Abstract
Mitochondrial rRNA modifications are essential for mitoribosome
assembly and its proper function. The m^4C methyltransferase METTL15
maintains mitochondrial homeostasis by catalyzing m^4C839 located in
12 S rRNA helix 44 (h44). This modification is essential to fine-tuning
the ribosomal decoding center and increasing decoding fidelity
according to studies of a conserved site in Escherichia coli. Here, we
reported a series of crystal structures of human METTL15–hsRBFA–h44–SAM
analog, METTL15–hsRBFA–SAM, METTL15–SAM and apo METTL15. The structures
presented specific interactions of METTL15 with different substrates
and revealed that hsRBFA recruits METTL15 to mitochondrial small
subunit for further modification instead of 12 S rRNA. Finally, we
found that METTL15 deficiency caused increased reactive oxygen species,
decreased membrane potential and altered cellular metabolic state.
Knocking down METTL15 caused an elevated lactate secretion and
increased levels of histone H4K12-lactylation and H3K9-lactylation.
METTL15 might be a suitable model to study the regulation between
mitochondrial metabolism and histone lactylation.
Subject terms: Post-translational modifications, Nanocrystallography
Introduction
As a major source of ATP, mitochondria are essential for the process of
cellular energy production through oxidative phosphorylation (OXPHOS).
Mitochondrial dysfunction causes defects in OXPHOS activity, disorder
of cell energy metabolism, increases in reactive oxygen species (ROS),
oxidative stress and cell death, which are closely related to aging,
diabetes, neurodegenerative diseases, cardiovascular diseases and
cancers^[38]1–[39]6.
Mammalian mitochondria retain a unique ~16.5 kb circular DNA genome,
which encodes 2 mitochondrial ribosomal RNAs (12 S and 16 S), 22
mitochondrially-encoded transfer RNAs and 13 protein components of the
OXPHOS system. Mammalian mitochondrial ribosomes (mitoribosomes), which
are responsible for the translation of the 13 protein component genes,
are composed of the 28 S mitochondrial small subunit (mt-SSU,
containing 12 S rRNA and 30 nucleus-encoded mitoribosomal proteins) and
39 S mitochondrial large subunit (mt-LSU, containing 16 S rRNA and 52
nucleus-encoded mitoribosomal proteins)^[40]7–[41]9. A series of
post-transcription modifications have been found in mitochondrial
rRNAs, which are located at the functionally critical regions of the
mitoribosome and play a crucial role in mitoribosome assembly and
efficient translation^[42]10–[43]12. Abnormal expression of mt-rRNA
modification enzymes directly affects modification levels of mt-rRNA
and mitoribosome assembly, which damages mitochondria function and
leads to a series of mitochondrial diseases^[44]13,[45]14. For example,
methyltransferase TFB1M, a potential risk gene for type 2 diabetes in
humans, dimethylates 12 S rRNA A936 and A937 and regulates the assembly
of mitoribosome^[46]15. Tfb1m homozygous knockout (KO) mice were
embryonically lethal, while mitochondrial damage and decreased insulin
secretion in response to glucose occurred in the islets of mice
heterozygous for Tfb1m deficiency^[47]16,[48]17.
Methyltransferase-like (METTL) family proteins are characterized by a
conserved Rossman-like fold S-adenosyl methionine (SAM)-binding domain,
which methylates proteins, nucleic acids, and other small molecule
metabolites, are involved in the regulation of mRNA stability and
translation efficiency^[49]18–[50]20. Several METTLs localize to
mitochondria, among which METTL9, METTL12 and METTL20 are responsible
for protein methylation, while METTL8 and METTL2A for mt-tRNA
methylation, METTL15 for mt-rRNA methylation^[51]10,[52]21–[53]23.
The N4-methylcytidine (m^4C) methyltransferase METTL15 has been
identified as responsible for specific recognition and modification of
m^4C839 in the 44th cervical ring structure of mt-SSU 12 S rRNA (helix
44)^[54]24–[55]26. The 12 S mt-rRNA m^4C839 modification is highly
conserved to that in Escherichia coli SSU 16 S rRNA m^4C1402, which is
modified by homology methyltransferase RsmH^[56]27–[57]29. In both
mitoribosomes and E. coli ribosomes, m^4C839 (of mitoribosomes) and
m^4C1402 (of E. coli ribosomes) modifications are located near the
P-site codon of mRNA, which was speculated to play a role in
fine-tuning P-site and increasing decoding fidelity^[58]30.
Inactivation of human METTL15 (or mouse Mettl15) perturbs the
translation of mitochondrial protein-coding mRNAs and decreases
mitochondrial respiration capacity^[59]24–[60]26. Studies have reported
that METTL15 may be highly correlated with childhood obesity^[61]31.
Moreover, the assembly and stability of mitoribosomes also depend on
mitochondrial ribosome-binding factors, such as hsRBFA. Acting as a
scaffold protein, hsRBFA is recruited to mt-SSU 12 S rRNA helix 44 and
helix 45, promoting proper decoration, assembly and maturation of
mt-SSU^[62]32. Recent studies have shown that hsRBFA adapts a large
conformational change, which ensures that TFB1M completes the
dimethylation of helix 45 and then recruits METTL15 for further
modification^[63]33. Our previous research has shown that hsRBFA
recognizes 12 S rRNA via a novel binding mode. Binding of hsRBFA to
12 S rRNA was not only dependent on its KH-like domain but also
dependent on the basic amino acid of its N-terminus, which effectively
promotes the proper assembly of mitoribosomes^[64]34.
Understanding the molecular mechanisms that govern the interaction
between METTL15, hsRBFA, and 12 S rRNA is critical for unraveling
regulatory role of METTL15 in mitochondrial gene expression. In this
study, we determined four crystal structures of the methyltransferase
METTL15 in complex with different substrates. By analyzing these
structures, we identified a series of key residues in METTL15 and
hsRBFA that are responsible for specific recognition. Moreover, METTL15
recognized RNA substrates using the second scaffold-like domain (domain
2) through electrostatic interactions between the domain 2 and the
sugar-phosphate backbone of RNA, indicating that hsRBFA might be the
key factor in recruiting METTL15 to mt-SSU, instead of 12 S rRNA.
Furthermore, knocking down METTL15 affected OXPHOS activity and
cellular metabolic state, leading to increased ROS and membrane
potential. Hence, through a combination of biochemical, structural, and
functional approaches, we provided detailed understanding of the
specific recognition mechanism underlying the interaction between
METTL15, hsRBFA and 12 S rRNA, shedding a new light on the regulatory
role of METTL15 in mitochondrial gene expression.
Results
Overall structure of human METTL15 in complex with SAM
Previous studies have revealed that METTL15 methyltransferase is
responsible for the formation of the m^4C839 residue of human
mitochondrial 12 S rRNA^[65]24–[66]26. To further investigate the
mechanism of m^4C839 modification, we determined the structures of apo
METTL15 and the binary complex METTL15–SAM (Fig. [67]1a, b;
Supplementary Fig. [68]S1b). Comparison of these two structures
revealed that the root-mean-square deviation (r.m.s.d.) of the Cα atom
was only 0.145 Å (269 to 269 atoms). Therefore, we mainly used the
METTL15–SAM complex for subsequent description. In the latter model,
the METTL15–SAM complex included two molecules in an asymmetric unit
both with observable electronic densities for SAM and most residues of
METTL15, except for part of the loop (Asn^362~Ala^371 and
Ser^384~Gly^396) (Fig. [69]1a). The detailed crystallographic
statistics were summarized in Table [70]1.
Fig. 1. Structural overview of METTL15 in complex with SAM.
[71]Fig. 1
[72]Open in a new tab
a Cartoon representation of the METTL15–SAM binary complex. The two
parts of MTases domain (domain 1) are colored in cyan and palecyan, the
scaffold-like domain (domain 2) in slate and SAM in magenta,
respectively. The missing loops of METTL15 are represented as cyan
dotted lines. b The electrostatic potential of the METTL15–SAM complex
is shown, in which positively charged, negatively charged and neutral
areas are represented in blue, red and white, respectively. The SAM
active pocket of METTL15 is highlighted as solid red circle. c The
detailed interactions between METTL15 (slate) and the SAM (magenta).
The 2Fo-Fc electron-density map of SAM is contoured at 1.0 σ (gray). d
Schematic representations of the recognition of SAM (colored purple and
labeled in red) by METTL15 (colored bright orange and labeled in
black). e The ITC fitting results of METTL15 WT and mutants by SAM.
Table 1.
Data collection and refinement statistics.
Data collection METTL15^70–407 apo METTL15^70–407–SAM
METTL15^70–407–RBFA^226–260–SAM METTL15^70–407–slRNA1–RBFA^226–260–SFG
Beamline 19U, SSRF 18U, SSRF 18U, SSRF 19U, SSRF
Space group P3[1]21 P3[1]21 P2[1]2[1]2[1] C222[1]
PDB code 8IPI 8IPK 8IPL 8IPM
Wavelength (Å) 0.9789 0.9789 0.9789 0.9789
Resolution (Å) 31.34-2.10 (2.14-2.10)^a 25.25-1.90 (1.93-1.90)^a
33.48-2.20 (2.28- 2.20) 33.53-3.10 (3.21-3.10)
Cell dimensions
a, b, c (Å) 70.61, 70.61, 136.063 70.50, 70.50, 134.674 63.17, 70.98,
75.93 71.22 134.11 93.03
α, β, γ (°) 90, 90, 120 90, 90, 120 90, 90, 90 90, 90, 90
Unique reflections 23694 (1134) 31270 (1647) 16587 (789) 8204 (888)
Completeness (%) 100 (100) 99.9 (100) 92.61 (97.65) 97.42 (98.90)
Redundancy 17.9 (17.8) 15.0 (15.4) 8.9 (8.9) 9.0 (9.7)
I/σI 15.8 (2.8) 23.2 (2.2) 16.5 (6.2) 14.7 (3.7)
R[merge] (%) 14.8 (80.0) 12.6 (100.3) 11.1 (50.7) 10.5 (61.6)
Refinement
R[work] (%) 19.32 18.23 17.99 25.25
R[free] (%) 22.91 22.07 23.46 28.78
No. of atoms 2307
Protein 2384 2373 2264
RNA 296
Water 110 113 32
Average B factors (Å^2)
Protein 29.05 32.91 32.26 84.49
RNA 178.65
Water 31.19 32.43 27.32
Root mean square deviations
Bond lengths (Å) 0.007 0.006 0.008 0.002
Bond angles (°) 0.793 0.784 0.979 0.537
Ramachandran plot
Favored (%) 97.67 98.36 97.66 97.29
Allowed (%) 2.33 1.64 2.34 2.71
Disallowed 0 0 0 0
[73]Open in a new tab
^aValues for the highest-resolution shell are shown in parentheses.
Similar to other members of rRNA methyltransferases, METTL15 has a
two-domain structure: an MTase domain (domain 1) and a scaffold-like
domain (domain 2). The domain 1 of METTL15 was slightly different from
the conserved Rossman-like fold. The structure still folded with a
central seven-stranded β-sheet flanked by five α-helices on each side,
but the sequence was composed of two parts: the N-terminal sequence
(Pro^74 to Thr^179) and the C-terminal sequence (Glu^295 to Leu^407)
(Fig. [74]1a). The domain 2 of METTL15 contained five α-helices and one
η-helix, which formed a scaffold for the binding of RNA substrates or
other functional proteins (Fig. [75]1a). The SAM ligand fit snugly in
the binding pocket with well-covered density (Fig. [76]1c). The
conserved active pocket of METTL15 to accommodate SAM was surrounded by
Thr^96, Ser^99, His^102, Asp^119, Arg^120, Gln^145, Phe^146, Asp^169,
Ser^173, Gln^176, Glu^298, and etc., which established hydrogen bonds,
salt bridges and hydrophobic interactions with SAM (Fig. [77]1c, d).
The base of SAM was sandwiched between Arg^120 and Phe^146, with its
adenine ring packed with the aromatic ring of Phe^146 via favorable π–π
interactions. Additionally, the adenine ring formed multiple hydrogen
bonds with the backbone nitrogen of Arg^120, Phe^146, and the carbonyl
group of Glu^298. Two hydroxyl groups in the ribose moiety of SAM
formed hydrogen bonds with the side chains of Asp^119 and Gln^176,
respectively. Finally, the carboxyl group of SAM formed several
hydrogen bonds with Thr^96, Ser^99, His^102 and Asp^169. The above
residues involved in SAM binding (either their sequences or their
structures) were conserved from prokaryotes to eukaryotes
(Supplementary Fig. [78]S1a). Comparison of the electrostatic potential
surface of these two structures showed that the positive charge at the
SAM-binding pocket of METTL15 was significantly enriched after SAM
binding (Fig. [79]1b; Supplementary Fig. [80]S1b). Then, we mutated
several residues of METTL15 SAM pocket (D119A/R120A, D169A and E298A)
and employed isothermal titration calorimetry (ITC) assays to evaluate
their contributions to SAM binding. The affinity of wild-type (WT)
METTL15 for SAM was in the micromolar range (2.99 ± 0.36 μM) (Fig.
[81]1e; Table [82]2). All these mutants weakened the interactions of
METTL15 with SAM, particularly D119A/R120A and E298A, which almost
eliminated the association capacity (Fig. [83]1e; Table [84]2;
Supplementary Fig. [85]S1c).
Table 2.
The thermodynamic parameters of the ITC experiments.
METTL15^70–407 RBFA SAM ΔH
kcal/mol -TΔS
kcal/mol K[D]
μM
METTL15[WT] / SAM –17.20 ± 1.82 9.78 2.99 ± 0.36
METTL15[D119A/R120A] / SAM N. D.
METTL15[D119A] / SAM –31.70 ± 6.14 25.3 17.30 ± 0.31
METTL15[E298A] / SAM N. D.
METTL15[WT] RBFA^226-260[WT] / –12.80 ± 0.11 4.42 0.62 ± 0.03
METTL15[WT] RBFA^226-276[WT] / –14.90 ± 0.27 6.44 0.46 ± 0.08
METTL15[WT] RBFA^226-260[C237A] / Weak binding
METTL15[WT] RBFA^226-260[H241A] / –14.90 ± 0.55 8.22 11.30 ± 0.85
METTL15[WT] RBFA^226-260[N245A] / –10.60 ± 0.20 2.26 0.61 ± 0.09
METTL15[WT] RBFA^226-260[Q247A] / –11.50 ± 0.20 2.92 0.41 ± 0.07
METTL15[WT] RBFA^226-260[I248A] / Weak binding
METTL15[WT] RBFA^226-260[Y251A] / Weak binding
METTL15[WT] RBFA^226-260[K252A] / –17.20 ± 3.10 10.90 19.60 ± 6.48
METTL15[WT] RBFA^226–260 [Y251A/K252A] / N. D.
METTL15[WT] RBFA^226–260 [H241A/Y251A/ K252A] / N. D.
METTL15[D209A] RBFA^226-260[WT] / –7.43 ± 0.32 0.583 7.94 ± 0.07
METTL15[N212A] RBFA^226-260[WT] / –9.50 ± 0.34 1.51 1.12 ± 0.03
METTL15[D215A] RBFA^226-260[WT] / Weak binding
METTL15[F291A] RBFA^226-260[WT] / Weak binding
METTL15[D209A/F291A] RBFA^226-260[WT] / N. D.
METTL15[D209A/N212A/D215A] RBFA^226-260[WT] / N. D.
[86]Open in a new tab
K[D] dissociation constant, ΔH binding enthalpy, -TΔS binding entropy,
N. D. Not Detected. Each K[D] value is presented as fitted value ±
error.
The C-terminal helix of hsRBFA is essential for METTL15 recognition
As a mitochondrial ribosomal assembly factor, hsRBFA binds to METTL15,
which promotes further rRNA maturation and a large conformational
change in hsRBFA^[87]25,[88]33. We validated the recognition of hsRBFA
by METTL15^70–407 using pull-down experiments with various hsRBFA
constructs in vitro. The results showed that the MBP-hsRBFA^42–343
could interact with METTL15^70–407, but the interaction was abolished
in MBP-hsRBFA^△226–260 (Fig. [89]2a, b; Supplementary Fig. [90]S2a).
Further pull-down assays revealed that the C-terminal helix of
hsRBFA^226–276 was responsible for METTL15 recognition (Fig. [91]2c;
Supplementary Fig. [92]S2a). Moreover, according to the ITC assays, the
K[D] values of METTL15 with hsRBFA^226–276 and hsRBFA^226–260 were
0.46 ± 0.08 μM and 0.62 ± 0.03 μM, respectively (Fig. [93]2d).
Fig. 2. METTL15 recognized the C-terminal helix of hsRBFA.
[94]Fig. 2
[95]Open in a new tab
a–c Interactions of different MBP-tagged hsRBFA constructs with METTL15
visualized by Coomassie blue staining. The indicated MBP-hsRBFA fusion
proteins or MBP alone were incubated with METTL15. The complexes were
collected with glutathione-agarose resin and bound proteins were eluted
and then subjected to SDS-PAGE. MBP or MBP-hsRBFA fusion proteins
without METTL15 are shown as a negative control. d The ITC fitting
results of WT METTL15 by the C-terminal helix of hsRBFA.
To further characterize the interactions between METTL15 and the
C-terminal helix of hsRBFA, we crystallized and determined the complex
structure of METTL15^70–407 with SAM and hsRBFA^226–260. The
METTL15^70–407–hsRBFA^226–260–SAM ternary complex structure was
subsequently refined to a resolution of 2.2 Å in space group P 2[1].
The crystal structure was solved by molecular replacement using the
structure of apo METTL15^70–407 solved previously as the search model.
Finally, the R[work] and R[free] of the ternary complex structure were
refined to 17.99% and 23.46%, respectively. The detailed
crystallographic statistics were summarized in Table [96]1.
In the ternary structure, METTL15 formed a 1:1:1 complex with SAM and
hsRBFA^226–260 in an asymmetric unit (Fig. [97]3a). METTL15 and SAM in
this ternary complex adopted a similar conformation with that in the
METTL15–SAM complex, displaying an overall r.m.s.d. for Cα atoms of
0.361 Å. Most of the hsRBFA^226–260 (Leu^236~Asp^255) were visible in
the electron density map (Supplementary Fig. [98]S2c), forming as a
helix and lying on the side of METTL15 domain 2 surface (Fig. [99]3a,
c). Superposition of the METTL15^70–407–hsRBFA^226–260–SAM ternary
complex and the METTL15–SAM complex showed that the η-helix
(Thr^339~Glu^361) of METTL15 domain 1 posed a physical barrier with
hsRBFA helix and was not observed in the ternary complex (Supplementary
Fig. [100]S2e), indicating that the η-helix of METTL15 domain 1 is
flexible, which may undergo a conformational change as a switch when
METTL15 binds to hsRBFA.
Fig. 3. Structure of METTL15 in complex with hsRBFA^226–260 and SAM.
[101]Fig. 3
[102]Open in a new tab
a Cartoon representation of the METTL15^70–407–hsRBFA^226–260–SAM
ternary complex in two orientations related by a 180° rotation around a
vertical axis. The asymmetric unit consists of the METTL15 (cyan),
hsRBFA (bright orange) and SAM (magenta). b Schematic of hsRBFA
interactions with METTL15, colored as described in (a). The red and
black dotted lines indicate contacts mediated by hydrogen bonds and
water-bridged hydrogen bonds, respectively. c The C-terminal helix of
hsRBFA are represented as sticks on the molecular face of METTL15. The
positively charged, negatively charged and neutral areas are
represented in blue, red and white, respectively. d Higher
magnification views of individual interactions between METTL15 (cyan,
labeled blue) and hsRBFA (bright orange, labeled black). Hydrogen bonds
are indicated with black dotted lines.
The interactions between METTL15 and hsRBFA^226–260 were shown in Fig.
[103]3b, including hydrogen-bonding interactions and hydrophobic
interactions. The helix of hsRBFA spanned one hydrophobic surface of
METTL15 (Fig. [104]3c). The side chains of hsRBFA Cys^237, Ile^239,
His^241, Asn^245 and Ile^248 inserted into several wide and shallow
hydrophobic pockets of METTL15, resulting in specific recognition
between METTL15 and hsRBFA (Fig. [105]3c; Supplementary Fig. [106]S2d).
Moreover, the main chain of hsRBFA Leu^236 formed a hydrogen bond with
the peptide carbonyl group of METTL15 Asn^297 (Fig. [107]3d). The side
chain of hsRBFA His^241 protruded into the aromatic cage surrounded by
METTL15 Tyr^226, Phe^291, Asn^294, and Asn^297, and formed a direct
hydrogen bond with the carbonyl of Phe^291 side chain (Fig. [108]3d).
The hydroxyl group of hsRBFA Gln^247 side chain and carbonyl group of
hsRBFA Tyr^251 side chain formed specific hydrogen bonds with the
carbonyl group of METTL15 Asp^215 side chain and hydroxyl group of the
side chain of METTL15 Asn^212, respectively (Fig. [109]3d). HsRBFA
Lys^252 contacted the side chain of METTL15 Asp^209 via the positively
charged side chain (Fig. [110]3d).
To confirm the specific recognition between METTL15 and hsRBFA^226–260
in vitro, we introduced alanine mutations into METTL15 and
hsRBFA^226–260, respectively. Subsequently, we performed ITC assays to
measure the binding affinities between WT METTL15 and hsRBFA^226–260
mutants or METTL15 mutants and WT hsRBFA^226–260 (Table [111]2;
Supplementary Figs. [112]S3, [113]S4). As expected, the introduction of
H241A, Y251A and K252A single mutations on hsRBFA^226–260 resulted in
more than 10-fold decrease in the binding affinities to hsRBFA^226–260
with METTL15 (K[D] = 11.30 ± 0.85 μM, 11.30 ± 0.85 μM and
11.30 ± 0.85 μM, respectively), compared with 0.62 ± 0.03 μM for WT
hsRBFA^226–260 (Supplementary Fig. [114]S3). HsRBFA^226–260 C237A and
I248A mutants also exhibited substantial reduction in binding
affinities to METTL15 (Supplementary Fig. [115]S3). It is not
surprising that hsRBFA^226–260 Y251A/K252A and H241A/Y251A/K252A
mutations abolished the binding to METTL15 completely (Supplementary
Fig. [116]S3). Similarly, METTL15 single mutations (D209A, D215A and
F291A) notably decreased the binding affinities between METTL15 and
hsRBFA^226–260 (Supplementary Fig. [117]S4). Moreover, METTL15 double
mutation (D209A/F291A) or triple mutation (D209A/N212A/D215A) also
abolished hsRBFA^226–260 binding (Supplementary Fig. [118]S4). Circular
dichroism (CD) spectral analyses confirmed that all mutants of METTL15
and hsRBFA^226–260 maintained a secondary structure composition similar
to that of the WT proteins.
METTL15 recognizes RNA substrates by electrostatic interactions in vitro
To further explore the RNA substrates, we employed fluorescent
polarization (FP) assays to test a series of RNAs for their binding
affinities to METTL15. METTL15 bound to dsRNA (h44_dsRNA1) or ssRNA
(h44_ssRNA3) derived from h44 around the 12 S rRNA m^4C839 residue,
with K[D] values of 1.10 ± 0.09 μM and 2.20 ± 0.33 μM, respectively
(Fig. [119]4b). Interestingly, the GAAG tetraloop substitution of the
native loop in dsRNA (h44_dsRNA2) didn’t affect the binding ability
between METTL15 and RNA (Fig. [120]4b; Supplementary Fig. [121]S5a).
Additionally, METTL15 could effectively bind to the ssRNA (polyA13)
with non-sequence-specificity (Fig. [122]4b). Furthermore, adding
hsRBFA^226–260 didn’t affect the binding affinities between METTL15 and
RNA substrates (Table [123]3).
Fig. 4. Structure of METTL15 in complex with RNA, hsRBFA^226–260 and SFG.
[124]Fig. 4
[125]Open in a new tab
a Cartoon representation of the METTL15^70–407–RNA–hsRBFA^226–260–SFG
quaternary complex in two orientations related by a 180° rotation
around a vertical axis. The asymmetric unit consisted of the METTL15
(cyan), h44_dsRNA1 (slate), hsRBFA (bright orange) and SFG (magenta).
In the left pannel, the MTases domain (domain 1) is colored in cyan and
the scaffold-like domain (domain 2) in slate, respectively. b The
binding affinities of 5’-FAM-labeled RNAs for METTL15 determined by FP
experiments are shown. c The RNA and C-terminal helix of hsRBFA are
represented as sticks binding to the either side of METTL15,
respectively. The positively charged, negatively charged and neutral
areas are represented in blue, red and white, respectively. d Higher
magnification views of individual interactions between METTL15 (cyan,
labeled blue) and h44_dsRNA1 (slate). Hydrogen bonds are indicated with
yellow dotted lines. e Superposition of the
METTL15^70–407–RNA–hsRBFA^226–260–SFG quaternary complex with the
cryo-electron structure of pre-mt-SSU (PDB ID 7PNX). The
METTL15^70–407–RNA–hsRBFA^226–260–SFG quaternary complex colored as
described in (a), except the h44_dsRNA1 colored red. The METTL15 and
hsRBFA in pre-mt-SSU are colored magenta and light pink, respectively.
Table 3.
Binding affinities of WT and mutant METTL15 proteins to different RNA
substrates by FP assays.
Protein RBFA RNA K[d] (μM)
METTL15 (wildtype) / h44_dsRNA1 1.10 ± 0.09
wildtype / h44_dsRNA2 0.77 ± 0.05
wildtype / h44_ssRNA3 2.20 ± 0.33
wildtype / h44_dsRNA4 2.33 ± 0.13
wildtype / polyA13 0.61 ± 0.11
wildtype +RBFA h44_dsRNA1 1.15 ± 0.08
wildtype +RBFA h44_ssRNA3 1.76 ± 0.16
H231A/K234A/F262A / h44_dsRNA1 N. D.
F262A/P263A/K270A / h44_dsRNA1 N. D.
R327A/K330A/R331A / h44_dsRNA1 1.03 ± 0.15
K380A/K381A / h44_dsRNA1 2.68 ± 0.29
[126]Open in a new tab
We then determined the binding mode between METTL15 and the RNA
substrate (h44_dsRNA1 in Supplementary Fig. [127]S5a), together with
hsRBFA^226–260 and S-adenosyl-methionine analog sinefungin (SFG). The
METTL15^70–407–RNA–hsRBFA^226–260–SFG quaternary complex structure was
subsequently refined to a resolution of 3.1 Å in space group C222[1].
The crystal structure was solved by molecular replacement using the
structure of ternary complex solved before as the search model.
Finally, the R[work] and R[free] of the quaternary complex structure
were refined to 25.25% and 28.79%, respectively. The detailed
crystallographic statistics were summarized in Table [128]1.
In the structure of the quaternary complex (Fig. [129]4a),
METTL15–RNA–hsRBFA^226–260–SFG adopted a similar conformation to that
in the METTL15^70–407–hsRBFA^226–260–SAM ternary complex, displaying an
overall r.m.s.d. for Cα atoms of 0.475 Å. The helix of hsRBFA^226–260
and the RNA substrate bound on either side of METTL15 domain2 (Fig.
[130]4c). Unexpectedly, both the 5’ end and the 3’ end of h44_dsRNA1
bases in the quaternary complex partially paired with the neighboring
RNA in the next symmetry equivalent for better crystallization,
respectively (Supplementary Fig. [131]S5b, c).
Consistent with the previous FP assays, METTL15 interacted with the
phosphodiester backbone of the RNA instead of the bases (Fig. [132]4d).
The residues in the N-terminus of helix α7 and in the loop between
helices α8 and α9 of METTL15 formed a positively charged channel to
accommodate and recognize the RNA substrate. METTL15 His^231, Lys^234
and Arg^270 formed several direct hydrogen bonds with the
phosphodiester backbone of the RNA mainly via the positively charged
side chain, while Phe^262 and Pro^263 recognized the phosphodiester
backbone of the RNA via the hydrophobic interactions (Fig. [133]4d).
The side chains of Pro^263, Ala^266 and Thr^269 recognized the ribose
of the RNA via the hydrophobic interactions (Fig. [134]4d).
Furthermore, the introduction of H231A/K234A/F262A or F262A/P263A/K270A
mutations to METTL15 abolished the binding affinities to RNA substrate
(Table [135]3). The electrostatic potential of the surface of METTL15
showed that domain1 of METTL15 also contained a positively charged
region, which was reported to recognize the h24 of 12 S rRNA in the
cryo-electron structures of pre-mt-SSU (PDB ID 7PNX, 7PNY and
7PNZ)^[136]33 (Fig. [137]4c, e). However, mutants R327A/K330A/R331A or
K380A/K381A with key residues of this region being replaced showed
comparable RNA binding affinities to WT METTL15 (Table [138]3).
Therefore, METTL15 domain 2 plays a key role in RNA substrate
recognition.
Taken together, our biochemical and structural results showed that
METTL15 recognizes RNA substrates without sequence specificity in
vitro, which indicated the binding between METTL15 and its RNA
substrate was contributed by the electrostatic interactions. HsRBFA,
instead of RNA, might be the key factor in recruiting METTL15 to mt-SSU
for m^4C839 modification.
METTL15 knockdown caused significantly altered cellular metabolic state
As our in vitro assays revealed the recognition mode between METTL15,
hsRBFA and RNA, we tried to further study the biological function of
METTL15 and hsRBFA on mitochondria in HEK293T cells. We knocked down
METTL15 in HEK293T cells using three different shRNAs (sh1, sh2 and
sh3, Supplementary Table [139]S1), among which sh2 downregulates the
levels of METTL15 mRNA and protein by more than 70%, according to the
qRT-PCR and western blotting results (Fig. [140]5a; Supplementary Fig.
[141]S6a). Moreover, m^4C in bacteria, mediated by the E. coli
homologous methyltransferase RsmH may play a role in fine-tuning
ribosomal decoding center and increasing decoding fidelity^[142]30.
Therefore, we explored the impact of METTL15 knockdown on the
translation of protein components of OXPHOS system encoded by the
mitochondrial genome. The results showed that mitochondrial
gene-encoded protein (mt-protein) expressions were decreased upon
METTL15 knockdown, especially in the sh2 group. Furthermore, to
investigate whether the key residues we identified previously
influenced the mt-protein translation in vivo, we expressed different
METTL15 mutants in cells to rescue deficiency of METTL15 induced by
knockdown (Supplementary Fig. [143]S6b). The results were consistent
with our in vitro studies showing that elimination of the interactions
between METTL15 with either hsRBFA or RNA affected mt-protein
translation.
Fig. 5. Effect of METTL15 knockdown on mitochondrial functions.
[144]Fig. 5
[145]Open in a new tab
a Western blot analyzing protein levels of METTL15 and mitochondrial
gene-encoded proteins in OXPHOS complexes in HEK293T cells transfected
with different shRNAs. β-tubulin served as a loading control. NTC,
non-targeting control. b Flow cytometry analyzing mitochondrial
membrane potential in HEK293T cells. The mean and standard deviation of
mean fluorescence intensity (MFI) for three repetitions are shown as
bars. shNTC, non-targeting control. c Flow cytometry analyzing
mitochondrial reactive oxygen species. The mean and standard deviation
of MFI for three repetitions are shown as bars. d Seahorse analyzing
oxidative phosphorylation levels upon METTL15 knockdown. e Seahorse
analyzing glycolysis upon METTL15 knockdown.
Besides, knockdown of METTL15 might lead to dysfunction of OXPHOS
complex by impacting functions of the proteins translated in
mitochondria, which results in malfunction of proton pump gradients
across inner mitochondrial membrane and imbalance of the ROS.
Therefore, we monitored the mitochondrial membrane potential and ROS
levels by flow cytometry (Supplementary Figs. [146]S6c, d). The
membrane potential strikingly decreased, while ROS were increased in
METTL15 knockdown cells (Fig. [147]5b, c). Furthermore, the levels of
mitochondrial OXPHOS and glycolysis were detected using the Seahorse
real-time cell metabolic analysis. The level of OXPHOS decreased, while
the level of glycolysis increased (Fig. [148]5d, e), which indicated
that cellular metabolic state in METTL15 knockdown cells was changed.
To determine the cellular metabolic alterations in METTL15 knockdown
cells, we applied untargeted metabolomic analysis to METTL15 knockdown
cells and used the shNC cells as a control. All samples were processed
and analyzed using UPLC-MS/MS following the standardized protocol. In
total, 3,525 metabolite peaks were identified, among which 1075
metabolites were identified based on MS/MS spectra through the BGI
reference library, MzCloud, KEGG and HMDB (Supplementary Table
[149]S3), with coefficients of relative standard deviation < 0.30
across quality control (QC) samples. 91 out of 1075 metabolites were
significantly associated with METTL15 knockdown (q value < 0.05) (Fig.
[150]6a). Compared to control cells, the volcano plots showed that 45
metabolites were significantly increased (fold change ≥ 1.2) and 46
were decreased (fold change ≤ 0.83) in METTL15 knockdown cells (Fig.
[151]6b). Consistent with our Seahorse real-time cell metabolic
analysis, the metabolites were implicated in the decreased citrate
cycle and increased glycolysis (2-oxoglutaric acid and L-(+)-lactic
acid), respectively (Fig. [152]6c), which indicated a metabolic switch
from mitochondrial OXPHOS to aerobic glycolysis. The level of lactate
secretion increased ~1.3-fold in METTL15 knockdown cells compare with
WT cells, which was more obvious in METTL15 KO cells as studied by
previous study^[153]26. As reported, the increase in intracellular
metabolite lactate might raise the level of lactylation modification of
certain proteins in the nucleus or cytoplasm, which in turn regulates
the transcription and expression of related genes^[154]35–[155]37. To
explore the effect of elevated lactate secretion in METTL15 knockdown
cells, we tested the level of several histone lactylation in METTL15
knockdown cells and WT cells, respectively. The western blotting
results showed that levels of H4K12-lactylation (H4K12-la) and
H3K9-lactylation (H3K9-la) were increased in METTL15 knockdown cells
(Supplementary Fig. [156]S7a). Moreover, metabolic pathway enrichment
analysis of differentially abundant metabolites based on KEGG database
showed that in METTL15 knockdown cells, many metabolic pathways were
changed (Supplementary Fig. [157]S8a). Some metabolites involved in
pathways such as tricarboxylic acid (TCA) cycle, pyruvate metabolism,
D-Glutamine and D-glutamate metabolism, D-arginine and D-ornithine
metabolism, were dysregulated upon METTL15 knockdown (Fig. [158]6c;
Supplementary Fig. [159]S8b). Taken together, a proper function of
METTL15 is critical to mitoribosome biogenesis and cellular metabolic
state maintenance.
Fig. 6. METTL15 knockdown alters cellular metabolic state.
[160]Fig. 6
[161]Open in a new tab
a 91 significantly differential metabolites (q value < 0.05) between
shNC cells and METTL15 sh2 knockdown cells are shown in the heatmap. b
The volcano diagram shows different metabolites of the two groups.
(Fold change ≥ 1.2 or Fold change ≤ 0.83). c Two metabolites implicated
in TCA cycle and glycolysis are compared and shown as histograms.
Discussion
METTL15 was reported as a methyltransferase responsible for the m^4C839
modification in the mt-SSU 12 S rRNA h44. A recent cryo-electron
microscopy study on the structure of the mt-SSU suggested that hsRBFA
promotes TFB1M binding and dimethylation of mt-SSU rRNA h45, which
enables METTL15 binding and promotes further rRNA maturation^[162]33.
In our study, we determined a series of crystal structures of METTL15
in complex with different substrates and analyzed specific interactions
of METTL15 with SAM, hsRBFA and 12 S rRNA in detail. Since METTL15
recognizes RNA without sequence specificity, hsRBFA might be the key
factor for METTL15 recruitment to mt-SSU. We also observed reduced
amounts of mitochondrial genome-coding proteins, detected decreased
membrane potential and increased ROS in METTL15 knockdown cells,
accompanied by an abnormal cellular metabolic state. An increased level
of glycolysis led to elevated lactate secretion (Fig. [163]6c), which
induced the altered histone lactylation in the nucleus. Among the
lactating sites, levels of H4K12-la and H3K9-la increased most
significantly in METTL15 knockdown cells (Supplementary Fig. [164]S7a).
A recent study revealed that elevated histone H4K12-la is the most
common form of histone lactylation in microglia of Alzheimer’s disease
mice, which activates transcription of glycolytic genes and exacerbates
microglial dysfunction in AD^[165]38. It was also found that lactate
not only functions as a metabolic substrate to provide energy but can
also functions as a signaling molecule to modulate cellular functions
under pathophysiological conditions. Lactylation highlights the novel
role of lactate in regulating transcription, cellular functions and
disease pathogenesis^[166]39. Intriguingly, homozygous Mettl15 KO mice
survived and exhibited a much milder phenotype than Tfb1m^−/− mice, yet
decreased exercise and learning capability were detected^[167]40.
Furthermore, H3K9-la was recently identified to promote tumorigenicity
of liver cancer stem^[168]41. More detailed studies exploring the
regulation between METTL15 and histone lactylation are needed and
Mettl15 KO mice might be a suitable model.
The METTL15 ortholog RsmH is responsible for catalyzing m^4C1402 of the
SSU 16 S rRNA in E. coli. The alignment of our binary complex
METTL15–SAM structure with the RsmH structure (PDB ID 3TKA) showed that
the r.m.s.d. for Cα atoms of the two methyltransferases was 1.216 Å
(210 to 210 atoms) (Supplementary Fig. [169]S9a). The two structures
shared highly similar conformations, except for some flexible loops.
The key residues of active pocket to accommodate SAM were conserved
between METTL15 and RsmH (Supplementary Fig. [170]S1a). Strikingly, the
alignment of our quaternary complex
METTL15^70–407–RNA–hsRBFA^226–260–SFG structure with the RsmH structure
showed that the helix of hsRBFA posed a physical barrier with the
inactive cytidine in RsmH structure (Supplementary Fig. [171]S9b,
indicated with a red circle). As we described before, the C-terminal
helix of hsRBFA is essential for METTL15 incorporation into mt-SSU.
However, the C-terminal helix was apparently missing in RBFA of E. coli
(Supplementary Fig. [172]S9c), indicating that the incorporation
mechanisms of METTL15 and RsmH into small ribosomal subunits are quite
different.
Notably, METTL17, another METTL family member located in mitochondria,
was also reported to affect levels of m^4C839 and m^5C841 modifications
in 12 S mt-rRNA^[173]42. Likewise, inactivation of METTL17 also
resulted in loss of OXPHOS activity and affected cellular metabolome in
mouse embryonic stem cells. Nathan J. Harper, et al. recently reported
a linear assembly pathway of mt-SSU intermediates bound by METTL17, in
which METTL17 acts as an assembly factor and mediates the formation and
compaction of the head domain of mt-SSU intermediates^[174]43.
Moreover, METTL17 binds to mt-SSU rRNA h31 and prevents premature
association of METTL15. Then, METTL15 outcompetes METTL17 to complete
the final modification of the decoding center. METTL17 KO leads to an
immature head domain of intermediates, failing to recruit METTL15,
further showing reductions in m^4C839 and m^5C841 of 12 S
mt-rRNA^[175]43. However, METTL17 is not just as an assembly factor,
its function is dependent on its SAM-binding ability, as the
SAM-binding-deficient mutant of METTL17 fails to rescue the decreased
level of m^4C839 12 S mt-rRNA caused by METTL17 KO^[176]42. Therefore,
how does METTL17 function as an enzyme during mt-SSU assembly? Is there
a functional correlation between METTL15 and METTL17 during mt-SSU
assembly? Or are alternative pathways present in the cell? The assembly
of mt-SSU does not occur in a strictly linear manner^[177]44. In
addition, in WT HAP1 cells, the m^4C839 methylation level is only ∼14%,
which is strikingly lower than that of other rRNA methylated sites in
mitoribosomes^[178]24,[179]45–[180]47. We also found that protein
expression level of METTL17 was slightly elevated in METTL15 knockdown
cells (Supplementary Fig. [181]S7b), implying a correlation between
METTL15 and METTL17. We are currently working on the possible relevance
between these two m^4C methyltransferases.
The proper assembly and maturation of mitoribosomes are critical for a
proper function of the mitochondria. rRNA and tRNA modifications play
key roles in mitoribosome biogenesis and efficient and accurate protein
translation, which are tightly regulated. The studies of mitoribosome
assembly and RNA modification mechanisms have broad implications for
understanding mitochondrion biology and the pathogenesis of
mitochondrial diseases.
Materials and methods
Protein expression and purification
DNA encoding human METTL15 (residues 70–407) was amplified and
incorporated into a modified pET28a (Novagen) plasmid. The modified
pET-28a plasmid contained an N-terminal SUMO-tag and a ULP1 protease
cleavage site. The protein was expressed in E. coli Rosetta (DE3) cells
(Novagen) cultured in LB medium at 37 °C to OD[600] = 1.0, and then
shifted to 16 °C for 24 h after induction with 0.2 mM
isopropyl-β-D-thiogalactopyranoside (IPTG). Bacterial pellets were
resuspended in buffer A (50 mM Na[2]HPO[4], 1 M NaCl, pH 7.5) and lysed
by sonication on ice. The crude lysate was then centrifuged at
14,000 rpm for 30 min at 4 °C. The supernatant was applied to a Ni-NTA
column (QIAGEN), followed by size exclusion chromatography using a
Superdex 200 (GE Healthcare) column. After cleavage with ULP1 protease
overnight at 16 °C to remove the SUMO-tag, an additional HiTrap SP FF
column purification step was employed. The purified protein was
concentrated to ~10 mg/mL in buffer B (50 mM Na[2]HPO[4], 150 mM NaCl,
pH 7.0) and stored at –80 °C. All mutants were generated using a
MutanBEST kit (TaKaRa) and confirmed by DNA sequencing. The mutant
proteins were purified using the protocol described above.
For maltose binding protein (MBP) pulldown assays, DNAs encoding human
hsRBFA (residues 42–343, 42–201, 226–276 and 226–260) were amplified
and incorporated into a modified pET22b (Novagen) plasmid,
respectively. The modified pET-22b plasmid contained N-terminal 6× His
and MBP tags. The 6× His-tagged MBP-hsRBFA fused protein was expressed
in E. coli BL21-Gold (DE3) cells (Novagen) cultured in LB medium at
37 °C to OD[600] = 0.8 ~ 1.0, and then shifted to 16 °C for 24 h after
induction with 0.5 mM IPTG. Bacterial pellets were resuspended in
buffer C (20 mM Tris, 1 M NaCl, pH 7.5) and lysed by sonication on ice.
The crude lysate was then centrifuged at 14,000 rpm for 30 min at 4 °C.
The supernatant was applied to a Ni-NTA column (QIAGEN), followed by
size exclusion chromatography using a Superdex 200 (GE Healthcare)
column. The purified hsRBFA protein was diluted in buffer D (20 mM
Tris, 200 mM NaCl, pH 7.5). The mutant hsRBFA^△226–260 was generated
using a MutanBEST kit (TaKaRa) and confirmed by DNA sequencing. The
mutant protein was purified using the protocol described above.
For crystallization and ITC assays, DNA encoding human hsRBFA (residues
226–276 or 226–260) was amplified and incorporated into a modified
pGEX-4T-1 (Novagen) plasmid, in which the thrombin protease site was
substituted for a tobacco etch virus (TEV) cleavage site. The proteins
were expressed in E. coli BL21-Gold (DE3) cells (Novagen) cultured in
LB medium at 37 °C till OD600 reaches ~1.0, then shifted to 16 °C and
induced with 0.2 mM IPTG overnight. Bacterial pellets were resuspended
in buffer C and lysed by sonication on ice. The fusion proteins were
purified on glutathione-Sepharose beads (GE Healthcare) and eluted with
buffer C containing 30 mM reduced L-glutathione. The GST tag was
cleaved by TEV protease overnight at 16 °C, the protein was then
purified by a Superdex 75 column (GE Healthcare). Finally, proteins
were concentrated to ~15 mg/mL in buffer D and stored at –80 °C. All
mutants were generated using a MutanBEST kit (TaKaRa) and confirmed by
DNA sequencing. The mutant proteins were purified using the protocol
described above.
RNA preparation
For crystallization, RNA oligomers were purchased from Accurate
Biotechnology (Hunan) Co.,Ltd, ChangSha, China and dissolved in diethyl
pyrocarbonate (DEPC)-treated water to a final concentration of 1 mM.
Prior to use, the RNA substrates were heat-denatured at 95 °C for 5 min
and annealed on ice for 5 min. All RNA oligomers used in this study
were listed in Supplementary Table [182]S1.
Pull-down assay
For the pull-down assay of the MBP-tagged or GST-tagged hsRBFA against
the METTL15 protein, MBP-tagged hsRBFA (residues 42–343, 42–201,
226–276) or GST-tagged hsRBFA (residues 226–260) were bound to MBP
beads or glutathione-Sepharose beads (GE Healthcare), respectively, and
then the beads were incubated with METTL15 protein in buffer B for 2 h
at 4 °C. After washing six times with buffer B containing 0.1%
TritonX-100, the results were analyzed on Coomassie stained
SDS-polyacrylamide gels.
Protein crystallization, data collection and structure determination
Apo METTL15^70–407 was concentrated to ~8 mg/mL in buffer B and
crystallized in 10% PEG monomethyl ether 2000, 0.2 M ammonium sulfate,
0.1 M sodium acetate (pH 5.5) at 16 °C by vapor diffusion in sitting
drops. METTL15^70–07 and S-adenosyl methionine (SAM), were mixed at a
1:5 molar ratio at a final concentration of ~8 mg/mL. Crystals of the
METTL15^70–407–SAM complex were grown at 16 °C using the hanging drop
vapor diffusion method by mixing 1 μL of mix with 1 μL reservoir buffer
(42% PEG 600, 0.2 M imidazole malate, pH 5.5). METTL15^70–407,
hsRBFA^226–260 and SAM were mixed at a 1:2:5 molar ratio, and incubated
overnight at 4 °C. The ternary complex was crystallized in 17% w/v PEG
4000, 0.05 M Potassium chloride, 0.1 M Lithium chloride, 0.012 M
Spermine tetrahydrochloride, 0.05 M MES pH 6.5 at 16 °C by vapor
diffusion in sitting drops. METTL15^70–407, hsRBFA^226–260, RNA
substrate and S-adenosyl methionine analog (sinefungin, SFG) were mixed
at a 1:2:1.2:5 molar ratio, and incubated overnight at 4 °C. The
quaternary complex was crystallized in 20% PEG 4000, 0.2 M ADA (pH 6.4)
at 16 °C by vapor diffusion in sitting drops.
Crystals were soaked in mother liquor supplemented with 20% glycerol
before being flash-frozen in liquid nitrogen. X-ray diffraction data
for the crystals were collected on beamline 19U of the Shanghai
Synchrotron Radiation Facility (SSRF). The data were processed with
HKL2000 and programs in the CCP4 suite. The apo structure of the
METTL15^70–407 was solved by molecular replacement with the program
MOLREP^[183]48, using AlphaFold model of METTL15 as the search model
(AF-[184]A6NJ78-F1). The apo METTL15^70–407 structure was then
recruited as the search model in the molecular replacement using
program MOLREP for the structures of METTL15^70–407–SAM complex,
METTL15^70–407–hsRBFA^226–260–SAM ternary complex and
METTL15^70–407–hsRBFA^226–260–RNA–SFG quaternary complex. The hsRBFA
peptide was then modeled in COOT^[185]49. All initial models were
refined using the maximum likelihood method implemented in
REFMAC5^[186]50 as part of CCP4i program suite and rebuilt
interactively using the program COOT. Final refinement strategies
included XYZ coordinates, individual B-factors, occupancies, and
automated correction of N/Q/H errors using PHENIX^[187]51.
Crystallographic parameters were listed in Table [188]1. All images of
the structures were prepared using PyMOL ([189]http://www.pymol.org/).
CD measurements
Far-UV CD spectra of METTL15 and its mutants were determined using an
Applied Photophysics Chirascan spectrometer at 298 K. The spectra were
recorded at wavelengths ranging from 195 to 260 nm using a 0.05 cm path
length cell. The protein samples were diluted to 0.1 mg/mL with the
buffer B. A buffer-only reference was subtracted from each curve. All
samples were tested in triplicate.
ITC assays
ITC assays were performed on a Microcal PEAQ-ITC instrument (Malvern)
at 20 °C. The concentrations of proteins were determined
spectrophotometrically. HsRBFA and METTL15 proteins were dialyzed
against buffer B and adjusted to 600 μM and 40 μM, respectively. SAM
and METTL15 proteins were dialyzed against buffer B and adjusted to
500 μM and 50 μM, respectively. Thermodynamic data were analyzed with a
single binding site model using MicroCal PEAQ-ITC Analysis Software
provided by the manufacturer. The thermodynamic parameters of the ITC
experiments were listed in Table [190]2.
FP assays
Different lyophilized 5’-FAM (carboxyfluorescein)-labeled RNA oligomers
were purchased from Accurate Biotechnology (Hunan) Co.,Ltd, ChangSha,
China, dissolved in DEPC-treated water to a final concentration of
100 μM and stored at –80 °C. The stock (100 μM) was diluted to 80 nM in
dilution buffer B. All RNA oligomers used in this study were listed in
Supplementary Table [191]S1. Equilibrium dissociation constants of
different RNAs with METTL15 and its mutants were determined by
measuring FP, as previously described. METTL15 and its mutants were
first diluted to 20 times the highest concentration used in the binding
system, and then successively diluted 2-fold until the lowest desired
concentration was reached. Before the assay, 100 μL of 80 nM
fluorescence-labeled RNA was mixed with 100 μL of protein stocks from
the diluted series and incubated for 15 min. Samples were then excited
at 485 nm, and FP was detected at 525 nm using a SpectraMax M5
(Molecular Devices) plate reader at 20 °C. All FP data were well fitted
to a 1:1 binding model and were expressed as follows:
[MATH:
FP=FPini+max
2nR×
Kd+P+nR−−4nPR+Kd+P+nR2
:MATH]
where FP is the observed total polarization, FP[ini] is the initial FP
of RNA without any protein, P is the protein concentration, R is the
concentration of labeled RNA, n is the binding stoichiometry (protein:
RNA ratio), and K[d] is the equilibrium dissociation constant. Standard
errors were obtained by fitting the data to the above equation.
Cell culture
HEK293T cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM,
Gibco) supplemented with 10% (v/v) fetal bovine serum and 1%
penicillin/streptomycin. Cells were grown at 37 °C in a humidified
atmosphere containing 5% CO[2].
Plasmids and stable cell lines
All short hairpin RNAs (shRNAs) against METTL15 were inserted into
PLKO.1 vectors. The modified PLKO.1 plasmid was packaged into
lentivirus in HEK293T cells by PSPAX2 and PMD.2 G plasmids. HEK293T
cells were infected with lentivirus followed by puromycin (1 μg/mL)
selection to establish stable cell lines.
Western blotting assay
Cells lysates were prepared in radioimmunoprecipitation assay buffer
(50 mM Tris–HCl, pH 8.0, containing 150 mM NaCl, 5 mM EDTA, 0.1% SDS,
and 1% NP-40) supplemented with protease inhibitor cocktails (Roche).
Equal amounts of total cell lysate were separated by SDS-PAGE.
To detect histone lactylation levels, the extracted nuclear fractions
from cells were lysed with nuclear buffer (50 mM Tris-HCl, pH 7.5,
500 mM NaCl, 1 mM EDTA, 0.2% NP-40, 10 mM β-mercaptoethanol, 10%
glycerol) supplemented with protease inhibitor cocktail, and the
nuclear lysates were sonicated using an Ultrasonic Cell Disruptor
(Scientz). Protein lysate supernatant was collected after
centrifugation at 15,000× g for 10 min at 4 °C, and was quantified with
a Bradford assay kit (Sangon Biotech, C000164-0200). Equal amounts of
protein were subjected to 5%–12% SDS-PAGE.
Primary antibodies against the following proteins were used in western
blotting: METTL15 (Abcam, Cat# ab307819), METTL17 (Thermo, Cat#
pA5-26973), hsRBFA (Thermo, Cat# pA5-59587), MT-CO1 (Abcam, Cat#
ab203912), MT-CO2 (Thermo, Cat# A-6404), MT-CYB (Abcam, Cat# ab219823),
MT-ND3 (Abcam, Cat# ab192306), MT-ATP8 (Proteintech, Cat# 26723-1-AP),
H4K12-lac (PTM, Cat# 1411RM), H3K9-lac (PTM, Cat# 1419RM), H3K18-lac
(PTM, Cat# 1406RM), H3K56-lac (PTM, Cat# 1421RM), H4K5-lac (PTM, Cat#
1407RM), H4K8-lac (PTM, Cat# 1415RM) and H4K16-lac (PTM, Cat# 1417RM).
β-Tubulin (Abcam, Cat# T0023) served as a loading control. Horseradish
peroxidase-conjugated anti-rabbit or anti-mouse (Bio-Rad) secondary
antibodies were used to detect primary antibody binding (Respective
primary antibodies), and the signal was detected using LAS4000mini (GE
bio. Inc.).
Real-time fluorescence quantitative PCR (qRT-PCR)
TRIzol (Invitrogen, Cat# 15596018) was used to extract total RNA from
HEK293T cells. The isolated RNA was reverse transcribed into cDNA for
qRT-PCR. SYBR Green I fluorescent dye and a Roche LC96P temperature
gradient fluorescence quantitative PCR instrument were used, with GAPDH
as an internal reference. All the primer sequences used in this study
were listed in the Supplementary Table [192]S2.
Mitochondrial respiratory and glycolysis activity assay
Mitochondrial respiratory and glycolytic activities in HEK293T cells
were measured by using the Seahorse Extracellular Flux Analyzer XFp
(Agilent Technologies) with the XF Cell Mito Stress Test Kit (Agilent
Technologies). Transfected cells (1.5 × 10^4) were counted using
ADAM-MC2 (NanoEntek) and plated into V3-PS 96-well plates the day
before performing the assay. Standard mitochondria stress tests were
performed by first measuring basal values followed by measurements
after sequential addition of 1 µM oligomycin, 1.25 µM FCCP, and 0.5 µM
rotenone/antimycin A. The glycolysis stress test was performed by first
measuring basal values followed by measurements after sequential
addition of 10 mM glucose, 1 µM oligomycin, and 50 mM 2-deoxy-glucose.
ROS and membrane potential detection
MitoSOX (Yeasen, Cat# 40778ES50) was used for the determination of
mitochondrial ROS. TMRE (Beyotime, Cat#C2001S) was used for the
determination of membrane potential. The corresponding channels were
detected by flow cytometry.
Untargeted liquid chromatography with tandem mass spectrometry analysis
For each sample, 7 × 10^6 HEK293T cells were added into 2 mL thickened
centrifuge tubes, simultaneously, 2 magnetic beads and 10 μL of the
prepared internal standard 1 were added to each sample at the same
time. Then 800 µL of precooled extraction reagent (methanol:
acetonitrile: water as 2:2:1 in volume) was added and ground
subsequently at 50 Hz for 5 min. Following incubation at −20 °C for
2 h, samples were centrifuged at 25,000× g for 15 min. The supernatant
was transferred into split-new EP tubes for vacuum freeze-drying. The
metabolites were resuspended in 120 µL of 50% methanol and centrifuged
for 15 min at 25,000× g, and the supernatants were transferred to
autosampler vials for analysis. 10 μL of each sample was taken and
mixed into QC samples to evaluate the reproducibility of the analysis.
In this experiment, Waters 2777c UPLC (Waters, USA) in line with Q
exactive HF high resolution mass spectrometer (Thermo Fisher
Scientific, USA) was used for the separation and detection of
metabolites. The mass spectrometry data was imported into Compound
Discoverer 3.0 (ThermoFisher Scientific) for data processing.
Metabolite identification was conducted against the BGI metabolome
database (Shenzhen, China), mzCloud (ThermoFisher Scientific),
ChemSpider, the Kyoto Encyclopedia of Genes and Genomes (KEGG) and the
Human Metabolome Database (HMDB). The metaX software was used for
further statistical analysis^[193]52.
Supplementary information
[194]Supplementary Information^ (1.7MB, pdf)
[195]Supplementary Tables S3^ (1.6MB, xlsx)
Acknowledgements