A Recent Insertion of an Alu Element on the Y Chromosome Is a Useful

Marker for Human Population Studies

Michael F. Hammer
Laboratory of Molecular Systematics and Evolution, University of Arizona

A member of the Alu family of repeated DNA elements has been identified on the long arm of the human Y
chromosome, Yq 11. This element, referred to as the Y Alu polymorphic (YAP) element, is present at a specific site
on the Y chromosome in some humans and is absent in others. Phylogenetic comparisons with other Alu sequences
reveal that the YAP element is a member of the polymorphic subfamily-3 (PSF-3), a previously undefined
subfamily of Alu elements. The evolutionary relationships of PSF-3 to other Alu subfamilies support the hypothesis
that recently inserted elements result from multiple source genes. The frequency of the YAP element is described
in 340 individuals from 14 populations, and the data are combined with those from other populations. There is
both significant heterogeneity among populations and a clear pattern in the frequencies of the insertion: sub-
Saharan Africans have the highest frequencies, followed by northern Africans, Europeans, Oceanians, and Asians.
An interesting exception is the relatively high frequency of the YAP element in Japanese. The greatest genetic
distance is observed between the African and non-African populations. The YAP is especially useful for studying
human population history from the perspective of male lineages.

Introduction
The human Y chromosome has useful properties
for evolutionary studies. With the exception of the pseu-
doautosomal region, it is inherited paternally and does
not recombine. As a consequence, the DNA sequence
of every Y chromosome preserves a unique record of
mutational events that occurred in previous generations.
Studies of polymorphisms in the nonrecombining por-
tion of the Y chromosome have been proposed for de-
tecting male-mediated migration events and for recon-
structing paternal history (Casanova et al. 1985; Ngo et
al. 1986; Malaspina et al. 1990). Unfortunately, there
are only a handful of Y-linked polymorphisms reported
in the literature (Jakubiczka et al. 1989; Malaspina et
al. 1990; Spurdle and Jenkins 1992)) and most of these
are based on repeated Y-DNA sequences that have
proved difficult to interpret. For example, Y-chromo-
some probes 49f and 49a detect multiple haplotypes
(Ngo et al. 1986), but the mechanism of polymorphism
is not understood, and there is evidence that fragments
of identical size have arisen more than once (Torroni
Key words: human, Alu, Y chromosome, polymorphism, parsi-
mony. Abbreviation: YAP = Y Alu polymorphism, Y Alu polymorphic
element.
Address for correspondence and reprints: Michael F. Hammer,
Laboratory of Molecular Systematics and Evolution, Biosciences West,
University of Arizona, Tucson, Arizona 8572 1.
Mol. Biol. Evol. 11(5):749-76 I. 1994.
0 1994 by The University of Chicago. All rights reserved.
0737-4038/94/i IOS-0004$02.00
et al. 1990; Spurdle and Jenkins 1992). Similar problems
exist with several other Y-specific polymorphisms
(Spurdle et al. 1994). Therefore, there is a need to iden-
tify Y-specific polymorphisms with well-characterized
genetic mechanisms.
In this report, a simple and stable polymorphism
resulting from the recent insertion of an Alu-family
member on the long arm of the Y chromosome is de-
scribed. As there is no evidence for a mechanism of spe-
cific removal of Alu family members, Alu elements are
stable over millions of years (Sawada et al. 1985; Dein-
inger 1989). Alu elements at specific loci that are not
fixed in human populations are likely to have resulted
from unique and recent insertion events. This type of
marker has been shown to be valuable for human pop-
ulation studies (Batzer and Deininger 199 1; Perna et al.
1992). In the present paper, the distribution and fre-
quency of the Y Alu insertion is investigated in 14 hu-
man populations.
The Alu family of short interspersed repeats
(SINES) is found in all primate species, with -500,000
copies per haploid genome present in humans (Dein-
inger 1989 ) . Alu-family members are believed to have
spread throughout primate genomes by a transposition
process involving an RNA intermediate. However, most
copies of Alu sequences are unable to be transcribed and
are considered to be pseudogenes by some authors
(Schmid and Shen 1985; Deininger et al. 1992). Com-

749
750 Hammer
parisons between Alu sequences in the database dem-
onstrate an average of -70% sequence identity. On the
basis of analyses of specific diagnostic positions, several
authors have placed Alu elements into specific subfam-
ilies (Slagel et al. 1987; Willard et al. 1987; Britten et
al. 1988; Jurka and Smith 1988; Quentin 1988). The
different degrees of divergence exhibited by the members
of each subfamily suggest that they appeared at different
evolutionary times.
It is generally agreed that these subfamilies are the
result of successive waves of retroposon events that orig-
inated from a small number of different source genes
(Labuda and Striker 1989). Consistent with this inter-
pretation is the finding that the few human Alus that
are known to be polymorphic-and that therefore are
known to be the most recently inserted-belong to
subfamilies with the lowest levels of sequence divergence
among their members (Matera et al. 1990; Shen et al.
199 1). However, controversy has surrounded the issue
of the number of source genes that are active. The sim-
plest model posits that there is only one transcriptionally
active source gene at any given evolutionary time and
that the different subfamilies have resulted from se-
quence changes in the lineage of this gene (Shen et al.
199 1). This model predicts that all polymorphic Alu
sequences belong to a single subfamily or a single lineage
of closely related subfamilies. A competing model posits
that multiple source genes, deriving from independently
evolving lineages, have given rise to unrelated subfam-
ilies ( Leeflang et al. 1992 ) .
To test hypotheses on the number of active source
genes, the sequence of the Y Alu Polymorphic (YAP)
element is compared with those of other recently inserted
Alu elements. A parsimony analysis is performed to de-
termine the branching pattern of these elements and the
phylogenetic relationships of recently formed subfami-
lies.
Subjects, Material, and Methods
Subjects
A total of 340 unrelated males were analyzed (table
1). DNA samples were isolated from volunteers living
in Tucson or were provided by other investigators as
follows: 2 1 English ( Persichetti et al. 1992)) 24 Italians,
30 Egyptians and 2 1 Saudi Arabians, by A. Novelletto;
6 Kenyans, 3 Zambians, and 8 Nigerians, by J. S.
Wainscoat ( Wainscoat et al. 1986); 49 Gambians by
A. V. S. Hill (Hill et al. 199 1); 14 Mbuti Pygmies from
Zaire, 12 Biaka Pygmies from the Central African Re-
public (Bowcock et al. 199 1 ), and 2 Ethiopians, by J.
Rogers, J. Kidd, and K. Kidd; 16 Western Bantu-speak-
ers from South Africa by Amanda Spurdle (Spurdle and
Jenkins 1992 ) ; 5 Chinese and 13 Japanese, by J. Rogers,
J. Kidd, and K. Kidd; and 48 samples from highland
Papua New Guinea and 7 samples from Australia, by
M. Stoneking (Stoneking and Wilson 1989).
DNA Probe
A 5.5-kb Hind111 restriction fragment (Page et al.
1987) was subcloned into the Hind111 site of pEMBL.
A 2.8-kb EcoRV fragment released from this vector was
further subcloned into the SmaI site of pBluescript SK
( Stratagene) and was named pYAP-2.8 (Genome Data
Base DYS287). On digestion with EcoRI and BglII, a
500-bp fragment was released from pYAP-2.8 (fig. 1C).
After isolation from a low-melting-temperature agarose
gel, the fragment was labeled with 32P by the method of
Feinberg and Vogelstein ( 1984).
DNA Extraction and Hybridization
Human genomic DNAs were prepared from whole
blood by standard procedures. Approximately 2 ug of
DNA was digested with restriction endonucleases, elec-
trophoresed on 0.8% agarose gels, and alkaline trans-
ferred to nylon membranes. Hybridization with radio-
actively labeled probes was carried out according to the
procedure of Hammer and Silver ( 1993 ).
DNA Sequencing
DNA sequences were determined by the dideoxy-
chain-termination method (Sanger et al. 1977) using
the Sequenase kit (U.S. Biochemical) according to
manufacturer’s specifications. Once sequences were ob-
tained for the ends of the pYAP-2.8 insert by using uni-
versal primers to the plasmid vector, internal sequences
were determined by primer walking (author’s unpub-
lished data). Genomic DNAs from individuals with a
300-bp indel were amplified with internal primers and
were directly sequenced by the method of Casanova et
al. (1990).
Data Analysis
The method of Reynolds et al. ( 1983), with a cor-
rection for sample size, was used to calculate F,, . Clus-
tering diagrams were generated using the neighbor-join-
ing method (Saitou and Nei 1987 ) and the unweighted-
pair-group method using arithmetic averages (UPGMA)
(Sokal and Michener 1958).
Phylogenetic analysis was performed on Alu se-
quences selected from the GenBank (release 76.0) and
EMBL (release 30.0) databanks by using a similarity-
analysis program (Fasta in GCG) that uses the method
of Pearson and Lipman ( 1988 ) . The YAP Alu sequence
and members of the HS- 1 subfamily (Shen et al. 199 1)
were used as query sequences with a k-tuple value of 6.
Thirty-three sequences from the top 50 scoring sequences
were selected. In addition, six more distantly related Alu
 Alu-Element Polymorphism on the Human Y Chromosome 751

Table 1
Frequency (%) of the Y Alu Insertion

Population (n) Frequency of Alu+ + SE Reference(s)

Europe:
United Kingdom (28) ........... 0.07 f 0.05 Persichetti et al. 1992
Present study
Italy (45) ...................... 0.11 IL 0.05 Persichetti et al. 1992
Present study
Sardinia (48) ................... 0.08 f 0.04 Persichetti et al. 1992
South African European (5 1) ...... 0.04 + 0.03 Spurdle et al. 1994
Other (20) ..................... 0.00 + 0.00 Present study
Total (192) .................. 0.07 + 0.02
Northern Africa:
Egypt (64) ..................... 0.53 + 0.06 Persichetti et al. 1992
Present study
Africa:
Eastern Africa (8) ............... 0.75 f 0.15 Present study
Zambia (3) .................... 1.00 f 0.00 Present study
Nigeria (8) .................... 1.00 I!I 0.00 Present study
Gambia (44) ................... 0.86 + 0.05 Present study
Eastern Pygmy ( 14) ............. 0.57 + 0.13 Spurdle et al. 1994
Present study
Western Pygmy (24) ............ 0.71 f 0.09 Spurdle et al. 1994
Present study
Bantu (442) ................... 0.78 f 0.02 Spurdle et al. 1994
Present study
Khoisan (68) ................... 0.46 + 0.06 Spurdle et al. 1994
Total (611) .................. 0.74 f 0.02
Near East:
Saudi Arabia (2 1) ............... 0.10 -t 0.07 Present study
Asia:
Japan(21) ..................... 0.24 + 0.09 Present study
China (2 1) .................... 0.00 + 0.00 Present study
India (63) ..................... 0.00f 0.00 Spurdle et al. 1994
Other(l1) ..................... 0.00f 0.00 Present study
Total (116) .................. 0.04 k 0.02
Oceania:
Australia (7) ................... 0.14 + 0.13 Present study
Papua New Guinea (48) ......... 0.00 + 0.00 Present study
Other (3) ...................... 0.00AI0.00 Present study
Total (58) ................... 0.02 f 0.02

sequences representing the J, Sx, Sp, and Sb subfamilies analyzed using the branch-and-bound option, which
(Jurka and Milosavljevic 199 1) were included as out- guarantees finding the most parsimonious trees. PAUP
groups. Alignments were carried out using the method also was used for computing strict-consensus trees (Sokal
of Feng and Doolittle ( 1987 ) and were refined manually and Rohlf 198 1), for performing bootstrap analyses as
to match published alignments as closely as possible. a means of estimating relative support for clades ( lOO-
Parsimony analyses were conducted using PAUP 1,000 replicates) (Felsenstein 1985 ) and for calculating
3.0 for the Macintosh computer (Swofford 1993). Two character and tree statistics (consistency index; Kluge
rounds of analyses were performed. In the first round, and Farris 1969). A neighbor-joining analysis was per-
the heuristic search option was performed on 43 se- formed on all 43 taxa ( Saitou and Nei 1987).
quences. In each of 100 replications, starting trees were
Nomenclature
obtained by a random stepwise addition sequence, and
branch swapping was performed on 100 trees by using The nomenclatures of Jurka and Milosavljevic
the tree bisection-reconnection algorithm. Only minimal (1991) (Sb), Shen et al. (1991) (CS/HS), and Matera
length trees found in each replication were kept. In the et al. ( 1990) (PV/Precise) are used unless otherwise
next round, a representative subset of 18 sequences was noted.
752 Hammer

A 1234565
Interval 5-O
A I
a I
" .. F------

Probe
C
R BT
I 1 1

FIG. 2.-Autoradiograph of three human DNA samples digested

Scale (kb)
FIG. 1.-Structure of YAP element region. A, Schematic represen-
tation of the human Y chromosome. The hatched box denotes the pseu-
doautosomal region, the shaded oval denotes the centromere, and the
blackened box indicates interval 5-O to which the YAP element has been
mapped ( Vollrath et al. 1992). B, Restriction map of the region encom-
passing the YAP element. Restriction-enzyme sites are indicated with
capital letters (T = TaqI; V = EcoRV; and B = BglII), and the Alu
element is represented by a triangle. C, Probe used in the survey. This
probe is a OS-kb EcoRI-BglII fragment that is released from the plasmid
vector, pYAP-2.8 (see Subjects, Material, and Methods).
with two restriction enzymes and hybridized with the OS-kb pYAP
probe. Genomic DNAs in lanes 2-4 were digested with TaqI, and those
in lanes 5-7 were digested with EcoRV. Lane 1, Lambda-Hind111 ladder.
Lanes 2 and 5, Gambian 1 ( Alu-). Lanes 3 and 6, Gambian 2 ( Alu+).
Lanes 4 and 7, Gambian 3 ( Alu-).
posed sites preferred for Alu integration (Daniels and
Deininger 1985). Only one copy of the 1 1-bp sequence
is present in individuals lacking an Alu element. The
Alu sequence terminates with an oligo-dA tail composed
of 26 adenine residues.

Results
1
PSF-3 GGCCGGGCGCGGTGGCTCAdGCCTGTAATCCA~ACTTT 61
YAP ............................................................
YAP cs ...................................................... ..c ...
HS-1 ...................................................... ..c ...
The hybridization of pYAP-0.5 to genomic DNAs HS-2 ...................................................... ..c ...

digested with different restriction enzymes gave evidence 8 . 2

PSF-3 TCATGAGGTCAGGAGATCGAGACCATCCTGGCTAACAAGd 121
of a 0.3-kb indel polymorphism (fig. 2). The 300-bp YAP . ..c........................................................
indel maps between a BgZII and a TuqI site within the cs
HS-1
. ..C . . . . . . . . . . . .. . . . . . . . . . .. . .. . . . . ..c......................
. . .C . . . . . . . . . . . . . . . . . . . . . . ..C......A.C......................
HS-2 . ..C........................C......A.C......................
2%kb EcoRV fragment (fig. 1B). When DNA se-
quences of the region between the BgZII and TaqI sites 3
PSF-3 AAATACAAAAAATTAGCCGdGCGCGGTGGCGGGCGCCTGT 181
were determined, individuals with the larger EcoRV re- YAP ............................................................
CS ..................... ..T ....................................
striction fragments were found to have an inserted copy HS-1 ..................... ..TA............................T ......
HS-2 ..C ....... ..-..........TA....................T.......T ......
of the Alu family of repeated sequences (fig. 3). The
Alu element was found to be present between the same PSF-3
9 4
CTGAGGCAG;;AGAATGGCGTGAACCCGGGAAGCGGAGCT
5
241
two base pairs in several individuals from different pop- YAP ........................ ..A .................................
cs ............................ ..G........................C ....
ulations, confirming the single origin of this insertion HS-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..G........................CC...
HS-2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..G........................CC...
(data not shown). The shorter TaqI fragment associated
with insertion is explained by the presence of an addi- PSF-3
10 6. 7. 11 .
CACTGCAGTCCGCAGTCCGGCCTGGGCGACAGAGCGAGAkCCGTCTCAiA
tional TaqI site within the Alu element. This member YAP
cs
..T........................A.......................
. . . . . . .C . . .-------A....... .. . .. . . . . .. . . . . . .. . . . . . . .
HS-1 C . _______A._
of the Alu family of repeated elements is named the HS-2
.

.
.

.
.

.
.

.
.

.
.

.
.

. C
.

.
.

. . ______-A . .
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

YAP element
FIG. 3.-Comparison of the YAP element sequence and four Alu
YAP-Element Sequence consensus sequences. The CS, HS-1, and HS-2 consensus sequences
are from Shen et al. ( 199 1). The PSF-3 consensus is described in the
The YAP element is flanked on each side by an text and is defined by diagnostic mutations 1-7 as well as by a 7-bp
exact A-rich 11-bp direct repeat, consistent with the pro- duplication (indicated by dashes) at the 3’ end of the sequence.
 Alu-Element Polymorphism on the Human Y Chromosome 753

By means of the diagnostic-site method of Jurka Sub-Family

TPA25
and Milosavljevic ( 199 1)) the YAP element is identified Consensus 1 PSFZ
C4N5
PV92
as a member of the Sb subfamily. Except for the A at
PlN5
position 98, it has all of the diagnostic sites (fig. 3, po- G15N2
CHPV
sitions 76, 86, 93, 132, 152, 196, 199, and 218) of the APOA
CS subfamily (Shen et al. 199 1). In addition, it has all C3Nl
C3N2
five diagnostic mutations (fig. 3, positions 89, 96, 145, C3N3
174, and 237) that distinguish the human CS/ Precise C3N4
C3N6
subfamily from the HS/PV subfamily. According to C3N7
C4N2 PSFl
these criteria, the YAP element belongs to the older CS/
C4N6
Precise / Sb subfamily. G18Nl
However, the YAP-element sequence differs at 10 G18N2
CS/F’recise/Sb
G19Nl
Consensus
positions ( l-7 and 9- 11 in fig. 3) from the CS/Precise H3Nl
MLVI
consensus sequences (3.5%), and there is a 7-bp dupli- C4N4
cation 30 bp from the 3’ end of the Alu sequence. The C4N8
NEUR
sequence differences at positions 98 and 248 are trans- PV83
versions, while the remaining five differences at positions PlN6
CHOL
57, 144,207,2 11,236,243,259, and 268 are transitions. LCAT
Positions 57, 207, 236, and 268 involve changes from LDLR
HDAB PSF3
CpG in the consensus sequence to TpG or CpA in the BGP
YAP-element sequence. CpG dinucleotides have been HD
YAP I
found to mutate to TpG or CpA at a rate - 10 times NE53
AFP
higher than the normal rate in recently inserted Alu VWFA
family members (Labuda and Striker 1989). VWFB
LDHB
ClINH
Phylogenetic Analysis GOR
TPAl4
To investigate the phylogenetic relationships among ADAG
recently inserted Alu elements, a parsimony analysis was APOE
GAL
performed on the YAP-element sequence and 42 Alu
sequences reported in the literature (Appendix). Other FIG. 4.-Parsimony tree for 43 Alu sequences. The tree is a strict
consensus of the 5,600 equally parsimonious trees of length 199, dis-
than the outgroup (galago) and older subfamily se-
covered via a heuristic search. Circles at the ends of the branches indicate
quences (TPA 14, ADAG, and APOE), the sequences polymorphism of the Alu element in human populations. Shaded ver-
were chosen on the basis of similarity to the YAP-ele- tical bars refer to the consensus sequence from which sets of Alu se-
ment sequence or to HS/ PV subfamily members. An quences are derived. Bootstrap percentages are given in circles within
effort was made to include all known polymorphic Alu the branches.
elements. A heuristic search, performed on 77 infor-
mative characters in the 43 sequences, yielded 5,600
equally parsimonious trees of length 199 (confidence branches from the HS/ PV consensus sequence. Al-
index [ CI ] = 0.53 ) . Two well-defined clusters of se- though there is not strong support for this clade in 1,000
quences appear in the strict-consensus tree shown in fig- bootstrap replications, this subfamily will be referred to
ure 4. Twenty-three sequences of the HS-1 and HS-2 as polymorphic subfamily-2 ( PSF-2 ) .
(Shen et al. 199 1) and PV subfamilies (Leeflang et al. The YAP element is a member of a second cluster
1992) form a clade that branches from the HS/PV con- of seven Alu sequences that branches from the CS/Pre-
sensus sequence. Two HS/PV subfamily members, cise/Sb consensus sequence. This clade, supported in
PV83 and P 1N6, diverge one step before the HS/PV 88% of the bootstrapping replications, is referred to as
consensus, suggesting that they were transcribed from a polymorphic subfamily-3 (PSF-3). Six of the Alu se-
more ancestral form of the HS/PV source gene. The quences branch from the PSF-3 consensus sequence,
HS/PV clade includes the chimpanzee PV sequence, whereas the YAP sequence diverges one step before the
consistent with the observation that Alu elements in the divergence of the other PSF-3 elements. The ancestral
HS/PV subfamily are not necessarily human specific C at nucleotide position 64 in the YAP sequence (fig.
(Leeflang et al. 1992). Therefore, this clade will be re- 3, position 8) mutated to a T in all of the other PSF-3
ferred to as the polymorphic subfamily-l (PSF- 1). The sequences.
TPA25 and C4N5 Alus, previously assigned to the HS- Two Alu elements, NE53 and AFP, branch from
2 subfamily ( Shen et al. 199 1), form a clade that also the lineage leading to the PSF-3 subfamily. Four lineages
754 Hammer

Ah
that branch independently from the CS / Precise consen-
Subfamik
sus sequence lead to the remaining five Alu sequences
in the analysis, one of which is the polymorphic element
within the C 1INH locus ( Stoppa-Lyonnet et al. 1990). PSF-1
A neighbor-joining analysis gave similar results.
Four branches led from the CS/ Precise consensus: the
first to the VWFA and VWFB sequences; the second to
the PSF-1 and PSF-2 sequences; the third branch to the
PSF-3 subfamily, including NE53 and AFP; and the CHOL -

fourth to the remaining LDBH, Cl INH, and GOR se- .5 LCAT

1
LDLR
quences.
HDAB PSF-3
A second parsimony analysis was performed on a BGP
subset of Alu sequences by using the branch-and-bound l’ 99 HD

option of PAUP, which guarantees finding the most par- YAP .

simonious trees. All of the PSF-3 Alus were included in ClINH

the analysis, as were all of the Alus that are known to 10 TPA14 SP
be polymorphic in human populations. The analysis of 24 ADAG sx
4 1 phylogenetically informative characters yielded 24 37 AP0.E J
equally parsimonious trees with 68 steps (CI = 0.72).
I I

A strict-consensus tree ( fig. 5 ) depicts three main lineages I

20 15 10
8

stemming from the CS / Precise / Sb consensus sequence.
The first leads to the PSF-1 and PSF-2 subfamilies, the
second leads to the PSF-3 subfamily, and the third is the
lineage leading to the ClINH Alu sequence. Bootstrap-
ping gave strong support for the assignment of all the
repeats (except C IINH) to only one of the three
subfamilies. The C 1INH sequence, diverging from the
CS/Precise source gene, clearly lies outside the other
PSF subfamilies.
There are eight substitutions along the branch
leading from the CS/Precise/Sb consensus to the com-
mon ancestral sequence, or source gene, for the PSF-3
subfamily (fig. 5). These sites correspond to the diag-
nostic sites for this subfamily (fig. 3, positions l-7). In
addition, all seven members have the 7-bp duplication
30 bp from the 3’ end of the Alu sequence. The lineage
leading to the YAP-element sequence diverges one step
before the PSF-3 consensus, as the YAP sequence lacks
one of the diagnostic sites of the PSF-3 subfamily (fig.
3, position 8). The eight diagnostic sites include two
transversions and six transitions, three of which involve
CpG dinucleotides that have mutated to TpG.
Sequence Divergence
The PSF-3 subfamily sequences differ from each
other by an average of 2.1%, which is similar to the av-
erage pairwise divergence of complete PSF-1 and PSF-
2 sequences examined in this survey. However, when
CpG-to-TpG/CpA substitutions are ignored, there is a
1. I%, 0.7%, and 1.2% sequence divergence within the
PSF- 1, PSF-2, and PSF-3 subfamilies, respectively (table
2). The PSF-3 member sequences differ from the PSF-
3 consensus by an average of 0.63%. This compares with
5
I
0
Millions of Years
FIG. 5.-Parsimony tree for 18 Alu sequences. The tree is a strict
consensus of the 24 equally parsimonious trees of length 68, discovered
via a branch-and-bound search. The number of steps is given above the
branches, and the percent sequence change (non-CpG to CpA/TpG
sites only) is shown below the branches. Other numbers and symbols
are as in fig. 4. The timescale is based on the number of CpG-to-TpG/
CpA changes and an evolutionary rate of 0.15% /Myr (see text).
0.54% and 0.4% for the PSF-1 and PSF-2 subfamilies,
respectively.
The mean intersubfamily divergences are also given
in table 2. Considering only non-CpG-to-TpG /CpA
substitutions (below the diagonal in table 2), the PSF-
1 and PSF-3 consensus sequences, as well as the C 1INH
sequence, have diverged from the CS consensus by 1.8%.
The PSF-2 consensus has diverged from the PSF-1 con-
sensus by 0.7%.
Population Data
The frequency of the YAP insertion has been de-
termined in a sample of 340 individuals from Europe,
Africa, the Far East, and Oceania (table 1) . The insertion
is present in 107 individuals ( 3 1%) representing all geo-
graphic regions. The frequency of the YAP element in
sub-Saharan Africa is - lo-fold higher than that in Eu-
rope, Oceania, and most regions of the Far East. When
these results are combined with other surveys (Persichetti
et al. 1992; Spurdle et al. 1994), the YAP element is
present in 509 / 1,062 individuals (48%). The YAP ele-
ment is widespread in Africa, with the highest frequency
(78%-86%) in South African Bantu-speakers and West
Africans. It is also present at high levels in smaller sam-
 Alu-Element Polymorphism on the Human Y Chromosome 755

Table 2
Average Percent Sequence Difference Between and Within Alu Subfamily Members and
Consensus Sequences

INTERMEMBER INTERCONSENSUS

PSF- 1 PSF-2 PSF-3 cs PSF- 1 PSF-2 PSF-3 CS

PSF-1.. .... (1.1\2.1) 2.8 6.3 2.7 0.7 4.6 1.8

PSF-2 ...... 1.6 (0.7\2.1) 7.3 3.6 0.7 5.4 2.5
PSF-3 ...... 4.5 5.3 (1.2\2.1) 3.8 3.6 4.3 2.8
cs ........ 2.2 2.9 2.4 (6.2) 1.8 2.5 1.8

NOTE.-Data above the diagonal include all sites; data below the diagonal exclude CpG-to-TpG/CpA changes. Data
in parentheses are average percent sequence differences within subfamilies.

ples of East Africans. The frequency is moderately high performed ( fig. 6A ) . The greatest amount of genetic dis-
in Pygmies (66%) and is at intermediate levels in the tance separates African and non-African populations.
Khoisan (46%) and in Egypt ( 53% ). The frequency of The European and Saudi populations are separated by
the YAP element is significantly lower (4%- 11% ) in Eu- small genetic distances. They form a cluster that inter-
ropeans and is absent in most Asian and Oceanian pop- venes between the Chinese, Indian, and Papua New
ulations (i.e., India, China, and Papua New Guinea). Guinea populations, on the tip of the non-African
Interestingly, the YAP element is present in 24% of the branch, and the Japanese, who lie closer to the Africans.
Japanese surveyed. A single Australian was found to have A UPGMA tree gives a similar picture (fig. 6B).
the Alu insertion. As it is not known if this is due to The African populations are divided into two subclusters,
recent admixture, a more extensive survey is needed to one with very high and the other with intermediate levels
determine if the YAP element is polymorphic in native of the YAP insertion. The non-African branch is divided
Australian populations. into three subclusters: ( 1) Asian and Oceanian popu-
The group mean frequency of African populations lations in which the insertion is absent, (2) Europeans
is significantly higher than the group means of European, and Saudis with low frequencies of the YAP insertion,
Asian, and Oceanian populations (Bonferroni P values and ( 3) Japanese with moderate levels. This pattern is
< 0.0 1). All sample frequencies are statistically equiv- consistent with other data suggesting a primary sepa-
alent within Africa, except the Khoisan frequency, which ration between Africans and non-Africans ( Wainscoat
is significantly lower (Fisher’s exact test, P < 0.000 1). et al. 1986; Cann et al. 1987; Bowcock et al. 199 1; Horai
The frequency in the eastern Pygmies (57%) is slightly et al. 199 1; Nei and Roychoudhury 1993 ) . It is unusual
lower (but not significantly) than that in the western in that the Japanese population is separated from the
Pygmies ( 7 1% ) from the Central African Republic. The other Asian and Oceanian populations.
western Pygmies, more admixed than the eastern-Pygmy
population, are probably a mixture of about three- Discussion
fourths non-Pygmy ancestry and one-fourth Pygmy an- Recent Insertion of an Alu Element on the Y
cestry (Bowcock et al. 199 1). Chromosome
The Fstvalue, calculated for the 14 populations with The Alu family member characterized here, YAP,
a sample size >20, is 0.50. This compares to an average has the hallmarks of a recently inserted Alu element:
Fstof 0.139 -t 0.0 10 for 100 DNA polymorphisms scored the flanking 1 I-nucleotide direct repeats are identical,
in five human populations ( Bowcock et al. 199 1)) with and there is a perfect run of 26 adenine residues at the
the highest Fsl being 0.49. When the frequencies from 3’ end of the element. Sequence divergence in these por-
similar populations studied by Bowcock et al. ( 199 1) tions of Alu elements is typically related to the age of
(eastern and western Pygmies, Melanesians, Chinese, the insertion (Schmid and Shen 1985). Furthermore,
and Europeans) are used to calculate Fst, the value is the Y Alu element was found to be absent from the
0.58. Therefore, the YAP represents the highest known homologous Y-chromosomal site in 14 chimpanzees and
Fstfor a single, biallelic locus. This approximately four- one gorilla (author’s unpublished data), supporting the
fold higher F,, is consistent with expectation, because hypothesis that the insertion event occurred after the
the effective population size (Ne) of Y-chromosomal divergence of humans from great apes.
DNA is about one-fourth of that for autosomal loci. The YAP element is a member of a set of seven
Pairwise Fstvalues among the same 14 populations Alu elements that constitute the subfamily here termed
were determined, and a neighbor-joining analysis was PSF-3. The recent origin of this subfamily is docu-
756 Hammer
YAP
Frequency
MI
PNG
IND
+++
I I
0.6 0.5
I
0.4 0.3 0.2
I
0.1 0 and Leeflang et al. ( 1992) favor the multiple-source-
FIG. 6.--A, Neighbor-joining tree for 14 populations. B, Den-
drogram constructed assuming constant evolutionary rates. The plus
(+) and minus (-) symbols refer to the relative frequencies of the
YAP element in each cluster. (JAP = Japan; SAU = Saudi Arabia;
ITA = Italy; UK = United Kingdom; SAR = Sardinia; SAE = South
African European; CHI = China; IND = India; PNG = Papua New
Guinea; KHO = Khoisan; EGY = Egypt; PYG = Pygmy; BAN
= Bantu; and GAM = Gambia).
mented by the low level of sequence divergence among
its members and by the polymorphism, in the human
population, of at least three of its members (YAP,
HDAB, and CHOL). The average sequence divergence
of PSF-3 (2.1%) is comparable to that of two other re-
cently formed subfamilies, PSF-1 and PSF-2.
Two research groups have identified the PSF- 1 and
PSF-2 subfamilies of Alu elements, but the nomenclature
used has varied. For example, Matera et al. ( 1990) and
Shen et al. ( 199 1) refer to PSF- 1 as the PI’ and HS
( human-specific ) subfamily, respectively. However,
Leeflang et al. ( 1992) detected a member of the PV/HS
subfamily in chimpanzee DNA, showing that this
subfamily is not specific to humans. Recently, two other
groups recognized PSF-3 Alu elements as defining a dis-
tinct subfamily, which they referred to as the Sb-2
subfamily (Hutchinson et al. 1993; Jurka 1993). It is
suggested here that the most recently formed subfamilies
containing polymorphic Alu elements be termed poly-
morphic subfamilies (PSF), to avoid problems of no-
menclature and to facilitate the naming of polymorphic
subfamilies identified in the future.
Number of Active Alu Source Genes
The existence of discernible Alu subfamilies implies
that the individual members of these different subfam-
ilies resulted from either a single source gene or a closely
related subset of source genes (Willard et al. 1987; Britten
et al. 1988; Jurka and Smith 1988; Quentin 1988; La-
buda and Striker 1989 ) . There are two competing mod-
els in the literature, with regard to the number of Alu
source genes. Shen et al. ( 199 1) support a sequential
-
lineage model that posits that there is only one tran-
scriptionally active source gene at any given evolutionary
time and that the different subfamilies result from se-
+
quence changes in the lineage of this gene. In this model,
most Alu copies are silent, but, occasionally, a retro-
position event forms an Alu copy that mutates and be-
++ comes a new source gene, forming a new subfamily. The
previous source gene must be silenced relatively soon,
or there would be parallel lines of Alu evolution. The
alternate model posits that there are multiple source
genes, deriving from independently evolving lineages,
that give rise to new subfamilies. Matera et al. ( 1990)
gene model based on the sequence diversity of recently
inserted Alus in the PSF-1 subfamily and on the poly-
morphism, in the human population, of a CS/Precise
Alu subfamily member in the Cl inhibitor locus
(Stoppa-Lyonnet et al. 1990). In contrast, Shen et al.
( 199 1) suggest that this is an ancient polymorphism re-
sulting from an insertion event that preceded the ap-
pearance of the putative single source gene for the PSF-
1/PSF-2 subfamily. This necessitates the activity of the
CS source sequence for some period of time after the
chimp/human divergence, ending with the occurrence
of the PSF-1 subfamily changes. Furthermore, the over-
lapping presence of the PSF- 1 and PSF-2 subfamilies is
explained by allelic variation in a single Alu source gene.
The existence of PSF-3, a third recently formed Alu
subfamily, poses problems for the single-source-gene
model. First, it increases the number of active source-
gene alleles that must exist contemporaneously in the
human population. But, more significantly, the topol-
ogies of the trees in figures 4 and 5 indicate that the three
PSF source genes are not on a single lineage but are
related through the CS consensus sequence. This is in-
consistent with of Shen et al.‘s ( 199 1) explanation of
the appearance of all new subfamilies by the sequential
 Alu-Element Polymorphism on the Human Y Chromosome 757

accumulation of diagnostic mutations in a single lineage history from the perspective of male lineages. This
of source genes. A completely different set of diagnostic serves to complement mitochondrial DNA studies and
sites accumulates on the lineage leading from CS to PSF- to give a more complete view of the origin and mi-
l/-2 than on the lineage leading from CS to PSF-3 gration patterns of Homo sapiens sapiens. Ultimately,
(fig. 5). Y-specific polymorphisms can be used to build trees
To rescue the model of a single source gene, one that contain information about the order of branching
must consider CS, PSF- 1, PSF-2, and PSF-3 to be alleles of Y lineages, linking ancestral types to modern types,
of a single locus. Sequence differences between the source and the approximate times at which the branching
“alleles” range from 0.7% to 4.3% (table 2 and fig. 5). events occurred. The YAP is a useful marker because
This is much greater than the average number of nu- the ancestral state, the absence of the YAP element,
cleotide differences between two randomly chosen alleles is known and because the insertion event occurred a
from the human population. For example, in their com- single time after the divergence of humans and chimps.
parison of two or more allele sequences from 49 loci, Li The insertion site is located on the long arm of the Y
and Sadler ( 199 1) found that the nucleotide diversity chromosome, linked to millions of base pairs that do
of the fastest-evolving sites is only 0.1%. They suggested not undergo meiotic recombination. Furthermore,
that the low levels of diversity are due to a small long- there is a clear pattern in the frequencies of the inser-
term effective population size for humans (i.e., - 104). tion and significant heterogeneity among human pop-
The only known case of alleles differing by as much as ulations. As yet, few other Y-linked polymorphisms
4% is the HLA locus, on which overdominant selection with these properties have been reported in the liter-
has probably maintained alleles in the population since ature (Spurdle et al. 1994).
before the human/chimp divergence (Nei and Hughes The frequency data are consistent with ‘the hy-
1991). pothesis that the YAP element originally inserted on
The divergence times for the CS and PSF source the Y chromosome of a sub-Saharan African. Because
genes/alleles can be determined using the average the frequency of the YAP element is relatively high
substitution rate in noncoding DNA, O.l5%/Myr in eastern Africans, western Africans, Pygmies,
(Miyamoto et al. 1987; Takahata 1993). The PSF-1 Bantu-speaking South Africans, and Khoisan, the
and PSF-2 sequences have been diverging for -4 Myr, insertion event is likely to predate the divergence of
and both the PSF-1 and PSF-3 sequences have been the major African groups. Depending on the exact
diverging from CS for - 10 Myr. Under a neutral timing of the Y Alu insertion event, the distribution
model, it is unlikely that source alleles could persist pattern may help to distinguish among hypotheses
for these lengths of time. On the basis of the formula on the origin and migration of human populations
4N,g, where N, is the effective population size ( 104) (fig. 7). For example, if the insertion event occurred
and g is the number of years ( 15-20) per generation at the time of H. erectus, 2 1 Mya, then the presence
along the lineage leading from apes to humans, it has
been estimated that a two-allele polymorphism is likely
to last ~0.8 Myr (Chapman et al. 1986; Takahata Time
1993). For a three-allele polymorphism, the persis- mya H. erectus H. sapiens H. erectus H. sapiens H. erectus H. sapiens
,~ ++_V_
tence time should be even shorter (Jeffreys et al. 1985). 0.0 +- +- +-
It should be pointed out that, even if the population o.5
was much larger at some point in the past, there would ,,.
be little effect on the persistence time, because N, is
the harmonic mean of the population sizes (Wright Mode,: 1 2A 28
1938 ) . In the absence of evidence for balancing selec- z~;;$;,~~: H. erectus H. sapiens H. sapiens
no complete partial
tion or some other force to maintain Alu source alleles
in the human population, the results presented here %sequence divergence
YAP+IYAP-: High Low High/Low
strongly favor the multiple-source-gene model. How- YAP+IYAP+: High/Low Low Low
High/Low Low High/Low
ever, it should be noted that the overall picture of a YAP-‘YAP-’
very small number of active source genes, put forward FIG. 7.-Three hypotheses on the timing of the Alu insertion
by the Deininger et al. ( 1992) group, is supported in event. The branching diagram illustrates the divergence of Y-chro-
mosome lineages in Homo erectus and II?. sapiens populations. The
the analysis presented here.
small intersecting line (+) refers to the Alu insertion event, and the
Human Population Studies longer intersecting line refers to the loss of H. erectus lineages. Model
1 is based on the insertion of the Alu element in H. erectus, and model
Polymorphisms on the nonrecombining portion 2 is based on the insertion occurring in the lineage leading to, but
of the Y chromosome can be used to study population before the divergence of, modern H. sapiens.
758 Hammer

of the YAP element in populations as separated as group migrated to the Indian Subcontinent and then
Africans and Japanese can be explained by migrations to Southeast Asia. In this model, the southern group
of H. erectus populations (model 1). On the other retained some genes of African origin, which even-
hand, if the insertion event occurred more recently tually were maintained in populations that moved to
in a modern sub-Saharan population, the distribution Australia and New Guinea -40,000 years ago. The
pattern would reflect population movements of H. discovery of high frequencies of the recently diverged
sapiens sapiens. One explanation, based on the hy- YAP+ chromosomes in populations along either of
pothesis of an African origin of modern humans these routes would identify a genetic trail and support
(Cann et al. 1987), posits that the Alu element was the migration hypothesis. Furthermore, the discovery
polymorphic before modern human populations mi- of the YAP element in certain populations may help
grated out of Africa and that male migrants carrying to unravel the history of the settlement of Japan. The
Y chromosomes with the YAP element (YAP+) and Japanese population is considered to be descended
without the YAP element (YAP-) completely re- from two separate populations: the Yayoi and the Jo-
placed H. erectus populations in Asia (model 2A). mon people (Horai et al. 199 1). The Yayoi, originally
An alternative form of this hypothesis, based on par- from northeastern Asia, started migrating to Japan
tial replacement, suggests that some YAP- chro- from the Korean peninsula -2,300 years ago. The
mosomes from H. erectus survive in contemporary Jomon lived in Japan for > 10,000 years and may have
populations (model 2B). Results supporting either come originally from Southeast Asia (Turner 1990).
model 1 or model 2B may be more consistent with Preliminary data suggest that the YAP element may
the hypothesis of a multiregional origin of modern have been present in Japan before the Yayoi migration
humans (Wolpoff et al. 1984). (author’s unpublished data).
These models make certain predictions that can Because the frequency of the YAP element is sig-
be tested with DNA sequences from appropriate Y nificantly lower in Khoisan than in the other African
chromosomes surveyed in human and outgroup pop- populations tested, the possibility exists that the YAP
ulations (e.g., chimpanzee). For example, the first element inserted after the divergence of the major hu-
model posits that the YAP polymorphism, arising in man groups and has been introduced into the Khoisan
H. erectus, has persisted for > 1 Myr. Therefore, YAP+ by admixture during the Bantu expansion. This pre-
chromosomes should differ from YAP- chromosomes dicts that there should be very little, if any, difference
by nearly 20% of the average sequence divergence be- either between YAP+ and YAP- chromosomes or
tween humans and chimps (if a human-chimp diver- among YAP+ chromosomes. In this scenario, the YAP
gence time of 5 Mya is assumed). In fact, in modern element could have spread from Africa to Europe ei-
humans there should be a range of sequence differences ther by means of trans-Saharan trade routes (Nagel
among YAP+ chromosomes, as well as among YAP- 1984) or by earlier migration events across the fertile
chromosomes, a range in which the highest level of Sahara 2,000- 10,000 B .C . ( Adekile 1992 ) . However,
divergence approaches that between YAP+ and YAP- recent gene flow between Africa and Japan seems less
chromosomes. On the other hand, model 2 (complete likely. Another consideration is the possibility of a
replacement) predicts a low level of sequence diver- non-African origin, which has been suggested by some
gence between YAP+ and YAP- chromosomes, as well authors (Excoffier et al. 1987). This predicts that pop-
as in “within” comparisons among YAP+ chromo- ulations with the polymorphism migrated to Africa
somes and among YAP- chromosomes. The level of some time after the original insertion event. Further
sequence divergence should be nearly an order of studies of Y-chromosome polymorphisms in human
magnitude less than the highest levels predicted in the populations are necessary to distinguish among these
first model. Model 2B (partial replacement) predicts hypotheses.
both a low level of sequence divergence among YAP+ Acknowledgments
chromosomes and a range of divergences from low to I would like to thank R. C. Lewontin and D. C.
high among YAP- chromosomes. Page for support, lab space, DNA probes, and en-
One possible consequence of the complete-re- couragement. I thank Stephen Zegura, M. Roxane
placement hypothesis is a genetic trail of recently di- Bonner, and Susanne Kaplan for comments on the
verged YAP+ chromosomes somewhere between Af- manuscript; Junhyong Kim for assistance with the
rica and Japan. According to Nei and Roychoudhury neighbor-joining analysis, and the following people
( 1993), there were two migration routes from Africa for DNA samples: A. Spurdle, A. V. S. Hill, J. S.
to the Far East. One group of people moved to north- Wainscoat, A. Novelletto, J. Kidd, J. Rogers, K. Kidd,
eastern Asia through the Middle East, and a second and M. Stoneking.
 Alu-Element Polymorphism on the Human Y Chromosome 759

APPENDIX

Alu Sequences and Their Sources

Sequence Symbol GenBank Name Sequence Source or Reference

1. ADAG Humadag5 Human adenosine deaminase gene

2. AFP . . . . Humafp Human alpha-fetoprotein gene
3. APOA . Humalapoa Matera et al. (1990)
4. APOE Humapoe4 Human apolipoprotein E gene
5. BGP Hsbgpex 1 Homo sapiens BGP gene for biliary glycoprotein
6. ClINH . . . . HumCINHA Matera et al. ( 1990)
7. C3Nl . . . . . ... Batzer et al. ( 1990)
8. C3N2 . . Hsalu003 Batzer et al. ( 1990)
9. C3N3 . . H&u004 Batzer et al. ( 1990)
10. C3N4 . Hsalu005 Batzer et al. ( 1990)
11. C3N6 . . Hsalu006 Batzer et al. ( 1990)
12. C3N7 . . . Hsalu007 Batzer et al. ( 1990)
13. C4N2 . . . . . Hsc4n2 Batzer et al. ( 1990)
14. C4N4 . . ... Batzer et al. (1990)
15. C4N5 . . ... Batzer and Deininger (199 1)
16. C4N6 . . . . ... Batzer et al. (1990)
17. C4N8 . . . . ... Batzer et al. (1990)
18. CHOL S75201 Cholinesterase {Alu element) [human insertion . . .
19. CHPV . . . . S43650 Leeflang et al. (1992)
20. Gl5N2 . Hsalu008 Batzer et al. ( 1990)
21. Gl8Nl . . Hsalu009 Batzer et al. (1990)
22. G18N2 . . . HsaulO 10 Batzer et al. ( 1990)
23. G19Nl . . . . HsaluO 11 Batzer et al. (1990)
24. GAL . . . . . ... Daniels and Deininger ( 1983)
25. GOR . GORALU Gorilla Alu-repetitive sequence in beta-globin
26. H3N1 . HsaluO 12 Batzer et al. ( 1990)
27. HD . . . . . . . S56967 Alu insertion in proximity of HD gene
28. HDAB . Humcc 1s 149 Human cosmid clone HDAB
29. LCAT ... Hslcatg Human gene for lecithin-cholesterol
30. LDHB Hsldhb 1 Human lactate dehydrogenase B gene
31. LDLR . Humldlr 18 Human low-density-lipoprotein receptor
32. MLVI . . . . Hummlvi2 Batzer et al. (1990)
33. NE53 . Humalne53 Human carcinoma cell-derived Alu
34. NEURO . . ... Wallace et al. ( 199 1)
35. PlN5 . . . . . HsaluOO1 Batzer et al. (1990)
36. PlN6 .. Hsalu002 Batzer et al. (1990)
37. PV83 . . . . . Humalpv83 Batzer et al. (1990)
38. PV92 . . . . . Humalurp7 Matera et al. (1990)
39. TPA14 . Humtpa 14 Human tissue plasminogen activator gene
40. TPA25 . . Humtpa25 Batzer et al. ( 1990)
41. VWFA . . . Humvwfaa Human von Willebrand factor gene
42. VWFB . . Humvwfab Human von Willebrand factor pseudogene

LITERATURE CITED BOWCOCK,A. M., J. R. KIDD, J. L. MOUNTAIN, J. M. HER-

ADEKILE,A. D. 1992. Anthropology of the PS gene-flow from
West Africa to North Africa, the Mediterranean, and
Southern Europe. Hemoglobin 16: 105- 12 1.
BATZER,M. A., and P. L. DEININGER. 199 1. A human-specific
subfamily of Alu sequences. Genomics 9:48 l-487.
BATZER,M. A., G. E. KILROY, P. E. RICHARD, T. H. SHAIKH,
T. D. DESSELLE,C. L. HOPPENS, and P. L. DEININGER.
1990. Structure and variability of recently inserted Alu
family members. Nucleic Acids Res. 8:6793-6798.
BERT, L. CAROTENUTO,K. K. KIDD, and L. L. CAVALLI-
SFORZA. 199 I. Drift, admixture, and selection in human
evolution: a study with DNA polymorphisms. Proc. Natl.
Acad. Sci. USA 88:839-843.
BRITTEN, R. J., W. F. BARON, D. B. STOUT, and E. H. DA-
VIDSON. 1988. Sources and evolution of human Alu re-
peated sequences. Proc. Natl. Acad. Sci. USA 85:4770-
4774.
CANN, R. L., M. STONEKING,and A. C. WILSON. 1987. Mito-
chondrial DNA and human evolution. Nature 3253 l-36.
760 Hammer
CASANOVA,J. L., C. PANNETIER, C. JAULIN, and P. KOUR-
ILSKY. 1990. Optimal conditions for directly sequencing
double-stranded PCR products with Sequenase. Nucleic
Acids Res. l&4028.
CASANOVA,M., P. LEROY, C. BOUCEKKINE,J. WEISSENBACH,
C. BISHOP,M. FELLOUS,M. PURRELLO,G. FIORI, and M.
SINISCALCO.1985.A human Y-linked DNA polymorphism
and its potential for estimating genetic and evolutionary
distance. Science 230: 1403- 1406.
CHAPMAN, B. S., K. A. VINCENT, and A. C. WILSON. 1986.
Persistence or rapid generation of DNA length polymor-
phism at the zeta-globin locus of humans. Genetics 112:
79-92.
DANIELS,G. R., and P. L. DEININGER. 1983. A second major
class of Alu family repeated DNA sequences in a primate
genome. Nucleic Acids Res. 11:7595-76 10.
-. 1985. Integration site preference of the Alu family
and similar repetitive DNA sequences. Nucleic Acids Res.
13:8939-8954.
DEININGER, P. L. 1989. SINES: short interspersed repeated
DNA elements in higher eukaryotes. Pp. 6 19-636 in D. E.
BERG and M. M. HOWES, eds. Mobile DNA. American
Society of Microbiology, Washington, D.C.
DEININGER, P. L, M. A. BATZER, C. A. HUTCHISON, and
M. H. EDGELL. 1992. Master genes in mammalian repet-
itive DNA amplification. Trends Genet. 8:307-3 11.
EXCOFFIER,L., B. PELLEGRINI,A. SANCHEZ-MAZAS, C. SIMON
and A. LANGANEY. 1987. Genetics and history of sub-Sa-
haran Africa. Yearbook Phys. Anthropol. 30: 15 1- 194.
FEINBERG,A. P., and B. VOGELSTEIN. 1984. A technique for
radiolabeling DNA restriction endonuclease fragments to
high specific activity. Anal. Biochem. 137:266-267.
FELSENSTEIN,J. 1985. Confidence limits on phylogenies: an
approach using the bootstrap. Evolution 39:783-79 1.
FENG, D.-F., and R. F. DOOLITTLE.1987. Progressive sequence
alignment as a prerequisite to correct phylogenetic trees. J.
Mol. Evol. 25:35 l-360.
HAMMER, M. F., and L. M. SILVER. 1993. Phylogenetic anal-
ysis of the alpha-globin pseudogene- (Hba-ps4) locus in
the house mouse species complex reveals a stepwise evo-
lution of t haplotypes. Mol. Biol. Evol. lo:97 1- 100 1.
HILL, A. V. S., C. E. M. ALLSOPP,D. KWIATKOWSKI,N. M.
ANSTEY,P. TWUMASI,P. A. ROWE, S. BENNETT,D. BREW-
STER, A. J. MCMICHAEL, and B. M. GREENWOOD. 199 1.
Common West African HLA antigens are associated with
protection from severe malaria. Nature 352:595-600.
HORAI, S., R. KONODO, K. MURAYAMA, S. HAYASHI, H.
KOIKE and N. NAKAI . 199 1. Phylogenetic affiliation of an-
cient and contemporary humans inferred from mitochon-
drial DNA. Philos. Trans. R. Sot. Lond. Biol. 333:409-
417.
HUTCHINSON, G. B., S. E. ANDREW, H. MCDONALD, Y. P.
GOLDBERG,R. GRAHAM,J. M. ROMMENSand M. R. HAY-
DEN. 1993. An Alu element retroposition in two families
with Huntington disease defines a new active Alu family.
Nucleic Acids Res. 21:3379-3383.
JAKUBICZKA,S., J. ARNEMANN,H. J. COOKE,M. KRAWCZAK,
and J. SCHMIDTKE. 1989. A search for restriction fragment
length polymorphism on the human Y chromosome. Hum.
Genet. 84:86-88.
JEFFREYS,A. J., V. WILSON, and S. L. THEIN. 1985. Hyper-
variable ‘minisatellite’ regions in human DNA. Nature 314:
67-73.
JURKA, J. 1993. A new subfamily of recently transposed human
Alu repeats. Nucleic Acids Res. 21:2252.
JURKA, J., and A. MILOSAVLJEVIC.199 1. Reconstruction and
analysis of human Alu genes. J. Mol. Evol. 32: 105- 12 1.
JURKA, J., and T. SMITH. 1988. A fundamental division in
the Alu family of repeated sequences. Proc. Natl. Acad. Sci.
USA 85:4775-4778.
KLUGE, A. G., and J. S. FARRIS. 1969. Quantitative phyletics
and the evolution of anurans. Syst. Zool. 18:1-32.
LABUDA, D., and G. STRIKER. 1989. Sequence conservation
in Alu evolution. Nucleic Acids Res. 17:2477-249 1.
LEEFLANG,E. P., W.-M. LIU, C. HASHIMOTO,P. V. CHOU-
DARY,and C. W. SCHMID. 1992. Phylogenetic evidence for
multiple Alu source genes. J. Mol. Evol. 35:7-16.
Ll, W.-H., and L. A. SADLER. 1991. Low nucleotide diversity
in man. Genetics 129:5 13-523.
MALASPINA,P., F. PERSICHETTI,A. NOVELLETTO,C. IODICE,
L. TERRANATO,J. WOLFE, M. FERRARO, and G. PRAN-
TERA. 1990. The human Y chromosome shows a low level
of DNA polymorphism. Ann. Hum. Genet. 54:297-305.
MATERA, G. A., U. HELLMANN, M. F. HINTZ, and C. W.
SCHMID. 1990. Recently transposed Alu repeats result from
multiple source genes. Nucleic Acids Res. 18:6019-6023.
MIYAMOTO,M. M., J. L. SLIGHTOM,and M. GOODMAN. 1987.
Phylogenetic relations of humans and African apes from
DNA sequences in the pseudo-eta-globin region. Science
238:369-373.
NAGEL, R. L. 1984. The origin of the hemoglobin S gene:
clinical, genetic, and anthropological consequences. Einstein
Q. J. Biol. Med. 2:53-62.
NEI, M., and A. HUGHES. 199 1. Polymorphism and evolution
of the major histocompatibility complex in mammals. Pp.
222-247 in R. K. SELANDER,A. G. CLARK AND T. S.
WHITTAM, eds. Evolution at the molecular level. Sinauer,
Sunderland, Mass.
NEI, M., and A. K. ROYCHOUDHURY.1993. Evolutionary re-
lationships of human populations on a global scale. Mol.
Biol. Evol. 10:927-943.
NGO, K. Y., G. VERNAUD, C. JOHNSSON,G. LUCOTTE, and
J. WEISSENBACH . 1986. A DNA probe detecting multiple
haplotypes of the human Y chromosome. Am. J. Hum.
Genet. 38:407-4 18.
PAGE, D. C., K. BIEKER, L. G. BROWN, S. HINTON, M. LEP-
PERT,J.-M. LALOUEL,M. LATHROP,M. NYSTROM-LAHTI,
A. DE LA CHAPELLE, and R. WHITE. 1987. Linkage, physical
mapping, and DNA sequence analysis of pseudoautosomal
loci on the human X and Y chromosomes. Genomics 1:
243-256.
PEARSON,W. R., and D. J. LIPMAN. 1988. Improved tools for
biological sequence analysis. Proc. Natl. Acad. Sci. USA
85:2444-2448.
PERNA, N. T., M. A BATZER, P. L. DEININGER,and M. STO-
NEKING. 1992. Alu insertion polymorphism: a new type of
marker for human population studies. Hum. Biol. 64:64 l-
648.
PERSICHETTI,F., P. BLASI, M. HAMMER, P. MALASPINA,C.
IODICE, L. TERRANATO,and A. NOVELLETTO.1992. Dis-

equilibrium of multiple markers on the human Y chro-
mosome. Ann. Hum. Genet. 56:303-310.
QUENTIN, Y. 1988. The Alu family developed through suc-
cessive waves of fixation closely connected with primate
lineage history. J. Mol. Evol. 27: 194-202.
RAMSAY, M., and T. JENKINS. 1988. Alpha-globin gene cluster
haplotypes in the Kalahari San and southern African Bantu-
speaking blacks.. Am. J. Hum. Genet. 43:527-533.
REYNOLDS,J., B. S. WEIR, and C. C. COCKERHAM.1983. Es-
timation of coancestry coefficient: basis for short-term ge-
netic distance. Genetics 105:767-779.
SAITOU,N., and M. NEI. 1987. The neighbor-joining method:
a new method for reconstructing phylogenetic trees. Mol.
Biol. Evol. 4:406-425.
SANGER, F., S. NICKLEN, and A. R. COULSON. 1977. DNA
sequencing with chain-terminating inhibitors. Proc. Natl.
Acad. Sci. USA 74:5463-5467.
SAWADA,I., C. WILLARD, C.-K. J. SHEN, B. CHAPMAN,A. C.
WILSON,and C. W. SCHMID. 1985. Evolution of Alu family
repeats since the divergence of human and chimpanzee. J.
Mol. Evol. 22:3 16-322.
SCHMID, C. W., and C.-K. J. SHEN. 1985. The evolution of
interspersed repetitive DNA sequences in mammals and
other vertebrates. Pp. 323-358 in R. J. MACINTYRE, ed.
Molecular evolutionary genetics. Plenum, New York.
SHEN, M. R., M. A. BATZER, and P. L. DEININGER. 199 1.
Evolution of the master Alu gene. J. Mol. Evol. 33:3 1 l-
320.
SLAGEL, V., E. FLEMINGTON,V. TRAINA-DORGE, H. BRAD-
SHAW,and P. DEININGER. 1987. Clustering and subfamily
relationships of the Ah family in the human genome. Mol.
Biol. Evol. 4: 19-29.
SOKAL,R. R., and C. D. MICHENER. 1958. A statistical method
for evaluating systematic relationships. Univ. Kans. Sci.
Bull. 28:1409-1438.
SOKAL, R. R., and F. J. ROHLF. 198 1. Taxonomic congruence
in the Leptopodomorpha re-examined. Syst. Zool. 30:309-
325.
SPURDLE,A. B., M. F. HAMMER, and T. JENKINS. 1994. The
Y Ah polymorphism in southern African populations and
its relationship to other Y-specific polymorphisms. Am. J.
Hum. Genet. 54:3 19-330.
SPURDLE, A., and T. JENKINS. 1992. Y chromosome probe
49a detects complex Pvu II haplotypes and many TaqI hap-
lotypes in southern African populations. Am. J. Hum. Ge-
net. 50:107-125.
STONEKING, M., and A. C. WILSON. 1989. Mitochondrial
DNA. Pp. 215-245 in A. V. S. HILL and S. SERJEANTSONS,
eds. The colonization of the Pacific: a genetic trail. Oxford
University Press, Oxford.
STOPPA-LYONNET,D., P. E. CARTER, T. MEO, and M. TOSI.
Alu-Element Polymorphism on the Human Y Chromosome 761
1990. Clusters of intragenic Alu repeats predispose the hu-
man C 1 inhibitor locus to deleterious rearrangements. Proc.
Natl. Acad. Sci. USA 87:1551-1555.
SWOFFORD,D. 1993. PAUP: phylogenetic analysis using par-
simony, version 3.1.1. Illinois Natural History Museum,
Champaign.
TAKAHATA,N. 1993. Allelic genealogy and human evolution.
Mol. Biol. Evol. 10:2-22.
TORRONI, A., 0. SEMINO,R. SCOZZARI,G. SIRUGO, G. SPE-
DINI, A. ABBAS,M. FELLOUS,and A. S. SANTACHIARABE-
NERECETTI. 1990. Y chromosome DNA polymorphisms in
human populations: differences between Caucasoids and
Africans detected by 49a and 49f probes. Ann. Hum. Genet.
54:287-296.
TURNER, C. 1990. Major features of sundadonty and sino-
donty, including suggestions about East Asian microevo-
lution, population history and late Pleistocene relationships
with Australian aboriginals. Am. J. Phys. Anthropol. 82:
295-3 17.
VOLLRATH,D., S. FOOTE, A. HILTON, L. G. BROWN,P. BEER-
ROMERO,J. S. BOGAN,and D. C. PAGE. 1992. The human
Y chromosome: a 43-interval map based on naturally oc-
curring deletions. Science 258:52-59.
WAINSCOAT,J. S., A. V. S. HILL, A. L. BOYCE, J. FLINT, M.
HERNANDEZ, S. L. THIEN, J. M. OLD, J. R. LYNCH, A. G.
FALUSI, D. J. WEATHERALL, and J. B. CLEGG. 1986, Evo-
lutionary relationships of human populations from an
analysis of nuclear DNA polymorphisms. Nature 319:49 l-
493.
WALLACE, M. R., L. B. ANDERSON, A. M. SAULINO, P. E.
GREGORY, T. W. GLOVER, and F. S. COLLINS. 199 1. A de
novo Alu insertion results in neurofibromatosis type 1. Na-
ture 353:864-866.
WILLARD, C., H. T. NGUYEN, and C. W. SCHMID. 1987. Ex-
istence of at least three distinct Alu subfamilies. J. Mol.
Biol. 26: 180- 186.
WOLPOFF,M. H., X. Z. WV and A. G. THORNE. 1984. Modem
Homo sapiens origins: a general theory of hominid evolution
involving the fossil evidence from East Asia. Pp. 4 1 l-484
in F. H. SMITH and F. SPENCER,eds. Origins of modem
humans: a world survey of the fossil evidence. A. R. Liss,
New York.
WRIGHT, S. 1938. Size of population and breeding structure
in relation to evolution. Science 87:430-43 1.
SIMON EASTEAL, reviewing editor
Received November 12, 1993
Accepted April 1, 1994