Any number of amino acids can chain together by successive peptide bonds. For example, the linkage of three amino acids by two peptide bonds
The alpha carbons from each amino acid alternate with the peptide bonds to form the “backbone” of the peptide.
A similar linkage between a large number of amino acids forms polypeptides, which are called proteins when they are large enough and have a defined three-dimensional structure.
This ball-and-stick model of a tetrapeptide displays only what we call the peptide backbone: the alpha carbons, the atoms that take part in the peptide bonds (—CO—NH—) and the amino and carboxyl terminal groups. Neither the sidechains nor the other hydrogens are displayed, and only one of the oxygens of the terminal carboxyl group is shown. By moving the model around you can confirm that each peptide bond is a planar structure ; in it, the oxygen and the hydrogen are located in a ... all cis all trans different for each one configuration
At position nº 1 (terminal amino), let's add as sidechain a -CH2-CH2-CH2-CH2-NH2 group. The resulting amino acid is
At position 2 let's put as sidechain a methyl group, giving the amino acid
For the third amino acid, the sidechain is -CH(CH3)-CH2-CH3; so it is
Finally, at position 4 (terminal carboxyl) let's put -CH(OH)-CH3 to form
you can see the shape of the tetrapeptide using the spacefilling rendering. (Once again, not all hydrogen atoms are displayed).
Spacefilling, ball-and-stick and even sticks-only renderings show too much detail to be useful with large proteins. That's the reason why it is common to use more schematic models where the peptide chain is represented by a line that connects the alpha carbons of all the amino acids. Such a line (with diverse formats) represents the peptide backbone: just the schematic rendering of the peptide backbone.