A biotin-HaloTag ligand enables efficient affinity capture of protein variants from live cells

Proteins serve as the primary driving force behind nearly all cellular processes, and their proper functioning is essential for the health and fitness of living organisms. Therefore, exploring protein function is fundamental to unveiling the molecular nature of vital cellular processes and their related changes underlying disease development. A diverse array of biochemical, cell biology, proteomic, and genetic methodologies has been devised to visualize proteins and study their function through the attachment of a functional tag to the protein of interest. Over the years, a wide range of functional tags has been developed to suit different types of experimental approaches. For instance, the polyhistidine tag (His-tag) is one of the most popular small tags for protein isolation and purification, while GFP and its derivatives are widely utilized in imaging studies (Giepmans et al., 2006; Terpe, 2003). More recently, small inducible tags, such as auxin-inducible degron, have emerged as prominent tools for achieving rapid protein degradation in functional studies (Bond et al., 2021; Nabet et al., 2018; Yesbolatova et al., 2020). Nevertheless, a comprehensive understanding of protein function typically requires the combination of multiple tailor-made tags, which is a major limitation of the traditional protein-tagging approach. The elegant solution for this bottleneck came with the discovery of self-labeling protein tags, exemplified by the HaloTag and SNAP-tag (Keppler et al., 2003; Los et al., 2008). These tags, fused to the protein of interest, are based on the covalent interaction of an enzyme with a small synthetic ligand comprising a reactive group and a functional reporter group. While the SNAP-tag is the mutant of the DNA repair protein O6-alkylguanine-DNA alkyltransferase enzyme reacting with benzylguanine derivatives, the HaloTag is the mutant of the bacterial haloalkane dehalogenase enzyme binding the reactive linear chloroalkanes (Keppler et al., 2003; Los et al., 2008). Both enzymes react specifically and rapidly with their respective small-molecule ligands, leading to irreversible covalent labeling of HaloTag or SNAP-tag. A wide variety of functional reporter groups can be linked to the enzyme-reactive group of the ligands, lending the much-needed versatility to protein tagging. Nowadays, the commercially available HaloTag or SNAP-tag systems include a wide spectrum of fluorescent dyes for in vitro and in vivo imaging studies as well as conjugation of specific protein fragments, e.g., E3 ubiquitin ligase, enabling proximity-driven degradation of Halo-tagged proteins (Buckley et al., 2015; Deo et al., 2021; Frei et al., 2022; Grimm et al., 2020; Wang et al., 2020). The operational simplicity of working with fully assembled bifunctional ligands can offer distinct experimental advantages over alternative two-step labeling approaches based on click chemistry (Murrey et al., 2015).

All these advances highlight the Halo and SNAP tags as powerful tools for studying protein function using complex biological approaches. The unique feature of these systems is the option for sequential labeling of proteins with different ligands using a pulse-chase approach to monitor protein dynamics, allowing for the real-time tracking of protein trafficking, synthesis, and overall turnover (Adam et al., 2016; Torne et al., 2018). In particular, this method has emerged as a powerful approach for studying the maintenance of chromatin landscape through histone variants, one of the most prominent carriers of epigenetic memory necessary for preserving cellular identity over generations of dividing cells (Stewart-Morgan et al., 2020). The Halo/SNAP-tag–based imaging systems have been successfully used to distinguish between old and new histones and visualize their dynamics throughout the cell cycle at the single-cell level (Adam et al., 2016; Torne et al., 2018). The discoveries of various dedicated histone chaperones guiding the timing and mode of histone deposition during vital cellular processes, such as DNA replication, repair, and transcription, represent a major advance in our understanding of the molecular pathways responsible for chromatin landscape maintenance (Adam et al., 2013; Jansen et al., 2007; Saredi et al., 2024; Torne et al., 2020). Beyond histones, distinct protein variants have been observed through HaloTag-based imaging within cohesin and, more recently, minichromosome maintenance (MCM) protein complexes (Rhodes et al., 2017; Sedlackova et al., 2020; Srinivasan et al., 2020). Cohesin protein complexes play a pivotal role in maintaining the chromatin architecture by organizing the genome into dynamic chromatin loops and facilitating the cohesion of sister chromatids generated during genome duplication (Yatskevich et al., 2019). Recent research has delineated two discrete forms of cohesin protein complexes (Srinivasan et al., 2020). One variant is linked with chromatin postmitotic exit and subsequently transforms into cohesive forms behind the replication forks during the S phase, while the other variant is newly loaded onto nascent DNA at the replication fork. Furthermore, the transition and de novo loading of cohesin complexes are facilitated by a distinct set of proteins reminiscent of histone dynamics (Alonso-Gil and Losada, 2023). MCM protein complexes are essential precursors of genome duplication (Yadav and Polasek-Sedlackova, 2024). During the G1 phase, a massive amount of MCM complexes is loaded on chromatin, a portion of which is converted to active replicative helicase, unwinding duplex DNA, during DNA replication in the S phase. Using the Halo-tag imaging system, our work revealed that MCM complexes exist in different protein forms—parental and nascent MCMs (Sedlackova et al., 2020). These protein forms serve distinct functions during the DNA replication program and are maintained by specific pathways with dedicated chaperones.

Despite the considerable progress in understanding the functional differences among protein variants of histone, cohesin, and MCM complexes, several fundamental questions remain unexplored. For instance, what determines the functional properties of different protein forms, and what chaperones and molecular pathways are involved in generating these variants? Additionally, can protein variants other than histones carry epigenetic information through cellular generations to sustain the regulatory settings of chromatin and DNA replication? The experimental approach and robust technology enabling the efficient labeling (maximum labeling of a given protein pool within short time periods) of different protein variants in the cellular environment and their subsequent capture by affinity methods to answer these questions are currently not well-developed. HaloTag or Snap-tag probes based on biotin–streptavidin interaction represent a logical extension of the self-labeling protein tag technology. For studies of distinct protein variants in the cell, such probes need to display good live-cell permeability, rapid protein labeling, and efficient affinity capture. Meeting these requirements has been challenging thus far, as highlighted by recent studies of the various biotin-HaloTag constructs (Pratik et al., 2022, Preprint; Promega, 2016). In this manuscript, we present the development and comprehensive characterization of a new cell-permeable biotin-HaloTag ligand, which allows for efficient labeling of Halo-tagged proteins in living cells and biochemical separation of the protein variants using streptavidin pull-down.

Various beads and resins have been developed to capture and purify Halo- or SNAP-tagged proteins (Payne et al., 2021; Sridharan et al., 2022). However, these tools are designed to interact with Halo or SNAP tags irrespective of the occupancy of the active site by specific ligands, thereby lacking the capability to separate distinct protein variants. In addition, new and old protein variants of histones, cohesins, and MCMs are generated in a cell cycle–dependent manner (Rhodes et al., 2017; Sedlackova et al., 2020; Stewart-Morgan et al., 2020). Therefore, a maximum labeling efficiency of a given protein pool in live cells within a sufficiently short period of time (≤2 h in the context of the 24-h–long cell cycle) is the critical criterion when considering the experimental setup. Recognizing these limitations, we set out to explore a HaloTag-based experimental approach using biotin–streptavidin interaction for efficient labeling and affinity capture of Halo-tagged proteins (Fig. 1 A). Since two biotin-containing HaloTag ligands are commercially available (Fig. 1 B), namely HaloTag-biotin ligand (here referred to as biotin-[0]-HaloTag ligand 1, or shortly ligand 1) and HaloTag PEG-biotin ligand (here referred to as biotin-[16]-HaloTag ligand 2, or shortly ligand 2), we first evaluated whether they meet the required properties outlined above. To this end, we employed human U2OS cells expressing endogenously Halo-tagged MCM4 subunit using CRISPR-Cas9 genome editing. The MCM4-Halo cell line was thoroughly tested to validate the homozygous tagging of all MCM4 alleles (Fig. S1, A–C). To test the labeling efficiency of commercial biotin-HaloTag ligands, the MCM4-Halo cells were incubated with respective ligands at a final labeling concentration of 2.5 μM for 2 h, followed by a collection of cell lysates and detection of labeled MCM4 by western blotting (WB) (Fig. 1 C). While ligand 1 effectively labeled MCM4-Halo, ligand 2 showed only a modest capacity for labeling (∼20–40%) of the endogenous MCM4 protein under the specified conditions (Fig. 1 C). To further determine whether ligand 1 can efficiently label the entire available MCM4 protein pool present in the cells, we performed a pulse-chase labeling protocol (Fig. 1 D). First, cells were pulsed with biotin-HaloTag ligands, and after a brief wash, fluorescence-based Janelia Fluor (JF)X554-HaloTag ligand was added after additional time elapsed. If ligand 1 labels the entire MCM4 protein pool, then no fluorescence signal should be detected by the JFX554-HaloTag ligand. Indeed, the pulse-chase experiment revealed that ligand 1 effectively labels the entire cellular fraction of MCM4 protein, while ligand 2 shows only a modest labeling capacity (Fig. 1 D). The same observations were reproduced by an orthogonal approach, during which the labeling efficiency of individual ligands was measured as a residual MCM fraction labeled by JFX554-HaloTag ligand at the single-cell level using fully automated quantitative image-based cytometry (QIBC) (Fig. 1 E and Fig. S1 D). The labeling kinetics of both commercial ligands were additionally measured in a concentration- and time-dependent manner (Fig. S1, E and F). While ligand 1 efficiently labeled almost the entire cellular fraction of MCM4-Halo at a final labeling concentration of 1 μM, the major fraction of MCM4 protein, ∼50%, remained unlabeled at the highest concentration of ligand 2 (Fig. S1 E). QIBC of residual MCM4 fraction visualized by fluorescent ligand revealed that extending the treatment time to 6 h is not sufficient to label the entire MCM4 fraction by ligand 2 (Fig. S1 F). The observed slow kinetics of HaloTag labeling by ligand 2 in the context of the 24-h doubling time for the majority of human cells disqualifies this ligand for the purpose of affinity capture of different protein variants.

The second criterion in our assessment of the two commercial biotin-HaloTag ligands is their ability to engage with streptavidin beads. Such interaction enables the capture of biotin-labeled protein complexes, which can be further subjected to proteomic analysis to identify interacting partners or specific posttranslational modifications. Strikingly, a streptavidin-based pull-down experiment revealed that MCM4-Halo labeled by commercial ligand 1 was unable to engage with streptavidin beads, as the entire pool of labeled protein was present in the flow-through fraction (lane 5; Fig. 1 F). In sharp contrast to ligand 1, ligand 2, despite its reduced labeling efficiency (compare input fractions, lanes 2 and 3), was able to interact with streptavidin beads. Pulling down the MCM2 and MCM7 subunits, along with MCM4-Halo, validates the efficacy of this approach in capturing functional MCM complexes (Fig. 1 F). Taken together, our comprehensive characterization of biotin-HaloTag ligands revealed that none of the currently commercially available ligands possess the necessary properties to biochemically study distinct protein variants (Fig. 1 G).

Since commercial ligand 1 failed in the streptavidin pull-down, we explored several strategies to enhance the poor labeling capacity of ligand 2. We argued that the poor labeling of HaloTag by ligand 2 may be attributed to low cell permeability (Promega, 2016). To investigate this aspect, we employed verapamil, a broad-spectrum efflux pump inhibitor previously shown to enhance the labeling efficiency of fluorogenic probes for live-cell imaging of the cytoskeleton (Lukinavicius et al., 2014). Pulse-chase experiments, analyzed either by WB (Fig. 2 A) or QIBC (Fig. 2 B and Fig. S2 A), indeed demonstrated an increase in the labeling efficiency of ligand 2 in the presence of verapamil. This supports the notion that the poor labeling efficiency is attributable to the ligand’s impaired cell permeability rather than its degradation within the cellular environment. However, it is important to note that while verapamil can be advantageous for short-term experiments, it may not be suitable for long-term studies, such as studying different protein forms produced in two successive cellular generations. As shown previously and in this study, the use of verapamil can affect cell growth (Fig. 2 C; and Fig. S2, B and C) (Kania et al., 2017).

To develop a new biotin-based HaloTag ligand capable of efficient labeling and streptavidin pull-down with minimal cell toxicity, we sought to modify the structure of commercial biotin-HaloTag ligands (Fig. 1 B). We hypothesized that the inability of the cell-permeable ligand 1 to bind the streptavidin beads after its covalent attachment to the HaloTag protein might be due to an insufficient atom spacer between the biotin group and the reactive chloroalkane tail. The analysis of available crystal structures revealed that the chloroalkane chain is buried deep within the HaloTag active site. For ligand 1, this could render the pendant biotin group sterically inaccessible and compromise the interaction with streptavidin. Conversely, the 16-atom spacer in ligand 2 allows ample space between the biotin and chloroalkane groups necessary for simultaneous interaction with HaloTag and streptavidin; however, it comes at the expense of low labeling efficiency due to presumed poor cell permeability (see above). Following the rationale, we used chemical synthesis to prepare a set of biotin-HaloTag ligands featuring variable-length amino acid spacers, which, we hoped, could combine the desired characteristics of the two commercial ligands. Our synthetic approach is overviewed in Fig. 2 D. Inspired by prior literature (Los et al., 2008; So et al., 2008), the known amine-substituted chloroalkane component was coupled to various N-protected amino acid spacers. Following the removal of the nitrogen-protecting group, biotin was attached via its active (NHS) ester. Using the outlined synthetic approach, we first produced two biotin-HaloTag ligands, denoted as ligands 3 and 4, each containing a 10-atom spacer (Fig. 2 E), representing the atom count midpoint. Ligand 4, prepared from the hydroxy-substituted chloroalkane component, is an ester analog of the known ligand 3 (So et al., 2008) and was chosen in light of the recent report on amide-to-ester substitution as a strategy to increase cell permeability of PROTAC constructs (Klein et al., 2021). The labeling efficiency of ligands 3 and 4 was examined by WB (Fig. 2, F and G). While ligand 3 showed slightly increased labeling efficiency, its ester analog (ligand 4) was inferior relative to the commercial ligand 2. The modest increase in labeling efficiency seen with ligand 3 suggested that shortening the atom spacer may be used to improve the labeling properties. Next, to investigate whether ligand 3, having the shorter atom spacer (relative to ligand 2), retains its ability to bind streptavidin, we performed the streptavidin pull-down (Fig. 2, H and I). The results confirmed that the ligand 3-HaloTag conjugate maintained the capacity to interact with streptavidin beads. The functional relevance of this interaction was demonstrated by the pull-down of other MCM subunits, specifically MCM7 and MCM2, indicating the successful capture of functional MCM complexes. Overall, these findings suggested that a judicious design of the atom spacer in the biotin-HaloTag ligands is a viable strategy to enhance labeling capacity while preserving the ability to interact with streptavidin beads.

Motivated by observations in Fig. 2, we decided to shorten the 10-atom spacer further to attain the labeling efficiency of commercial ligand 1. To help determine the appropriate spacer length, we created a computational model for the visualization of the HaloTag–ligand–streptavidin ternary complexes (Fig. 3 A and Fig. S2, D–G). These were assembled from the available crystal structures of HaloTag–chloroalkane and streptavidin–biotin binary complexes, respectively. First, an atom spacer of a chosen length was attached to the carboxylate of biotin within the streptavidin–biotin complex (PDB: 3RY2; a single subunit of streptavidin tetramer was used) (Le Trong et al., 2011), and multiple conformers of the atom spacer were generated by constrained embedding in RDKit. Only the lowest energy spacer conformers (within 1 kcal/mol) were kept. Secondly, the HaloTag–chloroalkane complex (PDB: 6U32, modified) (Deo et al., 2021) was attached to the carboxy group of each of the computed spacer conformations using the least squares fit of three specific atoms (Fig. S2 E). Finally, the computed ternary complexes were ranked using the number of overlapping heavy atoms, including and excluding flexible loops, to identify all valid complexes (Fig. S2, F and G). The representative structural visualizations of the commercial ligands corroborate our earlier experimental findings (Fig. 1). No viable ternary complex could be identified for ligand 1, as it is evidently too short to accommodate both HaloTag and streptavidin proteins (Fig. 3 A; and Fig. S2, F and G). This is in contrast to ligand 2 having a 16-atom spacer (Fig. 3 A; and Fig. S2, F and G) and ligand 3 with a 10-atom spacer (Fig. S2, D–G). Based on the structural analysis, we designed a new ligand with a 7-atom spacer (ligand 5, middle in Fig. 3 A), representing a midpoint comparison between 4-atom and 10-atom spacers (Fig. S2, F and G). Despite being nine atoms shorter than the commercial ligand 2, ligand 5 was predicted by the computation to give multiple valid ternary HaloTag–ligand–streptavidin complexes (Fig. S2, F and G).

Following our synthetic scheme (Fig. 2 D), we prepared the above-designed ligand 5 (Fig. 3 B). Notably, western blot analysis demonstrated that ligand 5 achieves high labeling of Halo-tagged MCM4 comparable with that of commercial ligand 1 (Fig. 3, C and D), demonstrating that shortening the spacer from 16 to 7 atoms substantially enhanced the labeling efficiency. In addition, the labeling efficiency was tested by pulse-chase experiments and evaluated by WB and QIBC, indicating that ligand 5 proficiently labels nearly the entire fraction of MCM4 protein in the cellular environment (Fig. 3, E and F; Fig. S2 H; and Fig. S3, A and B). The robustness of our synthesis and data was confirmed in experiments assessing the labeling kinetics in a concentration- and time-dependent manner using two independent batches of the synthetic ligand 5 (Fig. S3, C and D). Both batches of ligand 5 showed the same labeling kinetics, achieving ∼90% labeling of the cellular MCM4 fraction at a concentration of 2.5 μM within 2 h. Altogether, our findings unequivocally demonstrate that the introduction of a 7-atom spacer into the biotin-HaloTag ligand substantially increased the labeling efficiency of Halo-tagged proteins.

As ligand 5 demonstrates a comparable labeling efficiency to commercial ligand 1, we next evaluated the newly synthesized ligand for its ability to interact with streptavidin beads as predicted by our structural visualizations. This was done using a streptavidin-based pull-down assay (Fig. 3 G), which demonstrated that ligand 5, upon conjugation to HaloTag, retains its ability to bind streptavidin. Importantly, interaction with other MCM subunits was observed, indicating the successful capture of functional MCM complexes through ligand 5-HaloTag conjugation. Although the efficiency of streptavidin pull-down mirrored that of commercial ligand 2, a substantial portion of MCM4-Halo labeled by ligand 5 was evident in the flow-through fraction (compare lanes 5 and 6 of the western blot in Fig. 3 G). This observation may be rationalized by the approximately fivefold higher labeling efficiency of ligand 5 compared with commercial ligand 2 (for further details, see quantification in Fig. 3 D, or compare lanes 2 and 3 in the input fraction of Fig. 3 G). Such increased labeling efficiency could potentially exhaust the pull-down capacity of the streptavidin beads. In support of this argument, upon increasing the amount of streptavidin beads, we enhanced the pull-down efficiency for MCM4-Halo labeled by ligand 5 (compare Fig. 3, G and H). Notably, the efficient binding to streptavidin can be utilized beyond affinity pull-down, for instance, for the visualization of protein complexes directly in their natural environment. Thus, we employed Alexa Fluor 647 (A647)-conjugated streptavidin to detect ligand 5–bound MCM–Halo complexes within the cell nucleus using QIBC. The analysis confirmed the capacity of ligand 5 to label nuclear as well as chromatin-bound fractions of MCM complexes (Fig. S4, A and B). No signal was observed in naïve U2OS cells, indicating that labeling by ligand 5 is HaloTag-specific. Additionally, in alignment with previous literature (Polasek-Sedlackova et al., 2022), typical chromatin dynamics of MCM complexes labeled by ligand 5 were evident throughout the cell cycle, indicating the visualization of MCM complexes involved in the genome duplication process (Fig. S4, B and C).

The successful generation and favorable properties of biotin-HaloTag ligand 5 prompted us to examine an analogous ligand having only a 4-atom spacer (ligand 6, Fig. S4 D). The structural visualization model introduced above (Fig. S2, F and G; and Fig. S4 E) suggested the 4-atom spacer to be a borderline case and had to be assessed experimentally. The synthesis of ligand 6 was straightforward following our approach (Fig. 2 D), as the spacer came in the form of commercially available N-protected β-alanine. While ligand 6 showed marginal improvement in the labeling efficiency of Halo-tagged MCM4 compared with ligand 5, it lost the ability to interact with streptavidin (i.e., ternary complex formation; Fig. S4, F and G). These observations underscore the importance of judicious optimization of the atom spacer in the design of biotin-HaloTag ligands.

To further emphasize the superior labeling efficiency of ligand 5 in comparison to commercial ligand 2, we have conducted a direct assessment of the labeling kinetics of ligand 5 alongside both commercial ligands (Fig. S5, A–D). This analysis was performed in a concentration- and time-dependent manner through pulse-chase experiments, as well as direct visualization with A647-conjugated streptavidin using QIBC. Our comprehensive set of experiments demonstrates that even at its highest concentration, ligand 2 achieves only 40–50% labeling efficiency in direct comparison to the 90–100% labeling capacity of ligand 5. Notably, as shown in previous experiments, ligand 5 displays a labeling efficiency of the cellular MCM4 fraction comparable with that of commercial ligand 1. Overall, the newly identified ligand 5 with a 7-atom spacer is proficient in cell labeling assays and streptavidin pull-down experiments and superior to the commercially available biotin-HaloTag ligands. These properties render ligand 5 a candidate small-molecule tool for studying the biochemical properties of distinct protein variants.

As mentioned in the preceding paragraphs, studying different protein forms across successive cellular generations may require longer exposure of cells to the new biotin-HaloTag ligand 5. Therefore, we determined the effects of ligand 5 on cell cycle progression and growth using QIBC and cell survival assays. The results demonstrated that extended exposure to ligand 5 did not elicit alterations in the cell cycle (Fig. S6, A and B), nor did it impede the growth of U2OS and MCF7 cells (Fig. 4 A; and Fig. S6, C and D). These findings, along with prior characterizations, suggest that ligand 5 is well suited for long-term studies, such as investigating the various protein forms generated across multiple cell divisions (Fig. 4 B).

Throughout the manuscript, we systematically evaluated the labeling efficiency and streptavidin pull-down capacity of commercial and newly synthesized biotin-HaloTag ligands in U2OS cells, endogenously expressing the Halo-tagged MCM4 subunit. To reinforce our findings beyond a single protein target, we opted to broaden our exploration of biotin-HaloTag ligands with different cell systems expressing distinct Halo-tagged proteins. To achieve this, we implemented CRISPR-Cas9 endogenous tagging to establish a U2OS cell line expressing Halo-tagged MCM2. The resulting cell line was extensively validated for the homozygous tagging of all MCM2 alleles (Fig. S6, E–G). Importantly, ligand 5 exhibited high labeling efficiency of Halo-tagged MCM2 and demonstrated the capacity to effectively pull down functional MCM complexes using streptavidin beads (Fig. S6, H and I; and Fig. S7 A). These results validate ligand 5 as a tool for efficient labeling and biochemical separation of different protein variants within MCM complexes. Next, we employed human HeLa cells that ectopically expressed a Halo-tagged histone variant H2B, a commonly used tool for studying chromatin architecture (Saxton et al., 2023). The expression level of H2B-Halo in the cells was validated by WB and QIBC (Fig. S7, B and C). Notably, ligand 5 exhibited a high labeling capacity for Halo-tagged H2B, well comparable with that of commercial ligand 1, as measured by WB and QIBC (Fig. 4 C; and Fig. S7, D and E). Moreover, using ligand 5 in conjugation with Halo-tagged H2B, we successfully captured functional nucleosomes through streptavidin pull-down (Fig. 4 D). Collectively, we conclude that the biotin-HaloTag ligand 5 is a new ligand encompassing efficient labeling and streptavidin pull-down capacity, properties that are unmatched by the commercially available ligands.

Based on our observations, we propose utilizing the new ligand 5 in pulse-chase experiments to investigate the biological properties of different protein variants through their biochemical separation. Previous studies have provided a detailed description of pulse-chase protocols to visualize different protein variants of Halo-tagged histones, cohesins, and MCM complexes generated through two successive cellular generations using one or two fluorescent HaloTag ligands (Rhodes et al., 2017; Sedlackova et al., 2020; Torne et al., 2018). We posit that the previously established cell lines and protocols can be readily adapted for the biochemical capture and isolation of diverse protein variants. To support this claim experimentally, we employed the MCM4-Halo cell line and adopted a previously established labeling protocol (Sedlackova et al., 2020) to label parental and nascent MCMs with ligand 5 (Fig. 4 E). Compound S26 (Data S1) was used as a blocking ligand. As previously reported, parental MCMs are preferentially converted into active replicative CMG helicases (CDC45-MCM2-7-GINS1-4), whereas nascent MCMs predominantly remain inactive during DNA replication (Sedlackova et al., 2020). Consequently, the successful separation of parental and nascent MCM protein variants can be effectively distinguished through the specific pull-down of CMG components, such as CDC45 and GINS4. Results from the streptavidin pull-down showed that comparable amounts of parental and nascent MCM4 were captured (Fig. 4, E and F). However, CDC45 and GINS4 were present exclusively in the parental fraction, indicating successful biochemical separation of the parental and nascent MCMs (Fig. 4, E and F). We anticipate that this approach and methodology can be applied to proteins beyond MCM complexes. Nonetheless, the optimization of labeling intervals in pulse-chase protocols will be essential to align with the turnover rates of the specific proteins of interest.

In summary, the outlined experimental strategy can be employed to biochemically separate distinct protein variants (Fig. 4 G), facilitating the investigation of their respective interactomes, posttranslational modifications, and chromatin occupancy using advanced proteomic and next-generation sequencing techniques. This, in turn, will provide mechanistic insights into fundamental cellular processes, including genome duplication, chromatin maintenance, and transmission of epigenetic information across generations of dividing cells. We elaborate on the specific biological questions that could be addressed using the ligand 5–based approach in the last paragraphs of the discussion below.

HaloTag technology represents a versatile platform for studying protein function using a range of biological methods (England et al., 2015). Its versatility stems from the modular nature and synthetic accessibility of HaloTag ligands, each incorporating different functional reporter groups conjugated to the reactive chloroalkane. Fluorescent HaloTag probes, particularly those based on the series of rhodamine dyes, are now widely utilized in various biochemical and biological imaging studies (Frei et al., 2022; Grimm et al., 2020; Grimm et al., 2021; Kompa et al., 2023). A recently developed HaloPROTAC ligand, which facilitates the rapid degradation of Halo-tagged proteins via von Hippel–Lindau E3 ubiquitin ligase-based ligand, greatly expands the use of HaloTag technology beyond imaging studies (Buckley et al., 2015). However, a HaloTag ligand that would enable effective labeling of cellular proteins and their pull-down by affinity methods has been missing from the available HaloTag toolbox.

HaloTag affinity ligands based on the biotin–streptavidin interaction would represent a potentially powerful tool for the affinity capture and purification of Halo-tagged proteins and, in particular, for the biochemical separation of various protein variants generated in successive cellular generations. Currently, two biotin-based HaloTag ligands are available commercially (ligand 1 and ligand 2). Unfortunately, our comprehensive characterization of these two ligands revealed that they do not display optimal properties: (1) efficient in vivo labeling, (2) ability to engage with streptavidin beads, and (3) minimal cell toxicity. While ligand 1 achieves high labeling efficiency of the cellular fraction of Halo-tagged proteins, it fails in streptavidin pull-down experiments. Conversely, ligand 2, after conjugation with HaloTag, effectively interacts with streptavidin but shows poor labeling efficiency of Halo-tagged proteins in the cellular environment due to limited cell permeability. These experiments, paired with computer-assisted structural visualizations, pointed to the critical nature of the atom spacer length situated between the biotin functional group and the reactive chloroalkane group. Our findings aligned well with a previous study in which this issue was noted but not addressed experimentally (Pratik et al., 2022, Preprint). The comprehensive characterization of the new ligands prepared in this work demonstrates how the judicious choice of the linker length can fine-tune the properties of the HaloTag ligands.

The newly identified ligand 5 (biotin-[7]-HaloTag ligand 5) effectively labels Halo-tagged proteins in living cells and enables their pull-down through streptavidin beads without inducing cytotoxicity. We conducted extensive validation of the properties of ligand 5 for various proteins and cell lines using a comprehensive array of techniques encompassing labeling, pulse-chase procedures, and streptavidin pull-downs, assessed via WB and high-content imaging. Although ligand 5 exhibited slower kinetics in labeling Halo-tagged proteins in the cellular environment compared with commercial ligand 1, it still efficiently labeled the entire protein pool of Halo-tagged proteins in a relatively short time (2 h) and at low concentrations (2.5 μM). In sharp contrast to commercial ligand 1, however, ligand 5 can be readily captured by streptavidin beads. Based on these data, we believe that ligand 5 effectively fills the gap in the palette of available HaloTag ligands for labeling and affinity pull-down of Halo-tagged proteins in their cellular environment. The tool empowers the HaloTag technology to investigate the functional differences among protein variants generated through two successive cellular generations (Fig. 4 G). In the past, Halo- or SNAP-tag–based pulse-chase experiments using fluorescent ligands across cellular generations have proven to be an invaluable method for uncovering distinct protein variants within histone, cohesin, and MCM complexes (Rhodes et al., 2017; Sedlackova et al., 2020; Torne et al., 2018). In the last paragraphs, we outline how the ligand 5–based pulse-chase experiments can be used to capture and biochemically separate different protein pools produced by successive cellular generations and how this approach can help answer fundamental questions about the regulation of genome duplication and chromatin maintenance in the future.

Genome maintenance processes, such as DNA replication and repair, are vital for preserving the integrity of genetic information throughout cellular generations. Although essential, these processes disrupt the chromatin landscape, which is crucial for preserving the gene expression program that governs cell identity. The nucleosome, a histone octamer, constitutes the fundamental unit of chromatin architecture and is required to be dismantled by an incoming replication fork to duplicate DNA and subsequently reassembled at the newly synthesized DNA (Stewart-Morgan et al., 2020). More than 40 years of extensive research have unveiled that old histones preceding the DNA replication fork are recycled within postreplicative chromatin alongside the deposition of newly synthesized histones. These old and new histone variants differ in molecular pathways responsible for their deposition, as well as through carried posttranslational modifications, representing a central source of cellular epigenetic memory (Stewart-Morgan et al., 2020). Recent advancements in sequencing-based methodologies have furnished valuable insights into the molecular mechanisms underpinning the recycling of old histones at the replication fork (Flury et al., 2023; Charlton et al., 2024; Li et al., 2020; Petryk et al., 2018; Yu et al., 2018). In view of these recent breakthroughs, we hypothesize that employing ligand 5–based pulse-chase experiments in combination with the isolation of proteins on nascent DNA method may serve as an intriguing approach to elucidate the interactomes of new and old histones at the replication fork (Fig. 4 G). Such an approach holds promise for unveiling novel histone chaperones and molecular mechanisms responsible for the maintenance of chromatin landscape and, thereby, the preservation of epigenetic memory.

Similar to histones, cohesin complexes play a crucial role in maintaining chromatin architecture, and their behavior during the cell cycle is reminiscent of that of histones. Initially, the cohesin complex was known for its function in holding sister chromatids during genome duplication; nevertheless, later on, its major role in organizing the topology of interphase chromatin was discovered (Yatskevich et al., 2019). These observations raised an important conundrum, namely, what governs the regulation of cohesin activity to transition from loop extrusion of prereplicative chromatin to sister chromatid cohesion in postreplicative chromatin (Srinivasan et al., 2020)? Previous research indicated that the answer may lie in understanding the dynamics of different forms of cohesin during the cell cycle. While one cohesin variant promotes loop extrusion after postmitotic exit and, during the S phase, is transformed into cohesive forms behind the replication forks, the other variant is newly loaded onto nascent DNA at the replication fork (Srinivasan et al., 2020). Furthermore, recent studies revealed that transition and de novo loading of cohesin complexes are facilitated by a distinct set of proteins, reviewed by Alonso-Gil and Losada (2023). In light of these recent findings, we speculate that employing ligand 5 in pulse-chase experiments in combination with the mass spectrometry method may serve as a potent approach to explore the interactomes and posttranslational modifications of different cohesin variants (Fig. 4 G), thereby offering novel and valuable insights into the regulation of cohesin enzymatic activities on pre- and postreplicative chromatin.

Finally, our recent work has unveiled distinct protein variants within MCM complexes, the central players of genome duplication (Sedlackova et al., 2020). Upon inheritance, daughter cells receive both parental and nascent protein forms, each fulfilling specific functions during the DNA replication process, and their levels are preserved through distinct pathways. While parental MCMs are preferentially converted to active replisomes, the nascent MCMs remain largely inactive but serve as natural replisome pausing sites, sustaining physiological replication fork speed. Despite the valuable insights provided by our work into MCM biology, several key questions remain unresolved. Specifically, what defines the functional properties of parental and nascent MCMs that govern their functions during the DNA replication process? Furthermore, what protein chaperones are responsible for MCM biogenesis and recycling pathways? Building on our findings in this work, we propose that combining ligand 5–based pulse-chase experiments with mass spectrometry may represent an effective strategy to explore the posttranslational modifications of different MCM protein variants and identify novel protein chaperones crucial for maintaining MCM equilibrium (Fig. 4 G). Furthermore, the integration of next-generation sequencing methods holds promise for yielding valuable insights into replication origin activation and their mapping in the mammalian genome, a fundamental yet largely enigmatic inquiry.

In previous research, the HaloTag-based sequential labeling technique employing one or two fluorescent ligands proved invaluable in uncovering distinct protein variants within histone, cohesin, and MCM complexes. These discoveries have advanced our understanding of the molecular pathways involved in maintaining genome stability, with direct implications for elucidating disease pathogenesis. Following the imaging studies, the next essential step requires the biochemical isolation of the distinct protein variants to facilitate a comprehensive examination of their biological properties. While the experimental HaloTag approach based on biotin–streptavidin interaction is well suited for such studies, the practical implementation has proven to be challenging. In this work, we combined chemical synthesis, computer-assisted structural visualizations, and comprehensive cell biology–based characterization to examine several new biotin-HaloTag ligands. Our results underscore the importance of the atom spacer (linker) length in effectively fine-tuning the HaloTag ligands and developing superior variants. The effort led to a new cell-permeable biotin-[7]-HaloTag ligand 5 (ligand 5), which exhibits efficient labeling of Halo-tagged proteins in the cellular environment of live cells and enables their pull-down through streptavidin beads without inducing cytotoxicity. Ligand 5 is demonstrated to effectively fill the gap in the spectrum of available HaloTag ligands for efficient cellular labeling and affinity pull-down of Halo-tagged proteins. Most notably, ligand 5 will enable researchers to apply HaloTag technology to studies of functional differences among protein variants using various biochemical approaches. Our work presents the proof of concept of how to utilize the herein-developed biotin-HaloTag ligand for the biochemical separation of parental and nascent MCM complexes. Additionally, we outline an experimental design for investigating the biological properties of distinct protein variants. This encompasses the examination of interacting partners, posttranslational modifications, and chromatin occupancy using advanced proteomic and next-generation sequencing techniques. Overall, we anticipate that this methodology will emerge as a robust approach for gaining mechanistic insights into fundamental cellular processes essential for organismal fitness and survival.

A detailed description of the experimental procedures for the synthesis of individual biotin-HaloTag ligands and their structural analysis can be found in Data S1.

The ternary complexes were constructed using a two-step procedure that combined information from two previously determined crystal structures: a modified HaloTag–chloroalkane complex (PDB: 6U32) and a single subunit of the streptavidin–biotin complex (PDB: 3RY2). In the first step, a linker of a selected length (0, 4, 7, 10, or 16 heavy atoms) was covalently attached to the carboxylate group of the biotin ligand within the streptavidin–biotin complex. Constrained embedding via the RDKit software (version 2023.03.3) was then set to generate up to 100,000 conformations from which energetically favorable conformations of the ligand were selected—specifically, those with energy below 1 kcal/mol (1,000 steps of energy minimization, calculated using the MMFF94 force field)—ensuring that the linker adopts a realistic geometry. In the second step, each computed conformation of the ligand was attached to the HaloTag–chloroalkane complex. This attachment was performed by aligning three corresponding atoms (3-atom alignment in blue in Fig. S2 E) of the two ligands using a least-squares fitting method, which optimizes the spatial overlap in a way that preserves realistic molecular geometry. Once assembled, the ternary complexes were evaluated and ranked based on the number of overlapping heavy atoms between the HaloTag and streptavidin (Fig. S2 F).

The ranking of modelled ternary complexes is based on the extent of steric clashes (overlaps) between the HaloTag and streptavidin components. In this analysis, each line corresponds to a different ligand that bridges the two proteins, with the linker portion of the ligand comprising 0, 4, 7, 10, or 16 heavy atoms. Here, an “overlap” is defined as any instance where a heavy atom from one protein lies within 3 Å of a heavy atom from the other protein—a condition that indicates a potential steric clash. The inset of Fig. S2 F highlights the 25 top-ranking complexes that exhibit the fewest or no overlapping heavy atoms. Within this selection, ligands with linker lengths of 10 and 16 heavy atoms (i.e., ligands 2 and 3) produced ternary complexes that were free of such overlaps, consistent with the expectation that longer linkers are better at accommodating the spatial requirements of the two proteins. Ligand 5 displayed only a limited number of overlaps, and valid complexes were obtained after considering the flexible protein loops (Fig. S2 G). No viable complexes could be identified for ligand 1 using the analysis described above. The number of valid ternary complexes is defined as those in which no heavy atom overlaps occur between HaloTag and streptavidin monomer. The analysis was performed considering both the entire protein structure, including flexible loops, and a reduced model, excluding flexible loops (Fig. S2 G). The distinction was made because protein flexibility and dynamics were not explicitly accounted for in the ternary complex assembly. Linker length refers to the number of heavy atoms in the spacer. The total number of evaluated complexes in the last column of Fig. S2 G stems from the number of linker conformations with energy below 1 kcal/mol.

The human osteosarcoma cell line U2OS (RRID:CVCL_0042, ATCC, HTB-96), the adenocarcinoma mammary epithelial cell line MCF7 (RRID:CVCL_0031, ATCC, HTB-22), and the cervical HeLa Kyoto cell line (RRID:CVCL_0030, obtained from S. Narumiya) (Ochs et al., 2019) used in this study were cultured in DMEM containing high glucose and GlutaMAX (31966047; Thermo Fischer Scientific) and supplemented with 10% FBS (A5256801; Thermo Fischer Scientific) along with 0.5% penicillin-streptomycin (15140122; Thermo Fischer Scientific). The cells were grown under standard sterile conditions at 37°C and 5% CO₂ level. CRISPR-Cas9–mediated derivatives of U2OS cells expressing C-terminal endogenously tagged MCM2-Halo and MCM4-Halo were generated and validated in the previous study (Sedlackova et al., 2020). Additional validation of these cell lines by WB, junction PCR, and QIBC is provided in Fig. S1, A–C and Fig. S6, E–G. The primers used for junction PCR were as follows: for MCM4 (forward 5′-CCC​CTT​TCA​TAG​TTG​TTA​TAA​GTT​T-3′ and reverse 5′-CTA​CAA​ATA​TTG​AAC​CCA​AGT​TAG​A-3′); for MCM2 (forward 5′-GCC​CAG​CAG​GAC​ACT​ATT​GA-3′ and reverse 5′-TGA​ATA​CGC​AAC​TCA​CGC​CA-3′).

To generate a cell line stably expressing H2B-Halo, the HeLa cells were transfected with a plasmid (pHCT-CMV-neo-H2B-Halo) (Ochs et al., 2019) using Lipofectamine LTX with Plus reagent (15338100; Thermo Fischer Scientific). The transfected cells were selected with DMEM containing geneticin (G-418) (0.4 mg ml⁻¹, 4727878001; Sigma-Aldrich) for 7 days, serially diluted and seeded onto 100-mm dishes, and grown under G-418 selection for an additional 12 days. The isolated colonies were expanded and tested for H2B-Halo expression by labeling with fluorescent JFX554 HaloTag ligand (the labeling protocol is described in the section “HaloTag ligands and labeling protocol”). The generated H2B-Halo HeLa cell line was validated via SDS-PAGE and QIBC. All cell lines and their derivatives were regularly tested for mycoplasma using the Mycoplasma Detection Kit (InvivoGen, rep-mys-50) and were always found negative.

Verapamil (331-20649-1; RayBiotech) was used as indicated in the figure panels. For the detection of biotin-HaloTag ligand-tagged proteins, HRP-conjugated streptavidin (1:1,000, 89880D; Thermo Fischer Scientific) or A647-conjugated streptavidin (1:500, S32357; Thermo Fischer Scientific) was used. Primary antibodies used for immunofluorescence (IF) were as follows: MCM2 (rabbit, 1:1,000, 10513-1-AP; Proteintech, RRID:AB_2142131), MCM4 (rabbit, 1:1,000, 13043-1-AP; Proteintech, RRID:AB_2142293), and PCNA (human, 1:1,000, 2037; Immuno Concepts). Primary antibodies for WB: α-tubulin (rabbit, 1:3,000, 11224-1-AP; Proteintech, RRID:AB_2210206), Halo (mouse, 1:1,000, 28A8; Proteintech, RRID:AB_2827565), H3 (rabbit, 1:5,000, 17168-1-AP; Proteintech, RRID:AB_2716755), MCM2 (mouse, 1:1,000, H00004171-M01; Novus Biologicals, RRID:AB_490071), MCM4 (rabbit, 1:2,000, 13043-1-AP; Proteintech, RRID:AB_2142293), and MCM7 (mouse, 1:1,000, sc-9966; Santa Cruz, RRID:AB_627235). For IF, secondary antibody conjugates were goat anti-rabbit A647 (1:2,000, A-21245; Thermo Fischer Scientific, RRID:AB_2535813) and donkey anti-human A647 (1:2,000, 709-605-149; Jackson Immuno Research, RRID:AB_2340578). For WB, secondary antibody conjugates were HRP horse anti-mouse IgG antibody (1:10,000, PI-2000; Vector Laboratories, RRID:AB_2336177) and HRP goat anti-rabbit IgG antibody (1:10,000, PI-1000-1; Vector Laboratories, RRID:AB_2916034).

For single labeling by biotin-HaloTag ligands, MCM4-Halo U2OS cells were incubated with indicated ligands at a final labeling concentration of 2.5 μM for 2 h or at increased concentrations as specified in the figure panels for 2 h. For concentration-dependent experiments, increasing ligand concentration was used as specified in the figure panels for 2 h. For all dual-HaloTag labeling utilizing biotin-HaloTag ligands and JFX554-HaloTag ligand, MCM4-Halo U2OS, MCM2-Halo U2OS, or H2B-Halo HeLa cells were pulsed with indicated biotin-HaloTag ligands at a final concentration of 2.5 μM for 2 h or for the indicated time points as specified in figure panels. The biotin-HaloTag ligands were then removed by washing cells three times with PBS buffer (14190250; Thermo Fisher Scientific), and the cells were incubated with the JFX554-HaloTag ligand in a final concentration of 25 nM for 30 min based on kinetics experiments. For dual-HaloTag labeling utilizing ligand 5 and S26 compound (Fig. 4 E), MCM4-Halo U2OS cells were pulsed with ligand 5 (for parental MCM interactome) or S26 (for nascent MCM interactome) in a final concentration of 2.5 μM for 2 h. After pulse, cells were washed three times with 1× PBS buffer and incubated with S26 (for parental MCM interactome) or ligand 5 (for nascent MCM interactome) in a final concentration of 5 μM for 24 h. For all single JFX554-HaloTag labeling, MCM4-Halo U2OS, MCM2-Halo U2OS, or H2B-Halo HeLa cells were incubated with the JFX554-HaloTag ligand in a final concentration of 25 nM for 30 min. After HaloTag labeling, cell samples were further processed for WB or IF staining.

Whole-cell lysates were obtained by incubating cells in lysis buffer (10 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, and 0.5% NP-40) supplemented with protease and phosphatase inhibitors (4693116001 and 4906845001; ROCHE) and 250 U per ml of benzonase (E1014-25KU; Merck) on ice for 45 min. For the preparation of whole-cell lysates from H2B-Halo HeLa cells, cell lysis on ice was followed by sonication. The cell lysates were denatured at 95°C for 5 min in Laemmli loading buffer containing NuPAGE LDS Sample Buffer (NP0007; Thermo Fisher Scientific) and NuPAGE Sample Reducing Agent (NP0009; Thermo Fisher Scientific) and separated on mPAGE 4–12% Bis-Tris Precast Gels (MP41G10 or MP41G12; Merck) through standard SDS-PAGE protocol. After SDS-PAGE, gels were scanned with a Cy3 wavelength to detect MCM4-Halo or H2B-Halo labeled with the fluorescent JFX554 HaloTag ligand. Afterward, the separated proteins were transferred from the gel to a nitrocellulose membrane (IB23002X3; Thermo Fisher Scientific) using iBlot (IB21001; Thermo Fisher Scientific). Ponceau S solution (P7170-1L; Merck) was used for total protein staining. To detect MCM4-Halo, MCM2-Halo, or H2B-Halo labeled by nonfluorescent biotin-HaloTag ligands, the membrane was blocked with Nucleic Acid Detection Blocking Buffer (NADB) (89880A; Thermo Fischer Scientific) for 30 min, followed by incubation with HRP-coupled streptavidin (89880D; Thermo Fischer Scientific) in NABD for 1 h at room temperature. To detect the target proteins, the membrane was blocked with PBS buffer containing 5% milk (70166-500G; Merck) and 0.1% Tween-20 (P1379-500ML; Merck) for 1 h at room temperature, followed by primary and secondary antibody staining. Both primary and secondary antibodies were diluted in PBS buffer containing 5% milk and 0.1% Tween-20 and then incubated overnight at 4°C or 2 h at room temperature, respectively. After incubation, the HRP signal was detected by ECL Select Western Blotting Detection Reagent (GERPN2235; Merck). After the detection of target proteins, the membrane was stripped using Restore PLUS Western Blot Stripping Buffer (46430; Thermo Fischer Scientific) and stained with MCM7 (Fig. S1 E), α-tubulin (Fig. S1 A and Fig. S6 E), or H3 (Fig. 4 C) antibodies as loading controls. For processing controls (Fig. 1 C, Fig. 2 F, Fig. 3 C, and Fig. S3 B), the same lysates and loading concentrations were used, and the samples were run in parallel to detect target proteins. All western blot images were acquired using a FUJIFILM LAS-3000 Image Analyzer equipped with a 3.2-megapixel CCD camera. Image acquisition was performed using the manufacturer’s software (LAS-3000; Image Reader, version 2.2, used in “Lite” mode), with exposure settings optimized to avoid pixel saturation. Western blot band intensities were quantified using the Gel Analyzer tool in ImageJ software (FIJI, version 1.54i). Each lane was defined manually, and the area under the curve for each band was measured. Relative band intensities were calculated and expressed as percentages, as detailed in the figure legends.

Whole-cell lysates obtained from indicated cells were incubated with either 40 μl of anti-streptavidin agarose beads (20347; Thermo Fisher Scientific) for 2 h at 4°C (Fig. 1 F, Fig. 2 H, Fig. 3 G, Fig. 4 D, Fig. S4 G, and Fig. S6 I) or 100 μl of anti-streptavidin agarose beads overnight at 4°C (Fig. 3 H and Fig. 4 E). Beads were then washed twice with low-salt wash buffer (10 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, and 0.5% NP-40) supplemented with protease and phosphatase inhibitors and twice with high-salt wash buffer (10 mM Tris-Cl, pH 7.5, 500 mM NaCl, 0.5 mM EDTA, and 0.5% NP-40) supplemented with protease and phosphatase inhibitors. To elute bound proteins, beads were incubated with 60 μl of Laemmli loading buffer for 20 min at 95°C. The elutes were analyzed by WB using either streptavidin-HRP to detect BIOTIN-HaloTag ligand-labeled proteins or specific antibodies, as indicated in the figures.

For IF staining, cells were grown on round 12-mm diameter, 1.5-mm–thick glass coverslips (MENZCB00120RAC20; VWR). For staining of chromatin-bound proteins, cells were pre-extracted with ice-cold cytoskeleton buffer (10 mM Hepes, pH 7.5, 300 mM sucrose, 100 mM NaCl, 3 mM MgCl₂, and 0.5% Triton X-100) for 10 min at room temperature, washed three times with PBS, and fixed using 4% buffered formaldehyde (9713.1000; VWR) for 20 min at room temperature. To stain total nuclear proteins, cells were fixed directly with 4% buffered formaldehyde for 20 min at room temperature, washed three times with PBS, and incubated with ice-cold PBS containing 0.2% Triton X-100 for 5 min at room temperature. Both primary and secondary antibodies were diluted in fresh DMEM supplemented with 10% FBS and incubated with coverslips at room temperature for 90 and 45 min, respectively. To counterstain nuclear DNA, the secondary antibody cocktail was supplemented with 0.5 µg per ml DAPI (D1306; Thermo Fisher Scientific, RRID:AB_2629482). After secondary antibody incubation, the coverslips were washed three times with PBS and twice with distilled water, air-dried, and mounted on slides using Mowiol-based mounting medium (12% Mowiol 4–88 [81381-250G; Merck], 30% glycerol, and 0.12 M Tris-HCl, pH 8.5). To detect Halo-tagged proteins labeled by biotin-HaloTag ligands, cells were fixed or pre-extracted as described above. Coverslips were blocked with NADB (89880A; Thermo Fischer Scientific) for 30 min at room temperature, washed three times with PBS, and incubated with A647-conjugated streptavidin (S32357; Thermo Fischer Scientific) and DAPI in PBS buffer for 1 h at room temperature. After incubation, the coverslips were processed as described above.

Images were acquired using an inverted screening microscope ScanR (IX83; Evident), equipped with a UPLXAPO dry objective (20X, 0.8 NA); a fast excitation and emission filter wheel for DAPI, FITC, Cy3, and Cy5 wavelengths; a Lumencor Spectra X LED fluorescence light source; and digital monochromatic sCMOS ORCA-Flash 4.0 LT Plus camera (C11440-42U30; Hamamatsu, resolution: 2,048 × 2,048, pixel size: 6.5 × 6.5 μm, effective area: 13.312 × 13.312 mm). The images were acquired with the ScanR acquisition software (Evident, v.3.4.1) in an automated fashion (at the laboratory temperature) with constant laser intensity at 100% and exposure times adjusted individually for each fluorophore specified in the QIBC plots. The automated image analysis was subsequently performed in ScanR analysis software (Evident, v.3.4.1). Automated dynamic background correction thresholding at least fivefold pixel intensity above background levels was applied for each fluorescent channel separately to maintain the same conditions within a single experiment. An intensity-threshold–based mask was generated based on the DAPI signal to identify individual nuclei as main objects. This mask was then applied to analyze pixel intensities in different channels for each nucleus. The multiparameter analysis was exported as a table and further analyzed in Spotfire software (TIBCO v.11.8.0, RRID:SCR_008858). A similar number of cells was analyzed and compared within each experiment.

U2OS or MCF7 cells were seeded onto 6-well plates (400 cells per well) containing complete DMEM supplemented with increasing concentrations of verapamil or ligand 5. After 10 days, cells were fixed using 4% buffered formaldehyde for 20 min at room temperature and stained with 0.1% crystal violet diluted in 20% ethanol for 30 min at room temperature. The plates were washed and air-dried overnight before counting the colonies.

Statistical analysis was performed using GraphPad Prism 10.1.2 (RRID:SCR_002798). Experiments were not randomized, and no blinding was used during data analysis. No statistical methods were used to predetermine sample size. For all the experiments, the sample sizes were determined based on experimental experience, technical expertise, and currently accepted standards in the field of high-content imaging. The sample size, statistical tests, and number of replicates for all experiments are specified in the figure legends. In Fig. 3 H, P values were calculated using a two-tailed unpaired t test; in Fig. S5, ordinary one-way ANOVA followed by Šídák’s multiple comparisons test was used for pairwise comparisons. All statistical tests were parametric, assuming normal distribution and equal variance; however, these assumptions were not formally tested.

Fig. S1 shows the MCM4-Halo cell line validation and labeling kinetics of commercial biotin-HaloTag ligands. Fig. S2 shows the ranking of modelled HaloTag–ligand–streptavidin ternary complexes and the efficiency of ligand 5 in vivo labeling. Fig. S3 shows the labeling kinetics of commercial ligand 1 in comparison to ligand 5. Fig. S4 shows the dynamics of MCM complexes during the cell cycle using ligand 5 and streptavidin conjugated with A647 dye. Fig. S5 shows the labeling kinetics of commercial ligands in comparison to ligand 5 in a concentration- and time-dependent manner. Fig. S6 shows no impact of ligand 5 on cell cycle and cell growth. Fig. S7 shows the ligand 5–based labeling efficiency of H2B-Halo in HeLa cells. Data S1 provides a detailed description of the experimental procedure employed for the synthesis of individual biotin-HaloTag ligands, along with an analysis of their structural characteristics.

All the data are available in the main text or the supplementary materials. Any additional data or information in support of this study will be available from the corresponding authors upon request.

We sincerely thank Jiri Polasek, Tomas Pop, and Tomas Jendrulek for the technical maintenance of high-content imaging microscopes. We are also deeply grateful to Maj-Britt Rask and Claudia Lukas (University of Copenhagen, Copenhagen, Denmark) for their support and for generously providing essential reagents. The HeLa Kyoto cell line was a gift from Shuh Narumiya (Kyoto University, Kyoto, Japan). JFX554-HaloTag ligand was a kind gift from Luke Lavis (Janelia Research Campus, Ashburn, VA, USA). Special thanks to Claudia Lukas and Yimon Aye for their critical reading of the manuscript. We also express our gratitude to all members of the Sedlackova and Svenda laboratories for their stimulating discussions and insightful comments on the manuscript.

The research work in the Sedlackova laboratory was supported by the Czech Science Foundation Junior Star (grant no. 22-20303M), the European Union’s Horizon 2022 Widera Talent program (ERA grant agreement no. 101090292), the EMBO Installation Grant (grant no. IG-5689-2024), and the Jihomoravske centrum pro mezinarodni mobilitu project scholarship for foreign students. The research work in the Svenda laboratory was supported by the Bader Philanthropies and the National Infrastructure for Chemical Biology (CZ-OPENSCREEN, LM2023052). Vacha laboratory acknowledges funding from the project National Institute of Virology and Bacteriology (Program EXCELES, ID Project No. LX22NPO5103)—funded by the European Union - Next Generation EU.

Author contributions: A.K. Yadav: conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization, and writing—review and editing. A.S. Jadhav: data curation, formal analysis, investigation, methodology, validation, visualization, and writing—review and editing. P.M. Szczepanik: conceptualization, data curation, formal analysis, investigation, methodology, validation, and writing—review and editing. P. Fagherazzi: formal analysis, investigation, validation, and writing—review and editing. I. Kabelka: formal analysis, visualization, and writing—review and editing. R. Vácha: resources, supervision, and writing—review and editing. J. Svenda: conceptualization, funding acquisition, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing. H. Polasek-Sedlackova: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing.