Semen-Derived Amyloid Fibrils
Drastically Enhance HIV Infection
Jan Mu ¨nch,1Elke Ru ¨cker,1Ludger Sta ¨ndker,2,3Knut Adermann,2,3,4Christine Goffinet,5Michael Schindler,1
Steffen Wildum,1Raghavan Chinnadurai,1Devi Rajan,1Anke Specht,1Guillermo Gime ´nez-Gallego,6
Pedro Cuevas Sa ´nchez,7Douglas M. Fowler,8Atanas Koulov,8Jeffery W. Kelly,8Walther Mothes,9
Jean-Charles Grivel,10Leonid Margolis,10Oliver T. Keppler,5Wolf-Georg Forssmann,2,3,*
and Frank Kirchhoff1,*
1Institute of Virology, University Clinic of Ulm, 89081 Ulm, Germany
2IPF PharmaCeuticals GmbH, 30625 Hannover, Germany
3Hannover Medical School, Center of Pharmacology, 30625 Hannover, Germany
4VIRO Pharmaceuticals GmbH & Co. KG, 30625 Hannover, Germany
5Department of Virology, University of Heidelberg, 69120 Heidelberg, Germany
6Centro de Investigaciones Biolo ´gicas (CIB/CSIC), Madrid 28040, Spain
7Hospital Ramo ´n y Cajal, Madrid 28034, Spain
8Department of Chemistry and the Skaggs Institute of Chemical Biology, The Scripps Research Institute, La Jolla, CA 92037, USA
9Section of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT 06536, USA
10National Institute of Child Health and Human Development, Bethesda, MD 20892, USA
*Correspondence: email@example.com (W.-G.F.), firstname.lastname@example.org (F.K.)
Sexual intercourse is the major route of HIV
transmission. To identify endogenous factors
sion, we screened a complex peptide/protein li-
brary derived from human semen. We show that
naturally occurring fragments of the abundant
semen marker prostatic acidic phosphatase
(PAP) form amyloid fibrils. These fibrils, termed
Semen-derived Enhancer of Virus Infection
(SEVI), capture HIV virions and promote their at-
tachment to target cells, thereby enhancing the
infectious virus titer by several orders of mag-
nitude. Physiological concentrations of SEVI
amplified HIV infection of T cells, macrophages,
ex vivo human tonsillar tissues, and transgenic
rats in vivo, as well as trans-HIV infection of
T cells by dendritic or epithelial cells. Amyloido-
genic PAP fragments are abundant in seminal
fluid and boost semen-mediated enhancement
of HIV infection. Thus, they may play an impor-
tant rolein sexual transmission of HIV and could
represent new targets for its prevention.
HIV-1, the causative agent of AIDS, has infected about 60
million people and caused over 20 million deaths. More
than 80% of these HIV-1 infections are acquired through
sexual intercourse. Despite its dramatic spread in the hu-
man population, the efficiency of HIV-1 transmission via
the sexual route is surprisingly poor. For instance, the
risk of male-to-female intravaginal HIV-1 transmission is
estimated at about 1 event per 200–2000 coital acts
ing acute infection when the viral load is particularly high
(Pilcher et al., 2004). Moreover, the presence of other sex-
ually transmitted diseases and sexual practices associ-
ated with bleeding and lesions of the mucosal barrier
can increase this risk to up to 3% per sexual contact (Gal-
vin and Cohen, 2004). Nevertheless, the poor transmissi-
bility of HIV-1 is clearly a major factor restricting the
Globally, most infections result from genital exposure
to semen (SE) of HIV-positive men (Royce et al., 1997).
Women who acquired HIV-1 through vaginal intercourse
constitute almost 60% of new infections in Africa (re-
viewed in Haase, 2005). The infectivity of HIV-1 in male
genital fluid together with the susceptibility of the host,
the type of sexual practice, and the viral load are major
determinants of sexual transmission (Chakraborty et al.,
2001; Gupta et al., 1997; Pilcher et al., 2004). The factors
modulating HIV infectiousness in SE are poorly under-
stood (reviewed in Miller and Shattock, 2003).
To identify natural agents that might play a role in sexual
transmission of HIV/AIDS, we screened a complex pep-
tide/protein library derived from human seminal fluid
(SE-F) for novel inhibitors and enhancers of HIV infection.
We found that fragments of prostatic acidic phosphatase
(PAP) drastically enhance HIV infection. Functional and
structural analyses showed that these peptides form
amyloid fibrils that capture HIV particles and strongly en-
hance the number of productively infected cells by pro-
moting virion-cell attachment and fusion. In agreement
with a relevant role in vivo we found that semen and
Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc. 1059
seminal fluid also drastically enhance HIV infection and
provide evidence that fibril-forming PAP fragments con-
tribute to this effect. Our data support that amyloidogenic
peptides are abundant in semen and promote sexual
transmission of HIV/AIDS.
Purification of an Enhancer of HIV-1 Infection
Our recentisolation ofanovel HIV-1entryinhibitor (Mu ¨nch
et al., 2007) shows that screening of peptide libraries from
unknown molecules modulating HIV-1 infection. To iden-
tify factors that play a role in sexual transmission of HIV-1,
we analyzed a complex peptide/protein library derived
from pooled human SE. This library encompassed 294
fractions and ought to represent all peptides and small
proteins (MW < 50 kDa) present in seminal fluid (SE-F).
We found that fraction 29 of pH pool 7 significantly en-
only small amounts of peptide/protein (Figure 1A, left in-
set). Mass spectrometry (MS) after one additional round
Peptide sequencing identified them as fragments of PAP
(TableS1).Thisproteinisproduced bytheprostatic gland,
secreted in large amounts (1–2 mg/ml) into SE (Ro ¨nnberg
All peptides mapped to the same region of PAP but dif-
fered in length from 34 to 40 residues. The predominant
form of 4551 Da (Figure 1A, right inset) corresponds to
amino acids 248 to 286 of PAP (EMBL accession number
Next, we verified that chemically synthesized PAP pep-
tides also enhance HIV-1 infection, whereas control pep-
tides, including a sequence scrambled variant of the pre-
and Table S1). The PAP peptides promoted HIV-1 infec-
tion in the absence of fetal bovine serum (Figure S1) indi-
cating that no serum cofactor is required. If not mentioned
otherwise, all subsequent experiments were performed
with the synthetic peptide corresponding to the major
form detected in the SE-F (PAP248-286). Examination of
CEMx174 5.25 M7 (CEMx M7) cells containing the GFP
reporter gene under the control of the HIV-1 promoter
(Hsu et al., 2003) by fluorescence microscopy (Figure 1C)
and flow cytometry (Figure 1D) confirmed that PAP248-
286 drastically increases the number of HIV-1-infected
(Figure S2). Fresh PAP248-286 solutions were inactive in
promoting HIV-1 infection. After overnight incubation,
however, they enhanced HIV-1 infection more efficiently
and were less cytotoxic than the polycation polybrene
roviral gene transfer. In comparison, full-length PAP nei-
ther promoted HIV-1 infection nor inhibited the enhancing
activity of PAP248-286 (Figure 1F).
PAP Fragments form Amyloid Fibrils that Promote
We noted that PAP248-286 and other PAP fragments in-
creased HIV-1 infection only when the solutions became
turbid either spontaneously during storage or after agita-
tion and found that the precipitate contains the active
fibrils associated with Alzheimer’s disease enhance HIV-1
infection (Wojtowicz et al., 2002). Thus, we examined
that agitation of fresh PAP248-286 solutions induced
a strong increase in Thioflavin T binding (Figure 2B), in
green birefringence upon staining with Congo red (data
croscopy confirmed effective fibril formation (Figure 2C),
and X-ray powder diffraction demonstrated reflections at
spacing and intersheetdistances, respectively(Figure 2D).
All these properties are typical for amyloid fibrils (reviewed
in Nilsson,2004). LengthvariationsattheN terminusof the
PAP fragments did not impair fibril formation or enhance-
ment of HIV-1 infectivity, whereas deletion of the four
C-terminal (LIMY) residues reduced both effects (Figures
1B, S4,and S5). However, the PAP247-282fragment lack-
ing the LIMY region did not exert efficient transdominant-
negative effects on fibril formation by other PAP fragments
(Figure S4). Notably, PAP fragments were substantially
more potent in enhancing HIV-1 infection than other amy-
loidogenic peptides (Figure S5). Since the amyloid fibrils
strongly enhanced the infectivity of HIV-1 we refer to
them herein after as Semen-derived Enhancer of Virus
To assess whether SEVI interacts directly with HIV-1
particles, we preincubated virus stocks with the fibrils
for 5 min in a small volume and subsequently added the
virus/SEVI mixture to the cell culture, thereby diluting it
50-fold. Diluting the HIV-1/SEVI mixture did not reduce
the magnitude of infectivity enhancement (Figure S6A),
implying that SEVI efficiently bound to the virions. How-
ever, preincubation of the target cells with SEVI also
enhanced HIV-1 infection, even after extensive washing
(Figure S6B). To clarify whether SEVI promotes a physical
interaction of virions with target cells, we incubated TZM-
bl cells with HIV-1 particles in the presence or absence of
sociated p24 after extensive washing. SEVI significantly
enhanced binding of both wild-type HIV-1 particles and
virionslackingEnv,althoughthe absolute levelsof cell-as-
sociated p24 were about 30-fold lower in the absence of
Env (Figure 2E). To directly visualize the effects of SEVI,
we monitored infection of TZM-bl cells microscopically
using fluorescently labeled HIV-1 virions. We found that
SEVI drastically enhanced virion binding to the cells and
the coverslips (Figure 2F). Parallel bright-field phase
images demonstrated that SEVI fibrils are loaded with
1060 Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc.
Figure 1. Purification of an Enhancer of HIV-1 Infection from Seminal Fluid
(A) SE-derived fraction 29 of pH pool 7 (red) promotes HIV-1 infection of P4-CCR5 cells. The inlets show the HPLC chromatograms of pH pool 7 (left)
and of the active fraction 29 (right). Fraction 44 used for peptide sequencing is shown in red and the molecular masses of the major peaks are indi-
cated.+,nopeptide addedand?,uninfectedcells.Average values(±standard deviation [SD])of triplicate measurementsare showninpanels (A)and
(B) Synthetic PAP fragments enhance HIV-1 infection. Infectivities were determined in TZM-bl cells and are shown relative to those measured in the
absence of peptide (100%). Data represent mean values obtained from quintuple infections. Numbers correspond to the amino acid positions in
full-length PAP; PAPscr represents a scrambled form of PAP248-286. C1 to C4 are structurally and functionally unrelated control peptides. Active
peptides are indicated in red.
(C and D) Analysis of CEMx M7 cells infected with HIV-1 in the presence of PAP248-286by (C) fluorescence microscopy and (D) flow cytometric anal-
ysis. The numbers in (D) indicate the percentages of HIV-1-infected (GFP+) cells.
(E) Active PAP248-286 enhances HIV-1 infection of TZM-bl cells more efficiently and is less cytotoxic than Polybrene. The effects of PAP248-286 that
was freshly diluted or incubated overnight and Polybrene on viral infectivity (left) and metabolic activity (right) are shown relative to those measured in
the absence of peptide (100%) and were derived from sextuple measurements.
(F) Full-length PAP does not enhance HIV-1 infection. TZM-bl cells were infected in the presence of PAP248-286 and full-length PAP agitated over-
night or mixtures thereof. Similar results were obtained with full-length PAP agitated for up to one week.
Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc. 1061
Figure 2. PAP248-286 Forms Fibrils that Enhance Virion Attachment and Fusion
(A) Effect of PAP248-286 agitated overnight, the clear supernatant, and the redissolved pellet on HIV-1 infection of CEMx M7 cells. Average values
obtained from sextuple infections are shown.
(B) PAP248-286 aggregation to active fibrils monitored by Thioflavin T fluorescence. RFU, relative fluorescence units.
(C) Electron micrographs of freshly diluted PAP248-286 and after agitation for 16 hr.
(D) SEVI was subjected to X-ray powder diffraction. The resulting diffraction pattern, exhibiting strong reflectionsat 4.7 and 10.6 A˚,is characteristic of
an amyloid cross b sheet structure.
(E) SEVI enhances binding of wild-type (left) and Env-defective (right) HIV-1 particles to TZM-bl cells. Shown are average values (±SD) of sextuple
measurements. The numbers above the bars give n-fold enhancement of p24 binding relative to that measured in the absence of SEVI.
1062 Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc.
sequestered HIV-1 virions, and timelapse microscopy re-
vealed that they are efficiently captured and internalized
by cellular protrusions (Movies S1 and S2). Electron mi-
croscopy confirmed uptake of PAP248-286 fibrils into
the cells (Figure S7). However, despite evidence for the
internalization of fibrils into cells, enhancement of HIV-1
infection by SEVI was not affected when TZM-bl cells
sitol 3-kinase and contractile actin microfilaments that
all block phagocytosis (Figure S8). Thus, phagocytosis
is not critical for the ability of SEVI to promote HIV-1
To further elucidate how SEVI promotes viral infection
we monitored both HIV-1 virion fusion and viral gene ex-
pression from the same infection (Cavrois et al., 2002;
Goffinet et al., 2007). We found that SEVI enhanced HIV-1
virion fusion and subsequent gene expression about
10-fold (Figures 2G and S9). The fusion inhibitor T20
and the CCR5 antagonist TAK779 blocked CCR5-tropic
HIV-1 YU2 virion fusion, whereas the CXCR4 ligand
AMD3100 had no inhibitory effect. Predictably, the RT in-
hibitor Efavirenz blocked viral gene expression but not vi-
rion fusion (Figure S9). Altogether, these results suggest
that SEVI enhances HIV-1 infection by capturing virions
and promoting their physical interaction and fusion with
target cells but does not bypass the requirement for the
appropriate coreceptor, e.g., by disrupting the integrity
of the cell membrane.
SEVI Is a General Enhancer of HIV-1 Infection
infection depends on the viral geno- or phenotype we an-
alyzed its effect on a large panel of HIV-1 variants. We
found that SEVI enhanced infection by R5- X4-, and
dual-tropic HIV-1 clones at concentrations R 0.8 mg/ml
(Figure S10). Irrespective of the viral pheno- or genotype,
were observed after infection with low viral doses
fectivity enhancement was inversely correlated to the vi-
rus inoculum or infectious dose (Figures S11 and 3C).
Thus, the enhancing activity of SEVI is most pronounced
when the levels of infectious virus are low and hence re-
semble theconditions ofsexual HIV-1 transmission.Nota-
magnitude of viral infectivity enhancement by SEVI was
stronger for individual HIV-1 molecular clones than for dif-
ferentprimary HIV-1 isolates(Figures3Cand S11B).Thus,
viral properties, such as CD4 and coreceptor affinities,
may also impact SEVI-mediated enhancement of HIV-1
cell type dependent. Our data demonstrate that SEVI
enhanced infection of various indicator cell lines with sim-
ilar efficiency (Figure S12). More importantly, SEVI also ef-
ficiently increased HIV-1 infection of peripheral blood
mononuclear cells (PBMC) and monocyte-derived macro-
phages (MDM) by both R5 and X4 HIV-1 (Figures 4A and
4B). The effect of SEVI on X4 HIV-1 infection of macro-
phages wasrelativelyweak, mostlikely becauseMDMex-
press low levels of CXCR4 and the amyloid fibrils do not
bypass the requirement for the appropriate coreceptor.
In addition to the infection of cells at the site of exposure,
binding of viral particles to DC-SIGN expressed on den-
dritic cells (DCs) and their subsequent dissemination to
lymphatic organs may play a role in sexual transmission
and systemic spread of HIV (Geijtenbeek et al., 2000).
Therefore, weexaminedwhetherSEVIaffects HIV-1infec-
tion in DC-T cell cocultures. Predictably, binding of HIV-1
to DCs strongly enhanced trans-HIV-1 infection of T cells.
Remarkably, SEVI amplified infection of T cells by HIV-1
particles bound to DCs even further by up to 24-fold
(Figure 4C). It has been observed that SE-F enhances
binding of virions to epithelial cells in ex vivo cervicovagi-
nal tissue (Maher et al., 2005). Following this observation,
infection in the absence of DC-SIGN. The HeLa human
epithelial carcinoma cell line (Scherer et al., 1953) re-
mained nonpermissive for HIV-1 infection even in the
presence of SEVI (data not shown). However, SEVI
increased the ability of HeLa cells to transmit R5- and
X4-tropic HIV-1 to T cells by 30- to 70-fold (Figure 4D).
Thus, SEVI may promote virus attachment to genital sur-
faces, penetration of the mucosal barrier, and subsequent
dissemination to lymphoid organs by increasing HIV-1
virion binding to epithelial cells and to migrating DCs.
SEVI Increases the Infectious Titer of HIV-1
up to Five Orders of Magnitude
The magnitude of HIV-1 entry enhancement in infectivity
assays may underestimate the real potency of SEVI be-
cause 0.5% to 5% of the target cells already became in-
fected in its absence. To determine the effect of SEVI on
the TCID50(50% tissue culture infectious dose) of virus
stocks more accurately, we performed thorough limiting
dilution infection assays. Treatment with SEVI enhanced
the TCID50of HIV-1 in CEMx M7 cells by four orders of
magnitude (Figures 5A and 5B). Amazingly, up to
400,000-fold enhancement was measured in PBMC cul-
tures (Figures 5A and 5B). Notably, these assays detected
spreading HIV-1 infection, implying that the detected vi-
rions were replication competent. We calculated the num-
ber of virions in our virus stocks assuming that there are
(F) Effect of SEVI on binding of HIV-1 virions to TZM-bl cells. Cells were infected for 30 min with equal doses of infectious fluorescently labeled HIV-1
virions in the absence or presence of SEVI (20 ug/ml). The arrows indicate the positions of fibrils.
mock-infected or infected with untreated or SEVI-treated HIV-1R7/3YU-2 Env GFP BlaM-Vpr. The CCR5 antagonist TAK779, the CXCR4 antagonist
AMD3100, and the NNRTI Efavirenz were added to the cells 1 hr prior to virus challenge.
Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc. 1063
5000 p24 CA proteins per viral particle (Briggs et al., 2004)
and determinedthe genomic HIV-1RNAcopynumbersby
quantitative real-time PCR analysis. The results obtained
by both methods showed that just one to three virions
are usually sufficient for productive HIV-1 infection of
PBMC in the presence of SEVI.
sion, we inoculated unstimulated ex vivo human tonsillary
tissue (HLT) with a dose of HIV-1 (0.1 pg p24) that is usu-
ally subinfectious. Efficient replication of HIV-1 was ob-
served, however, in the presence of SEVI (Figure 5C).
Control experiments showed that AZT blocked p24 pro-
duction by HLT infected ex vivo with X4 and R5 HIV-1
clones both in the absence or presence of amyloid fibrils
(Figure 5D). Thus, SEVI enhances HIV-1 replication in
ex vivo infected HLT and not the desorption of inoculated
virions. These results demonstrate that SEVI lowers the
viral threshold required for productive HIV-1 infection of
human tonsillar tissues ex vivo.
PAP Fragments Boost Semen-Mediated
Enhancement of HIV Infection
pooled SE-F, which corresponds to a yield of about
35 mg/ml. PAP, the precursor of SEVI, is secreted into SE
at quantities of 1 to 2 mg/ml (Ro ¨nnberg et al., 1981). Thus,
this concentration of SEVI/PAP fragments can be
achieved by degradation of about 20% of its precursor.
Importantly, SEVI already enhanced HIV-1 infection at
concentrations R 2 mg/ml (Figures 1E, 4, and 5B). To
directly test whether SE contains PAP fragments in suffi-
cient quantity for HIV-1 infectivity enhancement, we ana-
lyzed partially purified peptide samples from individual
SE donors. Proteins > 50 kDa were removed from the
SE-F by ultracentrifugation and the resulting peptide/pro-
tein mixture was separated by a single RP chromatogra-
phy step. Infection of TZM-bl cells with HIV-1 in the pres-
ence of fractions predicted to contain amyloidogenic
PAP fragments revealed that they efficiently enhance
HIV-1 infection after short-term storage without agitation
Next, we assessed whether SEVI is active under condi-
tions resembling the deposition of HIV-infected SE in the
genital tract. As expected from published data (Kiessling,
2005), SE and SE-F were highly cytotoxic (Figure S14). To
minimize these toxic effects we mixed virus stocks with
equal volumes of fresh SE spiked with SEVI and added
25 ml of these mixtures to 975 ml CEMx M7 cell cultures.
ofHIV-1 inSE andSE-F ina dose-dependent manner(Fig-
ures S15). To assess whether active PAP aggregates may
form in vivo, we spiked SE-F with unassembled PAP248-
286. Unexpectedly, the PAP248-286 fragments boosted
the ability of SE-F to enhance HIV-1 infection even at the
first time point analyzed (Figure S16A), indicating that
fibrils or other active aggregates already formed during
freezing and thawing of the SE-F samples. At an acidic
pH of 4.2, which resembles the vaginal environment, this
enhancing effect became apparent after 30 min and was
more pronounced after 2 hr (Figure S16A). Our finding
Figure 3. SEVI Enhances HIV-1 Infection
Independently of the Viral Geno- and
(A) Relative effect of SEVI (5 mg/ml) on infection
of CEMx M7 cells by primary R5- (blue), X4-
(black), and dual-tropic (red) HIV-1 M and O
isolates. (A) and (C) show the average en-
hancement (n = 2) of the infectivity levels rela-
tive to those measured in the absence of SEVI.
(B) SEVI enhances HIV-1 infection most effi-
ciently after low-dose infection. TZM-bl cells
were infected with the indicated doses of X4
(black) or R5 (blue) HIV-1 in the presence
(squares) or absence (triangles) of SEVI (10
mg/ml). Each symbol represents the average
b-galactosidase activity (n = 3) measured 3
days after virus exposure. RLU/s: relative light
units per second.
(C) Correlation between the magnitude of
SEVI-mediated enhancement of HIV-1 infec-
tion and the p24 content (left) or the infectivity
(right) of the viral stocks used for infection.
Shown is the average (n = 3) enhancement of
HIV-1 infection by SEVI (10 mg/ml) measured
after inoculation of TZM-bl cells with virus
stocks containing the indicated quantities of
p24 antigen content (left) or as a function of
the b-galactosidase reporter gene activities
measured after infection with the untreated
virus stocks (right).
1064 Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc.
that amyloidogenic PAP fragments rapidly acquire the
ability to increase HIV-1 infection suggests that smaller
aggregates that typically precede mature fibril formation
may also enhance HIV-1 infection. In agreement with
this possibility, we found that fresh PAP248-286 solutions
already moderately enhanced HIV-1 infection after 1 hr
of agitation and that this effect saturated after 7 hr
(Figure S16B). In comparison, Congo Red or Thioflavin T
binding increased with slower kinetics and did not satu-
rate within the 24 hr investigation period. Different assay
sensitivities and the possibility that mature fibril formation
may occur during the 2 hr infection period prevent defini-
tive conclusions. Nevertheless, these data support that
PAP248-286 aggregates forming prior to mature fibrils
also promote HIV infection.
Semen Increases HIV-1 Infection
To analyze whether semen directly affects the infectious-
ness of HIV-1, we mixed virus stocks with SE, SE-F, or the
semen pellet (SE-P) containing the spermatocytes and
large protein aggregates. Subsequently, these mixtures
were added to TZM-bl cell cultures thereby diluting them
15-fold. After 3 hr the virus/semen-containing medium
was replaced by fresh medium. Under these conditions,
mixtures containing up to 10% SE, SE-F, or the SE-P dur-
ingvirion incubation displayed onlyweak cytotoxic effects
particularly at high cell density (Figure S17A).Treatment of
virusstockswithSE, SE-F,andSE-P drasticallyenhanced
HIV-1 infection (Figures 6A and S17B) and virus-induced
cytopathic effects (Figure 6B). Similarly to SEVI-mediated
infectivity enhancement (Figure 3B), the most dramatic
effects of SE were observed after inoculation with low viral
doses (Figure 6C). SE, SE-F, and the SE-P also increased
transmission of R5 HIV-1 variants from HeLa cells to TZM-
bl cells and infection of CEMx M7 cells (Figures 6D and
6E). Acidification enhanced Thioflavin T binding by SE
HIV-1 infection (Figure S18). Thus, acidification of semen
and seminal fluids most likely results in the precipitation
of proteins/peptides that increase nonspecific Thioflavin
T binding, rather than in the formation of amyloid fibrils.
More importantly, the results demonstrate that an acidic
pH, resembling the conditions in the vaginal tract, does
not inhibit enhancement of HIV-1 infection by SE and
To further analyze the enhancing factor(s), we agitated
fresh, clear SE-F overnight. The solution became slightly
turbid and newly aggregated peptides and proteins were
preciptitated by centrifugation. The supernatant was dis-
carded and the pellet dissolved in the same volume of
PBS. Treatment of virus stocks with an equal volume of
this solution potently enhanced HIV-1 infection (7.2 ±
1.7-fold, n = 3) showing that the precipitate contains a
significant proportion of the enhancing activity. Further
Figure 4. The Effect of SEVI on HIV-1 Infection Is Cell Type Independent
(A and B) Effect of SEVI on R5- or X4-tropic HIV-1 infection of PBMCs (A) and macrophages (B).
(C and D) SEVI enhances in trans-infection of T cells by viral particles bound to DCs (C) or HeLa cells (D). The numbers indicate n-fold infectivity
enhancement observed in the presence of the indicated concentrations of SEVI compared to those measured in the absence of SEVI. U, uninfected
cells; HIV-1, wash control; DCs, HeLa and CEMx M7, indicated cell type only. All data shown in this figure give average values ± SD obtained from
triplicate or quintuple infections.
Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc. 1065
analysis of the compounds present in this small precipi-
tate by RP-HPLC showed that fractions eluting at reten-
tion times similar to synthetic PAP fragments contained
primarily a peptide of 4553 Da, corresponding to
PAP248-286 (Figure 6F). These results clearly support
that endogenous PAP248-286 fragments form precipita-
ble aggregates in SE-F that contribute to the enhancing
effect of semen on HIV-1 infection.
SEVI Enhances HIV-1 Infection In Vivo
It has been established that rats transgenically expressing
human CD4 and CCR5 on T cells and macrophages are
susceptible to HIV-1 infection (Keppler et al., 2002; Goffi-
net et al., 2007). To explore the ability of SEVI to enhance
HIV-1 infection in vivo, hCD4/hCCR5-transgenic rats were
challenged with HIV-1 YU2 by tail vein injection. Four days
after inoculation between 3 and 12 copies of HIV-1 cDNA
per ng of total DNA were detected in splenocyte extracts
derived from four rats infected with untreated virus stocks
(Figure 7). In comparison, the relative viral cDNA copy
numbers were about 5-fold higher ranging from 21 to 49
of cDNA per ng of total DNA in five rats that received the
same dose of SEVI-treated HIV-1. Thus, SEVI significantly
enhanced the infectivity of R5 HIV-1 in vivo. Notably, this
result demonstrates that SEVI does not only enhance tar-
edly increases the levels of HIV-1 infection in a lymphatic
organ by viral particles transported via the blood stream.
SEVI—a Key Factor in Sexual HIV-1 Transmission?
The limited ability of HIV-1 to cross the mucosal barrier
and to infect sufficient numbers of cells in the genital tract
to establish a sustained infection constitutes a major bar-
rier for sexual transmission (reviewed in Haase, 2005). In
this study, we show that fragments of PAP, a highly abun-
dant SE marker, form amyloid fibrils (SEVI) that drastically
Figure 5. SEVI Lowers the Threshold Required for Productive HIV-1 Infection
in the presence of the indicated concentrations of SEVI. Indicated is the number of cultures that became productively infected. Similar results were
obtained using different X4- and R5-tropic virus stocks.
(B) Effect of SEVI on the TCID50of HIV-1. TCID50values were calculated based on the infectivity data shown in (A). The numbers above the bars
indicate n-fold enhancement compared to the titer measured in the absence of SEVI.
(C) Effect of SEVI on HIV-1 replication in ex vivo-infected human lymphoid tissue. Tonsillary tissue was infected ex vivo with a very low viral dose
(0.1 pg p24) of X4-tropic HIV-1 that has been preincubated with different concentrations of SEVI. Similar results were obtained in two independent
experiments using tissues derived from different donors.
(D) Influence of AZT on HIV-1 replication in ex vivo HLT. Tissues were infected with untreated (blue symbols) or SEVI-treated (50 mg/ml; red symbols)
X4 (left panel) or R5 (right panel) HIV-1 virus stocks containing 1.0 or 0.1 ng of p24 antigen, respectively, and subsequently cultured in the presence
(open symbols) or absence (filled symbols) of AZT.
1066 Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc.
enhance the infectiousness of HIV-1 by promoting virion-
infection of CD4+T lymphocytes and macrophages (most
in Pope and Haase, 2003; Haase, 2005) and increased
trans-HIV-1 infection of CD4+T cells by an epithelial cell
line and by primary DCs (Figure 4). An intact mucosal ep-
ithelium provides a strong physical barrier to HIV-1 infec-
tion (Miller and Shattock, 2003). However, epithelium in-
tegrity is often compromised after sexual intercourse as
well as in the presence of ulcerative sexually transmitted
diseases. Moreover, semen itself might cause local in-
flammation or epithelial breaks and induce DC projections
to the luminal surface (Sharkey et al., 2007). Thus, SE
components—such as amyloidogenic PAP fragments—
might frequently be able to access CD4+T cells, macro-
phages, and DCs in the subepithelium to enhance HIV-1
attachment, infection, and dissemination. Our data sug-
gest that amyloidogenic peptides in SE may help HIV to
pass the early ‘‘bottleneck’’ in infection by assisting the vi-
rus to attach to genital surfaces, to establish a self-propa-
gating infection at the point of entry, and to cross the
mucosal barrier with migrating DCs.
Although the magnitudes of the effects measured in in-
fectivity assays were quite remarkable, they underesti-
mated the real potency of SEVI. More quantitative limiting
dilution-infection assays demonstrated that SEVI en-
hanced the TCID50of HIV-1 by three to five orders of mag-
of detectable infectious units to virus particles was on the
order of 1:1,000 to 1:100,000, which is in agreement with
published data (Dimitrov et al., 1993; Rusert et al., 2004).
Figure 6. Semen and Seminal Fluid Enhance HIV-1 Infection
(A) Effect of SE, SE-F, and SE-P on R5-tropic HIV-1 infection of TZM-bl cells. The numbers indicate n-fold infectivity enhancement observed after
treatment of HIV-1 with the indicated concentrations of SE, SE-F, and SE-P compared to those measured after infection with the PBS-treated virus
stock. U, uninfected cells. Average values ± SD (n = 3) are shown in (A) and (C)–(E).
(B) Microscopic examination of TZM-bl cells infected with equal doses of HIV-1 that was either untreated or treated with semen.
(C) Semen enhances HIV-1 infection most efficiently after low-dose infection. TZM-bl cells were infected with 10-fold dilutions of R5-topic HIV-1
treated with the indicated concentrations of semen. The numbers indicate n-fold enhancement compared to the infectivity measured using PBS-
treated virus stocks. The X-axes in (C) and (E) indicate the percentage of semen during virus incubation (blue) and the final concentration in the
cell culture (black).
(DandE)Semenenhances intrans-infectionofTZM-blbyvirus-exposedHeLacellsandinfection ofCEMxM7cells.TZM-blcellswereexposedtothe
virus/semen mixture for 3 hr and CEMx M7 cells for 1 hr, respectively. Numbers indicate n-fold infectivity enhancement observed in the presence of
the indicated concentrations of SE, SE-F, and SE-P. U, uninfected cells; HeLa, HIV-1-exposed HeLa cells only.
(F) MALDI-MS analysis of aggregated protein/peptides pelleted from seminal fluid. Molecular masses and the N-terminal sequence of the major peak
Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc. 1067
Incontrast, onlyoneto three virionswere usuallysufficient
for productive infection in the presence of SEVI. Recent
data suggest that the apparent high frequency of ‘‘nonin-
fectious’’ HIV-1 particles is largely due to the low fre-
quency of successful virus-cell interactions (Thomas
et al., 2007). In agreement with this possibility, our data
demonstrate that a few HIV-1 particles are sufficient for
spreading infection in the presence of an effective attach-
HIV-1 is detected in SE and cell-free SE-F of most in-
fected men (Tachet et al., 1999), even under HAART
(Zhang et al., 1998). However, the quantity of HIV-1 trans-
mitted during sexual intercourse is usually subinfectious
(Gray et al., 2001). On average, about 11,000 RNA cop-
ies/ml have been detected in SE of HIV-1-infected men
(Gupta et al., 1997). Thus, an ejaculate of 4 ml SE would
deposit ?22,000 virions in the genital tract, which corre-
sponds to approximately 5 pg of p24 antigen. Interest-
ingly, productive HIV-1 infection of PBMC cultures and
of ex vivo tonsillar tissues after exposure to such low viral
doses was only observed in the presence of SEVI (exam-
ples shown in Figure 5).
Semen-Mediated Enhancement of HIV-1 Infection
worldwide, surprisingly little is known about its effects on
viral infectivity. One of the most striking findings of our
study is that semen potently enhances HIV-1 infection
(Figures 6, S17, and S18). Perhaps this effect was missed
in previous studies because analysis of semen is compli-
cated due to its high cytotoxicity and frequent bacterial
contaminations. We circumvented these problems, at
least in part, by adding relatively small volumes of virus/
semen mixtures to monolayer cell cultures for limited
time periods. To some extent this approach resembles
the deposition of the HIV-1-infected ejaculate in the geni-
infection was not reduced by an acidic pH (Figures S16
and S18B), which represents a nonspecific vaginal de-
fense mechanism against various pathogens including
HIV-1 (Tevi-Benissan et al., 1997). Thus, the observed ef-
fects of semen on HIV-1 infection are most likely relevant
for sexual viral transmission.
Our data support that amyloidogenic PAP fragments
account for a significant fraction of the enhancing activity
in semen. First, the yield of amyloidogenic PAP fragments
from pooled SE (?35 mg/ml) considerably exceeds the
concentration required to enhance HIV-1 infection (R2
mg/ml). Second, freshly diluted PAP fragments were highly
prone to fibril formation and rapidly achieved the ability to
enhance HIV-1 infection in the context of SE and SE-F
(Figure S16). The rapid kinetics of HIV-1 infectivity en-
hancement by freshly diluted PAP248-286 suggest that
preexisting aggregates or fibrils in SE-F accelerate amy-
loid formation. Third, PAP248-286 was detected in the
precipitate of SE-F that contains a large fraction of the en-
hancing activity (Figure 6F). Fourth, amyloidogenic PAP
fragments were isolated from a complex semen-derived
library that ought to represent all peptides and proteins
< 50 kDa. Finally, our preliminary data show that a second
semen-derived peptide fraction enhancing HIV-1 infection
contains a different PAP fragment that also forms amyloid
fibrils. It is noteworthy that seminal vesicle amyloid is
a well-known form of localized amyloidosis (Maroun
et al., 2003) and that PAP can readily be detected in vag-
inal washings and hence has been used as a SE marker,
e.g., in the case of alleged sexual assaults (Graves et al.,
1985). After sexual intercourse, elevated levels of PAP
can usually be detected in the vagina for about 24 hr,
but not after 48 hr (Collins and Bennett, 2001). Thus, it is
easy to envision that the concentrations of SEVI may in-
crease after sexual intercourse because intact PAP is de-
graded in the vaginal, rectal, or oral environment. Since
the concentration of SEVI during virion incubation deter-
mines the magnitude of infectivity enhancement (Figures
S6A and S15), its subsequent dilution by vaginal fluids
would not diminish its potency. Accordingly, SEVI fibrils
that form prior to or after sexual intercourse may all in-
crease virion attachment and infectivity. Amyloid fibrils
are known to be highly stable and could potentially facili-
tate HIV-1 infection in the genital tract for relatively long
genic PAP fragments also serve a purpose in normal SE
physiology. Moreover, it will be interesting to further in-
vestigate how efficiently other possible aggregates, such
as low-molecular-weight oligomers, pores, spheres, and
protofilaments that typically precede fibril formation
(Walsh et al., 1999), promote HIV-1 infection. However,
the relative contribution of such intermediates to HIV-1 in-
fectivity enhancement will be difficult to assess because
even if they are initially separated from monomers and
Figure 7. SEVI Enhances HIV-1 Infection In Vivo
hCD4/hCCR5-transgenic rats were challenged intravenously with un-
treated HIV-1YU-2 (animals A–D) or with virus stocks treated with
SEVI (animals E–I). As control, a hCD4-single transgenic rat (J) was in-
fected withuntreated virus.On4 dayspost-challenge, allanimals were
sacrificed and the spleens were removed for determination of the HIV-1
cDNA load relative to the amount of cellular DNA. The right panel de-
picts arithmetic mean values ± SEM for the two groups of animals.
1068 Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc.
mature fibrils, the equilibrium between these species may
rapidly be re-established. Notably, our results clearly sup-
port a role of immature PAP aggregates in HIV-1 infection.
For example, we did not observe large fibrils in SE,
although amyloidogenic PAP fragments clearly formed
aggregates boosting the potency of semen-mediated
enhancement of HIV-1 infection (Figure S16A), and SE
showed birefringence under polarized light upon staining
with Congo red (data not shown). Moreover, fresh solu-
tions of PAP fragments acquired the ability to enhance
HIV-1 infection very rapidly in comparison to other indica-
tors of amyloid formation, such as increased Congo red
and Thioflavin T binding (Figure S16B). It has been sug-
ing the infectiousness of HIV-1 (Wojtowicz et al., 2002).
Thus, smaller aggregates that form in semen could be
even more potent in assisting HIV-1 infection than the
large SEVI fibrils obtained in vitro.
Possible Role of Amyloid Fibrils in Viral Infections
Recent studies show that amyloid fibrils are much more
common than previously recognized. To date, about 30
human diseases are known to be associated with amyloid
deposits (reviewed in Termussi et al., 2003; Westermark,
are also formed by proteins unrelated to disease (Fowler
et al., 2006) and produced by bacteria and fungi (reviewed
in Gebbink et al., 2005; Kelly and Balch, 2003). It has been
previously shown that b-amyloid fibrils associated with
fection (Wojtowicz et al., 2002). We did not observe an en-
hancing effect of Aß1-40 (Figure S5A). This apparent dis-
crepancy could be due to the fact that amyloid fibers
composed of the same protein can show different confor-
mations with distinct phenotypes (reviewed in Chien et al.,
2004). Our results confirmed, however, that other amy-
loidogenic peptides, i.e., PPI-2480 and a-synuclein, also
enhance HIV-1 infection, albeit with much lower efficiency
than PAP fragments (Figure S5A). Thus, the ability to en-
hance HIV-1 infection seems to be a common feature of
amyloid fibrils. Moreover, the capability to promote the in-
teraction between virions and the cell surface is indepen-
dent of the viral Env glycoprotein and hence not restricted
to retroviruses. Thus, further studies on the role of amy-
loids in the transmission and pathogenesis of enveloped
est to clarify whether some fungal and bacterial infections
may enhance the risk of HIV transmission (and other sexu-
ally transmitted viral infections) because they produce am-
yloid fibrils that facilitate virus adsorption and infection.
ity factor in SE offers exciting new avenues for further
investigation. The high potency of SEVI in promoting viral
infection together with its relatively low cytotoxicity sug-
gests that it may not only play a relevant role in sexual
HIV transmission but could also help to improve vaccine
ample, pretreatment with SEVI may amplify the potency of
in mediating virus-specific immune responses and hence
mucosal immunity. Finally, agents blocking the generation
tachment factors may offer new prospects for preventive
Identification of SEVI
The peptide library was generated and screened for factors affecting
See the Supplemental Experimental Procedures for further details.
Peptide Synthesis and Fibril Formation
Peptides were produced by standard Fmoc solid-phase peptide syn-
thesis, purified by preparative RP HPLC, and analyzed by HPLC and
MS. Lyophilized synthetic peptides were resuspended in serum-free
DMEM or PBS or in SE-F at concentrations of 5 to 10 mg/ml. Fibril for-
mation was induced by overnight agitation at 37?C at 1400 rpm using
an Eppendorf Thermomixer and verified by Congo red staining or elec-
HIV-1 Variants, Virus Stocks, and Infectivity
HIV-1 clones and primary isolates were obtained through the NIH
viously (Mu ¨nch et al., 2007). The TCID50was determined as described
by the ACTG Laboratory Technologist Committee (http://aactg.s-3.
See the Supplemental Experimental Procedures for further details.
PAP250-286 was dissolved in PBS at 2 mg/ml and incubated at 37?C
on a rotating mixer to induce fibril formation. A 0.02 mg/ml suspension
of SEVI was adsorbed for 60 s onto 200-mesh carbon-coated copper
grids (Electron Microscopy Sciences, Hatfield, PA, USA). Grids were
washed with distilled water and subsequently stained with 1% aque-
ous uranyl acetate (Electron Microscopy Sciences) for 60 s. Fibrils
were visualized with a Philips (New York, USA) CM100 transmission
Thioflavin T Binding Assay
SEVI aggregation was performed as described in the electron micros-
copy section. Aliquots were withdrawn from the aggregation reaction
nm,emission 485nm)was monitoredusing aCaryEclipsefluorimeter.
X-Ray Diffraction Analysis
SEVI wasaggregatedasdescribedintheelectron microscopysection.
tion of about 1.0 mg SEVI in quartz capillaries was recorded using a 6
KW Bruker Direct Drive Rotating anode X-ray generator with a Xenocs
focusing mirror (50 kV3 100 mA, 0.3 3 3 mm focus, 0.5 mm slits, Cop-
per[Cu] Target) and aMar 345mmIPscanner. Thedistance from sam-
ple to scanner was 250 mm and CuK radiation (1.5418 A˚) was utilized.
HIV-1 Infection In trans
1 3 104HeLa or DCs, isolated as described previously (Ru ¨cker et al.,
Thereafter, 10 ml SEVI dilutions and 0.5 ng p24 antigen of HIV-1 NL4-3
or HIV-1 005-pf-103 in a volume of 40 ml were added. After 2 hr of
incubation the cells were washed twice in RPMI and resuspended in
Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc. 1069
150 ml of CEMx M7 cell cultures. Luciferase assay was performed
2 days post-cocultivation.
Infection of Ex Vivo Tonsillary Tissues
Human tonsillar tissues removed during routine tonsillectomies were
dissected, maintained, and infected within 5 hr of excision, as de-
scribed previously (Glushakova et al., 1995; Ru ¨cker et al., 2004), ex-
cept that aliquots of the virus stocks were preincubated with SEVI.
Briefly, the tonsils were dissected into 2 to 3 mm3blocks and infected
by inoculating each block with viral stock suspensions derived from
transfected 293T cells. Infection doses were normalized based on
p24 content. Productive HIV-1 infection was evaluated by measuring
the amount of p24core antigen released into themedium as described
(Ru ¨cker et al., 2004). To assess the effect of AZT each block was in-
fected with wild-type HIV-1 NL4-3 or an R5-tropic 92UG037 V3 re-
combinant (Mu ¨nch et al., 2007), respectively, that were either preincu-
bated for 5 min with various concentration of SEVI or with PBS only.
After overnight incubation, the virus inoculum was removed by exten-
sive washing and the tissues were cultured in fresh medium with or
without AZT (10 mM).
First, 5 3 103TZM-bl cells were sown in 96-well dishes in a volume of
50 ml. Next, 10 ml of DMEM containing different concentrations of SEVI
and 40 mlof 10-fold dilutionsof the viral stocks wereadded to the cells.
After 3 hr incubation at 37?C unbound virus was removed by washing
with DMEM. Thereafter, the cells were lysed in DMEM containing 1%
Triton X-100. Cell-associated HIV-1 core antigen was detected using
the p24 antigen ELISA obtained by the NIH ARRRP.
Imaging Fluorescently Labeled HIV-1
Fluorescently labeled HIV-1 was generated and infection of TZM-bl
cells was monitored essentially as described (Sherer et al., 2003).
See the Supplemental Experimental Procedures for further details.
Detection of PAP248-286 in Seminal Fluid
Freshly collected semen was centrifuged (5 min, 14000 g) to remove
spermatocytes and the SE-F was agitated (1300 rpm) overnight at
37?C in the presence of gentamycin (50 mg/ml). Aggregated peptide
and protein was precipitated by centrifugation for 15 min at 14000 g.
The supernatant was discarded and the pellet was dissolved in 6 M
Guanidin HCl, diluted 1:5 with chromatography buffer A (0.1% TFA
in water), and subjected to an analytical RP HPLC. The peptides/pro-
tein mixture was separated by a linear gradient of buffer B (80% ace-
tonitril in A) and the resulting fractions were analyzed by MALDI-MS
and Edman sequencing.
Effect of Semen on HIV-1 Infection
Semen and seminal fluid were obtained and analyzed for their effects
on HIV-1 infection and transmission as described in the Supplemental
Transgenic Rat Model
The hCD4/hCCR5-transgenic rat model and the generation of replica-
tion-competent HIV-1YU-2 stocks have been reported previously
(Keppler et al., 2002). For details, see the Supplemental Experimental
HIV-1 Virion-Fusion and Gene Expression
This sensitive flow cytometry-based HIV-1 virion-fusion assay was
2007). For details, see the Supplemental Experimental Procedures.
ThePRISM packageversion4.0 (AbacusConcepts, Berkeley,CA)was
used for all statistical calculations. If not specified otherwise, data
present mean values ± SD obtained from at least triplicate measure-
ments. Nonparametric statistical analyses were performed by using
the Mann-Whitney u test.
Supplemental Data include Supplemental Experimental Procedures,
eighteen figures, one table, and two movies and can be found with
this article online at http://www.cell.com/cgi/content/full/131/6/1059/
We thank Thomas Mertens and Bernhard Fleckenstein for support;
Nicola Bailer, Rolf Kopittke, Aleksandra Heidtland, Andreas Zgraja,
Dirk Pape-Lange, Wolfgang Posselt, Daniela Krnavek, Martha Mayer,
and Ina Allespach for expert technical assistance; Reinhold Schmitt
and Silvio Krasemann (IBF Heidelberg) for animal handling; Julia
Lenz and Blanche Schwappach (ZMBH, Heidelberg) for TBD FACSA-
ria analysis; and Nat Landau for CEMx174 5.25 cells. The authors also
thank Beatrice H. Hahn and Ronald C. Desrosiers for discussion; Carl-
and Harald John for cooperation; Xiaoping Dai in Ian Wilson’s labora-
tory at the Scripps Research Institute for help acquiring the X-ray pow-
der diffraction data; and the ‘‘Kinderwunschzentrum’’ Go ¨ttingen for
providing semen. This work was supported by the European TRIoH
consortium (EUproject LSGH-2003-503480) toO.T.K. and the govern-
mentof Lower Saxonyand theVWFoundationtoW.G.F.and bygrants
from the DFG and the Wilhelm-Sander Foundation and NIH grant
1R01AI067057-01A2 to F.K. The work of J.-C.G. and L.M. was sup-
ported by the NICHD Intramural Program.
Received: February 6, 2007
Revised: June 18, 2007
Accepted: October 4, 2007
Published: December 13, 2007
V.M., and Johnson, M.C. (2004). The stoichiometry of Gag protein in
HIV-1. Nat. Struct. Mol. Biol. 11, 672–675.
Cavrois, M., De Noronha, C., and Greene, W.C. (2002). A sensitive and
specific enzyme-based assay detecting HIV-1 virion fusion in primary
T lymphocytes. Nat. Biotechnol. 20, 1151–1154.
Chakraborty, H., Sen, P.K., Helms, R.W., Vernazza, P.L., Fiscus, S.A.,
Eron, J.J., Patterson, B.K., Coombs, R.W., Krieger, J.N., and Cohen,
M.S. (2001). Viral burden in genital secretions determines male-to-
female sexual transmission of HIV-1: a probabilistic empiric model.
AIDS 15, 621–627.
Chien, P., Weissman, J.S., and DePace, A.H. (2004). Emerging princi-
ples of conformation-based prion inheritance. Annu. Rev. Biochem.
Collins, K.A., and Bennett, A.T. (2001). Persistence of spermatozoa
and prostatic acid phosphatase in specimens from deceased individ-
uals during varied postmortem intervals. Am. J. Forensic Med. Pathol.
Dimitrov, D.S., Willey, R.L., Sato, H., Chang, L.J., Blumenthal, R., and
Martin, M.A. (1993). Quantitation of human immunodeficiency virus
type 1 infection kinetics. J. Virol. 67, 2182–2190.
Dittmar, M.T., Zekeng, L., Kaptue, L., Eberle, J., Kra ¨usslich, H.G., and
O isolates from Cameroon. AIDS Res. Hum. Retroviruses 15, 707–712.
son, B.A., Piatak, M., Jr., Lifson, J.D., Mansfield, K.G., and Desrosiers,
R.C. (2005). Immunization of macaques with single-cycle simian im-
munodeficiency virus (SIV) stimulates diverse virus-specific immune
1070 Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc.
responses and reduces viral loads after challenge with SIVmac239. J. Download full-text
Virol. 79, 7707–7720.
Fowler, D.M., Koulov, A.V., Alory-Jost, C., Marks, M.S., Balch, W.E.,
tissue. PLoS Biol. 4, e6. 10.1371/journal.pbio.0040006.
Galvin, S.R., and Cohen, M.S. (2004). The role of sexually transmitted
diseases in HIV transmission. Nat. Rev. Microbiol. 2, 33–42.
Gebbink, M.F., Claessen, D., Bouma, B., Dijkhuizen, L., and Wosten,
H.A. (2005). Amyloids - a functional coat for microorganisms. Nat.
Rev. Microbiol. 3, 333–341.
Geijtenbeek, T.B., Kwon, D.S., Torensma, R., van Vliet, S.J., van Duijn-
hoven, G.C., Middel, J., Cornelissen, I.L., Nottet, H.S., KewalRamani,
V.N., Littman, D.R., et al. (2000). DC-SIGN, a dendritic cell-specific
HIV-1-binding protein that enhances trans-infection of T cells. Cell
Glushakova, S., Baibakov, B., Margolis, L.B., and Zimmerberg, J.
(1995). Infection of human tonsil histocultures: a model for HIV patho-
genesis. Nat. Med. 1, 1320–1322.
Goffinet, C., Allespach, I., and Keppler, O.T. (2007). HIV-susceptible
transgenic rats allow rapid preclinical testing of antiviral compounds
targeting virus entry or reverse transcription. Proc. Natl. Acad. Sci.
USA 104, 1015–1020.
Graves, H.C., Sensabaugh, G.F., and Blake,E.T. (1985). Postcoital de-
tection of a male-specific semen protein. Application to the investiga-
tion of rape. N. Engl. J. Med. 312, 338–343.
Gray, R.H., Wawer, M.J., Brookmeyer, R., Sewankambo, N.K., Ser-
wadda, D., Wabwire-Mangen, F., Lutalo, T., Li, X., van Cott, T., Quinn,
T.C., et al. (2001). Probability of HIV-1 transmission per coital act in
monogamous, heterosexual, HIV-1-discordant couples in Rakai,
Uganda. Lancet 357, 1149–1153.
Gupta, P., Mellors, J., Kingsley, L., Riddler, S., Singh, M.K., Schreiber,
S., Cronin, M., and Rinaldo, C.R. (1997). High viral load in semen of
human immunodeficiency virus type 1-infected men at all stages of
disease and its reduction by therapy with protease and nonnucleoside
reverse transcriptase inhibitors. J. Virol. 71, 6271–6275.
Haase, A.T. (2005). Perils at mucosal front lines for HIV and SIV and
their hosts. Nat. Rev. Immunol. 5, 783–792.
Hsu, M., Harouse, J.M., Gettie, A., Buckner, C., Blanchard, J., and
Cheng-Mayer, C. (2003). Increased mucosal transmission but not en-
hanced pathogenicity of the CCR5-tropic, simian AIDS-inducing sim-
ian/human immunodeficiency virus SHIVSF162P3 maps to envelope
gp120. J. Virol. 77, 989–998.
Kelly, J.W., and Balch, W.E. (2003). Amyloid as a natural product. J.
Cell Biol. 161, 461–462.
Keppler, O.T., Welte, F.J., Ngo, T.A., Chin, P.S., Patton, K.S., Tsou,
C.L., Abbey, N.W., Sharkey, M.E., Grant, R.M., You, Y., et al. (2002).
Progress toward a human CD4/CCR5 transgenic rat model for de
novo infection by human immunodeficiency virus type 1. J. Exp.
Med. 195, 719–736.
Kiessling, A.A. (2005). Isolation of human immunodeficiency virus type
1 from semen and vaginal fluids. Methods Mol. Biol. 304, 71–86.
Maher, D., Wu, X., Schacker, T., Horbul, J., and Southern, P. (2005).
HIV binding, penetration, and primary infection in human cervicovagi-
nal tissue. Proc. Natl. Acad. Sci. USA 102, 11504–11509.
Maroun, L., Jakobsen, H., Kromann-Andersen, B., and Horn, T. (2003).
Amyloidosis of the seminal vesicle - a case report and review of the
literature. Scand. J. Urol. Nephrol. 37, 519–521.
Miller, C.J., and Shattock, R.J. (2003). Target cells in vaginal HIV trans-
mission. Microbes Infect. 5, 59–67.
Mu ¨nch,J., Sta ¨ndker, L., Adermann, K., Schulz, A., Schindler, M.,Chin-
nadurai, R.,Po ¨hlmann,S.,Chaipan,C.,Biet,T.,Peters,T.,etal.(2007).
Discovery and optimization of a natural HIV-1 entry inhibitor targeting
the gp41 fusion peptide. Cell 129, 263–275.
Nilsson, M.R. (2004). Techniques to study amyloid fibril formation
in vitro. Methods 34, 151–160.
art, P.W., Goh, L.E., Cohen, M.S., and Quest Study, Duke-UNC-Emory
Acute HIV Consortium (2004). Brief but efficient: acute HIV infection
and the sexual transmission of HIV. J. Infect. Dis. 189, 1785–1792.
Pope, M., and Haase, A.T. (2003). Transmission, acute HIV-1 infection
and the quest for strategies to prevent infection. Nat. Med. 7, 847–852.
Ro ¨nnberg, L., Vihko, P., Sajanti, E., and Vihko, R. (1981). Clomiphene
citrate administration to normogonadotropic subfertile men: blood
hormone changes and activation of acid phosphatase in seminal fluid.
Int. J. Androl. 4, 372–378.
Royce, R.A., Sena, A., Cates, W., Jr., and Cohen, M.S. (1997). Sexual
transmission of HIV. N. Engl. J. Med. 336, 1072–1078.
Ru ¨cker, E., Mu ¨nch, J., Wildum, S., Brenner, M., Eisemann, J., Margo-
lis, L., and Kirchhoff, F. (2004). A naturally occurring variation in the
proline-rich region does not attenuate human immunodeficiency virus
type 1 nef function. J. Virol. 78, 10197–10201.
Rusert, P., Fischer, M., Joos, B., Leemann, C., Kuster, H., Flepp, M.,
Bonhoeffer, S., Gunthard, H.F., and Trkola, A. (2004). Quantification
of infectious HIV-1 plasma viral load using a boosted in vitro infection
protocol. Virology 326, 113–129.
Scherer, W.F., Syverton, J.T., and Gey, G.O. (1953). Studies on the
propagation in vitro of poliomyelitis viruses. IV. Viral multiplication in
a stable strain of human malignant epithelial cells (strain HeLa) derived
from anepidermoidcarcinomaofthecervix. J.Exp.Med.97,695–710.
Sharkey, D.J., Macpherson, A.M., Tremellen, K.P., and Robertson,
S.A. (2007). Seminal plasma differentially regulates inflammatory cyto-
kine gene expression in human cervical and vaginal epithelial cells.
Mol. Hum. Reprod. 13, 491–501.
Sherer, N.M., Lehmann, M.J., Jimenez-Soto, L.F., Ingmundson, A.,
Horner, S.M., Cicchetti, G., Allen, P.G., Pypaert, M., Cunningham,
J.M., and Mothes, W. (2003). Visualization of retroviral replication in
Tachet, A., Dulioust, E., Salmon, D., De Almeida, M., Rivalland, S., Fin-
kielsztejn, L., Heard, I., Jouannet, P., Sicard, D., and Rouzioux, C.
(1999). Detection and quantification of HIV-1 in semen: identification
ofasubpopulationofmenathighpotentialrisk ofviralsexual transmis-
sion. AIDS 13, 823–831.
Termussi, P.A., Masino, L., and Pastore, A. (2003). From Alzheimer to
Huntington:whyis astructural understandingsodifficult? EMBO J.22,
Tevi-Benissan, C., Belec, L., Levy, M., Schneider-Fauveau, V., Si Mo-
hamed, A., Hallouin, M.C., Matta, M., and Gresenguet, G. (1997). In
vivo semen-associated pH neutralization of cervicovaginal secretions.
Clin. Diagn. Lab. Immunol. 4, 367–374.
Thomas, J.A., Ott, D.E., and Gorelick, R.J. (2007). Efficiency of human
immunodeficiency virus type 1 postentry infection processes: evi-
dence against disproportionate numbers of defective virions. J. Virol.
Walsh, D.M., Hartley, D.M., Kusumoto, Y., Fezoui, Y., Condron, M.M.,
Lomakin, A., Benedek, G.B., Selkoe, D.J., and Teplow, D.B. (1999).
Amyloid beta-protein fibrillogenesis. Structure and biological activity
of protofibrillar intermediates. J. Biol. Chem. 274, 25945–25952.
Westermark, P. (2005). Aspects on human amyloid forms and their
fibril polypeptides. FEBS J. 272, 5942–5949.
D.I., Sodroski, J., and Mirzabekov, T. (2002). Stimulation of enveloped
virus infection by b-amyloid fibrils. J. Biol. Chem. 277, 35019–35024.
Zhang, H., Dornadula, G., Beumont, M., Livornese, L., Jr., Van Uitert,
B., Henning, K., and Pomerantz, R.J. (1998). Human immunodefi-
ciency virus type 1 in the semen of men receiving highly active antire-
troviral therapy. N. Engl. J. Med. 339, 1803–1809.
Cell 131, 1059–1071, December 14, 2007 ª2007 Elsevier Inc. 1071