Biophysical Society On-line Textbook

PROTEINS

CHAPTER 1. PROTEIN STRUCTURE

Section 1.

Primary structure, secondary motifs,

tertiary architecture, and quaternary organization

Jannette Carey* and Vanessa Hanley^

*Department of Chemistry

^Department of Chemical Engineering

Princeton University

Princeton, NJ 08544-1009

*corresponding author

(609) 258-1631 phone

(609) 258-6746 FAX

carey@chemvax.princeton.edu

1. The amino acid building blocks

Proteins are polymeric chains that are built from monomers called amino

acids. All structural and functional properties of proteins derive from the chemical

properties of the polypeptide chain. There are four levels of protein structural

organization: primary (1°), secondary (2°), tertiary (3°), and quaternary (4°). Primary

structure is defined as the linear sequence of amino acids in a polypeptide chain.

The secondary structure refers to certain regular geometric figures of the chain.

Tertiary structure results from long-range contacts within the chain. The quaternary

structure is the organization of protein subunits, or two or more independent

polypeptide chains.

Amino acids are the chemical constituents of proteins, and are characterized

by a central alpha carbon atom. The alpha indicates the priority position from which

the numbering follows for all subordinate groups. Four substituents are connected

to this Cα: one substituent is the alpha proton -H, another is the side chain -R that

gives rise to the chemical variety of the amino acids, the third is the carboxylic acid

functional group (-COOH), and the fourth is the amino functional group (-NH). The

α carbon is the asymmetric center of the molecule for all 20 amino acids except

glycine, which has only a proton as its side chain. The configuration about the α

carbon center must be the L-isomer for proteins synthesized on the ribosome. This is

probably an accident of chemical evolution where the L-isomer happens to be the

one chosen for early prebiotic systems and fixed into evolutionary history.

figure 1

The side chains have a wide chemical variety which is vital for the unique

functions of biological proteins (Figure 1). These side chains can be grouped into

three categories: nonpolar, uncharged polar, and charged polar. The simplest amino

acid is glycine. Alanine, valine, leucine, isoleucine, and proline are amino acids

whose side chains are entirely aliphatic. Alanine has a methyl group as its side

chain. Valine has two methyl groups connected to the β carbon, and this residue is

said to be β-branched. Leucine has one more carbon atom in the side chain than

valine, so that two methyl groups are attached to Cγ. Leucine and isoleucine are

isomers whose only difference in structure is the position of the methyl groups.

Isoleucine is a β-branched amino acid and has a second asymmetric center at the β−

carbon. Proline contains an aliphatic side chain that is covalently bonded to the

nitrogen atom of the α-amino group, forming an imide bond and leading to a

constrained 5-membered ring.

Side chains that are generally nonpolar have low solubility in water because

they can form only van der Waals interactions with water molecules. On the other

hand, the rest of the amino acids contain heteroatoms in their side chains, opening

many bonding possibilities. The uncharged members of this group include: serine,

threonine, asparagine, glutamine, tyrosine, and tryptophan. Serine, threonine, and

tyrosine contain hydroxyl groups so they can function as both hydrogen bond

donors and acceptors, and threonine also has a methyl group, making it β-branched.

The benzene ring of tyrosine permits stabilization of the anionic phenolate form

upon loss of the hydroxyl proton, which has a pKa near 10. Serine and threonine

cannot be deprotonated at ordinary pH values. Asparagine and glutamine side

chains are relatively polar in that they can both donate and accept hydrogen bonds.

The nitrogen and proton of the tryptophan indole side chain can also participate in

hydrogen bond interactions.

Another group of polar residues can bear a full, formal charge depending on

pH, but their pKa values are such that the charged form is largely populated near

neutral pH's. These include lysine, arginine, histidine, aspartic acid, glutamic acid,

and cysteine. Lysine and arginine are two basic amino acids that can bear a positive

charge at the end of their side chain. The lysine α-amino group has a pKa value near

10, while the arginine guanidino group has a pKa value of ~12. Histidine is another

basic residue with its side chain organized into a closed ring structure that contains 2

nitrogen atoms. One of these nitrogens already has a proton on it, but the other one

has an available position that can take up an extra proton and form a positively

charged histidine group with a pKa of about 6. Aspartic acid and glutamic acid differ

only in the number of methylene (-CH2-) groups in the side chain, with one and two

methylene groups, respectively. Their carboxylate groups are extremely polar and

can both donate and accept hydrogen bonds, and have pKa values near 4.5.

The sulfhydryl (thiol) group of cysteine can ionize at slightly alkaline pH

values, with a pKa near 9. The thiolate form can react with a second sulfhydryl to

form a disulfide bond that is reversible by reduction. Methionine has a long alkyl

side chain that also contains a sulfur heteroatom but is hydrophobic. This sulfur

atom is relatively inert as a hydrogen bond acceptor.

A group of three amino acids that all have aromatic side chains are

phenylalanine, tryptophan, and tyrosine. The aromatic ring of phenylalanine is like

that of benzene or toluene. It is very hydrophobic and chemically reactive only

under extreme conditions, though its ring electrons are readily polarized. The side

chain of histidine is arguably considered aromatic; it meets the electron rule in one

of its protonation states, but does not have the characteristic strong near-UV

absorption of the other three aromatic amino acids. The UV spectra of the three

aromatic groups are distinctive, as are their extinction coefficients, and these

properties are reflected in the electronic spectra of polypeptides.

Finally, the α-amino and α-carboxylate groups of amino acids can also ionize,

with pKa's of 6.8-7.9 and 3.5-4.3, respectively, for the aliphatic amino acids; nearby

charged side chains can alter the pKa's of these groups. Each amino acid

incorporated into a polypeptide chain is referred to as a residue. Thus, only the

amino- and carboxyl- terminal residues possess available α−amino and α−

carboxylate groups, respectively.

2. The polypeptide chain

In order to form the amino acid monomers into a polymeric chain, amino

acids are condensed with one another through dehydration synthesis. This reaction

occurs when water is lost between the carboxylic functional group of one amino acid

and the amino functional group of the next to form a C-N bond. These

polymerization reactions are not spontaneous; however they can be arranged to

occur through the energy-driven action of the ribosome. Ribosomes are complexes

of proteins and RNA that translate a gene sequence in the form of mRNA into a

protein sequence. The 20 amino acids listed above are encoded by the genes and

incorporated by the ribosomal machinery during protein synthesis. Other minor

amino acids are incorporated by ribosomes, but are derived by post-translation

modifications.

The reverse reaction, involving hydrolysis of the peptide bond, is not

spontaneous either. It can be accomplished chemically, but only under very

vigorous conditions. For example, treatment with very strong acid (1 molar HCl)

and boiling at 100°C overnight can hydrolyze the peptide bonds. So, the reverse

hydrolysis reaction actually happens very slowly under normal conditions. Thus,

proteins are chemically and biologically stable unless they are deliberately

depolymerized. The decomposition of a polypeptide chain into individual amino

acids can also be facilitated by hydrolytic enzymes.

Most proteins are heteropolymeric (i.e., they contain most or all the different

amino acids). Only rarely do regions of proteins consist of sequences composed of

just a few amino acids. Any region of a typical protein will therefore have a

chemically heterogeneous environment. This heterogeneity is further amplified by

the higher levels of protein structure, as we will see.

3. The peptide bond

The peptide bond between two amino acids is a special case of an amide bond

flanked on both sides by α-carbon atoms. Peptide bond angles and lengths are well-

known from many direct observations of protein and peptide structures. The

peptide bond (C-N) length is observed to be 1.33Å (Figure 2A). This is considerably

shorter than the adjacent (nonpeptide) C-N bond length of 1.45Å, but longer than

the C=O bond length of 1.23Å.

Figure 2A

Figure 2B

These bond lengths and angles reflect the distribution of electrons between

atoms due to differences in polarity of the atoms, and the hybridization of their

bonding orbitals. The two more electronegative atoms, O and N, can bear partial

negative charges, and the two less electronegative atoms, C and H, can bear partial

positive charges. The peptide group consisting of these four atoms can be thought of

as a resonance structure. (Figure 2B) Thus, the peptide bond has partial double bond

character, accounting for its intermediate bond length.

Like any double bond, rotation about the peptide bond angle ω is restricted,

with an energy barrier of ˜3 kcal/mole between cis and trans forms. These two

isomers are defined by the path of the polypeptide chain across the bond. (Figure3)

Successive α−carbons in the chain (i, i+1) are on the same side of the bond in the cis

isomer as opposed to the staggered conformation of the trans isomer. For all amino

acids but proline, the cis configuration is greatly disfavored because of steric

hindrance between adjacent side chains. Ring closure in the proline side chain

draws the β−carbon away from the preceding residue, leading to lower steric

hindrance across the X-pro peptide bond. In most residues, the trans to cis

distribution about this bond is about 90 - 10, but with proline, the trans to cis

distribution is about 70 - 30.

Figure 3

Also like any other double bond, certain atoms are confined to a single plane

about the peptide bond. The group of six atoms between successive α-carbon atoms,

inclusive, lie in one plane exactly as do the six atoms of ethene. These six atoms are

shown in figure 3. In the trans configuration of the peptide bond, the combined

effects of polarity and planarity result in a permanent small dipole moment across

the peptide bond, with its negative end on the side of the carbonyl oxygen. The

planarity of the peptide bond has additional profound consequences for polypeptide

structure, as we shall see.

4. Restrictions on bond rotations

While there is restricted rotation about the peptide bond, there is free rotation

about the four bonds to the α-carbon of each residue. Two of these rotations are of

particular relevance for the structure of the polypeptide backbone. To fully

appreciate these rotations, we must shift our perspective from the peptide-bond-

centered view of figure 3 to the Cα-centered view of figure 4A. The bond from the

α−carbon to the carbonyl carbon of that residue is given the name ψ. Similarly, the

bond from the α−carbon to the amino group of that residue is given the name φ.

Because Cα is one of the six planar atoms of the peptide group, rotation about φ or ψ

flanking Cα rotates the entire plane of the peptide group (Figure 4B).

Figure 4A Figure 4B

Since the entire plane rotates on either side of Cα, certain values of the angles

φ and ψ cannot be achieved due to steric occlusion. The allowed regions of φ,ψ space

differ for each amino acid because of the restriction due to Cα and its substituents.

However, even for glycine, some angles are not allowed.

Figure 5A

Figure 5B

The allowed regions of φ,ψ space for each amino acid are displayed on

Ramachandran plots. The allowed regions can be defined in terms of the energetic

cost that must be paid to enter a disallowed region (Figure 5B), or in terms of the

limiting so-called hard-sphere boundary when atoms clash (Figure 5A). For β-

branched residues the restrictions are severe, and only a small fraction of φ,ψ space is

allowed. Valine and isoleucine have access to only about 5% of all φ,ψ space.

However, all residues have access to at least part of the most favorable regions of φ,ψ

space in the upper and lower left of the plot. As we will return to shortly, it turns

out that these two regions correspond to combinations of φ and ψ angles that

characterize the two common regular secondary structures that can be adopted by

the polypeptide backbone, the α-helix and β-strand.

Note that there is an energy barrier between the α-helical region of φ,ψ space

and the β-strand region. Thus, direct conversion between α- and β- structures is

restricted even though most residues are allowed in both regions. Two conclusions

from recent structural analysis of proteins and peptides are relevant to this point.

First, the peptide bond deviates slightly from planarity in a surprisingly large

fraction of cases (10). Presumably, the observed range of peptide bond angles has the

effect of slightly enlarging the allowed φ,ψ space, and perhaps reducing the α/β

barrier. Second, in protein structures certain residues are overrepresented outside

the allowed regions, and these tend to be the small polar residues (11). Presumably,

these can form favorable local interactions that compensate for the energetic penalty

in those φ,ψ regions.

5. Secondary structures

Since the restrictions on φ,ψ space arise in part from steric hindrance between

side chain and backbone, this same steric hindrance is the origin of α and β

secondary structures. There is no sequence dependence on the steric restrictions of

the α and β space because φ,ψ restrictions arise within each residue rather than

between residues. However, a sequence of residues that all have similar allowed φ,ψ

space can give rise to a chain segment that forms α or β structures. Thus, these

secondary structures owe their formation to both backbone and side chain steric

restrictions. This analysis provides an important insight into the origins of protein

secondary structures: these structures are intrinsically favorable for the chain under

all conditions, independent of considerations about bonding.

The helix structure looks like a spring. The most common shape is a right

handed α−helix defined by the repeat length of 3.6 amino acid residues and a rise of

5.4 Å per turn. Thus residues (i+3) and (i+4) are closest to residue (i) in the helix

(Figure 6A). The pitch and dimensions of the helix also bring the peptide dipole

moments of successive residues into proximity such that their opposite charges

neutralize each other substantially in the middle of a helix (Figure 6B). At the ends,

the peptide dipole cannot be neutralized by this mechanism, resulting in a net helix

macrodipole of approximately one-half unit of charge at each end. This charge may

be neutralized by nearby side chains.

Figure 6A

Figure 6B

The pitch and dimensions of the helix also bring the amide proton of residue

(i+3) or (i+4) into proximity to the carbonyl oxygen of residue (i) such that a

hydrogen bond can form. All peptide group hydrogen bond donors and acceptors are

satisified in the central part of the helical segment, but not at the ends. While

structural evidence clearly indicates that these hydrogen bonds are highly populated

in helical segments of proteins, their contribution to helix stability is less clear since

donors and acceptors would be satisfied by hydrogen bonding to water in nonhelical

structures. However, φ,ψ restrictions can have the effect of preorganizing the chain

into a helical conformation, which may favor hydrogen bonding by enhancing the

local concentration of donors and acceptors.

β strands are the other regular secondary structure that proteins form (Figure

7A). These are extended structures in which successive peptide dipole moments

alternate direction along the chain. Because it is an extended structure, φ,ψ steric

hindrance is reduced in the β strand, and the β region of φ,ψ space is larger than the

α region. Two or more strand segments can pair by hydrogen bonding and dipolar

interactions to form a β-sheet. Unlike helical segments, all peptide group hydrogen

bond donors and acceptors are satisfied not within but between β-strand segments;

thus individual β-strands do not have an independent existence.

Also unlike a helical segment, adjacent strands of a sheet can come from

sequentially distant segments of the chain; rarely, this can occur even within one

strand of a sheet. β sheets can consist of either parallel or antiparallel strands, or a

mixture of the two. In purely antiparallel sheets, segments that are sequentially next

to each other in the primary structure often form adjacent strands.

Figure 7A

Figure 7B

However, even when forming a hairpin from contiguous chain segments, linearly

distant residues are brought into proximity at the N-and C- terminal ends of the

hairpin (Figure 7B). Thus, while a β-strand is a secondary structure element because

of its geometrically regular features, a β-sheet can be thought of as a tertiary

structural feature because it is intrinsically nonlocal. This example illustrates that

the distinctions between secondary and tertiary structural features are not entirely

clear.

So-called turn structures are also classified as secondary structural elements,

but unlike helices and strands, they do not have a repeating, regular geometry.

Rather, they can have well-defined spatial dispositions defined by certain values of φ

and ψ angles that often require specific residue types and/or sequences, as well as

fixed hydrogen bonding patterns. Most turns are local in the primary structure, but

omega loops (12) can have a large number of intervening residues lacking defined

geometries, with the turn being defined by the conformations of residues that form

the constriction that gives this turn its name (Ω).

Turns are essential for allowing the polypeptide chain to fold back upon itself

to form tertiary interactions. Such interactions are generally long-range, and result

in compaction of the protein into a globular, often approximately spherical, form.

The turn regions are thus generally located on the outside of the globular structure,

with helices and/or sheets forming its core. Turns on the surfaces of proteins have a

wide range of dynamics, from quite mobile in cases where they form few

interactions with the underlying protein surface to quite fixed due to extensive

tertiary contacts. Thus, turns are also ambiguously classified as secondary structure

elements.

6. Tertiary structures

The side chains project outward from both α-helical (figure 6A) and β-strand

(figure 7A) structures, and are therefore available for interactions with other

surfaces through hydrophobic contacts and various kinds of bonding interactions to

form the tertiary structure. In a helix, the side chains project radially outward, and

in a strand successive side chains project alternately up and down. Rotation about

bonds in the side chain are also restricted, however. The same steric hindrance that

limits the backbone conformation also limits the side chain conformation about the

Cα-Cβ bond to preferred rotamers defined by rotation angle χ1. Rotation angles

beyond χ1 are restricted by side chain packing in the tertiary structure.

If secondary structural elements result from steric restrictions in φ,ψ space, it

is less obvious why tertiary structures form. Proteins with highly organized tertiary

structures generally have a well-developed core of hydrophobic residues contributed

from most or all of the secondary structure elements in the chain. Thus, secondary

and tertiary structures are in general intimately and explicitly interconnected. These

buried residues do not form merely a liquid-like oily interior, but rather are usually

well-packed, with extensive rotamer restrictions. In aqueous solvents, the

hydrophobic effect drives the chain toward compaction to relieve unfavorable

solvation of these exposed side chains, but compaction and internal organization are

entropically costly due to loss of chain flexibility, and it is likely that these competing

effects nearly cancel each other energetically.

On the other hand, upon compaction, bonding interactions with solvent

molecules are replaced by intramolecular partners, with a likely net gain in

favorable energetic contributions due to several effects, including lower dielectric

constant in the incipient interior. Hydrogen bonding is favored within secondary

structures because these are partially preorganized by φ,ψ restrictions into

configurations that permit bonding at little additional entropic cost. In the case of β-

sheet formation, an additional favorable effect may result when two β-strands are

brought into register, much like DNA duplex formation.

The view developed in these paragraphs suggests that protein secondary and

tertiary structures are not independent of each other, but rather interdependent. It

seems likely that this interdependence is the molecular origin of the extraordinary

cooperativity of protein structural stability, which is reflected in the observation that

protein secondary and tertiary structures are lost concomitantly and in an all-or-

none manner upon changes in environment that disfavor the folded state, such as

higher temperature or solvent additives.

7. Quaternary structure

The highest level of protein structural organization is the quaternary structure. The

subunits that associate may be identical or not, and their organization may or may

not be symmetric. In general, quaternary structure results from association of

independent tertiary structural units through surface interactions, such as

formation of the hemoglobin tetramer from myoglobin-like monomers. However,

an increasing number of examples illustrates that tertiary structure can also be

formed concomitantly with quaternary association in some cases. A notable example

is the tryptophan repressor protein, which forms a highly intertwined dimer in

which essentially all tertiary contacts are satisfied only across the subunit interface,

rather than within each polypeptide chain (13). Thus, subunit assembly is

necessarily a step in tertiary structure formation. Another example is the cyclin/Cdk

inhibitor, which like Trp repressor has a well-formed secondary structure but no

intramolecular tertiary structure; rather, all tertiary interactions are formed through

its contacts to the binary cyclin/Cdk complex (14). These examples show that the co-

dependence of tertiary and quaternary structures parallels the co-dependence

between secondary and tertiary structures, and suggest that the distinction among

these levels of the protein structure organizational hierarchy are blurry at best, and

perhaps even misleading for our understanding of protein structural stability and

folding.

8. Literature cited

1. Creighton, T. E. (1983) in Proteins: Structures and Molecular Properties, W.H.

Freeman and Company, New York. pg. 3.

2. ibid., pg. 5.

3. Cantor, C. R., and Schimmel, P. R. (1980) in Biophysical Chemistry, W.H.

Freeman and Company, New York. pg. 41.

4. Creighton, T. E. (1983) in Proteins: Structures and Molecular Properties, W.H.

Freeman and Company, New York. pg. 174

5. Cantor, C. R., and Schimmel, P. R. (1980) in Biophysical Chemistry, W.H.

Freeman and Company, New York. pg. pg. 165.

6. Creighton, T. E. (1983) in Proteins: Structures and Molecular Properties, W.H.

Freeman and Company, New York. pg. 7.

7. ibid., pg. 160.

8. Cantor, C. R., and Schimmel, P. R. (1980) in Biophysical Chemistry, W.H.

Freeman and Company, New York. pg. pg. 256.

9. Creighton, T. E. (1983) in Proteins: Structures and Molecular Properties, W.H.

Freeman and Company, New York. pg. 167.

10. MacArthur, M.W., and Thornton, J.M. (1996) J. Mol. Biol. 264, 1180-1195.

11. Gunasekaran, K., Ramakrishnan, C., and Balaram, P. (1996) J. Mol. Biol. 264, 191-

198.

12. Fetrow, J. S. (1995) FASEB J. 9, 708-717.

13. Schevitz, R.W., Otwinowski, Z., Joachimiak, A., Lawson, C.L., and Sigler, P.B.

(1985) Nature 317, 782-786.

14. Russo, A.A., Jeffrey, P.D., Patten, A.K., Massague, J., and Pavletich, N.P. Nature

382, 325-331.