To identify the ZFs that play roles in NFκB signaling, the full list of reviewed H. sapiens proteins in Uniprot (n = 20,428 total proteins) was filtered based upon the presence of at least one annotated ZF domain (n = 1,814 total proteins with ZF annotation). This list of ZFs was then analyzed using DAVID software (NIH), a web-based server for functional enrichment analysis of gene lists, using H. sapiens as the background gene list to match the Uniprot accessions (Sherman et al., 2022). With the gene list in DAVID, pathway analysis by the Kyoto Encyclopedia of Genes and Genomes (KEGG) knowledge database was selected (Kanehisa and Goto, 2000). KEGG analysis was employed because it yielded the highest number of gene hits (17) for ZFs in the NF-kappa B signaling pathway (KEGG designation hsa04064; Table 1; Supplementary Table S1).

Of the 17 associated gene hits, 13 are pro-inflammatory in their NFκB function, indicating that the KEGG pathway analysis is limited in its included proteins to early signaling/pathway activation. There are additional NFκB-responsive ZFs that were not identified in the KEGG analysis which are included in Table 1 and are discussed in this review. These proteins include Tristetraprolin (TTP), MCP-1-induced protein-1 (MCPIP1), and Roquin, which negatively regulate the pro-inflammatory cytokines and chemokines produced by NFκB activation (Fu and Blackshear, 2017; Makita et al., 2021; Xu et al., 2012). Ubiquitin-editing ZFs are also included, such as the anti-inflammatory mediators optineurin (OPTN) and OTU domain-containing protein 7A (OTU7A), as well as the pro-inflammatory ZFs SHARPIN, HOIL-1L, and HOIP. The latter three form the linear ubiquitination chain assembly complex (LUBAC) (Fennell et al., 2018; Tokunaga et al., 2011) which catalyzes many forms of protein ubiquitination, enabling protein-protein interactions (PPIs) required for NFκB activation. M1-linked linear polyubiquitination (LUBAC) and K63-linked ubiquitination (LUBAC and others) are forms of ubiquitination that are unique to canonical NFκB activation (Sun, 2017; Tao et al., 2022). Eight other ZFs in the list from Table 1 have ubiquitin E3 ligase activity, demonstrating the essentiality of this PTM, and ZFs, in NFκB signaling.

All of the ZF gene hits from the DAVID analysis, along with the additional ZFs described above, are included in this holistic review that focuses on the importance of ZFs in NFκB signaling and related pathways (i.e., apoptosis, cancer, etc.).

The tumor necrosis factor receptor (TNFR)-associated factors (TRAFs) are a family of ZFs that act as early transducers of NFκB signaling. All members contain multiple CCCC ZF domains and one RING finger domain which connects various stimulated receptors (for example, TNFRs and toll-like receptors, TLRs) with their respective adaptor complexes (TRADD and MyD88, respectively) (Shi and Sun, 2018). In TLR initiation, activated MyD88 forms a complex with and activates IRAK1/2, which is recognized by TRAF6 (Figure 5). The RING finger domains of TRAF6 function as E3 ligases, in tandem with the E2 complex Uev1A:Ubc13, to catalyze the K63-linked ubiquitination of TRAF6 and other substrates (Xie, 2013). K63-linked TRAF6 is recognized by the ZF complex of TGF-beta activated kinase 1 (TAK1) through its subunits of TAK1-binding proteins 2 and 3 (TAB2/3), both containing a single RanBP2 (CCCC-type) ZF which recognizes K63-linked ubiquitin (Park, 2018). This complex is then recognized by NFκB-essential modulator (NEMO) through NEMO’s lone CCHC ZF domain on the N-terminus (Maubach et al., 2017). Interaction of NEMO with the ubiquitinated upstream elements activates the IκBα kinase complex (IKK), which leads to phosphorylation and degradation of IκBα by IKK, and nuclear translocation of the p50/p65 NFκB complex.

For TNF-mediated activation of NFκB signaling, TNFR forms a signaling complex with TNF receptor-associated death domain (TRADD), TRAF2 or TRAF5, receptor-interacting protein kinase (RIP1), and the cellular inhibitors of apoptosis (cIAP1 and cIAP2) (Shi and Sun, 2018) (Figure 5). The IAP family are ZFs with E3 ubiquitin ligase activity which ubiquitinate RIP1 and recruit LUBAC (Gyrd-Hansen and Meier, 2010). LUBAC is comprised of 3 ZFs: SHANK-associated RH domain-interacting protein (SHARPIN), heme-oxidized iron regulatory protein ubiquitin ligase (HOIL-1), and HOIL-1 interacting protein (HOIP) (Fuseya and Iwai, 2021; Kasirer-Friede et al., 2019). Both HOIL-1 and HOIP have an arrangement of the following ZF domains: RING1, In-Between-RING (IBR), and RING2; these domains confer the E3 ubiquitin ligase functionality. LUBAC’s subunits also have domains that promote and stabilize their trimeric association, such as the ubiquitin-associating domain of HOIP and the ubiquitin-like domains of HOIL-1 and SHARPIN. LUBAC can uniquely conjugate linear ubiquitin to substrates via the Met1 residue of free ubiquitin; this serves as an additional scaffold for NEMO and others to recognize which is distinct from the branched chains of K63-ubiquitination (Fuseya and Iwai, 2021; Tokunaga et al., 2011).

In both TNFR-mediated or TLR-mediated activation, ubiquitin serves as both a scaffold to relay the signal from receptor to effector (NFκB complex) and to catalyze the degradation of the NFκB inhibitor (IκBα). IKK-mediated degradation of IκBα, and subsequent liberation of p50 and p65 (canonical NFκB subunits) for transcriptional activity, occurs within minutes, leading to a rapid transcriptional activation of NFκB-associated genes (Beg and Baldwin, 1993; Chen et al., 1995). Elevated levels of cytokine mRNAs (e.g., TNFα, interleukins, etc.) are efficiently translated into proteins, acting as positive feedback of this pathway, as RNA-regulatory ZFs, such as tristetraprolin (TTP), are not yet fully responsive (Rappl et al., 2021). Expressed cytokines and chemokines can also participate in crosstalk with adjacent cells, tissues, etc. This pro-inflammatotory signaling is required to activate the immune system and/or combat invading pathogens.

It is important that levels of pro-inflammatory stimuli are controlled by repressors of the pathway, which are also induced by NFκB, to dampen signaling. As shown in Figure 6, one subset of anti-inflammatory mediators are the ubiquitin-editing proteins, including TNFα-induced protein 3 (TNFAIP-3, A20). A20 is a CCCC-type ZF which is rapidly upregulated to act as both a deubiquitinase (DUB) of M1 and K63-ubiquitin linkages and an E3 ligase of K48-ubiquitin linkages (Mooney and Sahingur, 2021; Priem et al., 2020). A20 recognizes many forms of ubiquitinated proteins and disrupts multiple pathway transducers by inhibiting their interactions through ubiquitin. Additionally, it can conjugate K48-ubiquitin chains to early pathway inducers, degrading the substrate, further disrupting the pro-inflammatory signal (Verstrepen et al., 2010). A20 has seven CCCC-type ZF domains, where ZF 4 confers the K48-linked ubiquitin ligase activity and ZF 7 is crucial for polyubiquitin recognition and deubiquitination (Priem et al., 2020). A20 quickly functions to disassemble the ubiquitin frameworks that are the structural basis for continued signaling.

Additional ubiquitin-associated proteins are the ZFs Optineurin (OPTN) and OTUD7, as well as the non-ZFs Cylindromatosis (CYLD), and OTULIN (Fennell et al., 2018). OPTN’s C-terminal UBAN and CCHC ZF domains are required for interaction with ubiquitin-like structures, including linear polyubiquitin and CYLD (Guo et al., 2020). OPTN and NEMO have high sequence homology and signal the NFκB pathway antagonistically through their competition for both linear and K63-linked polyubiquitin (Qiu et al., 2022; Slowicka and van Loo, 2018). CYLD and OTULIN negatively regulate NFκB activity through recognition and de-ubiquitination of polyubiquitinated substrates, including the TRAFs and HOIP (Fennell et al., 2018; Shi and Sun, 2018; Xie, 2013). Collectively, these ubiquitin-editing enzymes regulate ubiquitination antagonistically to NEMO, the LUBAC complex, and the TRAF family.

While the DUBs act to prevent further NFκB translocation and activation, RNA-binding ZFs are induced to respond to the NFκB-dependent mRNAs (Behrens and Heissmeyer, 2022; Maeda and Akira, 2017) (Figure 7). TTP and others regulate the positive feedback of cytokine mRNAs by limiting their translation into mature proteins which are capable of further stimulating NFκB-related receptors (ex: TNFα and various interleukins). These RNA-binding ZFs recognize and bind specific ribonucleotide sequences, destabilizing the mRNAs for translation, and ultimately leading to their degradation (Hudson et al., 2004; Kedar et al., 2012; Lai et al., 2000). The complete resolution of these cytokines to basal levels can take upwards of 4 h, depending on the initial concentration and identity of the stimuli (e.g., LPS, TNFα, etc.) (Mulvey et al., 2021).

TTP, MCPIP, and Roquin are CCCH-type ZFs which negatively regulate excessive pro-inflammatory mRNAs and are essential for proper immune function (Maeda and Akira, 2017; Makita et al., 2021; Rappl et al., 2021). TTP contains two CCCH ZF domains which are required for specific recognition of AU-rich sequences in the 3′-UTR of mRNAs (Hudson et al., 2004). TTP’s ZFs are also necessary for localization to the nucleus and mRNA-processing bodies (Lai et al., 2000). Degradation of cytokine mRNAs requires an evolutionarily conserved C-terminal domain which recruits the CCR4-NOT deadenylase complex to TTP-bound mRNA (Fabian et al., 2013). In addition to mRNA destabilization, TTP disrupts NFκB nuclear translocation and participates in alternative splicing in innate immunity (Gu et al., 2013; Schichl et al., 2009; Tu et al., 2020). In canonical NFκB signaling, TTP is one of the main anti-inflammatory mediators due to its simultaneous regulation of thousands of ARE-containing mRNAs (Brooks and Blackshear, 2013; Rappl et al., 2021; Tiedje et al., 2016).

MCPIP1, has an N-terminal ubiquitin-associated domain, a single CCCH ZF domain for mRNA recognition, and an adjacent PIN-like RNase domain for mRNA degradation (Fu and Blackshear, 2017). MCPIP can regulate the NFκB pathway by deubiquitination of TRAF complexes, via its ubiquitin-associated domain, and by degrading cytokine mRNAs, with its inherent endonuclease activity from the adjacent ZF-RNase domains (Xu et al., 2012). Roquin contains two different ZFs: one CCCH domain adjacent to a ROQ domain and one RING ZF domain. The CCCH ZF and ROQ domains are required for mRNA target recognition and recruitment of the decapping complex of mRNA-decapping protein 4 (EDC4) and RCK. The RING ZF is proposed to be important for E3-ubiquitin ligase activity and stress granule association (Zhang et al., 2015). Roquin recognizes conserved stem-loop elements in cytokine mRNAs, such as TNFα, which are spatially distinct from the ARE’s regulated by TTP and represent an overlap in regulation (Maeda and Akira, 2017). MCPIP1 and Roquin also interact in a way that affects the autoregulation of their own mRNAs (Behrens et al., 2021). In addition to the autoregulation of mcpip1 mRNA by MCPIP1, ttp mRNA has multiple ARE elements in its 3′-UTR, which serve as regulatory sites for TTP-mediated degradation. Collectively, these CCCH ZFs are induced by NFκB activation and disrupt the positive feedback loop through degradation of cytokine mRNAs. At some point in the resolution process, activity of these RNA-binding ZFs decreases, and downregulation of their levels restores pathway homeostasis.

ZFs are signaled via PTMs of the ZF domain, or adjacent motifs in the full-length sequence, which affects overall protein function and cell signaling (Fennell et al., 2018; Vignane and Filipovic, 2023; Vilas et al., 2018). In the NFκB pathway, phosphorylation of Ser or Thr residues signals through several ZFs, fine-tuning their activity for a regulated inflammatory response. ZF kinases such as the TAB family (TAB, subunits of TAK1) and protein kinase C family (PKC), in addition to the Ca(II)-mobilizing ZF bruton tyrosine kinase (BTK), phosphorylate key substrate proteins to facilitate the early NFκB signal (Altman and Kong, 2014; Kramer et al., 2023; Newton, 2018). For example, the upstream kinases phosphorylate and activate the IKK complex, which then phosphorylates IκBα, leading to ubiquitination and degradation of IκBα, liberating NFκB to translocate to the nucleus (Beg et al., 1993; Chen et al., 1995). Phosphorylation of TTP by MAPK kinases (e.g., MK-2) at S60 and S186 inhibits TTP function via recruitment of the 14-3-3 adaptor complex, permitting the transient increase of cytokine mRNAs during the early inflammatory response (Sandler and Stoecklin, 2008; Tiedje et al., 2014). TTP is then dephosphorylated by PP2A, which directly competes with the 14-3-3 complex, releasing TTP for mRNA regulation.

Another common PTM communicated through ZFs in the NFκB pathway is ubiquitination, which occurs in at least three distinct forms, namely, K48, K63, and M1-linked ubiquitination (Maubach et al., 2017). K48-linked ubiquitination of proteins signals them for degradation via the proteasome. This PTM of IκBα is required to de-repress NFκB and allow for nuclear translocation (Beg and Baldwin, 1993; Beg et al., 1993; Chen et al., 1995). K48-linked ubiquitination is also used by A20, the IAPs, and the TRAFs to regulate early signaling complexes. In contrast, K63-linked ubiquitin and the linearly linked M1-ubiquitin stabilize their substrates and signal as structural scaffolding for downstream interactions of co-activators (i.e., NEMO and SHARPIN) or co-repressors (i.e., TRAF6 and A20) (Fuseya and Iwai, 2021; Kasirer-Friede et al., 2019). In total, the variety of ubiquitination is intertwined with ZF regulation of the NFκB signaling, particularly the RING ZF E3 ligase domains.

Much like phosphorylation and ubiquitination, cysteine-specific PTMs play a role in modifying protein activity during inflammatory signaling, particularly for cysteine-rich ZFs (Chantzoura et al., 2010; Kukulage et al., 2022; Oppong et al., 2023). Human proteins have the highest abundance of cysteines relative to other life forms, and ZF motifs are particularly enriched in their cysteine content (Wiedemann et al., 2020) (See Table 2; Supplementary Table S2). Cysteine residues can be oxidized to disulfides, and while Zn(II) binding to these cysteine residues decreases the susceptibility of cysteine oxidation, such redox changes can occur affecting ZF integrity and function during stress (Doka et al., 2020; Hartle et al., 2016; Kimura, 2015; Kumar and Banerjee, 2021). Adequate redox balance through cellular glutathione and other redox mediators is essential for mitigating oxidative damage to ZFs and other metalloproteins (Fukuto et al., 2020). One cysteine-specific PTM is persulfidation (Cys-SH to Cys-SSH), which modulates protein function and protects the cysteine thiols from oxidation during oxidative stress (Chantzoura et al., 2010; Luo et al., 2023).

Persulfidation is mediated through the gaseous signaling molecule, hydrogen sulfide (H₂S). Cystathionine-gamma-lyase (CSE), cystathionine-beta-synthase (CBS), and 3-mercaptopyruvate sulfurtransferase (3-MST) are the main H₂S synthase enzymes. These enzymes constitute the transsulfuration pathway governing the homeostasis of H₂S and related sulfur-containing small molecules (Kimura, 2015; Vignane and Filipovic, 2023). H₂S can directly form persulfides via reaction with oxidized cysteine residues (disulfides and sulfenic acids) as well with as the Cys-S-Zn sites of ZFs (Cuevasanta et al., 2015; Lange et al., 2019). Protein-persulfides can also be formed co-translationally with cysteinyl-tRNA synthetases (CARS) which functions cooperatively with the transsulfuration pathway, particularly during oxidative stress (Akaike et al., 2017; Sawa et al., 2020). Persulfidation impacts protein structure and can enhance or inhibit activity depending on the specific Cys residue that is modified. As such, the ‘signaling’ property of H₂S can be described as modulating regulators and their function. ZFs are directly and indirectly affected by protein persulfidation. Directly, protein stability and activity are altered in the emerging examples of ZF persulfidation (Saha et al., 2016; Vandiver et al., 2013). ZFs are also indirectly induced by persulfidation of other proteins, such as p65 and MEK1, which both contribute to downstream upregulation of PARP-1 and other NFκB-associated ZFs (Dai et al., 2019; Sen et al., 2012; Zhao et al., 2014a).

Our laboratory previously reported that the persulfidation of the tandem CCCH ZF domains of TTP leads to disruption of the protein’s structure and obviates TTP/RNA-binding (Lange et al., 2019). TTP-2D reacts with H₂S when Zn(II) is bound to the protein and O₂ is present. Cryo-electrospray-ionization Mass Spectrometry identified the persulfidated ZF. This method along with orthogonal techniques led to the identification of sulfur- and oxygen-based radicals formed during the persulfidation reaction. A mechanism whereby Zn(II) acts as a conduit for electron transfer between H₂S and O₂, activating O₂ to form superoxide and other reactive species, leading to eventual disulfide formation and Zn(II) ejection was proposed. Cell experiments showed that MEF cells harvested from both WT and ΔCSE mice have similar basal levels of TTP; CSE contributes the majority of protein persulfidation in cells (Filipovic et al., 2018; Paul and Snyder, 2015). Thus, absence of persulfidation did not greatly affect TTP protein stability/abundance, as has been shown in some other examples of ZF-persulfides (Kalous et al., 2021; Saha et al., 2016; Vandiver et al., 2013). TNFα mRNA levels were decreased in the CSE knockout cells, connecting in vitro findings that persulfidation of TTP restricts its RNA binding activity. This loss of structure and function due to TTP persulfidation appears to be a regulatory mechanism for its activity, with H₂S serving to “signal” mRNA processing.

We subsequently reported that persulfidation of ZFs is a common PTM. Analysis of persulfide specific proteomics data in multiple mammalian cell lines led to the identification of a large number of persulfidated ZFs. These include ZFs with different ligand sets and regulatory functions (e.g., transcription, translation, ubiquitination, etc.) (Li et al., 2024; Stoltzfus et al., 2024). The finding that ZFs with various ligand sets are persulfidated led us to investigate whether the chemistry that drives persulfidation is affected by ligand set. Using the TTP ZF peptide scaffold as the starting point, a series of mutants that offer one, two, three or four cysteine ligands per ZF domain were prepared and their reactivity with H₂S evaluated. In all cases, the TTP ZF peptides were persulfidated by H₂S, as long as Zn(II) was bound and O₂ was present. These findings suggest a common chemistry amongst ZFs for persulfidation.

In addition to these proteomics data, there are scattered reports of isolated ZF-persulfides, shown in Figure 8, the earliest of which is Parkin (Vandiver et al., 2013). Parkin contains multiple RING ZF domains which confer E3 ubiquitin ligase activity. This study demonstrated functional activation of Parkin by persulfidation. Parkin was found to be physiologically persulfidated in healthy tissues but depleted in Parkinson’s disease (PD) patients and mice models, indicating the loss of Parkin persulfidation and activity as a pathogenic mechanism for sporadic Parkinson’s disease. A slow-releasing H₂S-donor (GYY4137) improved Parkin-mediated ubiquitination and degradation of target protein, AIMP2, improving cell viability. As Parkin contains four ZF domains of various classes, this early work highlights the nuance of ZF persulfidation and the need for site-specific determination to consider the full protein sequence and the functional implications of persulfidation. The authors determined the specific residues of Parkin which are persulfidated using mammalian transfection, mutations of Parkin Cys residues, and mass spectrometry. Five Cys residues were determined to be persulfidated but modification of the catalytic C678, involved in ubiquitin transfer, was not observed, suggesting an allosteric effect on Parkin activity by persulfidation of the RING-type 0 (C182 and C212) and IBR-type (C377) ZF domains (Figure 8).

The classical ZF, SP1 (CCHH domain) has also been reported to be persulfidated. SPI acts as a transcriptional activator or repressor of GC-rich gene promoters (Beishline and Azizkhan-Clifford, 2015; Kaczynski et al., 2003; Vizcaino et al., 2015; Yin and Wang, 2021). SP1 was found to be upregulated during NFκB pathway stimulation by TNFα, leading to increased H₂S production and persulfidation through SP1-mediated transcriptional activation of CSE (Sen et al., 2012). More recently, Saha and coworkers showed that SP1 is itself persulfidated, dependent on functional CBS, leading to enhanced transcription of VEGF-1 by SP1 (Saha et al., 2016). The authors again used an MS/MS technique to show that persulfidation of SP1 at C68 and C755 are necessary for optimal transcriptional activity in the maintenance of endothelial cell function. Although the persulfidated residues are outside of the ZF domains of SP1, they play a role in stabilizing the protein, resulting in higher SP1 protein levels, leading to optimal binding of its transcriptional elements. As CBS converts homocysteine to H₂S in the transsulfuration pathway, knockdown of CBS led to hyperhomocysteinemia and a reduction in H₂S and GSH levels (Miles and Kraus, 2004). Furthermore, SP1 had decreased binding to the VEGFR-2 promoter and endothelial cells displayed a phenotype with compromised chemotaxis. All outcomes were ameliorated by treatment with H₂S, but not glutathione, demonstrating the importance of SP1 persulfidation mediated by proper CBS activity.

The MYND-type ZF (CCCC-CCHC) prolyl hydroxylase domain-containing protein 2 (PHD2) constitutively hydroxylates HIF-1α leading to ubiquitination and degradation (Dey et al., 2020; Maxwell et al., 1999). Hypoxic conditions lead to inhibition of PHD2 activity, stabilization of HIF-1α, complexation with HIF-1beta, and transcription of hypoxia-related survival genes. Dey and coworkers discovered that PHD2 is persulfidated at C21 and C33 residues of its MYND-type ZF domain (Figure 8), leading to enhanced hydroxylation of HIF-1α by PHD2 (Dey et al., 2020). Disruption of CBS in zebrafish lead to abnormal development, diminished persulfidation of PHD2, and stabilized HIF-1α. All phenotypes were rescued by H₂S supplementation. These data support a role for persulfidation in healthy endothelial cell development.

The sirtuin family of ZFs has also been shown to be persulfidated (Kalous et al., 2021). This ZF family features a singular, conserved CCCC-type ZF domain that provides a structural element to deacetylate histones and activate transcription of stress-response genes. The SIRT1 and SIRT3 proteins have been shown to undergo persulfidation during various types of cell stress. Persulfidation improves SIRT1 protein levels in the cell by increasing Zn-binding affinity and reducing ubiquitination, both of which lead to enhanced protein stability and deacetylase activity (Dong et al., 2023; Du et al., 2019; Li et al., 2020; Wu et al., 2023). Similar observations have been made for SIRT3, localized in the mitochondria (Liu F. et al., 2023; Liu et al., 2020; Xiong et al., 2023; Yuan et al., 2019). Liu et al. established that persulfidation is essential for optimal SIRT3 activity in two ways: first, by the direct stabilization and activation of SIRT3 by persulfidation, and second by the indirect upregulation of SIRT3 transcription by Nrf-2, which is also indirectly activated by persulfidation of Keap1 (Liu et al., 2020). This Nrf-2/Keap1/SIRT3 persulfidation axis is perturbed during conditions of oxidative stress and can be rescued by exogenous NaHS supplementation. For both SIRT1 and SIRT3, the cumulative evidence highlights a role for ZF persulfidation leading to enhanced deacetylation activity, which is an essential factor for increased antioxidant gene expression by the sirtuin ZF family.

In contrast to the ZF-persulfides discussed so far, which have enhanced protein activity, there are also examples of protein inactivation by persulfidation, such as TTP-2D and androgen receptor (AR). Zhao and coworkers focused on AR, a CCCC type ZF that functions as a transcriptional activator of androgen-responsive gene elements, in androgen-resistance and the development of prostate cancer (Zhao et al., 2014b). They found that CSE expression was diminished in human prostate cancer tissue and androgen-resistant cell lines, suggesting a possible deficiency in H₂S-associated signaling in this disease state. CSE overexpression repressed the expression of AR-target genes and remediated pathogenic phenotypes in mice models. They also found that mutation of the ZF domains of AR at C611 and C614 abrogated the rescue potential of CSE overexpression, indicating that persulfidation of these residues is required for regulation of AR via inhibition of AR dimerization and gene transcription.

The broader biological prevalence and significance of ZF persulfidation have not been fully appreciated and reviewed. To connect the signaling role of H₂S and persulfidation of Cys-rich ZFs, it is necessary to consider the local environment and location of specific residues found to be persulfidated. Of the six ZF-SSHs listed in Section 5.1 which were studied in cell (Parkin, SP1, PHD2, SIRT1, SIRT3, and AR), five ZFs had enhanced activity with persulfidation, as well as ameliorative effects in cells and tissue. What’s left to unravel are the connections between modified Cys residues and the fate of overall protein activity, although the commonality is activation by persulfidation. This may prove helpful in screening and characterizing potentially pathogenic mutations which lead to some loss-of-function due to diminished protein persulfidation.

One example is Parkin which was found to be persulfidated but to a diminished degree in Parkinson’s patients. Only three of five Cys mutants could be expressed and functionally assessed by iteratively mutating these residues from Cys to Ser. The three mutants (C59, C95, and C182) that were assessed are integral to the auto-regulatory inhibition of Parkin in its inactive, basal state. This suggests that persulfidation of these residues may contribute to the “opening” and activation of Parkin, as seen in the well-studied phosphorylation of Parkin by PINK1 (Barazzuol et al., 2020). Persulfidation of these residues likely does not directly enhance Parkin’s catalytic activity (as catalytic C678 was not modified), but rather it contributes to a greater pool of active Parkin and indirectly increases the ubiquitation of substrates.

As described in Section 5.1, SP1 persulfidation at C68 and C755 enhances its binding and transcriptional activation of VEGF (Saha et al., 2016). Separately, persulfidation of the second ZF domain at C664 is required for SP1-mediated inhibition of Krüppel-like factor 5 (KLF5) expression (Meng G. et al., 2016). These two cases illustrate that further characterization and partner/target binding studies are required for novel ZF-SSHs. Persulfidation of certain residues might enhance interactions with either a corepressing or a coactivating protein, depending on factors of the local environment such as surface accessibility. Further in-depth studies of ZF-SSHs can utilize the approaches from these examples and expand our basis of knowledge concerning the signals of persulfidation through ZFs.

A powerful tool in cell and animal models are persulfide-specific proteomic approaches which can determine enriched signaling pathways and new ZF-SSHs of interest. There are several persulfide-labeling approaches for proteomic identification (Li et al., 2024) and many can be used to corroborate protein-persulfides with non-reducing gel electrophoretic methods of labeled cell lysates. These labeling methods may be applied to other methods of visualization, such as microscopy and flow cytometry, to inform on cellular localization (Zivanovic et al., 2019).

As presented by the case studies mentioned here, experiments which identify residue-specific Cys persulfidation (i.e., mass spectrometry and/or protein mutation) are essential for ZF-SSHs. Furthermore, mapping these modifications to reported protein structures can aid in understanding the structural and functional consequences of ZF persulfidation. In cell, applying the techniques cited in Section 5.1 to newly identified ZF-SSHs will be helpful in uncovering the broader implications of ZF persulfidation and H₂S signaling. Identified ZF-SSHs can be tested for site-specific persulfidation, which requires mammalian cell transfection, overexpression, and Cys-SSH identification by MS techniques. The cell systems are then utilized to iteratively mutate identified Cys residues and determine the functional fidelity of mutants (ex: Parkin-C182S). Additionally, knockout mice models can broadly (or specifically) reduce persulfidation and provide information about which H₂S-generating enzymes contribute to the ZF-SSH of interest. By knocking out or inhibiting any of CSE, CBS, 3-MST, or CARS, direct connections between an enzyme and a ZF’s persulfidation can be elucidated. Similarly, biochemical studies on isolated ZF proteins and peptides, to further characterize persulfidation sites and determine the effects of persulfidation on protein activity will further broaden our understanding of ZF protein persulfidation and the role of H₂S.

While ZFs are now understood as ubiquitous regulators of transcription and translation, interest in gene editing precedes their discovery (Helene and Toulme, 1990; Kawasaki et al., 1996). ZFs have been studied for their potential in gene editing therapy due to their innate function of oligonucleotide binding. This specificity of binding is tunable for different targets (Bibikova et al., 2001; Kawasaki et al., 1996). ZF nucleases (ZFNs) are chimeric constructs containing multiple ZF domains for recognition of target genes, where each domain recognizes a triplet of nucleotide bases, fused with a nuclease domain (Porteus and Baltimore, 2003). The small recognition patterns provide ZFNs an advantage over other endonucleases, as the fingers can be modularly designed to increase sequence specificity (Hauschild-Quintern et al., 2013). ZFNs have been implemented in human and other mammalian cells, but translation to animal models and human clinical trials is still developing (Jabalameli et al., 2015).

The first human clinical trial using ZFNs was conducted in 2022 to treat mucopolysaccharidosis (n=12) and hemophilia B (n=1) (Harmatz et al., 2022). The authors found adequate safety and tolerance at all tested doses with evidence for successful gene editing and enhanced enzyme expression in liver tissues. However, the hemophilia B subject could not be assessed for gene-editing, and no long-term enzyme expression was observed in any patients. RNA-binding CCCH ZFs, such as MCPIP, are also in development as a potential ZFN design strategy for chronically elevated cytokines which may be disrupted by the stability and translational efficiency of their mRNA (Gaj et al., 2012; Garg et al., 2021; Liu et al., 2021; Paschon et al., 2019). Ultimately, the long-term efficacy of ZFN treatment, as seen in the clinical trial, needs improvement.

Glutathione and N-acetyl cysteine (NAC) have been popularized as antioxidant supplements in sports medicine and elsewhere, presumably acting by replenishing a depleted sulfur pool (Fernández-Lázaro et al., 2023). Dick and coworkers recently showed that the beneficial effects of NAC are due in part to endogenous H₂S production from the provided Cys (Ezerina et al., 2018). As a result of increased H₂S concentrations, persulfide species, including GSSH, were increased. Sulfane sulfur and persulfides have garnered increased interest over the last decade due to the ability of the former to store sulfur and generate various forms of protein- and small molecule-persulfides, which are associated with a multitude of positive biological outcomes in cell and animal studies (Dai et al., 2019; Doka et al., 2016; Donnarumma et al., 2017; Sun et al., 2020; Yu et al., 2023; Zhang et al., 2019; Zivanovic et al., 2019). The highly reactive nature of a persulfide makes it a better radical scavenger than other cellular thiols but also presents an obstacle to therapy as it may react before reaching its target (Filipovic et al., 2018; Fukuto et al., 2020). To overcome this, efforts have been made to create persulfide donor prodrugs which activate upon reaction with specific stimuli, such as superoxide, hydrogen peroxide, and photons (Bora et al., 2018; Chaudhuri et al., 2019; Wang et al., 2020). Other groups are dissecting the chemical reactivity of persulfides based on their local chemical environment, adding to both our fundamental understandings of protein persulfides and the tunable stability of persulfide-related therapies (Fosnacht et al., 2024). These therapeutic methods may restore ZF-SSH function and more broadly aid in biomarker discovery of diseases characterized by chronic inflammation.