Protein Secondary Structure PDF

 Protein secondary structure   #### Secondary structure Leninger Fig 3-13 - Secondary structure reflects the arrangement of residues close in sequence - Most residues in most proteins can be seen to be forming either α...

 Protein secondary structure   #### Secondary structure Leninger Fig 3-13 - Secondary structure reflects the arrangement of residues close in sequence - Most residues in most proteins can be seen to be forming either α-helices or β-strands - But what dictates that these arrangements predominate, and not others? Protein backbone conformation ================================================== - The conformation of the backbone can be understood by examining the bonds that join one residue to the next - Three bonds join each Cα to the next Cα, so a protein's conformation can be described by three angles per residue - These angles are named omega (Ω), phi (Φ), and psi (Ψ) Dihedral angles --definition =================================================  - Given four atoms A-B-C-D connected in a row - The dihedral angle is the angle made between the A-B bond and the C-D bond when looking down the B-C bond - Measured in a clockwise direction - Single bonds are short (\ - The omega angle is the dihedral angle between C of the previous residue and nitrogen of this residue - It is measured clockwise from C~α(-1)~ to C~α~ 3.8 Å #### character - In amide bonds, the C-N bond is **resonance stabilized** - The partial double bond character of the C-N bond imposes a high energetic penalty for distorting the Cα-C-N- Cα dihedral (the Ω angle) away from 180° (or 0°) - In small peptides, Ω is always very close to 180° - Note that this means that the groups joining Cα to Cα are rigid, holding these atoms 3.8 Å apart  - Very precise high-resolution structures show that ω is often slightly twisted (5 to 10˚) from the ideal 180˚ - This distortion causes local strain, but better packing interactions elsewhere within the protein can compensate *cis* (Ω = 0˚) peptide bonds are much less stable than *trans* (Ω = 180˚) -------------------------------------------------------------------------  - For Xaa~1~-Xaa~2~ (where Xaa~2~ is [not] Proline), *cis* peptide bonds are around 20 kJ/mol [less] stable than the corresponding *trans* peptide bond - Consequently, only 0.03% (\~1 in 3000) of amino acids (proline not in the second position) are in the *cis* configuration - Given that the average protein has \~375 a.a., one in tent proteins ### Proline more readily adopts *cis*   - For Xaa-Pro, the Cδ atom clashes [in *trans* conformation] - Therefore, *cis* is only 2 kJ/mol less stable than trans - About 5.4% (\~1 in 18) of Xaa-Pro peptides are *cis* - *cis-*Prolines have important structural roles, e.g. in forming certain types of turns  ### Phi & Psi define backbone conformation - Generally, we can assume peptide bonds are essentially flat, with Ω - Cα(-1), C(-1), O(-1), NH and Cα (\*) [are confined to one plane] - Backbone conformation is then a function of two dihedral angles per residue: phi (φ) and psi (ψ) - I.e. knowing just phi and psi angles for every amino acid is enough to describe the overall structure of a protein (side chains excluded) - Phi (Greek letter φ) is the dihedral angle subtended by the N-Cα bond - It measures the angle between carbonyl carbon (CO) groups in a clockwise fashion - Possible mnemonic: CO-Phi (coffee) - Psi (Greek letter Cα -CO bond - It measures the angle between NH groups in a clockwise fashion - Possible mnemonic: PoSeIdoN (and hey, is that not a trident?) Ramachandran plots ========================================= - Ramachandran plots show the stability of an amino acid in a protein as a function of [Phi] and [Psi] angle - The green areas correspond to conformations where strain and van der Waals clashing is minimal - Note that positive phi values are largely disallowed because the carbonyl oxygen groups tend to clash (on right with C~β~) #### Ramachandran describe allowed backbone conformations - Amino acids need to avoid strongly strained conformations - However, they also need to maximize stabilizing interactions (e.g. H-bonds) - Actual observed conformations (black dots) are a compromise - Typically, \< 2 % of amino acids will be outside the blue areas - Ramachandran plots differ slightly for different amino acids - E.g. Val, Ile and Thr are slightly more restricted by the C~β~ branching Lehninger, 6e 4.9b Ramachandran plot for Proline ==================================================  - The imine ring of proline restricts Phi to values of - Proline is much less flexible than other amino acids Ramachandran plot for Glycine ==================================================== - Glycine (H- sidechain) has no hand and therefore no preference for L- or R- conformations - Glycine is therefore can adopt conformations that would be very strained for any other amino acid  The extended state ----------------------------------------- - If φ = -ψ, this results in a residue with an extended conformation - φ = -135; ψ = 135 is extended and minimally strained - In this maximally extended state, **peptides span 3.5 Å per residue** - Attaining the minimal energy state of the backbone requires minimizing conformational strain, while maximizing interactions - Protein interiors are generally non-polar, so buried backbone amide groups generally hydrogen bond with other backbone amides - Hydrogen bonds are much stronger with optimal geometry, which strongly constrains favourable conformations - The fixed omega angle limits the flexibility of the polypeptide chain - It is only possible for multiple successive buried peptide groups to make good interactions if all adopt a similar pattern of repeated interactions - These **runs of residues in (near) identical conformations, stabilized by repeated patterns of hydrogen bonds,** are the **secondary structure elements** that are characteristic of protein structure #### Favoured regions of the Ramachandran plot are characteristic of specific secondary structures - We can distinguish three general regions of the Ramachandran plot that are unstrained - α (characteristic of α−helices) - β (characteristic of β−strands) - L: left-handed α-helix, infrequent and never forms extended runs Lehninger, 6e 4.9a  α-helix ======= - Helices are characterized by a repeating pattern of hydrogen bonds between residues a small set number apart in sequence - In α-helices, the carbonyl oxygen of **each residue H-bonds with the NH of the residue 4 residues C-terminal** to it (e.g. 1 with 5) - 3.6 residues per turn (=100˚ rotation per a.a.) - The α-helix length increases by 1.5 Å for each - Residues (C~α~-C~β~ bond) point towards N- terminus (so side chains "hang down" like a Christmas tree) - There are 3.6 residues per turn - There are 13 atoms that covalently link the two H- bonded atoms (including the hydrogen atom) - The α-helix is therefore also called a 3.6~13~ helix α-Helix Phi - Psi angles =============================================== - α-helices occur with φ around - Note that the backbone conformation is also unstrained with more negative - This is because despite the lack of clashes, hydrogen bonding is compromised - α-helices are very common, accounting for about one third of a.a. in protein structures  The α- helix is compact and stable ---------------------------------- - The α-helix is tightly packed - Residues on the inside being in van der Waals contacts with each other (note - no hole in the middle) - The α-helix is therefore stabilized by: - tight packing - strong hydrogen bonds between turns - a lack of backbone strain (Ramachandran) α-helices are right-handed =============================================== - Thumb points along the axis - Fingers curl around and "up" the helix -  α-helices, 3~10~ and pi helices are all right- handed Left-handed helices are very rare ================================= - In principle, it is possible to have a run of residues in "L" conformation - These then form a left-handed helix - Such helices are very rare in real proteins - Since the "L" conformation is only marginally stable, the surrounding protein is needed to stabilize such regions - Such helices are very rare (\ - The CαH in β-sheets forms a weak H- bond with the carbonyl oxygen of the opposite strand (orange dash line) - Changes in the atom's chemical shift (measured by NMR) show that this interaction is a hydrogen bond - Crystallography confirms that the hydrogen atom is shifted towards the carbonyl oxygen - These hydrogen bonds help stabilize both parallel and anti-parallel β-sheets β-strand types prefer to not mix ================================ - Parallel and antiparallel β-sheets have slightly different backbone geometry preferences - The geometric compromises required to mix parallel and antiparallel strands in the same sheet creates a small energetic cost - As a result, only about 20% of strands have parallel interactions on one side, and anti-parallel on the other #### Parallel β-strands connect almost exclusively through right-handed connections - This applies whether there is a loop, an α-helix or a whole sub-domain between the strands C N - Rossmann fold domain Polyproline II helices =========================================== - Polyproline II helices are right-handed, with three residues per turn - They make no internal hydrogen bonds - They are almost as extended as β-strands - Generally, the preferred conformation of extended proline tracts - Prolines favour this structure, but it can also occur in sequences with no proline residues - Comprise about 2 % of protein residues - Polyproline I helices are all-*cis* and left-handed -- and even rarer PPII helix as a unique ligand ========================================================================= - Four copies of the acetylcholine esterase oligomerization domain (white) assemble around a single copy of a poly-proline membrane anchor protein (blue) - The anchor protein is proline rich and forms a PPII helix - This complex is important for recruiting acetylcholine esterase to the membrane PPII bundles ============ - Snow flea ice binding protein uses an extended flat surface to bind ice - The structure is built from six co-linear polyproline II helices - The structure is stabilized by hydrogen bonds between poly-Pro II helices that extend out at 120˚ intervals - All interior amino acids are glycine to allow chains to pack closely on all sides - There is no hydrophobic core in this protein - Very few proteins are built this way - Turns are short motifs that allow the protein chain to do a quick 180˚ turn - These very commonly connect adjacent anti-parallel β-strands - β-turns are comprised of four residues and are held together by a main chain hydrogen bond - Type II turns require that the residue in position 3 be in L (left- handed helical conformation) - for this reason, it is generally a glycine Loops =====  - Successive secondary structural elements (helices and strands) are connected by loops - Loops are irregular and can adopt almost any conformation as that is stabilized by available local interactions - Loop residues do need to adopt conformations from [favoured areas of the Ramachandran plot] - Backbone is a mix of α-like and β-like conformations, with no regular order - Loops range from short (e.g. 1 residue) to tens of residues - Loops rarely pack between secondary structural elements - Generally, they are on the surface of the [protein] ### Loops are stabilized by interactions with the rest of the structure - Loops do not have a regular inbuilt set of H-bond interactions - They make H-bonds wherever they can - with other main chain atoms, side chains - Many residues hydrogen bond with water -- hence the preference for surface locations - The lack of prior constraints allows enormous conformational diversity - This is critical -- loops are often the "business end" of a protein - Proteins average \~30 % of residues  - As with the main chain, side chains prefer to adopt conformations that minimize internal steric clashes - Consequently, residues prefer a set of specific conformations known as rotamers - Most non-polar residues have few single bonds (e.g. Leu), and can adopt only a few rotamers - Long hydrophilic residues (e.g. Arg) can adopt many more conformations, but still avoid strain Different amino acids preferentially stabilize different secondary structure ------------------------------------------------------------------------------------------------- - Side chains in α-helices pack relatively closely - This favours amino acids that are linear - MALEK (Met Ala etc.) residues most strongly favour α-helices - Hydrophobic interactions are the primary factor stabilizing β- sheets, so large nonpolar amino acids have the largest sheet- forming propensities - β-branched amino acids (V,I,T) also favour β-strands - Asp, Asn and Ser can form local hydrogen bonds with the main chain, and are preferred at the N-terminus of α-helices and C- terminus of β-strands - Pro breaks the hydrogen bonding pattern, and favours loops - Gly is very flexible and has no methyl group to pack, and favours loops too  Predicting secondary structure ============================== - Chou & Fasman (1970) came up with a simple algorithm for predicting the secondary structure of a protein given the amino acid sequence - They noticed from surveys of available structures that some secondary structure elements are enriched in certain amino acids. - By counting the secondary structural context of each amino acid in all known structures, they generated a table of trends - They found that you could (fairly) reliably predict the secondary structural of new proteins by looking at the s.s. preferences of short (5+) runs of amino acids in the sequence Chou Fasman a.a. preference table ------------------------------------------------------ - P(α) is the preference for that amino acid to form and - P(β) - β-strand preference - P(turn) - preference to be in a turn - A score \ 100 indicates that a residue favours that conformation - Thus P(α) \ 100 are helix called formers e.g. glutamate - Residues with P(α) \ - Using a multiple sequence alignment provides a great deal more information than a single sequence for prediction algorithms - [Secondary structure tends to be more conserved than the] [sequence] - Insertions and deletions generally do not occur in the middle of secondary structural elements (what are two "extra" residues going to do in the middle of a helix?) but rather in loops - The predictions from all sequences can be pooled and averaged to increase confidence - E.g. an amino acid that is generally helix breaking can more confidently be added to a helix if the residue is generally helix promoting in homologous sequences Consensus predictions ===================== - Using multiple tools for predicting secondary structure and then using the consensus gives the most robust results - Jpred will generate a multiple sequence alignment, run several independent secondary structure prediction programs for you and give a consensus secondary structure prediction - [[http://www.compbio.dundee.ac.uk/\~www-jpred/submit.html]](http://www.compbio.dundee.ac.uk/~www-jpred/submit.html) - In general, the agreement between jpred consensus and the real s.s. pretty good (often \\> 75%) - The exact beginning/end of s.s. elements are often off by 1-2 a.a. - Some elements are missed entirely e.g. last helix - Methods that consider local sequence can give good, but not perfect predictions - Recent methods that predict the whole structure (e.g. Alphafold) also predict secondary structure - These predictions are more accurate than secondary structure alone predictions (\~98% accuracy)