Glycine transporter GlyT1 is the main regulator of neuronal excitation and inhibition mediated by neurotransmitter glycine, which is a co-agonist of the N-methyl-D-aspartate (NMDA) receptor at excitatory synapses and a direct agonist of inhibitory glycine receptors. GlyT1 inhibition prolongs glycinergic signalling, and it has been extensively studied over the past two decades in the search for an effective treatment of schizophrenia, a severe, chronic mental illness associated with impaired NMDA receptor function¹. Although, clinical studies found several potent and selective GlyT1 inhibitors that achieved antipsychotic and pro-cognitive activity for the treatment of schizophrenia, a successful drug candidate has yet to emerge. In this study, we investigate one of the most advanced GlyT1 inhibitors that however failed in a phase III clinical trial study². To understand why potent inhibitors of GlyT1 have fallen short to show efficacy in late stages of clinical studies and elucidate mechanism of inhibition, we set out to determine the structure of GlyT1 with a benzoylpiperazine chemotype inhibitor from Roche library of compounds³. Catching the transporter in a clinically relevant inhibition-state conformation can help re-evaluate the efforts for developing selective inhibitors and provide insight in finding new strategies to target glycine reuptake transporter.

Wild-type human GlyT1 is unstable under experimental conditions and contains unstructured termini and a large, flexible extracellular loop 2 (EL2). To enable structure determination, we performed several iterations of protein engineering to identify the optimal protein construct, featuring shortest boundaries of GlyT1 with acceptable expression levels (minimal construct), and single-point mutations to increase thermal stability of the transporter. In the next step of construct design, we combined thermostabilizing mutations and shortened the flexible extracellular loop. To increase the hydrophilic surface area of the transporter and improve the chance of crystal contact formation, we generated constructs where a soluble fusion protein was inserted at different locations of the transporter, e.g. N terminus, or EL2, or both at N terminus and EL2. A total number of 154 constructs were generated in these rounds of construct design over the course of five years.

Using lipid-based crystallization methods, HiLiDe⁴ and LCP⁵, we began our crystallization trials back in 2015 and with the minimal GlyT1 construct containing a single point mutation. However, we only started to get diffracting protein crystals in the later iterations of construct designs inserting a fusion protein and combining thermostabilizing mutations with a shortened EL2. The crystals, however, did not diffract better than 7.5 Å, exhausting any attempts for further optimization. In the winter of 2017, the team designed Sybody® molecule Sb_GlyT1#7 (Linkster Therapeutics’ Sybody® technology)⁶ against a novel cell surface-exposed binding site of the transporter with the goal to address its inhibition-state. Just a year later in spring of 2018, Sybody® Sb_GlyT1#7 enabled crystallization out of one of 960 conditions used to derive well-diffracting protein crystals. However, crystal size was unexpectedly small even after optimization with the longest dimension measuring 2–5 µm. The focus was then put on a Serial Synchrotron Crystallography (SSX) approach that enabled the collection of diffraction data from hundreds of < 5 µm³ microcrystals, that under normal circumstances would not provide any usable data set as individual crystals. A few hundreds partial mini data sets were collected on the state-of-the-art P14 beamline operated by EMBL Hamburg at the PETRA III storage ring (DESY, Hamburg). This is a very bright source with a microfocus beam putting a total photon flux of 1.3 × 10¹³ photons/second at the sample position. It took almost two years to develop and implement a successful SSX data collection strategy that finally yielded a complete data set at 3.4 Å resolution on the morning of October 5th 2019. Merging of various partial mini data sets combining the best correlating ones to a complete high-quality data set succeeded using an in-house developed script, Ctrl-d.

The inhibitor-bound GlyT1 complex captures the transporter in an inward-open state and reveals a unique mode of binding for neurotransmitter transporters. The structure unravels how a bulky inhibitor lodges between transmembrane helices in the middle of the transporter and extends into the intracellular release pathway for ions and substrate. The non-competitive mechanism of action occurs through shifting the conformational equilibrium from outward-open to inward-open by stabilizing the inward-open conformation of GlyT1 associated with release of ions and substrate. The sybody Sb_GlyT1#7 is also highly selective for the inhibited, inward-open conformation of GlyT1 and represents an alternative approach to small molecule inhibitors of GlyT1. The structure provides blueprints for the rational design of new small molecule inhibitors and antibodies for targeting the glycine reuptake transporter.

Working on challenging research questions, such as our adventurous project, always requires a diverse team of people with different scientific backgrounds, and we as well combined expertise of research groups of Poul Nissen (Aarhus University), Roger J. P. Dawson (back then in Roche, now Linkster Therapeutics), Thomas R. Schneider (EMBL-Hamburg), Markus A. Seeger (University of Zurich), and Steffen Sinning (Dept. of Forensic Medicine) to tackle this research challenge.

We would like to say thanks to all that have contributed, colleagues and friends from all the nodes of our team for fruitful discussions and support.

Read more about our study “Structural insights into the inhibition of glycine reuptake”.

References

1. Harvey, R. J. & Yee, B. K. Glycine transporters as novel therapeutic targets in schizophrenia, alcohol dependence and pain. Nat. Rev. Drug Discov. 12, 866–885 (2013).

2. Pinard, E., Borroni, E., Koerner, A., Umbricht, D. & Alberati, D. Glycine transporter type I (GlyT1) inhibitor, bitopertin: a journey from lab to patient. CHIMIA Int. J. Chem. 72, 477–484 (2018).

3. Pinard, E. et al. Discovery of benzoylisoindolines as a novel class of potent, selective and orally active GlyT1 inhibitors. Bioorg. Med. Chem. Lett. 20, 6960–6965 (2010).

4. Gourdon, P. et al. HiLiDe—Systematic Approach to Membrane Protein Crystallization in Lipid and Detergent. Cryst. Growth Des. 11, 2098-2106, (2011).

5. Caffrey, M. & Cherezov, V. Crystallizing membrane proteins using lipidic mesophases. Nat Protoc 4, 706-731, (2009).

6. Zimmermann, I. et al. Synthetic single domain antibodies for the conformational trapping of membrane proteins. eLife 7, e34317 (2018).