Structure, Volume 21

Supplemental Information

Functional Classification

of Immune Regulatory Proteins

Rotem Rubinstein, Udupi A. Ramagopal, Stanley G. Nathenson, Steven C. Almo, and Andras Fiser Inventory of Supplemental Information List of Supplementary Figures: Figure S1, related to Figure 1 Figure S2, related to Figure 1 Figure S3, related to Figure 1 and 5 Figure S4, related to Figure 1 and 3 Figure S5, related to Figure 3 Figure S6, related to Figure 3 Figure S7, related to Figure 5 Supplementary Tables Table S1, related to Figure 5

Supplementary Information List of clusters with at least 6 members identified by the Brotherhood method (these are in addition to the 14 representative families described in the main text):

1. CMRF35-like molecule-1 (CLM1, CLM2, CLM4, CLM6, CLM7, CLM8, CLM9), Triggering receptor expressed on myeloid cells – 2 (TREM2), TREM like 2 (TRML2), TRML3, TRML4, Fas apoptotic inhibitory molecule-3 (fam3), High affinity immunoglobulin alpha and immunoglobulin mu Fc receptor (FCAMR), Polymeric immunoglobulin receptor (PIGR), Natural cytotoxicity triggering receptor -2 (NCTR2).

2. Leucine-rich repeat and immunoglobulin-like domain-containing nogo receptor-interacting protein- 1 (LIGO1), LIGO2, LIGO3, LIGO4, Leucine-rich repeat-containing protein 4B (LRC4B), LRC4C, Leucine-rich repeat and fibronectin type III domain-containing protein 1 (LRFN1), LRFN2, LRFN3, LRFN4, LRFN5, LRRC4, Leucine-rich repeat neuronal protein 1 (LRRN1), LRRN2, LRRN3 ISLR2, Immunoglobulin superfamily containing leucine-rich repeat protein 2 (ISLR), LRC24, Leucine-rich repeat, immunoglobulin-like domain and transmembrane domain-containing protein - 1 (LRIT1), LRIT2, LRIT3

3. VCAN (CSPG2), NCAN, ACAN (PGCA), BCAN (PGCB), Hyaluronan and proteoglycan link protein 1 (HPLN1), HPLN2, HPLN3, HPLN4,

4. Interleukin 1 receptor family: I18RA, IL18R, IL1AP, IL1R1, IL1R2, ILRL1, ILRL2, IRPL1, IRPL2. (Single Ig IL-1-related receptor (SIGIR) is singleton in the Brotherhood)

5. Sodium channel subunit SCN1B, SCN2B, SCN3B, SCN4B, Myelin protein zero MYP0, MPZL1, MPZL2, MPZL3, JAML1

6. Contactins (Bairoch et al., 2005): CNTN1, CNTN2, CNTN3, CNTN4, CNTN5, CNTN6

7. BOC and DCC family (Bairoch et al., 2005): DCC, IGDC3, IGDC4, NEO1, PRTG. 8. Roundabout ROBO1, ROBO2, ROBO3, ROBO4, Inactive tyrosine-protein kinase

7 PTK7, Muscle, skeletal receptor tyrosine-protein kinase (MUSK) 9. SHPS1, Signal-regulatory protein beta-1 (SIRB1), SIRB2, SIRBL, SIRPD,

SIRPG 10. AXL/UFO subfamily (Bairoch et al., 2005): Tyrosine-protein kinase Mer

(MERTK), Tyrosine-protein kinase receptor TYRO3, Tyrosine-protein kinase receptor UFO, protein-tyrosine phosphatase family - Receptor class 2A subfamily (Bairoch et al., 2005): Receptor-type tyrosine-protein phosphatase delta (PTPRD), PTPRF, PTPRS.

Supplementary Figures Figure S1, related to Figure 1: The number of correct assignments (true positive (TP), in blue), missing assignments (false negative (FN), in green), and incorrect assignments (false positive (FP), in red) of 238 proteins are plotted for 14 representative families. Two methods are compared: 1) CD-HIT at four different sequence identity thresholds 40%, 30%, 25% and 20%, and 2) Brotherhood at 45% overlap threshold. The Table inset shows the number of actual clusters (including singletons) for the 14 families generated by different methods.

Figure S2, related to Figure 1: A) Domain organization in the nectin/nectin-like family. Nectins, nectin-like proteins, and CD96 contain two Ig-C2 domains and one Ig-V domain. CRTAM and CD226 contain one Ig-C2 domain and one Ig-V domain. TIGIT contains one Ig-V domain. The red and orange boxes located at the cytoplasmic tails of the proteins represent PDZ binding and ITIM (immunoreceptor tyrosine-based inhibition motif), respectively.

Figure S3, related to Figure 1 and 5: Multiple sequence alignment of the Ig-V domains of the nectin/nectin-like family. Secondary structures from the crystal structures of CRTAM and Necl-5 is shown at the top and the bottom of the alignment, respectively. The conservation analysis was calculated for three subgroups in the family: the nectin-like (necl-1 to necl-4), the nectin (necl-5 and nectin-1 to nectin-4) and nectin-associated proteins (CD96, TIGIT, CD226, and CD200). Invariant positions across the whole alignment are highlighted in red, such as the 2 cysteines forming the Ig canonical disulfide bond. Positions that are invariant in one subgroup but not in the others are highlighted in blue, and conserved positions are highlighted in green. Blue circles mark residues that are at the CRTAM Ig-V homodimer interface. According to this alignment, CRTAM interface residues are conserved in the nectin-like subgroup but not in the nectin or nectin-associated protein subgroups. AMAS was used to format the multiple sequence alignment (Livingstone and Barton, 1993).

Figure S4, related to Figure 1 and 3: The gene structure of CRTAM, nectins, nectin-like proteins and nectin-related proteins. DNA sequence is illustrated as a black line. Exons are depicted by light blue boxes. Each exon is described by two parameters: the length of the exon (shown in the boxes), and the exon phase (the intron interrupting codon position, which can be either 0, 1 or 2, shown above each box). Exons that encode for Ig-V and Ig-C2 domains are highlighted in blue and in red, respectively. A) Gene structure alignment of CRTAM with Necl-1 – Necl-4 proteins. B) Gene structure alignment of nectin-1 – nectin-4, and Necl-5 proteins. C) Gene structure alignment of the nectin-related proteins CD226, CD200, TIGIT AND CD96.

Figure S5, related to Figure 3: Gel filtration profile of CRTAM. Elution profile demonstrating a peak corresponding to the dimer (~28kD) with a shoulder corresponding to the monomer (~14kD). The blue and the red triangle denote standard markers of ribonuclease-A (13.7kDa) and chymotrypsinogen-A (25kDa), respectively.

Figure S6, related to Figure 3: Sedimentation equilibrium experiments of the Ig-V domain of CRTAM was performed at rotor speeds of 20000 or 25000 RPM (filled and empty circle, respectively) and at three different protein concentrations 38 M, 21 M and 5 M (A, B, and C, respectively). Fitting the equilibrium sedimentation data to the monomer-dimer model yields a dissociation constant value of 10 M. The fitting residual is shown in D.

Figure S7, related to Figure 5: Multiple sequence alignment of the Ig-V domain of CRTAM from different species. Invariant and conserved sequence positions are highlighted in red and in green, respectively. Residues forming the homophilic trans-interaction protein interface are marked with blue circles above the alignment. The black vertical arrow marks a potential N-glycosylation site at sequence position N45 in the human ortholog. AMAS was used to format the multiple sequence alignment (Livingstone and Barton, 1993).

Supplementary Tables Monomers Dimers RMSD (Å) C RMSD (Å) C

Conserved interface –residues

Nec-l1:Necl-1 1.6 95 1.9 192 11 Nectin1:Nectin1 2.3 92 4.2 186 1 CTLA-4:B7.1 2.7 90 5.35 92 0 CD84:CD84 2.3 91 2.86 177 0 Table S1, related to Figure 5: Results of structural superimposition of CRTAM monomer and dimer with other IgSF proteins. Compared parameters are RMSD, the number of structurally equivalent residues (C) and the number of conserved interface residues with respect to CRTAM amino acid sequence. In the column of Monomers the CRTAM monomer was superimposed with the molecule highlighted in bold and in the column of Dimers the CRTAM:CRTAM dimer was aligned to the complexes. Bairoch, A., Apweiler, R., Wu, C.H., Barker, W.C., Boeckmann, B., Ferro, S., Gasteiger, E., Huang, H., Lopez, R., Magrane, M., et al.  (2005). The Universal Protein Resource (UniProt). Nucleic acids research 33, D154‐159. Livingstone,  C.D.,  and Barton, G.J.  (1993).  Protein  sequence  alignments:  a  strategy for  the hierarchical analysis of  residue conservation. Computer applications  in  the biosciences : CABIOS 9, 745‐756.