Notes on protein-DNA interaction:

General themes:

Particular motifs illustrated in class

• Helix-turn-helix motif:

  
  The HTH was the first DNA recognition motif discovered and crystallized. It is found in many bacterial and phage repressors and activators (Lac repressor, CAP, lambda CI and cro, Trp repressor). The eukaryotic homeodomain proteins have a related motif, but it is oriented somewhat differently in the groove. We looked at the lambda phage cI repressor.

  The recognition helix is an example of the most common recognition motif, an alpha helix making hydrogen bonding contacts in the major groove. The packing helix in the helix-turn helix motif holds it into groove. Just as in enzymes, much of the rest of the protein is scaffolding to fold the business end and keep it in place.

  There are stabilizing contacts from other parts of the protein to neighboring grooves, often in the form of arms which reach into a neighboring groove or make backbone contacts. These are more elaborate and obvious in other related proteins like the "winged helix" family.

  Bacterial/phage repressors are typically homodimers which bind to inverted repeats -- apparently biology typically can't get enough affinity/specificity from just one alpha helix/major groove interaction. Similarly, eukaryotic homeodomain proteins typically bind as heterodimers with other proteins, for example the yeast Mat a1/Mat alpha 2 complex.

  On the lambda repressor, we zoomed in and looked at an Asn-A contact, with associated water molecules contacting a neighboring T. The hydrogen bonding groups tend to buttressed into a network that makes many contacts within itself and to DNA.

• Zinc fingers:

  
  The zinc fingers are an extremely common class of transcription factor, for example there are hundreds in the C. elegans genome. The zinc finger consensus sequence is (F,Y)-X-C-X2,4-C-X3-F-X5-L-X2-H-X3-H-X5. Classic examples are the Sp1 (3 fingers) and TFIIIA (9 fingers) transcription factors. We looked at the Zif268 protein, 1aay.pdb. The Zn++ in each module is coordinated to the two His and the two Cys in the consensus. The metal is there to hold the protein together, not to contact DNA. The alpha helix makes most contacts, again in major groove, but the orientation of the helix is very different from the HTH motif: the zinc finger helix sort of noses in, making contacts primarily with one strand. From the connectivity between modules, the zinc finger really should have been called the zinc sausage. There are always at least two modules per protein, otherwise the binding affinity/specificity would not be sufficient. TFIIIA has nine fingers, although since it doesn't wrap around the DNA one or two of the fingers are skipped.

  Zinc fingers are good for engineering: the fingers are semi-independent and they contact primarily one strand. Designed Zinc fingers can bind to any sequence, though it is easier to make them bind to G-rich strands. Sangamo Biosciences is trying to make a business of delivering "payloads" that control transcription using engineered zinc fingers, but this is not necessarily a stock tip -- remember all those car companies in the early twentieth century. We showed an example of an Arg (buttressed by an Asp) interacting with a G from the Zif268 structure.

• Leucine zippers:

  
  The chopstick model for DNA recognition, using basic regions held together by a dimerization domain (the zipper) characterized by leucines every seventh amino acid. It was thought the leucines might interdigitate, hence the name, but they don't - it's just a hydrophobic interface. Presented the idea of the helical wheel representation for the protein alpha helix. The "a" and "d" positions form the interface, with the "e" and "g" positions controlling specificity via complementary or opposing charge-charge interactions (fos and jun). It's a left-handed coiled coil, not perfectly parallel because 7 amino acid repeat = 700 degrees around a circle, not 720. Characterized by DNA induced folding - the protein is monomeric and unfolded at low concentration in the absence of DNA, cooperative interactions bring two protein monomers together on the surface.DNA binding occurs through the N-terminal basic regions, alpha helices extended off the leucine zipper in major grooves on each side of the DNA. Potential for combinatorial regulation because changing dimerization partners can change DNA binding affinity/specificity, for example fos/jun is more stable than jun/jun is more stable than fos/fos. This is especially well characterized in related helix-loop-helix proteins myc/max/mad. We showed an example of Asn binding to A.

• TBP:

  And now for something completely different. Recognition of widened and flattened minor groove mainly by a ten-stranded beta sheet, with Phe's on the stirrups at the sides partially intercalating between base pairs. Alpha helices on top of the saddle hold the beta sheet in place, extensive interactions with other protein factors occur on the top side of the saddle. Very few specific H-bonding contacts, recognition is via deformability of the TATA box DNA sequence.

• Nucleosome:

  DNA wraps around 1.65 turns or so. There's a high concentration of Lys and Arg on the sides of the tuna can to bind electrostatically to DNA. The histone tails reach out between the DNA gyres. See the web page on nucleosome structure.

• Tat-Arg, Rev-RRE, U1A-RNA:

  The general theme in RNA-protein recognition is that sequence-specific proteins don't bind to undisturbed duplex RNA because there's not much information in the minor groove and the major groove is inaccessible. The TAR-Arg (model for TAR-tat protein) and Rev-RRE structures illustrate protein binding in amajor groove widened by bulging and/or non-WC base pairs. The U1A-RNA structure illustrates binding to splayed out base pairs in the loop of a stem-loop. See the new web page on RNA-protein interaction.