Copyright 0 1992 by the Genetics Society of America

Native American MitochondrialDNA Analysis Indicates

That the Amerind and the Nadene Populations Were Founded byTwo
Independent Migrations

Antonio Torroni,* TheodoreG. Schurr,* Chi-Chuan Yang,* EmokeJ. E. Szathmary?

Robert C. Williams? Moses S. Schanfield? Gary A. Troup,** WilliamC. Knowler,w
Dale N. Lawrence,*$ KennethM. Weisss and Douglas C. Wallace*
*Centerfor Genetics and Molecular Medicine and Departments of Biochemistry and Anthropology, Emory University School of
Medicine, Atlanta, Georgia 30322, +Officeof the Dean, Faculty of Social Science, University of Western Ontario, London, Ontario,
Canada N6A 5C2,*Department of Anthropology, Arizona State University, Tempe, Arizona 85281, @AnalyticalGenetic Testing
Center, Inc., Denver, Colorado80231, **Department of Pathology, University of New Mexico, Albuquerque, New Mexico 87131,
-Diabetes and Arthritis Epidemiology Section, National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK),
Phoenix, Arizona 85014, "Division of Host Factors, Center for Infectious Diseases, Centers for Disease Control,
Atlanta, Georgia 30333, and BDepartment of Anthropology and Graduate Program in Genetics, Pennsylvania State University,
University Park, Pennsylvania 16802
Manuscript received April 19, 1991
Accepted for publication September 13, 1991

ABSTRACT
Mitochondrial DNAs (mtDNAs) from 167 American Indians including 87 Amerind-speakers
(Amerinds) and 80 Nadene-speakers (Nadene) were surveyed for sequence variation by detailed
restriction analysis. All Native American mtDNAs clustered into oneof four distinct lineages, defined
by the restriction site variants: HincII site loss at np 13,259, AluI site loss at np 5,176, 9-base pair (9-
bp) COII-tRNALYsintergenic deletion and HaeIII site gain at np 663. The HincII np 13,259 and AluI
np 5,176 lineages were observed exclusively in Amerinds and were shared by all such tribal groups
analyzed, thus demonstrating that North, Central and South American Amerinds originated from a
common ancestral genetic stock. The 9-bp deletion and HaeIII np 663 lineages were found in both
the Amerinds and Nadene but the Nadene HaeIII np 663 lineage had a unique sublineage defined
by an RsaI site loss at np 16,329. The amount of sequence variation accumulated in the Amerind
HincII np 13,259 and AluI np 5,176 lineages and that in the Amerind portion of the HaeIII np 663
lineage all gave divergence times in the order of 20,000 years before present. The divergence time
for the Nadene portion of the HaeIII np 663 lineage was about 6,000-10,000 years. Hence, the
ancestral Nadene migrated from Asia independently and considerably more recently than the
progenitors of the Amerinds. The divergence times of both the Amerind and Nadene branches of
the COII-tRNALY"deletion lineage were intermediate between the Amerind and Nadene specific
lineages, raising the possibility of a third source of mtDNA in American Indians.

T HEconsiderablecultural
complexity (SPENCER
diversity, linguistic
et al. 1977) andbiological
variation (NEEL1978; NEEL and THOMPSON 1978;
SZATHMARY 1984) of Native Americans has been the
subject of much speculation. T h e observed diversity
has at times been attributed to multiple migrations of
Asiatic peoples to the Americas, a few migrations of
ethnically distinct peoples, or in situ differentiation of
ancestral Native Americans occurring after theinitial
colonization of the New World(LAUGHLIN 1988;
SZATHMARY 1991). A recenthypothesis proposes that
three waves of migrationpopulated the Americas
(WILLIAMS et al. 1985; GREENBERG, TURNER and ZE-
GURA 1986), corresponding to the tripartite division
of Native American languages [Amerind, Nadeneand
Eskaleut] (GREENBERG 1987). However, the common
origin of the numerous Amerind languages and the
time depthrequired t o develop this plurality of
(;enetics 130: 153-162 uanuary, 1992)
tongues is disputed (DIAMOND 1990; LEWIN 1990).
Estimates of the time of arrival of the first migrants
have also varied widely (MARSHALL 1990; MORELL
1990). Dating of skeletal remains and Clovis lithic
artifacts yield values of 13-14,000 years before pres-
ent (YBP) (TAYLOR et al. 1985; NELSONet al. 1986)
while archaeological studies have reported North and
South American sites which date to more than 33,000
YBP (GUIDONand DELIBRIAS1986;DILLEHAYand
COLLINS1988), with the best North American sites
startingat 20,000-25,000 YBP(ADOVASIO et al.
1983).
To further investigate the populating of the Amer-
icas, we have examined the mtDNA variation of Na-
tive American populations. Becauseof its maternal
inheritance (GILESet al. 1980), the mtDNA accumu-
lates sequential mutations along radiating female lin-
eages. The high mutation rate of the mtDNA
154 A. Torroni et al.

(BROWN, GEORGEand WILSON1979; WALLACE et al. cell enzyme genes(SZATHMARY, FERRELL and GERSHOWITZ
1987)permits discrimination between even closely 1983).
TLe Navajo are southern Athapaskans who are closely
related populations. In previous studies of Amerind related genetically (SZATHMARY 1983) and linguistically to
mtDNAs, we reported that North, Central and South the northern Athapaskan Indians living in Alaska and Can-
American populations exhibited high frequencies of ada. Their ancestors were hunting-gathering people from
the same set of rare Asian variants which suggested northwestern Canada who migrated to the southwest US
that they derived from a common ancestral populationafter 1000 AD (HASKELL 1987). Having increased to the
present 150,000 from a total of no more than 15,000 people
(WALLACE, GARRISON and KNOWLER1985; SCHURR et in 1868 (SPENCER et al. 1977), theNavajo are now dispersed
al. 1990). Because these mtDNAs represented only a throughout the 30,000 square mile reservation in relatively
small subset of all haplotypes observed in Asia (JOHN- isolated family units (WILLIAMS et al. 1981; TROUP et al.
SON et al. 1983; HORAI, GOJOBORI and MATSUNAGA 1982). The bloodsanalyzed in this study represent two
1984; HORAI and MATSUNAGA1986;HARTHARA, segments of the population, a northern one centeredat the
Public Health Service Indian Hospital inKeams Canyon,
HIRAI andOMOTO1986; CANN, STONEKING and WIL- Arizona, inwhichCaucasian admixture was estimated at
SON 1987; HARIHARA et al.. 1988; STONEKING et al. 1.1% (WILLIAMS et al. 1981; 1985), and a southern one
1990; BALLINGER et al. 1992), we hypothesized that based in Albuquerque, New Mexico, where individuals from
all Amerinds derivedfroman Asiatic population a wider region of the reservation were sampled and Cauca-
which experienced a dramatic reduction of its effec- sian admixture was estimated at 5.0% (TROUP et al. 1982).
The Pima of southwestern Arizona are descendants of
tive size and genetic variation before or during its the Hohokam who migrated north from northwestern Mex-
migration to the Americas. We now report a quanti- ico around 300 B.C., and belong to the Uto-Aztecan lan-
tative analysis of mtDNA sequencevariation for some guage family which includes the Papago, Hopi and several
Nadene and Amerind tribeswhich indicates that these northern Mexican tribes (MATSONet al. 1968; HAURY 1976).
two groups derived from distinct populations which Upon entering and settling the area, they are believed not
to have intermingled with the Puebloan peoples who already
originated at very different times. occupied the region, but may have mixed considerablywith
other Southwest Indians (MATSON et al. 1968). In 1977, the
MATERIALS AND METHODS Pima numbered some 8,000 people (SPENCER et al. 1977).
Gm allotype studies indicated that Caucasian admixture in
Samples: Blood cells in the form of lymphoblasts,platelets the Pima was 1.0% (WILLIAMS et al. 1985, 1986). Bloods
or buffy coats were obtained from 167 subjects including were collected atthe Gila River Indian Community in
80 Nadene and 87 Amerinds. * The Nadene were comprised Sacaton, Arizona, by the NIDDK as part of a diabetes study
of 30 Dogrib (NW Canada), 48 Navajo (Arizona and New (KNOWLERet al. 1978).
Mexico) and 2 Tlingits (Alaska), and their bloods represent The lowland Maya occupy parts of Mexico, Guatemala
a random sampling of individuals within each respective and Honduras, and are known to have inhabited the area
tribal group. The Amerinds were comprised of 30 Pima continuously for at least 5,000 years (MACNEISH 1983).
(Arizona), 1 Hopi (Arizona), 1 Pomo (California), 27 Maya Bloodsamples were drawn from individualsliving in an
(Mexico) and 28 Ticuna (Brazil). The Pima and Ticuna isolated villagein the central Yucatan Peninsula. Becauseof
individuals from whom blood was taken were known to be the smallsize and remoteness of the village,most of its
unrelated for at least three generations, whereas the Maya inhabitants were related at least at the second cousin level,
represent a random sample of that tribal group. however, only one member per family cluster was analyzed
The Dogrib, speakers of anorthern Athapaskan lan- in this study. These people speak Yucatec, one ofmany
guage, are the largest Indian group in the Canadian North- Maya languages belonging to the Mexican Penutian
west Territories. Their traditional subsistence economywas subgroup (GREENBERG 1987). Ananalysis of blood types
based on hunting and fishing in the boreal forest and adja- and serum proteins in nearby villages revealed approxi-
cent barrenlands. Aboriginally they were subdivided into mately 10% European admixture inthis population (our
several regional bands, ie., socioterritorial groupsthat unpublished data).
moved withinparticular regions for much ofthe year (HELM The Ticuna are a linguisticallydistinct and geographically
1981). The bloods used in this study were obtained from isolated tribe living inthe Amazonian rain forest of western
adult members of the Rae Band (SZATHMARY, RITENBAUGH Brazil (MESTRINER, SIMOESand SALZANO 1980; NEEL et al.
and GOODBY 1987). In 1979 about 1600people (75% of all 1980). A study of blood group markers indicated only 2.2%
Dogrib) were members of this band. Although traditionally non-native admixture in this tribe (NEELet al. 1980). Since
nomadic, by 1960 themajority of the Rae Band had settled 1942 the Brazilian Ticuna population has increased from
into permanent communities located within regional band 2,000 to 1,000
1 persons (SALZANO, CALLEGARI~ACQUES and
areas. Consanguineal and affinal tiescontinue to link people NEEL 1980), and this recent rapid population growth has
in different settlements. Accordingly the Dogrib form a been accompanied by migrations from jungle to river settle-
recognizable genetic unit among other Athapaskan-speak- ments (LAWRENCE, BODMER and BODMER1980; NEEL et al.
ers, asshown by genetic distance analyses (SZATHMARY 1980;SALZANO, CALLEGARI-JACQUESand NEEL 1980).
1983). The maximum European admixture in the group Blood samples were taken from individuals living in three
was 8.7%, based on blood group, serum protein and red villages located along the Rio Solimoes(LAWRENCE, BODMER
and BODMER 1980).
’ The Native American tribal groups termed “Amerinds”are the descend- For convenience, in someof our analyses we have
ants of the original migrants (ie., Paleoindianderived),not the migrant grouped populations linguistically. Thus, the Amerinds in-
group itself; this term is borrowed from the linguistic classificatory term clude the Pima,Maya, Ticuna, Pomo and Hopi, and the
“Amerind”(GREENBERG 1987). to permit us to group the non-Nadenetribes.
By using “Amerind we do not imply any agreement or disagreement with Nadene include the Dogrib, Tlingit and Navajo.
linguistic issues surrounding the term. Methods: mtDNAs were extracted from blood cells using
 Mitochondrial DNA Analysis 155

procedures in SCHURR et al. 1990 or a modification of those. 000 ooo-o-

In the modified method, buffy coats or platelet pellets were 000 ooo-o-
incubated in a 50O-pl digestion solution containing 100 mM
000 00000000
NaCI, 10 mM Tris-CI, pH 7.4, 5 mM NaZEDTA, pH 8.0,
0.5% sodium dodecyl sulfate, and 500 pg/ml proteinase K, 0 0 0 oooocvmc
at 55" for 10-1 6 hr, mixed with 250 pl cold 5 M KOAc 0 0 0 oooo-r*
with vigorous shaking for 5 min, then incubated at 0" for 0 0 0 0000--

10 min, and thelysed celldebris pelleted through microcen- 0 0 0 oo-o*m

trifugation. The supernatant was then twice extracted with
phenol-chloroform, and genomic DNA within the aqueous 0 0 0 o o o m o m
phase ethanol precipitated. 0 0 0 oooomm
All mtDNAs were polymerase chain reaction (PCR) am- 0 0 0 o o o o m m
plified (SAIKIet al. 1985)in nine overlapping segments which
0 0 0 00000000
encompassed the entiremtDNA genome (APPENDIX A). Each
0 0 0 o-ooo-
PCR segment was independently digested with 14 restric-
tion endonucleases ( A M , AvaII, DdeI, HaeIII, HhaI, HinfI, 0 0 0 o-ooo-
HpaI, HpaII, MboI, RsaI, TaqI, BamHI, HaeII, HincII), and 0 0 0 00000000
c;
the resulting fragments resolved by electrophoresis in 1 .O- 0 0 0 0 0 0 0 - -

+
2.5% NuSieve 1.0% SeaKem agarose (FMC BioProducts) 0 0 0 omooom
gels and detected by ethidium bromide fluorescence. The 0 0 0 o-ooo-
fragments observed for each PCR segment were restriction 0 0 0 o-ooo-
mapped by the sequence comparison method (JOHNSON et 0 0 0 o-ooo-
al. 1983; CANN,BROWN and WIJSON1984), the composite
0 0 0 ooo-o-
restriction maps of these sets of segments representing in-
0 0 0 ooo-o-
dividual haplotypes is given inAPPENDIX B. The distribution
and frequency of haplotypes among tribal groups is shown o o m ooooom
in Table 1.
0 0 0 ooo-o-
Sequence divergences: Mean intra- and intergroup se-
quence divergences were estimated with NEI and TAJIMA'S 0 0 0 - 0 0 0 0 -
(1983) maximum likelihood procedure using the computer 0 0 0 0 0 0 0 0 m
program DREST (graciously provided by L. JIN). This pro-
0 0 0 o-ooo-
cedure considers the ratioof shared sites to the total number
of sites between two haplotypes, and the mean length of the 0 0 0 o-ooo-
restriction enzyme recognition sequences to calculate an 0 0 0 omooom
initial estimate of x (the probability that the two mtDNAs 0 0 0 o-ooo-
have different nucleotides at a given nucleotide position). 0 0 0 ooo-o-
Using this initialestimate, A is solved iteratively using equa- 0 0 0 o-ooo-m
tion 28 and the sequence divergence (6) is estimated by 0 0 0 o-ooo-
equation 21 (NEI and TAJIMA 1983). 0 0 - 0 0 0 0 0 -
Sequence divergence values within and among the tribal 0 0 - 0 0 0 0 0 -
groups are presented in Table 2. Divergence values shown
0 0 - ooooom
in standard type represent interpopulational comparisons,
0 0 - 0 0 0 0 0 -
those underlined represent intrapopulational comparisons,
and those inbold type represent interpopulation diver- o o m ooooom

gences corrected for intrapopulational variation, 6 = 6, - oo= o m o m o -

mI
0.5(6, + 6J, where 6, is the mean pairwise divergence be-
tween individuals within a single population ( X ) , 6, is the
corresponding value for a second population (Y),and 6, is 0 0 0 ooo-o-
the mean pairwise divergence between individuals belong- 0 0 0 ooo-o-
ing to the two different populations ( X and Y) (NEI and 0 0 0 ooo-o-
TAJIMA 1983). The Tlingit, Hopi and Pomo samples were
too small in number to be used as separate units for diver-
coo* -
o o o m m -

0 0 0 0 0 0 0 - -
gence calculations among tribal groups, but were included
in estimates of overall Nadene and Amerind divergences. 0 0 0 o o o o m m

Phylogeneticanalysis: The evolutionary relationships - 0 0 0 0 0 0 0 - *

among the 50 Native American haplotypes were inferred

using parsimony analysis (PAUP 3.0m; SWOFFORD1989).
p"m - 0 0 0 0 0 -
m
0 0 0 o o o m o m
All but two haplotypes(28 and29) segregated into fourwell
0 0 0 o o o m o m
defined clusters, A, B, C andD, each of whichwas delineated
0 0 0 o m o o o m
by a distinctive genetic marker (Table 1; Figure 1). The
dendrogram presented in Figure 1 is rootedfrom HYP-
ANC, "hypothetical ancestor a"(CANN, STONEKING and
2" - o o o m o m
m

WILSON1987), and is 77 mutational steps in length with a

consistency index of 0.833. When mtDNA haplotypes from
modern Africans (our unpublished data) were substituted
from HYPANC and used as outgroup, the branching order
did not change. This dendrogram represents a tree gener-
ated by the subtree pruning and regrafting (SPR) branch
156 A. Torroni et al.
TABLE 2
mtDNA divergences of Amerinds and Nadene

Population Dogrib Navajo Pima Amerinds

Maya Nadene Ticuna

Dogrib 1.6 -C 1.5 5.2 f 3.5 13.3 -t 6.5 8.0 f 4.7 17.4 f 7.4
Navajo 1.3 f 0.9 6.2 f 3.8 13.2 -t 6.3 9.0 f 5.0 17.9 2 7.3

Pima 6.0 f3.6

2.8 f 1.4 13.0 -t 6.0 14.9 f 6.8 16.8 f 7.3
Maya 2.0 k 1.4 0.7 f 3.2
0.3 k 1.3 10.4 f 5.1 17.9 f 7.5
Ticuna 9.7 25.8
3.6 f3.4
2.3 2 1.2 5.8 f 1.9 13.8 -t 6.2

Nadene 5.0 f 3.3 13.3 f 6.3

Amerinds 3.3 +: 1.3 15.1 f 6.7
Intrapopulation divergences are along the diagonal (underlined), interpopulation divergence above the diagonal in standard type, and
interpopulation divergences corrected for intrapopulation variation (NEI and TAJIMA 1983) below the diagonal in bold type. Divergence
values and standard errors are presented X lo4. The Nadene include 2 Tlingit, and theAmerinds include 1 Hopi and 1 Pomo.

swapping method. Although no shorter trees were found,

they could exist and the numberof trees equal in length to
Figure 1 is probably large.
Bootstrap analysis of the branching order of the dendro-
gram was also performed. However, most American Native
haplotypes are separated by a few site differences. Hence,
the resampling of haplotypes for tree-building did not create
enough definition in the branching order ofall trees to
permit a statistically significant test of their relationships
and the analysis could not be completed unless the MUL-
PARS (multiple parsimony) option was turned off. This
meant that branch swapping was carried out on only one
randomly chosen tree of minimum length at each resam-
pling of the data. Nevertheless, given this reduced level of
robustness of testing, 100 bootstrap replications produced
a consensus tree nearly identical to Figure 1 with a consist-
ency index of 0.844 and a length of 77 mutational steps.
The major internal nodes distinguishing each cluster were
present in 63-78% of the treesgenerated.Hence, even
- IO M though no statistically significant levelof association of
31 haplotypes could be obtained, the bootstrap analysis sup-
033
0% ported the segregation of mtDNAs into the observed clus-
0 35 ters.
-C 0% Intragroup sequence divergence for each cluster were
-
04.3
032 73.1 a
. calculated according toNEIand TAJIMA (1983). Divergence
I 0 39 times for each group were estimated using a mtDNA evo-
om lutionary rate of 2.0-4.0% per million years (MYR), a range
I
J 042
0 41
which encompasses most estimates of the average mtDNA
0 45 evolution rate2 (BROWN1980; MIYATAet al. 1982; STONE-
D ’
KING, BHATIAand WILSON 1986; CANN, STONEKINC and
044 ,046 0 47 WILSON1987; WALLACE et al. 1987; STONEKINC et al. 1990)
048 (Table 3).
I 1.49
OM
HVPANC
RESULTS
AMERiND 0 NADENE 8 SHAREDBETWEENAMERINDSANDNADENE
“HYPANC HYPOTHETICAL ANCESTOR Restriction analysis: A total of 67 polymorphic
A = +Haell1 np 663 B = 9-BPDeletion C = -H/ncilnp13259 D = - A M np 5176
Haplotypes 5 and 6 have theRsal np 16239 siteloss characteristicof the Nadene
sites and the 9-base pair (9-bp) COII/tRNALYs inter-
genic deletion (CANNand WILSON1983; HORAIand
FIGURE1 .-A phylogenetic tree of 50 Native American mtDNA MATSUNACA1986; WRISCHNIKet al. 1987; HERTZ-
haplotypes, A, B, C and D indicate the four haplotype clusters or BERG et al. 1989; SCHURR et al. 1990) were observed
lineages, with the key below the dendrogram indicating the char-
acteristic mutation of each one. The numbers at the end of each in this analysis. These defined 50 haplotypes (Table
branch refer to distinct mtDNA haplotypes, with black circles
For themtDNAcodingregions, WALLACEel al. (1987) estimateda
indicating those present in Amerinds, white circles those present in
replacement mutation rateof about 1.8 X lo-’ substitutions/site/year (SSY),
Nadene, grey circles those present in both groups, and“HYPANC” or 0.18%/myr, and a synonymous mutation rate of about 32.3 X lO-’SSY,
the hypothetical ancestor. Individual branch lengths are propor- or 3.2%/myr. The non-coding D-loop sequence, which represents approxi-
tional to the number of mutational steps between haplotypes. mately 7.0% of the total mitochondrial genome, evolves much faster, about
8.4%/myr (GREENBERG, NEWFJORD and SUGINO1983; VIGILANT et al. 1989;
HORAIand HAYASAKA 1990).
 Mitochondrial DNA Analysis 157
TABLE 3 be closer to the Nadene (0.080-0.090%) than they
Sequence divergence (SD) and divergence time (DT) of Native are to the Pima (0.149%)and Ticuna (0.179%).The
American mtDNAhaplotype clusters reduced divergence between the Maya and the Na-
dene groups is primarily due to their sharing haplo-
Cluster
Population n N S D (XlO‘) D T (YBP) type 1 . This haplotype was present in 38.8% of the
A Amerinds 10 21 7.1 k 4 . 2 17,750-35,500 Nadene and 18.5% of the Maya but is not found in
Nadene 4 60 2.1 f 2.05,250-10,500 either the Pima or the Ticuna. This result could be
B Amerinds 10 22
3.7 f 3.0
9,250-18,500 explained by either gene flow from the Nadene to
Nadene (Navajo) 6 18 3.2 k 2.5
8,000-16,000 some Amerind populations, the loss of haplotype 1 in
C Amerinds 14 10.2
26 f 25,500-51,000
5.2 the Pima and Ticuna or admixture with a third mi-
D Amerinds 7 17 7.3 f18,250-36,500
4.2 gratory group.
Evolutionary relationships: Analysis of the phylo-
For each population, n = the number of haplotypes in each
cluster; N = the total number of mtDNAs observed in each cluster; genetic relationships among Native American haplo-
SD = sequence divergence presented X104; and DT = divergence types revealed four well-defined mtDNA lineages,’
time in YBP calculated by using the evolutionary rate of 2-4%/ designated clusters A-D (Figure 1).
myr. I f all cluster B haplotypes are considered to have derivedfrom
a single migration and are pooledtogether, the corresponding The haplotypes in cluster A are associated with the
values are: n = 15; N = 40; SD = 3.8 k 3.0; and D T = 9,500- Hue111 np 663 site. This site has previously been
19,000 YBP. observed at low frequencies in mtDNAs of East Asians
[Han Chinese and Koreans (BALLINGER et al. 1992)]
1). T h e 14 restriction endonucleases used in this sur- and Chicanos4 (CANN,STONEKINC and WILSON1987).
vey screened an average of 373 sites/genome or over ClusterA is divided into two subclusters radiating
10% of the mtDNA sequence per individual. from haplotypes 1 and 9. Haplotype 1 was found in
Genetic variation within Native American popu- 38.8% of the Nadene, 18.5%of the Maya and 5.0%
lations: Intra- and intergroup sequence divergences of Taiwanese Han (haplotype 56 in the Asian phylo-
divide the populations into two groups that parallel
geny of BALLINCER et al. 1992). Haplotype 9 differs
the postulated linguistic divisions (Table 2). The mean
fromhaplotype 1 by the presence of a HaeIII np
intragroup divergence value for the
Nadene is
16,5 17 site and was detected in 8.8%of the Nadene,
0.050%, while that of the Amerinds is three times
7.4%of the Maya, 7.1 % of the Ticuna. Thishaplotype
higher at 0.151%. This result confirms the general
is only one mutationalstep away fromthe Asian
genetic cohesiveness of the Nadeneobserved in other
haplotypes 28 and 103 found in 7.7% of Koreans,
studies(SZATHMARY and OSSENBERG 1978; SZATH-
7.1% of Han Chinese and 5.0% of Taiwanese Han
MARY 1979, 1983;WILLIAMS et al. 1985) and suggests
(BALLINGER et al. 1992). Therefore, haplotypes 1 and
thatthe Amerinds have radiatedlongerthanthe
9 are likely to have been the founding haplotypes of
Nadene.
this lineage since they are nodal within the cluster,
Within the Nadene, the intragroup divergence of
the most prevalent cluster A haplotypes in both the
the Navajo is much higher (0.062%)than that of the
Dogrib (0.016%) (Table 2). This differenceresults Nadene and Amerinds and the only ones found in
from 37.5% of the Navajo having haplotypes associ- Asia. Within cluster A two haplotypes, 5 and 6, have
ated with the 9-bp deletion which is not seen in those lost the RsaI np 16,329 site, a polymorphism not
of the Dogrib or Tlingits. In fact, 6 out of the 10 previously observed. This mutation occurs in all Na-
Navajo haplotypes (haplotypes 13-1 8) are associated dene populations examined (50.0% of the Tlingits,
with the 9-bpdeletion.However, all non-deletion 26.7%of the Dogrib, and 27.1%of the Navajo), but
haplotypes observed in the Navajo ( 1 , 5 and 9) are not in Amerinds. Therefore, the RsaI np 16,329 site
shared with the Dogrib and Tlingits (Table 1). loss appears to be a specific genetic marker for the
Among the Amerinds, the Pima and Ticuna have Nadene.
similar intragroup divergences (0.132-0.138%) The haplotypes in Cluster B are defined by the 9-
whereas the Maya are somewhat less divergent bp deletion. The Hue111 site gain at np 16,517 was
(0.104%). always found in association with the deletion. Deleted
T h e interpopulational divergence values revealed mtDNAs were observed in 24.0%of theNative Amer-
intriguing relationships between tribal groups. The ican mtDNAs analyzed, with haplotype 13 represent-
South American Ticuna have similarly high diver- ing 52.5% of the mtDNAs within the cluster. This
gence values relative to both theNadeneandthe ’ When using the term “lineage” and “haplotype,” we define “lineage” as
other Amerindgroups (0.168-0.179%). T h e Pima aset of mtDNAsrelated to eachother by sharedcommonandunique
also have similar divergence values relative to all other mutations not observed in other mtDNAs. We define “haplotype”as a distinct
association of restriction endonuclease site polymorphisms for 14 enzymes
groups analyzed (0.132-0.168%), although they are within one mtDNA genome.
less divergent from the Navajo and Dogrib than are ‘T h e Occurrence of the HaeIIl np 663 site in Chicano mtDNAs (CANN,
STONEKING and WILSON1987) indicates Native American admixture in this
the Ticuna. On the other hand, the Maya appear to population; for this reason it cannot be considered a “Caucasian” marker.
158 A. Torroni et al.
haplotype is also positioned at the nodeof the cluster,
is the only one of the Amerindian deleted haplotypes
which is found in East Asians [2.9-7.1% (haplotype
54 in BALLINCER et al. 1992)], and is the only cluster
B haplotype shared by morethanonepopulation
(Navajo, Pima and Maya). Consequently,haplotype
13 is the most likely ancestral mtDNA of this cluster.
The haplotypes of cluster Care characterized by an
A to G transition at np 13,263 which simultaneously
eliminates a HincII site at np 13,259 and creates an
AluI site at np 13,262 (morph-6, WALLACE, GARRISON
and KNOWLER 1985; SCHURRet al. 1990). This mu-
tation was always found in association with a DdeI site
gain at np 10,394 and anAh1 site gain at np 10,397.
Haplotypes within this cluster are absentfrom the
Nadene but are present in all three Amerind groups
(43.3% of the Pima, 14.8% of the Maya and 32.1%
of the Ticuna)at an overall frequency of 29.9%. This
mutation is also found at low frequencies inAsian
mtDNAs, being observed in 1.8% ofJapanese (HORAI,
GOJOBORI and MATSUNACA1984), 1.6% of Orientals
(BLANCet al. 1983) and 1 out of 153 East Asians (the
only population samples analyzed with the same set of
enzymes, BALLINCER et al. 1992). The phylogenetic
analysis indicated haplotype 43 as being central to the
radiation of the cluster, and since this haplotype is
identical to the one observed in East Asians(haplotype
65) it islikely to be the founding mtDNA for this
lineage.
The haplotypes of cluster D are distinguished by a
loss of the AluI site at np 5,176 and this mutation is
virtually always associated with the DdeI np 10,394
and AZuI np 10,397 site gains. Haplotypes within this
cluster are found exclusively amongtheAmerinds
(Pomo, Maya, Ticuna), comprising 19.5% of the sam-
ples analyzed. This mutation has also been observed
in mtDNAs of 23.1% of Koreans, 14.3% of Chinese
Han and 10.0% of Taiwanese Han (BALLINGER et al.
1992). T h e phylogenetic analysis positioned haplotype
44 at the nodeof cluster D. This haplotype is the only
one shared by two tribal groups (Pomo and Ticuna)
and is only a single mutational step away from Asian
haplotype 25 of BALLINCER et al. 1992. Hence, it is
probably the founding haplotype of this lineage.
Caucasian admixture:Haplotypes 28 and 29 of the
Maya and the Navajo, respectively, lack characteristic
Native American mutations and do not cluster in any
of the four Native American lineages (Figure 1). Hap-
lotype 28 has an AluI site loss at np 7,025 found
predominantly in Caucasian mtDNAs (ANDERSON et
al. 1981; CANN, STONEKINC and WILSON1987; SHOFF-
NER et al. 1990; BROWNet al. 1992), and haplotype
29 has a DdeI site loss at np 1,7 15 found in about 10%
of Caucasian mtDNAs (SHOFFNER et al. 1990; BROWN
et al. 1992).Consequently, these haplotypes were
probably introduced by European admixture.
DISCUSSION
Native American mtDNA markers: Our analysis
shows that all Native American mtDNAs are charac-
terized by one of the fourfollowing mutations: HaeIII
np 663 site gain; 9-bp deletion; HincII np 13,259 site
loss/AluI np 13,262 site gain; and AluI np 5,176 site
loss.All of these mutations have been observed at
much lower frequencies in Asians (HORAI,GOJOBORI
and MATSUNACA1984; HARIHARA, HIRAIOMOTO and
1986; HORAI and MATSUNAGA 1986; CANN,
STONEKING and WILSON1987; HARIHARA et al. 1988;
STONEKINC et al. 1990; BALLINCER et al. 1992). The
first three of these mutations have not been observed
in Africans and Caucasians, whereas thelatter has
been observed at low frequency in Africans but not
in association with the DdeI np 10,394 and AZuI np
10,397 site gains as observed in Amerind cluster D
haplotypes (CANN,STONEKING and WILSON 1987;
BROWNet al. 1992; A. TORRONI, unpublished data).
Hence, these are highly informative genetic markers
which can be used to confirm American Indian ad-
mixture in other American ethnic groups.
Founding haplotypes:This study suggests that only
five haplotypes (1,9,13,43and 44)foundedthe
Amerinds, whereas only two or three of these (1, 9
and 13) founded the Nadene. In general, these are
the most prevalent among the populations, are shared
between different tribes, are the only ones also de-
tected in Asians, and are positioned at the nodes of
the phylogenetic haplotype clusters. Except for the
nodal haplotypes and Nadene haplotype 5, all other
mtDNA haplotypes are present at low frequencies and
are tribal specific (“private polymorphisms”; NEEL
1978), suggesting that they developed in situ in the
Americas after the initial radiation of the founding
population/s in the Americas or in Siberia.
Geneticdivergencetimes: Calculations of intra-
cluster divergences yielded similarly high values for
AmerindclustersC and D (0.102% and 0.073%;
Table 3). Likewise, the divergence value for the C ~ U S -
ter A haplotypes in the Amerinds was 0.071% which
is essentially identical to that calculated for cluster D.
T h e sequence divergenceof the Nadene cluster A was
much lower than the Amerind specific clusters, being
only 0.02 1%. The overall divergence of the cluster B
deletion haplotypes in Navajo, Pima and Maya, was of
0.038% and independent cluster B calculations for
the Navajo and Amerinds gave very similar values of
0.032% and 0.037%, respectively. Hence, the diver-
gence of cluster Bhaplotypes is intermediate between
the Amerind andNadene specific lineages.
The similarity of the sequence divergences of the
Amerind clusters A, C and D suggests that these were
the founding lineages of the Paleoindians. Using the
weighted mean value of 0.084%, the Amerind
mtDNA lineages are estimated to have begun radiat-
 Mitochondrial DNA Analysis
ing between 21,000-42,000 YBP.Ifmost of this
radiation occurred after entry into theAmericas, this
timerange would suggest thatthe Americas were
colonized before the dates associated with the oldest
Paleoindian skeletal remains (14,000 YBP) and Clovis
lithic artifacts (1 1,500YBP), and would be consistent
with the estimated ages of the oldest American ar-
chaeological sites (35,000 YBP).
The Nadene clusters A and B (Table 3) show a
much more limited divergence (0.021% and 0.032%)
than the Amerind clusters A, C and D. If both these
lineages observed in modern Nadene were present in
the founding Nadene, then the radiation time of the
Nadene would fall into the range of 5,250-16,000
YBP.
However, the only mtDNA lineage shared by all
Nadene tribesis cluster A. The overall Nadene cluster
A divergence is 0.021%, much lower thanthat of
cluster A in the Amerinds. Since the founder cluster
A haplotypes must have predated the Nadene radia-
tion, this implies that the Nadene became genetically
distinct about 5,250-10,500 YBP. These estimates of
the age of the Nadene lineages are concordant with
the 9,000 YBP date calculated from glottochronol-
ogical analyses of Nadene linguistic diversification
(GREENBERG, TURNER and ZECURA1986).
The relationship of cluster B to the Nadene and
Amerinds is unclear. Cluster Bhaplotypes are present
in 37.5% of the Navajo, 50.0% of the Pima and 22.2%
of the Maya but areabsent within the Dogrib and the
Ticuna. Moreover, the divergence of cluster B hap-
lotypes is greater than the rest of the Nadene and
lower than the rest of the Amerind lineages.
The presence of cluster B haplotypes in the Navajo
but not in the Dogrib could have occurred in three
ways: deletion haplotypes could have been part of the
original Nadene genetic stock and then lostby the
Dogrib and the Tlingit through genetic drift; they
could have beenacquired by theNadenethrough
admixture of their ancestors with the Amerinds; or
they could have arrived in an independent migration
and mixed with both Nadene andAmerinds.
The first explanation would require that theDogrib
lost all deletion haplotypes without losing any of the
nondeletion haplotypes (1, 5 and 9) which they share
with the Navajo.
The second possibility is that the Navajo acquired
the deletion haplotypes by admixture with Amerinds.
This is more probable than the first hypothesis since
the ancestors of the Navajo are believed to have
separated from other northern Athapaskan peoples
and migrated south in small familial groups through
regions exclusively inhabited by Amerinds about 1000
A.D.(HASKELL1987).Aftertheirmovement into
what is now north-central New Mexico, the Navajo
became semisedentary and mixed both genetically and
159
culturally with Puebloan Indians, borrowing both ag-
ricultural practices and ceremonial rituals(SPENCER et
al. 1977). Their genetic admixture with surrounding
peoples has also been confirmed by the presence of
Albumin variants specific to southern Native Ameri-
cans (Albumin Mexico; SCHELL and BLUMBERC 1988).
However, this scenario does not explain how the Na-
vajo acquired deletion haplotypes from the Amerinds
without obtaining any haplotypes from Amerind clus-
ters C and D (Table 1, Figure Similarly,
1). the limited
amount of divergence observed for Amerind cluster
B haplotypes (0.037%) can not be reconciled with the
higher divergence values determined for the other
Amerind clusters.
B haplotypes
The third alternativeis that the cluster
arrived in the Americas in a separate migration that
was intermediate between those of the Paleoindians
and ancestral Nadene. As this migration moved south,
perhaps along the continental divide, it would have
mixed with some Amerindtribes. The subsequent
arrival and limited southward expansion by the an-
cient Nadene would explain the preferential admix-
ture of the deletion haplotype in the Nadene and the
similar and intermediate divergence of cluster B hap-
lotypes in both modern Nadene and Amerinds. Inter-
estingly enough, inAsia the frequency of mtDNAs
characterized by the 9-bp deletion reaches its highest
value in coastal populations (BALLINCER et al. 1992)
and deletion mtDNAs were virtually the only ones
present in the migratory groupsof Asianancestry that
began to colonize Polynesia around5,000 YBP
(HERTZBERC et al. 1989). Hence this could have been
a separate migration. Clearly, further clarification of
the origin and migrations of American Indian mt-
DNAs will require characterization of the nature and
frequency of these mtDNA lineages in Asia, Siberia
and the Pacific.
The authors would like to thank C. RONALDSCOTT for contrib-
uting the Tlingit samples and JAM= DERRfor assistance with the
PAUP bootstrap analysis. This workwas supported by National
Science Foundation grant BNS8718775 awarded to D.C.W.
LITERATURE CITED
ADOVASIO, J. L., J.DONAHUE, K. CUSHMAN,R. C. CARLISLE,R.
STUKENRATH, J. D. GUNNand W. C.JOHNSON,1983 Evidence
from Meadowcroft Rockshelter, pp. 163-189 in Early Man in
the New World, edited by R. SHUTLER, JR.Sage Publications,
Beverly Hills, Calif.
ANDERSON, S., S. T. BANKIER, B.G . BARRELL,
M. H. L. DE BRUIJN,
A . R. COULSON,J. DROUIN,I. C. EPERON,D. P. NIERLICH, B.
A. ROE, F. SANGER, P. SCHREIER, A. J. H. SMITH, R. STADEN
and I. G. YOUNG, 1981 Sequence and organization of the
human mitochondrial genome. Nature 290: 457-465.
BALLINGER, S. W., T. G. SCHURR,A . TORRONI, Y. Y. GAN, J. A.
HODGE, K. HASSAN, K. H. CHEN and D. C. WALLACE,
1992 Southeast Asian mitochondrial DNA analysis reveals
genetic continuity of ancient mongoloid migrations. Genetics
1 3 0 139-152.
BLANC,H . , K. H. CHEN, M. A.D’AMORE and D. C. WALLACE.
160 A. Torroni et al.
1983 Amino acid change associated with the major polymor- KNOWLER, W. C., P. H. BENNETT, R. F. HAMMAN and M. MILLER,
phic HincIII site of oriental and Caucasian mitochondrial 1978 Diabetes incidence and prevalence inPima Indians: a
DNAs. Am. J. Hum. Genet. 3 5 167-176. 19-fold greater incidence than Rochester, Minnesota. Am. J.
BROWN, M. D., A. S. VOLJAVEC, M. T. LOTT, A. TORRONI,C.-C. Epidemiol. 108: 497-505.
YANGand D.C. WALLACE, 1992 Mitochondrial DNA Com- LAUGHLIN, W. S., 1988 From Ammassalik to Attu: 10,000 years
plex I and I11 mutations associated with Leber’s hereditary of divergent evolution. Objets et Mondes 25: 141-148.
optic neuropathy. Genetics. 1 3 0 163-173. LAWRENCE, D.N., j. A. BODMERand W. F. BODMER,1980
BROWN,W. M., 1980 Polymorphism in mitochondrial DNAof Distribution of HLA antigens in Ticuna Indians ofBrazil:
humans as revealed by restriction endonuclease analysis. Proc. results of typing a leprosy-affected family. Tissue Antigens 1 6
Natl. Acad. Sci. USA 77: 3605-3609. 152-160.
BROWN,W. M., M. GEORGE, JR. AND A. C. WILSON,1979 Rapid LEWIN,R., 1990 American Indian language dispute. Science 242:
evolution of animal mitochondrial DNA. Proc. Natl. Acad. Sci. 1632-1633.
USA 7 6 1967-1971. MACNEISH,R., 1983 Mesoamerica, pp. 125-135 in Early Man in
CANN, R. L., and A. C. WIUON, 1983 Length mutations in human the New World, edited byR. SHUTLER, JR. Sage Publications,
mitochondrial DNA. Genetics 104: 699-7 11. Beverly Hills, Calif.
CANN, R. L., W. M. BROWN and A. C. WILSON,1984 Polymorphic MARSHALL, E., 1990 Clovis counterrevolution. Science 2 4 9 738-
sites and the mechanism of evolution in human mitochondrial 741.
DNA. Genetics 104: 699-7 1 1. MATSON,G . A,, T . A. BURCH,H. F. POLFSKY,J. SWANSON, H. E.
CANN,R. L., M. STONEKING and A.C.
WILSON, 1987 Mito- SUTTONand A. ROBINSON, 1968 Distribution of hereditary
chondrial DNA and human evolution. Nature 3 2 5 31-36. factors in the blood of Indians of the Giles River, Arizona. Am.
DIAMOND, J. M., 1990 The talk of the Americas. Nature 344: J. Phys. Anthropol. 2 9 31 1-337.
589-590. MESTRINER, M. A., A. L. SIMOES and F. M. SALZANO, 1980New
DILLEHAY, T. D., and M. B. COLLINS, 1988 Early cultural evidence studies on the esterase D polymorphism in South American
from Monte Verde in Chile. Nature 3 3 2 150-152. Indians. Am. J. Phys. Anthropol. 5 2 95-1 0 1.
GILES,R. E., H. BLANC,H. M. CANN andD.C. WALLACE, 1980 MIYATA,T., H. HAYASHIDA, R. KIKUM,M. HASEGAWA, M. KOBA-
Maternal inheritance of human mitochondrial DNA. Proc. YASHI and K. KOIKE,1982 Molecular clock of silent substitu-
Natl. Acad. Sci. USA 77: 67 15-67 19. tion: at least a six-fold preponderance of silent changes in
GREENBERG, D.,
B. J. E. NEWBOLDand A. SUGINO, mitochondrial genes over those in nuclear genes. J. Mol. Evol.
1983 Intraspecific nucleotide sequence variability surround- 19: 28-35.
ing the origin of replication in human mitochondrial DNA. MORELL,V., 1990 Confusion in earliest America. Science 2 4 8
Gene 21: 33-49. 439-441.
GREENBERG, J., 1987 Language in the Americas. Stanford Univer- NEEL,j.V., 1978 Rare variants, Privatepolymorphisms, and locus
sity Press, Stanford, Calif. heterogeneity in Amerindian populations. Am. J. Hum. Genet.
GREENBERG, C. J., G. TURNER I1 and S. L. ZEGURA,1986 The 30: 465-490.
settlement of the Americas: a comparison of the linguistic, NEEL,J. V., and E. A. THOMPSON, 1978 Founder effect and the
dental and genetic evidence. Curr. Anthropol. 4: 477-497. number of private polymorphisms observed in Amerindian
GUIDON,N., and G. DELIBRIAS, 1986 Carbon-14 dates point to tribes. Proc. Natl. Acad. Sci. USA 75: 1904-1908.
man in the Americas 32,000 years ago. Nature 321: 769-77 1. NEEL,J. V., H. GERSHOWITZ, H. W. MOHRENWEISER, B. AMOS, D.
HARIHARA, S., M. HIRAI and K. OMOTO,1986 Mitochondrial D. KOSTYU, F. M. SALZANO, M. A. MESTRINER,D. LAWRENCE,
DNA polymorphism in Japanese living in Hokkaido. Jpn. J. A. L. SIMOES, P. E. SMOUSE, W. J. OLIVER,R. S. SPIELMAN and
Hum. Genet. 31: 73-83. J. V. NEEL,JR.,1980 Genetic studies ontheTicuna, an
HARIHARA, S., N. SAITOU.M. HIRAI,T. GOJOBORI,K. S. PARK,S. enigmatic tribe of Central Amazonias. Ann. Hum. Genet. 4 4
MISAWA, S. B. ELLEPOLA,T. ISHIDA and K. OMOTO, 37-54.
1988 Mitochondrial DNA polymorphism among fiveAsian NEI, M., and F. TAJIMA, 1983 Maximum likelihood estimation of
populations. Am. J. Hum. Genet. 43: 134-143. the number of nucleotide substitutions from restriction site
HASKELL, J. L., 1987 SouthernAthapaskanMigration A B . 200- data. Genetics 105: 207-217.
1750. Navajo Community College Press, Tsaile, Ariz. NELSON, D. E., R. E. MORLAN,J. S. VOGEL, J. R. SOUTHON and C.
HAURY, E. W .( 1976 The Hohokam: Desert Farmers and Craftsmen. R. HARINGTON, 1986 New datesonNorthern Yukon arti-
University of Arizona Press, Tuscon, Ariz. facts: Holocene not Upper Pleistocene. Science 232: 749-751.
HELM,J., 1981 The Dogrib, pp. 291-309 in Handbook of North SAIKI,R. K., S. SCHARF, F. FALOONA, K. B. MULLIS, G.T. HORN,
AmericanIndians, Vol. 6: Subarctic. Smithsonian Institution, H. A. ERLICHand N. ARNHEIM, 1985 Enzymaticamplifica-
Washington, D.C. tion of a-globin genomic sequences and restriction site analyses
HERZBERG, M., K. N. P. MICKLESON, S. W. SERJEANTSON, J. F. for diagnoses of sickle cellanemia. Science 2 3 0 1350-1354.
PRIORand R. J. TRENT,1989 An Asian-specific 9-base pair SALZANO,F. M., S. M. CALLEGARI~ACQUESJ. andV. NEEL,
deletion of mitochondrial DNA is frequently found in Polyne- 1980 Genetic demography of the Amazonian Ticuna Indians.
sians. Am. J. Hum. Genet. 44: 504-510. J. Hum. Evol. 9 179-191.
HORAI,S., and K. HAYASAKA, 1990Intraspecific nucleotide se- SCHELL, L. M., and B. S. BLUMBERG, 1988 Alloalbuminemia and
quence differences in the major noncoding region of human the migration of Native Americans. Yearb. Phys. Anthropol.
mitochondrial DNA. Am. J. Hum. Genet. 4 6 828-842. 31: 1-14.
HORAI,S., T . GOJOBORI and E. MATSUNAGA, 1984 Mitochondrial SCHURR, T. G., S. W. BALLINGER,Y . Y .CAN,J. A. HODCE,D. A.
DNA polymorphism in Japanese. Hum. Genet. 6 8 324-332. MERRIWETHER, D.N. LAWRENCE, W.C. KNOWLER,K. M.
HORAI,S:, and E. MATSUNAGA,1986 Mitochondrial DNApoly- WEIS and D.C. WALLACE, 1990 Amerindian mitochondrial
morphism in Japanese. 11. Analysis with restriction enzymes of DNAs have rare Asian mutations at high frequencies, suggest-
four or five base pair recognition. Hum. Genet. 72: 105-1 17. ing they derived from four primary maternal lineages. Am. J.
JOHNSON, M. J., D. C. WALLACE,S. D. FERRIS,M. C. RATTAZZI and Hum. Genet. 4 6 613-623.
L. L. CAVALLI-SFORZA, 1983 Radiation of human mitochon- SHOFFNER, J. M., IV, M. T. LOTT, A. M. S. LEZZA,P. SEIBEL,s.
drial DNA types analyzed by restriction endonuclease cleavage W. BALLINGER and D. C. WALLACE, 1990 Myoclonic epilepsy
patterns. J. Mol. Evol. 1 9 255-27 1. andragged-red fiber disease (MERRF) is associated with a
 Mitochondrial DNA Analysis 161
mitochondrial DNA tRNALYsmutation. Cell 61: 931-937. mitochondrial genes sustain seventeen times more mutations.
SPENCER, R. F., J. D. JENNINGS, E. JOHNSON, A. R. KING,T . STERN, Curr. Genet. 12: 81-90.
K. M. STEWART and W. J. WALLACE, 1977The Native Ameri- WALLACE D. C., G. SINGH,M. T. LOTT,J. A. HODGE, T . G. SCHURR,
cans. Harper & Row, New York. A. M. S. LEZZA,L. J. EISAS I1 and E. NIKOSKELAINEN,
STONEKING, M., K. BHATIAand A. C. WILSON,1986 Rate of 1988 Mitochondrial DNA mutation associated with Leber’s
sequence divergence estimated from restriction maps of mito- hereditary optic neuropathy. Science 242: 1427-1430.
chondrial DNAs from Papua New Guinea. Cold Spring Harbor WILLIAMS, R. C., H. G. MORSE, M . D. BONNELL, R. G. RATE and
Symp. Quant. Biol. 51: 433-439. T. T. KUBERSKI, 1981 The HLA loci of the Hopi and Navajo.
STONEKING, M.,L. B. JORDE,K. BHATIAand A. C. WILSON, Am. J. Phys. Anthropol. 5 6 291-296.
1990 Geographic variation in human mitochondrial DNA WILLIAMS, R. C., A. G. STEINBERG, H. GERSHOWITZ, P. H. BEN-
from Papua New Guinea. Genetics 124: 717-733. NETT, W. C. KNOWLER,D. J. PETTITT,W. BUTLER, R. BAIRD,
SWOFFORD, D., 1989 Phylogenetic Analysis Using Parsimony (PAUP), L. DOWDA-REA, T. A. BURCH,H. G. MORSEand C. G . SMITH,
Version 3.0m. Illinois Natural History Survey, Champaign, Ill. I985 GM allotypes in Native Americans: evidence for three
SZATHMARY, E. J. E., 1979 Blood groups of Siberians, Eskimos, distinct migrations across the Bering land bridge. Am. J. Phys.
subarctic and northwest coast Indians: the problem of origins Anthropol. 66: 1-19.
and genetic relationship, pp. 185-210 in The First Americans: WILLIAMS, R. C., A.G. STEINBERG, W. C. KNOWLERand D. J.
Origins, Afjnities and Adaptations, edited by W. S. LAUGHLIN PETTITT, 1986 Gm 3;5,13,14and stated admixture: inde-
and A. B. HARPER. Gustav Fischer, New York. pendent estimates of admixture in American Indians. Am. J.
SZATHMARY, E. J. E., 1983 Dogrib Indians of the Northwest Hum. Genet. 39: 409-413.
Territories, Canada: genetic diversity and genetic relationship WRISCHNIK, L. A., R. G. HIGUCHI,M. STONEKING, H. A. ERLICH,
among subarctic Indians. Ann. Hum. Biol. 1 0 147-162. N. ARNHEIM and A. C. WILSON,1987 Length mutations in
SZATHMARY, E. J. E., 1984 Peopling of northern North America: human mitochondrial DNA: direct sequencing of enzymatically
clues from genetic studies. Acta Anthropol. 8 79-109. amplified DNA. Nucleic Acids Res. 15: 529-542.
SZATHMARY, E. J. E., 1991 Modeling ancient population relation- Communicating editor: V. G. FINNERTY
ships from modern population genetics, in Method and Theory
for Investigating the Peopling to the Americas, edited by R. BON-
NICHSEN and D. G. STEELE. Texas A&M Press, College Station,
APPENDIXA
Tex. (in press). Oligonucleotide primers for PCR amplifications of
SZATHMARY, E. J. E., R. E. FERRELLand H. GERSHOWITZ,
1983 Genetic differentiation in Dogrib Indians: serumpro-
Native American mtDNAs are shown in Table 4.
tein and erythrocyte enzyme variation. Am.J. Phys. Anthropol.
62: 249-254. TABLE 4
SZATHMARY, E. J. E., and N. S. OSSENBERG, 1978 Are the biolog-
ical differences between North American Indians and Eskimos Oligonucleotide primers for PCR amplifications of Native
truly profound? Curr. Anthropol. 1 9 694-701. American mtDNAs
SZATHMARY, E. J- E.,C. RITENBAUGH, and C. S. GOODBY,
1987 Dietary changes and plasma glucose levels in an Amer- Primer coordinates TH(“C) Segment size (bp)
indian population undergoing cultural transition. SOC.Sci.
1562-1581,3717-3701 51 2155
Med. 24: 791-804.
3007-3023,5917-5898 55 2910
TAYLOR, R. E., L.A. PAYEN,C. A. PRIOR,P. J. SLOTA,JR., R.
5317-5333,7608-7588 57 2291
GILLESPIE, J. A. J. GOWLETT, R. E. M. HEDGES, A. J. T . HULL,
7392-74 10,8921-8902 57 1529
T . H. ZABEL,D. J. DONAHUE and R. BERGER,1985 Major
8282-8305,10107-10088 57 1825
revisions in the Pleistocene age assignments for North Ameri-
9911-9932,11873-11851 69 1962
can human skeletons by C-14 accelerator mass spectrometry:
11673-11691,13950-13932 57 2277
none older that 11,000 C-14 years B.P. Am. Antiq. 50: 136-
13914-13930,16547-16527 47 2633
140.
16453-16472.1696-1677 61 1812
TROUP, G. M., M. S. SCHANFIELD, C. H. SINGARAJU, R. L. HARVEY,
J. JAMESON, J. CAPPERand B. BAKER,1982 Study of HLA Primer pair are numbered according to ANDERSON et al. (1 98 I),
alloantigens of the Navajo Indians of North America. Tissue with the 5’ + 3’ coordinates before the comma corresponding to
Antigens 2 0 339-35 1. the forward primer and those afterthe comma to the reverse
VIGILANT, L., R. PENNINGTON, H. HARPENDING, T. D. KXHER primer. The THused for annealing was the lowest for the primer
pair as calculated from the nucleotide sequence of each primer,
and A. C. WILSON,1989 Mitochondrial DNA sequences in
where TH= 4(C + G) + 2(T + A) - 5’. All primers were synthesized
single hairs from a southern African population. Proc. Natl. at the Emory University Microchemical Facility.
Acad. Sci. USA 86: 9350-9354.
WALLACE,D. C., K. GARRISONand W. C. KNOWLER, 1985
Dramatic founder effects in Amerindian mitochondrial DNAs.
Am. J. Phys. Anthropol. 68: 149-155. APPENDIX B
WALLACE, D. C., J. YE, S. N. NECKELMANN, G . SINGH,K. A.
WEBSTER and B. D. GREENBERG, 1987 Sequence analysis of Figure 2 shows the polymorphicrestriction sites
cDNAs for the human and bovine ATP synthase p subunit: observed in Native American mtDNA haplotypes.
162 A. Torroni et al.
SITES HAPLOTYPES

11111111112222222222333333333344444444445
12345678901234567890123456789012345678901234567890

29c
6630
9311
1:M
171%
3192c
3337k
3371k
3388e/3391.
3403t
38991
3981.
4310a
4546g
4769a*
5176a
5584a
5983g
6204a
7025a+
7497e
78530
78593
8569c/8572s
87741
8858f*
9052n/9053f
9504.
9714.
9942j
10091.
10254j
10394~
10397..
108931
111ooa
11321.
11793c
119241
12406h/124060
13031g
132590
132590/13262a
13633e/136340
13702e*
14015a
141990*
14268g*
14304a

15073c
15172e
16049k
161450
:::I
16329k
16389 rn/16390b
163d+g/1639Ob
16398j
16456~
16511.
Region V

FIGURE2.-Table presenting the polymorphic restriction sites

observed in Native American mtDNA haplotypes. Haplotypes are
numbered according to Table 1. A “ 1 ” indicates the presence of a
site and a “0” indicates the absence of a site except for region V
where ‘‘ 1 ” indicates a single copy of the 9-bp repeat (deletion) and
“2”indicates two copies ofthe repeat. Sites are numbered from the
first nucleotide of the recognition sequence according to the pub-
lished sequence (ANDERSONet al. 1981). Boldface numbers indicate
site gains relative to the published sequence and standard type
numbers indicate site losses. The 14 restriction enzymes used in the
analysis are designated by the following single-letter code (after
CANN,BROWNand WILSON1984): a, AluI; b, AuaII; c, DdeI; e,
HaeII1; f, HhaI; g, HznfI; h, HpaI; i, HpalI; j , MboI; k, RsaI; 1, TaqI;
111, BamHI; n, HaeII; 0,HZncII. Sites separated by a diagonal line
indicate either simultaneous site gains or site losses for two different
enzymes or a site gain for one enzyme and a site loss for another
because of a single inferred nucleotide substitution; these sites are
considered to be only one restriction site polymorphism in the
statistical analysis. Sites marked with an asterisk were found to be
present or absent in all samples (except where polymorphic) con-
trary to the published sequence, and were confirmed in mtDNAs
from all major ethnic subdivisions by DNA sequencing (WALLACE
et al. 1988; SHOFFNER et al. 1990; BROWN et al. 1992; A. TORRONI,
unpublished data).