pH-susceptibility of HLA-DO tunes DO/DM ratios to regulate HLA-DM catalytic activity

The peptide-exchange catalyst, HLA-DM, and its inhibitor, HLA-DO control endosomal generation of peptide/class II major histocompatibility protein (MHC-II) complexes; these complexes traffic to the cell surface for inspection by CD4+ T cells. Some evidence suggests that pH influences DO regulation of DM function, but pH also affects the stability of polymorphic MHC-II proteins, spontaneous peptide loading, DM/MHC-II interactions and DM catalytic activity, imposing challenges on approaches to determine pH effects on DM-DO function and their mechanistic basis. Using optimized biochemical methods, we dissected pH-dependence of spontaneous and DM-DO-mediated class II peptide exchange and identified an MHC-II allele-independent relationship between pH, DO/DM ratio and efficient peptide exchange. We demonstrate that active, free DM is generated from DM-DO complexes at late endosomal/lysosomal pH due to irreversible, acid-promoted DO destruction rather than DO/DM molecular dissociation. Any soluble DM that remains in complex with DO stays inert. pH-exposure of DM-DO in cell lysates corroborates such a pH-regulated mechanism, suggesting acid-activated generation of functional DM in DO-expressing cells.

Similar constructs were used in plasmids directing the expression of soluble DQ1 in complex with CLIP in S2. The regions in DQA1*01:01 and DQB1*05:01 encoding the ectodomains of α and β chains of DQ1, respectively were amplified by PCR using cDNAs made from a B lymphoblastoid cell line 3.1.3. The two amplicons were subcloned into plasmids pRmHa3-sDQ2A and pRmHa3-sDQ2B, respectively, directing expression of soluble DQ2 (ref 3 ), in place of the DQ2A and DQ2B genes. The resulting plasmids are pRmHa3-sDQ1A and pRmHa3-sDQ1B, respectively. The latter allowed the expression of a flag-tag (DYKDDDDK) at the C-terminal end of DQ1 β chain for purification and detection purposes.
S2 cells were co-transfected following the standard protocol (Life Technologies) with three plasmids directing the expression of recombinant DQ α chain , β chain, and the neomycin (geneticin)-resistance gene, respectively, at a ratio of 20:20:1. The transfected S2 cells were selectively cultured in complete Schneider Drosophila medium (with 10% FBS and 2mM Glutamine, Life Technologies) containing 1.5 mg/ml Geneticin (G418, Life Technologies) for 4-5 weeks to construct a stable cell line secreting recombinant DQ6 or DQ1 proteins.
To make Alexa fluor dye conjugated mAbs, MagsDO5 and MapDM1 were first purified from the culture supernatant of mouse hybridoma (a gift from Dr. Lisa Denzin, Rutgers, the State University of New Jersey) and ascites, respectively, using protein G beads (GE Healthcare), and then labeled with corresponding Alexa Fluor dye carboxylic acid, succinimidyl ester (Life Technologies), followed by size exclusion chromatography using a Superdex 200 gel filtration column to remove extra free dye.
MapDM1 mAbs were coupled with CNBr-activated sepharose 4B (GE Healthcare) before packing into a column for DM purification. DM was purified from the clarified T2DM lysate (from 500 million cells) by affinity chromatography using the customized MapDM1 column and eluted from the column using an acidic buffer (100 mM glycine-HCl, 150 mM NaCl, pH 3) + 0.5% (w/v) CHAPS. The eluent was immediately neutralized with 1M Tris-Cl (pH 8.5) + 0.5% (w/v) CHAPS and analyzed by western blot and size exclusion chromatography.

Purification of recombinant proteins
The supernatant of S2 cell culture containing soluble proteins were filtered through a 0.22 um membrane to remove cell debris, and proteins were purified from the supernatant by affinity chromatography, followed by size exclusion chromatography, using a Superdex 200 gel filtration column (GE Healthcare). Fractions containing monomeric forms of each protein were pooled and analyzed by Coomassie staining or western blotting. For affinity chromatography, customized L243 (DR-specific mAb) columns 6 were used to purify DR1, DR3, and DR4; a customized SPV-L3 (DQ-specific mAb) 7 column was used to purify DQ6; and M2 (a flag-tag-specific mAb) columns (Sigma) were used to purify DQ1, DM, DO v , DO mutants and wild type DM-DO complexes, which have flag-tag epitopes covalently attached at the C-terminus of one chain. A low buffer-strength running buffer (e.g., 20-25 mM Tris-Cl, 150 mM NaCl, pH 7.4) typically was used for size exclusion chromatography to minimize the dilution/neutralization effect on the downstream pH-sensitive assays. The SPV-L3 mAb was purified from the culture supernatant of mouse hybridoma using gammabind G sepharoses (GE Healthcare) and then coupled with CNBr-activated sepharose 4B (GE Healthcare) before packing into a column. Some soluble DR4 with covalently tethered CLIP was kindly provided by Drs. Laura Su and Mark Davis (Stanford). Ectodomains of wild type DM-DO complexes were gel filtration-purified twice to eliminate any free DM.

Protein biotinylation and thrombin cleavage
An avitag (GLNDIFEAQKIEWHE) was added to the C-terminus of DM, DO v , and wild type DM-DO ectodomains, respectively, allowing site-specific biotylation using BirA ligase (Avidity). The biotinylation reaction was performed in the presence of BirA enzyme at RT for 1 h and then kept at 4°C overnight before the removal of excess free biotins by repeated concentration and buffer-exchange using an Amicon Ultra-15 (10 KDa MW cut off, Millipore). The resultant proteins are designated as bio-DM, bio-DO v , and bio-DM-DO, respectively. Recombinant MHC-II allelic proteins covalently tethered with CLIP 87-101 were treated with thrombin (at 1 U/100 µg protein, Novagen) for 2 h at RT to cut the linker between the CLIP peptide and the MHC-II α/β dimer.

Immunoassay for quantification of DM effect at varied pH
A series of high strength (100 mM) buffers were made for the peptide exchange reaction, and pH was monitored before and after the introduction of reaction components to ensure precise pH. These reaction buffers contain 150 mM NaCl, 1% (w/v) BSA, 0.5% (v/v) IGEPAL CA-630 (Sigma), 0.1% (w/v) NaN3, and 100 mM of one of the following buffers: acetate buffer (acetic acid and sodium acetate), pH 4.6; citrate buffer (citric acid and sodium citrate), pH from 4.56 to 5.41; phosphate buffer (monosodium phosphate and disodium phosphate), pH 6.3 or 7.2; (see Supplementary Table 4 for a comparison of pH measured at different temperatures). 1X EDTA-free complete protease inhibitor (Santa Cruz Biotechnology, Inc.) was added to each reaction buffer before use. Separate neutralization buffers, all consisting of 150 mM NaCl, 1% (w/v) BSA, 0.5% (v/v) IGEPAL CA-630, 0.1% (w/v) NaN 3 , and 100 mM of an appropriate pH buffer, were made for reactions at different pHs (e.g., neutralization buffer in 100 mM Tris-Cl, pH 8.4 can be used for reactions at pH 4.6 to pH 4.9).
In the time course experiment, bulk reaction mixtures containing recombinant MHC-II allelic proteins or native MHC-II in clarified cell lystaes and various concentrations of peptides with or without DM were prepared in reaction buffers at different pHs and incubated at 37°C. Affinity purified soluble and native DM were used to catalyze peptide loading onto recombinant and native MHC-II, respectively. At different time points, a fraction of each reaction was collected and stopped by adding two volumes of corresponding ice-cold neutralization buffer. MHC-II α/β dimers loaded with biotinylated peptides were quantified by capture ELISA with streptavidin-europium detection, as previously described 3,8,9 . For the capture ELISA, L243 and SPV-L3 mAbs from ascites were coupled to the high adsorption 96-well plate (Thermo Scientific) for capture of DR and DQ proteins, respectively.

In vitro peptide loading with pre-treated DM and DO
Recombinant accessary molecules, including DM, DO v , and wild type DM-DO complexes, were pre-diluted into the reaction buffer (described above) or bufferexchanged to a low buffer-strength pH buffer (see below) with 1X EDTA-free complete protease inhibitor and incubated at 4°C or 37°C for an indicated time before being mixed with the MHC-II protein and the biotinylated peptide. These low strength pH buffers include 150 mM NaCl and 10 mM of one of the following buffers: acetate buffer, pH 4.8; citrate buffer, pH from 4.6 to 5.4; MES buffer, pH 6.3; PBS, pH 7.4. The peptide exchange reaction in the presence of different combinations of accessary molecules (pretreated versus non-treated) took place at 37°C for 1-3 h as indicated. After incubation, reactions were neutralized and MHC-II loaded with biotinylated peptides were quantified as described above by capture ELISA.
In case of pre-incubation of regulatory molecules with DR4, 2x concentrated mixtures containing 10 nM of recombinant DR4 and various concentrations of soluble DM or DM-DO wt were prepared in reaction buffers (with protease inhibitor) at varied pH and preincubated at 37°C. When full-length proteins were utilized, clarified lysates of 0.2 million T2 cells or T2 MHC-II-transfectants were diluted in reaction buffers (with protease inhibitor) at different pHs and pre-incubated at 37°C. After pre-incubation, bio-HA 306-318 prepared in reaction buffers (with protease inhibitor) with the same pH was added to a final concentration of 1.5 µM and peptide exchange reactions were then took place at 37°C for 1.5 h before neutralization and quantification of HA in association with DR4, as described above. Samples were all prepared and mixed on ice in order to maintain the stability of proteins and to minimize the variation in incubation time at higher temperature. Neutralized reactions were frozen at -20°C before analysis by the capture ELISA if necessary.

Capture of DM-DO by ELISA
200 nM of bio-DM and various concentrations of DO v were mixed in reaction buffers at different pHs (detailed above) and incubated at 37°C for 1.5 h. After neutralization, the reaction mixture was transferred to a 96-well plate pre-coated with Mags.DO5 mAb for capture of DO v or DM-DO v at RT for 1 h. The plate was then washed and the captured complexes were detected by europium-labeled streptavidin (PerkinElmer), using timeresolved fluorescence (Wallac).

Fluorescent dye labeling and FRET analysis
DM, DO v βS63C, and DO v αR80C were buffer-exchanged to TBS buffer (50 mM Tris-Cl buffer, 150 mM NaCl, 5 mM EDTA, pH 7.4) using Amicon Ultra-15 (10 KDa MW cut off) and mixed with freshly dissolved DTT to a final concentration of 0.3-0.4 mg protein/ml to reduce the single free surface cysteine for conjugation of CyDye maleimides (GE Healthcare). DM was incubated with 45 mM DTT at RT and then left overnight at 4°C, while DO mutants were incubated with 20 mM DTT plus 1X complete protease inhibitor at RT for 45 min to maintain the stability of DO molecules. Reduced DM and DO mutants were repeatedly diluted in TBS and concentrated using Amicon Ultra-15 (10 KDa MW cut off) to remove excess DTT and then incubated with a 20-40 fold molar excess of Cy5 maleimide and Cy3 maleimide (GE Healthcare), respectively, at RT for 2-3 h and left overnight at 4°C, protected from light. After incubation, the reaction was quenched with 1 mM DTT. Cy5 labeled DM (DM-Cy5) or Cy3 labeled DO mutants (DO v βS63C-Cy3 and DO v αR80C-Cy3) were separated from free CyDye maleimides by size exclusion chromatography, using a Superdex 200 gel filtration column. The activity of CyDye labeled proteins was verified using peptide-loading assays ( Supplementary  Fig. 5). CyDye/protein ratios for DM and DO mutants were 0.5 and 0.8, respectively. An irrelevant protein (MHC-I) that contains a free cysteine was also labeled with Cy5 similarly; and the CyDye/protein ratio was 0.5. MHC-I was kindly provided by Dr. Larry Stern (UMass Medical School). 100 µl of 50 mM phosphate buffer, 150 mM NaCl, pH 7.0 or 50 mM acetate buffer, 150 mM NaCl, pH 4.7, containing various concentrations of Cy3, Cy5, DM-Cy5, or MHC-I-Cy5, or DO v βS63C-Cy3, or DO v αR80C-Cy3 alone, or any combinations of donor (Cy3) / acceptor (Cy5) pairs were prepared in a black 96-well flat-bottom Costar assay plate (Corning Inc.). Samples were excited by laser at a wavelength of 500 nm and the emission spectrum in between 560 nm and 700 nm was acquired at 25°C using a Gemini XS fluorescent microplate reader (Molecular Devices), equipped with a 550 nm long-pass filter. Emission values measured for acceptor alone samples were subtracted from those measured for corresponding donor/acceptor samples to calculate the FRET-associated fluorescence signal.

Western blotting analysis
Non-treated proteins or proteins reduced by β-mercaptoethanol followed by boiling at 100°C for 10 min were separated by gel electrophoresis using 12% precast polyacrylamide gels (Bio-Rad). Proteins were then transferred to Immobilon-P membrane (Millipore) for antibody detection. TAL 18.1 (0.2 mg/ml, DMα-specific mAb, Santa Cruz Biotechnology, Inc; used at 1:600 dilution) and 27E8 (His-tag-specific mAb; used at 1:1000 dilution) followed by HRP-goat anti-mouse IgG1 Ab (1mg/ml, Invitrogen; used at 1:10,000 dilution) were used to detect DMα and His-tag at the C-terminus of recombinant DQα, respectively. Rabbit anti-KT3-tag Ab (1mg/ml, Genscript; used at 1:4000 dilution) and rabbit anti-Avitag Ab (0.5 mg/ml, GenScript, used at 1:2000 dilution) followed by HRP-donkey anti-rabbit IgG Ab (GE healthcare; used at 1:20,000 dilution) were used to detect KT3-tag at the C-terminus of recombinant DMβ and Avitag at the C-terminus of recombinant DOβ, respectively. KT3 and Avitag are both unique in DM-DO wt . HRP-anti-flag-tag mAb (1 mg/ml, Sigma; used at 1:10,000 dilution) was used to detect flag-tag at the C-terminus of recombinant DQβ. Sufficient stripping of prebound antibodies was performed using western blot stripping buffer (Restore;Thermo Scientific) when blotting of the same membrane was needed. Eluents at 12-13 ml, expected to contain intact DO v based on their elution at the volumes where non-treated DO v (gray area) elute, were pooled and used for SPR analysis in Figure 2f and peptide loading studies in Figure 3c. To prepare acid-treated DO v, it was incubated at pH 4.6, 37°C for 8 min and separated by size exclusion chromatography using a Superdex 200 gel filtration column. The normalized UV absorbance of pulsed DO v (red line) is plotted against eluent volume with an overlay of the UV trace for nontreated DO v .