Chemoselective tryptophan labeling with rhodium carbenoids at mild pH

Abstract

Significant improvements have been made to a previously reported tryptophan modification method using rhodium carbenoids in aqueous solution, allowing the reaction to proceed at pH 6-7. This technique is based on the discovery that N-(tert-butyl)hydroxylamine promotes indole modification with rhodium carbenoids over a broad pH range (2-7). This methodology was demonstrated on peptide and protein substrates, generally yielding 40-60% conversion with excellent tryptophan chemoselectivity. The solvent accessibility of the indole side chains was found to be a key factor in successful carbenoid addition, as demonstrated by conducting the reaction at temperatures high enough to cause thermal denaturation of the protein substrate. Progress toward the expression of proteins bearing solvent accessible tryptophan residues as reactive handles for modification with rhodium carbenoids is also reported.

Figure 1

Modification of tryptophan residues with metallocarbenes. a) Horse heart myoglobin was exposed to 1 and Rh₂(OAc)₄ for 7 h. b) A representative ESI-MS spectrum of modified myoglobin. c) No modification is observed without Rh₂(OAc)₄. All assigned species agree to within 0.1% of the expected mass values.

Figure 2

The effect of reaction pH with using H₂NOH as a reaction additive was monitored for (i) the catalytic degradation of diazo 1 and for the modification of (ii) horse heart myoglobin and (iii) subtilisin Carlsberg. In all cases, no reaction was observed at pH ≥ 6. The degradation of 1 was assessed by RP-HPLC. The extent of modification for myoglobin was estimated by ESI-MS, and the values represent the sum total of both singly and doubly modified species. The extent of modification for subtilisin was estimated qualitatively by MALDI-TOF-MS (primarily single modification was observed). ^aThe extent of conversion for each data point is expressed as a percentage of the maximum conversion observed for that substrate across the entire pH range. The absolute conversion was estimated to be 90% for myoglobin and 50% for subtilisin Carlsberg.

Figure 3

pH-dependent binding of H₂NOH to Rh₂(OAc)₄ in aqueous solution. At low pH the color of Rh₂(OAc)₄ in solution is consistent with coordination by water or possibly the oxygen of protonated hydroxylamine, while above pH 6.0 the color is consistent with coordination via the nitrogen atom of hydroxylamine. Samples contain 1 mM Rh₂(OAc)₄ in 80% water / 20% ethylene glycol in the presence of 75 mM H₂NOH at pH 3.0 and pH 6.0.

Figure 4

Modification of melittin. (a) Crystal structure of melittin (PDB ID: 2MLT, W in orange) and sequence. (b) Typical MALDI-TOF-MS spectrum for the modification of melittin at pH 6.0 with tBuHNOH. Expected masses are m/z 2846 for unmodified, 3152 for singly modified, and 3458 for doubly modified peptides.

Figure 5

The use of tBuNHOH substantially improved the pH range over which successful modification of melittin could be achieved. Note: Percent conversion was estimated by MALDI-TOF-MS (sinapinic acid matrix) and values represent the average of three independent analyses of the same sample (standard deviation indicated by error bars). Conversion values represent the sum total of all modified species.

Figure 6

MS/MS analysis of doubly modified species. (a) Melittin sequence and expected y" and b ions assuming double modification of tryptophan. (b) MS/MS spectrum of the (M+5H)⁵⁺ ion of the mass 3450 Da. Multiply charged ions have been transformed onto a singly charged x-axis to simplify spectral interpretation.

Figure 7

Effect of buffer/additive identity on the modification of myoglobin and melittin. Modification of horse heart myoglobin in the presence of (a) 75 mM H₂NOH•HCl (pH ~3.5) and (b) 75 mM phosphate (pH 3.5). Modification of melittin in the presence of (c) 75 mM tBuNHOH (pH 6.0) and (d) 75 mM phosphate (pH 6.0). Note: (a) and (b) are ESI-MS spectra, while (c) and (d) are MALDI-TOF-MS spectra.

Figure 8

Proposed binding of tBuNHOH with Rh₂(OAc)₄ at pH 6.0.

Figure 9

Modification of lysozyme requires thermally denaturing conditions. (a) The six tryptophan side chains are shown in green. A 100 µM solution of lysozyme was modified and following removal of the small molecules via gel filtration, the sample was analyzed by ESI-MS. No modification was observed in the absence of Rh₂(OAc)₄ under otherwise identical reaction conditions. (b) Modification of lysozyme using the same conditions as in (a) at various temperatures. Percent conversion is calculated as the combined area of singly and doubly modified peaks as a ratio to the total area for all protein peaks. Conversion was only observed at 75 °C and above.

Figure 10

Addition of tryptophan containing “tags” to allow protein modification. (a) Crystal structure of wild type FKBP with rapamycin bound (PDB ID: 1A7X). FKBP contains a single, buried tryptophan residue (in orange) located at the base of the binding pocket. (b) Wild type FKBP (~10 µM) showed only trace levels of modification (expect m/z of 11920 for unmodified protein), while (c) lin-FKBP fusion (~90 µM) (expect m/z of 13031 for unmodified protein) and (d) mel-FKBP (~30 µM) (expect m/z of 13131 for unmodified protein) were modified when treated with 20 mM 1 and 100 µM Rh₂(OAc)₄ in 75 mM tBuNHOH at pH 6.0 (2% tBuOH v/v). The reaction was run for 15 h at RT and analyzed by MALDI-TOF-MS. The expected mass increase for each modification is 306. In both c and d, the total level of modification was estimated to be in excess of 40%

Equation