DOI: 10.1126/science.280.5362.427, 427 (1998);280 Science

, et al.Ronald S. VeazeyViral Replication in SIV Infection

T Cell Depletion and+Gastrointestinal Tract as a Major Site of CD4

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): February 23, 2011 www.sciencemag.org (this infomation is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/280/5362/427.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/280/5362/427.full.html#ref-list-1, 7 of which can be accessed free:cites 15 articlesThis article

570 article(s) on the ISI Web of Sciencecited by This article has been

http://www.sciencemag.org/content/280/5362/427.full.html#related-urls100 articles hosted by HighWire Press; see:cited by This article has been

http://www.sciencemag.org/cgi/collection/medicineMedicine, Diseases

subject collections:This article appears in the following

registered trademark of AAAS. is aScience1998 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Feb

ruar

y 23

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

photocarriers are generated in the top andbottom cells. However, when the photocur-rent-matched configuration is placed in anaqueous environment, the system eitherproduces no current in the external circuitor decomposes (27). A thicker, top p-typelayer and the resultant mismatch of elec-tron-hole formation in the two-junctionregion appear to be the key for properoperation.

REFERENCES AND NOTES___________________________

1. A. J. Bard and M. A. Fox, Acc. Chem. Res. 28, 141(1995).

2. H. Gerischer, in Topics in Applied Physics, vol. 31,Solar Energy Conversion, B. O. Seraphin, Ed.(Springer-Verlag, Berlin, 1979), pp. 115–172.

3. The ability of a semiconductor electrode to drive theelectrochemical reaction of interest is determined byits band gap (the energy separation between thevalence and conduction band edges) and the posi-tion of the valence and conduction band edges rela-tive to the vacuum level (or other reference elec-trode). In contrast to metal electrodes, semiconduc-tor electrodes in contact with liquid electrolytes havefixed energies where the charge carriers enter thesolution. This fixed energy is given by the energeticposition of the semiconductor’s valence and con-duction bands at the surface (where these bandsterminate at the semiconductor/electrolyte inter-face). The energetic position of these band edges isdetermined by the chemistry of the semiconductor/electrolyte interface, which is controlled by the com-position of the semiconductor, the nature of the sur-face, and the electrolyte composition. So eventhough a semiconductor electrode may generatesufficient energy to effect an electrochemical reac-tion, the energetic position of the band edges mayprevent it from doing so. For spontaneous watersplitting, the oxygen and hydrogen reactions must liebetween the valence and conduction band edges,and this is almost never the case.

4. G. E. Shakhnazaryan et al., Russ. J. Electrochem.30, 610 (1994).

5. E. Aharon-Shalom and A. Heller, J. Electrochem.Soc. 129, 2865 (1982).

6. A. Heller, Science 223, 1141 (1984).7. S. Kocha, J. Turner, A. J. Nozik, J. Electroanal.

Chem. 367, 27 (1991).8. S. Kocha and J. A. Turner, J. Electrochem. Soc. 142,

2625 (1995).9. H. Yoneyama, H. Sakamoto, H. Tumura, Electro-

chem. Acta 20, 341 (1975).10. A. J. Nozik, Appl. Phys. Lett. 29, 150 (1976).11. R. C. Kainthla, B. Zelenay, J. O’M. Bockris, J. Elec-

trochem. Soc. 134, 841 (1987 ).12. Y. Sakai et al., Can. J. Chem. 66, 1853 (1988).13. G. Lin et al., Appl. Phys. Lett. 55, 386 (1989).14. J. Augustynski, G. Calzaferri, J. Courvoisier, M.

Graetzel, in Proceedings of the 11th World HydrogenEnergy Conference, T. N. Veziroglu, C.-J. Winter,J. P. Baselt, G. Kreysa, Eds. (DECHEMA, Frankfurt,Germany, 1996), pp. 2378–2387.

15. K. A. Bertness et al., App. Phys. Lett. 65, 989 (1994).16. J. R. Bolton et al., Nature 316, 495 (1985).17. S. R. Kurtz et al., J. Appl. Phys. 68, 1890 (1990).18. K. A. Bertness et al., in AIP Conference Proceedings

No. 306, R. Noufi and H. Ullal, Eds. (American Insti-tute of Physics, New York, 1994), p. 100.

19. O. Khaselev and J. A. Turner, in preparation.20. For electrochemical measurements, the wafers were

cleaved into samples ;0.2 cm2 in area. The sampleswith gold ohmic contacts were mounted on a Teflon-covered screw electrode with silver epoxy and heat-ed to 80°C for 1 hour. The electrical contact wasinsulated from the electrolyte by an epoxy coatingthat also covered the samples’ edges. A convention-al two-electrode configuration was used with a plat-inum gauze or foil counter-electrode coated with ru-thenium metal. Photoelectrochemical characteris-

tics were measured with an EG&G 263A poten-tiostat. The electrolyte, 3 M H2SO4, was freshlyprepared from deionized water having resistivity of18 megohm/cm. All solutions were made of analyti-cal-grade reagents. Before electrochemical mea-surements, the samples were etched in 1:20 :1 HCl:CH3COOH:H2O2 solution. A single-compartmentcell was used. Although the electrodes were spatiallyseparated, no attempt was made to separate theproducts of the anode and cathode reactions.

21. To reduce the overvoltage losses associated withthe noncatalytic surface of the semiconductor, a thinlayer of platinum catalyst was electrochemically de-posited on the surface of the semiconductor elec-trodes from a 20 mM H2PtCl6 solution. Photoas-sisted galvanostatic deposition was performed at acathodic current density of 1 mA/cm2, with a plati-num quantity corresponding to a charge of 10 mC/cm2. A recent publication (28) has shown that theplatinum not only acts as a catalyst for hydrogenevolution but also drastically reduces the corrosionreaction of a similar III-V compound, indium phos-phide. We would expect that platinum would offer acomparable corrosion-inhibiting mechanism here.

22. We used concentrated light for two reasons: (i) Thephotocurrent and the amount of gases produced arehigher and therefore easier to quantify, and (ii) be-cause of their cost, the major terrestrial application ofthe solid-state analog of these tandem cells is in con-centrator systems (.100 suns). Therefore, any realapplication using the present solid-state design forphotoelectrolysis would also use concentrated light.However, the PEC systems will probably be limited toless than 100 suns, because that would give rise to acurrent of about 1A/cm2, a practical limit for electrol-ysis. The evolution of gas bubbles on the illuminatedsurface is also an issue, as these bubbles will scatterlight, reducing the efficiency. Given this limitation aswell as problems regarding the amount of catalyst and

system design issues, a more realistic limit as to theamount of light concentration that can be used forphotoelectrolysis is in the range of 10 to 20 suns.

23. The gas products of the photoelectrolysis were an-alyzed with a UTI-100C mass spectrometer. Asealed cell was used, connected directly to the high-vacuum chamber of the mass spectrometer via agas inlet system. Before being sealed, the cell waspurged for several minutes with nitrogen to removeany oxygen. The cell was then sealed, and the pho-toelectrolysis reaction was run for several hours. Theexternal current was monitored with an ammeter.Periodically, the gas mixture from the upper part ofthe cell was sampled by the mass spectrometer sys-tem. The same cell was used and the same experi-mental technique was repeated with the use of twoplatinum electrodes at the same external current.Within experimental error (;620%), the results werethe same, indicating stoichiometric water splitting.

24. For 100% electrolysis efficiency, this calculation isidentical to Bolton’s calculation of efficiency (29). Forthe size of these samples and the amount of hydro-gen and oxygen being generated, a calculation usingexternal current flow is far more accurate.

25. M. F. Weber and M. J. Dignam, Int. J. HydrogenEnergy 11, 225 (1986).

26. iiii, J. Electrochem. Soc. 131, 1258 (1984).27. S. Kocha, D. Montgomery, M. Peterson, J. A. Turn-

er, Sol. Energy Mater. Sol. Cells, in press.28. P. Bogdanoff, P. Friebe, N. Alonso-Vante, J. Electro-

chem. Soc. 145, 576 (1998).29. J. R. Bolton, Solar Energy 57, 37 (1996).30. We thank S. Kurtz and J. Olson for the tandem cell

samples and A. Dillon for help with the mass spec-trometer measurements. This work was supportedby the Hydrogen Program of the U.S. Department ofEnergy.

18 November 1997; accepted 23 February 1998

Gastrointestinal Tract as a Major Site ofCD41 T Cell Depletion and Viral Replication

in SIV InfectionRonald S. Veazey, MaryAnn DeMaria, Laura V. Chalifoux,Daniel E. Shvetz, Douglas R. Pauley, Heather L. Knight,

Michael Rosenzweig, R. Paul Johnson, Ronald C. Desrosiers,Andrew A. Lackner*

Human and simian immunodeficiency virus (HIV and SIV ) replicate optimally in activatedmemory CD41 T cells, a cell type that is abundant in the intestine. SIV infection of rhesusmonkeys resulted in profound and selective depletion of CD41 T cells in the intestinewithin days of infection, before any such changes in peripheral lymphoid tissues. The lossof CD41 T cells in the intestine occurred coincident with productive infection of largenumbers of mononuclear cells at this site. The intestine appears to be a major target forSIV replication and the major site of CD41 T cell loss in early SIV infection.

It is now thought that ongoing HIV rep-lication results in a continual loss ofCD41 T lymphocytes that is nearly bal-anced by the production of new CD41 Tlymphocytes (1). This model explainssome of the puzzles of HIV infection, butthe events that occur in the initial stageof infection remain largely unexplored.Although it is clear that HIV targets lym-phoid tissue, nearly all studies in this areahave focused on peripheral blood andlymph nodes. These studies overlook the

fact that the gastrointestinal tract con-tains most of the lymphoid tissue in thebody (2, 3). Furthermore, it is likely thatthe behavior of HIV in the unique immu-nologic environment of the intestinalmucosa differs from that observed in theperiphery.

The gut-associated lymphoid tissue(GALT) consists of organized lymphoid tis-sue (Peyer’s patches and solitary lymphoidfollicles) as well as large numbers of acti-vated memory T lymphocytes diffusely dis-

REPORTS

www.sciencemag.org z SCIENCE z VOL. 280 z 17 APRIL 1998 427

on

Feb

ruar

y 23

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

tributed throughout both the intestinallamina propria and epithelium. The propor-tion of activated memory CD41 T cells ismuch greater in the intestinal lamina pro-pria than in peripheral blood or lymphnodes (2, 4–6). Because HIV replicatesmost efficiently in activated memory CD41

T cells (7), the intestinal tract may be thepreferred target for initial infection and rep-lication of HIV. The effects of primary HIVinfection on GALT are difficult to assess inhumans, largely because of the difficultiesinvolved in examining the intestinal tractwithin days of infection. The SIV macaquemodel provides a useful tool for examiningthe pathogenesis of acquired immunodefi-ciency syndrome and the early events ininfection (8).

Groups of macaques were inoculatedintravenously with the following molecu-lar clones of SIV (9): pathogenic SIV-mac239 (10); a macrophage-competentderivative of SIVmac239, designated SIV-mac239/316 (11); and an attenuated, nef-deletion derivative of SIVmac239, desig-nated SIVmac239Dnef (12). Two animalsin each group were killed at 7, 14, 21, and50 days after inoculation. Lymphocyteswere separately harvested from the intes-tinal epithelium and lamina propria aswell as from the blood, spleen, and bothmesenteric and axillary lymph nodes ofeach animal (13). Although intestinal in-traepithelial lymphocytes (IELs) and lam-ina propria lymphocytes (LPLs) were col-lected and analyzed separately, only minorchanges were detected in the IEL popula-tion (14); thus, only the LPL data arepresented here. Lymphocytes from all tis-sues were examined for expression of CD3(a pan–T cell marker), CD4 (T helpercells), CD8 (cytotoxic/suppressor cells),CD25 (interleukin-2 receptor, activated Tcells), and CD45RA (naı̈ve T cells) bymultiparameter flow cytometry (13). Tolocalize virus and examine the numbersand phenotypes of infected cells, we per-formed in situ hybridization for SIV andimmunohistochemistry on intestine andperipheral lymphoid tissues (15).

Infection with pathogenic SIV (SIV-mac239 or SIVmac239/316) consistentlyresulted in rapid and profound depletionof CD41 cells exclusively in the laminapropria of the intestinal tract (jejunum,ileum, and colon) (Fig. 1A). These de-clines were evident by 7 days after infec-tion, as compared to controls or animalsinfected with SIVmac239Dnef (P , 0.01)(16). The nadir of intestinal CD4 deple-

tion was reached by 14 to 21 days afterinfection in SIVmac239-infected animalsbut was delayed to 50 days in SIVmac239/316-infected animals.

Immunohistochemistry on intestinalsegments confirmed the CD41 T cell de-pletion in the lamina propria and to a lesserextent in the Peyer’s patches and solitarylymphoid nodules (14). Immunohistochem-istry and quantitative analysis of CD3 didnot reveal any significant change in thetotal number of T cells in the intestine(14). These results, in conjunction with theobserved increase in the percentage ofCD81 T cells (Fig. 1), indicate that theCD41 T cell depletion was accompanied byan increase in absolute numbers of CD81 Tcells.

In marked contrast, there were minimalchanges in the percentages of CD41 lym-phocytes in the blood, spleen, and lymphnodes from these same animals at the sametime points. Slight decreases in CD41 cellswere observed in the axillary and mesenter-

ic lymph nodes of animals infected withpathogenic clones of SIV, but these differ-ences were not significant (Fig. 1). Percent-ages of CD41 T cells in the blood wereincreased at days 7 and 14 after infection,but the absolute numbers were not signifi-cantly altered at these time points.

To rule out the possibility of CD4 down-modulation or gp120-mediated CD4 mask-ing in the intestinal CD4 depletion, weperformed three-color analysis with anti-bodies to CD3, CD4, and CD8. No signif-icant increase in the proportion ofCD31CD4–CD8– (double-negative) T cellswas observed, which would have beenexpected if either CD4 masking or down-modulation were involved (Fig. 2). Thisstrongly suggests that the CD41 T cellswere eliminated from the intestinal tractby lytic or apoptotic mechanisms or, alter-natively, by altered trafficking of mucosallymphocytes.

To determine whether these resultswere unique to molecular clones of SIV-

New England Regional Primate Research Center, Har-vard Medical School, 1 Pine Hill Drive, Post Office Box9102, Southborough, MA 01772, USA.

*To whom correspondence should be addressed. E-mail:alackner@warren.med.harvard.edu

Fig. 1. Sequential changes in CD41 and CD81 T cell populations in lymphoid tissues of macaques 7, 14,21, and 50 days after infection with molecular clones of SIV. (A) Intestinal lamina propria; (B) lymphnodes, spleen, and blood. Each bar represents the mean of two (infected) or four (uninfected control)animals.

SCIENCE z VOL. 280 z 17 APRIL 1998 z www.sciencemag.org428

on

Feb

ruar

y 23

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

mac239, we inoculated two additional an-imals with uncloned SIVmac251. Whenthese animals were killed at 14 days afterinfection, they also had marked CD41 Tcell depletion in the intestinal lamina pro-pria (CD41 T cell percentages decreasedto 8% or less in the jejunum and colon and15% or less in the ileum) without signifi-cant changes in peripheral lymphoid tis-sues (14). The speed and extent of theCD41 T cell depletion in these animalswas indistinguishable from what was ob-served in animals infected with SIV-

mac239 (Fig. 1A). An additional threeanimals inoculated with SIVmac251 werekilled 5 months after infection to deter-mine whether the CD41 cells returnedafter the acute phase of infection. CD41 Tcells were markedly decreased in the in-testine of these animals as well (to lessthan 10% of LPLs) (14). Thus, intestinalCD41 T cell depletion is consistent fordifferent pathogenic strains of SIV andappears to persist throughout the course ofinfection. Selective depletion of CD41 Tcells from the intestine also appears to

occur in HIV-infected humans. Severalstudies have suggested that CD41 T celldepletion is more pronounced or occurssooner in the intestine than in peripheralblood, even in the relatively early stages(first several months) of HIV infection (5,17, 18).

In contrast to pathogenic SIV, infectionwith attenuated virus (SIVmac239Dnef) didnot result in a significant loss of intestinalCD41 cells (Fig. 1) but was associated withmarked T cell activation, as determined byup-regulation of CD25 (interleukin-2 re-ceptor). Marked increases in CD25 coex-pression were detected in both CD41 andCD81 T cell subsets in SIVmac239Dnef-infected animals (Fig. 3).

The virulence of the molecular clonesused here correlated with the rapidity anddegree of intestinal CD41 T cell deple-tion, and correlated inversely with CD25expression on the remaining CD41 cells.Infection with SIVmac239 resulted in rap-id and profound CD41 depletion in theintestine, and it decreased the relativeexpression of CD25 on the remainingCD41 intestinal lymphocytes. Infectionwith SIVmac239Dnef was associated withmarked up-regulation of CD25 on intesti-nal CD41 T cells and to a lesser extent onCD81 cells (Figs. 3 and 4). The effects ofinfection with SIVmac239/316 on CD25expression were intermediate between theothers, with the highest expression 21 daysafter infection. Infection with SIV-mac239/316 resulted in more CD25 ex-pression on CD81 cells and on the re-maining CD41 T cells than did infectionwith SIVmac239 (Fig. 4). Expression ofCD25 was consistently higher in the jeju-num of both normal and SIV-infected ma-caques than in the ileum or colon. Fur-thermore, expression of CD25 was consis-tently higher on CD41 cells than onCD81 lymphocytes (Fig. 4). Althoughmore organized lymphoid tissue is presentin the ileum, these data suggest that jeju-nal lymphocytes (which are mostly LPLs)are more activated than ileal lymphocytes,which contain many organized lymphoidnodules containing fewer activated cells.This is consistent with the recent findingthat CD4 depletion in the intestine ofHIV-infected patients selectively occursin the lamina propria rather than in theorganized lymphoid tissue of the intestine(18).

Our findings also confirm that the nor-mal intestine contains larger numbers ofactivated memory CD41 T cells than doperipheral lymphoid tissues. The expres-sion of CD25 is consistently higher, andthe expression of CD45RA (naı̈ve cells) isconsistently lower, on intestinal lympho-cytes than on lymphocytes obtained from

Fig. 2. Three-color flow cytometry dot plots comparing lymphocytes isolated from the intestinal laminapropria ( jejunum LPL) with those from the axillary lymph node (LN) from a normal uninfected macaqueand from macaques infected with SIVmac239 at 7 and 21 days after infection (pi). Each set of top andbottom panels corresponds to the same animal. Plots were generated by first gating through lympho-cytes and then through CD31 T cells.

Fig. 3. Flow cytometry dot plots comparing CD25 expression on intestinal LPLs from an uninfectedmacaque and from macaques infected with SIVmac239 or SIVmac239Dnef. Plots were generated bygating on lymphocytes.

REPORTS

www.sciencemag.org z SCIENCE z VOL. 280 z 17 APRIL 1998 429

on

Feb

ruar

y 23

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

the peripheral lymphoid tissues or mesen-teric lymph nodes (2, 4–6). Although asuitable memory cell marker has not beenfound that cross-reacts with the rhesusmacaque, these data confirm that the in-testinal mucosal lymphoid tissue of rhesusmacaques is similar in composition to thatof humans. Moreover, these data indicatethat the major target cells of HIV and SIVare far more abundant in the intestinaltract than in peripheral tissues.

Further insights into the early patho-genesis of SIV were gained by examiningthe distribution and phenotype of SIV-infected cells in the intestine by in situhybridization, alone and combined withimmunohistochemistry, at sequential timepoints after SIVmac239 infection. Coin-ciding with the onset of CD4 depletion (7days after infection), many SIV-infectedlymphocytes were present in the intestinaltract of animals infected with pathogenicclones of SIV (Fig. 5A). Also, there weremore virus-infected cells in the intestinethan in the peripheral lymphoid tissues, aspreviously described (19). Initially, manyinfected cells were present throughout theintestinal lamina propria as well as in theT-dependent areas of organized lymphoidnodules (Peyer’s patches and solitary lym-phoid follicles). At later time points (21and 50 days after infection) when CD41 Tcell depletion was marked, the number ofvirus-infected cells was greatly diminishedand the remaining infected cells weremainly limited to the organized lymphoidnodules. Conversely, the percentage of in-

fected macrophages increased at later timepoints. Although infected macrophageswere rare in animals at day 7 after infec-tion, combined immunohistochemistry formacrophages (HAM-56) and in situ hy-bridization for SIV demonstrated an in-crease in the percentage of infected mac-rophages at later time points, which weremainly located within organized lymphoidnodules (Fig. 5B). This is consistent withthe results of previous experiments usinguncloned SIVmac251 (20).

Combined, the in situ hybridizationand flow cytometry data suggest the fol-lowing course of events: Initially, largenumbers of activated memory CD41 Tcells are constitutively present throughoutthe intestinal lamina propria. These initialtarget cells are rapidly infected and serveas sources for viral amplification and dis-semination. However, this large pool ofactivated memory CD41 T cells in theintestine is rapidly depleted, leaving onlynaı̈ve lymphocytes and macrophages toserve as viral host cells. The decrease inthe pool of susceptible cells results in adecline in viral load in the tissue. Howev-er, new lymphocytes are continually re-cruited to or produced in the organizedlymphoid tissues (GALT and lymphnodes), and these may become activatedby antigenic stimulation to serve as freshhost cells for viral replication, thus perpet-uating the infection. This would explainwhy the numbers of virus-infected lym-phocytes are reduced in GALT at the latertime points of infection.

These observations strongly suggestthat intestinal lymphoid tissue is a crucialtarget organ in the initial pathogenesis ofSIV infection. On the basis of this evi-dence, we hypothesize that the gastroin-testinal tract, and not the peripheral lym-phoid tissue, is the major site of early SIVand HIV replication and amplification,resulting in profound and rapid CD41 Tcell loss. If so, then the acute phase ofinfection with SIV or HIV should beviewed primarily as a disease of the muco-sal immune system. There are importantramifications of considering SIV/HIV as amucosal disease, including the design oftherapies that target the intestinal tractand vaccines that stimulate an effectivemucosal immune response. These resultssuggest that an important advantage of amodified live virus vaccine that targetsthe intestinal mucosa would be its ability

Fig. 4. CD25 expression by CD41 and CD81 lymphocytes in the intestinal lamina propria of macaques7, 14, 21, and 50 days after infection, as measured by flow cytometry. Data were generated by firstgating on CD41 or CD81 lymphocytes. Each bar represents the mean of two (infected) or four (control)animals.

Fig. 5. Localization of SIV-infected cells in theintestine of macaques. (A) In situ hybridization forSIV in the jejunum showing large numbers of in-fected lymphocytes (black) in the lamina propria 7days after infection with SIVmac239. Scale bar,100 mm. (B) Intestinal lymphoid nodule double-labeled by in situ hybridization for SIV (black) andimmunohistochemistry (brown) for HAM-56 (mac-rophages) in a macaque 21 days after infectionwith SIVmac239. Note the presence of severalSIV-infected macrophages having large black nu-clei and brown cytoplasm (arrowheads). A fewSIV-infected lymphocytes are also visible in thisfield (arrows). Scale bar, 50 mm.

SCIENCE z VOL. 280 z 17 APRIL 1998 z www.sciencemag.org430

on

Feb

ruar

y 23

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

to stimulate an appropriate immune re-sponse at the site involved in the earlieststages of viral infection.

REFERENCES AND NOTES___________________________

1. D. D. Ho et al., Nature 373, 123 (1995); X. Wei et al.,ibid., p. 117; A. Amadori, R. Zamarchi, L. Chieco-Bianchi, Immunol. Today 17, 414 (1996).

2. T. T. MacDonald and J. Spencer, in Gastrointestinaland Hepatic Immunology, R. H. Heatley, Ed. (Cam-bridge Univ. Press, Cambridge, 1994), pp. 1–23.

3. N. Cerf-Bensussan and D. Guy-Grand, Gastroen-terol. Clin. N. Am. 20, 549 (1991); J.-P. Kraehenbuhland M. R. Neutra, Physiol. Rev. 72, 853 (1992).

4. M. Zeitz, W. C. Greene, N. J. Peffer, S. P. James,Gastroenterology 94, 647 (1988); H. L. Schiefer-decker, R. Ullrich, H. Hirseland, M. Zeitz, J. Immunol.149, 2816 (1992).

5. T. Schneider et al., Clin. Exp. Immunol. 94, 430(1994).

6. R. S. Veazey et al., Clin. Immunol. Immunopathol.82, 230 (1997).

7. J. S. McDougal et al., J. Immunol. 135, 3151 (1985);G. Nabel and D. Baltimore, Nature 326, 711 (1987);C. J. Van Noesel et al., J. Clin. Invest. 86 (1990);S. M. Schnittman et al., Proc. Natl. Acad. Sci. U.S.A.87, 6058 (1990); Z. F. Rosenberg and A. S. Fauci,FASEB J. 5, 2382 (1991).

8. R. C. Desrosiers and D. J. Ringler, Intervirology 30,301 (1989); R. C. Desrosiers, Annu. Rev. Immunol. 8,557 (1990); K. A. Reimann et al., J. Virol. 68, 2362(1994); A. A. Lackner, Curr. Top. Microbiol. Immunol.188, 35 (1994).

9. In separate experiments, groups (n 5 8) of juvenile(age 1 to 1.5 years) male rhesus macaques wereintravenously inoculated with equal doses (50 ng ofp27) of molecularly cloned SIVmac239, SIVmac239/316 (a weakly pathogenic, macrophage-competentmolecular clone derived from SIVmac239), orSIVmac239Dnef (an attenuated virus derived fromSIVmac239 by deleting a large portion of the nefgene). Four age- and sex-matched, uninfected ma-caques were used as controls. All animals weremaintained in accordance with the standards of theAmerican Association for Accreditation of LaboratoryAnimal Care and the guidelines of the Committee onAnimals of Harvard Medical School.

10. H. Kestler et al., Science 248, 1109 (1990).11. K. Mori, D. J. Ringler, T. Kodama, R. C. Desrosiers,

J. Virol. 66, 2067 (1992).12. H. W. Kestler et al., Cell 65, 651 (1991).13. Lymphocytes were isolated from the intestinal epi-

thelium, lamina propria, spleen, lymph nodes, andblood and stained for flow cytometric analysis asdescribed (6). Samples were stained with monoclo-nal antibodies (mAbs) to human CD4 (Ortho Diag-nostics), CD8, CD25, CD45RA (Becton Dickinson),and rhesus CD3 [T. Kawai et al., Transplant. Proc.26, 1845 (1994)] directly conjugated to fluoresceinisothiocyanate, phycoerythrin, peridinin chlorophyllprotein, or allophycocyanin. Samples were analyzedusing either a FACScan or Vantage flow cytometerand Cell Quest software (Becton Dickinson).

14. R. S. Veazey et al., data not shown.15. In situ hybridization for SIV and immunohistochem-

istry for macrophages (HAM-56, Dako), CD3, andCD4 were performed on intestinal tissues, peripherallymph nodes, and spleen as described [C. J. Horvathet al., J. Leukocyte Biol. 53, 532 (1993); P. Ilyinskii etal., J. Virol. 68, 5933 (1994); J. H. Lane et al., Am. J.Pathol. 149, 1097 (1996)]. Absolute numbers ofCD31 T cells were quantitated per square microme-ter of intestinal lamina propria by computer-assistedimage analysis, as described (6). Only LPLs werequantitated because of the technical difficulties in-volved in discriminating positive cells in Peyer’spatches and lymphoid follicles.

16. Statistical analyses were accomplished by meansof a Kruskal-Wallis one-way analysis of varianceusing commercial computer software (Sigma Plot,Jandel). Analyses were performed by comparingthe percentages of CD41 cells of all lamina propria

samples (three per animal; jejunum, ileum, and co-lon) from uninfected normal (n 5 12), SIVmac239-infected (n 5 6), and SIVmac239Dnef-infected (n 56) macaques.

17. S. G. Lim et al., Clin. Exp. Immunol. 92, 448 (1993);M. Schrappe-Bacher et al., J. Acquired ImmuneDefic. Syndr. Hum. Retrovirol. 3, 238 (1990); T.Schneider et al., Gut 37, 524 (1995); V. D. Rodgers,R. Fassett, M. F. Kagnoff, Gastroenterology 90, 552(1986); F. Clayton et al., ibid. 103, 919 (1992).

18. F. Clayton, G. Snow, S. Reka, D. P. Kotler, Clin. Exp.Immunol. 107, 288 (1997).

19. V. G. Sasseville et al., Am. J. Pathol. 149, 163 (1996).

20. C. Heise, P. Vogel, C. J. Miller, C. H. Halsted, S.Dandekar, ibid. 142, 1759 (1993).

21. We thank K. Toohey and A. Hampson for photo-graphic assistance, J. Wong for rhesus CD3 mAb,and P. Sehgal and the animal care staff at the NewEngland Regional Primate Research Center for theirexcellent care of the animals. Supported by NIHgrants DK50550, AI38559, AI25328, RR07000, andRR00168. A.A.L. is the recipient of an Elizabeth Gla-ser Scientist Award.

7 November 1997; accepted 20 February 1998

Association of the AP-3 Adaptor Complexwith Clathrin

Esteban C. Dell’Angelica, Judith Klumperman,Willem Stoorvogel, Juan S. Bonifacino*

A heterotetrameric complex termed AP-3 is involved in signal-mediated protein sortingto endosomal-lysosomal organelles. AP-3 has been proposed to be a component of anonclathrin coat. In vitro binding assays showed that mammalian AP-3 did associate withclathrin by interaction of the appendage domain of its b3 subunit with the amino-terminaldomain of the clathrin heavy chain. The b3 appendage domain contained a conservedconsensus motif for clathrin binding. AP-3 colocalized with clathrin in cells as observedby immunofluorescence and immunoelectron microscopy. Thus, AP-3 function in proteinsorting may depend on clathrin.

The formation of vesicles for transport be-tween organelles of the endocytic and se-cretory pathways and the selection of cargofor packaging into those vesicles are medi-ated by protein coats associated with thecytosolic face of the organelles (1). The bestcharacterized coats contain clathrin andprotein complexes termed “adaptors” (2).Clathrin is a complex of three heavy chainsand three light chains that polymerizes toform the scaffold of the coats. The adaptorsmediate attachment of clathrin to mem-branes and recruit integral membrane pro-teins to the coats. Two clathrin adaptorshave been described to date—AP-1 andAP-2, which participate in protein trans-port to the endosomal-lysosomal systemfrom the trans-Golgi network (TGN) andthe plasma membrane, respectively.

Recently, another complex related toAP-1 and AP-2 has been identified. Thiscomplex, termed AP-3, is composed of twolarge subunits (d and b3A or b3B), a me-dium-sized subunit (m3A or m3B), and asmall subunit (s3A or s3B) (3–8). AP-3has been localized to the TGN and endo-somes (4, 5) and is involved in protein

trafficking to lysosomes or specialized endo-somal-lysosomal organelles such as pigmentgranules, melanosomes, and platelet densegranules (8, 9). Previous studies suggestedthat AP-3 is a component of a nonclathrincoat (3, 4). However, a critical question hasremained unanswered: does AP-3 assemblewith a structural coat protein that plays therole of clathrin?

We used glutathione S-transferase(GST) fusion proteins bearing the H(hinge) and C (COOH-terminal) segmentsof the “appendage” region of human b3A(6) to search for this putative protein (Fig.1A). The reason for selection of these con-structs was that both AP-1 and AP-2 bindclathrin via the appendage regions of theb1 and b2 subunits (10, 11). Affinity puri-fication from a cytosolic extract of a humanT cell line, Jurkat, resulted in isolation of aprominent 180-kD protein on GST-b3ACbut not GST-b3AH or GST columns (Fig.1B). Unexpectedly, microsequencing ofeight tryptic peptides derived from the 180-kD protein identified it as the clathrinheavy chain (Fig. 1B) (12). Binding ofclathrin to GST-b3AC could also be dem-onstrated by immunoblotting with antibod-ies to either the heavy (Fig. 1C, IB, and Fig.1D, TD.1) or light (Fig. 1D, CON.1) chainof clathrin as well as by immunoprecipita-tion of metabolically labeled proteins (Fig.1C, IP). Furthermore, the extent of clathrinbinding to GST-b3AC was comparable tothat of a GST fusion protein having the

E. C. Dell’Angelica and J. S. Bonifacino, Cell Biology andMetabolism Branch, National Institute of Child Health andHuman Development, National Institutes of Health, Be-thesda, MD 20892, USA.J. Klumperman and W. Stoorvogel, Department of CellBiology, School of Medicine, Institute of Biomembranes,Utrecht University, 3584 CX Utrecht, The Netherlands.

*To whom correspondence should be addressed. E-mail:juan@helix.nih.gov

REPORTS

www.sciencemag.org z SCIENCE z VOL. 280 z 17 APRIL 1998 431

on

Feb

ruar

y 23

, 201

1w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from