References and Notes
1. N. B. Simmons, in Mammalian Species of the World: A
Taxonomic and Geographic Reference,D.E.Wilson,
D. M. Reeder, Eds. (Johns Hopkins Univ. Press, Baltimore,
2. B. D. Patterson, M. R. Willig, R. D. Stevens, in Bat
Ecology, T. H. Kunz, M. B. Fenton, Eds. (Univ. of
Chicago Press, Chicago, 2003), pp. 536–579.
Evolution of Tertiary Mammals of North America, vol.
2, Marine Mammals and Smaller Terrestrial Mammals,
C. Janis, G. Gunnell, M. Uhen, Eds. (Cambridge Univ.
Press, Cambridge, in press).
4. N. B. Simmons, J. H. Geisler, Bull. Am. Mus. Nat. Hist.
235, 1–152 (9 March 1998).
5. J. M. Hutcheon, J. A. W. Kirsh, J. D. Pettigrew, Philos.
Trans. R. Soc. London Ser. B 353, 607 (1998).
6. E. C. Teeling et al., Nature 403, 188 (2000).
7. E. C. Teeling et al., Proc. Natl. Acad. Sci. U.S.A. 99,
8. E. C. Teeling, O. Madsen, W. J. Murphy, M. S.
Springer,S.J.O’Brien,Mol. Phylogenet. Evol. 28,
9. R. J. Baker, J. L. Longmire, M. Maltbie, M. J. Hamilton,
R. Van Den Bussche, Syst. Biol. 46, 579 (1997).
10. R. A. Van Den Bussche, S. R. Hoofer, J. Mammal. 85,
11. Materials and methods are available as supporting
material on Science Online and at the Web site of
the Laboratory of Genomic Diversity, National
12. J. E. Hill, Bull. Br. Mus. (Nat. Hist.) 32, 29 (1974).
13. P. J. J. Bates, T. Nwe, K. M. Swe, S. S. H. Bu, Acta
Chiropt. 3, 33 (2001).
14. N. B. Simmons, Am. Mus. Novit. 3077, 1 (1993).
15. S. R. Hoofer, S. A. Reeder, E. W. Hansen, R. A. Van
Den Bussche, J. Mammal. 84, 809 (2003).
16. M. S. Springer, W. J. Murphy, E. Eizirik, S. J. O’Brien,
Proc. Natl. Acad. Sci. U.S.A. 100, 1056 (2003).
17. P. Wilf, C. C. Ladanderia, Science 284, 2153 (1999).
18. C. C. Labanderia, J. Sepkoski, Science 261, 310 (1993).
19. S. L. Wing, H. Sues, in Mesozoic and Early Cenozioc
Terrestrial Ecosystems in Terrestrial Ecosystems Through
Time, A. K. Behrensmeyer et al. Ed. (Univ. of Chicago
Press, Chicago, 1992), pp. 327–418.
20. G. L. Jepsen, Science 154, 1333 (1966).
21. J. Habersetzer, G. Storch, Naturwissenschaften 79,
22. S. Hand, M. Novacek, H. Godthelp, M. Archer, J.
Vertebr. Paleontol. 14, 375 (1994).
23. G. Storch, B. Sige
, J. Habersetzer, Palaeontol. Z. 76,
24. M. C. McKenna, S. K. Bell, Classifcation of Mammals
above the Species Level (Columbia Univ. Press, New
Stanhope,W.W.deJong,Proc. Natl. Acad. Sci.
U.S.A. 98, 6241 (2001).
26. S. J. Hand, J. A. Kirsch, in Bat Biology and Conserva-
tion, T. H. Kunz, P. A. Racey, Eds. (Smithsonian Insti-
tution Press, Washington, DC, 1998), pp. 72–90.
27. J. J. Flynn, A. R. Wyss, Trends Ecol. Evol. 13,449
28. A. Cooper et al., Nature 409, 704 (2001).
29. M. O. Woodburne, J. A. Case, J. Mamm. Evol. 3, 121
30. W. J. Murphy et al., Science 294, 2348 (2001).
32. The research was funded and supported by federal
funds from the National Cancer Institute, NIH, under
contract no. N01-CO-12400, and the Darwin Initia-
tive Grant no. 162-11-09. We would like to thank K.
Jones, I. Hora
cek, I. Mackie, C. Mainstone, F. Catzeflis,
N. Czaplewski, G. Morgan, Tin Nwe, N. Crumpler,
A. Roca, and R. Stanyon for samples, technical sup-
port, and advice.
Supporting Online Material
Materials and Methods
Figs. S1 to S3
Tables S1 to S7
References and Notes
10 September 2004; accepted 24 November 2004
Interindividual Variation in
Posture Allocation: Possible Role
in Human Obesity
James A. Levine,
Lorraine M. Lanningham-Foster,
Shelly K. McCrady, Alisa C. Krizan, Leslie R. Olson, Paul H. Kane,
Michael D. Jensen, Matthew M. Clark
Obesity occurs when energy intake exceeds energy expenditure. Humans
expend energy through purposeful exercise and through changes in posture
and movement that are associated with the routines of daily life [called non-
exercise activity thermogenesis (NEAT)]. To examine NEAT’s role in obesity,
we recruited 10 lean and 10 mildly obese sedentary volunteers and measured
their body postures and movements every half-second for 10 days. Obese
individuals were seated, on average, 2 hours longer per day than lean
individuals. Posture allocation did not change when the obese individuals lost
weight or when lean individuals gained weight, suggesting that it is
biologically determined. If obese individuals adopted the NEAT-enhanced
behaviors of their lean counterparts, they might expend an additional 350
calories (kcal) per day.
Obesity is epidemic in high-income countries.
In the United States alone poor diet and
physical inactivity are associated with 400,000
deaths per year (1) and obesity-related medical
expenditures in 2003 approximated $75 billion
middle- and low-income countries, where the
health and fiscal costs are likely to be dev-
As the impact of obesity on health
escalates, so too does the need to understand
its pathogenesis. Weight gain and obesity
occur when energy intake exceeds energy
expenditure. We are interested in a specific
component of energy expenditure called
NEAT and the role it might play in human
obesity. NEAT is distinct from purposeful
exercise and includes the energy expenditure
of daily activities such as sitting, standing,
walking, and talking.
We have previously shown that when
humans overeat, activation of NEAT helps to
prevent weight gain (4). To better understand
NEAT and its role in obesity, we separated
NEAT into the thermogenesis associated with
posture (standing, sitting, and lying) and that
associated with movement (ambulation).
To investigate whether the obese state has
an effect on NEAT, we first developed and
validated a sensitive and reliable technology
for measuring the postural allocation of
NEAT in human volunteers (5, 6). This
physical activity monitoring system uses
inclinometers and triaxial accelerometers to
capture data on body position and motion
120 times each minute. By combining these
measurements with laboratory measures of
energy expenditure, we can summate NEAT
and define its components (7).
To compare body posture and body
motion in lean and obese people, we re-
cruited 20 healthy volunteers who were self-
proclaimed Bcouch potatoes.[ Ten participants
(five females and five males) were lean Ebody
mass index (BMI) 23 T 2 kg/m
^ and 10
participants (five females and five males)
were mildly obese (BMI 33 T 2kg/m
(table S1). We deliberately selected mildly
obese subjects who were not incapacitated
by their obesity and who had no joint
problems or other medical complications
of obesity. The volunteers agreed to have all
of their movements measured for 10 days
the use of a stable isotope technique (9).
They were instructed to continue their usual
daily activities and occupations and not
to adopt new exercise practices. Over the
10-day period, we collected È25 million
data points on posture and movement for
Our analysis revealed that obese partic-
ipants were seated for 164 min longer per
day than were lean participants (Fig. 1A).
Correspondingly, lean participants were
upright for 152 min longer per day than
obese participants. Sleep times (lying) were
almost identical between the groups. Total
Endocrine Research Unit, Mayo Clinic, Rochester, MN
*To whom correspondence should be addressed.
28 JANUARY 2005 VOL 307 SCIENCE www.sciencemag.org
body movement, 89% of which was ambu-
lation, was negatively correlated with fat
mass (Fig. 1, B and C). Notably, if the obese
subjects had the same posture allocation as
the lean subjects, they would have expended
an additional 352 T 65 (TSD) (range, 269 to
477) calories (kcal) per day (Fig. 1C).
To investigate whether these differences
in posture allocation are a cause or conse-
quence of obesity, we asked seven of the
original obese volunteers (four females and
three males, BMI 33 T 2 kg/m
) to undergo
supervised weight loss over a period of 8
weeks. The average weight loss was 8 kg.
Likewise, we recruited nine of the original
lean volunteers and one additional lean
volunteer (six females and four males, BMI
23 T 2 kg/m
) to undergo supervised over-
feeding over a period of 8 weeks. The av-
erage weight gain was 4 kg. After these
weight perturbations, we studied posture
allocation in these subjects for another 10
days. Interestingly, both the obese subjects
losing weight and the lean subjects gaining
weight maintained their original posture
allocation (Fig. 2). Thus, it appears that
interindividual differences in posture alloca-
tion are biologically determined.
It should be emphasized that this was a
pilot study and that the results need to be
confirmed in larger studies. Nevertheless, the
current data may be important for under-
Fig. 1. (A) Time allocation for different postures for
10 obese and 10 lean sedentary subjects. Data are
shown as mean þ SEM. Significant differences
between lean and obese are indicated: *, P 0 0.001;
**, P 0 0.0005. There were no statistically significant
differences between females (n 0 10) and males (n 0
10): Females stood 470 T 35 min/day and males
stood 429 T 40 min/day. (B)Relationshipbetween
total body movement and body fat content. Body
fat, determined from dual x-ray absorptiometry, is
expressed as a percentage (left) and mass (right)
plotted against the total 10-day accelerometer
output [accelerometer units (AU)] for 20 (10 obese
and 10 lean) sedentary subjects. The open diamonds
are data for females and the filled diamonds are data for males. There was
no significant relationship between fat-free mass and accelerometer
output (fig. S1). The relationship between NEAT by doubly labeled water
adjusted by weight versus accelerometer output is shown in fig. S2. (C)
(Left) Total daily energy expenditure and (right) NEAT in 10 obese and 10
lean sedentary subjects. The uppermost segments of the bars for obese
individuals (vertical arrows) represent the additional energy that could be
expended if these subjects were ambulatory for the same amount of time
as lean subjects. BMR, basal metabolic rate; TEF, thermic effect of food.
There was no significant difference in sleeping time between the lean
group (423 T 15 min) and the obese group (434 T 17 min). The energy
expenditure data and standard deviations appear in table S2. The rela-
tionship between NEAT measured with doubly labeled water and NEAT
measured with the instruments is shown in fig. S3.
Fig. 2. (A) Posture allocation in seven obese sedentary subjects who underwent caloric restriction
(8). (Left) Posture allocation data at baseline and after weight loss of 8 T 2 kg. (Right) The time the
subjects spent standing/ambulating at baseline is plotted against the time the subjects spent
standing/ambulating after weight loss. (B) Posture allocation in 10 lean sedentary subjects who
underwent experimental weight gain (8). (Left) The posture allocation data for baseline and after
weight gain of 4 T 2 kg. (Right) The time the subjects spent standing/ambulating at baseline is
plotted against the time the subjects spent standing/ambulating after weight gain. Data are shown
as mean þ SEM.
www.sciencemag.org SCIENCE VOL 307 28 JANUARY 2005
standing the biology of obesity and how best
to treat it. The propensity of obese persons to
sit more than lean individuals has several
potential explanations. Rodent studies sup-
port the concept that there are central and
humoral mediators of NEAT (10, 11). For
example, we have shown that a neuropeptide
associated with arousal, orexin (12), in-
creases NEAT in rats when injected into
the paraventricular nucleus (PVN) of the
hypothalamus. Preliminary data suggest that
PVN injections of orexin also cause dose-
dependent increases in standing posture
allocation in rats (13). Thus, there may be
central and humoral mediators that drive the
sedentary behavior of obese individuals. The
negative relationship between fat mass and
movement (Fig. 1B) raises the intriguing
possibility that body fat releases a factor that
slows physical activity in obesity. However,
these data also demonstrate that posture
allocation is not the mechanism by which
NEAT is modulated with short-term over-
feeding. One hypothesis is that this occurs
through modulation of energy efficiency;
this is an area worthy of future investigation.
These data may also have implications
for obesity intervention. One could argue
that obese individuals have a biologically
determined posture allocation and therefore
are destined to become obese. If this were
true, obesity would have been as common 50
years ago as it is today. However, obesity
rates have increased and continue to do so
(14). We speculate that obese and lean
individuals respond differently to the envi-
ronmental cues that promote sedentary be-
havior. If the obese volunteers adopted the
NEAT-enhanced behavior of their lean
counterparts, they could expend an addition-
al 350 kcal per day. Over a year, this alone
could result in a weight loss of È15 kg, if
energy intake remained unchanged. Herein
lies the rationale behind nationwide ap-
proaches to promote NEAT in small incre-
ments (15). For example, in Rochester,
Minnesota, in 1920 before car use was com-
monplace the average walk to and from work
was 1.6 miles (16). If walking this distance
to work were reinstituted by our obese
subjects, all of whom currently drive to
work, an extra 150 kcal per day could be
expended. We will need to use similar
measures to promote NEAT as an impetus
to create an active and dynamic environment
in which, for example, dancing supersedes
television as a leisure activity. Approaches
that succeed in getting people out of their
chairs and moving could have substantial
impact on the obesity epidemic.
References and Notes
Gerberding, JAMA 291, 1238 (2004).
2. E. A. Finkelstein, I. C. Fiebelkorn, G. Wang, Obes. Res.
12, 18 (2004).
3. World Health Organization, Obesity: Preventing and
Managing the Global Epidemic (Geneva, Switzerland,
4. J. A. Levine, N. L. Eberhardt, M. D. Jensen, Science
283, 212 (1999).
5. J. A. Levine, P. A. Baukol, K. R. Westerterp, Med. Sci.
Sports Exerc. 33, 1593 (2001).
6. J. A. Levine, E. L. Melanson, K. R. Westerterp, J. O.
Hill, Eur. J. Clin. Nutr. 57, 1176 (2003).
7. J. A. Levine, E. L. Melanson, K. R. Westerterp, J. O. Hill,
Am. J. Physiol. Endocrinol. Metab. 281, E670 (2001).
8. Materials and methods are available as supporting
material on Science Online.
9. D. A. Schoeller, C. A. Leitch, C. Brown, Am. J. Physiol.
251, R1137 (1986).
10. J. A. Levine, J. Nygren, K. R. Short, K. S. Nair, J. Appl.
Physiol. 94, 165 (2003).
11. K. Kiwaki, C. M. Kotz, C. Wang, L. Lanningham-Foster,
J. A. Levine, Am. J. Physiol. 286, E551 (2004).
12. J. G. Sutcliffe, L. de Lecea, Nature Med. 10, 673
13. C. M. Kotz, personal communication.
14. J. O. Hill, H. R. Wyatt, G. W. Reed, J. C. Peters,
Science 299, 853 (2003).
15. More information about promoting NEAT in small
increments can be found at www.smallstep.gov and
16. L. Lanningham-Foster, L. J. Nysse, J. A. Levine, Obes.
Res. 11, 1178 (2003).
17. We thank the volunteers, dietitians, food technicians,
nursing staff, and the Mass Spectrometer Core at the
General Clinical Research Center, A. Oberg for
assistance with statistics, and P. Baukol for techni-
cal support. Supported by NIH grants DK56650,
DK63226, DK66270, and M01 RR00585, by T. S. and
D. B. Ward, and by the Mayo Foundation.
Supporting Online Material
Materials and Methods
Figs. S1 to S4
Tables S1 and S2
20 October 2004; accepted 7 December 2004
Translocation by Purified FtsK
Paul J. Pease,
Gregory J. Cost,
Jerod L. Ptacin,
Nicholas R. Cozzarelli
DNA translocases are molecular motors that move rapidly along DNA using
adenosine triphosphate as the source of energy. We directly observed the
movement of purified FtsK, an Escherichia coli translocase, on single DNA
molecules. The protein moves at 5 kilobases per second and against forces up
to 60 piconewtons, and locally reverses direction without dissociation. On
three natural substrates, independent of its initial binding position, FtsK
efficiently translocates over long distances to the terminal region of the
E. coli chromosome, as it does in vivo. Our results imply that FtsK is a
bidirectional motor that changes direction in response to short, asymmetric
directing DNA sequences.
DNA translocases are adenosine triphosphate
(ATP)–driven machines required for DNA
replication, recombination, and transfer
within and between cells (1–6). FtsK is a
membrane-bound and septum-localized E.
coli translocase that coordinates cell division
with chromosome segregation (7). At times,
the product of chromosome replication is a
circular dimer rather than two monomers.
These dimers are resolved by the XerCD
site-specific recombinase at a site near the
terminus of replication, termed dif (8).
Recombination between the distant dif sites
requires FtsK. In addition, at the time of cell
division some DNA may remain in the septal
region, and FtsK appears to act as a pump to
clear this region of DNA (9, 10). FtsK may
also promote disentanglement of chromo-
some termini by means of an interaction with
topoisomerase IV (11).
Translocases move along DNA or move
the DNA if the translocase is anchored. In
either case, translocases must be able to move
in a specific direction relative to the DNA.
Translocase directionality could be determined
by strand polarity for single-strand tracking
enzymes, nonrandom orientation of trans-
locase binding through localized accessory
factors or binding sites, or DNA sequences
that affect the enzyme during translocation. In
E. coli, provocative genetic experiments
showed that lambda-phage DNA inserted
near dif disrupted chromosome dimer reso-
lution in a manner dependent on insertion
orientation (9). Thus, DNA sequences could
be directing FtsK toward dif so that it can
activate XerCD recombination.
Department of Molecular and Cell Biology,
Department of Physics,
Hughes Medical Institute, University of California,
Berkeley, CA 94720–3204, USA.
Division of Molecular
Genetics, Department of Biochemistry, University of
Oxford, OX1 3QU, UK.
*These authors contributed equally to this paper.
.To whom correspondence should be addressed.
E-mail: firstname.lastname@example.org (N.R.C.);
28 JANUARY 2005 VOL 307 SCIENCE www.sciencemag.org