http://orcid.org/0000-0001-6826-6401
Figures
Abstract
Many archaeological science studies use the concept of “provenance”, where the origins of cultural material can be determined through physical or chemical properties that relate back to the origins of the material. Recent studies using DNA profiling of bacteria have been used for the forensic determination of soils, towards determination of geographic origin. This manuscript presents a novel approach to the provenance of archaeological minerals and related materials through the use of 16S rRNA sequencing analysis of microbial DNA. Through the microbial DNA characterization from ochre and multivariate statistics, we have demonstrated the clear discrimination between four distinct Australian cultural ochre sites.
Citation: Lenehan CE, Tobe SS, Smith RJ, Popelka-Filcoff RS (2017) Microbial composition analyses by 16S rRNA sequencing: A proof of concept approach to provenance determination of archaeological ochre. PLoS ONE 12(10): e0185252. https://doi.org/10.1371/journal.pone.0185252
Editor: Dong Hoon Shin, Seoul National University College of Medicine, REPUBLIC OF KOREA
Received: September 29, 2016; Accepted: September 8, 2017; Published: October 18, 2017
Copyright: © 2017 Lenehan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files. In addition, a full data set is available at the Harvard Dataverse: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/FUB62C.
Funding: This work was supported by the Australian Institute of Nuclear Science and Engineering (AINSE) Fellowship (R.P.) www.ainse.edu.au. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
A fundamental question in archaeological studies is the concept of “provenance”, where the origins of a material or artefact can be characterized and determined [1]. Ascertaining provenance can provide information on the object’s movement from the original source due to cultural exchange. Analytical methods capable of determining inherent characteristics of the original deposit or source, give valuable information that can assist with source attribution and in reconstructing past exchange. Ochre is a complex mineral pigment frequently observed in archaeological sites worldwide, and is often of unknown provenance. The term “ochre” is broadly used to describe natural iron oxide based pigments that range in color from red to purple, as well as yellow to brown, depending on the type and amount of iron oxide present and the mixture of other minerals and organic materials [2–7]. The major iron oxides present are hematite [Fe2O3] and goethite [FeO(OH)x or α-FeOOH]. It is well known that ochres were widely traded in the Australian Aboriginal context, and recent studies have demonstrated trace elemental analysis by neutron activation analysis (NAA) of ochre for determining this “trace elemental fingerprint” with a view for provenancing ochre samples [8–13]. In addition to these studies, researchers have also used techniques such as various forms of mass spectrometry, X-ray diffraction and quartz-isotope ratios, especially for Australian ochre [14–17], Elemental and mineralogical profiling have been established as a viable methodology for provenance studies of complex mineralogical samples. However, to date no research to date has examined the possibility of the use of soil microbiology to provenance complex archaeological minerals. These novel 16S rRNA sequencing methods offer potential to understand the microbial “fingerprint” of a source that also includes the source history. Data from 16S rRNA sequencing studies provide profiles of ochre sources that are both independent of and complementary to elemental and mineralogical analyses.
The DNA profile of microrganisms within samples is a rich source of information that can be exploited to further understand archaeological problems. DNA profiling has been used for paleoclimatic reconstruction [18] and understanding environmental change [19]. To our knowledge, DNA profiling has not been applied to archaeological provenancing. Recent research has demonstrated the utility of DNA profiling of microbial communities for forensic discrimination (provenance) of visually similar soils [20, 21]. DNA profiles from the entire genetic composition (bacteria, fungi, plant material etc.) within the soil generate a rich data set for meaningful comparison and discrimination of samples. These results can be combined with those of trace elemental analysis for further resolution [20].
In order to investigate the powerful potential of using microbial DNA profiles for archaeological provenancing, a proof of concept study utilizing the 16S rRNA profile of ochres from four well-documented Australian sites were sequenced and compared [9, 10, 16, 22, 23]. This study focuses on the provenance of ochre minerals, however this method has potential for characterising and understanding other related cultural material.
These sites represent varying geological characteristics as well as varying site histories, both geological and cultural. (Fig 1) Wilgie Mia (sample OCH117), a weathered banded iron formation in Western Australia, is one of the most well known Australian ochre deposits due to the quality of the ochre and the extensive scale of the ochre excavations over the past thousands of years [24]. Similarly, Bookartoo (sample OCH255, Flinders Ranges, South Australia), and Karrku (sample OCH095, Northern Territory) are also banded formations of high quality ochre [25]. In contrast, Moana (sample OCH037) is an exposed coastal site in South Australia and subject to weathering.
Modified from http://d-maps.com/carte.php?num_car=3289&lang=en.
Microbiological profiling of soils has been demonstrated to allow discrimination between different, but visually similar, soil types in forensic science as well as environmental microbial ecology [21, 26, 27]. The factors that contribute to bacterial community composition have been shown to include: pollutants, aquatic or terrestrial substrate, and environmental factors [28–32]. Based on successful discrimination of soils for forensic and ecological purposes, this avenue of investigations offers an exciting possibility for archaeological provenance.
Materials and methods
Site description and sampling
Four samples from the four sites (Bookartoo [OCH255], Karrku [OCH095], Moana [OCH037] and Wilgie Mia [OCH117]) were selected. All samples used were obtained either through the South Australian Museum or directly from the researcher who collected the sample. In either case, samples had been packaged individually. No permits were required for the described study. The project has approval number 4670 from the Social and Behavioural Research Ethics Committee of Flinders University.
Once in the preparation laboratory, samples were handled with nitrile gloves that were changed between samples. The surface material was removed prior to preparation and the sample preparation continued with a freshly exposed area of sample [16, 22, 23].
Sample preparation, microbial community DNA extraction and sequencing
The ochre samples were powdered in an agate mortar and pestle and dried in a desiccator. Stringent methods were used in cleaning materials between samples, which included disposable paper and plastic materials where possible; and remaining items cleaned thoroughly with pyroneg detergent, deionized water, ethanol, and dried between samples. Approximately 3 grams of each was divided into three 1-gram portions, for a total of 12 samples (3 samples times the 4 sites investigated). These samples were prepared into individual sterile glass vials for storage prior to molecular work.
DNA extraction took place in a dedicated extraction and PCR set-up laboratory with protocols in place to ensure no PCR products were present. All benches and equipment were sterilized prior to use with a 3% bleach wash followed by pure molecular grade ethanol. All reagents and consumables were sterilized with ultra-violet light before use in a crosslinker. To ensure quality control, negative samples were processed at the same time as the ochre samples and subjected to the same scientific procedures prior to sequencing.
Microbial community DNA was extracted using the PowerSoil DNA Isolation Kit (MoBio laboratories, Inc., Carlsbad, CA, USA). DNA quality and concentration were determined by 1.5% TBE agarose gel electrophoresis and by using a Qubit fluorometer (Qubit dsDNA HS Assay Kit; Life Technologies). Control samples did not show any gel product and their Qubit reading was below detection limits/extremely low. Some carry over DNA from the bacteria, from which the DNA polymerase is derived, is expected and unavoidable, so this very low signal from the negative control was within acceptable laboratory limits.
High molecular weight DNA from the ochre samples was then sent to the Molecular Research LP (MR DNA; Texas, USA) for 16S rRNA gene based sequences on the Illumina MiSeq platform using the MiSeq Reagent Kit v3 (600 cycle) (Illumina). Bacterial diversity of ochre samples was determined by amplification of the 16SS rRNA gene using the primers 27F (5’-AGRGTTTGATCMTGGCTCAG -3’) and 519R (5’-GTNTTACNGCGGCKGCTG -3’).
The Paired-End reAd merger (PEAR) v.0.9.5 [33] was used to pair forward and reverse reads from each ochre sample. Merged DNA sequences were processed using Quantitative Insights Into Microbial Ecology (QIIME) [34] v.1.8.0 and UPARSE [35] as previously described, however without the removal of singletons The quality filtering criteria were a minimum 200bp in length, minimum quality score of 30, no mismatches in the primers sequence and no more than 6 ambiguous bases. USEARCH [36] was used to perform filtering of duplicate sequences and chimera removal. The remaining sequences were clustered into operational taxonomic units (OTUs) based on sequence similarity using uclust and Greengenes database (13_08) as a references in QIIME, with a minimum identity cutoff of 97%. Sequences were rarefied to the lowest sequence number of 5199 to remove any bias based on differences in sequencing depth.
The full data set is available at the Harvard Dataverse: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/FUB62C
Data analysis
Differences in overall taxonomic composition between the ochre samples were analyzed using the PERMANOVA+ version 1.0.3 3 add-on to PRIMER [37, 38]. Non-metric Multi-Dimensional scaling (NMDS) of Bray-Curtis similarities was performed as an unconstrained ordination method to graphically visualize the multivariate patterns in the taxa associated with the ochre samples.
Canonical analysis of principal coordinates (CAP) [38] on the sum of squared canonical correlations was used as a constrained ordination to determine whether there were any significant differences between microbial community compositions between the ochre samples. The a priori hypothesis that the taxonomic compositions between the ochre samples were different was tested in CAP by obtaining a P-value using 9999 permutations. CAP ordinations were generated using order level taxonomic classifications.
Where significant differences were found using CAP, the percent contribution of each taxa to the separation between ochre samples were assessed using similarity percentage (SIMPER) analysis [39]. The resulting top 90 percent of all taxa were used to determine those taxa that were driving the dissimilarity between ochre samples.
Results/discussion
All samples had data returned with the exception of one of the Karrku samples (OCH95), of which one replicate did not amplify (S1 and S2 Tables). Data were analyzed using CAP to examine the hypothesis that the ochre samples were the same, as well as the NMDS as an unconstrained ordination (no hypothesis to constrain the data). Unconstrained (NMDS) and constrained (CAP) multivariate analysis demonstrated a clear separation of data (P-value = 0.0003) between the ochre samples (Figs 2 and 3 and Tables 1 and 2).
NMDS ordination derived from a Bray-Curtis similarity matrix calculated from the square-root transformed abundance of 16S rRNA sequences matching the Greengenes database (13_08), order level.
CAP analysis (using m = 4 principal coordinate axes) is derived from the sum of squared canonical correlations of 16S rRNA sequences matching the Greengenes database (13_08), order level.
The clear separation in statistical space between data points in the NMDS and the CAP plots indicate that the ochre samples can be readily distinguished based on taxonomic composition. In addition, the replicate samples from the ochre sources cluster very closely within the groups, further supporting the evidence of low levels of variability between replicates of the same sample [21]. Similar results have been observed in related soil types. The differences in the ochre groups can also be observed as reflecting the origins of their original sites and site genesis. For instance, while both Bookartoo and Wilgie Mia are banded iron formations, each has a distinct site history over geological time. In the case of Bookartoo, the ochre developed as the weathering of pyrite in dolomitic beds, while Wilgie Mia ochre was thought to be hematite oxidized from magnetite [14]. In contrast, Moana is from a coastal, weathered sedimentary formation [14]. Each of these sites, due to their geographic origin and subsequent genesis over geological time, will have a different microbial DNA signature [21]. Separation seen in both NMDS and CAP suggests that the data points are not simply conforming to the more hypothesis-driven CAP analysis.
SIMPER analysis supports the separation of data with clear shifts in the taxonomic composition between ochre samples (Fig 4 and S1 Fig).
Only taxa that were consistently contributing (i.e. Diss/SD > 1.4) to the dissimilarity have been included. Cut off percentage is 90%. Percent contributions were calculated based on square root transformed relative abundance values. The legend can be found in the SI.
Previous analysis shows that environmental factors are a major contributor to the composition of microbial communities [30, 32, 40]. Here, SIMPER revealed 46 taxa that were driving the dissimilarity between ochre samples, demonstrating a powerful discriminatory method for separating ochre samples. Differences in environmental conditions could influence the microbial communities structure for each of the four sites investigated [32].
Conclusion
This study demonstrates the novel application of 16S rRNA sequencing analysis of microbial DNA towards identifying the provenance of archaeological samples, in this case Indigenous ochre samples from four locations. The use of the microbial data and statistical analysis reveals that despite time and sample movement away from its origins, microbial information in a given soil or geological sample can be used as a “signature” for the original source of the material. The 16S rRNA sequencing data in conjunction with elemental data has the potential to enhance provenance determination for complex archaeological samples such as ochre and soils. More case studies can be examined to evaluate larger data sets and samples from archaeological contexts, including multiple samples from the same/differing site and strata, however we have demonstrated that trace microbial content in archaeological samples provide a yet unexploited source of information for archaeological provenance studies.
Supporting information
S1 Fig. Plot showing the relative abundance in each sample of 70 different orders.
https://doi.org/10.1371/journal.pone.0185252.s001
(PDF)
S1 Table. Number of sequences obtained for each ochre sample.
https://doi.org/10.1371/journal.pone.0185252.s002
(PDF)
S2 Table. Order level relative proportion of matches to the Greengenes (13_08) database.
https://doi.org/10.1371/journal.pone.0185252.s003
(PDF)
Acknowledgments
We gratefully acknowledge the South Australian Museum Board and South Australian Museum Aboriginal Advisory Group for support to access and analyze the collections. We also thank the staff at the South Australian Museum Aboriginal Australian Collections for access to the collections. We are also grateful to Professor Mike Smith (National Museum of Australia) and other collaborators for access to their collections for analysis. Funding is gratefully acknowledged from Australian Institute of Nuclear Science and Engineering (AINSE) Research Fellowship (Popelka-Filcoff). Data are available in the Supplementary Materials.
References
- 1.
Weigand PC, Harbottle G, Sayre EV, editors. Turquoise Sources and Source Analysis: Mesoamerica and the Southwestern U.S.A. New York: Academic Press; 1977.
- 2.
Cornell RM, Schewertmann U. The Iron Oxides. Weinheim: Wiley-VCH Verlag GmbH &Co.; 2003.
- 3. Elias M, Chartier C, Pre´vot G, Garay H, Vignaud C. The colour of ochres explained by their composition. Materials Science and Engineering B: Solid-State Materials for Advanced Technology. 2006;127(1):70–80.
- 4. Genestar C, Pons C. Earth pigments in painting: Characterisation and differentiation by means of FTIR spectroscopy and SEM-EDS microanalysis. Analytical and Bioanalytical Chemistry. 2005;382(2):269–74. pmid:15761735
- 5. Hradil D, Grygar T, Hradilova J, Bezdicka P. Clay and iron oxide pigments in the history of painting. Applied Clay Science. 2003;22(5):223–36.
- 6. Montagner C, Sanches D, Pedroso J, Melo MJ, Vilarigues M. Ochres and earths: Matrix and chromophores characterization of 19th and 20th century artist materials. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy. 2013;103(0):409–16. http://dx.doi.org/10.1016/j.saa.2012.10.064.
- 7. Mortimore JL, Marshall LJR, Almond MJ, Hollins P, Matthews W. Analysis of red and yellow ochre samples from Clearwell Caves and Cätalhöyük by vibrational spectroscopy and other techniques. Spectrochimica Acta—Part A: Molecular and Biomolecular Spectroscopy. 2004;60(5):1179–88.
- 8.
Popelka-Filcoff RS, Craig N, Glascock MD, Robertson JD, Aldenderfer M, Speakman RJ. INAA of Ochre Artifacts from Jiskairumoko, Peru. In: Glascock MD, Speakman RJ, Popelka-Filcoff RS, editors. Archaeological Chemistry: Analytical Techniques and Archaeological Interpretation. ACS Symposium Series. Washington D.C.: American Chemical Society; 2008.
- 9. Popelka-Filcoff RS, Lenehan C, Walshe K, Bennett JW, Stopic A, Jones P, et al. Characterisation of Ochre Sources in South Australia by Neutron Activation Analysis (NAA). Journal of the Anthropological Society of South Australia. 2012;35(Special Edition: Archaeological Science in South Australia):81–90.
- 10. Popelka-Filcoff RS, Lenehan CE, Glascock MD, Bennett JW, Stopic A, Quinton JS, et al. Evaluation of relative comparator and k 0-NAA for characterization of Aboriginal Australian ochre. Journal of Radioanalytical and Nuclear Chemistry. 2012;291(1):19–24.
- 11. Popelka-Filcoff RS, Lenehan CE, Lombi E, Donner E, Howard DL, Jonge MDd, et al. Microelemental characterisation of Aboriginal Australian natural Fe oxide pigments. Analytical Methods. 2015;7(Emerging Investigators):7363–80
- 12. Popelka-Filcoff RS, Miksa E, Robertson JD, Glascock MD, Wallace H. Elemental Analysis and Characterization of Ochre Sources from Southern Arizona. Journal of Archaeological Science. 2008;35:752–62.
- 13. Popelka-Filcoff RS, Robertson JD, Glascock MD, Descantes C. Trace Element Characterization of Ochre from Geological Sources Journal of Radioanalytical and Nuclear Chemistry. 2007;272(1):17–27.
- 14. Jercher M, Pring A, Jones PG, Raven MD. Rietveld X-Ray Diffraction And X-Ray Fluorescence Analysis Of Australian Aboriginal Ochres. Archaeometry. 1998;40(2):383–401.
- 15. Scadding R, Winton V, Brown V. An LA-ICP-MS trace element classification of ochres in the Weld Range environ, Mid West region, Western Australia. Journal of Archaeological Science. 2015;54(0):300–12. http://dx.doi.org/10.1016/j.jas.2014.11.017.
- 16.
Smith M, Fankhauser B. Geochemistry and identification of Australian red ochre deposits Canberra: National Museum of Australia and Centre for Archaeological Research, 2009 September 2009. Report No.
- 17. Smith MA, Pell S. Oxygen-Isotope Ratios in Quartz as Indicators of the Provenance of Archaeological Ochres. Journal of Archaeological Science. 1997;24:773–8.
- 18. Grund BS, Williams SE, Surovell TA. Viable paleosol microorganisms, paleoclimatic reconstruction, and relative dating in archaeology: a test case from Hell Gap, Wyoming, USA. Journal of Archaeological Science. 2014;46(0):217–28. http://dx.doi.org/10.1016/j.jas.2014.02.010.
- 19. Mitusov AV, Mitusova OE, Pustovoytov K, Lubos CC-M, Dreibrodt S, Bork H-R. Palaeoclimatic indicators in soils buried under archaeological monuments in the Eurasian steppe: a review. The Holocene. 2009;19(8):1153–60.
- 20. Concheri G, Bertoldi D, Polone E, Otto S, Larcher R, Squartini A. Chemical Elemental Distribution and Soil DNA Fingerprints Provide the Critical Evidence in Murder Case Investigation. PLoS ONE. 2011;6(6):e20222. pmid:21674041
- 21. Khodakova AS, Smith RJ, Burgoyne L, Abarno D, Linacre A. Random Whole Metagenomic Sequencing for Forensic Discrimination of Soils. PLoS ONE. 2014;9(8):e104996. pmid:25111003
- 22.
Nobbs MF. Aboriginal painters of the Olary District of South Australia an ethnoecological study of the Lake Frome Plains and the adjoining uplands, with particular reference to the granite area of the Olary Uplands / By Margaret Nobbs.—History: 1995.; 1996.
- 23.
Smith MA, Fankhauser B. An archaeological perspective on the geochemistry of Australian red ochre deposits: Prospects for fingerprinting major sources. Canberra: Australian Institute of Aboriginal and Torres Strait Islander Studies, 1996.
- 24. Clarke J. Two Aboriginal Rock Art Pigments from Western Australia: Their Properties, Use and Durability. Studies in Conservation. 1976;21:134–42.
- 25. Paterson N, Lampert R. A Central Australian Ochre Mine. Records of the Australian Museum. 1985;37:1–9.
- 26. Horswell J, Cordiner SJ, Maas EW, Martin TM, Sutherland KBW, Speir TW, et al. Forensic comparison of soils by bacterial community DNA profiling. Journal of forensic sciences. 2002;2(47):350–3.
- 27. Meyers MS, Foran DR. Spatial and Temporal Influences on Bacterial Profiling of Forensic Soil Samples*. Journal of forensic sciences. 2008;53(3):652–60. pmid:18471210
- 28. Smith RJ, Jeffries TC, Adetutu EM, Fairweather PG, Mitchell JG. Determining the metabolic footprints of hydrocarbon degradation using multivariate analysis. 2013.
- 29. Chau JF, Bagtzoglou AC, Willig MR. The effect of soil texture on richness and diversity of bacterial communities. Environmental Forensics. 2011;12(4):333–41.
- 30. Dinsdale EA, Edwards RA, Hall D, Angly F, Breitbart M, Brulc JM, et al. Functional metagenomic profiling of nine biomes. Nature. 2008;452(7187):629–32. pmid:18337718
- 31. Willner D, Thurber RV, Rohwer F. Metagenomic signatures of 86 microbial and viral metagenomes. Environmental microbiology. 2009;11(7):1752–66. pmid:19302541
- 32. Jeffries TC, Seymour JR, Gilbert JA, Dinsdale EA, Newton K, Leterme SS, et al. Substrate type determines metagenomic profiles from diverse chemical habitats. PloS one. 2011;6(9):e25173. pmid:21966446
- 33. Zhang J, Kobert K, Flouri T, Stamatakis S. PEAR: a fast and accurate Illumina Paired-End reAd mergeR. Bioinformatics. 2014;30:614–20. pmid:24142950
- 34. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7(5):335–6. pmid:20383131
- 35. Edgar RC. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Meth. 2013;10(10):996–8. http://www.nature.com/nmeth/journal/v10/n10/abs/nmeth.2604.html—supplementary-information. pmid:23955772
- 36. Edgar RC. Search and clustering orders of magnitude faster than BLAST. Bioinformatics. 2010;26:2460–1. pmid:20709691
- 37. Anderson MJ, Robinson J. Permutation tests for linear models. Australian & New Zealand Journal of Statistics. 2001;43(1):75–88.
- 38.
Anderson MJ, Gorley RN, Clarke KR. PERMANOVA+ for PRIMER: Guide to Software and Statistical Methods. PRIMER-E, Plymouth UK. 2008.
- 39. Clarke KR. Nonparametric multivariate analysis of changes in community structure. Australian Journal of Ecology. 1993;18(1):117–43.
- 40. Smith RJ, Jeffries TC, Adetutu EM, Fairweather PG, Mitchell JG. Determining the metabolic footprints of hydrocarbon degradation using multivariate analysis. PLoS ONE. 2013;8(11):e81910. pmid:24282619