Carbohydrate Research 389 (2014) 165–173

Contents lists available at ScienceDirect

Carbohydrate Research journal homepage: www.elsevier.com/locate/carres

Sialo-CEST: chemical exchange saturation transfer NMR of oligo- and poly-sialic acids and the assignment of their hydroxyl groups using selective- and HSQC-TOCSY Hadassah Shinar a, Marcos D. Battistel b, Michael Mandler b, Flora Lichaa b, Darón I. Freedberg b, Gil Navon a,⇑ a

School of Chemistry, Tel Aviv University, Levanon Street, Ramat Aviv, Tel Aviv 69978, Israel Laboratory of Bacterial Polysaccharides, Center for Biologics Evaluation and Research, Food and Drug Administration, 1401 Rockville Pike, HFM-428, Rockville, MD 20852, United States b

a r t i c l e

i n f o

Article history: Received 19 September 2013 Received in revised form 3 February 2014 Accepted 6 February 2014 Available online 18 February 2014 Keywords: Sialic acid Polysialic acid Chemical exchange saturation transfer (CEST) NMR Exchange rate constants Hydroxyl groups assignments

a b s t r a c t Chemical exchange saturation transfer (CEST) is an NMR method that takes advantage of proton exchange between solute and solvent molecules in dynamic equilibrium, enabling the detection of the solute NMR signals with enhanced sensitivity. Herein, we report that the hydroxyl groups in a naturally occurring polysaccharide, a-2,8 polysialic acid in aqueous solution, yield very significant CEST effects even at 37 °C where the resonances of the hydroxyl groups are not directly observed. We also report the assignments of the hydroxyl groups for the polymer and its oligomeric building blocks, from monomer to hexamer. We show that the same assignments can be made by either 1H–1H TOCSY methods or 1 H–13C HSQC-TOCSY methods, to alleviate spectral overlap. Finally, we report the exchange rates of the OH groups with water and show how these rates can be used to select and fine-tune CEST effects. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Chemical exchange saturation transfer (CEST) is an evolving method currently being used to obtain contrast in MRI. This technique enables the detection of endogenous as well as exogenous agents containing labile protons such as hydroxyl, amine, or amide protons with enhanced sensitivity. In this technique, a selective RF pulse saturates the exchangeable protons and the reduction of the water signal, caused by magnetization transfer, is monitored. Enhancement factors between 102 and 106 relative to the concentration of the molecules of interest have been reported for certain systems (for reviews see Refs. 1–3). The CEST approach has been used to image tissue pH,4 map brain proteins through their –NH residues,5 monitor glycogen concentration in the liver,6 track glucose accumulation in tumors,7,8 image myo-inositol and glutamic acid in the brain9–11 and map a specific gene expression in vivo.12 Using CEST we also imaged changes in glycosaminoglycan (GAG) levels in human joints in vivo following injury13 followed by changes in the GAG contents in the nucleus pulposus part of intervertebral disks

⇑ Corresponding author. Tel.: +972 3 6408156; fax: +972 3 6410665. E-mail address: [email protected] (G. Navon). http://dx.doi.org/10.1016/j.carres.2014.02.008 0008-6215/Ó 2014 Elsevier Ltd. All rights reserved.

in vitro.14 In this paper we demonstrate that CEST is a very sensitive method for the detection of polysialic acid and related oligomers. Sialic acids (SA) are a family of nine carbon acidic monosaccharides that occur naturally at the end of sugar chains. They are attached to cell-surfaces and proteins as in glycoproteins, and peptides as in mucopolysaccharides. SAs are ubiquitous in nature, especially in mammals, playing a central role in many physiological and pathological processes.15 The mammalian central nervous system has the highest concentration of SA, most of which is present in gangliosides and glycoproteins.16 SA also plays a role in the regulation of immune response, and alteration in sialylation is associated with malignancy and metastatic phenotypes of various types of tumors.17,18 For instance, it has been shown that SA is overexpressed in cells from colon cancer tissues, and its mucosal expression correlated to the metastatic stage.19 Therefore, SA is an important molecular target for diagnostic and therapeutic approaches.20 SA can form linear homo-polymers (pSA) of various lengths, between 8 and 280 residues depending on the biological context, linked through a-2,8 and/or a-2,9 glycosidic bonds. Due to its location at cell surfaces, pSA plays a vital role in biological processes such as embryogenesis,21 neural cell growth, differentiation, cell–cell interactions, and membrane transport.22 pSA is also

166

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

implicated in pathological processes such as tumor progression, especially nodal and distant metastases in lung cancer cells,23,24 and is an important virulence factor of Neisseria meningitidis and Escherichia coli K1.25,26 Recently the structure of pSA was determined in living bacteria by NMR spectroscopy.27 SA has six exchangeable hydrogen atoms (Fig. 1A) in five hydroxyl groups and one NH group. In the case of pSA, nearly all the residues have only three hydroxyl groups since two are lost when the glycosidic linkage is formed (Fig. 1B). As we show in this article, the proton exchange of the hydroxyl groups makes CEST a sensitive method to detect SA and pSA in vivo. Furthermore, since pSA is present in so many contexts, CEST has great potential in cancer detection and tracking human development, because both involve changes in the quantities of pSA being expressed. 2. Materials and methods SA (N-acetylneuraminic acid) was purchased from Sigma–Aldrich. a-2,8-linked sialic acid oligomers (one to six units) were purchased from Nacalai Inc. (San Diego, CA.). pSA was obtained from a variant of E. coli K12, as previously reported.28,29 The polymer we studied is a linear homo-polymer with a length varying from 170 to 280 SA residues, linked a-2,8 (Fig. 1B). 2.1. Extraction of pSA from bacterial cell paste A 10 g sample of E. coli K1230 bacterial paste was mashed with a spatula, while 150 mL of a filtered (0.22 lm StericupÒ filter, Millipore Corporation, Billerica, MA), 1 M calcium chloride solution was added to the paste drop-wise over the course of one hour. The homogeneous mixture was then centrifuged at 7500g and 4 °C30 for 30 min and the cell debris pellet was discarded. The supernatant was adjusted to 25% (v/v) ethanol and stirred on ice for 30 min to precipitate nucleic acids. The resulting mixture was centrifuged at 7500g and 4 °C for 30 min, and the pellet was discarded. The clear supernatant was adjusted to 80% (v/v) ethanol, mixed, and stored at 20 °C freezer for one hour to precipitate crude polysaccharides. The appearance of a cloudy white solid marked the precipitation of pSA material; if the solution remained clear, 1 M CaCl2 was added dropwise (to induce salting-out). The

mixture was then centrifuged in a Corex tube at 7500g and 4 °C for 30 min, and the alcohol supernatant was discarded. The crude polysaccharide pellets were combined and dissolved in 40 mL of filtered 10% (v/v) saturated sodium acetate solution and placed on ice. An equal volume of buffered phenol solution (3.57 g phenol/1 mL 10% (v/v) saturated sodium acetate solution) was added for extraction of protein material, and the mixture was stirred for 30 min on ice. The resulting emulsion was resolved by centrifugation at 7500g and 4 °C for 30 min, yielding a biphasic mixture. The upper aqueous layer containing the pSA was carefully decanted by vacuum aspiration. Remaining nucleic acids and proteins were removed by at least one additional cycle of 25% (v/v) ethanol precipitation, followed by re-precipitation of pSA material at 80% (v/v) ethanol, and phenol extraction, as described above. The whiter, purified pSA was dissolved in 10 mL of filtered water and dialyzed against filtered water overnight (Fisher, Pittsburgh, PA; MWCO = 12,000–14,000 kDa). The dialysate was ultracentrifuged at 170,000g and 4 °C for 90 min, and the supernatant was decanted into a pre-weighed conical tube. The solution was shell-frozen on dry ice and then lyophilized to yield 100–300 mg of pSA. All NMR measurements were performed on Bruker 500 MHz DRX, and Avance III spectrometers. 2.2. CEST measurements CEST spectra were recorded in aqueous solutions containing 10% D2O after detuning of the probes to avoid radiation damping. B1 fields were in the range of 0.6–3.0 lT (25–125 Hz), irradiation time 3 s, and relaxation delay of 7 s (except for amide proton measurements where irradiation time of 10 s was used). The read pulse was in most cases a hard pulse, but using a selective pulse instead gave the same results. The %CEST was quantified by measuring the asymmetry in the magnetization transfer ratio (MTRasym) according to the following equation:

MTRasym ¼ ½MðxÞ  MðxÞ=MðxÞ where M(x) and M(x) are the water signal intensities with saturation at x and x from the water signal, respectively.

A

B Figure 1. (A) Structure of sialic acid (SA) monomer showing that the monomer is present in dynamic equilibrium between the b and a anomers in a 94:6 ratio, respectively. (B) a-2,8 Linked polymer (pSA), the structure of the subject of the present study, a-2,8 linked pSA.

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

where c is the concentration of the exchanging protons, kb is the exchange rate and Ra1 is 1/T1 of the water protons.

2.3. Hydroxyl proton assignments via HSQC-TOCSY

a-2,8-Linked sialic acid oligomers (one to six units) were purchased from Nacalai Inc. (San Diego, CA). Aliphatic proton and carbon chemical shifts were assigned as previously described.31 SA oligosaccharides were dissolved in deionized water to a final volume of 300 lL containing 0.1% DSS and 0.05% NaN3. All raw NMR data from HSQC-TOCSY experiments were processed with nmrPipe software32 applying Gaussian multiplication in both dimensions. Linewidths were extracted from fitting utilizing the fitTab.tcl script. For each oligomer, the pH was adjusted to maximize OH peak intensity and HSQC-TOCSY experiments were carried out once the optimal pH was reached (pH 7–7.5). The spectral window and carrier frequencies were set to 10 and 22 ppm, and 4.75 and 69 ppm for proton and carbon, respectively. A 10 ms DIPSI mixing time was utilized in all experiments. In the case of SA dimer and trimer, 128 experiments were acquired in the indirect dimension, whereas for the tetramer, pentamer, and hexamer, 256 t1 points were acquired. For SA monomer, a 13C labeled sample was utilized, for which a constant time 13C evolution was incorporated to the HSQC-TOCSY pulse sequence. The spectral window and carrier frequencies were set to 10 and 3 ppm, and 4.75 and 69 ppm for proton and carbon, respectively. The exchange rates for each hydroxyl proton at 10 °C were estimated from the line width of each cross peak in the 1H dimension utilizing the equation: kex ¼ p ðW 1=2

observed

 W 1=2

natural Þ

3. Results and discussion The one dimensional NMR spectrum of a 300 mM solution of SA in water (pH 6.7, at 25 °C) is given in Figure 2. The NH and aliphatic protons were previously assigned,34,35 however, the hydroxyl protons were not. The peaks of the a anomer of the H3(ax) and the H3(eq) protons are clearly resolved and correspond to concentration ratio of 94:6 between the b and a anomers, respectively. The

ð1Þ

where W1/2 is the line width at half height in Hz. Natural line width was obtained from averaged aliphatic proton line widths, which ranged from 5 to 6 Hz. The exchange rates at 37 °C were estimated from the CEST effect as a function of the irradiation field B1 (x1 in units of rad/s) according to the following equation:33

1 55:5 1 1 ¼ kb Ra1 2 þ 2 MTRasym c x kb 1

167

!

ð2Þ

Figure 3. One dimensional 1H NMR spectra of the hydroxyl region of a 50 mM solution of sialic acid in 10% D2O, at 5 °C as a function of pH (256 scans). The 1H probe coil was detuned for these experiments.

Figure 2. 1H NMR spectra of a solution of 300 mM SA in 10% D2O, pH 6.7, 25 °C, acquired with a 45° pulse, 256 scans and a relaxation delay of 1.0 s. Aliphatic proton assignment is taken from Ref. 34. The 1H probe coil frequency was detuned to avoid radiation damping. Asterisks denote impurities.

168

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

Figure 4. 1H and Watergate NMR spectra of 300 mM SA in 10% D2O, pH 6.1 at 5 °C. Number of scans 128, relaxation delay 1 s, 90° pulse. Watergate was measured with the Bruker zggpw5 pulse sequence using Watergate W5 pulse sequence with gradients using double echo. The OH residues were assigned by 1D selective TOCSY measurements.

hydroxyl and water chemical shifts and line widths depend strongly on pH (Fig. 3) and temperature. Figure 4 shows that the resolution and the S/N are dramatically improved when the WATERGATE pulse sequence was used to suppress the water signal.36 Only four of the expected five hydroxyl protons are resolved in the 1D spectrum. In order to assign the signals arising from the OH protons we used two methodologies: (1) selective 1D TOCSY (total correlation spectroscopy) and (2) HSQC-TOCSY. All our assignments for the sialic acid refer to the b-anomer, which is the predominant (94%) isomer. In this paper, we report the chemical shifts of the hydroxyl protons relative to the water chemical shift, which is set to zero ppm, unless noted otherwise. Since the chemical shift of water is temperature-dependent, calculation of the absolute chemical shifts must take the temperature into consideration. 3.1. Hydroxyl peak assignments Hydroxyl proton assignments via 1D selective TOCSY: in the selective 1D TOCSY measurements one peak is selectively excited and, after different mixing times, its magnetization is transferred

to the J-coupled protons in a stepwise process. This technique is useful for dividing the proton signals into groups of coupled spin systems. For example, as a result of the selective excitation of the aliphatic H3(eq) (Fig. 2), the OH signals at 1.7 and 1.2 ppm downfield from the water increased in intensity at short mixing times (20–60 ms) and gradually decayed at longer times (80–120 ms) (Fig. 5A). Similar results were obtained upon selective excitation of the aliphatic H3(ax) (not shown). Selective excitation of H4 and H6 increased the signal intensity at 1.2 ppm. These results indicate that the 1.7 and 1.2 ppm signals correspond to the OH groups on C2 and C4, respectively. A confirmation of this assignment was obtained from selective excitation of the OH proton at 1.7 ppm, which increased the intensity of both H3(eq) and H3(ax) signals. After selective excitation of H90 and H9 the magnetization is mostly transferred to the OH signals at 0.76 and 0.54 ppm (Fig. 5B), whereas, selective excitation of H8, transfers the magnetization only to the OH signal at 0.76 ppm. As H9 is the only signal that transfers its magnetization to the OH at 0.54 ppm, we assign this signal to the OH on C9. Accordingly the signal at 0.76 ppm is assigned to both OH on C8 and OH on C7. Indeed, above pH 7, this signal is a superposition of two signals (Fig. 3).

Figure 5. One dimensional 1H selective TOCSY measurements as a function of TOCSY mixing time, performed on a 300 mM solution of SA, pH 6.1, 10% D2O, 5 °C (256 scans). The selective pulses are centered at the frequency of H3(eq) (A) and H90 (B). The upper traces are the spectra without off-resonance irradiation.

169

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

Hydroxyl Region

Aliphatic Region I-7

I-7

67.5

I-4

I-4 68.0

II-4

VI-7

VI-7

II-4

II-7

68.5

III/IV/V-4

III-7 III-7

IV-7

VI-4

IV-7

13 C ppm

II-7

III/IV/V-4 VI-4

69.0

V-7 OH

HO

69.5

V-7

O

70.0 6.4

6.3

6.2

6.1

6.0

5.9

5.8

5.7

5.6

4.0

5.5

3.9

3.7

3.8

3.6

3.5

1H ppm

V

VI

IV O

O

O

HO HO

OH

C

OH

O

O HO

H N

O

HO

OH

OH

C

H N

O

O

O HO

HO

O C

OH

H N

O

OH

OH

O HO

HO

H N

O

O OH

OH

C

O

C

O

O HO

HO

I

II

III O

H N

O

OH

9

2

8

11 10

OH C

7 6

O HO

HO

OH

OH

O O

H N

5

HO

4

1

3

O

Figure 6. Comparison between the HSQC-TOCSY spectra of the hydroxyl and the aliphatic proton signal regions at 5 °C, using a 10 ms TOCSY mixing time (right panel). The spectra show that correlations from hydroxyl 1H’s to their covalently bonded carbons provide additional signal dispersion that is not provided by the aliphatic 1H’s covalently attached to the same carbons (right panel). Table 1 Chemical shift assignments of a-2,8 SiA oligomers

Monomer Dimer Trimer Tetramer Pentamer Hexamer Assignment 13C (ppm) 1H (ppm) 13C (ppm) 1H (ppm) 13C (ppm) 1H (ppm) 13C (ppm) 1H (ppm) 13C (ppm) 1H (ppm) 13C (ppm) 1H (ppm) I-OH2,C3 39.49 6.26 39.62 6.43 39.46 6.65 39.31 6.67 39.37 6.71 39.30 6.66 I-OH4,C4 67.46 6.20 67.70 6.34 67.75 6.32 67.61 6.31 67.68 6.31 67.60 6.31 I-OH7,C7 68.56 5.86 66.57 5.90 67.49 5.89 67.49 5.89 67.68 5.90 67.43 5.88 I-OH8,C8 70.31 5.77 I-OH9,C9 63.40 5.58 60.67 5.47 61.10 5.47 60.88 5.48 61.00 5.47 60.85 5.49 II-OH4,C4 68.62 6.48 68.26 6.40 68.23 6.38 68.38 6.38 68.24 6.39 II-OH7,C7 67.95 5.85 69.14 5.74 68.56 5.77 68.75 5.76 68.65 5.77 II-OH8,C8 72.47 6.43 II-OH9,C9 62.59 5.85 61.49 5.78 61.06 5.81 61.09 5.83 60.98 5.81 IIII-OH4,C4 68.83 6.45 68.23 6.38 68.38 6.38 68.34 6.38 IIII-OH7,C7 68.28 5.78 69.30 5.56 68.89 5.58 68.79 5.55 IIII-OH8,C8 72.17 6.40 IIII-OH9,C9 62.78 5.82 61.60 5.66 no no 61.28 5.78 IV-OH4,C4 68.76 6.43 68.38 6.38 IV-OH7,C7 68.17 5.79 69.44 5.55 68.82 5.57 IV-OH8,C8 72.01 6.43 IV-OH9,C9 62.62 5.81 61.61 5.73 61.28 5.78 V-OH4,C4 68.80 6.43 68.34 6.38 V- OH7,C7 68.28 5.79 69.34 5.55 V- OH8,C8 72.05 6.47 V-OH9,C9 62.75 5.82 no no VI-OH4,C4 68.74 6.43 VI-OH7,C7 68.17 5.79 VI-OH8,C8 72.00 6.43 VI-OH9,C9 62.63 5.81 The data were collected in a Bruker Avance III 500 MHz instrument at 5 °C, the samples used were 13C natural abundance SiA monomer sample was 13C labeled and the data were collected in a Bruker Avance 700 MHz instrument. DSS was used as reference.

170

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

Hydroxyl proton assignments via HSQC-TOCSY: the optimal pH for OH assignment was found to be in the 7.0–7.5 range. Even slight changes in pH (0.5 units) dramatically affect OH signal intensity (Fig. 6). Interestingly, the optimal pH for OH detection and for their utilization in CEST experiments, closely matches that of the physiological milieux. Efficient hydroxyl assignments for SA monomer and a-2,8 SA oligomers with degree of polymerization (DP) of 2–6 was realized by utilizing a recently published technique,31 which provides sufficient resolution to resolve

70

Monomer Dimer Trimer Tetramer Pentamer Hexamer

OH7

66

13

39.6

OH2

39.4

13

64 OH9 62

6.2

6.4 1

60 5.4

3.2. CEST effects

OH4

68

C ppm

OH8

OH8

5.6

C ppm

72

5.8

6.0

6.2

6.6

39.2

H ppm

6.4

overlapping signals in the proton dimension performed at 5 °C (Fig. 6). It is clear from Figure 6 that OH assignment will be more complicated for larger oligomers because the selective 1D TOCSY experiment relies on correlating only the aliphatic proton signals, whose overlap increases as the oligomerization increases. Assignment of hydroxyl protons of longer SA oligomers (Table 1) also affords the assignment of those in the polymer, because it provides a range of chemical shifts for each hydroxyl proton at each position (Fig. 7) that can be extrapolated from what is observed in pSA. Utilizing this strategy we confirmed the assignment obtained for SA monomer based on selective TOCSY experiments, in a single experiment.

6.6

1

H ppm

Figure 7. Hydroxyl 1H and 13C signals for SAs ranging from one to six units. Chemical shifts (relative to DSS as a reference) were obtained from HSQC-TOCSY experiments performed at 5 °C on a 500 MHz Avance III instrument, as described in Materials and methods. Hydroxyl signals from different residues are color-coded and the atom position is indicated by Arabic numerals. The number of correlations depends on the number of SA residues in a given oligomer, that is, the monomer shows one marker for each OH peak, the dimer shows two for each OH peak, the trimer shows three for each OH. The pattern continues up to six for the hexamer. For the cross peak labeled OH2, the 13C chemical shifts correspond to couplings between OH2 to hydrogen atoms attached at C3.

The MTR asymmetry plots (MTRasym) for a solution of 300 mM SA, pH 6.7 at 25 °C and B1 fields of 25, 30 and 50 Hz are given in Figure 8. Well-resolved CEST effects are observed from OH groups on C2 and C4 (1.68 and 1.2 ppm downfield from the water). The broad CEST signal at 0.76 ppm and closer to the water, corresponding to the OH residues on C7, C8 and C9, is partially resolved only at the low B1 field. At neutral pH values, the amide proton at 3.28 ppm downfield from the water did not exhibit a CEST effect, indicating that this NH group is in slow exchange with water. However, at pH = 9.1, CEST was also obtained for the amide protons, reaching a value of 25% for a 100 mM solution, at B1 of 54 Hz and irradiation time of 10 s. In the 1H NMR spectrum of pSA at pH 7.4, no hydroxyl protons are resolved (Fig. 9). Nevertheless, a significant CEST effect was observed. Figure 10 shows MTRasym plots at different B1 values for a pSA solution at a concentration corresponding to 20 mM of SA residues, at pH 7.4 and 36 °C. A very broad CEST signal is observed between 1.2 and 0.6 ppm. This broad signal consists of a partially resolved peak at 1.2 ppm and an unresolved broad peak between 0.2 and 0.6 ppm downfield from the water. The spectrum is similar to that obtained for the SA except that the peak at 1.7 ppm is absent in the pSA MTRasym plot. This is expected since the linkage for

Figure 8. The percent CEST of a solution of 300 mM SA (pH 6.7, 10% D2O, 25 °C) as a function of the distance from the water for B1 fields of 50, 30 and 25 Hz. The B1 saturation pulse lasted 3 s, relaxation delay 3 s and number of scans 8.

171

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

Figure 9. 1H NMR spectrum of an aqueous solution of pSA at a concentration corresponding to 20 mM sialic units, in 10% D2O, pH 7.4, 25 °C.

2

1.5

1

0.5

0

Figure 10. The percent CEST of an aqueous solution of pSA at a concentration corresponding to 20 mM sialic units, pH 7.4, 10% D2O, 36 °C as a function of the distance from the water for B1 fields of 94, 60 and 47 Hz. The duration of the B1 saturation pulse was 3 s, relaxation delay 7 s and number of scans 8.

this pSA is a-2,8, thus, there is no OH group at C2. Consequently, the observed signal at 1.2 ppm is assigned to the OH on C4, as in SA and its oligomers. The data obtained from hydroxyl group assignment in SA oligomers indicate that the broad peak between 0.2 and 0.6 ppm downfield from the water corresponds to the overlapped resonances of OH7 and OH9. This broad peak is similar to

that of SA, except that in the region between 0.6 and 0.8 the signal intensity is significantly decreased in the 1H spectrum of pSA, in line with our previous assignment of the contribution of the OH8 to the signal in that region. The MTR asymmetry was measured for seven preparations of pSA. The average %CEST at 1.2 ppm downfield to the water

172

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

obtained for solutions containing 20 mM sialic acid residues at pH 7.4, 36 °C and a B1 field of 115 Hz (2.7 lT) was 14 ± 2%. For comparison, for 20 mM SA, pH 7.5, and 36 °C the percent CEST at 1.2 ppm was 11.5%, which is similar to the average obtained for the various preparations of pSA under the same conditions. Like the case of SA, for pSA at neutral pH, the amide proton at 3.28 ppm downfield from the water did not exhibit a CEST effect. However, at pH = 9.4 CEST was obtained from the amide protons, reaching a value of 10% for a 100 mM solution, at B1 of 51 Hz and irradiation time of 10 s. This result is lower than that obtained for SA under similar conditions discussed earlier, indicating slower exchange rate for NH group of the pSA. CEST efficiency depends on the rate constant for chemical exchange between the irradiated proton and the bulk water. Dixon et al. recently showed33 that the exchange rate could be measured in a concentration independent manner by measuring the CEST effect as a function of the irradiation field B1 in units of rad/s rate. The CEST data obtained at 1.2 ppm (OH4) downfield from the water for three different pSA preparations yield an exchange rate constant of 540 s1 (Fig. 11). The peak at 1.2 ppm is not completely resolved and has an increasing contribution from (x1) (Eq. 2). The x intercept in a plot of (CEST)1 versus (1/x1)2 yields the exchange the other peaks with higher values of B1, therefore, the rate constant obtained from this figure is only a rough estimation. This value is at the lower end of the range of exchange rates found for a variety of mono- and polysaccharides.37–39 However these studies were performed by measuring the CPMG dispersion of water that is governed by the exchange rate of the fastest exchanging OH groups whose identity is usually unknown. We also estimated the exchange rate for the hydroxyl protons of SA oligomers from linewidths at 10 °C (Table 2 and Fig. 12). When comparing the estimated exchange rates at 10 °C to that of OH4 at 37 °C obtained for the pSA from the CEST experiment, that is, 540 s1, one may consider the exchange rates of OH4 of the central residues of the oligomers that were 15, 36, 46 and 13 s1 for the trimer, tetramer, pentamer and hexamer, respectively. The increase of exchange rates by factors of 12–42 following a temperature change of 47 degrees is reasonable. These results

Table 2 Exchange rate constants for OH groups of SiA oligomers at pH 7–7.5, 5 °C Monomer ( ) Residue I

2 4 7 8 9

101.90 128.62 132.90 137.84 117.90

Dimer Trimer Tetramer Pentamer Exchange Rate Constants (s-1)

Hexamer

3.41 4.01 19.64 21.07 49.35 52.06

6.46 31.07 22.96

7.07 33.27 32.18

3.15 16.09 38.54

30.39 40.35

23.28

29.96

20.22

2 4 7 8 9

16.35 14.70 9.98 11.76 6.38 42.80 34.76

36.20 7.67

43.57 9.88

11.35 8.94

21.67

28.43

32.78

2 4 7 8 9

19.59 19.55 8.14 no

36.20 8.74

43.57 13.83

13.47 15.64

43.16

no

no

45.79 11.84 3.50 47.08

43.57 11.57

13.47 13.70

41.49

no

36.03 12.97 6.00 34.79

13.47 10.35

Residue II

Residue III

Residue IV

2 4 7 8 9 Residue V

2 4 7 8 9

no

Residue VI

2 4 7 8 9

16.63 10.18 4.81 54.24

The data were collected in a Bruker Avance III 500 MHz instrument.

60

20

OH9

50

18 40

16

1/CEST

14

30

OH4

12 10

20

OH7

8

OH2 OH 8

6

PSA1 PSA2 PSA3

4 2

1.65 (6.75)

-20

10

1.25 (6.35)

1.05 (6.05)

0.85 (5.95)

0.65 (5.75)

0.45 (5.55)

0 0.25 (5.35)

Offset from H2O

0 -40

1.45 (6.55)

Exchange rate constants (s-1)

Dimer Trimer Tetramer Pentamer Hexamer

0

20

2

40

60

-2

(1/ω1) (rad/sec) x 10

80

100

120

1

( H ppm)

-7

Figure 11. A linear plot relating the CEST with the irradiation power B1 designed to evaluate the exchange rate from the negative intercept on the x-axis. The line was fitted to the combined results from solutions of three pSA preparations containing 20 mM sialic residues pH 7.4, 37 °C.

Figure 12. Averaged hydroxyl exchange rate as a function of averaged chemical shifts of SAs. Exchange rates were estimated from proton line-widths extracted from slices (t1 points) of HSQC-TOCSY experiments. Rates were calculated utilizing the equation Kex = p⁄(W1/2observed  W1/2natural), where W1/2 is the line width at half height. An averaged natural line width of 6 Hz was used for all hydroxyl groups and was calculated from the line-widths of all aliphatic protons.

H. Shinar et al. / Carbohydrate Research 389 (2014) 165–173

imply that assignment of OH groups, in combination with the determination of the exchange rate profile can be used to design and predict CEST experiments in larger molecules. 4. Concluding remarks In this work, we show that CEST effects can be observed for OH groups and proton exchange rates can be measured at physiological conditions (pH = 7.4 and 37 °C), in spite of the fact that the OH groups are not directly detected in the NMR spectrum under these conditions. The sensitivity of the CEST experiments was about 0.7% for 1 mM of sialic acid residues. Considering that pSA length may reach 280 sialic residues, one could detect a 1% CEST for 5 lM polymer. This value should be considered when the potential of the technique for the in vivo measurements of bacteria or metastatic potential of tumors will be assessed. In vivo measurements will no doubt be complicated by additional resonances from other glycans and metabolites. However, two points should be considered with regard to in vivo practicality. First, it is expected that the density of PSA on cells will be higher than the density of other glycans or metabolites. Second, the H3 resonances in sialic acid and PSA are unique among glycans and can therefore be used as NMR ‘handles’ to selectively excite or detect resonances that are distinct from other metabolites in the NMR spectrum. Nevertheless, we can only verify these points through future in vivo studies. Sialo-CEST will also provide a means to differentiate tumors that display PSA on their surfaces from those that do not. Assignment of individual OH resonances in SA oligomers and estimation of the OH exchange rates in vitro are important because: (1) they explain the dependence of the CEST enhancement on the offset of the observed resonances for the polymer; (2) the exchange rate of labile protons directly relates to the percent CEST observed; (3) because NMR enables discrimination of frequencies within the same molecule, selective irradiation of selected frequencies can be utilized for contrast fine-tuning using the same endogenous contrast agent in vivo; and (4) the pH directly affects OH exchange rates and the properties of the observed signal; temperature and salt concentration impact the exchange rate as well.40 Therefore, knowledge of how exchange rates are affected in various conditions in vitro can provide valuable information of molecular environment in vivo. Many glycans are ideal for CEST experiments since they possess numerous exchangeable hydroxyl hydrogens. Furthermore, in polysaccharides, these resonances overlap, allowing one to simultaneously excite many OH groups at a single resonance position without the technical difficulty of selecting only certain OHs. We expect that in the future, glycans will be a useful class of molecules for CEST and hence contrast agents. This also provides the opportunity to design new polysaccharides that may have improved CEST profiles. Acknowledgment We gratefully acknowledge Dr. Willie Vann and Dr. Eric Vimr, who provided us with the E. coli variant necessary to prepare the a-2,8 polysialic acid used in this study.

173

References 1. 2. 3. 4. 5. 6. 7.

8.

9. 10. 11. 12.

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

24.

25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

Zhou, J.; Zijl, P. C. M. v. Prog. Nucl. Magn. Reson. Spectrosc. 2006, 48, 109–136. Vinogradov, E.; Sherry, A. D.; Lenkinski, R. E. J. Magn. Reson. 2013, 229, 155–172. Liu, G.; Song, X.; Chan, K. W.; McMahon, M. T. NMR Biomed. 2013, 26, 810–828. Zhou, J.; Payen, J. F.; Wilson, D. A.; Traystman, R. J.; van Zijl, P. C. Nat. Med. 2003, 9, 1085–1090. Zhou, J.; Lal, B.; Wilson, D. A.; Laterra, J.; van Zijl, P. C. Magn. Reson. Med. 2003, 50, 1120–1126. van Zijl, P. C.; Jones, C. K.; Ren, J.; Malloy, C. R.; Sherry, A. D. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 4359–4364. Kannie, W. Y.; Chan, M. T.; McMahon, Y. K.; Liu, G.; Jeff, W. M.; Bulte, Z. M.; Zaver, M. B.; Dimitry, A.; van Zijl, P. C. M. Magn. Reson. Med. 2012, 68, 1764– 1773. Walker-Samuel, S.; Ramasawmy, R.; Torrealdea, F.; Rega, M.; Rajkumar, V.; Johnson, S. P.; Richardson, S.; Goncalves, M.; Parkes, H. G.; Arstad, E.; Thomas, D. L.; Pedley, R. B.; Lythgoe, M. F.; Golay, X. Nat. Med. 2013, 19, 1067–1072. Haris, M.; Cai, K.; Singh, A.; Hariharan, H.; Reddy, R. Neuroimage 2011, 54, 2079–2085. Haris, M.; Nath, K.; Cai, K.; Singh, A.; Crescenzi, R.; Kogan, F.; Verma, G.; Reddy, S.; Hariharan, H.; Melhem, E. R.; Reddy, R. NMR Biomed. 2012, 26, 386–391. Cai, K.; Haris, M.; Singh, A.; Kogan, F.; Greenberg, J. H.; Hariharan, H.; Detre, J. A.; Reddy, R. Nat. Med. 2012, 18, 302–306. Gilad, A. A.; McMahon, M. T.; Walczak, P.; Winnard, P. T., Jr.; Raman, V.; van Laarhoven, H. W.; Skoglund, C. M.; Bulte, J. W.; van Zijl, P. C. Nat. Biotechnol. 2007, 25, 217–219. Ling, W.; Regatte, R. R.; Navon, G.; Jerschow, A. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 2266–2270. Saar, G.; Zhang, B.; Ling, W.; Regatte, R. R.; Navon, G.; Jerschow, A. NMR Biomed. 2012, 25, 255–261. Varki, A. Trends Mol. Med. 2008, 14, 351–360. Schnaar, R. L. FEBS Lett. 2010, 584, 1741–1747. Fukuda, M. Cancer Res. 1996, 56, 2237–2244. Dall’Olio, F.; Chiricolo, M. Glycoconjugate J. 2001, 18, 841–850. Azab, A. K.; Kleinstern, J.; Srebnik, M.; Rubinstein, A. Pharm. Res. 2008, 25, 379– 386. Matsumoto, A.; Cabral, H.; Sato, N.; Kataoka, K.; Miyahara, Y. Angew. Chem., Int. Ed. 2010, 49, 5494–5497. Probstmeier, R.; Bilz, A.; Schneider-Schaulies, J. J. Neurosci. Res. 1994, 37, 324– 335. Rutishauser, U.; Landmesser, L. Trends Neurosci. 1966, 19, 422–427. Tanaka, F.; Otake, Y.; Nakagawa, T.; Kawano, Y.; Miyahara, R.; Li, M.; Yanagihara, K.; Inui, K.; Oyanagi, H.; Yamada, T.; Nakayama, J.; Fujimoto, I.; Ikenaka, K.; Wada, H. Cancer Res. 2001, 61, 1666–1670. Tanaka, F.; Otake, Y.; Nakagawa, T.; Kawano, Y.; Miyahara, R.; Li, M.; Yanagihara, K.; Nakayama, J.; Fujimoto, I.; Ikenaka, K.; Wada, H. Cancer Res. 2000, 60, 3072–3080. Swartley, J. S.; Ahn, S. H.; Liu, L. J.; Kahler, C. M.; Stephens, D. S. J. Bacteriol. 1996, 178, 4052–4059. Vogel, U.; Hammerschmidt, S.; Frosch, M. Med. Microbiol. Immun. 1996, 185, 81–87. Azurmendi, H. F.; Vionnet, J.; Wrightson, L.; Trinh, L. B.; Shiloach, J.; Freedberg, D. I. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 11557–11561. Methods Enzymol.; Vann, W. F., Freese, S., Eds.; Academic Press: San Diego, 1994; Vol. 235, pp 304–311. Ferrero, M. A.; Aparicio, L. R. Appl. Microbiol. Biotechnol. 2010, 86, 1621–1635. Vimr, E. R. J. Bacteriol. 1992, 174, 6191–6197. Battistel, M. D.; Pendrill, R.; Widmalm, G.; Freedberg, D. I. J. Phys. Chem. B 2013, 117, 4860–4869. Delaglio, F.; Grzesiek, S.; Vuister, G. W.; Zhu, G.; Pfeiffer, J.; Bax, A. J. Biomol. NMR 1995, 6, 277–293. Dixon, W. T.; Ren, J.; Lubag, A. J. M.; Ratnakar, J.; Vinogradov, E.; Hancu, I.; Lenkinski, R. E.; Sherry, A. D. Magn. Reson. Med. 2010, 63, 625–632. Otesuka, H.; Uchimura, E.; Koshino, H.; Okano, T.; Kataoka, K. J. Am. Chem. Soc. 2003, 125, 3493–3502. Battistel, M. D.; Shangold, M.; Trinh, L. B.; Shiloach, J.; Freedberg, D. I. J. Am. Chem. Soc. 2012, 134, 10717–10720. Piotto, M.; Saudek, V.; Sklenar, V. J. Biomol. NMR 1992, 2, 661–665. Hills, B. P.; Cano, C.; Belton, P. S. Macromolecules 1991, 24, 2944–2950. Fabri, D.; Williams, M. A. K.; Halstead, T. K. Carbohydr. Res. 2005, 340, 889–905. Aroulmoji, V.; Mathlouthi, M.; Feruglio, L.; Mjurano, E.; Grassi, M. Food Chem. 2012, 132, 1644–1650. Battistel, M. D.; Azurmendi, H. F.; Yu, B.; Freedberg, D. I. Prog. NMR Spec., 2013. in press.