From Proteopedia

Jump to: navigation, search

Note: This entry on keratins has been published in Biochem. Mol. Biol. Educ.[1]. Please cite it as Biochem. Mol. Biol. Educ. 42:93-4, 2014.

Keratin is the name given to a large family of homologous proteins that have a filamentous (fibrous) structure. These proteins are expressed in epithelial cells and in epidermal cells where they are assembled forming cytoskeletal structures within the cell and epidermal derivatives such as hair, nail and horn [2].

The keratins represent the largest branch within the super-family of intermediate-filament (IF) proteins [3] [4]. Keratins are grouped into two families termed as type I and type II keratins based on their sequence homology [5]. Similarly, other IF proteins are also grouped into families termed consecutively as types III, IV, V and VI IF proteins, based on their sequence homology [6]. These families include desmin, vimentin, neurofilament protein and GFAP that are expressed in specific tissues and cell types [3]. The IF family of lamins are located on the nuclear lamina and are ubiquitously expressed [3].


Intermediate filaments

In most eukaryotic cells there are three major cytoskeletal systems: [7]

  • Microfilaments composed of actin subunits
  • Intermediate filaments
  • Microtubules composed of tubulin subunits

The name "intermediate filament" reflects the comparative morphology of these filaments as their diameter is about 8-12 nm; a value that is "intermediate" between microfilaments with a diameter of 6-7 nm and microtubules with a diameter of 25 nm [8].

Both microfilaments and microtubules are assembled from globular subunits of actin and tubulin respectively. In contrast, intermediate filaments (IFs) are composed of proteins that have a long fibrous structure that results from long stretches of alpha helical domains.

The basic building block of each intermediate filament is a dimer of a coiled-coil pair of IF proteins. Each keratin filament is assembled as a hetero-dimer of a type I keratin coiled together with a type II keratin. [5]. Other types of IFs are mostly composed of homo-dimers [3].

Primary structures of keratins

In humans there are 54 functional genes that code for keratins [9] [10]. The first sequences of human keratin cDNAs revealed that there are two distinct but homologous keratin families [5] [11]. These two distinct types were named as Type I keratin and Type II keratin [5].

Human genome sequencing revealed that type I and type II keratin genes are located in two clusters each of which includes 27 genes on chromosome 17q21 and on chromosome 12q13 respectively [10] [12]. The juxtaposed location of the genes indicate that these gene clusters evolved by a series of gene duplication events.

Determination of the sequences of type I and type keratins revealed that the two types of keratins have a central ~310 residue long segment that share ~30% homology, but the amino and carboxy terminal regions of these proteins show great diversity [11]. Consistent with the initial observations, sequencing of keratins and other intermediate filament proteins showed that all IF proteins have a conserved central domain and widely divergent amino and carboxy terminal regions [13].

Sequencing and two dimensional gel electrophoresis of the complete family of keratins revealed that the type I and type II keratins differ in their size and isoelectric points [14] [15]. Type I keratins are generally smaller (average length 460 aa's), and acidic (isoelectric point 4.4-5.4), while type II keratins are longer (average length 545 aa's) and basic (isoelectric point 5-8.3). As noted, the size differences among keratins result from differences in the amino and carboxy terminals of the proteins [5].

Secondary structures of keratins

Fig. 1. The locations of the α-helical domains (1A, 1B, 2A and 2B) in the central rod of a keratin subunit.
Fig. 1. The locations of the α-helical domains (1A, 1B, 2A and 2B) in the central rod of a keratin subunit.

The first model of alpha-helix was proposed by Pauling based on the crystallography of wool fibers [16] that were shown to have long helical segments [17].

Analysis of the first cytoskeletal keratin sequence revealed that this protein contains a central domain of ~310 residues that was predicted to be mostly in α-helix conformation [11]. By comparative analysis of the predicted structures of a type I keratin, a type II keratin, desmin and vimentin, Hanukoglu and Fuchs suggested that all IF proteins have a central ~310 residue domain that contains four segments in α-helical conformation that are separated by three short linker segments predicted to be in beta-turn conformation [5]. This model has been confirmed by analysis of the crystal structure of segments of keratin coiled-coil [18].

The structures of the head and tail domains of keratins are highly variable and have not been elucidated. Based on their sequences, these domains are predicted to be non-helical, probably forming globular structures that participate in interactions between subunits and other proteins in the scaffold of cellular cytoskeleton [14].

Tertiary and quaternary structures of keratins

Keratin fibers are difficult to solubilize and so far it has not been possible to crystallize a whole keratin or a combination of keratin polymers. In the face of this difficulty, soluble segments of keratins have been generated both by proteolytic digestion and gene engineering to study the structural properties of keratins [19] .

As noted above, keratin filaments are composed of hetero-dimers. To express the long 2B segment of hetero-dimer of keratins K5 and K14, Lee et al. transformed two cDNAs into E. coli, isolated the heteromeric complex, and crystallized it. Structural analysis revealed a coiled-coil hetero-dimer structure of K5 and K14 intertwined around one another. These findings establish that keratin filament is composed of a coiled-coil hetero-dimer wherein the 2B segments are intertwined in parallel [18].

All evidence to date indicates that the basic unit of a keratin filament is a left-handed hetero-dimer of a matched pair of keratins aligned in parallel. The ~10 nm wide keratin filament is assembled in several steps [14]:

  • Hetero-dimer: Formed by the twining of a matched pair of type I and type II keratins that form a coiled-coil.
  • Tetramer: Formed by binding of two hetero-dimers in anti-parallel orientation. The exact mode alignment of the proteins, i.e. which helical domains lie side-by-side, is not known.
  • Octamer: Formed by side-by-side binding of two tetramers containing overall eight keratin molecules. Such an octamer is named a protofibril.
  • Unit length filaments (ULF): Formed by lateral - side by side - association of four protofibrils. In cross-section a protofibril has 32 keratin chains. ULFs are ~60 nm long and ~20 nm wide.
  • Keratin filament: Formed by end-to-end association of ULFs. After assembly, the filament is compacted to a width of 10-12 nm.

Thus, in a general picture, the helical domains of keratins form the backbone of the filaments, and the head and tail domains are involved in the end-to-end linking of the proteins.

Bonds that hold the coiled-coil structure

Fig. 2. Crystal structure of the 2B helical domain of coiled-coil dimer of type I keratin K14 (chain A) and type II keratin K5 (chain B) (residues Ser332-Gly421 of K14 and Thr382-Gly476 of K5). PDB ID: 3tnu. Please click the green colored links in the text in order to view highlighted features of the structure.

The basic building unit of keratin filaments is a hetero-dimer of a type I and a type II keratin. The crystal structure of coiled-coil 2B helical domains of keratins K5 and K14, have revealed the bonds that are involved in tight binding of the two subunits [1].

Prior to enumeration of the bonds involved in keratin-keratin binding, it is essential to understand the structure of α-helix. The backbone of α-helix consists of the atoms that participate in the formation of the peptide bonds that connect the amino acid residues. The helix structure can be visualized as a cylinder around which the chain of residues are wrapped. The central axis of this cylinder defines the central axis of the helix. The R-groups of the residues are positioned perpendicular to the central axis. Thus, the helical surface is covered by the R-groups that protrude outward of the central axis of helix.

Binding of two helical domains in an intertwined structure requires that the surfaces of the helical domains contain atoms or groups that participate in the binding of the two chains.

Protein chains can bind to one another by several types of bonds:

  • Covalent bonds. Example: disulfide S-S bond between two cysteines.
  • Ionic bonds between charged residues with complementary charge. Example: Glu-Arg.
  • Hydrophobic interactions between hydrophobic residues. Example: Leu-Val.
  • Hydrogen bonds between suitable groups.

In the 2B domains of keratins type I K14 and type II K5 shown in Fig. 2, there are two and a single cysteine respectively. These cysteines are far apart and cannot form disulfide bridges.

  • (Wait a few moments for change of scene)

Thus, disulfide bridges cannot be responsible for the binding of K14 and K5.

The second option is ionic bonds, or salt bridges between the two keratins.

Both acidic and basic residues are seen to protrude mostly towards the outside surface of the two keratins and hardly in the space between the two keratins. The contact surface between the two keratins in a coiled-coil is located between the two keratins. Thus, the charged residues do not play a predominant role in the formation of the coiled-coil. In the K14-K5 dimer only 3-4 residues are involved in inter-strand interactions. Nonetheless, these residues are essential for normal function of keratin [1].

Hydrophobic residues: Main points of contact between chains

The third option noted above is hydrophobic interactions between the two keratins.

It can be seen that the hydrophobic residues are predominantly located in the interface between the two chains and essentially occupy the space between these chains. Thus, hydrophobic residues that can associate with one another in the aqueous environment of cell are the main points of contact between the chains in the coiled-coil.

As the two chains of keratins are intertwined in parallel, the contact points along the entire coiled-coil represents a seam along the two proteins. Coiled-coil structures are found in many types of proteins. In two-chained coiled-coil proteins hydrophobic residues appear in a periodic pattern that has been named a heptad-repeat [2]. In a regular α-helix there are 3.6 residues per turn of the helix. In a left-handed coiled-coil there are 3.5 residues per turn. Thus, in a two chained coiled-coil there is a repeat pattern of seven residues that are represented by the letters a-b-c-d-e-f-g. Residues a and d in this pattern are hydrophobic. These two residues define a hydrophobic flank for each protein. This periodic pattern was first reported on both type I and type II wool keratins [3] and later observed on cytoskeletal keratins as well [4]. The crystal structures of the 2B segment of keratins K14 and K5 provided final confirmation for the role of these hydophobic residues in coiled-coil formation [1].

3D structure of keratin

3tnu - hKRT14 residues 295-422 + hKRT5 residues 350-477 - human
3asw, 4f1z - hKRT10 peptide + clumping factor B


  1. Cite error: Invalid <ref> tag; no text was provided for refs named Lee-2012
  2. Woolfson DN. The design of coiled-coil structures and assemblies. Adv Protein Chem. 2005;70:79-112. PMID:15837514 doi:10.1016/S0065-3233(05)70004-8
  3. Elleman TC, Crewther WG, Van Der Touw J. Amino acid sequences of alpha-helical segments from S-carboxymethylkerateine-A. Statistical analysis. Biochem J. 1978 Aug 1;173(2):387-91. PMID:697726
  4. Cite error: Invalid <ref> tag; no text was provided for refs named Hanukoglu-1983

Proteopedia Page Contributors and Editors (what is this?)

Israel Hanukoglu, Michal Harel, Liora Ezra

Personal tools