Supplementary Materials for

Transcription

1 advances.sciencemag.org/cgi/content/full/2/6/e /dc1 Supplementary Materials for Synergistic roles of climate warming and human occupation in Patagonian megafaunal extinctions during the Last Deglaciation Jessica L. Metcalf, Chris Turney, Ross Barnett, Fabiana Martin, Sarah C. Bray, Julia T. Vilstrup, Ludovic Orlando, Rodolfo Salas-Gismondi, Daniel Loponte, Matías Medina, Mariana De Nigris, Teresa Civalero, Pablo Marcelo Fernández, Alejandra Gasco, Victor Duran, Kevin L. Seymour, Clara Otaola, Adolfo Gil, Rafael Paunero, Francisco J. Prevosti, Corey J. A. Bradshaw, Jane C. Wheeler, Luis Borrero, Jeremy J. Austin, Alan Cooper Published 17 June 2016, Sci. Adv. 2, e (2016) DOI: /sciadv This PDF file includes: Supplementary Methods Supplementary Results fig. S1. Map of South America showing sites where genetic data were recovered. fig. S2. Phylogeny of camelids based on 432 bp of mitochondrial control region data. fig. S3. Mitochondrial sequence data for Arctotherium. Legends for tables S1 to S6 References (61 65) Other Supplementary Material for this manuscript includes the following: (available at advances.sciencemag.org/cgi/content/full/2/6/e /dc1) table S1 (Microsoft Excel format). Fossil samples included in this study. table S2 (Microsoft Excel format). Samples for which DNA extraction failed and 14 C failed or was not attempted. table S3 (Microsoft Excel format). Primer sequences used to amplify mitochondrial genes. table S4 (Microsoft Excel format). Megafaunal 14 C ages were calibrated against the SHCal13 data set. table S5 (Microsoft Excel format). Published human 14 C ages were calibrated against the SHCal13 data set.

2 table S6 (Microsoft Excel format). Results of independent replication of ancient DNA sequences.

3 Supplementary Methods An example of the OxCal 4.2 code used (Hippidion) is given below: Plot() Curve("ShCal13","ShCal13.14c"); Outlier_Model("General",T(5),U(0,4),"t"); Sequence() Boundary("Hippidion arrival"); Phase("Hippidion") R_Date("OXA-9504", 10310, 160) Outlier("General", 0.05); R_Date("NUTA-1811", 10710, 100) Outlier("General", 0.05); R_Date("ACAD3613", 11570, 50) Outlier("General", 0.05); Boundary("Hippidion extinction "); Supplementary Results Phylogenetic reconstructions Phylogenetic reconstructions based on mitochondrial DNA (mtdna) sequence data revealed that the camelids present in southern Patagonia during the Late Pleistocene were genetically distinct from extant L. guanicoe (Fig. 2, fig. S2). This suggests that a previously unknown lineage of L. guanicoe (guanaco) became regionally extinct in southern Patagonia at the end of the Pleistocene, and was then replaced by a discrete population of guanaco from elsewhere, presumably to the north outside the sampling area used in this study. This finding contrasts with current models assuming a direct line of descent between Patagonian Late Pleistocene and modern L. guanicoe. These data suggest the regional extinction of guanaco in Patagonia almost led to the global extinction of the species, and challenges the long-held perception that the species survived relatively unscathed despite human hunting in the Late Pleistocene and early Holocene (61).

4 Mitochondrial sequence data also revealed a second, distinct group of camelids that formed a robust monophyletic clade separate from either guanaco or modern vicuña (Fig. 2). We confirmed the identity of the new group as L. gracilis, the gracile llama, using sequences generated from morphologically identified specimens from the province of Buenos Aires, Argentina (table S1). In contrast to earlier molecular analysis (17), our results suggest that L. gracilis was phylogenetically divergent from modern V. vicugna, the vicuña, consistent with morphological differences identified between the extinct and extant species (62, 63,64). We confirm the large jaguar Panthera onca mesembrina is an extinct subspecies (18, 65), and clearly divergent from extant jaguars (Fig. 2). In contrast, a haplotype network of modern and ancient puma (Puma concolor) revealed that Late Pleistocene individuals fall within the haplotype diversity of modern South American populations (Fig. 2). The newly analyzed five samples of Smilodon all exactly matched previously reported data (297 bp of mitochondrial DNA ND5) for this species (22). We compared the control region sequence for the two Arctotherium sp. samples and found only two substitutions (< 1% difference). Phylogenetic reconstructions revealed high support (BPP: 1.0, ML alrt: 97%) for the giant short-faced South American bear Arctotherium sp. within a monophyletic Tremarctine clade (fig. S3); however, our data were not sufficient to further resolve the phylogenetic relationships of Tremarctine bears. Independent replication of results Camelid samples: A comparison of independently replicated sequence data for 8 samples, spanning 3 camelid species (including extinct southern Patagonian L. guanicoe and L. gracilis), revealed only 5 mismatches between 3,070 bp of the original and independently replicated mitochondrial sequence data (error frequency of ) (table S6). Bear samples: One specimen (ACAD3599; Arctotherium sp; Cueva del Puma) was sampled on two separate occasions, and one of these two samples was extracted, cloned and sequenced at the Henry Wellcome Ancient Biomolecules Centre, Oxford (Jacobo Weinstock; unpublished data). The consensus of these clones was identical to the same fragment amplified and directly sequenced from the second sample at ACAD.

5 fig. S1. Map of South America showing sites where genetic data was recovered from Late Pleistocene (stars) and Holocene (circles) samples of bone or teeth. Sites include: 1- Tres Arroyos, 2 - Cueva de los Chingues, 3 - Cueva del puma, 4 - Cueva Lago Sofia 4, 5 - Cueva del Medio, 6 - Cueva del Milodón, 7 - Casa de Piedra, 8 - Campo Moncada, 9 - Agua de Pérez, 10 - Cueva Arroyo Colorado, 11 - Arroyo Malo #3, 12 - Agua de la Cueva, 13 Pasqualama, 14 Telarmachay, 15 - Cueva Rosello.

6 fig. S2. Phylogeny of camelids based on 432 bp of mitochondrial control region data. Mitochondrial DNA sequences identify 17 bone and teeth fragments from the Pleistocene Patagonian environment representing L. gracilis, and 25 bone and teeth fragments representing an extinct lineage of L. guanicoe. We also surveyed 41 bone and teeth from Holocene camelids across present-day Argentina, Chile, and Perú, and discovered only haplotypes falling within the diversity of modern L. guanicoe. We also included sequence data from three Holocene vicugna samples.

7 fig. S3. Mitochondrial sequence data for Arctotherium confirms a phylogenetic placement within Tremarctinae. Additional sequencing effort is required to clarify the sister taxon.

8 table S1. Fossil samples included in this study. We include radiocarbon data, lab number, 13 C, 15 N, C:N ratio, and whether DNA was successfully recovered. We also include museum or institute from where the fossil was subsampled, as well as other notes such as museum ID of fossil. At the bottom of the table, we sum number of samples with radiocarbon data and DNA sequence data. table S2. Samples for which DNA extraction failed and 14 C failed or was not attempted (samples not included in any analysis). table S3. Primer sequences used to amplify mitochondrial genes (CR = control region, ATP8 = Adenosine triphosphate synthase 8, Cytb = cytochrome b, trna Pro and Thr = trna Proline and Threonine). table S4. Megafauna 14 C ages were calibrated against the Southern Hemisphere calibration (SHCal13) dataset. For extinct Pleistocene megafauna, the radiocarbon ages were used to constrain the estimates of extinction ages for megafaunal taxa using the Phase model option in OxCal 4.2 with General Outlier analysis detection (probability=0.05). table S5. Published human 14 C ages were calibrated against the Southern Hemisphere calibration (SHCal13) dataset, and arrival ages were estimated using the Phase model option in OxCal 4.2 with General Outlier analysis detection (probability=0.05). table S6. We show the number of differences between samples sequenced at the Australian Centre for Ancient DNA, and independently sequenced mitochondrial DNA fragments for a subset of samples. Overall, sequencing results were confirmed with few discrepancies between labs (average error per base ).