Is nitrocellulose filter binding really a universal assay for protein-DNA interactions?

The ability to bind to nitrocellulose is commonly accepted as being a universal property of proteins and has been widely used in many different fields of study. This property was first exploited in the study of DNA-binding proteins 30 years ago, in studies involving DNA binding by the lactose repressor (LacR) of Escherichia coli. Termed the filter-binding assay, it remains the quickest and easiest assay available for the study of protein-DNA interactions. However, the exact mechanism by which proteins bind to nitrocellulose remains uncertain. Given the supposedly universal nature of the interaction, we were surprised to notice that certain LacR variants were completely unable to bind simultaneously to DNA containing a single lac operator and nitrocellulose. Investigation of this loss of binding suggests that LacR requires a protein region that is both hydrophobic in nature and more or less unstructured, in order to bind to both nitrocellulose and DNA. In the case of wild-type, tetrameric LacR, the DNA-recognition domain that is not bound to DNA suffices. Dimeric LacR variants will only bind if they have certain C-terminal extensions. These experiments sound a cautionary note for the use of filter binding as an assay of choice, particularly in applications involving screening for the DNA-binding site of putative DNA-binding proteins.     

DNA-protein interactions at the telomeric repeats of Schizosaccharomyces pombe.

Gel retardation assays using a probe containing the repeat region of a Schizosaccharomyces pombe chromosomal telomere identified four specific DNA- protein complexes in S. pombe total protein extracts (I, I', IIa and IIb). The proteins responsible for these complexes bound to the telomeric repeat region irrespective of whether or not the repeats were in close proximity to the end of a DNA molecule, and none of them bound strongly to single-stranded DNA. The protein responsible for complex I (TeRF I) was separated from the activity responsible for complexes IIa and IIb (TeRF II) using heparin-Sepharose chromatography. Both factors were efficiently cross-competed by an oligonucleotide containing the 18 bp sequence 5'-GGTTACAGGTTACAGGTT-3', which corresponds to two complete telomeric repeat units. Mutation of the T residues at positions 4 and 11 in the oligonucleotide dramatically reduced binding by TeRF II, but had no affect on binding by TeRF I. The protein responsible for complex I' did not bind strongly to either the wild-type or mutant oligonucleotide.     

DNA binding activities of c-Myc purified from eukaryotic cells.

c-Myc is a nuclear phosphoprotein which contains both a leucine zipper and a helix-loop-helix dimerization motif. These are adjacent to a basic region believed to make specific contacts with DNA upon dimerization. We report the purification of full-length c-Myc to near homogeneity from two independent eukaryotic systems: the baculovirus overexpression system using an insect cell host, and Chinese hamster ovary cells containing heat-inducible c-myc genes. The DNA binding capabilities of these preparations were characterized. Both preparations contain two distinct activities that bind specifically to sequences with a core of CACGTG. The Myc protein is solely responsible for one of these binding activities. Specific sequences that bound to c-Myc were selected from a large pool of random DNA sequence. Sequencing of individual binding sites selected by this procedure yielded a 12-base consensus, PuACCACGTGCTC, for c-Myc binding. Both protein preparations additionally demonstrated a distinct complex, containing both c-Myc and a copurifying 26-29-kDa protein, that bound to DNA with higher affinity than Myc alone. Selection of specific DNA sequences by this complex revealed a consensus binding site similar to the 12-base consensus described above. These data demonstrate that c-Myc isolated from eukaryotic cells is capable of sequence-specific DNA binding and further refine the optimal sequence for c-Myc binding. These protein preparations should prove useful in further characterizing the biochemical properties of c-Myc.     

Interaction of wild-type and mutant adeno-associated virus (AAV) Rep proteins on AAV hairpin DNA.

Both the Rep68 and Rep78 proteins of adeno-associated virus type 2 (AAV) bind to AAV terminal repeat hairpin DNA and can mediate site-specific nicking in vitro at the terminal resolution site (trs) within the terminal repeats. To define the regions of the Rep proteins required for these functions, a series of truncated Rep78 derivatives was created. Wild-type and mutant proteins were synthesized by in vitro translation and analyzed for AAV hairpin DNA binding, trs endonuclease activity, and interaction on hairpin DNA. Amino-terminal deletion mutants which lacked the first 29 or 79 amino acid residues of Rep78 did not bind hairpin DNA, which is consistent with our previous identification of a DNA-binding domain in this region. Progressive truncation of the carboxyl-terminal region of Rep78 did not eliminate hairpin DNA binding until the deletion reached amino acid 443. The electrophoretic mobility of the Rep-specific protein-DNA complexes was inversely related to the molecular weight of the Rep derivative. Analysis of the C-terminal deletion mutants by the trs endonuclease assay identified a region (amino acids 467 to 476) that is essential for nicking but is not necessary for DNA binding. When endonuclease-positive, truncated Rep proteins that bound hairpin DNA were mixed with full-length Rep78 or Rep68 protein in electrophoretic mobility shift assays, a smear of protein-DNA complexes was observed. This smear migrated at an intermediate position with respect to the bands generated by the proteins individually. An antibody recognizing only the full-length protein produced a novel supershift band when included in a mixed binding assay containing Rep68 and a truncated Rep mutant. These experiments suggest that the Rep proteins can form hetero-oligomers on the AAV hairpin DNA.     

Mapping of a DNA binding region of the PI-sceI homing endonuclease by affinity cleavage and alanine-scanning mutagenesis.

The PI-SceI protein is a member of the LAGLIDADG family of homing endonucleases that is generated by a protein splicing reaction. PI-SceI has a bipartite domain structure, and the protein splicing and endonucleolytic reactions are catalyzed by residues in domains I and II, respectively. Structural and mutational evidence indicates that both domains mediate DNA binding. Treatment of the protein with trypsin breaks a peptide bond within a disordered region of the endonuclease domain situated between residues Val-270 and Leu-280 and interferes with the ability of this domain to bind DNA. To identify specific residues in this region that are involved in DNA binding and/or catalysis, alanine-scanning mutagenesis was used to create a set of PI-SceI mutant proteins that were assayed for activity. One of these mutants, N281A, was >300-fold less active than wild-type PI-SceI, and two other proteins, R277A and N284A, were completely inactive. These decreases in cleavage activity parallel similar decreases in substrate binding by the endonuclease domains of these mutant proteins. We mapped the approximate position of the disordered region to one of the ends of the 31 base pair PI-SceI recognition sequence using mutant proteins that were substituted with cysteine at residues Asn-274 and Glu-283 and tethered to the chemical nuclease FeBABE. These mutational and affinity cleavage data strongly support a model of PI-SceI docked to its DNA substrate that suggests that one or more residues identified here are responsible for contacting base pair A/T(-)(9), which is essential for substrate binding.     

DNA-binding domain of bovine papillomavirus type 1 E1 helicase: structural and functional aspects.

The E1 protein of bovine papillomavirus type 1 is a multifunctional enzyme required for papillomaviral DNA replication. It assists in the initiation of replication both as a site-specific DNA-binding protein and as a DNA helicase. Previous work has indicated that at limiting E1 concentrations, the E2 protein is required for efficient E1 binding to the replication origin. In this study, we have defined the domain of the E1 protein required for site-specific DNA binding. Experiments with a series of truncated proteins have shown that the first amino-terminal 299 amino acids contain the DNA-binding domain; however, the coterminal M protein, which is homologous to E1 for the first 129 amino acids, does not bind origin DNA. A series of small internal deletions and substitution mutations in the DNA-binding domain of E1 show that specific basic residues in this region of the protein, which are conserved in all E1 proteins of the papillomavirus family, likely play a direct role in binding DNA and that a flanking conserved hydrophobic subdomain is also important for DNA binding. A region of E1 that interacts with E2 for cooperative DNA binding is also retained in carboxy-terminal truncated proteins, and we show that the ability of full-length E1 to complex with E2 is sensitive to cold. The E1 substitution mutant proteins were expressed from mammalian expression vectors to ascertain whether site-specific DNA binding by E1 is required for transient DNA replication in the cell. These E1 proteins display a range of mutant phenotypes, consistent with the suggestion that site-specific binding by E1 is important. Interestingly, one E1 mutant which is defective for origin binding but can be rescued for such activity by E2 supports significant replication in the cell.     

Mutational analysis of an extracytoplasmic-function sigma factor to investigate its interactions with RNA polymerase and DNA

The extracytoplasmic-function (ECF) family of sigma factors comprises a large group of proteins required for synthesis of a wide variety of extracytoplasmic products by bacteria. Residues important for core RNA polymerase (RNAP) binding, DNA melting, and promoter recognition have been identified in conserved regions 2 and 4.2 of primary sigma factors. Seventeen residues in region 2 and eight residues in region 4.2 of an ECF sigma factor, PvdS from Pseudomonas aeruginosa, were selected for alanine-scanning mutagenesis on the basis of sequence alignments with other sigma factors. Fourteen of the mutations in region 2 had a significant effect on protein function in an in vivo assay. Four proteins with alterations in regions 2.1 and 2.2 were purified as His-tagged fusions, and all showed a reduced affinity for core RNAP in vitro, consistent with a role in core binding. Region 2.3 and 2.4 mutant proteins retained the ability to bind core RNAP, but four mutants had reduced or no ability to cause core RNA polymerase to bind promoter DNA in a band-shift assay, identifying residues important for DNA binding. All mutations in region 4.2 reduced the activity of PvdS in vivo. Two of the region 4.2 mutant proteins were purified, and each showed a reduced ability to cause core RNA polymerase to bind to promoter DNA. The results show that some residues in PvdS have functions equivalent to those of corresponding residues in primary sigma factors; however, they also show that several residues not shared with primary sigma factors contribute to protein function.     

Nuclear proteins from Drosophila melanogaster embryos which specifically bind to simple homopolymeric sequences poly [(dT-dG).(dC-dA)]

The binding of nuclear proteins from Drosophila melanogaster embryos to simple homopolymeric DNA sequences was studied. Nuclear proteins were electrophoresed, transferred onto nitrocellulose and incubated with labelled synthetic homopolymers or natural fragment containing simple sequences. Several protein bands were found in the 65-72 KDa region, which specifically bind both poly [(dG-dT).(dA-dC)] and a natural fragment containing 40 bp of this sequence. These proteins do not bind to homopolymers poly [(dA).(dT)] and poly [(dG-dA).(dC-dT)], or other foreign DNAs.     

Identification and characterization of Saccharomyces cerevisiae Cdc6 DNA-binding properties

Recent studies have shown that Cdc6 is an essential regulator in the formation of DNA replication complexes. However, the biochemical nature of the Cdc6 molecule is still largely unknown. In this report, we present evidence that the Saccharomyces cerevisiae Cdc6 protein is a double-stranded DNA-binding protein. First, we have demonstrated that the purified yeast Cdc6 can bind to double-stranded DNA (dissociation constant approximately 1 x 10(-7) M), not to single-stranded DNA, and that the Cdc6 molecule is a homodimer in its native form. Second, we show that GST-Cdc6 fusion proteins expressed in Escherichia coli bind DNA in an electrophoretic mobility shift assay. Cdc6 antibodies and GST antibodies, but not preimmune serum, induce supershifts of GST-Cdc6 and DNA complexes in these assays, which also showed that GST-Cdc6 binds to various DNA probes without apparent sequence specificity. Third, the minimal requirement for the binding of Cdc6 to DNA has been mapped within its N-terminal 47-amino acid sequence (the NP6 region). This minimal binding domain shows identical DNA-binding properties to those possessed by full-length Cdc6. Fourth, the GST-NP6 protein competes for DNA binding with distamycin A, an antibiotic that chelates DNA within the minor groove of the A+T-rich region. Finally, site-direct mutagenesis studies revealed that the (29)KRKK region of Cdc6 is essential for Cdc6 DNA-binding activity. To further elucidate the function of Cdc6 DNA binding in vivo, we demonstrated that a binding mutant of Cdc6 fails to complement either cdc6-1 temperature-sensitive mutant cells or Deltacdc6 null mutant cells at the nonpermissive temperature. The mutant gene also conferred growth impairments and increased the plasmid loss in its host, indicative of defects in DNA synthesis. Because the mutant defective in DNA binding also fails to stimulate Abf1 ARS1 DNA-binding activity, our results suggest that Cdc6 DNA-binding activity may play a pivotal role in the initiation of DNA replication.     

Nonspecific DNA binding activity of simian virus 40 large T antigen: evidence for the cooperation of two regions for full activity.

We generated a series of COOH-terminal truncated simian virus 40 large tumor (T) antigens by using oligonucleotide-directed site-specific mutagenesis. The mutant proteins [T(1-650) to T(1-516)] were expressed in insect cells infected with recombinant baculoviruses. T(1-623) and shorter proteins [T(1-621) to T(1-516)] appeared to be structurally changed in a region between residues 269 and 522, as determined by increased sensitivities to trypsin digestion and by altered reactivities to several monoclonal antibodies. These same mutant proteins bound significantly less nonorigin plasmid DNA (15%) and calf thymus DNA (25%) than longer proteins [T(1-625) to T(1-708)]. However, all mutant T antigens exhibited a nearly wild-type level of viral origin-specific DNA binding and binding to a helicase substrate DNA. This indicated that binding to origin and helicase substrate DNAs is separable from about 85% of nonspecific binding to double-stranded DNA. As an independent confirmation that a region distinct from the origin-binding domain (amino acids 147 to 247) is involved in nonspecific DNA binding, we found that up to 96% of this latter activity was specifically inhibited in wild-type T antigen by several monoclonal antibodies which collectively bind to the region between residues 269 and 522. In order to investigate the relationship between the origin-binding domain and the second region, we performed origin-specific DNA binding assays with increasing amounts of calf thymus DNA as competitor. The results suggest that this second region is not an independent nonspecific DNA binding domain. Rather, it most likely cooperates with the origin-binding domain to give rise to wild-type levels of nonspecific DNA binding. Our results further suggest that most of the nonspecific binding to double-stranded DNA is involved in a function other than direct recognition and binding to the pentanucleotides at the replication origin on simian virus 40 DNA.     

A mutation in the C-terminal putative Zn2+ finger motif of UL52 severely affects the biochemical activities of the HSV-1 helicase-primase subcomplex

Herpes simplex virus type 1 encodes a heterotrimeric helicase-primase complex that is composed of the products of the UL5, UL52, and UL8 genes. A subcomplex consisting of the UL5 and UL52 proteins retains all the enzymatic activities exhibited by the holoenzyme in vitro. The UL52 protein contains a putative zinc finger at its C terminus which is highly conserved among both prokaryotic and eukaryotic primases. We constructed a mutation in which two highly conserved cysteine residues in the zinc finger motif were replaced with alanine residues. A UL52 expression plasmid containing the mutation in the zinc finger region is unable to support the growth of a UL52 mutant virus in a transient complementation assay. Wild type and mutant UL5.UL52 subcomplexes were purified from insect cells infected with recombinant baculoviruses. Surprisingly, the mutant protein was severely affected in all biochemical activities tested; no helicase or primase activities could be detected, and the mutant protein retains only about 9% of wild type levels of single-stranded DNA-dependent ATPase activity. Gel mobility shift assays showed that DNA binding is severely affected as well; the mutant subcomplex only retains approximately 8% of wild type levels of binding to a forked substrate. On the other hand, the mutant protein retains its ability to interact with UL5 as indicated by copurification and with UL8 as indicated by a supershifted band in the gel mobility shift assay. In addition, the ability of individual subunits to bind single-stranded DNA was examined by photo cross-linking. In the wild type UL5.UL52 subcomplex, both subunits are able to bind an 18-mer of oligo(dT). The mutant subcomplex was severely compromised in the ability of both UL5 and UL52 to bind the oligonucleotide; total cross-linking was only 2% of wild type levels. These results are consistent with the proposal that the putative zinc binding motif of UL52 is required not only for binding of the UL52 subunit to DNA and for primase activity but also for optimal binding of UL5 to DNA and for the subsequent ATPase and helicase activities.     

Simian virus 40 origin DNA-binding domain on large T antigen.

Fifty variant forms of simian virus 40 (SV40) large T antigen bearing point, multiple point, deletion, or termination mutations within a region of the protein thought to be involved in DNA binding were tested for their ability to bind to SV40 origin DNA. A number of the mutant large T species including some with point mutations were unable to bind, whereas many were wild type in this activity. The clustering of the mutations that are defective in origin DNA binding both reported here and by others suggests a DNA-binding domain on large T maps between residues 139 and approximately 220, with a particularly sensitive sequence between amino acids 147 and 166. The results indicate that the domain is involved in binding to both site I and site II on SV40 DNA, but it remains unclear whether it is responsible for binding to cellular DNA. Since all the mutants retain the ability to transform Rat-1 cells, we conclude that the ability of large T to bind to SV40 origin DNA is not a prerequisite for its transforming activity.     

Interaction between gene II protein and the DNA replication origin of bacteriophage f1

The origin of DNA replication of the filamentous bacteriophage f1 binds its initiator protein (gene II protein) in vitro to form a complex that can be trapped on nitrocellulose filters. The binding occurs with both superhelical form DNA and linear DNA fragments. A number of defective mutants of the origin were tested for the ability to bind gene II protein. The region of DNA required for the binding is around a second palindrome downstream from the palindrome that contains the DNA replication initiation site. It overlaps, but is not identical to, the region required for the nicking reaction by the protein. The nicking site itself was dispensable for the binding. In vivo, a number of defective deletion mutants of the origin, when in a plasmid, inhibited growth of superinfecting phage if the intracellular level of gene II protein was low. In addition, these defective origins inhibited the activity of the functional phage origin located on the same replicon. The domain of the DNA sequence required for inhibition in vivo was consistent with that for the binding in vitro.