*3.5. Determination of the IC<sup>50</sup> Value by AlphaLISA*®

#### 3.5.1. Principle of the Competitive Binding Assay

The binding affinity of the inhibitors to PD-L1 were measured using the AlphaLISA® assay kit (AL356 HV/C/F, PerkinElmer) according to the manufacturer's instructions, with the anti-PD-1 mAb nivolumab (Selleck Chemicals, Houston, TX, USA) included as a positive control [41]. In this assay, direct binding of an inhibitor to PD-L1 is detected by photoinduced energy transfer (Figure S6). Biotin-conjugated PD-1 is attached to streptavidin-coated donor beads and histidine (His)-tagged PD-L1 is attached to anti-His-conjugated acceptor beads. Photoexcitation of the donor beads at 680 nm yields singlet oxygen. If PD-L1–PD-1 binding is successful, energy is transferred through singlet oxygen, leading to an increase in fluorescence intensity at 615 nm (Figure S6).

#### 3.5.2. Preparation of Samples

**BMS-8** was purchased from AA Blocks LLC (San Diego, CA, USA). Stock solutions of inhibitors in DMSO (stock solution A, 5 mM) were serially diluted (Figure S5A) to obtain 10 assay solutions (1–10) with concentrations ranging from 5.0 mM to 2.6 nM (Table S2). An aliquot of solution 1–10 (2 µL) was mixed with His-tagged PD-L1 (25 nM, 2 µL), biotin-conjugated PD-1 (25 nM, 2 µL), anti-His acceptor beads (0.55 g L–1, 2 µL), and streptavidin-coated donor beads (1.1 g L−<sup>1</sup> , 2 µL) (Figure S5B) in a final volume of 10 µL and incubated at 25 ◦C for 90 min. Positive and negative technical controls were included in parallel. Positive controls contained buffer (2 µL) in place of solution 1–10, and negative controls contained only the beads (2 µL each) and buffer (6 µL).

#### 3.5.3. AlphaLISA® Measurement and Analysis

The reaction samples (10 µM) were placed in a 384-well microplate and photoirradiated at 680 nm from the top. Fluorescence at 615 nm was detected using an EnSpire multimode plate reader (Perkin Elmer, Waltham, MA, USA). IC<sup>50</sup> values were estimated from a sigmoidal curve of fluorescence intensity vs. inhibitor concentration using a relative weighting method (1/Y <sup>2</sup> weighting) with GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA, USA).

#### *3.6. Docking Simulation of Compounds*

The docking simulation software ICM 3.8-7 [33] was used to investigate the binding modes of **X** and **amino-X**s to the PD-L1 homodimer complexed with **BMS-8** (PDB ID: 5J8O) [31]. We performed docking without template docking [37] or introducing flexibility [37] to avoid over-fitting of the ligands into the pocket. The docking simulation supposed Monte Carlo pseudo-Brownian motion [46]. In the simulation, the score suggests goodness of docking, defined as follows [45]:

$$\text{Score} = \Delta E\_{\text{Int}\text{FF}} + T\Delta S\_{\text{Tor}} + \alpha\_1 \Delta E\_{\text{HBond}} + \alpha\_2 \Delta E\_{\text{HBDesol}} + \alpha\_3 \Delta E\_{\text{soil}} + \alpha\_4 \Delta E\_{\text{HFob}} + \alpha\_5 Q\_{\text{size}} \tag{1}$$

where α1–α<sup>5</sup> = weight, ∆*E*IntFF = ligand–target van der Waals interactions and internal force field energy of the ligand, *T*∆*S*Tor = free energy changes due to conformational energy loss upon ligand binding, ∆*E*HBond = hydrogen bonding interactions, ∆*E*HBDesol = hydrogen bond donor–acceptor desolvation energy, ∆*E*solEl = solvation electrostatic energy upon ligand binding, ∆*E*HPob = hydrophobic free energy gain, and *Q*size = a size correction term proportional to the number of ligand atoms [45,47,48]. We calculated RMSD values by using CORREL function in the Microsoft Excel.

#### **4. Conclusions**

This study reports that we prepared the new biphenyl-conjugated bromotyrosine, which inhibits the PD-1/PD-L1 interaction with better effect than that of **BMS-8**. In addition, the **amino-X**s, which are conjugates of **X** with a variety of amino acids, provide the molecular mechanism how amino acid modifications of **X** affects inhibition of PD-1/PD-L1 interactions. Binding of the **X** without the Cα, NH2, and COOH atoms portion of **amino-X**s into the PD-L1 binding pocket is required to promote transient homodimerization of PD-L1A/PD-L1B, leading to formation of a stable ternary complex composed of **X** and PD-L1AB. Amino acid conjugation, however, alters the **X** docking conformation in the PD-L1 pocket, reducing the IC<sup>50</sup> values dramatically. We conclude that improper interactions between amino acids conjugated to **X** and those in the binding pocket induced displacement of the compounds, thereby reducing inhibitory effect. In the future, we plan to design conjugates with amino acids that do not disturb the conformation of **X** in the PD-L1 binding pocket.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/1422-0067/21/10/3639/s1.

**Author Contributions:** H.M., M.K., Y.I., and E.K. conceived the project. E.-H.K. and M.K. conducted the organic syntheses. H.M. and E.-H.K. performed the in silico simulations. E.-H.K., R.D., and H.M. measured the IC<sup>50</sup> values by the AlphaLISA®. E.-H.K., M.K., and H.M. wrote the manuscript. All of the authors discussed the results and approved the content of the final manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was partly supported by the Incentive Research Fund of RIKEN (FY2017-2018, H.M. and M.K.) and the Japan Agency for Medical Research and Development (AMED: FY2018-2020, Y.I.). E.-H.K. was financially supported by the Junior Research Associate Program at RIKEN.

**Acknowledgments:** We thank the Bio-material Analysis Support Unit, RIKEN Center for Brain Science for synthesis of biphenyl-conjugated amino acids. We thank Anne M. O'Rourke, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.

**Conflicts of Interest:** The authors declare no conflicts of interest.
