Functions of Early (AP-2) and Late (AIP1/ALIX) Endocytic Proteins in Equine Infectious Anemia Virus Budding*

The proline-rich L domains of human immunodeficiency virus 1 (HIV-1) and other retroviruses interact with late endocytic proteins during virion assembly and budding. In contrast, the YPDL L domain of equine infectious anemia virus (EIAV) is apparently unique in its reported ability to interact both with the μ2 subunit of the AP-2 adaptor protein complex and with ALG-2-interacting protein 1 (AIP1/Alix) protein factors involved in early and late endosome formation, respectively. To define further the mechanisms by which EIAV adapts vesicle trafficking machinery to facilitate virion production, we have examined the specificity of EIAV p9 binding to endocytic factors and the effects on virion production of alterations in early and late endocytic protein expression. The results of these studies demonstrated that (i) an ∼300-residue region of AIP1/Alix-(409-715) was sufficient for binding to the EIAV YPDL motif; (ii) overexpression of AIP1/Alix or AP-2 μ2 subunit specifically inhibited YPDL-mediated EIAV budding; (iii) virion budding from a replication-competent EIAV variant with its L domain replaced by the HIV PTAP sequence was inhibited by wild type or mutant μ2 to a level similar to that observed when a dominant-negative mutant of Tsg101 was expressed; and (iv) overexpression or siRNA silencing of AIP1/Alix and AP-2 revealed additive suppression of YPDL-mediated EIAV budding. Taken together, these results indicated that both early and late endocytic proteins facilitate EIAV production mediated by either YPDL or PTAP L domains, suggesting a comprehensive involvement of endocytic factors in retroviral assembly and budding that can be accessed by distinct L domain specificities.

Equine infectious anemia virus (EIAV) 2 is a member of the lentivirus subfamily of retroviruses, which also includes human immunodeficiency virus 1 (HIV-1). The EIAV genome is the simplest among the lentivirus family and encodes three major structural proteins (Gag, Pol, and Env) along with three accessory gene products (Tat, Rev, S2). Like all retroviruses, EIAV Gag protein is synthesized as a polyprotein that upon virion maturation is processed by virus-encoded protease to yield four major structural proteins: matrix, capsid, nucleocapsid, and p9 protein. All four structural proteins of Gag polyprotein play critical and distinctive roles in retrovirus assembly and budding. Matrix proteins direct targeting of Gag polyproteins to cell membranes for viral assembly and budding (1,2). Both capsid and nucleocapsid have been shown to mediate Gag-Gag interactions during viral assembly, with capsid responsible for the dimerization of homologous Gag polyproteins (3,4) and nucleocapsid essential for multimerization of Gag molecules (5,6). The EIAV p9 protein contains a YPDL late (L) domain that is critical for virion release during the late stage of virus budding. Proline-rich L domains such as PTAP and PPPY with similar functions have also been identified from HIV-1, Rous sarcoma virus, and a variety of other enveloped viruses (7).
L domains appear to serve as docking sites to recruit cellular factors, providing essential functions for virion release. The HIV-1 PTAP motif specifically interacts with the N terminus of tumor suppression gene 101 products (Tsg101), an essential component of the ESCRT I complex responsible for multivesicular body (MVB) formation (8,9). The PPPY motif binds to ubiquitin ligase Nedd4 (10,11). Interestingly, the EIAV YPDL motif has been shown to interact with two cellular proteins, ALG-2-interacting-protein-1 (AIP1/Alix) (AIP1 is used hereinafter) (12,13) and the 2 subunit of the AP-2 adaptor protein complex (14). Although these L domains appear to interact with different cellular proteins, certain functional interchangeability was previously reported. For example, both PTAP and PPPY motifs can substitute for the YPDL domain to support EIAV replication (15). The PTAP function for HIV-1 budding and replication, however, can be replaced by the PPPY but not the YPDL motif (16). These observations indicate that retroviruses, along with other enveloped viruses, may have evolved different L domains to exploit redundant cellular endocytic machineries to achieve virus budding and release.
The L domain-interacting cellular proteins identified to date are endocytic components associated with the formation of either early endosomes (e.g. ubiquitin ligase, AP-2) or late MVBs (e.g. Tsg101, AIP1). Although it is currently assumed that MVBs play a central role in retrovirus budding, it remains uncertain whether and how components of other endosomal vesicles, including early endosomes, contribute to retroviral budding. It is reported that HLTV-1 Gag polyproteins interact first with Nedd4.1 at the plasma membrane and then with Tsg101 in late endosomes/MVB, suggesting a successive assembly and budding process through the endocytic pathway (17,18). However, neither early nor late endosomes are optimally evolved for viral assembly and release. For example, the formation of early endosomes is topologically opposite that of viral budding, whereas budding into late endosome is topologically correct but imposes subsequent obstacles to virion release from infected cells. Therefore, it remains to be determined whether and how retroviruses combine and/or integrate early and late endosomal machinery to facilitate virion assembly and budding. In the light of previous observations that the EIAV YPDL L domain interacts with both AP-2 and AIP1 in different model systems, the current study was designed to examine the specificity of EIAV p9 binding to early and late endocytic factors and the effects of alterations in early and late endocytic protein expression on virion production. The results of these studies demonstrated a previously unrecognized involvement of both early and late endocytic machinery in EIAV virion assembly and budding, whether accessed by YPDL or PTAP L domains.
Cell Culture and Transfection-COS-1 cells were cultured in minimum essential medium as described previously (20). HeLa and 293T cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum. GenePorter II (Gene Therapy Systems, San Diego, CA) or FuGENE 6 (Roche Diagnostics) was used to transfect cells according to the manufacturer's recommendation. SiRNA duplexes were synthesized in vitro and transfected into HeLa cells using GeneEraser TM (Stratagene, La Jolla, CA).
Yeast Two-hybrid Analysis-The Saccharomyces cerevisiae strain used for two-hybrid studies was CTY10.5d (Mata ade2-101 his3-⌬200leu2-⌬1 trp1-⌬901gal4 gal80 URA3::lexAop-lacZ). Yeast cells were grown in synthetic dextrose medium lacking appropriate supplements to maintain selection for plasmids. ␤-Galactosidase activity was assayed on permeabilized cells grown to mid-log phase in the selective synthetic dextrose medium and expressed in Miller units. Yeast protein extracts for immunoblot analysis were prepared as described previously (21) and analyzed by immunoblotting with a monoclonal anti-HA antibody (Roche Diagnostics).
Reverse Transcriptase (RT) and Western Blotting Assays-To measure virion production, culture medium collected from 24 to 48 h after transfection was clarified of cell debris by low speed centrifugation and then centrifuged at 20,800 ϫ g for 3 h at 4°C to pellet virions. Virion pellets were resuspended in 50 l of phosphate-buffered saline, and 10-l samples were assayed in RT assays as described previously (20).
To examine EIAV-specific protein content in transfected cells and pelleted virions, total cell lysates prepared as described previously (20) and pelleted virions were resolved by electrophoresis through 4 -15% gradient gels (Bio-Rad) and immunoblotted using a reference immune serum from a field-infected horse (20). Horseradish peroxidase-conjugated goat anti-horse IgG F(abЈ) 2 (Jackson ImmunoResearch, West Grove, PA) was used as secondary antibody. The immunoblots were developed by incubation with SuperSignal West Pico chemiluminescent substrate (Pierce). Horseradish peroxidase-conjugated high affinity anti-HA antibody (Roche Applied Science) was used to specifically detect HA-tagged proteins.

RESULTS
Interaction between AIP1 and EIAV p9 Proteins-Interaction between AIP1 and EIAV p9 was previously demonstrated using yeast two-hybrid and GST pull-down assays (12,13,23). To serve as a starting point for a comprehensive comparison, we initially examined the p9-AIP1 interaction using similar approaches. We previously demonstrated interaction between the 2 subunit of the AP-2 adaptor protein complex and EIAV p9 using GST pull-down assay (24). In that study, a recombinant protein containing only the N-terminal 30 amino acids of EIAV p9 was fused to GST, in contrast to the full-length p9 protein (51 residues) used in studies of p9-AIP1 binding (12,13,23). Therefore, we performed p9-AIP1 binding assays using the N terminus of p9 (designated as *p9) in this study. Fig. 1 summarizes the yeast two-hybrid results using *p9 as bait to examine binding activities of full-length and two truncated AIP1 proteins. Full-length AIP1 bound to *p9 as expected (Fig. 1A), and interaction of two truncated fragments (AIP1 372-869 and AIP1 409 -715 ) with *p9 was also detected. The stronger interaction of full-length AIP1 was not due to higher expression of the fusion protein as shown by Western blot analysis (Fig. 1B). The observed binding properties are consistent with previous studies showing that both the N terminus (1-175) and the C-terminal proline-rich domain (717-869) of AIP1 are not required for this interaction (12,13,23). Additionally, the current studies further defined the p9 binding domain of AIP1 to residues 409-715. It is interesting to note that the EIAV p9 binding domain of AIP1 (residues 409-715) does not contain the PSAP motif (AIP1 717-720 ) involved in interaction with Tsg101 (13).
To compare the binding properties of AIP1 proteins and the 2 subunit in parallel, we performed GST pull-down assays using GST-*p9 and the AIP1 409 -715 construct. Our data revealed that similar levels of GST fusion protein were produced from each construct ( Fig. 2A) and that *p9 evidently bound to the AIP1 409 -715 . Single alanine substitutions in the EIAV L domain at residues Tyr-23, Pro-24, or Leu-26 abolished detectable AIP1 409 -715 binding, suggesting their critical role for AIP1 interaction, whereas mutation of D25A reduced AIP1 409 -715 binding by about DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 70% when compared with *p9. The D25A substitution in p9 has previously been reported to have no effect on EIAV viral budding (24). Our data confirmed that the YPDL sequence is a critical determinant of AIP1-p9 interaction, consistent with other reports (12,13,23). Interestingly, the YPDL sequence specificity for AIP1 binding reported here is similar to that for p9 binding to 2 that we described previously (24). The *p6 protein (SRPEPTAPPEES) did not bind to AIP1 409 -715 as expected due to the absence of the p6 LYP motif responsible for AIP1 interaction in this construct. The observation that RSV *p2b failed to bind to AIP1 409 -715 further confirmed the specificity of AIP1 interaction. Taken together, these data demonstrated that interaction of EIAV p9 and AIP1 is dependent on the p9 YPDL sequence and on a small region (residues 409 -715) of the AIP1 protein. Thus, it appeared that in vitro, both AIP1 and 2 equally interact with EIAV p9 protein via the YPDL sequence, suggesting possible interactions with early and late endocytic proteins in infected cells.

AP-2 and AIP1/Alix in EIAV Budding
Effects of AIP1 Proteins on EIAV Budding-Previous studies of AIP1 interaction with the YPDL sequence of EIAV p9 protein were performed using Gag expression vectors and not with full-length EIAV proviral constructs. To examine the functional role of AIP1-p9 interactions in virion production in the context of complete proviral expression, we co-transfected COS-1 cells with an EIAV proviral construct and AIP1 expression vectors. As a specificity control, we also included an expression vector encoding the N-terminal moiety of Tsg101 1-204 that contains the HIV-1 PTAP binding domain (25). As summarized in Fig. 3, EIAV production was reduced by 50 -70% from cells expressing AIP1 proteins when compared with cells transfected with vector DNA (Fig. 3A). Various AIP1 proteins differed slightly in their inhibitory effect on EIAV budding, with the full-length AIP1 apparently the most potent inhibitor despite the relatively low levels of full-length AIP1 detected in transfected cells. Another interesting observation is that co-expression of Tsg101 1-204 reproducibly enhanced EIAV production by an average of 30 -40% when compared with the vector control. Overexpression of AIP1 proteins and Tsg101 1-204 seemed not to significantly affect the steady state levels of EIAV-specific protein as shown in Fig. 3B  (upper panel). Thus, these data demonstrated that EIAV budding mediated by the YPDL domain is specifically inhibited by AIP1 protein overexpression.
To further verify the specificity of the interactions observed above, we next examined the presence of HA-tagged AIP1 proteins in pelleted EIAV virions. Our data showed that all three AIP1 proteins as well as Tsg101 were detected in pelleted virions (Fig. 3B, lower panel). To exclude the possibility that overexpression of these endocytic proteins results in their appearance in the pelletable fraction of the culture medium, we transfected COS-1 cells with the various endocytic protein  expression vectors in the absence of EIAV proviral DNA and centrifuged the culture medium under conditions used to pellet EIAV virions. No AIP1 or Tsg101 could be detected by immunoblotting in these control pelleted preparations in the absence of virion production (data not shown). Therefore, our results further confirmed specific incorporation of AIP1 into EIAV virions. The incorporation of Tsg101 into EIAV virions that we observed also suggested its involvement in EIAV budding, perhaps through indirect interactions mediated by AIP1-Tsg101 interaction (13).
Effects of AIP1 and 2 Expression on EIAV Budding-Results from the preceding experiments and previous reports have confirmed that both AIP1 and 2 bind to the YPDL sequence of EIAV p9 protein in GST pull-down assays and that overexpression of AIP1 inhibits EIAV budding from transfected COS-1 cells. We next compared the roles of AIP1 and 2 in EIAV budding using wild type EIAV (CMVuk YPDL ) and a mutant EIAV (CMVuk PTAP ) with its YPDL motif replaced by the HIV-1-derived PTAP motif. AIP1 and 2 inhibited EIAV budding by 60 and 50%, respectively, as measured by pelleted virion-associated RT activity (Fig. 4A). The specificity of 2 inhibition was further confirmed by the observation that overexpression of a mutant 2 (2 DW ) that is unable to bind to EIAV p9 protein had no significant effect on EIAV production (Fig. 4A). Thus, both AIP1 and 2 contributed to EIAV budding.
Overexpression of Tsg101 1-204 enhanced CMV uk (YPDL) budding by about 30% (Fig. 4A) as observed in Fig. 3A, and budding of CMV uk (PTAP) was inhibited about 55% by overexpressed Tsg101 1-204 as expected (Fig. 4B). These two proviral constructs only differ from each other in the specificity of the L domain sequences, and both are previously demonstrated to be replication-competent (15). Our data indicated that Tsg101 is involved in both PTAP-and YPDL-mediated retrovirus budding but probably through different interaction mechanisms. Interestingly, expression of both wild type and mutant 2 reduced CMV uk (PTAP) budding by about 80% (Fig. 4B), indicating that the AP-2 adaptor protein complex might also participate in PTAP-mediated budding, although direct interaction with the PTAP motif may not be required.
The involvement of both early and late endocytic proteins in retrovirus budding is also supported by the observation that these proteins could be co-pelleted with the EIAV virion (Fig. 4, C and D). It appeared that there was more Tsg101 1-204 and AIP1 409 -715 incorporated into pelleted EIAV virions when compared with the levels of 2 subunit. There are several possible explanations for the observed differences. Although all proteins were detected via their HA epitope tags, the positions of these tags differ (N-terminal for Tsg101 and AIP1, internal for 2), possibly affecting the level of antibody reactivity of each protein to the anti-HA conjugate used in our immunoblots. If equal reactivity is assumed, different incorporation of these proteins might be the result of varied levels of these proteins expressed in the transfected cells (cell lysate data not shown). In particular, AP-2 expression is under such a tight regulation that loss of one subunit leads to complete depletion of AP-2 (26), and robust expression of AP-2 complex is difficult to achieve by overexpressing only one subunit of the complex. Taken together, our results demonstrated that both AIP1 and the 2 subunit of AP-2 contribute to YPDL-mediated EIAV budding and that AP-2 also participates in PTAP-mediated budding.
Combined Effect of AIP1 and 2 Overexpression on EIAV Budding-As both AIP1 409 -715 and 2 overexpression inhibited EIAV budding, we sought to determine whether they function independently or in cooperation. For this purpose, we individually titrated the minimum amounts of AIP1 409 -715 or 2 expression vector for detectable inhibitory effect (data not shown). Based on these assays, we cotransfected 293T cells using 0.8 g of 2 and 0.2 g of AIP1 409 -715 , individually or in combination, whereas keeping the total amount of DNA constant with blank vector plus EIAV proviral DNA. Virions released into the culture medium during the period from 24 to 48 h after transfection were pelleted, and RT activity associated with the resuspended pellet was measured to quantify EIAV budding (Fig.  5). Co-expression of 2 and AIP1 409 -715 inhibited EIAV budding more than either protein expressed individually (80% versus 40 and 50% inhibition, respectively), indicating an additive effect of AIP1 and 2 and further supporting the concept that both AIP1 and 2 contribute to EIAV budding. To test whether the combined effect was cell type-dependent, we also carried out the same experiment in COS-1 cells and found similar results (Fig.  5B). Incorporation of AIP1 409 -715 and 2 was also detected in pelleted EIAV virions (Fig. 5C).  DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 siRNA and shRNA Silencing of 2 and AIP1 on EIAV Budding-It has been cautioned that overexpression of certain proteins might nonspecifically affect global cellular functions rather than specifically disturbing the precise cellular pathway under investigation. To complement our overexpression study on the roles of 2 and AIP1 in EIAV budding, we examined the effects of 2 and AIP1 down-regulation on EIAV budding. Fraile-Ramos et al. (26) previously tested specific siRNA sequences against the 2 subunit of AP-2 and demonstrated an average 50% reduction in 2 expression in HeLa cells at 48 h after transfection. Based on the same siRNA sequence, we generated a pSilencer-derived DNA construct that produces small hairpin RNA (shRNA2) upon transfection, which is spontaneously processed into siRNA in the transfected cells. The silencing effect of this construct on COS-1 cells cotransfected with HA-tagged 2 was examined (Fig. 6A). Consistent with the results of Fraile-Ramos et al. (26), shRNA2 reduced 2 expression by about 50% at 48 h after transfection. To knock down AIP1 expression, we cloned several sequences targeting various regions of AIP1 into the pSilencer vector and identified a specific sequence (5Јgctcaagatggtgtgataa-3Ј) to achieve efficient silencing (Fig. 6B). At 48 h after co-transfection, expression of full-length AIP1 was reduced by about 80% and expression of AIP1 409 -715 was reduced by about 70%, respectively.

AP-2 and AIP1/Alix in EIAV Budding
Our results revealed that 70 -80% of AIP1 silencing could be achieved by shRNA AIP1, but only about 50% of 2 silencing was observed using either siRNA or shRNA at 48 h after transfection. The long half-life of 2 subunit (ϳ24 h) (27) might partially account for the relative inefficiency of knockdown procedures. Therefore, to more rigorously suppress AP-2 expression, assays at later time points (72-96 h after transfection) or two consecutive transfections were used to achieve higher levels of 2 silencing (28). We therefore designed our experiments to first transfect 293T cells with siRNA duplex at day 0, and at day 2, cells were split, reseeded, and incubated for ϳ12 h prior to transfection using the EIAV proviral DNA with or without pSilencer vectors encoding the specific shRNA. RT activity in culture medium collected from 24 to 48 h after the second transfection was quantitated as a measure of EIAV budding. In Fig. 6C, we first transfected 293T cells with increasing concentrations of 2 siRNA and used only EIAV proviral DNA during the second transfection. Under these conditions, EIAV budding was inhib- along with plasmids encoding for the indicated proteins at a 1:2 mass ratio using FuGENE 6. Culture medium from 24 to 48 h after transfection was collected. RT activity of pelleted virions was assayed to represent virion production. Data represent means Ϯ ranges for duplicate samples from one of two independent experiments. Approximately equal amounts of resuspended virions normalized by RT activity were also subjected to Western blot to examine the content of HA-tagged endocytic proteins associated with the pelleted virions produced from CMV uk (YPDL) (C) or CMV uk (PTAP) (D) transfection. FIGURE 5. Additive effect of overexpressed 2 and AIP1 409 -712 proteins on EIAV budding. 293T (A) and COS-1 (B) cells grown on 6-well plates were co-transfected with 1 g of CMV uk proviral DNA and a total of 1 g of indicated plasmids using FuGENE 6. Culture medium from 24 to 48 h after transfection was collected to pellet virions. RT activity of 10 l of resuspended virions was measured to represent virion production. Data represent means Ϯ ranges for duplicate samples from one of three independent experiments. C, approximately equal amounts of resuspended virions produced from the transfected COS-1 cells normalized by RT activity were also examined for the content of HA-tagged proteins.
ited by 2 siRNA in a dose-dependent manner (Fig. 6C), i.e. no inhibition at 0 nM, 40% inhibition at 10 nM, and 60% inhibition at 25 nM 2 siRNA. To assay the effect of the specific shRNA encoded by the pSilencers, we mock-transfected the 293T cells with the GeneEraser TM . Following splitting and reseeding of the cells, we transfected them with EIAV proviral DNA along with the shRNA constructs. Our results in Fig. 6D showed that EIAV budding was reduced by about 15-20% by shRNA 2 and shRNA AIP1 , respectively, indicating that the shRNAs we generated affect EIAV budding as expected. In another experiment (Fig.  6E), 293T cells were first transfected with 10 nM 2 siRNA and then transfected with EIAV proviral DNA and pSilencer constructs at 60 h after the first transfection. Expression of shRNA 2 further reduced EIAV production by about 50% and shRNA AIP1 by about 60%, respectively. The inhibitory effect was cumulative for 2 and AIP1 downregulation, further confirming that both 2 and AIP1 are involved in EIAV budding.

DISCUSSION
It has become increasingly appreciated that retroviruses have evolved to adapt various host cellular machinery for their replication, and the number of identified cellular factors that interact with viral proteins is rapidly increasing (7, 29 -31). Recently, cellular proteins that specifically bind to retrovirus L domains have been extensively characterized (9,10,23,24), providing strong evidence to link retrovirus budding with vesicular trafficking. However, the detailed molecular and cellular mechanisms that underlie the adaptation of endocytic proteins by retroviruses to facilitate budding remain to be illustrated.
In the current study, we compared the functions of the 2 subunit of AP-2 and AIP1 in EIAV budding using complementary approaches. Our results revealed that disruption of 2 or AIP1 expression, either up-regulation by transient transfection or down-regulation by siRNA silencing, reduced EIAV budding from cells cotransfected with an EIAV proviral construct. Furthermore, the effects of 2 and AIP1 were cumulative, if not synergistic, as overexpression or knockdown of both resulted in more potent inhibition than that observed for either individual protein. Interestingly, our data demonstrated that HIV-1 PTAP L domain-mediated budding was also inhibited when 2 subunit was overexpressed in the same cells. Therefore, the current report has pro-vided evidence for a contribution of the AP-2 adaptor protein complex in both PTAP-and YPDL-mediated budding, in addition to the previously reported role of AIP1 in this process.
The 2 subunit of AP-2 is a critical component for the formation of early endosomes, and identification of its interaction with the EIAV L domain suggested the involvement of the early endocytic pathway in retrovirus budding (24). However, the precise role of early endocytic proteins in viral budding has been controversial as the topology of endocytosis is opposite to that of retrovirus budding. The past few years have seen a collection of evidence that AIP1, a cellular component involved in late endosome/MVB formation, interacts with the EIAV YPDL L domain and an LYP motif located downstream of the PTAP motif in HIV p6 (12,13,23). Along with the identification of Tsg101 (a component of the ESCRT I complex essential for MVB formation) as the binding partner of HIV PTAP motif, these results suggest that retroviruses can recruit late endocytic machinery for budding as both processes share the same topology. However, it remains unclear whether and how retroviruses bud into late endosomes before they exit host cells. The current results that both 2 and AIP1 contribute to YPDL-and PTAPmediated budding strongly indicated that retrovirus might actually combine both early and late endosomal machinery to facilitate budding. Consistent with this model, MLV Gag polyproteins have been shown to traffic between the plasma membrane and the late endosomes, although the functional significance of this trafficking remains to be defined (17).
Here we have reported that PTAP-mediated budding is also sensitive to 2 overexpression and that both wild type and mutant 2 had similar effects on budding. These data indicated that the AP-2 adaptor protein complex might contribute to retrovirus budding via a general mechanism that remains to be defined. It is also worth noting that when the 2 subunit was depleted using high concentrations of siRNA oligonucleotides (25 nM), no further reduction in EIAV budding was observed upon introduction of a second shRNA (shRNA2 or shRNA AIP1 ) by transfection (data not shown). These observations indicated that adaptation of 2 and AIP1 might be sequential such that complete blockade of one interaction abolishes the entire chain. Further characterizations of the apparently complex interactions between the Gag proteins of different retroviruses and endocytic proteins involved in assembly and budding can provide further insights into the exact role of the endocytic machin-FIGURE 6. Effect of siRNA and shRNA targeting 2 and AIP1 409 -715 on EIAV budding. An siRNA sequence known to down-regulate 2 expression (26) was used to generate pSilencer plasmid, denoted as shRNA2 here. A sequence (5Ј-gctcaagatggtgtgataa-3Ј) targeting to 1765-1783 nt downstream of AIP1 start codon was selected to make shRNAAIP1 pSilencer constructs. Down-regulation of 2 (A) and AIP1 (B) was verified in COS-1 cells transfected with the indicated plasmids. Cell lysates were prepared at 48 h after transfection. Approximately equal amounts of cell lysates (normalized to ␤-actin) were immunoblotted with anti-HA antibody to detect 2 and AIP1 proteins (arrows), respectively. C, the effect of 2 siRNA on EIAV budding. 293T cells were transfected with siRNA targeting 2 at the indicated concentrations. At 48 h after siRNA transfection, cells were split and reseeded and then incubated for another 12 h before they were transfected with EIAV proviral DNA. Culture medium from 24 to 48 h after the second transfection was collected to pellet virions, and RT activity of 10 l of resuspended virions was measured to represent virion production. D, to evaluate the effect of shRNA alone on EIAV budding, 293T cells were transfected with the indicated constructs along with the EIAV proviral DNA. Virion production was measured as in C to quantify budding efficiency. E, the effects of siRNA and shRNA on EIAV budding. The same protocol used in C was followed except that 293T cells were first transfected with 10 nM siRNA targeting 2, and EIAV proviral DNA along with the indicated shRNA constructs were included during the second transfection. Data represent means Ϯ ranges for duplicate samples from one of two independent experiments. DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 ery in Gag trafficking, assembly, and budding. In addition, the characterization of these critical Gag-endocytic protein interactions can provide new targets for the development of antiviral drugs.