Peptides.Vol. 13, pp. 787-793, 1992

0196-9781/92 $5.00 + .00 Copyright© 1992 PergamonPressLtd.

Printed in the USA.

Synthesis and Biological Activity of Novel CTerminal-Extended and Biotinylated Growth Hormone-Releasing Factor (GRF) Analogs ROBERT

M. C A M P B E L L , *l Y U N G LEE,* T H O M A S F. M O W L E S , * K I M W. M c I N T Y R E , t M U S H T A Q AHMAD,:]: A R T H U R M. FELIX:I: A N D E D G A R P. HEIMER:~

Departments of*Animal Science, -~Immunopharmacology, and ~cPeptide Research, Hoffmann-La Roche Inc., Nutley, NJ 07110 R e c e i v e d 27 F e b r u a r y 1992 CAMPBELL, R. M., Y. LEE, T. F. MOWLES, K. W. McINTYRE, M. AHMAD, A. M. FELIX AND E. P. HEIMER. Synthesis and biologicalactivityof novelC-terminal-extendedand biotinylatedgrowthhormone-releasingfactor (GRF)analogs. PEPTIDES 13(4) 787-793, 1992.--A series of novel hGRF(1-29)-NH2 analogs were synthesized and biotinylated. The immunological and biological activities of these analogs were then characterized. To distance the biotin moiety from the putative bioactive core, a C-terminal spacer arm consisting of -GIy-Gly-Cys-NH2 (-GGC) was added to hGRF(I-29)-NH2 (hGRF29) and analogs, with subsequent biotinylation performed at the cysteine residue. Neither addition of the C-terminal spacer arm nor biotinylation affected affinity of these analogs for GRF antibody. Relative to hGRF(I-44)-NH2 (hGRF44: potency = 1.0), the biotinylated analogs were equipotent in vitro to their nonbiotinylated, parent compounds: [desNH2Ty?,D-Ala2,Ala~5]hGRF29-GGC-(tpBiocytin)NH2 (4.7) = [Ala~5]hGRF29-GGC-(tpBiocytin)-NH2(3.9) > hGRF29-GGC-(tpBiocytin)-NH2 (0.8). Based upon cumulative GH release data in vivo (0-60 min postinjection), [desNH2Tyr~,D-Ala2,AlatS]hGRF29-GGC-(tpBiocytin)-NH2,[Ala~S]hGRF29-GGC(tpBiocytin)-NH2, and hGRF29-GGC-(tpBiocytin)-NH2 displayed 8.6, 5.5, and 0.8 times, respectively, the potency of hGRF44. These in vivo potency values were not significantly different from the corresponding parent compounds (i.e., with or without the C-terminal spacer arm). In summary, biotinylated hGRF analogs have been developed that retain full immunoreactivity and potent bioactivity (in vitro and in vivo), thus permitting their use in GRF receptor isolation, ELISA, and histochemical procedures. Growth hormone-releasing factor (GRF) Receptor probe

Biotinylated GRF analogs

BIOTIN, characterized as a water-soluble B vitamin, binds with extremely high affinity to avidin and streptavidin. As avidin/ streptavidin can be labeled with a variety of fluorescent, chromogenic, and isotopic agents, the biotin-avidin coupling may be utilized for diagnostic and microdetection applications. Recently, peptide growth factors/hormones have been biotinylated (1-5,16-19) for use in ELISAs, flow cytometry, histochemistry, and receptor isolation. Growth hormone-releasing factor (GRF) is a hypothalamic peptide whose receptor has not been isolated from any species. A potent biotinylated G R F analog could greatly facilitate this isolation. For receptor binding/isolation studies, it is essential to ensure that a G R F agonist retains growth hormone (GH)releasing activity following biotinylation (e.g., at the N-terminus, C-terminus, or free lysine residues). However, G R F bioactivity is relatively intolerant to N-terminal (Tyr ~) modifications (14). Furthermore, native lysine residues (Lys 12and Lys 21 of hGRF), preferred by biotin, are located in regions

Growth hormone (GH)

Rat anterior pituitary

associated with G R F receptor binding and biological activity (15). Acetylation of these lysine residues has been found to greatly reduce (+84%) the in vitro activity of G R F analogs (Campbell, Mowles, Heimer, and Felix, Unpublished data). Based upon these observations, it was postulated that G R F would require biotinylation at a site distant from the bioactive core in order to retain GH-releasing activity. Theoretically, a short spacer arm (-GIy-Gly-Cys-NH2; -GGC) could be added to the C-terminus of each parent compound and subsequently biotinylated at the distal cysteine residue (via thiol oxidation) without hindering biological activity. In the present studies, the objective was to synthesize and characterize novel biotinylated hGRF (human sequence) analogs for potential use as G R F receptor probes. This was accomplished using a three-step methodology: 1. Parent compound synthesis: hGRF29, [AlaJS]hGRF29, and [desNH2Tyr ~,D-Ala2,AlalS]hGRF29;

1Requests for reprints should be addressed to Dr. Robert M. Campbell, Hoffmann-La Roche, Inc., Department of Animal Science Research, 340 Kingsland Street/Bldg. 86, Nutley, NJ 07110.

787

788

CAMPBELL ET AL.

2. C-terminal extension: hGRF29-GGC-NH2, [AlalS]hGRF29 GGC-NH2, and [desNH2Tyrl,o-Ala2,AlalS]hGRF29-GGCNH2; and 3. Biotinylation of the C-terminal extension [addition of thiopropionyl-Biocytin-NH2, -(tpBiocytin)-NH2]: hGRF29-GC~(tpBiocytin)-NH2, [Alal 5]hGRF29-GGC-(tpBiocytin)-NH2, and [deSNHETyr~,D-Ala2,Ala~5]hGRF29-GGC-(tpBiocytin)NH2.

GRFa_c -Gly-Gly-Cys-NH 2

÷ H

Q

o coo o u t , , S-S-CH2.CH2.C.NH-CH.(CH214-NH-C.ICH2)4I

~

I

- NH

Following each step, the immunoreactivity and biological activity (in vitro and in vivo) of each analog were determined. Optimally, all compounds (nine in total) would be expected to retain full immunoreactivity and high GH-releasing activity. HS

METHOD

~N~

H

Parent and C-Terminal-Extended GRF Analog Synthesis Parent (hGRF29, [AlalS]hGRF29, and [desNH2Tyr~,D-Ala2,Ala~S]hGRF29) and C-terminal-extended (hGRF29-GCd2-NH2, [AI#5]hGRF29-GGC-NH2, and [desNH2Tyr~,D-Ala2,AIaIS]hGRF29-GGC-NH2) hGRF analogs were synthesized by solidphase methodology (6). The crude peptides were purified by preparative HPLC, determined to be homogeneous (>99% purity) by analytical HPLC, and the structure confirmed by amino acid analysis and FAB mass spectrometry.

GRF Analog Biotinylation For each of the C-terminal-extended hGRF29 analogs, biotinylation was performed using the following procedure (outlined in Fig. 1). hGRF29-GGC-NH2 analogs (3.35 #moo were each dissolved in 50% trifluoroethanol/H20 (4 ml). N(-2-pyridyl-dithiopropionyl)-Biocytin-NH2 (5.85 #mol; Calbiochem) in N,Ndimethylformamide (2 ml) was combined with the hGRF29GC_d2analog solution and stirred for 10 min; 0.2 M NH4HCO3 (4 ml; pH 8.0) was then added and stirred for an additional 45 min. The reaction mixture was acidified (to pH 4.0) by dropwise addition of glacial acetic acid and subsequently lyophilized (yield = 12.5 mg crude product and excess reagent). The residue was dissolved in 0.5 ml methanol, 0.1% trifluoroacetic acid (TFA)/ H20 (10 ml) added, and filtered (0.45 ~ Millex). The biotinylated hGRF29 product was purified by reverse-phase HPLC [C3 column, DuPont Protein Plus, 2.2 × 25 cm, 10 # particle size; mobile phase: (A) 0.1% trifluoroacetic acid in H20, (B) acetonitrile, in a linear gradient from 20-50% (B) in 60 min; flow rate: 1 ml/min, detection ~, = 206 nm). Eluent fractions containing hGRF29-GGC-(thiopropionyl-Biocytin)-NH2 were collected and evaporated, resulting in ~39% yield. hGRF(1-29)-Gly-Gly-Cys-(thiopropionyl-Biocytin)-NH2 [hGRF29-GCd2-(tpBiocytin)-NH2] was determined to be homogeneous by analytical HPLC and displayed the expected amino acid composition after acid hydrolysis (6 N HCI, 110°C for 24 h): Asp 2.8 (3); Thr 0.9 (1); Ser 2.9 (3); Glu 2.1 (2); Gly 3.1 (3); Ala 3.0 (3); Val 1.0 (1); Met 1.0 (1); lie 1.8 (2); Leu 3.9 (4); Tyr 1.7 (2); Phe 0.9 (1); Lys 2.8 (3); and Arg 2.9 (3). Further confirmation of structure was provided by HI-NMR (DMSOdr) and FAB mass spectroscopy [calculated: (M + H) ÷ 4033.8; found: 4034.0]. [AIaIS]hGRF(1-29)-Gly-Gly-Cys-(thiopropionyl-Biocytin)NH2 {[AlalS]hGRF29-GCd2-(tpBiocytin)-NH2} was homogeneous by analytical HPLC, exhibiting the expected amino acid composition after acid hydrolysis (6 N HC1, 110°C for 24 h): Asp 3.0 (3); Thr 1.0 (1); Ser 2.7 (3); Glu 2.0 (2); Gly 2.4 (2); Ala 4.0 (4); Val 1.1 (l); Met 1.1 (l); Ile 1.9 (2); Leu 3.9 (4); Tyr 2.5 (2); Phe 0.9 (1); Lys 2.8 (3); and Arg 2.8 (3). Additional confirmation of structure was provided by HI-NMR (DMSO-~) and

GRFa-c-GIy-GIy-Cys-NH2 O COOH O S IJ~"N%ffr " N~~ O II I II 1 / / S-CH2-CH2-C-NH-CH-CI-12)4oNH.C-(CH2)4[ NH FIG. I.Biotinylationscheme for C-terminal-extended h G R F 2 9 analogs: a = hGRF29-, b = [AlaIS]hGRF29 - and c = [desNH2Tyrl,D-Ala2,AIatS]hGRF29-. To each of the C-terminal-extended h G R F analogs (GRF,.c-GIy-GIy-Cys-NH2), an excess of N(-2-pyridyl-dithiopropionyl)(tpBiocytin)-NH2 was added. Under oxidativeconditions,thiopmpionylBiocytin was directedto the singledistalcysteineresidueof GRF,.c-GIyGIy-Cys-NH2, with 2-pyridinc thiol acting as a leaving group. The resultant product(s) were GRFa_c-Gly-Gly-Cys-(thiopropionyl-Biocytin)NH2.

FAB mass spectroscopy [calculated: (M + H) + 4048.0; found: 4047.8]. [desNH2Tyrl,D - Ala2,Ata 15]hGRF29 - GGC - (thiopropionyl Biocytin)-NH2 {[desNH2Tyrl,D-Ala2,AIa~S]hGRF29-GGC(tpBiocytin)-NH2} was homogeneous by analytical HPLC and gave the expected amino acid composition after acid hydrolysis (6 N HCI, 110°C for 24 h): Asp 3.2 (3); Thr 0.9 (1); Ser 2.4 (3); Glu 2.1 (2); Gly 1.9 (2); Ala 3.9 (4); Val 1.1 (1); Met 0.9 (1); Ile 1.9 (2); Leu 4.2 (4); Phe 1.0 (1); Lys 3.1 (3); and Arg 3.3 (3). This structure was also confirmed by Ht-NMR (DMSO-d6) and FAB mass spectroscopy [calculated: (M + H) + 4032.8; found: 4032.9].

GRF Iodination Human GRF29 was radioiodinated (8.5 min reaction time) by the Iodogen (Pierce Biochemical) method (11). The radioligand, [125I]-hGRF29, was purified by column chromatography (CM-52 carboxymethyl cellulose, Whatman) and collected in tubes containing 10 ug/ml bovine serum albumin (Fraction V, Sigma Chemical) in 0.1 M sodium phosphate buffer. The specific activity of the iodinated radioligand was typically 100-150 uCi/ ug and was stable for ~ 10 days at 4°C.

Immunological Identity of GRF Analogs: GRF Radioimmunoassay (RIA) Recognition of hGRF analogs (24-50,000 pg/100 ~tl) by polyclonal hGRF antisera [HLR lot #2; generated in rabbits immunized against hGPdr29-GC~-NH2 conjugated to bovine serum albumin via MBS (m-maleimidobenzoyl-N-suceinimide ester)] was determined by competitive displaeement. This antibody cross-reacts with hGRF44 and hGRF29, but not with vasoactive intestinal peptide, PHI(1-27), secretin, glueagon, somatostatin, or rat GH. All hGRF standards, analogs, primary antibody, second antibody, and normal sera were diluted in GRF

BIOTINYLATED G R F ANALOGS

789

10096

--

--

4,..~,

B.

+"-+,

~

C. ,,.

".~. ~,~

2o~-I

",-

.I I01

",~.

102

103

104

105

101

hGRF Concentration (pM]

102

103

"~,

104

105

102

101

hGRF ConcentmUon ~]d)

103

104

105

hGRF ConcentrsUon (]Did)

FIG. 2. (A-C) Antibody binding to parent, C-terminal-extended, and biotinylated hGRF analogs. Recognition of hGRF analogs (24-50,000 pg/100 tA) by hGRF antisera was determined by competitive displacement in a GRF RIA (i.e., for ability to inhibit [~251]-GRF29binding). Displacement curves (% inhibition of radioligand binding) were compared to determine whether C-terminal extension and/or biotinylation altered recognition by this antibody. For all curves (A-C), hGRF44 (+) is included as a reference standard. (A) Effect of unsubstituted hGRF analogs: hGRF29 (©), hGRF29-GGC-NH2 (@), and hGRF29-GGC-(tpBiocytin)-NH2 (11). (B) Effect of monosubstituted hGRF analogs: [AlatS]hGRF29 (©), [AIa~5]hGRF29-GGC-NH2(@), and [Ala~5]hGRF29-GGC-(tpBiocytin)-NH2(11). (C) Effect of multisubstituted hGRF analogs: [desNH2Tyr',D-AIa2,Ala~5]hGRF29(©), [desNH2Tyri ,D-AIa2 ,Ala15]hGRF29-GC~-NH2 (@), and [desNH2TyrI ,D-Ala2,Ala15]hGRF29GGC-(tpBiocytin)-NH2 (11).

RIA buffer consisting of: 0.5 M sodium phosphate, 0.15 M NaCI, 0.03 M EDTA, 0.1% sodium azide, 0.1% human serum albumin (Fraction V, ICN Biochemicals), 0.1% Triton X-100 (v/v), and 0.1% porcine gelatin, hGRF44, hGRF29, and hGRF29 analogs (24-50,000 pg/100 #1) and [lzsI]-hGRF29 (10-11,000 cpm/100 #1) were added to rabbit anti-hGRF sera (1:50,000 initial dilution) and allowed to incubate for 24 h at room temperature. Goat anti-rabbit IgG (100 tA of 1:5 initial dilution; Antibodies Inc.) and normal rabbit serum (100 ul of 4% solution; Calbiochem) were subsequently added, followed by 1 ml PEG-8000 (6%). All tubes were centrifuged (40 min, 1670 × g), the supernatant decanted, and the pellet counted (ICN/Micromedic 10/600 PLUS "r-counter).

Pituitary Cell Dispersion, Culture, and GRF Analog Treatment Anterior pituitary cells from male Sprague-Dawley rats (150175 g b.wt.) were dispersed with 0.4% collagenase/0.2% dispase (80 min), followed by a 10-min incubation with neuraminidase

+ hGRF44 Q [AIalSl.hGRF29 , [AI.151hGRF29GGCN1~ • IAlalSIhGRF29GGC Itl~oqt~ NI~

+ hGRF44 0 hGRF29

]

'

. hGRF29GGCNI~ 0 hGRF29GGC[tpBIocyttnlN142 .

/

|

/

~

(4 #g/ml) as previously described (7). After 96-h culture, cells were incubated in the presence of hGRF analogs for 4 h (37°C) in l ml of defined serum-free Dulbecco's MEM (7) per well at concentrations of 3. l --* 200 p M (1:2 serial dilutions). Each dose of G R F analog was tested in quadruplicate wells. Media aliquots (75 ttl, diluted 1:40 in RIA buffer) were later assayed for GH by standard double antibody RIA employing reagents (NIADDK Rat GH standard GH-RP-2) supplied by Dr. A. Parlow (Torrance, CA).

In Vivo GRF Analog Administration Female C57BL/6 mice ( 2 2 0 g b.wt.) were used in all studies following at least 1 week acclimation prior to experimentation. hGRF44 and hGRF29 analogs were injected IV (retro-orbital, 0-30 ug/kg as noted) in 0.2 ml volume. Mice were then bled (retro-orbital) at designated time intervals: 0, 5, 10, 15, 30, and 60 min postinjection (n = 5 mice per time point). Serum was separated and frozen (-20°C) for subsequent GH RIA (see Pituitary Cell Dispersion and Culture for details).

/

+

/

~

/

hGRF44

A lc~.~1.~2. ~t~-ht~,~

~ "

/

/

/

o

2~

]°°Ib ] ....... ibo " GRle~

/.-.

..... ih2 ..... 10'-i

Concentrafloa (p]~

. . . . . . .I 0. . . . . . . . . .10. . . . . . . . . .10. . . . ORF A~lol[ Cmm=eatraflo~ (pM)

-t . . . . . . . ib° . . . . . . . i ~ . . . . . . i62 . . . . . . . . GRF Analog Comcent~Uoa

FIG. 3. Effect of hGRF analogs on GH release by cultured rat anterior pituitary cells, hGRF analogs were added at the indicated concentrations in serum-free medium and allowed to incubate for 4 h. Figures illustrate only the linear portion of the log[dose]-response curve, determined (95% confidence limits) by a Parallel Line Bioassay computer program (PARATL) (8) for potency assessment. Computer-generated simple regression lines (p >_ 0.90) are shown. All results are mean values (SEM < 10%) for n ~ 3 independent determinations. Left panel: unsubstituted hGRF analogs; center panel: monosubstituted hGRF analogs; and right panel: multisubstituted hGRF analogs.

790

CAMPBELL ET AL. TABLE 1 IN VITRO AND IN VIVO POTENCIES OF HUMAN GRF(l-29)-NH2 (hGRF29) ANALOGS RELATIVE TO hGRF(I-44)-NH2 RelativePotency hGRF Analog

In Vitro*

In Vivo~"

hGRF( 1-44)-NH2 (standard) hGRF( 1-29)-NH 2 (hGRF29) hGRF29-Gly-Gly-Cys-NH 2 hGRF29-Gly-Gly-Cys-(tpBiocytin)-NH 2 [Ala 15]-hGRF29 [Ala J5]hGRF29-Gly-Gly-Cys-NH 2 [Ala ~5]hGRF29-Gly-Gly-Cys-(tpBiocytin)-NH 2 [desNH 2TyrI,D-Ala2,Ala15]hGRF29 [desNH2Tyr~,D-Ala2, AlalS]hGRF29-Gly-Gly-Cys-NH2 [desNH2Tyrl,D-Ala2, AlalS]hGRF29-Gly-Gly-Cys-(tpBiocytin)-NH 2

1.00 0.71 0.81 0.79 3.81 3.68 3.92 4.70 4.94 4.70

1.00 0.83 0.76 0.78 5.70 5.55 5.47 9.53 8.72 8.64

* For in vitro assessment of potency, cultured (96 h) rat anterior pituitary cells were incubated for 4 h in the presence of hGRF analogs (3. l --~ 200 pM) and compared for GH-reteasing ability. t For in vivo determination of GH-releasing activity, hGRF44 and hGRF29 analogs were injected IV. (retro-orbital, 0-30 ~g/kg) into mice and bled (retro-orbital) at designated time intervals: 0, 5, l 0, 15, 30 and 60 min postinjection. Total GH area under the curve (0--60min) was used for calculation of potency. Relative in vitro and in vivo potencies (hGRF(l-44)-NH 2 = 1.00) were determined by a Parallel Line Bioassay computer program (PARATL) (8).

Data Analysis

GH-Releasing Activity In Vitro

Relative in vitro and in vivo potencies were calculated on a molar basis using a Parallel Line Bioassay Program (PARATL) (8), such that hGRF44 was assigned a potency of 1.0. PARATL plots log[dose] vs. response (linear regression) and tests (95% confidence) for both parallelism (i.e., to the standard) and linearity (i.e., excluding points significantly departing from linearity, such as doses producing no effect or saturation). Therefore, calculations represent the linear portion of the log[dose]-response curve. For in vivo comparisons of hGRF29 analogs, potencies were calculated using integrated GH area under the curve (GH AUC; determined by trapezoidal summation) data for total (060 min) treatment period (Fig. 5). The peak (0-10 min) and postpeak (10-60 min) GH AUC were also calculated to ascertain relative duration of G H response (Fig. 5).

The relative in vitro potency of each of the parent h G R F compounds was compared with their C-terminal-extended and biotinylated counterparts to ascertain whether GH-releasing activity was influenced by these derivatizations (Fig. 3, Table 1). Relative to hGRF44 (potency = 1.00), similar potencies (0.71, 0.81, and 0.79, respectively) were observed with hGRF29, hGRF29-GGC-NH2, and hGRF29-GGC-(tpBioeytin)-NH2. [AlalS]hGRF29, [AIaIS]hGRF29-GGC-NH2, and [AlalS]hGRF29-GGC-(tpBiocytin)-NH2 were significantly more potent than the above nonsubstituted hGRF analogs, but were equipotent to each other (3.81, 3.68, and 3.92, respectively). The multisubstituted analogs [desNH2Tyr~,D-AIa2,AlaIS]hGRF29, [desNH2Tyrt,D-AlaZ,Ala~5]hGRF29-GGC-NH2, and [desNH2TyrI,D-AlaZ,AlalS]hGRF29-GGC-(tpBiocytin)-NH2were equally active and displayed the greatest GH-releasing potencies in vitro (4.70, 4.94, and 4.70, respectively).

RESULTS

Antibody Recognition of GRF Analogs

GH-Releasing Activity In Vivo

All analogs synthesized were examined in a G R F radioimmunoassay (RIA) for ability to inhibit [~25I]-GRF29 binding to G R F antisera (Fig. 2). Displacement curves (% inhibition ofradioligand binding) were compared to determine whether C-terminal extension and/or biotinylation altered recognition by this antibody. On a molar basis, the displacement curves for hGRF44, hGRF29, hGRF29-GGC-NH2, and hGRF29-GC~-(tpBiocytin)NH2 were virtually superimposable. Substitution of Ala t5 (for Gly JS) similarly reduced (2.5-3.0-fold) the antibody affinity for each of these hGRF29 analogs: [Ala~5]hGRF29, [AIa~S]hGRF29GGC-NHz, and [Ala~5]hGRF29-GC~-(tpBiocytin)-NH~, resulting in comparable binding profiles. Additional desNH2Tyr ~ and D-AIa2 substitution (for Tyr ~ and L-Ala2) did not further alter the ability of these analogs {[desNH2Tyr~,D-Ala2,Ala~5] hGRF29, [desNH2Tyr~,D.AIa2,AIaIS]hGRF29_GC~.NH2, and [desNH2Tyr~,D-Ala2,Alat 5]hGRF29-GCC-(~Bioeytin)-NH2 } to displace the radioligand. In all cases, neither C-terminal extension nor biotinylation affected the relative immunoreactivity o f h G R F analogs.

The acute effects (0-60 min postinjection) of the C-terminal-extended and biotinylated h G R F analogs on circulating GH concentrations were compared in vivo to their corresponding parent compounds (Fig. 4). Potencies were calculated on a molar basis using total G H area under the curve (GH AUC) data (Table 1, Fig. 5). hGRF29, hGRF29-GGCNH2, and hGRF29-GGC-(tpBiocytin)-NH2 elicited similar effects on circulating G H profiles and cumulative G H AUC (potencies = 0.83, 0.76, and 0.78, respectively, relative to hGRF44). [Ala ~5] substitution resulted in enhanced potency for all three analogs examined, with [Ala~5]hGRF29, [AIaIS]hGRF29-GGC-NH2, and [ A l a l S ] - h G R F 2 9 - G G C (tpBiocytin)-NH2 being 5.70, 5.55, and 5.47 more potent than hGRF44. [desNH2Tyr~,D-Ala2] substitution, in addition to [Ala~5], further augmented GH-releasing potency of the three m u l t i s u b s t i t u t e d analogs: [desNH2Tyr~,D-Ala2, AlalS]hGRF29, [desNH:TyrI,D-AIa2,AlatS]hGRF29-GGCNH2, and [desNHzTyr~,D-AIa2,AIa~S]hGRF29-GGC-(tpBiocytin)-NH2 (potencies = 9.53, 8.72, and 8.64, respectively).

BIOTINYLATED G R F ANALOGS

791

GRF~

il

[AlalSl-GlqF,m

e°°l

ill

t ~

i

..... o

15

so

.-~--

,. o a ~

:

45

6o

o

Time ~ l a t i v e to In]ecUon (rot.)

15 so 45 6o Time l ~ e l a t i v e t o I n j e c t i o n (m/n)

o

i

_o.o,,,,

i'

.--6--.

::.

.......

,

.

o.

t/~

,.o~.,

o.o

o.s~

_+

IdesNH;lyr I. D-AIa2,AIalS]-GRF~-GGC

,

.o.o

..

t A

o.~.~

--~

o.

is so 45 ~o Time R e l a t i v e t o I n j e c t i o n (mi.I

[Ala15]-GRF29-GGC

GI~F~-C,GC o.o

IdesNl~'l~/r1, D-AIa2, AIalSI"GIqF~*

.o.o

"

~

F.

o.~,~.~

t

! t

100

--

o~

.

:1~>

3b

"

4g

"

~

"

" o

Time RelaUve to Injection (ml.)

" lb

GRF~-GGC-tpBiocytln

i

i~

.--0-,--

,. o

. . . .

0

'

t~ 4 IlXX 1

S O i~]ill

. . . . . . . .

"

"ok

"

~

"

IAla!51-GIqF~-GGC-tpBlocyUn

--2-,.o~ 3o~,, ,o,,~,

4O0

sb

ill

-

15 30 45 60 T i m e R e l a t i v e t o I n j e c t i o n (rain}

" o

-

~ o~,~, - - . - ,.o,,~., - 3.o,,,,,

llkC__ 0

15 30 45 60 Time Relative to Injection {ml.)

-

_

_. Ill

.~..

SO

. . . . .

45

73

.

.

60

Time Relative to Injection (~h~)

Time Relative to Injection (min)

[desNl-~'l~yr 1. D-AIa 2, Alalsl-Gl~w-GGC-tpBiooytln

- . ~ o.~,,~ --o- ,.o,,,~, . ~.o~,

I~ 4 rxi, 1

.

0

-

.

.

.

.

°

~i

.

15 30 45 60 T i m e R e l a t i v e t o I n J e e t i o n {mln)

FIG. 4. Effect of hGRF analogs on serum GH concentrations in mice. hGRF analogs were injected IV (retro-orbital, 0-30 pg/kg as noted) and bled (retro-orbital) at designated time intervals: 0, 5, 10, 15, 30, and 60 min postinjection. Values represent mean ___SEM (n = 5 mice per time point). (A-C) parent compounds; (D-F) C-terminal-extended hGRF29 analogs; and (G-I) biotinylated hGRF29 analogs.

The in vivo potencies for these multisubstituted h G R F analogs were not significantly different from each other (within 95% confidence limits). The total GH AUC response (0-60 min) was subdivided into peak (0-10 min) and postpeak (10-60 min) GH AUC responses (Fig. 5). Peak GH AUC responses were substantially greater for monosubstituted ([Ala~5]) and multisubstituted (desNH2Tyr~,D-Ala2,AlatS]) hGRF29 analogs than for unsubstituted analogs. Although the potencies (calculated using total G H AUC) of all [desNH2Tyr~,D-Ala2,Ala~5]-substituted analogs were higher than that of the [AlatS]-substituted counterparts, their peak G H AUC did not significantly differ. However, the postpeak G H AUC response was more elevated for the [desNH2Tyr~,D-Ala2,Ala~]-substituted analogs relative to the [AlatS]-substituted analogs..

DISCUSSION Three parent compounds representing unsubstituted (hGRF29), monosubstituted ([AlatS]hGRF29) and multisubsti-. tuted ([desNH2Tyr~,D-Ala2,AlalS]hGRF29) hGRF29 analogs were successfully synthesized. Each of these analogs was also C-terminal extended (with -GIy-Gly-Cys-NH2) and biotinylated (using N-2-pyridyl-dithiopropionyl)-(tpBiocytin)-NH2)to yield six additional compounds. Unlike previously reported methods (1-5,16-19), the procedure utilized in the present studies (Fig. 1) is advantageous for C-terminal biotinylation ofpeptides (i.e., where N-terminal biotinylation reduces bioactivity) and results in a single quantifiable biotinylation site (i.e., per Cys residue). Dimerization of the hGRF29-GGC-NH2 analogs during biotinylation is avoided by availability of excess reactant [N(-2pyridyl-dithiopropionyl)-Biocytin]. The thiopropionyl-Biocytin

792

CAMPBELL ET AL.

0.3 ~ / k g

0 - I 0 ntln I n t e g r a t i o n

~ Lo~/~ .q.o~g/kg

Treatment

o.3 ~ M

10410 men Integration

B lo~t/M ao ~/M

0

Treatment

10000

8000

ii~ w

2000

0

Trtmtmat FIG. 5. Effect ofhGRF analogs on GH area under the curve (GH AUC) in mice: peak (0-10 min), postpeak ( 10-60 min), and total GH AUC responses, hGRF analogs were injected IV (retro-orbital, 0-30 #g/kg as noted) and bled (retro-orbital) at designated time intervals: 0, 5, 10, 15, 30, and 60 min postinjection. Treatments: saline control (1), hGRF44 (2), hGRF29 (3), hGRF29-GC~-NH2 (4), hGRF29-C~-(tpBiocytin)NH2 (5), [AlaJS]hGRF29 (6), [AIaIS]hGRF29-GGC-NH2 (7), [AlaI~]hGRF29-GGC(tpBiocytin)-NH2 (8), [desNH2Tyrt,D-Ala2,AlatS]hGRF29 (9), [desNH2TyrI,DAIa2,AlatS]hGRF29-GGC-NH2 (10), and [desNH2TyrI,D-Ala2,AlalS]hGRF29-GGC(tpBiocytin)-NH2 (11).

hGRF29 derivatives obtained were homogeneous, fully characterized, and obtained in good yield (~40%). The resulting C-terminal-extended and biotinylated analogs of hGRF29, [AlalS]hGRF29, and [desNH2Tyrl,D-Ala2,AlatS]hGRF29 exhibited nearly identical immunoreactivities and bioactivities (in vitro and in vivo) to the parent compounds. Therefore, neither C-terminal extension nor biotinylation affected these parameters. The high affinities of the biotinylated hGRF29 analogs for G R F antibody imply that the respective

secondary structures, within the binding epitope (residues 2 9 - 2 0 for this antibody), were not significantly altered by these derivatizations. Furthermore, the data suggest that these biotinylated analogs may be useful for histochemistry and/ or G R F ELISA (i.e., in combination with chromogenically/ fluorogenically labeled avidin or streptavidin). The potent GH-releasing activities o f the biotinylated hGRF29 analogs further indicates that these peptides may be utilized as G R F receptor probes, for ultimate isolation of the G R F receptor.

BIOTINYLATED G R F ANALOGS

793

However, adequate receptor solubilization techniques, which retain the functional integrity of the G R F receptor, have yet to be elucidated. During the course of these studies, it was also revealed that [Ala 15] and [desNH2Tyrl,D-Ala2,Ala15] substitution imparted greater GH-releasing activity to hGRF29, regardless of C-terminal extension or biotinylation. However, some disparity exists between the in vitro and in vivo potencies of [AlalS]hGRF29 and [desNH2Tyrl,D-AlaZ,Al#5]hGRF29 (3.8 and 4.7 in vitro vs. 5.7 and 9.5 in vivo, respectively, relative to hGRF44). The similar in vitro activities, also previously observed with nonderivatized [Ala ls]hGRF29 and [desNHzTyr l,D-Ala2,Alals]hGRF29 (9,10), are largely due to enhanced G R F receptor affinity by [Ala 15] substitution (8). This is presumably also reflected in the peak GH AUC response in vivo (Fig. 5), where no differences were observed between [AlatS]- and [desNH2Tyrt,D-Ala2,AlalS]-substituted analogs. The incremental effect of additional [desNH2TyrI,D-Ala2] substitution on in vivo potency is, however, directly correlated with postpeak GH AUC, producing a greater duration of GH response (Figs. 4, 5). Therefore, [desNH2Tyrt,DAla2] substitution allows more bioactive peptide to be made available to the pituitary per unit time. The mode of action for

this sustained response may involve increased resistance to enzymatic degradation, as demonstrated in plasma in vitro ( 12,13,20), although a decrease in metabolic clearance rate cannot be precluded. In summary, novel C-terminal and biotinylated hGRF29 analogs have been synthesized that retain full immunological and biological activities. These biotinylated hGRF29 analogs may be employed for a variety of experimental and/or diagnostic purposes, such as ELISA, histochemistry, and receptor isolation. The C-terminal-extended intermediates demonstrate unique utility, as a vast array of conjugations/derivatizations can take advantage of the distal cysteine residue. For example, hGRF29GGC-NH2 conjugated to bovine serum albumin via MBS was used in the present studies for generation of G R F antibodies in rabbits. ACKNOWLEDGEMENTS The authors wish to thank P. Stricker for GRF/GH RIA and T. Lambros for HPLC purifications. We also wish to thank B. Tytla for amino acid analysis, Dr. W. Benz for mass spectrometry, Dr. T. Williams for NMR, and Dr. Y.-C. Pan for sequence analysis. The critical review of this manuscript by Dr. R. Cordts is gratefully acknowledged.

REFERENCES

1. Abou-Samra, A.-B.; Freeman, M.; Jiippner, H.; Uneno, S.; Segre, G. V. Characterization of fully active biotinylated parathyroid hormone analogs. Application to fluorescence-activated cell sorting of parathyroid hormone receptor bearing cells. J. Biol. Chem. 265:5862; 1990. 2. Andersson, M.; Marie, J.-C.; Carlquist, M.; Mutt, V. The preparation of biotinyl-~-aminocaproylated forms of the vasoactive intestinal polypeptide (VIP) as probes for the VIP receptor. FEBS Lett. 282: 35-40; 1991. 3. Anton, P. A.; Reeve, J. P., Jr.; Vidrich, A.; Mayer, E.; Shanahan, F. Development of a biotinylated analog of substance P for use as a receptor probe. Lab. Invest. 64:703-707; 1991. 4. Anton, P. A.; Reeve, J. P., Jr.; Rivier, J. E.; Vidrich, A.; Schepp, W.; Shanahan, F. Biotinylation of a bombesin/gastrin-releasing peptide analogue for use as a receptor probe. Peptides 12:375-381; 1991. 5. Balasubramaniam, A.; Sheriff, S.; Ferguson, D. G.; Stein, M.; Rigel, D. F. N-a-biotinylated-neuropeptideY analogs: Syntheses, cardiovascular properties and application to cardiac NPY receptor visualization. Peptides 11:1151-1156; 1990. 6. Barany, G.; Merrifield, R. B. Solid-phase peptide synthesis. In: Gross, E.; Meienhofer, J., eds. The peptides: Analysis, synthesis, biology, vol. 2. New York: Academic Press; 1980:1-284. 7. Brazeau, P.; Ling, N.; BOhlen, P.; Esch, F.; Ying, S.-Y.; Guillemin, R. Growth hormone-releasing factor, somatocrinin, releasespituitary growth hormone in vitro. Proc. Natl. Acad. Sci. USA 79:7909-7913; 1982. 8. Campbell, R. M.; Lee, Y.; Rivier, J.; Heimer, E. P.; Felix, A. M.; Mowles, T. F. GRF analogs and fragments: Correlation between receptor binding, activity and structure. Peptides 12:569-574; 199 I. 9. Felix, A. M.; Heimer, E. P.; Mowles, T. F.; Eisenbeis, H.; Leung, P.; Lambros, T. J.; Ahmad, M.; Wang, C.-T.; Brazeau, P. Synthesis and biological activity of novel growth hormone releasing factor analogs. In: Theodoropoulus, D., ed. Peptides 1986. New York: Walter de Gruyter & Co.; 1987:481-484. 10. Felix, A. M.; Wang, C.-T.; Heimer, E.; Fournier, A.; Bolin, D. R.; Ahmad, M.; Lambros, T. J.; Mowles, T. F.; Miller, L. Synthesis and biological activity of novel linear and cyclic GRF analogs, In: Mar-

11.

12.

13.

14.

15.

16. 17. 18. 19. 20.

shall, G. R., ed. Peptides: Chemistry and biology. Leiden: ESCOM; 1988:465-467. Fraker, P. J.; Speck, J. C. Protein and cell membrane iodinations with a sparingly soluble chloroamide 1,3,4,6-tetrachloro-3a,6a-diphenyl-glycoluril. Biochem. Biophys. Res. Commun. 80:849-857; 1978. Frohman, L. A.; Downs, T. R.; Williams, T. C.; Heimer, E. P.; Pan, Y.-C. E.; Felix, A. M. Rapid enzymatic degradation of growth hormone-releasing hormone by plasma in vitro and in vivo to a biologicallyinactive product cleaved at the NH2 terminus. J. Clin. Invest. 78:906-913; 1986. Frohman, L. A.; Downs, T. R.; Helmet, E. P.; Felix, A. M. Dipeptidylpeptidase IV and trypsin-like enzymatic degradation of human growth hormone-releasing hormone in plasma. J. Clin. lnvest. 83: 1533-1540; 1989. Ling, N.; Baird, A.; Wehrenberg, W. B.; Ueno, N.; Munegami, T.; Chiang, T.-C.; Regno, M.; Brazaeu, P. Synthesis and in vitro bioactivity of human growth hormone-releasing factor analogs substituted at position- I. Biochem. Biophys. Res. Commun. 122:304-310; 1984. Ling, N.; Baird, A.; Wehrenberg, W. B.; Ueno, N.; Munegami, T.; Brazeau, P. Synthesis and in vitro bioactivity of C-terminal deleted analogs of human growth hormone-releasing factor. Biochem. Biophys. Res. Commun. 123:854-861; 1984. Lutz, W. H.; Londowski, J. M.; Kumar, R. Synthesis and biological activity of a biotinylated vasopressin analog. Peptides 11:687-691; 1990. Pieri, I.; Barritault, D. Biotinylated basic fibroblast growth factor is biologically active. Anal. Biochem. 195:214-219; 1991. Sawutz, D. G.; Yanni, J.; Kelley, M.; Wolfe, H. Synthesis and molecular characterization of a biotinylated analog of [Lys]bradykinin. Peptides 12:1019-1024; 1991. Schvartz, I.; Gitlin, G.; Amarant, T.; Ittloop, O.; Hazum, E. Biotinylated endothelin as a probe for the endothelin receptor. Peptides 12:1229-1233; 1991. Su, C.-M.; Jensen, L. R.; Heimer, E. P.; Felix, A. M.; Pan, Y.-C. E.; Mowles, T. F. In vitro stability of growth hormone releasing factor (GRF) analogs in porcine plasma. Horm. Metab. Res. 23:15-21; 1991.