Hybrid Multimodal Imaging Synthons for Chemoselective and Efficient Biomolecule Modification with Chelator and Near-Infrared Fluorescent Cyanine Dye

The development of hybrid multimodal imaging synthons (MIS), carrying in addition to a chelator for radiometal labeling also a near-infrared (NIR) fluorescent cyanine dye was the aim of this work. The MIS should be introducible into biomolecules of choice via an efficient and chemoselective click chemistry reaction. After chemical optimization, a successful synthetic strategy towards such hybrid MIS was developed, based on solid phase-based synthesis techniques and applying different near-infrared fluorescent cyanine dyes. The developed hybrid agents were shown to be easily introducible into a model homobivalent peptidic gastrin-releasing peptide receptor- (GRPR)-specific carrier without forming any side products and the MIS as well as their bioconjugates were radiolabeled with the positron-emitter 68Ga3+. The hybrid multimodal agents were characterized with regard to their logDs, GRPR target affinities and photophysical characteristics. It could be shown that the properties of the bioconjugates were not per se affected by the introduction of the MIS but that the cyanine dye used and specifically the number of comprised negative charges per dye molecule can have a considerable influence on target receptor binding. Thus, the molecular toolbox described here enables the synthesis of tailored hybrid multimodal imaging synthons for biomolecule modification, meeting the specific need and envisioned application of the combined imaging agent.


Additional Information Regarding the Initial, Suboptimal Synthetic Attempts towards the Target Multimodal Imaging Synthons
In the first attempt, we used a Fmoc-Cys(Trt)-loaded Wang resin to conjugate first, Fmoc-Lys(Mtt)-OH and second (after Nε-Mtt-deprotection under mild acidic conditions), the protected chelator NODA-GA(tBu)3 in this Nε-position, followed by conjugation of the respective 2-(1H-benzotriazol-1-yl)-1,1,3,3tetramethyluronium hexafluorophosphate (HBTU)-activated near-infrared (NIR) dye (Scheme S1A). However, the target molecules could not be obtained after acidic cleavage from the resin and deprotection. Only compounds of low molecular weight not carrying the dye moieties were obtained.
As we assumed that this problem resulted from the dye conjugation on resin, we repeated the synthesis using standard reaction conditions (as detailed in the Materials and Methods section) but also varied solvent, coupling agent and reaction time (DMF→DMSO or NMP, HBTU→1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU) or Benzotriazole-1-yl-oxy-tris-pyrrolidino-phosphonium hexafluorophosphate (PyBOP), 30 min standard coupling time → up to 16h reaction time). We intended to cleave the intermediate 4 from the resin and to perform the conjugation of the NIR dyes in solution (Scheme S1B). Intermediate 4 could, however, not be obtained after cleavage from the resin and only products of lower molecular weight were isolated.
We hypothesized that this phenomenon was caused by the Lys-Nε-Mtt-deprotection and chelator conjugation in this position on the resin. Hence, we first synthesized the NODA-GA-lysine-conjugate Fmoc-Lys(NODA-GA(tBu)3)-OH (5) in solution and used it for the following solid-supported synthesis of 4 (Scheme S1C). Unfortunately, we were still not able to isolate 4 but again only products of lower molecular weight, which were not further characterized. This was unexpected as this synthetic approach comprised only one conjugation and one deprotection/cleavage step, and usually proceeds with high efficiency. It is thus unclear what were the reasons for the observed synthetic difficulties. Scheme S1. Schematic depiction of the unsuccessful synthetic strategies towards the synthesis of the target multimodal imaging synthons (MIS). At first, standard reaction conditions were used (reaction in DMF for 30 min using 4 eq. of amino acid, 3.9 eq. of 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU) as coupling reagent and 4 eq. of N,N-diisopropylethylamine (DIPEA) as base, Fmoc protecting groups were removed using 50% (v/v) piperidine in DMF within seven min and the products were cleaved from the solid support using a mixture of trifluoroacetic acid (TFA) : TIS triisopropylsilane (TIS) : H2O (95:2.5:2.5) for 60 min) but also other solvents (DMSO, NMP), coupling agents (1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate: HATU, Benzotriazole-1-yl-oxy-tris-pyrrolidino-phosphonium hexafluorophosphate: PyBOP) and prolonged reaction times (up to 16h) were tested, giving the same results.
The reasons for these negative results are not obvious. We changed the cysteine-comprising solid support to a thio-ethanol amine-modified cysteamine 2-chlorotrityl resin, liberating the free thiol to see if this influences the synthesis outcome. Similar to before, no product formation could be observed. After further replacement of the resin with a highly acid-sensitive cysteamine super acid sensitive resin (SASRIN) resin and further change of the position of the chelator introduction from the Lys-Nε side chain functionality to the Nα amino functionality, intermediate 6 could be obtained in at least moderate yields of 23%. Then, 6 could be further reacted with different activated model fluorescent dyes (5(6)carboxyfluorescein-pfp ester [1], coumarin 343-pfp ester (7), dansyl chloride, and Py1 [2]), producing the respective model chelator-dye-conjugates 8a-d in yields of 14-24% (Scheme S2). Scheme S2. Schematic depiction of the synthetic pathway producing different model MIS comprising derivatives of fluorescein, coumarin, dansyl and pyridinium dyes (8a-d).
The relatively low yields observed for the second synthesis step of 6 with the activated model dyes to 8a-d following this route can be attributable to the presence of a second unprotected functional group (-NH2 and -SH) being able to react with the activated dyes. Furthermore, the free thiol functionalities of the molecules 6 and 8a-d can result in the formation of -S-S-bridges under neutral or slightly basic conditions, resulting in the formation of further side products.
To circumvent these problems and to determine if the complete synthesis of the multimodal imaging synthons (MIS) can be performed on resin, we further modified the synthesis route. Using coumarin 343 as a model, we studied the possibility to first conjugate the dye in its pfp ester form to the free Lys-Nε amino functionality of Fmoc-Lys-OH to obtain Fmoc-Lys(coumarin 343)-OH (9) and to directly use this amino acid derivative in the synthesis of 8b on solid support (Scheme S3). This procedure omits the cleavage of intermediate 6 from the resin and its purification prior to dye conjugation and should furthermore reduce the number of possible side products significantly. Indeed, this approach gave 8b somewhat higher product yields of 24% instead of 15%. However, this relatively small increase in product yields does not seem to justify the additional synthetic efforts for the solutionphase preparation of the Fmoc-Lys(dye)-OH conjugates such as 9.
Thus, we tried another possibility to introduce the NIR dyes into the Lys-Nε side chain amino functionality of the lysine building block, omitting a Mtt-deprotection of this functionality on resin by mild acids (as this might result in a loss of material to due a cleavage of the whole synthon from the solid phase) and not requiring the synthesis of the respective Fmoc-Lys(dye)-OH conjugates.
Details of this synthesis pathway, finally yielding the target MIS, are provided in the manuscript.