A Comparison of Derivatives of Alanine and d-Alanine Used in Gas Chromatography–Mass Spectrometry Analysis for Protein Kinetics

Abstract

Stable isotope techniques serve as invaluable tools for kinetic measurements in metabolic research. In particular, deuterated water $\left({\text{D}}_2\text{O}\right)$ administration is increasingly being applied in human health research. For use in protein kinetic studies, this includes measurements on gas chromatography–mass spectrometry (GC–MS) analysis of alanine (ALA) and deuterium-labeled alanines (d-ALAs) coming from ${\text{D}}_2\text{O}$ administration. However, the choice of the derivative of ALA and d-ALAs used in such analyses has not been evaluated thoroughly. Hence, we conducted a comprehensive head-to-head comparison to determine the most effective and reliable derivative. Two derivatization reagents, N,N-dimethylformamide dimethyl acetal (methyl-8 reagent) and N-methyl-N-tert-butyldimethylsilyltrifluoroacetamide (MtBSTFA), were considered as candidates. Using chemical standards and available rodent muscle tissue, both reagents underwent testing, including the standard curve linear regression fit, sensitivity, reproducibility, and, importantly, column effectiveness. Our findings indicate that both reagents were suitable for ALA/d-ALAs analyses. However, the MtBSTFA derivative exhibited a better linear regression fit, higher sensitivity, and greater reproducibility than methyl-8. More importantly, the methyl-8 derivative resulted in severe column damage. In conclusion, our study highlights the MtBSTFA derivative as a preferred choice for ALA and d-ALAs GC–MS analysis, contributing to a reliable and sensitive analytical method for ${\text{D}}_2\text{O}$ administration studies for measurements of in vivo metabolic rates.

Graphical Abstract

INTRODUCTION

Stable isotope techniques are powerful tools for conducting kinetic measurements in metabolic research.¹ Among these techniques, deuterated (²H or D) water $\left({\text{D}}_2\text{O}\right)$ is experiencing a surge in popularity as a stable tracer. ${\text{D}}_2\text{O}$ rapidly equilibrates throughout all body water pools such as blood, intracellular water, and saliva, making it an ideal surrogate precursor for metabolic studies.² Deuterium in body water is endogenously exchanged onto different substrates at stable C–H positions and further incorporated into the end products such as glucose, lipid, DNA, RNA, and proteins during de novo synthesis.^3–5 Thus, ${\text{D}}_2\text{O}$ serves as a versatile tracer capable of elucidating dynamics across diverse metabolic pathways, including the metabolism of protein, fat, glucose, and cholesterol in free-living subjects. Particularly, in protein kinetic analysis, this principle is widely accepted. Briefly, orally administered ${\text{D}}_2\text{O}$ intracellularly exchanges deuterium onto carbon-bound hydrogens of amino acids, resulting in deuterium-labeled amino acids, which subsequently are incorporated into protein. Among these amino acids, alanine (ALA) undergoes a rapid turnover.⁶ ALA possesses four potential sites for deuterium exchange (alpha $\left(\alpha \right)$- and beta $\left(\beta \right)$-hydrogens) via transamination,⁷ which makes detection easier. All deuterium-labeled ALAs (d-ALAs) are assumed to be incorporated evenly into protein synthesis, diluted similarly by endogenous ALA from protein breakdown and distributed equally between the intracellular and extracellular compartments.⁶ In addition, the carbon-bound hydrogens on ALA do not undergo nonspecific exchange post-translationally or during protein hydrolysis.⁸ From this, it can be concluded that ALA is a well-suited amino acid for studies of protein kinetics using ${\text{D}}_2\text{O}$ administration. As part of such studies, the detection and measurement of ALA and/or d-ALAs are central. The common methodological approach is to extract, derivatize, and finally analyze ALA and d-ALAs on a gas chromatograph–mass spectrometer. In our investigations, we identified the choice of the derivative of ALA/d-ALAs as a key issue. Notably, prior evaluation of derivatization reagents has not been done for the measurement of ALA and its isotopologues in the study of protein kinetics. Therefore, we aimed to conduct a comprehensive head-to-head comparison to determine the most effective and reliable derivatization reagent.

N,N-Dimethylformamide dimethyl acetal (methyl-8 reagent) is a derivatization reagent of the functional groups, such as carboxylic and amino groups, used in gas chromatography–mass spectrometry (GC–MS). Normally the derivatization occurs through esterifying the carboxylic group and acylating the amino group.⁹ The adoption of methyl-8 originated from the protocol of Thenot and Horning.⁹ It is applied in ALA and d-ALA analyses as reported in several studies.^8,10–13 It reacts both with carboxylic and amino groups of amino acids to form N-dimethylaminomethylene methyl ester derivatives with a high yield, resulting in a mixture that is ready for subsequent GC–MS analysis.⁹

N-Methyl-N-tert-butyldimethylsilyltrifluoroacetamide (MtBSTFA) is a well-known derivatization reagent and widely used in various amino acid analyses, although less so in ALA and d-ALAs analyses with only a few reports.^14,15 MtBSTFA is a common derivatization reagent for different alcohols, thiols, phenols, carboxylic acids, amines, and amides via silylation reactions. Silylation is the most prevalent derivatization technique. In a silylation reaction, a trimethylsilyl group replaces a labile hydrogen from alcohols, thiols, phenols, carboxylic acids, amines, or amides through a nucleophilic attack (SN₂) mechanism and provides a great yield. The products are generally more volatile and less reactive than the original analyte, thus improving its chromatographic behavior. In contrast to acylation, silylation normally does not require a purification step, and the derivatives can be injected directly into the gas chromatograph–mass spectrometer.¹⁶

In this study, our objective was to assess the suitability of methyl-8 and MtBSTFA derivatives of ALA and d-ALAs for GC–MS analyses for protein kinetics, focusing on various parameters such as the standard curve linear regression fit, sensitivity, reproducibility, and, importantly, column lifetime.

MATERIALS AND METHODS

Chemicals and Materials

Standard unlabeled l-ALA (≥99%, later called ALA) was purchased from Sigma (St. Louis, MO). Standard one deuterium labeled at the $\alpha$ position l-ALA (2-d₁, 96–98%, later called d₁-ALA) and four deuterium labeled at the $\alpha$ and $\beta$ positions dl-ALA (2,3,3,3-d₄, 98%, later called d₄-ALA) were purchased from Cambridge Isotope Laboratories Inc. (Teuksbury, MA). The derivative reagent MtBSTFA with 1% tert-butyldimethylchlorosilane (t-BDMCS) (grade: silylation reagent for GC; purity: 97%) was purchased from Regis (Morton Grove, IL), and the derivative reagent methyl-8 (grade: derivatization for GC; purity: >95%) was purchased from Sigma (St. Louis, MO). Solvents (methanol and acetonitrile (ACN)) were purchased from Thermo Fisher (Waltham, MA). The solid-phase extraction (SPE) columns (Strata X-C 33um Polymeric Strong Cation 60 mg/3 mL, Phenomenex 8B–S029_UBJ) for amino acid purification were purchased from Phenomenex (Torrance, CA).

Derivatization

Methyl-8.

Standards/samples underwent reactions with methyl-8, methanol, and ACN at a 3:2:1 (vol/vol/vol) ratio at 75 °C for 15 min. Thereafter, the reaction mixture was ready for GC–MS analysis.¹¹

MtBSTFA.

Standards/samples underwent reactions with MtBSTFA + 1% t-BDMCS and ACN at a 1:1 (v/v) ratio at 100 °C for 30 min before the reaction mixture was ready for GC–MS analysis.

GC–MS Analysis

Methyl-8.

An Agilent 7890A GC system equipped with a 5975C mass spectrometer and a DB17-MS capillary column $(30\text{m}\times 0.25\text{mm}\times 0.25\mu \text{m})$ were used. An amount of $0.5\mu \text{L}$ of derivatized standard (STD) containing about 0.5 nmol of total derivatized ALA (ALA + d₁-ALA) or $2\mu \text{L}$ of the derivatized sample was injected in splitless mode through a $900\mu \text{L}$ liner. Helium was used as the carrier gas at a flow rate of 1 mL/min. The injector temperature was set to 265 °C. The employed temperature program was as follows: an initial temperature of 90 °C was held for 5 min, followed by an increase of 5 °C/min until reaching 130 °C, and then a rapid increase by 40 °C/min to 270 °C, where it was held for another 5 min. The ALA methyl-8 derivative eluted at approximately 11.5 min. The mass spectrometer was operated in the electron ionization (EI) mode at 70 eV. The target ion was determined by using full-scan mode, while the target ion was measured by using selected ion monitoring (SIM) mode.¹¹

MtBSTFA.

An Agilent 7890A GC system equipped with a 5975C mass spectrometer and a DB17-MS capillary column $(30\text{m}\times 0.25\text{mm}\times 0.25\mu \text{m})$ were used. An amount of $0.5\mu \text{L}$ of derivatized STD containing about 0.5 nmol of total derivatized ALA (ALA + d₁-ALA) or $1\mu \text{L}$ of the derivatized sample was injected at a split ratio of 1:1 through a $900\mu \text{L}$ liner. Helium was used as the carrier gas at a flow rate of 1 mL/min. The injector temperature was set to 265 °C. The employed temperature program was as follows: an initial temperature of 90 °C was held for 3 min, followed by an increase of 5 °C/min until reaching 130 °C and then a rapid increase by 40 °C/min to 285 °C, where it was held for another 5 min. The ALA MtBSTFA derivative eluted at approximately 12.3 min. The mass spectrometer was operated in the EI mode at 70 eV. The target ion was determined by using full-scan mode, while the target ion was measured by using SIM mode.

Experimental Design

Tests Using Standards.

A set of enrichment standards was used. The enrichment standard was d₁-ALA having one deuterium at the alpha $\left(\alpha \right)$ carbon position. The percentage enrichments included 0.5, 0.25, 0.125, and 0.0625% d₁-ALA (1 mM) in ALA (1 mM). The enrichment number, position, and range of the standards were determined based on the observation that deuterium-enriched body water up to one deuterium at 0.5% was usually reached and maintained in humans by the common deuterium water $\left({\text{D}}_2\text{O}\right)$ administration method.⁸ Unlabeled ALA (1 mM) was used for the background control as an isotope natural abundance. Each standard $\left(100\mu \text{L}\right)$ was dried with speed VAC and was thereafter derivatized with $100\mu \text{L}$ of the derivatization reagent mixture, upon which $0.5\mu \text{L}$ of the derivatized mixture containing about 0.5 nmol of total derivatized ALA (ALA + d₁-ALA) was injected into the gas chromatograph–mass spectrometer to analyze unlabeled ALA and one $\alpha$-deuterium-labeled ALA (d₁-ALA). The injection amount was determined by testing different volumes to ensure it remained within the instrument’s limitations.

Tests Using Biological Samples.

A set of mouse tibial muscle tissues was utilized for the study. The tissue was collected in a study approved by the University of Arkansas for Medical Sciences’ Institutional Animal Care and Use Committee. One group of mice $(n=6)$ received an i.p. dosage of $20\mu \text{L}$ of ${\text{D}}_2\text{O}/\text{g}$ body mass followed by 5% ${\text{D}}_2\text{O}$-enriched drinking water for 10–14 days prior to sacrifice. Meanwhile, another group of mice $(n=4)$ did not receive ${\text{D}}_2\text{O}$ supplementation, serving as a background control for natural abundance levels of tissue deuterium enrichment. For each mouse, about 40 mg of tibialis muscle tissue was used and processed through a series of 5-sulfosalicylic acid extractions and centrifugations to obtain separately intracellular free (IF) amino acids and proteins (about 10 mg). Thereafter, proteins were further hydrolyzed using 6 N hydrochloric acid (HCl) to amino acids, here called protein-bound (PB) amino acids. The processed samples (consisting of either the entire resulting IF or the equivalent of 1 mg of PB hydrolysate) went through the acidified strong cation SPE column purification for amino acids and then were dried by speed VAC and were ultimately derivatized with $100\mu \text{L}$ of the derivatization reagent mixture. The derivatized mixtures ($2\mu \text{L}$ for methyl-8; $1\mu \text{L}$ for MtBSTFA) were then subjected to GC–MS analysis to measure unlabeled ALA and all possible labeled ALAs (one to four deuteriums (d₁ to d₄) at any position), which were the metabolic results from ${\text{D}}_2\text{O}$ administration. The injection amount was determined by testing different volumes to ensure it remained within the instrument’s limitations. The sample processing followed standard procedures commonly employed in protein kinetics analysis.^14,15,17

Statistical Analysis

Statistical analyses were conducted using GraphPad Prism V5, with values presented as means ± standard deviation (SD). Linear regression was used to analyze standard curves. The relative standard deviation (%RSD) was used to compare the reproducibility and sensitivity of the data between methyl-8 and MtBSTFA derivatives. Student’s $t$ test was used to compare the differences between groups with the significance set to $p<0.05$.

RESULTS

Tests Using Standards

All enrichment standards were made from one deuterium-labeled ALA. Therefore, in the final MS measurement, the mass-to-charge ratio $(m/z)$, the unlabeled ion $m/z$ was termed ${\text{M}}_0$, while one labeled ion $m/z$ was termed ${\text{M}}_1$. The enrichment ratio would be $M_1/{\text{M}}_0$. Furthermore, considering the natural abundances of all isotopes in the tested target ion and the effect from adjacent detected ions, the skewed abundance correction and the overlapping spectra correction could be made for the raw enrichment ratio using the following formula¹⁷

$${\text{M}}_{\text{n}}/{\text{M}}_0=\left[{\left({\text{M}}_{\text{n}}/{\text{M}}_0\right)}_{\text{sample}}-{\left({\text{M}}_{\text{n}}/{\text{M}}_0\right)}_{\text{background}}\right.\left.-d\left({\text{M}}_{n-1}/{\text{M}}_0,{\text{M}}_{\text{n}}/{\text{M}}_0\right)\times {\text{M}}_{n-1}/{\text{M}}_0\right]\times {(1-A)}^n$$

where $(1-A{)}^n$ is the skew correction factor, while $A$ is the natural abundance of the tracer. $d\left({\text{M}}_{n-1}/{\text{M}}_0,{\text{M}}_n/{\text{M}}_0\right)$ is the relative contribution of the tracer with the base ion ${\text{M}}_{n-1}$ to the tracer with the base ion ${\text{M}}_n$. In this case, $A$ is the natural abundance of deuterium; the background is unlabeled ALA.¹⁷

Methyl-8.

The target ion was determined based on Gasier et al.¹¹ Briefly, by analyzing the MS spectrum with the possible fragmented ions of methyl-8 ALA displayed in Figure 1A, the ion with $m/z=158$ represents the molecular ion; the ion with $m/z=143$ includes only the $\alpha$-hydrogen of ALA, while $m/z=99$ covers one $\alpha$- and three $\beta$-hydrogens of ALA. Therefore, the ion $m/z$ at 99 was selected as the target ion $\left({\text{M}}_0\right)$, and then the ion $m/z$ at 100 $\left({\text{M}}_1\right)$ was selected as one deuterium-labeled ALA (d₁-ALA). A sample GC chromatogram is illustrated in Figure 2A. The enrichment ratio $\left({\text{M}}_1/{\text{M}}_0\right)$ was calculated from the peak area ratio at $m/z$ 100 and $m/z$ 99 and was further corrected by the skewed correction factor and overlapping effector.

Figure 1.

Full-scan EI mass spectrum with tentative identifications of in-source fragment ions. (A) ALA methyl-8 derivative and (B) ALA MtBSTFA derivative.

Figure 2.

Typical gas chromatography of the derivatized standard ALA. (A) Methyl-8 ALA $m/z$ 99.1 (black) and d₁-ALA $m/z$ 100.1 (blue). (B) MtBSTFA ALA $m/z$ 232.1 (black) and d₁-ALA $m/z$ 233.1 (blue).

A serial dilution of d₁-ALA standards was analyzed. Within the tested d₁-ALA percentage range, the corrected enrichment ratio to percentage enrichment showed a linear correlation with $y=0.0085x+0.00052$ and $R^2=0.9337$ (Figure 3A). For the precision assessment, the relative standard deviation (%RSD) was calculated for each enrichment percentage. The %RSD was 0.9195, 0.1189, 0.0613, 0.1831, and 0.2474 for 0, 0.0625, 0.125, 0.25, and 0.5% d₁-ALA, respectively (detailed data can be found in the Supporting Information).

Figure 3.

Standard (STD) curve correlation between the enrichment ratio and the standard percentage of enrichment. (A) Methyl-8 derivative: a linear correlation equation $y=0.0085x+0.00052$ and $R^2=0.9337$. (B) MtBSTFA derivative: a linear correlation equation $y=0.0086x-0.000059$ with $R^2=0.9861$.

MtBSTFA.

To determine the target ion, the MS spectrum was acquired, as shown in Figure 1B. Among the possible fragmented ions, the ions at $m/z=158.1,232.2$, and 260.1 were detected. These ions were also reported in previous publications.^17,18 Further, the formation of these ions may be explained as mentioned by Taylor¹⁹ that during EI processes, the excess energy might induce further fragmentation or rearrangement to result in these fragment ions. By analyzing the GC chromatography and MS spectrum with the possible detected fragmentation pattern, the ion with $m/z=232.1$ and retention time around 12.3 min showed great abundance with good chromatography properties and a better signal-to-noise ratio and importantly includes one $\alpha$- and three $\beta$-hydrogens of ALA. Thus, this ion with $m/z=232.1$ was selected as the target ion as ${\text{M}}_0$, while the $m/z=233.1$ was for the MtBSTFA derivative of one deuterium-labeled ALA (d₁-ALA) as ${\text{M}}_1$. A typical chromatography is illustrated in Figure 2B. The enrichment ratio $({\text{M}}_1/{\text{M}}_0)$ was calculated from the peak area at $m/z$ 233.1 over that at $m/z$ 232.1 and was further corrected by the skewed correction factor and overlapping effector.

A serial dilution of d₁-ALA standards was analyzed, and the corrected enrichment ratio was plotted against the percentage enrichment range. The results within the tested d₁-ALA percentage range showed a linear relationship with $y=0.0086x-0.000059$ and $R^2=0.9861$ (Figure 3B). For precision assessment, the relative standard deviation (%RSD) was calculated for each enrichment percentage. The %RSD was 0.1363, 0.0172, 0.0459, 0.1053, and 0.0576 for 0, 0.0625, 0.125, 0.25, and 0.5% d₁-ALA, respectively (detailed data can be found in the Supporting Information).

Tests Using Biological Samples

The enrichment ratio is a very useful measurement for kinetic studies in metabolic research because the enrichment and tracer positions are normally unknown in biological samples. Therefore, the enrichment ratios of ALA were measured in mixed muscle protein as PB ALA and in muscle intracellular fluid as IF ALA, using groups with and without ${\text{D}}_2\text{O}$ administration. With ${\text{D}}_2\text{O}$ administration, there are 4 possible sites of hydrogens available for exchanging with deuterium. Therefore, the ion without deuterium was defined as ${\text{M}}_0$ ($m/\text{z}$ at $M$); the ion with deuterium on (i) any one site was defined as ${\text{M}}_1$ ($m/z$ at M + 1), (ii) any two sites as ${\text{M}}_2$ ($m/z$ at M + 2), (iii) any three sites as ${\text{M}}_3$ ($m/z$ at M + 3), and (iv) all four sites as ${\text{M}}_4$ ($m/z$ at M + 4), whereas their enrichment ratio would be ${\text{M}}_1/{\text{M}}_0$, ${\text{M}}_2/{\text{M}}_0$, ${\text{M}}_3/{\text{M}}_0$, and ${\text{M}}_4/{\text{M}}_0$. Without ${\text{D}}_2\text{O}$ administration, the enrichment ratios ${\text{M}}_1/{\text{M}}_0$, ${\text{M}}_2/{\text{M}}_0$, ${\text{M}}_3/{\text{M}}_0$, and ${\text{M}}_4/{\text{M}}_0$ are displaying the natural abundances of the various isotopes in each ion with $m/z$ at M, M + 1, M + 2, M + 3, or M + 4.

Methyl-8.

With the methyl-8 derivative, in both ${\text{D}}_2\text{O}$-untreated and -treated mouse muscle tissues, the ions ${\text{M}}_0$, ${\text{M}}_1$, and ${\text{M}}_2$ were detectable, while ${\text{M}}_3$ and ${\text{M}}_4$ were absent; within the detection range, ${\text{D}}_2\text{O}$-treated tissues had greater enrichment ratios than untreated ones, as expected (Table 1). An example of chromatography is presented in Figure 4A. In the group without ${\text{D}}_2\text{O}$, muscle PB and IF ALA ${\text{M}}_1$ enrichment ratios were similar and close to unlabeled ALA (Table 2A). In the group supplemented with ${\text{D}}_2\text{O}$, both muscle PB and IF ALA ${\text{M}}_1$ enrichment ratios were significantly higher than in the group without ${\text{D}}_2\text{O}$ $(p<0.0001)$. Furthermore, the IF ALA ${\text{M}}_1$ enrichment ratio (0.1635 ± 1.867 × 10⁻²) was significantly higher than PB (0.09934 ± 3.693 × 10⁻³) $(p<0.0001)$ (Figure 5A). For reproducibility, the average %RSD of repeated GC injections was 1.14 for ${\text{M}}_1$ enrichment ratios and up to 5.46 for ${\text{M}}_2$ enrichment ratios; the average %RSD of replicated mice data was 8.06 for ${\text{M}}_1$ enrichment ratios and even reached to 19.77 for ${\text{M}}_2$ enrichment ratios (detailed data can be found in the Supporting Information).

Table 1.

Deuterium Enrichment Ratios of ${\mathbf{D}}_2\mathbf{O}$-Untreated or -Treated Mouse Muscle Protein-Bound (PB) and Intracellular Free (IF) ALA Using Either Methyl-8 or MtBSTFA as the Derivatization Reagent^a,b

	${\mathbf{M}}_2/{\mathbf{M}}_0$		${\mathbf{M}}_3/{\mathbf{M}}_0$		${\mathbf{M}}_4/{\mathbf{M}}_0$
${\text{D}}_2\text{O}$-untreated	PB	IF	PB	IF	PB	IF
methyl-8	0.0018 ± 0.00038	0.0007 ± 0.00028	ND	ND	ND	ND
MtBSTFA	0.0907 ± 0.00017	0.0908 ± 0.00012	0.0163 ± 0.00056	0.0168 ± 0.00211	0.0030 ± 0.00012	0.0032 ± 0.00051
${\text{D}}_2\text{O}$-treated	PB	IF	PB	IF	PB	IF
methyl-8	0.0045 ± 0.00056	0.0082 ± 0.00036	ND	ND	ND	ND
MtBSTFA	0.0976 ± 0.00089	0.1140 ± 0.00096	0.0190 ± 0.00068	0.0280 ± 0.00216	0.0036 ± 0.00018	0.0055 ± 0.00048

^a${\text{M}}_2$ denotes two deuterium-labeled ALA; ${\text{M}}_3$ denotes three deuterium-labeled ALA; ${\text{M}}_4$ denotes four deuterium-labeled ALA.

^bND = not detectable.

Figure 4.

Example of gas chromatography of derivatized ALA in a mouse muscle sample. (A) Methyl-8 ALA $m/z$ 99.1 (black), d₁-ALA $m/z$ 100.1 (blue), d₂-ALA $m/z$ 101.1 (green), d₃-ALA $m/z$ 102.1 (pink), and d₄-ALA $m/z$ 103.1 (red). The top image is on full scale. The zoomed-in image (bottom) only shows d₁-ALA (blue) and barely d₂-ALA (green) peaks and no peak for d₃-ALA (pink) and d₄-ALA (red). (B) MtBSTFA ALA $m/z$ 232.1 (black), d₁-ALA $m/z$ 233.1 (blue), d₂-ALA $m/z$ 234.1 (green), d₃-ALA $m/z$ 235.1 (pink), and d₄-ALA $m/z$ 236.1 (red). The top image is on full scale. In the zoomed-in image (bottom), all four peaks are clearly displayed (blue d₁-ALA, green d₂-ALA, pink d₃-ALA, red d₄-ALA).

Table 2.

Enrichment Ratios ${\mathbf{M}}_1/{\mathbf{M}}_0$ in Standard Unlabeled ALA (STD) and in Muscle Protein-Bound (PB) and Intracellular Free (IF) ALA from Mice without (w/o) ${\mathbf{D}}_2\mathbf{O}$ Supplementation^a,b,c

A. methyl-8 derivative
${\mathbf{M}}_1/{\mathbf{M}}_0$ ratio	mean	SD	% error (PB/IF vs unlabeled)
unlabeled STD	0.06402	1.08 × 10−3
PB w/o ${\text{D}}_2\text{O}$	0.06509	2.09 × 10−3	1.67
IF w/o ${\text{D}}_2\text{O}$	0.07277	1.01 × 10−2	13.67
B. MtBSTFA derivative
${\mathbf{M}}_1/{\mathbf{M}}_0$ ratio	mean	SD	% error (PB/IF vs unlabeled)
unlabeled STD	0.2231	2.89 × 10−4
PB w/o ${\text{D}}_2\text{O}$	0.2211	2.82 × 10−4	0.90
IF w/o ${\text{D}}_2\text{O}$	0.2219	2.24 × 10−4	0.54

^a% Error = (Measured Value – Referred Value)/Referred Value × 100%.

^bReferred value is the unlabeled STD mean.

^cMeasured value is PB/IF w/o the ${\text{D}}_2\text{O}$ mean.

Figure 5.

Enrichment ratio ${\text{M}}_1/{\text{M}}_0$ in muscle protein-bound (PB) and intracellular free (IF) ALA in mice with (w)/without (w/o) ${\text{D}}_2\text{O}$. (A) Using methyl-8 derivatization: enrichment ratio ${\text{M}}_1/{\text{M}}_0$ in the group w/o ${\text{D}}_2\text{O}$ PB: mean 0.065 with SD 2.09 × 10⁻³ and %RSD 3.22; in the group w ${\text{D}}_2\text{O}$ PB: mean 0.099 with SD 3.70 × 10⁻³ and %RSD 3.72; in the group w/o ${\text{D}}_2\text{O}$ IF: mean 0.073 with SD 1.01 × 10⁻² and %RSD 13.87; in the group w ${\text{D}}_2\text{O}$ IF: mean 0.16 with SD 1.87 × 10⁻² and %RSD 11.42. (B) Using MtBSTFA derivatization: the enrichment ratio ${\text{M}}_1/{\text{M}}_0$ in the group w/o ${\text{D}}_2\text{O}$ PB: mean 0.2211 with SD 2.82 × 10⁻⁴ and %RSD 0.13; in the group w ${\text{D}}_2\text{O}$ PB: mean 0.2470 with SD 3.26 × 10⁻³ and %RSD 1.32; in the group w/o ${\text{D}}_2\text{O}$ IF: mean 0.2219 with SD 2.24 × 10⁻⁴ and %RSD 0.10; in the group w ${\text{D}}_2\text{O}$ IF: mean 0.3123 with SD 2.22 × 10⁻³ and %RSD 0.71.

MtBSTFA.

With the MtBSTFA derivative, in both ${\text{D}}_2\text{O}$-untreated and -treated mouse muscle tissues, all ions including ${\text{M}}_0$, ${\text{M}}_1$, ${\text{M}}_2$, ${\text{M}}_3$, and ${\text{M}}_4$ were detected; all detected ions had higher enrichment ratios in ${\text{D}}_2\text{O}$-treated tissues compared to the untreated tissues, as expected (Table 1). A chromatography example is presented in Figure 4B. In the group without ${\text{D}}_2\text{O}$, the enrichment ratio $\left({\text{M}}_1/{\text{M}}_0\right)$ was similar between muscle PB and IF ALA and the unlabeled standard (Table 2B). In the group with ${\text{D}}_2\text{O}$, the enrichment ratio $\left({\text{M}}_1/{\text{M}}_0\right)$ in muscle PB and IF ALA was significantly higher than those in mice without ${\text{D}}_2\text{O}$ $(p<0.0001)$ and also in the unlabeled standard $(p<0.0001)$. Furthermore, the IF ALA enrichment ratio $\left({\text{M}}_1/{\text{M}}_0\right)$ (0.3123 ± 2.221 × 10⁻³) was significantly higher than in PB (0.2470 ± 3.258 × 10⁻³) $(p<0.0001)$ (Figure 5B). For reproducibility, the average %RSD of repeated GC injections was 0.075 for ${\text{M}}_1$ enrichment ratios and 0.094 for ${\text{M}}_2$ enrichment ratios; the average %RSD of replicated mice data was 0.56 for ${\text{M}}_1$ enrichment ratios and similarly 0.52 for ${\text{M}}_2$ enrichment ratios (detailed data can be found in the Supporting Information). In addition, by using a commercially available d₄-ALA standard, we validated the ${\text{M}}_4$ peak with $m/z=236.1$ in the biological sample by spiking a standard into the sample (data not shown).

Impact on the Column Lifetime

Methyl-8.

The utilization of the methyl-8 derivative resulted in severe column damage with a dramatically reduced column lifetime when compared to the use of the MtBSTFA derivative. A new DB17-MS column lasted less than 100 injections before it stopped producing chromatography.

MtBSTFA.

With the MtBSTFA derivative, the column successfully completed about 600 injections for d-ALA analysis and still provided excellent chromatography. Of note, the MtBSTFA derivative has been used to analyze phenylalanine, tyrosine, urea, leucine, isoleucine, and 3-methyl histidine for thousands of injections in our lab, yet the column continues to perform impeccably.

DISCUSSION

In this methodological study, we aimed to determine the most effective and reliable derivative for analyzing ALA/d-ALAs using GC–MS. We conducted a comprehensive head-to-head comparison of two different derivatization reagent methods. One is a popular GC–MS method with the methyl-8 derivatization reagent, frequently used for alanine analysis using stable isotopes. It was published by Previs et al.^8,10–13 based on Thenot and Horning⁹. The other method that we tested is a GC–MS method with the MtBSTFA derivatization reagent, commonly used for other amino acids.¹⁷ It has also been used to determine protein kinetics after ${\text{D}}_2\text{O}$ administration,^14,15 but only briefly described in the papers and with a variation in GC–MS methods. Thus, we set up the GC–MS method as described in the Materials and Methods section.

First, this study showed some similarities between methyl-8 and MtBSTFA derivatives used for GC–MS analyses of ALA and d-ALAs, as part of the ${\text{D}}_2\text{O}$ administration method. The current experimental design corresponds to settings commonly used in protein kinetic studies with ${\text{D}}_2\text{O}$ administration. (1) The ${\text{D}}_2\text{O}$ administration in humans usually leads to a ratio of d₁-ALA over ALA in the body water of approximately 0.5%,^8,13,14 which corresponds to the standard concentration range in the current study. Around this percentage, our results showed that the corrected enrichment ratio $\left({\text{M}}_1/{\text{M}}_0\right)$ was linearly proportional to the enrichment percentage (Figure 3), indicating that these two derivatization reagents are valid for analytical analyses. (2) Similar enrichment ratios $\left({\text{M}}_1/{\text{M}}_0\right)$ in muscle tissue protein-bound (PB) ALA and muscle intracellular free (IF) ALA were found in the group that did not receive ${\text{D}}_2\text{O}$ vs the unlabeled standard (Table 2), which closely aligned with the calculated theoretical values (6.455% for the ALA methyl-8 derivative C₅H₁₁N₂ and 21.697% for the ALA MtBSTFA derivative C₁₀H₂₆ONSi₂),¹⁷ thus demonstrating no detectable isotope enrichment in those samples. On the other hand, as expected, in muscle tissue from the group with ${\text{D}}_2\text{O}$, the enrichment ratios $\left({\text{M}}_1/{\text{M}}_0\right)$ in both muscle tissue PB ALA and muscle IF ALA were higher than the unlabeled standard. Moreover, it is accepted that the administered isotope equilibrates faster with tissue intracellular fluid than tissue protein as it takes a longer time for incorporation into protein. In line with this, the enrichment ratio $\left({\text{M}}_1/{\text{M}}_0\right)$ of muscle IF ALA was higher than that of muscle tissue PB ALA (Figure 5). (3) Quality peaks were seen in GC–MS chromatography of both derivatives with standards (Figure 2).

Second, this study also showed significantly different performances between the methyl-8 and MtBSTFA derivatives. (1) Although both derivatives showed a linear standard curve, the coefficient of determination $\left(R^2=0.9861\right)$ for MtBSTFA was considerably higher than that for methyl-8 $\left(R^2=0.9337\right)$. In other words, the unexplained variation $\left(1-R^2\right)$ for the methyl-8 derivative was 0.0663, which was almost 5 times higher than that for the MtBSTFA derivative (0.0139) (Figure 2). This result indicates that the standard curve of the methyl-8 derivative has a poorer fit with linear regression. (2) Examining isotope-labeled ALA in biological samples from both ${\text{D}}_2\text{O}$-untreated and—treated mice, the methyl-8 derivative failed to detect multilabeled ALA, while the MtBSTFA derivative successfully led to the detection of all possible labeled ALA (Table 1 and Figure 4), which was confirmed by observing the greater enrichment ratios in ${\text{D}}_2\text{O}$-treated muscle tissues than untreated tissue and the increased ${\text{M}}_4/{\text{M}}_0$ enrichment ratio in the MtBSTFA-derivatized sample when spiked with standard d₄-ALA. This indicates that the MtBSTFA derivative is more sensitive than methyl-8. (3) Although the ${\text{M}}_1/{\text{M}}_0$ enrichment ratios of both derivatives produced the expected results with the ${\text{M}}_1/{\text{M}}_0$ ratio of the sample without ${\text{D}}_2\text{O}$ being similar to unlabeled ALA and lower than the sample with ${\text{D}}_2\text{O}$ and the ${\text{M}}_1/{\text{M}}_0$ ratio of the sample IF with ${\text{D}}_2\text{O}$ being higher than the sample PB with ${\text{D}}_2\text{O}$, as listed in Table 2, the mean ${\text{M}}_1/{\text{M}}_0$ enrichment ratios of the sample PB and IF without ${\text{D}}_2\text{O}$ obtained using MtBSTFA derivatives were much closer to that of unlabeled STD (shown with a low % error <1%) than using methyl-8 derivatives (shown with a high % error up to 13.67%). This indicates that MtBSTFA derivatives provided more accurate data. On the other hand, as shown in Figure 5, the enrichment ratios $\left({\text{M}}_1/{\text{M}}_0\right)$ for MtBSTFA derivatives (SD down to 10⁻⁴; %RSD low to 0.10) had much lower SDs and %RSDs than for methyl-8 (SD up to 10⁻²; %RSD high to 13.87). This indicates that the MtBSTFA derivative data had less variation and more precise values, again meaning that it better differentiates enrichment from nonenrichment (with ${\text{D}}_2\text{O}$ PB/IF vs without ${\text{D}}_2\text{O}$ PB/IF MtBSTFA, the $p$ value ranged from 10⁻⁶ to 10⁻⁹, while with methyl-8, from 10⁻⁵ to 10⁻⁸) and better differentiates various enrichments across a range in biological samples (with ${\text{D}}_2\text{O}$ PB vs with ${\text{D}}_2\text{O}$ IF MtBSTFA, $p=1.99\times {10}^{-12}$, while with methyl-8, $p=2.94\times {10}^{-04}$). It is known that both accuracy and precision of data are very important for any study. (4) Furthermore, for repeated GC injections, the MtBSTFA derivative produced a much smaller SD than the methyl-8 derivative. For example, as listed in Table 2, unlabeled STD repeated injections exhibited an SD of 1.08 × 10⁻⁰³ for methyl-8, whereas it was 2.89 × 10⁻⁰⁴ for MtBSTFA, indicating a notable exponential difference between the two compounds. Data from replicated mice showed the same evidence: the enrichment ratios $\left({\text{M}}_1/{\text{M}}_0\right)$ in all PB or IF groups obtained by the MtBSTFA derivative exhibited SDs in the range of 10⁻⁴, whereas those from the methyl-8 derivative exhibited SDs of 10⁻³ and even 10⁻² (refer to Table 2 and Figure 5 for PB or IF in various mice/groups). Similarly, %RSD, as the coefficient of variation, was used, and it showed a much smaller average %RSD of each STD enrichment percentage with MtBSTFA derivatives (0.072) than methyl-8 (0.31) while also showing significantly smaller average %RSDs for repeated biological sample GC injection data obtained with MtBSTFA derivatives compared to methyl-8. The repeated injection data showed that MtBSTFA %RSD values were much less than 1 (0.075 for ${\text{M}}_1$ enrichment ratios and 0.094 for ${\text{M}}_2$); while with methyl-8, %RSD values were above 1 (1.14 for ${\text{M}}_1$ enrichment ratios and 5.46 for ${\text{M}}_2$) ($p=8.56\times {10}^{-05}$ for ${\text{M}}_1$ and $p=0.0052$ for ${\text{M}}_2$). Similar results were obtained for replicated mice data. With MtBSTFA, %RSD values were less than 1 (0.56 for ${\text{M}}_1$ enrichment ratios and 0.52 for ${\text{M}}_2$); while with methyl-8, %RSD values were up to 19.77 (8.06 for ${\text{M}}_1$ enrichment ratios and 19.77 for ${\text{M}}_2$) ($p=0.032$ for ${\text{M}}_1$ and $p=0.038$ for ${\text{M}}_2$) (detailed data are given in the Supporting Information). These findings indicate that the MtBSTFA derivative provides more reliable and reproducible data. (5) The methyl-8 derivative chromatographic peak was losing the symmetric property and adding an extra tail, while the MtBSTFA derivative kept the peak integrity. It indicates that the MtBSTFA derivative produces higher-quality chromatographic peaks than the methyl-8 derivative (Figure 4). (6) Of note, the methyl-8 derivative severely degraded the column, thus substantially shortening the column lifetime. To the best of our knowledge, this is the first report of this serious issue. We speculate that the methyl-8 derivative might have a chemical interaction with the column material and degrade the column. Thus, we hypothesize that the poor performance of the methyl-8 derivative might largely be ascribed to column degradation. In addition, since the methyl-8 derivative has lower sensitivity, the data analysis requires a bias integration as developed by Katanik et al.²⁰ for when the enrichment ratio is very low. Taken together, the MtBSTFA derivative has several crucial advantages over the methyl-8 derivative.

There are a few other derivatives of ALA that have been used. One of them is N-methoxycarbonyl methyl ester (MCME)^21,22 developed from Huŝek’s method.²³ This derivative requires extraction using chloroform prior to GC–MS analysis. Another derivative that has been used is 2,3,4,5,6-pentafluorobenzyl bromide,²⁴ which also needs to be extracted using ethyl acetate prior to GC–MS analysis. Again, the MtBSTFA derivative does not require additional extraction, which is a significant advantage.

CONCLUSIONS

The use of the MtBSTFA derivative enhances ALA and d-ALAs GC–MS analyses, offering a reliable and sensitive analytical method for assessing in vivo metabolic rates using the increasingly popular ${\text{D}}_2\text{O}$ administration technique.