In!. J. Pepride Protein Res. 40, 1992, 41-48

Conformational studies of N-Tyr-MIF-1 in aqueous solution by lH nuclear magnetic resonance spectroscopy M. PETERSHEIM, R.L. MOLDOW, H.N. HALLADAY, A.J. KASTIN and A.J. FISCHMAN

Departments of Chemistry and Biology, Seton Hall University, South Oranje, NJ; Veterans Aflairs Medical Center and Tulane University School of Medicine. New Orleans, LA; Division of' Nuclear Medicine of the Department of Radiology. Massachusetts General Hospital, and the Department of Radiology Harvard Medical School, Boston, MA, USA

Received 1 May 1991, accepted for publication 22 February 1992 N-Tyr-MIF-1 (Tyr-Pro-Leu-Gly-NHZ) is an endogenous brain peptide with multiple effects on animal behavior. However, there have been no studies on the conformation of this tetrapeptide. In this report, we studied the conformation of N-Tyr-MIF-1 in aqueous solution by conventional one-dimensionaland two-dimensional (COSY and NOESY) 'H nuclear magnetic resonance spectroscopy at 300 MHz. A complete set of assignments for the resolved resonances and approximate assignments for the overlapping resonances were made. The results demonstrate that N-Tyr-MIF-1 is in slow exchange between two conformers, most likely determined by the cis and trans states of the proline residue. The minor conformation represents 30 3% of the population over the temperature range from 3" to 73". In the major conformation, the tyrosine aromatic ring appears to be close enough to interact directly with the proline pyrrolidine ring, as indicated by a strong temperature dependence of the proline CPH, C6H and CaH' chemical shifts. In contrast, this interaction of the tyrosine and proline rings is not present in the minor conformation.

*

Key words; conformation; COSY; NOESY; NMR; N-Tyr-MIF-1

N-Tyr-MIF-1 (Tyr-Pro-Leu-Gly-NH2) is a peptide with multiple behavioral effects, including antiopiate activity (1,2), anti-immobility effects (3), increased passive avoidance behavior (4, 5), and antidepressant effects (6). Even before the peptide was isolated from brain tissue and its structure unambiguously defined, it was localized to multiple areas of the brain by radioimmunoassay (7). This immunoreactivity has been verified to be N-Tyr-MIF- I by gel filtration chromatography and high performance liquid chromatography (2). In addition, high affinity saturable binding sites for N-TyrMIF-1 have been found in brain tissue (8). Recently, the peptide was isolated from bovine brain tissue and the sequence was verified (9). It is also present in human brain cortex (10). Although N-Tyr-MIF- 1 has been extensively studied in numerous models of animals behavior (2-7) the number of reports of structure-activity studies has been extremely limited (4). Many conformational studies have been performed with the related peptide, MIF-1 (1 1-24); however, there have been no reports of the conformation in solution of N-Tyr-MIF-1. The presence of an X-Pro linkage in N-Tyr-MIF-1 introduces

many additional conformational possibilities, including conformers in which the Tyr-Pro amide bond is in a cis conformation. These potential conformational differences could help to explain the difference in the spectrum of the biological activities of the two peptides. In this report, we use conventional one-dimensional and two-dimensional (COSY and NOESY) 'H nuclear magnetic resonance spectroscopy to evaluate the conformational states of N-Tyr-MIF-1 in aqueous solution. The results of these studies may be of value in designing new analogs of N-Tyr-MIF-1 for further behavioral analysis. MATERIALS A N D METHODS Peptide synthesis Tyr-Pro-Leu-Gly-NH2 (N-Tyr-MIF- 1) was prepared by solid phase methods with benzhydrylamine resin to yield the C-terminal amide after cleavage from the resin with liquid hydrogen fluoride and concomitant deblocking. The crude synthetic peptide was desalted by gel filtration chromatography on Sephadex G-10 (2.5 x 60 cm) eluted with 1.0 M acetic acid. Final puri41

M. Petersheim er al. fication of the desalted product was achieved bq chromatography on the same column eluted Ivith 0.20 M acetic acid. Thin layer chromatography revealed single spots in the following sqstems: n-butanol: acetic acid: water 4:1:5 (upper phase). ethyl acetate: acetic acid: n-butanol: water 1: 1: 1:1, and chloroform: methanol: water 8:5:1. Reversed phase HPLC ( C I ~column, 0.45 x 15 cm) with a linear gradient from 2"" methanol 0.10," TFA to 35", methanol/O.l", TFA showed a single peak. NMR spectroscopj' The two-dimensional and most of the conventional NMR experiments were performed with 5.2 mg of the tctrapeptide dissolved in 0.75 mL of 99.96', D.0 (Aldrich) after lyophilization twice from D20. The final pH was 4.8 and the solution contained about 0.1 niM TSP as an internal reference. Studies of the exchangeable amide protons were performed in 0.52 niL of distilled water and 0.13 mL of 99.96", D.0 (Aldrich). All experiments were performed on a General Electric QE-300 NMR spectrometer with a 5 n m lH:13C dual isotope probe equipped with variable temperature accessories. The temperature at the sample ivas determined by the relative chemical shifts of the hydroxyl and aliphatic resonances of neat ethylene glycol. COSY experiments were performed with the standard n.2-tl7r/2-t2 pulse sequence and phasing (25). The data sets were 1K by 2K with a full spectral width of 3400 Hz. 16 acquisitions for each value of t l and 3.5 s total time per cycle. The data were apodized in both dimensions with the half-sine function and converted to pure amplitude spectra. NOESY experiments were performed at 8' and 23' using a 7r/2-t1-n/2-tI,,2-t1 4-7r-t1, 2-n'2t 2 pulse sequence (26), with mixing times (tl,,) ranging from 0.2 to 1.0 s. Selective saturation NOE experiments were performed by alternating on- and offresonance irradiation every 8 acquisitions. Water suppression in the studies of the exchangeable ainide protons was accomplished with the 1-3-3-1 soft n 2 pulse (27). The composite pulse used consisted of 10 and 30 ps pulses centered on the water resonance and separated by 200 ps delays. Molecular modeling Unconstrained energy minimization of both cis-N-TyrMIF-1 and rrcms-N-Tyr-MIF-1 was performed by the MMX modification of the MM2 force field (28) using the program PCMODEL (version 2.0; Serena Software, Bloomington, IN). This modification of the MM2 force field permits energy minimization of structures containing a variety of functional groups. The starting structures were derived from previously reported data on the structure and conformation of model peptides with extended conformations (29). Initially, all calculations were performed without hydrogens. In the final refinement of the structures all hydrogens and solvent (water) were included in the minimizations. 42

RESULTS Assigrmenr of proton resoiinnces Fig. 1 sh0u.s a COSY contour plot and corresponding spectrum for the N-Tyr-MIF-1 resonances upfield of the HDO peak. The remaining aromatic tyrosine resonances and the exhangeable amide protons are presented in Fig. 2. Although this is a relatively simple peptide, there is enough spectral overlap at any one tempcrature to complicate assignments. The assignments presented in Table 1 were determined by COSY experiments at both 23.3 and 73 in conjunction with direct decoupling experiments. O,

FIGURE 1 COSY contour p l o ~of the aliphatic protons of N-Tyr-MIF-1 in Y9.96",, D20 at 23.3 '. The cross peak assignments are: a = LeuC".'H (major and minor); b = ProCB-'H (minor); c = ProCQl,H' and ProC''-rH (major); d = ProC'."H (minor); e = Proc'H, ProC"', (niinor): f = L c L I C ~ J N(major); g = ProCIH, ProCm, (major); h = ProC~-'iH(major); i = P I - O C ~ J H (minor); . j = ProC'H, ProC"H', (major): k = ProCZ4I, (major); I = TyrC'H, T y r C m ' , (major); ni = GI!C'H TlrHI'C ("ridgc" artifact); n = TyrC".W, (minor); o = T!rC".BH (major); p = ProCx*-"H(major); q = ProC"H, ProC"H' (major): r:ProC"H,H' (major); s = ProCjH,H' (minor): t = LeuC","H (major): u = unknown contaminant.

Conformation of N-Tyr-MIF- 1 Tyr-m(trans1 Tyr-o(trans)

TABLE 1

Resonance assignments for Tvr-Pro-Leu-Gb-NH: (A'-Tvr-MIF-I) ut 23.3" andpH4.8 Proton

6 (ppm)*

69.8'C

I I

Z ?

Tyr-o(trans)

1

Gly-NH, Leu-NH

Tyr-dtrans)

Tyr-o(trans)

Tyr C"H CPH CPH] C'H CD

4.49 3.22 3.08 1.23 6.90

Pro CzH CPH C4€' C'H C'H' CDH COH'

4.50 2.30 1.90 2.05 2.05 3.73 3.40

Leu NxH CXH C4I C4I' C 'H CDH CDH'

8.47 4.35 1.5-1.8 1.5-1.8 1.5- 1.8 0.98 0.93

Gly N"H NHdt) NHdc) C"H C=H'

8.51 7.47 7.1 1 3.93 3.91

Cross peaks + 0 3 1

0

I

it it kg

k,c g,c h,ic h,ic h,r j,r -

f f f

a,f a a

-

-

6 (ppm)*

Cross peaks

+

3.85 3.1 3.1 7.15 6.92

it it

3.64 1.87 1.87 1.75 1.75 3.58 3.45

i,b 1.b b,ed b,c.d e.s d.s

4.40 1.5-1.8 1.5-1.8 1.5-1.8 0.98 0.90

n n n

1

f f f a,f 7

a

-

-

3.90 3.88

-

-

*

ryr-rdtrans)

sistent with the results of a previous study of Z-ProLeu-Gly(C00-) (13).

--i--l 7. 0

Chemical shifts reported relative to 2,2,3,3,-d~-3(trimethylsi1yl)propionate (TS P). + Cross peaks representing scalar coupling among resonances in the COSY experimcnt at 23.3". it These assignments were made by comparison with the spectra of Tyr-Pro (34).

PPM

300 MHz IH NMR spectra of the amide and tyrosine ring resonances of N-Tyr-MIF-I in water (20% DzO) a1 pH 4.8 3.8" (bottom), 23.3" (center) and 69.8" (top).

Cis-trans isomerism From Fig. 2 it is apparent that at pH 4.8, N-Tyr-MIF-1 is in slow exchange between two conformers. The most obvious indications of the two structures are the second set of resonances for the protons of the tyrosine ring at 7.15 ppm, theglycine C"H protons at 3.85 ppm and the leucine C6H protons at 0.95 ppm. These two conformations are very likely determined by the cis and trans states of the proline residue. The minor conformation remains at 30 & 3 % of the population over the temperature range from 3 ' to 73', This observation is con-

Temperature dependence of NH proton chemical shifrs The temperature dependence of amide proton chemical shifts can be used to distinguish between protons exposed to solvent and intramolecularly hydrogcn bonded protons; exposcd protons have larger temperature coefficients than hydrogen bonded protons (30). The amide and aromatic regions of the 300 MHz 'H NMR spectrum of N-Tyr-MIF-1 at 3.8, 23.3 and 69.8" are shown in Fig. 2. In the low temperature spectrum, the cis and trans carboxamide protons are sharp singlets at 7.21 and 7.55 ppm. In the intermediate temperature spectrum, these resonances have broadened and shifted upfield. In the high temperature spectrum, they continue to broaden and are not resolved from the tyrosine aromatic resonances. The cis and trans carboxamide protons of the trans isomer of N-Tyr-MIF- 1 have the lowest temperature 43

M. Petersheim et al. 7.6 .06

2 7.2 - 0 7.4

.04

.02

I W

0

-.02 -.04

~

I

-.a6

t3

-.08 -.lo

r 0

7 r)

0

0

o o o o

0

0

0

6.8 6.6

6.4

6.0

dependence: -5.25 & 0.49 x and -4.55 0.03 x 10- ppm/"C. The leucine and glycine amide protons have higher temperature coefficients: -9.97 & 0.07 x and -7.82 0.10 x ppm/cC respectively. The magnitude of temperature dependence

0

0

0

7.0

6.2

FIGURE 3 Changes In chemical shlft with temperature for the noncxchangeable protons of the major conformer of N-Tkr-MIF-1 relative to 3 5

0

0

*.

-

* a * *

e -

Tyrl

I

of the N H resonances argues against intramolecular hydrogen bonding.

Temperature dependence of nonexchangeable proton cheniical shifts and vicinal coupling constants The temperature dependence of the chemical shifts of nonexchangeable protons is a sensitive marker of conformational transitions. This is particularly true for protons that are in close proximity to an aromatic ring. The temperature dependence of the aliphatic proton chemical shifts of the resonances corresponding to the major and minor conformers of N-Tyr-MIF-1 is plotted in Fig. 3 (major conformer) and Fig. 4 (minor conformer). Fig. 5 shows the temperature dependence of the three bond C"H-C/.?-I scalar coupling constants, 3J,fi, for the

20.0

N

16.0

9

n ' W

I . l . I I L I ( / . . . . . I l / 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 TEMPERATURE ( " C ) FIGURE 4 Changes in chemical shift with tempcrature for the nonexchangeable protons in the minor N-Tyr-MIF-1 conformer.

44

I

I

I

8.0 I 0

10

20

30

40

50

60

70

80

TEMPERATURE (OC) FIGURE 6 Leucine C"H-CsH scalar coupling versus temperature for the major conformer. The C4f and ClrH' chemical shifts differ above 6 0 " .

Conformation of N-Tyr-MIF- 1

a

b FIGURE 7 Representative structures of trans-N-Tyr-MIF-1 from molecular modeling. A) Stereo view of a conformation with the proline and leucine hydrogens positioned relative to the aromatic ring in a manner consistent with observed chemical shifts. B) Stereo view of an energetically equivalent conformation with the leucine group too close to the aromatic ring to be consistent with observed chemical shifts. The proline ring is in roughly the same orientation for both structures.

tyrosine residue in the major conformer of N-TyrMIF-I. Assuming an ABX coupling pattern (31), the geminal coupling constant, 2Jp,p,, was determined to be -14.74 0.15 Hz. At high temperatures, the two vicinal coupling constants converge to 6.9 Hz. The low

field resonances (3.2 ppm) undergo the larger change in coupling (0.9 Hz) and the smaller change in chemical shift (0.4 ppm) of the two CQ protons. Due to the second order characteristics of the leucine and proline spin systems, individual couplings could not be accu45

M. Petersheim et al.

FIGURE 8 Representative structure of cir-N-Tyr-MIF-1 from molecular modeling. The orientation of the aromatic ring is consistent with the sniallcr shifts in the proline delta and gamma resonances and a greater temperature depcndence in the alpha resonance observed for the minor component. The proline ring is in rough]! the same orientation as in Fig. 7.

rately determined. As Fig. 6 indicates, the sum of the CIH-C&I vicinal coupling constants in leucine is essentially constant up to about 60” where there is a sharp increase. Nuclear Overhauser efects Several one- and two-dimensional nuclear Overhauser experiments were performed at 8 and 23 ‘ with mixing times of up to 1.0 s. In none of these experiments was there consistent evidence of interresidue interactions. It thus seems likely that the rotational correlation time of N-Tyr-MIF-I is close to the null value for the NOE’s at this field strength. Molecular modeling Figs. 7 and 8 are representative of structures obtained by unconstrained energy minimization of trans- and cis-N-Tyr-MIF- 1, respectively. In Fig. 7a for the framisomer, the position of the tyrosine aromatic ring is far from the glycine, leucine and those protons of proline that show relatively little change in chemical shift with temperature, The plane of the aromatic ring is directed toward the C{”H and C6H protons of proline, which is consistent with the large upfield chemical shift of these resonances as the temperature is increased. Fig. 7b is a conformation obtained with roughly the same energy 46

but it places the leucine side chain in a position that should have given a temperature dependence comparable to that observed for the proline resonances. Fig. 8 shows a cis-N-Tyr-MIF- 1 conformation that places the aromatic ring well away from the other side chains, as may be expected from the small temperature dependence observed for the resonances of this conformer. These structures are offered as “reasonable” but not necessarily globally preferred conformations. The flexibility of the peptide and the lack of NMR constraints makes it difficult to justify a global search.

DISCUSSION The major conformer of N-Tyr-MIF-1 appears to have the plane of the tyrosine aromatic ring directed toward the proline pyrrolidine ring, as indicated the strong temperature dependence of the proline C&I, C6H and C6H‘ chemical shifts. The large downfield shifts at low temperatures is consistent with the plane of the tyrosine ring being roughly normal to the proline ring and between the C“‘H and C6H protons. The invariance of the ProC”H shift with temperature indicates that the tyrosine ring interacts with the side of the proline ring opposite from the C”H proton. The leucine C”H and C6H resonances undergo much

Conformation of N-Tyr-MIF- 1 smaller chemical shift changes over the temperature range studied, suggesting that they are not in the vicinity of the tyrosine ring. The glycine and tyrosine resonances undergo changes of 0.04 ppm, or less, except for one of the tyrosine Cm protons, which shifts downfield by 0.07ppm. Although fewer of the resonances of the minor conformation could be followed over the entire temperature range, none showed the large chemical shift changes observed for the more prominent structure (Fig. 4). The nearly sigmoidal shape of the temperature dependence of one of the tyrosine C"H-C/3H couplings supports the idea of a preferred conformation at low temperatures. Assuming that the X1 dihedral angle of tyrosine is partitioned over three staggered rotamers (rotamer I: side chain gauche to NH; and C"H, rotamer 11: side chain gauche to C=O and C"H, rotamer 111: side chain gauche to NH,' and CO) and using previously reported value for 3Jlrun,sand 3JgNwrhe(32), we calculate the rotamer populations (P) to be: PI = 0.42, PII= 0.31 and PIII= 0.27 at low temperature. In the high temperature spectrum, the populations are: PI = 0.39, PII= 0.39 and PIII= 0.22. Although PIII is uniquely defined by the two homonuclear couplings, PI and PI, could be reversed (33). For leucine, PIII is calculated to be 0.32 at low temperature and decreases to -zero at high temperature. This suggests that the conformation of the leucine side chain may be constrained, possibly through interaction with the proline ring. The tyrosine and leucine structural transitions are apparently uncoupled, since the change in tyrosine C"H-Cm is essentially complete by 40". One of the comformations of the trans form of N-TyrMIF-1 (Fig. 7a) predicted by molecular modeling is qualitatively consistent with the chemical shift trends presented in Fig. 3. Since the edge of the aromatic ring is oriented toward the proline delta and gamma hydrogens, the downfield shift of these resonances and the strong temperature dependence of the tyrosine conformation can be explained. The relative temperature independence of all the leucine resonances and that of the proline C"H is consistent with the predicted distance from the aromatic ring to these hydrogens. In this conformation, the tyrosine side chain is gauche to the GO and C"H, i.e., rotomer 11. The other two rotomers would result in smaller ring current shifts for the proline hydrogens, thereby contributing less to the average chemical shifts observed. Thus, although the rotamer form shown here is not necessarily the most stable of the three, it is consistent with the chemical shift data. In the conformation for trans-N-Tyr-MIF- 1 shown in Fig. 7b, the leucine, proline and tyrosine side chains form a hydrophobic face. Although this structure is appealing in terms of receptor binding, it is inconsistent with the chemical shift data. The aromatic ring is oriented in a manner that would cause upfield shifts of the proline C6H and CTHresonances and one of the leucine methyl groups is as close to the aromatic ring as the

proline C6H hydrogens. For this structure, increasing temperature would be expected to cause a shift in the proline resonances opposite to that observed in Fig. 3 and a much greater shift in the leucine C6H resonances. Although fewer of the resonances of the minor conformation could be followed over the entire temperature range, none seems to undergo the very large chemical shift changes observed for the more prominent structure (Fig. 4). In particular, the interaction between tyrosine and the proline ring does not appear to be as great as in the major conformation. However, the bimodal temperature dependence of the tyrosine and proline C"H resonances (Fig. 4) does provide some evidence for a structural transition. One of the conformations for cis-N-Tyr-MIF-1 is presented in Fig. 8. In this conformer, the aromatic ring is on the opposite face of the proline ring, which could explain the greater temperature dependence of the proline alpha hydrogen in the minor component (Fig. 4). REFERENCES 1. Kastin, A.J., Stephens, E., Ehrensing, R.H. & Fischman, A.J. (1984) Pharmacol. Biocheni. Behav. 21, 937-941 2. Kastin, A.J., Stcphcns, E., Zadina, J.E., Coy, D.H. & Fischman, A.J. (1985) Pharmacol. Biochem. Behav. 23, 1045- I049 3. Pulvircnti, L. & Kastin, A.J. (1988) European J . Pharrirucoi. 151, 289-292 4. Hayashi, T., Fischman, A.J., Kastin, A.J. & Coy, D. (1984) Pharmacol. Biochem. Behav. 21, 809-812 5. Hayashi, T., Kastin, A.J., Coy, D. & Olson, R.D. (1983) Brairi Res. Bull. 11, 659-662 6. Kastin, A.J., Abel, D.A., Ehrensing, R.H., Coy, D.H. & Graf. M.V. (1984) Pharmacol. Biochem. Behav. 21, 767-771 7. Kastin,A.J., Lawrence, S.P.&Coy,D.H. (1981)BruirrRes.Bid/. 1,697-702 8. Zadina, J.E., Kastin, A.J., Krieg, E.F. & Coy, D.H. (1982) Pharmacol. Biochen?. Behav. 17, 1193-1 198 9. Horvath,A. & Kastin, A.J. (1989)J. Biol. Chrm. 264,2175-2179 10. Horvath, A. & Kastin, A.J. (1990) f n t . J . Peptide Proreir?Res. 36. 281-284 11. Deslauriers, R., Walter, R. & Smith, I.C.P. (1973) Biochmi. Biophys. Res. Comm. Un. 53, 244-250 12. Deslauriers, R., Somorjai, R.L. & Ralslon, E. (IY77) Nanrre 266. 746-748 13. Higashijima, T., Tasumi, M., Miyazawa T. & Miyoshi. M. ( I 978) European J . Biochem. 89, 543-556 14. Walter, R., Deslauriers, R. & Smith, I.C.P. (1978) FEBS Lett. 95, 357-360 15. Deslauriers, R., Walter, R. & Smith, I.C.P. (1973) FEBS Leu. 31, 27-32 16. Hruby, V.J., Brewster, A.T. & Glassel, J . A . (1971) Proc. Nu//. Acad. Sci. USA 68,450-453 17. Vander Elst, P., Elseviers, M., De Cock, B., Van Marsenille, M.. Tourwe, D. & Van Binst, G . (1986) fnr. J . Pepride Prorein Res. 27, 633-642 18. Kang, S. & Walter, R. (1976) Proc. Natl. Acud. Sri. USA 13, 1203-1206 19. Walter, R., Bernal, I. & Johnson, L.F. (1972) Chem. Bid. Peptides, 131- 140

47

M. Petersheim et al. 20. Higashijirna T., Tasumi, M . & Milazaaa. T. (1975) FEES Leu. 57, 175-178 21. Naik, V.M. & Krimm, S. (1985) fnr. J . Pepride Proreiri R e x 23, 1-24 22. Ralston, E., De Coen, J.L. & Waltcr, R. (1974) Proc. ,\arl. Acad. Sci. USA. 71, 1142-1144 23. Schwartz, R.W., Mattice, W.L. & Spirtes, M.A. (1979) Biopoli,meis 18, 1835-1848 24. Reed, L.I. & J o h n , P.L. (1973)J. Aru. Chern. Soc., 7523-7524 25. Bax, A,, Freeman, R. & Morris, J. (1981) J . Magif. Resori. 42. 164- 168 26. Widcr, G., Macura, S., Kumar, A., Ernst, R.R. & Wuthrich, K. (1984) J . Magn. Reson. 56, 207-234 27. Hore, P.J. (1983) J. Magn. Reson. 55, 283-300 28. Buckcrt, U. & Allinger, N.L. (1982) Molecular Mechairics. American Chemical Society, Washington, D C 29. Benedetti, E. (1977) feprides, Proceedings of the Fifth American Peptide Symposium, pp. 257-273. John Wile) & Sons, New York 30. Kopple, K.D., Ohnishi, M. & Go, A. (1969) Biocheniisrq 8, 4087-4095

48

31. Corio, PL. (1966) Structure of High Resolution N M R Spectra. Academic Press, New York 3 2 . Pachler, K.G.R. (1963) Specrrochim. Acta 19, 2085-2092 33. Fischman. A.J., Live, D.H., Wysshrod, H.R., Agosta, W.C. & Cowburn, D. (1980) $1.A m . Chem. Sac. 102, 2533-2539 34. Wagner, G., Bodenhausen, G., Muller, N., Rance, M. Sorenscn, O.W., Ernst, R.R. & Wuthrich, K. (1985)J.Am. Chem. SOC.107, 6440-6446

Address: Dr. Alun J . Fischnian Division of Nuclear Medicine Department of Radiology hlassachusetts General Hospital 32 Fruit Street Boston. MA 02 114 USA