NMR studies on structure and dynamics of the monomeric derivative of BS-RNase: new insights for 3D domain swapping

Abstract

Three-dimensional domain swapping is a common phenomenon in pancreatic-like ribonucleases. In the aggregated state, these proteins acquire new biological functions, including selective cytotoxicity against tumour cells. RNase A is able to dislocate both N- and C-termini, but usually this process requires denaturing conditions. In contrast, bovine seminal ribonuclease (BS-RNase), which is a homo-dimeric protein sharing 80% of sequence identity with RNase A, occurs natively as a mixture of swapped and unswapped isoforms. The presence of two disulfides bridging the subunits, indeed, ensures a dimeric structure also to the unswapped molecule. In vitro, the two BS-RNase isoforms interconvert under physiological conditions. Since the tendency to swap is often related to the instability of the monomeric proteins, in these paper we have analysed in detail the stability in solution of the monomeric derivative of BS-RNase (mBS) by a combination of NMR studies and Molecular Dynamics Simulations. The refinement of NMR structure and relaxation data indicate a close similarity with RNase A, without any evidence of aggregation or partial opening. The high compactness of mBS structure is confirmed also by H/D exchange, urea denaturation, and TEMPOL mapping of the protein surface. The present extensive structural and dynamic investigation of (monomeric) mBS did not show any experimental evidence that could explain the known differences in swapping between BS-RNase and RNase A. Hence, we conclude that the swapping in BS-RNase must be influenced by the distinct features of the dimers, suggesting a prominent role for the interchain disulfide bridges.

Figure 1. NMR derived structure of mBS.

A) Bundle of the best 10 structures of mBS calculated out of 2252 distance constraints (pdb ID code 2lfj); B) protection factors mapped on mBS structure with the following colour code: residues with a P>1×10⁵ in red, the one with 10×10³<P<1×10⁵ in orange and the one with P<10×10³ in yellow. Residues that couldn't be analyzed are shown in grey.

Figure 2. Relaxation data of mBS.

S² order parameters of mBS as obtained from the model-free analysis of Lipari and Szabo. (Details are given in the text.) Error bars within the 90% confidence limit were calculated with Montecarlo method Outliers according to R2/R1 rates (residues 70–72) were intentionally omitted from the simultaneous fit of the global correlation time, effective correlation times and order parameters. The mobility of these residues is better demonstrated with the reduced spectral density mapping (see supplementary material).

Figure 3. Theoretical Molecular Dynamics of mBS.

Root mean square fluctuations (rmsf) values vs. mBS sequence.

Figure 4. Urea perturbation of mBS structure.

A) On mBS structure are shown in red residues whose intensities were still detectable at 6 M urea; B) changes in the normalized NMR peak intensities for selected residues across the range of urea concentration studied. The line/symbol code for each residue is the following:▪ K39, ♦ S77, ▴ L32, ▵ K26, * H12, ▴ G111, dashed line V118, ▪ V54, dashed-dotted line G88, ♦ D83, ○ E49, □- - T36.

Figure 5. Tempol surface mapping of mBS.

Paramagnetic attenuations, Ai, are reported for each well resolved ¹H-¹⁵N HSQC signal of mBS (filled triangles). Histogram heights refer to the fractional freedom from intramolecular hydrogen bonding predicted by the 100 ns MD simulation in explicit water (grey bars).