Atlas of Macromolecules for MolviZ.Org (Atlas.MolviZ.Org) Suggestions to Eric Martz. All images copyright © 2002-2006 by Eric Martz. Click on any image for more information.

Contents Lesson Plans Enzymes Soluble Proteins (Not Enzymes) Signal Cascade Proteins (Intracellular) Structural & Motility Proteins Calcium-Binding Proteins Lipid Bilayers & Water Integral Membrane Proteins Myristoylated Proteins DNA, RNA Proteins Complexed to Nucleic Acids (Transcription Factors, Polymerases etc.) Virus Capsids Virus Components Immune System & Defense Molecules (Antibody, MHC, etc.) Toxins Carbohydrates Magnificent Molecular Machines Unusual Tertiary/Quaternary Structures Animated Morphs of Conformational Changes Evolutionary Conservation Protein Crystals History: Earliest Crystallographic Structures Other Browsable Lists of Molecules

STRAIGHTFORWARD
• Lysozyme, human, complexed to inhibitor, (1965-1994). Hen egg white lysozyme was the first enzyme structure solved crystallographically (see history). Catalytic mechanism, 2001.
• Trypsin (porcine pancreatic) complexed to soybean trypsin inhibitor, (1998).
• Carboxypeptidase A complexed with inhibitor, (1992).
• Acetylcholinesterase complexed with huperzine inhibitor, (1993-1997). SELECT Protein, DISPLAY Surface, then rotate until you can see the deep catalytic gorge containing the inhibitor.
• Lipase, (1994). See animation of catalytic pocket opening.
• HIV protease with inhibitor, (1991).
• Angiotensin-converting enzyme complexed with inhibitory drug lisinopril (licensed for treating hypertension), (2003). Such inhibitors were developed without knowing the target protein structure, and on the assumption of mechanistic homology with carboxypeptidase A. This structure showed little if any such homology, but rather structural (but not sequence) homology to a family of zinc metalloproteases.
• Intramembrane protease: see Integral Membrane Proteins.
   
  CHALLENGING
• Animated simulation of the process of HIV protease inhibitor Ritonavir binding to the protease. This simulation is fictional, but chemically possible. The simulation was generated, starting with (1997), by pulling the inhibitor out of the "side" of the protease with concurrent real-time energy minimization by molecular mechanics using the Sculpt software package. The animation can be viewed in the article Molecular Playground/HIV Protease Inhibitor: there, press the button play animation.
• Succinyl Co-A synthetase, (1994). (The mmol tetramer was constructed as being a more probable specific oligomer than the tetramer in the asymmetric unit of
• Calcineurin, a serine/threonine phosphatase with two metals in its catalytic site, a myristoylated regulatory chain with four bound calcium ions, complexed with the immunosuppresant drug FK506 and FK506-binding protein, (1995). What is the state of the catalytic site?
• Glutamine synthetase, a twelve-chain homo-dodecamer, (1995). See the overview of this molecule at Molecule of the Month.
• Particulate methane monooxygenase converts methane to methanol: (2005). Can you guess the environment in which this enzyme functions by examining its surface polarity? Understanding the mechanism of this metallo-enzyme may facilitate the development of synthetic catalysts useful in alternative energy strategies.

STRAIGHTFORWARD
• Inositol 1,4,5-triphosphate bound to the ligand-binding domain of the endoplasmic-reticulum-membrane-associated calcium release channel, (2002).
• Phytochrome (definition) chromophore-binding domain with covalently bound biliverdin, (2005). Surprisingly, the structure contains a knot. (Knots are quite rare in proteins.) The knot appears to stabilize the region near the chromophore. View the knot.
  
  
• Postsynaptic density master scaffolding protein, under structural proteins.

STRAIGHTFORWARD
• Myoglobin, oxy, (1958-1980). The first protein structure solved crystallographically (see history).
• Myoglobin, deoxy, with all hydrogen atoms! (2000). What was the experimental method? (See Note #1.)
• Hemoglobin, deoxy, (1968-1992).
  See Oxygen-Binding Proteins tutorial at Biochemistry in 3D, and hemoglobin tutorial at MolviZ.Org in various languages.
• Insulin, (1982-1992).
• Green fluorescent protein, (1997).
• Bone morphogenetic protein complexed with its inhibitor Noggin, described as "back-to-back butterflies" (Nature 420:613), (2002). These secreted proteins control events in embryonic morphogenesis. Don't miss the cystine knots! (The complex was generated from the crystal asymmetric unit

  See (under other categories)

• Calmodulin

  CHALLENGING

• Sickle cell hemoglobin (dimer of tetramers bound via mutant valine), (1985). See hemoglobin tours at molvisindex.org. See procedure for aligning sickle Hb with wild type and viewing the mutant residue in the Guide to Using MSA3D.

Anthrax Toxins
• Review of the anthrax disease and its toxins, Mock & Fouet, Ann. Rev. Med. 55:647, 2001. See also Nature's Focus on Anthrax.
• Anthrax protective antigen, PA ( 1997) transports the lethal and edema factors through the cell membrane. The full length PA 735 AA chain, PA83, is cleaved to form the active fragment PA63, which forms heptamers. The publication linked to 1ACC describes the PA monomer and heptmer, but the 4.5A heptamer model was not deposited in the PDB. Robert Liddington kindly provided the with permission to share it for educational purposes.
• Anthrax lethal factor alone (2001), and complexed to its target MAPKK2, (2001). (1j7n_1.mmol was extracted from the crystal dimer,
• Anthrax edema factor (2002) vs. edema factor complexed with calmodulin, (2002). Calmodulin activates the adenylyl cyclase exotoxin activity. (1k93_2.mmol is one dimer extracted from the original hexamer See also other calmodulin structures.
  
  Botulinum Neurotoxins, "BoTox®"
• Botulinum toxins, some of the most deadly toxins known, bind strongly and specifically to protein and ganglioside receptors localized to nerve cell membranes. Toxin molecules are thought to enter temporary openings in synaptic vesicles. As the vesicles are recycled, toxin gains access to the cytoplasm, where its light chain inactivates neuromuscular junctions, causing paralysis.
• Botulinum neurotoxin serotype B, carboxy-terminal domain of the heavy chain, complexed to the toxin-binding region of its protein receptor, the luminal portion of synaptotagmin II, residues 44-60 ( 2006). This portion of synamtotagmin II is natively unstructured, but forms an alpha helix by induced fit to the BoTox.
• Botulinum neurotoxin serotype A ("BoTox®") light chain, inactivated by two mutations, complexed to the carboxy terminal domain, residues 141-204, of human SNAP-25 ( 2004). SNAP-25 is a SNARE protein involved in neurotransmitter release. The light chain of botulinum toxin is a zinc metalloprotease that inactivates SNAREs at neuromuscular junctions, producing paralysis. This structure confirms that the protease uses exosites, remote from the catalytic site, to recognize its substrate.
• Botulinum neurotoxin serotype B complexed to sialyllactose ( 2000). Sialyllactose is a partial mimic of a ganglioside. Gangliosides and protein receptors on neurons are critical for toxin binding. This structure has the entire toxin molecule, including its catalytic, translocation, and binding domains.

CHALLENGING
• Collagen fiber, (1994).
  
  
  
  
• Bacterial Flagellar Filament
  
  
  
• The postsynaptic density (PSD) is a large protein complex involved in neural signalling. Master scaffolding proteins in the Shank family are believed to form sheets of helical fibers. These sheets are believed to nucleate the PSD, a disk-shaped protein assembly roughly 45 nm thick and 500 nm wide, containing over 100 different proteins. In the crystal of one Shank protein (mutated to increase solubility), protein chains assembled into helical fibers, which in turn self-assembled into flat sheets resembling those seen in electron micrographs. The asymmetric unit of this crystal, (2006), contains only monomers lacking interchain contacts. Marisa Baron and coauthors kindly provided coordinates for the of helical fibers. All the contacts in the sheet are represented in this in which the helix-forming contacts occur between chains B and W, while the inter-helix sheet-forming contacts occur between the B-W dimer, and chains C and D (which are in the contacting helix).

• EF Hand binding calcium, animated, in Proteopedia's article on Recoverin.
• Calmodulin:
  • With calcium, (1994).
  • With calcium and bound portion of myosin light chain kinase, (1992).
  • With calcium and bound portion of calcium-activated potassium channel, (2001). First instance of calmodulin binding three alpha helices, with the N vs. C ends of calmodulin contacting different bound domains. (The complete tetramer was generated from the crystal asymmetric unit dimer

  See (under other categories)

• Anthrax edema toxin bound to calmodulin, with calmodulin in a new, extended conformation.
• Recoverin

  CHALLENGING

• Calmodulin, with calcium, showing flexibility of interdomain linker, (1995).
• Calmodulin, with calcium, complexed to fragment of myosin light chain kinase, (1992).

STRAIGHTFORWARD
• 200-phospholipid hydrated bilayer theoretical models (2 megabytes each!) in
  conformations (Heller, Schaefer & Schulten, J. Phys. Chem. 97:8343, 1993; PDB files redistributed with permission).
  Figure at right (© 1996) "Five Bakers Dancing".
  
  
  
• Lipid Bilayers and the Gramicidin Channel, Modelo de bicapa lipídica y canal de gramicidina.
• Visualizations of the gramicidin channel passing water through a lipid bilayer.

• Water: theoretical simulation of 10 water molecules assembling into a hydrogen-bonded network. Also models of water freezing, and larger assemblies of liquid water.
  

STRAIGHTFORWARD
• Potassium channel, (1998, 3.2 Å). See another version below under "CHALLENGING".
• Aquaporin, a water channel, (2001). (1j4n.mmol was assembled as a tetramer of the asymmetric unit in
• A homo-heptameric membrane pore, Staphlococcal alpha-hemolysin, (1996).
• Intramembrane protease of the rhomboid family, core catalytic domain of E. coli GlpG, (2006). A transmembrane protein itself, GlpG cleaves substrate transmembrane protein within the hydrophobic membrane lipid environment. The catalytic site includes Ser201, His254, and water.
• Bacteriorhodopsin, a light-driven ion pump, (1990 by electron diffraction, 1995 by X-ray, 1999). Has seven trans-membrane alpha helices with retinal and associated lipid. Note the buried water.
  See Bacteriorhodopsin tutorial at Biochemistry in 3D.
• Outer membrane iron transporter of E. coli, (2002).
   
                       

  CHALLENGING

• theoretical model (Crouzy, Woolf & Roux, Biophys. J. 67:1370, 1994; PDB file redistributed with permission).
  • Visualizations of the gramicidin channel passing water through a lipid bilayer.
• Potassium channel with four antibody Fab fragments bound (Caution, 1.2 megabytes) (2001, 2.0 Å). (1k4c.mmol was assembled as a tetramer of the asymmetric unit in
• Photosynthetic reaction center, purple bacterium, (reported 1984, deposited in the Protein Data Bank 1988). This is the first high-resolution trans-membrane structure obtained by X-ray crystallography, earning its authors the 1988 Nobel Prize in Chemistry. (Here is a link to their original 1984 paper, not linked to the PDB record.)
• Photosynthetic reaction center with bacteriochlorophyll a, ubiquinone and iron, (1994).
• Calcium pumping ATPase, with cytoplasmic and transmembrane domains, (2000). Another conformation, based upon cryo-electron microscopy, provides a model for the catalytic cycle, (2002).
   
  
• Porin, sucrose specific, Salmonella, water and sucrose in the pores, (1998).
• Mechanosensitive channel, gated, "large", MscL homolog from Mycobacterium tuberculosis, (1998).
• Mechanosensitive channel, voltage sensitive, E. coli MscS, (2002).
• Outer membrane protein A of E. coli (2000). Compare with showing flexibility, (2001).
• Chloride channel, (2002). (1kpl_1.mmol is one dimer, extracted from the crystal tetramer
• Succinate:quinone oxidoreductase (SQR), also the succinate dehydrogenase of the Krebs cycle, composed of four protein chains and five prosthetic groups, including a heme buried in the membrane, (2003). This structure "reveals that electrons are tranferred in a path more than 40 Å in length from the succinate oxidation site at the FAD, by way of the three iron sulfur clusters, to the ubiquinone binding site".

Recoverin 1iku...1jsa

CHALLENGING
• Recoverin, a calcium-activated myristoyl switch, (Caution: NMR ensembles, >1 megabyte each; see Note #2) without calcium, myristate buried, (1995), and with calcium, myristate exposed, (1997).
  Brief introduction to recoverin and morphs of the transition.

  See also (under other categories)

• Calcineurin.

Genes were shown to reside in DNA in 1944 (Avery et al.) and this became widely accepted after the 1952 experiments of Hershey and Chase. The double helical structure of the DNA was predicted by James Watson and Francis Crick in 1953 (Nobel Prize, 1962). Their prediction was based in part upon X-ray diffraction studies by Rosalind Franklin, to whom Watson and Maurice Wilkins gave inadequate credit (see Rosalind Franklin: Dark Lady of DNA by Brenda Maddox, HarperCollins, 2002). The predicted B-form double helix was not confirmed with atomic-resolution crystal structures until 1973, first by using dinucleotides of RNA (Rosenberg et al.). The first crystal structure containing more than a full turn of the double helix was not solved until 1980 (Wing et al. 1981, 12 base pairs). The lag of more than a quarter century between prediction and empirical confirmation involved development of X-ray crystallography for macromolecules, and the need to produce a short, defined sequence of DNA for crystallization. This brief account is based upon a review by Berman, Gelbin, and Westbrook (Prog. Biophys. molec. Biol. 66:255, 1996), where the references will be found.

STRAIGHTFORWARD
• DNA double helix (B form), 12 base pairs, one turn, (first full turn solved crystallographically, see above). This is an empirical crystallographic structure: notice the substantial and variable deviations from coplanarity for the base pairs. (1986) has almost the same sequence, but can you find the mismatched base pair? (See Brown et al.).
  See tutorial Nucleotides at Biochemistry in 3D, and interactive DNA molecule at UMass in English, auf Deutsch, en Espagñol, in Portuguese.

  

• DNA double helix (B form), 36 base pairs (3 full turns), ideal theoretical model,
• DNA double helix (A form), 36 base pairs, ideal theoretical model,
• DNA double helix, 12 base pairs,
• Make your own sequence into A or B form DNA (or A form RNA), courtesy Pittsburgh Supercomputer Web Tools, Carnegie-Mellon University.
• Nucleotides, base pairs, and DNA forms at Glactone/Georgia State U.
• Atlas of Nucleic Acid-Containing Structures.

• DNA double helix, two turns of chromatin theoretical model, 171 base pairs,
• DNA/RNA hybrid double helix, (1995).
• Transfer RNA (Phe), (1974-1978). ( [2000] is a more challenging tRNA.)

  CHALLENGING

• DNA B form to Z form transition, (2005). The cover of the issue of Nature reporting this structure showed a theoretical model with longer B- and Z-form ends (shown at right), kindly provided for this Atlas by Kyeong Kyu Kim. View this model in MolSlides in Jmol; or view model directly, In order to solve this structure, the Z-DNA portion was stabilized by Z-DNA-binding proteins. The surprise is that base pairing is disrupted at only a single pair at the transition point. Sequences near promotors often favor Z-DNA, enabling them to trap negative supercoiling that occurs behind a moving polymerase, or during nucleosome unwrapping. Z-DNA cannot participate in a nucleosome, hence exposing it to transcription factors.
• Hammerhead ribozyme, (1994). (1hmh_1.mmol is one dimer, extracted from the crystal hexamer
  See Hammerhead Ribozyme tutorial at Biochemistry in 3D.
• RNA self-splicing group I intron with both exons, (2004). While most introns are removed by a large ribonucleoprotein complex, the spliceosome, a subset catalyze their own excision without accessory factors. 1U6B is the first structure to include an active site (with magnesium) plus an intron and two exons, representing the splicing intermediate before the exon ligation step.

STRAIGHTFORWARD
• Lac repressor bound to DNA: Animations of nonspecific binding converting to specific binding and bending the DNA, as in the image at right. Questions for students are provided.
• DNA polymerase complexed with DNA, (1996).
• Lambda repressor complexed to DNA, (1992).
  See Lac Repressor tutorial at Biochemistry in 3D.
• Zinc finger (Zif268) complexed to DNA, (1996).
• TATA-box binding protein, human, complexed to DNA, (1996).
• EcoRV endonuclease complexed to DNA, (1995).
  See Endonuclease tutorial at Biochemistry in 3D.
• SarA transcriptional regulator of virulence factors in Staphylococcus aureus, (2001).
  Morph "movie" of conformational change when SarA binds to DNA, see Animations in PE.
• HIN recombinase with specific contacts in both the major and minor grooves of the DNA double helix, (1994).

  CHALLENGING

• RNA polymerase (T7 bacteriophage), with DNA template and nascent RNA, (1999).
• Recombinase (Flp) with enclosed Holliday junction of DNA, (2000).
• DNA mismatch repair protein MutS binding to a mismatch, (2000).
  • See if you can find the mismatch in the DNA double helix! To help, here is (about 1,700 of the about 13,000 atoms in 1E3M).
  • Another case of MutS with bound DNA is (2000): what is wrong with the DNA there? To help, here is only the
  • [If you would like to know how to save subsets of atoms from a model, contact emartz@microbio.umass.edu. Generally, using either the Jmol Application or the signed Jmol applet, you select the desired subset, then write a PDB file to disk.]
• Ku heterodimer bound to DNA, (2001). Facilitates repair of double-strand DNA breaks by non-homologous end-joining. Encircles the DNA double helix with a steric fit, avoiding base contacts.

  ENORMOUS

• Nucleosomes 
  • Nucleosome (DNA histone complex), (1997).
  • Tetranucleosome, (2006). The authors built a model of a chromatin fiber (not shown in this Atlas) in which the H4 tail can interact with H2A-H2B across an internucleosomal interface (as chemical crosslinking evidence suggests that it does).
  • Kaposi's sarcoma Herpesvirus protein bound to nucleosome, (2006). This virus protein (LANA) attaches viral genes (episomes) to mitotic chromosomes, ensuring faithful partitioning into progeny cells. This crystal structure shows that the LANA N-terminus (residues 1-23) binds to histones H2A-H2B in the same region thought to be occupied by the H4 tail in chromatin fibers. In fact, in these crystals, the H4 tail bound to some of the same regions through a crystal contact (not shown in this Atlas).
• Ribosomes 
  • Ribosome, 70S, with 3 tRNA's and mRNA (phosphorus and alpha carbon atoms only). The 49 chains are provided in 2 PDB files ( 2001). A combined model (a pseudo with the sidechains missing is This is a pseudo-NMR file, with the small subunit in model 1 and the large subunit in model 2. Find tours of the ribosome at molvisindex.org.
  • Ribosome, small subunit (30S), 22 chains with sidechains (CAUTION, 1.1 megabytes*), (2000).
  • Ribosome, large subunit (50S), 29 chains with sidechains (CAUTION, 1.3 megabytes*), (2000).

Images of virions elsewhere:
• VIPERdb Virus Particle Explorer at The Scripps Research Institute. (This site does not work in Netscape 4.)
• Virus World at the University of Wisconsin.

  CHALLENGING

• Simplified model of SV40 capsid: Each of the 360 chains has been reduced to a single "atom", and the enclosing icosahedron is represented.
   
  ENORMOUS
• (CAUTION: 3 megabytes*, alpha carbons only), based on the homopentamer (+1) of the VP1 capsid protein in (1996).
  • The whole capsid (about 70 megabytes with all sidechains) was assembled After removing all atoms except the alpha carbons (with PDB Tools), the resulting 12 megabyte PDB file was gzipped to less than 3 megabytes in file (alpha carbons only, but default java memory may be too small).
  • The 360 chains in this mmol file greatly exceed the number Jmol can handle gracefully, but all are displayed showing a complete capsid. In FirstGlance, they are shown not as chains, but as alpha carbon atoms. The 360 chains are arranged in 72 pentamers, which in turn are arranged to form a 12-faced icosahedron.
  
   
  

• "Protein chainmail" virus capsid, bacteriophage HK97, (CAUTION: 2 megabytes*) entire capsid, alpha carbon atoms only, This capsid matures by autocatalytic formation of 420 covalent isopeptide bonds between the side chains of lysine 169 and asparagine 356. (These isopeptide bonds are not easily visualized with the PDB files available here.) (1fh6_cao.mmol was assembled from
  
• sTALL-1, a cytokine proposed to function as a 60-chain homomultimer very similar to a virus capsid. The possible evolutionary relation to capsids of small RNA plant viruses is intriguing.

STRAIGHTFORWARD
• HIV protease with inhibitor, (1991).
• Animation of HIV protease binding inhibitor: see Enzymes.
• SARS coronavirus spike protein complexed to its receptor, angiotensin-converting enzyme 2 (ACE2), (2005). (One copy of this complex was extracted from the crystal dimer in the asymmetric unit of
  
  ENORMOUS
• See Kaposi's sarcoma protein bound to a nucleosome.

CHALLENGING

• The bacterial flagellar hook, model of a molecular universal joint.
  
  See also Bacterial Flagellar Structures at Atomic Resolution.
  
  
• Bacterial cell wall puncturing device of bacteriophage T4, aka tail lysozyme, (Caution: 1.6 megabytes) (2002). (This hexamer was generated from the dimer in the asymmetric unit of
• Proteasome (28 chains, 1.2 megabytes*), (1995). From the archeon Thermoplasma acidophilum.
• HslU protease/HslV chaperone complex (24 chains, 1.0 megabytes*), (2000). From Hemophilus influenzae.
• Chaperonin, GroEL-GroES (21 chains, 1.3 megabytes*), (1997). From E. coli.

  SEE ALSO (under other categories)

• The Ribosome.
• The Nucleosome.

STRAIGHTFORWARD

    Antibody
• Complete IgG antibody, (1997).
  See interactive antibody tutorial at UMass in English, auf Deutsch.
• Complete IgG antibody with finger-like loop able to fit into the crevice in HIV gp120 that accomodates CD4, (2001). The finger-like loop is heavy chain CDR3, 18 residues numbered 94-101 (including inserted residues numbered 100A-100J). A synthetic peptide mimicking this CDR H3 neutralized HIV.
• Fab fragment of antibody bound to lysozyme antigen, (1991).
  
      MHC
• Major histocompatibility complex, class I, with viral peptide, (1992). To compare with a different viral peptide, see
  See MHC tutorials at Biochemistry in 3D and at MolviZ.Org.
• Major histocompatibility complex, class II, with peptide from lysozyme, (2000). (One copy of this complex was extracted from the crystal dimer in the asymmetric unit of
• Major histocompatibility complex, class II, with peptide from HIV Gag protein, (2004). T cell activation sometimes requires residues in the peptide that extend beyond the MHC groove. Here, a previously unknown peptide conformation is observed in which the carboxy terminus folds up and back, providing a required portion of the epitope. This could turn out to be a common situation.
• CD1b complexed to ganglioside GM2, (2002). CD1b is thought to present microbial lipids, including mycobacterial mycolates, to T lymphocytes, aiding in immune defenses against microbes. CD1's are structurally very similar to MHC proteins.
• CD1d complexed to alpha-galactosylceramide, (2005). This complex stimulates human and mouse natural killer T cells. Alpha-stereoisomers of glycosphingolipids, absent from mammals, are found in microbes and marine sponges. CD1's are structurally very similar to MHC proteins. (One copy of this complex was extracted from the crystal dimer in the asymmetric unit of
  
      Cytokines
• Interleukin-1 receptor complexed to IL-1 receptor antagonist, human, expressed in E. coli, (1997).
• Interleukin-2 complexed to the Alpha, Beta, and Gamma chains of its receptor, (2006). (One copy of this complex was extracted from the crystal dimer in the asymmetric unit of
• Interleukin-4 complexed to alpha-chain of receptor, human, expressed in E. coli, (1999).
• Interferon Gamma, bovine, expressed in E. coli, (2000).
• Chemokine IL-8, human (mutated), expressed in E. coli, (1997).  
• sTALL-1 cytokine of the TNF superfamily, human, soluble fragment expressed in E. coli, (2001; alpha carbons only).
  This 60-chain homo-multimer, proposed to be the functional form of this cytokine, was generated from the crystal asymmetric unit of This virus-capsid-like assembly is very similar to capsids of the smallest plant RNA viruses, raising the question of which came first.
  
  
      Immuno-Receptors
• CTLA-4 bound to B7-1 (external domains only), (2001). The divalent homodimers of both molecules are believed to form chains in the immunological synapse.  
• Defensin (Rhesus theta defensin one, RTD-1), an antibacterial cyclic 18-amino acid peptide, (20-model NMR strcture, 2001). Each half is coded by a different gene. Cyclization is thought to increase resistance to exoproteases (Trabi & Craik, TiBS 27:132).
  

  CHALLENGING

• CD4 (human) complexed to HIV gp120 and Fab (human), (2000).
• CD8 complexed to MHC class I, (1998). Only the terminal domains of the CD8 alpha-alpha homodimer are present. The MHC class I in this complex is H-2Kb with VSV8 peptide (same as 2VAA above).
• T Cell Receptor (ab), complexed to HTLV1 viral peptide presented by HLA-A2, (1996).
• T Cell Receptor (gd), (2001). (One copy of the heterodimer was extracted from four copies in the crystal asymmetric unit of

• Capsular polysaccharide, E. coli, (1977).
• Chondroitin-4-sulfate, sodium salt, (1978).
• Heparin (2-model NMR structure) (1993).

• A deeply knotted protein chain, (2000). You won't be able to discern the knot by simple observation. See Knots in Proteins for an explanation and convincing visualizations.
• The cyclic cystine knot (CCK) motif involves penetration of a small ring, formed by two disulfide bonds and their connecting peptide backbone segments, by a third disulfide bond (Trabi & Craik, TiBS 27:132). Found in plant defense proteins called cyclotides, the first CCK structure reported was that of (10-model NMR structure, 1995).
• Intrinsically disordered chains that fit together in a stable heterodimer via "mutual synergistic folding". The heterodimer between p160 receptor coactivator and CBP/p300 transcriptional coactivator, (2002), is a member of a family of ligand-activated transcription factors involved in nuclear hormone receptors.
• Chaperone SicP maintaining Salmonella SptP virulence effector protein (105 residues) in a largely "unfolded" state, presumably to facilitate type III secretion, (2001).
• Naturally-occurring cyclic peptides.
• sTALL-1, a cytokine proposed to function as a 60-chain homomultimer very similar to a virus capsid. The possible evolutionary relation to capsids of small RNA plant viruses is intriguing.

Suggestions to Eric Martz.

1. To find the experimental method:
   Use Proteopedia:
   • If a resolution is given just below the molecule, the method is usually X-ray crystallography. To be sure, click on the link to RCSB ( under Resources).
   • If a number of "NMR models" is given, the method is usually solution nuclear magnetic resonance.
   
   
2. To see an NMR ensemble of models:
   • Proteopedia shows you all models by default, and (just beneath the molecule) gives the number of NMR models.
   • Note that FirstGlance shows you only the first model of the ensemble, until you ask to see all models by clicking All Models.
   
   
3. Most of the biological units in this Atlas (filenames ending in .mmol) were obtained from the Probable Quaternary Structure server of the European Bioinformatics Institute. However, this server stopped adding new structures in August, 2009.

   Currently (2012), the Biological Unit (functional quaternary structure) can be obtained most easily from the MakeMultimer Server.

   • Proteopedia: Click on the link to FirstGlance (under Resources beneath the molecule), then see the next item here:
   • FirstGlance: Click Key Resources, then Probable Quaternary Structure. Several options will be offered and explained.