Arginine shortage induces replication stress and confers genotoxic resistance by inhibiting histone H4 translation and promoting PCNA ubiquitination

SUMMARY

The arginine dependency of cancer cells creates metabolic vulnerability. In this study, we examine the impact of arginine availability on DNA replication and genotoxicity resistance. Using DNA combing assays, we find that limiting extracellular arginine results in the arrest of cancer cells at S phase and a slowing or stalling of DNA replication. The translation of new histone H4 is arginine dependent and influences DNA replication. Increased proliferating cell nuclear antigen (PCNA) occupancy and helicase-like transcription factor (HLTF)-catalyzed PCNA K63-linked polyubiquitination protect arginine-starved cells from DNA damage. Arginine-deprived cancer cells display tolerance to genotoxicity in a PCNA K63-linked polyubiquitination-dependent manner. Our findings highlight the crucial role of extracellular arginine in nutrient-regulated DNA replication and provide potential avenues for the development of cancer treatments.

Graphical abstract

In brief

Wang et al. find that arginine shortage can temporarily place DNA replication on hold via inhibiting histone H4 translation and stalling fork movement. They show that this is associated with HLTF-mediated K63-linked polyubiquitination of PCNA, which increases PCNA’s presence on nascent DNA strands, and helps cells tolerate DNA damage.

INTRODUCTION

A major hallmark of tumor progression is unchecked DNA replication and adaptation to the high metabolic demands of rapid cell division in a nutritionally deprived microenvironment.^1,2 Arginine, a semi-essential amino acid, is a key nutrient for arginine auxotrophic tumors.³ These tumors become reliant on external sources of arginine due to widespread urea cycle dysregulation.^4,5 External arginine shortages in the tumor core can result from poor vascularization and dysfunctional blood flow in the tumor microenvironment (TME).^6–8 Furthermore, limiting dietary arginine or use of arginine-deprivation agents leads to the inhibition of arginine auxotroph tumor growth.^5,9–12 Moreover, the spatiotemporal regulation of DNA replication during S phase of the cell cycle is synchronized with nutrient availability to cells.^13–15 However, whether extracellular arginine shortage represents a widespread obstacle to DNA replication and how DNA replication contends with insufficient extracellular arginine availability remain largely unclear.

During DNA replication, nucleosomes are disrupted as the replication forks pass, followed by reassembly of recycled parental histones and incorporation of newly synthesized histones to restore chromatin state.¹⁶ As such, the synthesis of replication-dependent histones, essential building blocks for chromatin reassembly behind the replication fork, is tightly coupled to ongoing DNA synthesis.¹⁷ Indeed, histone biosynthesis is S-phase specific to govern S-phase timing and cell-cycle progression.^18–20 Given that arginine is a proteinogenic amino acid, it is clearly important to understand the role of arginine depletion on replication-dependent histone biosynthesis.

DNA replication stress comprises a multitude of cellular conditions, including physical blockage of DNA replication fork progression along the template, deregulation of the replication initiation or elongation complexes, or deoxyribonucleotide depletion.²¹ Proliferating cell nuclear antigen (PCNA) plays a vital role in coordinating DNA replication activity and maintaining genome integrity.^22,23 PCNA serves as a hub for proteins involved in DNA synthesis, cell-cycle control, and DNA damage response and repair.^24–27 PCNA is also subjected to multiple post-translational modifications, including ubiquitination, contributing to the coordination of DNA replication and genotoxic damage tolerance.^28,29 Whereas PCNA monoubiquitination in response to DNA damage promotes the bypass of replication-blocking lesions,³⁰ PCNA polyubiquitination has been associated with fork reversal.³¹ In particular, the mammalian RAD5 ortholog helicase-like transcription factor (HLTF), a ubiquitin ligase for PCNA, has also been shown to be required for replication fork reversal.^32,33 However, mechanisms by which arginine levels affect PCNA levels on replicating chromatin and the factors involved in this process remain elusive. In addition, it is currently unclear how arginine availability impacts DNA replication stress response and adaptation to genotoxic insult.

Here, we report that extracellular arginine shortage suppresses DNA replication elongation by reducing newly synthesized histone H4. Transient transfection of recombinant histone H4 protein can partially rescue DNA replication in arginine-starved cells. We present evidence that both HLTF and PCNA play key roles in fork protection in the absence of arginine. Arginine shortage increases HLTF-mediated PCNA K63-linked polyubiquitination and PCNA occupancy on the stalled nascent DNA strands. Next, we find that under replication stress induced by hydroxyurea (HU), arginine-starved cells exhibit vigorous nascent strand degradation, when HLTF is depleted, in a manner dependent on DNA2. Finally, HLTF depletion or treatment with PCNA ligand AOH996 results in a higher DNA damage signaling and sensitivity to genotoxicity under arginine shortage. Our results highlight a previously undescribed replication stress that arginine shortage induces defects on de novo chromatin assembly to interfere with DNA replication and S-phase progression. On the basis of these results, we conclude that extracellular arginine shortage triggers replication stress responses and confers tumor escape from a variety of genotoxicities.

RESULTS

Extracellular arginine regulates DNA replication fork progression

To determine the fate of cells that experience extracellular arginine shortage during cell-cycle progression, asynchronously cycling arginine auxotrophic MDA-MB-231 breast cancer cells were pulse labeled with bromodeoxyuridine (BrdU; t = 0 h), an analog of thymidine, and then chased in medium ± arginine for an additional 10 h. Using flow cytometry, we found that in the presence of arginine, most BrdU-labeled MDA-MB-231 cells were not at S phase (Figure 1A, bottom middle panel), while cells in arginine-free media were retained at S phase (Figure 1A, bottom right panel). Multiple cell lines, irrespective of arginine auxotrophy, displayed a similar impediment to S-phase exit in response to extracellular arginine shortage (Figure 1A, right panel). The overall cell-cycle distribution of G1, S, and G2/M were analyzed, and increased S-phase and decreased M-phase populations were found in arginine-deprived cells (Figure S1A). To determine if this stalled S-phase phenomenon occurs within the S phase, we synchronized MDA-MB-231 cells in early S phase using a double thymidine block. BrdU-labeled synchronized cells did not exit S phase when deprived of arginine (Figure 1B). Further, hallmarks of cell-cycle progression such as the upregulation of cyclin A, cyclin B1, phosphorylated H3S10, and H4 were absent in cells released in arginine-free medium, further confirming the inhibitory effect of arginine shortage on S-phase progression (Figure S1B). This observed S-phase blockade induced by arginine removal was reversible when 2.6–5.3 mg/L arginine was added to the medium (Figure S1C). In parallel, the overall incorporation of a 5-ethynyl-2’-deoxyuridine (EdU) analog in breast cancer-derived BT-549 spheroids decreased in the absence of arginine in a time-dependent manner (Figure S1D). We conclude that arginine is required for the cell-cycle progression of S-phase cells.

Figure 1. Essential role of arginine in DNA replication.

(A) Flow cytometry analysis of cell-cycle progression after BrdU labeling in the presence and absence of arginine. Top left: experimental design. Bottom left: analysis of BrdU-positive MDA-MB-231 cells in G1 phase (purple boxes). Right: percentage of BrdU-positive G1 cells in various cancer cell lines, quantified from three independent experiments with 50,000 events analyzed per experiment.

(B) Flow cytometry analysis of S-phase progression in double thymidine-synchronized MDA-MB-231 cells pulse-labeled with BrdU under ±arginine. Top: experimental design. Bottom: percentage of BrdU-positive S-phase cells at different time periods after thymidine block release, quantified from three independent experiments with 50,000 events analyzed per experiment.

(C and D) DNA fiber analysis using CldU (red) or IdU (green) antibodies in MDA-MB-231 (C) and BT-549 (D) cells under ±arginine. Top left: experimental design. Bottom left: representative images of DNA fibers. Middle: histograms shows percentage frequency of IdU track length. Right: quantitation of IdU track length, elongation ratio (IdU/CldU) and relative new replication origin firing. Each dot represents one fiber or one field with at least 100 tracts, or 40 fields were analyzed per sample. Scale bars, 10 μm.

(E) DNA combing analysis in MDA-MB-231 cells deprived of arginine for varying time frames and then resupplemented with arginine (84 mg/L, the same amount of arginine in full medium). Left: experimental design. Right: quantitation of CldU and IdU track lengths, with at least 100 tracts analyzed per sample.

Mean ± SEM is shown; ns: not significant; ***p < 0.005 (two-tailed unpaired Student’s t test). R, arginine.

We hypothesized that extracellular arginine is required for DNA replication fork progression. To address this possibility, we studied replication fork dynamics in cells at single-molecule resolution under ±arginine conditions. First, MDA-MB-231 and BT-549 cells were pulse labeled with a thymidine analog, 5-chloro-2’-deoxyuridine (CldU), for 40 min to establish the baseline tract lengths. After washing, cells were concomitantly labeled with the second thymidine analog iododeoxyuridine (IdU) under ±arginine conditions for 40 min (Figures 1C and 1D, top left diagrams). The track length of ongoing replication forks (with adjacent CldU [first labeling, red] and IdU [second labeling, green] signals) was examined and quantified (Figures 1C and 1D, bottom left panels). Quantitative analyses confirmed that the frequency of IdU-labeled tracts with reduced lengths increased in both cell lines upon arginine shortage, but this was not due to the new initiation of DNA replication (Figures 1C and 1D, right four panels).

Next, we determined whether the duration of arginine shortage affects DNA replication fork resumption after arginine is resupplied. Cells that were deprived of arginine for up to 120 min were first pulse labeled with CldU in arginine-free medium for an additional 40 min (i.e., up to 160 min) and subsequently pulse labeled with IdU in arginine-containing medium for 40 min (Figure 1E, left diagram). Notably, we found a time-dependent shortening of CldU-labeled tracts up to 100 min of arginine shortage (Figure 1E, middle panel). However, arginine replenishment restored DNA fork progression, visualized with IdU-labeled tracks and track length, despite varying arginine-free exposure times (Figure 1E, right panel). This indicates that transient extracellular arginine shortage is not a catastrophic DNA replication stress event, and replication forks are primed to resume upon arginine readdition. Together, we propose that arginine shortage holds DNA replication forks in a ‘‘standby’’ mode until the arginine shortage is resolved, and concomitantly, arginine shortage decelerates DNA elongation in an immediate, yet reversible, manner.

Extracellular arginine controls DNA replication by promoting the synthesis of histone H4 protein

Since arginine is a proteogenic amino acid and the biosynthesis of core histone proteins are tightly coupled to the cell cycle and peaks specifically during S phase, concomitant with DNA replication,^18,34 we surveyed whether arginine promotes DNA replication by regulating translation. We adopted the AHA (L-azidohomoalaine; a methionine analog) incorporation assay to assess the direct impact of arginine shortage on newly synthesized proteins in S-phase cells (Figure S2A). While those Coomassie blue-stained signals were only subtly affected when cells were released in arginine-free medium (Figure S2B, left two panels), AHA-labeled proteins were markedly reduced upon arginine shortage (Figure S2B, right two panels). To determine the effect of arginine availability on newly synthesized histone H2B, H3, and H4, AHA-labeled proteins were isolated using a biotin pull-down strategy followed by western analyses. As shown in Figure 2A, histone H4, unlike H2B and H3, was not synthesized upon arginine shortage. In stark contrast, AHA-labeled ATF4 was induced, serving as an arginine shortage-induced endoplasmic reticulum (ER) stress control.⁵ To rule out the possibility that arginine shortage also reduced histone H4 mRNA abundance, qRT-PCR analysis was performed. We found that H4 mRNA abundance began to rise at 30 min after release from the double thymidine block, irrespective of arginine availability, validating that cells are in S phase (Figure S2C). However, H4 mRNA level subtly leveled off between 4 and 8 h post-release in full medium, rendering it significantly lower than H4 mRNA in arginine-deprived cells (Figure S2C). These results suggest that arginine shortage leads to a preferential depletion of newly synthesized histone H4 via a translational control.

Figure 2. The transfected recombinant histone H4 protein rescues DNA elongation during arginine shortage.

(A) Analysis of nascent AHA-labeled histones in S-phase-enriched MDA-MB-231 cells. Top: experimental design. Middle: representative western blot. Bottom: densitometric quantification of relative histone and ATF4 levels. The relative protein levels were calculated by dividing the intensity of the target band with that in cells released to +arginine medium for 60 min (set as 1). n = 3 independent experiments.

(B) DNA combing assay using MDA-MB-231 cells transfected with His-tagged histone H4 under ±arginine. Top: experimental design. Middle: non-linear regression curve analysis for IdU track length. Bottom: quantification of IdU track lengths and IdU/CldU ratio. At least 200 tracts were analyzed per sample.

Mean ± SEM is shown; ns: not significant; ***p < 0.005 (two-tailed unpaired Student’s t test).

Next, we tested whether increased newly synthesized H4 rescues DNA replication in arginine-deprived cells. To this end, we used a protein transfection reagent to introduce recombinant human His-tagged histone H4 proteins into cells, modeling an increase in newly synthesized H4, prior to arginine removal (Figure S2D, left panel). After confirming the incorporation of recombinant histone H4 protein into chromatin (Figure S2D, right panel), DNA combing assays showed that transfected recombinant H4 protein, at least in part, rescued both IdU-labeled track lengths and IdU/CldU ratio in arginine-depleted cells (Figure 2B). Together, our data suggest that while histone H4 is not the only protein involved in DNA replication whose synthesis is selectively suppressed by arginine depletion, the transfected recombinant histone H4 is able to partially phenocopy extracellular arginine to promote DNA replication. This also explains why extracellular arginine shortage prevents S-phase exit (Figures 1A and 1B).

Newly synthesized histone H4 marks are surrogate markers of arginine availability within the TME

To search for a marker for arginine availability and/or DNA replication speed, we exposed cells to various arginine concentrations for different durations and assessed histone H4 marks via western blot analysis. Both H4K5ac and H4K12ac marks, which are enriched at newly assembled chromatin,¹⁹ decreased in MDA-MB-231 (Figure 3A) and BT-549 cells (Figure S3A) when exposed to suboptimal extracellular levels of arginine. Multiple cell lines displayed similar results upon 2 h of extracellular arginine shortage, regardless of their de novo arginine biosynthesis capacity via ASS1 (Figure S3B).³⁵ Because newly translated histone H4 is acetylated on lysines 5 and 12 in the cytoplasm prior to nuclear translocation, accumulation, and nucleosome deposition,^36–38 we assessed the steady-state levels of H4K5ac and H4K12ac in the cytoplasmic and nuclear subfractions of cells released into ±arginine medium for 4 h post-double thymidine block. The lack of nuclear lamin A/C and cell surface CD44 signals in the cytoplasmic and nuclear fractions, respectively, confirmed the relative purity of each fraction. Following arginine shortage, the overall levels of lamin A/C, CD44, histone H3, and actin remained unchanged, and the nuclear accumulation of ATF4 served as an arginine shortage control (Figures 3B and S3C).⁵ Consistently, a universal decrease of newly synthesized histone H4 marks in total cell lysates and cytoplasmic and nuclear fractions was observed with arginine shortage (Figures 3B and S3C). This arginine removal-mediated decline of histone marks was more notable in H4K5ac and H4K12ac (~50%) but not in H4K16ac, H3K9me3, and H3K4me3 levels (Figure S3D). To support that the arginine shortage contributes to the reduction of newly synthesized histone H4 marks, supplementation of arginine rapidly restored H4K5ac and H4K12ac signals (Figure S3E), correlating well with DNA replication resumption (Figure 1E).

Figure 3. The newly synthesized histone H4 marks show a strong correlation with arginine availability.

(A) Analysis of H4K5ac and H4K12ac in MDA-MB-231 cells grown under varying arginine concentrations. Left: representative western blot. Right: densitometric quantification of relative levels of H4K5ac and H4K12ac normalized to control (84 mg/L) from three independent experiments.

(B) Subcellular fraction analysis of H4K5ac and H4K12ac levels in synchronized S-phase MDA-MB-231 cells grown under ±arginine. Left: representative western blot. Right: densitometric quantification of relative levels of H4K5ac and H4K12ac normalized to control (+arginine) from three independent experiments.

(C) H4K12ac and H4K5ac expression in MDA-MB-231 xenograft tumors. Top: representative IHC image of periphery and non-necrotic core. The corresponding enlarged view of inset area (black square) shows staining for H4K12ac and H4K5ac (brown). Dashed line indicate the boundary between periphery and non-necrotic core. Bottom: quantification of the intensities of H4K12ac-positive and H4K5ac-positive MDA-MB-231 cells from representative IHC images (n = 12) and the quantification of the respective percentage of cells with indicated H4K5ac or H4K12ac intensity. Insets in images show +++ strong, ++ medium, and + weak positive staining. Scale bars, 200 μm (50 μm in zoom images).

(D) Representative IHC images of CD31 (purple) and H4K12ac (brown) expression in MDA-MB-231 and BT-549 xenograft tumors. An overview of full tumor section is shown (top). A corresponding enlarged view of boxed area is shown below for the periphery and the non-necrotic core. Scale bars, 2.5 mm, 50 μm (20 μm in zoom images).

(E) Representative immunofluorescence image of a BT-549 tumor spheroid cryosection stained for H4K12ac (green), EdU (red), and Hoechst 33342 (blue). Scale bars, 100 μm (50 μm, 20 μm and 10 μm in zoom images).

Mean ± SEM is shown; ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.005; two-tailed unpaired Student’s t test. le, long exposure; R, arginine.

A typical feature of solid tumors is their heterogeneous distribution of blood vessels and flow.³⁹ Consequently, suboptimal nutrient availability develops within the avascularized TME. Previous studies have shown that arginine is one of the least available amino acids in the tumor core regions.⁶ To determine how DNA replication is perturbed by the lack of arginine supply in vivo, we sought to determine whether the distribution of H4K5ac and H4K12ac marks are located together with cell proliferation marker Ki67 in breast cancer MDA-MB-231 and BT-549 xenograft tumors using immunohistochemical (IHC) staining. Quantitative analyses confirmed, in general, the gradual decrease of H4K5ac and H4K12ac signals within the non-necrotic tumor core compared with the peripheral regions of MDA-MB-231 (Figure 3C) and BT-549 (Figure S4A) xenograft tumors. In addition, CD31-positive cells, an endothelial cell marker, and H4K12ac-positive cells were detected in close proximity at the periphery and non-necrotic core regions (Figure 3D), and H4K12ac-negative cells were largely enriched in the CD31-negative regions (Figure S4B). In addition, S-phase marker Ki67 was also enriched at the peripheral region, correlating well with H4K12ac (Figure S4C) in MDA-MB-231 and BT-549 xenografts. Consistent with these observations, most EdU-labeled cells were localized to the periphery of BT-549 spheroids (Figure 3E). Likewise, most H4K12ac-positive cells were found at the periphery of Hs578T and MCF-7 tumor spheroids (Figure S4D). To test the uniqueness of arginine shortage, we deprived other nutrients, such as glutamine, which is limited in tumor core regions, and serum, as well as leucine, which are indispensable for cell proliferation. Results showed that neither removal of serum, glutamine, or leucine affected H4K5ac and H4K12ac signals (Figures S4E–S4G).^6,40 Collectively, these data suggest that extracellular arginine availability preferentially affects H4K5ac and H4K12ac abundance in vitro and in vivo.

Extracellular arginine shortage creates genotoxic tolerance

The maintenance of genome stability relies on successful DNA replication and proper DNA damage response in S phase. Next, we hypothesize that arginine availability might impact the DNA replication stress response and/or cell recovery from genotoxicity. When replication fork stress or DNA damage occurs, single-stranded DNA (ssDNA) is created and elicits the DNA damage response.^21,41 RPA coating of ssDNA results in the activation of the ATR/Chk1 pathway to promote replication fork stability and genome stability.^21,41 To test our hypothesis, we monitored the replication stress response via the phosphorylation of RPA (p-RPA(S33)) and Chk1 (p-Chk1(S345)) in arginine-starved cells. We used HU and camptothecin (CPT) as positive controls; both are known agents that cause DNA replication stress and activate the ATR/Chk1 pathway (Figure 4A).^21,41 Arginine shortage (2 h) failed to induce p-Chk1(S345) and phosphorylation of RPA in multiple cell lines (Figures 4A and S5A). Consistent with this, we failed to observe an increase in the chromatin-bound RPA levels, a reflection of the degree of ssDNA exposure, in pre-extracted S-phase cells via flow cytometry post-arginine shortage (4 h) (Figure 4B). This is in contrast to the observations following HU and CPT treatment.⁴² The inability of extracellular arginine shortage to activate Chk1 phosphorylation and RPA chromatin binding suggests that arginine shortage did not lead to significant ssDNA accumulation while deaccelerating DNA replication.

Figure 4. Arginine shortage promotes resistance to genotoxic stress.

(A) One representative western blot for RPA (S33) and Chk1 (S345) phosphorylation in HU-treated MDA-MB-231 cells under ±arginine from three independent experiments.

(B and C) Quantification of chromatin-bound RPA-positive (B) and γH2AX-positive (C) EdU-labeled MDA-MB-231 cells grown under stated conditions for 4 h, as analyzed by flow cytometry. n = 3 independent experiments. At least 10,000 events were collected and analyzed.

(D) γH2AX expression analysis in BT-549 spheroids treated under stated conditions for 4 h. Top: representative fluorescent images of BT-549 spheroid cryosections stained with DAPI (blue) and γH2AX (red). Bottom: quantification of relative γH2AX intensity normalized to tumor area. n = 4 independent experiments. Scale bars, 100 μm.

(E) Analysis of relative cell recovery of MDA-MB-231 cells under stated conditions for 4 h, followed by recovery in regular full medium. Cell viability was analyzed at indicated time points.

Mean ± SEM is shown; two-tailed unpaired Student’s t test; ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.005. CPT, camptothecin; HU, hydroxyurea; R, arginine.

We next determined whether the DNA damage response to double-strand DNA breaks (DSBs) was impacted under ±arginine conditions by assessing the histone variant H2AX phosphorylation on S139 (γH2AX) in MDA-MB-231 cells.⁴³ To induce DSBs, we used (1) CPT alone and (2) HU, a recognized source of DNA damage by misincorporation, in combination with either Chk1 inhibitor (Chk1i; rabusertib) or an ATR inhibitor (ATRi; BAY1895344).^44–47 Chk1i (1 μM) alone, but not ATRi (100 nM) alone, induced p-Chk1(S345) without a marked γH2AX accumulation (Figure S5B, lane #7 vs. #5), and co-treatment of ATRi (100 nM) blocked HU-induced p-Chk1(S345) but induced γH2AX (Figure S5B, lane #9 vs. #3) when serving as a positive control.⁴⁸ As expected, arginine depletion reduced γH2AX chromatin binding in S-phase cells treated with genotoxic agents as analyzed by flow cytometry (Figure 4C). Notably, arginine shortage attenuated ATRi- or Chk1i-sensitized γH2AX accumulation in HU-treated MDA-MB-231 cells (Figure S5B, lane #9 vs. #10 and lane #11 vs. #12) and γH2AX signals in tumor spheroids treated with CPT (Figure 4D). Taken together, cells exposed to arginine shortage appeared to exhibit reduced DNA replication stress responses than cells maintained in full medium upon treatment with the same genotoxic agents.

To determine the impact of arginine on cell recovery from genotoxic treatments, MDA-MB-231 cells were subjected to CPT (2 μM), HU (2 mM)+ATRi (100 μM), or HU (2 mM)+Chk1i (1 μM) under ±arginine conditions for 4 h and then cultured in drug-free full medium for the indicated time periods, with or without S-phase synchronization (Figure 4E). Similar results were found in CPT-treated BT-549 and MCF-7 cells (Figure S5C). There was a significant gain in resistance to genotoxic agents tested in arginine-starved cells at day 3 post-treatment. A similar result was observed in tumor spheroids (Figure S5D). Collectively, unlike other DNA replication stalling chemicals that induce DNA damage responses, arginine shortage slows down DNA replication without causing DNA damage responses while enabling tolerance to genotoxic insults.

Distinct protein dynamics at arginine shortage-stressed replication forks

We next sought to identify arginine shortage-induced and ATR-independent signaling events that mediate DNA replication slowing in MDA-MB-231 cells using the accelerated isolation of proteins on nascent DNA (aniPOND) method.⁴⁹ Due to the generation of shorter EdU-incorporated nascent DNA strands under arginine removal (Figures 1C and 1D), we expected less H4K5ac, H4K12ac, H2B, H3, and H4 to be pulled down from EdU-labeled cells exposed to arginine shortage compared with cells grown in full medium (Figure 5A). Furthermore, arginine replenishment following shortage restored these histone levels to similar levels as those observed in cells grown in full medium (Figure 5A). Decreased levels of H4K5ac and H4K12ac marks showed that there was less newly synthesized H4 occupying nascent DNA strands during DNA replication under arginine shortage (Figure 5A). In stark contrast to reduced histone levels, we found that more PCNA was bound to EdU-labeled DNA after arginine removal (Figure 5B), and this did not occur with arginine resupply following arginine shortage (Figure 5B). As previously reported,^43,50 CPT and HU treatment did not increase nascent DNA-bound PCNA levels (Figure 5B) when serving as controls. Furthermore, chromatin fractionation experiments revealed that a higher level of chromatin-bound PCNA was exclusively detected in arginine-deprived cells, and both CPT and HU treatments failed to affect occupancy of chromatin-bound PCNA under ±arginine conditions (Figure S6A). These results suggested that there was increased occupancy of PCNA on nascent DNA strands at stalled replication forks upon arginine withdrawal.

Figure 5. Arginine shortage-dependent regulation of PCNA and HLTF.

(A and B) Representative western blots of fork-associated proteins isolated using aniPOND method from EdU-positive MDA-MB-231 cells grown under stated conditions. EdU (−), negative control. n = 3 independent experiments.

(C–F) Western blot analysis of immunoprecipitated PCNA-associated proteins from Myc-ubiquitin-transfected FLAG-PCNA-overexpressed HEK293T cells subjected to indicated treatments. Left: representative western blots are shown for the indicated antibodies. Right: densitometric quantification of relative levels of indicated proteins normalized to control (+arginine) are shown. n = 3 independent experiments.

Mean ± SEM is shown; two-tailed unpaired Student’s t test; ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.005. R, arginine; le, long exposure; −R + R, arginine depletion followed by resupply.

During DNA replication, PCNA is repeatedly loaded and unloaded by the replicating clamp loader, replication factor C (RFC) complex, and an alternative PCNA ring opener, the ATAD5 (ELG1 in yeast)-RFC-like complex (ATAD5-RLC). ATAD5-RLC unloads replication-coupled PCNA.⁵¹ Maintaining the balance of PCNA loaded onto nascent DNA strands is crucial for DNA replication.⁵² To further test the effect of arginine shortage on chromatin-bound PCNA levels, we performed chromatin fractionation to determine the chromatin-associated fraction of subunits of the PCNA unloader, the RLC (ATAD5/RFC2–5) complex.⁵¹ Consistent with the increased PCNA levels, chromatin-associated ATAD5, RFC2, and RFC5, subunits of the RLC complex, were significantly reduced in arginine-deprived MDA-MB-231 cells. In addition, the total level of these proteins as shown in whole-cell extracts was also reduced upon arginine removal (Figure S6B). A similar observation was made in arginine-deprived BT-549 cells, except that the ATAD5 level was almost undetectable (Figure S6B). While consistent with the reduced H4K5Ac and H4K12Ac deposition, a marked reduction in chromatin-bound BRD4 (Figure S6B), an acetyl-histone-binding chromatin reader, seems to conflict with increased PCNA loading. A previous study showed that ATAD5, the major PCNA unloader, had dominant effects on PCNA unloading, and overexpression of BRD4 significantly induced PCNA accumulation on chromatin. However, depletion of BRD4 can only slightly decrease the levels of PCNA on nascent DNA.⁵³ Therefore, even though the downregulation of BRD4 should promote PCNA unloading during arginine shortage, the decreased levels of ATAD5, RFC2, and RFC5 could eventually lead to insufficient unloading of PCNA. Together, these findings suggest that the reduced chromatin binding of the PCNA-unloader ATAD5-RLC causes the increased retention of PCNA in the arginine-deprived cells. It remains to be elucidated how arginine removal affects the cross-talk among BRD4, ATAD5, and H4K5ac/H4K12ac to fine-tune PCNA retention on nascent DNA strands.

Given that altered ubiquitination of replication machinery is a hallmark of stressed replication forks,⁵⁴ we investigated if elevated levels of PCNA might correspond to altered PCNA ubiquitination. PCNA can be either monoubiquitinated or polyubiquitinated to activate translesion synthesis or fork reversal.^30,31 To test whether PCNA is ubiquitinated upon arginine removal in a manner similar to that observed under canonical replication stress, HEK-293T cells stably expressing FLAG-tagged PCNA were transiently transfected with Myc-tagged ubiquitin in our biochemical analyses. Following arginine removal (4 h) and immunoprecipitation with an anti-FLAG antibody, increased K63-linked polyubiquitination and K164 monoubiquitination of PCNA were detected (Figure 5C). HLTF is one of the E3 ligases that catalyzes PCNA K63-linked polyubiquitination in a RAD6-RAD18-dependent manner.^32,33 Immunoprecipitation followed by western analyses further revealed that arginine removal enhanced the recruitment of RAD6, RAD18, and HLTF to PCNA (Figure 5D). The replenishment of arginine reversed arginine shortage-increased K63-linked polyubiquitination and recruitment of RAD6, RAD18, and HLTF to PCNA without affecting PCNA levels (Figure 5E). Lastly, knockdown of HLTF abolished arginine shortage-induced K63-linked polyubiquitination of PCNA (Figure 5F). Next, we performed a chromatin fractionation and found the levels of chromatin-bound PCNA, HLTF, RAD18, and RAD6 were all significantly higher in arginine-deprived cells, indirectly supporting that increased associations among RAD6, RAD18, HLTF, and PCNA on chromatin resulted in more PCNA K63-linked polyubiquitination at forks (Figure S6A). Based on these observations, we propose that HLTF-mediated K63-linked polyubiquitination of PCNA and/or its associated proteins increases in response to extracellular arginine shortage in a reversible manner. To examine the possibility whether other K63-linked E3 ligases, such as SHPRH,⁵⁵ are involved in arginine shortage-induced PCNA K63-linked polyubiquitination, SHPRH was knocked down in arginine-deprived cells (Figure S6C). However, depletion of SHPRH failed to reverse arginine shortage-induced K63-linked polyubiquitination. These findings demonstrate that shortage of arginine induces HLTF- , but not SHPRH-, mediated K63-linked polyubiquitination of PCNA. To examine whether there is a link between histone H4 synthesis and PCNA K63-linked polyubiquitination, we impaired histone H4 biosynthesis by knocking down FLASH or SLBP, both of which are essential for histone biosynthesis (Figure S6D).²⁰ Indeed, knockdown of FLASH or SLBP decreased newly synthesized H4 marks, as previously reported, and increased K63-linked polyubiquitination of PCNA (Figure S6D). This observation links arginine shortage-suppressed de novo histone H4 synthesis to HLTF-mediated PCNA K63-ubiquitination.

Arginine shortage promotes HLTF-dependent genotoxin adaption

To mechanistically explore the role of HLTF-mediated K63-linked polyubiquitination of PCNA and/or its associated proteins in response to arginine shortage, we first sought to analyze how HLTF loss altered cellular responses to arginine removal. Previously, HLTF loss reportedly led to an unrestrained fork progression.^32,33 To confirm and extend this finding in arginine-starved cells, we used two different small interfering RNAs (siRNAs) to knock down HLTF and monitored fork progression under ±arginine in MDA-MB-231 cells as described in Figure 1C. Compared with siRNA control (siCtrl)-treated cells, knockdown of HLTF alone did not markedly affect the level of H4K5ac and H4K12ac marks (Figure S7A) and the ratio of IdU/CldU track lengths (Figure 6A) in unstarved cells, respectively. However, knockdown of HLTF partially rescued the arginine shortage-suppressed IdU/CldU ratio (Figure 6A). To provide additional evidence supporting the effect of PCNA ubiquitination on fork slowing in arginine-starved cells, isogenic paired HEK-293T wild-type (WT) and HEK-293T-K164R cells were employed.⁵⁶ Figure 6B shows that the PCNA K164R mutation, at least in part, rescued the IdU/CldU ratio in arginine-starved cells, and inactivating HLTF in the K164R cells had no further effect. Consistent with the idea that HLTF has an activity of slowing fork progression in cells experiencing replication stress,³³ our observation confirmed that HLTF restrains replication fork progression in arginine-starved cells.

Figure 6. HLTF and PCNA protects genome stability in arginine-starved cells.

(A) Analysis following MDA-MB-231 cells treated with indicated siRNAs targeting HLTF and ±arginine. Top left: experimental outline. Bottom left: representative western blot analysis of HLTF. Right: quantitation of IdU/CldU ratio. n = 3 independent experiments. Each dot represents one fiber. At least 100 tracts were analyzed per sample.

(B) Analysis of HEK-239T cells overexpressing wild-type (WT) or mutated (K164R) PCNA treated with siRNA targeting HLTF ±arginine. Top left: experimental outline. Bottom left: representative western blot analysis of K164 ubiquitinated PCNA (ub-PCNA) and HLTF. Right: quantitation of IdU/CldU ratio. n = 3 independent experiments. Each dot represents one fiber. At least 100 tracts were analyzed per sample.

(C) Analysis of the effect of DNA2 inhibitor C5 or MRE11 inhibitor mirin on replication fork stability under ±arginine and exposure to HU in HLTF-knockdown MDA-MB-231 cells. Top: experimental outline. Bottom: quantification of IdU/CldU ratio. n = 3 independent experiments. Each dot represents one fiber. At least 100 tracts were analyzed per sample.

(D) Analysis of HLTF knockdown on chromatin-bound RPA and γH2AX in MDA-MB-231 cells grown under ±arginine and subjected to genotoxic insult. Top: experimental outline. Bottom: quantification of chromatin-bound RPA-positive or γH2AX-positive cells following flow analysis. At least 10,000 events were collected and analyzed. n = 3 independent experiments.

(E) Relative cell recovery of control or HLTF-knockdown MDA-MB-231 cells after treatment with indicated genotoxic agents and ±arginine. n = 3 independent experiments.

(F) Relative cell recovery of WT or mutated (K164R) PCNA-expressing HEK-293T cells after treatment with indicated genotoxic agents and ±arginine. n = 3 independent experiments.

Mean ± SEM is shown; ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.005; two-tailed unpaired Student’s t test. R, arginine.

Next, we sought to address the contribution of HLTF to the integrity of arginine shortage-stressed replication forks. In order to do this, we treated HLTF-depleted MDA-MB-231 cells with a high concentration of HU (4 mM), which has been shown to promote fork reversal,⁵⁷ and measured the IdU/CldU-labeled fiber ratio under ±arginine conditions (Figure 6C). siRNA depletion of HLTF alone had no effect on fork degradation in cells cultured with HU (4 mM) and arginine. In contrast, the removal of arginine and exposure to HU (4 mM) notably accelerated fork degradation in HLTF-depleted cells (Figure 6C, lane #4 vs. #3). Collectively, these findings suggest that HLTF is critical to protect nascent DNA from degradation in arginine-deprived cells when exposed to HU (4 mM). Mirin, an MRE11 inhibitor, did not suppress nascent strand degradation in arginine-starved, HLTF-depleted cells (Figure 6C). In contrast, treatment with C5, an inhibitor of the nuclease DNA2, restored the nascent track length (Figure 6C), indicating the possible involvement of DNA2-mediated nascent DNA nucleolysis in arginine-starved, HLTF-depleted cells. To determine whether the protective function of HLTF relies on its Hiran domain to bind specifically to the 3′ end of ssDNA as previously reported, we tested the fork integrity in cells overexpressing siRNA-resistant WT HLTF and Hiran-mutated (R71E) HLTF.⁵⁸ Figure S7B shows that knockdown of endogenous HLTF in MDA-MB-231 cells induced the degradation of replication fork; however, both WT and Hiran-mutated HLTF can almost fully restore nascent DNA strands (Figure S7B), indicating that HLTF exerts its protective role in arginine-depleted cells independent of its ssDNA binding ability. To further investigate whether other known translocases share the protective function of HLTF, the fork integrity was tested in SMARCAL1- or ZRANB3-knockdown cells (Figure S7C).^59,60 Results showed that knockdown of SMARCAL1 or ZRANB3 did not induce degradation of replication fork in arginine-starved cells (Figure S7C). These observations indicated that HLTF-dependent fork reversal is not involved in resolving arginine shortage-induced fork stalling.

To determine the impact of HLTF on DNA damage signaling upon arginine shortage, we then used a flow cytometry-based method to quantitatively monitor RPA and phospho-H2AX or γH2AX chromatin binding in S-phase cells (Figure 6D, top diagram).⁴² As expected, brief exposure to a high concentration of CPT (2 μM, 4 h) increased both RPA and γH2AX chromatin binding in MDA-MB-231 S-phase cells cultured in full medium, and arginine shortage attenuated both signals (Figure 6D, bottom panels). Notably, HLTF loss increased both RPA and γH2AX chromatin binding in EdU-positive cells under arginine shortage (Figure 6D, bottom panels, lanes #10 and #12 vs. #8). A similar trend was noted (Figure 6D, bottom panels) in cells treated with HU+ATRi (lanes #16 and #18 vs. #14) or HU+CHK1i (lanes #22 and #24 vs. #20) under arginine shortage. Lastly, we determined the role of HLTF in protecting MDA-MB-231 cells from genotoxins under ±arginine conditions. Compared with Figure 4E, HLTF knockdown mitigated the advantage conferred by arginine removal on cell recovery from indicated genotoxic treatments (Figure 6E). Similar observations, albeit less pronounced, were made in PCNA-mutated (K164R) HEK-293T cells (Figure 6F). While PCNA K63-linked polyubiquitination is likely associated with poor recovery of HLTF-knockdown cells from CPT, HU+ATRi, or HU+Chk1i, the exact contribution to the better recovery remains unclear. It is possible that HLTF-catalyzed PCNA K164 ubiquitination (Figure 5F) is indispensable for arginine shortage to enable cells to better cope with DNA damage induced by a variety of genotoxic agents.

AOH1996 is the lead compound targeting the L126-Y133 region of PCNA to interfere with its interaction with partners.⁶¹ To address PCNA function in further detail, we determined the effect of AOH1996 on the fork integrity in arginine-deprived cells. ani-POND assays show that AOH1996 suppressed PCNA accumulation on replication forks (Figure S8A) without affecting the levels of newly synthesized H4 marks (Figure S8B). Like HLTF knockdown, a fork protection assay showed that AOH1996 accelerated fork degradation in arginine-deprived cells (Figure S8C, lane #4 vs. #3). Next, to address whether AOH1996 could reverse arginine shortage-induced K63-linked polyubiquitination of PCNA and genotoxin resistance, immunoprecipitation, a flow cytometry-based assay, and a cell recovery assay were performed.⁴² We showed that AOH1996 reversed arginine shortage-induced K63-linked polyubiquitination of PCNA (Figure S8D, lane #5 vs. #3). AOH1996 also increased γH2AX accumulation on chromatin in arginine-starved, EdU-positive cells treated with genotoxins (Figure S8E). Additionally, AOH1996 abolished the protection provided by arginine shortage on cells recovered from genotoxic insult (Figure S8F). Taken together, we suggest that PCNA loading and its ubiquitination are important for maintaining fork integrity and recovery from genotoxicity in arginine-deprived cells. Collectively, we propose a model, based on our findings, that arginine shortage inhibits H4 translation to stall DNA replication and enables HLTF-mediated K63-linked polyubiquitination of PCNA and/or its associated proteins and hyper-PCNA accumulation on nascent DNA to protect arginine-depleted cells against genotoxins (Figure S9).

DISCUSSION

We have discovered that arginine, a nutrient that is often lacking in TME,⁶ plays a crucial role in modulating DNA replication. Our findings indicate that a shortage of arginine significantly impairs the progression of replication forks by inhibiting the translation of histone H4 and leading to an excessive buildup and HLTF-mediated ubiquitination of PCNA on newly synthesized DNA strands. This establishes a clear connection between arginine shortage and the development of genotoxic adaptation.

Replication-dependent histone biosynthesis has long been considered as a crucial factor in regulating DNA replication and cell-cycle progression.^18–20 Our studies find that increased levels of histone H4 alone can promote DNA replication elongation under conditions of arginine shortage. Despite being only a moderately abundant amino acid, arginine plays an important role in histone H4 synthesis. This is due to the fact that histone H4 contains a higher percentage of arginine codons (CGC and CGU) compared with other histones, making it more sensitive to changes in arginine availability (Tables 1, 2, and 3). The observed stalling of DNA replication in response to arginine shortage is thought to be regulated by the selective loss of arginine tRNA charging and ribosome pausing at specific arginine codons. Darnell et al. reported that the selective loss of arginine tRNA charging during arginine shortage regulates translation via ribosome pausing at specific arginine codons.⁶² Altogether, our observations raise the possibility that arginine residues in histone H4 may carry out a nutrient-sensing function by changing its synthesis in response to extracellular arginine availability. This is an important area of investigation, as it may shed light on the mechanisms by which cells respond to changes in nutrient availability.

Table 1.

The mRNA and amino acids composition of Homo sapiens histone H4 (H4C1 H4)

mRNA	AUG	UCU	GGA	CGU	GGU	AAG	GGC	GGG	AAG	GGU	UUG	GGU	AAG	GGG	GGU	GCC	AAG	CGC	CAC	CGC
Protein	MET	SER	GLY	ARG	GLY	LYS	GLY	GLY	LYS	GLY	LEU	GLY	LYS	GLY	GLY	ALA	LYS	ARG	HIS	ARG
	AAG	GUG	UUG	CGU	GAC	AAC	AUC	CAG	GGC	AUC	ACC	AAG	CCG	GCC	AUC	CGG	CGU	CUG	GCC	CGG
	LYS	VAL	LEU	ARG	ASP	ASN	ILE	GLN	GLY	ILE	THR	LYS	PRO	ALA	ILE	ARG	ARG	LEU	ALA	ARG
	CGU	GGC	GGU	GUG	AAG	CGG	AUC	UCU	GGU	CUG	AUC	UAC	GAG	GAG	ACU	CGC	GGG	GUG	CUC	AAG
	ARG	GLY	GLY	VAL	LYS	ARG	ILE	SER	GLY	LEU	ILE	TYR	GLU	GLU	THR	ARG	GLY	VAL	LEU	LYS
	GUG	UUU	UUG	GAG	AAC	GUG	AUC	CGU	GAC	GCU	GUC	ACC	UAU	ACG	GAG	CAC	GCC	AAG	CGC	AAG
	VAL	PHE	LEU	GLU	ASN	VAL	ILE	ARG	ASP	ALA	VAL	THR	TYR	THR	GLU	HIS	ALA	LYS	ARG	LYS
	ACA	GUC	ACU	GCC	AUG	GAC	GUG	GUC	UAC	GCG	CUU	AAG	CGC	CAG	GGA	CGC	ACC	CUU	UAU	GGC
	THR	VAL	THR	ALA	MET	ASP	VAL	VAL	TYR	ALA	LEU	LYS	ARG	GLN	GLY	ARG	THR	LEU	TYR	GLY
	UUU	GGC	GGU	UAA	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
	PHE	GLY	GLY	STOP	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–

Ribosome pause-site arginine codons: CGU, CGC; non-ribosome pause-site arginine codons: CGG, CGA, AGA, AGG. Arginine pausing codon/all codons: 11/104 (10.58%). CGU/all codons: 5/104 (CGU/arginine codon: 5/14; CUG/ribosome pausing arginine codon: 5/11). CGC/all codons: 6/104 (CGC/arginine codon 6/14; CGG/ribosome pausing arginine codon: 6/11).

Table 2.

The mRNA and amino acids composition of Homo sapiens histone H3 (H3C14)

mRNA	AUG	GCC	CGU	ACU	AAG	CAG	ACU	GCU	CGC	AAG	UCG	ACC	GGC	GGC	AAG	GCC	CCG	AGG	AAG	CAG
Protein	MET	ALA	ARG	THR	LYS	GLN	THR	ALA	ARG	LYS	SER	THR	GLY	GLY	LYS	ALA	PRO	ARG	LYS	GLN
	CUG	GCC	ACC	AAG	GCG	GCC	CGC	AAG	AGC	GCG	CCG	GCC	ACG	GGC	GGG	GUG	AAG	AAG	CCG	CAC
	LEU	ALA	THR	LYS	ALA	ALA	ARG	LYS	SER	ALA	PRO	ALA	THR	GLY	GLY	VAL	LYS	LYS	PRO	HIS
	CGC	UAC	CGG	CCC	GGC	ACC	GUA	GCC	CUG	CGG	GAG	AUC	CGG	CGC	UAC	CAG	AAG	UCC	ACG	GAG
	ARG	TYR	ARG	PRO	GLY	THR	VAL	ALA	LEU	ARG	GLU	ILE	ARG	ARG	TYR	GLN	LYS	SER	THR	GLU
	CUG	CUG	AUC	CGC	AAG	CUG	CCC	UUC	CAG	CGG	CUG	GUA	CGC	GAG	AUC	GCG	CAG	GAC	UUU	AAG
	LEU	LEU	ILE	ARG	LYS	LEU	PRO	PHE	GLN	ARG	LEU	VAL	ARG	GLU	ILE	ALA	GLN	ASP	PHE	LYS
	ACG	GAC	CUG	CGC	UUC	CAG	AGC	UCG	GCC	GUG	AUG	GCG	CUG	CAG	GAG	GCC	AGC	GAG	GCC	UAC
	THR	ASP	LEU	ARG	PHE	GLN	SER	SER	ALA	VAL	MET	ALA	LEU	GLN	GLU	ALA	SER	GLU	ALA	TYR
	CUG	GUG	GGG	CUG	UUC	GAA	GAC	ACG	AAC	CUG	UGC	GCC	AUC	CAC	GCC	AAG	CGC	GUG	ACC	AUU
	LEU	VAL	GLY	LEU	PHE	GLU	ASP	THR	ASN	LEU	CYS	ALA	ILE	HIS	ALA	LYS	ARG	VAL	THR	ILE
	AUG	CCC	AAG	GAC	AUC	CAG	CUG	GCC	CGC	CGC	AUC	CGU	GGA	GAG	CGG	GCU	UAA	–	–	–
	MET	PRO	LYS	ASP	ILE	GLN	LEU	ALA	ARG	ARG	ILE	ARG	GLY	GLU	ARG	ALA	STOP	–	–	–

Ribosome pause-site arginine codons: CGU, CGC; non-ribosome pause-site arginine codons: CGG, CGA, AGA, AGG. Arginine pausing codon/all codons: 12/137 (8.76%). CGU/all codons: 2/137 (CGU/arginine codon: 2/18; CUG/ribosome pausing arginine codon: 2/12). CGC/all codons: 10/137 (CGC/arginine codon 10/18; CGG/ribosome pausing arginine codon: 10/12).

Table 3.

The mRNA and amino acids composition of Homo sapiens histone H2 (H2BC21)

mRNA	AUG	CCU	GAA	CCG	GCA	AAA	UCC	GCU	CCG	GCC	CCU	AAA	AAG	GGC	UCC	AAG	AAA	GCC	GUC	ACC
Protein	MET	PRO	GLU	PRO	ALA	LYS	SER	ALA	PRO	ALA	PRO	LYS	LYS	GLY	SER	LYS	LYS	ALA	VAL	THR
	AAA	GCC	CAG	AAG	AAA	GAC	GGC	AAG	AAG	CGC	AAG	CGC	AGC	CGC	AAA	GAG	AGC	UAC	UCC	AUC
	LYS	ALA	GLN	LYS	LYS	ASP	GLY	LYS	LYS	ARG	LYS	ARG	SER	ARG	LYS	GLU	SER	TYR	SER	ILE
	UAC	GUG	UAC	AAG	GUG	CUG	AAG	CAG	GUC	CAC	CCC	GAC	ACC	GGC	AUC	UCG	UCC	AAG	GCC	AUG
	TYR	VAL	TYR	LYS	VAL	LEU	LYS	GLN	VAL	HIS	PRO	ASP	THR	GLY	ILE	SER	SER	LYS	ALA	MET
	GGC	AUC	AUG	AAC	UCC	UUC	GUC	AAC	GAC	AUC	UUC	GAG	CGC	AUC	GCG	GGA	GAG	GCU	UCC	CGC
	GLY	ILE	MET	ASN	SER	PHE	VAL	ASN	ASP	ILE	PHE	GLU	ARG	ILE	ALA	GLY	GLU	ALA	SER	ARG
	CUG	GCG	CAC	UAC	AAC	AAG	CGC	UCC	ACC	AUC	ACA	UCC	CGC	GAG	AUC	CAG	ACG	GCC	GUG	CGC
	LEU	ALA	HIS	TYR	ASN	LYS	ARG	SER	THR	ILE	THR	SER	ARG	GLU	ILE	GLN	THR	ALA	VAL	ARG
	CUG	CUG	CUG	CCC	GGC	GAG	CUG	GCC	AAG	CAC	GCC	GUG	UCC	GAG	GGC	ACC	AAG	GCG	GUC	ACC
	LEU	LEU	LEU	PRO	GLY	GLU	LEU	ALA	LYS	HIS	ALA	VAL	SER	GLU	GLY	THR	LYS	ALA	VAL	THR
	AAG	UAC	ACC	AGC	UCC	AAG	UGA	–	–	–	–	–	–	–	–	–	–	–	–	–
	LYS	TYR	THR	SER	SER	LYS	STOP	–	–	–	–	–	–	–	–	–	–	–	–	–

Ribosome pause-site arginine codons: CGU, CGC; non-ribosome pause-site arginine codons: CGG, CGA, AGA, AGG. Ribosome pause-site encoded arginine: ARG; non-ribosome pause-site encoded arginine: ARG. Arginine pausing codon/all codons: 8/127 (6.3%). CGU/all codons: 0/127 (CGU/arginine codon: 0/8; CUG/ribosome pausing arginine codon: 0/8). CGC/all codons: 8/127 (CGC/arginine codon 8/8; CGG/ribosome pausing arginine codon: 8/8).

In addition to the relationship between extracellular arginine and histone H4, our studies also indicate that arginine is crucial for maintaining DNA replication through its link with PCNA (Figure S9). It is not possible to exclude the involvement of other factors, such as chromatin assembly factor-1 (CAF-1) or other histone chaperones (i.e., FACT and ASF1) in connecting arginine-dependent H4 translation with PCNA’s K63-linked polyubiquitination.^63,64 Additionally, changes in arginine availability may impact the interaction between PCNA and other key components involved in DNA replication, such as DNA polymerases, RFC, and RFC-like complexes. These proteins work together to coordinate the recycling of histones and PCNA, ensuring the proper assembly of chromatin on newly replicated DNA.¹⁶ However, our results do not suggest the presence of large ssDNA tracts during arginine shortage, although DNA nicks cannot be ruled out, as PCNA interacts with DNA ligase I.⁵² Increased levels of PCNA K63-linked polyubiquitination in response to arginine deficiency have the potential to serve as a marker for DNA replication stress. Further studies are needed to better understand the role of arginine in regulating PCNA’s residence on nascent DNA strands and maintaining genome stability.

An open question concerns the belief that PCNA is the only protein that significantly accumulates on nascent DNA strands in arginine-deprived cells. However, it is possible that technical issues, such as the efficiency of cross-linking or the conditions used for chromatin purification, may have led to the loss of PCNA-associated factors that are loosely bound. These studies have led to a surprising conclusion: arginine shortage activates Rad6/Rad18- mediated PCNA K164 monoubiquitination and HLTF-catalyzed PCNA K63-linked polyubiquitination. Given that PCNA must be repeatedly recycled during replication to allow for continuous DNA replication,⁶⁵ it is intriguing to find that arginine plays a role in regulating the timely removal of PCNA from nascent DNA strands by controlling the levels of unloading machinery proteins. The exact role of HLTF-mediated K63-linked polyubiquitination of PCNA and associated proteins in stalling DNA replication remains to be determined. We surmise that the K63-linked E3 ligase activity of HLTF, but not SHPRH, is crucial in maintaining the association of PCNA with the nascent strands and potentially affecting its interactions with other proteins. Another possibility is that K63-linked polyubiquitination of PCNA on newly synthesized DNA strands restricts its sliding, leading to the accumulation of PCNA behind stalled replication forks in arginine-deficient cells. Further research is needed to clarify the regulation of HLTF activity and its relationship with the DNA replication machinery, including the role of K63-linked polyubiquitination in stalling fork progression and enabling cells to withstand genotoxins. Additionally, it will be interesting to examine the coordination between PCNA unloading and histone deposition during DNA replication in response to arginine shortage, as well as the regulation of PCNA activity by arginine to maintain genome stability.

The full effects of arginine shortage on genome stability remain unclear. While the observed S-phase arrest in arginine-deficient cells alone is not enough to account for their increased resistance to genotoxicity, arginine shortage does slow down DNA replication without triggering a damage response. Our study suggests that HLTF plays a role in genome protection in an arginine-dependent manner, potentially through its interactions with other proteins or its E3 ligase activity. As demonstrated by Peng et al., long-term treatment with HU (4 mM) can result in fork degradation in FANCJ-knockout cells, but HLTF depletion significantly reduces this degradation.⁶⁶ However, we observe that the absence of HLTF exacerbates HU (4 mM)-induced fork degradation in arginine-deficient cells. Furthermore, K63-linked polyubiquitination of PCNA blocks the loading of DNA2 onto replication forks.⁵⁶ This aligns with our findings that arginine deficiency leads to DNA2-dependent, but MRE11-independent, nucleolytic degradation of nascent DNA at stalled replication forks in HLTF-deficient cells (Figure 6C). Our results are consistent with previous studies, which have shown that HLTF confers resistance to DNA damaging agents and that its loss sensitizes arginine-starved cells.^67–70 HLTF is widely recognized for its role in post-replication repair, where it facilitates replication fork reversal to bypass blocks and restart replication in response to stress.⁷¹ Additionally, van Toorn et al. recently reported that HLTF is necessary for efficient nucleotide excision repair.⁷² Our study excluded the role of HLTF and other translocases in fork reversal as a mechanism for managing replication stress caused by arginine shortage (Figures S7B and S7C). Nevertheless, the exact mechanism by which HLTF-mediated K63-linked polyubiquitination of PCNA maintains genome stability independent of fork reversal in the face of arginine shortage remains uncertain and deserves further investigation.

In summary, we have uncovered a hitherto unidentified pathway, the histone H4 translation-arrest and HLTF-mediated PCNA K63-linked polyubiquitination pathway, which provides an immediate and reversible response to slow DNA replication when external arginine availability changes (Figure S9). Our results indicate that the amount of arginine produced endogenously by ASS1 is not enough to support DNA replication and that mechanisms regulating DNA replication, elongation, and preventing replication fork disruption may be more widespread than previously understood. The instant slowing of the replication fork in response to a shortage of extracellular arginine helps maintain genome stability against various agents and prevents harmful consequences in rapidly dividing cells. This arrest of replication also gives any existing lesions time to be repaired, allowing replication to resume once arginine becomes available again. Our findings of H4 translation inhibition and PCNA K63-linked polyubiquitination-mediated arrest preventing genotoxin-induced genomic instability highlight the significance of local nutrient availability in the TME and its impact on tumor cell heterogeneity. Our study emphasizes the protective effect of arginine shortage against genotoxicity, which may be a common phenomenon in the central regions of various types of solid tumors, regardless of their genetic, epigenetic (i.e., as histone and DNA modification), and metabolic characteristics. Additionally, the fact that administering arginine can make tumors more sensitive to radiation therapy supports the importance of considering arginine levels when developing therapeutic approaches and the potential impact on the response to anti-cancer therapy.⁷³

Limitations of the study

Our study sheds light on the role of extracellular arginine in regulating DNA replication through promoting histone H4 synthesis and facilitating PCNA recycling. Our findings are supported by the results of our histone H4 transfection experiment and the involvement of the HLTF-mediated PCNA K63-linked polyubiquitination pathway. However, a limitation of the study is that we were unable to determine the precise mechanism by which arginine shortage leads to the accumulation of PCNA on nascent strands. The results of our immunoprecipitation assay may not provide conclusive evidence for the promotion of K63-linked polyubiquitination of chromatin-bound PCNA in response to arginine shortage. Further research is needed to understand the role of ZRANB3 in this process and to clarify the transition of unmodified PCNA to the K63-linked polyubiquitinated state. Additionally, the impact of newly synthesized histone H3 in DNA replication during arginine shortage remains an area of future investigation.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, David Ann (dann@coh.org).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• All data reported in this paper will be shared by the lead contact upon request.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

Animal experiments were approved by the Institutional Animal Care and Use Committee at City of Hope. MDA-MB-231 or BT-549 cells (3.3 × 3 10⁵) were injected into the mammary fat pads of 6-week-old female NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) mice. All mice were assigned to the same experimental group. Xenograft tumors were formalin-fixed, paraffin embedded (FFPE) sections were prepared by City of Hope Pathology Core.

Cell lines

Human breast cancer MDA-MB-231, MDA-MB-468, MCF7, BT-549, Hs578T and embryonic kidney HEK-293T cells were purchased from American Type Culture Collection and cultured with Dulbecco’s Modified Eagle’s Medium (DMEM, Cat# 10–013-CV, Corning) containing fetal bovine serum (10%, Cat# 10437028, ThermoFisher) and 1X antibiotic and antimycotic solution (Cat# 30–004-CI, Corning).

METHOD DETAILS

Treatment

Arginine-free media for BT-549, HEK-293T, Hs578T, MCF7, MDA-MB-231 and MDA-MB-468 cells were prepared from DMEM for SILAC (Cat# 88364, ThermoFisher) supplemented with dialyzed FBS (10%, Cat# 26400–044, ThermoFisher), 1X antibiotic and antimycotic solution (Cat# 30–004-CI, Corning), and L-lysine (146.2 μg/ml, Cat# L-9037, Millipore Sigma). For arginine removal, cells were washed with warm arginine-free medium once then incubated in arginine-free medium for indicated times.

Lentivirus preparation

pMD2.G (1 μg; Addgene, Cat# 12259) and psPAX2 (2 μg; Addgene, Cat# 12260) were co-transfected with lentiviral expression construct (3 μg) using Lipofectamine 2000 (14 μL; Life Technologies, 11668–019) into HEK-293T cells to generate lentivirus. Medium was replaced at 24 h post transfection, and viral medium was collected at 72 h post transfection. Viral medium was filtrated with 0.45 μm filter (Millipore Sigma, Cat# SLHAM33SS) and concentrated with Ultra-15 centrifugal filter device (Millipore Sigma, Cat# UFC903024). Concentrated lentivirus-containing medium was recovered and used to transform desired cell lines in the presence of polybrene (10 μg/mL, Millipore Sigma, Cat# TR-1003-G). Clonal cell lines were selected by pre-determined concentration of puromycin (0.5, 1 and 2 μg/ml; Sigma, Cat# P8833) for respective cell line and overexpression of desired protein was confirmed by Western blotting.

Western blotting

Total cell lysates were prepared by lysing cells with sample buffer and analyzed by Western blotting. Following polyacrylamide gel electrophoresis, separated proteins were transferred to polyvinylidene difluoride (PVDF, Millipore) membranes. In general, the PVDF membranes with transferred proteins were blocked with nonfat milk (5%) in PBST (137 mM NaCl, 10 mM phosphate, 2.7 mM KCl, 0.1% Tween 20, pH 7.4) for 40 mins. Proteins of interest were identified with antibodies described in key resources table. Following incubation with primary antibodies (in PBST with non-fat milk (5%)) at 4 °C overnight, PVDF membranes were washed with PBST 3 times, and then incubated with horse radish peroxidase-linked anti-mouse or anti-rabbit secondary immunoglobulin antibodies (1:10000; Cat# 7076S, Cat# 7074S Cell signaling) in PBST for 40 mins, Blots were developed with homemade ECL (2.5 mM luminol, 044 mM p-coumaric acid, 0.1 M Tris-HCl pH 8.5, 0.1% hydrogen peroxide in 100 ml distilled water). ChemiDoc Touch imaging system (Bio-Rad) was used to visualize signals. The captured images were analyzed with Image Lab Software (Bio-Rad, version 5.2.1).

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-Ki67	abcam	Cat#ab15580; RRID:AB_443209
Rabbit monoclonal anti-CD31	Cell signaling	Cat#77699; RRID:AB_2722705
Mouse monoclonal anti-CD44	Cell Signaling	Cat#3570; RRID:AB_2076465
Rabbit polyclonal anti-acetyl-histone H4 (Lys5)	Millipore	Cat#07–327; RRID:AB_310523
Rabbit polyclonal anti-acetyl-histone H4 (Lys12)	Millipore	Cat#07–595; RRID:AB_310740
Rabbit polyclonal anti-acetyl-histone H4 (Lys16)	Millipore	Cat#07–329; RRID:AB_310525
Mouse monoclonal anti-Actin	Millipore	Cat#MAB1501; RRID:AB_2223041
Rabbit monoclonal anti-ATF4	Cell Signaling	Cat#11815; RRID:AB_2616025
Mouse monoclonal anti-Chk1	Cell Signaling	Cat#2345; RRID:AB_10693648
Rabbit polyclonal anti-cyclin A	Santa Cruz	Cat#sc-751; RRID:AB_631329
Mouse monoclonal anti-cyclin B1	Enzo	Cat#ADI-KAm-CC195; RRID:AB_10615085
Rabbit polyclonal anti-eIF2α	Cell Signaling	Cat#9722; RRID:AB_2230924
Rabbit polyclonal anti-γH2AX	Bethyl	Cat#A300–081; RRID:AB_203288
Mouse monoclonal anti-histone H3	Active Motif	Cat#39763; RRID:AB_2650522
Rabbit monoclonal anti-histone H4	abcam	Cat#ab177840; RRID:AB_2650469
Mouse monoclonal anti-lamin A/C	Cell Signaling	Cat#4777; RRID:AB_10545756
Rabbit polyclonal anti-BRCA2	abcam	Cat#ab27976; RRID:AB_2067760
Rabbit polyclonal anti-PCNA	Santa Cruz	Cat#sc-7907; RRID:AB_2160375
Rabbit polyclonal anti-phospho-eIF2α	Cell Signaling	Cat#9721; RRID:AB_330951
Rabbit monoclonal anti-phospho-Chk1 (Ser296)	abcam	Cat#ab79758; RRID:AB_2244917
Rabbit monoclonal anti-phospho-Chk1 (Ser345)	Cell Signaling	Cat#2348; RRID:AB_331212
Rabbit polyclonal anti-phospho-histone H3 (Ser10)	Millipore Sigma	Cat#06–570; RRID:AB_310177
Mouse monoclonal anti-RPA32/RPA2	abcam	Cat#ab2175; RRID:AB_302873
Rabbit polyclonal anti-phospho-RPA32 (Ser33)	Bethyl	Cat#A300–246A; RRID:AB_2180847
Rabbit monoclonal anti-trimethyl-histone H3 (Lys4)	Cell Signaling	Cat#9751; RRID:AB_2616028
Mouse monoclonal anti-trimethyl-histone H3 (Lys9)	Active Motif	Cat#61013; RRID:AB_2687870
Rabbit monoclonal anti-ubiquityl-PCNA (Lys164)	Cell Signaling	Cat#13439; RRID:AB_2798219
Rabbit monoclonal anti-ubiquitin Lys63-specific	Cell Signaling	Cat#5621; RRID:AB_10827985
Rabbit polyclonal anti-SMARCA3 (HLTF)	Bethyl	Cat# A300–230A; RRID:AB_2117307
Rabbit polyclonal anti-Rad6	abcam	Cat#ab31917; RRID:AB_777604
Rabbit polyclonal anti-Rad18	abcam	Cat#ab188235; RRID:AB_2935813
Rabbit polyclonal anti-RFC5	Proteintech	Cat#10385; RRID:AB_2178603
Rabbit polyclonal anti-RFC2	Proteintech	Cat#10410; RRID:AB_2285035
Rabbit polyclonal anti-BRD4	Proteintech	Cat#28486; RRID:AB_2918170
Rabbit polyclonal anti-ATAD5	abcam	Cat#ab72111; RRID:AB_1209390
Mouse monoclonal anti-flag	Proteintech	Cat#66008–3-Ig; RRID:AB_2749837
Rabbit monoclonal anti-FLASH/CASP8AP2	Cell Signaling	Cat#43339; RRID:AB_2935814
Rabbit polyclonal anti-SLBP	Millipore Sigma	Cat#HPA019254; RRID:AB_1856951
Mouse monoclonal anti-SHPRH	Santa Cruz	Cat#sc-514395; RRID:AB_2935815
Mouse monoclonal anti-SMARCAL1	Santa Cruz	Cat#sc-376377; RRID:AB_10987841
Rabbit polyclonal anti-ZRANB3	Novus Biologicals	Cat#NBP2–93301; RRID:AB_2935816
Rabbit polyclonal anti-phospho-Akt (Ser473)	Cell Signaling	Cat#9271; RRID:AB_329825
Rabbit polyclonal anti-Akt	Cell Signaling	Cat#9272; RRID:AB_329827
Rat monoclonal anti-BrdU (BU1/75)	abcam	Cat#ab6326; RRID:AB_305426
Mouse monoclonal anti-BrdU (B44)	BD Biosciences	Cat#347580; RRID:AB_10015219
Chemicals, peptides, and recombinant proteins
L-azidohomoalanine	Thermo Fisher	Cat#C10102
L-arginine	Millipore Sigma	Cat# A8094
Thymidine 5’-monophosphate disodium salt hydrate	Millipore Sigma	Cat# T7004
Hydroxyurea	MP Biomedicals	Cat#102023
(+)-Camptothecin	ACROS Organics	Cat#AC276721000
BAY 1895344	Selleckchem	Cat#S8666
Rabusertib	Selleckchem	Cat#S2626
AOH1996	Gu et al.61	N/A
Recombinant human His-tagged histone H4 produced in E. coli	Active Motif	Cat#31493
Critical commercial assays
FiberPrep	Genomic Vision	Cat#EXT-001
Combicoverslips	Genomic Vision	Cat#COV-002-RUO
Experimental models: Cell lines
Human: MDA-MB-231	ATCC	HTB-26
Human: MDA-MB-468	ATCC	HTB-132
Human: MCF7	ATCC	HTB-22
Human: BT-549	ATCC	HTB-122
Human: Hs578T	ATCC	HTB-126
Human: HEK-293T	ATCC	CRL-11268
Human: HEK-293T-PCNA-WT	Thakar et al.56	N/A
Human: HEK-293T-PCNA-K164R	Thakar et al.56	N/A
Experimental models: Organisms/strains
Mouse:NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ(NSG)	The Jackson Laboratory	RRID: IMSR_JAX:005557
Oligonucleotides
siHLTF	Sigma Aldrich	Cat#SASI_Hs01_00162286 and SASI_Hs01_00029458
siFLASH	Sigma Aldrich	Cat#SASI_Hs02_00343647 and SASI_Hs01_00152963
siSLBP	Sigma Aldrich	Cat#SASI_Hs01_00147022 and SASI_Hs01_00147023
siSHPRH	Sigma Aldrich	Cat#SASI_Hs02_00310973 and SASI_Hs02_00310974
siSMARCAL1	Sigma Aldrich	Cat#SASI_Hs02_00328762 and SASI_Hs02_00328763
siZRANB3	Sigma Aldrich	Cat#SASI_Hs01_00109802 and SASI_Hs01_00109803
H4C5 forward primer 5’-GGTGTCAAGCGCATTTCTGGTC-3′	This paper	N/A
H4C5 reverse primer 5’-CGTAGACCACATCCATCGCTGT-3′	This paper	N/A
Recombinant DNA
pMD2.G	Addgene	Cat#12259
psPAX2	Addgene	Cat#12260
CMV-FLAG-SV40-eGFP-IRES-puromycin	GeneCopoeia	EX-NEG-Lv203
CMV-FLAG-PCNA-SV40-eGFP-IRES-puromycin	GeneCopoeia	EX-B0066-Lv203
pHAGE-HLTF-WT	Taglialatela et al.58	N/A
pHAGE-HLTF R71E Hiran mutant	Taglialatela et al.58	N/A
Software and algorithms
ImageJ	NIH	https://imagej.nih.gov/ij/
FCS Express	De Novo software	https://denovosoftware.com/
GraphPad Prism 7.0 software	GraphPad Prism Software Inc.	https://www.graphpad.com/
Image Lab Software 5.2.1	Bio-rad	https://www.bio-rad.com/en-us/product/image-lab-software?ID=KRE6P5E8Z

Subcellular fractionation

Equal number (3×10⁶) of cells were harvested in ice-cold nuclear extraction buffer (NEB) (400 μL, 10 mM HEPES pH 7.9, 10 mM KCl, 1.5 mM MgCl₂, 0.34 M sucrose, 10% glycerol, 1 mM dithiothreitol, 0.1% Triton X-100) containing protease/phosphatase cocktail inhibitors (Thermo Scientific, Cat# 78446) and incubated on ice for 5 min. 300 μl of whole cell lysates were centrifuged at 1300 × g for 4 min at 4°C. Supernatants were collected as cytoplasmic fractions and nuclei pellets were washed with NEB twice and re-suspended in 300 μl of NEB. Total cell lysates (remaining 100 μl of whole cell lysates), cytoplasmic and nuclear fractions were mixed with 4× Laemli buffer (v/v) and analyzed with desired antibodies by Western blotting.

Cell cycle analyses (flow cytometry)

For double thymidine block, cells were treated with thymidine (2.5 mM; Cat# 194754, MP Biomedicals) for 16 h and washed twice with PBS. Released cells were incubated in complete medium for 8 h and blocked with the 2^nd-thymidine treatment. Following 2^nd-thymidine block (16 h), cells were washed twice with PBS and released in regular medium for analyses. To enrich mitotic cells, cells were treated with thymidine (2.5 mM, 24 h) and released in regular medium. At 3 h post-releasing in regular medium, cells were treated with nocodazole (100 ng/ml, Cat# M1404, Millipore Sigma) for 12 h and released in regular medium for analysis. For BrdU pulse-labeling, cells were pulse-labeled with BrdU (10 μM, Cat# B9285, Millipore Sigma, 30 min) before or after treatment under indicated conditions. Cells were collected and fixed in ethanol (70%) at −20°C overnight. To analyze BrdU incorporated DNA, cells were washed with PBS to remove ethanol and denatured with HCl (2.5 M, 20 min). Cells were washed twice with PBS then stained with FITC conjugated anti-BrdU antibody (BioLegend, Cat# 364104) (1:100) at 37°C for 1 h. After incubation, cells were washed 3 times with PBS and DNA content were visualized by DAPI (1 μM; Cat# D9542, Millipore Sigma, I h) staining at room temperature. Samples were monitored by Attune NxT flow cytometer (Thermofisher) and results were analyzed and quantified by FCS Express (De Novo software). At least 50000 events were collected and 3 independent experiments were performed and analyzed.

DNA combing assay

Equal number (5×10⁴) of MDA-MB-231 or BT-549 cells were seeded in 60-mm dishes and incubated overnight. Cells were pulse-labeled with CldU (50 μM; Cat# 105478, MP Biomedicals), IdU (250 μM; Cat# 54–42-2, ACROS Organics) and treated as indicated. Cells were washed with PBS and collected. DNA plugs were generated using FiberPrep (Cat# EXT-001, Genomic Vision) and DNA combing was performed using Combicoverslips (Cat# COV-002-RUO, Genomic Vision), according to the manufacturer instructions. DNA combed coverslips were dehydrated with 70%, 90%, and 100% ethanol for 1 min each. DNA combed coverslips were denatured by 0.5 M NaOH plus 1 M NaCl solution for 8 min at room temperature then washed with PBS. The coverslips were dehydrated again with 70%, 90%, and 100% ethanol for 1 min each and air-dried at room temperature for 10 min. Following blocking with 5% BSA in PBS for 5 mins at 37°C in a humidity chamber, coverslips were incubated with primary antibodies (rat anti-BrdU (BU1/75) (1:500; recognizing CldU; Cat# NB500–169, Novus) and mouse anti-BrdU (B44) (1:500; recognizing ldU; Cat# 347580, BD) with 5% BSA in PBS) for 2 h at 37°C in a humidity chamber. Following incubation, coverslips were washed 3 times with PBST (0.05% Tween 20) then dehydrated with 70%, 90%, and 100% ethanol stepwise. Coverslips were incubated in PBS containing both goat anti-rat 647 secondary antibody (1:500; Cat# A-21247, ThermoFisher) and goat anti-mouse 488 secondary antibody (1:500; Cat# A-11001, ThermoFisher) for 1 h at 37°C in a humidity chamber. After incubation, coverslips were washed 3 times with PBST (0.05% Tween 20) then dehydrated with 70%, 90%, and 100% ethanol stepwise. Coverslips were mounted with prolong diamond antifade mountant (Cat# P36970, ThermoFisher), and sealed with nail polish. DNA replication tracts were visualized with a fluorescence microscope (Observer II, Zeiss) and ImageJ was used for measuring the length of individual labeled DNA tract.

Analysis of newly synthesized proteins with AHA labeling

Double thymidine synchronized MDA-MB-231 cells were released in L-Azidohomoalanine (AHA, 200 μM, Cat# C10102, ThermoFisher) containing full or arginine-free medium and cells were harvest in indicated time. Cells were collected and washed with PBS for 3 times. AHA-labeled newly synthesized proteins were analyzed by electrophoresis and detected with anti-biotin antibody (Cat# ab201341, abcam). For visualizing total proteins, whole cell lysates were analyzed by electrophoresis and gels were stained with Coomassie blue and imaged on ChemiDoc Touch imaging system (Bio-Rad). The captured images were analyzed with Image Lab Software (Bio-Rad, version 5.2.1).

To assess levels of specific AHA-labeled proteins, MDA-MB-231 cells were labeled with AHA as described above and lysed in RIPA buffer (1% sodium deoxycholate, 50 mM HEPES, 150 mM NaCl, 1% NP-40, 0.1% SDS, 2.5 mM MgCl₂, 10 mM sodium glycerophosphate, 10 mM sodium biphosphate pH 7.2) containing protease and phosphatase inhibitors (Thermo Scientific, Cat# 78446). Approximately 1.5 mg of proteins from each sample was reduced with TCEP (5 mM, room temperature for 10 min, Cat# 580561, Millipore Sigma) then alkylated with chloroacetamide (20 mM, room temperature for 15 min, Cat# C8625G, TCI America). Alkylated proteins were precipitated with methanol/chloroform/H₂O and washed 2 times with methanol to remove residual AHA. The precipitated proteins were suspended in suspension buffer (2% SDS, 50 mM HEPES, 150 mM NaCl, pH 7.2, 2.5 mM TCEP) and sonicated briefly with a probe sonicator. Click chemistry reaction was performed with biotin-PEG4-alkyne (100 μM, Cat# 764213, Millipore Sigma), CuSO4 (1 mM, Cat# 422870050, ACROS organics), and L-ascorbate (1 mM, Cat# 352680050, ACROS organics), THPTA ligand (100 μM, Cat# 762342, Millipore Sigma) and mixed on a rotor at room temperature for 2 h. Biotin-alkyne-reacted proteins were precipitated with methanol/chloroform/H₂O and washed 2 times with methanol to remove residual biotin-alkyne. The precipitated proteins were suspended in suspension buffer (200 μl, 2% SDS, 50 mM HEPES, 150 mM NaCl, pH 7.2, 2.5 mM TCEP) and sonicated briefly with a probe sonicator and diluted to 600 μl with RIPA buffer. High-capacity streptavidin agarose (20 μl, Cat# 20359, ThermoFisher Scientific) was added to each sample and mixed on a rotor at room temperature overnight. The beads were spun down and washed sequentially twice with RIPA buffer (1 ml), twice with 1M KCl, 0.1 M Na₂CO₃, 2 M urea in 50 mM HEPES buffer pH 7.5 (1 ml), twice with RIPA buffer (1 ml) and twice with ddH₂O (1 ml). After washing, samples were mixed with 2× sample buffer and boiled at 100°C for 15 min to extract the proteins. Elution was analyzed by Western blotting with indicated antibodies.

Protein transfection

Human recombinant proteins were introduced into cells with ProteoJuice protein transfection reagent (Cat# 71281–3, MilliporeSigma) according to manufacturer’s instruction. Briefly, recipient cells were seeded in a 6-well plate, washed with PBS for 3 times and maintained in OPTI-MEM. Approximately 5 μg of recombinant human His-tagged histone H4 produced in E. coli (Cat# 31493, Active Motif) was mixed with 5 ml of ProteoJuice in 100 μl of OPI-MEM and incubated at room temperature for 20 min. After 20 min of incubation, 900 μl of OPI-MEM was added to make a 1000 μl of mixture. After aspirating the OPTI-MEM from the cells, cells were incubated with the transfection mixture at 37°C (5% CO₂) for 2–3 h. After incubation, cells were washed 3 times with PBS and incubated in full or arginine-free medium as indicated prior to harvesting or subjected to DNA combing assays.

Immunoprecipitation

Stable expressing FLAG or FLAG-tagged PCNA HEK293T cells were overexpressed with Myc-tagged ubiquitin plasmid (gift from Dr. Hsiu-Ming Shih) for 24 h prior to the indicated treatments. After treatment, cells were harvested by modified RIPA buffer (50mM Tris-HCl pH7.5, 150 mM NaCl, 5 mM EDTA, 0.1% Triton X-100, and 0.1% NP-40) with protease/phosphatase inhibitor (Cat# 78446, ThermoFisher) and incubated in 4°C on rotating mixer for 15 min. Cell debris were removed by 12000 rpm centrifuge for 10 min in 4°C and supernatants were mixed with anti-DYKDDDDK magnetic agarose (Cat# A36797, ThermoFisher) in 4°C overnight on rotating mixer according to manufacturer’s protocol. Magnetic beads were captured with magnetic rack and washed with modified RIPA buffer for 3 times. Samples were eluted by adding 2X SDS-PAGE Sample Buffer and incubated at 100°C for 15 min. Levels of specific proteins were analyzed by Western blotting. To knockdown HLTF, cells were transfected with siRNA targeting HLTF 24 h prior to Myc-ubiquitin overexpression.

Accelerated native iPOND (aniPOND)

A rapid and accelerated method modified from isolation of protein on nascent DNA (iPOND) was performed following previous study.⁴⁹ In short, MDA-MB-231 cells were pulse labeled by EdU (10 μM) for indicated time along with indicated treatments. Cells were harvested and nuclear extracts were prepared by using the nuclear extraction buffer as described.⁴⁹ Click reaction was performed to conjugate biotin-azide (Cat# 1265, Click Chemistry Tools) to EdU labeled nascent DNA and sonicated (Bioruptor Pico sonication device). Samples were centrifuged and supernatant was collected and mixed with streptavidin-coated beads (Cat# 20359, ThermoFisher) to pull down biotin-EdU labeled nascent DNA. Beads were centrifuged and washed for 3 times and proteins were eluted by adding 2× sample buffer and boiled at 100˚C for 15 min. Western blotting was performed to analyze the proteins on nascent DNA. Cells without Edu labeling (EdU(−)) served as negative control. To compensate for the reduced EdU-labeled track in perturbed cells, the EdU labeling time of -R, HU, and CPT treated cells were extended to 25 min while +R cells were labeled for 15 min in Figure 5B.

Spheroid formation

Equal number (1.2×10³) of cells were seeded in ultralow attachment 96-well plates (Cat# CLS3474, Corning) and incubated for 72 h in a 37°C humidity chamber to form spheroids. Spheroids were washed with medium containing indicated chemical twice then incubated in medium containing same chemical for 2 h at 37°C. After 2 h of incubation, spheroids were washed with complete medium three times and then cultured in complete medium for 8 days. Spheroids were visualized at day 0 and day 8 post-treatment by using Cytation 5 imaging reader (BioTek) and ImageJ was used to process the areas of spheroids.

Cell viability assay

Equal number (5×10³) cells were seeded in 96-well plate and cultured under indicated conditions for different time periods. Acid phosphatase (ACP) assay was performed to measure cell viability. Cells were lysed with ACP1 solution (100 μl, 0.1 M sodium acetate, 0.1% Triton X-100, pH 5.7) containing p-Nitrophenyl Phosphate (12 mM) per well and kept in a 37°C incubator. Following 30 min of incubation, the reaction was stopped with NaOH (10 μl, 1 M) and absorbance at 410 nm was recorded using Synergy H1 hybrid multi-mode reader (BioTek). The value of each group measured at day 0 was set as 1.

Immunofluorescence staining

To visualize H4K12ac and EdU incorporation in BT-549 spheroids, equal number (8×10³) cells were seeded in ultralow attachment 96-well plates for 72 h. Spheroids were incubated in EdU (10 μM)- and Hoechest33342 (10 μM)-containing medium for 1 h, collected, fixed with paraformaldehyde (4%) for 10 min and washed twice with PBS. To generate blocks for cryosection, spheroids were incubated in sucrose (15%) for 30 min, followed by sucrose (30%) for 30 min, and sucrose (30%) containing optimal cutting temperature compound (50%, O.C.T., Tissue-TeK, Cat# 4583) for 30 min. Spheroids were then embedded in O.C.T. (100%) in cryomold and 8 μm thick sections were cryosectioned using a cryostat (Leica, CM3050S). EdU incorporation was stained according to manufacturer’s protocol (EdU kit for imaging, Cat# C10640, ThermoFisher). Following washing with PBS 3 times and blocking with BSA (4%) in 0.2% Tween-20 in PBS (PBST (0.2%)) for 30 min in a 37°C humidity chamber, slides were stained with an anti-H4K12ac antibody (1:500; Cat# 07–595, Millipore Sigma) in PBST with BSA (2%) slides in a 37°C humidity chamber for 2 h. Slides were washed 3 times with PBST (0.2%) then incubated with a goat anti-rabbit 488 secondary antibody (1:500; Cat# A-11034, ThermoFisher) in PBST with BSA (2%) in a 37°C humidity chamber. Slides were washed 3 times with PBST (0.2%) and mounted with prolong diamond antifade mountant (Cat# P36970, ThermoFisher), and then sealed with nail polish.

To assess H4K12ac signals in BT-549 spheroids, equal number (8×10³) cells were seeded in ultralow attachment 96-well plates and incubated with complete medium in a 37°C humidity chamber for 72 h to form spheroids. Cryosectioning and immunofluorescence staining were performed as described above. Briefly, cryosections were probed with an anti-H4K12ac antibody (1:500; Cat# 07–595, Millipore Sigma) and followed with incubation with a goat anti-rabbit 488 secondary antibody (1:500; Cat# A-11034, ThermoFisher). H4K12ac-positive cells were visualized using a fluorescence microscope (Observer II, Zeiss; LSM 700 Confocal microscope, Zeiss). ImageJ was used for processing the images in each cryosection. For visualizing histone H4K12ac signals in Hs578T or MCF7 spheroids, 3×10⁴ cells were seeded in ultralow attachment 96-well plates and incubated in complete medium to form spheroids in a 37°C humidity chamber. After 5 days of incubation, tumor spheroids were collected and fixed with paraformaldehyde (4%) for 10 min and washed twice with PBS. Cryosectioning and immunofluorescence staining were performed as described above.

To assess H4K12ac and γH2AX signals in BT-549 spheres, 8×10³ cells were seeded in ultralow attachment 96-well plates and incubated in complete medium for 72 h to form spheres. Cryosection and immunofluorescence staining were performed as described above. Briefly, sections were probed with an anti-H4K12ac antibody (1:500; Cat# 07–595, Millipore Sigma) or an anti-H2A.X phosphor (Ser139) antibody conjugated with Alexa Fluor 647 (1:500; Cat# 613408, BioLegend). For H4K12ac visualization, after primary antibody incubation, sections were incubated with a goat anti-rabbit 488 secondary antibody (1:500; Cat# A-11034, ThermoFisher). Samples were imaged by a fluorescence microscope (Observer II, Zeiss; LSM 700 Confocal microscope, Zeiss), and ImageJ was used for evaluating the overall intensity of indicated signals in each sphere.

Chromatin-bound RPA and γH2AX detection

MDA-MB-231 cells were pulse labeled with EdU (10 μM) for 1 h then treated with full, arginine-free, or medium containing HU (2 mM), CPT (2 μM), ATR inhibitor (100 nM) or Chk1 inhibitor (1 μM) for 4 h. Cells were trypsinized, washed with PBS and extracted with cold 0.2% Triton X-100 in PBS for 7 min on ice. After extraction, cells were washed with PBS and fixed with 4% paraformaldehyde in PBS (Cat# BM-155, Boston BioProducts). Fixed cells were blocked with 0.1% BSA in PBS and RPA signals were stained followed by DAPI staining. Percentage of RPA-positive or γH2AX-positive cells in S-phase (EdU-positive) was analyzed by Attune NxT flow cytometer (Thermofisher) and results were quantified by FCS Express (De Novo software).

Immunohistochemistry

Animal experiments were approved by the Institutional Animal Care and Use Committee at City of Hope. Xenograft tumors were formalin-fixed, paraffin embedded (FFPE) sections were prepared by City of Hope Pathology Core. For detecting the proteins of interest in xenograft tumors, anti-histone 4 lysine 12 acetylation (H4K12ac) antibody (1:500; Cat# ab177793, abcam), anti-histone 4 lysine 5 acetylation (H4K5ac) antibody (1:4000; Cat# ab232507, abcam), anti-Ki67 antibody (1:200, Cat# ab15580, abcam), and anti-CD31 antibody (1:100, Cat# 77699, Cell signaling), respectively, were used for immunohistochemical staining by the pathology core at City of Hope. IHC stain was performed on Ventana Discovery Ultra (Ventana Medical Systems, Roche Diagnostics, Indianapolis, USA) IHC automated stainer. Briefly, tissue samples were sectioned at a thickness of 5 μm and mounted on positively charged glass slides. The sections were deparaffinized and rehydrated. The endogenous peroxidase activity was blocked before antigen retrieval. The antigens were sequentially detected with primary antibody incubation and heat inactivation was performed to prevent antibody cross-reactivity between the same species. Following each primary antibody, DISCOVERY anti-Rabbit HQ and DISCOVERY anti-HQ-HRP were incubated. The positive signals were visualized with DISCOVERY ChromoMap DAB and DISCOVERY Purple Kit, respectively, counterstained with hematoxylin (Ventana) and then coverslipped. Slides were scanned by bright-field microscopy and 4 fields of peripheral and non-necrotic core regions were photograph in each section. H4K5ac and H4K12ac signal intensities were quantified by ImagePro Premier 9.0.

RNA extraction and qRT-PCR

Total RNA was extracted from cells using the Quick-RNA MiniPrep (Zymo Research, Cat# R1055). qRT-PCR was performed by as described previously.⁵ In brief, the RNA concentration and purity were analyzed by Synergy H1 hybrid multi-mode reader (BioTek). Complementary DNA (cDNA) was synthesized with 1 μg of total RNA by using the iScript cDNA Synthesis Kit (Bio-Rad, Cat# 1708891). cDNA was diluted 20 times and real-time quantitative PCR was performed by using iTaq Universal SYBR Green Supermix (BioRad, Cat# 1725122) with CFX-96 Real-Time PCR Detection System (BioRad). The primers used in this study targeting histone H4C5 are, forward: 5’-GGTGTCAAGCGCATTTCTGGTC-3’, and reverse: 5’-CGTAGACCACATCCATCGCTGT-3’. qRT-PCR data from 3 biological replicates were calculated using 2^−ΔΔCt method and normalized to 18S RNA.⁷⁴ The mean mRNA abundance in the first time point (0 h) was set as 1.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data are presented as mean and standard error of mean (Mean ± SEM). Statistical analyses were performed using GraphPad Prism 7.0 software (GraphPad Prism Software Inc., San Diego, CA). Normal distribution was confirmed using Shapiro-Wilk normality test before performing statistical analyses. For normally distributed data, comparison between two means were assessed by unpaired two-tailed Student’s t test and that between three or more groups were evaluated using one-way analysis of variance followed by Tukey’s post hoc test. A p-value of <0.05 was considered statistically significant.

Highlights.

• Arginine shortage halts H4 histone synthesis, stalling DNA replication

• Fork stalling creates vulnerability of nascent DNA strands to DNA2-mediated degradation

• Cells respond to arginine deprivation by HLTF-mediated K63-polyubiquitination of PCNA

• HLTF and PCNA polyubiquitylation protect arginine-starved cells from DNA damage agents