Abstract
Lysine L-lactylation [K(L-la)] is a newly discovered histone mark stimulated under conditions of high glycolysis, such as the Warburg effect. K(L-la) is associated with functions that are different from the widely studied histone acetylation. While K(L-la) can be introduced by the acetyltransferase p300, histone delactylases enzymes remained unknown. Here, we report the systematic evaluation of zinc- and nicotinamide adenine dinucleotide–dependent histone deacetylases (HDACs) for their ability to cleave ε-N-L-lactyllysine marks. Our screens identified HDAC1–3 and SIRT1–3 as delactylases in vitro. HDAC1–3 show robust activity toward not only K(L-la) but also K(D-la) and diverse short-chain acyl modifications. We further confirmed the de-L-lactylase activity of HDACs 1 and 3 in cells. Together, these data suggest that histone lactylation is installed and removed by regulatory enzymes as opposed to spontaneous chemical reactivity. Our results therefore represent an important step toward full characterization of this pathway’s regulatory elements.
INTRODUCTION
Emerging lines of evidence suggest that metabolic end products and intermediates can have signaling functions in addition to their cognate roles. A metabolite can exert its function covalently through intrinsic chemical reactivity (1, 2) or enzyme-catalyzed reactions (3). Classic examples of the latter include acetyl–coenzyme A (CoA) and S-adenosylmethionine (SAM), which can be used by acetyltransferases for lysine acetylation and by methyltransferases for lysine methylation, respectively (3, 4). Other metabolites, such as nicotinamide adenine dinucleotide (NAD+) and α-ketoglutarate, can serve as cofactors and regulate the activities of corresponding deacetylases and demethylases (5). L-Lactate, traditionally known as a metabolic waste product, has recently been found to play important roles in metabolism. L-Lactate production regenerates the NAD+ consumed by glycolysis within cells. The shuttling of L-lactate between different organs and cells serves as a major circulating carbohydrate source that plays important roles in normal physiology and in cancer (6–8). L-Lactate is massively induced under hypoxia and during the Warburg effect (7, 9, 10), which are associated with many cellular processes and are closely linked to diverse diseases including neoplasia, sepsis, and autoimmune diseases (11). Nevertheless, the nonmetabolic functions of L-lactate in physiology and disease, especially during Warburg effect, remain largely unknown.We recently reported that L-lactate is a precursor that can label and stimulate histone lysine ε-N-L-lactylation [K(L-la)] (12). Data suggest that L-lactate is transformed into L-lactyl-CoA (13) and transferred onto histones by acetyltransferases such as p300 (Fig. 1A) (12). Therefore, like acetyl-CoA and histone lysine acetylation (Kac) (14), histone K(L-la) represents another example showing that acyl-CoA species can affect gene expression directly via histone posttranslational modification (PTM). In addition, histone K(L-la) has different kinetics from those of histone (Kac) during glycolysis (12). Histone K(L-la) is induced by hypoxia and the Warburg effect, serving as a feedback mechanism to promote homeostatic gene expression in the late phase of macrophage polarization (12). Our data therefore indicate that histone K(L-la) is a physiologically relevant histone mark and has unique biological functions.
In addition to L-lactate [with (S) configuration], its structural isomer D-lactate [with (R) configuration] is also found in cells, although at much lower concentration (11 to 70 nM concentration, compared to typical millimolar concentration of L-lactate, which can reach up to 40 mM in some cancer cells) (15, 16). D-Lactate is formed primarily from methylglyoxal (MGO) through the glyoxalase pathway (17, 18) and is overproduced in rare conditions, including certain cases of short bowel syndrome. Under these unusual circumstances, D-lactate can reach 3 mM concentration or higher in plasma (16, 19). S-D-(R)-Lactylglutathione, an intermediate in the glyoxalase pathway, can react nonenzymatically to install K(D-la) PTMs on glycolytic enzymes (Fig. 1A) (20). Nevertheless, five lines of evidence suggest that histones are L-lactylated rather than D-lactylated: (i) the huge difference in cellular concentration of the precursor metabolites, L-lactate versus D-lactate; (ii) the specificity of the antibodies used in our work toward K(L-la) (~100-fold or higher; Fig. 1B); (iii) the fact that histone lactylation can be stimulated and labeled by L-lactate (12); (iv) the specific genomic localization of histone K(L-la) that excludes the possibility of random, spontaneous chemical installation (12); and (v) the fact that K(D-la) is enriched solely in cytosolic proteins that are in close contact with S-D-lactylglutathione (20), which is not the case for histones. Despite the progress, key regulatory mechanisms of histone K(L-la) remain unknown, including the enzymes that remove this modification in the cell (Fig. 1A).Mammals express two families of lysine deacylases with 18 enzymes in total. Histone deacetylases (HDAC1–11, grouped into classes I, II, and IV) are Zn2+ dependent (21), while the sirtuins (SIRT1–7, class III HDACs) are dependent on NAD+ as cosubstrate (22). Among these, several isozymes exhibit preferential enzymatic activities against non-Kac acylations. For example, SIRT5 is an efficient lysine demalonylase, desuccinylase, and deglutarylase but not deacetylase (23–26), and HDAC8, HDAC11, SIRT2, and SIRT6 harbor substantial activities against long-chain acyl modifications (27–30). HDAC1–3, in addition to their efficient deacetylase activity, also remove ε-N-crotonyllysine (Kcr) and ε-N-D-β-hydroxybutyryllysine [K(D-bhb)] PTMs (31–34). Therefore, it is likely that enzymes from the lysine deacylase classes may be able to catalyze the removal of histone lactyl modifications.Here, we report a screening of all 18 HDACs for potential delactylase activities using fluorophore-coupled peptides, fluorophore-free histone peptides, and extracted histones as substrates. We show that class I HDAC1–3 are the most efficient lysine delactylases in vitro and that HDAC1 and HDAC3 have site-specific delactylase activity in cells. These findings support that histone K(L-la) modification is a regulatory and dynamic epigenetic mechanism. HDAC1–3 harbor activities in vitro toward diverse acyllysine groups including aliphatic and hydroxylated modifications of two to five carbons. In addition, even though we detect minor delactylase activity by SIRT1–3 in vitro, the pan-sirtuin inhibitor nicotinamide (NAM) does not affect the delactylase activity of a cell lysate or the histone K(L-la) levels in cells. Our data suggest that HDAC1–3 are the main delactylases in the cell….
