Coordinated DNA polymerization by Polγ and the region of LonP1 regulated proteolysis

Abstract

The replicative mitochondrial DNA polymerase, Polγ, and its protein regulation are essential for the integrity of the mitochondrial genome. The intricacies of Polγ regulation and its interactions with regulatory proteins, which are essential for fine-tuning polymerase function, remain poorly understood. Misregulation of the Polγ heterotrimer, consisting of (i) PolG, the polymerase catalytic subunit and (ii) PolG2, the accessory subunit, ultimately results in mitochondrial diseases. Here, we used single particle cryo-electron microscopy to resolve the structure of PolG in its apoprotein state and we captured Polγ at three intermediates within the catalytic cycle: DNA bound, engaged, and an active polymerization state. Chemical crosslinking mass spectrometry, and site-directed mutagenesis uncovered the region of LonP1 engagement of PolG, which promoted proteolysis and regulation of PolG protein levels. PolG2 clinical variants, which disrupted a stable Polγ complex, led to enhanced LonP1-mediated PolG degradation. Overall, this insight into Polγ aids in an understanding of mitochondrial DNA replication and characterizes how machinery of the replication fork may be targeted for proteolytic degradation when improperly functioning.

Graphical Abstract

Figure 1.

Biochemical Analysis of LonP1-dependent PolG degradation. (A) Schematic of PolG and LonP1 domain structures. PolG colored according to domain boundaries: NTD 1–170 (residues 1–25, MTS), green; Exo domain (171–441), navy blue; polymerase domain (442–476 and 787–1239), teal; linker (477–786), purple. LonP1 is colored according to domain boundaries: MTS (1–67 or 1–114), gray; N-domain (68–485 or 115–485), salmon; ATPase domain (486–729), vermillion; Proteolytic domain (730–959), maroon. Exo, exonuclease; NTD, N-terminal domain; MTS, mitochondrial targeting sequence; IP, intrinsic processivity; and AID, an accessory interacting determinant. (B) Left: Quantification of PolG (1.25 μM) degradation, over 60 min in the presence of either Δ67 LonP1^WT (squares) or Δ114 LonP1^WT (circles) (0.42 μM, hexamer). Right: representative Coomassie stained SDS-PAGE gel of LonP1-dependent PolG digestion. MW, molecular weight. BSA, bovine serum albumin. (C) Quantification of PolG (1.25 μM) degradation at 60 minutes in the presence of either Δ114 LonP1^WT, Δ114 LonP1^Y565A, or Δ114 LonP1^S855A (0.42 μM, hexamer). All errors bars in figures B and C are displayed as standard deviations of three independent experiments.

Figure 2.

XL-MS Analysis of PolG:LonP1 Interaction and PolG monomer instability by Cryo-EM. (A) 4–12% Coomassie stained SDS-PAGE gel of Δ114 LonP1^S855A or Δ114 LonP1^WT (0.625 μM, hexamer) and PolG (0.625 μM) in the presence and absence of Bs3d0 crosslinker (0.625 mM). Molecular weight, MW; kilodaltons (KDa). (B) 2D representation of crosslinking-mass spectrometry of Δ114 LonP1^WT and PolG displayed using XiNET. A gray circle highlights K504 (PolG) crosslink with K405 and K411 (LonP1) (C). A Side (left) and top view (right) of the LonP1 cartoon, cylindrical helices representation of PDB ID: 7NFY (50). The bound, unknown substrate (mauve) is shown in ball representation. Residues K411 (green) and K405 (purple) are highlighted in ball representation. (D) Sample cryo-EM 2D class averages of PolG monomer. (E) Cartoon representation of monomeric PolG (PDB ID: 8V5D, EMD-42982, this study). (F) Top: Global, all atom alignment of PDB ID: 4ZTZ (18), chain A (light orange) and 8V5D, colored according to Figure 2B and shown in cartoon representation. Dashed boxes highlight two regions of missing density in PDB ID: 8V5D when the PolG2 homodimer is absent. Bottom: Zoom-in on the region of PolG missing density. Right: The region located near the PolG: LonP1 crosslink.

Figure 3.

Probing the structures of PolG/PolG2 which occlude LonP1-mediated degradation. (A, B) PolG colored according to domain boundaries in Figure 1A schematic. PolG2 monomer 1 (light orange) and PolG2 monomer 2 (pale yellow). (A) Cartoon representation of (1) Polγ^open-DNA bound complex (PDB ID: 8V55, EMD-42980) (2) Polγ^engaged-DNA bound complex (PDB ID: 8V54, EMD-42979) (3) Polγ^active-DNA bound complex (PDB ID: 8V5R, EMD-42984). A dotted line marks a central x-axis. The arrows indication the conformational movement of the proximal protein domains when aligning to chain B (PolG2). The dashed circle highlights a region of conformational change. (B) A surface representation of Polγ^open, as colored in Figure 1A, and Polγ^engaged (gray) aligned in the region of PolG2 (chain B in PDB ID: 8V55 and 8V54). Dashed circles highlight a region of conformational change near the loop containing Y233 PolG. (C) Alignment of chains A (PolG) in Polγ^open (colored according to Figure 1A) and Polγ^engaged (grey cartoon representation), dashed box highlights the L-helix of PolG which interacts with PolG2. Distances highlight the movement of DNA between the open and engaged conformations. Right: A zoom-in of a region of interaction between PolG and PolG2. Alignment performed using residues 530–555 of PolG. The region of PolG2^G451 is highlighted along with the distances corresponding to local movement between the open and engaged conformations.

Figure 4.

Catalytic intermediates of the POLγ catalytic cycle. Monomer (PDB ID: 8V5D, this study, pale yellow), holoenzyme (PDB ID: 3IKM (17), deep olive), open (PDB ID: 8V55, this study, light magenta), engaged (PDB ID: 8V54, this study, cyan), active (PDB ID: 8V5R, this study, lime green), X-ray crystallography ΔI4 inhibited (PDB ID:4ZTZ (18), wheat), replication complex with Ca²⁺ (PDB ID: 8D33 (53), bright orange), mismatch intermediate with no metal modeled (PDB ID: 8G5I (65), salmon), mismatch intermediate Ca²⁺ (PDB ID: 8D3R (53), slate), mismatch intermediate 2 with Ca²⁺ (PDB ID: 8D37 (53), limon), editing complex with exonuclease site occupied (PDB ID: 8D42 (53), chocolate). Additional editing states with no metal bound, states referred to Buchel et al. as mismatch uncoupling (PDB ID: 8G5J (65), deepteal), mismatch locking (PDB ID: 8G5M (65), raspberry), G-loop engagement intermediates (PDB ID: 8G50 (65), yellow; PDB ID: 8G5P gray cartoon representation not shown; PDB ID: 8G5N, gray cartoon representation not shown), backtracking initiation (PDB ID: 8T7E (65), bluewhite), wedge alignment (PDB ID: 8G5K (65), bright orange), primer separation (PDB ID: 8G5L (65), sand). All PDB accession codes are color matching with the cartoon representation of the structure. RMSDs are globally calculated with all atoms as performed in PyMol (66). The inset which describes the PolG L-helix alignments were performed as in Figure 3C. Inset images of active sites were aligned using chain A in the structures, i.e. all atoms alignment of PolG.

Figure 5.

PolG2 clinical variants and implications of LonP1-dependent digestion of PolG. (A) Cylindrical representation and zoom-in on PDB ID: 8V55 (Polγ^open). Residues shown in stick or ball representation are PolG2 clinical mutations. PolG is colored according to Figure 1A and the proximal PolG2 is colored light orange the distal PolG2 is colored dark yellow. (B) Left: Quantification of LonP1 (0.42 μM, hexamer)-mediated PolG (1.25 μM) degradation over 60 min in the absence of PolG2 or in the presence of PolG2^WT, PolG2^G451E and PolG2^R369G (PolG2, 2.5 μM, dimer). The final salt of this reaction was 78 mM NaCl and 47 mM KCl. The amount of PolG remaining was normalized according to the following formula: x_normalized= (x – x_minimum) / range of x. The error bars are displayed as the coefficient of variation in three independent experiments. Right: Representative SDS-PAGE gel from the experimental quantification. (C) Quantification of LonP1 (0.42 μM, hexamer)-dependent degradation of PolG (0.625 μM) over 60 min in the absence of PolG2 (square) or in the presence of PolG2^WT (triangle), PolG2^G451E (diamond), or PolG2^R369G (circle), where PolG2 monomeric concentration is varied between 0.625 and 10 μM, or between 1:0.5 molar ratio to 1:8 PolG:PolG2 in a heterotrimer complex. The final salt of the reaction was 46 mM NaCl and 64 mM KCl. The error bars are displayed as standard deviation of three independent experiments.

Figure 6.

Proposed model of LonP1-mediated degradation of PolG in the presence and absence of PolG2 clinical mutations. PolG (blue) (EMD-42982, this study), in the presence, of a functional PolG2 (yellow) (EMD-25764) (33) forms the active Polγ complex (burnt orange), first, binding DNA and forming Polγ^open (1—EMD-42980), then engaging DNA forming Polγ^engaged (2—42979) and when provided the appropriate nucleotides, forms an actively replicating Polγ^active (3-EMD-42984). However, in the presence of a dysfunctional PolG2, such as a PolG2 clinical mutations, PolG2^R369G or PolG2^G451E, the Polγ complex association is weakened or unable to form and monomeric APO PolG (blue) becomes a target for LonP1-mediated degradation (maroon, EMD-12315(50)). All cryo-EM maps shown in surface representation and generated by ChimeraX.