Sirtuin 6 is a histone delactylase

Abstract

Histone lactylation (Kla) is a post-translational modification (PTM) that is derived from metabolic lactate. Histone Kla has been extensively studied in the field of inflammation resolution and macrophage polarization but has also been implicated in diverse cellular processes including differentiation, various wound repair phenotypes, and oncogenesis in several cancer models. While mechanistic connections between histone Kla and transcriptional changes have been studied in very limited contexts, general mechanistic details describing how regulation of gene expression by histone Kla occurs are scarce. It is hypothesized that histone Kla may be installed either through nonenzymatic means or by acetyltransferases like p300, and it is known that Class I HDACs and Sirtuins 1 to 3 can remove histone Kla. Here, we identified histone delactylase activity of the deacylase enzyme Sirtuin 6 (Sirt6), a member of the Class III HDAC family known to have roles in regulating metabolic homeostasis. We characterized the ability of Sirt6 to delactylate histones in vitro and in a mammalian cell culture model. We identified H3K9 and H3K18, canonical histone sites of Sirt6-catalyzed deacetylase activity, as sites of its delactylase activity. We also demonstrated that Sirt6 and the Class I HDACs exhibit some degree of non-overlapping delactylase activity, suggesting that they represent different cellular axes of regulating gene expression via controlling levels of histone Kla.

Keywords: epigenetics; histone deacetylase; histone lactylation; histone modifications; lactic acid; sirtuin.

Figure 1

Sirt6 exhibits histone delactylase activity in vitro.A, Western blot of histone Kla levels on nucleosomes isolated from HEK-293T cells that were incubated with recombinant Sirt6 and/or NAD⁺. B, quantitation of the blot from panel A. Densitometry data were corrected based on the total protein stain, then normalized to the “no NAD⁺/no Sirt6” control condition. Groups were compared using a one-way ANOVA followed by Tukey’s post hoc test. n = 3 technical replicates, error plotted as S.D. p (−/− v. +/+) = 0.006. C, Western blot analysis of time courses of enzymatic deacylation reactions using 50 nM substrate nucleosome, 100 nM Sirt6, and 1 mM NAD⁺. D, quantitation of Western blot data for the H3K9Ac substrate. Densitometry data were corrected based on the H3 loading control, then normalized to the t = 0 time point. E, quantitation of Western blot data for the H3K9La substrate. Data was processed as in panel D. F, Michaelis–Menten plot showing V_I values plotted as a function of substrate concentration for each PTM substrate. V_I was calculated for each substrate concentration by fitting the first three time points using standard linear regression. For panels D, E, and F, n = 3 technical replicates, error plotted as S.D.

Figure 2

Sirt6 depletion leads to histone Kla accumulation in cells.A, Western blot using a pan-lactyllysine antibody to analyze Kla levels on acid-extracted histones from wild type (“WT”), Sirt6 knockout (“S6KO”), and Sirt7 knockout (“S7KO”) U2OS cells in the presence of a titration of sodium L-lactate. Total protein was measured using a fluorescent total protein stain. B, as in panel A but using a pan-acetyllysine antibody to analyze levels of histone Kac. C, Western blots from panel A quantified by normalizing Kla signal to total protein then represented as a fold-change from the untreated condition. n = 3 biological replicates, error plotted as S.D. Slopes of linear regressions were compared using Welch’s t test. p (slope, S6KO v. WT) < 0.0001. p (slope, S7KO v. WT) = 0.39. D, Western blots from panel B quantified by normalizing Kac signal to total protein then represented as a fold-change from the untreated condition. For WT and S6KO, n = 3 biological replicates. For S7KO, n = 2 biological replicates. Error plotted as S.D. E, baseline histone lactylation in untreated cells was measured using a pan-Kla antibody as in panel A and normalized to the loading control. Note: these same data are plotted again in Figure 4C. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test. Error plotted as S.D. F, Western blot using a pan-lactyllysine antibody (PTM BIO) to analyze Kla levels on acid-extracted histones from WT and S6KO U2OS cells in the presence of a titration of rotenone, a mitochondrial Complex I inhibitor. Total protein was measured using a fluorescent total protein stain (LI-COR). G, antibody signal was corrected based on the loading control signal, then normalized as a fold-change from the untreated condition. n = 3 biological replicates, error plotted as S.D. Slopes of linear regressions were compared using Welch’s t test, p = 0.0005. H, as in panel F but using the pan-acetyllysine antibody to measure levels of histone Kac. (I) Quantitation of data from panel H. Data were processed as in panel G. n = 3 biological replicates, error plotted as S.D.

Figure 3

Sites of Sirt6-catalyzed histone delactylation in cells.A, levels of various histone lactyl PTMs on acid-extracted histones from WT or S6KO U2OS cells were measured using site-specific antibodies (CST and PTM Bio). Total protein in each sample was measured using a fluorescent total protein stain (LI-COR). Samples from S6KO and WT cells were analyzed on the same membrane and are spliced here only to make direct comparisons more evident. The complete blots can be found in the Supporting Information. B, quantitation of data from panel A. Data were corrected based on the Revert 700 total protein stain, then presented as a fold-change from the untreated control condition. Statistical analysis was performed using Student’s 2-tailed t test and corrected for multiple hypothesis testing using the Holm-Šídák correction. n = 3 biological replicates, error plotted as S.D. H3K9La p = 0.013, H3K18La p = 0.019. Note: The same data in panel B is also plotted in Fig. S4.

Figure 4

Cellular histone Kla is reduced by overexpression of Sirt6.A, Western blot measuring histone Kla on acid-extracted histones in WT and S6KO U2OS cell lines not treated with supplemental L-lactate with and without overexpression of human Sirt6 under the control of a CMV promoter. WT histone H3 (CST) is used as a loading control. B, Western blot measuring Sirt6 in the soluble protein fraction from cell lines from panel A. β-Actin is used as a loading control. C, Kla signal from panel A was quantified and normalized to the loading control. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test. n = 3 biological replicates, error plotted as S.D. D, histone Kac was measured using the same method and controls as in panel A. E, Kac data from panel D was quantified as in panel C. n = 3 biological replicates, error plotted as S.D. F, as in panel A, but the cells were treated with 25 mM sodium L-lactate for 24 h prior to histone extraction. G, As in panel B, but the cells were treated with 25 mM sodium L-lactate for 24 h prior to cell lysis. H, Kla signal from panel F was quantified and normalized to the loading control. I, inverse correlation of data from panels F and G. Sirt6 signal was normalized by dividing by the β-actin loading control. Kla signal was normalized as in panel H. The data were analyzed using a standard linear regression, R² = 0.84. For all panels, p-values < 0.05 have been printed in the figure, p-values > 0.05 marked as ns.

Figure 5

Sirt6-and HDAC-catalyzed histone delactylation are additive.A, Western blot measuring histone Kla in WT and S6KO U2OS cells treated with sodium L-lactate and panobinostat, as indicated. A total protein stain (LI-COR) was used as a loading control. B, quantitation of data from panel A. Histone Kla signal was quantified using densitometry and corrected based on the total protein stain. Data are presented as a fold change from the untreated (-lactate, -panobinostat) condition. Statistical analysis was performed using Student’s 2-tailed t test and corrected for multiple comparisons using the Holm-Šídák correction. p₃ and p₄ were calculated using Student’s 2-tailed t test. p₁ = 0.03, p₂ = 0.02, p₃ = 0.005, p₄ = 0.002. n = 3 biological replicates, error plotted as S.D. C, histone Kac was measured using the same methods and controls as in panel A. D, quantitation of data from panel C. Data was processed as in panel B. p₁ = 0.001, p₂ = 0.002. n = 3 biological replicates, error plotted as S.D.