Python Programming for Biology: Bioinformatics and Beyond

Protein structure

When we determine the amino acid sequence of a protein we gain the knowledge of which

types of residue have been linked into a polypeptide chain. Because we know the chemical

structure of the individual amino acid components, and because protein chains are formed

in a regular and predictable way, we therefore know virtually all of the atoms and covalent

bond  connections

2

 that  are  present  in  the  entire  protein  molecule.  In  general  the  only

deviations  from  this  overall  chemical  structure  will  occur  where  small  parts  of  specific

chemical groups are not static, as when hydrogen ions hop on and off acidic residues, or if

the  protein  is  subsequently  modified  by  enzymes.  Such  post-translational  modifications

include  the  formation  of  cross-links  (between  cysteine  residues),  cutting  of  the  peptide

backbone and the addition of other moieties like sugars, fats and phosphate groups. While

modifications  complicate  the  affair,  if  we  are  unable  to  do  specific  experiments  to

determine  what  has  happened  (e.g.  mass  spectrometry)  then  we  can  often  discover  what

has occurred once we determine the overall three-dimensional structure.

---

We will now consider why proteins fold into their respective shapes. Protein folding is a

deep topic because the number of potential conformations for a typical polypeptide chain

is  vast  and  the  relationship  between  protein  sequence  and  structure  is  generally  not

predictable.  Even  where  it  is  possible  to  investigate  a  sufficient  number  of  hypothetical

three-dimensional  arrangements,  knowing  which  arrangement  is  correct,  the  native

conformation  observed  in  nature,  from  purely  theoretical  considerations  requires

exceedingly  long  computational  calculations.  Fortunately,  in  molecular  biology  we

generally  don’t  have  to  make  such  tricky  predictions  because  we  can  determine  protein

structure by performing experiments and making observations. Because protein folding is

an exceedingly complex topic, most discussions about its mechanisms are well beyond the

remit  of  this  programming  book.  Nevertheless  we  will  describe  some  of  the  basic

principles,  specifically  what  kind  of  forces  are  involved  in  holding  a  protein  structure

together, because this helps us understand the features we observe in structure data.

Overall the folding of molecules can be thought of in terms of energy. The atoms of a

molecule,  because  they  are  in  constant  thermal  motion,

3

 are  able  to  change  relative

position  so  that  the  overall  conformation  moves  towards  the  lowest,  most  stable  energy.

Generally you can think of this as the three-dimensional arrangement that forms the most

stabilising  interactions  between  atoms.  Strictly  speaking  a  molecule  will  not  be  static,  at

its energy minimum, because it will move about due to temperature (it has kinetic energy).

Accordingly,  we  often  think  of  a  molecule’s  native  state  as  being  a  set  of  similar

conformations  that  are  close  to  the  energy  minimum,  albeit  bumbling  about.  It  should

always be remembered that the higher the temperature the wider are a molecule’s motions

and the further it can stray.

Proteins fold into compact, globular structures because of the way amino acids interact

with one another and whether they interact (or do not interact) with water molecules, the

primary  biological  solvent  that  surrounds  them.  Sometimes  a  protein  will  have  cysteine

residues that form covalent disulphide links (under oxidising conditions) that tie different

parts of the protein together, but most of the compactness and precision of folding is due

to  weaker,  non-covalent  interactions,  including  those  with  water  molecules.  In  simple

terms the residues that can form stabilising interactions with water lie on the outside and

those  that  cannot  lie  on  the  inside  (in  the  core).  Admittedly  there  are  some  kinds  of

proteins that aren’t really dissolved in water directly, including those that are embedded in

lipid  bilayers  (the  fatty  membranes  that  surround  cells  and  their  internal  compartments).

However,  even  here  it  is  the  ability  of  particular  amino  acids  to  interact  with  or  avoid

water that is behind the formation of a compact structure.

The atoms around the peptide links, which form the backbone of a protein’s amino acid

chain, are capable of interacting in a stabilising way with water and amongst themselves;

the amide (N-H) and carboxyl (C=O) groups form polar hydrogen bonds. All things being

equal the interaction with water is stronger, but the other parts of the amino acids that stick

out  from  their  backbone,  the  side  chains,  tip  the  balance  so  that  the  protein  backbone  is

mostly stabilised by the backbone atoms hydrogen bonding with each other, and not water.

The  different  amino  acids  have  chemical  structures  that  govern  whether  their  side  chain

can  make  a  significant  interaction  with  water.  Side  chains  containing  atomic  groups  that

can  form  relatively  strong  hydrogen  bonds  (O-H,  N-H,  C=O)  and  those  that  carry  an

electric charge are said to be hydrophilic (water-loving), because they can make stabilising

---

interactions  with  water.  Those  that  do  not  are  described  as  hydrophobic  (water-hating).

Strictly speaking there is not a set dividing line between hydrophobic and hydrophilic; it is

more  a  matter  of  degree.  In  an  aqueous  (water)  environment,  the  hydrophobic  and

hydrophilic  residues  segregate  when  a  protein  folds,  to  form  a  hydrophobic  core  and

hydrophilic exterior, i.e. a globule. This is just a general trend though; the protein globule

is stabilised further by the hydrogen bonds along the backbone, which tend to form regular

patterns  of  hydrogen-bonding  networks,  called  secondary  structure.  Also,  the  electric

charges  and  polarities  will  push  and  pull  the  structure  into  the  final  shape.  This  final

conformation is one where the core residues (mostly hydrophobic) come together and give

rise to another weaker, but widespread, kind of interaction described as the van der Waals

force,  and  thus  the  core  packs  tightly.  This  weak  non-bonding  interaction  is  actually

present  all  the  time  between  close  atoms,  including  those  from  water,  but  in  many

situations it is swamped by other, stronger interactions. As a final point on protein folding,

it  should  be  noted  that  some  large  sections  of  amino  acid  sequences  do  not  have  a

significant hydrophobic component. These regions will typically not form a single stable,

folded structure because they don’t have the ability to form a hydrophobic core. Usually

this results in the region being highly dynamic or unstructured and is commonly seen at

the  ends  of  protein  chains  and  as  flexible  linkers  between  folded  domains,  which  are

compact and globular.

Protein  structure  is  often  described  in  terms  of  a  structural  hierarchy,  which  helps  us

understand  the  final  form  as  a  combination  of  smaller  elements;  which  is  to  say  nothing

about the actual mechanism of folding. This hierarchy is roughly described as follows: