An efficient protocol for the complete incorporation of methyl-protonated alanine in perdeuterated protein

Abstract

A strategy for the introduction of (¹H,¹³C-methyl)-alanine into perdeuterated proteins is described. Specific protonation of alanine methyl groups to a level of 95% can be achieved by overexpressing proteins in M9/D₂O based bacterial growth medium supplemented with 800 mg/l of 2-[²H], 3-[¹³C] l-alanine. However, though simple, this approach results in undesired, non-specific background labeling due to isotope scrambling via different amino acid metabolic pathways. Following a careful analysis of known metabolic pathways we found that co-addition of perdeuterated forms of α-ketoisovalerate-d₇, succinate-d₄ and l-isoleucine-d₁₀ with labeled l-alanine, reduces undesired background labeling to <1%. When combined with recently developed methyl TROSY experiments, this methyl-specific labeling protocol permits the acquisition of excellent quality correlation spectra of alanine methyl groups in high molecular weight proteins. Our cost effective strategy offers a significant enhancement in the level of incorporation of methyl-labeled alanine in overexpressed proteins over previously reported methods.

Introduction

Recent progress in NMR spectroscopy of high molecular weight proteins have been strongly connected to the development of new isotopic labeling schemes. High level perdeuteration of proteins combined with optimized experiments has significantly increased the size of proteins for which a NMR study can be realized (Salzmann et al. 2000; Fernandez et al. 2001; Arora et al. 2001; Fiaux et al. 2002; Tugarinov et al. 2005). To date, transverse relaxation optimized spectroscopy (TROSY) has been developed for several atoms in biological molecules including backbone amide groups (Pervushin et al. 1997; Riek et al. 1999), aromatic methines (Pervushin et al. 1998; Brutscher et al. 1998), methylenes (Miclet et al. 2004, 2005) and methyl groups (Tugarinov et al. 2003). Because of their proton multiplicity and fast rotation, methyl groups are characterized by favorable spectroscopic properties. High sensitivity and resolution methyl ¹H-¹³C correlation spectra can be recorded of protein complexes above 100 kDa, provided that these groups are selectively protonated in an otherwise highly perdeuderated protein background (Tugarinov et al. 2003). Methyl groups have been shown to be particularly useful to probe dynamics, interactions and structure properties in large protein assemblies (Gardner and Kay 1998; Tugarinov et al. 2005; Hamel and Dahlquist 2005; Sprangers et al. 2005, 2007; Sprangers and Kay 2007).

Initial protocols for the production of selectively methyl-protonated proteins used ¹H,¹³C-pyruvate as the sole carbon source with growth media prepared in D₂O (Rosen et al. 1996). With this method partially protonated Ala, Leu, Val, Ile (Ile-γ₂ only) methyl groups were obtained with proton incorporation levels ranging from 40% to 80%. However, due to the partial exchange of protons, the final sample contains four different methyl group isotopomers (CH₃, CH₂D, CHD₂ and CD₃), which can significantly complicate subsequent NMR spectra. Later strategies based on the use of selectively protonated α-keto acid precursors in M9/D₂O cultures enabled specific incorporation of ¹³CH₃ groups in selected methyl sites. Using the approach developed in the Kay laboratory, the level of incorporation is very high without detectable scrambling of protons (for review see Tugarinov et al. 2006). Protocols using 2-keto, 3,3-[²H₂], 4-[¹³C]-butyrate as the sole proton source produce proteins protonated at isoleucine-δ₁ methyl sites only (Gardner and Kay 1997). Similarly, α-ketoisovalerate with specifically protonated methyl sites can be used to produce proteins with only the methyl groups of leucine and valine protonated (Goto et al. 1999; Tugarinov and Kay 2004). Alanine is another methyl site of great interest as the β-methyl group can report directly on the structure and dynamics of the protein backbone. Alanine is also the most abundant amino acid found in proteins (McCaldon and Argos 1988). Unlike isoleucine, leucine and valine residues, which are all synthesized by irreversible metabolic pathways in Escherichia coli, the synthesis of l-alanine is a single step reversible reaction with pyruvate as a precursor. This reaction can be catalyzed by three different transaminases: alanine-oxo acid amino transferase (EC 2.6.1.12), alanine-glycine transaminase (EC 2.6.1.44) and alanine transaminase (EC 2.6.1.2). As pyruvate is a key metabolite involved in a large number of biosynthetic pathways producing proteins with specific protonation of alanine methyl groups and without significant isotopic scrambling is more challenging. The best way to ensure specific labeling is to supplement labeled alanine directly to the culture medium. Isaacson et al. (2007) recently proposed a strategy to specifically incorporate methyl-protonated alanine using rich perdeuterated medium. In rich medium the high background level of perdeuterated amino acids serves to greatly reduce isotopic scrambling. However, this amino acid reservoir also dilutes the methyl-protonated alanine added to the culture medium and consequently reduces the level of incorporation in the target protein. Here we propose an efficient labeling protocol based on M9 culture medium in D₂O to accomplish a near complete incorporation of [¹³C,¹H₃]-methyl groups in to alanine without detectable scrambling of protons to other sites. This new strategy relies on the introduction of perdeuterated precursors in combination with exogenous labeled alanine to abolish the metabolic scrambling of protons. Particular attention has been paid to the choice of the precursors and carbon sources in order to reduce costs and optimize the level of deuteration.

Materials and methods

Preparation of 2-[²H] l-alanine

2-[²H], 3-[¹³C] l-alanine and 2-[²H], 2,3-[¹³C₂] l-alanine were synthesized using a modification of the procedure described by Goss and Newill for l-halotryptophan synthesis (Goss and Newill 2006). The exchange of the α-proton was achieved using a tryptophan synthase-rich cell lysate prepared from an LB culture of E. coli transformed with pSTB7, a high copy number plasmid that constitutively expresses Salmonella enterica tryptophan synthase (commercially available ATCC 37845; Goss and Newill 2006). For the proton to deuterium exchange, 1 g ¹³C-labeled l-alanine was dissolved in 60 ml of a 50 mM TRIS solution in D₂O at pH 7.6, in presence of 1 mg of pyridoxal phosphate (PLP or vitamin B6). Lyophilized lysate, corresponding to 1 l of E. coli/pSTB7 culture (Goss and Newill 2006), was added and the mixture incubated at 35°C for 24 h. The extent of proton-deuterium exchange was monitored by NMR. After 24 h the solution was lyophilized. The resulting lysate was resuspended in 60 ml D₂O and incubated for a further 24 h to complete the exchange. The reaction was stopped and tryptophan synthase precipitated by heating the solution to 80°C for 30 min. Following centrifugation, the supernatant containing labeled alanine was collected and used without further purification. The level of deuteriation was estimated to be over 98% from the analysis of the α proton NMR signal.

Optimization of the incorporation of alanine in overexpressed ubiquitin

Initial experiments to determine the level of protonated alanine incorporation into ubiquitin were performed using 3-[¹³C] l-alanine (CortecNet) to reduce costs. Escherichia coli BL21(DE3) cells were transformed with a pET41c plasmid carrying the human His-tagged ubiquitin (pET41c-His-Ubi) gene and transformants were grown in 70 ml unlabeled M9/H₂O media containing 1 g/l ¹⁵NH₄Cl, and 5 g/l glycerol (unlabeled) as the sole carbon source. When the optical density (OD) at 600 nm reached 0.8, a 30 ml solution containing 3-[¹³C] l-alanine was added. After an additional 1 h, protein expression was induced by the addition of IPTG to a final concentration of 1 mM. Induction was performed for 3 h at 37°C. Ubiquitin was purified by Ni-NTA (Qiagen) chromatography in a single step.

The optimal quantity of alanine required to achieve near complete incorporation in the overexpressed protein was assessed in a series of cultures (100 ml each) in which different amounts of 3-[¹³C] l-alanine were added 1 h prior induction (final concentration ranging from 0 to 800 mg/l). To quantify the reduction of isotope scrambling, several samples of ubiquitin were produced in the presence of 800 mg/l of 3-[¹³C] l-alanine with the addition of different metabolites: α-ketoisovalerate 200 mg/l; succinate (2.5 g/l); pyruvate (3 g/l) and acetate (3 g/l); and l-isoleucine with concentrations ranging from 0 to 100 mg/l.

Production of U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-Ala-β proteins

Escherichia coli BL21(DE3) carrying the plasmid of the overexpressed protein (ubiquitin or MSG) were progressively adapted, in three stages over 24 h to M9/D₂O media containing 1 g/l ¹⁵ND₄Cl, 2.5 g/l glycerol-d₅ (Cambridge Isotope Laboratories, Inc.) and 1 g/l d-glucose-d₇ (Isotec). In the final culture, the bacteria were grown at 37°C in M9 media prepared with 99.85% D₂O (Euriso-top). When the OD (600 nm) reached 0.8, a solution containing methyl-protonated alanine (prepared with the protocol described above) and different perdeuterated precursors in D₂O was added. After addition the concentrations in culture medium were: 800 mg/l of 2-[²H], 3-[¹³C] l-alanine; 2.5 g/l of succinate-d₄ (Cambridge Isotope Laboratories, Inc.); 200 mg/l for α-ketoisovalerate-d₇ (CDN Isotopes Inc); and 60 mg/l for isoleucine-d₁₀ (Cambridge Isotope Laboratories, Inc.). Protein expression was induced by the addition of IPTG to a final concentration of 1 mM 1 h later. Ubiquitin was purified by Ni-NTA chromatography a single step. Malate Synthase G (MSG) was purified in initially by Chelating Sepharose chromatography (GE Healthcare) followed by gel filtration chromatography (Superdex 200 pg GE Healthcare). All the perdeuterated precursors and carbon sources used in this protocol were chosen based on their availability from standard suppliers at relatively low cost. [¹³C,¹H₃] labeling of the C_β position of alanine in U-[²H], U-[¹⁵N]-enriched proteins using the approach described here increases the isotope cost by ~400 € per liter of culture compared to standard protocols for the production of U-[²H], U-[¹⁵N]-proteins, (estimation based on the price of isotope purchased for this study in 2007 and 2008).

NMR spectroscopy

All spectra of ubiquitin were recorded on a Varian DirectDrive spectrometer operating at a proton frequency of 600 MHz equipped with a cryogenic triple resonance pulsed field gradient probehead. Data were acquired at 25°C in 100% D₂O buffer containing 100 mM sodium phosphate (pH = 6.7, uncorrected). Protein concentrations were within the range of 0.5 to 2.0 mM. The ¹³C enrichment level at the C_β position of alanine was determined by analyzing the multiplet structure of cross-peaks between alanine H_α and H_β protons observed in a 2D-TOCSY experiments recorded of ubiquitin overexpressed in M9/H₂O medium in presence of 3-[¹³C] l-alanine. The intensity of the central line of the multiplet corresponds to the proportion of H_β bound to ¹²C_β, while the two outer-lines (separated by 125 Hz) correspond to the proportion of ¹³C incorporated at the β position of alanines. The level of ¹³C labeling on other methyl, methylene and methine positions was determined by comparing the intensities of the resonances in ¹³C-HSQC spectra acquired in constant time (CT) mode. Signal intensities were corrected to account for different sample concentrations, and each cross peak normalized, as a percentage, against a comparable reference spectrum acquired of U-[¹³C]-ubiquitin.

All spectra of MSG were recorded on a Varian Inova spectrometer operating at a proton frequency of 800 MHz equipped with a cryogenic triple resonance pulsed field gradient probehead. Data were acquired on a 0.8 mM U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-Ala-β MSG sample at 37°C in 100% D₂O buffer containing 20 mM MES (pH = 7.0 uncorrected), 20 mM MgCl₂, 5 mM DTT. The ¹H-¹³C HMQC was recorded with 2576 complex data points in direct dimension (maximum t₂ = 198 ms) and 160 points in carbon dimension (maximum t₁ = 80 ms), with two scans by increment. In order to facilitate assignment of the alanine methyl resonances a 3D HCC experiment was recorded of 1 mM U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C²H]-Ala-α, U-[¹³C¹H₃]-Ala-βMSG sample prepared using the protocol described here. The experiment was collected with 192 points in the C_α dimension (maximum t₁ = 30 ms), 104 points in the C_β dimension (maximum t₂ = 16 ms), and 1820 points in the direct dimension (maximum t₃ = 140 ms) and 8 scans per increment, for a total acquisition time of 2.5 days. All data were processed and analyzed with the nmrPipe/nmrDraw suite of programs (Delaglio et al. 1995).

Results and discussion

Alanine incorporation

The first parameter optimized was the amount of exogenous alanine required to ensure complete incorporation of labeled alanine in proteins overexpressed in E. coli culture. To minimize isotope scrambling, labeled alanine was added to the culture one hour prior to induction. The level of incorporation of alanine into ubiquitin was followed by adding varying amounts of 3-[¹³C] l-alanine (concentrations from 50 to 800 mg/l of M9/H₂O culture media). Figure 1 shows the level of ¹³C incorporation at the C_β position of alanine in ubiquitin as a function of the amount of labeled alanine added to the culture medium. The incorporation level reaches a plateau above 90% for a concentration of exogenous l-alanine in excess of 600-mg/l. This result is in agreement with level of inhibition previously reported in literature (Muchmore et al. 1989).

Fig. 1

Level of incorporation of ¹³C at the C_β-alanine position in overexpressed ubiquitin as a function of the amount of exogeneous 3-[¹³C] l-alanine added. The ¹³C enrichment level at the C_β position of alanine was determined by analyzing the multiplet structure of cross-peaks between alanine H_α and H_β protons observed in 2D-TOCSY experiments

Figure 2A represents an extract of a ¹³C-HSQC spectrum of ubiquitin produced in E. coli in M9/H₂O medium supplemented by 800 mg/l of 3-[¹³C] l-alanine one hour prior to induction. As expected, intense signals for the two alanine residues of ubiquitin are observed, but several additional signals could also be detected for other methyl, methylene and methine sites. Comparison of signal intensities with a ¹³C-HSQC spectrum recorded on a uniformly ¹³C-labeled ubiquitin sample confirms that the C_β atoms of alanine are labeled at more than 95% (data not shown). The level of ¹³C labeling of leucine, valine and isoleucine (Ile-γ₂ only) methyl groups is approximately 25%. The incorporation of ¹³C in all other methyl (Ile-δ₁, Met and Thr), methylene and methine sites ranges from 5% to 15%.

Fig. 2

A 2D ¹H-¹³C HSQC spectrum of ubiquitin overexpressed in presence of ¹³C-alanine. The spectrum was recorded at 25°C in D₂O on a NMR spectrometer operating at a proton frequency of 600 MHz. The first contour level corresponds to 5 % of the maximal intensities of C_β-alanine correlation peaks. The two cross-peaks corresponding to two expected C_β-alanine signals are labeled. The spectrum shows that a significant amount of ¹³C from exogeneous ¹³C-alanine has been incorporated in other amino acids. The scrambling level reaches 25% in methyls of Leucine, Valine and Isoleucine (Ile-γ₂ only) and 5–15% for other methyl, methylene and methines sites. The locations of two selected traces, (a) and (b) presented below the HSQC spectra, are indicated by a dashed line on the 2D spectrum. On the 1D traces the first contour level plotted in the 2D spectrum is represented by a dashed line. The intensities of trace (b) relative to trace (a) has been multiply by a factor 12. B Schematic illustration of the major metabolic pathways involved in amino acid biosynthesis in E. coli. The level of the scrambling of ¹³C from 3-[¹³C] l-alanine is represented by the thickness of the arrows. Three transaminases are responsible for the conversion of ¹³C-alanine into pyruvate. The methyl group of pyruvate is directly utilised for the synthesis of Leucine, Valine and Isoleucine (Ile-γ₂ only), and gives rise to an important (~25%) ¹³C scrambling in these methyl groups. Pyruvate is also incorporated in the TCA cycle and glycolytic pathways. These reactions are responsible of the background ¹³C labeling of most of the amino-acids sites (5–15%)

Reduction of isotopic scrambling

The scrambling of ¹³C observed in Fig. 2A arises from the metabolic interconversion of alanine and pyruvate (Fig. 2B). The simplest way to reduce the scrambling would be to add unlabeled pyruvate to the cell culture at the same time as the exogenous 3-[¹³C] l-alanine. Indeed, addition of 3 g/l of pyruvate with 3 g/l of acetate reduced scrambling to a level lower than 5% for all methine, methylene and methyl sites (Figure S1). However, due to the efficiency of the three transaminases, pyruvate is still converted into alanine. The addition of exogenous unlabeled pyruvate also reduces the level of incorporation of 3-[¹³C] l-alanine to approximately 60% (Figure S1). Therefore, to maintain a high level of incorporation of labeled alanine, the reduction of isotopic scrambling should be achieved without the addition of exogenous pyruvate.

The methyl groups of leucine, valine and isoleucine (Ile-γ₂ only) are directly transferred from pyruvate (Fig. 2B) and therefore these sites are more severely affected by isotopic scrambling. Addition of α-ketoisovalerate, at a concentration higher than 100 mg/l of culture, has been shown to give near complete labeling of leucine and valine methyl groups (Goto et al. 1999). To attempt to reduce isotope scrambling in leucine and valine methyl groups, the medium was supplemented with unlabeled α-ketoisovalerate in addition to ¹³C-alanine. α-ketobutyrate has also been used for specific labeling of the isoleucine-δ₁ methyl group (Gardner and Kay 1997). However, as the Ile-γ₂ methyl group is derived directly from the methyl group of pyruvate, the use of deuterated α-ketobutyrate will not reduce isotope scrambling on Ile-γ₂ methyl sites. To overcome this issue, isoleucine can be added directly to the culture to further reduce isotope scrambling. As illustrated in Figure S2, the addition of unlabeled isoleucine at 60 mg/l reduces the undesired secondary labeling of isoleucine methyl groups to a level lower than 3%.

¹³C labeled atoms of pyruvate are also incorporated in other amino acids (Fig. 2B) via the glycolytic pathway (H, S, G, C, W, F, Y) and the tricarboxylic acid (TCA) cycle (E, N, Q, P, R, D, K, M, T, I). Efficient deuteration of proteins can be realized using perdeuterated glucose, glycerol, acetate (Venters et al. 1995) or succinate (Vanatalu et al. 1993) as the source of carbon. Glucose and glycerol are incorporated in to the glycolytic pathway (Fig. 2B), while acetate (converted in to acyl-CoA) and succinate are directly incorporated in TCA cycle. Perdeuterated forms of the three last compounds are available at very competitive price, compared to glucose-d₇. These three metabolites are thus good candidates for carbon sources to complement glucose in order to reduce background isotope scrambling in methine, methylene and methyl groups of isoleucine (Ile-δ₁), threonine and methionine.

Figure 3A shows a region of an ¹H-¹³C-HSQC spectrum of ubiquitin produced in E. coli in M9/H₂O medium using a combination of 1 g/l of glucose, 2.5 g/l glycerol as sources of carbon. One hour prior to induction, 800 mg/l of 3-[¹³C] l-alanine together with 2.5 g/l of succinate, 60 mg/l of isoleucine and 200 mg/l of α-ketoisovalerate were added to the culture. NMR spectra and extracted 1D traces show near complete incorporation of labeled alanine is achieved with a concomitant reduction in ¹³C scrambling of all other methine, methylene and methyl sites to a level lower than 5%.

Fig. 3

2D ¹H-¹³C HSQC spectra of ubiquitin with methyl-specific alanine labeling. The spectra were recorded at 25°C in D₂O on a NMR spectrometer operating at a proton frequency of 600 MHz. Ubiquitin has been overexpressed in E. coli in presence of labeled alanine at 800 mg/l, 2.5 g/l of succinate, 60 mg/l of isoleucine and 200 mg/l of α-ketoisovalerate. In A U-[¹H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-Ala-β ubiquitin overexpressed in M9/H₂O medium using protonated glucose (1 g/l) glycerol (5 g/l), and protonated precursors. B U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-Ala-β ubiquitin overexpresed in M9/D₂O medium using glucose-d₇, glycerol-d₅ with the addition one hour prior to induction of 2-[²H], 3-[¹³C] l-alanine, together with succinate-d₄, isoleucine-d₁₀ and α-ketoisovalerate-d₇. In both spectra, the first contour level corresponds to a signal equal to 5% of the maximal intensities of C_β-alanine correlation peaks. The two cross-peaks corresponding to C_β-alanine signals are labeled. Two selected traces (a) and (b) located by a dashed line on the 2D spectra, are presented below the HSQC spectra. On each 1D trace the first contour level plotted in the 2D spectrum is represented by a dashed line. The intensities of trace (b) relative to trace (a) has been multiply by a factor 12

Production of U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-β-Ala proteins in M9/D₂O culture medium

Commercially available 3-[¹³C] l-alanine was used to report isotope scrambling in order to develop an efficient protocol to achieve a high level of incorporation of methyl-labeled alanine with reduced recycling into other amino-acids. To ensure that perdeuterated proteins are specifically protonated on the alanine methyl group, the culture should be prepared in 100% D₂O using perdeuterated carbon sources and alanine deuterated at the H_α position. For this purpose the use of fully protonated 3-[¹³C] l-alanine in M9/D₂O growth medium would be inadequate and would result in the production of proteins with a high level of protonation at the alanine H_α position. All the precursors and carbon sources in the optimized protocol described above have been selected not only for their efficiency to reduce isotope scrambling but also for their commercial availability in perdeuterated form at a relatively low cost (see section “Materials and Methods” for details). As 2-[²H], 3-[¹³C] l-alanine is not available commercially, a simple and cost effective protocol was established to specifically exchange the alpha proton of the 3-[¹³C] l-alanine by deuterium, with retention of configuration, using tryptophan synthase (Nibeilliu and Malthouse 2004; Isaacson et al. 2007; see section “Materials and Methods” for further details).

Specific methyl-protonated alanine was used to produce U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-Ala-β ubiquitin. Figure 3B shows an extract of a ¹H-¹³C-HSQC spectrum of ubiquitin produced in M9/D₂O with 2-[²H], 3-[¹³C] l-alanine and perdeuterated forms of all supplemented precursors and carbon sources (a detailed protocol is given in the Materials and Methods section). As it can be seen, complete incorporation of 2-[²H], 3-[¹³C] l-alanine is achieved with negligible labeling at other sites. Quantitative analysis of NMR spectra demonstrated that the level of alanine methyl group labeling is more than 95% without any detectable level ¹³CDH₂ or ¹³CD₂H isotopomers. Isotopic scrambling has been considerably reduced and none of the other methine, methylene or methyl sites give rise to peaks intensities higher than 0.5%. The reduction of background labeling (compare Fig. 3A, B) is primarily due to the fact that in deuterated M9, the only source of protons is the alanine methyl group. In order to be incorporated in other residues, alanine methyl protons would have to be transferred to pyruvate. In vivo the protons of pyruvate spontaneously exchange with deuterium from the D₂O used to prepare the culture medium. Moreover, protonated pyruvate is recycled through the glycolysis pathway or TCA cycle for use in the synthesis of other amino acids. These reactions involve enzyme-catalyzed exchange of protons with deuterium from the D₂O solvent. As a result the few ¹³C-methyl groups derived from exogenous 2-[²H], 3-[¹³C] l-alanine (<5% of all amino acid carbon sites) lose their protons during the process of being incorporated into other amino acids.

Comparison with previously proposed Alanine labeling protocol

Isaacson et al. (2007) proposed a method for specifically protonating the methyl group of alanine in which proteins are overexpressed in perdeutered rich culture medium supplemented with methyl-protonated alanine (50 mg/l). The presence of a mix of perdeuterated amino-acids in the culture medium ensures that the recycling of ¹³CH₃ groups to other residues is considerably reduced. However, rich medium also contains a source of fully deuterated alanine, which would also be incorporated into the target protein. Therefore, the overall result is a protein containing a mixture of alanine labeling with incorporation levels of the desired 3-[¹³C,¹H]-labeled alanine diluted by the background levels of 3-[¹²C,²H]-labeled alanine.

In order to compare the protocol presented here with the method proposed by Isaacson et al. (2007), the same protein (i.e. ubiquitin) was overexpressed using both protocols and analyzed by NMR spectroscopy. Production of perdeuterated ubiqutin using rich perdeuterated culture medium as previously described (Isaacson et al. 2007) substantially enhanced protein expression. Ubiquitin prepared by either method yielded high quality NMR spectra, with narrow lines for the two alanine methyl groups and without noticeable isotope scrambling in other residues. Nevertheless, at identical protein concentration the intensity of the alanine signals in the protein prepared in rich perdeuterated medium is substantially weaker (Fig. 4). This difference of intensity suggests that the incorporation level of methyl-specific labeled alanine when [¹³C]-labeled alanine is used in combination with deuterated Spectra 9 based rich medium is only ~30% (Isaacson et al. 2007). This result is consistent with the low level of incorporation observed in our systematic study where various amounts of ¹³C-alanine were added in M9/H₂O culture medium (Fig. 1). While isotope incorporation levels are dependent on the different growth/expression conditions used, the primary reason for the lower of isotope incorporation is the presence of perdeuterated alanine in the Spectra 9 rich deuterated medium (a ²H-Celtone/D₂O derived culture medium). According to the manufacturer’s analysis Spectra 9 medium contains 60 ± 10 mg of deuterated alanine (for additional information see the manufacturer’s website: www.spectrastableisotopes.com/Catalog/Celtone_Powder_Information.aspx). Furthermore, we have established that when 2-keto, 3-[²H₂], 4-[¹³C]-butyrate precursor (50 mg/l) is added to Spectra 9 perdeuterated rich medium the level of incorporation of isoleucine is reduced to 70% compared to the level achieved using α-keto-acid in M9 medium (Gardner and Kay 1997; Sounier et al. 2007).

Fig. 4

Comparison of 2D ¹H-¹³C HSQC spectra of U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-Ala-β ubiquitin overexpressed in M9/D₂O medium (A) or Spectra-9/D₂O rich medium (B). Spectra have been recorded at 25°C in D₂O on a NMR spectrometer operating at a proton frequency of 600 MHz. Ubiquitin used for spectra A has been produced in E. coli in M9/D₂O medium using glucose-d₇ (1 g/l) and glycerol-d₅ (2.5 g/l) with an addition one hour prior to the induction of 2-[²H], 3-[¹³C] l-alanine at 800 mg/l, together with 2.5 g/l of succinate-d₄, 60 mg/l of isoleucine-d₁₀ and 200 mg/l of α-ketoisovalerate-d₇. Ubiquitin used for spectra B has been produced in Spectra-9/D₂O rich medium with an addition 1 h prior to induction of 2-²H, 3-¹³C l-alanine at 50 mg/l as described by Isaacson et al. (2007). Spectra B has been scaled to take into account the difference in sample concentration used for acquisition of spectra A. The first contour level (of both spectra A and B) corresponds to a signal equal at 5% of the maximal intensities of C_β-alanine correlation peaks detected in spectra A. The two cross-peaks corresponding to C_β-alanine signals are labeled. Selected traces located by a dashed line on the 2D spectra, are presented below the HSQC spectra. On each 1D trace the first contour level plotted in the 2D spectrum is represented by a dashed line. The intensities of trace (b) relative to trace (a) has been multiply by a factor 12

A reduction in the level of isotope incorporation will directly affect the sensitivity of NMR experiments. The active alanine concentration (i.e. the amount of sites with the desired isotope incorporation) in proteins prepared using the method of Issacson et al. is essentially 30% of the total protein concentration. Protein samples with methyl-specific protonation have been shown to be useful for extracting precise long-range inter-methyl group distances (Sounier et al. 2007). The signal of an NOE cross-peak between two protonated alanines methyl groups decreases with the square of the level of incorporation. Therefore the preparation of proteins with methyl-specific protonated alanines prepared using the M9-based method presented here will enhance the sensitivity of NOESY experiments by one order of magnitude compared to the scheme previously proposed (Isaacson et al. 2007).

Application to malate synthase G

One of the major applications of methyl-specific labeling strategy is the study of large proteins and protein complexes. MSG, a monomeric protein of 82 kDa, was produced using the new protocol presented here. Pure MSG specifically ¹H and ¹³C labeled on the methyl group of alanine was produced at a yield of 80 mg/l. A ¹³C-¹H-HMQC spectrum of MSG recorded at 800 MHz is presented in Fig. 5A. The signals of the 73 alanine residues of MSG can be detected with high signal to noise ratio and resolution. The method described in this manuscript is equally applicable for the introduction of 2-3-[¹³C₂] l-alanine or U-[¹³C] l-alanine in perdeuterated proteins. One should note that acquisition of [¹³C,¹H]-HMQC spectra acquired of samples produced using 2-3-[¹³C₂] l-alanine or [¹³C₃] l-alanine requires CT carbon frequency editing or homonuclear carbon-carbon decoupling to maintain the high resolution of the spectra. C_α and C_β resonance assignments available from standard triple resonance experiments acquired on a U-[¹³C,¹⁵N, ²H]-labeled protein sample can be used a starting point for alanine methyl resonance assignment. These data in combination with, for example, HCC spectra recorded of a perdeuterated protein specifically labeled with 2-[²H], 2-3-[¹³C₂] L-alanine would permit residue-specific assignment of alanine methyl [¹³C,¹H] resonances. Introduction of U-[¹³C] l-alanine would also be possible using the protocol outlined here but at significantly higher cost. To facilitate the assignment of alanine methyl resonances of MSG, 2-[²H], 2-3-[¹³C₂] l-alanine was incorporated in to perdeuterated MSG. An HCC experiment was used to correlate ¹H_β and ¹³C_β frequencies to the corresponding ¹³C_α resonance. 2D extracts of this experiment are presented in Fig. 5B. Using this experiment together with C_α and C_β assignments of alanine, previously reported by Tugarinov et al. (2002), it was possible to assign all the alanine methyl resonances observed in HMQC spectra of MSG (Fig. 5A).

Fig. 5

A 2D ¹H-¹³C HMQC spectrum of alanine methyl group in Malate Synthase G (MSG—82 kDa) recorded on a 0.8 mM sample of U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C¹H₃]-Ala-β MSG. An expansion of the central part of the spectrum is displayed in the bottom right corner. B 2D-(F₁,F₃) extracts of 3D HCC experiment correlating C_α (F₁), C_β (F₂) and H_β (F₃) frequencies of MSG alanine residues recorded on a 1 mM sample of U-[²H], U-[¹⁵N], U-[¹²C], U-[¹³C²H]-Ala-α, U-[¹³C¹H₃]-Ala-β MSG. Both spectra have been recorded at 37°C in D₂O on a NMR spectrometer operating at a proton frequency of 800 MHz. All the cross-peaks in spectra A and B have been labeled with the corresponding residue number. Asterisk corresponds to minor form of MSG as previously described by Tugarinov et al. (2002)

Conclusion

A cost effective method for the preparation of methyl-specific ¹³C,¹H-labeled alanine in M9/D₂O media was developed. Simultaneous introduction in the culture medium of specifically labeled alanine with perdeuterated precursors allowed isotope scrambling to be reduced to <0.5%. The protocol presented here will allow incorporation of specifically methyl-protonated alanine at a level higher than 95% in an otherwise perdeuterated protein. Combined with recently developed methyl TROSY approach, this specific methyl-labeling protocol allows the acquisition of high quality correlation spectra of alanine methyl groups in high molecular weight proteins. This new labeling protocol is expected to be particularly useful as the incorporation of protonated alanine methyl groups offers a unique opportunity to probe structural and dynamical properties of protein backbone in large biomolecular assemblies.