Proteins are built of Lamino acids.

1
Proteins are built of L-amino acids. 20 different
amino acids are encoded by specific DNA base
triplets. The amino acids are linked together by
amide bonds. Proteins are linear chains of amino
acids. Peptides are short proteins (residues) Peptide bond is another name for the
amide bond connecting two amino acids.
dipeptide
Peptide bonds are planar due to partial
double- bond character. The backbone of a
protein consists of the atoms N, Ca and C. Side
chain carbons are labelled b, g, d, etc. Bond
lengths and bond angles are invariant. Only
dihedral angles vary. By convention, a protein
starts at the N-terminus and ends at the
C-terminus. (The N-terminus is synthesized first
during translation.)
Carl Branden John Tooze, Introduction to
Protein Structure, Garland, 1998
2
glycine, Gly, G
alanine, Ala, A
hydrophobic A I L M F P W V positively charged
R K (H) negatively charged D E polar C N Q S T
Y tiny G
arginine, Arg, R
aspartic acid, Asp, D
asparagine, Asn, N
cysteine, Cys, C
glutamic acid, Glu, E
glutamine, Gln, Q
isoleucine, Ile, I
leucine, Leu, L
histidine, His, H
lysine, Lys, K
methionine, Met, M
phenylalanine, Phe, F
proline, Pro, P
serine, Ser, S
threonine, Thr, T
tryptophan, Trp, W
tyrosine, Tyr, Y
valine, Val, V
3
Hydrophobic cores are tightly packed
The interior of protein structures is tightly
packed. Water is excluded, except for very few
buried hydration water molecules. Almost all
residues in the interior are hydrophobic or, at
least, uncharged. Charged residues are almost
always on the protein surface. The same rules
apply to protein-protein binding
surfaces. Regular secondary structures form,
because amide groups are polar, seeking H-bonding
partners when buried in a hydrophobic environment.
CPK model of ubiquitin yellow hydrophobic grey
polar but uncharged blue positively charged red
negatively charged green backbone atoms
Backbone H-bonds
Only backbone and hydrophobic sidechains
retained
4
Since peptide bonds are planar (and virtually
always trans), the backbone conformation of
each amino acid is determined by only two
dihedral angles, f and y. Knowledge of the f/y
pairs of each residue is sufficient to define
the 3D structure of the entire backbone!
5
Ramachandran plot
a-helix b-sheet left-handed a-helix
non-Gly residues
Gly
The Ramachandran plot displays experimentally
observed f/y combinations (one dot per
residue). Steric clashes between the side chains
of neighboring amino acids limit the accessible
conformational space. Glycine can access larger
regions in the Ramachandran plot than residues
with longer side chains.
6
Primary, secondary, tertiary, quarternary
structure
Primary structure amino acid sequence Secondary
structure helices, sheets, turns (i.e. regular
sub-structures defined by H-bonds between
backbone amides) Tertiary structure 3D
structure Quarternary structure complex between
different protein molecules (e.g. dimer, trimer,
tetramer)
2 Cys residues can form a disulfide
bridge -CH2-SH 1/2 O2
-CH2-S-S-CH2- H2O
7
a-helix 3.6 residues per turn, H-bonds between
residues i and i4
8
antiparallel b-sheet
parallel b-sheet
in both types of b-sheets, the side-chains point
alternatingly above and below the plane of the
sheet
9
mixed b-sheet
example thioredoxin
10
Leventhals paradox
Assume a small protein with 100 amino acids, each
one of them can access 3 different
conformations 3100 5 x 1047 conformations Fastes
t motions 10-15 sec, so sampling
all conformations would take 5 x 1032 sec 60 x 60
x 24 x 365 31536000 3.1536 x 107 seconds in a
year Sampling all conformations will take 1.6 x
1025 years, much longer than the age of the
universe
In nature, proteins fold correctly within seconds!
The 3D structure is unambiguously encoded in the
amino-acid sequence, but protein structures are
very hard to predict from amino acid sequence,
unless the structure of a similar protein ( 20
amino-acid sequence identity) is known.
Different proteins fold by different,
unpredictable mechanisms (some of them even need
helper proteins (chaperones) to fold. The
current picture is that of a folding funnel,
where the vertical axis displays energy and the
width of the funnel represents the accessible
conformational space.
11
Secondary structure prediction
PHD is best http//npsa-pbil.ibcp.fr/cgi-bin/npsa
_automat.pl?page/NPSA/npsa_phd.html Expected
accuracy 72
10 20 30 40
50 60 70

MGRARDAILDALNLTAEEKLKKPKLELLSVPLREGYGRIPRGALLSMD
ALDLTDKLVSFYLETYGAELTA CCHHHHHHHHHHHHHHHHHHHHHHHHH
hhcchhHhcCcCcHHHHHhcCHHHHHHHHHHHHHHHHhHHHHH NVLRDM
GLQEMAGQLQAATHQ HHHHHHHHHHHHHHHHHHccC   Sequence
length 91   PHD Alpha helix (Hh)
77 is 84.62 310 helix (Gg) 0
is 0.00 Pi helix (Ii) 0 is
0.00 Beta bridge (Bb) 0 is 0.00
Extended strand (Ee) 0 is 0.00 Beta
turn (Tt) 0 is 0.00 Bend region
(Ss) 0 is 0.00 Random coil
(Cc) 14 is 15.38 Ambiguous states(?)
0 is 0.00 Other states 0
is 0.00
MGRARDAILDALNLTAEEKLKKPKLELLSVPLREGYGRIPRGALLSMDAL
DLTDKLVSFYLETYGAELTANVLRDMGLQEMAGQLQAATHQ CCHHHHHH
HHHHHHHHHHHHHHHHHHHhhcchhHhcCcCcHHHHHhcCHHHHHHHHHH
HHHHHHhHHHHHHHHHHHHHHHHHHHHHHHccC --HHHHHHHHHHHHH
HHHHHHHHH---------------HHHHHHHHHHHHHHHHHHHHHHHH
HHHHHHHHHH---HHHHHHHHHH--
12
ca. 20 sequence identity can still result in
similar 3D structure (statistical sequence
identity 6)
Bars identify side-chains with less than 5
solvent accessibility in PYRIN domain Boxes
delineate helix boundaries Vertical grey-shading
conserved core residues
PYRIN DED
CARD
DD
Death domains (DD), death effector domains (DED),
caspase activation and recruitment domains (CARD)
and PYRIN domains form 4 branches of the death
domain superfamily, i.e. their 3D structures are
related while their sequence similarity is
limited (J. Mol. Biol. 332, 1155 (2003)).
13
3D structure prediction
Swiss-Model www.expasy.org/swissmod/SWISS-MODEL.h
tml If a template structure with 25 amino acid
sequence identity is available in the Swiss-Model
database, only the amino acid sequence needs to
be submitted. Otherwise, the 3D coordinates of
the desired template structure must be submitted
too.
Length of target sequence 91 residuesSearching
sequences of known 3D structuresNo suitable
target found Exit
Swiss-Model and any other modelling software
(best regarded is Modeller, www.salilab.org/modell
er/modeller.html) depend crucially on the
sequence alignment. Any model has the same
coordinates for backbone and Cb atoms as the
template. Insertions and deletions are handled
gentlemanly.
14
Identification of similar structures with Dali
FSSP      FAMILIES OF STRUCTURALLY SIMILAR
PROTEINS, VERSION 1.0 (Apr 1 1995) CREATED   Fri
Nov  1 011514 GMT 2002 for dali on
sputnik2-node68.ebi.ac.uk METHOD    Dali ver.
2.0 Holm, L., Sander, C. (1993) J.Mol.Biol.
233,123-138 DATABASE  3241 protein
chains PDBID     6340  HEADER    Structure from
MOLMOL COMPND    pyrin SOURCE    AUTHOR   
SEQLENGTH    90 NALIGN       54 WARNING   pairs
with ZSUMMARY PDB/chain identifiers and structural
alignment statistics   NR. STRID1 STRID2  Z  
RMSD LALI LSEQ2 IDE REVERS PERMUT NFRAG TOPO
PROTEIN    1 6340   6340   23.2  0.0   90    90 
100      0      0     1 S    pyrin    2 6340  
1a1z    8.5  2.2   79    83   19      0     
0     5 S    fadd protein fragment
(fas-associating death domain-con    3 6340  
1ich-A  7.3  2.3   75    87   19      0     
0     6 S    tumor necrosis factor receptor-1
fragment (tnf-1) Muta    4 6340   3ygs-P  6.9 
2.5   78    97   19      0      0     6 S   
apoptotic protease activating factor 1 fragment
procasp    5 6340   1ngr    6.6  2.6   73   
85   10      0      0     5 S    p75 low affinity
neurotrophin receptor fragment    6 6340  
1dgn-A  5.9  2.8   77    89   13      0     
0     6 S    iceberg (protease inhibitor)
fragment    7 6340   1cy5-A  5.5  3.4   79   
92   15      0      0     6 S    apoptotic
protease activating factor 1 fragment (apaf-1   
8 6340   1d2z-B  5.0  2.8   80   150    9     
0      0     7 S    death domain of pelle death
domain of tube    9 6340   3crd    4.6  3.0  
75   100   15      0      0     7 S    raidd
fragment   10 6340   1ddf    4.3  2.9   73  
127   12      0      0     4 S    fas   11
6340   1g71-A  4.0  3.1   68   344   12     
0      0     7 S    DNA primase   12 6340  
1au7-A  3.8  3.5   62   130    6      0     
0     7 S    pit-1 fragment (ghf-1) Mutant
biological_unit DNA   13 6340   1d2z-A  3.7 
2.5   63   102    5      0      0     6 S   
death domain of pelle death domain of tube   14
6340   1dly-A  3.5  3.2   65   121   12     
0      0     5 S    hemoglobin
http//www.ebi.ac.uk/dali/
alternative http//cl.sdsc.edu/
15
Large ribosomal subunit from Haloarcula
marismortui
Science 289, 905 (2000)
16
From Structure to Function
Convergent evolution the overall structures of
chymotrypsin and subtilisin are very different,
but the catalytic triade (Asp, His, Ser
side-chains shown in blue) is conserved
subtilisin
chymotrypsin
17
A mechanism for the evolution of proteolytic
function
FMN-binding protein from Desulfovibrio vulgaris
protease from hepatitis virus
Views differ by a 90o rotation around a vertical
axis.
Catalytically active residues shown as
spheres. The structure is composed of two
domains, each of which is similar to the
FMN-binding protein. The binding site of FMN is
at the site corresponding to the substrate
binding site in the protease.
Two orthogonal views of chymotrypsin backbone
allosteric regulation site
Chymotrypsin consists of two subdomains of
similar structure. The active site is at the
interface. The residues of the catalytic triade
are contributed by both domains.
substrate binding site

• Chymotrypsin is activated by proteolytic cleavage
  of the N-terminal end, resulting in altered
  binding of the new N-terminus to the
• C-terminal domain.

Nat. Struct. Biol. 4, 975 (1997)
18

• Steps for evolution of proteolytic function
• a primordial peptide-binding protein (similar to
  the FMN-binding protein)
• gene duplication, resulting in 2 domains linked
  by a polypeptide chain
• any proteolysis, even if inefficient, decreases
  cooperativity of binding,
• hence peptide fragments dissociate, enabling
  capture of new, uncleaved peptide

gene duplication
No increased chance for proteolytic activity
Proteolysis could occur, if the correct residues
approach the peptide in the cleft between
both domains.
If this evolutionary pathway is correct, what
happened to the peptide binding site that is not
at the interface between the two domains? It
became an allosteric regulation site, i.e.is
still a peptide binding site in a way.