ABSTRACT
DNA restriction fragments, 120-650 base pairs (bp) in length, with 5'-GCGC-3', 5'-GGCC-3' or 3'-GCGC-5' single-stranded overhanging termini, give rise to diffuse bands of unusual electrophoretic mobility in non-denaturing polyacrylamide gels. This shift in electrophoretic mobility can be observed at 4-12°C, not at higher temperatures, but is stabilized by 5-10 mM Mg2+, even at 37°C. The nucleotide sequence in the abutting double-stranded part of the fragment does not affect this phenomenon, which is not caused by dimerization. The altered mobility may be due to the unusual terminal DNA structure, which is dependent on co-operative interactions among more than two neighboring G and C residues. The structure is stabilized by cytidine methylation. The biological role of such fragment structures in DNA repair and recombination is presently unknown.
Current theories of DNA electrophoresis are based on the reptation model that postulates that the uniformly charged, flexible DNA chain moves in a snake-like fashion between the fibers of a gel (1-3). DNA molecules are thought to migrate through the matrix with their leading end entering new segments of tubules, which are presumably formed by gel pores, and abandoning these segments at the trailing end. The electric field promotes the movement of the leading DNA segment in a downfield direction. The DNA molecule follows this path, overcoming the friction created by the interactions with the gel fibers (4-6). DNA molecules with sequence-dependent modulations, such as intrinsically curved, cruciform or gapped fragments, migrate in a non-denaturing polyacrylamide gel more slowly than straight molecules (7-9).
The greatest deviation in migration for curved DNA fragments is observed when the bending sequence is located near the center of the molecule (10,11). The reptation model of DNA electrophoresis implies that the ends and the stem of the molecule interact with the gel matrix in different ways. It is assumed that the polyacrylamide gel matrix, with the pore diameters smaller than the persistent length of the DNA molecule (12), accommodates deformations more easily for molecules bent near the end than for those bent near the center (13,14).
The impact of the DNA fragment end structures upon their performance in gel electrophoresis has not yet been studied. Here we demonstrate that DNA fragments with four bases, namely 5'-GGCC-3', 5'-GCGC-3' or 3'-CGCG-5', in single-stranded overhang sequences produce specific migration anomalies during polyacrylamide gel electrophoresis (PAGE).
The construction of the plasmids pd364, pd412 and pd490 was described elsewhere (15). Plasmid pd364 contained a tandem dimer of the 364 bp PstI fragment from the adenovirus type 2 (Ad2) late E2A gene promoter, which had been cloned into the PstI site of the vector pBS(+) (Stratagene, La Jolla, CA). This promoter DNA contained a unique ApaI-Bsp120I site. The cutting of pd364 with these enzymes created a 364 bp fragment with the permuted sequence. The plasmid pd490 contained the tandem dimer of the 490 bp HindIII fragment of the Ad2 late E2A promoter cloned into the plasmid pGL2-Basic (Promega, Madison, WI) (15). The promoter DNA contained a unique ApaI-Bsp120I site. Cleavage by these enzymes generated a 490 bp fragment with the permuted promoter sequence. Plasmid pd412 contained a tandem dimer of the 412 bp RsaI fragment from the 5'-flanking region of the human angiogenin gene, which had been cloned into the SmaI site of pBS(+) (15). This fragment carried a unique restriction site for NarI, which was also cleaved by BanI and HaeII. Additional recognition sites for BanI and HaeII were located in the vector DNA.
For restriction, 5 µg of Qiagen (Hilden, Germany) midiprep column-purified plasmid DNA were cleaved for 2 h with 20 U of different restriction endonucleases (60 U of BanI) in a 100 µl volume under the conditions recommended by the manufacturers. Subsequently, the sample was divided into two equal portions of which one was treated for 30 min with 5 U of calf intestine alkaline phosphatase (CIAP; Boehringer-Mannheim, Mannheim, Germany). Both samples were then heat-inactivated for 20 min at 72°C, phenol-chloroform and chloroform extracted, reprecipitated and dissolved in 25 µl of de-ionized water. Samples of 3-5 µl were applied per lane in the polyacrylamide gel for subsequent electrophoresis.
Restriction enzymes were purchased from New England Biolabs (Beverly, MA) or from MBI Fermentas (Vilnius, Lithuania). The isoschizomers Bsp120I and ApaI cleaved at 5'-GGGCCC-3' sequences generating four-base 5'- or 3'-overhangs, respectively. HaeII cuts the sequence 5'-PuGGCCPy-3' producing four-base 3'-overhangs. BanI cuts the sequence 5'-GGPyPuCC-3' and yields four-base 5'-overhangs. Both enzymes recognize 5'-GGCGCC-3' sequences, as do KasI, EheI, NarI. The Klenow fragment of the Escherichia coli DNA polymerase I was purchased from Boehringer Mannheim, and ultrapure dNTPs from Pharmacia (Uppsala, Sweden).
Standard 150 × 220 × 1.2 mm 10% polyacrylamide gels (acrylamide:bisacrylamide ratio 39:1), in 40 mM Tris-acetate, pH 8.0 and 1 mM EDTA (TAE), were poured. The vertical electrophoresis chamber was placed into a thermostated environment and pre-electrophoresed for a minimum of 3 h at 210-240 V until current and temperature in the gel had stabilized. After application of the samples, the gels were run at 210-240 V for 12-14 h. During pre-electrophoresis and separation, the buffer was continuously recycled. After electrophoresis, the gels were stained for 10 min in 1 µg of ethidium bromide per ml TAE, destained for 5 min in water and photographed on a shortwave UV transilluminator with a Polaroid camera using an orange filter.
Rc values and apparent length were determined in two to five independently run gels. RL values did not fluctuate by more than ± 0.02; apparent length measurements remained within a margin of ±2%.
Migration deviations of restriction fragments attracted our attention during the analysis of plasmid pd412 fragments by PAGE. The plasmid was cleaved with both BanI and HindIII endonucleases. BanI generated fragments with different terminal overhang sequences in the same restriction reaction. When the 412 bp fragment of pd412 was electrophoresed without prior phosphatase treatment, a fuzzy band was observed (Fig. 1a). The 412 bp fragment carried a 5'-GCGC-3' overhang on both termini. Dephosphorylation of the 5'-ends by CIAP further changed the shape and mobility of this fragment (Fig. 1a, lane 2). The 243 and 204 bp fragments, which carried the 5'-GCGC-3' overhangs on one terminus only, yielded fuzzy bands with reduced electrophoretic mobility in comparison with the DNA with phosphorylated termini. The treatment with alkaline phosphatase did not affect the appearance of fragments without 5'-GCGC-3' overhangs. The smaller dephosphorylated fragment of 136 bp migrated slightly more slowly than its phosphorylated counterpart, possibly due to the relatively big loss in net charge.
Next, we tested to what extent the anomaly in electrophoretic migration was dependent on the nucleotide sequence of the single-stranded overhangs of various restriction fragments of plasmid DNA. These fragments were investigated for electrophoretic mobility prior to, and subsequent to, CIAP treatment. In total, 30 different restriction endonucleases were tested, which generated the following fragments.
Fragments with four-base 5'-overhangs: AATT (EcoRI, MunI), AGCT (HindIII), GATC (BamHI), CCGG (XmaI), CGCG (BssHI, MluI), CWWG (StyI), GCGC (KasI), GGCC (Bspl120I, EagI, EaeI), GYRC (BanI), TCGA (XhoI), TRYA (SfcI).
Fragments with four-base 3'-overhangs: ACGT (AatII), AGCT (SacI), CATG (SphI), GGCC (ApaI), GTAC (KpnI), GYRC (HaeII), TGCA (NsiI, PstI).
Fragments with two-base 3'-GC overhangs (SacII) and 5'-CG overhangs (HpaII, NarI).
Fragments with three-base 5'-overhangs: GWC (AvaII) and GNC (EcoO109I); blunt ends (Ecl136II, EheI, SmaI, HaeIII).
Four types of electrophoretic mobility could be assigned to these classes of fragments, respectively. (i) No migration anomaly. The dephosphorylated and phosphorylated fragments exhibited ratios of <= 1.02 ± 0.01 in their apparent migration rates. The fragments with four-base 5'-CCGG (XmaI) or 3'-AGCT (SacI) overhangs migrated identically to blunt-ended fragments, which were generated by their isoschizomers SmaI or Ecl136II, respectively. Fragments with A- or T-residues in the single-stranded overhang sequence or with overhangs of less than four bases showed no migration anomaly. (ii) Marginal migration anomaly. Phosphorylated fragments generated by Bsp120I or EagI (5'-GGCC overhang) exhibited a migration ratio in the gel of >= 1.02, compared with the same fragments generated by the isoschizomer ApaI (3'-GGCC overhang) or by the enzyme HaeIII, which removed the overhanging sequences. (iii) Pronounced migration anomaly. The CIAP-treated fragments with 5'-GGCC overhangs generated by Bsp120I migrated more slowly than the phosphorylated fragments and yielded fuzzy bands. The CIAP-untreated fragments with 5'- and 3'-GCGC overhangs in the BanI and HaeII cleavage patterns yielded fuzzy bands. (iv) Strong migration anomaly in electrophoretic mobility. The CIAP-treated fragments gave rise to extremely diffuse bands, e.g. the BanI fragments, which carried 5'-GCGC-3' overhangs (compare with Fig. 1b).
Figure 1b presents typical examples of the fragment overhang sequence-dependent migration anomalies with various DNA samples. The 364 bp fragment in lanes 1-4 was generated by ApaI or Bsp120I cleavage. The 336 bp fragment resulted from the double cleavages with ApaI and HindIII or Bsp120I and HindIII. Dephosphorylation affected the mobility of the 364 bp Bsp120I fragment only, which carried 5'-GGCC overhangs on both ends (Fig. 1b, lane 4). Mobility and appearance of the fragments that carried 3'-GGCC overhangs or 5'-GGCC overhangs only on one teminus remained unaltered by dephosphorylation. The 481, 412 and 370 bp bands (Fig. 1b, lanes 5 and 6) carried 3'-GCGC overhangs on both ends, whereas the 270 and 243 bp bands had 3'-GCGC overhangs on one end and 5'-AGCT overhangs on the opposite end. Dephosphorylation caused fuzzy borders and reduced the mobility of all fragments. The 370 bp plasmid fragment (Fig. 1b, lanes 5 and 6) contained the origin of replication of the cloning vector pBS(+) with a strong, sequence-dependent DNA curvature. Under the separation conditions applied, the curvature of the fragments reduced their electrophoretic mobility in polyacrylamide gels, e.g. the curved 370 bp fragment migrated more slowly than the 395 bp fragment, which does not contain a static bend (Fig. 1, lanes 7-9).
The 412 bp BanI and HaeII fragments of pd412 were identical in their sequences, unless the first enzyme had generated 5'-overhangs and the second 3'-overhangs. The deviation in migration was more apparent for 5'-GCGC than for 3'-GCGC overhangs (Fig. 1b, compare lanes 5 and 7 with lanes 6 and 8). Filling in the first nucleotide in the single-stranded four-base overhang generated by BanI, abolished the deviation in electrophoretic mobility (Fig. 1b, lane 9).
One might surmise that the anomalies in fragment migration were due to the di- or multimerization of DNA fragments via their cohesive ends. In this case, the electrophoretic migration anomaly of the fragments with specific overhang sequences should disappear in diluted samples. The control experiments, described as follows, demonstrated that this interpretation was not valid. The plasmid pd412 DNA was cleaved with BanI and HindIII, and pd364 DNA with Bsp120I and HindIII. Subsequently, the generated fragments were dephosphorylated and 5'-labeled with polynucleotide kinase and [[gamma]-32P]ATP. In some of the experiments, the 5'-overhang was partly filled in with dGTP using Klenow polymerase. The fragments were separated by electrophoresis in polyacrylamide gels using buffer without EDTA. The data presented in Figure 2 demonstrated the migration anomaly for the 412 bp fragment with a 5'-GCGC overhang and for the 364 bp fragment with a 5'-GGCC overhang on both ends (lanes 2 and 4, arrows). This deviation in migration was still apparent when the DNA samples were diluted 7-fold (Fig. 2B) or 50-fold (Fig. 2C), as compared with the undiluted DNA samples (Fig. 2A).
The aberrant electrophoretic mobility was still apparent at 12-14°C but vanished at higher temperatures, as demonstrated at 37°C (Fig. 1c). The dephosphorylated and phosphorylated fragments migrated equally as sharp bands. Elevated temperatures also diminished the migration anomalies of the bent DNA fragments that were 336, 364, 370 and 412 bp in length. These fragments migrated as expected for their lengths when the polyacrylamide gels were electrophoresed at 55°C (data not shown). Magnesium ion, NaCl and polyacrylamide concentrations also influenced the DNA overhang sequence and curvature-dependent migration anomaly (Figs 1d and 2).
The presence of Mg2+ ions in the electrophoresis buffer stabilized the structures responsible for the migration anomalies of 5'-phosphorylated fragments with specific single-stranded termini. The CIAP-untreated 364 bp Bsp120I fragment with 5'-GGCC overhangs of plasmid pd364 (Fig. 1d, lane 3) migrated significantly more slowly than its dephosphorylated counterpart (lane 4). The phosphorylation status did not influence the gel electrophoretic mobility of the same fragment with a 3'-GGCC overhang (Fig. 1d, lanes 1 and 2). Fragments with HaeII 3'-GCGC-5' overhangs (lanes 5 and 6) and the BanI fragments with 5'-GCGC-3' overhangs (lanes 7 and 8) also showed phosphorylation-dependent, reduced electrophoretic mobility in the presence of Mg2+ ions. At 37°C, the migration anomaly was already apparent at 2 mM Mg2+ concentrations (Fig. 3a). The influence of the Mg2+ concentration upon the migration of the bent DNA fragments was shown to be sequence-dependent (16-19). The 364 bp ApaI fragment of plasmid pd364 was also curved (15), and increased Mg2+ concentrations retarded its electrophoretic mobility regardless of the terminal sequence configuration. However, the migration anomaly of the phosphorylated Bsp120I fragment exceeded the average reduction of the migration rate due to the intrinsic DNA curvature (Fig. 3a). The phosphorylation and Mg2+ dependence of the single-stranded termini-dependent migration anomalies emphasized the importance of charge at the DNA fragment termini for their electrophoretic mobilities.
Like the formation of Z-DNA (23-25), the generation of DNA structures which produce single-stranded termini-dependent migration anomalies was enhanced by the methylation of the cytosine-residues in the 5'-CG-3' dinucleotides. Comparisons of the electrophoretic migration rates between CIAP-treated methylated and unmethylated DNA fragments showed the migration anomaly to be enhanced by the methylation of C-residues in the overhangs (Fig. 1f). The methylated 412, 243 and 204 bp fragments were more delayed and more heterogeneous in their apparent lengths (Fig. 1f, lanes 2 and 4) than their unmethylated counterparts (Fig. 1f, lane 6). DNA methylation by the DNA methyltransferases M-HhaI (5'-GCGC-3') and M-SssI (5'-CG-3') yielded the same results, although the latter enzyme introduced many more methyl groups.
Overhangs with identical lengths and nucleotide sequences but different flanking nucleotides in the double-stranded part of the molecule were generated by cleavage with Bsp120I (G <=> GGCCC), EaeI (Py <=> GGCCPu), EagI (C <=> GGCCG) or HaeII (PuGCGC <=> Py). The results of a series of experiments demonstrated that the nucleotide pair flanking the single-stranded termini of the DNA fragment influenced the migration anomaly of the fragment.
EagI fragments with the 3'-terminal C-residue in the double-stranded part of the fragment opposing the first C in the overhang, showed only minor migration deviations when compared with the Bsp120I fragments in which a 3'-terminal G complements the first C in the overhang. Dephosphorylated EaeI fragments produced easily detectable shifts when the 3'-T in the double-stranded part of the fragment opposed the first C in the overhang. However, the shifts were much weaker for the fragments with 3'-C residues. Migration anomalies of the HaeII fragments were most apparent with 5'-terminal C-residues in the double-stranded part of the fragment opposing the first G in the overhang, but weaker with 5'-T-residues. Consequently, the migration anomaly was apparent when the first nucleotide in the overhang was opposed by a complementary nucleotide at the end of the opposite DNA strand (Table 1).
The overall nucleotide sequence of the double-stranded DNA fragments apparently did not influence the expression of the migration anomalies. DNA fragments with 5'-GCGC and 5'-GGCC overhangs in the vector plasmid pBR322 and in the RET proto-oncogene promoter generated identical aberrations in band appearance and electrophoretic migration rates (data not shown). In the range of 100-500 bp fragments, the migration anomalies for the larger molecules were more pronounced (Fig. 3d).
Table 1.
I.M. was a fellow of the Alexander-von-Humboldt Foundation, Bonn, Germany and on leave from the University of Latvia. This research was supported by the Deutsche Forschungsgemeinschaft through grant SFB274-A1 and by the Latvian Council of Science through grant 96.0621.
Nucleic Acids Research
Pages
Introduction
Materials And Methods
Plasmid DNA
Phosphatase treatment of DNA fragments
Restriction endonucleases
Detection of migration aberrations by electrophoresis in polyacrylamide gels
Results
The structure of restriction fragment termini and electrophoretic mobility
Dependence on the nucleotide sequence of the single-stranded overhangs
Migration delay is not caused by multimerization of fragments
Temperature, ion and gel composition effects
Consequences of DNA methylation of the 5'-CG-3' sequences in the termini
Structure of the single-stranded termini and impact of the adjacent nucleotide pair
Discussion
Acknowledgements
References
Restriction enzymes
Fragment end structure
Mobility anomaly
EagI
5'-GGCCG-
3'-Cmarginal/none
Bsp120I
5'-GGCCC-
3'-Gpronounced
EaeI
5'-GGCCG-
3'-C-marginal/none
EaeI
5'-GGCCA-
3'-T-pronounced
HaeII
-GGCGC-3'
-C-5'pronounced/strong
HaeII
-AGCGC-3'
-T-5'pronounced
REFERENCES
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 27 Mar 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Q. Liu, W. A. Scaringe, and S. S. Sommer Discrete mobility of single-stranded DNA in non-denaturing gel electrophoresis Nucleic Acids Res., February 15, 2000; 28(4): 940 - 943. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||