ABSTRACT
Sulfolobus spp. possess a great metabolic versatility and grow heterotrophically on various carbon sources, such as different sugars and peptides. Known sugar transporters in Archaea predominantly belong to ABC transport systems. Although several ABC transporters for sugar uptake have been characterized in the crenarchaeon Sulfolobus solfataricus, only one homologue of these transporters, the maltose/maltooligomer transporter, could be identified in the closely related Sulfolobus acidocaldarius. Comparison of the transcriptome of S. acidocaldarius MW001 grown on peptides alone and peptides in the presence of d-xylose allowed for the identification of the ABC transporter for d-xylose and l-arabinose transport and the gaining of deeper insights into pentose catabolism under the respective growth conditions. The d-xylose/l-arabinose substrate binding protein (SBP) (Saci_2122) of the ABC transporter is unique in Archaea and shares more similarity to bacterial SBPs of the carbohydrate uptake transporter-2 (CUT2) family than to any characterized archaeal one. The identified pentose transporter is the first CUT2 family ABC transporter analyzed in the domain of Archaea. Single-gene deletion mutants of the ABC transporter subunits exemplified the importance of the transport system for d-xylose and l-arabinose uptake. Next to the transporter operon, enzymes of the aldolase-independent pentose catabolism branch were found to be upregulated in N-Z-Amine and d-xylose medium. The α-ketoglutarate semialdehyde dehydrogenase (KGSADH; Saci_1938) seemed not to be essential for growth on pentoses. However, the deletion mutant of the 2-keto-3-deoxyarabinoate/xylonate dehydratase (KDXD [also known as KDAD]; Saci_1939) was no longer able to catabolize d-xylose or l-arabinose, suggesting the absence of the aldolase-dependent branch in S. acidocaldarius.
IMPORTANCE Thermoacidophilic microorganisms are emerging model organisms for biotechnological applications, as their optimal growth conditions resemble conditions used in certain biotechnologies such as industrial plant waste degradation. Because of its high genome stability, Sulfolobus acidocaldarius is especially suited as a platform organism for such applications. For use in (ligno)cellulose degradation, it was important to understand pentose uptake and metabolism in S. acidocaldarius. This study revealed that only the aldolase-independent Weimberg pathway is required for growth of S. acidocaldarius MW001 on d-xylose and l-arabinose. Moreover, S. acidocaldarius employs a CUT2 ABC transporter for pentose uptake, which is more similar to bacterial than to archaeal ABC transporters. The identification of pentose-inducible promoters will expedite the metabolic engineering of S. acidocaldarius for its development into a platform organism for (ligno)cellulose degradation.
INTRODUCTION
In Archaea, most characterized sugar transporters predominantly belong to the class of ATP-binding cassette (ABC) transporters (1). So far, the only major facilitator superfamily (MFS) transporter studied in Archaea is the glucose transporter in the euryarchaeon Haloferax volcanii (2). A phosphotransferase system (PTS) system for d-fructose has been characterized in the same organism, and homologues were found in a few haloarchaea (3).
In the phylum of crenarchaeota, Sulfolobus solfataricus is the best-studied organism for sugar transport, and a number of sugar ABC transporters have been examined (4, 5). Five different sugar-binding proteins were identified belonging to CUT1 or the peptide/opine/nickel uptake transporter (PepT) family of ABC transporters: (i) d-/l-arabinose/d-fructose/d-xylose, (ii) d-galactose/d-glucose/d-mannose, (iii) cellobiose and higher cellooligosaccharides, (iv) maltose/maltodextrin, and (v) trehalose. The glucose-binding protein GlcS and the trehalose-binding protein TreS are constitutively expressed independently of the growth substrate. The maltose- and cellobiose-binding proteins MalE and CbtA showed moderate induction in the presence of the corresponding substrates. Expression of the arabinose binding protein AraS was strongly induced when S. solfataricus cells were grown on d-arabinose (4, 6–8). Although the promoter architecture of AraS was studied intensively and the cis-regulatory element “ara box” was identified in several promoter regions of genes upregulated in the presence of d-arabinose as the growth substrate, the corresponding transcriptional regulator has not been identified so far (9–11).
In the closely related crenarchaeon Sulfolobus acidocaldarius, only one homologue of the S. solfataricus sugar ABC transporters has been identified and functionally characterized as a maltose and maltooligomer transporter (12, 13). Co-utilization studies of d-glucose and d-xylose in S. acidocaldarius showed no carbon catabolite repression, as was reported for the glucose-induced inhibition of d-arabinose uptake in S. solfataricus (6, 14). Using microarray data of the co-utilization studies in S. acidocaldarius, the respective d-glucose and d-xylose transport systems were predicted (14) but not experimentally confirmed.
Once l-arabinose and d-xylose have entered the cell, Sulfolobus spp. oxidize both C5 sugars by means of the same glucose dehydrogenase that also has been shown to be promiscuous for the oxidation of d-glucose and d-galactose in the branched Entner-Doudoroff pathway (15–17). The resulting C5 sugar acids are further converted to 2-keto-3-deoxy-d-xylonate/l-arabinoate (KDX/KDA) by a d-xylonate/l-arabinoate dehydratase (l-XAD/l-AraD). KDX/KDA has been described to be further converted by an aldolase-dependent and an aldolase-independent pathway. In the aldolase-dependent pathway, KDA/KDX is cleaved to pyruvate and glycolaldehyde, which is further oxidized to glyoxylate via glycolaldehyde oxidoreductase and glycolate dehydrogenase. Malate synthase converts glyoxylate to malate, which enters the citric acid cycle. In the aldolase-independent pathway, KDX/KDA is first dehydrated to α-ketoglutarate semialdehyde (KGSA) and further converted to α-ketoglutarate (α-KG) via a KDX/KDA dehydratase and a KGSA dehydrogenase, respectively. α-KG then enters the citric acid cycle. Most enzyme activities of the branched pathway have been confirmed in cell extracts of S. solfataricus and S. acidocaldarius, and several of the enzymes have been biochemically characterized from S. solfataricus (some also from S. acidocaldarius). 13C labeling experiments in S. acidocaldarius grown on d-xylose revealed that both routes are operative in vivo at a 1:1 ratio (15).
The aim of this study was to investigate the uptake of d-xylose and l-arabinose as well as the regulation of the assimilation pathway for both pentoses in S. acidocaldarius MW001 by using a transcriptomic and mutational approach.
RESULTS
Identification of xylose-inducible transcripts in S. acidocaldarius using RNA-seq analysis.Growth experiments with S. acidocaldarius MW001 in liquid Brock medium with 0.2% (wt/vol) N-Z-Amine and 0.4% (wt/vol) d-xylose as the carbon and energy source compared to 0.2% (wt/vol) N-Z-Amine alone revealed a significant growth enhancement, as indicated by higher final cell densities in the presence of d-xylose (Fig. 1). To identify genes transcriptionally regulated in response to d-xylose and thereby to identify potential xylose transporters, RNA sequencing (RNA-seq) was performed with S. acidocaldarius cells grown under both nutrient conditions. Cells were harvested and RNA was isolated at mid-exponential growth phase (optical density at 600 nm [OD600], ∼0.5) (Fig. 1, arrow), where the growth rate of the two growth conditions was comparable.
Growth curves of S. acidocaldarius MW001 in Brock medium containing 0.2% (wt/vol) N-Z-Amine alone (closed circles) and 0.2% (wt/vol) N-Z-Amine and 0.4% (wt/vol) d-xylose (open circles). For the experiment, six replicates were performed for each growth condition. Cells of three replicates were harvested for RNA isolation at mid-exponential growth phase (OD600 of ∼0.5; marked with an arrow); growth of three additional replicates was further monitored.
Quantitative RNA determination and RNA data analysis using DESeq (23) revealed that only 10 genes were significantly (2-fold; P < 0.01) upregulated, and 17 genes were significantly downregulated in the presence of 0.4% (wt/vol) d-xylose compared to levels in cells grown only on 0.2% (wt/vol) N-Z-Amine. Genes more than 2.5-fold upregulated (log2-fold change, >1.35) or downregulated (log2-fold change, <−1.35) are summarized in Table 1. The highest transcriptional upregulation (>10-fold) was observed for the four genes saci_2122, saci_2121, saci_1938, and saci_1939.
Changes in transcript levels of S. acidocaldarius MW001 in response to d-xylose identified by RNA sequencinga
Bioinformatic analysis of saci_2121 and saci_2122 using BLASTP 2.3.0 (34) indicated that these genes encode an ABC transporter-associated transmembrane domain (TMD) and a substrate binding protein (SBP), respectively. An ABC transporter-associated nucleotide binding domain (NBD) was not annotated in the genomic neighborhood. However, an open reading frame (ORF) prediction search (https://www.ncbi.nlm.nih.gov/orffinder/) in the nearby genomic region of saci_2121 and saci_2122 revealed the presence of a gene located downstream of saci_2121. This gene encodes a putative ATPase and here is called saci_2120. The RNA-seq data confirmed that saci_2120 was cotranscribed with saci_2122 and saci_2121 as a single mRNA. saci_2120 transcripts were 8.97-fold (3.17 log2-fold) upregulated in cells growing in the presence of d-xylose (Table 1) compared to the cells grown only on N-Z-Amine.
Interestingly, synteny studies using the Archaeal and Bacterial Synteny Explorer (Absynte) webserver (http://archaea.u-psud.fr/absynte/) (35) showed that this operon is only conserved among all sequenced S. acidocaldarius strains (Fig. 2) and a few sequenced archaeal species, like some Thermophilum species, the two S. islandicus strains L S 2 15 and M 16 27, and the haloarchaea Halopiger xanaduensis and Halorhabdus tiamatea. However, the Absynte scores for archaeal species other than S. acidocaldarius all were below 15, which means very low synteny. Instead, the synteny of this operon appeared more conserved in bacterial species such as Sulfobacillus acidophilus, different Burkholderia species, Acidiphilium species, Rubrobacter xylanophilus, or Gluconacetobacter diazotrophicus, with Absynte scores of up to 40, suggesting that the ABC transport system was acquired by horizontal gene transfer from bacteria. The newly identified ABC transporter can be classified according to the Transporter Classification Database (http://www.tcdb.org/) (36) to the carbohydrate uptake transporter-2 (CUT2) family with a 10-transmembrane-spanning TMD subunit and only a single NBD, probably forming a homodimer.
Synteny of the d-xylose-induced ABC transporter-encoding gene cluster. Absynte scores represent TBLASTN scores of the 15-kb chromosome segments normalized to the BLASTP score of the query protein sequence as maximal bit score (35). Only bacterial sequences are closely related to the S. acidocaldarius genes, while archaeal hits only had a very low similarity that was below the value of relevant similarity, which is indicated by the dashed line. The numbers in the arrows indicate the respective ORF number.
ABC SBP expression is induced in the presence of pentoses.To confirm the transcriptional upregulation of the ABC substrate-binding protein (SBP) (saci_2122) and to identify other inducing sugars, quantitative RT-PCR (qRT-PCR) was performed. qRT-PCR was performed with RNA extracted from cultures grown in the presence of d-arabinose, l-arabinose, dextrin, d-glucose, sucrose, or d-xylose. SBP transcripts were highly upregulated in the same log2-fold range observed with DESeq when cells were grown in the presence of all tested pentoses, i.e., d-arabinose, l-arabinose, and d-xylose, but were not induced by the other tested sugars (Fig. 3). Together with the cotranscription revealed by RNA-seq data, this indicates that transcription of the entire ABC transport system is responsive to pentoses but not to hexoses (d-glucose), disaccharides (sucrose), or polysaccharides (dextrin). d-Arabinose was not tested in further experiments due to the lack of growth enhancement by this sugar.
Changes in SBP (saci_2122) transcript levels in response to the presence of different sugars in the growth medium. Differences in saci_2122 transcript amounts were determined by qRT-PCR. Bars indicate the sugar-specific transcript levels compared to transcription of cells grown only with N-Z-Amine on a log2-fold scale.
Isolation of the d-xylose-binding protein of S. acidocaldarius ABC transporter.It has been shown previously that sugar-binding proteins of S. solfataricus are N-glycosylated and therefore can be isolated via lectin affinity chromatography (4, 5). To confirm the induction of the predicted sugar binding protein on the cell surface of S. acidocaldarius at the translational level, glycosylated surface/membrane proteins were isolated from d-xylose- and l-arabinose-grown cells as well as from cells grown on N-Z-Amine by affinity chromatography on concanavalin A (ConA) Sepharose (Fig. 4).
Changes in the composition of glycosylated membrane proteins of S. acidocaldarius in response to the offered carbon source. NZ, N-Z-Amine; d-xyl, d-xylose; l-ara, l-arabinose.
SDS-PAGE analysis of the ConA elution fractions revealed several differences between the composition of surface proteins of MW001 cells grown on N-Z-Amine alone and that in the presence of d-xylose or l-arabinose. In the elution fraction of the d-xylose-grown cells, four protein bands were highly abundant, with molecular masses of ∼90 kDa (Fig. 4, band 1), ∼80 kDa (Fig. 4, band 2), ∼50 kDa (Fig. 4, band 3), and ∼46 kDa (Fig. 4, band 4). These protein bands were not visible in the N-Z-Amine-grown cells. Tandem mass spectrometry (MS/MS) analysis of these bands identified the ∼46-kDa protein under the d-xylose as well as the l-arabinose condition as the predicted sugar binding protein Saci_2122 (Fig. 4, bands 5 and 6, and Table 2). The ∼90-kDa and ∼50-kDa proteins also were detected in the arabinose fraction. The ∼90-kDa protein corresponded to Saci_1760, a predicted periplasmic solute-binding protein possibly belonging to an oligopeptide transport system, and the ∼50-kDa band was identified as Saci_2316, a predicted Fe-S-oxidoreductase and probably part of the aap pilus expressed by S. acidocaldarius (37). Additionally, within the 50-kDa band in the arabinose fraction, another predicted periplasmic solute-binding protein, Saci_1737, could be identified which is not located within an ABC transport operon (albeit with 4% coverage only). The 80-kDa protein (Fig. 4, band 2) was identified as predicted periplasmic solute-binding protein Saci_1668, which seemed to be present only under the d-xylose condition. This protein is not located in the genomic neighborhood of an ABC transport system.
MS analysis of glycosylated membrane proteins that were absent from N-Z-Amine but present under d-xylose and l-arabinose growth conditionsa
Deletion of predicted ABC transporter genes results in a growth defect on d-xylose and l-arabinose.To confirm the function of the transcriptionally strongly upregulated ABC transporter genes in vivo (i.e., saci_2120, saci_2121, and saci_2122), markerless deletions of each of the encoding genes were constructed. The mutant strains were tested for growth on N-Z-Amine alone or on N-Z-Amine with the addition of d-xylose, l-arabinose, or dextrin (0.1% [wt/vol] each) (see Fig. 5). All three ABC transporter mutants showed a significant growth defect compared to the wild type when grown on l-arabinose and d-xylose (Fig. 5A to D). No changes in growth were observed on N-Z-Amine alone, excluding a general growth defect of the mutants. Although impaired, the growth promotion caused by l-arabinose or d-xylose is not completely abolished in the transporter mutants, suggesting that the ABC transporter is important for d-xylose and l-arabinose uptake but does not represent the only transport system for pentoses. Additionally, growth phenotypes were analyzed on rich medium containing 0.2% (wt/vol) N-Z-Amine and 0.4% (wt/vol) d-xylose or minimal medium containing 0.4% d-xylose (see Fig. S1A and C in the supplemental material). The growth defect of the ABC transporter mutants observed in rich medium containing 0.1% (wt/vol) d-xylose could be complemented in rich medium with a higher xylose concentration (0.4% [wt/vol] d-xylose) (Fig. S1A). In contrast, the ABC transporter mutants were unable to grow on minimal medium containing 0.4% (wt/vol) d-xylose as the sole carbon and energy source (Fig. S1C). The transport system seems to be essential for d-xylose uptake under these conditions.
Growth of S. acidocaldarius MW001 wild-type (wt) and ΔABC-TMD, ΔABC-SBP, ΔABC-NBD, ΔαKGSADH, and ΔKDXD/KDAD gene deletion strains in the presence of different carbon sources. Cultures were grown on 0.1% (wt/vol) N-Z-Amine alone (black closed circles) or with the addition of 0.1% (wt/vol) l-arabinose (orange open squares), dextrin (green open triangles), or d-xylose (blue open circles). (A) Wild-type S. acidocaldarius MW001; (B) ΔABC-TMD, ABC type transmembrane domain (saci_2121) mutant; (C) ΔABC-SBP, ABC type substrate binding domain (saci_2122) mutant; (D) ΔABC-NBD, ABC type nucleotide binding domain (saci_2120) mutant; (E) ΔαKGSADH, α-ketoglutarate semialdehyde dehydrogenase (saci_1938) mutant; (F) ΔKDXD/KDAD, 2-keto-3-deoxyarabinoate/xylonate dehydratase (saci_1939) mutant.
Deletion of predicted KDXD/KDAD gene illustrates the importance of the aldolase-independent pentose degradation branch.Two additional genes, saci_1938 and saci_1939, were found to be highly upregulated in response to d-xylose (Table 2). Saci_1938 and Saci_1939 are close homologues of αKGSADH (SSO3117; 84% sequence identity) and KDXD (also known as KDAD) (SSO3118; 78% sequence identity), involved in the aldolase-independent pentose degradation pathway in S. solfataricus (9, 15). Deletion of saci_1939 (Δsaci_1939; ΔKDXD/KDAD) completely abolished growth of S. acidocaldarius MW001 on d-xylose minimal medium (Fig. S1D). A similar phenotype was also observed in medium containing l-arabinose and d-xylose in addition to N-Z-Amine, where the mutation nearly abolished the growth promotion observed in wild-type MW001 (Fig. 5F). The growth of Δd-KDXD/l-KDAD remained unchanged in N-Z-Amine alone and dextrin-containing medium, suggesting that growth defects are specific for l-arabinose and d-xylose. This suggests that pentose degradation mainly proceeds via the aldolase-independent pathway and cannot be replaced by the aldolase-dependent pathway in vivo. However, the deletion of saci_1938 (α-KGSADH) resulted in a strong but not complete growth inhibition on minimal medium and caused only a slight growth defect in the presence of N-Z-Amine and l-arabinose or d-xylose, indicating that alternative dehydrogenase enzymes substitute for KGSADH (Fig. 5E and Fig. S1B and D).
l-arabinose- and d-xylose-specific enzyme activities of crude extracts.To further analyze the regulation of the pentose degradation pathway on the protein level and the effect of the mutations, enzyme measurements were performed in crude extracts of the S. acidocaldarius wild type (MW001) and the ΔKGSADH, ΔKDXD/KDAD, and ΔABC-SBP mutant strains grown on different carbon sources. MW001 and all mutant strains showed similar sugar dehydrogenase activity (i.e., 500 to 900 mU mg−1) irrespective of the offered carbon source for growth (Fig. S2A to D). Additionally, the dehydrogenase activity toward the substrates d-xylose, l-arabinose, and d-glucose was almost identical in all strains. No activity was observed with d-arabinose as the substrate. Furthermore, all strains showed comparable sugar acid dehydratase activities (Fig. S3). Thus, in agreement with the transcriptomics data, the upper part common to the aldolase-dependent and -independent pentose degradation pathway also appears not to be regulated at the enzyme activity level in response to the offered carbon and energy sources. The ΔKGSADH (Δsaci_1938) and ΔKDXD/KDAD (Δsaci_1939) deletions also had no effect on the sugar acid dehydratase activities. Similarly, the KD(P)G aldolase measured in the cleavage direction with KDG as the substrate (Fig. 6A) appeared not to be regulated or, if at all, only slightly downregulated when pentoses were used as the carbon and energy source for growth. In regard to KDG aldolase activity, no differences were observed between wild-type MW001 and mutant strains (Fig. 6A). Importantly, complementation of the saci_2120 and saci_1939 deletion mutants showed that the deletion of the respective genes did not cause any polar downstream effects (Fig. S4).
Enzymatic activities of KDG aldolase, KDXD/KDAD, and αKGSADH in crude extracts of S. acidocaldarius wild-type (MW001) and ΔKDXD/KDAD, ΔKGSADH, and ΔABC-SBP mutant strains. The columns denote the mean enzyme activity of three biological replicates, and the error bars show standard deviations. αKGSADH, α-ketoglutarate semialdehyde dehydrogenase (Saci_1938); KDXD/KDAD, 2-keto-3-deoxyarabinoate/xylonate dehydratase (Saci_1939); ABC-SBP, ABC type substrate binding domain (Saci_2122).
In contrast, the enzyme activities of the lower part of the aldolase-independent route of the pentose degradation pathway were upregulated in response to pentoses. The KDXD/KDAD activity in the wild-type MW001 and ΔKGSADH mutant strains and the ΔABC-SBP strain was induced to 35 mU mg−1 when d-xylose or l-arabinose was used as the growth substrate (Fig. 6B). In the absence of these sugars, i.e., on N-Z-Amine alone and on N-Z-Amine plus dextrin as the carbon and energy source, only negligible KDXD/KDAD activity was observed. In the ΔKDXD/KDAD mutant, as expected, no KDXD/KDAD activity could be detected on the four different carbon and energy sources, confirming the absence of an alternative pathway/enzyme for KDX/KDA conversion under these conditions in S. acidocaldarius MW001.
The second enzyme activity of the aldolase-independent branch, the αKGSADH activity, was induced in wild-type and ΔABC-SBP mutant cells in the presence of pentoses (300 to 500 mU mg−1) but not with 100 mU mg−1 in N-Z-Amine- and N-Z-Amine-plus-dextrin-grown cells (Fig. 6C). In the ΔKGSADH mutant, basal αKGSADH activity was observed in the range of 50 mU mg−1, which was only slightly reduced compared to that of the uninduced state (in the absence of pentoses). This indicated the presence of an alternative dehydrogenase(s) catalyzing the KGSA oxidation. The ΔKDXD/KDAD mutation, however, unexpectedly also appeared to have an inhibitory effect on the αKGSADH activity, since only a very slight induction (if at all) was observed compared to the basal activity in the uninduced state on N-Z-Amine alone and N-Z-Amine plus dextrin as growth substrates.
Identification of the ara box in xylose-inducible promoters.Bioinformatic analyses of the three promoter regions of the highly upregulated genes saci_1938, saci_1939, and saci_2122 were performed. A motif search in the 80-bp upstream region of each transcription start site (TSS) using the Regulatory Sequence Analyzing Tools (RSAT) for prokaryotes (29) revealed a motif, AACATGTT, located −40 to −50 bp upstream of the TSS with a score of 1.42. This sequence was identical to the ara box motif AMCWWGTT (M = A or C; W = A or T) found in the promoter regions of d-arabinose-upregulated transcripts in S. solfataricus (9). A genome-wide promoter screening was performed in S. acidocaldarius, using the S. solfataricus ara box motif AMCWWGTT as a query sequence in which the regions −35 to −60 bp upstream of all transcription start sites were analyzed for the presence of this sequence motif. In addition to the three already-mentioned ara box-containing promoters, three other regions could be identified harboring that motif: the upstream regions of saci_1782, saci_2057, and saci_2295. saci_1782 encodes a predicted MFS transporter, saci_2057 encodes an alcohol dehydrogenase, and saci_2295 encodes a 3,4-dihydroxyphenylacetate-2,3-dioxygenase. However, these genes were not (saci_1782 and saci_2057) or were only slightly (saci_2295, <2-fold) upregulated in response to d-xylose in the RNA-seq data set. The ara box motifs upstream of these genes were slightly changed with respect to sequence or distance to the TATA box: positions 4 and 5 of the motif are inverted (saci_1782), A at position 4 is replaced by T (saci_2295), or the sequence remains unchanged but the distance to the TATA box is shortened by 1 bp (saci_2057) (highlighted in Fig. 7). These changes might influence the pentose-mediated induction or might be another layer of regulation.
Identification of a cis-acting element in d-xylose-inducible promoter regions. Shown is WEBLogo and promoter alignment of ara box-containing promoters in the genome of S. acidocaldarius. Sequences containing the ara box motif AMCWWGTT (M = A or C; W = A or T) in the region −40 to −50 bp upstream of the transcription start site were used to generate a sequence logo by WEBLogo (http://weblogo.berkeley.edu/) (30).
Promoter activity.To analyze the actual role of the ara box motif in vivo, promoter-reporter gene fusion assays were performed as described before (12). The β-galactosidase gene lacS of S. solfataricus (SSO3019), as the reporter gene, was cloned under the control of the promoters of saci_1938, saci_1939, and saci_2122. Additionally, the promoter region of saci_1938 was mutated with a purine-to-purine and pyrimidine-to-pyrimidine exchange in the 8-bp-long motif region, resulting in the Psaci_1938Δara-box mutation (depicted in Fig. 8). Cell extracts of transformants growing in the presence of l-arabinose, dextrin, or d-xylose were analyzed for their β-galactosidase activity. The three natural promoters were highly inducible in the presence of l-arabinose and d-xylose (Fig. 8). Both sugars led to comparable promoter activity levels for each single construct. The β-galactosidase activity was up to 12 times higher in the presence of l-arabinose or d-xylose than dextrin. The ara box-deficient promoter of saci_1938 was significantly less active than the natural promoter. The resulting β-galactosidase activity was reduced by a factor of 34 or 27 for l-arabinose or d-xylose induction, respectively.
β-Galactosidase (β-gal) activity of promoter-reporter fusions in cells grown in the presence of l-arabinose, dextrin, and d-xylose. The columns in panel A denote the mean enzyme activity of three biological replicates, and the error bars show the standard deviations. The dextrin condition represents the negative control. Sequences of the different promoters are shown in panel B.
DISCUSSION
The crenarchaeon S. acidocaldarius utilizes different carbohydrates, including pentoses such as d-xylose and l-arabinose (14, 15, 38). However, so far no transport system for d-xylose could be identified, and whether pentose assimilation strictly relies on both the aldolase-dependent and the aldolase-independent pathways is not yet fully understood. In this study, we compared growth on complex, proteinaceous media in the absence or presence of d-xylose. RNA sequencing of S. acidocaldarius MW001 revealed only five highly upregulated transcripts in response to d-xylose. Three of them encode an unknown and uncharacterized ABC transport system, and two represent enzymes of the previously reported aldolase-independent pathway.
The identified ABC transport system is not conserved within the Sulfolobales or found in any other archaea. The SBP Saci_2122 revealed a closer relationship to bacterial SBPs. Additionally, the NBD-encoding gene saci_2120 was previously not annotated. Therefore, the d-xylose/l-arabinose ABC transporter remained undiscovered for a long time. In this study, we identified the NBD-encoding gene by ORF prediction and confirmed the ABC transporter function in vivo.
All carbohydrate ABC transporters identified so far in Archaea belong either to the CUT1 (3.A.1.1) family or to the PepT (3.A.1.5) family, as classified by the Transporter Classification Database (TCDB; http://www.tcdb.org/) (1, 36, 39). However, the pentose ABC transporter consisting of Saci_2120 to Saci_2122, identified in S. acidocaldarius, belongs to the CUT2 (3.A.1.2) family. Transporters of the CUT2 family usually consist of one SBP, one or two TMD with 7 to 12 transmembrane segments each, and one motor protein, which represents a natural fusion of two NBDs (40, 41). In this fusion ATPase, only the N-terminal Walker A motif is active, whereas the C-terminal Walker A motif has no more ATP hydrolyzing activity due to the replacement of an invariant lysine residue within the Walker A motif (40). However, the CUT2 family d-xylose/l-arabinose ABC transporter of S. acidocaldarius harbors an ATPase consisting of only one NBD. Two other CUT2 family members with only one NBD have been identified in Streptomyces coelicolor and Sinorhizobium meliloti that transport xylose and fructose, respectively (42, 43).
In S. solfataricus P2, the best-studied crenarchaeal organism with regard to sugar transport and metabolism, no homologous system of the d-xylose/l-arabinose CUT2 ABC transporter (consisting of Saci_2120 to Saci_2122) could be identified. Instead, S. solfataricus uses a CUT1 family ABC system (SSO3066-3069) to transport the pentoses d-arabinose, l-arabinose, and d-xylose, which in turn is absent from S. acidocaldarius (4).
Interestingly, the ABC transport gene cluster consisting of saci_2120 to saci_2122 was not found to be differentially regulated in former microarray analyses using S. acidocaldarius grown on either d-glucose or d-xylose as the sole carbon source (14). The authors predicted Saci_0880 to Saci_0883 to be the transport system responsible for d-xylose transport. Transcriptional induction of Saci_0883 in response to d-xylose could be observed by RNA sequencing as well, but the transcriptional changes were only marginal compared to those of the ABC transporter consisting of saci_2120 to saci_2122. Saci_0880 to Saci_0883 is predicted to be an ammonium transporter, but its function has not yet been tested.
To functionally characterize the ABC transporter operon, Δsaci_2120, Δsaci_2121, and Δsaci_2122 single ABC transporter mutants were phenotypically analyzed on different carbon sources. The mutants grew identically to wild-type MW001 on dextrin, confirming that the mutations did not lead to a general growth defect. All three ABC transporter mutants revealed a distinct defect for growth on d-xylose minimal medium or on rich medium containing 0.1% d-xylose or l-arabinose. However, growth enhancement due to the presence of these pentoses in rich medium was not completely abolished, and the growth defect could be complemented by supplementation of higher pentose concentrations, indicating the presence of an additional transport system. We hypothesize that the newly identified CUT2 ABC transporter consisting of Saci_2120 to Saci_2122 is crucial for the uptake of pentoses from minimal medium or low-concentration carbon source medium. The observed stimulation of growth of the Δsaci_2121 and Δsaci_2122 deletion mutants in contrast to the Δsaci_2120 mutant at high xylose concentrations was unexpected and cannot be explained without further experiments (see Fig. S1A in the supplemental material).
The pentose assimilation pathway of Sulfolobus has been studied previously in detail (9, 15). It was demonstrated that d-xylose and l-arabinose are degraded via an aldolase-dependent route and an aldolase-independent route. Both pathways have the initial steps in common, i.e., sugar oxidation catalyzed by sugar dehydrogenase and sugar acid dehydration catalyzed by sugar acid dehydratase (Fig. 9). The absence of transcriptional or translational regulation of these genes by different sugars is in agreement with earlier studies (14) (Fig. S2A to C and S3A to C). These results are in agreement with former enzyme activity measurements in crude extracts of the close relative S. solfataricus grown on rich medium supplemented with either d-glucose or d-xylose (15) and with the broad substrate specificity of glucose dehydrogenase-1 (GDH1), which is promiscuous for d-glucose, d-galactose, d-xylose, and l-arabinose (46).
Model of d-xylose and l-arabinose transport and metabolism in S. acidocaldarius MW001. The ABC transporter consists of the SBD Saci_2122 XylF (purple), the TMD Saci_2121 XylH (orange), and the NBD Saci_2120 XylG (green) and transports d-xylose and l-arabinose into the cell. An additional unknown transporter can import pentoses in the presence of peptides. Within the cells, these pentose sugars were shown to be degraded via the aldolase-independent Weimberg pathway to α-ketoglutarate, which enters the citric acid cycle. Conversely, the aldolase-dependent Dahms pathway converting the pentoses to pyruvate and glycolaldehyde (entering the citric acid cycle via glyoxylate) is not utilized under the chosen growth conditions. Enzymes transcriptionally upregulated on d-xylose are in blue, and stars mark promiscuous enzymes, which function simultaneously in hexose and pentose metabolism. Abbreviations: GDH-1, glucose dehydrogenase (isoenzyme 1); d-XAD/l-AraD, xylonate/arabinoate dehydratase; d-KDXD/l-KDAD, 2-keto-3-deoxy-xylonate/arabinoate dehydratase; α-KGSADH, α-ketoglutarate semialdehyde dehydrogenase.
The KD(P)G aldolase of S. solfataricus, as the first enzyme of the aldolase-dependent branch, is promiscuous for the assimilation of hexose- and pentose-derived 2-keto-3-deoxy sugar acids (9, 15, 47). Crude extract measurements in S. solfataricus revealed no change of KD(P)G aldolase activity between d-xylose- and d-glucose-grown cells, whereas the downstream pentose-specific enzymes of the aldolase-dependent branch, glycolate reductase and malate synthase, were more active in d-xylose cells (15). In S. acidocaldarius MW001, the homologues for the proposed genes of the aldolase-dependent branch were not significantly regulated at the transcript level when grown on d-xylose. The S. acidocaldarius KD(P)G aldolase, which exhibits a substrate promiscuity similar to that of the S. solfataricus enzyme, was active in d-xylose- and l-arabinose-grown cultures, demonstrating that the key activity of the aldolase-dependent pathway is present.
In contrast to the noninduced aldolase-dependent branch, the genes saci_1938 and saci_1939, encoding the enzymes of the aldolase-independent pathway KDXD/KDAD and α-KGSADH, were significantly upregulated when cells were grown on d-xylose or l-arabinose, which is in accordance with the previous microarray results (14).
The significant growth defects of the ΔKDXD/KDAD deletion mutant strain in the presence of l-arabinose and d-xylose suggest that this enzyme and the whole aldolase-independent pathway are essential for S. acidocaldarius MW001 growth on these pentoses. However, S. acidocaldarius DSM639, which was used in the earlier studies (14), can use these sugars as a sole carbon source. This could explain the differences in the obtained results.
The growth of the ΔKGSADH strain on d-xylose and l-arabinose, as well as the αKGSADH activity observed in crude extracts, indicate the presence of alternative enzymes with αKGSADH activity in S. acidocaldarius. In the S. acidocaldarius genome, six additional aldehyde dehydrogenase homologues were identified, including Saci_1700, with an identity of 47% to αKGSADH. Saci_1738 shows 38% identity to αKGSADH but even higher identities (79%) to two aldehyde dehydrogenases of S. solfataricus previously characterized as succinic semialdehyde (SSA) dehydrogenases SSADH-1 (SSO1842) and SSADH-2 (SSO1629), which also were shown to catalyze the conversion of αKGSA to 2-oxoglutarate in addition to the canonical KGSADH (SSO3117) (48). Therefore, it is likely that these aldehyde dehydrogenase homologues account for the basal α-KGSA oxidizing activity which enables the S. acidocaldarius ΔKGSADH mutant to grow on d-xylose.
Unexpectedly, the αKGSADH activity was also reduced in the ΔKDXD/KDAD mutant strain. Both genes are located directly adjacent to each other in the S. acidocaldarius genome but are oriented in opposite directions and separated by 107 bp. However, the promoter assays using this region as the saci_1938 promoter region showed results similar to those of the other promoter constructs, pointing at a possible posttranslational regulation. Interestingly, KGSADH was previously found to be phosphorylated in a proteomic study (26).
The promoters of the five xylose-upregulated genes saci_2120 to saci_2122, saci_1938, and saci_1939 had the highly conserved ara box, 5′-AACATGTT-3′, as their upstream regulatory element, located −40 to −49 bp upstream of the TSS. Promoter activity could be abolished by exchanging the conserved sequence motif with an unspecific DNA sequence in the promoter of saci_1938, as was demonstrated by promoter-reporter gene assays. The same conserved sequence motif was found in the upstream regions of S. solfataricus genes (9), which were upregulated when cells were grown on d-arabinose. The genes encoding KDAD/KDXD (SSO3118), KGSADH (SSO3117), and the arabinose ABC transporter (SSO3066-3069) were among these upregulated transcripts. The genes encoding KDXD/KDAD and KGSADH show synteny among S. solfataricus and S. acidocaldarius, but the operons of the pentose ABC transporter share no synteny between the two Sulfolobus species. It is quite remarkable that the two Sulfolobus species share the same transcription regulation mechanism, even though inducible sugar transport systems are not alike.
In conclusion, we propose a model, as described in Fig. 9, for d-xylose and l-arabinose transport and metabolism in S. acidocaldarius. Pentoses are transported into the cell via the newly identified ABC transporter. Since pentose-dependent growth enhancement was not completely abolished in the ABC transporter mutants (Fig. 5), d-xylose and l-arabinose can be imported via other unknown transport systems. Imported d-xylose and l-arabinose are metabolized exclusively via the aldolase-independent Weimberg pathway. The aldolase-dependent Dahms pathway seems not to be active under these conditions in S. acidocaldarius MW001. The pentose CUT2-ABC transporter consisting of Saci_2120 to Saci_2122, the KDXD/KDAD Saci_1939, and the KGSADH Saci_1938 are transcriptionally upregulated in cells grown on pentoses to enable efficient conversion of d-xylose and l-arabinose to finally form α-ketoglutarate, which enters the citric acid cycle.
MATERIALS AND METHODS
Strains, growth conditions, and media.Sulfolobus acidocaldarius strain MW001 (18), the resulting gene deletion mutants, and expression plasmid-harboring strains were grown aerobically in Brock medium at a pH of 3 in a 76°C incubator. The medium was supplemented with 0.2% (wt/vol) N-Z-Amine or with 0.2% (wt/vol) N-Z-Amine and 0.4% (wt/vol) d-arabinose, 0.4% (wt/vol) l-arabinose, 0.4% (wt/vol) dextrin, 0.4% (wt/vol) fructose, 0.4% (wt/vol) glucose, 0.4% (wt/vol) sucrose, or 0.4% (wt/vol) d-xylose, unless specified otherwise. MW001 and the deletion mutants were supplemented with 10 μg/ml uracil. Growth of cells was monitored by measuring the optical density at 600 nm (OD600) or by measuring light scattering with a CGQ system (Aquila-Biolabs, Badesweiler, Germany).
Escherichia coli Top10 cells (Invitrogen, Carlsbad, CA) used for cloning were grown aerobically in LB medium supplemented with 50 μg/ml ampicillin when transformed with a plasmid at 37°C. E. coli ER1821 (New England BioLabs, Ipswich, MA) bearing the additional plasmid pM.EsaBC4I (New England BioLabs, Ipswich, MA) was used for methylation of the S. acidocaldarius expression plasmids. This propagation was done in LB medium containing 50 μg/ml ampicillin and 25 μg/ml kanamycin.
RNA isolation.S. acidocaldarius cells were grown on 0.2% (wt/vol) N-Z-Amine or 0.2% (wt/vol) N-Z-Amine supplemented with 0.4% (wt/vol) d-xylose and uracil as described above. Cells of three biological replicates for each condition were harvested at mid-exponential phase (OD600 of ≈0.5; 22.5 h). Additionally, three other biological replicates for each condition were further incubated, and growth was monitored until stationary phase/beginning of death phase (54.5 h) (Fig. 1). RNA was isolated for all replicates using TRIzol (ThermoFisher, Waltham, MA) as described earlier (19). The obtained RNA samples were treated with RNase-free DNase (Qiagen, Venlo, Netherlands) before the purification, and another DNase treatment was performed according to the RNeasy minikit cleanup protocol.
rRNA was depleted using a RiboZero magnetic kit for bacteria (Epicentre, Madison, WI) with a modified protocol. Only 90 μl, instead of 225 μl, of magnetic beads was used, and for the rRNA removal reaction only 1 μg RNA was mixed with 4 μl removal solution in a total volume of 20 μl instead of 40 μl.
RNA sequencing and data analysis.Sequencing libraries for all six samples were prepared with the TruSeq stranded mRNA HT kit (Illumina, San Diego, CA), starting with the RNA fragmentation step after elution of precipitated RNA in 19 μl of the Elute, Fragment, Prime Mix (Illumina). Sequencing libraries were quantified using a high-sensitivity chip in a Bioanalyzer (Agilent, Böblingen, Germany). Sequencing was performed with a HiSeq1500 instrument (Illumina) in rapid mode with a read length of 2 × 25 nucleotides (nt). Sequencing reads were mapped with Bowtie2 (20) against the reference genome.
Since S. acidocaldarius MW001 is a deletion mutant of S. acidocaldarius DSM639, the latter was chosen as a reference (S. acidocaldarius DSM639 genome size, 2,225,959 nt; RefSeq identifier NC_007181.1). Mapped reads were counted using the software ReadXplorer (21). Subsequently, RPKM (reads per kilobase of gene per million reads of experiment) values were manually calculated for each gene (22). The calculation of RPKM comprises a normalization regarding library size (i.e., one million reads). The following stochastical treatment of RPKM data then was performed with standard software (DESeq) (23). Only reads mapping to coding sequences were added together in order to get the total number of reads. For determination of regulated genes, statistical analysis was performed using DESeq.
Quantitative RT-PCR.S. acidocaldarius MW001 cells were grown in triplicate, as described above, in Brock medium supplemented with only 0.2% (wt/vol) N-Z-Amine or supplemented with 0.4% (wt/vol) sugar, i.e., d-arabinose, l-arabinose, dextrin, glucose, sucrose, or d-xylose. Cell cultures (14 ml) were harvested at mid-exponential phase (OD600, ≈0.5).
Total RNA isolation and cDNA synthesis were carried out as described previously (24). Quantitative RT-PCR was performed in a Rotor Gene Q (Qiagen, Venlo, Netherlands) using the 2xqPCRBIO SyGreen mix Lo-Rox (PCR Biosystems, London, United Kingdom) by following the manufacturer's instruction. Gene-specific primer sets (Table 3) were used to analyze the transcript levels of genes encoding the periplasmic binding protein Saci_2122. All quantification cycle (Cq) values for the efficiency of amplification were calculated for each primer set and were standardized to the Cq value of the housekeeping gene saci_0574 (secY) (25).
Primers used in this study
ConA purification of glycosylated membrane proteins of S. acidocaldarius.S. acidocaldarius strain MW001 was grown as described above and supplemented with 0.2% (wt/vol) N-Z-Amine alone or 0.2% (wt/vol) N-Z-Amine with 0.4% (wt/vol) l-arabinose or d-xylose in a volume of 400 ml. Cells were harvested at mid-exponential phase (OD600, ≈0.5) by centrifugation for 15 min at 4°C at 6,675 × g. The pellet was resuspended with 15 ml ConA buffer A (20 mM Tris-HCl, pH 7.5, 500 mM NaCl) containing 2 mM phenylmethylsulfonyl fluoride. Cells were disrupted by sonication with a Sonotrode KE-76 (20 min; 20 s on, 20 s off; amplitude, 50%). Cell debris and unbroken cells were removed by centrifugation for 20 min at 4°C at 2,000 × g. Membranes and S-layer were pelleted at 140,000 × g for 30 min at 4°C in an Optima Max-XP Ultracentrifuge (Beckman Coulter, Brea, CA). The membrane pellet was resuspended with 5 ml ConA buffer. ConA affinity purification of glycosylated proteins was further performed as described in reference 4. ConA elution fractions were loaded on an 11% SDS-PAGE gel and analyzed by mass spectrometry as described before (26).
Construction and characterization of deletion mutants.For markerless deletion of saci_1938, saci_1939, saci_2120, saci_2121, and saci_2122, the last ∼500 bp of the coding region of each gene was amplified by PCR using the primers described in the primer list (Table 3). The purified PCR products were cloned into MCSII of pSVA431 (18), resulting in plasmids pSVA553, pSVA555, pSVA549, pSVA503, and pSVA551. Upstream and downstream regions (∼500 bp) of all genes were amplified by PCR using primers as depicted in the primer list. With the purified PCR products an overlap extension PCR was performed and cloned into MCSI of pSVA431 derivatives pSVA553, pSVA555, pSVA549, pSVA503, and pSVA551, resulting in deletion plasmids pSV554, pSVA556, pSVA550, pSVA504, and pSVA552, respectively.
Linear markerless deletion fragments were obtained by PCR using the gene forward primer and the downstream reverse primer, in combination with the deletion plasmids pSV554, pSVA556, pSVA550, pSVA504, and pSVA552 as templates. One hundred nanograms of the purified PCR product was used for electroporation of 50 μl competent MW001 cells (theoretical OD600 of 10 in 20 mM sucrose) as described previously (18). Correct deletion was confirmed by sequencing with primers binding outside the cloning regions (Table 3).
For the construction of the complementation plasmids, PCR products were amplified using the primers described in Table 3 from genomic DNA of S. acidocaldarius. The products were purified and cloned into the NcoI/XhoI site of pSVAxylFX-H6, a pRN1-based E.coli-Sulfolobus shuttle vector (M. Wagner, unpublished data), yielding pSVA5328 (saci_1939) and pSVA5329 (saci_2120). As a negative control, pSVAxylFX-H6 was digested with NcoI and XhoI, treated with Klenow, and religated, yielding pSVA5351. All plasmids were first methylated in E. coli ER1821 and then transformed into the respective deletion mutants as described above.
Growth comparison of S. acidocaldarius MW001, ΔABC NBD (saci_2120), ΔABC TMD (saci_2121), ΔABC SBP (saci_2122), ΔKGSADH (saci_1938) and ΔKDXD/KDAD (saci_1939) strains.S. acidocaldarius parental strain MW001 and the ΔABC NBD (saci_2120), ΔABC TMD (saci_2121), ΔABC SBP (saci_2122), ΔKGSADH (saci_1938), and ΔKDXD/KDAD (saci_1939) deletion mutant precultures were grown in Brock medium containing 10 μg/ml uracil and supplemented with 0.2% (wt/vol) N-Z-Amine. From each non-sugar-adapted exponential-phase-grown preculture, the sugar-containing Brock media were inoculated with a starting OD600 of 0.01. The tested culture conditions were 0.1% (wt/vol) N-Z-Amine alone and 0.1% (wt/vol) N-Z-Amine together with 0.1% (wt/vol) l-arabinose, dextrin, or d-xylose. Growth was monitored by measuring the optical density at 600 nm until stationary phase. Additionally, cell growth was monitored in Brock medium containing 0.2% (wt/vol) N-Z-Amine together with 0.4% (wt/vol) sugar solution or in minimal Brock medium containing 0.4% d-xylose using CGQ online backscatter sensors (Aquila Biolabs, Badesweiler, Germany).
Bioinformatic analyses of xylose-regulated promoter sequences.To identify possible binding sites of the putative l-arabinose and d-xylose responsive regulator, bioinformatic analysis of the three promoter regions of saci_1938, saci_1939, and saci_2122 was performed. An 80-bp upstream region of each transcription start site was analyzed for possible motifs using Regulatory Sequence Analyzing Tools (RSAT) for prokaryotes (29). The resulting motif and variation of that motif were used for a genome-wide search of possible promoter sequences. Sequences of hits in the region −40 to −50 bp upstream of the next transcription start site were used to generate a sequence logo by WEBLogo (http://weblogo.berkeley.edu/) (30). Transcription start sites were obtained from reference 49.
Promoter-reporter gene assays.For the construction of the promoter activity plasmids, the different upstream regions of saci_1938, saci_1939, and saci_2122 were amplified by PCR with lengths of 106 bp, 106 bp, and 430 bp, respectively (related primers are listed in Table 3). Additionally, the promoter region of saci_1938 was mutated using the primers 6837 and 6838 (Table 3), resulting in a purine-for-purine and pyrimidine-for-pyrimidine exchange in the sequence of the ara box region. These four promoter regions were cloned into pSVA1450 (12) using SacII and NcoI to control transcription of the β-galactosidase gene lacS (sso3019) from S. solfataricus P2. All promoter activity constructs were confirmed by sequencing. The promoter activity constructs were methylated by propagation in E. coli ER1821 (New England BioLabs, Ipswich, MA) bearing the methylation plasmid pM.EsaBC4I (New England BioLabs, Ipswich, MA), as described previously (18). The methylated plasmids were transformed into electrocompetent S. acidocaldarius MW001 as described above. Obtained transformants were used in o-nitrophenyl-β-d-galactopyranoside assays as described in reference 31.
Crude extract measurement of S. acidocaldarius MW001, ΔABC-SBP (Δsaci_2122), ΔKGSADH (Δsaci_1938), and ΔKDXD/KDAD (Δsaci_1939) strains.S. acidocaldarius parental strain MW001 and the ΔABC-SBP (saci_2122), ΔKGSADH (saci_1938), and ΔKDXD/KDAD (saci_1939) deletion mutant strains were preinoculated in Brock medium containing 10 μg/ml uracil and 0.2% (wt/vol) N-Z-Amine. Cells then were inoculated with a starting OD600 of 0.01 in 50 ml Brock medium containing either 0.2% (wt/vol) N-Z-Amine alone or 0.2% (wt/vol) N-Z-Amine and 0.4% (wt/vol) of l-arabinose, dextrin, or d-xylose. Experiments were performed in triplicate. Cells were harvested in the mid-exponential growth phase (OD600 of ≈0.5) and stored at −70°C until use.
Cell pellets were resuspended in 0.1 M HEPES-NaOH (pH 7.3) to an OD600 of 25, and the cells were disrupted in a Precellys24 homogenizer three times at 6,500 rpm for 15 s with 1 min of cooling on ice after each run. Cell debris was removed by centrifugation (21,130 × g, 4°C for 30 min), and the protein content of the supernatant, referred to as crude extract, was determined in a modified Bradford assay (32) using the Bio-Rad protein assay kit with bovine serum albumin as a standard.
The sugar dehydrogenase activity in crude extracts was determined photometrically at 340 nm in a continuous enzyme assay at 70°C as sugar-dependent reduction of NADP+. The assay mixture (total volume of 500 μl) contained 0.1 M Tris-HCl (pH 6.5 at 70°C), 5 mM MgCl2, 5 mM NADP+, 10 mM d-glucose, d-xylose, l-arabinose, or d-arabinose, and 10 μg of protein. The assay mixture was preincubated at 70°C for 2 min, and the reaction was started by addition of the sugar substrate.
The sugar acid dehydratase activity in crude extracts was determined at 70°C as 2-keto-3-deoxy sugar acid formation from the respective sugars using NADP+ as a cosubstrate. The assay (total volume, 500 μl) was performed in 0.1 M Tris-HCl (pH 7.5 at 70°C), 5 mM MgCl2, 5 mM NADP+, 10 mM sugar (d-glucose, d-xylose, l-arabinose, or d-arabinose) with 100 μg of protein. The reaction mixture was preincubated at 70°C for 2 min, and the reaction was started by addition of the sugar substrate. Sample (100 μl) was removed at different time points, and the reaction was stopped by addition of 10 μl 12% (wt/vol) trichloroacetic acid. Formation of 2-keto-3-deoxy sugar acids was quantified using the thiobarbituric acid (TBA) assay as described before (33).
The KDXD/KDAD activity in crude extracts was assayed as αKGSA formation from d-KDX by coupling the reaction to the reduction of NADP+ using an excess amount of the purified recombinant αKGSADH (Saci_1938; L. Shen and B. Siebers, unpublished data). The continuous assay (total volume of 500 μl) was performed in 0.1 M Tris-HCl (pH 9.0 at 75°C), 5 mM NADP+, 1 mM d-KDX with 30 μg αKGSADH as auxiliary enzyme and 25 μg of protein. The assay mixture was preincubated at 75°C for 2 min, and the reaction was started by addition of d-KDX. Formation of NADPH was determined photometrically as the increase of absorbance at 340 nm. d-KDX was prepared enzymatically from d-xylonic acid (Carbosynth Ltd., United Kingdom) with the d-xylonate dehydratase (CCNA_00862) from Caulobacter crescentus (Shen and Siebers, unpublished).
αKGSADH activity in the crude extract was measured in 0.1 M Tris-HCl (pH 9.0 at 75°C), 5 mM NADP+, 10 μg crude extract protein, and 1 mM αKGSA at 75°C. After preincubation at 75°C for 2 min, the reaction was started by addition of αKGSA. Formation of NADPH was continuously monitored photometrically as the increase of absorbance at 340 nm. αKGSA was prepared enzymatically from d-xylonic acid (Carbosynth Ltd., United Kingdom) with the d-xylonate dehydratase (CCNA_00862) and 2-keto-3-deoxy-d-xylonate dehydratase (CCNA_00866) from Caulobacter crescentus (Shen and Siebers, unpublished).
The KDG aldolase activity in crude extracts was determined in a continuous assay by coupling the formation of pyruvate from d-KDG to the oxidation of NADH via l-lactate dehydrogenase from rabbit muscle (Sigma-Aldrich). The assay (total volume, 500 μl) was performed at 55°C in 50 mM sodium phosphate buffer (pH 6.0 at 55°C) containing 0.2 mM NADH, 2.5 mM d-KDG with 3 U l-lactic dehydrogenase as auxiliary enzyme and 200 μg of protein. The assay mixture was preheated to 55°C, and the reaction was started by the addition of d-KDG. NADH oxidation was monitored by the decrease of absorbance at 340 nm.
ACKNOWLEDGMENTS
M.W. was supported by the SFB987. S.V.A. received support from intramural funds of the Max Planck Society. L.S. was supported by the Mercator foundation with a Mercur startup grant (Pr-2013-0010) and by the Federal Ministry of Education and Research (BMBF) (e:Bio initiative, HotSysAPP, 03120078A). A.A. and J.K. acknowledge funding by the German Federal Ministry of Education and Research (BMBF; grants 0316188D and 031L0078C). N.V.D.K. was supported by the Federal Ministry of Education and Research (BMBF) (e:Bio initiative, HotSysApp, 0316188C).
Mass spectrometry analysis was performed by Veronica I. Dumit (Core Facility Proteomics, Center for Biological Systems Analysis [ZBSA], Freiburg University, Freiburg, Germany).
FOOTNOTES
- Received 9 June 2017.
- Accepted 12 November 2017.
- Accepted manuscript posted online 17 November 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/AEM.01273-17.
- Copyright © 2018 American Society for Microbiology.
