Design, synthesis and pharmacological characterization of fluorescent peptides for imaging human V1b vasopressin or oxytocin receptors

Abstract

Among the four known vasopressin and oxytocin receptors, the specific localization of the V1b isoform is poorly described due to the lack of selective pharmacological tools. In an attempt to address this need, we decided to design, synthesize and characterize fluorescent selective V1b analogues. Starting with the selective V1b agonist, [deamino-Cys¹, Leu⁴, Lys⁸]vasopressin (d[Leu⁴,Lys⁸]VP) synthesized earlier, we added blue, green or red fluorophores to the lysine residue at position 8, either directly or by the use of linkers of different lengths. Among the nine analogues synthesized, two exhibited very promising properties. These are d[Leu⁴, Lys (Alexa 647)⁸]VP (3) and d[Leu⁴, Lys (11-aminoundecanoyl-Alexa 647)⁸]VP (9). They remained full V1b agonists with nanomolar affinity and specifically decorated the plasma membrane of CHO cells stably transfected with the human V1b receptor. These new selective fluorescent peptides will allow the cellular localisation of V1b or OT receptor isoforms in native tissues.

Introduction

Vasopressin (VP) and the VP receptor family including V1a, V1b, V2 and the related oxytocin (OT) receptor, are involved in many different physiological functions. Among them, the V1b receptor is well known for the activation of adrenocorticotropin hormone (ACTH) secretion in the pituitary, thus strongly participating, with corticotrophin releasing hormone (CRF), in the activation of the pituitary/corticotrope axis in stress and anxiety¹. Vasopressin also regulates adrenal function by mediating catecholamines secretion² and induces pancreatic insulin secretion via V1b receptors³. Beside this neuroendocrine effect, V1b receptors have been suspected to be also localized centrally and to participate in cognitive and behavioural functions involved in rewarding, cognition and sociality (see⁴ for review).

Although several publications describe the vasopressin V1a, V2 and OT receptor distribution by using autoradiography^5,6,7 or immunodetection⁸ techniques, the lack of selective radio labelled VP analogues or of receptor antibodies has hindered progress in the detection of V1b receptor distribution in native tissues. Results obtained by molecular approaches such as polymerase chain reaction after reverse transcription (RT-PCR)^9,10 or messenger ribonucleic acid (mRNA) detection by in situ hybridization^11,12,13, although more accurate, did not provide clear information regarding the brain regions detected by immunostaining. This was partially due to the fact that the probes shared common sequences with V1a and OT receptors⁹.

Fluorescent tools have been synthesized for other receptors of the VP/OT family^14,15,16. However, although good fluorescent V1a and OT ligands have been produced; to date, no good fluorescent specific ligand is available to selectively detect central and peripheral V1b receptors. Here we describe the synthesis of nine new fluorescent analogues for the human V1b receptor (hV1b-R). Two of them are promising new tools for the detection of V1b or OT receptors in native tissue. In our previous work^17,18,3,19, by replacing the glutamine⁴ of the natural arginine vasopressin (AVP) with a cyclohexylalanine or a leucine, the arginine⁸ by a lysine and by removing the NH₂ of the cysteine¹ to increase stability towards aminopeptidases, we produced analogues showing an increased selectivity for the rat V1b receptor (rV1b-R). Namely, the [deamino-Cys¹,4-leucine,8-lysine]vasopressin (d[Leu⁴,Lys⁸]VP) analogue, previously synthesized in the du Vigneaud laboratory²⁰, was found to be selective for the rV1b-R^3,19 and, to a lesser extent, for the hV1b-R. It conserved a nanomolar affinity for these receptor isoforms¹⁸. By studying the structure/activity of the ligand/V1b receptor complex, we have also been able to propose a model of the ligand-receptor interaction²¹. This model predicts that the first six amino acids would be located into a “binding pocket” inside the membrane, whereas the residues at positions 7, 8 and 9 would be outside. Thus, the Lys⁸ residue provides an epsilon NH₂ group which could be used for further additions. Accordingly, we have taken advantage of the stability and good selectivity of d[Leu⁴,Lys⁸]VP^3,18 to introduce fluorophores on the side chain of the Lys⁸ residue to create fluorescent tools that would conserve the pharmacology of d[Leu⁴,Lys⁸]VP, resist degradation and selectively decorate with an excellent resolution, the plasma membrane of CHO cells expressing V1b and/or OT receptors.

Rationale for fluorescent selective V1b synthesis

We used d[Leu⁴, Lys⁸]VP (Peptide A in Fig 1, Tables I, II and III) as the parent molecule for designing fluorescent V1b analogues, since we had previously shown that this peptide is a V1b selective agonist both for rat and human VP/OT receptors^3,18,19. First, we carried out the addition of an hydroxyl (OH) group at the N terminal position of d[Leu⁴,Lys⁸]VP to give [1-L-(−)-2-hydroxy-3-thiopropanoic acid, 4-leucine, 8-lysine]vasopressin ([HO¹][Leu⁴,Lys⁸]VP) (Peptide B in Tables I and II). We had shown previously that this modification was helpful in the design of fluorescent agonists for the hOT-R^16,22. We observed that the OH modification at position one didn’t affect the ligand’s affinity but strongly increased its selectivity for the hV1b-R (see Table II). We thus also synthesized a short series of OH-modified fluorescent compounds. We selected different fluorophores to attach to the d[Leu⁴,Lys⁸]VP (Peptide A, Fig 1): First, we have chosen the anthraniloyl (Atn) group, a small fluorescent molecule with a MW=97 daltons and a quantum yield (R) of 0.358, used for the design of analogue 1 and highly sensitive to microenvironmental changes²³. We have also selected the Alexa fluorophores (Molecular Probes) for their brightness and their resistance to photobleaching²⁴. We have used Alexa Fluor 488 (Alexa 488) (MM= 643; Quantum yield R=0.60; Molar extinction coefficient ε=71,000) and Alexa Fluor 647 (Alexa 647) (MM~1300; R=not determined, ε=239,000), the latter being one of the brightest fluorescent molecules reported so far²⁵. Moreover Alexa fluorophores are charged molecules, soluble in water and not prone to create molecular stacking as the tetramethylrhodamine does (which they are derived from). The aliphatic omega-amino carboxylic acids, 3-aminopropionic acid (β-Ala), 7-aminohexanoic acid (Aha) and 11-aminoundecanoic acid (Aud) were selected as linkers of various sizes to optimize the interaction of the fluorescent ligand to the V1b receptor. Their spacer arms consist of 4, 8 and 12 atoms respectively, and will be referred to as 4A, 8A and 12A further in the text. The structures of the fluorescent peptides and of their parent analogues, designed according to the above rationale, are given in Fig 1 and Table I.

Fig 1.

General structure of fluorescent peptides (1–3, 5, 7, 9, 10, 11, 13) and parent peptides (A, 4, 6, 8, B, 12). X represents the modification on Cys¹ either by H atom or OH group, and Y the types of linker and fluorophore attached at position Lys⁸ of the parent peptides d[Leu⁴,Lys⁸]VP (A) and HO¹[Leu⁴,Lys⁸]VP (B).

Table 1.

Physicochemical Properties of Fluorescent Peptides 1–3, 5, 7, 9 – 11, 13, Parent Peptides A and B and of intermediates 4, 6, 8, 12.

No.	Peptide	Yielda %	TLC, Rfb				HPLCc TR, min	Formula	MW (calcd.)	MW (MS) (found)
			a	b	c	d
A	d[Leu4,Lys8]VPd	60.6	0.35	0.37	0.30	0.34	12.4	C47H67O11N11S2	1,026.3	1,026.4
1	d[Leu4,Lys(Atn)8]VPe	24.3	0.44	0.55	0.66	0.47	13.6	C54H72O12N12S2	1,145.4	1,145.8
2	d[Leu4,Lys(Alexa 488)8]VP	74.4	0.20	0.14	0.07	0.13	11.8	C68H77O21N13S4	1,540.7	1,540.5
3	d[Leu4,Lys(Alexa 647)8]VP	43.1	0.08	0.22	0.03	0.04	10.1	-	-	-
4	d[Leu4,Lys(β-Ala)8]VPe	67.9	0.21	0.40	0.12	0.48	15.0	C50H72O12N12S2	1,097.3	1,097.7
5	d[Leu4,Lys(β-Ala-Alexa 647)8]VP	64.4	0.17	0.29	0.02	0.12	10.1	-	-	-
6	d[Leu4,Lys(Aha)8]VPe	36.5	0.25	0.23	0.30	0.38	13.1	C54H80O12N12S2	1,153.5	1,153.4
7	d[Leu4,Lys(Aha-Alexa 647)8]VP 35.0	35.0	0.11	0.02	0.03	0.06	10.5	-	-	-
8	d[Leu4,Lys(Aud)8]VPe	41.9	0.40	0.51	0.53	0.53	14.4	C58H88O12N12S2	1,209.5	1,209.9
9	d[Leu4,Lys(Aud-Alexa 647)8]VP	55.9	0.18	0.26	0.12	0.17	11.7	-	-	-
B	[HO1][Leu4,Lys8]VP	54.2	0.26	0.22	0.18	0.38	11.1	C47H67O12N11S2	1,042.3	1,042.6
10	[HO1][Leu4,Lys(Alexa 488)8]VP	39.8	0.17	0.25	0.13	0.22	11.3	C68H77O22N13S4	1,556.7	1,556.6
11	[HO1][Leu4,Lys(Alexa 647)8]VP	49.6	0.16	0.28	0.03	0.14	11.0	-	-	-
12	[HO1][Leu4,Lys(Aud)8]VPe	54.4	0.38	0.51	0.53	0.50	12.7	C58H88O13N12S2	1,225.6	1,226.0
13	[HO1][Leu4,Lys(Aud-Alexa 647)8]VP	55.1	0.14	0.28	0.09	0.15	11.2	-	-	-

^aYields are based on the amount of peptide used in the reaction and are uncorrected for acetic acid, TFA and water content. Yields of peptides 3, 5, 7, 9, 11 and 13 are approximated and based on the MW (~ 1250) of Alexa 647 given by Invitrogen/Molecular Probes (pending patent).

^bSolvent systems and conditions are given in Experimental section.

^cAll peptides were at least 95% pure. For elution, a linear gradient 90: 10 to 30: 70 (0.05% aqueous TFA: 0.05% TFA in CH₃CN) over 30 min, with flow rate 1.0 mL/min was applied.

^dd[Leu⁴,Lys⁸]VP (A) was resynthesized as previously described^3,19: for original synthesis see Ref. ²⁰.

Abbreviations of position 8 substituents and linkers are: Atn (anthraniloyl), β-Ala (β-Alanine), Aha (7-aminoheptanoic acid) and Aud (11-aminoundecanoyl).

Table II.

Binding properties of fluorescent vasopressin analogues for human vasopressin and oxytocin receptors.

Analogue	Peptide	Affinity Ki for [3H]AVP (nM)				hV1b-R Selectivity Index (S.I.)
		hV1a-R	hV1b-R	hV2-R	hOT-R	V1a/V1b	V2/V1b	OT/V1b
AVP	AVPa	1.1 ± 0.1	0.68 ± 0.01	1.2 ± 0.2	1.7 ± 0.5	1.6	1.8	2.5
A	d[Leu4, Lys8]VPa	69 ± 16	0.52 ± 0.07	6,714 ± 806	29 ± 6	133	12,912	56
1	d[Leu4, Lys(Atn)8]VP	246 ± 53	0.65 ± 0.22	4,918 ± 1,447	21 ± 5	378	7566	32
2	d[Leu4, Lys(Alexa488)8]VP	258 ± 36	1.2 ± 0.2	> 100,000	30 ± 9	215	> 83,333	25
3	d[Leu4, Lys(Alexa647) 8]VP	12,033 ± 1,903	165 ± 16	> 100,000	36 ± 7.5	73	> 606	0.2
4	d[Leu4, Lys(β-Ala)8]VP	215 ± 50	1,3 ± 0,5	5,200 ± 190	12.6 ± 2.1	165	4,000	9.7
5	d[Leu4, Lys(β-Ala–Alexa647)8]VP	8,011 ± 2,035	65 ± 11	> 100,000	62 ± 9	123	> 1,538	1
6	d[Leu4, Lys(Aha) 8]VP	266 ± 28	0.61 ± 0.05	22,010 ± 3,310	22 ± 7	436	36,082	36
7	d[Leu4, Lys(Aha–Alexa647)8]VP	2,158 ± 369	74 ± 6	70,705 ± 17,325	121 ± 16	29	955	1.6
8	d[Leu4, Lys(Aud)8]VP	384 ± 93	3.7 ± 1.5	31,100 ± 7,900	210 ± 67	104	8,405	57
9	d[Leu4, Lys(Aud–Alexa647)8]VP	1,930 ± 326	13.0 ± 3.9	> 100,000	86 ± 27	148	> 7,692	6.6
B	HO[Leu4, Lys8]VP	879 ± 123	1.7 ± 0.5	1,558 ± 137	179 ± 20	517	934	105
10	HO[Leu4, Lys(Alexa488)8]VP	55,300 ± 11,250	149 ± 26	> 100,000	963 ± 197	371	> 336	671
11	HO[Leu4, Lys(Alexa647)8]VP	29,300 ± 5,700	1,799 ± 255	> 100,000	324 ± 29	16	> 56	0.2
12	HO[Leu4, Lys(Aud)8]VP	2,790 ± 381	8.4 ± 2.7	45,000 ± 13,400	3,742 ± 713	332	5,357	445
13	HO[Leu4, Lys(Aud–Alexa647)8]VP	> 100,000	668 ± 89	> 100,000	3,415 ± 965	> 150	> 150	5.1

Binding assays were performed on plasma membranes derived from CHO cells stably expressing the different VP/OT receptors as described in the experimental section. K_i values are the mean ± SEM of at least three independent experiments, each performed in triplicate. For each analogue, the hV_1b-R S.I. was calculated as follows: S.I. = (K_i analogue for hV_x-R)/(K_i analogue for hV_1b-R), where hV_x-R is the hV_1a-R, hV₂-R or hOT-R. Abbreviations of the linkers were β-Ala (β-Alanine, 4-A), Aha (7-aminoheptanoic acid, 7-A) and Aud (11-aminoundecanoyl, 12-A) and Atn is for Anthraniloyl.

^aData from reference 3.

Table III.

Functional properties of some fluorescent analogues on phospholipase C activity in CHO cells stably expressing human V1b or OT receptors.

Analogue #	Peptide	CHO hV1b-R		CHO hOT-R
		EC50 (nM)	Emax (%)	EC50 (nM)	Emax (%)
AVPa OT		2.0 ± 0.6 24.5 ± 2.6	100 31 ± 1	8.3 ± 2.2 10.8 ± 1.3	27.5 ± 1.8 100
A	d[Leu4,Lys8]-VPa	1.5 ± 0.4	99 ± 6	nd
1	d[Leu4,Lys(Atn)8]-VP	2.48 ± 2.1	108 ± 8.5	nd
2	d[Leu4,Lys(Alexa 488)8]-VP	6.93 ± 4	93.1 ± 4.6	nd
3	d[Leu4,Lys(Alexa 647)8]-VP	401 ± 14	112 ± 9	70.6 ± 19	73.4 ± 5.2
7	d[Leu4,Lys(Aha-Alexa 647)8]-VP	267 ± 44	106.7 ± 4.3	nd
9	d[Leu4,Lys(Aud-Alexa 647)8]-VP	11.4 ± 2.6	102 ± 2	172 ± 22	69.7 ± 3.05

Phospholipase C stimulation was performed on CHO cells stably expressing hV1b-R or hOT-R. For each analogue, EC₅₀ values and E_max were calculated as described in Methods. Results are the mean ± SEM of at least 3 independent experiments, each performed in triplicate.

^aData from reference ³.

Results

Binding pharmacological properties of fluorescent analogues

We first characterized the binding properties of each analogue and compared their affinities to those of AVP or of the corresponding parent peptides A or B taken as references (Fig 2, Table II). Attaching the Atn fluorophore directly to the d[Leu⁴,Lys⁸]VP (A) to give d[Leu⁴,Lys(Atn)⁸]VP (analogue 1) did not affect the affinity for the hV1b-R (0.65 vs 0.52 nM) and poorly changed the selectivity versus the hV1a, hV2 or hOT receptors. Thus, since there was no major decrease in affinity or selectivity, we did not perform further modifications of analogue 1.

Fig 2.

Binding pharmacological profiles of d[Leu⁴, Lys(Alexa 647)⁸]VP (analogue 3) and d[Leu⁴,Lys(Aud-Alexa 647)⁸]VP analogue 9 for human vasopressin and oxytocin receptors stably expressed in CHO cells lines.

The binding properties of analogues 3 and 9 were determined by competition experiments on plasma membranes prepared from CHO cells stably expressing hV1a-R, hV2-R, hV1b-R or hOT-R. Plasma membranes expressing the different VP/OT receptor isoforms were incubated 1 h at 30°C with ~1 nM ³[H]AVP in the presence or absence (control: C) of increasing concentrations of unlabeled analogues 3 or 9. Non specific and total binding were determined respectively with and without 1 μM of either AVP or OT, depending on the nature of the receptor studied. Specific binding calculated as the difference between total and non specific binding is expressed as percent of the corresponding control specific binding and is the mean ± SEM of at least 3 independent experiments each performed in triplicate.

For analogue 2, d[Leu⁴,Lys(Alexa 488)⁸]VP, with Alexa 488 replacing Atn, the affinity for the hV1b receptor was still very good, with a K_i of 1.2 nM compared to 0.52 nM for the parent analogue A. Its selectivities towards the hV1a-R and hV2-R were excellent (Table II). For analogue 3, d[Leu⁴, Lys (Alexa 647)⁸]VP, the affinity for the hV1b-R dropped dramatically to 165 nM (Table II and Fig 2), probably due to the relative size of the fluorescent moiety. The Alexa 647 fluorophore is indeed about twice as big as Alexa 488, thus conferring a MW of ~ 2200 to analogue 3 vs 1500 for analogue 2. In spite of its loss of affinity, analogue 3 retained a good selectivity for the hV1b versus the hV1a and hV2 receptors. Moreover, both fluorescent V1b analogues 2 and 3 exhibited a good affinity for the hOT-R: 30 nM for analogue 2 (Alexa 488) and 36 nM for analogue 3 (Alexa 647). Introducing an OH group at position 1 of analogue A to give B ([HO] [Leu⁴,Lys⁸]VP) reduced the affinity of this analogue for the hOT-R (K_i of 179 nM for analogue B vs 29 nM for analogue A. Thus, we synthesized analogues of B coupled to Alexa 488 (analogue 10) and to Alexa 647 (analogue 11). Introducing an Alexa 488 or Alexa 647 to (B) led to decreases of the affinities for the hOT-R of the resultant fluorescent peptides 10 and 11. For the Alexa 488 analogue (10), the K_i for hOT-R was 963 nM compared with 30 nM for analogue 2. The Alexa 647 derivative (analogue 11) exhibited a K_i for hOT-R of 324 nM compared with 36 nM for analogue 3. However, the affinities of peptides 10 and 11 for hV1b-R also dropped dramatically (from 1.2 to 149 nM for the Alexa 488 analogues (2 and 10) and from 165 to 1799 nM for the Alexa 647 analogues (3 and 11).

In order to improve the affinity of the Alexa 647 analogues for the hV1b-R, we introduced linkers of different lengths. We first verified that attaching linkers to the reference analogue d[Leu⁴,Lys⁸]VP (A) (Table I) maintained the affinity of the non fluorescent analogues for the hV1b-R within the nanomolar range (Fig 3 and Table II). Adding a 4A-spacer (3-aminopropionyl) βAla to d[Leu⁴,Lys(Alexa647)⁸]VP analogue 3 to give d[Leu⁴,Lys(β-Ala-Alexa 647)⁸]VP (analogue 5) improved its affinity (65 nM for analogue 5 compared with 165 nM for analogue 3) but did not improve its selectivity for hV1b versus hOT receptors (Selectivity Index of 1 for analogue 5 vs 0.2 for analogue 3). Increasing the length of the spacer to 8 atoms with a 7-aminohexanoyl (Aha) linker to give d[Leu⁴,Lys(Aha-Alexa 647)⁸]VP (analogue 7) modestly improved the affinity and selectivity of analogue 7. Further increasing the length of the spacer to 12 atoms with an 11-aminoundecanoyl (Aud) linker to give d[Leu⁴,Lys(Aud-Alexa 647)⁸]VP (analogue 9) significantly improved the affinity and selectivity of analogue 9. The K_i was 13 nM for analogue 9 as compared to 165 nM for analogue 3. This improvement might result from the fact that the high MW fluorescent moiety is now more distant from the core peptide (A) and does not hamper the ligand-receptor interaction as much (Fig 3). The selectivity was also much better (S.I. of 6.6 compared to 0.2, Table II and Fig 2). In order to try to gain more V1b-R selectivity vs OT-R, an OH-version of analogue 9, (HO [Leu⁴,Lys(Aud-Alexa 647)⁸]VP, analogue 13), was also synthesized. The affinity severely dropped (from 13 nM to 668 nM), without any improvement in its selectivity. We thus decided to forego using HO [Leu⁴,Lys⁸]VP (Peptide B) for any further designs of fluorescent V1b ligands.

Fig 3.

Influence on the size of the linker between Alexa 647 and peptide A on its V1b affinity and selectivity for human vasopressin and oxytocin receptors.

Panel A: The V1b affinity of parent analogues A, 4, 6, and 8 and of their corresponding peptides with Alexa 647 attached to linkers of different lengths (analogues 3, 5, 7, 9) are plotted as the number of aliphatic carbons of the spacer bound to Lys⁸. Values are from Table II.

Panel B: The hV1b-R Selectivity Index of analogues described above is plotted as the number of atoms of the spacer bound to Lys⁸. Values are from Table II with V1b Selectivity index calculated as log [K_i hVx-R/K_i hV1b-R]) where Vx is hV2-R, hV1a-R or hOT-R.

Functional properties of fluorescent peptides: capacity to stimulate phospholipase C

As both the hV1b and hOT receptors are directly coupled to phospholipase C (PLC)²⁶, we measured the ability of each fluorescent analog to activate PLC in stably transfected CHO cells. First we looked at the capacity of the most promising analogues to stimulate PLC in a CHO hV1b-R cell line. As illustrated in Figure 4 and summarized in Table III, with the exception of OT, which behaved as a partial agonist of the hV1b-R (31% of maximal VP-stimulated PLC activation), all analogues tested were full agonists with analogue 3 having a higher EC₅₀. The Atn-analogue (1) and the Alexa 488 analogue (2) showed EC₅₀ in the same range as analogue A and AVP (Table III). The Alexa 647 analogues 3, 7 and 9, although full agonists, showed EC₅₀ values of 401 nM for analogue 3 (Alexa 647 directly attached to analogue A) and 267 nM for analogue 7 (having the 8A, Aha linker). Only the 12A Aud-peptide (analogue 9) displayed a good EC₅₀ of 11.4 nM for hV1b-R coupling. Secondly, as the Alexa 647 compounds displayed some affinity for the hOT-R too, we also compared their capacities to activate PLC in a CHO hOT-R cell line. We observed that, as compared to OT, the natural ligand AVP was a partial agonist (27.5% of maximal OT-stimulated PLC activation). Analogues 3 and 9 also displayed partial agonism, with ~70% of PLC activation as compared to the one induced by the natural agonist OT. In general, the respective EC₅₀s of all analogues for either the hV1b-R or the hOT-R remained well correlated with their binding properties (compare Tables II and III).

Fig 4.

Functional properties of d[Leu⁴,Lys(Alexa 647)⁸]VP (analogue 3) and d[Leu⁴,Lys(Aud-Alexa 647)⁸]VP (analogue 9) on phospholipase C activity in CHO cells stably expressing the human V1b and OT receptors.

Left panel: CHO cells stably expressing the hV_1b-R were incubated with or without (control) increasing amounts of fluorescent or parent peptides. Total Inositol phosphates InPs which accumulated were measured as described in Methods and expressed as % of maximal AVP stimulation (1 μM). Data are the mean ± SEM of 3 independent experiments, each performed in triplicate.

Right panel: Similar experiments were performed with CHO cells stably expressing the hOT-R. Total Inositol phosphates which accumulated were measured as described in Methods and expressed as the % of maximal OT stimulation (1 μM). Data are the mean ± SEM of 3 independent experiments, each performed in triplicate.

Labelling properties of fluorescent peptides evaluated by spectrofluorimetry and cell imaging

The Atn analogue 1, d[Leu⁴,Lys(Atn)⁸]VP, was designed to provide a fluorescent V1b selective ligand with blue emission that could be a possible partner of an Alexa 488 analogue to record experimental fluorescence resonance energy transfer (FRET) between V1b homodimers. Since UV confocal microscopes are not easily available, the Atn analogue 1 cannot be tested by imaging but by spectrofluorimetry of cell suspensions in cuvettes. Living CHO cells, stably, expressing hV1b receptors, were exposed to analogue 1 for 1h at 12°C, washed and evaluated for the fluorescence emitted at the wavelength (λ) of 412 nm by spectrofluorimetry. Complete excitation and emission spectra were established previously and the value of 412 nm was found to be the λ_max of experimental emission for analogue 1 whereas the λ_max of excitation was confirmed to be 343 nm (data not shown). As shown in Figure 5 (panel A), analogue 1 can display a saturable fluorescent binding in Dulbecco’s modified Eagles medium/bovine serum albumin/4-2-hydroxyl ethyl-l-piperazineethane sulfonic acid (DMEM/BSA/HEPES) buffer with a fluorescence (EC₅₀) of 1.7 nM, a value very close to the one obtained by binding experiments performed on membrane preparations derived from CHO cells stably expressing hV1b-R (0.65 nM). This binding (analogue 1 used at the saturating concentration of 10 nM) can be completely suppressed with an excess of the non fluorescent analogue A (1μM) demonstrating its specificity (Fig 5, panel A).

Fig 5.

Fluorescent properties of d[Leu⁴,Lys(Atn)⁸]VP (analogue 1), d[Leu⁴, Lys(Alexa 488)⁸]VP (analogue 2) and d[Leu⁴, Lys(Alexa 647)⁸]VP (analogue 3) on CHO cells stably expressing the human V1b receptor.

Column A: Histogram (top): Cells in suspension were incubated for 1h at 12°C with 10 nM of analogue 1 in the absence (dark bar) or in the presence (hatched bar) of 1 μM of d[Leu⁴,Lys⁸]VP (analogue A) and compared to control cells incubated without fluorescent ligand (white bar). Cells were washed 3 times with cold PBS and measured for fluorescent emission at 412 nm in cuvettes of spectrofluorimetry. Saturation curve (bottom): Aliquots of cells were incubated with increasing concentrations of analogue 1 for 1h at 12°C, washed 3 times and counted for fluorescence emission at 412 nm. The endogenous fluorescence at 412 nm of control cells, not exposed to fluorescent ligand, was deduced for each concentration. Column B: CHO hV1b-R cells grown on coverslips were incubated in the presence of analogue 2 (50 nM) or of analogue 3 (250 nM) for 1h at 12°C in the absence or in the presence of 1 μM of d[Leu⁴,Lys⁸]VP (analogue A). Confocal imaging was performed with Zeiss LSM510 Meta microscope using a x63 (NA1.4) objective with a Argon laser excitation at 488 nm and a BP505–530 emission filter for analogue 2 (green) and with a Helium/Neon laser excitation at 633 nm with a LP650 for emission for analogue 3 (purple). The bar represents 10 μm (zoom 4x).

Fluorescent derivatives of the V1b agonist d[Leu⁴,Lys⁸]VP (A) were designed to provide tools for imaging the V1b-R on cells in culture. To verify this point, confocal imaging was conducted on living CHO stable cell lines, expressing either one of the 4 human receptors i.e. V1a-R, V1b-R, V2-R or OT-R. Only CHO cells expressing the hV1b-R and the hOT-R could be labelled. The CHO hV1a-R and the CHO hV2-R remained unlabelled, even at the concentration of 500 nM of either Alexa 488- or Alexa 647 analogues (data not shown). In CHO hV1b-R, the binding appeared to be patchy and mainly localized at the plasma membrane (Fig 5 panel B and Fig 6 right panels). This “patchy” effect was not due to internalization since the experiments were conducted at 12°C, a temperature below the chain-melting transition temperature of the membrane lipids (16°C). This aspect was also observed after one night of incubation at 4°C, a condition where internalization cannot take place either (data not shown). Moreover, the distribution of the fluorescent labelling with either Alexa 488 or Alexa 647 compounds was found to be similar and was completely suppressed in the presence of 1 μM of analogue A (Fig 5 panel B).

Fig 6.

Displacement of d[Leu⁴,Lys(Alexa 647)⁸]VP (analogue 3) and of d[Leu⁴,Lys(Aud-Alexa 647)⁸]VP (analogue 9) binding with non fluorescent ligands on CHO hV1b-R and hOT-R stable cell lines.

CHO cells on coverslips were incubated in the presence of analogue 3 (250 nM) or of analogue 9 (150 nM) for 1h at 12°C in the absence or in the presence of 100 nM of the selective V1b agonist d[Leu⁴,Lys⁸]VP (analogue A), or of 1μM of the non peptide AVP V2 antagonist 15²⁹ or of 500 nM of the V1a/OT peptide antagonist Manning Compound27a. Confocal imaging was performed with Zeiss LSM510 Meta microscope using a x63 (NA1.4) with laser excitation at 633 nm and LP650 for emission. The bar represents 10 μm (zoom 4x). Images are representative of 3 independent experiments.

Fig 6 further shows the ability of the fluorescent analogues 3 and 9 to selectively label hV1b or hOT receptors at the plasma membrane of stably transfected CHO cells. Plasma membrane labelling observed on CHO hV1b-R cells incubated with 250 nM of analogue 3 (a mixed OT/V1b agonist, see Table II) is totally suppressed by co-incubation with 100 nM of analogue A but only very partially when similar experiments were performed on CHO hOT-R cells (Fig 6, top rows). The complete displacement of analogue 3 from CHO hOT-R was obtained only with 500 nM or more of A, in agreement with its respective affinities for hV1b-R (0.52 nM) as compared to hOT-R (29 nM)³. As expected, [1-β-mercapto-β,β-cyclopentamethylenepropionyl, 2-0-methyl tyrosine, 8-arginine]vasopressin (Manning Compound) (500 nM), a mixed OT/V1a antagonist^26,27a.b, totally displaced analogue 3 from CHO hOT–R cells but not from CHO hV1b-R cells. Similarly, total displacement of analogue 3 from CHO hOT-R cells was also obtained with a specific non peptide OT antagonist, 4-Chloro-3-[(3R)-(+)-5-chloro-1-(2,4-dimethoxybenzyl)-3-methyl-2-oxo-2,3-dihydro-1H-indol-3-yl]-N-ethyl-N-(3-pyridylmethyl)-benzamide, hydrochloride (14) (SR126768A)²⁸ (100 nM), and this displacement was not seen on CHO hV1b- R cells (not shown). Analogue 3 labelling was not displaced by the selective non peptide AVP V2 antagonist (1-[4-N-tert-butylcarbamoyl)-2-methoxybenzene.sulfonyl]-5-ethoxy-3-spiro-[4-(2-morpholinoethoxy)cyclo-hexane]indol-2-one, phosphate monohydrate, (15) (SR121463)²⁹ (1 μM), neither from CHO hV1b-R nor from CHO hOT-R cells.

The labelling of hV1b-R CHO cells with 150 nM of analogue 9 (a selective V1b agonist, see Table II) was intense and completely displaced by 100 nM of A, but neither by 500 nM of Manning Compound nor by 1 μM of 15 (Figure 6, bottom row left). When similar experiments were performed on CHO hOT-R cells, the labelling was very low and both A (V1b selective) and Manning Compound (OT/V1a peptide antagonist) totally displaced analogue 9 labelling (Fig 6, bottom row right). However, this labelling was not displaced by the non peptide AVP V2 antagonist 15 (1 μM) neither from CHO hV1b-R nor from CHO hOT-R cells as previously shown for analogue 3.

In order to fully characterize the imaging properties of the two analogues 3 and 9, we have established saturation curves of each analogue in the conditions of living cells (DMEM/BSA/HEPES) on CHO cells expressing either V1b-R or OT-R. The membrane fluorescence was quantified with a laser scanning microscope (LSM) browser after confocal imaging using Helium/Neon laser 633 nm for detecting Alexa 647 analogues (Fig 7). Analogue 9, which displays a better affinity for the hV1b-R as compared to analogue 3 (as measured by competition studies on membrane preparations with ³[H]-AVP, see Table II), also showed a better capacity to bind membrane V1b-R sites on living cells at low concentration (affinity of the fluorescent ligand of 14.23 ± 2.19 nM for analogue 9 as compared to 172.3 ± 19 nM for analogue 3, Fig 7 left panel). On CHO hOT-R cells, the affinities of the two analogues were quite similar, with a slight preference for analogue 3 (109 ± 9.7 nM) as compared to analogue 9 (162 ± 17 nM) (Fig 7, right panel), reflecting the results obtained in the displacement studies performed with ³[H]-AVP (Table II).

Fig 7.

Fluorescent binding properties of d[Leu⁴,Lys(Alexa 647)⁸]VP (analogue 3) and d[Leu⁴, Lys(Aud-Alexa 647)⁸]VP (analogue 9) on CHO cells stably expressing the human V1b and OT receptors.

CHO hV1b-R cells on coverslips were incubated in the presence of increasing concentrations of analogue 3 or of analogue 9 for 1h at 12°C, washed 3 times with PBS and fixed with 4% PFA. Confocal imaging was performed with a Zeiss LSM510 Meta microscope using a x63 objective with a Helium/Neon laser excitation at 633 nm with a LP650 for emission. Membrane fluorescence was evaluated on 30–40 cells for each concentration of analogue by LSM Browser and ImageJ using the 1–250 fluorescent unit dynamics. Saturation curves were constructed with the means of 3 distinct experiments in which membrane fluorescence of 30–40 cells was evaluated for each concentration (after deduction of the control unlabelled cells) and expressed as % of the maximal labelling.

Discussion and Conclusion

Receptors of the vasopressin and OT family are important in the regulation of the stress processes (review in³⁰). Centrally, the V1b receptors have been involved in stress and especially in learning and memory processes. Important data have been obtained by the use of knock out (KO) animals but, after a period of cloning and pharmacological characterization in the last decade, it became necessary to elucidate the distribution of these receptors to better understand their central functions in vivo.

To do so, tracers are necessary and the radiolabelled natural hormones were the first candidates to detect receptors in situ. However, due to the close affinity between AVP and OT for the OT receptors^26,31, it is very difficult to discriminate between their receptor sites. Other studies were conducted using receptor antibodies but, since no good antibody against the V1b-R is available, the information about V1b receptor localization remains uncertain. Moreover, data obtained by in situ hybridization experiments were not always reliable, due to the partial similarities between the VP and OT receptor probes,^9,10 and reflect mRNA levels encoding the receptor and not the expression of the receptor itself.

Agonists and antagonists for different classes of receptors have been produced but, still, all were not completely selective, especially towards the V1b-R27b. A selective V1b agonist was produced by replacement in AVP of the glutamine residue in position 4 by a leucine¹⁸ and made highly resistant to degradation by deamination at the Cys¹ position. The subsequent replacement of Arg⁸ by a Lys⁸, leading to the V1b agonist A^3,19,20, did not modify its V1b selectivity. This allowed the covalent addition of a fluorophore to the free epsilon NH₂ group of the Lys residue at position 8 in A, thus making possible the synthesis of new fluorescent V1b agonists.

Taking advantage of the progress in confocal microscopy and in the production of new fluorophores, we elected to design and synthesize the fluorescent V1b analogues reported here. We selected 3 different colours for complementary purposes. The Atn, successfully attached to the mu opioid receptor agonist DALDA (H-Dmt-D-Arg-Phe-Lys-NH₂ with Dmt=2′,6′-dimethyltyrosine)^32,33, is a good reporter of environment²³ displaying a relatively small size. Thus, we confirm here that the Atn fluorophore does not prevent d[Leu⁴,Lys(Atn)⁸]VP (1) to bind to the V1b-R with a similar affinity as the parent peptide A (0.65 vs 0.52 nM). It may be a good compound to probe ligand binding interactions with V1b receptors and to confirm the data about the vasopressin V1b binding pocket we have obtained by modelling²¹. It may also be a good donor to a green acceptor in FRET experiments for identification of V1b homodimers in vivo. Thus, although other receptors of the vasopressin/oxytocin family are know to be expressed as homo-(OT-R/OT-R)^34,35 or heterodimers (V1a-R/V2-R)^36,37, nothing is known about V1b monomers or dimers. Only heterodimers of V1b-R with corticotrophin-releasing hormone receptor type 1 (CRHR1) have been described³⁸. We have also chosen the “Alexa family” of fluorophores (Molecular Probes) that develop very good brightness, good resistance to degradation and to photobleaching, as compared to Cyanine 3 and 5 for example^24,39, and that are relatively easy to attach to peptides. The Alexa 488 was selected because it is excited in the 488 nm range, a wavelength very common on any fluorescent microscope, including the electrophysiological settings, where it can be an excellent tool to reveal cells to be recorded (data not shown). It was also a good acceptor for energy transfer with Atn (good overlapping of the emission of Atn/excitation of Alexa 488 spectra), so that we could conduct studies on V1b-R homodimers in natural tissues later. Alexa 647 was also selected because it provides an extremely bright signal (ε=239 000) as compared to other Alexa fluorophores including Alexa 488 (ε=165 000), a quality that will be essential for detecting low levels of receptors in tissue slices. As Alexa 647 excitation is in the far red, cell auto-fluorescence is also very low.

The pharmacological properties (binding, coupling to PLC) were determined for the thirteen analogues and compared with the parent peptide d[Leu⁴,Lys⁸]VP (analogue A) and to the two natural peptides AVP and OT (Table II). First it appears that all analogues conserved a very good selectivity for V1b-R vs V1a-R and V2-R. Attaching the fluorophore directly, reduced slightly the selectivity, from 56 fold in d[Leu⁴,Lys⁸]VP (analogue A) to 32 fold (analogue 1, Atn) and to 25 fold in analogue 2 (Alexa 488). The selectivity was totally inverted for analogue 3 (Alexa 647), becoming 5 times better for the OT-R as compared to the V1b-R. As a possible explanation, it has been previously proposed that in the OT molecule, compared to AVP, the neutral Leu⁸ is an important amino acid to keep OT-R selectivity; while this position has to be positively charged to interact with vasopressin receptors^40,41. As the Lys⁸ of d[Leu⁴,Lys⁸]VP (A) (that has replaced the Arg⁸ naturally present in AVP) is now neutralized by attaching a bulky residue on the free epsilon NH², it may explain the modification of V1b-R towards OT-R selectivity. It is also possible that the bulky fluorescent molecule (especially big in the case of Alexa 647) prevents the 1–6 sequence of the peptide A to fit into the binding pocket situated at the top of TM5 and TM7 as described in our V1b-R model^21,42. To overcome this problem, we added linkers of different sizes, 4A (βAla), 8A (Aha) and 12A (Aud). These linkers, comprised of carbon chains of different lengths, all have a C-terminal NH2 group, to which a fluorophore could be covalently attached. We observed that, as shown in Fig. 3A, the longer the linker, the more the affinity of the Alexa 647 analogue for the V1b-R becomes closer to that of the parent peptide A. This affinity reaches 13 nM for analogue 9, as compared to 0.52 nM for A and compared to 3.7 nM observed for analogue 8, which corresponds to an analogue of A simply connected to the Aud linker with no fluorophore. The selectivity towards the hV1b-R also increased. Thus analogue 3 (Alexa 647 without a linker) is more selective for the hOT-R compared to the hV1b-R (36 nM vs 165 nM). With the Aud linker (Analogue 9), it became more selective for the hV1b-R (13 nM) as compared to the hOT-R (86 nM). Thus, we totally inverted the compound selectivity (Fig 3B). The selectivity for the hV1a-R and the hV2-R was not strongly modified and both analogues 3 and 9 remained low-affinity agonists for these two receptor isoforms (over 2000-fold).

We also noticed that replacing one of the hydrogens on the alpha carbon at position 1 by a hydroxyl group in peptide A, to give Peptide B, enhanced the hV1bR/hOTR selectivity. Thus, for analogue B, the affinity for the hV1b-R was conserved (1.7 nM vs 0.52 nM), whereas the hV1bR/hOTR selectivity was doubled (V1b-R S.I. of 105 instead of 56). We thus synthesized OH-substituted versions of the analogues coupled to Alexa 488 or Alexa 647 (Peptides 10–13). Unfortunately, although the affinity for the hOT-R remained weak, the affinity for the hV1b-R also dropped (to 149 nM for the Alexa 488 analogue (10) and to 1799 nM for the Alexa 647 analogue (11). Adding a 12A linker (analogue 13) did not improve the affinities, neither for the hOT-R (3415 nM) nor for the hV1b-R (668 nM). So, we didn’t pursue this track any further.

To fully characterize the analogues, functional coupling was evaluated for the most promising compounds (Table III). The activation of PLC was measured for the parent peptide A and the fluorophore-coupled analogues: Atn (analogue 1), Alexa 488 (analogue 2) and Alexa 647 (analogue 3) and with linkers (analogues 7 and 9). The activity was compared to the corresponding native peptides AVP and OT in 2 different stable cell lines: one expressing hV1b- R, the other hOT-R. First, it appears that the EC₅₀s were in the same range as compared to the K_i values. All analogues (original A, Atn-, Alexa 488- or Alexa 647-coupled analogues) were full agonists at the hV1b-R compared to the natural peptide AVP (Table III). Only OT displayed partial agonism. The two Alexa 647 analogues 3 and 9 were also tested on CHO hOT-R and both displayed partial agonism, with ~70% of activity as compared to 100% observed with OT. AVP was also a partial agonist at hOT-R with an E_max of only 27.5 %. It is well known that the OT-R can couple to G_q (PLC) and also to G_i^40,41, which is not the case for the V1b-R. These data confirm that OT can be a partial agonist for AVP receptors and on the contrary, AVP can be a partial agonist for the OT-R. Since it is known that there is a functional antagonism between AVP and OT, the first (AVP) being released during the stress and the second (OT) being present in nursing and calming behaviours, each acting at its own receptors, the partial agonism shown here could play a role in this functional antagonism by competing for the opposite site and preventing the “opposite effect” hormone to interact with its own receptor, thus reinforcing the physiological effects.

One can notice that the membrane hV1b receptors are labelled as patches. This patchy effect could reveal domains of the membrane where the receptors are concentrated. However, it doesn’t seem to be due to the fact that the ligand is an agonist, since it also happened with other peptides from the vasopressin family including the V1a antagonist Rhm⁸-PVA (4OH-Ph(CH₂)₂CO-DTyr(Me)-Phe-Gln-Asn-Arg-Pro-Lys(5-carboxytetramethylrhodamyl)-NH₂) on rat hepatocytes cells¹⁵, V1a/OT derivatives attached to Rhm (tetramethylrhodamyl)²² and with other ligands such as Alexa 488-opioid dermorphin and deltorphin⁴³ on CHO cells. However, this aspect was similar to the one observed on living cells before paraformaldehyde (PFA) fixation (not shown) and we preferred this technique to the one using antibodies, where fixation and Triton permeation are necessary for receptor detection and may give distortions. Thus we depict here functional receptors, capable to activate PLC and to internalize.

In this study, in contrast to studies performed only on membranes, we have validated the capacity of the fluorescent ligands 3 and 9 to label directly specific living cells and we have verified that the pharmacological and functional data fit well with the imaging data. Often, many new compounds lack full characterization on all VP and OT receptor isoforms and, as they display only a partial specificity (V1b S.I < 50), they could induce non specific labelling of several OT/AVP receptors27b. From our data, analogues 3 and 9 seem really promising for detecting human V1b or OT receptors. d[Leu⁴,Lys(Aud-Alexa 647)⁸]VP analogue 9 seems more selective for the hV1b-R but also presents some affinity for hOT-R. Utilized at the concentration of 150 nM and in the presence of 100 nM of the OT-R selective OT antagonist 14 (or of the Manning Compound), it provides a very selective labelling of the hV1b-R. d[Leu⁴,Lys(Alexa 647)⁸]VP (Analogue 3) presents a mixed affinity for hV1b and hOT receptors and can also be used in the presence of 100 nM of the non peptide OT antagonist 14 or of the Manning Compound to selectively label V1b sites. It can also be used to label OT sites in the presence of the selective non peptide AVP V1b antagonist (2S,4R)-1-[5-chloro-1-[2,4-dimethoxyphenyl)sulfonyl]-3-(2-methoxyphenyl)-2-oxo-2,3-dihydro-1H-indol-3yl]-4-hydroxy-N,N-dimethyl-2-pyrrolidine carboxamine (16) (SR149415)⁴⁴. Thus, the labelling of analogue 3 on CHO hOT-R cells was only completely suppressed with high concentrations of analogue A, more than 500 nM or with the non peptide V1b antagonist 16 at a concentration higher than 500 nM (not shown) at which it can also compete with hOT-R as described⁴⁴.

To conclude, by inserting different fluorophores at position 8 in d[Leu⁴,Lys⁸]VP (A), we generated two new promising fluorescent analogues; d[Leu⁴,Lys(Alexa 647)⁸]VP (3) and d[Leu⁴,Lys(Aud-Alexa 647)⁸]VP (9), capable of selectively labelling hV1b-R or hOT-R expressed on CHO cells. We completely characterized their conditions of utilization. In natural conditions on non fixed tissue, these fluorescent peptides 3 and 9, having the Alexa 647 fluorophore attached to the lysine residue at position 8, will allow the detection of very low levels of V1b or OT binding sites on native tissues in the CNS, with almost no background, and will permit, for the first time, the cartography of the V1b receptor within the brain.

Experimental Section

Materials

All reagents used were analytical grade. Most standard chemicals were purchased from Sigma (St. Louis, MO) or Merck (Darmstadt, Germany) unless otherwise indicated. AVP and OT were from Bachem (Bubendorf, Switzerland), [³H] AVP (44 Ci/mmol) and myo-[2-³H] inositol (20 Ci/mmol) from PerkinElmer (Courtaboeuf, France). Fetal calf serum and poly-ornithine came from Sigma (Saint-Quentin Fallavier, France). Dulbecco’s modified Eagle medium (DMEM) and penicillin-streptomycin were purchased from Invitrogen (Cergy Pontoise, France). Inositol-free DMEM came from ICN Biochemicals (Orsay, France). Dowex AG1-X8 formate form 200–400 mesh was purchased from Bio-Rad (Munich, Germany). All peptides and fluorescent derivatives listed in Table 1 and Manning Compound27a were synthesized in the laboratory of Dr. M. Manning (College of Medicine, University of Toledo, USA) as described below. The non peptide OT antagonist 14²⁸, the non peptide AVP V2 antagonist 15²⁹ and the non peptide AVP V1b antagonist 16⁴⁴ were from Sanofi-Synthelabo (Toulouse, France).

L-(−)-2-Hydroxy-3-methoxybenzylthiopropanoic acid (HO-Mpr(Mob) was synthesized as previously described for HO-Mpr(Bzl)⁴⁵. The Mpr(Meb) and Boc-11-aminoundecanoic acid were from Chem-Impex International (Wood Dale, IL). The N^α-Boc protected amino acids were purchased from Bachem (Torrance, CA) and Chem-Impex International (Wood Dale, IL). The Boc-Gly-resin was from Calbiochem-Novabiochem Corp. (San Diego, CA). The Alexa Fluor 488 carboxylic acid, 2,3,5,6-tetrafluorophenyl ester (Alexa Fluor 488 5-TFP) (5-isomer) and Alexa Fluor 647 carboxylic acid succinimidyl ester were from Invitrogen/Molecular Probes, Inc. (Eugene, OR). The N-Carboxyanthranilic anhydride (isatoic anhydride) and Boc-β-alanine (Boc-3-aminopropionic acid) were from Sigma-Aldrich, Inc. (St. Louis, MO). The Boc-7- aminoheptanoic acid (BOC-Aha) was from Bachem, (Torrance, CA). Peptide d[Leu⁴,Lys⁸]VP²⁰ (A) was resynthesized as previously described^3,19. Peptide [HO¹]Leu⁴,Lys⁸]VP (B) was synthesized specifically for this study by the Merrifield solid-phase method^46,47 with modifications previously described^48,49,3,19 as shown in Solid-Phase Synthesis Procedures below. TLC was run on precoated silica gel plates (60F-254, 3 Merck) with the following solvent systems: (a) 1-butanol/AcOH/H₂O (upper phase) (4:1:5); (b) 1-butanol/AcOH/H₂O (4:1:2); (c) 1-butanol/AcOH/H²O/pyridine (15:3:3:10); (d) 1-butanol/AcOH/H₂O (2:1:1). Loads of 10–15 μg each were applied and chromatograms were developed at a minimal length of 10 cm. For detection, a combination of monitoring on a UV lamp (model UVGL-58, UVP Inc., San Gabriel, CA) and the chlorine gas procedure for the KI-starch reagent⁴⁷ was used. Analytical HPLC was performed on a Hitachi D-7000 HPLC system under the following conditions: 90:10 to 30:70 0.05% aqueous TFA/0.05% TFA in MeCN, linear gradient over 30 min at 1.0 mL/min (λ = 254 or 214 nm), on a Vydac 218TP54 C18 column (Grace Vydac, Hesperia, CA). For semipreparative HPLC the same apparatus and gradient were used over 30 min at 5.0 mL/min (λ = 214 nm) on a Vydac 218TP510 C18 column. All peptides were at least 95% pure. Mass spectra (MS) were done by the Tuft’s Core Facility, Physiology Department (Boston, MA) on a MALDI-TOF Voyager mass spectrometer (Perspective Biosystems/Applied Biosystems) using dihydrobenzoic acid as the matrix.

Solid-Phase Synthesis Procedures

The schematic structures of d[Leu⁴,Lys⁸]VP (A) and parent molecules including all fluorescent analogues are illustrated in figure 1. Their synthesis will be described below.

HO-Mpr(Mob)-Tyr(Bzl)-Phe-Leu-Asn-Cys(Mob)-Pro-Lys(2CIZ)-Gly-NH₂ (Protected precursor for peptide B (PrB)

Boc-Gly-resin (0.7 g. 0.5 mmol) was subjected to eight cycles of deprotection, neutralization and coupling with Boc-Lys(2CIZ), Boc-Pro, Boc-Cys(Mob), Boc-Asn-ONp, Boc-Leu, Boc-Phe, Boc-Tyr(Bzl), and HO-Mpr(Mob), respectively. A 1 M HCl/AcOH mixture was used in all the deprotection steps⁴⁶. Neutralizations were carried out with 10% Et₃N/CH₂Cl₂. All coupling reactions (except when Boc-Asn was involved) were performed by the DCC/HOBt procedure⁵⁰ in CH₂Cl₂/DMF (9:1, v/v). Boc-asparagine was coupled as its nitrophenyl esters⁵¹ in DMF. The resulting protected peptidyl resin was cleaved by ammonolysis in methanol⁵². The protected peptide was extracted with hot DMF (30 mL), and the product was precipitated by the addition of hot water (ca 300 mL). After cooling, the product was collected, dried in vacuo over P₂O₅, and reprecipitated from DMF (30 mL) and ether (ca. 200 mL). Collection and drying in vacuo over P₂O₅ gave the required protected nonapeptide amide (PrB) 0.53 g 68.5% yield. M.p.112–114°C. TLC, R_f (solvent system): 0.73 (a); 0.78 (b); 0.70 (c); 0.77 (d).

[HO¹][Leu⁴,Lys⁸]VP (Free peptide (B)

The Na/liquid NH₃ procedure⁵³ with the modifications previously described^5,48,49,54 was used for the deprotection of the protected precursor PrB. A solution of PrB in sodium-dried ammonia (ca. 400 mL) was treated at the boiling point and with stirring with sodium from a stick of metal contained in a small-bore glass tube until a light-blue color persisted in the solution for about 30 s. NH₄Cl was added to discharge the colour. Reoxidation of the resulting deblocked disulphydryl peptide (B) was performed with potassium ferricyanide (K₃[Fe(CN)₆]) using the modified reverse procedure⁵⁵ as follows; the peptide residue was dissolved in 25 mL of 50% AcOH, and the solution was diluted with 50 mL of H₂O. The peptide solution was added dropwise with stirring over a period of 15–30 min to a 600 mL aqueous solution that contained 20 mL of a 0.01 M solution of K₃[Fe(CN)₆]. Meanwhile, the pH was adjusted to approximately 7.0 with concentrated ammonium hydroxide. Following oxidation, the free peptide (B) was isolated and purified as follows: after acidification with AcOH to pH 4.5 and stirring for 20 min with an anion exchange resin (Bio-Rad, AG 3 × 4, Cl⁻ form, 5 g damp weight), the suspension was slowly filtered and washed with 0.2 M AcOH (3 × 30 mL). The combined filtrate and washings were lyophilized. The resulting powder was desalted on a Sephadex G-15 column (110 × 2.7 cm), eluting with aqueous AcOH (50%), with a flow rate of 5 mL/h⁵⁶. The eluate was fractionated and monitored for absorbance at 254 nm. The fractions making up the major peak were checked by TLC, pooled, and lyophilized. The residue was further subjected to gel filtration on Sephadex LH-20 using 100 × 1.5 cm column, eluting with aqueous AcOH (2 M) with a flow rate of 4 mL/min. The peptide was eluted in a single peak (absorbance at 254 nm). After lyophilization of the pertinent fractions, the collected peptide was subjected to a final semi-preparative HPLC purification to give 54.9 mg (54.2%) of desired [HO¹]Leu⁴,Lys⁸]VP (B). TLC, R_f (solvent system): 0.26(a); 0.22(b); 0.18(c); 0.38(d). HPLC, T_R: 11.1 min. M.W. (calculated): 1042.3; M.W. (found by MS): 1042.6.

Fluorescent Analogue Synthesis

Peptide Labeling with Anthraniloyl (Atn) Fluorophore (Peptide 1, Table I)

We adapted the synthesis of Boc-Dap(Atn)²³ and Boc-Lys(Atn)³² for the preparation of d[Leu⁴,Lys(Atn)⁸]VP (peptide 1) as follows: To a stirred solution of 12.5 mg (12.2 μM) of d[Leu⁴,Lys⁸]VP (A) and 1.3 mg (12.2 μM) of dry Na₂CO₃ in water/acetonitrile (1:1 v/v, 150 μL), isatoic anhydride 2.38 mg (14.6 μM) in 100 μL of acetonitrile were added. The mixture was stirred overnight at room temperature with pH ~ 8 and TLC monitoring. The mixture was acidified to pH 2 with 5% aqueous KHSO₄ and was subjected to a two step purification procedure, first on a Sephadex G-15 column with 25% acetic acid as eluent and finally be semipreparative HPLC to give the desired peptide 1.

General Procedure for Peptide Labelling with Alexa 488 and Alexa 647 Fluorophores

Fluorescent peptides with Alexa fluorophore directly attached at position 8 of parent peptide d[Leu⁴,Lys⁸]VP (A) and [HO¹][Leu⁴,Lys⁸]VP (B) (Peptides 2, 3, 10 and 11 (Table I)

The preparation of the fluorescent peptides 2, 3, 10 and 11 was performed as described by Albizu and co-workers⁵⁷ with some modifications as follows: 1.2 μM (in excess) of parent peptide d[Leu⁴,Lys⁸]VP (A) or [HO¹][Leu⁴,Lys⁸]VP (B) was dissolved in 100 μL of anhydrous DMF, mixed with 5 μL of DIPEA and the solution was added to 1 mg of the appropriate fluorophore. For fluorescent peptides 2 and 3 the parent peptide was d[Leu⁴,Lys⁸]VP (A). For fluorescent peptides 10 and 11 the parent peptide was [HO¹][Leu⁴,Lys⁸]VP (B). The fluorophores were as follows: for peptides 2 and 10 Alexa 488 tetrafluorophenyl ester (5-carboxy pure isomer) was used. For peptides 3 and 11, Alexa 647 succinimidyl ester was utilized. The reaction mixture was whirled using vortex for 1 hour at room temperature in the dark with pH, TLC and HPLC monitoring. After the reaction was over, the mixture was acidified (6 μL TFA) and evaporated in vacuo. The fluorescent peptide was isolated by semi-preparative HPLC and lyophilized. Yields of 2, 3, 10 and 11 are given in Table I.

Fluorescent peptides with the Alexa 647 fluorophore attached at position 8 of parent peptide d[Leu⁴, Lys⁸]VP (A) by means of spacer (Peptides 5, 7, 9 and 13 (Table I)

All peptides 5, 7, 9 and 13 were prepared by a general two-step procedure as follows. First, the spacer-containing peptides d[Leu⁴,Lys(β-Ala)⁸]VP (4), d[Leu⁴,Lys(Aha)⁸]VP (6), d[Leu⁴,Lys(Aud)⁸]VP (8) and [HO¹][Leu⁴,Lys(Aud)⁸]VP (12) (Table I) were synthesized as described below. Next, the Alexa 647 fluorophore was attached to peptides 4, 6, 8 and 12 as described above to give the desired peptides 5, 7, 9 and 13.

Preparation of d[Leu⁴,Lys(Aha)⁸VP (6)

7.35 mg (30 μM) of Boc-7 aminoheptanoic acid (Boc-Aha), 7.35 mg (30 μM) of DCC and 4.1 mg (30 μM) of HOBt were dissolved in ca. 200 μL of dry DMF and the mixture was stirred for 30 min. Next, 23.6 mg (23 μM) of d[Leu⁴Lys⁸]VP (A) as a solution in ca. 150 μL of dry DMF and neutralized in advance with DIPEA was added and the reaction mixture was incubated overnight at room temperature with continuous stirring and pH~8.5 monitoring. The desired pH-value was kept using DIPEA. After the reaction was over (TLC and HPLC monitoring), the reaction mixture was separated on Sephadex G-15 column using 25% AcOH as an eluent. The fraction containing the peptide was collected and lyophilized. The Boc-protecting group of the peptide was removed by treatment with 50% TFA in CH₂Cl₂ for 30 min at room temperature. The reaction mixture was evaporated by a flow of N₂ and the resulting residue was subjected to a final purification by semi-preparative HPLC, to give the spacer-containing peptide d[Leu⁴,Lys(Aha)⁸]VP (6) with a yield of 9.7 mg (36.5%) (Table I). Spacer-containing peptides d[Leu⁴, Lys(β-Ala)⁸]VP (4) and d[Leu⁴, Lys(Aud)⁸]VP (8) were prepared by the same manner as peptide (6), using the parent peptide d[Leu⁴,Lys⁸]VP (A) and Boc-β-Ala and Boc-Aud, respectively. Peptide [HO¹]Leu⁴,Lys(Aud)⁸]VP (12) was prepared by the same approach, utilizing Boc-Aud and the parent peptide [HO¹][Leu⁴,Lys⁸]VP (B). Yields of (4), (8) and (12) are given in Table I.

Synthesis of d[Leu⁴,Lys(Aha-Alexa 647)⁸VP (7)

1.38 mg (1.2 μM) (in excess) of peptide d[Leu⁴,Lys(Aha)⁸]VP (6) was dissolved in 100 μL of anhydrous DMF, mixed with 5 μL of DIPEA. The solution was added to 1 mg of Alexa 647 succinimidyl ester and the reaction mixture was whirled using Vortex for 1 hour at room temperature in the dark at pH (~ 8), TLC and HPLC monitoring. After the reaction was over, the mixture was acidified (6 μL TFA) and evaporated in vacuo. The fluorescent peptide d[Leu⁴, Lys(Aha-Alexa 647)⁸]VP (7) was isolated by semi-preparative HPLC and lyophilized. Yield 1 mg (~ 35.0%) (Table I).

Peptides d[Leu⁴, Lys(β-Ala)⁸]VP (4), d[Leu⁴, Lys(Aud)⁸]VP (8) and [HO¹][Leu⁴, Lys(Aud)⁸]VP (12) were labelled utilizing Alexa 647 succinimidyl ester as described above for peptide (7) to give peptides d[Leu⁴, Lys(β-Ala-Alexa 647)⁸]VP (5), d[Leu⁴, Lys(Aud-Alexa 647)⁸]VP (9) and [HO¹][Leu⁴, Lys(Aud-Alexa 647)⁸]VP (13). Yields of (5), (9) and (13) are given in Table I. The structures of peptides 1, 2, 4, 6, 8, 10, and 12 were confirmed by MALDI-TOF mass spectroscopy (MS). The structures of the Alexa 647 containing peptides 3, 5, 7, 9, 11 and 13 were not confirmed by mass spectrometry; since the structure of the Alexa 647 fluorophore has not been revealed by Invitrogen/Molecular Probes, due to a pending patent application. The analytical and some other physicochemical data of the fluorescent peptides 1 – 3, 5, 7, 9 – 11, 13 and of the intermediates 4, 6, 8, 12 are presented in Table I.

Membrane preparation

Four different Chinese hamster ovary (CHO) cell lines stably expressing the human V1a, V1b, V2 or OT receptors were washed twice in phosphate buffer saline (PBS) without CaCl₂ and MgCl₂, harvested in lysis buffer (15 mM Tris-HCl, pH 7.4; 2 mM MgCl₂; 0.3 mM ethylene diaminetetraacetic acid (EDTA), polytron homogenized, and centrifuged at 44,000× g for 20 min at 4°C as previously described³. Pellets were washed in 50 mM Tris-HCl (pH 7.4), 3 mM MgCl₂ and centrifuged at 44,000 × g for 20 min at 4°C. Membranes were re-suspended in a small volume of the same buffer. Protein concentration was determined by the method of Bradford (Bio-Rad protein assay kit) using bovine serum albumin (BSA) as a standard and membranes were stored in liquid nitrogen.

Binding assays

Membrane incubations with [³H] AVP were performed as previously described³. Briefly, 10–20 μg membrane proteins were incubated for 60 min at 30°C in a medium containing 50 mM Tris-HCl (pH 7.4), 3 mM MgCl₂, 1 mg/ml BSA, 0.01 mg/ml leupeptine and 1 nM ³[H]-AVP in the presence (non specific binding) or in the absence (total binding) of 1 μM of unlabelled AVP in a final volume of 200 μL. For K_i determinations, increasing amounts of the unlabelled analogue to be tested were added in the incubation medium to a fixed concentration of tritiated ligand. The membrane-associated radioactivity was collected by filtration through GF/C filters and counted. Specific binding was calculated as the difference between total and non-specific binding values and expressed as percent of the specific binding determined without unlabelled analogue.

Phospholipase C assay

Inositol phosphate (InsPs) accumulation was determined as described previously³. CHO cells were seeded at 100,000 cells per well in 24-well plates and grown for 24h in their regular culture medium, then incubated for another 24h period in serum- and inositol-free medium supplemented with 2 μCi/ml myo-[2-³H] inositol. Cells were then washed twice with HBS, incubated for 15 min in HBS supplemented with 20 mM LiCl, and further stimulated for 15 min at 37°C in the same buffer with increasing concentrations of vasopressin analogues. The reaction was stopped by perchloric acid (5% vol/vol). Total InsPs were extracted, purified on Dowex AG1-X8 anion exchange chromatography column and counted.

Cell labelling and imaging

CHO cells, seeded on 12-mm glass cover slips pre-coated with poly-ornithine, were incubated in DMEM, BSA 0.2 mg/ml, HEPES 25mM, pH 7.4 for 1h at 12°C with fluorescent analogues in the absence or in the presence of non fluorescent VP analogues. In the latter case, a 30 min pre-incubation of the cells in the same medium containing only these ligands was performed. After 3 washes with cold PBS, the cells were finally fixed in 4% PFA at 4°C overnight and mounted with mowiol. The fluorescent cells were imaged using a confocal microscope Zeiss LSM510 Meta equipped with an Axiovert200M microscope. The objective 63x (NA 1.4) for oil immersion was used. For detecting Alexa 488-labelled cells, the excitation was performed with an Argon laser at the λ488 nm. The green emission was collected using a BP 505–530 emission filter. For detection of Alexa 647, Helium/Neon laser 633 nm was used. The emission was collected with a LP650 nm filter. Images were captured in multi-tracking mode to avoid channel crosstalk and the settings were established on unlabelled cells used as negative control. For fluorescence quantification, the Zeiss LSM Browser was used as previously described⁵⁸ with 30–40 cells evaluated for each fluorescent binding condition.

Alternatively, the binding of the Atn-labelled peptide (analogue 1) was evaluated by cuve spectrofluorimetry. Cells were detached with PBS, 10 mM EDTA and incubated with the fluorescent analogue in DMEM, 0.2 mg/ml BSA, 25 mM HEPES pH 7.4 at 12°C as previously described. After 3 cold washes with PBS and gentle centrifugation (900 g, 4 min), the cells were resuspended in PBS and placed in a cuvette at 4°C. The fluorescence was measured using a λ 343 nm excitation and a λ 412 nm emission in a Photon Technology International 810 spectrometer equipped with an UV lamp. These values were chosen after full excitation and emission curves had been established with soluble Atn-analogue 1. The specific detection was calculated by deducing the fluorescence emission measured in CHO hV1b-R cells incubated with an excess of non fluorescent ligand.

Data analysis

All data were were analyzed using the GraphPad Software, Inc. Prism (GraphPad Software, Inc., San Diego, CA). For the radioligand binding data, the inhibitory dissociation constant (K_i) for unlabeled vasopressin analogues was calculated from binding competition experiments according to the Cheng and Prusoff equation: K_i= IC₅₀ × (1+[L]/K_d), where IC₅₀ is the concentration of unlabeled analogue leading to half-maximal inhibition of specific binding, [L] the concentration of the radioligand present in the assay and K_d its affinity for the receptor studied. The respective Kd values of ³[H]-AVP for hV1a-R (1.1 nM), hV1b-R (0.68 nM), hV2-R (1.2 nM) and hOT-R (1.7 nM) had been previously determined in Pena et al (2007)³ and are listed in Table II (lane 1). For the PLC assay, results are expressed as the mean ± SEM of the number of distinct experiments indicated (n). Statistical analysis of the data was performed using the one-way analysis of variance test.

Abbreviations

Symbols and abbreviations are in accordance with the recommendations of the IUPAC-IUB Commission on Biochemical Nomenclature (Eur. J. Biochem. 1989, 180, A9–A11) and IUPHAR (Trends Pharmacol. Sci. 2001). All amino acids are in the L-configuration unless otherwise noted. Other abbreviations are the following

AcOH

acetic acid

ACTH

adrenocorticotropin hormone

β-Ala

3-aminopropionyl

Alexa 488

Alexa Fluor 488 carboxylic acid, 2,3,5,6-tetrafluorophenyl ester (5-isomer) and the related acyl group

Alexa 647

Alexa Fluor 647 carboxylic acid succinimidyl ester and the related acyl group (structure unknown, pending patent application)

Aha

7-aminohexanoyl

Atn

anthraniloyl

Aud

11-aminoundecanoyl

AVP

arginine vasopressin

Boc

tert-butyloxycarbonyl

Bzl

benzyl

cAMP

cyclic adenosine monophosphate

BSA

bovine serum albumin

CHO cells

Chinese hamster ovary cells

CHO hV1a-R, CHO hV1b-R, CHO hV2-R, CHO hOT-R

CHO cells stably expressing human vasopressin (VP), V1a, V1b, V2 or oxytocin (OT) receptors respectively

CRH (or CRF)

corticotrophin releasing hormone

CRHR1

corticotrophin-releasing hormone receptor type 1

2CLZ

2-chlorobenzyloxycarbonyl

DALDA

H-Dmt-D-Arg-Phe-Lys-NH₂, Dmt, 2′,6′-dimethyltyrosine

Dap

L-2,3-diaminopropionic acid

DCC

dicyclohexylcarbodiimide

DIPEA

N,N-diisopropylethylamine

d[Leu⁴

Lys⁸]VP, [deamino-Cys¹, 4-leucine, 8-lysine]vasopressin

DMF

dimethylformamide

DMEM

Dulbecco’s Modified Eagle’s Medium

ε

molar extinction coefficient

EC₅₀ or K_act

concentration of agonist leading to half-maximal activity

E_max

maximal efficiency

EDTA

ethylenediaminetetraacetic acid

EGFP

Enhanced green fluorescent protein

ESMS

electron spray mass spectrometry

Et₃N

triethylamine

FRET

Fluorescence Resonance Energy Transfer

Gi

G-protein type i

Go

G-protein type o

HOBt

1-hydroxybenzotriazole

HBS

HEPES buffer saline

HO¹[Leu⁴

Lys⁸]VP, [1-L-(−)-2-hydroxy-3-thiopropanoic acid, 4-leucine, 8-lysine]vasopressin

HO-Mpr

L-(−)-2-hydroxy-3-thiopropanoyl

HPLC

high-performance liquid chromatography

InsPs

total inositol phosphates

K_d

concentration of ligand leading to half-maximal specific binding deduced from saturation experiments

K_i

concentration of peptide leading to half-maximal specific binding deduced from competition experiments

KO

knock-out animals

λ

wavelength

MeCN

acetonitrile

LSM

laser scanning microscope

mowiol

mounting medium for fluorescence microscopy, mRNA, messenger ribonucleic acid

Manning compound

[1-β-mercapto-β,β-cyclopentamethylenepropionyl, 2-O-methyltyrosine, 8-arginine]-vasopressin (VP V1a/OT antagonist)

Mob

p-methoxybenzyl

Mpr

3-mercaptopropionyl

ONp

p-nitrophenyl ester

OT

oxytocin

OT-R

oxytocin receptor

PBS

Phosphate Buffer Saline

PFA

paraformaldehyde

PLC

phospholipase C

Rhm

tetramethylrhodamyl group

Rhm⁸-PVA

4-HO-Ph(CH₂)₂-CO-DTyr(Me)-Phe-Gln-Asn-Arg-Pro-Lys(5-carboxytetramethylrhodamyl]-NH₂, D-Tyr(Me), O-methyl-D-Tyrosine, RT-PCR, polymerase chain reaction after reverse transcription

SEM

Standard Error to the Mean

SI

selectivity index

TFA

trifluoroacetic acid

TLC

thin-layer chromatography

VP

vasopressin

V1a-R, V1b-R, V2-R

vasopressin receptor isoforms V1a (vascular), V1b (pituitary) and V2 (renal) respectively