The aromatic 1H-NMR spectrum of plasminogen kringle 4

Share

Transcript

1 Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/20135978 Thearomatic1H-NMRspectrumof plasminogenkringle4.Acomparativestudyof human,porcineandbovinehomologs ArticleinEuropeanJournalofBiochemistrySeptember1986 DOI:10.1111/j.1432-1033.1986.tb09925.xSource:PubMed CITATIONS READS 23 26 4authors,including: LszlPatthy MiguelLlins HungarianAcademyofSciences CarnegieMellonUniversity 302PUBLICATIONS7,102CITATIONS 76PUBLICATIONS2,494CITATIONS SEEPROFILE SEEPROFILE AllcontentfollowingthispagewasuploadedbyLszlPatthyon26January2015. Theuserhasrequestedenhancementofthedownloadedfile.

2 Eur. J. Biochem. 159, 581 -595 (1986) 0FEBS 1986 The aromatic H-NMR spectrum of plasminogen kringle 4 A comparative study of human, porcine and bovine homologs Vasudevan RAMESH Marianne GYENES , Laszl6 PATTHY and Miguel LLINAS Department of Chemistry, Carnegie-Mellon University, Pittsburgh Institute of Enzymology, Biological Research Center, Budapest (Received April 14/July 1, 1986) - EJB 86 0370 The isolated kringle 4 domain of human plasminogen has been compared with homologous structures from bovine and porcine sources, both free and in the presence of the ligand 6-aminohexanoic acid, by two-dimensional H-NMR spectroscopies at 300 MHz and 600 MHz. The chemical-shift-correlated, spin-echo-correlated, and double-quantum-correlated aromatic spectra of the three proteins reveal that the globular conformation of the fourth kringle is closely maintained throughout the set of homologs. Direct comparison shows that the three conserved Trp residues (at sites 25, 62 and 72) which exhibit highly non-degenerate subspectra, find themselves in similar intramolecular environments. In particular, proton Overhauser experiments reveal that the close steric interaction between the Trp-I1 (Trp62 or TrpZ5)indole group and the aromatic ring at site 74 ( T Y ~or ~Phe74)is strictly preserved. This feature forces the kringle inner loop, closed by the Cys- Cy~ ~link, to fold back onto itself so as to place the site 74 residue proximal to the C y ~ ~ ~ - bridge. C ys ~ ~ Single-residue substitutions enable unambiguous assignments of His4 to His3, Tyr-I11 to Tyr4 and Tyr-IV to From this direct evidence, comparison with the kringle 1 spectrum, and the previously reported chemical modlfication of Tyr-I1 (Tyr) [Trexler M., Banyai L., Patthy L., Pluck N. D. & Williams R. J. P. (1985) Euv. J . Biochem. 152, 439-4461, Tyr-I and Tyr-V (the latter, an immobile ring on the 600-MHz time scale) could be assigned to Tyr2 and Tyr, respectively. Since Trp-111 has previously been assigned to Trp7 at the lysine-binding site, the present study completes the assignment of 10 out of 12 aromatic spin systems in the kringle 4 H-NMR spectrum; the only ambiguity which remains concerns the Trp-I and Trp-I1 indole spin systems, which are totally identified but as yet only tentatively assigned to Trp25and Trp6, respectively. Although the advent of two-dimensional N MR techniques vine, and porcine plasminogen sources, whose primary [l - 31 has greatly facilitated the analysis of complex protein structures have recently been reported [4 -61. spectra in terms of well-defined, connected spin systems, the Plasmin is a blood plasma serine protease primarily re- problem of unambiguously assigning the latter to specific sponsible for the dissolution of fibrin clots deposited on the residues in the primary structure remains a challenge for a walls of blood vessels. It is a glycoprotein composed of two structural interpretation of the spectroscopic data. This prob- chains, covalently linked by two cystine bridges [4]. The light lem can be approached via a number of ways but probably chain, of M , z 25000, carries the catalytic function, while the one of the most rewarding methods is to compare spectra heavy chain, of M , x 57000, is formed by a tandem array of of closely related homologous polypeptides, that differ in a iive highly homologous structural domains, of M , x 10000 limited number of amino acid residues and which can be each, known as kringles, whose role appears to be the binding unambiguously established to possess very similar, if not the of plasmin as well as of plasminogen, the inactive precursor same, solution conformations. In this paper we exploit this of plasmin, to the fibrin matrix that structures the clot. The idea to derive assignments for most of the aromatic signals in intact fourth kringle can be readily isolated via elastase diges- the spectrum of the intact human plasminogen kringle 4. tion of plasminogen 141. The kringle 4 fragment is known to Specifically, we have compared, using 1D and 2D H-NMR fold as a globular polypeptide [7-91, to be thermally stable spectroscopies, kringle 4 variants isolated from human, bo- up to z 330 K [lo- 121 and to carry a lysine-binding site which is likely to contribute to the attachment of plasminogen Correspondence ta M. LlinCs, Department of Chemistry, Car- to the blood clot [13, 141. The kringle lysine-binding site is negie-Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylva- known to exhibit affinity for other w-amino acids, ligand nia, USA-I5213 dipole analogs of L-lysine, such as 6-aminohexanoic acid Abbreviations. CIDNP, chemically-induced dynamic nuclear (cAhx) [13, 15, 161. polarization; D Q two-dimensional double-quantum-correlated A number of H-NMR studies have been published which spectroscopy; COSY, two-dimensional chemical-shift-correlated focus on the fourth kringle from human plasminogen [7, 8, spectroscopy; EAhx, 6-aminohexanoic acid; NOE, nuclear 12,17 - 221. Main emphases of those investigations have been Overhauser effect; rf, radio frequency; SECSY, two-dimensional spin the analysis of the kringle spectroscopic data vis d vis the echo chemical shift correlated spectroscopy; 1D, one-dimensional; 2D, two-dimensional; b, chemical shift in ppm; pH *, glass electrode amino acid constituents, conceptualizing the information in pH reading uncorrected for H isotope effect. terms of structural features pertaining to the lysine-binding Enzymes. Pancreatic elastase (EC 3.4.21.36); plasmin (EC site, and deriving rough conformational models for the kringle 3.4.21.7). in solution [8, 17, 191. As active participants in these

3 582 developments, we have remained concerned with achieving a total identification of the aromatic signals [18, 211. Human plasminogen kringle 4 contains one Phe, three His, three Trp and five Tyr residues which mostly are in close contact with each other and appear to provide important hydrophobic components to the lysine-binding site [8, 17, 19, 20,221. Thus far, besides Phe64, only the aromatic resonances arising from T ~ P ' ~Tyr4' , and Tyr5' rings have been identified by direct chemical modifications [8, 17, 211 while the imidazole spectrum of His3' could be unambiguously assigned by comparison with kringle 1, which conserves this residue 1171. The comparative criteria also served to substantiate the assignment of the Trp7' indole multiplets, as the side-chain group becomes replaced by a Tyr phenol ring in the first 10 kringle [7, 171. In this paper we further confirm many of the Fig. 1. Plasminogen kringle 4 : outline of the primary structure. The assignments listed above while we are able to complete the diagram shows the positions of aromatic amino acid residues in the identification of the His and Tyr aromatic components. This sequence of the human kringle. The numbering of residues begins at leaves the indole spectra of Trp-I and Trp-I1 (TrpZ5and the half-cystine closest to the N terminus. Two deletions, at sites 35 Trp6') as the sole aromatic spin systems still unassigned. and 59, have been introduced to facilitate comparison with other plasminogen kringles. Fully conserved aromatic residues in human, Fig. 1 outlines the kringle 4 contour on a planar projection; porcine and bovine homologs are denoted by solid circles. The re- also indicated are the aromatic residues found in the species placement of a histidine at position 3 and a substitution of a derived from human plasminogen. Of the three His residues, phenylalanine for a tyrosine at position 41 in the porcine variety are His3 is lost on going to the porcine homolog which serves to shown by solid semicircles. Substitution of a phenylalanine for a identify the signals from this residue in the 'H-NMR spectrum tyrosine at position 74 in the bovine homolog is indicated in a similar and, since the His3' signals have been assigned [17, 181, the manner. Site 64 is filled with a Phe residue only in the human kringle 4; His33spin system becomes automatically identified. A similar a Tyr residue substitutes for Phe64 in the bovine and porcine species argument applies to the Tyr41 and Tyr74 phenolic doublets, which are substituted by Phe multiplets in the porcine and in the bovine homologs, respectively. Since Tyr50 has already of the National NMR Facility for Biomedical Research at been assigned [21], this leaves us with the problem of re- Carnegie-Mellon University, Pittsburgh. Both data acquisi- cognizing Tyr' and Tyr9 subspectra. The latter we can tion and processing were done using the Aspect 2000 accomplish by comparing kringle 4 with the (human) minicomputers and Diablo 44 disk storage systems interfaced kringle 1, which lacks Tyr', and given that all the other to the spectrometers. The temperature inside the spectrometer common Tyr residues have been identified [17] (and Motta probes was measured with a calibrated methanol sample. A., Laursen R. A. & LlinLs M., unpublished observations), 1D NMR spectra were routinely acquired over 16 K data the identities of the phenol ring resonances from Tyr' and, by points to afford a digital resolution of 0.6 Hz (300 MHz) and exclusion, Tyr', become established. 1.1 Hz (600 MHz). The residual 'H'HO signal was sup- pressed by pre-irradiading with a selective, gated pulse for 1 s. Typically 4000- 8000 transients were collected for each experiment to achieve a satisfactory signal/noise ratio. Reso- MATERIALS AND METHODS lution enhancement was achieved via Gaussian multiplication Bovine and porcine plasminogens were isolated from fresh of the free induction decay [25]. Chemical shifts are referred citrated plasma by affinity chromatography on lysine- to the sodium trimethyl~ilyl-(2,2,3,3-~H)propionateline Sepharose 4B according to the method of Deutsch and Mertz using p-dioxane as internal standard [26]. [23]. Plasminogens were subjected to limited proteolysis with 2D spin-echo-correlated spectroscopy (SECSY) [27], porcine pancreatic elastase (Serva) and kringle 4 was isolated chemical-shift-correlated spectroscopy (COSY) [2, 281 and by affinity chromatography on lysine-Sepharose 4B and gel- double-quantum (DQ) experiments [29] at 300 MHz were exe- filtration on Sephadex G-75 according to described methods cuted by using the standard DISNMR program (version [4 - 61. The kringle 4 sample from human plasminogen was 830701) of the Bruker software applications package [l]. kindly provided to us by Prof. Laursen R. A., who generated COSY experiments at 600 MHz were performed using the it from plasminogen extracted from commercial (Miles labs.) FTNMR2D program (version 810515) of the same package. Cohn fraction I11 as published [15]. Both SECSY (300 MHz) and COSY (300 MHz and Deuterated solvents were products of Merck, Sharp and 600 MHz) experimental data were acquired with 512 equally Dohme Ltd, Canada. Kringle 4 samples for NMR measure- spaced evolution time period t l values and each t l spectrum ments were dissolved in 5-mm NMR tubes (Wilmad, Buena, averaged between 128 - 352 transients. The spectral width in NJ) after exchanging the labile amide protons of the protein the tz direction for 300-MHz experiments was usually about in 'H20, pH* 7, 310 K, for 3 h and lyophilizing the sample 3200 Hz with a block size of 2048 data points, while for 600- three times. The pH * of the samples were adjusted with dilute MHz COSY spectra the spectral width was usually about 'HCI or Na02H and the values quoted are uncorrected for 7250 Hz with a block size of 1024 data points. Quadrature deuterium isotope effects [24]. The sample concentration was detection was used in the f l and f 2 directions with the carrier usually about 1 mM in 0.35 ml ' H 2 0 for both 1D and 2 D placed in the middle of the spectrum. Appropriate phase- NMR experiments. cycling procedures were adopted to render N-type peak selec- 'H-NMR spectra were recorded in the Fourier mode both tion [l]. The residual 'H'HO resonance was suppressed at at 300 MHz (7.05 T) using a Bruker WM-300 NMR spec- 300 MHz by gated irradiation during the relaxation delay of trometer and at 600 MHz (14.10 T) using the spectrometer 2.3 s introduced between scans while at 600 MHz the same

4 583 peak was suppressed by irradiating the transition at all times tion. A relaxation delay of 1.25 s was introduced between except during data acquisition. The time domain data matrix successive scans. The difference spectrum was obtained by size of 512 x 2048 (300 MHz) was zero-filled once in the t l subtracting the spectrum with transient NOES (10000 scans) direction and resolution-enhanced with an appropriate from the reference spectrum (10000 scans). weighting function before Fourier transformation to yield a frequency domain data matrix size of 1024 x 1024. This gave RESULTS AND DISCUSSION a digital resolution of about 3.2 Hz along both t l and t2 The homology in aromatic amino acid residues of human, directions for COSY experiments and 1.8 Hz (tl) and 3.2 HZ ( t 2 ) for SECSY experiments. Similarly, at 600 MHz the time porcine and bovine plasminogen kringle 4 [4 -61 and the sub- domain data matrix size of 512 x 1024 was zero-filled in both stitutions in their sequence are highlighted in Fig.1 and in t , and tz directions and resolution-enhanced with an ap- Species Residue in kringle 4 site propriate weighting function before Fourier transformation to yield a frequency domain data matrix size of 1024 x 1024. 2 3 9 25 31 33 41 50 62 64 12 14 This gave a digital resolution of 7.25 Hz along both tl and t z directions. Frequency domain spectra in the absolute value mode are shown in the figures either as symmetrized (300 MHz) or unsymmetrized (600 MHz) contour plots. Bovine The DQ spectra at 300 MHz were measured by imple- menting the pulse scheme proposed by Mareci and Freeman the chart above. The assignment of aromatic ring proton [29] which is as follows: resonances in the 'H-NMR spectrum of human kringle 4 to x(x) -z - n(y) - z - YZ(x)- t l - 3n/(x) - tz(acq). specific residues in the sequence will be based mainly on comparison with the spectra of porcine and bovine kringle 4 A theoretical analysis of this pulse sequence to exite and detect varieties. For the purpose of uniformity with our previously double-quantum coherence has been given by Braunschweiler published NMR study on human plasminogen kringle 4 [18], et al. [30]. Quadrature detection along the t1 direction was the convention of labelling resonances of each type of aromat- implemented by repeating the pulse sequence twice at each ic residue by the standard IUPAC-IUB one-letter amino acid evolution time period t l setting; the second experiment code followed by a Roman numeral to indicate its number is consisted of a Y4 pulse with appropriate phase before the third retained. pulse in the sequence 111. This enabled the carrier to be set near the center of the spectrum and thus save data storage space [31]. Wagner and Zuiderweg [32] and Boyd et al. 1331 Histidyl spectra have clearly demonstrated the suitability of this pulse se- Fig.2 shows the 300-MHz COSY spectrum of human quence to simplify resolving scalar connectivities within the kringle 4 containing the ligand EAhx at pH * 7.2, together with aromatic spin systems of tyrosine, phenylalanine and the 1D reference spectrum. Presence of ligand narrows the tryphophan residues in protein NMR spectra, as a sequel spectrum to the extent of enabling detection of the long-range to the well-established COSY pulse sequence. The foremost scalar coupling between the previously identified H2 and H4 advantage of the DQ spectrum has been the absence of strong ring protons of the three histidyl residues H-I, H-I1 and H-111 diagonal peaks which cause complication in COSY spectra [18]. Cross-peak connectivities, denoted by solid lines, signify due to overlap with coupled cross-peaks with small chemical coupling between the cross-peaks with imidazole protons by shift differences. In particular, all the singlet peaks are reference to the diagonal; these correspond well with the suppressed as they do not give rise to DQ coherences. Further, position of the six singlets in the 6.8 -8.3-ppm region of the choice in the selection of the fixed preparation time period 1D reference spectrum and enable the pairing of the singlets delay z, which depends on the coupling constant J, provides as shown, i.e. at 7.68 and 7.05 ppm (H-I), 7.54 and 8.32 ppm a means of screening different multiplets or emphasizing the (H-II), and 8.29 and 6.86ppm (H-111). In an earlier paper intensity of a particular spin system [33]. we have reported the imidazole cross-peaks in the COSY Our DQ measurements at 300 MHz were acquired with a spectrum of human kringle 4 containing the ligand fixed preparation period delay z = 15 ms (y' J - ') optimized for aminomethyl-bicyclooctane-1-carboxylic acid (AMBOC) coupling between protons with J x 8 Hz and 256 equally [18]. However, that 2D spectrum was measured with an spaced evolution time period tl values. Each tl value was additional fixed delay of 50 ms following each evolution time averaged over 320 transients. The spectral width in the t2 increment in the COSY pulse sequence to emphasize the long- direction was usually about 3000 Hz with a block size of 2048 range coupling [36] whereas in Fig. 2B no such delay has been data points. Quadrature detection in both time domains was implemented. used with the carrier placed in the middle of the spectrum. We have previously assigned the spin system of H-I1 A 32-step phase cycling was used to select double-quantum (6 = 7.54 and 8.32 ppm) in the spectrum of human kringle 4 coherence and N-type peaks as prescribed [31]. The time to His3' by comparison with the spectrum of kringle 1 in domain data matrix size of 256 x 2048 was zero-filled in the which only His3' is conserved 117, 181. We adopt a similar tl direction and resolution-enhanced with an appropriate strategy in this paper to assign unambiguously the proton window function before Fourier transformation to yield a spin system of H-I (6 = 7.68 and 7.05 ppm) in the COSY frequency domain data matrix size of 1024 x 1024. This gave spectrum of human kringle 4 (Fig. 2) by comparison with the a digital resolution of 6.0 Hz along tl and 3.0 Hz along t 2 . 300-MHz COSY spectrum of porcine kringle 4 (Fig.3) in Contour plots of the frequency domain spectra in the absolute which His3 is deleted. Inspection of the latter spectrum shows value mode are shown in the figures. that the cross-peaks due to H-I are absent while those due to Transient 1D NOE difference experiments [34, 351 at H-II(6 = 7.56 and 8.37 ppm) and H-III(6 = 8.25 and 6.95 ppm) 600 MHz were executed by selectively irradiating the peak of are retained. The deletion of the cross-peaks is also manifested interest for 43 ms and waiting for 250 ms before data acquisi- in the absence of the two resonances of H-I in the 1D reference

5 584 fi H4 mH5 w-I hH7 THZ H6 1H4 W-II TH2 H 7 h hH6 H5 H7t mH 6 W-II H2bH4 H5 1i A H2 1 H-III IH4 H21 H-I IH4 ~ ~ ~ " ~ ~ ~ ~ ' " ~ ' l ~ ~ ~ ' ~ ' ~ ~ ' " ~ 8.0 7.0 6.0 5.0 SP( PPm Fig.2.300-MHz 'H-NMRspectra of human kringle 4 in the presence ofzdhx: aromatic rexion. (A) 1D reference spectrum (resolution-enhanced). Ring proton spin systems belonging to the three His residues, labelled H-I, H-I1 and H-111, and to the three Trp residues, labelled W-I, W-I1 and W-111, are indicated by reference to the scalar connectivities in the COSY contour plot (B). (B) Contour plot of the 2D COSY spectrum with sine-bell filtering along both dimensions. Off-diagonal cross-peaks, labelled by two-digit numbers, represent pairs of ring protons connected via scalar coupling. The 'matrix' convention (first digit labels row proton, second digit labels column proton) is followed to mark individual connectivities: thus, for example, connectivity 5-4 for W-I in the upper-left triangle means indole H5 resonates at higher field position relative to the J-coupled indole H4. According to this convention the same connectivity is labelled with reversed numbering if indicated on the symmetric cross-peak within the lower-right triangle. J-connectivities of the three ring protons for the Trp and the three His residues are indicated by solid lines. Notice that connectivities are indicated only once, either in the upper-left or lower-right triangle although the COSY spectrum shows both. Kringle concentration 1 mM, [ligand]/[protein] z 2, pH* 7.2, 310 K spectrum of the porcine homolog (Fig. 3A) and leads to its signals: it agrees fairly well with that of the human variety assignment as His3. Following the unambiguous assignment reflecting conservancy of the three His residues. It should be of the spin system of both H-I and H-I1 resonances (Fig. 2) to noticed that among the nine sharp peaks in the aromatic His3 and His31, respectively, the only remaining His spin spectrum of human kringle4 [17, 181, three represent Trp system, that of H-I11 (6 = 8.29 and 6.86 ppm) (Fig.2), indole H2 singlets. These are conserved in the three homologs becomes automatically assigned to His33. and can readily be distinguished from the His resonances as The spectrum of bovine kringle 4 (not shown) does not they exhibit similar chemical shifts in their corresponding add new information useful for the assignment of imidazole spectra (Table 1).

6 585 hn4 fin5 w-I h ~ 7T H ~ H6 hH 4 W-II fH2 n H7 mH6 H5 l " " l " ~ ' l " " l ~ " ~ ~ ' " ' ~ " ' ~ ~ ' ' ~ ' ~ ~ 8.0 7.0 6.0 5.0 B,(PPm) Fig.3. 300-MHz 'H-NMR spectra of porcine kringIe 4 in the presence of cAhx: aromatic region. (A) ID reference spectrum (resolution- enhanced). Ring proton spin systems belonging to the two His residues, labelled H-I1 and H-111, and the three Trp residues, labelled W-I. W-I1 and W-111, are indicated by reference to the scalar connectivities in the COSY contour plot (B). The low field signal marked with an asterisk (*) denotes an impurity singlet. (B) Contour plot of the 2D COSY spectrum with sine-bell filtering along both dimensions. Ring proton connectivities of the two His and three Trp residues are indicated. Cross-peaks due to H-I coupling are absent, reflecting deletion of His3 in the porcine homolog. COSY connectivities are labelled with two digits, following the convention in Fig. 2. Kringle concentration 1 mM, [ligand]/[protein] x 1, pH* 7.2, 310 K Tryptophanyl spectra [18]. The assignment of the W-I11 spin system (Fig.2) to Trp7' has been made earlier both by chemical modification of the The identification of the full spin system of W-I, W-I1 and residue [8, 171 and by spectral comparison with the human W-I11 in the 300-MHz spectra of human kringle 4 containing kringle 1 homolog [7, 171in which T Y ~substitutes '~ for Trp7' the ligand ~ A h xis shown in Fig. 2. Cross-peaks in the COSY [4]. For the latter, indole-group-specific proton identification spectrum show the couplings connecting the indole has been achieved via photo-CIDNP experiments at 360 MHz ring protons; their respective line positions, including the (De Marco A., Petros A., Kaptein R. & Llinas, M., coupling network, are marked above the ID reference unpublished results). spectrum (Fig. 2A). Except for the shifts induced by the addi- Overhauser experiments in l H 2 0 centered on the indole tion of the ligand, the COSY connectivities shown in Fig.2 NHI proton produced enhancements on the H2 (singlet) and concur with our previously published SECSY connectivities H7 (doublet) protons of both W-I and W-I1 of human of the Trp residues of human kringle 4 in the absence of ligand kringle 4 [18] (and Motta, A,, Laursen, R. A. & Llinas, M.,

7 Table 1. Chemical shifts and resonance assignments of the aromatic proton spin systems in the H - N M R spectra of human, porcine und bovine krinele 4 homolom Spectra were recorded in the presence of saturating amounts of the ligand, 6-aminohexanoic acid, at pH * 7.2, 310 K. The chemical shifts (ppm) and resonance assignments are based on 2D NMR experiments both at 300 and 600 MHz discussed in this paper Residue type Spin system Species 6 for Assignmcnt H2 H4 Histidine H-I human 7.683 7.046 His3 porcine - - bovine 7.733 7.079 H-I1 human 7.543 8.321 His3' porcine 7.556 8.367 bovine 7.563 8.365 H-Ill human 8.291 6.863 His3' porcine 8.248 6.948 bovine 8.249 6.880 d for H3,5" H2,6" Tyrosine Y-I human 6.547 6.806 Tyr ' porcine 6.560 6.794 bovine 6.545 6.772 Y-I1 human 6.870 6.970 Tyr5' porcine 6.881 6.955 bovine 6.870 6.963 Y-111 human 7.053 7.205 Tyr"' porcine - - bovine 7.009 7.219 Y-I11 human - - Tyr64 porcine 6.828 6.868 bovine 6.804 6.924 Y -1v human 6.854 7.319 TY~'~ porcine 6.828 7.334 bovine - - Y-Vb human 6.620 7.199 7.290 7.502 Tyr' porcine 6.566 7.214 7.266 7.491 bovine - - - - 6 for H2 H4 H5 H6 H7 Tryptophan W-I human 7.173 8.282 6.501 4.973 7.465 .i'rp25 (7) porcine 7.164 8.273 6.515 5.098 7.480 bovine 7.171 8.280 6.517 5.028 7.497 w-I1 human 7.809 7.053 4.963 6.628 6.806 Trp6' (?) porcine 7.722 7.012 5.152 6.913 7.203 bovine 7.788 6.658 4.742 6.613 7.144 w-I11 human 6.669 6.637 5.034 6.754 7.164 Trp7' porcine 6.654 6.684 5.205 6.726 7.160 bovine 6.647 7.199 5.205 6.897 7.497 H2,6 H3,5 H4 Phenylalanine human 6.902 7.046 7.327 Phe64 porcine 7.389 7.593 7.480 Phe4' bovine 7.497 7.450 7.650 Phe74 a Ortho (H2,6) and meta (H3,5) Tyr phenol ring identification is tentative, based on their relative chemical shift ordering. The chemical shift of the Y-V spin system could be measured only at 600 MHz. 293 K. At 600 MHz, 310 K, the resonances disappear. In the case of bovine kringle 4 the signals are too broad to be identifiable. The human kringle 4 contains no ligand while the porcine homolog represents a 1:1 complex with EAhx.

8 587 hH4 dlH5 w-I AH7 TH2 H6 AH4 W-II 1H2 hH 7 dlH6 H5 hH4 1H2 1 8, ( PPm) 9.o b y , ,, , , , , , , , , , , , , , , , , , , , , , , , , , , , , 8 .O 7.0 6.0 5.0 EJPPm) Fig.4. 3OO-MHz ' H - N M R spectra of bovine kringle 4 in the presence of eAhx. (A) 1D reference spectrum (resolution-enhanced). Spin systems of the ring protons belonging to the three Trp residue, labelled W-I, W-I1 and W-111, are indicated by reference to the scalar connectivities in the DQ contour plot (B). (B) Contour plot of the 2D DQ spectrum at t = 15ms (I/* J - ' ) with shifted sine-bell filtering along both dimensions. Scalar connectivities for the indole protons of the three Trp residues are indicated by horizontal bars linking cross-peaks which are equidistant from the diagonal. Double quantum frequencies appear along the B,-axis and are referred to the carrier frequency. Cross-peaks are labelled according to the COSY convention used in Fig.2. Notice, however, that only the second ('column') digit labels an observable frequency. Kringle concentration x 1 mM, [ligand]/[protein] x 4, pH* 7.2, 310 K unpublished observations). This result, in conjunction with ( A 6 z - 0.08 ppm for the H2 singlet). This we attribute to the COSY spectrum (Fig. 2), enabled the full identification of disaggregation of the protein which also results in a the specific ring proton spin systems of W-I and W-11. How- sharpening of the spectrum. ever, sequence-specificassignment of the aromatic resonances Similar comparison of human and bovine kringle 4 spectra of both W-I and W-I1 still remains to be conclusively estab- lead to the identification of the W-I, W-I1 and W-I11 spin lished. Comparisons among the homologs do not provide systems of the latter homolog. Again, despite some small insights as these two Trp residues are conserved. differences in chemical shifts, the overall connectivity patterns The characterization of the full spin systems of W-I, W-I1 of the three tryptophanyl residues in the two spectra are and W-111 in the spectra of human kringle4 leads to similar. Unfortunately, in the COSY spectrum of bovine identifying resonances from the conserved residues in the kringle 4, cross-peaks due to coupling between W-I1 H4 and spectra of porcine and bovine kringle 4 containing the EAhx H5 protons overlap with those between H6 and H5. This ligand. The COSY spectrum of porcine kringle 4 (Fig.3) overlap was partially removed in the DQ spectrum of the shows the cross-peaks arising from the coupling of the indole same sample shown in Fig.4. The small difference in the ring protons of W-I, W-I1 and W-111; their peak positions are chemical shift of the H4 and H6 protons of W-I1 are dis- marked in the 1D reference spectrum. Comparison of the cernible due to the separate cross-peaks observed for these human and porcine kringle 4 spectra indicates that except for two protons in the DQ spectrum which enabled the spin slight differences in the chemical shifts of the W-I1 and W-111 coupling route to be mapped. Similarly, cross-peaks resonances, the overall proton connectivity pattern of the connecting the H6 and H7 of W-111, could be readily discerned various indole rings appears identical. Another interesting in the DQ spectrum due to the absence of a diagonal (Fig. 4B). observation regarding porcine kringle 4 was the effect of dilu- tion on its NMR spectrum (not shown). A fivefold dilution Tyrosyl and phenylalanyl spectra of the sample, 1.O mM to 0.2 mM, resulted in a small upfield shift ( A 6 < - 0.02 ppm) of most of the aromatic resonances, The assignment of previously identified [18, 211 tyrosyl especially noticeable for W-I1 which experiences a larger shift resonances to specific residues in the sequence was made by

9 588 A n, IJ D Y-v y u C B Y-v v A Fig. 5 . 600-MHz 'H - N M R spectra of human plasminogen kringle 4 : Fig.6. 600-MHz ' H - N M R spectra of porcine kringle 4 in the presence aromatic region. (A) 1D reference spectrum (resolution-enhanced). of EAhx: aromatic region. (A) 1D reference spectrum (resolution- Ring proton spin systems belonging to the five Tyr residues, labelled enhanced). Ring proton spin systems belonging to the five Tyr residues Y-I, Y-11,Y-111,Y-IV and Y-V, and the lone phenylalanine residue, labelled Y-I, Y-11, Y-111' ( T Y ~ ~Y-IV ~ ) , and Y-V, and the lone Phe F64, are indicated by reference to the scalar connectivities in the residue, Phe41, are indicated by reference to the scalar connectivities COSY contour plot (B). (B) Contour plot of the 2D COSY spectrum in the COSY contour plot (B). (B) Contour plot of the 2D COSY with sine-bell filtering along both dimensions. Scalar proton spectrum with sine-bell filtering along both dimensions. Scalar connectivities for the five Tyr residues and Phe64 (shaded black) are connectivities of the aromatic protons for the five Tyr residues and indicated by solid lines. Connectivities between ring protons H6 and Phe41 (shaded black) are indicated by solid lines. The connectivity H7 of W-I1 and W-I11 (Trp'*) are indicated by broken lines. Kringle between ring protons H6 and H7 of W-111 (Trp7*)is indicated by a concentration 1 mM, pH * 7.2,293 K broken line. Note the deletion of cross-peaks due to Y-111 (Tyr4', see Fig. 5). Kringle concentration 1 mM, [ligand]/Lprotein] z 1, pH * 7.2, 293 K comparing the NMR spectra of the three kringle 4 homologs phenylalanyl residue, Phe64, (shaded black) and of the and human kringle 1 with each other. Fig. 5 shows 600-MHz tryptophanyl residues W-I1 and W-111. 'H-NMR spectra of human kringle 4. The ring proton re- Comparison of the 600-MHz COSY spectrum of human sonances of Y-I, Y-11, Y-I11 and Y-IV exhibit AA'BB' spin kringle 4 (Fig. 5) with that of porcine kringle 4 (Fig. 6) enabled pattern; cross-peaks due to scalar coupling between their re- the unambiguous assignment of the Y-111 spin system to spective meta (H3,5) and ortho (H2,6) protons are indicated Tyr4': cross-peaks due to coupling between meta (H3,5) and by solid line connectivities in the COSY spectrum (B). The ortho (H2,6) protons of Y-111 seen in the COSY spectrum of connectivity scheme shown by the 2D spectrum (B) enables the human kringle 4 (Fig. 5) are clearly absent in the COSY the pairing of the two-proton doublet spins of each residue as spectrum of the porcine variety (Fig.6) since a Phe residue shown in the 1D reference spectrum (A). In contrast to the substitutes for Tyr41in the porcine sequence. New cross-peaks spectra of these four Tyr rings, the spin system of Y-V exhibits (shaded black) due to coupling between the ring protons of and ABCD pattern at 600 MHz, 293 K; thus it generates two the substituent, Phe4', in the porcine spectrum are shown by sets of COSY cross-peaks (Fig.S), each one representing a solid line connectivities (Fig. 6). Similarly, another aromatic pair of one-proton doublets. Fig. 5 also shows connectivities substitution, that of Phe64for a tyrosine in the same porcine due to coupling between aromatic protons of the lone sequence, is manifested by the absence of cross-peaks due to

10 589 Y -1v I -6.0 -6.5 . G,(PPm) -7.0 I I I I I I I I 7.4 7.0 6.6 Fig. 7.300-MHz H-NMR spectra of human kringle 4 in thepresence of eAhx: aromatic region. (A) 1D reference spectrum (resolution-enhanced). Ring proton spin systems belonging to the four Tyr residues, labelled Y-I, Y-11, Y-I11 and Y-IV, and the lone phenylalanine residue, F64, are indicated by reference to the scalar connectivities in the DQ contour plot (B). (B) Contour plot of the 2D DQ spectrum at z = 15 ms (y8J - ) with sine-bell filtering along both dimensions. The diagonal is indicated (- . -). Double quantum frequencies appear along the 6, axis and are referred to the carrier frequency. Scalar connectivities between ring protons of the Tyr residues and Phe64(shaded black) are indicated by horizontal bars connecting cross-peaks equidistant from the diagonal. Connectivities between the ring protons H6 and H7 of W-I1 and W-111 TI^'^) are indicated by broken horizontal lines. Kringle concentration x 1 mM, [ligand]/[protein] x 2, pH * 7.2, 310 K the Phe64 ring protons and the appearance of new from the line widths and chemical shifts of the Y-111 re- connectivities due to the Tyr64 aromatic protons which lie sonances in the porcine and bovine homologs that the side close to the diagonal in the porcine spectrum (Y-III, barely chain of TyP4 is rather constrained by the structure and visible in Fig.6). Support for these assignments has been probably not significantly exposed. Consistently, the Phe64 derived by comparing the 300-MHz DQ spectrum of human phenyl group in human kringle 4 (discussed below) also dis- and porcine kringle 4 varieties (Figs 7 and 8). The pair of plays broad multiplets of non-random characteristics doublet peaks about the DQ diagonal due to Y-I11 seen in the (Fig. 5). This may bear functional relevance for the kringle 4 human spectrum (Fig. 7) is found to be absent in the porcine lysine-binding site structure as it is known that Phe64 spectrum (Fig. 8) and thereby confirms the assignment of establishes close contact with the ligand [20, 221. Y-I11 to Tyr4. In an analogous way, peaksdue to Phe64 are In the 600-MHz COSY spectrum of bovine kringle4 found deleted and replaced by peaks of Y-111 ( T Y ~in~the ~) (Fig. 9), cross-peaks due to Y-IV are found to be missing and porcine spectrum (Fig. 8). This analysis, including detection hence lead to their assignment to Tyr74after comparison with of Y-111 cross-peaks, was substantiated by a SECSY exper- the COSY spectrum of human kringle 4 (Fig. 5). This deletion iment at 300 MHz (not shown). arises as a consequence of the substitution of Phe74 for Tyr74 The identification of a close-to-diagonal position for the in the bovine sequence (Fig. 1). New cross-peaks (shaded Y-111 resonances in the porcine kringle 4 spectrum (Figs 6 black) due to the substituent, Phe74,can be found in the same and 8) is confirmed by the pattern shown by the Tyr64 aro- figure (Fig. 9). This assignment was verified by comparing the matic signals in the 600-MHz spectrum of the bovine homo- 300-MHz DQ spectrum of bovine kringle 4 (Fig. 10) with the log, which happens to display slightly better-resolved (less de- spectrum of the human homolog (Fig.7). Peaks due to cou- generate) H2,6/H3,5 doublets for Y-111 (Fig. 9). It is suggested pling between the ring protons of Y-IV seen in the latter were

11 590 F Y-IV I I I rh rtl h h V-lT h v-1 I II I It lB Y-I -6.0 81 (PPm) Y-IV -7.0 Fig.8. 300-MHz H - N M R spectra of porcine kringle 4 in the presence of EAhx: aromatic region. (A) 1D reference spectrum (resolution- enhanced). Ring proton spin systems belonging to the four Tyr residues, labelled Y-I, Y-11, Y-111 ( T Y ~and ~ ~Y-IV, ) and the lone Phe residue, F4. are indicated by rcference to the scalar connectivities in the DQ contour plot (B). (B) Contour plot of the 2D DQ spectrum at 7: = 15 ms (!4~J - )with shifted sine-bell filtering along both dimensions. Scalar connectivities between ring protons of the Tyr residues and Phe41 (shaded black) are indicated by horizontal bars and the connectivities between the ring protons H6 and H7 of W-I1 and W-111 ( T ~ P are ~ ) indicated by broken horizontal lines. Kringle concentration 1 mM,[ligand]/[protein] z 1, pH* 7.2, 310 K found absent in the former and this supports the assignment sequence) (not shown). Consistency in the results of the above of Y-IV to Tyr74. The new peaks due to Phe74 ring protons Overhauser experiments support the unambiguous assign- in the bovine DQ spectrum are shaded black (Fig. 10). ment of Y-IV to Tyr74 while it verifies a preserved steric The assignment of the Y-IV spin system to T Y ~ was ~ relationship between the Trp-I1 side chain and the site 74 tested for consistency with a previously observed NOE in aromatic ring. human kringle 4. We have shown that the 600-MHz irradia- Assignment of the Y-I spin system was established by tion of the W-I1 indole H2 singlet transition produced NOEs comparing spectra of human plasminogen kringle 4 and on the Y-IV aromatic doublets of human kringle 4 [20]. kringle 1. The pair of doublets due to Y-I seen in the 1D Similar Overhauser experiments at 600 MHz were performed reference spectrum of kringle 4 (Fig. 5A) was found deleted on porcine and bovine kringle 4 homologs to observe the in the kringle 1 spectrum and accordingly 2D cross-peaks effects due to irradiation of the W-I1 H2 singlet. Irradiation arising from the ring protons of Y-I are not seen in its COSY of the same W-I1 singlet in the spectrum of porcine kringle 4 spectrum either (Motta, A., Laursen, R. A. & Llinas, M., (Fig. 11A) produced NOEs on both Y-IV and F4 ring pro- unpublished results). (A doublet at w 6.6 ppm in the spectrum tons as shown (Fig. 11 B). The latter may result from rf power of kringle 1, previously indentified as Y-I1 in kringle 1, re- spill over on the Phe4 H3,5 transition at 7.59 ppm. A control sonating at similar chemical shift as Y-I in kringle 4 [17], does experiment was implemented to remove the ambiguity that in fact arise from Phe3; Motta, A., Laursen, R. A. & Llinb, the NOE on Y-IV is a consequence of rf power spill over on M., unpublished observations.) As human kringle 1 lacks Tyr2 the adjacent Phe41 resonance and not due to the W-I1 singlet [4] the Y-I spin system (Fig. 5 ) can be assigned to this residue. irradiation. Thus, we irradiated the H3,5 transitions of Phe4 Tyr, which is conserved in all the three kringle 4 homologs, in the porcine spectrum (Fig. 11 C) and no NOE was observed produces cross-peaks in the 2D spectra with similar chemical on Y-IV. Similarly, irradiation of the same W-I1 singlet of the shifts (Figs 5, 6 and 9; Table 1). This assignment is in agree- bovine kringle 4 spectrum produced NOEs of the ring proton ment with the study of Trexler et d . [S], who have shown by multiplets of Phe74(substituting for Tyr74,Y-IV, in the bovine Cu2 titration the selective paramagnetic broadening of Y-1 +

12 591 A Y-IU V-T r 74 Y-IU h I l l I I I I , A r u I I I It W-II 16.5 B Q---------- ! F 74 I I I I I I I ( I I I 1 T 7:5 7.0 6:5 %( PPm) Fig.9. 600-MHZ H - N M R spectra of bovine kringle 4 in the presence of cAhx: aromatic region. (A) 1D reference spectrum (resolution- enhanced). Ring proton spin systems belonging to the four Tyr residues, labelled Y-I, Y-11,Y-I11 and Y-111 ( T Y ~ and ~ ~ )the , lone Phe residue, F74,are indicated by reference to the scalar connectivities in the COSY contour plot (B). (B) Contour plot of the 2D COSY spectrum with sine-bell filtering along both dimensions. Scalar connectivities between ring protons of the four Tyr residues and Phe74 (shaded black) are indicated by solid lines. A connectivity between ring protons H6 and H7 of W-I1 is indicated with a broken line. Note the absence of cross- ~ , 5). Kringle concentration 1 mM, [ligand]/[protein] = 4, pH * 7.2.293 K peaks due to Y-IV ( T Y ~see~ Fig. (Y-A in the nomenclature of Trexler et al. [S]) doublets, proton doublets in the respective 1D reference spectra. How- suggesting proximity of the phenol ring to the N-terminus ever, even at 600 MHz resonances due to Y-V broaden and metal-binding site and hence tentatively assigned the Y-I spin disappear when the temperature is raised to 310 K (Fig. 12). system to Tyr. On the other hand, nitration of the TyrSo In the case of bovine kringle 4 the Y-V resonances are not residue of human kringle 4 [S, 211 resulted in the loss of the identifiable at 600 MHz, neither at 293 nor 310 K, as they are ring proton doublets of Y-I1 in the spectrum (Y-B in the very broad (Figs 9 and 12C). Recently, Esnouf et al. [37] have nomenclature of Trexler et al. [8]). Since TyrSo is conserved reported 500-MHz COSY spectra of human kringle 4 (pH 7.7) in the three kringle 4 homologs considered in this study, their showing the Tyr-V (Tyr-E in the authors nomenclature) aro- spectra consistently exhibit cross-peaks (Figs 5, 6 and 9) due matic resonances as an AABB spin system of 6 z 6.85 to coupling between ring protons with similar chemical shifts and 7.20 ppm in the presence of ligand at 310 K. At this (Table 1) lending support to the assignment of Y-I1 to TyrS0. temperature, we have confirmed such a pattern at 300 MHz, It was mentioned above that the ring protons of the Tyr in the presence of L-LYSor EAhx, as ligand presence favors residues Y-I, Y-11, Y-111 and Y-IV exhibit AABB spin the observation of intrinsically broad signals by narrowing patterns while Y-V produces an ABCD spectrum in the the spectrum. Thus, the appearance of the Tyr-V aromatic human kringle 4 spectrum. At 600 MHz the four Y-V one- signals as either ABCD or AABB spin system, would depend proton doublets were observed with b = 6.62ppm (A), on both temperature and the observational frequency. This 7.20 ppm (B), 7.29 ppm (C) and 7.50 ppm (D) (Fig.5). At indicates that at ambient temperature the Tyr-V side chain 300 MHz these four resonances are too broad to be identified finds itself in a regime somewhere in between slow (+ ABCD in the absence of ligand. The 600-MHz COSY spectra of both pattern) and fast (+ AABB pattern) exchange, i.e. of inter- human and porcine kringle 4 (Figs 5 and 6 ) produce cross- mediate ring flip dynamics in the NMR time scale. In view of peaks arising from Y-V and enabled the pairing of the one- the assignment of the spin systems of Y-I, Y-11, Y-111 and

13 592 F B -6.0 -7.0 - 8, (PPm) w-I F74 /- / I I I I I I 7.5 7.0 6.5 a,(PPm) Fig. 10. 300-MHz H-NMR spectra of bovine kringle 4 in the presence of EAhx: aromatic region. (A) 1D reference spectrum (resolution- enhanced). Ring proton spin systems belonging to the four Tyr residues, labelled Y-I, Y-11, Y-111 and Y-111 ( T Y ~ ~ and~ )the , lone Phe residue, F74,are indicated by reference to the scalar connectivities in the DQ contour plot (B). (B) Contour plot of the 2D DQ spectrum at T = 15 ms (Y8J) with shifted sine-bell filtering along both dimensions. The diagonal is indicated (- . -). Scalar connectivities between ring protons of the four Tyr residues and Phe74 (shaded black) are indicated by horizontal bars; the correlations between the indole protons H6 and H7 of W-I1 and W-111 (Trp) are marked by broken horizontal lines. This figure is an expansion of Fig.4 Y-IV to residues Tyr, Tyr50, Tyr4 and Tyr74, respectively, CONCLUSIONS the above four one-proton doublets of Y-V are automatically assigned to the only remaining tyrosyl residue, Tyr9. The coupled ring proton spin systems of the aromatic The identification of the ring proton multiplets of the residues in the human kringle 4 H-NMR spectrum has been unique phenylalanine residue, Phe64, in the spectrum of totally unravelled via a combined use of 2D NMR methods human kringle 4 was established from the COSY experiment such as COSY, SECSY and DQ experiments. Subsequent at 600 MHz (Fig. 5). Cross-peaks due to the ring protons of spectral comparison of kringle 4 homologs enabled the unam- this residue can be identified (shaded black) and their line biguous assignment of His3, Tyr41, Tyr74, Tyr9 and Phe64 positions determined. The connectivities of the Phe64 ring resonances. The spin systems of W-I and W-11 are the only proton multiplets are shown in the 1D reference spectrum remaining ones to be firmly assigned to either Trp or Trp6 (Fig. 5 A). Although the phenylalanyl multiplets are still in the sequence. In view of the large perturbations induced on clustered with other nearby resonances, the degeneracy is W-I1 upon ligand-binding and the results of ligand-kringle partially removed when the spectrum of kringle 4 is measured saturation transfer (Overhauser) experiments, the latter spin in the presence of a ligand. This assignment was checked system most probably arises from Trp6 [20]. Thus, by ex- by recording a 300-MHz DQ spectrum of human kringle 4 clusion, we tentatively assign W-I to Trp. containing the ligand EAhx (Fig.7). Due to the absence of Overhauser experiments centered on the ligand-bound diagonal peaks, peaks due to the coupling between ring pro- protein complex reveal efficient cross-relaxation between the tons of Phe64 are more clearly identified in this spectrum, indole rings of W-I1 and W-I11 and the ligand [20]. Several including their multiplicity pattern. This analysis reinforces other Overhauser experiments have implicated residues within our earlier assignment made at 300 MHz [18]. the C y ~ ~ - C loop y s ~ (Fig. ~ l), rich in aromatic side chains, as

14 593 NOE I I I I I 9.0 8.0 7.0 6.0 5.0 6 (ppm) Fig. 11. Porcine kringle 4 in the presence of EAhx: proton Overhauser experiments at 600 MHz. (A) Reference spectrum (resolution-enhanced). (B) NOE difference spectrum after irradiation of the W-11 H2 singlet transition, indicated by wavy arrow b in spectrum A. The perturbed ring proton resonances of Phe4 (F41) and T Y ~(Y-IV) ~ are indicated. (C) NOE difference spectrum after irradiation of the Phe41 H3,5 triplet transitions, indicated by wavy arrow c in spectrum A. The perturbed adjacent ring proton resonances of the same Phe ring, an H2.6 doublet and a (broad) H4 triplet, as well as the H2 singlet of W-11,are indicated. Kringle concentration 1 m M ,[ligand]/[protein] x 2, pH * 7.2, 310 K being involved in ligand binding [20]. We have already noted tertiary folding places the CysZ2- Cys6 bridge into close van the cross-relaxation of the site 74 ring protons following the der Waals contact with the C y ~ ~ - C y link, s ~ ~ the two S-S irradiation of the W-I1 indole H2 singlet in the spectra of both bonds becoming adjacent and roughly orthogonal to each porcine (Fig. 11) and bovine kringle 4 homologs. These are in other. The model is also in harmony with our earlier studies agreement with Overhauser experiments indicating that W-I1 showing that, in the process of aerobic restoration of reduced (Trp62?) and W-I11 (Trp7) are close together in the folded kringle 4, the Cys - C Y S ~and ~ Cys -Cysb3 disulfide structure [I91while both participate in ligand-binding [20,22]. bonds are the first to form [39] and thus may potentially serve The interaction of the site 74 residue with W-I1 (Trp62?) to nucleate the reversible kringle folding. implies a folding of the kringle inner loop, which extends from The identification and assignment, including the coupling Cys5 to Cys7, such that the T r ~ ~ ~ - C y s ~ -stretch P h e ~ is ~ network, of the various aromatic spin systems of human, brought into the proximity of the CysS1- bridge. This porcine and bovine kringle 4 containing the ligand EAhx as folding would place both Trp62and Phe64 neighboring Trp, discussed in this paper are summarized in the 600-MHz ID thus contributing to establish a hydrophobic environment at NMR spectra of the three homologs (Fig.12) and their the lysine-binding site: the Trp-I1 (Trp62?), Phe64 and Trp72 chemical shifts are listed in Table 1. The spin system of Y-V rings are known to participate directly in ligand-binding [20, (Tyr) is not observable at 37C and thus not indicated above 221. Such a structural feature, clearly of utmost relevance for the spectra (Fig. 12). establishing the plasminogen kringles function, would be in The near conservancy in the chemical shifts of the aromatic total accord with a recent crystallographic model of kringle residues in the spectra of kringle 4 homologs (Table 1) folding based on the structure at 0.28-nm resolution of the suggests that their structures in solution are similar. Sizable prothrombin fragment 1 [38]. Thus, although the prothrom- variations in chemical shifts are noted mainly for the spin bin kringle is not known to carry a lysine-binding site, its systems of W-I1 and W-I11 which suggests that they are more

15 594 w-I A h I m m w-II I h r n h m w72 A m lh m h Y" ~ 3I 3 Y50' H3' Y4' h r L l Y2 H' FS46-37 I 1 w-I h h I m m w-II I h l + h m w72 A mhl m mY74 F4'rLIJ Y"f;;i h Y Z H3' I 1 fhY64 w-I h H3' - w-I H F" r rn h Wl2 h 3 h I Y4'RK h m r 1 h l m Y2 dl m m I I I I I 9.0 8.0 7.0 6.0 5.0 MPPm) Fig. 12. Comparison of 600-MHz ' H - N M R spectra of kringle 4 homologs in the presence of EAhx. Spectra A, B and C arise from fragments derived from human, porcine and bovine plasminogens, respectively. Assigned aromatic spin systems are indicated for each homolog. Ring proton resonances of Y-V (Tyr') were broadened at 310 K but can be seen a t 293 K in the human and porcine kringle 4 spectra (Figs 5 and 6). An asterisk in spectrum C, denotes an impurity singlet. Kringle concentration 1 mM, pH* 7.2, 310 K

16 595 sensitive to minor changes in their environments. An analysis 17. Llinis, M., De Marco, A., Hochschwender, S. M. & Laursen, R. of the aliphatic region of the spectra of kringle 4 homologs is A. (1983) Eur. J . Biochem. 135, 379-391. in progress and will be reported in a later publication. 18. De Marco, A., Pluck, N. D., Banyai, L., Trexler, M., Laursen, R. A., Patthy, L., Llinas, M. &Williams, R. J. P. (1985) Bio- chemistry 24,748 - 753. We are grateful to Dr A. Motta for his enthusiastic assistance 19. De Marco, A., Laursen, R. A. & Llinis, M. (1985) Biochim. while implementing the DQ experiments. This work was supported Biophys. Actu 827, 369 - 380. by the U.S. Department of Health and Human Services, National Institutes of Health grant HL-29409. The 600-MHz NMR facility is 20. Llinas, M., Motta, A,, De Marco, A. & Laursen, R. A. (1985) J . supported by Nationdl Institutes of Health grant RR-00292. Biosci. Suppl. 8, 121 - 139. 21. Trexler, M., Banyai, L., Patthy, L., Pluck, N. D. & Williams, R. J. P. (1985) Eur. J . Biochem. 152,439-446. 22. De Marco, A., Laursen, R. A. & Llinas, M. (1986) Arch. Biochem. Biophys. 244,727 - 741. REFERENCES 23. Deutsch, D. G . & Mertz, E. T. (1970) Science (Wash. D C ) 170, 1. Hull, W. E. (1982) Two dimensional N M R , Bruker, Karlsruhe. 1095- 1096. 2. Wider, G., Macura, S., Kumar, A., Ernst, R. & Wuthrich, K. 24. Glasoe, P. K. & Long, F. A. (1960) J . Phys. Chem. 64,188- 190. (1984) J . Mugn. Reson. 56, 207-234. 25. Ferrige, A. G. & Lindon, J. C. (1978) J . Mugn. Reson. 31, 337- 3. Bax, A. (1985) Bull. Mugn. Reson. 7, 167-183. 340. 4. Sottrup-Jensen, L., Claeys, H., Zdjdel, M., Peterson, T. E. & 26. De Marco, A. (1977) J . Mugn. Reson. 26, 527-528. Magnusson, S. (1978) in Chemicul fibrinolysis and thrombosis 27. Nagayama, K., Wiithrich, K. & Ernst, R. (1979) Biochem. Bio- (Davidson, J. F. et al., eds) ~01.3,pp. 191-209, Raven Press, phys. Res. Commun. 90,305 -311. New York. 28. Nagayama, K., Kumar, A,, Wuthrich, K. & Ernst, R. (1980) J . 5. Brunisholz, R. A. & Rickli, E. E. (1981) Eur. J. Biochem. 119, Mugn. Reson. 40, 321 - 334. 15-22. 29. Mareci, T. H. & Freeman, R. (1983) J . Mugn. Reson. 51, 531 - 6. Schaller, J., Moser, P. W., Dannegger-Miiller, G. A. K., Rosselet, 535. S. J., Kampfer, U. & Rickli, E. E. (1985) Eur. J . Biochem. 149, 30. Braunschweiler, L., Bodenhausen, G. & Ernst, R. (1983) Mol. 267-278. Phy.7. 48, 535 - 560. 7. Hochschwender, S. M., Laursen, R. A., De Marco, A. & Llinas, 31. Bax, A., Freeman, R., Frenkiel, T. A. & Levitt, M. H. (1981) J . M. (1983) Arch. Biochem. Biophys. 223, 58-67. Magn. Reson. 43,478-483. 8. Trexler, M., Bhnyai, L., Patthy, L., Pluck, N. D. &Williams, R. 32. Wagner, G. & Zuiderweg, E. R. P. (1983) Biochem. Biophys. Res. J. P. (1983) FEBS Lett. 154, 311 -318. Commun. 113, 854-860. 9. Trexler, M. & Patthy, L. (1983) Proc. Nutl Acad. Sci. USA 80, 33. Boyd, J., Dobson, C. M. & Redfield, C. (1983) J. Mugn. Reson. 2457 - 2461. 55, 170-176. 10. Castellino, F. J., Ploplis, V. A., Powell, J. R. & Strickland, D. K. 34. Gordon, S. L. & Wiithrich, K. (1978) J . Am. Chem. Soc. 100, (1981) J . Biol. Chem. 256,4778-4782. 7094 - 7098. 11. Novokhatny, V. V., Kudinov, S. A. & Privalov, P. L. (1984) J. 35. Szeverenyi, N. M., Bothner-By, A. A. & Bittner, R. (1980) 1. Mol. Biol. 179,215- 232. Phys. Chem. 84,2880-2883. 12. De Marco, A., Motta, A., Llinas, M. & Laursen, R. A. (1985) 36. Kink G. &Wright, P. E. (1982) Biochem. Biophys. Res. Commun. Biophys. J. 48,411 -422. 106, 559-565. 13. Lerch, P. G., Rickli, E. E., Lergier, W. & Gillessen, D. (1980) 37. Esnouf, P., Lawrence, M. P., Mabbutt, B. C., Patthy, L., Pluck, Eur. J . Biochem. 107, 7 - 13. N. D. & Williams, R. J. P. (1985) Bull. Soc. Chim. Belg. 94, 14. Thorsen, S., Clemmenscn, I., Sottrup-Jensen, L. & Magnusson, 883 - 896. S. (1981) Biochim. Biophys. Acta 668, 377-387. 38. Park, C . H. & Tulinsky, A. (1986) Biochemistry 25, 3977-3982. 15. Winn, E. S., Hu, S. P., Hochschwender, S. M. & Laursen, R. A. 39. Trexler, M. & Patthy, L. (1983) Proc. Nail Acad. Sci. USA 80, (1980) Eur. J . Biochem. 104, 579-586. 2457 - 2461. 16. Vali, Z. & Patthy, L. (1982) J. Biol. Chem. 257, 7401 -7406. View publication stats

Load More