Triple Labeling Resolves a GPCR Intermediate State by Using Three-Color Single Molecule FRET

Abstract

The correlation of individual conformational changes in dynamic protein complexes remains challenging as most structural methods rely on averaged information over a large number of molecules. Single molecule FRET is a powerful tool for monitoring such conformational changes. When performed using three distinct probes, it enables the correlation of domain movements by providing up to three simultaneous distance measurements with high temporal resolution. Nevertheless, a major challenge lies in the site-specific attachment of three probes to unique positions within the target protein. Here, we propose an orthogonal triple-labeling strategy that is not compromised by native, reactive amino acid functionalities. It combines genetic code expansion and bioorthogonal labeling of two different noncanonical amino acids with an enzymatic self-labeling SNAP tag. We demonstrate its application by establishment of a 3-color sensor on the human metabotropic glutamate receptor 2, a dimeric, multidomain G protein-coupled neuroreceptor, and describe a previously unknown conformational intermediate state using 3-color single molecule FRET.

1

Major conformational changes during mGlu2 activation. Cryo-EM structures of the mGlu2 receptor (left) in the resting open (R _O) state with Venus flytrap (VFT) domains open, the lower lobes and cysteine-rich domains (CRDs) separated, and an interface mediated by the transmembrane domain helices IV (PDB 7EPA) and (right) in the active closed (A _C) state with the VFT domains closed, the lower lobes and CRDs in closer proximity and a slightly twisted dimer interface mediated by 7TM helices VI and VII (PDB 7EPB). The upper and lower lobes of the VFT domains are shaded in gray. The labeling positions used to establish a 3-color smFRET sensor are indicated by colored circles (Donor 1 (D1): purple; Acceptor 1/Donor 2 (A1/D2): green; Acceptor 2 (A2): red).

2

Orthogonal triple labeling strategy. (a) Top: Schematic representation of the mGlu2 protomer genes expressed to produce the 3-color smFRET sensor. Important features of the protomers are highlighted, including ATG start codons, SNAP tag, premature TAG and TAA stop codons within the mGlu2 gene, the GABA_B C1 and C2 tails and the final TGA stop codons. Bottom: Schematic representation of cotranslational double ncAA incorporation in HEK cells. (b) Three-step site-specific labeling reactions of mGlu2 receptors in the membrane of living cells prior to solubilization: (1) SNAP-tag labeling using Atto488–0⁶-benzylguanine; (2) TCOK labeling using Cy3B-tetrazine; (3) PrF labeling using AF647-picolyl-azide. (c) Orthogonal suppression of TAG and TAA stop codons. Fluorescence signal of HEK cells expressing the mGlu2 receptor bearing either a TAG or a TAA premature stop codon and an N-terminal SNAP fusion protein, labeled at the cell surface with Lumi4-Tb-BG. Only full-length receptors expressed through successful ncAA incorporation are presented at the cell-surface and labeled. All expressions were performed with both synthetase/tRNA pairs in the absence or presence of either ncAA. Data are given as the mean ± standard deviation from biological triplicates. (d) Evaluation of the labeling specificity. Receptors with incorporated TCOK, PrF or SNAP fusion protein were subjected to the respective labeling reactions as indicated with (+) using Cy3B-tetrazine, AF546-picoly zide and/or Cy3B-benzylguanine and detected by in-gel fluorescence after solubilization and SDS-PAGE.

3

2-Color smFRET experiments suggest efficient VFT domain closure but incomplete stabilization of the reoriented state. Histograms of VFT domain closure (a) and reorientation (b) sensors in the apo condition and in the presence of a saturating glutamate concentration (10 mM). The closure sensor (a) was labeled with Atto488-O⁶-benzylguanine via the SNAP-tag on the upper lobe and with AF647-picolyl-azide via PrF incorporated in response to a TAG stop codon at position 248 within the lower lobe. Sensors bearing only a single labeled protomer were achieved using the engineered GABA_B receptor quality control system. The reorientation sensor (b) was obtained by labeling of PrF incorporated in response to a TAG stop codon at position 248 within the lower lobe of protomer 1 with AF647-picolyl-azide and TCOK incorporated in response to a TAA stop codon at position 258 within the lower lobe of protomer 2 with Cy3B-tetrazine. The engineered GABAB receptor quality control system was used to specifically label the desired dimer populations on the cell surface. Data are given as the mean ± standard deviation from biological triplicates.

4

Three-color smFRET experiments unveil coordinated rearrangements of the VFT domains of mGlu2 and an unknown conformational intermediate. (a, b) Two-dimensional projections of VFT domain closure (FRET efficiency BR) versus reorientation (FRET efficiency GR) in the absence (a) and presence of a saturating glutamate concentration (b). (c) One-dimensional projections of the proximity ratio for VFT domain closure (PRBR, left) and reorientation (PRGR, middle) and two-dimensional apparent distance distribution histogram (right) extracted from PDA analysis for the apo state. (d) One-dimensional projections of the proximity ratio for VFT domain closure (PRBR, left) and reorientation (PRGR, middle) and two-dimensional apparent distance distribution histogram (right) extracted from PDA analysis under the influence of a saturating glutamate concentration (10 mM). Two populations are sufficient to describe the 3-color data, which is clearly observable in the PRGR projection (d, middle) but more difficult to detect for the PRBR data (d, left). However, the PDA analysis reveals a clear reduction in the distance between upper and lower lobes (PRBR) between the apo (c, right) and glutamate condition (d, right) indicating a slight closure of the VFTs. Shown is a representative data set. A second data set can be found in Figure S4.

5

Constrained analysis of 2-color smFRET data on the VFT domain closure recovers the intermediate state. (a, b) SmFRET histograms from 2-color experiments in the presence of 10 mM glutamate are shown for the SNAP-PrF VFT closure sensor (a, corresponds to data in Figure a) and the closure sensor obtained through incorporation and labeling of TCOK at position 358 with Atto488-tetrazine and PrF at position 248 with AF647-pycolyl-azide within the same protomer using the GABAB quality control system (b). The high FRET values, corresponding to the fully closed state, were fixed to 0.47 (a) and 0.64 (b), respectively, as obtained from control measurements performed in the presence of glutamate and allosteric modulator BINA (see Figure S3). The ratio between high and low FRET states were constrained to 55%/45% as obtained from 3-color measurements (see Figure b). The low FRET values extracted from this procedure were found to be intermediate between the fully closed, high FRET values and the low FRET values obtained for the apo state (see Figure S3). Data are given as the mean ± standard deviation from biological triplicates.