Analysis of Bacterial DNA in Skin and Muscle of the TyroleanIceman Offers New Insight Into the Mummification Process

FRANCO ROLLO,1* STEFANIA LUCIANI,1 ADRIANA CANAPA,2AND ISOLINA MAROTA1

1Dipartimento di Biologia Molecolare, Cellulare e Animale,Universita di Camerino, I-62032, Camerino, Italy2Istituto di Biologia e Genetica, Universita di Ancona,I-60131, Ancona, Italy

KEY WORDS iceman; mummification; cadaveric microflora;C. algidicarnis; DNA; Alps

ABSTRACT About 80 sequences (16s ribosomal RNA gene) of bacterialDNA in samples of skin and muscle taken directly from the Tyrolean iceman(3350–3100 years B.C.) or recovered during the 1992 archaeological expedi-tion at the Alpine site were analyzed to obtain clues to the natural mummifi-cation process that allowed the corpse of the Neolithic shepherd/hunter to bepreserved for more than 5,000 years. The investigation was made morecomplex by the fact that the surface of the mummy had been swabbed withphenol soon after the discovery (September 19, 1991). Our results show thatno trace of microbial DNA is left on the actual surface of the body, while theuntreated skin still bears the remains of large numbers of bacteria belongingto the genera Sphingomonas, Afipia, Curtobacterium, Microbacterium, Agro-myces, and others. Compared to the untreated skin, the iceman’s muscle isalso very rich in bacterial DNA. However, this DNA comes, with fewexceptions, from the species Clostridium algidicarnis. The sharp difference inthe bacterial DNA composition of skin and muscle suggests that the remainsof the original cadaveric microflora of the latter have not disappeared duringthe iceman’s taphonomic history. On the other hand, the massive presence ofC. algidicarnis, a cold-adapted sporigenous, the DNA of which was previously(Ubaldi et al. [1998] Am. J. Phys. Anthropol. 107:285–295) found in the softtissue of a naturally desiccated Andean mummy, indicates that the hypothesisthat the iceman’s corpse underwent rapid dehydration by the effect of a warmwind (fohn) is no longer plausible. The results best fit with the hypothesis(Bereuter et al. [1997] Chem. Eur. J. 7:1032–1038) that the body was firstcovered by snow and ice, and then underwent thawing and, finally, desicca-tion. Am J Phys Anthropol 111:211–219, 2000. r 2000 Wiley-Liss, Inc.

The so-called Tyrolean iceman or ‘‘Simi-laun man’’ or ‘‘Otzi’’ is a mummified humancorpse found in an Alpine glacier on Septem-ber 19, 1991 close to the Austrian/Italianborder (actually 92.6 m inside Italian terri-tory). The conventional radiocarbon age ofthe remains is 4,546 6 17 years B.P. basedon nine independent measurements (Pri-noth-Fornwagner and Niklaus, 1995; Rolloet al., 1995a). This results in a real age

range (calibrated) between 3350–3100 B.C.,corresponding to Late Neolithic times. Oneof the most relevant features of this archaeo-

Grant sponsor: Italian Ministry of University and ScientificResearch; Italian National Research Council; Grant number:97.00611.PF36.

*Correspondence to: Dr. Franco Rollo, Dipartimento di Biolo-gia Molecolare, Cellulare e Animale, Universita di Camerino,I-62032 Camerino, Italy. E-mail: rollo@cambio.unicam.it

Received 8 March 1999; accepted 31 August 1999.

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 111:211–219 (2000)

r 2000 WILEY-LISS, INC.

logical discovery is the exceptional state ofpreservation of the mummy and of the cloth-ing and equipment found on and near it.Another feature of extraordinary interestfor the anthropologist is that the man, possi-bly a shepherd/hunter, died accidentally (andalone in all probability) during his journeyacross the Alps. As a direct consequence, theaccompanying goods are not the reflection ofany cult ritual, but are fully representativeof things in daily use at the time. Theyinclude, among many other items, extremelyperishable ones such as a large fragment of acloak made of knotted tufts of grass, a furhat, and yew bow with arrows and quiver(Spindler, 1995; De Marinis and Brillante,1998).

The taphonomic history of the iceman hasbeen the object of much speculation. Inparticular, several hypotheses have beenmade regarding the process of mummifica-tion. Some of them, such as the fantasticclaim that the iceman might represent anelaborate hoax in which a pre-Dynastic Egyp-tian mummy had been planted togetherwith artifacts at the site, were soon aban-doned (Bahn and Everett, 1993). Othershave been kept and are discussed below.

At the moment of discovery by two Ger-man tourists, the corpse lay in a chamber-like depression at 3,213 m above sea level,below a rocky ledge, sheltered from theshearing flow of glacial ice. Presumably, thisparticular situation had prevented the bodyfrom being crushed and then expelled withthe regular glacial turnover in the course ofcenturies, as normally happens to glacier-trapped corpses and artifacts (Meyer, 1992).As concerns the mummification process, oneof the first hypotheses was that the corpsehad undergone rapid dehydration by a warmwind (an autumn fohn) and had been subse-quently covered by snow. A warm windyspell soon followed by a mimetizing snowfallseemed necessary in order to explain howthe body could become air-mummified fastenough not to be discovered and attacked byscavengers (vultures or raven-type birds). Awind of that heat at more than 3,000 m ofaltitude, however, has seemed unlikely tosome meteorologists, as the fohn blows at amuch lower altitude. An alternative specula-tion was that the body and equipment be-

came rapidly frozen, and were then coveredby a porous layer of snow which allowed thebody to slowly desiccate through combinedfreeze-drying (Spindler, 1995). This recon-struction is mainly based on the glaciologi-cal observation that snow can remain air-permeable for several years (Spindler, 1995).Bahn (1996), on the other hand, suggestedthat the corpse was preserved in the sameway as the many frozen carcasses of mam-moths and other Ice Age animals in Siberiaand Alaska. They were preserved by thebuildup of ice in the sediments that envel-oped the bodies: the ice layers desiccated thesoil and dehydrated the carcasses. Unlikefreeze-drying, where the original form re-mains intact, this process shrivels the body.Fascinating though this model may be, itdoes not take into account the fact that theiceman’s body was not embedded in soil.

The picture is further complicated by anumber of features which have been progres-sively revealed by the investigations carriedout independently in several laboratoriessince the first exploration of the iceman’ssite. For example, histological and biochemi-cal analyses have shown an almost completeloss of the iceman’s epidermis, accompaniedby postmortem alterations of skin triacylglyc-erols, which seems to imply a prolongedpermanence in a humid or waterlogged site(Bereuter et al., 1997).

In a previous paper (Ubaldi et al., 1998),we proposed a method based on the analysisof bacterial DNA sequences in soft-tissuesamples of a Peruvian mummy as a reliableway of identifying the major bacterial taxain the intestinal microflora of a person wholived in the Andes about 1,000 years ago.The same method can also be used to deter-mine the composition of the cadaveric florain corpses and carcasses of any age (Rollo,1998; Rollo and Marota, 1999).

For the last 8 years, the Tyrolean icemanhas been the object of many different typesof investigation. In particular, DNA analy-ses performed by research teams in Munichand Oxford (Handt et al., 1994) have shownthat the original DNA of the mummy is, tosome extent, preserved. Further evidence onthe preservation of genuinely ancient DNAat the iceman’s site has come from theanalysis of the grass clothing (cloak and

212 F. ROLLO ET AL.

boots) conducted in our laboratory (Rollo etal., 1995a–c).

Practically, we could use specimens ofmuscle as a potential source of cadavericflora DNA and samples of skin as a controlmaterial to identify the microorganisms ofthe glacier with which the mummy had beenkept in contact through the centuries. Anymolecular analysis of the iceman’s skin,however, should take into account that thismay hardly be representative of the originalsituation. During the 24 hr that followed thetransfer from the glacier to the mortuary atInnsbruck University’s Institute of ForensicMedicine, the body was photographed andfingered by members of the local press. Theday after (September 24, 1991), KonradSpindler identified the accompanying metalaxe as bronze, and the body was thereforeascribed to the early second millenniumB.C. (Bahn and Everett, 1993). Within hours,green bluish spots, perhaps too hastily iden-tified as rapidly growing molds, were no-ticed all over the corpse; therefore, it washanded over to the Institute of Anatomy,whose staff swabbed the entire surface of themummy with a sheet daubed with phenol.The pustules were later recognized as beingof nonbiological origin: they were actuallydeposits of vivianite (Fe3P2O8 1 H2O) pro-duced by contact of the body with iron-richsediments (Tiefenbrunner, 1992; Spindler,1995; De Marinis and Brillante, 1998).

If any trace of the ancient microbial coloni-zation of the skin was left, this has mostprobably been removed. It was thereforefortunate that, during the very first at-tempts to pull the mummy out of the ice(September 20, 1991) with the help of apneumatic chisel (at that time no one sus-pected the antiquity of the find), the leftthigh underwent severe damage with conse-quent release of skin, muscle, and bloodvessel portions into the snow. Several tissuefragments could thus be recovered and iden-tified (Capasso et al., 1995) during the 1992archaeological expedition.

A fragment of untreated skin from the1992 survey, together with one of musculartissue taken directly from the mummy, wereused in the present research to reconstructthe exact composition of the original (pri-mary) cadaveric flora of the iceman by isolat-

ing and identifying short fragments of bacte-rial chromosomal DNA.

MATERIALS AND METHODSThe iceman’s soft-tissue specimens

One sample of muscular tissue and one ofskin were personally taken from the mummyby F.R. at the Institute of Anatomy, Univer-sity of Innsbruck, in 1992, under the supervi-sion of a staff member (Othmar Gaber). Theskin comes from the coccygeal region, andthe muscular tissue from the left thigh(Gaber et al., 1992). Both regions were dam-aged following the first attempt to pull thebody out of the ice.

Another skin sample (specimen no. 183)was recovered during the 1992 archaeologi-cal exploration of the iceman site conductedby an Austro-Italian interdisciplinary re-search team lead by Andreas Lippert fromthe University of Vienna. The ice was meltedoff gradually in the rock gully by the help ofsteam jet blowers. The melting water wasfiltered in order to trace small objects. About400 different fragments of artificial, anatomi-cal, and botanical origin were discovered(Bagolini et al., 1995). Six anatomical speci-mens of interest, including a fingernail(specimen no. 168), two fragments of skin(specimen nos. 183 and 187), and remains ofhair (specimen nos. 158, 189, and 324), wereaccurately identified and described by Ca-passo et al. (1995).

Since 1992, all the samples have beenconstantly kept frozen at 220°C.

Bacterial DNA analysisDNA extraction. The details of the phe-nol-based DNA extraction procedure havebeen reported previously (Ubaldi et al.,1998). The samples were left overnight in anextraction medium composed of sodium do-decyl sulfate (SDS), Na2EDTA, Tris-HCl (pH8.0), and phenol. Each tissue sample (ap-proximately 100 mg) was then milled in amortar with a pestle in the presence of theextraction medium. The homogenate wasthen extracted stepwise using phenol, phe-nol/chloroform, chloroform, and ether, andthe DNA was finally precipitated using etha-nol at 220°C. The DNA was further purifiedby a MerMaid kit (BIO 101, Vista, CA) to

213ICEMAN MUMMIFICATION

eliminate polymerase chain reaction (PCR)inhibitors.

PCR amplification of DNA. Enzymaticamplifications by PCR were performed in 50ml of a reaction mixture containing 10 mMTris-HCl, pH 8.3, 1.5 mM MgCl2, 50 mMKCl, 0.1 mg/ml gelatin, 200 mM (each)dNTPs, 300 ng of each oligonucleotideprimer, 2.5 units Taq polymerase, and 1 µlDNA. The 338f/531r primer-pair (338f, 58-AACTGAGACACGGTCCAGAC-38; 531r, 58-ACGCTTGCACCCTCCGTATT-38) behavesas ‘‘universal’’ primer-pair for bacterial 16sribosomal RNA genes. Following an initialstep at 94°C for 5 min, the thermal cyclerwas set as follows: 94°C for 60 sec (denatur-ation); 56°C for 30 sec (annealing); and 72°Cfor 60 sec (elongation). Forty amplificationcycles were performed. To eliminate smallamounts of E. coli DNA possibly present inthe commercial enzyme preparations, and,in general, any types of contaminant DNA,the Taq polymerase and the reaction compo-nents were pretreated with 3 units of DNaseI (from bovine pancreas) for 30 min at roomtemperature. After incubation, the DNase Iwas inactivated by boiling for 10 min at94°C.

Precautions to avoid contamination. Inthe course of DNAextraction and PCR ampli-fication steps, we routinely employed thestandard precautions for ancient DNA work(e.g., wearing of sterile gloves, pretreatmentof mortars, pestles, and homogenizers withHCl, use of ultraviolet (UV)-irradiated safetycabinets, and dedicated gel trays and tanks).We also implemented the different steps inphysically separated laboratories and used aDNA-free chamber.

Traditional contamination controls arerepresented by mock extractions and blankamplification reactions. We routinely per-form both types of control. In our laborato-ries, however, monitoring of contaminantDNA does not stop at this point, as we cloneamplified DNA and sequence large numbersof bacterial amplicons. In this way we areable to detect low copy number contaminantDNA species which might escape detectionby other means (Ubaldi et al., 1998). Finally,it must be pointed out that the contamina-tion problems are not relevant if the amplifi-

cation target is a very high copy number.This is precisely the case of the low molecu-lar-weight bacterial chromosomal DNA inthe iceman’s bioptic samples.

DNA cloning and sequencing. Amplifi-cation products were cloned using a pMOS-Blue T-vector kit (Amersham, ArlingtonHeights, IL) according to the manufactur-er’s instructions. Plasmid DNA was ex-tracted using an alkaline lysis based methodfollowed by silica gel purification (Plasmix,Talent, Trieste, Italy). Sequencing was per-formed by the use of a Dye Terminator CycleSequencing Kit/AmpliTaq FS (PE AppliedBiosystems, Foster City, CA). Cycle sequenc-ing products were subsequently purified byCentri-sep spin columns (Princeton Separa-tions, Adelphia, NJ) and analyzed using anABI-PRISM 310 Sequencer (Perkin-Elmer,Oak Brook, IL).

Analysis of DNA sequences. Chimericsequences were detected using the CheckChimera program available from the Ribo-somal Database Project (Maidak et al., 1997).Comparisons with reference sequences inthe National Center for Biotechnology Infor-mation (NCBI) data library were performedusing the BLAST (Altschul et al., 1990)search program.

Quantitative PCR test. To quantify thenumber of bacterial chromosomal DNA mol-ecules in the iceman’s tissue (skin andmuscle) extracts, we employed a competitivePCR (cPCR) system. The assay was per-formed by mixing an equal volume (1 µl) ofthe iceman’s DNA with a stated copy num-ber (1,000) of a competitor (a 232-bp-longfragment of Variovorax paradoxus 16s rDNAwhich had previously been PCR-amplifiedusing the 338f/531r primer-pair and clonedinto a pMOSBlue T-vector) and by submit-ting the mixture to enzymatic amplification,using the 338f/531r primer pair (Ubaldi etal., 1998).

RESULTSBacterial DNA in the iceman’s

soft-tissue specimens

DNA preparations obtained from the ice-man’s soft-tissue specimens (treated and

214 F. ROLLO ET AL.

untreated skin, and muscle) were initiallychecked for the presence and copy number ofbacterial DNA, using a competitive PCRsystem designed to amplify a short (196-bp)fragment of the prokaryotic 16s ribosomalRNA gene (16s rDNA). The oligonucleotideprimers (338f/531r) employed bind to a broadrange of bacterial DNAs (‘‘universal’’ prim-ers).

The result is shown in Figure 1. WhenDNA extracts from the iceman’s tissues arePCR-amplified in the presence of a 232-bp-long competitor, virtually no amplificationsignal is produced by the skin sample takendirectly from the phenol-swabbed mummy,thus showing the absence of remnants ofbacteria. Conversely, the signal produced bythe muscle extract is equivalent to approxi-mately 2,000 copies of bacterial DNA; this,in turn, corresponds to about 100,000 copiesof bacterial DNA per milligram of desiccatedmuscle. The extract obtained from the skinsample collected at the site in 1992 (un-treated skin) contains bacterial DNA at aneven higher copy number than that in themuscle, as witnessed by the disappearanceof the competitor (232-bp) band.

Identification of bacterial DNA

To identify the bacteria, the extracts fromuntreated skin and muscle were PCR-ampli-fied with the 338f/531r primers, and theproducts cloned into a plasmid vector asdescribed in Materials and Methods. Subse-quently, we determined the nucleotide se-quence of about 40 amplicons from eachcollection, and the sequences were used toscan the NCBI data library. The results areshown in Tables 1 and 2.

The skin microflora is mainly composed ofmicroorganisms belonging to five genera:Sphingomonas, Afipia, Curtobacterium, Mi-crobacterium, and Agromyces. The membersof the genus Sphingomonas are ubiquitous;they can be found in soil, water, and sedi-ments. Afipia genosp. is also a soil microor-ganism. Curtobacterium luteum is a plantpathogen related to Clavibacter michiganen-sis, and to Corynebacterium aquaticum(Komagata and Suzuki, 1986; Funke et al.,1994). The Coryneform bacteria group in-cludes a vast range of microorganisms whichare animal or plant pathogens or grow sapro-

bically on decaying organic matter. Withinthe iceman’s skin microflora, the genus Mi-crobacterium (Collins et al., 1983) is repre-sented by the species M. lacticum. Thesemicrobacteria are thin, irregular rods ca-pable of growing at temperatures up to 40°Cin raw milk, cheese, and other dairy prod-ucts. Agromyces cerinus and A. ramosum aresoil-dwellers. Other species, such as Lentzeaalbidocapillata and Arthrobacter sp., arerepresented by a single amplicon each. Thefirst is a recently discovered, rare species ofthe order Actinomycetales, isolated fromclinical specimens. The members of the ge-nus Arthrobacter, on the other hand, areinhabitants of the soil (Hagedorn and Holt,1975). Psychrophilic strains of Arthrobacterhave been isolated from glacier sedimentsand Arctic cavern muds (Gounot, 1968).

Contrary to that of the skin, the musclemicroflora is dominated by a single species,

Fig. 1. Competitive PCR test (negative picture of a3% agarose gel) designed to quantify the bacterial DNAfragments (196 bp, lower arrowhead) in preparations ofthe iceman’s phenol-treated and untreated skin, andmuscle. Lane 1, molecular weight standard (50–2,000-bp ladder, Bio-Rad, Richmond, CA); lane 2, phenol-treated skin; lane 3, muscle; lane 4, untreated skin;lane 5, negative control (no DNA). A competitor (232 bp,upper arrowhead) at a multiplicity of 1,000 copies wasadded to each reaction mixture.

215ICEMAN MUMMIFICATION

Clostridium algidicarnis. This Gram-posi-tive, anaerobic, spore-forming rod was iso-lated for the first time from chilled, vacuum-packed raw beef (Cato et al., 1986; Lawsonet al., 1994). In the laboratory, this microor-ganism grows down to at least 4°C. Frag-ments of 16s rDNA of C. algidicarnis havealso been obtained, as the prevailing bacte-rial DNA, following PCR amplification, us-ing the universal 338f/531r primer pair, ofthe DNA isolated from the colon of an An-dean mummy (Ubaldi et al., 1998). To ourknowledge, these are the first demonstra-tions of the presence of this species of Clos-tridium in corpses. The composition of thecadaveric flora is still largerly unknown. Inthe initial stages of decay of a corpse, thebacteria are commonly assumed to comefrom the intestines.

DISCUSSION

The iceman’s corpse shows no major signsof putrefaction (e.g., eyeballs and dermis arepreserved), though his epidermis, hair, and

nails are absent (Spindler, 1995; De Marinisand Brillante, 1998). Molecular analysesperformed in the past few years have shownthat the soft tissues still contain traces ofthe original human genetic material (Handtet al., 1994). Actually, the quantitation(Handt et al., 1996) of mitochondrial DNA(mtDNA) fragments showed that 1 mg drysoft tissue contained approximately 2,000fragments of 103 bp in length; the same testindicated that longer (189-bp) fragmentswere also present, though at a much lowercopy number (10 copies/mg). Lipid analysisof skin samples recovered during the 1992archaeological expedition, performed by Be-reuter et al. (1997), on the other hand,demonstrated a declining gradient of triacyl-glycerols and unsaturated fatty acids fromthe outer to the inner side. This is in agree-ment with adipocere formation, which be-gins in the subcutaneous fat deposits andproceeds upwards to the neighboring tissues(Mayer et al., 1996). Transformation of hu-man fat by microbes into adipocere usually

TABLE 1. Identification of 16s rDNA amplicons from iceman’s untreated skin1

Amplicon BLAST identification NCBI accession no. Similarity (%)

183S-11 Sphingomonas mali Y09638 100183S-20 Sphingomonas mali Y09638 100183S-21 Sphingomonas mali Y09638 100183S-25 Sphingomonas mali Y09638 100183S-28 Sphingomonas mali Y09638 100183S-29 Sphingomonas mali Y09638 100183S-37 Sphingomonas mali Y09638 100183S-44 Sphingomonas mali Y09638 100183S-47 Sphingomonas mali Y09638 100183S-27 Sphingomonas mali Y09638 100183S-37 Afipia genosp. U87782 100183S-39 Afipia genosp. U87782 100183S-48 Afipia genosp. U87782 98183S-61 Afipia genosp. U87782 100183S-65 Afipia genosp. U87782 100183S-30 Curtobacterium luteum X77437 99183S-35 Curtobacterium luteum X77437 99183S-38 Curtobacterium luteum X77437 98183S-49 Curtobacterium luteum X77437 100183S-3 Microbacterium lacticum D21343 98183S-35 Microbacterium lacticum D21343 98183S-52 Microbacterium lacticum D21343 98183S-71 Microbacterium lacticum D21343 98183S-31 Agromyces cerinus X72723 92183S-40 Agromyces cerinus X72723 96183S-42 Agromyces cerinus X72723 96183S-43 Agromyces cerinus X72723 96183S-24 Agromyces ramosum X77447 96183S-23 Rathaybacter iranicus U96184 96183S-32 Corynebacterium aquaticum X77450 99183S-45 Lentzea albidocapillata X84321 98183S-58 Arthrobacter sp. MB90 99

1 Not included are the chimeric (Wang and Wang, 1996) amplicons 183S-46, 183S-60, and 183S-66.

216 F. ROLLO ET AL.

takes several months and requires humidconditions as well as temperatures abovefreezing.

It seems logical to assume that the mum-mification process responsible for the preser-vation of the human DNA, irrespective ofthe way this process took place, allowedpreservation of the DNA of the bacteria(cadaveric microflora) that grew within thecorpse for a certain time span followingdeath. In view of this, we employed a PCRprimer pair designed to bind to fragments ofbacterial chromosomal 16s rDNA, compa-rable to the longest mtDNA fragments de-tected by Handt et al. (1996), i.e., about 196bp. Of course, as the body lay in close contactwith ice and sediments for most of its tapho-nomic history, we can conceive that, throughthe millennia, a certain number of microor-ganisms of the environment seeped throughthe mummified tissues, with the result ofmasking and erasing any traces of the primi-tive cadaveric flora. Operatively, we can put

forward the two following working hypoth-eses: 1) the flow of microorganisms from theouter environment has been massive; 2) ithas been negligible compared to the extentof primary proliferation.

If the first hypothesis were true, we shouldfind a substantial uniformity in the bacte-rial DNA composition of skin (which hasbeen constantly in contact with the outerenvironment) and muscle. Conversely, if thesecond were true, we would expect to find asignificant difference between the bacterialDNA composition of muscle and skin.

The result of the sequence analysis ofbacterial amplicons shown in Tables 1 and 2clearly speaks in favor of the second hypoth-esis. This conclusion, i.e., negligible or verylow secondary penetration of environmentalmicroorganisms, is further strengthened bythe observation that small fragments ofgrass clothing recovered from the glaciercontain only very tiny amounts of bacterialDNA (Cano et al., 1999). A further element of

TABLE 2. Identification of 16s rDNA amplicons from iceman’s muscle1

Amplicon BLAST identification NCBI accession no. Similarity (%)

FS-6 Clostridium algidicarnis X77676 100FS-7 Clostridium algidicarnis X77676 100FS-9 Clostridium algidicarnis X77676 96FS-13 Clostridium algidicarnis X77676 100FS-14 Clostridium algidicarnis X77676 100FS-15 Clostridium algidicarnis X77676 100FS-17 Clostridium algidicarnis X77676 100FS-18 Clostridium algidicarnis X77676 100FS-22 Clostridium algidicarnis X77676 100FS-25 Clostridium algidicarnis X77676 99FS-40 Clostridium algidicarnis X77676 95FS-42 Clostridium algidicarnis X77676 100FS-46 Clostridium algidicarnis X77676 100FS-47 Clostridium algidicarnis X77676 100FS-49 Clostridium algidicarnis X77676 100FS-51 Clostridium algidicarnis X77676 100FS-55 Clostridium algidicarnis X77676 99FS-57 Clostridium algidicarnis X77676 100FS-58 Clostridium algidicarnis X77676 100FS-62 Clostridium algidicarnis X77676 100FS-64 Clostridium algidicarnis X77676 100FS-65 Clostridium algidicarnis X77676 99FS-76 Clostridium algidicarnis X77676 100FS-77 Clostridium algidicarnis X77676 100FS-80 Clostridium algidicarnis X77676 100FS-81 Clostridium algidicarnis X77676 100FS-20 Clostridium paraputrificum X73445 96FS-50 Clostridium paraputrificum X73445 95FS-53 Clostridium paraputrificum X73445 96FS-71 Clostridium paraputrificum X73445 96FS-75 Clostridium paraputrificum X73445 95FS-78 Clostridium paraputrificum X73445 96FS-82 Clostridium paraputrificum X73445 96FS-66 Clostridium sp. X95274 99

1 Not included are the chimeric (Wang and Wang, 1996) amplicons FS-2, FS-3, FS-21, FS-29, FS-35, FS-38, FS-63, FS-65, and FS-84.

217ICEMAN MUMMIFICATION

interest is the observation that, actually,skin and muscle share no common species.The composition and eco-physiological char-acteristics of the skin bacterial flora suggesta history of close contact with aerated sedi-ments and plant debris at relatively mildtemperatures. The muscle, conversely, con-tains the DNA of the anaerobic psychrotrophClostridium algidicarnis at a very high copynumber, thus demonstrating that this micro-organism once found the appropriate condi-tions for its growth.

Regarding the above two hypotheses onthe mummification process, the prolifera-tion of C. algidicarnis within the iceman’stissues does not seem compatible with thehypothesis of rapid desiccation of the bodyby effect of a warm wind shortly after death.On the other hand, one is struck by thesimilarity of the bacterial DNA compositionof the iceman’s muscle with that of the softtissue of a natural mummy from Cuzco, i.e.,54% of the PCR amplicons identified as C.algidicarnis (Ubaldi et al., 1998). The Cuzcomummy, as many other Andean mummies,is a corpse that desiccated by effect of the drycold air of the heights.

The presence of adipocere, however, indi-cates that the mummification process of theiceman cannot be so simple. According toBereuter et al. (1997), ‘‘Shortly after hisdeath at the site of discovery in that prehis-toric early autumn, iceman’s corpse wascovered by snow and ice. At some later stage,perhaps during Roman times, his shelteredcorpse thawed and became immersed in thewater collecting within the depression whereit lay. During this brief warm period, most ofthe physical damage incurred by the corpse,as well as the formation of adipocere andloss of epidermis took place. In the light ofits excellent preservation, it would seemlikely that desiccation of the corpse occurredshortly after its immersion in water.’’ Ourresults fit rather well with this picture. Inparticular, the composition of the skin micro-flora, as inferred from the analysis of bacte-rial DNA, is best explained if we accept thehypothesis of a prolonged permanence of thebody in a humid site at a relatively mildtemperature. The absence of identifiablerepresentatives of the normal skin micro-flora is probably the result of the loss of

epidermis. The growth of C. algidicarnismay have taken place before the warmperiod, shortly after it, or in both periods.Incidentally, the spoiling activity of thismicroorganism may be responsible for themarked degradation and loss of human DNAobserved by Handt et al. (1994).

ACKNOWLEDGMENTS

The authors thank Werner Platzer andOthmar Gaber, Institute ofAnatomy, Univer-sity of Innsbruck (Innsbruck, Austria) andLuigi Capasso, Servizio Tecnico perl’Antropologia e la Paleopatologia, Ministerodei Beni Culturali (Chieti, Italy) for provid-ing the iceman’s bioptic samples. This workwas cofunded by the Italian Ministry ofUniversity and Scientific Research throughthe project ‘‘Disease, Health, and Socioeco-nomic Conditions in Ancient Egypt. A Mul-tidisciplinary Project,’’ and by the ItalianNational Research Council ‘‘Beni Culturali’’program, grant 97.00611.PF36.

LITERATURE CITED

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ.1990. Basic local alignment search tools. J Mol Biol215:403–410.

Bagolini B, Dal Rı L, Lippert A, Nothdurfter H. 1995.Der Mann im Eis: die Fundbergung 1992 am Tisen-joch, Gem. Schnals, Sudtirol. In: Spindler K, Rast-bichler-Zissernig E, Wilfing H, zur Nedden D, Noth-durfter H, editors. Der Mann im Eis. Wien, New York:Springer-Verlag. p 3–22.

Bahn PG. 1996. Tombs, graves and mummies. NewYork: Barnes & Noble. 213 p.

Bahn PG, Everett K. 1993. Iceman in the cold light ofday. Nature 362:11–12.

Bereuter TL, Mikenda W, Reiter C. 1997. Iceman’smummification—implications from infrared spectro-scopical and histological studies. Chem Eur J 7:1032–1038.

Cano RJ, Tiefenbrunner F, Ubaldi M, del Cueto C,Luciani S, Cox T, Orkand P, Kunzel KH, Rollo F. 1999.Sequence analysis of bacterial DNA in the colon of theneolithic glacier mummy from the Alps. Manuscript inpreparation.

Capasso L, Capelli A, Dal Rı L, Frati L, Mariani-Costantini R, Nothdurfter H. 1995. I resti umanidell’Hauslabjoch raccolti durante la campagna di scavodel 1992: rapporto preliminare. In: Spindler K, Rast-bichler-Zissernig E, Wilfing H, zur Nedden D, Noth-durfter H, editors. Der Mann im Eis. Wien, New York:Springer-Verlag. p 241–245.

Cato EP, Lance George W, Finegold SM. 1986. GenusClostridium. In: Sneath PHA, Mair NS, Sharpe ME,Holt JG, editors. Bergey’s manual of systematic bacte-riology, volume II. Baltimore: Williams and Wilkins. p1157–1160.

Collins MD, Jones D, Kroppenstedt RM. 1983. GenusMicrobacterium. Syst Appl Microbiol 4:65–78.

De Marinis RC, Brillante G. 1998. Otzi, l’uomo venutodal ghiaccio. Venice: Marsilio Editore. 189 p.

218 F. ROLLO ET AL.

Funke G, von Graevenitz A, Weiss N. 1994. Primaryidentification of Aerobacterium spp. isolated fromclinical specimens as ‘‘Corynebacterium aquaticum.’’ JClin Microbiol 11:2686–2691.

Gaber O, Kunzel KH, Maurer H, Platzer W. 1992.Konservierung und Lagerung der Gletschermummie.In: Hopfel F, Platzer W, Spindler K, editors. Der Mannim Eis. Innsbruck: Veroffentlichungen der UniversitatInnsbruck. 187 p 92–99.

GounotAM. 1968. Etude microbiologique de deux grottesantiques. C R Acad Sci Paris [D] 266:1613–1620.

Hagedorn C, Holt JG. 1975. Ecology of soil arthrobactersin Clarion-Webster toposequences of Iowa.Appl Micro-biol 29:211–218.

Handt O, Richards M, Trommsdorff M, Kilger C, Sima-nainen J, Georgiev O, Bauer K, Stone A, Hedges R,Schaffner W, Utermann G, Sykes B, Paabo S. 1994.Molecular genetic analyses of the Tyrolean Ice Man.Science 264:1775–1778.

Handt O, Krings M, Ward RH, Paabo S. 1996. Theretrieval of ancient human DNA sequences. Am JHum Genet 59:368–376.

Komagata K, Suzuki K. 1986. Genus Curtobacterium.In: Holt JG, editor. Bergey’s manual of systematicbacteriology. Baltimore: Williams and Wilkins. p 1313–1317.

Lawson P, Dainty RH, Kristiansen N, Berg J, CollinsMD. 1994. Characterization of a psychrotrophic Clos-tridium causing spoilage in vacuum-packed cookedpork: description of Clostridium algidicarnis sp. nov.Lett Appl Microbiol 19:153–157.

Maidak BL, Olsen GJ, Larsen N, Overbeek R, Mc-Caughey M, Woese CR. 1997. The RDP RibosomalDatabase Project. Nucleic Acids Res 25:109–111.

Mayer BX, Reiter C, Bereuter TL. 1996. Investigation ofthe triacylglycerol composition of iceman’s mummi-fied tissue by high-temperature gas chromatography.J Chromatogr 692:1–6.

Meyer W. 1992. Der Soldner wom Theodulpaß undandere Gletscherfunde aus der Schweiz. In: Hopfel F,Platzer W, Spindler K, editors. Der Mann im Eis.Innsbruck: Veroffentlichungen der Universitat Inns-bruck. 187 p 321–333.

Prinoth-Fornwagner R, Niklaus TR. 1995. Der Mann imEis. Resultate der Radiokarbon-Datierung. In: Spin-dler K, Rastbichler-Zissernig E, Wilfing H, zur Ned-den D, Nothdurfter H, editors. Der Mann im Eis.Wien, New York: Springer-Verlag. p 77–89.

Rollo F. 1998. Ancient DNA: problems and perspectivesfor molecular microbial palaeoecology. In: CarvalhoGR, editor. Advances in molecular ecology (NATOsciences series). Amsterdam: IOS Press. p 133–149.

Rollo F, Marota I. 1999. How microbial ancient DNA,found in association with human remains, can beinterpreted. Philos Trans R Soc Lond [Biol] 354:111–119.

Rollo F, Asci W, Marota I, Sassaroli S. 1995a. DNAanalysis of grass remains found at the iceman’s ar-chaeological site. In: Spindler K, Rastbichler-Zis-sernig E, Wilfing H, zur Nedden D, Nothdurfter H,editors. Der Mann im Eis. Wien, New York: Springer-Verlag. p 91–105.

Rollo F, Antonini S, Ubaldi M, Asci W. 1995b. The‘‘Neolithic’’ microbial flora of the iceman’s grass: mor-phological description and DNA analysis. In: SpindlerK, Rastbichler-Zissernig E, Wilfing H, zur Nedden D,Nothdurfter H, editors. Der Mann im Eis. Wien, NewYork: Springer-Verlag. p. 107–114.

Rollo F, Sassaroli S, Ubaldi M. 1995c. Molecular phylog-eny of the fungi of the Iceman’s grass clothing. CurrGenet 28:289–297.

Spindler K. 1995. The man in the ice. London: Phoenix.305 p.

Tiefenbrunner F. 1992. Bakterien und Pilze, ein Prob-lem fur unseren altesten Tiroler? In: Hopfel F, PlatzerW, Spindler K, editors. Der Mann im Eis. Innsbruck:Veroffentlichungen der Universitat Innsbruck. 187 p100–107.

Ubaldi M, Luciani S, Marota I, Fornaciari G, Cano RJ,Rollo F. 1998. Sequence analysis of bacterial DNA inthe colon of an Andean mummy. Am J Phys Anthropol107:285–295.

Wang GCY, Wang Y. 1996. The frequency of chimericmolecules as a consequence of PCR co-amplification of16s rRNAgenes from different bacterial species. Micro-biology 142:1107–1114.

219ICEMAN MUMMIFICATION