Cycling of the E. coli lagging strand polymerase is triggered exclusively by the availability of a new primer at the replication fork

Two models have been proposed for triggering release of the lagging strand polymerase at the replication fork, enabling cycling to the primer for the next Okazaki fragment—either collision with the 5′-end of the preceding fragment (collision model) or synthesis of a new primer by primase (signaling model). Specific perturbation of lagging strand elongation on minicircles with a highly asymmetric G:C distribution with ddGTP or dGDPNP yielded results that confirmed the signaling model and ruled out the collision model. We demonstrated that the presence of a primer, not primase per se, provides the signal that triggers cycling. Lagging strand synthesis proceeds much faster than leading strand synthesis, explaining why gaps between Okazaki fragments are not found under physiological conditions.


Preincubation of reaction components minimizes the lag phase and permits synchronization of the rolling circle reaction
Simply mixing all components together leads to a rolling circle reaction with a lag phase at the beginning due to unsynchronized DNA synthesis. To minimize this problem, a pre-initiation complex of all protein components was assembled on the template in the presence of ATPS, CTP, GTP, and UTP. Omitting ATP from the pre-incubation stage ensures that helicase can be loaded but not translocated. Once dNTPs and ATP were added, synthesis of each strand initiates with a reduced lag phase ( Figure S4). A final concentration of 5 M ATPS was required for reactions containing 20 nM and 10 nM template with preincubation times of 5 and 11 min, respectively. A final concentration of 10 M ATPS was required for reactions containing 1 nM template with a 17 min pre-incubation time. As the concentration of ddGTP is increased beyond the level shown in Figure 2C in the main article, the level of incorporation of dNTPs in the lagging strand product decreases ( Figure S5). This is, in part, the result of the synthesis of shorter Okazaki fragments ( Figure 2A). But, there is a reduction of the overall molar level of Okazaki fragment synthesis, indicating some level of perturbation. Nevertheless, the rate of Okazaki fragment synthesis remains linear for over two minutes, even in the presence of the highest ddGTP concentrations.

dGDPNP displays a higher apparent K m for the DNA polymerase III holoenzyme-catalyzed reaction and can be used to modulate the elongation rate
The DNA Pol III HE incorporates dGDPNP with a 20-fold higher apparent K m than dGTP ( Figure   S6A,B). The K m for dGDPNP (40 M) is sufficiently high that reactions can be conducted at sub-K m concentrations without danger of depleting nucleotides. This allowed modulation of the rate of elongation on cytosine-containing templates and allowed us to selectively slow the rate of lagging strand synthesis when templates with an asymmetric G:C distribution were used. The measured K m is an estimate and expressed as an apparent value, because the reaction kinetics monitored included both initiation complex formation and elongation stages.
On a template containing approximately 25% C, the rate of elongation in the presence of nearsaturating dGDPNP was reduced to 57 nt/s, 10-fold slower than the 570 nt/s measured in the presence of saturating dGTP ( Figure S6C). Thus, dGDPNP may also be slowing the chemistry step of the reaction sufficiently that it becomes rate-limiting. Decreasing dGDPNP to a level of 0.75 K m decreased the observed elongation rate about 25-fold relative to dGTP. In contrast, the rate of leading strand elongation on a minicircle template containing only two G residues is not slowed significantly ( Figure S6D).  s, and an extra 10 min at 72°C at the end. A 1.5% agarose gel showed that more than 95% amplified products were the target DNA ( Figure S8B lane 1, Figure S8A step 1). The PCR product (80 mg) was extracted with one volume of phenol/chloroform/isoamyl alcohol (25:24:1) and one volume of chloroform, and precipitated by the addition of 0.5 volumes of 5 M ammonium acetate and 1.5 volumes of isopropyl alcohol. The pellet was washed with 70% ethanol and dissolved in 10 mM Tris-HCl buffer (pH 8) to 1 g/l. The purified PCR product (40 mg) was digested with EcoRI (666 U/mg DNA) at 37°C for 9 h, and the digestion was stopped by heating at 65°C for 40 min. Digestion was >90% completed ( Figure S8B lane 2, Figure S8A step 2a). The linear minicircle DNA was separated from biotincontaining terminal fragments created by EcoRI digestion by passing the digested DNA over a high capacity streptavidin resin (Pierce, 10 ml). Electrophoresis in 1.5% agarose showed that more than 95% of the product was the linear minicircle DNA ( Figure S8C lane 2, Figure S8A step 2b). The purified linear DNA (27 mg) was diluted to 2.5 g/ml and ligated using ligase (0.4 U/ml DNA solution, Epicentre) in 33 mM Tris-acetate (pH 7.8), 66 mM potassium acetate, 10 mM magnesium acetate, 5 mM DTT, and 1 mM ATP at 16°C for 20 h ( Figure S8D lane 2, S8A step 3a). The unligated fragment and linear multiligated product were digested completely from both 3' and 5' ends by the combination of lambda exonuclease (0.3 U/ml ligation mixture), exonuclease I (0.3 U/ml ligation mixture), and exonuclease III (1 U/ml ligation mixture) at 37°C for 8 h. All enzymes were thermally inactivated at 80°C for 20 min.
Electrophoresis in 2% agarose showed that more than 90% of the product was the ligated minicircle DNA ( Figure S8D lane 3, Figure S8A step 3b). NaCl was added to the ligation reaction mixture to 0.5 M final concentration. Then the ligated minicircle DNA was loaded onto three QIAGEN-tip 10000 columns.
The column was washed with Qiagen Buffer QC and the DNA eluted with Qiagen Buffer QF. DNA was precipitated by adding 0.7 volumes of isopropanol, washed with 70% ethanol, and dissolved in TE buffer to 1 g/l ( Figure S8A step 3c). The purified product (7 mg) was nicked at the single recognition site with Nt. BstNBI nicking enzyme (2 U/g DNA) at 55°C for 3 h. Nt. BstNBI was thermally inactivated at 80°C for 30 min ( Figure S8A step 4). The nicked DNA was incubated with Vent polymerase (0.75 U/g DNA) and 300 M dATP, dCTP, and dTTP at 75°C for 2 h to form a 394-bp-long DNA flap (37). EDTA was added to a final concentration of 25 mM. The tailed DNA was purified by phenol chloroform extraction and isopropyl alcohol purification as described above. Electrophoresis in 2% agarose indicated a yield >80% ( Figure S8E lane 2, Figure S8A   NaCl gradient. DnaT eluted in Buffer T+220 mM NaCl and was pooled and loaded onto a Heparin Sepharose column (110 ml) equilibrated with Buffer T+100 mM NaCl. The column was washed with 7 volumes of Buffer T+100 mM NaCl, and proteins were eluted with 12 volumes of Buffer T with a 100 mM-500 mM NaCl gradient. DnaT eluted with Buffer T+280 mM NaCl was pooled and precipitated by addition of ammonium sulfate to 65% saturation. The pellet was resuspended by 2 ml of Buffer T+150 mM NaCl and 30% glycerol, and loaded onto a Sephacryl 200 column (105 ml) equilibrated with Buffer T+150mM NaCl and 30% glycerol. The eluate containing DnaT (35 mg) was collected, aliquoted, frozen in liquid N 2 , and stored at -80°C.
DnaC Fr II was prepared by addition of 0.075% polyethyleneimine and 50% saturated ammonium sulfate to Fr I (generated from 110 g cells). The pellet was resuspended in Buffer B and the conductivity of the resulting solution was similar to that of Buffer B+20 mM NaCl. The resulting solution was loaded onto a Q Sepharose column (210 ml) equilibrated with Buffer B+20 mM NaCl. The column was washed with 3 volumes of Buffer B+20 mM NaCl, and the flow-through was collected and loaded onto a phosphocellulose column (40 ml) equilibrated with Buffer B+20 mM NaCl. The column was washed with 3 volumes of Buffer B+20 mM NaCl, and proteins were eluted with 10 volumes of Buffer B with a 20 mM-300 mM NaCl gradient. DnaC eluted with Buffer B+180 mM NaCl was pooled and loaded onto a hydroxyapatite column (11 ml) equilibrated with Buffer B+50 mM NaCl. The column was washed with 6 volumes of Buffer B+50 mM NaCl, and proteins were eluted with 15 volumes of Buffer C with a 0 mM-300 mM ammonium sulfate gradient. DnaC eluted with Buffer C+90 mM ammonium sulfate was pooled and dialyzed against Buffer B+150mM NaCl and 30% glycerol. Purified DnaC (10 mg) was collected, aliquoted, frozen in liquid N 2 , and stored at -80°C.
To prepare Pol III* Fr II, Fr I (generated from 125 g cells) was adjusted to 40% ammonium sulfate.
Contaminants were removed by backwashing with decreasing amounts of ammonium sulfate first using a 0.20 ammonium sulfate backwash (specified as g added to each ml buffer), followed by a 0.17 ammonium sulfate backwash to generate Fr II as described (46). The pellet (147 mg protein; 8.6x10 7 units) was resuspended in Buffer A and the conductivity of the resulting solution was adjusted to that of Buffer A+20 mM NaCl. The resulting solution was loaded onto an SP Sepharose column (100 ml) equilibrated with Buffer A+20 mM NaCl. The column was washed with 3 volumes of Buffer A+20 mM NaCl, and proteins were eluted with 10 volumes of Buffer A with a 20 mM-200 mM NaCl gradient. Pol III* eluted with Buffer A+110 mM NaCl and was pooled and precipitated by addition of ammonium sulfate to 55% saturation. The pellet was resuspended by 0.85 ml of Buffer F (5.1 mg protein; 1.5x10 7 units), and loaded onto a Sephacryl S-300 column (10 ml) equilibrated with Buffer F. The eluate containing Pol III* (3.8 mg, 1.1x10 7 units) was collected, aliquoted, frozen in liquid N 2 , and stored at -

80°C.
Concentrations of all proteins were measured with a Bradford protein assay using the Albumin Standard from Pierce (47).

Alkaline agarose gel electrophoresis
For the analysis of the size of lagging strand products, samples were digested with protease K (30 min, 37 o C, 25 g/ml); mixed with 30 mM NaOH, 2 mM EDTA, 2% glycerol, and 0.02% bromophenol blue; and fractionated on 0.6% alkaline agarose gels for approximately 18 h at 24 V in a running buffer of 30 mM NaOH and 2 mM EDTA. Gels were fixed in 8% (w/v) trichloroacetic acid, dried onto DEAE paper, autoradiographed on storage phosphor screens, and scanned with a PhosphorImager.

Development of a method to fill gaps between Okazaki fragments without strand displacement of the downstream Okazaki fragment
Gaps between incomplete Okazaki fragments were filled by thermophilic polymerases to minimize issues that might result from secondary structure within gaps. To obtain accurate quantification of gap size, we needed to ensure that the polymerase used did not catalyze strand displacement synthesis into the downstream Okazaki fragment. Under the reaction conditions we employed, Pfu catalyzed an unacceptable level of strand displacement synthesis. We pursued additional polymerases and tried Phusion because the supplier (New England Biolabs) indicated it did not strand displace. We observed moderate strand displacement (<150 nt) above one unit of polymerase per 20 l reaction, but not at lower levels ( Figure S9). We thus used 0.2 U of Phusion for gap filling of purified products resulting from rolling circle replication.
After the rolling circle product from each 25 l reaction was extracted with phenol-chloroform and precipitated with isopropanol, it was incubated with 100 M dNTPs, 0.2 U Phusion polymerase, and 32 P-dATP (2 Ci/reaction) at 72°C for 15 min. Gap filling products with 0.2 U and 1 U of Phusion were compared side by side, and no difference was found. According to the supplier, Phusion extends DNA at a rate of 15-30 s/kb, which is 10 times faster than Pfu. Since 30 min was chosen for gap filling reactions with Pfu (25), 15 min should be long enough to fill all gaps with Phusion.
Normalizing the lengths of gap-filled Okazaki fragments requires the same specific activity of the radioactive nucleotides as in the original Okazaki fragments with gaps between them. Without addition of radioactivity in the gap filling reaction, the lengths might be biased towards the lower molecular weight. However, we also need to make sure that not too many unused primers are elongated during gap filling, which may obscure the true size of gaps. Therefore, a comparison of gap filling was made in the presence and absence of the same amount of 32 P-dATP as the rolling circle reaction. No significantly different lengths in two conditions were found, which relieved our concerns ( Figure S9).

Determination of primer utilization and the amounts of GMP+UMP per Okazaki fragment
Rolling circle reactions were carried out in the presence of -[ 32 P] GTP and -[ 32 P] UTP (specific activity 12,000 cpm/pmol) or -[ 32 P] dCTP (specific activity 400 cpm/pmol). The sample was loaded onto a 20% denaturing polyacrylamide gel in 50% w/v urea. -[ 32 P] ATP labeled 12-mer RNA and 20-mer DNA were loaded as markers. The gel was prerun at 14 W for 30 min and then run at 12 W for 3.5 h. The gel was dried directly with no pre-treatment on DEAE paper, exposed to a phosphorimage screen for 18 hours, and scanned with a PhosphorImager. DNA fragments of 10-14 nt in length were quantified as free primers, and the dark bands at the top of the gel were quantified as elongated primers.
The quantification from scintillation counting was directly proportional to the pixel density from phosphorimaging. Thus, 1 l of -[ 32 P] G/UTP and 1 l of -[ 32 P] dCTP were spotted on a GFC filter paper and the counts were determined with a scintillation counter. The specific activity of -[ 32 P] G/UTP and -[ 32 P] dCTP were calculated by the count divided by the amount of each nucleotide in a reaction. A