Ancient viral genomes reveal introduction of human pathogenic viruses into Mexico during the transatlantic slave trade
Targeted ancient DNA capture applied to skeletal remains from a Colonial hospital and chapel in Mexico City recovers three human parvovirus B19 genomes and one hepatitis B virus genome whose closest relatives are African strains, in hosts whose genetic ancestry and strontium isotope signatures point to African birth, giving direct molecular evidence that these viruses reached New Spain during the transatlantic slave trade.
- Senior authors
- María C Ávila Arcos; Daniel Blanco-Melo
- Correspondence
- Daniel Blanco-Melo; María C Ávila Arcos
Research areas & themes
Citation
Guzmán-Solís AA, Villa-Islas V, Bravo-López MJ, Sandoval-Velasco M, Wesp JK, Gómez-Valdés JA, Moreno-Cabrera ML, Meraz A, Solís-Pichardo G, Schaaf P, TenOever BR, Blanco-Melo D, Ávila Arcos MC. Ancient viral genomes reveal introduction of human pathogenic viruses into Mexico during the transatlantic slave trade. eLife 2021, volume 10, article e68612.
DOI 10.7554/elife.68612. PMID 34350829. PMCID PMC8423449.
One-sentence contribution
Targeted ancient DNA capture applied to skeletal remains from a Colonial hospital and chapel in Mexico City recovers three human parvovirus B19 genomes and one hepatitis B virus genome whose closest relatives are African strains, in hosts whose genetic ancestry and strontium isotope signatures point to African birth, giving direct molecular evidence that these viruses reached New Spain during the transatlantic slave trade.
Executive summary
The collapse of Indigenous populations after European colonisation of the Americas is attributed in part to introduced pathogens, but the identity of the agents responsible for the Colonial epidemics of New Spain has rested on historical accounts and on inference from modern viral strains rather than on material from the period. This study asks what can be recovered directly. Dental remains from 26 individuals were sampled at two Colonial sites in Mexico City, a hospital where mass burials indicate urgent simultaneous disposal of bodies and a chapel ten kilometres away. Shotgun sequencing with metagenomic classification identified samples carrying viral signal, and a designed panel of RNA probes targeting clinically important human viruses was used to enrich those libraries, with a virus-negative library carried through as a control. Enrichment yielded three human parvovirus B19 genomes and one hepatitis B virus genome, authenticated by best-hit filtering against the full nucleotide database and by the terminal deamination damage characteristic of ancient DNA. The hepatitis B virus genome belongs to sub-genotype A4, not previously reported in the Americas, and the parvovirus B19 genomes belong to genotype 3, which predominates in West Africa, whereas only genotype 1 had been described molecularly in Mexico. Ancestry analysis of the same sequencing data placed three hosts within African genetic variation with L mitochondrial haplogroups, strontium ratios indicated West African birth with later life in the Mexico City valley, and radiocarbon dates fell in the early Colonial period. A fourth individual, from the chapel site and of Indigenous ancestry, also carried genotype 3.
Scientific context
European colonisation of the Americas was followed by a severe collapse of Indigenous populations, attributed in part to introduced pathogens. Smallpox, measles and mumps have long been proposed as agents of the Colonial epidemics, and the deadliest outbreaks in New Spain, referred to as Cocoliztli, have no established cause. Prior work on those epidemics has rested on historical accounts, including autopsy descriptions from the Hospital Real de San José de los Naturales, on bioarchaeology and on phylogenetic inference from modern viral strains, not on direct recovery of viral material from the period. Ancient viral genomics had by then produced results for several viral families from Eurasian and some American contexts, including ancient hepatitis B virus and parvovirus B19 genomes, and tooth roots had been shown to be a usable source of pathogen DNA for blood-borne agents. Bioarchaeological and paleogenomic work on the hospital collection had already identified individuals of sub-Saharan African ancestry. The gap this study addresses is the absence of direct molecular evidence for specific viruses circulating in Colonial Mexico.
Central question
Which viral pathogens can be recovered directly from human remains buried in Colonial epidemic contexts in Mexico City, and do their genome sequences, read alongside the ancestry and geographic origin of the individuals who carried them, indicate introduction from Africa during the transatlantic slave trade?
Experimental strategy
The design joins three independent lines of evidence about the same individuals, which is what allows a claim about geographic introduction rather than merely about viral presence. Sampling targets dental remains, chosen because vascularised tooth roots can retain DNA from blood-borne agents, from two archaeologically distinct Colonial sites, a hospital associated with mass burials suggestive of urgent simultaneous disposal and a chapel ten kilometres away. Shotgun sequencing with metagenomic classification first identifies which samples carry any plausible viral signal, and a biotinylated RNA probe set then enriches for clinically important human viruses chosen on three criteria, DNA viruses already recovered from archaeological remains, representatives of families that integrate into the human genome, and RNA viruses with a DNA intermediate. A virus-negative library is carried through capture as the control that measures enrichment specificity. Authentication proceeds by several separate tests rather than one, including retention only of reads whose best database hit is the virus in question, characteristic deamination damage at read termini, and coverage structure across the genome. Human ancestry is then read from the same shotgun data by mitochondrial haplogroup, principal component analysis against a reference panel and admixture analysis, while strontium isotope ratios in enamel against bone distinguish place of birth from place of later life, and radiocarbon dating places the individuals in time. Viral origin is finally tested by phylogenetic placement against modern genotype and sub-genotype diversity of known geographic distribution.
Key findings
- Metagenomic screening of 26 dental samples identified 17 with at least one normalised viral hit, with signals resembling Hepadnaviridae, Herpesviridae, Parvoviridae and Poxviridae, and twelve samples were carried forward to capture (Figure 1c).
- After capture, one library showed roughly a hundredfold increase in hepatitis B virus hits and three showed roughly fifty to two hundredfold increases in parvovirus B19 hits, while the negative control library showed negligible enrichment (Figure 1c).
- All recovered reads displayed cytosine to thymine accumulation toward the five prime terminus with a near-symmetrical guanine to adenine pattern at the three prime end, the damage signature expected of ancient DNA (Figure 2a).
- Three parvovirus B19 genomes were reconstructed at 92.37 to 99.1 percent coverage of the single-stranded coding region with average depths of 2.98 to 15.36, and one hepatitis B virus genome at 89.9 percent coverage and 30.8 average depth (Figure 2, b and c).
- Coverage was markedly higher over the double-stranded inverted terminal repeats of parvovirus B19 and lower over the single-stranded region of hepatitis B virus. The authors interpret this uneven structure two ways, as reflecting better postmortem survival of double-stranded DNA and as an argument that the reads come from replication intermediates rather than from virus integrated into the human genome, since integration would give even coverage. Both are interpretations of the coverage pattern.
- The hepatitis B virus genome carries the six-nucleotide core gene insertion characteristic of genotype A and clusters phylogenetically with sub-genotype A4, which had previously been recovered only from African individuals in Belgium and never in the Americas (Figure 3a).
- All three parvovirus B19 genomes fall in genotype 3. Two are similar to sub-genotype 3b sequences from immigrants in Germany and one is closer to a divergent sub-genotype 3a strain isolated from a child with severe anaemia in France (Figure 3b). Genotype 3 is predominant in West Africa, and only genotype 1 had previously been described molecularly in Mexico.
- Testing for temporal structure found little for the full hepatitis B virus phylogeny, a stronger signal within genotype A analysed alone, and a usable signal for parvovirus B19 across all three genotypes, corroborated by date randomisation. A dated coalescent analysis gave a parvovirus B19 substitution rate of 1.03 by 10 to the minus five substitutions per site per year under a strict clock, with divergence times from the most recent common ancestor of genotypes 1, 2 and 3 of 7.19, 2.11 and 3.64 thousand years.
- The three individuals from the hospital fall within African genetic variation by principal component analysis and carry L mitochondrial haplogroups (Figure 4a).
- Strontium ratios in tooth enamel from two hospital individuals, 0.71098 and 0.71109, resemble West African soils and rocks and values reported for first-generation Africans in the Americas, while their parietal and phalanx values, 0.70672 and 0.70755, resemble the Trans Mexican Volcanic Belt where Mexico City lies. The isotope data therefore support birth in West Africa and later life in New Spain.
- Radiocarbon dating placed two individuals at 1442 to 1608 and 1472 to 1625 CE, the early Colonial period when arrivals of enslaved people were highest.
- The individual from the chapel site, who carried a genotype 3 parvovirus B19 genome, clusters with present-day Mexicans and Peruvians and shows a Native American component by admixture analysis (Figure 4b). The authors describe this as the first genotype 3 genome obtained from a non-African individual and read it as evidence that the virus spread across ancestries after introduction.
Mechanistic model
This is a historical and phylogeographic study rather than a mechanistic one, and it establishes no disease mechanism. What the evidence supports is a chain of inference about origin. Viral genomes of African-associated lineages were recovered from individuals whose nuclear ancestry, mitochondrial haplogroup and enamel strontium signature indicate African birth, at a date when forced transport from Africa to New Spain was at its height, which the authors read as direct molecular evidence of introduction during the transatlantic slave trade. The recovery of one of those lineages from an individual of Indigenous ancestry at a separate site is read as onward transmission between populations. Several further propositions in the discussion are offered explicitly as possibilities rather than findings. The authors state that they cannot establish where the African-born individuals acquired either infection, whether in Africa, in the Americas or during the crossing, nor whether infection caused their deaths, nor the age at acquisition, nor whether the hepatitis B virus infection was acute or chronic or acquired vertically. The suggestion that parvovirus B19 infection might explain the skeletal anaemia markers in one individual is framed as possible and is accompanied by the acknowledgement that genetic anaemias cannot be excluded because the relevant loci were not covered. The suggestion that some cases recorded as measles in sixteenth-century Mexico might have been parvovirus B19 is presented as a hypothesis requiring further study and is accompanied by an explicit statement that the study does not reject the role of measles. The overlap between the radiocarbon dates and the hepatitis-like symptoms recorded in Cocoliztli autopsies is noted, with the authors stating that additional analyses are needed before any link is established.
Conceptual or technical advance
The study demonstrates that a designed capture panel spanning clinically important human DNA viruses, applied to dental remains from documented Colonial burial contexts, can recover near-complete viral genomes and that those genomes can be authenticated and placed phylogenetically with enough resolution to assign sub-genotype. Combining that recovery with host nuclear and mitochondrial ancestry, strontium isotopes and radiocarbon dating on the same individuals is what converts a viral sequence into evidence about human movement, and it provides a template for investigating introduced pathogens in other Colonial contexts. The ancient parvovirus B19 genotype 3 genomes also serve as calibration points that had not previously existed for that genotype, allowing dated coalescent inference. Substantively, the work moves the discussion of Colonial epidemic aetiology from historical inference and modern-strain phylogeography onto direct molecular evidence, while stopping short of assigning causation.
Relationship to the broader research program
The paper is led by the Ávila Arcos group at the Universidad Nacional Autónoma de México, with Daniel Blanco-Melo, formerly of the tenOever laboratory and here affiliated with Mount Sinai and the Fred Hutchinson Cancer Research Center, as co-corresponding author and a principal conceptual contributor. The stated contribution of Benjamin tenOever is resources and manuscript review, and he is neither a corresponding author nor a senior author, so this record is classified as collaborative rather than lab-led. The connection to the wider corpus runs through the alumnus rather than through a research theme, and the paper's questions, methods and material are those of ancient DNA and paleovirology rather than of the laboratory's work on virus-host interaction in living systems. Asserting a thematic continuity would be category 3 synthesis and is not supported by this paper alone.
Related publications
- Blanco-Melo and colleagues, 2020, and other tenOever laboratory work, no relationship asserted. Nothing in this paper's reference list or content links it to the tenOever laboratory's own published research, and the connection is one of personnel rather than of scientific lineage.
The list is deliberately empty of corpus links. The references this paper builds on are from the ancient DNA, paleovirology, bioarchaeology and virus phylogenetics literature produced by other groups.
Limitations and boundaries
Four viral genomes from four individuals, drawn from 26 sampled, is a small basis for statements about what circulated in Colonial Mexico City, and the authors say that larger sample sizes from varied contexts are needed. Coverage is thin for two of the parvovirus B19 genomes, with average depths under four over the coding region. The capture panel covers only DNA viruses and RNA viruses with a DNA intermediate, and for some families only a few genes were tiled because of probe kit size limits, so the leading candidate agents of the Colonial epidemics, which are RNA viruses, could not be sought at all, a constraint the authors state and attribute to rapid RNA degradation. Sampling was biased by design, since hospital individuals were selected on morphological and dental indicators previously interpreted as African ancestry, which shapes what ancestry and viral lineage combinations could be found. Strontium and radiocarbon data exist for only two individuals each, and not for all four virus-positive individuals. Temporal structure was weak for hepatitis B virus across genotypes, so no dated analysis was performed for that virus, and the parvovirus B19 divergence estimates rest on one clock model choice among two reported. Genotype geography is inferred from modern sampling that is itself uneven, and sub-genotype A4 in particular has a contested origin resting on sequences from African immigrants in Europe. No causal link to any epidemic, to Cocoliztli or to the individuals' deaths is established, and the authors repeatedly say so. The authors also note that both viruses have been found in ancient DNA datasets with no association to disease or epidemics.
Audience summaries
25 words
Viral DNA recovered from teeth in Colonial Mexico City burials yields African-lineage parvovirus B19 and hepatitis B virus genomes, direct evidence of introduction during the transatlantic slave trade.
75 words
Which pathogens reached the Americas after European colonisation has been argued mostly from historical records. Using a capture assay for ancient viral DNA on teeth from a Colonial hospital and chapel in Mexico City, the authors reconstructed three parvovirus B19 genomes and one hepatitis B virus genome. All belong to lineages associated with Africa, and the individuals carrying them show African ancestry and West African birth signatures. One Indigenous individual also carried the virus.
150 words
Targeted hybridisation capture applied to ancient DNA from 26 dental samples at two Colonial sites in Mexico City yielded three human parvovirus B19 genomes at 92 to 99 percent coverage of the coding region and one hepatitis B virus genome at 90 percent coverage, all authenticated by capture-negative controls, best-hit filtering and terminal deamination damage. The hepatitis B virus genome carries the genotype A core insertion and clusters with sub-genotype A4, never before reported in the Americas. The parvovirus B19 genomes fall in genotype 3, predominant in West Africa, whereas only genotype 1 had been described molecularly in Mexico. Three hosts fall within African genetic variation and carry L mitochondrial haplogroups, and enamel strontium ratios indicate West African birth against bone values matching the Mexico City valley, with radiocarbon dates in the early Colonial period. A fourth individual, of Indigenous ancestry from a separate site, also carried genotype 3.