Alternative protein-protein interfaces are frequent exceptions

Abstract

The intricate molecular details of protein-protein interactions (PPIs) are crucial for function. Therefore, measuring the same interacting protein pair again, we expect the same result. This work measured the similarity in the molecular details of interaction for the same and for homologous protein pairs between different experiments. All scores analyzed suggested that different experiments often find exceptions in the interfaces of similar PPIs: up to 22% of all comparisons revealed some differences even for sequence-identical pairs of proteins. The corresponding number for pairs of close homologs reached 68%. Conversely, the interfaces differed entirely for 12-29% of all comparisons. All these estimates were calculated after redundancy reduction. The magnitude of interface differences ranged from subtle to the extreme, as illustrated by a few examples. An extreme case was a change of the interacting domains between two observations of the same biological interaction. One reason for different interfaces was the number of copies of an interaction in the same complex: the probability of observing alternative binding modes increases with the number of copies. Even after removing the special cases with alternative hetero-interfaces to the same homomer, a substantial variability remained. Our results strongly support the surprising notion that there are many alternative solutions to make the intricate molecular details of PPIs crucial for function.

Figure 1

(A) Sketch for interface comparison. Two proteins Px and Py always interact in the same way, do they? We compared pairs of proteins for which we found several experimental solutions for their interaction. Assume that we have two high-resolution protein complexes C1 and C2. From these, we pick two hetero-dimers (Structure A and Structure B) for the interaction between proteins Px and Py (identified by the chains X and Y in Structure A, and by X′ and Y′ in Structure B). We then compared the interface of the same interaction between those two experimental solutions. (B) The PPI network induced by complexes C1 and C2. Complexes C1 and C2 contain two protein-protein interactions: Px-Py and Px-Pz. We differentiated between three types of interface comparisons. First, we only compared interactions corresponding to same pair of sequences (SameSeq; red; shown in A). Then, sequences could change as long as the original proteins remained the same (SameProt; blue; interfaces S1/S3 are compared to S2/S3; both sequences S1 and S3 are variants of protein Px). Finally, we compared interologous interactions (green; interfaces Px/Py are compared to Pz/Py; Px and Pz come from the same family).

Figure 2. Filtering out interface diversity introduced by homomers.

Assume you want to compare an interface Px-Py in complex 1 to the interfaces in complex 2. Usually, you will calculate two similarities (0.1 and 0.6), because there are two Px-Py interfaces in complex 2. Looking for homomers, you will find the two sequence identical Px chains in complex 2 interacting and connecting the two Px-Py interfaces. Now, you can correct the comparison by using only the one best match (0.6). The comparisons of the “worse” alternatives are discarded.

Figure 3. Faces are similar yet different.

For two different ways to measure interface similarity - Face Position Similarity [A] and L_rms [B] - we present the similarity distribution for all interfaces. The rightmost interval shows largely identical faces, the leftmost completely different faces. For each similarity range and measure, there are three bars: one for each type of sequence divergence (D-SameSeq to D-Interolog). For example, Face Position Similarity finds about 7% of all the interface similarities at D-SameProt to fall in the range 0.0–0.1, i.e. suggests in 7% of the cases completely different outcomes when experimentally measuring the same interaction again. The error bars show standard errors and are explained in Section S1.3 in Text S1. The inlet displays the cumulative distribution giving the fraction of all similarities that differ by a certain value. For instance, 21% of all interface comparisons result in a value above 2 Å according to the L_rms in D-SameProt. In these cases, the two smaller proteins are clearly not in the same position after superimposing the two larger proteins.

Figure 4. Three typical interactions exhibiting surprising variety.

(A) Protein ‘ras’ binds to ‘son of sevenless’ (1NVV): alternative binding for sequence-identical pairs of proteins and without a multimeric context; the lower left panel shows the residues of the two interfaces in purple and red. (B) Natural dimeric interactions between proteins from the protein kinase and cyclin families (interface copy number 1; e.g. 1OI9). Cyclin chains (green) have been structurally aligned and superimposed. Protein kinases (cyan and blue) were subject to the same geometric translations. The blue chain has a recently discovered outlier interface (see text). (C) Superimposition of entire sequence-identical F1-ATPase complexes. Complexes were aligned and superimposed with the gamma chains (green). Alpha (orange) and beta (cyan) subunits were subject to the same geometric translations. In the main panel, we look at the complexes from the top. The inlet displays an interaction between a beta and a gamma subunit from the side.