Aligning fermentation conditions with non-canonical amino acid addition strategy is essential for Nε-((2-azidoethoxy)carbonyl)-L-lysine uptake and incorporation into the target protein | Scientific Reports

Scientific Reports volume 14, Article number: 25375 (2024) Cite this article Metrics details Protein engineering with non-canonical amino acids (ncAAs) holds great promises for diverse applications,

Scientific Reports volume 14, Article number: 25375 (2024) Cite this article

Metrics details

Protein engineering with non-canonical amino acids (ncAAs) holds great promises for diverse applications, however, there are still limitations in the implementation of this technology at manufacturing scale. The know-how to efficiently produce ncAA-incorporated proteins in a scalable manner is still very limited. In the present study, we incorporated the ncAA N6-[(2-azidoethoxy)carbonyl]-L-lysine (Azk) into an antigen binding fragment (Fab) in Escherichia coli. We used the orthogonal pyrrolysyl-tRNA synthetase/suppressor tRNACUAPyl pair from Methanosarcina mazei to incorporate Azk site-specifically. We characterized Azk uptake and Fab production at bench-scale under different fermentation conditions, varying timing and mode of Azk addition, Azk-to-cell ratio and induction time. Our results indicate that Azk uptake is comparatively efficient in the batch phase. We discovered that the time between Azk uptake and inducing its incorporation into the Fab must be kept short, which suggests that intracellular Azk is consumed and/or degraded. The results obtained in this study are an important step towards the development of efficient production methods for Azk-incorporated proteins in E. coli. The developed process is scalable and provides excellent yields of 2.95 mg functionalized Fab per g CDM, which corresponds to 80% of yield obtained with the wild type Fab. We also identified the cellular uptake of Azk being dependent on the physiological state of the cell as a potential bottleneck in production.

Genetically encoded protein modification through the incorporation of a ncAA has been considered as a promising strategy in both basic and clinical research. ncAAs carry chemically diverse side chain moieties, e.g. bioorthogonal reactive groups, which are absent in the canonical amino acids prescribed by the standard genetic code1. Since the standard genetic code allows only canonical amino acids for ribosomal translation, the admission of ncAAs requires an orthogonal translation system. The orthogonal translation system consists of an aminoacyl-tRNA synthase (aaRS) and its cognate tRNA that neither interacts with tRNAs nor aaRSs of the host. This orthogonality prevents misincorporation of the canonical amino acids into protein of interest1,2. Usually, a stop codon serves as the coding unit for the ncAA. The amber stop codon (TAG in DNA, UAG in mRNA) has been routinely redefined for the incorporation of an ncAA of interest because it is the rarest stop codon in many expression hosts such as E. coli3. The aaRS charges its cognate suppressor tRNACUA with the ncAA, the ncAA-charged tRNACUA recognizes UAG as a sense codon and consequently this process leads to amber suppression and site-specific ncAA incorporation1,2.

So far, more than 200 ncAAs have been incorporated site-specifically into proteins of interest (reviewed in Tang et al.4). There are many ncAAs with different physicochemical features available, however, the choice of ncAA is dependent on the ultimate application5. If the directed conjugation of a protein with a particular compound such as small molecule drugs or biopolymers (reviewed in Kim et al.6) is intended, ncAAs with bioorthogonal reactive groups that enable the required chemical reaction are chosen and incorporated at pre-defined positions. The ncAA used in this study, N6-[(2-Azidoethoxy) carbonyl]-L-lysine (Azk) (Fig. S1), carries an azide group on its side chain, which allows site-specific conjugation through strain-promoted or copper(I)-catalyzed azide-alkyne cycloaddition chemistry7.

ncAA-labeled antibodies and antigen binding fragments (Fab) can be conjugated to a drug in a site-specific and selective manner. The unique bioorthogonal moiety avoids the problems associated with canonical amino acids for drug conjugation which are mainly based on lysine and cysteine residues. Conjugation with these residues is a challenging process due to the heterogeneity of the population caused by difficulty in controlling the ligand-to-antibody ratio and the site specificity of the conjugation8,9. Despite promising applications, there have been few reports of successful production of ncAA-labeled antibodies/Fabs10,11. Production of recombinant protein having posttranslational modification is commonly done using eukaryotic expression systems, however for incorporating ncAAs into these proteins, the used expression systems should be genetically developed12,13. In vitro approaches tend to be used for studies on the functionalization of complex proteins with exotic or toxic ncAAs, however, these systems are less advanced than in vivo incorporation systems (reviewed in Cui et al.14). Therefore, antibody fragments which can be produced using simple bacterial expression systems offer an attractive alternative15. For instance, Fabs labeled in vivo with bioorthogonal reactive groups can be used for ncAA-based drug conjugation studies and also for the production of highly potent protein-drug conjugates16,17.

Very often, the incorporation of ncAAs leads to significant reduction in the production yield of the target proteins of interest in bacteria18. The efficiency of amber suppression is influenced by several key factors including the abundance of the orthogonal aaRS/tRNACUA pair, which depends on their promoters and the origin of replication on their expression plasmid; the affinity of the aaRS for the target ncAA; the site of ncAA incorporation in the sequence context on the mRNA and its effects on protein stability; as well as the competition by the releaser factor 1 in E. coli19,20,21,22,23. Furthermore, ncAAs uptake by the cells is inefficient and most of the internalized content may be consumed by and/or excreted by the host cells24,25. Thus, only a considerably small fraction remains for synthesis of the recombinant protein.

Although E. coli can synthesize all the canonical amino acids, it has two amino acid importer super-families dedicated for import of canonical amino acids, the ATP-binding cassette (ABC) and the amino acid/polyamine/organocation (APC)26. Some of the APC transporters have the highest specificity for certain canonical amino acids as their substrate, however, they can facilitate import of other canonical amino acids and metabolites that have similar chemical structures to the specific substrate26. This allows the hypothesis that some ncAAs can be transported into the cell via the canonical amino acid importers, albeit with different km values. However, this remains a theoretical assumption for all those ncAAs for which experimental data on their uptake into the cell are not yet available. Which also applies to the uptake of Azk the ncAA used in this study.

Various strategies have been devised to mitigate inefficient ncAA uptake. ncAAs can be supplied as di-peptides in conjugation with a canonical amino acid24 or in esterified form25 for improved uptake. As well, strains with engineered periplasmic binding protein, which facilitates the internalization of specific ncAAs were used27. The intracellular biosynthesis of the ncAA of interest represents another strategy for improving the yield of ncAA-labeled protein28. Poor uptake of ncAAs into the cell can also be counteracted by supplementing increased quantities in the medium and thus creating a steep concentration gradient29. However, this strategy also has its limits, as high ncAA concentrations can also have growth-inhibiting or toxic effects206, 106679. https://doi.org/10.1016/j.mimet.2023.106679 (2023)." href="/articles/s41598-024-73162-9#ref-CR30" id="ref-link-section-d133197542e607">30. Consequently, a suitable ncAA concentration range must be determined for each cell system/process/ncAA combination. Another limitation is the cost of ncAAs, which can be very high, and the limited availability of larger quantities31.

At the scientific laboratory scale, there is a considerable knowledge base on how proteins with incorporated ncAAs can be produced using shake flask or titer plate cultures. The next logical step on the path to full technological maturity is to move from scientific experiments to scalable processes to produce larger quantities of ncAA-incorporated proteins. This scalability can only be achieved through fully controlled bioreactor cultivation and defined synthetic media. However, little is known about how media composition and variations in culture and production conditions can affect the uptake of ncAAs and their incorporation into the target protein in this process environment. Optimizing and adjusting the process parameters that influence the incorporation of ncAA into the target protein and the protein yield is the central task of process engineering. For example, the production yield can be significantly improved by determining the best possible combination of type, time and amount of ncAA addition. Another approach is to arrest cell growth and decouple recombinant protein synthesis, either by genetic means or by reducing the media supply during the production phase. This strategy allows the release of intracellular resources and capacities that can then be used for product formation. At the same time, the misincorporation of ncAA into host proteins derived from genes with amber stop codons is reduced32,33,34. Considering the inefficient ncAA uptake, a potential strategy might be to design the cultivation conditions such that the cells allocate the ncAA predominantly for protein synthesis rather than consuming or secreting it.

In the present study, we investigated the site-specific incorporation of Azk into an anti-tumor-necrosis-factor-α-Fab (FTN2), which is an industrially relevant Fab that tolerates Azk incorporation without losing its structure and function35. We explored the recombinant production of Azk-labeled FTN2 in 1 mL fed-batch-like µ-bioreactor cultivations in titer-plate format and in fully controlled 1 L stirred tank bench-top bioreactor systems operated in C-limited fed-batch mode. Furthermore, the uptake of Azk was measured at different phases of fermentation using a RP-HPLC method.

The main objective of this study was to develop a scalable bench-scale bioreactor process to produce an Azk-incorporated FTN2 hereafter referred to as SHC133Z as the model protein. SHC133Z contained Azk in its HC at the permissive position 133 instead of serine while the light chain (LC) remained unaltered. Wild-type (wt) FTN2 was used as the reference. In pursuit of an improved production yield, we explored various cultivation parameters including time and mode (pulse or feed) of Azk addition, Azk uptake, duration of the production phase, and the time of induction.

With the µ-scale cultivations, our primary objective was to verify whether the SHC133 is permissive for Azk incorporation and whether it is feasible to produce SHC133Z with an acceptable titer. We used E. coli BL21(DE3)(pT7 × 3_oFTN2) expressing wt FTN2 as a reference and E. coli BL21(DE3)(pT7 × 3_oFTN2.SHC133am) that expressed the FTN2 amber mutant as the test strain. In both constructs, HC and LC were expressed as a bicistron under the control of the T7 promoter (Fig. S2). Both strains were grown in fed-batch like µ-scale culture under carbon limited conditions in the production phase. The Azk concentration in the medium was set to 10 mM in all µ-scale experiments with E. coli BL21(DE3)(pT7 × 3_oFTN2.SHC133am) either at the start of the culture or after 10 h of batch cultivation.

The results for two of these fed-batch like µ-scale cultivations are presented in Fig. 1A. With respect to growth characteristics, no significant differences were observed between the two variants. Data from other experiments presented in Fig. S3 also confirmed that Azk addition in the concentrations used and time point of addition did not affect cell growth. The batch phase lasted about 12 h and a final cell dry mass (CDM) of approximately 5 g/L was obtained in all experiments (Fig. 1A and Fig. S4). In all studied conditions, only the pattern of Azk concentration during the fermentation phases, concerning the time or mode of Azk addition, is presented, without considering the amount that can be internalized. For the Azk-incorporated FTN2 variant, SHC133Z, a product yield of 3.7 ± 0.5 mg/g CDM was reached which is about 80% of the wt FTN2 production yield of 4.7 ± 0.2 mg/g CDM under these conditions (Fig. 1B, Fig. S4). Presence of correctly translated SHC133Z was confirmed via the HC specific Western blot (Fig. S4) and quantification of the Fab was done via the sandwich ELISA that only detects LC-HC hetero dimer. LC-ESI-MS/MS analysis of trypsinized samples confirmed the site-specific incorporation of Azk at position S133 in the HC without any indication of misincorporation of cAAs (Fig. 1C).

µ-scale cultivations of BL21(DE3) strains producing wt FTN2 or SHC133Z. (A) Growth curves for wt FTN2 (black curve) and SHC133Z (blue curve) and calculated Azk concentration (blue line). The black arrow indicates the time of induction. (B) Fab production yield quantified via ELISA (n = 3 biological replicates).(C) LC-ESI-MS/MS analysis of the samples digested with trypsin. The target fragment GPSVFPLAPSXK for the SHC133Z (X = Azk, 1340.73 Da, blue) and its corresponding wt FTN2 fragment (X = S,1186.65 Da, black) are shown.

In a first approach (run #1), we adapted our established routine process for Fab production at laboratory scale using again the test and the reference strain. The medium was designed for a total of 6 g CDM in the batch phase and 34 g CDM in the feed phase. Since the µ-bioreactor results had shown that the time of Azk addition made no difference in CDM nor Fab yield, we added Azk in two pulses, the first 30 min before and the second 4 h after induction to set a concentration of around 15 mM (trend is shown in Fig. 2A) only to BL21(DE3)(pT7 × 3_oFTN2.SHC133am). Additionally, we kept the time interval between Azk addition and induction of target gene expression short to avoid possible Azk degradation/consumption in the non-induced phase of the process. The production phase was set to last for about 8 h, equivalent to approximately 0.6 cell generations.

Bench-scale bioreactor cultivations of E. coli BL21(DE3) strains producing wt FTN2 or SHC133Z. (A,B) run #1; (C,D) run #2; (E,F) run #3. (A,C,E) Calculated CDM (grey lines) and calculated Azk concentrations (blue lines), measured CDM for wt FTN2 (black triangles) and SHC133Z (blue circles), black arrows indicate induction times. (B,D,F) Fab production yield quantified via ELISA (n = 3 technical replicates, SDM).

In these experiments, the observed final CDM for both, the reference BL21(DE3)(pT7 × 3_oFTN2) and the test strain BL21(DE3)(pT7 × 3_oFTN2.SHC133am), was as expected (Fig. 2A). ELISA measurements revealed that wt FTN2 was produced up to 2.2 mg/g CDM under set conditions, but no SHC133Z was detected (Fig. 2B). Anti-HC Western blots of end of fermentation samples revealed no visible Fab band for SHC133Z (Fig. S5). However, anti-LC Western blot results clearly demonstrated that LC was efficiently produced (Fig. S5).

In the next experiment (run #2) with the test and the reference strain, we kept the cultivation conditions constant, but we changed the induction and the Azk addition strategies. To circumvent the potential problem of insufficient intracellular availability of Azk, we added the ncAA in one pulse 12 h before induction (still in fed-batch phase). In addition, induction was done earlier than in run #1 to allow for an extended production phase of 14 h, which corresponds to one generation at a growth rate of 0.05 h-1, to potentially increase the Fab yield (Fig. 2C).

Adding all the Azk in one pulse at an earlier point in time led to a notably high initial concentration of 25.4 mM Azk (Fig. 2C). Even this elevated concentration did not impact cell growth, as evidenced by the similar growth patterns for both the test and the reference strain with values close to theoretical biomass yields (Fig. 2C). With respect to Fab production, the experiment with wt FTN2 yielded 3.2 mg/g CDM representing an increase of ~ 50% compared to run #1 with a shorter production period. ELISA quantification of the SHC133Z variant resulted in a very low specific concentration of only 0.35 mg/g CDM (Fig. 2D), which was also confirmed by anti-HC western blot (Fig. S5).

To tackle the question whether the Azk to biomass ratio impacts the SHC133Z yield, we reduced the theoretical final biomass concentration to 10 g/L (run #3), which was four times lower than in the previous two runs. Azk was added to the culture in a pulse 12 h before induction, and the rest was added to the medium at induction and continuously fed to the culture during the production phase (Fig. 2E).

For both processes (wt FTN2 and SHC133Z) the cell growth was according to the expected kinetic (Fig. 2E). However, the specific product yields for wt FTN2 (2.48 mg/g CDM) and for SHC133Z (0.33 mg/g CDM) were lower than in the previous experiments (Fig. 2D, S5).

All bench-scale approaches tested so far had resulted in only very low SHC133Z yields (Table 1). Therefore, in the next experiment (run #4), Azk was added already at the beginning of the process to provide more time for Azk uptake into the cells. In addition, we used an HPLC method to quantify the amount of Azk in the cell supernatant for the direct assessment of Azk uptake.

With this experimental setup, again we observed no negative effects on cell growth (Fig. 3A). The yield of SHC133Z increased to 0.57 mg/g CDM, which was still much lower than the wt FTN2 production (Fig. 3B). The quantification of Azk in the supernatant by HPLC showed a decrease of 14% from 4.4 to 3.8 g in the batch phase and no further significant changes during the feed phase (Fig. 3C).

Bench-scale bioreactor cultivations of E. coli BL21(DE3) strains producing wt FTN2 or SHC133Z (run #4) with Azk addition at start of the culture. (A) Calculated CDM (grey line), calculated Azk concentration (blue line), measured CDM for wt FTN2 (black triangles) and SHC133Z (blue circles), black arrow indicates induction timepoint. (B) Fab production yield quantified by ELISA. (C) Total amount of Azk in the supernatant quantified by HPLC.

Based on the HPLC findings, we designed the next process differently (run #5). The induction was set to the end of the batch phase to eliminate the carbon limitation under non-induced conditions and the potentially associated degradation/consumption of Azk in the cell so that Azk could be immediately allocated for target protein synthesis. For this run, the batch biomass was set to a final concentration of 20 g/L in order to increase biomass, as early induction leads to having lower CDM (Fig. 4A).

Bench-scale bioreactor cultivation of E. coli BL21(DE3) strains producing wt FTN2 or SHC133Z (run #5) with Azk addition at the start of the culture and induction at feed start. (A) Calculated CDM (grey line) and calculated Azk concentration (blue line), measured CDM for wt FTN2 (black triangles) and SHC133Z (blue circles), black arrow indicates induction timepoints. (B) Fab production yield quantified by ELISA (n = 3 technical replicates).

The significantly increased CDM concentration in the batch phase did not cause any oxygen limitations (Fig. S6). In both cultivations (reference and test strain), the early induction led to reduced growth and thus to a final biomass 14% lower than calculated (Fig. 4A). Encouragingly, the implemented strategy led to a significant improvement in SHC133Z production yield to 2.95 mg/g CDM (Fig. 4B). This yield corresponded to approximately 80% of the wt FTN2 with 3.68 mg/g CDM achieved in the reference experiment. The site-specific and correct incorporation of Azk in the HC of the FTN2 at position 133 was confirmed by mass spectrometry analyses (Fig. S7) and no misincorporation of cAAs was observed. The additional peaks in the spectrogram are LC degradation products.

The µ-scale cultures showed high productivity of fully Azk-labeled SHC133Z, which represented 80% of the yield obtained with wt FTN2 in the same setting (Fig. 1B). This was rather unexpected because with genetic code expansion, such incorporation efficiencies are still rather uncommon19,34,36. However, Fabs are generally produced at comparably low yields in E. coli. Their low production appears beneficial as inefficiencies in Azk uptake may not have a strong impact on such low product titers. As the quantification by ELISA only detects LC-HC hetero dimer there is strong indication that our Fab is correctly processed and MS data from other experiments also confirm correct mass (Fig. S7).This conclusion is also supported by the Western blot data against the HC, with bands in the 50 kD range corresponding in size to the correctly formed Fab and bands in the 25 kD range corresponding to the size of the HC. Azk addition to E. coli BL21(DE3)(pT7 × 3_oFTN2.SHC133am) did not result in any changes in the growth curve compared to the reference experiment with no Azk added (Fig. 1A). There is also no impact on cell growth if time point of addition varies (Fig. S3). Production clones encoding wt FTN2 on a standard pET expression vector (pET30acer) showed identical growth behavior but about 50% higher yields with 7.2 mg/g for wt FTN237. This reduction in yield indicates that plasmid pT7 × 3, which encodes the orthogonal pair in addition to the protein of interest significantly affects the expression capacity of the cell even for wt FTN2.

The results from the standard Fab production process (run #1) showed that the yield for wt FTN2 with 2.2 mg/g was about 40% lower than that obtained in µ-scale experiments (Fig. 2A, B). As the cells in bench-scale cultivations with a growth rate of 0.05 h-1 are exposed to harsher and tightly carbon limiting conditions, an increased level of plasmid loss and formation of a non-producer population could be a potential reason for lower yields. Under these conditions, SHC133Z was not detectable at all. However, based on the Western blot results we can exclude that this is caused by total loss of plasmid or by a non-functional mutated SHC133am plasmid as the LC expression was confirmed by anti-LC Western blot (Fig. S5). It should be noted that only LC monomers are capable of making a homodimer (~ 50 kDa). Based on our observation this cannot be seen for HC, therefore, staining against HC could confirm whether HC-LC heterodimer has been formed.

So, we conclude that the timespan of only 30 min before induction and 4 h after induction was inappropriate to provide sufficient intracellular Azk to facilitate SHC133Z production.

The strategy to induce the Fab production earlier in the feed phase for an extended production phase and thus increased product yields (run #2) led to a ~ 50% increased specific wt FTN2 titer. Changing the Azk addition strategy to 12 h before induction still resulted in low (0.35 mg/g CDM) but detectable concentrations of correctly folded SHC133Z (Fig. 2C-D, S5). This yield was nevertheless far lower than what we achieved in µ-scale cultivations, and far from about 80% yield relative to wt FTN2 (Fig. 1B). One possible reason for this result could be the difference in cell densities with about eight times higher CDM/L in bench-scale fed-batch culture than in the µ-bioreactor cultures. However, the experiments with higher and constant Azk/CDM ratio shown in Fig. 2E-F, S5 clearly demonstrated that changes in the Azk/CDM ratios did not noticeably influence the SHC133Z production, at least not in the ratios that were encountered in our experiments.

The addition of Azk at the beginning of the process (run #4) was only partially successful, as the SHC133Z yield of 0.57 mg/g CDM was a significant improvement compared to the previous experiments (Fig. 3B). However, it was still far below the wt FTN2 yield. In general, the experiments discussed so far (Figs. 2 and 3) revealed that all strategies with different Azk addition times and variation of the Azk/CDM ratio did not lead to SHC133Z yields observed in µ-scale cultivations (Fig. 1). However, these unexpected results can be partially explained by the data on Azk amounts in the cell supernatant. In general, Azk uptake was inefficient, but worked better in the batch than the feed-phase. This observation may be related to the high growth rate in unlimited growth conditions. The constant amount of Azk in the supernatant during the feed phase may indicate that Azk uptake no longer occurred or that a steady state between Azk uptake and excretion had been established.

In the bench-scale experiments, in which SHC133Z was produced, there was a gap between the batch phase and the induction of Fab production. This is a remarkable difference compared to the µ-scale experiments, in which induction always occurred directly at the end of the batch phase. The high SHC133Z yields achieved in the µ-scale experiments suggest that elimination of a time gap between batch end and production can prevent or minimize potential Azk degradation/consumption.

The experiment with a high final batch CDM and induction of SHC133Z synthesis directly at the end of the batch phase (run #5) resulted in a specific product titer of 2.95 mg/g CDM (Fig. 4A, B) which was close to that obtained in µ-scale experiments (Fig. 1B). This result strongly points at the time window between batch end and induction being the reason for the poor yields of Azk-incorporated Fab in the earlier experiments. Obviously, no delay between the uptake of Azk and its consumption in synthesis of Azk -incorporated Fab works best. The reduced glucose yield coefficient observed for both the reference and test strain of run #5 can be explained by the long production phase which in case of Fabs usually leads to metabolic overload of the cells, which respond with reduced or no further growth37.

Since the charge of ncAAs might prevent its intracellular uptake, we assessed Azk charge in physiological pH range by computational calculations. We found that Azk is presumably neutral within the pH range of 4–12 (Fig. S5). Although Azk is neutral, its internalization is highly inefficient. Furthermore, the uptake of Azk is comparatively better under non-limiting growth conditions; however, its precise mechanism must await future elucidation.

We successfully developed a scalable production process for the Azk-incorporated Fab SHC133Z and identified several bottlenecks limiting yield and process flexibility. Uptake of Azk into the cell is limited to the batch phase and the induction of protein production must be close to the batch end to avoid Azk degradation or conversion during the C-limited fed-batch phase, where there is no significant sign of Azk uptake.

From a process engineering point of view, the finding that SHC133Z production must be induced shortly after the batch end for efficient Azk incorporation is disadvantageous. The high growth rates in the batch phase lead to high oxygen consumption rates yet generate limited amounts of biomass because the oxygen transfer in the bioreactor is limited. In turn, this results in lower product yields of intracellular proteins because these are directly correlated with biomass formation. We will use the knowledge gained in the current study as a base for a more detailed characterization of Azk’s intracellular fate. A basic prerequisite for this is the development of a sensitive analytical method that allows quantification of intracellular Azk. The influence of non-producing subpopulations is another important issue in the context of process optimization. Finally, we will validate our strategies with other target proteins and host strains.

The cassette for SHC133Z was constructed by substituting the serine codon at amino acid position 133 with TAG (amber) stop codon in the coding sequence. The wildtype cassette of FTN2 on the pET30a plasmid was used as starting material. The substitution was introduced by whole-plasmid PCR using Q5 high fidelity DNA polymerase (New England Biolabs, NEB) and pair of primer (IDT) with the forward primer having TAG overhang at its 5’. The ligated plasmid (T4 DNA ligase from NEB) was transferred into NEB5α (NEB). After plasmid isolation, new restriction sites for NdeI and BglII (NEB) for each cassette were introduced, and afterwards, the cassette was inserted into pT7 × 3 plasmid34. This plasmid encodes the orthogonal MmPylRS/MmtRNACUAPyl pair requited for Azk incorporation. Codon substitution for each cassette was confirmed by sequencing (Microsynth, Vienna, Austria). The resulting plasmids for the production of wt FTN2 and SHC133Z were dubbed (pT7 × 3_oFTN2) and (pT7 × 3_oFTN2.SHC133am) where “o” indicates the leader sequence (ompA) for periplasmic expression and “am” refers to the in-frame amber stop codon. Competent BL21(DE3) (NEB) cells were transformed with the plasmids and the expression of SHC133Z was assessed. Master cell banks were prepared from positive clones according to38. Plasmids, strains, and primers are presented in Table S1.

The BioLector system (m2p-labs GmbH, Germany) was used for fed-batch-like µ-cultivations in 48-well Flowerplates® as described by Toeroek et al.39. The fed-batch like condition was provided through continuous degradation of dextran by a mixture of amylases which were provided by the manufacturer38. Precultures were prepared according to40 and used for inoculation to achieve a final volume of 800 µL with an initial OD600 of 0.3. Induction was performed with a concentration of 0.5 mM of IPTG (Carl Roth, Germany) The production phase was set to be 8 h, and endpoint samples were used for analyses. All cultivation conditions were set for a shaking speed of 1400 rpm, at 30 °C and relative humidity of 80%. Online scattered light measurement at 620 nm was converted to CDM using a pre-defined calibration curve. Azk was only added to cultures producing the modified Fab SHC133Z either from the beginning or 2 h before induction, that is 10 h after the start of cultivation, to reach a final concentration of 10 mM.

For the bench-scale fed-batch cultivations, we used a DASGIP® parallel bioreactor system (Eppendorf AG, Germany) enabling four parallel cultivations. The total vessel volume was 2.1 L with a maximum working volume of 1.8 L The bioreactors were equipped with a pH probe (Hamilton Bonaduz AG, Switzerland), an optical dissolved oxygen (DO) probe (Hamilton Bonaduz AG). For media preparation, all chemicals were purchased from Carl Roth GmbH (Germany) if not stated otherwise. The volumes used for the batch phase was fixed to 600 mL. The bioreactors were inoculated with an initial OD600 of 0.3 with cells from a pre-culture. For pre-cultures, 250 mL shake flasks with 50 mL semisynthetic medium (SSM, Table S2) were used, inoculated with cells from working cell banks and incubated in an orbital shaker at 37 °C at 180 rpm. The amount of glucose for the batch and feed phases of the bench-scale cultures was calculated based on the biomass to be produced by using a yield coefficient (Yx/s) of 0.3 g/g for glucose monohydrate. All other compounds of the medium were calculated accordingly in relation to the planned biomass. Due to solubility issues the amounts of KH2PO4, 85% H3PO4, C6H5Na3O7·2H2O calculated for the whole process were already added to the batch medium (see Tables S3 and S4).

For all runs, an exponential feed providing a specific growth rate of 0.05 h-1 was applied. The temperature was kept at 37 °C ± 0.2 during batch and at 30 °C ± 0.2 in the feed and production phase, respectively. During the whole process, the pH was maintained at 7.0 ± 0.2 by the addition of 12.5% (w/w) ammonia. The dissolved oxygen saturation was set to 30% by cascade control of the stirrer speed, aeration rate, and addition of oxygen to the gassing air. Foaming was suppressed by the addition of 750 µL of PPG 2000 (BASF, Germany) to the batch medium and by the automatic addition of 1:10 diluted PPG 2000 controlled by a conductivity operated level sensor.

Protein expression was induced by addition of 2 µmol IPTG/g of theoretical CDM. A part of the calculated amount was added at the induction time considering the g CDM at that instance, and the rest was added to the feed medium. In bench-top bioreactor cultivations Azk addition was explicitly limited to cultures producing the modified Fab SHC133Z. There was no Azk addition to reference cultures producing wtFTN2. Time and type of addition (pulse or feed) are described in detail in the respective experiments. Amount and time of Azk (Iris Biotech GmbH, Marktredwitz, Germany) addition was as explained in the Results and Discussion sections.

CDM was determined gravimetrically as described by40. Briefly, 1 mL of cell suspension was transferred to pre-weighed 2.0 mL tubes and centrifuged for 10 min at 16,100 rcf at 4 °C. The pellet was washed with reverse osmosis H2O, and after another centrifugation step as above, it was resuspended in 1.8 mL reverse osmosis H2O, dried at 105 °C for 24 h and weighed.

To confirm the production of the SHC133Z variant both quantitatively and qualitatively, samples from bench-scale and µ-scale cultivations were lysed in 400 µL of lysis buffer (30 mM Tris-HCl, pH 8.2 containing 25 mM EDTA, 250 mM MgCl2, lysozyme (50 µl, 0.5 mg/mL), benzonase (50 µl, 50 U/mL) and Triton-100X) according to37. After centrifugation at 15,000 rpm at 4 °C for 15 min, the supernatant was collected to be used for ELISA.

For sandwich ELISA briefly, a 96-well plate (Nunc MaxiSorp, Denmark) was coated with anti-human IgG (Fab specific) goat antibody (I5260, 1:400, Sigma-Aldrich). Anti-human IgG mouse antibody (SA19255, 1:1000, sigma, ) and peroxidase-labeled anti-mouse IgG (Fab specific) goat antibody (A2304; 1:1000, Sigma-Aldrich) were used as primary and secondary antibodies. The capture antibody binds only the LC, whereas the detection antibody specifically recognizes the hinge region of the HC. Consequently, only LC-HC dimer can be detected via ELISA. Details are provided by37.

For quantification of Azk in the supernatant, we employed an HPLC-based method routinely used for quantifying free amino acids in their derivatized form with OPA41. To quantify the Azk concentration in cell supernatants first a calibration curve was prepared. Known varying concentrations of Azk and a constant concentration of 3-(2-thienyl)-DL-alanine (Fluka) as the internal derivatization control were used as a standard. To be in linear part of calibration curve, all samples were diluted 20 times. Using an automated pre-column derivatization program, the primary amine groups in the mixture were derivatized with ortho-phthalaldehyde (OPA, Agilent) followed by separation on a Poroshell HPH-C18 reversed-phase column (Agilent) at 40 °C using a flow rate of 0.64 mL/min. After gradient elution with Buffer A and B the OPA-derivatized compounds were excited at 230 nm and the signal was detected at 450 nm. Buffer A was composed of 10 mM K2HPO4, 10 mM K2B4O7 pH 8.2 and buffer B contained a mixture of acetonitrile/methanol/water (45:45:10, v/v/v; Merck). Samples from different points in time were collected and centrifuged (15,000 rpm, 4 °C, 15 min). The supernatants were filtered (0.2 μm filter, Sartorius) and used for HPLC measurement as outlined above.

Site-specific incorporation of Azk was monitored by mass spectrometry. First, the samples were purified using Protein G-based affinity chromatography (ÄKTA™ pure workstation, Cytiva, Sweden) in combination with a KanCap™ G (Kaneka, Japan) prepacked column according to42. The protein G-purified samples of µ-scale and bench-scale cultivations were used for peptide mapping and intact protein analysis, respectively.

For peptide mapping, the sample was S-alkylated with iodoacetamide and digested with trypsin. The digested samples were loaded on a nanoEase C18 column (nanoEase M/Z HSS T3 Column, 100 Å, 1.8 μm, 300 μm x 150 mm, Waters) using 0.1% (v/v) formic acid as the aqueous solvent. A gradient from 3.5% B (B: 80% acetonitrile, 20% A) to 40% B in 30 min was applied, followed by a 5 min gradient from 40% B to 95% B that facilitates elution of large peptides, at a flow rate of 6 µL/min. Detection was performed with an iontrap MS (amazon speed ETD, Bruker) equipped with the standard ESI source in positive ion, DDA mode (= switching to MSMS mode for eluting peaks). MS-scans were recorded (range: 150–2200 Da) and the 8 highest peaks were selected for fragmentation. Instrument calibration was performed using ESI calibration mixture (Agilent). The analysis files were converted (using Data Analysis, Bruker) to mgf files, which are suitable for performing an MS/MS ion search with X!-Tandem.

For intact protein analysis (samples of bench-scale studies), 5 µL of the protein solution was directly injected to a LC-ESI-MS system (LC: Agilent 1290 Infinity II UPLC). A gradient from 15 to 80% acetonitrile in 0.1% formic acid (using a Waters BioResolve column (2.1 × 5 mm)) at a flow rate of 400 µL/min was applied. Detection was performed with a TOF instrument (Agilent Series 6230 LC-TOFMS) equipped with the Jetstream ESI source in positive ion, MS mode (range: 400–3000 Da). Data was processed using MassHunter BioConfirm B.08.00 (Agilent) and the spectrum was deconvoluted by MaxEnt.

All data generated or analyzed during this study are included in this published article and its supplementary information.

Wang, Q., Parrish, A. R. & Wang, L. Expanding the genetic code for biological studies. Chem. Biol. 16, 323–336. https://doi.org/10.1016/j.chembiol.2009.03.001 (2009).

Article CAS PubMed PubMed Central Google Scholar

Chin, J. W. Expanding and reprogramming the genetic code. Nature. 550, 53–60. https://doi.org/10.1038/nature24031 (2017).

Article ADS CAS PubMed Google Scholar

Smolskaya, S. & Andreev, Y. A. Site-specific incorporation of unnatural amino acids into Escherichia coli recombinant protein: Methodology development and recent achievement. Biomolecules 9 (2019). https://doi.org/10.3390/biom9070255

Tang, H., Zhang, P. & Luo, X. Recent technologies for Genetic Code Expansion and their implications on Synthetic Biology Applications. J. Mol. Biol. 434, 167382. https://doi.org/10.1016/j.jmb.2021.167382 (2022).

Article CAS PubMed Google Scholar

Rezhdo, A., Islam, M., Huang, M. & Van Deventer, J. A. Future prospects for noncanonical amino acids in biological therapeutics. Curr. Opin. Biotechnol. 60, 168–178. https://doi.org/10.1016/j.copbio.2019.02.020 (2019).

Article CAS PubMed PubMed Central Google Scholar

Kim, C. H., Axup, J. Y. & Schultz, P. G. Protein conjugation with genetically encoded unnatural amino acids. Curr. Opin. Chem. Biol. 17, 412–419. https://doi.org/10.1016/j.cbpa.2013.04.017 (2013).

Article CAS PubMed PubMed Central Google Scholar

Mbua, N. E., Guo, J., Wolfert, M. A., Steet, R. & Boons, G. J. Strain-promoted alkyne-azide cycloadditions (SPAAC) reveal new features of glycoconjugate biosynthesis. Chembiochem. 12, 1912–1921. https://doi.org/10.1002/cbic.201100117 (2011).

Article CAS PubMed PubMed Central Google Scholar

Sun, S. B., Schultz, P. G. & Kim, C. H. Therapeutic applications of an expanded genetic code. Chembiochem. 15, 1721–1729. https://doi.org/10.1002/cbic.201402154 (2014).

Article CAS PubMed PubMed Central Google Scholar

Dennler, P., Fischer, E. & Schibli, R. Antibody conjugates: From heterogeneous populations to defined reagents. Antibodies. 4, 197–224 (2015).

Article Google Scholar

Axup, J. Y. et al. Synthesis of site-specific antibody-drug conjugates using unnatural amino acids. Proc. Natl. Acad. Sci. USA. 109, 16101–16106. https://doi.org/10.1073/pnas.1211023109 (2012).

Article ADS PubMed PubMed Central Google Scholar

VanBrunt, M. P. et al. Genetically encoded Azide containing amino acid in mammalian cells enables site-specific antibody-drug Conjugates using click Cycloaddition Chemistry. Bioconjug. Chem. 26, 2249–2260. https://doi.org/10.1021/acs.bioconjchem.5b00359 (2015).

Article CAS PubMed Google Scholar

Guo, Q. & Cao, Y. J. Applications of genetic code expansion technology in eukaryotes. Protein Cell. 15, 331–363. https://doi.org/10.1093/procel/pwad051 (2023).

Article CAS PubMed Central Google Scholar

Zhou, W., Wesalo, J. S., Liu, J. & Deiters, A. Genetic code expansion in mammalian cells: A plasmid system comparison. Bioorg. Med. Chem. 28, 115772. https://doi.org/10.1016/j.bmc.2020.115772 (2020).

Article CAS PubMed PubMed Central Google Scholar

Cui, Z., Johnston, W. A. & Alexandrov, K. Cell-Free Approach for non-canonical amino acids incorporation into polypeptides. Front. Bioeng. Biotechnol. 8https://doi.org/10.3389/fbioe.2020.01031 (2020).

Pirkalkhoran, S. et al. Bioengineering of antibody fragments: Challenges and opportunities. Bioengineering. 10, 122 (2023).

Article CAS PubMed PubMed Central Google Scholar

Yu, X. et al. PD-L1/TLR7 dual-targeting nanobody-drug conjugate mediates potent tumor regression via elevating tumor immunogenicity in a host-expressed PD-L1 bias-dependent way. J. Immunother. Cancer 10https://doi.org/10.1136/jitc-2022-004590 (2022).

Bordeau, B. M., Nguyen, T. D., Polli, J. R., Chen, P. & Balthasar, J. P. Payload-binding fab fragments increase the Therapeutic Index of MMAE antibody-drug conjugates. Mol. Cancer Ther. 22, 459–470. https://doi.org/10.1158/1535-7163.Mct-22-0440 (2023).

Article CAS PubMed PubMed Central Google Scholar

Martin, R. W. et al. Cell-free protein synthesis from genomically recoded bacteria enables multisite incorporation of noncanonical amino acids. Nat. Commun. 9, 1203. https://doi.org/10.1038/s41467-018-03469-5 (2018).

Article ADS CAS PubMed PubMed Central Google Scholar

Cho, H. et al. Optimized clinical performance of growth hormone with an expanded genetic code. Proc. Natl. Acad. Sci. 108, 9060–9065 (2011). https://doi.org/10.1073/pnas.1100387108

Johnson, D. B. et al. RF1 knockout allows ribosomal incorporation of unnatural amino acids at multiple sites. Nat. Chem. Biol. 7, 779–786. https://doi.org/10.1038/nchembio.657 (2011).

Article CAS PubMed PubMed Central Google Scholar

Chatterjee, A., Sun, S. B., Furman, J. L., Xiao, H. & Schultz, P. G. A versatile platform for single- and multiple-unnatural amino acid mutagenesis in Escherichia coli. Biochemistry 52, 1828–1837. https://doi.org/10.1021/bi4000244 (2013).

Article CAS PubMed Google Scholar

Guo, J., Melançon, C. E. 3rd, Lee, H. S., Groff, D. & Schultz, P. G. Evolution of amber suppressor tRNAs for efficient bacterial production of proteins containing nonnatural amino acids. Angew Chem. Int. Ed. Engl. 48, 9148–9151. https://doi.org/10.1002/anie.200904035 (2009).

Young, T. S., Ahmad, I., Yin, J. A. & Schultz, P. G. An enhanced system for unnatural amino acid mutagenesis in E. coli. J. Mol. Biol. 395, 361–374. https://doi.org/10.1016/j.jmb.2009.10.030 (2010).

Article CAS PubMed Google Scholar

Luo, X. et al. Genetically encoding phosphotyrosine and its nonhydrolyzable analog in bacteria. Nat. Chem. Biol. 13, 845–849. https://doi.org/10.1038/nchembio.2405 (2017).

Article CAS PubMed PubMed Central Google Scholar

Takimoto, J. K., Xiang, Z., Kang, J. Y. & Wang, L. Esterification of an unnatural amino acid structurally deviating from canonical amino acids promotes its uptake and incorporation into proteins in mammalian cells. Chembiochem. 11, 2268–2272. https://doi.org/10.1002/cbic.201000436 (2010).

Article CAS PubMed PubMed Central Google Scholar

Hook, C., Eremina, N., Zaytsev, P., Varlamova, D. & Stoynova, N. The Escherichia coli amino acid uptake protein CycA: Regulation of its synthesis and practical application in l-Isoleucine production. Microorganisms 10, 647 (2022).

Article CAS PubMed PubMed Central Google Scholar

Ko, W., Kumar, R., Kim, S. & Lee, H. S. Construction of bacterial cells with an active transport system for unnatural amino acids. ACS Synth. Biol. 8, 1195–1203. https://doi.org/10.1021/acssynbio.9b00076 (2019).

Article CAS PubMed Google Scholar

Chen, L. et al. Advances in biosynthesis of non-canonical amino acids (ncAAs) and the methods of ncAAs incorporation into proteins. Molecules. 28, 6745 (2023).

Article CAS PubMed PubMed Central Google Scholar

Chen, Y. et al. Unleashing the potential of noncanonical amino acid biosynthesis to create cells with precision tyrosine sulfation. Nat. Commun. 13, 5434. https://doi.org/10.1038/s41467-022-33111-4 (2022).

Article ADS CAS PubMed PubMed Central Google Scholar

Landor, L. A. I., Bratbak, G., Larsen, A., Tjendra, J. & Våge, S. Differential toxicity of bioorthogonal non-canonical amino acids (BONCAT) in Escherichia coli. J. Microbiol. MethodsBold">206, 106679. https://doi.org/10.1016/j.mimet.2023.106679 (2023).

Article CAS PubMed Google Scholar

Hickey, J. L., Sindhikara, D., Zultanski, S. L. & Schultz, D. M. Beyond 20 in the 21st Century: Prospects and challenges of non-canonical amino acids in peptide drug Discovery. ACS Med. Chem. Lett. 14, 557–565. https://doi.org/10.1021/acsmedchemlett.3c00037 (2023).

Article CAS PubMed PubMed Central Google Scholar

Stargardt, P., Feuchtenhofer, L., Cserjan-Puschmann, M., Striedner, G. & Mairhofer, J. Bacteriophage inspired growth-decoupled recombinant protein production in Escherichia coli. ACS Synth. Biol. 9, 1336–1348. https://doi.org/10.1021/acssynbio.0c00028 (2020).

Article CAS PubMed Google Scholar

Yu, T. et al. Reprogramming yeast metabolism from alcoholic fermentation to Lipogenesis. Cell 174https://doi.org/10.1016/j.cell.2018.07.013 (2018).

Galindo Casas, M., Stargardt, P., Mairhofer, J. & Wiltschi, B. Decoupling protein production from cell growth enhances the site-specific incorporation of noncanonical amino acids in E. coli. ACS Synth. Biol. 9, 3052–3066. https://doi.org/10.1021/acssynbio.0c00298 (2020).

Article CAS PubMed Google Scholar

Hanaee-Ahvaz, H. et al. Antibody fragments functionalized with non-canonical amino acids preserving structure and functionality—A door opener for new biological and therapeutic applications. Heliyon 9, e22463. https://doi.org/10.1016/j.heliyon.2023.e22463 (2023).

Article CAS PubMed PubMed Central Google Scholar

Hohl, A. et al. Engineering a polyspecific Pyrrolysyl-tRNA synthetase by a high throughput FACS screen. Sci. Rep. 9, 11971. https://doi.org/10.1038/s41598-019-48357-0 (2019).

Article ADS CAS PubMed PubMed Central Google Scholar

Fink, M. et al. Microbioreactor cultivations of Fab-Producing Escherichia coli Reveal Genome-Integrated systems as suitable for prospective studies on direct Fab expression effects. Biotechnol. J. 14, 1800637. https://doi.org/10.1002/biot.201800637 (2019).

Article CAS Google Scholar

Vazulka, S., Schiavinato, M., Wagenknecht, M., Cserjan-Puschmann, M. & Striedner, G. Interaction of Periplasmic Fab production and intracellular redox balance in Escherichia coli affects product yield. ACS Synth. Biol. 11, 820–834. https://doi.org/10.1021/acssynbio.1c00502 (2022).

Article CAS PubMed PubMed Central Google Scholar

Toeroek, C., Cserjan-Puschmann, M., Bayer, K. & Striedner, G. Fed-batch like cultivation in a micro-bioreactor: Screening conditions relevant for Escherichia coli based production processes. Springerplus 4, 490. https://doi.org/10.1186/s40064-015-1313-z (2015).

Article CAS PubMed PubMed Central Google Scholar

Fink, M., Cserjan-Puschmann, M., Reinisch, D. & Striedner, G. High-throughput microbioreactor provides a capable tool for early stage bioprocess development. Sci. Rep. 11, 2056. https://doi.org/10.1038/s41598-021-81633-6 (2021).

Article CAS PubMed PubMed Central Google Scholar

Lemoine, A., Martίnez-Iturralde, M., Spann, N., Neubauer, R., Junne, S. & P. & Response of Corynebacterium glutamicum exposed to oscillating cultivation conditions in a two- and a novel three-compartment scale-down bioreactor. Biotechnol. Bioeng. 112, 1220–1231. https://doi.org/10.1002/bit.25543 (2015).

Article CAS PubMed Google Scholar

Schimek, C. et al. Three-dimensional chromatography for purification and characterization of antibody fragments and related impurities from Escherichia coli crude extracts. J. Chromatogr. A 1638, 461702. https://doi.org/10.1016/j.chroma.2020.461702 (2021).

Article CAS PubMed Google Scholar

Download references

The financial support provided by the Austrian Federal Ministry for Digital and Economic Affairs, the National Foundation for Research, Technology, and Development, and the Christian Doppler Research Association is gratefully acknowledged. Furthermore, we gratefully acknowledge the financial support received from our industry partner, Boehringer Ingelheim RCV GmbH & Co KG. We thank the colleagues of BI RCV for the scientific input and their continuous scientific support. Furthermore, we thank Alexander Doleschal for providing help with fermentations, and Anton Sphylovyi for providing help with protein purification. Also, we gratefully thank Dr. Younes Valadbeigi for doing Computational calculations.

Christian Doppler Laboratory for Production of Next-Level Biopharmaceuticals in E. coli, Department of Biotechnology, BOKU University, Institute of Bioprocess Science and Engineering, Vienna, Austria

Hana Hanaee-Ahvaz, Marina Alexandra Baumann, Christopher Tauer, Birgit Wiltschi, Monika Cserjan-Puschmann & Gerald Striedner

Biopharma Austria, Process Science, Boehringer Ingelheim Regional Center Vienna GmbH & Co KG, Vienna, Austria

Bernd Albrecht

Acib - Austrian Centre of Industrial Biotechnology, Vienna, Austria

Birgit Wiltschi

You can also search for this author in PubMed Google Scholar

H. H-A. contributed to conception and design of the experiments, data analysis, and interpretation, performed the practical work, and drafted the manuscript; B. M performed part of practical work. C. T performed molecular biology experiments. B. (A) data interpretation and paper revision; (B) W. contributed to study design, data interpretation, and paper revision; M.C.-P. contributed to study design, data interpretation, and paper revision; G.S. designed the study, revised the paper, and secured funding. All authors reviewed the manuscript.

Correspondence to Monika Cserjan-Puschmann.

The authors declare no competing interests.

No ethical aspects need to be considered in this study. No animal experiments were performed.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Below is the link to the electronic supplementary material.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Hanaee-Ahvaz, H., Baumann, M.A., Tauer, C. et al. Aligning fermentation conditions with non-canonical amino acid addition strategy is essential for Nε-((2-azidoethoxy)carbonyl)-L-lysine uptake and incorporation into the target protein. Sci Rep 14, 25375 (2024). https://doi.org/10.1038/s41598-024-73162-9

Download citation

Received: 21 March 2024

Accepted: 16 September 2024

Published: 25 October 2024

DOI: https://doi.org/10.1038/s41598-024-73162-9

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

I'm a doctor — these are the 3 worst chemicals in your food that could damage your health

Bora Pharmaceuticals to Expand Rare Disease Portfolio with Acquisition of US-Based Pyros Pharmaceuticals

News

Aligning fermentation conditions with non-canonical amino acid addition strategy is essential for Nε-((2-azidoethoxy)carbonyl)-L-lysine uptake and incorporation into the target protein | Scientific Reports