NCI Program for Natural Product Discovery: A ... - ACS Publications

May 29, 2018 - Library of Natural Product Fractions for High-Throughput Screening ... Data Management Services, Inc., Frederick National Laboratory fo...
2 downloads 0 Views 2MB Size
Subscriber access provided by NORTH CAROLINA A&T UNIV

Article

NCI Program for Natural Product Discovery: A Publicly-Accessible Library of Natural Product Fractions for High-Throughput Screening Christopher C Thornburg, John R. Britt, Jason R. Evans, Rhone K. Akee, James A. Whitt, Spencer K. Trinh, Matthew J. Harris, Jerell R. Thompson, Teresa L. Ewing, Suzanne M. Shipley, Paul G. Grothaus, David J. Newman, Tanja Grkovic, and Barry R. O'Keefe ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.8b00389 • Publication Date (Web): 29 May 2018 Downloaded from http://pubs.acs.org on May 30, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 42 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

ACS Chemical Biology

NCI Program for Natural Product Discovery: A PubliclyAccessible Library of Natural Product Fractions for HighThroughput Screening Christopher C. Thornburg,† John R. Britt,† Jason R. Evans,‡,§ Rhone K. Akee,† James A. Whitt,† Spencer K. Trinh,† Matthew J. Harris,† Jerell R. Thompson,† Teresa L. Ewing,† Suzanne M. Shipley,† Paul G. Grothaus,§ David J. Newman, § Tanja Grkovic,† and Barry R. O’Keefe*,§,⊥



Natural Products Support Group, Leidos Biomedical Research, Inc., Frederick National

Laboratory for Cancer Research sponsored by the National Cancer Institute, Frederick, Maryland 21702-1201, United States, ‡Data Management Services, Inc., Frederick National Laboratory for Cancer Research sponsored by the National Cancer Institute, Frederick, Maryland 21702-1201, United States, §Natural Products Branch, Developmental Therapeutics Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute, Frederick, Maryland 21702-1201, United States,



Molecular Targets Program, Center for Cancer Research, National Cancer

Institute, Frederick, Maryland 21702-1201, United States KEYWORDS. Prefractionation, high-throughput screening, natural product library. Heading: NCI Prefractionated Natural Product Library

ACS Paragon Plus Environment

1

ACS Chemical Biology 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 42

ABSTRACT

The US National Cancer Institute’s (NCI) Natural Product Repository is one of the world’s largest, most diverse collections of natural products containing over 230,000 unique extracts derived from plant, marine and microbial organisms that have been collected from biodiverse regions throughout the world. Importantly, this national resource is available to the research community for the screening of extracts and the isolation of bioactive natural products. However, despite the success of natural products in drug discovery, compatibility issues that make extracts challenging for liquid handling systems, extended timelines that complicate natural product-based drug discovery efforts and the presence of pan-assay interfering compounds have reduced enthusiasm for the highthroughput screening (HTS) of crude natural product extract libraries in targeted assay systems. To address these limitations, the NCI Program for Natural Product Discovery (NPNPD), a newly launched, national program to advance natural product discovery technologies and facilitate the discovery of structurally defined, validated lead molecules ready for translation, will create a prefractionated library from over 125,000 natural product extracts with the aim of producing a publicly-accessible, HTS-amenable library of >1,000,000 fractions. This library, representing perhaps the largest accumulation of natural-product based fractions in the world, will be made available free of charge in 384-well plates for screening against all disease states in an effort to reinvigorate natural product-based drug discovery.

ACS Paragon Plus Environment

2

Page 3 of 42 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

ACS Chemical Biology

Natural products are composed of relatively few elements—carbon, hydrogen, oxygen and an occasional nitrogen, sulfur or halogen—yet the molecular complexity and collective diversity of natural product scaffolds, which is driven by the continual evolution of the biosynthetic genes that produce compounds that interact with the unique three-dimensional structures of nucleic acids and proteins in cellular pathways, provides unparalleled opportunity for drug discovery efforts.1 The value is further demonstrated by the significant proportion of small-molecule approved drugs from 1981 to 2014 in disease areas such as cancer (>60%) that were developed from a natural product or were based on a natural product pharmacophore.2 This ratio is especially impressive given the relatively low percentage (125,000 extracts from the NCI’s Natural Product Repository extract library as part of the NCI Program for Natural Product Discovery (NPNPD) with the aim of accelerating natural-product based drug discovery by producing a publicly-accessible, HTS-amenable library of natural product samples (1,000,000 fractions), which will be made available free of charge in 384-well plates to researchers for screening against any disease.

ACS Paragon Plus Environment

4

Page 5 of 42 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

ACS Chemical Biology

Figure 1. (A) Collection locations of marine organisms (blue), plants (green) and microbial specimens (red) that comprise the US National Cancer Institute’s Natural Product Repository. The size of the hexagon is indicative of the relative size of the individual collection and locations with more than one type of collected organism are shaded in gray. Since 1986, samples collected through the NCI Natural Products Collection program have been acquired through collection agreements based on the NCI Letter of Collection with each participating source country or their representatives, which stipulates equitable benefit sharing from commercial products derived from discoveries made through these collections. Plant specimens have been collected in over 25 countries through contracts with the Missouri Botanical Garden (Africa and Madagascar), the New York Botanical Garden (Central and South America), the University of Illinois at Chicago (Southeast Asia), and the Morton Arboretum and World Botanical Associates (United States and territories). Marine organisms have been collected throughout the world through contracts with SeaPharm, the Harbor Branch Oceanographic Institute, the Australian Institute of Marine Science, the University of Canterbury, New Zealand, and the Coral Reef Research Foundation based in Palau in Micronesia. Microbial collections were first obtained through an initial contract with the University of Connecticut and then from the middle 1990s, through collaborations with the United States Department of Agriculture.17,18 (B) Overview of the automation procedure developed to facilitate the production of the NPNPD fraction library: (1) Extracts from the NCI Natural Products Repository are weighed and reconstituted in either organic solvent or water. (2) Dissolved extracts are adsorbed onto cotton rolls contained within an empty SPE cartridge, while 2D-barcoded tubes (10 mL) are preweighed on an automated weighing station (2 h). (3) Extracts (n = 88) are

ACS Paragon Plus Environment

5

ACS Chemical Biology 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 42

prefractionated on a customized positive pressure solid phase extraction workstation (ppSPE) with two robotic arms working in parallel to produce seven fractions per extract (3.5 h; n = 616 fractions). (4) Fractions are dried using high-performance centrifugal evaporation systems (18 h) and the final mass of each fraction is determined on an automated weighing station (A; 2 h). (5) Assay plates containing fractions and unprocessed extracts are generated from each set of 88 extracts (2 × 384-well plates).

RESULTS AND DISCUSSION NCI Natural Products Repository (NPR) Extracts. Extracts in the NCI NPR are prepared using both an aqueous and organic solvent extraction process, resulting in two sequential extracts per collected specimen.19 Importantly, depending on the source material, the order in which the aqueous extract is generated significantly influences the final composition of the extract. For plant specimens, the organic extract is produced prior to the water extract using a solvent mix of 1:1 dichloromethane (DCM) / methanol (MeOH), followed by 100% MeOH, resulting in an organic extract enriched in both non-polar and mid-polarity components. In contrast, marine organisms are first extracted with water, which removes salts and polysaccharides, followed by an organic solvent extraction (1:1 DCM/MeOH) of the lyophilized biomass. There are currently more than 30,000 microbial extracts in the NCI NPR that have been derived from fungal and bacterial fermentations grown under varying conditions. To disrupt cells and mat-like material, organic extracts are prepared from high-shear homogenized cultures in 10% MeOH (v/v), followed by partitioning with 50% DCM. The filtered aqueous layer (~1 L) is then lyophilized to create an aqueous extract. NPR Prefractionated Library Design. To develop a natural-products based fraction library that is not only well-adapted to HTS, but also captures the diversity of the NCI NPR, several essential criteria were considered during the methods development process. Importantly, platforms that could accommodate 0.2–1.0 g loading of each extract, provide enough material within each

ACS Paragon Plus Environment

6

Page 7 of 42 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

ACS Chemical Biology

fraction for generating a large number of HTS plates, and be adapted for high-throughput processing were needed. In this regard, the versatility and high-throughput potential of solid-phase extraction (SPE) methods such as those typically used for sample extraction, concentration and cleanup were considered over high-performance liquid chromatography (HPLC) or supercritical fluid chromatography (SFC) prefractionation strategies that, by comparison, have a decreased loading capacity, lower throughput, and limited sample collection formats for determining the mass of each fraction. Another advantage of using SPE is the potential to generate a small set of fractions (5 to 10) for each extract that cover a diverse range of metabolites with regards to polarity, and thus maximize the coverage of chemical and biological diversity screened within a given HTS campaign. As the goal was the pre-fractionation of >125,000 crude extracts, the methodologies chosen also had to be optimized for both cost and throughput. However, prior to developing a SPEbased prefractionation method that could be readily adapted to an automated, high-throughput robotics platform (Figure 1B), techniques to address the logistics of sample loading, elution, drying and weighing were developed. First, the relatively large starting mass of each extract presented significant challenges with regards to sample reconstitution and automated liquid handling loading due to sample viscosity and precipitation, which also impacted the reproducibility of the SPE sample loading and elution techniques. However, after some preliminary experimentation, dissolving 200–250 mg of the organic solvent extracts or 400–1000 mg of the aqueous extracts in 4.5 mL MeOH-EtOAc-MTBE (6:3:1) or 100% H2O, respectively, followed by directly adsorbing onto a cotton plug and freeze-drying, resulted in a high-throughput amenable loading technique with minimal clogging of the SPE frit and matrix. This dry-loading technique, whereby the freezedried, cotton-adsorbed samples are housed in a separate cartridge also allows the SPE adsorbent to be pre-equilibrated following the manufacturer’s recommended guidelines. Next, it was

ACS Paragon Plus Environment

7

ACS Chemical Biology 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 42

determined that a controlled rate of elution ( 90%) and of sufficient quantity for hit validation, dereplication and potential secondary screens.

Table 3. Average LCMS Peak Count by Fraction for All of the Extracts Prefractionated in the NPNPD Fraction Library Pilot Study Using the C8, RP-2.7 Method. Avg Peak Count per C8 SPE Fraction (s) 1

2

3

4

5

6

7

2 (1)

2 (2)

2 (2)

4 (3)

7 (3)

11 (8)

9 (8)

Total MSb

11 (7)

9 (8)

12 (8)

22 (16)

43 (44)

40 (34)

23 (12)

Majorsc

2 (3)

2 (3)

3 (3)

7 (7)

17 (21)

12 (12)

6 (3)

Minorsd

9 (6)

8 (5)

9 (6)

15 (13)

26 (25)

28 (27)

17 (11)

ELSD

a

a

Total number of analytes detected using an evaporative light scattering detector (ELSD). bTotal number of analytes estimated from the LC-HRMS data and defined by m/z value, retention time and intensity (MS buckets). cTotal number of MS buckets within each detectable ELSD retention time window. cTotal number of MS buckets that were not detected in the corresponding ELSD chromatogram. Standard deviation (s).

The organic extract of the plant F. virosa was selected for scale-up isolation work based on the observation that while the organic extract was inactive in the NCI-60 screen a number of

ACS Paragon Plus Environment

22

Page 23 of 42 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

ACS Chemical Biology

fractions showed potent growth inhibition across most NCI-60 cell lines at 10 µg/mL and, in the C8, RP-2.7 series, cell-line selectivity toward several melanoma cell lines (Figure 2A and Supplementary Tables S5–S9). The major active principle, phyllanthusmin D (1),45 was purified in a single preparative HPLC separation, while a minor active triterpenoid, dichapetalin P (2),46 was isolated in two steps. Although both compounds were found to be cytotoxic at low micromolar concentrations with mean GI50 values of 0.43 µM (1) and 1.74 µM (2), phyllanthusmin D was highly selective and 200-fold more potent against melanoma cell lines at the TGI level (Supplementary Figures S22 and S23). In the microbial extracts, prefractionation of a Penicillium griseofulvum (strain ID 0G0S1555, NSC number F250369) organic extract using the developed C8 SPE-based method showed the presence of two separable regions of bioactivity when tested against the NCI-60 human tumor cell lines panel (Figure 5A). The PLS scores plot showed a separation of the active fractions (F1–F6) from the inactive fractions (F4 and F7), with F1–3 grouped closely together, while F5 and F6 were clustered away from one another and the rest of the P. griseofulvum metabolome, suggesting the presence of two major active principles or a single minor active component in both fractions (Figure 5B). Subsequent bioassay-guided isolation from a single preparative HPLC purification step identified the active principle in fractions F1–3 to be the mycotoxin patulin (3),47 while the small molecule responsible for the activity of fraction F5 and F6 was identified as a prenylated diketopiperazine, mycelianamide (4).48 In the NCI-60 assay, compounds 3 and 4 showed low micromolar cytotoxicity with mean GI50 values of 2.95 µM and 1.95 µM, respectively (Supplementary Figures S24–S26). Although the NCI-60 dose response pattern of the organic solvent extract was similar to 3 by a COMPARE analysis (GI50 = 0.71, TGI = 0.65), it was not significantly correlated with 4 (COMPARE at GI50 = 0.16, TGI = 0.15). Thus, under the growth

ACS Paragon Plus Environment

23

ACS Chemical Biology 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 42

conditions employed for P. griseofulvum 0G0S1555, 3 is likely the major active metabolite responsible for the initial activity profile of the organic extract, while 4 is present as a minor active metabolite, demonstrating the potential of the C8 SPE-based method to separate confounding toxicity from other, perhaps more interesting, activities.

Figure 5. Prefractionation of a fungal extract of Penicllium griseofulvum using the developed C8 SPE-based method. (A) Partial least-squares (PLS) scores plot based on the UPLC-HRMS and NCI-60 bioassay data shows a separation of the bioactive (F1, F2, F3, F5 and F6) and inactive fractions (F4 and F7). (B) A presence absence analysis of the UPLC-HRMS data for the active principles, patulin (3, m/z 153.02 [M H]–, RT = 0.5 min) and mycelianamide (4, m/z 369.23 [M 2H2O H]+, RT = 5.5 min) shows that the separation of NCI-60 bioactive fractions is driven by the presence of 3 (F1–F3) and 4 (F5–F6).

ACS Paragon Plus Environment

24

Page 25 of 42 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

ACS Chemical Biology

O

O

O H O

H H AcO

An organic extract from the sponge

O H O

H H AcO

H O OAc

H O OAc

5

6 H OAc O H OAc

7

Spongionella sp. was selected for further proofof-concept studies based on the potent antiproliferative activity observed against the NCI-60 screen (GI50 = 2.88 µg/mL, TGI = 11.75 µg/mL, and LC50 = 61.66 µg/mL, Supplementary Figure S27). Prefractionation

using the developed C8, RP-2.7 SPE method and subsequent preparative HPLC separation led to the isolation of the known gracilin-type trisnorditerpenes gracilins H and I (5 and 6)49 as a 3:2 mixture from a single fraction, while a minor analog, gracilin A (7),50 was purified in a successive HPLC step. In the NCI-60 assay, the 3:2 mixture of 5 and 6 exhibited cytotoxicity against the NCI60 human tumor cell lines panel with a GI50 value of 0.93 µM (Supplementary Figure S28), while 7 had a GI50 value of 1.05 µM (Supplementary Figure S29). Additional purification of the two stereoisomers showed that the two compounds had GI50 values of 0.76 µM and 3.39 µM for gracilin H (5) and I (6), respectively (Supplementary Figures S3 and S31). Collectively, these examples, representing plant, marine and microbial sourced secondary metabolites, emphasize the value of the prefractionation methodology to efficiently identify multiple active constituents in a natural product extract, as well as concentrate minor, biologically active natural products in an automated fractionation and purification workflow to accelerate the dereplication process.

ACS Paragon Plus Environment

25

ACS Chemical Biology 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 42

Figure 6. Structures of additional bioactive compounds isolated from the NPNPD model fraction library and corresponding validation set using high-throughput semi-preparative HPLC methods. Bengamide B (8),51 bengazole A (9),52 trichothecinol A (10),53 trichothecin (11),54 8deoxytrichothecin (12),55 8-isocyano-11(20)-ene-15-amphilectaformamide (13),56 8,15-

ACS Paragon Plus Environment

26

Page 27 of 42 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

ACS Chemical Biology

diisocyano-11(20)-amphilectene (14),56,57 luffariellolide (15),58 epinuapapuin B (16),59 muqubilin (17),60,61 aaptamine (18),62 demethyloxyaaptamine (19),62 3,5-dibromo-2-(2,463 dibromophenoxy)phenol (20), 2,3,5-tribromo-6-(2,4-dibromophenoxy)phenol (21),63 hippuristanol (22),64 erythrolide F (23),65 mycgranol (24),66 agelasine D (25),67 plakinidine A (26).68

Chemoinformatic Analysis of NP Compounds Arising from the NPNPD Model Prefractionated Library. In addition to the compounds isolated on a preparative scale, a range of biologically active natural products, representing various biosynthetic pathways from the plant, fungal and marine extracts used in this pilot study, were isolated on a high-throughput semipreparative scale (1-5 mg/separation; Onyx Monolithic C18, 100 x 10 mm) with purities greater than 90% (Figure 6). The compounds, their select physicochemical properties and measured accurate mass values are presented in the Supporting Information (Tables S52–S54). Overall, the isolated compounds ranged in molecular weight from 154.1 to 656.8 Da, with a spread of clogP values ranging from

0.3 to 7.0. A chemoinformatics analysis was conducted using a

previously defined set of structural and physiochemical parameters to compare the chemical properties of the isolated compounds to the biologically relevant chemical space occupied by NPsourced drugs approved between 1981 and 2010.69 As shown in Figure 7, the distribution of the isolated compounds from this study with that of approved natural product-based drugs demonstrates that the prefractionation methodology described here, using the medium retentive characteristics of the selected C8 SPE matrix, can capture drug-like biologically active natural products with a range of different physiochemical properties, comparable to that of natural product-sourced and –inspired drugs currently approved for clinical use.

ACS Paragon Plus Environment

27

ACS Chemical Biology 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 42

Figure 7. (A) PCA plot showing the distribution of the isolated natural products from the NPNPD prefractionated library methods development set with that of approved natural product-based drugs. The first three principal components (PC1-PC3) contain 77% of the covariance in the set of 20 structural and physiochemical described previously in the analysis of NCEs from 1981–2010.69 As shown in the component loadings plot for the PCA (B), distribution of compounds along the axis of PC1 is largely driven by vectors related to molecular weight (MW), such as the number of heteroatoms (N, O), hydrogen bond donor/acceptors (HBD, HBA), rotatable bonds (RotB), the topological polar surface area (tPSA), and Van der Waals surface area (VWSA). Positioning along the axis of PC2 is influenced in the positive direction by the calculated n-octanol/water partition coefficient (ALOGPs) and in the negative direction by the calculated aqueous solubility (ALOGpS) and relative polar surface area (relPSA). PC3 represents molecular complexity with regards to the number of stereocenters (nStereo) and fraction of sp3 carbons (Fsp3). In general, molecules with greater molecular complexity are shown in the positive direction. See Supplementary Tables S52 and S53, for the calculated structural and physiochemical parameters of the NPNPD compounds that were projected onto the model generated by Stratton et al.69

CONCLUSIONS In summary, the prefractionation methodology developed to generate the NPNPD fraction library resulted a high-throughput, automated SPE platform that was shown to produce partially purified fractions containing an even distribution of metabolites, to concentrate minor, biologically active natural products, and to accommodate a sufficient amount of starting material (0.2–1.0 g)

ACS Paragon Plus Environment

28

Page 29 of 42 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

ACS Chemical Biology

to support screening and subsequent streamlined downstream processes in drug discovery programs. Although improvements in dereplication strategies and recent advances in NMR spectroscopy and mass spectrometry instrumentation now allow the structure elucidation of complex natural product scaffolds at the microgram level,44,70 a significant challenge remains in many modern drug discovery programs is access to the chemical diversity of natural products. This is largely due to the desire to screen libraries of pure compounds, which presents challenges regarding compound availability and supply. Of the more than 250,000 natural products reported, only a small fraction are commercially available.71 Furthermore, it is estimated that 90% of the collective biodiversity of marine, microbial and plant species have yet to be investigated in drug discovery campaigns.72 The NCI effort reported here—the production of a large, publicly-available prefractionated natural product library—will enhance the chemical diversity accessible for screening and hopefully stimulate the continued success of natural products as biological probes and lead compounds in drug discovery. To this end, the aim of the NPNPD is to expand the availability of high quality, well annotated natural product-based screening libraries derived from the NCI Natural Products Repository through the generation of a HTS-amenable library containing over 1,000,000 fractions from more than 125,000 marine, microbial and plant extracts that have been collected from biodiverse regions throughout the world. It is anticipated that the first set of 150,000 plated prefractionated extracts will be available for screening by January 2019.

ACS Paragon Plus Environment

29

ACS Chemical Biology 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 42

EXPERIMENTAL SECTION General Experimental Procedures. NMR spectra were recorded at 25 °C on a 500 MHz Bruker Avance spectrometer, equipped with a triple resonance 5mm CPTCI cryo-probe. The 1H and 13C NMR chemical shifts were referenced to the solvent peaks for CD3OD at δH 3.30 and δC 49.05. NMR FID processing and data interpretation was done using MestReNova software, version 11.0. UPLC-HRMS data were acquired using a Phenomenex Kinetex C8 [1.7 µm, 50

2.1 mm]

column on a Waters Acquity UPLC system coupled to a Waters LCT Premier TOF mass spectrometer with an electrospray ionization source. The mass spectrometric data for compounds 1−26 were recorded on an Agilent 6545 Accurate-Mass Q-TOF LC/MS system (1290 Infinity II) equipped with a dual AJS ESI source. Preparative-scale HPLC purification was performed with a Waters Prep LC system, equipped with a Delta 600 pump and a 996-photodiode array detector or an Agilent 1200 series, equipped with a 6130-single quad mass spectrometer, using either a Luna C18 [10 µm, 250

21.2 mm], Phenomenex Kinetex C8 HPLC columns [5 µm, 150

21.2 mm],

or a Phenomenex Onyx C18 HPLC columns [100 × 10 mm]. Semipreparative-scale HPLC purification was performed with a Gilson HPLC purification system equipped with a GX-281 liquid handler, a 322-binary pump, a 172- photodiode array detector, and a Verity 1900 MS detector. All solvents used for SPE, HPLC, and MS were GC/LC-MS grade, and the H2O for preparative HPLC was Millipore Milli-Q PF filtered. Extraction and Prefractionation of Natural Product Extracts for NPR Model Library. Plant, marine and microbial extracts included in the model prefractionated library methods development, validation and proof-of-principle studies were obtained from the NCI Natural Products Repository and were prepared according to the extraction procedures detailed in McCloud,19 unless otherwise noted. A portion of the organic solvent extracts (200–250 mg) and

ACS Paragon Plus Environment

30

Page 31 of 42 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

ACS Chemical Biology

aqueous extracts (400–1000 mg) were weighed into barcoded tubes and dissolved in 4.5 mL MeOH-EtOAc-MTBE (6:3:1) or H2O, respectively. Dissolved samples were adsorbed onto cotton rolls (1.27 cm

3.81 cm, TIDI Products, LLC) contained within an empty SPE cartridge, followed

by freeze-drying to remove solvents. Prior to prefractionation, each SPE cartridge was washed with three column volumes of 100% MeOH, followed by equilibration with three column volumes of the first eluent of the respective solvent series employed. For the normal phase prefractionation series, organic solvent extracts were prefractionated on 2 g Isolute® Diol SPE cartridges (2,3dihydroxypropoxypropyl, non-endcapped functionalized silica: 50 µm, 60 Å) using four different gradients: NP-1.7 [Hex/CH2Cl2 (9:1), Hex/CH2Cl2 (7:3), CH2Cl2/EtOAc (5:1), 100% EtOAc, EtOAc/MeOH (5:1), MeOH/EtOAc (1:1), and 100% MeOH]; NP-1.10 [100% Hex, Hex/ CH2Cl2 (9:1), Hex/ CH2Cl2 (7:3), CH2Cl2/EtOAc (5:1), EtOAc/CH2Cl2 (7:3), 100% EtOAc, EtOAc/MeOH (5:1), MeOH/EtOAc (1:1), MeOH/EtOAc (5:1), and 100% MeOH]; NP-2.7 [100% Hex, Hex/EtOAc (5:1), Hex/EtOAc (1:1), EtOAc/Hex (7:3), 100% EtOAc, EtOAc/MeOH (7:3), and 100% MeOH]; and NP-2.10 [100% Hex, Hex/EtOAc (5:1), Hex/EtOAc (3:2), Hex/EtOAc (2:3), EtOAc/Hex (5:1), 100% EtOAc, EtOAc/MeOH (5:1), EtOAc/MeOH (1:1), MeOH/EtOAc (5:1), and 100% MeOH]. For the reversed-phase prefractionation series, the organic solvent extracts were prefractionated on the following 2 g SPE cartridges: Isolute® C8 SPE (octyl, non-endcapped: 50 µm, 60 Å); HyperSepTM C8 (octyl, non-endcapped: 50 µm, 60 Å); Oasis® HLB (hydrophiliclipophilic-balanced divinylbenzene, N-vinylpyrrolidone copolymer; 60 µm); and Isolute® C8, Oasis® HLB mixed-bed SPE column (1.5 g and 0.5 g, respectively). The aqueous extracts were prefractionated on 2 g SPE cartridges that included Bakerbond Wide-PoreTM C4 (butyl, nonendcapped: 50 µm, 60 Å); Isolute® C8 and HyperSepTM C8; Oasis® HLB; Bondesil® ENV (hydroxylated polystyrene divinylbenzene copolymer; 125 µm); and Bakerbond Wide-PoreTM C4,

ACS Paragon Plus Environment

31

ACS Chemical Biology 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 42

Oasis® HLB mixed-bed SPE (1.5 g and 0.5 g, respectively). The reversed-phase solvent series for organic solvent extracts included four different gradients: RP-1.7 [ACN/H2O (5:95), ACN/H2O (3:7), ACN/H2O (1:1), ACN/H2O (7:3), ACN/MeOH/H2O (85:5:10), ACN/MeOH (1:1), and ACN/CH2Cl2 (3:7) or ACN/MeOH (1:1) for ppSPE system]; RP-1.10 [100% H2O, ACN/H2O (5:95), ACN/H2O (15:85), ACN/H2O (3:7), ACN/H2O (1:1), ACN/H2O (7:3), ACN/MeOH/H2O (70:20:10), ACN/MeOH/H2O (60:30:10), ACN/MeOH (1:1), and ACN/CH2Cl2 (3:7) or ACN/MeOH (1:1) for ppSPE system]; RP-2.7 [MeOH/H2O (5:95), MeOH/H2O (1:4), MeOH/H2O (2:3), MeOH/H2O (3:2), MeOH/H2O (4:1), 100% MeOH, and MeOH/CH2Cl2 (3:7) or ACN/MeOH (1:1) for ppSPE system]; and RP-2.10 [MeOH/H2O (5:95), MeOH/H2O (15:85), MeOH/H2O (1:3), MeOH/H2O (35:65), MeOH/H2O (1:1), MeOH/H2O (65:35), MeOH/H2O (3:1), MeOH/H2O (85:15), 100% MeOH, and MeOH/CH2Cl2 (3:7) or ACN/MeOH (1:1) for the automated SPE system]. The aqueous extracts were prefractionated with the reversed-phase solvent schemes, with two additional 100% H2O washes (2

8 mL) included prior to the collection

of the first eluent. For each prefractionation scheme, the cotton-adsorbed, lyophilized SPE sample cartridge was stacked above an adsorbent-containing SPE cartridge using a SPE tube adapter (Supelco, Sigma-Aldrich) for individualized fractionation or equipped with barbed luer (Nordson Medical) and SPE sealing cap (Gilson Inc.) and placed within a custom 48-postion manifold printed out of polylactic acid using a Series 1 Pro 3D printer (Type A Machines) for automated ppSPE fractionation. Manual prefractionation was performed using a top-loaded syringe and the automated prefractionation was carried out with a customized positive pressure solid phase extraction (ppSPE) workstation (Tecan Freedom Evo®) with two robotic arms working together to process 44 extracts in parallel. A single ppSPE run could accommodate up to 88 natural product extracts and generate 616 fractions in 3.5 h. A controlled rate of elution (