleas and body lice) during plague pandemics in Europe, the plausibility of which has previously been demonstrated (17, 22, 30), while the vector competence was supposed to be low (31).

See the rest here:
Assessing the origins of the European Plagues following the Black Death: A synthesis of genomic, historical, and ecological information - pnas.org