Loss of function of metabolic traits in typhoidal Salmonella without apparent genome degradation

ABSTRACT Salmonella enterica serovar Typhi and Paratyphi A are the cause of typhoid and paratyphoid fever in humans, which are systemic life-threatening illnesses. Both serovars are exclusively adapted to the human host, where they can cause life-long persistent infection. A distinct feature of these serovars is the presence of a relatively high number of degraded coding sequences coding for metabolic pathways, most likely a consequence of their adaptation to a single host. As a result of convergent evolution, these serovars shared many of the degraded coding sequences although often affecting different genes in the same metabolic pathway. However, there are several coding sequences that appear intact in one serovar while clearly degraded in the other, suggesting differences in their metabolic capabilities. Here, we examined the functionality of metabolic pathways that appear intact in S. Typhi but that show clear signs of degradation in S. Paratyphi A. We found that, in all cases, the existence of single amino acid substitutions in S. Typhi metabolic enzymes, transporters, or transcription regulators resulted in the inactivation of these metabolic pathways. Thus, the inability of S. Typhi to metabolize Glucose-6-Phosphate or 3-phosphoglyceric acid is due to the silencing of the expression of the genes encoding the transporters for these compounds due to point mutations in the transcriptional regulatory proteins. In contrast, its inability to utilize glucarate or galactarate is due to the presence of point mutations in the transporter and enzymes necessary for the metabolism of these sugars. These studies provide additional support for the concept of adaptive convergent evolution of these two human-adapted S. enterica serovars and highlight a limitation of bioinformatic approaches to predict metabolic capabilities. IMPORTANCE Salmonella enterica serovar Typhi and Paratyphi A are the cause of typhoid and paratyphoid fever in humans, which are systemic life-threatening illnesses. Both serovars can only infect the human host, where they can cause life-long persistent infection. Because of their adaptation to the human host, these bacterial pathogens have changed their metabolism, leading to the loss of their ability to utilize certain nutrients. In this study we examined the functionality of metabolic pathways that appear intact in S. Typhi but that show clear signs of degradation in S. Paratyphi A. We found that, in all cases, the existence of single amino acid substitutions in S. Typhi metabolic enzymes, transporters, or transcription regulators resulted in the inactivation of these metabolic pathways. These studies provide additional support for the concept of adaptive convergent evolution of these two human-adapted S. enterica serovars.

which exhibit different host specificity and pathogenicity.In addition, according to their pathogenesis, S. enterica serovars are often grouped into those that cause systemic infection (i.e., typhoidal serovars) and those that cause self-limiting gastroenteritis (i.e., non-typhoidal serovars) (6)(7)(8)(9).Salmonella Typhi and Salmonella Paratyphi A are the most common typhoidal serovars that affect humans, where they cause typhoid and paratyphoid fever, which affect approximately 20 million people every year resulting in approximately 150,000 deaths worldwide (3,6,(10)(11)(12).A subset of individuals who recover from these illnesses harbor these organisms in the gall bladder for life, where they serve as reservoirs for future infections.
Salmonella Typhi and Salmonella Paratyphi A arose independently through evolution from their non-typhoidal common ancestors (13,14).However, up to a quarter of their genomic material was subsequently exchanged through recombination (15).The process of their adaptation to a single host combined with their unique lifestyle is partially reflected by the presence of a significantly higher number of degraded coding sequen ces or pseudogenes coding for virulence factors (e.g., type III secretion effectors) or various metabolic pathways (e.g., central anaerobic metabolism) when compared to broad host S. enterica serovars (13)(14)(15).Importantly, genome degradation is observed in areas of the chromosome that were not subject to exchange, thus exhibiting evidence of adaptive convergent evolution.Comparison of degraded coding sequences in different serovars can be a useful tool to gain insight into specific adaptations to a given niche.In fact, this type of analysis has allowed the identification of metabolic signatures specifically associated with serovars that cause gastroenteritis or systemic infection (16,17).
S. Typhi and S. Paratyphi A exhibit significant differences in the set of degraded coding sequences, in particular, several genes that encode metabolic pathways (13,14).This is surprising because these organisms share the same niche (i.e., the human host) and exhibit indistinguishable clinical presentation and pathogenic features (i.e., both serovars cause typhoid fever and systemic and persistent infection) (18).Here we have examined coding sequences that are degraded in one typhoidal serovar (i.e., S. Paratyphi A) but that are apparently intact in another (i.e., S. Typhi).We found that in the cases examined, specific point mutations in the S. Typhi metabolic genes or their regulators have rendered them non-functional.These results further support the concept of adaptive convergent evolution shaping the metabolic pathways of typhoidal Salmonella.In addition, these findings uncover a limitation of comparative genomic approaches to detect phenotypic differences between related bacteria.

S. Typhi and S. Paratyphi A show differences in degraded coding sequences for metabolic pathways
Based on previously published comparisons between typhoidal and non-typhoidal Salmonellae (13,14,16,17), and using the S. Typhimurium LT2 strain as reference, we compiled a list of genes encoding well-characterized or putative metabolic pathways that are either absent or annotated as pseudogenes in S. Typhi (strains CT18 and Ty2) or Paratyphi A (strains ATCC 9150 and AKU12601) (Table S1).As previously noted (13)(14)(15)19), and in line with the concept of convergent evolution during adaptation to extraintestinal infection in the human host, there is significant overlap in the metabolic pathways that have been silenced in both organisms.However, this inactivation often occurs through the degradation of different genes within the same pathway.Intriguingly and as noted before (13), there are instances of metabolic pathways lost in one serovar but not in the other, where the coding genes appear to be intact.For example, this is the case for the glucose-6-phosphate, 3-phosphoglyceric acid, and glucarate/galacta rate utilization pathways in S. Typhi, which have been degraded in S. Paratyphi A, or the galactitol and L-rhamnose utilization pathways in S. Paratyphi A, which have been degraded in S. Typhi (Table S1).This observation is intriguing as it suggests potential metabolic differences in these bacterial pathogens.In turn, these differences may reflect partially distinct environments during their pathogenic cycles, which could have potentially influenced the selection of different metabolic capabilities.This would be surprising considering their shared mechanisms of pathogenesis, clinical manifestations, and the ability to cause persistent infection.On the other hand, this observation raises the question whether these apparently intact pathways are truly operational.To address this issue, we investigated the functionality of a set of metabolic pathways where the coding genes are degraded in S. Paratyphi A but are putatively intact in S. Typhi.
Glucose-6-phosphate sensing and transcription regulation are impaired in S.

Typhi despite the absence of gene degradation
Glucose-6-phosphate (G6P) is abundantly present within the cytosol of eukaryotic host cells (20), although it is absent within the Salmonella-containing vacuole (SCV) (21).As some S. enterica serovars such as S. Typhimurium can gain access to the cytosolic environment (22), the observation of potential differences in the ability to utilize G6P by S. Typhi and S. Paratyphi A is of potential physiological significance.The import of G6P is mediated by the UhpT antiporter, whose expression is strictly controlled by an uncon ventional two-component regulatory system composed of the membrane-localized UhpC sensor and UhpB kinase/phosphatase, and the cytoplasmic response regulator UhpA (23,24).In this system, UhpC directly senses G6P triggering a conformational change in the interacting kinase UhpB, which in turn phosphorylates and activates the response regulator UhpA.Phosphorylation increases the affinity of UhpA for the uhpT promoter (PuhpT) thus initiating transcription (Fig. 1A).Although the uhpA, uhpB, uhpC, and uhpT coding sequences are apparently intact in S. Typhi, they are degraded in S. Paratyphi A (Table S1).
To investigate whether the G6P pathway was operational in S. Typhi we constructed a G6P biosensor/transcriptional reporter in which the PuhpT promoter drives the expres sion of super-folder GFP (sfGFP), as previously reported in S. Typhimurium (25-27) (Fig. S1A).We found that, as previously reported, the S. Typhimurium biosensor produced a robust fluorescence signal when the bacteria were grown in the presence of G6P.However, no signal was detected when S. Typhi encoding an equivalent biosensor was grown in the same media (Fig. 1B; Fig. S1B).Consistent with these observations, a bright fluorescence signal was detected in cytosolic S. Typhimurium after infection of HeLa cells, but no fluorescence signal was detected in cytosolic S. Typhi encoding an equivalent G6P biosensor (Fig. S1C).These results imply that despite the presence of apparently intact coding sequences, G6P sensing, and/or transcriptional regulation is impaired in S. Typhi.
Comparison of the nucleotide sequence of the uhp loci in S. Typhi and S. Typhimu rium detected an 80 bp insertion on the S. Typhi PuhpABC promoter as well as point mutations on the uhpA, uhpB, and uhpC coding sequences, which could be responsible for the observed differences in the expression of the G6P biosensor.We first investigated whether the 80 bp insertion in the S. Typhi PuhpABC promoter was responsible for the observed differences in the expression of the G6P biosensor.To this aim, we swapped the PuhpABC promoters in the S. Typhi and S. Typhimurium biosensors and quantified their expression in the presence of G6P by flow cytometry.We found robust expression of the G6P biosensor driven by the PuhpABC S. Typhi promoter in S. Typhimurium, but we observed no expression of the biosensor driven by the PuhpABC S. Typhimurium promoter in S. Typhi (Fig. S2A).Similar results were obtained with equivalent transcrip tional reporters driving the expression of nanoluciferase (Fig. S2B and C).These results indicated differences in the promoter sequences could not explain the lack of expression of the G6P biosensor in S. Typhi.
We then examined whether point mutations in the S. Typhi UhpC sensor, its interact ing kinase UhpB, or the response regulator UhpA were responsible for the inability of S. Typhi to respond to the presence of G6P.We constructed transcriptional reporters in which the S. Typhi uhpA, uhpB, or uhpC coding sequences were alternatively replaced with those from S. Typhimurium and expressed them in the S. Typhi and S. Typhimurium ∆uhpABC mutant backgrounds.We found that all combinations expressing the UhpA response regulator from S. Typhi were unable to activate the transcriptional reporter in the presence of G6P.In addition, the expression in S. Typhi or S. Typhimurium of uhpA from S. Typhimurium in the context of uhpB and uhpC from S. Typhi led to robust expression of the transcriptional reporter in the presence of G6P (Fig. 1C).These results indicated that the inability of the G6P transcriptional reporter to operate in S. Typhi is at least in part due to mutations within the response regulator UhpA.
In comparison to its S. Typhimurium homolog, S. Typhi UhpA (UhpA STy ) exhibits three amino acid changes: Val77Met, Ala146Glu, and Thr168Met (Fig. S3).We changed each one of the S. Typhi UhpA amino acids to the S. Typhimurium residues in the equivalent positions and examined the expression of the G6P transcriptional reporter.We found that simultaneously introducing in S. Typhi the uhpA Glu146Ala and uhpA Met168Thr mutations rescued the expression of the G6P transcriptional reporter indicating that these point mutations inactivated UhpA in S. Typhi (Fig. 1D).We also found that this S. Typhi mutant strain as well as an S. Typhi strain encoding the S. Typhimurium uhpA allele showed expression of the transcriptional reporter when gaining access to the cytosol of infected HeLa cells (Fig. 1E).These observations indicate that the mutations present in the S. Typhi uhpA allele have rendered this response regulator inactive.Alignment of the structures of UhpA STy (blue in Fig. 1F) and UhpA STm (pink in Fig. 1F) as predicted by AlphaFold2 revealed that the two mutations that led to the loss of function of UhpA in S. Typhi reside within its helix-turn-helix DNA-binding domain, indicating that, most likely, these point mutations impaired the ability of UhpA STy to activate the PuhpT promoter.Taken together, our results indicate that two amino acid substitutions within the response regulator UhpA have rendered it inactive in S. Typhi thus impairing G6P sensing and transcriptional regulation of the uhp genes.

Glucose-6-phosphate transport is impaired in S. Typhi due to a loss-of-func tion mutation in UhpT
Since the transcriptional regulation of the G6P utilization genes is impaired in S. Typhi due to mutations in the response regulator UhpA, it is predicted that the metabolism of this sugar would be impaired in S. Typhi.Consistent with this prediction and in contrast to S. Typhimurium, S. Typhi was unable to grow in the M9 minimal medium containing G6P as the sole carbon source.Intriguingly, S. Typhi expressing the S. Typhimurium UhpA response regulator was also unable to grow in minimal medium with G6P (Fig. 2A) suggesting that defects other than transcription regulation must also be contributing to the inability of S. Typhi to utilize this sugar.We therefore swapped the gene encoding the UhpT transporter for its counterpart from S. Typhimurium in the context of an S. Typhi strain that expressed the UhpA response regulator from S. Typhimurium.We found that the resulting mutant strain was able to robustly grow in the M9 medium supple mented with G6P as the only carbon source (Fig. 2B).In comparison to S. Typhimurium, S. Typhi UhpT exhibits four amino acid substitutions: Asp18Asn, Pro27Ser, Gly61Arg, and Ala395Asp (Fig. S4).To investigate which of the mutations resulted in the loss of function, we changed each one of the S. Typhi amino acids to those present in its S. Typhimurium homolog and expressed the resulting constructs in the context of an S. Typhi strain encoding the S. Typhimurium UhpA response regulator (which in S. Typhi carries a loss of function mutation, see above).We found that introducing in S. Typhi the uhpT Asp395Ala allele rescued the ability of S. Typhi to grow in the M9 minimal medium supplemented with G6P as the only carbon source (Fig. 2C).Interestingly, the structure of UhpT predicted by AlphaFold2 indicates that alanine 395 resides within an alpha helix that makes up the lumen of the predicted channel of the UhpT transporter (Fig. 2D).Therefore, mutation of this residue most likely impairs the transport function of UhpT.Taken together, these results indicate that the ability to utilize G6P by S. Typhi is not only impaired by a mutation in the response regulator that controls expression of the transporter, but also by a mutation in the transporter itself that impairs its function.
G6P has been hypothesized to serve as an important carbon source for S. Typhimu rium growth within the cytosol of infected cells.Since S. Typhi is also able to gain access to the cytosol, we investigated whether the inability to utilize G6P could affect its ability to grow within this compartment.To address this issue, we made use of the S. Typhi strain encoding UhpA and UhpT from S. Typhimurium, which can utilize G6P.Since growth within the cytosolic compartment is much faster than within the SCV, bacteria that can utilize the metabolites available in the cytosol may be able to grow better in this compartment, particularly at later times after infection.We found no difference in the total number of CFU of wild-type S. Typhi (unable to utilize G6P) with that of the strain expressing the S. Typhimurium UhpA and UhpT alleles, which can utilize this sugar (Fig. S5).Although this experiment did not directly measure growth in the cytosol, given that in HeLa cells late in infection, most of the CFU are generated in the cell cytosol (replication in the vacuole is substantially slower), these results argue that the inability to use G6P does not seem to substantially handicap the ability of S. Typhi to grow within the cytosol of cultured epithelial cells.

A loss of function mutation in the PgtA response regulator impairs 3-phos phoglyceric acid metabolism in S. Typhi
The metabolite 3-phosphoglyceric acid (3PG) is an important intermediate in the aerobic glycolysis pathway in eukaryotic cells (28).Importantly, stimulation of innate immune receptors in macrophages by bacterial ligands promotes the switch of glucose metab olism from oxidative phosphorylation to aerobic glycolysis (i.e., the "Warburg effect") with accumulation of the glycolytic intermediate 3PG (29,30).It has been recently reported that 3PG is essential for S. Typhimurium intracellular replication and systemic virulence (31).Consequently, potential differences in the ability to utilize 3PG by the typhoidal serovars S. Typhi and S. Paratyphi A could have important implications for their pathogenesis.The transport of 3PG in S. enterica is mediated by the PgtP trans porter, whose expression is controlled by the unconventional two-component regulatory system composed of the membrane-localized PgtC sensor and PgtB kinase/phosphatase, and the cytoplasmic PgtA response regulator (32, 33) (Fig. 3A).While the pgtB coding sequence is clearly degraded in S. Paratyphi A, all the genes required for 3PG utilization are apparently intact in S. Typhi (Table S1).
To investigate if the 3PG pathway is functional in S. Typhi, we designed a 3PG biosensor in which the S. Typhi PpgtP promoter drives the expression of sfGFP on a plasmid system.We then introduced the plasmid into both S. Typhimurium and S. Typhi.We observed that while S. Typhimurium harboring the 3PG biosensor produced a high fluorescence signal when the bacteria were grown in the presence of 3PG, no signal was detected when S. Typhi harboring the same plasmid was grown under identical condi tions (Fig. 3B).These results indicate that, while the S. Typhi PpgtP promoter is functional (there are only five nucleotide differences between S. Typhi and S. Typhimurium in the 435 bp region that contains the PpgtABC and PpgtP promoters), the sensing and/or transcriptional regulation of the 3PG utilization genes are impaired in S. Typhi.
We sought to identify which components of the 3PG sensing and metabolism were impaired in S. Typhi (Fig. 3A).Compared to its S. Typhimurium homologs, there are three amino acid substitutions in the S. Typhi PgtB kinase and the PgtA response regulator, while the PgtC sensor is identical in both serovars.Thus, we investigated whether the point mutations in PgtA or PgtB were the cause of the absence of signal for S. Typhi in the presence of 3PG.We constructed versions of the 3PG biosensor where the S. Typhi pgtA and pgtB coding sequences were replaced with those from S. Typhimurium and expressed them in an S. Typhi ΔpgtAB mutant strain.Expression in S. Typhi of pgtA from S. Typhimurium in the context of pgtB and pgtC from S. Typhi led to robust expression of the transcriptional reporter in the presence of 3PG (Fig. 3C; Fig. S6), confirming that mutations in pgtA impair the transcription regulation of 3PG utilization genes in S. Typhi.
In comparison to S. Typhimurium, the response regulator PgtA in S. Typhi (PgtA STy ) exhibits three amino acid substitutions: Glu141Asp, Gly173Asp, and His223Tyr (Fig. 3D; Fig. S7).We changed each one of these amino acids in S. Typhi for those present in S. Typhimurium and tested the expression of the 3PG transcriptional reporter.We observed that the introduction of the pgtA Asp173Gly mutation rescued the expression of the 3PG transcriptional reporter indicating that this point mutation inactivated PgtA in S. Typhi (Fig. S8A and B).The AlphaFold2 model of PgtA indicates that Gly173 is located within a domain that is predicted to interact with σ54 suggesting that mutations in this residue may prevent the interaction of the S. Typhi response regulator with the sigma factor thus precluding expression of the 3PG utilization genes (Fig. 3D).
In the case of the glucose-6-phosphate, the inability of S. Typhi to utilize this sugar is due not only to defects in sugar sensing and transcription regulation but also to mutations in the transporter itself.We therefore tested whether 3PG transport was impaired in S. Typhi by examining the growth of the S. Typhi pgtA STm or the S. Typhi pgtA Asp173Gly mutants in M9 containing 3PG as the only carbon source (Fig. 3E).Both strains were able to grow in this medium in a manner that was indistinguishable from S. Typhimurium wild type or an S. Typhimurium mutant expressing the S. Typhi 3PG transporter PgtP (Fig. 3F).Collectively, these results indicate that the inability of S. Typhi to utilize 3PG is due to the presence of a loss-of-function mutation in the response regulator PgtA.

S. Typhi is unable to metabolize glucarate and galactarate due to lossof-function mutations in the GudT transporter and the GudD and GarD dehydratases
In S. Paratyphi A, the genes involved in glucarate and galactarate utilization, gudT, ygcY, gudD, and garD, show clear signs of degradation and therefore are considered to be pseudogenes.In contrast, these coding sequences remain apparently intact in S. Typhi (Table S1).This observation is intriguing since utilization of these sugars has been proposed to be only important for the growth of S. enterica in the inflamed gut (34), which would not be relevant for S. Typhi, an extraintestinal pathogen.Therefore, we investigated whether these S. Typhi genes encode functional proteins.Glucarate and galactarate in S. enterica is transported by the permease GudT and subsequently dehydrated to 5-dehydro-4-deoxy-D-glucarate by the dehydratases GudD and GarD, respectively (35,36).Subsequently, 5-dehydro-4-deoxy-D-glucarate is metabolized into D-glycerate by the aldolase GarL and the reductase GarR, thus feeding the central metabolism.The expression of these metabolic genes is controlled by the transcription factor CdaR, which senses these sugars in the cytosol and activates the promoters for the genes responsible for their uptake and metabolism (Fig. 4A).To study the sensing and metabolism of glucarate and galactarate in S. Typhi, we first constructed biosensors consisting of the S. Typhi promoters for gudTYD (PgudTYD), gardD (PgarD), or garLRK (PgarLRK) driving the expression of sfGFP (Fig. S9A).We found that when these reporters were introduced into S. Typhimurium, they readily responded to the presence of glucarate and galactarate.In particular, the PgarLRK-sfGFP reporter showed the most robust response.However, this reporter did not respond to the presence of glucarate or galactarate in S. Typhi (Fig. S9B).This observation indicated that although the promoters for these genes seem to be functional, either their transcriptional control by the CdaR regulator, or the uptake of these sugars by the GudT transporter must be defective in S. Typhi.In addition to glucarate and galactarate, CdaR also responds to the presence of glycerate, which unlike the sugars, does not require specific transport as it diffuses through the membrane (35).Therefore, we tested the response of the PgarLRK-sfGFP reporter when S. Typhi was grown in the presence of glycerate (Fig. 4B).We found that under these growth conditions, the reporter displayed a robust response indicating that CdaR is functional in S. Typhi, and therefore the inability of this reporter strain to respond to glucarate and galactarate must be due to defects in the transport of these sugars through GudT.
To test this hypothesis, we constructed a strain of S. Typhi expressing the GudT homolog from S. Typhimurium and examined the response of the PgarLRK-sfGFP transcriptional reporter to the presence of glucarate or galactarate.In contrast to wildtype or the ∆gudT mutant strains, the S. Typhi strain expressing S. Typhimurium GudT showed a robust response of the PgarLRK-sfGFP transcriptional reporter in the presence of these sugars (Fig. 4C; Fig. S9C).The S. Typhi strain expressing the GudT transporter from S. Typhimurium was able to grow in M9 minimal medium containing glucarate or galactarate as the sole source of carbon.However, in comparison to S. Typhimurium, growth in galactarate was limited and growth in glucarate was preceded by an extended lag period (Fig. 4D; Fig. S9D).These results indicated that the glucarate and galactarate GudT transporter is defective in S. Typhi.Comparison of the amino acid sequence of GudT in S. Typhi and S. Typhimurium showed the presence of two amino acid changes at positions Ser321 (S.Typhi GudT Ser321Phe ) and Ala357 (S.Typhi GudT Ala357Val ) (Fig. S10).The structure of the GudT transporter predicted by AlphaFold2 showed that Ala 357 is located on an alpha helix that forms the putative channel of the GudT transporter (Fig. 4E).It is Typhimurium and S. Typhi or their isogenic mutants ΔgudT, ΔgudT, and gudT STm as indicated.Growth was monitored in M9 + 0.05% casamino acids medium (Continued on next page) therefore possible that the presence of a valine at this position may disrupt the GudT channel activity in S. Typhi.
The observation that, relative to S. Typhimurium, the S. Typhi strain expressing the GudT transporter from S. Typhimurium showed an extended lag period in media with galactarate or glucarate as the only carbon source raised the issue as to whether other elements of the glucarate/galactarate metabolism pathway may also been defective in S. Typhi.After its transport through the GudT transporter, glucarate and galactarate are metabolized by the GudD and GarD dehydratases, respectively.When compared to S. Typhimurium homologs, the amino acid sequences of the S. Typhi GudD and GarD exhibit several amino acid substitutions that could account for the observed deficiencies in the ability of S. Typhi to utilize these sugars: Ile43Val, Gly78Asp and Ala159Val in GudD, and Ser237Asn in GarD (Fig S11).To explore this possibility, we expressed S. Typhimurium GudD or GarD in an S. Typhi strain expressing the GudT transporter from S. Typhimurium.We found that the resulting strains were able to grow in M9 minimal media supplemented with glucarate or galactarate, respectively, as the only carbon source to levels equivalent to those of S. Typhimurium, although a lag period was also apparent (Fig. 4F; Fig. S12).Taken together these results indicate that S. Typhi is unable to utilize galactarate and glutarate due to mutations in the GudT transporter, as well as in GudD and GarD, which are critical enzymes for the metabolism of these sugars.Overall, these results provide further demonstration that the metabolic changes that occurred in S. Typhi presumably through the process of adaptation to a single host and a systemic type of infection included not only gene degradation but more subtle, specific changes that have led to the inactivation of critical metabolic genes.

DISCUSSION
The process of adaptation to persistent infection in a single host has resulted in a significant accumulation of degraded coding sequences or pseudogenes in the typhoidal S. enterica serovars S. Typhi and S. Paratyphi A (13,14).While on aver age broad-host non-typhoidal Salmonella serovars exhibit ~1% of degraded coding sequences, typhoidal Salmonellae carry significant higher number of pseudogenes (>4%).Comparison of the degraded coding sequences in Salmonella serovars that cause systemic or gastrointestinal infection has been useful to glean specific metabolic traits necessary for replication in these two different environments (17).This comparison has also revealed several coding sequences that although clearly degraded in S. Paratyphi A, appear to be intact in S. Typhi (13).Many of these genes encode important metabolic pathways, which could suggest potential differences in the host niches occupied by these two typhoidal S. enterica serovars.Given the high similarity between the pathogen esis of these two human-adapted typhoidal serovars, this observation was surprising.
To gain insight into this issue, we closely examined the functionality of a subset of S. Typhi genes that encode functions necessary for the utilization of key carbon sources: glucose-6-phosphate, which is abundant within the cytosol of mammalian cells (20); 3-phosphoglyceric acid, a product of aerobic glycolysis that accumulates in activated macrophages and has been shown to be a key signal to reprogram the metabolism of S. Typhimurium within macrophages (31); and glucarate and galactarate, which have been shown to be critical for the growth of S. Typhimurium in the inflamed gut (34).The genes related to the utilization of these carbon sources appear intact in S. Typhi, with none or very limited number of substitutions when their sequences are compared to the S. Typhimurium homologs.We found that, in all cases, while all the promoters of the transcriptional units are intact, the metabolic pathways encoded by these genes are inactive due to point mutations in one or more of the regulatory components, transporters, or metabolic enzymes.These mutations are highly conserved across the more than 1,000 S. Typhi isolates whose genome sequences are currently available in the NCBI databases.Thus, the inability to utilize G6P is due to two amino acid substitutions in the DNA-binding domain of the response regulator UhpA and a single amino acid substitution in the G6P transporter UhpT.Similarly, the inability to metabolize 3PG is due to a single amino acid substitution in the response regulator PgtA although the PgtP transporter retains its ability to function at wild-type levels, both when expressed in S. Typhi or S. Typhimurium.Therefore, the inability to utilize both G6P and 3PG by S. Typhi is due to the silencing of the expression of the genes encoding the necessary transporters due to point mutations in the regulatory components.In contrast, the presence of point mutations in the transporter and enzymes necessary for their metabolism is responsible for the inability of S. Typhi to utilize glucarate and galactarate.It is noteworthy that in this case, S. Typhi is still able to express these inactive enzymes and transporter.It is thought that genes that are either not expressed or that encode inactive products are eventually degraded due to the accumulation of mutations.This appears to be the case for the genes that encode the G6P, 3PG, and glucarate/galactarate utilization pathway in S. Paratyphi A. However, the observation that the homologous sequences remained intact in S. Typhi or that in one instance (i.e., the glucarate utilization pathway) are still expressed despite the presence of loss-of-function mutations suggests that the inactivation of these pathways must have been a relatively recent event.
Although our analyses involved only a subset of genes that are degraded in S. Paratyphi A but not in S. Typhi, it is likely that a similar situation may apply to other metabolic pathways that are degraded in one serovar but not in the other.For example, sty3536, which codes for a putative tartrate transporter, which is critical for tartrate utilization, is clearly degraded in S. Paratyphi A but appears intact in S. Typhi.However, both S. Typhi and S. Paratyphi A are unable to metabolize tartaric acid (Fig. S13).Comparison of the amino acid sequence of S. Typhi STY3536 with its homolog in S. Typhimurium, which can utilize tartaric acid, reveals the presence of one amino acid substitution that may lead to loss of function (Fig. S14).Furthermore, amino acid differences between the S. Typhi and S. Typhimurium homologs are also present within STY3537 and STY3538, which encode putative transcription regulators of this metabolic pathway (Fig. S14B and C).Similarly, genes coding for regulators, transporters, and enzymes important for the metabolism of aspartate are clearly degraded in S. Paraty phi A but remain putatively intact in S. Typhi.For example, dcuS, which encodes the sensor-kinase of the two-component regulatory system that regulates the DcuABC-medi ated uptake of aspartate (37), has point mutations in S. Typhi compared to its homolog in S. Typhimurium (Fig. S15A).Amino acid differences between the S. Typhi and S. Typhimurium homologs are also present on the aspartate chemo-attractant tar and on the enzyme AsnB (Fig. S15B and C) (38).Point mutations in these proteins may lead to loss of function and could explain S. Typhi's inability to utilize aspartate.Another example is hutU, which encodes a hydratase that is essential for histidine utilization (39).This gene is degraded in S. Paratyphi A but appears intact in S. Typhi.However, S. Typhi was not able to utilize histidine, suggesting that the presence of single amino acid substitutions in S. Typhi HutU and/or the transcription regulator HutC relative to functional homologs are likely to be responsible for its inability to metabolize this amino acid (Fig. S16).
Our analyses confirm the remarkable convergence of the metabolism of the typhoidal serovar S. Typhi and S. Paratyphi A during the process of adaptation to the human host, suggesting a strong selection for the elimination of specific metabolic pathways.Interestingly, these metabolic pathways also appear degraded in other host-adapted serovars such as S. Cholerasuis, S. Gallinarum, and S. Dublin (Table S1).It is possible that the absence of these metabolites in the specific environments of the specific hosts where these pathogens may reside during persistent infection may provide the selection pressure to result in the loss of these metabolic capabilities.However, this remarkable convergence may indicate that the utilization of these carbon sources may somehow be deleterious for some not understood aspects of their physiology when in their specific hosts.More experiments will be required to test this hypothesis.Finally, our analyses also highlight a limitation in the prediction of metabolic capabilities based only on bioinformatic analyses.

Bacteria strains and cell lines
The Salmonella strains used in the experiments conducted in this study were derived from the S. enterica serovar Typhimurium strain SL1344 (40), and S. enterica serovar Typhi strain ISP2825 (41) and S. enterica serovar Paratyphi A strain 3343A-5 (Roy Curtiss III, strain collection) listed in Table S4.Mutants were constructed using standard recombinant DNA and allelic exchange using R6K suicide plasmids (42) in Escherichia coli ß2163 Δnic35 as donor strain (43).Strains were routinely cultured in LB broth at 37°C.The experiments using cultured cells were conducted using the HeLa human epithelial cell line.The cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 10% bovine calf serum (BCS) at 37°C with 5% CO 2 in a humidified incubator.coding sequence of each operon) from the S. Typhi ISP2825 genome and cloning them upstream to an sfGFP gene into the pAJM.657p15A ORI backbone, generating the vectors pSB6407, pSB6408, and pSB6409, respectively.To improve the signal levels, the PgarLRK STy -sfGFP biosensor was modified by replacing the native RBS with the phage G10 RBS, generating the glucarate/galactarate biosensor PgarLRK.G10-sfGFP vector pSB6534.For the expression of gudD, its coding sequences from S. Typhi or S. Typhimurium were cloned in the pAJM.657backbone downstream from the PgudTYD promoter consisting of 520 nucleotides upstream of the gudT plus 36 nucleotides of gudT's actual coding sequence.For the expression of garD, its coding and promoter sequences (i.e., 520 bp upstream of its initiation codon) from S. Typhi or S. Typhimurium were cloned in the pAJM.657backbone.

Transcriptional reporter assays
Wild-type and mutant strains of S. Typhimurium and S. Typhi harboring the different transcriptional reporters were grown at 37°C in LB supplemented with G6P (Roche) (concentrations ranging from 7.6 µM to 10 mM, as indicated), D-3-phosphoglyceric acid (3PG) (Sigma) (concentrations ranging from 1 mM to 10 mM, as indicated), glucaric acid (Sigma) (concentrations ranging from 1 mM to 10 mM, as indicated), or galactaric acid (Sigma) (concentrations ranging from 1 mM to 10 mM, as indicated) for at least 12 h.Incubation and agitation were carried out in final 200 µL volume in flat-bottom black wall 96-well plates (Costar) using a Spark multimode microplate reader (Tecan).The growth and sfGFP fluorescence signal were monitored with measurements at 30 min intervals (excitation wavelength of 485 nm, bandwidth of 15 nm; emission wavelength of 530 nm, bandwidth of 15 nm).The background fluorescence of the medium without bacteria was subtracted from all readings, which were then normalized by the bacterial OD 600 growth to give the final sfGFP RFU/OD 600 values.Alternatively, S. Typhimurium and S. Typhi and indicated mutants harboring the PuhpABC STm -NLuc or PuhpABC STy -NLuc transcriptional were grown for 20 h with agitation at 37°C, diluted 10× into sterile water and processed with the Nano-Glo Luciferase Assay System (Promega).The bioluminescence was measured using a Spark multimode microplate reader (Tecan).

Analytical flow cytometry
Wild-type S. Typhimurium and S. Typhi, as well as the indicated mutant derivatives harboring the G6P biosensor PuhpT-sfGFP were grown in the presence of G6P as described above.The cultures were then washed two times and re-suspended in phosphate-buffered saline (PBS) to a final concentration of ~10 6 bacteria/mL and the fluorescence intensity measured at 488 nm excitation, 530/33 emission was analyzed for 10,000 bacteria using a BD Accuri C6 flow cytometer.

Bacterial growth in minimal medium
Wild-type and mutant strains of S. Typhimurium and S. Typhi were grown in M9 minimal medium (M9 salts 6 g Na 2 HPO 4 , 3 g KH 2 PO 4 , 0.5 g NaCl, 1 g NH 4 Cl per liter; 2 mM MgSO 4 ; 0.1 mM CaCl 2; 50 µg/mL L-tryptophan; 50 µg/mL L-cysteine hydrochloride; and 50 µg/mL L-histidine) supplemented with 10 mM G6P, 10 mM 3 PG, 10 mM glucarate, or 10 mM galactarate as the only carbon source, as indicated.Except when strains were grown in the presence of G6P, the media were supplemented with casamino acids 0.05% (wt/vol).In all cases, the Salmonella strains were grown in LB medium overnight at 37°C until stationary phase.Cultures were washed two times in PBS and diluted to an OD 600 of 0.025 in the M9 minimal medium containing one of the specific carbon sources, and incubated on 96-well plates (Costar) at 37°C with orbital agitation for at least 12 h using the Spark multimode microplate reader (Tecan).Culture growth was monitored with measurements at 30 min intervals.The background of the medium without bacteria was subtracted from all readings to give the final OD 600 values.

Salmonella infections
For cultured HeLa cell infections, the different Salmonella strains were grown in a modified LB broth containing 0.3 M NaCl to induce the expression of the SPI-1 T3SS.Overnight cultures were diluted 1:20 and grown to an OD 600 of 0.9 (48).HeLa cells at a confluency of 80% were infected for 1 h with S. Typhimurium or S. Typhi in Hank's balanced salt solution (HBSS) at multiplicities of infection (MOIs) of 25 and 50, respec tively.The infected cells were washed three times with PBS and incubated in DMEM containing 100 µg/mL gentamicin to kill extracellular bacteria.After 1 h, the cells were washed once with PBS and further incubated in DMEM containing 10 µg/mL gentamicin for the remaining time of infection, as indicated in the figure legends.At the indicated times, cultures were lysed in 0.5 mL of PBS containing 0.2% Na-deoxycholate and the number of colony forming units (CFU) was determined by dilution plating in LB agar plates.Bacterial intracellular growth was determined by dividing the CFU at 8 h by the CFU at 2 h after infection.

Fluorescence microscopy
HeLa cells grown on glass coverslips were infected with S. Typhimurium or S. Typhi carrying a plasmid with the G6P biosensor/transcriptional reporter PuhpT-sfGFP as indicated above.At 6 h and 20 h post-infection, the samples were washed three times with PBS, fixed in 4% paraformaldehyde (PFA) for 15 min and washed with PBS again.The cells were then permeabilized with a buffer containing PBS, 3% BSA, and 0.3% Triton-X, stained with DAPI (Sigma) for 5 min and washed three times with PBS.Finally, the slides were mounted using ProLong antifade Mountaunt (Thermo) and left to dry for 24 h.Slides were imaged using an Eclipse TE2000 inverted microscope (Nikon) with an Andor Zyla 5.5 sCMOS camera controlled by the Micromanager software (https://www.micromanager.org).Brightness and contrast were optimized for each of the individual color channels to maximize visual clarity.

Genomic analysis
Degraded coding sequences (pseudogenes) for metabolic genes in S. Paratyphi A (strains ATCC 9150 and AKU_12601) that are putatively intact in S. Typhi (strains CT18 and Ty2) compared to the sequences in the reference S. Typhimurium strain LT2 were listed at Table S1 based on previously reported genomic comparisons (13,14,16,17).

Protein structural models
Protein structures were modeled using the AlphaFold2 notebook from ColabFold executing the default parameters (49).Visualization and structural alignment were carried out using PyMol (50).

Statistical analysis
Statistical significance was calculated either by ANOVA with Tukey's multiple compari sons test or Student's unpaired two-tailed t tests.For most experiments, values indicate the mean ± SEM with n = 3-4 replicates per condition.Further details are provided in the figure legends.

FIG 1
FIG 1 Glucose-6-phosphate sensing and transcriptional regulation are impaired in Salmonella Typhi.(A) Diagram of the glucose-6-phosphate (G6P) sensing, transcription regulation, and transport in Salmonella enterica.G6P (purple hexagon) is imported into the bacterial cell by the UhpT antiporter, whose expression is strictly controlled by an unconventional two-component regulatory system composed of the membrane-associated UhpC sensor and UhpB kinase/phospha tase, and the cytoplasmic response regulator UhpA.In this system, UhpC directly senses G6P triggering a conformational change in the interacting kinase UhpB, which in turn phosphorylates and activates the response regulator UhpA.Phosphorylation increases the affinity of UhpA for the uhpT promoter (PuhpT) thus initiating transcription.The potential steps where the sensing, regulation and transport could be impaired in S. Typhi are noted.(B) Transcriptional response (Continued on next page)

FIG 1 (
FIG 1 (Continued)of the S. Typhimurium and S. Typhi G6P biosensors after growth in LB medium containing G6P (500 µM).The sfGFP fluorescence signal for S. Typhimurium (circle on red bar) or S. Typhi (square on blue bar) in the presence or absence of G6P is shown.(C) Investigation of the functionality of UhpA, UhpB, and UhpC regulatory proteins.Each S. Typhi gene was swapped for its S. Typhimurium homolog in the plasmid backbone containing the G6P biosensor as indicated in the diagram.The resulting plasmids were then introduced into the S. Typhi or S. Typhimurium ΔuhpABC mutant backgrounds, and the resulting strains were grown in M9 medium in the presence or absence of G6P (1 mM).(D) Investigation of the functionality of each of the different uhpA mutations present in S. Typhi in comparison of its functional homolog in S. Typhimurium.Plasmids encoding the G6P biosensor and expressing S. Typhi UhpA mutants in which residues had been individually changed to the amino acid present in its S. Typhimurium homolog as indicated (i.e., UhpA Met77Val , UhpA Glu146Ala , or UhpA Met168Thr ) were introduced into the S. Typhi ΔuhpABC mutant background.The resulting strains were grown into M9 medium in the presence of G6P (1 mM).In panels B-D, the sfGFP fluorescence signal normalized to cell density (sfGFP RFU/OD 600 ) was measured.Values indicate the mean ± SEM of n = 3 or 4 replicates per condition.Asterisks denote statistically significant differences relative to the corresponding uninduced sample determined using ANOVA with Tukey's multiple comparisons test ***P < 0.001 (B and D), or unpaired two-tailed t-test ****P < 0.0001 (C).(E) Functionality of different UhpA variants expressed in S. Typhi in the context of cultured epithelial cell infection.HeLa cells were infected with wild-type S. Typhi or derivative strains expressing the uhpA allele from S. Typhimurium (uhpA STm ) or an S. Typhi allele, which had been rendered functional by introducing changes based on the S. Typhimurium sequence (i.e., uhpAGlu146Ala,Met168Thr ).Hela cells were infected with the different S. Typhi strains for 6 h, fixed, stained with DAPI, and examined under a fluorescence microscope.In all the images, brightness and contrast were optimized for each of the individual color channels to maximize visual clarity.Scale bars, 10 µm.(F) Alignment of the modeled atomic structures of S. Typhi UhpA (light blue) and S. Typhimurium UhpA (light pink) showing structural variations in the C-terminal portion at the helix-turn-helix motif (boxed).The models were predicted using AlphaFold2 and aligned with Pymol.The amino acid differences between the two structures (Val77Met, Ala146Glu, and Thr168Met) are shown in green.

FIG 2
FIG 2 Glucose-6-phosphate transport is impaired in S. Typhi due to a loss-of-function mutation in UhpT.(A-C) Growth kinetics of wild-type (WT) S. Typhimurium and S. Typhi or the indicated mutant strains.Bacteria were grown in the M9 medium containing 10 mM G6P as the only carbon source.The OD 600 values measured at 30 min intervals are shown.Values indicate the mean ± SEM of n = 3 replicates per condition.(D) Alignment of the modeled atomic structures of S. Typhi (dark blue) and S. Typhimurium (pink) UhpT predicted with AlphaFold2 and aligned with Pymol.The amino acids that differ between the two structures (Asp18Asn, Pro27Ser, Gly61Arg, and Ala395Asp) are shown in green.

FIG 3
FIG 3 The metabolism of 3-phosphoglycerate is impaired in Salmonella Typhi due to a loss of function mutation in the PgtA response regulator.(A) Diagram of 3PG sensing, transcription regulation, and transport in Salmonella enterica.PgtC senses 3PG (red trapezoid) triggering its interaction with the PgtB histidine kinase/phosphatase that phosphorylates the PgtA response regulator.Phosphorylated PgtA, in turn, activates the PgtpP promoter leading to the expression of the PgtP transporter.The potential steps where the sensing, regulation, and transport could be impaired in S. Typhi are noted.(B) Transcriptional response of the S. Typhimurium and S. Typhi strains encoding the 3PG biosensors after growth in LB medium containing 3PG (1 mM).The sfGFP fluorescence signal for S. Typhimurium (circle on red bar) or S. Typhi (square on blue bar) in the presence or absence of 3PG is shown.Values indicate the mean ± SEM of n = 3 replicates per condition.(C) Investigation of the functionality of the S. Typhi PgtA and PgtB regulatory proteins.The indicated S. Typhi genes were swapped for their S. Typhimurium homologs and the resulting mutant strains were tested for their ability to sense 3PG with the PpgtP-sfGFP biosensor.Bacteria were grown in LB medium for 20 h in the presence or absence of 3PG (1 mM), and the fluorescence signal (normalized to cell density) was measured (sfGFP RFU/OD 600 ) for each strain.Values in panels B and C are the mean ± SEM of n = 3-4 replicates per condition.Asterisks denote statistically significant differences relative to the corresponding uninduced sample determined using ANOVA with Tukey's multiple comparisons test.***P < 0.001.(D) Alignment of the modeled atomic structures of S. Typhi (blue) and S. Typhimurium (pink) PgtA predicted with AlphaFold2 and aligned with Pymol.The amino acids that differ between the two structures (Glu141Asp, Gly173Asp, and His223Tyr) are shown in green.(E and F) Growth kinetics of wild-type (WT) S. Typhimurium and S. Typhi or the indicated mutant strains.Bacteria were grown in M9 + 0.05% casamino acids medium containing 10 mM 3PG as the only carbon source.The OD 600 values measured at 30 min intervals are shown.Values indicate the mean ± SEM of n = 3 replicates per condition.

FIG 4
FIG 4 The glucarate and galactarate metabolism are impaired in Salmonella Typhi due to loss of function mutations in the GudT transporter and the GudD dehydratase.(A) Diagram of glucarate and galactarate sensing, transcription regulation, transport, and metabolism in Salmonella enterica.Glucarate (green circles) and galactarate (yellow squares) are transported into the bacterial cell by the permease GudT and converted to 5-dehydro-4-deoxy-D-glucarate (orange star) by the dehydratases GudD and GarD, respectively.5-Dehydro-4-deoxy-D-glucarate is further metabolized into D-glycerate (red triangles) by the aldolase GarL and the reductase GarR.The expression of these genes is controlled by the transcription factor CdaR, which senses glucarate, galactarate and D-glycerate in the cytosol and activates the PgudTYD, PgarD, and PgarLRK promoters.The potential steps where the sensing, regulation, transport, or metabolism could be impaired in S. Typhi are noted.(B and C) Wild-type S. Typhimurium and S. Typhi or the indicated mutant strains all harboring the PgarL_sfGFP transcriptional reporter were grown in LB medium containing glucarate, galactarate, or glycerate (as indicated) at 10 mM (or water as a negative control) and the fluorescence signal (normalized to cell density) was measured (sfGFP RFU/OD 600 ) for each strain.Values indicate the mean ± SEM of n = 3 replicates per condition.Asterisks denote statistically significant differences determined using ANOVA with Tukey's multiple comparisons test.***P < 0.001.(D) Growth kinetics of wild-type S.

FIG 4 (
FIG 4 (Continued) containing 10 mM glucarate as the only carbon source.The OD 600 measured at 30 min intervals are shown and values represent the mean ± SEM of n = 3-4 replicates per condition.(E) Alignment of the modeled atomic structures of S. Typhi (teal) and S. Typhimurium (beige) GudT predicted with AlphaFold2 and aligned with Pymol.The amino acids that differ between the two structures (Ser321Phe and Ala357Val) are shown in green.(F) Growth kinetics of S. Typhi gudT STm expressing plasmid-borne gudD STm and gudD STy , as indicated.Growth was monitored in M9 medium + 0.05% casamino acids containing 10 mM glucarate as the only carbon source.The OD 600 measured at 30 min intervals are shown and values represent the mean ± SEM of n = 3-4 replicates per condition.