The following queries are supported. You may click on an example query to see its results

Definition of used Terms

Feature overview

Column headers


The multiple sequence alignment of interaction interfaces of complete families in combination with the classification of distinct interaction sites is the unique feature of SCOPPI. The following examples will illustrate this feature and show how it can provide interesting insight into the available data.

Viruses mimick interfaces

Chemokines play a key role in leukocyte recruitment and migration. Consulting SCOPPI for the Interleukin-8 like chemokines family (d.9.1.1) for interacting partners reveals interactions with the viral chemokine binding protein M3 (b.116.1.1) and with d.9.1.1 itself. Interestingly, the same interface type, namely 1, is used for both viral protein M3 and its homodimer:

Alexander et al. [2] report that viral protein M3 sequestrates chemokines with high affinity due to conformational flexibility and electrostatic complementation. Knowing that the binding occurs at the same binding site, interface similarities between the ligands are expected. Indeed, as seen in figures below, there is an optimal fitting to a binding site that is utilised by chemokines to form homodimers. To achieve this, the virus has evolved the SVSPLP motif which can play the role of the native motif SSDTTP.

Left: The viral protein M3 (yellow) adapting to human chemokine (blue), mimicking chemokines homodimer (green). The backbone of M3 between Ser269 to Pro274 (motif SVSPLP) fits nicely to Ser4 to Pro9 (motif SSDTTP) in chemokine. (Source: PDB structures 1ml0 and 1cm9 superimposed.)
Right: similarity between another instance of a chemokine homodimer and M3, both utilizing prolin to fill the cleft (1ml0 and 1b3a superimposed).

Where do cytokines bind to their receptors?

Some families display a considerable variety of different interface types when interacting with members of another family. If we query SCOPPI for the Long-chain cytokines (a.26.1.1), we see such an example. Domains of this family appear as part of human cytokines, human growth hormone and prolactin. When interacting with their receptors, this domain interacts with the fibronection type III domain of the receptor. As we can see from the SCOPPI result (Interface Type 1 column on the right), there are 11 different interaction sites where fibronectin domain binds to the cytokine domain:

To confirm this we displaying the structures of the above PDB files in PyMOL and structurally align the cytokine domains (shown in red) with multiprot. As we see can see, there are indeed many distinct interaction sites:

How well conserved is the trypsin pocket?

The family of eukaroytic proteases (b.47.1.2) contains trypsin-like serine proteases. The active site of these enzymes involves three residues forming a catalytic triad: His - Asp - Ser (in sequential order). The triad is H-bonded: Asp is H-bonded to His to keep it oriented, His is H-bonded to Ser. Because of the importance of these three residues, we would expect them to be conserved through all members of the serine protease family.
To verify this, we enter b.47.1.2 in SCOPPI, select a redundancy level of 50% for a better overview and select "Conservation" from the color select box.
Scrolling down to the interacting partner g.15.1.1 (Animal Kazal-type inhibitors), we see
The conservation percentage is calculated by counting the number of residues of the same type in that column divided by all residues in that column. Residues with a value above 90% will display in red, those with a value below 10% in purple. A color legend popup is available through a hyperlink next to the selection boxes.

Considering the serine proteases, we can spot a highly conserved region AAHC with the catalytic His residue (catalytic residues are markes with an asterisk in the screenshot above). Asp (D) is also well conserved. Serine is right in a conserved GDSGGP motif. We notice, however, that the serine is not fully conserved -- between 10% and 20% of the family members are missing the serine at this position. Closer examination reveals that in these cases, serine has been mutated to alanine. Therefore, we get a blurred picture of what we consider as conservation.


Gene fusion

With SCOPPI, you can check if interactions between two domain families occur as inter as well as intra interactions (i. e., if there are examples where the interacting domains are on a single polypeptide chain and where they are on separate polypeptide chains.) If at the same time one observes the same interface type (i. e. the same structural orientation, please see above for detailed definition), the reason behind this observation might be a gene fusion event.

Here we see an example of domains of the c.1.2.1 family (Histidine biosynthesis enzymes) interacting with domains of family c.23.16.1 (Class I glutamine amidotransferases (GAT)). As you can see in the Interaction Type column on the right, there are inter and intra cases that share the same interface types (1, 3). This is also seen nicely by looking at the highlighted interacting residues of the aligned sequences (you may click on the picture to see the results as they appear in SCOPPI).

Let's take a look at the structural level. We take two PDB files of the listed cases above, 1gpw and 1ox4, and display the interacting domains. Both PDB files describe the crystal structures of Imidazole Glycerolphosphate Synthase, which catalyses formation of the imidazole ring in histidine biosyntheseis. 1gpw shows the enzyme in T. maritima (A), and 1ox4 the enzyme in S. cerevisiae (B). The functional enzyme consists of a glutamine amidotransferase domain (top), and a cyclase domain (bottom). In T. maritima, a hyperthermophile bacterium, these domains are located on two separate polypeptide chains, forming a heterodimeric protein. In yeast, the two domains are fused together. It is known that this fusion is common in plants and fungi (Chaudhuri et al.). SCOPPI nicely picks up this example by its classification of geometrically distinct interface types between domain familes. In total, we identify 59 of such examples.

Imidazole Glycerolphosphate Synthase as heterodimeric protein in bacteria (A) and as fusion protein in yeast (B). The Polypeptide chains of 1gpw (A) and 1ox4 (B) are shown in rainbow colors with blue N-Terminus and red C-terminus. Screenshots were generated using PyMOL (

Data sources used


  1. Kottha S, Kuehn M, Schroeder M. Size matters: How to classify permanent and transient protein interactions. (submitted)
  2. Chaudhuri BN, Lange SC, Myers RS, Davisson VJ, Smith JL. Toward understanding the mechanism of the complex cyclization reaction catalyzed by imidazole glycerolphosphate synthase: crystal structures of a ternary complex and the free enzyme., Biochemistry 2003 Jun 17;42(23):7003-12.
  3. Alexander JM, Nelson CA, van Berkel V, Lau EK, Studts JM, Brett TJ, Speck SH, Handel TM, Virgin HW, Fremont DH. Structural basis of chemokine sequestration by a herpesvirus decoy receptor. Cell 2002 Nov 1;111(3):343-56.