Methods to Study the Structure and Catalytic Activity of cis-Splicing Inteins

Abstract

The autocatalytic process of protein splicing is facilitated by an intein, which interrupts flanking polypeptides called exteins. The mechanism of protein splicing has been studied by overexpression in E. coli of intein fusion proteins with nonnative exteins. Inteins can be used to generate reactive α-thioesters, as well as proteins with N-terminal Cys residues, to facilitate expressed protein ligation. As such, a more detailed understanding of the function of inteins can have significant impact for biotechnology applications. Here, we provide biochemical methods to study splicing activity and NMR methods to study intein structure and the catalytic mechanism.

1. Introduction

Protein splicing is a posttranslational modification facilitated by an intein [1–3]. Inteins are intervening protein sequences that catalyze their own excision from the flanking polypeptides, or exteins, concomitant to extein ligation. Some inteins are also interrupted by an endonuclease domain that can initiate the homing of intein-containing genes into inteinless alleles. Other inteins are the so-called mini-inteins, in which the homing endonuclease domain is lost and replaced by a short linker. Yet other inteins are split in this region, with N- and C-terminal extein–intein segments separated in the genome, requiring reassociation of the inteins to facilitate splicing in trans. In this chapter we focus on methods to study the mechanisms of cis-splicing inteins.

Most studies of the mechanism of protein splicing in cis have been undertaken with inteins that are overexpressed in E. coli, interrupting nonnative exteins that serve dual roles as solubility enhancers and affinity tags. Usually, a number of native extein residues flanking the intein are retained, as these residues can influence the efficiency of splicing [4–9]. Given that most inteins that are well expressed in E. coli splice efficiently on translation, in vitro study of protein splicing or of intein structures is usually performed either with inteins that have mutations at conserved catalytic residues or with inteins from extremophiles that can be purified as unspliced precursors and induced to splice by a physical or chemical cue, including change in temperature, pH, oxidation state, or salinity. Alternatively, some inteins have been induced to splice on a conditional basis by directed evolution or random mutagenesis [10–15].

The canonical protein splicing mechanism consists of four steps (Fig. 1): (1) an amide-ester (or thioester) rearrangement of the peptide bond that links the N-extein and intein, called an N–O (or N–S) acyl shift; (2) a transesterification that results in transfer of the N-extein from the side chain of the first residue of the intein (usually a Ser or Cys) to the side chain of the first residue of the C-extein (usually a Ser, Thr, or Cys); (3) cyclization of the C-terminal Asn residue of the intein that cleaves the peptide bond between the intein and C-extein; and (4) spontaneous conversion of the ester or thioester bond linking the excised exteins back to a peptide bond [1–3].

Fig. 1.

Schematic representation of the four steps of the protein splicing mechanism

Inteins can be manipulated to create tools for protein biotechnology both in vitro and in vivo [16–18]. In the case of expressed protein ligation (EPL), the reactive α-thioester can be created if steps 2 or 3 of the splicing reaction are blocked by mutation. Induced C-terminal cleavage via Asn cyclization uncoupled from splicing can generate a C-extein with an N-terminal Cys residue, which can serve as the other reactant for EPL in place of a synthetic peptide [16, 19].

Our labs have studied the activity and structure of inteins using assays via SDS-PAGE [20, 21] and NMR spectroscopy [21, 22]. Below, we describe and discuss methods for intein fusion protein expression in E. coli, analysis of splicing activity via SDS-PAGE, and analysis of protein structure and pK_a determination of key catalytic residues by NMR spectroscopy.

2. Materials

All solutions for protein analysis should be prepared using ultra-pure water. All solutions for bacterial media should be sterilized before use.

1. LB Media: 20 g/L Tryptone, 10 g/L Yeast Extract, 10 g/L NaCl, pH 7.5. Sterilize by autoclave.

2. LB Agar: LB media supplemented with 15 g/L agar. Sterilize by autoclave.

3. E. coli NovaBlue and E. coli BL21(DE3).

4. 100 mg/mL carbenicillin or ampicillin. Sterilize by filtration.

5. 1.0 M Isopropyl β-d-1-thiogalactopyranoside (IPTG): 238 mg/mL IPTG in water.

6. 0.25 M Phenylmethylsulfonyl fluoride in isopropanol.

7. Benzonase nuclease.

8. EDTA-free protease inhibitor cocktail.

9. Cobalt- or Nickel-based metal affinity resin.

10. Centrifugal filter unit for protein concentration with appropriate molecular weight cut off.

11. 10 × M9 Media: 30 g/L KH₂PO4, 68 g/L Na₂HPO₄ and 10 g/L NaCl. Sterilize by autoclave.

12. 1 M MgSO₄. Sterilize by autoclave.

13. 1 M CaCl_2. Sterilize by autoclave.

14. 1 g/mL ¹⁵NH₄Cl. Sterilize by filtration.

15. 1 g/mL glucose. Sterilize by filtration.

16. 1 g/mL ¹³C₆-D-glucose. Sterilize by filtration.

17. Protein splicing buffer solution: 20 mM HEPES buffer, pH 7.5, 500 mM NaCl.

18. Precast 4–20% gradient gels with stacking for SDS-PAGE.

19. 3× SDS loading buffer.

20. 1.2 M dithiothreitol (DTT).

21. 20 mM tris(2-carboxyethyl) phosphine (TCEP).

22. 50 mM EDTA in protein splicing buffer solution.

23. 6% aqueous acetic acid.

24. Coomassie Blue staining solution.

25. High-resolution flatbed scanner.

26. Protein buffer solution for NMR analysis: 20 mM sodium phosphate, pH 7.1, 5 mM TCEP.

27. Plasmids: pET21-b(+), pMal-c2X, pMal-c5X, pETM-44.

28. Bruker Advance II 800 MHz or Bruker Advance II 600 MHz spectrometer, equipped with a triple-resonance cryogenic probe.

29. A pH electrode for small volume NMR samples.

30. French Pressure Cell Press with appropriate cell.

3. Methods

3.1. Construction of Expression Vectors for Protein Expression in E. coli

We utilize one of two strategies for the design of E. coli expression vectors. In order to study protein splicing via SDS-PAGE, it is useful to have N- and C-terminal exteins that are soluble, can facilitate affinity chromatography, can be used to distinguish spliced products from precursor inteins via SDS-PAGE, and can be detected by commercially available antibodies. We use E. coli maltose binding protein (MBP) as the N-extein and a polypeptide with a C-terminal His-tag as the C-extein. Alternatively, for applications where we plan to study the structure or folding of the intein, we use expression vectors that encode for a very short segment of the N-extein in place of the MBP.

1. In order to generate an expression vector with N-terminal MBP, insert the intein gene in-frame following the gene for E. coli maltose binding protein (MBP) and a short linker sequence that also encodes for a protease-recognition site. (See Note 1 for possible plasmid vectors to use.) Fuse the 3′-end of the intein gene in-frame to a sequence that codes for another affinity tag, such as a poly-His tag or the C-terminal domain of I-TevI.

2. To generate an expression vector without the MBP, either delete the MBP segment so that expression is still under control of the Tac promoter in the pMal plasmid, or transfer of the intein gene into an expression vector such as pET-21b(+), so that expression is under the control of the T7 promoter (see Note 2 and Note 3).

3.2. Protein Overexpression and Purification of Unlabeled Intein Fusion Protein

Intein fusion proteins can be expressed in E. coli. Many inteins splice in E. coli, and their activity can be analyzed by comparing the extent of splicing or splicing side reactions of wild type inteins to those modified by mutation to key active site residues. Other inteins, such as those described here, can be purified from E. coli as unspliced precursors and induced to splice after purification.

1. Transform the expression plasmid into E. coli BL21(DE3) on an LB agar plate with appropriate antibiotic selection (100 μg/mL carbenicillin or ampicillin) and incubate overnight at 37 °C (see Note 4).

2. Pick a single colony and incubate overnight in 2 mL of LB media at 37 °C with shaking.

3. Inoculate 1 mL of overnight culture into 100 mL of LB media and grow with shaking at 37 °C until OD₆₀₀ reaches about 0.6.

4. Induce expression with 1.0 mM Isopropyl β-d-1-thiogalactopyranoside (IPTG) and incubate with shaking at 20 °C for 16 h (see Note 5).

5. Harvest cells by centrifugation at 3000 × g for 10 min. Decant supernatant and either store cell pellets at −20 °C or process immediately.

6. Resuspend cells in 2.5 mL of protein splicing buffer solution supplemented with 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 10 units of benzonase nuclease, and EDTA-free protease inhibitor cocktail.

7. Disrupt cells via passage through a French Pressure cell at approximately 18,000 psi (with 900 psi gauge pressure using a 3/8 inch diameter piston) (see Note 6).

8. Centrifuge clarified lysate at 12,000 × g for 30 min at 4 °C.

9. Purify protein via C-terminal His-tag using immobilized metal affinity chromatography. Add 0.7 mL of a 50% slurry of the resin to a 15 mL centrifuge tube, and centrifuge at 500 × g to settle resin. Decant liquid and resuspend resin with 5 mL cold protein splicing buffer solution, repeating twice. Add clarified cell lysate to resin and incubate with rotation for 30 min at 4 °C. Centrifuge at 500 × g for 5 min, decant liquid, resuspend resin in 5 mL of cold protein splicing buffer solution, and add to a small chromatography column at 4 °C. Wash one time with 10 column volumes (CV) of protein splicing buffer solution supplemented with 10 mM imidazole and 0.1% Tween-20, two times with 10 CV of protein splicing buffer solution supplemented only with 10 mM imidazole, and elute with three 500 μL fractions of protein splicing buffer solution supplemented with 200 mM imidazole (see Notes 7 and Notes 8).

10. In order to remove the high concentration of imidazole from the protein solution, use a centrifugal filter with appropriate molecular weight cut off. First add the elution to the filter, centrifuge at 14,000 × g for 10 min, add protein splicing buffer solution to 500 μL, centrifuge and repeat six times. Recover the final buffer-exchanged protein by inverting the filter into a1.5 mL Eppendorf tube and centrifuging at 14,000 × g for 2 min.

11. Determine concentration of protein in exchanged fraction using a Bradford assay. Using the reagent brought to room temperature, add 20 μL of protein sample to 980 μL of reagent, measure absorbance at 595 nm, and use a standard curve to estimate protein concentration.

3.3. Standard Method for Observing Splicing Activity

For intein fusion proteins that can be purified as unspliced precursors, we can induce splicing or protein splicing side reactions by incubation at conditions similar to those native to the host organism. For inteins with active site Cys residues, this includes incubation with a reducing agent such as DTT or TCEP.

1. Create a reaction cocktail of 16 μL total volume on ice. Prepare stock solutions of EDTA (50 mM) and TCEP (20 mM) in protein splicing buffer solution. Create a master mix such that 16 μL aliquots of the mix have 3 μg of protein, and are 5 mM EDTA and 2 mM TCEP, all in protein splicing buffer solution. Transfer 16 μL aliquots into thin-walled 0.5 mL PCR tubes, and incubate the reaction for the appropriate temperature or time for the intein under study (see Notes 9–11).

2. Analyze the extent of protein splicing via SDS-PAGE, using precast 4–20% gradient gels with stacking, which provide good separation in the molecular weight region of interest. For each sample, add 8 μL of 3× SDS loading buffer, 1 μL of 1.2 M DTT, heat for 3 min at 100 °C, and load. Include a sample lane that contains the same reaction cocktail, but stored at 4 °C as a negative control for splicing.

3. Stain the SDS-PAGE gel in Coomassie Blue overnight, and destain in 6% acetic acid (see Note 12) [23].

4. Scan the gel with a high resolution flat-bed scanner. Protein splicing converts the precursor protein (MBP-Intein-C-extein) into MBP-C-extein and Intein (see Note 13). See Fig. 2 for an example of a SDS-PAGE analysis of an intein from Pyrococcus abyssi. Densitometry of the intensity of the bands can be used as a measure of the extent of protein splicing. Using the free Java-based program ImageJ [24], estimate the relative intensity of each band, and correct for the M_r of each protein. Given that protein splicing results in the conversion of the unspliced precursor (MIH) to the spliced product (MH) and excised intein (I), one can estimate the percentage of splicing using the ratio of $\frac{(\text{MH})}{(\text{MH})+(\text{MIH})}$.

Fig. 2.

In vitro assay of protein splicing activity. (a) Schematic representation of the process of protein splicing, in which the precursor protein (MIH in panel B) is converted to the spliced product and excised intein (MH and I in panel B, respectively.) Peptide bond cleavage uncoupled from splicing can result in the side reactions of N-terminal or C-terminal cleavage. For the Pyrococcus abyssi PolII intein, the fusion protein contains M [residues 27–392 of E. coli MBP, a 23 residue Asn-rich linker, and the 7 C-terminal residues of the N-extein (H-A-A-K-R-R-N)], I [the 185 residue intein], and H [the first six residues of the C-extein (C-D-G-D-E-D) and a His-6 tag] [56]. (b) SDS-PAGE analysis of the in vitro splicing of the Pyrococcus abyssi PolII intein. Lanes 1 and 3 contain purified, unspliced intein fusion protein, whereas lanes 2 and 4 are the result of incubation for 18 h at 70 °C. The intein has a native C-terminal Gln (PolWT); splicing efficiency was compared to the intein with conversion to the more highly conserved Asn (PolQN). (This research was originally published in The Journal of Biological Chemistry. Mills, K.V.; Manning, J.S.; Garcia, A.M., and Wuerdeman, L.A. Protein Splicing of a Pyrococcus abyssi Intein with a C-terminal Glutamine. J. Biol. Chem. 2004, 279 (20), 20,685–20,691. © The American Society for Biochemistry and Molecular Biology)

3.4. NMR Analysis of Intein Structure and Dynamics

In order to gain structural and mechanistic insight into protein splicing, NMR spectroscopy has been used to study the structure and dynamics of a hyperthermophilic intein. The ionization states of active site residues make crucial contributions to the atom-level mechanism of the catalysis, with highly elevated or depressed pK_a values observed in many enzymes [25]. We used a pH titration coupled with NMR measurement to determine the pK_a of catalytic intein residues in protein splicing. The methods presented are based on our studies of Mtu RecA intein [22, 26], Pab PolII intein [21] and Hint domain of Hedgehog protein [27] and can be generally applicable to other inteins and other proteins.

3.4.1. Overexpression of Isotopically Labeled Proteins for NMR Analysis

For NMR analysis, clone the intein of interest into a pETM-44 vector, which expresses a fusion protein with an N-terminal His-tag fused to MBP and the intein. There is a linker sequence (TPGSLEVLKQGPM) between MBP and the intein. The sample data shown are from the Pab PolII intein (Pyrococcus abyssi DNA polymerase II intein).

1. Transform the pETM-44 plasmid with the intein gene into E. coli BL21(DE3) on an LB agar plate with appropriate antibiotic selection (100 μg/mL carbenicillin or ampicillin) and incubate overnight at 37 °C.

2. Pick single colony and inoculate overnight in LB media at 37 °C with shaking.

3. Prepare 10× M9 salt stock solution, 1 M MgSO₄, and 1 M CaCl₂ stock solutions. All stock solutions should be prepared in water and sterilized before use.

4. Dilute 10× M9 salt 1:10 with sterilized water in a 1 L flask. Add 2 mL of MgSO₄ stock solution and 0.1 mL of CaCl₂ stock solution.

5. Add ampicillin or carbenicillin to a final concentration of 100 μg/mL.

6. Add isotopes. For ¹⁵N labeled samples, add 1 mL 1 g/mL of ¹⁵NH₄Cl (final concentration is 1 g/L) and 4 mL 1 g/mL of glucose (final concentration is 4 g/L). For ¹⁵N and ¹³C doubly labeled samples, add 1 mL 1 g/mL of ¹⁵NH₄Cl (final concentration is 1 g/L) and 4 mL 1 g/mL of ¹³C₆-d-glucose (final concentration is 4 g/L) (see Note 14).

7. Swirl flask until all components are mixed well.

8. Resuspend the pellets obtained from an overnight culture into M9 medium to obtain a starting OD₆₀₀ of 0.1.

9. Incubate the M9 cultures at 37 °C until OD₆₀₀ reaches 0.3–0.4 and induce with 1 mM IPTG at 20 °C for an additional 16 h.

3.4.2. NMR Resonance Assignment and Structure Determination of Intein

To solve the solution structure of the intein, carry out the assignment of NMR resonance and structure determination of intein using a series of 2D and 3D NMR spectra (see Notes 15 and 16).

1. As described in Subheading 3.4.1, overexpress and purify ¹⁵N labeled and ¹⁵N,¹³C doubly labeled intein samples.

2. Record a 2D ¹H–¹⁵N HSQC spectrum using ¹⁵N labeled intein.

3. Record a 2D ¹H–¹³C HSQC spectrum using ¹³C labeled intein.

4. To assign the backbone atoms (H_N, N_H, C_α, C_β, and C′), record 3D HN(CA)CO, 3D HNCO, 3D HNCACB, and 3D HN(CO)CACB spectra using ¹⁵N,¹³C doubly labeled intein.

5. To assign the sidechain atoms (H_α, H_β, H_γ, H_δ, C_γ, C_δ) of the intein, record a 3D ¹⁵N-TOCSY (τ_m = 55 ms) spectrum using ¹⁵N labeled intein, and 3D (H)CC(CO)NH-TOCSY, 3D H (CC)(CO)NH-TOCSY, and 3D HCCH-TOCSY (τ_m = 15 ms) spectra using ¹⁵N,¹³C doubly labeled intein.

6. To determine the structure of the intein, record a 3D ¹⁵N-NOESY spectrum in 90% H2O, and two 3D ¹³C-NOESY spectra in 100% D2O and 90% H2O, respectively.

7. Generate peak lists from 3D ¹⁵N-NOESY and 3D ¹³C-NOESY spectra. Use the peak lists, along with chemical shifts from the resonance assignment, as input for a program such as CYANA3.0 [28, 29], which can generate automatic NOE assignments and 3D structure through iterative assignment and structural calculation (see Note 17).

8. Derive angle restraints from TALOS+ [30, 31] prediction based on chemical shifts or from three-bond J-coupling using the Karplus relationship [32, 33] (see Note 18).

9. Subject the structure from CYANA with the lowest target function values to refinement with residual dipolar coupling in explicit water in Xplor-NIH [34].

10. Assess the quality of the final structures with PSVS (see Note 19) [35].

    See Table 1 for the statistics of experimental NMR data for the structure calculation.

    See Fig. 3 for the ensemble of 20 energy-minimized conformers of Pab PolII intein and the ribbon presentation of the structure closest to average NMR structure.

Table 1.

The experimental NMR data for the structure calculation and the structural statistics of the 20 energy-minimized conformers of Pab PolII intein

(a) NMR constraints	Value
Distance constraints	3231
Intra-residue [i = j]	774
Sequential [| i — j| = 1]	843
Medium range [1 < | i — j| < 5]	345
Long range [|i — j|1≥5]	1269
Dihedral-angle constraints	223
Φ	110
Ψ	113
Hydrogen bonds	51
Residual dipolar couplings	90
Total number of restricting constraints	3595
(b) RMSD from ideal geometry
Bond lengths, Å	0.016
Bond angle, °	0.9
(c) RMSD to mean coordinates, Å
All backbone atoms excluding disordered loop residues 121–146	0.46 ± 0.10
All heavy atoms excluding disordered loop residues 121–146	0.93 ± 0.12
(d) Ramachandran plot statistica
Residues in most favored regions	82.0%
Residues in additionally allowed regions	15.4%
Residues in generally allowed regions	2.6%
Residues in disallowed regions	0.0%
(e) Structure quality factors	Raw score	Z-score of Pab PolII intein structureb	Z-score of pdb_nmrc
Verify3D	0.43	−0.48	−2.15
Prosall	0.67	0.08	−0.62
Procheck (phi-psi)	−0.67	−0.32	−2.55
Procheck (all dihedrals)	−0.63	−3.73	−5.08
MolProbityClashscore	54.5	−7.82	−10.74

This research was originally published in The Journal of Biological Chemistry. Du, Z.; Liu, J.; Albracht, C. D.; Hsu, A.; Chen, W.; Marieni, M. D.; Colelli, K. M.; Williams, J. E.; Reitter, J. N.; Mills, K. V.; and Wang, C. Structural and Mutational Studies of a Hyperthermophilic Intein from DNA Polymerase II of Pyrococcus Abyssi. J. Biol. Chem. 2011, 286 (44), 38,638–38,648. © The American Society for Biochemistry and Molecular Biology

^aResidues for Ramachandran statistics are 5–90, 93–120, 147–158, 161–170, 174–183

^bWith respect to mean and standard deviation for a set of 252 X-ray structures <500 residues, of resolution ≦ 1.80Å, R-factor ≦ 0.25 and R-free ≦ 0.28; a positive value indicates a “better” score

^c183 NMR structures determined by nonstructural genomics consortia groups

Fig. 3.

Solution NMR structure of Pab PolII Intein. (a) Ensemble of 20 conformers representing the Pab PoIII intein structure. (b) Ribbon presentation of the structure closest to the average Pab PoIII intein NMR structure. (This research was originally published in The Journal of Biological Chemistry. Du, Z.; Liu, J.; Albracht, C. D.; Hsu, A.; Chen, W.; Marieni, M. D.; Colelli, K. M.; Williams, J. E.; Reitter, J. N.; Mills, K. V.; and Wang,C. Structural and Mutational Studies of a Hyperthermophilic Intein from DNA Polymerase II of Pyrococcus Abyssi. J. Biol. Chem. 2011, 286 (44), 38,638–38,648. © the American Society for Biochemistry and Molecular Biology)

3.4.3. Using NMR to Determine the pK_a of Key Catalytic Residues

The ionization states of amino acid residues are determined by their pK_a and the pH of the environment. Elevated or depressed pK_a has been found in many catalytic residues of enzymes [25]. One can use solution NMR to measure the pK_a of key catalytic residues in enzymes such as an intein.

Histidine pK_a Determination

1. As described in Subheading 3.4.1, overexpress and purify the protein sample labeled with ¹⁵N.

2. Start with a sample in a buffer at pH 7, and split it in two: one for pH titration in the range from 7 to 10.4, and the other for pH titration from 7 to 3.5. Adjust pH values with 0.1 M HCl or 0.1 M NaOH, and measure with a pH meter equipped with an electrode for small volume NMR samples.

3. Record a long-range ¹H-¹⁵N HMQC spectrum of the intein at each pH point (see Note 20).

4. Plot the observed ¹⁵N chemical shifts of the histidine imidazole ring against pH (see Note 21). See Fig. 4 for the pK_a determination of histidine residues by long range ¹H-¹⁵N HMQC NMR spectra.

5. Fit with nonlinear least-squares regression analysis according to the Henderson–Hasselbalch equation. The Henderson–Hasselbalch equation using a Hill parameter (n_H) to account for nonideality is given by

Fig. 4.

pK_a determination of conserved histidine residues in excised intein. (a) Plot of the ring ¹⁵N chemical shift (N_δ1) versus pH for conserved histidines, B-block H73 and F-block H429. (b) HMQC spectrum for intein precursor with an N-extein at pH 7.1. (Figure reprinted with permission from Du, Z.; Shemella, P. T.; Liu, Y.; McCallum, S. A.; Pereira, B.; Nayak, S. K.; Belfort, G.; Belfort, M.; Wang, C. Highly Conserved Histidine Plays a Dual Catalytic Role in Protein Splicing: A pK_a Shift Mechanism. Journal of the American Chemical Society 2009, 131 (32), 11581–11589. Copyright 2009 American Chemical Society)

$${\delta }_{\text{obs}}=\frac{({\delta }_{\text{HA}}+{\delta }_{{\text{A}}^{-}}\times {10}^{n_{\text{H}\left(\text{pH}-\text{p}{K}_{\text{a}}\right)}})}{(1+{10}^{n_{\text{H}\left(\text{pH}-\text{p}{K}_{\text{a}}\right)}})}$$

where δ_obs, δ_HA, and δ_A− are the chemical shifts for the observed, protonated, and deprotonated species, respectively [36].

Cysteine and Aspartate pK_a Determination

1. As described in Subheading 3.4.1, overexpress and purify the protein sample labeled with ¹³C.

2. pH Titration. Same as in Subheading 3.4.3.1.

3. For cysteine, use a 2D ¹H-¹³C HSQC spectrum to monitor the ¹³C chemical shift change of the β-carbon (¹³C^β) (see Note 22).

4. For aspartate, use a 2D ¹H-¹³C HSQC or 2D HB(CB)CO spectrum to monitor the 13C chemical shift change of the β-carbon (¹³C^β) or side chain carbonyl (¹³C′), respectively (see Note 23).

5. Plot the observed ¹³C chemical shifts (¹³C^β or 13C′) as a function of sample pH value.

6. Fit with nonlinear least-squares regression analysis according to the Henderson–Hasselbalch equations as given above in Subheading 3.4.3.1.

3.4.4. ¹⁵N Relaxation Rates and Analysis of Intein

Both conformational changes and protein dynamics are important in intein catalysis [37;38]. Dynamic properties can be probed by ¹⁵N spin relaxation rates, R₁, R₂, and heteronuclear steady-state NOE experiments (see Note 24). These relaxation parameters are sensitive to motions occurring at the time scale faster the time scale of overall protein tumbling, on the order of picosecond to nanosecond [39].

1. As described in Subheading 3.4.1, overexpress and purify the protein sample labeled with ¹⁵N.

2. In order to measure the ¹⁵N R₁ parameter, set approximately 10 relaxation time points with at least two duplicate time points. The longest delay should cause about a 30% decrease in signal intensity. For example, for the Pab PolII intein, we used the following time points: 10, 100, 200, 300, 400 (×2), 500, 600, 700, 800, and 900 ms.

3. In order to measure the ¹⁵N R₂ parameter, again set about 10 relaxation time points with at least two duplicate time points. The longest delay should cause about a 30% decrease in signal intensity. For example, for the Pab PolII intein, we used the following time points: 2, 16 (×2), 30, 44, 58, 72, 86 (×2), and 100 ms.

4. For ¹H-¹⁵N steady-state heteronuclear NOEs, interleave the proton saturation experiment and no proton saturation experiment at each t1 point (see Note 25).

5. For each R₁ and R₂ experiment, pick peaks in the spectrum with the shortest relaxation time (highest intensities) with Sparky. Use the program CURVEFIT (see Note 26) to fit the NMR data to obtain relaxation rates.

6. For heteronuclear NOE, calculate the ratio of the peak heights for saturated and unsaturated spectra.

7. Analyze NMR relaxation data using a r_NH of 1.02 Å as the mean amide nitrogen-hydrogen bond length, and Δσ = −172 ppm, which is the chemical shift anisotropy of backbone ¹⁵N.

8. Determine the amplitudes and time scales of the internal motions of the protein according to the model-free formalism [40, 41] using the program Fast-Modelfree [42] (see Note 27).

   See Fig. 5 for the ¹⁵N backbone dynamics of intein detected by R₁, R₂, and heteronuclear steady-state NOE experiments, demonstrating the unusual rigidity at the termini of the intein, which also is the location of the active site.

Fig. 5.

¹⁵N backbone dynamics for the Pab PolII intein. (a) Longitudinal relaxation rates R₁ (s⁻¹) with regular secondary structure indicated on the top of the panel. (b) Transverse relaxation rates R₂ (s⁻¹). (c) ¹H-¹⁵N steady-state heteronuclear NOEs. (d) Generalized order parameters. (This research was originally published in The Journal of Biological Chemistry. Du, Z.; Liu, J.; Albracht, C. D.; Hsu, A.; Chen, W.; Marieni, M. D.; Colelli, K. M.; Williams, J. E.; Reitter, J. N.; Mills, K. V.; and Wang, C. Structural and Mutational Studies of a Hyperthermophilic Intein from DNA Polymerase II of Pyrococcus Abyssi. J. Biol. Chem. 2011, 286 (44), 38,638–38,648. © The American Society for Biochemistry and Molecular Biology)

4. Notes

1. Although other vectors are available, we use expression vectors based on the pMal-c2X or pMal-c5X vectors from New England Biolabs because they already include the gene for MBP. Users should consider vectors with antibiotic resistance compatible with their E. coli expression strain, and may wish to consider whether to use a T7 or Tac promoter (see Note 2). In the case of the P. abyssi PolII intein, we included seven and 6 residues of both the native N- and C-exteins, respectively; extein residues flanking the intein can have a significant influence on the folding of the active site and on splicing activity [4, 6, 8].

2. We have found that if the intein fusion protein expressed poorly, which can be an issue particularly for inteins with active homing endonuclease domains or those from extreme halophiles, we can achieve better levels of expression from pET expression vectors.

3. We have found that some inteins with active homing endonuclease domains express either poorly or into inclusion bodies, and that the soluble expression of these inteins can be enhanced with mutation of key conserved homing endonuclease residues to Ala, which is unlikely to affect splicing activity [43, 44].

4. For expression of inteins whose splicing is conditional on reduction of a native disulfide bond, we have used E. coli BL21(DE3) Origami cells [45] to prevent or reduce in vivo splicing. Either carbenicillin or ampicillin is appropriate; carbenicillin may reduce the number of satellite colonies but also is usually more expensive.

5. We generally overexpress at 20 °C in order to promote intein solubility as well as to prevent splicing of intein precursors. We have found that the fraction of unspliced to spliced precursor can be higher at shorter expression times.

6. Alternatively, we use Novagen’s Bugbuster protocol, although we have found that some intein fusion proteins fractionate into inclusion bodies using the detergent protocol and that the French Press technique yields more soluble protein. For the detergent lysis protocol, we add 250 μL of Novagen Bugbuster (10×) to Subheading 3.2, step 6, incubate on ice for 30 min, and skip the French Press step.

7. We find that Cobalt-based affinity resin does not result in co-purification of nonspecific E. coli proteins of similar size to E. coli MBP and is useful for this application. (We utilize Talon metal-affinity resin from Takara Bio and base our protocol, in part, on their instructions.) For purification of His-tag containing proteins lacking E. coli MBP, we often use instead the nickel-based HisLink resin, which has fewer nonspecific binding proteins near the size of the isolated intein. Our lab uses a similar method for either resin. We have also used an amylose resin to purify fusion proteins via their N-terminal MBP.

8. It is important that the pH of stock imidazole solutions is adjusted to pH 7.5, and that no solutions containing EDTA or reducing agents are used in the purification.

9. To study inteins from halophiles, we also prepare salt solutions at high ionic strength in 20 mM HEPES, pH 7.5, to vary the salt concentration as appropriate, usually in a range from 0.5 M to 2.5 M. We usually incubate inteins from halophiles at the growth temperature of their host. For inteins from thermophiles, we initially survey a range of temperatures from 30 °C to 70 °C, as their activities can vary and higher temperatures can result in protein precipitation. We initially screen for activity at an extended time (16 h), and then use increasingly shorter incubations to determine the optimal time to study activity.

10. EDTA serves to chelate any residual divalent cation, as these have been shown to inhibit splicing [46–49]. TCEP is a nonnucleophilic reducing agent, and is important for inteins that have disulfide bonds involving catalytic Cys residues.

11. We find that the reactivity of some intein fusion proteins declines rapidly on storage at 4 °C, indicating that these inteins may be susceptible to oxidative or other damage. When this is the case, we purify the protein and run the reaction on the same day.

12. We have observed two conditions that result in the aberrant migration of bands on SDS-PAGE. Some inteins have disulfide bonds that are particularly resistant to reduction, and a disulfide bond between Cys residues near the termini of the intein results in aberrant migration [13, 50–52]. We also have observed that inteins from halophiles that have particularly low pI values migrate aberrantly on SDS-PAGE [20], which is common for proteins with a highly negative surface charge [53].

13. We also verify the identity of the protein bands on SDS-PAGE via Western blot. When using either commercial antibodies against poly-His tags or direct reagents such as KPL’s HisDetector Nickel-AP, we have found when analyzing inteins from halophiles that the performance of the reagents varies and a trial-and-error technique is required to determine which antibody is appropriate.

14. Isotopic labeling was accomplished by growing cultures in M9 minimal medium containing either 1 g/L of ¹⁵NH₄Cl for uniformly ¹⁵N-labeled samples or 1 g/L of ¹⁵NH₄Cl and 1 g/L of ¹³C₆-D-glucose for uniformly ¹⁵N- and ¹³C-labeled samples.

15. All 2D/3D spectra were acquired at an appropriate temperature on a Bruker Advance II 800 MHz or Bruker Advance II 600 MHz spectrometer, each equipped with a triple-resonance cryogenic probe. For the Pab PolII intein, temperature up to 47 °C can be utilized to take advantage of the enhanced NMR sensitivity at high temperature due to the extreme thermostability of proteins from hyperthermophiles.

16. The ¹H chemical shifts were referenced by 4,4-dimethyl-4-silapentane-1-sulfonic acid and the ¹⁵N and ¹³C chemical shifts were referenced using frequency ratios between ¹⁵N, ¹³C, and ¹H (¹⁵N/¹H = 0.101329118, ¹³C/¹H 0.251449530). All chemical shifts were deposited in the BioMagResBank under accession number 17418 [21].

17. CYANA carries out automatic NOE assignment and uses distance and angular constraints from NMR data to calculate the structure of a protein, using torsion angle dynamics [28, 29].

18. TALOS+ uses chemical shifts of N_H, C′, C^α, C^β, H^α, and H^N for a given residue to predict protein backbone torsion angles [30, 31].

19. PSVS analyzes standard NMR constraint and NMR structures, and provides structure quality scores [35].

20. In our experiment [22], long range ¹H-¹⁵N HMQC spectra [54] were acquired with 2048 × 288 complex data points, spectral widths of 100 ppm in ¹⁵ N and 16 ppm in ¹H and 128 scans. The spectrum center was set as 200.0 ppm for ¹⁵N and 4.76 ppm for ¹H.

21. The assignments of histidine residues are based on the (HB)CB (CGCD)HD spectrum, which connects backbone C_β and H_δ.

22. The assignments of cysteine side chains in our experiment [27] were based on H_β chemical shifts from ¹⁵N-TOCSY, C_β chemical shifts from HNCACB and ¹H-¹³C aliphatic HSQC.

23. ¹⁵N-TOCSY, 1H-¹³C HSQC, and HBCBCO spectra were recorded for the assignment of aspartate side chains.

24. All spectra were recorded at an appropriate temperature on a Bruker SB 600 MHz Spectrometer equipped with a ¹H/²H/¹³C/¹⁵N cryoprobe with z-axis gradients. For Pab PolII intein, temperature up to 47 °C can be utilized to take advantage of the enhanced NMR sensitivity at high temperature due to the extreme thermostability of proteins from hyperthermophiles.

25. The NOE experiment was performed with a recycle delay of 7.5 s to ensure that ¹⁵N magnetization reaches equilibration before each scan starts. A 120° proton pulse at 5 ms delay was applied to achieve proton saturation.

26. CURVEFIT is a program that computes a nonlinear least square fit with an arbitrary number of parameters, using a gradient-expansion algorithm.

27. Fast-Modelfree [42] is a graphic interface for Modelfree [55], a program for rapid automated analysis of NMR relaxation data. The generalized order parameters (S²) can describe the amplitude of the internal motion for individual amide bonds at the picosecond to nanosecond time scale.