Finding Spirocyclic and Strongly sp3 Cores as Scaffold and R-group Replacements

We show how Spark™ can focus searches on sp3 rich fragments to achieve compound designs with improved properties
SHARE

Introduction

In 1900, Baeyer established the nomenclature of spirocycles, a chemical moiety that features a quaternary carbon atom between two rings1. The first spirocyclic drug, Griseofulvin, an antifungal agent, was approved in 19592. Since then, spirocyclic moieties have often been seen in drug design, such as CRGP inhibitor Ubrogepant3, DNA gyrase inhibitor ETX09144 and newly discovered pan-CoV inhibitor ASAP-00174455 (Figure 1).

Figure 1. The structures of Griseofulvin (1), Ubrogepant (2) and ETX0914 (3), represented in 3D in Flare™. The quaternary spiro carbon is shown in cyan.

Recently, spirocycles have been increasingly investigated in drug design programs, as they can provide numerous benefits when designing molecules6. They can be used as bioisosteres of common heterocycles to modulate properties such as lipophilicity and pKa. Due to their three-dimensionality, they can increase solubility by reducing crystal packing. Metabolic stability issues can be alleviated by the inclusion of additional steric hindrance near the site of metabolic susceptibility7. As spirocycles feature numerous sp3 centers, they can provide new exit vectors to grow into hard-to-reach pockets. With all these qualities in mind, integrating spirocycles into compounds is a potentially lucrative design route.

Fraction of sp3 carbon atoms, Fsp3, is a common metric to measure the amount of 3D complexity a molecule possesses which, as outlined in the previous paragraph, can benefit drug discovery8. Whilst this can be done by introducing spirocycles to a molecule, sp3 rich bioisosteric replacements of phenyl groups are also a popular route of compound optimization. Replacing these flat rings with a more three dimensional substituent can introduce similar benefits as introducing a spirocyclic core9. Examples of the inclusion of sp3 rich moieties to generate new vectors for interactions and rigidification to increase activity can be seen in BCL-XL inhibitor A-133185210 and autotaxin inhibitor ONO-843050611 (Figure 2).

Figure 2. A-1331852 (4) introduced an adamantane (cyan) to generate highly productive interactions with a key binding pocket of BCL-XL. ONO-8430506 (5) introduced a bicyclo[2.2.1]heptane (cyan) to rigidify the molecule, increasing its plasma activity.

Cresset’s bioisostere replacement tool, Spark™, can suggest sp3 rich design elements to introduce these benefits. By focusing the search of libraries with the correct SMARTS filters, Spark will suggest moieties with high globularity that closely match the shape similarity and 3D similarity (represented by Cresset’s field points to model electrostatics, Van der Waal’s and hydrophobics12) of the input molecule.

Spirocycles from bicycles: an illustrative example

Ikeda et al. sought to find how spiro derivatives of Monoacylglycerol Lipase (MAGL) inhibitors might improve compound properties13. Hit molecule 6 (Figure 3) was identified as a starting point owing to the presence of key His121 and Arg57 interactions with the bicyclic moiety. With a view to retaining these interactions, alternative groups were substituted in iteratively and tested for activity and lipophilic ligand efficiency. This culminated in the inclusion of spiro scaffolds that are novel bioisosteres of the 3-oxo-3,4-dihydro-2H-benzo[b][1,4]oxazin-6-yl moiety.

Figure 3. The hit ligand (6) from Ikeda et al., displayed with the two hydrogen bond interactions with His121 and Arg57 which were to be preserved. The portion of the molecule to be replaced is highlighted in cyan.

Spirocycles and aliphatic rings from phenyl: an illustrative example

Levterov et al. investigated 2-oxabicyclo[2.2.2]octane as a bioisostere for a para-substituted phenyl ring14. Phenyl rings are ubiquitous throughout natural products and bioactive compounds but can lead to undesirable properties, such as poor solubility.

Crystallographic analysis of 2-oxabicyclo[2.2.2]octanes showed very similar cross-ring distances to the phenyl ring in Imatinib (Figure 4), which is an issue with phenyl bioisosteres such as bicyclo[1.1.1]pentane. 2-oxabicyclo[2.2.2]octanes also closely retain the collinearity of exit vectors, maintaining molecular shape. The inclusion of the oxygen atom into the bicyclo[2.2.2]octane ring improves the solubility of the group.

When substituted into Imatinib, the 2-oxabicyclo[2.2.2]octane containing compound showed much improved solubility, a prolonged metabolic half-life, and generally improved ADME profile.

Figure 4. Imatinib (7), the para-phenyl ring is highlighted in cyan. The compound with its bioisosteric replacements are also shown (8, 9), with the phenyl bioisosteres also highlighted in cyan.

Discovering spirocyclic candidates with Spark

We aimed to explore these bioisosteric substitutions by performing two separate Spark experiments. Through inputting the starting molecules into Spark and then running searches focused upon spirocyclic and sp3 rich moieties, we aimed to find the replacements that the authors chose along with alternative valid options (Figure 5).

Figure 5. The Spark input for replacement of a phenyl ring in Imatinib. The portion of the molecule to be replaced, a para-phenyl ring in this case, has been selected. On the right-hand menu, we are allowing all options for the attachment points and are not defining any ring constraints.

Monoacylglycerol lipase Inhibitor replacing aromatic rings

For investigating Ikeda et al.’s study searches, where they aimed to find reversible MAGL inhibitors through the introduction of spirocyclic moieties, we kept the default Spark settings whilst introducing these SMARTS filters in the Advanced Filters tab:

Table 1. The SMARTS strings used to instruct Spark to only output replacement moieties that follow these patterns. These SMARTS used also allow searching of closely related compounds. As an example, the SMARTS shown for bicyclohexanes will also give structures for bicyclohexenes, bicyclohexones, diazabicyclohexanes, and so on.

ComponentSMARTS string
Spirocycles[X4;R2;r](@[r])(@[r])@[r]
AdamantanesA~1~A~2~A~A~3~A~A(~A~A~1~A~3)~A~2
CubanesA~1~2~A~3~A~4~A~1~A~5~A~2~A~3~A~4~5
BicyclooctanesA~1~2~A~A~A(~A~A~1)~A~A~2
BicyclohexanesA~1~2~A~A(~A~A~1)~A~2
BicyclopentanesA~1~2~A~A(~A~1)~A~2

Cresset’s Spark was able to identify bioisosteric replacements shown in this study and no custom libraries were needed for these searches; the sp3 rich fragments were available in the default ChEMBL and SureChEMBL databases distributed for use in Spark by Cresset (Figure 6).

Figure 6. The presence of the named substructures in the ChEMBL and SureChEMBL databases, using the SMARTS strings shown in Table 1.

Further treatment can help to optimize the results output from Spark. In the case of Ikeda et al., we performed Conf Hunt & Align on the results. For this, we selected the Quick Conf Hunt and MCS Alignment, using the input molecule as the reference.

Ikeda et al. note that hydrogen bonds to His121 and Arg 57 are essential for potent binding. Therefore, we used a script in the Python interpreter in Flare, to add a tag to results that make these essential contacts (please reach out to Cresset support if you would like this Python script). The tags could then be used to generate a list of compounds that maintained these essential interactions to focus on. This identified twelve results out of the five hundred generated, including the molecule the authors synthesized (Figure 7).

Figure 7. Our input molecule (6) with the R-group replacement highlighted in light cyan, overlaid with the Spark result showing Ikeda et al.’s molecule (10), with the replacement R-group in yellow.

Replacing phenyl ring with 2-oxabicyclo[2.2.2]octane in known drugs

For the results generated from Levterov et al.’s study14, whereby the effect of substituting in a 2-oxabicyclo[2.2.2]octane core in place of the phenyl ring in Imatinib and Vorinostat was studied, we added some additional constraints to the Spark run, alongside including the aforementioned SMARTS patterns. A key component of this study was to maintain the distance across, and exit vectors of, the bioisosteric replacements compared to the phenyl ring of the input molecule, so we edited Sparks advanced settings to reduce the geometric tolerance of the search. This ensured that the exit vectors of the output molecules spatially resembled those of the input phenyl ring. This was done by modifying the “Allowed attachment point distance error” and “Allowed attachment point angle errors” in the Advanced options of the Spark wizard to the lowest recommended settings (Figure 8).

Figure 8. The Advanced options in the Spark Scaffold Hopping or R-group Replacement Wizard. We have changed the “Allowed attachment point distance error” to 0.25 Å and the “Allowed attachment point angle error” to 5 deg. This will cause the Spark output to more closely mirror the physical distances and angle of the input moiety.

The output of this experiment gave us 59 scaffold replacement suggestions from the ChEMBL and SureChEMBL databases (Figure 9). This was filtered down to 51 suggestions by omitting those molecules that were flagged as possessing an unstable moiety. Both 2-oxabicyclo[2.2.2]octane and bicyclo[2.2.2]octane were present in these results, showing positive BIF% scores (bioisostere factor, a measure of how closely a bioisostere replacement reflects the input moiety).

Figure 9. The top 4 results output (11, 12, 13, 14) from the Spark experiment, alongside some associated scores and metrics. The replaced portions of the scaffolds are shown in cyan.

Conclusion

By utilizing SMARTS filters available in Spark, chemists can focus searches on sp3 rich fragments for bioisosteric ideas, exploiting the properties of these types of sp3 rich rings to achieve compound designs with improved properties. Here, we showed how Spark reproduced results from two investigative studies which aimed to introduce sp3 rich moieties to improve the properties of known active molecules.

Through judicious use of filters, Spark can provide ideas not just for hit generation, but also for lead optimization to improve compound qualities such as solubility and metabolic stability.

References

  1. Baeyer, A. Systematik Und Nomenclatur Bicyclischer Kohlenwasserstoffe. Berichte der deutschen chemischen Gesellschaft 1900, 33 (3), 3771–3775. https://doi.org/10.1002/cber.190003303187.
  2. Martin, A. R. For the Treatment of Animals Especially, by Oral Admin Istration, the Griseofulvin May First Be Incorporated into GRISEOFULVIN USES AND COMPOSITIONS Food Premixes and the Said Premixes Then Used for Incor.
  3. Moore, E.; Fraley, M. E.; Bell, I. M.; Burgey, C. S.; White, R. B.; Li, C.-C.; Regan, C. P.; Danziger, A.; Michener, M. S.; Hostetler, E.; Banerjee, P.; Salvatore, C. Characterization of Ubrogepant: A Potent and Selective Antagonist of the Human Calcitonin Gene‒Related Peptide Receptor. The Journal of Pharmacology and Experimental Therapeutics 2020, 373 (1), 160–166. https://doi.org/10.1124/jpet.119.261065.
  4. Basarab, G. S.; Doig, P.; Galullo, V.; Kern, G.; Kimzey, A.; Kutschke, A.; Newman, J. P.; Morningstar, M.; Mueller, J.; Otterson, L.; Vishwanathan, K.; Zhou, F.; Gowravaram, M. Discovery of Novel DNA Gyrase Inhibiting Spiropyrimidinetriones: Benzisoxazole Fusion with N-Linked Oxazolidinone Substituents Leading to a Clinical Candidate (ETX0914). J. Med. Chem. 2015, 58 (15), 6264–6282. https://doi.org/10.1021/acs.jmedchem.5b00863.
  5. Antiviral unveiled that goes after multiple coronaviruses. Chemical & Engineering News. https://cen.acs.org/acs-news/acs-meeting-news/Antiviral-unveiled-goes-multiple-coronaviruses/103/web/2025/03 (accessed 2025-03-31).
  6. Varela, M. T.; Dias, G. G.; de Oliveira, L. F. N.; de Oliveira, R. G.; Aguiar, F. D.; Nogueira, J. P.; Cruz, L. R.; Dias, L. C. Spirocyclic Compounds as Innovative Tools in Drug Discovery for Medicinal Chemists. European Journal of Medicinal Chemistry 2025, 287, 117368. https://doi.org/10.1016/j.ejmech.2025.117368.
  7. Griffith, D. A.; Kung, D. W.; Esler, W. P.; Amor, P. A.; Bagley, S. W.; Beysen, C.; Carvajal-Gonzalez, S.; Doran, S. D.; Limberakis, C.; Mathiowetz, A. M.; McPherson, K.; Price, D. A.; Ravussin, E.; Sonnenberg, G. E.; Southers, J. A.; Sweet, L. J.; Turner, S. M.; Vajdos, F. F. Decreasing the Rate of Metabolic Ketone Reduction in the Discovery of a Clinical Acetyl-CoA Carboxylase Inhibitor for the Treatment of Diabetes. J. Med. Chem. 2014, 57 (24), 10512–10526. https://doi.org/10.1021/jm5016022.
  8. Wei, W.; Cherukupalli, S.; Jing, L.; Liu, X.; Zhan, P. Fsp3: A New Parameter for Drug-Likeness. Drug Discovery Today 2020, 25 (10), 1839–1845. https://doi.org/10.1016/j.drudis.2020.07.017.
  9. Tsien, J.; Hu, C.; Merchant, R. R.; Qin, T. Three-Dimensional Saturated C(Sp3)-Rich Bioisosteres for Benzene. Nat Rev Chem 2024, 8 (8), 605–627. https://doi.org/10.1038/s41570-024-00623-0.
  10. Wang, L.; Doherty, G. A.; Judd, A. S.; Tao, Z.-F.; Hansen, T. M.; Frey, R. R.; Song, X.; Bruncko, M.; Kunzer, A. R.; Wang, X.; Wendt, M. D.; Flygare, J. A.; Catron, N. D.; Judge, R. A.; Park, C. H.; Shekhar, S.; Phillips, D. C.; Nimmer, P.; Smith, M. L.; Tahir, S. K.; Xiao, Y.; Xue, J.; Zhang, H.; Le, P. N.; Mitten, M. J.; Boghaert, E. R.; Gao, W.; Kovar, P.; Choo, E. F.; Diaz, D.; Fairbrother, W. J.; Elmore, S. W.; Sampath, D.; Leverson, J. D.; Souers, A. J. Discovery of A-1331852, a First-in-Class, Potent, and Orally-Bioavailable BCL-XL Inhibitor. ACS Med. Chem. Lett. 2020, 11 (10), 1829–1836. https://doi.org/10.1021/acsmedchemlett.9b00568.
  11. Iwaki, Y.; Ohhata, A.; Nakatani, S.; Hisaichi, K.; Okabe, Y.; Hiramatsu, A.; Watanabe, T.; Yamamoto, S.; Nishiyama, T.; Kobayashi, J.; Hirooka, Y.; Moriguchi, H.; Maeda, T.; Katoh, M.; Komichi, Y.; Ota, H.; Matsumura, N.; Okada, M.; Sugiyama, T.; Saga, H.; Imagawa, A. ONO-8430506: A Novel Autotaxin Inhibitor That Enhances the Antitumor Effect of Paclitaxel in a Breast Cancer Model. ACS Med. Chem. Lett. 2020, 11 (6), 1335–1341. https://doi.org/10.1021/acsmedchemlett.0c00200.
  12. Science overview | Cresset. https://cresset-group.com/science/overview/ (accessed 2025-07-01).
  13. Ikeda, S.; Sugiyama, H.; Tokuhara, H.; Murakami, M.; Nakamura, M.; Oguro, Y.; Aida, J.; Morishita, N.; Sogabe, S.; Dougan, D. R.; Gay, S. C.; Qin, L.; Arimura, N.; Takahashi, Y.; Sasaki, M.; Kamada, Y.; Aoyama, K.; Kimoto, K.; Kamata, M. Design and Synthesis of Novel Spiro Derivatives as Potent and Reversible Monoacylglycerol Lipase (MAGL) Inhibitors: Bioisosteric Transformation from 3-Oxo-3,4-Dihydro-2H-Benzo[b][1,4]Oxazin-6-Yl Moiety. J. Med. Chem. 2021, 64 (15), 11014–11044. https://doi.org/10.1021/acs.jmedchem.1c00432.
  14. Levterov, V. V.; Panasiuk, Y.; Sahun, K.; Stashkevych, O.; Badlo, V.; Shablykin, O.; Sadkova, I.; Bortnichuk, L.; Klymenko-Ulianov, O.; Holota, Y.; Lachmann, L.; Borysko, P.; Horbatok, K.; Bodenchuk, I.; Bas, Y.; Dudenko, D.; Mykhailiuk, P. K. 2-Oxabicyclo[2.2.2]Octane as a New Bioisostere of the Phenyl Ring. Nat Commun 2023, 14 (1), 5608. https://doi.org/10.1038/s41467-023-41298-3.

Related Science Resources

Application of generative AI to design covalent inhibitors of prolyl oligopeptidase
Generative chemistry is revolutionizing drug discovery, using AI to efficiently explore entirely new areas of chemical space in the search for...
Modality Agnostic Drug Discovery Strategies for Data Sparse Targets
This webinar in collaboration with Avammune, focuses on pursuing novel drug discovery targets with sparse or fragmented data...

Subscribe & Don't Miss Out

Receive our newsletter to be among the first to hear about product releases, case studies, opinion articles, events and more.