T. Yagami
`K. Kitagawa
`C. Aida
`H. Fujiwara
`S. Futaki
`
`Stabilization of a tyrosine
`O-sulfate residue by a
`cationic functional group:
`formation of a conjugate
`acid±base pair
`
`Authors' af®liations:
`
`Key words: cholecystokinin; conjugate acid±base pair;
`
`T. Yagami, National Institute of Health Sciences,
`
`desulfation; gastrin II; liquid secondary-ion mass spectrometry;
`
`Tokyo, Japan.
`
`MALDI-TOFMS; sulfated tyrosine
`
`K. Kitagawa, C. Aida and H. Fujiwara, Niigata
`
`College of Pharmacy, Niigata, Japan.
`
`S. Futaki, Institute for Chemical Research, Kyoto
`
`University, Japan.
`
`Correspondence to:
`Correspondence to: Dr Kouki Kitagawa
`
`Niigata College of Pharmacy
`
`5-13-2 Kamishin'ei-cho
`
`Niigata
`
`Niigata 950-2081
`
`Japan
`
`Tel.: 81-25-268-1241
`
`Fax: 81-25-268-1230
`
`E-mail: kouki@niigata-pharm.ac.jp
`
`Dates:
`
`Received 17 December 1999;
`
`Revised 7 March 2000;
`
`Accepted 19 May 2000
`
`To cite this article:
`
`Yagami, T., Kitagawa, K., Aida, C., Fujiwara, H. &
`
`Futaki, S. Stabilization of a tyrosine O-sulfate residue by a
`
`cationic functional group: formation of a conjugate acid±
`
`base pair.
`
`J. Peptide Res., 2000, 56, 239±249.
`
`Copyright Munksgaard International Publishers Ltd, 2000
`
`ISSN 1397±002X
`
`Abstract: Sulfated tyrosine [Tyr(SO3H)]-containing peptides
`showed characteristic peak patterns in their liquid secondary-ion
`
`mass spectrometry (LSIMS) spectra. Protonated molecules were
`
`desulfated more easily than their deprotonated counterparts.
`
`Therefore, the stabilities of the Tyr(SO3H) residues were well-
`re¯ected by peak patterns in their positive-ion spectra. These
`
`intrinsic peak patterns were investigated by comparing the
`
`behavior of each Tyr(SO3H) residue in acidic solution. As the
`peptide chain was lengthened and the number of cationic
`functional groups increased, the peak representing the [MH]+ of a
`
`Tyr(SO3H)-containing peptide became more prominent than that
`representing the desulfated [MH±SO3]+. These alterations in
`peptide structure also increased the stability of the Tyr(SO3H)
`residue in acidic solution. Based on the desulfation mechanism of
`
`an aryl monosulfate, we predicted that intramolecular cationic
`
`functional groups would stabilize Tyr(SO3H) residues by forming
`conjugate acid±base pairs (or salt bridges) both in the gaseous
`
`phase and in acidic solution. In accordance with this theory, Arg
`
`residues would take primary responsibility for this self-stabilization
`
`within Tyr(SO3H)-containing peptides. Moreover, a long peptide
`backbone was expected to have a weak protective effect against
`desulfation of the [MH]+ in the gaseous phase. Tyr(SO3H) residues
`were also stabilized by adding an external basic peptide containing
`
`multiple Arg residues. Formation of such intermolecular acid±base
`
`pairs was demonstrated by matrix-assisted laser desorption/
`
`ionization time-of-¯ight mass spectrometry (MALDI-TOFMS) which
`
`detected conjugated peptide ions. The energetically favorable
`
`formation of conjugate acid±base pairs prompted by Tyr(SO3H)
`residues might be a driving force for protein folding and protein±
`
`protein interaction.
`
`239
`
`-239-
`
`
`
`
`MAIA Exhibit 1021
`MAIA V. BRACCO
`IPR PETITION
`
`

`

`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`Abbreviations: AcOH, acetic acid; AcONH4, ammonium acetate;
`Boc, tert-butyloxycarbonyl; CCK, cholecystokinin; DIPCDI,
`
`diisopropylcarbodiimide; DMF, N,N-dimethylformamide; Fmoc,
`
`¯uoren-9-ylmethyloxycarbonyl; FT-IR, Fourier transform infrared
`
`spectrometry; G-II, gastrin-II; HOBt, 1-hydroxybenzotriazole;
`
`LSIMS, liquid secondary-ion mass spectrometry; MALDI-TOFMS,
`
`matrix-assisted laser desorption/ionization time-of-¯ight mass
`
`spectrometry; NMM, N-methylmorpholine; Pbf,
`
`2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl; RP-HPLC,
`
`reversed-phase high-performance liquid chromatography; tBu,
`
`tert-butyl; TFA, tri¯uoroacetic acid
`
`The sulfation of tyrosine residues is one of the ubiquitous
`
`post-translational modi®cations of peptides and proteins
`
`(1±3). This modi®cation is catalyzed by tyrosylprotein
`
`sulfotransferase, a membrane-bound enzyme within the
`
`trans-Golgi compartment. It has been proposed that sulfate
`
`plays an important role in speci®c protein±protein interac-
`
`tions during protein transport. According to this theory,
`negative charges on tyrosine O-sulfate [Tyr(SO3H)] residues
`might participate in intermolecular recognition (3). They may
`
`also affect protein folding by interacting with cationic
`residues within Tyr(SO3H)-containing peptides. Despite its
`biological importance, this post-translational modi®cation is
`hard to detect. It is well known that Tyr(SO3H) residues tend
`to rapidly desulfate to Tyr under acidic conditions. Thus, due
`
`to their intrinsic acid lability, traditional analytical methods,
`
`for example amino acid analysis of the acid hydrolysate or
`
`Edman degradation, cannnot be performed on peptides and
`proteins containing Tyr(SO3H) residues.
`Since the late 1980s, mass spectrometry using soft-
`
`ionization techniques, such as fast atom bombardment
`
`ionization, MALDI and electrospray ionization, has seemed
`
`to be a very promising analytical tool for the examination of
`peptides and proteins containing Tyr(SO3H) residues (4±10).
`In previous studies, we observed characteristic fragmenta-
`
`tion patterns in the positive- and negative-ion LSIMS spectra
`of peptides containing multiple Tyr(SO3H) residues (11).
`Moreover, we selectively detected the [MH]+ of Tyr(SO3H)-
`containing peptides from crude peptide mixtures using
`
`_F\_ i
`_F\_
`R~°i""°H ..,..!.. R ~ 0 ~ - - R~OH
`+
`
`0
`
`-0-H '7
`-H· ~•H· 0 X
`+X -0 11
`-0- 11
`
`R
`
`0
`0
`+
`0-S--0- ~ R ~ g 0-S-0- X +
`x•
`II
`II
`o
`-
`o
`(X• : counterion)
`Figure 1. Plausible mechanism for desulfation and stabilization of a
`Tyr(SO3H) residue.
`
`Tyr(SO3H) residues rapidly decompose to Tyr residues in
`nonpolar organic solvents, as well as in strongly acidic
`
`solutions (14,15). Accordingly, we can rationally predict that
`the proton-surplus [MH]+ of Tyr(SO3H)-containing peptides
`rapidly loses its sulfate in the gaseous phase (a nonpolar
`
`the completely desulfated
`environment). As a result,
`[MH±nSO3]+, where n represents
`fragment
`ion,
`the
`number of Tyr(SO3H) residues, should appear prominently
`in the positive-ion spectrum. The remarkable lability of the
`[MH]+ also allows its selective detection by constant
`neutral-loss (80 amu) scanning. In contrast to [MH]+, the
`proton-de®cient [M±H]± of Tyr(SO3H)-containing peptides
`must be relatively stable even in the gaseous phase. We
`expect that at least one Tyr(SO3H) residue should remain
`intact without becoming an amphoteric ion subject to
`
`desulfation. As a result, the characteristic ladder fragmenta-
`tion pattern, [M±H±mSO3]± (m51,2,..., n±1), should appear
`in the negative-ion spectrum.
`
`From the same desulfation mechanism (Fig. 1), we
`predicted that Tyr(SO3H) residues become stable by forming
`conjugate acid±base pairs or ion pairs with cations other
`
`than protons. Intramolecular cationic functional groups are
`
`candidates for the conjugate base. If such conjugate acid±
`base pairs are actually formed, Tyr(SO3H) residues would
`become stable not only in acidic solutions, but also in the
`
`gaseous phase. In order to investigate this assumption, we
`®rst compared the stability of Tyr(SO3H) residues within
`various sulfated peptides in the gaseous phase and in acidic
`
`solution.
`
`constant neutral-loss (80 amu) scanning (12).
`
`Experimental Procedures
`
`Both the appearance of characteristic fragmentation
`patterns and the ability to selectively detect Tyr(SO3H)-
`containing peptides could be explained by a proposed
`
`Fluoren-9-ylmethyloxycarbonyl (Fmoc) amino acid deriva-
`
`tives and 2-chlorotrityl (Clt) resin (16) (substituted level;
`
`desulfation mechanism (13)
`
`(Fig. 1). According to this
`
`1.47 mmol/g) were purchased from Watanabe Chemical Co.,
`
`model,
`desulfation of
`aryl monosulfates,
`including
`Tyr(SO3H) residues, is catalyzed by protons and accelerated
`under nonpolar conditions. Indeed, it has been reported that
`
`Ltd. (Hiroshima, Japan). Other chemicals were of analytical
`
`grade. Authentic samples of tyrosine-O-sulfated Leu-enke-
`phalin [H-Tyr(SO3H)-Gly-Gly-Phe-Leu-OH] (17) and CCK-8
`
`240 | J. Peptide Res. 56, 2000 / 239±249
`
`-240-
`
`
`
`
`

`

`were obtained from the Peptide Institute, Inc. (Osaka, Japan).
`
`Peptides used in the MALDI-TOFMS study [H-RRLSSLRA-
`OH (S6-1), Ac-MLF-OH and H-KHG-NH2 (Bursin)] were
`purchased from Novabiochem (Tokyo, Japan). The amino
`
`acid composition of an acid hydrolysate was determined
`
`using a Hitachi 8500 model amino acid analyzer. FT-IR
`
`spectra were recorded on a Perkin±Elmer 1720 spectrometer.
`
`RP-HPLC was performed on a Hitachi L-6200 model.
`
`Synthesis of sulfated peptides
`
`CCK peptides and G-II peptides
`
`Human cholecystokinin (CCK) and gastrin-II (G-II) peptides
`
`(Table 1) were prepared using the facile solid-phase method
`
`developed by Kitagawa et al. (18). The purity of each peptide
`
`exceeded 98% on RP-HPLC. Details of their syntheses are
`
`reported elsewhere.
`
`N-Acetyl tyrosine-O-sulfated Leu-enkephalin
`
`N-Acetyl tyrosine-O-sulfated Leu-enkephalin was prepared
`
`using Fmoc-Leu-Clt-resin (0.15 mmol) as a starting mate-
`rial. Each residue including Fmoc-Tyr(SO3Na)-OH (19,20)
`was introduced using the DIPCDI/HOBt coupling protocol
`
`[Fmoc-amino acid (3 eq.), DIPCDI (3 eq.) and HOBt (3 eq.);
`90 min]. After introducing the N-terminal Tyr(SO3Na)
`residue, the free a-amino group was acetylated with acetic
`
`anhydride
`
`(1.0 mmol)
`
`in the
`
`presence
`
`of
`
`pyridine
`
`(1.0 mmol). The peptide-resin obtained (50 mg) was treated
`with a mixture of AcOH/tri¯uoroethanol/CH2Cl2 (1 : 1 : 3 v/
`v, 5 mL) for 20 min at 208C, then ®ltered. After the ®ltrate
`had been concentrated with a stream of N2, dry ether
`(50 mL) was added to it. The resultant precipitate was
`
`collected by centrifugation and lyophilized from 0.025 m
`NH4HCO3 (30 mL); weight 9.80 mg. Part of the crude
`peptide was puri®ed by RP-HPLC [HPLC conditions:
`column, Cosmosil 5C18-AR (103250 mm); elution system,
`
`Table 1. Amino acid sequences of CCK and G-II peptides
`
`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`a linear gradient of CH3CN/0.1 m AcONH4 (20±25% in
`60 min); ¯ow rate, 2.5 mL/min; absorbance was detected
`
`at 265 nm]. The integrity of the sulfate was con®rmed by
`FT-IR (1050 cm±1).
`
`CCK-12 derivatives
`[Lys4],
`[Leu4] and [Glu4]CCK-12,
`[H-Ile-Ser-Asp-X-Asp-
`Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2; X5Lys, Leu,
`or Glu], were prepared from 0.1 mmol of Fmoc-Asp (Clt
`resin)-Phe-NH2 (18,21) (Fig. 2). The DIPCDI/HOBt coupling
`protocol was used to elongate each derivative. After
`assembly, the protected peptide-resin (< 80 mg) was treated
`with a mixture of AcOH/tri¯uoroethanol/CH2Cl2 (1 : 1 : 3,
`4 mL) at 258C for 20 min; the detached peptide was then
`treated with 90% aqueous TFA (2 mL) at 08C for 7 h. The
`crude peptide was puri®ed by RP-HPLC [column, mBonda-
`sphere 5C18 100 AÊ
`(193150 mm); elution system, a linear
`gradient of CH3CN/0.1 m AcONH4 at a ¯ow rate of
`1.75 mL/min; absorbance was detected at 265 nm]. Yields
`
`of the puri®ed peptides were between 31 and 37% based on
`
`the protected peptide-resin. Amino acid ratios of the acid
`hydrolyzates were as follows: [Lys4]CCK-12, Asp 3.03, Ser
`0.90, Gly 1.00, Met 1.74, Ile 1.05, Tyr 1.00, Phe 0.96, Lys
`1.04, Trp ND (not determined); [Leu4]CCK-12, Asp 2.96, Ser
`0.88, Gly 1.00, Met 1.43, Ile 1.01, Leu 1.00, Tyr 0.98, Phe
`0.95, Trp ND; [Glu4]CCK-12, Asp 3.03, Ser 0.89, Glu 1.03,
`Gly 1.00, Met 1.53, Ile 1.04, Tyr 1.01, Phe 0.96, Trp ND. The
`integrity of the sulfate was con®rmed by FT-IR (1049 cm±1).
`The detected molecular mass (monoisotopic mass) of each
`
`peptide coincided with calculated values.
`
`Cionin
`
`The
`protochordate-derived
`cionin
`[H-Asn-Tyr(SO3H)-
`Tyr(SO3H)-Gly-Trp-Met-Asp-Phe-NH2] was prepared using
`the Fmoc-based solid-phase approach (22).
`
`CCK Peptides
`
`CCK-8
`
`CCK-12
`
`CCK-22
`
`CCK-33
`
`G-II Peptides
`
`H-Asp-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2
`
`H-Ile-Ser-Asp-Arg-Asp-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2
`
`H-Asn-Leu-Gln-Asn-Leu-Asp-Pro-Ser-His-Arg-Ile-Ser-Asp-Arg-Asp-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2
`
`H-Lys-Ala-Pro-Ser-Gly-Arg-Met-Ser-Ile-Val-Lys-Asn-Leu-Gln-Asn-Leu-Asp-Pro-Ser-His-Arg-Ile-Ser-Asp-Arg-Asp-Tyr(SO3H)-Met-Gly-Trp-
`
`Met-Asp-Phe-NH2
`
`mini gastrin-II (G-14)
`
`H-Trp-Leu-Glu-Glu-Glu-Glu-Glu-Ala-Tyr(SO3H)-Gly-Trp-Met-Asp-Phe-NH2
`
`little gastrin-II (G-17)
`
`Pyr-Gly-Pro-Trp-Leu-Glu-Glu-Glu-Glu-Glu-Ala-Tyr(SO3H)-Gly-Trp-Met-Asp-Phe-NH2
`
`big gastrin-II (G-34)
`
`Pyr-Leu-Gly-Pro-Gln-Gly-Pro-Pro-His-Leu-Val-Ala-Asp-Pro-Ser-Lys-Lys-Gln-Gly-Pro-Trp-Leu-Glu-Glu-Glu-Glu-Glu-Ala-Tyr(SO3H)-Gly-Trp-
`
`Met-Asp-Phe-NH2
`
`J. Peptide Res. 56, 2000 / 239±249 | 241
`
`-241-
`
`
`
`
`

`

`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`9
`
`Fmoc-Asp-Phe-NH2
`
`0
`
`1~-
`i 1) AcOH I CF3CH~H I CH202
`
`CCK-12
`
`225nm
`
`Figure 2. Synthetic scheme for CCK-12 and
`its derivatives. Inset: RP-HPLC
`chromatogram of crude CCK-12 (X5Arg)
`obtained after deprotection. An asterisk
`shows the peak representing the desulfated
`peptide (CCK-12 nonsulfate). HPLC
`conditions [column, Cosmosil 5C18 AR
`(4.63150 mm); elution system, a linear
`gradient of CH3CN/0.1 M AcONH4
`(20±35% in 30 min); ¯ow rate, 1 mL/min;
`absorbance was detected at 225 nm].
`
`H-lle-Ser('eu)-Asp(O'eu)- X -Asp{O'Bu)-Tyr(S031-Met-Gly-Trp-Met-l-Phe-NH2
`
`2) 90% aqueous TFA (O "C)
`3) Preparative RP-HPLC
`H-lle-Ser-Asp- X -Asp-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2
`
`[ X = Arg(Pbf), Lys(Boc), Leu, or Glu(O'eu). ]
`O = 2-chlorotrityl resin
`IX = Arg, Lys, Leu, or Glu]
`
`Desulfation rates of Tyr(SO3H)-containing peptides in acidic
`solution
`
`counterparts are listed in Table 2, along with the molecular
`
`masses of the sulfated forms.
`
`HPLC-puri®ed G-17 (540 mg) was dissolved in an ice-cooled
`mixture of CH3CN and H2O (2 : 3 v/v, 500 mL) containing
`5% TFA, and then incubated at 378C. An aliquot of the
`solution was withdrawn at appropriate intervals (0.5, 1, 2, 3
`
`and 5 h) and immediately lyophilized. Each lyophilizate was
`reconstituted with ice-cooled 15% CH3CN in 0.1 m
`AcONH4 (100 mL), 20 mL of this solution was then applied
`to RP-HPLC [column, Cosmosil 5C18-AR (3.93150 mm);
`elution system, a linear gradient of CH3CN/0.1 m AcONH4
`(17±32% in 30 min); ¯ow rate,1 mL/min; absorbance was
`
`detected at 225 nm].
`
`Desulfation of the other peptides was similarly examined.
`
`Retention times of intact peptides and their nonsulfated
`
`Mass spectrometry
`
`LSIMS spectra were obtained using a VG ZAB-2SE double-
`
`focusing mass spectrometer (VG Analytical, Ltd, Manche-
`
`ster, UK) with an OPUS operating data system. Each HPLC-
`puri®ed peptide was dissolved in a mixture of H2O/CH3CN
`(1 : 1) at a concentration of < 10 mg/mL. Glycerol, thiogly-
`cerol and m-nitrobenzyl alcohol were used as the matrix,
`either neat or in combination. Typically, 2 mL of matrix was
`deposited on the target, and 1 mL of the peptide solution was
`added and mixed with the matrix. Ionization of the sample
`was performed using < 1 mA of cesium ions which were
`applied with a cesium ion gun. All spectra were recorded at
`
`Table 2. Molecular masses and retention times of CCK and G-II peptides
`
`tR on HPLC (min)a
`
`Temp
`
`Formula and molecular mass
`
`Sulfate
`
`Nonsulfate
`
`CCK Peptides
`
`20±35% in 30 min
`
`CCK-8
`
`CCK-12
`
`CCK-22
`
`CCK-33
`
`C49H62N10O16S3
`
`C68H95N17O23S3
`
`C117H173N35O39S3
`
`C167H263N51O52S4
`
`[Lys4]CCK-12
`
`C68H95N15O23S3
`
`[Leu4]CCK-12
`
`C68H94N14O23S3
`
`[Glu4]CCK-12
`
`C67H90N14O25S3
`
`G-II Peptides
`
`G-14
`
`G-17
`
`G-34
`
`C85H109N17O30S2
`
`C97H123N20O34S2
`
`C176H251N43O56S2
`
`1142.4b
`
`1613.6b
`
`2790.0c
`
`13.4
`
`12.0
`
`15.1
`
`3945.5c
`
`24.7d
`
`1585.6b
`
`8.7
`
`1570.6b
`
`11.8
`
`1586.5b
`
`5.5
`
`17±32% in 30 min
`
`1913.0c
`
`2177.3c
`
`3929.3c
`
`14.3
`
`15.0
`
`17.7
`
`16.4
`
`15.4
`
`17.3
`
`26.5d
`
`12.4
`
`16.2
`
`9.8
`
`18.0
`
`17.9
`
`19.5
`
`a. HPLC conditions: column, Cosmosil 5C18 AR (4.63150 mm); elution system, a linear gradient of
`CH3CN/0.1 M AcONH4; ¯ow rate,1 mL/min; absorbance was detected at 225 nm. b. Calculated as
`monoisotopic mass. c. Calculated as average mass. d. 20±40% in 60 min at a ¯ow rate of 1 mL/min.
`
`242 | J. Peptide Res. 56, 2000 / 239±249
`
`-242-
`
`
`
`
`

`

`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`an accelerating voltage of 8 kV and at a resolution of 1000.
`The scan rate was chosen so that each scan took < 20 s, and
`data from several scans were generally summed over the
`
`22 and 3946.5 for CCK-33) became more prominent than
`that of the desulfated [MH±SO3]+ which exhibited negligible
`intensity. These peak patterns were not signi®cantly
`
`course of 1±2 min. Data accumulated in multichannnel
`
`analyzer (MCA) mode were processed by standard practices
`
`employed by the software. No corrections were made to
`
`affected by changing the matrix for ionization. In contrast,
`the peak representing the [M±H]± of each peptide (m/z
`1141.3 for CCK-8, 1612.5 for CCK-12, 2788.9 for CCK-22
`
`account for background from the matrix. Mass calibration
`
`and 3944.6 for CCK-33) appeared consistently as the base
`
`was carried out using clusters of cesium iodide to set an
`
`peak in the negative-ion LSIMS spectra. We could not detect
`
`external standard. The pseudo-molecular ions of peptides
`
`any desulfated CCK-8 on a RP-HPLC chromatogram even
`
`that exceed a molecular mass of 1700 Da, are shown as
`
`after CCK-8 was exposed to a 3-min measurement of its
`
`average masses (Table 2).
`
`LSIMS spectrum in the positive-ion mode.
`
`In order to examine the possibility that desulfation could
`
`occur in the matrix during mass spectrometric analysis, CCK-
`
`8 was dissolved in glycerol and measured. After being
`
`bombarded for 3 min with cesium ions, the sample was
`
`withdrawn from the target and examined by RP-HPLC
`[column, Wakosil-II 5C18 (4.63150 mm); elution system, a
`linear gradient of CH3CN/0.1 m AcONH4 (15±45% in
`30 min); ¯ow rate, 1 mL/min; absorbance was detected at
`
`280 nm].
`
`MALDI-TOFMS
`
`MALDI-TOFMS spectra were obtained using a Shimadzu/
`
`Kratos Kompact MALDI IV time-of-¯ight mass spectro-
`
`meter ®tted with a nitrogen laser (337 nm). Either CCK-8 or
`cionin was dissolved in H2O/CH3CN (1 : 1, v/v) and mixed
`with 5 eq. of a peptide (S6-1, Ac-MLF-OH or H-KHG-NH2).
`The mass spectrum of each mixture was recorded in linear
`
`mode at a 20-kV accelerating voltage using sinapinic acid as
`
`the matrix for ionization. Pseudo-molecular ions registered
`
`on MALDI-TOFMS spectra as average masses.
`
`Results
`
`LSIMS spectra of CCK peptides
`
`The positive- and negative-ion LSIMS spectra of CCK
`
`peptides are shown in Fig. 3. In the positive-ion spectrum
`
`of CCK-8 (the shortest CCK peptide examined), a prominent
`[MH±SO3]+ peak (m/z 1063.3) accompanied a weak [MH]+
`peak (m/z 1143.3). However, there were signi®cant changes
`
`within the peak patterns in the positive-ion spectra of CCK
`
`peptides as peptide lengths were altered. In the spectrum of
`CCK-12, the [MH]+ (m/z 1614.5) and the [MH±SO3]+ (m/z
`1534.6) were detected at an almost equal intensity. How-
`
`ever, in the spectra of CCK-22 and CCK-33, the peak
`representing the [MH]+ of each peptide (m/z 2791.1 for CCK-
`
`LSIMS spectra of G-II peptides
`
`The positive- and negative-ion LSIMS spectra of G-II
`
`peptides are shown in Fig. 4. In the positive-ion spectra of
`
`G-14 and G-17, the base peaks were the desulfated [MH±
`SO3]+ (m/z 1833.7 for G-14 and 2098.4 for G-17). The peaks
`corresponding to the [MH]+ (m/z 1913.8 for G-14 and 2178.1
`for G-17) were very weak. In the positive-ion spectrum of G-
`34, a prominent [MH]+ peak (m/z 3929.9) accompanied a
`weak [MH±SO3]+ peak (m/z 3849.0). The [M±H]± peak of
`each peptide (m/z 1911.9 for G-14, 2176.3 for G-17, and
`
`3927.8 for G-34) appeared consistently as the base peak in
`
`the negative-ion LSIMS spectra.
`
`Stability of the Tyr(SO3H) residues in acidic solution
`
`In general, short peptides tended to desulfate more easily
`
`than longer ones in acidic solution (Fig. 5). This tendency
`
`was especially pronounced in CCK peptides; the order of the
`
`desulfation rate was CCK-8.CCK-12.CCK-22.CCK-33.
`
`However, size dependency could not be substantiated in G-II
`
`peptides; G-17 was most susceptible to desulfation and G-14
`
`was exceptionally stable in acidic solution.
`
`Effects of cationic functional groups in Tyr(SO3H)-containing
`peptides on the positive-ion LSIMS spectra and desulfation rates
`
`First, the positive-ion LSIMS spectra of tyrosine-O-sulfated
`
`Leu-enkephalin and its N-acetylated derivative were com-
`
`pared. In the spectrum of sulfated Leu-enkephalin, the
`[MH]+ peak (m/z 636.2) and the desulfated [MH±SO3]+ peak
`(m/z 556.3) were detected at comparable intensities. In
`contrast, a prominent [MH±SO3]+ peak (m/z 598.3) and a
`weak [MH]+ peak (m/z 678.3) were detected in the spectrum
`of the N-acetylated derivative. The relative dominance of
`[MH]+ over [MH±SO3]+ was well correlated with the stability
`
`J. Peptide Res. 56, 2000 / 239±249 | 243
`
`-243-
`
`
`
`
`

`

`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`10
`BO CCK-8
`60
`
`1063. 3
`[M+H-S03]+
`
`40
`
`20
`
`[M+H-S04]' ..._
`
`[M+H]'
`1143 .3
`
`CCK-8
`
`10
`
`BO
`
`60
`,o
`20
`
`1141.3
`-[M-H]"
`
`Figure 3. LSIMS spectra of human CCK
`peptides. A mixture of glycerol/thioglycerol
`(1 : 1) was used as the matrix for ionization.
`
`9 0
`
`9 0
`
`1000
`
`10
`BO CCK-12
`60
`,o
`20
`
`[M+H-S03]+
`
`161'. 5
`-[M+H]'
`
`1350 1'00 1,so 1500 1550 1600 1650
`2791.1
`-[M+H]'
`
`10
`BO CCK-22
`60
`•o
`20
`
`0
`
`2550 2600 2650 2700
`
`10
`BO CCK-33
`60
`
`40
`
`20
`
`m/z
`
`10
`
`BO
`
`60
`,o
`20
`
`m./z
`
`10
`
`BO
`
`60
`
`40
`
`20
`
`m/ z
`
`10
`
`BO
`
`60
`,o
`20
`
`9 0
`
`9 0
`
`CCK-12
`
`1000 1050 1100 1150
`16 2 . 5
`-[M-H)"
`
`m/z
`
`1350 1'00 1'50 1500 1550 1600 1650
`2788. 9
`-[M-H)"
`
`CCK-22
`
`m/z
`
`2550 2600 2650 2700
`
`m/z
`
`3700 3750 3800 3850 3900 3950
`
`m/z
`
`3700 3750 3800 3850 3900 3950
`
`m/z
`
`Positive
`
`Negative
`
`of the Tyr(SO3H) residues in acidic solution (data not
`shown).
`
`Second, CCK-12 derivatives with different residues at the
`X position [H-Ile-Ser-Asp-X-Asp-Tyr(SO3H)-Met-Gly-Trp-
`Met-Asp-Phe-NH2] were examined. The [M±H]± peak of
`each peptide consistently formed the base peak in their
`
`negative-ion LSIMS spectra (data not shown). In contrast,
`
`their peak patterns changed signi®cantly depending on the X
`
`residue in their positive-ion spectra (Fig. 6A, and Fig. 3 for
`
`CCK-12: X5Arg). When the X residue was a basic amino
`acid (Arg or Lys), an intense peak corresponding to [MH]+
`was recorded. Whereas, the desulfated [MH±SO3]+ became
`the prominent peak if Leu or Glu composed the X residue.
`The relative dominance of [MH]+ over [MH±SO3]+ correlated
`well with the acid stability of the Tyr(SO3H) residue in each
`CCK-12 derivative (Fig. 6B).
`
`10~
`90
`BO G-14
`70
`60
`so
`,o
`30
`20
`10
`
`1833. 7
`[M+H-SO:i]+
`
`[M+H-S04]'
`
`[M+H]'
`1913 . B
`
`1650 1700 1750 1800 1850 1900 1950
`2098 . 4
`
`10
`
`m/z
`
`G-14
`
`10
`90
`BO
`70
`60
`so
`,o
`30
`20
`10
`0
`
`[M-H-SOa]-
`
`2096.7
`
`1650 1700 1750 1800 1850 1900 1950
`2176.3
`-[M-H)"
`
`G-17
`
`10
`go·
`BO
`70
`60
`so
`,o
`30
`20
`10
`0
`1900 1950 2000 2050 2100 2150 2200
`3927. 8
`10
`90
`-[M-H]"
`BO G-34
`70
`60
`so
`,o
`30
`20
`10
`0
`3650 3700 3750 3800 3850 3900 3950
`
`Figure 4. LSIMS spectra of human G-II
`peptides. A mixture of glycerol/thioglycerol
`(1 : 1) was used as the matrix for ionization.
`
`m/z
`
`m/z
`
`m/z
`
`,.,.
`
`m/z
`
`[M+H-SO:i]+
`
`[M+H]'
`
`2178 .1
`
`90 G-17
`BO
`70
`60
`so
`•o
`30
`20
`10
`0
`1900 1950
`10
`90
`BO G-34
`70
`60
`so
`,o
`30
`20
`10
`0
`3650 3700 3750 3800 3850 3900 3950
`
`2000 2050 2100 2150 2200
`3929 . 9
`-[M+H] '
`
`[M+H-SOaJ'
`3849 . 0
`
`Positive
`
`Negative
`
`244 | J. Peptide Res. 56, 2000 / 239±249
`
`-244-
`
`
`
`
`

`

`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`B
`
`100
`
`..
`..
`
`-0- G •14
`
`-0--- G-34
`
`G-17
`
`70
`
`..
`. ..
`
`...
`
`. ..
`
`, ..
`Time (h)
`
`50
`
`, ..
`
`2.0
`
`...
`
`..•
`
`, ..
`
`A
`'"'
`..
`..
`.. -- CCK-12
`
`---- CCK-33
`
`-0- CCK-22
`
`-0- CCK-8
`
`70
`
`50
`
`..•
`
`Remaining sulfated peptide (%)
`
`Figure 5. Desulfation rates of (A) CCK and
`(B) G-II peptides in a solution containing 5%
`TFA. Each experiment was run twice and
`the averages plotted.
`
`1.0
`
`2.0
`
`Effects of externally added peptides on the positive-ion MALDI-
`
`TOFMS spectra of Tyr(SO3H)-containing peptides
`
`The positive-ion MALDI-TOFMS spectrum of CCK-8 mixed
`
`in advance with a cationic peptide (S6±1) is shown in
`
`Fig. 7A. The intermolecularly conjugated peptide-ions,
`[CCK-8+S6-1+H]+ (m/z 2102.4) and [23CCK-8+S6-1+H]+
`(m/z 3245.1), were intensely detected. Moreover, these
`
`[cionin+ S6-1+H]+
`[23cionin
`and
`2214.9)
`(m/z
`S6-1;
`+ S6-1+H]+ (m/z 3311.1; Fig. 8A). In contrast, when cionin
`was premixed with Bursin, a completely desulfated fragment
`ion, [cionin + H±2SO3]+ (m/z 1097.4), formed the base peak.
`No conjugated peptide ions were detected in the MALDI-
`
`TOFMS spectrum of this mixture (Fig. 8B).
`
`conjugated peptide ions did not lose their sulfates.
`
`In
`
`Discussion
`
`contrast, no conjugated peptide ions were detected in the
`
`MALDI-TOFMS spectrum of a mixture of nonsulfated CCK-
`
`8 and S6-1 (Fig. 7B). Conjugated peptide ions were also not
`
`detected in the MALDI-TOFMS spectra of CCK-8 when it
`
`Very recently, we developed a facile solid-phase method for
`the synthesis of Tyr(SO3H)-containing peptides (18). Using
`this method, we prepared human CCK and G-II peptides
`
`was premixed with Ac-MLF-OH or Bursin (data not shown).
`Cionin, a peptide containing two Tyr(SO3H) residues, also
`formed conjugated peptide ions when it was premixed with
`
`without any dif®culties. These peptides with various chain
`
`lengths and cationic functional groups were studied with
`
`regard to their degree of
`
`intramolecular conjugation,
`
`(M+H-S03t 1506.7 -
`
`1586.6
`-(M+Ht
`
`A
`10
`
`X=Lys
`
`B
`
`100
`
`90
`
`80
`
`70
`
`-0-- CCK-12
`60 -+-- (Lys)CCK-12
`-0-- (Glu)CCK-12
`---½-
`(Leu)CCK-12
`
`Remaining sulfated peptide (%)
`
`50
`0.0
`
`1.0
`
`2.0
`3.0
`Time (h)
`
`4.0
`
`5.0
`
`J. Peptide Res. 56, 2000 / 239±249 | 245
`
`10
`80
`60
`40
`20
`0
`1300
`
`X=Leu
`
`1350 1400 1450
`
`1500 1550 1600
`1491.7
`
`m/z
`
`(M+H-S03t _
`
`[M+Ht
`1571.6
`
`1350 1400 1450 1500 1550 1600
`1507.7
`(M+H-S03t_
`
`m/z
`
`X=Glu
`
`1459 .6
`
`[M+Ht
`1588.6
`
`10
`80
`60
`40
`20
`0
`
`Per cent
`
`1500 1550 1600
`m/z
`1350 1400 1450
`Figure 6. (A) Positive-ion LSIMS spectra of CCK-12 derivatives [H-Ile-Ser-Asp-X-Asp-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2]. See Fig. 3 for
`the positive-ion LSIMS spectra of CCK-12 (X5Arg). (B) Desulfation rates of CCK-12 and its derivatives in a solution containing 5% TFA.
`
`-245-
`
`
`
`
`

`

`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`(S6-1 + Ht
`
`·~··
`
`,/ [CCK-8 + H- S03)+
`
`[CCK-8 + S6-1 + HJ+
`2102.4
`
`(2 x CCK-8 + S6-1 + Ht
`32~5.1
`
`1000
`
`1500
`
`2000
`
`2500
`
`3000
`
`[S6-1 + HJ+
`15~.3
`
`Mass/Charge
`
`,t [CCK-8(N.S.) + Ht
`
`A
`
`B
`
`%Int
`100
`
`..
`..
`..
`
`20
`
`%Int
`100
`
`..
`..
`..
`
`20
`
`1000
`
`1500
`
`2000
`
`2500
`
`3000
`
`3500
`
`4000
`
`Mass/Charge
`Figure 7. Positive-ion MALDI-TOFMS spectra of (A) CCK-8 premixed with S6-1 (H-RRLSSLRA-OH) and (B) nonsulfated CCK-8 premixed with
`S6±1.
`
`speci®cally, that involving the Tyr(SO3H) residue. Peak
`patterns in their LSIMS spectra, especially those in their
`
`positive-ion spectra, were not due to preferential desorbance
`
`of contaminating impurities; homogeneity of the peptides
`and the integrity of each Tyr(SO3H) residue were con®rmed
`using several analytical methods in addition to mass
`
`spectrometry (HPLC, FT-IR). The possibility of desulfation
`
`in the matrix was also excluded.
`
`Several notable features were observed in the positive-ion
`
`LSIMS spectra of CCK and G-II peptides (Figs 3, 4). One
`distinct characteristic was that the [MH]+ peak became
`dominant over the peak representing the desulfated [MH±
`SO3]+ as the peptide chains were lengthened. This can be
`explained in part by speculating that long peptide chains
`protect labile Tyr(SO3H) residues against proton-catalyzed
`desulfation. The `protecting effect' of long peptide chains
`
`[Cionin + S6-1 + Hi+
`
`221_4.t
`
`[S6-1 + HJ+
`K~.I
`
`I
`,~ ..
`
`[Clonln + H- 2S03i+
`
`(2 x Clonln + S6-1 + HJ+
`J:3~1.1
`
`1000
`
`1500
`
`2000
`
`2500
`
`,~ ..
`
`[Clonln + H-2S03i+
`
`Mass/Charge
`
`....
`
`A
`
`B
`
`%Int.
`100
`
`..
`..
`..
`
`20
`
`%Int.
`100
`
`..
`..
`..
`
`20
`
`3500
`
`3500
`
`....
`
`1000
`
`1500
`
`2000
`
`2500
`
`....
`
`Mass/Charge
`Figure 8. Positive-ion MALDI-TOFMS spectra of (A) cionin premixed with S6-1 and (B) cionin premixed with H-KHG-NH2 (Bursin).
`
`246 | J. Peptide Res. 56, 2000 / 239±249
`
`-246-
`
`
`
`
`

`

`may be attributable to their large number of carbonyl oxygen
`
`which may result from a decrease in the effectiveness of
`
`atoms. These might collectively capture surplus protons
`from [MH]+ in the gaseous phase and prevent the formation
`of amphoteric ions which are required for desulfation
`
`ionic interactions due to the absence of Arg residues (Fig. 4).
`Intramolecular conjugation of Tyr(SO3H) residues with Arg
`residues should be effective in the gaseous phase as well as
`
`Yagami et al . Stabilization of a tyrosine O-sulfate in a peptide
`
`(Fig. 1). A second characteristic of their LSIMS spectra was
`that G-II peptides produced a desulfated [MH±SO3]+ more
`readily than CCK peptides. This observation cannot be
`
`The disparate stability of
`
`explained solely by the `protecting effect' described above.
`respective Tyr(SO3H)
`residues in the gaseous phase must be due to structural
`
`their
`
`in solution.
`
`The stabilization afforded by the formation of conjugate
`
`acid±base pairs (or salt bridges) is also expected to result
`from the conjugation of Tyr(SO3H) residues with a- or e-
`amino groups within peptides. Two sets of experiments, in
`
`which the lengths of the peptide chains were ®xed to
`
`differences between CCK peptides and G-II peptides
`
`exclude the `protecting effect' of the peptide backbones, gave
`
`(Table 1).
`We also examined the stability of Tyr(SO3H) residues in
`acidic solution (Fig. 5). The observed results can be
`
`summarized as follows:
`
`(i)
`
`in general, desulfation was
`
`retarded as peptide chains were lengthened, and (ii)
`Tyr(SO3H) residues in G-II peptides tended to desulfate
`more rapidly than those in CCK peptides. The `protecting
`
`results that were consistent with this prediction. First, the
`
`peak patterns of sulfated Leu-enkephalin and its N-
`
`acetylated derivative were compared. As expected, desul-
`fated [MH±SO3]+ formed the dominant peak in the spectrum
`of the N-acetylated derivative in which the Tyr(SO3H)
`residue cannot form any stabilizing ionic interactions.
`
`Second,
`
`four CCK-12 derivatives were examined with
`
`effect' of long peptide chains is probably not effective in
`
`respect
`
`to their peak patterns in LSIMS spectra and
`
`acidic solutions where desulfation-catalyzing protons are
`
`desulfation rates in acidic solution. The positive-ion spectra
`
`present in excess. We again need to consider the structural
`
`of peptides having cationic residues (X5Arg or Lys) were in
`
`differences between these two series of peptides to explain
`the diverse susceptibilities of their Tyr(SO3H) residues to
`desulfation.
`
`As already mentioned, a proton-catalyzed mechanism has
`
`been proposed which is based on the desulfation of aryl
`monosulfates (ArO±SO3H) in acidic solutions (13) (Fig. 1).
`According to this theory, if aryl monosulfates form ion pairs
`±X+), they
`with counterions other than protons (ArO±SO3
`should be less apt to decompose than their proton-donating
`forms (ArO±SO3H). This stabilization can be explained by
`assuming that a higher energy barrier would have to be
`
`overcome in order to form intermediate amphoteric ions
`
`contrast to those peptides having neutral residues (X5Leu)
`
`or acidic residues (X5Glu) (Fig. 6 and Fig. 3 for CCK-12).
`Stabilities of the Tyr(SO3H) residues within the peptides in
`acidic solution correlated with the relative dominance of the
`[MH]+ peak over the [MH±SO3]+ peak in each LSIMS
`spectrum. These results reveal the `stabilizing effect' of
`
`neighboring guanidino groups and a- or e-amino groups on
`Tyr(SO3H) residues.
`The `stabilizing effect' provided by these cationic func-
`
`tional groups is based on their ability to form conjugate acid±
`base pairs with Tyr(SO3H) residues. Our results indicate that
`the guanidino group of Arg is preferentially chosen by the
`
`during the course of decomposition. Therefore, the different
`
`sulfate as its conjugate base. Indeed, guanidino cations were
`
`acid stabilities of CCK and G-II peptides can be mainly
`
`shown to have greater af®nities for the sulfate in glycosa-
`
`ascribed to the formation of conjugate acid-base pairs
`between Tyr(SO3H)
`groups that are in close proximity. Moroder & WuÈ nsch
`
`residues and cationic functional
`
`minoglycans than ammonium cations by Linhardt and co-
`
`workers (24,25). Furthermore, Arg is known to have the
`
`highest proton af®nity among the 20 common amino acids
`
`(23) also pointed out intramolecular ionic interactions
`between Tyr(SO3H) residues and Arg residues in CCK
`peptides.
`By referring to the acid stability of Tyr(SO3H) residues in
`solutio