DONSON facilitates Cdc45 and GINS chromatin association and is essential for DNA replication initiation

Abstract Faithful cell division is the basis for the propagation of life and DNA replication must be precisely regulated. DNA replication stress is a prominent endogenous source of genome instability that not only leads to ageing, but also neuropathology and cancer development in humans. Specifically, the issues of how vertebrate cells select and activate origins of replication are of importance as, for example, insufficient origin firing leads to genomic instability and mutations in replication initiation factors lead to the rare human disease Meier-Gorlin syndrome. The mechanism of origin activation has been well characterised and reconstituted in yeast, however, an equal understanding of this process in higher eukaryotes is lacking. The firing of replication origins is driven by S-phase kinases (CDKs and DDK) and results in the activation of the replicative helicase and generation of two bi-directional replication forks. Our data, generated from cell-free Xenopus laevis egg extracts, show that DONSON is required for assembly of the active replicative helicase (CMG complex) at origins during replication initiation. DONSON has previously been shown to be essential during DNA replication, both in human cells and in Drosophila, but the mechanism of DONSON’s action was unknown. Here we show that DONSON’s presence is essential for replication initiation as it is required for Cdc45 and GINS association with Mcm2–7 complexes and helicase activation. To fulfil this role, DONSON interacts with the initiation factor, TopBP1, in a CDK-dependent manner. Following its initiation role, DONSON also forms a part of the replisome during the elongation stage of DNA replication. Mutations in DONSON have recently been shown to lead to the Meier-Gorlin syndrome; this novel replication initiation role of DONSON therefore provides the explanation for the phenotypes caused by DONSON mutations in patients.


INTRODUCTION
Faithful cell division is the basis for the propagation of life. The r eplication, r epair and epigenetic r egulation of the human genome are three of nine interrelated pathways involved in ageing ( 1 , 2 ) and defects in the maintenance of genome integrity are the foremost dri v ers of cellular deterioration ( 2 ). The mechanisms of origin firing during DNA replica tion initia tion and replica tion for k progression hav e been well characterised and reconstituted in yeast, howe v er, equal understanding of these processes in higher eukaryotes is lacking. Specifically, the issue of how vertebrates, including humans, select and activate origins of r eplication ar e not resolv ed conclusi v ely and are of importance as, for e xample, areas with a paucity of replication origins are more prone to genomic instability ( 3 ), while mutations in factors involved in origin firing lead to the rare human disease Meier-Gorlin syndrome ( 4 ).
DNA replication occurs in three stages: initiation, elongation and termination. DNA replication initiates from thousands of replication origins. These are the positions within the genome wher e r eplicati v e helicases become activated and start unwinding DN A w hile moving in opposite directions, away from each other, creating two DNA replication forks. The replicati v e helicase is composed of C dc45, the M cm2-7 hexamer and the G INS complex (CMG complex) ( 5 ); it is positioned at the tip of replication forks and f orms a platf orm f or replisome assembly ( 6 ). Activation of the CMG helicase (initiation stage) is divided into two cell cycle stages (Figure 1 A): loading of the inacti v e core of Mcm2-7 in the form of double hexamers in the G1 stage of the cell cycle (origin licensing) and subsequent activation of a proportion of the double hexamers by Cdc45 and GINS during S-phase (origin firing). Once established, the replication forks replicate chromatin (elongation stage) until they encounter forks, moving in an opposing direction, which initiated from neighboring origins (Figure 1 A). At this point, the termination of replication forks takes place.
Mutations in DONSON lead to microcephalic primordial dwarfism (MPD), which is a collecti v e term for a group of human disorders characterised by intra-uterine and postnatal growth delay alongside marked microcephaly ( 4 , 25 ). DONSON was shown to interact with the replisome and is proposed to protect stalled DNA replication forks from the action of nucleases ( 25 ). Cells with siRNA downregulated DONSON (siDONSON) exhibit accumulation of stalled and asymmetric DNA replication forks and generation of double-stranded DN A breaks, w hich can be partially rescued by co-depletion of nucleases that can act on stalled forks ( 25 ). Moreover, DONSON was shown to be important for proper activation of the S-phase checkpoint response ( 25 ). Intriguingly, DONSON was shown to interact with only a subset of replisomes --specifically ones that were activated in early S-phase rather than late S-phase ( 26 ). During early S-phase, DONSON was shown to help replica-tion forks traverse inter-strand crosslinks (ICL) ( 26 ). Interestingly, the Drosophila homologue of DONSON, humpty dumpty ( hd ), was also found to be essential for DNA replication, especially in early cleavage divisions of fly embryos, as hd knockout embryos arrested during the first few nuclear cleavage cycles with incomplete chromosome segregation ( 27 , 28 ). This suggests the evolutionary conservation of DONSON importance for cell proliferation in the de v eloping embryo.
Altogether, we know se v eral processes during DNA replica tion tha t DONSON is involved in and the consequences of DONSON's do wnregulation, ho we v er, the mechanisms by which DONSON deli v ers its functions and how they are r egulated ar e unknown. We have ther efor e set out to determine the molecular mechanism of DONSON's action during DNA replication in the Xenopus laevis egg extract system, which is the only higher eukaryotic cell-free system able to undertake a full round of cell-cycle regulated DNA replication. We have found that X.l. DONSON indeed interacts with replisomes during DNA replication but surprisingly, we found that DONSON is r equir ed for initiation of DNA r eplication -mor e specifically for the activation of replicati v e helicase by loading of GINS and Cdc45.

Inhibitors
MLN4924 (A01139, Acti v e Biochem) was dissolv ed in DMSO at 5 mM and added to the extract 15 min after addition of sperm chroma tin a t 10 M. NMS873 (17674, Cayman Chemical Company) was dissolved in DMSO at 10 mM and added to the extract 15 min after addition of sperm nuclei at 50 M. Caffeine (C8960, Sigma) was dissolved in water at 100 mM and added to the extract along with demembranated sperm chromatin at 5 mM. Aphidicolin was dissolved in DMSO at 8 mM and added to the extract along with demembranated sperm chromatin at 40 M. EcoR1 (R6011, Promega) was purchased at stock 12 U / l and added to the extract at 0.05 U / l. Mitomycin C (Calbiochem) was dissolved in water at 5 mM in water and used at 500 M. Etoposide (Calbiochem) was dissolved in DMSO at 10 mM and used at 200 M. Camptothecin was dissolved in water at 10 mM and used at 500 M. Cdc7i PHA-767491 was dissolved in water and added at 100 M.
Xenopus full-length DONSON protein was purified as described above and antibodies raised against such prepared antigens in sheep or rabbit. The resulting antibody sera was purified in-house against the purified antigen. The specificity of each new antibody is presented in Supplementary Figure S1.

Xenopus laevis egg extract preparation
All of the work with Xenopus laevis was approved by Animal Welfare and Ethical Re vie w Bod y (AWERB) a t Uni v ersity of Birmingham and approved by UK Home Office in form of Project License issued for Dr Agnieszka Gambus. Xenopus laevis egg extract was pr epar ed as pr eviously described ( 35 ).

Immunodepletion
DONSON immunodepletions with sheep antibodies were performed using Dynabeads Protein G (10004D, Life Technologies) coupled to antibodies against DONSON or nonspecific sheep IgGs (I5131, Sigma), with two rounds of 45 min incubation at 4 • C. The DONSON antibodies were coupled at 600 g per 1 ml of beads. Effecti v e immunodepletion r equir ed two rounds of 45 min incubation of egg extract with antibody coupled beads at 50% beads ratio.
DONSON immunodepletions with rabbit antibodies were performed using Dynabeads Protein A (10008D, Life Technologies) coupled to Xenopus DONSON antibodies r aised in r abbit and affinity purified or nonspecific rabbit IgG (I5006, Sigma). The DONSON antibodies were coupled at 600 g per 1 ml of beads. Effecti v e immunodepletion r equir ed two rounds of 45 min incubation of egg extract with antibody coupled beads at 50% beads ratio.
TopBP1 immunodepletions were performed using Dynabeads Protein G (10004D, Life Technologies) coupled to antibodies against TopBP1 or nonspecific sheep IgGs (I5131, Sigma). The TopBP1 antibodies were coupled at 600 g per 1 ml of beads. Effecti v e immunodepletion r equir ed two rounds of 45 min incubation of egg extract with antibody coupled beads at 50% beads ratio.

DNA synthesis assay
Interphase X. laevis egg extract was supplemented with 10 ng / l of demembranated sperm chromatin and incuba ted a t 23 • C for indica ted time. Synthesis of nascent DNA was then measured by quantification of ␣ 32 P-dATP (PerkinElmer) incorporation into newly synthesised DNA, as described before ( 35 ). The extract contains endogenous dNTP pools of ∼50 M ( 36 ). The total amount of DNA synthesized, expressed as ng DNA / l extract, can then be calculated by m ultipl ying percent total 32 P incorporated by a factor of 0.654 ( 36 ). This calculation assumes an average molecular weight of 327 Da for dNMPs and equal quantities of all four dNTPs incorporated into DNA (weight of dNMP incorporated in ng / l = percent total 32 P incorporated / 100 × 50 × 10 −6 × 4 × 327 × 10 3 ) ( 36 ).
For quantification of replication efficiency between different experimental repeats in different extracts, the quantity of DNA replicated at the end of the reaction in IgGdepleted extract was set as 100% and the remaining values normalised to this.
During the chroma tin isola tion procedure, a sample without addition of sperm DNA (no DNA) is processed in an analo gous way, usuall y at the end of the time course, to serve as a chromatin specificity control. The bottom of the PAGE gel on which the chromatin samples were resolved is cut off and stained with Colloidal Coomassie (SimplyBlue, Life Technologies) to stain histones which provide loading controls and indications of sample contamination with egg extract (cytoplasm).

Western blot quantification
The quantification of western blots is provided to indicate reproducibility of trends in experiments rather than to provide absolute values of increases or decreases in a signal. The quantified experiments were performed in different preparations of extracts and independently immunodepleted extracts to confirm that observed phenotypes are not specific for one extract prep.
The density of pixels of each band of the western blot and scanned Coomassie stained histones within the gel were quantified using Image J software. The numeric value in arbitrary units for each western blot band was normalised to appr opriate loading contr ol (bands of Coomassie stained histones from the same sample). The analysis of IgGdepleted (control) and treatment samples was always done together and the fold difference between them calculated. Fold change from three repeated experiments is plotted on the graph with mean value and standard error of the mean (SEM) as calculated by GraphPad PRISM.

Immunoprecipitation from chromatin
275 l of egg extract per IP was induced into interphase and mixed with 10-15 ng / l demembranated sperm nuclei and optionally supplemented with the indicated treatments. The reaction was incubated at 23 • C for the indicated time. Chromatin was isolated in ANIB100 (50 mM HEPES pH 7.6, 100 mM KOAc, 10 mM MgOAc, 2.5 mM Mg-ATP, 0.5 mM sper midine, 0.3 mM sper mine, 1 g / ml of each aprotinin, leupeptin and pepstatin, 25 mM ␤-glycerophosphate, 0.1 mM Na 3 VO 4 , 0.1% triton and 10 mM 2-chloroacetamide), and the chromatin pellets re-suspended in the same volume of original extract of ANIB100 containing 20% sucr ose. Pr otein complexes were released from chromatin by digestion with 2 U / l of Benzonase nuclease (E1014-25KU, Sigma) and sonicated for 5 min using a Diagenode sonicator with settings: 30 s on, 30 s off, medium setting. The insoluble fraction was then spun in a microfuge at 4 • C, 10 min, 16k g.
Pr epar ed beads: were incubated with 220 l digested chromatin at 4 • C for 1-2 hours with rotation. Following the incubation time, beads were washed for 5 min rota ting a t 4 • C twice with ANIB100, once with ANIB100 containing an additional 0.1% Triton X-100 and finally twice with ANIB100 buffer. Each sample was pr epar ed by boiling in 50 l of 2x Nu-PAGE LDS loading buffer (Life Technologies) for 5 min.

Immunoprecipitation from nucleoplasm
550 l of egg extract per IP was induced into interphase and mixed with 10-15 ng / l demembranated sperm chromatin and optionally supplemented with the indicated treatments. CDK inhibitor (p27 Kip1 ) and CDC7 inhibitor (PHA-767491) were added immediately after sperm DNA. The reaction was incubated at 23 • C for 45 min. Nucleoplasm was isolated in ANIB100 (50 mM HEPES pH 7.6, 100 mM KOAc, 10 mM MgOAc, 2.5 mM Mg-ATP, 0.5 mM sper midine, 0.3 mM sper mine, 1 g / ml of each aprotinin, leupeptin and pepstatin, 25 mM ␤-glycerophosphate and 10 mM 2-chloroacetamide), and the nuclear pellets re-suspended in the same volume of original extract of ANIB100 containing 20% sucrose and 0.1% triton. Protein complexes were released from nucleoplasm by digestion with 2 U / l of Benzonase nuclease (E1014-25KU, Sigma) and sonicated for 5 min using a Diagenode sonicator with settings: 30s on, 30s off, medium setting. The insoluble fraction was then spun in a microfuge at 4 • C, 10 min, 16 000 rcf. 100 l of Dynabeads Protein A (10008D, Life Technologies) covalently coupled to 20 g of affinity purified DON-SON or IgG from rabbit serum (I5006, Sigma) using BS3 crosslinker (S5799, Sigma); were incubated with 400 l digested nucleoplasm at 4 • C for 1-2 h with rotation. Following the incubation time, beads were washed for 5 min rota ting a t 4 • C twice with ANIB100, once with LFB1 / 50 and finally twice with ANIB100 buffer. Each sample was prepared by boiling in 50 l of 2 × NuPAGE LDS loading buffer (Life Technologies) for 5 min.

Large scale immunoprecipitation of DONSON for mass spectrometry and CHROMASS
3.9 ml of X. laevis egg extract was activated and supplemented with 10 ng / l of demembranated sperm DNA, 50 M p97 inhibitor NMS873 or 40 M aphidicolin + 5 mM caffeine and incubated at 23 • C for 60 min. Chroma tin was isola ted in ANIB / 100 buf fer. Immunoprecipitation of DONSON was performed as described previously for p97 ( 37 ) and the immunoprecipitated material was analysed by mass spectrometry with Dr Richard Jones from MS Bioworks LLC.
Similarly, chroma tin isola ted from 10 l of egg extract was processed in analogous way for mass spectrometry.

Sample pr epar ation
Each sample was run on a 5-20% gradient gel (Invitrogen) for 1 cm and cut into 10 bands. Samples were submitted pre-plated for ten fraction analysis. Gel pieces were processed using a r obot (Pr oGest, DigiLab) with the following protocol: • Washed with 25 mM ammonium bicarbonate followed by acetonitrile. • Reduced with 10 mM dithiothreitol at 60 • C followed by alkylation with 50 mM iodoacetamide at RT. • Digested with trypsin (Promega) at 37 • C for 4 h.
• Quenched with formic acid and the supernatant was analysed directly without further processing.

Mass spectrometry
The gel digests were analysed by nano LC / MS / MS with a Waters M-class HPLC system interfaced to a Ther-moFisher Oribitrap Fusion Lumos. Peptides were loaded on a trapping column and eluted over a 75 m analytical column at 350 nl / min; both columns were packed with Luna C18 resin (Phenomenex). A 30 min gradient was employed. The mass spectrometer was operated in datadependent mode, with MS and MS / MS performed in the Orbitrap at 60 000 resolution and 15 000 r esolution, r especti v el y. Ad vanced Peak Detection was turned on. The instrument was run with a 3 s cycle for MS and MS / MS. Proteome Discoverer v1.4 was used for peak generation. Peptide identifications were accepted if they could be established at greater than 34.0% probability to achie v e an FDR < 1.0% by the Percolator posterior error probability calculation ( 38 ). Protein identifications were accepted if they could be established a t grea ter than 99.0% probability to achie v e an FDR less than 1.0% and contained at least two identified peptides. Pr otein pr obabilities were assigned by the Pr otein Pr ophet algorithm ( 39 ). Pr oteins that contained similar peptides and could not be dif ferentia ted based on MS / MS analysis alone were grouped to satisfy the principles of parsimony.
For calculation of fold enrichment for proteins with 0 peptides detected in control immunoprecipita tion, tha t number was changed to 1 to allow for fold enrichment calculation.

AlphaFold analyses
AlphaFold (AF) modelling was carried out using the AFmultimer software ( 40 , 41 ) installed on BlueBEAR, the high-performance computing (HPC) cluster at the University of Birmingham. Xenopus laevis DONSON protein structures were predicted using the standalone version of AlphaFold (2.1.1-foss-2021a-CUDA-11.3.1). The Al-phaFold models used were selected after producing fiv e conformations ranked by their per-residue confidence scores (pLDDT) for each conforma tion ( 42 ). Dif ferently, Human DONSON AF structure was downloaded from UniProt. UniPr ot pr otein sequences used are Q5U4U4 and Q9NYP3 for Xenopus and human respecti v el y. Molecular gra phics and analyses performed with UCSF ChimeraX, ( 43 , 44 ).

DONSON interacts with replisomes on chromatin in egg extract
DONSON was pr eviously r eported to interact with replisomes in human cells, especially those activated in early S-phase ( 25 , 26 ) and we have found DONSON interacting with terminated replisomes on chromatin in Xenopus egg extract (Supplementary Figure S1A). In this experiment, (described previously in ( 37 )), a DNA replication reaction was set up in the egg extract, supplemented with inhibitors to block unloading of ubiquitylated terminated replisomes: cullin E3 ubiquitin ligases inhibitor (MLN4924) to block Cul2 Lrr1 from ubiquitylating Mcm7 within the terminated CMG helicases, and an ATPase-dead mutant of p97 segregase (p97mut). Such treatments led to the accumulation of large quantities of post-termination replisomes on chromatin. They wer e immunopr ecipitated with Mcm3 antibodies and co-immunoprecipitated proteins were analysed by mass spectrometry ( 37 ). Many replisome components were identified this way and, interestingly, we also found multiple peptides of DONSON. As DONSON was reported to protect stalled replication forks in human cells ( 25 ), we hypothesised that it may also play a role in regulation of replisome disassembly during DNA r eplication. Ther efor e, we raised two antibodies, in sheep ( ␣-DONSONs) and in rabbit ( ␣-DONSONr) against recombinant X.l. DONSON purified from bacteria (Supplementary Figure S1B). We used these antibodies to analyse the pattern of DONSON chromatin binding during DNA replication in Xenopus egg extract.
Firstl y, we anal ysed an unperturbed DN A r eplication r eaction (Figure 1 B). This reaction was initiated in egg extract by addition of the substrate DNA (demembranated sperm chromatin), and the progress of the replication reaction followed by incorporation of ␣ 32 P-dATP into the synthesised nascent DNA was measured in parallel to isolation of the chromatin bound proteins for western blot analysis (Figure 1 B). As expected, we could observe Mcm7 binding to chromatin from the beginning of the reaction as origin licensing takes place within the first 5 min of the r eplication r eaction in our system. The replisome components wer e pr esent on chroma tin a t 60 min, which coincides with the fast rate of DNA synthesis in this experiment and thus a large number of replication forks replicating DN A. Finall y, we could observe DONSON binding to chromatin at a similar time as the replisome factors but found that it was retained at a lower le v el for a longer time (Figure 1 B). We decided therefore to test whether the observed DONSON interaction with chromatin depends on acti v e r eplication (Figur e 1 C). For this, we inhibited DNA replication either through inhibition of origin licensing with geminin (Mcm4 and Mcm7 not present on chromatin), or through inhibition of CDK activity with p27 Kip1 (Mcm4 loaded and phosphorylated by DDK; no Cdc45, PCNA and Psf2 loading), or through inhibition of DDK activity with Cdc7i PHA-767491 (Mcm4 loaded but not phosphoryla ted; replica tion fork factors not present). In all situations where the DNA replication reaction was inhibited, we observed reduced binding of DONSON to chromatin (Figure 1 C), suggesting that DONSON's interaction with chromatin is mostly dependent on acti v e replication. The reduction of binding was reproducibly less profound upon Cdc7 inhibition suggesting that DONSON's chromatin interaction depends more on CDK activity. We verified also that once on chromatin during S-phase, DONSON indeed forms part of the replisome, as it can co-immunoprecipitate with GINS (Psf2) and Cdc45 from a replica ting chroma tin fraction ( Figure 1 D and E).
As discussed previously, we observed DONSON interacting with immunoprecipita ted termina ted replisomes (Supplementary Figure S1A). To follow up this observation, we analysed DONSON chromatin binding when replisome unloading was inhibited either by inhibition of Cul2 Lrr1 activity using MLN4924 nedd yla tion inhibitor (Supplementary Figure S1D) or by inhibiting the activity of p97 segregase with the NMS973 inhibitor (Supplementary Figure S1C). In both cases, we observed that DONSON accumulates on chromatin with similar timings as terminated replisomes. Altogether, these data suggest that DONSON continues interacting with replisomes at later stages of the replication.
As DONSON was reported in human cells to be important for cell survival when challenged with replication stress, and to fully activate S-phase checkpoint ( 25 , 26 ), we next anal ysed w hether DONSON's chromatin binding activity is stimulated by different forms of replication stress and DNA damage (Figure 1 F). In this case, the DNA replication reactions were set up in the presence of different stressors: aphidicolin --inhibitor of replicati v e polymerases; camptothecin -inhibitor of Topoisomerase 1, causing ssDNA breaks and DNA-protein crosslinks (DPCs); mitomycin C (MMC) --causing interstrand crosslinks (ICLs); EcoRIrestriction enzyme inducing double-stranded DNA breaks (DSBs); etoposide --inhibitor of Topoisomerase 2, causing DSBs and DPCs. All these treatments reduced the treated extract's ability to synthesise nascent DNA, albeit to different le v els, and showed differing le v els of reduction in replisome components binding to replicating chromatin. In all trea tments, DONSON's chroma tin binding pa ttern was most similar to that of other replisome components (PCNA and GINS), and we did not observe stimulation of DON-SON chroma tin associa tion upon any of the treatments (Figure 1 F).
To learn more about DONSON's interaction with chroma tin upon replica tion stress and S-phase checkpoint activation, we also set up replication reactions in the presence of aphidicolin alone or with aphidicolin and caffeine together. Caffeine is a potent A TM / A TR inhibitor and in the presence of aphidicolin and caf feine, replica tion forks fire and stall due to inhibition of polymerases; howe v er, as the checkpoint is not acti v e, origins fire uncontrollably leading to an accumulation of high numbers of stalled replication forks on chromatin, while very little nascent DNA is being synthesised (Figure 1 G). Importantly, under these conditions we can observe that DONSON accumulates on chromatin together with the stalled replication forks, suggesting that S-phase checkpoint activity is not needed for DONSON chromatin binding.
Finally, we wanted to determine whether DONSON still interacts with replisomes upon S-phase checkpoint inhibition. To achie v e this, we isola ted replica ting chromatin challenged with aphidicolin / caffeine as above, digested DNA to release protein complexes from chromatin and immunoprecipitated GINS complex or Cdc45 (Figure 1 H). Reassuringly, we observed that DONSON coimmunoprecipitated with GINS and Cdc45 together with other r eplisome factors, irr especti v ely of checkpoint inhibition. Comparing DONSON interactions with the replisome components on normal S-phase chromatin and chromatin from aphidicolin / caf feine trea ted extract, we can detect the same interactions, although there are more replisomes and more DONSON accumulated upon aphidicolin / caffeine treatment and thus we detect their interactions more robustly.

DONSON is essential f or r eplication in Xenopus egg extract
Having established that DONSON is a component of normal replisomes in the Xenopus system, we aimed to determine the importance of DONSON for replication in the egg extract. To this end, we immunodepleted DON-SON from egg extract using each of the raised antibodies. Both antibodies were able to immunodeplete DON-SON to a le v el lower than 5% of DONSON remaining in the extract (Figure 2 A, B). We then analysed the ability of IgG-depleted and DONSON-depleted extracts to synthesise nascent DNA. In both cases, the replication capacity of DONSON-depleted extract was very strongly diminished (Figure 2 A, B). We also analysed chromatin binding of DNA replication factors in IgG-and DONSONdepleted extracts and could observe no DONSON, but also no Cdc45 and GINS binding to chromatin in the DONSON-depleted extract, which went hand in hand with the inhibition of DNA synthesis (Figure 2 A, B). Importantly, we could rescue both the DNA synthesis and Cdc45 and GINS chromatin biding in DONSON-depleted extract through addition of recombinant X.l. DONSON purified from bacteria (Figure 2 C and Supplementary Figure S1E). This rescue confirms that the only essential replication factor depleted in DONSON-depleted extract was DONSON itself. Addition of recombinant DONSON to DONSONdepleted egg extract readily rescues the le v els of DNA synthesis in the extract, but fewer replication forks seem to be established to support this DNA synthesis. It is likely that DONSON purified from bacteria lacks important posttransla tional modifica tions and is not as acti v e as the endogenous DONSON.

DONSON is essential for replication initiation
As DONSON was proposed to stabilise stalled replication forks in human cells ( 25 ), we next tested whether the inhibition of DNA replication in DONSON-depleted extract is due to the collapse of early replication forks, which could lead to S-phase checkpoint activation and inhibition of any further origin firing. To do so, we supplemented IgG-or DONSON-depleted extract with caffeine to block potential checkpoint activation (Figure 3 A). Such treatment of DONSON-depleted extract did not rescue its ability to synthesise DNA, and importantly, did not rescue Cdc45 and GINS chromatin binding (Figure 3 A) suggesting that it is not checkpoint activation that inhibits Cdc45 and GINS chromatin binding in DONSON-depleted extract. We also verified that DONSON depletion is not blocking the formation of the nuclear envelope around chromatin in the egg extract, which is essential for replication in our system ( 45 ) (data not shown). During

DONSON's cooperation with other replication initiation factors
As DONSON seems to be involved in the initiation stage of DNA replication, we next examined in more detail the timings of chromatin association for DONSON and other initiation factors (Figure 3 D). We set up an unchallenged DNA r eplication r eaction in the egg extract, but this time analysed a pattern of protein binding to chromatin early on during the r eaction, befor e onset of replication forks firing. As shown previously (Figure 1 B), we can see binding of Mcm7 from the earliest time point (10 min) due to rapid origin licensing. Treslin starts binding to chromatin early (10-20 min), which is consistent with a pr evious r eport showing that the Treslin / MTBP complex binds chromatin befor e r eplica tion initia tion ( 18 ). In this replica tion reaction, Cdc45 and GINS start binding chromatin at 20 min with a peak at 30-40 min. DONSON's pattern of chromatin binding mostly resembles that of Cdc45 and GINS (Figure 3 D).
We also compared chromatin binding of DONSON and other initiation factors upon inhibition of replication initia tion a t dif fer ent stages of the r eaction: inhibition of licensing (geminin treatment), inhibition of DDK activity (PHA-767491 Cdc7i) and inhibition of CDK activity (p27 Kip1 inhibitor) (Figure 3 E). Treatment with geminin was shown previously to significantly reduce chromatin binding of Treslin / MTBP and RecQ4 ( 18 , 20 , 46 ), while inhibition of CDK activity with p27 Kip1 led to Treslin / MTBP hyper-loading onto chromatin ( 18 , 46 ) and partially reduced loading of RecQ4 ( 18 , 20 , 46 ). Inhibition of DDK activ-ity was shown to have a less profound of an effect on Treslin / MTBP chromatin binding ( 18 ). Our observations agree with these previous findings. Previous studies concerning TopBP1 howe v er, showed tha t chroma tin binding was either unaffected by all such treatments ( 47 ) or reduced to different le v els by all of them ( 18 , 46 ). In our hands, we observe that TopBP1 chromatin binding is partially reduced after CDK inhibition. Some of these differences are likely to be the effect of different antibodies used in different studies. Interestingly, DONSON chromatin binding is reduced upon inhibition of S-phase kinases DDK and CDK, especially with the latter, and its chromatin binding pattern resembles much more the replisome components rather than other initiation factors (Figure 3 E), suggesting that it is acting relati v ely la te in the initia tion cascade.
In an effort to understand the mechanism by which DONSON exerts its initiation function, we first compared interactors of DONSON on chromatin at early or late replication stages. To do so, we accumulated DONSON and other replisome components on chromatin either at the termination stage, by inhibiting replisome unloading with p97i (akin to the experiment in Supplementary Figure S1C), or at an earlier stage of replication by supplementing extract with aphidicolin and caffeine (akin to Figur e 1 F); her e, the r eplisomes ar e stalled, but new origins keep firing due to checkpoint inhibition. We immunoprecipitated DONSON and analysed co-immunoprecipitated factors by mass spectrometry. We then calculated the enrichment of replisome and replication initiation factors in both IPs (Supplementary Figure S2A). Interestingly, we observed enrichment of peptides of TopBP1 and RecQ4 in the DONSON IP from aphidicolin / caffeine treated sample. We then analysed interactions between DONSON and initiation factors within the aphidicolin / caffeine chromatin fraction by immunoblotting (Figure 4 A). Here we observed that DONSON interacts with the replisome components, as we have seen in Figur e 1 , but inter estingly also with TopBP1, RecQ4 and Tr eslin / MTBP (Figur e 4 A). Importantly, r eciprocal immunoprecipitation of TopBP1 and Treslin from aphidicolin / caf feine trea ted chroma tin fraction shows analogous interactions with DONSON (Figure 4 B). Altogether these data suggest again that DONSON is not only a part of the matur e r eplisome, but also plays a role at the initiation stage.

DONSON interacts with TopBP1
We have shown that DONSON forms a part of the replisome during the initiation stage of DNA replication as it can interact with replication initiation factors on chromatin ( Figure 4 A and B). We next wanted to understand whether any of these interactions can occur directly with DONSON, outside the context of a r eplisome. We ther efore isolated nuclei containing nucleoplasm (chromatin and nucleoplasm), in a buffer without addition of detergent, and immunoprecipitated DONSON. Interestingly, out of all potential interacting proteins tested, DONSON could interact with TopBP1 and this interaction was strongly reduced by p27 Kip1 CDK inhibitor (Figure 4 C). We therefore partially depleted TopBP1 from egg extract to 25% of its original concentration (Supplementary Figure S2B) and Chromatin was isolated after 60 min of replication reaction, protein complexes were released from chromatin by digestion of DNA with benzonase and non-specific control antibodies or ␣-DONSON(s) antibodies used for immunoprecipitation. The input and immunoprecipitated samples were analysed by immunoblotting with indicated antibodies. ( B ) As (A) but ␣-TopBP1 or ␣-Treslin antibodies were used for imm unoprecipitation. ( C ) DN A replication reaction was set up in egg extract with optional addition of replica tion initia tion inhibitors (p27 KIP1 or Cdc7i). In the middle of S-phase, replicating nuclei were isolated in a buffer without detergent, which allows to retain part of the n ucleoplasm. F rom this input material, DONSON was immunoprecipitated and immunoprecipitation samples analysed by immunoblotting with indicated antibodies. As a control, non-specific antibody immunoprecipitation from unchallenged S-phase input was performed in parallel. ( D ) TopBP1 was immunodepleted from egg extract to 25% of the original le v el. Replication reaction was set up in IgG-and TopBP1-depleted extract and chromatin isolated at indicated times of reaction and chromatin fractions analysed as above.
anal ysed w hether chromatin binding of DONSON depends on TopBP1 (Figure 4 D). With this, we observed that the le v el of chromatin-bound DONSON was strongly reduced. Altogether, all these data suggest that DONSON interacts with chromatin in a TopBP1-and CDK-dependent manner and acts as a replica tion initia tion factor, essential for Cdc45 and GINS recruitment to the CMG complex.

DONSON's protein structure and mutations
We next wanted to explore whether DONSON's function may be conserved through evolution. The AlphaFold prediction of X.l. DONSON's structure suggests that DON-SON is formed of a long, mostly unstructured N-terminus (D1), wrapping around a folded globular core with one protruding fle xib le loop (D3) (Figur e 5 A-C). Inter estingly, the folded globular core of DONSON is highly conserved throughout evolution, while the unstructured N-terminus is not (Supplementary Figure S3A Figure S3C). We chose two mutations leading to MPD: M446T ( Xenopus M463T), which is a homozygous mutation found in 3 patients, E504K ( Xenopus E521K), which was found in 2 patients as a compound heterozygous mutation combined with a stop codon introducing m utation. Interestingl y, we could not efficiently express X.l. DONSON-E521K in bacteria, most likely due to protein destabilisation (Supplementary Figure S3C), while X.l. DONSON-M446T purified well and could fully r estor e the ability of the DONSONdepleted extract to synthesise nascent DNA (Supplementary Figure S3C and Figure 5 F). We also chose three mutations leading to Meier-Gorlin syndrome: R211C ( Xenopus R217C), which is a homozygous mutation found in two patients, and F165S ( Xenopus F171S), which was found in one patient as a compound heterozygous mutation combined with an intronic splice site mutation, and another compound heterozygous mutation P224S ( Xenopus P230S).
All three mutants were readily purified ( Supplementary Figure S3C) and while X.l. DONSON-F171S and DONSON-P230S could rescue the ability of DONSON-depleted extract to synthesise nascent DNA, X.l. DONSON-R217C was partially defecti v e ( Figure 5 F), suggesting that this mutation may disrupt the functionality of DONSON during replica tion initia tion.
Altogether, these investigations revealed that flexible and poor ly conserved fr agments of DONSON are ne v ertheless essential for DONSON activity and that some of the patient mutations leading to the Meier-Gorlin syndrome phenotype can indeed disrupt the essential function of DON-SON during DNA replication initiation.

DISCUSSION
Based on the data pr esented her e we would like to propose that DONSON is a novel DNA replica tion initia tion factor, as it is clearly essential for loading of Cdc45 and GINS onto origins during initiation. In the absence of DONSON, the replicati v e helicase, the CMG complex, cannot be formed and without the helicase activity, replication forks cannot be assembled. Mechanistically, our data suggest that DON-SON acts through direct interaction with TopBP1, and this interaction is dependent on CDK activity. TopBP1, and its yeast homologue Dpb11 / Cut5, is a master regulator of replica tion initia tion. TopBP1 it is thought to bridge together Treslin / MTBP in complex with Cdc45, and RecQ4 in complex with GINS (Figure 1 A) (14)(15)(16)(17)(18)(19)(20). Howe v er, immunodepletion of RecQ4 from Xenopus egg extract does not inhibit Cdc45 and GINS chromatin binding ( 20 ), suggesting that RecQ4, despite the partial homology to Sld2, does not play an equivalent role in higher eukaryotes.
TopBP1 acts as a scaffold protein, able to make numerous pr otein-pr otein interactions, facilitated in most cases through the multiple BR CT domains. BR CT domains most often interact with phosphorylated partners and often act in pairs. Yeast Dpb11 ( S. cerevisiae ) and Cut5 ( S. pombe ) contain 4 BRCT domains each, with BRCT1 + 2 interacting with phosphorylated Sld3 ( 9 ) and BRCT3 + 4 interacting with phosphorylated Sld2 ( 49 , 50 ). Howe v er, the human and Xenopus TopBP1 homologues are much larger proteins, containing eight BRCT domains (1-8), with an additional N-terminal BRCT domain identified in human TopBP1, named BRCT0 ( 51 ). This evolution of multiple additional BRCT domains suggests that higher eukaryotic TopBP1 is likely to have additional interactors, functions and regulations than the yeast counterparts. Like the yeast protein howe v er, TopBP1 binds with phosphorylated Treslin (at S1001 in human and S976 in Xenopus ) through its BRCT1 + 2 domains ( 14 ). TopBP1 also contains a GINS interacting motif (GINI), which lies between BRCT3 and BRCT4 domains ( 52 ), and was proposed recently to interact directly with GINS using BRCT4 + 5 domains ( 53 ). Interestingly, although BRCT4 + 5 are homologous to fungal BRCT3 + 4, it is BRCT7 + 8 that binds RecQ4, but this interaction is not dependent on CDK phosphorylation in either species ( 21 , 54 ). Moreover, TopBP1 BRCT7 + 8 are not essential for replica tion initia tion a t all ( 46 ), strengthening the argument that RecQ4 cannot be the functional homologue of yeast Sld2.
AlphaFold prediction of the DONSON-TopBP1 interaction (not shown) suggests that DONSON interacts with BRCT3 of TopBP1. Interestingly, it was shown that although the N-terminal fragment of TopBP1 spanning BRCT1 + 2 is sufficient to interact with Treslin, it is not sufficient to support DNA replication in the Xenopus egg extr act. Instead, the TopBP1 fr agment including BRCT1-3 is essential and sufficient to support replication initiation ( 46 ). This suggests an essential initiation role for BRCT3 in TopBP1, which thus far was not assigned to interact with any specific initiation factor. DONSON could ther efor e be such a factor. Interestingly, DONSON's interaction with TopBP1 is dependent on CDK activity, as would be expected for the functional homologue of Sld2. It remains to be determined howe v er, whether this interaction is dri v en by direct phosphorylation of DONSON by CDKs, or perhaps phosphorylation of another factor.
We propose here that DONSON likely fills the role of a functional homologue of Sld2 in higher eukaryotes. In yeast, Dpb11-Sld2 forms a 'pre-landing complex' together with GINS and DNA Pol ε ( 55 ). In our analyses of DON-SON interactions, we could detect DONSON interacting with GINS and DNA Pol ε , but only from chromatin fractions, where many interactions can be stimulated by formation of replisomes around the CMG complex. The only robust interaction that we found for DONSON within the nucleoplasmic fraction was that with TopBP1. It is not yet clear, ther efor e, whether DONSON and TopBP1 indeed form an analogous pre-landing complex. How else could DONSON act during origin firing to stimulate Cdc45 and GINS interaction with Mcm2-7? The molecular mechanism underpinning how TopBP1 and Treslin / MTBP deli v er GINS and Cdc45 to Mcm2-7 remains unknown. It is clear from the work of many groups that Treslin / MTBP bind Mcm2-7 early in the initiation reaction. Their interaction is strengthened by DDK activity but does not r equir e CDK activity; on the contrary, they are hyperloaded upon inhibition of CDKs ( 18 , 46 ). Phosphorylation of Treslin / MTBP by CDKs stimulates their interaction with N-terminal BRCT repeats of TopBP1 and is essential for helicase activation ( 14-15 , 46 ). TopBP1 can interact with GINS and is clearly essential for Cdc45 and GINS chromatin binding ( 47 , 53 ). We show here that DONSON interaction with TopBP1 depends on CDK activity. It is likely ther efor e that upon CDK activation, both TopBP1 and DONSON ar e r ecruited to phosphorylated Treslin / MTBP at Mcm2-7 double hexamers. The presence of all these initiation factors together allows for stable recruitment of Cdc45 and GINS into a CMG complex and helicase activation. TopBP1 is a limiting factor for origin firing ( 56 ) and is likely to be released following helicase activation, to facilitate firing of further origins, while DONSON is likely to be retained as part of the replisome in Xenopus egg extract.
Our data clearly indicate that DONSON plays an essential role in initiation of DNA replication, but DONSON has been shown previously in human cells to regulate the progr ession of r eplication forks, especially in r esponse to r eplication str ess ( 25 , 26 ). Howe v er, the molecular function of DONSON has largely been investigated in highly unstable HeLa cells ( 25 , 26 ), which often over expr ess r eplication factors to support fast proliferation. Moreover, DONSON w as alw ays downregulated using siRN A (w hich takes several days to ef ficiently downregula te the protein le v el) or through analysis of patient-deri v ed mutants, which destabilise the protein and lead to continuous lower le v els of DONSON ( 25 , 26 ). Finally, the effects of DONSON downregulation on DNA replication were analysed by DNA fibre analyses or creation of DN A damage, w hich would miss cells that do not initiate DNA r eplication. Ther efor e, it is possible that the initiation phenotype of DONSON downregulation had been missed thus far in human cells and awaits careful dissection and determination. Interestingly, DONSON was also proposed to specifically form part of early firing replisomes, and not the ones in late S-phase ( 26 ). This observation raises the possibility that DONSON specificall y activates earl y firing origins, in contrast to Rif1, which is proposed to regulate firing of mainly late replication origins ( 57 ). Such a specified role for DONSON would also explain why it is so essential during embryogenesis in both Xenopus egg extract and in Drosophila embryos ( 27 , 28 ), as both of these exemplify embryonic systems, which have very fast DNA replication and lack replication timing programmes, so all origins can be regarded as early firing.
Downregulation of DONSON in human cells has been shown to diminish the ability of the fir ed r eplication forks to progress past replication obstacles and causes activation of S-phase checkpoint ( 25 , 26 ). We observed here that DON-SON interacts with replisomes not only during replication initiation but also at later stages of the DNA replication reaction. DONSON accumulates with the rest of the replisome on chromatin upon inhibition of unloading of posttermination replisomes and interacts with post-termination replisomes. It is very likely therefore that after facilitating the initiation of replication machinery, DONSON travels with the replisomes and regulates fork progression. To study such a function of DONSON, we need to de v elop separa tion-of-function mutants, tha t can deli v er the initiation function but are defecti v e in the fork progression function. It is possible that some patient mutations may deli v er such tools for the future.
Mutations in DONSON have been known for a while to be the underlying genetic problems, which lead to microcephalic primordial dwarfism (MPD). Interestingl y, m utations in DONSON were also recently shown to specifically cause Meier-Gorlin syndrome ( 48 , 58 ), the subtype of MPD caused by d ysregula tion of DNA origin firing and replication initiation. The discovery of a replication initiation role f or DONSON theref ore explains the phenotypes caused by these mutations. It is interesting to speculate that if DON-SON has multiple roles during DNA replica tion, dif ferent mutations within the protein can specifically disrupt these different functions. Most DONSON patient mutations have been shown to destabilise the protein and lead to much lower le v els of DONSON protein in the cells ( 25 ). Ne v ertheless, it is possible that some point mutations (e.g. R211C) disrupt specifically its interaction with TopBP1, while others may destabilise interactions with other replication factors important for fork progression and checkpoint activation. Mor e r esear ch is needed to answer all these questions.
Finally, in the last two years, r esear chers have r eached a better understanding of the importance of DONSON in cancer de v elopment. DONSON was found to be a biomarker for risk stra tifica tion in clear cell renal carcinoma (ccRCC) ( 59 ), and in prostate carcinoma (PCa) ( 60 ). It has also been linked with a malignant phenotype in ccRCC cell culture models ( 61 ). Furthermore, it was shown to be a dri v er of breast cancer progression and a potential target for subsequent therapies ( 62 ). Understanding the different molecular functions of DONSON during DNA replication provides us with a solid foundation for translation of DON-SON's potential as a biomarker and as a therapy target in the future.

DA T A A V AILABILITY
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE repository with the dataset identifier: Pr oject Name: Chroma tin proteome of S-phase in IgGand DONSON-depleted egg extract Project accession : PXD042435 Project DOI: 10.6019 / PXD042435