Resume Jurnal Fixed.pdf

RESUME JURNAL “Efficient Ligand Discovery From Natural Herbs By Integrating Virtual Screen, Affinity Mass Spectrometry And Targeted Metabolomics”

Kelompok 4 (Drug From Natural Source) Anggota : 1. Intan Ayu Permata F.

(162210101136)

2. Siti Rokhayah

(172210101042)

3. Devina Aulia Zulfa

(172210101045)

4. Rindi Valent Sebatines

(172210101054)

5. Adelia Nadyana Arief P.

(172210101153)

6. Erna Putri Iliyin

(172210101160)

Dosen Pengampuh : Dwi Koko Pratoko, M.Sc.,Apt.

FAKULTAS FARMASI UNIVERSITAS JEMBER 2019

Pendahuluan Jamu telah menjadi sumber senyawa penemuan obat baru. Ada penelitian lanjut tentang pengobatan dari herbal (TCM). Banyak penemuan baru untuk memudahkan hal itu contohnya layar virtual berbaasis struktur untuk identifikasi senyawa bioaktifnya. Namun dalam menggunakan layar virtual penemuan ligan memiliki hit rate yg terbatas sehingga menghambat skrining bioaktif dari obat. Teknik percobaan yg baru dikembangkan afinitas spektrometri massa (AMS) memungkinkan mendeteksi ligan dari campuran senyawa, dikembangkan penemuan kombinasi dari kekuatan layar virtual, afinitas MS , dan metabolisme dengan target untuk menyaring ligan dari tanaman. Penerapan dilakukian untuk skrining ligan terhadap nucleoprotein(NP) virus Ebola(EBOV) sehingga teridentifikasi ligan yang menargetkan kantong hidrofobik EBOV NP untuk mengkarakterisasi bioaktivitasnya. Saku hidrofobik penting untuk eplikasi EBOV dan ditargetkan untuk pengembangan terapi. Hanya satu ligan dari protein virus yang terdeteksi berinteraksi dengan kantong hidrofobik EBOV, sehingga penelitian ini bertujuan mengidentifikasi ligan baru yang menargetkan kantong hidrofobik EBOV agar pengembangan terapi semaakin luas. 1) Bahan dan Reagen Tris, NaCl, gliserol, DTT, asam format. Ammonium, pepton, ragi bubuk, IPTG, ampisilin, imidazole, unit ultrafiltrasi sentrifugal (30 KD), ptoein thermal shift dye kitTM, 4 jamu, standar untuk senyawa MD-1, MD-2, MD-4, GC-7, pelarut organic berstandar HPLC, basis data ramuan TCM, 272 ramuan. 2) Ekspresi dan pemurnian protein Gen EBOV NP di klon dengan 6x penyatuan tag pada terminal C, dan diekspresikan dalam E. Coli rantai BL21. Pemurnian afinitas dilakukan dengan kolom Ni-NTA dan protein dielusi dengan lisis buffer 0.5-1 M imidazole. Protein dimurnikan pada kolom Superdex200. Protein NP yang dimurnikan adalah >95% murni menurut analisis SDS-Page 3) Preparasi ekstrak ramuan TCM 4 jamu TCM (Piper kadsura, Piper nigrum, Ophiopogon japonicas dan Salvia miltiorrhiza) dihaluskan menjadi serbuk lalu diekstraksi dengan etanol 70%, dan diekstraksi 2x melalui pemanasan refluks. Setelah penyaringan supernatan, pelarut organik dibuang dengan

penguapan vakum dan bahan yang tersisa dikeringkan dengan liofilisasi. Serbuk yang dihasilkan disimpan pada -20°C. 4) Skrinning ligan ekstrak herbal dengan afinitas MS Larutan stok (100 mg/mL) dari setiap ekstrak TCM dibuat dengan melarutkan bubuk dengan DMSO 95%. Stok diencerkan dengan buffer inkubasi EBOV NP (20 mM Tris-HCl, 200 mM NaCl, 5 mM DTT, pH 8,5) hingga konsentrasi 2,5 mg/mL. Lalu EBOV NP dicampur dengan larutan ekstrak encer pada 1:1 (v/v) ke volume akhir 50 μL. Inkubasi disimpan pada suhu 16°C selama 40 menit. disaring dengan ultrafiltrasi dengan sentrifugasi. Larutan yang tertahan pada membran dicuci 2x dengan amonium asetat 0.15 M. Komponen yang terikat pada protein target dipisahkan dengan 90% methanol. Setelah sentrifugasi, senyawa dalam supernatan diperoleh dan diuapkan dengan vakum, dilarutkan dalam 20% metanol, dan dianalisis dengan LC-MS/MS. Protein bebas kontrol disiapkan menggunakan pengganti buffer untuk EBOV NP selama inkubasi. Semua sampel disiapkan dalam 4 replikasi dan dianalisis secara terpisah. Sampel dianalisis pada sistem Shimazu L30A UPLC digabungkan ke spektrometer massa TripleTOF 6600 yang beroperasi dalam mode ion positif. 5) Pemrosesan data metabolic untuk ligan NP Sampel ekstrak TCM diidentifikasi dengan mengekstraksi ion kromatogram (XIC) menggunakan Peakview 2.2 (AB SCIEX) berdasarkan massa akurat pengukuran (<10 ppm deviasi) dan pencocokan amplop isotop (<10% perbedaan dari amplop teoretis) sesuai dengan rumus majemuk di TCMHD. H+ dan Na+ digunakan untuk deteksi senyawa. Senyawa dalam target dan sampel kontrol diidentifikasi dengan memenuhi di atas kriteria plus waktu retensi (RT) yang cocok dengan puncak yang sesuai dalam TCM ekstrak (<0,2 mnt shift). Senyawa diidentifikasi dalam ekstrak dan sampel target juga memiliki intensitas XIC>103. Untuk setiap senyawa yang diidentifikasi dalam target dan sampel kontrol, indeks pengikatannya (BI) dihitung dengan membagi intensitas XIC dari senyawa ditarget oleh yang dikontrol. Hasil skrinning dipilih berdasarkan rata-rata BI>2 dan p <0,05 dari repilkasi 4x. Nilai-p dihungan dengan MSstat, menggambarkan perbedaan signifikan dalam intensitas XIC antara kelompok target dan control. 6) Isolasi senyawa dari jamu TCM

Seri senyawa HJ, DS dan HFT pada Tabel 1 diisolasi dari ramuan yang sesuai. Protokol isolasi terperinci disediakan dalam Informasi Pendukung. Spektrum NMR untuk semua senyawa yang terisolasi 7) Analisis afinitas MS dari ikatan ligan dengan senyawa murni 14 standar senyawa dicampur pada konsentrasi akhir masing-masing 2 μM, dan diinkubasi dengan EBOV NP. Dalam uji pengikatan senyawa individu, setiap standar pada 10 μM secara terpisah diinkubasi dengan protein pada 10 μM dalam buffer NP. Inkubasi dilakukan pada 16°C selama 40 menit. Setelah melalui ultrafiltrasi dan disosiasi ligan, dianalisis campuran senyawa yang dihasilkan/ligan spesifik dengan LC-MS/MS menggunakan instrumen pada penyaringan ligan TCM. XIC dari senyawa spesifik dalam sampel target dan kontrol diekstraksi menggunakan PeakView 2.2. 8) Uji Ligan dengan Peptida VP35 Protein EBOV NP 10 μM diinkubasi dengan setiap tes ligan sebanyak 5 μM dan meningkatkan jumlah VP35 peptida 20-48 (0, 5, 10, 20, 40 μM) pada 16°C selama 40 menit. Kemudian analisis afinitas MS dilakukan untuk mengukur fraksi pengikat setiap tes ligan untuk protein seperti yang dijelaskan sebelumnya. Fraksi pengikat yang mengacu pada persentase ligan terikat protein lebih jumlah total di inkubasi diplot terhadap konsentrasi VP35 peptida. Setiap pasang target dan kontrol sampel disiapkan dan dianalisis dalam rangkap tiga. 9) Layar Virtual dan Docking Molekul Konsensus docking di layar virtual dilakukan dengan tiga alat docking. Ktub hidrogen dan biaya parsial Gasteiger ditambahkan ke target dan senyawa TCMHD oleh AutoDockTools. Konformasi protein dan ligan kaku dan fleksibel. Untuk langkah docking, baik presisi standar (SP) dan presisi ekstra (XP) dilakukan. Metode pengambilan sampel ligan fleksibel tanpa tambahan dan minimalisasi pasca-docking dilakukan. Akhirnya, GlideScore digunakan sebagai mencetak fungsi. Glide XP docking digunakan untuk menampilkan interaksi antara HJ-4 dan EBOV NP. 10) Analisis SEC Protein EBOV NP dimurnikan dengan buffer rendah garam 20 mM Tris (PH 8,5), 200 mM NaCl dan 5 mM DTT, atau buffer garam dari 20 mM Tris (pH 8,5), 1 M NaCl dan 5 mM DTT pada kolom Superdex-200. Setiap ligan uji dicampur dengan protein NP di garam dan

sampel diinkubasi 24 jam pada 16°C. Percobaan SEC dilakukan dengan menggunakan sistem AKTA FPLC. Sampel disuntikkan sebanyak 2ml. 11) Protein Assay Pergeseran Thermal Uji pergeseran termal dilakukan dengan Protein Thermal Pergeseran Dye Kit TM. Kedua protein EBOV NP dan senyawa uji diencerkan dalam inkubasi buffer yang sesuai. NP protein pada 10 μM kemudian dicampur dengan masing-masing senyawa pada 100 μM dan pewarna (1X) pada 16 ° C dalam volume total 20 μL. Semua reaksi disiapkan di plate 96 yang kemudian disegel, diguncang, dan disentrifugasi pada 1000 rpm selama 1 menit. Panas pemindaian dilakukan pada mesin PCR real-time dan intensitas fluoresensi terus dipantau. Kurva fitting, perhitungan titik leleh protein (Tm) dan laporan data dibuat menggunakan LightCycler® 480 software. 3. HASIL 3.1. Prosedur kerja ramuan berbasis layar virtual terhadap pocket hidrofobik dari EBOV NP Proses dimulai dengan layar virtual ramuan senyawa di TCM Database herbal (TCMHD) terhadap protein yang ditentukan. Herbal dipilih kaya akan senyawa aktif. Herbal yang terpilih diuji afinitas layar MS menggunakan ekstrak kasar mereka. Dengan menggabungkan metabolomik , mampu membedakan herbal aktif dari pandangan holistik dan kemudian mengidentifikasi semua konstituen yang diketahui dalam herbal aktif untuk menemukan ligan diduga untuk target. Terapi khasiat herbal TCM dikaitkan dengan baik kelimpahan dan pengayaan senyawa aktif terpilih empat herbal ( Piper kadsura, piper nigrum,japonicas Ophiopogon dan Salvia miltiorrhiza) mengandung setidaknya enam ligan diduga untuk percobaan berikutnya. 3.2. Afinitas MS dikombinasikan dengan metabolomik ditargetkan untuk layar ekstrak herbal dengan dimurnikan EBOV NP analisis afinitas MS dikombinasikan dengan metabolomik untuk menyaring herbal dan mengevaluasi kapasitas pengikatan ligan dan konstituen di setiap ramuan. Tujuannya untuk, domain inti rekombinan dari EBVO NP (Residu 36-351) dimurnikan dan diinkubasi secara terpisah dengan ekstrak herbal TCM. Ligan terikat NP kompleks kemudian diisolasi dengan ultrafiltrasi, dan senyawa dipisahkan dari target NP atau protein bebas sampel kontrol dianalisis dengan LC-MS / MS menggunakan spektrometri massa resolusi tinggi. Kombinasi dari afinitas

MS dan metabolomik memungkinkan mendapat hasil analisis konstituen keseluruhan herbal yang akurat. 3.3. Perbandingan hits awal dari layar virtual dan afinitas layar MS Di antara 29 peringkat teratas hits menunjukkan sampel mengikat saling diverifikasi oleh dua pendekatan screening. Di antara 15 hits unik untuk layar virtual, 13 senyawa tidak diverifikasi oleh layar MS afinitas karena nilai BI rendah di bawah ambang. Dengan demikian mereka cenderung hit positif palsu dari layar virtual. di sisi lain, 12 senyawa dari tiga herbal yang menghindari layar virtual yang ditemukan oleh afinitas layar MS. Pemilihan hit dengan kriteria yang sangat ketat untuk memaksimalkan hit rate layar virtual. 3.4. Isolasi senyawa dan ligan mengikat validasi Di antara 14 ligan, kami memperoleh 7 standar murni dari pemurnian dari tumbuhan yang sesuai .Analis Semua tujuh senyawa uji co-diidentifikasi oleh dua pendekatan screening yang bermakna dikaitkan dengan target NP dan dianggap ligan divalidasi. Untuk valigasi ligan, kami memperkirakan afinitas pengikatan masing-masing senyawa murni menggunakan metode yang dikembangkan sebelumnya afinitas MS berbasis kesetimbangan menggunakan metode yang dikembangkan sebelumnya afinitas MS berbasis konstanta disosiasi (K d) dihitung untuk ligan yang berbeda terutama dalam medium berbagai mikro-molar, yang biasanya dilaporkan untuk produk turunan ligan (Tabel 1). 3.5. Penilaian bioaktivitas dan mekanisme molekuler dari ligan kimia baru EBOV NP protein merupakan bagian integral dari nukleokapsid virus yang berkaitan erat dengan template virus RNA, oligomer dari NP protein berperan penting dalam menjaga integritas Template dan menyediakan template akses ke RNA polimerase virus. Untuk menilai bioaktivitas ligan NP baru, akses ke RNA polimerase virus menggunakan protein NP (residu 36-450). Struktur kristal EBOV NP (PDB4Z9P) mengindikasikan lipatan Helix-20 NP menuju dan berinteraksi dengan pocket hidrofobik dalam C-lobus untuk membentuk struktur kompak. Ligan kimia baru hanya dapat mengikat struktur kompak. Sebagai peptida VP35 dilipat dalam bentuk V dengan ujung dua heliks orthogonal terbuka luas, memungkinkan akses mudah dari HJ-4 ke pocket hidrofobik seperti yang ditunjukkan oleh docking molekuler. Karena afinitas tinggi dari peptida VP35, interaksinya dengan NP dapat mendorong protein untuk membuka pocket hidrofobik menjadi lebih besar, sehingga meningkatkan persentase ikatan ligan kimia ke bagian yang sama.

View Article Online View Journal

Analyst Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: Z. Wang, H. Liang, H. Cao, B. Zhang, J. Li, W. Wang, S. Qin, Y. Wang, L. Xuan, L. Lai and W. Shui, Analyst, 2019, DOI: 10.1039/C8AN02482K. Volume 141 Number 1 7 January 2016 Pages 1–354

Analyst www.rsc.org/analyst

This is an Accepted Manuscript, which has been through the Royal Society of Chemistry peer review process and has been accepted for publication. Accepted Manuscripts are published online shortly after acceptance, before technical editing, formatting and proof reading. Using this free service, authors can make their results available to the community, in citable form, before we publish the edited article. We will replace this Accepted Manuscript with the edited and formatted Advance Article as soon as it is available. You can find more information about Accepted Manuscripts in the author guidelines.

ISSN 0003-2654

PAPER Michele Zagnoni et al. Emulsion technologies for multicellular tumour spheroid radiation assays

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the ethical guidelines, outlined in our author and reviewer resource centre, still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

rsc.li/analyst

Page 1 of 35

View Article Online

10.1039/C8AN02482K Efficient ligand discovery from natural herbs by integrating virtual DOI: screen,

affinity mass spectrometry and targeted metabolomics Zhihua Wang[a]#, Hao Liang[b]#, Haijie Cao[a], [c]#, Bingjie Zhang[c], Jun Li[d], Wenqiong Wang[d], Shanshan Qin[c], Yuefei Wang[e], Lijiang Xuan[d]*, Luhua Lai[b], [f], [g]* & Wenqing Shui[c]*

[a]

College of Pharmacy, Nankai University, Tianjin 300071, China; High-throughput

Molecular Drug Discovery Center, Tianjin Joint Academy of Biotechnology and Medicine, Tianjin 300457, China [b]

Peking-Tsinghua Center for Life Sciences, Academy for Advanced Interdisciplinary

Studies, Peking University, Beijing 100871, China [c]

iHuman Institute, ShanghaiTech University, Shanghai 201210, China

[d]

State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica,

Chinese Academy of Sciences, 501 Haike Road, Zhangjiang Hi-Tech Park, Shanghai 201203, P. R. China [e]

Tianjin State Key Laboratory of Modern Chinese Medicine, Tianjin University of

Traditional Chinese Medicine, Tianjin 300193, China [f]

Center for Quantitative Biology, Academy for Advanced Interdisciplinary Studies,

Peking University, Beijing 100871, China. [g]

Beijing National Laboratory for Molecular Sciences, State Key Laboratory for

Structural Chemistry of Unstable and Stable Species, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China.

1

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

#These authors contribute equally to this work

DOI: 10.1039/C8AN02482K

*To whom correspondence should be addressed to: Lijiang Xuan, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 501 Haike Road, Zhangjiang Hi-Tech Park, Shanghai 201203, China; E-mail: [email protected]; Tel: 86-21-20231968 Luhua Lai, Peking-Tsinghua Center for Life Sciences, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing 100871, China; E-mail: [email protected]; Tel: 86-10-62757486 Wenqing Shui, iHuman Institute, ShanghaiTech University, Shanghai 201210, China; E-mail: [email protected]; Tel: 86-21-20685595

2

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 35

Page 3 of 35

View Article Online

DOI: 10.1039/C8AN02482K

Abstract

Although natural herbs have been a rich source of compounds for drug discovery, identification of bioactive components from natural herbs suffers from low efficiency and prohibitive cost of the conventional bioassay-based screening platforms. Here we developed a new strategy that integrates virtual screen, affinity mass spectrometry (MS) and targeted metabolomics for efficient discovery of herb-derived ligands towards a specific protein target site. Herb-based virtual screen conveniently selects herbs of potential bioactivity whereas affinity MS combined with targeted metabolomics readily screens candidate compounds in a high-throughput manner. This new integrated approach was benchmarked on screening chemical ligands that target the hydrophobic pocket of nucleoprotein (NP) of Ebola viruses for which no small molecule ligands have been reported. Seven compounds identified by this approach from the crude extracts of three natural herbs were all validated to bind to the NP target in pure ligand binding assays. Among them, three compounds isolated from Piper nigrum (HJ-1, HJ-4 and HJ-6) strongly promoted formation of large NP oligomers and reduced protein thermal stability. In addition, cooperative binding between these chemical ligands and an endogenous peptide ligand was observed, and molecular docking was employed to propose a possible mechanism. Taken together, we established a platform integrating in silico and experimental screening approaches for efficient discovery of herb-derived bioactive ligands especially towards non-enzyme protein targets.

3

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

1. INTRODUCTION

DOI: 10.1039/C8AN02482K

Natural herbs have been a rich source of compounds for drug discovery especially as anticancer and antimicrobial agents1, 2. There is a sustaining interest in herbal medicine, especially Traditional Chinese Medicine (TCM), for generation of lead compounds, as they possess more diverse chemical scaffolds and more drug-like properties than synthetic compounds3. However, the traditional method of screening drug leads from natural herbs is labor-intensive and cost-prohibitive. It requires multi-step isolation of individual compounds from crude plant extracts, structural characterization and bioactivity test of each compound. Various computational approaches have emerged as valuable tools in assisting the holistic understanding of molecular mechanism4-6 and the cost-efficient discovery of bioactive natural compounds or herbs7, 8. Particularly, structure-based virtual screen is now widely employed for identification of bioactive compounds from natural products9, 10 as well as chemical libraries11. However, most studies applying virtual screen alone for ligand discovery have limited hit rate (5-30%) and cannot rank order screening hits due to the difficulties in treating protein flexibility and accurate prediction of protein-ligand binding free energy with current docking algorithms12, 13. Furthermore, a number of individual hit compounds yielded by virtual screen need to be synthesized or isolated in the initial stage for functional tests, which presents a key bottleneck for bioactive screening from herbal medicine. Among the newly developed experimental screening techniques, affinity mass spectrometry (MS) has demonstrated distinct advantages for ligand discovery from crude extracts of natural products towards specific protein targets. Affinity MS allows 4

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 35

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

View Article Online

DOI: 10.1039/C8AN02482K for direct capture and detection of ligands from compound mixtures, thus eliminating

the need of pure compound preparation in the initial screening phase14-16. In order to pick up a few “active” herbs for affinity MS screen, the total or fractionated extracts of different herbs need to be first assayed for their inhibitory activity against a specific enzyme target17. However, to address novel targets related to newly emerged diseases that have no bio-activity assays available, the task of testing a large number of herbal extracts becomes formidable. In addition, bio-assays of complex compound mixtures can be confounded by interference or nuisance compounds2. Thus, new strategies are required to select candidate herbs for the affinity MS screen especially towards challenging non-enzyme targets. Recently affinity MS has been combined with metabolomics techniques in several studies for mapping protein-metabolite interactions in animal cells and tissues18-20. We have established a workflow to combine affinity MS with untargeted metabolomics for identification of ligands interacting with a protein target from metabolite extracts of Traditional Chinese Medicine (TCM)16. Compared to most other studies that evaluate the binding capacity of only dozens of most abundant compounds in natural herbs by HPLC analysis21-24, our metabolomics-incorporated workflow considerably increased the throughput and sensitivity by simultaneously screening hundreds to thousands of components associated with a TCM herb. However, it is noteworthy that in the earlier study we had to randomly analyze a number of TCM extracts (>20 types) before we could find the right species that contained active ligands for our protein target.

5

Analyst Accepted Manuscript

Page 5 of 35

Analyst

View Article Online

DOI: 10.1039/C8AN02482K In this study, we developed a comprehensive strategy that combines the

strengths of virtual screen, affinity MS and targeted metabolomics for screening chemical ligands from natural herbs. As a proof of concept, we applied our new approach to ligand screening towards the nucleoprotein (NP) of Ebola virus (EBOV). EBOV causes severe hemorrhagic fever with extremely high fatality rates and has posed significant threats to global human health25, 26. EBOV NP possesses a unique hydrophobic pocket at the surface of its C-lobe27. This hydrophobic pocket is proposed to provide a viral interface that is essential for EBOV replication and can be targeted for therapeutic development28. Only one endogenous peptide ligand derived from the viral protein VP35 has been reported to interact with the hydrophobic pocket28, 29 and no chemical ligands has been reported. Our study aims to identify new chemical ligands targeting the hydrophobic pocket of EBOV NP using this integrated approach and characterize the bioactivity profiles of verified ligands.

6

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 35

Page 7 of 35

View Article Online

2. MATERIALS AND METHODS

DOI: 10.1039/C8AN02482K

2.1. Materials and reagents Tris, NaCl, glycerol, DTT, formic acid and ammonium acetate were all purchased from Sigma Aldrich (St. Louis, MO, USA). Peptone, yeast powder, IPTG, ampicillin and imidazole were purchased from Sangon Biotech Shanghai Co., Ltd (Shanghai, China). Centrifugal ultrafiltration units (30 KD) were purchased from Sartorius (Germany). Protein Thermal Shift Dye KitTM was purchased from Thermo Fisher Scientific (Massachusetts, USA). Four TCM herbs were purchased from Tongrentang Chinese Medicine Co., Ltd (Beijing, China). Standards for compounds MD-1, MD-2, MD-4 and GC-7 listed in Table 1 were purchased from Chengdu Herbpurify Co., Ltd (Chengdu, China). All organic solvents are HPLC grade or better and were purchased from Merck (Darmstadt, Germany). TCM herb database (TCMHD) was constructed based on Chinese Pharmacopoeia (2010 Edition) and 3D structures from Traditional Chinese Medicine Database (TCMD, NeoTrident Co., Ltd.)30. A total of 272 well-studied and widely used herbs and 4851 natural compounds from them were chosen with multiple criteria to construct TCMHD, as described in our previous study31.

2.2. Protein expression and purification The detailed methods for cloning and purification of EBOV NP were described previously16, 27. In brief, the gene of EBOV NP (residues 36–351 or 36–450) was cloned with a 6× His tag fused at the C terminus, and expressed in E. coli strain BL21 (DE3) (BioMed, China). The affinity purification was performed using a Ni-NTA 7

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

10.1039/C8AN02482K column (GE Healthcare) and proteins were eluted with lysis buffer containing DOI: 0.5–1 M

imidazole. Proteins were then purified on a Superdex-200 column (GE Healthcare). The purified NP protein was > 95% pure according to SDS-PAGE analysis.

2.3. Preparation of TCM herbal extracts All TCM extraction was carried out according to Chinese Pharmacopoeia extraction rules. Briefly, four TCM herbs (Piper kadsura, Piper nigrum, Ophiopogon japonicas and Salvia miltiorrhiza) were first pulverized into powder. The powder (50 g) was first extracted with 200 mL of 70% ethanol, and then extracted twice through heating reflux or waterbath ultrasonication. After filtration of the supernatant, the organic solvent was removed by vacuum evaporation at 55 °C, and the remaining material was dried out by lyophilization. The resulting powder was stored at -20°C.

2.4. Ligand screening of herbal extracts by affinity MS A stock solution (100 mg/mL) of each TCM extract was prepared by dissolving the powder with 95% DMSO. The stock was diluted with the EBOV NP incubation buffer (20 mM Tris-HCl，200 mM NaCl, 5 mM DTT, pH 8.5) to a concentration of 2.5 mg/mL. Then EBOV NP at 40 μM was mixed with the diluted extract solution at 1:1 (v/v) to a final volume of 50 μL. Incubation was kept at 16 °C for 40 min. The protein incubation solution was then filtered through an ultrafiltration unit (30 KD) by centrifugation at 13000g for 10 min at 4 °C. The retained solution on the membrane was washed twice with 150 mM ammonium acetate. Compounds bound to the protein target were 8

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 35

Page 9 of 35

View Article Online

DOI: 10.1039/C8AN02482K dissociated with 90% methanol. After centrifugation, compounds in the supernatant

were obtained and evaporated by speed vacuum, reconstituted in 20% methanol, and analyzed by LC-MS/MS. The protein-free control was prepared using the buffer substitute for EBOV NP during incubation. All samples were prepared in four replicates and analyzed separately. Samples were analyzed on a Shimazu L30A UPLC system (Shimazu) coupled to a TripleTOF 6600 mass spectrometer (AB SCIEX) operating in the positive ion mode. Chromatographic separation was performed on a ZORBAX Eclipse Plus C18 column (3.5 µm, 2.1×100mm, Agilent) at a flow rate of 200 uL/min and maintained at 40°C with the mobile phases of water/0.1% formic acid (A) and acetonitrile/0.1% formic acid (B). The LC method was as follows: 0−2 min, B at 2%; 2−6 min, B at 2%−35%; 6−10 min, B at 35%-45%; 10−21 min, B at 45%-50%; 21−23 min, B at 50%-75%; 23−25 min, B at 75%-90%; 25−30 min, B at 90%-90%, then re-equilibrate for 5 min. Full-scan mass spectra were acquired in the range of 100-1000 m/z with major ESI source settings: voltage 5.0 kv; gas temperature 500°C; curtain gas 35 psi; nebulizer gas 55 psi; and heater gas 55 psi. MSMS spectra were acquired on top 10 compound precursors with collision energy set at 45 eV with a CE spread of 20 eV and other ion source conditions identical to MS full scans.

2.5 Metabolomics data processing for NP ligand identification Compounds in the TCM extract samples were first identified by extracting selected ion chromatograms (XICs) using Peakview 2.2 (AB SCIEX) based on accurate mass 9

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K measurement (<10 ppm deviation) and isotope envelop matching (<10% difference

from the theoretical envelop) in accordance with the compound formula registered in TCMHD. H+ and Na+ adducts were considered for compound detection. Then compounds in the target and control samples were identified by meeting the above criteria plus retention time (RT) matching with corresponding peaks in the TCM extract (<0.2 min shift). Compounds identified in the extract and target samples should also have XIC intensity >103. For each compound identified in the target and control samples, its binding index (BI) was calculated by dividing the XIC intensity of the compound detected in the target by that in the control. Screening hits were selected based on a mean BI >2 and p <0.05 from four experimental replicates. P-values calculated by MSstats32 reflect the statistical significance of differences in XIC intensities between the target and the control groups.

2.6. Compound isolation from TCM herbs The compound series of HJ, DS and HFT in Table 1 were isolated from the corresponding herbs. Detailed isolation protocols are provided in Supporting Information. NMR spectra for all isolated compounds are shown in Figure S2.

2.7. Affinity MS analysis of ligand binding with pure compounds Fourteen compound standards were mixed at a final concentration of 2 μM each, and incubated with EBOV NP at 10 μM. In the individual compound binding assay, each standard at 10 μM was separately incubated with the protein at 10 μM in the NP buffer 10

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 35

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

View Article Online

10.1039/C8AN02482K described earlier. Incubation was conducted at 16 °C for 40 min. After goingDOI: through

ultrafiltration and ligand dissociation, we analyzed the resulting compound mixture or a specific ligand by LC-MS/MS using the same instrument as in the TCM ligand screening experiment. A short LC gradient (5% to 80% B, 0-2 min; 80% B, 2-8 min) was adopted for compound elution. Each incubation sample and the corresponding protein-free control was prepared and analyzed in triplicate. XICs of specific compounds in target and control samples were extracted using PeakView 2.2 (AB SCIEX) based on the accurate mass measurement (<10 ppm) and RT matching with the standards (<0.2 min shift). BI values were calculated for each compound to assess binding specificity. In the individual ligand binding assay, we calculated the dissociate constant (Kd) of each ligand bound to NP based on the MS quantification data using the method reported previously15.

2.8. Ligand competition assay with VP35 peptide EBOV NP protein at 10 μM was incubated with each test ligand at 5 μM and an increasing amount of VP35 peptide 20-48 (0, 5, 10, 20, 40 μM) at 16 °C for 40 min. Affinity MS analysis was then performed to measure binding fraction of each test ligand to the protein as described earlier14. Binding fraction which refers to the percentage of the ligand bound to the protein over its total amount in incubation was plotted against the concentration of VP35 peptide. Each pair of target and control samples was prepared and analyzed in triplicate.

11

Analyst Accepted Manuscript

Page 11 of 35

Analyst

View Article Online

2.9. Virtual screen and molecular docking

DOI: 10.1039/C8AN02482K

Consensus docking in virtual screen was conducted by using three widely accepted docking tools, including AutodDock 4.2, AutoDock Vina and Glide. For AutoDock 4.2, we first created a grid box to encompass the entire hydrophobic pocket of EBOV NP. Polar hydrogens and Gasteiger partial charges were added to the target site and TCMHD compounds by AutoDockTools. The conformations of proteins and ligands were set to rigid and flexible, respectively. Compounds in TCMHD were then docked to the hydrophobic pocket by genetic algorithm and ranked by the empirical free energy function. These input files, including protein structure, grid box and Gasteiger partial charges, together with default parameter settings, were used in AutoDock Vina docking process. The Glide docking modules were implemented in Schrödinger version 2015-4. We first used protein preparation wizard to assign bond orders and add hydrogens for ENP. Then, the binding site was defined as a rectangular box centered on the coordinates of residues in the hydrophobic pocket. Meanwhile, all TCMHD compounds were treated by LigPrep module with default parameters. For Glide docking step, both the standard precision (SP) and extra precision (XP) were chosen. The ligand sampling method was flexible with no additional constraints and post-docking minimization was performed. Finally, GlideScore was used as the scoring function. Glide XP docking pose was used to display the interactions between HJ-4 and EBOV NP.

12

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 35

Page 13 of 35

View Article Online

DOI: 10.1039/C8AN02482K

2.10. SEC analysis

EBOV NP protein (residues 36-450) was purified with a low salt buffer of 20 mM Tris (pH 8.5), 200 mM NaCl and 5 mM DTT, or a high salt buffer of 20 mM Tris (pH 8.5), 1 M NaCl and 5 mM DTT on a Superdex-200 column (GE Healthcare). Each test ligand was mixed with the NP protein at the low salt or high salt condition in 2-fold molar excess, and the samples were incubated overnight at 16 °C. SEC experiments were performed using an AKTA FPLC system in line with a Superdex-200 column (GE Healthcare) in the corresponding low salt or high salt buffer. 2 mg/ml sample was injected each time.

2.11. Protein thermal shift assay The thermal shift assay was performed with Protein Thermal Shift Dye KitTM according to the user protocol (Thermo Fisher, USA). Both the EBOV NP protein (residues 36-450) and the test compound were diluted in the corresponding incubation buffer. The NP protein at 10 μM was then mixed with each compound at 100 μM and the dye (1X) at 16 °C in a total volume of 20 μL. All reactions were prepared in a 96-well plate which was then sealed, shaken, and centrifuged at 1000 rpm for 1 min. Thermal scanning (16 to 95 °C at 0.03 °C/s) was conducted on a real-time PCR machine (LightCycler®480, Roche) and fluorescence intensity was continuously monitored. Curve fitting, calculation of the protein melting temperature (Tm) and generation of data reports were carried out using LightCycler® 480 software (v1.5.0, Roche).

13

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K

3. RESULTS

3.1. The overall workflow and herb-based virtual screen against the hydrophobic pocket of EBOV NP Our overall workflow starts with herb-based virtual screen of compounds in our TCM herbal database (TCMHD) against the specified protein target (Figure 1 left). Unlike regular virtual screen that identifies individual ligands with best docking scores, our herb-based screening approach favors selection of TCM herbs that are highly enriched in active compounds31. A group of candidate herbs selected from the virtual screen are then subjected to the affinity MS screen using their crude extracts (Figure 1 right). By incorporating targeted metabolomics into this experimental screen, we are able to distinguish the active herbs from a holistic view and then interrogate all known constituents in the active herbs to discover putative ligands for the target. Overlapping hits identified by both virtual and experimental screens are selected for downstream ligand validation and functional characterization. Our virtual screen was performed based on the high-resolution crystal structure of EBOV NP core domain (PDB code: 4Z9P)27 . The hydrophobic pocket in the C-lobe of EBOV NP was defined as the screening target site. Helix-20 of EBOV NP which occupies this hydrophobic pocket at the apo state was removed to expose the target site before molecular docking was conducted for all 14851 compounds in TCMHD30, 31. Consensus docking using four different tools (AutoDock 4.2, AutoDock Vina, standard precision (SP) and extra precision (XP) modes in Glide) were implemented to improve the accuracy and hit rate of virtual screen33, 34. A total of 108 compounds ranked 14

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 35

Page 15 of 35

View Article Online

DOI: 10.1039/C8AN02482K among top 30% by all four docking tools were identified as putative ligands (docking

hits), and they were mapped to their corresponding herbs according to the herb-compound association in TCMHD. Although 58 herbs were identified in this way, most of them contain merely one or two putative ligands. Given that the therapeutic efficacy of TCM herbs is attributed to both the abundance and the enrichment of active compounds31, we selected four candidate herbs (Piper kadsura, Piper nigrum, Ophiopogon japonicas and Salvia miltiorrhiza) containing at least six putative ligands for subsequent experiments.

3.2. Affinity MS combined with targeted metabolomcis to screen crude herbal extracts with purified EBOV NP Affinity MS analysis was combined with targeted metabolomics to screen the candidate herbs and evaluate the binding capacity of putative ligands and other constituents in each herb. To this end, the recombinant core domain of EBVO NP (residues 36–351) was purified and separately incubated with the crude extracts of individual TCM herbs. The ligand-bound NP complexes were then isolated by ultrafiltration, and the compounds dissociated from the NP target or the protein-free control samples were analyzed by LC-MS/MS using high-resolution mass spectrometry (Figure 1 right). Representative total ion chromatograms for the total extract, the NP target and control samples for each herb are shown in Figure 2A. As the virtual screen designated a few candidate herbs for the experimental screen, we devised a targeted metabolomics-based data processing strategy to verify the active 15

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K herb and search for active ligands present in the herbal extract. According to the

compound inventory of each candidate herb in TCMHD, LC-MS features were assigned to be constituents of a specific TCM herb if they met all criteria in terms of mass accuracy, retention time, isotope envelop and MSMS fragmentation patterns. For these assigned features, we then performed multivariate analysis of the NP target and control groups based on orthogonal partial least-squares discriminant analysis (OPLS-DA). Multivariate analysis revealed distinct separation of the NP target and control groups for three herbs Piper nigrum, Ophiopogon japonicas and Salvia miltiorrhiza, suggesting they probably contain active compounds associated with the NP target (Figure 2B). In contrast, much less segregation of the compound profiles was seen between the target and control groups for the fourth herb Piper kadsura (Figure 2B). Thus we considered Piper nigrum, Ophiopogon japonicas and Salvia miltiorrhiza to be active herbs enriched in active compounds whereas Piper kadsura to be inactive. To identify active compounds specifically interacting with NP, we defined a binding index (BI) for each structurally assigned feature. BI is the ratio of the MS response of a feature detected in the target vs the control samples. Screening hits were selected if their mean BI values were above 2.0 (p < 0.05, n = 4)16. A total of 8, 9 and 9 hits were identified from screening the crude extracts of Piper nigrum, Ophiopogon japonicas and Salvia miltiorrhiza, respectively and no hits were found from the extract of Piper kadsura (full data set in Supporting Information Table S1). These results fully agreed with differentiation of active vs inactive herbs based on the 16

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 35

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

View Article Online

10.1039/C8AN02482K previous multivariate analysis. The LC-MS data for three screening hitsDOI:and a

non-specific binder from different herbs are illustrated in Figure 3. Therefore combination of affinity MS and targeted metabolomics enables high-throughput interrogation of the majority of known herbal constituents including low-abundance ones, with hit selection and compound annotation completed in one experiment.

3.3. Comparison of initial hits from virtual screen and affinity MS screen Among the 29 top-ranking hits from the virtual screen and 26 initial hits from the affinity MS screen, 14 compounds from three TCM herbs overlapped, suggesting their binding to the target is mutually verified by two screening approaches (Figure 4). Chemical structures of these hits are shown in Figure S1. Their docking scores and BI values are listed in Table S2. Among the 15 hits unique to virtual screen, 13 compounds were not verified by affinity MS screen due to low BI values below the threshold. Thus they are likely to be false positive hits from the virtual screen. On the other side, 12 compounds from three herbs that eluded the virtual screen were recovered by the affinity MS screen (Figure 4). Ten of them were predicted by multiple docking algorithms to bind the NP hydrophobic pocket in similar modes. But they were excluded in the virtual screen because they only passed the selection criteria in two or three docking simulations (Table S2). We purposely set very stringent hit selection criteria (ranking as top 30% by all four docking programs) to maximize the hit rate of the virtual screen. As a trade-off, false negatives would have increased, yet luckily they could be recovered in the affinity MS screen. One hit from the affinity MS screen 17

Analyst Accepted Manuscript

Page 17 of 35

Analyst

View Article Online

10.1039/C8AN02482K not justified by any docking software (ID 11272, Table S2) is speculated toDOI: interact

with another site different from the hydrophobic pocket.

3.4. Compound isolation and ligand binding validation Among the 14 putative ligands co-identified by both the virtual screen and the affinity MS screen, we obtained 7 pure standards from commercial sources or through purification from corresponding herbs (NMR spectra for home-purified compounds shown in Figure S2). In addition, we isolated three compounds (HJ-6, MD-4 and DS-3) that were only selected by the affinity MS screen yet missed in the virtual screen, and two compounds (HFT-1, HFT-2) that were positive hits in the virtual screen yet negative in the affinity MS screen for further evaluation. These twelve compounds together with a known ligand 18β-glycyrrhetinic acid (GC-7) for EBOV NP16 were pooled and incubated with the protein target. Then an affinity MS assay was performed on this pure standard mixture to verify individual ligand binding to the target. All seven test compounds co-identified by two screening approaches were significantly associated with the NP target and considered validated ligands (Table 1). For the validated ligands, we estimated the binding affinity of each pure compound using a previously developed affinity MS-based method14,

15

. The equilibrium

dissociation constants (Kd) calculated for different ligands are mainly in the medium micro-molar range, which is typically reported for natural product-derived ligands (Table 1). Notably, compounds HJ-6, MD-4 and DS-3 identified in the affinity MS 18

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 35

Page 19 of 35

View Article Online

10.1039/C8AN02482K screen yet missed in the virtual screen were also verified in the pure ligandDOI: binding

assay. Furthermore, compounds HFT-1 and HFT-2 were confirmed to be negative, which is consistent with our affinity MS screen results (Table 1). Ligand validation results clearly revealed the distinct advantages of this integrated screening approach. First, herb-based virtual screen conveniently picks up TCM herbs to be assayed in the subsequent affinity MS screen, which is especially useful to address targets like NP with no simple bio-assays available. On the other hand, virtual screen hits can be readily verified by the affinity MS screen of the crude herbal extracts without the need of tedious compound isolation. Both false positives and false negatives in the virtual screen can be discerned by the affinity MS screen. Finally, justification of the binding poses of the affinity MS screen hits by docking simulations strengthened the confidence of ligand identification towards the targeted functional site.

3.5. Bioactivity assessment and molecular mechanism of new chemical ligands EBOV NP protein constitutes an integral part of the viral nucleocapsid and it is intimately associated with the viral RNA template, thus the oligomeric state of NP protein plays a vital role in maintaining template integrity and providing template access to the viral RNA polymerase28, 29. To assess the bioactivity of new NP ligands, we used the NP protein (residues 36-450) with a longer C-terminal tail that can shift between a monomeric state at the high salt condition and an oligomeric state at the low salt condition. Three ligands that we discovered, HJ-1, HJ-4 and HJ-6, were found 19

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K to significantly induce oligomerization of the NP protein at the high salt condition as

shown by the monodispersed peaks eluting in the void volume of SEC analysis (Figure 5A upper). But none of the ligands altered the status of the homo-oligomeric complexes of NP formed at the low salt condition (Figure 5A lower). In contrast, the endogenous peptide ligand of NP that is derived from viral protein VP35 (residues 20-48) maintained NP mostly in the monomeric form at the high salt condition and largely prevented its oligomerization at the low salt condition (Figure 5B). Concordantly, the fluorescent thermal shift assay revealed that incubation of NP with HJ-1, HJ-4 or HJ-6 reduced the thermal stability of NP purified at the high salt condition, yet incubation with the VP35 peptide increased the NP thermal stability (Figure 5C, Table S3). These results indicated that three chemical ligands are able to shift the equilibrium of EBOV NP towards the native oligomeric state of NPs, which is opposite to the disassembling effect of the VP35 peptide ligand28. Strikingly, when we titrated the VP35 peptide into the reaction while holding constant concentrations of NP and each chemical ligand, we observed a dose-dependent increase of binding of HJ-1, HJ-4 and HJ-6 to NP with an increasing amount of the VP35 peptide (Figure 6A). This is the first discovery of chemical ligands that show cooperative binding with the VP35 peptide to EBOV NP. Molecular docking was employed to interpret the possible differences in ligand binding to EBOV NP in the absence or presence of the VP35 peptide. In silico docking analysis implied that ligand HJ-4 is inserted into the hydrophobic pocket by interacting with certain residues well conserved among strains in the filovirus family27 (Figure 6B). Importantly, the 20

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 35

Page 21 of 35

View Article Online

DOI: 10.1039/C8AN02482K crystal structure of apo EBOV NP (PDB4Z9P) reveals that Helix-20 of NP folds

towards and interacts with the hydrophobic pocket in the C-lobe to form a compact structure27 (Figure 6C left). We speculate that the new chemical ligands can only bind to a small fraction of proteins with their Helix-20 swinging outward and no longer blocking the hydrophobic pocket (Figure 6C middle). When NP is in complex with the VP35 peptide which also interacts with the hydrophobic pocket, Helix-20 transits to the opposite side of the C-lobe of NP28. As the VP35 peptide is folded in a V shape with the ends of two orthogonal helices widely open, it allows easy access of HJ-4 into the hydrophobic pocket as shown by molecular docking (Figure 6C right). Due to the high affinity of the VP35 peptide, its interaction with NP would possibly push a larger fraction of proteins to open the hydrophobic pocket, thus increasing the binding percentage of chemical ligands to the same site.

4. Conclusions In summary, we developed a workflow of integrating virtual screen, affinity MS and targeted metabolomics for efficient discovery of chemical ligands from natural herbs. Thirteen compounds were identified from three TCM herbs to be new ligands targeting the hydrophobic pocket of EBOV NP. Among them, three compounds isolated from Piper nigrum (HJ-1, HJ-4 and HJ-6) were further verified in pure ligand binding assays and showed strong activities of promoting formation of large NP oligomers and reducing protein thermal stability. In addition, cooperative binding between these chemical ligands and an endogenous peptide ligand was reported for 21

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K the first time, and molecular docking was employed to propose a possible mechanism.

Taken together, this integrated approach largely reduces labor consumption in the affinity MS experiment and improves the overall success rate for screening bioactive ligands from natural products.

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 35

22

Page 23 of 35

View Article Online

Acknowledgements

DOI: 10.1039/C8AN02482K

This work was supported by ShanghaiTech University, the National Key Research and Development Program of China (2018YFA0507004 to W.S.), the Ministry of Science and Technology of China (2016YFA0502303 to L.L.), the National Natural Science Foundation of China (21633001 to L.L., 81473111, 21702219, 81773863 to L.X.), and the State Key Laboratory of Drug Research (SIMM1601ZZ-03 to L.X.). We thank the MS facility of National Center for Protein Science Shanghai (Chinese Academy of Sciences) and the MS facility of School of Life Science and Technology at ShanghaiTech University for critical technical support. Part of the docking analysis was performed on the High Performance Computing Platform of the Center for Life Science at Peking University.

Conflict of Interest Disclosure The authors declare no competing financial interest.

23

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K

References 1.

D. J. Newman; G. M. Cragg, J Nat Prod 2012, 75, 311-335.

2.

A. L. Harvey; R. Edrada-Ebel; R. J. Quinn, Nat Rev Drug Discov 2015, 14, 111-129.

3.

J. Rosen; J. Gottfries; S. Muresan; A. Backlund; T. I. Oprea, J Med Chem 2009, 52, 1953-1962.

4.

S. Gu; N. Yin; J. Pei; L. Lai, Mol Biosyst 2013, 9, 2696-2700.

5.

S. Tian; Y. Li; D. Li; X. Xu; J. Wang; Q. Zhang; T. Hou, J Chem Inf Model 2013, 53, 1787-1803.

6.

X. Liang; H. Li; S. Li, Mol Biosyst 2014, 10, 1014-1022.

7.

J. Gu; Y. Gui; L. Chen; G. Yuan; H. Z. Lu; X. Xu, PLoS One 2013, 8, e62839.

8.

J. Fang; Z. Wu; C. Cai; Q. Wang; Y. Tang; F. Cheng, J Chem Inf Model 2017.

9.

D. S. Chan; H. M. Lee; F. Yang; C. M. Che; C. C. Wong; R. Abagyan; C. H. Leung; D. L. Ma, Angew Chem Int Ed Engl 2010, 49, 2860-2864.

10.

D. L. Ma; D. S. Chan; G. Wei; H. J. Zhong; H. Yang; L. T. Leung; E. A. Gullen; P. Chiu; Y. C. Cheng; C. H. Leung, Chem Commun (Camb) 2014, 50, 13885-13888.

11.

S. Kar; K. Roy, Expert Opin Drug Discov 2013, 8, 245-261.

12.

G. Schneider, Nat Rev Drug Discov 2010, 9, 273-276.

13.

J. J. Irwin; B. K. Shoichet, J Med Chem 2016, 59, 4103-4120.

14.

X. Chen; S. Qin; S. Chen; J. Li; L. Li; Z. Wang; Q. Wang; J. Lin; C. Yang; W. Shui, Sci Rep 2015, 5,

15.

S. Qin; Y. Ren; X. Fu; J. Shen; X. Chen; Q. Wang; X. Bi; W. Liu; L. Li; G. Liang; C. Yang; W. Shui,

8361. Anal Chim Acta 2015, 886, 98-106. 16.

X. Fu; Z. Wang; L. Li; S. Dong; Z. Li; Z. Jiang; Y. Wang; W. Shui, Sci Rep 2016, 6, 29680.

17.

Z. Wang; X. Li; M. Chen; F. Liu; C. Han; L. Kong; J. Luo, Talanta 2018, 180, 337-345.

18.

N. Vinayavekhin; A. Saghatelian, J Am Chem Soc 2011, 133, 17168-17171.

19.

X. Li; T. A. Gianoulis; K. Y. Yip; M. Gerstein; M. Snyder, Cell 2010, 143, 639-650.

20.

Q. Qin; B. Wang; J. Wang; M. Chang; T. Xia; X. Shi; G. Xu, Talanta 2019, 194, 63-72.

21.

Z. Yang; Y. Zhang; L. Sun; Y. Wang; X. Gao; Y. Cheng, Anal Chim Acta 2012, 719, 87-95.

22.

H. P. Song; H. Zhang; Y. Fu; H. Y. Mo; M. Zhang; J. Chen; P. Li, J Chromatogr B Analyt Technol

23.

H. P. Song; J. Chen; J. Y. Hong; H. Hao; L. W. Qi; J. Lu; Y. Fu; B. Wu; H. Yang; P. Li, Chem

Biomed Life Sci 2014, 961, 56-61. Commun (Camb) 2015, 51, 1494-1497. 24.

H. P. Song; S. Q. Wu; L. W. Qi; F. Long; L. F. Jiang; K. Liu; H. Zeng; Z. M. Xu; P. Li; H. Yang, J Chromatogr A 2016, 1456, 176-186.

25.

E. C. Holmes; G. Dudas; A. Rambaut; K. G. Andersen, Nature 2016, 538, 193-200.

26.

S. Khan; Muhammad; A. Rauf; A. Khan; M. Rizwan; S. Patel; H. Khan; A. M. Mahasneh; M. S. Mubarak, Infect Disord Drug Targets 2017.

27.

S. S. Dong; P. Yang; G. B. Li; B. C. Liu; W. M. Wang; X. Liu; B. R. Xia; C. Yang; Z. Y. Lou; Y. Guo; Z. H. Rao, Protein & Cell 2015, 6, 351-362.

28.

D. W. Leung; D. Borek; P. Luthra; J. M. Binning; M. Anantpadma; G. Liu; I. B. Harvey; Z. Su; A. Endlich-Frazier; J. Pan; R. S. Shabman; W. Chiu; R. A. Davey; Z. Otwinowski; C. F. Basler; G. K. Amarasinghe, Cell Rep 2015, 11, 376-389.

29.

R. N. Kirchdoerfer; D. M. Abelson; S. Li; M. R. Wood; E. O. Saphire, Cell Rep 2015, 12, 140-149.

30.

M. He; X. Yan; J. Zhou; G. Xie, J Chem Inf Comput Sci 2001, 41, 273-277. 24

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 35

Page 25 of 35

View Article Online

31.

H. Liang; H. Ruan; Q. Ouyang; L. Lai, Scientific reports 2016, 6, 36767.

32.

M. Choi; C. Y. Chang; T. Clough; D. Broudy; T. Killeen; B. MacLean; O. Vitek, Bioinformatics

DOI: 10.1039/C8AN02482K

2014, 30, 2524-2526. 33.

T. Cheng; X. Li; Y. Li; Z. Liu; R. Wang, J Chem Inf Model 2009, 49, 1079-1093.

34.

T. Tuccinardi; G. Poli; V. Romboli; A. Giordano; A. Martinelli, J Chem Inf Model 2014, 54, 2980-2986.

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

25

Analyst

View Article Online

Table 1 Ligand binding validation of each compound by affinity MS analysis

a b

TCMHD ID

Name

16384

HJ-1

17437

b

a

Formula

BI

a

Kd (μM)

C21H27NO3

7.48± 0.11

24.4±3.4

HJ-2

C20H23NO3

2.77 ±0.39

73.1±9.2

17469

HJ-3

C21H29NO3

7.61 ±0.05

16.7±2.3

17436

HJ-4

C20H25NO3

3.20 ±0.10

20.1±3.6

17435

HJ-6

C20H27NO3

3.80 ±0.11

33.8±4.2

14635

MD-1

C19H18O6

4.33 ±0.08

44.8±3.0

14636

MD-2

C19H20O5

4.10 ±0.04

88.5±2.0

19072

MD-4

C27H42O4

3.44 ±0.22

>100

4950

DS-1

C19H20O2

2.28 ±0.02

73.2 ±0.5

4629

DS-3

C17H16O3

2.32 ±0.02

>100

8038

HFT-1

C21H24O5

1.58 ±0.06

NA

8039

HFT-2

C21H24O5

1.61 ±0.21

NA

8841

GC-7

C30H46O4

3.66 ±0.06

47.1±7.3

c

Mean values and standard deviation from experimental triplicates. Compounds in bold are co-identified by the virtual screen and the affinity MS screen.

c

GC-7 is a known ligand of EBVO NP used as positive control.

26

DOI: 10.1039/C8AN02482K

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 35

Page 27 of 35

View Article Online

DOI: 10.1039/C8AN02482K

Figure legends

Figure 1 The overall workflow of ligand discovery from natural herbs by integrating virtual screen, affinity MS and targeted metabolomics.

Figure 2 Representative LC-MS chromatograms of the crude extracts, the NP target sample and the control for each TCM herb (A), and multivariate analysis of the LC-MS data from the target and control with an OPLS-DA model (B, n=4). The results exhibited a well segregated pattern between the target and control groups for three herbs (HJ, MD and DS) and poor separation of the target and control groups for the fourth herb HFT. HJ, Piper nigrum; MD, Ophiopogon japonicas; DS, Salvia miltiorrhiza; HFT, Piper kadsura.

Figure 3 Selected ion chromatograms (left) and high-resolution MS spectra (right) for three screening hits HJ-1, HJ-4, DS-1 and a non-specific binder HFT-1. Binding index (BI), chemical structure and formula are shown for each compound.

Figure 4 Hits from the virtual screen and the affinity MS screen. Shown are numbers of total hits (A) or hits from each TCM herb (B) with two screening methods.

Figure 5 Influences of new chemical ligands on the oligomerization and thermal stability of EBOV NP. (A) SEC analysis of NP alone or NP incubated with a chemical ligand at the high salt (upper) or low salt (lower) condition. Peaks eluting at ~ 8 mL 27

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K (void volume) are large oligomers of NP and peaks at ~15 mL are monomers of NP.

(B) SEC analysis of NP alone or NP incubated with the VP35 peptide at the high salt (upper) or low salt (lower) condition. (C) Thermal denaturation curves of NP with a chemical ligand (upper) or the VP35 peptide (lower) at the high salt condition.

Figure 6 The VP35 peptide promotes binding of three chemical ligands to EBOV NP. (A) Binding fraction of each chemical ligand to EBOV NP titrated with an increasing amount of the VP35 peptide. Error bars represent SD from independent experiments (n=3). (B) Docking model of HJ-4 interacting with the hydrophobic pocket in the C-lobe of NP. (C) Orientation of Helix-20 in EBOV NP apo structure (left), in the docking model of HJ-4 bound NP (middle), and in the docking model of NP bound with both HJ-4 and VP35 peptide (right).

28

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 35

Page 29 of 35

View Article Online

DOI: 10.1039/C8AN02482K

Figure 1 Protein expression

Herb

TCMDB

Crude extracts Target site of NP

14851 NPS

Purification

Incubation

Ultrafiltration enrichment

Precise docking

MS/MS

Intensity

Docking hits

LC-MS/MS

RT

MS

m/z

Identification of putative ligands

Herb-compound association

Bioactivity Ligand assessment validation Affinity MS + targeted metabolomics

Candidate herbs Virtual screen

29

Analyst Accepted Manuscript

Consensus docking

Intensity

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K

Figure 2

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 35

30

Page 31 of 35

View Article Online

DOI: 10.1039/C8AN02482K

Figure 3

80%

Target

HJ-1

Control

BI=8.25

60% 40%

342.2101

100% 80%

MS intensity (%)

100%

MS intensity (%)

A

60%

C21H27NO3

40%

343.2175

20%

20%

344.2264 0%

17.6 Time (min)

B 80%

BI=8.46

60% 40%

60%

329.1925

330.2085 327

16.3

MS intensity (%)

BI=12.63

60% 40% 20%

330

60%

40%

24.3

282.1193 283.2016 280

MS intensity (%)

BI=1.34

60% 40% 20%

281

282 m/z

283

284

357.1687

100%

HFT-1

331

C19H20O2

0% 23.8 Time (min)

D 80%

329 m/z

80%

20%

0% 23.3 100%

328

281.1167

100%

DS-1

345

40%

0% 15.8 Time (min)

100% MS intensity (%)

344

C20H25NO3

20%

0% 15.3

80%

343 m/z

80%

20%

C

342 328.1907

100%

HJ-4

MS intensity (%)

MS intensity (%)

100%

341

18.1

80% 60% C21H24O5 40% 358.1740

20%

359.1700

0% 11.6

0% 12 Time (min)

12.4

356

31

357

358 m/z

359

360

Analyst Accepted Manuscript

0% 17.1

MS intensity (%)

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

Analyst

View Article Online

DOI: 10.1039/C8AN02482K

Figure 4

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 35

32

Page 33 of 35

View Article Online

DOI: 10.1039/C8AN02482K

Figure 5

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analyst

33

Analyst

View Article Online

DOI: 10.1039/C8AN02482K

Figure 6

Analyst Accepted Manuscript

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 35

34

Page 35 of 35

Analyst View Article Online

Published on 26 January 2019. Downloaded by Iowa State University on 1/26/2019 6:36:05 PM.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

77x42mm (300 x 300 DPI)

Analyst Accepted Manuscript

DOI: 10.1039/C8AN02482K