A simple and cost-effective protocol for high-yield expression of deuterated and selectively isoleucine/leucine/valine methyl protonated proteins in Escherichia coli grown in shaker flasks

Abstract

A simple and cost-effective protocol is presented for expression of perdeuterated, Ile/Leu/Val ¹H/¹³C methyl protonated proteins from 100 ml cultures in M9 ++ /D₂O medium induced at high (OD₆₀₀ ~ 10) cell density in shaker flasks. This protocol, which is an extension of our previous protocols for expression of ²H/¹⁵N/¹³C and ¹H/¹³C labeled proteins, yields comparable quantities of protein from 100 ml cell culture to those obtained using a conventional 1 L culture with M9/D₂O medium, while using three-fold less α-ketoisovaleric (1,2,3,4-¹³C₄; 3,4′,4′,4′-d₄) and α-ketobutyric (¹³C₄; 3,3-d₂) acid precursors.

Isotope labeling, including uniform ¹⁵N, ¹³C and ²H labeling, is the mainstay of modern protein NMR spectroscopy. In two recent papers we presented a simple, cost-effective protocol for high yield expression of ¹⁵N/¹³C/²H (Cai et al. 2016) and ¹⁵N/¹³C (Cai et al. 2019a) labeled proteins in E.coli grown to high cell density in shaker flasks. These protocols entail the use of much lower volumes of cultures (100 ml) than conventional protocols (typically 1 L), thereby saving on D₂O usage as well as the relevant carbon (¹³2 C-glucose, ¹³C/²H glucose or ¹²C/²H glucose) and nitrogen (¹⁵NH₄Cl) sources. In addition to ¹⁵N, ¹³C and ²H labeling, specific methyl group labeling in a perdeuterated background has permitted studies of large proteins and protein complexes and yielded considerable biological insights (Rosenzweig and Kay 2014; Rossi et al. 2019; Schutz and Sprangers 2020; Tugarinov et al. 2004). One of the most common methyl labeling schemes involves ¹³C/¹H-methyl labeling of Ile/Leu/Val using α-ketoisovaleric (1,2,3,4-¹³C₄; 3,4′,4′,4′-d₄) and α-ketobutyric (¹³C₄; 3,3-d₂) acid precursors (Goto et al. 1999). The cost of such precursors is non-trivial and we therefore thought it would be worthwhile to extend our previous high density, shaker flask protocols to the expression of ¹³C/¹H-methyl Ile/Leu/Val labeled proteins in a perdeuterated background. The procedure described produces the same amount of labeled protein as the standard 1 L expression protocol for a third of the quantity of methyl labeling precursors.

The results reported here make use of the R10Q mutant of the ngMinE protein cloned with a C-terminal His tag in pET11a and expressed in E coli BL21(DE3) cells (Cai et al. 2019b). The growth medium comprises a modified M9 ++ medium described previously (Cai et al. 2019a) and includes a trace amount of Luria-Bertini (LB) medium to facilitate growth in D₂O (Table 1).

Table 1.

Composition of regular M9 and the new M9 ++ medium for expression of U-[¹⁵N/²H]/Ile^δ1-[¹³CH3];Leu/Val-[¹³CH₃] labeled proteins

	Regular M9 medium (1 L)	M9 ++ medium (100 ml)
K2HPO4	10.0 g	1.9 g
KH2PO4	13.0 g	0.5 g
Na2HPO4	9.0 g	0.9 g
K2SO4	2.4 g	0.24 g
D-d7-Glucosea	2.5 g	1.8 g
NH4Cla	1.0 g	0.5 g
Trace element solutionb	10.0 ml	100 μl
MEMc	10.0 ml	–
1 M MgCl4	10.0 ml	100 μl
LB/D2O	–	0.1 ml
α-ketoisovaleric acidd	100 mg	32 mg
α-ketobutyric acidd	50 mg	16 mg

^aFor perdeuteration d₇-D-glucose is the sole carbon source; for perdeuteration and uniform ¹³C-labeling ¹³C₆/d₇-D-glucose would be used instead.¹⁵NH₄Cl is used as the sole nitrogen source

^bThe trace element solution is that described by (Cai et al. 1998) and comprises the following per 100 ml: 0.6 g FeSO₄ (7H₂O), 0.6 g CaCl₂ (2H₂O), 0.12 g MnCl₂ (4H₂O), 0.08 g CoCl₂ (6H₂O), 0.07 g ZnSO₄ (7H₂O), 0.03 g CuCl₂ (2H₂O), 2 mg H₃BO₄, 0.025 g (NH₄)₆Mo₇O₂₄ (4H₂O), 0.5 g ethylenediaminetetraacetic acid (EDTA)

^cMEM vitamins 100 × solution from Corning (VWR catalog number 45000–702) was lyophilized and dissolved back in the same volume of D₂O

^dα-ketoisovaleric acid (1,2,3,4-¹³C₄, 98 atom %; 3,4′,4′,4′-d₄, 98 atom %) and α-ketobutyric (¹³C₄, 98%; 3,3-d₂, 98 atom %) are only added 1 h prior to induction

We first tested the effect of inclusion of trace amounts of LB on deuterium incorporation by expressing ngMinE in 100 ml M9 ++ medium containing 0.5 g ¹⁵NH₄Cl, 1.8 g d₇-D-glucose (98% D), and 0, 1, 2 or 3% (v/v) LB. The cell cultures were allowed to grow to OD₆₀₀ ~ 10 at 37 °C and induced with 1 mM isopropyl β-D-galactopyranoside (IPTG) at 25 °C for 20 h. After purification using a single Ni-affinity column, deuterium incorporation was evaluated by electrospray mass spectrometry using a q-TOF MS/MS mass spectrometer (model Xevo Gs-XS, Waters). The molecular weights obtained for the proteins grown in the presence of 0, 1, 2 and 3% (v/v) LB are 11,736.3, 11,735.7, 11,735.1 and 11,732.4 (with a mass error of ± 0.1), indicating that inclusion of LB even up to 3% (v/v) has only a minimal effect on deuterium incorporation. For our purposes, however, 0.1% (v/v) LB is sufficient.

The optimal concentrations of α-ketoisovaleric acid (1,2,3,4-¹³C₄, 98 atom %; 3,4′,4′,4′-d₄, 98 atom %) and α-ketobutyric (¹³C₄, 98%; 3,3-d₂, 98 atom %) were determined by carrying out a series of 11 trial cultures as summarized in Table 2 and Fig. 1. ngMinE contains 7 isoleucines, 13 leucines and 6 valines. When all 45 methyl groups (both methyl groups of Leu and Val, and the δ-methyl group of Ile) are protonated, the molecular weight should decrease by 135 unit relative to that of the perdeuterated protein. However, since the methyl labeling precursors are also ¹³C-labeled, the expected decrease in molecular weight is ~ 90 units. As shown in Fig. 1a, the molecular weight of purified ngMinE decreases as the amount of added methyl precursors is increased. The difference reaches slightly more than 90 units at 31.3 mg/100 ml and 15.7 mg/100 ml of added α-ketoisovaleric and α-ketobutyric acid, respectively (sample 6). The further small decrease in molecular weight beyond sample 6 is indicative of precursor scrambling to other sites. Examination of the integrated intensity of the ¹H methyl envelope (normalized relative to the amide proton envelope) indicates that saturation of protonated methyl labeling is reached at sample 6 (Figs. 1b and c), and this is confirmed from the evolution of cross-peak intensities in 2D ¹H-¹³C correlation spectra (Fig. 2). On the basis of these results, we therefore determined that the optimal concentrations of α-ketoisovaleric and α-ketobutyric acid were 32 and 16 mg, respectively, per 100 ml cell culture.

Table 2.

Optimization of precursor concentration (mg per 100 ml culture)

	Sample number
	1	2	3	4	5	6	7	8	9	10	11
α-ketoisovaleric (mg/100 ml):	0	5.6	11.2	18.6	24.8	31.3	38.2	45.0	51.7	58.5	67.5
α-ketobutyric acid (mg/100 ml):	0	1.8	5.6	9.3	12.4	15.7	19.1	22.5	25.9	29.3	33.3

α-ketoisovaleric acid, sodium salt (1,2,3,4-¹³C₄, 98%; 3,4′,4′,4′-d₄, 98%) and α-ketobutyric acid, sodium salt (¹³C₄, 98%; 3,3-d₂, 98%) were purchased from Cambridge isotope laboratories (catalog numbers, CDLM-8100-PK and CDLM-4611-PK, respectively). Precursor stock solutions containing 225 μg/μl α-ketoisovaleric acid and 112.5 μg/μl α-ketobutyric acid were prepared. A 150 ml culture in M9 ++/D₂O medium containing 0.75 g ¹⁵NH₄Cl and 2.7 g d₇-D-glucose was grown to an OD₆₀₀ ~ 8 in a 2.8 L flask, as described previously (Cai et al. 2016). Eleven 10 ml cultures were prepared by aliquoting 10 ml of the above culture to eleven 250 ml Corning disposable flasks to which were added 0, 2.5, 5, 8, 11, 14, 17, 20, 23, 26 and 30 μl of the above precursor stock solution. The 11 cell cultures were allowed to continue growing at 37 °C for 1 h (at which point the OD₆₀₀ ~ 10) and induced with 1 mM IPTG for 20 h at 25 °C. Cell pellets from these cultures were resuspended in 10 ml lysis buffer containing 50 mM Tris–HCl, pH 7.5 and 200 mM NaCl. The cells suspensions were sonicated for three minutes in a cold room with the samples immersed in ice water. The supernatant from each cell lysate was loaded onto a 1 ml Ni-affinity column, and washed with 20 mL of buffer A (50 mM Tris–HCl, pH 7.5, 200 mM NaCl and 40 mM imidazole). The ngMinE was eluted with an imidazole gradient from 40 mM (buffer A) to 1 M imidazole (buffer B = buffer A plus 1 M imidazole). The protein fractions were pooled, and further purified by gel filtration on a Superdex75 column in 25 mM potassium phosphate buffer, pH 6.5 and 200 mM NaCl. The fractions containing ngMinE were pooled and concentrated using a centriprep with a 3.5 kDa cutoff

Fig. 1.

Evaluation of Ile^δ1-[¹³CH₃];Leu/Val-[¹³CH₃] incorporation into ngMinE using the M9 ++/D₂O medium from a 100 ml culture in a shaker flask for a series of 11 concentrations of α-ketoisovaleric and α-ketobutyric acid precursors as summarized in Table 2. a Bar plot of molecular masses for samples 1 through 11 determined by electrospray mass spectrometry. The error in the mass determination is 0.1 Da. b Bar plot of the relative ¹H methyl envelope intensity for samples 1 through 11, referenced to sample 11, determined from a 1D ¹H-NMR spectrum recorded with a jump-return water suppression sequence. The data were normalized to the amide proton envelope to account for any small variations in sample concentration. The estimated error in the relative methyl envelope intensities is 5–10%, given any uncertainties in baseline correction. c Examples of 1D spectra for samples 2, 4, 6 and 11. (D) ¹H-¹³C HSQC spectra for samples 2, 4, 6 and 11. NMR spectra were collected at 25 °C on ~ 60 μM ngMinE samples (samples 1–11) in 25 mM potassium phosphate, pH 6.5, 200 mM NaCl (95% H₂O/5% D₂O, v/v) on a Bruker 600 MHz spectrometer equipped with a z-gradient, triple resonance cryoprobe

Fig. 2.

Schematic of the protocol used for high level expression of U-[¹⁵N/²H]/Ile^δ1-[¹³CH3];Leu/Val-[¹³CH₃] proteins in M9 ++ / D₂O medium from a 100 ml culture in a shaker flask. For reference, the standard 1 L cell culture protocol that makes use of regular M9 medium (see Table 1) is as follows: a 1 ml LB starter culture in a 15 ml culture tube, inoculated from a freshly transfected agar plate or glycerol stock, is grown at 37 °C for 5 h with the OD₆₀₀ reaching a value between 1 and 2; 100 μl of the LB starter culture is inoculated into a 25 ml M9/D₂O pre-culture and grown at 37 °C for 20 h; subsequently the 25 ml pre-culture is transferred to a 4 L (or larger) flask containing 1 L M9/D₂O medium and grown to an OD₆₀₀ of 0.7–0.8, after which the α-ketovaleric and α-ketobutyric acid precursors (in the amounts listed in Table 1) are added; the culture is allowed to grow at 37 °C for a further 1 h before induction with IPTG; the cells are then harvested 3 to 4 h after induction. The shaker speed is set to 200 rpm for all cultures

Our final optimized protocol for the expression of U-[¹⁵N/²H]/Ile^δ1-[¹³CH3];Leu/Val-[¹³CH₃] labeled proteins is shown schematically in Fig. 2. The protocol is similar to the protocol described for perdeuterated protein expression (Cai et al. 2016) and comprises five major steps: cell adaptation from H₂O medium to D₂O medium, pre-culturing, cell growth, cell growth with methyl precursors added and IPTG induction. To grow a 100 mL D₂O cell culture, a 1 ml LB/H₂O culture is started from a fresh agar plate or glycerol stock early in the morning. After about 5 h growth at 37 °C, 2.0 ml of LB/D₂O culture is inoculated with 100 μl of the LB/H₂O culture. The cell density of this culture should reach an OD₆₀₀ between 1 and 3 in about 5 h. The overnight pre-culture is initiated by transferring the entire 2.0 ml LB/D₂O culture into a 250 ml baffled flask containing 20 ml M9 ++ / D₂O medium and allowing it to grow at 37 °C for about 15 h. The OD₆₀₀ of this pre-culture should reach a value of ~ 6 early next morning. All 20 ml of the pre-culture are then poured into a 2.8 L baffled bottom flask containing 80 ml M9 ++ /D₂O medium and allowed to grow at 37 °C until the cell density reaches an OD₆₀₀ of ~ 8 (which takes 6 to 8 h). Following addition of 32 mg α -ketoisovaleric acid and 16 mg α-ketobutyric acid to the above 100 ml cell culture, the cell culture is allowed to grow at 37 °C for one more hour. The shaker temperature is then lowered to 25 °C and the cells induced by addition of IPTG (at a final concentration of 1.0 mM). Cells are then harvested by centrifugation 20 h after. In our hands, we obtain sufficient ngMinE protein to provide slightly more than 1 ml of a 1 mM protein solution (i.e. ≥ 11 mg of purified protein) from a 100 ml cell culture. If one needs to grow more than a 100 ml cell culture, the volumes of the preculture and final culture need to be increased proportionally. One should keep in mind, however, that the maximum culture volume is 25 ml in a 250 ml preculture flask and 280 ml in the final 2.8 L culture flask.

The cells in the current shake flask protocol are induced at an OD₆₀₀ of ~ 10, which represents a ten-fold higher cell density compared to a conventional protocol where cells are induced at an OD₆₀₀ of ~ 1 (1 L cell culture, growth at 37 °C to an OD₆₀₀ of ~ 0.7 to 0.8 prior to addition of precursors and induction; see Fig. 2 legend). The cell mass produced using our current protocol from a 100 ml cell culture is the same as that obtained from a 1 L cell culture using the conventional protocol. As shown previously for perdeuterated protein expression (Cai et al. 2016), protein expression levels are the same per unit cell for cells induced at an OD₆₀₀ of 1 or an OD₆₀₀ of 10. The total amount of protein obtained from a 100 ml cell culture using the current protocol is therefore the same as that from a 1 L cell culture using the conventional protocol. At current average prices for isotopes and precursors, the cost of producing a perdeuterated, Ile/Leu/Val ¹³C/¹H-methyl protonated protein is approximately $750 per one liter using the conventional M₉/D₂O growth medium and protocol, compared to about $250 for 100 ml M9 ++ / D₂O growth medium of the current high cell density protocol in a shaker flask, a three-fold saving in cost.

Data availability

All data is freely available from the authors upon request.