Previous Article | Next Article ![]()
Journal of Virology, April 2003, p. 4985-4991, Vol. 77, No. 8
0022-538X/03/$08.00+0 DOI: 10.1128/JVI.77.8.4985-4991.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Department of Microbiology and Immunology, University of Illinois at Chicago, Chicago, Illinois
Received 18 October 2002/ Accepted 13 January 2003
|
|
|---|
|
|
|---|
The need for these cooperative interactions has brought about considerable debate regarding the extent of association between the CD4 and chemokine receptor molecules in the target cell membrane. Two recent studies have found that the addition of purified gp120 to target cell cultures induces actin-dependent colocalization of CD4 and chemokine receptor in the membrane (11, 16). Others have described a constitutive association between CD4 and chemokine receptors in complexes (22), lipid rafts (16, 18), or specific cell types (13). To gain additional insight into the membrane localization and association of these proteins, we conducted immunofluorescent analysis of CD4 and CCR5 by high-resolution deconvolution microscopy. Deconvolution microscopy permits detailed, quantitative analysis of biological materials. Deconvolution uses computer processing to compensate for distortion caused by the optical path. Through the observation of fluorescent beads, an algorithm is developed which corrects the distortion caused by lenses and removes out-of-focus light. In this way, it is possible to remove the out-of-focus light, like a confocal microscope would do, but without the noise that is a consequence of photomultiplier tubes necessary for confocal detection. Therefore, deconvolution microscopy reveals unparalleled details of cell substructures using fluorescent microscopy. To avoid the dilemma of fixation artifacts, fusion proteins were used, allowing localization to be observed in both live and fixed cells. Fluorescent constructs were created by fusing yellow fluorescent protein (YFP) to the cytoplasmic tail of human CD4 (CD4-YFP) in plasmid EYFP-N1 (Clontech, Palo Alto, Calif.) and cyan fluorescent protein (CFP) or green fluorescent protein (GFP) to the cytoplasmic tail of human CCR5 (CCR5-CFP and CCR5-GFP [4], respectively, in plasmid pECFP-N1 or pEGFP-N1 [Clontech]).
Initial studies were focused on characterization and validation of the fusion proteins to ensure that they were characteristically and functionally similar to the native proteins. Figure 1 shows the localization of native CD4, CD4-YFP, and CCR5-CFP in transfected target cells as observed by deconvolution microscopy. After transfection of HeLa cells with a native CD4 construct, cells were fixed and stained with an anti-CD4 antibody (Sigma, St. Louis, Mo.) and then with a Cy3 secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, Pa.). Interestingly, we found that CD4 localization is not random in the plasma membrane. Rather, CD4 expressed on the cell surface was concentrated in protruding membrane structures (Fig. 1A). A similar experiment was conducted utilizing live cells expressing the CD4-YFP construct. Three-dimensional reconstruction (which allows rotation of images) revealed that CD4-YFP in living cells accumulated in cellular protrusions similar to those in which unfused CD4 was found in fixed cells (Fig. 1B). This is clear when the boxed region shown in the figure is rotated 45 or 90 degrees (Fig. 1B, left panels). The resulting side view reveals that the projections are sticking out of the cell surface. In some cases, accumulation was also observed at sites of cell-cell contact. Comparably, GFP- and CFP-conjugated CCR5 had a similar localization in the membranes of fixed cells (Fig. 1C and E).
![]() View larger version (98K): [in a new window] |
FIG. 1. Characterization of fluorescent fusion proteins. (A) Selective distribution of CD4 in membrane protrusions in transfected HeLa cells. Cells were fixed and stained for CD4. (B) Left, a similar distribution of CD4-YFP was observed during live cell imaging of transfected HeLa cells. Bright spots toward the center of the cell depict CD4 in intracellular vesicles. The white box indicates the specific area of interest depicted at the right. Right, volume projection depicts 0-, 45-, and 90-degree rotation of the boxed image, demonstrating clear localization of CD4 in cellular protrusions. (C) Fixed cell imaging of 293T cells expressing CCR5-GFP demonstrated localization of CCR5 in similar membrane protrusions. (D) Rapid internalization of CD4-YFP was observed after the addition of 50 ng of PMA per ml to transfected HeLa cells. Cells were imaged before (left) and 20 min after (right) PMA treatment. (E) Ligand-specific internalization was observed in 293T cells expressing CCR5-CFP. Cells were imaged before (left) and 1 h after (right) the addition of 100 nM AOP-RANTES. All cells were imaged using deconvolution microscopy and are shown as volume projections (created with API Softworx software). Bars, 15 µm.
|
The ability of the receptor and coreceptor fusion proteins to function as their native counterparts in HIV entry was confirmed using a ß-galactosidase assay for infection. HeLa cells containing the HIV long terminal repeat linked to a ß-galactosidase reporter gene were transfected with CD4-YFP or CCR5-CFP. At 2 days posttransfection, the cells were infected with appropriate strains of virus for 2 h. Unbound virus was washed, and cells were cultured for 48 h prior to assay. Cells were then fixed and stained, and infected foci were counted using light microscopy. Substantial infection was detected in cultures expressing both CD4 and CCR5 fluorescent fusion constructs, while no infection was observed in nontransfected control cultures (data not shown). The CCR5-GFP derivative used here has been previously reported to be functional for HIV infection (4).
Thus, both CD4 and CCR5 fluorescent fusion constructs looked and behaved identically to what would be expected for the native proteins, as demonstrated by the pattern of surface expression, response to appropriate ligand stimulation, and functional behavior as receptor and coreceptor. However, we observed that both CD4 and CCR5 accumulated in certain cell surface structures. Therefore, further investigation was conducted to determine the nature of these membrane protrusions.
Primary T cells were fixed and stained for actin and CD4 using Oregon Green phalloidin (Molecular Probes, Eugene, Oreg.), anti-CD4 from the SIM.4 hybridoma (provided by James Hildreth, NIH AIDS Research and Reference Reagent Program, National Institute of Allergy and Infectious Diseases [NIAID]), and a Cy3 secondary antibody (Jackson ImmunoResearch Laboratories). Image analysis revealed that much of the CD4 (red) accumulated at sites of actin staining (green) (Fig. 2A). The CD4 that was not associated with actin could be sequestered in intracellular vesicles, which appear to be coincident with the plasma membrane in primary T cells because of their small cytoplasm. Figure 2A shows staining of actin alone, CD4 alone, and CD4 and actin together, with Hoechst staining for cell nuclei. Similar results were observed when phytohemagglutinin-activated T cells were stained for actin (green) and CCR5 (red) (antibodies from R&D Systems were provided by NIH AIDS Research and Reference Reagent Program, NIAID) (Fig. 2B). Expression of CCR5 was heterogeneous, with many cells expressing little or no CCR5. Because it was somewhat difficult to obtain information regarding the colocalization of proteins due to the small size of the primary cells, additional experiments were conducted using various cell lines in order to gain a clearer perspective of CD4, CCR5, and actin localization. CHO K1 cells growing on glass coverslips were transfected with CD4-YFP, fixed, stained for actin by use of Texas Red phalloidin (Molecular Probes), and imaged in three dimensions using deconvolution microscopy. Analysis of individual image slices taken as part of the z-series revealed that CD4 was not constitutively colocalized with actin. Figure 2C shows a representative image taken toward the bottom of the cell (proximal to the interface between the coverslip and the cell membrane), where CD4 localization (green) is clearly separate from the actin stress fibers (red). A better perspective can be gained when looking closely at actin and CD4 staining individually in the same section of the cell, which is also shown in Fig. 2C (bottom panels).
![]() View larger version (40K): [in a new window] |
FIG. 2. CD4 and CCR5 localize to actin-dependent structures. (A) Colocalization of CD4 and actin was imaged in resting primary T cells that were adhered onto coverslips by use of poly-L-lysine for 5 min, fixed, and stained for actin (left) with phalloidin, CD4 (middle) with anti-CD4 antibody, and cell nuclei (merged image, right) with Hoechst stain. (B) Activated T cells were stained for actin (left), surface CCR5 (middle), or cell nuclei (merged image, right). (C) Representative image from a z-series of CHO K1 cells transfected with CD4-YFP and stained for actin taken proximal to the coverslip, where actin stress fibers predominate. The boxed region indicates the area of interest shown in detail at the bottom. Bottom, individual imaging of actin (left) and CD4 (right) revealed that CD4 is not constitutively associated with cellular actin. (D) An alternative plane from the same z-series taken distal to the coverslip interface shows CD4 and actin colocalization in membrane protrusions. Arrows in the upper panel point to microvilli, while arrows in the lower panel point to ruffles. Bars, 5 µm. A movie showing the z-series, including Fig. 2C and D, is available at http://www.uic.edu/depts/mcmi/hope.
|
293T cells were transfected with CD4-YFP, and 48 h later they were fixed and stained for actin by use of Texas Red phalloidin (Molecular Probes) and for ezrin by use of a monoclonal antibody against ezrin (Cytoskeleton, Inc., Denver, Colo.) and a Cy5 secondary antibody (Jackson ImmunoResearch Laboratories). As shown in Fig. 3A, all three proteins accumulated in the same membrane structures in fixed cells. When a closer view is taken of the boxed region shown in Fig. 3A, extensive CD4 (green), actin (red), and ezrin (blue) colocalization can be observed (white). The separated color images of CD4, actin, and ezrin demonstrate the localization of these proteins in a ruffling area of the membrane (Fig. 3B). Similar imaging studies were performed to analyze CCR5-GFP and ezrin expression. Considerable colocalization of CCR5-GFP (green) and ezrin (red) can be observed in apparent microvilli (Fig. 3C and D). Additional individual analysis of CCR5-GFP and ezrin demonstrated enrichment of these proteins in the same membrane protrusions (Fig. 3D). When this image is rotated 90 degrees on its side, obvious colocalization of CCR5 and ezrin is observed in microvilli and other membrane structures (Fig. 3E).
![]() View larger version (29K): [in a new window] |
FIG. 3. CD4 and CCR5 accumulate in ezrin-enriched structures. (A) 293T cells transfected with CD4-YFP were fixed and stained for actin and ezrin. Fixed cell imaging shows CD4 (green), actin (red), and ezrin (blue) colocalization in membrane ruffles and protrusions in the volume projection of a z-series. The white box depicts the area of interest shown in panel B. (B) A single plane of the series from the boxed area of the membrane is shown in greater detail. The single plane is distal to the coverslip and focuses on an area of extensive membrane ruffling. Combined expression (white) in the area of interest is shown as well as expression of each individual protein. (C) 293T cells transfected with CCR5-GFP (green) were fixed, permeabilized, stained for ezrin (red), and shown as a volume projection of a z-series. The white box indicates the area of interest shown in panels D and E. (D) A volume projection of the region proximal to the coverslip included in the area of interest is shown in greater detail, with areas of CCR5 and ezrin colocalization shown in yellow (top), CCR5 shown in green (middle), and ezrin shown in red (bottom). (E) Volume projection of area of interest rotated 90 degrees to demonstrate localization of CCR5 and ezrin in membrane protrusions. Bars, 5 µm. A movie showing the rotation of Fig. 3D is available at http://www.uic.edu/depts/mcmi/hope.
|
![]() View larger version (58K): [in a new window] |
FIG. 4. CD4 and CCR5 colocalize to cellular structures but do not exist as a complex in the cell membrane. (A) Colocalization of receptor and coreceptor molecules was observed in a living 293T cell expressing both CD4-YFP (green) and CCR5-CFP (blue). The white box depicts an area of interest in the membrane with microvilli and ruffles. Expression of each individual protein is shown in greater detail at upper left. (B to D) 293T cells coexpressing CD4-YFP (green) and CCR5-CFP (blue) were imaged before the addition of AOP-RANTES or PMA (B), 1 h after the addition of 100 nM AOP-RANTES (C), and 20 min after the addition of 50 ng of PMA per ml (added after 1 h of incubation with AOP-RANTES) (D). White boxes indicate areas of interest shown in greater detail at the bottom of each panel, demonstrating the individual ligand-specific internalization of each protein. All images in this figure are volume projections. Bars, 15 µm. White arrows, areas in the cell membrane where membrane-associated CCR5-CFP expression (B) or loss of expression (C and D) can be observed after addition of AOP-RANTES.
|
In summary, the localization of the HIV receptor CD4 and coreceptor CCR5 in the cell membrane is dynamic. However, visualization of fluorescent fusion proteins using high-resolution deconvolution microscopy demonstrated that these proteins tend to accumulate in actin-dependent, ezrin-enriched cellular structures. Additional ligand-specific internalization studies revealed that CD4 and CCR5 are not constitutively associated with one another. Our findings may explain the discrepancies reported in previous studies, as these structures are inherently dynamic and are influenced by variations in cellular activity and culture conditions. For example, the accumulation of CD4 and CCR5 in ezrin-containing structures is consistent with reports that cross-linking with gp120 induces colocalization, often observed in membrane structures (11, 16). Cross-linking of receptor and coreceptor induces signaling (21) which may result in the formation of ezrin-containing structures and hence the observed recruitment. Interestingly, CXCR4 has also been reported to be recruited (11), suggesting that CXCR4 may behave similarly to CCR5 on the cell surface. Thus, the preferential localization of both CD4 and CCR5 in actin-dependent structures and at sites of cell-cell contact may have broad implications for how HIV interacts with receptor and coreceptor to generate a potential site of fusion pore formation. Further investigation is necessary to determine the impact of protein localization and mobility on the ability of target cells expressing both receptor and coreceptor to be infected by HIV.
This work was supported by amfAR fellowship 70572-31-RF to C.M.S. and by NIH grants RO1 AI47770 and R21 AI052051 to T.J.H.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»