The standard cell culture medium (glutamine-supplemented medium) comprised Dulbecco’s Modified Eagle Medium (DMEM, Thermo Fisher) without glucose (Glc), glutamine (Gln), phenol red or sodium pyruvate, supplemented with 10% dialyzed fetal bovine serum (dFBS), 2.5 g/L Glc, and 2 mM Gln. HEK293, HCT116 and RKO cells were grown in 10 cm plates at 37°C, 5% CO₂, 21% O_2, and 85% relative humidity, and were passaged every 3,4 days to avoid contact inhibition and supply new media. When a confluency of at least 70% was reached, cells were washed once with 1x PBS and detached from the plate via trypsinization with TrypLE (GIBCO). Pre-warmed medium was added to cease trypsinization. The volume of medium added was calculated according to the desired splitting ratio and the cells were resuspended before the appropriate fraction of the cell suspension was transferred to a new plate. For cell growth assays and subsequent experiments, cells were harvested at a confluency of 80%–90% before being transferred to new plates at a seeding density of 2×10⁶ cells which prevents contact inhibition.

For the cell growth assays, pre-cultivated cells were seeded on 10 cm plates. The following day, the viable cell count was measured for the 0 hour time point and the cell culture medium was changed to that containing the appropriate condition (Gln: 2 mM; Alanine (Ala), Asparagine (Asn), Aspartate (Asp), Glutamate (Glu): all 1 mM, NH₄ ⁺: 0.8 mM). Cells were passaged once they reached a confluency of at least 60%, upon which the cell count was determined. Media was replaced every 3,4 days to avoid limiting nutrients. Viability and cell number were monitored using the TC20 automated cell counter (Biorad).

Pre-cultivated cells were seeded on 6-well plates at a seeding density of 3×10⁵ cells and 12×10⁵ cells for HCT116 and HEK293 cells, respectively. The following day, the viable cell count was measured for the 0 hour time point and the cell culture medium was changed to that containing the appropriate culture condition (see Section 2.1) treated with either 500 µM Methionine Sulfoximine (MSO, Sigma Aldrich) or, as a negative control, sterile water (H₂O). The viable cell count was determined at 24, 48, 72, and 96 h post-treatment. Media was replaced daily to replenish substrates and remove secreted reaction products.

Cells grown in standard media conditions (not starved for glutamine) were washed with PBS and harvested in 1 ml ice-cold RIPA buffer. Cell lysates of HCT116, RKO and HEK293 were denatured in loading buffer for 5 min at 95°C. 40 µg of proteins were loaded and separated on a 10% SDS gel and run for 1 h at 70 V and 1 h at 120 V. The gel was transferred to a nitrocellulose membrane (0.2 µm, Biorad) at 25 V, 1 A for 30 min (Biorad TransBlot Turbo V1.02). The membrane was blocked for 1.5 h in 5% milk in TBS-T at room temperature and cut below the 70 kDa band. The membranes were incubated with primary antibodies against Vinculin (1:2,000 dilution, Sigma, V9131) and against GLUL (1:1,000 dilution, Thermo Fisher, PA1-46165) in 5% milk in TBS-T over-night at 4°C. After washing the membranes in TBST, the membranes were incubated in the HRP-conjugated secondary antibodies (NEB, 7074S; NEB, 7076S) for 1 h at room temperature. After washing the membranes in TBST and TBS, the membrane was developed using an ECL Western Blotting detection reagent (Amersham, RPN2109) according to the manufacturer’s protocol. The Vilber FX gel system was used to record the luminescence (Vilber Lourmat, France).

Cells were pre-cultivated in Glu + NH₄ ⁺-supplemented medium for at least 3 days prior to stable SIRM analysis. HEK293 and HCT116 cells were able to proliferate in Glu + NH₄ ⁺-supplemented dFBS medium and were therefore pre-cultivated for over one month for cell growth assays before being seeded at a density of 2E+6 cells on 10 cm plates. RKO cells were unable to proliferate in Glu + NH₄ ⁺-supplemented medium and were therefore pre-cultivated in Gln-supplemented dFBS medium and changed to medium containing Glu + NH₄ ⁺ performed 3 days prior to the SIRM experiment.

For SIRM experiments, cells were then labelled for 24 h with ¹³C labelled glutamate and ¹⁵N labelled ammonium and treated in parallel with either 500 µM Methionine Sulfoximine (MSO, Sigma Aldrich) or, as a negative control, sterile water (H₂O). For pSIRM experiments, cells were pre-treated for 6 h with either 1 mM MSO or, as a negative control, sterile water, in fresh Glu + NH₄ ⁺-supplemented dFBS medium. Afterwards, cells were labelled for 15 min, 30 min, 1 h, and 3 h with ¹³C labelled glutamate and ¹⁵N labelled ammonium, with or without 1 mM MSO. In both experiments, SIRM and pSIRM, 1 mM ¹³C labelled glutamate was used. However, for SIRM experiments 96 µM ¹⁵N labelled ammonium were used, whereas for pSIRM 0.8 mM ¹⁵N labelled ammonium were used. Cells were harvested and extracted in a methanol-chloroform-water solution as described elsewhere [DOI: 10.1016/B978-0-12-801329-8.00009-X].

Intracellular amino acids were measured as TBDMS derivatives by high-resolution GC-MS. Dried cellular extracts were mixed with 25 µl MTBSTFA (Sigma) and 25 µl ACN and incubated at constant shaking for 1 h at 80°C. Derivatization was automatized on a TriPlus RSH auto-sampler (Thermo Fisher) and each sample was injected immediately after the derivatization. Samples were injected into a Q Exactive GC Orbitrap system (Thermo Fisher) with a splitof 1:5 (1 µl injection volume) in a temperature-controlled injector (TriPlus RSH auto-sampler, Thermo Fisher) with a baffled glass liner. The initial temperature was 80°C for 15 s, followed by an increase of 7°C/s up to 260°C, which is held for 3 min at the end of the temperature program. Gas chromatographic separation was carried out on a Trace 1,300 GC (Thermo Fisher) equipped with a TG-5SILMS column (30 m length, 250 µm inner diameter, 0.25 µm film thickness (Thermo Fisher). Helium was used as the carrier gas (1.2 ml/min flow rate). Gas chromatography was performed with an initial temperature of 68°C for 2 min, followed by an increase of 5°C/min up to 120°C, followed by an increase of 7°C/min up to 200°C, followed by an increase of 12°C/min up to 320°C which is held for 6 min. The spectra were recorded in a mass range of m/z = 60––600 with resolution at 200 m/z set at 120,000.

The elemental composition of different fragments for glutamine were calculated based on the exact mass and compared with known literature-values. To extract the intensities for the different isotopic masses we constructed a compound library including the mass shifts induced by ¹³C and ¹⁵N. Mass shifts were calculated via a custom R-Script based on the known masses for the fragments and the number of potentially incorporated carbon and nitrogen atoms. Each incorporated ¹³C or ¹⁵N increased the target mass by 1.0033548 or 0.99693689, respectively.

For the SIRM experiment, samples were then processed and peaks were integrated with Tracefinder 5.0 (Thermo Fisher), by importing this target list as a Tracefinder Compound database and extracting the extracted ion chromatograms (EIC) within a 5 ppm window. For the pSIRM experiment, samples were processed and peaks integrated in Xcalibur Quanbrowser (Thermo Fisher), extracting EIC with a mass tolerance of 2.5 ppm. For both, peak integration quality was visually checked and finally all peak areas were exported.

Free nucleotides were measured by direct infusion MS on a Q Exactive HF (Thermo Fisher) coupled to a Triversa Nanomate (Advion) nanoESI ion source. The Triversa Nanomate was operated in negative mode, with 1.5 kV spray voltage and 0.5 psi head gas pressure. The spectra were recorded for a duration of 3 min in a mass range of m/z = 140–850 m/z mass units with resolution at 200 m/z set at 240,000. A target list with 48 compounds was prepared in a similar way as described above. The M-H fragment was further calculated by subtracting 1.00728 from the exact mass of the uncharged molecule. For the extraction of the peak intensities the raw files were first converted to.dta2d files using TOPPAS FileConverter tool (Kohlbacher et al., 2007a; Kohlbacher et al., 2007b; Sturm et al., 2008).

The.dta2d files were then processed with a custom R script. Briefly, zero intensities and TIC intensities were removed from the datafiles as well the first and last five scans as these scans tend to be instable. All masses that fit into a 5 ppm window for each mass in the target list was associated to that specific compound. To extract only the apex the most intense mass per compound and scan was kept. Finally, the median and the standard deviation for all the scans was calculated to obtain a single readout per compound and sample.

Natural abundance correction for both types of experiment was performed using the Accucor package (URL: https://doi.org/10.1021/acs.analchem.7b00396).

Dejure and Royla, tested the effect of glutamine starvation on GEO, HCT116, HEK293, HT29, RKO and SW480 cells and found that all cell lines were not able to proliferate except HEK293 (Dejure et al., 2017).

We investigated as to whether two colon cancer cell lines HCT116 and RKO can can adapt to glutamine depletion and removed glutamine from the medium for several days (Figure 1). Also in this experiment HEK293 cells exhibited glutamine independence, but RKO and HCT116 cells were not able to proliferate. In order to exclude that the glutamine independence of HEK293 cells is not caused by small molecules provided by the fetal bovine serum (FBS) we used dialyzed FBS and repeated the proliferation experiment (Figure 2). Interestingly, in this case also HEK293 cells were not able to proliferate when glutamine was deprived. Thus, glutamine was also essential for HEK293 cells when using dialyzed FBS.

In order to identify which metabolic pathways are efficiently utilised in glutamine-depleted condition, we monitored cell survival and growth upon supplementation with substrates of the glutamine-centric metabolic network. To achieve this, we supplemented: Alanine (Ala), Ala + ammonium (NH₄ ⁺), Asparagine (Asn), Aspartate (Asp), Asp + NH₄ ⁺, Glutamate (Glu), Glu + NH₄ ⁺ in dialyzed FBS medium (Figure 3).

Viable cell count was determined at every passage over the course of 31 days. Cell count data were log2-transformed and graphically represented. Cell doubling time was calculated based on the division of culture duration by delta in log2-transformed cell counts. All three tested cell lines exhibit the highest proliferation rate and lowest doubling time in Gln-supplemented medium (Figure 3). In HEK293 and HCT116 cells the proliferation rate in Glu + NH₄ ⁺- supplemented medium is close to that in Gln-supplemented medium. For HCT116 cells proliferation in Asp + NH₄ ⁺-supplemented media is remarkably high. In HEK293 cells also the addition of glutamate leads to intermediate cell proliferation rates. Contrary, RKO cells can not compensate glutamine withdrawal under any condition. RKO cell viability decreased and cell death occurred, preventing the possibility of obtaining viable cell count data after 5 days onwards. Therefore, the proliferation rate for Ala, Ala + NH₄ ⁺, Asn, Asp- Asp + NH₄ ⁺, Glu, Glu + NH₄ ⁺-supplemented media is 0.

The cell growth assays show that in glutamine-depleted dialyzed FBS conditions, HEK293 and HCT116 cells proliferate best when supplemented with the substrates of GLUL: Glu and NH₄ ⁺(Figure 3). Based on this result, an inhibitor assay was performed to assess the effect of blocking de novo glutamine synthesis. Therefore, cells were treated with MSO, a competitive inhibitor of GLUL. As RKO cells are unable to proliferate in glutamine-depleted conditions, they were not subjected to this assay.

A pilot experiment (data not shown) was performed in HEK293 and HCT116 cells and an inhibitor concentration of 500 µM was found to be effective. The inhibitor-containing medium was refreshed and viable cell count was determined every 24 h over a 96 h time period. A parallel assay was performed using water, the solvent control, instead of MSO. Each measurement was taken from three biological replicates. The mean and standard deviation of the viable cell counts for each time point were graphically represented (Figure 4). Untreated HEK293 and HCT116 cells exhibit similar proliferation rates to those observed in the previous growth assay for all tested different conditions. However, MSO-treated HEK293 and HCT116 cells only show proliferation in Gln-supplemented medium. Showing that the chosen inhibitor concentration is not harmful to cells if glutamine is provided.

We designed a dual tracer stable isotope resolved metabolomics (SIRM) study using ¹³C and ¹⁵N labelled substrates (¹³C glutamate, ¹⁵N ammonium) to determine whether the cells utilise extracellular substrates for de novo glutamine synthesis and subsequent nucleotide biosynthesis. Based on the results of the cell growth assays, Glu + NH₄ ⁺-supplemented medium was chosen to trace nitrogen and carbon incorporation into glutamine. The experiment was performed in three biological replicates.

Glutamine can be synthesised by GLUL-mediated ligation of glutamate and ammonium, with ammonium providing the amido-group. The ligation of ¹³C₅-glutamate and ¹⁵N-ammonium was monitored by GC-MS detection of ¹³C₅-, ¹⁴N₁-, and ¹³C₅- ¹⁵N₁-glutamine isotopologues. Automated peak extraction from GC-MS spectra was performed via Tracefinder (Thermo Fisher) and mean ¹³C and ¹⁵N enrichment calculations were performed using custom R scripts. In HCT116 and HEK293 cells, ¹³C- and ¹⁵N-incorporation into glutamine was detected in several characteristic glutamine-3TBDMS fragments, as indicated by the corresponding ¹³C- and ¹⁵N-induced mass shift of the peaks. Glutamine-3TBDMS fragments comprising a 5C skeleton and 2N atoms underwent a mass shift of approximately 6 Da (m+6) while fragments comprising a 4C skeleton and 2N atoms underwent a mass shift of 5 Da (m+5), corresponding to ¹³C₅-¹⁵N₁ and ¹³C₄- ¹⁵N₁ isotopologues, respectively. The cleanest signal was obtained for the fragment at 431 m/z and therefore this fragment was used for further analysis.

In the pilot experiment HCT116 and HEK293 cells were labelled for 24 h with 13C glutamate and 15N ammonium. In HEK293 cells 51% enrichment of ¹³C and 2% enrichment of ¹⁵N in the glutamine-3TBDMS fragment were monitored. HCT116 cells have almost twice as much ¹³C enrichment at 87% and ¹⁵N enrichment at 22%. In both HEK293 and HCT116 cells MSO treatment abolished ¹³C and ¹⁵N enrichment to 0%. RKO cells do not demonstrate ¹³C or ¹⁵N enrichment in untreated and MSO-treated conditions (Figure 5).

The contribution of newly synthesised glutamine isotopologues to nucleotide biosynthesis was monitored by direct-infusion MS detection of nucleotide isotopologues. Peak extraction from direct infusion-MS spectra was performed manually with XCalibur Qualbrowser (Thermo Fisher). In the de novo purine biosynthesis pathway, glutamine donates two nitrogen atoms to IMP and AMP, and three nitrogen atoms to GMP. HEK293 and HCT116 demonstrate ¹⁵N enrichment in AMP and GMP while RKO cells do not. In HEK293 and HCT116 cells, one ¹⁵N atom (N1) is incorporated into each AMP and GMP. HEK293 cells exhibit 25% enrichment of ¹⁵N₁-AMP and 20% enrichment of ¹⁵N₁-GMP, while HCT116 cells exhibit 35% enrichment of ¹⁵N₁-AMP and 15% enrichment of ¹⁵N₁GMP (Figure 6).

In a second step we analyzed the dynamics of glutamine synthesis in HCT116 and HEK293 cells in a time course manner. The cell lines were incubated for 15 min, 30 min 1 h and 3 h with ¹³C labeled glutamate and ¹⁵N labeled ammonium. Label incorporation in glutamine was analyzed as described above (Figure 7). Interestingly label incorporation in glutamine can be found already after 15 min in HCT116 cells and after 30 min in HEK293 cells. Both cell lines show a fast glutamine synthesis but the kinetics are different. In HCT116 cells the incorporation of ¹⁵N labeled ammonium into glutamine exceeds the formation of carbon and nitrogen labeled glutamine, this argues for faster ammonium import into HCT116 cells compared to HEK293 cells.

Interestingly, we performed a western blot analysis to analyze GLUL protein expression in the three cell lines and found that GLUL protein is present in all cell line even under normal conditions (Supplementary Material). We compared the western blot result with proteomics data (not shown) and could also find specific peptides for GLUL in all cell lines. Thus, the reason for the lacking GLUL activity in RKO cells cannot be explained by missing GLUL protein levels but must be caused by other reasons, e.g., mutations in the GLUL gene or impaired transport of substrates for GLUL reaction.