High-Affinity and Proteolytically Stable Peptidic Fluorescent NTS₁R Ligands

Abstract

Labeled ligands for the neurotensin receptor 1 (NTS₁R), which is expressed in the CNS, the gastrointestinal tract, and in malignant tumors, are needed to investigate NTS₁R-ligand binding and NTS₁R expression. Aiming for fluorescence-labeled neurotensin(8–13)-derived NTS₁R ligands with high affinity and proteolytic stability, several previous approaches were combined: (1) replacement of Arg⁸ by an amino-functionalized carbamoylated arginine, allowing conjugation to a fluorophore, (2) N ^α-methylation of Arg⁸ and replacement of Tyr by β,β-dimethyl-l-Tyr¹¹, conferring proteolytic stability, and (3) replacement of Leu¹³ by trimethylsilyl-Ala, boosting binding affinity. This strategy gave fluorescent NTS₁R ligands with unprecedented NTS₁R binding affinity (5-TAMRA-labeled ligand 19: K _i 0.14 nM, sulfo-Cy5-labeled probe 21: K _i 0.094 nM) and high stability in human plasma (t _1/2 ≫ 48 h). Their suitability for competition binding studies (flow cytometry; 19, 21) and the imaging of NTS₁R expression in living cells (confocal microscopy, biomolecular imaging; 19, 21) and tumor tissue (biomolecular imaging; 21) is demonstrated.

Introduction

The neurotensin receptor 1 (NTS₁R), a class A G-protein coupled receptor (GPCR), is one of three receptors of the neurotensin receptor family. The hexapeptide neuromedin N and the tridecapeptide neurotensin (NT) represent endogenous NTS₁R agonists (Figure A). The bioactivity of neurotensin is conferred by the C-terminal hexapeptide sequence, NT(8–13). Thus, NT(8–13) has served as a lead structure for the development of various NTS₁R ligands including radio- and fluorescence labeled analogs. In healthy tissues, the NTS₁R is mainly expressed in the brain and gastrointestinal tract. , With regard to neoplastic diseases, the NTS₁R is (over)­expressed, e.g., in pancreatic adenocarcinoma, prostate and colon carcinoma, and in about one-third of primary breast cancer tumors. − Therefore, the NTS₁R is considered a promising target for tumor diagnostics and treatment using metabolically stable, appropriately radiolabeled NTS₁R ligands.

1.

(A) Peptide sequences and NTS₁R binding affinities of the endogenous NTS₁R ligands neuromedin N and neurotensin, and the lead compound NT(8–13). (B) Reported fluorescent NTS₁R ligands with binding affinities in the low nanomolar range. Fluorescent dyes are shown in red. ^a Skrzydelski et al., ^b Granier et al., ^c Keller et al., ^d Keller et al., ^e Ertl et al., ^f Renard et al., ^g Maes et al.

Besides radioligands, fluorescently labeled NTS₁R ligands also represent useful tool compounds which can substitute radioligands, e.g., in competition binding assays needed to determine binding affinities of new NTS₁R ligands. Advantages of fluorescent probes over radioligands are mild safety issues, unproblematic waste disposal, and accessibility to a broad range of methods such as flow cytometry, high-content imaging, fluorescence anisotropy or NanoBRET, most of them being compatible with multimode plate reader measurements. Moreover, fluorescent probes can be used to study receptor–ligand binding by fluorescence microscopy. Unlike radiochemical assays, most fluorescence-based techniques allow for measurement under homogeneous conditions, meaning that a separation of bound from free ligand is not required.

To date, several fluorescent NTS₁R ligands, all derived from NT(8–13), have been reported (e.g., 1–4, Figure B). These probes were characterized by high-content imaging, flow cytometry, fluorescence anisotropy, NanoBRET or fluorescence microscopy. − Among the reported fluorescently labeled NTS₁R ligands, compounds 1–3 exhibit the highest binding affinities (K _i = 0.76–1.1 nM) allowing an application as probes for cell-based binding assays.

Yet, there is interest in fluorescently labeled NTS₁R ligands with even higher binding affinity since this comes along with reduced nonspecific binding and potentially with longer residence time, which is particularly advantageous for fluorescence-based imaging of receptor expression, e.g., in tumor tissue.

The design of the novel fluorescently labeled NT(8–13)-derivatives in this study was guided by different reported studies. First, Leu¹³ is replaced by (trimethylsilyl)­alanine (TMSAla) in the synthesized NT(8–13) analogs because Fanelli and coworkers reported that this structural modification leads to markedly increased NTS₁R binding affinity of hexapeptides derived from NT(8–13) (see compounds 5 and 6, Figure ).

2.

(A) Schematic presentation of reported amino acid replacements in NT(8–13) or NT(8–13) derivatives useful to increase NTS₁R binding affinity (replacement of Leu¹³ by TMSAla), for peptide conjugation (replacement of Arg⁸ by an amino-functionalized carbamoylated arginine), and to achieve proteolytic stability (methylation of the N-terminus and replacement of Tyr¹¹ by β,β-dimethyl-l-Tyr). (B) Aim of the present study: combination of approaches shown in A in one molecule followed by the synthesis of fluorescently labeled NTS₁R ligands. ^a Vivancos et al., ^b Fanelli et al., ^c Keller et al., ^d Schindler et al.

Second, Arg⁸ is bioisosterically exchanged by an amino-functionalized N ^ω-carbamoylated arginine (Arg­(carb)) (e.g., compound 7, Figure ). Attachment of fluorescent dyes to the side chain of this carbamoylated arginine was previously demonstrated to be well tolerated with respect to NTS₁R binding. , Third, the N-terminus is methylated and Tyr¹¹ is replaced by β,β-dimethyltyrosine, in order to obtain proteolytically stable peptides. The N-terminal methylation was reported to prevent N-terminal degradation by proteases and the incorporation of β,β-dimethyl-Tyr was recently shown to result in an effective stabilization of the C-terminus against proteolytic cleavage (see compounds 8 and 9, Figure ). Notably, the replacement of Tyr¹¹ by β,β-dimethyl-Tyr represents a favorable alternative to the replacement of Ile¹² by α-tert-butylglycine (standard modification for the stabilization of the C-terminus), since NTS₁R binding is affected by the latter but not by the former modification.

For fluorescence labeling of the peptides, the rhodamine-based dye 5-TAMRA and the indolinium-type cyanine dye sulfo-Cy5 were used. The precursor compounds and the labeled peptides were characterized in a radiochemical NTS₁R competition binding assay. The labeled probes were also studied in a functional NTS₁R Ca²⁺ assay and were photophysically characterized (fluorescent quantum yields and fluorescence lifetimes). Moreover, NTS₁R binding of the fluorescent ligands was investigated by flow cytometry (saturation binding, kinetic studies, and competition binding), confocal microscopy, and in tumor sections using a biomolecular imager.

Results and Discussion

Synthesis

The unnatural, Fmoc-protected amino acids (10–13) used for the peptide synthesis are shown in Scheme A. N ^α-methylated Fmoc/Pbf-protected arginine (11) and racemic Fmoc/tBu-protected β,β-dimethyl-Tyr (12) were commercially available (note: the desired l-configured stereoisomer of 12 was not available as a pure enantiomer). The Fmoc/Boc-protected amino-functionalized, carbamoylated arginine (13) was prepared as reported. The Fmoc-protected TMSAla (10) was synthesized according to a published procedure yielding (R)-10 with a reported ee of 98%. Here, this asymmetric synthesis, using (1R,2R,5R)-2-hydroxy-3-pinanone (23) as chiral auxiliary (see Scheme S1), gave (R)-10 with an ee of 86% (determined with capillary electrophoresis, electropherograms shown in the Supporting Information).

1. Structures of the Unnatural Amino acids 10–13, Sequences of the Synthesized Peptides 6 and 14–17b, and Exemplary Schematic Presentation of the Syntheses of 14 and 17 .

^a (A) Structures of used unnatural amino acids. (B) Amino acid sequences of the synthesized NT(8–13) derivatives. Overall yields for SPPS: 13–52%. (C) SPPS shown for 14 and 17. Reagents and conditions: (a) anchoring of 10 to a 2-ClTrt resin: 10/DIPEA (1/2.5 equiv), CH₂Cl₂, rt, overnight, Fmoc deprotection: 20% piperidine in DMF/NMP 8:2 v/v, rt, 2 × 10 min; (b) amino acid coupling: Fmoc-amino acid/HOBt/HBTU/DIPEA (5/5/4.9/10 equiv), DMF/NMP 8:2 v/v, 35 °C, 2 × 45 min (“double” coupling), Fmoc deprotection: as under (a); (c) (1) TFA/CH₂Cl₂ 1:3 v/v, rt, 2 × 20 min; (2) TFA/H₂O 95:5 v/v, rt, 5 h; (d) 12 or 13/HOBt/HBTU/DIPEA (3/3/2.95/6 equiv), DMF/NMP 8:2 v/v, 35 °C, 16 h (“single” coupling), Fmoc deprotection: as under (a); (e) N-terminal methylation according to reported procedures: , (1) 2-nitrobenzenesulfonylchloride/collidine (3/5 equiv), CH₂Cl₂, rt, 2 h; (2) MTBD/methyl-4-nitrobenzenesulfonate (4/5 equiv), DMF, rt, 30 min; (3) DBU/2-mercaptoethanol (5/10 equiv), DMF, rt, 30 min. The separation of the diastereomers 17a and 17b was performed by achiral preparative RP-HPLC. The wavy bond in 17a/b indicates either configuration, l or d.

All peptides (6, 14–17b; sequences shown in Scheme B) were synthesized by solid-phase peptide synthesis (SPPS) on a 2-ClTrt resin. Peptide 14 represents the NT(8–13) derivative containing TMSAla instead of Leu¹³. To our best knowledge, this peptide has not been described to date. Peptide 15 represents the N-terminally methylated congener of 14 and in 16 Tyr¹¹ is additionally replaced by β,β-dimethyl-Tyr (16a: β,β-dimethyl-l-Tyr, 16b: β,β-dimethyl-d-Tyr). Finally, 17, representing the precursor for fluorescence labeling, contains the N ^α-methylated, amino-functionalized N ^ω-carbamoylated arginine in position 8, β,β-dimethyl-l-Tyr (17a) or β,β-dimethyl-d-Tyr (17b) in position 11, and TMSAla in position 13 (note: the amino acid positions in the synthesized hexapeptides were assigned with the numbers corresponding to the respective amino acid positions in NT(8–13)). For 14 and 17, the solid-phase synthesis is exemplarily illustrated in Scheme C. Regarding the separation of the diastereomers 16a and 16b as well as 17a and 17b, this could be conveniently achieved by achiral preparative RP-HPLC. The assignment of the absolute configuration to β,β-dimethyl-Tyr in 16a–17b was carried out as recently reported for structurally related NT(8–13) derivatives. This approach is based on NT(8–13) analogs also containing β,β-dimethyl-l-Tyr or β,β-dimethyl-d-Tyr in position 11, for which the configuration was assigned using CD spectroscopy (structures shown in Figure S1A). As described by Ertl et al., the elution order in RP-HPLC and the NTS₁R binding affinities served as criteria for the assignment. Derivatives containing β,β-dimethyl-l-Tyr consistently elute first (RP-HPLC) and show markedly higher NTS₁R binding affinities compared to the analogs containing β,β-dimethyl-d-Tyr (see Figure S1).

The fluorescently labeled peptides 19 and 21 were obtained by treatment of the precursor 17a with the succinimidyl ester of the 5-TAMRA dye (18) or sulfo-Cy5 dye (20) in the presence of a base (Scheme ). The fluorophores were chosen based on different criteria. Both dyes are commercially available at low to moderate costs and filter sets matching the excitation and emission wavelengths of these dyes are available for most instruments used for fluorescence detection. A favorable feature of the 5-TAMRA dye is its high photostability, whereas the sulfo-Cy5 dye is excitable at higher wavelengths (>600 nm) which is advantageous with respect to an application at cells and tissues. Moreover, both dyes contain a negatively charged acid function conveying increased polarity and solubility.

2. Synthesis of the Fluorescently Labeled Peptides 19 and 21 .

^a Reagents and conditions: (a) DIPEA, DMF/NMP 8:2 v/v, rt, 1 h; 40% (19), 30% (21).

Chemical and Proteolytic Stability

The chemical stability of the fluorescently labeled ligands 19 and 21 was studied in phosphate-buffered saline (PBS, pH = 7.4) at room temperature over 48 h. Both compounds showed excellent stability (Figures S2 and S3). In the case of 21, adsorption of the fluorescent ligand to the vessel (visible by eye) resulted in a moderate decrease (ca. 33%) in peptide concentration over 48 h (cf. Figure S3). The adsorption as the cause for the decrease in concentration is also supported by the fact that no additional peak was observed in the RP-HPLC chromatograms. The proteolytic stability of 15, 17a, 19, and 21 was investigated in human plasma at 37 °C over 48 h. NT(8–13), studied as a control, showed very low plasma stability (complete degradation after 30 min, Table ). The NT(8–13) analog 15, being N-terminally methylated and containing TMSAla instead of Leu¹³ (cf. Scheme B), showed moderate proteolytic stability (ca. 50% intact peptide after 48 h). In contrast, the labeling precursor 17a and the fluorescent ligands 19 and 21, containing β,β-dimethyl-l-Tyr in position 11, exhibited high proteolytic stability (Table ). The apparent low degradation of peptide 21 (10% after 48 h) can be attributed to the adsorption of 21 to the vessel material as also observed in the case of stability studies in PBS (discussed above). These results show that replacement of Leu¹³ by TMSAla as the only C-terminal modification in the NT(8–13) sequence is not sufficient to achieve high proteolytic stability. Consequently, the results confirm that N-terminal methylation in combination with the incorporation of β,β-dimethyl-Tyr in position 11 is highly effective in terms of proteolytic stability of NT(8–13) derivatives.

1. Stabilities of NT(8–13), 15, 17a, 19, and 21 in Human Plasma/PBS 1:2 v/v (37 °C).

	% intact compound in plasma after the specified incubation times
compd.	10 min	30 min	2 h	6 h	24 h	48 h
NT(8–13)	17 ± 1	<1	<1	<1	n.d.	n.d.
15	n.d.	n.d.	83 ± 1	80 ± 1	65 ± 1	47 ± 1
17a	n.d.	n.d.	>99	>99	>99	>99
19	n.d.	n.d.	>99	95 ± 2	93 ± 2	97 ± 1
21	n.d.	n.d.	97 ± 1	93 ± 1	90 ± 1	90 ± 2

^aThe initial concentration of the peptide in human plasma/PBS (1:2 v/v) was 80 μM. Data represent mean values ± SEM from two or three independent experiments (SEM not given when no degradation was observed).

Neurotensin Receptor Binding Studied in Radiochemical Competition Binding Assays

NTS₁R binding affinities were determined in a radioligand competition binding assay as reported using intact HT-29 colon carcinoma cells which express NTS₁R endogenously but no NTS₂R. The tritium-labeled NT(8–13) analog [³H]­UR-MK300 (structure shown in Figure S4) was used as radioligand. The obtained K _i values of 6, 14-17b, 19, and 21 are presented in Table and competition binding curves are shown in Figure S5A. The previously described peptide 6, containing two lysines, exhibited high NTS₁R binding affinity with a K _i value of 0.055 nM (Table ) confirming the reported high binding affinity (IC₅₀ = 0.02 nM). Peptide 14, containing TMSAla instead of Leu¹³ as the only structural difference compared to NT(8–13), displayed similar binding affinity as 6. This confirmed the former observation that replacement of Arg⁸ and Arg⁹ by Lys⁸ and Lys⁹ does not affect NTS₁R binding. ,− N-methylation of 14 (peptide 15) had no impact on NTS₁R binding as shown in Table . This finding is consistent with a previously reported N ^α-methyl scan of NT(8–13). Replacement of Tyr¹¹ by β,β-dimethyl-l-Tyr in 15, resulting in 16a, did also not affect binding to the NTS₁R (K _i of 16a: 0.050 nM). Likewise, replacement of Arg⁸ in 16a by the amino-functionalized N ^ω-carbamoylated arginine (peptide 17a) did not influence the NTS₁R binding affinity (K _i of 17a: 0.055 nM). As stated before, the peptides containing β,β-dimethyl-d-Tyr (16b and 17b) showed markedly lower binding affinities (Table ). The fluorescently labeled peptides 19 and 21, bearing a bulky fluorescent dye at the side chain of Arg⁸, exhibited only slightly higher K _i values (K _i = 0.14 and 0.094 nM, respectively) compared to the precursor 17a. This confirms that conjugation of NT(8–13) derivatives via an N ^ω-carbamoylated arginine, incorporated in position 8, is a favorable approach in terms of preserving high NTS₁R binding affinity. As a consequence, 19 and 21 bind to the NTS₁R with similar binding affinity to NT(8–13) (K _i = 0.093 nM).

2. NTS₁R Binding Affinities of NT(8–13), 6, 14-17b, 19, and 21, and NTS₁R Agonistic Potencies and NTS₂R Binding Affinities of 19 and 21 .

	NTS1R		NTS2R
compd.	pK i ± SEM/K i [nM]	pEC50 ± SEM/EC50 [nM]	pK i ± SEM/K i [nM]
NT(8–13)	10.03 ± 0.03/0.093	8.96 ± 0.07/1.2	9.23/0.62
6	10.28 ± 0.06/0.055	n.d.	n.d.
14	10.10 ± 0.02/0.079	n.d.	n.d.
15	10.43 ± 0.1/0.041	n.d.	n.d.
16a	10.30 ± 0.05/0.050	n.d.	n.d.
16b	7.91 ± 0.06/12	n.d.	n.d.
17a	10.39 ± 0.2/0.055	n.d.	n.d.
17b	8.42 ± 0.1/4.2	n.d.	n.d.
19	9.86 ± 0.09/0.14	9.00 ± 0.09/1.1	9.11 ± 0.02/0.78
21	10.04 ± 0.06/0.094	8.89 ± 0.09/1.4	9.14 ± 0.07/0.75

^aDetermined by radioligand competition binding with [³H]­UR-MK300 at HT-29 colon carcinoma cells.

^bDetermined in a Fura-2 Ca²⁺ assay at HT-29 cells.

^cDetermined by radioligand competition binding with [³H]­UR-MK300 at membranes of HEK293T-hNTS₂R cells. Data represent mean values ± SEM (pK _i, pEC₅₀) or mean values (K _i, EC₅₀) from at least three independent experiments performed in triplicate. n.d. not determined.

In addition to NTS₁R binding, NTS₂R binding affinities were determined for the fluorescent ligands 19 and 21 in a radiochemical competition binding assay using membranes of HEK293T-hNTS₂R cells and [³H]­UR-MK300 as radioligand (competition binding curves see Figure S5B). With pK _i values of 9.11 and 9.14, respectively, both probes bind with slightly lower affinity to NTS₂R than to NTS₁R (Table ).

NTS₁R Agonism

The agonistic potencies of NT(8–13) and the fluorescently labeled NTS₁R ligands 19 and 21 were determined in a Fura-2 Ca²⁺ assay using HT-29 cells (concentration–response curves shown in Figure S6). Peptides 19 and 21 proved to be full agonists with potencies comparable to that of NT(8–13) (Table ). The pEC₅₀ values were consistently circa one log unit lower than the corresponding pK _i values, which can be explained by the nonequilibrium conditions in the case of the Fura-2 Ca²⁺ assay: as the signal occurs immediately after addition of the agonist, the receptor occupancy at the time of the readout is considerably lower than the hypothetical receptor occupancy that would be achieved in equilibrium for the respective concentrations of receptor and receptor ligand. An impairment of the optical readout of the assay by the 5-TAMRA label, whose absorption spectrum (λ_max = 555–558 nm, cf. Table ) largely overlaps with the emission spectrum of the Fura-2 Ca²⁺ complex (λ_max = 505 nm), can be excluded based on the results of previous studies using a 5-TAMRA dummy ligand.

3. Maxima of the Absorption, Excitation, and Emission Spectra, Fluorescence Lifetimes and Fluorescence Quantum Yields of 19 and 21 in PBS in the Presence or Absence of BSA (1% w/w).

	λabs/λex/λem [nm]|Δ[eV]		τfl [ns]		Φfl (%)
compd.	PBS	PBS + 1% BSA	PBS	PBS + 1% BSA	PBS	PBS + 1% BSA
19	558/552/583|0.12	555/553/578|0.097	2.7	3.5	41	34
21	647/648/665|0.049	650/651/669|0.051	1.4	2.7	29	39

^aCalculated based on λ_ex and λ_em.

^bOnly the main component of a biexponential fit with a contribution >90% is presented.

Photophysical Characterization

The absorption of the first electronic transition and the emission spectrum of both compounds are their corresponding mirror images indicating a rather constrained molecular scaffold both in the electronic ground and first excited state (Figure S7A,C). However, in the presence of BSA (1% w/w) the spectral distributions shift. In the case of 19, the spectra blue-shift, while in the case of 21 the spectra red-shift. This already indicates an interaction with the protein’s surface. Further, while in the case of 21 the excitation spectrum always resembles the corresponding absorption spectrum with or without BSA, the excitation spectrum of 19 recorded in the absence of BSA already shows a blueshift compared to the corresponding absorption spectrum. This indicates a significant structural rearrangement of the molecular framework in the excited state compared to the ground state. The impact of BSA in form of interactions to the chromophore is further manifested in altered fluorescence quantum yields and lifetimes (Table ). Each chromophoric system shows a prolonged excited singlet state lifetime in the presence of BSA. To note, the emission decay is biexponential and the second lifetime is of small amplitude (cf. Figure S7B,D), which may arise either from small impurities or an equilibrium between different conformers. Interestingly, while 19 shows a decreased fluorescence quantum yield, 21 shows an increased fluorescence quantum yield. In the case of 19, this may indicate a deactivation process induced by the interaction with the protein. An electron transfer from a redox active amino acid to the chromophore or the hindrance for structural changes in the excited state, which could result in a brighter emissive configuration, may represent two potential explanations. This will be addressed by transient absorption spectroscopy and presented elsewhere. In the case of 21, the emissive character of the electronically excited chromophore is apparently inversely effected. Here, the data suggest a situation in which, in the excited state, the molecule undergoes structural changes forming a less emissive configuration that is hindered when attached to the protein’s surface. Also, in this case the transient absorption will inform on the mechanistic scenario in more detail, which will be presented elsewhere. The observed changes in quantum yield are in accordance with reported studies on 5-TAMRA-labeled and sulfo-Cy5-labeled fluorescent GPCR ligands. ,−

Investigation of NTS₁R Binding of 19 and 21 by Flow Cytometry

The flow cytometric binding experiments were carried out at 23 °C using a BD FACSCanto II cytometer equipped with an argon laser (488 nm) and a red laser (640 nm). Saturation binding experiments with 19 and 21 were performed at live human HT-29 colon carcinoma cells and stably transfected CHO-hNTS₁R cells. For both ligands, specific binding was saturable (Figures A and S8), however, the apparent K _d values determined at CHO-hNTS₁R cells were higher (10–20-fold) compared to the apparent K _d values obtained from HT-29 cells (Table ) which were in good agreement with the K _i values determined by competition binding with [³H]­UR-MK300 at HT-29 cells (Table ). The higher K _d values determined at CHO-hNTS₁R cells can be explained by the markedly higher NTS₁R expression in CHO-hNTS₁R cells (ca. 300,000 receptors/cell) compared to HT-29 cells (ca. 45,000 receptors/cell). As the studied ligands 19 and 21 show subnanomolar binding affinities, a high receptor concentration in saturation binding experiments leads to ligand depletion and consequently results in higher apparent K _d values. This effect can be reduced by using less cells in the experiment. Typically, for this kind of flow cytometric analysis, cell densities ≥100,000 cells/mL are used to obtain a reasonable number of gated events (≥2000), required for the calculation of representative fluorescence intensity mean values, within a reasonable measuring time. To explore the influence of the receptor concentration and associated with this the effect of ligand depletion on the discrepancy between the apparent K _d and the true K _d, additional saturation binding experiments were performed with CHO-hNTS₁R cells using a considerably lower cell density of 15,000 cells/mL. These conditions resulted in 200–300 gated events, just high enough to enable a statistically reasonable data analysis. The apparent K _d values of 19 and 21 obtained from these saturation binding experiments were markedly lower (factor > 10) than the apparent K _d values obtained from saturation binding studies using 150,000 cells/mL (Table and Figure S8) and similar to the K _d values determined at HT-29 cells. The apparent K _d values determined at low receptor concentrations (HT-29 cells, CHO-hNTS₁R cells at a density of 15,000 cells/mL) should better resemble the true K _d, although the latter is potentially even lower than the experimentally determined dissociation constants.

3.

NTS₁R equilibrium binding and kinetic studies of 19 and 21 determined by flow cytometry at intact HT-29 cells. (A) Representative binding isotherms (specific binding, open symbols) of 19 and 21 obtained from saturation binding experiments (incubation: 120 min at 23 °C). Nonspecific binding was determined in the presence of 1 μM NT(8–13). K _d values are presented in Table . Data represent mean values ± SEM (total and nonspecific binding) or calculated values ± propagated error (specific binding) from a representative experiment performed in triplicate. (B) Association and dissociation of 19 and 21 studied at 23 °C. Concentrations of 19 and 21 used for the association: 1 nM and 0.1 nM, respectively; concentrations of 19 and 21 used for the preincubation period (19: 120 min, 21: 180 min) of dissociation experiments: 1 nM and 0.25 nM, respectively. Dissociation and association rate constants are presented in Table . Data represent mean values ± SEM from at least three independent experiments performed in triplicate.

4. NTS₁R Binding Data of 19 and 21 Determined by Flow Cytometry at HT-29 and CHO-hNTS₁R Cells.

	saturation binding		binding kinetics (HT-29 cells)
	apparent K d [nM]
compd.	HT-29 cells	CHO-hNTS1R cells	k obs [min–1]	k on [nM–1min–1]	k off [min–1]	K d(kin) [nM]
19	0.11 ± 0.04	1.2 ± 0.3	k obs,mono = 0.039 ± 0.005	k on,mono = 0.017 ± 0.01	0.021 ± 0.007	1.2 ± 1.3
		0.034 ± 0.009	k obs(bi,fast) = 0.19 ± 0.03	k on(bi,fast) = 0.16 ± 0.04		0.13 ± 0.07
		0.88 ± 0.1	k obs(bi,slow) = 0.016 ± 0.003	k on(bi,slow) = −0.0055 ± 0.01		n.a.
21	0.046 ± 0.02	0.022 ± 0.005	0.014 ± 0.002	0.11 ± 0.02	0.0033 ± 0.0002	0.030 ± 0.007

^aAssociation rate constant ± propagated error calculated from k _off, k _obs, and the ligand concentration used for the association experiments.

^bKinetically derived dissociation constant ± propagated error calculated from k _off and k _on values. K _d, k _obs, and k _off values represent mean values ± SEM from at least three independent experiments; k _on and K _d(kin) values represent calculated values ± propagated error.

^cExperiments performed with a cell density of 150,000 cells/mL.

^dExperiments performed with a cell density of 15,000 cells/mL. n.a. not applicable.

Noteworthy, the excitation laser line (640 nm) used for the excitation of the sulfo-Cy5 dye in 21 is close to its absorption maximum (647/650 nm, see Table ). In contrast, the excitation of the 5-TAMRA label in 19 by the 488 nm argon laser is not ideal with respect to its absorption maximum (558/555 nm, Table ). This is also reflected by the markedly lower B _max values obtained from saturation binding experiments with 19 compared to 21 (Figure A).

Kinetic binding studies, performed at HT-29 cells, revealed notable differences between 19 and 21. While 21 showed a monophasic association, the 5-TAMRA-labeled ligand 19 displayed a clear biphasic association (two-phase association favored over one-phase association according to the F-test, P < 0.005, GraphPad Prism 5) (Figure B). The proportion of the fast and slow association phase was 42:58 (based on the fit to the data shown in Figure B). The observed association rate constants k _obs and the calculated association rate constants k _on are shown in Table . The k _on(bi,slow) value, calculated from k _obs(bi,slow), k _off, and the concentration of 19 used for the association studies, was negative. This could be due to internalization of ligand–receptor complex during the association process (cf. Figure ), but could also be attributed to the low ratios of total over nonspecific binding in the case of 19 particularly during the initial phase of the association (discussed in more detail below). As the k _on(bi,slow) was negative, precluding a calculation of the kinetically derived dissociation constant K _d(kin), the association data of 19 were additionally fitted with a monophasic exponential fit covering the whole association process. This enabled the calculation of a K _d(kin) value which comprises the entire association (see below).

6.

Visualization of binding of 19 (2 nM) to intact HT-29 cells (temperature: 22 °C) by confocal microscopy. Shown is total binding (A, B), nonspecific binding (C), and autofluorescence (D). Nuclei were stained with H33342 (2 μg/mL). (A) Merged fluorescence of 19 (green) and nuclei (blue). (B) Fluorescence of 19, without nuclei. (C) Merged fluorescence of 19 and nuclei acquired in the presence of 1 μM NT(8–13).

The fact that one ligand shows a monophasic and the other ligand a biphasic association for the same cellular system suggests that the biphasic association cannot be explained by the existence of two subpopulations of NTS₁R (e.g., one coupled to and the other uncoupled from G-protein). It should be mentioned that for the association studies of 19, a fluorescent ligand concentration corresponding to 9-fold the K _d value of 19 had to be used since with lower ligand concentrations, correlating with lower receptor occupancies, the difference between total and nonspecific binding was too low for a reliable data analysis. This was particularly due to the aforementioned mismatch of the excitation wavelength and the absorption maximum of the 5-TAMRA dye in 19. In contrast, for association studies with 21, a ligand concentration corresponding to 3-fold the K _d value could be used because of the optimal excitation of the sulfo-Cy5 dye in 21 with the red laser (640 nm) resulting in the detection of high fluorescence intensities. It is noteworthy that, in the case of experiments with 19, the excitation with 488 nm light gave a high autofluorescence of the cells which also affected data analysis. The excitation of the cells with 640 nm (used for 21) resulted in very low autofluorescence. The different ligand concentrations used for the association experiments might be the reason for the observed differences in the association kinetics of 19 and 21 (biphasic vs monophasic). Moreover, it could be explained by the internalization of ligand–receptor complex (as confirmed by confocal microscopy, see below), occurring already during the association process. Upon internalization, ligand-dependent differences could occur with respect to changes of fluorescent properties in the cell interior and retrafficking of ligand–receptor complex or free receptor to the plasma membrane.

Dissociation studies at HT-29 cells, initiated by adding an excess of NT(8–13), gave monophasic dissociation curves revealing an incomplete dissociation (plateau significantly higher than zero, P < 0.05, t-test) as also previously reported for fluorescently labeled NT(8–13) derivatives. , The plateau values amounted to 77% (19) and 44% (21). This can be explained by the NTS₁R agonism of 19 and 21 (cf. Figure S6) meaning that the ligand–receptor complex is internalized, potentially followed by intracellular dissociation of the fluorescent ligand from the receptor. The k _off values of 19 and 21 differed by a factor of 7 (k _off = 0.021 min^–1 vs 0.0033 min^–1, see Table ). Compared to previously described fluorescently labeled NT(8–13) derivatives, 19 and 21 show a slower dissociation from NTS₁R which is favorable for imaging of NTS₁R expression in cells or tissues. The kinetically derived dissociation constants K _d(kin) (Table ) were calculated according to K _d(kin) = k _off/k _on. In the case of 19, the K _d(kin) was calculated from k _on,mono and k _on(bi,fast) yielding K _d(kin) values of 1.2 and 0.13 nM, respectively. The value of 0.13 nM is in excellent agreement with the K _d obtained from saturation binding (Table ). Likewise, the K _d(kin) value obtained for 21 is in good agreement with the K _d from saturation binding studies (0.030 nM vs 0.046 nM), indicating a binding process largely following the law of mass action. Yet, the determined kinetic parameters of 19 and 21 must be interpreted with care since both ligands induce receptor endocytosis (cf. Figures and ) and show an incomplete dissociation. Moreover, as discussed before, flow cytometric kinetic studies with the 5-TAMRA-labeled ligand 19 were compromised by the nonoptimal excitation and higher autofluorescence compared to 21.

7.

Visualization of binding of 21 (2 nM) to intact HT-29 cells (temperature: 22 °C) by confocal microscopy. Shown is total binding (A, B), nonspecific binding (C), and autofluorescence (D). Nuclei were stained with H33342 (2 μg/mL). (A) Merged fluorescence of 21 (red) and nuclei (blue). (B) Fluorescence of 21, without nuclei. (C) Merged fluorescence of 21 and nuclei acquired in the presence of 1 μM NT(8–13). Note: as no nonspecific binding of 21 was observed, images of nonspecific binding were only acquired after 45 min.

To investigate the suitability of 19 and 21 to serve as probes for the determination of NTS₁R binding affinities of unlabeled NTS₁R ligands, the K _i values of reported NTS₁R ligands (NT­(8-13), SR142948, SR48692, UR-MK300) were determined in flow cytometric competition binding studies at HT-29 cells using 19 or 21 as labeled probe (displacement curves shown in Figure ).

4.

Displacement curves from flow cytometric competition binding studies using 19 (A) or 21 (B) as a labeled probe and NT(8–13), SR142948, SR48692, or UR-MK300 as competitors. Experiments were performed at intact HT-29 cells at 23 °C. The concentrations of 19 and 21 were 0.3 nM and 0.1 nM, respectively. Incubation times: 120 min (A) and 180 min (B). Data represent mean values ± SEM from at least three individual experiments performed in triplicate.

Although the dissociation kinetics of 19 and 21, indicating long-lasting binding, is not ideal in terms of competition binding experiments, the pK _i values obtained for the reference ligands were generally in good agreement with reported data (Table ).

5. NTS₁R Binding Affinities of Reported NTS₁R Ligands Determined by Flow Cytometry at HT-29 Cells Using 19 and 21 as Labeled Probes .

	pK i
compound	19	21	literature
NT(8–13)	10.20 ± 0.2	10.25 ± 0.09	9.85, 9.62, 9.54, 9.00, 9.55, 10.09
SR142948	9.70 ± 0.1	10.14 ± 0.08	8.96, 9.24, 9.74, 9.00
SR48692	8.25 ± 0.1	8.49 ± 0.2	8.07
UR-MK300	9.85 ± 0.1	10.05 ± 0.1	9.33

^aReported K _i values were converted to pK _i values. Data represent mean values ± SEM from at least three individual experiments performed in triplicate.

The Hill coefficients of the displacement curves (4-parameter logistic fit) were not significantly different from −1 (two-tailed t test, P > 0.05) except for SR142948 and SR48692 studied with 21 (Hill coefficients: −0.63 ± 0.02 and −0.55 ± 0.06, respectively; mean values ± SEM). The low slope factors might be attributed to the (pseudo)­irreversible binding of 21 (cf. Figure B) or to the internalization and putative externalization of NTS₁R induced by the fluorescent agonists, potentially resulting in a retarded displacement of fluorescent ligand. However, for the other competitors (NT(8–13), MK300) and for all competitors studied with 19, also showing (pseudo)­irreversible binding, the Hill coefficients were not significantly lower than unity, which contradicts the aforementioned hypothesis. Since slope factors different from unity could also arise from the existence of ligand-dependent multiple conformational receptor states, , the low slopes observed for SR142948 and SR48692 could be caused by the preference of different receptor conformations of the used NTS₁R ligands: whereas the agonist 21 prefers the active conformation, the antagonists do not discriminate between receptor states or even prefer the inactive conformation. ,

Cooperativity between 21 and the Allosteric NTS₁R Modulator SBI-553

In recent years, several allosteric NTS₁R modulators, targeting an intracellular binding site, have been reported. − In vivo studies suggested that allosteric potentiators of NTS₁R binding of neurotensin can potentially be used as therapeutics for the treatment of pain. A prominent example is the positive allosteric modulator SBI-553 which was demonstrated to potentiate binding of radiolabeled neurotensin or NT(8–13). , To explore if SBI-553 also enhances binding of fluorescently labeled NT(8–13) derivatives such as 21, the effect of SBI-553 on NTS₁R binding of 21 was studied by flow cytometry using HT-29 tumor cells. For these experiments, 21 was used at a concentration of 0.01 nM, corresponding to approximately one-fifth of its K _d, and these samples were titrated with SBI-553. As previously observed in the radiochemical assays, , SBI-553 clearly enhanced NTS₁R binding of 21 (Figure ). However, in contrast to the reported studies, a biphasic CEC of SBI-553 was obtained, yielding a pEC₅₀(high) of 10.37 and a pEC₅₀(low) of 7.28. The latter is in excellent agreement with the reported pEC₅₀ value of SBI-553 of 7.30 and with the described EC₅₀ value of 140 nM. To note, in contrast to the pEC₅₀(low), the pEC₅₀(high) was not well reproducible in the individual experiments which is reflected by the high SEM of 0.40 (cf. Figure ). The E _max observed for SBI-553 (E _max = 88%) was lower than the reported E _max of 140% and 242%.

5.

(A) Structure of the allosteric NTS₁R modulator SBI-553. (B) Effect of SBI-553 on NTS₁R binding of the fluorescent NT(8–13) derivative 21 (c = 0.01 nM) determined in a flow cytometric binding assay performed with intact HT-29 cells at 23 °C. Cells were preincubated with SBI-553 for 30 min prior to the addition of 21 and continued incubation for 180 min. SBI-553 enhanced binding of 21 in a biphasic manner (fraction high/low: 0.33/0.67). 100% specifically bound 21 corresponds to specific binding of 21 in the absence of SBI-553. Data represent mean values ± SEM from six individual experiments performed in triplicate. pEC₅₀(high) = 10.37 ± 0.40. pEC₅₀(low) = 7.28 ± 0.16. E _max = 88 ± 8% (difference between lower and upper plateau).

It is a matter of speculation why SBI-553 shows a biphasic CEC when studied with the fluorescent ligand 21. The major differences compared to the reported radiochemical assays , are the type of labeled ligand (radiolabeled vs fluorescently labeled), the use of intact cells (in contrast to membrane preparations in the reported studies), and a measurement under equilibrium conditions in the present study (flow cytometry). Supposedly, the biphasic course of the CEC of SBI-553 can be attributed to one or more of these differences in the technical setup of the binding assays. The elucidation of the underlying mechanisms requires further studies. Likewise, the lower E _max compared to the reported E _max values , could also arise from the different assay parameters.

Confocal Microscopy

Binding of 19 and 21 to both HT-29 cells, natively expressing NTS₁R, and stably transfected CHO-hNTS₁R cells, overexpressing the NTS₁R, was studied by confocal microscopy at 22 °C using a confocal laser scanning microscope. In all cases, a marked difference between total and nonspecific binding was found (Figures and ; Figures S9 and S10), showing that the NTS₁R expression can be visualized with both fluorescent ligands, even at low receptor expression levels (HT-29 cells, Figures and ).

The excitation with 561 nm (19) resulted in higher autofluorescence compared to an excitation with 633 nm (21). This was only observable when using HT-29 cells (cf. Figures D and D) because the lower receptor expression in these cells compared to CHO-hNTS₁R cells (cf. Figures S9D and S10D) necessitates higher laser power and higher detector gain (see Experimental Section). As becomes obvious from Figures C and C, showing nonspecific binding of 19 and 21, respectively, the 5-TAMRA-labeled ligand 19 exhibited considerably higher nonspecific binding compared to the sulfo-Cy5-labeled probe 21.

The microscopic studies with 19 and 21 revealed that the NTS₁R is internalized by endocytosis upon binding of the fluorescent ligands, occurring already during the association process (Figures and ; Figures S9 and S10). The intracellular fluorescent ligand appeared to be located in vesicles. The internalization rates were estimated by defining an outer ROI, representing the total cellular fluorescence, and an inner ROI, representing the intracellular fluorescence, and quantification of fluorescence (Figure S11). This analysis suggests that the fraction of internalized ligand only slightly increases over the studied time (5–60 min). This could be explained by recycling of NTS₁R to the plasma membrane after internalization as also suggested by previous studies of fluorescently labeled NT(8–13) analogs. It should be noted that the determined internalization rate represents only a rough estimation due to the visual definition of the ROIs and potentially biased fluorescence intensities caused by changes in the photophysical properties of the fluorescent ligands in the intracellular environment.

Biomolecular Imaging of NTS₁R Expression in Cells and Tumor Tissue

In another approach, the suitability of 19 and 21 to serve as probes for the investigation of NTS₁R expression in living cells and tumor tissue was studied using an Azure Sapphire Biomolecular Imager. For these experiments, HT-29 cells, showing a low NTS₁R expression, and HT-29 tumors subcutaneously grown in nude mice were used. Saturation binding experiments with the sulfo-Cy5-labeled ligand 21 at adherent HT-29 cells gave satisfactory fluorescence images revealing low nonspecific binding of 21 (Figure A). Based on the fluorescence intensities, saturation binding curves could be generated (Figure B) yielding a K _d value of 1.7 ± 1 nM (mean value ± SEM from four individual experiments performed in triplicate), which was higher than the K _d value obtained from flow cytometric saturation binding experiments at suspended HT-29 cells (cf. Table ).

8.

(A) Fluorescence image of a 96-well plate with adherent HT-29 cells acquired with an Azure Sapphire Biomolecular Imager (laser: 658 nm; pixel size: 20 μm) 2 h after incubation with 21 at rt. Nonspecific binding was determined in the presence of SR142948 (1 μM). Shortly before image acquisition, cells were washed three times with cold buffer. (B) Fluorescence intensities from (A) plotted against the concentration of 21. Specific binding was calculated from total and nonspecific binding. Data represent means ± SEM (total and nonspecific binding) or calculated values ± propagated error (specific binding) from four independent experiments performed in triplicate.

This discrepancy could be explained by a lower sensitivity of the biomolecular imaging technique compared to the flow cytometric analysis, meaning that cell-bound fluorescence is underestimated for low ligand concentrations in the case of biomolecular imaging. Moreover, a higher adsorption of 21 to the plate material (polystyrene) and the required washing step in the case of the biomolecular imaging could account for the higher K _d value compared to the flow cytometric experiment which uses polypropylene plates and allows a measurement under equilibrium conditions (no washing required).

In contrast to 21, the fluorescence images obtained from the same experiment performed with the 5-TAMRA-labeled ligand 19 were not suited for the determination of a K _d value due to high autofluorescence and high nonspecific binding (Figure S12).

Using the same biomolecular imager, visualization of NTS₁R expression in HT-29 tumors was explored by imaging 10 μm cryosections of tumor tissue after incubation with 19 and 21. As observed for the saturation binding experiment with HT-29 cells, a detection of NTS₁R expression in the tumor was possible with the red-emitting probe 21 (Figure ), but failed when using the 5-TAMRA-labeled ligand 19 as probe mainly due to high autofluorescence relative to specific binding (Figure S13). The regions of the tumor sections showing high nonspecific binding of 21 (Figure ) likely represent necrotic tumor tissue. It should be noted that the HT-29 tumors originated from a former research project and had been stored for more than eight years at −80 °C. This long-term storage might have led to partial unfolding or degradation of the receptor protein.

9.

Fluorescence images of cryosections (size: 11–28 mm², thickness: 10 μm) of a HT-29 tumor acquired after incubation with 21 (5 nM) at rt for 60 min with an Azure Sapphire FL Biomolecular Imager (laser: 658 nm; pixel size: 20 μm). Nonspecific binding was determined in the presence of SR142948 (5 μM). Blank samples were incubated with neat buffer to determine background signals. Shortly before image acquisition, tumor sections were washed three times with cold buffer.

Conclusion

N-terminal methylation and the replacement of Arg⁸, Tyr¹¹, and Leu¹³ in NT(8–13) by an amino-functionalized carbamoylated arginine, β,β-dimethyl-l-Tyr, and TMSAla, respectively, resulted in a NT(8–13) analog (17a, UR-FE083-I) that was used as a precursor for the preparation of a 5-TAMRA-labeled (19, UR-FE093) and a sulfo-Cy5-labeled (21, UR-FE094) NTS₁R ligand. Both fluorescent ligands show subnanomolar NTS₁R binding affinities with K _i values of 0.14 nM and 0.094 nM. The high binding affinity allows an application of these probes for flow cytometric binding assays using cells with low NTS₁R expression levels, which was demonstrated for nontransfected HT-29 cells natively expressing the NTS₁R.

The fluorescent peptides 19 and 21 display high in vitro proteolytic stability (t _1/2 ≫ 48 h, human plasma), which is favorable with respect to binding studies in living cells potentially featuring proteolytic activities. 19 and 21 are anticipated to exhibit also high stability in vivo since recently described NT(8–13)-derived PET ligands, containing an N-terminal methyl group and β,β-dimethyl-l-Tyr in position 11, showed high in vivo stability in mice.

Fluorescence imaging of receptor expression in tissues represents an attractive alternative to autoradiography that requires radiolabeled receptor ligands. The present study reveals that the 5-TAMRA-labeled ligand 19 is not suited for fluorescence-based imaging of the NTS₁R expression in tumor tissue. This shows that the 5-TAMRA label, although characteristic of high photostability, does not meet the demands for this kind of imaging, in particular due to the excitation wavelength of <600 nm, resulting in interfering autofluorescence. Contrarily, the near-infrared sulfo-Cy5-labeled probe 21, differing from 19 only with respect to the fluorescent dye, turned out to be an excellent molecular tool for NTS₁R imaging in tumor tissue sections with high signal-to-noise ratio (used excitation: 658 nm), suggesting its application in tumor tissue screening procedures. As 19 and 21 display also high NTS₂R binding affinity, they can potentially serve as probes for NTS₂R binding studies using cellular systems solely expressing NTS₂R.

One of the major challenges with respect to the in vivo use of labeled peptidic NTS₁R ligands has been the development of tracers which exhibit both high proteolytic stability and high NTS₁R binding affinity (K _i < 1 nM). With the development of the amine-functionalized precursor 17a, which can be conjugated to various molecular labels, the present study paved the way to the future synthesis and use of high-affinity and proteolytically stable probes for the NTS₁R with a functionality of choice.

Experimental Section

Materials

The protected amino acids Fmoc-Tyr­(tBu)–OH, Fmoc-Arg­(Pbf)–OH, and glycine tert-butyl ester hydrochloride (22) were purchased from Carbolution Chemicals (St. Ingbert, Germany). Fmoc-N-Me-Arg­(Pbf)–OH (11), Fmoc-Lys­(Boc)–OH, Fmoc-Pro-OH, Fmoc-Ile-OH, and HBTU were from Merck (Darmstadt, Germany). Racemic Fmoc-β,β-dimethyl-Tyr­(tBu)–OH (12) and 2-ClTrt resin were obtained from Iris Biotech (Marktredwitz, Germany). HOBt, 7-methyl-1,5,7-triazabicyclo[4.4.0]­dec-5-ene (MTBD), methyl-4-nitrobenzenesulfonate, 2-mercaptoethanol, and 1-methyl-d-Trp were from Sigma-Aldrich (Taufkirchen, Germany). Collidine, 2-nitrobenzenesulfonyl chloride, 1,8-diazabicyclo[5.4.0]­undec-7-ene (DBU), NMP and DMF for peptide synthesis, anhydrous DMF and NMP, dichloromethane, piperidine, and TFA were purchased from Fisher Scientific (Schwerte, Germany). (1R,2R,5R)-2-Hydroxy-3-pinanone (23) and (iodomethyl)­trimethylsilane (25) were from TCI (Eschborn, Germany). DIPEA was obtained from ABCR (Karslruhe, Germany). Acetonitrile (HPLC gradient grade) was from VWR (Ismaning, Germany). NT(8–13) was synthesized via SPPS in-house. SR142948 and SR48692 were purchased from Tocris Bioscience (Bristol, UK). SBI-553 was from Biocat (Heidelberg, Germany). Bacitracin, HEPES, and bovine serum albumin (BSA) were obtained from Serva (Heidelberg, Germany). Fetal bovine serum (FBS) was purchased from Pan-Biotech (Aidenbach, Germany). Fura-2 AM and Pluronic F-127 were obtained from Calbiochem/Merck Biosciences (Beeston, UK). The radioligand [³H]­UR-MK300 (molar activity: 2.41 TBq/mmol) and its unlabeled analog UR-MK300 were prepared according to a described procedure. Compounds 10 and 13 were prepared according to the reported procedures. 5-TAMRA succinimidyl ester (18) was purchased from Carl Roth (Karlsruhe, Germany). Sulfo-Cyanine5 NHS ester (S0586, 20) was obtained from FEW Chemicals (Bitterfeld, Germany). Millipore water was consistently used for the preparation of stock solutions, buffers, and aqueous eluents for HPLC. Polypropylene reaction vessels with screw cap (1.5 or 2 mL) from Sarstedt (Nümbrecht, Germany) were used for small-scale reactions (e.g., activation of Fmoc-protected amino acids) and to keep stock solutions.

NMR Spectroscopy

NMR spectra were recorded on an AVANCE 600 instrument with cryogenic probe (¹H: 600 MHz; ¹³C: 150 MHz) and an AVANCE 400 instrument (¹H: 400 MHz; ¹³C: 100 MHz) (Bruker, Karlsruhe, Germany). NMR spectra were calibrated based on the solvent residual peaks (¹H NMR, DMSO-d ₆: δ = 2.50 ppm, CDCl₃: δ = 7.26 ppm; ¹³C NMR, DMSO-d ₆: δ = 39.50 ppm, CDCl₃: δ = 77.16 ppm) and data are reported as follows: ¹H NMR: chemical shift δ in ppm (multiplicity [s = singlet, d = doublet, t = triplet, m = multiplet, br s = broad singlet], integral, coupling constant J in Hz); ¹³C NMR: chemical shift δ in ppm.

Mass Spectrometry

High-resolution mass spectrometry (HRMS) was performed with an Agilent 6540 UHD accurate-mass Q-TOF LC/MS system coupled to an Agilent 1290 analytical HPLC system (Agilent Technologies. Santa Clara, CA) using an ESI source and the following LC method: column: Agilent Zorbax Eclipse Plus C18, 1.8 μm, 50 × 2.1 mm, column temperature: 40 °C, solvent/linear gradient: 0–4 min: 0.1% aqueous HCOOH/acetonitrile supplemented with 0.1% HCOOH 95:5–2:98, 4–5 min: 2:98, flow: 0.6 mL/min.

Preparative HPLC

Preparative HPLC was performed with a system from Knauer (Berlin, Germany) consisting of two K-1800 pumps and a K-2001 detector (in the following referred to as “system 1”) or with a Prep 150 LC system from Waters (Eschborn, Germany) comprising a Waters 2545 binary gradient module, a Waters 2489 UV/vis detector, and a Waters fraction collector III (in the following referred to as “system 2”). A Gemini NX-C18, 5 μm, 250 mm × 21 mm (Phenomenex, Aschaffenburg, Germany) was used as stationary phase at a flow rate of 20 mL/min using mixtures of 0.1% aqueous TFA and acetonitrile as the mobile phase. A detection wavelength of 220 nm was used throughout. Collected fractions were lyophilized using a Scanvac CoolSafe 100–9 freeze-dryer (Labogene, Allerød, Denmark) equipped with a RZ 6 rotary vane vacuum pump (Vacuubrand, Wertheim, Germany).

Analytical HPLC

Analytical HPLC analysis was performed with a system from Agilent Technologies composed of a 1290 Infinity binary pump equipped with a degasser, a 1290 Infinity autosampler, a 1290 Infinity thermostated column compartment, a 1260 Infinity diode array detector, and a 1260 Infinity fluorescence detector. A Kinetex-XB C18, 2.6 μm, 100 × 3 mm (Phenomenex) served as stationary phase at a flow rate of 0.6 mL/min. Detection was performed at 220 nm and the temperature of the column compartment was set to 25 °C. Mixtures of acetonitrile (A) and 0.04% aqueous TFA (B) were used as mobile phase. The following linear gradients were applied: compounds 6 and 14–17b: 0–14 min: A/B 10:90–30:70, 14–15 min: 30:70–95:5, 15–18 min: 95:5 (isocratic); compounds 10, 24, 26, and 27: 0–20 min: A/B 10:90–90:10, 20–21 min: 90:10–95:5, 21–25 min: 95:5 (isocratic); compounds 19 and 21: 0–14 min: A/B 20:80–40:60, 14–15 min: 40:60–95:5, 15–18 min: 95:5 (isocratic). The injection volume was 20 μL. Retention (capacity) factors k were calculated from the retention times t _R according to k = (t _R – t ₀)/t ₀ (t ₀ = dead time, 0.76 min for the used system and column).

General Procedure for Solid Phase Peptide Synthesis

Peptides were synthesized manually by SPPS on a 2-ClTrt resin according to the Fmoc strategy. 5 mL NORM-JECT syringes (B. Braun-Melsungen, Melsungen, Germany), equipped with a 35-μm polypropylene frit (Roland Vetter Laborbedarf, Ammerbuch, Germany), were used as reaction vessels. DMF/NMP 4:1 v/v was used as solvent for all amino acid coupling and Fmoc deprotection steps. Prior to the coupling of the first amino acid (10), the 2-ClTrt resin (loading: 1.60 mmol/g) was allowed to swell in CH₂Cl₂ for 30 min at rt followed by washing of the resin with CH₂Cl₂ (2×). Amino acid 10 was attached to the resin according to a reported procedure. The loading of the resulting Fmoc-TMSAla-2-ClTrt resin, which was used for the synthesis of peptides 6 and 14–17b, amounted to 1.00 mmol/g (determined photometrically via absorbance at 304 nm in DMF according to a reported protocol). Fmoc deprotection of the Fmoc-TMSAla-2-ClTrt resin was carried out using 20% piperidine in DMF/NMP 4:1 v/v (2× 10 min at rt) followed by washing with solvent (6× ca. 1 mL). The following Fmoc-amino acids (except for 11-13), used in 5-fold excess, were preactivated with HOBt/HBTU/DIPEA (5/4.9/10 equiv) in solvent (about 2.2 mL/mmol amino acid) for at least 5 min before addition to the resin. The Fmoc-protected unnatural amino acids 11-13 were used in 3-fold excess and were preactivated with HOBt/HBTU/DIPEA (3/3/6 equiv) in anhydrous solvent (about 1.6 mL/mmol amino acid) for 5–10 min prior to addition to the resin. Amino acid coupling was carried out on a shaker (Heidolph Multi Reax; Heidolph Instruments, Schwabach, Germany) covered with a thermostat controlled (35 °C) box. In the case of standard amino acids, “double” coupling (2× 45 min) was performed. 11–13 were attached by a single coupling procedure (35 °C, 16 h). After coupling of an Fmoc-amino acid, the resin was washed with solvent (4×) followed by Fmoc deprotection using 20% piperidine in solvent (2× 10 min at rt) and subsequent washing of the resin with solvent (6× ca. 1 mL). After coupling of the last amino acid and final Fmoc deprotection, the resin was washed with solvent (6×) and CH₂Cl₂ (3×) (treated with potassium carbonate). Peptides were cleaved off the resin using CH₂Cl₂/TFA 3:1 v/v (2× 20 min at rt). The liquids (2× ca. 2 mL) were collected in a 100 mL round-bottom flask and the resin was washed once with CH₂Cl₂/TFA 3:1 v/v (2 mL). The volatiles of the combined liquids were removed by evaporation, TFA/H₂O 95:5 v/v (2 mL per 100 mg resin) was added to the residue and the mixture was stirred at rt for 5 h. The volatiles were removed by evaporation followed by the addition of H₂O (ca. 50 mL) and lyophilization to obtain the crude peptide, which was subjected to purification by preparative HPLC.

Compound Characterization

Amino acid 7, peptides 14–17b, and compounds 24 and 26 were characterized by HRMS, ¹H-, ¹³C-, and 2D-NMR spectroscopy (2D: ¹H-COSY, HSQC, HMBC), and RP-HPLC. Compounds 19 and 21 were characterized by HRMS, ¹H NMR spectroscopy, and RP-HPLC. 6 and 27 were characterized by HRMS and RP-HPLC. HPLC purities of all target compounds were ≥97% (UV detection, 220 nm).

Experimental Protocols and Analytical Data

Lys-Lys-Pro-Tyr-Ile-β-trimethylsilyl-Ala Tris­(hydrotrifluoroacetate) (6)

Peptide 6 was synthesized on a Fmoc-TMSAla-2-ClTrt resin (50 mg, 1.00 mmol/g) according to the general procedure. Purification by preparative RP-HPLC (system 2, gradient: 0–30 min: acetonitrile/0.1% aqueous TFA 22:78–27:73, t _R = 7 min) yielded 6 as a white fluffy solid (29.4 mg, 52%). HRMS (ESI): m/z [M+3H]³⁺ calcd. for [C₃₈H₆₉N₈O₈Si]³⁺ 264.4997, found: 264.5001. RP-HPLC (220 nm): >99% (t _R = 8.6 min, k = 10.3). C₃₈H₆₆N₈O₈Si·C₆H₃F₉O₆ (794.10 + 342.07).

Arg-Arg-Pro-Tyr-Ile-β-trimethylsilyl-Ala Tris­(Hydrotrifluoroacetate) (14)

Peptide 14 was synthesized on a Fmoc-TMSAla-2-ClTrt resin (75 mg, 1.00 mmol/g) according to the general procedure. Purification by preparative RP-HPLC (system 1, gradient: 0–25 min: acetonitrile/0.1% aqueous TFA 15:85–45:55, t _R = 13 min) yielded 14 as a white fluffy solid (34.1 mg, 38%). ¹H NMR (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.76–0.85 (m, 6H), 0.91–0.98 (m, 1H), 0.98–1.08 (m, 2H), 1.35–1.44 (m, 1H), 1.44–1.62 (m, 5H), 1.64–1.76 (m, 4H), 1.77–1.90 (m, 3H), 1.93–2.03 (m, 1H), 2.62–2.70 (m, 1H), 2.80–2.93 (m, 1H), 3.03–3.12 (m, 4H), 3.54–3.61 (m, 2H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.81–3.84 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 4.18–4.26 (m, 2H), 4.31–4.36 (m, 1H), 4.37–4.43 (m, 1H), 4.44–4.53 (m, 1H), 6.59–6.64 (m, 2H), 6.84–7.28 (br s, 4H, interfering with next listed signal), 6.99–7.03 (m, 2H), 7.28–7.65 (br s, 4H), 7.73 (d, 1H, J 9.2 Hz), 7.81 (t, 1H, 5.5 Hz), 7.86 (t, 1H, J 6.2 Hz), 7.94 (d, 1H, J 8.1 Hz), 8.13–8.29 (m, 4H), 8.67 (d, 1H, J 7.3 Hz), 9.23 (s, 1H), 12.45 (s, 1H). ¹³C NMR (150 MHz, DMSO-d ₆): δ (ppm) 0.00 (3 carbon atoms), 10.89, 15.13, 19.28, 24.03, 24.09, 24.17, 24.61, 28.21, 28.42, 29.10, 36.43, 37.27, 40.16, 40.53, 46.85, 48.63, 50.46, 51.69, 54.28, 56.32, 59.15, 113.96 (TFA), 114.85 (2 carbon atoms), 115.93 (TFA), 117.90 (TFA), 119.88 (TFA), 127.80, 130.11 (2 carbon atoms), 155.79, 156.92, 156.94, 158.88 (q, J 32 Hz) (TFA), 168.31, 169.24, 170.54, 170.75, 171.37, 174.61. HRMS (ESI): m/z [M + 3H]³⁺ calcd for [C₃₈H₆₉N₁₂O₈Si]³⁺ 283.1705, found: 283.1716. RP-HPLC (220 nm): >99% (t _R = 9.0 min, k = 10.8). C₃₈H₆₆N₁₂O₈Si·C₆H₃F₉O₆ (847.11 + 342.07).

N ^α-Methyl-Arg-Arg-Pro-Tyr-Ile-β-trimethylsilyl-Ala Tris­(hydrotrifluoroacetate) (15)

Peptide 15 was synthesized on a Fmoc-TMSAla-2-ClTrt resin (75 mg, 1.00 mmol/g) according to the general procedure. Purification by preparative HPLC (system 1, gradient: 0–25 min: acetonitrile/0.1% aqueous TFA 15:85–45:55, t _R = 14 min) yielded 15 as a white fluffy solid (40.3 mg, 44%).¹H NMR (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.76–0.86 (m, 6H), 0.91–0.98 (m, 1H), 0.98–1.07 (m, 2H), 1.38–1.48 (m, 3H), 1.51–1.59 (m, 3H), 1.66–1.76 (m, 4H), 1.79–1.87 (m, 3H), 1.94–2.04 (m, 1H), 2.46–2.48 (m, 3H), 2.62–2.69 (m, 1H), 2.82–2.88 (m, 1H), 3.06–3.15 (m, 4H), 3.55–3.60 (m, 2H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.78–3.81 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 4.17–4.28 (m, 2H), 4.32–4.36 (m, 1H), 4.38–4.43 (m, 1H), 4.52–4.57 (m, 1H), 6.60–6.63 (m, 2H), 6.82–7.29 (br s, 4H, interfering with next listed signal), 7.00–7.02 (m, 2H), 7.29–7.65 (br s, 4H), 7.71–7.75 (m, 1H), 7.84–7.90 (m, 2H), 7.93 (d, 1H, J 8.1 Hz), 8.20 (d, 1H, J 7.9 Hz), 8.87 (d, 1H, J 7.3 Hz), 8.95 (s, 1H), 9.04–9.70 (m, 2H), 12.46 (s, 1H). ¹³C NMR (150 MHz, DMSO-d ₆): δ (ppm) 0.00 (3 carbon atoms), 10.89, 15.13, 19.28, 23.92, 24.03, 24.18, 24.62, 26.99, 28.12, 29.09, 31.16, 36.42, 37.27, 40.08, 40.44, 46.86, 48.63, 50.58, 54.27, 56.32, 59.16, 59.92, 113.94 (TFA), 114.85 (2 carbon atoms), 115.91 (TFA), 117.88 (TFA), 119.86 (TFA), 127.80, 130.11 (2 carbon atoms), 155.78, 156.93 (2 carbon atoms), 158.91 (q, J 33 Hz) (TFA), 167.00, 169.06, 170.54, 170.74, 171.33, 174.60. HRMS (ESI): m/z [M + 3H]³⁺ calcd for [C₃₉H₇₁N₁₂O₈Si]³⁺ 287.8424, found: 287.8434. RP-HPLC (220 nm): >99% (t _R = 9.2 min, k = 11.1). C₃₉H₆₈N₁₂O₈Si·C₆H₃F₉O₆ (861.13 + 342.07).

N ^α-Methyl-Arg-Arg-Pro-β,β-dimethyl-l-Tyr-Ile-β-trimethylsilyl-Ala Tris­(hydrotrifluoroacetate) (16a) and N ^α-Methyl-Arg-Arg-Pro-β,β-dimethyl-d-Tyr-Ile-β-trimethylsilyl-Ala Tris­(hydrotrifluoroacetate) (16b)

Peptides 16a and 16b, representing diastereomers, were synthesized on a Fmoc-TMSAla-2-ClTrt resin (50 mg, 1.00 mmol/g) according to the general procedure. The unavailability of enantiomerically pure Fmoc-β,β-dimethyl-l-Tyr­(tBu)–OH (this Fmoc amino acid was only available as racemic mixture) necessitated the synthesis of both diastereomers, which were separated by preparative HPLC (system 1, gradient: 0–25 min: acetonitrile/0.1% aqueous TFA 15:85–45:55, t _R (16a) = 14 min, t _R (16b) = 16 min). Lyophilization of the eluates yielded 16a and 16b as white fluffy solids (16a: 23.0 mg, 37%; 16b: 21.1 mg, 34%). ¹H NMR of 16a (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.73–0.85 (m, 6H), 0.90–1.04 (m, 3H), 1.19–1.30 (m, 6H), 1.33–1.42 (m, 1H), 1.42–1.50 (m, 2H), 1.50–1.61 (m, 3H), 1.61–1.84 (m, 7H), 1.84–1.95 (m, 1H), 2.48 (s, 3H), 3.04–3.17 (m, 4H), 3.48–3.53 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.55–3.59 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.82 (s, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 4.14 (t, 1H, J 8.4 Hz), 4.14–4.24 (m, 1H), 4.36–4.43 (m, 1H), 4.52–4.58 (m, 1H), 4.60–4.69 (m, 1H), 6.59–6.65 (m, 2H), 6.93–7.30 (br s, 4H, interfering with next listed signal), 7.11–7.14 (m, 2H), 7.30–7.81 (m, 6H), 7.84–7.94 (m, 2H), 8.08 (d, 1H, J 7.8 Hz), 8.92 (d, 1H, J 8.0 Hz), 8.95 (s, 1H), 9.04–9.41 (m, 2H), 12.42 (s, 1H). ¹³C NMR of 16a (150 MHz, DMSO-d ₆): δ (ppm) 0.00 (3 carbon atoms), 10.94, 15.13, 19.22, 23.92, 24.11, 24.14, 24.45, 24.55, 26.99, 27.09, 28.13, 28.63, 31.16, 37.08, 40.06, 40.16, 40.43, 46.86, 48.68, 50.51, 56.43, 59.25, 59.88, 59.91, 114.07 (TFA), 114.34 (2 carbon atoms), 116.05 (TFA), 118.02 (TFA), 120.00 (TFA), 127.42 (2 carbon atoms), 136.42, 155.26, 156.90, 156.95, 158.83 (q, J 32 Hz) (TFA), 167.05, 169.35, 169.42, 170.43, 170.56, 174.66. ¹H NMR of 16b (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.71–0.82 (m, 6H), 0.88–1.03 (m, 3H), 1.17–1.34 (m, 8H), 1.40–1.49 (m, 2H), 1.49–1.64 (m, 5H), 1.64–1.85 (m, 5H), 2.47 (s, 3H), 3.01–3.17 (m, 4H), 3.41–3.50 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.51–3.58 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.80 (s, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 4.10 (q, 1H, J 7.3 Hz), 4.13–4.19 (m, 1H), 4.35–4.42 (m, 1H), 4.47–4.54 (m, 1H), 4.87 (d, 1H, J 9.9 Hz), 6.57–6.67 (m, 2H), 6.74–7.26 (br s, 4H, interfering with next listed signal), 7.14–7.17 (m, 2H), 7.26–7.59 (br s, 4H), 7.62 (d, 1H, J 8.5 Hz), 7.69 (d, 1H, J 9.8 Hz), 7.79–7.90 (m, 2H), 8.19 (d, 1H, J 7.0 Hz), 8.85 (d, 1H, J 7.1 Hz), 8.94 (s, 1H), 9.00–9.56 (m, 2H), 12.43 (s, 1H). ¹³C NMR of 16b (150 MHz, DMSO-d ₆): δ (ppm) 0.00 (3 carbon atoms), 11.09, 15.26, 19.25, 23.84, 23.99, 24.01, 24.15, 24.55, 26.53, 26.97, 28.02, 29.40, 31.14, 37.09, 40.09, 40.43, 41.00, 46.71, 49.31, 50.60, 56.41, 59.15, 59.49, 59.86, 114.06 (TFA), 114.35 (2 carbon atoms), 116.03 (TFA), 118.02 (TFA), 120.00 (TFA), 127.29 (2 carbon atoms), 136.86, 155.38, 156.92 (2 carbon atoms), 158.74 (q, J 32 Hz) (TFA), 166.95, 168.97, 169.40, 170.17, 170.84, 174.48. HRMS (ESI): m/z [M + 3H]³⁺ calcd for [C₄₁H₇₅N₁₂O₈Si]³⁺ 297.1861, found: 297.1866 (16a), 297.1866 (16b). RP-HPLC (220 nm): 16a: 99% (t _R = 10.7 min, k = 13.1), 16b: 97% (t _R = 12.2 min, k = 15.1). C₄₁H₇₂N₁₂O₈Si·C₆H₃F₉O₆ (889.19 + 342.07).

N ^α-Methyl-N ^ω-[(4-aminobutyl)­aminocarbonyl]­Arg-Arg-Pro-β,β-dimethyl-l-Tyr-Ile-β-trimethylsilyl-Ala Tetrakis­(hydrotrifluoroacetate) (17a) and N ^α-Methyl-N ^ω-[(4-aminobutyl)­aminocarbonyl]­Arg-Arg-Pro-β,β-dimethyl-d-Tyr-Ile-β-trimethylsilyl-Ala Tetrakis­(hydrotrifluoroacetate) (17b)

Peptides 17a and 17b, representing diastereomers, were synthesized on a Fmoc-TMSAla-2-ClTrt resin (150 mg, 1.00 mmol/g) according to the general procedure with the following modification: after coupling of the last amino acid and Fmoc-deprotection, the resin was washed with CH₂Cl₂ (5×), a solution of 2-nitrobenzenesulfonyl chloride (100 mg, 0.45 mmol) and collidine (100 μL, 0.75 mmol) in CH₂Cl₂ (3.5 mL) was added and the mixture was shaken at rt for 2 h. The resin was washed with DMF (5×) and a solution of MTBD (87 μL, 0.6 mmol) and methyl-4-nitrobenzenesulfonate (163 mg, 0.75 mmol) in DMF (4.2 mL) was added followed by shaking at rt for 30 min. After washing the resin with DMF (3×), a solution of DBU (112 μL, 0.75 mmol) and 2-mercaptoethanol (104 μL, 1.5 mmol) in DMF (3.5 mL) was added and the mixture was shaken at rt for 30 min. The resin was washed with DMF (5×) followed by cleavage from the resin as described in the general procedure for SPPS. The unavailability of enantiomerically pure Fmoc-β,β-dimethyl-l-Tyr­(tBu)–OH (only available as racemic mixture) necessitated the synthesis of both diastereomers, which were separated by preparative RP-HPLC (system 1, gradient: 0–25 min: acetonitrile/0.1% aqueous TFA 15:85–45:55, t _R (17a) = 12 min, t _R (17b) = 14 min). Lyophilization of the eluates yielded 17a and 17b as white fluffy solids (17a: 30.1 mg, 14%; 17b: 27.8 mg, 13%). ¹H NMR of 17a (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.72–0.89 (m, 6H), 0.90–1.04 (m, 3H), 1.18–1.31 (m, 6H), 1.33–1.42 (m, 1H), 1.43–1.61 (m, 9H), 1.62–1.94 (m, 8H), 2.47 (s, 3H), 2.73–2.83 (m, 2H), 3.03–3.18 (m, 4H), 3.21–3.30 (m, 2H), 3.46–3.54 (m, 1H), 3.54–3.63 (m, 1H), 3.76–3.91 (m, 1H), 4.12 (t, 1H, J 7.1 Hz), 4.16–4.25 (m, 1H), 4.36–4.43 (m, 1H), 4.48–4.59 (m, 1H), 4.66 (d, 1H, J 9.6 Hz), 6.57–6.65 (m, 2H), 6.71–7.28 (br s, 2H, interfering with next listed signal), 7.11–7.14 (m, 2H), 7.28–7.70 (m, 5H), 7.71–7.99 (m, 4H), 8.06 (d, 1H, J 7.9 Hz), 8.52 (s, 2H), 8.68–9.15 (m, 3H), 9.17 (s, 1H), 10.78 (s, 1H), 12.41 (s, 1H). One exchangeable proton (NH) of the presumably 4-fold protonated molecule could not be identified. ¹³C NMR of 17a (150 MHz, DMSO-d ₆): δ (ppm) 0.00 (3 carbon atoms), 10.95, 15.14, 19.25, 23.43, 24.11, 24.16, 24.38, 24.54 (2 carbon atoms), 25.99, 27.00 (2 carbon atoms), 28.16, 28.64, 32.16, 37.08, 38.49, 38.64, 40.06, 40.14, 40.43, 46.86, 48.71, 50.50, 56.45, 59.25, 59.88 (2 carbon atoms), 114.12 (TFA), 114.35 (2 carbon atoms), 116.10 (TFA), 118.08 (TFA), 120.06 (TFA), 127.42 (2 carbon atoms), 136.40, 153.93 (2 carbon atoms), 155.27, 156.94, 158.90 (q, J 32 Hz) (TFA), 167.14, 169.32, 169.39, 170.42, 170.58, 174.69. ¹H NMR of 17b (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.70–0.82 (m, 6H), 0.87–1.02 (m, 3H), 1.19–1.35 (m, 8H), 1.44–1.65 (m, 11H), 1.65–1.83 (m, 5H), 2.47 (s, 3H), 2.75–2.83 (m, 2H), 3.04–3.16 (m, 4H), 3.21–3.30 (m, 2H), 3.43–3.50 (m, 1H), 3.50–3.58 (m, 1H), 3.81 (s, 1H), 4.11 (q, 1H, J 7.6 Hz), 4.15 (t, 1H, J 7.4 Hz), 4.36–4.41 (m, 1H), 4.49–4.55 (m, 1H), 4.86 (d, 1H, J 9.8 Hz), 6.57–6.65 (m, 2H), 6.73–7.32 (br s, 2H, interfering with next listed signal), 7.14–7.17 (m, 2H), 7.32–7.72 (m, 5H), 7.72–8.04 (m, 4H), 8.17 (d, 1H, J 7.0 Hz), 8.53 (s, 2H), 8.87 (d, 1H, J 7.3 Hz), 8.96 (s, 1H), 9.05–9.23 (m, 2H), 10.88 (s, 1H), 12.43 (s, 1H). One exchangeable proton (NH) of the presumably 4-fold protonated molecule could not be identified. ¹³C NMR of 17b (150 MHz, DMSO-d ₆): δ (ppm) 0.00 (3 carbon atoms), 11.09, 15.26, 19.26, 23.37, 23.99, 24.02, 24.20, 24.39, 24.55, 25.99, 26.48, 26.91, 28.10, 29.32, 31.14, 37.10, 38.49, 38.66, 40.11, 40.44, 40.98, 46.71, 49.26, 50.59, 56.39, 59.17, 59.55, 59.81, 114.07 (TFA), 114.37 (2 carbon atoms), 116.04 (TFA), 118.01 (TFA), 119.98 (TFA), 127.29 (2 carbon atoms), 136.86, 153.94 (2 carbon atoms), 155.40, 156.97, 158.97 (q, J 32 Hz) (TFA), 166.92, 168.99, 169.34, 170.19, 170.81, 174.50. HRMS (ESI): m/z [M + 3H]³⁺ calcd. for [C₄₆H₈₅N₁₄O₉Si]³⁺ 335.2126, found: 335.2135 (17a), 335.2132 (17b). RP-HPLC (220 nm): 17a: 97% (t _R = 9.5 min, k = 11.5), 17b: 99% (t _R = 10.9 min, k = 13.3). C₄₆H₈₂N₁₄O₉Si·C₈H₄F₁₂O₈ (1003.34 + 456.09).

N ^α-Methyl-N ^ω-{[4-(N-{1-carboxylato­[2-(6-(dimethylamino)-3-(dimethyliminio))-3H-xanthen-9-yl]­phen-5-yl}­carbonyl)-aminobutyl]­aminocarbonyl}­Arg-Arg-Pro-β,β-dimethyl-l-Tyr-Ile-β-trimethylsilyl-Ala Tris­(hydro-trifluoroacetate) (19)

Peptide 17a (3.4 mg, 2.3 μmol) and DIPEA (4.7 μL, 27 μmol) were dissolved in DMF/NMP 8:2 v/v (100 μL) followed by the addition of a solution of 5-TAMRA succinimidyl ester 18 (1.1 mg, 2.1 μmol) in DMF/NMP 8:2 v/v (30 μL). After stirring at room temperature in the dark for 1 h, the mixture was diluted with 10% aqueous TFA (14 μL) and acetonitrile/0.1% aqueous TFA 20:80 v/v (1 mL). Isolation by preparative RP-HPLC (system 1, gradient: 0–35 min: acetonitrile/0.1% aqueous TFA 20:80–60:40, t _R = 16 min) gave 19 as a purple fluffy solid (1.6 mg, 40%) ¹H NMR (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.73–0.83 (m, 6H), 0.92–1.03 (m, 3H), 1.22–1.29 (m, 6H), 1.34–1.41 (m, 1H), 1.49–1.61 (m, 9H), 1.62–1.68 (m, 1H), 1.68–1.86 (m, 6H), 1.86–1.96 (m, 1H), 2.45–2.48 (m, 3H), 3.07–3.18 (m, 6H), 3.27 (s, 12H), 3.33–3.36 (m, 2H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.51–3.52 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.58–3.59 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.81–3.84 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 4.13 (t, 1H, J 7.8 Hz), 4.18–4.23 (m, 1H), 4.39–4.44 (m, 1H), 4.54–4.60 (m, 1H), 4.66 (d, 1H, J 9.1 Hz), 6.31–6.72 (br s, 1H, interfering with next listed signal), 6.59–6.64 (m, 2H), 6.82–7.17 (m, 9H), 7.18–7.48 (br s, 2H, interfering with next listed signal), 7.39 (d, 1H, J 9.3 Hz), 7.48–7.62 (m, 3H), 7.68 (t, 1H, J 5.1 Hz), 8.08 (d, 1H, J 7.6 Hz), 8.28 (d, 1H, J 7.6 Hz), 8.47 (s, 2H), 8.61–8.74 (m, 1H), 8.81–9.24 (m, 6H), 10.30 (s, 1H), 12.40 (s, 1H), 13.40 (s, 1H). HRMS (ESI): m/z [M + 3H]³⁺ calcd. for [C₇₁H₁₀₅N₁₆O₁₃Si]³⁺ 472.5933, found: 472.5942. RP-HPLC (220 nm): 99% (t _R = 8.8 min, k = 10.6). C₇₁H₁₀₂N₁₆O₁₃Si·C₆H₃F₉O₆ (1415.78 + 342.07).

N ^α-Methyl-N ^ω-{[4-(N-{6-3H-[2-(5–3H-(1-(4-sulfonatobutyl)-3,3-dimethyl)­indol-2-ylidenepent-1,3-dienyl)-3,3-dimethyl-5-sulfo]­indol-1-ium-1-yl}­hexanoyl)-aminobutyl]­aminocarbonyl}­Arg-Arg-Pro-β,β-dimethyl-l-Tyr-Ile-β-trimethylsilyl-Ala Tris­(hydrotrifluoroacetate) (21)

Peptide 17a (3.3 mg, 2.3 μmol) and DIPEA (4.6 μL, 26 μmol) were dissolved in DMF/NMP 8:2 v/v (100 μL) followed by the addition of the sulfo-Cy5 succinimidyl ester 20 (1.1 mg, 2.1 μmol) in DMF/NMP 8:2 v/v (30 μL). After stirring at room temperature in the dark for 1 h, the mixture was diluted with 10% aqueous TFA (14 μL) and 0.1% acetonitrile/aqueous TFA 20:80 v/v (1 mL). Isolation by preparative RP-HPLC (system 1, gradient: 0–35 min: acetonitrile/0.1% aqueous TFA 20:80–60:40, t _R = 17 min) gave 21 as a purple fluffy solid (1.4 mg, 30%). ¹H NMR (600 MHz, DMSO-d ₆): δ (ppm) 0.00 (s, 9H), 0.73–0.83 (m, 6H), 0.91–1.01 (m, 3H), 1.21–1.29 (m, 6H), 1.28–1.44 (m, 8H), 1.48–1.59 (m, 7H), 1.63–1.71 (m, 15H), 1.71–1.84 (m, 9H), 1.85–1.95 (m, 1H), 2.04 (t, 2H, J 6.4 Hz), 2.46–2.48 (m, 3H), 2.64 (t, 2H, J 6.8 Hz), 2.98–3.03 (m, 2H), 3.06–3.14 (m, 4H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.22–3.26 (m, 2H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.55–3.57 (m, 2H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 3.80–3.84 (m, 1H, interfering with the water signal, quantified in the spectrum acquired after the addition of D₂O), 4.03–4.09 (m, 2H), 4.10–4.16 (m, 3H), 4.17–4.23 (m, 1H), 4.39–4.44 (m, 1H), 4.52–4.58 (m, 1H), 4.64 (d, 1H, J 9.3 Hz), 6.24 (d, 1H, J 13.7 Hz), 6.39 (d, 1H, J 14.0 Hz), 6.53–6.65 (m, 3H), 6.68–7.06 (br s, 2H), 7.09–7.16 (m, 2H), 7.20–7.32 (m, 3H), 7.32–7.58 (m, 7H), 7.61–7.67 (m, 2H), 7.78 (t, 1H, J 5.5 Hz), 7.81–7.84 (m, 1H), 8.08 (d, 1H, J 7.7 Hz), 8.25–8.48 (m, 4H), 8.74–9.05 (m, 4H), 9.06–9.18 (m, 1H), 9.99 (s, 1H), 12.40 (s, 1H). HRMS (ESI): m/z [M + 3H]³⁺ calcd. for [C₈₁H₁₂₆N₁₆O₁₆S₂Si]³⁺ 557.2937, found: 557.2943. RP-HPLC (220 nm): 99% (t _R = 10.1 min, k = 12.3). C₈₁H₁₂₃N₁₆O₁₆S₂Si·C₆H₃F₉O₆ (1669.18 + 342.07).

Capillary Electrophoresis

Capillary electrophoresis was performed with an Agilent 7100 CE system using a bare fused silica capillary with an inner diameter of 0.05 mm and a length of 72 cm (G1600–62211, Agilent). 50 mM α-cyclodextrin (Honeywell, Charlotte, United States) in a 125 mM sodium phosphate buffer (pH = 7.0) served as electrolyte solution. Injections were performed hydrodynamically applying a pressure of 100 mbar for 10 s. The temperature of the capillary housing was set to 30 °C. A voltage of 20 kV was applied for 35 min and the detection was performed at 220 nm.

Chemical Stability

The chemical stability of peptides 19 and 21 was investigated in PBS (adjusted to pH = 7.4) at 22 °C in the dark. The incubation was started by the addition of 3 μL of a 5 mM stock solution (solvent: DMSO) to 147 μL of PBS to yield a concentration of 100 μM. After periods of 0, 6, 24, and 48 h, an aliquot (25 μL) was removed and added to 25 μL of acetonitrile/0.04% aq. TFA 2:8 v/v to obtain a peptide solution with a concentration of 50 μM. 20 μL of this solution were subjected to analytical RP-HPLC analysis using the same system and conditions as described under Analytical HPLC. A 1:1 mixture of PBS and acetonitrile/0.04% aq. TFA 2:8 v/v (20 μL) was analyzed to obtain the blank chromatogram.

Stability in Human Plasma

The proteolytic stabilities of NT(8–13), 15, 17a, 19 and 21 were investigated in human blood plasma/PBS 1:2 v/v according to a described procedure. Samples were analyzed using the RP-HPLC system and conditions as described under Analytical HPLC with the following gradient: 0–6 min: acetonitrile/0.04% aq. TFA 10:90–21:79, 6–12 min: 21:79–40:60, 12–13 min: 40:60–95:5, 13–16 min: 95:5 (isocratic). Data analysis was based on UV detection at 220 nm. Reference samples, representing 100% recovery, were prepared in quadruplicate. Recovery ratios were obtained by dividing the recovery of the peptide by the recovery of IS for each individual sample (n = 3–5). The obtained recoveries and the recovery ratios are summarized in Table S1.

Excitation Spectra, Emission Spectra, and Fluorescence Quantum Yields

Absorption spectra were recorded with a referenced single beam spectrometer (Cary 60, Agilent) and the emission and excitation spectra were recorded in an orthogonal configuration in an emission spectrometer (Fluorolog, Horiba) setting the resolution to 1 nm for excitation and emission detection. The fluorescence decays were recorded with a self-constructed Time Correlated Single Photon Counting (TCSPC) setup with single detection wavelength at room temperature. The sample was excited along a 10 mm path length at λ_ex = 280 nm and the emission was recorded orthogonally to this along a 2 mm path length at λ_obs as indicated in Figure S7. The optical density of the sample was set to ca. 0.1 at the excitation wavelength over 10 mm path length. The fluorescence quantum yields of compounds 19 and 21 were determined in PBS and PBS supplemented with 1% BSA via an absolute method using an Ulbricht sphere with an inaccuracy of ca. 3% (Hamamatsu C9920–02 system equipped with a Spectralon integrating sphere) at room temperature using a 10 mm × 10 mm quartz cuvette. The optical density at the excitation wavelength of the sample was <0.1 along an optical path length of 1 cm.

Cell Culture

Mammalian cells were cultured in T75 or T175 culture flasks (Sarstedt). Chinese hamster ovary (CHO) cells stably expressing hNTS₁R were cultured in DMEM/HAM’s F12 (Sigma, Taufkirchen, Germany) medium (1:1) supplemented with 7.5% FBS, l-glutamine (Sigma) (630 μg/mL), and hygromycin B (Carl Roth, Karlsruhe, Germany) (250 μg/mL). HT-29 colon carcinoma cells (DSMZ-no. ACC 299) were grown in antibiotic-free RPMI-164 medium (Sigma) supplemented with 7.5% FBS or in MEM Earles medium supplemented with 10% FBS, 1% l-glutamine, 1% NEAA, and 1% pyruvate. HEK293T-hNTS₂R cells were cultured in Dulbecco’s modified Eagle’s medium (Sigma-Aldrich) supplemented with 10% FBS, l-glutamine (2 mM) and penicillin–streptomycin (100 IU/mL and 0.1 mg/mL, respectively) (Sigma-Aldrich). All cell lines were cultured in a humidified atmosphere (95% air and 5% CO₂) at 37 °C.

Buffers Used for Binding and Functional Assays

DPBS: Dulbecco’s phosphate-buffered saline with calcium and magnesium (1.8 mM CaCl₂, 2.68 mM KCl, 1.47 mM KH₂PO₄, 3.98 mM MgSO₄, 137 mM NaCl, 8.06 mM Na₂HPO₄, pH = 7.4) supplemented with 0.1% BSA (flow cytometry) or 1% BSA (radiochemical assays) and 0.1 mg/mL bacitracin. For flow cytometric binding studies, DPBS was filtrated using 0.2 μm nylon syringe filters (Phenomenex).

Fura-2 assay buffer: HEPES buffer (120 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 1.5 mM CaCl₂, 25 mM HEPES, and 10 mM glucose at pH = 7.4) supplemented with 2% BSA and 2.5 mM Probenecid (Sigma).

Radiochemical Competition Binding Assays with [³H]­UR-MK300

NTS₁R radioligand competition binding experiments with [³H]­UR-MK300 (molar activity: 2.41 TBq/mmol) and the peptides under study were performed with intact hNTS₁R-expressing human HT-29 colon carcinoma cells at 23 ± 1 °C using a previously reported protocol. The K _d value of the radioligand for the used batch of [³H]­UR-MK300 was determined recently (K _d = 0.41 nM). Cells were seeded 1 day before the assay, yielding a confluency of at least 90% on the day of the experiment. Total binding data (including total binding in the absence of competitor) were plotted as dpm values against log­(concentration competitor) and analyzed by a four-parameter logistic equation (log­(inhibitor) vs response – variable slope, GraphPad Prism 5, GraphPad Software, San Diego, CA) to obtain pIC₅₀ values. Individual pIC₅₀ values were converted to pK _i values according to the Cheng-Prusoff equation (logarithmic form). To plot average data from individual binding experiments, data were normalized (100% = “top” of the four-parameter logistic fit, 0% = nonspecifically bound radioligand).

NTS₂R radioligand competition binding experiments with [³H]­UR-MK300 were performed with membranes of HEK293T-hNTS₂R. Membrane preparations were prepared using a recently reported protocol for membrane preparations of CHO-hNTS₁R cells. The protein concentration amounted to 1.8 ± 0.2 mg/mL (mean value ± SEM from three different sample dilutions). NTS₂R saturation and competition binding experiments with [³H]­UR-MK300 were carried out as described for saturation and competition binding studies with [³H]­UR-FE051 at membrane preparations of CHO-hNTS₁R cells. The soluble protein concentration in the assays was 135 μg/mL. The K _d value of [³H]­UR-MK300 amounted to 3.1 ± 0.2 nM (mean value ± SEM from three independent experiments, each performed in triplicate). The concentration of [³H]­UR-MK300 in the competition binding assays was 5 nM. Data from competition binding assays were processed as the data from the NTS₁R binding assays.

Fura-2 Ca²⁺ Assay

The Fura-2 Ca²⁺ assay was conducted as previously described with the following modifications: HT-29 cells were used in place of CHO-hNTS₁R cells and 1 μM NT(8–13) was used to determine the maximal response. Measurements were performed in black 96-well plates (Greiner 655076, Greiner Bio-One, Frickenhausen, Germany) and a measurement for one well comprised 40 cycles (instead of 44 cycles) with a cycle duration of 1.4 s (total time: 56 s).

Flow Cytometric Binding Experiments

Flow cytometry-based NTS₁R binding studies were performed with intact HT-29 (all kind of experiments) and CHO-hNTS₁R cells (only saturation binding) using a BD FACSCanto II flow cytometer (Becton Dickinson, Heidelberg, Germany), equipped with an argon laser (488 nm), a red diode laser (640 nm), and a BD High Throughput Sampler (HTS unit) for microtiter plates. Saturation and competition binding experiments were performed in triplicate in 96-well polypropylene plates (Brand 701330, Wertheim, Germany) using the HTS unit for sample injection. Association and dissociation experiments were performed in triplicate in 5 mL polypropylene tubes (VWR, Radnor, USA), from which samples were directly injected into the flow cytometer. The following gain settings for forward and sideward scatter were applied: FSC, 0 V; SSC, 252 V. Fluorescence was recorded using the PE channel (excitation: 488 nm, emission: 585 ± 21 nm) with a PMT gain of 400–500 V (19) or the APC channel (excitation: 640 nm, emission: 660 ± 10 nm) with a PMT gain of 380–580 V (21). For measurements using the HTS unit, 45 μL of the sample were injected with a speed of 1.5 μL/s. For measurements, using an injection from 5 mL sample tubes, the medium flow rate (60 μL/min) was used. Measurements were stopped after 30 s (HTS unit) or after counting of at least 3000 gated events (injection from sample tubes).

Cells were seeded in T75 culture flasks 3–4 days (CHO-hNTS₁R cells) or 5–6 days (HT-29 cells) prior to the experiment. On the day of the experiment, cells were detached by trypsinization, suspended in DPBS and centrifuged at 200 g at rt for 5 min. The cell pellet was resuspended in binding buffer and the cell density was adjusted to 1.5 × 10⁴ or 1.5 × 10⁵ cells/mL (CHO-hNTS1R cells) or 5.0 × 10⁵ cells/mL (HT-29 cells). Nonspecific binding was determined in the presence of 1 μM NT(8–13).

Saturation Binding Experiments

A 96-well polypropylene plate was prefilled with 200 μL of cell suspension. For total binding, H₂O (2 μL) and DMSO/H₂O 2:8 v/v (2 μL) containing the fluorescent ligand (100-fold concentrated compared to the final concentration) were added. To determine nonspecific binding, a 100 μM solution of NT(8–13) in H₂O (2 μL) and DMSO/H₂O 2:8 v/v (2 μL) containing the fluorescent ligand (100-fold concentrated) were added. Samples were incubated at 23 ± 1 °C in the dark under gentle shaking for 2 h followed by measurement via the HTS unit. Specific binding data, obtained by subtracting triplicate mean values of nonspecific binding from triplicate mean values of total binding, were plotted against the fluorescent ligand concentration and analyzed by a two-parameter equation describing hyperbolic single-site binding (one site, specific binding, GraphPad Prism 5) to obtain K _d values.

Association Experiments

5 mL polypropylene tubes were prefilled with 2000 μL of cell suspension and H₂O (20 μL) was added (determination of total binding). To start the association, a 100-fold concentrated solution (compared to the final concentration) of the fluorescent ligand in DMSO/H₂O 2:8 v/v (20 μL) was added (final concentrations: 19: 1 nM, 21: 0.1 nM). The sample tubes were gently shaken in the dark at 23 ± 1 °C. Measurements were conducted within different periods of time (1–240 min) by placing the tubes in the injection port of the cytometer. To determine nonspecific binding, samples were set up as in the case of total binding, but instead of 20 μL of H₂O, 20 μL of a 100 μM solution of NT(8–13) in H₂O were added. It should be noted that this experimental setup corresponds to a pseudo first-order measurement as the concentration of the free ligand, being markedly higher than the receptor concentration, can be considered constant by approximation during the association reaction. Specific binding data of 21, obtained by subtracting triplicate mean values of nonspecific binding from triplicate mean values of total binding, were plotted against the time and analyzed by a three-parameter equation describing an exponential rise to a maximum (one-phase association, Y₀ constrained to zero, GraphPad Prism 5) to yield the observed association rate constant k _obs. To calculate mean values in %, specific binding data were normalized based on the corresponding B _eq value. Since specific binding data of 19 indicated a biphasic association, data of 19 were analyzed using a five-parameter two-phase association fit (two-phase association, Y₀ constrained to zero, GraphPad Prism 5) to obtain the observed association rate constants k _obs(bi,fast) and k _obs(bi,slow) for the fast and slow phase of the association, respectively. To calculate mean values in % (cf. Figure B), specific binding data were normalized based on the corresponding B _eq values from the two-phase association fit (set to 100%). In addition to this analysis, the whole specific binding data of 19 were analyzed by a three-parameter equation describing an exponential rise to a maximum (one-phase association, Y₀ constrained to zero, GraphPad Prism 5) to yield the observed association rate constant k _obs(mono).

Dissociation Experiments

5 mL polypropylene tubes were prefilled with 2000 μL of cell suspension. For the determination of total binding, the preincubation was started by the addition of H₂O (20 μL) and a 100 nM solution of 19 or a 25 nM solution of 21 in DMSO/H₂O 2:8 v/v (20 μL) (final fluorescent ligand concentration: 1 nM (19) or 0.25 nM (21)). To determine nonspecific binding, a 100 μM solution of NT(8–13) in H₂O (20 μL) and a 100 nM solution of 19 or a 25 nM solution of 21 in DMSO/H₂O 2:8 v/v were added. The samples were gently shaken in the dark at 23 ± 1 °C for 120 min (19) or for 180 min (21). The dissociation process was initiated by the addition of a 2.5 mM solution of NT(8–13) in H₂O (20 μL) (final concentration: approximately 25 μM). After different periods of time (1–300 min (19) or 5–600 min (21)), sample aliquots were measured by placing the tube in the injection port of the flow cytometer. Specific binding data, obtained by subtracting triplicate mean values of nonspecific binding from triplicate mean values of total binding, were plotted against the time and analyzed by a three-parameter equation describing an incomplete monophasic exponential decline (one phase decay, GraphPad Prism 5) to obtain k _off values. For both fluorescent ligands, the mean ± SEM of the plateau values from individual experiments proved to be significantly different from zero (unpaired one-tailed t test, P > 0.05). To calculate mean values in %, binding data were normalized based on the specifically bound ligand measured immediately before the start of the dissociation.

Calculation of Association Rate Constants (k _on) and Kinetically Derived Dissociation Constants K _d(kin)

The association rate constants were calculated from k _obs mean values, k _off mean values, and the fluorescent ligand concentration used for the association experiments ([FL]) according to the equation k _on = (k _obs – k _off)/[FL]. The kinetically derived dissociation constants K _d(kin) were calculated from the respective k _on value and the k _off mean value according to K _d(kin) = k _off/k _on.

Competition Binding Experiments

A 96-well polypropylene plate was prefilled with 200 μL of cell suspension. 2 μL of DMSO/H₂O 2:8 v/v (for the determination of total binding in the absence of competitor), 2 μL of a 100 μM solution of NT(8–13) in H₂O (determination of nonspecific binding) or 2 μL of a 100-fold concentrated solution (compared to the final concentration) of the compound of interest (NT(8–13), SR142948, SR48692 or UR-MK300; used at varying concentrations) in DMSO/H₂O 2:8 v/v were added and the plate was shortly shaken. Subsequently, 2 μL of a 30 nM solution of 19 or a 10 nM solution of 21 in DMSO/H₂O 2:8 v/v were added to each well and the plate was gently shaken in the dark at 23 ± 1 °C for 120 min (19) or 180 min (21) followed by measurement via the HTS unit. The final concentrations of 19 and 21 corresponded to their 3-fold or 2-fold K _d values determined by equilibrium saturation binding (K _d = 0.11 nM (19) or 0.046 nM (21)). Total binding fluorescence intensities (including total binding in the absence of competitor) were plotted against log (concentration inhibitor) and analyzed by a four-parameter logistic equation (log­(inhibitor) vs response-variable slope, GraphPad Prism 5) to obtain pIC₅₀ values. Individual pIC₅₀ values were converted to pK _i values according to the Cheng-Prusoff equation (logarithmic form). To plot average data from individual binding experiments, data were normalized (100% = “top” of the four-parameter logistic fit, 0% = nonspecifically bound fluorescent ligand).

Binding Experiments with SBI-553

A 96-well polypropylene plate was prefilled with 200 μL of cell suspension. 2 μL of DMSO/H₂O 2:8 v/v (determination of total binding in the absence of SBI-553) or 2 μL of a 100-fold concentrated solution (compared to the final concentration) of SBI-553 (used at varying concentrations) in DMSO/H₂O 2:8 v/v (determination of total binding in the presence of SBI-553) were added and the plate was incubated at 23 ± 1 °C for 30 min. Subsequently, 2 μL of a 1 nM solution of 21 in DMSO/H₂O 2:8 v/v (final concentration: 0.01 nM) were added. To determine nonspecific binding, 2 μL of a 100 μM solution of NT(8–13) in H₂O and 2 μL of a 1 nM solution of 21 in DMSO/H₂O 2:8 v/v were added. The plate was then gently shaken in the dark at 23 ± 1 °C for 180 min followed by measurement via the HTS unit. Total binding fluorescence intensities (including total binding in the absence of the allosteric modulator) were plotted against log (concentration SBI-553). As the data clearly indicated a biphasic course, data were analyzed by a five-parameter two sitesfit logIC₅₀ (GraphPad Prism 5) to obtain pIC₅₀ (corresponding to pEC₅₀) values for the high- and low-affinity state (the biphasic fit was favored over the four-parameter logistic fit according to the F-test, P < 0.05, GraphPad Prism 5). To plot average data from individual binding experiments, data were normalized (100% = lower plateau of the initial biphasic fit, 0% = nonspecifically bound fluorescent ligand). The differences between the lower and upper plateaus of the individual experiments gave the E _max values which were averaged to give the E _max mean ± SEM.

Confocal Microscopy

Confocal microscopy was performed with a Zeiss LSM 710 confocal laser scanning microscope (Zeiss, Jena, Germany). The objective was 63× magnification with oil (1.4 NA). One day prior to the experiment, CHO-hNTS₁R (30,000 to 35,000 cells per chamber) and HT-29 cells (45,000 to 50,000 cells per chamber) were seeded in Nunc LabTekTM II cover glasses with 8 chambers (Thermo Fisher Scientific). On the day of the experiment, the confluency of the cells was 60–80%. After removal of the culture medium, cells were washed with Leibovitz’s L15 medium (200 μL) and covered with L15 medium (150 μL) containing H33342 (Sigma-Aldrich) (2 μg/mL). To study total binding, L15 medium (150 μL) containing H33342 (2 μg/mL) and 19 or 21 (final concentration: 2 nM) was added. To determine nonspecific binding, L15 medium (150 μL) containing H33342 (2 μg/mL), NT(8–13) (final concentration: 1 μM) and 19 or 21 was added. The first image was acquired 5 min after addition of the fluorescent ligand. The pinhole was set to 1.0 airy unit. Laser powers and gains for the investigation of 19: 405 nm, 2% (gain: 700 V); 561 nm, 8% (gain: 800 V) (CHO-hNTS₁R cells) or 10% (gain: 850 V) (HT-29 cells). Laser powers and gains for the study of 21: 405 nm, 2% (gain: 650 V (HT-29 cells) or 750 V (CHO-hNTS₁R cells)); 633 nm, 20% (gain: 1050 V) (CHO-hNTS₁R cells) or 40% (gain: 1100 V) (HT-29 cells). The following filter settings for fluorescence detection were applied: 410–549 nm (H33342, experiments with 19), 410–585 nm (H33342, experiments with 21), 562–649 nm (19), and 638–759 nm (21).

Biomolecular Imaging

Saturation Binding Studies at HT-29 Cells

HT-29 cells were seeded 1 day prior to the experiment (1.0 × 10⁵ cells/well) in black clear bottom 96-well plates (Greiner 655090, Greiner Bio-One). After washing the cells with Leibovitz’s L-15 Medium (Thermo Fisher Scientific, Waltham, MA, USA), 100 μL of L-15 containing 19 or 21 in various concentrations (0.1–100 nM (19) or 0.1–25 nM (21), each in triplicates) followed by incubation at rt in the dark for 2 h. To determine nonspecific binding, SR142948 (1 μM; Sigma-Aldrich) was added to the samples. For the determination of the autofluorescence only L-15 was added to the cells. In addition, CellTag 520 or 700 Stain (LI-COR Biotechnology, Bad Homburg, Germany) was added to each well (final concentrations: 1 μM and 0.2 μM, respectively). After incubation, the plate was set on ice, the cells were washed three times with ice-cold DPBS (Sigma-Aldrich), the buffer was removed, and fluorescence images were acquired using the Sapphire FL Biomolecular Imager (520 and 658 nm laser; pixel size: 20 μm; Azure Biosystems, Dublin, CA, USA). The following bandpass filters were used for fluorescence detection: 565/24 nm (19) or 710/40 nm (21). Fluorescence intensities for each well were normalized to the respective CellTag fluorescence. Specific binding data, obtained by subtracting triplicate mean values of nonspecific binding from triplicate mean values of total binding, were plotted against the ligand concentration and K _d values were obtained by fitting of the data with a two-parameter equation describing hyperbolic single-site binding (one site, specific binding, GraphPad Prism 9). Four independent experiments were performed in triplicate.

Fluorescent Imaging of Tumor Slices

NMRI (nu/nu) mice were bred in the animal facility of the University of Regensburg. All animal experiments were performed following the protocols evaluated and approved by the local veterinary medicine authorityRegierung der Oberpfalz, Bavaria, Germany (approval number 2532.4-11/11). HT-29 tumor slices (10 μm) were prepared from HT-29 tumors (subcutaneously grown in 3–6 months old male NMRI nude (nu/nu) mice) using a cryostat microtome HM 500 O (Microm, Walldorf, Germany) and thaw-mounted on HistoBond adhesive glass slides (Marienfeld, Lauda-Königshofen, Germany). Tumor slices used for the determination of total binding were covered with buffer (DPBS supplemented with 1.8 mM CaCl₂ and 3.98 mM MgCl₂) at rt for 15 min. Tumor slices used for the determination of nonspecific binding were preincubated with buffer containing SR142948 (5 μM in buffer). The tumor slices were then incubated with 19 (10 nM) or 21 (5 nM) in buffer to determine total binding. For the determination of nonspecific binding, the samples contained additionally SR142948 (5 μM). Tumor slices were incubated with neat buffer to determine background signals (autofluorescence). After 1 h of incubation, the tumor slices were washed three times with ice-cold buffer, once with cold water, and were allowed to dry prior to acquisition of fluorescence images as described under Saturation binding studies at HT-29 cells. Data analysis was performed using the Aida Image Analyzer (v.5.1, Elysia-raytest GmbH, Straubenhardt, Germany).

Calculation of Propagated Errors

Propagated errors (applying to specifically bound fluorescent ligand (saturation binding)), association rate constants k _on, and kinetically derived dissociation constants K _d(kin) were calculated as described elsewhere.

Glossary

Abbreviations

AF

autofluorescence

Arg­(carb)

N ^ω-carbamoylated arginine

CE

capillary electrophoresis

CEC

concentration-effect curve

CH₂Cl₂

dichloromethane

CHO cells

Chinese hamster ovary cells

2-ClTrt

2-chlorotrityl

DPBS

Dulbecco’s phosphate buffered saline

DIPEA

diisopropylethylamine

FBS

fetal bovine serum

FC

flow cytometry

HBTU

[bis­(dimethylamino)­methyliumyl]-3H-benzotriazol-1-oxide hexafluorophosphate

HOBt

hydroxybenzotriazole

K _d

dissociation constant from a saturation binding experiment

k _obs

observed association rate constant

k _off

dissociation rate constant

k _on

association rate constant

MTBD

methyl-1,5,7-triazabicyclo­[4.4.0]­dec-5-ene

NEAA

non essential amino acid solution

NTS₁R

neurotensin receptor type 1

pEC₅₀

negative decadic logarithm of the half maximal effective concentration (functional assays)

pK _i

negative decadic logarithm of the dissociation constant K _i (in nM) obtained from a competition binding experiment

ROI

region of interest

SEM

standard error of the mean

SPPS

solid phase peptide synthesis

TMSAla

(trimethylsilyl)­alanine

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.5c01701.

• Synthesis of compounds 10, 24, 26, and 27 (Scheme S1) and experimental protocols and analytical data of 10, 24, 26, and 27; structures and NTS₁R binding affinities of reported NT(8–13) derivatives containing β,β-dimethyl-l-Tyr or β,β-dimethyl-d-Tyr in position 11 (Figure S1); chemical stability of 19 and 21: chromatograms of the HPLC analyses (Figures S2 and S3); structure of [³H]­UR-MK300 (Figure S4); radioligand displacement curves from competition binding experiments with [³H]­UR-MK300 (Figure S5); concentration response curves of NT(8–13), 19, and 21 obtained from a Fura-2 Ca²⁺ assay (Figure S6); absorption, emission, and excitation spectra, and graphical presentation of the fluorescence decay of 19 and 21 (Figure S7); representative binding isotherms of 19 and 21 obtained from flow cytometric saturation binding experiments performed at intact CHO-hNTS₁R cells (Figure S8); visualization of binding of 19 to intact CHO-hNTS₁R cells by confocal microscopy (Figure S9); visualization of binding of 21 to intact CHO-hNTS₁R cells by confocal microscopy (Figure S10); confocal microscopy: estimation of total cellular fluorescence and intracellular fluorescence (Figure S11); fluorescence image of a 96-well plate with adherent HT-29 cells acquired with a Biomolecular Imager after incubation with 19 (Figure S12); fluorescence images of cryosections of a HT-29 tumor acquired after incubation with 19 (Figure S13); recoveries of 19 and 21 from human plasma/PBS 1:2 v/v and ratios of compound-recovery over recovery of internal standard (Table S1); electropherograms of the CE analysis of (R,S)-10 and (R)-10; RP-HPLC chromatograms of 6, 10, 14–17b, 19, and 21 (purity controls); ¹H NMR spectrum and¹³C NMR spectrum of compound 10; ¹H NMR spectra of compounds 14–17b, 19, and 21 and ¹³C NMR spectra of compounds 14–17b (PDF)

• Molecular formula strings and NTS₁R receptor affinities (CSV)

∇.

Roche Diagnostics GmbH, Nonnenwald 2, 82377 Penzberg, Germany

F.J.E. and E.J.S. performed the syntheses. F.J.E. performed radiochemical binding assays, stability studies, and flow cytometric binding assays. F.J.E. and C.H. performed functional assays. N.A. and R.J.K. recorded the absorption, emission, and excitation spectra including the fluorescence decays and fluorescence quantum yields. F.J.E. and M.K. performed confocal microscopy imaging experiments. A.F. performed biomolecular imaging experiments. M.K. initiated and planned the project. M.K., O.P., P.K., S.M., and R.J.K. supervised the research. F.J.E. and M.K. wrote the manuscript with support from all co-authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.