1. Supplementary figure S1. Sequence and secondary structure of wt and R416A influenza A virus nucleoprotein. The difference between the sequences at position 416 is indicated in red. The secondary structure elements derived from the monomer structure are indicated above the sequence and those from the trimer structure below. The structures for the monomer and the trimer are very similar up to residue 386. After this residue the secondary structures are somewhat different although helices 19 and 20 are maintained. Residues 386-401 are represented in green in the monomer and trimer structures in figures 2, 3 and 5, residues 402-428 in yellow and 429-498 in red.

2. Supplementary Figure S2. Characterization by SPR of S165D binding to a 24-mer RNA. A streptavidin sensor chip (SA, Biacore) was functionalized with the 24-mer biotinylated RNA used in a previous work (Flu1 RNA, see reference 20 main text). The protein was injected at increasing concentrations (highest concentration 8 µM, dilution factor 2). The sensorgram with the lowest signal corresponds to the injection of buffer. The sensor chip surface was regenerated with a mixture of 40% formamide, 3.6 M urea and 30 mM EDTA prepared in milliQ water. The sensorgrams could not be fitted by global analysis to determine the rate constants. The signals at the end of the injection were first plotted as the function of the concentration (not shown). Then, in order to determine an apparent dissociation equilibrium constant, a linear function that describes a non specific binding mode was added to the standard Langmuir equation (see reference 51 main text). The insert shows the difference between the total SPR signal and that describing non specific binding as the function of the concentration, yielding K D = 732 ± 39 nM.

3. Supplementary figure S3. Filter binding curves of a radioactive 81 nt panhandle RNA (see reference 1 main text) binding to NP isolated from virus, recombinant wild type NP and recombinant S165D mutant NP. The KD values, given by the protein concentrations at the mid points of the curves, are 30 nM for the viral and recombinant wt NP’s and 250 nM for the S165D mutant NP.

4. AB Supplementary figure S4. Comparison of the distances between nucleoprotein protomers in A the wt trimer crystals of H1N1 A/WSN/33 nucleoprotein (reference 16 main text) and B protomers docked in the cryo electron microscopy reconstruction of A/WSN/33 virion RNPs (reference 33 main text). In the trimer structure in A, the protein densities between the exchange domain and the core of the protein (residues 397-401 and 429-437) are not visible. The distances between the core and the exchange domain residues are: In A (2iqh.pdb) d(396-CA to 402-CA) = 14 Å and d(428-CA to 438-CA) = 21 Å, whereas in B (4bbl.pdb) the distance are: d(395-CA to 405-CA) = 44 Å d(418-CA to 438-CA) = 25 Å.