Control of three-carbon amino acid homeostasis by promiscuous importers and exporters in Bacillus subtilis: role of the “sleeping beauty” amino acid exporters

ABSTRACT The Gram-positive model bacterium Bacillus subtilis can acquire amino acids by import, de novo biosynthesis, or degradation of proteins and peptides. The accumulation of several amino acids inhibits the growth of B. subtilis, probably due to misincorporation into cellular macromolecules such as proteins or peptidoglycan or due to interference with other amino acid biosynthetic pathways. Here, we studied the adaptation of B. subtilis to toxic concentrations of the three-carbon amino acids L-alanine, β-alanine, and 2,3-diaminopropionic acid, as well as the two-carbon amino acid glycine. Resistance to the non-proteinogenic amino acid β-alanine, which is a precursor for coenzyme A biosynthesis, is achieved by mutations that either activate a cryptic amino acid exporter, AexA (previously YdeD), or inactivate the amino acid importers AimA, AimB (previously YbxG), and BcaP. The aexA gene is very poorly expressed under most conditions studied. However, mutations affecting the transcription factor AerA (previously YdeC) can result in strong constitutive aexA expression. AexA is the first characterized member of a group of amino acid exporters in B. subtilis, which are all very poorly expressed. Therefore, we suggest to call this group “sleeping beauty amino acid exporters.” 2,3-Diaminopropionic acid can also be exported by AexA, and this amino acid also seems to be a natural substrate of AerA/AexA, as it can cause a slight but significant induction of aexA expression, and AexA also provides some natural resistance toward 2,3-diaminopropionic acid. Moreover, our work shows how low-specificity amino acid transporters contribute to amino acid homeostasis in B. subtilis. IMPORTANCE Even though Bacillus subtilis is one of the most-studied bacteria, amino acid homeostasis in this organism is not fully understood. We have identified import and export systems for the C2 and C3 amino acids. Our work demonstrates that the responsible amino acid permeases contribute in a rather promiscuitive way to amino acid uptake. In addition, we have discovered AexA, the first member of a group of very poorly expressed amino acid exporters in B. subtilis that we call “sleeping beauty amino acid exporters.” The expression of these transporters is typically triggered by mutations in corresponding regulator genes that are acquired upon exposure to toxic amino acids. These exporters are ubiquitous in all domains of life. It is tempting to speculate that many of them are not expressed until the cells experience selective pressure by toxic compounds, and they protect the cells from rare but potentially dangerous encounters with such compounds.

may have important functions for the cell.In bacteria, meso-diaminopimelic acid and ornithine, as well as D-glutamate and D-alanine, are essential building blocks of the peptidoglycan cell wall (1).β-alanine, the only naturally occurring β-amino acid, is required as a precursor for the biosynthesis of coenzyme A (2,3).Citrulline, ornithine, and γ-aminobutyrate can be used as carbon and nitrogen sources, and the latter two also serve as inducers for transcription factors in Bacillus subtilis (4)(5)(6).Finally, citrulline serves as a compatible solute to protect the cell from osmotic stress (7).
While amino acids are not only important to the cells, they can also be harmful.Often, the addition of amino acids results in growth inhibition.This is the case for some of the natural proteinogenic and non-proteinogenic amino acids but also for artificial analogs (8).As an example, serine can give rise to α-aminoacrylate and the highly reactive, and therefore toxic, β-hydroxypyruvate through deamination and transamina tion, respectively (9,10).The growth inhibition can also be caused by the reactivity of the amino acids, leading to the formation of toxic metabolites.In the case of natural amino acids, their toxicity may also be the result of chemical similarity to other amino acids that may result in binding to enzymes of non-cognate biosynthetic pathways and thus inhibition of biosynthesis and subsequent limitation of a different amino acid.This effect is well studied in the case of serine, which, in Escherichia coli, can bind to and inactivate the bifunctional enzyme aspartate kinase/homoserine dehydrogenase (ThrA) (11), thus interfering with threonine biosynthesis.The effect of serine interference with threonine synthesis also occurs in B. subtilis (12).Non-natural amino acids can also be toxic for the same reasons, but in addition, they can also be misincorporated into proteins due to the poor discrimination of amino acyl-tRNA synthetases against these amino acid analogs, which was never required during the evolution of these systems (13).
Bacteria have several mechanisms to protect themselves against the action of growth-inhibiting amino acids.For some of them, the bacteria have degradative pathways that can convert the toxic amino acids to non-inhibiting metabolites.However, these degradation mechanisms are often not sufficiently active in wild-type strains, and only their mutational activation allows the effective disposal of some of the amino acids, as shown for serine in B. subtilis and for B. subtilis strains that are sensitive to gluta mate (12,14,15).In some cases, bacteria have specific protective mechanisms against the misincorporation of D-amino acids into proteins, e.g., by producing a D-aminoacyl-tRNA deacylase that can recycle misaminoacylated D-Tyr-tRNA(Tyr) and some other D-aminoacyl-tRNAs (13).In addition, a few bacteria, including B. subtilis, possess even a second tyrosine-specific amino acyl-tRNA synthetase that strongly discriminates against D-tyrosine.However, the expression of the corresponding gene requires a mutation that inactivates the transcription repressor (16).Interestingly, the D-aminoacyl-tRNA deacylase is inactive in many B. subtilis strains; they thus have to rely on the more selective Tyr-tRNA synthetase (17).A third way to deal with toxic amino acids is their export; this has been described for the leucine analog 4-azaleucine and histidine.In both cases, export by the AzlCD amino acid exporter only becomes possible when the transcription repressor AzlB which normally prevents the expression of the exporter is inactivated (18,19).
In addition to the active strategies to deal with harmful amino acids, the bacteria can also deny them access to the cell by the inactivation of the corresponding transporters.The use of toxic amino acid analogs or the identification of conditions under which a given amino acid becomes growth-inhibiting is a powerful tool for the identification of amino acid transporters.This approach has been used to identify transporters for serine, threonine, proline, glutamate, and asparagine, as well as for the toxic inhibitors of amino acid biosyntheses, glyphosate and glufosinate, in B. subtilis (12,15,(20)(21)(22)(23).This strategy, however, has limited application as some amino acids are imported by multiple transporters (19,24), and thus, the simultaneous inactivation of all of them would be required to prevent the uptake.Moreover, in natural ecosystems, the loss of the ability to exploit important resources such as amino acids from the environment would certainly compromise the competitiveness of such bacteria.
We have a long-standing interest in the control of amino acid homeostasis in the model bacterium B. subtilis (5,19,(25)(26)(27).Although B. subtilis is one of the best-studied bacteria, the transporters for several amino acids as well as the substrates of several potential amino acid transporters have not yet been elucidated (24,28,29).With the exception of γ-aminobutyrate (30), the transport of non-proteogenic amino acids has not been studied in B. subtilis.In this work, we have studied the homeostasis of C3 amino acids in B. subtilis, α-L-alanine, β-alanine, and 2,3-diaminopropionic acid.Using the fact that all three amino acids inhibit the growth of B. subtilis, we identified the underlying importers and exporters.All three amino acids are transported into the cell by the broad-range amino acid transporter AimA.In addition, α-L-alanine and β-alanine are taken up by AlaP, whereas diaminopropionic acid can also be transported by BcaP and YbxG.The latter protein also acts as a broad-range amino acid permease, which we, therefore, rename AimB (amino acid importer B).Again, all three C3 amino acids can be exported by the poorly expressed amino acid exporter YdeD, which we rename AexA (amino acid exporter A).AexA is the first member of a group of "sleeping beauty" amino acid exporters.Expression of the corresponding aexA gene can either be induced by the presence of 2,3-diaminopropionic acid or by mutational activation of the transcription factor YdeC, which we rename AerA (amino acid export regulator A).

Three-carbon amino acids inhibit the growth B. subtilis
To test how B. subtilis can cope with aminated propionic acid derivatives, we tested the growth of the wild-type strain 168 in the presence of L-α-alanine, β-alanine, and 2,3-diaminopropionic acid (Fig. 1A) in Luria-Bertani (LB) complex medium and C minimal medium.As shown in Fig. 1B, the strain grew on both media in the absence of any added amino acid.Similarly, the bacteria were efficiently plated on an LB medium irrespective of the presence of the amino acids.In contrast, we observed a concentrationdependent growth inhibition by all three C3 amino acids in a minimal medium (Fig. 1C).This inhibition was most severe with diaminopropionic acid, where 2 mM caused a nearly total growth inhibition.In contrast, the inhibitive effect was less pronounced for β-alanine and even less for the proteinogenic amino acid L-α-alanine.Since β-alanine is of biological relevance as a precursor for the biosynthesis of coenzyme A, we focussed first on this amino acid.

A suppressor screen identifies a regulator involved in tolerance to β-alanine
To get more insights into the toxicity of β-alanine, we isolated suppressor mutants that were able to grow at increased β-alanine concentrations.In the first step, mutants were isolated at 30 mM β-alanine.Of six isolated suppressor mutants that were viable in the presence of β-alanine, two were subjected to whole-genome sequencing.In both mutant strains, we identified single-point mutations in the aerA (ydeC) gene, encoding a so far unknown transcription factor of the AraC family (31).In strain GP4117, Tyr-179 is replaced by an aspartate residue, and in GP4119, Leu-122 is replaced by a tryptophan.The fact that aerA was the only affected gene in both cases suggested an important role for this gene in the acquisition of resistance to β-alanine.We, therefore, sequenced the aerA allele in the remaining four mutants and found that all had the Y179D substitution as in GP4117.Thus, the potential AraC-type transcription factor plays a key role in the adaptation of B. subtilis to β-alanine (see Fig. 2 for an overview of all isolated suppressor mutants).
Since AerA is a member of the AraC family of transcriptional regulators, the muta tions affecting AerA might result in the inactivation of the protein, or they might have an activating effect.To distinguish between these possibilities, we tested the plating efficiency of the aerA deletion mutant BKK05150 on a medium containing β-alanine.As shown in Fig. 3A, the mutant was as sensitive as the wild type suggesting that the suppressor mutations resulted in an activation of the AerA protein.In the case of a repressor protein, this would result in even stronger repression of the target genes.In contrast, for an activator, the suppressor mutation might allow transcription activation even in the absence of a putative inducer.As the members of the AraC family are commonly transcription activators (32), it seems most plausible that the latter hypothesis is true.Mutations that result in the activation of otherwise inactive activators are often designated as * mutations.Therefore, we have used the designation aerA* for these suppressor mutations hereafter.

Overexpression of the EamA family protein AexA is responsible for the tolerance to β-alanine
AraC regulators often control the expression of single genes or operons in their direct genomic vicinity (32).The inspection of the aerA genomic context revealed the presence of the aexA (ydeD) and ydzE genes up-and downstream of aerA, respectively.The AexA protein is a protein of unknown function that consists of two conserved EamA domains.In the COG database (33), these proteins are annotated as permeases of the drug/metabolite transporter superfamily.In contrast, the YdzE gene consists of only one truncated EamA domain, and the Uniprot database suggests that the ydzE gene might be a pseudogene.As a putative permease, AexA might be involved in β-alanine homeostasis as an importer or an exporter.
To test whether AexA is involved in β-alanine transport, and whether it acts as an importer or an exporter, we deleted the aexA gene in the wild-type strain and in the aerA* suppressor mutant and tested the growth of the resulting mutants GP3955 and GP3958 in the presence of β-alanine.Typically, importer mutants show improved growth in the presence of toxic substrates, whereas exporter mutants are as sensitive as the parent strain (12,15,18,19).As shown in Fig. 3A, the deletion of aexA in the wild-type background had no effect on the sensitivity of the bacteria toward β-alanine.In contrast, the deletion of aexA in the aerA* mutant resulted in the complete loss of the resistance to β-alanine.These observations suggest that AexA most likely acts as an exporter.
If the aerA* mutation would result in the expression of the otherwise silent aexA gene, artificial overexpression of the aexA gene would be expected to provide tolerance to β-alanine.Therefore, we tested the effect of aexA overexpression on the growth in the presence of β-alanine.For this purpose, we used a strain carrying plasmid pGP3727 that allows plasmid-borne constitutive expression of aexA in a strain carrying the wild-type was tested.The cells were grown in C-Glc minimal medium to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing no β-alanine or 100 mM β-alanine and incubated at 37°C for 48 h.(B) The sensitivity of the wild type harboring either the empty vector pBQ200 or pGP3727 (aexA hy ) to β-alanine was tested.The cells were grown in C-Glc minimal medium (with erythromycin and lincomycin) to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing no β-alanine or 100 mM β-alanine and incubated at 37°C for 48 h.aerA allele.As shown in Fig. 3B, the overexpression of AexA indeed resulted in aerA-inde pendent tolerance to β-alanine.Taken together, our findings suggest that AexA is a β-alanine exporter that depends on the mutant AerA* activator protein for expression.

The aerA* mutation allows inducer-independent activation of the aexA promoter
The results presented above suggest that the mutations affecting the AerA protein might result in an inducer-independent activation of AerA that is required for the expression of the otherwise silent aexA gene.Previously, the expression of each B. subtilis gene has been analyzed under 104 different conditions, including different complex and minimal media (34).The aexA gene is barely expressed under all tested conditions (see Fig. 4A).Based on the average expression under all 104 conditions, aexA is ranked 3,799th of 4,266 genes.Thus, the condition under which aexA is actively expressed has yet to be identified.
To test the role of AerA* in the control of AexA expression, we fused the promoter regions of both the aerA and aexA genes to a promoterless lacZ reporter gene in the background of a wild-type strain and the isogenic aerA* and ΔaerA mutant strains.The bacteria were grown in C glucose minimal medium, and the β-galactosidase activities were determined.Similar results were obtained with the aexA-lacZ fusion that reports the activity of the aexA promoter (see Fig. 4B).In agreement with the global expression data (Fig. 4A), nearly no expression was observed in the wild-type strain GP4132.In contrast, the AerA* protein in GP4438 allowed very high promoter activity (3,210 units per mg of protein) as compared to the expression of genes of central carbon metabolism (35).The deletion of the aerA gene in GP4134 resulted in a complete loss of aexA expression (Fig. 4B).These data confirm our hypothesis that AerA is a transcription factor that, upon acquisition of an activating * mutation, activates transcription of its own gene and the divergent aexA gene.Thus, the suppressor mutation causes tolerance to β-alanine by causing the very high expression of the otherwise silent aexA gene.
Often, gene expression can be induced by specific substrates.We, therefore, also tested the activity of the aexA promoter in the three strains during growth in the presence of an amino acid mixture (Casamino acids) and of β-alanine.As shown in Fig.
FIG 4 The aexA gene is only expressed in the aerA * mutant.A. The Expression Browser of SubtiWiki allows a direct comparison of the expres sion levels of two or more genes (here aexA, cspD, and aimA).The expression level of cspD serves as an example of a gene with high constit utive expression.For aimA, a moderate expression is observed, while the aexA gene is virtually never expressed.For details on the conditions displayed on the X-axis, please consult the interactive expression browser of the SubtiWiki database (http://www.subtiwiki.uni-goettingen.de/v4/expression?gene=3BD05DE3ED5B293980F47EB621C54C7339F78DC7, 29).(B) The expression of the aexA promoter was monitored in strains that harbor the aexA-lacZ reporter gene fusion integrated into the chromosomal amyE gene.The tested strains were wild type, the aerA * mutant, as well as the deletion mutant ΔaerA.
Cultures were grown in C-Glc minimal medium supplemented with or without casaminoacids (CAA; magenta) or β-alanine (cyan) to the early exponential phase (OD 578 of about 0.6-0.8)and then harvested for β-galactosidase enzyme activity assays.The values for the β-galactosidase activity indicated for each strain represent three independently grown cultures, and for each sample, enzyme activity was determined twice.
4B, the addition of β-alanine resulted in a slight increase of promoter activity (6 units per mg of protein to 12 units for the wild-type strain GP4132).In contrast, the presence of casamino acids had no effect.Thus, the expression of aexA remained very low even in the presence of β-alanine.
Many regulators control the expression of their own genes.The aerA promoter showed little activity in the wild-type strain GP4439 (14 units per mg of protein).A significant increase to 42 units was observed in the aerA* mutant GP4440.These data demonstrate that the presence of the AerA* protein results in significant transcriptional activation of its own promoter.

AexA is a β-alanine exporter
All our genetic data presented above strongly suggest that AexA, if expressed, is an exporter for β-alanine.To verify this idea, we investigated the intracellular accumulation of this amino acid by using a radioactively labeled β-alanine.As shown in Fig. 5, both the wild-type strain 168 and the isogenic aerA* ΔaexA mutant GP3958 accumulated β-alanine in their cells.This observation is also a direct evidence for the existence of uptake system(s) for β-alanine in B. subtilis.In contrast, the aerA* mutant GP4117 that exhibits a very high level of aexA expression did not accumulate intracellular β-alanine.This is in excellent agreement with the fact that the aexA gene is required for β-alanine resistance even in the aerA* mutant (see Fig. 3).Together with the observation that the overexpression of AexA is causative for the tolerance to β-alanine (see above, Fig. 3), the lack of intracellular accumulation of β-alanine upon AexA overexpression clearly demonstrates that AexA indeed acts as an exporter for β-alanine.To the best of our knowledge, this is the first identified exporter for a β-amino acid.

AexA is a general exporter for pyruvate-based C3-amino acids
Amino acid transporters, both importers and exporters, tend to be promiscuous and can transport multiple amino acids (24).This has been extensively studied for the amino acid importer AimA and the exporter AzlCD which are involved in either the uptake of serine, glutamate, and asparagine or in the export of 4-azaleucine, histidine, and asparagine (12,15,18,19,23).We, therefore, considered the possibility that AexA might also export amino acids other than β-alanine.As all amino acids that are chemically derived from pyruvate are toxic for B. subtilis [(12); Fig. 1], we tested the tolerance of a strain overexpressing plasmid-borne AexA to these amino acids, i.e., serine, L-alanine, and 2,3-diaminopropionic acid.As shown in Fig. 6, AexA expression resulted in increased tolerance to all three amino acids indicating that AexA is also active in the export of these amino acids.The contribution of an aerA* mutation to serine resistance has been described in a recent study (23), and the results reported here demonstrate that this is caused by the increased expression of AexA in the mutant.We can thus conclude that AexA is a non-specific amino acid exporter for C3-amino acids.

2,3-Diaminopropionic acid acts as a weak natural inducer of AerA activity
Of the pyruvate-based amino acids, 2,3-diaminopropionic acid has the most severe growth-inhibiting effect (see Fig. 1).We, therefore, analyzed the functional link between this amino acid and AerA/AexA in more detail.First, we recorded the growth of the wild-type strain 168 and the ΔaexA mutant GP3955 in the absence and presence of 1 mM 2,3-diaminopropionic acid.As shown in Fig. 7A, the growth of both strains was identical in the absence of the amino acid.In the presence of 2,3-diaminopropionic acid, the wild-type strain was able to grow, however, more slowly than in its absence.In contrast, the ΔaexA mutant did not grow at all even at this low concentration of 2,3-diaminopropionic acid.This observation confirms the role of AexA in conferring tolerance to 2,3-diaminopropionic acid but also suggests that the AerA is able to provide sufficient activation of AexA expression to allow growth of B. subtilis in the presence of 2,3-diaminopropionic acid.To test this idea, we determined the activity of the AerA-dependent aexA-lacZ fusion in strain GP4132 in the absence and presence of 2,3-diaminopropionic acid (Fig. 7B).While the aexA promoter was inactive in the absence of 2,3-diaminopropionic acid, we observed a 20-fold induction in the presence of the amino acid to 120 units per mg of protein.Even though this promoter activity is much lower than what has been observed with the mutant AerA* protein, it is statistically significant and biologically relevant since this expression is sufficient to provide a basic resistance against 2,3-diaminopropionic acid in B. subtilis.Thus, 2,3-diaminopropionic acid might not only be the inducer for AerA but also a natural substrate for AexA.

Several amino acids protect the cell from the toxic effect of β-alanine
As described above (see Fig. 1), β-alanine is highly toxic in minimal medium, but it is tolerated to high concentrations in complex medium that contains all amino acids.This may indicate that β-alanine interferes with some amino acid biosynthetic pathways, as observed for serine (12).Alternatively, the other proteinogenic amino acids might compete with β-alanine for uptake capacity, thus preventing it from entering the cell.To get further insights into the protection of the cell from toxic β-alanine, we cultivated B. subtilis 168 in the presence of β-alanine and each of the proteinogenic amino acids.This analysis revealed that alanine, arginine, isoleucine, and methionine significantly protected the bacteria from β-alanine toxicity (see Fig. 8A).In addition, glycine, proline, and valine had a weak protective effect (see Fig. 8A), whereas the other amino acids did not provide protection.The fact that multiple amino acids protect the cell from β-alanine toxicity suggests that the protection results from competition for the uptake FIG 6 Overexpression of AexA confers resistance against L-serine, L-alanine, and 2,3-diaminopropionic acid.The sensitivity of the wild type harboring the empty vector pBQ200 or pGP3727 aexA hy to L-serine, L-alanine, or 2,3-diaminopropionic acid was tested.The cells were grown in C-Glc minimal medium to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing 1 mM L-serine, 80 mM L-alanine, or 1 mM 2,3-diaminopropionic acid and incubated at 37°C for 48 h.capacity rather than from inhibition of specific biosynthetic pathways.We, therefore, tested the effect of several amino acids on the uptake of β-alanine into the cell.Indeed, we observed a reduced uptake of β-alanine in the presence of alanine, arginine, isoleucine, and methionine, whereas glycine, proline, and valine failed to interfere with β-alanine incorporation (see Fig. 8B).These results correlate well with the results of the plating assay: those amino acids that provide the strongest protection against β-alanine compete efficiently for the transporter capacity.Thus, β-alanine shares the uptake systems with alanine, arginine, isoleucine, and methionine.

AlaP and AimA are the importers for both L-α-alanine and β-alanine
As described above, β-alanine and alanine seem to share their transporter(s).A recent study has identified the AlaP protein as the transporter for D-α-alanine (36).Based on the structural similarity of both amino acids and on the observed competition for uptake capacity between them (see Fig. 8B), we hypothesized that AlaP might also be involved in the transport of β-alanine.Moreover, our suppressor screen already identified a mutation in the low-affinity amino acid importer AimA that results in resistance to very high concentrations of β-alanine that are not even tolerated if the amino acid exporter AexA is expressed due to the aerA* mutation (GP4121, see Fig. 2).This observation suggests that AimA also imports β-alanine in addition to serine, glutamate, and aspartate (12,15,23).
To test the possible involvement of AlaP and AimA in the transport of both amino acids, we tested the growth of a wild-type strain and single and double mutants, alaP and aimA, in the presence of otherwise toxic concentrations of both amino acids.As shown in Fig. 8C, the alaP mutant GP4146 has a slightly higher tolerance toward both β-alanine and L-alanine than the wild-type strain, indicating that AlaP contributes to the uptake of both amino acids.The aimA mutant GP3590 exhibited tolerance toward both amino acids, even more pronounced as compared to the alaP mutant.While the lack of AimA did not provide full resistance to β-alanine as observed for the aerA* mutant, the aimA mutant fully tolerated the presence of L-alanine.Thus, our results suggest that both AlaP and AimA contribute to the uptake of both β-alanine and L-alanine.Indeed, the double mutant GP4135 showed increased resistance to both amino acids as compared to the two isogenic single mutants (see Fig. 8C).This confirms the idea that both AlaP and AimA contribute to the uptake of β-alanine and L-alanine.
To provide direct evidence for β-alanine uptake by AlaP and AimA, we measured the β-alanine transport of a wild-type strain (B.subtilis 168), an alaP mutant (GP4146), an  aimA mutant (GP3590), and an alaP aimA double mutant (GP4135).As shown in Fig. 8D, β-alanine was efficiently transported by the wild-type strain, whereas a severe reduction was observed in the alaP and aimA mutants.In the alaP aimA double mutant GP4135, the transport of β-alanine was nearly completely abolished (see Fig. 8D), confirming that AlaP and AimA are the main players in β-alanine uptake in B. subtilis.

AimA is responsible for the uptake of glycine
With the results presented above, amino acid importers have been identified in B. subtilis for all proteinogenic amino acids, with the exception of glycine, phenylalanine, and tyrosine.As AimA transports all C3 amino acids tested so far, we considered the possibility that it might also be involved in the uptake of the C2 amino acid glycine (see Fig. 8E).As observed for alanine, glycine also inhibited the growth of B. subtilis 168 on minimal medium.In contrast, the isogenic aimA mutant GP3590 was resistant to this amino acid, indicating that AimA is indeed the main glycine permease in B. subtilis.

AimA, AimB, and BcaP are the importers for 2,3-diaminopropionic acid
To get first insights into the uptake of 2,3-diaminopropionic acid, we isolated suppressor mutants that are able to grow in the presence of 2 mM of this amino acid.Since the exporter AexA already provides a weak natural resistance toward diaminopropionic acid, we used the aexA mutant GP3955 for suppressor isolation to ensure that we identify AexA-independent players in diaminopropionate homeostasis.The determination of the genome sequences of the five independent suppressor mutants revealed the presence of multiple mutations in each mutant (see Fig. 2A).Interestingly, three of the mutants had genomic deletions of 21-74 kb.In all these mutants, the deletion affected the region from ybdO to ybfI as the common denominator (see Fig. 2B).This region encodes two known amino acid transporters, AimA and AimB (previously YbxG).Moreover, these three strains had mutations affecting the amino acid transporter BcaP, in one case, a frameshift mutation that results in the synthesis of a truncated inactive protein.To study the role of these three amino acid transporters in providing resistance to 2,3-diaminopropionic acid, we tested the growth of the corresponding single, double, and triple mutants in the presence of this amino acid.As shown in Fig. 9A, the single mutants were still strongly inhibited by 2,3-diaminopropionic acid.The double mutants showed increased resistance which was most pronounced in the mutants lacking AimA in addition to either BcaP or AimB.The triple ΔaimA ΔbcaP ΔaimB mutant GP4495 had the highest resistance, with growth only weakly affected.These observations demonstrate that all three transporters, AimA, BcaP, and AimB, transport 2,3-diaminopropionic acid into the cell; thus, the inactivation of all three, as observed in all suppressor mutants, is required for a high-level resistance of B. subtilis to 2,3-diaminopropionic acid.The identification of diaminopropionate as an additional substrate for AimB suggests that this protein also is a general broad-spectrum amino acid permease.
The remaining two 2,3-diaminopropionic acid-tolerant mutants both had mutations affecting the transcription repressor AzlB that controls the expression of the well-studied amino acid exporter AzlCD (18,19,23) (see Fig. 2).The azlB mutations result in high expression of the otherwise poorly expressed broad-spectrum amino acid exporter AzlCD.To test the role of AzlB and the AzlCD amino acid transporter in conferring resistance to 2,3-diaminopropionic acid, we compared the growth of the azlB and azlBCD mutants in the genetic backgrounds of a wild-type strain and of an aexA mutant that is unable to export 2,3-diaminopropionic acid via AexA.(Fig. 9B).The latter mutant had been used to isolate the resistant mutants (see Fig. 2).In the wild-type background, the Azl system had no effect on the resistance toward 2,3-diaminopropionic acid, probably due to the presence of the natural exporter AexA.Indeed, an azlB mutant only confer red increased 2,3-diaminopropionic tolerance to the aexA mutant, and this tolerance depends on the presence of the azlCD exporter genes, indicating that the amino acid exporter AzlCD can contribute to 2,3-diaminopropionic acid export in addition to AexA.

DISCUSSION
The results presented in this study identify the transport systems for C2 and C3 amino acids in B. subtilis (see Fig. 10).All these amino acids can become toxic for B. subtilis if they are present at high concentrations.Thus, mutant characterization and suppressor screens are excellent tools to get insights into both the active and dormant transport systems.

Broad-range amino acid permeases can transport multiple amino acids
Our work identifies AimA as a major broad-range amino acid transporter in this B. subtilis.This permease is able to take up glycine, L-alanine, β-alanine, and 2,3-diaminopropionic acid in addition to glutamate, asparagine, and serine (12,15,23).This role is in good agreement with the constitutive expression of the aimA gene under a broad variety of conditions (see Fig. 4A) (34).As already observed for several other amino acids, the C3 amino acids are all transported by more than one permeases.In the case of L-alanine and β-alanine, this is AlaP which was also identified as a transporter for D-alanine (36).Surprisingly, 2,3-diaminopropionic acid can even be transported into the cell by three amino acid permeases, the promiscuous transporters AimB (previously YbxG) and BcaP in addition to AimA.These transporters have previously been shown to be required for the uptake of threonine and serine (12,20,37).In addition, BcaP also acts in the acquisition of asparagine and the branched-chain amino acids isoleucine and valine (23,37).It is tempting to speculate that 2,3-diaminopropionic acid, as a non-natural growth-inhibit ing amino acid, can be transported by multiple transporters, as no selectivity against this molecule is required in natural ecosystems.Interestingly, all four of these amino acid permeases involved in the uptake of alanine, β-alanine, and 2,3-diaminopropionic acid are members of the large amino acid-polyamine-organocation (APC) superfamily (38).These proteins typically transport amino acids, but they may also be involved in the uptake of unrelated substrates such as methyl-thioribose (MtrA) or potassium ions (KimA) (39,40).The APC family transporters are ubiquitous in all domains of life with typically between 20 and 30 members in each organism.Poor specificity seems to be a rather general characteristic of this family, suggesting that they are important for the acquisition of a large range of different amino acids and other compounds in archaea, bacteria, yeast, and man.However, the individual melange of these proteins differs from species to species.Interestingly, AimA seems to be limited to B. subtilis, as close orthologs of the protein are missing in most species, even in other Bacillus species.ΔaimA ΔbcaP (GP2949), ΔaimA ΔaimB (GP2951), ΔaimB ΔbcaP (GP2952), and ΔaimA ΔaimB ΔbcaP (GP2950) to 2,3-diaminopropionic acid was tested.The cells were grown in C-Glc minimal medium to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing no, 100 mM, or 200 mM β-alanine and incubated at 37°C for 48 h.(B) The sensitivity of the wild type (168), ΔazlB (GP3600), ΔazlBCD (GP3623), ΔaexA (GP3955), ΔaexA ΔazlB (GP4384), and ΔaexA ΔazlBCD (GP4385) to 2,3-diaminopropionic acid was tested.The cells were grown in C-Glc minimal medium to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing no, 100 mM, or 200 mM β-alanine and incubated at 37°C for 48 h.

The power of genetic screens: you get what you select for
It is interesting to note that three of the suppressor mutants that were resistant to 2,3-diaminopropionic acid contained large genomic deletions.In each case, the deletions resulted in the loss of AimA and AimB (see Fig. 2B).In two cases, GltP, a third putative amino acid permease that was shown to be involved in the transport of the antimetabo lite glyphosate (22), was also missing due to the deletions.However, in GP4456, which had the smallest deletion, the gltP gene was still present suggesting that it was not involved in the sensitivity to 2,3-diaminopropionic acid.In prior studies, numerous aimA mutants that were resistant to either serine, glutamate, or asparagine have been isolated.However, all of them just had single nucleotide polymorphisms or small deletions or insertions within the aimA gene but not deletions of complete genomic regions (12,15,23).The ubiquitous occurrence of deletions of the aimA-aimB genomic region indicates that the selective pressure was really directed to the loss of both transporters (in addition to BcaP) and underlines the power of genetic screens for the identification of genes and proteins that are involved in poorly studied functions so far.

Mutational activation of otherwise cryptic genes is a common theme in the adaptation to toxic amino acids
Previous studies have identified the AzlCD amino acid exporter to be required for the disposal of harmful amino acids in the cases of 4-azaleucine, histidine, and asparagine in B. subtilis, and this study indicates that AzlCD overexpression also contributes to the resistance to 2,3-diaminopropionic acid (18,19,23,41).In each case, the expression of the AzlCD exporter and subsequent resistance to these amino acids was acquired by the mutational inactivation of the AzlB repressor either by amino acid substitutions or by frame-shift mutations.Similarly, mutations in the AerA transcription activation trigger the expression of the AexA amino acid exporter.In this case, only point muta tions that allow DNA binding and subsequent transcription activation in the absence of any cofactor are possible.Similar mutations that render transcription activators effector-independent have been identified for the Crp and PrfA regulators in E. coli and Listeria monocytogenes that control the expression of genes for the utilization of secondary carbon sources and of virulence factors, respectively (42)(43)(44).Interestingly, in a recent study aimed at the identification of B. subtilis genes and proteins involved in the sensitivity to asparagine, a yetL mutant was observed (23).Strikingly, YetL is a MarR-type transcription factor that is encoded in the genome next to YetK, another EAM family transporter.It is tempting to speculate that the mutation resulted again in the effector-independent activity of YetL.

A strongly conserved group of poorly expressed "sleeping beauty amino acid exporters"
As mentioned above, AexA is a member of the EamA transporter family, named after the E. coli cysteine/O-acetylserine exporter EamA (45).This family is also classified as transporter family 2.A.7.3.(Drug/Metabolite Exporter Family with 10 membrane-span ning domains) (46).These proteins consist of two conserved EamA domains (47).According to the COG database (33), members of this family are ubiquitous in all organisms, with multiple proteins in most species.Characterized members of the family include EamA and YddG from E. coli, which export cysteine/O-acetylserine and aromatic amino acids, respectively (45,48).B. subtilis encodes eight members of the EamA transporter family.One particularly striking common feature of the EamA type transporter genes in B. subtilis is their low expression under all studied conditions [see Fig. 4A with aexA as a representative example; (34)].Most of the corresponding genes belong to 10% of the most poorly expressed genes in B. subtilis, and the ydfC gene, encoding one of these uncharacterized transporters, has the third position in the list of the most poorly expressed genes in B. subtilis.It is possible that the conditions that induce the expression of the EamA transporters have not been identified, as is the case for AexA which is induced in the presence of 2,3-diaminopropionic acid.Alternatively, it is possible that they are never expressed unless the cells experience a selective pressure that results in the acquisition of mutations that allow expression of the transporters, such as the presence of a toxic substrate.Another example is the EAM family member YetK, which may become expressed upon the acquisition of YetL* mutations in response to asparagine stress [ (23); see above].It is tempting to speculate that YetK is active in asparagine export.Based on amino acid export as the common function of these transporters and their commonly very low and only occasionally induced expression, we suggest dubbing this group of proteins "sleeping beauty amino acid exporters." With the identification of transporters for the C2 and C3 amino acids in B. subtilis, phenylalanine and tyrosine remain the last amino acids, for which no permease has so far been identified in this bacterium.While the field of amino acid uptake is thus coming to an end, the export of amino acid still remains an interesting area of research.It will be interesting to see how B. subtilis responds to the presence of other non-natural amino acids and whether the activation of sleeping beauty amino acid exporters helps in the adaptation to such molecules.

Strains, media, and growth conditions
E. coli DH5α (49) was used for cloning.All B. subtilis strains used in this study are derivatives of the laboratory strain 168.They are listed in Table 1.B. subtilis and E. coli were grown in LB or in sporulation medium (49,50).For growth assays, B. subtilis was cultivated in C minimal medium containing 0.5% glucose (51) or MSSM medium (52).MSSM is a modified SM medium in which KH 2 PO 4 was replaced by NaH 2 PO 4 , and KCl was added as indicated (52).The media were supplemented with ampicillin (100 µg/ mL), kanamycin (10 µg/mL), chloramphenicol (5 µg/mL), spectinomycin (150 µg/mL), or erythromycin and lincomycin (2 and 25 µg/mL, respectively) if required.

DNA manipulation and transformation
Transformation of E. coli and plasmid DNA extraction were performed using standard procedures (49).All commercially available plasmids, restriction enzymes, T4 DNA ligase, and DNA polymerases were used as recommended by the manufacturers.B. subtilis was transformed with plasmids, genomic DNA, or PCR products according to the two-step protocol (50).Transformants were selected on LB plates containing antibiotics suitable for the introduced construct.DNA fragments were purified using the QIAquick PCR Purification Kit (Qiagen, Hilden, Germany).DNA sequences were determined by the dideoxy chain termination method (49).

Construction of mutant strains by allelic replacement
Deletion of the aimA, aimB, and aexA genes was achieved by transformation of B. subtilis 168 with a PCR product constructed using oligonucleotides to amplify DNA fragments flanking the target genes and an appropriate intervening resistance cassette as described previously (54).The integrity of the regions flanking the integrated resistance cassette was verified by sequencing PCR products of about 1,100 bp amplified from the chromosomal DNA of the resulting mutant strains (see Table 1).

Phenotypic analysis
In B. subtilis, amylase activity was detected after growth on plates containing nutrient broth (7.5 g/L), 17 g Bacto agar per liter (Difco), and 5 g hydrolyzed starch per liter (Connaught).Starch degradation was detected by sublimating iodine onto the plates.Quantitative studies of lacZ expression in B. subtilis were performed as follows: cells were grown in C-glucose minimal medium.Cells were harvested at OD 600 of 0.5-0.8.β-Galactosidase-specific activities were determined with cell extracts obtained by lysozyme treatment as described previously (50).One unit of β-galactosidase is defined as the amount of enzyme that produces 1 nmol of o-nitrophenol per minute at 28°C.

Genome sequencing
To identify the mutations in the suppressor mutant strains (see Fig. 2; Table 1), the genomic DNA was subjected to whole-genome sequencing.Concentration and purity of the isolated DNA were first checked with a Nanodrop ND-1000 (PeqLab Erlangen, Germany), and the precise concentration was determined using the Qubit dsDNA HS Assay Kit as recommended by the manufacturer (Life Technologies GmbH, Darmstadt, Germany).Illumina shotgun libraries were prepared using the Nextera XT DNA Sample Preparation Kit and subsequently sequenced on a MiSeq system with the reagent kit v3 with 600 cycles (Illumina, San Diego, CA, USA) as recommended by the manufacturer.The reads were mapped on the reference genome of B. subtilis 168 (GenBank accession number: NC_000964) (55).Mapping of the reads was performed using the Geneious software package (Biomatters Ltd., New Zealand) (56).Frequently occurring hitchhiker mutations (57) and silent mutations were omitted from the screen.The resulting genome sequences were compared to that of our in-house wild-type strain.Single nucleotide polymorphisms were considered as significant when the total coverage depth exceeded 25 reads with a variant frequency of ≥90%.All identified mutations were verified by PCR amplification and Sanger sequencing.

Plasmid constructions
The plasmids pAC6 and pAC7 (58,59) were used to construct translational fusions of the potential aexA and aerA promoter regions, respectively, to the promoterless lacZ gene.For this purpose, the promoter regions were amplified using oligonucleotides that attached EcoRI and BamHI restriction to the ends of the products.The fragments were cloned between the EcoRI and BamHI sites of the plasmids.The resulting plasmids were pGP3729 (aexA-lacZ) and pGP4010 (aerA-lacZ).
To allow for the ectopic expression of the aexA gene, we constructed the plasmid pGP3727, respectively.The corresponding gene was amplified using oligonucleotides that added SphI and SalI sites to the ends of the fragments and cloned into the replicative expression vector pBQ200 (60), which was linearized with the same enzymes.

Intracellular accumulation of radiolabeled β-alanine in B. subtilis
To study the intracellular accumulation of [1-14 C]β-alanine in cells of the B. subtilis, cultures were grown in C minimal medium containing glucose to exponential growth phase (OD 600 , 0.7-0.9),and β-alanine was spiked in with radiolabeled [1-14 C]β-ala nine (specific activity for both solutes, 50-60 mCi/mmol), which was added to a final concentration of 30 mM.Uptake of the amino acid was followed at 0, 2, 4, 6, 8, 10, 20, and 40 minutes by pipetting 500 µL of the culture on a filter.The cells were washed three times with 1 mL of C minimal medium, and the filter was transferred to a scintillation vial.The filters were completely dried for 30 minutes at 70°C before 5 mL of Ecoscint A liquid scintillation cocktail (National Diagnostics, USA) was added, and the radioactivity was measured using a HIDEX 300SL β-particle scintillation detector.For the controls, 500 µL of C Glc with 30 mM β-alanine with and without radiolabeled [1-14 C]β-alanine was pipetted directly into the scintillation vial and let dried completely at 70°C.

Competitive uptake of radiolabeled β-alanine in B. subtilis in the presence of amino acids
The competitive uptake of β-alanine in the presence of other amino acids L-alanine, L-arginine, L-isoleucine, L-methionine, L-glutamate, glycine, L-proline, and L-valine was measured.The B. subtilis wild-type strain 168 was cultivated in C minimal medium containing glucose to exponential growth phase (OD 600 , 0.7-0.9),and β-alanine spiked with radiolabeled [1-14 C]β-alanine (specific activity for both solutes, 50-60 mCi/mmol), together with or without the respective amino acid, was added to a final concentration of 30 mM.A sample was taken after 40 minutes by pipetting 500 µL of the culture on a filter.The cells were washed three times with 1 mL of C minimal medium, and the filter was transferred to a scintillation vial.The filters were completely dried for 30 minutes at 70°C before 5 mL of Ecoscint A liquid scintillation cocktail (National Diagnostics, USA) was added, and the radioactivity was measured using a HIDEX 300SL β-particle scintillation detector.For the controls, 500 µL of C Glc with 30 mM β-alanine with and without radiolabeled [1-14 C]β-alanine was pipetted directly into the scintillation vial and let dried completely at 70°C.

FIG 2 A
FIG 2 A suppressor screen in the presence of alanine derivates.(A) Evolutionary trajectory of the wild-type strain exposed to toxic concentrations of β-alanine or the ΔaexA mutant exposed to toxic concentrations of 2,3-diaminopropionic acid.(B) Genetic regions affected by deletions in the suppressor strains (GP4454, GP4455, and GP4456) under 2,3-diaminopropionic acid stress.Each strain's deleted regions are demarcated, and an overlay highlights the shared deleted region, denoted by a blue-striped rectangle.This shared deleted region suggests a common genetic adaptation in response to the specific stress condition.

FIG 7 2 , 3 -
FIG 7 2,3-Diaminopropionic acid is a weak inducer of aexA expression.(A) The growth of the wild type and the ΔaexA mutant (GP3955) on C Glc minimal medium was tested in the presence or absence of 1 mM 2,3-diaminopropionic acid.(B) The expression of the aexA promoter was monitored in strains that harbor the aexA-lacZ reporter gene fusion integrated into the chromosomal amyE gene.Cultures were grown in C Glc minimal medium in the presence or absence of 2 mM 2,3-diaminopropionic acid to the early exponential phase (OD 578 of about 0.6-0.8)and then harvested for β-galactosidase enzyme activity assays.The values for the β-galactosidase activity indicated for each strain represent three independently grown cultures, and for each sample, enzyme activity was determined twice.

FIG 8
FIG 8 The presence of different amino acids can prevent β-alanine toxicity.(A) The growth of the wild-type strain in the presence of β-alanine and either alanine, arginine, isoleucine, methionine, glutamate, glycine, proline, or valine was checked.The cells were grown in C-Glc minimal medium to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing 100 mM β-alanine together with 30 mM of the respective amino acid and incubated at 37°C for 48 h.(B) To check the relative uptake of β-alanine in the presence or absence of the tested amino acid, the wild type was grown in C-Glc minimal medium containing 30 mM β-alanine spiked with radiolabeled [1-14 C]β-alanine with or without 5 mM of the respective amino acid to an OD 600 of approximately 0.8.The accumulation of intracellular [1-14 C]β-alanine was assayed after 40 minutes.n = 3 (C) The sensitivity of the wild type 168, ΔalaP (GP4136), ΔaimA (GP3590), and ΔalaP ΔaimA (GP4135) to β-alanine or L-alanine was tested.The cells were grown in C-Glc minimal medium to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing 100 mM β-alanine or L-alanine and incubated at 37°C for 48 h.(D) For the quantification of intracellular accumulation of [1-14 C]β-alanine, cells of the wild type (168, black circles), ΔalaP (GP4136, cyan triangles pointing upward), ΔaimA (GP3590, magenta squares), and ΔalaP ΔaimA (GP4135, purple triangles pointing downward) mutant were cultivated in C-Glc minimal medium to an OD 600 of approximately 0.8.The cells were then diluted with C-Glc minimal medium containing [1-14 C]β-alanine to reach a final concentration of 30 mM.The accumulation of intracellular [1-14 C]β-alanine was assayed after 0, 2, 4, 6, 8, 10, 20, and 40 minutes.n = 3 (E) The sensitivity of the wild type 168 andthe ΔaimA (GP3590) to glycine was tested.The cells were grown in C-Glc minimal medium to an OD 600 of 1.0, and serial dilutions (10-fold) were prepared.These samples were plated on C-Glc minimal plates containing 200 mM glycine and incubated at 37°C for 48 h.

FIG 10
FIG 10 Transport of C2 and C3 amino acids in B. subtilis.The uptake and export systems for the C2 and C3 amino acids are shown in the left and right parts of the figure, respectively.The amino acid exporters AexA and AzlCD are only expressed if the transcription factors AerA and AzlB have acquired mutations that allow their DNA binding in the absence of the inducer or in the presence of 2,3-diaminopropionic acid (for AerA).