Chemical shift

Figure 9.29

only has one coupling: to H-g2. If we make a horizontal slice of this crosspeak we get an antiphase doublet with J = 3.8 Hz. The symmetry-related crosspeak (Fig. 9.28, lower right inset) represents transfer of magnetization from H-g1 (F1) to H-g2 (F2). The H-g2 resonance in the 1D spectrum is a double doublet (J = 10.0, 3.8) because it is coupled to H-g1 (3.8 Hz, axial-equatorial) and to H-g3 (10.0 Hz, axial-axial). The horizontal (F2) slice through this crosspeak shows the double-doublet structure, but it is antiphase (+,—) with respect to the 3.8 Hz coupling and in-phase (+,+ or —,—) with respect to the 10.0

Hz coupling. The antiphase H-g1 coherence,2l|.I2, undergoes coherence transfer to give

2 1 y z the antiphase H-g2 coherence, 2Iy I^, but because H-g2 is also coupled to H-g3 we see this coupling as well. Any coupling that does not appear as a multiplier (x Iz) in the operator product must be in-phase. This is recorded in the FID and after the F2 Fourier transform we see that any pair of lines separated by the "passive" coupling J23 will have the same sign (+,+ or —,—) and any pair of lines separated by the "active" coupling J12 will have the opposite sign (+,— or —,+). The "active" coupling is simply the coupling that gave rise to the crosspeak. If the crosspeak is located at the intersection of the H-g1 chemical shift and the H-g2 chemical shift, the active coupling in its fine structure will be J12, the coupling that led to the H-g1/H-g2 antiphase state and allowed coherence transfer to occur. The passive coupling (J23) does not appear at all in the F1 dimension of this crosspeak because the F1 resonance is H-g1, which is only coupled to H-g2.

Figure 9.30 shows the upfield region of the 300 MHz DQF-COSY spectrum of sucrose. Starting at the upper left, we follow horizontally from the H-g2 (F1)/H-g1 (F2) crosspeak (from Fig. 9.28, upper left) to the right along the dotted line, passing through a crosspeak and on to the H-g2 diagonal peak. From here we go straight down to a crosspeak and then to the left back to the diagonal: this is the H-g3 diagonal peak. From here we reverse direction, moving horizontally to the right and passing the H-g2/H-g3 crosspeak to stop at another crosspeak, enclosed in a rectangular box. From here we move straight up, returning to the diagonal at the most upfield peak of the spectrum. We can assign this peak to H-g4 (a triplet

Proton (ppm) Figure 9.30

in the 1D spectrum). From here we move to the left, passing the H-g3/H-g4 crosspeak to a faint crosspeak just beyond it (rectangular box labeled "g5"). Moving up to the 1H spectrum at the top of the COSY we can pinpoint the chemical shift of H-g5 in the overlapped region. This completes our walk through the glucose spin system. H-g6 (two protons) is too close in chemical shift to H-g5 to provide any useful crosspeak, but we can guess that its resonance is in the tall peak just upfield of H-g5.

Starting again with the H-f3 diagonal peak (Fig. 9.30, lower left), we move up and then right to the H-f4 diagonal peak, and then up again and right to the H-f5 diagonal peak ("X" shape). If we strain our eyes a bit, we can see a blob of intensity above the H-f5 diagonal peak (rectangular box) that leads us to the right side to the H-f6 diagonal peak. This corresponds to the same tall, overlapped peak in the 1D spectrum that we assigned to H-g6. This confirms the assignments we made in Chapter 8 based on selective 1D TOCSY experiments, and it also confirms our 13C assignments made through the HETCOR correlations. Sucrose is assigned!

In the 300 MHz DQF-COSY spectrum of sucrose the crosspeaks appear very large because the chemical-shift range is small, particularly for the nonanomeric protons. One way to "shrink" the size of the crosspeaks is to move to a higher-field instrument. We saw in Chapter 2 how the "footprint" of a 1H resonance (multiplet) gets smaller on the ppm scale as we increase the field strength. Figure 9.31 shows the upfield region of the 600 MHz DQF-COSY spectrum of sucrose. Comparing to Figure 9.30, all of the splittings have been cut in half and the crosspeaks and diagonal peaks are one half the size in both dimensions.

I start

Figure 9.31

I start

Figure 9.31

This is because 1 ppm now corresponds to 600 Hz rather than 300 Hz and the coupling constants in hertz have not changed. It is now much easier to follow the spin system from H-g2 to H-g5 and from H-f3 to H-f6 (Fig. 9.31, arrows). At 600 MHz the H-f5 resonance is completely resolved (a ddd coupled to H-f4 and the two H-f6 protons) and the H-f5 to H-f6 crosspeak is clearly visible.

In Chapter 5 we looked at the presaturation 1H spectrum of a cyclic peptide in 90% H2O/10% D2O (Fig. 5.20). Figure 9.32 shows a portion of the DQF-COSY of the same cyclic peptide in D2O, also using presaturation of the HOD resonance. The Ha resonances are seen on the diagonal in the region 3.9-5.4 ppm, with connections to the Hp protons shown by dotted lines. One Ha resonance in particular is labeled with crosspeaks to two p protons. This type of spin system is typical of a large number of amino acids and is called "three-spin" or "AMX": ND-CH—CHp -Y, where Y is either a heteroatom or a quaternary carbon (aromatic ring or carbonyl). Note that the amide NH is exchanged with deuterium in D2O so it is no longer part of the spin system. Another AMX spin system is found at Ha = 4.77 ppm, slightly downfield of the HOD streak. Notice that the Ha peak is missing on the diagonal for this spin system, and there are no symmetry-related Ha ^ Hp crosspeaks below the diagonal. In this case, the presaturation of HOD "wiped out" (saturated) the Ha proton so it could not transfer magnetization to Hp or Hp/. This kills the Ha diagonal peak (failed transfer from Ha) and the F2 = Hp, Hp/ crosspeaks (transfer from Ha). But the F1 = Hp, Hp/ crosspeaks are fine because they are far from HOD and are not affected by the presaturation. They transfer magnetization to Ha (F2 = Ha), which is observed in the FID.

Figure 9.32

It might be possible to measure the side-chain dihedral angle xi (chi-1), defined by the N-Ca-Cp-Y angle, if we can measure the Ha-Hp /-coupling constants. In the AMX spin system, there are three protons and each appears as a double doublet: Ha is split by Hp and Hp/ ;Hp is split by Hp/ and Ha; and Hp/ is split by Hp and Ha. Assuming that Hp and Hp/ have significantly different chemical shifts, we can diagram the expected crosspeak fine structure in the COSY spectrum (Fig. 9.33, left). For example, in F2 we have the Ha resonance with a double doublet defined by Jap and Jap/. The Ha-Hp crosspeak (bottom) will have the Jap couplings antiphase (active coupling) and the Jap/ couplings in-phase (passive coupling). The Ha-Hp/ crosspeak (top) will have the Jap/ couplings antiphase (active coupling) and the Jap couplings in-phase (passive coupling). In all, each of the two crosspeaks will have 16 peaks in the fine structure (double doublet in F1 and dd in F2). We could make a horizontal (F2) slice through one of the rows and try to extract the active and passive couplings, but usually this is difficult due to crowding and the complexity of the pattern. A modified COSY experiment, called COSY-35, greatly reduces the intensity of every other peak in the fine structure, leaving only eight peaks in each crosspeak (Fig. 9.33, right). This is accomplished by simply reducing the pulse width of the final pulse in the COSY sequence from 90o to 35° (Fig. 9.34), reducing 8 of the peaks to an intensity of 10% ((1 - cos ©)/(1 + cos ©)) of the other 8. Now an F2 slice gives a simple antiphase doublet pattern representing the active coupling, or a direct measurement in F2 can be made between peaks of the same sign (+ to +, or — to —) to measure the passive coupling (Fig. 9.33, right). Figure 9.35 shows an expansion of the F1 = Hp, Hp//F2 = Ha crosspeak highlighted with a rectangle in Figure 9.32. The left side is the DQF-COSY spectrum, with 16 peaks in each crosspeak,

and the right side is the COSY-35 spectrum, with only eight peaks in each crosspeak. The contour threshold is set high enough to completely reject the other eight peaks that are much lower in intensity. An F2 slice of the upper (H^/) crosspeak yields the active coupling Jap/, which can be accurately measured by simulation and curve fitting. An F2 slice of the lower (H) crosspeak shows an antiphase doublet with the active coupling Jap. Alternatively, direct measurement from peaks of like sign gives us the passive coupling Jap from the upper (H^/) crosspeak and the passive coupling Jap from the lower (H^) crosspeak.

Figure 9.36 shows Newman projections of the three low-energy conformers for the Ca-C^ bond of an "AMX" amino acid residue within a peptide or protein. The gauche relationship should give a small coupling (<6 Hz) and the anti relationship should give a

0 0

Post a comment