A screen for endocytic motifs

Abstract

Sorting signals for cargo selection into coated vesicles are usually in the form of short linear motifs. Three motifs for clathrin-mediated endocytosis have been identified: YXXPhi, [D/E]XXXL[L/I] and FXNPXY. To search for new endocytic motifs, we made a library of CD8 chimeras with random sequences in their cytoplasmic tails, and used a novel fluorescence-activated cell sorting (FACS)-based assay to select for endocytosed constructs. Out of the five tails that were most efficiently internalized, only one was found to contain a conventional motif. Two contain dileucine-like sequences that appear to be variations on the [D/E]XXXL[L/I] motif. Another contains a novel internalization signal, YXXXPhiN, which is able to function in cells expressing a mutant mu2 that cannot bind YXXPhi, indicating that it is not a variation on the YXXPhi motif. Similar sequences are present in endogenous proteins, including a functional YXXXPhiN (in addition to a classical YXXPhi) in cytotoxic T-lymphocyte-associated protein 4 (CTLA-4). Thus, the repertoire of endocytic motifs is more extensive than the three well-characterized sorting signals.

Figure 1. A CD8 reporter-based system to study internalization signals

A) Cytoplasmic tail sequences of wtCD8, CD8 with the cytoplasmic tail of CIMPR, and the four ‘designer’ CD8 chimeras. Known trafficking motifs are indicated in colour. B) HeLa cells were transfected with the six constructs listed in (A), and the steady-state localization of the constructs was examined by immunofluorescence microscopy. Scale bar: 20 µm. C) Schematic representation of the assay used to measure the efficiency of internalization of CD8 chimeras. Cells are harvested 24 h post-transfection and allowed to internalize anti-CD8 conjugated to Alexa Fluor 488 for 40 min at 37°C. The cells are then shifted to 4°C and surface Alexa Fluor 488 is quenched with an antibody against the fluorescent dye. Surface anti-CD8 is also relabelled with a Cy5-conjugated anti-mouse IgG. The cells are analysed by flow cytometry. D) FACS profiles of cells transiently transfected with the six CD8 constructs.

Figure 2. A CD8 library-based approach to identify novel endocytic motifs

A) Schematic representation of the approach. (1) Size-selected fragments of yeast genomic DNA are ligated into a CD8-encoding plasmid and transformed into bacteria to generate the library. (2) HeLa cells are transfected with the library and stably transfected cells are selected with G418. (3) The stable transfectants are pooled and endocytosis efficiency is assayed by FACS (see Figure 1). Cells with a high green-to-red fluorescence ratio (green gate, indicated by a plus sign) and with a low green-to-red fluorescence ratio (red gate, indicated by a minus sign) are collected, and tail sequences are determined using a cell-based PCR method. B) Steady-state distribution of the constructs in the two populations of cells. Scale bar: 20 µm.

Figure 3. CD8 chimeras with internalization signals isolated from the library

A) Sequences of the five cytoplasmic tail inserts that promoted the most efficient internalization of anti-CD8. Vector sequences are shown in italics. The nomenclature refers to the first three residues of the insert. B) Steady-state localization of transiently transfected constructs. Scale bar: 20 µm. C) To determine the efficiency of internalization of the library constructs, the antibody uptake assay was carried out on transiently transfected cells. D) Bar graphs showing the amount of internalized antibody (mean fluorescence of Alexa Fluor 488) for cells expressing equal amounts of surface CD8 [gated as shown in (C) for CD8-STS].

Figure 4. Mutational analysis of CD8-PIA and CD8-PNT suggests that they contain atypical dileucine motifs

A) To identify the residues in CD8-PIA that mediate its internalization, the amino acids in the insert were mutated in groups of three or four. The mutated constructs were expressed in HeLa cells and the antibody uptake assay was carried out. B) Mutational analysis of CD8-PNT. To determine which residues from the vector sequence (in italics) contribute to the internalization of CD8-PIA, C-terminally truncated constructs were studied. Alanine-scanning mutagenesis was then carried out on the full-length tail. C) Internalization of designer constructs with variations on acidic dileucine motifs.

Figure 5. Mutational analysis of CD8-WPK

Alanine-scanning mutagenesis was carried out to identify the residues involved in the internalization of CD8-WPK.

Figure 6. CD8-RYR contains a novel tyrosine-based motif

A) Alanine-scanning mutagenesis was carried out to identify the residues involved in the internalization of CD8-RYR. The most important residues are Y2, L6 and N7. B) Endocytosis of CD8-RYR in cells expressing µ2 with a mutation (F174A/D176S) in the tyrosine-binding pocket. Wild-type HeLa cells, or HeLa cells expressing siRNA-resistant µ2 F174A/D176S, were depleted of endogenous µ2, transfected with the CD8 constructs and processed for the antibody uptake assay. The mutant µ2 rescues the uptake of CD8-EAAALL and CD8-RYR, but not of CD8-YAAL.

Figure 7. Sorting signals in endogenous proteins

A) To search for trafficking motifs in naturally occurring proteins, we generated a database of 4794 cytoplasmic tail sequences of single-pass transmembrane proteins (1266 type I and 3528 type II proteins). We then searched the database for tails from type I proteins that contain known endocytic motifs (YXXΦ, [D/E]XXXL[L/I], FXNPXY) or our candidate novel endocytic motif (YXXXΦN), as well as motifs recognized by GGAs and COPI (DXXLL and KKXX, respectively). The table lists the frequency of the various sorting signals. The same sorting signals are shown on both the x- and y-axis. Where a sorting signal on the x-axis intersects with the same sorting signal on the y-axis, the number in the (blue) box is the number of proteins with at least one copy of that motif. Where a sorting signal on the x-axis intersects with a different sorting signal on the y-axis, the number in the box is the number of proteins that have both motifs (e.g. there are 143 proteins with both YXXΦ and [D/E]XXXL[L/I] motifs). The red numbers indicate how many of the nine tails with the post-Golgi motif, DXXLL, or of the seven tails with the ER-retention motif, KKXX, also contain internalization signals (YXXΦ, [D/E]XXXL[L/I], FXNPXY or YXXXΦN). The database can be searched for known or novel motifs and is accessible online at http://www.bioinformatics.cimr.cam.ac.uk/transtype/. B) Examples of type I membrane proteins with potential YXXXΦN motifs. The numbers of residues in the cytoplasmic tails both before and after each sequence are indicated. C) A C-terminal fragment of the cytoplasmic tail of CTLA-4 bearing a YXXXΦN motif was transplanted onto CD8, and the antibody uptake assay was carried out. D) Full-length CTLA-4, either wild-type or with the indicated mutations, was transiently transfected into HeLa cells and the antibody uptake assay was carried out using anti-CTLA-4 instead of anti-CD8.

Figure 8. Effect of AP-2 depletion on endocytosis

A) Antibody uptake assay carried out to compare the effects of clathrin depletion and AP-2 depletion on the various constructs. B) Uptake of the same constructs investigated using a different assay. Cells were stably transfected with the constructs and antibody followed by ¹²⁵I-labelled protein A was prebound to the cells at 4°C, then the cells were warmed to 37°C for 10 min to measure a single round of endocytosis.