The K/HDEL receptor does not recycle but instead acts as a Golgi-gatekeeper

Abstract

Accurately measuring the ability of the K/HDEL receptor (ERD2) to retain the ER cargo Amy-HDEL has questioned earlier results on which the popular receptor recycling model is based upon. Here we demonstrate that ERD2 Golgi-retention, rather than fast ER export supports its function. Ligand-induced ERD2 redistribution is only observed when the C-terminus is masked or mutated, compromising the signal that prevents Golgi-to-ER transport of the receptor. Forcing COPI mediated retrograde transport destroys receptor function, but introducing ER-to-Golgi export or cis-Golgi retention signals re-activate ERD2 when its endogenous Golgi-retention signal is masked or deleted. We propose that ERD2 remains fixed as a Golgi gatekeeper, capturing K/HDEL proteins when they arrive and releasing them again into a subdomain for retrograde transport back to the ER. An in vivo ligand:receptor ratio far greater than 100 to 1 strongly supports this model, and the underlying mechanism appears to be extremely conserved across kingdoms.

Fig. 1. Genetic validation of YFP-TM-ERD2 by stable transformation in Nicotiana benthamiana and Physcomitrium patens.

a Schematic of N. benthamiana ERD2ab antisense (AS) construct driven by the strong constitutive CaMV35S promoter (35S), combined with YFP constructs expressed under the weak TR2 promoter on the same T-DNA. b Confocal laser scanning microscopy (CLSM) of stably transformed N. benthamiana tissues, expressing all three fusions in regenerating shoots in tissue culture. Only YFP-TM-ERD2 led to fertile plants allowing us to image this fusion in root cortex cells from next-generation seedlings. Notice that ST-YFP-HDEL labels the ER, ERD2-YFP labels the ER and weak Golgi bodies, while YFP-TM-ERD2 only labels Golgi bodies. In roots, Golgi-stacks are either viewed from the side (arrow heads) or from top/bottom (stars), giving rise to the typical donut shapes. Even with high detector gain, YFP-TM-ERD2 cannot be detected in the ER. Size marker 10 μm. c Schematic of YFP-TM targeted gene knock-in onto PpERD2B-1 (Pp3c9_13230V3.9), leading to expression of a YFP-TM-ERD2 derivative under the transcriptional control of the native promoter in P. patens. d Schematic of PpERDB2-2 (Pp3c15_12830) knockout by complete deletion of the second ERD2 gene. e YFP-TM-PpERD2 expression under its native promoter in P. patens. (e1) Notice stronger expression near growing tips and newly formed cell plates (white stars). Size marker 50 μm. (e2) At high magnification, distinguish punctate structures (white arrow heads) from weak autofluorescence of chloroplasts (Chl.). Size marker 10 μm.

Fig. 2. ERD2 function and Golgi residency is conserved amongst eukaryotes.

a Retention assay using protoplasts showing the secretion index (ratio extra/intracellular Amy-HDEL activity) with cargo alone (either Amy-HDEL or Amy-KDEL) or with co-expressed A. thaliana ERD2b (At) and 12 further ERD2 orthologs from the eukaryotes Ostreococcus lucimarinus (Oi), Acanthamoeba castellanii (Ac), Phytophthora infestans (Pi), Chondrus crispus (Cc), Galdieria sulphuraria (Gs), Homo sapiens (HsERD2), Hypsibius dujardini (Hd), Thalassiosira pseudonana (Tp), Puccinia graminis (Pg), Kluyveromyces lactis (KlERD2), Trypanosoma brucei (Tb) and Saccharomyces cerevisiae (Sc). Transfection efficiencies were normalised by the internal marker GUS established at 5 standard OD units as described in materials and methods. Percentages in brackets refer to the sequence identity with AtERD2b. Error bars are standard errors. Source data are provided as a Source Data file. b CLSM analysis of two separate Arabidopsis thaliana ERD2b fluorescent fusions (YFP-TM-AtERD2 and AtERD2-YFP, shown in green), each co-expressed with the ER marker RFP-KDEL (shown in red) in HeLa cells. The endogenous cis-Golgi marker GM130 was detected via immunocytochemistry (shown in light blue) and the nucleoplasm is stained with DAPI (dark blue). Notice that in contrast to the C-terminal AtERD2-YFP fusion, YFP-TM-AtERD2 is not detected in the ER and shows the best co-localisation with GM130. The size marker bar is 10 microns. c CLSM analysis at higher magnification to compare YFP-TM-AtERD2 with two different Golgi markers. Notice that the trans-Golgi marker TGN46 is clearly distinct from the ERD2-fusion and GM130. Size marker 10 microns. d As in (b), but fluorescent fusions contain human ERD2 (HsERD2). Size marker 10 microns.

Fig. 3. Analysis of C-terminal residues and sensitivity to C-terminal epitope tagging in Hs and At ERD2.

a CLSM analysis of two human ERD2 fusions (YFP-TM-HsERD2 and HsERD2-YFP constructed as described (Silva-Alvim et al., 2018) imaged in tobacco leaf epidermis cells. Notice that only the C-terminal YFP fusion causes partial ER localisation. Size marker 10 microns. b Amy-HDEL retention assays as in Fig. 2a but either comparing the two HsERD2 fluorescent variants from panel (a), or a comparison of untagged HsERD2 (WT) with the point–mutations in the HsERD2 C-terminus indicated above each lane. Notice that the C-terminal YFP fusion has completely lost biological activity. Notice also that only the LLGG mutant has lost biological activity when untagged HsERD2 is analysed. Error bars are standard deviations from 3 biological replicas. A full dose response for KKAA is provided in Supplementary Fig. 3. c C-terminal amino acid sequences of human (KDELR2) and A. thaliana ERD2B. Conserved residues are highlighted grey and the conserved di-leucine motif is highlighted bold. d CLSM analysis of selected mutants from (b) but in the YFP-TM-HsERD2 configuration. Silent mutations in panel (b) retain the Golgi localisation, while the inactive LLGG mutant displays partial ER localisation. e Schematic of C-terminal fusions to At and Hs ERD2 for functional assays (upper) and the fluorescent derivative for CLSM analysis (lower schematic). f Secretion index of Amy-HDEL, co-expressed with either wild type human or plant ERD2 compared to the three different C-terminal modifications (FLAG, c-myc, HA) in each case. Constant levels of ERD2 encoding plasmids were co-transfected (yielding 5 standard OD units). In both instances, the addition of a FLAG or c-myc tag strongly reduced function, whilst most of the activity was maintained for each ortholog after adding the HA tag. Error bars are standard deviations from 3 biological replicas. Source data are provided as a Source data file. g Localisation of human and plant ERD2 fluorescent fusions with C-terminal FLAG, c-myc and HA tags. Notice that FLAG and c-myc additions cause an ER-Golgi localisation, whilst the addition of an HA tag does not affect the Golgi localisation of YFP-TM-ERD2 for both orthologs.

Fig. 4. FRAP and redistribution assays identify a Golgi-retention signal at the ERD2 C-terminus.

a Fluorescence recovery after photobleaching (FRAP) comparing wild type ERD2 and the LLGG mutant. ERD2 recovery to 50% (400 s) took almost twice the time of the LLGG mutant (240 s). LLGG mutant recovery reached 85%, whereas wild type ERD2 remained at around 50%. Error bars are standard deviations from at least 6 biological replicas. Source data are provided as a Source data file. b Schematic of dual expression T-DNA constructs used to co-express fluorescent ERD2 fusions with either secreted Amy or the ERD2-ligand Amy-HDEL. c Golgi localisation of YFP-TM-ERD2 co-expressed with Amy and Amy-HDEL. Distribution remains unchanged for both cargo. Size bars 10 microns. d Dual ER-Golgi localisation of YFP-TM-ERD2ΔC5 co-expressed with Amy and a more prominent ER localisation when co-expressed with Amy-HDEL. Size bars 10 microns. e Dual ER-Golgi localisation of ERD2-YFP co-expressed with Amy. The re-distribution of ERD2-YFP to the ER by co-expressed Amy-HDEL is even more drastic compared to that of the deletion mutant in panel (d). Size bars 10 microns. f Schematic including the T-DNA construct used to study ERD2-YFP localisation, which is variable depending on expression levels. Golgi bodies (arrows) are labelled by ERD2-YFP more visibly during high cellular expression, whereas punctae are much fainter relative to the ER fluorescence at low expression levels (imaged at higher detector gain). Size bars 10 microns.

Fig. 5. A canonical COPI transport motif (KKXX) causes ER Localisation and abolishes Amy-HDEL retention of ERD2.

a C-terminal amino acid sequences of ERD2 wild type (WT) and two variants in which the last 9 amino acids of ERD2 is replaced by the corresponding region of p24 (underlined). The proposed COPII ER export signal of p24 (Contreras et al., 2004b) is in bold, as is the dileucine motif in the WT sequence, the relevant lysines of the canonical COPI transport motif in the p24 variant and finally the mutant serines in the KKSS variant. Size bars 10 microns. b Dose–response assay measuring the influence of co-transfected C-terminal ERD2 variants (given in standard GUS OD units below each lane) on Amy-HDEL secretion. ERD2-WT mediates strong cell retention whilst the p24 fusion shows no retention activity. The KKSS mutant of the p24 fusion restores the retention activity at the highest dose. Error bars are standard deviations from at least 4 biological replicas. c The effect of Brefeldin A on the transport of RFP-TM-ERD2 compared to ST-YFP. Notice that both fusions have re-distributed to the ER after 3 h of Brefeldin A treatment. Size bars 10 microns. d Sequence of the ERD2 C-terminus, the p24 fusion and the KKSS mutant thereof. e Amy-HDEL retention activity of constructs presented in (d). Fusing the p24 C-terminus renders ERD2 completely inactive, yet mutating the KKXX motif restores the bulk of biological activity. However, KKSS cannot meet the activity of the wild type ERD2 at lower doses. Error bars are standard deviations from at least 4 biological replicas. Source data are provided as a Source data file. f Localisation of p24 and KKSS hybrids incorporated into fluorescent ERD2 fusion proteins. The p24 C-terminus mediates complete ER localisation of the resulting ERD2 fusion whilst the KKSS mutant shows high steady-state levels at the Golgi. Size bars 10 microns.

Fig. 6. Reactivation of ERD2-YFP with a novel cis-Golgi retention motif.

a Comparative Golgi distribution between ST-RFP and YFP-TM-ERD2. Although both labelling punctae, an rS value of 0.54 highlights a cis-trans Golgi segregation of YFP-TM-ERD2 and ST-RFP, respectively. This is most visible in regions with mostly red fluorescence (white arrow heads), also apparent from two populations in the scatterplot. Size bars 10 microns. b When compared with MNS3-RFP instead, the shared cis-Golgi localisation of YFP-TM-ERD2 is clearly demonstrated by the much higher rS value of 0.92 and a single population in the scatterplot. Size bars 10 microns. c Retention of Amy-HDEL by ERD2 fusion variants at increasing concentrations. As previously published, TM-ERD2 effectively retains Amy-HDEL at low and high concentrations. Meanwhile, the C-terminal addition of YFP completely abolishes retention, regardless of increasing concentration. However, the N-terminal addition of the LPYS Golgi retention motif does allow significant activity to return with increasing concentration. Error bars are standard deviations from at least 2 biological replicas. Source data are provided as a Source data file. d N-terminal addition of LPYS causes redistribution of TM-ERD2-YFP exclusively to the Golgi, in agreement with reactivation in secretion assays. Size bars 10 microns.

Fig. 7. Receptor-recycling cannot explain the observed receptor-ligand stoichiometry.

Transient expression in which both receptor and ligand were expressed from dual expression vectors harbouring GUS as reference marker for transfection efficiency and expressed at identical GUS units. a Immunoprecipitated ERD2-HA (α-HA) and co-expressed Amy-HDEL (α-Amylase) from mock-transfected cells (-), cells transfected with Amy-HDEL alone (A) and cells with Amy-HDEL and ERD2-HA (AE). Immunoprecipitated proteins were separated by SDS-page, followed by blotting on nitrocellulose and visualisation via phosphorimaging. Molecular weight markers are given on the right in kilo-daltons. Source data are provided as a Source data file. b Table showing the total number of cysteine and methionine residues in cargo and receptor. Relative radioactivity units measured for ERD2-HA by phosphorimaging must be multiplied by 1.5 to permit comparison with units from Amy-HDEL to permit calculation of relative number of molecules. Source data are provided as a Source data file. c Phosphorimaging quantification (arbitrary relative units) of signals from 3 different nitrocellulose blots as in (a) showing the radioactivity from transiently expressed ERD2-HA, Amy-HDEL (A) alone in cells and medium and Amy-HDEL co-expressed with ERD2-HA (AE) in cells and medium. Error bars are standard errors from 3 biological replicas. d Retention of Amy-HDEL where the maximum receptor levels from panel a are co-transfected (second lane), followed by consecutive dilution of the receptor plasmid up to 100-fold (last lane). Notice that a strong reduction in Amy-HDEL secretion compared to the control (first lane) is still observed even after 100-fold dilution of the receptor plasmid (last lane). Error bars are standard deviations from at least 6 biological replicas. Source data are provided as a Source data file. e Schematic of the ER-Golgi interface, summarizing the current findings: K/HDEL cargo is several orders of magnitude more abundant than ERD2 and reaches the Golgi together with other cargo. A concentrated array of ERD2 molecules recognise K/HDEL cargo in the cis-Golgi/ERGIC lumen and releases them close to the COPI vesicle budding site. ERD2 itself is retained by an unknown ERD2-Golgi-retention (EGR) complex that interacts with the cytosolic di-leucine motif. The findings support a specific identity for the cis-Golgi/ERGIC and are difficult to reconcile with the cisternal progression model for intra-Golgi anterograde transport.