We report a hybrid ensemble and single-molecule assay to directly image and quantify the motion of fluorescently labeled, fully reconstituted Cdc45/Mcm2-7/GINS (CMG) helicase on linear DNA molecules held in place in an optical trap.
Eukaryotes have one replicative helicase known as CMG, which centrally organizes and drives the replisome, and leads the way at the front of replication forks. Obtaining a deep mechanistic understanding of the dynamics of CMG is critical to elucidating how cells achieve the enormous task of efficiently and accurately replicating their entire genome once per cell cycle. Single-molecule techniques are uniquely suited to quantify the dynamics of CMG due to their unparalleled temporal and spatial resolution. Nevertheless, single-molecule studies of CMG motion have thus far relied on pre-formed CMG purified from cells as a complex, which precludes the study of the steps leading up to its activation. Here, we describe a hybrid ensemble and single-molecule assay that allowed imaging at the single-molecule level of the motion of fluorescently labeled CMG after fully reconstituting its assembly and activation from 36 different purified S. cerevisiae polypeptides. This assay relies on the double functionalization of the ends of a linear DNA substrate with two orthogonal attachment moieties, and can be adapted to study similarly complex DNA-processing mechanisms at the single-molecule level.
DNA replication in eukaryotes is carried out by a dynamic protein complex known as the replisome1. A key component of this complex is the eukaryotic replicative helicase Cdc45/Mcm2-7/GINS (CMG), which drives and centrally organizes the replisome, leading the way at the front of replication forks1,2. Obtaining a deep quantitative understanding of the dynamics of CMG is therefore critical to understanding the dynamics of the replisome. Such an understanding could be acquired with single-molecule techniques, which are uniquely suited to study molecular motors, such as CMG, due to their unmatched spatial and temporal resolution, and can provide us with an unparalleled quantitative understanding of their function, stochasticity, and dynamics2,3,4,5,6,7,8,9.
In vivo, CMG is loaded and activated in temporally separated fashion to ensure that replication occurs only once per cell cycle1,10,11. First, in the G1 phase of the cell cycle, a set of proteins known as loading factors loads the first component of CMG, the Mcm2-7 hexameric complex, onto dsDNA12,13,14,15,16 in the form of double hexamers with a head-to-head configuration15,17,18. In the specific case of yeast, this initial process occurs at specific DNA sequences known as origins of replication1. Although Mcm2-7 is the motor core of the replicative helicase, it is by itself unable to unwind DNA19 without the two helicase-activating factors Cdc45 and GINS, which need to be recruited to the loaded Mcm2-7 to give rise to fully active CMG11,19,20,21. The process of helicase activation takes place in the S phase of the cell cycle and starts with the selective phosphorylation of Mcm2-7 double hexamers by the cell cycle-regulated kinase DDK22,23,24. These phosphorylation events facilitate the recruitment of Cdc45 and GINS to the Mcm2-7 double hexamers10,22,23,24,25,26 by a second set of proteins known as firing factors10,11,26. The binding of Cdc45 and GINS gives rise to two sister CMG helicases, which initially enclose both strands of the parental DNA and are still located in a head-to-head configuration11,27. In the final activation step, the firing factor Mcm10 catalyzes the ATP hydrolysis-dependent extrusion of one DNA strand from each sister CMG11. After strand extrusion, sister CMG helicases bypass and separate from each other by translocating along ssDNA in an ATP hydrolysis-dependent manner11,20,21,28, unwinding DNA by sterically excluding the non-translocation strand29. This entire process has been fully reconstituted in vitro from a minimal set of 36 purified S. cerevisiae polypeptides10,11.
Despite the exquisite in vivo regulation of CMG assembly and activation described above, in vitro reconstituted single-molecule motion studies of CMG2,30,31,32,33,34 have thus far relied on pre-activated CMG purified as a complex from cells20,21, missing all the steps prior to its activation and the bidirectional nature of its motion. This pre-activated CMG approach has been the gold standard in the single-molecule field partly due to the biochemical complexity of the fully reconstituted CMG assembly reaction10,11. This biochemical reaction has been challenging to translate from the bulk biochemical level to the single-molecule level for several reasons. First, to maximize reaction efficiencies, the loading and firing factors needed for CMG assembly and activation are required at concentrations in the range of 10-200 nM10,11,27. These ranges of concentration correspond to the high end of what most single-molecule techniques can tolerate, especially when using fluorescently labeled components35. Finally, CMG has evolved to cruise through thousands of base pairs in a cell36,37,38,39. Therefore, to study its motion at a biologically relevant spatial scale, one requires long DNA substrates (typically of lengths in the order of tens of kilobases)30,31,34,40,41,42. Employing such long DNA substrates poses the additional challenge that the longer the DNA substrate is, the more potential non-specific binding sites for proteins and protein aggregates it has. In the case of CMG, the latter point is particularly important, as several of the loading and firing factors involved in CMG assembly and activation contain intrinsically disordered regions43 and are aggregation-prone.
Here, we report a hybrid ensemble and single-molecule assay that allowed the observation and quantification of the motion of CMG after fully reconstituting its assembly and activation from 36 purified S. cerevisiae polypeptides28. This assay relies on the double functionalization of both ends of a DNA substrate with two orthogonal attachment moieties: desthiobiotin and digoxigenin2 (Figure 1A). The first moiety, desthiobiotin, is used to reversibly bind the DNA substrate to streptavidin-coated magnetic beads44 (Figure 1B). Following this, fluorescently labeled CMG is assembled and activated onto the bead-bound DNA, and a magnetic rack is used to purify and wash the resulting magnetic bead-bound DNA:CMG complexes (Figure 1C). In doing so, the excess protein that would otherwise aggregate on the DNA substrate is removed; this provides virtually aggregation-free DNA:CMG complexes. Intact complexes are then eluted from the magnetic beads by the addition of a molar excess of free biotin, which can outcompete the desthiobiotin-streptavidin interaction (Figure 1D). Individual DNA:CMG complexes are then bound between two optically trapped polystyrene beads coated with anti-digoxigenin antibody (anti-Dig); for this step, the second moiety on the DNA, digoxigenin, is used as it can bind to anti-Dig even in a buffer solution containing an excess of free biotin (Figure 1E). Once the DNA:CMG complex is held in place in the optical trap, the motion of fluorescently labeled CMG is imaged with a confocal scanning laser (Figure 1F). We anticipate that this assay can be easily adapted for the study at the single-molecule level of similarly complex DNA:protein interactions.
Critical steps and important reagent quality checks
The critical steps and biological reagent quality checks in the assay are highlighted here. First, the purity of the proteins used is important because DNA degradation caused by even small nuclease contaminants in the protein samples will adversely affect the data. This is because only intact (or partially nicked) DNA molecules can be trapped in the dual-beam optical tweezers. More importantly, nicks on the DNA will cause CMG to dissociate41, complicating the observation of CMG's long-range motion. We strongly recommend testing each purified protein for nuclease activity, as well as constantly monitoring the integrity of the starting plasmid substrate to ensure that nicking is reduced to a minimum. The second important step is the careful removal of the magnetic beads following the elution of DNA:CMG complexes. Supernatant removal in this step should be conducted slowly so as not to perturb the collected beads. If magnetic beads are left in the sample flown into the optical tweezers, they will often hit the optically trapped polystyrene beads, causing them to escape the optical trap and complicating the data acquisition. Finally, DNA:CMG complexes should be handled carefully in the optical tweezers. To this end, we recommend not increasing the tension of the DNA above 10 pN, as the application of force may dissociate CMG from the DNA. Furthermore, moving between channels in the microfluidic flow cell should be done as slowly as possible (~ 0.2 mm/s) to prevent the resulting drag forces from dissociating CMG from the DNA.
Modifications of the method
There are several steps of the assay that could be modified. For instance, we have shown that the elution time can be reduced from 60 min to 30 min without significantly affecting the elution yield. In addition, we recommend supplementing the elution buffer with a low (below 1 mM) concentration of either ATP or ATPγS to prevent CMG from diffusing off the DNA ends as well as to generally stabilize CMG28. Further, although the buffer compositions and protein concentrations we report here are based on those employed in prior ensemble biochemical and single-molecule work11,18, the assay we describe is fully compatible with other protocols to assemble CMG26,27. Therefore, any biochemical advancement reported to increase the efficiency of CMG assembly or activation could and should be implemented in the bulk part of the assay to increase the yield. Finally, increasing the time between frames increases the total time in which CMG can be imaged, facilitating the observation of long-range CMG motion before fluorophore bleaching.
Limitations of the method
The hybrid method we describe is limited in that one can only image CMG following its activation in bulk. Further work will be required to observe the activation of CMG in real-time. Another important limitation is that, while we expect CMG to be assembled in pairs17,26,27, the total number of CMG complexes per DNA that we observe is mostly one28, suggesting that CMG, or at least Cdc45 is dissociating from the DNA during the handling of the sensitive DNA:CMG complexes. Reducing the number of handling steps prior to the single-molecule imaging, as well as developing better passivation strategies for the plastic tubing and glass of the microfluidic flow cell are poised to increase this yield.
Significance of the method
Single-molecule motion studies of CMG have thus far employed pre-activated CMG purified as a complex from cells. While relatively simpler, this pre-activated CMG approach is limited in that it misses any of the steps leading up to CMG activation, as well as the bidirectional nature of CMG and replisome motion. On the other hand, the full reconstitution of CMG assembly and activation has the potential to study any pre-activation steps, as well as to study CMG motion in a bidirectional manner. Nevertheless, this approach is harder to translate from the bulk biochemical level to the single-molecule level, as it involves a lot more purified protein factors and steps. The assay we describe here has helped to overcome these challenges by allowing us to image the motion of fully reconstituted CMG at the single-molecule level, allowing us to access some previously missed pre-activation dynamics28. Additionally, although we mostly see one CMG per DNA, we were able to observe several instances of two CMG complexes moving in opposite directions, and we could even capture the initial separation of sister CMGs from one another28, providing some insights into the establishment of bidirectional replication.
Another advantage of this assay compared to previous CMG motion lies in the fully double-stranded nature of the DNA substrate we employ (Figure 1A). In previous pre-activated CMG work, the most common way of binding CMG to the DNA substrate is through a 3' ssDNA flap. This results in a DNA construct that cannot be easily torsionally constrained and thus prohibits the study of the role of supercoiling in replisome progression. Conversely, the new approach we describe here could have the potential to be adapted to study the role of torque in this process, as the DNA substrate used is completely double stranded.
Broader applications of the method
The hybrid assay described will pave the way toward the full reconstitution of a complete eukaryotic replisome, allowing us and others to observe and quantify the important dynamics that allow the replisome to succeed at all its different tasks. DNA replication aside, the assay we report represents an important advancement in translating a complicated biochemical reaction from the bulk biochemical to the single-molecule level. We anticipate that this assay can be easily modified to study similarly complex DNA:protein interactions involved in different DNA processing mechanisms.
The authors have nothing to disclose.
The authors thank Anne Early, Lucy Drury, and Max Douglas for providing yeast strains for the overexpression of unlabeled proteins, as well as N.D. lab members Anuj Kumar, Katinka Ligthart, and Julien Gros for their help purifying loading factors and DDK. The authors also thank Kaley McCluskey, Dorian Mikolajczak, Joseph Yeeles, Jacob Lewis, Alessandro Costa, Hasan Yardimci, and Taekjip Ha for useful scientific discussions. D.R.M. acknowledges funding from a Boehringer Ingelheim Fonds PhD Fellowship. N.D. acknowledges funding from the Netherlands Organisation for Scientific Research (NWO) through Top grant 714.017.002 and from the European Research Council through an Advanced Grant (REPLICHROMA; grant number 789267).
AflII | NEB | R0520L | |
Anti-digoxigenin coated polystyrene beads | Spheroteck | DIGP-20-2 | |
ATP solution | Thermo Fisher | R0441 | |
ATPγS | Roche | 11162306001 | |
BSA | NEB | B9000S | |
C-Trap | Lumicks | ||
CutSmart Buffer | NEB | B6004S | Provided with AflII |
dCTP | Promega | U122B | |
D-Desthiobiotin-7-dATP | Jena Bioscience | NU-835-Desthiobio | |
dGTP | Thermo Fisher | 10218014 | |
Digoxigenin-11-dUTP | Jena Bioscience | NU-803-DIGXL | |
Dynabeads M-280 Streptavidin magnetic beads | Invitrogen | 11205D | |
Klenow Fragment (3'→5' exo-) | NEB | M0212L | |
Microspin S-400 HR spin columns | GE Healthcare | GE27-5140-01 | |
NEBuffer2 | NEB | B7002S | Provided with Klenow Fragment |
Nonidet P 40 Substitute | Sigma | 74385 | |
Pluronic F-127 | Sigma | P2443 |
.