To obtain ssDNA, we chose lambda exonuclease as it has processive 5′→3′ exoDNase activity selectively digesting the 5′-phosphorylated strand of dsDNA with 20 times more affinity for a 5′-phosphorylated terminus than a 5′-hydroxylated terminus (44). Although lambda exonuclease selectively digest its preferred substrate, the 5′-phosphate group, it also possesses very low background activity on single-stranded and non-phosphorylated DNA templates. Thus, it is crucial to optimize enzyme concentration and incubation time to generate sufficient single-stranded material.
For enzyme optimization, 1 μg of a dsDNA template phosphorylated at only one end was incubated with 5 U or 10 U of lambda exonuclease at 37°C for either 30 min or 1 h. Analysis on a 1% agarose gel showed that 5 U enzyme was insufficient to produce an adequate amount of single-stranded material as a strong residual band of dsDNA was observed. In contrast, the amount of ssDNA template generated using 10 U of enzyme was greater with a lower band intensity of dsDNA template observed. Variations in incubation time (30 min and 1 hr) did not show any difference in the amount of ssDNA product generated (Figure 3A). In the end, we determined the best conditions for ssDNA generation were 1 μg of dsDNA template incubated with 10 U of lambda exonuclease at 37°C for 30min. After digestion with lambda exonuclease, all single-stranded templates have free 3′-OH overhangs for filling in reaction.
To further confirm the successful generation of ssDNA using our optimized lambda exonuclease digest method, PAGE analyses and fluorescence measurements in the presence of SYBR green were performed. Difference between double-strand and single-strand DNA could be clearly observed using 6 M urea PAGE (Figure 3B) as ssDNA shows faster migration (45). Hence, lambda exonuclease successfully generates sufficient ssDNA using the optimized conditions. As additional confirmation, we generated fluorescence measurements in the presence of SYBR green. SYBR green strongly binds dsDNA molecules with weaker binding to ssDNA, emitting a strong fluorescence when bound to dsDNA (46). We determined the relative fluorescence reading of SYBR green alone to be 0.2241 rfu. Control samples of oligo20 dsDNA and oligo20 ssDNA gave intensity values of 0.7540 rfu and 0.3821 rfu, respectively. The dsDNA sample treated with lambda exonuclease produced a fluorescence reading of 0.3600 rfu (Figure 3C). This drop in fluorescence signal upon lambda exonuclease treatment to levels comparable to the control ssDNA oligo further confirms the generation of ssDNA template in agreement with the urea PAGE gel results.Extension of ssDNA using Klenow fragment
To anneal ssDNA fragments prior to Klenow treatment, a mixture of 400 ng of lambda exonuclease digested ssDNA molecules of VH and VL with complementary overlapping ends was heated at 90°C for 5 min and cooled to 30°C for reannealing. This mixture was then treated with KF, which is able to synthesize new complementary DNA strands in the presence of dNTPs (47). KF works by adding bases to the 3′ hydroxyl groups of a blunt-ended DNA duplex, working in a 5′ to 3′ direction (48). To determine the optimal conditions for synthesis of dsDNA, the double-stranded products of different KF titrations were evaluated on a 1% agarose gel.
10 U and 20 U of KF were evaluated for either 4 h or 8 h at 37°C with 400 ng of ssDNA templates. The amount of dsDNA generated with 10 U was insufficient to complete the synthesis of full dsDNA (data not shown).
For 20 U enzyme, a comparison between 4 h and 8 h incubation at 37°C was also tested (Figure 4). Treatment with 20 U enzyme for 4 h and 8 h did not yield a significant difference in the amount of dsDNA template generated.
When applying non-cycling conditions, we did note a significant generation of dsDNA from the individual 350 bp ssDNA templates (Figure 4). We rationalized that this increase in band intensity was likely caused by residual dsDNA from the exonuclease reaction. We examined this effect by incubating VH and VL ssDNA templates separately with 20 U of KF for 4 h. Similar amounts of ssDNA template (400 ng) were used, and band intensity of KF treated samples visualized on a 1% agarose gel increased (Figure 5A). This demonstrates the presence of residual dsDNA template after exonuclease treament. However, we hypothesized that even with a small amount of leftover dsDNA template, an increase in annealing efficiency would reduce background dsDNA without the need for sample clean up.